# RECENT PROGRESSES IN AMEBIASIS

EDITED BY : Anjan Debnath, Mario Alberto Rodriguez and Serge Ankri PUBLISHED IN : Frontiers in Cellular and Infection Microbiology

#### Frontiers Copyright Statement

© Copyright 2007-2019 Frontiers Media SA. All rights reserved. All content included on this site, such as text, graphics, logos, button icons, images, video/audio clips, downloads, data compilations and software, is the property of or is licensed to Frontiers Media SA ("Frontiers") or its licensees and/or subcontractors. The copyright in the text of individual articles is the property of their respective authors, subject to a license granted to Frontiers.

The compilation of articles constituting this e-book, wherever published, as well as the compilation of all other content on this site, is the exclusive property of Frontiers. For the conditions for downloading and copying of e-books from Frontiers' website, please see the Terms for Website Use. If purchasing Frontiers e-books from other websites or sources, the conditions of the website concerned apply.

Images and graphics not forming part of user-contributed materials may not be downloaded or copied without permission.

Individual articles may be downloaded and reproduced in accordance with the principles of the CC-BY licence subject to any copyright or other notices. They may not be re-sold as an e-book.

As author or other contributor you grant a CC-BY licence to others to reproduce your articles, including any graphics and third-party materials supplied by you, in accordance with the Conditions for Website Use and subject to any copyright notices which you include in connection with your articles and materials.

All copyright, and all rights therein, are protected by national and international copyright laws.

The above represents a summary only. For the full conditions see the Conditions for Authors and the Conditions for Website Use. ISSN 1664-8714 ISBN 978-2-88963-006-6 DOI 10.3389/978-2-88963-006-6

### About Frontiers

Frontiers is more than just an open-access publisher of scholarly articles: it is a pioneering approach to the world of academia, radically improving the way scholarly research is managed. The grand vision of Frontiers is a world where all people have an equal opportunity to seek, share and generate knowledge. Frontiers provides immediate and permanent online open access to all its publications, but this alone is not enough to realize our grand goals.

### Frontiers Journal Series

The Frontiers Journal Series is a multi-tier and interdisciplinary set of open-access, online journals, promising a paradigm shift from the current review, selection and dissemination processes in academic publishing. All Frontiers journals are driven by researchers for researchers; therefore, they constitute a service to the scholarly community. At the same time, the Frontiers Journal Series operates on a revolutionary invention, the tiered publishing system, initially addressing specific communities of scholars, and gradually climbing up to broader public understanding, thus serving the interests of the lay society, too.

### Dedication to Quality

Each Frontiers article is a landmark of the highest quality, thanks to genuinely collaborative interactions between authors and review editors, who include some of the world's best academicians. Research must be certified by peers before entering a stream of knowledge that may eventually reach the public - and shape society; therefore, Frontiers only applies the most rigorous and unbiased reviews.

Frontiers revolutionizes research publishing by freely delivering the most outstanding research, evaluated with no bias from both the academic and social point of view. By applying the most advanced information technologies, Frontiers is catapulting scholarly publishing into a new generation.

### What are Frontiers Research Topics?

Frontiers Research Topics are very popular trademarks of the Frontiers Journals Series: they are collections of at least ten articles, all centered on a particular subject. With their unique mix of varied contributions from Original Research to Review Articles, Frontiers Research Topics unify the most influential researchers, the latest key findings and historical advances in a hot research area! Find out more on how to host your own Frontiers Research Topic or contribute to one as an author by contacting the Frontiers Editorial Office: researchtopics@frontiersin.org

# RECENT PROGRESSES IN AMEBIASIS

Topic Editors:

Anjan Debnath, University of California, San Diego, United States Mario Alberto Rodriguez, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional (CINVESTAV-IPN), Mexico Serge Ankri, Technion, Israel

Figure 6 from Sierra-López et al.

Sierra-López F, Baylón-Pacheco L, Espíritu-Gordillo P, Lagunes-Guillén A, Chávez-Munguía B and Rosales-Encina JL (2018) Influence of Micropatterned Grill Lines on *Entamoeba histolytica* Trophozoites Morphology and Migration. *Front. Cell. Infect. Microbiol*. 8:295. doi: 10.3389/fcimb.2018.00295

Amebiasis, a parasitic disease transmitted by the unicellular protozoan parasite *Entamoeba histolytica*, is the cause of at least 100,000 deaths each year. The disease is mostly prevalent in developing countries and is one of the three common causes of death from parasitic diseases. The parasite has two stages in its life cycle in the host: the infective cyst and the invasive trophozoite. In the large intestine, the parasite feeds on bacteria and on cellular debris. No vaccine against amebiasis currently exists. Although metronidazole is the drug of choice for treating amebiasis, adverse effects in patients and potential resistance to metronidazole in other protozoa exist. About nine out of 10 people who are infected with *E. histolytica* are asymptomatic and in those individuals who develop symptoms, bloody diarrhea (amebic colitis) and liver abscess are the most common symptoms.

One possible explanation for this observation is the difference in the gut microbiota between individuals that may significantly influence the host's immune response in amebiasis and *E. histolytica*'s virulence. Amebiasis is characterized by acute inflammation of the intestine with release of pro-inflammatory cytokines, reactive oxygen species and reactive nitrogen species from activated cells of the host's immune system. In recent years, significant advances on the cell biology of *Entamoeba*  infection have been achieved through the development of new genetic tools to manipulate gene expression in the parasite and through the application of Omics tools.

In this Research Topic, we welcome high quality original research articles, as well as review, opinion or method articles, on amebiasis including but not limited to the regulation of gene expression, cell biology and signaling, adaptation and resistance to environmental stresses, metabolism, pathogenesis and immunity, pathogenesis and microbiome, drug discovery and drug resistance.

Citation: Debnath, A., Rodriguez, M. A., Ankri, S., eds. (2019). Recent Progresses in Amebiasis. Lausanne: Frontiers Media. doi: 10.3389/978-2-88963-006-6

# Table of Contents

*07 Editorial: Recent Progresses in Amebiasis* Anjan Debnath, Mario Alberto Rodriguez and Serge Ankri

# DRUG DISCOVERY

*11 Highly Potent 1*H*-1,2,3-Triazole-Tethered Isatin-Metronidazole Conjugates Against Anaerobic Foodborne, Waterborne, and Sexually-Transmitted Protozoal Parasites*

Sumit Kumar, Trpta Bains, Ashley Sae Won Kim, Christina Tam, Jong Kim, Luisa W. Cheng, Kirkwood M. Land, Anjan Debnath and Vipan Kumar

*20 Discovery of Antiamebic Compounds That Inhibit Cysteine Synthase From the Enteric Parasitic Protist* Entamoeba histolytica *by Screening of Microbial Secondary Metabolites*

Mihoko Mori, Satoshi Tsuge, Wataru Fukasawa, Ghulam Jeelani, Kumiko Nakada-Tsukui, Kenichi Nonaka, Atsuko Matsumoto, Satoshi Ōmura, Tomoyoshi Nozaki and Kazuro Shiomi

*29 High-Throughput Screening of* Entamoeba *Identifies Compounds Which Target Both Life Cycle Stages and Which are Effective Against Metronidazole Resistant Parasites*

Gretchen M. Ehrenkaufer, Susmitha Suresh, David Solow-Cordero and Upinder Singh

*39 Bioactivity of Farnesyltransferase Inhibitors Against* Entamoeba histolytica *and* Schistosoma mansoni

Alexandra Probst, Thi N. Nguyen, Nelly El-Sakkary, Danielle Skinner, Brian M. Suzuki, Frederick S. Buckner, Michael H. Gelb, Conor R. Caffrey and Anjan Debnath

*51 Flavonoids as a Natural Treatment Against* Entamoeba histolytica Moisés Martínez-Castillo, Judith Pacheco-Yepez, Nadia Flores-Huerta, Paula Guzmán-Téllez, Rosa A. Jarillo-Luna, Luz M. Cárdenas-Jaramillo, Rafael Campos-Rodríguez and Mineko Shibayama

# CELL BIOLOGY AND SIGNALING


Yunuen Avalos-Padilla, Roland L. Knorr, Rosario Javier-Reyna, Guillermina García-Rivera, Reinhard Lipowsky, Rumiana Dimova and Esther Orozco

*108 Phagocytosis of Gut Bacteria by* Entamoeba histolytica Lakshmi Rani Iyer, Anil Kumar Verma, Jaishree Paul and Alok Bhattacharya *117 A Flow Cytometry Method for Dissecting the Cell Differentiation Process of* Entamoeba *Encystation*

Fumika Mi-ichi, Yasunobu Miyake, Vo Kha Tam and Hiroki Yoshida


Elisa Azuara-Liceaga, Abigail Betanzos, Cesar S. Cardona-Felix, Elizabeth J. Castañeda-Ortiz, Helios Cárdenas, Rosa E. Cárdenas-Guerra, Guillermo Pastor-Palacios, Guillermina García-Rivera,

David Hernández-Álvarez, Carlos H. Trasviña-Arenas, Corina Diaz-Quezada, Esther Orozco and Luis G. Brieba

*176 A Calpain-Like Protein is Involved in the Execution Phase of Programmed Cell Death of* Entamoeba histolytica

Tania Domínguez-Fernández, Mario Alberto Rodríguez, Virginia Sánchez Monroy, Consuelo Gómez García, Olivia Medel and David Guillermo Pérez Ishiwara


# REGULATION OF GENE EXPRESSION


# *280 Postsplicing-Derived Full-Length Intron Circles in the Protozoan Parasite*  Entamoeba histolytica

María S. Mendoza-Figueroa, Eddy E. Alfonso-Maqueira, Cristina Vélez, Elisa I. Azuara-Liceaga, Selene Zárate, Nicolás Villegas-Sepúlveda, Odila Saucedo-Cárdenas and Jesús Valdés

*292 Telomeric Repeat-Binding Factor Homologs in* Entamoeba histolytica*: New Clues for Telomeric Research*

Francisco Javier Rendón-Gandarilla, Víctor Álvarez-Hernández, Elizabeth J. Castañeda-Ortiz, Helios Cárdenas-Hernández, Rosa Elena Cárdenas-Guerra, Jesús Valdés, Abigail Betanzos, Bibiana Chávez-Munguía, Anel Lagunes-Guillen, Esther Orozco, Lilia López-Canovas and Elisa Azuara-Liceaga

# PATHOGENESIS AND IMMUNITY


Enrique Gonzalez Rivas, Cecilia Ximenez, Miriam Enriqueta Nieves-Ramirez, Patricia Moran Silva, Oswaldo Partida-Rodríguez, Eric Hernandez Hernandez, Liliana Rojas Velázquez, Angelica Serrano Vázquez and Ulises Magaña Nuñez

*367* Entamoeba histolytica *Induce Signaling via Raf/MEK/ERK for Neutrophil Extracellular Trap (NET) Formation*

Zayda Fonseca, César Díaz-Godínez, Nancy Mora, Omar R. Alemán, Eileen Uribe-Querol, Julio C. Carrero and Carlos Rosales

*387* Entamoeba histolytica *Trophozoites Induce a Rapid Non-Classical NETosis Mechanism Independent of NOX2-Derived Reactive Oxygen Species and PAD4 Activity*

César Díaz-Godínez, Zayda Fonseca, Mario Néquiz, Juan P. Laclette, Carlos Rosales and Julio C. Carrero

*404 Reassessing the Role of* Entamoeba gingivalis *in Periodontitis* Mark Bonner, Manuel Fresno, Núria Gironès, Nancy Guillén and Julien Santi-Rocca

# Editorial: Recent Progresses in Amebiasis

#### Anjan Debnath<sup>1</sup> \*, Mario Alberto Rodriguez <sup>2</sup> \* and Serge Ankri <sup>3</sup> \*

<sup>1</sup> Center for Discovery and Innovation in Parasitic Diseases, Skaggs School of Pharmacy and Pharmaceutical Sciences, University of California, San Diego, La Jolla, CA, United States, <sup>2</sup> Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional (CINVESTAV-IPN), Mexico City, Mexico, <sup>3</sup> Department of Molecular Microbiology, Ruth and Bruce Rappaport Faculty of Medicine, Technion, Haifa, Israel

Keywords: Entamoeba histolytica, adaptation to stress, metabolism, pathogenesis, microbiome, drug discovery and resistance

### **Editorial on the Research Topic**

### **Recent Progresses in Amebiasis**

The aim of this research topic is to provide an overview of our current knowledge on amebiasis. This major neglected tropical disease which is transmitted by the unicellular protozoan parasite Entamoeba histolytica accounted for 55,500 deaths and 2.237 million disability-adjusted life years (i.e., the sum of years of life lost and years lived with disability) in 2010 (Turkeltaub et al., 2015). The parasite has two stages in its life cycle in the host: the infective cyst and the invasive trophozoite. About nine out of 10 people who are infected with E. histolytica are asymptomatic and in those individuals who develop symptoms, bloody diarrhea (amebic colitis), and liver abscess are the most common symptoms. Although the exact conditions, which trigger the onset of invasive disease, are still unknown, the interaction between the parasite's virulence factors and the host's response contribute to the development of disease (Huston and Petri, 1998). In recent years, significant advances on the cell biology of Entamoeba infection have been achieved through the development of new genetic tools to manipulate gene expression in the parasite and through the application of omics tools. Since the publication of the last book on amebiasis five years ago (Nozaki and Bhattacharya, 2015), more than one thousand publications on this subject have been published. This research topic (RT) which collates 29 articles by 167 authors presents the most recent development in this field through original articles and reviews which we introduce here briefly by classifying them according to four sections along the path from drug discovery, cell biology and signaling, regulation of gene expression, and pathogenesis and immunity.

# DRUG DISCOVERY

E. histolytica is capable of acquiring resistance to amebicidal concentrations of metronidazole (MTZ), the current drug of choice, under laboratory conditions, and this drug resistance has been associated with an increased expression of iron-containing superoxide dismutase and peroxiredoxin (Wassmann et al., 1999). Moreover, partial resistance to MTZ has also been described in some clinical strains of E. histolytica, suggesting the emergence of MTZ resistant strains (Bansal et al., 2004; Iyer et al., 2017). Thus, these observations have led to the search for new drugs with targets and modes of action distinct from those of MTZ [for a recent review see (Nagaraja and Ankri, 2019)]. In this RT, three studies based on drug screening have been published. The first one by the Kumar et al. has led to the identification of 1H-1,2,3-triazole-tethered isatin-metronidazole conjugates, which are more efficient against E. histolytica and Giardia lamblia than MTZ.

#### Edited and reviewed by:

Jeroen P. J. Saeij, University of California, Davis, United States

#### \*Correspondence:

Anjan Debnath adebnath@ucsd.edu Mario Alberto Rodriguez marodri@cinvestav.mx Serge Ankri sankri@technion.ac.il

#### Specialty section:

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

> Received: 15 June 2019 Accepted: 25 June 2019 Published: 09 July 2019

#### Citation:

Debnath A, Rodriguez MA and Ankri S (2019) Editorial: Recent Progresses in Amebiasis. Front. Cell. Infect. Microbiol. 9:247. doi: 10.3389/fcimb.2019.00247 The second one, from the Mori et al., targets the essential cysteine biosynthetic pathway of the parasite. The authors have described a fungal metabolite pencolide as the first compound that inhibits cysteine synthase and amebic cell growth in a cysteine-dependent manner with relatively low mammalian cytotoxicity. The third one from the Ehrenkaufer et al. identified anisomycin and prodigiosin as drugs able to kill mature cysts and MNZ resistant E. histolytica. Another study by the Probst et al. targets farnesyltransferase (FT), the last common enzyme for products derived from the mevalonate pathway. This enzyme is vital for diverse functions, including cell differentiation and growth. The authors found that a synergistic combination of metronidazole and an FT inhibitor lonafarnib offers a promising treatment strategy for amebiasis. Plants and their extracts are currently used to treat gastrointestinal diseases in many different parts of the world (Kelber et al., 2017). The Martínez -Castillo et al. summarizes our current knowledge on the antiamebic properties of flavonoids, a class of antioxidants compounds with variable phenolic structures which are found in many plants and vegetables.

# CELL BIOLOGY AND SIGNALING

E. histolytica acquires most of its nutrients by phagocytosis of bacteria present in the gut microbiota of the colon and by phagocytosis of host's cells. This essential event for the development of the parasite is directly involved in its pathogenesis. Actin is a fundamental component of the cytoskeleton which is responsible for changes in cell shape and other pivotal processes, including motility, phagocytosis of human cells, and parasite-substrate interactions. In this RT, the Manich et al. describes the 3D structural major divergences of E. histolytica actin compared to human actin. The authors also provide new information on the genesis of actin-enriched structures in E. histolytica, and on the role of the Arp2/3 actin-nucleation complex in the dynamics of the actin-rich cytoskeleton. Studying changes occurring in the parasite's cytoskeleton during migration may be challenging. A new method from Sierra-López et al. used glass or plastic surfaces covered with a substrate in a "micropatterned grill line." This method stimulated adhesion, migration, and an efficient formation of different membrane protrusions of E. histolytica trophozoites. On the phagocytosis side, the Avalos-Padilla et al. used an epigenetic silencing approach (Bracha et al., 2006) to demonstrate the essential role of Ehvps20 and Ehvps24, two proteins of the endosomal sorting complex required for transport, in erythrophagocytosis. The Rani Iyer et al. gives a glimpse of phagocytosis when E. histolytica is incubated with its favorite food, bacteria, isolated from human fecal material. The authors used a rRNA based metagenomic approach to point out that the parasite prefers to phagocytose a few bacterial species including Lactobacillus ruminus which was never shown to be associated with E. histolytica.

Two articles dealt with encystation of Entamoeba invadens. In the first one from the Mi-ichi et al., the authors used flow cytometry analysis to monitor cell differentiation. The second one from the Krishnan and Ghosh describes the formation of multinucleated giant cells which are formed during encystation by repeated cellular fusion with fusion-competent trophozoites. These observations indicate the possibility of Entamoeba undergoing sexual or parasexual reproduction.

During the colonization of the host, E. histolytica triggers an acute inflammatory process with the release of cytokines, reactive oxygen species (ROS), and nitric oxide (NO) from activated cells of the immune system. ROS and NO have been reported to trigger stress responses. A number of articles in this RT are dealing with stress responses of the parasite: the Nagaraja and Ankri provides a review of all the studies that used omics approaches to study the parasite's response to stresses. The Azuara-Liceaga et al. provides insight into the role of a DNA ligase involved in repairing DNA following oxidative DNA damage. ROS and NO cause programmed cell death (PCD) in E. histolytica. The Domínguez-Fernández et al. investigated the role of a calpain-like protein in triggering PCD in the parasite exposed to the aminoglycoside G418. Another work on PCD is provided by the Valle-Solis et al. The authors identified a protein of the CCX family (EhCCX) which is involved in PCD. Interestingly, the overexpression of EhCCX increased the in vitro virulence of trophozoites.

The outcome of two in silico genome-wide survey analyses performed by the Nakada-Tsukui et al. are described in this RT. The first one involved the identification of phosphatidylinositol (PI) kinases and PI phosphatases. Although it appears that E. histolytica possesses 10 PI kinases and 23 PI phosphatases, class II PI 3-kinases, type II PI 4-kinases, type III PI 5-phosphatases, and PI 4P-specific phosphatases are not represented in the parasite. There are several kinases and phosphatases that have the nuclear localization signal suggesting that PI metabolism also has conserved roles related to nuclear functions in E. histolytica, as it does in model organisms.

Regulated trafficking and secretion of pathogenic factors (Nakada-Tsukui et al., 2009; Somlata et al., 2017) and cyst wall proteins (Herman et al., 2017) have been extensively studied in this parasite. In a second review, the Das and Nozaki provides in silico data supporting the existence of a well-organized lipid transport in E. histolytica. It has been suggested that lipids play central roles in parasite growth, proliferation, differentiation, and virulence (Serrano-Luna et al., 2010).

# REGULATION OF GENE EXPRESSION

A number of articles focused on various aspects of RNA processing during the life cycle of E. histolytica. The Valdés-Flores et al. offers a review about our current knowledge on the molecular basis for splicing, 3′ end formation and mRNA degradation in ameba. Insights into the mechanism of splicing of AG-dependent and AG-independent transcripts and the post splicing of full length intron circles are provided by the Torres-Cifuentes et al. These events are involved in the regulation of gene expression of different regulatory processes including virulence traits in the parasite. Finally, the Valdés et al. also provided the first characterization of the RNA lariat debranching enzyme (Dbr1). This enzyme hydrolyzes the 2′ -5′

linkage in intron lariats, and modulates snRNP recycling during splicing reactions.

It is now well-documented that telomere length affects gene expression even for genes located far up to 1.2 Mb from the telomeres (Robin et al., 2014; Kim and Shay, 2018). In E. histolytica, our knowledge about the parasite's telomeres and their role in controlling gene expression is scanty. It is known that E. histolytica telomeres are non-conventional and they are formed by long tandem arrays that contain between 1 and 5 tRNA types per repeat unit and STRs which resemble microsatellites (Clark et al., 2006; Tawari et al., 2008). In the work presented by the Rendón-Gandarilla et al., the authors described the first three E. histolytica Telomeric Repeat Binding Factors. One of them, EhTRF-like III, forms specific DNA-protein complexes with telomeric related sequences.

# PATHOGENESIS AND IMMUNITY

The mechanisms leading to the interaction of E. histolytica with the intestinal epithelium of the host and its colonization have been recently reviewed (Cornick and Chadee, 2017). In a work from the Betanzos et al., the authors proposed that the EhADH adhesin protein altered tight junction of the host epithelium and that it may consequently make epithelial cells more susceptible to other E. histolytica effector proteins.

Immune reaction and immune evasion during amebiasis has been recently reviewed. It includes the suppression by the parasite of IFN-γ production, the elimination of immune cells and soluble immune mediators, and the neutralization of the cytotoxic effect of ROS and NOS produced during the inflammatory process (Nakada-Tsukui and Nozaki, 2016). In this RT, a number of amebic factors that regulate the host immune response are described. A new mechanism used by E. histolytica to prevent the triggering of inflammation during invasion of the gut via the production of a cyclooxygenase-like protein (EhCox) is described by the Shahi et al. EhCox alters the activity of E. histolytica cysteine proteases that are known to trigger inflammation (Hou et al., 2010). The López-Rosas et al. provided evidence that up-regulation of host microRNA-643 by E. histolytica promoted apoptosis of human epithelial colon cells. The Gonzalez Rivas et al., described the role of E. histolytica calreticulin (EhCRT) which acts as a mitogen. Indeed, EhCRT specifically activated a Th2 cytokine profile during the acute phase of liver abscess and a Th1 profile during the resolution phase of liver abscess. Two articles from the Fonseca et al.

# REFERENCES


and Díaz-Godínez et al. provided insights into the mechanism of induction of neutrophil extracellular traps (NETs) by E. histolytica. NETs are formed by DNA fibers decorated with histones and neutrophils constitute a first line of defense against invading pathogens. The authors showed that NETs formation is induced by Raf/MEK/ERK, extracellular calcium and serineprotease activity but does not depend on PKC, TAK1, ROS, NOX2-derived ROS, and PAD4 activity.

Finally, this RT includes a comprehensive review on the role of Entamoeba gingivalis in periodontitis by the Bonner et al. E. gingivalis shares a number of features with E. histolytica including the ability to phagocytose bacteria and human cells.

# CONCLUSIONS

Although considered a neglected disease, research on amebiasis continued unabated. Progress on research in amebiasis encompassed studying the life cycle stages and biology of E. histolytica, use of modern tools of genomics and metabolomics to analyze the responses of the parasite to different host stimuli and involvement of chemistry to develop new drug leads to control the parasite. Because of the lack of compilation of a comprehensive report on the progress that happened in the last 5 years in this field, our RT on the recent progress in amebiasis is timely. The RT articles covering different aspects of current amebiasis research may provide an important resource to the scientific community involved in the research of protozoan parasites. This may also serve as a guide for developing new areas of research in the field of amebiasis.

# AUTHOR CONTRIBUTIONS

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

# ACKNOWLEDGMENTS

We would like to thank the contributing authors and the reviewers for their helpful comments and suggestions that have helped us to achieve high standard for this Research Topic. AD was supported by the National Institutes of Health, Grant no. 1KL2TR001444 and UCSD Academic Senate Grant. SA was supported by the Israel Science Foundation (260/16), the Rappaport Institute, and the US-Israel Binational Science Foundation (2015211).

Clark, C. G., Ali, I. K., Zaki, M., Loftus, B. J., and Hall, N. (2006). Unique organisation of tRNA genes in Entamoeba histolytica. Mol. Biochem. Parasitol. 146, 24–29. doi: 10.1016/j.molbiopara.2005.10.013


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Debnath, Rodriguez and Ankri. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Highly Potent 1*H*-1,2,3-Triazole-Tethered Isatin-Metronidazole Conjugates Against Anaerobic Foodborne, Waterborne, and Sexually-Transmitted Protozoal Parasites

### *Edited by:*

Herbert Leonel de Matos Guedes, Universidade Federal do Rio de Janeiro, Brazil

#### *Reviewed by:*

Veeranoot Nissapatorn, Walailak University, Thailand Abeer Elhenawy, Mansoura University, Egypt Floriano Paes Silva-Jr, Fundação Oswaldo Cruz (Fiocruz), Brazil

### *\*Correspondence:*

Anjan Debnath adebnath@ucsd.edu Vipan Kumar vipan\_org@yahoo.com

#### *Specialty section:*

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

*Received:* 15 June 2018 *Accepted:* 09 October 2018 *Published:* 30 October 2018

#### *Citation:*

Kumar S, Bains T, Won Kim AS, Tam C, Kim J, Cheng LW, Land KM, Debnath A and Kumar V (2018) Highly Potent 1H-1,2,3-Triazole-Tethered Isatin-Metronidazole Conjugates Against Anaerobic Foodborne, Waterborne, and Sexually-Transmitted Protozoal Parasites. Front. Cell. Infect. Microbiol. 8:380. doi: 10.3389/fcimb.2018.00380 Sumit Kumar <sup>1</sup> , Trpta Bains <sup>2</sup> , Ashley Sae Won Kim<sup>3</sup> , Christina Tam<sup>4</sup> , Jong Kim<sup>4</sup> , Luisa W. Cheng<sup>4</sup> , Kirkwood M. Land<sup>3</sup> , Anjan Debnath<sup>2</sup> \* and Vipan Kumar <sup>1</sup> \*

<sup>1</sup> Department of Chemistry, Guru Nanak Dev University, Amritsar, India, <sup>2</sup> Center for Discovery and Innovation in Parasitic Diseases, Skaggs School of Pharmacy and Pharmaceutical Sciences, University of California, San Diego, La Jolla, CA, United States, <sup>3</sup> Department of Biological Sciences, University of the Pacific, Stockton, CA, United States, <sup>4</sup> Foodborne Toxin Detection and Prevention Research Unit, Agricultural Research Service, United States Department of Agriculture, Albany, CA, United States

Parasitic infections like amebiasis, trichomoniasis, and giardiasis are major health threats in tropical and subtropical regions of the world. Metronidazole (MTZ) is the current drug of choice for amebiasis, giardiasis, and trichomoniasis but it has several adverse effects and potential resistance is a concern. In order to develop alternative antimicrobials, a library of 1H-1,2,3-triazole-tethered metronidazole-isatin conjugates was synthesized using Huisgen's azide-alkyne cycloaddition reaction and evaluated for their amebicidal, anti-trichomonal, and anti-giardial potential. Most of the synthesized conjugates exhibited activities against Trichomonas vaginalis, Tritrichomonas foetus, Entamoeba histolytica, and Giardia lamblia. While activities against T. vaginalis and T. foetus were comparable to that of the standard drug MTZ, better activities were observed against E. histolytica and G. lamblia. Conjugates 9d and 10a were found to be 2–3-folds more potent than MTZ against E. histolytica and 8–16-folds more potent than MTZ against G. lamblia. Further analysis of these compounds on fungi and bacteria did not show inhibitory activity, demonstrating their specific anti-protozoal properties.

Keywords: *Entamoeba histolytica*, *Trichomonas vaginalis*, *Tritrichomonas foetus*, *Giardia lamblia*, metronidazole, cytotoxicity, isatin-metronidazole conjugates

# INTRODUCTION

Anaerobic protozoan parasites Trichomonas vaginalis and Tritrichomonas foetus are the major causes of reproductive tract infections viz. trichomoniasis and bovine trichomoniasis (Kumar et al., 2015; Ravaee et al., 2015). Though men remain asymptomatic, trichomoniasis leads to urogenital infection in women via sexual transmission, resulting infertility, urethritis, vaginitis, preterm delivery and low birth weight, Bovine trichomoniasis, acquired by direct sexual intercourse, is asymptomatic in nature but earlier death of developing foetus was observed in some cases. Multilocus genotyping confirmed that the isolates obtained from cattle and pig represented the "bovine genotype" of T. foetus and the cat isolates represented a closely related "feline genotype" of T. foetus (Fang et al., 2015). The prevalence of T. vaginalis infection in the United States is estimated to be 2.3 million among women of 14–49 years and this increases with age (Sutton et al., 2007; Conrad et al., 2013). Moreover, women without any past history of sexual intercourse can still be affected with trichomoniasis (Kumar et al., 2012). Major health threats associated with trichomoniasis include transmission of HIV-1 (Fichorova, 2009), benign prostatic hyperplasia (Mitteregger et al., 2012), prostate cancer and pelvic inflammatory disease (Stark et al., 2009).

Two other anaerobic intestinal protozoans, Giardia lamblia and Entamoeba histolytica cause the gastrointestinal diarrheal diseases (Halliez and Buret, 2013), giardiasis and amebiasis. Acute gastroenteritis is considered as one of the leading causes of illnesses and deaths in children under the age of 5 years (Heresi et al., 2000; Youssef et al., 2000). Infection occurs through the ingestion of cysts in contaminated water or food and direct person-to-person contact. Upon ingestion of G. lamblia or E. histolytica cysts, trophozoitese merge from the cysts and multiply in the lumen of the small intestine, where G. lamblia attach to the intestinal mucosa (Navaneethan and Giannella, 2008). E. histolytica trophozoites invade the colon and causes amoebic colitis. While 50% of G. lamblia infection is asymptomatic, major symptoms of amebiasis and giardiasis include weight loss, loss of appetite, watery or bloody diarrhea, dehydration, bloating and abdominal cramps, cognitive impairment in children, and chronic fatigue in adults (Berkman et al., 2002; Hanevik et al., 2007).

Nitro-imidazoles such as ornidazole (OZ), benznidazole (BZ), and secnidazole (SZ) (**Figure 1**) are the widely used medicament to treat anaerobic infections but have lower efficacies than MTZ (Nash, 2002). MTZ, an effective synthetic drug introduced in 1960, showed strong inhibitory efficacies against Gram-negative anaerobic bacteria like Helicobacter pylori and protozoans such as G. lamblia and E. histolytica. It is the only therapeutic drug available till date against trichomoniasis (Cudmore et al., 2004; Sutherland et al., 2010). These outstanding achievements have encouraged researchers to focus on the development of nitro-imidazoles with imminent medicinal application. However,

MTZ resistance has been observed in E. histolytica, G. lamblia, and T. vaginalis and thus the development of novel, noncytotoxic and efficient scaffolds against amebiasis, giardiasis, trichomoniasis, and bacterial infections is desirable (Meri et al., 2000; Upcroft et al., 2005; Debnath et al., 2013).

Isatin is one of the important pharmacophores with wide application in drug discovery and it is the core constituent of many alkaloids, dyes, pesticides, and analytical reagents. Various studies from the literature show the anti-bacterial (Sarangapani and Reddy, 1994), anti-fungal (Pandeya et al., 1999), anti-inflammatory (Bhattacharya and Chakrabarti, 1998) and anticonvulsant (Popp et al., 1980) properties of isatinschiff bases. Thiosemicarbazones are the class of Schiff bases that exhibit anti-parastic and anti-bacterial properties (Chellan et al., 2012). Moreover, the spiro compounds of isatin are also known to exhibit versatile biological properties. Earlier results based on 1H-1,2,3-triazole and β-amino-alcohol tethered isatin-β-lactam conjugates showed efficient inhibitory activities against T. vaginalis (Nisha et al., 2013). Results obtained with N-propargylated-isatin-Mannich adducts and N-propargylated isatin-quinoline Mannich adducts against T. foetus have also prompted the use of isatin as a pharmacophore (Nisha et al., 2014). Based on the anticancer (Vine et al., 2007; Kumar et al., 2018; Singh et al., 2018), anticonvulsant (Verma et al., 2004), antidepressant (Singh et al., 1997), anti-HIV (Bal et al., 2005), and anti-bacterial (Pandeya et al., 2000) properties of isatin, the present study explicates the synthesis and biological evaluation of 1H-1,2,3-triazole-tethered isatin-metronidazole conjugates as shown in **Figure 2**. Substitutions at C-5 position of isatin-ring as well as the introduction of thiosemicarbazide and ketocarbonyl at C-3 carbonyl of isatin have been carried out to ascertain the structure-activity relationship of the synthesized conjugates.

# EXPERIMENTAL SECTION

# General Information

Melting points were determined with open capillaries using a Veego Precision Digital Melting Point apparatus (MP-D) and are uncorrected. <sup>1</sup>H and <sup>13</sup>C NMR were recorded on JEOL 400, Bruker 500 MHz spectrometer, respectively using CDCl<sup>3</sup> as solvent. Chemical shifts were reported in parts per million (ppm) using tetramethylsilane (TMS) as an internal standard and coupling constants J indicated in Hertz. Splitting patterns were indicated as s: singlet, d: doublet, t: triplet, m: multiplet, dd: double doublet, ddd: doublet of a doublet of a doublet, and br: broad peak. Mass spectrometric analysis was carried out on BrukermicrOTOF QII equipment using ESI as the source. Column chromatography was performed on a silica gel (60–120 mesh) using an ethyl acetate: hexane mixture as eluent (Kumar et al., 2017).

# General Procedure for the Synthesis of C-5 Substituted Isatin-Spiro-Ketals (2)

A solution of isatin **1** (1 mmol) in dry DMF was added dropwise to a stirred suspension of NaH (1.2 mmol) in dry DMF at 0 ◦C. The solution was then stirred for 15 min. followed by the drop-wise addition of 2-bromoethanol (1.2 mmol). The reaction

mixture was allowed to stir at room temperature for 20 min. with subsequent heating at 80◦C for 4 h. The progress was monitored by TLC. Upon completion, the reaction mixture was extracted with ethyl acetate (2 × 25 mL), washed with brine water (2 × 20 mL) and the combined organic layers were dried over anhydrous Na2SO4. The organic layer was concentrated under reduced pressure to afford the desired precursors.

# General Procedure for the Synthesis of *N*-Propargylated C-5 Substituted Isatin (3)/Isatin Spiroketal (4)

To a stirred suspension of NaH (1 mmol) in dry DMF was added drop-wise a solution of C-5 substituted isatin (1 mmol) **3** /isatinspiroketal **4** in dry DMF at 0◦C. The mixture was stirred for 15 min followed by the drop-wise addition of propargyl bromide (1.2 mmol). The reaction mixture was allowed to stir at room temperature for 10 min. with subsequent heating at 60◦C for 4 h with progress being monitored by TLC. Upon completion, the reaction mixture was extracted with ethyl acetate (2 × 25 mL) and washed with brine water (2 × 20 mL). The combined organic layers were dried over anhydrous Na2SO4, concentrated under vacuum to afford the desired precursors.

# General Procedure for the Synthesis of Conjugates, 8a–e and 9a–e

CuSO4.5H2O (0.055 mmol) and sodium ascorbate (0.143 mmol) were added to a well stirred solution of 5-substituted 1-(prop-2 yn-1-yl)indoline-2,3-dione **3** (1 mmol) or 5-substituted 1-(prop-2-yn-1-yl)spiro[indoline-3,2/ -[1,3]dioxolan]-2-one **4** and 1-(2 azidoethyl)-2-methyl-5-nitro-1H-imidazole **7** (1 mmol) in an ethanol:water (85: 15) mixture. The reaction mixture was allowed to stir at room temperature for 6–7 h and the progress was monitored by TLC. Upon completion, the reaction mixture was extracted with ethyl acetate and water and the combined organic layers were dried over anhydrous Na2SO<sup>4</sup> and concentrated under vacuum to afford the desired conjugates which were purified via column chromatography using an ethyl acetate: hexane (70:30) mixture.

# General Procedure for the Synthesis of Conjugate, 10a–e

Thiosemicarbazide (1 mmol) and glacial acetic acid in catalytic amount were added to a stirred solution of 5-substituted1-((1-(2-(2-methyl-5-nitro-1H-imidazol-1 yl)ethyl)-1H-1,2,3-triazol-4-yl)methyl)indoline-2,3-dione (**8**) in ethanol. The reaction mixture was heated to reflux for

2 h and the progress was monitored using TLC. On completion, yellow colored solid resulted and this was re-crystallized using an absolute ethanol to afford the desired conjugates.

# MATERIALS AND METHODS

# Biological Evaluation

# In vitro Susceptibility Assay Against the Bovine Trichomonad T. foetus Strain D1 and the Feline Trichomonad T. foetus-Like Strain C1

To perform the initial susceptibility screens on T. foetus D1 (from Lynette Corbeil, University of California at San Diego, School of Medicine, San Diego, CA, USA) and T. foetus-like C1 (from Stanley Marks, University of California at Davis, School of Veterinary Medicine, Davis, CA, USA), compounds were dissolved in 100% DMSO to obtain concentrations of 100 mM. Stock solutions were kept at −20◦C. 5 microliter aliquots of these solutions were diluted in 5 mL of TYM Diamond's media (Hardy Diagnostics, Santa Maria, CA, USA) to obtain a final concentration of 100µM. Metronidazole-sensitive D1 and C1 strains were cultured under anaerobic conditions. After 24 h, cells were counted using a hemacytometer. The IC<sup>50</sup> value for the series of compounds was determined by inoculating a constant number of parasite cells in TYM medium and running assays of increasing drug concentrations, 0.02–100µM, and performing a regression analysis on percentage growth inhibition relative to DMSO control, using Prism software, from GraphPad. Predicted IC<sup>50</sup> values of compounds were then confirmed by testing again using the same assay described above. The sample size consisted of four independent trials carried out on four different days (to account for possible variation in the parasite population).

# In vitro Susceptibility Assay Against the Human Trichomonad T. vaginalis Strain G3

T. vaginalis G3 trophozoites (from Patricia Johnson, University of California at Los Angeles, CA, USA) were maintained in TYM media (pH 6.2) at 37◦C for 24 h and every 24 h, 1000 <sup>µ</sup>L of cells were passed into 10 ml of TYM media to maintain the culture. To perform the initial susceptibility screens on metronidazolesensitive T. vaginalis G3, compounds were dissolved in DMSO to obtain concentrations of 100 mM; 5 µL aliquots of these solutions were diluted in 5 mL of TYM medium to obtain a final concentration of 100µM. After 24 h, cells were counted using a hemacytometer. The assays were performed in 15 mL culture tubes with T. vaginalis G3 strain. 0.1% DMSO-only treated parasites served as control to normalize the effects of the solvent and in vitro conditions. TYM media only control was also included in the assays. After 24 h, cells were counted using a hemacytometer. The IC<sup>50</sup> values were determined by inoculating a constant number of parasite cells in TYM medium and running assays of increasing compound concentrations, 0.02–100µM, and performing a regression analysis on percentage growth inhibition relative to DMSO control, using Prism software from GraphPad. Predicted IC<sup>50</sup> values of compounds were then confirmed by testing again using the same assay described above. The sample size consisted of four independent trials carried out on four different days (to account for possible variation in the parasite population).

# G. lamblia and E. histolytica IC<sup>50</sup> Assays

Axenic trophozoites of metronidazole-sensitive G. lamblia WB and E. histolytica HM1:IMSS were grown in TYI-S-33 medium supplemented with penicillin (100 U/mL) and streptomycin (100µg/mL). (Diamod et al., 1978; Keister, 1983) For anti-E. histolytica and anti-G. lamblia IC50assays, 10 mM stocks of the test compounds were serially diluted in DMSO to achieve a concentration range of 10 mM−78µM. 0.5 microliter of compound from this concentration range was transferred in triplicate to each well of 96-well plates and 5,000 trophozoites/well were added in a final volume of 100 <sup>µ</sup>L/well. Cultures were grown for 2 days at 37◦C under anaerobic conditions (GasPak EZ Anaerobe Gas Generating Pouch System (VWR). Cell growth and viability were determined by adding Cell Titer-Glo cell viability assay reagent (Promega) and measuring ATP-dependent luminescence in a microplate reader. Percent inhibition relative to maximum and minimum reference signal controls was calculated using the formula:

% Inhibition = [(mean of Maximum Signal Reference Control—Experimental Value)/(mean of Maximum Signal Reference Control—mean of Minimum Signal Reference Control)] × 100.

The 50% inhibitory concentration (IC50) and standard error (SE) was derived from the concentration-response curves using Prism software (GraphPad).

# Anti-bacterial and Anti-fungal Susceptibility Analyses

To determine if the compound library had other antimicrobial activities, several fungal and bacterial species were analyzed for susceptibility to these compounds. Antifungal activity of compounds was tested in the filamentous fungal pathogen Aspergillus parasiticus 5862 (National Center for Agricultural Utilization and Research, USDA-ARS, Peoria, IL, USA) and the model yeast Saccharomyces cerevisiae BY4741 wild type (Mat a his311 leu210met1510ura310) (Open Biosystems, Huntsville, AL, USA). A. parasiticus was cultured at 35◦C on potato dextrose agar (PDA), and S. cerevisiae was grown on Synthetic Glucose (SG; Yeast nitrogen base without amino acids 0.67%, glucose 2% with appropriate supplements: uracil 0.02 mg/mL, amino acids 0.03 mg/mL) or Yeast Peptone Dextrose (YPD; Bacto yeast extract 1%, Bacto peptone 2%, glucose 2%) medium at 30◦C. All chemicals for culturing fungi were procured from Sigma Co. (St. Louis, MO, USA).

To evaluate antifungal activity of compounds in A. parasiticus, bioassays were performed in microtiter plates (triplicate wells) (3 × 10<sup>4</sup> to 5 × 10<sup>4</sup> CFU/mL) with RPMI 1640 medium (Sigma Co., St. Louis, MO, USA). Compounds were tested at 500µM, where fungal growth was monitored at 24–48 h after inoculation. A numerical score from 0 to 4 was provided to each well according to the protocol outlined by the Clinical and Laboratory Standards Institute (CLSI) M38-A2 (CLSI, 2008) as follows: 0 = optically clear/no visible growth, 1 = slight growth (25% of no treatment control), 2 = prominent growth reduction (50% of no treatment control), 3 = slight growth reduction (75% of no treatment control), 4 = no growth reduction. To test antifungal activity of compounds in S. cerevisiae, bioassays were performed in SG liquid medium (triplicate wells in microtiter plates). Compounds were examined at 500µM, where antifungal activity was assessed 24–48 h after inoculation. A numerical score from 0 to 4 (See above) was provided to each well by the modified protocols outlined by European Committee on Antimicrobial Susceptibility Testing (EUCAST) (Keister, 1983).

For anti-bacterial susceptibility testing, disc diffusion methods were used. Vehicle control (DMSO) and 100 mM stock compounds were diluted to 100µM in media and incubated with

empty BDL-sensi-discs (6 mm) for 20 min at room temperature. These discs were placed upon plates streaked either with Lactobacillus reuteri (ATCC 23272), Lactobacillus acidophilus (ATCC 43560), Lactobacillus rhamnosus (ATCC 53103), Listeria monocytogenes 10403 (RM2194), Salmonella enterica pGFP, or Escherichia coli K-12 MG 1655. Additionally, various antibiotic discs [levofloxacin (5 µg), gentamicin (10 µg), and gentamicin (120 µg)] were placed upon these plates as controls for sensitivity. Plates were streaked for growth of Lactobacilli, grown in Lactobacilli MRS at 37◦C under anaerobic conditions whereas the rest of the strains was grown at 37◦C aerobically in Luria Broth or Brain Heart Infusion Broth. Zones of inhibition measured in millimeters were measured for each disc.

# RESULTS AND DISCUSSION

# Synthetic Chemistry

The methodology for the synthesis of isatin-spiroketal **2** involved sodium hydride promoted reaction of C-5 substituted isatin **1** with 1-bromoethanol (Sigma-Aldrich, Cat No. 48874, CAS No. 540-51-2) in dry DMF. N-propargylated C-5 substituted isatin and isatin-spiroketal were obtained via base mediated reaction of isatin/spiroisatin with propargyl bromide (Sigma-Aldrich, 80% weight in toluene, Cat No. P51001, CAS No. 106-96-7) in dry DMF at 60◦C (**Scheme 1**). The precursor viz. 1-(2-azidoethyl)-2-methyl-5-nitro-1-H-imidazole **7**, was synthesized via initial mesylation of metronidazole 5 (Sigma, St. Louis, MO, USA, Cat. No. M3761) in dry THF at 0◦C to form the corresponding methane-sulfonic acid 2-(2-methyl-5 nitro-imidazol-1-yl)-ethyl ester 6 followed by its nucleophilic substitution reaction with sodium azide in dry DMF at 60◦C to afford the corresponding precursor 1-(2-Azido-ethyl)-2 methyl-5-nitro-1H-imidazole **7** in good yields (**Scheme 2**).Cu promoted azide-alkyne cycloaddition reaction of **7** with Npropargylated-isatin **3** and N-propargylatedspiro-isatin 4 led to the isolation of desired conjugates **8** and **9** (**Scheme 3**). Conjugate **8** was further treated with thiosemicarbazide to obtain metronidazole-isatin-thiosemicarbazone conjugates 10. On the bases of spectral data and analytical evidence, structures were assigned to the synthesized 1H-1,2,3-triazole-tethered


TABLE 1 | In vitro activity of isatin-metronidazole conjugates against T. vaginalis, T. foetus, T. foetus-like pathogen, G. lamblia, and E. histolytica.

N.D., not determined.

isatin-metronidazole conjugates (see **Supplementary Material**). The compound 8c, for example was characterized as 5-chloro-1-{1-[2-(2-methyl-5-nitro-imidazol-1-yl)-ethyl]-

1H-[1,2,3]triazol-4-ylmethyl}-1H-indole-2,3-dione analyzed for C17H14ClN7O<sup>4</sup> and showed molecular ion peak at m/z 416.4302 ([M+H]+) and 417.4318 ([M+2]+) in its mass spectrum. The salient feature of its (Ravaee et al., 2015) H NMR spectra include charterstick peak at δ 4.60 (t, J = 4.8 Hz, 2H); δ 4.75 (t, J = 5.8 Hz, 2H) and one singlet at 4.90 (2H) ppm corresponding to methylene group, respectively. Two singlets also appeared at δ 7.93 and 7.97 which correspond to triazole and imidazole proton (Navaneethan and Giannella, 2008). C NMR spectrum of compound **8c** exhibited the appearance of characteristic peaks at δ 182.40 and 157.90 which correspond to isatin carbonyl (C=O) and amidic carbonyl of isatin (=N-C=O). Three methylenic carbons show characteristic peaks at δ 35.3, 46.5 and 49.3 along with one single peak at 13.3 corresponding to—CH<sup>3</sup> group of imidazole ring.

# *In vitro* Evaluation Against *T. vaginalis*, *T. foetus*, *T. foetus*-Like, *G. lamblia*, *E. histolytica*, Fungal, and Bacterial Species

The chemical library of metronidazole-isatin conjugates was evaluated in a general inhibitory screen against protozoal pathogen T. vaginalis and the percentage inhibition results at different concentrations are enlisted in **Table 1**. As evident most of the conjugates exhibited 100% growth inhibition at 100µM except **8a, 8d**, and **10e** (**Table 1**). The potent conjugates were further evaluated for their IC<sup>50</sup> against different strains of T. vaginalis and T. foetus and compared with metronidazole, (**Table 1**). A closer inspection of **Table 1** revealed an interesting structure activity relationship with activity being dependent upon the nature of substituent at C-3 and C-5 positions of isatin ring. In case of MTZ susceptible G-3 strain, introduction of a ketal and thiosemicarbazone substituent at C-3 position improved the activity profiles as evident from conjugates 9a–e and 10a–d. Further, the presence of halogen substituent (F, Cl, Br) improved the activity profiles with the most potent conjugate **10d** (R=Br) exhibiting an IC<sup>50</sup> value of 1.2µM against G-3 strain, which is comparable to MTZ.

Similar SAR has been observed against C1 strain of T. foetuslike pathogen with activity mainly dependent upon the nature of substituent at C-3 position. The replacement of ketocarbonyl with ketal in general improved the activity profiles with conjugate **9c** (R=Cl) exhibiting an IC<sup>50</sup> value of 1.1 µM**.** However, the introduction of thiosemicarbazide functionality substantially improved the efficacy of the conjugates with compound **10a** (R=H) displaying an IC<sup>50</sup> value of 0.6µM, better than the standard drug MTZ. The synthesized conjugates were also evaluated against D1 strain of T. foetus. The conjugate **10a** again proved to be the most potent among the series, exhibiting an IC<sup>50</sup> value of 0.9µM. The generalized SAR of the synthesized conjugates against T. vaginalis and T. foetus has been provided in **Figure 3**.

Encouraged with these biological results, it was considered worthwhile to examine the activities of synthesized conjugates against other anaerobic protozoans such as WB strain of G. lamblia and HM1 strain of E. histolytica. Since trophozoites are the relevant forms that cause trichomoniasis, amebiasis and giardiasis, we tested the activity of compounds against trophozoite forms. As evident from IC<sup>50</sup> values enlisted in

**Table 1**, all conjugates exhibited good activity profiles with IC<sup>50</sup> values ranging from 0.4 to 16.2µM against WB strain of G. lamblia. Activity profiles showed a preference for electron withdrawing substituent (R= F, Cl, Br, NO2) at C-5 position of isatin except **10a** (R=H) which exhibited an IC<sup>50</sup> value of 0.4µM. The replacement of ketocarbonyl with ketal and thiosemicarbazide improved the activities except **8e** which displayed an IC<sup>50</sup> value of 1.6µM. Conjugates **8c, 8e, 9a, 9b, 9c, 9d, 9e, 10a, 10b,** and **10d** proved to be more potent than the standard drug MTZ (IC<sup>50</sup> = 6.4µM) (Arendrup et al., 2012), with conjugates **9d, 10a, and 10b** exhibiting IC<sup>50</sup> values in submicromolar concentration. Similarly, conjugates **8e, 9c, 9d, 9e, 10a,** and **10b** showed better potency than the current standard of care MTZ (IC<sup>50</sup> = 5µM)(Bashyal et al., 2017) against E. histolytica (**Table 1**). The generalized SAR of the synthesized conjugates against E. histolytica and G. lamblia has been provided in **Figure 4**.

To determine if these compounds were specific for protozoal parasites, we also tested a number of different fungal and bacterial species. In all assays, the activities of these compounds were specific for protozoa and exerted no detectable antifungal and antibacterial properties.

In conclusion, the present study undertook the synthesis of 1H-1,2,3-triazole-tethered metronidazole-isatin conjugates and evaluated their activity against multiple protozoan parasites, namely T. vaginalis, T. foetus, T. foetus-like pathogen, G. lamblia and E. histolytica. SAR studies revealed the dependence of activity profiles on the nature of substituent at C-3 and C-5 positions of isatin ring with most of the synthesized conjugates exhibiting comparable efficacies to that of MTZ against different trichomonal strains while better inhibitory activities were observed against G. lamblia. Interestingly, introduction of thiosemicarbazide into the basic core of the synthesized conjugates improved the activities with most potent conjugate **10a** being ∼16 folds more active than MTZ against G. lamblia, ∼2 folds more active than MTZ against E. histolytica and equipotent to MTZ against T. vaginalis and T. foetus. Further work on improving the activities of most promising conjugates of the synthesized library, is presently underway.

# AUTHOR CONTRIBUTIONS

SK synthesized and characterized metronidazole-isatin, metronidazole-spiroisatin, and metronidazole-isatinthiosemicarbazones. TB performed E. histolytica and G. lamblia experiments. AD conceptualized the E. histolytica and G. lamblia studies, analyzed the data, reviewed, and edited the manuscript. KL conceptualized the T. vaginalis and T. foetus, and T. foetuslike pathogen studies, analyzed the data, reviewed, and edited the

# REFERENCES


manuscript. AW performed T. vaginalis, T. foetus, and T. foetuslike pathogen experiments. LC conceptualized the anti-fungal and anti-bacterial studies, analyzed the data, reviewed, and edited the manuscript. JK and CT performed the anti-fungal and anti-bacterial experiments, respectively. VK designed and characterized the described conjugates, reviewed, and edited the manuscript.

# FUNDING

VK acknowledges the Council of Scientific and Industrial Research (CSIR, New Delhi) for financial assistance [grant no. 02(0293)/17/EMR-I]. The project described was partially supported by the National Institutes of Health, 1KL2TR001444 (to AD). KL and AW were supported by the Department of Biological Sciences at the University of the Pacific. LC, JK, and CT were funded by the United States Department of Agriculture, Agricultural Research Service (National Program 108, Project #5325-42000-039-00D).

# ACKNOWLEDGMENTS

We thank the Department of Biological Sciences at the University of the Pacific for their continuous support of research in antimicrobials.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00380/full#supplementary-material

Supplementary Data Sheet 1 | <sup>1</sup>H, <sup>13</sup>C NMR data of all the synthesized compounds along with scanned <sup>1</sup>H, <sup>13</sup>C and <sup>13</sup>C (DEPT) NMR of 8a, 8b, 8c, 9c, 10a, 10b.

as antitumor and antiparasitic agents. Organometallics 31, 5791–5799. doi: 10.1021/om300334z


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Kumar, Bains, Won Kim, Tam, Kim, Cheng, Land, Debnath and Kumar. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Discovery of Antiamebic Compounds That Inhibit Cysteine Synthase From the Enteric Parasitic Protist *Entamoeba histolytica* by Screening of Microbial Secondary Metabolites

Mihoko Mori 1,2, Satoshi Tsuge<sup>1</sup> , Wataru Fukasawa<sup>1</sup> , Ghulam Jeelani <sup>3</sup> , Kumiko Nakada-Tsukui <sup>4</sup> , Kenichi Nonaka1,2, Atsuko Matsumoto1,2, Satoshi Omura ¯ <sup>2</sup> , Tomoyoshi Nozaki <sup>3</sup> \* and Kazuro Shiomi 1,2 \*

#### *Edited by:*

Anjan Debnath, University of California, San Diego, United States

#### *Reviewed by:*

Elisa Azuara-Liceaga, Universidad Autónoma de la Ciudad de México, Mexico Mark Butler, The University of Queensland, Australia

#### *\*Correspondence:*

Tomoyoshi Nozaki nozaki@m.u-tokyo.ac.jp Kazuro Shiomi shiomi@lisci.kitasato-u.ac.jp

#### *Specialty section:*

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

*Received:* 29 August 2018 *Accepted:* 01 November 2018 *Published:* 05 December 2018

#### *Citation:*

Mori M, Tsuge S, Fukasawa W, Jeelani G, Nakada-Tsukui K, Nonaka K, Matsumoto A, Omura S, ¯ Nozaki T and Shiomi K (2018) Discovery of Antiamebic Compounds That Inhibit Cysteine Synthase From the Enteric Parasitic Protist Entamoeba histolytica by Screening of Microbial Secondary Metabolites. Front. Cell. Infect. Microbiol. 8:409. doi: 10.3389/fcimb.2018.00409 <sup>1</sup> Graduate School of Infection Control Sciences, Kitasato University, Tokyo, Japan, <sup>2</sup> Kitasato Institute for Life Sciences, Kitasato University, Tokyo, Japan, <sup>3</sup> Graduate School of Medicine, The University of Tokyo, Tokyo, Japan, <sup>4</sup> Department of Parasitology, National Institute of Infectious Diseases, Tokyo, Japan

Amebiasis is caused by infection with the protozoan parasite Entamoeba histolytica. Although metronidazole has been a drug of choice against amebiasis for decades, it shows side effects and low efficacy against asymptomatic cyst carriers. In addition, metronidazole resistance has been documented for bacteria and protozoa that share its targets, anaerobic energy metabolism. Therefore, drugs with new mode of action or targets are urgently needed. L-cysteine is the major thiol and an essential amino acid for proliferation and anti-oxidative defense of E. histolytica trophozoites. E. histolytica possesses the de novo L-cysteine biosynthetic pathway, consisting of two reactions catalyzed by serine acetyltransferase and cysteine synthase (CS, Oacetylserine sulfhydrylase). As the pathway is missing in humans, it is considered to be a rational drug target against amebiasis. In this study, we established a protocol to screen both a library of structurally known compounds and microbial culture extracts to discover compounds that target de novo cysteine biosynthesis of E. histolytica. The new screening system allowed us to identify the compounds that differentially affect the growth of the trophozoites in the cysteine-deprived medium compared to the cysteinecontaining medium. A total of 431 structurally defined compounds of the Kitasato Natural Products Library and 6,900 microbial culture broth extracts were screened on the system described above. Five compounds, aspochalasin B, chaetoglobosin A, prochaetoglobosin III, cerulenin, and deoxyfrenolicin, from the Kitasato Natural Products Library, showed differential antiamebic activities in the cysteine-deprived medium when compared to the growth in the cysteine-containing medium. The selectivity of three cytochalasans apparently depends on their structural instability. Eleven microbial extracts showed selective antiamebic activities, and one fungal secondary metabolite, pencolide, was isolated. Pencolide showed cysteine deprivation-dependent antiamebic activity (7.6 times lower IC<sup>50</sup> in the absence of cysteine than that in the presence of cysteine), although the IC<sup>50</sup> value in the cysteine-deprived medium was rather high (283 µM). Pencolide also showed inhibitory activity against both CS1 and CS3 isoenzymes with comparable IC<sup>50</sup> values (233 and 217µM, respectively). These results indicated that antiamebic activity of pencolide is attributable to inhibition of CS. Cytotoxicity of pencolide was 6.7 times weaker against mammalian MRC-5 cell line than E. histotytica. Pencolide has the maleimide structure, which is easily attacked by Michael donors including the thiol moiety of cysteine. The cysteine-adducts of pencolide were detected by mass spectrometric analysis as predicted. As CS inhibition by the pencolide adducts was weak and their IC<sup>50</sup> values to CS was comparable to that to the parasite in the cysteinecontaining medium, the cysteine-adducts of pencolide likely contribute to toxicity of pencolide to the parasite in the cysteine-rich conditions. However, we cannot exclude a possibility that pencolide inactivates a variety of targets other than CSs in the absence of cysteine. Taken together, pencolide is the first compound that inhibits CS and amebic cell growth in a cysteine-dependent manner with relatively low mammalian cytotoxicity.

Keywords: amebiasis, antiamebic compounds, cysteine synthase, *Entamoeba histolytica*, microbial secondary metabolites, natural products

# INTRODUCTION

Amebiasis is a diarrheal disease in humans, caused by infection with a protozoan parasite Entamoeba histolytica. WHO estimates 50 million cases of amebiasis, resulting in 40,000– 100,000 deaths annually worldwide (Harque et al., 2003; Stanley, 2003; Ximénez et al., 2009). Transmission occurs via the fecal-oral route, either directly by person-to-person contact or indirectly through ingestion of contaminated food or water. In East Asia and Australia, domestic infections are increasing among men who have sex with men, particularly those infected with HIV (James et al., 2010; Watanabe et al., 2011; Hung et al., 2012; Lo et al., 2014; Ishikane et al., 2016; Yanagawa et al., 2016). Metronidazole has been used as a drug of choice against amebiasis for decades despite its low efficacy against asymptomatic cyst carriers (Ali and Nozaki, 2007). For the treatment of cyst carriers, combination of metronidazole and paromomycin is currently recommended as a standard regimen. However, both compounds have a teratogenic effect and several adverse side effects, such as nausea, vomiting, and a metallic taste (Ohnishi et al., 2014). Moreover, it has been shown that E. histolytica is capable of surviving sub-therapeutic levels of metronidazole in vitro (Samarawickrema et al., 1997; Wassmann et al., 1999). Therefore, new drugs with different targets or modes of action have been needed.

L-Cysteine is essential for E. histolytica, and implicated in various important biological processes including attachment, motility, proliferation, and anti-oxidative defense (Gillin and Diamond, 1981; Fahey et al., 1984; Jeelani et al., 2010, 2014; Hussain et al., 2011). E. histolytica is cysteine autotrophic, and capable of de novo biosynthesis of L-cysteine by so called the sulfur-assimilatory de novo L-cysteine synthetic pathway, which does not exist in humans. This pathway consists of two reactions catalyzed by serine acetyltransferase (SAT, EC 2.3.1.30) and cysteine synthase (CS, O-acetylserine sulfhydrylase, EC 2.5.1.47) (Nozaki et al., 1998, 1999, 2005), and exists in bacteria, plants, and some parasitic protozoa (Ali and Nozaki, 2007). E. histolytica has three SAT (EhSAT1-3) and two CS (EhCS1 and 3) isotypes, and these enzymes have unique features when compared to those from bacteria and plants (Nozaki et al., 1998, 1999; Hussain et al., 2009). Recently, it was demonstrated that CS gene-silenced lines displayed a severe growth effect in an L-cysteine-lacking medium, whereas SAT1/2 or SAT3 gene silencing caused no or only mild growth defect in L-cysteine lacking medium (Jeelani et al., 2017), suggestive of the central role CS plays. As CS is essential for the parasite survival, this enzyme could be a promising drug target. In a few previous studies, EhCS inhibitors have been discovered. In silico prediction assisted to discover one compound that showed EhCS inhibition at the IC<sup>50</sup> value of 100µM (Nagpal et al., 2012), and a series of pyrazolo[3,4 d]pyrimidines with in vitro antiamebic activity at the IC<sup>50</sup> value of around 1µM might inhibit EhCS (Yadava et al., 2015). We previously conducted an enzymatic screening of the Kitasato Natural Products Library and microbial culture broth extracts (Mori et al., 2015). We found nine EhCS inhibitors from the Kitasato Natural Products Library and isolated two compounds (xanthofulvin and exophillic acid) from microbial culture broth extracts with the IC<sup>50</sup> value ranging 0.31–490µM (Mori et al., 2015). However, none of these compounds showed in vitro amebicidal activity with the comparable IC<sup>50</sup> values except for deacetylkinamycin C, which showed cytotoxicity to MRC-5.

In order to overcome this problem, we took advantage of a cell-based phenotypic screening using two culture conditions (cysteine-deprived and cysteine-supplemented conditions). If a compound inhibits CS leading to intracellular cysteine deprivation, trophozoites are not able to grow or killed without externally supplied L-cysteine; in other words, trophozoites should grow only in the cysteine-supplemented medium, but not in the cysteine-deprived medium. Using this strategy, we obtained pencolide from a fungal culture broth extract, which inhibits CS activity and cell growth at a comparable range of the IC<sup>50</sup> values, from microbial culture broth extracts.

# MATERIALS AND METHODS

# Organism and Culture

Trophozoites of E. histolytica clonal strain HM-1:IMSS cl6 were cultured axenically in Diamond's BI-S-33 medium at 37◦C (Clark and Diamond, 2002).

# Production of Recombinant *E. histolytica* CSs and Measurement of CS Inhibitory Activity

Expression and purification of recombinant E. histolytica CS1 and CS3, and measurement of inhibitory activity against both enzymes were described previously (Mori et al., 2015).

# *In vitro* Evaluation of Antiamebic Activity

Trophozoites were harvested in a logarithmic growth phase 3 or 4 days after inoculation of 1/30–1/12 volume of the seed culture. After the cultures were chilled on ice for 5 min, trophozoites were collected by centrifugation at 500 × g for 10 min at 4 ◦C. The trophozoites were resuspended in BI-S-33 medium containing 1% (v/v) penicillin/streptomycin (Life Technologies, Grand Islands, NY, U.S.A.), with a final concentration of 5 × 10<sup>4</sup> cells/ml. Approximately 200 µl of the trophozoite suspension (1 × 10<sup>4</sup> trophozoites) was dispensed into each well of a 96 well plate and incubated at 35.5◦C for 2 h under anaerobic conditions with Anaeropak (Mitsubishi Chemical, Tokyo, Japan). After incubation, the medium was removed and 180 µl of either L-cysteine-deprived or L-cysteine (9 mM) supplemented BI-S-33 medium, 10 µl of 2% (v/v) penicillin/streptomycin together, and 10 µl of sample solution [dissolved in 20% dimethylsulfoxide (DMSO) and 80% H2O] were added to each well. The plates were incubated under anaerobic conditions with Anaeropak for 48 hrs. After incubation, the medium was removed and 90 µl of pre-warmed Opti-MEM I (Life Technologies) and 10 µl of WST-1 solution (Dojindo, Kumamoto, Japan) were added to each well. Optical absorbance at 450 nm was measured on a photospectrometer (SH-9000Lab, Corona Electric, Ibaraki, Japan). Metronidazole (Sigma-Aldrich, MO, USA) was used as positive control. Antiamebic assays were done in triplicate. ED<sup>50</sup> values were calculated by the equation described previously (Arita-Morioka et al., 2015). The cysteine dependence of the antiamebic activities was calculated as EDCys(+) <sup>50</sup> /EDCys(−) <sup>50</sup> and expressed as "cysteine dependence index."

# Evaluation of Cytotoxicity Against MRC-5 Cells

Human fibroblast cell line, MRC-5, was used for cytotoxicity evaluation. MRC-5 cells were plated on 96-well flat bottom plates at a density of 1.5 × 10<sup>4</sup> cells/well in 100 µl of MEM medium each well (Life Technologies) containing 10% fetal bovine serum (Hana-nesco Bio, Tokyo, Japan) and 1% penicillin-streptomycin (Life Technologies) and incubated at 37◦C with 5% CO<sup>2</sup> for 2 days. Test compounds in 5 µl of 50% DMSO and 50% H2O and 100 µl of MEM medium were mixed and added to each well. After cultivation at 37◦C with 5% CO<sup>2</sup> for 2 days, cell density and morphological changes were observed under a microscope. After observation, 10 µl of WST-8 solution (Dojindo) was added to the cells and the plate was incubated at 37◦C with 5% CO<sup>2</sup> for 2 h. Then, absorbance at 450 nm was measured as above. Cytotoxicity against MRC-5 cells was measured in triplicate. ED<sup>50</sup> values were calculated as previously described (Arita-Morioka et al., 2015).

# Screening Sources

The Kitasato Natural Products Library consists of 431 natural and semisynthetic compounds. The compounds were dissolved at 1 mg/ml with DMSO and kept at −20◦C. The microbes for the cell-based screening were collected in Japan. Fungal strains were isolated from soil samples that had been collected from a close proximity to plants. Actinomycetes were isolated from plants and soil samples attached to plant roots. Each microbe was cultured in 10 ml of two to four different media whose main carbon sources were glucose, sucrose, starch, and rice, using 50-ml glass or disposable plastic tubes for 6 days (shaking culture for liquid media) or 13 days (static culture for rice medium). In total, 6,900 broth extract samples (3,224 fungi and 3,676 actinomycetes) were prepared. After cultivation, the same amount of EtOH was added to each broth, and the mixtures were then centrifuged at 1,630 × g for 10 min. The supernatants were used for screening.

# Primary Screening of Antiamebic Compounds From the Kitasato Natural Products Library and Microbial Culture Broths for Antiamebic Acitivities

Approximately 2 µl of the compound solution (1 mg/ml, in DMSO) of the Kitasato Natural Products Library or 10 µl of EtOH extracts of microbial culture broth extracts was added to each well of E. histolytica trophozoite-seeded 96-well microtiter plates (final concentrations, 10µg/ml). Compounds showing more than 80% growth inhibition and having no cytotoxicity against MRC-5 cells were selected for further evaluation.

# Isolation of Pencolide From a Culture Broth of Unidentified Fungus FKI-7363 Producing Strain and Cultivation

The unidentified fungal strain FKI-7363 was isolated from a soil sample collected in Tokushima city, Tokushima Prefecture, Japan. FKI-7363 strain was isolated after cultivation on a malt extract agar plate consisting 2% malt extract, 2% glucose, 0.1% peptone, and 2% agar (pH was adjusted to 6.0 before sterilization) for 7 days. The genus of FKI-7363 strain could not be assigned by its morphological features. FKI-7363 strain was maintained on an LcA slant consisting of 0.1% glycerol, 0.08% KH2PO4, 0.02% K2HPO4, 0.02% MgSO4·7H2O, 0.02% KCl, 0.2% NaNO3, 0.02% yeast extract**,** and 1.5% agar (pH 6.0). A loopful of spores were inoculated into six 50-ml tubes containing each 10 ml of the seed medium consisting of 2% glucose, 0.5% Polypepton (Nihon Pharmaceutical, Tokyo, Japan), 0.2% yeast extract, 0.2% KH2PO4, 0.05% MgSO4·7H2O**,** and 0.1% agar (pH 6.0). The tubes were shaken on a reciprocal shaker at 27◦C for 3 days. Thirty milliliters of the seed culture were inoculated into each of two Ulpack 47 bags (Hokken Co. Ltd, Tochigi, Japan), each containing 300 g of sodden rice with 3 g of kobucha (Japanese kelp tea) and the cultures were incubated statically at 25◦C for 14 days.

# Isolation of Pencolide From the Culture Broth

The obtained culture (600 g) was mixed and extracted with 600 ml of EtOH with an electric mixer for 30 min. The mixture was centrifuged at 3,000 × g for 10 min and the supernatant was filtered with a filter paper (No.2; Toyo Roshi Kaisha, Ltd., Tokyo, Japan). After the filtrate was concentrated and EtOH was removed in vacuo, the residue was applied to an ODS column (25 ϕ × 110 mm). The samples were eluted with a H2O-MeCN gradient system to give 7 fractions (100:0, 90:10, 80:20, 70:30, 60:40, 40:60 and 0:100, 500 ml each). The active 90:10 fraction (733 mg) was applied to an ODS column (20 ϕ × 110 mm). An H2O-MeCN system described above but containing 0.1% trifluoroacetic acid (TFA) was used as eluent. The active component was eluted in 10–20% MeCN fractions. The active fractions were mixed and dried, and then yielded 225 mg amorphous residues. This crude material was purified by preparative HPLC (column, Develosil ODS-C30-UG-5, 20 ϕ × 250 mm, Nomura Chemical, Japan; solvent, 15% MeCN-H2O containing 0.1%TFA; flow rate, 7.0 ml/min; detection, UV at 210 nm). The peak eluted at 40 min was collected and concentrated to dryness to afford pencolide (70.3 mg). The NMR and MS data shown below were in agreement with those reported previously (Sutherland, 1963; Wang et al., 2010).

# Pencolide

White amorphous. <sup>13</sup>C NMR (100 MHz, CDCl3) δ ppm: 11.2, 14.6, 122.6, 128.1, 146.0, 146.5, 167.4, 168.8, 169.9 ppm. <sup>1</sup>H NMR (400 MHz, CDCl3) δ ppm (Int., mult., J in Hz): 7.42 (1H, q, 7.0), 6.45 (1H, q, 2.0), 2.14 (3H, d, 2.0), 1.82 (3H, d, 7.0). ESI-MS: 194.0646 ([M–H]−), 196.0622 ([M+H]+).

# Evaluation of the Effect of L-Cysteine on the Antiamebic Effects of Epoxide-Containing Compounds

Each 5 µl of 1 mg/ml DMSO solution of cytochalasan compounds and cerulenin was dispensed into each well of a 96-well microtiter plate and 200 µl of 9 mM L-cysteine solution (L-cysteine HCl salt in pure water) or water was added to each well. After mixing well with plate mixer for 3 min, the plate was incubated at 37◦C axenically with Anaeropak (Mitsubishi Chemical, Tokyo, Japan) for 2 days. The reaction mixtures were evaluated by HPLC analysis (column, Senshupak Pegasil ODS SP100, 4.6 ϕ × 250 mm, Senshu Scientific Co. Ltd., Japan; solvent, MeCN-H2O gradient system, 10–90% MeCN per 0–30 min; flow rate, 1.0 ml/min; detection, UV at 210 nm) and LC-ESI-MS analysis (column, Capcellcore C18, 2.1 ϕ × 50 mm, Osaka soda, Japan; solvent, MeCN-H2O-0.1% HCO2H gradient system, 5–100% MeCN per 0–8 min, 100%MeCN per 8–10 min; flow rate, 0.4 ml/min; ESI-MS data was obtained by JEOL JMS-T100LP, JEOL, Japan).

# Purification of L-Cysteine Adducts of Pencolide and Measurement of the CS1 Inhibitory Activity

Pencolide (5.5 mg) was dissolved in 1 ml of 0.25 M L-cysteine aqueous solution (L-cysteine HCl salt in H2O) and left at room temperature for overnight. The reaction mixture was purified by Seppak ODS (Waters co., Massachusetts, USA) and the cysteine adducts were eluted with 30%MeCN-70%H2O containing 0.1%TFA. The obtained fraction was further purified by preparative HPLC (column, Senshupak Pegasil ODS SP100, 20 ϕ × 250 mm, Senshu Scientific Co. Ltd.; solvent, 0.1%TFA-MeCN-H2O gradient system, 10–50% MeCN per 0–30 min; flow rate, 7.0 ml/min; detection, UV at 210 nm). The cysteine adducts were eluted at 12 min. The fraction was dried in vacuo and colorless amorphous compounds (7.3 mg) were obtained. The methanol solution of the obtained adducts was used for measurement of the CS1 inhibitory activity.

# RESULTS AND DISCUSSION

# A New Screening System to Discover Antiamebic Compounds That Inhibit *de novo* Cysteine Biosynthesis in *E. histolytica*

Although in our previous study, we identified, by an enzyme assay-based screening, two natural fungal secondary metabolites, xanthofulvin and exophillic acid, which inhibited CS, these compounds did not show the antiamebic activity in vitro (Mori et al., 2015). The failure can be due to several possible reasons including low permeability of the compounds to the cell and low solubility of the compounds in aqueous culture media.

We aimed to improve the screening system in order to specifically identify the compounds that inhibit growth of or kill the amebic trophozoites through inhibition of the de novo cysteine biosynthetic pathway. The rationale was as follows: the amebicidal activity and/or the growth inhibition should be more pronounced when the cells are cultured in the absence of L-cysteine than in the L-cysteine-supplemented conditions because under L-cysteine deprived conditions, the cells depends on the de novo cysteine biosynthesis for survival. Thus, we carried out primary cell-based phenotypic screening in both the presence and absence of L-cysteine and selected hits that showed differential amebicidal effects between the two conditions. The essentiality of CS genes for the growth and survival of trophozoites in the L-cysteine-lacking conditions has been demonstrated (Jeelani et al., 2017).

# Discovery of CS-Dependent Amebicidal Compounds From the Kitasato Natural Compounds Library

We screened the Kitasato Natural Compounds Library composed of 431 compounds. In the first screening, the compounds that showed 80% growth inhibition against E. histolytica trophozoites at 10µg/ml in the L-cysteine-deprived medium were selected. In the second screening, the compounds that showed more than two-fold difference in the IC<sup>50</sup> value of antiamebic activities between the L-cysteine-supplemented [Cys (+)] medium and the cysteine-deprived [Cys (–)] medium (i.e., higher static or amebicidal activity without L-cysteine than with L-cysteine). Five compounds were identified: aspochalasin B, chaetoglobosin A, prochaetoglobosin III, cerulenin, and deoxyfrenolicin (**Figure 1**, **Table 1**). Three of these five hits, aspochalasin B, chaetoglobosin A, and prochaetoglobosin III, were cytochalasans. Cytochalasans are fungal metabolites and known to inhibit actin polymerization (Scherlach et al., 2010). The growth inhibitory effect of cytochalasans against the reptilian E. invadens trophozoites were previously demonstrated; most of them showed growth inhibition at 10µM (Makioka et al., 2004). Cerulenin is known to inhibit fatty acid biosynthesis (Vance et al., 1972). Deoxyfrenolicin has antifungal and anti-Mycoplasma activity (Iwai et al., 1978), and mammalian cytotoxicity of deoxyfrenolicin was also reported (Wang et al., 2013). Next, inhibitory effects of these five antiamebic compounds on the CS activity were examined. We used two E. histolytica enzymes, CS1 and CS3, which are 83% identical at the amino acid levels, but likely plays distinct physiological roles (Jeelani et al., 2017). Deoxyfrenolicin showed mild inhibition toward both CS1 and CS3 as previously reported (Mori et al., 2015); its antiamebic activity was, however, much higher than the CS inhibitory activity, suggesting (1) that the observed antiamebic activity was unlikely attributable to the CS inhibition, or (2) deoxyfrenolicin is converted to more CS-inhibitory secondary derivative(s) after metabolized in the amebas. Interestingly, deoxyfrenolicin has a naphthoquinone moiety in the structure, which was frequently found in CS inhibitors as previously reported (Mori et al., 2015). On the other hand, neither three cytochalasans nor cerulenin showed the inhibitory activity at 100µg/ml against both EhCSs (**Table 1**).

# The Cysteine-Dependent Cell-Based Assay Is Prone to False Hits for L-Cysteine-Sensitive Compounds

Two of the three cytochalasans showed good L-cysteine dependence, but did not show CS inhibition. We speculated if the growth inhibition by these cytochalasans was masked or abolished by L-cysteine supplementation in the assay. Three cytochalasans that were identified as hits have α,β-unsaturated ketone moiety in their structures (**Figure 1**). In previous studies, involvement of α,β-unsaturated ketone moiety in the biological activity and cytotoxicity of cytochalasans was indicated (Flashner et al., 1982; Tomikawa et al., 2002). This moiety is easily attacked by the thiol moiety of L-cysteine, resulting in 1,4-addition (Michael addition) of cysteine (Flashner et al., 1982). Therefore, we considered a possibility that in the Cys (+) medium, abundantly present cysteine attacked on the α,β-unsaturated moiety of each cytochalasan and consequently canceled their

TABLE 1 | CS inhibitory activity and anti-amebic activity of five compounds screened from Kitasato Natural Compounds Library.


amebicidal activity. To test this hypothesis, two representative cytochalasans (chaetoglobosin A and prochaetoglobosin III) were incubated in both cysteine-supplemented (9 mM) and cysteine-lacking water at 37◦C for 2 days. After incubation, components of each solution were determined by HPLC analysis. In cysteine-supplemented water, the peaks of cytochalasans were totally lost (**Supplemental Figures 1**, **2**), although the peaks corresponding to their cysteine adducts were not detected by LC-MS (**Supplemental Figures 5**–**20**). Newly appeared peaks showed the same molecular weight as corresponding parent cytochalasans. It is known that chaetoglobosin A is unstable and isomerized into another form of chaetoglobosin in acidic and basic conditions (Sekita et al., 1973, 1982a,b). As our cysteinesupplemented water was acidic (because L-cysteine hydrochloric acid was used in the assay), chaetoglobosin A was likely derived into chaetoglobosins B and D as demonstrated (Sekita et al., 1982a,b). Since prochaetoglobosin III and chaetoglobosin A structurally resemble (**Figure 1**), they were likely converted into their corresponding isomers in cysteine-rich water. On the other hand, the pH of the amebic culture medium was neutral (∼6.8). Thus, the dependence of cytochalasans' amebicidal activity on cysteine deprivation cannot be well-explained, although the structural unstability of cytochalasans seems to be a possible reason. The epoxide moiety of cytochalasans was also implicated in mammalian cytotoxicity (Li et al., 2014). Since some isomers of chaetoglobosin A lack an epoxide in their structures, these environment-driven structural changes of cytochalasans could be a cause of the selectivity in the cysteine-dependent antiamebic activity.

Cerulenin does not contain α,β-unsaturated ketone moiety; instead, it has an epoxide moiety in the structure. We also tested reactivity of cerulenin to cysteine. Although cerulenin was relatively stable in this condition (**Supplemental Figure 3**), the IC<sup>50</sup> values of the compound against CS1 and CS3 were >10- to 100-fold higher than the ED<sup>50</sup> values against the cells. A possible mode of action of cerulenin and the dependence of antiamebic activity on L-cysteine remain elusive.

# Discovery of Microbial Broth Extracts Showing CS-Dependent Amebicidal Activities

Next, we screened microbial broth extracts to discover antiamebic agents with CS inhibitory activity. The extracts of 3,224 fungal broths and 3,676 actinomycetes broths were screened. The results are shown in **Table 2**. In the first screening, 214 microbial broth extracts (152 fungal and 62 actinomycetes) showed antiamebic activity in the Cys (–) medium. In the second screening, only 11 fungal broth extracts showed >10 times differentially reduced antiamebic activity in the Cys (+) medium compared to that in the Cys (–) medium. Inhibitory activities of these 11 fungal broth extracts were evaluated against CS1 and CS3, and 6 samples showed inhibitory activity against both CSs. We roughly chromatographed these samples with small-scale ODS column and measured antiamebic activity of obtained fractions. As five samples showed the same trend, we selected one fungal strain FKI-7363 as a representative. This strain was further subjected to purification of active compounds. Purification of active compounds from remaining strains will be described elsewhere.

# Pencolide Was Identified as an CS-Inhibiting Antiamebic Agent Isolated From a Cultured Broth of FKI-7363

One broth extract produced by an unidentified fungus, FKI-7363, showed selective inhibitory activity in the Cys (–) medium. Five µl of the broth extract (50% EtOH broth extract) showed 50% amebicidal activity in the Cys (–) medium, on the other hand, 50 µl of the extract showed no measurable antiamebic activity in the Cys (+) medium, suggesting that the cysteine dependence index was >10.

An extract of FKI-7363 was produced from the 14-day large size cultures on the rice medium for purification. The culture was extracted with EtOH, and the concentrated extract was purified by repetitive column chromatography with acidic MeCN-H2O system. The active fractions were purified by preparative HPLC, and finally, 70 mg of an active compound were obtained from 600 g of cultured rice. The structure of active compound was determined by <sup>1</sup>H and <sup>13</sup>C NMR spectra and ESI-MS spectra (see Materials and Methods), and the active compound was identified as pencolide (**Figure 2**). The reported producer of pencolide was Penicillium multicolor (Birkinshaw et al., 1963), and its weak antimicrobial biological activity was previously described against some Gram positive/negative bacteria and Candida albicans (Lucas et al., 2007).

# Antiamebic and CS Inhibitory Activities of Pencolide

We measured antiamebic activity of pencolide in both the Cys (+) and Cys (–) media (**Table 3**). The IC<sup>50</sup> values of pencolide

TABLE 2 | EhCS inhibitory samples found in microbial broth extracts.


are 283µM in the Cys (–) medium and 2,140µM in the Cys (+) medium, respectively, and the cysteine dependence index of 7.6 [7.6-fold more potent in the Cys (–) medium than in the Cys (+) medium]. Pencolide caused apoptosis-like cell death against E. histolytica in Cys (–) medium as observed under a light microscope; the pencolide-treated trophozoites was morphologically similar to those treated with metronidazole (**Supplemental Figure 27**). Pencolide showed low cytotoxicity (EC<sup>50</sup> = 1,900µM) against a mammalian cell line, MRC-5. The selectivity index of pencolide against the ameba relative to MRC-5 cells is 6.7. CS inhibitory activity of pencolide was also measured. The IC<sup>50</sup> values of pencolide against CS1 and CS3 are 233 and 217µM, respectively, which are comparable to the levels of the ED<sup>50</sup> value against the ameba cultured without L-cysteine (**Table 4**), reinforcing that CS inhibition is likely the cause of antiamebic activity.

# Pencolide Is Able to Become the Michael Acceptor in Cys (+) Medium

Pencolide showed antiamebic activity in a cysteine-dependent manner and CS inhibitory activity. However, pencolide has a maleimide structure which is known to be reactive with the thiol moiety. The Michael-type addition reaction between the thiol moiety of cysteine and maleimide is fast; this reaction is often utilized for labeling biomolecules (Boyatzis et al., 2017). We tested the reactivity and stability of pencolide in cysteine-supplemented conditions. After incubation with 9 mM aqueous cysteine solution for 2 days, we confirmed that most of pencolide was changed to the cysteine adducts by Michaeltype addition reaction (**Supplemental Figures 4**, **21**–**26**). Thus, the cysteine-dependent loss of anti-amebic activity of pencolide was apparently caused by adduct formation of pencolide. Pencolide inhibited CS and showed antiamebic activity in cysteine-deprived conditions with comparable concentrations. Furthermore, we purified cysteine adducts of pencolide and measured their CS inhibitory activity. The IC<sup>50</sup> value of the

TABLE 3 | In vitro antiamebic activities and cytotoxicity against MRC-5 cells of pencolide.


TABLE 4 | Cysteine synthase inhibitory activities of pencolide and its cysteine adducts.


cysteine adducts against CS1 was estimated to be 2,340µM (740µg/ml; the apparent molecular weight of cysteine adduct was 316; **Supplemental Figures 25**, **26**). This IC<sup>50</sup> value is comparable to ED<sup>50</sup> value against E. histolytica trophozoites of pencolide in Cys (+) medium. From these results, we concluded that the amebicidal activity of pencolide was attributable to CS inhibition. Since the maleimide moiety of pencolide easily accepts Michael donors, this labile moiety must be substituted with another stable structure in order to further utilize pencolide as a seed compound of new antiamebic drugs with a novel mode of action.

# The First Identification of the Antiamebic Compounds With CS Inhibitory Activity by Screening of Microbial Culture Broths

In this study, we have identified for the first time pencolide as an antiamebic compound with comparable CS inhibitory activity by screening about 7,000 microbial culture broths using a new protocol for L-cysteine-dependent antiamebic assay. In the previous study, we identified deacetylkinamycin C showing the ED<sup>50</sup> value against amebic trophozoites and the IC<sup>50</sup> value against CSs of ca. 20µM, from the Kitasato Natural Products Library. However, the compound showed strong (>10 times higher) cytotoxicity against MRC-5 cells (Mori et al., 2015). Furthermore, no new CS inhibitors that showed anti-amebic activity in an L-cysteine-dependent manner, were discovered through CS-based screening of the microbial culture broth extracts (Mori et al., 2015). Thus, we claim that our new screening protocol of the L-cysteine-dependent phenotypic screening has provided a first proof-of-concept for a screening of CS-inhibiting antiamebic compounds. Although we successfully screened both a chemical library and microbial culture extracts using our Lcysteine-dependent screening platform to identify pencolide, it should be noted that our proposed protocol for L-cysteinedependent cell-based screening is prone to an artifact for the compounds that are inactivated or structurally altered by Lcysteine, as shown for some cytochalasans.

Although pencolide showed only moderate potency at this stage, it has good potential for full organic synthesis and further derivatizations based on its low molecular weight. For rational modifications of pencolide for better efficacy, it is necessary to elucidate the co-crystal structures of CS together with pencolide. We should also emphasize the fact that natural products, especially microbial secondary metabolites, have rich structural diversity (Newman and Cragg, 2016). Furthermore, there has still existed many undiscovered microbes which have potential to produce novel compounds (Ling et al., 2015). A few natural compounds whose structures are related to that of pencolide have been reported. Fungal metabolites, farinomaleins, have a maleimide structure similar in pencolide (Putri et al., 2009; El Amrani et al., 2012). Farinomalein showed antifungal activity against a plant pathogen Phytophthora sojae (Putri et al., 2009). Farinomaleins also showed no cytotoxicity against mammalian cells, and therefore, pencolide-related natural compounds appear to be a safe seed compounds to further develop antiamebic agents

although we have to conquer the problem that the maleimide moiety tends to be attacked by various Michael donors.

In summary, we introduced a new L-cysteine- (and CS-) dependent cell-based phenotypic screening system for the discovery of antiamebic agents. We screened 431 compounds of the Kitasato Natural Products Library and 6,900 samples of microbial broth extracts. We identified from the Kitasato Natural Products Library, a naphtoquinone compound, deoxyfrenolicin, inhibited E. histolytica CS among six compounds showing L-cysteine-dependent anti-amebic activity. However, CS inhibition-mediated anti-amebic activity by the compound was not validated. Pencolide was purified and identified from microbial broth extracts to be a fair antiamebic lead compound as it showed L-cysteine-dependent anti-amebic activity against E. histolytica trophozoites in cysteine-deprived medium and CS inhibition at comparable concentrations with lesser cytotoxicity against mammalian cells. Therefore, pencolide may be a seed compound to develop a new class of antiamebic agents that target CS.

# AUTHOR CONTRIBUTIONS

MM designed and conducted this study. ST and WF also conducted this study under MM. GJ and KN-T contributed to

# REFERENCES


design the assay system. KN and AM contributed to identify microbes and provided microbial cultured broths. SO provided ¯ advices and facilities of the study. TN and KS were supervisors of this study.

# ACKNOWLEDGMENTS

We thank Dr. Kenichiro Nagai, School of Pharmacy, Kitasato University for the measurement of mass spectra. This work was supported by JSPS Grant-in-Aid for Scientific Research (C) (26460129 and 17K08344) and Kitasato University Research Grant for Young Researchers (to MM), a grant from The Japan Agency for Medical Research and Development (AMED) (JP17fk0108119), a grant for Science and Technology Research Partnership for Sustainable Development (SATREPS) from AMED (17jm0110009h0004) and Japan International Cooperation Agency (to MM, KS, and TN).

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00409/full#supplementary-material


of L-cysteine in the enteric protozoan parasite Entamoeba histolytica. MBio 5, e01995–e01914. doi: 10.1128/mBio.01995-14


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Mori, Tsuge, Fukasawa, Jeelani, Nakada-Tsukui, Nonaka, Matsumoto, Omura, Nozaki and Shiomi. This is an open-access article distributed ¯ under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# High-Throughput Screening of *Entamoeba* Identifies Compounds Which Target Both Life Cycle Stages and Which Are Effective Against Metronidazole Resistant Parasites

Gretchen M. Ehrenkaufer <sup>1</sup> , Susmitha Suresh<sup>1</sup> , David Solow-Cordero<sup>2</sup> and Upinder Singh1,3 \*

<sup>1</sup> Division of Infectious Diseases, Department of Internal Medicine, Stanford University School of Medicine, Stanford, CA, United States, <sup>2</sup> High-Throughput Bioscience Center, Department of Chemical and Systems Biology, Stanford University, Stanford, CA, United States, <sup>3</sup> Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, CA, United States

### *Edited by:*

Serge Ankri, Technion – Israel Institute of Technology, Israel

### *Reviewed by:*

Esther Orozco, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional (CINVESTAV-IPN), Mexico Fuhang Song, Key Laboratory of Pathogenic Microbiology & Immunology, Institute of Microbiology (CAS), China

> *\*Correspondence:* Upinder Singh usingh@stanford.edu

*Received:* 20 May 2018 *Accepted:* 25 July 2018 *Published:* 17 August 2018

#### *Citation:*

Ehrenkaufer GM, Suresh S, Solow-Cordero D and Singh U (2018) High-Throughput Screening of Entamoeba Identifies Compounds Which Target Both Life Cycle Stages and Which Are Effective Against Metronidazole Resistant Parasites. Front. Cell. Infect. Microbiol. 8:276. doi: 10.3389/fcimb.2018.00276 Neglected tropical diseases, especially those caused by parasites, are significantly underserved by current drug development efforts, mostly due to the high costs and low economic returns. One method for lowering the costs of drug discovery and development for these diseases is to repurpose drugs developed for other indications. Here, we present the results of a screen of five repurposed drug libraries to identify potential new lead compounds to treat amebiasis, a disease that affects tens of millions of people and causes ∼100,000 deaths annually. E. histolytica, the causative agent of amebiasis, has two major life cycle stages, the trophozoite and the cyst. The current primary treatment for amebiasis, nitroimidazole compounds, do not eliminate parasites from the colonic lumen, necessitating a multi-drug treatment regimen. We aimed to address this problem by screening against both life stages, with the aim of identifying a single drug that targets both. We successfully identified eleven compounds with activity against both cysts and trophozoites, as well as multiple compounds that killed trophozoites with improved efficacy over existing drugs. Two lead compounds (anisomycin and prodigiosin) were further characterized for activity against metronidazole (MNZ) resistant parasites and mature cysts. Anisomycin and prodigiosin were both able to kill MNZ resistant parasites while prodigiosin and its analog obatoclax were active against mature cysts. This work confirms the feasibility of identifying drugs that target both Entamoeba trophozoites and cysts, and is an important step toward developing improved treatment regimens for Entamoeba infection.

Keywords: *Entamoeba*, amebiasis, drug screen, drug repurposing, cyst

# INTRODUCTION

The protozoan parasite Entamoeba histolytica contributes significantly to morbidity and mortality in the developing world and is a leading parasitic cause of death (World Health Organization, 1997). The infectious cycle of Entamoeba begins with the ingestion of the cyst, which undergoes excystation in the small intestine to produce the invasive trophozoite form. Trophozoites invade tissue and cause disease symptoms of colitis and liver abscess. Some parasites in the colon convert to the cyst form, which is excreted in the stool and can infect new hosts. Thus, both the cyst and trophozoite forms are important to pathogenesis: cysts transmit infection and trophozoites cause disease symptoms.

All invasive disease with E. histolytica should be treated. Additionally in non-endemic countries, the World Health Organization also recommends treatment of asymptomatic colonization with the goal of preventing cyst shedding, as these can be transmitted to household members or close contacts (Stanley, 2003). Treatment for amebiasis is reliant on a single class of agents, the nitroimidazole compounds (i.e., MNZ and tinidazole) (Haque et al., 2003). Advantages of MNZ are that it is highly effective at killing invasive trophozoites, gets to systemic levels to treat liver abscesses, is cheap, and can be orally dosed. However, because MNZ is rapidly absorbed and has poor activity against cysts, it is ineffective in fully treating luminal disease and cysts. Thus, 40% of patients treated with MNZ will continue to have parasites in the colonic lumen (Haque et al., 2003) and a second agent (paramomycin or iodoquinol) must be given to completely clear remaining trophozoites and cysts from the colonic lumen. Other drawbacks to MNZ include unpleasant side effects, alcohol intolerance, and problems with use during pregnancy and lactation (Roe, 1977). The complexity of the treatment regimen increases likelihood of patient noncompliance especially when the second agent is needed at a time when the patient may have clinically improved. Infection with and successful treatment of Entamoeba does not confer protective immunity; thus, in countries where the pathogen is endemic, individuals get repeated episodes of invasive disease, and require repeated treatment (Haque et al., 2003). Given that patients are given repeated episodes of treatment and that resistance to MNZ can easily be induced in the lab, it raises concerns that resistant strains may arise (Wassmann et al., 1999).

Given their significant impact on human health, discovery of new therapeutics for parasitic diseases such as amebiasis and other neglected tropical diseases is vital. However, the costs of developing a new drug, which can top one billion dollars (DiMasi et al., 2003), can be prohibitive for diseases mostly found in developing countries. For this reason, screening of repurposed drug libraries, consisting of compounds with known bioactivities and toxicity profiles, has been gaining popularity. Identification of a new indication for a known drug or compound with previously established pre-clinical or clinical data can greatly reduce the costs and time required of bringing a drug to market.

A recent example of this approach in amebiasis is the drug auranofin (Debnath et al., 2012), which recently completed a Phase I trial (Capparelli et al., 2017) to establish safety and pharmacokinetic profiles. Auranofin is a gold-containing compound originally developed to treat rheumatoid arthritis (Minigh, 2007). Screening of a ∼700 compound library revealed potent activity of auranofin against E. histolytica trophozoites, and in vivo efficacy was demonstrated in an animal model of amebic colitis (Debnath et al., 2012). Auranofin was subsequently found to be active against Giardia (Tejman-Yarden et al., 2013), suggesting the possibility of its use for a more general treatment of gastrointestinal parasites. Despite the promise of this new potential anti-amebic agent, the path to clinical use is not guaranteed and it is evident that more lead compounds are needed to increase the likelihood of an effective treatment making it to the clinic.

In order to find new drugs targeting Entamoeba with a simplified treatment regimen, we aimed to identify compounds that could target both the trophozoite and cyst forms. Because E. histolytica cannot be induced to encyst in vitro, we used Entamoeba invadens, a related Entamoeba in which high efficiency encystation can be induced to perform the screen. Both trophozoite and cyst stages of E. invadens were screened simultaneously using five libraries, totaling ∼3,400 unique compounds; these compounds are enriched for known bioactives and drugs with clinical data, including some FDAapproved compounds. Screening both life stages, we identified three categories of compounds: those that target trophozoites only, those that target cysts only, and those that target both trophozoites and cysts. Following a second round of confirmation in E. invadens, select hits were further screened against E. histolytica trophozoites. A total of nine compounds had significant activity at ≤10µM concentration in E. histolytica. Two promising lead compounds, anisomycin and prodigiosin, were chosen for further study and characterized for activity against MNZ resistant parasites and mature cysts. This study represents the first successful high throughput screen for compounds targeting multiple life-cycle stages of Entamoeba. We have shown that it is possible to identify promising drug candidates that have activity against both trophozoites and cysts and our results give confidence that simplified treatment regimens for Entamoeba can be developed.

# MATERIALS AND METHODS

# Parasite Culture and Strains Used

Entamoeba invadens (strain IP-1)was grown and maintained at 25◦C in LYI media under standard conditions (Ehrenkaufer et al., 2013). For the trophozoite assay, a stable transgenic E. invadens cell line expressing luciferase (CK-luc) was established by transfection (Ehrenkaufer and Singh, 2012) and maintained in 40µg/ml G418 in LYI media. Entamoeba histolytica strain HM1:IMSS was grown and maintained at 37◦C in TYI media under standard conditions (Diamond et al., 1978).

# Compound Libraries

Compounds from Sigma Library of Pharmacologically active compounds (LOPAC 1280: sigmaaldrich.com/life-science/ cell-biology/bioactive-small-molecules/lopac1280-navigator), NIH clinical collection (NIHCC: pubchem.ncbi.nlm.nih.gov/ source/NIH Clinical Collection), Biomol known bioactives: enzolifesciences.com/BML-2840/iccb-known-bioactives-library, Biomol FDA approved: enzolifesciences.com/BML-2843/screenwell-fda-approved-drug-library and Microsource Spectrum msdiscovery.com/spectrum.html were obtained from the Highthroughput Biosciences center at Stanford. A majority of the compounds were maintained in DMSO at a stock concentration of 10 mM. The NIHCC was screened in duplicates while the rest of the libraries were screened in seven-point dose curve with the highest concentration in duplicate for the trophozoite screen. Only the top three concentrations were tested in the cyst screen with the highest concentration in duplicate. Compounds that were active against trophozoites and cysts were re-confirmed in a 384-well format before the secondary assay.

# Compound Screening

Compounds (200 nl) were pinned using the SciClone into 384 well plates (EK-30080 white plates for trophozoite screening and EK-30091 clear-bottomed black plates for encystation). CKluc trophozoites were harvested mid-log phase and seeded with 100 µl media and drug or DMSO control to a final density of 5,000 parasites per well. DMSO only wells were used as a positive control, and wells lacking parasites were used to establish background signal. Plates were sealed to create an anaerobic environment, and allowed to grow for 48 h at 25◦C. For the trophozoite killing assay, plates were spun briefly then inverted to pour off media. Bright-GLO luciferase reagent (Promega) was added (20 µl of a 50:50 dilution in PBS) and plates were allowed to incubate 30 min at room temperature to ensure lysis. Luminescence was read on a Tecan Infinite M1000 pro.

For the encystation assay, IP-1 trophozoites were harvested in mid-log phase, washed once in encystation media (47%LG) (Sanchez et al., 1994), and plated at a density of 30,000 parasites per well. DMSO only wells were used as a positive control, and trophozoites were used to establish background signal. Sealed plates were incubated for 48 h at 25◦C, then spun and the media removed as with the trophozoite plates. Cysts were stained by addition of 40 µl of 50µM calcofluor white (Sigma) in PBS, followed by imaging in an ImageXpress micro (Molecular Devices). Number of cysts per well was quantified using MetaXpress software.

# Secondary Screen in *E. histolytica* Using Cell Viability Assay

Entamoeba histolytica trophozoites were seeded in 96-well plates (Corning 3904; black clear bottom) containing 350 µl media and either drug or DMSO at a density of 10,000 cells per well. The plates were then sealed and incubated at 37◦C. After 72 h, the media was aspirated and a cell viability dye, fluorescein diacetate (FDA) was added to each well at a concentration of 20µg/ml, diluted in media. The plates were incubated for 20–30 min at 37◦C. The FDA was removed by aspiration and 100 µl of 1X PBS added to each well. Fluorescence was read using the Tecan Infinite M1000 PRO. Effect of the drug was calculated by comparison to DMSO control, after subtraction of background signal. Sources for all compounds are listed in **Supplemental Table 1**.

# Metronidazole Resistant *E. histolytica*

E. histolytica resistant to MNZ were generated by continuous growth of E. histolytica HM-1:IMSS in MNZ, with steadily increasing concentrations from 1 to 15µM. Parasites growing stably at 15µM MNZ were used for further experiments. 20,000 amoebae from the resistant strain were seeded into a 96-well plate with media and 5µM drug, 20µM MNZ, or DMSO control and incubated for 72 h. Viability was assayed by FDA fluorescence as detailed above. Two independently generated MNZ resistant lines were used. A wild-type, MNZ sensitive strain was included as a control, with parasite seeding at 10,000 per well as in previous experiments.

# Assay for Effect of Drug on Viability on *E. invadens* Mature cysts

CK-luc parasites were induced to encyst by incubation in encystation media (47% LG). After 72 h, cells were harvested, washed once in distilled water, resuspended in water and incubated at 25◦C for 4–5 h to lyse trophozoites. Purified cysts were pelleted, counted to ensure equal cyst numbers and resuspended in encystation media at a concentration of 1– 5 × 10<sup>5</sup> cells per ml. One ml suspension per replicate was transferred to glass tubes containing encystation media and drug or DMSO, and incubated at 25◦C for 72 h. On the day of the assay, cysts were pelleted and treated once more with distilled water (5 h) to lyse any trophozoites that had emerged during treatment. Purified cysts were then resuspended in 75 µl Cell Lysis buffer (Promega) and sonicated for 2 × 10 s to break the cyst wall. Luciferase assay was performed using the Promega luciferase assay kit according to the manufacturer's instructions. Assays were performed on equal volume of lysate (35 µl) and not normalized to protein content. Effect of the drug was calculated by comparison to DMSO control, after subtraction of background signal.

# RESULTS

# Development and Validation of a High-Throughput Screen Against *Entamoeba* Trophozoites and Cysts

In order to discover novel drugs with activity against different life stages of Entamoeba, we designed a screen to simultaneously test compounds for activity against trophozoites and cysts. As E. histolytica cannot undergo encystation in vitro, we took advantage of a well-characterized model system for Entamoeba stage conversion, the reptile parasite E. invadens (Eichinger, 1997). A diagram of the workflow used for screening and identification of lead compounds is shown in **Figure 1.** To assay trophozoite growth, we used a strain with constitutive expression of luciferase. Parasites were seeded in 384-well plates with drug or vehicle control, sealed with plate sealers to reduce exposure to oxygen. After growth for 48 h, media was removed and luciferase activity was measured. Validation of the assay for well-to-well consistency and ability to detect parasite killing was performed using plates with no drug and plates with spike-in controls of varying concentrations of MNZ. We found significant (>3-fold) reduction in luciferase activity in wells treated with an active compound. Results are shown in **Figure 2A**; all wells with > 3-fold reduction in luciferase activity, compared to the plate median, were those treated with MNZ (indicated by L). No false positives were noted and little effect was seen from addition of up to 1% DMSO. The Z-factor, calculated using a plate with no drug, was 0.35.

For the encystation assay, mid-log stage trophozoites were harvested and resuspended in encystation media (47% LG), and seeded into 384-well plates containing drug or vehicle MNZ resistant E. histolytica parasites and mature E. invadens cysts.

FIGURE 2 | Assay optimization. Validation of trophozoite and cyst growth and drug screening assays. (A) Trophozoite growth: E. invadens trophozoites were seeded into 384-well plates containing media and either DMSO (0.2%) or MNZ (50 or 100µM) and assayed for luciferase activity at 48 h. Signal in each well was compared to the plate median and the percent inhibition plotted. Wells containing MNZ are indicated by: L). (B) Encystation: E. invadens trophozoites were pelleted and resuspended in either LYI or encystation media, then seeded at varying cell densities. After 48 h, plates were spun down, media removed, and calcofluor added to label cyst walls. The plate was imaged and quantified using MetaExpress software. Results of the quantification, expressed as integrated intensity, along with representative images from cyst and trophozoite wells are shown. Background signal coming from the trophozoite wells is very low, and signal in wells with cysts is proportional to the number of parasites seeded.

control. Parasites were incubated in this media for 48 h to allow cysts to form; the number of cysts was assayed by staining the cyst walls with calcofluor and performing high-throughput imaging to quantify fluorescently labeled cells. Trophozoites, which do not bind calcofluor, were used as a negative control. No signal was seen from wells with trophozoites (image shown in **Supplemental Figure 1**); wells with cysts showed increasing numbers of cysts based on the number of parasites in the initial seeding **(Figure 2B**). False positives were minimized by screening plates in duplicate.

Based on these results, we concluded that the assays were sufficiently robust and proceeded with high-throughput screening. Five libraries were screened: NIH clinical collection, Lopac 1280, Biomol FDA, Biomol known bioactives and Microsource spectrum, for a total of 3,444 unique compounds. NIHCC was screened in duplicate at ∼10µM, while all other libraries were screened in a 7-pt dose curve from 20 to 0.31µM (for the majority of the compounds), with the 20µM plate duplicated, to allow calculation of an EC50. Positive hits were defined as compounds with EC<sup>50</sup> <30µM (or >50% inhibition in both plates for NIHCC). Interestingly, we found that some compounds known to kill Entamoeba trophozoites, such as the recently identified drug auranofin (Debnath et al., 2012) produced anomalously high calcofluor staining. Visual inspection of phase images captured indicated that cell killing had occurred but with an increase in calcofluor staining, likely due to a stress response leading to chitin upregulation, as has been seen previously in other stress conditions (Field et al., 2000; Aguilar-Díaz et al., 2010) (**Supplemental Figure 1**). Based on these findings, we considered compounds with low signal (<50% control) as well as compounds with signal >150% over control to be hits for our encystation screen.

After a single round of screening, 167 compounds were found to have activity in at least one assay, with 22 compounds active against both trophozoites and cysts. Significantly, auranofin and MNZ were both present in the Biomol FDA approved library and identified as hits against trophozoites, providing further validation that our screen was able to identify compounds with anti-amebic activity. A total of 63 compounds, including MNZ, were chosen for further testing based on multiple factors including specificity, toxicity, low EC<sup>50</sup> and activity in both trophozoite and cyst assays; a second round of confirmation screening was performed on these compounds. Of the 63 compounds tested for confirmation, a total of 48 hits were confirmed positive, with the majority (32) being active only in the trophozoite assay. A total of 11 compounds with both trophozoite and cyst activity were identified. Results are in **Supplemental Table 2**.

# Confirmation of Hits in *E. histolytica* Trophozoites

As our original and confirmatory screens were performed using E. invadens, the next step in identifying potential antiamebic drugs was to determine efficacy in the human pathogen. E. histolytica trophozoites (strain HM-1:IMSS) were grown in sealed 96-well plates with varying levels of drug. After 72 h incubation, cells viability was assayed using the vital dye fluorescein diacetate (FDA). FDA is a cell permeant esterase substrate that is converted by viable cells to yield fluorescein. This probe measures both the enzymatic activity that is required to activate its fluorescence and also the cell-membrane integrity that is required to retain the fluorescence (Medzon and Brady, 1969). We chose FDA as a readout of parasite survival in the E. histolytica assay to ensure that the compounds we selected in the luciferase-based assay in E. invadens were not selected simply based on luciferase inhibition alone. Percent survival of parasites at 5 and 10µM drug (compared to DMSO control) are listed in **Table 1**. Using this assay, we calculated percent survival of parasites treated with MNZ as 96 and 24% at 5 and 10µM, respectively with a calculated EC<sup>50</sup> of ∼8–9µM. This is consistent with published reports of MNZ EC50, giving us confidence in our assay (Jarrad et al., 2016). Importantly, several drug classes that had promising results in E. invadens, including the tetracycline class compounds, showed very poor activity against E. histolytica. This discordance between trophozoite susceptibility in E. invadens and E. histolytica was surprising, and may be due to biological differences between the two species. Of interest, E. invadens is usually more robust (for example it needs a higher levels of drug treatment for drug selection with neomycin ; Singh et al., 2012), so the finding that some compounds that kill E. invadens have no activity against E. histolytica was somewhat unexpected. However, robust cell killing was seen with other compounds, including, as expected, the nitroimidazoles (nithiamide, ronidazole, ornidazole). Especially interesting were compounds, such as anisomycin and lycorine, that killed E. histolytica trophozoites and were also potent in our screen of E. invadens cysts.

In addition to candidate compounds selected from highthroughput screening, we also examined several compounds which were not in the libraries, but which have historical evidence of use against amebiasis (Balamuth and Brent, 1950; Woolfe, 1963). These include prodigiosin, conessine, and emetine. Of these, prodigiosin, a natural product isolated from the bacterium Serratia marcescens, had very strong activity against E. histolytica, (0% FDA signal at 5µM) and we elected to include this compound in further experiments.

# Target Validation of Protein Phosphatase 2a

Among the most potent compounds with confirmed activity against E. histolytica, was okadaic acid, a toxin derived from dinoflagellates, which has been shown to inhibit protein phosphatase 2a (PP2a) in mammalian systems (Xing et al., 2006). While okadaic acid itself is not a promising drug candidate due to its high toxicity and price, a different molecule targeting the same protein could potentially be developed as a therapeutic agent. With this in mind, we attempted to validate PP2a as the target of okadaic acid in Entamoeba. First, we searched the E. histolytica genome (amoebadb.org) for homologs of PP2a. We identified seven proteins with high similarity (e-value < 1e-100) to human PP2a, of which four had all amino acid residues shown to contact bound okadaic acid, including those



E. histolytica trophozoites were assayed for viability after 72 h treatment with 5 or 10µM of each drug using the vital dye FDA. Percent survival compared to DMSO, calculated EC<sup>50</sup> of E. invadens trophozoites, and percent inhibition (calcofluor signal compared to DMSO controls) of E. invadens cysts (mean of 20 and 10 µM data) are shown. Compounds with high calcofluor staining and confirmed killing by visual inspection are indicated as "enhancer."

thought to be responsible for its specificity for inhibition of PP2a vs. PP1 (∼100-fold difference in EC50). Gene names and alignments with Homo sapiens PP2a and PP1 are shown in **Supplemental Figure 2**.

We then tested the ability of other known PP2a inhibitors, cantharidin, calyculin A, and fostriecin to kill E. histolytica parasites (**Table 2**). Both fostriecin and calyculin A had antiamebic activity with EC<sup>50</sup> of 0.001 and 9.4µM respectively. A third compound, cantharidin, did not kill at concentrations up to 10µM; this may not be surprising as its activity against human PP2a is the weakest of all the compounds tested. Of particular interest was the activity of fostriecin, which does not inhibit HsPP1 (Swingle et al., 2007) but does have activity against E. histolytica. This result may indicate that PP2a, and not other phosphatases, is the target of calyculin, fostriecin and okadaic acid in Entamoeba. While further in vitro and genetic assays would be required to prove that PP2a is the true target of these compounds in Entamoeba, these results are promising and point to PP2a as a potential drug target for Entamoeba. PP2a inhibitors, including the compounds we tested, have potent activity against human PP2a, and have been noted for cytotoxic effects against cancer cells; for example, fostriecin has an EC<sup>50</sup> of ∼5µM against HCT-8 cells, a human adenocarcinoma cell line (Jackson et al., 1985). For this reason, further development efforts may be required to identify compounds that selectively inhibit E. histolytica PP2a.

# EC<sup>50</sup> Calculation of Select Lead Compounds

At this point, we removed two compounds from further experiments: okadaic acid, due to its high toxicity and general unsuitability as a human drug, and ornidazole, which is very similar to the two other nitroimidazole hits. All other compounds with <50% FDA signal at 10µM concentration (vs. control) were re-tested for E. histolytica killing in a dose curve ranging from 10 to 0.01µM, and EC<sup>50</sup> and EC<sup>90</sup> values calculated using the R package drc (Ritz et al., 2015). Results are shown in **Table 3**. Two non-nitroimidazole compounds, anisomycin and prodigiosin, were found to have EC<sup>50</sup> values below 1, significantly better than that of MNZ, potentially indicating that a lower dose would be needed for effective treatment. These positive results led us to test two additional compounds in our assay: preussin, an analog of anisomycin (Kasahara et al., 1997), and obatoclax, a synthetic derivative of prodigiosin, which is under development as a drug for lymphoma (Oki et al., 2012). Assaying derivatives of lead compounds can aid drug discovery efforts by identifying compounds with greater efficacy or improved medicinal chemical characteristics. In addition, it can help provide information about the drug target. While preussin had no activity at concentrations up to 10µM, obatoclax was similar in potency to prodigiosin (**Table 3)**. This result, along with the previous established use in human subjects, where it was well tolerated with mostly transient side effects (Schimmer et al., 2014), indicates the possibility of

TABLE 2 | PP2a inhibitors and their activity against E. histolytica trophozoites.


E. histolytica trophozoites were assayed for viability after 72 h treatment with varying concentrations (0.001–10µM) of the phosphatase inhibitors calyculin A, cantharidin, fostriecin or okadaic acid drug using the vital dye FDA and EC50s were calculated based on results from 3 independent experiments using the R package drc (Ritz et al., 2015). Potencies of each compound for inhibition of human PP2a and PP1 are included as reference (Swingle et al., 2007). Activity against HsPP2a is a good predictor of anti-amebic activity.

TABLE 3 | Potencies of lead compounds against E. histolytica trophozoites.


Selected lead compounds were used to treat E. histolytica trophozoites in a dose curve from 0.01 to 20µM for 72 h; parasites were assayed for viability using the vital dye FDA. EC<sup>50</sup> and EC<sup>90</sup> were calculated based on results from at least 3 independent experiments using the R package drc (Ritz et al., 2015).

developing a drug based on the prodigiosin structure as these compounds appear to be highly effective for killing Entamoeba, with a potential for a reasonable therapeutic index. Structures for all eight compounds are shown in **Supplemental Figure 3**.

# Characterization of Activity Against MNZ Resistant Parasites and Mature Cysts

Development of hits from screening into viable drug candidates for neglected diseases depends on multiple factors including in vitro efficacy, host toxicity, pharmacokinetic properties, and cost. Using these principles, we chose two main lead compounds: anisomycin and prodigiosin, for further characterization. Anisomycin was of particular interest based on reports of historical use as an anti-amebic agent, including human use (Gonzalez Constandse, 1956; Martin Abreu, 1962), indicating that it likely has an acceptable safety profile. Prodigiosin was chosen due to its potency at low doses and observed rapidity of action. These two compounds were assayed for activity against parasite strains and conditions refractory to current drugs: MNZ resistant trophozoites and mature cysts.

To determine activity of our lead compounds against drug resistant parasites, we grew E. histolytica trophozoites in incrementally increasing concentrations of MNZ (from 1 to 15µM). Once parasites had consistent growth at 15µM MNZ, they were subjected to our FDA viability assay as previously described. Cellular viability was compared to that of parasites treated with DMSO; a MNZ sensitive line was assayed at the same time and both strains were assayed for MNZ sensitivity to demonstrate that MNZ resistance has been achieved. Anisomycin and prodigiosin both killed MNZ resistant parasites at 5µM concentrations (**Figure 3**). The identification of two compounds with activity against MNZ resistant parasites is promising and suggests the possibility that anisomycin or prodigiosin, or related molecules, can be options in cases of clinical MNZ resistance.

Several compounds found in our screen, including anisomycin, strongly inhibited encysting parasites. We hypothesized that some compounds that inhibited early cysts may not be active against mature (48–72 h) cysts, as mature cysts have thick, chitin containing walls, which are impervious to many chemicals (Chatterjee et al., 2009). Therefore, we developed an assay for killing of E. invadens mature cysts based on loss of luciferase activity in parasites constitutively expressing luciferase. After allowing the parasites to develop in encystation media for 72 h, cysts were isolated and transferred to fresh encystation media containing drug or DMSO. After 3 days, cysts were treated with water for ∼5 h, to lyse any trophozoites that had emerged during drug incubation, and luciferase activity was assayed. Activity was standardized based on number of cysts at beginning of drug treatment, rather than total protein, to account for potential cell lysis during treatment.

Using this assay, we tested the activity of anisomycin, prodigiosin, and its analog obatoclax against mature cysts. We found that both prodigiosin and obatoclax significantly reduced cyst viability at 10µM concentrations (**Figure 4**). In contrast, MNZ did not affect mature cysts. Anisomycin, when tested at 10µM concentration, did not seem to have a substantial effect on mature cysts, despite its inhibition of encysting parasites at this concentration. When tested at 20µM, a slight reduction in luciferase signal was seen, although this was inconsistent between experiments (**Figure 4**). Overall, the results suggest that we can identify compounds, such as prodigiosin, that kill mature cysts; however, some drugs with activity against encysting parasites will have reduced potency against fully developed cysts.

# DISCUSSION

The treatment of parasitic diseases is frequently complicated by the fact that many parasites exist in multiple life stages with differing sensitivities to chemical agents. In the case of E. histolytica, the dormant cyst form is highly resistant to environmental stresses, as well as to the major drug used to treat amebiasis, MNZ. It would be ideal to have one drug that would potentially treat both forms of the parasite. To this end, we have performed a high throughput screen for anti-amebic compounds directed at both trophozoites and cysts. Using the reptile parasite E. invadens, a model for encystation, we identified 11 compounds with activity against both the trophozoite and cyst life cycle stages. After confirmation in E. histolytica trophozoites, we obtained five compounds

with good activity (EC<sup>50</sup> <25µM, >50% inhibition of cysts) against both E. histolytica trophozoites and E. invadens cysts. Surprisingly, although multiple examples of the tetracycline and quinolone groups of antibiotics were found in the initial screening with E. invadens, all of these compounds had poor activity in assays against E. histolytica. The two species do have significant differences at the genomic level, including little synteny and variant genome sizes. However, these two classes of antibiotics target biological processes likely to be highly conserved (protein synthesis and DNA topoisomerase). Differences in observed potency therefore may be due to factors such as uptake of the drug by the parasite, or possibly by assay conditions, such as temperature or pH.

One significant aim of a drug screening effort is the identification of proteins that are viable drug targets; this

#### TABLE 4 | Summary of results.


EC<sup>50</sup> of E. histolytica trophozoites, activity against E. invadens cysts and activity against MNZ resistant parasites are shown. Prodigiosin and obatoclax were considered to have the best activity against cysts, given their ability to kill mature cysts. nt, Not tested.

can lead to the discovery of additional molecules that have both the desired activity and have improved toxicity, cost, or other characteristics that make them superior drug candidates. In this screen, we characterized PP2a as a potential drug target in Entamoeba, due to the potent parasite killing of okadaic acid as well as several other known PP2a inhibitors. Interestingly, several PP2a inhibitors, including fostriecin and LB-100 (Lê et al., 2004; Hong et al., 2015), have been in Phase I clinical trials due to their potential as cancer treatments. Thus, there is the potential for further development of one of these compounds or other PP2a inhibitors as successful drugs, though delivery and dosage would need to be determined for any use against amebiasis. Alternatively, identification of compounds with greater affinity for EhPP2a than the human enzyme could lead to effective treatments with lower host toxicity.

Aside from the PP2a inhibitors, our screen produced a number of intriguing lead compounds **(Table 4)**, in particular the antibiotic anisomycin, a protein synthesis inhibitor isolated from Streptomyces. Anisomycin was ∼10-fold more potent than MNZ against E. histolytica trophozoites, as well as having activity against E. invadens cysts. Further research into this compound revealed previous historical use in humans, for both amebiasis and giardiasis, with good results, although in small cohorts (Gonzalez Constandse, 1956; Martin Abreu, 1962). Side effects were comparable to MNZ: vomiting, nausea, etc. Given the efficacy of anisomycin against MNZ resistant parasites, it could be a useful tool in the event of the emergence of drug resistance in Entamoeba. An additional compound, prodigiosin, had very potent activity including against mature cysts. Like anisomycin, there are historical references to prodigiosin as an anti-amebic agent (Balamuth and Brent, 1950); taken together these two cases make an argument for better mining of historical literature for potential anti-parasitic compounds, an approach that has recently proven very fruitful in the treatment of malaria (Tu, 2011). The actual target of prodigiosin in Entamoeba is unclear. It has been shown to induce apoptosis in cancer cells, and to alter mitochondrial membrane potential (Montaner et al., 2000; Francisco et al., 2007). While neither of these mechanisms is likely to be active in Entamoeba, the target proteins of these pathways may be conserved. Importantly, an analog of prodigiosin (obatoclax), which has been in human clinical trials, had significant activity against both trophozoites and cysts. Development of this molecule as a therapeutic may be possible, given the established safety record in patients.

In this work we present a high-throughput screen of Entamoeba encystation and demonstrate that compounds that are highly active against both Entamoeba trophozoites and cysts can be identified. Furthermore, we identified multiple compounds with improved efficacy compared to MNZ, and show that some of the compounds are active against MNZ resistant parasites indicating that they may target an alternate pathway. The work opens up the ability to screen high-value libraries against a neglected parasitic disease and increases the chances that new compounds that are highly efficacious against trophozoites, cysts, and drug resistant strains can be identified. The possibility of using new drugs that affect trophozoites only or cysts only could be enhanced by synergistic effect of the drugs on each. However, further studies are required to determine if using a combination of drugs that affect trophozoite only and cyst only could enhance the potency of the drugs against both forms of the parasite. Future work will focus on further characterization of the lead compounds as development as drug candidates, including lead optimization for toxicological and pharmacokinetic properties.

# AUTHOR CONTRIBUTIONS

GE and SS designed and performed experiments. DS-C provided libraries and helped with data analysis. US conceived of the project and aided in manuscript preparation.

# FUNDING

Funding for this project was provided by SPARK Translational Research Program at Stanford University; a training grant from the Stanford Translational Research and applied Medicine Center; The Child Health Research Institute, Lucile Packard Foundation for Children's Health; and through the NIH (R21-AI123594 to US, and National Center for Advancing Translational Science Clinical and Translational Science Award UL1-TR001085). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

# ACKNOWLEDGMENTS

We would like to thank Jason Wu in the Stanford HTBC for technical help, and Monica Kangussu-Marcolino, Dipak Manna and other members of the Singh lab for advice. In addition, invaluable consultation was provided by SPARK advisors, especially Steve Schow, Robert Greenhouse, Nancy Federspiel, and Kevin Grimes.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00276/full#supplementary-material

# REFERENCES


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Ehrenkaufer, Suresh, Solow-Cordero and Singh. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Bioactivity of Farnesyltransferase Inhibitors Against Entamoeba histolytica and Schistosoma mansoni

Alexandra Probst <sup>1</sup> , Thi N. Nguyen<sup>1</sup> , Nelly El-Sakkary <sup>1</sup> , Danielle Skinner <sup>1</sup> , Brian M. Suzuki <sup>1</sup> , Frederick S. Buckner <sup>2</sup> , Michael H. Gelb<sup>3</sup> , Conor R. Caffrey <sup>1</sup> \* and Anjan Debnath<sup>1</sup> \*

<sup>1</sup> Center for Discovery and Innovation in Parasitic Diseases, Skaggs School of Pharmacy and Pharmaceutical Sciences, University of California, San Diego, La Jolla, CA, United States, <sup>2</sup> Division of Allergy and Infectious Diseases, Department of Medicine, Center for Emerging and Reemerging Infectious Diseases, University of Washington, Seattle, WA, United States, <sup>3</sup> Departments of Chemistry and Biochemistry, University of Washington, Seattle, WA, United States

The protozoan parasite Entamoeba histolytica can induce amebic colitis and amebic liver abscess. First-line drugs for the treatment of amebiasis are nitroimidazoles, particularly metronidazole. Metronidazole has side effects and potential drug resistance is a concern. Schistosomiasis, a chronic and painful infection, is caused by various species of the Schistosoma flatworm. There is only one partially effective drug, praziquantel, a worrisome situation should drug resistance emerge. As many essential metabolic pathways and enzymes are shared between eukaryotic organisms, it is possible to conceive of small molecule interventions that target more than one organism or target, particularly when chemical matter is already available. Farnesyltransferase (FT), the last common enzyme for products derived from the mevalonate pathway, is vital for diverse functions, including cell differentiation and growth. Both E. histolytica and Schistosoma mansoni genomes encode FT genes. In this study, we phenotypically screened E. histolytica and S. mansoni in vitro with the established FT inhibitors, lonafarnib and tipifarnib, and with 125 tipifarnib analogs previously screened against both the whole organism and/or the FT of Trypanosoma brucei and Trypanosoma cruzi. For E. histolytica, we also explored whether synergy arises by combining lonafarnib and metronidazole or lonafarnib with statins that modulate protein prenylation. We demonstrate the anti-amebic and anti-schistosomal activities of lonafarnib and tipifarnib, and identify 17 tipifarnib analogs with more than 75% growth inhibition at 50µM against E. histolytica. Apart from five analogs of tipifarnib exhibiting activity against both E. histolytica and S. mansoni, 10 additional analogs demonstrated anti-schistosomal activity (severe degenerative changes at 10µM after 24 h). Analysis of the structure-activity relationship available for the T. brucei FT suggests that FT may not be the relevant target in E. histolytica and S. mansoni. For E. histolytica, combination of metronidazole and lonafarnib resulted in synergism for growth inhibition. Also, of a number of statins tested, simvastatin exhibited moderate anti-amebic activity which, when combined with lonafarnib, resulted in slight synergism. Even in the absence of a definitive molecular target, identification of potent anti-parasitic tipifarnib analogs encourages further exploration while the synergistic combination of metronidazole and lonafarnib offers a promising treatment strategy for amebiasis.

Keywords: Entamoeba histolytica, Schistosoma mansoni, farnesyltransferase, metronidazole, lonafarnib, tipifarnib, statin, chemotherapy

#### Edited by:

Tomoyoshi Nozaki, University of Tokyo, Japan

#### Reviewed by:

David Leitsch, Medical University of Vienna, Austria Elisa Azuara-Liceaga, Universidad Autónoma de la Ciudad de México, Mexico

#### \*Correspondence:

Conor R. Caffrey ccaffrey@ucsd.edu Anjan Debnath adebnath@ucsd.edu

#### Specialty section:

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

> Received: 08 January 2019 Accepted: 09 May 2019 Published: 29 May 2019

#### Citation:

Probst A, Nguyen TN, El-Sakkary N, Skinner D, Suzuki BM, Buckner FS, Gelb MH, Caffrey CR and Debnath A (2019) Bioactivity of Farnesyltransferase Inhibitors Against Entamoeba histolytica and Schistosoma mansoni. Front. Cell. Infect. Microbiol. 9:180. doi: 10.3389/fcimb.2019.00180

# INTRODUCTION

Entamoeba histolytica is a non-flagellated protozoan parasite exclusive to humans that has a simple life cycle comprising an infective cyst stage and an invasive trophozoite form (Petri and Singh, 1999; Stanley, 2003). Infection with E. histolytica can lead to three major outcomes: (a) asymptomatic colonization, (b) intestinal amebiasis, most commonly amebic colitis, and (c) extra-intestinal amebiasis with liver abscess being the most common complication (Petri and Singh, 1999). Amebiasis causes up to 110 thousand deaths annually and is estimated to be the second most common cause of parasite infection-related mortality worldwide (Petri and Singh, 1999; Lozano et al., 2012; Watanabe and Petri, 2015). Each year 40 to 50 million cases of amebic colitis and liver abscess are reported with high prevalences in Central and South America, Africa, and Asia (Petri and Singh, 1999).

Amebic infection is initiated by ingestion of E. histolytica cysts in fecally contaminated food or water. These cysts excyst in the intestine to form trophozoites, which degrade the mucous layer via cysteine protease activities, destroy and ingest epithelial cells via trogocytosis, and invade the lamina propria, which leads to colitis and liver abscesses in the case of invasion of the blood vessels (Petri, 2002; Stauffer and Ravdin, 2003; Watanabe and Petri, 2015).

First-line drugs for the treatment of invasive amebiasis are the nitroimidazoles, in particular metronidazole, which is given orally to adults in three doses of 750 mg (total 2,250 mg/day) per day for 7–10 days (Haque et al., 2003). Nitroimidazole compounds carry a nitro group on the 5-position of the imidazole ring. As prodrugs, that must be activated by reductases of the parasite. After entering the trophozoite, reduced ferredoxin donates electrons to the nitro group of the prodrug, which is then reduced to toxic radicals. Covalent binding to DNA macromolecules results in DNA damage and killing of the parasites (Muller, 1983; Edwards, 1993). Nitroreductases and thioredoxin reductase are also known to reduce nitroimidazole drugs in Entamoeba (Leitsch et al., 2007).

Potential resistance of E. histolytica to metronidazole remains a major concern (Samarawickrema et al., 1997; Wassmann et al., 1999) and in the absence of a back-up drug, it is important to search for alternative antimicrobials against E. histolytica.

Schistosomiasis is caused by various species of the Schistosoma flatworm that reside in the venous system. Infection is found in populations living close to freshwater bodies that harbor the appropriate vector snail. With as many as 200 million people infected (Hotez, 2018) and possibly over 700 million at risk (King, 2010), infections can be chronic and painful as a consequence of progressive tissue and organ damage due to the parasite's eggs. The disease impacts school attendance and performance, the ability to work, and, consequently, it has been considered a direct contributor to poverty (Hotez et al., 2008; Utzinger et al., 2011). Treatment and control of schistosomiasis relies on just one drug, praziquantel. Though safe and reasonably effective, the drug is rarely curative and is less effective against immature parasites (Caffrey, 2007, 2015). The possibility of resistance, particularly as dissemination of the drug is increasing (http://unitingtocombatntds.org/wp-content/themes/tetloose/ app/staticPages/fifthReport/files/fifth\_progress\_report\_english. pdf, 2014) is a constant concern, and alternative drugs would be welcome.

The mevalonate metabolic pathway is vital for diverse functions in parasitic protozoa and helminths such as sterol synthesis and cell growth (Li et al., 2013; Rojo-Arreola et al., 2014; Millerioux et al., 2018). 3-hydroxy-3-methyl-glutaryl-coenzyme A reductase (HMGR) is the rate-limiting enzyme in the pathway which catalyzes the conversion of HMG-CoA into mevalonate (Edwards and Ericsson, 1999). HMGR inhibitors, also known as statins, prevent the conversion of HMG-CoA to L-mevalonate resulting in the inhibition of the downstream sterol biosynthesis (Gazzerro et al., 2012). Mevalonate is also a precursor of isoprenoid groups, in addition to more than a dozen classes of end products (Goldstein and Brown, 1990). The last common substrate for the synthesis of the end products of the mevalonate pathway is farnesyl pyrophosphate (FPP, also referred to as farnesyl diphosphate FDP). It is the natural substrate of farnesyl transferase (FT), which catalyzes the transfer of a farnesyl moiety from FPP to proteins. Farnesylated proteins include Ras and Ras related GTP-binding proteins, nuclear lamins, centromereassociated proteins, tyrosine phosphatases, and co-chaperones (Zhang and Casey, 1996). FT catalyzes the prenylation of proteins through a thioether linkage and the prenylation may be concluded by the palmitoylation of cysteine residues for some proteins. Due to the lipids involved in the mechanism and their hydrophobicity, prenylation leads to membrane interactions by the proteins and plays an important role in the signal transduction pathway for cell differentiation (Zhang and Casey, 1996). A previous characterization of E. histolytica FT showed that the amebic FT did not utilize a majority of Ras and Rap as substrates, but only one Ras protein, Ras4, was farnesylated (Kumagai et al., 2004). Schistosoma mansoni Ras was also found to be farnesylated and inhibition of farnesylation in S. mansoni extract was achieved using an FT inhibitor (FTI) (Osman et al., 1999). The deduced amino acid sequences of the β-subunit of both E. histolytica and S. mansoni FT are 36 and 43% identical to the β-subunit of human FT. There is evidence that targeting the farnesyltransferase enzyme in protozoan parasites leads to inhibition of protein prenylation and severely impairing growth, including Plasmodium falciparum (Ibrahim et al., 2001; Buckner et al., 2002; Chakrabarti et al., 2002; Carrico et al., 2004; Esteva et al., 2005). Here, we demonstrate the anti-parasitic activities of known FTIs lonafarnib and tipifarnib against E. histolytica, and S. mansoni somules (post-infective larvae) and adults. In addition, we identified tipifarnib analogs with anti-amebic and anti-schistosomal activities. The combination of lonafarnib with the currently used anti-amebic drug, metronidazole, generated a synergistic growth inhibition of E. histolytica.

# MATERIALS AND METHODS

# Chemicals and Reagents

Assay plates (format includes: 96- and 24-well; flat and U-bottomed; and transparent and white) were purchased from VWR International (Radnor, PA). The CellTiter-Glo luminescent cell viability assay was acquired from Promega (Madison, WI); DMSO, metronidazole, lonafarnib and simvastatin were purchased from Sigma-Aldrich (St. Louis, MO). The tipifarnib analogs (**Figure 1**) were previously synthesized as part of a program to develop FTIs active against Trypanosoma brucei and Trypanosoma cruzi, the causative agents of Human African Trypanosomiasis and Chagas disease, respectively (Kraus et al., 2009, 2010). Compounds were dissolved in DMSO at a concentration of 10 mM and stored at −20◦C.

# Maintenance of Entamoeba histolytica

E. histolytica trophozoites (strain HM1:IMSS) were maintained axenically in TYI-S-33, supplemented with penicillin (100 U/mL), streptomycin (100µg/mL) and 10% heat inactivated adult bovine serum as previously described (Diamond et al., 1978). The cells were maintained in the logarithmic phase of growth by routine passage every 2 days and the logarithmic phase of growth was determined by counting the cells using a hemocytometer.

# Maintenance of Schistosoma mansoni

The NMRI isolate of S. mansoni was maintained by passage through Biomphalaria glabrata snails and 3–5 week-old, female Golden Syrian hamsters (Charles River, San Diego, CA) as intermediate and definite hosts, respectively (Abdulla et al., 2009; Long et al., 2016). A dose of 600 infective larvae (cercariae) was used to infect hamsters. The acquisition, preparation and in vitro maintenance of S. mansoni post-infective larvae (schistosomula or somules) and adults have been described (Abdulla et al., 2009; Štefanic et al., 2010 ´ ). Vertebrate maintenance and handling at the University of California San Diego Animal Care Facility were in accordance with protocols approved by the university's Institutional Animal Care and Use Committee (IACUC).

# First-Pass Screening of the FTIs and Statins for Activity Against Entamoeba histolytica

A first-pass cell viability assay was performed with compounds at a concentration of 50µM against E. histolytica (Debnath et al., 2014). Briefly, 0.5 µL of 10 mM (FTIs) or 20 mM (statins) stock compounds were plated into white 96-well flat-bottom plates in duplicate and 5,000 trophozoites in 99.5 µL TYI-S-33 medium were added to the 96-well plates. As a positive control, 50µM of metronidazole were plated; negative controls contained 0.5% of DMSO. Cell culture plates were incubated at 37◦C for 48 h in the GasPak EZ gas-generating anaerobe pouch system (VWR). After incubation, 50 <sup>µ</sup>L of CellTiter-Glo <sup>R</sup> were added. The luminescent signal, resulting from the lysis of the cells was measured by an EnVision luminometer, and can be converted into the percentage of inhibition of the cell growth relative to maximum and minimum reference signal controls by using the following equation (Debnath et al., 2012):

# Dose-Response Assays of the FTIs Lonafarnib and Tipifarnib, Simvastatin as Well as Metronidazole Against Entamoeba histolytica

Dose-response assays were implemented as confirmatory screens of FTIs, simvastatin and metronidazole against E. histolytica trophozoites. An 8-point EC<sup>50</sup> determination (half maximal effective concentration) was performed as follows: 0.5 µL of 10 mM (FTIs) or 20 mM (simvastatin) stock compounds was plated in triplicate into white 96-well flat-bottom plates to obtain a starting concentrations of 50 or 100µM. Compounds were then serially diluted in triplicate to obtain a concentration range of 50 to 0.39µM for FTIs and 100 to 0.78µM for simvastatin. The addition of the cells, the incubation and the reading were performed as described in the first-pass assay. Graph, EC<sup>50</sup> calculations and standard error (fitting method: least square (ordinary) fit) were obtained by using GraphPad Prism (San Diego, CA) (Debnath et al., 2012).

# Mammalian Cytotoxicity Assay

Tipifarnib was screened for cytotoxicity against the fibroblast 3T3 cell line. Cells were grown in the presence of tipifarnib for 48 h before growth was quantified using Alamar Blue (Alamar Biosciences Inc., Sacramento, CA) (Kraus et al., 2009). Tipifarnib was tested at final concentrations of 100, 50, 25, 12.5, 6.25, 3.13, 1.56, and 0.78 µM.

# Entamoeba histolytica Invasion Assay

This was performed as described previously (Emmanuel et al., 2015). Briefly, 75,000 E. histolytica trophozoites were preincubated for 3 h with 0.5% DMSO, 14 and 25µM of tipifarnib and after 3 h, cells were re-suspended in serum-free TYI medium and loaded in the upper chamber of a Corning BioCoat Matrigel Invasion Chamber (Corning). The lower chamber contained TYI medium supplemented with 10% adult bovine serum (Sigma-Aldrich). The Matrigel Invasion Chamber was incubated at 37◦C for 48 h in a GasPak EZ gas-generating anaerobe pouch system. At the end of incubation, images of trophozoites in the lower chamber were captured using a 10 × objective lens fitted in a Zeiss Axiovert A1 inverted microscope. Trophozoites that had migrated into the lower chamber were also counted using a hemocytometer. The data were obtained from three independent experiments each performed in triplicate and the percentage invasion of trophozoites was calculated and plotted by using GraphPad Prism.

# Combination Assays

Because the activity of tested FTIs and simvastatin against E. histolytica was moderate when administered alone, we tested if a combination of lonafarnib with metronidazole or an HMG-CoA reductase inhibitor (simvastatin) would have a better activity against E. histolytica in vitro than a single compound. Drugs were combined at constant ratios 1:1, 1:2, 1:4, 1:8,

$$\% \text{ Inh.} = \left( 1 - \left( \frac{\text{[hunnicsxent output (RLU) of compound x} - \text{man of lumimexent output (RLU) of positive control}}{\text{(mean of lumimexent output (RLU) of negative control } - \text{man of lumimexent output (RLU) of positive control)}} \right) \right) \times 100 \text{ cm}^2 \text{ [m]}^2 \text{ or [m]}^2 \text{ or [m]}^2 \text{ or [m]}^2 \text{ or [m]}^2 \text{ or [m]}^2 \text{ or [m]}^2 \text{ or [m]}^2 \text{]}^2$$

2:1, 4:1, 8:1, 16:1 [µM lonafarnib/µM metronidazole and µM lonafarnib/µM simvastatin]. Growth inhibition was determined with the CellTiter-Glo <sup>R</sup> luminescent cell viability assay, as described previously (Debnath et al., 2012). Quantitative drug interaction (synergy) was determined by using the software CompuSyn (Chou and Talalay, 1984; Chou, 2006).

# Phenotypic Evaluation of the Effect of Combination of Lonafarnib and Metronidazole on Entamoeba histolytica Trophozoites

The assay was performed using transparent 96-well flat-bottom plates. DMSO was used as a negative control at a final concentration of 0.5%. Metronidazole was plated at 50µM as a positive control. Lonafarnib and metronidazole were combined at ratios of 4:1, 2:1, 1:1, 1:2, and 1:4. Parasites were plated and incubated for 48 h as described in the first pass and dose response assay for E. histolytica (Debnath et al., 2014). Images were captured after 24 and 48 h post-incubation using a Zeiss Axiovert A1 inverted microscope (10×, 20×, 40× and 63× objective) and a Zeiss AxioCam 503 mono digital camera controlled by the Zen 2 lite software (Version 2.0.0.0).

# Phenotypic Screening of FTIs With Different Developmental Stages of Schistosoma mansoni

Screens were performed using post-infective larvae (somules) and adult parasites, as described (Abdulla et al., 2009; Rojo-Arreola et al., 2014; Long et al., 2016) using transparent Ushaped 96-well plates and flat-bottom 24-well plates, respectively. Given the relatively large number of compounds to screen, the entire collection was screened against somules to identify the most active compounds for subsequent screening against adult parasites, as conducted previously (Abdulla et al., 2009). Somules are harvestable in their thousands from vectors snails whereas adults are only recoverable from infected hamsters in limited numbers (∼10–20% return on the infecting 600 cercarial dose used).

Somules (40 units/well) were prepared in 100 µL Basch medium (Basch, 1981) supplemented with 5% FBS, 100 U/mL penicillin and 100µg/mL streptomycin. Compounds were then added at 2x the final concentration in 100 µL of the same medium to yield final concentrations of 5 or 10µM compound and 0.5% DMSO. Parasites were incubated at 37◦C in a 5% CO<sup>2</sup> environment and phenotypic changes recorded at 24 and 48 h. Adult 42-day-old S. mansoni (5 males and including approximately 2 pairs) was maintained in 2 ml of the same Basch medium under the same conditions in the presence of 10µM compound and 0.1% DMSO. Phenotypic changes were recorded at 1, 5, 24, and 48 h.

Phenotypic changes are recorded as described (Abdulla et al., 2009; Rojo-Arreola et al., 2014; Glaser et al., 2015; Weeks et al., 2018). Briefly, we employ simple descriptors that describe the effects of compounds on the parasites (changes in shape, motility, and density). To allow for comparisons of compound activity, each descriptor is awarded a "severity score" of 1 and these are added up to a maximum score of 4. Evidence of degeneracy or death is awarded the maximum score of 4, and, for adults specifically, non-adherence by the oral or ventral suckers to the well surface (a score of 1) is taken into account as is damage to the outer surface (tegument; a score of 4) on the understanding that such damage is lethal in vivo to the parasite (Andrews et al., 1983). Images were captured via a Zeiss AxioCam 105 color digital camera that was attached to a Zeiss Axiovert A1 inverted microscope and controlled by ZEN 2 lite software (Version 2.0.0.0).

# RESULTS

# First-Pass Screening of FTIs (Lonafarnib, Tipifarnib, 125 Analogs of Tipifarnib) and Statins Against Entamoeba histolytica

**Table 1** lists the most active statins and FTIs that were obtained from the first pass screening against E. histolytica trophozoites (>75% growth inhibition). From the five statins tested (mevastatin, atorvastatin, fluvastatin, pitavastatin, simvastatin), only simvastatin exhibited 94% growth inhibition at 100µM. Lonafarnib, tipifarnib and 17 tipifarnib analogs (**Table 1**) demonstrated >75% of inhibition at 50µM after an incubation period of 48 h. Both lonafarnib and tipifarnib demonstrated TABLE 1 | Activity of statins at 100µM and tipifarnib analogs at 50µM against E. histolytica.

TABLE 2 | EC50 values of simvastatin, FTIs, tipifarnib analogs, and metronidazole against E. histolytica.


almost 100% inhibitory activity and one tipifarnib analog, HB-24 showed 98.9% inhibition at 50µM. The results of the first-pass screening against E. histolytica trophozoites for all 125 tipifarnib analogs are shown in **Supplementary Table 1**. Activities were stratified into four groups depending on the percentage growth inhibition of E. histolytica trophozoites. Compounds showing 75–100% growth inhibition belonged to group A, group B showed 50–74% inhibition, group C had 25– 49% growth inhibition and group D exhibited 0–24% growth inhibition. In referencing to the generic structure shown in **Figure 1** (right), most of the active compounds contain R<sup>3</sup> = NH<sup>2</sup> (11 of 17). For the analogs, ring 1 was either left the same as tipifarnib (4-Cl) or modified to 4-CH3, 4-CF3, 3-Cl, or several other variants. There were no apparent trends associating specific ring 1 substituents with more potent E. histolytica activity. With respect to ring 2, compounds with the 3-Cl-phenyl (as occurs with tipifarnib) were found in groups 1, 2, 3, and 4.

# Entamoeba histolytica Concentration Response Assay

As simvastatin and FTIs tipifarnib and lonafarnib were among the most active compounds in the primary screen, they were selected for further testing. Some of the active tipifarnib analogs containing NH2 at R<sup>3</sup> were also assessed for EC50. The EC<sup>50</sup> of simvastatin was 50µM, whereas both tipifarnib and lonafarnib generated EC<sup>50</sup> values of about 14µM. The EC<sup>50</sup> value for tipifarnib on a mammalian fibroblast cell line 3T3 was 35.1µM. This provided a selectivity index of about 2.5. The most active tipifarnib analog in the primary screen, HB-24, generated an EC<sup>50</sup> value similar to those of lonafarnib or tipifarnib (**Table 2**). The dose response and EC<sup>50</sup> data for simvastatin, tipifarnib, lonafarnib, and metronidazole are displayed in **Figure 2** and **Table 2**, respectively.


<sup>a</sup>EC<sup>50</sup> minimum n = 3.

FIGURE 2 | Concentration-dependent inhibition of growth of E. histolytica by tipifarnib, lonafarnib, and simvastatin, as compared to metronidazole. Different concentrations of compounds were tested in triplicate for activity against E. histolytica trophozoites. The data points represent mean percentage growth inhibition and standard error of mean (SEM) of different concentrations of metronidazole, tipifarnib, lonafarnib and simvastatin. EC50 curves were generated from mean values of percentage growth inhibition ± SEM of metronidazole, tipifarnib, lonafarnib, and simvastatin against E. histolytica.

# Entamoeba histolytica Invasion Assay

To test whether tipifarnib influences invasiveness of E. histolytica, we used a transwell matrigel invasion assay. The data show that a 3 h pre-incubation of trophozoites with tipifarnib at its EC<sup>50</sup> concentration (14µM) or at 25µM decreased the invasion of trophozoites through the matrigel after 48 h compared to 0.5% DMSO-treated trophozoites (**Figure 3**). Specifically, whereas 68.7 ± 5.7% of the DMSO-treated trophozoites penetrated the matrigel and were recovered in the lower chamber, just 23.2 ± 5.7% (p = 0.005) and 7.3 ± 2.3% of the cells (p = 0.0005) that had been exposed to 14 and 25µM tipifarnib were recovered, respectively.

# Synergy Assay of Lonafarnib and Metronidazole Against Entamoeba histolytica

To measure the effect of combining lonafarnib and metronidazole, drugs were combined at constant micromolar ratios and the inhibition of E. histolytica trophozoites growth was

determined using the CellTiter-Glo <sup>R</sup> luminescent cell viability assay. Synergism was observed for all combinations of lonafarnib and metronidazole at 50, 75, 90, and 95% of growth inhibition, respectively. Synergism at high effect level is much more relevant than at low effect level, therefore a 95% growth inhibition of E.

histolytica trophozoites upon drug administration was chosen to be presented in **Table 3**. All tested combinations of lonafarnib and metronidazole showed synergism, given the combination index (CI) values being <1. According to Chou (Chou, 2006), CI values ranging from 0.85 to 0.9 indicate slight synergism,


TABLE 3 | Summary of synergism assay with lonafarnib and metronidazole, shown for 95% growth inhibition of E. histolytica trophozoites.

Mass-action law based computer software "CompuSyn" was used for automated data analysis. Experiments were carried out in triplicate. CI stands for combination index, DRI for dose reduction index.

values in the range of 0.7 to 0.85 and 0.3 to 0.7 indicate moderate synergism and synergism, respectively. Combining lonafarnib and metronidazole led to a reduction of each dose (Dose Reduction Index or DRI) at given effect level compared with the doses of each drug alone.

# Phenotypic Evaluation of the Effect of the Synergy Experiment With Lonafarnib and Metronidazole on Entamoeba histolytica Trophozoites

Trophozoites treated with 7.4µM of lonafarnib (**Figure 4A**) showed a similar morphology and movement to 0.5% DMSO-treated control trophozoites (**Figure 4D**) after 24 h of incubation. This concentration of lonafarnib is almost half of the concentration needed for a 50% cell growth inhibition (EC<sup>50</sup> = 14.5µM). Treatment with metronidazole alone (1.9µM; **Figure 4B**) evoked the same phenotypic response as reported for lonafarnib. When lonafarnib and metronidazole were administered in a combination ratio of 4:1 (7.4µM lonafarnib: 1.9µM metronidazole) (**Figure 4C**), lysis of the cells was induced and completed by 24 h post-incubation. The cell death observed at 24 h in lonafarnib-metronidazole combination experiment was comparable to the death induced by 50µM of metronidazole alone (**Figure 4E**).

# Effect of Combination of Lonafarnib and Simvastatin Against Entamoeba histolytica

The inhibitory effects of lonafarnib and simvastatin were estimated by ATP-bioluminescence assay at fixed concentration ratios, and their dose-effect relationships were assessed by Chou-Talalay combination index (CI) method using CompuSyn software. Only two ratios of lonafarnib and simvastatin (1:1 and 2:1) showed slight synergism to additivity with CI ≤ 1 for 50, 75, 90, and 95% of growth inhibition. The combination of lonafarnib and simvastatin at 1:1 and 2:1 achieved 95% growth inhibition with 1.2- to 1.4-fold dose reduction for lonafarnib and 4- to 6.4-fold dose reduction for simvastatin (**Table 4**).

# Phenotypic Screening of FTIs Against Schistosoma mansoni

Because the anti-schistosomal activity of statins, including mevastatin, atorvastatin, fluvastatin and simvastatin, has already been demonstrated against S. mansoni somules and adults (Rojo-Arreola et al., 2014; Asarnow et al., 2015), the phenotypic screening in this study focused on the FTIs.

For somules, phenotypic changes in the presence of the FTIs were assessed visually after 24 and 48 h at 5 and 10 µM (**Figure 5** as a representative example), and the descriptors recorded were converted to severity scores to allow for comparison of compound activities. Based on the combination of time- and dose-dependency of activity, the compounds could be ranked into four groups (Groups A-D; **Supplementary Table 2**). Group A comprised the 17 most active compounds that generated severity scores of 3 or 4 at either 5 or 10µM after 48 h (**Table 7**). These included lonafarnib and tipifarnib. Group B comprised 47 compounds that were maximally active (score of 4) at both 24 and 48 h at 10µM, but elicited little or no activity at 5µM. The 34 Group C compounds were those for which severity scores of 1–3 were recorded at 10µM after 24 h and, finally, the 28 Group D compounds were those that registered zero scores at 10µM after 24 h. Throughout, the most common phenotypic responses recorded were loss of translucency (darkening) and degeneracy sometimes leading to death.

The most active 17 Group A compounds were subsequently screened at 10µM against 42-day-old adult S. mansoni and phenotypes recorded at 1, 5, 24, and 48 h. All were active to varying degrees (**Table 7**, **Supplementary Table 2**, **Supplementary Videos 1**, **2**). Thus, the most common phenotypic responses of the parasites noted at 1 and 5 h was an uncoordinated motion coupled with an inability of the parasite to adhere via its oral or ventral suckers to the floor of the well (severity score of 2). By 24 and 48 h, for the 12 most active compounds, these responses had progressed to degeneracy, worm

shrinkage, decreased motility and blebbing of the surface tegument (severity scores of 3 or 4). Among the Group A compounds, and apart from lonafarnib and tipifarnib, only CHN-13, JK-02, JK-17, and JK-19 overlapped with those compounds that were most active against E. histolytica (**Table 1**).

Analysis of the physicochemical properties of all 125 compounds indicated that there was a statistically significant trend for compounds with higher cLogP values being associated with better E. histolytica and S. mansoni activity (**Tables 5**, **6**).

# DISCUSSION

We report the anti-parasitic activities of FTIs against two phylogenetically distinct parasites, the protozoan, E. histolytica, and the metazoan flatworm, S. mansoni. FT has gained much attention as a target for pharmaceutical development and a number of FTIs have been tested as anti-cancer agents (Bagchi et al., 2018). Both lonafarnib and tipifarnib are potent inhibitors of mammalian FT and have been extensively tested in clinical trials for malignancies (Martinelli et al., 2008; Moorthy et al.,


Mass-action law based computer software "CompuSyn" was used for automated data analysis. Experiments were carried out in triplicate. CI, combination index; DRI, dose reduction index.

FIGURE 5 | S. mansoni somules incubated for 48 h as described in the text. (A) DMSO control; (B) in the presence of 10µM tipifarnib—note the rounding and varying degrees of degeneracy of the parasites. Scale bar = 200 µm.

TABLE 5 | Average cLogP values for FTIs according to group\* in E. histolytica.


\*Groups A, B, and C each have cLogP values that are significantly higher than group D (P < 0.05, ANOVA).


\*Groups A, B, and C each have cLogP values that are significantly higher than group D (P < 0.001, ANOVA).

2013). Lonafarnib has also undergone clinical trials in children with progeria (Gordon et al., 2016). Thus, in an observational cohort study in patients with progeria, lonafarnib treatment led to a significantly reduced mortality rate after 2.2 years of follow-up (Gordon et al., 2018).

The role of protein farnesylation has been previously documented in E. histolytica (Kumagai et al., 2004). Specifically, the E. histolytica Ras4 (a small GTPase mediating cell proliferation/differentiation) was identified as a unique farnesyl acceptor for the E. histolytica FT (Kumagai et al., 2004). Molecular characterization of protein farnesyltransferase in E. histolytica demonstrated that α and β subunits are wellconserved signature domains shared by other organisms (Kumagai et al., 2004).

The studies reported herein demonstrate that two FTIs, lonafarnib and tipifarnib, were active against E. histolytica in culture with EC<sup>50</sup> values of ∼14µM. Because the compounds are less potent than the current standard metronidazole therapy, we screened a library of 125 tipifarnib analogs against E. histolytica in culture to identify compounds with greater potency and a structure-activity relationship (SAR). This library was originally assembled to target the protozoan parasite, Trypanosoma cruzi, and comprised compounds with known FTI activity as well as compounds with structural changes that abrogated that activity (Kraus et al., 2009, 2010). The changes were intended to reduce activity on the mammalian FT in order to avoid potential sideeffects, while retaining anti-parasitic activity that was mediated through inhibition of the T. cruzi CYP51 enzyme (essential for ergosterol metabolism) (Kraus et al., 2009). The methylimidazole attached to the central sp3 carbon (**Figure 1**) is integral to the activity against both FT and sterol 14-alpha-demethylase (CYP51), and is retained in all of the analogs. However, changes to other substituents at positions R1–R3 (**Figure 1**, right) resulted in decreased FTI activity compared to tipifarnib (mammalian FT IC<sup>50</sup> value of 0.7 nM). Specifically, introduction of a 2-methyl group on Ring 2 or replacing the tipifarnib Ring 2 with a naphthyl group yielded an IC<sup>50</sup> value of 294 or 485 nM, respectively, against mammalian FT (Kraus et al., 2009). Combining the Ring 2 changes with a change of the NH<sup>2</sup> to OMe at the R3 position essentially abrogated FTI activity (mammalian FT IC<sup>50</sup> of >5,000 nM).

Previously reported whole-organism studies in vitro with tipifarnib demonstrated moderate bioactivity (EC<sup>50</sup> = 6µM) against the protozoan parasite, Trypanosoma brucei (Buckner et al., 2005), but much more potent activity against Trypanosoma cruzi (EC<sup>50</sup> = 0.004µM) (Kraus et al., 2010). Against the Trypanosoma brucei FT recombinantly expressed in E. coli (Buckner et al., 2000), tipifarnib analogs with R<sup>3</sup> = NH<sup>2</sup> (e.g., JK-02, JK-12, JK-13, JK-17, and JK-19), i.e., those that were active against E. histolytica, possess weak to negligible activity (data not published). Further, we previously demonstrated that 11 of the 17 most active tipifarnib analogs against E. histolytica and which contained the same NH<sup>2</sup> R<sup>3</sup> group (**Supplementary Table 1**) exert their activity against T. cruzi via inhibition of CYP51 (Kraus et al., 2009), yet, a CYP51 ortholog is not found in the E. histolytica genome (searching E.C. #1.14.13.70). Overall, therefore, it seems that the activity against E. histolytica is not mediated through inhibition of an FT enzyme, but could involve another cellular target that is sensitive to this chemotype. Alternatively, FT could be the target in E. histolytica as a result of differences in binding preferences compared to other FT enzymes (found in mammals and trypanosomes), but this will need further exploration.

Apart from targeting HMGR in the mevalonate pathway (which is absent in E. histolytica), statins exhibit other pleiotropic effects including the modulation of protein prenylation that then leads to the inhibition of the activity of cell signaling molecules including the Rab family of small GTP-binding proteins (Greenwood et al., 2006; Wang et al., 2008). The E. histolytica genome encodes heterotrimeric G protein subunits and a large number of small G proteins, which are involved in vesicular trafficking (Bosch and Siderovski, 2013). Accordingly, we measured the effect of statins against E. histolytica. Of the statins, mevastatin, atorvastatin, fluvastatin, pitavastatin, and simvastatin tested at 100µM for 48 h, only simvastatin yielded moderate activity against E. histolytica. However, the combination of simvastatin and lonafarnib was slightly synergistic against E. histolytica, an intriguing finding given the apparent lack of a HMGR in the parasite.

Using microscopical observation and a constrained scoring system, lonafarnib, tipifarnib and the panel of 125 tipifarnib analogs were also screened in a time and/or concentrationdependent manner against S. mansoni somules and adults. Both lonafarnib and tipifarnib were among the most active Group A compounds, causing severe degenerative changes after 24 h in both developmental stages. FT activity has been documented in schistosomes and Ras farnesylation in S. mansoni extracts has been inhibited by the FT inhibitor, FTI-277 (Osman et al., 1999). However, like the situation with E. histolytica, the tipifarnib analogs that were most active against the schistosome are weak inhibitors of trypanosome and rat FTs (Buckner et al., 2000). This includes about half of the compounds shown in **Table 7** (CHN-15, JK-02, JK-17, JK-19, JK-25, JK-35, PN-077, and PN-149; unpublished data). It is possible that the S. mansoni FT ortholog is sufficiently distinct from the FTs of other species to limit the above SAR analysis. In this regard, it is worth noting that the three main schistosome species infecting humans all carry a FT gene, but the percentage identities are modest with approximately 36% identity across 79% of the human FT sequence (PF49354.1). In addition, schistosomes do not express CYP51 and, as might be anticipated, there was no correlation between the activity against S. mansoni and the EC<sup>50</sup> values previously measured for activity against T. cruzi in which the target is known to be CYP51 (Kraus et al., 2009). Thus, in spite of the potent anti-schistosomal activities measured, understanding the putative target(s) in S. mansoni for the tipifarnib analogs will need further investigation.

For both parasites, an analysis of the functional groups on the tipifarnib scaffold showed that large, hydrophobic groups such as 3-phenyl, 3-methyl-phenyl, 4-dimethyl, and naphthalene in the ring 2 position were more active. An analysis of the hydrophobicity of the 125 compounds (cLogP) confirmed a trend that more hydrophobic compounds were associated with stronger anti-amebic and anti-schistosomal activity (**Tables 5**, **6**). TABLE 7 | First pass screening of FTIs at 10µM against S. mansoni somules and adults.


Screens were performed twice each in duplicate and representative data are shown. Compounds with asterisks were also those most potent against E. histolytica in first pass screens (Table 1).

It is possible that the higher cLogP values (associated with greater lipophilicity) better facilitated permeability of E. histolytica and S. mansoni. Previously, for the latter parasite, the bioactivity of six statin analogs was positively associated with lipophilicity (Rojo-Arreola et al., 2014).

Although FTIs showed moderate activity against E. histolytica with EC<sup>50</sup> values ranging from 14.2 to 27.6µM, we see potential in the use of FTIs in combination with metronidazole. The combination of lonafarnib and metronidazole at 1:1, 1:2, 1:4, 1:8, 2:1, 4:1, 8:1, 16:1 elicited synergism with CI values ranging from 0.5 to 0.8. Thus, the dose of either one or both compounds necessary for complete growth inhibition of the parasite can be reduced.

To conclude, we report the potent anti-amebic and antischistosomal activities of the clinically developed FTIs, lonafarnib and tipifarnib. Screening of a previously developed tipifarnib analog library comprising 125 compounds identified subsets of potent compounds with little overlap between E. histolytica and S. mansoni. Also, scrutiny of the SAR data available for this set of compounds against trypanosomatid parasites, suggests that neither the E. histolytica nor the S. mansoni FT seems to be the target responsible for the bioactivities recorded. Nonetheless, the potent bioactivities measured encourage further study. Finally, the synergistic activity of the combination of lonafarnib and metronidazole against E. histolytica provides an opportunity to investigate a combination therapy for amebiasis.

# AUTHOR CONTRIBUTIONS

AP performed experiments, analyzed the data, and prepared the original draft. TN performed experiments and analyzed the data. NE-S, DS, and BS provided resources. FB analyzed the data, reviewed and edited the manuscript. MG synthesized the farnesyltransferase inhibitors. CC conceptualized the study, performed experiments, analyzed the data, reviewed and edited the manuscript. AD conceptualized the study, performed experiments, analyzed the data, wrote, reviewed and edited the manuscript.

# ACKNOWLEDGMENTS

AD was supported by the National Institutes of Health, Grant no. 1KL2TR001444 and UCSD Academic Senate Grant. TN was supported by 1TL1TR001443. Schistosome screens were made possible in part by an NIH-NIAID award R21AI126296 to CC. S. mansoni-infected and uninfected B. glabrata snails were supplied in part by the NIAID Schistosomiasis Resource Center (Biomedical Research Institute, Rockville, MD) through the NIH-NIAID Contract HHSN272201000005I. FB and MG were supported by NIH grants: R01AI070218 and R01AI106850. ChemAxon's JChem for Office (Excel; Version

# REFERENCES


15.1.1900.1773) was used to display the structures shown in **Supplementary Table 2** (http://www.chemaxon.com).

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2019.00180/full#supplementary-material

Supplementary Table 1 | Chemical structures and properties of analogs of tipifarnib and their activity against E. histolytica trophozoites. NA means compound was not available for testing.

Supplementary Table 2 | Chemical structures of tipifarnib analogs and their activity against S. mansoni somules and adults.

Supplementary Video 1 | DMSO control adult S. mansoni incubated for 24 h. Three male parasites are visible. Note the flexibility of the parasites and their ability to grasp the well bottom with their oral and ventral suckers.

Supplementary Video 2 | Adult S. mansoni incubated for 24 h in the presence of 10µM tipifarnib. Five male parasites are visible in the movie. Relative to the DMSO control parasites (Supplementary Video 1), note the uncoordinated movement, more constrained flexibility, darkened appearance and the inability of the parasites to adhere to the well bottom with their oral and ventral suckers.


quantitative structural analysis. Curr. Med. Chem. 20, 4888–4923. doi: 10.2174/09298673113206660262


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Probst, Nguyen, El-Sakkary, Skinner, Suzuki, Buckner, Gelb, Caffrey and Debnath. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Flavonoids as a Natural Treatment Against Entamoeba histolytica

Moisés Martínez-Castillo<sup>1</sup> , Judith Pacheco-Yepez <sup>2</sup> , Nadia Flores-Huerta<sup>1</sup> , Paula Guzmán-Téllez <sup>1</sup> , Rosa A. Jarillo-Luna<sup>2</sup> , Luz M. Cárdenas-Jaramillo<sup>3</sup> , Rafael Campos-Rodríguez <sup>2</sup> and Mineko Shibayama<sup>1</sup> \*

<sup>1</sup> Departamento de Infectómica y Patogénesis Molecular, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Ciudad de México, Mexico, <sup>2</sup> Sección de Estudios de Posgrado e Investigación, Instituto Politécnico Nacional, Escuela Superior de Medicina, Ciudad de México, Mexico, <sup>3</sup> Coordinación de Morfología, Departamento de Formación Básica Disciplinaria, Instituto Politécnico Nacional, Escuela Superior de Medicina, Ciudad de México, Mexico

Over the past 20 years, gastrointestinal infections in developing countries have been a serious health problem and are the second leading cause of morbidity among all age groups. Among pathogenic protozoans that cause diarrheal disease, the parasite Entamoeba histolytica produces amebic colitis as well as the most frequent extra-intestinal lesion, an amebic liver abscess (ALA). Usually, intestinal amebiasis and ALA are treated with synthetic chemical compounds (iodoquinol, paromomycin, diloxanide furoate, and nitroimidazoles). Metronidazole is the most common treatment for amebiasis. Although the efficacy of nitroimidazoles in killing amebas is known, the potential resistance of E. histolytica to this treatment is a concern. In addition, controversial studies have reported that metronidazole could induce mutagenic effects and cerebral toxicity. Therefore, natural and safe alternative drugs against this parasite are needed. Flavonoids are natural polyphenolic compounds. Flavonoids depend on malonyl-CoA and phenylalanine to be synthesized. Several flavonoids have anti-oxidant and anti-microbial properties. Since the 1990s, several works have focused on the identification and purification of different flavonoids with amebicidal effects, such as, -(-)epicatechin, kaempferol, and quercetin. In this review, we investigated the effects of flavonoids that have potential amebicidal activity and that can be used as complementary and/or specific therapeutic strategies against E. histolytica trophozoites. Interestingly, it was found that these natural compounds can induce morphological changes in the amebas, such as chromatin condensation and cytoskeletal protein re-organization, as well as the upregulation and downregulation of fructose-1,6-bisphosphate aldolase, glyceraldehyde-phosphate dehydrogenase, and pyruvate:ferredoxin oxidoreductase (enzymes of the glycolytic pathway). Although the specific molecular targets, bioavailability, route of administration, and doses of some of these natural compounds need to be determined, flavonoids represent a very promising and innocuous strategy that should be considered for use against E. histolytica in the era of microbial drug resistance.

Keywords: Entamoeba histolytica, flavonoids, alternative treatment, anti-oxidants, anti-inflammatory response, metronidazole

Edited by:

Serge Ankri, Technion – Israel Institute of Technology, Israel

#### Reviewed by:

Maryam Dadar, Razi Vaccine and Serum Research Institute, Iran Elisa Azuara-Liceaga, Universidad Autónoma de la Ciudad de México, Mexico

> \*Correspondence: Mineko Shibayama mineko@cinvestav.mx

Received: 06 March 2018 Accepted: 05 June 2018 Published: 22 June 2018

#### Citation:

Martínez-Castillo M, Pacheco-Yepez J, Flores-Huerta N, Guzmán-Téllez P, Jarillo-Luna RA, Cárdenas-Jaramillo LM, Campos-Rodríguez R and Shibayama M (2018) Flavonoids as a Natural Treatment Against Entamoeba histolytica. Front. Cell. Infect. Microbiol. 8:209. doi: 10.3389/fcimb.2018.00209

**51**

# INTRODUCTION

Entamoeba histolytica is a protozoan pathogen in humans and is the causative agent of amebiasis. This disease produces 100,000 deaths each year, particularly in developing countries. E. histolytica can manifest as a commensal or invasive microorganism in the large intestine, producing amebic colitis. The parasite has two cell cycle stages, the cyst and the trophozoite. The cysts are secreted in the stool from individuals who harbor E. histolytica. The trophozoites emerge from the cysts in the intestine, and this vegetative form is able to invade the intestinal mucosa and to disseminate to the liver, producing amebic liver abscess (ALA) (Sehgal et al., 1996; Haque et al., 2003). Few infected persons show symptoms with invasion to the bowel or extraintestinal sites, mainly in the liver (Stanley, 2003).

Diverse drugs have been employed against amebiasis. The treatment depends on the diagnosis and severity of the illness (**Table 1**). Usually, the drug can attack luminal forms and invasive amebiasis. However, in symptomatic patients and in invasive disease, the most widely used drugs against E. histolytica are the nitroimidazoles (metronidazole and tinidazole) (Marie and Petri, 2013; Ansari et al., 2015). Metronidazole (MTZ) kills amebas but does not cause damage to cysts. Although MTZ is the standard compound for treating amebiasis, it causes adverse effects, such as diarrhea, metallic flavor, loss of appetite, and nausea due to the doses and long-term treatment (1.5 g/day for 10 days) (Haque et al., 1997, 2003; Stanley, 2003). There are in vivo and in vitro studies where the nitroimidazoles induce genotoxic effects, which are related to the ability of cells to reduce these drugs; moreover, the position of the CH<sup>3</sup> and NO<sup>2</sup> groups of these compounds are involved in DNA damage (Boechat et al., 2015). In the case of MTZ, it is known that its biotransformation produces nitroso intermediates (e.g., hydroxy metabolite and acetic acid), which can form adducts in the DNA or inhibit the thioredoxin reductase-generating reactive oxygen species, causing oxidative cell damage (Mudry et al., 1994; Elizondo et al., 1996; Leitsch et al., 2007). Furthermore, MTZ can cross the blood-brain barrier producing cerebellar toxicity (Agarwal et al., 2016). Thus, it is important to consider the doses of MTZ administration, the individual susceptibility and to evaluate the risk-benefit in relation with the severity of the infection (Mudry et al., 1994; Elizondo et al., 1996).

Because of the undesirable effects of MTZ, it is necessary to develop an alternative treatment from biological or synthetic sources that can eliminate E. histolytica. Specifically, the development or discovery of novel compounds without toxicity and side effects is needed. In many countries, plant extracts have been employed ancestrally and are a good alternative to treat amebiasis. Furthermore, it is necessary to validate, identify and purify the active compounds with anti-amebic properties before employing them as a treatment.

Usually, in plants, polyphenolic metabolites can be present in roots, leaves, flowers, and fruits. Flavonoids are contained within these natural compounds, which have demonstrated promising results mainly in in vitro studies; however, it is necessary to evaluate their potential activity in in vivo models as a complementary and/or specific therapeutic strategy against E. histolytica trophozoites, as we describe below.

# FLAVONOID STRUCTURE AND CLASSIFICATION

Flavonoids are natural pigments present in vegetables. These compounds protect organisms against the damage produced by oxidant agents such as UV radiation, pollution, and the chemical substances present in the foods. Humans obtain these protector flavonoids through direct ingestion of aliments and supplements. The chemical structure of flavonoids is characterized by the presence of a variable number of phenolic hydroxyl groups (polyphenols). Flavonoids have low molecular weights (500– 4,000 Da) and share a common skeleton of diphenylpyranes (C6-C3-C6) composed of two phenyl rings linked through a ring C of pyran (**Figure 1**). These compounds present excellent iron-chelating ability that gives them great anti-oxidant capacity (Havsteen, 1983). The anti-oxidant activity of flavonoids depends on the redox properties of their phenolic groups (Bors et al., 1990). According to their structural characteristics (the nature of the C3 element), flavonoids can be classified in **Table 2**. Different chemical modifications may occur in each group, such as hydroxylation, hydrogenation, sulfuration, methylation, acylation, and glycosylation (Andersen and Markham, 2006; Wang et al., 2017).

# Flavones

The flavone family is a subgroup of flavonoids that is synthesized depending on whether they contain C- or O-glycosylation and a hydroxylated B-ring. These compounds have been mainly isolated from leaves, aerial parts, and the exudates of plants. Flavones are characterized by a double bond between C-2 and C-3 and a B ring in C-2 (Jiang et al., 2016). Flavones present antioxidant activities due to their ability to scavenge reactive oxygen species (ROS). For example, luteolin inhibits xanthine oxidase, which is an important enzyme that is involved in ROS production (Cos et al., 1998; Spanou et al., 2012). Apigenin reduces the phosphorylation of NF-κB/p65 in mouse macrophages and in human monocytes, inhibiting its transcriptional activity and the expression of pro-inflammatory cytokines (Nicholas et al., 2007). Acacetin has shown important anti-peroxidative and antiinflammatory activities, and these effects were determined by inhibiting iNOS and COX-2 activity in murine macrophages (Pan et al., 2006) (**Table 2**).

# Flavonols

Flavonols are composed of multiple phenol structural units. Examples of this group are kaempferol, kaempferol-3-methyl ether, quercetin, and quercetin-3-methyl ether among others. The flavonols present important health benefits, such as the anti-oxidant activity, by increasing the activity of catalase and glutathione peroxidase (Bai et al., 2016). They also act as radical scavengers (Tipoe et al., 2007; Bai et al., 2016). In addition, they present anti-inflammatory effects by inhibiting the activity of lipoxygenase and cyclooxygenase (Kim et al., 1998; Marunaka, 2017).

#### TABLE 1 | Pharmacological treatment of amebiasis.


# Flavanones

This group is characterized by the lack of a double bond between C-2 and C-3 in the C ring of the flavonoid skeleton. Thus, in these compounds, C-2 bears one hydrogen atom in addition to the phenolic B-ring, and C-3 has two hydrogen atoms (Grayer and Veitch, 2006). Flavanones have anti-oxidant, anti-inflammatory, and neuroprotective activity (Hernández-Aquino et al., 2017).

# Flavan-3-ols (Flavanols)

The flavanols or catechins, also referred to as flavan-3-ols, are non-glycosylated compounds that are present in plants in the form of monomers (catechins). The hydroxyl group is bound to position 3 of the C ring, and has no double bonds between positions 2 and 3. Another important characteristic is the high nucleophilicity of their A-rings to HO<sup>−</sup> and RO<sup>−</sup> (Andersen and Markham, 2006; Panche et al., 2016). These compounds have anti-oxidant, anti-inflammatory and anti-microbial properties. The most studied flavan-3-ol monomers are (+)-catechins, (–)-epicatechin, (–)-epigallocatechin, (–)-epicatechin gallate, (– )-epigallocatechin gallate, and (+)-gallocatechin (Panche et al., 2016; Borges et al., 2017; Kuhnle, 2017).

# Anthocyanidins

Anthocyanidins have an -OH group in position 3 but also have a double bound between carbons 3 and 4 of ring C. This group is the main flavonoid responsible for cyanic colors (red, purple, and blue) in vegetables, flowers and fruits. The most common examples are cyanidin, delphinidin, pelargonidin, malvidin, and peonidin (Andersen and Markham, 2006; Azzini et al., 2017). They also have anti-oxidant and anti-microbial


effects (Middleton et al., 2000; Pietta, 2000; Azzini et al., 2017; Khoo et al., 2017).

### Chalcones

The chalcones are referred to as open-chain flavonoids. Chalcones are the yellow to orange flower pigments of some plants (Andersen and Markham, 2006; Panche et al., 2016). The A and B rings of chalcones are linked by a three-carbon chain instead of a C ring, which is absent (Veitch and Grayer, 2006). Chalcones have nutritional and biological benefits owing to their anti-bacterial and anti-parasitic activities (Nowakowska, 2007; Costa et al., 2016).

# Isoflavonoids

The isoflavones are commonly known as ß-glucosides and have a B ring in position 3. These compounds have potent anti-oxidant activity. Genistein and daidzein are the main two soy isoflavones, whose main effects are the inhibition of lipid peroxidation (Lapcík et al., 1998; Yu et al., 2016). Genistein and daidzein present anti-oxidant activity in peripheral blood lymphocytes, increasing DNA protection against oxidative damage, which contributes to homeostasis in humans (Takahashi et al., 2009).

# FLAVONOIDS AND ANTI-OXIDANT PROPERTIES

The imbalance between the generation of ROS and reactive nitrogen species (RNS) and their elimination is classically described as "oxidative stress," which plays an important role in the pathophysiology of many diseases because ROS and RNS can react with lipids, proteins and DNA, inducing their oxidation and causing cell damage. The human body has several anti-oxidant systems, where anti-oxidants are understood as "any substance that retards, prevents or removes oxidative damage to target molecules" (Halliwell and Gutteridge, 2015). The most important anti-oxidant enzymes in mammalian cells are superoxide dismutase (SOD), catalase (CAT), and glutathione peroxidase (GPx) (Valko et al., 2007; Deponte, 2013; Losada-Barreiro and Bravo-Díaz, 2017). Non-enzymatic mechanisms include iron-binding proteins such as transferrin and ferritin, melatonin, and uric acid (Othman et al., 2008; Pizzino et al., 2017). These endogenous anti-oxidant systems are complementary and usually sufficient to prevent oxidative damage to the cells, but in certain conditions, such as when ROS production is excessive, the intake of exogenous anti-oxidants is convenient to reduce damage. These anti-oxidants derived from the diet are found in all vegetables and include phenols, phenolic acids, tannins, lignans, and flavonoids (Pizzino et al., 2017). The human diet contains between 50 mg and 800 mg of flavonoids per day, depending on the consumption of fruits and vegetables.

In general, the anti-oxidant activity of a flavonoid depends on three of its structural characteristics: (a) the presence of the catechol group in the B ring, (b) a 12 double bond and a 4-oxo group in the C ring, and (c) the hydroxyl groups on positions C-3 and C-5 (Wolfe and Liu, 2008; Kumar et al., 2013). These activities are modulated by three general mechanisms: (a) scavenging the ROS, (b) suppressing the formation of free radicals by enzymatic inhibition or chelating elements involved in the formation of them, and (c) protecting or upregulating the anti-oxidants. The ROS-scavenging activity is carried out by direct donation of hydrogen atoms, resulting in more stable and less reactive radicals.

Furthermore, some flavonoids can scavenge O<sup>−</sup> 2 and others scavenge NO (Hanasaki et al., 1994; Vanacker et al., 1995; Prior and Cao, 2000). Anthocyanidins present the most effective scavenger function, with an activity 10–1,000 times greater than glutathione (GSH) (Cao et al., 1997). The activity of scavenging O − 2 and NO prevents the formation of ONOO−, which is highly oxidizing. Moreover, flavonoids are direct chelators of ONOO−, as was reported for quercetin (Haenen et al., 1997; Heijnen et al., 2001; Spencer et al., 2003).

However, flavonoids that chelate Fe2<sup>+</sup> or Cu<sup>+</sup> can remove a causal factor in the formation of free radicals (e.g., quercetin) (de Groot and Rauen, 1998). These natural compounds stimulate the induction of anti-oxidant enzymes, such as glutathione-S transferase, UDP-glucuronosyl transferase, and NADH-quinone oxidoreductase, which are the main defense against toxic electrophilic and oxidant stress (Procházková et al., 2011). Other enzymes inhibited by flavonoids are those involved in the metabolism of arachidonic acid, such as lipoxygenase, cyclooxygenase, microsomal succinoxidase, and NADH oxidase (de Groot and Rauen, 1998). The anti-oxidant function of flavonoids includes the reduction of α-tocopherol, which represents the main anti-oxidant of membranes and low-density lipoproteins (LDL). Flavonoids can donate a hydrogen to αtocopherol radicals and thus protect from oxidation to LDL (Hirano et al., 2001), as observed for quercetin and catechins (Zhu et al., 2000).

Although the anti-oxidant role of flavonoids is welldocumented, many studies have reported their pro-oxidant activity, which seems to depend on their concentration and is directly proportional to the total number of OH groups in the molecule, especially those located in the B ring (Perron et al., 2011).

# OXIDATIVE AND ANTI-OXIDATIVE MICROENVIRONMENT IN AMEBIASIS

During ALA, E. histolytica is capable of inducing an important inflammatory response, which is mainly composed of neutrophils (PMNs) and macrophages. These cells create an oxidative stress environment. In the initial stages of ALA, the amebas are surrounded by neutrophils and posteriorly by macrophages. Chronic inflammation allows the formation of a granulomatous reaction. This exacerbated response can lead to extensive areas of necrosis (Tsutsumi et al., 1984). In this milieu, E. histolytica interacts with oxidative and non-oxidative metabolites produced by the inflammatory cells. In experimental ALA, NO concentration increases in a time dependent manner, which was found in serum samples from hamsters at different stages of the infection. In this study, the authors used histochemistry to determine the presence of the NADPH enzyme that correlated with the size and severity of the lesion. With these findings, they concluded that during the establishment of ALA, E. histolytica trophozoites resisted the increment of NO produced by inflammatory cells; therefore, this molecule is not sufficient to eliminate the parasite in in vivo studies (Pacheco-Yépez et al., 2001). The presence of NO was also evaluated in intestinal amebiasis; in these reports, a significant increase in the NO levels in patients with diarrhea was found compared with the control group. These results support that this oxidative molecule possesses a central role in the pathophysiological mechanisms underlying amebiasis (Pérez-Fuentes et al., 2000; Namiduru et al., 2011).

The relevance of NO is also related with its capacity to react with other molecules that are present in this oxidative environment, such as the superoxide anion (O<sup>−</sup> 2 ), which leads to the production of peroxynitrite (ONOO−), which has been described as one of the most potent oxidants produced in a biological system. To date, there are no reports related with peroxynitrites in amebiasis. It is necessary to evaluate the role of this molecule in the pathogenesis of amebiasis (Pacheco-Yepez et al., 2014).

In this oxidative milieu, the production of O<sup>−</sup> 2 depends of the activity of the NADPH-oxidase, besides the SOD enzyme converts the superoxide anion to H2O2, which activates the myeloperoxidase system (MPO). This enzyme catalyzed the production of hypochlorous acid (HOCl), a highly oxidant molecule, as cytotoxic effector synthetized mainly by neutrophils. The anti-amebic activity of the MPO was described in in vivo experiments. The MPO caused important damage in trophozoites. These results showed that this enzyme, which was produced by inflammatory cells, is capable of protecting host tissue against E. histolytica (Pacheco-Yépez et al., 2011).

As was previously mentioned, trophozoites are exposed to oxidative stress (ROS and RNS such as O<sup>−</sup> 2 , H2O2, ONOO−, and NO). In these conditions, the ameba displays endogenous anti-oxidant enzymes to avoid the effect of these toxic molecules; therefore, E. histolytica can survival in this environment. A trypanothione reductase in E. histolytica was identified (Ondarza et al., 2005; Tamayo et al., 2005). This enzyme possesses anti-oxidant properties similar to glutathione and thioredoxin enzymatic systems (Krauth-Siegel and Comini, 2008; Krauth-Siegel and Leroux, 2012).

A 29 kDa thiol-dependent peroxiredoxin protein (Eh2CysPrx or Eh29) has been reported. The authors demonstrated that the recombinant protein presented thiol peroxidase activity due to its ability to remove H2O2. In addition, they identified the protein in the ameba cytoplasm with a molecular weight of 29 kDa (Bruchhaus et al., 1997). In contrast, other research groups reported that peroxiredoxin was localized in the membrane, and the authors demonstrated that E. histolytica presented a larger amount of this enzyme compared with E. dispar. These results showed that peroxiredoxin can be a virulence factor of E. histolytica, protecting trophozoites from oxidative stress (Choi et al., 2005). When the amebas were exposed for 1 h to high oxygen concentrations, the expression of eh29 was increased 2.1 fold. The authors concluded that the enzyme is involved in the detoxification of peroxides and peroxynitrites (Akbar et al., 2004; Sen et al., 2007).

More recently, the thioredoxin system (EhTRXR/TRX), a group of anti-oxidant enzymes, has been described in E. histolytica. The recombinant protein is able to catalyze the reduction of NADPH or NADH and S-nitrosothiols. This enzyme exhibited NADPH dependent oxidase activity, which generates H2O<sup>2</sup> from O2. This protein represents an important mechanism to regulate intracellular and extracellular levels of oxidative molecules (Arias et al., 2012). It is important to mention that these anti-oxidant systems maintain a redox balance in the parasite, allowing their survival in adverse oxidative conditions (Arias et al., 2007, 2008).

Based on these findings, we strongly suggest that flavonoids could participate as natural anti-oxidant compounds in the highly oxidative environment during ALA, favoring the resolution of liver tissue damage caused not only by the presence of the ameba and its secretion products but also by the inflammatory response characteristic of this pathology. In addition, it is important to consider the role of these natural molecules that could act directly on trophozoites. Further studies are necessary to determine the effect of flavonoids in in vivo models to identify their possible molecular targets in amebiasis (**Figure 5**).

# FLAVONOIDS AND THEIR MOLECULAR TARGETS IN PROTOZOA

Flavonoids and their therapeutic applications in human health have been extensively investigated in recent years due to their use in traditional medicine. Interestingly, these studies demonstrated a possible correlation between the chemical structure of the flavonoid and the molecular target in different cell lines (Panche et al., 2016). However, the mechanisms of action and the multiple effects of these natural compounds on the cells are not fully understood.

The anti-protozoa effect of flavonoids and some of their molecular targets have been demonstrated in Plasmodium falciparum, Trypanosoma brucei brucei, T. brucei gambiense, T. cruzi, Leishmania donovani, Cryptosporidium parvum, Toxoplasma gondii, and Giardia lamblia. In P. falciparum the catechins acts against some identified molecular targets that include enzymes involved in the biosynthesis of fatty acids (FabG, FabZ, FabI, and enoyl-ACP reductase) (Sharma et al., 2007). In Trypanosoma cruzi, (–)-epicatechin has been described to affect the arginine kinase activity and NADH-oxidase activity (Paveto et al., 2004; Maya et al., 2007; Scotti et al., 2010; Dodson et al., 2011). In Leishmania donovani, it was reported that kaempferol promotes the inhibition of the activity of pyruvate kinase, the dihydroorotase enzyme (LdDHOase) and the cytidine deaminase, which impact the pyrimidine biosynthesis pathway, causing the death of the parasites (Scotti et al., 2015; Tiwari et al., 2016). In the case of Toxoplasma gondii, quercetin inhibits the synthesis of HSP90, HSP70, and HSP27, that have been described as virulence factors (Dobbin et al., 2002; Kerboeuf et al., 2008). These alterations promote reductions in the invasion to the host tissues, adhesion, proliferation and cell differentiation (Calzada et al., 1998; Mamani-Matsuda et al., 2004; Mead and McNair, 2006; Kerboeuf et al., 2008; Sen et al., 2008; Dodson et al., 2011; Jin et al., 2014; Cornelio et al., 2017) (**Figure 2**).

# FLAVONOIDS WITH ANTI-AMEBIC ACTIVITY

The diarrheal infections caused by E. histolytica represent a great problem in developing countries. They are responsible for a considerable number of morbidities and mortalities in these populations. The development of novel and effective antiamebic compounds without side effects is necessary. Since the 1990s, medicinal plants have gained popularity as potential therapeutic alternatives. Because of this, in the last three decades, natural drugs and their products have represented approximately 50% of all treatments that have come to market (Newman and Cragg, 2016), and flavonoids display significant anti-amebic activity in in vitro studies (**Figure 3** and **Table 3**).

In recent years, flavonoids have generated great interest in the scientific community. However, there are few studies concerning their molecular mechanisms against E. histolytica. The most studied are kaempferol, (–)-epicatechin and tiliroside. In these studies, it has been observed that the main molecular targets correspond to cytoskeleton related proteins (myosin II heavy chain, α-actinin, and actin). The authors also demonstrated a dysregulation of glycolytic enzymes and stress oxidative proteins. They concluded that all these changes modify the pathogenic mechanism, such as adhesion, cytolysis, phagocytosis, and migration (Bolaños et al., 2014, 2015; Calzada et al., 2017a) (**Figure 4** and **Table 4**). Below, we highlight the therapeutic effectiveness of the main flavonoids with anti-amebic activity and explain in more detail the molecular targets that could probably be affected in the ameba.

#### TABLE 3 | In vitro IC50 of flavonoids against E. histolytica.


\*IC<sup>50</sup> values correspond to the mean concentration of the highest anti-amebic activity reported in the references.

# IN VITRO STUDIES AND MOLECULAR TARGETS

# Catechin

Catechin has been isolated from plants that are used in traditional medicine for the treatment of gastrointestinal disorders. Plants with anti-amebic activity that have been used as sources of catechins and their derivatives include Helianthemum glomeratum, Osyris alba, Chiranthodendron pentadactylon, Geranium niveum, Geranium mexicanum, and Rubus coriifolius (Meckes et al., 1999; Alanis et al., 2003; Al-Jaber et al., 2010; Calzada et al., 2017c).

The (–)-epicatechin isoform in in vitro assays has shown the best IC<sup>50</sup> value of 1.9µg/ml in inhibiting amebic growth compared with that of other catechin derivatives [e.g., (–) epigallocatechin, (+)-catechin-3-O-α-L-rhamnopyranoside, and geranins A, B, C, and D] (Meckes et al., 1999; Calzada et al., 2001). The differences in the anti-amebic activity of the epicatechin derivatives can be related with the presence of hydroxy groups in the phenolic ring and galloyl moieties [e.g., (–)-epicatechin and (–)-epigallocatechin]. In addition, the presence of hydroxy groups in position 3, 4, and 5 in the B ring enhances the anti-oxidant and scavenging activities (Mizushina et al., 2005; Braicu et al., 2011) (**Figure 5**). The (–) epicatechin induced morphological changes in the trophozoites at the nuclear and cytoplasmic levels, causing programmed cell death in approximately 95% of amebas (Soto et al., 2010). Moreover, (–)-epicatechin caused alterations of cytoskeleton

proteins from E. histolytica (myosin II heavy chain, actin and alpha-actinin), affecting adhesion, cytolysis, migration, and phagocytosis (**Table 4**). In addition, (–)-epicatechin caused the dysregulation of enzymes that are involved in energy metabolism, such as glyceraldehyde-phosphate dehydrogenase and fructose-1,6-bisphosphate aldolase (G/FBPA). The (–) epicatechin concentration used did not promote cytotoxicity in intestinal mammalian cells (Caco-2) (Bolaños et al., 2014). Therefore, (–)-epicatechin can be considered the most promising alternative, safe flavonoid for treatment against amebiasis. It is necessary to evaluate in more detail the molecular mechanisms involved in the specific targets and their impact on the virulence of E. histolytica (**Figures 4**, **5**).

# Kaempferol

One of the first studies of the anti-amebic activity of kaempferol showed that the purified compound from the aerial parts of Cuphea pinetorum had an inhibitory effect on amebic culture proliferation. This work reported an IC<sup>50</sup> value near 7.9µg/ml (Calzada, 2005). Two later studies determined similar IC<sup>50</sup> values against E. histolytica. The specific anti-amebic activity of kaempferol is independent of the plant species [Cnidoscolus chayamansa (Mc Vaugh) and Annona cherimola Miller] (Calzada et al., 2017b; Pérez-González et al., 2017). Some of the molecular mechanisms triggered by kaempferol in amebas include the dysregulation of actin, myosin II heavy chain, cortexillin II, heat shock protein 70, glyceraldehyde-phosphate dehydrogenase and G/FBPA. These results are similar to the results for (–) epicatechin (Bolaños et al., 2014, 2015). The authors showed that at 27.7µM, kaempferol inhibited 77.1% of ameba growth in comparison with trophozoites untreated with this flavonoid. In addition, they reported the inhibition of adhesion to Caco-2 cells in amebas co-incubated with kaempferol. (Bolaños et al., 2014, 2015; **Table 4**).

Kaempferol has a high affinity for pyruvate ferredoxin oxidoreductase (PFOR), an important therapeutic target in E. histolytica (Samarawickrema et al., 1997; Jeelani and Nozaki, 2016). By docking analysis, the authors found that the molecular interaction of kaempferol with the amebic PFOR included nine amino acid residues (Phe 665, Pro 666, Leu 667, Gly 845, Ala 846, Met 851, Tyr 853, Trp 864, and Asn, 866) compared with the four different binding sites of MTZ (Phe 174, His 178, Lys 435, Phe 453, and Tyr 455) (Calzada et al., 2017b).

Kaempferol derivatives also exhibit considerable anti-amebic activity. In particular, tiliroside showed an IC<sup>50</sup> of 7.45µg/ml against E. histolytica trophozoites. In a similar manner to kaempferol, tiliroside can interact directly with PFOR and G/FBPA according to docking studies, showing an inhibition constant (KI) of 53.57 and 55.5µM, respectively. These data are comparable with the values of the KI for MTZ (47.64 and 44.01µM, respectively) (Calzada and Alanis, 2007; Calzada et al., 2017a) (**Figure 4** and **Table 4**). It is important to remark that the differences in IC values and the molecular mechanisms of action between kaempferol and its derivatives may be due to the presence of glucosyl moieties on position C3 of the principal structure or the number of hydroxyl groups in ring B, which probably limits the correct activity of tiliroside (Cimanga et al., 2006; Singh et al., 2009) (**Table 3**).

Based on previous reports, kaempferol, (–)-epicatechin, and (–)-epigallocatechin present different IC<sup>50</sup> for anti-amebic activity and presented the best IC<sup>50</sup> values that showed good activity against E. histolytica. These natural compounds

#### TABLE 4 | Flavonoids with anti-amebic activity.


FIGURE 5 | Possible effects of flavonoids in the regulation of biochemical and immunological responses against amebiasis. (A) Recruitment of neutrophils and macrophages by E. histolytica promoting the synthesis and production of pro-inflammatory cytokines and oxidative mediators. E. histolytica presents detoxifying enzymes. (B) Possible mechanisms of the flavonoids as direct scavenger of ROS, ONOO−, and NO, enhancer of CAT, SOD, and GPx enzymes and regulation of inflammation via STAT and NF-κB. Effect of the flavonoids in neutrophils (MPO production), macrophages (decrease of inflammatory mediators), and E. histolytica trophozoites (direct damage in different molecular targets).

reduced the amebic viability by 50% (1.9, 6.89, and 7.93µg/ml respectively). Due to the low IC50, these flavonoids could be considered good candidates to evaluate their in vivo activity. Other important advantages of the use of flavonoids are their anti-oxidant and anti-inflammatory activity, which could participate in the resolution of amebiasis.

# Quercetin

This flavonoid showed a slight anti-amebic effect in vitro (IC<sup>50</sup> 114.3µg/ml) compared with that of G. lamblia (IC<sup>50</sup> 26.6µg/ml). However, isoquercitin, a quercetin derivative, inhibited the viability and growth of E. histolytica trophozoites after 48 h of incubation (IC<sup>50</sup> 14.7µg/ml) (Calzada and Alanis, 2007). Although there are no reports of a specific target of isoquercetin in the ameba, it is imperative to mention that this flavonoid is metabolized in the bowel and the liver (Valentova et al., 2014); therefore, its administration in in vivo models could diminish the tissue damage induced by E. histolytica.

# Other Polyphenol Compounds With Anti-amebic Effects

Other flavonoids that have anti-amebic activity are luteolin and apigenin. These compounds, isolated from the total extract of Morinda morindoides, have shown IC<sup>50</sup> values of 17.8 and 10µg/ml, respectively, against the ameba. The authors suggested that glycosylation on the C-7 position decreased the anti-amebic activity of these flavonoids (Cimanga et al., 2006) (**Table 3**). Nevertheless, it is important to mention that the anti-amebic mechanisms of these flavonoids and their real potential activities have not been determined. Other compounds with anti-amebic effects are the chalcones. Leeza Zaidi et al. (2015) synthetized modified chalcones (N-substituted ethanamine). In vitro assays showed that their derivatives displayed a better anti-amebic activity than the MTZ reference drug (Leeza Zaidi et al., 2015).

# IN VIVO STUDIES

The flavonoids have shown extensive benefits and no side effects in in vivo models of different types of cancer, liver cirrhosis, neurodegenerative diseases and metabolic disorders (e.g., obesity and diabetes) (Goto et al., 2012; Nabavi et al., 2015; Panche et al., 2016; Hernández-Aquino et al., 2017).

In contrast, there are few reports of flavonoids in parasitic diseases using in vivo models. The efficacy of intragastric treatment with (–)-epicatechin at 0.072µmol/kg was demonstrated in CD-1 mice infected with G. lamblia trophozoites (1 × 10<sup>6</sup> ). The (–)-epicatechin administered 6 days post-infection decreased the number of parasites in the small intestine. Moreover, this natural compound displayed a higher activity than that of MTZ and emetine (Barbosa et al., 2007). The epigallocatechin was effective in BALB/c mice infected with Leishmania amazonensis. The results showed that the animals treated orally with 30 mg/kg/day of epigallocatechin for 5 days presented a reduction in the size of the ear lesion and a low parasite burden compared with that of the control group (Inacio et al., 2013).

It is important to mention that, until now, there have been no studies of amebiasis using flavonoids in in vivo models. Nevertheless, a polyphenol, resveratrol, has presented anti-amebic properties both in vitro and in vivo. In vitro incubation with resveratrol (72µM) for 48 h induced cell growth arrest, production of ROS, damage to membrane lipids, increased intracellular Ca2+, calpain activation, and decreased superoxide dismutase activity, leading to apoptosis in E. histolytica. Moreover, in hamsters with ALA, the pretreatment for 2 or 10 days with 30 mg/kg of resveratrol by oral gavage prevented liver damage by the trophozoites, whereas nonpretreated animals developed extensive ALA (80% of the total liver). Histopathological analysis showed that resveratrol-treated hamsters presented a healthy liver parenchyma with retraction of the granulomatous reaction, whereas the untreated animals displayed liver necrosis. Based on these results, the authors suggested that resveratrol could be used as alternative treatments for amebiasis (Pais-Morales et al., 2016). Although resveratrol is a polyphenol, it is necessary to perform in vivo studies to demonstrate the potential anti-amebic effect of flavonoids before their use as an alternative or complementary treatment.

# CLOSING REMARKS

In recent years, the use of natural compounds against infectious diseases caused by protozoan parasites has gained popularity among pharmaceutical corporations (Newman and Cragg, 2016). In 2015, the Nobel Prize in Medicine was awarded to Professor Youyou Tu for her valuable contributions to the discovery of artemisinin as a natural malaria treatment (Su and Miller, 2015). These advances have been a consequence of the various reports on drug resistance in microorganisms as well as the toxicity and side effects of many drugs on humans. The field of amebiasis is no exception. Promising advances have been made using different secondary metabolites from plant extracts (Procházková et al., 2011; Herrera-Martínez et al., 2016; Bashyal et al., 2017). The use of plants rich in flavonoids (grapes, cacao and flowers such as Geranium mexicanum) have been shown to exert anti-amebic activity, having cytoskeletal proteins and enzymes related to the glycolytic metabolism of E. histolytica as molecular targets and leading to alterations in DNA replication. These changes suggest that the ameba lost their virulence factors (Bolaños et al., 2014, 2015) (**Figure 4**). It is well known that during the establishment of amebiasis in the intestine and in the liver, the inflammatory reaction can promote the participation of pro-inflammatory cytokines. In this milieu, there are a highly oxidative stresses, constituted by NO, peroxynitrites (ONOO−), ROS, and hypochlorous acid. All these oxidant metabolites caused tissue damage. Flavonoids can regulate inflammation-activating anti-oxidants enzymes (CAT, SOD, and GPx). Additionally, flavonoids participate as scavengers that can remove these free radicals. We cannot discard that flavonoids act directly with neutrophils and macrophages to kill amebas (**Figures 5A,B**).

Considering all the benefits and probable therapeutic targets of flavonoids in the treatment of E. histolytica infection, it is important to investigate these natural compounds in basic research studies that will establish the doses, administration routes, bioavailability, and metabolic biotransformation (glycosidation, glucuronidation, sulfation, and O-methylation) as well as effects on the tissue microenvironment of these natural compounds in in vivo models of amebiasis. Considering all of these remarks, flavonoids can be considered good alternatives for the effective treatment of amebiasis.

# AUTHOR CONTRIBUTIONS

Authors that contributed to writing the manuscript: MS (introduction and closing remarks), JP-Y and RC-R (flavonoids structure and classification), RJ-L (flavonoids and anti-oxidant properties), NF-H and LC-J (flavonoids and their molecular targets in protozoa), MM-C and PG-T (flavonoids with antiamebic activity; in vitro and in vivo studies and molecular

# REFERENCES


targets). MS organize and revise the manuscript. All the authors prepared and edited the figures and tables. All authors contributed to manuscript revision, read and approved the submitted version.

# FUNDING

This work was supported by CONACyT grant number 237523 assigned to MS.

# ACKNOWLEDGMENTS

We would like to thank MSc Daniel Coronado-Velázquez for his valuable assistance in the performance of the schematic representations (**Figures 4**, **5**). English grammar editing was done by the American Journal Experts (https://www.aje.com/).


bearing N-substituted ethanamine tail. Eur. J. Med. Chem. 98, 179–189. doi: 10.1016/j.ejmech.2015.05.013


Entamoeba histolytica by mass spectrometry and chemical analysis. Biotechnol. Appl. Biochem. 42, 175–181. doi: 10.1042/BA20050023


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Martínez-Castillo, Pacheco-Yepez, Flores-Huerta, Guzmán-Téllez, Jarillo-Luna, Cárdenas-Jaramillo, Campos-Rodríguez and Shibayama. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Morphodynamics of the Actin-Rich Cytoskeleton in *Entamoeba histolytica*

Maria Manich1,2, Nora Hernandez-Cuevas 2†, Juan D. Ospina-Villa3†, Sylvie Syan<sup>2</sup> , Laurence A. Marchat <sup>3</sup> , Jean-Christophe Olivo-Marin<sup>1</sup> and Nancy Guillén2,4 \*

<sup>1</sup> BioImaging Unit, Institut Pasteur, Paris, France, <sup>2</sup> Cell Biology of Parasitism Unit, Institut Pasteur, Paris, France, <sup>3</sup> Instituto Politécnico Nacional, Escuela Nacional de Medicina y Homeopatía, Mexico City, Mexico, <sup>4</sup> Centre National de la Recherche Scientifique, CNRS-ERL9195, Paris, France

#### *Edited by:*

Mario Alberto Rodriguez, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional (CINVESTAV-IPN), Mexico

#### *Reviewed by:*

Anne-Marie Krachler, University of Texas Health Science Center at Houston, United States Eden Ramalho Ferreira, Federal University of São Paulo, Brazil

> *\*Correspondence:* Nancy Guillén

nguillen@pasteur.fr

#### *†Present Address:*

Nora Hernandez-Cuevas, Laboratory of Parasitology C.I.R. "Hideyo Noguchi", Universidad Autónoma de Yucatán, Merida, Mexico Juan D. Ospina-Villa, Grupo Biología y Control de Enfermedades Infecciosas-BCEI, Universidad de Antioquia, Medellín, Colombia

> *Received:* 21 March 2018 *Accepted:* 09 May 2018 *Published:* 29 May 2018

#### *Citation:*

Manich M, Hernandez-Cuevas N, Ospina-Villa JD, Syan S, Marchat LA, Olivo-Marin J-C and Guillén N (2018) Morphodynamics of the Actin-Rich Cytoskeleton in Entamoeba histolytica. Front. Cell. Infect. Microbiol. 8:179. doi: 10.3389/fcimb.2018.00179 Entamoeba histolytica is the anaerobic protozoan parasite responsible for human amoebiasis, the third most deadly parasitic disease worldwide. This highly motile eukaryotic cell invades human tissues and constitutes an excellent experimental model of cell motility and cell shape deformation. The absence of extranuclear microtubules in Entamoeba histolytica means that the actin-rich cytoskeleton takes on a crucial role in not only amoebic motility but also other processes sustaining pathogenesis, such as the phagocytosis of human cells and the parasite's resistance of host immune responses. Actin is highly conserved among eukaryotes, although diverse isoforms exist in almost all organisms studied to date. However, E. histolytica has a single actin protein, the structure of which differs significantly from those of its human homologs. Here, we studied the expression, structure and dynamics of actin in E. histolytica. We used molecular and cellular approaches to evaluate actin gene expression during intestinal invasion by E. histolytica trophozoites. Based on a three-dimensional structural bioinformatics analysis, we characterized protein domains differences between amoebic actin and human actin. Fine-tuned molecular dynamics simulations enabled us to examine protein motion and refine the three-dimensional structures of both actins, including elements potentially accounting for differences changes in the affinity properties of amoebic actin and deoxyribonuclease I. The dynamic, multifunctional nature of the amoebic cytoskeleton prompted us to examine the pleiotropic forms of actin structures within live E. histolytica cells; we observed the cortical cytoskeleton, stress fibers, "dot-like" structures, adhesion plates, and macropinosomes. In line with these data, a proteomics study of actin-binding proteins highlighted the Arp2/3 protein complex as a crucial element for the development of macropinosomes and adhesion plaques.

Keywords: *Entamoeba*, actin, macropinosome, HaloTag, Arp2/3

# INTRODUCTION

Actin is a fundamental component of the cytoskeleton. It is able to form robust cellular scaffolds (called microfilaments) that underpin the vast majority of motile events in eukaryotic cells, including changes in cell shape and in the morphology of the endomembrane system (Svitkina, 2018). The actin-rich cytoskeleton's ability to fulfill all these essential cellular functions depends on the precise spatiotemporal control of microfilament formation and turnover. Microfilament polymerization and depolymerization are tightly regulated by (i) the presence of diverse isoforms of actin in the same organism (there are three in humans, for example), and (ii) the existence of more than a hundred proteins associated with G-actin (monomeric actin) or F-actin and that regulate microfilament assembly or disassembly (for a review, see Pollard, 2016). These actinbinding proteins (ABPs) use subtle mechanisms of action to control microfilament organization in various networks. For instance, β-actin is the most strongly expressed of the three actin isoforms in mammalian cells, followed by γ-actin. The isoforms' intracellular localizations and functions differ, since β-actin is enriched in the frontal cell lamellipodia (forming dendritic scaffolds), while γ-actin is mostly present in actin arcs and/or stress fiber structures (mainly involved in the cell's adhesive properties). Furthermore, β-actin and γ-actin's functions require interactions with distinct ABPs (for a recent review, see Skruber et al., 2018).

In contrast to mammals, several unicellular organisms have a single actin protein that constitutes all the actin-rich cytoskeletal structures required for life. Here, we focused on Entamoeba histolytica, the protozoan parasite responsible for human amoebiasis. This infectious disease occurs at high incidence in large populations with limited modern sanitation systems. The infestation arises after ingestion of cysts contaminating water and food. Upon de-cystation, a vegetative cell, the trophozoite, is formed that colonizes the intestine or becomes invasive destroying the tissue during the disease process (Marie and Petri, 2014). Entamoeba histolytica's invasive behavior relies on three main activities: motility, adhesion, and cell lysis/toxicity. In this context, the cytoskeleton is responsible for changes in cell shape and other pivotal processes, including motility, phagocytosis of human cells, and parasite-substrate interactions (Guillén, 1993). E. histolytica is a highly motile single cell that can rapidly alter its shape (Dufour et al., 2015). The dynamic reorganization of the cytoskeleton is crucial for all these processes—highlighting its central role in amoebic pathogenesis.

The actin-rich cytoskeleton is the major skeletal component in E. histolytica because its microtubules are solely intranuclear (Vayssié et al., 2004) and intermediate filaments are absent (Clark et al., 2007). In contrast to other eukaryotes (including other amoebae like Dictyostelium discoideum) E. histolytica has a single actin protein only. The initial evidence for the existence of actin in E. histolytica (referred to here as EhActin) was obtained by immunostaining with an antibody against human actin (HsActin) (Kettis et al., 1977). The purified actin protein was unable to bind to DNAse I, in contrast to the majority of known actins (Lazarides and Lindberg, 1974; Gadasi, 1982; Meza et al., 1983). In E. histolytica, cell displacement and other cell functions are correlated with the presence of various actinenriched structures (Bailey et al., 1985; Talamás-Rohana and Meza, 1988; Dufour et al., 2015; Emmanuel et al., 2015). Very few of the many ABPs known to participate in the structural dynamics of actin filaments have been identified in E. histolytica (Meza et al., 2006; Hon et al., 2010). However, a few have been experimentally confirmed: ARPC1, a subunit of the Arp2/3 complex that participates in actin nucleation and filament dendritic organization (Babuta et al., 2015), and is involved in amoebic phagocytosis; formin, which stabilizes microfilaments (Majumder and Lohia, 2008); filamin (ABP-120), which organizes microfilaments into orthogonal networks (Vargas et al., 1996; Díaz-Valencia et al., 2007); and profilin, the G-actin-sequestering protein (Binder et al., 1995). Furthermore, coactosin binds and stabilizes F- actin and regulates microfilaments in E. histolytica (Kumar et al., 2014). Actin-binding protein 16 (a member of the actin depolymerizing factor/cofilin family) is necessary for E. histolytica motility (de la Cruz et al., 2014), whereas the Factin-binding protein NCABP166 translocates into the nucleus and participates in phagocytosis and cell motility (Campos-Parra et al., 2010; Uribe et al., 2012). Due to the presence of a single actin protein in E. histolytica, we hypothesized that the actin-rich cytoskeleton interacts with ABPs that vary according to the subcellular compartment and function. We used a variety of approaches (including bioinformatics analyses and protein structure modeling) to determine the major differences between actin in E. histolytica and human monomeric actins. In transcriptomic experiments, we determined the level of actin gene expression during pathogenesis. A proteomics analysis of the actin-rich cytoskeleton revealed a number of important ABPs. Furthermore, we investigated the origin, composition, and fate of actin-rich structures in E. histolytica (such as "dot-like" structures, adhesion plates, stress fibers and macropinosomes) and imaged their dynamics in living trophozoites. Our experiments provided new insights into (i) the 3D structural major divergences of EhActin compared to HsActin; (ii) the genesis of actin-enriched structures in E. histolytica, and (iii) highlighted an important role for the Arp2/3 actin-nucleation complex in the dynamics of the actinrich cytoskeleton.

# MATERIALS AND METHODS

# Cell Strain and Culture

The E. histolytica strain HM1:IMSS was cultured in TYI-S-33 medium at 37◦C (Diamond et al., 1978). Drug treatments included latrunculin B (100 nM, for 15 min; this compound binds to actin monomers and prevents actin polymerization), jasplakinolide (Sigma-Aldrich, USA, at 10µM, for 30 min; this is a filament polymerizing and stabilizing agent), and 2-fluoro-N- [2-(2-methyl-1H-indol-3-yl)ethyl]-benzamide (CK-666, Sigma-Aldrich, USA at 40µM for 2 h; this compound binds to the Arp2/3 complex, stabilizes it in an inactive state, and prevents the formation of the active conformation). The bacterium Escherichia coli (One Shot TOP10, Thermo Fisher Scientific, USA) was grown in Luria-Bertani medium supplemented with ampicillin (100µg/ml) and used for plasmids amplification.

# Expression of Actin-Encoding Genes in *E. histolytica*

The full-length nucleotide sequences of the eight copies of the actin gene (EHI\_182900, EHI\_159150, EHI\_142730, EHI\_126190, EHI\_140120, EHI\_107290, EHI\_163750, and EHI\_043640) were retrieved from AmoebaDB (the amoeba genomics resource at http://amoebadb.org/amoeba/). The DNA sequences corresponding to the open reading frame or the 5' end of the actin gene (150 bp upstream of the ATG initiation codon) were aligned using Clustal Omega (https://www.ebi.ac. uk/Tools/msa/clustalo/). Values of actin gene expression were obtained from RNA-Seq data of E. histolytica in culture and interacting with human colon during intestinal infection (Weber et al., 2016).

# Structure Modeling of Actin Proteins

The three-dimensional structure of EhActin (Uniprot P11426, 376 residues) or HsActin (Uniprot P68032, 377 residues) was predicted by homology modeling with the I-TASSER package (https://zhanglab.ccmb.med.umich.edu/I-TASSER/). The threedimensional (3D) models were validated with Procheck, Rampage, and Verify\_3D software tools (http://services.mbi. ucla.edu/Verify\_3D/), and visualized using Visual Molecular Dynamics (VMD) software.

# Molecular Dynamics Simulations

We used the CHARMM22 force field and the TIP3P water model from the GROMACS software package (Foloppe and MacKerell, 2000). Proteins were solvated in a cubic box with 1 nm edges; 27,654 water molecules and 13 sodium ions were added for HsActin, while 30,878 water molecules and 14 sodium ions were added for EhActin. Molecular dynamics (MD) simulations were conducted with periodic boundary conditions in an isobaricisothermal ensemble, at 300 K and 0.1 MPa for 150,000 ps. These parameters are typically used to mimic experimental conditions. The coordinates and energy data were stored every 1 ps. The atomic characteristics of HsActin and EhActin proteins were compared using the analysis tools included in the GROMACS software. The root mean square deviation (RMSD) and root mean square fluctuation (RMSF) of the backbone were calculated. The change over time in the secondary structures of both proteins was followed using the do\_dssp tool in GROMACS. For these analyses, the time at which the RMSD converged was considered to be the initial step (40,000 ps for HsActin, and 10,000 ps for EhActin) for the production of simulations.

# Immunofluorescence Assays

Trophozoites (4.5 × 10<sup>4</sup> ) on slides were incubated under anaerobic conditions overnight at 37◦C in TYI-S-33 medium using an Genbag Anaer (catalog number 45,534, Biomérieux, France). The medium was removed, and the cells were fixed with 37◦C-prewarmed buffer containing 10 mM PIPES pH 7.4, 3 mM MgCl2, 1 mM EGTA pH 8, 1 mM DTT and 4% PFA (for 30 min at room temperature (RT). The fixed trophozoites were permeabilized with 0.05% Triton X-100 in PBS for 1 min. Next, the slides were washed with PBS and quenched with 50 mM NH4Cl for 15 min. After blocking with 2% BSA for 1 h, the slides were incubated with primary antibody for 2 h at RT, washed with PBS, and incubated with secondary antibody or phalloidin for 1 h at RT. The primary antibodies included a mouse monoclonal anti-actin antibody (clone C4, Merck Millipore, Germany, 1:200 dilution) and a rabbit polyclonal anti-Arp3 antibody (1:200 dilution, generated in the present study - see below). The secondary antibodies were goat anti-rabbit Alexa Fluor-488 or Alexa Fluor-546 antibodies, and goat anti-mouse Alexa Fluor-546 or AlexaFluor-488 antibodies (Molecular Probes, 1:200 dilution). To decorate microfilaments, we used phalloidin Alexa Fluor-488 or phalloidin Alexa Fluor-546 (Sigma-Aldrich, USA, 1:200 dilution). Lastly, coverslips were washed with BSA-free PBS, mounted with ProLong antifading reagent containing DAPI (Molecular Probes, USA), and observed under a confocal microscope (LSM700, Zeiss, Germany). Confocal planes were acquired in Z-stacks (step size: 0.5µm) and structures were quantified in randomly selected cells from 10 fields, using a 63X objective, NA = 1.4.

The anti-Arp3 antibody was raised in rabbits (Eurogentec, Belgium) by immunization with the purified peptides 252- FKKHQAIDPISKKP and 334-LQRDURRFTDFRJQK from the amino acid sequence of E. histolytica Arp3 protein (locus tag: EHI\_198930 and XP\_647871).

# Image Analysis and Quantification

All image analyses were conducted using open-source Icy software (http://icy.bioimageanalysis.org; de Chaumont et al., 2012). The "Ruler Helper" plug-in was used to measure the size of structures. Colocalization of actin with the various cell markers was quantified using the "Colocalization Studio" plugin and calculating Pearson's correlation coefficient with p < 0.05. The strength of association is considered as small (PS = 0.1 to 0.3); medium (PS = 0.3–0.5) or large (PS = 0.5–1.0). Specific regions of interest (ROIs) within the cells (i.e., adhesion plates and macropinosomes) were taken from raw image data for 54 randomly selected trophozoites. A single, representative cell from this subset is shown the results presented below.

# Construction of a Recombinant Actin-HaloTag Plasmid

The nucleotide sequence of the E. histolytica HM1:IMSS strain's actin gene (EHI\_163750, 1,113 bp) with BglII restriction sites at the 5′ and 3′ ends was synthetized by Eurofins (Brussels, Belgium). The stop codon was silenced by replacing the TAA codon by an arginine AGA codon within the BglII endonuclease site. Furthermore, the internal BglII restriction site was eliminated by the silent G255A mutation, which does not affect the protein's amino acid sequence. The gene was cloned into a pEX-K4 vector and propagated in E. coli. The purified recombinant plasmid was digested with BglII restriction endonuclease, and the resulting fragment was ligated into the BglII restriction site of the pHalo-Tag vector (kindly provided by Professor Tomoyoshi Nozaki) so that the HaloTag (Promega, USA) would be fused to the C-terminus of the actin protein. The construct's orientation and sequence were checked by endonuclease digestion and DNA nucleotide sequencing. The recombinant pEhEx-ActinHaloTag plasmid and the empty pHalo-Tag vector were transfected into trophozoites, as described below.

# Transfection Assays

Entamoeba histolytica trophozoites were transfected according to a modified version of a previously published protocol (Penuliar et al., 2012). Briefly, 2 × 10<sup>5</sup> trophozoites in TYI-S-33 medium were seeded on a six-well plate, and incubated under anaerobic conditions using a Genbag Anaer (catalog number 45534, Biomerieux, France) at 37◦C until the cells reached 80% confluence (usually less than 24 h). Transfection was performed in OPTI-MEM media (Thermo Fisher Scientific, USA) containing L-cysteine (5 mg/ml) and ascorbic acid (1 mg/ml) adjusted to pH 6.8 (referred to as transfection medium, TM). Actin-HaloTag or HaloTag plasmid DNA (4 µg) diluted in TM (final volume: 30µl) was mixed with 15µl of Lipofectamine 3000 (Thermo Fisher Scientific, USA) and incubated at RT for 15 min. Next, 960µl of TM were added to the mixture. After removing the TYI-S-33 medium, the DNA/TM/Lipofectamine mixture was added to the trophozoites, and the plate was incubated at 37◦C for 3 h under anaerobic conditions. The mixtures were then transferred into a 14 ml glass tube containing TYI-S-33 medium. After overnight incubation, the transfection mixture was eliminated, fresh TYI-S-33 medium was added, and the cultures were incubated at 37◦C for 24 h. Lastly, increasing amounts (2, 5, and 10 µg) of G418 (Sigma-Aldrich, USA) were added for the drug selection of transfected cells.

# Protein Detection by Immunoblotting

Total protein extracts were obtained from 10<sup>6</sup> amoebae incubated in lysis buffer (10% SDS, 10 mM Tris/HCl) containing a protease inhibitor cocktail (20 mM leupeptine (Sigma-Aldrich, USA), 50 mM N-ethylmaleimide (Sigma-Aldrich, USA), 5 mM p-chloromercuribenzoate (Sigma-Aldrich, USA), 2 mM 4-(2 aminoethyl) benzenesulfonyl fluoride (Sigma-Aldrich, USA), 2× complete mini EDTA-free (protease inhibitor cocktail, Roche, Suisse), a tablet of phosSTOP (Roche, Suisse), 10 mM E-64, 2 mM Na3VO4, 100 mM NaF, 10 mM iodoacetamide and 1% SDS). Crude cell extracts were boiled for 5 min at 100◦C, cooled in ice for 1 min, aliquoted, and stored at −20◦C. Protein samples (equivalent to 80,000 cells per lane) were resolved by 10% SDS-PAGE and electrotransferred onto a 0.2µm PVDF membrane (Immobilon PSQ, Millipore). Proteins were detected by immunoblotting using mouse antiactin monoclonal antibody (1:50 dilution; Clone C4, Merck Millipore, Germany), anti-HaloTag mouse monoclonal antibody (1:200 dilution, Promega, USA) or polyclonal anti-Arp3 (1:500 dilution, this work); the secondary antibodies were sheep peroxidase-conjugated anti-mouse (1:10,000; G&E) or antirabbit (1:20,000; G&E, USA) IgG. Membranes were treated with ECL Western blotting detection reagent and then exposed to Kodak Biomax film.

# Actin-HaloTag Expression and Cell Imaging

Transfected trophozoites were grown overnight at 37◦C in TYI-S-33 medium. The medium was then replaced by incomplete (serum-free) TYI-S-33 medium (TYIi) and labeled with 1µM HaloTag tetramethylrhodamine (TMR) ligand (Promega, USA) for 15 min. After washing with TYIi, trophozoites were suspended in TYIi and seeded on 35 mm glass-bottomed imaging dishes (Ibidi, France). Images were recorded with a spinning disk confocal microscope (UltraVIEW VoX, Perkin Elmer, USA; excitation: 561 nm; objective: 63x; temperature control set to 37◦C. Images were acquired with Volocity 3D image analysis software (Perkin Elmer, USA). In some experiments, 40µM CK-666 (Baggett et al., 2012) was added for 2 h before image acquisition.

# Cell Fractionation and the Recovery of Actin and Its Partners by Immunochromatography

Initially, 7 × 10<sup>6</sup> trophozoites were washed twice with 4◦C cold PBS containing 5 mM EGTA, recovered, and lysed with 800µL of lysis buffer (60 mM PIPES pH 7, 25 mM HEPES, 125 mM KCl, 2 mM MgCl2, 5 mM EGTA pH 8, 1% Triton-X-100, 0.5 mM ATP) in the presence of protease inhibitors (as described above). The mixture was centrifuged at 500 × g for 15 min at 4 ◦C and the recovered supernatant was further centrifuged at 100,000 × g for 1 h at 4◦C. Next, the pellet (i.e., the fraction not soluble in Triton-X100) was suspended in 800µL of Trisglycerol buffer (125 mM Tris and 20% glycerol without SDS and β-mercaptoethanol). For the immunochromatography assay, 40µL of iron beads coupled to proteins A and G (PAG-beads, Ademtech, France) were placed in a tube on a magnetic unit until a pellet had been formed. The pellet was washed twice with 0.65% Tween 20 in PBS pH 7.5. Next, 7 µg of anti-actin C4 antibody were added for 60 min, at 4◦C with agitation (1,000 rpm). Anti-GST (an irrelevant antibody) was used as a control. The tube was placed on the magnetic unit until a pellet had formed, the supernatant was removed, and the beads were recovered in 40µl of 200 mM triethanolamine pH 9.0. The tube was again placed on the magnetic unit until a pellet had formed. The supernatant was removed, and the pellet of beads/antibody were suspended in 200µl of 20 mM dimethyl pimelimidate dihydrochloride dissolved in 200 mM triethanolamine pH 9.0, and incubated for 60 min with agitation (1,000 rpm) at 4◦C. The tube was placed on the magnetic unit, and the supernatant was removed. The reaction was stopped by adding 40µL of 50 mM Tris pH 7.5, and the mixture was incubated for 30 min with agitation (1,000 rpm) at 4◦C. The beads/antibody were washed twice and suspended in 40µl of 50 mM glycine, 0.65% Tween 20, pH 2.7. The tube was placed on the magnetic unit. The beads were recovered in the Tris-glycerol buffer previously described, and then mixed with the above-mentioned Triton X-100 insoluble fraction overnight at 4◦C with agitation (1,000 rpm). The tube was placed on the magnet unit, and the supernatant was recovered as a flowthrough sample. The beads/antibody/actin were washed three times, and the supernatant was recovered as wash samples. Fifteen microliters of PAG elution buffer were added to the beads and mixed for 2 min. The supernatant containing actin and its partners was recovered as the elution sample. Lastly, 10µl of 50 mM Tris pH 7.5 was added to neutralize the acidic samples.

# Protein Analysis by Liquid Chromatography Coupled to Tandem Mass Spectrometry (LC-MS/MS)

The elution samples were loaded on a 12% acrylamide gel with a 4% stacking gel and electrophoresis was run until the samples had reached the separating gel. The proteins were then excised and processed for identification with LC-MS/MS, using a standard protocol (Shevchenko et al., 2006; Perdomo et al., 2016) and an Orbitrap Velos instrument (Thermo Fisher Scientific, USA) connected to a nanoUltimate 3000 HPLC system (Dionex). Mass spectrometry peak lists were generated from the raw data files using Proteome Discoverer version 1.2 (Thermo Fisher Scientific, USA). The resulting peak lists were searched with Mascot v.2.2 (Matrix science, London, UK) against the E. histolytica HM1:IMSS protein database from Uniprot (http://www.uniprot.org/) concatenated with known contaminants and reversed sequences of all entries. Peptide identifications were accepted if the probability in the Peptide Prophet algorithm was greater than 95.0% (Keller et al., 2002). The list of proteins was visualized and retrieved using Scaffold software (http://www.proteomesoftware.com/products/ scaffold/). The various categories of identified proteins were obtained by searching with PANTHER tools (http://pantherdb. org) and InterProScan (https://www.ebi.ac.uk/interpro). Venn diagrams were generated using Venny software (http://bioinfogp. cnb.csic.es/tools/venny/index.html). The reference E. histolytica genome was retrieved from the AmoebaDB (http://amoebadb. org/amoeba/).

# RESULTS

# Actin-Encoding Genes in *E. histolytica*

We previously identified eight copies of the actin coding gene in E. histolytica (EHI\_182900, EHI\_159150, EHI\_142730, EHI\_126190, EHI\_140120, EHI\_107290, EHI\_163750, and EHI\_043640), and determined that the predicted protein is phylogenetically related to amoebazoan (e.g., Dictyostelium discoideum) and parabasalid actins (e.g., Trichomonas vaginalis) (Hon et al., 2010). To study these genes, we first retrieved the corresponding nucleotide sequences from AmoebaDB and aligned them using Clustal. The gene EHI\_043640 appeared as a truncated version, and so we discarded it from further analyses. The seven full-length nucleotide sequences gave a high alignment score—indicating that they are extremely conserved and showed 93–99.7% homology (**Supplemental Datasheet 1**) with a total of only 17 nucleotide mismatches. The seven genes encode a protein that is 100% homol. Moreover, alignment of the 150 bp region at the 5' end of each gene also highlighted major sequence similarities between the seven full-length genes. This was particularly true for the nucleotides near the transcription initiation site, which suggests the coordinated regulation of actin gene expression (**Supplemental Datasheet 1**). In previous work, it was concluded that the actin gene's 5' untranslated region (UTR) contains a cAMP-response element (CRE, with the palindromic sequence TGACGTCA) and a serum-response element (SRE) box (with the CC(A/T)nGG motif) (Ortiz et al., 2000). Both motifs are involved in actin gene transcription in various eukaryotic cells. Our present analysis demonstrated that only the EHI\_182900 locus carries the indicated motifs. Overall, these findings indicate that E. histolytica's actin genes may correspond to recent duplications of an ancestral gene within the amoeba genome.

To investigate actin gene expression, we took advantage of our literature transcriptome data concerning E. histolytica in culture or during parasite interaction with the human intestinal colon (Weber et al., 2016). Although the nucleotide sequence (reads) in the amoeba genome were mapped stringently as single reads (meaning that a read is allowed to map only to one gene), the high level of nucleotide sequence conservation in actin genes prevented us from determining the specific features of transcription for each individual gene. However, taking these data as a whole, we found that the seven full-length copies of actin gene are expressed in parasites cultured in vitro and

E. histolytica transcripts in cultured trophozoites (blue bars) or after seeding onto a human colon explant (red bars) by Weber et al. (2016) (the accession numbers are taken from AmoebaDB). (B) Counted reads. In cultured trophozoites, each of the seven actin-encoding genes gave similar numbers of reads. This was also true for trophozoites seeded onto a human colon explant. There was a 2.5-fold overall increase in actin transcription during pathogenesis. The numbers of reads are quoted with the respective probability values and adjusted probability values. FPKM: fragments (reads) per kilobase of transcript per million fragments mapped.

during infection of the human intestine. A marked (2.5-fold) overall upregulation of gene expression was observed during tissue invasion (**Figure 1**).

# Dynamic Structure of Monomeric EhActin vs. Monomeric HsActin

To gain insight into the structural differences between EhActin and HsActin, we first conducted in silico molecular modeling experiments. Given that the full-length amino-acid sequences of E. histolytica and human proteins share 86% identity, this modeling was relatively straightforward. Overall, the two proteins showed a highly conserved folding pattern, with an RMSF of 0.31 Å. As in the human protein, amoebic actin has 17 beta strands and 15 alpha helices, which form the four typical domains seen in actins. However, there are differences between the respective 3D structures—mainly in the subdomain II (**Figure 2A**), as a result of the substitution of non-polar residues in HsActin by polar residues in EhActin (V42/Q43; A46/S47) and, conversely, the substitution of polar residues by non-polar residues (S43/G44) (**Figure 2B**). Interestingly, Q43 and G44 in amoebic actin correspond to Q40 and G41 in HsActin, which are involved in the DNAse I interaction (Wriggers and Schulten, 1997). In addition, G41 is involved in HsActin interaction with Thymosin ß4, an actin sequestering protein that prevents G-actin association to microfilaments (Domanski et al., 2004).

To establish whether these amino acid sequence differences have an effect on protein behavior, we performed MD simulations. We notably described the proteins' flexibility and compared structural parameters obtained from the dynamic trajectories. Amoebic actin showed a rapid increase in the

FIGURE 2 | Comparison of 2D amino acid sequences and predicted 3D structures of EhActin vs. HsActin. (A) The 3D structure of EhActin (in blue, Uniprot P11426) was obtained with the I-TASSER package using the 3D structure of HsActin (in red, Uniprot P68032) as the template. Amino acid alignment starts at amino acid D3 in EhActin. The typical subdomains I, II, III, and IV are indicated. (B) Comparison of amino acid sequences of both actins after alignment using BLASTp. Note the major differences at the amino-terminus.

RMSD during the first 5,000 ps, whereas this production step required about 30,000 ps for HsActin. The proteins then reached equilibrium and oscillated within an interval of 0.25–0.35 nm; there were no significant interval differences between the two (**Figure 3A**) (**Supplemental Videos 1, 2**). The respective RMSF profiles revealed that subdomains II and IV are the most mobile parts in both proteins, followed by subdomain I (**Figure 3B**). The same conclusion can be drawn from observation of the average 3D structure in MD simulations (**Figure 3C**). Importantly, a detailed analysis of RMSD values showed that the region spanning residues 40–50 (corresponding to subdomain II) is more mobile in EhActin than in HsActin. The next most mobile region includes residues 190–220 and 225–250 of subdomain IV (**Figure 3B**). Overall, the 3D structural dynamics data indicate that subdomain II is the protein region with the greatest divergence between the two actins.

# Actin and Microfilaments Form Diverse Actin-Rich Structures

The presence of a single actin protein in E. histolytica and the marked 2D and 3D structural differences with respect to HsActin prompted us to investigate the dynamics of the actin-rich structures in this parasite. In particular, the stress fibers that are preponderant in eukaryotic cells are less visible in the highly mobile amoeba. To determine the distribution of actin and the various actin-rich structures in E. histolytica trophozoites, we used confocal microscopy and immunodetection with the C4 monoclonal anti-actin antibody that binds to all forms of actin in eukaryotic cells (including unicellular organisms like E. histolytica) (Lessard, 1988). We also used phalloidin to detect F-actin-containing microfilaments. In each scanned cell (approximately 45µm long and 22µm wide), we identified and counted actin-rich structures over a distance of 10µm (20 focal planes of 0.5µm; **Figure 4**). Two ROIs were defined in each trophozoite, i.e., a lower part (focal planes 1–10) and an upper part (focal planes 11–20). In both ROIs, the cortical cytoskeleton was clearly seen around the cell membrane. The lower part of the cell contained stress fibers, dot-like contact points, and large adhesion plates, whereas the upper part of the cell contained small cytoplasmic dots and multiform vesicular structures. The latter included large, endocytic vacuoles reminiscent of the macropinosomes through which cells internalize fluid. We then quantified the number and size of actin-rich adhesive

FIGURE 3 | Molecular dynamics (MD) simulation of EhActin and HsActin. Backbone RMSD (A) and RMSF (B) values of EhActin (in red) and HsActin (in black). (C) Representation of average 3D structures obtained from MD simulations of EhActin (left) and HsActin (right). The colored scale bar indicates the intensity of movements within protein 3D domains; these movements are presented in Supplemental Videos 1, 2.

and macropinosome-like structures in 54 trophozoites by analyzing the confocal microscopy images after fluorescence staining (**Figure 5**). We found that 23 of the 54 E. histolytica trophozoites displayed large adhesion plates (with an external diameter of up to 22µm and an internal diameter of up to 16µm) containing actin stained by both the anti-actin antibody and phalloidin. Furthermore, 51 of the 54 trophozoites contained macropinosomes, which measured (on average) 5µm in length and 3µm in width. When we used latrunculin B and jasplakinolide to inhibit microfilament dynamics, the actinrich structures disappeared; most of the actin aggregated inside the cells (**Supplemental Figure 1**), and the trophozoites became more spherical.

# Dynamics of Actin-Rich Structures in Living *E. histolytica*

To examine the dynamics of actin-rich structures in E. histolytica, we transfected trophozoites with the actin-HaloTag construct as described in the Material and Methods section. DNA plasmid sequencing confirmed the quality of the construct, in which HaloTag is fused to the carboxy-terminal end of actin. We used confocal microscopy to check that the actin-HaloTag fusion protein was present in trophozoites and that it colocalized with actin (**Figure 6**). As expected, actin-HaloTag was bound by both the anti-HaloTag antibody and HaloTag's TMR fluorescent ligand. Structures such as macropinosomes, "dot-like" structures, stress fibers and large adhesion plates contained the actin-HaloTag fusion protein and F-actin (as identified by staining with phalloidin or anti-actin antibody) (**Figure 6**). Transfected and non-transfected cells did not appear to differ with regard to the abundance of actin-rich structures. The presence of the actin-HaloTag fusion protein was further examined by immunoblotting crude protein extracts from transfected trophozoites (**Supplementary Figure 2A**). Although the fusion protein was recognized by the anti-HaloTag antibody and had the right molecular mass, we did not find a protein signal when the anti-actin antibody was used in the immunoblot suggesting that when extracted from cells, the fusion protein may have a lower affinity for the C4 antibody.

We next imaged actin-rich structures in live cells (transfected with actin-HaloTag or HaloTag) using video microscopy and a spinning disk laser microscope. Remarkably, the various actin-rich structures changed very rapidly over time.

**Figure 7A** shows a series of micrographs from various amoebae, highlighting the polymorphism of the actin-rich structures. **Figure 7B** depicts the course of events in a single amoeba (**Supplemental Video 3**). Actin-HaloTag was present, and gave a strong fluorescent signal in cell-adhesive structures (dots and plates), vacuoles, macropinosomes, and stress fibers. These fibers were present at the rear of the polarized cells and near to the plasma membrane. No signals were observed in control cells (e.g., wild-type cells or Halo-Tag-treated cells). Actin-HaloTag was recruited to the macropinocytic cup in less than 9 s. Next, the internalized macropinosomes that formed after 20–30 s were entirely covered by actin, and moved through the cytoplasm until actin disappeared from the vesicle. In some cases, the vesicle appeared to fuse with other resident endomembrane compartments. Overall, the macropinosomes' turnover time was roughly 50–60 s—indicating that actin-rich cytoskeleton is rapidly reorganized within these structures.

# Actin and ABPs in the *E. histolytica* Cytoskeleton

To gain insight into the ABPs associating with the amoebic cytoskeleton, we purified actin and its partners by immunoaffinity chromatography of the cytoskeleton-enriched protein fraction. The recovered proteins were identified using LC-MS/MS. Three independent experiments were performed, and only proteins with at least two peptides were considered in the subsequent bioinformatics analysis. Using a Venn diagram, we determined that 266 proteins were present in two or three independent experiments. The PANTHER tools and manual annotations enabled us to identify cytoskeletonrelated proteins, small GTPases involved in cytoskeleton regulation, and proteins related to membrane traffic. The protein families and classes are listed in **Supplemental Table 1**. By applying stringent criteria, we focused on 14 proteins that bound to the actin-rich cytoskeleton (**Table 1**). Along with actin, the most strongly represented family was the Arp2/3 complex involved in actin nucleation and which contains seven actin-related proteins: Arp2 (44 kDa), Arp3 (47 kDa), ARPC1 (40 kDa), ARPC2 (35 kDa), ARPC3 (21 kDa), ARPC4 (20 kDa), and ARPC5 (16 kDa). All the subunits other than ARPC3 and ARPC5 were found in the proteomic analysis. Arp2 and Arp3 fold into a structure that is similar to that of actin, and act as monomer nucleators. We identified a number of other proteins: the heavy chain of myosin II, which binds F-actin and is responsible for cell contraction and motility (Arhets et al., 1998); filopodin, which contains a FERM domain involved in protein membrane binding (Chishti et al., 1998) and was previously found in the uropod of moving E. histolytica (Marquay Markiewicz et al., 2011) and in phagosomes (Marion et al., 2005); profilin, which sequesters G-actin and thus blocks actin polymerization (Binder et al., 1995); coronin, which participates in cell migration and vesicular trafficking in eukaryotes (de Hostos et al., 1991) but has not been studied in E. histolytica; and cyclase-associated protein (CAP), a highly conserved protein that links nutritional response signaling to the cytoskeleton via its actin-binding carboxy-terminal end (Iwase and Ono, 2016). We also found 12 small GTPases (**Supplemental Table 1**), including 7 Rabs and 3 Rhos that are all expected to regulate cytoskeleton function and vesicle trafficking. Lastly, 11 proteins were associated

and show actin-HaloTag (in red) detected by TMR, and G-actin and F-actin (in green) detected using an anti-actin antibody or fluorescent phalloidin, respectively. Scale bar: 10µm. Both types of actin colocalized in actin-rich structures including macropinosomes (1 and 4), large adhesion plates (2 and 5), stress fibers, and "dot-like" structures (4 and 6). F-actin exhibited low anti-HaloTag staining. Colocalization (Pearson's correlation coefficient: 0.67, 0.51, 0.47, 0.48, 0.38, and 0.32, respectively) within the structure was assessed. These values indicate a significant level of colocalization, with the lowest level observed for stress fibers.

with the endomembrane system traffic, including endocytic compartments (**Supplemental Table 1**).

# Arp2/3 Is Involved in Macropinosome and Adhesion Plates Formation

Due to the prominence of Arp2/3 within the actin-rich cytoskeleton (as analyzed by proteomics), we studied the complex's involvement in the dynamics of the above-mentioned actin-rich structures. To this end, we first prepared an anti-Arp3 antibody (see the Materials and Methods section). In immunoblots of an E. histolytica crude extract, the antibody bound to a single 47 kDa protein as expected (**Supplemental Figure 2B**). To examine the localization of Arp3 protein in trophozoites, we performed confocal microscopy

FIGURE 7 | (diameter: 5µm). (2) Early-stage macropinosomes (diameter: 5µm). (3) A late-stage macropinosome and a new invagination (diameter: 5µm). (4) Adhesion plates (large structures; diameter: 10µm). (5) An intracellular vacuole (diameter: 2µm). (6) "Dot-like" structures; diameter: 1 to 3µm). (7) Actin filaments at the rear of a polarized cell. Scale bar: 10µm. (B) Micrographs of a single motile trophozoite (Supplemental Video 3) in which all the actin-containing structures appear one after the other. Five different events were selected as examples. Note (i) the initial accumulation of actin at the cortical cytoskeleton (1a, 2a, 4a), (ii) cell surface deformation and actin ring formation (all b lines), (iii) the closure of the macropinosome (1c, 3c, 4c), (iv) the migration of an actin-rich macropinosome to the internal cell compartments (2d, 3d, 4d) and (v) the detachment of actin from the macropinosome. In event 5, a large adhesive plate forms at the rear of the cell (5c). Scale bar: 10µm.

TABLE 1 | Proteins identified in the cytoskeleton enriched fraction.


and immunofluorescence experiments on wild-type cells and actin-HaloTag transfected cells. The results showed that Arp3 colocalizes with G-actin and F-actin in macropinosomes, dots, adhesion plates, and the cortical cytoskeleton but not in stress fibers (**Figure 8I**). CK-666 is a recently discovered small-molecule inhibitor of the Arp2/3 complex that binds at the interphase between the subunits and stabilizes an inactive conformation (Baggett et al., 2012). Treatment with CK-666 greatly reduced the number of adhesion plates and macropinosomes per cell (respectively seen in 3 and 6 of the 28 cells), whereas the number of stress fibers did not change (**Figure 8II**). These findings highlight the implication of Arp2/3 complex in adhesion plates and macropinosomes formation.

# DISCUSSION

Entamoeba histolytica's ability to move, divide, kill, and phagocytose human cells requires actin, which is widely distributed throughout the amoebic cytoplasm. The continuous polymerization/depolymerization of actin enables the deployment of several different structures. The E. histolytica genome carries seven full-length actin-encoding genes but (unlike higher organisms) expresses a single, conventional isoform. The selective evolutionary advantage of maintaining multiple copies of genes coding for the same protein has not been elucidated. Nevertheless, in view of the nucleotide sequence similarities between the actin genes, one can reasonably hypothesize that the actin gene family arose from duplications of

an ancestral gene. As in the case of actin genes in E. histolytica, gene duplication is also seen in E. dispar and E. invadens (Hon et al., 2010). In phylogenetic terms, amoebic actin is closely related to actin from Trichomonas vaginalis (another infectious parasite) and the free-living amoeba Dictyostelium discoideum.

The actin genes have sequence similarities in the 5'UTR, which suggests the existence of common transcription factor binding sites. In previous studies, potential regulatory DNA motifs (e.g., CRE and SRE) were described in actin's 5'UTR (Ortiz et al., 2000); here, we conclude that these motifs are only present at the EHI\_182900 actin locus. Nevertheless, all seven full-length loci share consensus DNA stretches at regions very close to the transcription start site. Potentially these common DNA stretches may regulate all actin genes transcription. This proximity is observed for many genes in E. histolytica's very compact genome (Loftus et al., 2005). Although actin gene expression is abundant in E. histolytica, it is difficult to draw conclusions as to the specific contribution of each gene to the actin mRNA levels. Nevertheless, invasion of the human intestine led to a 2.5-fold overall increase in actin gene transcription. The increase in actin mRNA abundance may be justified by the potential need for more protein during infection-related processes such as cell motility and phagocytosis. Furthermore, it is possible that actin itself regulates the gene expression profile during the infectious process; indeed, monomeric nuclear actin reported regulates gene expression, modifies nuclear content, and maintains genome integrity in eukaryotic cells (Virtanen and Vartiainen, 2017).

We found that most of the structural differences between EhActin and HsActin concerned domain II of the protein. In particular, EhActin and HsActin differ with regard to the amino acids involved in the interaction with DNAse I. Subdomain II is also involved in the G-actin interaction with thymosin β4 which sequesters actin monomers, how the instability of subdomain II can influence this interaction in E. histolytica is an open question. This observation suggests that the loop between amino acids 30 to 52 in monomeric actin is a prime potential target for exploitation in drug screening. Two loops in domain IV (at amino acids 190–220 and 225–250) are also good potential targets.

Live cell imaging enabled us to establish the first imagebased atlas of actin-containing structures in E. histolytica: actindots, adhesion plates, and macropinosomes. It has been reported that actin dots accumulate on the ventral side of trophozoites; this process has been linked to signal pathways that depend on extracellular matrix cell surface receptors (e.g., fibronectin (FN) receptors) regulated by the activity of the small GTPase Rab21 (Emmanuel et al., 2015). The presence of these dots in the non-FN-activated amoebae studied here suggests that they accumulate when the cell's displacement is slowed down by adhesive FN signaling. Adhesion plates have been previously observed in E. histolytica seeded on FN (Vázquez et al., 1995). The plates are enriched in cytoskeletal proteins such as actin and the associated myosin I, myosin II, alpha-actinin, and tropomyosin.

Entamoeba histolytica accumulates fluid-phase markers by macropinocytosis (Meza and Clarke, 2004); here, we observed a link between macropinocytosis and the actin-rich cytoskeleton. By taking advantage of the properties of a fusion protein between actin and HaloTag (a fluorescent tag that is well suited to use in anaerobes), we gained information on the dynamics of macropinocytosis in E. histolytica. The vacuoles form and circulate in as little as 50–60 s. Actin then detaches from the vesicle, which eventually fuses with the resident internal endomembrane system. Macropinocytosis differs from phagocytosis, and is conserved among amoeboid eukaryotic cells. A large body of evidence indicates that this feeding phenomenon depends on Ras signaling pathways. To form, macropinosomes start by accumulating actin beneath the plasma membrane in order to produce a ring. Upon membrane ruffling, the ring closes by membrane fusion to produce an internal vesicle filled with external fluid (for a recent review, see Bloomfield and Kay, 2016). These large structures (up to 5µm in diameter) occur in a wide range of cell types, including immune cells (e.g., macrophages and dendritic cells) and amoebae. It was recently observed that macropinocytosis and cell motility are not compatible in dendritic cells during antigen presentation (Chabaud et al., 2015) or in D. discoideum during chemotaxis (Veltman et al., 2014). This is not the case in E. histolytica, since our video microscopy studies showed that macropinosomes formed in highly motile cells. However, we have not yet looked at whether chemotaxis which is important for pathogenesis in E. histolytica—competes with macropinosome formation.

Our rigorous proteomic study highlighted the abundant presence of the Arp2/3 complex in the cytoskeleton fraction. This complex binds pre-existing microfilaments, induces an actin nucleation step, and this starts the filament branching process. The new branch forms at an angle of 70◦ to the supporting filament, leading to a so-called dendritic network of actin (Pollard, 2016). We found that the Arp3 subunit of the Arp2/3 complex in E. histolytica is associated with actin dots, adhesion plates and macropinosomes but not stress fibers. Experiments with the inhibitor CK-666 showed that the Arp 2/3 complex was essential for the formation of adhesion plates and macropinosomes. The functional study of macropinocytosis in E. histolytica is in its early stages. Based on the actin-HaloTag construction and the information gained from our proteomic studies, we hope to further elucidate the role of this nutrient pathway in the parasite's invasion of the intestine.

# AUTHOR CONTRIBUTIONS

MM, NH-C, JO-V, and SS performed the experiments; JO-V, LM, and NG performed the bioinformatics analyses; MM, J-CO-M, LM, and NG designed the study and wrote the manuscript.

# FUNDING

This study was funded by the French National Agency for Research (grant: ANR-10-INTB-1301, PARACTIN), the French Parasitology consortium Labex ParaFrap (grant: ANR-11- LABX0024), and the Labex IBEID (grant: ANR-10-LABX-62- IBEID). The Icy platform is part-funded by an ANR grant under the France-BioImaging national infrastructure program (ANR-10-INBS-04-06 France-BioImaging). We also warmly acknowledge funding from the Mexico-France ECOS NORD (M14S02) and SEP-CONACYT-ANUIES (249554) programs. JO-V received a scholarship from the Mexican BEIFI-IPN and CONACyT programs.

# ACKNOWLEDGMENTS

We thank Professor Tomoyoshi Nozaki (National Institute of Infectious Diseases, Tokyo, Japan) for the kind gift of the vector containing the HaloTag. We are grateful to Pascal Roux and Jean-Ives Tinevez (BioImagerie Photonique Technical Unit, Institut Pasteur) for their advice on and help with confocal and video microscopy experiment, and to Magalie Duchateau and Mariette Matondo (proteomics facility, Institut Pasteur) for their help with proteomics. We thank Professor Artur Sherf (Institut Pasteur) for his constant interest in our project, and the members of the Bioimage analysis unit (Institut Pasteur) for their advice on image analysis.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00179/full#supplementary-material

Supplemental Figure 1 | E. histolytica actin-rich structures belong to the actin-rich cytoskeleton. Confocal image micrographs from Z-Stack show G-actin (in red, detected by anti-actin antibody) and F-actin (in green, detected by fluorescent phalloidin). Scale bar: 10µm. (A) Wild-type trophozoites. (B) Trophozoites treated with latrunculin B or jasplakinolide. Scale bar: 10µm. Treated amoebae displayed an aggregation of actin filaments and adopted a spherical shape, due to the inhibition of actin cytoskeleton dynamics.

Supplemental Figure 2 | Expression of actin-HaloTag in E. histolytica trophozoites. A sample of amoebic extracts (equivalent to 80,000 cells) was resolved by SDS-PAGE and immunoblotted as described in the Material and Methods section. (A) Electrophoretic analysis of proteins from transfected and wild-type amoebae, for the detection of actin and actin-HaloTag. Actin (WT, 42 kDa), HaloTag (H, 34 kDa), and actin-HaloTag (AH, 76 kDa) were detected in the respective protein extracts. As expected, only endogenous actin was detected in both transfected and wild-type amoebae. The anti-HaloTag antibody detected an additional band (at roughly 30 kDa) in the actin-HaloTag protein extract; this may reflect degradation of the fusion protein. Line1 corresponds to Actin-HaloTag cells;

# REFERENCES


line 2 is HaloTag cells and line3 is wild-type amoeba. (B) Electrophoretic analysis of proteins from wild-type amoeba for the detection of Arp3. A single band was seen at 47 kDa.

Supplemental Table 1 | Proteins identified by proteomics.

Supplemental Video 1 | Molecular dynamics simulation of HsActin. The video was made with VMD software and shows the behavior of human actin during 150,000 ps of simulation.

Supplemental Video 2 | Molecular dynamics simulation of EhActin. This video was made with VMD software and shows the behavior of EhActin during 150,000 ps of simulation.

Supplemental Video 3 | Changes over time in actin-HaloTag structures. Transfected trophozoites (actin-HaloTag) in TYIi medium at 37◦C were imaged using a spinning disc confocal microscope. All images were acquired using a 63X objective in a selected confocal plane. For video reconstruction, 582 frames were taken into account and the video speed is 15 frames per second. The scale bar corresponds to 10µm.

Supplemental Datasheet 1 | Nucleotide sequence analysis of actin-encoding genes in E. histolytica HM1:IMSS strain. The DNA sequences were taken from AmoebaDB and aligned using CLUSTAL. The document indicate identity (stars) and homology (double dots) between the DNA sequences in coding regions (A) and promoter regions (150 bp upstream of the ATG initiation codon) (B).


protein KERP1 reveals features of endomembrane organization in Entamoeba histolytica. Cell. Microbiol. 18, 1134–1152. doi: 10.1111/cmi.12576


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Manich, Hernandez-Cuevas, Ospina-Villa, Syan, Marchat, Olivo-Marin and Guillén. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Influence of Micropatterned Grill Lines on *Entamoeba histolytica* Trophozoites Morphology and Migration

Francisco Sierra-López, Lidia Baylón-Pacheco, Patricia Espíritu-Gordillo, Anel Lagunes-Guillén, Bibiana Chávez-Munguía and José L. Rosales-Encina\*

Departamento de Infectómica y Patogénesis Molecular, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico City, Mexico

Entamoeba histolytica, the causal agent of human amoebiasis, has two morphologically different phases: a resistant cyst and a trophozoite responsible for the invasion of the host tissues such as the colonic mucosa and the intestinal epithelium. During in vitro migration, trophozoites usually produce protuberances such as pseudopods and rarely filopodia, structures that have been observed in the interaction of trophozoites with human colonic epithelial tissue. To study the different membrane projections produced by the trophozoites, including pseudopods, filopodia, uropods, blebs, and others, we designed an induction system using erythrocyte extract or fibronectin (FN) in micropatterned grill lines (each micro-line containing multiple micro-portions of FN or erythrocyte extract) on which the trophozoites were placed in culture for migration assays. Using light, confocal, and scanning electron microscopy, we established that E. histolytica trophozoites frequently produce short and long filopodia, large retractile uropods in the rear, pseudopods, blebs, and others structures, also showing continuous migration periods. The present study provides a simple migration method to induce trophozoites to generate abundant membrane protrusion structures that are rarely obtained in normal or induced cultures, such as long filopodia; this method will allow a–better understanding of the interactions of trophozoites with FN and cell debris. E. histolytica trophozoites motility plays an important role in invasive amoebiasis. It has been proposed that both physical forces and chemical signals are involved in the trophozoite motility and migration. However, the in vivo molecules that drive the chemotactic migration remain to be determined. We propose the present assay to study host molecules that guide chemotactic behavior because the method is highly reproducible, and a live image of cell movement and migration can be quantified.

Keywords: *Entamoeba histolytica*, micropatterned grill lines, migration, fibronectin, erythrocyte extract, filopodia, pseudopods, lamellipodia

#### *Edited by:*

Anjan Debnath, University of California, San Diego, United States

#### *Reviewed by:*

Carol A. Gilchrist, University of Virginia, United States Sudip K. Ghosh, Indian Institute of Technology Kharagpur, India

> *\*Correspondence:* José L. Rosales-Encina rosales@cinvestav.mx

*Received:* 28 February 2018 *Accepted:* 02 August 2018 *Published:* 24 August 2018

#### *Citation:*

Sierra-López F, Baylón-Pacheco L, Espíritu-Gordillo P, Lagunes-Guillén A, Chávez-Munguía B and Rosales-Encina JL (2018) Influence of Micropatterned Grill Lines on Entamoeba histolytica Trophozoites Morphology and Migration. Front. Cell. Infect. Microbiol. 8:295. doi: 10.3389/fcimb.2018.00295

# INTRODUCTION

Entamoeba histolytica, the causative agent of human amoebiasis, presents in its life cycle **two** stages, trophozoite and cyst (WHO, 1997; Ali et al., 2012). Trophozoites correspond to the motile and invasive form of the parasite that usually resides in the human large intestine, and occasionally penetrate intestinal mucosa and migrate to other organs such as the liver, lungs or brain (Petri and Haque, 2013). During intestinal infection, trophozoites disrupt, and degrade intestinal mucosa barrier, and can also trogocyte mucosal epithelial cells and promote cell death by multiple cytotoxic mechanisms (Ralston et al., 2014; Begum et al., 2015). In the course of invasion, trophozoites interact with extracellular matrix components (ECM) such as fibronectin (FN), laminin and collagen (Solaymani-Mohammadi and Petri, 2008), which in vitro induce actin cytoskeleton remodeling that is involved in adhesion, migration (Sengupta et al., 2009; Javier-Reyna et al., 2012) and motility (Aguilar-Rojas et al., 2016).

Migration, associated with changes in trophozoite morphology, is stimulated by different factors such as the need for nutrients or the chemoattractant environment, which induces the characteristic amoeboid movement of trophozoites (Aguilar-Rojas et al., 2016). In two-dimensional cell culture surfaces, the migration, and motility of trophozoite usually show a fast generation of blebs and pseudopodia protrusions at the leading edge (Maugis et al., 2010). Occasionally retracting uropod at the rear end and short or large filopodia are produced by trophozoites in normal culture (González-Robles and Martínez-Palomo, 1983; Marquay Markiewicz et al., 2011), therefore these structures are generally not mentioned in the description of the mobility of the parasite (Aguilar-Rojas et al., 2016). However, filopodia of 1-6 micrometers extending between the trophozoites and MDKC or Caco-2 cell monolayers were reported (Li et al., 1994). In addition, trophozoites in contact with the mucus and epithelial cells, in an "ex-vivo human intestinal model," show short filopodia (Bansal et al., 2009).

FN-coated surfaces induce a wide variety of cellular responses in trophozoites, such as focal binding, degradation in situ of FN, remodeling of actin and myosin cytoskeleton, adhesion structures formation (Meza, 2000; Emmanuel et al., 2015), and pseudopodia and lamellipodia formation (Talamás-Lara et al., 2015). Some stress or toxic conditions of culture such as the treatment of trophozoites with α-linoleic acid causes the formation of large filopodia (about 5µm) and loss of directional motility followed by cell death of the trophozoites (Manna et al., 2013).

Lysed red blood cells (RBCs), bacteria, ECM proteins, and TNF represent chemotactic stimulus in trophozoites (Zaki et al., 2006). Chemoattractant components such as ECM proteins have been used as substrate during in vitro adhesion and migration assays, methods that enable the study of confined cell migration (Paul et al., 2017), such as microcontact-printed and micro patterns of substrate (micropatterns symmetric or asymmetric) on different surfaces to study both cell morphology and protein expression (Jiang et al., 2005; Alamdari et al., 2013; Kim et al., 2013; Paul et al., 2016). Methods for micropatterning cells culture usually require a complex and specialized equipment that is not readily accessible in most laboratories but other simple and fast methods to obtain micropatterns have been performed, such as the "ParafilmTM insertion method" (plated cells into circular or striped micropatterns) used to culture ARPE-19 and MDCK epithelial cells (Javaherian et al., 2011). Actually, micropatterns of a substrate such as continuous micropattern lines have been used to study the morphology of cancer cells during migration. These micropatterns allow an efficient formation of different large membrane protrusion, directional migration, and identification of crucial proteins related to cellular mobility (Théry, 2010; Paul et al., 2016; Tocco et al., 2018), and even facilitate the characteristic amoeboid cell migration which frequently shows pseudopodia, uropods, lamellipodia and filopodia structures in these abnormal cells (Théry, 2010; Fruleux and Hawkins, 2016; Paul et al., 2017).

Here we present a method that uses glass or plastic surfaces covered with a substrate in a "micropatterned grill line" (MPGL), which spatially stimulate E. histolytica trophozoite adhesion, migration, and an efficient formation of different membrane protrusions.

# MATERIALS AND METHODS

# Human Samples for Migration Assays

Human blood was obtained from voluntary donors to purify fibronectin from plasma or erythrocytes; these materials were used to prepare Micropatterned Grill Lines for migration assays with trophozoites from Entamoeba histolytica. The procedure for obtaining fresh blood from volunteers was carried out under the international guidelines established for the study in human populations of the Declaration of Helsinki.

# *E. histolytica* Culture

E. histolytica trophozoites, strain HM1:IMSS (ATCC 30459), clone A (Ramírez-Tapia et al., 2015), were axenically cultured at 37◦C in TYI-S-33 medium supplemented with 10% heatinactivated adult bovine serum (ABS) and harvested during logarithmic growth phase (Diamond et al., 1978). All experiments presented here were performed on at least three separate occasions and in triplicate.

# Fibronectin Purification

FN was purified by the gelatin-sepharose affinity chromatography method (Ruoslahti et al., 1982) from fresh human blood collected in 5% sodium citrate and 10 mM phenylmethylsulfonyl fluoride (PMSF). Protein purity was monitored by SDS-PAGE. Affinity purified FN was dialyzed against 0.15 M NaCl, 0.05 M Tris-HCl (pH 7.4), and stored at −70◦C. FN was quantified using the extinction coefficient of 1.28 ml/mg-cm at 280 nm and was suspended to a final concentration of 0.2 µg/µl.

# Red Blood Cells Extract Preparation

Human blood (40–80 µl) was obtained by finger prick and transferred into 1 ml of PBS. Red blood cells (RBC) were washed three times with PBS (320 x g) and diluted to a final concentration of 1000 RBC/µl in PBS. RBC (10 ml) were sonicated at 90% amplitude for 8–10 s (Ultrasonic processor GE 100), and the obtained extracts used immediately.

# Micropatterned Grill Lines (MPGLs) Preparation

FN (1–2 µl, 0.2 µg/µl) or RBCs (1,000 RBCs/µl) were placed on glass or plastic surfaces and subsequently spread over an area of about 2 × 15 mm. From this area, several thin lines were extended that were dried with a continuous air flux and sterilized under UV for 5 min. Substrate excess in the 2 × 15 mm dry fringes was removed by scraping and this area was used as a site to seed cells. For motility assay by light microscopy, pieces of Parafilm "M" (Bemis Flexible Packaging, Neenah, WI 54856) were placed on each long side of the glass slide and the coverslips. The separation (depth) between glass slides and coverslips were 0.1–0.2 mm. Plastic barriers were placed on the periphery of both glass slider sides without a coverslip. The glass slider periphery in contact with Parafilm, coverslip, and plastic was sealed with nail varnish (**Supplementary Figure 1**) (MPGLs chamber).

# Light Microscopy of Living Cells

Medium TYI-S-33 was placed in the space between the coverslip and glass slider, then 2 × 10<sup>5</sup> trophozoites in 150–200 µl of TYI-S-33 medium were gently seeded on the start site of the MPGLs (dry fringe) and incubated for 15–20 min at 37◦C. Supernatant from the "seeding area" was removed and 400 µl of TYI-S-33 medium was added to the two sites of the chamber lined with plastic to continue incubation at 37◦C for 1–5 h (**Supplementary Figure 1B**). The MPGLs chambers were incubated at different times, using as reference the 2 h of incubation by light microscopy at 35–37◦C.

# Migration Assay

For the migration assay, trophozoites plated in culture media on the MPGLs start site, as above, were incubated for 15 min at 37◦C (**Supplementary Figure 1C**). Non-adherent trophozoites were washed out and then adherent trophozoites were preincubated for 2 h in complete medium and thereafter monitored every 20 sec for 2 min by light microscopy.

# Migration Rate

The migration rate was determined after 120 min of culturing the trophozoites on the MPGLs (FN or erythrocyte extract) at 37◦C. Images were captured from videos at a rate of six frames/min. We applied ImageJ (http://imagej.nih.gov/ij/) to obtain the region of interest (ROI) with the function "adjust ellipse," and the geometric center and the front of each cell was determined. Cells that had migrated more than 3 mm from the initial front of the seeding in the MPGLs were chosen. The velocity of a mobile cell and the time of migration was determined by plotting the geometric center movement every 10 s. Non-stimulated trophozoites were evaluated as the negative control.

# Scanning Electron Microscopy (SEM)

Trophozoites (2 × 10<sup>5</sup> ) in 150–200 µl of TYI-S-33 medium were seeded on the start site of the MPGLs (dry fringe) and incubated for 15–20 min at 37◦C, then the supernatant was removed and enough TYI-S-33 medium was added to cover the entire surfaces (**Figure 2B**). Trophozoites were incubated at 37◦C in a humid chamber for 1–5 h. Trophozoites were fixed with 2.5% (v/v) glutaraldehyde in 0.1 M sodium cacodylate buffer pH 7.2 and dehydrated with increasing concentrations of ethanol. Samples were critically point dried with CO<sup>2</sup> in a Samdri-780 Tousimis apparatus. Then, they were gold coated in an ionsputter device (Jeol-JFC-1100) and examined with a Jeol JSM-7100F field emission scanning electron microscope (Chávez-Munguía and Martínez-Palomo, 2011).

# Confocal Microscopy

Trophozoites were incubated at 37◦C on the MPGLs (FN or RBCs extract as substrate) or uncoated coverslips for 2 h in TYI-S-33 medium and fixed with 4% p-formaldehyde in PBS for 45 min at 37◦C; then they were washed with PBS and permeabilized with 0.1% SDS and 0.06% Triton X-100 in PBS for 8 min at room temperature and washed with PBS. F-actin was stained with rhodamine-phalloidin (Molecular Probes, Sigma, 1:200) for 30 min at room temperature. Samples were mounted onto glass slides with VectaShield with DAPI (Vector Laboratories Inc., Burlingame, CA, USA) and observed under a Carl Zeiss LSM 700 confocal microscope.

# RESULTS

# Elaboration of MPGLs

To explore how the MPGLs can regulate E. histolytica trophozoites migration, we used RBCs extract or FN as the substrate to construct pattern arrays on glass or plastic surfaces. We fabricate topographic patterns that were composed of parallels arrays with two different volume of substrate solutions respectively (**Supplementary Figure 1**). The dry lines of the MPGLs were 20–100µm wide and 1–5 cm long, and the amount of protein was 5 µg/cm<sup>2</sup> for FN and 1,000 ± 500 µg/cm<sup>2</sup> for RBCs extract. To ensure binding on glass or plastic surfaces and formation of the micrometric substrate arrays, it is crucial to apply the substrate solution under a continuous air flux. Assay chambers were made by forming a "coverslip-glass slider" sandwich to observe the mobility of the trophozoites stimulated by the MPGLs.

# Incubation of Trophozoites on the MPGLs

Trophozoites were seeded on the MPGLs start sites and cell morphology was monitored during the cell-substrate interaction. Trophozoites induced by MPGLs showed a response that was significantly different from those non-stimulated trophozoites (**Figure 1**). Trophozoites in contact with the MPGLs generally showed two phases alternating between them at different times. The first is a phase of non-migration in which they remained on the MPGLs (but showed plasma membrane movement, membrane protrusions as large filopodia), and the second in which trophozoites migrate rapidly on MPGLs (showing the characteristic pseudopod and others membrane protrusions).

The non-stimulated trophozoites usually migrated poorly and without direction, which is clearly observed when superimposing the schemes of the cellular periphery of the representative trophozoite monitored every 20 s (**Figures 1 A,B**). Trophozoites monitored during the migration phase showed fast projections of pseudopods, pseudopods with filopodia, blebs,

FIGURE 1 | Influence of the MPGLs on E. histolytica migration. Time-lapse images (in seconds) of trophozoites incubated without (A) or with stimulation (MPGLs) (C). Overlap of the schematic periphery of one trophozoite (asterisk) at the indicated times without (B) or with stimulation (D). The arrow indicates the direction of the trophozoite migration. (Scale bar in (A,C): 20µm; Scale bar in (B,D): 10µm).

control. The groups were analyzed using one-way Anova and Bonferroni posttest. \*\*\*P < 0.0001. large rear uropod-shape body drag with filopodia behind the cell's body, tufts of filopodia, and occasionally lamellipodia.

Error bars were computed from scattering of values stacked in each column. N/S: non-stimulated (non-substrate) trophozoites were used as a negative

These trophozoites migrated persistently in the direction of the substrate (**Figures 1 C,D**). The velocity of migration of stimulated trophozoites was close to 1 µm/s (**Figure 2**) with no statistical differences between both chemoattractants, but significantly different with no-stimulated trophozoites.

Trophozoites that were monitored by light microscopy during adhesion to the MPGLs chambers generally showed short and large filopodia, blebs, and occasionally lamellipodia, which were not usually observed in non-stimulated trophozoites (**Figure 3**). Membrane projections of low thickness were difficult to observe by light microscopy (including filopodia), but better results were obtained by phase contrast using DIC microscopy. On the contrary, trophozoites that were carefully monitored during the transition from non-migration to migration phases frequently showed projecting pseudopods at the sites with the highest number of filopodia and were significantly different to normal cultured trophozoites (non-stimulated) when showing pseudopods (**Figure 3C**). To obtain long membrane protrusions, mainly large filopodia, it was necessary to maintain the temperature at 35–37◦C during the process of observation of living cells as well as during fixation, because they showed to be sensitive to temperature variations and they retracted rapidly when the temperature was lowered.

Because the actin filaments constitute the physical backbone of membrane protrusions in migrating cells, non-stimulated and stimulated trophozoites on the MPGLs were processed for the detection of F-actin. Short and long filopodia, large retractile uropods in the rear, blebs, and lamellipodia were stained for F-actin (**Figure 4**). However, short filopodia were entirely stained whereas long filopodia were partially stained **(Figure 4A**), and the actin in lamellipodia was less structured (**Figure 4B**)

trophozoites showed them when cultured on the erythrocyte extracts (**Figure 5**).

# Study of Trophozoites Incubated on MPGLs by SEM

Trophozoites cultured on the MPGLs and analyzed by SEM showed pleomorphic cell morphology because they did not migrate in a synchronized manner. During MPGLs-trophozoites

With respect to the number of trophozoites that showed the different types of protrusions on the MPGLs, no statistically significant differences were found between the different types of protrusions induced by both components of the human host, except for the long filopodia, since a smaller number of

interaction (at 120 min), the trophozoites that moved on the MPGLs showed one or more combinations of the following structures: abundant filopodia (long, short, thin, and wide), pseudopods, lamellipodia, blebs, and a rear retractile zone (uroid with filopodia). **Figure 6** shows representative images of some of the combinations of structures such as lamellipodia with filopodia (B3, B4), lamellipodia with blebs (B4), pseudopods with abundant filopodia (B1-2, C1), pseudopods with irregular filopodia (D1-D2), and structures at the retractile rear. Lowmobility trophozoites showed abundant filopodia (**Figure 6A**).

# DISCUSSION

We set out to develop a method based on a substrate forming micropatterns on surfaces of glass or plastic and on E. histolytica trophozoite stimulation to produce a great variety of structures and membrane protrusions, including those that are generally observed with very low frequency. Micropatterning allows the control of cell adhesion on a surface with a substrate and proved to be a technique useful to answer several questions of cell biology (Scarpa et al., 2013), such as expression, polarization and compartmentalization of proteins; furthermore, cells internal organization (Lee et al., 2006; Théry et al., 2006), and control of cell-cell architecture (Bornens, 2008).

analyzed using one-way Anova and Bonferroni posttest. \*\*\*P < 0.0001; \*\*P < 0.001.

ECM components have been used to coat surfaces where trophozoites are grown, for example collagen and FN (Ramírez-Tapia et al., 2015; Talamás-Lara et al., 2015), and these interaction trophozoite-substrate leads to a response as the differential expression of proteins, protease secretion (Piña-Vázquez et al., 2012), and cytoskeletal rearrangements in which actin polymerization is actively involved (Sengupta et al., 2009; Javier-Reyna et al., 2012; Starke et al., 2014; Aguilar-Rojas et al., 2016). EMC components (such as FN) have been used to generate micropatterns in order to stimulate eukaryotic cells to migrate, produce abundant and different membrane protrusions (Alamdari et al., 2013; Starke et al., 2014). Based on this, we decided to use FN in discontinuous micropatterns in line to form what we denominated micropattern grill line (MPGL). We decided to use RBCs extracts in MPGL because E. histolytica trophozoites have been also shown to respond when in contact with erythrocytes and their residues (López-Revilla and Cano-Mancera, 1982; Boettner et al., 2005; Zaki et al., 2006). In recent years it has been known that cell protrusions and migration efficiency are dependent of the substrate geometry and the dose of the substrate (Starke et al., 2014) As a result we have obtained an optimal range of stimulation in our discontinuous design of micropatterns (MPGLs).

When moving, E. histolytica trophozoites usually show fast active pseudopodium protrusions and blebs in the front of the cell (Maugis et al., 2010), and occasionally retracting uropods at the rear (González-Robles and Martínez-Palomo, 1983; Marquay Markiewicz et al., 2011). Short and large filopodia are rarely produced, but filopodia of 1-6 micrometers extending between the trophozoites and MDKC or Caco-2 cell monolayers have been seen (Li et al., 1994), and in an "ex-vivo human intestinal model" (Bansal et al., 2009). In our design of MPGLs, when trophozoites were cultured for more than 1 h, the cells generated various membrane protrusions such as pseudopodia, lamellipodia, filopodia (including some larger than 10µm), uropods and blebs. These protrusions were observed in trophozoites cultured either with FN or with RBCs as chemoattractants. Therefore, the MPGLs developed here could be a model quite useful for the study of these structures and for the search of proteins that participate in their function. In some cases the production of many trophozoite filopodia has been due to toxic lethal environmental conditions (Manna et al., 2013). In our case, the MPGLs did not cause the death of trophozoitesor alter their viability.

FIGURE 6 | Scanning electron microscopy representative of membrane protuberances that eventually the trophozoites on MPGLs showed. Trophozoites were cultured on MPGLs for 120 min. (2-3) Maximization of the respective image (1) above. (A–D) Representative trophozoites morphology. Abundant long, thin, short and motile filopodia are marked with black arrows. L: lamellipodia. R: rear retractile zone. P: pseudopodia. B: blebs. Scale bar: 10µm (1) and 1µm (2, 3). Micrographs from right to left (A–D) respectively, and from top to bottom (1–3) respectively.

When making micropatterns on glass the components of these, such as ECM components, usually drop off of the surface after a day of cultivation (Alamdari et al., 2013); thus, for cultures designed to last only a few hours it was not necessary to stabilize them additionally, as in our MPGLs, where trophozoites were cultured less than 5 h. A wide variety of methods to make substrate micropatterns have been used as a tool to induce and analyze different cell types, which results in the induction of a variety of membrane protrusion in cancer cells (Théry, 2010; Paul et al., 2017; Tocco et al., 2018). Similarly, MPGLs were effective to increase significantly, a great amount and variety of trophozoites' membrane protrusions and to increase the motility in only a few hours of induction.

On the other hand, with the developed method it was possible to measure the velocity of migration with the chemoattractants used. This velocity was around 1 µm/s, which indicates that trophozoites are highly mobile cells compared to other cells analyzed in different systems. HT1080 fibrosarcoma cells displayed a mean velocity of 0.560–0.633 µm/min when plated in absence of FN to 0.695–0.761 µm/min on FN (Barry et al., 2015); D. discoideum seeded into a Dunn chemotaxis chamber with the external channel containing cAMP showed an average velocity of 5.9 µm/min (Nenasheva et al., 2012); NIH 3T3 cells presented a mean velocity of 2.96 µm/min on FN coated polyacrylamide gel and 4.23 µm/min on Collagen IV coated polyacrylamide gel (Pushkarsky et al., 2014); T24 cancer cells exhibited a velocity of 9.6 µm/h when cultured on Collagen gels (Laforgue et al., 2015); HT1080 fibrosarcoma cells presented an average migration velocity of 13.2 µm/h when visualized in vessels in live mice (Yamauchi et al., 2005).

Trophozoites, like other migratory cells (Yamaguchi and Condeelis, 2007; Jacquemet et al., 2015; Fritz-Laylin et al., 2017), showed filamentous actin in their membrane protrusions when stimulated in the MPGL, generating mainly long and short filipodia. The long filipodia did not show completely structured F-actin, possibly due to the fact that it was in the process of integration to stabilize the tubular filamentous structure (Karlsson et al., 2013).

Numerous membrane protrusions are produced in the trophozoites due to a migratory action that is promoted by some chemoattractants. Our results indicate that the way in which the substrate is imprinted can enhance the formation of structures such as abundant short and long filopodia since many trophozoites showed them when migrating over the MPGLs. Various cell types express high levels of long transient filopodia prior spreading (Partridge and Marcantonio, 2006), which suggests a highly conserved role of filopodia in mediating initial adhesion events and in exploring environmental features. In trophozoites, filopodia could be the sensory protrusion to detect chemoattractants to initiate directed migration.

Our experiments on MPGLs have also indicated that temperature of 36–37◦C is crucial for the formation and maintenance of abundant filopodia because when the temperature was reduced to 30◦C, the filopodia retracted (data not shown). During trophozoite migration assays on the MPGLs, two types of trophozoites were detected: those of rapid migration and those of slow or without migration. The last ones were found in the place where they were seeded and at a distance no greater than 3 mm from the chemoattractant. The highly mobile trophozoites showed the largest variety of membrane protrusions.

Trophozoites' motility plays an important role during invasive amoebiasis (Aguilar-Rojas et al., 2016). It has been proposed that both, physical forces and chemical signals are involved in the trophozoites' motility and migration (Leitch et al., 1988). However, the in vivo molecules that drive the chemotactic migration remain to be determined. We propose the MPGLs assay to study host's molecules that guide their chemotactic behavior, based on two considerations: this method has been shown to be reproducible, and the live image of cell movement and migration is quantifiable.

# REFERENCES


# AUTHOR CONTRIBUTIONS

FS-L designed, performed the experiments, analyzed results and participated in writing the manuscript. LB-P participated in the MPGLs assays. PE-G participated in experiments and was in charge of the E. histolytica cultures. AL-G participated in SEM experiments. BC-M participated in SEM experiments and the writing of the manuscript. JR-E, thesis director of Ph.D. student FS-L participated in the design, analysis and discussion of results, and in the writing of the manuscript.

# ACKNOWLEDGMENTS

We thank Mr. Enrique Martinez de Luna for his technical assistance. FS-L was a recipient of a Ph.D. fellowship from CONACyT, México. This work was supported by a grant from Consejo Nacional de Ciencia y Tecnología (CONACyT), México (Grant 104119) to JR-E.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00295/full#supplementary-material

Supplementary Figure 1 | Micropatterned Grill Lines. (A) Preparation of the MPGLs with FN or RBCs extracts at low and high concentrations, and observation of fresh dry MPGLs by light microscopy (40X, scale bar in µm). (B) Culture method on the MPGLs at 37◦C by 1–5 h in TYI-S-33 medium. Cells were fixed with 4% p-formaldehyde or 2.5% (v/v) glutaraldehyde in 0.1 M sodium cacodylate buffer pH 7.2. (C) Culture of trophozoites on the MPGLs and light microscopy observation of the trophozoites' movements at 35–37◦C.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Sierra-López, Baylón-Pacheco, Espíritu-Gordillo, Lagunes-Guillén, Chávez-Munguía and Rosales-Encina. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# The Conserved ESCRT-III Machinery Participates in the Phagocytosis of *Entamoeba histolytica*

Yunuen Avalos-Padilla1,2, Roland L. Knorr <sup>1</sup> , Rosario Javier-Reyna<sup>2</sup> , Guillermina García-Rivera<sup>2</sup> , Reinhard Lipowsky <sup>1</sup> , Rumiana Dimova<sup>1</sup> and Esther Orozco<sup>2</sup> \*

<sup>1</sup> Department of Theory and Bio-Systems, Max Planck Institute of Colloids and Interfaces, Potsdam, Germany, <sup>2</sup> Departamento de Infectómica y Patogénesis Molecular, CINVESTAV IPN, Mexico City, Mexico

The endosomal sorting complex required for transport (ESCRT) orchestrates cell membrane-remodeling mechanisms in eukaryotes, including endocytosis. However, ESCRT functions in phagocytosis (ingestion of ≥250 nm particles), has been poorly studied. In macrophages and amoebae, phagocytosis is required for cell nutrition and attack to other microorganisms and cells. In Entamoeba histolytica, the voracious protozoan responsible for human amoebiasis, phagocytosis is a land mark of virulence. Here, we have investigated the role of ESCRT-III in the phagocytosis of E. histolytica, using mutant trophozoites, recombinant proteins (rEhVps20, rEhVps32, rEhVps24, and rEhVps2) and giant unilamellar vesicles (GUVs). Confocal images displayed the four proteins located around the ingested erythrocytes, in erythrocytes-containing phagosomes and in multivesicular bodies. EhVps32 and EhVps2 proteins co-localized at the phagocytic cups. Protein association increased during phagocytosis. Immunoprecipitation and flow cytometry assays substantiated these associations. GUVs revealed that the protein assembly sequence is essential to form intraluminal vesicles (ILVs). First, the active rEhVps20 bound to membranes and recruited rEhVps32, promoting membrane invaginations. rEhVps24 allowed the detachment of nascent vesicles, forming ILVs; and rEhVps2 modulated their size. The knock down of Ehvps20 and Ehvps24genes diminished the rate of erythrophagocytosis demonstrating the importance of ESCRT-III in this event. In conclusion, we present here evidence of the ESCRT-III participation in phagocytosis and delimitate the putative function of proteins, according to the in vitro reconstruction of their assembling.

### *Edited by:*

Anjan Debnath, University of California, San Diego, United States

### *Reviewed by:*

Lesly Temesvari, Clemson University, United States César López-Camarillo, Universidad Autónoma de la Ciudad de México, Mexico

> *\*Correspondence:* Esther Orozco

esther@cinvestav.mx

*Received:* 11 December 2017 *Accepted:* 12 February 2018 *Published:* 01 March 2018

#### *Citation:*

Avalos-Padilla Y, Knorr RL, Javier-Reyna R, García-Rivera G, Lipowsky R, Dimova R and Orozco E (2018) The Conserved ESCRT-III Machinery Participates in the Phagocytosis of Entamoeba histolytica. Front. Cell. Infect. Microbiol. 8:53.

doi: 10.3389/fcimb.2018.00053

Keywords: protozoan parasites, ESCRT-III proteins, *Entamoeba histolytica*, phagocytosis, GUVs model

# INTRODUCTION

In eukaryotic cells, endocytosis is the universal process to ingest nutrients and separate proteins for digestion and recycling pathways. The endosomal sorting complex required for transport (ESCRT) orchestrates several important cell membrane-remodeling mechanisms. Endocytosis involves the ingestion of small particles (≤250 nm) and of large particles (≥250 nm), including cells. In the latter case, one usually speaks of phagocytosis, a mechanism also used as an attack strategy against bacteria and other target cells. Even though the ingestion of small and large particles has many common features, phagocytosis involves a more complex molecular machinery (including contractil actin myosin ring).

One important aspect of phagocytosis is that it is a key event in the virulence mechanism of Entamoeba histolytica, the protozoan cause of human amoebiasis, responsible for killing 100,000 people each year (Mortimer and Chadee, 2010).Some of the molecules involved in the phagocytosis of E. histolytica have been already discovered. They are localized in the plasma membrane, in phagocytic cups (Arroyo and Orozco, 1987; Petri et al., 2002; Seigneur et al., 2005; Jain et al., 2008), and in the endosomes and internal membranes (Saito-Nakano et al., 2004; Loftus et al., 2005; Castellanos-Castro et al., 2016). Earlier, we identified the EhADH protein (an ALIX family member) that together with EhCP112 (a cysteine protease), forms the EhCPADH virulence complex (García-Rivera et al., 1999), which interacts with the Gal/Gal lectin at the trophozoites surface (Seigneur et al., 2005). EhADH possesses an adherence epitope at the C-terminus, which interacts with target cells (Arroyo and Orozco, 1987; García-Rivera et al., 1999) and a Bro1 domain at the N-terminus that faces the cytoplasm (Bañuelos et al., 2005). The EhADH protein binds to the EhVps32 protein (Bañuelos et al., 2012; Avalos-Padilla et al., 2015), which is the most abundant member of the ESCRT machinery.

In eukaryotes, ESCRT-0, ESCRT-I, ESCRT-II, ESCRT-III complexes and accessory proteins, among them ALIX and the Vps4 ATPase, form the ESCRT machinery (Babst et al., 2002a,b; Katzmann et al., 2002). ESCRT also participate in cytokinesis, virus budding and other cellular processes that require membrane fusion and fission (Strack et al., 2003; Morita et al., 2007; Hurley, 2015). During endocytosis, ESCRT-0 contacts the cargo and recruits ESCRT-I and ESCRT-II. Together, they trigger membrane invaginations in an opposite topology to that described in clathrin-coated cargo-carrying vesicles (Williams and Urbé, 2007).Then, ESCRT-II recruits ESCRT-III, whose proteins oligomerise on the endosomal membrane, stretching the preformed necks and generating intraluminal vesicles (ILVs) in the multivesicular bodies (MVBs). Later, Vps4 ATPase disassembles the ESCRT-III complex and the proteins return to the cytoplasm to start a new ILVs formation round (Babst et al., 1998, 2002a).

E. histolytica possesses the majority of the ESCRT machinery genes (López-Reyes et al., 2011), but their role in phagocytosis is poorly understood. To further elucidate the phagocytosis puzzle, we performed in vivo and in vitro studies using mutant trophozoites, the E. histolytica ESCRT-III purified recombinant proteins (rEhVps20, rEhVps32, rEhVps24, and rEhVps2) and giant unilamellar vesicles (GUVs). Our results demonstrated that in trophozoites, ESCRT-III proteins are present in the phagocytic cups and then, surround the ingested erythrocytes. Later, they appear in the erythrocyte-containing phagosomes and MVB-like structures. In GUVs, rEhVps20 bound to the membrane and recruited rEhVps32 which promoted the formation of quasi-spherical buds connected to the membranes by narrow necks. rEhVps24 cleaved the buds provoking the formation of ILVs; whose size was modulated by rEhVps2. The knock down of Ehvps20 and EhVps24 genes diminished the rate of phagocytosis, providing evidence for the importance of the ESCRT machinery in this process.

# RESULTS

# *E. histolytica* Possesses the Four Orthologs of the ESCRT-III Complex

Earlier, we identified and partially characterized the Ehvps2, Ehvps24, and Ehvps32 genes and transcripts, as well as the EhVps32 protein, members of the ESCRT-III complex (López-Reyes et al., 2011; Avalos-Padilla et al., 2015). However, there are no studies regarding the Ehvps20 gene and its product and we have not dissected the proteins function. Here, we first detected and cloned the Ehvps20 gene and then, produced the four ESCRT-III recombinant proteins to elucidate their putative function.

We employed the SNF7 domain sequence, present in all ESCRT-III homologs (Winter and Hauser, 2006) to search for the Ehvps20 gene in the AmoebaDB (http://amoebadb.org/amoeba/). By this domain, we found the EHI\_114790 sequence with a 621 bp open reading frame carrying a 53 bp intron, and predicting a 206 amino acids protein (**Figure 1A**). This protein hereafter referred as EhVps20, has 45% homology to the Saccharomyces cerevisiae Scvps20 gene, 36% to the Homo sapiens Chmp6 gene and 23% identity to both genes.

The four putative E. histolytica ESCRT-III genes predicted proteins of 204–246 amino acids (**Figure 1A**). As their orthologs,

FIGURE 1 | E. histolytica ESCRT-III proteins present the same structure of their orthologs. (A) Schematic representation of the four ESCRT-III proteins. Black squares: α-helixes (α1, α2, α3, α4, and α5), gray squares: k-linker (k). Numbers at right: amino acid number of each protein. (B) 3D structures of ESCRT-III proteins from yeast (Sc), human (Hs), and E. histolytica (Eh). Right panels: overlapping of 3D structures of yeast or human and E. histolytica proteins. Arrows show the two long α-helices and arrowheads the two short α-helices in the core domain.

proteins exhibited positively charged α-helices in the N-terminal domains, and negatively charged α-helices in the C-termini (Babst et al., 2002a). Secondary structures of all proteins presented the highly hydrophobic "k-linker" sequence, located between the fourth and fifth α-helices (**Figure 1A**). The k-linker is essential for protein opening and closing to generate the protein active and inactive states, respectively (Henne et al., 2012). Comparison of the predicted E. histolytica 3D structures (retrieved from the Phyre2 server) with the S. cerevisiae and H. sapiens orthologs exposed in all proteins, a core domain formed by two long and two short α-helices (**Figure 1B**, arrows and arrowheads, respectively), like the unveiled ESCRT-III homolog crystals (Muzioł et al., 2006; Bajorek et al., 2009; Martinelli et al., 2012). Judging by their 3D structures, E. histolytica ESCRT-III proteins are more similar to their human than to their yeast orthologs (**Figure 1B**). The low homology in sequence and structure of ESCRT-III orthologs (**Table 1**), together with the high homology in key domains, point out to the relevance of the SNF7 domain, α-helices and k-linkers in the protein function.

# ESCRT-III Proteins Participate in the Erythrophagocytosis of E. histolytica

In yeast and mammals, the inactive forms of ESCRT-III proteins are soluble in the cytoplasm; whereas in their active state, they bind to the endosomal membrane to promote the continuity of endocytosis until cargo digestion and protein recycling occur (Shim et al., 2007). Given its participation in receptor-mediated endocytosis, the ESCRT machinery is candidate to have a central role in phagocytosis. Nevertheless, molecular mechanisms in both events could differ, mainly due to the size and complexity of the ingested cargo.

To study the role of the ESCRT-III in phagocytosis, we cloned the genes, expressed their products and generate specific antibodies against each protein. Coomassie blue stained gels confirmed the integrity of the proteins by their predicted molecular weight (**Figure 2A**). By western blot assays, the specific antibodies recognized the rEhVps20 and rEhVps24 recombinant proteins and a single band in trophozoite lysates (**Figure 2B**).Similarly to the EhVps32 protein, EhVps20 and EhVps24 consistently exhibited higher molecular weight (39 and 31 kDa, respectively) than the predicted ones by the amino acid sequence (24 kDa for both) (**Figure 2B**) (Avalos-Padilla et al., 2015). However, bacterially expressed EhVps20 and EhVps24 also migrated in an identical manner. As in their orthologs, the highly charged nature of the proteins could influence their migration (Babst et al., 2002a). Moreover, when we removed the last alpha helix from EhVps20 and EhVps32, where the acidic charges are concentrated, the recombinant proteins migrated at the expected molecular weight. This supports our hypothesis for the observed discrepancy of the molecular weights. In contrast, rEhVps2 migrated at its predicted molecular weight (28 kDa), but in trophozoite lysates, α-rEhVps2 antibody revealed 28 and 56 kDa bands. This last could be a dimer or a complex formed by EhVps2 with another unidentified protein (**Figure 2B**) but more experiments are need to prove this. Pre-immune serum did not react with any protein from bacterial or trophozoites lysates.

To localize the proteins in the cell, we performed confocal microscopy assays of trophozoites in basal conditions (0 time, no phagocytosis) and after phagocytosis. FITC-labeled secondary antibodies detected each one of the specific antibodies and proteins. In basal conditions, EhVps20, EhVps32, EhVps24, and EhVps2 appeared in punctuated structures dispersed in the cytoplasm. EhVps20, EhVps32, and EhVps24 localized adjacent to plasma membrane (**Figure 2C**). As the antibodies gave no

TABLE 1 | Percentage of identity and similarity and predicted functions of ESCRT-III proteins.


The e-value and the percentages of identity and similarity were obtained by comparing the sequences of E. histolytica ESCRT-III putative proteins with their orthologs in Human and Yeast.

signal in non-permeabilized trophozoites, we concluded that like EhVps32 (Avalos-Padilla et al., 2015), EhVps20 and EhVps24 were close to the inner leaflet of the plasma membrane.

After 15–30 min of phagocytosis, all four proteins appeared spread in the cytoplasm and in the membranes and lumen of phagosomes and surrounding the ingested erythrocytes (**Figure 2C**). Intriguingly, EhVps24 and EhVps2, consistently migrated to the nuclei at longer times of phagocytosis (60 min), co-localizing with the nuclear protein pCNA (Trasviña-Arenas et al., 2017) and DAPI staining (**Figures 2C,E**). Altogether, these results demonstrated that the ESCRT-III orthologs associate to erythrocytes and erythrocyte-containing phagosomes, suggesting that ESCRT-III proteins are a part of the machinery involved in phagocytosis. Yet, we have no data to explain the presence of ESCRT-III proteins in the nucleus. We speculate that they could be regulating or co-regulating some unidentified nuclear function; but further experiments are necessary to prove this.

α-pCNA protein (red) antibodies after 60 min of erythrophagocytosis. Nuclei were stained by DAPI. (F) Open field of trophozoites stained by αEhVps24 or α-rEhVps2

and α-pCNA with corresponding secondary antibodies after 60 min of erythrophagocytosis. phc, phase contrast. Scale bar: 10µm.

# ESCRT-III Proteins Interact With Each Other during Erythrophagocytosis

We explored the possible association among the ESCRT-III proteins in basal conditions and during erythrophagocytosis by immunofluorescence, immunoprecipitation, flow cytometry and in vitro assays using GUVs.

In basal conditions, confocal images showed all ESCRT-III proteins dispersed in the cytoplasm as clumps, or inside vesicles of different size. Proteins were grouped, co-localizing in distinct combinations (**Figure 3A**, see magnification square of merging image). The observed vesicles with more than one protein may be due to the active constitutive endocytosis in trophozoites.

After 10 min of phagocytosis, EhVps20 appeared around the ingested erythrocytes and dispersed in the cytoplasm (**Figure 3B**, cyan). A part of EhVps32 remained in the plasma membrane and the cytoplasm, but another part also migrated to the area close to the protrusion forming the phagocytic cup and to the membrane of vesicles with ingested erythrocytes (**Figure 3B**, green). EhVps24 appeared in the vesicles with ingested erythrocytes (**Figure 3**, blue), co-localizing with other proteins, but frequently, it was alone in a vesicle (**Figure 3B**, red arrow). EhVps2 strongly marked the phagocytic cup and appeared in the cytoplasm around the ingested erythrocytes (**Figure 3B**, red). Surprisingly, in many trophozoites, EhVps32 and EhVps2 co-localized at the phagocytic cup to a higher extent than in phagosomes (**Figure 3B**, see magnification square, yellow).

Subsequently, we explored the localization of the ESCRT-III proteins in phagosomes after 90 min of erythrophagocytosis. At this time, hemoglobin is in an advanced digestion process, but is still possible to distinguish the remnant erythrocytes. For these experiments, we used couples of antibodies for the same preparation:α-EhVps20/α-EhVps32 (sc1), α-EhVps32/α-EhVps24, and α-EhVps24/α-EhVps2 (sc2). The accumulation of erythrocytes inside the trophozoites, produced phagosomes of distinct shapes and size, depending on the number of erythrocytes they carried and on the ingestion time. Confocal images showed that EhVps20 covered the phagosomes with a uniform punctuate pattern (**Figure 3C**, see magnification in 3D) and it co-localized in some points with EhVps32. EhVps32 defined a honeycomb-like arrangement around the erythrocytes inside the huge phagosomes (**Figure 3C**, see magnification in 3D upper, right panel). EhVps24 and EhVps2 appeared diffuse in the cytoplasm and around the phagosomes, co-localizing with each other and surrounding the erythrocytes. In many images, EhVps2 appeared closer to the erythrocytes (**Figure 3C**, see magnification

(green) and α-rEhVps2 (red) primary antibodies, respectively; Alexa647 (cyan) and Pacific Blue (blue) directly labeled the α-rEhVps20 and α-rEhVps24 antibodies, respectively. (A) Basal conditions. At right: magnification of white square. (B) Trophozoites at 10 min of erythrophagocytosis. At right: magnification of white square. Pcup: phagocytic cup, e: erythrocytes, red arrow: vesicle stained only by α-rEhVps24 antibody. (C) Confocal representative image of a trophozoite after 90 min of phagocytosis. The areas are marked in squares are magnified in (D): upper panels: erythrocytes covered by EhVps20 and surrounded by EhVps32 and EhVps32 covering two erythrocytes (e) (arrowhead in C) in contrast to others that are surrounded by the protein. Lower panel left: EhVPs32 forming the honeycomb like panel around the erythrocytes together with EhVps24. Lower panel right: EhVps2 and EhVps24 together surrounding the partially digested erythroctyes. phc, phase contrast; sc1, subcomplex 1; sc2, subcomplex 2. Scale bar: 5µm.

in 3D, lower, right panel). These results exhibited the four ESCRT-III proteins in the erythrocyte-containing phagosomes, on the ingested erythrocytes and around them, but proteins displayed different patterns which varies according to the time of the erythrocytes ingestion.

Interestingly, in a single trophozoite, MVBs frequently appeared with distinct morphology, suggesting that these structures were in different maturation stage (**Figure 4**, red square). The four ESCRT-III proteins also appeared surrounding MVBs-like structures, but they were observed also inside them and in the nascent ILVs. The most abundant protein in these structures was EhVps32 (**Figure 4**, green). Magnification in **Figure 4** illustrates a representative image with at least three MVB-like structures (up to 10µm of diameter) in a different maturation process (**Figure 4**, white arrows). Frequently, in these structures, EhVps32 appeared around nascent vesicles (ILVs) inside MVB-like structures. EhVps2, EhVps20 and EhVps24 were also observed inside or surrounding these putative nascent vesicles (**Figure 4**).

## The ESCRT-III Proteins Associate during Phagocytosis

We further discern whether the ESCRT-III proteins only colocalized, as confocal images shown, or they associate during phagocytosis. For these experiments we used the α-rEhVps32 antibody to immunoprecipitate lysates from trophozoites in basal conditions and after phagocytosis. Western blot assays of the immunoprecipitates evidenced that the four ESCRT-III proteins associate in basal conditions and during erythrophagocytosis (**Figure 5A**). However, we did not detect significant differences in the protein amount before and after phagocytosis.

To quantify the protein association, we measured by flow cytometry the interactions between two proteins at the same time, using the sequence described above: (EhVps20/EhVps32, EhVps32/EhVps24, EhVps24/ EhVps2, and EhVps2/EhVps20). Results showed a low percentage of association at basal conditions, slightly higher for EhVps20/EhVps32 (8%) than for the other three couples (2–4%) (**Figure 5B**). All couples increased their association during phagocytosis. After 30 min of erythrophagocytosis, EhVps20/EhVps32 and EhVps32/EhVps24 reached about 47% interaction, whereas EhVps24/EhVps2 and EhVps2/EhVps20 presented 10 and 13% respectively (**Figure 5B**). Intriguingly, EhVps2/EhVps20 interaction increased from 2% at basal conditions to 24% at 5 min and decreased to about 12% after 30 min of phagocytosis (**Figure 5B**). These results suggested that EhVps20/EhVps32 as well as EhVps32/EhVps24 association is stronger than the ones in which EhVps2 is involved. We are aware that protein association is a dynamic process, according to the biological conditions of the trophozoites. Thus, from these experiments, it is not possible to deduce the precise order of protein assembly.

# The EhVps20 (Active Form) Binds More Efficiently to Artificial Membranes Than Other ESCRT-III Proteins

To resolve the sequence in which proteins bind to the membrane and monitor the associated changes in the membrane morphology after their binding, we employed GUVs as a model system (Wollert et al., 2009). The advantage of GUVs over other membrane models is that the membrane response can be directly observed under the microscope. We mimicked the endosomal membrane composition (Evans and Hardison, 1985; Kobayashi et al., 2003) (POPC:POPS:chol:PI(3)P in a 62:10:25:3 ratio) and also tested other lipid mixtures (only DOPC; DOPC:PI(3)P 95:5; DOPC:DOPS 90:10 or DOPC:DOPG 90:10) (for abbreviations see Materials and Methods section); and probed the binding

FIGURE 4 | ESCRT-III proteins are present in MVBs like structures of trophozoites. Representative confocal microscopy image of a trophozoite exhibiting MBVs like structures at distinct maturation phases, MVBs like structures are signaled by arrows in the α-rEhVps32 image. Red square of the phase contrast (phc) image is magnified below the picture. The three MVBs like structuresare in white squares in merging images and marked with 1, 2, and 3 numbers that correspond to the three images at right. Arrows in merging images signal the presence of EhVps32 in the putative MVBs structures. e, erythrocyte; phc, phase contrast. Scale bar: 10µm.

and membrane reshaping produced by E. histolytica ESCRT-III recombinant proteins. The GUVs were labeled with the fluorescent lipid analog TexasRed-DHPE to visualize membrane deformations.

Numerous studies have shown that ESCRT-III proteins share a common domain structure of four α-helices packed into an N-terminal core domain and a C-terminal auto-inhibitory region that blocks the main membrane contact sites present in α1 and α2. Truncation of the auto-inhibitory C-terminal region (corresponding to the last α-helix including the k-linker) produces proteins with exposed membrane binding domains (Zamborlini et al., 2006; Shim et al., 2007; Henne et al., 2012).

To avoid the use of other factors required for activation, we generated truncated recombinant protein versions of rEhVps20 (1-173 amino acids) and rEhVps32 (1-165 amino acids) (**Figure 2A**, lanes 3 and 6) to have them in an open conformation. As expected, fluorescently-labeled rEhVps20 and rEhVps32 full proteins presented low binding efficiency to GUVs with poor Pearson coefficients (PC) (**Figures 6A,B**). In contrast, the active form of fluorescentlylabeled rEhVps20 (1-173) exhibited 8.5 times higher binding efficiency to GUVs than the inactive protein (PC = 0.456 and 0.0534, respectively) (**Figures 6A,B**). The binding efficiency observed with rEhVps32 (1-165) was 3.3 times greater than the triggered by the full protein (PC = 0.1382 and 0.041, respectively) (**Figures 6A,B**). Interestingly, rEhVps20(1-173) presented 3.26 times greater binding efficiency to GUVs than rEhVps32(1-165) (**Figures 6A,B**). These experiments showed that only EhVps20(1-173) and EhVps32(1-165) active forms interacted with membranes, but with distinct affinity. Because of the higher binding efficiency exhibited by rEhVps20(1-173), we employed the truncated version of rEhVps20 for the rest of the experiments. It is important to point out that in vitro and in vivo assays give accuracy to the results obtained, however, results could slightly differ, mainly due to the distinct complexity of the endocytosis/phagocytosis processes and to the differences in membrane composition.

# Reconstruction of the *E. histolytica* ESCRT-III Machinery in GUVs

To rebuild the whole ESCRT-III machinery and establish the binding order of the proteins, we added in distinct sequence each of the ESCRT-III recombinant proteins to TexasRed-DHPE-labeled GUVs, giving 5 min intervals between the additions of each protein. As previously observed, confocal images showed that Alexa488 labeled rEhVps20(1-173) binds to GUV membranes (**Figure 7A**). After adding rEhVps32, we detected small fluorescent quasi-spherical buds attached to the rEhVps20-labeled membranes, suggesting invaginations induced by rEhVps32. After 10 min, the buds remained attached to GUV membranes via narrow membrane necks, but when we incorporated the rEhVps24 protein, small ILVs (∼0.5µm) were generated inside the GUVs (**Figures 7A,B**). It is quite remarkable that all quasi-spherical buds induced by the binding of rEhVps20 and rEhVps32 were similar in size. Such a uniform bud size indicates that the adsorbed protein layer undergoes phase separation into a rEhVps32-rich and a rEhVps32-poor phase and that the rEhVps32-rich domains have a significant spontaneous curvature that determines the bud size (Lipowsky, 1992).

The results presented so far could imply that EhVps2 is not necessary for ILVs generation. However, upon incubation of GUVs with EhVps20(1-173) and EhVps32 followed by the addition of rEhVps24 and rEhVps2 together, many larger ILVs appeared (∼3µm of diameter) (**Figure 7B**). Additionally, the percentage of GUVs with ILVs increased from 64 to 71% by EhVps2 presence (**Figure 7D**). These results strongly suggest that in E. histolytica, the core formed by active EhVps20, EhVps32, and EhVps24 is sufficient to generate ILVs, while EhVps2 acts as an additional regulator of ILVs size. The substitution of EhVps20(1-173) by EhVps32(1-165) did not produce ILVs, revealing that the function of EhVps20 and EhVps32 is not redundant (**Figure 7C**). Experiments changing the sequence of

the proteins (**Figure S1**) indicated that the order of protein addition to produce ILVs must be the following: active EhVps20, EhVps32, EhVps24, and EhVps2.

To prove the specificity of ESCRT-III binding toward negatively charged membranes, we performed reconstruction assays using GUVs entirely composed of DOPC (zwitterionic) or a combination of the same negative molar ratio (equivalent to 10% PS) of different negatively charged lipids: DOPS, PI(3)P and DOPG. In eukaryotes, PS and PI(3)P are present on early endosomes and the internal leaflet of the plasma membrane (Gillooly et al., 2000) while PG is present in non-endosomalrelated membranes such as mitochondrial. As expected, ILVs appeared only in negatively charged membranes (**Figure 7E**). We also found that rEhVps20(1-173) formed protein clusters on the surface of membranes with high negative charge (data not shown). The greatest percentage of ILV formation was achieved using GUVs with DOPS and PI(3)P (∼30% in both cases) (**Figure 7E**). These results suggested that even when ESCRT-III proteins are able to bind to a wide range of different negativelycharged lipids, there is an apparent affinity toward lipids present in endosomal membranes during the transition from early to late endosomes (Lemmon and Traub, 2000).

# Knock Down of *Ehvps20* and *Ehvps24* Genes Affects the Rate of Phagocytosis in Trophozoites

To get more evidence on the role of ESCRT-III complex in phagocytosis of trophozoites, first, we knocked down the Ehvps20 gene, whose product is the first protein that binds to membranes and triggers the assembly of the rest of ESCRT-III members. We employed trophozoites of the G3 strain to transcriptionally silence the Ehvps20 gene (Mirelman et al., 2006). Western blot assays revealed 60% protein reduction in EhVps20−silenced trophozoites compared with those transfected with the empty vector (**Figures 8A,B**). The expression of other ESCRT-III members, was significantly affected (**Figure 8B**). Immunofluorescence assays confirmed the reduction of the EhVps20 amount in Ehvps20<sup>−</sup> trophozoites (**Figures 8C,D**). These trophozoites, exhibited a reduced capacity to ingest erythrocytes (65–70% at 5 min and 50–65% at 15 min), compared with the rate of ingestion of G3 trophozoites transfected with the empty vector (**Figures 8E,F**).

The knocked down of Ehvps24 gene showed 25% remnant expression of the protein, using as control the actin protein (**Figures 9A,B**). Silencing of this gene did not affect the expression of the EhVps32 and the EhVps24 proteins (**Figures 9A,B**). Immunofluorescence assays confirmed the 75% reduction of fluorescence in Ehvps24<sup>−</sup> knocked down trophozoites, in comparison with trophozoites transfected only with the vector (**Figures 9C,D**). On the other hand, the rate of erythrophagocytosis of Ehvps24−knocked down trophozoites appeared affected in about 60% at 5 min and 66% at 15 min (**Figures 9E,F**). These results, together with the ones previously obtained with EhVps32 knocked down trophozoites (Avalos-Padilla et al., 2015) strengthened the hypothesis that ESCRT-III in E. histolytica has prior influence in the rate of phagocytosis in the trophozoites. Each one of its member carry out defined functions that are necessary to efficiently ingest and process target cells.

# DISCUSSION

The ancient ESCRT machinery is critical for several central cellular processes; among them, endocytosis, MVBs biogenesis, viral budding, cytokinesis, autophagy, exosomes secretion and others based on membrane deformation and scission (Carlton and Martin-Serrano, 2007; Filimonenko et al., 2007; Lee et al., 2007; Rusten et al., 2007). However, its role in phagocytosis, an event in which membrane deformation and scission is evident, has not been completely explored. Besides, many of the detailed molecular mechanisms to produce these events are unknown.

But it is well documented that process involving membrane remodeling are linked to the assembly-disassembly cycles of ESCRT-III proteins in membranes (Hurley, 2015). Thus, it is expected that ESCRT machinery has a relevant role in the phagocytosis of the E. histolytica trophozoites.

In E. histolytica, the precise function for the majority of the ESCRT products is unknown, even when almost all ESCRT genes are present in the genome (López-Reyes et al., 2011). EhVps32, the ortholog of Snf7 in yeast and CHMP4 in human, as well as the accessory proteins EhADH (an ALIX family protein) and the EhVps4 ATPase are involved in the phagocytosis of trophozoites (García-Rivera et al., 1999; López-Reyes et al., 2010; Avalos-Padilla et al., 2015). Additionally, EhVps32 and EhADH are also involved in pinocytosis of dextran and phagocytosis of latex-coated beads (Avalos-Padilla et al., 2015; Castellanos-Castro et al., 2016) which implies that ESCRT-III proteins participate in different types of endocytosis. Even when substantial work has been done in the study of the molecules involved in phagocytosis in E. histolytica (Petri et al., 2002; Saito-Nakano et al., 2004; Seigneur et al., 2005; Jain et al., 2008), many others, as well as the mechanisms participating from the interaction with cargo molecules to the digestion and recycling of proteins, are not well known yet. Here, we expanded our research on E. histolytica phagocytosis, using in silico analysis, mutant trophozoites, ESCRT-III recombinant proteins and the GUV model to unravel a part of the phagocytosis puzzle. Our results give evidence that the ESCRT-III proteins participate in the formation of phagocytic cups, phagosome maturation and MVB-like structure generation in this primeval voracious parasite.

In silico analysis showed that E. histolytica possesses all four ESCRT-III proteins with a 3D structure similar to the yeast and human orthologs. Although their amino acid sequences exhibit low homology (less than 32% in all cases, **Table 1**), the four proteins have all functional domains described in

their orthologs (**Figures 1A,B**), remarking the importance of these domains in distinct cellular events and their conserved function in eukaryotes. The high homology in the secondary protein structure together with their function performed during phagocytosis, reconstructed in GUVs, confirmed that they are bona fide E. histolytica orthologs of ESCRT-III proteins.

Immunofluorescence assays, using specific antibodies against each of the four ESCRT-III proteins evidenced that all they make contact with the erythrocytes since early times of erythrophagocytosis (**Figures 2C**, **3B**). Even in basal conditions, images revealed proteins co-localizing in all possible combinations, due to the active constitutive endocytosis of trophozoites (**Figure 3A**). Moreover, immunoprecipitation and flow cytometry experiments using α-rEhVps32 antibody revealed that, in basal conditions and during phagocytosis, the proteins associate with each other and association is enhanced during phagocytosis (**Figure 5**). Although we cannot discern whether it is a direct or indirect association, results strongly support their involvement in this event and evidence the dynamic of the protein association.

During the early stages of phagocytosis (10 min), EhVps32 and EhVps2 consistently appeared in the phagocytic cups and in the plasma membrane close to the contact with the erythrocytes (**Figure 10**). This location suggests a novel function for these two proteins, not described in other eukaryotes, during the capture of the cargo, in addition to their known role in the ESCRT-III-complex formation on endosomes. At the first contact of trophozoites with erythrocytes, EhVps32 co-localize with Gal/Gal lectin and EhADH, which can act as receptors for erythrocytes in the early phagosome maturation stages (Avalos-Padilla et al., 2015) (**Figure 10**). Thus, the assembling of the phagocytosis puzzle in E. histolytica has advanced with the results shown here.

Later in phagocytosis, when trophozoites have ingested a larger number of erythrocytes, EhVps20 covered the erythrocytes inside phagosomes, and EhVps32 together with EhVps24 surrounded them. EhVps2 also appeared around the phagosomes (**Figure 4**). Intriguingly, EhVps24 and EhVps2 consistently migrate to the nucleus during phagocytosis, when digestion of the prey is advanced (**Figure 2D**). The nuclear localization of these proteins was confirmed by co-localization with the nuclear protein pCNA (**Figure 2E**) So far, we have no experimental data to speculate about the nuclear function of these proteins, although ESCRT-III proteins participate in nuclear envelope repair in human cells (Raab et al., 2016; Ventimiglia and Martin-Serrano, 2016). As migration happens in more than 95% of the trophozoites, we conjectured that they could act as transcription factors or co-factors or to repair the nuclear envelope as in humans. Moreover, phagocytosis is an exhausting experience for trophozoites; it produces stress, and involves and consumes many cellular proteins. EhVps24 and EhVps2 could contribute by a yet unknown mechanism to recover the basal state of trophozoites

after long times of phagocytosis. This could involve transcripts synthesis to replace the proteins consumed during the process. However, further experiments will help to better understand this.

The ESCRT-III proteins in trophozoites in basal conditions and during phagocytosis showed a highly dynamic movement in the cell, accordingly to the synthesis and exchange of membranes in this insatiable parasite. Confocal images revealed the detection of ESCRT-III proteins in distinct forms of phagosomes and MVBs-like structures, which might correspond to different phases of phagosome maturation (**Figure 4**), as it is shown in **Figure 10**. In other systems, ESCRT-III proteins start ILVs formation at an intermediate stage, where the phagosome is no longer an early compartment but has not yet the characteristic features of a late phagosome (Flannagan et al., 2012). It is possible to hypothesize that this also occurs in trophozoites. At 60 and 90 min, when digestion is advanced, we detected ESCRT-III proteins in MVB-like structures (up to 10µm of diameter) with partially digested erythrocytes and a distinct number of ILVs and phagosomes with different shapes. In other eukaryotes, MVBs are smaller (400–500 nm of diameter); however, they are mainly formed during receptor-mediated endocytosis of molecules and small particles, whereas the erythrocytes are much larger than the ingested molecules during these types of endocytosis. Besides, phagosomes can content between one to twenty or more erythrocytes. Interestingly, in these MVBs-like structures, we consistently observed EhVps32 clusters in the membrane (**Figure 4**), which might correspond to the nascent vesicles observed in GUVs after contact with active rEhVps20 and rEhVps32 (**Figure 7**). We also detected the presence of EhVps24 and EhVps2 in these invaginations suggesting that the four proteins participate in ILVs formation in trophozoites.

Phagocytosis is a process that occurs in a series of chained events and a large number of proteins participate in it. This makes difficult to assign by in vivo experiments, a particular function to each protein. The GUV model and the recombinant proteins allowed us to precisely assign a function for each one of the E. histolytica ESCRT-III proteins during ILVs formation and to dissect the order in which proteins are assembled. We are aware that, as in other systems, ESCRT-III proteins could perform other functions, unidentified here. The fact that the recombinant proteins bind differentially to GUV membranes shows the adequacy of the model for this type of studies.

By the GUV model we confirmed that active rEhVps20(1- 173) efficiently binds only to negatively charged membranes, and exhibit no specific preference toward the charged lipid species which can explain its presence in different membranes. Then,

FIGURE 10 | Model for the participation of ESCRT-III in phagocytosis. (0 min) Attachment to human erythrocytes to trophozoites involves plasma membrane proteins including Gal/GalNAc lectin, TMK96 and EhADH. In particular, EhADH possesses a Bro1 domain in the cytoplasmic tail that recruits EhVps32 upon binding of erythrocytes to trophozoites. In this first phase of contact, it is also possible to observe EhVps20 in close contact with the plasma membrane. (2–10 min) The union of erythrocytes triggers signaling mechanisms that modulate actin cytoskeleton that together with other proteins remodels and produces the phagocytic cup. During this stage, EhADH-EhVps32 complex and EhVps20 independently recruit the other ESCRT-III molecules to the phagocytic cup and surround the nascent phagosome forming hetero-polymers that will finally lead to the scission and internalization of the phagosome. (30–90 min) After longer times, huge phagosomes with different number of erythrocytes inside them appear as a product of the fusion of several phagosomes. In such phagosomes, the formation of ILVs is triggered by the action of ESCRT-III polymers and the proteins are present either surrounding the phagosome, or inside them in close contact with the internalized erythrocytes. After 60 min of phagocytosis, EhVps2 and EhVps24 consistently migrated to the nucleus. E, erythrocyte.

rEhVps32 binds to rEhVps20 and produces quasi-spherical membrane buds connected to the GUVs by narrow membrane necks. The uniform size of these buds indicates that the adsorbed protein layer undergoes phase separation and forms rEhVps32 rich domains with a significant spontaneous curvature that determines the bud size (Lipowsky, 1992). The two proteins and the sequence of their binding are necessary to initiate ILV generation, but rEhVps20 and rEhVps32 do not cleave the neck to release ILVs. As in their orthologs, EhVps32 and EhVps24 are able to polymerize, and polymers could nucleate in the areas where the membrane is deformed. Then, the polymer growth can provoke the closure of the formed neck and split the nascent vesicle producing ILVs.

Our experiments also evidenced that EhVps24 in combination with rEhVps2 modulate the size of the ILVs. It is interesting that EhVps2 appears to increase the size of the generated ILV, this function is especially important in the case of phagocytosis, where the diameter of ILVs within phagosomes should be larger due to the proportion of the ingested cells. We do not know to what extent EhVps2 modulates size and EhVps24 cleaves the nascent vesicles in trophozoites, but we can speculate that they have analogous functions as revealed in GUV experiments.

Finally, the knock down of Ehvps20 and Ehvps24, presented in this paper, and Ehvps32, published earlier (Avalos-Padilla et al., 2015), genes in trophozoites confirmed the functional importance of the ESCRT-III complex in phagocytosis, since EhVps20 is necessarily the first protein that binds to membranes to recruits the others. Its silencing reduced the rate of phagocytosis in transfected trophozoites (**Figure 8**), evidencing its importance in this step of the phagocytosis. Interestingly, the knock-down of the Ehvps20 gene, affected the expression of the EhVps32 and EhVps24 proteins. Since EhVps20 is the first protein that recruits the rest of the ESCRT-III subunits, its absence can affect the expression of the rest of the ESCRT-III members. In the same way, the knocked down of EhVps24, an important player in ILVs formation, affected the rate of phagocytosis, but had no effect in the expression of the other ESCRT-III members. The remnant phagocytic activity in transfected trophozoites can be due to the action of the remnant protein and to other proteins involved in phagocytosis, among them the Rab family proteins, which act as regulators or carriers of other molecules (Rodríguez and Orozco, 2000; Saito-Nakano et al., 2004). Moreover, in previous experiments, the silencing of EhVps32 impaired the rate of phagocytosis in 80% in trophozoites (Avalos-Padilla et al., 2015), these differences could be due to the distinct participation of the proteins as seen in the in vitro studies in this paper. Considering that EhVps32 also act during capture of cargo and together with EhADH (Avalos-Padilla et al., 2015), it can participate in an alternative pathway of ESCRT-III proteins recruitment. Since phagocytosis is a complex process, the GUVs model is suitable to analyze the specific function of each of the proteins that participate in phagocytosis/endocytosis and their relationship to other molecules already studied.

In conclusion, we have been able to reconstruct the sequential assembling of ESCRT-III and give evidences for its participation in the phagocytosis of E. histolytica, as it is shown in the model in **Figure 10**. Additionally, we demonstrate here that the combination of in vivo and in vitro studies using GUVs is a good strategy to analyze with high detail the role of these proteins in membrane deforming. Also this is a good model to study virulence functions in parasites.

# MATERIALS AND METHODS

# *E. histolytica* Cultures

Trophozoites of E. histolytica (strain HM1:IMSS) clones A (Orozco et al., 1983) and G3 (Mirelman et al., 2008) were axenically cultured in TYI-S-33 medium at 37◦C and harvested in logarithmic growth phase (Diamond et al., 1978) to perform the experiments. Silenced G3 trophozoites were analyzed 24 h after transfection (Mirelman et al., 2008). Experiments were performed at least three times by duplicate for statistical analysis.

# *In Silico* Analysis of EhVps20

The SNF7 protein domain was searched in the E. histolytica genome database (www.amoebadb.org). Comparison between putative EhVps20 protein and human and yeast orthologs was determined using the Expert Protein Analysis System (ExPASy) Proteomics Server by the NCBI BLAST service program. Sequence alignments were generated using the ClustalW2 program (Blackshields et al., 2006).

# Tertiary (3D) Protein Modeling

EhVps2, EhVps20, and EhVps24 amino acid sequences were submitted to the Phyre2 Server (http://www.sbg.bio.ic.ac.uk/ phyre2/html/page.cgi?id=index) to obtain the proteins 3D predicted structures. Results obtained were documented and analyzed through the UCSF Chimera software (Pettersen et al., 2004).

# Cloning and Expression of Recombinant Proteins

The full-length Ehvps2, Ehvps20 and Ehvps24 genes, the first 173 residues from Ehvps20 and the first 165 residues from Ehvps32 were PCR-amplified using cDNA as template and specific primers (**Table 1**) which introduced unique BamHI and SalI sites in the sense and antisense primers, respectively (underlined in **Table 2**). Genes were cloned into the pJET1.2/blunt plasmid (Thermo Fisher, Waltham, MA, USA), accordingly to manufacturer's instructions. Then, genes were subcloned into the BamHI and SalI sites of the pGEX-6P-1 plasmid, generating pGEX6P-Ehvps2, pGEX6P-Ehvps20, pGEX6P-Ehvps20(1-173), pGEX6P-Ehvps24 and pGEX6P-Ehvps32(1-165) constructs. Escherichia coli pLys-S bacteria were transformed with the plasmids, and recombinant proteins were induced by 0.1 mM of IPTG to produce GST-rEhVps2, GST-rEhVps20, GST-rEhVps20(1- 173), GST-rEhVps24 and GST-rEhVps32(1-165) tagged proteins. GST-tagged proteins were dialyzed against the buffer for the PreScission protease enzyme (GE-healthcare, Freiburg, Germany) and the GST-tags were removed according to manufacturer's instructions. GST-free proteins were purified by size exclusion chromatography.

# Generation of Polyclonal Antibodies

rEhVps2 (60 µg) emulsified in Titer-Max Classic adjuvant (1:1) (Sigma, St. Louis, MO, USA) was subcutaneously and intramuscularly inoculated into Wistar rats. Two more doses of rEhVps2 (30 µg) were injected at 20 days intervals and then, animals were bled to obtain α-rEhVps2 antibodies. rEhVps20 and rEhVps24 proteins were immunized in New Zealand male rabbits following the protocol previously described using 100 µg of protein for the first dose and 50 µg for subsequent doses. α-rEhVps32 was previously generated in mice (Avalos-Padilla et al., 2015). Pre-immune serum was obtained before immunization in all cases.

# Western Blot Experiments

Trophozoites lysates (30 µg) were separated in 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), transferred to nitrocellulose membranes and probed with mouse α-rEhVps32 (1:15,000) (Avalos-Padilla et al., 2015), rat αrEhVps2 (1:15,000), rabbit α-rEhVps20 (1:20,000) or rabbit α-rEhVps24 (1:10,000) antibodies. Membranes were washed, incubated with the corresponding α-mouse, α-rat or α-rabbit HRP-labeled secondary antibodies (Zymed; 1:10,000), and revealed with ECL Prime western blotting detection reagent (GE-Healthcare).

# Laser Confocal Microscopy Assays

Trophozoites were grown on coverslips, fixed with 4% paraformaldehyde (PFA) at 37◦C for 1h, permeabilized with 0.2% Triton X-100 and blocked with 10% fetal bovine serum (FBS) in PBS. Then, cells were incubated with either mouse α-rEhVps32 (1:200), rat α-rEhVps2 (1:200), rabbit α-rEhVps20 (1:200) or rabbit <sup>α</sup>-rEhVps24 (1:200) antibodies at 37◦C for 1h, followed by incubation for 1h with α-rat TRITC-labeled, α-mouse or α-rabbit FITC-labeled secondary antibodies (Zymed-Thermo Fisher; 1:100) as appropriate. For multi-labeling experiments, secondary antibodies were used to detect EhVps32 and EhVps2 as above. For rEhVps20 and rEhVps24 detection, no secondary antibodies were used, instead, the α-rEhVps20 antibody was labeled with Alexa647 fluorochrome and the α-rEhVps24 antibody was labeled with AlexaPacific blue kit (Molecular Probes-Thermo Fisher), accordingly to the manufacturer's instructions. In the case of nuclear co-localization, rEhVps2 and EhVps24 were detected by secondary antibodies as described previously. pCNA (Trasviña-Arenas et al., 2017) protein was detected incubating cells with α-pCNA (Trasviña-Arenas et al., 2017) (1:200) for 1h at 37◦C, followed by incubation with <sup>α</sup>-mouse TRITC-labeled antibodies (1:100) at 37◦C for 1h. Nuclei were counterstained by 2.5µg/ml 4′ ,6-diamidino-2-phenylindole (DAPI; Sigma)

TABLE 2 | Primers used for genes amplification.


for 5 min. All preparations were preserved using Vectashield antifade reagent (Vector, Burlingame, CA, USA), examined through a Carl Zeiss LMS 700 confocal microscope in laser sections of 0.5µm and processed with ZEN 2009 Light Edition Software (Zeiss, San Diego, CA, USA).

# Phagocytosis Assays

Trophozoites were incubated at 37◦C with human erythrocytes (1:25 ratio) for different times at 37◦C and processed them for immunofluorescence (10, 15, 30, 60, and 90 min), immunoprecipitation (0 and 30 min) and flow cytometry (0, 5, and 30 min). For immunofluorescence experiments, non-ingested erythrocytes were removing using a mixture of TY1-S-33 medium and water (1:1) at 37◦C. This methodology permits to distinguish phagocytic cups from pseudopodia and other membrane arrangements, without producing significant damage to the adhered erythrocytes and membrane structures. After 4% PFA fixation, an aliquot of the cell mixture was put on coverslips and processed as described above. For some experiments, to distinguish erythrocytes, preparations were stained by Novikoff technique (Novikoff et al., 1972) and then, we counted the number of ingested erythrocytes per trophozoite in 100 trophozoites. Simultaneously, we measured the amount of hemoglobin inside trophozoites by spectrophotometry at 400 nm as described (Vacca et al., 1978).

# Immunoprecipitation Experiments

Trophozoites were lysed with 10 mM Tris-HCl, 50 mM NaCl and 100 mM protease inhibitors (PHMB, IA, NEM and TLCK), followed by cycles of freeze-thawing in liquid nitrogen and vortexing. In parallel, 200 µl of recombinant protein G-agarose (rProtein-G; Invitrogen) were incubated with 100 µg of mouse <sup>α</sup>-rEhVps32 antibody or pre-immune serum for 2 h at 4◦C, with gentle stirring. Then, rProtein-G beads were washed with 0.5% BSA in PBS, followed by additional washes with PBS for 5 min, under gentle stirring and centrifuged at 11,600 × g for 2 min. Trophozoites lysates (1 mg) were pre-cleared with 200 µl of rProtein-G (previously blocked with 2% BSA) and incubated 2 h at 4◦C under gentle stirring. Samples were centrifuged at 11,600 × g to obtain the supernatant that was added to rProtein-G previously incubated with the antibody. Preparations were incubated overnight (ON) at 4◦C and then, beads were recovered by centrifugation. After washes with PBS, 60 µl of 4X sample buffer (40% glycerol, 240 mM Tris-HCl pH 6.8, 8% SDS, 0.04% bromophenol blue and 5% β-mercaptoethanol) were added. Samples were boiled for 3 min and centrifuged again at 11,600 <sup>×</sup> g for 2 min at 4◦C. Supernatant (30 <sup>µ</sup>l) was loaded into 12% SDS–PAGE and subjected to western blot assays.

# Flow Cytometry Assays

Trophozoites (2 × 10<sup>6</sup> ) in 2 ml of TYI medium at 37◦Cwere incubated with erythrocytes for 0, 5, and 30 min as described above. At the end of each time, cell mixtures were centrifuged at 50 × g for 7 min and then, washed with 10 ml of PBS three times. Pellets were PFA (4%) fixed for 1 h and washed again as above. One ml of 0.2% TritonX-100 was added to each pellet for 10 min and subsequently, cells were washed again with PBS and incubated with 10% fetal bovine serum at 37◦C for 1 h. After this time, pellets were washed and four couples of the primary antibodies (1:100): α-rEhVps20/α-rEhVps32, α-rEhVps32/αrEhVps24, α-rEhVps24/α-rEhVps2 or α-rEhVps2/α-rEhVps20 were added separately to the cell mixtures and incubated ON at 4◦C. Cells were washed again three times with PBS and incubated for 1 h at 37◦C with the corresponding secondary antibodies (1:200) coupled to distinct fluorochromes as follows: α-rEhVps20and α-rEhVps24 with Cy5-labeled α-rabbit (cyan), α-rEhVps32 with Alexa488 α-mouse (green), and α-rEhVps2 with Alexa647 α-rat (red).Cell mixtures were washed three times with PBS, re-suspended in 500 µl of PBS and analyzed in a flow cytometer (Celesta, mod: BdFACS equipment). As controls, we used cell mixtures of 0, 5, and 30 min of phagocytosis free of antibodies (primary and secondary). Data analysis was performed using the Kaluza software.

# Labeling of Recombinant Proteins

For experiments using GUVs model, EhVps20, EhVps20(1- 173), EhVps32, and EhVps32(1-165) proteins were labeled using Alexa488 (Molecular Probes-Thermo Fisher) accordingly to manufacturer's instructions. The labeled and unlabeled proteins were separated by size exclusion chromatography. The degree of labeling was obtained accordingly to manufacturer's instructions. In all cases, we used 1:5 ratio of labeled: unlabeled proteins to maintain activity.

# Preparation of Giant Unilamellar Vesicles (GUVs)

The lipids 1-palmitoyl-2oleoyl-sn-glycero-3-phosphocholine (POPC), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPS), cholesterol (chol), 1,2-dioleoyl-sn-glycero-3-phospho-(1′ -myo-inositol-3′ -phosphate) (PI(3)P), 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1,2 dioleoyl-sn-glycero-3-phospho-L-serine (DOPS) and 1,2-dioleoyl-sn-glycero-3-phopho-(1′ rac′ glycerol) (DOPG) were purchased from Avanti Polar Lipids. In all cases, we added 1,2-Dihexadecanoyl-sn-Glycero-3-Phosphoethanolamine (TexasRed-DHPE) (Molecular Probes) at a concentration of 0.1 mol% in the lipid mixtures for the visualization of the membranes. Giant unilamellar vesicles of different lipid composition were grown using the electroformation method (Angelova and Dimitrov, 1986). Briefly, 10 µl of a 4 mM lipid stock solution in chloroform were spread on indium tin oxide (ITO) coated glasses. The excess of chloroform was eliminated under vacuum at room temperature (RT) for 1h. Then the glasses were assembled with a 2 mm-thick Teflon spacer between them to form the electroformation chamber, which was filled with a 600 mM sucrose solution that matched the osmolarity of the buffer containing the proteins (∼650 mOsmol). Finally, an electric AC-field (1.6V, 10 Hz) was applied for 1 h at different temperatures. GUVs were collected and cooled to RT before use.

# Protein Binding to the GUVs Membrane

GUVs composed of POPC (62 mol%), POPS (10 mol%), chol (25 mol%), and PI(3)P (3 mol%) were grown by electroformation at <sup>60</sup>◦C. Then, 100 <sup>µ</sup>l of the GUV suspension were placed in an observation chamber and incubated with either rEhVps20 (final concentration 125 nM), rEhVps20 (1-173) (125 nM), rEhVps32 (300 nM) or rEhVps32 (1-165) (300 nM) during 5 min at RT. In all GUVs experiments, the final volume was adjusted to a ratio of 1:1 with the protein buffer (50 mM Tris-HCl, 300 mM NaCl, pH 7.4) or proteins contained in the same buffer. The buffer and all proteins added to the GUVs were osmotically matched to the osmolarity of GUV solution. GUVs were observed with a Leica TCS SP5 confocal microscope. To quantify co-localization of the proteins, the Just Another Co-localization Plugin (JACoP) (Bolte and Cordelieres, 2006) was used in the Image J 1.48 software.

# Reconstitution OF ESCRT-III in GUVs

For the reconstitution experiments, 100 µl of GUVs with the same composition as above were placed in an observation chamber and mixed at a final ratio of 1:1 with the proteins or buffer. rEhVps20(1-173) was added to the GUVs to yield a final concentration of 125 nM, after 5 min of incubation at room temperature, rEhVps32 (300 nM) and rEhVps24 (100 nM) were added in that order, separated by 5 min incubation intervals. rEhVps2 (100 nM) was co-incubated with rEhVps24 (100 nM) and added subsequently to rEhVps20(1-173) and rEhVps32. Experiments omitting each protein or altering the order were also performed. Similarly, GUVs were incubated with five rounds of buffer as a negative control. GUVs were analyzed through a Leica TCS SP5 confocal microscope.

# ESCRT-III Reactions in Negatively Charged Lipids

GUVs composed of DOPC (100 mol%) or mixed with 5 mol% of PI(3)P or 10 mol% of either DOPS or DOPG were grown by electroformation at RT as described before. GUVs were mixed with rEhVps20(1-173) (125 nM), rEhVps32 (300 nM), and rEhVps24 (100 nM) and incubated for 5 min between de addition of each protein. The final ratio was maintained at 1:1. Percentage of GUVs with ILVs was measured in 100 randomly observed GUVs with a diameter in the range of 25 to 30µm. The mean and standard error was obtained from three independent experiments.

# Generation of EhVps20 and EhVps24 Knock Down Trophozoites

The first 444 bp from the 5′ end of the Ehvps20 gene were PCR-amplified and cloned into the pJET1.2/blunt plasmid and then, subcloned into pSAP2/Gunma plasmid, downstream of the 5' upstream segment of the EhAp-A gene, using a 5 ′ StuI site and a 3′ SacI site with the following primers: forward, 5′ -CCAAGGCCTATGTTAAATCGATTCATTGGAA AGA-3′ ; reverse, 5′ -CACGAGCTCTCTTGTGATAAAATGTCA CCAAATT-3′ (StuI and SacI restriction sites are underlined, respectively).Trophozoites of clone G3 were transfected as described (Bracha et al., 2006). Briefly, G3 trophozoites were cultured in 35-mm Petri dishes and transfected with 20 µg of each plasmid: pSAP2/GunmaEhVps20 (1–444 bp) or pSAP/Gunma, using SuperFect (Qiagen) reagent. The transfected parasites were incubated for 24 h at 37◦C and then, EhVps20 silencing was confirmed by western blot analysis, immunofluorescence assays and rate of erythrophagocytosis. In the case of the Ehvps24 gene, the first 426 bp were PCR-amplified using the following primers: forward, 5′ -GAGAGGCCTATGGG CAACCTTAATAGCCA-3′ ; reverse, 5′ -CCTGAGCTCTTGTTC ATACAATGAATCAATCTCC-3′ (StuI and SacI restriction sites are underlined, respectively). The products were cloned and treated following the protocol previously described.

# ETHICS STATEMENT

The Centre for Research and Advanced Studies (CINVESTAV) fulfill the standard of the Mexican Official Norm (NOM-062- ZOO-1999) "Technical Specifications for the Care and Use of Laboratory Animals" based on the Guide for the Care and Use of Laboratory Animals "The Guide," 2011, NRC, USA with the Federal Register Number BOO.02.03.02.01.908, awarded by the National Health Service, Food Safety and Quality (SENASICA) belong to the Animal Health Office of the Secretary of Agriculture, Livestock, Rural Development, Fisheries and Food (SAGARPA), an organization that verifies the state compliance of such NOM in Mexico. The Institutional Animal Care and Use Committee (IACUC/ethics committee) from CINVESTAV as the regulatory office for the approval of research protocols, involving the use of laboratory animals and in fulfillment of the Mexican Official Norm, has reviewed and approved all animal experiments (Protocol Number 0505-12, CICUAL 001).

# AUTHOR CONTRIBUTIONS

YA-P performed the experiments with GUVS, discussed experiments and results, wrote the manuscript; RK directed experiments with GUVS, discussed experiments and results; RJ-R performed the immunofluorescence and flow cytometry experiments, discussed experiments and results; GG-R performed experiments with E. histolytica, obtained mutants; RL directed the project related to GUVs, reviewed the manuscript, discussed strategies and experiments; RD directed the project related to GUVS, proposed and directed experiments, reviewed the manuscript; EO directed the project related to the biology of E. histolytica, proposed and directed experiments and wrote and reviewed the manuscript.

# FUNDING

This work was supported by Consejo Nacional de Ciencia y Tecnología, México. This work is part of the MaxSynBio consortium, which is jointly funded by the Federal Ministry of Education and Research of Germany and the Max Planck Society.

# REFERENCES


# ACKNOWLEDGMENTS

The authors thank Prof. Robert Seckler and Dr. Stefanie Barbirz for kindly providing their facilities for protein purification. They are also deeply grateful to Dr. Abigail Betanzos for the critical reading of the manuscript and to Tomas Sánchez and Carmen Remde for technical support. Finally, we acknowledge Eleanor Ewins for proofreading the manuscript.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00053/full#supplementary-material

Figure S1 | Effect of different combinations of ESCRT-III proteins in GUVs. GUVs prepared from POPC (62 mol%), POPS (10 mol%), chol (25 mol%), PI(3)P (3 mol%) and TR-DHPE (0.1 mol%) were incubated with different combinations of ESCRT-III proteins indicated at left during 5 min. GUVs were analyzed through confocal microscope. Scale bar: 20µm. Accession numbers/ID for proteins and genes mentioned in the text. EhADH (Q9U7F6/EHI\_181220), EhCp112 (QPI7F7/EHI\_181230), EhVps32 (C4M1A5/EHI\_169820), EhVps2 (C4LZV3/EHI\_194400), EhVps20 (C4M7T5/EHI\_114790), EhVps24 (C4M2Y2/EHI\_048690), EhVps4 (C4LYN8/EHI\_118900), Gal/Gal lectin (C4lTM0/EHI\_012270).


research and analysis. J. Comput. Chem. 25, 1605–1612. doi: 10.1002/jcc. 20084


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Avalos-Padilla, Knorr, Javier-Reyna, García-Rivera, Lipowsky, Dimova and Orozco. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Phagocytosis of Gut Bacteria by *Entamoeba histolytica*

#### Lakshmi Rani Iyer, Anil Kumar Verma† , Jaishree Paul and Alok Bhattacharya\*

School of Life Sciences, Jawaharlal Nehru University, New Delhi, India

The protist parasite Entamoeba histolytica causes amoebiasis, a major public health problem in developing countries. Only a small fraction of patients infected with the parasite display invasive disease involving colon or extra intestinal tissues such as liver. E. histolytica exists as two distinct forms, cysts, the infective form, and trophozoites, that are responsible for disease pathology. The latter multiply in the large intestine occasionally causing disease. The large intestine in humans is populated by a number of different bacterial communities and amoebic cells grow in their midst using some as food material. Several studies have shown relationship between bacteria and E. histolytica growth and virulence. However, an understanding of this relationship in human gut environment is not clear. We have investigated the possibility that there may be specific interaction of amoeba with different bacteria present in the gut environment by using a metagenomic pipe line. This was done by incubating bacteria isolated from human fecal material with E. histolytica and then identifying the bacterial population isolated from amoebic cells using a rRNA based metagenomic approach. Our results show that the parasite prefers a few bacterial species. One of these species is Lactobacillus ruminus which has never shown to be associated with E. histolytica.

#### Keywords: microbiota, phagocytosis, *Entamoeba histolytica*, metagenomics, *Lactobacillus ruminus*

# INTRODUCTION

Amoebiasis is caused by the protist parasite Entamoeba histolytica affecting both intestinal and extra intestinal tissues. It is the third leading cause of death from parasitic diseases worldwide, mainly in developing countries. It is estimated that approximately 50 million people worldwide are infected with this parasite, resulting in 40–100 thousand deaths annually (Haque et al., 2003). In the human gut, E. histolytica cells reside and multiply in the presence of a large variety of microorganisms. E. histolytica consumes bacteria by phagocytosis. Phagocytosis is an essential process in this organism and blocking this process leads to inhibition of proliferation and loss of pathogenicity of amoebic cells (Hirata et al., 2007). It is therefore believed that the interaction between bacteria and amoeba plays an important role in the growth and virulence of the parasite (Sahoo et al., 2004)

The human gut is made up of a complex community of hundreds of species of microbes, mostly commensal in nature. A high biodiversity of the gut microbiota is associated with a healthy state. Microbiota compositions can vary significantly from person to person, even within healthy individuals or twins in the same household (Lozupone et al., 2012; Smith et al., 2013). Several studies have noted that the bacterial microbiota may influence the behavior and virulence of individual pathogens, their immune response, and lead to variability in the outcome of parasitic infections (Marie and Petri, 2014; Bär et al., 2015). These organisms form a symbiotic relationship

#### *Edited by:*

Serge Ankri, Technion Israel Institute of Technology, Israel

#### *Reviewed by:*

Nancy Guillen, Centre National de la Recherche Scientifique (CNRS), France Julio César Carrero, Universidad Nacional Autónoma de México, Mexico

#### *\*Correspondence:*

Alok Bhattacharya alok.bhattacharya@gmail.com

#### *†Present Address:*

Anil Kumar Verma, Indian Council of Medical Research (ICMR), Institute of Research in Tribal Health (NIRTH), Jabalpur, India

#### *Specialty section:*

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

*Received:* 29 November 2018 *Accepted:* 04 February 2019 *Published:* 26 February 2019

#### *Citation:*

Iyer LR, Verma AK, Paul J and Bhattacharya A (2019) Phagocytosis of Gut Bacteria by Entamoeba histolytica. Front. Cell. Infect. Microbiol. 9:34. doi: 10.3389/fcimb.2019.00034 that influences human physiology and disease progression (Lozupone et al., 2012; Sassone-Corsi and Raffatellu, 2015). Metagenomic studies on gut microbiome have generated a wealth of information, that has enabled a critical examination of the

fluctuations in the gut microbiome during different disease

conditions for any possible medical intervention. A few studies have revealed that infection with E. histolytica is significantly correlated with fecal microbiome composition and diversity. Coculture with Escherichia coli can augment the virulence of an amoebic cell line (Padilla-Vaca et al., 1999). Epithelial monolayers exposed to enteropathogenic bacteria become more susceptible to damage by E. histolytica. Varet et al. (2018) have shown that live bacteria of the family Entereobacteriaceae help E. histolytica to survive oxidative stress and establish itself in the intestinal mucosa by causing changes in genes involved in glycolysis and proteasome activity. However, these changes did not occur in the presence of Lactobacillus acidophilus which could be explained by the fact that the probiotic effect of bacteria such as L. acidophilus is mediated by the ability to produce H2O<sup>2</sup> and to maintain a normal, homeostatic microbiota. Further it was shown that phagocytosis of pathogenic E. coli by amoebae increased epithelial cell damage (Galván-Moroyoqui et al., 2008). Two separate studies showed that individuals with symptomatic infection due to E. histolytica correlated with the variable levels of members belonging to the Prevotellaceae family. A cohort study conducted in south west Cameroon showed that the outcome of amoeba infections can be predicted accurately from the composition of an individual's gut microbiota, and members of Prevotellaceae was found to be one of the most prevailing taxa in individuals with asymptomatic infection whereas Prevotella copri and Prevotella stercorea are found to be downregulated in infected individuals (Morton et al., 2015). In another study conducted in children with diarrhea in Bangladesh, elevated levels of P. copri were found in patients with diarrheagenic E. histolytica infections suggesting the influence of the microbiome on the state of the gut (Gilchrist et al., 2016). Elevated levels of P. copri are also associated with risk of autoimmune disease, colitis and inflammation suggesting that inflammation driven by the gut microbiome has the potential to influence the outcome of an infection (Scher et al., 2013). A study on intestinal parasite infection in individuals from southern Côte d'Ivoire demonstrated a significant increase in the relative abundance of Bifidobacterium in Giardia duodenalis positive patients. This study suggested that intestinal protists such as E. histolytica, G. duodenalis, and Blastocystis hominus can induce significant changes in the intestinal microbiome that result in substantially different bacterial communities (Burgess et al., 2017). Real time analysis on the gut flora of patients infected with E. histolytica from our laboratory have shown significant changes in the indigenous gut flora of infected individuals showing increase in abundance of Bifidobacterium and depletion of a few predominant genera in the gut like Bacteroides, Lactobacillus, Clostridium leptum, Clostridium coccoides subgroup, Campylobacter, and Eubacterium. It was concluded that the decrease in beneficial bacterial population leads to dysbiosis of gut bacteria influencing the outcome of disease (Verma et al., 2012).

In the present study we have investigated the possibility that E. histolytica may prefer specific bacteria for phagocytosis in colonic lumen. We have identified the bacteria that are preferentially taken up by amoeba from total bacterial population prepared from human fecal material using rRNA based metagenomic analysis.

# METHODOLOGY

# Culture of *E. histolytica* and Bacterial Cells

Axenic strain of E. histolytica, HM-1:IMSS was cultured in TYI-S-33 medium (Clark and Diamond, 2002), in 250 ml culture flasks for 72 h at 35.5◦C. The spent medium was discarded and 5 ml of fresh prewarmed culture medium was added to the flask. The walls of the flasks were tapped to detach the trophozoites followed by incubation of the flasks in ice water for 5 min. The trophozoites were harvested by pelleting at 300 g for 5 min at 25◦C. The pelleted cells were resuspended in incomplete medium (TYI-S-33 medium without adding serum), number of cells were counted using a haemocytometer and then placed at 35.5◦C till used for incubation with bacteria (1 amoeba: 1000 bacterial cells).

E. coli cells (O55 and C41-DE3) were incubated in LB medium at 37◦C for 12–16 h and harvested by centrifugation at 14,000 g. The cells were then washed with Phosphate-buffered Saline (PBS) maintained at 7.4 pH, twice to remove the medium before use. The bacterial cells were counted by measuring the O.D at 600 nm.

# Uptake of Bacteria

The O55 E.coli cells were incubated with the trophozoites in incomplete TYI-S-33 medium at 37◦C in the ratio of 1:1000 for indicated time periods to allow phagocytosis to take place. The unphagocytosed bacteria were removed by washing twice with ice cold PBS containing 5 mM Na azide and 50 uM gentamycin at 300 × g for 5 min at 25◦C to remove the non-phagocytosed bacteria sticking to the surface of the parasite and finally with PBS to remove traces of the antibiotic. The absence of bacteria sticking to the surface was checked by plating an aliquot of the washed trophozoites before performing the lysis step following the protocol of (Galván-Moroyoqui et al., 2008). The trophozoites were lysed using 100 µl of 0.12% Triton X-100 in LB medium for 3 min. The lysate was plated onto LB agar plates after serial dilution and incubated overnight at 37◦C. Colonies were counted and the optimum time necessary for maximum phagocytosis to occur was inferred from the result. As a negative control, the bacteria were incubated with the trophozoites at 4◦C and plated under the same conditions.

# Assay of Phagocytosis of *E.coli* Cells by Fluorescent Microscopy

In order to visualize the phagocytosis of bacteria by the parasite, GFP expressing recombinant E.coli cells (C41-DE3) were incubated with E. histolytica (HM-1:IMSS ) cells in the ratio 1:1,000 for different time periods at 37◦C in 200 <sup>µ</sup>l of incomplete TYI-S-33 medium. After incubation, the amoebae were pelleted by centrifugation at 300 × g for 5 min at 25◦C. The unphagocytosed bacteria adhering to the outer surface of amoeba were removed by washing in 500 µl PBS containing 5 mM Na Azide and 50µM gentamycin at 300 × g for 5 min at 25◦C twice and finally with PBS once. Amoebic cells were finally suspended in 1 ml PBS. The cells were transferred to acetone cleaned coverslips placed in a prewarmed petridish and placed at 35.5◦C for 5–10 min. PBS was discarded, and the cells were then fixed using 3.7% paraformaldehyde for 30 min followed by mounting on slides using DABCO (1,4-diazbicyclo (2,2,2) octane (55) 2.5% in 80% glycerol. (Babuta et al., 2015). Fluorescence images were obtained, at 60X magnification using GFP filter set at 488 nm excitation and 510 nm emission in an Olympus Fluoview FV1000 laser scanning microscope.

# Phagocytosis of Bacterial Population Isolated From Whole Stool Sample of Healthy Volunteers

Whole stool sample was collected from two healthy age and sex matched volunteers, who were not suffering from any gastrointestinal disease and did not consume any antibiotics for 6–8 weeks prior to sample collection. Each volunteer gave informed consent for the study. About 200 mg of fresh stool sample was taken in 1 ml of sterile PBS containing 0.1% cysteine (PBSC), homogenized by vortexing for 5 min. Suspension was allowed to stand at room temperature (RT) for 5 min to allow large solid particles to settle down to the bottom of the tube. The vial was centrifuged at 1,200 × g for 5 min, and the supernatant containing all the bacteria harvested from stool was taken into a sterile tube, centrifuged at 14,000 × g and the supernatant was discarded. The pellet was then washed with PBSC and resuspended in 200 <sup>µ</sup>l PBSC and kept at 37◦C. The bacteria were pelleted again just before incubation with amoeba for the phagocytosis assay.

The bacterial pellet obtained from whole stool was incubated with E. histolytica trophozoites at the ratio of 1:1,000. (Amoeba: Bacteria) in incomplete TYI-S-33 medium in 1.5 ml Eppendorf tubes and incubated at 37◦C for 15 min to allow phagocytosis of the bacteria to occur. A negative control with bacterial pellet from one stool sample was also incubated at 4◦C for 15 min to inhibit phagocytosis. The amoebae were then washed with ice cold PBSC containing 5 mM Na azide and 50µM gentamycin and centrifuged at 300 × g for 5 min at 25◦C. This process was carried out 2–3 times to remove all the surplus bacteria. The pellet containing amoebic cells were finally lysed and DNA was isolated using QIAamp Fast DNA stool mini kit. The bacterial species phagocytosed by the amoeba were identified by 16S rRNA metagenomic analysis (V1-V5).

# Metagenomic Analysis

Amplification and sequencing of the 16S rDNA V1-V3 & V3- V5 regions was done using Illumina HiSeq 2500 Rapid/MiSeq sequencing platforms using the DNA of bacterial groups phagocytosed by amoeba and the DNA from the whole stool as control. Universal bacterial primers were used for PCR amplification. PCR reaction volume of 25 µl was set up using NEB Taq polymerase. PCR conditions were as follows: initial denaturation at 95◦C for 30 s, followed by 30 cycles of 95◦C for 30 s; annealing at specific temperatures for 45 s; and 68◦C TABLE 1 | Primer composition and annealing temperatures for the PCR amplifications.


for 30 s. The final extension temperature was 72◦C for 5 min. Primer compositions and annealing temperatures are given in **Table 1**. The PCR products were purified using Pure link kit and were used to build a library for sequencing on Illumina platforms. Libraries from each of the sample were tracked using a sequence barcode. Library quantification was done on Qubit 3.0. Short read sequences obtained from the NGS platform were analyzed by a standard pipeline for metagenomics analysis shown as a flow chart in **Figure 1**. Briefly, V1-V5 regions from Illumina paired-end sequences were extracted, read quality was checked followed by trimming of spacer and conserved regions. The paired end sequence length was 250 bp. The consensus V1-V5 region sequence is constructed using FLASH program. Filters such as conserved region filter and mismatch filter were performed to take further only the high quality V1-V5 region sequences for various downstream analyses. Chimera were also removed. The average Phred score was >30 and average GC content >50.

OTUs, Taxonomy classification and relative abundance analysis were performed using the pre-processed consensus V1- V5 sequences in QIIME. Pre-processed reads from all samples were pooled and clustered into Operational Taxonomic Units (OTUs) based on their sequence similarity using UCLUST algorithm (similarity cutoff = 0.97). Representative sequence was identified for each OTU and aligned against Greengenes core set of sequences using PyNAST program. Taxonomy classification was performed using RDP classifier against SILVA OTUs database. The phylum, class, order, family, genus, and species distribution for each sample based on OTU and reads were obtained. The taxa other than top 10 were categorized as others. The sequences not showing any alignment against taxonomic database were categorized as unknown.

# RESULTS

# Uptake of Bacteria by *E. histolytica* Cells During Phagocytosis

Uptake of bacteria by E. histolytica was initially studied using E.coli (O55) strain. **Figure 2** shows the number of colony forming units (CFUs) obtained at indicated times after incubation of E.coli with E. histolytica. The results showed that maximum bacterial uptake took place at 15 min and on further incubation the number of CFUs decreased. There was no colony formation

in the negative control when bacteria were incubated with the parasite at 4◦C. Therefore, further bacterial uptake studies were carried out at 37◦C for a time period of 15 min. Fluorescence microscopic examination of amoebic cells after incubation with GFP tagged E. coli, C41-DE3 showed the phagocytosis of GFPtagged bacteria by E. histolytica cells in **Figures 3A–E** from a time period of 0–15 min It is seen from the images that initially no E.coli cells are seen inside the parasite and later E. coli cells were engulfed by the parasite. It was also seen that there were no GFP tagged E. coli cells adhering to the surface of the amoeba suggesting that using the above protocol, no non-phagocytosed bacteria remained in the preparation after washing.

# Metagenomics Analysis

Bacteria engulfed by E. histolytica cells from a preparation of fecal material were identified by rRNA based metagenomic pipeline as described in Methodology. In both the replicate sets, a sequence library consisting of 250 × 2 base pairs was created. The raw read summary suggested a GC content of >50% and a Phred quality score >35. The read summary of the V1-V5 paired end sequences after trimming is shown in the **Table 2**.

# OTUs, Taxonomy Classification and Relative Abundance Analysis

All samples were pooled and clustered into Operational Taxonomic Units (OTUs) based on their sequence similarity (similarity cutoff = 0.97). **Table 3** shows the number of OTUs selected for further analysis in each set.

# Metagenomic Analysis of Phagocytosed Microbiota by the Parasite

The results of analysis of metagenomic sequencing data of bacterial population phagocytosed by E. histolytica (HM-1: IMSS) cells in replicate sets 1 and 2, respectively, after a period of 15 min incubation at 37◦C are shown in **Table 4**. The bacterial population preferentially phagocytosed was obtained by comparing the percent relative abundance of bacterial OTUs between the control and the phagocytosed samples. The abundance of bacterial population in the control stool samples were represented by (WS-1) and (WS-2) in Set-1 and Set-2, respectively. The bacterial population phagocytosed in both the sets are represented as (PP3715-1) and (PP3715- 2) in Set-1 and Set-2, respectively Enrichment of a bacterial group shown by increase in percent abundance of specific bacterial population in the phagocytosed sample in comparison to control indicate preferential phagocytosis. **Figures 4A–E** show the percent relative abundance of major bacterial OTUs that showed enrichment at the class, order, family, genus, and species levels, respectively. **Table 4** summarizes the bacterial groups preferentially phagocytosed from the class to the species level in the two sets. As can be seen in the table there is a significant fold increase in the percent relative abundance of these bacterial groups in the phagocytosed samples when compared

to their respective controls. The bacterial composition of the control stool samples were different however the bacterial groups enriched in the phagocytosed samples were common in both the experiment sets.

At the class level, (**Figure 4A**, **Table 4**), Bacilli, Erysipelotrichia, and Actinobacteria were preferentially phagocytosed in both the replicates and showed significant fold increases. Bacteroidia showed fold increase in Set-2. Members of Lactobacillales, Erysipelotrichales and Bifidobacteriales at the order level were enriched in both sets, Bacteroidales showed fold increase in Set-2 (**Figure 4B**, **Table 4**). At the family level, bacteria belonging to Lactobacillaceae, Erysipelotrichaeceae, Clostridiaceae-1, and Bifidobacteriaceae were selectively taken up in Set-1 and Set-2. Peptostreptococcaceae and Ruminococcaceae was taken up in Set-1, while Prevotellaceae and Enterobacteriaceae was taken up in Set-2 (**Figure 4C, Table 4**). At the Genus level, bacteria belonging to Lactobacillus, Faecalibacterium and Bifidobacterium were the predominant genera enriched in both the sets. (**Figure 4D**, **Table 4**). Genus Turicobacter was also enriched in Set-1 and Genus Prevotella and



TABLE 3 | Summary of Singleton OTUs.


Catenibacter were enriched in Set-2. Lactobacillus ruminus was the main species that showed significant enrichment consistently in both the sets. Faecalibacterium prausnitzii also was enriched in Set-1, while Catenibacterium mitsuokai and an uncultured Faecalibacterium showed fold increase in Set-2. Other species that showed fold increase in Set 1 were Bifidobacterium longum and Bifodobacterium ruminantium **Figure 4E**, **Table 4**. In a negative control incubated at 4◦C in Set-2, the percent relative abundance of Lactobacillus ruminus and uncultured Faecalibacterium is negligible (Data not shown).

# DISCUSSION

It has long been recognized that gut bacterial flora plays an important role in the biology of the protist parasite E. histolytica (Mirelman, 1987). However, it is not clear that if the parasite has specific interaction with only a subset of bacteria present in the gut. If this is true then the biology of amoeba gets modulated by different gut microbiome composition. We have investigated the relationship of amoeba and bacteria by characterizing the bacterial population phagocytosed by E. histolytica in comparison to the total gut microbiome population. In order to show specific interaction, we have estimated enrichment of a group or specific organism after phagocytosis. Our results do show enrichment of groups of organisms in phagocytosed population compared to total population. Though the starting bacterial population in the controls were very different our results suggest that the bacteria being enriched in both sets were similar. In both the sets Bacilli, Erysipelotrichia and Actinobacteria are the major class of bacteria being phagocytosed by the parasite. Further, members of the order Lactobacillales of Bacilli class, Erysipelotrichales of Erysipelotrichia class, and Bifidobacteriales of Actinobacteria class are taken up by E.histolytica. At the family level we observed that bacteria belonging to the family Lactobacillaceae, Clostridiaceae, Erysipelotrichaceae, and Bifidobacteriaceae dominated among the phagocytosed bacteria in both the sets of experiments. Our sequence analysis at the genus level revealed predominance of Lactobacillus members followed by members of Bifidobacterium and Faecalibacterium being preferentially phagocytosed. Catenibacterium also showed enrichment in Set-2 This was further confirmed at species level, where we identified species Lactobacillus ruminus in both the sets as the species being mainly phagocytosed by the parasite. Faecalibacterium prausnitzii, Bifidobacterium longum, and Bifodobacterium ruminantium were also enriched in set-1 and Catenibacterium mitsuokai was enriched in Set-2. In a negative control run with experiment set-2, incubated at 4◦C to inhibit phagocytosis, relative abundance of Lactobacillus ruminus was negligible in the sequence analysis at the species level.This evidence suggests that this organism is taken up by the parasite and is not attached to the surface of Entamoeba trophozoites. The preferential phagocytosis of L. ruminus in both experiment sets is a novel observation as this species has never been reported to be in association with E. histolytica.

Our findings reveal that the members of bacteria phagocytosed by Entamoeba constitute the healthy gut flora. It has been demonstrated earlier that the members belonging to the order Lactobacillales perform many important functions in the gut, such as production of Bacteriocins to supress the growth of pathogenic bacteria, synthesis of lactic acid and H2O2, fortification of epithelial barrier by induction of mucin secretion and enhancement of tight-junction functioning (Lebeer et al., 2008) and help in the prevention of infectious diarrhea, antibiotic associated diarrhea and diarrhea in children (Van Neil et al., 2002). Varet et al. (2018) have shown that the gene responses observed in E. histolytica in the presence of Enterobacteriaceae favoring their survival in the presence of the oxidative stress do not occur during coculture with L. acidophilus.

Lactobacillus ruminus, the major bacterial species phagocytosed by E. histolytica as observed here, is an autochthonous member of healthy gut and considered to be a potential probiotic. It can produce flagella (O'Donnell et al., 2015) and inhibits the growth of antibiotic resistant pathogens like vancomycin resistant Staphylococcus aureus and Enterococci. (Yun et al., 2005). Depletion of L. ruminus may promote colonization of the pathogenic bacteria in the gut. However further studies are needed in order to understand why L. ruminus is preferentially phagocytosed by E.histolytica.

Members of the family Erysipelotrichaceae and Clostridiaceae were found to be enriched in our phagocytosis experiment. Bacteria belonging to Erysipelotrichales are highly immunogenic and play a role in preventing gut inflammation (Kaakoush, 2015). Higher levels of Catenibacterium genus, a polysaccharide degrading genus belonging to Erysipelotrichales are found in


mediterranean diets which are in turn associated with reduced inflammation markers (Shankar et al., 2017). Phagocytosis of Catenibacterium by E. histolytica could lead to lower levels of beneficial Catenibacterium in the gut.

Members of Clostridiaceae family are known to produce butyrate as the end product of their fermentation. Butyrates are the preferred energy source for colonocytes and are important anti-inflammatory modulators (Lopetuso et al., 2013). Enrichment of Bifidobacterium ruminantum and Bifidobacterium longum at the species level suggest their phagocytosis by E. histolytica. B.longum protect the gut from enteropathogenic infection through the production of acetate (SCFA). Acetate is the major energy source for colonocytes. (Fukuda et al., 2011). Verma et al. (2012) have found significant increase of Bifidobacterium species in amoebic patients by Real time PCR. They suggest a cross talk between the intestinal epithelial cells, the bacteria and the parasite, leading to increased mucus secretion by the intestinal cells and colonization of the intestinal epithelium by Bifidobacterium species during E. histolytica infections.

FIGURE 4 | show the bacteria phagocytosed by E. histolytica at the class, order, family, genus, and species levels in replicates Set-1 and Set-2 respectively. Prefix (WS) represents the starting bacterial population (control) and prefix (PP) represents the phagocytosed bacteria in each set. 3715 represents incubation at 37◦C for 15 min.

In conclusion our study shows E. histolytica preferentially phagocytosed some of the beneficial bacteria that are required for the maintenance of a healthy gut amongst which are family Lactobacillales, Erysipelotrichales, Clostridales, and Bifidobacteriales. Phagocytosis of these bacteria may cause dysbiosis of gut bacteria and creates conditions for the proliferation of the parasite in the human intestinal lumen. At the species level our data shows that preferential phagocytosis of Lactobacillus ruminus by Entamoeba was quite significant.

# DATA AVAILABILITY

Raw reads data for this project were deposited at SRA (NCBI) data base with the following accession numbers:


The link to the data is: https://www.ncbi.nlm.nih.gov/Traces/ study/?acc=PRJNA515425.

# ETHICS STATEMENT

This study was carried out in accordance with the recommendations of IERB-JNU (Institutional Ethics review board of Jawaharlal Nehru University). (IERB Ref. No.2015/Faculty/82) with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the Institutional Ethics review board of Jawaharlal Nehru University.

# AUTHOR CONTRIBUTIONS

LI, AV, JP, and AB conceived, designed the experiments, and analyzed the data. LI and AV performed the experiments. JP and AB contributed reagents, materials, and analysis tools. LI, JP, and AB wrote the paper.

# FUNDING

This research was supported by a grant from the Department of Biotechnology, Government of India and European Union under Indo-European collaboration on biotechnology with Indian investigators (Order No: BT/IN/Infect-Era/AB/2015)**.**

# ACKNOWLEDGMENTS

The authors thank Department of Biotechnology, Government of India and European Union for financial support. They also thank Department of Science & Technology, Government of India for JC Bose Fellowship (AB) and FIST grant to the School of Life Sciences.

# REFERENCES


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Iyer, Verma, Paul and Bhattacharya. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# A Flow Cytometry Method for Dissecting the Cell Differentiation Process of Entamoeba Encystation

Fumika Mi-ichi\*, Yasunobu Miyake, Vo Kha Tam and Hiroki Yoshida

*Division of Molecular and Cellular Immunoscience, Department of Biomolecular Sciences, Faculty of Medicine, Saga University, Saga, Japan*

Amoebiasis is caused by *Entamoeba histolytica* infection, a protozoan parasite belonging to the phylum Amoebozoa. This parasite undergoes a fundamental cell differentiation process from proliferative trophozoite to dormant cyst, termed "encystation." The cysts formed by encystation are solely responsible for the transmission of amoebiasis; therefore, *Entamoeba* encystation is an important subject from both biological and medical perspectives. Here, we have established a flow cytometry strategy for not only determining the percentage of formed cysts but also for monitoring changes in cell populations during encystation. This strategy together with fluorescence microscopy enables visualization of the cell differentiation process of *Entamoeba* encystation. We also standardized another flow cytometry protocol for counting live trophozoites. These two different flow cytometry techniques could be integrated into 96-well plate-based bioassays for monitoring the processes of cyst formation and trophozoite proliferation, which are crucial to maintain the *Entamoeba* life cycle. The combined two systems enabled us to screen a chemical library, the Pathogen Box of the Medicine for Malaria Venture, to obtain compounds that inhibit either the formation of cysts or the proliferation of trophozoites, or both. This is a prerequisite for the development of new drugs against amoebiasis, a global public health problem. Collectively, the two different 96-well plate-based *Entamoeba* bioassay and flow cytometry analysis systems (cyst formation and trophozoite proliferation) provide a methodology that can not only overcome the limitations of standard microscopic counting but also is effective in applied as well as basic *Entamoeba* biology.

Keywords: flow cytometry, Entamoeba, encystation, cell differentiation, amoebiasis

# INTRODUCTION

Entamoeba histolytica is the causative agent of amoebiasis. Amoebiasis is a global public health problem owing to its high morbidity and mortality rates (Ralston and Petri, 2011; Watanabe and Petri, 2015). High numbers of individuals are infected with E. histolytica but are asymptomatic and do not require treatment. However, they are important because they unconsciously spread the disease (Watanabe and Petri, 2015). Clinical treatment is currently inadequate because only a few drugs are available, and an effective vaccine has not been developed (Haque et al., 2003; Quach et al., 2014).

#### Edited by:

*Anjan Debnath, University of California, San Diego, United States*

#### Reviewed by:

*Lesly Temesvari, Clemson University, United States Upinder Singh, Stanford University, United States*

#### \*Correspondence:

*Fumika Mi-ichi fumika@cc.saga-u.ac.jp*

Received: *19 March 2018* Accepted: *02 July 2018* Published: *24 July 2018*

#### Citation:

*Mi-ichi F, Miyake Y, Tam VK and Yoshida H (2018) A Flow Cytometry Method for Dissecting the Cell Differentiation Process of Entamoeba Encystation. Front. Cell. Infect. Microbiol. 8:250. doi: 10.3389/fcimb.2018.00250*

E. histolytica, a protozoan parasite belonging to the phylum Amoebozoa, survives drastic environmental changes outside as well as inside its human host, by alternating its form between proliferative trophozoite and dormant cyst. These two distinct stages are connected by two cell differentiation processes: "encystation" and "excystation." Encystation is a process for differentiation of trophozoite into cyst whereas excystation is that of cyst into trophozoite. Trophozoites colonize the large intestine and proliferate there. Some of them differentiate into cysts. These cysts are excreted, and are then ingested by new hosts and reach the small intestine, where they hatch into trophozoites (Watanabe and Petri, 2015; Miichi et al., 2016). Encystation, a parasitic strategy involving a fundamental cell differentiation process, appears simple but is closely associated with transmission of the disease. The transmission of amoebiasis is solely mediated by cysts formed by encystation; therefore, inhibition of encystation is an effective strategy against amoebiasis. Hence, Entamoeba encystation is an important subject from a medical as well as a biological perspective. Nevertheless, the underlying molecular and cellular mechanisms require further elucidation (Mi-ichi et al., 2016).

In this study, we describe a method for counting Entamoeba cysts, an indispensable procedure in Entamoeba encystation studies. It is a flow cytometry method using premixed calcofluor (CF) and Evans blue (EB) dyes, which is rapid and quantitative, providing reproducible and reliable data. By exploiting this method together with fluorescence microscopy, we visualized differentiating cells that appeared during Entamoeba encystation. We also standardized a flow cytometric protocol to separately count live and dead E. histolytica trophozoites. By combining these two different 96-well plate-based systems we were able to screen a chemical library for potential leads that inhibit Entamoeba encystation and/or trophozoite proliferation, which is a prerequisite step for the development of new drugs against amoebiasis. To confirm the effectiveness of this combined system, we screened 400 compounds exhibiting diverse scaffolds from the Pathogen Box of the Medicine for Malaria Venture (MMV; https://www.pathogenbox.org/).

# MATERIALS AND METHODS

# Chemicals

Calcofluor White Stain, a premixed CF (Fluorescent Brightener 28) and EB dye, was purchased from Sigma-Aldrich (St. Louis, Mo, USA). CF and EB were from Sigma-Aldrich and Nacalai Tesque (Kyoto, Japan), respectively. N-Lauroylsarcosine sodium salt (>94.0% purity) was from Sigma-Aldrich.

Lactacystin (>94.2% purity) and polyoxin D (>94.5% purity) were purchased from Biolinks Co. Ltd. (Tokyo, Japan) and Kaken Pharmaceutical Co. Ltd. (Tokyo, Japan), respectively, whereas metronidazole (>98.0% purity) and paromomycin (>98.0% purity) were both from Sigma-Aldrich. All four compounds were dissolved in sterilized water, respectively, at 1, 10, 50, and 50 mM as the stock solutions. Aliquots of 50 <sup>µ</sup>L were stored at <sup>−</sup>30◦C, and freeze-thaw cycles were less than two before use.

The Pathogen Box, 400 compounds exhibiting diverse scaffolds, was provided by MMV (https://www.pathogenbox. org/); each compound was dissolved in DMSO at 10 mM and distributed into individual wells of 96-well plates (10 µL/compound and 80 compounds/plate). Ninety microliters of DMSO were then added to each well and then 10 replicates were made (10 µL aliquots of all the compounds' stocks at 1 mM) and stored at −30◦C. When needed, a set of replicates covering all 400 compounds (10 µL aliquoted at 1 mM each) was thawed, and 1 µL for the trophozoite proliferation assay and 2.4 µL for the cyst formation assay was dispensed into wells of a 96-well culture plate to make a replicate. Auranofin (>98% purity) (which is identified as E-H-05 in the Pathogen Box) was also purchased from Sigma-Aldrich, dissolved in DMSO to 10 mM and dispensed into 50 µL aliquots for storage at −30◦C.

# Parasite Culture and Sample Preparations for Each Analysis

E. invadens (IP-1) and E. histolytica (HM-1:IMSS cl6) were routinely maintained as described (Mi-ichi et al., 2009, 2015). For the cyst formation assay using E. invadens, encystation inducing treatment was performed as described (Mi-ichi et al., 2015), except that E. invadens trophozoites suspended in encystation medium were seeded in a 96-well culture plate (240 µL per well) and the plate was sealed as described (Suresh et al., 2016) using Parafilm <sup>R</sup> from Bemis Flexible Packaging (Neenah, WI, USA). Note that the final cell density of 6 × 10<sup>5</sup> cells/mL was not different from that in Mi-ichi et al. (2015); therefore, the initial number of cells per well was 1.44 x 10<sup>5</sup> . After incubating at 26◦C for the period indicated, cells in 96-well culture plates were harvested by centrifugation at 440 × g for 5 min at 4◦C. Cell pellets were then suspended in 120 µL PBS containing an appropriate staining reagent, a premixed CF and EB, CF, or EB. The premixed CF and EB, Calcofluor White Stain, was diluted 5-fold with PBS just before use; final concentrations of CF and EB used were 0.2 and 0.1 mg/mL, respectively. The working solutions of CF and EB were prepared just before use by 5 fold dilution of the stocks with PBS to 0.2 and 0.1 mg/mL, respectively, and stored at room temperature. The obtained cell suspensions were held for 15 min at room temperature and then precipitated by centrifugation at 440 × g for 5 min at 4 ◦C, washed with 120 <sup>µ</sup>L PBS, and precipitated again. Finally, the cell pellet was resuspended in flow cytometry buffer (0.5% BSA, 2 mM EDTA, and 0.05% azide in PBS) for injection into a flow cytometer [MACSQuant from Miltenyi Biotec (Bergisch Gladbach, Germany)]. When needed, sarcosyl treatment was performed before the staining step, as described previously (De Cádiz et al., 2013; Mi-ichi et al., 2015).

For the E. histolytica trophozoite proliferation assay, trophozoites were harvested from a routine culture by centrifugation at 440 × g for 5 min at 4◦C, and the harvested cells were resuspended in fresh standard culture medium (Mi-ichi et al., 2009). A 96-well culture plate was then seeded with the obtained cell suspension (100 µL per well; final cell density, 1 × 10<sup>5</sup> cells/mL); therefore, the initial number of cells per well was 1 x 10<sup>4</sup> . After incubation at 37◦C for 24 h under anaerobic conditions using Anaerocult A (Merck), cells in the 96-well culture plate were harvested, stained with 1.0µg/mL PI and processed for flow cytometry analysis as described above for E. invadens.

For the treatment of E. invadens or E. histolytica with compounds, each stock solution of lactacystin, polyoxin D, metronidazole, paromomycin, or auranofin was serially diluted in the medium used for each treatment. As controls, water was used in place of the first four compound solutions and the final water content in a well was 1% (vol/vol) whereas DMSO was used as the control of auranofin and the final DMSO content was 1% (vol/vol). The highest control solvent content in any well was 1% (vol/vol). Each compound was added when the cells were seeded into 96-well culture plates and incubated for 72 h in the cyst formation assay or for 24 h in the trophozoite proliferation assay. When needed, compounds were added 48 h after induction of encystation and cells then analyzed by a flow cytometer as described above.

For screening the 400 Pathogen Box compounds, 1 mM of each compound stock solution or DMSO control were added to the culture medium for each assay at 1% (vol/vol).

# Conditions for Flow Cytometry Analysis

A MACSQuant from Miltenyi Biotec was used as a 96-well platebased flow cytometer. CF was excited using a 405 nm laser, and the fluorescence emission was collected using a 450/50 filter. EB and PI were both excited using a 488 nm laser and the fluorescence emission was collected using a 614/50 nm filter (Hed et al., 1983). The processing volume was set at 30 µL from the 120 µL fluorescent dye(s)-treated cell suspension in each well prepared as described above. The obtained data were analyzed using Flow Jo software (Tree Star, Ashland, OR, USA).

After the samples from either an encystation-inducing culture or a standard culture for trophozoite proliferation were treated with staining solutions, the flow cytometry analysis of the prepared samples was completed within 3 or 1 h, respectively; the stability of the prepared samples was confirmed by the reproducibility of data from the chemical library screening. It should be mentioned that samples prepared for the trophozoite proliferation assay were suspended by pipetting just before processing in the flow cytometer.

# Fluorescence Microscopy

A portion of the E. invadens sample prepared for flow cytometry analysis, as described above, was examined under a BZ-9000 fluorescence microscope (Keyence, Osaka, Japan). The obtained images were processed using BZ-II software (Keyence).

# RESULTS

# Establishing a Flow Cytometry Method for Measuring Entamoeba Cyst Numbers

In Entamoeba encystation studies, the in vitro culture of E. invadens, a reptilian parasite, and not that of E. histolytica, has been adopted as a model system. This is mainly because laboratory strains of E. histolytica do not encyst after adaptation to in vitro culture conditions; however, E. invadens strains are able to undergo in vitro encystation (Sanchez et al., 1994; Coppi and Eichinger, 1999; Mi-ichi et al., 2016). In vitro encystation is usually induced by transfer of proliferating trophozoites from a standard culture to encystation conditions (Sanchez et al., 1994; Mi-ichi et al., 2015). In encystation-inducing conditions, the number of cysts formed increases with the length of incubation and then reaches a plateau. In a standard encystation assay, measurement is manually performed under a microscope at 72 h post induction (Sanchez et al., 1994).

To obtain accurate data, reproducible distinction between trophozoites and cysts is essential. Flow cytometry has been recently reported as a suitable method for this (Welter et al., 2017). Here, we attempted to introduce and standardize conditions to achieve these criteria. Samples prepared from routine trophozoite proliferation cultures gave a single population by forward scatter (FSC)/side scatter (SSC) analysis (**Figure 1A**, upper panel). Samples prepared from encystationinducing cultures at 72 h after induction also only gave a single population, although its area was moved to a slightly lower FSC/SSC position (**Figure 1B**, upper panel).

We then focused on CF and EB dyes to differentially stain trophozoites and cysts. CF is a fluorochrome that binds to structures containing chitin, a component in the Entamoeba cyst wall, and has, therefore, been used in standard microscopy methods (Arroyo-Begovich et al., 1980; Herrera-Martínez et al., 2013). EB was tentatively used for staining live Entamoeba cells because it was shown to stain the plasma membrane of viable human neutrophils (Hed et al., 1983). To confirm that these two dyes can differentiate between trophozoites and cysts, trophozoites in a routine culture and cells in an encystationinducing culture at 72 h after induction were separately stained with a premixed reagent of CF and EB dyes, Calcofluor White Stain (Sigma-Aldrich) (**Figures 1A,B**, lower panels). As expected, the single FSC/SSC population detected in the trophozoite culture gave only a single CF negative (CF−) and EB positive (EB+) population, indicating that the trophozoites can be isolated as a CF−/EB<sup>+</sup> population (**Figure 1A**, lower panel). In contrast, the encystation-inducing culture consisted of several distinct populations. The largest population exhibited CF and not EB fluorescence (CF+/EB−). Another two populations also produced CF fluorescence and stronger EB fluorescence than trophozoites (CFlow/EBstrong and CF+/EBstrong) (**Figure 1B**, lower panel). Most cells in these three populations remained after sarcosyl treatment, which is used to eliminate trophozoites to facilitate the counting of cysts (**Figure 2A**; De Cádiz et al., 2013). Additionally, one population of cells exhibiting neither CF nor EB fluorescence, was detected (CF−/EB−) (**Figure 1B**, lower panel). This population partly overlapped with the CF−/EB<sup>+</sup> population, which consists of trophozoites (see **Figure 1A**, lower panel). Furthermore, the cells in this population were not resistant to sarcosyl treatment (**Figure 2A**).

All five populations (CF−/EB<sup>+</sup> in the trophozoite culture and CF+/EB−, CFlow/EBstrong, CF+/EBstrong, and CF−/EB<sup>−</sup> in the encystation-inducing culture) can be detected in a fluorescent reagent-binding dependent manner. Staining with either CF or EB alone provided cell distributions along only the CF- or EB-signal axis, respectively (**Figure 2B**). Under a fluorescence

*invadens* trophozoites prepared from the routine culture. (B) *E. invadens* cells from the encystation-inducing culture at 72 h after induction. Representative data are shown from three independent experiments.

microscope, the most frequently observed cells in encystationinducing culture at 72 h post induction gave CF and almost no EB fluorescence, consistent with the flow cytometry results. Furthermore, these cells were round and surrounded by a thick layer (see 72 h images in **Figure 5**). Round cells yielding EB as well as CF fluorescence, or yielding EB and almost no or no CF fluorescence were also be observed at much lower frequencies (**Figure 3**). Collectively, these results indicate that the majority of cells in the Entamoeba encystation-inducing culture at 72 h after induction are round, surrounded by a thick chitin cyst wall to which CF dye binds and are resistant to sarcosyl treatment, and emit CF and not EB fluorescence; in other words, based on their morphology, composition and properties against detergent treatment, the major CF+/EB<sup>−</sup> population consists of cells showing typical features of mature Entamoeba cysts.

Consistency of mature Entamoeba cyst counting obtained by the above flow cytometry analysis and standard microscopy counting was confirmed using an encystation-inducing culture at 72 h after induction. For instance, the flow cytometry analysis calculated mature cyst densities to be 4.47 × 10<sup>5</sup> cells/ml, whereas the microscopy method performed by counting at most a few hundred cells as described by (Mi-ichi et al., 2015) gave densities of 4.80 × 10<sup>5</sup> cells/ml.

Hence, these results demonstrate a method for counting mature Entamoeba cysts, which is rapid and quantitative, providing reproducible and reliable data. Furthermore, the method can be integrated into a 96-well plate-based Entamoeba encystation bioassay using flow cytometry and premixed CF and EB.

# Monitoring the Cell Differentiation Process of Entamoeba Encystation

Our 96-well plate-based Entamoeba bioassay can count the number of mature cysts formed by encystation (see the above section). This system is an alternative to the standard microscopy method used in the Entamoeba field. Moreover, from the results demonstrated in the above section, we predict that this methodology will also be effective in studying the molecular and cellular mechanisms underlying encystation. To confirm our prediction, we monitored the dynamics of population changes during the course

of encystation by characterization of populations using flow cytometry analysis in combination with fluorescence microscopy.

The only population detected at 0 h was the single CF−/EB<sup>+</sup> population, which is composed of proliferating trophozoites (**Figure 4**; see **Figure 1A**, lower panel). The cells giving EB- and

not CF-signal were confirmed to have trophozoite morphology by fluorescence microscopy (**Figure 5**). Similar results were obtained by both flow cytometry analysis and fluorescence microscopy at 4 h after induction of encystation (**Figures 4**, **5**).

The single CF−/EB<sup>+</sup> population detected up to 4 h after induction, then became smaller and smaller from 8 h on (**Figure 4**). Inversely, two new populations, CFlow/EB<sup>+</sup> and CF+/EB<sup>+</sup> were generated (**Figure 4**). Consistent with the flow cytometry results, fluorescence microscopy showed that from 8 h, CF fluorescent cells became evident, while cells giving almost no CF fluorescence were also still detected, and their sizes and morphologies were smaller and less motile than those of proliferating trophozoites. From 12 h, the most frequently observed cells showed very similar size and morphology to the mature cyst, but unlike mature cysts, both gave EB and CF fluorescence (**Figure 5**).

The CF+/EB<sup>+</sup> population moved lower and lower along the EB signal axis from 12 h and reached a position similar to that of the mature cyst population (CF+/EB−) at 72 h (**Figure 4**; see **Figure 1B**, lower panel and the 4th paragraph in the 1st section). Consistently, fluorescence microscopy revealed that cells mainly observed from 12 to 16 h were round and gave EB as well as CF fluorescence, but, from 16 h, cells lost the EB but retained the CF signal and the CF+/EB<sup>−</sup> population became dominant from 30 h (**Figure 5**). These results indicate that the CF+/EB<sup>+</sup> population sequentially becomes the CF+/EB−, or mature cyst population.

Taken together, analysis of the time course enabled visualization of the main differentiation process of Entamoeba encystation: proliferating trophozoites (CF−/EB<sup>+</sup> population) become dormant cysts (CF+/EB<sup>−</sup> population) via CFlow/EB<sup>+</sup> and CF+/EB<sup>+</sup> populations. Furthermore, it showed that cells in the CF+/EB<sup>+</sup> population were already becoming round and surrounded by chitin and that their cellular content had reached a plateau. These cells were losing permeability to solutes, such as EB, probably because they were acquiring a complete cyst wall structure like the mature cysts in the CF+/EB<sup>−</sup> population (**Figure 6A**).

# Effectiveness of Combining the Two 96-Well Plate-Based Entamoeba Bioassay and Flow Cytometry Systems for the Development of New Drugs Against Amoebiasis

### Assessment of the 96-Well Plate-Based E. invadens Encystation Assay Connected to Flow Cytometry for Chemical Library Screening

To develop new preventive measures against amoebiasis, such as anti-amoebic and amoebiasis transmission-blocking drugs, screening potential leads from chemical libraries is a prerequisite toward the ultimate goal. To assess the applicability of the above system, the E. invadens encystation bioassay was integrated with flow cytometry and the model compounds, lactacystin, and polyoxin D, which show a significant inhibitory effect on encystation (Avron et al., 1982; Gonzalez et al., 1999), were assayed. Metronidazole and paromomycin, which are clinically used to treat amoebiasis patients, were also assayed (Marie and Petri, 2013; Penuliar et al., 2015). In the present assay system, the inhibitory effect of different compounds on encystation was evaluated by the encystation rate expressed as the mature cyst population percentage (CF+/EB−) in each sample relative to that in a solvent-treated control (set as 100%).

Lactacystin inhibited cyst formation (reduction of the CF+/EB<sup>−</sup> population) with an IC<sup>50</sup> value of 1.56 <sup>±</sup> 0.251µM, while polyoxin D did not show any inhibitory effect (**Figure 7A**). The IC<sup>50</sup> value of lactacystin was close to the previously reported value of 1.25–2.5µM (Gonzalez et al., 1999). However, the result for polyoxin D was inconsistent with one previous study, which

showed a dose-dependent inhibition of cyst formation at 2– 500µg/mL (Avron et al., 1982), but was supported by another study that demonstrated no effect on chitin synthase activity, a predicted target of polyoxin D, in cyst lysate at 100µg/mL (Das and Gillin, 1991). Furthermore, at ∼IC99, lactacystin caused significant accumulation of the CF−/EB<sup>−</sup> population (**Figure 7A**, 10µM).

Metronidazole and paromomycin also inhibited cyst formation with IC<sup>50</sup> values of 14.7 ± 3.06 and 2.65 ± 0.311µM, respectively (**Figure 7A**). At ∼IC99s, they also caused irregularities in population distribution; metronidazole caused significant accumulation of the CFlow/EBstrong population (**Figure 7A**, 55.6µM) whereas paromomycin caused significant accumulation of the CF−/EB<sup>−</sup> and CFlow/EBstrong populations (**Figure 7A**, 18.5µM). These results suggest that both metronidazole and paromomycin, similar to lactacystin, possess inhibitory activity against encystation. Alternatively, the possibility exists that all three compounds affect only trophozoites in both culture conditions used for trophozoite proliferation and cyst formation assays; in the encystationinducing culture, there is a lag time to commit to trophozoite differentiation.

Metronidazole was added to Entamoeba encystation-inducing cultures at 48 h post induction at six different concentrations ranging from 2.06 to 500µM. Samples were then analyzed at 72 h post induction. None of the concentrations tested, even 500µM [∼10 times higher than the IC<sup>99</sup> concentration (see **Figure 7A**)], affected the number of cysts formed at 72 h compared with the present encystation assay (0–72 h) (**Figure 7B**). Flow cytometry and fluorescence microscopy indicated that the majority of cells in the culture at 48 h post induction were similar to mature cysts (see **Figures 4**, **5**). Collectively, these results indicate that the majority of Entamoeba cells in the encystation-inducing culture at 48 h post induction were tolerant to metronidazole. This finding indicates that the halting of cyst formation by metronidazole observed in the present encystation assay (0–72 h) is not a direct effect on cells that show similar characteristics to mature Entamoeba cysts.

images are demonstrated from two independent experiments.

# Standardizing the 96-Well Plate-Based Flow Cytometry to Count Live E. histolytica Trophozoites,

Which Can Be Integrated Into the Proliferation Assay Whether a compound identified in the encystation screen exerts an effect on trophozoite growth is among the most important questions that should be addressed to further characterize the compound. For this assessment, we attempted to standardize the flow cytometry analysis using propidium iodide (PI), a membrane-impermeable dye that specifically stains dying or dead cells (Chatterjee et al., 2015). In the established assay system, the inhibitory effect of different compounds on trophozoite proliferation is evaluated by the live cell rate, which is calculated using the live trophozoite cell number [PI negative (PI−) population] and the total cell number in each sample relative to those in solvent-treated controls (set as 100%) (**Figure 8**).

We then assayed the four model compounds, lactacystin, polyoxin D, metronidazole, and paromomycin, to assess the applicability of this system for screening a chemical library. Lactacystin, metronidazole, and paromomycin treatment of the in vitro E. histolytica culture dose-dependently decreased the PI<sup>−</sup> population (live trophozoites) while polyoxin D treatment did not; IC<sup>50</sup> values of lactacystin, metronidazole and paromomycin were determined as 3.6 ± 0.14, 12.3 ± 2.47, and 11.5 ± 0.71µM, respectively (**Figure 8**). The IC<sup>50</sup> value of lactacystin treatment of trophozoites for 72 h or those of metronidazole or paromomycin treatment of trophozoites for 48 h were previously reported as ∼1.0 (Makioka et al., 2002), 9.5 (Debnath et al., 2012; Penuliar et al., 2015) and ∼12.6µM (Penuliar et al., 2015), respectively. These values are close to those determined in the present study. The slightly higher IC<sup>50</sup> values of lactacystin and metronidazole in the present study were probably because of the shorter incubation time (24 h). These results indicate that lactacystin, metronidazole, and paromomycin, all of which significantly inhibit cyst formation (see **Figure 7A**), also exert a cytotoxic effect on the proliferating trophozoites.

Taken together, the flow cytometry-integrated Entamoeba bioassay for cyst formation or trophozoite proliferation rapidly and quantitatively evaluated the effects of different compounds on the maintenance of the Entamoeba life cycle. Nevertheless, exploiting only the encystation assay may give misleading results; for instance, the present encystation assay (0–72 h) could not discriminate whether a compound acted on the mature cyst itself, on differentiating cells that appeared during the course of encystation, or only on trophozoites because there was a lag time to commit to trophozoite differentiation. To compensate for this drawback, two different 96-well plate-based Entamoeba bioassay and flow cytometry analysis systems were combined for the primary screening of a chemical library: one was for E. invadens encystation and the other was for E. histolytica trophozoite proliferation.

# Screening a Chemical Library to Identify Compounds That Affect Entamoeba Trophozoite Proliferation and Cyst Formation

To validate the effectiveness of this methodology for the development of new drugs against amoebiasis, a chemical library was screened to obtain compounds that exert effect(s) on Entamoeba processes essential for maintenance of its life cycle: trophozoite proliferation, cyst formation, or both. The Pathogen Box provided by the MMV was chosen as a model chemical library, and contains 400 compounds exhibiting various scaffolds (https://www.pathogenbox.org/).

Among 400 compounds screened, 22 consistently showed a high negative effect at 10µM (>80% reduction) in either cyst formation or the trophozoite proliferation assay, or both (**Supplementary Figures S1A–F**, **Table 1**; see **Figure 9** for their structures); two compounds (C-F-08 and E-H-05) almost completely arrested both of these biological processes. Fourteen compounds (A-A-09, A-B-10, A-B-11, A-D-03, A-D-11, A-G-07, A-H-11, B-E-06, C-A-10, C-D-11, D-E-05, D-G-11, E-G-04, and E-G-08) almost completely halted cyst formation but only partially impaired trophozoite proliferation levels (35.0– 74.9%). In contrast, two compounds (B-A-03 and B-B-06) almost completely arrested trophozoite proliferation but inhibited the cyst formation only partially (23.7–55.1%). Four compounds (B-F-10, B-G-03, D-H-03, and E-A-02) showed a biased inhibitory pattern; B-F-10 almost completely inhibited cyst formation, but did not affect trophozoite proliferation, whereas B-G-03, D-H-03, and E-A-02 showed the inverse effect.

E-H-05 is auranofin, which is 10-times more potent than metronidazole against E. histolytica trophozoites and has been used as an FDA (food and drug administration in USA)-approved anti-rheumatoid drug (Debnath et al.,



*Data from two independent cyst formation assays or trophozoite proliferation assays are summarized.*

2012). IC<sup>50</sup> values of auranofin for Entamoeba cyst formation and trophozoite proliferation were determined as 1.73 ± 0.70 and 0.690 ± 0.139µM, respectively (**Figures 10A,B**; 0– 72 h). Furthermore, addition of auranofin to the Entamoeba encystation-inducing culture at 48 h post induction did not affect the number of cysts formed at 72 hr, similar to the effect of metronidazole (**Figure 10A**; see **Figure 7B**). These results can be interpreted as an indirect negative effect of auranofin on Entamoeba cyst formation by causing Entamoeba cell dysfunction; the cell population includes proliferating trophozoites and differentiating cells that do not yet show characteristics of mature cysts (see **Figures 4**, **5**; 48 h).

Hence, combining the two different 96-well plate-based Entamoeba bioassay and flow cytometry analysis systems (cyst formation and trophozoite proliferation) is effective for the development of new drugs against amoebiasis, such as antiamoebic and amoebiasis transmission-blocking drugs.

# DISCUSSION

We have developed two standardized 96-well plate-based Entamoeba bioassay and flow cytometry systems: one is for an encystation assay using E. invadens and the other is for a trophozoite proliferation assay using E. histolytica. Each system can rapidly, reproducibly and quantitatively analyze the biological processes essential for maintenance of the Entamoeba life cycle; therefore, studies exploiting each or both systems will enable investigations from both biological and medical perspectives. For example, visualization of the Entamoeba encystation cell differentiation process was achieved by identifying and characterizing distinct cell populations. The feasibility of providing potential leads for the development of amoebiasis transmission-blocking and anti-amoebic drugs was demonstrated by screening a chemical library, the Pathogen Box from MMV (https://www.pathogenbox.org/).

Recently, Welter et al. reported using flow cytometry for fixed E. invadens prepared from an encystation-inducing culture with the fluorescent dye, Congo Red. This method can separate the trophozoite and cyst populations and assess the effect of compounds on the dynamics of these two populations (Welter et al., 2017). In the present study, as well as a fluorescent dye (CF) that stains chitin (Arroyo-Begovich et al., 1980), a dye (EB) that stains membrane (Hed et al., 1983) were used and more time points during encystation were then

analyzed to monitor the dynamics of population changes. A time course study using the present system for E. invadens encystation combined with fluorescence microscopy revealed that differentiation of proliferating trophozoites into dormant, mature cysts involves at least two distinct cell types that exist as sequential precursors during encystation. One is morphologically similar to the proliferating trophozoite but appears less motile, and emits EB fluorescence at the same level as proliferating trophozoites. The other is a round cell that looks like the mature cyst, and emits CF and EB fluorescence at the same levels as the mature cyst and the proliferating trophozoite, respectively. Detecting these two cell types, each of which shows mixed characteristics of proliferating trophozoites and mature cysts, is intriguing because it provides strong evidence for the existence of transient forms during encystation, as has been suggested (Chatterjee et al., 2009). However, correlation of the identified precursor cells to the assumed transient forms remains undetermined. In addition, more detailed characterization of the two precursor cell types is needed at cellular and molecular levels, to provide new mechanistic insights into Entamoeba encystation.

Other intriguing cells include the CF−/EB−, CFlow/EBstrong , and CF+/EBstrong populations. The CF−/EB<sup>−</sup> and CFlow/EBstrong cells produced CF-signal at lower levels, suggesting that these cells don't synthesize or accumulate normal levels of chitin, the target of the CF dye. Meanwhile, the CFlow/EBstrong and CF+/EBstrong cells gave a much stronger EB-signal, indicating that these cells abnormally accumulate EB dye. Collectively, these findings indicate that the CF−/EB−, CFlow/EBstrong, and CF+/EBstrong populations are not normal forms, and that they appear to fail in the differentiation process of Entamoeba encystation. Based on the findings from the time course study, we propose a scenario for the cell differentiation process of Entamoeba encystation (**Figures 6A,B**).

The precise mechanism by which Entamoeba cells are stained with EB dye remains to be elucidated. The mechanism that damaged cells and disrupted blood vascular system or blood brain barrier are exclusively permeable to EB is widely recognized and is a basis for various mammalian biological studies, such as measurement of polymorphonuclear leukocyte infiltration (Griswold et al., 1989; Senaldi et al., 1994; Saunders et al., 2015). In Entamoeba study, however, it is unlikely because time course analysis of Entamoeba encystation indicates that EB<sup>+</sup> cells sequentially become EB<sup>−</sup> cells, indicating that EB<sup>+</sup> cells are neither dead nor dying. It is also unlikely that EB binds to serum albumin (SA) that is associated with Entamoeba cells because, despite cultivating Entamoeba cells in the presence of SA, the cells were washed several times with PBS before treating with EB, and SA is not internalized into Entamoeba cells. However, the presence of an SA-binding protein and/or an SA-like protein in Entamoeba cannot be ruled out. The most plausible mechanism is for EB to stain the membrane of Entamoeba cells, similarly to its action on human neutrophils (Hed et al., 1983). Molecular identification of the target to which EB binds and unraveling the molecular and cellular mechanisms underlying EB fluorescence changes during encystation will inform new topics on this differentiation process.

This study and that of Welter et al. (2017) for determining the effects of compounds on encystation by quantifying E. invadens cyst formation are essentially the same; therefore, the usefulness of flow cytometry to screen compounds that show a significant effect on Entamoeba encystation was confirmed by two distinct procedures. However, as is evident from the present study, an important issue to be addressed is whether each compound directly affects the encystation process, or has an indirect effect by damaging trophozoites. As a clue to solve this issue, the present study shows that compounds that halt Entamoeba cyst formation do not necessarily act on the mature cyst; for example, metronidazole and auranofin did not exert their effects on differentiated Entamoeba cells that possess similar characteristics to mature cysts. This result, however, does not provide an answer to the question of whether trophozoites themselves or differentiating cells that appeared during the course of encystation are targeted by the above compounds.

As a solution, in this study, we standardized a flow cytometry method to exclusively analyze proliferating E. histolytica trophozoites and combined it with that for analyzing cyst formation in E. invadens. The effectiveness of this combined system was demonstrated by providing very similar IC<sup>50</sup> values for lactacystin, metronidazole, and paromomycin for cyst formation and trophozoite proliferation. This finding indicates compounds that halt cyst formation by causing trophozoite dysfunction, and that the molecules targeted by them have fundamental roles in the Entamoeba life cycle; lactacystin and paromomycin impair the ubiquitin proteasome system and protein synthesis, respectively, and metronidazole induces damage of DNA or proteins (Liu and Weller, 1996; Makioka et al., 2002; Penuliar et al., 2015; Mi-ichi et al., 2016; Prokhorova et al., 2017). In addition, two compounds, which almost completely arrested both cyst formation and trophozoite proliferation, were identified using the combined system to screen the chemical library, the Pathogen Box of MMV (https://www.pathogenbox. org/). Importantly, one of these two compounds, auranofin was previously suggested to exert a cytotoxic effect on E. histolytica trophozoites by enhancing reactive oxygen-mediated cell killing via inhibition of thioredoxin reductase, an enzyme critical for preventing reactive oxygen generation (Debnath et al., 2012). However, the possibility still remains that cells undergoing differentiation from proliferative trophozoites are sensitive to compounds such as metronidazole and auranofin. Therefore, discrimination of these cells regarding drug sensitivity and elucidating the underlying molecular mechanisms present not only new topics in Entamoeba encystation research but also important issues to be addressed for the development of new drugs against amoebiasis.

Interestingly, the combined system-based screening of the chemical library also identified four compounds that show an effect on either cyst formation (one compound) or trophozoite proliferation (three compounds). This finding is intriguing because the molecules targeted by these compounds are suggested to be stage-specifically expressed to exert an essential role in distinct stages of the Entamoeba life cycle: the proliferative trophozoite or dormant cyst stages. In addition, 16 compounds were identified that almost completely abolished either cyst Mi-ichi et al. A Methodology for Entamoeba Biology

formation or trophozoite proliferation, but inhibited the other process only partially. This finding indicates that the target molecules of these 16 compounds also play important roles in the both trophozoite and cyst stages, but that their expression levels differ in each stage. However, the possibility cannot be ruled out that distinct levels of inhibition by the 20 compounds are attributed to the diversity of their target molecules between E. histolytica and E. invadens; for the above case of one compound that inhibited cyst formation, its target molecule may exist only in E. invadens, and not in E. histolytica. Hence, identifying and characterizing the molecules targeted by the screened compounds will provide new mechanistic insight for Entamoeba encystation, a cell differentiation process.

The effectiveness of the combined system of the E. invadens cyst formation and E. histolytica trophozoites proliferation assays is also demonstrated for applied perspective. This combined system identified twenty-two compounds that showed significant, negative effects on Entamoeba cyst formation, trophozoite proliferation, or both by screening 400 compounds with diverse scaffolds, the Pathogen Box from MMV (https:// www.pathogenbox.org/). All in all, auranofin (which is identified as E-H-05 in the Pathogen Box) is a promising lead for the development of new anti-amoebic drugs because of almost completely arresting the both biological processes. The argument is essentially the same with that of Debnath et al. (2012), in which the different assay system was exploited. In view of inhibitory profile observed in the present assay system, C-F-08 may also be a promising lead compound because of its similarity to auranofin. Among 14 compounds that almost completely halted cyst formation but only partially impaired trophozoite proliferation, iodoquinol (which is identified as B-E-06 in the Pathogen Box) was included. This is used as a luminal amoebicide and recommended to administer asymptomatic cyst carriers (Levine, 1991; Marie and Petri, 2013). Considering the similarity observed in their inhibitory profiles, the rest 13 compounds also be a potential lead for the development of new amoebiasis transmission-blocking drugs. However, determination of cells affected by the each compound is needed; for instance, trophozoite, differentiating cells appeared during encystation, and/or cyst. B-F-10 may be the same type with the above 14 compounds, but, as discussed in the previous paragraph, we should be aware that its target may exist only in E. invadens, and not in E. histolytica. Among other five compounds that exert more severe effect on trophozoite proliferation than cyst formation, nitazoxamide (which is identified as B-G-03 in the Pathogen Box) was included. Nitazoxamide is proposed to be a noncompetitive inhibitor of pyruvate:ferredoxin oxidoreductase, a critical enzyme in the main route for ATP supply in E. histolytica (Hoffman et al., 2007); therefore, it is plausible that nitazoxamide-susceptible cell requires pyruvate:ferredoxin oxidoreductase activity to fulfill high ATP demand, such as proliferating E. histolytica trophozoite. Furthermore, all 22 compounds screened exhibit bactericidal as well as parasiticidal activities (**Table 2**); therefore, the success of this new trial confirms that the Pathogen Box is an appropriate resource for the development of new drugs against a wide range of human pathogens. Importantly, it also confirms relevance of TABLE 2 | Summary of the 22 screened compounds: specificities against various pathogens.


*Compound IDs are taken from Medicine for Malaria Venture (MMV; https://www. pathogenbox.org/).* # *Information was obtained from MMV and references.*

the presented combined system, in which the threshold is set as >80% reduction by 10µM compound, for a primary screen of a chemical library to provide potential leads for the development of new anti-amoebic and amoebiasis transmission-blocking drugs.

Meanwhile, this system can also be applied to vaccine development; it can be used to screen antibodies that inhibit the Entamoeba trophozoite proliferation, cyst formation, or both, which will lead to the development of anti-amoebic and amoebiasis transmission-blocking vaccines. Attainment of the ultimate goals—developing new drugs and vaccines against amoebiasis—by exploiting the methodology demonstrated in this study is urgently needed in amoebiasis medicine because of limited available clinical options. Contribution of the demonstrated methodology to this important medical issue will become more substantial, if an automated, high-throughput system can be integrated with it, such as a 384-well system.

In conclusion, we present a methodology that can be widely adopted for Entamoeba study. Developing methodologies, such as that described here, is still required for various areas of the Entamoeba field.

# AUTHOR CONTRIBUTIONS

FM designed the experiments and analyzed the data. FM, YM, and VT performed the experiments. FM, YM, VT, and HY interpreted the data. FM and HY wrote the manuscript.

# ACKNOWLEDGMENTS

We would like to acknowledge MMV for their support and for designing and providing the Pathogen Box. We thank Ms. Shizuko Furukawa for technical assistance. The flow cytometric analysis was performed using a MACSQuant Analyzer at the Analytical Research Center for Experimental Sciences, Saga University. This study was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (26117719, 16H01365, and 16K19117) to FM, (17K08888) to YM and (16K08842) to HY, AMED-J-PRIDE (JP18fm0208025) to FM and HY, and by

# REFERENCES


Cooperative Research Grants of NEKKEN, 2016, 2017 to FM. This work was also supported by the Naito Foundation to FM and HY, respectively, by the Ohyama Health Foundation Inc. to FM, and by the Takeda Science Foundation to FM. and YM, respectively. We thank Jeremy Allen, Ph.D., from Edanz Group (www.edanzediting.com/ac) for editing a draft of this manuscript.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00250/full#supplementary-material

Supplementary Figure S1 | Secreening 400 compounds with diverse scaffolds from the Pathogen Box of MMV. (A–E) The effect(s) on *Entamoeba* cyst formation (upper) or trophozoite proliferation (lower) of 80 compounds from the Pathogen Box in plates (A–E), respectively. In the trophozoite proliferation assay, the total cell number counted is also indicated. Representative data are shown from two independent experiments. (F) Controls of cyst formation (upper) and trophozoite proliferation (lower) assays. Average percentages of mature cysts and average cell numbers of live trophozoites were 70.2 ± 0.21% and 6824 ± 324 (*n* = 3), respectively. These data were obtained from each assay and used as controls at 100% to calculate the inhibition rate of each compound tested.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Mi-ichi, Miyake, Tam and Yoshida. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Cellular Events of Multinucleated Giant Cells Formation During the Encystation of Entamoeba invadens

Deepak Krishnan and Sudip K. Ghosh\*

Department of Biotechnology, Indian Institute of Technology Kharagpur, Kharagpur, India

Entamoeba histolytica, the causative agent of amoebiasis, does not form cysts in vitro, so reptilian pathogen Entamoeba invadens is used as an Entamoeba encystation model. During the in vitro encystation of E. invadens, a few multinucleated giant cells (MGC) were also appeared in the culture along with cysts. Like the cyst, these MGC's were also formed in the multicellular aggregates found in the encystation culture. Time-lapse live cell imaging revealed that MGC's were the result of repeated cellular fusion with fusioncompetent trophozoites as a starting point. The early MGC were non-adherent, and they moved slowly and randomly in the media, but under confinement, MGC became highly motile and directionally persistent. The increased motility resulted in rapid cytoplasmic fissions, which indicated the possibility of continuous cell fusion and division taking place inside the compact multicellular aggregates. Following cell fusion, each nucleus obtained from the fusion-competent trophozoites gave rise to four nuclei with half genomic content. All the haploid nuclei in MGC later aggregated and fused to form a polyploid nucleus. These observations have important implications on Entamoeba biology as they point toward the possibility of E. invadens undergoing sexual or parasexual reproduction.

### Edited by:

Anjan Debnath, University of California, San Diego, United States

### Reviewed by:

Esther Orozco, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico Nancy Guillen, Centre National de la Recherche Scientifique (CNRS), France

> \*Correspondence: Sudip K. Ghosh sudip@bt.iitkgp.ac.in

Received: 24 April 2018 Accepted: 13 July 2018 Published: 31 July 2018

#### Citation:

Krishnan D and Ghosh SK (2018) Cellular Events of Multinucleated Giant Cells Formation During the Encystation of Entamoeba invadens. Front. Cell. Infect. Microbiol. 8:262. doi: 10.3389/fcimb.2018.00262 Keywords: Entamoeba, haploidization, cell fusion, cytofission, nuclear fusion

# INTRODUCTION

The Entamoeba species, like most parasitic microbes, is considered to be asexual as the mechanism enabled the parasites to generate clonal populations that have well adapted to host. Sexual or non-meiotic parasexual recombination could disrupt the allelic combinations required for the success of the parasite, and thus a clonal population structure was proposed initially for parasitic protozoa (Tibayrenc et al., 1990). Harmful mutations are supposed to accumulate in asexually reproducing organisms, and the sexual pathway is necessary to prevent this process known as Muller's ratchet. A clonal population from asexual reproduction reduces the genetic variability in the population required for adaptation and evolution. So the present understanding of sexual nature of parasitic protozoa is that they produce a clonal population by asexual pathway but retains sexual reproduction (Heitman, 2006). When met with environmental stress like antimicrobial therapy or host immune reaction sexual pathway produces a diverse progeny from which a new favorable trait can be selected and get fixed in the population through subsequent clonal lineages. Such strategy was reported in fungal pathogens where they show a clonal population structure but have evolved various sexual or parasexual mechanisms, and even rare sexual events were observed to change their pathogenicity and lifestyle (Ene and Bennett, 2014). Alteration of pathogenicity by sex was observed in Toxoplasma with the sexual process giving rise to hyper-virulent strains from avirulent parents (Grigg et al., 2001). The reason behind Vancouver Island Cryptococcus gattii outbreak was found to be a highly virulent strain produced by a cryptic unisexual mating (Fraser et al., 2005). These observations show that even rare events of sex could alter the lifestyle of a parasite and create public health problems. Thus, the understanding of the sexual pathway is of enormous medical importance especially in vaccine and drug development.

Observation of sexual or parasexual reproduction is difficult in most protozoan parasites as it was infrequent or occurred under unknown conditions, or it was not recognized as a sexual mechanism (Birky, 2005). But lately such mechanisms have been reported in important protozoan parasites like Giardia (Poxleitner et al., 2008), Leishmania (Akopyants et al., 2009), and Trypanosoma (Peacock et al., 2014). Generation of hybrids, detection of meiotic genes and population genetics (Weedall and Hall, 2015) have been used so far to find the presence of sexual reproduction in protozoa. Genome data analysis showed that Entamoeba histolytica and its reptilian counterpart Entamoeba invadens have most of the meiotic genes required for sexual/parasexual reproduction (Ramesh et al., 2005; Ehrenkaufer et al., 2013). Gene conversion by homologous recombination in Gal/GalNac lectin genes, which could help the parasite in immune evasion, has also been reported in E. histolytica (Weedall et al., 2011). Isolated parasites from the intestine and liver abscess of the same patient showed genetic variation indicating the presence of genomic reorganization and formation of parasites with invasive characteristics (Ali et al., 2008). Also, multi-locus sequence typing of E. histolytica isolates from the same geographic origin showed very high genomic diversity indicating DNA recombination (Gilchrist et al., 2012). All these observations indicate Entamoeba undergoes sexual or parasexual reproduction at some stage but how and when it takes place is not yet understood. Meiotic genes were found to be up-regulated (Ehrenkaufer et al., 2013) and homologous recombination was observed to be enhanced (Singh et al., 2013) during the stage conversion of reptilian parasite Entamoeba invadens which is used as a model for studying encystation process as E. histolytica does not form a cyst in vitro. Starvation triggered the sexual pathway in many eukaryotes like yeast and Dictyostelium, and E. invadens in vitro encystation is also a response to starvation. Cell fusion, nuclear fusion, and ploidy transitions are the hallmarks of sexual or parasexual events and so using microscopy, the encystation process of E. invadens was investigated for the presence of these events.

# MATERIALS AND METHODS

# Cells and Reagents

Entamoeba invadens strain IP-1 was maintained in TYI-S-33 medium containing 10% adult bovine serum (HiMedia) and 3% Diamond vitamin mix at 25◦C. DAPI, Propidium iodide, Hoechst 33342, Fluorescein diacetate, and calcofluor white were purchased from Sigma-Aldrich. Alexafluor 488 conjugated phalloidin was purchased from Molecular Probes, Invitrogen, USA.

# Encystation

To prepare the encystation induction (LG 47) medium which contained 47 % of nutrients, TYI medium without glucose was prepared and diluted to 2.12 times and completed with 5% heat inactivated adult bovine serum, 1.5% vitamin mix and antibiotics, penicillin and streptomycin. Mid log phase trophozoites were chilled on ice for 10 min to detach the cells from the culture tube wall and harvested by centrifugation at 500 × g for 5 min at 4◦C. The cells were washed multiple times with LG media and 5 × 10<sup>5</sup> trophozoites per ml were counted and transferred into encystation induction medium (LG) and incubated at 25◦C. These cultures were observed for the presence of giant cells and used for further experiments.

# Cell Staining

Cells were fixed with 4% (w/v) paraformaldehyde in PBS for 10 min and then permeabilized in 0.1% (v/v) Triton X-100 in PBS for 5 min. DAPI and PI were used to stain the nucleus. Chitin wall was stained with calcofluor white (Arroyo-Begovich et al., 1980). For actin localization permeabilized cells were blocked with 2% (w/v) BSA and stained with Alexafluor 488 conjugated phalloidin.

# Microscopy

Olympus FV1000 confocal microscope and Olympus IX 51 fluorescence microscope with camera attachment and photoediting software (Image Pro Discovery) were used for imaging. From the images, cell cross sectional area and number of nuclei were measured using ImageJ software (NIH).

# Cell Viability Determination

Cells from encystation culture were pelleted and resuspended in PBS containing 10µg/ml Fluorescein diacetate. After incubating at room temperature for 5 min cells were washed with PBS. The cells were then observed under a fluorescence microscope for the fluorescence produced by live cells.

# Live Cell Imaging

Live cell imaging was done using Olympus 1X51 inverted light microscope. Hoechst 33342 was used to stain nuclei for live cell imaging. Because of the high cell density and cell aggregates of encystation culture which obscured the live cell imaging, the cell aggregates were dispersed by adding galactose and the culture was diluted and then taken in tissue culture plates for imaging. The raw images were then processed using ImageJ software (http:// rsb.info.nih.gov/ij/)

# Analysis of Cell Motility

Time-lapse video microscopy was performed using Olympus 1X 51 inverted microscope with camera attachment and photoediting software (Image Pro Discovery). From the time-lapse recording, the trajectories of the cells were visualized using the ImageJ software (National Institutes of Health, Bethesda, USA) with plugins Manual Tracking and MTrack J. The tracking data was then analyzed using Chemotaxis and Migration Tool (ibidi) to obtain velocity and directness.

# Measurement of Nuclear DNA Content

Encystation cultures containing MGC and cyst were chilled to detach the cells, pelleted by centrifugation at 500 g, washed with ice-cold PBS, and then fixed in 4% (w/v) paraformaldehyde in PBS for 10 min. It was then permeabilized in 0.2% (v/v) Triton X-100 in PBS for 5 min and blocked with 2% (w/v) BSA in PBS. After treating with 20µg/ml RNase A, the nuclei were stained with 10µg/ml Propidium Iodide. Olympus FV1000 confocal microscope was then used to take image of the cell. From this image fluorescence intensity of trophozoite and MGC nuclei were measured using ImageJ (NIH). The results were represented using Box whisker plot. Boxes indicate 25–75 percentiles; the line 50th percentile and the solid square indicate the mean. The whiskers mark the 10th and 90th percentile.

# Statistical Analysis

Quantitative data are shown as the mean ± standard deviation (SD). The significance of the experimental data was calculated using unpaired, two-tailed Student's t-test. Differences between groups were considered statistically significant if p < 0.05.

# RESULTS

# Multinucleated Giant Cells Are Formed During Encystation by Cell Fusion

Encystation model organism Entamoeba invadens undergoes stage conversion to produce chitin walled cyst when subjected to starvation and osmotic stress (Sanchez et al., 1994). It was observed that during encystation, a few giant cells also appeared in the culture (**Figures 1A,B** and Movie S1) along with cyst, and nuclear staining showed these cells were multinucleated (**Figure 1C**). The MGC were observed at a very low frequency (1 in 10<sup>4</sup> cells), and size of observed giant cells showed considerable variation. Entamoeba invadens cysts were formed only inside the cell aggregates found in the encystation culture (**Figure 1D**) and MGC were usually observed emerging from or associated with these cell aggregates (**Figure 1E**). The cell aggregates were galactose ligand-mediated (Cho and Eichinger, 1998) and the addition of galactose to the LG medium disrupted of the cell aggregates and released these multinucleated giant cells (MGCs) into the medium. To study their behavior released MGC were observed using time-lapse live cell imaging. The MGC obtained from early encystation cultures (24–48 h) did not show any cell adhesion and moved randomly in the media. When these early MGC came into contact with each other, their membrane fused very rapidly at the point of contact, and they became one cell. When kept together in the media MGC fused continuously to produce bigger and bigger cells (**Figure 2A** and Movie S2). The cell fusion continued even after transferring the MGC to growth media (TYI-S-33). The cell fusion was very fast in early MGC (24–36 h) with the cells fusing instantaneously on membrane contact. Larger MGC from 48 to 72 h of encystation took longer time, and MGC obtained after 96-h culture did not fuse. To find how the cell size changes with time, MGCs were obtained from encystation cultures at different time points and their cross-sectional area, which is an approximate measure of cell size, was compared to that of trophozoites (**Figure 2B**). MGC became observable by 24th hour, and these early giant cells were nearly twice the size of trophozoites, and by 48th hour their size again increased two-fold by fusing among themselves. MGC obtained from 72 to 96 h showed a broad distribution in size with a few of them reaching sizes 10 times to that of trophozoites, but MGC from week old cultures were comparatively smaller. The doubling of cell size in the early hours and the ability to undergo continuous fusion implies that the starting point of MGC could be the trophozoites which acquired fusion competency. The MGC's were never observed to interact or fuse with trophozoites indicating cell surfaces changes like the expression of fusion proteins. Cell signaling required cyst formation take place inside the cell aggregates (Coppi et al., 2002) and as MGC were often seen associated with these aggregates, the formation of fusion- competent trophozoites and the initial cell fusions may also be occurring inside them. Entamoeba genome was searched for all known proteins involved in cell fusion, but none could be detected yet. Fluorescein diacetate (FDA) hydrolysis assays showed that most MGC, especially those remained inside cell aggregates retained their viability even in week-old cultures (**Figure 2C**) though a few MGC released into the media sometimes underwent vacuolation and cell lysis (**Figure 2D**). Multinucleated cells are regularly found in Entamoeba histolytica culture (Supplementary Figure 1), but time-lapse live-cell imaging did not show the presence of cell fusion. These cells were probably formed due to the delinking of nuclear division and cytokinesis as cell cycle checkpoints are absent in Entamoeba histolytica (Das and Lohia, 2002).

# Multinucleated Giant Cells Undergo Cytofission in Confinement

The MGC released from the aggregates moved very slowly in the culture media (Movie S3) but when compressed between slide and cover slip the MGC became highly motile (**Figure 3A**). To analyze the changes in cell motility, its velocity and directness was measured (Zengel et al., 2011). Directness is the ratio of the displacement and the total distance traversed by the cell and is a measure of directional persistence (lower values indicting random motility and values closer to one indicating directional motility). Velocity of the cell in the liquid media was found to be 0.15 ± 0.05 µm/s, and the directness value of 0.26 ± 0.10 indicting random motility (**Figure 3B**). In confinement velocity of the MGC increased to 0.66 ± 0.09 µm/s and the motility became highly directional as shown by the directness value of 0.77 ± 0.19. Staining the early MGC with Alexafluor 488 phalloidin showed no cell adhesion structures but the presence of actin under the plasma membrane (**Figure 3C**). In the absence of cell–substrate adhesions, early MGC could be using chimneying motility, in which the cell generate traction using the opposing surfaces to crawl in a 3D environment and that required the actomyosin cortex (Malawista et al., 2000; Hawkins et al., 2009; Liu et al., 2015).The cell aggregates in the encystation culture could be providing the confinement as MGC were associated with them.

The consequence of such increased motility was the rapid cytoplasmic fission of MGC due to movement toward opposite

directions and tearing itself into daughter cells (**Figure 4A** and Movie S4). The resulting daughter cells were also capable of further cytofission (**Figure 4B** and Movie S5), and these divisions often continued until all the daughter cells reached trophozoite size (Movie S6). When the distribution of nuclei into daughter cells during cytofission was observed using cellpermeable nuclear stain Hoechst 33342, it was found that there is no correlation between daughter cell size and the number of nuclei per cell (**Figure 4C** and Movie S7). Cytofission started at random sites, and nuclei were unevenly distributed, so that daughter cells with variable numbers of nuclei were formed. The daughter cells of cytofission sometimes remained connected through the cytoplasmic bridge (**Figure 5A**). It was observed that the nearby trophozoites converged on the dividing area probably helping the cytofission by severing cytoplasmic bridge between two daughter cells (**Figure 5B** and Movie S8, S9) as reported to occur during E. invadens cell division (Biron et al., 2001).

Myosin II inhibitor, 2,3-butanedione monoxime (BDM) was observed to inhibit the encystation when added to the encystation culture at a concentration 20 mM (**Figure 6A**) but it greatly increased the number of MGC to 1 in 10<sup>2</sup> (**Figure 6B**). In growth media, BDM treatment did not cause any multinucleation, so the development of giant cells was not due to cytokinesis failure and it also did not inhibit the cell aggregation step during encystation. As actomyosin contractile ring is responsible for cleaving the dividing cell into two daughter cells, BDM might have blocked the cytofission without stopping the cell fusion, thus the number of MGC was found to be increased. These observations indicate the possibility of continuous fusion and cytofission occurring inside the cell aggregates and the MGC's observed in older encystation cultures could be the final product of such cyclic fusion and cytofission. The galactose ligand mediated cell aggregate is required only in the early hours of encystation in encystation medium (LG) to provide the cell signaling (Turner and Eichinger, 2007). During the in vitro encystation most of the cells in the multicellular aggregates were converted to cysts. Once the cysts were fully formed, aggregates lost their compact structure and release the cyst and MGC, if there is any into the media.

# MGC Nuclei Undergo Haploidization After Cell Fusion

To find the changes in MGC nuclei following cell fusion, their number, size and genomic content were measured and compared with that of trophozoites. Since each MGC were formed by the continuous fusion of the trophozoites which gained fusion competency, the number of nuclei must be proportional to cell size. When the increase in MGC size relative to trophozoite estimated from the cross-sectional area was plotted against the

Fluorescein (b) indicating their viability. (D) Very large MGC formed by continuous fusion in the media sometimes underwent vacuolation and cell lysis. Scale bars: 50µm.

FIGURE 3 | Motility of MGC in medium and under confinement. (A) Migration tracks of MGC in the medium and under confinement. (B) Velocity and Directness (ratio of distance between the start and the end point and the total distance traversed by the cell) of the MGC motility. Velocity in the medium was 0.15 ± 0.05 µm/s and under confinement it increased to 0.66 ± 0.09 µm/s. The MGC moved randomly in the media as shown by the directness value 0.26 ± 0.10 and under confinement the migration became directionally persistent with the directness value of 0.77 ± 0.19. (C) Alexafluor 488-phalloidin staining of early motile MGC showed actin was present in the cortical region of the MGC. Scale bars: 5µm.

number of nuclei per cell (**Figure 7A**), their relationship could be explained as

Number of nuclei per MGC ≈ 4 × fold increase in MGC size (relative to size of trophozoites)

This meant that for each fusion competent trophozoite participated in the formation of MGC; there was a fourfold increase in the number of nuclei. The early MGC were usually found to contain nuclei of different sizes, large nuclei with size similar to that of trophozoites, and clusters of small nuclei (**Figure 7B** and Movie S10). To find the genomic content of these nuclei, MGC were isolated from encystation culture at different time points, their nuclei were stained with Propidium Iodide (PI), and images were taken using confocal microscope. From the image, the fluorescence intensity of each nucleus which is a measure of genomic DNA content was calculated. E. invadens trophozoites exhibited a broad distribution of DNA content within a population due to lack of cell cycle control (Byers and Eichinger, 2005; Mukherjee et al., 2009). The distribution of nuclear DNA content of the larger nuclei observed in 24th hour MGC was similar to that of trophozoites, but in smaller nuclei, it reduced by half (**Figure 7C**). The large nuclei were only observed in MGC from 24th hour and small nuclei were prevalent in MGC obtained from 48 to 72 h (**Figure 7D**), so nuclear division in MGC probably started after the initial cell fusions between fusion competent trophozoites.

# Nuclear Aggregation and Fusion Occur in Late MGC

MGC taken from encystation cultures after 72–96 h of incubation were non-motile, relatively smaller, and adherent and nuclear

FIGURE 4 | Increase in velocity under confinement led to cytofission in MGC. (A) MGC underwent cytoplasmic fission in confinement by moving in opposite direction and tearing the cell in to two.(B) The daughter cells were also capable of continuous cytofission. See also Movies S4–S6 (C) Random nuclear distribution during cytofission process shown by Hoechst 33348 staining. (Scale bars: 50µm, time in minutes). See also Movie S7.

FIGURE 5 | MGC cytofission was assisted by trophozoites. (A) When the membrane bridge linking the dividing cells (arrow mark) was not severed, the daughter cells came back and re-joined. Scale bars: 50µm (B) Trophozoites converged on the dividing area (a1,b1) and severed membrane bridge between two daughter cells (a2,b2). See also Movies S8 and S9. Scale bars: 20µm.

staining showed all the nuclei had aggregated to one point (**Figure 8A** and Movie S11). As MGC became non-motile, the cortical actin reorganized into ring-like structures on the cell surface (**Figure 8Ba**). These structures were similar to podosome rosettes involved in adhesion and composed of individual podosomes arranged in a ring like shape, similar to those found on the ventral side of adherent trophozoites (**Figure 8Bb**). In week old encystation contained highly condensed clusters of nuclei (**Figures 8Ca,b**) and a few cells contained a single large nucleus (**Figures 8Cc,d**) which was found to be polyploid on nuclear DNA measurement (**Figure 8D**). This shows the presence of interaction among nuclei in older MGC leading to nuclear fusion (**Figure 8E**). These interactions start with nuclei clustering to one location (**Figures 8Ea,a**′ **,b,b**′ ) and later these aggregates became tightly packed structure with individual nuclei becoming indistinguishable (**Figures 8Ec,c**′ ). Inside these clusters, a single large nucleus was sometimes observed (**Figures 8Ed,d**′ ) possibly resulting from nuclear fusion but it is unclear how many nuclei participated in fusion to produce the giant cell with a polyploid nucleus (**Figures 8Ee,e**′ ).

# DISCUSSION

Life cycle of Entamoeba histolytica and Entamoeba invadens contains two stages, uninucleate motile trophozoites, and tetranucleate dormant cysts. Microscopic analysis of the in vitro encystation culture of E. invadens revealed rare encystation specific multinucleated giant cells formed by cell fusion. Multinucleated cells were also found in the stationary phase of E. histolytica culture, but it is reported to be due to cytokinesis failure, not cell fusion (Das and Lohia, 2002).The starting point for MGC was the conversion of a few trophozoites into fusion competent cells which may have involved the expression of fusion proteins since MGC never fused with normal trophozoites. It has been already reported that the encystation is initiated by the cell signaling taking place inside the cellular aggregates

(Coppi et al., 2002). Similarly signaling within the aggregates may play some role in the formation of MGC. The high population density of the cell aggregate may also be required to increase the chance of cell fusion (Woznica et al., 2017).The MGC released from cell aggregates moved slowly and randomly in the media, but under confined conditions, they became highly motile and directionally persistent. The increased motility led to larger MGC tearing itself into smaller cells by cytoplasmic fission. Inhibition the cytofission by BDM greatly increased the number of MGC in the encystation media indicating sequential fusion and cytofission may be taking place inside the cell aggregates. Such cyclic fusion and cytofission could be the reason for variation observed in the number and size of MGC during encystation experiments. Formation of multinucleated cells that underwent cell fusion and cytofission was also observed during early stages of macrocyst formation which is the sexual stage of social amoeba Dictyostelium discoideum (Ishida et al., 2005) and also during the parasexual stage of amoeba Cochliopodium (Tekle et al., 2014). The exact relevance of such a random cell fusion and cytofission process is unclear, but it could help to share nutrients among cells during starvation and also gather different traits from multiple cells so that multinucleated cell which accumulated the best combination could survive during stress.

During the encystation, uninucleate trophozoite undergoes nuclear division to form tetranucleate cyst (Supplementary Figure 2). Correlation between increase in MGC size relative to trophozoites and number of nuclei it contained showed that for each trophozoite participated in the cell fusion to form MGC there was a fourfold increase in the number of nuclei indicating MGC also underwent similar nuclear division. But the final ploidy was different as the cyst nucleus contained one-fourth DNA (Lohia, 2003; Supplementary Figure 2) and MGC nuclei contained half genomic DNA compared to trophozoites. Ploidy changes can occur through meiotic or non-meiotic mechanisms

was similar to trophozoite nucleus but the smaller nuclei were haploid relative to trophozoites. Boxes indicate 25–75 percentiles; the line 50th percentile and the solid square indicate the mean. The whiskers mark the 10th and 90th percentile (NS: Not significant). (D) Representative images of MGC from different hours of encystation showing the changes in number and size of nuclei after staining with PI. MGC from 24 h of encystation contained larger nuclei along with haploid nuclei (Upper). MGC took from 48 h of encystation show mainly haploid nuclei (Middle). MGC from 48 to 72 h of encystation were formed by the fusion of smaller MGC (Lower). These cells also contained only haploid nuclei (See also Movie S11). Scale bars: 10 µm.

like chromosome loss (Bennett and Johnson, 2003; Bennett et al., 2014). Meiotic genes were also reported to be expressed during encystation (Ehrenkaufer et al., 2013); but it is not yet clear whether the nuclear division in MGC is meiotic or not. Expression of meiotic genes also took place during the parasexual reproduction of Giardia (Carpenter et al., 2012) and Candida albicans (Forche et al., 2008). So it is important to determine the exact nature of ploidy transition in MGC. Nonetheless, fusion competency, fast motility, and presence of haploid nuclei make MGC the gamete equivalent of Entamoeba. In the late, nonmotile MGC, the nuclei aggregated and underwent multiple nuclear fusions to form a polyploid nucleus. Both E. invadens and E. histolytica have been shown to alter ploidy depending on the growth conditions and different stages of life cycles (Mukherjee et al., 2008). It is yet to be determined whether the polyploid giant cells are returning to the trophozoite stage with respect to cell size and DNA content with time or not. Cell fusion or fertilization is the most important step in the sexual/parasexual pathways as it allows genetic exchange within the population and recombines beneficial traits from different lineages. The observation of cell fusion at any time during life cycle alone provides indirect evidence for the presence of sexual/parasexual pathway (Lahr et al., 2011). Even without any genomic recombination such cell fusion events (agamic cell fusion) can be important because through the increased nutrient reserves and resulting polyploidy and hybrid vigor, it can increase cell survival in adverse conditions (Comai, 2005; Goodenough and Heitman, 2014). The formation of multinucleated and polyploid cells by cell fusion was also reported in many cancers, and the resulting genome reorganizations facilitated metastasis

FIGURE 8 | Nuclear aggregation and fusion in MGC. (A) PI staining shows nuclear aggregation observed in MGC taken from encystation cultures after 72–96 h of incubation. See also Movie S12. (B) Staining the non-motile MGC with Alexafluor 488 phalloidin showed the presence of actin structures on the cell surface (a) which could be podosome rosettes found on the ventral side of adherent E. invadens trophozoites (b). (C) In week old cultures most MGC contained closely packed nuclei (a,b) or a single polyploid nucleus (c,d) were found in encystation cultures after week long incubation. (D) Comparison of the nuclear content of trophozoite (N = 355), MGC (N = 563) and the giant cell with single nucleus (N = 20). (E) Stages of nuclear interactions leading to nuclear fusion. After 72 h nuclei started aggregating (a–b') to form a compact nuclear cluster (c,c') in which formation of a single large nuclei (d,d', arrow mark) was observed. The nuclei which did not participate in fusion then disappeared to produce a giant cell with a single large nucleus (e,e'). Scale bars: 10µm.

and drug resistance in cancer (Weihua et al., 2011; Niu et al., 2016). Similarly, the MGC also possesses the potential to induce phenotypic variations in Entamoeba though this hypothesis is yet to be tested.

The three main developmental pathways of amoebozoans: sporulation, encystation and sexual macrocyst pathway were controlled by the environmental conditions (O'Day and Keszei, 2012). It could be possible that like encystation, MGC pathway is a stress response mechanism and activated by the nutrient and osmotic stress caused by the encystation media. Sexual/parasexual recombination causes mixing and reshuffling of genes and creates genetic variations in the population, but such events may be detrimental to a parasite as it disrupts the allelic combinations required for survival in the host. Most parasites have thus limited the sexual/parasexual reproduction to during stress or dispersal to new hosts (Weedall and Hall, 2015). While encystation helps to survive adverse conditions by forming a resistant cyst, MGC pathway can introduce hybrid fitness and beneficial genomic changes and ensure the survival of Entamoeba in a changing environment. But so far no other stress conditions tested like oxidative or heat shock, induced MGC formation. The in vitro encystation is conducted using an axenic culture but the intestinal bacteria have been shown to influence characteristics like DNA content and virulence (Bracha and Mirelman, 1984; Mukherjee et al., 2008) and it may also have an influence on the sexual/parasexual pathway. For example in the Choanoflagellate Salpingoeca rosetta, sexual reproduction is enhanced by bacteria Vibrio fischeri by inducing cell aggregates in which the cells underwent extensive fusion (Woznica et al., 2017).

Like cysts, MGC were also formed inside cell aggregates, so the initial cell signaling associated with encystation or the expression of meiotic genes may have caused a few cells to gain fusion competency and start the MGC pathway. Encystation is

# REFERENCES


the ancestral survival mechanism of all amoebozoa (Kawabe et al., 2009) and the cysts themselves can be asexual or sexual where a zygote formed by cell fusion undergoes encystation (Schaap and Schilde, 2018). Copromyxa protea and Sappinia diploidea form double walled sexual cyst by fusion of two amoebae (Walochnik et al., 2010; Brown et al., 2011) but no such wall formation was observed in MGC. Sexual reproductions and encystations were probably present in the life cycle of last eukaryotic common ancestor (LECA).Cell fusion and meiosis during encystation might have helped them to survive as dormant cysts in adverse environmental conditions by providing genetic redundancy and recombinational DNA repair, and that may be associated with the evolution of sex (Cavalier-Smith, 2010). Entamoeba is regarded as a primitive eukaryote (Bakker-Grunwald and Wöstmann, 1993) and even if the MGC cell fusion is agamic, the study of its cellular events could be useful in understanding the origin and evolution of sexual reproduction.

# AUTHOR CONTRIBUTIONS

DK and SG: designed research; DK: performed research; DK and SG: wrote the paper.

# ACKNOWLEDGMENTS

DK is recipient of Senior Research Fellowship from Indian Council of Medical Research, India. Authors thank FIST, DST, Government of India for confocal facility.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00262/full#supplementary-material


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Krishnan and Ghosh. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Utilization of Different Omic Approaches to Unravel Stress Response Mechanisms in the Parasite *Entamoeba histolytica*

Shruti Nagaraja and Serge Ankri\*

Department of Molecular Microbiology, Ruth and Bruce Rappaport Faculty of Medicine, Technion, Haifa, Israel

#### *Edited by:*

Mario Alberto Rodriguez, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional (CINVESTAV-IPN), Mexico

#### *Reviewed by:*

César López-Camarillo, Universidad Autónoma de la Ciudad de México, Mexico Adriane Regina Todeschini, Universidade Federal do Rio de Janeiro, Brazil Siddhartha Das, University of Texas at El Paso, United States Humberto Lanz-Mendoza, Instituto Nacional de Salud Pública, Mexico

#### *\*Correspondence:*

Serge Ankri sankri@technion.ac.il

*Received:* 06 November 2017 *Accepted:* 16 January 2018 *Published:* 08 February 2018

#### *Citation:*

Nagaraja S and Ankri S (2018) Utilization of Different Omic Approaches to Unravel Stress Response Mechanisms in the Parasite Entamoeba histolytica. Front. Cell. Infect. Microbiol. 8:19. doi: 10.3389/fcimb.2018.00019 During its life cycle, the unicellular parasite Entamoeba histolytica is challenged by a wide variety of environmental stresses, such as fluctuation in glucose concentration, changes in gut microbiota composition, and the release of oxidative and nitrosative species from neutrophils and macrophages. The best mode of survival for this parasite is to continuously adapt itself to the dynamic environment of the host. Our ability to study the stress-induced responses and adaptive mechanisms of this parasite has been transformed through the development of genomics, proteomics or metabolomics (omics sciences). These studies provide insights into different facets of the parasite's behavior in the host. However, there is a dire need for multi-omics data integration to better understand its pathogenic nature, ultimately paving the way to identify new chemotherapeutic targets against amebiasis. This review provides an integration of the most relevant omics information on the mechanisms that are used by E. histolytica to resist environmental stresses.

Keywords: *Entamoeba histolytica*, omics, oxidative stress, nitrosative stress, iron starvation, glucose starvation, microbiota, virulence

# INTRODUCTION

Amebiasis, caused by the eukaryotic parasite Entamoeba histolytica, is an enormous global medical problem that still exists due to poor sanitary and unhygienic conditions. According to the World Health Organization, 50 million people in India, Southeast Asia, Africa, and Latin America suffer from amebic dysentery and liver abscesses, and amebiasis causes the death of at least 100,000 individuals each year. In 90% of the infected patients, E. histolytica trophozoites normally inhabit the colon and spend their time in the host as a non-pathogenic commensal. However, the reasons why these trophozoites become virulent and invasive are unknown. Anti-amoebic drugs are the preferred choice due to the unavailability of vaccines. Based on their site of action, two categories of anti-amoebic drugs are used, namely, luminal amebicides (diloxanide furoate, and Iodoquinol) (Marie and Petri, 2013) and tissue amebicides (metronidazole) (Salles et al., 2007; Tazreiter et al., 2008; Marie and Petri, 2013) and potential resistance of the parasite to metronidazole is a real concern. Moreover, metronidazole is not effective in eliminating cysts inside the lumen and thus a combination of luminal and tissue amebicides is generally recommended (Marie and Petri, 2013). Recently, auranofin has been identified as a potent drug that targets redox enzymes in the parasite, eventually leading to oxidative stress in the parasite and it has found to be more effective than metronidazole (Debnath et al., 2012). Nevertheless, it is still a need of the hour to identify more potential drug targets to treat amebiasis. E. histolytica is challenged in the host environment due to fluctuations in partial pressure of oxygen, changes in glucose concentration and changes in the composition of the microbiota. The activation of innate immune responses against the parasite leads to the production of reactive oxygen species (ROS), nitric oxide (NO) by macrophages, complement activation and phagocytosis, and heat shock responses (Mortimer and Chadee, 2010; Moonah et al., 2013; Nakada-Tsukui and Nozaki, 2016; Olivos-Garcia et al., 2016). The parasite must be capable of adapting to the demand of surrounding environment in order to survive. This adaptive response of the parasite provides a shield against the host response as well as aids in their survival (**Figure 1**). In eukaryotic cells, the general stress response mechanism is a tightly orchestrated process. The first step involves the role of a stress-sensor proteins (heat shock proteins, nutrient sensing proteins, antioxidant proteins and also chromatin—proteins) to relay the message to the cells to adapt to stress (De Nadal et al., 2011; Walter and Ron, 2011; Santi-Rocca et al., 2012; Smith and Workman, 2012; Shahi et al., 2016a; **Figure 2**). The transfer of this stress signal to downstream proteins leads to a signal transduction cascade. This cascade begins with the phosphorylation of effector proteins [eIF2 kinases and Mitogen Activating Protein Kinases (MAPK)], and these proteins are known to play a role during stress. This eventually helps the cells to adapt to the stress by either attenuating translation (by phosphorylation of serine residue of the α subunit of eIF2 leading to its inactivation) (Jiang and Wek, 2005; Hendrick et al., 2016; Sharma et al., 2016) or through the

FIGURE 1 | Strategies used by E. histolytica when challenged with different stresses. E. histolytica faces threat in a number of ways: 1. Following colonic invasion, trophozoites penetrate the intestinal epithelial layer of the host where they are challenged by immune cell during invasion. Exposure to ROS (Rastew et al., 2012) and RNS (Kolios et al., 2004) released by macrophages and other immune cells eventually triggers the parasite defense mechanisms as shown in the box. 2. The amoeba is also threatened by low glucose concentration in the colon and it survives this condition by inhibiting glycolysis and by degrading stored glycogen and converting it to free glucose. Moreover, low glucose concentration triggers the parasite to become more virulent by upregulating Gal/GalNAc lectins (Baumel-Alterzon et al., 2013). 3. Changes occurring during the absence of L-cysteine, which is an important thiol required for antioxidant activity in the parasite. During cysteine starvation, there is an increase of phospholipids and other metabolites such as S-methyl cytosine, S-adenosine methionine etc. (Jeelani et al., 2014). 4. The absence of iron also increases virulence of the parasite by increasing the expression of CP-A5, CP-A7, and leads to the upregulation of transport proteins to scavenge iron from other external sources (Hernandez-Cuevas et al., 2014). SOD, superoxide dismutase; CLS, cyst like structure; NAOD, N-acetyl ornithine deactylase; Gal/GalNAc, Galactose/N-acetylgalactosamine binding lectin; CP, cysteine proteinases; DPD, dihydropyramidine dehydrogenase; AIG, Androgen Inducible Gene; ABC transporters; ATP, binding cassette transporters.

modulation of gene expression and metabolism (Vonlaufen et al., 2008; Darling and Cook, 2014). While conventional molecular techniques provided an outline, the gradual development and utilization of "omics" technologies and bioinformatics to study E. histolytica open new avenues to understand the complexity of its behavior under different conditions. For example, the field of DNA microarrays and proteomics have revolutionized our manner to assess the virulence of the parasite and its ability to cope with various stresses (Gilchrist et al., 2006, 2010; López-Camarillo et al., 2009), study the expression of different genes in the parasite exposed to UV radiation (Gilchrist et al., 2006; Weber et al., 2009), and assess the effects of metronidazole as a chemotherapeutic agent (Tazreiter et al., 2008). Moreover, there are other studies investigating the proteome of cell surface proteins and the excretory-secretory protein system of the parasite that may help in understanding its pathogenicity (Biller et al., 2014; Ujang et al., 2016).With the help of different (transcriptomic, genomic, and metabolomics) omics analysis, it has now become possible to study the responses of the parasite to various stresses during host invasion and these studies can provide several critical pieces of evidence as to how the parasite manages to survive inside the host (**Table 1**). Thus, it is essential to review all the data in order to characterize the mechanisms essential for stress response and identify potential drug targets against this parasite. This article presents an overview of recent advances that have been made by using various omics approaches to investigate stress response in E. histolytica, with focus on oxidative stress (OS) and nitrosative stress (NS).

# OXIDATIVE STRESS—WHAT TRIGGERS IT?

The key players that are involved in OS are a variety of ROS. They are capable of damaging essential biomolecules in the cell such as DNA, proteins, lipids and they lead to the fragmentation of the endoplasmic reticulum (ER) due to the accumulation of misfolded proteins (Imlay, 2003; Apel and Hirt, 2004; Pineda and Perdomo, 2017). In the large intestine, the invading E. histolytica trophozoites encounter OS. The sources of these stresses are fluctuations in oxygen tension in the intestinal lumen and the generation of ROS by cells of the immune system. Hydrogen peroxide (H2O2) is capable of damaging the proteins by its interaction with thiol groups, which are present in the cysteine side chains as well as with metal cofactors. Once formed, ROS leads to the oxidative damage of proteins thereby affecting their structure and functional properties (Shacter, 2000; Wu et al., 2006; Aiken et al., 2011). OS resistance contributes to the pathogenic potential of E. histolytica (Rastew et al., 2012).

# How Did Omics Help Us Understand Oxidative Stress Response in *E. histolytica*?

Analysis of the parasite's transcriptome in response to OS showed that this parasite copes with this stress by a complex modulation of a broad set of genes encoding proteins that are mainly involved in protein folding [Heat shock proteins (Hsps)], amino acid catabolism (MGL-1), signaling/regulatory pathways, and also in pathways involved in repair during DNA damage



MGL-1, methyl-gamma-lyase-1; ATP, Adenosine-triphosphate; ER, Endoplasmic reticulum; DPD, dihydropyramidine dehydrogenase; Gal/GalNAc lectin, Galactose Nacetylgalactosamine lectin; AIG-1, Androgen Inducible gene; Fe-S, Iron sulfur cluster proteins.

and metabolism (Vicente et al., 2009). The authors reported an upregulation of Rad3 helicase, Rad50, and DNA excision base repair proteins. These proteins along with Rad52, are also involved in DNA damage response in the parasite during UV irradiation (Weber et al., 2009) **Table 1**. Two genes coding for deoxyuridine triphosphate nucletotidehydrolase (dUTPase) were upregulated in the parasite exposed to OS. These genes are considered essential for the stability of DNA and have been proposed as a potential drug target against parasites (Nguyen et al., 2005). A 4-fold increase in the homolog of polynucleotide kinase-3 phosphatase along with a 2-fold increase in MutS DNA repair proteins was also reported. These proteins are known for their role in repairing DNA breaks during formed during OS (Chang et al., 2002; Blondal et al., 2005). These proteins were also found to be upregulated when the parasite was exposed to NS.

A functional study of genes responsive to OS revealed the role of the stress-induced adhesion factor, the phospholipidtransporting P-type ATPase and the EhFdp1 oxygen reductase in the resistance of the parasite to OS (Rastew et al., 2012).

The role of Hsps in the resistance of the parasite to stresses has been widely credited as they facilitate the stabilization/sequestration of damaged or misfolded proteins (Kaul and Thippeswamy, 2011). Hsps are very well conserved in all organisms and they function as molecular chaperons during any stressful event (Perez-Morales and Espinoza, 2015). Hsp70 is known to aid in the refolding of denatured and misfolded proteins, and translocation of secretory proteins (Voisine et al., 1999). E. histolytica Hsp70 (EhHsp70) is essential for the resistance of the parasite to OS, the formation of liver abscess, and its levels are also upregulated during heat shock response in the parasite (Akbar et al., 2004; Weber et al., 2006; Santos et al., 2015) **Table 1**. EhHsp70 expression is also upregulated when the E. histolytica Dnmt2 homolog (Ehmeth) is overexpressed. Ehmeth catalyzes the methylation of C<sup>38</sup> present in the anticodon loop of tRNAAsp (Tovy et al., 2010). Ehmeth-overexpressing trophozoites exhibit significantly greater survivability to H2O<sup>2</sup> exposure, which emphasizes the role of EhHsp70 in OS resistance suggesting that EhHsp70 expression is under epigenetic control (Fisher et al., 2006).

It has been recently reported that the gene expression under OS influence is regulated by a transcription factor that binds to a specific motif (AAACCTCAATGAAGA) in the promoters of the genes receptive to H2O<sup>2</sup> (Pearson et al., 2013). It is interesting to note that there is an association between the expression of E. histolytica OS responsive genes and the parasite's virulence (Rastew et al., 2012). Antioxidant enzymes such as glutathione peroxidase, glutathione, and catalase provide a shield for an organism to defend the harsh conditions of OS. E. histolytica lacks the presence of these antioxidative enzymes (Tekwani and Mehlotra, 1999) and it relies mostly on two proteins for its defense against OS. These are the 29-kDa peroxiredoxin (Sen et al., 2007) and the iron peroxide dismutase (Fe-SOD) (Bruchhaus and Tannich, 1994). E. histolytica also relies on exclusive variants of antioxidants that have a low molecular mass like cysteine (Krauth-Siegel and Leroux, 2012). In contrast to E. histolytica, which utilizes superoxide dismutases (SOD), Giardia intestinalis—another common intestinal parasite, has also developed defense mechanisms to cope with OS. This parasite utilizes superoxide reductases (SORs) for the elimination of superoxide anions by reducing H2O<sup>2</sup> (Testa et al., 2011). Moreover, this parasite contains high levels of NADH-dependent oxidase for detoxification of oxygen (Tekwani and Mehlotra, 1999). Pyruvate has also been identified as another alternate potential antioxidant protein that detoxifies H2O<sup>2</sup> inside Giardia spp (Biagini et al., 2001).

An interesting study using capillary electrophoresis–mass spectrometry was performed to determine the effects of OS on the metabolism of the parasite (Husain et al., 2012). OS inactivates the glycolytic pathway and increases the production of glycerol along with changes in nucleotide metabolism, and activation of the chitin biosynthetic pathway. These data suggest that the glycerol synthesis pathway defends the parasite against OS and that oxidized proteins may be crucial constituents of the parasite's machinery to cope with OS. E. histolytica accounts on the thiol-dependent redox metabolism to resist OS (recently reviewed in Jeelani and Nozaki, 2016). The thioredoxindependent system has been extensively studied in E. histolytica and it consists of Fe-superoxide dismutase, rubrerythrin, peroxiredoxin, flavodiiron proteins, and amino acids including L-cysteine, and thioprolines (thiazolidine-4-carboxylic acids), and S-methyl-l-cysteine respectively (Jeelani and Nozaki, 2016). Some of the thiol-dependent redox metabolism proteins like the iron-containing superoxide dismutase have been found to be oxidized in trophozoites exposed to H2O<sup>2</sup> (Shahi et al., 2016a). It was also reported that the presence of superoxide radical anions, which cause OS, lead to the expression of superoxide dismutase and these enzymes are known to contain iron (Bruchhaus and Tannich, 1994). These enzymes further interact with the metabolites of the drug metronidazole and form covalent adducts (Leitsch et al., 2007). The effect of oxidation on the activity of the amebic iron-containing superoxide dismutase has still not been determined yet. Proteomics has helped us to strengthen our understanding of the parasite's response to OS. Comparative proteomics of the virulent strain—HM1:IMSS and the avirulent strain—Rahman showed that Rahman is deficient in two proteins with antioxidative properties (peroxiredoxin and superoxide dismutase). Overexpression of peroxiredoxin in Rahman restores its resistance to OS and its ability to cause colitis in human colonic xenografts (Davis et al., 2006). Proteomics has also been extensively used to study the composition of cysts (Ali et al., 2012) and cyst-like structures (CLS) (Luna-Nacar et al., 2016). CLS is formed in response to OS in the parasite, and they may be part of the parasite's mechanisms to stand OS. CLS share features found in cysts like the presence of a chitin-like resistant coat on their surface and features of trophozoites exposed to OS like the expression of stress response proteins and redox homeostasis (Rastew et al., 2012) and a downregulation of glycolysis and metabolism-related proteins (Jeelani and Nozaki, 2016). Surprisingly, there is no proteomics data available in the literature about the parasite's response to OS and we are currently working to fill this knowledge gap. However, redox proteomics has been recently performed by using resin-assisted capture of oxidized proteins (OX-RAC) (Shahi et al., 2016a). OX-RAC does not provide information about the differential level of proteins before, and after exposure to OS, however, it somewhat identifies oxidized cysteines in proteins following the exposure of the parasite exposed to OS (Kohr et al., 2011). Some of the oxidized proteins identified by OX-RAC like superoxide dismutase have been associated with antioxidant activity. However, a very weak overlapping was found between genes that have their expression changed during OS (Vicente et al., 2009) and oxidized proteins identified by OX-RAC (Shahi et al., 2016a).

Taken together, these findings emphasize the importance of performing multi-omics approaches to fully understand the response of the parasite exposed to stress.

# NITROSATIVE STRESS—HOW DOES THE PARASITE COPE WITH IT?

A second challenge encountered by E. histolytica presents itself in the form of reactive nitrogen species (RNS) in the large intestine. Here the parasite is exposed to NO at nanomolar concentrations by the cells of intestinal epithelium (Kolios et al., 2004). During acute inflammation, the activation of specific immune cells of the host's immune system comprising of natural killer cells, macrophages, and neutrophils releases NO in micromolar amount (Thomas et al., 2008; Begum et al., 2015). S-nitrosylation is a post-translational modification that occurs by covalent attachment of NO group to the thiol side chain of cysteine residues in proteins (Hess et al., 2005). NO-mediated cytotoxicity is partly due to the formation of S-nitrosylated proteins which results in aberrant protein function. For example, the inactivation of key glycolytic enzymes in the parasite exposed to NO leads to the fragmentation of the endoplasmic reticulum, low levels of ATP and the death of the parasite (Santi-Rocca et al., 2012). S-nitrosylation of E. histolytica cysteine proteinases, which are essential virulence factors, inactivate their activity (Siman-Tov and Ankri, 2003). S-nitrosylated (SNO) proteins may also be involved in the regulation of protein activity and function (Gould et al., 2013). Insights into the formation of SNO proteins and their regulation can be achieved by performing SNO-RAC analysis. This technique involves the capture of SNO proteins by chromatography on thiopropyl sepharose and mass spectrometry for their identification (Forrester et al., 2009). The heavy subunit of the E. histolytica Gal/GalNAc lectins was one of the SNO proteins identified by SNO RAC (Hertz et al., 2014a). S-nitrosylation occurred in the cysteine residues of the Gal/GalNAc lectin—carbohydrate recognition domain, which in turn disabled E. histolytica to attach to mammalian cells (Hertz et al., 2014a). The effects of NO on E. histolytica were also studied by transcriptomics. The most extensive groups of genes induced by NS are those related with signaling/regulatory processes, DNA repair, redox-regulation, glycolysis-related genes, and lipids (Vicente et al., 2009; Santi-Rocca et al., 2012). There is a significant overlap of genes modulated under NS and OS including genes involved in degradation and repair of misfolded proteins, lipid metabolism, transport and glycosylation, and DNA repair (Vicente et al., 2009). This overlap shows that the parasite uses common strategies to overcome the cytotoxicity of ROS and NOS (Vicente et al., 2009). Functional studies have demonstrated that Ehmeth is involved in the resistance of E. histolytica to NS (Hertz et al., 2014b). The authors proposed that enolase, a glycolytic protein that inhibits Ehmeth activity (Tovy et al., 2010), cannot bind to SNO-Ehmeth, thereby leading to higher tRNAAsp methylation. NO resistance can also be achieved by the methylation of tRNAAsp (Hertz et al., 2014b) and this requires more studies to elucidate the mechanism.

Infection with E. histolytica leads to non-symptomatic amoebiasis in 90% of the cases (Nath et al., 2015; Ishikane et al., 2016). Despite the absence of apparent inflammation, the parasite is nevertheless exposed to nanomolar levels of NO (Kolios et al., 2004). The exposure to low NO concentration can also occur due to molecules produced during the mechanism of denitrification by the microbiome present in the gut (Vermeiren et al., 2009). This non-toxic concentration of NO may prepare the parasite to resist higher concentration of NO. A recent study (Shahi et al., 2016b) has tested this hypothesis by adapting E. histolytica to a progressive amount of the NO donor Snitrosoglutathione (GSNO) up to 120µM. Here, the authors identified and studied the role of N-acetyl-ornithine deacetylase (NAOD) in aiding the parasite to adapt to NS. NAOD catalyzes the deacetylation of N-acetyl-L-ornithine to acetate and ornithine respectively. However, they found that the catalytic activity of this enzyme is not necessary to confer this protective effect on the parasite. Rather, NAOD has an interacting partner; glyceraldehyde 3-phosphate dehydrogenase (GAPDH). GAPDH has many roles apart from its role in metabolism (Nicholls et al., 2012). Its expression is increased in trophozoites exposed to NS and a high expression of this enzyme is detrimental to the parasite (Shahi et al., 2016b). Consequently, NAOD serves as a moonlighting protein that neutralizes the detrimental effect of GAPDH and helps the parasite to adapt to NS. Exciting pieces of evidence about the adaptation of the parasite to NS stem from SNO-RAC analysis (Trebicz-Geffen et al., 2017). Out of the SNO proteins in trophozoites adapted to NS, a significant enrichment of actin family cytoskeleton proteins was found. Trophozoites adapted to NO have their ability to form actin filaments impaired and consequently have their virulence reduced. These phenotypes are reversed upon removal of GSNO from the medium which suggests that the parasite has to compromise on some level for its adaptation to NS.

While E. histolytica depends on the NAOD-GAPDH interaction to neutralize the toxic effects NO, flavohemoglobin (flavoHb) in Giardia intestinalis reported to aerobically metabolize NO efficiently (Rafferty and Dayer, 2015). During NS, the expression of this protein is increased, and hence it aids in the breakdown of NO, whereas under normal conditions its expression is low.

# GLOBAL CHANGES IN THE PARASITE IN THE ABSENCE OF L-CYSTEINE, IRON, AND GLUCOSE

# L-Cysteine Starvation

The parasite requires L-cysteine for its growth and survival. It obtains L-cysteine either through the extracellular medium or produces it via the de novo synthesis pathway with the help of Serine acetyl transferase (SAT) and cysteine synthase (CS) (Pye et al., 2004; Hussain et al., 2009). It is well known that cysteine takes part into the post-translational modifications of several proteins and its role also extends to redox mechanisms, electron transfer reactions, and many more processes (Beinert et al., 1997). The unique feature of this amino acid lies in the ability of its thiol group to endure redox reactions owing to its antioxidant property (Krauth-Siegel and Leroux, 2012). This feature makes cysteine extremely beneficial for the parasite which is deficient of antioxidant enzymes such as catalase and glutathione S-transferases. Cysteine is also essential for its growth, its adherence, and its resistance to OS (Jeelani et al., 2014) and also protects the parasite from the oxidative stress induced by the anti-amebic drug metronidazole (Leitsch et al., 2007). During the course of NS, the requirement of L-cysteine is increased, and the de novo pathway tries to compensate for increasing the production of cysteine. Yet, the amount of cysteine remains low in the cell and requires the addition of L-cysteine in the medium (Jeelani and Nozaki, 2014). Transcriptome analysis performed in the absence of L-cysteine revealed the upregulation of several genes belonging to the Fe-S cluster family of proteins. Among them, three iron-sulfur flavoproteins (EHI\_025710, EHI\_067720, and EHI\_138480) showed higher expression in the absence of L-cysteine. Downregulation of EHI\_025710 expression severely affected the growth of the parasite, whereas the downregulation of EHI\_138480 showed a mild defect in the growth of E. histolytica in the absence of cysteine (Husain et al., 2011).

The deprivation of cysteine has also led to an increase in the intracellular concentration of metabolites such as O-acetyl serine sulfhydralase, glycerol-3-phosphate, and isopropanolamine (**Table 1**). Moreover, S-methyl cytosine levels were also increased in the cysteine-starved trophozoites, whereas under normal growth condition, S-methyl cytosine was undetectable (Husain et al., 2010). Cysteine deprivation has also caused an accumulation of phosphatidylisopropanolamine (Husain et al., 2011). The role of this unusual phospholipid is still unknown.

# Iron Starvation

Iron is essential for the growth of E. histolytica (Park et al., 2001). Ferric ammonium citrate is provided to the parasite as the source of iron in the medium under laboratory conditions. In the host, the trophozoites scavenge iron from the bacterial flora and the iron-binding proteins of the host such as hemoglobin and ferratin by either phagocytosis or through hemophores or siderophores (Wandersman and Delepelaire, 2004). These proteins are released once the trophozoites phagocytose the cells (Lopez-Soto et al., 2009). Not much is known about the uptake, storage, and utilization of iron in E. histolytica. However, only a few proteins such as Rubrerythrin, NifS, and NifU are known to be involved in the iron-sulfur cluster formation (Ali et al., 2004; Maralikova et al., 2010). Iron is essential for the activity of proteins such as alcohol dehydrogenase 2, superoxide dismutase and ferredoxin (Tannich et al., 1991). Exposure to low iron concentrations in vitro impairs the parasite's adherence and cytopathic activity (Lee et al., 2008; Espinosa et al., 2009) suggesting the central role of iron toward the pathogenicity of the parasite. Iron may also affect the attachment of the parasite to the host cells along with cytotoxicity. Interestingly, there is a definite correlation between adherence, cytopathic activity and increasing concentration of iron. This effect is solely specific to iron and replacing the iron with other cationic salts did not reproduce this result (Lee et al., 2008). Transcriptome analysis of trophozoites growing in the absence of iron shows an increase in the transcription level of cysteine proteinases (CP-A5, CP-A7, and CP-EHI\_01850), ribosomal proteins and elongation factors which are required for translation (**Table 1**). Iron deprivation also upregulates the expression of acyl-CoA synthetase, Androgen-Inducible Gene 1 (AIG1), ComEC protein, NADPH-dependent oxidoreductase (EhNO2) (Hernandez-Cuevas et al., 2014). The authors have also reported the higher expression of three genes belonging to the AIG1 family (EHI\_022500, EHI\_115160, and EHI\_195260). A comparative study between two Entamoeba histolytica cell lines showed that these AIG1 genes were highly expressed in one of the cell lines that caused liver abscess in the gerbil model compared to the other which did not produce abscess (Biller et al., 2010). This observation suggests that AIG1 has a role in the virulence of the parasite. EhNO<sup>2</sup> is involved in reducing cystine to cysteine as cysteine is normally present in the oxidized state inside the cells. Thus, its reduction is necessary before its utilization for various functions. EhNO<sup>2</sup> also takes part in the reduction of metronidazole to activate this drug to produce reactive species that are toxic to the parasite (Jeelani et al., 2010). Three families of transporters were also upregulated in the absence of iron. ABC-family of transport proteins, P-glycoprotein-5 and Major family transporters (Hernandez-Cuevas et al., 2014). A similar family of transporters present in Leishmania is known for its ability to scavenge heme/iron from the medium (Perez-Victoria et al., 2001). An increase in the expression of these transporters in E. histolytica suggests that the parasite tries to acquire iron/heme from different sources during iron starvation. However, the role of these transporters in E. histolytica needs to be determined.

# Glucose Starvation

E. histolytica resides the human colon, a niche where the amount of available glucose for fermentation is low (around 0.2 gr glucose/kg-1 tissue) due to the high absorptive capacity of the glucose transporters in the small intestine (Cummings and Macfarlane, 1997; Kellett et al., 2008; Hirayama et al., 2009). As a result, the parasite is facing glucose starvation (GS) in the human gut. The phenotypic and metabolic responses of E. histolytica to GS have been recently reviewed (Baumel-Alterzon and Ankri, 2014). When exposed to GS, the parasite downregulates the expression of genes involved in glycolysis and upregulates the genes involved in the degradation of stored products such as glycogen granules to make free glucose (**Table 1**). The research of new sources of energy by the parasite is illustrated by the upregulation of dihydropyrimidine dehydrogenase (DPD) expression. DPD is involved in the degradation of pyrimidines, and it is essential for the adaptation of E. histolytica to GS (Baumel-Alterzon et al., 2013). Although the exact role of DPD in the adaptation of E. histolytica to GS is still not understood, it is possible that the induction of DPD expression during glucose starvation contributes to energy production through the degradation of pyrimidines (Girjes et al., 1995; Beaulande et al., 1998). Metabolomics approaches that have been recently adapted to the study of E. histolytica metabolism (Jeelani and Nozaki, 2014) will help establish the existence of these pyrimidine degradation pathways in E. histolytica. Another example that illustrates the ability of the parasite to seek for an alternative source of carbon is the β-amylase-mediated degradation of mucin present in the colon (Thibeaux et al., 2013).

Another behavior associated with the exposure of the parasite to GS is its enhanced virulence. This phenotype correlates with the upregulation of some virulence factors like the Gal/GalNAc lectins and EhCP-A4, which is a cysteine protease that aids the parasite to invade the host by destroying the intracellular matrix (Baumel-Alterzon et al., 2013).

GS is also involved in epigenetic regulation by promoting the shuttling of the glycolytic enzyme enolase in the nucleus and the inhibition of Ehmeth (Tovy et al., 2011). The consequence of Ehmeth inhibition on the adaption of the parasite to GS is not yet understood.

# HUMAN INTESTINAL GUT FLORA

The gut hosts a plethora of microbes, and their number exceeds 10<sup>14</sup> (Thursby and Juge, 2017). It is estimated that about 400–1,000 bacterial species are present in the intestine, and 97% of the total population is made up by 30–40 species of bacteria (Xu et al., 2003; Sonnenburg et al., 2004; Jandhyala et al., 2015). The heterogeneous population of microbes inside the human gastrointestinal (GI) tract gives them the potential to influence human physiology in terms of promoting health and causing a wide variety of diseases such as obesity, irritable bowel syndrome and diarrhea, cardiovascular disorders, liver cirrhosis, to name a few (Chang and Lin, 2016). Upon entry into the host, E. histolytica finds itself in contact with a plethora of bacterial species. The intestinal lumen is the site of proliferation of the E. histolytica trophozoites, and they phagocytose the resident bacteria. Bacteria possessing specific recognition motifs were able to adhere to the parasite and to undergo ingestion (Bracha and Mirelman, 1984). The intricate relationship that exists between E. histolytica and the gut flora was the subject of many studies which concluded that it affects greatly several aspects of E. histolytica physiology. An axenic culture of E. histolytica replenished with E. coli O:55, was found to be more virulent or less virulent compared to the axenic culture alone depending on the time of interaction between the parasite and E. coli O:55 (Bracha et al., 1982; Mirelman, 1987; Padilla-Vaca et al., 1999). Moreover, it was found that bacteria exert their effect on E. histolytica virulence through cell surface entities such as lectin that attach to specific carbohydrate domains. One such gene coding for a 35-kDa lectin subunit of a heterodimeric Gal/GalNAc lectin molecule was found to be involved in promoting pathogenicity of the parasite, whereas the heavy chain subunit of 170 kDa was involved in adherence and attachment to the bacteria (Bhattacharya et al., 1992, 1998; Ankri et al., 1998; Padilla-Vaca et al., 1999). Apart from virulence, incubation with intestinal bacterial flora was suggested to trigger encystations in this parasite, which may explain why E. histolytica axenic cultures are unable to encyst in vitro (Ehrenkaufer et al., 2007). E. histolytica also has a direct influence on the gut microbiota composition (Verma et al., 2012). The presence of enteropathogenic bacteria (Paniagua et al., 2007) or the presence of Prevotella copri (Gilchrist et al., 2016), a normal component of the gut microbiota, has been correlated with E. histolytica infection. In contrast, the presence of Clostridia segmented filamentous bacteria is harmful to the parasite (Burgess et al., 2014).

# REFERENCES


# CONCLUDING REMARKS

Appropriate adaptation to stress is essential for E. histolytica's survival in harsh environments. In this review, we have shown that the parasite has developed multiple mechanisms regulated at the transcriptomics or at the post-transcriptomics level that allow an adequate response to a specific stress. While some features of the response of E. histolytica to stress are stress specific, some features such as the extensive reorganization of gene expression and expression of Hsps or antioxidant proteins such as thioredoxin are shared among different stresses. For years, studies about E. histolytica's response to stresses have almost exclusively focused on the instant response of the parasite to acute stresses, but its ability to adapt to these stresses has not been sufficiently considered. Recent reports are illustrating the exceptional ability of this parasite to adapt to various stresses like GS, NS, and the role of DPD and NAOD proteins in the mechanism of adaptation (Baumel-Alterzon et al., 2013; Shahi et al., 2016b). We believe that strategies that counteract the protective effect of these proteins may be valuable in the struggle against this parasite. The identification of new targets for antiamebic therapeutics will also beneficiate from a better integration of omics data available in the literature and of the ones to come. Multi-omic processing and combination of these data is a general challenge (Palsson and Zengler, 2010) that can only be overcome by training competent people to perform this task. Finally, it is essential not to neglect the procedures downstream to omics analysis of E. histolytica by improving current animal models of amoebiasis (Tsutsumi and Shibayama, 2006). This will enable us to reproduce the clinical manifestations observed in the human host in a better way. Moreover, the tools to knock out/downregulate gene expression can be enhanced (Morgado et al., 2016) by adapting CRISPR/Cas9 system to E. histolytica (Cui and Yu, 2016).

# AUTHOR CONTRIBUTIONS

All authors listed, have made substantial, direct and intellectual contribution to the work, and approved it for publication.

# FUNDING

This study was supported by the Israel Ministry of Health within the framework ERA-NET Infect-ERA (031L0004) (AMOEBAC project) and grants from the Israel Science Foundation (ISF) (260/16) and U.S.-Israel Binational Science Foundation (BSF) (2015211).


and extracellular signal-regulated kinase for migration of human hepatoma cell Hepg2. Mol. Cancer Res. 4, 747–758. doi: 10.1158/1541-7786.MCR-0 6-0096

Xu, J., Bjursell, M. K., Himrod, J., Deng, S., Carmichael, L. K., Chiang, H. C., et al. (2003). A genomic view of the human-Bacteroides thetaiotaomicron symbiosis. Science 299, 2074–2076. doi: 10.1126/science.1080029

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The handling editor is currently editing co-organizing a Research Topic with one of the author SA, and confirms the absence of any other collaboration.

Copyright © 2018 Nagaraja and Ankri. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

#### *Edited by:*

Serge Ankri, Technion – Israel Institute of Technology, Israel

#### *Reviewed by:*

Andrei A. Zimin, Institute of Biochemistry and Physiology of Microorganisms (RAS), Russia Nicholas A. Wallace, Kansas State University, United States

#### *\*Correspondence:*

Elisa Azuara-Liceaga elisa.azuara@uacm.edu.mx Luis G. Brieba luis.brieba@cinvestav.mx

#### *†Present Address:*

Cesar S. Cardona-Felix, Centro Interdisciplinario de Ciencias Marinas de Instituto Politécnico Nacional, La Paz, Mexico Guillermo Pastor-Palacios, División de Biología Molecular, Instituto Potosino de Investigación Científica y Tecnológica, San Luis Potosí, Mexico

> *Received:* 20 March 2018 *Accepted:* 07 June 2018 *Published:* 12 July 2018

#### *Citation:*

Azuara-Liceaga E, Betanzos A, Cardona-Felix CS, Castañeda-Ortiz EJ, Cárdenas H, Cárdenas-Guerra RE, Pastor-Palacios G, García-Rivera G, Hernández-Álvarez D, Trasviña-Arenas CH, Diaz-Quezada C, Orozco E and Brieba LG (2018) The Sole DNA Ligase in Entamoeba histolytica Is a High-Fidelity DNA Ligase Involved in DNA Damage Repair. Front. Cell. Infect. Microbiol. 8:214. doi: 10.3389/fcimb.2018.00214

# The Sole DNA Ligase in *Entamoeba histolytica* Is a High-Fidelity DNA Ligase Involved in DNA Damage Repair

Elisa Azuara-Liceaga<sup>1</sup> \*, Abigail Betanzos 2,3, Cesar S. Cardona-Felix 2,4† , Elizabeth J. Castañeda-Ortiz <sup>1</sup> , Helios Cárdenas <sup>1</sup> , Rosa E. Cárdenas-Guerra<sup>1</sup> , Guillermo Pastor-Palacios 4†, Guillermina García-Rivera<sup>3</sup> , David Hernández-Álvarez <sup>1</sup> , Carlos H. Trasviña-Arenas <sup>4</sup> , Corina Diaz-Quezada<sup>4</sup> , Esther Orozco<sup>3</sup> and Luis G. Brieba<sup>4</sup> \*

<sup>1</sup> Posgrado en Ciencias Genómicas, Universidad Autónoma de la Ciudad de México, Mexico City, Mexico, <sup>2</sup> Consejo Nacional de Ciencia y Tecnología, Mexico City, Mexico, <sup>3</sup> Departamento de Infectómica y Patogénesis Molecular, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico City, Mexico, <sup>4</sup> Laboratorio Nacional de Genómica para la Biodiversidad, Centro de Investigación y de Estudios Avanzados, Irapuato, Mexico

The protozoan parasite Entamoeba histolytica is exposed to reactive oxygen and nitric oxide species that have the potential to damage its genome. E. histolytica harbors enzymes involved in DNA repair pathways like Base and Nucleotide Excision Repair. The majority of DNA repairs pathways converge in their final step in which a DNA ligase seals the DNA nicks. In contrast to other eukaryotes, the genome of E. histolytica encodes only one DNA ligase (EhDNAligI), suggesting that this ligase is involved in both DNA replication and DNA repair. Therefore, the aim of this work was to characterize EhDNAligI, its ligation fidelity and its ability to ligate opposite DNA mismatches and oxidative DNA lesions, and to study its expression changes and localization during and after recovery from UV and H2O<sup>2</sup> treatment. We found that EhDNAligI is a high-fidelity DNA ligase on canonical substrates and is able to discriminate erroneous base-pairing opposite DNA lesions. EhDNAligI expression decreases after DNA damage induced by UV and H2O<sup>2</sup> treatments, but it was upregulated during recovery time. Upon oxidative DNA damage, EhDNAligI relocates into the nucleus where it co-localizes with EhPCNA and the 8-oxoG adduct. The appearance and disappearance of 8-oxoG during and after both treatments suggest that DNA damaged was efficiently repaired because the mainly NER and BER components are expressed in this parasite and some of them were modulated after DNA insults. All these data disclose the relevance of EhDNAligI as a specialized and unique ligase in E. histolytica that may be involved in DNA repair of the 8-oxoG lesions.

Keywords: EhDNAligI, protozoan, DNA insults, ligation, repairing, 8-oxoG adduct, NER and BER pathways

# INTRODUCTION

Entamoeba histolytica, the protozoan parasite responsible for human amoebiasis, must cope with reactive oxygen (ROS) and nitric oxide (NOS) species derived from human immune cells, during colonic tissue invasion and metronidazole drug treatment (Vicente et al., 2009; Wilson et al., 2012). ROS and NOS species generate DNA lesions such as 8-oxoguanine (8 oxoG), thymine glycol (Tg) and pyrimidine dimers (Demple and Harrison, 1994). 8-oxoG is the most abundant DNA lesion formed during oxidative stress. During UV and H2O<sup>2</sup> treatments, oxygen radicals generate 8-oxoG (Dianov et al., 1998). To avoid deleterious mutations this lesion is repaired primarily via the base excision repair (BER) pathway, which is present in two modalities: short and long-path (Lindahl, 1993). 8-oxoG is also recognized and processed by the nucleotide excision repair (NER) pathway, suggesting a cross talk between NER and BER to repair non-bulky oxidative DNA lesions (Menoni et al., 2012; Parlanti et al., 2012). DNA replication and repair are multienzymatic reactions that require the orchestrated participation of several enzymes, each with a specific task. For instance, in eukaryotic cells the interactions between DNA ligase I with PCNA and DNA polymerase β are critical for Okazaki fragment maturation and DNA repair, respectively (Dianov et al., 1998). DNA ligase I is recruited to UV-damage sites only in proliferating cells and it is implicated in BER and NER pathways (Moser et al., 2007). In humans the deficiency in DNA ligase I induces sunlight sensitivity, suggesting a role of this enzyme in DNA double-strand break repair (Bhat et al., 2006). DNA ligase III is involved in NER following UV-damage in quiescent cells, singlestrand, and double stranded break repair; whereas DNA ligase IV is involved in DNA double-strand breaks repaired by the NHEJ pathway (Tomkinson et al., 2006).

In contrast to vertebrates, the protozoan parasite E. histolytica contains only one DNA ligase dubbed EhDNAligI (Cardona-Felix et al., 2010). EhDNAligI is similar to DNA ligase I from higher eukaryotes, however, its N-terminal is approximately 150 amino acids shorter than its human counterpart. The enzymatic activity of EhDNAligI is stimulated by the DNA polymerase processivity factor PCNA (Buguliskis et al., 2007; Cardona-Felix et al., 2010, 2011). The lack of other ATP or NAD+-dependent DNA ligases in the genome of E. histolytica suggests that EhDNAligI could participate in repairing diverse DNA lesions and sealing Okazaki fragments. Additionally, DNA sequencing of the genome of E. histolytica reveals that this parasite contains genes of the BER and NER pathways, including glycosylases, an apurinic endonuclease, excision repair proteins and helicases (Loftus et al., 2005; Clark et al., 2007; López-Camarillo et al., 2009). Some NER proteins of E. histolytica are overexpressed in response to UV irradiation (Weber et al., 2009).

In E. histolytica few proteins involved in DNA metabolism such as EhDNAligI, EhPCNA, EhMutY, specialized DNA polymerases, Myb transcription factors, and EhRad51 have been biochemically characterized (López-Casamichana et al., 2008; Cardona-Felix et al., 2010, 2011; Meneses et al., 2010; Pastor-Palacios et al., 2010, 2012; Trasviña-Arenas et al., 2016). Because EhDNAligI is the sole ligase in E. histolytica, we addressed its role during DNA damage induced by UV and H2O<sup>2</sup> treatments. We tested its fidelity and ligation capabilities in sealing different DNA lesions in vitro and analyzed its changes in expression and localization upon DNA damage.

# MATERIALS AND METHODS

# Nick-Sealing Ligation Reactions

Recombinant EhDNAligI (rEhDNAligI) was heterologously expressed, purified and deanylated as previously described (Cardona-Felix et al., 2010). T4 DNA ligase was purchased from New England Biolabs and deadenylated using molar excess of pyrophosphate. Ligation reactions were performed in 50 mM Tris pH 7.5, 1 mM ATP, 10 mM MgCl2, and 5 mM DTT. Double stranded DNA nicked substrates were annealed accordingly to the figure legends. Nick-sealing reactions were carried out at 37◦C using 50 fmol of EhDNAligI and 250 fmol of nicked <sup>32</sup>P labeled substrates. Reactions were stopped by adding equal volumes of stop buffer (95% formamide, 0.01% bromophenol blue, 1 mM EDTA). Samples were run on a 15% polyacrylamide, 8 M urea denaturing gel and analyzed on a phosphorimager (Storm; GE Healthcare) to quantify the results. All experiments were performed by triplicate.

# Electrophoretic Mobility Shifts Assays

To determine the DNA binding specificity of rEhDNAligI to different lesions in comparison to a canonical substrate, we prepared three different substrates for DNA mobility shift assays. Oligonucleotides were mixed at equimolar amounts in an annealing buffer (20 mM Tris pH 7.5 and 150 mM NaCl), heated at 95◦C for 5 min and slowly cooled to room temperature. Binding experiments were carried out with 300 fmol of DNA substrate and 26, 52, and 104 fmol of rEhDNAligI. Complex formation was analyzed by EMSA.

# Steady-State Kinetics

Michaelis-Menten steady-state kinetics for ATP consumption was measured using 500 fmol of each of the three substrates. In all cases, the 5′ substrate was phosphorylated with γ-<sup>32</sup>P ATP. ATP concentrations varied from 3.125 to 800 nM. Ligation reactions were separated by electrophoresis and quantified on the phosphorimager. The Km and Kcat values were determined by fitting to the Michaelis–Menten equation as previously described (Buguliskis et al., 2007; Cardona-Felix et al., 2010).

# Survey of BER and NER Gene Encoding Proteins in *E. histolytica* Genome

Sequences annotated as NER and BER proteins were obtained from E. histolytica database (http://amoebadb.org/). For each component of the amoeba NER and BER proteins we searched for its putative orthologous in Homo sapiens, Saccharomyces cerevisiae, and other species using blast (http://blast.ncbi.nlm. nih.gov) against NCBI RefSeq protein data base restricting to the corresponding organism or the whole data base. The Percentage Identity (PID) was calculated using amino acids sequences from NER and BER proteins from E. histolytica, H. sapiens, S. cerevisiae and the organism corresponding to the best hit, taking gaps into account using the following equation: PID = (identical positions/length of the alignment) ×100. Oligonucleotides of the selected NER and BER genes (**Table S1**) were designed according to the DNA sequence obtained from NCBI RefSeq protein data base.

# *E. histolytica* Cultures

Trophozoites of HM1:IMSS strain were axenically cultured in TYI-S-33 medium supplemented with 15% of bovine serum at previously described (Diamond et al., 1978) and used in logarithmic growth phase for all experiments.

# Trophozoites Treatments

Trophozoites from 24 h cultures were subjected to two different types of stress: UV light irradiation and exposure to H2O2. For the first treatment, 5 × 10<sup>5</sup> trophozoites were irradiated with 254 nm UV-C light at 150 J/m<sup>2</sup> using a UV Stratalinker 1800 device (Stratagene) at previously described (López-Casamichana et al., 2008). For H2O<sup>2</sup> treatment, 8 × 10<sup>5</sup> trophozoites were exposed to 2.5 mM of H2O<sup>2</sup> diluted in serum-depleted medium to avoid its protective effect during oxidative stress derived from peroxide, at 37◦C for 1 h, as previously described (Shahi et al., 2016). After treatments, cells were allowed to recover in fresh TYI-S-33 complete medium at 37◦C during 1, 3, 6, and 24 h. Trophozoites were chilled for 10 min and parasites were harvested at 1,500 rpm. Cell viability was determined using a TC20 Automated Cell Counter (Bio-Rad) using Trypan blue dye exclusion assays (Bio-Rad) and trophozoites with 80% viability were used for further experiments.

# RT-PCR and qRT-PCR Assays

Total RNA from trophozoites either from the untreated control or immediately after treatment with UV or H2O<sup>2</sup> was isolated using TRIzol (Life Technologies), following the manufacturer's protocol. cDNA was synthesized using 500 ng of total RNA with an oligo(dT)15 (Promega). The PCR reactions contained 1 µl of cDNA and 0.4 µmol of each specific oligonucleotide combination (**Table S1**). RT-PCR products were resolved in 2% agarose gels, stained with ethidium bromide and visualized with a FLA-5100 Imaging System (FUJIFILM). qRT-PCR was performed using the QuantiFast SYBR Green RT-PCR kit (Qiagen) and 100 ng of the total RNA, according to the manufacturer's instructions. Relative changes in gene expression were calculated using 40s rps2 as an internal gene calculated by 11CT method according to the Guide to Performing Relative Quantitation of Gene Expression Using Real-Time Quantitative PCR from Applied Biosystems. All experiments were performed in triplicate and repeated in independent experiments by duplicate.

# Production of Polyclonal Antibodies Against EhDNAligI and EhPCNA

rEhDNAligI and rEhPCNA recombinant proteins were heterologously expressed and purified as previously described (Cardona-Felix et al., 2010, 2011) and subsequently used as antigens to immunize New Zealand male rabbits. Before immunization, the pre-immune serum from rabbits was obtained; rabbits were subcutaneously inoculated with an initial dose of 130 µg rEhDNAligI or EhPCNA diluted in Freund's complete adjuvant (Sigma-Aldrich) and 4 more booster injections (130 µg each) at 21-day intervals. After 1 week of the last immunization, rabbits were bleed and polyclonal antisera were obtained. Reactivity of the generated antibodies was tested by Western blot assays, using rEhDNAligI and rEhPCNA recombinant proteins (**Figure S1**).

# Western Blot Assays

Lysates from trophozoites either untreated control, immediately after treatment with UV or H2O<sup>2</sup> or during different recovery times (1, 3, and 6 h) were obtained using trichloroacetic acid (TCA; Sigma-Aldrich) method. Proteins were separated by 12% denaturing polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto nitrocellulose membranes (Millipore). Membranes were blocked with 5% nonfat dried milk and incubated for 18 h at 4◦C with rabbit <sup>α</sup>-EhDNAligI (1:1,000) or α-γH2AX (1:7,000 dilution; Cell Signaling) antibodies or rabbit pre-immune serum (1:1,000) or mouse α-actin (1:1,500; antibody; Manning-Cela et al., 1994) followed by incubation for 2 h at room temperature with peroxidase-conjugated goat αrabbit (1:5,000 dilution; Santa Cruz Biotechnology) or α-mouse (1:10,000 dilution; Invitrogen) secondary antibodies. Membranes were developed using an enhanced chemiluminescence kit (Luminata Forte Western HRP Substrate, Millipore), in the Image Lab 5.2.1 System (Bio-Rad). All experiments were performed in triplicate and repeated in independent experiments by duplicate.

# Immunofluorescence and Determination of the 8-oxoguanine (8-oxoG) Formation by Confocal Microscopy

The 8-oxoG formation was analyzed using a direct fluorescencebased binding assay with avidin-FITC (fluorescent dye, Sigma-Aldrich). Avidin binds with high affinity to 8-oxoG as Struthers et al. described (Struthers et al., 1998). Trophozoites untreated or treated with UV or H2O<sup>2</sup> or recovered at different times (1, 3, 6, and 24 h) were fixed with absolute ethanol for 20 min at −20◦C, washed with PBS pH 6.8 and blocked with 50 mM NH4Cl for 30 min at 37◦C, followed by incubation of 1% BSA for 30 min at 37◦C. Cells were incubated with 5µg/mL FITCconjugated avidin (1:50) for 1 h at room temperature in darkness. Then, samples were incubated with rabbit α-EhDNAligI (1:200) antibody overnight at 4◦C. Next day, cells were washed with PBS pH 6.8 and incubated with α-rabbit TRITC-labeled secondary antibody (1:100; Jackson Immuno Research). Nucleic acids were counterstained with 2.5µg/mL DAPI (4′ ,6′ -diamidino-2-phenylindole; Zymed) for 5 min. Preparations were washed with PBS and fluorescence was preserved using antifade reagent Vectashield (Vector Lab). For some experiments, α-EhDNAligI (1:200) and α-EhPCNA (1:200) antibodies were Alexafluor 647 and 555 labeled, respectively using the Antibodies Labeling Kit (Molecular Probes), following the manufacturer's instructions. Light optical sections were obtained through a Nikon inverted microscope attached to a laser confocal scanning system (Leica Microsystems) and analyzed by Confocal Assistant Software Image. The integrated density fluorescence of the nucleus was quantified using the ImageJ software and it was normalized regarding the fluorescence of the whole cell in 5 different fields of three Z-stack sections (0.5µm) of independent experiments. Nuclear co-localization was quantified in 1µm z-stacks confocal images to obtain the Pearson Coefficient using the Just Another Co-localization Plugin (JACoP) in the Image J 1.48i software. Each point represented an average of 10–20 cells and values are given as means and standard error.

# Statistical Analyses

All data shown are displayed as mean with standard error in triplicate and repeated in independent experiments by duplicate. GraphPad Prism 6.0e software was used for student t-test by two-tailed analyses.

# Ethics Statement

The Institutional Animal Care and Use Committee (Cinvestav IACUC/ethics committee) reviewed and approved the animal care and use of rabbits and employed to produce antibodies (Protocol Number 0313-06, CICUAL 001). All steps were taken to ameliorate the welfare and to avoid the suffering of the animals. Rabbits were housed under controlled conditions of humidity, temperature, and light (12 h light/12 h dark cycles). Food and water were available ad libitum. Animals were monitored pre and post-inoculation. All procedures were conducted by trained personnel under the supervision of veterinarians and all invasive clinical procedures were performed while animals were anesthetized and when it was required animals were humanely euthanized. The ethics committee verified that our institute fulfills the NOM-062-ZOO-1999 regarding the Technical Specifications for Production, Care, and Use of Laboratory Animals given by the General Direction of Animal Health of the Minister of Agriculture of Mexico (SAGARPA-Mexico). The technical specifications approved by SAGARPA-Mexico fulfill of the international regulations/guidelines for the use and care of animals used in laboratory and were verified and approved by Cinvestav IACUC/ethics committee (Verification Approval Number: BOO.02.03.02.01.908).

# RESULTS

# rEhDNAligI Is a High-Fidelity DNA Ligase

To initiate the biochemical characterization of EhDNAligI, we purified the rEhDNAligI recombinant protein and performed a nick-sealing in vitro assay. Nick-sealing fidelity in DNA ligases relates to the extent to which the enzyme can ligate substrates containing mismatched bases on either side of the nick. Thus, we tested the 12 possible DNA mismatches at the 3′ and 5′ sides of a double-stranded DNA nick (**Figure 1A**). We found that rEhDNAligI was able to discriminate almost all the noncanonical Watson-Crick base pairs either at the 3′ -OH or 5′ - PO<sup>4</sup> of the nicked DNA. rEhDNAligI only efficiently ligated the T:G mismatch at the 5′ - PO<sup>4</sup> strand (**Figure 1A**, lane 20). DNA ligation kinetics for a G:T mismatch in comparison to ligation employing G:C and T:A canonical pairs indicated that the mismatch and the canonical pairs are ligated with similar catalytic efficiencies that differ by no more than twofold in their Km. This data indicates the among the 12 possible Watson-Crick errors, EhDNAligI only efficiently recognizes the T:G mismatch as a substrate for ligation (**Figure 1B**; **Table 1**). G:T mismatches occurs by deamination of 5-methylcytosine to thymine, the lack of substrate discrimination by EhDNAlig I is also observed in Thermus thermophilus DNA ligase (Luo et al., 1996).

# rEhDNAligI Is Able to Discriminate Across From DNA Lesions

DNA is susceptible to damage from endogenous or exogenous sources. Oxidative damage, including species such as singlet oxygen atoms, hydroxyl radicals, and superoxide radicals, alters the coding specificity of the nitrogenous bases. Thus, we tested four possible substrates (**Figure 2A**) with different base modifications related to different oxidative stress and compared the ability of rEhDNAligI and T4 DNA ligase to use those substrates. As 8-oxoG is the most common oxidative lesion generated by ROS, we tested the ligation efficiency of different substrates that contained 8-oxoG either at the 3′ -OH or 5′ - PO<sup>4</sup> (**Figure 2B**). rEhDNAligI was able to ligate a dATP:8-oxoG pair but not a dCTP:8-oxoG pair at the 3′ -OH (**Figure 2B** lanes 1, 2), in contrast, T4 DNA ligase was able to use as a substrate dCTP and dATP paired to an 8-oxoG lesion (**Figure 2B** lanes 5, 6). This substrate discrimination is similar to the one present in human DNA ligase I (Hashimoto et al., 2004). 8 oxoG lesion, at the 5′ -PO<sup>4</sup> of the lesion, is efficiently ligated when paired with dATP or dCTP by EhDNAligI and T4 DNA ligase (**Figure 2B** lanes 3, 4, 7, 8). 5,6-dihydroxy-5,6 dihydrothymine (thymine glycol), is another important DNA lesion induced by oxidation and ionizing radiation. In this work, we used two thymine glycol isomers, the 5R-6S and the 56- 6R isomers and we tested whether rEhDNAligI discriminates between them. rEhDNAligI and T4 DNA ligase efficiently ligated a thymine glycol 5R-6S isomer and a 56-6R isomer at the 3′ -OH position (**Figure 2C** lanes 1, 3 and 5, 7, respectively). However, rEhDNAligI and T4 DNA ligase were inefficient to perform a nick sealing reaction when thymine glycol is located at the 5′ - PO<sup>4</sup> (**Figure 2C** lanes 2, 4 and 6, 8, respectively). Finally, an abasic site occurs when the base is excised from the chain by spontaneous hydrolysis of the N-glycosylic bond. rEhDNAligI was unable to ligate an abasic site 3′ -OH of the nick (**Figure 2D** lanes 1, 2); however, T4 DNA ligase moderately catalyzed the reaction, when dATP was present opposite to the lesion (**Figure 2D** lane 5). rEhDNAligI efficiently ligated an abasic site 5 ′ -PO<sup>4</sup> to the lesion only when dATP is complementary to the lesion but not dCTP (**Figure 2D** lanes 3, 4). In contrast, T4 DNA ligase efficiently ligated an abasic site 5′ -PO<sup>4</sup> to the nick opposite a purine and a pyrimidine (**Figure 2D** lanes 7, 8).

We decided to further investigate the ligation capabilities of rEhDNAligI in comparison with commercial T4 DNA ligase, to perform nick sealing of different DNA lesions located at the "bottom" or "template" strand of a synthetic DNA nick containing thymine dimers. Thymine dimers as a cyclobutane

TABLE 1 | Steady-state kinetics constants for canonical bases and G:T mismatch.


pyrimidine dimer (CPD) or a 6-4 photo product pose a strong block to replicate DNA polymerases and are only bypassed by specialized DNA polymerases. Thus, we tested three possible substrates with different base modification related to thymine dimers (**Figure 3A**). Our results indicate that rEhDNAligI was able to ligate a CPD when this lesion is located between the DNA junction, but not when it is located at the 3′ -OH or 5′ -PO<sup>4</sup> (**Figure 3B** lane 2). In contrast T4 DNA ligase was also able to ligate a 6-4 photoproduct when the lesion was in the 3′ -OH of the junction (**Figure 3C** lanes 2, 4). Neither of both DNA ligases were able to seal a thymine dimer, when the lesion was present at the 3′ -OH side of the nick (**Figures 3B,C** lanes 3, 6).

# Survey of BER and NER Gene Encoding Proteins in *E. histolytica* Genome

E. histolytica encodes genes involved in the NER pathway, suggesting that this mechanism could be used by this parasite (López-Camarillo et al., 2009; Marchat et al., 2011). In our in silico analysis, we found orthologs for genes encoding the protein Cul4 and XPF and ERCC1 nucleases, which are involved in the release of DNA damage during the repair process (**Table 2**). Regarding BER pathway, we identified the monofunctional DNA glycosylase EhMUTY, EhAPEX, and EhFEN1 the strand resolving exonucleases (**Table 2**). Founding a homolog to the AP endonuclease APEX is relevant as previously it was reported the absence of a gene encoding an AP endonuclease in this parasite (López-Camarillo et al., 2009). All proteins from NER and BER pathways encoded in the E. histolytica genome described here, presented high percent of homology with proteins from H. sapiens, S. cerevisiae, and other eukaryotes (**Table S2**). Nevertheless, we found that some components from the BER pathway have high identity with proteins from bacteria suggesting that were acquired by lateral gene transfer (LGT).

# DNA Damage Induces Expression Changes of NER and BER Genes

In E. histolytica, UV and H2O<sup>2</sup> treatments have been used to produce oxidative stress, generating changes in gene expression (Vicente et al., 2009; Weber et al., 2009; Pearson et al., 2013). In order to investigate if BER and NER genes were regulated upon these treatments, we irradiated trophozoites with UV (150 J/m<sup>2</sup> ) or incubated them with H2O<sup>2</sup> (2.5 mM for 1 h at nonlethal conditions, as previously described (López-Casamichana et al., 2008; Shahi et al., 2016). Data showed that the viability of the trophozoites was 85% immediately after treatments, similar to those previously reported. However, between 1 and 2 h of recovery after treatments, we observed a decrease in the number of trophozoites (>50% of the viability) and eventually the viability was recovered after 3 h (**Figure S2**), suggesting that treatments produce DNA lesions that could be repaired allowing trophozoites to continue divide and multiply. First, in order to determine the expression patterns of NER and BER genes in E. histolytica, we performed RT-PCR assays of representative genes from each step of both pathways, in basal conditions. In basal culture conditions, trophozoites expressed all of the genes selected from the bioinformatics analysis (**Figure 4A**), suggesting the presence of both active DNA repair pathways in this parasite. We included minus reverse transcriptase controls, demonstrating that no genomic contamination was present in amplified genes. Additionally, positive PCR controls using genomic DNA were added, showing correspondence between the length of the amplicons obtained from genomic DNA and cDNA.

We also included, the non-template control which eliminates possible false-positive results. Then, we analyzed changes in expression of NER and BER genes immediately after UV and H2O<sup>2</sup> treatments using qRT-PCR. Results indicated that most of the NER genes increased their expression after irradiation treatment; ehr23.1, ehxpb, ehrpa, ehxpg, ehercc, and ehpcna genes presented 3.7, 3.6, 34, 2.8, 6.5, and 1.8-fold change, respectively compared with their basal condition (**Figure 4B**). Regarding genes of the BER pathway, only ehnth-like, ehpcna, and ehdnaligI increased their expression upon H2O<sup>2</sup> treatment exhibiting 1.5, 1.6, and 1.5-fold change, respectively compared with their basal condition (**Figure 4C**). Furthermore, upregulation of the NER and BER genes suggest the activation of the DNA repair pathways. As positive control, we included the ehmutS gene (NCBI sequence ID: XM\_647442) which encode for the DNA mismatch repair protein EhMutS. This gene was selected as a positive control because its expression is increased in response to both treatments in this parasite. The ehmutS gene is upregulated 1.55-fold after 3 h of recovery from UV injury and 2-fold during 1 mM H2O<sup>2</sup> treatment (Vicente et al., 2009; Weber et al., 2009). In our conditions this gene was also upregulated after 1 h of recovery from UV and 2.5 mM H2O<sup>2</sup> treatments, 2.4 and 2.1-fold respectively (**Figures 4D,E**), showing a response to DNA damage. Interestingly, ehdnaligI was upregulated immediately after H2O<sup>2</sup> treatment (**Figure 4C**), and in UV injury ehdnaligI exhibited 2.1 fold-change after 1 h recovery (**Figure 4D**). This data showed that DNA ligases participate at the final steps of DNA repair pathways, therefore is important evaluate its expression during recovery from DNA damage.

# EhDNAligI Induction During Recovery From UV and H2O<sup>2</sup> Treatments

To analyze the expression EhDNAligI protein in total extracts of trophozoites in basal condition and with both treatments at different times of recovery, we performed Western blot using a specific α-EhDNAligI antibody. This antibody detected a 75 kDa protein in trophozoite lysates (**Figures 5A,B**; **Figure S1A**) as previously described (Cardona-Felix et al., 2011). Immediately, after both treatments EhDNAligI level did not significantly raise, contrary to the expected we observed a decreased approximately 25–33%. An EhDNAligI increase of 30 and 44% at 6 h recovery of UV and H2O<sup>2</sup> insults, respectively, was noticed. This late response could indicate that there could be a more complex regulatory pathway that would require additional experiments to understand it. As positive control, we included the rEhDNAligI recombinant protein, where the α-EhDNAligI antibody recognized a 78 kDa band, corresponding to the predicted molecular weight plus the histidine tag. We also included the anti-human γ-H2AX antibody that detects serine-phosphorylated EhH2AX homologue as an indicator of DNA damage. We identified a 17 kDa band, which corresponds to the expected molecular weight of γ-EhH2AX histones in E. histolytica (López-Casamichana et al., 2008). Trophozoites exhibited low levels of γ-EhH2AX in basal condition and increased after 1 h of recovery from both treatments and then it was diminished, indicating recovery from DNA damage. As a loading control, the α-actin antibody was used, and the relative protein expression was calculated after normalizing the EhDNAligI levels over the actin control (**Figures 5C,D**). The observed increase of EhDNAligI in response to UV and H2O<sup>2</sup> treatments could be important during DNA repair following DNA insults.

# EhDNAligI Co-localizes in Nuclear Regions Where 8-oxoG Is Formed in Trophozoites Exposed to UV and H2O<sup>2</sup> Treatments

Here we have demonstrated that UV and H2O<sup>2</sup> treatments produced changes in EhDNAligI expression. Nevertheless, the effects of these agents on the DNA integrity in this parasite have not been determined. Even though we are using two different

#### TABLE 2 | Proteins from the NER and BER pathways in Entamoeba histolytica.


<sup>a</sup>Bedez et al., 2013, <sup>b</sup>Marchat et al., 2011, <sup>c</sup>López-Camarillo et al., 2009, <sup>d</sup>Alseth et al., 2006, <sup>e</sup>Cardona-Felix et al., 2010, <sup>f</sup> Cardona-Felix et al., 2011, <sup>g</sup>Pastor-Palacios et al., 2010. \*NCBI reference sequence database.

&According to AmoebaDB.

treatments, it is well documented that UVC radiation may generate ROS, which consequently induce DNA damage (Wei et al., 1997; Zhang et al., 1997). Guanine is attacked preferentially upon oxidative DNA damage because it has the lowest oxidation potential of the four bases, resulting in the formation of 7,8 dihydro-8-oxoguanine (8-oxoguanine; 8-oxoG). This lesion is abundantly produced in vivo and it is used as a biomarker of oxidative DNA damage (Shigenaga et al., 1990; Shigenaga and Ames, 1991). In order to verify whether UV and H2O<sup>2</sup> induced 8-oxoG formation in this parasite, we assessed 8 oxoG accumulation in E. histolytica trophozoites using avidinconjugated FITC. Avidin binds with high specificity to 8-oxoG and it has been used to detect oxidative DNA damage in different cell types, including parasites (Struthers et al., 1998; Furtado et al., 2012). In basal conditions, no avidin-FITC staining was observed in the nucleus of trophozoites (**Figures 6A**, **7A**). However, in trophozoites treated with UV or H2O2, the avidin-FITC signal was markedly observed at nuclei reaching the maximum 8-oxoG induction at 1 h recovey for both treatments, co-localizing with the DAPI staining (**Figures 6B**, **7B**). These results suggest that both treatments induced the formation of an 8-oxoG adduct, indicative of oxidative DNA damage. To explore possible cellular localization changes of EhDNAligI during and after DNA damage, we carried out immunofluorescence assays in trophozoites during and recovered from the UV and H2O<sup>2</sup> treatments, using the α-EhDNAligI antibody. In basal conditions,

confocal images evidenced the presence of EhDNAligI at the cytoplasm, whereas only a weak signal in the nucleus of the trophozoites was observed. Interestingly, immediately after UV treatment, EhDNAligI was translocated to the nuclear periphery and co-localized with the 8-oxoG adduct detected by the avidin-FITC, in some trophozoites (**Figure 6A**). After 1 h recovery, EhDNAligI was located at the nuclear periphery in the majority of the trophozoites, co-localizing with the avidin-FITC signal. From 3 to 6 h, EhDNAligI increased its presence at the cytoplasm and nucleus, and some trophozoites showed a diffuse staining in the whole nucleus not only at the periphery. At this time, a decrease in the nuclear DNA avidin-FITC intensity was observed. After a 24 h recovery, only a few trophozoites showed EhDNAligI at the nuclear periphery and it was more abundant at cytoplasm (**Figure 6A**), probably because the 8 oxoG was eliminated or repaired. The avidin-FITC intensity and the nuclear localization of EhDNAligI were quantified in several fields. After 1 h treatment, trophozoites exhibited the higher DNA damage which correlates with the presence of the phosphorylated histone γ-H2AX. Meanwhile, the mayor nuclear presence of EhDNAligI occurred after 6 h of recovery (**Figure 6B**).

Immediately, after H2O<sup>2</sup> treatment, EhDNAligI translocated from the cytoplasm to the nuclear periphery, in comparison with the basal condition. At 1 to 6 h of recovery, EhDNAligI and avidin-FITC signals co-localized at the nuclear periphery, however the avidin-FITC staining diminished during the recovery. After a recovery of 6 h, EhDNAligI diffusely localized into the nucleus and at 24 h its cytoplasmic and membrane localization indicated a basal localization pattern (**Figure 7A**). These nuclear localizations were corroborated by fluorescence quantification (**Figure 7B**). Data from UV and H2O<sup>2</sup> treatments are consistent with the protein changes observed by Western blot analysis. Altogether these findings suggest that both treatments induce 8-oxoG formation, which reflects the DNA damage was produced and also suggests that the 8-oxoG is being repaired. The decrease in trophozoites viability after treatments, and its recovery from 3 to 12 h, also reinforce the notion that 8-oxoG is being repaired (**Figure S2**). Furthermore, the EhDNAligI co-localization with the DNA damage suggests

that this enzyme could be involved in the DNA repair process.

# EhDNAligI Co-localizes With EhPCNA During Recovery From DNA Damage by UV and H2O<sup>2</sup> Exposure

In several systems, DNA ligase I is recruited to DNA damaged sites by its interaction with PCNA (Mortusewicz et al., 2006). Moreover, EhDNAligI activity is enhanced by EhPCNA and these proteins present a functional interaction in vitro (Cardona-Felix et al., 2011; Trasviña-Arenas et al., 2017). Thus, we tested whether during recovery from UV and H2O<sup>2</sup> treatments, EhDNAligI co-localizes with EhPCNA using an antibody generated against the rEhDNAligI recombinant protein (Cardona-Felix et al., 2010, 2011). By Western blot assays, this antibody recognized two bands in E. histolytica lysates, one of 28.5 kDa corresponding to the predicted molecular weight of EhPCNA, and another band of 75 kDa that may correspond to the trimeric ring assembly of EhPCNA (**Figure S1B** ). The same bands were immunodetected when the rEhPCNA recombinant protein was used, demonstrating the specificity of the antibody (**Figure S1B**). In basal conditions, EhPCNA showed poor colocalization with EhDNAligI, where EhDNAligI presented a predominantly cytoplasmic stain while EhPCNA presented a more diffused pattern into the nuclei (**Figures 8A**, **9A**). However, after DNA insults, EhDNAligI is relocated into the nucleus colocalizing with EhPCNA, with the higher colocalization rate immediately after UV treatment and 1 h recovery (**Figure 8B**) and 3 h recovery for H2O<sup>2</sup> injury (**Figure 9B**). Additionally, EhDNAligI co-localized with avidin at the nuclei after insults, at the same times that it colocalized with EhPCNA. The EhDNAligI and EhPCNA co-localization inside the nuclei was mainly observed in the nuclear periphery and in some nuclear foci-like structure (**Figures 8A**, **9A**). The colocalization of both proteins at nuclear foci-like structures during treatments recovery, suggest that EhPCNA recruits EhDNAligI toward DNA damage sites, indicating that EhPCNA is involved in targeting EhDNAlig to damaged sites but additional coimmunoprecipitation assays will necessary to strengthen the present results.

# DISCUSSION

E. histolytica genome is subject to continuous DNA damaging agents. In order to survive and propagate, the parasite must cope with them. DNA repair pathways converge on the step of nick sealing, and in higher eukaryotes specialized DNA ligases are involved either in DNA repair or replication. DNA ligases are sensitive to mispairs on the 3′ side of the nick, but tolerant to mispairs on the 5′ or acceptor side (Sriskanda and Shuman, 1998a,b). The base pair geometry of 3′ base pairs is recognized by hydrogen bond interactions at the minor groove between DNA ligase and DNA duplex (Levin et al., 2004). Here, we found that rEhDNAligI only efficiently ligates the T:G mismatch at the

5 ′ -phosphate strand, indicating a high discrimination level to non-canonical pairs. Thermus thermophilus and E. coli's DNA ligase are also able to ligase a T:G mismatch (Chauleau and Shuman, 2016; Lohman et al., 2016). The wooble T:G pair creates less steric constrictions and is one of the most common mistakes by DNA polymerase β (Beard and Wilson, 1998) indicating that at the minor groove both enzymes are able to T:G pair as C:G, maybe by a water mediated interaction between the O<sup>4</sup> of thymine and guanosine as the O<sup>2</sup> of a cytosine. Thus, the

biochemical data indicates that EhDNAligI is a high-fidelity DNA ligase. As this enzyme is the sole ligase present in the genome of E. histolytica, its high fidelity maybe a consequence of its dual role in DNA replication and DNA repair. Therefore, the EhDNAligI study is relevant as it could be a potential target for anti-parasitic agent.

Oxidative damage, alkylating agents, and UV light are the most prone agents to modify DNA and alter its coding capabilities (For reviews: Evans et al., 2004; Reardon and Sancar,

images show a xy-section. ph c: Phase-contrast. Arrows: co-localization of EhDNAligI and avidin-FITC. Squares were magnified in right panels. (B) Quantification of green (Avidin), red (EhDNALigI) and green/red fluorescence channels using DAPI staining as reference for the cell number.

2004). Nick-sealing fidelity in DNA ligases relates to the extent to which the enzyme can ligate substrates containing mismatched bases on either side of the nick. The mutagenic potential of DNA lesions has been extensively studied in DNA polymerases, however few studies have been carried out with DNA ligases. Reactive oxygen species originates 8-oxoG and thymine glycol. DNA ligase III efficiently sealed a Tg lesion in the 3′ OH (Budworth and Dianov, 2003), however its ligation efficiency was not tested at the 5′ OH. EhDNAligI and T4 DNA ligase efficiently ligate a thymine glycol at a 3′OH, however they are inefficient in sealing a nick at the 5′OH. Crystal structures of thymine glycol with DNA polymerases indicate that the methyl group of thymine glycol affects catalysis by altering the orientation of the 3 ′ OH (Aller et al., 2007). Human DNA ligase is able to ligate a dATP or dCTP opposite 8-oxoG at the 3′ OH; however, dATP is ligated more efficiently (Hashimoto et al., 2004). This contrasts

with rEhDNAligI, that discriminates a dCTP-8-oxoG pair. This discrimination is reminiscent to the one observed by family A of DNA polymerases that are more efficient in extending an 8 oxoG:dATP mismatch than an 8-oxoG:dCTP base pair (Brieba et al., 2004; Hsu et al., 2004). EhDNAlig I is unable to ligate an abasic site 3′ -OH of the nick, however T4 DNA ligase moderately ligates an AP site having dATP, but not dCTP, opposite the lesion. EhDNAlig I efficiently ligates an abasic site 5′ -PO4 to the lesion only when dATP is complementary to the lesion. In contrast, T4 DNA ligase efficiently ligates an abasic site 5′ -PO4 to the nick opposite a purine and a pyrimidine. This discrimination could be important for repairing abasic sites in the Entamoeba genome, as abasic sites are highly mutagenic and require a rapid and efficient repair. CPD and 6-4 photoproducts are not efficiently ligated by EhDNAligI. EhDNAligI was able to ligate a CPD when this lesion is located between the DNA junction, but not when it is at the 3′ -OH or 5′ -PO4, this correlates with the lack of binding for EhDNAligI observed by EMSA. In contrast, weak binding was observed in the 6-4 photoproduct when the lesion is present at the 3′ -OH or 5′ -PO4, indicating that thymine dimers distort the DNA helix and the catalytic step (**Figure S3**). Eukaryotic and archaeal DNA ligase contain an extra N-terminal domain dubbed DNA binding domain that considered essential to detect a properly organized substrate. The lack of a DBD in a T4 ligase may be involved in its low fidelity and its ability to use damaged bases as substrates (Zhao et al., 2007).

Different DNA repair systems are involved in the preservation of the genome by correcting DNA lesions originated by damaging agents, such as UV light and H2O2. DNA ligases are involved in BER and NER, DNA double-strand breaks, DNA non-homologous end-joining and homologous recombination pathways. In this work, we focused in BER and NER repair pathways, which can be induced by H2O<sup>2</sup> and UV light treatments, respectively. Even though different genes involved in NER and BER pathways have been identified in the E. histolytica genome (Alseth et al., 2006; López-Camarillo et al., 2009; Cardona-Felix et al., 2010; Pastor-Palacios et al., 2012; Bedez et al., 2013), in this study we extended these findings including

new genes that we identified in our survey. The E. histolytica genome encodes for proteins involved in the DNA damage recognition, as Rad23 and DDB1, which could be involved in UV excision repair. The gene encoding for Cul4 is also present in the E. histolytica genome; this protein is part of the core involved in DNA damage response. The nine subunits of the TFIIH complex and the nucleases XPF and ERCC1 are also encoded by the Entamoeba genome. In the case of genes encoding proteins for the BER pathway, it genome contains putative monofunctional (UDG and AlkD) and bifunctional (NTH) glycosilases (Loftus et al., 2005). Our survey revealed genes encoding for the MutY glycosilase and strand resolving exonucleases as Apex and Fen1. Interestingly, EhUDG and EhMutY glycosylases presented high identity with bacterial proteins, suggesting that the horizontal gene transfer had a role in their acquisitions (**Table S2**). Functional studies of EhMutY revealed that this enzyme, like its bacterial counterparts, is functional, even it lacks an iron-sulfur cluster (Trasviña-Arenas et al., 2016). Apparently, the genome of E. histolytica does not encode for proteins required in short-patch BER, as DNA polymerase β, XRCC1 and DNA ligase III; thus, we speculate that this parasite uses the long-patch BER to cope with the oxidative stress produced in the colonic tissue or during the amebic liver abscess. In both NER and BER pathways, DNA gaps are filled using the information available on the complementary strand by a DNA polymerase, that in E. histolytica could be a DNA polymerase with high amino acid identity to bacterial DNA polymerase I (Pastor-Palacios et al., 2010). Eukaryotes express three DNA ligases, dubbed I, III and IV. DNA ligase I is the main nuclear ligase, DNA ligase III is involved in DNA repair at the mitochondria and nucleus, and DNA ligase IV is involved in non-homologous end-joining (NHEJ) (Nash et al., 1997; Ellenberger and Tomkinson, 2008) and DNA ligase I can be substituted for DNA ligase IV in NHEJ (Lu et al., 2016). Because Entamoeba does not contain associated organelles with DNA like apicoplast or mitochondria is plausible that EhDNAligI is involved in all the nick-sealing reactions during DNA replication or repair of this parasite. Genome reduction is observed in protozoan parasites, for instance E. histolytica and Plasmodium falciparum code for half and one quarter of the proteins reported for H. sapiens. In these protozoan parasites only one DNA ligase

with similitude to DNA ligase I is reported (Buguliskis et al., 2007; Cardona-Felix et al., 2010). The lack of other ATP-dependent DNA ligases encoding genes involved in the DNA repair in the E. histolytica genome, suggests that EhDNAligI participates in BER, NER, and Okazaki fragment end-joining. But we cannot exclude the presence of noncanonical RNA ligases of the RtcB family that could directly ligate DNA 3′ -phosphate with DNA 5′ - OH ends (Das et al., 2013). Interestingly, E. histolytica genome encodes two proteins related the RctB family (NCBI sequence ID: XP\_648794.1 and XP\_651200.1). These proteins are annotated as hypothetical proteins and until now its RNA ligation function has not been characterized in this parasite. Remarkably, the genome of Naegleria gruberi, an amoeboflagellate, encodes for 13 canonical RNA ligases (seven from the Rnl1 family, five from the Rnl2 family, and one Rnl5 enzyme) and two noncanonical RNA ligases of the RtcB family. This plethora of RNA repair enzymes in this organism reflects a functional specialization of the ligase families for different types of DNA/RNA damages (Unciuleac and Shuman, 2015). In E. histolytica genome we didn't find encoded canonical RNA ligases using bacterial or Trypanosome RNAediting ligases as a bait, leaving only the RNA ligases capable of binding DNA fragments.

In E. histolytica, few DNA binding proteins like EhDNAligI, EhPCNA and EhMuyY have been characterized, however little is known about how these proteins interact during DNA repair [10, 12]. Here, we shown that representative genes of BER and NER pathways are expressed in trophozoites, under basal culture conditions. In order to evaluate expression changes of NER and BER genes during different injuries, we selected UV and H2O<sup>2</sup> treatments to produce DNA damage. Basal expression of NER genes was compared against trophozoites treated with UV irradiation. This treatment caused upregulation of most NER genes, but no changes in ehdnaligI transcript were observed. Nevertheless, during 1 h recovery ehdnaligI was upregulated. Regarding the 2.5 mM H2O<sup>2</sup> treatment, we found upregulation of the ehnthlike, ehpcna, and ehdnaligI genes. An explanation for this difference could due to UV treatment involved a short time period, in contrast to 2.5 mM H2O<sup>2</sup> injury which comprise 1 h incubation, suggesting that the induction of repair mechanisms was differentially activated. In another study, trophozoites treated for 1 h with 1 mM H2O2, 286 genes belonging to the detoxification systems were modulated; however, not significant changes in ehdnaligI or BER genes were observed (Vicente et al., 2009; Pearson et al., 2013). Even though, 1 mM of H2O<sup>2</sup> is a physiologically concentration present in the gastrointestinal lumen (Mayol et al., 2006), we do not know if during the oxidative stress in colon or hepatic abscess, the H2O<sup>2</sup> concentration is elevated and it produces expression changes in BER genes in response to DNA damage. Furthermore, we cannot discard modifications of BER genes or proteins or their recruitment to the DNA damage sites, during different DNA recovery treatments. Our findings suggest the presence and participation of NER and BER pathways during DNA damage. We included as a positive DNA damage control, the ehmutS gene, because in this parasite is upregulated during both DNA insults, as it has been widely studied for Escherichia coli (Vicente et al., 2009; Weber et al., 2009). Prokaryotic MutS and eukaryotic homologs are involved in the DNA mismatch repair (MMR) and have a major role in the mismatched DNA recognition; however, MMR has been also implicated in the response to oxidative response (Khil and Camerini-Otero, 2002; Mazurek et al., 2002; Germann et al., 2012). The 8-oxoG recognition systems involved several overlapping pathways common in bacteria, yeast, and human cells (Mazurek et al., 2002), thus we considered that ehmutS gene is a suitable positive control for determining induction of DNA damages responses in E. histolytica.

The ehdnaligI transcript was slightly upregulated only immediately after peroxide treatment, however its protein levels decreased after peroxide and UV treatments and increased from 1 to 6 h, reinforcing the idea about its participation during DNA repair. Similar results were obtained for DNA ligase I in fibroblasts exposed to UV irradiation by Montecucco and coworkers, where DNA lig I expression decreases immediately after treatment and then increases reaching a maximum 24 h after damaging treatment (Montecucco et al., 1992). Also, free radicals produced by UV and H2O<sup>2</sup> treatments caused oxidative modifications of proteins, leading to changes in their physical and chemical properties, including conformation, structure, solubility, susceptibility to proteolysis, and enzymatic activities (Davies, 2001, 2005). Recently, a large-scale oxidized proteins have been identified in oxidatively stressed trophozoites treated for 1 h with 2.5 mM H2O<sup>2</sup> (Shahi et al., 2016). This could explain the EhDNAligI diminish observed during UV and H2O<sup>2</sup> treatments, assuming that these injures generated physical and chemical changes in EhDNAligI. Then, at 3 and 6 h recovery from both treatments, it is possible that the ehdnaligI transcription be stimulated to increase the amount of protein observed.

In previous reports, E. histolytica trophozoites have been exposed to several H2O<sup>2</sup> concentration and UV irradiation (López-Casamichana et al., 2008; Vicente et al., 2009); however, oxidative damage of DNA has not been demonstrated. In this work, we determined that immediately after 2.5 mM H2O<sup>2</sup> and UV (150 J/m<sup>2</sup> ) light treatments the formation of the 8 oxoG adduct was induced, as a result of the DNA damage. The DNA damage was evaluated by avidin-FITC, which is widely used to assess the 8-oxoG accumulation due to structural similarities between biotin and 8-oxoG and the high affinity of avidin for biotin (Struthers et al., 1998). This approach has been also used in other parasites as T. cruzi to detect the adduct induction during treatment with 200µM of H2O<sup>2</sup> (Furtado et al., 2012). On the other hand, cells treated with UV light provide an experimental system for depicting the biological consequences of DNA damage. It has been previously demonstrated that irradiation of trophozoites with 254 nm UV light at 150 J/m<sup>2</sup> , induced double-stranded DNA breaks (López-Casamichana et al., 2008); however, formation of neither DNA photoproducts or 8-oxoG has been not determined. It is also reported that the 8-oxoG adduct is induced in human cells exposed to different wavelengths throughout the UV spectrum, causing formation of ROS through the generation singlet oxygen, H2O2, superoxide and hydroxyl free radicals (Batista et al., 2009). The genotoxic effect of UV light has been mainly attributed to the induction these oxidative species (Heck et al., 2003).The 8-oxoG, and other forms of oxidative damage, may play an important role in the induction of the biological effects caused by UV light treatment. In this work, we found that in E. histolytica both treatments (UV and H2O2) produced the 8 oxoG adduct, possibly due to free radicals that were formed during treatments. Further analyses are needed to quantify the ROS species generated by these treatments.

In human cells, specific recruitment of BER and NER components to DNA damage sites, where 8-oxoG, is present has been reported (Melis et al., 2013). Then, it is possible that in E. histolytica, when DNA is damaged, repairing proteins such as NER and BER components could be recruited to the lesion sites. Here, we observed co-localization of EhDNAlig and 8 oxoG, suggesting that this enzyme is recruited to the DNA damage sites. These results agree with the reported in HeLa cells treated with laser micro-irradiated or 5mM H2O2, where BER components (MUTYH glycosylase and DNA polymerase λ) co-localized with this injury (van Loon and Hubscher, 2009). DNA ligase I is recruited to DNA repair sites by interaction with PCNA. The interaction between PCNA and DNA ligase I is not only critical for the subnuclear targeting of the ligase, but also for coordination of the molecular transactions that occur during lagging-strand synthesis (Mortusewicz et al., 2006). This interaction is mediated by a conserved peptide interacting motif (PIP box) present at the N-terminal region of EhDNAligI (Cardona-Felix et al., 2010). Additionally, it has been demonstrated that EhPCNA assembles as a homotrimer to interact with and stimulate EhDNAligI (Cardona-Felix et al., 2011). Our microscopy observations revealed co-localization of these proteins at nuclei in the same time that 8-oxoG lesion co-localized with EhDNAligI, suggesting a direct participation of EhDNAligI and EhPCNA during DNA repairing. However, elucidate whether they are participating in both BER and NER pathways remains to be analyzed. The biological importance of the 8-oxoG adduct formation is due to its propensity to mispair with adenine residues, leading to an increased frequency of spontaneous G:C to T:A transversions. This fact may predispose E. histolytica to evolve toward an AT rich genome. Although BER remains as the preferred pathway for repairing the 8-oxoG lesion, NER factors contribute to repair this lesion (Parlanti et al., 2012). Thus, the interaction of EhDNAligI and EhPCNA proteins could be an indicator of the possible activation of NER and BER pathways in E. histolytica, however co-localization assays of EhDNAligI with proteins involved in these routes, could clarify the ligase participation. The cytoplasmic abundance of this enzyme could indicate another possible role in nicking-closing reaction not only in nucleus but also in cytoplasm, in response to mutagenic effects.

Taken altogether our results showed that EhDNAligI is a highfidelity DNA ligase in comparison to T4 DNA ligase. EhDNAligI also discriminated erroneous base pairs from opposite DNA lesions. EhDNAligI decreased after DNA insults but it recovered at 6 h and co-localized with EhPCNA, where DNA lesions remain. Additional studies are needed to better understand the role of EhDNAligI or other NER and BER components during DNA repair and recombination processes in this parasite.

# AUTHOR CONTRIBUTIONS

All authors contributed equally to design and conception of this work. CC-F, GP-P, CT-A, CD-Q, and LB conducted the biochemical experiments. EA-L, AB, EC-O, RC-G, GG-R, and DH-Á collected E. histolytica experimental data. HC performed the in silico analysis. EA-L, AB, EO, and LB contributed to experimental design, intellectual input, interpreting data, and in writing the manuscript.

# FUNDING

LB thanks the Miguel Aleman foundation for a research grant to study DNA ligases. EA-L thanks CONACYT for a research grant to study proteins that preserve genome integrity in Entamoeba histolytica (222956).

# ACKNOWLEDGMENTS

HC thanks fellowships BI 080530161142 and ICyTDF/179/2011. We thank Professor Shigenori Iwai for thymine glycol and thymine dimer substrates and Dr. Cesar López-Camarillo for reading the manuscript and providing constructive comments. The authors are grateful to Alfredo Barberi for graphical design. Authors also thank Karen Becerril Puente and Brenda Herrera Villalobos for technical assistance.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00214/full#supplementary-material

Figure S1 | Immunoreactivity of the antibodies generated against rEhDNAligI and rEhPCNA proteins. Trophozoites extracts, rEhDNAligI (A) and rEhPCNA (B) were separated in SDS-PAGE gels. Gels were Coomassie stained or processed for Western blot analysis, using α-EhDNAligI (A) or α-EhPCNA (B) antibodies. TE: total cell extracts from E. histolytica trophozoites. Black arrowheads: molecular weight predicted size of the detected proteins, 28.5 and 75 kDa for EhPCNA and EDNAligI, respectively. Gray arrowhead: trimeric form of rEhPCNA. Numbers at the left: molecular weight of standards. <sup>∗</sup>EhDNAligI degraded products.

Figure S2 | Viability of E. histolytica trophozoites exposed to UV and H2O2 treatments. Trophozoites treated with (A) UV (150 J/m<sup>2</sup> ) or (B) H2O2 (2.5 mM for 1 h) were incubated in TYI-S-33 medium for different times and then stained with Trypan blue dye. The number of living trophozoites was counted in a counting chamber under a light microscope.

Figure S3 | EhDNAligI is able to bind to DNA damaged substrates. EMSA experiments with radiolabeled DNA substrate containing lesions at the 3′ -OH end or 5′ -PO4 end in comparison to a canonical substrate using increasing concentrations of EhDNAlig I (26, 52, and 104 fmol) and 300 fmol of the indicates DNA substrates. Reaction mixtures were loaded onto a 6% native polyacrylamide gel and revealed by phosphorimagery. The position of the EhDNAligI-DNA substrate complexes and the free substrate are indicated by arrows. (A) EMSA experiments using an undamaged DNA substrate. Lane 1 shows the migration of radioactive probe in the absence of EhDNAligI. (B) EMSA experiments with a CPD the lesion at the 3′ -OH (lanes 1–3) or 5′ -PO4 end (lanes 4–6) and covering both positions (lanes 7–9). (C) EMSA experiments with a 6–4 PP DNA lesion, the lines are represented as in (B). (D) EMSA experiment with thymine glycol 5R-6S isomer (lanes 1–6) or a 56-6R isomer (lanes 7–12). The damaged bases are located either at the 3′ -OH (lanes 1–3 and 7–9) or at the 5′ -PO4 end (lanes 4–6 and 10–12).

(E) EMSA experiment with an abasic site. The nucleotide opposite to the abasic site is either a cytidine (lanes 1–6) or an adenine (lanes 7–12). The damaged base is located as in (C). (E) EMSA harboring an 8-oxo guanosine adduct. The nucleotide opposite to the 8-oxo guanosine is either a cytidine (lanes 1–6) or an adenine (lanes 7–12).

# REFERENCES


Table S1 | Paired primers used for assessing the transcript levels of the NER and BER genes in E. histolytica.

Table S2 | Comparison of NER and BER proteins from E. histolytica, H. sapiens, S. cerevisiae, and other organisms.


of Entamoeba invadens. J. Eukaryot. Microbiol. 41, 360–365. doi: 10.1111/j.1550-7408.1994.tb06090.x


chromatography with electrochemical detection. Meth. Enzymol. 186, 521–530. doi: 10.1016/0076-6879(90)86146-M


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Azuara-Liceaga, Betanzos, Cardona-Felix, Castañeda-Ortiz, Cárdenas, Cárdenas-Guerra, Pastor-Palacios, García-Rivera, Hernández-Álvarez, Trasviña-Arenas, Diaz-Quezada, Orozco and Brieba. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# A Calpain-Like Protein Is Involved in the Execution Phase of Programmed Cell Death of *Entamoeba histolytica*

Tania Domínguez-Fernández <sup>1</sup> , Mario Alberto Rodríguez <sup>1</sup> , Virginia Sánchez Monroy <sup>2</sup> , Consuelo Gómez García<sup>2</sup> , Olivia Medel <sup>2</sup> and David Guillermo Pérez Ishiwara<sup>2</sup> \*

<sup>1</sup> Departamento de Infectómica y Patogénesis Molecular, CINVESTAV, Ciudad de México, Mexico, <sup>2</sup> Programa de Biomedicina Molecular, Escuela Nacional de Medicina y Homeopatía (ENMyH), Instituto Politécnico Nacional, Ciudad de México, Mexico

Oxygen or nitrogen oxidative species and chemical stress induce the programmed cell death (PCD) of Entamoeba histolytica trophozoites. PCD caused by the aminoglycoside G418 is reduced by incubation with the cysteine protease inhibitor E-64; however, no typical caspases or metacaspases have been detected in this parasite. Calpain, a cysteine protease activated by calcium, has been suggested to be part of a specific PCD pathway in this parasite because the specific calpain inhibitor Z-Leu-Leu-Leu-al diminishes the PCD of trophozoites. Here, we predicted the hypothetical 3D structure of a calpain-like protein of E. histolytica and produced specific antibodies against it. We detected the protein in the cytoplasm and near the nucleus. Its expression gradually increased during incubation with G418, with the highest level after 9 h of treatment. In addition, a specific calpain-like siRNA sequence reduced the cell death rate by 65%. All these results support the hypothesis that the calpain-like protein is one of the proteases involved in the execution phase of PCD in E. histolytica. The hypothetical interactome of the calpain-like protein suggests that it may activate or regulate other proteins that probably participate in PCD, including those with EF-hand domains or other calcium-binding sites.

Keywords: *Entamoeba histolytica*, cysteine proteases, programmed cell death, protein overexpression, calcium binding sites, calpain

# INTRODUCTION

Programmed cell death (PCD) plays crucial roles in a multitude of physiological processes, such as embryogenesis, aging, and maintenance of cell populations in tissues (Ameisen, 1996). It involves the activation of a group of calcium-dependent cysteine proteases called "caspases" and a complex cascade of events characterized by distinct morphological and biochemical changes (Elmore, 2007). In Entamoeba histolytica, PCD had been induced in vitro by nitric oxide (Ramos et al., 2007), hydrogen peroxide (Nandi et al., 2010) and the aminoglycoside G418 (Villalba et al., 2007). Amoeba PCD is characterized by some typical biochemical and morphological changes described in other organisms, such as a considerable increase in cytosolic calcium, a reduction in intracellular potassium, intracellular pH acidification and chromatin condensation (Villalba et al., 2007). The PCD phenotype diminish when trophozoites are incubated with E-64, a specific cysteine protease inhibitor; nevertheless, genes encoding for caspases have not been identified in E. histolytica (Villalba et al., 2007). Interestingly, Nandi et al. (2010) demonstrated that the activity of calpain-like

#### *Edited by:*

Brice Rotureau, Institut Pasteur, France

#### *Reviewed by:*

Michael Duchene, Medizinische Universität Wien, Austria Friedrich Frischknecht, Universität Heidelberg, Germany Marcos L. Gazarini, Federal University of São Paulo, Brazil

#### *\*Correspondence:*

David Guillermo Pérez Ishiwara ishiwaramx@yahoo.com.mx

#### *Specialty section:*

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

*Received:* 24 May 2018 *Accepted:* 06 September 2018 *Published:* 25 September 2018

#### *Citation:*

Domínguez-Fernández T, Rodríguez MA, Sánchez Monroy V, Gómez García C, Medel O and Pérez Ishiwara DG (2018) A Calpain-Like Protein Is Involved in the Execution Phase of Programmed Cell Death of Entamoeba histolytica. Front. Cell. Infect. Microbiol. 8:339. doi: 10.3389/fcimb.2018.00339 increase during H2O2-induced PCD. Moreover, a specific inhibitor of calpain activity (Z-Leu-Leu-Leu-al) decrease DNA fragmentation and increase cellular viability of trophozoites during G418-induced PCD (Sanchez-Monroy et al., 2015).

Calpains, a group of non-lysosomal Ca2+-dependent cysteine proteases, have been identified in almost all eukaryotes and bacteria, but not in archaebacteria. Among the 15 members of the calpain family found in human, the ubiquitous calpains 1 and 2 are the most intensely studied. Calpains have been involved in various physiological processes such as cell proliferation, cell cycle progression (Glading et al., 2002), signal transduction (Carafoli and Molinari, 1998), cell migration, cytoskeletal remodeling (Zhang et al., 2011) and in the regulation of cell death (Squìer et al., 1994; Tagliarino et al., 2001; Harwood et al., 2005). In fact, calpain was the first protease identified in initiating apoptosis (Squìer et al., 1994). Several studies have notably highlighted how closely these proteases are linked to caspases. Calpains 1 and 2 cleave several members of caspase family, activating the caspase-3,−7, and−12 and inactivating the caspase-8 and -9 (Chua et al., 2000; Nakagawa and Yuan, 2000). By regulating caspases, calpains can thus control indirectly apoptosis. Also, in situations of mass calcium influx, membrane transection or ischemia/reperfusion injury, the ubiquitous calpains are activated and in turn trigger caspase-3 (Wang, 2000).

In addition to the typical morphological events related to nuclear and membrane changes (Kerr et al., 1972), apoptosis accompanies a dramatic reorganization of the cytoskeleton due to the selective proteolysis of vital cellular substrates. Thus, regulated proteolysis by calpain is required for the control of fundamental cellular processes including cytoskeletal remodeling, and activation of proteolytical cascades leading to apoptosis (Saido et al., 1994).

Calpains are heterodimeric proteins, consisting of two subunits of 80 and 28 kDa (Croall and Ersfeld, 2007). The large subunit in classical calpains consists of four conserved domains: An N-terminal anchor helix (Nter), a catalytic protease core domain (CysPc) with the two subdomains PC1 and PC2, a C2 like domain (C2L), and a penta-EF-hand domain (PEF). Nonclassical calpains lack both the Nter and the PEF domain and may contain additional domains in combination with CysPc (Hosfield et al., 1999; Strobl et al., 2000; Joyce et al., 2012).

Atypical or unconventional calpains are described as calpainlike proteins that contain only a CysPc consensus signature with variations in the catalytic triad and no PEF-containing domain is present, and they may also contain additional domains in combination with CysPc (Sorimachi et al., 2010). Calpainlike proteins have mainly been found in invertebrates and lower eukaryotes. In Trypanosoma and Blastocystis hominis, calpain-like proteins have been involved in the life cycle, the differentiation process and the regulation of PCD (Hertz-Fowler et al., 2001; Giese et al., 2008; Yin et al., 2010). In Entamoeba histolytica upregulation of calpain-like gene very early during PCD induction, correlates with the release of cytosolic calcium (Villalba et al., 2007; Sanchez-Monroy et al., 2015) and calpain activity increased after 6 h of PCD induction (Sanchez-Monroy et al., 2015).

In this study, we modeled a hypothetical 3D structure of the calpain-like protein of E. histolytica based on the conserved domains previously identified in the primary protein sequence (Sanchez-Monroy et al., 2015). By Western blot (WB) and confocal microscopy analyses, we demonstrated that the expression of calpain-like protein increased during PCD induction, localizing the protein in the cytoplasm and near the nucleus. Knockdown of the calpain-like gene by a specific small interference RNA sequence (siRNA) provoked a 65% decrease in PCD. The results presented here support the hypothesis that the calpain-like protein plays an important role in the execution phase of E. histolytica PCD. In addition, a hypothetical interactome of the calpain-like protein suggests that other proteins, including some with calcium-binding domains, also participate in the PCD pathway of this parasite.

# MATERIALS AND METHODS

# *Entamoeba histolytica* Culture

Trophozoites of clone A (Orozco et al., 1983), which is a virulent subclone of strain HM1: IMSS, were axenically cultured in TYI-S-33 medium at 37◦C, harvested at a logarithmic growth phase, as described (Diamond et al., 1978). PCD was induced by exposure to 10µg/ml G418 for different periods of time, as indicated.

# *In silico* Analysis of the *E. histolytica* Calpain-Like Protein

The sequence of the calpain-like protein from E. histolytica was obtained from NCBI (XP\_657312.1) (EHI\_045290). Then, a hypothetical 3D structure was predicted by the I-TASSER program (Protein Structure and Function Predictions). The UCSF CHIMERA program was used to compare the tertiary structure of the calpain-like from E. histolytica with the tertiary structure of the calpain-1, (mu/I) large subunit [Homo sapiens] (AAH08751) and the PyMOL software tools were used to visualize the preserved calpain-like domains (Sanchez-Monroy et al., 2015). The similarity of the proteins included in our study was compared with the available protein homologs against non-redundant databases like BLASTP program of NCBI.

# qRT-PCR Assays

Total RNAs from trophozoites without treatment or treated with 10µg/ml G418 for 0.5, 1.5, 3, 6, and 9 h were isolated using the Trizol reagent (Invitrogen) according to the manufacturer's protocol. cDNA was synthesized using an oligo(dT) primer and the Superscript II reverse transcriptase (Invitrogen). For qRT-PCR assays, specific primers for the calpain-like gene (XM\_652220.1) were designed by the Primer Express Software for Real-Time PCR version 3.0 (Applied Biosystems), (sense primer: 5′ -GTTTCAATATCACAACCTCGTTGTG-3′ and

**Abbreviations:** PCD, programmed cell death; PBS, phosphate-buffered saline; WB, Western blot; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; NRS, non-related sequence; CaM, calmodulin; CaMKIV, calcium/calmodulindependent protein kinase IV; AR, androgen receptor; LDCD, lysosome-dependent cell death; ROS, reactive oxygen specieS; LMP, lysosomal membrane permeabilization.

antisense primer: 5′ -AAAGTCTCTCCAGAATCACCTCCA-3′ ). As an internal control, we used specific primers for the gapdh gene (XM\_651327.2, XM\_645264.2 and XM\_649267.2) (sense 5′ - CCGTCCACAGACAATTCGAA-3′ ; antisense 5′TTGAGCTG GATCTCTTTCAGCTT-3′ primers). Reactions were performed in the ABI PRISM 7000 Sequence Detection System (Applied Biosystems) by monitoring the real-time increase in fluorescence using the SYBR Green PCR Master Mix (Applied Biosystems). The relative quantification was calculated using the CT method, which uses the formula 2−11CT (Livak and Schmittgen, 2001). Statistically significant differences in gene expression between non-induced and PCD-induced trophozoites were analyzed by comparisons of the means of three independent biological replicates in triplicate using the Tukey's test with GraphPad Prism statistical software version 6.0.

# Production of Calpain-Like Antibodies

Antigenic peptides from calpain-like protein were analyzed by the ABCpred program (Saha and Raghava, 2006). Peptides with higher antigenic scores were used as probes for BLAST search of the E. histolytica genome. The CCEWKGKWRDDDPAWT polypeptide, situated at positions 230–246, was synthesized coupled to the KLH (Keyhole Limpet Hemocyanin) tag to increase its immunogenicity (GL Biochem). Four BALB/c mice were immunized by the subcutaneous and intramuscular routes using 50 µg of the polypeptide emulsified in Titer-Max Gold adjuvant (1:1) (Sigma). Then, the animals were immunized with two more doses (100 µg) of the polypeptide resuspended in the adjuvant at 15-day intervals followed by bleeding to obtain antibodies. The pre-immune serum was obtained before the first immunization.

The experimental protocol was approved by the institutional committee for animal care and provided all technical specifications for the production, care and use of laboratory animals (NOM-062-ZOO-1999).

# Western Blot Assays

Trophozoites (2 × 10<sup>7</sup> ) without treatment or treated with 10µg/ml G418 for 0.5, 1.5, 3, 6, and 9 h were harvested and washed twice with cold PBS. For the extraction of total proteins, cells were hypotonically lysed for 20 min at 4◦C in the presence of a mixture of protease and phosphatase inhibitors (PMSF, 1 mM; leupeptin, 10µM; N-ethylmaleimide, 25 mM; PHMB, 2.5 mM; E-64, 5µM; Na3VO4, 1 mM; NaF, 50 mM; iodoacetamide, 5 mM) and 1X Complete Protease inhibitor (Roche). Proteins were separated by 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), transferred to nitrocellulose membranes, and non-specific binding sites were blocked with 5% fat-free milk. Membranes were incubated with calpain-like (dilution 1:3,500) or GAPDH (dilution 1:5,000, Santa Cruz Biotechnology) antibodies. Afterwards, membranes were incubated with a secondary HRP-labeled antibody (Invitrogen). Finally, protein bands were visualized by chemiluminescence (ECL, GE Healthcare, United Kingdom), and the MicroChemi system (Biostep). Pre-immune serum was used as a control.

# Immunofluorescence and Confocal Microscopy

After treatments in test tubes (brand Pyrex), trophozoites (1 × 10<sup>6</sup> ) were placed on cover slides for approximately 20 min, fixed and permeabilized with 100% methanol for 5 min. Non-specific binding sites were blocked with 1% horse serum for 1 h at 37◦C. Cell slides were incubated overnight at 4◦C with <sup>α</sup>-calpain-like antibodies (1:40), rinsed with PBS three times, incubated with anti-mouse secondary antibody conjugated to Alexa 488 (1:400) for 1 h at room temperature and rinsed again three times with PBS. Nuclei were stained with 4′ ,6-diamidino-2-phenylindole (DAPI), and samples were observed in a confocal microscope (Carl Zeiss LSM 700) using the ZEN 2009 software. About 10 amoebas were chosen at random from each slide, fluorescence of each amoeba represents the intensity mean value of 25 optical sections from the top to the bottom of each cell. The images were representative of three independent experiments.

# Knockdown of the Calpain-Like Gene

The complete mRNA sequence (XM\_652220.1) of the calpainlike gene was analyzed by the online program Target finder (https://www.genscript.com/tools/sirna-target-finder) to obtain potential small interference RNA (siRNA) sequences, which were evaluated by a nucleotide BLAST on NCBI (https://blast.ncbi. nlm.nih.gov). A specific siRNA sequence of the calpain-like mRNA corresponding to nucleotides 6-31 (sense: 5′ - UACUGA CGAUGAAUUUCCAGCUGA-3′ ; antisense: 5′ - UUCAGCUGG AAAUUCAUCGUCAGUA−3 ′ ) was synthesized (Ambion). As a negative control, an additional siRNA sequence of non-related sequence (NRS) using the following set of sense and antisense primers was synthesized (sense: 5′ -CAAGCUGACCCUGAAGU UCdTdT−3 ′ ; antisense: 5′ -GAACUUCAGGGUCAGCUUGdTd T-3′ ).

The uptake of siRNAs by trophozoites was carried out by the soaking method as previously described (Ocádiz-Ruiz et al., 2013). Briefly, trophozoites (1 × 10<sup>6</sup> ) collected from 90% confluent cultures were inoculated in 25-ml culture plastic flasks (Corning) containing TYI-S-33 medium and incubated at 37◦C for 24 h. Then, the double-stranded siRNAs calpain-like or NRS sequences (50 nM) were added to the cultures and incubated at 37◦C for 24 h. To confirm the knockdown of the calpain-like gene, qRT-PCR and WB assays were performed as described above.

# Terminal

# Deoxynucleotidyltransferase-Mediated dUTP Nick-End Labeling (TUNEL) Assays

Trophozoites were fixed in 4% formaldehyde for 2 h at 4◦C. After the samples were washed twice with PBS, 50 µl of TUNEL reaction mixture (Roche) was added and incubated for 60 min at 37◦C in a humidified atmosphere in the dark. Then, trophozoites were rinsed four times with PBS, and nuclei were counterstained with DAPI and mounted with VECTASHIELD (Vector Laboratories). The samples were observed through a confocal microscope (Carl Zeiss LSM 700) using the ZEN 2009 software. As a positive control, trophozoites were treated with

20 mg/ml DNase I endonuclease (Invitrogen) for 30 min, and non-treated trophozoites were used as a negative control.

# Flow Cytometry

To determine the effect of calpain-like gene silencing on the viability of trophozoites subjected to PCD, cells were stained with propidium iodide (PI) and analyzed by flow cytometry. Briefly, after 24 h of treatment with the siRNA, trophozoites were incubated for 9 h with G418; then, cells were suspended in 1 ml of PBS and incubated for 5 min on ice in the dark with 1 mg/ml PI (Invitrogen). Finally, fluorescence was examined in a BD LSRFortessa cell analyser (BD Biosciences).

# Bioinformatic Analysis

The STRING 10.5 protein-protein interaction database (https:// string-db.org/) (Szklarczyk et al., 2015) was used to determine the potentially involved known and predicted protein networks. For these analyses, we used a medium confidence value and non-filtered disconnected nodes.

# Statistical Analysis

Statistically significant differences in WB, immunofluorescence and trophozoites viability were analyzed by comparisons of the means of three independent biological replicates by ANOVA using the GraphPad Prism statistical software version 6.0.

# RESULTS

# Identification of Conserved Domains in the Protein Architecture and 3D Structure of Calpain-Like

The I-TASSER program database was used to predict the threedimensional molecular structure of calpain-like (XP\_657312.1) from E. histolytica (**Figure 1**, pink structure) and human (**Figure 1**, blue structure) calpain-1 protein obtained from NCBI (AAH08751). One of the templates used by I-TASSER to predict both structures was a catalytic domain of human calpain-1 crystal (PDB ID: 2ARY). Using the UCSF CHIMERA, we obtained a 30.4% identity between the complete proteins. Furthermore, the 3D structure of catalytic domain of human calpain-1 (the domain II that contain amino acid residues that form the catalytic triad were identified and indicated in yellow by the PyMOL software tools) showed a higher identity (34.88%) (**Figure 1**). The BLASTP analysis of calpain-1 (mu/I) large subunit [Homo sapiens] yielded 33% identity with calpain-like of E. histolytica.

FIGURE 2 | Expression and localization of the calpain-like protein in trophozoites. (A) Total extracts of E. histolytica were separated by 12% SDS-PAGE and analyzed by WB assays using pre-immune serum (PS, 2) or mouse α -calpain-like antibody (3). FPL-007 Flash protein ladder molecular marker (Gel company) (1). (B) Representative image of laser confocal microscopy of methanol-fixed trophozoites using mouse α-calpain-like antibody. PS, pre-immune serum.

FIGURE 1 | 3D structures of the calpain-like protein of E. histolytica and human calpain-1. The diagram represents the 3D structures of human calpain (blue) and the calpain-like protein of E. histolytica (pink). Upper panels show the 3D models of the complete calpains. Lower show the 3D structures of the domain II of both calpains. Right shows the overlapping in the 3D calpain structures, the numbers indicate the percentage of identity. The amino acids that make up the catalytic triad of calpain are represented in yellow.

# Expression and Location of Calpain-Like Protein Within Trophozoites

By performing WB analysis on trophozoite extracts using specific antibodies against calpain-like protein, we detected a single 53 kDa band, whereas the pre-immune serum did not recognize any band (**Figure 2A**). Immunofluorescence assays using the same antibodies showed that the calpain-like protein is located throughout the cytoplasm in non-treated trophozoites (parasites cultured in TYI-S-33 medium, without G418) (**Figure 2B**).

# Expression of the Calpain-Like Protein During PCD Induced by G418

Calpain-like protein expression increased slightly in a timedependent manner (1.7-, 2.3-, 3.0-, and 3.5-fold) after 1.5, 3, 6, and 9 h of PCD induction by G418, respectively (**Figure 3A**). On the other hand, the distribution and location of the calpain-like protein in trophozoites during the PCD process was almost the same as in un-treated trophozoites, but the signal was increased, with the maximum expression at 9 h. Interestingly, at 6 and 9 h, the calpain-like protein was also re-located closer to the nucleus (**Figure 3B**).

# Silencing of the Calpain-Like Gene

To evaluate the role of the calpain-like protein in the execution phase of PCD, we silenced its gene expression through a small interference RNA (siRNA) sequence. First, using qRT-PCR, we analyzed the effect of specific calpain-like silencing with 50 nM/ml siRNAs for 16 and 24 h. The results showed 37 and 67% decrement of the calpain-like gene expression after 16

and 24 h, respectively (**Figure 4A**). As expected, incubation with NRS siRNA sequence for the same time (24 h) did not affect calpain-like gene expression (**Figure 4A**). By WB we observed that calpain-like protein expression increased 3.0- and 3.5-fold times in control trophozoites induced to PCD or in trophozoites treated with NRS siRNA sequence, respectively, in comparison to the basal expression of control trophozoites. Interestingly, PCD induced trophozoites treated with calpain-like siRNA sequence significantly reduced (80%) the calpain-like protein expression (**Figure 4B**).

# Effect of Calpain-Like Silencing on PCD

The trophozoites incubated with the siRNA for 24 h and subsequently with G418 for 9 h were analyzed by TUNEL assay (In Situ Cell Death Detection Kit, AP). The results showed that silencing of the calpain-like gene caused a considerable decrease in positive tunnel staining compared to treatment with G418 alone for 9 h or NRS treatment for 24 h and PCD induction for 9 h (**Figure 5A**). Densitometric analysis of three independent experiments, showed that calpain silencing trophozoites displayed a significant reduction of fluorescence (65%) in comparison to PCD induced trophozoites or those treated with NRS siRNA sequence (**Figure 5B**). Consistently, silencing the calpain-like gene expression increased 90% cell viability of trophozoites after 9-h PCD induction, while trophozoites treated only with G418 for 9 h or treated with NRS followed by a 9-h PCD induction, cell viability was approximately 63% (**Figure 5C**).

# Searching for Putative Protein Interaction Networks of Calpain-Like Protein

We analyzed the calpain-like protein in the STRING 10.5 database to create a putative network of proteins that interact with it. To make the prediction, we selected the interactome based on text mining, experiments, databases, coexpression, neighborhood, gene fusion, and co-occurrence. The results suggested that 20 proteins could have several possible interactions among them with a p-value of 1.23e-07 (**Figure 6**). This interactome contains proteins related to endocytosis, such as vesicular trafficking proteins, proteins related to vacuolar transport, proteins with zinc-finger domains related to DNA binding, and proteins with EF-hand calcium binding domains, such as grainins 1 and 2 (**Table 1**).

# DISCUSSION

Programmed cell death involves a complex cascade of events characterized by distinct morphological and biochemical changes triggered by a group of cysteine proteases (Samali et al., 1999). In eukaryotes, the expression and activation of cysteine proteases such as caspases, metacaspases, or calpain occur when the cytosolic Ca2<sup>+</sup> concentration increases (Elmore, 2007). Calpains, enzymes belonging to the family of calcium-dependent cysteine proteases, have also been implicated in pro-apoptotic pathways by the cleavage of apoptosis-associated proteins, such as caspase 7 (Gafni et al., 2009), Bax, Bcl-2 (Gao and Dou, 2000), Jun and Fos (Hirai et al., 1991), caspase 12 and Jnk (Tan et al., 2006). Several

calpain-like gene from trophozoites incubated with 50µg/ml of the calpain-like siRNA sequence for 16 and 24 h. As control, trophozoites were incubated with a non-related sequence (NRS). Negative control represents trophozoites without siRNA incubation. (B) WB of calpain-like protein. (A) Negative control, untreated trophozoites; (B) Trophozoites after a 9-h incubation with 10µg/ml of G418; (C,D) Trophozoites pre-incubated for 24 h with NRS (C) or calpain-like (D) siRNAs sequences and incubated 9 h with 10µg/ml G418. The graph shows the densitometry analysis of calpain-like protein expression levels. \*indicate statistically significant difference (P < 0.05).

findings in higher eukaryotes have suggested a role of calpains in the execution phase of PCD. For instance, an inhibitor of calpains has been reported to prevent apoptosis of glial cells induced by

silibinin, a natural polyphenolic flavonoid (Jeong et al., 2011). In addition, calpain-2 has been demonstrated to play a crucial role in hydrogen peroxide-induced apoptosis in pancreatic AR42J cells (Hiemer et al., 2017).

E. histolytica has no canonical caspases. However, PCD is inhibited by E-64, a specific inhibitor of cysteine proteases (Villalba et al., 2007), and calpain-like activity increases when trophozoites are induced to PCD by nitric oxide species (Nandi et al., 2010). Typical calpains contain four structural domains: domain I, which is cleaved after Ca2<sup>+</sup> activation; domain II, which contains the active site (the catalytic triad of cysteine, asparagine and histidine) conserved throughout the family and residues that can bind two Ca2<sup>+</sup> atoms implicated in the enzymatic activation; domain III, which contains a phospholipidbinding motif in the area of C2 (Clan 2 within the classification of the cysteine protease family); and domain IV, which contains five EF-hand motifs that bind Ca2<sup>+</sup> (Smith and Schnellmann, 2012). Interestingly, typical caspases have not been identified in E. histolytica, but a caspase-like enzyme has been suggested to be involved in PCD (Sanchez-Monroy et al., 2015). In this work, we predicted the 3D structure of the calpain-like protein of E. histolytica, and within this structure, we located domain II, containing the catalytic triad residues (Cys57, His206, and Asn226). Despite being phylogenetically distant from typical calpain, domain II of the calpain-like protein of E. histolytica shows a 34.88% identity with the same domain of human calpain-1.

Using a specific antibody against the calpain-like protein, we detected a 53-kDa band that displayed a time-dependent induction during PCD, obtaining a maximum expression after 9 h of incubation with G418. This fact is in concordance with the enzymatic activity, which showed the highest calpain activity at longer times of PCD induction (Sanchez-Monroy et al., 2015).

Subcellular localization of calpain-like protein by confocal microscopy showed that this enzyme was distributed diffusely in the cytoplasm of non-treated trophozoites, but after PCD induction, in addition to its higher expression, it apparently relocated closer to the nucleus.

Regulated proteolysis by calpain is required for the control of fundamental cellular processes including cytoskeletal remodeling, membrane fusion, cell proliferation and differentiation, and activation of proteolytical cascades leading to apoptosis (Saido et al., 1994). This fact is in concordance with our findings that calpain-like protein is expressed basally

in the cytoplasm. On the other hand, when trophozoites are exposed to G418 calpain-like protein also increased, probably due to its participation in the cytoskeleton rearrangement, during apoptosis. Once activated, calpains degrade membrane, cytoplasmic and nuclear substrates, leading to the breakdown of cellular architecture (Momeni, 2011).

In higher eukaryotes, calpains have been suggested to be involved in DNA fragmentation via endonuclease activation (Squier and Cohen, 1997; Villa et al., 1998), and as effector proteases that cleave cellular proteins involved in DNA repair. For instance, upon activation, human µ-calpain is translocated to the nucleus, where it cleaves PARP and p53 (Tagliarino et al., 2001). On the other hand, calpain located in the nucleus of Plasmodium falciparum has been associated with the progression of the cell cycle (Russo et al., 2009). Thus, the observation of the calpain-like protein in the peri-nuclear area of trophozoites 9 h after PCD induction suggests that this protease may participate in the downstream activation of proteins related to DNA fragmentation. This hypothesis is supported by an increase in trophozoite viability and by the blockage of DNA fragmentation in parasites treated with the specific calpain inhibitor Z-Leu-Leu-Leu-al during PCD induction (Sanchez-Monroy et al., 2015).

Here, we showed that knockdown of the calpain-like protein diminished DNA fragmentation and increased cell viability of trophozoites incubated for 9 h with G418, supporting the hypothesis that this cysteine protease participates in the execution phase of PCD. Similarly, the siRNA-mediated knockdown of calpain-1 in neuronal cultures submitted to apoptosis prevented the translocation of the apoptosis-inducing factor (AIF) to the nucleus, thus increasing cell viability (Cao et al., 2007; Jeong et al., 2011), and the knockdown of calpain-2 increased the viability of pancreatic AR42J cells treated with hydrogen peroxide (Hiemer et al., 2017).

Protein-protein interaction experiments currently in progress will allow us to identify other proteins that participate in the PCD of this parasite. We initiated this study by analyzing the putative interactome using the String 10.5 database. The results suggest that at least 20 proteins may have various possible interactions among them with a p-value of 1.23e-07. Interactions of these proteins with the calpain-like protein have not yet been reported in the E. histolytica database. We identified proteins that participate in endocytosis, such as vps36 and snf7, zinc binding proteins, as well as proteins that can bind calcium, including the grainins 1 and 2, which contain EF-hand calcium binding domains. Interestingly, previous studies have demonstrated an increase in the expression of grainins 1 and 2 at 30 min of PCD induction with G418, suggesting that these proteins could act as negative regulators of this event attempting to regulate the cytosolic concentration of free calcium related to the activation of PCD (Monroy et al., 2010).

This assumption is because other proteins with EF-hands, such as calmodulins (CaM) and calmodulin-dependent protein kinase (CaMKIV), that interact with calpain are critical for improved survival. In addition to CaM, other Ca2+-binding proteins, including calpains, play important roles in signal transduction leading to the control of cell proliferation as well as cell death. In vitro, Ca2<sup>+</sup> is known to change calpain conformation, affecting its autocatalytic cleavage and activation (Strobl et al., 2000). Additionally, the Ca2+-activated protease calpain has been shown to be a pro-apoptotic factor by its ability TABLE 1 | Component proteins of the predicted interactome of the calpain-like protein.


to cleave and activate various proteins implicated in the apoptotic process (Momeni, 2011). Interestingly, virtually all proteins that bind to CaM are also calpain substrates (Wang et al., 1989); thus, an increase in intracellular Ca+<sup>2</sup> stimulates the breakdown of proteins necessary for proliferation and viability and triggers programmed cell death (Berridge et al., 2003; Smyth and Putney, 2012).

Other proteins such as the androgen receptor (AR), which is required for growth in both androgen-sensitive and androgeninsensitive prostate cancer are cleaved by calpain in the presence of high Ca2<sup>+</sup> concentrations due to the release of calpain from the calpain-CaM-calpastatin complex. In addition, the calpain-mediated proteolysis of CaMKIV has been demonstrated to trigger apoptosis in cultured neurons (Tremper-Wells and Vallano, 2005).

Considering an update classification of cell death focusing on mechanistic and essential morphological, biochemical, and functional aspects of the process (Galluzzi and Vitale, 2018), we should expeculate that amoeba PCD could be related to the lysosome-dependent cell death (LDCD). This type of cell death is induced by similar stimulus such as perturbations of intracellular homeostasis, cellular stress, ROS (reactive oxygen species) and Ca2<sup>+</sup> imbalance. LDCD is demarcated by the permeabilization of lysosomal membranes and the release of lysosomal contents, including proteolytic enzymes of the cathepsin family, to the cytoplasm (Aits and Jäättelä, 2013). Other triggers include lysosomotropic agents (e.g., sphingosine), and the activation of calpains (Gómez-Sintes et al., 2016). ROS play a prominent causal role in LMP (lysosomal membrane permeabilization), not only

# REFERENCES


as the H2O2-driven luminal production of hydroxyl radicals (Kurz et al., 2008), but also as ROS favor the activation of lysosomal Ca <sup>2</sup><sup>+</sup> channels (Sumoza-Toledo and Penner, 2011). Overall, the results presented here provide strong evidence that the calpain-like protein plays an important role in the execution phase of PCD in E. histolytica. We identified here in silico, some proteins that could interact with calpain-like of E. histolytica. Molecular and biochemical experiments currently in progress will allow us to discover new insights about the lysosomal or other pathways roles in controlling E. histolytica PCD.

# AUTHOR CONTRIBUTIONS

TD-F conceived and carried out experiments, analyzed data and drafted the manuscript. MR and CG participated in the design of the study, analyzed data and drafted the manuscript. VS and OM collaborated in the qRT-PCR and WB assays. DPI conceived and designed the study, analyzed data and drafted the manuscript.

# ACKNOWLEDGMENTS

This work was supported by CONACyT and SIP-IPN grants to DPI. TD-F was a recipient of a fellowship from CONACyT (Grant 237293). We are also deeply grateful to Carlos Vázquez-Calzada (CINVESTAV-IPN) and Maria de Jesus Perea-Flores (Centro de Nanociencias y Micro y Nanotecnologías-IPN) for their technical support in confocal microscopy experiments.


Momeni, H. R. (2011). Role of calpain in apoptosis. Cell J. 13, 65–72.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Domínguez-Fernández, Rodríguez, Sánchez Monroy, Gómez García, Medel and Pérez Ishiwara. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# A Calcium/Cation Exchanger Participates in the Programmed Cell Death and *in vitro* Virulence of *Entamoeba histolytica*

Martha Valle-Solis <sup>1</sup> , Jeni Bolaños <sup>1</sup> , Esther Orozco<sup>1</sup> , Miriam Huerta<sup>1</sup> , Guillermina García-Rivera<sup>1</sup> , Andrés Salas-Casas <sup>2</sup> , Bibiana Chávez-Munguía<sup>1</sup> and Mario A. Rodríguez <sup>1</sup> \*

#### *Edited by:*

Margaret E. Bauer, School of Medicine, Indiana University Bloomington, United States

#### *Reviewed by:*

César López-Camarillo, Universidad Autónoma de la Ciudad de México, Mexico Abhijeet Anil Bakre, University of Georgia, United States

> *\*Correspondence:* Mario A. Rodríguez marodri@cinvestav.mx

#### *Specialty section:*

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

*Received:* 27 April 2018 *Accepted:* 10 September 2018 *Published:* 01 October 2018

#### *Citation:*

Valle-Solis M, Bolaños J, Orozco E, Huerta M, García-Rivera G, Salas-Casas A, Chávez-Munguía B and Rodríguez MA (2018) A Calcium/Cation Exchanger Participates in the Programmed Cell Death and in vitro Virulence of Entamoeba histolytica. Front. Cell. Infect. Microbiol. 8:342. doi: 10.3389/fcimb.2018.00342 <sup>1</sup> Departamento de Infectómica y Patogénesis Molecular, CINVESTAV-IPN, Mexico City, Mexico, <sup>2</sup> Área Académica de Gerontología, Instituto de Ciencias de la Salud, Universidad Autónoma del Estado de Hidalgo, Pachuca, Mexico

Entamoeba histolytica is the etiologic agent of human amoebiasis, disease that causes 40,000 to 100,000 deaths annually worldwide. The cytopathic activity as well as the growth and differentiation of this microorganism is dependent on both, extracellular and free cytoplasmic calcium. However, few is known about the proteins that regulate the calcium flux in this parasite. In many cells, the calcium extrusion from the cytosol is performed by plasma membrane Ca2+-ATPases and calcium/cation exchangers. The aim of this work was to identify a calcium/cation exchanger of E. histolytica and to analyze its possible role in some cellular processes triggered by calcium flux, such as the programmed cell death and in vitro virulence. By searching putative calcium/cation exchangers in the genome database of E. histolyica we identified a protein belonging to the CCX family (EhCCX). We generated a specific antibody against EhCCX, which showed that this protein was expressed in higher levels in E. histolytica than its orthologous in the non-pathogenic amoeba E. dispar. In addition, the expression of EhCCX was increased in trophozoites incubated with hydrogen peroxide. This E. histolytica exchanger was localized in the plasma membrane and in the membrane of some cytoplasmic vesicles. However, after 10 min of erythrophagocytosis, EhCCX was found predominantly in the plasma membrane of the trophozoites. On the other hand, the parasites that overexpress this exchanger contained higher cytosolic calcium levels than control, but the extrusion of calcium after the addition of hydrogen peroxide was more efficient in EhCCX-overexpressing trophozoites; consequently, the programmed cell death was retarded in these parasites. Interestingly, the overexpression of EhCCX increased the in vitro virulence of trophozoites. These results suggest that EhCCX plays important roles in the programmed cell death and in the in vitro virulence of E. histolytica.

Keywords: *Entamoeba histolytica*, calcium/cation exchanger, protein overexpression, programmed cell death, virulence

# INTRODUCTION

The ion calcium (Ca2+) is a highly versatile secondary messenger that operates over a wide temporal range to regulate many different cellular processes (Berridge et al., 2003). Thus, the concentration of Ca2<sup>+</sup> in cytosol and some organelles, such as the endoplasmic reticulum, the Golgi apparatus, and nucleus, is strongly regulated by channels, transporters and pumps, which dynamically adjust the Ca2<sup>+</sup> amount in agreement with a specific physiological demand (Berridge et al., 2003). In many cells Ca2<sup>+</sup> extrusion from the cytosol is performed by plasma membrane Ca2+-ATPases (PMCAs) and by Ca2+/cations exchangers (Berridge et al., 2003). PMCAs have high Ca2<sup>+</sup> affinity, but low turnover rates, while exchangers have a lower Ca2<sup>+</sup> affinity, but higher turnover rates (Blaustein and Lederer, 1999). Consequently, PMCAs are involved in the maintenance of the resting cytoplasmic concentration of Ca2+, whereas exchangers are important in the restoration of cytoplasmic concentration of Ca2<sup>+</sup> after it has been elevated during signaling (Harper and Sage, 2016). Interestingly, exchangers also participate in the entry of Ca <sup>2</sup><sup>+</sup> to cytoplasm, since they can reverse its transport in response to changes in the concentration of the transported ions and membrane potential (Harper and Sage, 2016). The superfamily of calcium exchangers comprises five branches (Lytton, 2007; Emery et al., 2012): (i) H+/ Ca2<sup>+</sup> exchangers (CAX), mainly found in plants; (ii) bacteria exchangers (YRGB); (iii) Na+/Ca2<sup>+</sup> exchangers (NCX or SLC8); (iv) Na+/Ca2<sup>+</sup> + K <sup>+</sup> exchangers (NCKX or SLC24); and (v) Ca2+/cation exchangers (CCX), which contains one mammalian member, the Na+/Ca2<sup>+</sup> or Li<sup>+</sup> exchanger (NCLX) that is also named as NCKX6.

Entamoeba histolytica is the etiological agent of human amoebiasis, a disease that produces 40,000 to 100,000 deaths per year worldwide (Stanley, 2003). The cytolytic activity of this parasite is dependent on both, extracellular and free cytoplasmic Ca2<sup>+</sup> (Ravdin et al., 1982, 1985). In addition, Ca2<sup>+</sup> flux participates in the adherence of trophozoites to fibronectin (Carbajal et al., 1996), as well as in growth and differentiation of E. histolytica and E. invadens (Makioka et al., 2001, 2002; Martínez-Higuera et al., 2015). However, little is known about the proteins that regulate the Ca2<sup>+</sup> flux in this parasite. Our previous studies showed that E. histolytica contains at least five Ca2+- ATPases: three related to PMCAs, one to Sarco/Endoplasmic reticulum Ca2+-ATPases (SERCA), and another to Secretory Pathway Ca2+-ATPases (SPCA) (Martinez-Higuera et al., 2013; Rodríguez et al., 2018). Indeed, we detected the SPCA-related pump in the Golgi apparatus of E. histolytica (Rodríguez et al., 2018) and the SERCA-related pump in the endoplasmic reticulum of E. histolytica and E. invadens (Martinez-Higuera et al., 2013; Martínez-Higuera et al., 2015). In addition, we demonstrated that specific inhibitors of SERCA affected the encystation of E. invadens (Martínez-Higuera et al., 2015), suggesting that calcium flux through SERCA is involved in the development of Entamoeba sp. However, other proteins involved in the Ca2<sup>+</sup> movement, such as channels or exchangers, and their role in the Enatamoeba biology have not been described.

In this work, we identified a calcium/cation exchanger of E. histolytica related to members of the CCX family (EhCCX). This exchanger was higher expressed in E. histolytica that its orthologous in the non-pathogenic amoeba E. dispar. EhCCX expression was increased in trophozoites incubated with hydrogen peroxide. The exchanger was located in the plasma membrane of trophozoites and in the membrane of some cytoplasmic vesicles, but after 10 min of erythrophagocytosis, EhCCX was mainly detected in the plasma membrane. On the other hand, the overexpression of EhCCX augmented the cytosolic calcium levels under basal conditions, but the calcium extrusion after the addition of hydrogen peroxide was more efficient. In addition, the overexpression of EhCCX retarded the programmed cell death and increased the in vitro virulence. These results suggest that the Ca2<sup>+</sup> flux through EhCCX plays an important role in the programmed cell death and the virulence of E. histolytica.

# MATERIALS AND METHODS

# *Entamoeba* Cultures

Trophozoites of E. histolytica clone A, strain HM1:IMSS (Orozco et al., 1983) were axenically cultured in TYI-S-33 medium (Diamond et al., 1978), whereas trophozoites of E. dispar (strain SAW 760) were axenically cultured in YI-S medium (Diamond et al., 1995). Cells were harvested during the logarithmic growth phase as previously described (Diamond et al., 1978).

# Identification and *in silico* Characterization of a Calcium/Cation Exchanger of *E. histolytica*

To identify possible calcium/cation exchangers in Entamoeba spp., a BLAST search was performed on the databases of the Amoeba Genomics Resource (http://amoebaDb.org/amoeba/) using as probes the α1 and α2 repeats of human calcium/sodium exchangers (NCX, NCKX, and NCLX). These motifs are characteristics of these transporters and participate in the ions transport (Philipson and Nicoll, 2000). Then, the retrieved protein of E. histolytica was characterized in silico using the software deposited in the Expasy Bioinformatics Resource Portal (http://expasy.org) and in the NCBI Home Page (http://www. ncbi.nlm.nih.gov). The amino acid sequence of this protein was compared, by ClustalW, with sequences of proteins belonging to the different families of calcium exchangers (CAX, YRGB, NCX, NCKX, and CCX). Then, a phylogenetic analysis was performed using the Unweighted Pair Group Method with Arithmetic Mean

**Abbreviations:** CCX, calcium/cation exchanger; EhCCX, calcium/cation exchanger of E. histolytica; EdCCX, calcium/cation exchanger of E. dispar; EhNPC1, Niemann-Pick type C protein 1 of E. histolytica; EhRabB, RabB protein of E. histolytica; EhSERCA, sarco/endoplasmic reticulum Ca2+-ATPase of E. histolytica; Gal/GalNac lectin, lectin of E. histolytica inhibited by galactose and N-acetyl-D-galactosamine; NCLX, sodium/calcium/lithium exchanger; NCX, sodium/calcium exchanger; MjNCX, sodium/calcium exchanger of Methanococcus jannaschii; PMCA, plama membrane Ca2+-ATPase; PCD, programmed cell death.

(UPGMA) employing the MEGA 5.05 software package (Tamura et al., 2011). Bootstrapping was performed for 1000 replicates.

The 3D molecular model was built with the I-TASSER server (http://zhanglab.ccmb.med.umich.edu/I-TASSER) (Zhang, 2008) and with the Raptor X Structure Prediction (http:// raptorx.uchicago.edu) using as a template the crystalized structure of the Methanococcus jannaschii NCX (Protein Data Bank: 3V5S) (Liao et al., 2012).

# PCR and RT-PCR

The genomic DNA of E. histolytica was obtained with the Wizard Genomic DNA purification kit (Promega), following the manufacturer's recommendations. Total RNA was isolated using the Trizol reagent (Invitrogen), following the manufacturer's recommendations, and cDNA was synthesized using an oligo dT primer and the Superscrip II reverse transcriptase (Invitrogen). Amplifications of the Ehccx full-length gene was performed in a select Cycler (Select BioProducts) using 200 ng of DNA or cDNA and specific primers situated at the 5′ - and 3′ -ends of the gene and containing the BamHI and KpnI restriction sites, respectively (forward, 5′ - CCCCGGTACCATGAAACAGATGAATAAAAT TTATATTATATTA-3′ ; reverse, 5′ -CCCCGGATCCTTAACCAA ACAGTTTAAAAACGTTAA-3′ ). The assays were performed in a 50 µl volume reaction containing 1µM of each primer, dNTPs 1.5 mM, MgCl<sup>2</sup> 2 mM, and 1 U of the high-fidelity enzyme KAPA HiFi DNA polymerase (KAPABIOSYSTEMS). Amplification cycles comprised: (i) 1 min of denaturing step at 94◦C; (ii) 35 cycles of 1 min of denaturing step at 94◦C, 1 min of annealing step at 55◦C; and 3 min of elongation step at 72◦C; and (iii) 10 min of elongation step at 72◦C. The amplified products were analyzed by electrophoresis in 1% agarose gels.

# Production of the α-EhCCX Antibody and Western Blot

To obtain antigenic peptides of the predicted calcium/cation exchanger of E. histolytica (EhCCX), its amino acid sequence was analyzed by the ABCpred program (http://www.imtech.res. in/raghava/abcpred/). Peptides with higher scores were used as probes in BLAST searches on the E. histolytica genome database; then, a specific peptide at position 216-229 (ISEQLDSENKTKLI) was synthesized (GL Biochem) linked to the KLH (Keyhole limpet Hemocyanin) tag to increase its immunogenicity. Next, Wistar rats were intradermally immunized three times, in an interval of 2 weeks, with 100 µg of the synthetic polypeptide resuspended in Titermax Gold adjuvant (1:1) (Sigma). All animals used in this study were handled in accordance with the guidelines of the Institutional Animal Care and Use Committee. Our institution fulfills all the technical specifications for the production, care and use of laboratory animals and it is certified by national law (NOM-062-ZOO-1999).

For Western blot assays, total extracts of E. histolytica or E. dispar trophozoites obtained in the presence of protease inhibitors (Complete Mini, Roche-Mannheim) were separated by 10% SDS-PAGE, transferred to nitrocellulose membranes and probed with the α-EhCCX antibody (1: 3,000). Then membranes were incubated with a peroxidase-coupled secondary antibody (1: 10,000) (ZYMED) and finally, the reaction was developed by chemiluminescence (ECL Plus GE-Healthcare). To compare the expression level of EhCCX and EdCCX or to analyze the expression level of EhCCX in trophozoites under different conditions (in the presence of H20<sup>2</sup> 1 mM for 10 min, or incubated at 42◦C for 30 and 60 min) we performed Western blot assays using the α-EhCCX antibody. Then, membranes were exposed to an α-actin antibody (1: 20,000) (kindly provided by Dr. Jose Manuel Hernandez-Hernandez at CINVESTAV-IPN, Mexico), used as an internal control of loading. The band detected by the α-EhCCX antibody was analyzed by scanning densitometry and the data were normalized to the actin content according to the reactivity of the α-actin antibody. For semiquantitative comparisons, the protein level in E. histolytica trophozoites under basal conditions was arbitrary taken as 1.

# Immunofluorescence and Confocal Microscopy

Trophozoites grown on coverslides were fixed and permeabilized with methanol for 10 min and non-specific binding sites were blocked with 10% FBS in phosphate buffered saline (PBS). Then, cells were incubated for 1 h at 37◦C with the rat antibody against EhCCX (1:50 dilution). After several washes with PBS, samples were incubated with an Alexa 555 conjugated secondary antibody (Zymed) (1:400). Finally, the nuclei were stained with 4',6-diamidino-2-Phenylindole (DAPI) and samples were observed through a confocal microscope (Carl Zeiss LSM 700). Observations were performed in ∼20 optical sections from the top to the bottom of each sample. For the immunolocalization of EhCCX during phagocytosis, before performing the immunofluorescence, trophozoites were incubated with fresh human red blood cells (RBCs) from healthy donors (1:25 ratio) for 5, 10, and 15 min at 37◦C. As a control, cells were incubated in the presence of EGTA 0.3 mM.

For co-localization assays, after incubation with α-EhCCX, the samples were incubated with rabbit antibodies against EhSERCA (Martinez-Higuera et al., 2013) (dilution 1:50), EhRabB (Rodriguez et al., 2000) (dilution 1:25), or EhNPC-1 (Bolaños et al., 2016) (dilution 1:100); or with mouse antibodies against the Gal/GalNac lectin (Petri et al., 1989) (dilution 1:50). Subsequently, the samples were incubated with anti-rat IgGs labeled with Alexa 555, and with anti-rabbit IgGs labeled with Alexa 488, or anti-mouse IgGs labeled with FITC (Zymed; 1:100) as appropriate. For detection of the Golgi apparatus, trophozoites were incubated with 5µM of NBD C6-Ceramide (Thermo Fisher Scientific) for 90 min at 37◦C. To evaluate the co-localization between molecules, Pearson coefficients were obtained from at least 15 confocal independent images (laser sections: 0.5µm) using the ImageJ 1.45v software and the JACoP plugin.

# Immunoelectron Microscopy

Immunoelectron microscopy assays were performed as described (Segovia-Gamboa et al., 2011). E. histolytica trophozoites were fixed in 4% paraformaldehyde and 0.1% glutaraldehyde in serum-free DMEM for 1 h at room temperature. Samples were embedded in LR white resin (London Resin Co) and polymerized under UV at 4◦C for 48 h. Then, thin sections (60 nm) were mounted on formvar-covered nickel grids followed by overnight incubation with the α-EhNCX (1:20) antibody and later, by overnight incubation with the secondary antibody (1:60) conjugated to 15 nm gold particles (Ted Pella Inc.). Finally, sections were observed with a morgana 268 D Philips transmission electron microscope (FEI Company).

# Overexpression of the *Ehccx* Gene

The full-length Ehccx gene was obtained by PCR using genomic DNA as template and using the conditions described above. Then, the amplicon was sequenced and cloned into the BamHI and Kpnl restriction sites of the pExEhNeo (pNeo) plasmid (Hamann et al., 1995), obtaining the pNeo/Ehccx construct.

For transfection, 3 × 10<sup>5</sup> trophozoites were cultivated overnight at 37◦C in a 5% CO<sup>2</sup> environment; then, cells were washed with M199 medium (Sigma), resuspended in 1.8 ml of M199 medium supplemented with 15% of fetal bovine serum (FBS), and placed in 35-mm Petri dishes. Subsequently, a mix of 10 µg of plasmid (pNeo or pNeo/Ehccx) and 10 µg of Superfect (Qiagen) in 100 µl of M199 medium was added to the cultures and incubated for 4 h at 37◦C in a 5% CO<sup>2</sup> environment. Afterwards, cells were transferred to a tube containing 10 ml of TYI-S-33 pre-warmed medium and incubated 48 h at 37◦C. Finally, transfected parasites were selected in the presence of 10µg/ml of G-418 (Sigma-Aldrich). Overexpression was confirmed by qRT-PCR and Western blot assays.

For qRT-PCR, 100 ng of cDNA, 0.15µM of each primer (forward, 5′ -CACTGAAACACAAATCCCTTC-3′ ; reverse, 5′ - CCAACTGAATTTCCCCAACA-3′ ) and the KAPA SYBR FAST PCR Master Mix (Kapa Biosystems) were used in a StepOne TM Real-Time PCR System (Applied Biosystem). As normalizer we used the gene encoding for the protein S2 of the 40S ribosomal subunit (primers: forward, 5′ -ATTCGGAAATAG AAGAGGAGG-3′ ; reverse, 5′ - CTATTCTTCCAAGCTTGGT-3 ′ ). Data from three independent cDNA preparations were analyzed using the 2−11CT method.

Western blot assays and semi-quantitative comparisons were performed as described above to analyze the expression level of the EhCCX protein in transfected trophozoites. The protein level in trophozoites transfected with the empty vector (pNeo) was arbitrary taken as 1.

# Intracellular Ca2<sup>+</sup> Levels

The trophozoites (1 × 10<sup>6</sup> ) were suspended in 1 ml of PBS and incubated for 30 min at 37◦C in the dark with 4µM of the Ca2+-sensitive dye Fluo-4 AM (Invitrogen). To remove the extracellular dye, the parasites were washed three times with PBS and resuspended in 1 ml of Ca2+/Mg2<sup>+</sup> PBS (CaCl<sup>2</sup> 0.1 mM, MgCl<sup>2</sup> 1 mM). The fluorescence signal was detected by flow cytometry using the 520/40 filters in a BD FACSCalibur analyzer (BD Biosciences). To analyze the cytosolic calcium levels during the programmed cell death, H2O<sup>2</sup> 1 mM was added to the samples and the fluorescence intensity was measured every 40 s.

# Cell Death Induced by Hydrogen Peroxide

The trophozoites (5 × 10<sup>5</sup> ) were treated with H2O<sup>2</sup> 1 mM in serum-free TYI medium and incubated at 37◦C. Then, the cell viability was measured at different times by Trypan Blue exclusion.

# Erythrophagocytosis

Trophozoites suspended in serum-free TYI medium were incubated with fresh RBCs (1:100 ratio) at 37◦C with slight agitation for 5, 10, and 15 min. Non-ingested erythrocytes were lysed by incubation with distilled water for 10 min at room temperature. Then, the samples were washed three times with PBS and parasites with internalized RBCs, were lysed with 1 ml of concentrated formic acid (J.T. Baker). Finally, phagocytosis was determined by measuring the hemoglobin released by the ingested erythrocytes (absorbance at 405 nm using a Beckman Coulter DU800 spectrophotometer).

# Cytopathic Assays

Cytopathic assays, defined as the ability of live trophozoites to destroy cultured cells, were carried out as previously described (Rodríguez and Orozco, 1986). Briefly, MDCK cell monolayers (1 × 10<sup>5</sup> cells) were incubated for 2 h at 37◦C with 1 × 10<sup>5</sup> trophozoites. Then, the trophozoites were eliminated by incubation at 4◦C during 10 min, and the remaining mammalian cells were fixed with 2.5% (v/v) glutaraldehyde and stained with 1% (w/v) methylene blue. After exhaustive washes, the dye captured by cells was extracted with 0.1 M HCl and measured in a spectrophotometer (Beckman Coulter DU800) at 660 nm.

# Migration Assays

Serum-starved (3 h) trophozoites (7.5 × 10<sup>4</sup> ) were placed in the upper chamber of transwell inserts (5µm pore size, 24 well, Costar) and 500 µl of adult bovine serum was added to the lower chamber. Then, trophozoites were incubated for 3 h at 37◦C and migration was determined by counting the number of trophozoites at the lower chamber of the well.

# Statistical Analysis

Values of all assays were expressed as the mean ± standard error of three independent experiments by duplicate. Statistical analyzes were carried out using the GraphPad Prism V5.01 software by two-way ANOVA and Student's t-test.

# RESULTS

# *E. histolytica* Has a Calcium Exchanger Belonging to the CCX Family

We search for putative calcium/cation exchangers in the Amoeba Genomics Resource (http://www.amoebaDb.org) using as probes the α1 and α2 domains of the different human calcium/cation exchangers (NCX, NCKX, or NCLX). By this analysis, the α1 domains of NCLX and NCX2, as well as the α2 domains of NCLX, NCKX1, and NCKX2 retrieved just one putative cation/calcium exchanger in E. histolytica (EHI\_001770), E. nuttali (ENVI\_134420), E. dispar (EDI\_138590), and E. invadens (EIN\_083040), whereas they retrieved two exchangers of E. moshkovskii (EMO\_063580 and EMO\_049830). The identity of the E. histolytica exchanger with the orthologous proteins of the other Entamoeba species ranged from 69 to 99%, whereas similarity varied from 83 to 99% (**Table 1**). On the other hand, the E. histolytica exchanger displayed 16–30% identity and 42– 51% similarity with calcium/sodium exchangers from plants, mammals, amphibians, insects, and bacteria (**Table 1**).

The putative calcium/cation exchanger of E. histolytica has 513 amino acid residues. It contains a signal peptide in positions 1 to 19 and its secondary structure model showed a similar organization to the human NCLX protein (NP\_079235.2) (**Figure 1A**). Both proteins contain: (i) two sodium/calcium exchanger (NCX) domains (positions 62-206 and 352-503 in EHI\_001770) that include the α1 and α2 repeats (positions 101-141 and 391-447 in EHI\_001770); and (ii)10 putative transmembrane segments (positions 65-75, 108-126, 127-147, 165-181, 187-200, 363-382, 387-403, 422-444, 461-479, and 484-501 in EHI\_001770) (**Figure 1A**). Then, we constructed a predicted 3D model of the putative E. histolytica exchanger using the crystallized structure of the sodium/calcium exchanger of Methanococcus jannaschii (MjNCX) as template. This 3D model confirmed the presence of 10 transmembrane segments in the amoebic protein, but it showed a modest structural identity with MjNCX (9.38%) and with the human NCLX (11.13%) (**Figure 1B**). However, the structural identity between human and bacteria exchangers was also moderate (17.19%) (**Figure 1B**). Nevertheless, the predicted structures of the α1 and α2 repeats of the E. histolytica protein showed a higher identity with those of NCLX (43.9 and 43.14%, respectively) (**Figure 1B**).

To determine the family of calcium exchangers to which the putative E. histolytica protein belongs, we compared its amino acid sequence with those of different members of the YRGB (bacterial exchangers), NCX (Ca2+/Na<sup>+</sup> exchangers), NCKX (Ca2+/Na<sup>+</sup> + K <sup>+</sup> exchangers), CAX (Ca2+/H<sup>+</sup> exchangers), and CCX (Ca2+/cation exchangers) families. Then, we constructed a phylogenetic tree, where the proteins of the different Entamoeba species were clustered into the clade comprising the CCX family, whereas the exchangers of other protozoa parasites were grouped in the clade of the CAX family (**Figure 2A**). Moreover, the E. histolytica exchanger contains the sequences GNG(A/T)PD and GNSIGD, corresponding to distinctive motives of the CCX family, which include the Ca2+/Na<sup>+</sup> or Li<sup>+</sup> exchangers (NCLX) (Sharma and O'Halloran, 2014) (**Figure 2B**). Concordantly, the E. histolytica protein shares all the amino acid residues that were identified as responsible of transporting Ca2+, Na+, or Li<sup>+</sup> in the human NCLX (Roy et al., 2017) (**Figure 2B**); therefore, we named the E. histolytica exchanger as EhCCX.

# Expression and Localization of EhCCX

To investigate whether the transcript of the Ehccx gene is found in the trophozoites we performed RT-PCR assays. Results showed the amplification of a cDNA fragment with the expected molecular size (**Figure 3A**), indicating that the Ehccx gene is transcriptionally active under basal conditions. Next, with the purpose to immunodetect the EhCCX protein, we produced an antibody against a specific peptide located at position 216-229. Based on the 3D model, this peptide is into an exposed segment (**Figure 2A**). In Western blot assays, the antibody detected a single band of 56 kDa (**Figure 3B**), the expected molecular weight for EhCCX. By immunofluorescence assays, this protein was detected in the plasma membrane and in the membrane of numerous cytoplasmic vesicles (**Figure 3C**). Immunoelectron microscopy confirmed the presence of EhCCX in the plasma membrane and in the membrane of some cytoplasmic vesicles (**Figure 4**).

To identify the possible cytoplasmic compartments in which EhCCX might be situated, we analyzed its co-localization with markers for endoplasmic reticulum (EhSERCA), Golgi apparatus (NBD C6-Ceramide), late endosomes (NPC-1), and endocytic vesicles (EhRabB). We also confirmed the presence of EhCCX in the plasma membrane by its co-localization with the Gal/GalNac lectin. In these assays, EhCCX showed some spots of co-localization with the Gal/GalNac lectin and EhRabB into or close to the plasma membrane (**Figure 5A**). This exchanger also showed co-localization with NBD C6-Ceramide in some vesicles close to the plasma membrane (**Figure 5A**),

TABLE 1 | Comparison of the calcium/sodium exchanger of E. histolytica with orthologous proteins of different organisms.


suggesting its presence in the trans-Golgi network. On the other hand, EhCCX did not display significant co-localization with EhSERCA, NPC-1 or cytoplasmic EhRabB (**Figure 5A**). The Pearson coefficient correlation confirmed a significant colocalization of EhCCX only with the Gal/GalNac lectin, EhRabB and NBD C6-Ceramide (**Figure 5B**). These results suggested

that EhCCX in the plasma membrane could be involved in the extrusion of Ca2<sup>+</sup> from cytosol, whereas its presence in in trans-Golgi could be due to its cytoplasmic transport. Alternatively, the posttranslational modifications needed for the maturation process of EhCCX could explain its presence in the Golgi apparatus. Nevertheless, we do not know the identity of most of the cytoplasmic vesicles that contain EhCCX nor its function in these structures.

The non-pathogenic amoeba E. dispar has a lower erythrophagocytic capacity that the pathogenic E. histolytica (Talamás-Lara et al., 2014). To analyze whether EhCCX could participate in virulence, we compared the location and

PCR using RNA (omitting the cDNA synthesis) as template. (B) A specific peptide of EhCCX was synthesized and inoculated in rats to obtain antibodies against this exchanger. Then, to analyze their specificity, these antibodies were used in Western blot assays on total extracts of trophozoites. Lane 1, total proteins of E. histolytica trophozoites stained with Coomassie blue; lane 2, Western blot assay using the pre-immune serum; lane 3, Western blot assay using α-EhCCX; (C) E. histolytica trophozoites were fixed, permeabilized and incubated with rat antibodies against EhCCX. Subsequently, samples were incubated with Alexa 555-conjugated secondary antibodies. Finally, nuclei were stained with DAPI and samples were analyzed by confocal microscopy. PI, pre-immune serum.

expression of the exchanger in E. histolytica (EhCCX) and E. dispar (EdCCX). Western blot assays showed that E. histolytica expresses a higher amount of the exchanger compared with E. dispar (**Figure 6A**). Immunofluorescence assays using the α-EhCCX antibody showed that the localization of the exchanger in the trophozoites of E. histolytica and E. dispar is similar, but the fluorescent signal was apparently lower in E. dispar (**Figure 6B**). The minor expression of the exchanger in E. dispar supports the hypothesis that EhCCX could participates in the virulence mechanism of E. histolytica.

# EhCCX Is Stress-Inducible

Calcium/cation exchangers are important in the restoration of the cytoplasmic concentration of Ca2<sup>+</sup> after it has been elevated by different stimuli, including oxidative stress. Thus, we analyzed the expression level and localization of EhCCX in trophozoites exposed to H2O<sup>2</sup> 1 mM during 10 min. Western blot assays showed that this exchanger slightly increased its expression (∼0.4 times) under this condition (**Figure 7A**), suggesting that an augment of this protein could be needed to enhance the calcium extrusion trying to avoid the cellular damage. By immunofluorescence we observed a small accumulation of EhCCX in the plasma membrane and in some big vacuoles (**Figure 7B**).

We also analyzed the expression of EhCCX in trophozoites submitted to heat shock. After 30 min of incubation at 42◦C, the expression of EhCCX showed a slight increase (about 0.2 times) that descend to approximately the basal levels after 60 min of heat shock (**Figure 7C**). We do not observe any change of the location of this exchanger under 30 and 60 min of incubation at 42◦C (**Figure 7D**). Results suggest that EhCCX could has a minor participation or does not participate in the response to heat shock.

# Phagocytosis Alters EhCCX Localization

It has been described that the accumulation of cytosolic calcium plays a major role in several cellular processes of E. histolytica trophozoites, including phagocytosis (Jain et al., 2008). Therefore, to investigate the possible role of EhCCX in this cellular event we performed Western blot assays on total extracts of trophozoites obtained at different times of erythrophagocytosis. No differences in the protein level were detected in these experiments (**Figure 8A**). However, we observed changes in its localization during the erythrophagocytosis. At 5 min, the localization of EhCCX was similar to the basal conditions; but at 10 min the exchanger was mainly detected in the plasma membrane, and after 15 min the presence of this protein diminished in the plasma membrane and it was concentrated in some cytoplasmic spots (**Figure 8B**). The accumulation of EhCCX in the plasma membrane at 10 min of phagocytosis suggested that this exchanger could participate in the signaling triggered by the interaction with the target cells and this signaling could be interrupted at 15 min by the internalization of EhCCX. In the presence of the calcium chelator EGTA, few erythrocytes were phagocyted by trophozoites, even after 30 min (**Figure 8B**), confirming that calcium flux is involved in this process. In this condition the movement of EhCCX to the plasma membrane was not observed, but it was concentrated in some small points in the cytoplasm (**Figure 8B**). These results suggest that the signaling pathway triggered by the presence of target cells promotes the recruitment of EhCCX in the amoeba surface.

# EhCCX Overexpression Modifies the Cytoplasmic Calcium Level and PCD

To confirm the role of EhCCX in calcium movement and virulence we overexpressed this putative exchanger in

trophozoites. Thus, we amplified the Ehccx gene by PCR using genomic DNA as a template and a high-fidelity DNA polymerase. The sequence of the amplicon showed two nucleotide changes with respect to the sequence EHI\_001770 deposited in the genome database of E. histolytica: T to C in the position 220 and G to A in the position 328 (data not shown). These nucleotide substitutions modify the amino acid residues 74 (Ser to Pro) and 110 (Ala to Thr), which are into the α1 repeat (**Supplementary Figure 1A**); however, they did not produce significant alterations in the predicted 3D structure of this motif (**Supplementary Figure 1B**). In addition, although the residue in position 110 is into one of the distinctive sequences of the CCX members (**Figure 2B**), the amino acid change fits into the GNG(A/T)PD consensus sequence (Sharma and O'Halloran, 2014), and previous functional studies indicated that the residue situated in this place in the human NCLX does not participate in the ions transport (Roy et al., 2017). Thus, the amplicon was cloned into the pExEhNeo (pNeo) vector (Hamann et al., 1995) and trophozoites were transfected with the resultant construct (pNeo/Ehccx) to overexpress this exchanger. qRT-PCR assays showed that parasites transfected with the pNeo/Ehccx plasmid (pNeo/EhCCX trophozoites) overexpressed about 3.5 times the Ehccx transcript (**Figure 9A**). On the other hand, Western blot assays revealed that the EhCCX protein was overexpressed 2.8 times in pNeo/EhCCX trophozoites compared with the control (**Figure 9B**).

Then, to confirm that EhCCX participates in Ca2<sup>+</sup> flux, we analyzed the cytosolic calcium levels in pNeo and pNeo/EhCCX cells using the Ca2+-sensitive dye Fluo-4 AM and flow cytometry. In these assays the fluorescence intensity in pNeo/EhCCX cells was about twice higher with respect to that showed by the pNeo trophozoites (**Figure 10A**). Then, to analyze the calcium flux in these trophozoites, we added H2O<sup>2</sup> 1 mM and analyzed the fluorescence intensity every 40 s. In these assays, the signal gradually increased in pNeo cells (**Figure 10B**). In contrast, in

pNeo/EhCCX trophozoites the florescence intensity augmented after 40 s of incubation with hydrogen peroxide, but subsequently the signal progressively diminished until at 400 s the fluorescence intensity was similar to that of pNeo cells under basal conditions (**Figure 10B**). These results confirm the role of EhCCX in Ca2<sup>+</sup> transport, because the higher expression of the exchanger produced a greater influx of calcium under basal conditions and improve the extrusion of the cytosolic calcium during the programmed cell death (PCD) induced by H2O2.

In concordance with the augmented extrusion of calcium during the PCD-induction, after 30 and 60 min of incubation with H2O2, the cell death of pNeo/EhCCX was significantly lower than pNeo (**Figure 10C**); however, we did not observe differences in the cell death after 90 min of treatment (**Figure 10C**). These results suggest that the higher extrusion of cytosolic calcium driven by the overexpressed exchanger retarded the PCD of trophozoites.

FIGURE 6 | Expression and localization of CCX in E. dispar. (A) The comparison of the expression level of the CCX protein in E. histolytica (EhCCX) and E. dispar (EdCCX) was performed by Western blot using the α-EhCCX antibody. As an internal control, the membranes were probed with an α-actin antibody. The band detected by the α-EhCCX antibody was analyzed by densitometry and the data were normalized to the actin content. The relative expression in E. histolytica was taken as 1. Data are expressed as the mean ± standard error of three independent experiments. (B) E. histolytica and E. dispar trophozoites were fixed, permeabilized and incubated with α-EhCCX antibodies. Next, they were incubated with Alexa 555-conjugated secondary antibodies, nuclei were stained with DAPI, and samples were analyzed by confocal microscopy.

The band detected by the α-EhCCX antibody was analyzed by densitometry and the data were normalized to the actin content. The relative expression in trophozoites at 0 min was taken as 1. Data are expressed as the mean ± standard error of three independent experiments. (B,D) Cells were fixed, permeabilized and incubated with α-EhCCX antibodies. Next, they were incubated with Alexa 555-conjugated secondary antibodies, nuclei were stained with DAPI, and samples were analyzed by confocal microscopy.

# EhCCX Overexpression Is Increasing the *in vitro* Virulence of *E. histolytica*

We analyzed the erythrophagocytosis, cytopathic activity and migration of pNeo/EhCCX to corroborate that EhCCX is also involved in the in vitro virulence of E. histolytica. Results showed that pNeo/EhCCX cells displayed a significant increase in erythrophagocytosis, compared with pNeo (**Figure 11A**). EhCCX-overexpressing trophozoites also enhanced more than twice their ability to destroy mammalian cells (**Figure 11B**), and remarkably, migration of trophozoites augmented almost four times compared to pNeo (**Figure 11C**). All these results support the hypothesis that Ca2<sup>+</sup> flux mediated by EhCCX participates in the in vitro virulence of E. histolytica.

# DISCUSSION

The Ca2<sup>+</sup> extrusion from the cytosol is carried out mainly by Ca2+-ATPases of the PMCA family and calcium/cations exchangers, therefore, these proteins play a critical role in Ca2<sup>+</sup> homeostasis (Berridge et al., 2003). Calcium/cation exchangers

probed with an α-actin antibody. The band detected by the α-EhCCX antibody was analyzed by densitometry and the data were normalized to the actin content. The relative expression in trophozoites at 0 min was taken as 1. Data are expressed as the mean ± standard error of three independent experiments. (B) Cells were fixed, permeabilized and incubated with α-EhCCX antibodies. Next, they were incubated with Alexa 555-conjugated secondary antibodies, nuclei were stained with DAPI, and samples were analyzed by confocal microscopy. Images at the right correspond to assays performed in the presence of EGTA 0.3 mM.

transport Ca2<sup>+</sup> against its electrochemical gradient coupled to the exchange of different cations, such as H<sup>+</sup> (YRGB and CAX), Na<sup>+</sup> (NCX), Na<sup>+</sup> + K <sup>+</sup> (NCKX), and Na<sup>+</sup> or Li<sup>+</sup> (CCX) (Lytton, 2007; Emery et al., 2012). All these exchangers share structural similarities, including two α-repeats involved in the ion binding/transport events (Giladi et al., 2016). The differences in the ion-coordinating residues among the members of the Ca2+/cation superfamily are responsible of the divergences in ion selectivity (Refaeli et al., 2016).

Ca2<sup>+</sup> homeostasis is involved in the host invasion by different protozoa parasites (Ravdin et al., 1985; Lu et al., 1997; Vieira and Moreno, 2000; Lovett and Sibley, 2003; Moreno et al., 2007). It has been described the presence of PMCAs in most of these microorganisms (Moreno and Docampo, 2003), however, until now there are not reports about the existence of Ca2+/Na<sup>+</sup> exchangers for any parasite protozoa. A genomic analysis in Apicomplexans revealed that Toxoplasma gondii, Cryptosporidium spp., and Plasmodium spp. contain orthologs of Ca2+/H<sup>+</sup> exchangers (CAX) found in plants and yeast, but not in animals. Conversely, Ca2+/Na<sup>+</sup> exchangers, which are common in animals, are not found in these parasites (Nagamune and Sibley, 2006). In Plasmodium berghei, the CAX protein participates in ookinete development and differentiation (Guttery et al., 2013). On the other hand, biochemical evidence suggests that trypanosomatids also have a Ca2+/H<sup>+</sup> antiporter (Verseci et al., 1994), which was proposed to be involved in Ca2<sup>+</sup> release by the increase of sodium mediated by a Na+/H<sup>+</sup> antiporter (Versesi and Docampo, 1996). Concordantly, by a in silico analysis we found a gene coding for a putative Ca2+/H<sup>+</sup> exchanger in Leishmania mexiana. We also identify a gene coding for a putative Ca2+/H<sup>+</sup> exchanger in Trichomonas vaginalis, but we did not discover Ca2+/Na<sup>+</sup> exchangers in any parasite protozoa different to Entamoeba spp.

The E. histolytica exchanger showed a higher similarity with the CCX family and contains the same amino acid residues involved in the transport of calcium, sodium, and lithium of human NCLX. It is known that the intracellular concentration of Li<sup>+</sup> is very low (0.6–0.8 mM); thus, the Li<sup>+</sup> transport by NCLX members just could be related to some functional properties. For instance, Ca2<sup>+</sup> transport by NCX is driven by a steep Na<sup>+</sup> gradient and a moderate (∼-70 mV) membrane potential,

FIGURE 9 | Overexpression of EhCCX. The Ehccx full-length gene was cloned into the pNeo vector and trophozoites were transfected with this construct (pNeo/EhCCX). Trophozoites transfected with the empty vector (pNeo) were used as controls. (A) RNA isolated from trophozoites was used for qRT-PCR assays. As gene normalizer we used the gene coding for the S2 protein of the 40S ribosome subunit. Relative expression was determined using the 2−11CT method. Inset showed the melt curve that demonstrates the specificity of the Ehccx amplification. (B) Western bolt assays using the α-EhCCX antibody. As an internal control, the membranes were probed with an α-actin antibody. The band detected by the α-EhCCX antibody was analyzed by densitometry and the data were normalized to the actin content. The relative expression in pNeo trophozoites was taken as 1. Data are expressed as the mean ± standard error of three independent experiments.

treated with hydrogen peroxide 1 mM and the fluorescence intensity was measured every 40 s. (C) The cell viability of trophozoites incubated with hydrogen peroxide 1 mM was determined by Trypan blue exclusion at 5, 10, 30, 60, and 90 min. Data are expressed as the mean ± standard error of three independent experiments performed by duplicate.

whereas Ca2<sup>+</sup> efflux by NCLX is primarily driven by a much steeper (∼−200 mV) membrane potential (Roy et al., 2017).

Based on the disulfide cross-linking data, a model for Ca2+/Na<sup>+</sup> exchangers was proposed (Nicoll et al., 1999). In this model, the transmembrane segments 5 and 6 (TMS5 and TMS6) are separated by a large (∼500 amino acids) floop that faces the cytosolic side. This loop is responsible for the regulation of the ions transport activity by the binding of several cytoplasmic messengers, protons, NO, PIP2, phosphoarginine, phosphocreatinine, ATP, and endogenous inhibitors (Khananshvili, 2013). Interestingly, NCLX members have a very short f-loop with no regulatory domains on it (Khananshvili, 2013). Similarly, EhCCX has a small f-loop (163 aa), suggesting that this protein, as NCLX members, has no controlling domains for "secondary" regulation.

The human NCLX is found in the inner membrane of the mitochondria (Palty et al., 2010); however, it was also detected in the plasma membrane of ventricular myocytes and pancreatic β-cells (Cai and Lytton, 2004; Han et al., 2015). Indeed, it has been reported that in pancreatic β-cells the NCLX of plasma

membrane is involved in the calcium flux that contributes to the vesicle recruitment for sustained exocytosis in response to repetitive depolarization (Han et al., 2015). Likewise, the EhCCX located in the surface of trophozoites may participates in the calcium flux through the plasma membrane. This hypothesis is supported by the increase of its expression in trophozoites exposed to H20<sup>2</sup> and by the changes in the cytosolic calcium levels in EhCCX-overexpressing trophozoites during the incubation with H202.

On the other hand, EhCCX seems to be involved in the amoeba virulence. This hypothesis is supported by: (i) the expression level of this protein is higher in the pathogenic E. histolytica than in the non-pathogenic E. dispar; and (ii) the overexpression of EhCCX augmented the in vitro virulence properties of trophozoites, such as phagocytosis, destruction of mammalian cell monolayers and migration. Similarly, the overexpression of NCX proteins enhances the heart rate mediated by beta-adrenergic (Kaese et al., 2017), and the relaxation of gastric fundus (Fujimoto et al., 2016). We are hypothesizing that, as calcium/cation exchangers can reverse the ions transport under certain conditions (Harper and Sage, 2016), the different stimuli that are involved in the E. histolytica virulence could increase the Ca2<sup>+</sup> influx via EhCCX, extending the period of high Ca2<sup>+</sup> levels in cytoplasm, maintaining the response to the signal. Such mechanism of Ca2<sup>+</sup> influx via NCX has also been proposed to explain the enhanced contraction displayed by the urinary bladder smooth muscles of transgenic mice overexpressing NCX1.3 (Yamamura et al., 2013).

In conclusion, in this work we identified a calcium/cation exchanger of E. histolytica (EhCCX). The expression of this exchanger, belonging to the CCX family, is higher in pathogenic than in non-pathogenic amoeba. Moreover, the cellular localization of EhCCX changed during phagocytosis and its overexpression increased the in vitro virulence and retarded the PCD induced by H2O2. These results suggest that Ca2<sup>+</sup> flux through EhCCX participates in the PCD and the in vitro virulence of E. histolytica.

# AUTHOR CONTRIBUTIONS

MV-S conceived and carried out experiments, analyzed data, and drafted the manuscript. JB performed experiments and analyzed data. EO and MH participated in the design of the study and analyzed data. GG-R carried out the transfection of trophozoites and participates in their characterization. AS-C realized the measurements of cytosolic calcium levels. BC-M performed the immunoelectron microscopy. MR conceived and designed the study, analyzed data and drafted the manuscript.

# ACKNOWLEDGMENTS

MV-S was a recipient of a fellowship from Consejo Nacional de Ciencia y Tecnología (Grant 418500). We are grateful for the excellent technical assistance of Carlos Vázquez-Calzada and Victor Rosales-García. We also thank Dr. José Manuel Hernández-Hernández (CINVESTAV-IPN) for providing the αactin antibody.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00342/full#supplementary-material

# REFERENCES


of Entamoeba histolytica. Parasitol. Res. doi: 10.1007/s00436-018- 6030-4. [Epub ahead of print].


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Valle-Solis, Bolaños, Orozco, Huerta, García-Rivera, Salas-Casas, Chávez-Munguía and Rodríguez. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Phosphatidylinositol Kinases and Phosphatases in *Entamoeba histolytica*

Kumiko Nakada-Tsukui <sup>1</sup> \* † , Natsuki Watanabe1,2†, Tomohiko Maehama<sup>3</sup> and Tomoyoshi Nozaki <sup>4</sup> \*

#### *Edited by:*

Serge Ankri, Technion Israel Institute of Technology, Israel

#### *Reviewed by:*

Elisa Azuara-Liceaga, Universidad Autónoma de la Ciudad de México, Mexico Patricia Talamás-Rohana, Centro de Investigación y de Estudios Avanzados (CINVESTAV), Mexico

#### *\*Correspondence:*

Kumiko Nakada-Tsukui kumiko@nih.go.jp Tomoyoshi Nozaki nozaki@m.u-tokyo.ac.jp

†These authors have contributed equally to this work

#### *Specialty section:*

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

*Received:* 15 December 2018 *Accepted:* 23 April 2019 *Published:* 06 June 2019

#### *Citation:*

Nakada-Tsukui K, Watanabe N, Maehama T and Nozaki T (2019) Phosphatidylinositol Kinases and Phosphatases in Entamoeba histolytica. Front. Cell. Infect. Microbiol. 9:150. doi: 10.3389/fcimb.2019.00150 <sup>1</sup> Department of Parasitology, National Institute of Infectious Diseases, Tokyo, Japan, <sup>2</sup> Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Japan, <sup>3</sup> Division of Molecular and Cellular Biology, Graduate School of Medicine, Kobe University, Kobe, Japan, <sup>4</sup> Department of Biomedical Chemistry, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan

Phosphatidylinositol (PtdIns) metabolism is indispensable in eukaryotes. Phosphoinositides (PIs) are phosphorylated derivatives of PtdIns and consist of seven species generated by reversible phosphorylation of the inositol moieties at the positions 3, 4, and 5. Each of the seven PIs has a unique subcellular and membrane domain distribution. In the enteric protozoan parasite Entamoeba histolytica, it has been previously shown that the PIs phosphatidylinositol 3-phosphate (PtdIns3P), PtdIns(4,5)P2, and PtdIns(3,4,5)P<sup>3</sup> are localized to phagosomes/phagocytic cups, plasma membrane, and phagocytic cups, respectively. The localization of these PIs in E. histolytica is similar to that in mammalian cells, suggesting that PIs have orthologous functions in E. histolytica. In contrast, the conservation of the enzymes that metabolize PIs in this organism has not been well-documented. In this review, we summarized the full repertoire of the PI kinases and PI phosphatases found in E. histolytica via a genome-wide survey of the current genomic information. E. histolytica appears to have 10 PI kinases and 23 PI phosphatases. It has a panel of evolutionarily conserved enzymes that generate all the seven PI species. However, class II PI 3-kinases, type II PI 4-kinases, type III PI 5-phosphatases, and PI 4P-specific phosphatases are not present. Additionally, regulatory subunits of class I PI 3-kinases and type III PI 4-kinases have not been identified. Instead, homologs of class I PI 3-kinases and PTEN, a PI 3-phosphatase, exist as multiple isoforms, which likely reflects that elaborate signaling cascades mediated by PtdIns(3,4,5)P<sup>3</sup> are present in this organism. There are several enzymes that have the nuclear localization signal: one phosphatidylinositol phosphate (PIP) kinase, two PI 3-phosphatases, and one PI 5-phosphatase; this suggests that PI metabolism also has conserved roles related to nuclear functions in E. histolytica, as it does in model organisms.

Keywords: *Entamoeba histolytica*, phosphoinositide, kinase, phosphatase, signaling

# 1. INTRODUCTION

Phosphoinositides (PIs) are phosphorylatedphosphatidylinositol (PtdIns) derivatives and play pivotal roles in a variety of biological processes such as receptormediated signaling, vesicular traffic, cytoskeleton rearrangement, and regulation of channels and transporters (Sasaki et al., 2009; Balla, 2013). Spatiotemporal regulation of PI-mediated biological processes is achieved by interconversion of the phosphorylation states of PIs by specific kinases and phosphatases, followed by recruitment of PI-specific effectors. Phospholipids are ubiquitous in all three domains of life. Nevertheless, the complexity of PIs and enzymes that interconvert them appears to have increased in eukaryotes (Michell, 2008, 2011). It has been suggested that the PI metabolism developed in the last common eukaryotic ancestor (Michell, 2008) and diverged during eukaryotic evolution.

Human amebiasis is a common infection caused by the protozoan parasite Entamoeba histolytica in both developing and developed countries (Taniuchi et al., 2013; Lo et al., 2014; Ishikane et al., 2016), causing as far as 73,800 deaths annually (Lozano et al., 2012). The transmission usually occurs upon ingestion of water or food contaminated with E. histolytica cysts. The ingested cysts pass through the stomach and differentiate into trophozoites that colonize the colon. It is estimated that only 10–20% of individuals who are infected with E. histolytica develop symptoms (Gathiram and Jackson, 1985; Marie and Petri, 2014). The most common clinical manifestations in symptomatic cases are colitis and dysentery, and 5–10% of these are accompanied by invasive extraintestinal amebiasis, which is mostly amoebic liver abscess (Walsh, 1986).

Entamoeba histolytica belongs to the eukaryotic supergroup Amoebozoa, which is only distantly related to the eukaryotic model organisms in the Opisthokonta clade, including Saccharomyces cerevisiae, Caenorhabditis elegans, Drosophila melanogaster, and Homo sapiens. Various unique features of E. histolytica have been described due to its anaerobic/microaerophilic and parasitic life style, including metabolism of sulfur-containing amino acids, anaerobic energy generation, anti-oxidative stress mechanisms, and compartmentalization of sulfate activation to mitosomes, a unique mitochondria-related organelle (Ali and Nozaki, 2007; Müller et al., 2012; Makiuchi and Nozaki, 2014; Jeelani and Nozaki, 2016; Mi-Ichi et al., 2017; Pineda and Perdomo, 2017). Furthermore, the mechanisms regulating membrane trafficking in E. histolytica appear to be at least as complex as those found in higher eukaryotes. While most of the machineries underlying membrane-trafficking such as clathrin coats, coatomers, SNAREs, ESCRTs, and the retromer complex are conserved in E. histolytica (Nakada-Tsukui et al., 2005; Clark et al., 2007; Leung et al., 2008), unique evolutionary features in membrane trafficking are also apparent. For example, E. histolytica has numerous extremely diversified Rab small GTPases (104 genes) despite its unicellularity throughout its life cycle (Saito-Nakano et al., 2005; Nakada-Tsukui et al., 2010). In addition, a family of unique receptors that transport lysosomal hydrolase emerged in Entamoeba and related lineages during evolution (Furukawa et al., 2012, 2013; Nakada-Tsukui et al., 2012; Marumo et al., 2014). Although membrane trafficking in E. histolytica has been well-studied in the last few decades, E. histolytica PIs and PI metabolism are still relatively elusive despite the fact that they likely play critical roles in the physiology, especially in membrane trafficking, and pathogenicity of this organism (Raha et al., 1994, 1995; Giri et al., 1996; Makioka et al., 2001; Powell et al., 2006; Blazquez et al., 2008; Nakada-Tsukui et al., 2009; Byekova et al., 2010; Goldston et al., 2012; Koushik et al., 2013, 2014; López-Contreras et al., 2013; Lee et al., 2014; Bharadwaj et al., 2017). A previous genome-wide survey suggested that PI effectors found in other eukaryotes are not well-conserved in E. histolytica (Nakada-Tsukui et al., 2009). In this particular study, in order to better understand the level of conservation, elimination or diversification of the enzymes involved in the metabolism of E. histolytica PIs, we performed an extensive search for the potential kinases and phosphatases specific for the PIs found in the genome of this pathogen. Additionally, we summarized the known structural features and functions of similar enzymes in other organisms. To find and weigh the significance of possible homologs, we primarily used the E-values in the BLAST search. This was because E. histolytica homologs often differ in domain configurations and protein lengths to homologs in model organisms and the E-values better reflect both local and entire protein similarity. Such a comprehensive understanding of PI kinases and phosphatases will help us construct new hypotheses in future research.

# 2. GENERAL OVERVIEW ON INTRACELLULAR LOCALIZATION AND ROLES OF PIs

# 2.1. Definition, Structure, Synthesis, Transport, and Localization of PIs

### 2.1.1. Definition, Structure, Synthesis, and Transport of PIs

PtdIns consists of a glycerol backbone with two covalently bound fatty acids at the stereospecifically numbered (sn)-1 and 2 positions, and a D-myo-inositol head group linked via the sn-3 phosphate of glycerol. Three hydroxyl groups of the D-myoinositol head group (D3–5) are independently phosphorylated or dephosphorylated to form seven kinds of phosphorylated PtdIns (PIs) (**Figure 1**). PtdIns is synthesized in the endoplasmic reticulum (ER) from cytidine diphosphate diacylglycerol (CDP-DAG) and myo-inositol by PtdIns synthase (PIS) and transported to other cellular compartments either by vesicular transport or by PI transfer proteins (PITPs) (Di Paolo and De Camilli, 2006; Lev, 2010; Das and Nozaki, 2018). PtdInss are further metabolized to a variety of PIs on the membranes of these organelles (**Figure 1**).

# 2.1.2. Localization of PIs

PtdIns and PIs are concentrated at the cytosolic surface of the plasma membrane. Each PI type is enriched in a specific compartment(s) or sub-compartment(s) (Balla, 2013; Schink et al., 2016) (**Figure 2**). This disequilibrium in the type and distribution of PIs serves as a molecular tag to recruit specific effectors (Hammond and Balla, 2015; Várnai et al., 2017). In the

**Abbreviations:** see **Supplementary Table S6**.

model organisms, the distribution of PtdIns and PIs has been well-characterized. PtdIns4P and PtdIns(4,5)P<sup>2</sup> are enriched on the plasma membrane, where PtdIns(3,4)P<sup>2</sup> and PtdIns(3,4,5)P<sup>3</sup> are transiently generated in situ in response to extracellular stimuli or intracellular signaling (Di Paolo and De Camilli, 2006). PtdIns4P is enriched in the Golgi apparatus, where it regulates both intra-Golgi trafficking and the subsequent transport to the plasma membrane or the endosomal system (De Matteis et al., 2013). PtdIns3P is enriched in early endosomes and is known to trigger the recruitment of a number of effector proteins important for early endosomal identity and function (Di Paolo and De Camilli, 2006; Marat and Haucke, 2016; Schink et al., 2016). PtdIns(3,5)P2, converted from PtdIns3P, accumulates in the multivesicular bodies (MVBs) and late endosomes/lysosomes as early endosomes mature (Marat and Haucke, 2016). PtdIns5P is present in the nucleus, plasma membrane, and endomembranes including autophagosomes (Hammond and Balla, 2015; Vicinanza et al., 2015; Várnai et al., 2017), and functions in cytoskeleton regulation, and stress signaling pathways (Viaud et al., 2014). Except for PtdIns(3,4)P<sup>2</sup> and PtdIns(3,5)P2, nuclear localization of all the PIs has been reported (Ye and Ahn, 2008). Although PI metabolism in the nucleus is not fully understood, the involvement of nuclear PIs in transcription and chromatin remodeling in mammals, fly, yeast, and plant has been reported (Cheng and Shearn, 2004; Blind et al., 2012; Dieck et al., 2012; Shah et al., 2013; Poli et al., 2016).

# 2.2. Physiological Roles of PIs 2.2.1. Signaling via Phospholipase C-PtdIns(4,5)P<sup>2</sup> Breakdown

PIs are involved in signaling via two major pathways: as precursors of second messengers, and as regulators of various PI-specific effectors. The role of phospholipase C (PLC), which breaks down PtdIns(4,5)P<sup>2</sup> to inositol 1,4,5 trisphosphate[Ins(1,4,5)P3] and DAG, in the receptor-mediated growth signal pathway was first demonstrated in the early '80s (Michell et al., 1981; Berridge, 1983; Nishizuka, 1984; Michell, 1995). The role of PI turnover and PI-mediated signaling in cell proliferation is well-established (Berridge, 1984, 1987). PI turnover has also been implicated in the upstream signaling of Ca2<sup>+</sup> fluxes (Fain and Berridge, 1979). Given that the primary target of PLC is PtdIns(4,5)P<sup>2</sup> but not PtdIns (Berridge, 1983; Berridge et al., 1983; Creba et al., 1983), and Ins(1,4,5)P<sup>3</sup> is involved in Ca2<sup>+</sup> release from the ER, PLC and PtdIns(4,5)P<sup>2</sup> indirectly affect the regulation of non-mitochondrial Ca2<sup>+</sup> storage (Streb et al., 1983, 1984; Volpe et al., 1985). DAG activates the phospholipid-dependent kinase family, protein kinase C (PKC), and subsequently, the downstream signaling cascades (Nishizuka, 1984, 1995).

Ins(1,4,5)P3-mediated calcium signaling is conserved in a wide range of eukaryotes (Plattner and Verkhratsky, 2013). Interestingly, Ins(1,4,5)P<sup>3</sup> can be generated by an alternative pathway independent of PLC, and many protists do not have orthologous genes for the canonical Ins(1,4,5)P<sup>3</sup> receptor, which regulates Ca2<sup>+</sup> release from the ER (Kortholt et al., 2007; Plattner and Verkhratsky, 2013; Artemenko et al., 2014; Garcia et al., 2017). However, Ca2<sup>+</sup> release by Ins(1,4,5)P<sup>3</sup> has been observed

FIGURE 1 | Structures of phosphatidylinositol (PtdIns) and phosphoinositides (PI), and the routes of their interconversion. PtdIns consists of a glycerol backbone with two covalently attached fatty acids at the sn-1 and sn-2 positions, and a D-myo-inositol head group linked via the phosphate at the sn-3 position. Three hydroxyl groups of the D-myo-inositol head group (D3-5) are independently phosphorylated or dephosphorylated to form the seven kinds of phosphorylated PtdIns, PIs. Solid and broken arrows indicate kinase and phosphate reactions, respectively.

even in the organisms without an Ins(1,4,5)P<sup>3</sup> receptor. Besides, an orthologous gene has been identified in Trypanosoma cruzi, which is responsible for Chagas disease, suggesting some extent of conservation of the signaling pathway among eukaryotes (Hashimoto et al., 2013; Plattner and Verkhratsky, 2013).

# 2.2.2. Vesicular Traffic

PIs are involved in a variety of processes that involve vesicular trafficking, including secretion, recycling, endocytosis/phagocytosis, and autophagy (Frere et al., 2012; Balla, 2013; Swanson, 2014; Klinkert and Echard, 2016; Makowski et al., 2017; Wallroth and Haucke, 2018). A majority of secretory proteins are first transported into the ER lumen through the translocon on the ER membrane, then to the Golgi, where they are packaged into transport vesicles to be dispatched to endosomes or the plasma membrane. PtdIns4P, which is enriched in the Golgi, cooperatively works with PtdIns4P effectors such as GGA (Golgi-localized, gamma adaptin ear-containing, ARF-binding; a clathrin adaptor protein), Arf1, Ypt32p/Rab11 (small GTPase), and Sec2 (RabGEF) to form and target transport vesicles to the plasma membrane (De Matteis et al., 2013; Makowski et al., 2017). At the plasma membrane, PtdIns(4,5)P<sup>2</sup> cooperates with its effectors and promotes fusion of secretory vesicles with the plasma membrane (Li and Chin, 2003; Balla, 2013; Martin, 2015). PtdIns(4,5)P<sup>2</sup> at the plasma membrane is involved in the initiation of internalization processes such as endocytosis, micropinocytosis, and phagocytosis (Swanson, 2014; Wallroth and Haucke, 2018). During clathrin-mediated endocytosis, local synthesis of PtdIns(4,5)P<sup>2</sup> from PtdIns4P

by PIP kinases initiates clathrin-coated pit (CCP) formation. Subsequent conversion of PtdIns(4,5)P<sup>2</sup> to PtdIns(3,4)P<sup>2</sup> is necessary for CCP maturation. It has been demonstrated that elimination of PtdIns(4,5)P2, and generation of PtdIns(3,4)P<sup>2</sup>

and PtdIns3P on CCPs by PI 5-phosphatases and PI 3-kinases are the key events for maturation of endosomes (Nakatsu et al., 2010). Generation of PtdIns4P on endosomes and recruitment of PtdIns4P effectors have been reported to be necessary for recycling the plasma membrane proteins (Henmi et al., 2016). Both macropinocytosis and phagocytosis depend on actin reorganization, in which PI metabolism is known to be involved (Yeung and Grinstein, 2007; Swanson, 2014). Briefly, local accumulation of PtdIns(4,5)P<sup>2</sup> stimulates actin rearrangement to form the phagocytic/macropinocytic cup. The accumulated PtdIns(4,5)P<sup>2</sup> is then removed from the cup via three different mechanisms: hydrolysis by PLC to generate Ins(1,4,5)P<sup>3</sup> and DAG, phosphorylation by PI 3-kinase to generate PtdIns(3,4,5)P3, and dephosphorylation by PI 5-phosphatases (OCRL1 and INPP5B, see below sections) to generate PtdIns4P. Removal of PtdIns(4,5)P<sup>2</sup> from the cup causes actin dissociation and cup closure. On the nascent phagosomes, PtdIns3P accumulates by the action of type III PI 3-kinase and SHIP1/2 phosphatases. The generated PtdIns3P, alongside its effectors, engages in the early phase of phagosome/macropinosome maturation (Birkeland and Stenmark, 2004). In the later phase of the maturation, PtdIns3P is converted to PtdIns(3,5)P2, which drives sorting of cargos, such as carboxypeptidase S in yeast, and EGF receptor in human, into MVBs in cooperation with the ESCRT (endosomal sorting complex required for transport) complex (Odorizzi et al., 1998; Whitley et al., 2003). Autophagy is a mechanism necessary for bulk breakdown of cytoplasmic proteins and organelles (Mizushima et al., 2011). The unique serine/threonine kinase ULK1 (unc-51-like kinase 1, Atg1 in yeast) is activated during autophagy, and it subsequently activates the class III PI 3-kinase Vps34 complex to generate PtdIns3P on the autophagic membrane. This, in turn, recruits a variety of proteins involved in autophagosome formation (Marat and Haucke, 2016). PtdIns(3,5)P<sup>2</sup> synthesis has been reported to be required at the later phase of autophagosome maturation (Ferguson et al., 2009; Zou et al., 2012; Al-Qusairi et al., 2013).

# 2.2.3. Cytoskeletal Rearrangement, Motility, and Regulation of Transporters

As mentioned above, PtdIns4P and PtdIns(4,5)P<sup>2</sup> are the major PIs on the plasma membrane, and PtdIns(3,4,5)P<sup>3</sup> is transiently generated to provide a temporary signal. The importance of PtdIns(4,5)P<sup>2</sup> has been well-established by a number of studies, and it has been shown that the level of PtdIns(4,5)P<sup>2</sup> on the plasma membrane changes. As discussed above, PtdIns(4,5)P<sup>2</sup> is involved in signal transduction and endocytosis/phagocytosis (see sections 2.2.1 and 2.2.2). PtdIns(4,5)P<sup>2</sup> is also involved in the regulation of actin cytoskeleton and membrane channel activity (Balla, 2013; Hille et al., 2015; Schink et al., 2016).

During chemotaxis, chemoattractants are recognized by G-protein-coupled receptors (GPCRs) on the plasma membrane. This interaction leads to dissociation of the Gα heterodimer, which in turn activates PI 3-kinase to generate PtdIns(3,4,5)P<sup>3</sup> from PtdIns(4,5)P<sup>2</sup> on the cytoplasmic side of the plasma membrane. Local accumulation of PtdIns(3,4,5)P<sup>3</sup> causes translocation of actin-binding proteins (ABP) that interact with PtdIns(3,4,5)P3, and activates actin remodeling at the leading edge of the cell. On the contrary to these events at the leading edge, the PI 3-phosphatase PTEN (phosphatase and tensin homology located on chromosome 10), which converts PtdIns(3,4,5)P<sup>3</sup> to PtdIns(4.5)P<sup>2</sup> to cease the signal, has been shown to accumulate at the posterior side of the cell.

There are many ion channels regulated by PIs (Hilgemann and Ball, 1996; Hille et al., 2015). Kir2.2 is a member of the inwardly rectifying potassium channel family localized on the plasma membrane, and it is known to be activated upon interaction with PtdIns(4,5)P<sup>2</sup> (Hansen et al., 2011). Crystal structure analysis revealed that a direct interaction of PtdIns(4,5)P<sup>2</sup> with Kir2.2 induces a structural change on this channel. This, in turn, induces the channel to compress by pulling its cytoplasmic domain toward the potassium-selective pore on the membrane, shifting the channel to the active conformation (Rohács et al., 2003; Whorton and MacKinnon, 2011). Two possible advantages of PI dependence of ion channels have been suggested: (1) to achieve local activation of the channels depending on the lipid composition (i.e., no or decreased activity during synthesis and trafficking of the lipids to the target membrane) and (2) to swiftly regulate the channel activity by lipid modifying enzymes such as PLC, PI kinases, and PI phosphatases.

# 2.2.4. Nuclear Functions

Besides the various roles of PIs in the cytoplasm and the plasma membrane described above, PIs play indispensable roles in the nucleus. Localization of PIs, except for PtdIns(3,4)P<sup>2</sup> and PtdIns(3,5)P2, in the nuclear matrix has been demonstrated (Payrastre et al., 1992; Vann et al., 1997; Tanaka et al., 1999; Gillooly et al., 2000; Clarke et al., 2001). Since the nuclear matrix is hydrophilic, it is not well-understood how PIs remain soluble in the nucleus (York, 2006). The significance of PtdIns(4,5)P<sup>2</sup> and PtdIns5P has been well-demonstrated (Irvine, 2003; Poli et al., 2016; Hamann and Blind, 2018). PtdIns(4,5)P<sup>2</sup> is involved in the transcriptional regulation and chromatin remodeling by interacting with histones (Yu et al., 1998; Cheng and Shearn, 2004; Shah et al., 2013). PtdIns(4,5)P<sup>2</sup> also regulates cell cycle and differentiation through DAG generated by PLC-induced hydrolysis (Clarke et al., 2001; Newton, 2010; Poli et al., 2013, 2014). Nuclear DAG accumulation is followed by translocation of PKC to the nucleus for the phosphorylation of the target proteins (Neri et al., 1998). PtdIns5P is known to interact with TAF3, a component of the TATA box-binding protein complex, TFIID, and the chromatin-associating protein ING2, to regulate transcription and chromatin remodeling (Shi et al., 2006; Bua et al., 2013; Stijf-Bultsma et al., 2015). Interestingly, nuclear PI metabolism is regulated independently from cytoplasmic PI metabolism (Lindsay et al., 2006).

# 2.3. Spatiotemporal Regulation of PI-Mediated Signaling

# 2.3.1. PI-Specific Binding Proteins

Spatiotemporal regulation of PI-mediated signaling occurs in a variety of biological processes by various PI-specific effectors and enzymes that mediate interconversion of PIs. The seven phosphorylated PI species are enriched on specific membrane regions in both the cytoplasm and nucleus (Balla, 2013), and specifically recognized by PI effectors. This specific recognition occurs through the interaction of the PI-specific binding domains

of the effector proteins with the head groups of PIs (**Figure 1**). All the distinct PI-binding domains, consisting of a total of 24, have already been reported (Várnai et al., 2017).

# 2.3.2. Major Players of PI Interconversion

Each PI interconversion reaction is regulated by specific kinases or phosphatases (**Figure 1**). In mammals, 18 PI interconversion reactions have been identified, and these reactions are mediated by 19 PI kinases and 28 PI phosphatases (**Supplementary Tables S1**, **S2**, Sasaki et al., 2009). PI 3-, PI 4-, and PIP kinases use PtdIns as a substrate to generate PtdIns3P, PtdIns4P, and PtdIns5P, respectively. Mono-phosphorylated PIs are further phosphorylated by PIP kinases to generate PtdIns(3,4)P2, PtdIns(3,5)P2, and PtdIns(4,5)P2, which is further phosphorylated to generate PtdIns(3,4,5)P3. Each PI is dephosphorylated by a series of PI phosphatases such as PI 3-phosphatases (PTEN, MTM) PI 4-phosphatases (INPP4, TMEM55), and PI 5-phosphatases (Synaptojanin, OCRL1, INPP5, SHIP). It has been suggested that unique expression and localization patterns of PI kinases and PI phosphatases influence the local accumulation of PIs (Balla, 2013; Schink et al., 2016).

# 3. PREVIOUS FINDINGS ON THE ROLE OF PIs IN *E. histolytica*

E. histolytica trophozoites have been reported to have phospholipid compositions similar to those of mammalian cells except for the unique ceramide, ceramide aminoethylphosphonate (CEAP), which constitutes ∼15% of the total phospholipids (Aley et al., 1980). PI is a minor phospholipid component, constituting ∼5% of all phospholipids (Aley et al., 1980). Similarly, PI content of intracellular vesicles and the plasma membrane is <5%. While the plasma membrane contains less phosphatidylcoline than other membranes, it has a high content (40%) of CEAP (Aley et al., 1980). The resistance of the plasma membrane of trophozoites to the intrinsic pore forming peptide, amoebapores, is attributable to CEAP (Andrä et al., 2004).

Several previous studies demonstrated that PIs are involved in pathogenesis related processes such as adhesion, secretion, and phagocytosis. When the amebic trophozoites adhere to host cells, Gal/GalNAc lectin serves as a major adhesion molecule and transduces the signals. It is composed of heavy (Hgl), intermediate (Igl), and light (Lgl) subunits, of which Igl and Lgl are GPI-anchored. The downstream cytosolic signals transmitted from the lectin have not been well-investigated except for one example (Hughes et al., 2003). However, it has been shown that PtdIns(4,5)P2- and cholesterol-dependent enrichment of Gal/GalNAc lectin subunits to lipid rafts causes an increment of Ca2<sup>+</sup> level followed by adhesion to the mammalian cell (Welter et al., 2011; Goldston et al., 2012). These results suggest involvement of PtdIns(4,5)P2-mediated Ca2<sup>+</sup> signaling during cell adhesion (Goldston et al., 2012).

Cysteine proteases (CPs) are the major virulence factors. They are secreted via the default brefeldin A-sensitive or unique brefeldin A-insensitive pathways, and Rab11B-dependent pathways (Manning-Cela et al., 2003; Mitra et al., 2007). In the model organisms, it has been established that Rab11 on the secretory vesicles, and Sec3 of the exocyst complex interact, leading to tethering of the Rab11 vesicles to the plasma membrane in a PtdIns(4,5)P2-dependent manner (He et al., 2007; Zhang et al., 2008; Wu and Guo, 2015). The components of the exocyst complex including PtdIns(4,5)P2 binding Sec3 and Exo70 are mostly conserved in E. histolytica. Thus, it is conceivable that PI-regulated secretion takes place in E. histolytica.

PIs are also involved in phagocytosis in E. histolytica as in mammals. Several studies in which amebic transformants that expressed PI-binding proteins fused with green fluorescent protein (GFP) or in which recombinant glutathione S-transferase (GST) were used as bioprobes demonstrated that PtdIns(4,5)P<sup>2</sup> were localized on the plasma membrane, while PtdIns(3,4,5)P<sup>3</sup> were localized on the extended pseudopodia, and phagocytic cups, and PtdIns3P on the phagocytic cups, nascent phagosomes, and internal vesicles (Powell et al., 2006; Nakada-Tsukui et al., 2009; Byekova et al., 2010; Koushik et al., 2013). Localization of PtdIns(3,4,5)P<sup>3</sup> and PtdIns3P is similar during E. histolytica phagocytosis and macrophage phagocytosis (Yeung and Grinstein, 2007). It has been recently shown that AGC kinases 1 and 2 that bind to PtdIns(3,4,5)P<sup>3</sup> or PtdIns(3,4)P<sup>2</sup> are localized to the contact site upon interaction with mammalian cells (Somlata et al., 2017). Interestingly, AGC kinases 1 and 2 have different localization patterns, although their apparent PtdIns specificities have been demonstrated with lipid overlay assay. AGC kinase 2 localizes to a tunnel-like structure proximal to the primary trogocytic cup and adjacent to the contact site on the plasma membrane during trogocytosis ("trogo" means "nibble" or "chew," and trogocytosis is the process of internalizing live cells by nibbling them). In contrast, AGC kinase 1 is confined to the intermediate part of the trogocytic tunnel. Such an observation has been made in E. histolytica but not in professional phagocytes of multicellular organisms, including mammals.

Cytoskeletons also play indispensable roles during phagocytosis and trogocytosis. EhRho1, which is involved in actin rearrangement via EhFormin1 and EhProfilin1 (Bharadwaj et al., 2018), has been shown to regulate membrane blebbing to initiate internalization of the prey through PI 3-kinases (Bharadwaj et al., 2017). Inducible expression of a constitutively active EhRho1 increased the PtdIns(3,4,5)P<sup>3</sup> level and reduced PtdIns(4,5)P<sup>2</sup> level, whereas expression of a dominant negative EhRho1 caused opposite effects (Bharadwaj et al., 2017). EhRho1 is considered to be orthologous to HsRhoA as the amino acid sequences of their ROCK-binding domains are 65% identical. Moreover, EhRho1 complements HsRhoA activity in HEK 293T cells (Bharadwaj et al., 2017). Interestingly, HsRhoA does not localize to the phagocytic cup in mammalian cells unlike EhRho1. Signaling transduced downstream of PtdIns3P in mammals appears to be different in E. histolytica, because E. histolytica does not seem to be equipped with the orthologs of known mammalian PtdIns3P effectors. Transformation of E. histolytica with GFP-fused human Hrs showed that PtdIns3P was concentrated on phagosomes, more specifically at the bottom of the phagocytic cup during the early phase of phagocytosis (prior to the closure of the phagosome) (Nakada-Tsukui et al., 2009). Two PtdIns3P-binding domains are known: Phox homology (PX) and Fab-1–YGL023–Vps27–EEA1 (FYVE) domains. E. histolytica apparently has two PX and twelve FYVE domain-containing proteins (Nakada-Tsukui et al., 2009; N. Watanabe, data not shown). It has been demonstrated that eleven out of the twelve E. histolytica FYVE domain-containing proteins (EhFPs) also have a RhoGEF domain, and one of EhFP (EhFP4) preferentially binds to PtdIns4P and localizes to the plasma membrane proximal to the phagosome that is not yet closed (Nakada-Tsukui et al., 2009). Surprisingly, the C-terminal domain instead of the FYVE domain of EhFP4 binds to PtdIns3P, PtdIns4P, and PtdIns5P (Nakada-Tsukui et al., 2009). In model organisms, phagosomal PtdIns3P has been shown to recruit FYVE domain-containing proteins, which are subsequently involved in maturation of the phagosomes. However, like RhoGEF, EhFP4 appears to be primarily involved in actin rearrangement during phagocytosis eventhough full length EhFP4 does not seem to recognize PtdIns3P. Additional PtdIns3P effectors on phagosomes, most likely to be PX domain-containing proteins, still remain elusive.

# 4. PI 3-KINASES

Phosphorylation of PtdIns and PIs is initially observed as conversion of PtdIns to PtdIns4P, and PtdIns4P to PtdIns(4,5)P<sup>2</sup> (Balla, 2013). The enzymes responsible for these activities are named PtdIns kinases and PI kinases, respectively. Currently, it is known that some enzymes can phosphorylate both PtdIns and PIs. The enzyme classification given in this review is based on the position of the hydroxyl group that the enzymes can phosphorylate.

PI 3-kinases phosphorylate the hydroxyl group at the D3 position of the inositol ring of PtdIns, PtdIns4P, and PtdIns(4,5)P<sup>2</sup> to generate PtdIns3P, PtdIns(3,4)P2, and PtdIns (3,4,5)P3, respectively. There are three subfamilies of PI 3 kinases: class I, II, and III (Sasaki et al., 2009). In general, class I enzymes preferentially generate PtdIns(3,4,5)P<sup>3</sup> from PtdIns(4,5)P2. Class II enzymes mostly generate PtdIns(3,4)P<sup>2</sup> from PtdIns4P and also generate PtdIns3P from PtdIns. Class III enzymes almost exclusively generate PtdIns3P from PtdIns. In mammals, there are a total of 8 members of PI 3-kinases. All PI 3-kinases contain a "signature motif " consisting of the catalytic kinase domain, a helical domain, also called "lipid kinase unique (LKU) domain," and a membrane-binding C2 domain (Vanhaesebroeck et al., 2010a; Balla, 2013; Marat and Haucke, 2016). The class I to III classification of PI 3 kinases is mainly based on the presence of additional protein domains and their interactions with regulatory subunits. Class I enzymes have an adaptor binding domain (ABD), and regulatory subunit-binding and Ras binding domains (RBD). Class II enzymes have an N-terminal extension, which is involved in clathrin binding, and a C-terminal PX and extra C2 domains. It is of note that this PX domain in class II PI 3-kinase is known to preferentially bind to PtdIns(4,5)P<sup>2</sup>

(Stahelin et al., 2006). These domains are involved in subcellular localization and activity of the enzyme, and downstream effector selection. The class I and class III enzymes have regulatory subunits which modulate localization and activity of these enzymes.

# 4.1. Class I PI 3-Kinase

# 4.1.1. General Description of Class I PI 3-Kinase

Class I PI 3-kinases predominantly produce PtdIns(3,4,5)P<sup>3</sup> from PtdIns(4,5)P2. There are two kinds of class I PI 3-kinases based on their composition of catalytic and regulatory subunits. One of the three class IA catalytic subunits (p110α, β, and δ) associates with one of the five p85 class regulatory subunits (p85α, p85β, p55α, p55γ, and p50α), while the class IB catalytic subunit (p110γ) associates with one of the two P101/p87 class regulatory subunits (p101 and p87) (Vadas et al., 2011; Jean and Kiger, 2014) (**Figure 3**). In mammals, p110α and p110β are expressed ubiquitously, while p110δ and p110γ seem to be restricted to hematopoietic cells. The class IA catalytic subunits (p110α, β, and δ) are activated via receptor tyrosine kinases and generate PtdIns(3,4,5)P<sup>3</sup> at the plasma membrane. On the other hand, the p110β and p110γ catalytic subunits are activated downstream of the GPCR (Stephens et al., 1994; Stoyanov et al., 1995; Vanhaesebroeck et al., 2010a). PtdIns(3,4,5)P<sup>3</sup> generation causes recruitment of PI effectors, such as the protein kinase Akt, also named protein kinase B (PKB) (James et al., 1996; Ma et al., 2008; Rodgers et al., 2017). Activated Akt is involved in cell survival and metabolism via various cellular processes, including those that involve mammalian target of rapamycin complex-1 (mTORC1), the pro-apoptotic factor BAD, and FOXO transcription factors (Vanhaesebroeck et al., 2010a; Dibble and Cantley, 2015). Due to their crucial roles in cell growth and proliferation, dominant activating mutations of the class I PI 3-kinases are known to be associated with cancers, making PI 3-kinases potential drug targets (Vanhaesebroeck et al., 2010b).

# 4.1.2. Class I PI 3-Kinase of E. histolytica

In the E. histolytica genome (http://amoebadb.org/amoeba/), six potential catalytic subunits of class I PI 3-kinases were identified by probing the genome with the catalytic subunits of human class I PI 3-kinases as queries (NP\_006209, NP\_006210, AAH35683, and NP\_005017, corresponding to p110α, β, γ, and δ, respectively). Independent of the query used, ten proteins were identified to share significant overall similarity, reflecting a possible redundancy among them (E-value < 1 × 10−10). Six of them have conserved domains, such as RBD, C2, LKU, and PI 3-kinase catalytic domains, but no protein with the ABD was identified (**Figure 3**; **Supplementary Table S1**). Four additional proteins were also identified during this survey: a class III PI 3-kinase (EHI\_096560) and type III PI 4 kinase (EHI\_148700) (see below), a catalytic domain-only protein (EHI\_127850), and a protein that lacked the LKU domain (EHI\_073560). As PI 3-kinase catalytic domain is also conserved in PI 3- and PI 4-kinases (Vogt et al., 2007), the identification of both class III PI 3-kinase and type III PI 4-kinase homologs during this search is understandable. In the present review, we tentatively designated the proteins that


FIGURE 3 | Structural features of PI kinases of H. sapiens and E. histolytica. Structural features and domain organization of PI kinases, including their regulatory subunits are shown. Numbers showing at the end of the protein indicates amino acid length. ABD, adaptor binding domain; AR, acidic region; BH, Bcl Homology; C2,C2 domain; Catalytic, lipid kinase domain of PI 3- and PI 4-kinases; DEP, disheveled, Egl-10 and pleckstrin domain; FYVE, Fab1, YOTB, Vac1, and EEA1 domain; HEAT, Huntington, Elongation factor3, PR65/A, and TOR; Kinase, Ser/Thr kinase domain; LKU, lipid kinase unique domain; P, Proline-rich; PDZ-B, PDZ domain binding domain; PH, Pleckstrin-homology; PIPKc, kinase core domain of PIP kinases; PX, Phox homology; Rab-BD, Rab binding domain; RBD, Ras binding domain; SH2, Src homology 2; SH3, Src homology 3; TPC-1, T-complex 1 homology; TPR, tetratricopeptide repeat; WD40, WD40 repeat. Myristoylation, palmitoylation sites, and the nuclear localization signal are also depicted with "\*", " ", or "H", respectively.

contain LKU and catalytic domains as class I PI 3-kinases, which excluded the four additional proteins mentioned above (**Figure 3**; **Supplementary Table S1**). Among the six class I PI 3-kinases, EHI\_040690 showed a lower E-value to class II PI 3 kinase. The E-values with p110β and PI 3-kinase-C2α were 1 × 10−<sup>107</sup> and 2 × 10−110, respectively (**Supplementary Table S3** and N. Watanabe, data not shown). However, because it lacks the C-terminal domains and catalytic domain, we included this gene among the class I PI 3-kinases (also refer section 4.2.2). The six proteins with conserved RBD, C2, LKU, and PI 3-kinase catalytic domains cannot be further classified into p110α, β, γ, or δ, as none of them has the ABD and show only marginal E-value to the ABD-containing proteins p110α, β, or δ (**Supplementary Table S3**; **Supplementary Figure S4**). Furthermore, all potential class I PI 3-kinase catalytic subunit homologs showed the lowest E-value to p110β, but not for p110γ despite the fact that all the amebic homologs lack the ABD and structurally resemble p110γ (**Supplementary Table S3**). No homologs of the regulatory subunits that contain Src homology 2 (SH2) domain were identified in the E. histolytica genome database when p85α, β, p55α, p50α, and p55γ were used as the queries. Furthermore, only five proteins were predicted to have an SH2 domain and four of them were annotated as protein kinases, while the remaining protein was predicted to have a role in RNA stability and/or transcriptional regulation, with no possible link to PI 3-kinase regulatory subunits. These data suggest the possibility that the regulatory subunits of class I PI 3-kinase have been lost or replaced with a lineage-specific protein in E. histolytica during evolution. In Saccharomyces cerevisiae, class I and II PI 3-kinases are not conserved. Dictyostelium discoideum has catalytic but not regulatory subunits of three class I PI 3-kinases and lacks class II PI 3-kinases (Engelman et al., 2006). The catalytic subunits of D. discoideum class I PI 3-kinases also lack the ABD as in E. histolytica. Such lineage-specific modifications of the catalytic subunits and loss of the regulatory subunits of class I PI 3-kinases likely suggest divergence of PtdIns(3,4,5)P3 mediated lipid signaling in eukaryotes. It should be noted that E. histolytica has six PtdIns(3,4,5)P<sup>3</sup> phosphatase homologs of PTEN, while there is only one PTEN gene in the human genome (see section 7.1).

# 4.2. Class II PI 3-Kinase

# 4.2.1. General Description of Class II PI 3-Kinase

Class II PI 3-kinases are monomeric enzymes that generate PtdIns(3,4)P<sup>2</sup> and PtdIns3P from PtdIns4P and PtdIns, respectively (Balla, 2013; Maffucci and Falasca, 2014). There are three subtypes: PI 3-kinase C2α, β, and γ, among which PI 3-kinase C2α and β have N-terminal extensions that are likely involved in autoinhibition and protein-protein interactions with clathrin (Marat and Haucke, 2016). Except for the N-terminal extensions, all class II PI 3-kinases contain one RBD, two C2, one LKU, one catalytic, and one PX domains. PI 3-kinase C2α and β isoforms are ubiquitously expressed, whereas the γ isoform is largely restricted to the liver. This class of PI 3-kinases does not have the regulatory subunit; however, they are regulated by interacting with proteins such as clathrin and Rab5 small GTPase. Clathrin associates with PI 3-kinase C2α and β isoforms through the N-terminal extension, and Rab5 interacts with the γ isoform via the RBD (Gaidarov et al., 2001, 2005; Braccini et al., 2015). Accumulating evidence suggests that class II PI 3-kinases are involved in the regulation of membrane trafficking from the plasma membrane via PtdIns(3,4)P<sup>2</sup> synthesis. PI 3-kinase C2α is involved in clathrin-mediated endocytosis by the formation of PtdIns(3,4)P2, which constricts the membrane by recruiting PX and BAR domain-containing sorting nexin (SNX) SNX9 (Posor et al., 2013; Schöneberg et al., 2017). PI 3-kinase C2γ is recruited to endosomes as Rab5 effector for PtdIns(3,4)P<sup>2</sup> synthesis, which is indispensable for delayed and sustained activation of Akt2 in the liver (Braccini et al., 2015). It was also suggested that PI 3-kinase C2α and β also play a role in the regulation of intracellular PtdIns3P levels and directly or indirectly regulate membrane traffic and autophagy (Jean et al., 2012; Devereaux et al., 2013; Franco et al., 2014).

# 4.2.2. Class II PI 3-Kinase From E. histolytica

In E. histolytica, we concluded that there are no class II PI 3-kinases. When three human class II PI 3-kinases were used as queries, the best hits we obtained were the same proteins identified as class I PI 3-kinases (see above). As described above, because of the low similarity to class II PI 3-kinases in five out of six candidates and the absence of the PX domain in all the six, they were classified into class I PI 3-kinases. It is of note that EHI\_040690 showed a lower E-value with PI3KC2α (2 <sup>×</sup> <sup>10</sup>−<sup>110</sup> with PI3KC2<sup>α</sup> and 1 <sup>×</sup> <sup>10</sup>−<sup>107</sup> with class I PI 3-kinase, p110β). Additionally, class II PI 3-kinases evolved after Metazoa, and another amoeboid organism, D. discoideum, lacks this class of PI 3-kinases (Engelman et al., 2006; Brown and Auger, 2011). According to these contexts, we decided to conclude that there are no class II PI 3-kinases in E. histolytica. However, the conservation of a gene showed low E-value with the class II PI 3-kinase, suggesting the possibility that some of the class I PI 3-kinases have a role similar to that of class II PI 3-kinases.

# 4.3. Class III PI 3-Kinase

# 4.3.1. General Description of Class III PI 3-Kinase

The human genome has one class III PI 3-kinase, vacuolar protein sorting (Vps) 34, which phosphorylates the D3 position of PtdIns. Vps34 gene was first identified as a temperature-sensitive mutation that impairs the sorting of vacuolar hydrolases in yeast (Herman and Emr, 1990; Schu et al., 1993). Vps34 consists of one of each of the C2, LKU, and catalytic domains, and forms a dimer with the p150 regulatory subunit (Vps15 in yeast). p150 constitutively interacts with Vps34, and the myristoyl modification in its amino terminal links Vps34 to the membrane (Stack et al., 1993; Vanhaesebroeck et al., 2010a). Since Vps34 is the only PI 3-kinase in yeast, and also widely conserved in Eukaryota, Vps34 is considered to be the ancestral PI 3-kinase (Schu et al., 1993; Engelman et al., 2006; Brown and Auger, 2011). Vps34 participate in membrane trafficking, endocytosis, phagocytosis, and autophagy through the synthesis of PtdIns3P (Sasaki et al., 2009; Swanson, 2014; Wallroth and Haucke,

2018; also see section 2.2.2). In the endocytic pathway, early endosomes mature as PtdIns3P is synthesized in situ, subsequently recruiting Rab5 and Rab7 to early and late endosomes, respectively. Vps34 was identified as one of the mutual effectors of Rab5 and Rab7, involved in spatiotemporal generation of PtdIns3P on endosomal membranes (Christoforidis et al., 1999; Stein et al., 2003; Shin et al., 2005).

Vps34 has been shown to form two kinds of complexes that differ in localization and function (Marat and Haucke, 2016). Complex I consists of p150, and the mammalian orthologs of yeast Vps30, Atg14, and Atg38 (Beclin-1, ATG14L, and NRBF2, respectively). It is involved in autophagy (Itakura et al., 2008; Cao et al., 2014; Lu et al., 2014). In contrast, complex II consists of p150, Beclin-1, and UVRAG, which is the mammalian ortholog of yeast Vps38. Complex II is involved in the regulation of endosome and autophagosome maturation (Kihara et al., 2001; Matsunaga et al., 2009; Funderburk et al., 2010; Sun et al., 2010; Rostislavleva et al., 2015). It is also known that, in addition to its role in autophagy Vps34 functions as an amino acid sensor, and regulates mTORC1 activity and localization (Munson et al., 2015; Hong et al., 2017). These observations suggest multiple roles of Vps34 at the cross road of nutrient sensing and membrane trafficking. Vps34 is also involved in the negative regulation of autophagy through amino acid sensing (Furuya et al., 2005; Gulati and Thomas, 2007) and mTORC1 activation mediated by PtdIns3P-dependent recruitment of phospholipase D1 (PLD1) (Yoon et al., 2011; Bridges et al., 2012). Activated mTORC1 inhibits the autophagy-promoting activity of the Complex I by phosphorylating Atg14L in the complex (Yuan et al., 2013), while it activates the Complex II by phosphorylating UVRAG. Activation of the Complex II, in turn, leads to activation of Vps34 during the reformation of lysosomes from autophagosomes following recovery from starvation (Yu et al., 2010; Munson et al., 2015; Chen and Yu, 2017). Thus, Vps34-containing complexes are interactive and involved in eliciting opposite effects in the cell.

# 4.3.2. Class III PI 3-Kinase of E. histolytica

In E. histolytica, there are one of each Vps34 and p150 homolog (EHI\_096560 and EHI\_044190, respectively). Although neither their localization nor function have been demonstrated, roles of PtdIns3P are well-established as previously described (Powell et al., 2006; Nakada-Tsukui et al., 2009). During trogocytosis, which is ingestion by nibbling live mammalian cells (Ralston et al., 2014; Somlata et al., 2017), unclosed and nascent trogosomes are decorated with PtdIns3P. While localization of PtdIns3P to endosomes per se has not been well-documented, its localization to MVB-containing endosomes has been demonstrated (Nakada-Tsukui et al., 2009), suggesting a conserved role of PtdIns3P in the endocytic pathway in E. histolytica. It is conceivable that Vps34 is involved in the synthesis of PtdIns3P on trogosomes. E. histolytica has two TOR (EHI\_155160 and EHI\_169320) and two Atg8 homologs (EHI\_130660 and EHI\_172140). It is thus expected that E. histolytica Vps34 may also play a role in the response to starvation.

# 5. PI 4-KINASES

Among seven PtdIns isotypes, PtdIns(4,5)P<sup>2</sup> is the most abundant and well-studied in the context of PI turnover (see section 2.2.1). Since PI 4-kinase is one of the major enzymes responsible for producing the precursor of PtdIns(4,5)P2, it plays a significant role by producing PtdIns4P (Wang et al., 2003; D'Angelo et al., 2008). Various roles have been suggested for PtdIns4P and PI 4-kinases, including signaling on the plasma membrane (Tan and Brill, 2014). Two types of PI 4-kinases are currently known in humans: type II and type III. Type I PI 4-kinase, which was initially identified in a bovine brain homogenate chromatography fraction that showed PI kinase activity has turned out to be identical to PI 3-kinase, and thus it is no longer referred (Whitman et al., 1988). The human genome encodes two isotypes of both type II and type III PI 4-kinase. Type II and III PI 4-kinases differ in their domain structure and sensitivity to wortmannin, since the former is insensitive unlike the latter.

# 5.1. Type II PI 4-Kinase

### 5.1.1. General Description of Type II PI 4-Kinase

Type II PI 4-kinases (PI4KII) contain a large lipid kinase domain that is separated by a long non-conserved insert. This structure is significantly different from that of type III PI 4-kinases (PI4KIII) whose catalytic domain consists of LKU and catalytic domains. It was inferred by phylogenetic analyses that type II PI 4 kinases are evolutionarily different from type III PI 4-kinases. Furthermore, type III PI 4-kinases share significant homology with the typical protein kinase PKA and PI 3-kinases (Minogue and Waugh, 2012). The catalytic domains of PI4KIIα and β are highly similar, but their N-terminal regions are divergent. The N-terminal proline-rich region (P) in PI4KIIα and acidic region (AR) in β have been shown to interact with AP-3 and AP-1 adaptor complexes, respectively (Salazar et al., 2005; Wieffer et al., 2013). Initially, PI 4-kinases were expected to have a role in the generation of PtdIns4P as a precursor of PtdIns(4,5)P2, whereby they were thought to regulate signal transduction from the plasma membrane. However, it has recently been suggested that type II PI 4-kinases are mostly involved in the regulation of endomembrane sorting machinery. They do so mostly in the trans-Golgi network (TGN), which functions as a sorting hub. To date, there are four suggested roles of PI4KIIα and β during membrane trafficking: (1) cargo trafficking between the TGN and internal vesicles via interaction with adaptor proteins such as AP-1 and AP-3 (Wang et al., 2003; Salazar et al., 2005; Minogue et al., 2006; Wieffer et al., 2013); (2) synthesis of PtdIns4P on mature phagosomes/autophagosomes and regulation of fusion with lysosomes (Jeschke et al., 2015; Levin et al., 2017); (3) outbound traffic toward the plasma membrane (Husebye et al., 1990; Barylko et al., 2001; Xu et al., 2006); and (4) regulation of actin-dependent trafficking by interacting with actin regulatory proteins, such as RhoGEF1, and Wiskott-Aldrich Syndrome and SCAR homolog (WASH) complex components (Mössinger et al., 2012; Ryder et al., 2013; Gokhale et al., 2016). It has been shown in mammalian cells that the two isotypes of PI4KII are differently regulated due to differences in the regulatory proteins they interact with. Both PI4KII isotypes are palmitoylated at the CCPCC motif in the catalytic domain; however, only PI4KIIβ has the ability to bind to HSP90, and the interaction is disrupted upon stimulation by epidermal and platelet-derived growth factors (Jung et al., 2011). This association with HSP90 enables stabilization of the lipid-modified PI4KIIβ in the cytosol by preventing its proteasomal degradation (Jung et al., 2011). Such elaborate mechanisms enable isotype-specific regulation of PI4KII.

# 5.1.2. Type II PI 4-Kinase of E. histolytica

In E. histolytica, no ortholog with the E-value < 1 × 10−<sup>4</sup> was identified by using human type II PI 4-kinases as a query in the BLAST search. It has been reported that a majority of parasitic protists including Trypanosoma, Leishmania, Trichomonas vaginalis, Giardia lamblia, and E. histolytica, apparently lack type II PI 4-kinases (Brown and Auger, 2011). However, apicomplexans such as Plasmodium exceptionally conserve a type II PI 4-kinase. Fungal and apicomplexan type II PI 4-kinase orthologs are closely related to those found in metazoans and plants, respectively. This observation is consistent with the current understanding of the evolution scheme that fungi are closely related to metazoans, and Apicomplexa acquired a plant-associated enzyme together with the plastid-like apicoplast by endosymbios (Baldauf and Palmer, 1993; McFadden, 2000). Although Entamoeba appears to lack type II PI 4-kinases, there is a possibility that Entamoeba and other organisms that lack type II PI 4-kinase may have a novel type of PI 4-kinase that is yet to be identified.

# 5.2. Type III PI 4-Kinase

# 5.2.1. General Description of Type III PI 4-Kinase

Type III PI 4-kinases contain a continuous (uninterrupted) catalytic domain like PI 3-kinases, and both kinase types similarly show wortmannin sensitivity. Different from type II PI 4 kinases, the type III enzymes have a lipid kinase unique (LKU) domain, which is conserved among PI 3-kinases (Balla, 2013). As described above, the primary role of type III PI 4-kinases is generation of PtdIns4P, a precursor of PtdIns(4,5)P2, at the plasma membrane. PI4KIIIα has been shown to be recruited to the plasma membrane by interacting with two binding proteins, EFR3B and TTC7B, which are the mammalian homologs of yeast Efr3 and Ypp1 (Baird et al., 2008, see below). Additionally, knocking down PI4KIIIα causes reduction in PtdIns4P and PtdIns(4,5)P<sup>2</sup> level at the plasma membrane (Nakatsu et al., 2012). Notably, in the PI4KIIIα knockout mouse embryonic fibroblast (MEF) cells, the total cellular level of PtdIns(4,5)P<sup>2</sup> did not change due to the compensatory upregulation of PIPKIβ and γ, which also generate PtdIns(4,5)P<sup>2</sup> from PtdIns4P. However, the level of PtdIns(4,5)P<sup>2</sup> in the internal vesicles increased in the PI4KIIIα-knockout MEF cells. Several plasma membrane proteins such as M1 muscarinic receptor, and myristoylated/palmitoylated N-terminal anchor of LCK have been demonstrated to be concentrated in the internal vesicles where PtdIns(4,5)P<sup>2</sup> is also enriched. These results suggest that PI4KIIIα gives unique properties to the plasma membrane, and thus lack of PI4KIIIα perturbs the membrane identity (Nakatsu et al., 2012).

Of two isotypes in humans, PI4KIIIα contains a bipartite nuclear localization sequence (NLS) and PH domain (Heilmeyer et al., 2003). In contrast, PI4KIIIβ does not have either of these domains; however, it contains several stretches rich in basic amino acids and leucine-rich sequences that can potentially serve as nuclear localization and export signals, overall suggesting their nuclear localization (Heilmeyer et al., 2003). Both PI4KIIIs have indeed been detected in the nucleus, and the yeast homolog of PI4KIIIβ, Pik1p, has been shown to shuttle between the cytosol and nucleus, suggesting its contribution to the PI pools in the nuclear speckles (Garcia-Bustos et al., 1994; de Graaf et al., 2002; Heilmeyer et al., 2003; Demmel et al., 2008; Mellman et al., 2008; Barlow et al., 2010). PI4KIIIβ plays a role as Rab11 effector, and participate in the recruitment of Rab11 to the Golgi and TGN (de Graaf et al., 2004). The crystal structures of PI4KIIIβ, Rab11, and Rab11 effector FIP3 revealed that PI4KIIIβ-Rab11 binding is independent of the kinase activity of PI4KIIIβ, which suggests a role of PI4KIIIβ other than PI phosphorylation (Burke et al., 2014). While type II PI 4-kinases are palmitoylated, type III PI 4-kinases are soluble and present in the cytosol. For membrane association, they interact with other proteins that have membrane affinity. PI4KIIIα has been shown to bind to TTC7 and EFR3 (Baird et al., 2008; Nakatsu et al., 2012). These proteins function as a scaffold for PI4KIIIα (Wu et al., 2014). For instance, EFR3 binds to acidic phospholipids, whereby it recruits the enzyme complex to the plasma membrane (Nakatsu et al., 2012). On the other hand, PIK4IIIβ binds to neuronal calcium sensor 1 (NCS-1), acyl-CoA-binding domain containing protein 3 (ACBD3), 14– 3–3, and ADP-ribosylation factor 1 (Arf1) (Zhao et al., 2001; Hausser et al., 2006; Hsu et al., 2010; Sasaki et al., 2012; Klima et al., 2016). NCS-1 is a myristoylated calcium binding protein involved in membrane recruitment and activation of PI4KIIIβ. ACBD3 is a Golgi adaptor protein involved in the recruitment of PI4KIIIβ to the Golgi. Arf1 is a Golgi-localized small GTPase and its activation enhances binding and activity of PI4KIIIβ. 14–3–3 is a phosphoserine/threonine-binding protein. It binds to protein kinase D-phosphorylated PI4KIIIβ and this interaction stabilizes PI4KIIIβ activity. These binding proteins are the key regulators of type III PI 4-kinases.

# 5.2.2. Type III PI 4-Kinase of E. histolytica

Entamoeba histolytica has only one homolog of PI4KIIIα and PI4KIIIβ (EHI\_148700). As described above, E. histolytica does not have a type II PI 4-kinases, and EHI\_148700 is the only potential PI 4-kinase in this organism. NLS search did not indicate presence of NLS on EHI\_1478700 (http://nls-mapper. iab.keio.ac.jp/cgi-bin/NLS\_Mapper\_form.cgi) (Heilmeyer et al., 2003; Kosugi et al., 2009). Considering the analogy of E. histolytica to other organisms and also the fact that its type I PIP kinase is predicted to have NLS (see section 6.1.2 below), it is reasonable to speculate that E. histolytica also has a PtdIns4P pool in the nucleus. If so, as described above for PI4KIIIβ, basic amino acid-stretches and leucine-rich sequences in EHI\_148700 may function as a nuclear localization signal. This potential E. histolytica type III PI 4-kinase is a soluble

protein and predicted to associate with the plasma membrane through its binding proteins. However, no orthologs for the known PI4KIIIα-binding proteins TTC7 and EFR3 have been identified with an E-value lower than 1 × 10−10. EHI\_118850 has been identified during a similarity search using human TTC7 as the query, and thus the two proteins are thought to be homologs. However, although TTC7 has three tetratricopeptide repeat (TPR) domains, EHI\_118850 has two. No homologs of human/yeast ERF3 have been identified in E. histolytica; however, one should note that human and yeast ERF3 share only low (19.4%) amino acid homology according to Clustal Omega alignment, and no DNA similarity was detected by BLAST search. Murine homologs of EFR3 and TTC7 were identified from the PI4KIIIα immunoprecipitates of mouse brain extract (Nakatsu et al., 2012). Thus, biochemical approaches must be pursued in E. histolytica for the identification of its proteins that are functionally homologous to TTC7 and ERF3. It is also worth mentioning that type III PI4K of P. falciparum has been exploited for the development of antimalarials (McNamara et al., 2013; Kandepedu et al., 2018).

# 6. PHOSPHATIDYLINOSITOL PHOSPHATE KINASES (PIP KINASES)

PIP kinases have a unique catalytic domain that is not homologous to any other known lipid or protein kinases. There are three types of PIP kinases based on the substrate specificities. Type I and II PIP kinases generate PtdIns(4,5)P<sup>2</sup> from PtdIns4P and PtdIns5P, respectively. Type III PIP kinases generate PtdIns(3,5)P<sup>2</sup> form PtdIns3P. The only recognizable domain present in all PIPKs is the highly conserved kinase core domain (PIPKc) (Sasaki et al., 2009; Balla, 2013).

# 6.1. Type I PIP Kinase

# 6.1.1. General Description of Type I PIP Kinase (PIP 5-Kinase)

Three type I PIP kinases (PIPKI) have been identified in humans: PIPKIα, β, and γ. The PIPKIγ mRNA transcript has been shown to be alternatively spliced to encode multiple forms of PIPKIγ isoforms: PIPKIγ-i1 to 6 (Ishihara et al., 1998; Giudici et al., 2004; Schill and Anderson, 2009; Xia et al., 2011). All PIPKI isoforms share a central kinase core (PIPKc) domain with 80% amino acid homology (Ishihara et al., 1998). The C-terminal part of the PIPKc domain contains a 25 amino acid activation loop that is critical for both the substrate specificity and subcellular targeting of PIPKs (Kunz et al., 2000, 2002; Liu et al., 2016). PIPKIs are the major PtdIns(4,5)P2-generating enzymes, which phosphorylate the hydroxyl group at the D5 position of the inositol ring of PtdIns4P, and have a wide variety of roles relating to PtdIns(4,5)P<sup>2</sup> synthesis (Rameh et al., 1997). Since one of the major roles of PtdIns(4,5)P<sup>2</sup> is the actin-mediated processes, PIPKIs are also indispensable for the actin dynamics. In fact, yeast has a single PIPKI, Mss4p, and the mss4 mutant showed a phenotype similar to actin deficiency (Desrivières et al., 1998; Homma et al., 1998). In mammals, PIPKIs have also been shown to be involved in actin dynamics and membrane activities by generating PtdIns(4,5)P<sup>2</sup> from PtdIns4P, and the specific role of each PIPK seems to vary depending on their expression levels and the cell type (Balla, 2013). PIPKIs are widely distributed in the cell, and each isoform shows a unique localization pattern, whereby it regulates a specialized (compartmentalized) pool of PtdIns(4,5)P<sup>2</sup> (Doughman et al., 2003; Tan et al., 2015). PIPKIγi1–3 and 5 have been shown to be localized on the plasma membrane (Balla, 2013), while PIPKIγi2 is also localized in the recycling endosomes and focal adhesions (Di Paolo et al., 2002; Ling et al., 2002). It has been also shown that PIPKIα and β are targeted to autolysosomes (Rong et al., 2012). Additionally, it has been independently demonstrated that PIPKIα and PIPKIγi4 are found in nuclear speckles (Li et al., 2013), and PIPKIβ accumulates at the perinuclear regions (Doughman et al., 2003). PIPKIs apparently play redundant roles, and only a single copy of PIPKIγ is sufficient to support the development and growth of mice to the adulthood (Volpicelli-Daley et al., 2010). All three PIPKI isozymes have been linked to endosomal traffic (Galiano et al., 2002; Shinozaki-Narikawa et al., 2006). PIPKIα and β are known to initiate lysosomal reformation during autophagy (Yu et al., 2010; Rong et al., 2012; Chen and Yu, 2017). PIPKIα and γ are also implicated in chemotaxis (Lacalle et al., 2007; Lokuta et al., 2007). PIPKIγi1 is involved in the generation of pools of PtdIns(4,5)P<sup>2</sup> for Ins(1,4,5)P3, which regulates calcium release in histamine-stimulated HeLa cells (Wang et al., 2004). PIPKIα has been shown to be involved in pre-mRNA processing in association with non-canonical poly(A) polymerase, Star-PAP, which is specifically stimulated by PtdIns(4,5)P<sup>2</sup> (Mellman et al., 2008).

Activation of PIPKI differs from that of type II PIP kinase (PIPKII). PIPKI activity is stimulated by phosphatidic acid (PA) (Jenkins et al., 1994). Phospholipase D (PLD) and diacylglycerol kinase produce PA and are thought to be involved in the activation of PIPKIs (Tolias et al., 1998; Honda et al., 1999; Divecha et al., 2000). In mammals, two PLD isotypes use PtdIns(4,5)P<sup>2</sup> as a cofactor (Cockcroft, 2001), and thus, locally accumulated PIPKI and PLD mutually activate each other through their products, which results in a positive feedback loop (Mahankali et al., 2015). It has been shown that Arf6, a small GTPase which regulates membrane traffic, recruits PLD2 to membrane ruffles and stimulates PIPKI activity (Skippen et al., 2002). PLD1 is involved in initiation of autophagy by stimulating the PIPKI activity to generate necessary PtdIns(4,5)P<sup>2</sup> pool for the formation of the isolation membrane (Jenkins and Frohman, 2005; Dall'Armi et al., 2010; Fan et al., 2011; He et al., 2013).

6.1.2. Type I PIP Kinase (PIP 5-Kinase) of E. histolytica In E. histolytica, only one possible ortholog (EHI\_153770) with the E-value of < 1 × 10−<sup>10</sup> was found (E-value, 1 × 10−38) by using three H. sapiens type I PIP kinases (PIPKI) (NP\_001129110, AAH30587, and NP\_001287778) as queries. EHI\_153770 has a PIPKc domain based on pfam search. Furthermore, human and yeast PIPKIs, PIPKIβ (NP\_001265182) and Mss4p (BAA02869), respectively, were identified with the corresponding E-values of 8 × 10−<sup>39</sup> and 5 × 10−37, when EHI\_153770 was used as a query to search for a human or yeast ortholog, respectively. Since no PIPKII ortholog has been identified in E. histolytica (see below section 6.2.2), EHI\_153770 likely has a major role in PtdIns(4,5)P<sup>2</sup> generation from PtdIns4P. This single type I PIP kinase should have a wide variety of roles in E. histolytica. It is of note that the NLS is conserved in EHI\_153770 and the putative PLD of E. histolytica also has an NLS (K. Das, data not shown). Sharma and colleagues recentrly demonstrated of EHI\_15377 (Sharma et al., 2019).

# 6.2. Type II PIP Kinase

## 6.2.1. General Description of Type II PIP Kinase (PIP 4-Kinase)

Type II PIP kinase is the oldest PIP kinase identified among the others. However, the role of this class of enzymes is not as well-understood as that of type I PIP kinases (Boronenkov and Anderson, 1995; Divecha et al., 1995). Although PIPKII was initially thought to be responsible for generation of PtdIns(4,5)P<sup>2</sup> from PtdIns5P, this enzyme is currently considered to play a role in the regulation of the PtdIns5P levels (Clarke et al., 2010). Three PIPKII isotypes, PIPKIIα, β, and γ, are known, all of which contain the conserved PIPKc kinase domain bisected in the center by a non-conserved inserted sequence (Loijens et al., 1996; Itoh et al., 1998; Sasaki et al., 2009). All three PIPKIIs contain the ∼25 amino acid activation loop at the C-terminal part of the PIPKc domain as in PIPKIs (Sasaki et al., 2009). PIPKII α, β, and γ isotypes differ in their relative enzymatic activities in the order listed, with α being the most active (Clarke et al., 2008; Bultsma et al., 2010). It was speculated that the weak forms of PIPKII (β and γ) dimerize with the strong enzyme PIPKIIα and serve as adaptor proteins that bring it to specific membrane compartments (Clarke et al., 2010). Nuclear localization of PIPKIIα and PIPKIIβ has been reported (Bultsma et al., 2010; Wang et al., 2010). PIPKIIβ, which lacks an NLS, is targeted to the nucleus by a unique nuclear localization sequence consisting of an acidic α helix present in its unique insertion region in the kinase domain (Ciruela et al., 2000; Bunce et al., 2008). In contrast, PIPKIIγ has been detected in the ER by immunochemistry and subcellular fractionation, and also in other compartments of the endomembrane system (Itoh et al., 1998; Clarke et al., 2009).

Type II PIP kinase does not seem to play a role in the regulation of actin dynamics, as human type II PIP kinase failed to rescue yeast mss4 (type I PIP kinase) deficiency (Homma et al., 1998; Ishihara et al., 1998). It has been shown that PIPKIIα is involved in the formation and secretion of alpha granules in platelets (Rozenvayn and Flaumenhaft, 2001, 2003; Schulze et al., 2006). PIPKIIβ knockout mice show increased insulin sensitivity, likely through enhanced Akt activity (Carricaburu et al., 2003; Lamia et al., 2004). This is tentatively explained by slow degradation of PtdIns(3,4,5)P<sup>3</sup> in the knockout mice; however, the molecular basis remains elusive. The hypothesis that an excess amount of PtdIns5P inhibits phosphatase activity has been rejected (Campbell et al., 2003; Schaletzky et al., 2003). It has been shown that increasing the PtdIns5P level by overexpressing a bacterial PtdIns(4,5)P24-phosphatase (IpgD) enhanced Akt activity (Pendaries et al., 2006). It has also been shown that nuclear PIPKIIβ is involved in nuclear stress response. For instance, UV irradiation induces phosphorylation of Ser236 in PIPKIIβ, and this phosphorylation inactivates PIPKIIα kinaseactivity, which is associated with PIPKIIβ and accumulation of PtdIns5P (Jones et al., 2006; Bultsma et al., 2010; Wang et al., 2010) (also see section 2.2.4).

# 6.2.2. Type II PIP Kinase (PIP 4-Kinase) of E. histolytica

The E. histolytica genome was found to encode one possible PIPKI with a reasonable E-value of 6 × 10−<sup>26</sup> when three human PIPKII isotypes were used as queries in BLAST search. This protein (EHI\_153770) was identified as a top hit with a lower Evalue of 6 × 10−<sup>37</sup> when PIPKI was used as a query, as described above. Because of this higher similarity to PIPKIs, EHI\_153770 is categorized as a PIPKI.

# 6.3. Type III PIP Kinase

## 6.3.1. General Description of Type III PIP Kinase (PIP 5-Kinase)

Type III PIP kinases (PIPKIII) phosphorylate PtdIns3P to PtdIns(3,5)P2, which is one of the least abundant PIs (Hasegawa et al., 2017). PIPKIII was initially found in yeast by a genetic screening for defects in nuclear segregation (Yamamoto et al., 1995). A PIPKIII-deficient yeast line showed enlargement of vacuoles and retardation in vacuole delivery of hydrolases, such as carboxypeptidase Y (CPY) (Gary et al., 1998). This observation suggests the primary effect of PIPKIII deficiency is on membrane trafficking. PIPKIII is a large protein of >2000 amino acids and contains the C-terminal catalytic domain of PIPKc, which is similar to that of PIPKI and PIPKII. Distinct from other PIP kinases, PIPKIIIs have multiple domains in humans: FYVE (Fab1p, YOTB, Vac1p, and EEA1), DEP (disheveled, Egl-10, and pleckstrin), and TCP-1 (t-complex polypeptide-1) domains (Sasaki et al., 2009; Balla, 2013).These domains are involved in PtdIns3P binding, membrane association, and actin/tubulin binding, respectively (Cabezas et al., 2006). As PtdIns(3,5)P<sup>2</sup> has a critical role in endosome/lysosome biogenesis, PtdIns3P, which is highly used in endocytic pathways, is converted to PtdIns(3,5)P<sup>2</sup> by PIPKIII on endosomes to initiate MVB formation (Odorizzi et al., 1998). PIPKIII deficiency in mammalian cells has also been shown to cause massive vacuolization (enlarged endosomes) due to defective MVB formation (Ikonomov et al., 2003). PIPKIIIs are involved in a variety of membrane traffic pathways and signaling, such as retrograde transport of cation-independent mannose 6 phosphate receptor and sortilin, lysosomal localization and activity of mTORC1, and autophagosome-lysosome fusion (Rutherford et al., 2006; Zhang et al., 2007; Bridges et al., 2012; Hasegawa et al., 2016; Jin et al., 2016). It is of note that suppression of PIPKIII hampered phagosome maturation but did not inhibit acidification of lysosomes in macrophages (Kim et al., 2014). Additionally, PIPKIIIs are involved in nutrient (or macromolecule) import via vacuoles (Krishna et al., 2016). PIPKIII thus plays multiple roles by tightly regulating PtdIns3P and PtdIns(3,5)P<sup>2</sup> concentrations in vesicular trafficking. In mammals and yeast, PIPKIIIs are known to form a complex with the Sac1-related PI phosphatase, Fig4/Sac3, and a scaffold protein, Vac14/ArPIKfyve. In yeast, PIPKIII is also known to interact with the WD domain protein Atg18 and Vac7 (Duex et al., 2006; Chow et al., 2007; Sbrissa et al., 2007; Botelho et al., 2008; Jin et al., 2008). Deletion of Fig4/Sac3 and Vac14/ArPIKfyve reduces PtdIns(3,5)P<sup>2</sup> level (Gary et al., 1998, 2002; Bonangelino et al., 2002; Dove et al., 2002; Rudge et al., 2004; Duex et al., 2006; Zhang et al., 2007; Zolov et al., 2012). Thus, PIPKIII activation and stabilization is also regulated by the associated proteins, such as phosphatases and scaffold proteins in the complex, to tightly control the PtdIns(3,5)P<sup>2</sup> level.

# 6.3.2. Type III PIP Kinase of E. histolytica

A genome survey of E. histolytica by a BLAST search with H. sapiens PIPKIII as the query identified one ortholog candidate (EHI\_049480). EHI\_049480 shows similarities to human PIPKI and PIPKIII with the E-values of 5 × 10−<sup>4</sup> and 4 × 10−<sup>45</sup> , respectively. EHI\_049480 is considerably smaller than the potential human homolog, and has not been predicted to contain any additional domains such as FYVE. However, based on the highest similarity between the catalytic domains of EHI\_049480 and human PIPKIII, we tentatively annotated EHI\_049480 as a E. histolytica PIPKIII. Additionally, a BLASTP search against the H. sapiens and S. cerevisiae genomes with EHI\_049480 as the query identified PIPKIII and Fab1 with the E-values of 2 × 10−<sup>45</sup> and 9 × 10−40, respectively.

We further searched for other PIPKIII complex components such as Vac7, Vac14, Atg18, and Fig4 (the yeast homolog of the mammalian Sac3). Potential orthologs for Atg18 and Fig4 were identified with the E-values of 1 × 10−<sup>20</sup> and 1 × 10−<sup>34</sup> , respectively. This possible Fig4 ortholog showed a lower E-value with Sac1 (1 × 10−82). Thus, it is reasonable to tentatively assign this protein as a Sac1 ortholog, although it is not possible to specifically categorize Sac1 among Sac orthologs (see section 10). A human ortholog of Atg18 (WIPI), which is a WD repeat-containing protein that interacts with phosphoinositides, recognizes PtdIns3P on the nascent autophagosome and recruits the lipidation machinery to the autophagosome for LC3. However, the role of WIPI in PtdIns(3,5)P<sup>2</sup> metabolism remains unknown. It is unknown whether E. histolytica has PtdIns(3,5)P<sup>2</sup> metabolic pathways similar to other organisms. However, conservation of ESCRT and MVB, and the localization of PtdIns3P on phagosomes indicate that PtdIn(3,5)P2-mediated vesicular trafficking is also conserved in E. histolytica. We failed to identify potential Vac7 and Vac14 homologs with an E-value < 1 × 10−<sup>1</sup> . As 19 HEAT repeat-containing proteins are present in the E. histolytica genome, it is possible that some of them function in lieu of Vac14.

# 7. PI 3-PHOSPHATASES

PI phosphatases are also important regulators of PI signaling. Because identification of lipid kinases and PLC-mediated second messengers had a significant impact, studies on phosphatase activity in the early '80s was largely focused on Ins(1,4,5)P<sup>3</sup> decomposition. A number of inositol phosphatases and PI phosphatases were identified and characterized (Majerus et al., 1986). In the '90s, it was revealed that mutations in PI phosphatases are responsible for human genetic disorders, including Oculo-Cerebro-Renal Syndrome of Lowe (OCRL), human X-linked centromyotubular myopathy, and Charcot-Marie-Tooth disease type 4B. OCRL1, myotubularin-related (MTMR) 1, and MTMR2 were identified as the genes responsible for the above-mentioned human genetic disorders, respectively (Attree et al., 1992; Myers et al., 1997; Maehama and Dixon, 1998, 1999; Blondeau et al., 2000; Taylor et al., 2000; Kim S. A. et al., 2002). One of the most studied PI phosphatases, PTEN, was identified as a tumor suppressor gene (Maehama and Dixon, 1998). Similar to PI kinases, classification of PI phosphatases is primarily based on the position of the hydroxyl group that they dephosphorylate. The human and yeast genomes are known to encode twenty-eight and six PI phosphatases, respectively (Odorizzi et al., 2000; Sasaki et al., 2009).

PI 3-phosphatases are categorized into two groups based on substrate specificities. One group includes PTEN, TPTE (transmembrane phosphatase with tensin homology), and TPIP (TPTE and PTEN homologs inositol lipid phosphatase), while the other group includes myotubularins (MTMs). Since TPTEs have no phosphatase activity, and their functional role remains unclear, they are not discussed in this review. However, TPTE has been reported to be associated with cancers. For instance, it has been demonstrated that TPTE is upregulated in prostate cancer, and autoantibody production against TPTE is observed in lung cancer (Walker et al., 2001; Tapparel et al., 2003; Bansal et al., 2015; Kuemmel et al., 2015). PTEN and TPIP have similar catalytic domains but differ in substrate specificity. PTEN removes the phosphate moiety at D3 position of PtdIns(3,4,5)P<sup>3</sup> and PtdIns(3,4)P2, while TPIPs dephosphorylate any of the 3′ phosphorylated inositides at D3 position (Walker et al., 2001; Malek et al., 2017). On the other hand, MTMs remove the D3 phosphate from PtdIns(3,5)P<sup>2</sup> and PtdIns3P. The catalytic center of both groups of PI 3-phosphatases contains the CX5R motif, which is also found in protein tyrosine phosphatases (Hsu and Mao, 2015).

# 7.1. PTEN and TPIP

# 7.1.1. General Descriptions of PTEN and TPIP

PTEN was initially identified as a tumor suppressor gene located on chromosome 10 (Li et al., 1997; Steck et al., 1997; Maehama and Dixon, 1998). It is among the most frequently mutated genes in various cancers in humans (Guldberg et al., 1997; Li et al., 1997; Steck et al., 1997; Tashiro et al., 1997; Cairns et al., 1998; Kohno et al., 1998; Maehama, 2007), and in hereditary cancer predisposition syndromes, such as Cowden disease (Myers et al., 1997; Furnari et al., 1998). Human TPIP was bioinformatically identified as a protein encoded by PTEN-related genes (Walker et al., 2001). PTEN consists of a protein tyrosine phosphatase (PTP)-related lipid phosphatase domain, a C2 domain, two PEST (proline, glutamine, serine, threonine) sequences, and a PDZ domain (**Figure 4**). The C2 domain is known to be involved in lipid binding and protein stability. PEST sequences are known to enhance proteolytic sensitivity, and the PDZ domain is involved in protein-protein interactions (Maehama, 2007; Sasaki et al., 2009; Balla, 2013). The human genome encodes three isotypes of TPIP, and only two of them have an N-terminal transmembrane domain (**Figure 4**). PTEN and TPIP have a CX5R motifcontaining PTP-related lipid phosphatase domain, whose core sequence is CKAGKGR and CKGGKGR, respectively. This domain forms the catalytic cleft of PTEN, and it is wider than the corresponding domain of protein tyrosine phosphatases. This allows the bulky PtdIns(3,4,5)P<sup>3</sup> head group to access the active center of the enzyme (Lee et al., 1999). The principal role of PTEN is to cease the cell proliferation signal by inactivating Akt though dephosphorylation of PtdIns(3,4,5)P3, whereby it serves as a tumor suppressor.

PTEN is predominantly localized to the cytosol, and also dynamically associated with the plasma membrane, where it hydrolyzes PtdIns(3,4,5)P<sup>3</sup> (Billcliff and Lowe, 2014). Although PTEN lacks an NLS, it also localizes to the nucleus and plays important roles in chromosome stability by directly interacting with centromere specific binding protein C (CENP-C), DNA repair by interacting with p53 and Rad51, and cell cycle regulation by interacting with APC and MAP kinases (Freeman et al., 2003; Chung and Eng, 2005; Chung et al., 2006; Tang and Eng, 2006a,b; Shen et al., 2007; Trotman et al., 2007; Song et al., 2011). This nuclear translocation depends on the cytoplasmic localization signal and ubiquitination (Denning et al., 2007; Trotman et al., 2007; Wang et al., 2007; Drinjakovic et al., 2010). Mutations in the N-terminal cytoplasmic localization signal, which is found in familial Cowden disease patients, increases nuclear localization (Denning et al., 2007). Mono-ubiquitination of K13 and K298 serves as the nuclear translocation signal, while poly-ubiquitination causes degradation of PTEN (Trotman et al., 2007; Wang et al., 2007; Drinjakovic et al., 2010). PTEN has also been shown to participate in phagocytosis, autophagy and determination of cell polarity through dephosphorylation of PtdIns(3,4,5)P<sup>3</sup> (Arico et al., 2001; Kim J.S et al., 2002; Martin-Belmonte et al., 2007). In contrast, the physiological role of TPIP is not well-understood. TPIPα and γ both have the transmembrane domain unlike TPIPβ. Accordingly, they are localized on internal membranes, whereas TPIPβ remains in the cytosol (Walker et al., 2001).

# 7.1.2. PTEN and TPIP of E. histolytica

The E. histolytica genome potentially encodes six PTEN orthologs with E-values < 1 × 10−<sup>10</sup> to the human PTEN (**Figure 4**; **Supplementary Table S2**; **Supplementary Figure S5**). Domain search by pfam showed that these six PTEN orthologs contain the CX5R motif-containing phosphatase domain. Furthermore, the same protein candidates were detected when the three human TPIP isoforms were used as queries. The E-values against TPIP were around 1 × 10−16-10−17, which is higher than those against PTEN (**Supplementary Table S2**). Thus, we concluded that they are likely PTEN orthologs. Furthermore, five out of the six orthologs have a CX5R motif identical to that found in PTEN (CKAGKGR). The CX5R sequence of the remaining ortholog (EHI\_131070) is CLAGRGR. Three of the E. histolytica PTEN orthologs also had a C2 domain. PTENs commonly have the cytosol localization signal (**Supplementary Figure S2**) (Denning et al., 2007). Among the three C2 domain-containing candidates, all six consensus amino acids of the cytosol localization signal are conserved in EHI\_131070, and all but one amino acid are also conserved in EHI\_197010 and EHI\_098450 (**Supplementary Figure S2**). In the three E. histolytica PTEN candidates, which lack the C2 domain, only a few residues are conserved (four in EHI\_010360, EHI\_054460; three in EHI\_041900) (**Supplementary Figure S2**). All the three candidates have lysine or tyrosine instead of phenylalanine that is found in the amino acid 22 position of human PTEN. It has been shown that F21A mutation causes nuclear localization of PTEN, which in turn fails to activate Akt even though the phosphatase activity is retained (Denning et al., 2007). Thus, it is not clear if the three E. histolytica PTEN candidates which lack the C2 domain have functional cytosol localization signals. According to the gene expression profile, two of the PTEN homologs, EHI\_197010 and EHI\_098450, are actively transcribed (**Supplementary Figure S1**). Both of them have a C2 domain and the cytosol localization signal.

# 7.2. Myotubularin (MTM)

# 7.2.1. General Description of MTM

The human myotubularin (MTM) family consists of 15 members [MTM1 and MTM related (MTMR) 1–14]. As in PTEN, the CX5R motif of their phosphatase domain is overall well-conserved. Among the 15 MTMs, six of them (MTMR5 and MTMR9–13) substituted the conserved cysteine and arginine residues within the CX5R motif with other amino acids, and thus these MTMs are catalytically inactive (Laporte et al., 2003; Hnia et al., 2012; Hsu and Mao, 2015). The CX5R motif in MTMs (CXXGWDR) is slightly different from that in PTEN (CKAGKGR). While PTEN prefers PtdIns(3,4,5)P<sup>3</sup> and PtdIns(3,4)P<sup>2</sup> as substrates, MTMs preferentially dephosphorylate PtdIns3P and PtdIns(3,5)P2. MTMs are categorized into six groups based on their domain configurations and catalytic activity: MTM1 and MTMR1– 2; MTMR3–4; MTMR6–8; MTMR14; MTMR5 and 13; and MTMR9–12 (**Figure 4**). In all MTMs, except for MTMR14, the PH-GRAM (pleckstrin homology-glucosyltransferase, Rab-like GTPase activator and myotubularin) domain is conserved. This domain is involved in PI binding. Additionally, they all have an active or inactive catalytic core, and coiled-coil domain, which is involved in homo- or hetero- dimerization. In addition, some members have the C-terminal PDZ binding sequence and FYVE domain. The N-terminal DENN and C-terminal PH domains are conserved in two of the catalytically inactive MTM members, MTMR5 and MTMR13. Although all the catalytically inactive members (MTMR5 and MTMR9–13) lack an active phosphatase domain, they can heterodimerize with active MTMs, whereby inactive MTMs likely regulate the activity and localization of active MTMs (Kim et al., 2003; Mochizuki and Majerus, 2003; Lorenzo et al., 2006). The role of MTMs and their preferred substrates [PtdIns3P and PtdIns(3,5)P2] in endocytosis and membrane traffic have been well-characterized (Robinson and Dixon, 2006; Hohendahl et al., 2016). Besides these, MTMs have been suggested to have other roles in cellular processes, including cell proliferation and differentiation, autophagy, phagocytosis, organelle positioning, cytokinesis, cytoskeletal rearrangement, and cell junction dynamics (Hnia et al., 2012; Lawlor et al.,


Vac1, and EEA1 domain; GRAM, glucosyltransferases Rab-like GTPase activators and myotubularins; GAP, GTPase-activating domain; kinase, protein kinase domain; LRR, leucine-rich repeats; NPF, asparagine- proline-phenylalanine repeats; P, Proline-rich; PDZ-B, PDZ domain binding domain; PEST, Proline, glutamine, serine, threonine; PH, Pleckstrin-homology; (Ptase; gray), inactive Ptase domain; (Ptase; yellow), active CX5R motif containing PI 3- or PI 4-phosphatase domain; ROCO, comprised of a ROC (Ras of complex proteins) and COR (C-terminal of ROC) region; SAC, Sac domain; SAM, sterile alpha motif; SH2, Src homology 2; SKICH, SKIP carboxylhomology; TM, transmembrane domain. Clathrin-binding domain and the nuclear localization signal are also labeled with "" and "H", respectively.

2016). It is of note that some of the cellular functions performed by MTMs depend on the tissue-specific expression patterns of their binding proteins and do not involve a phosphatase activity. For instance, disrupting the interaction between MTM1 and its intermediate filament, desmin, causes the formation of desmin aggregates, and this impairment is associated with myofibrillar myopathies and cardiomyopathies (Hnia et al., 2011).

# 7.2.2. MTM of E. histolytica

Genome-wide survey against the E. histolytica genome using human MTM1, MTMR3, 5, 6, 9, or 14 as the query identified an identical set of 11 proteins with E-values < 1 × 10−<sup>58</sup> . We concluded that 10 out of these 11 potential homologs are E. histolytica MTM orthologs by pfam search, because they contain a myotubularin-like phosphatase domain (**Figure 4**; **Supplementary Tables S1**, **S4**). All of them, except EHI\_140980 and EHI\_188050, contain the conserved catalytic domain. The phosphatase domains of these two exceptions conserved the cysteine residue in their C(S/T)DGWDR motifs, but arginine is replaced with serine or isoleucine (CRNGWDS and CIDGTGI, respectively). Although E. histolytica MTMs appear to have a simpler domain organization, they also are thought to heterodimerize as in model organisms given that the genome contains both active and inactive MTMs. Among all the PI phosphatases in E. histolytica, MTMs are the most diverged ones.

It should be noted that Amoebozoa supergroup members exclusively have a protein family that contain multiple inactive myotubularin domains. These proteins have been designated as inactive myotubularin/LRR/ROCO/kinase (IMLRK) proteins (Kerk and Moorhead, 2010). Nine IMLRK proteins have previously been identified in the E. histolytica genome by the FFAS03 (Fold and Function Assignment System) sequence: profile method and the HHPred [Hidden Markov Model (HMM)-HMM structure prediction] profile: profile method (Kerk and Moorhead, 2010). The D. discoideum homologs of IMLRKs, Pats1 and GbpC, have been identified as a cytokinesisrelated protein and cGMP-binding protein, respectively (Goldberg et al., 2002; Abysalh et al., 2003). Since both D. discoideum IMLRK homologs are involved in cytoskeletonrelated processes such as cytokinesis and chemotaxis, E. histolytica IMLRKs may also be involved in the similar processes (Bosgraaf et al., 2002, 2005; Abysalh et al., 2003; Lewis, 2009). However, it is not clear why E. histolytica has inactive myotubularin domains, and more IMLRK proteins than D. discoideum (nine vs. two, respectively).

# 8. PI 4-PHOSPHATASES

Dissimilar to PI 3- and PI 5-phosphatases, which belong to multiple families of enzymes, PI 4-phosphatases consist of only four proteins. All members of human PI 4-phosphatases have the CX5R motif in the catalytic domain. There are two groups of PI 4-phosphatases: inositol polyphosphate-4-phosphatase (INPP4) and transmembrane protein 55 (TMEM55) (**Figure 4**). INPP4 dephosphorylates PtdIns(3,4)P2, and TMEM55 dephosphorylates PtdIns(4,5)P2. No PI 4 phosphatases that use PtdIns(3,4,5)P<sup>3</sup> as a substrate have been identified yet. E. histolytica has no homologs of PI 4-phosphatases in this class. Here, we will discuss only the general aspects of PI 4-phosphatases.

# 8.1. INPP4

Two INPP4 proteins have been identified in the human genome and named INPP4A and B. They both contain a conserved catalytic domain and N-terminal C2 domain; however only INPP4A contains the PEST sequence (Sasaki et al., 2009; Balla, 2013) (**Figure 4**). The C2 domains of these proteins show different binding specificities. The C2 domain of INPP4A preferentially binds to PtdIns(3,4)P2, PtdIns3P, phosphatidylserine, and calcium (Ivetac et al., 2005, 2009; Shearn and Norris, 2007), while that of INPP4B prefers phosphatidic acid and PtdIns(3,4,5)P<sup>3</sup> (Ferron and Vacher, 2006). INPP4A is cleaved and inactivated by calpain via recognition of the PEST sequence in INPP4A (Norris et al., 1995), as shown in platelets stimulated with thrombin and calcium ionophores (Norris et al., 1997). INPP4A is known to be localized in recycling and early endosomes in resting cells and translocated to the plasma membrane upon serum stimulation (Ivetac et al., 2005). In contrast, INPP4B has been shown to be localized diffusely in the cytoplasm. INPP4A is involved in the regulation of membrane traffic, which is consistent with its endosomal localization. It has been shown that INPP4A is activated by Rab5, and knocking down Rab5 inhibits transferrin uptake (Shin et al., 2005). On the other hand, overexpression of INPP4A suppresses the enlarged endosome morphology caused by PtdIns3P deficiency in INPP4A knock-out MEF cells (Ivetac et al., 2009), suggesting that INPP4A produces PtdIns3P from PtdIns(3,4)P2. A significant role of INPP4A in the generation of PtdIns3P to recruit SNX9 and actin polymerization machineries during clathrin-mediated endocytosis has also been reported (Malek et al., 2017). No potential orthologous genes with the E-value < 1 × 10−<sup>1</sup> have been identified in E. histolytica.

# 8.2. TMEM55

TMEM55 (transmembrane protein 55), which was named after the transmembrane regions it contains, was originally identified based on its homology to the virulence factor of Burkholderia pseudomallei, BopB, a putative phosphatase that contains a CX5R motif (Ungewickell et al., 2005). There are two isotypes of TMEM55, termed TMEM55A and B. They both consist of a CX5R motif-containing phosphatase domain and two putative transmembrane domains at the C-terminus (Rynkiewicz et al., 2012). Based on in vitro and in vivo observations, TMEM55 specifically hydrolyzes the D4 phosphate of PtdIns(4,5)P2. Both TMEM55A and TMEM55B show cytosolic and late endosomal membrane localization (Ungewickell et al., 2005). Overexpression of TMEM55A enhances EGFR degradation induced by EGF stimulation, suggesting that TMEM55A is involved in the endocytic and recycling pathways (Ungewickell et al., 2005). Additionally, TMEM55A has been reported to be involved in macrophage phagocytosis (Morioka et al., 2018). On the other hand, TMEM55B translocates from the cytosol to the nucleus and increases the PtdIns5P level in response to DNA damage (Zou et al., 2007). This upregulation of PtdIns5P in the nucleus is accompanied by activation of ING2 (Gozani et al., 2003) and enhanced p53-mediated cell death (Zou et al., 2007). In E. histolytica, no potential orthologous genes with the E-value < 1 × 10−<sup>1</sup> have been identified. It is possible that E. histolytica utilizes Sac phosphatase (see below) for dephosphorylation of the D4 position of PIs.

# 9. PI 5-PHOSPHATASES

Inositol polyphosphate 5-phosphatases (INPP5s) have an inositol 5-phosphatase (5-Ptase) domain that contains two signature motifs (F/Y)WXGDXN(F/Y)R and P(A/S)(W/Y)(C/T)DR(I/V)L(W/Y) separated by ∼60–75 residues (Majerus et al., 1999). INPP5s are Mg2+-dependent enzymes and share homology with apurinic/apyrimidinic family of endonucleases (Whisstock et al., 2000). There are four classes of PI 5-phosphatases (types I–IV). However, the type I enzyme (INPP5A) does not have lipid phosphatase activity, and dephosphorylate Ins(1,4,5)P<sup>3</sup> and Ins(1,3,4,5)P<sup>4</sup> (Laxminarayan et al., 1993, 1994; De Smedt et al., 1994). Type II PI 5-phosphatases include the synaptojanins OCRL1, INPP5B, INPP5J, and SKIP. The type III enzymes are two SHIPs, namely SHIP1 and 2. Interestingly, there is only one type IV PI 5 phosphatase, named INPP5E. Because E. histolytica has a low conservation of PI 5-phosphatases, we will describe the search result at the end of this section. However, the enzyme Sac can act as a PI 5-phosphatase. Although it does not contain the 5-Ptase domain, it contains the CX5R motif-containing phosphatase domain (CKAGRSR). As described below, like PTEN, it differs from PI 5-phosphatases in its active center configuration and substrate specificity.

# 9.1. Type II PI 5-Phosphatase

There are five kinds of type II PI 5-phosphatases, such as synaptojanins, OCRL1, INPP5B, INPP5J, and SKIP. They differ in structure and function.

# 9.1.1. General Description of Synaptojanins

There are two synaptojanins in mammals, and each of them has multiple splice variants (McPherson et al., 1996). All the synaptojanins have a conserved 5-Ptase catalytic domain and N-terminal Sac domain. The Sac domain also has inositol phosphatase activity. The domain configuration of the Cterminal region varies in each splice form (**Figure 4**). Most synaptojanin splice forms encode proteins that contain a proline-rich (P) region, and SYNJ1-170 additionally has an asparagine-proline-phenylalanine (NPF) repeat. NPF repeat is involved in the association of these proteins with the endocytic protein Eps15 (Haffner et al., 1997). SYNJ2A is ubiquitously expressed, and SYNJ1-170 shows a broad tissue distribution. Interestingly, SYNJ1-145, SYNEJ2B1, and SYNEJ2B2 are highly expressed in the brain, and SYNEJ2B1 and 2 are also abundant in the testis (Nemoto et al., 2001). The 5-Ptase domain of all SYNJ1 and 2 proteins enables them to dephosphorylate Ins(1,4,5)P3, Ins(1,3,4,5)P4, PtdIns(4,5)P2, and PtdIns(3,4,5)P<sup>3</sup> at the D5 position (McPherson et al., 1996; Sakisaka et al., 1997). Furthermore, the Sac domain of Synaptojanins enable them to dephosphorylate PtdIns3P, PtdIns4P and PtdIns(3,5)P<sup>2</sup> (Guo et al., 1999). Sac domain has been suggested to use the product generated by the PI 5-phosphatase domain to finally generate PtdIns (Guo et al., 1999; Nemoto et al., 2001). SYNJ1 has been found to be involved in synaptic vesicle exocytosis and recycling (McPherson et al., 1996). Additionally, synaptojanins have a critical role in the fate determination of clathrin-coated vesicles (CCVs). Knocking out SYNJ1 in the mouse caused accumulation of endocytosed CCVs and poor recycling of vesicles into the fusion-competent synaptic vesicle pool (Kim W. T. et al., 2002). Consequently, it caused neurological defects, and death after birth (Cremona et al., 1999). Synaptojanins are involved in endocytosis and synaptic vesicle recycling in concert with a variety of binding proteins such as endocytic proteins amphiphysin, endophilin syndapin/pacsin, and intersectin/Dap160 (McPherson et al., 1996; Bauerfeind et al., 1997; de Heuvel et al., 1997; Ringstad et al., 1997; Roos and Kelly, 1998; Qualmann et al., 1999). Also, the Cterminal NPF region of SYNJ1-170 enables the interaction with the EH (Eps15 homology) domain of Eps15, ear domain of the α-adaptin component of the adaptor protein (AP) 2 complex, and N-terminal domain of the clathrin heavy chain (Barbieri et al., 2001; Krauss et al., 2006).

# 9.1.2. General Descriptions of OCRL1 and INPP5B

OCRL1 was originally identified as a protein responsible for the X-linked human disease OCRL. It shares a high amino acid sequence homology (45%) with INPP5B (Attree et al., 1992; Jefferson and Majerus, 1995; Speed et al., 1995; Matzaris et al., 1998). OCRL1 and INPP5B consist of a PH domain, 5-Ptase domain, ASH (ASPM, SPD2, hydin) domain, and RhoGAP domain. Only OCRL1 has two clathrin binding domains (CB), and only INPP5B has a C-terminal CAAX extension for prenylation of the cysteine residue (Jefferson and Majerus, 1995). These two proteins are the only RhoGAP domain-containing PI 5-phosphatases in the human and mouse (Jefferson and Majerus, 1995; Speed et al., 1995; Matzaris et al., 1998; Lowe, 2005). However, their RhoGAP domains appear to lack the catalytic activity because of amino acid substitutions at the catalytic region (Peck et al., 2002). Even though these domains lack a catalytic activity, RhoGAP domain of OCRL1 can interact with Rac1, Cdc42, ARF1, and ARF6 (Faucherre et al., 2003; Lichter-Konecki et al., 2006; Erdmann et al., 2007; Choudhury et al., 2009). OCRL1 prefer PtdIns(4,5)P2, Ins(1,4,5)P3, Ins(1,3,4,5)P4, and PtdIns(3,4,5)P<sup>3</sup> as dephosphorylation substrates (Zhang et al., 1995; Schmid et al., 2004). OCRL1 can also dephosphorylate PtdIns(3,5)P<sup>2</sup> (Schmid et al., 2004). On the other hand, INPP5B can remove the D5 phosphate of PtdIns(4,5)P2, PtdIns(3,4,5)P3, Ins(1,4,5)P3, and Ins(1,3,4,5)P<sup>4</sup> with comparable efficiencies, but it is ineffective in the dephosphorylation of PtdIns(3,5)P<sup>2</sup> (Jefferson and Majerus, 1995; Schmid et al., 2004). OCRL1 is involved in the regulation of vesicular trafficking between endosomes and the TGN (Choudhury et al., 2005). This is in line with OCRL1's localization on the TGN, early endosomes, membrane ruffles, and clathrin-coated trafficking intermediates (Olivos-Glander et al., 1995; Dressman et al., 2000; Ungewickell et al., 2004; Choudhury et al., 2005; Faucherre et al., 2005; Erdmann et al., 2007). OCRL1 deficiency causes abnormal cytoskeletal reorganization (Suchy and Nussbaum, 2002; Faucherre et al., 2005). This is not only due to perturbation of the functions exerted by the RhoGAP-like domain, but also due to perturbed modulation of PtdIns(4,5)P<sup>2</sup> levels, suggesting the importance of OCRL1 in PtdIns(4,5)P<sup>2</sup> homeostasis. INPP5B is associated with the Golgi apparatus, ER exit sites, and Rab (Williams et al., 2007). Its localization is affected by interactions with Rab proteins (Vyas et al., 2000; Shin et al., 2005; Williams et al., 2007). These data suggest that INPP5B is involved in the early secretory pathway.

# 9.1.3. General Descriptions of INPP5J and SKIP

INPP5J and SKIP (skeletal muscle- and kidney-enriched inositol polyphosphate phosphatase) have been identified by virtue of two conserved consensus motifs characteristic of PI 5-phosphatases (Mochizuki and Takenawa, 1999; Gurung et al., 2003). INPP5J is also called proline-rich inositol polyphosphate 5-phosphatase (PIPP) because of the two proline-rich regions at the Nand C- terminal of the protein. The N-terminal prolinerich region of INPP5J contains a putative SH3-binding motif (PRSPSR) and six 14–3–3 binding motifs (Gurung et al., 2003), and a SKICH (SKIP carboxyl homology) domain which mediates localization to the plasma membrane (Ooms et al., 2006). The C-terminus of the 5-Ptase domain is overall conserved in both INPP5J and SKIP. INPP5J removes the D5 phosphate of PtdIns(4,5)P2, PtdIns(3,4,5)P3, Ins(1,4,5)P3, and Ins(1,3,4,5)P4, while SKIP prefers PtdIns(3,4,5)P3, but can also dephosphorylate PtdIns(4,5)P2, Ins(1,4,5)P3, and Ins(1,3,4,5)P<sup>4</sup> (Ijuin and Takenawa, 2003). INPP5J and SKIP are localized on the plasma membrane, and also in the TGN and ER of resting cells, respectively. Following insulin stimulation, SKIP translocates to the plasma membrane via the SKICH domain (Gurung et al., 2003). Both enzymes are considered to regulate Akt activation by negatively regulating PtdIns(3,4,5)P<sup>3</sup> (Ijuin and Takenawa, 2003; Ooms et al., 2006). Furthermore, SKIP participates in insulin response by regulating GULT4 translocation and glucose uptake in skeletal muscle.

# 9.2. Type III PI 5-Phosphatase

Type III PI 5-phosphatases, termed SHIP (SH2-containing inositol phosphatase), consists of SHIP1 and SHIP2. SHIP1 has several splice variants: a full-length isoform (SHIP1α) and shorter isoforms (SHIP1β, SHIP1γ, and s-SHIP1) (Lucas and Rohrschneider, 1999; Tu et al., 2001). SHIP1 was originally identified as a binding protein of several adaptor proteins such as Shc, Grab2, and DOK (Liu et al., 1994), and immunoreceptor tyrosine-based inhibitory (ITIM) or activation (ITAM) motifs of immune receptors (Osborne et al., 1996; Kimura et al., 1997). Both enzymes are composed of an N-terminal SH2 domain, a 5- Ptase catalytic domain followed by a C2 domain, and an NPXY motif. All SHIP isoforms except for SHIP1γ contain a C-terminal proline-rich domain. SHIP2 has a sterile alpha (SAM) motif at the C-terminus of the proline rich region. SHIP1 is exclusively expressed in hematopoietic and spermatogenic lineages, whereas SHIP2 expression is ubiquitous (Liu et al., 1998). SHIP1 isoforms hydrolyze the D5 phosphate of both PtdIns(3,4,5)P<sup>3</sup> and Ins(1,3,4,5)P4, whereas SHIP2 activity appears to be more specific for PtdIns(3,4,5)P<sup>3</sup> (Wisniewski et al., 1999). SHIP1 is involved in myeloid cell homeostasis, chemotaxis, bone metabolism, and mast cell activation (Helgason et al., 1998; Huber et al., 1998; Liu et al., 1999; Jiang et al., 2005; Nishio et al., 2007). SHIP2 is involved in the negative regulation of insulin-mediated cellular response (Wada et al., 2001; Kaisaki et al., 2004; Sleeman et al., 2005). These phenotypes are explicable given the high level of PtdIns(3,4,5)P<sup>3</sup> in the cells, also pointing to the pivotal role of SHIPs in PtdIns(3,4,5)P3-mediated signaling.

# 9.3. Type IV PI 5-Phosphatase

There is only one type IV PI 5-phosphatase, which is also called INPP5E, 72-kDa polyphosphate 5-phosphatase, or Pharbin (Kisseleva et al., 2000; Kong et al., 2000). INPP5E consists of the C-terminal farnesylation CAAX motif and the central 5-phosphatase domain, flanked by an N-terminal prolinerich region. PtdIns(4,5)P2, PtdIns(3,4,5)P3, and PtdIns(3,5)P<sup>2</sup> have been reported to be substrates of INPP5E (Kisseleva et al., 2000; Kong et al., 2000). Notably, INPP5E has the highest affinity to PtdIns(3,4,5)P<sup>3</sup> and shows ten times higher affinity for PtdIns(3,4,5)P<sup>3</sup> than SHIP (Kisseleva et al., 2000). It localizes to the cytosol, and Golgi in proliferating cells (Kong et al., 2000). In macrophages, INPP5E localizes to the phagocytic cup and regulates FcγR1-mediated, but not complement receptor 3-mediated, phagocytosis (Horan et al., 2007). In adipocytes, INPP5E has been shown to hydrolyze PtdIns(3,5)P<sup>2</sup> to PtdIns3P and enhance GULT4 translocation; however, insulin signaling does not decrease upon INPP5E overexpression (Kong et al., 2000). Importantly, INPP5E also plays a critical role in ciliopathies (Bielas et al., 2009; Jacoby et al., 2009). The primary cilium is a microtubule-based organelle, which forms protrusions on the plasma membrane and functions as a sensor for the environmental factors such as light, chemicals, and mechanical stress (Goetz and Anderson, 2010). Additionally, primary cilia are central in hedgehog signaling, which is a major pathway involved in the structural organization of the body, organogenesis, and tumorigenesis (Elliott and Brugmann, 2019). PtdIns4P and PtdIns(4,5)P<sup>2</sup> have been shown to be localized on the primary cilium membrane and ciliary base, respectively. This position-specific PI distribution pattern is maintained by INPP5E localized at the primary cilium (Chávez et al., 2015; Garcia-Gonzalo et al., 2015). Disrupting the PI compartmentalization pattern or increasing the PtdIns(4,5)P<sup>2</sup> level in the primary cilium causes cilial accumulation of Tulp3, which contains a PtdIns(4,5)P2-binding domain. In turn, Tulp3's binding proteins, Gpr161 and IFT-A, are recruited to the primary cilium. Gpr161 is a negative regulator of the hedgehog signaling, and IFT-A is a flagellar transporting protein, respectively. Accumulation of this complex has been shown to perturb hedgehog signaling from the primary cilium. This observation emphasizes the significance of PI metabolism in cell physiology (Nakatsu, 2015).

# 9.4. PI 5-Phosphatases of *E. histolytica*

The possible E. histolytica PI 5-phosphatases all have a simple domain configuration: a 5-phosphatase domain alone or together with one more domain (**Figure 4**). PI 5-phosphatases have divergent and specialized roles in the human. For instance, synaptojanins are involved in neurotransmitter secretion, while SHIPs participate in hematopoietic cell signaling. Thus, the simplification of 5-phosphatases in E. histolytica may indicate the importance of IP5-phosphatase in multicellularity, and also the possibility of E. histolytica having only the ancestral type of PI 5-phosphatases.

Six candidate proteins showed E-values lower than 1 × 10−<sup>30</sup> when searched with human OCRL1 as the query (AAB03839). One candidate, EHI\_153490, has the 5-Ptase and RhoGAP domains, and the other protein, EHI\_159880, contains the 5-Ptase and GAP domains. They both lack the C-terminus CAAX motif conserved in INPP5B. The apparent similarity in domain configurations suggests that these two proteins may have homologous roles to those of OCRL1. The rest of the hits (four) only have the 5-Ptase domain (**Supplementary Figure S3**) and showed only low levels of E-values to OCRL1 (**Supplementary Table S5**). Searching with type III and type IV PI 5-phosphatases detected the six candidates with higher E-values. Based on this result, we tentatively assigned the four proteins as type II PI 5-phosphatases homologous to OCRL1 (**Figure 4**; **Supplementary Table S2**). When searched with synaptojanin1 and synaptojanin2, nine candidates with E-values lower than 1 × 10−<sup>25</sup> were detected. Six of them are identical to the above mentioned OCRL1 homologs and showed lower E-values to OCRL1. Other three candidates have a SAC domain, and thus they are considered as SAC homologs (see section 10). EHI\_040380, as one of these SAC domain-containing proteins, was found to also harbor a DNase I domain by InterPro analysis (https://www.ebi.ac.uk/interpro/). The domain is classified in the endonuclease/exonuclease/phosphatase superfamily (IPR036691). As mentioned above, the 5-Ptase domain shares homology to the apurinic/apyrimidinic family of endonucleases (Whisstock et al., 2000), but EHI\_040380 lacks the residues critical for the 5-Ptase activity (**Supplementary Figure S3**). Consequently, we classified this domain as "SAC with DNase I or 5-Ptase" domain (**Figure 4**). If this protein has a IP 5-phosphatase activity, it may act as synaptojanin in E. histolytica.

# 10. Sac FAMILY PHOSPHATASES

# 10.1. General Description of Sac

The first member of Sac (suppressor of actin), Sac1, was originally discovered in yeast by two independent genetic suppressor screens, which searched for the genes that rescued either actin cytoskeleton defects (Novick et al., 1989) or secretion defects caused by sec14 mutation (Cleves et al., 1989). The catalytic domain of the Sac family phosphatases conserves the CX5R motif, which is commonly found in protein and PI phosphatases. According to the crystal structure, configuration of Sac1 catalytic center has a unique feature compared with PTEN and MTMs (Lee et al., 1999; Begley et al., 2003, 2006; Manford et al., 2010). The catalytic cysteine is oriented away from the conserved arginine in Sac1, while the corresponding residue in PTEN and MTMs faces toward the arginine and generates a narrow active center. This observation suggests that Sac1's catalytic center probably undergoes a conformational change during the catalysis. This premise appears to agree with the fact that Sac1 is an allosteric enzyme, and its activity is stimulated by anionic phospholipids (Zhong et al., 2012). The Sac domain is also found in synaptojanins (see section 9.1). It is responsible for the removal of phosphate from D3, D4, and/or D5 positions of various PIs, and thus, Sac is considered to be a PI phosphatase with broad specificity. Sac1 dephosphorylates PtdIns3P, PtdIns4P, and PtdIns(3,5)P2, but not PtdIns(4,5)P<sup>2</sup> (Nemoto et al., 2000), and Sac2 acts on D5 position of PtdIns(4,5)P<sup>2</sup> and PtdIns(3,4,5)P<sup>3</sup> (Minagawa et al., 2001), whereas Sac3 hydrolyzes only PtdIns(3,5)P<sup>2</sup> (Botelho et al., 2008). Sac1 is mostly localized on the ER and shuttles between the Golgi and ER. The C-terminus of Sac1 mediates an association with the COPI complex via a conserved KXKXX motif and this association induces the retrieval of Sac1 to the ER (Blagoveshchenskaya et al., 2008). Sac1 preferentially utilizes PtdIns4P as its substrate, and mutations that downregulate Sac1 cause the cellular PtdIns4P levels to increase in yeast and mammals (Guo et al., 1999; Nemoto et al., 2000). Sac1 has been also shown to be involved in the maintenance of the plasma membrane PtdIns4P levels at the ER-plasma membrane junctions (Stefan et al., 2011). Sac2 is a mammalianspecific negative regulator of the Akt pathway (Trivedi et al., 2007), and involved in the endocytic pathway as a PI 4 phosphatase (Nakatsu et al., 2015). Sac3 is important for the regulation of the endocytic pathway given that it regulates PtdIns(3,5)P<sup>2</sup> levels. Sac3 deficiency causes PtdIns(3,5)P<sup>2</sup> levels to increase and impairs late-endosome to lysosome transition. It is also involved in the regulation of PIPKIII enzymatic activity (see section 6.3.1).

# 10.2. Sac of *E. histolytica*

In the E. histolytica genome database, three proteins (EHI\_141860, EHI\_040380, EHI\_048570) showed E-values lower than 1 × 10−<sup>10</sup> with human Sac1 in a blastp search. These proteins were also detected when Sac2 and Sac3 were used as queries, and the E-values were lower than those obtained with Sac1. Therefore, these three proteins seem to be homologous to the human Sac1. EHI\_141860 has the highest homology to the human Sac1, conserves the two transmembrane domains (**Figure 4**), and thus, it is considered to be the E. histolytica Sac1 ortholog. EHI\_141860 also has the C-terminal COPI complex binding motif, KXKXX. As mentioned above, EHI\_040380 has a DNaseI domain. Since 5-Ptase domain shares homology to the apurinic/apyrimidinic family of endonucleases (Whisstock et al., 2000), there is a possibility that this protein functions like synaptojanin. However, as the domain in EHI\_040380 lacks the residues critical for the active center of PI 5-phosphatase (**Supplementary Figure S3**), we classified it as "SAC with DNase I or PI 5-phosphatase" domain (**Figure 4**). Nevertheless, further experimental evidence on its PI 5-phosphatase potential is necessary. Since E. histolytica appears to lack PI 4-phosphatases (see section 8), Sac orthologs may act as PI 4-phosphatase in this organism.

# 11. POSSIBLE PATHWAY-DEPENDENT COORDINATED REGULATION OF KEY PI METABOLIZING ENZYMES

# 11.1. Biosynthetic, Secretory, and Exocytotic Pathways

While a majority of interconversion steps between specific PIs are catalyzed by >1 enzymes, some steps are catalyzed by a single enzyme encoded by a single gene, which may suggest the significance of the enzyme. Furthermore, mRNA expression levels, inferred by transcriptomic analyses, often suggest biological importance of the enzyme(s) for the reaction under given conditions. In the biosynthetic and secretory pathways, the PtdIns4P level is maintained in an organelle specific fashion: low in the ER and high in the Golgi, by the coordinated action of PI 4-kinases, PI4KIIα and PI4KIIIβ, and Sac1 (Blumental-Perry et al., 2006; Graham and Burd, 2011; Bajaj Pahuja et al., 2015). E. histolytica possesses only one PI 4-kinase, EHI\_148700, which is most likely involved in this pathway. Role of Sac1 to maintain the low PtdIns4P level in the ER was shown (Bajaj Pahuja et al., 2015). It is conceivable to assume a single two transmembrane domain-containing PI-phosphatase Sac, EHI\_141860, likely plays important role to maintain the low PtdIns4P level in the Golgi to regulate biosynthetic and secretory pathways. The transcriptome data also suggest the robust expression (**Supplementary Figure S1**) and thus the significance of EHI\_141860.

Since clathrin-mediated trafficking machinery for the exocytic pathway is well-conserved in this organism (Clark et al., 2007), PtdIns4P is likely used in the Golgi apparatus as a key signaling molecule to recruit effector molecules, such as clathrin binding proteins AP, Arf, and Rab11for cargo selection and packaging. Following the generation of transport vesicles in the Golgi, PtdIns4P on the PtdIns4P-rich transport vesicles is replaced with sterol by oxysterol binding protein (OSBP), to form sterol-rich vesicles (Schink et al., 2016). This exchange is necessary to recruit exocyst complex onto the transfer vesicles and also function as sterol transfer mechanism from the ER to the plasma membrane (Schink et al., 2016). It remains elusive if this mechanism also works in E. histolytica, although its genome encodes 2-4 possible OSBP (Das and Nozaki, 2018).

At the plasma membrane, PtdIns(4,5)P<sup>2</sup> generated from PtdIns4P by single type I PIP kinase, EHI\_153770, likely determines the site of exocytosis, where the exocyst complex mediates a release of the content of the sterol-rich secretory vesicles. Sec3 and Exo70 of the exocyst complex are known PtdIns(4,5)P<sup>2</sup> effectors on secretory vesicles. On the plasma membrane, Syntaxin-1, CAPS, Munc13-1/2, and Synaptotagmin-1 involved in the fusion of transport vesicles and the plasma membrane, leading to secretion (Martin, 2015). E. histolytica conserves a homolog of Syntaxin (EHI\_052830, E-value 1 × 10−18), Sec3 (EHI\_148590, E-value 9 × 10−<sup>6</sup> ), and Exo70 (EHI\_142040, E-value 2 × 10−<sup>8</sup> ). The functionality of the apparently conserved basic amino acids implicated for PtdIns(4,5)P<sup>2</sup> binding should be verified for the amebic homolog (Martin, 2015; K. Nakada-Tsukui data not shown).

# 11.2. Endocytic Pathways

In the clathrin-dependent endocytic pathway, AP complex connects membrane cargo receptors and clathrin via PtdIns(4,5)P<sup>2</sup> by recognizing the cytoplasmic region of the cargo receptors and PtdIns(4,5)P2. In this process, single type I PIP kinase, EHI\_153770, is necessary to synthesize PtdIns(4,5)P<sup>2</sup> from PtdIns4P. During the scission of the vesicles from the plasma membrane, generation of PtdIns(3,4)P<sup>2</sup> from PtdIns4P by class II PI 3-kinase is necessary to recruit SNX9 (sorting nexin that recognizes membrane curvature and PIs) in mammals. However, E. histolytica does not possess either PX-domain containing class II PI 3-kinases or BAR domain containing SNXs, e.g., SNX9. However, it is conceivable that one or some of six class I PI 3-kinases also have PtdIns4P 3-kinase activity and some of putative SNXs lacking BAR domain have ability to

recognize PIs (N. Watanabe et al., data not shown). Therefore, it is expected that as enclosed endosomes mature after closure, PtdIns(4,5)P<sup>2</sup> is subsequently dephosphorylated into PtdIns and then further phosphorylated to PtdIns3P by the action of type III PI 3-kinase, Vps34, which is present in E. histolytica as a single protein. In mammals, a series of dephosphorylation reactions involving PtdIns(4,5)P<sup>2</sup> are regulated by OCRL1 and Sac2 (Nakatsu et al., 2015) and Synaptojanins (Cremona et al., 1999). Since Synaptojanins, which contain PI 5-phosphatase and PI 4-phosphatase domains, are not conserved in E. histolytica, the most highly transcribed PI 5-phosphatase, EHI\_160860, out of 6 type II PI 5-phosphatases, and one of two transmembrane lacking Sac proteins, EHI\_040380 and EHI\_048570, are likely involved in this process (**Figure 4; Supplementary Figure S1**).

# 11.3. Phago- and Trogocytic Pathways

During phagocytosis, one single type I PIP-kinase, EHI\_153770, and/or one or more of six class I PI 3-kinases are likely involved in local enrichment of PtdIns(4,5)P<sup>2</sup> and PtdIns(3,4,5)P<sup>3</sup> from PtdIns4P in response to the signal from a not-yet-identified ligand receptor (most likely galatose/N-acetylgalactosamine specific lectin) for phagocytosis at the plasma membrane. Based on the mRNA expression levels, five out of 6 class I PI 3 kinase genes appear to be abundantly expressed at similar levels and thus, it is not clear which is predominantly involved in this process (**Supplementary Figure S1**). In mammalian THP-1 cells, isoform-specific roles of class I PI 3-kinases were reported: p110α is involved in FcγR-mediated phagocytosis and oxidative burst mediated by PMA or opsonized zymosan, but not in CR3 mediated phagocytosis (Lee et al., 2007). On the other hand, p110β is involved in Rab5 recruitment and activation during phagosome maturation, while p110δ is involved in adhesion to VCAM-1 (Kurosu and Katada, 2001; Ferreira et al., 2006; Thi et al., 2012; Whitecross and Anderson, 2017).

In E. histolytica, two different modes of ingestion for target uptake, phagocytosis and trogocytosis, have been observed, and they have been shown to be regulated by different AGC kinases in an isotype-specific manner. It is conceivable that different receptors and class I PI 3-kinases are differentially involved in these processes. Elucidation of the isoform-specific involvement of class I PI 3-kinases in trogocytosis and phagocytosis shall be important to understand the pathogenesis of E. histolytica.

On the enclosed phagosomes, PtdIns(4,5)P<sup>2</sup> and PtdIns(3,4,5)P<sup>3</sup> are, as also seen in endocytosis, dephosphorylated to PtdIns by PI-phosphatases and then phosphorylated again to form PtdIns3P by class III PI 3 kinase, such as OCRL1, NPP5B, SHIP, INPP5E, TMEM55a, myotubularin, Sac2, and Vps34, in mammals (Cox et al., 2001; Ai et al., 2006; Horan et al., 2007; Neukomm et al., 2011; Bohdanowicz et al., 2012; Levin et al., 2017; Morioka et al., 2018). Similarly, conversion of PIs on the E. histolyica phagosomes is expected to proceed in a similar but possibly modified fashion. Among a number of type II PI 5-phosphatases, Sacs, myotubularins, and type III PI 3-kinase, we assume the following members are likely involved in this process. Among eight myotubularins that appear to be catalytically active, three isotypes, EHI\_016430, EHI\_104710, and, EHI\_024380, show relatively high expression and are likely involved in this maturation process. Once phagosomes are decorated with PtdIns3P, PtdIns3P is further phosphorylated to PtdIns(3,5)P<sup>2</sup> as phagosomes are further maturated. While in mammals, PIKfyve is involved in this process, type III PIP-kinase lacking FYVE domain, EHI\_049480, may be responsible for this reaction in E. histolytica. As described above (section 6.3.1), type III PIP-kinases form complex with Sac phosphatase and scaffold proteins (Sbrissa et al., 2007; Botelho et al., 2008; Jin et al., 2008). Since two Sacs lacking the transmembrane domain, EHI\_040380 and EHI\_048570, are present in E. histolytica, it is possible that one of them forms complex with EHI\_049480, similar to mammalian Sac3, while the other independently works in endocytic and phagocytic pathways similar to mammalian Sac2 (Nakatsu et al., 2015; Levin et al., 2017).

# 11.4. Motility

In the regulation of cell motility, local accumulation of PtdIns(4,5)P<sup>2</sup> and PtdIns(3,4,5)P<sup>3</sup> at the leading edge is the key initial event. Similar to phagocytosis, type I PIPkinase, EHI\_153770, and some of six type I PI 3-kinases are likely involved in this process. Also, dephosphorylation of PtdIns(3,4,5)P<sup>3</sup> by PI 3-phosphatases, PTEN, and PI 5 phosphatase, SHIP, at the side and the rear of the cell is known to be indispensable to regulate local accumulation of the lipid signal in mammals. In E. histolytica, some of six PTEN homologs and six PI 5-phosphatases are likely involved in this process. Among three of six amebic PTENs that contain C2 domain and the putative cytosol localization signal (see section 7.1.2), two showed significantly higher expression levels than four other PTEN isotypes (**Supplementary Figure S1**). Altogether, these data suggest that these two PTENs, EHI\_197010 and EHI\_098450, may be involved in the formation and maintenance of cellular polarity in E. histolytica.

# 11.5. Nuclear Functions

It is conceivable that PI kinases and PIP phosphatases that contain the nuclear localization signal have specific roles in the nucleus such as chromatin regulation and transcription. Such PI kinases and PIP phosphatases include type I PIP-kinase, EHI\_153770; PTEN, EHI\_041900; MTM, EHI\_070120; and PI 5-phosphatase, EHI\_046590. Also, PTENs that lack the cytosol localization signal, EHI\_041900, EHI\_010360, and EHI\_054460, may also be involved in nuclear functions.

# 12. CONCLUSION AND FUTURE PERSPECTIVE

One of the hallmarks of E. histolytica as an invasive eukaryotic pathogen is its extremely active cell motility accompanied with elaborate cytoskeletal rearrangement and membrane traffic. To enable such activities, spaciotemporal regulation of PI-mediated signaling that controls transient association with effector molecules is indispensable and accomplished via concerted regulation of PI metabolism. The E. histolytica genome encodes the majority of PI kinases and PI phosphatases conserved in model organisms. Strikingly, significant diversity of PI 3-kinases and PI 3-phosphatases was observed in E. histolytica, as represented by a higher level of complexity of class I PI 3-kinases, PTEN, MTM/MTMR, and IMLRKs in this unicellular eukaryote relative to human, which has a 100 times larger genome. The dependence of E. histolytica on the complexity of the D3 phosphate metabolism emphasizes the significance of PtdIns(3,4,5)P3 centric pathways for pathogenesis and physiology of E. histolytica.

On the other hand, the regulatory subunit of PI kinases, except for class III PI 3-kinase, was not identified, suggesting that their regulatory mechanisms had been gained only in higher eukaryotes or had differently evolved in E. histolytica in a lineage-specific fashion. The latter was also observed as an example with PI 4-kinase regulator EFR3 which is not conserved between the yeast and human. PtdIns4P metabolism also appears to have uniquely evolved in E. histolytica. No PtdIns4P-specific phosphatases that show similarity to the canonical enzymes are conserved in E. histolytica. It harbors only one PIP kinase, type I PIP kinase, which generate PtdIns(4,5)P<sup>2</sup> from PtdIns4P. PtdIns4P is known as one of the major PIs and is important as it is the precursor of the most abundant PI, PtdIns(4,5)P2. Since PtdIns(4,5)P<sup>2</sup> is indispensable for the regulation of actin cytoskeleton-dependent processes, which is vital for the pathogenesis of the amoeba, type I PIP kinase appears to be a rational drug target. Uniquely expanded gene families, such as class I PI 3-kinases and PTENs, may also be potential drug target. However, multiple enzymes may have a redundant role as shown for mammalian PIPKI (see section 6.1.1).

Besides expansion, certain families of PI kinases and PI phosphatases in E. histolytica are structurally unique in the sense that they have simpler domain configurations, especially type III PIP kinases and PI phosphatases, relative to their human counterparts with a exceptions such as inactive myotubularin/LRR/ROCO/kinase (IMLRK). Lineage-specific expansions of PIP phosphatases are found in particular for OCRL1 type II PI 5-phosphatase and IMLRK, some of which should be listed in the roster of rational drug targets once their functions are determined.

Extremely higher expression levels of two PTEN and one PI 5-phosphatase genes relative to other genes involved in PI metabolism may reflect the importance of phosphatases rather than kinases in stopping the PI signals. PTEN has also been reported to function as a protein phosphatase, and thus, it is also possible that the high expression of PTEN is because it has roles other than PI signaling (Shinde and Maddika, 2016; Wozniak et al., 2017).

Furthermore, it has recently been demonstrated that besides the phosphorylation status, type pf the acyl-chains in the lipids are important in the regulation of lipid functions (Choy et al., 2017). In mammals, the predominant type found in PIs is 1-stearoyl-2-arachidonoyl (18:0/20:4) (Traynor-Kaplan et al., 2017). This acyl chain type is generated by lysocardiolipin acyltransferase (LYCAT) (Imae et al., 2012), whose deficiency persturbes PI-mediated-membrane traffic (Bone et al., 2017). Mechanisms underlying such acyl-dependent regulation of PI signaling and downstream cascades are just about to unveil in the lipid research field and should be explored in amebiasis research. Furthermore, a new family of lipid transport proteins that mediate the regulation of PI metabolisms in both the cytoplasm and the nucleus have been described (Das and Nozaki, 2018). It may also be of interest that several PI kinases and PI phosphatases as well as several PI species are localized in the nucleus. However, PI metabolism and physiological roles of PI in the nucleus is poorly understood.

# AUTHOR CONTRIBUTIONS

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

# FUNDING

This work was supported in part by a grant for Research on Emerging and Re-emerging Infectious Diseases from Japan Agency for Medical Research and Development (AMED) (JP18fk0108046 and JP19fk0108046 to TN; JP18fk0108049 and JP19fk0108049 to KN-T), a grant for US-Japan Cooperative Medical Science Program from AMED (JP17jk0210018 to KN-T), grants from Japan Ministry of Education, Culture, Sports, Science, and Technology (MEXT)/Japan Society for the Promotion of Science (JSPS) (KAKENHI JP17K19416, and JP18H02650 to TN; JP16K08766, JP18H02650, and JP19H03463 to KN-T), a grant from the National Center for Global Health and Medicine (29–2013 to TN), and a grant from Science and Technology Research Partnership for Sustainable Development (SATREPS) from AMED and Japan International Cooperation Agency (JICA) to TN.

# REFERENCES


# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2019.00150/full#supplementary-material

Supplementary Figure S1 (Image 1) | Expression levels of PI kinases and PI phosphatases in E. histolytica trophozoites. Relative expression levels of indicated PI kinase and PI phosphatase genes of the HM-1:IMSS cl6 reference strain during in vitro cultivation. Signal intensity was normalized against the transcript of RNA polymerase II gene. Two bars represent the data from two independent experiments (Husain et al., 2011; Penuliar et al., 2015).

Supplementary Figure S2 (Image 2) | Cytosolic localization signal in human and E. histolytica PTEN. The cytosol localization signal of human PTEN was aligned with the corresponding region of E. histolytica PTEN orthologs. Green underlines depict the key amino acids for the signal and the amino acids conserved in PTENs from human and E. histolytica are indicated with green boxes.

Supplementary Figure S3 (Image 3) | Conservation of consensus amino acid sequences in PI 5-phosphatases. Amino acids of two consensus regions of PI 5-phosphatases, human OCRL1 and E. histolytica orthologs. Gray boxes indicate conserved amino acids.

Supplementary Figure S4 (1–4) (Images 4–7) | Multiple alignment of class I PI 3-kinases. LKU domain and catalytic core are indicated by green and yellow bar, respectively.

Supplementary Figure S5 (1–3) (Images 8–10) | Multiple alignment of PTEN. Catalytic core is indicated by yellow bar.

Supplementary Table S1 (Tab 1 in Table 1) | Potential PI kinase orthologs in E. histolytica. H. sapiens enzymes used as queries, and E-values are shown.

Supplementary Table S2 (Tab 2 in Table 1) | Potential PI phosphatase orthologs in E. histolytica. H. sapiens enzymes used as queries, and E-values are also shown.

Supplementary Table S3 (Tab 3 in Table 1) | Similarity between H. sapiens and E. histolytica class I PI 3-kinases. H. sapiens isotypes used as queries, and E-values are also shown.

Supplementary Table S4 (Tab 4 in Table 1) | Similarity between H. sapiens and E. histolytica MTM and MTMRs. H. sapiens isotypes used as queries, and E-values are also shown.

Supplementary Table S5 (Tab 5 in Table 1) | Similarity between H. sapiens and E. histolytica PI 5-phosphatases.

Supplementary Table S6 (Tab 6 in Table 1) | Abbreviation list.


transport from the endoplasmic reticulum. Proc. Natl. Acad. Sci. U.S.A. 112, E3199–E3206. doi: 10.1073/pnas.1509119112


J. Physiol. Renal. Physiol. 295, F1422–1430. doi: 10.1152/ajprenal.903 10.2008


phosphatase by MTMR5, a catalytically inactive phosphatase. Proc. Natl. Acad. Sci. U.S.A. 100, 4492–4497. doi: 10.1073/pnas.04310 52100


subunits with the small GTPase Rab5. J. Biochem. 130, 73–78. doi: 10.1093/oxfordjournals.jbchem.a002964


dynamin. J. Biol. Chem. 273, 19108–19119. doi: 10.1074/jbc.273.30. 19108


in chemotaxis and cell migration. Br. J. Pharmacol. 171, 5541–5554. doi: 10.1111/bph.12777


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Nakada-Tsukui, Watanabe, Maehama and Nozaki. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Non-vesicular Lipid Transport Machinery in *Entamoeba histolytica*

### Koushik Das and Tomoyoshi Nozaki\*

Graduate School of Medicine, The University of Tokyo, Bunkyo, Japan ¯

Eukaryotic cells are organized into separate membrane-bound compartments that have specialized biochemical signature and function. Maintenance and regulation of distinct identity of each compartment is governed by the uneven distribution and intra-cellular movement of two essential biomolecules, lipids, and proteins. Non-vesicular lipid transport mediated by lipid transfer proteins plays a pivotal role in intra-cellular lipid trafficking and homeostasis whereas vesicular transport plays a central role in protein trafficking. Comparative study of lipid transport machinery in protist helps to better understand the pathogenesis and parasitism, and provides insight into eukaryotic evolution. Amebiasis, which is caused by Entamoeba histolytica, is one of the major enteric infections in humans, resulting in 40–100 thousand deaths annually. This protist has undergone remarkable alterations in the content and function of its sub-cellular compartments as well represented by its unique diversification of mitochondrion-related organelle, mitosome. We conducted domain-based search on AmoebaDB coupled with bioinformatics analyses and identified 22 potential lipid transfer protein homologs in E. histolytica, which are grouped into several sub-classes. Such in silico analyses have demonstrated the existence of well-organized lipid transport machinery in this parasite. We summarized and discussed the conservation and unique features of the whole repertoire of lipid transport proteins in E. histolytica.

#### *Edited by:*

Anjan Debnath, University of California, San Diego, United States

### *Reviewed by:*

Esther Orozco, Centro de Investigación y de Estudios Avanzados (CINVESTAV), Mexico Nancy Guillen, Centre National de la Recherche Scientifique (CNRS), France

#### *\*Correspondence:*

Tomoyoshi Nozaki nozaki@m.u-tokyo.ac.jp

#### *Specialty section:*

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

*Received:* 27 April 2018 *Accepted:* 20 August 2018 *Published:* 19 September 2018

#### *Citation:*

Das K and Nozaki T (2018) Non-vesicular Lipid Transport Machinery in Entamoeba histolytica. Front. Cell. Infect. Microbiol. 8:315. doi: 10.3389/fcimb.2018.00315 Keywords: lipid, *E. histolytica*, signaling, protozoan parasites, pathogenesis

# INTRODUCTION

Eukaryotic cells are organized into distinct membrane-enclosed organelles or compartments. Each organelle has distinctive lipid and protein signature and dedicated function (Holthuis et al., 2003; Lev, 2010). Each cell organelle also assures proper segregation of complex cellular processes catalyzed by metabolic enzymes, structural and regulatory proteins (Lev, 2010). Proteins are precisely distributed to different cell organelles either by their intrinsic signal peptides or through post-translational modifications (Lev, 2010). In contrast, lipids do not have any such signal sequence that determine their accurate intracellular distribution (Lev, 2010). Nevertheless, each organelle differs in its lipid composition (Voelker, 1991; Sprong et al., 2001; Lev, 2010). In general, ER is the main site for lipid synthesis (Blom et al., 2011) and lipids are then transported from ER to site of function. Previous studies suggest that both vesicular and non-vesicular transport machinery are responsible for the delivery of lipids to their final destinations (Lev, 2010). Vesicular transport has an important role in protein trafficking, endocytic and exocytic (secretory) pathways (Lev, 2010). It is an energy dependent process and involves cytoskeletal reorganization (Lev, 2010). However, significant amount of lipids can be transferred by vesicular transport as lipids are the major component of transport vesicles (Lev, 2010). Nonetheless, lipid transport was still identified when vesicular transport was impaired by either depletion of ATP, reduced temperature, or treatment with pharmacological inhibitors (such as brefeldin A and colchicine) (Kaplan and Simoni, 1985; Vance et al., 1991; Li et al., 2014). Lipid transportation was also detected among cell organelles, those are not linked by vesicular transport machinery (e.g., ER/mitochondria and ER/peroxisomes) (Levine, 2004; Holthuis and Levine, 2005). These observations suggest that non-vesicular transport mechanisms have a significant role in intracellular lipid trafficking. Non-vesicular lipid transport in and between organellar membranes is mostly facilitated by three possible methods: lateral diffusion, trans-bilayer flip-flop, and monomeric lipid exchange (Sleight, 1987; Van Meer, 1989; Lev, 2010). Lateral diffusion is responsible for the lateral movement of lipid in a membrane bilayer (Lev, 2010). Although lateral diffusion mostly transport lipid within membranes, this process was also identified between membranes which are linked via membrane bridges (Lev, 2010). Lipids are moved between two layers of the membrane bilayer by the process called trans-bilayer flip-flop (Lev, 2010). This type of movement takes place either spontaneously or mediates by flippases and translocases (Sprong et al., 2001; Lev, 2010). Trans-bilayer flip-flop do not participate directly in inter-organelle lipid transport (Lev, 2010). It can either encourage non-vesicular lipid transport by monomeric lipid exchange or influence vesicular transport through the alteration of membrane curvature, vesicle budding and fusion (Sprong et al., 2001; Lev, 2006, 2010). Monomeric lipid exchange, the primary mechanism of intra-cellular lipid transport is an energy-independent process (Lev, 2010). In this process, lipid monomer is transported from a donor membrane to an acceptor membrane through the cytosol either spontaneously or facilitated by lipid transfer proteins (LTPs) (Lev, 2010). Spontaneous transport of lipid molecule from donor to the acceptor membrane through cytosol is a time- consuming process and insufficient for substantial transport of major lipids (Jones and Thompson, 1989; Mesmin and Maxfield, 2009; Lev, 2010, 2012).

Non-vesicular lipid transport between cellular membranes are greatly facilitated by LTPs, which are the key contributor of organelle-specific lipid distribution and cellular lipid homeostasis (Helle et al., 2013). LTP-mediated lipid transport locally modulates the lipid composition of membranes and consequently regulates various cellular processes including vesicular trafficking, lipid metabolism, and signal transduction (Ohashi et al., 1995; Kim et al., 2015). Furthermore, LTPs can also act as membrane contact sites (MCSs) between the endoplasmic reticulum (ER) and virtually all other organelles, and are involved in the transport of Ca2+, metabolites, and lipids (Lev, 2010; Helle et al., 2013).

Amebiasis is caused by a protozoan parasite Entamoeba histolytica and one of the major enteric infections in humans. An estimated 50 million people are infected with E. histolytica worldwide, resulting in 40–100 thousand human deaths annually (Haque et al., 2003; Stanley, 2003; Hung et al., 2012). Invivo growth and survival of E. histolytica as an enteric parasite depends on its efficient cellular processes to thrive in adverse host environment (Mittal et al., 2008; Nakada-Tsukui et al., 2009; Vicente et al., 2009). The parasite has two interchangeable stages consisting of the infective dormant cyst and the proliferative motile trophozoite stages in the life cycle. Interconversion between cysts and trophozoites involves complex cellular metabolic processes and essential for transmission of the disease. Lipid and its metabolism apparently play important roles in stage conversion (De Cádiz et al., 2013). Furthermore, this anaerobic or microaerophilic parasite has to overcome a wide variety of environmental and host-derived oxidative and nitrosative stresses during its life cycle (Vicente et al., 2009; Baumel-Alterzon and Ankri, 2014; Pineda and Perdomo, 2017). The pathogenic behavior of this parasite depends on their ability to uptake host nutrients through ingestion and degradation of mammalian cells and tissues. E. histolytica trophozoites engulf live and dead host cells through trogocytosis and phagocytosis, respectively (Nakada-Tsukui et al., 2009; Ralston et al., 2014). Contact dependent cytolysis of host cells depends on the secretion of cytolytic proteins, including cysteine proteases (CP) (Que and Reed, 2000) and pore forming peptides (Zhang et al., 2004). Such contact dependent killing and uptake of host cells are likely primed with a receptor-ligand interaction on the surface, which leads to activation of lipid signaling cascades to downstream effectors (Nakada-Tsukui et al., 2009; Somlata et al., 2017).

Lipids can either directly affect the membrane charge and the curvature by their altered local distribution or act as precursors to accelerate the local lipid synthesis by lipid metabolizing enzymes. Furthermore, it has also been established that in the nucleus, distribution and metabolism of lipids and lipidmediated signaling play essential roles in cell proliferation, differentiation, and stress adaptation in higher eukaryotes (Shah et al., 2013). By analogy, non-vesicular lipid transport mediated by LTPs are likely indispensable for E. histolytica. As E. histolytica resides in the anaerobic or microaerophilic environment, its mitochondrial functions have highly diverged, resulting in the atypical mitochondrion-related organelles (MROs), called mitosomes, which have drastically modified morphological, structural, and functional features compared to those of the aerobic mitochondria (Makiuchi and Nozaki, 2014). Likewise, E. histolytica also lacks well-defined morphologically discernible ER and Golgi apparatus (Perdomo et al., 2016). Lipid transport machinery between such highly divergent organelles in E. histolytica could be significantly different from those in the well characterized higher eukaryotes. Thus, investigation on the LTP-mediated lipid transfer in E. histolytica should certainly help us to better understand conservation and diversity of this important biological process. Here, we conducted an InterPro domain search (PFAM) analysis on AmoebaDB in order to identify the potential LTP candidates in E. histolytica. We also discussed the conservation and unique features of LTPs, and their possible mechanisms and roles in biology and pathophysiology of amebiasis.

# GENERAL BACKGROUND OF LIPID TRANSPORT BY LIPID TRANSFER PROTEINS (LTPs)

# Domains of LTPs

Non-vesicular lipid transport is catalyzed by LTPs. LTPs can extract lipids from the donor membrane and deliver them to the acceptor membrane. This type of transfer involves a special lipid-transfer domain (LTD) that can form a lipid binding cavity, which binds and accommodate the hydrophobic moieties of lipid molecules from the aqueous environment (Helle et al., 2013). A combination of hydrogen bonds and hydrophobic interactions secure the lipid binding and influence the binding affinity (Lev, 2010). In addition to LTDs, LTPs often possess different combinations of membrane/organelle targeting domains, which direct them to special cellular compartments (Helle et al., 2013), such as pleckstrin homology (PH), protein kinase C (PKC) conserved 1 (C1) and PKC conserved 2 (C2) domains. C1 domain (∼50 amino acids, a cysteine-rich compact structure) was first identified in protein kinase C (PKC) as the binding site for di-acyl-glycerol (DAG) and phorbol ester (Cho, 2001). The binding site has cationic residues, which accelerate the Ca+<sup>2</sup> dependent recruitment and adsorption of the C1 domain to the anionic membrane surfaces (Cho, 2001), while the hydrophobic tip of the domain penetrates the membrane to bind with DAG that is partially buried in the membrane (Cho, 2001). C2 domain (∼130 amino acids, Ca+<sup>2</sup> binding site) was found in PKC, cytosolic phospholipase A<sup>2</sup> (PLA2), phospholipase C (PLC), phospholipase D (PLD) and phosphoinositide (PtdIns phosphate, PI) 3-Kinase (Cho, 2001). Ca+<sup>2</sup> ions assist the membrane targeting of C2 domain either through providing a bridge between the C2 domain and anionic phospholipids or induce intra-domain conformational change which in turn triggers membrane protein interactions (Cho, 2001). Sub-cellular localization of C2 domain depends on its phospholipid specificities. C2 domain of PKC prefers anionic phospholipids rapidly translocate to PM, while the C2 domain of PLA<sup>2</sup> is selective to phosphatidylcholine (PC) localized to the perinuclear region in response to Ca+<sup>2</sup> import (Cho, 2001). LTPs are broadly classified into two major classes based on their domain architecture: (i) cytosolic LTPs, which lack any membrane binding domain, and (ii) membrane anchored LTPs, which contain some membrane binding domain(s) and function to form a MCS. Cytosolic LTPs could possibly facilitate intracellular transport of lipid through a sequential process involving interaction of the LTP with donor membrane (Lev, 2012) followed by the opening of hydrophobic cavity, lipid extraction and dissociation of LTP from donor membrane, movement through cytosol in a "closed" transport competent conformation (Kasper and Helmkamp, 1981; Helmkamp, 1986; Nichols, 1988; Rueckert and Schmidt, 1990; Wirtz, 1991; Gadella and Wirtz, 1994; Wirtz et al., 2005; Lev, 2012). The transport is completed by the interaction of LTP with an acceptor membrane, opening of the lipid binding cavity, and desorption of lipid molecule (Lev, 2012) **(Figure 1A)**. In contrast to cytosolic LTPs, LTPs with two targeting domains/motifs for two different organelle membranes are naturally directed to membrane contact sites (MCSs) between

donor membrane via its membrane binding domain (purple) and exposes its LTD (green) to the cytoplasm. LTD can extract the lipid molecule (yellow) from the donor membrane and deliver it to a closely positioned acceptor membrane.

these two cellular compartments (Lev, 2010; Helle et al., 2013). Such MCSs bring two membranes from different organelles in a close vicinity (at 10-20 nm), which favor lipid exchange between such closely apposed membranes **(Figure 1B)** and also regulate intra-cellular Ca+<sup>2</sup> and signaling processes (Levine, 2004; Voelker, 2005; Levine and Loewen, 2006; Giorgi et al., 2009; Lebiedzinska et al., 2009; Lev, 2010).

# Biological Roles of LTPs

LTPs have different modes of action: First, LTPs can facilitate vectorial and often bidirectional lipid transfer as shown in intact living cells and in vitro (Helle et al., 2013). Second, LTPs have been recognized as lipid presenting proteins for lipid metabolizing enzymes (Kular et al., 1997, 2002; Cockcroft and Garner, 2013). It can transiently modulates the intracellular lipid metabolism by providing lipid substrates to lipid metabolizing enzymes during signal transduction pathways under various physiological and cellular conditions (Cockcroft and Garner, 2013; Fayngerts et al., 2014). Third, LTPs can act as lipid sensors by altering their affinity to other associated proteins in response to the binding to lipids or bio-membranes (Lev, 2010). Fourth, LTPs can facilitate transient alterations in lipid distribution of a bio-membrane by extracting or delivering of lipid molecules to a certain region of the membrane, or through changing the lipid phase in a particular membrane portion to which it is bound (Lev, 2010). LTP can employ more than one of these mechanisms and regulates the following cellular process (Ohashi et al., 1995; Kim et al., 2015).

# Intra-Cellular Lipid Trafficking

Inter-organelle lipid transport facilitated by steroidogenic acute regulatory protein-related lipid transfer (START) domain containing proteins is mostly studied. Mammalian START domain containing proteins are sub-divided into broad subfamilies based on their domain organization (membrane targeted and soluble START) and lipid specificities (cholesterol/oxysterol binding proteins and phospholipid/ sphingolipid binding proteins). STARD1 sub-family (contain of STARD1 and STARD3/MLN64) and STARD4 sub-family (comprised of STARD4, STARD5, and STARD6) transport cholesterol to various cell organelles (Alpy and Tomasetto, 2005; Clark, 2012). STARD2 sub-family is composed of STARD2, STARD7, STARD10, and STARD11/CERT (Alpy and Tomasetto, 2005; Clark, 2012). STARD2, 7 and 10 transport PC/PE (Alpy and Tomasetto, 2005; Clark, 2012), while STARD11/CERT transport ceramide from ER to Golgi complex (Alpy and Tomasetto, 2005; Clark, 2012; Kumagai et al., 2014). Among the above mentioned START proteins, STARD3/MLN64 and STARD11/CERT have additional membrane targeted domain, rest possess only STARD domain and are cytosolic (Alpy and Tomasetto, 2005; Clark, 2012).

Lipid Supply for Metabolism and Signal Transduction The receptor-ligand mediated signaling processes such as trogocytosis, phagocytosis, pinocytosis and exocytosis involves an array of PIs and their metabolizing enzymes (PI 4-kinases, PtdIns4P 5-kinases, PI 3-phosphatases, PI 5-phosphatases, and non-specific phosphatases), residing in various cell organelles (Kölsch et al., 2008; Thomas, 2012; Haastert et al., 2013; Levin et al., 2015). However, PtdIns, the main precursor of PIs is synthesized in the ER and needs to be transported by LTPs (also known as PtdIns transfer proteins, PITPs) to cell organelles for the generation of PIs pools during these cellular processes. PA is produced from diacylglycerol (DAG) by DAG kinases at the plasma membrane (PM), also transported back to the ER by PITPs for replenishment of PtdIns at the ER (Cockcroft and Garner, 2013). This example illustrates that LTPs can function as lipid presenting proteins for lipid metabolizing enzymes and subsequently modulates the lipid metabolism associated with signal transduction pathways under various physiological and cellular conditions.

# Lipid Sensing, and Regulation of Vesicular Trafficking

LTPs can function as lipid sensors and regulate Golgi-mediated vesicular trafficking, exocytosis (Litvak et al., 2005; Peretti et al., 2008; Mattjus, 2009), as well explained for Sec14 in S. cerevisiae (Curwin et al., 2009). Sec14 involved in intra-cellular transport of either phosphatidylcholine (PC) or phosphatidylinositol (PtdIns) between the ER and the Golgi complex (Bankaitis et al., 1990; Lev, 2010). However, the PtdIns–PC exchange activity of Sec14 is not involved in the Golgi secretory function. Instead, Sec14 functions as PC sensor, can sense the PC level in Golgi and respond to increased PC level by inhibiting its production from diacylglycerol (DAG) through cytidine diphosphate (CDP) choline pathway (McGee et al., 1994; Skinner et al., 1995; Lev, 2010). In this way, Sec14 regulates a critical level of DAG and PC in Golgi, which is crucial for Golgi mediated vesicular trafficking, exocytosis and viability of S. cerevisiae (Bankaitis et al., 1989; Lev, 2010). Sec14 can function as both as a PC sensor and as a PtdInspresenting protein, which transmits PC metabolic information to PI synthesis (Schaaf et al., 2008; Lev, 2010). Oxysterol-bindingprotein-related proteins (ORPs) interact with Rab GTPases and control intra-cellular movement of transport vesicles as described for ORP1L from higher eukaryotes (Johansson et al., 2007; Rocha et al., 2009; Lev, 2010). ORP1L can induce the formation of the membrane contact site (MCS) between the ER and late endosomes, by undergoing conformational changes in response to lower cholesterol content in late endosomes (Johansson et al., 2007; Rocha et al., 2009; Lev, 2010). Priming and docking of the exocytic vesicle complex with the PM also requires an PITP mediated transient alteration in lipid [e.g., PtdIns(4,5)P2] distribution at the site of exocytosis (Lev, 2010; Thomas, 2012).

# Modulation of Nuclear Lipid Signaling and Associated Nuclear Functions

A repertoire of lipid (PI, PA, and DAG) metabolizing enzymes, their lipid substrates, byproducts and downstream effectors, involved in various aspects of transcription, chromatin remodeling, mRNA maturation, cell proliferation, differentiation, and stress management (Tanaka et al., 1999; Martelli et al., 2002; Audhya and Scott, 2003; Irvine, 2003; Balla and Balla, 2006; Matsubara et al., 2006; Carman and Henry, 2007; Demmel et al., 2008; Mishkind et al., 2009; Ren et al., 2010; Jang and Min, 2011; Shah et al., 2013; Symeon, 2013; Jülke and Ludwig-Müller, 2015; Karlsson et al., 2016), have been identified in eukaryotic nucleus. In order to maintain the critical level of lipid in the nucleus, the precursor for lipid biosynthetic enzymes needs to be transported to the nucleus by LTPs. Other than function as a lipid exchanger, LTP can interacts with other nuclear associated proteins and regulates the nuclear functions as described for microsomal triglyceride transfer protein, which regulate lipid homeostasis and interacts with RNA helicase DDX3, hepatocyte nuclear factor 4 (HNF4) and small heterodimer partner (SHP) (Tsai et al., 2017). An LTP can also modify nuclear transport via its interaction with a nuclear pore component (Nup62) as described for sterol transporter, ORP8 (Zhou et al., 2011; Béaslas et al., 2012).

# Cytoskeleton Organization, Adhesion, and Motility

Several LTP homologs in higher eukaryotes were reported to be associated with cytoskeleton organization, cell adhesion, and motility. For instances, ORP3 and its close relative ORP7 interacts with R-Ras and regulates cytoskeleton organization, cell adhesion, migration (Goldfinger et al., 2007; Weber-Boyvat et al., 2013), while STARD8/12/13 (START proteins with Rho GTPase activating protein (Rho-GAP) domain) are also involved in cytoskeleton organization and migration of a cancer cell line (Alpy and Tomasetto, 2005).

# IDENTIFICATION OF LIPID TRANSFER PROTEIN (LTP) HOMOLOGS IN *E. histolytica*

# Identification, Domain Organization of *E. histolytica* LTP Homologs

We conducted an InterPro domain search (PFAM) analysis on AmoebaDB version 38 (released 5th July, 2018) in order to identify the potential LTP candidates in E. histolytica HM-1:IMSS. The E. histolytica HM-1:IMSS genome encodes a diverse repertoire of 22 potential lipid transfer protein (LTP) homologs [four oxysterol-binding-protein-related (ORP) proteins, 15 steroidogenic acute regulatory protein-related lipid transfer (START) proteins, two Sec14 like proteins, and one protein of relevant evolutionary and lymphoid interest (PRELI) domain containing protein] **(Figure 2)** with different E-value (as E-value provided by AmoebaDB). We further verify each of these twenty two potential LTP homologs for possessing of the indicated lipid transfer domain (LTD) [for instances, STARD, ORD, Sec14 and PRELI domains] by NCBI conserved domain search analysis and position of individual domain in each homolog were defined. Twenty two potential LTP homologs were then compared with LTP homologs previously studied in other eukaryotes, in particular, human and Saccharomyces sp., and grouped based on their domain organization **(Figure 2)**. Note that information on non-vesicular lipid transport are mostly available in human (Curwin and McMaster, 2008), yeast (Bankaitis et al., 1989, 1990; Im et al., 2005), plant (Li et al., 2016) and little in Plasmodium sp. (Van Ooij et al., 2013; Hill et al., 2016)]. Mutual amino acid identities among LTP homologs from E. histolytica and their counterparts in human and yeast (as per the classification in **Figure 2**), calculated with ClustalW, revealed that E. histolytica LTPs are significantly divergent from those of human and other eukaryotes (Data not shown in this review). A summary table indicates the repertoire of LTP homologs will be analyzed in E. histolytica **(Table 1)**.

# ORPs

E. histolytica possess four ORP homologs (EHI\_086250, EHI\_050360, EHI\_023470, and EHI\_074110), which could mediate sterol transport between the ER and other organelles **(Figure 2)**. Among four ORP homologs, two of them (EHI\_023470 and EHI\_074110) contain only ORP-related domain (ORD) and lacks any potential membrane anchoring domain, which indicates that they are likely localized in the cytosol. The two other ORP homologs (EHI\_086250 and EHI\_050360) possess the PH domain and the diphenylalaninein-an-acidic-tract (FFAT) motif in the amino terminus and a single TM domain in the carboxyl terminus **(Figure 2)**. PH domain is known to interact with specific PIs such as PtdIns4P found on the Golgi membrane and the PM (Helle et al., 2013), while FFAT is known to bind to ER proteins (Helle et al., 2013). Thus, these two TM domain-containing ORP candidates likely act to form a potential MCS between the ER and the Golgi and/or the ER and the PM. They potentially bring these two organelle membranes close enough to facilitate lipid transport as observed in OSBP-mediated sterol transport (Raychaudhuri et al., 2006; Schulz and Prinz, 2007; Raychaudhuri and Prinz, 2010; Olkkonen, 2015) and CERT-mediated ceramide transport at the ER–Golgi MCSs in mammals (Hanada, 2010; Lev, 2010; Kumagai et al., 2014).

# START Domain Containing Proteins

Among 15 START domain containing proteins, most of them (except EHI\_155260) contain only START domain, which likely indicate their cytosolic localization **(Figure 2)**. Only one protein (EHI\_155260) has an additional BUD13 domain **(Figure 2),** which was previously reported to have a role in mRNA splicing and retention in higher eukaryotes (Scherrer and Spingola, 2006). EHI\_155260 also has the potential NLS and NES, which indicates that it can potentially relocate between the nucleus and the cytoplasm, depending upon physiological and environmental conditions. This domain organization of EHI\_155260 is unique to E. histolytica. Low mutual similarity among E. histolytica START protein homologs (Data not shown in this review) indicate their diverse ligand specificities and functions, as observed in higher eukaryotes (Alpy and Tomasetto, 2005).

# Sec14s

E. histolytica has two potential Sec14 homologs (EHI\_146930 and EHI\_158090), which also possess Ras GTPase activating protein (Ras-GAP) or Rho-GAP, respectively **(Figure 2)**. Sec14 protein was first identified in budding yeast, essential for the transport of secretory proteins from the Golgi complex (Mousley et al., 2006; Sirokmány et al., 2006; Curwin et al., 2009). Sec14 homologs possessing Ras- GAP or Rho-GAP domain has also been identified in human (Curwin and McMaster, 2008). p50Rho-GAP/ARH-GAP1 (Sec14 homolog with Rho-GAP domain) **(Figure 2)** is present on the endosomal membrane, where it co-localizes with internalized transferrin receptor (Sirokmány et al., 2006). The Sec14 domain of p50Rho-GAP is essential for its endosomal targeting. p50Rho-GAP forms in vivo a complex with Rab5 and Rab11 on endosomal membranes through its Sec14 domain. Thus, Sec14 domain, which was previously known as a phospholipid binding module, mediates proteinprotein interactions with Rab and Rho-GTPases and regulates receptor-mediated endocytosis (Sirokmány et al., 2006). E. histolytica Sec14 homolog with Rho-GAP domain (EHI\_158090) could potentially play a similar function in endocytic and trogo/phagocytic processes.

# PRELI Domain Containing Proteins

The E. histolytica genome contains a single PRELI-like domain containing protein (EHI\_143630) **(Figure 2)**, which may be involved in lipid homeostasis of its highly divergent mitochondrion-related organelle, as previously described for aerobic mitochondria from higher eukaryotes (Miliara et al., 2015; Tatsuta and Langer, 2016). E. histolytica possesses a highly divergent form of the mitochondrion called mitosome, which is unique in its content and function to Entamoeba (Makiuchi and Nozaki, 2014). It is mainly involved in sulfate activation, and important for parasite growth and differentiation (Mi-ichi et al., 2009, 2015). Since, E. histolytica do not possess any canonical

mitochondria (Makiuchi and Nozaki, 2014), it will be interesting to study whether function of PRELI homolog in E. histolytica (EHI\_143630) is evolutionary conserved or it has a distinct cellular function unique to this protozoan parasite.

# Other Proteins Known to Be Involved in Lipid Transfer in Other Organisms, but Missing in E. histolytica

E. histolytica lacks some of LTP homologs known to be present and functional in other eukaryotes **(Figure 2)**. E. histolytica has no homologs for human PITPs, similar as Saccharomyces sp. (Nile et al., 2010). The Saccharomyces genomes contain several Sec14 homologs that function as PITPs (Phillips et al., 2006). E. histolytica possesses a panel of START domain protein homologs, some of which could potentially function as PITP as reported in Plasmodium falciparum (Van Ooij et al., 2013; Hill et al., 2016). E. histolytica also lacks a homolog of eukaryotic synaptotagmin-like, mitochondrial and PH domain (SMP) containing proteins (Helle et al., 2013). In yeast SMP-containing homologs (Nvj2, Tcb1, Tcb2 and Tcb3) are localized at various MCSs, which indicates their common function at MCSs (Helle et al., 2013). Moreover, the ER–mitochondria encounter structure (ERMES) components (the ER protein, Mmm1, the cytosolic protein, Mdm12, and the OMM protein, Mdm34) also possess SMP domain (Helle et al., 2013). Since the E. histolytica genome encodes none of ERMES components and other SMP homologs **(Figure 2)**, it is plausible that some LTP homologs could potentially function at the MCSs. It is known in other organisms that some LTP homologs with unique domain organization are unique to a particular organism. For instance, Osh3 (of ORP sub-class), Sec14, Pdr16, Pdr 17, Sfh1 and Sfh5 (of Sec14 sub-class) are unique to yeast **(Figure 2)**. Similarly, EHI\_155260 (of START sub-class), EHI\_086250 and EHI\_050360 (of ORP sub-class) are unique to E. histolytica **(Figure 2)**.

# mRNA Expression of LTP Homologs in *E. histolytica* HM-1:IMSS

Relative steady-state levels of mRNA expression of a panel of 22 potential LTP homologs from E. histolytica (15 START proteins, 4 ORPs, 2 Sec14s and 1 PRELI domain proteins) were investigated using data available at AmoebaDB (Hon et al., 2013). Three members of E. histolytica START protein homologs showed higher mRNA expression in HM-1:IMSS compared to other LTP candidates (in a descending order of EHI\_080260, EHI\_161070, and EHI\_173480) **(Figure 3)**. Three START protein homologs (EHI\_182510, EHI\_155260 and EHI\_130730) showed very low levels of expression **(Figure 3)**. Interestingly, the reptilian sibling Entamoeba species, E. invadens, has two EHI\_155260 homologs (EIN\_257190, EIN\_107840). Similar patterns of upregulation of these two gene transcripts were observed during encystation


Twenty two potential LTP homologs of E. histolytica, their accession numbers and domain structure. Refer to *Figure 2* for classification and domain organization. <sup>a</sup>LTP homolog in E. histolytica as per AmoebaDB database.

<sup>b</sup>AmoebaDB ID.

<sup>c</sup>Domain organization based on NCBI conserved domain search analysis.

e steroidogenic acute regulatory protein-related lipid transfer (START) domain.

<sup>f</sup>BUD (BurrH domain) 13 domain.

<sup>g</sup>oxysterol-binding-protein-related (ORP) domain.

<sup>h</sup>pleckstrin-homology (PH) domain.

<sup>i</sup>diphenylalanine-in-an-acidic-tract (FFAT) motif.

<sup>j</sup>Transmembrane (TM) domain.

<sup>k</sup>Ras-GAP domain.

<sup>l</sup>Sec14 domain.

<sup>m</sup>Rho-GAP domain.

<sup>n</sup>protein of relevant evolutionary and lymphoid interest (PRELI) domain.

(data not shown) (De Cádiz et al., 2013), which indicates their potential roles in cell differentiation, as previously shown for LTP during somatic embryogenesis in Arabidopsis thaliana (Potocka et al., 2012). Among 4 ORP candidates, EHI\_023470 and EHI\_074110 are more highly expressed than EHI\_086250 and EHI\_050360 **(Figure 3)**. Two Sec14 homologs (EHI\_146930 and EHI\_158090) showed relatively low expression levels among all identified LTP candidates **(Figure 3)**.

# BIOLOGICAL SIGNIFICANCE OF LIPID TRANSFER IN *E. histolytica*

# Previous Reports on Lipids and Their Trafficking in *E. histolytica*

The structure and PM components of E. histolytica were studied previously (Espinosa-Cantellano and Martínez-Palomo, 1991), which has identified several surface antigens, adherence proteins, amoebic enzymes (for instances, collagenase, phospholipase A, neuraminidase, cysteine proteases) and fibronectin receptor. PM of E. histolytica, enriched in cholesterol, phosphatidylethanolamine (PE), ceramide aminoethyl phosphonate (CAEP) was also reported (Espinosa-Cantellano and Martínez-Palomo, 1991). Previous studies on lipids in E. histolytica mostly suggested their roles in parasite growth, proliferation, differentiation, and virulence. It was shown that PC-cholesterol liposomes enabled in vitro cultured trophozoites to retain their virulence-associated biological functions such as endocytosis, erythrophagocytosis, expression of surface molecules, protease activity, and liver abscess formation in hamsters. This indicates the contribution of lipids to parasite virulence (Serrano-Luna et al., 2010). Both the unique phospholipid compositions and the high cholesterol content in the amoebic membranes were shown to protect the parasite from self-destruction by its own pore-forming toxins (Andrä et al., 2004). Castellanos-Castro et al. has recently shown that lysobisphosphatidic acid is generally involved in endocytosis (i.e., pinocytosis and erythrophagocytosis). Lysobisphosphatidic acid was demonstrated to be localized in Rab7A positive vesicles in quiescent (non-phagoctytic) conditions and during a late phase of erythrophagocytosis (Castellanos-Castro et al., 2015). It was shown that lipopeptidophosphoglycan (EhLPPG) is also involved in adherence of E. histolytica trophozoites to intestinal epithelial cells, similar to other adhesive molecules like, Gal/GalNAc lectin, serine rich E. histolytica proteins (SREHP), and lysine glutamic acid rich protein 1 (KERP1) (Stanley et al., 1990; Dodson et al., 1999; Lauwaet et al., 2004; Seigneur et al., 2005). EhLPPG is also recognized by both the innate and the adaptive immune systems and stimulates cytokine production from human monocytes, macrophages, and dendritic cells (Wong-Baeza et al., 2010). EhLPPG induces in vitro formation of human neutrophil extracellular traps (NETs) (Ávila et al., 2016). Sulfolipids are one of the terminally synthesized bio-molecules of sulfur metabolism, shown to play an important role in trophozoites proliferation and differentiation processes (Mi-Ichi et al., 2017). For instance, fatty alcohol disulfates plays a crucial role in trophozoites proliferation (Mi-Ichi et al., 2017). Cholesteryl sulfate, another sulfolipid plays a central role in encystation, a differentiation process from the motile trophozoites to the dormant cysts (Mi-Ichi et al., 2017). Furthermore, sulfur metabolism, in which sulfolipids are generated, is not conserved in other free-living amoebae, indicating a causal relationship of sulfur metabolism with parasitism (Mi-Ichi et al., 2017). Lysophosphatidylinositol was shown to stimulate natural killer T (NKT) cells and to induce selective production of IFN-γ but not IL-4 in a CD1-d restricted manner in murine systems (Aiba et al., 2016).The

localization of PI (3,4,5)-trisphosphate [PI(3,4,5)P3 or PIP(3)] in E. histolytica during various endocytic processes was studied using glutathione S-transferase (GST)- and green fluorescent protein (GFP)-labeled PH domains as lipid biosensors (Byekova et al., 2010). PIP(3) specific biosensor was accumulated at extending pseudopods and also localized in the phagocytic cup during erythrophagocytosis. However, no such localization of the biosensor was observed in pinocytic compartment during pinocytosis. E. histolytica maintains a high steady state level of PIP(3) in its PM irrespective of serum concentration (Byekova et al., 2010).

However, non-vesicular lipid transport machinery and the function of LTPs remained largely unexplored in E. histolytica. There are two previous reports where the lipid trafficking in E. histolytica was described (Pina-Vázquez et al., 2014; Bolaños et al., 2016). Piña- Vázquez et al. identified a START domain containing protein in E. histolytica (EHI\_110720, also present in our list of E. histolytica LTP homologs) and named it as E. histolytica phosphatidylcholine transfer protein-like (EhPCTP-L). They identified EhPCTP-L by virtue of interaction with anti-chicken embryo caveolin-1 monoclonal antibody, which indicates their potential role in caveola-mediated endocytosis. EhPCTP-L mainly binds to anionic phospholipids phosphatidylserine (PS) and PA, and is localized to the PM and the cytosol (Pina-Vázquez et al., 2014). However, the essential biochemical characteristics as LTPs, i.e., lipid transport activity, was not reported in this study. START protein homolog (EHI\_178560) of E. histolytica is essential for parasite growth as observed in the previous study by Solis et al. Double stranded RNA (dsRNA) mediated silencing of EHI\_178560 causes growth retardation of E. histolytica trophozoites (Solis et al., 2009). Mfotie et al. also reported that plant derivatives that showed anti-amoebic activity also caused the repression of gene expression of another START homolog (EHI\_161070) of E. histolytica (Mfotie Njoya et al., 2014). Bolaños et al. recently reported that E. histolytica NPC1 (EhNPC1) and EhNPC2 proteins responsible for the trafficking of exogenous cholesterol in E. histolytica trophozoites and also influence the phagocytosis process (Bolaños et al., 2016). However, these previous studies only partially characterized the roles of LTPs in the trafficking of ingested lipids in Entamoeba. The complex network of intra-cellular lipid transport machinery mediated by diverse LTP homologs in E. histolytica remains elusive.

## Predicted Roles of LTPs in *E. histolytica* Role of E. histolytica LTPs in Phagocytosis, Trogocytosis, Endocytosis, Signal Transduction, and Lipid Presentation

E. histolytica is highly capable of engulfment of the host cells and microorganisms by two distinct processes. In phagocytosis, the parasite engulfs the dead host cells and bacteria as whole, while in trogocytosis the parasite nibbles parts of the live host cells (Nakada-Tsukui et al., 2009; Ralston et al., 2014; Somlata et al., 2017). These two distinct cellular processes are initiated by being triggered by different ligands (on dead/live host cells and microorganisms), and likely activate similar but different cascades of events (Somlata et al., 2017). It is possible that different LTPs may be selectively involved in phagocytosis and trogocytosis. Similarly, Entamoeba also depends on receptormediated endocytosis/pinocytosis for the transport of nutrients from the extracellular environment (Avalos-Padilla et al., 2015). Such receptor-ligand mediated signaling process is often initiated by PIs (Nakada-Tsukui et al., 2009; Somlata et al., 2017).

E. histolytica possess several classes of PI metabolizing enzymes such as: PtdIns 4-kinase (EHI\_148700) and PtdIns4P 5-kinases (EHI\_153770, EHI\_049480), and a panel of PI phosphatases (14 isotypes of PI 3-phosphatase, 6 isotypes of PI 5- phosphatase, and 3 isotypes of non-specific PI phosphatase). Identification of a repertoire of PI kinases and phosphatases enforces the notion that E. histolytica LTPs also participate in the constant replenishment of PIs at all cellular compartments at various physiological conditions. E. histolytica also possesses 5 putative DAG kinases (data not shown), thus is able to produce PA from DAG. This is consistent with the hypothesis that the reciprocal transport of PtdIns and PA between the ER and the PM (Cockcroft and Garner, 2013) occurs and is mediated by some members of LTPs as lipid presenting proteins.

# Role of E. histolytica LTPs in the Secretion of Hydrolytic Enzymes and Regulation of Vesicular Trafficking

Trafficking and secretion of lysosomal hyodrolases such as CPs contributes to both cytolysis of host tissues and degradation of internalized host cells and microorganisms (Nakada-Tsukui et al., 2005; Mitra et al., 2007). Thus, vesicular trafficking that regulates intracellular CP transport plays a pivotal role in virulence and parasitism of E. histolytica. CP trafficking and secretion are regulated via cysteine protease binding family protein (CPBF) 1 (Furukawa et al., 2012; Nakada-Tsukui et al., 2012; Marumo et al., 2014), Rab GTPases (Saito-Nakano et al., 2004, 2005; Mitra et al., 2007; Hanadate et al., 2016), the retromer-like complex (Nakada-Tsukui et al., 2005), and intrinsic CP inhibitors (Sato et al., 2006), and also presumably by priming and docking of CP-containing vesicles with the exo-cyst complex that tethers at the site of exocytosis on the PM (Nakada-Tsukui et al., 2005). As LTPs can function as lipid sensors to regulate Golgi-mediated (or post-Golgi) vesicular trafficking, as above explained for Sec14, it is possible that two Sec14 homologs (EHI\_146930 and EHI\_158090) that also possess either Rho-GAP or Ras-GAP domain **(Figure 2)** can regulate the Golgimediated secretory function, in a similar mechanism as described previously (Curwin et al., 2009). The four ORP homologs **(Figure 2)** may interact with Rab GTPase and control trafficking of the transport vesicles as described for ORP1L in higher eukaryotes (Johansson et al., 2007).

# Role of E. histolytica LTPs in Nuclear Lipid Transport and Signaling

The nucleus plays indispensable roles in cell proliferation, differentiation, and stress management (Shah et al., 2013). Stage conversion between the two forms in the life cycle requires remarkable alterations in cellular components, metabolism, transcriptional, and post- transcriptional/translational regulations of gene expression, and involves a complex and dynamic signaling events induced by extracellular stimuli (Mittal et al., 2008; Vicente et al., 2009; De Cádiz et al., 2013). Moreover, E. histolytica is also harassed by a wide variety of environmental and host-derived stresses, such as fluctuation in glucose concentrations, changes in pH, pO2, temperature, and attack by oxidative and nitrosative stresses from neutrophils and macrophages (Husain et al, 2010; Husain et al., 2012; Nagaraja and Ankri, 2018). E. histolytica senses extracellular stress and accordingly makes necessary amendment in its physiology and metabolism for survival and transmission.

Although the lipid transport, metabolism and signaling in the nucleus and the roles of LTPs in nuclear associated functions are largely unknown, it is conceivable that one START domain containing protein (EHI\_155260), which possess BUD13 domain and potential NLS and NES, is involved in nuclear lipid transport, more specifically shuttling between the nucleus and the cytoplasm, and signaling, and thus may be important for growth, stage conversion, and/or evasion from various stresses. In addition, some of the PI and PA metabolizing enzymes in E. histolytica (data not shown here) contain putative NLS and NES, indicating their potential roles in nuclear lipid homeostasis by nuclear-cytoplasmic shuttling, as well described in higher eukaryotes (Davis et al., 2015). The E. histolytica genome encodes a few PI-binding downstream effectors including a plant homeodomain (PHD) finger-containing protein (EHI\_138970), which also contains NLS. Thus, it is conceivable that a panel of these proteins coordinately function in the nucleus. As mentioned above (3.2), E. invadens apparently possesses two homologs (EIN\_257190, EIN\_107840) of the BUD13/NLS/NEScontaining START domain containing protein (a single protein in E. histolytica, EHI\_155260). These E. invadens genes showed upregulation of gene expression during encystation (De Cádiz et al., 2013), which indicates their potential role in the nucleus during differentiation, as observed in other eukaryotes (Potocka et al., 2012).

### Role of E. histolytica LTPs in Motility, Adherence, and Cytoskeletal Reorganization

It is conceivable that two SEC14 homologs containing Rho-GAP and Ras-GAP (EHI\_146930 and EHI\_158090, respectively) are involved in cytoskeletal reorganization associated with cell motility and adherence to the host cells and microorganisms. The genome of E. histolytica also contains four ORP (EHI\_086250, EHI\_050360, EHI\_023470, and EHI\_074110), 15 START domain homologs other than Sec14 **(Figure 2)**. ORP homologs might be associated with cytoskeleton organization, cell adhesion, and motility, as reported in higher eukaryotes (Goldfinger et al., 2007) and previously discussed in section Cytoskeleton organization, adhesion, and motility Other than ORP, a few START domain homologs could also be associated with cytoskeletal reorganization and migration, as previously reported in a cancer cell line (Alpy and Tomasetto, 2005). Adhesion with the host gut epithelia and migration through the tight junction between host cells are two key pathogenic processes presented by E. histolytica trophozoites (Tavares et al., 2005; Franco-Barraza et al., 2006), which also involve the receptor-ligand mediated signaling cascades, in which PIs play indispensable roles.

# CONCLUDING REMARKS

We have discovered a panel of conserved and lineage-specific LTPs in E. histolytica by domain-based survey of LTP homologs in AmoebaDB. The E. histolytica genome possess single PRELI domain containing protein (EHI\_143630). It is worth investigating whether this PRELI domain containing protein is involved in mitosomal transport. E. histolytica possess a single START domain containing protein (EHI\_155260) with potential NLS and NES, indicating its potential role as a nuclear lipid transporter and its ability of nucleocytoplasmic shuttling depending upon physiological conditions. Upregulation of gene expression of its two homologs (EIN\_257190, EIN\_107840) in E. invadens during encystation (De Cádiz et al., 2013), indicates its potential role in cell differentiation as well as disease transmission. The E. histolytica genome possess a repertoire of START domain containing proteins. Some of them can function as potential PITP similarly as in Plasmodium sp. (Van Ooij et al., 2013; Hill et al., 2016). This is also conceivable because phosphoinositides are the essential signaling phospholipids for E. histolytica, which depends greatly on receptor-ligand mediated signaling processes. On the other hand, E. histolytica lacks canonical PITP homologs from higher eukaryotes. Furthermore, E. histolytica lacks any homologs of SMP containing proteins, which are involved in the organization of various MCSs and ERMES complex (Helle et al., 2013). The absence of canonical mitochondria, ER, and Golgi apparatus, in the forms found in higher eukaryotes, could possibly justify this observation. The identified LTP repertoire encompassing ORP, START, SEC14, and PRELI present in E. histolytica

# REFERENCES


ensures the complexity and biological significance of LTPs in a variety of cellular processes: lipid transport between cellular compartments, adherence, phagocytosis, trogocytosis, endocytosis, signal transduction, vesicular traffic, cytoskeletal reorganization, nuclear regulation, lipid presentation, and sensing. Further studies are needed to elucidate the role of individual LTPs on the biology and pathogenesis of this parasite. Once E. histolytica-specific LTPs and lipid transfer mechanisms are identified, they may potentially provide a novel drug target against this medically important parasite. Since most of existent chemical interventions against lipid signaling pathways target particular lipid metabolizing enzymes (Nile et al., 2014; Khan et al., 2016), such intervention often exerts specific and limited overall effects to the eukaryotic cells, e.g., cancer cells. In addition, parasitic organisms often have highly adaptable nature and a bypass mechanism to overcome the effects caused by various chemical insults, results in the generation of drug resistance. Thus, chemical inhibition of parasite-specific LTPs might be a reasonable solution to the problem because LTPs are involved in a wide range of multiple cellular processes described above.

# AUTHOR CONTRIBUTIONS

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

# ACKNOWLEDGMENTS

We acknowledge The Tokyo Biochemical Research Foundation (TBRF) for providing TBRF postdoctoral fellowship for foreign researchers for KD (TBRF-RF17-105). This work was supported in part by Grants-in-Aid from TBRF (TBRF-RF17-105), Grantin- Aid for Challenging Exploratory Research (17K19416) and Grant-in-Aid for Scientific Research (B) (18H02650) from the Japan Society for the Promotion of Science, and a grant for Research Program on Emerging and Re-emerging Infectious Diseases from the Japan Agency for Medical Research and Development (AMED) (JP17fk0108119) to TN.

associated protein involved in pinocytosis and phagocytosis of Entamoeaba histolytica. PLoS Pathog. 11:e1005079. doi: 10.1371/journal.ppat.10 05079


stimulated during invasive behavior. Cell Motil. Cytoskeleton 63, 117–131. doi: 10.1002/cm.20105


lipin1 function in lipid metabolism and adipogenesis. MBoC 21, 3171–3181. doi: 10.1091/mbc.E10-01-0073


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Das and Nozaki. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Life and Death of mRNA Molecules in *Entamoeba histolytica*

Jesús Valdés-Flores <sup>1</sup> , Itzel López-Rosas <sup>2</sup> , César López-Camarillo<sup>3</sup> , Esther Ramírez-Moreno<sup>4</sup> , Juan D. Ospina-Villa4† and Laurence A. Marchat <sup>4</sup> \*

<sup>1</sup> Departamento de Bioquímica, CINVESTAV, Ciudad de Mexico, Mexico City, Mexico, <sup>2</sup> CONACyT Research Fellow – Colegio de Postgraduados Campus Campeche, Campeche, Mexico, <sup>3</sup> Posgrado en Ciencias Genómicas, Universidad Autónoma de la Ciudad de México, Ciudad de Mexico, Mexico City, Mexico, <sup>4</sup> Escuela Nacional de Medicina y Homeopatía, Instituto Politécnico Nacional, Ciudad de Mexico, Mexico City, Mexico

In eukaryotic cells, the life cycle of mRNA molecules is modulated in response to environmental signals and cell-cell communication in order to support cellular homeostasis. Capping, splicing and polyadenylation in the nucleus lead to the formation of transcripts that are suitable for translation in cytoplasm, until mRNA decay occurs in P-bodies. Although pre-mRNA processing and degradation mechanisms have usually been studied separately, they occur simultaneously and in a coordinated manner through protein-protein interactions, maintaining the integrity of gene expression. In the past few years, the availability of the genome sequence of Entamoeba histolytica, the protozoan parasite responsible for human amoebiasis, coupled to the development of the so-called "omics" technologies provided new opportunities for the study of mRNA processing and turnover in this pathogen. Here, we review the current knowledge about the molecular basis for splicing, 3′ end formation and mRNA degradation in amoeba, which suggest the conservation of events related to mRNA life throughout evolution. We also present the functional characterization of some key proteins and describe some interactions that indicate the relevance of cooperative regulatory events for gene expression in this human parasite.

#### *Edited by:*

Mario Alberto Rodriguez, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional (CINVESTAV-IPN), Mexico

#### *Reviewed by:*

Mark R. Macbeth, Butler University, United States Michael G. Sehorn, Clemson University, United States Chi Ming Wong, University of Hong Kong, Hong Kong

### *\*Correspondence:*

Laurence A. Marchat lmarchat@gmail.com; lmarchat@ipn.mx

### *†Present Address:*

Juan D. Ospina-Villa, Grupo Biología y Control de Enfermedades Infecciosas, Universidad de Antioquia, Medellín, Colombia

> *Received:* 03 April 2018 *Accepted:* 28 May 2018 *Published:* 19 June 2018

#### *Citation:*

Valdés-Flores J, López-Rosas I, López-Camarillo C, Ramírez-Moreno E, Ospina-Villa JD and Marchat LA (2018) Life and Death of mRNA Molecules in Entamoeba histolytica. Front. Cell. Infect. Microbiol. 8:199. doi: 10.3389/fcimb.2018.00199 Keywords: *Entamoeba*, mRNA decay, mRNA processing, P-bodies, polyadenylation, protozoan parasite, splicing

# INTRODUCTION

The metabolism of messenger RNA (mRNA) is a complex process that is essential for gene expression regulation and mRNA turnover in response to environmental signals and cell-cell communication in eukaryotic cells. During pre-mRNA synthesis by RNA polymerase II (RNA Pol II) in the nucleus, they are modified to generate mature transcripts that can be exported to the cytoplasm and translated to proteins. First, the 5′ end of nascent mRNA is capped by a 7-methyl guanosine linked by a 5′ -5′ triphosphate bridge to the first nucleoside of the transcript (capping). These reactions are catalyzed by three enzymes: RNA triphosphatase, guanylyltransferase, and RNA (guanine-7-)-methyltransferase (RNMT) (Cowling, 2010). Then, introns are removed and exons are ligated by the catalytic activity of the spliceosome components that include five small nuclear RNA (snRNAs), namely U1, U2, U4, U5, and U6, and small nuclear ribonucleic proteins (snRNPs) (splicing) (Shi, 2017). Finally, a phosphodiester bond is hydrolyzed at the 3′ end of mRNA and a poly(A) tail is added by the coordinated activity of a large set of polyadenylation factors that recognize specific motifs in RNA 3′ untranslated region (3′UTR) (cleavage/polyadenylation) (Xiang et al., 2014). After translation, the elimination of mRNA molecules is necessary to ensure proper course of gene expression and prevent the accumulation of transcripts (Christie et al., 2011). Pathways of mRNA decay depend on the formation of RNAprotein complexes in microscopically detectable cytoplasmic structures, called processing bodies (P-bodies) (Sheth and Parker, 2003), in which mRNAs are translationally repressed (silenced) or degraded; their re-incorporation into ribosomes is also possible (Eulalio et al., 2007). Transcript decay involves 3 ′ end deadenylation by CAF1 and CCR4/NOT1–5 complex (or by PARN, PAN2 and PAN3 deadenylases) followed by 5 ′ end decapping by DCP1–DCP2 complex and Lsm1–7 proteins, and 5′ -3′ digestion by exonuclease XRN1; alternatively, deadenylated transcripts can be degraded from the 5′ end by the exosome complex, while the scavenger-decapping DCP enzyme hydrolyzes the remaining cap structure (Łabno et al., 2016). During translation, aberrant mRNAs with premature termination codons can be detected and eliminated through the nonsense-mediated decay (NMD) pathway (Rebbapragada and Lykke-Andersen, 2009).

Although pre-mRNA processing reactions have usually been studied separately, they occur co-transcriptionally, simultaneously and in a coordinated manner. Moreover, a large set of data has shown that they are interconnected with transcription, translation, and mRNA degradation; proteinprotein interactions establish a functional link between the different molecular machineries and promote reciprocal regulation events to maintain the integrity of gene expression. Consequently, each of these processes plays a major role throughout the life cycle of mRNA. Thus, in addition to protect mRNA from 5′ to 3′ exonuclease cleavage, the m7G cap interacts with the cap-binding complex (CBC), which regulates spliceosome assembly, transcription termination, 3 ′ end processing, RNA export, and NMD in the nucleus. In the cytoplasm, CBC recruits eIF4G, RNA helicase eIF4A, and other proteins to promote translation initiation. Moreover, eIF4G interacts with poly(A) binding protein PABP1 bound to the poly(A) tail to create a mRNA pseudo-circularization and enhance the processivity of the ribosome. Furthermore, it has been recently demonstrated that 2′O methylated cap (cap 1) acts as a signature of self RNA molecules (Ramanathan et al., 2016). Several data indicate that U1 snRNP, the more abundant splicing factor, inhibits 3′ end processing. Notably, its interaction with PAP inhibits poly(A) tail synthesis and promotes degradation of U1A pre-mRNA (Gunderson et al., 1994, 1997). Moreover, its binding to the 5′ splice site (5′ ss) of the terminal intron, avoids the use of premature cleavage and polyadenylation to protect the integrity of the transcriptome (Furth et al., 1994). Other data indicate that splicing and 3′ end processing factors may recruit each other and form a stabilized complex on the target pre-mRNA, resulting in reciprocal stimulation of efficiency. Thus, interactions between U2AF65 and CFIm59 (Millevoi et al., 2006), or U1A and CPSF160 (Lutz et al., 1996), enhance the polyadenylation reaction, while CPSF (Kyburz et al., 2006) and PAP (Vagner et al., 2000) stabilize U2AF65 to the terminal intron to stimulate splicing. On the other hand, exon–exon junction complexes (EJC) participate in mRNA degradation, as part of the CBC whose CBP80 component interacts directly with the NMD factor, up-frameshift 1 (UPF1), enhancing the efficiency of this process (Isken and Maquat, 2008). CFIm may bridge 3′ processing with capping through the binding of CFIm25 with CBP20 (Yang et al., 2011).

Until recently, little was known about mRNA metabolism in Entamoeba histolytica, the protozoan responsible for human amoebiasis. The availability of the E. histolytica genome sequence and the development of the so-called "omics" technologies have provided new opportunities for the study of mRNA processing and turnover in this parasite. To our knowledge, capping has not been described in E. histolytica, although preliminary searches in parasite genome database suggest the presence of genes that encode proteins with similarities to human capping enzymes. In this review, we focus on the current knowledge about the molecular basis for splicing, 3' end formation and mRNA degradation, and describe some interactions between these events.

# WHAT IS KNOWN ABOUT SPLICING IN *E. HISTOLYTICA*

# Splicing Factors

There are nearly four thousand introns in the 8333 annotated genes of E. histolytica (Weedall and Hall, 2011), most of them flanked by highly conserved 5′ and 3′ splice sites (ss), GUUUGU and UAG, respectively, but their branch point sequences (BS) lack such degree of conservation (Wilihoeft et al., 2001; Hon et al., 2013). Whereas, no minor U12 introns have been identified in amoeba and most likely neither in the eukaryotic ancestor (Collins and Penny, 2005; Bartschat and Samuelsson, 2010), the majority of the main spliceosome components have been predicted and identified.

Molecular evidence and cloning confirmed the presence of U2, U4, U5, and U6 snRNAs (Miranda et al., 1996; Davis et al., 2007), however no significant homology with eukaryotic U1 snRNAs has led to the conclusion that such small nuclear RNA is absent in Entamoeba (Dávila et al., 2008). Nonetheless bioinformatic analyses predicted the presence of the three U1 snRNP U1-A, U1-C, and U1-70k factors, suggesting that activation of the 5′ ss might be due to direct interaction of snRNP proteins or by U6 snRNA-5′ ss complementarity substitution as demonstrated in other systems (Kandels-Lewis and Seraphin, 1993; Förch et al., 2002; Rhode et al., 2006; Huang et al., 2012). In vivo expression of tag-cloned U1-A and cross-linking immunoprecipitation (CLIP) assays of nuclear proteins coupled to mass spectroscopy allowed the identification of at least 32 splicing factors in trophozoites (**Table 1**), namely U2, U4, and U5 snRNP, integral SmD1, SmD3, and SmF proteins; the U1 snRNP components and auxiliary factors U1-70k and TIA-1/TIAR; the U2 snRNP and related components U2-A', SF3a120, SF3a60/Prp9, SF3b1, SF3b3, U2AF65, and U2AF35; the U5 snRNP components Prp8 and Prp6 [which was previously identified and cloned (Hernandez-Rivas et al., 2000)]; the U6 snRNP integral components LSm2 and LSm5; two alleles TABLE 1 | Comparison of splicing factors in Entamoeba histolytica vs. human and yeast.


<sup>a</sup>AmoebaDB. E. histolytica splicing factors, U snRNAs and components of the post-catalytic/intron lariat spliceosome complexes (PILS in superscript) were described by Miranda et al. (1996), Hernandez-Rivas et al. (2000), Davis et al. (2007), Dávila et al. (2008), Fourmann et al. (2013), and Valdés et al. (2014). Previously undetected additional Entamoeba PILS components (shaded) were identified in the ProteomeXchange repository PXD001080. # indicates that CWC22 is also part of the Prp19C. Transitions of spliceosome complexes remodeling by the respective DExH/D-box helicases are indicated (Liu, 2002; Marchat et al., 2008; Hahn et al., 2012; Wahl and Luhrmann, 2015).

FIGURE 1 | Co-transcriptional pre-mRNA processing in E. histolytica: focus on splicing factors. The model summarizes the data available to date (Valdés et al., 2014). During transcript elongation by RNA polymerase II (RNAPII; purple), Ser2 residues of the few heptapeptide repeats of its carboxy-terminal domain become phosphorylated (PSer2-CTD; maroon circles) and apt to recruit the spliceosomal (salmon triangle) and polyadenylation machineries. The large subunit of the U2 Auxiliary Splicing Factor U2AF65 (of 84 kDa in E. histolytica; orange oval) is a major player in pre-mRNA processing by tethering the spliceosome and the pre-mRNA (light green boxes) to RNAPII. U2AF65 interacts with the RNAPII-PSer2-CTD and with splicing factors conforming the Prp19 Complex (NTC; blue circle). The NTC regulates the formation and progression of essential spliceosome conformations required for the two steps of spacing. Splicing complex E formation occurs when the snRNP U1-A (yellow oval) binds to the 5′ ss (splice site) and the splicing factor TIA-1/TIAR (yellow box) binds to the U-rich sequence just downstream the 5′ ss. Splicing complex E also involves the 3′ ss definition (not shown). When RNAPII releases the 3′ ss from the transcription site, splicing factor 1 binds the branch site at the same time that U2AF65 binds the intron's polypyrimidine tract located between the branch site and the 3′ ss; also simultaneously, the small subunit of U2AF (U2AF35, of 29 kDa in E. histolytica; pale orange circle) recognizes the 3 ′ ss. The interaction of U2AF65 with splicing factor 1 and U2AF65 at the 3′ ss and with the CTD of RNAPII ensures that U2AF65 also tethers the pre-mRNA to RNAPII. Finally, in addition to the previously reported interactions of RNAPII with the polyadenylation complex (vide infra), U1-A directly or indirectly interacts with the splicing complexes B-C, and more importantly with CstF77 (dark green oval), a member of the polyadenylation machinery.

of the U4/U6 di-snRNP component CP6; the U4/U6.U5 trisnRNP components SAD1 and Prp38; and the nineteen complex (NTC) components Prp19, KIAA0560/Aquarius intron-binding spliceosomal factor, DDX5, and Abstrakt/DDX41 (Valdés et al., 2014) (**Figure 1**).

Splicing E (early) complex formation involves 5′ ss recognition by the U1 snRNP (Larson and Hoskins, 2017). However, the less conserved and poorly recognized (weak) 5′ ss are activated by 5'ss-U1-C interactions or when TIA-1/TIAR binds to Utracts localized in front of the 5′ ss (Förch et al., 2002). Only U1-70k and TIA-1/TIAR were detected in the U1-A CLIP assays, therefore the most likely scenario for Entamoeba 5 ′ ss activation involves direct interaction of U1-A/U1-70k with the 5′ ss with the participation of TIA-1/TIAR bound to the U-rich most often spliced Entamoeba 5 ′ ss (GUUUGUUU) (Hon et al., 2013) as described for weak 5′ ss.

Because cross-linking was carried out with UV, the number of factors identified is limited but it represents all complexes formed during spliceosome assembly, first and second steps of splicing, disassembly, turnover, exon junction complex, and mRNA transport. Moreover, the presence of the core protein of the NTC, Prp19, and U2AF65, which interact with the PSer2 CTD of the large subunit of RNA pol II, ensure proper co-transcriptional activation of the spliceosome, splicing catalysis, termination factors recruitment, and extranuclear mRNA transport factors (David et al., 2011; Gu et al., 2013).

Also, DExH/D RNA helicases involved in the proofreading of the sequential steps of spliceosome assembly and catalysis were identified (**Table 1**): Prp5 and Sub2/UAP56 that facilitate E to A (pre-spliceosome) complex transition; p68, and Prp28, that promote transition from A complex to pre-catalytic (B) spliceosome; two alleles of Brr2, and Snu114, required for spliceosome activation (B to Bact); Prp2, that catalytically activates the spliceosome (Bact to B<sup>∗</sup> complex); Prp16, that proof-reads the second step of splicing from the catalytic complex (C); the post-splicing complex helicase Prp22; and three alleles of the disassembly Prp43 helicase (Marchat et al., 2008; Valdés et al., 2014). The fact that all proofreading RNA helicases are present in E. histolytica, contrasts with the multiple splicing products repeatedly detected in deep RNA-seq data (Hon et al., 2013) indicating that additional cues control the splicing and alternative splicing of the majority of Entamoeba introns.

The components of the post-catalytic and of the intron lariat splicing complexes (collectively, PILC) have been recently identified (Fourmann et al., 2013). From our published data and the ProteomeXchange repository PXD001080, we identified the corresponding E. histolytica PILC factors: SF3a60/Prp9, SF3b1, SF3b3, Snu114, Brr2, 220K/Prp8, Prp19, CDC5, Syf1, RBM22, CWC22, MGC13125, and Prp22 (**Table 1**). Furthermore, the HA-tagged CLIP assays (Valdés et al., 2014) allowed the unprecedented identification of numerous messenger ribonucleoparticles factors, among them additional splicingrelated factors, as well as components of the transcription (large subunit of RNA pol II, and various transcription factors), and polyadenylation (EhCstF64) machineries, evidencing the complexity of this co-transcriptional process (**Figure 1**).

Finally, since most Entamoeba pre-mRNAs are monointronic, and bioinformatic predictions and deep RNA-Seq data indicate that intron retention is the main route for alternative splicing (Davis et al., 2007; McGuire et al., 2008; Hon et al., 2013), splicing events impact both proteome expansion and gene expression regulation in this parasite.

# Intron Lariat Debranching Enzyme

Intron lariat debranching enzyme, or Dbr1, is a member of the calcineurin-like metallophosphoesterases (MPEs) superfamily of binuclear metal-ion-center-containing enzymes that hydrolyse phosphomono-, phosphodi-, or phosphotriesters in a metal-dependent manner. From bacteriophages to humans, the MPE domain is found in Mre11/SbcD DNA-repair enzymes, mammalian phosphoprotein phosphatases, acid sphingomyelinases, purple acid phosphatases, nucleotidases, and bacterial cyclic nucleotide phosphodiesterases. Despite this functional diversity, MPEs show a remarkably similar structural fold and active-site architecture composed of five sequence blocks that allow metal coordination in the conserved motif D[X]H[x]nGD[x]nGNH[D/E] [x]nH[x]nGH[X]H (Matange et al., 2015). Alanine scanning assays identified the yeast Dbr1 RNA debranching active-site in vivo and in vitro, which in E. histolytica correspond to residues Cys14, His16, Asp45, Asn90, His91, His180, His230, and His232; in all debranching enzymes in nature, cysteine substitutes an aspartic acid residue in position 14 (Matange et al., 2015; Schwer et al., 2016).

Compared to different Dbr1 orthologs, E. histolytica Dbr1 is the shortest protein due to its truncated C-terminus; i.e., it is ≈50 residues shorter that S. cerevisiae Dbr1. Thanks to its small size, Entamoeba Dbr1 has been able to co-crystalize with synthetic branched RNAs and analogs, providing insights of enzyme-lariat interactions. Initial structural studies confirmed the MPE β metal binding pocket (Montemayor et al., 2014) however there was no metal in the A pocket. Subsequent structures revealed the 2nd metal ion in the A pocket (Clark et al., 2016; Ransey et al., 2017). Importantly, this approach identified within the MPE the lariat recognition loop (LRL) whose recognition module, unique to Dbr1 enzymes, comprises residues Ile132, Tyr133, Glu138, Pro141, Tyr144, and Pro148. Dbr1-analogs co-crystals showed that Dbr1 prefers 2′ purines in the branch, although the pocket could accommodate pyrimidines. In addition, there are few sequence-specific interactions at the BS, confirming recognition of atypical BS. Finally, the interactions between RNA and the LRL are stabilized by secondary contacts between residues 141–144 of the LRL and residues Phe292, Pro293, and Phe337 of the carboxy-terminal domain (CTD) of Dbr1; and intricate hydrogen bonds centered in Arg158 aid to stabilize further the conformation of the LRL (Montemayor et al., 2014). The structural data was confirmed by Dbr1 activity in vivo. Whereas, E. histolytica wild type Dbr1 was able to complement Saccharomyces cerevisiae Dbr1-deletant strains relieving intron lariat accumulation, none of the constructs carrying Cys14Ala/Ser substitutions or 141-146Ala substitutions, or CTD or LRL deletions relieved intron lariat accumulation (Montemayor et al., 2014). The presence of E. histolytica Dbr1 in intron large post-spliceosomal complexes along with U2, U5 and U6 snRNP components, and the proteins Ntr1/TFIP11 and Prp43 (Yoshimoto et al., 2009) or the Drn1/Ygr093w protein that transiently binds Dbr1 to post-spliceosomal complexes is still unproven (Garrey et al., 2014), although our previous findings point to this possibility (Valdés et al., 2014).

# UNDERSTANDING POLYADENYLATION IN *E. HISTOLYTICA*

# Cis-Elements for Pre-mRNA 3′ End Formation

One of the first reports about mRNA polyadenylation in E. histolytica was published in 1993 and describes the existence of a putative polyadenylation motif TAATT and a 12 pyrimidine stretch in the 3′UTR of parasite genes (Bruchhaus et al., 1993). Then, other groups showed that alternative polyadenylation sites and poly(A) tail size represent efficient postranscriptional mechanisms for gene expression regulation (Urban et al., 1996; López-Camarillo et al., 2003). But the publication of the first version of the E. histolytica genome sequence in 2005 (assembly of

∼23 Mb that predicted 9938 coding genes comprising 49% region of the genome) (Loftus et al., 2005) represented the critical step to identify motifs in the mRNA 3′UTR and the polyadenylation machinery in this parasite.

A small-scale in silico analysis of cDNA and genomic sequences revealed that E. histolytica 3 ′ UTRs contain three conserved motifs: (i) the consensus UA(A/U)UU polyadenylation signal or variants located 10–30 nt upstream the poly(A) site, (ii) the U-rich tract located 1–30 nt upstream the poly(A) site, and (iii) a U-rich element located 3–30 nt downstream the poly(A) site (López-Camarillo et al., 2005). Computational examination of a larger number of cDNA and genomic sequences confirmed this molecular array and suggested the presence of an additional distal A-rich element (**Figure 2**) (Zamorano et al., 2008). Study of the alternative usage of poly(A) sites using RNA-Seq indicated that microheterogeneity in poly(A) sites is likely to be stochastic in E. histolytica and only a small fraction of alternative polyadenylation isoforms appeared to be genuine (Hon et al., 2013). Interestingly, genes with alternative poly(A) sites may have a large impact on global gene expression in E. histolytica since most of them participate in DNA condensation, DNA binding, translation, splicing, mRNA binding, protein folding and protein transport; other genes are related to signaling, oxidation/reduction, calcium ion binding, cell cycle, and intracellular transport. Indeed, the upstream shift in poly(A) site selection resulting from the silencing of the polyadenylation factor EhCFIm25, was confirmed for thioredoxin and 60S ribosomal protein L7 transcripts and related to specific phenotypical changes and parasite death (Ospina-Villa et al., 2017).

# Polyadenylation Factors

Analyses of the 9938 coding genes predicted in the first version of the genome indicated that E. histolytica has genes that encode proteins with homology to the majority of polyadenylation factors described in human and yeast (**Table 2**): the cleavage and specificity factor (CPSF160, 100, 73, and 30), the cleavage stimulating factor (CstF77, 64, and 50), the 25 kDa subunit of the cleavage factor Im (CFIm) and both C1P1 and PCF11 subunits of CFIIm, as well as FIP1, poly(A) polymerase (PAP), poly(A) binding protein (PABP), RBBP6 (Mpe1 in yeast), WDR33 (Psf2 in yeast), PNAS-120 (Ssu72 in yeast), and PC4 (Sub1 in yeast) (**Figure 2**) (López-Camarillo et al., 2005, 2014). In human cells, WDR33 and CPSF30 are the CPSF subunits that binds the polyadenylation signal (Chan et al., 2014; Schonemann et al., 2014). CPSF73 is the endonuclease responsible for RNA cleavage (Mandel et al., 2006). CstF77 interacts with CPSF160, promoting their cooperative RNA binding during the assembly process (Murthy and Manley, 1995). CFIm25 regulates the selection of distal poly(A) sites, contributes to the recruitment of polyadenylation factors and is necessary for poly(A) tail synthesis (Brown and Gilmartin, 2003; Kubo et al., 2006). Although CFIIm subunits are the less characterized polyadenylation factors, evidence suggest that Pcf11 is required for degradation of the 3′ product following cleavage (West and Proudfoot, 2008), while Clp1 interacts with both CPSF and CFIm and likely tethers them to CFIIm (de Vries et al., 2000). PAP is responsible



<sup>a</sup>UniProtKB; <sup>b</sup>AmoebaDB. Data about parasite proteins were obtained from (López-Camarillo et al., 2005, 2014).

for the addition of the polyadenosine tail (Raabe et al., 1991) and its activity is accelerated by PABP (Wahle, 1991). RBBP6 associates with other core polyadenylation factors through its unusual ubiquitin-like domain and modulates expression of mRNAs with AU-rich 3' UTR (Di Giammartino et al., 2014). PC4 (Sub1 in yeast) associates with the polyadenylation factor CstF64 to modulate transcription termination and polyadenylation initiation (Calvo and Manley, 2001, 2003).

The presence of these proteins in E. histolytica suggests that 3 ′ end processing of parasite mRNA could be performed as it has been described in other eukaryotic cells. Accordingly, EhPAP has the conserved PAP central catalytic domain with the three invariant aspartate residues involved in nucleotide transfer and the F/YGS motif responsible for ATP binding, confirming that it belongs to the polymerases-like superfamily of nucleotidyl transferases (García-Vivas et al., 2005). In the predicted three-dimensional model, both functional domains fold as a U-shaped structure that likely orients the incoming ATP and RNA molecules for poly(A) tail extension. In agreement with its expected role in the poly(A) tail synthesis, EhPAP was able to bind RNA 3′ end although it has a divergent RNAbinding domain (RBD). Moreover, it was located in punctuate nuclear foci and in cytoplasmic dots, as it has been described for other polyadenylation factors (García-Vivas et al., 2005; López-Camarillo et al., 2007).

As the human homolog, the EhCFIm25 protein is an unconventional Nudix protein lacking three of the four conserved E residues and the last G residue of the Nudix box, which suggests that it is not able to cleave RNA. Despite the absence of a classical RBD, EhCFIm25 is able to bind the 3′ UTR of E. histolytica transcripts and this interaction involves the participation of the conserved Leu135 and Tyr236 residues (Ospina-Villa et al., 2015). EhCFIm25 recognizes the GUUG motif in mRNA 3′ end, while the human protein binds the UGUA sequence; consequently, aptamers containing this motif specifically identify the parasite protein, suggesting that they could be used as molecular tools for diagnosis or therapeutic purposes (Ospina-Villa et al., 2018). Interestingly, EhCFIm25 interacts with EhPAP (Pezet-Valdez et al., 2013), which indicates that it may be introduced into the processing complex in the early steps of the cleavage/polyadenylation reaction. EhCFIm25 controls the selection of the distal poly(A) site during mRNA 3 ′ end formation; consequently, its silencing by dsRNA or its blocking by aptamers alter parasite proliferation and virulence, demonstrating for the first time the relevance of polyadenylation factors as biochemical targets in E. histolytica (Ospina-Villa et al., 2017, 2018). Other works have also recently reported that targeting polyadenylation factors represents a valuable strategy for the control of the protozoan parasites Trypanosoma brucei (CPSF-30), Toxoplama gondii, and Plasmodium falciparum (CPSF-73) (Hendriks et al., 2003; Sidik et al., 2016; Palencia et al., 2017; Sonoiki et al., 2017).

EhPC4 is the homolog of the human positive coactivator 4, a multifunctional protein that establishes an important link between transcription and polyadenylation. On one hand, its binding to promoters facilitates the recruitment of transcription factors to stimulate the pre-initiation complex assembly (Conesa and Acker, 2010); on the other hand, its interaction with EhCstF64 avoids premature transcription termination and polyadenylation initiation until the polyadenylation motifs have been transcribed (Calvo and Manley, 2001). Moreover, it mediates chromatin organization and heterochromatin gene silencing by interacting with histones H3 and H2B (Das et al., 2006, 2010). As homologous proteins, the EhPC4 protein contains a single-strand DNA (ssDNA) binding region whose residue K127 is required for DNA interaction, and a dimerization domain in the so-called PC4 domain at the C-terminus (Hernandez de la Cruz et al., 2014). Interestingly EhPC4 and its potential partner, EhCstf-64, were significantly up-regulated in virulent trophozoites (Santi-Rocca et al., 2008). Consistently,

the overexpression of EhPC4 induced the modulation of proteins with key functions in cytoskeleton dynamics, cell migration and invasion in trophozoites. Among them, the up-regulation of a 16-kDa actin-binding protein (EhABP16) which is a putative member of the cofilin/tropomyosin family involved in actin polymerization was associated with an increase in parasite migration of trophozoites and destruction of human SW480 colon cells, confirming that EhPC4 has an impact on parasite virulence (Hernandez de la Cruz et al., 2014). On the other hand, the overexpression of EhPC4 significantly increased cell proliferation, DNA replication and DNA content of trophozoites, promoting the formation of giant multinucleated trophozoites. EhPC4 modulates the expression of genes involved in carbohydrate and nucleic acid metabolism, chromosome segregation and cytokinesis, evidencing the relevance of this factor in polyploidy and genome stability in E. histolytica (Hernández de la Cruz et al., 2016). The role of EhPC4 in mRNA 3 ′ end formation and its relevance for the events mentioned above remain to be investigated.

# MOLECULAR EVENTS FOR MRNA DECAY IN *E. HISTOLYTICA*

# mRNA Degradation Machineries

E. histolytica has most of the factors that are involved in mRNA degradation in eukaryotic cells, including proteins involved in deadenylation, decapping, and exonuclease activity, but it lacks several components (**Table 3**) (López-Rosas et al., 2012). The reduced mRNA deadenylation machinery includes the CAF1/NOT complex with the five NOT proteins and the poly-A specific ribonucleases CAF1 and CAF1-like, but the carbon catabolite repressor 4 (CCR4) described in yeast and human, as well as PAN2, PAN3, and PARN deadenylases, are missing (**Figure 3**). EhCAF1 is a ribonuclease D family member, having the CAF1 nuclease domain and the conserved DEDD residues (D84E86D206D276) that are important for 3′ to 5′ exonuclease activity in homologous proteins (Daugeron et al., 2001). Consequently, EhCAF1 is a functional deadenylase that binds 3′UTR and degrades the poly(A) tail of parasite transcripts in in vitro assays (López-Rosas et al., 2014).

Although capping has not been described in E. histolytica, bioinformatics analyses revealed the existence of a decapping complex that is formed by the catalytic subunit EhDCP2, and EhXRN2, EhLSM1–6, EhEDC3, and EhDHH1 as decapping associated proteins, whereas it also includes DCP1, SCD6, PAT1, and LSM7 in yeast and animals (**Table 3**; **Figure 3**) (López-Rosas et al., 2012). EhXRN2 and EhDCP2 have the typical architecture of homologous proteins. Notably, EhXRN2 has the XRN\_N nuclease domain and the internal tower domain with the active site motif KX2QQX2RR, which is critical for ribonuclease function (Xiang et al., 2009). EhDCP2 has the conserved DCP2 box A domain and the conserved nudix Box (GX5EX7REUXEEXGU) that are both responsible for cap structure removal (She et al., 2008). In eukaryotic cells, the elimination of the 5′ cap compromises mRNA to 5′ to 3′

TABLE 3 | Comparison of mRNA decay machineries between human and E. histolytica.


<sup>a</sup>UniProtKB; <sup>b</sup>AmoebaDB.

exonucleolytic decay, apparently in an irreversible way, hence, decapping activity is tightly regulated (Li and Kiledjian, 2010). The heptameric Lsm1–7 complex associates with the 3' end of deadenylated mRNAs and promotes decapping (Tharun et al., 2000; Tharun and Parker, 2001). The activity of the decapping enzyme is stimulated by accessory proteins, such as Edc proteins, the DExD/H-box RNA helicases Dhh1 and Pat1 (Bonnerot et al., 2000; Schwartz et al., 2003).

E. histolytica contains seven exosome encoding genes including Rrp41, Rrp43, Rrp46, Mtr3-Rrp42 and the catalytic subunit Dis3, as well as accessory stabilizing Rrp4, and Rrp40 proteins; but it lacks Rrp45 and Csl4 genes (**Table 3**; **Figure 3**) (López-Rosas et al., 2012). The EhRRP41 protein colocalizes and physically interacts with EhL-PSP, which also interacts and colocalizes with the EhCAF1 deadenylase. But the fact that EhRRP41 did not coimmunoprecipitate with EhCAF1, suggests the existence of two EhL-PSP-containing complexes. The colocalization of exosome factors (EhRrp41) with EhCAF1 and EhL-PSP in trophozoites showed novel interactions between mRNA degradation protein and suggests the existence of cooperative interactions between mRNA decay machineries in E. histolytica (López-Rosas et al., 2014). In yeast and human, the nine subunits of the exosome complex form a ring structure in which Dis3 (RRP44) is the key player in mRNA turnover being the catalytic subunit responsible for exonucleolytic and endonucleolytic activities in the 3′ -5′ decay of deadenylated transcripts in cytoplasm (Ibrahim et al., 2008). Additionally, the nuclear exosome is involved in 3′ -end trimming of rRNA, snRNA, and snoRNA, as well as mRNA surveillance and degradation of cryptic unstable transcripts (Parker and Song, 2004).

E. histolytica genome also contains genes for components of the NMD and RNA interference (RNAi) pathways, namely three Ehupf genes (López-Rosas et al., 2012), as well as two EhRdRP, one EhRNAseIII and three EhAGO2 proteins (Abed and Ankri, 2005; Zhang et al., 2008, 2011), respectively (**Table 3**). The absence of DICER and GW182 homologs suggests that RNA interference may use DICER-independent mechanisms in E. histolytica (Zhang et al., 2011). Pompey et al. (2015) recently showed that EhRNAseIII is able to cleave dsRNA to generate shorter fragments in a heterologous system. This suggests that EhRNAseIII in conjunction with other amoebic factors might reconstitute an active DICER-like complex. Congruently, numerous reports involving gene-silencing assays confirmed the functionality of the RNAi pathway in E. histolytica.

# P-Body-Like Structures

Several experiments suggest that mRNA decay reactions, namely deadenylation, decapping, and 5′ -exonucleolytic decay, take place in microscopically detectable cytoplasmic P-bodies like structures in E. histolytica (López-Rosas et al., 2012), as it has been described in other eukaryotic cells (Sheth and Parker, 2003). The EhCAF1 deadenylase, EhXRN2 exoribonuclease and EhDCP2 decapping proteins, as well as the EhAGO2-2 protein, were detected in cytoplasmic foci in immunofluorescence and confocal microscopy experiments (López-Rosas et al., 2012). Additionally biochemical analysis revealed that EhCAF1 coimmunoprecipitated with EhXRN2, thus linking deadenylation to 5'-to-3' mRNA degradation. Interestingly, these cytoplasmic structures also contain polyadenylated transcripts and dsRNA, which is congruent with their role in RNA decay. Moreover their formation depends on the presence of active transcription and translation (López-Rosas et al., 2012), as well as cellular stress, such as DNA damage, heat shock, and nitric oxide (López-Rosas et al., 2014), which make them bona fide P-body structures (**Figure 3**). Altogether, these data suggest that, as in human cells, the accumulation of transcripts in cytoplasmic P-bodies like structures for silencing or decay, represents a key regulatory process for gene expression regulation in response to specific conditions or signals in E. histolytica.

# CONCLUSION

Besides the evolutionary distance between E. histolytica and its human host, the screening of parasite genome sequences and the functional characterization of specific factors, revealed that molecular mechanisms regulating mRNA processing and degradation seem to be roughly similar in both organisms. Several subtle differences exist, but canonical factors involved in splicing, polyadenylation and decay are generally conserved in this primitive eukaryote, which highlights that these events are key players for gene expression regulation in eukaryotic cells. The study of a larger number of factors involved in splicing, polyadenylation or mRNA degradation remains to be addressed to elucidate all the relationship among these reactions. In addition to contribute to the better understanding of posttranscriptional regulation in E. histolytica, the characterization of these factors and events may also lead to the identification of a biochemical target involved in various mRNA processing pathways, whose inhibition would have a massive impact on parasite survival. On the other hand, recent data indicated the potential of factors involved in polyadenylation as biochemical targets for parasite control, which may open the way for the design of new molecules for the control of this parasitic disease. In this context, the results of the proof-of-concept study in E. histolytica may promote the use of aptamers to control E. histolytica during the development of amoebiasis or to eradicate residual trophozoites during antibiotic treatment. Further experiments are required to confirm their affinity, evaluate their effect in vivo and improve their bioavailability.

# FUNDING

This study was supported by the Mexico-France program grants ECOS NORD (M14S02) and SEP-CONACYT-ANUIES (249554), and Mexican grants from SIP-IPN (20170969) and CONACyT (178550), Mexico. JO-V was a scholarship recipient of Mexican BEIFI-IPN and CONACyT programs. LM is supported by COFAA-IPN.

# AUTHOR CONTRIBUTIONS

JV-F, IL-R, CL-C, ER-M, and JO-V reviewed data about mRNA splicing, polyadenylation and decay, respectively. JO-V designed the figures. CL-C and LM designed the review organization, revised and integrated the different parts of the manuscript.

# REFERENCES


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The handling Editor declared a shared affiliation, though no other collaboration, with one of the authors JV-F.

Copyright © 2018 Valdés-Flores, López-Rosas, López-Camarillo, Ramírez-Moreno, Ospina-Villa and Marchat. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# The Entamoeba histolytica Syf1 Homolog Is Involved in the Splicing of AG-Dependent and AG-Independent Transcripts

Diana M. Torres-Cifuentes <sup>1</sup> , José M. Galindo-Rosales <sup>1</sup> , Odila Saucedo-Cárdenas 2,3 and Jesús Valdés <sup>1</sup> \*

<sup>1</sup> RNA Laboratory, Department of Biochemistry, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico City, Mexico, <sup>2</sup> Departamento de Histología, Facultad de Medicina, Universidad Autónoma de Nuevo León, Monterrey, Mexico, <sup>3</sup> División de Genética, Centro de Investigación Biomédica del Noreste, Instituto Mexicano del Seguro Social, Monterrey, Mexico

Syf1 is a tetratricopeptide repeat (TPR) protein implicated in transcription elongation, spliceosome conformation, mRNA nuclear-cytoplasmic export and transcription-coupled DNA repair. Recently, we identified the spliceosomal components of the human parasite Entamoeba histolytica, among them is EhSyf. Molecular predictions confirmed that EhSyf contains 15 type 1 TPR tandem α-antiparallel array motifs. Amoeba transformants carrying plasmids overexpressing HA-tagged or EhSyf silencing plasmids were established to monitor the impact of EhSyf on the splicing of several test Entamoeba transcripts. EhSyf Entamoeba transformants efficiently silenced or overexpressed the proteins in the nucleus. The overexpression or absence of EhSyf notably enhanced or blocked splicing of transcripts irrespective of the strength of their 3′ splice site. Finally, the absence of EhSyf negatively affected the transcription of an intron-less transcript. Altogether our data suggest that EhSyf is a bona fide Syf1 ortholog involved in transcription and splicing.

### Edited by:

Anjan Debnath, University of California, San Diego, United States

### Reviewed by:

César López-Camarillo, Universidad Autónoma de la Ciudad de México, Mexico Gretchen Marie Ehrenkaufer, Stanford University, United States

\*Correspondence:

Jesús Valdés jvaldes@cinvestav.mx

Received: 23 March 2018 Accepted: 18 June 2018 Published: 09 July 2018

#### Citation:

Torres-Cifuentes DM, Galindo-Rosales JM, Saucedo-Cárdenas O and Valdés J (2018) The Entamoeba histolytica Syf1 Homolog Is Involved in the Splicing of AG-Dependent and AG-Independent Transcripts. Front. Cell. Infect. Microbiol. 8:229. doi: 10.3389/fcimb.2018.00229

Keywords: Xab2, Prp19C, splice sites, transcription, nuclear, CTD, human parasite

# INTRODUCTION

The spliceosome carries out co-transcriptional pre-mRNA (pre-messenger RNA) splicing (Wahl et al., 2009) with dynamical and sequential assembly of subunits on the pre-mRNA substrate for every splicing episode (Wahl and Lührmann, 2015) by, identifying conserved intronic elements. Whereas the U1 snRNP (small nuclear ribonucleoprotein) first binds to the 5′ ss (splice site), SF1 (splicing factor 1) and U2AF65/35 (U2 snRNP auxiliary factors of 65 and 35 kDa) bind the BS (branch point sequence), the pY (polypyrimidine tract) and the 3′ ss, respectively, thus forming complex A. Next, the small RNA molecule of U2 snRNP replaces SF1 forming a double stranded RNA hybrid with the BS, bulging the conserved A of the branch necessary for the first of the two transesterification reactions involved in the elimination of the intron-lariat and splicing of exons. Upon entry of the U4/U6.U5 tri-snRNP complex B is formed, which becomes an active spliceosome (BACT) when U4/U6 basepairing is unwound allowing U6 snRNA to form two mutually exclusive interactions with the 5′ ss and U2 snRNA, destabilizing U1 and U4 snRNPs. The first transesterification reaction occurs when U6/U2 and U2/BS RNA/RNA interactions position the bulged BS adenosine to attack the 5′ ss, producing 5 ′ exon and intron lariat-3′ exon intermediates. The active spliceosome undergoes various rearrangements to form complex C that carries out the second transesterification reactions, ligating the exons and removing the intron-lariat (Wahl et al., 2009).

Other non-snRNPs factors are major players in the splicing process (Wahl et al., 2009), among them the components of the Prp19 complex or NTC in yeast: Prp19, Cef1, Syf1, Syf2, Syf3, Snt309, Isy1, and Ntc20 (Fabrizio et al., 2009). Prp19 is the scaffold for NTC formation, and it is essential for splicing but does not constitute any of the individual spliceosomal snRNPs (Tarn et al., 1993a,b, 1994). The NTC joins the spliceosome before, or during, U4/U6 unwinding and remain associated with the spliceosome during the two steps of splicing, marking the change from the inactive B to the active BACT spliceosome (Chan et al., 2003). Recent evidence shows that NTC regulates spliceosome conformations and fidelity (Hogg et al., 2010), and as part of the intron lariat complex participates in spliceosome disassembly (Fourmann et al., 2013).

It has been shown that U2AF65 and NTC interact both in vitro and in vivo, and this interaction is required for activation of splicing (David et al., 2011). U2AF65 binds directly to the Ser2 phosphorylated carboxy terminus domain of RNA Pol II (CTD), strengthening U2AF65 and NTC recruitment to the pre-mRNA (David et al., 2011). Thus U2AF65/NTC tether the spliceosome and pre-mRNA to the CTD of elongating Pol II. By this mechanism the CTD enhances splicing, and describes interactions important for splicing and its coupling to transcription (David et al., 2011). In addition, such mechanism links Syf1 to transcription elongation, mRNA export, and to the nuclear excision DNA repair machinery in a transcription-coupled manner.

The human-to-yeast conserved TREX (TRanscription /EXport) complex shuttles mRNA from the nucleus to the cytoplasm. TREX is formed by the THO complex along with mRNA export proteins UAP56/Sub2 and Aly/Yra1. Strong evidence links TREX to transcription elongation in yeast. Strains carrying mutations in any of the four THO members fail to export mRNA and are defective in transcription elongation accumulating transcripts in or near their transcription start sites (Carmody and Wente, 2009). Furthermore, it has been shown that Syf1 acts as a novel transcription elongation factor required for TREX occupancy at transcribed genes (Chanarat et al., 2011). In mammals, TREX binds the mRNA co-transcriptionally suggesting that TREX is associated with the late stages of splicing. Whereas HPR1, a member of the THO complex, directly interacts with UAP56/Sub2 and is essential for UAP56/Sub2 and Aly/Yra1 recruiting to the mRNA (Zenklusen et al., 2002), in humans all THO components co-purify with spliceosomes (Jurica and Moore, 2003; Reed and Cheng, 2005).

Additionally, damage/alterations in DNA structure can interfere with DNA and RNA polymerases or compromise replication and transcription fidelity. Damaged DNA is restored by nuclear excision repair (NER). Two major NER pathways exist: global genome repair (GGR) and transcription-coupled repair (TCR). In the former, damage recognition is carried out by a DDB heterodimer which binds to the XPC-RAD23B-CEN2 complex (Hamann et al., 1995). In the later, proteins CSA and CSB together with Syf1 recognize DNA damage and recruit the DNA repairing machinery (Venema et al., 1990; Nakatsu et al., 2000). Both pathways converge in the following steps of DNA damage repair (Hanawalt and Spivak, 2008).

The protozoan parasite Entamoeba histolytica is the causative agent of amebiasis. E. histolytica infects approximately 1% of the human population, resulting in approximately 100,000 deaths annually (Nakada-Tsukui and Nozaki, 2016). Because pre-mRNA splicing mechanisms are almost unknown in this parasite, despite the limited amount of data, significant advances have been made to identify the Entamoeba splicing machinery and particular machineries (Miranda et al., 1996; Hernández-Rivas et al., 2000; Davis et al., 2007; Dávila López et al., 2008; Marchat et al., 2008; Hon et al., 2013; Valdés et al., 2014). Previous work using molecular biology and bioinformatic approaches revealed: splicing factor Prp6 (Hernández-Rivas et al., 2000), U2, U4, U5, and U6 snRNAs (Miranda et al., 1996; Davis et al., 2007), the possible lack of U1 snRNA in E. histolytica (Dávila López et al., 2008), the DExH/D RNA helicases involved in the proofreading of the sequential steps of spliceosome assembly and catalysis (Marchat et al., 2008), the preferred route of alternative splicing and the wide variety of alternative splice sites utilized by the Entamoeba spliceosome (Hon et al., 2013). Other proteins involved in pre-mRNA processing and maturation such as cleavage and specificity factors (CPSF160, 100, 73, and 30), cleavage stimulating factors (CstF77, 64, and 50), the cleavage factor Im of 25 kDa, both subunits of CFIIm (C1P1 and PCF11), FIP1, poly(A) polymerase, poly(A) binding protein, RBBP6 (Mpe1 in yeast), WDR33, PNAS-120, and PC4 (López-Camarillo et al., 2005, 2014). We have cloned HA-tagged snRNP component U1A and immunoprecipitated in vivo assembled pre-mRNA splicing complexes (Valdés et al., 2014). Among the nearly forty splicing factors identified by mass spectrometry were members of every splicing event, including factors involved in the formation of the aforementioned complexes A, B, and BACT, as well as catalytic, post-catalytic, intron lariat and disassembly complexes (reviewed by Valdés et al., this issue). Of note is the E. histolytica putative ortholog of U2AF65, which participates in spliceosome/pre-mRNA (at the 3′ ss of the intron)/CTD of Pol II. Factors of the Entamoeba histolytica NTC, Prp19 (EHI\_13870), Cwc2 (EHI\_126150), Cef1 (EHI\_000550) and Syf (EHI\_073300) were also identified. Furthermore, microarray data showed that EhSyf transcripts were overexpressed in amoebas transformed with HA-U1A (not shown) and splicing factor 1 (SF1). Trying to understand EhSyf participation in fundamental biological processes and to provide evidence about a key molecule that links transcription and splicing in a deep branched eukaryote, we cloned and characterized the general impact of EhSyf on splicing irrespective of 3′ ss strength of Entamoeba virulence-related and virulence-unrelated introns.

# MATERIALS AND METHODS

# Entamoeba Cultures

Axenic cultures of E. histolytica trophozoites strain HM-1:IMSS Cl-6 were incubated at 37◦C in 13 × 100 mm screwcapped Pyrex glass tubes or plastic culture flasks in BI-S-33 medium as described (Diamond et al., 1972, 1978, 1995).

# Plasmid Constructs and Amoeba Transfectants

The EhSyf gene was amplified by PCR using oligonucleotides containing appropriate restriction sites. PCR products were cloned in plasmids pEhExHA (Saito-Nakano et al., 2004) and pKT3M 04-trigger (Morf et al., 2013). To facilitate EhSyf cloning, a Sma Isite was inserted in the pKT3M 04-trigger plasmid by sitedirected mutagenesis. Constructs were verified by sequencing. Trophozoites were transformed with plasmids by liposomemediated transfection as previously described (Nozaki et al., 1999). Transformants were selected with 5 or 10µg/ml of Geneticin.

# Western Blot and Immunofluorescence

Protein extracts (60 µg) of amoeba transformants were separated by 10% SDS-PAGE and analyzed by western blotting using mouse anti-αSyf (human XAB2; Thermo Scientific), goat anti-Pol II (Santa Cruz Biotechnology), mouse anti-Actin, and secondary horseradish peroxidase-conjugated anti-IgG antibody (Sigma-Aldrich). The human anti-αSyf antibody target in EhSyf was validated by CLUSTALW alignment (Figure S1A). Enhanced chemiluminescent reagent (Perkin Elmer) was used to detect the proteins. To immunolocalize HA-EhSyf in amoeba transformants, cells were fixed and permeabilized in methanol at −20◦C using anti-HA Alexa Fluor Labeled for viewing in a Zeiss LSM700 confocal microscope. All experiments were conducted in experimental and biological triplicates.

# RT-PCR

For overexpression, silencing and in vivo splicing assays EhSyf, Sam50, RabX13, MybS6, ClcB1, Cdc2, actin, and RNA polymerase II expression was monitored by RT-PCR. Total RNA of amoeba transformants was extracted with Trizol (Invitrogen) and the synthesis of cDNA was performed using the SuperScript III First Strand Synthesis System (Invitrogen) according to the manufacturer's instructions. Unspliced pre-mRNA/intron retained (IR) and mRNA molecules were detected with their respective primers (Table S1). Three biological replicates were done for each experiment. PCR products were resolved by electrophoresis and images of the gels were acquired in color inversion mode.

# Statistical and Bioinformatics Analyses

Phylogenetic trees were constructed with MEGA 7.0 (http:// www.megasoftware.net/) using the neighbor-joining method; the significance of nodes was assessed using a bootstrap test with 1,000 (Tamura et al., 2013). To identify the type of TPR repeat consensus of EhSyf, CLUSTALW and CTREE alignments of the C-terminus of EhSyf (EHI\_073300), Homo sapiens XAB2 (accession number NP\_064581), Saccharomyces cerevisiae (NP\_010704), Drosophila melanogaster ORF de CG6197 accession number NP\_610891, Caenorhabditis elegans ORF C50F2.3, accession number NP\_491250, Schizosaccharomyces pombe ORF SPBC211.02c, accession number NP\_596612 were carried out. EhSyf modeling was carried out with I-Tasser. For RT-PCR and western blots, differences between means were determined by t-test with GraphPad Prism 5.0. p-values were calculated comparing all pairs of the expression percentages obtained from the normalized IntDen values of densitometries using ImageJ. To eliminate plasmid transfection-effect, differences of transcriptexpression values were normalized with respect to 18S rRNA (EhSyf) or RNA polymerase II (monitor transcripts) and were expressed as percent of the total expression (unspliced plus spliced gene products) as previously reported (Goren et al., 2006; de la Mata et al., 2010). Comparisons were carried out between empty vector and EhSyf constructs, prior verification of comparable copy number of plasmids by RT-PCR.

# RESULTS

We first performed phylogenetic inference of EhSyf. All EhSyf hits in the BLAST analysis were used to construct the rooted phylogeny tree (**Figure 1A**). As expected for E. histolytica proteins, the identity values of its relatives ranged from 20 to 30% (Table S2). The node of EhSyf relatives include the angiosperm Phalaenopsis equestris, the aphid Myzus persicae, the calanoid copepod Eurytemora affinis, the animal pathogen Basidiobolus meristosporus, E. nuttalli, E. invadens, E. dispar, other E. histolytica strains. Most importantly such node includes a representative of the deepest branching clade of fungi Rozella allomycis, and the only extant representative of basal metazoans Trichoplax adhaerens suggesting an early origin of EhSyf. This node appears somewhat distant from the rest of EhSyf relatives, however no other tetratricopeptide repeats (TPR) proteins were identified in the initial BLAST.

TPR proteins participate in a wide variety of functions involving protein-protein interactions (Das et al., 1998; Blatch and Lässle, 1999; Makiuchi et al., 2013). As deduced from the last six TPR motifs alignment with human, yeasts, insects and worms (**Figure 1B**), and in silico structural models, EhSyf contains 15 TPR class I (Sikorski et al., 1991) motifs in antiparallel α-helix arrays (**Figures 1C,D**). Interestingly, the Cterminus domain of EhSyf keeps low identity (I) but significant similarity (S) percentages with human (I 16.6%; S 36.4%), S. cerevisiae (I 19.3%; S 36.8%), S. pombe (I 25.2%; S 42.8%), D. melanogaster (I 16.9%; S 33.7%), and C. elegans (I 17.6%; S 35.2%).

To partially understand EhSyf functions, pKT3M plasmidbased silencing (Morf et al., 2013) and HA-tagged over

alignments of the last 6 TPR motifs of EhSyf with Syf proteins from Drosophila melanogaster, Homo sapiens, Caenorhabditis elegans, Schizosaccharomyces pombe and Saccharomyces cerevisiae. (C) I-Tasser structure model of the 699 amino acids of EhSyf. (D) CTREE alignments of human TPR Types I-III motifs consensus sequences with EhSyf TPRs, showing that EhSyf possess Type I TPRs (Sikorski et al., 1991). (E) EhSyf silenced (pKTSyf) and over expressed (pHASyf) amoeba (Continued)

FIGURE 1 | transformants were established by selection with the indicated amounts of G418 and Syf transcripts were monitored by RT-PCR. Basal post-transfection expression was compared to the Splicing Factor 1 silenced (pKTSF1) transfectants, and 18S rRNA was used for normalization. (F) Western blots were carried out with anti-αSyf (human) antibodies to assess protein overexpression and silencing of EhSyf in amoeba transformants, compared to the human Hek293 cell extracts. For normalization, actin and RNA polymerase II proteins were used. (G) Plot of normalized EhSyf protein overexpression (80%) and silencing (32%) in amoeba transformants. Error bars indicate the SD of three independent experiments, asterisks show statistically significant differences (t-Student P \* < 0.05 \*\*\* < 0.001) compared to empty vectors.

expression plasmid pEhExHA (Saito-Nakano et al., 2004) Syf Entamoeba transformants were established by selection with different amounts of geneticin, and their respective transcription and protein expression were monitored by RT-PCR and western blot. Contrary to transcriptomic analyses reported in AmoebaDB, we were not able to detect wild type EhSyf mRNA (**Figure 1E**), suggesting that Entamoeba expresses very low amounts of EhSyf. This is consistent with our previous findings in which EhSyf was not detected in MS2 aptamer-tagged RabX13 intron immunoprecipitates, which contained components of splicing B and C complexes, although other members of U2 snRNP and NTC were detected (Valdés et al., 2014; Valdes et al. this issue), reflecting NTC rearrangements within the spliceosome during pre-spliceosome to pre-catalytic complex transition. In spite of our observations, and because we expect constitutive EhSyf expression, we ensured the lack of EhSyf by means of pKTSyf-mediated silencing. To eliminate off-target effect another splicingrelated protein SF1 (pKTSF1) was silenced, and we observed that EhSyf mRNA expression was enhanced, instead of reduced. Conversely, more EhSyf mRNA was detected in pHASyf overexpression transformants. In agreement with this, whereas the ≈ 84 kDa EhSyf was robustly expressed in mocktransfected amoebae (pHA), pHASyf transformants expressed 80% more, nearly as abundant as that of Hek293 human cells (**Figure 1F**, Figures S1A–D). Although we cannot ascertain EhSyf mRNA reduction, only 32% reduction of EhSyf protein expression was achieved in pKTSyf amoeba transformants (**Figure 1G**). Finally, anti-HA confocal microscopy immunofluorescence experiments showed that EhSyf localized to the nucleus similar to the HA-U1A splicing factor control (**Figure 2A** and Figure S1E).

To determine its role in splicing, we reasoned that slight changes of EhSyf expression, being part of the NTC, would imbalance wild type pre-mRNA/mRNA ratios in a general manner. Despite the capacity of U1 snRNP components to define the 5′ ss and of the U2AF dimer to define the 3′ ss. These protein/RNA interactions must be taken into consideration since most 5′ ss (GUUUGUUU) are strong and conserved in Entamoeba (Hon et al., 2013), in the absence of U1 snRNA RNA (Dávila López et al., 2008). Therefore 5′ ss definition must be carried out by U1 snRNP components and splicing auxiliary factors as described for other systems (Förch et al., 2002; Huang et al., 2012). For 3'ss definition, whereas strong (AG-independent) poly-pyrimidine tracts do not always require U2AF35, both U2AF35/65 are required to define AG-dependent 3'ss with weak poly-pyrimidine tracts (Wu et al., 1999). A collection of introns was analyzed with respect to their AGdependency (**Figure 2B**) and several were chosen to test our hypothesis. AG-dependent introns of Sam50 and Cdc2, and AGindependent introns of RabX13, MybS6 and intron 1 of ClcB. As a control the intron-less Actin transcript was analyzed. These introns were also selected because they represent both virulencerelated [Sam50, MybS6, and ClcB are downregulated after trophozoites recuperated from hamster liver abscesses (Weber et al., 2016)] and virulence-unrelated transcripts. Furthermore, in steady state conditions, they always produce unspliced premRNA/intron retained (IR) and mRNA variants facilitating the monitoring of splicing variats. In vivo splicing assays were carried out using RNA extracts from the EhSyf expresser and silenced transformants. Compared to the HA control, spliced mRNA increased in all HASyf expresser transformants at the expense of pre-mRNA/IR variants. Correspondingly, EHSyf silencing resulted in pre-mRNA/IR accumulation at the expense of mRNA, compared to the pKT silencing controls (**Figure 2C**). Statistically different transcription levels are shown (**Figure 2D**). Unexpectedly this approach allowed us to detect that high selective pressure EhSyf silencing affected Actin gene expression too. This change was not observed in U2AF-silenced ameba transformants (psU2AF) selected with the same amount of G418 (**Figures 2B,C**), indicating that low amounts of EhSyf impact transcription of intron-less genes too.

# DISCUSSION

EhSyf gene encodes a tetratricopeptide repeats protein involved in several nuclear functions. The Entamoeba Syf gene products appear to have low copy number that can be boosted during transfection of unrelated genes. This might reflect the direct/indirect participation of EhSyf in transcription, or it might reflect that the introduction of plasmid DNA evokes general transcription/DNA-repair mechanisms, increasing in turn EhSyf transcripts. The impact of plasmid transfection on endogenous genes has been documented since the pioneering works attempting heterologous expression of proteins in E. histolytica (Hamann et al., 1995).

EhSyf expression directly impacts splicing in a general manner. This reflects its relationship with the large subunit of U2AF which binds to poly-pyrimidine tracts in front of the 3′ ss, and tethers the pre-mRNA/spliceosome to the CTD. Interestingly, EhSyf silencing but not overexpression, negatively affected transcription even of intron-less genes indicating its role in transcription elongation.

Together, our data suggest the coordinated functions of EhSyf in transcription and splicing in this early branched protist.

FIGURE 2 | pre-mRNA splicing is favored in amoebae overexpressing EhSyf and defective with the knockdown. (A) Entamoeba histolytica HA-Syf transformants were plated onto microscopy slides, incubated with FITC-coupled anti-HA antibodies to detect appropriate nuclear localization of EhSyf protein by confocal microscopy. Nuclei were stained with DAPI. (B) WebLogos of the 3′ ss of E. histolytica AG-independent (AG-i) and AG-dependent (AG-d) introns (Wu et al., 1999). (C) Representative semi-quantitative RT-PCR assays with specific primers to amplify Sam50, RabX13, MybS6, Clcb1, Cdc2 and the intron-less Actin gene products in 5 or 10µg/ml G418 selected amoeba transformants. RNA polymerase II transcript was used to normalize the expression levels. For comparison, Actin gene expression was also monitored in U2AF splicing silenced amoebae (psU2AF). (D) Error bars indicate the SD of three independent experiments. The asterisks show significant differences (t-Student P \* < 0.05 \*\* < 0.01) compared to empty vector transfectants. Whereas for Sam50, RabX13 and Actin the results shown correspond to HASyf expresser and knockdown transformants selected with 5 and 10µg/ml G418, respectively; MybS6 gene selection of HASyf expression and knockdown figures show the data of transformants selected with 10 and 5µg/ml G418, respectively. Clcb1 and Cdc2 gene products showed no significant differences in any selection (shown 5µg/ml G418).

Although the impact of EhSyf in transcription termination and mRNA export still remains to be tested. We conclude that EhSyf is a bona fide Entamoeba histolytica Syf1 ortholog that participates in splicing and transcription elongation. To our knowledge this is the first evidence of the molecular link between transcription and splicing in Entamoeba histolytica.

# AUTHOR CONTRIBUTIONS

JV, DT-C conception, experimental design, manuscript preparation; DT-C, JG-R, OS-C, and JV data acquisition and interpretation.

# REFERENCES


# ACKNOWLEDGMENTS

The authors are grateful to Tomoyoshi Nozaki and Upinder Singh for plasmids. Funding was provided by CONACYT (Grants 127557-M and 236104 to JV; Scholarship 141030016 to DT-C).

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00229/full#supplementary-material


formation of the functional spliceosome. Mol. Cell. Biol. 13, 1883–1891. doi: 10.1128/MCB.13.3.1883


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Torres-Cifuentes, Galindo-Rosales, Saucedo-Cárdenas and Valdés. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Unexplored Molecular Features of the Entamoeba histolytica RNA Lariat Debranching Enzyme Dbr<sup>1</sup> Expression Profile

Jesús Valdés <sup>1</sup> , Carlos Ortuño-Pineda<sup>2</sup> , Odila Saucedo-Cárdenas 3,4 and María S. Mendoza-Figueroa1,5 \*

<sup>1</sup> Departamento de Bioquímica, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico City, Mexico, <sup>2</sup> Unidad Académica de Ciencias Químico Biológicas, Universidad Autónoma de Guerrero, Chilpancingo, Mexico, <sup>3</sup> Histología, Facultad de Medicina, Universidad Autónoma de Nuevo León, San Nicolás de los Garza, Mexico, <sup>4</sup> División de Genética, Centro de Investigación Biomédica del Noreste, Instituto Mexicano del Seguro Social, Monterrey, Mexico, <sup>5</sup> Departamento de Atención a la Salud, Universidad Autónoma Metropolitana-Xochimilco, Mexico City, Mexico

The RNA lariat debranching enzyme (Dbr1) has different functions in RNA metabolism, such as hydrolyzing the 2′ -5′ linkage in intron lariats, positively influencing Ty1 and HIV-1 retrotransposition, and modulating snRNP recycling during splicing reactions. It seems that Dbr1 is one of the major players in RNA turnover. It is remarkable that of all the studies carried out to date with Dbr1, to our knowledge, none of them have evaluated the expression profile of the endogenous Dbr1 gene. In this work, we describe, for the first time, that Entamoeba histolytica EhDbr1 mRNA has a very short half-life (less than 30 min) and encodes a very stable protein that is present until trophozoite cultures die. We also show that the EhDbr1 protein is present in the nuclear periphery on the cytoplasmic basal side, contrary to the localization of human Dbr1. Comparing these results with previous hypotheses and with results from different organisms suggests that Dbr1 gene expression is finely tuned and conserved across eukaryotes. Experiments describing the aspects of Dbr1 gene expression and Dbr1 mRNA turnover as well as other functions of the protein need to be performed. Particularly, a special emphasis is needed on the protozoan parasite E. histolytica, the causative agent of amoebiasis, since even though it is a unicellular organism, it is an intron-rich eukaryote whose intron lariats seem to be open to avoid intron lariat accumulation and to process them in non-coding RNAs that might be involved in its virulence.

Keywords: Entamoeba histolytica, lariat, Dbr1, mRNA, splicing

# INTRODUCTION

The spliceosome mediates intron removal and exon ligation of pre-mRNA through two consecutive trans-esterification reactions, resulting in a 2′ -5′ linked lariat and in mRNA molecules (Padgett et al., 1984; Ruskin et al., 1984; Ruskin and Green, 1985; Konarska et al., 2006; Smith et al., 2009). Intron turnover is carried out by the RNA lariat debranching enzyme (Dbr1) that hydrolyzes the 2 ′ -5′ linkage, opening the lariats (Ruskin and Green, 1985) either to be degraded or processed in non-coding RNAs (Ooi et al., 1998). After splicing, lariat RNA is present in two different

#### Edited by:

Anjan Debnath, University of California, San Diego, United States

#### Reviewed by:

Marco A. Ramos, Universidad Autónoma de Baja California, Mexico Mark R. Macbeth, Butler University, United States

#### \*Correspondence:

María S. Mendoza-Figueroa msmendozafigueroa@gmail.com

> Received: 25 March 2018 Accepted: 18 June 2018 Published: 04 July 2018

#### Citation:

Valdés J, Ortuño-Pineda C, Saucedo-Cárdenas O and Mendoza-Figueroa MS (2018) Unexplored Molecular Features of the Entamoeba histolytica RNA Lariat Debranching Enzyme Dbr<sup>1</sup> Expression Profile. Front. Cell. Infect. Microbiol. 8:228.

doi: 10.3389/fcimb.2018.00228

post-splicing complexes, Intron Large (IL) and Intron Small (IS), but Dbr1 is not associated with either of them, indicating that Dbr1 association with lariat molecules is transitory (Yoshimoto et al., 2009; Garrey et al., 2014). Consistent with its function, the Homo sapiens Dbr1 (HsDbr1) in HeLa cells is localized in the nucleoplasm (Kataoka et al., 2013).

Dbr1 protein sequences from different organisms show high homology (Nam et al., 1997; Kim et al., 2000; Kataoka et al., 2013), suggesting that Dbr1 both in structure and in function is phylogenetically conserved. Dbr1 enzymes belong to the metallophosphoesterase (MPE) family (Koonin, 1994), showing a conserved N-terminal domain, a C-terminal domain (CTD) that does not show sequence similarity to any other class of proteins; a third domain between them, the lariat recognition loop (LRL) adjacent to the active site, is not present in other MPEs (Kim et al., 2000). As MPEs, Dbr1 enzymes use divalent metals for their activity (Arenas and Hurwitz, 1987), such as manganese in the case of the Saccharomyces cerevisiae Dbr1 (ScDbr1) (Khalid et al., 2005) or Fe2+, Zn2+, or Mn2<sup>+</sup> in the case of Entamoeba histolytica Dbr1 (EhDbr1) (Clark et al., 2016; Ransey et al., 2017), suggesting a different requirement of metal cofactors.

To date, only the protozoan EhDbr1 has been crystallized (Montemayor et al., 2014; Clark et al., 2016; Ransey et al., 2017) from these structures, with the observation that a trinucleotide formed by the adenosine branch point flanked by 2′ -5′ and 5′ -3′ bonds is the minimal substrate for the debranching enzyme (Arenas and Hurwitz, 1987), Montemayor and coworkers proposed that the lariat molecule seems to be the determinant instead of a specific sequence for substrate recognition by Dbr1, allowing for the debranching of a diverse set of lariat RNA molecules (Montemayor et al., 2014). It is proposed that the hydrolysis of the debranching reaction is a SN2 mechanism via a trigonal bipyramidal pentacoordinate intermediate that results in an investment of the configuration, after which the 2′ -O leaving group is protonated (Clark et al., 2016).

It has been observed that in S. cerevisiae Dbr1 mutants that exhibit intron accumulation (Chapman and Boeke, 1991) and complementation with HsDbr1 (Kim et al., 2000), with the Caenorhabditis elegans Dbr1 (CeDbr1) (Nam et al., 1997), or with EhDbr1 (Montemayor et al., 2014), such a phenotype is rescued.

In the present work, we describe unstudied traits of the E. histolytica Dbr1 protein, whose structure has been determined recently (Montemayor et al., 2014). We found that the EhDbr1 mRNA half-life is less than 30 min, but in contrast, it encodes a very stable protein that seems to be present until trophozoite cell cultures die. Additionally, we show that contrary to HsDbr1, EhDbr1 protein is present in the nuclear periphery on the cytoplasmic basal side, indicating that lariat opening in E. histolytica is carried out in a different cell compartment and suggesting that intron turnover in this organism might be regulated in a different way than humans. These data and data from other organisms suggest that Dbr1 gene expression is finely tuned and conserved across eukaryotes. Thus, studies with an emphasis on Dbr1 gene expression need to be performed. Particularly, in the protozoan parasite E. histolytica, EhDbr1 is possibly needed to open intron lariats that seem to be precursors of non-coding RNAs with unknown functions and is expressed from virulence genes (Mendoza-Figueroa et al. this issue).

# MATERIALS AND METHODS

# Entamoeba histolytica Cell Cultures

E. histolytica trophozoites from HMI: IMSS strain were axenically grown at 37◦C in TYI-S-33 medium (Diamond et al., 1978) until they reached the exponential growth phase. Cells were harvested first by incubating them in an ice-water bath and then collected by centrifugation. Trophozoites were used immediately to extract total RNA or proteins by the TRIzol method (Invitrogen) according to the manufacturer's instructions.

# HeLa and MRC-5 Cell Cultures

Human cervical carcinoma (HeLa) cells were grown on coverslips inside Petri dishes and cultured in DMEM (Life Technologies). MRC-5 cells were grown in RPMI (Life Technologies). Both cultures were supplemented with 10% fetal bovine serum and were incubated at 37◦C in a humid atmosphere of 5% CO2.

# RT-PCRs

Retrotranscription (RT) reactions were carried out with the M-MLV retrotranscriptase (Invitrogen) according to manufacturer's instruction. As template 10, 0.5, and 0.5 µg of total RNA were used in order to produce the EhDbr1, 18S rRNA, and EhActin cDNAs, respectively. Polymerase Chain Reactions (PCR) were carried out with 10% cDNA and the Taq DNA Polymerase (Invitrogen) according to manufacturer's instruction. Primers and PCR conditions are indicated in Table S1. As control in each experiment, PCR reactions with total RNA as template were carried out. Amplicons were analyzed in agarose gels and stained with ethidium bromide.

# EhDbr1 mRNA Half-life

Eighty percent confluent cell cultures of E. histolytica trophozoites were incubated in the presence of Actinomycin D to a final concentration of 1.3 nM (Ayala et al., 1990; Lopez-Camarillo et al., 2003) during the periods indicated in **Figure 1A**. Then, total RNA was extracted and EhDbr1 gene expression was analyzed by RT-PCR. As control EhActin and 18S rRNA gene expression were analyzed. The treatment was made by triplicate.

# EhDbr1 Protein Half-life

Eighty percent confluent cell cultures of E. histolytica trophozoites were incubated with the protein synthesis inhibitor emetine (Sigma-Aldrich) to a final concentration of 25µM (Grollman, 1966) during the periods indicated in **Figures 1C,D**. After that, cells were counted (**Figure 1C**) and total proteins extracted by incubating cells in extraction buffer [10 mM HEPES-KOH, pH 7.2, 24 mM KCl, 10 mM MgCl2, 1 mM PMSF, 2 mM DTT, 1% NP-40, in the presence of the protease inhibitor cocktail "Complete" (Roche)] during 30 min with gentle agitation on ice. Samples were centrifuged at 6,000 × g 10 min at 4◦C and proteins in the supernatant were quantitated with the DC Protein Assay Kit (Bio-Rad). Western blot assays were carried out in order to detect EhDbr1 and EhGAPDH proteins. The

antibody Dbr1 (ThermoFisher Scientific) was used in a dilution 1:1,500 and the anti GAPDH (Santa Cruz Biotechnology) in a dilution 1: 10,000. The treatment was made by triplicate.

# Nuclear and Cytoplasmic Extracts

Nuclear and cytoplasmic extracts were obtained following the Dignam (Dignam et al., 1983) and Rio protocols (Rio et al., 2011). Both extracts were quantitated by the Bradford method and used for western blot assays in order to detect the EhDbr1 protein. Also western blots using the antibody EhCPADH (Garcia-Rivera et al., 1999), a gift from Dra. Esther Orozco, in a dilution 1: 40,000 and the antibody Tri-methyl-histone H3 (Lys4) (Cell Signaling) in a dilution 1: 1,000 were carried out as cytoplasmic and nuclear controls, respectively.

# EhDbr1 Localization by Confocal Microscopy

E. histolytica trophozoites grown on coverslips were fixed and permeabilized with cold 100% methanol for 5 min and then blocked with 10% adult bovine serum for 1 h at room temperature. Cells were incubated overnight at 4◦C with the first antibody Dbr1 (1:150) or with PBS (Phosphate Buffered Saline) as negative control. After several washes, samples were incubated with the Goat anti-Rabbit IgG rhodamine conjugated (1:100) (ThermoFischer Scientific). Nuclei were stained with 4′ ,6- Diamidino-2-Phenylindole (DAPI) and samples were observed through a confocal microscope (Carl Zeiss LSM 700) using the ZEN 2009 software. Observations were performed in approximately 20 optical sections from the top to the bottom of each sample.

# RESULTS

# EhDbr1 mRNA Abundance Is Low, But the Enzyme Has High Stability

Dbr1 cDNA has been isolated and has been used in complementation assays in different organisms (Nam et al., 1997; Kim et al., 2000). However, until now, knowledge about the Dbr1 mRNA half-life has been missing. To determine the half-life of the EhDbr1 mRNA, RNA polymerase II transcription was inhibited in E. histolityca trophozoites using Actinomycin D (Ayala et al., 1990; Lopez-Camarillo et al., 2003); after different incubation times with the drug, the total RNA was isolated, and the expression of EhDbr1 was analyzed by RT-PCR. As **Figure 1A** shows, EhDbr1 mRNA disappeared almost immediately after addition of the drug to the cultures. EhDbr1 mRNA recuperated after 8 h.

Importantly, to amplify the EhDbr1 transcript, 10 µg of total RNA was needed in the RT reaction, and extra PCR cycles were employed to better visualize the product at 8 h of Actynomycin D treatment (**Figure 1A**, lanes 8 and 10 to the right of the main panel), suggesting a low level of EhDbr1 transcripts. This result is consistent with data from Arabidopsis thaliana, where amplification of the AtDbr1 transcript needed 40 PCR cycles compared to 24 for AhTUB2, indicating a low AtDbr1 mRNA abundance (Wang et al., 2004).

The identity of the PCR product was corroborated by restriction analysis of the amplicon (**Figure 1B**). EhActin gene expression was analyzed to control RNA polymerase II inhibition and, the results were consistent with previous work (Lopez-Camarillo et al., 2003); its transcription was not inhibited (half-life >12 h). The expression of the 18S rRNA was analyzed as a loading control and to control the resistance of RNA polymerase I to Actinomycin D (Oakes et al., 1993).

The EhDbr1 protein half-life was evaluated by inhibition of protein synthesis of E. histolytica trophozoites with emetine (Grollman, 1966). After incubation with the drug, the total protein was extracted and analyzed by western blots to detect the enzyme as well as to detect EhGAPDH as a loading control. In **Figure 1C**, the growth curve of E. histolytica trophozoites after emetine treatment is shown. Consistent with previous findings (Grollman, 1966), trophozoite growth was inhibited by the drug. Notably, EhDbr1 synthesis was inhibited only after 24 h of treatment (**Figure 1D**); however, protein disappearance coincided with the near absence of cells, suggesting that EhDbr1 disappearance is due to cell death instead of synthesis inhibition. This result indicates that EhDbr1 is present in the cells until trophozoites cultures die. It is interesting to note that the EhDbr1 signal disappears after 48 h, even in the absence of emetine. This result most likely occurs because starting at this time point, cell cultures are over 70% lysed and because at this time point EhDbr1 could be degraded faster than EhGAPDH.

# EhDbr1 Localization

Consistent with its function, the HsDbr1 protein is mainly localized in the nucleus (**Figure 2**), but in certain circumstances, it has been observed in the cytoplasm (Kataoka et al., 2013). We set to determine EhDbr1 localization in E. histolytica trophozoites. First, nuclear- and cytoplasmic-enriched fractions were obtained, and EhDbr1 was detected by western blot. Unexpectedly, as seen in **Figure 2A**, the EhDbr1 enzyme was localized mainly in the cytoplasm and was nearly absent in the nucleus. As controls of nuclear and cytoplasmic fractionation, histones and the EhCPADH complex (Garcia-Rivera et al., 1999), respectively, were evaluated by western blot, confirming appropriate fractionation.

The Z planes of confocal images suggest that EhDbr1 is in the nuclear periphery on the cytoplasmic side (**Figure 2Ba**), toward the basal-substrate region. Taken together, these results suggest that EhDbr1 might be associated with the nuclear envelope on the cytoplasmic basal side of the E. histolytica trophozoites.

# DISCUSSION

One of the natural products of pre-mRNA splicing are intron lariat molecules, whose functions are less characterized compared to mRNA. It seems that independent of its function, the 2′ -5′ phosphodiester linkage in the intron lariats needs to be opened by the Dbr1 enzyme to process the lariats. It is remarkable that of all the studies carried out to date with Dbr1, to our knowledge, none of them have evaluated the expression profile of the endogenous Dbr1 gene.

EhDbr1 mRNA appears to have a half-life of less than 30 min, indicating a low abundance of the transcript in the cell under steady state conditions, possibly due to low EhDbr1 gene transcription, which is consistent with the recuperation of transcription after 8 h of Actinomycin D treatment. Then, once translated, its degradation rate might be increased, resulting in a low accumulation of the EhDbr1 transcript. In this scenario, the resultant protein should have high stability, such that high quantities of mRNA are not required. In A. thaliana, Wang and coworkers observed that the accumulation level of the AtDbr1 transcript was low (Wang et al., 2004). As in this work, the authors needed a high number of PCR cycles to amplify the AtDbr1 mRNA. Furthermore, because S. cerevisiae, C. elegans and A. thaliana contain a single copy of the Dbr1 gene and because H. sapiens contain, at most, two copies (Kim et al., 2000; Wang et al., 2004), it has been proposed that HsDbr1 mRNA abundance should be extremely low and that the encoded protein should have a high specific activity, perhaps being very stable (Kim et al., 2000). In E. histolytica, the EhDbr1 gene copy number has not yet been determined, but we observed that the EhDbr1 protein is very stable.

First, low Dbr1 mRNA levels in E. histolytica, as in A. thaliana, suggest that their RNA turnover is fast. Particularly, in amoeba, the mRNA half-life is influenced by the length of the 3′ end poly(A) tail, modified by stress conditions (Lopez-Camarillo et al., 2003, 2014). The poly(A) tail length of EhDbr1 mRNA is unknown, but it will be important to determine this signature in the transcript to know if it is a determinant for the low accumulation observed. We do not know if other E. histolytica genes carry out a similar rate of RNA turnover.

Second, compared with the forward K<sup>M</sup> values of constitutive glycolytic amoebic enzymes such as hexokinase (∼33 µM for glucose), GAPDH (∼33µM for glyceraldehyde-3-phosphate) or the fructose bisphosphate aldolases that shows the highest affinities for their substrate Fru(1,6)P<sup>2</sup> (∼4µM) (Saavedra et al., 2005), the K<sup>M</sup> value for Dbr1 is 0.2–0.5µM for short-branched RNAs (Garrey et al., 2014; Clark et al., 2016). These K<sup>M</sup> values suggest that debranching enzymes have a high specific activity, as previously proposed (Kim et al., 2000), with turnover rates of ∼3 s −1 (Clark et al., 2016). This activity is enhanced by Dnr1, a Dbr1 homolog protein, that directly interacts with the lariat (Garrey et al., 2014).

Thus, all of these data support the idea that EhDbr1 gene expression regulation seems to be finely tuned and conserved across eukaryotes, such that cells synthesize few mRNA molecules that are translated into robust Dbr1 enzymes, constituting one of the key (and probably the most critical) factors that regulate intron lariat turnover.

Surprisingly, the EhDbr1 enzyme is present until trophozoite cultures die, suggesting that the half-life of the EhDbr1 enzyme may be longer than the trophozoite cell doubling time. We suspect that the longer half-life of this enzyme could be due to the absence of additional ubiquitination sites present in HsDbr1 (Figure S1), although this possibility remains to be tested. However, we can envisage the need of a robust EhDbr1 since, even though E. histolytica is a unicellular organism, it is an intron-rich eukaryote (Hon et al., 2013); thus, amoebic intron lariats need to be immediately open to degrade them, avoiding lariat accumulation, or to process them in ncRNAs, such as circular RNA molecules (Mendoza-Figueroa et al. this issue).

Because of the difference in the localization pattern between HsDbr1 and EhDbr1 shown by western blot analysis, we hypothesized that EhDbr1 has a weak nuclear localization signal (NLS), allowing its passage to the nucleus in minute quantities. The EhDbr1 protein sequence alignments with the bipartite NLS (VPLKRLSDEHEPEQRKKIKRRNQAIYAAVDDDDDDAA) of the HsDbr1 C-terminal domain (HsNLS), constituted by 37 amino acid residues (Kataoka et al., 2013) and the three portions (EhNLS1, 2, and 3) of the E. histolytica EhNCABP166 tripartite NLS (Uribe et al., 2012), showed that the single EhDbr1 NLS does not match perfectly to any of the NLS tested (Figure S2), supporting the notion of a weak NLS in EhDbr1 compared with the bipartite HsNLS and the tripartite EhNLS. This situation prevented the EhDbr1 from entering the nucleus as efficiently as HsDbr1 does. We do not discard the possibility that in other growth phases, EhDbr1 could be present abundantly inside the nucleus. These possibilities need further evaluation. EhDbr1 localization was analyzed by confocal microscopy. For comparison, nuclear speckles localization of the HsDbr1 in HeLa cells was included (Kataoka et al., 2013). Our results suggest that the EhDbr1 is localized on the cytoplasmic side of the nuclear periphery, appearing concentrated on the basal (toward the substrate) region. Taken together, these results suggest that the EhDbr1 might be associated with the nuclear envelope on the cytoplasmic side of the E. histolytica trophozoites, arguing for an inefficient NLS C-terminal EhDbr1 and allowing for minute amounts of nuclear EhDbr1 to be transported. It is also possible that in the growth phase analyzed, EhDbr1 mainly rests on the cytoplasmic side, while in other growth phases, it could be mainly nuclear. These hypotheses need to be evaluated.

The main function attributed to Dbr1 is to hydrolyze the 2′ - 5 ′ linkage in RNA lariats (Ruskin and Green, 1985); however, other functions have been associated with Dbr1. For example, the retrotransposon Ty1 replicates by retrotranscription, and the resultant cDNA then is then integrated into the host genome. In 1dbr1 S. cerevisiae cells, Ty1 cDNA accumulates (Chapman and Boeke, 1991; Salem et al., 2003), suggesting that ScDbr1 is involved in retrotranscription or cDNA stability (Salem et al., 2003). Experiments carried out by Salem and coworkers ruled out the possibility that the Ty1 cDNA had 2′ -5′ linkages at the 5 ′ or 3′ ends since it integrates into the genome in the 1dbr1 cells. Additionally, Pratico and Silverman demonstrated that a 2 ′ -5′ -branched RNA is not a retrotransposition intermediate, suggesting that Dbr1 binds RNA only to stabilize it or that it carries other functions different from the 2′ -5′ bond cleavage during the Ty1 retrotransposition (Pratico and Silverman, 2007). To gain more insights into the role of Dbr1 in Ty1 retrotransposition, it will be important to determine if Dbr1 associates with Ty1 RNA directly in the nucleus or in the retrosomes, the cytoplasmic foci where the virus-like particles are assembled (Pachulska-Wieczorek et al., 2016). Moreover, HIV-1 is a retrovirus resembling retrotransposons. It seems that HIV-1 initiates its cDNA synthesis at the cytoplasm and then completes it in the nucleus or in the perinuclear region. The initial synthesis does not depend on HsDbr1, but its completion does. Currently, it is unknown if a lariat-like structure with a 2′ -5′ bond is formed in the HIV-1 RNA such that HsDbr1 needs to open it (Galvis et al., 2014).

Recently, Han and coworkers showed that HsDbr1 modulates snRNP recycling, allowing the enzymes to dissociate from post-catalytic complexes and to be available to form new spliceosomes complexes. If snRNP availability is compromised, the resulting effect is a change in the alternative splicing pattern that leads to tumorigenesis (Han et al., 2017).

To date, the role of Dbr1 in the cytoplasm is not clear; however, from Ty1 and HIV-1 data, it seems that Dbr1 might have a function different from that of 2′ -5′ bond cleavage. EhDbr1, as well as other Dbr1 enzymes, may have a conserved cytoplasmic function not described thus far.

# REFERENCES


# CONCLUSION AND FUTURE DIRECTIONS

In this work, we describe that EhDbr1 mRNA has a very short half-life (less than 30 min) and encodes a very stable protein that is present until trophozoites cultures die. Moreover, in a steady state, E. histolytica trophozoite EhDbr1 seems to be localized in the nuclear periphery, probably because of the lack of a well-defined nuclear localization signal. These data and data from other organisms suggest that Dbr1 gene expression is finely tuned and conserved across eukaryotes. Experiments that describe the aspects of EhDbr1 gene transcription and EhDbr1 mRNA turnover, as well as those describing other functions of the protein both in the nucleus and cytosol, need to be performed.

# AUTHOR CONTRIBUTIONS

MM-F conceived, designed, acquired data, carried out the experiments, analyzed data and drafted the manuscript; CO-P participated in the design of the study and drafted the manuscript; OS-C drafted the manuscript; JV conceived and designed the study, analyzed data and drafted the manuscript. All authors read and approved the final manuscript.

# ACKNOWLEDGMENTS

The authors thank Dr. Esther Orozco from Departamento de Infectómica y Patogénesis Molecular-CINVESTAV for the antibody anti-EhCPADH.

This work was supported by Consejo Nacional de Ciencia y Tecnología (CONACyT, México) grants 127557-M and 236104. MM-F received the CONACyT fellowship 46417.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00228/full#supplementary-material


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Valdés, Ortuño-Pineda, Saucedo-Cárdenas and Mendoza-Figueroa. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Postsplicing-Derived Full-Length Intron Circles in the Protozoan Parasite *Entamoeba histolytica*

María S. Mendoza-Figueroa<sup>1</sup> , Eddy E. Alfonso-Maqueira<sup>1</sup> , Cristina Vélez <sup>2</sup> , Elisa I. Azuara-Liceaga<sup>3</sup> , Selene Zárate<sup>3</sup> , Nicolás Villegas-Sepúlveda<sup>4</sup> , Odila Saucedo-Cárdenas 5,6 and Jesús Valdés <sup>1</sup> \*

<sup>1</sup> Departamento de Bioquímica, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico City, Mexico, <sup>2</sup> Departamento de Biología Celular, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico City, Mexico, <sup>3</sup> Posgrado en Ciencias Genómicas, Universidad Autónoma de la Ciudad de Mexico, Mexico City, Mexico, <sup>4</sup> Departamento de Biomedicina Molecular, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico City, Mexico, <sup>5</sup> Departamento de Histología, Facultad de Medicina, Universidad Autónoma de Nuevo León, Monterrey, Mexico, <sup>6</sup> División de Genética, Centro de Investigación Biomédica del Noreste, Instituto Mexicano del Seguro Social, Monterrey, Mexico

Noncoding circular RNAs are widespread in the tree of life. Particularly, intron-containing circular RNAs which apparently upregulate their parental gene expression. Entamoeba histolytica, the causative agent of dysentery and liver abscesses in humans, codes for several noncoding RNAs, including circular ribosomal RNAs, but no intron containing circular RNAs have been described to date. Divergent RT-PCR and diverse molecular approaches, allowed us to detect bona fide full-length intronic circular RNA (flicRNA) molecules. Self-splicing reactions, RNA polymerase II inhibition with Actinomycin D, and second step of splicing-inhibition with boric acid showed that the production of flicRX13 (one of the flicRNAs found in this work, and our test model) depends on mRNA synthesis and pre-mRNA processing instead of self-splicing. To explore the cues and factors involved in flicRX13 biogenesis in vivo, splicing assays were carried out in amoeba transformants where splicing factors and Dbr1 (intron lariat debranching enzyme 1) were silenced or overexpressed, or where Rabx13 wild-type and mutant 5′ ss (splice site) and branch site minigene constructs were overexpressed. Whereas SF1 (splicing factor 1) is not involved, the U2 auxiliary splicing factor, Dbr1, and the GU-rich 5′ ss are involved in postsplicing flicRX13 biogenesis, probably by Dbr1 stalling, in a similar fashion to the formation of ciRNAs (circular intronic RNAs), but with distinctive 5′ -3′ ss ligation points. Different from the reported functions of ciRNAs, the 5′ ss GU-rich element of flicRX13 possibly interacts with transcription machinery to silence its own gene in cis. Furthermore, introns of E. histolytica virulence-related genes are also processed as flicRNAs.

Keywords: ncRNA, circular RNA, Dbr1, virulence, splicing, 5′ss, branch point sequence

# INTRODUCTION

Circular RNAs (circRNAs) are a group of noncoding RNAs (ncRNAs) that originated from protein coding genes. They were described nearly 40 years ago through electron microscopy studies (Hsu and Coca-Prados, 1979); however, their prevalence in different cellular conditions, biogenesis and possible functions have been explored only recently with the advent of massive sequencing of the transcriptomes.

### *Edited by:*

Serge Ankri, Technion–Israel Institute of Technology, Israel

#### *Reviewed by:*

Xiuzhu Dong, Institute of Microbiology (CAS), China Maria Del Consuelo Gomez, Instituto Politécnico Nacional, Mexico

> *\*Correspondence:* Jesús Valdés jvaldes@cinvestav.mx

*Received:* 20 March 2018 *Accepted:* 04 July 2018 *Published:* 03 August 2018

#### *Citation:*

Mendoza-Figueroa MS, Alfonso-Maqueira EE, Vélez C, Azuara-Liceaga EI, Zárate S, Villegas-Sepúlveda N, Saucedo-Cárdenas O and Valdés J (2018) Postsplicing-Derived Full-Length Intron Circles in the Protozoan Parasite Entamoeba histolytica. Front. Cell. Infect. Microbiol. 8:255. doi: 10.3389/fcimb.2018.00255

Most of them still lack a specific function but it has been described that some may act as postranscriptional regulators in the cytoplasm or may act as cotranscriptional regulators in the nucleus (Hansen et al., 2013; Memczak et al., 2013; Ashwal-Fluss et al., 2014; Li et al., 2015), suggesting that circRNAs have important functions in modifying gene expression.

The best studied circRNAs are exon-coded cytoplasmic microRNA sponges (Hansen et al., 2013; Memczak et al., 2013), while others consist of exons that include full intronic sequences, named EIciRNAs, which compete with splicing to control their parental gene expression in cis via their interaction with RNA Pol II and the U1 small nuclear RNA at their respective promoter regions (Ashwal-Fluss et al., 2014; Li et al., 2015). The biogenesis of circRNAs and EIciRNAs involves backsplicing or is mediated by an exon-containing lariat precursor, pathways favored by complementary sequences, such as ALU repeats, in the introns bracketing the circularized exons and in some cases mediated by the splicing factors muscleblind or quaking (Jeck et al., 2013; Memczak et al., 2013; Ashwal-Fluss et al., 2014; Barrett et al., 2015; Conn et al., 2015; Ivanov et al., 2015).

Cytoplasmic (and some nuclear) stable intronic sequence RNAs (sisRNAs) and ciRNAs conform an additional type of splicing-related circular RNA molecules that escape debranching reaction of intron lariats (Zhang et al., 2013; Talhouarne and Gall, 2014). Cytoplasmic sisRNAs are lariats lacking the 3′ terminal tail (approximately 30 nt upstream the 3′ ss) and are present during early embryogenesis in the cytoplasm of Xenopus tropicalis. It has been proposed that they avoid debranching before leaving the nucleus since they are opened by nuclear, but not cytoplasmic, extracts of Xenopus laevis (Talhouarne and Gall, 2014). Conversely, ciRNA production in human cells depends on the presence of a GU rich sequence element close to the 5′ ss and a C-rich element near to the branch site (BP). These elements are not present in introns that are not circularized, and more importantly they enable the lariats that contain such elements to escape debranching. Therefore, Dbr1 appears to be involved in the biogenesis of sisRNAs and ciRNAs through particular mechanisms not yet described. The ciRNAs interact with the phosphorylated RNA Pol II to activate the transcription of its parental genes (Zhang et al., 2013).

To date, the best characterized circular RNA molecules have been described in multicellular eukaryotes, yeasts, Plasmodium falciparum and Dictyostelium discoideum (Wang et al., 2014), but not in other protists of clinical relevance, such as Entamoeba histolytica, the etiologic agent of amoebiasis which still causes over 55,000 annual deaths worldwide, and 2.2 million disability-adjusted life years (Lozano et al., 2012; Murray et al., 2012). However, E. histolytica is not devoid of regulatory ncRNAs which include stressed-induced selfcircularized 5′ -external transcribed spacer rRNAs (Gupta et al., 2012), and microRNAs (De et al., 2006; Mar-Aguilar et al., 2013).

Entamoeba histolytica possess small introns of ≈75 nt in length with extremely conserved 5′ and 3′ ss, GUUUGUU and UAG, respectively (Wilihoeft et al., 2001; Davis et al., 2007; Lorenzi et al., 2010; Hon et al., 2013). The BP sequence (YNYYRAY) of amoebic introns has been identified only in silico and lacks the conservation maintained in other protozoans (Wilihoeft et al., 2001; Vanácová et al., 2005; Davis et al., 2007).

Here, we explored the existence of E. histolytica splicing products similar to the vertebrate circRNAs, EIcircRNAs, or ciRNAs. For the first time we report circular ncRNA molecules from different amoeba gene products, including some overexpressed in virulence. Different from their vertebrate counterparts, these are conformed by full-length intronic circular RNAs (flicRNAs). Characterization of the intron circle of RabX13 (flicRX13) showed that it originates from transcription/splicing of its parental gene and is more stable than the IR (intron retained), spliced and lariat variants that originated thereof. FlicRX13 biogenesis can be attributed to the intronic 5′ ss GUrich element and possibly to Dbr1 activity, which we are still investigating. Due to their distinctive 5′ -3′ ss ligation points, we propose that flicRNAs might arise from stalled Dbr1 with the aid of additional factors. Different from the reported functions of ciRNA, minigene intron constructs carrying 5′ ss GU-rich element mutants evidenced a cis-regulatory role of flicRX13 silencing its parental gene, possibly via interaction with factors of the transcription machinery.

# MATERIALS AND METHODS

# *E. histolytica* Cultures and Drug Treatments

Trophozoites of the E. histolytica strain HM1:IMSS were axenically cultured in Trypticase-yeast extract-iron serum (TYI-S-33) medium at 37◦C and harvested as described (Diamond et al., 1995). For in vivo assays (splicing, inhibition of second step of splicing or Dbr1 inhibition), 10<sup>6</sup> log phase (48 h) wild type or transfectant trophozoites per experimental point, either alone or treated with 1.3µM Actinomycin D (Sigma Aldrich) as described (López-Camarillo et al., 2003), or with different concentrations of boric acid (pH 7.9) (Shomron and Ast, 2003) during 1.5 h. After treatments, cells were harvested, and total RNA was isolated and analyzed by reverse transcription (RT) followed by polymerase chain reaction (PCR) amplification of the DNA.

# RNA and DNA Isolation

Genomic DNA (gDNA) and total RNA were isolated using the TRIzol Reagent as specified by the manufacturer (Invitrogen). For RNA isolation, preparations were treated with RQ1 DNase (Promega) or with RNase R (Epicentre), when specified. For gDNA purification, preparations were treated with RNases T1 and A.

# Retro Transcription and Polymerase Chain Reactions

All primers and PCR conditions are listed in **Table S1**. RabX13 (EHI\_065790), U6 snRNA (U43841), 18S rRNA (AB426549), RNA Pol II (EHI\_087360), and EhActin (EHI\_107290) gene expression were monitored by RT-PCR using specific primer pairs, M-MLV retro transcriptase, and Taq DNA Polymerase, as specified by the manufacturer (Invitrogen). All PCR products were cloned into the pCR2.1 plasmid vector (Invitrogen) and sequenced.

To detect circular RNA molecules, specific outward facing primer pairs targeted to the introns of the E. histolytica genes RabX13, rpL12 (EHI\_191750), rpS14 (EHI\_074090), Cdc2 (L03810), and intron 2 of ClcB (EHI\_186860), respectively, were designed as described (Vogel et al., 1997). The primers were used in circular RT-PCR using the aforementioned conditions. In some cases, PCR reactions were carried out with the 5′ end radiolabeled Rab2BSs oligonucleotide. Radiolabeling was carried out with the T4 PNK enzyme (New England BioLabs) as suggested by the manufacturer. Where indicated, RT reactions were carried out in the presence of 5µM Actinomycin D (Sigma-Aldrich), adding the drug immediately after the denaturing step (Houseley and Tollervey, 2010). PCR products were resolved in 8% urea-polyacrylamide gels. The Kappa Syber Fast Universal One-Step qRT-PCR kit (Sigma Aldrich) was used for quantitative RT-PCR with 10 ng of cDNA input in 10 µL.

# SF1, U2AF, and DBR1 Plasmid Constructs and Entamoeba Transformants

Using E. histolytica gDNA as a template and the respective primer pairs in which the SmaI and XhoI restriction sites were included, the full-length and C-terminus deleted (1C) U2AF and Dbr11C, were amplified by PCR and ligated into the SmaI/XhoI-digested pEhExHA expression plasmid able to express N-terminal hemagglutinin (HA)-tagged fusion protein (Saito-Nakano et al., 2007). The wild type and mutant RabX13 minigenes were purchased to GeneScript (**Table S2**) and subcloned into pEhExHA. The full-length EhSF1 gene was amplified by PCR and ligated in inverted orientation into the SmaI/XhoI-digested silencing pKT3M plasmid (Morf et al., 2013). Plasmids were transfected into E. histolytica HM1:IMSS trophozoites as described (Nozaki et al., 1999). To avoid lethality, 24 h after transfection the expression of HA-Dbr11C fusion proteins in transient transfectants was induced by the addition of 1.5µg/mL during 48 h. Stable HA-U2AF, HA-U2AF1C, and pK-SF1 transfectants were established by culturing in 3, 6, or 10µg/mL G418 as above. In vivo splicing assays were carried out with purified RNA from log phase cultures of transformants to detect RabX13 unspliced, spliced, lariats and circular RNA variants.

# Western Blots

HA-Dbr1, HA-U2AF1C proteins as well as histone H3 were detected by western blot assays using the anti-HA (Covance) and anti-histone H3 (Cell Signaling) antibodies diluted 1:25,000 and 1:1,000, respectively. SF1 was detected with anti-human SF1 antibodies (Santa Cruz) diluted 1:1,000.

# Statistical Analyses

The intensity of each splice isoform was measured using NIH ImageJ software from at least three RT-PCR experiments and analyzed using Student's t-test and One-Way ANOVA with SigmaPlot software. For reference, the mean value of each isoform in the absence of Boric acid or at the zero time points was set to 1. qRT-PCR data was analyzed using the 2−11C<sup>T</sup> method (Livak and Schmittgen, 2001).

# RESULTS

# Identification and Validation of Full-Length Intron Circular RNAs

To detect circular intronic noncoding RNAs, outward facing primers (**Figure 1A** and **Table S1**) targeted to five genes of different functions and intron sizes were used in circular RT-PCR (Vogel et al., 1997): the Rab GTPase family member RabX13 (136 nt intron); the ribosomal protein S14 gene rpS14 (73 nt intron); the ribosomal protein L12 gene rpL12 (106 nt intron); the yeast homolog p34cdc<sup>2</sup> Cdc2 (79 nt intron); and the chloride anion channel ClcB (57 nt intron 1 and 81 nt intron 2) (Salas-Casas et al., 2006). These genes were selected because their intron size is largerthan the mean of E. histolytica introns, therefore facilitating their study. Like intron 1 of ClcB, the majority of introns initially chosen for this study were too small and/or were A-T rich and were left out because were not amenable to such analyses.

Surprisingly no sequence reads of the introns analyzed corresponded to branched molecules showing a typical missincorporation of an A for T at the lariat's BP (Vogel et al., 1997; Barrett et al., 2015). Instead, the sequences corresponded to full-length intronic circularized RNAs (flicRNAs), with ligated G residues of the ends of the introns, 6–15 nt larger than the products expected from the lariat amplicons (**Figures 1B**,**C**, black arrowheads). No 3′ tail-less lariat structures such as those observed in vertebrates or other types of branched lariats (Li-Pook-Than and Bonen, 2006; Gao et al., 2008; Taggart et al., 2012, 2017; Zhang et al., 2013) were detected. The remaining amplicons (white asterisks) did not match their respective intronic sequences. None of the Cdc2 reads were informative; therefore, this intron was not analyzed further. Since the Rabx13 flicRNA (flicRX13) is readily detected in different Entamoeba strains, it was chosen for further characterization and analyses.

Experiments were carried out to discard any given condition that could render artifactual 5′ -3′ ss joined amplicons. First, to rule out intron tandem genomic templates, we compared the products obtained in circular RT-PCR experiments using DNase I-treated total RNA, with and without RT, and RNase A+T1 treated gDNA without RT reactions. FlicRX13 was detected only in plus RT DNase I-treated total RNA samples (**Figure 2A**). Second, to eliminate the possible inter-molecular trans-splicing origin of flicRNAs, cDNA synthesis was carried out with the primer EhActR or Rab2BSas, targeted to the EhActin gene and to the RabX13 intron, respectively. Then, independent PCR reactions were carried out with each cDNA, using the respective reverse primers employed in the RT reaction (EhActR or Rab2BSas) and the forward primers targeted to different transcripts (e.g., rpS14) as shown in **Figure 2B**. Except for the EhActin controls (lane 1), no PCR products were detected. Third, to discard possible self-ligation events due to diluted RNA samples, we used increasing amounts of total RNA input in the circular RT-PCR experiments and observed a proportional increase in flicRX13 amplification (**Figure 2C**). To eliminate artifactual inter- and intra-molecular RT jumps, cDNA was synthesized in the presence of Actinomycin D (Houseley and Tollervey, 2010) with increasing amounts of total RNA inputs. As expected, Actinomycin D increased flicRX13 specific

amplification as a function of the RNA input (**Figure 2D**). Together these experiments suggest that at least flicRX13 does not result from noncanonical trans-splicing events.

To verify the circularity of the intronic RNA molecules, additional criteria were considered. First, stringent northern blot hybridization was performed with a probe spanning the 3 ′ ss G-5′ ss G ligation point of flicRX13 (dashed arrow in **Figure 2E**). Next, total RNA was treated with RNAse R to degrade linear RNA (Suzuki et al., 2006; Burd et al., 2010; Salzman et al., 2012; Jeck et al., 2013; Ashwal-Fluss et al., 2014) and was used to perform circular RT-PCR using primer Rab2BSas (black arrow in **Figure 2F**) for cDNA synthesis. Only a specific PCR product corresponding to flicRX13 was amplified with the probe and sense primer (dashed and gray arrows respectively). Furthermore, as expected for intronic circular ncRNA (Zhang et al., 2013) we were able to amplify flicRNAs from nuclear RNA and from RNA extracted with miRNA isolation kits. Altogether, our results support the notion that E. histolytica flicRNAs are bona fide circular intronic RNA molecules.

# Postsplicing Origin of flicRNAs

Circular RNA molecules are more stable than their corresponding splicing variants (Jeck et al., 2013). This is also true for flicRX13, since this product appears early in the growth phase and is maintained throughout a 72-h standard growth culture (not shown) incubated in the presence of Actinomycin D at different time points. After 3 h of transcription inhibition, RabX13 pre-mRNA and spliced variants nearly disappear while the flicRNA population remained up to 6 h of treatment (**Figures 3A,B**). As previously observed (López-Camarillo et al., 2003), the expression of EhActin remained unchanged. U6 snRNA and 18S rRNA were also not affected, indicating that RNA polymerases III (Miranda et al., 1996) and I (Oakes et al., 1993), respectively, were not affected (**Figure 3C**). These results show the different stability between pre-mRNA, mRNA and flicRX13, and that the half-life of flicRX13 is greater than that of its parental mRNA.

To explore the splicing or postsplicing origin of flicRNAs we first silenced the E complex factor SF1 (Loftus et al., 2005; Valdés et al., 2014) and analyzed flicRX13 abundance

in the transfected amoebas. After confirming SF1 mRNA depletion by RT-PCR, we observed that reduced levels of SF1 moderately increased intron retained/spliced precursor and RabX13 spliced variants but not flicRX13, demonstrating its role in splicing, but not in flicRNA production (**Figure 4A**). We next overexpressed the E complex U2 auxiliary splicing factor, U2AF. Full-length U2AF slightly increased flicRNA formation in response to increased overall synthesis of unspliced and spliced RabX13 transcripts; however, overexpression of the C-terminus deleted U2AF strongly increased both splicing and flicRNA production (**Figure 4B**), suggesting that flicRNAs are byproducts of splicing after complex E has been assembled.

To seek whether flicRNA formation occurs during or after the first step of splicing, we compared the abundance of unspliced pre-mRNA/intron retained, spliced mRNA, lariats and flicRNAs in RNA extracts from Entamoeba cultured in boric acid, an inhibitor of the second step of splicing (Shomron and Ast, 2003). By circular RT-PCR using a radioactive primer, we detected RabX13 and rpS14 lariats with the predicted BP (**Figure S1** and **Figure 4C**) and with the expected 6–11 nt size, which is smaller compared to flicRNAs. As expected, 5 mM boric acid significantly affected the amplification of all splicing products (ANOVA P ≤ 0.001), particularly flicRX13. However, with 10 mM boric acid, a slight recovery of flicRX13 and lariat molecules (including actual

and cryptic BP lariats, and lariat-3′ exon intermediaries) was observed at the expense of the spliced products (**Figure 4C**). These results suggest that flicRX13 is another final product of splicing, most likely originated by the circularization of lariats. Because in vivo boric acid treatment is a novel experimental approach, the accumulation of lariats in boric acid treated mammalian cell cultures was validated as well. Lariats of the HPV16 proximal 3′ ss accumulated in infected HeLa cells only (**Figure S1**).

FIGURE 4 | Inhibition of the second step of splicing alters flicRNA production. (A) Stable SF1-silencing (pK-SF1) and empty pKT3M vector (pK) transfectants were established at 3 and 6µg/mL G418. Silencing was assessed by western blot (top) achieving 40–90% efficiency. Absence of SF1 slightly increased pre-mRNA/unspliced and mRNA Rabx13 transcripts (middle), and flicRX13 production (bottom); however, such increments were not statistically significant (right plot). (B) Amoeba transformants overexpressing wt (HA-U2AF) and carboxy-terminal deleted (HA-U2AF1C) U2 auxiliary splicing factor were established. Appropriate expression of the hemagglutinin-U2AF fusion proteins were monitored and compared to histone H3 expression by western blot (top). Compared to the empty Entamoeba HA expression plasmid (pHA) pHA-U2AF increased the overall expression of all RabX13 transcription products. However, HA-U2AF1C substantially increased mRNA (spliced) products and flicRX13. RNA polymerase II expression was equivalent between the transformants. (C) Expression of the RabX13 gene transcript variants (unspliced, spliced, flicRX13 and intron lariat) after incubation of E. histolytica trophozoites with different concentrations of boric acid, an inhibitor of the second step of splicing. 18S rRNA expression was used for normalization purposes. To detect intron lariats, circular RT-PCR was carried out with the <sup>32</sup>P-labeled Rab2BSas primer (Rab2BSas\*), which was used as the loading control. Bottom panel, products in the upper panel were quantitated and statistically analyzed. Decrease of spliced RabX13 variant and flicRX13 were statistically significant (ANOVA P ≤ 0.001, asterisks in bold). Values are the average of three independent experiments.

# Elements and Factors Involved in flicRNAs Biogenesis

Structurally, flicRNAs resemble circular group I and group II introns. We did not explore the possible origin of flicRNAs by self-splicing reactions homologous to group I introns because E. histolytica introns have no discernible internal guide sequences that facilitate self-splicing (Winter et al., 1990), and we did not observe circle ligation points containing guanine triplets, that would result by the attack of an exogenous αG to the 5 ′ ss, followed by the 5′ exon attack to the 3′ ss as previously reported (Hausner et al., 2014). To explore whether flicRNAs are self-spliced such as group II introns, synthetic radiolabeled RabX13 and ai5γ group II intron transcripts were incubated in the appropriate buffers. Whereas the intron-3′ exon intermediate and mRNA were produced from the ai5γ intron transcript, only intact RabX13 transcripts were observed suggesting that flicRNAs might not originate through this mechanism (**Figure S2**). However, since RabX13 intron does not have the RNA elements required for auto excision (data not shown), we cannot completely discard this route of intron circularization in E. histolytica.

To escape from the debranching reaction, ciRNA biogenesis requires a GU-rich 5′ ss and a C-rich element upstream of the BP (Zhang et al., 2013). Therefore, to explore whether these elements influence flicRNA formation, minigene intron Entamoeba transfectants were established harboring the wt intron, a minus GU-rich 5′ ss (1GU) intron, an intron containing a C-rich element 11 nt upstream of the BP (C; uncommon in the Entamoeba genome), and an intron with both modifications (1GU-C), all of which are flanked with partial exon sequences tagged with plasmid-specific sequences. The different RNA species transcribed from the transfected plasmids were monitored by qRT-PCR with primers CS+116 and Rab-REx2. The elimination of the GU-rich element at the 5′ ss affected flicRNA formation even in the presence of the synthetic Crich element. The C-rich element alone has a similar effect but not as strong as the 5′ ss-GU element indicating that the GU-rich element is sufficient to promote intron circularization (**Figure 5A**).

Dbr1 participates in sisRNAs and ciRNAs biogenesis (Zhang et al., 2013; Talhouarne and Gall, 2014). To explore the possible E. histolytica Dbr1 involvement in flicRNAs biogenesis, we attempted siRNA-mediated Dbr1 silencing without success. Also, the establishment of the catalytically deficient Dbr1 (Dbr11C) overexpressor amoebas was unsuccessful. Therefore, in vivo splicing assays were conducted using Dbr11C transiently transfected trophozoites under low selection pressure (1.5µg/mL G418). After assessing Dbr11C expression by western blot, the different RabX13 splicing products were monitored (**Figure 6**). Surprisingly, compared to the control Dbr11C overexpression increased flicRNA levels. Taken together, our data indicate that E. histolytica flicRNAs originate from lariat molecules after the second step of splicing has taken place, and that the GU-rich 5′ ss and Dbr1 are implicated in their biogenesis. However, the implication of Dbr1 in this process might be indirect or require additional factors.

# Insights Into flicRNAs Possible Functions

circRNAs bound to the U1 snRNA promote the transcription of their parental genes, which interact with the RNA polymerase II at the promoter or during elongation (Zhang et al., 2013; Li et al., 2015). Our experimental design with the minigene constructs allowed us to explore the possible function of

Entamoeba flicRNAs irrespective of the snRNA involved. Again, in vivo splicing assays were performed using RNA from minigenes transformed amoebas and the different endogenous RNA species were monitored. Unexpectedly, mutations in the 5′ ss-GU element resulted in an increase of endogenous RabX13 expression (**Figure 5B**), indicating that such elements of flicRX13 silences the expression of its parental gene in cis.

Finally, to explore the general nature of flicRNAs in the Entamoeba ncRNA repertoire, we sought additional flicRNAs, particularly those involved in liver abscess formation (Meyer et al., 2016). We detected flicRNAs from the gene products in the loci EHI\_014170 (intron 1), EHI\_169670 (both introns), and the mono-intronic EHI\_192510 (**Figure 7**), suggesting that flicRNAs are common ncRNA species in E. histolytica involved in gene expression regulation.

# DISCUSSION

Here, we identified full-length intronic circular RNAs or flicRNAs, whose 5′ ss is ligated to the 3′ ss, in the protozoan parasite E. histolytica, which differs from other parasitic or mammalian circular RNAs. For example, D. discoideum and P. falciparum circular RNAs contain exons and some, are related to exon-skipping splicing events (Wang et al., 2014), indicating that in protozoa there might exist a wide repertoire of circular noncoding RNAs that originated by distinct mechanisms in protozoa.

flicRNA production. E. histolytica trophozoites were transiently transfected with the empty vector pEhExHA (pHA), or with the C-terminus deleted Dbr1 (pHA-Dbr11C) plasmids and 24 h latter were tested under low selection pressure (1.5µg/mL G418). Overexpression of Dbr1 (wt or 1C) was monitored by western blot using anti-HA antibody compared to anti-histone H3 (loading control). The effect of Dbr1 (wt or 1C) overexpression on flicRX13, RabX13 mRNA, and RNA Pol II (loading control) gene expression was analyzed by RT-PCRs.

# flicRNAs Biogenesis in *E. histolytica*

Exon-intron-containing circular RNAs are generated by backsplicing, mainly favored by protein-mediated or sequencemediated approximation of intronic repetitive sequences that flank nonconsecutive circularized exons (Ashwal-Fluss et al., 2014; Jeck and Sharpless, 2014; Conn et al., 2015; Ivanov et al., 2015). The molecular mechanisms involved are not fully understood, but it is clear that circRNA formation competes with splicing (Ashwal-Fluss et al., 2014) and, in spite of the lack of direct biochemical evidence (Barrett et al., 2015), the occurrence of some circular RNAs have been correlated to exon-skipping splicing events (Zaphiropoulos, 1996; Surono et al., 1999; Zhang et al., 2014), as reported for the exon-derived circular RNAs found in D. discoideum and P. falciparum (Wang et al., 2014). The biogenesis of E. histolytica flicRNAs characterized here does not depend on complex E splicing formation, possibly due to the strong (Larson and Hoskins, 2017) and most frequent Entamoeba 5′ ss (GUUUGUU) (Davis et al., 2007; Hon et al., 2013) since SF1 is not related to splice site definition nor to flicRNA formation. The impact of the U2 auxiliary factor in flicRNA formation indicates a closer relationship between complex A factors and the requirements for intron circularization, particularly those recruited with the entrance of the U2 snRNP, facilitated by U2AF, that not only defines the 3′ ss (Shepard et al., 2002; Förch and Valcarcel, 2003), but also contacts both RNA polymerase II and the nineteen complex (David et al., 2011), which is part of the spliceosomal intron lariat complex, along with other components of the U2 snRNP, and remain attached to the lariat after spliceosomal disassembly (Yoshimoto et al., 2009; Fourmann et al., 2013). Likewise, such impact of U2AF on flicRX13 formation supports the notion that flicRNAs originate by postsplicing events, ruling out the autosplicing route of flicRNA biogenesis.

The circular RNAs in E. histolytica are similar to some human ncRNAs. They are full-length intronic circles with joined 5′ and 3′ ss and are, thus flicRNAs. Although functionally related, mammalian EIcircRNAs are dissimilar to flicRNAs since they contain both intronic and exonic sequences (Li et al., 2015).

Circular ncRNAs are structurally similar to flicRNAs, which are 1) the tail-less ciRNAs consisting of circular intron lariats derived from splicing, which escape lariat-debranching enzyme activity, due to a 7 nt GU-rich element at the 5′ ss and a C-rich element 11 nt upstream of the BP (Zhang et al., 2013), and 2) the senescence-related and translation-inhibiting human intronoriginated circular RNAs, similar to cirRNAs and flicRNAs, whose biogenesis remains unknown (Taggart et al., 2012, 2017; Abdelmohsen et al., 2017; Panda et al., 2017).

C-rich elements are absent in the genome and the BP environment of Entamoeba introns explaining why even the addition of synthetic C elements fails to elicit flicRNA production. The closest resemblances of such elements are the NYYUAY pyrimidine-rich elements in the vicinity of the BP (Wilihoeft et al., 2001), suggesting that other cis elements are necessary for their biogenesis. The GUUUGUU 5′ ss consensus sequence of E. histolytica introns resembles the GU-rich element of ciRNAs. Here, we showed that nucleotide substitutions of this 5′ ss for a vertebrate 5′ ss consensus (GUAAGAA) almost abolishes the formation of flicRNAs, suggesting that such GUrich elements are essential for splicing and flicRNA biogenesis, probably by allowing escape from the debranching reaction. Interestingly, the artificial introduction of a C-rich BP element had no ostensible effect in flicRNA formation indicating that the Entamoeba genome adapted to a C-poor environment to carry out flicRNA biogenesis using only strong 5′ ss GU-rich elements. During splicing complex E formation, U1 snRNA binds to the 5 ′ ss (Larson and Hoskins, 2017). In E. histolytica, the U1 snRNP proteins U1A, U1C, and U170k are able to recruit the pre-mRNA processing machinery (Valdés et al., 2014) even in the absence of the U1 snRNA (Dávila López et al., 2008). However, as reported in yeast and in vitro systems (Kandels-Lewis and Seraphin, 1993; Rhode et al., 2006), it is possible that the Entamoeba U6 snRNA activates the 5′ ss and splicing due to their high degree of base complementarity. Therefore, compared to the wild type 5′ ss the changes introduced in the 1GU mutant elicited a different splicing response due to impaired 5′ ss activation. The link of the GU-rich element to debranching reaction escape is still under investigation.

Taggart and coworkers detected the first intron-derived circular molecules similar to flicRNAs by deep sequencing (Taggart et al., 2012, 2017). They observed that approximately 3% of the BP corresponded to the 3′ ss of the respective transcripts, thus circular 5′ -3′ ss ligated molecules, and, although they did not characterize such molecules further, they proposed their postsplicing origin either by a third nucleophilic attack or by a debranching-ligation event. Our results using boric acid as the second step of splicing inhibitor (Shomron and Ast, 2003) suggest a postsplicing origin of flicRX13. We cannot discard that flicRNAs may originate via a third nucleophilic attack as proposed for Group II introns (Murray et al., 2001). However, a more complex debranching-ligation reactions scenario is more likely to occur since it must involve additional factors. Whereas in situations where 2′ -5′ phosphodiester bond hydrolysis is impaired, such as in human and yeast Dbr1 deficient cells, intron lariats accumulate (Montemayor et al., 2014; Han et al., 2017), overexpression of an equally impaired Dbr1 in amoeba transformants, unexpectedly increased flicRX13 formation, strengthening the notion that flicRNAs originate from accumulated lariats and that other factors might participate in circularization of Entamoeba introns.

BP conformation is essential for stable recognition and catalysis by Dbr1 CTD and lariat recognition loop (LRL) domains; therefore it is possible that CTD-less Dbr1 might have hydrolyzed the BP and, due to the unstable LRL-BP interaction and the 5′ ss GU-rich element, the intron was now exposed to yet to be identified circularization factors, RNA ligase for example, in addition to the intron lariat large complex (U2, U5 and U6 snRNPs, Ntr1, Prp43, and Dnr1) (Yoshimoto et al., 2009; Fourmann et al., 2013; Garrey et al., 2014). Finally, removal of the 3 ′ tail of the lariat precedes debranching (Chapman and Boeke, 1991; Salem et al., 2003), and we detected that the 3′ ss of RabX13 and rps14 lariats are less than 10 nt downstream of the BP, placing them within the LRL-BP recognition groove and protecting them from exonucleolytic degradation. Altogether, these observations support the role of Dbr1 in flicRNA biogenesis and suggest that debranching limits access/competes with other necessary factors for intron circularization. We cannot rule out the possibility that flicRNAs arise from a third nucleophilic attack even though such reaction has not been reported in postsplicing events.

# flicRNAs Possible Functions

Cytoplasmic exon-containing circRNAs function as miRNA sponges (Jeck and Sharpless, 2014; Chen, 2016). Conversely, the small (thus without space to accommodate miRNA target sites), and nuclear E. histolytica flicRNAs most likely have different functions. It is currently accepted that intron containing linear and circular RNAs are retained in the nuclei by a similar mechanism (Chen, 2016). Although by mechanisms not fully understood, intronic circular RNAs participate in gene transcription regulation (Chen, 2016). CLIP-Seq experiments identified intronic circular RNAs associated with Pol II (Li et al., 2015), and the depletion of the abundantly expressed ciRNAs such as ci-ankrd52 and ci-sirt7 results in lower transcription of their parental genes (ANKRD52 y SIRT7) indicating that they promote their transcription by interacting with the Pol II elongation complex (Zhang et al., 2013). Other nuclear circRNAs interact with the U1 snRNP and Pol II on the promoter of their parental genes, stimulating their transcription. Both knockdown of these circRNAs as well as disruption of U1 snRNA-circRNA binding with morpholinos anti-U1 diminishes transcription of their parental genes (Li et al., 2015). In contrast, expression of 1GU minigenes showed that Entamoeba flicRX13 apparently inhibits transcription of its parental gene, in which the 5′ ss GU-rich element again seems to be actively involved. Although direct evidence of flicRNA-RNA Pol II interaction still lacks, our findings suggest an inhibitory role of flicRNAs on transcription. Recently, it has been shown that small regulatory RNAs inhibit RNA polymerase during elongation in the nematode C. elegans (Guang et al., 2010). Interestingly the addition of a C-rich BP element positively affected transcription in cis probably due to poly C binding hnRNPs (Chaudhury et al., 2010; Han et al., 2010), which are able to bind to such elements in the transcripts,

favoring their expression and transcription over other genes (Kim et al., 2005; Meng et al., 2007).

Current experiments are directed to support our present models in which silencing of Rabx13 transcription might occur via flicRX13-U6 snRNA binding or flicRNA-U1A/TIA-1/TIAR binding prior interaction with Pol II at the promoter or during elongation. In conclusion flicRNAs are common ncRNA species in E. histolytica that are involved in gene expression regulation of different regulatory processes including virulence traits.

# AUTHOR CONTRIBUTIONS

MM-F, EA-M, NV-S, and JV conception and experimental design. MM-F, EA-M, CV, EA-L, SZ, NV-S, OS-C, and JV

# REFERENCES


data acquisition and interpretation. MM-F and JV manuscript preparation.

# FUNDING

The authors are grateful to Tomoyoshi Nozaki and Upinder Singh for plasmids. Funding was provided by CONACYT (Grants 127557-M and 236104).

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00255/full#supplementary-material


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer MDCG declared a shared affiliation, with no collaboration, with several of the authors, MM-F, EA-M, CV, JV, NV-S, to the handling Editor.

Copyright © 2018 Mendoza-Figueroa, Alfonso-Maqueira, Vélez, Azuara-Liceaga, Zárate, Villegas-Sepúlveda, Saucedo-Cárdenas and Valdés. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

, Lilia López-Canovas <sup>1</sup>

,

,

,

, Abigail Betanzos 3,4

, Esther Orozco<sup>4</sup>

# Telomeric Repeat-Binding Factor Homologs in *Entamoeba histolytica*: New Clues for Telomeric Research

Francisco Javier Rendón-Gandarilla1†, Víctor Álvarez-Hernández <sup>1</sup>

, Jesús Valdés <sup>2</sup>

, Anel Lagunes-Guillen<sup>4</sup>

Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico City, Mexico

Elizabeth J. Castañeda-Ortiz <sup>1</sup>

Rosa Elena Cárdenas-Guerra<sup>1</sup>

Bibiana Chávez-Munguía<sup>4</sup>

#### *Edited by:*

Anjan Debnath, University of California, San Diego, United States

#### *Reviewed by:*

Richard McCulloch, University of Glasgow, United Kingdom Dipak Manna, Stanford University, United States

#### *\*Correspondence:*

Elisa Azuara-Liceaga elisa.azuara@uacm.edu.mx

#### *†Present Address:*

Francisco Javier Rendón-Gandarilla, Centro Regional de Educación Superior, Campus Zona Norte, Universidad Autónoma de Guerrero, Taxco Guerrero, Mexico

#### *Specialty section:*

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

*Received:* 30 April 2018 *Accepted:* 10 September 2018 *Published:* 02 October 2018

#### *Citation:*

Rendón-Gandarilla FJ, Álvarez-Hernández V, Castañeda-Ortiz EJ, Cárdenas-Hernández H, Cárdenas-Guerra RE, Valdés J, Betanzos A, Chávez-Munguía B, Lagunes-Guillen A, Orozco E, López-Canovas L and Azuara-Liceaga E (2018) Telomeric Repeat-Binding Factor Homologs in Entamoeba histolytica: New Clues for Telomeric Research. Front. Cell. Infect. Microbiol. 8:341. doi: 10.3389/fcimb.2018.00341 and Elisa Azuara-Liceaga<sup>1</sup> \* <sup>1</sup> Posgrado en Ciencias Genómicas, Universidad Autónoma de la Ciudad de Mexico, Mexico City, Mexico, <sup>2</sup> Departamento de Bioquímica, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico City, Mexico, <sup>3</sup> Consejo Nacional de Ciencia y Tecnología, Mexico City, Mexico, <sup>4</sup> Departamento de Infectómica y Patogénesis Molecular,

, Helios Cárdenas-Hernández <sup>1</sup>

Telomeric Repeat Binding Factors (TRFs) are architectural nuclear proteins with critical roles in telomere-length regulation, chromosome end protection and, fusion prevention, DNA damage detection, and senescence regulation. Entamoeba histolytica, the parasite responsible of human amoebiasis, harbors three homologs of human TRFs, based on sequence similarities to their Myb DNA binding domain. These proteins were dubbed EhTRF-like I, II and III. In this work, we revealed that EhTRF-like I and II share similarity with human TRF1, while EhTRF-like III shares similarity with human TRF2 by in silico approach. The analysis of ehtrf-like genes showed they are expressed differentially under basal culture conditions. We also studied the cellular localization of EhTRF-like I and III proteins using subcellular fractionation and western blot assays. EhTRF-like I and III proteins were enriched in the nuclear fraction, but they were also present in the cytoplasm. Indirect immunofluorescence showed that these proteins were located at the nuclear periphery co-localizing with Lamin B1 and trimethylated H4K20, which is a characteristic mark of heterochromatic regions and telomeres. We found by transmission electron microscopy that EhTRF-like III was located in regions of more condensed chromatin. Finally, EMSA assays showed that EhTRF-like III forms specific DNA-protein complexes with telomeric related sequences. Our data suggested that EhTRF-like proteins play a role in the maintenance of the chromosome ends in this parasite.

Keywords: TRF, Myb-like DNA binding domain, lamin B1, H4K20, chromosome, DNA sequence

# INTRODUCTION

Telomeres are specialized protein-DNA complexes localized at the end of eukaryotic chromosomes (Blackburn and Gall, 1978; Meyne et al., 1989; Giraud-Panis et al., 2013). Telomeric DNA consists of tandem arrays of G + T-rich repetitive sequences ending in a single-stranded G-rich overhang which is added by the ribonucleoprotein enzyme telomerase (O'Sullivan and Karlseder, 2010; Giraud-Panis et al., 2013). The length of these arrays varies among species, ranking from few hundred base pairs disposed in irregular tandem repeats in Saccharomyces cerevisiae (Lue, 2010), to thousands of base pairs of a TTAGGG repeat in vertebrate telomeres(Moyzis et al., 1988; Giraud-Panis et al., 2013). Proteins that bind to telomeric DNA play critical roles in telomere length regulation and chromosomal end protection in eukaryotic organisms. They conform a machinery known as Telosome or Shelterin complex (Palm and de Lange, 2008). These protein complexes include members or functional homologs of the TTAGGG Repeat Binding Factors, Telomeric Repeat Binding Factors (TRF) and telobox family members (Chong et al., 1995; Broccoli et al., 1997b). TRF proteins are architectural nuclear proteins involved in diverse roles, such as telomere length regulation, chromosome end protection, prevention of chromosomes fusion, sense of DNA damage, and regulation of senescence (de Lange, 2005; Palm and de Lange, 2008). These proteins are conserved from lower eukaryotes, to plants and mammals (Horvath, 2000–2013). The genome of H. sapiens contains two genes coding for TRF proteins: TRF1 and TRF2, these proteins bind as homodimers to the double-stranded DNA telomeric sequence (Chong et al., 1995; Broccoli et al., 1997b). TRF1 controls the length of telomeric repeats, whereas TRF2 is involved in the assembly of the terminal t-loop, negative telomere length regulation and chromosomal end protection (Palm and de Lange, 2008). Proteins of the TRF family have similar architectures, defined by the presence of two domains: (i) the conserved single MYB-type helix-turn-helix (HTH) DNAbinding domain (MYB DBD; 55 amino acids) located in their C-terminal; this domain contains three evenly spaced tryptophan residues and in the third α-helix presents a telobox motif (VDLKDKWRT, consensus VxxKDxxR) (Bilaud et al., 1996). (ii) The TRF-homology domain (TRFH; 200 amino acids) situated in the N-terminal, which is unique for members of this family (Broccoli et al., 1997b); and its function is related to homodimerization and protein-protein interactions with other telomeric proteins (Fairall et al., 2001). Additionally, TRF1 and TRF2 proteins diverge in their N-terminal domain, which is rich in acidic or basic residues, respectively (Broccoli et al., 1997a; Palm and de Lange, 2008). Gene encoding homologs to TRF1 and TRF2 have been found in the genomes of Trypanosoma brucei, Trypanosoma cruzi and Leishmania. major based on similarities to the C-terminal MYB DBD (Li et al., 2005; da Silva et al., 2010). Besides TRF1 and TRF2, telomeric DNA also requires the binding of other specific proteins, such as Rap1 (replication protein A 1), POT1 (protection of telomeres 1), TIN2 (TRF1 interacting nuclear factor 2) and TPP which together conform the Shelterin complex in H. sapiens (Palm and de Lange, 2008). In Trypanosomes besides the TRF2, a homolog of Rpa-1 has also been identified, suggesting that the telomeric function is conserved and that the telomeric machinery evolved early in eukaryotes (Lira et al., 2007).

In Entamoeba histolytica, the causative protozoan of human amoebiasis, the MYB DBD is the most abundant domain related to transcriptional regulation (Clark et al., 2007). MYB DBD-containing proteins in this parasite are clustered into three monophyletic groups. Families I and III are related to transcriptional factors and were dubbed as EhMybR2R3 and EhMybSHAQKYF, respectively (Meneses et al., 2010). Family II includes single-repeat proteins related to human telomeric binding proteins (Meneses et al., 2010). In E. histolytica the identification of telomeric signatures has been a challenging task. Although the first draft of E. histolytica genome was published in 2005, it has not been possible to identify sequences that referenced the terminal ends of the chromosomes neither canonical telomeric sequences nor orthologs of telomerase genes have been identified (Loftus et al., 2005; Clark et al., 2007; Lorenzi et al., 2010). However, 10% of the E. histolytica genome corresponds to tRNA genes which are associated with short tandem repeats (Loftus et al., 2005). There are 25 different types of long tandem arrays that contain between 1 and 5 tRNA types per repeat unit and STRs which resembles microsatellites (Clark et al., 2006; Tawari et al., 2008). It has been proposed that these arrays could localize at the chromosome ends, acting as telomeric regions that fulfill a structural role in the genome (Clark et al., 2006; Tawari et al., 2008). In addition, E. histolytica chromosomes do not completely condense and there is a considerable variation in their chromosome size, maybe due to expansion and contraction of telomeric repeats, as in other protists (Patarapotikul and Langsley, 1988; Melville et al., 1999; Willhoeft and Tannich, 2000). Until now, no telomeric sequences or protein complexes implicated in telomere function have been described in this parasite. The study of TRF homologs will help to gain insight into telomere biology of E. histolytica. Thus, in this work, we identified and characterized the TRF-like proteins of E. histolytica as homologous to human TRF1 and TRF2. We observed their nuclear localization in condensed chromatin regions, their co-localization with Lamin B1 and trimethylated H4K20, and their binding capacity to form DNA-protein complexes with telomeric related sequences. Our results suggest that the TRF-like proteins of E. histolytica play similar function as in their human counterparts. However, further experiments are still needed to address their role in the maintenance of telomeres in this parasite.

# MATERIALS AND METHODS

# *E. histolytica* Cultures

E. histolytica trophozoites (strain HM1:IMSS) were axenically cultured at 37◦C in TYI-S-33 medium, supplemented with 15% bovine serum (Microlab), 1% Diamond vitamin mixture (Sigma-Aldrich), 100 <sup>µ</sup>g ml−<sup>1</sup> streptomycin sulfate and 100 U ml−<sup>1</sup> penicillin (Diamond et al., 1978). Trophozoites were harvested during logarithmic growth phase.

# *In silico* Analysis of the EhTRF-Like Proteins of *E. Histolytica*

Amino acid sequences of proteins coded by genes from locus EHI\_001090, EHI\_148140 and EHI\_001110, EHI\_009820 and EHI\_074810 were obtained from AmoebaDB database (http:// amoebadb.org/amoeba/). Such sequences were used as a bait to perform queries using the DELTA-BLAST (Domain Enhanced Lookup Time Accelerated BLAST) algorithm to identify orthologs in other members of Entamoeba or other eukaryotes (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The Percent of Identity (PID) was calculated using the amino acids sequences from E. histolytica proteins coded by the above-mentioned genes, TRF1, TRF2, and proteins from organisms corresponding with the best hits, taking gaps into account using the following equation: PID = (identical positions/length of the alignment) × 100. The predicted secondary structure of complete TRF1 and EhTRF-like I, II and III (corresponding to AmoebaDB ID: EHI\_001090, EHI\_001110, and EHI\_148140, respectively) proteins was determined using PSIPRED (http://bioinf.cs.ucl.ac. uk/psipred/) tool and aligned by ClustalW2 software (https:// www.ebi.ac.uk/Tools/msa/clustalw2/). Molecular weight, pI and post-translational modifications were analyzed using the ExPaSy: Compute pI/Mw (https://web.expasy.org/compute\_pi/), ProtPi (https://www.protpi.ch/) and Mod Pred tools (http:// www.modpred.org/). Nuclear localization signals (NLS) were located through SeqNLS (http://mleg.cse.sc.edu/seqNLS/) and cNLS (http://nls-mapper.iab.keio.ac.jp/cgi-bin/NLS\_Mapper\_ help.cgi) tools. The amino acid sequence of DBD MYB from TRF proteins was aligned using ClustalW2 and phylogenetic analysis was inferred using Neighbor-Joining method. The evolutionary distances were computed using the Poisson correction method. Evolutionary analyses were conducted using MEGA 7 (Kumar et al., 2016). Phylogenetic tree was constructed through a bootstrap of 1,000 replicates. To identify orthologous components of the Sheltering machinery in E. histolytica genome, the amino acid sequences of Homo sapiens Rap1 (Q9NYB0), TIN2 (Q9BSI4), TPP1 (Q96AP0), and POT1 (Q9NUX5) were obtained from the UniProt database (http://www.uniprot.org/) and used as a bait to make queries with Blast algorithm (http:// blast.ncbi.nlm.nih.gov). The presence of conserved functional domains in the identified proteins was analyzed in the Pfam database (http://pfam.xfam.org/).

# RT-PCR Assays

Total RNA from E. histolytica trophozoites was isolated following the Trizol <sup>R</sup> LS Reagent (Invitrogen) protocol and then semiquantitative RT-PCR assays were performed with 100 ng of total RNA. Primers used to amplify the ehtrf-like I, ehtrf-like II, ehtrf-like III and 40s rps2 genes were designed using Primer-BLAST (https://www.ncbi.nlm.nih.gov/tools/primer-blast/) and manually corrected (**Table S1**). To validate the specificity of these primers, we amplified the specific sequence from 50 ng of genomic DNA, obtained with the Wizard Genomic purification kit (Promega). PCR products were separated by gel electrophoresis in 2% agarose gels, stained with RedGelTM Nucleic AcidGel Stain 10,000X (BIOTUM) and visualized in a standard UV transilluminator. Relative quantification by qRT-PCR was performed using the QuantiFast SYBR Green RT-PCR kit (Qiagen) and 50 ng of the total RNA, according to the manufacturer's instructions. Relative changes in gene expression were calculated using 40s rps2 as an internal gene calculated by 11CT method. Absolute quantitation was performed using a 10-fold serially diluted standard curve of each ehtrf-like I, II, III and 40s rps2 genes in parallel with qRT-PCR of RNA from trophozoites grown in basal conditions. Reaction volumes were set with 25 µl QuantiFast SYBR Green RT-PCR kit (Qiagen) and qRT-PCR was performed using 50 ng of the total RNA and 1µM each primer, according to the manufacturer's instructions. Initial thermal cycling conditions were 1 cycle of 50◦C for 10 mins for reverse transcription, followed by 1 cycle of 95◦C for 5 min and 40 cycles of denaturation at 95◦C for 10 s and annealing/extension temperature of 55◦C for 30 s. Plotting Ct values vs. copy number of the different genes in a standard curve allowed to approximate copies ehtrf-like genes from Ct values. The data shown was displayed as mean with standard error in triplicate and repeated in independent experiments by duplicate. GraphPad Prism 6.0e software was used for student t-test by two-tailed analyses.

# Production of Polyclonal Antibodies Against EhTRF-Like I and EhTRF-Like III

The complete amino acid sequences from EhTRF-like I and EhTRF-like III proteins were aligned using the ClustalW2 tool to identify unique regions of these proteins. Subsequently, to determine hydrophobic regions, EhTRF-like I and III amino acid sequences were analyzed using the Hopp-Woods program (Hopp and Woods, 1981). Finally, a prediction of B epitopes was made using the ABCpred server (http://crdd.osdd.net/raghava/ abcpred/). Differential peptides, CTLPSVGNALIPPS and CNKQKVQPQVSQPH for EhTRF-like I and III, respectively, were synthesized, conjugated to Keyhole Limpet Hemocyanin (KLH) (GL Biochem, Shanghai) and used to immunize New Zealand male rabbits. Before immunization, the pre-immune serum (PS) was obtained and then rabbits were subcutaneously inoculated with 400 <sup>µ</sup>g of each peptide diluted in TiterMax <sup>R</sup> Gold (Sigma-Aldrich). Four booster injections (500 µg each) at 15-days intervals, were given. After 1 week of the last immunization, rabbits were bled, and polyclonal antisera were obtained and tested by western blot assays against total extracts of E. histolytica trophozoites.

# Subcellular Fractionation

E. histolytica trophozoites were recovered by centrifugation and lysed as follows to obtain subcellular fractions. Soluble nuclear extracts (Ns) were prepared as described by Schreiber et al. (1989). Briefly, 1 × 10<sup>6</sup> trophozoites were resuspended by gentle pipetting in 400 µl of cold buffer A (10 mM HEPES pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, 0.5 mM PMSF). Trophozoites were allowed to swell on ice for 15 min, after which 25 µl of 10% NP-40 solution (Sigma-Aldrich) was added and vigorously vortexed for 30 s. Homogenate was centrifuged for 10 min at 14,000 rpm, to separate the supernatant containing cytoplasmic extracts (C), and the pellet containing nuclei. Pellet was resuspended in 1 ml of buffer A and layered on 1 ml of buffer A containing 0.34 M sucrose (Sigma-Aldrich), then mixed and centrifuged for 10 min at 14,000 rpm. The nuclear pellet was resuspended in 150 µl ice-cold buffer C (20 mM HEPES pH 7.9; 0.4 M NaCl; 1 mM EDTA; 1 mM EGTA; 1 mM DTT; 1 mM PMSF) and vigorously rocked at 4◦C for 15 min. Nuclear soluble extracts (Ns) were obtained after centrifugation for 10 min at 13,000 rpm at 4◦C and frozen in aliquots at −70◦C until used. The remaining pellet was lysed in RIPA buffer (50 mM Tris-Cl, pH 8, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100, 0.1% SDS) and sonicated with three pulses of 30 s (30% power) and 2 min refractory period each on ice, to disrupt the chromatin. Later, samples were centrifuged at 14,000 rpm at 4◦C for 10 min and supernatants were labeled as nuclear insoluble fractions (Ni). Total extracts (T) were obtained by freezethawing trophozoites in the presence of 100 mM Tris-HCl pH 7.9, 100 mM p-hydroxymercuribenzoate (PHMB) (Sigma-Aldrich), CompleteTM protease inhibitor cocktail (Sigma-Aldrich), 10µM E64 (Sigma-Aldrich) and 100 mM PMSF (Sigma-Aldrich). Protein concentration was determined by Bradford method (Bradford, 1976) using the Bio-Rad protein assay or the DC Protein Assay (BIO-RAD).

# SDS-PAGE Gels and Western Blot Assays

Protein extracts obtained as previously indicated (T, Ns, Ni or C), were separated by sodium dodecyl sulfate polyacrylamide gel, electrophoresed by (SDS-PAGE). Briefly, 40 µg of each extract was loaded in a 12% SDS-PAGE and transferred to 0.45µm nitrocellulose membranes (BIO-RAD). For western blot (WB) assays, membranes were stained with Ponceau S Red staining solution (Sigma-Aldrich) and incubated 1 h with 5% non-fat milk. Then were incubated overnight (ON) with α-TRF-like I (1:2,500), α-TRF-like III (1:1,000), α-EhKMT4 (1:1,000) (Borbolla-Vázquez et al., 2015); α-H3K27me3 (1:1,000, Cell Signaling Technology), α-H4K20me3 (1:1,000, Abcam), α-Lamin B1 (1:1,000, Abcam), α-actin (1:1,000, Sigma-Aldrich) or α-Myc (1:1,000, Cell Signaling Technology) antibodies and then, incubated for 2 h with the corresponding horseradish peroxidase (HRP) labeled secondary antibodies (1:8,000, Santa Cruz Biotechnology). Protein bands were detected and visualized by chemiluminescence on X-ray films. All experiments were repeated at least three times.

# Subcellular Localization by Confocal Microscopy

Trophozoites cultured on cover slides were fixed with absolute ethanol for 20 min at −20◦C and washed with PBS pH 6.8. Then, coverslips were incubated with 50 mM NH4Cl for 30 min at 37◦C and blocked with 1% bovine serum albumin for 30 min. For co-localization experiments, α-EhTRF-like I, α-Lamin B1, α-EhTRF-like III and α-H4K20me3 antibodies were labeled with Alexa Flour 647, 555, 488 and 488, respectively, using Molecular Probes <sup>R</sup> Antibody Labeling kit (ThermoFisher Scientific) and following the manufacturer's instructions. In other experiments, samples were incubated ON at 4◦C with <sup>α</sup>-TRF-like III (1:200) and α-Myc (1:100, Cell Signaling) antibodies. Cells were washed and then, incubated with fluorescein labeled secondary antibody (1:200, Jackson Immuno Research). Slides were mounted with Vectashield containing DAPI (Vector Lab), visualized through a Nikon inverted microscope attached to a laser Confocal scanning system (Leica TCS SP2) and analyzed by Confocal Assistant software image.

# Construction of *pehtrf-like Iox*, *pehtrf-like IIIox*, and *pCold-ehtrf-like I* Plasmids

The ehtrf-like I (1,210 pb) and ehtrf-like III (1,383 bp) complete open reading frames (ORF) were cloned in the pKT-3M (Saito-Nakano et al., 2004) plasmid to express the EhTRF-like I and III proteins tagged with Myc sequence at their N-terminal. The ehtrf-like I gene was amplified by PCR using the following oligonucleotides: sense 5′ - TCCCCCCGGGATGAATAACCCTCAGTTGC-3′ and antisense 5′ - GCGCGCCTCGAGTTATTGAGAAAG ATCCAATTGTTTAAAT-3′ . The ehtrf-like III gene was amplified using the following oligonucleotides: sense 5′ - TCCCCCCCGGGATGGAGAAAAAACTAA-3′ and antisense 5 ′ -GGGGCCTCGAGTTAAAATTATCAGAATTA-3′ . Forward primers contained a SmaI site and the reverse primers contained a XhoI site (underlined). PCR was performed with 100 ng of E. histolytica genomic DNA, using the following conditions: 92◦C for 5 min, 92◦C for 1 min, 55◦C for 1 min, and 70◦C for 1 min (28 cycles). PCR products were digested with SmaI and XhoI and cloned into a previously digested pKT-3M vector. For the EhTRF-like I recombinant protein (rEhTRF-like I), the 1,210 pb entire ORF was amplified using the forward primer 5- CGCGGATCCTTATTGAGAAAGATCCAATTG-3 and reverse primer 5-GGAATTCCATATGAATAACCCTCAGTTTGC-3, which contained BamHI and NdeI restriction sites, respectively. The 1,210 bp amplicon was digested and cloned into a previously linearized pCold I vector (Takara). All plasmids obtained were confirmed by sequencing using the BigDye Terminator v3.1 Cycle Sequencing Kit and the ABI Prism <sup>R</sup> 310 Genetic Analyzer (Applied Biosystems).

# Transfection of Trophozoites

Trophozoites were transfected by lipofection as previously described (Olvera et al., 1997; Abhyankar et al., 2008). Briefly, 8 × 10<sup>5</sup> trophozoites were placed for adhesion in 6-well plates during 20 min, and then washed with M-199 medium (Invitrogen) supplemented with 5.7 mM cysteine, 1 mM ascorbic acid and 25 mM HEPES pH 6.8. Later, 20 µg of pKT-3M or pehtrf-like Iox or pehtrf-like IIIox plasmids were added to a tube containing 20 µl of Superfect (Qiagen), incubated at room temperature (RT) for 20 to 30 min, mixed with 0.8 ml of M199 supplemented medium plus 15% bovine serum and added to each well-plate. Trophozoites were incubated for 4 h at 37◦C and then harvested and added to a 125 mm cell culture tube, containing pre-warmed TYI-S-33 medium. Transfected trophozoites were grown for 48 h and selected initially with 3 µg ml−<sup>1</sup> G-418 (Thermofisher Scientific) and gradually increased to <sup>20</sup> <sup>µ</sup>g ml−<sup>1</sup> .

# Immunoelectron Microscopy

Immunoelectron microscopy was performed as described (Segovia-Gamboa et al., 2011). Briefly, transfected trophozoites were washed twice with PBS, pH 7.4 and fixed in 4% paraformaldehyde and 0.1% glutaraldehyde in serum-free medium for 1 h at RT. Samples were embedded in LR White resin and polymerized under UV at 4◦C ON. Sections were obtained and mounted on formvar-covered nickel grids. Then, they were incubated with α-EhTRF-like III (1:1,000) and α-Myc (1:1,000) antibodies, followed by incubation with 30 nm gold-conjugated particles (Ted Pella Inc.) secondary antibodies (1:60). Thin sections (60–90 nm) were observed in a transmission electron microscope (JEM 1011).

# Expression and Purification of rEhtRF-Like I

Escherichia coli Rosetta (DE3) competent cells (Novagen) were transformed with the pCold-ehtrf-like I construct. The expression of rEhTRF-like I was induced with 1 mM isopropyl-1-thio-β-D-galactopyranoside (IPTG) at 16◦C for 18 h and analyzed by SDS-PAGE and WB assays. The WB assays were performed by using an anti-histidine monoclonal antibody (α-His) (Santa Cruz Biotechnology, Inc) (1:1,000) as the primary antibody and a peroxidase-conjugated goat anti-mouse polyclonal secondary antibody (Invitrogen-Gibco) at a dilution of 1:3,000. The recombinant protein was obtained from the soluble fraction and purified using an affinity chromatography by Ni Sepharose High Performance agarose (GE Healthcare), following the manufacturer's instructions. The rEhTRF-like I was dialyzed with 25 mM Tris–HCl pH 7.5 and 300 mM NaCl. Protein was quantified and used as described below.

# Electrophoresis Mobility Shift Assays (EMSA)

DNA-binding activity was tested with four different DNA probes: HsTEL [TTAGGG, human telomeric sequence (Hwang et al., 2001)], EhTEL (the STR sequence TTAGTATT derived from the NK2 unit from the tRNA arrays of E. histolytica; Clark et al., 2006; Tawari et al., 2008), a mut-TEL (mutated telomeric sequence) and non-Rel (non-related sequence) probe (**Table S2**). Double-stranded oligonucleotides were end-labeled with 10 µCi [α-<sup>32</sup>P] dATP (Perkin Elmer) or 10 µCi [γ-<sup>32</sup>P] dATP, using Klenow fragment or a polynucleotide kinase (New England Biolabs) and purified with the QIAquick nucleotide removal kit (Qiagen). Nuclear soluble extracts (10 µg) from wild type or transfected trophozoites or 106.86µM rEhTRF-like I were incubated for 20 min at 4◦C with 1 ng of labeled oligonucleotides in binding buffer (12 mM HEPES pH 7.9, 60 mM KCl, 5 mM MgCl2, 1 mM EDTA, 4 mM Tris–HCl, 5 mM DTT, 10% glycerol, 4 mM Spermidine, 4 mM MgCl, 50 ng of poly dI·dC and protease inhibitors cocktail). Competitions were carried out in the presence of a 50, 100 and 200-fold molar excess of unlabeled specific HsTEL, EhTEL, mut-Tel and non-Rel probes. Supershifts were performed by pre-incubating nuclear extracts from transfected trophozoites with α-EhTRF-like III or PS (2–4 µg/reaction) for 30 min on ice, before adding the probe. DNA– protein complexes were resolved on a 6% non-denaturing PAGE gel by electrophoresis at 160 V for 2 h at RT in 0.5× TBE buffer (44.5 mM Tris-borate, 44.5 mM boric acid, 1 mM EDTA). Gels were vacuum-dried and radioactive complexes were detected in a Phosphor Imager apparatus (BIO-RAD) or revealed by autoradiography using Kodak X-omat film (Sigma) exposed at −70◦C.

# Ethic Statement

The Institutional Animal Care and Use Committee (Cinvestav IACUC/ethics committee) reviewed and approved our protocol for the animal care and use of rabbits employed to produce antibodies (Protocol Number 0313-06, CICUAL 001). All steps were taken to ameliorate the welfare and to avoid the animals suffering. Food and water were available ad libitum. Animals were monitored pre- and post-inoculation. All procedures were conducted by trained personnel under the supervision of veterinarians and, all invasive clinical procedures were performed while animals were anesthetized and when it was required, animals were humanely euthanized. The ethics committee verified that our Institute fulfills the NOM-062-ZOO-1999, regarding the Technical Specifications for Production, Care and Use of Laboratory Animals given by the General Direction of Animal Health of the Minister of Agriculture of Mexico (SAGARPA-Mexico). The technical specifications approved by SAGARPA-Mexico fulfill of the international regulations/guidelines for use and care of animals used in laboratory and were verified and approved by Cinvestav IACUC/ethics committee (Verification Approval Number: BOO.02.03.02.01.908).

# RESULTS

# *In silico* Characterization of *E. Histolytica* EhTRF-Like Proteins

In this work, we focus on the family II of the MYB DBDcontaining proteins of E. histolytica (Meneses et al., 2010). This family comprises genes encoding hypothetical proteins with AmoebaDB accession numbers (AE): EHI\_001090, EHI\_001110, EHI\_148140, EHI\_009820, and EHI\_074810, which were reported as single-repeat telomeric proteins related to human telomeric binding proteins (Meneses et al., 2010). On one hand, analysis of the amino acid sequences of the proteins coded by genes from locus EHI\_009820 and EHI\_074810 showed identities with the MYB DBD-containing proteins related to c-Myb, with narrowed identity to TRF-related proteins (**Table 1**). In contrast, the amino acid sequences of proteins coded by genes from locus EHI\_001090, EHI\_001110, and EHI\_148140 present about 25.8 to 29% identity with TRF1 from H. sapiens (**Table 1**). Moreover, these proteins share 29.6 to 30.3% identity with TRF proteins from other organisms, such as Pan troglodytes or Desmus rotundus (**Table 1**). Therefore, we considered these E. histolytica hypothetical proteins as TRF-like proteins and from now we named them as EhTRF-like I (EHI\_001090), II (EHI\_001110), and III (EHI\_148140). E. histolytica EhTRF-like proteins showed higher homology between each other than with human homologs. The sequence conservation between full length EhTRF-like proteins was 35.89 to 63.94% identity, in contrast to human TRF1 and TRF2 which had a 27.77% identity between them. Interestingly, the Myb DBD domain, which characterizes TRF proteins, showed 80.32 to 93.44% identity in EhTRF-like proteins, exhibiting a higher identity between EhTRF-like I and II (**Table S3**).

The alignment of the MYB DBD of the EhTRF-like proteins with TRF1 and TRF2 from H. sapiens showed that the sequences conserved the three α-helices characteristics of the MYB DBD, as well as the positions of the second and third tryptophan residues responsible of HTH conformation (**Figure 1A**; **Figure S1**). In addition, EhTRF-like I, II and III have a telebox motif (VxKDxxR) in the third α-helix, and a lysine and arginine


<sup>∧</sup>According to AmoebaDB data base. \*According to Uniprot. &Percentage of identity in global alignment. #Hypothetical protein.

binding domain; A, acidic N-terminus; B, basic N-terminus; MW, molecular weight; and pI, isoelectric point.

residues involved in the DNA telomeric recognition in human TRF1 and TRF2 (**Figure 1A**; **Figure S1**).

Even though there was scarce amino acid identity among the TRFH domains from EhTRF-like proteins and human TRFs, several hydrophobic residues required for the dimer interface, such as leucines, isoleucines, and phenylalanines were conserved (**Figure S1**). In addition, we deduced the secondary structure of the complete amino acid sequences for EhTRF-like proteins and human TRF1 and we aligned them to have a structural comparison (**Figure S2**). Based on their secondary structure, we proposed the presence of a TRFH domain located in the Nterminal of EhTRF-like proteins, because we found at the same position, the nine α-helices essential for domain dimerization in human TRF1 (**Figure 1B**; **Figure S2**).

According to the ORFs size (1,215, 1,326, and 1,383 pb for ehtrf-like I, II and III, respectively), the predicted molecular

XP\_001913661.1), and EHI\_074810 (E. histolytica, XP\_649576.1).

weight of EhTRF-like I, II and III is 46.8, 51. 5, and 53.1 kDa, respectively, showing similarity to their human counterparts (**Figure 1B**). TRF1 and TRF2 have isoelectric points (pI) of 5.99 and 9.22, respectively, which are related to their amino acids content in their N-terminal end. Theoretical pI of EhTRF-like I and II were predicted as 5.84 and 6.68, respectively, similar to that of TRF1, which is an acidic protein. On the contrary, EhTRF-like III presented a predicted pI of 8.39, alike to the basic protein TRF2. Consequently, the first 50 amino acids of EhTRFlike I and II were acidic residues, similar to those found at the same region of TRF1. On the other hand, the EhTRF-like III Nterminal was composed of basic residues as in TRF2 (**Figure 1B**). TRF2 also contains binding sites for the Shelterin proteins, Rap1 and TIN2 (de Lange, 2005). In concordance, in EhTRF-like III we identified the presence of two α-helices with conserved hydrophobic residues, corresponding to a Rap1-binding domain (RBM domain; **Figure 1B**). Likewise, all EhTRF-like proteins showed a nuclear localization signal (NLS) and susceptible sites of phosphorylation and SUMOylation, present in their human counterparts (**Figure 1B**). Altogether, we predict that EhTRF-like proteins from E. histolytica have a similar architecture to TRF1 and TRF2 from H. sapiens and are homologs to mammalian telomeric repeat-binding factors. In summary, EhTRF-like I and II share common properties with TRF1, while EhTRF-like III does with TRF2.

# Phylogenetic Analysis of TRF Proteins With MYB DBD or Telebox Domain

In order to shed light into the evolutionary relationships of EhTRF-like proteins, we aligned the MYB DBD amino acid sequence of TRFs from E. histolytica, H. sapiens, representative vertebrates, plants and deep branching protozoa, including other members of Entamoeba genus and Trypanosomatids (**Figure S3**). Proteins coded by genes from locus EHI\_009820 and EHI\_074810 from E. histolytica and TvTBP protein from Trichomonas vaginalis served as outgroup for tree reconstruction. Interestingly, the alignment with the MYB DBD showed that the first tryptophan residue in TRF-like proteins from Entamoeba genus was replaced with the aromatic amino acid phenylalanine, and unlike to TRF1 and TRF2 they present serine or cysteine residues in the third α-helix of their MYB DBD. In addition, all Entamoeba TRF-like proteins conserved the Telebox signature in a greater extent than other unicellular organisms like Trypanosomatids, where homologs to TRF have been already characterized (Li et al., 2005; da Silva et al., 2010). The phylogenetic inference showed that the TRF proteins are separated into two branches, one of them included the vertebrate proteins related to TRF1 and TRF2 and the TRF-like proteins from plants, and in the other branch contained TRFlike proteins from protozoan parasites including members from the Entamoeba genus (**Figure 2**). It is important to highlight that all TRF-like proteins from Entamoeba genus were clustered into a unique clade and subdivided in two groups that separated members of the EhTRF-like I and II (group A) from members of the EhTRF-like III (group B). The topology observed in the phylogenetic analysis could be the result of a gene duplication event. Interestingly, genes that encoded EhTRF-like I and II proteins are contiguous located in the same genomic location (DS571146). These results suggest that TRF proteins evolved from a common ancestor before vertebrate TRFs diverged and that in the case of E. histolytica, a gene duplication events could occur and therefore increased the TRF gene number.

# Shelterin Machinery Survey

To determine whether E. histolytica contains other components of Shelterin machinery besides TRFs, we searched for proteins homologous to Rap1, TIN2, TPP1 and POT1 in its genome. Through this analysis, we identified a hypothetical protein (AmoebaDB AN: EHI\_064550, 156 residues) with 26% identity to H. sapiens Rap1 (399 amino acids; **Table 2**). Sequence analysis of this protein revealed the presence of a TRF2-interacting telomeric protein/Rap1 C-terminal domain (e-value of 2e-19), suggesting that in E. histolytica this hypothetical protein could bind to EhTRF-like III. For these reasons, we decided to name EHI\_064550 as EhRap1-like protein. In H. sapiens, RAP1 binds to DNA through a MYB-type domain, which was absent in EhRap1-like protein, suggesting that its function is limited to TABLE 2 | Candidate proteins to be part of the Shelterin-like complex of E. histolytica.


<sup>∧</sup>According to Amoeba data base. #Putative uncharacterized protein.

telomeres protection. By using H. sapiens proteins as a bait, no other members of Shelterin machinery were identified in the E. histolytica genome. Nevertheless, the E. histolytica genome encodes a protein which is annotated in AmoebaDB database as the Replication Factor A1 (AN: EHI\_062980), therefore this protein was dubbed EhRpa1-like. EhRpa1-like conserves an oligonucleotide/oligosaccharide (OB) domain (e-value of 9e-50 according to Pfam), which could be used for single-stranded telomeric DNA recognition in the absence of POT1 (**Table 2**). These findings, support the idea that E. histolytica exhibits a rudimentary Shelterin-like complex conformed by EhTRF-like I, II, III, and EhRap1-like. In addition, EhRpa1-like could have a possible role in telomere maintenance.

# Expression of the EhTRF-Like Proteins in Trophozoites of *E. Histolytica*

To determine whether all three trf-like genes were constitutively expressed by E. histolytica, the mRNA expression patterns were analyzed using qRT-PCR assays derived from trophozoites grown in basal culture conditions. Results showed that three trf-like genes had differential expression levels, being ehtrf-like III more apparently expressed than the other two (**Figure 3A**).

In order to analyse their protein expression, we produced polyclonal antibodies only against EhTRF-like I and III proteins, using synthetic peptides (N-term-CTLPSVGNALIPPS and Nterm-CNKQKVQPQVSQPH, respectively) to immunize rabbits. In western blot analysis using trophozoites lysates, these antibodies revealed two bands of ∼55 and 65 kDa, respectively (**Figure 3B**). Those bands are of higher molecular weight than the expected for EhTRF-like I (46 kDa) and III (53 kDa), respectively. Differences between theoretical and experimental molecular weights could be explained by post-translational modifications, such as phosphorylation, ubiquitination and SUMOylation of EhTRF-like I and III (**Figure 1B**). To determine the subcellular localization of both proteins, we carried out a cellular fractionation according to (Schreiber et al., 1989), in which cytoplasmic and soluble nuclear fractions were isolated. To extract the insoluble nuclear proteins, RIPA buffer was added to the remaining nuclei pellet. Results evidenced that EhTRFlike I was present in cytoplasmic and nuclear fractions; however, in cytoplasmic and soluble nuclear fractions the antibody

FIGURE 3 | WB were also performed using α-H3K27me3 (as a control for Ns fraction), α-H4K20me3 (as a telomeric label in Ns fraction), α-Lamin B1 (as a control for Ni fraction), α-EhKMT4 (as control for Ns and C fractions), and α-actin (as loading control). (C,D) Localization of EhTRF-like proteins in E. histolytica. Trophozoites were processed for immunofluorescence and incubated with α-EhTRF-like I (Alexa-647, magenta), α-EhTRF-like III (Alexa-488, green), α-Lamin B1 (Alexa-555, red) and α-H4K20me3 (Alexa-488, green) antibodies. Nuclei were counterstained with DAPI and preparations were analyzed by confocal microscopy. The white arrowheads indicate colocalization in the nuclear periphery. ph c, phase contrast. \*p < 0.05, \*\*\*p < 0.0002, \*\*\*\*p < 0.0001.

recognized a 55 kDa band, while in insoluble nuclear fractions, it only detected a 36 kDa band (**Figure 3B**). Otherwise, EhTRFlike III protein was only observed in the soluble nuclear fraction, with the same 65 kDa molecular weight than in total lysates. In these assays, to probe the fractions purity, different cellular fractionation markers were included. As a control of the soluble nuclear fraction, we detected the K27me3 modification of H3 in the total and soluble nuclear fractions previously identified in E. histolytica as a repressive epigenetic mark by Foda and Singh (Foda and Singh, 2015; **Figure 3B**). We also detected the K20me3 modification in the H4 histone, which is specific for heterochromatin and related to telomeric regions (Blasco, 2007) and was previously identified in E. histolytica by Borbolla-Vázquez (Borbolla-Vázquez et al., 2015). H4K20me3 was also present in the total soluble nuclear fractions. Then, we included the detection of the enzyme methyltransferase EhKMT4 as a control of cytoplasmic and soluble nuclear fractions. This protein was detected in total, cytoplasmic and soluble nuclear extracts (**Figure 3B**). For the insoluble nuclear fraction, we used Lamin B1, which has been reported at the nuclear periphery in contact with the inner side of nuclear envelope in this parasite (Lozano-Amado et al., 2016). This protein was only detected in the total and insoluble nuclear fractions (**Figure 3B**), as expected. Additionally, we used actin as loading control for all cellular fractionations; however, a lesser amount of protein was detected in insoluble and soluble nuclear fractions than in cytoplasm, maybe due to different polymerization state of the actin within the nucleus. Despite, gel stained with Coomassie blue of the fractions obtained, demonstrated a similar protein amount in all samples (**Figure 3B**). Our results revealed that E. histolytica trophozoites differentially express ehtrf-like I, ehtrf-like II and ehtrf-like III transcripts, and the EhTRF-like I and III proteins are concentrated at the nucleus.

# EhTRF-Like I and III Are Nuclear Proteins That Co-localize With Lamin B1 and H4k20me3

In order to confirm the EhTRFs-like proteins localization, we performed immunofluorescence assays. Trophozoites cultured in basal conditions were processed for immunofluorescence using α-EhTRF-like I, α-EhTRF-like III, α-Lamin B1 and α-H4K20me3 antibodies coupled to Alexa-647,−488,−555 and−488, respectively. Confocal images evidenced EhTRF-like I and III mainly at trophozoite nuclei and EhTRF-like I was also localized at cytoplasm (**Figure 3C**). Staining of both proteins appeared at nuclear periphery, thus we employed Lamin B1 as a specific marker of this localization (Goldman et al., 2002; Lozano-Amado et al., 2016). Images revealed that EhTRF-like I and III co-localized with Lamin B1 at nuclear periphery, but EhTRF-like I presented a more diffused stain pattern inside nuclei. Therefore, we also investigated if EhTRF-like I was present in telomeric regions, employing H4K20me3 as a telomeric chromatin marker (Blasco, 2007). We found that EhTRF-like I colocalized with H4K20me3,showing sometimes diffused patterns or well-defined foci (**Figure 3D**). These results indicated that EhTRF-like I and III proteins are localized in specific nuclear regions as nuclear periphery or distributed in foci, co-localizing with Lamin B1 and H4K20me3. Considering the localization of EhTRF-like proteins they could be participating in the protection of the chromosome terminal ends of E. histolytica.

To validate the localization and to gain insight into the functional effect of EhTRF-like proteins, in E. histolytica trophozoites, we over-expressed the ehtrf-like I and III genes fused to the Myc tag and cloned in the pKT-3M vector to generate the pTRF-like Iox and pTRF-like IIIox overexpression vectors. Trophozoites were transfected with the pTRF-like Iox, pTRF-like IIIox or empty (pKT-3M) plasmid and stably selected in medium supplemented with 20 <sup>µ</sup>g ml−<sup>1</sup> G-418. According to semi-quantitative and quantitative RT-PCR results, the expression of ehtrf-like I and ehtrf-like III was 10.2 and 8 fold higher, respectively compared to the expression in pKT-3M transfected trophozoites (**Figures 4A,B**, **5A,B**). In agreement, WB experiments of lysates from pEhTRF-like Iox showed that EhTRF-like I was overexpressed (55 kDa band corresponding to EhTRF-like I), when they are compared with lysates derived from trophozoites transfected with the empty vector (**Figure 4B**). Similarly, WB experiments of lysates from pEhTRF-like IIIox transfected trophozoites and using α-Myc antibody, detected a 65 kDa band corresponding to EhTRF-like III (**Figure 5C**). As expected, in pKT-3M transfected trophozoites this antibody did not recognize any protein. EhTRF-like I and EhTRF-like III overexpression was confirmed in confocal images of pEhTRFlike Iox and pEhTRF-like IIIox transfected trophozoites and using the α-EhTRF-like I or α-EhTRF-like III antibody. In these parasites, proteins were more abundant at the nucleus, but they were also observed at cytoplasm (**Figures 4D**, **5D**,j–l), in comparison to the lesser staining of trophozoites transfected with empty vector (**Figures 4D**, **5D**,f–h). Similar nuclear and perinuclear staining were obtained using the α-Myc antibody, which detected only the heterologous proteins in trophozoites overexpressing EhTRF-like I and EhTRF-like III (**Figures 4D**, **5D**,r–t). Comparison with trophozoites transfected with the empty vector no staining was detected (**Figures 4D**, **5D**,n–p). No signal was obtained in trophozoites incubated with both preimmune serums (**Figures 4**, **5**,b–d). All of these data confirm the nuclear localization pattern of the EhTRF-like I and III proteins.

# EhTRF-Like III Localizes at Nuclear Heterochromatin Regions

Results of EhTRF-like III localization at nuclear foci (**Figures 3**–**5**), suggested this protein could act in specific and functional regions of the nuclei. Thus, we analyzed the

FIGURE 4 | using the 40s rsp2 gene as control. (C) Coomassie blue stained SDS-PAGE showing total extracts of transfected trophozoites (empty plasmid pKT-3M or pEhTRF-like Iox). A duplicate gel was transferred to nitrocellulose membrane and submitted to WB using α-EhTRF-like I and α-actin antibodies. Anti-Actin antibody was used as loading control. (D) Transfected trophozoites were processed for immunofluorescence and incubated with pre-immune serum (PS), α-EhTRF-like I or α-Myc antibodies, followed by the α-rabbit FITC-coupled secondary antibody. Nuclei were stained with DAPI and preparations were visualized by confocal microscopy. Arrowheads: EhTRF-like I location at nuclei. ph c, phase contrast.

localization of this protein by transmission electron microscopy (TEM) in pEhTRF-like IIIox transfected trophozoites. We found EhTRF-like III abundantly in the trophozoite nuclei, enriched in the heterochromatin or highly condensed chromatin regions, close to the nuclear periphery (**Figures 6C,D**,g,h), using the α-EhTRF-like III antibody. This location was corroborated using the α-Myc antibody (**Figure 6F**,i). No signal was obtained in trophozoites incubated with pre-immune serum (**Figures 6A,B**) or in pKT-3M transfected trophozoites stained with the anti-Myc antibody (**Figure 6E**). These data showed that EhTRFlike III is found in chromatin regions with high degree of compaction, which is suggestive of telomeric areas, where EhTRF-like proteins could protect the chromosomes terminal ends.

# EhTRF-Like III Forms DNA-Protein Complexes With Telomeres Related Sequences

TRF proteins bind as homodimers to double-stranded DNA telomeric sequences (Broccoli et al., 1997b; Blasco, 2007). Hence, we analyzed if in the nuclear extracts obtained from this parasite there were proteins that could recognize telomeric canonical sequences. We employed EMSA assays using nuclear extracts obtained from wild type trophozoites and the human telomeric sequence (HsTRF). The nuclear extracts from wild-type trophozoites formed three DNA-protein complexes (**Figure 7A**, lane 2). The presence of these three complexes is possibly due by the TRF homologs binding to the HsTEL DNA. Presumably, the formation of these complexes are due to the binding of the three EhTRF-like proteins with the HsTEL probe. In the absence of nuclear extract, no DNA-proteins complexes were formed (**Figure 7A**, lane 1). The HsTEL probe competed with the formation of the three complexes when it was added at increased concentrations (50, 100 and 200 molar). These results suggested that proteins forming these complexes are related to telomeric sequences (**Figure 7A**, lanes 3–5). Even more, when the competition was performed with an E. histolytica sequence, the EhTRF probe, it was more efficient (**Figure 7A**, lanes 6–8), showing a greater affinity for this sequence. Other mutated or non-related probes as mut-Tel or non-REL, did not competed with any of DNA-protein complexes (**Figure 7A**, lanes 9–14).

To investigate whether EhTRF-like proteins were able to bind to double-stranded telomeric DNA, we purified the EhTRFlike I recombinant protein (rEhTRF-like I) (**Figure S4**). As shown in **Figures 7B,C**, rEhTRF-like I bound specifically to the HsTEL and EhTEL probes. Competition assays showed that the formed complex by the recombinant protein was abolished in the presence of unlabeled HsTEL and EhTEL excess probes (**Figures 7B,C**, lanes 3–8). There was no competition for binding when the mut-TEL probe was used with the same molar excess (**Figures 7B,C**, lanes 9–11).

Next, we used the EhTEL probe and nuclear extracts derived from transfected trophozoites. In these EMSA assays, we also obtained three DNA-protein complexes, but the complex III was enriched when nuclear extracts from pEhTRF-like IIIox transfected trophozoites were used (**Figure 7D**, lane 3). To corroborate the identity of the proteins that formed the DNAprotein complex, a super-shift assay was performed using the oligonucleotide EhTEL, nuclear extracts from pEhTRF-like IIIox transfected trophozoites and the α-EhTRF-like III antibody. We observed a shifted complex in the presence of the α-EhTRFlike III antibody (**Figure 7E**, lane 3) and the pre-immune serum did not modify the formation of any DNA-protein complexes (**Figure 7E**, lane 4). We also included an anti-Myc antibody but in our conditions, we didn't obtain slower migration complexes, probably the Myc epitope become hidden upon binding to DNA (inducing a conformational change). This data indicated that pTRF-like III is able to form DNA-protein complexes with EhTEL sequences. Overall, these results indicated that in the nucleus of this parasite there are proteins interacting with telomeric sequences.

# DISCUSSION

All eukaryotes protect their chromosome ends through telomere binding proteins, which are well-conserved among these organisms (Linger and Price, 2009). These proteins conform a protein complex dubbed Shelterin and bind specifically to telomeric regions in mammalian organisms (de Lange, 2005). However, less complex machineries are present in fission yeast, protozoan ciliates, and plants (Watson and Riha, 2010). The stable interaction of Shelterin with telomeres depends on the association of two proteins, TRF1 and TRF2 to double-stranded telomeric repeats. The presence of TRF-like proteins in primitive unicellular eukaryotes is outstanding and might represent the ancestral scenario during evolution of telomeres and its protein counterparts.

In silico analysis showed that Entamoeba histolytica has three genes coding for TRF-like proteins. These proteins conserve the Telobox motif in their MYB DBD, which is highly conserved as in higher eukaryotes, and showed high identity (25 to 35%) with the amino acid sequences of the human Sheltering proteins TRF1 and TRF2. It is very interesting that this parasite has three genes that encode for TRF proteins, which could be the result of a gene duplication event. It has been proposed that the selective pressure through mechanisms of recombination were involved during TRF paralog formation. Stress responses is a selection pressure which generate elevated paralog formation

FIGURE 5 | qRT-PCR and the 40s rsp2 gene as control. (C) Coomassie-blue stained SDS-PAGE showing total extracts of transfected trophozoites (empty plasmid pKT-3M or pEhTRF-like IIIox). A duplicate gel was transferred to nitrocellulose membrane and submitted to WB using α-Myc antibody. (D) Transfected trophozoites were processed for immunofluorescence and incubated with pre-immune serum (PS), α-EhTRF-like III or α-Myc antibodies, followed by the α-rabbit TRITC-coupled secondary antibody. Nuclei were stained with DAPI and preparations were visualized by confocal microscopy. Arrowheads: EhTRF-like III location at nuclei. ph c, phase contrast.

FIGURE 6 | Detection by TEM of EhTRF-like III at E. histolytica trophozoites nuclei. TEM of transfected trophozoites (pKT-3M and pEhTRF-like IIIox) incubated with pre-immune serum (PS; A,B), α-EhTRF-like III (C,D), or α-Myc antibodies (E,F), followed by the incubation with α-rabbit gold-labeled secondary antibody. Panels g–i: magnifications of squares on (C,D,F), respectively. N, nuclei and C, cytoplasm. Arrows: gold particles.

and lead to an exceedingly high rate of Telomeric Binding Proteins evolution (Lustig, 2016). In the case of E. histolytica, the host's environment submits the parasite to a variety of stress conditions (oxidative stress derived from the immune response, tissue invasion, migration, or the simply need of persistence in the host). It has been proposed that gene duplication is the main process by which new genetic material is obtained by an organism. Our qRT-PCR results showed a differential expression pattern of Ehtrf-like genes. ehtrf-like III is more expressed in basal conditions while ehtrf-like II has the minor expression. These results suggest that the mechanisms that control gene expression could have changed depending on the environment conditions. This differential behavior of the trophozoite has been observed changing the growing conditions of the parasite (Weber et al., 2016). The presence of three ehtrflike genes may result in a differential functionality of telomeric binding proteins improving the organisms' responses to the environmental challenges and protecting their chromosome ends.

Few telomeric binding proteins have been identified in protozoan parasites. Our analyses suggested that E. histolytica has a simpler machinery to protect their telomeric DNA. This machinery Shelterin-like could be similar to other unicellular parasites, such as Trypanosoma and Leishmania, were orthologs of TRF-2, Rpa-1 (replication protein A subunit 1) and RAP1 have been identified and characterized, suggesting that the telomeric

FIGURE 7 | TRF-like III forms DNA-protein complexes with telomeric sequences. (A) EMSA was done using radiolabeled double-stranded human canonical telomeric DNA (HsTEL) as probe and nuclear extracts obtained from E. histolytica trophozoites. Competition assays were done in the presence of 50, 100, and 200-fold excess of non-labeled sequences: HsTEL specific competitor, E. histolytica sequence (EhTEL), mutated telomeric sequence (mut-TEL) or non-related sequence (non-Rel). (B) EMSA assay using rEHTRF-like I recombinant protein and HsTEL probe competition assays were done in the presence of 50, 100, and 200-fold excess of non-labeled sequences (HsTEL, EhTEL and mut-TEL). (C) EMSA assay using rEHTRF-like I and EhTEL competition assays were done in the presence of 50, 100, and 200-fold excess of non-labeled sequences (HsTEL, EhTEL and mut-TEL). (D) EMSA assay using EhTEL sequence and nuclear extracts obtained from trophozoites transfected with the pKT-3M or pEhTRF-like IIIox plasmids. (E) Super-shift assay was done using radiolabeled double-stranded EhTEL sequence as probe, nuclear extracts obtained from transfected trophozoites and α-EhTRF-like III antibody or pre-immune serum (PS). Protein-DNA complexes were separated in a 6% PAGE. Black arrowheads: specific protein-DNA complexes. Open arrowhead: super shifted protein-DNA complex.

machinery evolved early in eukaryotes (Lira et al., 2007; Yang et al., 2009).

EhTRF-like I and III proteins span from 404 to 460 amino acids with theoretical molecular weight (MW) of 46.8 to 53.1 kDa, respectively; However, in our experiments, EhTRF-like I was recognized at a 55 or a 36 kDa band and EhTRFlike III was observed in all fractions at 60 kDa band. The MW increase observed in both proteins could be related to posttranslational modifications (PTMs), such as phosphorylation, ubiquitinations and SUMOylations that change protein mobility (Audagnotto and Dal Peraro, 2017). This is relevant since the function of the TRF proteins, their ability to bind to the telomeric DNA, their dimerization and location as well as their degradation and interaction with other proteins (Walker and Zhu, 2012) is regulated through PTMs. Therefore, we performed an in silico analysis and found that EhTRF-like proteins can be modified by SUMOylations in different residues, some of them conserved with respect to TRF-1 and TRF-2. The K302 and K303 SUMOylation sites of EhTRF-like I were conserved with respect to TRF1 (K338 and K339), which explain the MW difference since it has been reported that SUMOylated proteins increase their MW from 8 to 17 kDa for each unit of the bound SUMO peptide (Hilgarth and Sarge, 2005). We propose similar scenario for EhTRF-like III. SUMOylation regulates proteins with nuclear function since it is related to nuclear transport, transcription, location in subnuclear compartments, chromatin organization, DNA damage repair of DNA and is linked to cell cycle regulation, growth and apoptosis (Flotho and Melchior, 2013). In TRF1,

SUMOylation is related to telomere maintenance through the ALT pathway (Alternative Lengthening of Telomeres), which occurs in the absence of telomerase. SUMOylated TRF-1 is recruited to the PML bodies where telomeric regions have been identified DNA (Yu et al., 2007; Royle et al., 2008). Therefore, the increase in the MW of the EhTRF-like I and III proteins could be explained by means of this PTM that suggest a similar role. Interestingly, all the enzymes involved in SUMOylation have been identified in E. histolytica (Bosch and Siderovski, 2013). This allows us to propose that EhTRF-like I could have similar PTMs as TRF-1 to modulate its activity and protect telomeric DNA. Detection of the EhTRF-like I protein in the insoluble nuclear fraction with a molecular weight of 36 kDa was also obtained. This molecular weight is lower than the predicted for EhTRF-like I. EhTRF-like I protein contains different residues susceptible to proteolytic cleavage, E352 was predicted as proteolytic cleavage site. If this site is functional, it would generate a peptide with a similar weight to that we found in the insoluble fraction. Interestingly, this residue is also conserved in H. sapiens TRF1. Finally, it has been reported that the presence of high content of acidic residues might affect the gel mobility shift of a protein and thereby explain the molecular weight variations found (Guan et al., 2015).

Human cells telomeres are tethered to the nuclear envelope during post-mitotic nuclear assembly. TRF proteins are associated with the nuclear membrane through Lamin B1. Binding of lamins to telomeres is partially mediated by TRF2, via its interaction with RAP1 which interacts with the nuclear envelope protein Sun1 (Hediger et al., 2002; Crabbe et al., 2012; Gonzalo and Eissenberg, 2016). In agreement, we found that EhTRF-like I and III colocalized with Lamin B1 at the nuclear periphery suggesting that they occupy a similar position. In accordance with this localization, in EhTRF-like III overexpressing trophozoites, this protein was localized also in the nuclear periphery proximal to the nuclear membrane in condensed heterochromatin regions. Likewise, telomeres have been found located in interphase nuclei close to the nuclear envelope; However, not in all organisms telomeres occupy this position, for example in plants like A. thaliana they have been reported close to the nucleolus (Schrumpfová et al., 2014). Therefore, the subnuclear location of telomeres is species-, cell type- and cell-phase dependent (Giraud-Panis et al., 2013). Telomeres carry features of repressive chromatin associated with constitutive heterochromatin. Different histone signatures have recently been identified associated with mammalian telomeres: trimethylation of H3K9 and H4K20 (Blasco, 2007). Here we selected H4K20me3 because it was previously identified and described in this parasite (Borbolla-Vázquez et al., 2015) and could suggest its participation in the organization and regulation of telomeric DNA in the E. histolytica nuclear periphery. Consistent with this, we observed that EhTRF-like I colocalized with the H4K20me3 mark suggesting that EhTRF-like proteins occupy regions of silenced compacted chromatin, in consistence with the telomere compacted structure (Giraud-Panis et al., 2013). Given the position of EhTRF-like I and III at the nuclear periphery and their colocalization with Lamin B and the trimethylated H4K20, we propose that these proteins could participate in the protection of chromosome ends.

Finally, we explored whether nuclear proteins from E. histolytica recognized the human telomeric sequence (HuTRF). Interestingly, we found three DNA-protein specific complexes from nuclear extracts of trophozoites that were competed with the human canonical telomeric sequence and with the probe derived from the STR repeat of the NK2 array that could be related to telomeric DNA (EhTEL). These DNA-protein complexes were not competed with mutated telomeric sequence (mut-TEL) or non-related (non- REL), showing specificity for telomeric sequences. Moreover, in an attempt to evaluate the telomeric DNA binding properties of EhTRF-like proteins, we observed that rEhTRF-like I recognized both HuTRF and EhTEL sequences. Since E. histolytica has three genes that code for three proteins with telomeric MYB DBD it would be interesting to determine if DNA-protein complexes correspond to these three proteins. Using nuclear extracts from trophozoites overexpressing EhTRF-like III we found that this protein formed DNA protein complexes with the STR sequence of E. histolytica. In our conditions, only complex III was enriched when extracts from TRF-like III overexpressing clones were used with the EhTEL probe indicating that the TRF-like III protein interacts with the STR repeat of the tRNA genes. This is the first report were a sequence derived from the tRNA arrays ([NK2]) is used to determine its recognition by nuclear proteins and rTRF-like I in EMSA assays. Previously, the YE array was used in fluorescence in situ hybridization (FISH) analysis, detecting six distinct signals at the parasite nucleus (Willhoeft and Tannich, 2000). However, as the ploidy of E. histolytica remains to be determined the interpretation of this evidence was difficult. It will be interest to determine the in vivo association of EhTRF-like proteins with sequences derived from the tRNA arrays.

In conclusion, the protection of the chromosome ends is critical to the survival of any cell as their disruption can induce genomic instability and consequently compromise the viability of the organism. Although no canonical Shelterin proteins have been identified in this protozoan, TRF-like proteins might accomplish this role for their ability to recognize and interact with telomeric DNA through their MYB-DBD. Therefore, this work shows the first evidence of telomeric proteins in E. histolytica able to interact with the proposed Entamoeba telomeric DNA sequences conforming a simple Shelterin complex as in other protozoans (**Figure 8**). This work raises several interesting questions and further investigation will contribute to a better understanding of the role of EhTRF-like proteins in the telomeric function in E. histolytica.

# REFERENCES


# AUTHOR CONTRIBUTIONS

All authors contributed equally to design and conception of this work. FR-G, VÁ-H, EC-O, and RC-G collected E. histolytica experimental data. HC-H performed the in-silico analysis; BC-M and AL-G performed MET analysis. AB contributes with confocal microscopy. FR-G, AB, JV, EO, LL-C, and EA-L contributed to experimental design, intellectual input, interpreting data and in writing the manuscript.

# FUNDING

EA-L thanks CONACYT for a research grant to study proteins that preserve genome integrity in Entamoeba histolytica (Ciencia Básica # 222956).

# ACKNOWLEDGMENTS

VÁ-H thanks fellowship 586852 from CONACYT. HC-H thanks fellowships BI 080530161142 and ICyTDF/179/2011. We thank Dr. Luis Brieba for reading the manuscript and providing constructive comments. The authors are grateful to Alfredo Barberi for graphical design and Brenda Herrera Villalobos for technical assistance.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00341/full#supplementary-material


amazonensis TRF (TTAGGG repeat-binding factor) homologue binds and co-localizes with telomeres. BMC Microbiol. 10:136. doi: 10.1186/1471- 2180-10-136


content information. PLoS Negl. Trop. Dis. 4:e716. doi: 10.1371/journal.pntd. 0000716


4169–4169. Available online at: http://www.bloodjournal.org/content/110/11/ 4169

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Rendón-Gandarilla, Álvarez-Hernández, Castañeda-Ortiz, Cárdenas-Hernández, Cárdenas-Guerra, Valdés, Betanzos, Chávez-Munguía, Lagunes-Guillen, Orozco, López-Canovas and Azuara-Liceaga. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Epithelial Cells Expressing EhADH, An *Entamoeba histolytica* Adhesin, Exhibit Increased Tight Junction Proteins

#### Abigail Betanzos 1,2 \*, Dxinegueela Zanatta<sup>2</sup> , Cecilia Bañuelos <sup>3</sup> , Elizabeth Hernández-Nava<sup>4</sup> , Patricia Cuellar <sup>5</sup> and Esther Orozco<sup>2</sup> \*

<sup>1</sup> Consejo Nacional de Ciencia y Tecnología, Mexico City, Mexico, <sup>2</sup> Departamento de Infectómica y Patogénesis Molecular, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico City, Mexico, <sup>3</sup> Coordinación General de Programas de Posgrado Multidisciplinarios, Programa de Doctorado Transdisciplinario en Desarrollo Científico y Tecnológico para la Sociedad, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico City, Mexico, <sup>4</sup> Programa de Cáncer de Ovario, Instituto Nacional de Cancerología, Mexico City, Mexico, <sup>5</sup> Centro Regional de Educación Superior, Universidad Autónoma de Guerrero, Chilpancingo, Mexico

#### *Edited by:*

Serge Ankri, Technion – Israel Institute of Technology, Israel

#### *Reviewed by:*

César López-Camarillo, Universidad Autónoma de la Ciudad de México, Mexico Elisabeth Labruyere, Institut Pasteur, France

> *\*Correspondence:* Abigail Betanzos

abetanzosfe@conacyt.mx Esther Orozco esther@cinvestav.mx

#### *Specialty section:*

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

*Received:* 12 April 2018 *Accepted:* 07 September 2018 *Published:* 28 September 2018

### *Citation:*

Betanzos A, Zanatta D, Bañuelos C, Hernández-Nava E, Cuellar P and Orozco E (2018) Epithelial Cells Expressing EhADH, An Entamoeba histolytica Adhesin, Exhibit Increased Tight Junction Proteins. Front. Cell. Infect. Microbiol. 8:340. doi: 10.3389/fcimb.2018.00340 In Entamoeba histolytica, the EhADH adhesin together with the EhCP112 cysteine protease, form a 124 kDa complex named EhCPADH. This complex participates in trophozoite adherence, phagocytosis and cytolysis of target cells. EhCPADH and EhCP112 are both involved on epithelium damage, by opening tight junctions (TJ) and reaching other intercellular junctions. EhADH is a scaffold protein belonging to the ALIX family that contains a Bro1 domain, expresses at plasma membrane, endosomes and cytoplasm of trophozoites, and is also secreted to the medium. Contribution of EhADH to TJ opening still remains unknown. In this paper, to elucidate the role of EhADH on epithelium injury, we followed two strategies: producing a recombinant protein (rEhADH) and transfecting the ehadh gene in MDCK cells. Results from the first strategy revealed that rEhADH reached the intercellular space of epithelial cells and colocalized with claudin-1 and occludin at TJ region; later, rEhADH was mainly internalized by clathrin-coated vesicles. In the second strategy, MDCK cells expressing EhADH (MDCK-EhADH) showed the adhesin at plasma membrane. In addition, MDCK-EHADH cells exhibited adhesive features, producing epithelial aggregation and adherence to erythrocytes, as described in trophozoites. Surprisingly, the adhesin expression produced an increase of claudin-1, occludin, ZO-1 and ZO-2 at TJ, and also the transepithelial electric resistance (TEER), which is a measure of TJ gate function. Moreover, MDCK-EhADH cells resulted more susceptible to trophozoites attack, as showed by TEER and cytopathic experiments. Overall, our results indicated that EhADH disturbed TJ from the extracellular space and also intracellularly, suggesting that EhADH affects by itself TJ proteins, and possibly synergizes the action of other parasite molecules during epithelial invasion.

Keywords: MDCK cells, EhCPADH, amoebiasis, trophozoites, epithelial barrier, transepithelial electrical resistance, intercellular junctions

Entamoeba histolytica is the protozoan responsible for human amoebiasis that infects 50 million people and kills between 30 and 100 thousand individuals around the world (Singh et al., 2016). Amoebiasis is characterized by acute diarrhea due to the substantial damage of the colonic epithelium produced by E. histolytica trophozoites (Cornick and Chadee, 2017). Trophozoites attach to and displace over the epithelium, contacting the epithelial cell surface. Then, they open the intercellular spaces by gradual separation of adjacent cells. Subsequently, epithelial cells are detached from the substrate and phagocytosed by the parasite (Martínez-Palomo et al., 1985). Several molecules are involved in this process, such as Gal/GalNAc lectin, amoebapores, cysteine and serine proteases, prostaglandin E2 (PGE2), the EhCPADH complex, among others (Chadee et al., 1987; Leippe, 1997; García-Rivera et al., 1999; Meléndez-López et al., 2007; Lejeune et al., 2011; Cornick et al., 2016).

Tight junctions (TJ) regulate ion and macromolecules flux across the epithelium, and also constitute the first barrier that pathogens face during host invasion. TJ are composed by integral proteins (e.g., claudins, occludin and junctional adhesion molecules) bound to the actin-cytoskeleton by cortical proteins, such as ZO-1,−2, and−3 (Capaldo et al., 2014).

The initial epithelial damage produced by E. histolytica is characterized by TJ opening, reflected as a dramatic drop of transepithelial electrical resistance (TEER) (Martínez-Palomo et al., 1985; Leroy et al., 2000; Betanzos et al., 2013), with the participation of PGE2 (Lejeune et al., 2011) and EhCPADH (Betanzos et al., 2013). PGE2 increases ion permeability by altering claudin-4 (Lejeune et al., 2011), while the EhCPADH complex affects claudin-1 and occludin (Betanzos et al., 2013). EhCPADH also damages adherens junctions (AJ) and desmosomes (DSM) (Hernández-Nava et al., 2017), structures that reinforce adhesion among epithelial cells, participate in cell polarity establishment and constitute centers of intracellular signaling (Capaldo et al., 2014).

The EhCPADH complex (Arroyo and Orozco, 1987), formed by an adhesin (EhADH) and a cysteine protease (EhCP112), participates in adhesion, cytolysis and phagocytosis of target cells (García-Rivera et al., 1999). EhCPADH, EhADH, and EhCP112 are secreted during trophozoite attack (Ocádiz et al., 2005; Bolaños et al., 2016). Moreover, an EhCP112 recombinant protein drops TEER of epithelial cells, and dislocates and degrades junctional molecules, including claudin-1, claudin-2, βcatenin, E-cadherin, desmoplakin-I/II and desmoglein-2 (Cuellar et al., 2017; Hernández-Nava et al., 2017).

EhADH contains a Bro1 domain (residues 9–349), characteristic of ALIX family members which are scaffold and multifunctional proteins (Odorizzi, 2006; Morita et al., 2007; Bissig and Gruenberg, 2014). Besides to its adhesive properties, EhADH is also an accessory protein of the endosomal sorting complex required for transport (ESCRT) machinery, whose components are pivotal players during phagocytosis in trophozoites (Avalos-Padilla et al., 2015, 2018). EhADH is localized at plasma membrane and endosomal compartments, and together with ESCRT members, contributes to multivesicular bodies formation (Bañuelos et al., 2012; Avalos-Padilla et al., 2015). Moreover, EhADH associates to cholesterol-trafficking proteins EhNPC1 and EhNPC2, suggesting an extra role in the uptake and transport of this essential lipid toward cellular membranes (Bolaños et al., 2016). Monoclonal antibodies (mAbAdh) against the C-terminal adherence domain (residues 480–600) of this protein (Montaño et al., 2017), inhibit trophozoite adhesion to and phagocytosis of erythrocytes, as well as destruction of MDCK cell monolayers (García-Rivera et al., 1999). However, the specific role of EhADH on epithelium damage has not been fully studied. What does the parasite protein do when it reaches the epithelium? Does it penetrate the target cell? If so, what does the adhesin carry out inside the cell? To approach these questions, we followed two different strategies: (i) we produced a recombinant protein (rEhADH) to scrutinize the effects of EhADH alone on epithelial cell monolayers, and (ii) we generated epithelial cells stably transfected with the ehadh gene (MDCK-EhADH) to evaluate EhADH effects within the cells.

Our findings showed rEhADH reaching the intercellular space of epithelial cells, co-localizing at TJ with claudin-1 and occludin. This protein was mainly internalized by clathrin-coated vesicles to MDCK cells. Meanwhile, MDCK cells expressing EhADH, exhibited epithelial aggregation and an increased adhesion to erythrocytes. Furthermore, EhADH mainly altered the amount of claudin-1 and occludin, reflected as an increase of TEER. Interestingly, we found that MDCK-EhADH cells resulted more susceptible to live trophozoites during TEER and cytopathic assays, than control cells. Thus, we suspect that EhADH, which preserved its adhesive properties within MDCK-EhADH cells, could be in some way preparing to epithelial cells for the trophozoites attack mediated by other virulence factors, whose identity and role on the concerted mechanism of epithelium damage should be further addressed.

# MATERIALS AND METHODS

# Antibodies

For EhADH immunodetection, we obtained rabbit polyclonal antibodies (α-EhADH) against a specific EhADH peptide (N-566 QCVINLLKEFDNTKNI 582-C) localized within the adherence domain. New Zealand male rabbits were immunized three times (each 2 weeks) with 300 <sup>µ</sup>g of this peptide diluted in TiterMax <sup>R</sup> Gold Adjuvant liquid (Sigma). Other primary antibodies used were: mouse against EhADH (mAbAdh) (Arroyo and Orozco, 1987), α-actin (kindly donated by Dr. José Manuel Hernández from Department of Cellular Biology, CINVESTAV, Mexico), αclaudin-1 (Invitrogen), α-occludin (Invitrogen) and α-caveolin-1 (Santa Cruz Biotechnology); rabbit α-ZO-2 (Invitrogen), α-ZO-1 (Invitrogen), α-occludin (Invitrogen), α-α/β tubulin (Cell Signaling), and α-PCNA (Azuara-Liceaga et al., 2018); and goat α-clathrin (Santa Cruz Biotechnology) and α-GAPDH (Santa Cruz Biotechnology). For some experiments, mouse IgM isotype control (Thermo Fisher) was used. Secondary antibodies included: α-rabbit, α-mouse and α-goat HRP-labeled IgG (1:10,000) (Life technologies); and α-rabbit, α-mouse and α-goat FITC-, TRITC- and Cy5-labeled IgM, and IgG (1:100) (Zymed) antibodies.

# Cell Cultures

Trophozoites of E. histolytica strain HM1:IMSS clone A (Orozco et al., 1983) were axenically cultured at 37◦C in TYI-S-33 medium and harvested during the logarithmic growth phase by chilling the culture tubes for 10 min in an ice-water bath. Then, trophozoites were collected by centrifugation at 360 × g for 5 min (Diamond et al., 1978).

Madin Darby canine kidney (MDCK) epithelial cells type I (Cereijido et al., 1978) and human colorectal adenocarcinoma (Caco-2) from the C2BBe1 lineage (Sambuy et al., 2005) were grown in DMEM medium (Gibco) supplemented with 100 IU/ml penicillin (in vitro), 100 mg/ml streptomycin (in vitro), 10% fetal bovine serum (Gibco), and 0.08 U/ml rapid-acting insulin (Eli Lilly), at 37◦C in a 95% air and 5% CO<sup>2</sup> atmosphere (Betanzos et al., 2013). Transfected MDCK cells were cultured in DMEM medium supplemented with 0.5 mg/ml G-418 (Gibco), a neomycin derivative.

# Construction of Plasmids and Production of rEhADH

A DNA fragment of 2061 bp encoding the full-length of E. histolytica ehadh gene, was PCR amplified from the pExEhNeo-Ehadh112 plasmid (Bañuelos et al., 2005) using the oligonucleotides described in **Table 1**. The ehadh gene was cloned into the pGEX6P1 and pcDNA3 plasmids (Invitrogen) between BamHI and Xho1, or KpnI and BamH1 digestion sites, respectively. Escherichia coli C43 (DE3) and -DH5α bacteria (Invitrogen) were transformed with pGEX6P1-ehadh and pcDNA3-ehadh, respectively. Plasmids were purified by an affinity column (Qiagen) and automatically sequenced to corroborate the ehadh gene sequence.

To produce an EhADH recombinant protein (rEhADH), E. coli C43 (DE3) bacteria were transformed with the pGEX6P1 ehadh plasmid. The recombinant protein was induced with 1 mM IPTG and purified as described (Bañuelos et al., 2012).

# Transfection Assays

Transfection of pcDNA3-ehadh or pcDNA3 plasmids into MDCK cells was performed with the Lipofectamine <sup>R</sup> 2,000 transfection reagent (Invitrogen), following the manufacturer's instructions. Positive clones were selected after 48 h transfection using 1 mg/ml of G-418 in the culture medium. After 3 weeks, the antibiotic was diminished to 0.5 mg/ml.

# RT-PCR Experiments

Total RNA from non-transfected and transfected MDCK cells was isolated by TRIzol <sup>R</sup> Reagent (Invitrogen). cDNAs were reverse-transcribed from 1 µg of total RNAs using M-MLV Reverse Transcriptase (Promega) and following the manufacturer's instructions. Then, PCR for ehadh, neo, and gapdh (as internal control) genes were performed with the oligonucleotides described in **Table 1**. PCR amplifications were done following standard procedures for cycling conditions that included an initial denaturing step at 94◦C for 1 min, followed by 30 cycles of 94◦C for 1 min, 50, 60, or 65◦C (according to respective Tm) for 1 min, and 72◦C for 1 min, with a final extension step at 72◦C for 7 min. Products were separated by electrophoresis in 1% agarose gels and revealed by ethidium bromide staining (Bolaños et al., 2016).

# Cellular Extracts

Bacteria were lysed with 2% sarcosyl and 0.5% Triton X-100 in PBS by sonication at 4◦C.

Trophozoites were washed twice with ice-cold PBS (140 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, pH 7.4) and lysed by freeze-thawing in the presence of 100 mM phydroxymercuribenzoate (PHMB) and 40µg/ml of E-64 (García-Rivera et al., 1999).

MDCK cells were collected with a rubber policeman, washed three times with ice-cold PBS and lysed for 30 min in RIPA buffer (40 mM Tris-HCl pH 7.6, 150 mM NaCl, 2 mM EDTA, 10% glycerol, 1% Triton X-100, 0.5% sodium deoxycholate, 0.2% SDS, 1 mM PMSF and the CompleteTM [Roche] protease inhibitor cocktail) under continuous and vigorous shaking. Extracts were


Restriction enzyme sites are underlined. F, forward; R, reverse.

sonicated three times for 30 s and centrifuged for 15 min at 25,000 × g to eliminate undissolved cellular debris (Betanzos et al., 2013).

# Western Blot Assays

Protein samples were separated by 6, 8, 10, or 15% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), subsequently transferred onto nitrocellulose membranes and incubated 1 h with 5% non-fat milk. Blotting were performed overnight (ON) with mouse mAbAdh (1:50), α-claudin-1 (1:1,000) or α-actin (1:1,500); rabbit α-EhADH (1:3,000), αoccludin (1:1,000), α-ZO-1 (1:500), α-ZO-2 (1:800) or α-tubulin (1:3,000); or goat α-GAPDH (1:10,000) antibodies. Then, we used HRP-conjugated secondary antibodies against mouse IgM and IgG, rabbit IgG or goat IgG, followed by a chemiluminiscence detection system (ECL-Plus kit; Amersham Pharmacia Biotech). Protein bands were visualized on a MicroChemi System (DNR Bio-Imaging), and densitometry analysis were performed using the ImageJ software.

# Interaction of Epithelial Cells With Recombinant Proteins

Confluent and sparse MDCK and Caco-2 cell monolayers were apically incubated with 10 µg/cm<sup>2</sup> of rEhADH or rEhPCNA (Cardona-Felix et al., 2011) for different times at 37◦C. After interaction, epithelial cells were washed five times with PBS (140 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, pH 6.8) to eliminate unbound molecules, and treated for immunofluorescence assays as described below.

In the case of rEhADH, it was Alexa Fluor 647-labeled using the Antibodies Labeling Kit (Molecular Probes), following the manufacturer's instructions. Briefly, 100 µg of rEhADH were equilibrated with 0.1 M sodium bicarbonate to allow the adequate succinimidyl ester reaction with primary amines of the protein, to form stable dye-protein conjugates (Cuellar et al., 2017).

# Immunofluorescence Assays

MDCK and Caco-2 cell monolayers grown on glass coverslips were fixed and permeabilized with absolute ethanol for 30 min at −20◦C. Cells were blocked for 30 min with 0.5% BSA and 0.05% saponin, and then incubated ON at 4◦C with mouse <sup>α</sup>-claudin-1 (1:25), mouse α-occludin (1:1,000), rabbit α-EhADH (1:100), rabbit α-ZO-1 (1:100) or rabbit α-ZO-2 (1:100) antibodies. For co-localization experiments, monolayers were incubated with mouse mAbAdh (1:10), rabbit α-EhADH (1:100) or rabbit α-PCNA (1:100), and mouse α-caveolin-1 (1:100), goat αclathrin (1:100), mouse α-claudin-1 (1:25), mouse α-occludin (1:1,000), rabbit α-ZO-1 (1:100) or rabbit α-ZO-2 (1:100) antibodies. After three-times washing with PBS, preparations were accordingly incubated for 1 h at room temperature (RT) with FITC-, TRITC- and Cy5-conjugated secondary antibodies. For cells in suspension, the procedure was performed in Eppendorf tubes and finally placed on coverslips. In some cases, nuclei were counterstained with 2.5µg/ml of 4′ ,6-diamidino-2-phenylindole (DAPI) (Zymed) during 5 min. Preparations were mounted with the antifade reagent Vectashield (Vector laboratories) and examined through a confocal microscope (Leica TCS\_SP5\_MO) in Z-stack optical sections of 0.5µm and xz- and zy-planes. In all cases, 10 fields per condition were analyzed and representative images were selected for each time.

For inhibition of clathrin-coated vesicles transport, MDCK cells were pre-incubated with 300 mM sucrose diluted in DMEM medium at 37◦C for 1 h or DMEM medium as a control (Mosso et al., 2008). Then, cells were processed for immunofluorescence as above.

# Aggregation and Adhesion Assays

Transfected MDCK cells were trypsinysed, washed twice in PBS, diluted in DMEM and suspended as hanging drops from the lid of a 24 well culture dish. Wells were filled with sterile water to prevent drops drying (Thoreson et al., 2000). Culture dishes were kept in a humid 5% CO<sup>2</sup> incubator at 37◦C and cell aggregation was determined 4 h after plating. Cells in each drop were passed 10 times through a standard 200 µl Gilson pipet tip and photographed through a Nikon E600 microscope, using a 20X phase contrast objective. The number of isolated cells and cells forming aggregates were counted in five random fields from three independent experiments.

Adhesion assays were lightly modified from previously described (Orozco et al., 1983). Briefly, MDCK cells in suspension were mixed with washed erythrocytes (1:100 ratio). Cell mixture was incubated for 0.5, 1 and 2 h at 37◦C, fixed with 2.5% glutaraldehyde for 30 min at 37◦C and washed three-times with PBS. Erythrocytes were counterstained with 4.5 mM diaminobenzidine for 30 min at 37◦C. Finally, adhered erythrocytes to each MDCK cell in ten random fields from three independent experiments were counted through a Nikon E600 microscope.

For inhibiting cell aggregation and erythrocytes adhesion, before experiments, 2 × 10<sup>5</sup> MDCK cells were incubated with <sup>10</sup> <sup>µ</sup>g mAbAdh antibody or IgM isotype for 30 min at 37◦C.

# Measurement of Transepithelial Electrical Resistance (TEER)

Transfected MDCK cells were seeded on Transwell filter supports (6.5 mm diameter and 0.4µm pore; Corning). Three days after plating, and after confirming through an inverted microscope that monolayers reached confluency, the TEER was measured using an EVOM epithelial voltmeter (World Precision Instruments) (Betanzos et al., 2013). TEER values were obtained by subtracting cell-free filter readings.

For some experiments, transfected MDCK cells were apically incubated with live trophozoites (10<sup>5</sup> /cm<sup>2</sup> ) and TEER was monitored during 90 min.

# Paracellular Flux Assays

TRITC-dextran (3 mg/ml) of 4 kDa (Sigma Aldrich) was added to the apical side of transfected epithelial cells in confluency, seeded in Transwell filters. After 90 min incubation at 37◦C with gentle shaking and darkness, samples from the basal chamber were collected and the diffused fluorescent tracer was measured in a fluorimeter (excitation λ = 547 nm; emission λ = 572 nm). Emission values were converted to TRITC-dextran Betanzos et al. EhADH Expression in MDCK Cells

concentration, using a standard curve (Cuellar et al., 2017). As a positive control, before tracer addition, cells were incubated for 30 min with 5 mM EDTA.

# Cytopathic Assays

Transfected MDCK cells in confluency, seeded in 24-well plates, were twice-washed with PBS to remove traces of serum and then incubated with live trophozoites (50, 100, and 250 × 10<sup>3</sup> ) suspended in TYI-S-33 medium without serum. Incubation was carried out for 2 h at 37◦C in a CO<sup>2</sup> containing incubator. The reaction was stopped by cooling cell culture plates in an icewater bath, to release adhered trophozoites. Epithelial cells were carefully washed with cold PBS and monolayer destruction was measured as described (Bracha and Mirelman, 1984).

# Statistical Analysis

All data shown in this work, were representative from three independent experiments performed at least by triplicate. Results were displayed as mean and standard error. For statistical analysis, we followed the two-tailed Student t-test and twoways ANOVA using the GraphPad Prism 5 software. Statistical significance was assumed when <sup>∗</sup>p < 0.05, ∗∗p < 0.01, or ∗∗∗p < 0.001.

# Ethics Statement

The Centre for Research and Advanced Studies (CINVESTAV) fulfilled the standard of the Mexican Official Norm (NOM-062-ZOO-1999) and Technical Specifications for the Care and Use of Laboratory Animals based on the Guide for the Care and Use of Laboratory Animals The Guide, 2011, NRC, USA with the Federal Register Number BOO.02.03.02.01.908. This is awarded by the National Health Service, Food Safety and Quality (SENASICA) belonging to the Animal Health Office of the Secretary of Agriculture, Livestock, Rural Development, Fisheries and Food (SAGARPA), an organization that verifies the state compliance of such Mexican Official Norm (NOM) in Mexico. The Institutional Animal Care and Use Committee (IACUC/ethics committee) of CINVESTAV, as the regulatory office for the approval of research protocols involving the use of laboratory animals and, in fulfillment of the NOM, has reviewed and approved all animal experiments (Protocol Number 0505-12, CICUAL 001).

# RESULTS

# Recombinant EhADH (rEhADH) Penetrates Epithelial Cells Through the Paracellular Route

The EhCPADH complex is a virulence factor involved in adhesion, phagocytosis and cytolysis of target cells by E. histolytica trophozoites. This protein, as well as EhCP112 and EhADH, are secreted during host invasion, and from the medium they reach the target cell surface (García-Rivera et al., 1999; Ocádiz et al., 2005). To investigate the role of EhADH on epithelium damage without the interference of other trophozoite proteins, we produced an EhADH recombinant protein (rEhADH). The ehadh gene was cloned into the pGEX6P1 plasmid, and pGEX6P1-ehadh transformed bacteria were induced for EhADH expression as a recombinant protein tagged to GST (**Figure 1A**). After purification by a glutathione-sepharose resin, the purity of rEhADH was visualized in silver-stained gels. A single 101 kDa band corresponding to the expected molecular weight for EhADH (75 kDa) (García-Rivera et al., 1999) plus GST (26 kDa) was evidenced (**Figure 1B**). The identity of the recombinant protein was probed by western blot assays using α-GST and mAbAdh antibodies. Both antibodies recognized the same 101 kDa band (**Figure 1B**); whereas in trophozoites lysates, the polyclonal α-EhADH antibody directed against a EhADH polypeptide (residues 566–582) located within the adherence domain, recognized the 75 kDa band (**Figure 1C**). Purified rEhADH coupled to the Alexa 647 fluorescent dye, was added to the apical side of confluent MDCK cells. Confocal microscopy images showed that immediately after rEhADH addition, the protein located at cellular borders, then, it appeared co-localizing with occludin and claudin-1 at the TJ region (**Figures 2**, **3**). xz-planes images revealed that rEhADH was posed firstly on the apical surface of cells. Next, it penetrated through the intercellular space to be later found inside the cells. During these experiments, monolayers seemed intact under phase contrast images, even at 60 min rEhADH incubation. However, confocal images revealed that after cells contact with rEhADH (1 and 5 min), occludin and claudin-1 were delocalized from the cellular borders and eventually, both proteins were internalized to the cytoplasm (15 min). Afterwards, at 30 and 60 min these proteins partially recovered their TJ localization. Another E. histolytica recombinant protein, rEhPCNA (Cardona-Felix et al., 2011), did not bind to epithelial cells and neither produced modifications on TJ proteins (**Figure 2B**), indicating that the rEhADH effects were specific. We also investigated the effect of rEhADH on sparse MDCK cultures, where TJ structure and functions are not yet well stablished. Results evidenced claudin-1 at cellular borders in a discontinuous pattern, and this protein was not diminished neither internalized during all rEhADH incubation times (**Figure S1**). Whilst, at the beginning of incubation, rEhADH was mainly localized at the borders of cellular groups in growth. Nevertheless, at 30 min incubation, rEhADH was found at plasma membrane and cytoplasm (**Figure S1**), in accordance to the effect of this protein on confluent cells at 15 min (**Figure 3**). To analyse the effect produced by rEhADH on the natural colonization site in the host, we assessed its localization on human intestinal Caco-2 cells. Confocal images showed similar results in Caco-2 (**Figure 4**) and MDCK cells, confirming that rEhADH contacts epithelial cells, then displaces along the intercellular space, penetrates the cells and delocalizes occludin and claudin-1.

# rEhADH Is Internalized by Clathrin-Coated Vesicles

Some pathogens and their secreted molecules enter the host cell by lipid microdomains known as caveolae and clathrin-coated vesicles (Rosenberger et al., 2000; Moreno-Ruiz et al., 2009; Machado et al., 2012). In fact, we have reported that EhCP112, the other part of the EhCPADH complex, is introduced to epithelial

zy-planes. Arrows: co-localization at cellular borders. Full arrowheads: co-localization at lateral membrane. Empty arrowheads: localization of EhADH (green) and occludin (red) at cytoplasm. Asterisks: EhPCNA localization. Bar = 20µm.

cells by these kind of vesicles (Hernández-Nava et al., 2017). To explore whether rEhADH was internalized into MDCK cells through any of these mechanisms, we used antibodies against human clathrin−1 and caveolin. MDCK cells were incubated with fluorescence-labeled rEhADH and followed by confocal microscopy. Images evidenced that rEhADH strongly localized at clathrin-coated vesicles, whereas its localization at caveolae was poor (**Figure 5A**). Next, to confirm rEhADH endocytosis mediated by clathrin-coated vesicles, we used sucrose to inhibit this type of transport (Mosso et al., 2008). MDCK cells treated for 1 h with sucrose diminished clathrin-coated vesicles (**Figure 5B**). These cells now incubated for 1 (**Figure S2**) and 5 min (**Figure 5B**) with rEhADH, poorly bound it and were not able to internalize this protein, while sucrose-untreated cells presented

rEhADH at cellular borders and inside the cells, co-localizing with clathrin (**Figure 5B**). In rEhADH-untreated MDCK cells, the sucrose treatment also affected the occludin pattern at TJ region, lightly reducing its localization at cellular borders. As mentioned above, in sucrose-untreated MDCK cells and incubated with rEhADH, occludin was delocalized (**Figure 2A**), however after 5 min rEhADH incubation, when clathrintransport is inhibited, occludin remained in its characteristic localization (**Figure 5B**). These findings robustly suggested that rEhADH is internalized to target cells by clathrin-coated vesicles.

# *E. histolytica* EhADH Is Expressed by *pcDNA3-ehadh* Transfected MDCK Cells

To study the effect produced by EhADH inside epithelial cells, we expressed this protein in MDCK cells that were transfected with the pcDNA3-ehadh plasmid, which contains a promoter for expressing exogenous genes in mammalian cells. Stably transfected cells were selected with G-418. To evidence the presence of the ehadh transcript, we performed RT-PCR assays. Results revealed that only MDCK cells transfected with pcDNA3 ehadh (MDCK-EhADH), expressed the ehadh transcript, in contrast to pcDNA3 transfected (MDCK-3) or non-transfected (MDCK) cells (**Figure 6A**). The stable gene transfection in these cells was also demonstrated by RNA expression of the neomycin resistance gene, which was also expressed by MDCK-3 cells. The expression of the EhADH protein was verified by western blot assays, using total extracts of MDCK-EhADH cells and a specific polyclonal antibody against the adhesin (α-EhADH). The α-EhADH antibody recognized a single 75 kDa band (**Figure 6B**), as reported (García-Rivera et al., 1999), and did not detect any protein in MDCK-3 and non-transfected cells (**Figure 6B**). In general, growth rates for all cell types were similar (**Figure 6C**). Thus, MDCK cells constitutively expressed E. histolytica EhADH and apparently, by itself the adhesin did not affect the viability and growth of epithelial cells.

# EhADH Is Localized at the MDCK Cellular Membrane

Protein function is closely related to its location in cellular structures. In trophozoites, EhADH is localized at plasma membrane, endosomes and cytoplasm (Avalos-Padilla et al., 2015; Montaño et al., 2017). The location of the EhADH protein in MDCK-EhADH cells was studied at different stages of the cell monolayer formation through immunofluorescence assays using the mAbAdh antibody. In MDCK-EhADH cells in suspension, EhADH mainly appeared at the plasma membrane (**Figure 6D**). In sparse cultures, when cell polarization begins, EhADH was predominantly concentrated at borders of the growing cellular groups, where the cell monolayer was extending (**Figure 6D**). This pattern correlated with that observed in sparse non-transfected cultures incubated with rEhADH (**Figure S1**). Interestingly, MDCK-EhADH confluent monolayers presented EhADH at cellular borders (**Figure 6D**), which is in concordance with the localization of its rat homolog (Alix), present at cellular borders of immortalized epithelial cells from rat choroid plexus

(Campos et al., 2016). EhADH was also abundantly detected around nuclei, probably in the endoplasmic reticulum, where adhesin is being synthesized (**Figure 6D**). As expected, EhADH was not present in MDCK-3 cells. These results indicated that as in trophozoites, EhADH localized at the plasma membrane and cytoplasm of transfected epithelial cells. Additionally, these experiments corroborated that EhADH stable expression did not modify the morphology of MDCK cells.

# EhADH Induces Aggregation of Target Cells

In E. histolytica trophozoites, EhADH is involved in target cell adherence (García-Rivera et al., 1999; Madriz et al., 2004; Martinez-Lopez et al., 2004). Therefore, we analyzed whether EhADH expressed by MDCK cells, evoked its adhesin function to target cells. For these experiments, transfected epithelial cells were trypsinysed and incubated in PBS at 37◦C for 4 h, then, we counted the number of clumps formed and the number of cells contained in each one. Optical microscopy images clearly evidenced that MDCK-EhADH cells formed more clumps with a higher number of cells than MDCK-3 cells that did not express the adhesin (**Figure 7A**). For quantification, clumps were grouped according to the number of contained cells (1–5, 6– 10, 11–20, 21–40, and more than 40 cells) and their frequencies were determined. Whilst MDCK-3 cells formed 82% of clumps containing 1–5 cells, only 43% of clumps were constituted by MDCK-EhADH cells (**Figure 7B**). In contrast, MDCK-EhADH and MDCK-3 cells formed 30% and 6% of clumps containing more than 40 cells, respectively. To confirm the adhesive role of EhADH in MDCK cells, before aggregation assays, transfected cells were pre-incubated with the mAbAdh antibody to inhibit clumps formation. Results showed that the antibody treatment, reduced the frequency of clumps with a higher number of cells in MDCK-EhADH, but not in MDCK-3 cells (**Figure 7B**). In contrast, an IgM isotype used as control, did not modify clumps frequency. These results suggested that MDCK-EhADH cells displayed stronger associations among them than MDCK-3 cells.

In trophozoites, EhADH also participates in erythrocytes adhesion (Avalos-Padilla et al., 2015; Bolaños et al., 2016). Thus, we studied whether MDCK-EhADH cells were able to bind erythrocytes. Transfected MDCK cells were incubated for different times with red blood cells and adhered erythrocytes were counted. Surprisingly, both types of transfected MDCK cells were able to attach erythrocytes (**Figure 7C**). However, after 1 and 2 h incubation, MDCK-EhADH cells adhered twice more erythrocytes than MDCK-3 cells (**Figure 7D**). To inhibit erythrocytes adhesion, again, before assays, transfected cells were pre-incubated with the mAbAdh antibody, resulting in a decreased number of red blood cells adhered to MDCK-EhADH, but not to MDCK-3 cells (**Figure 7D**). The IgM isotype presented a similar effect than in control transfected MDCK cells. These experiments showed that

EhADH preserved its adhesive properties within MDCK-EhADH cells.

# MDCK-EhADH Cells Present an Increased TEER

Previously, we reported that EhCPADH and EhCP112 affect the gate function of TJ in epithelial cells (Cuellar et al., 2017; Hernández-Nava et al., 2017). Therefore, we wondered if EhADH expression on MDCK cells may alter this function, characterized by the regulation of ion and macromolecules flux. To evaluate this, we used transfected epithelial cells cultured in Transwell filters. MDCK-EHADH cells exhibited approximately twice higher TEER values compared to MDCK-3 cells (**Figure 8A**). To measure the macromolecules flux, the non-ionic dextran marker coupled to the TRITC-fluorescent dye was added to the apical side of epithelial monolayers and then, we quantified the tracer diffusion across the paracellular pathway from the apical to the basolateral side. MDCK-EhADH and MDCK-3 cells presented a low dextran permeability, similar to nontransfected MDCK cells (**Figure 8B**). As a positive control, cells were incubated with EDTA that disassembles epithelial junctions (Deli, 2009). MDCK cells treated with EDTA allowed the free passage of dextran between intercellular spaces. Our data showed that EhADH enhanced the electrical tightness of MDCK monolayers, but it had not impact on macromolecules permeability.

# EhADH Co-localizes and Increases the Amount of TJ Proteins in MDCK-EhADH Cells

The above results evidenced that MDCK-EhADH monolayers have a higher TEER than MDCK-3 cells. TJ functions depend on the localization and amount of proteins in the intercellular space. Thus, to analyse the effect of EhADH expression in the localization of junctional proteins, we performed immunofluorescence assays using α-claudin-1, α-occludin, α-ZO-1, α- ZO-2, and α-EhADH antibodies. In contrast to the results obtained using the rEhADH added to the apical side of cell monolayers (**Figures 2**, **3**), confocal images showed that EhADH expression in MDCK cells did not affect the localization of proteins at TJ region, as it was clearly evidenced at xzplanes (**Figure 9A**). However, MDCK-EhADH cells presented an increase of the fluorescence corresponding to claudin-1, occludin, ZO-1 and ZO-2 (**Figure 9A**). As expected, all TJ proteins studied here co-localized with EhADH at the TJ region (**Figure 9B**). These results suggested that the stable presence of EhADH in transfected cells provoked an increase of claudin-1, occludin, ZO-1, and ZO-2 at TJ.

To further confirm the increased amount of TJ proteins in transfected cells, we analyzed their expression by western blot assays using the corresponding antibodies. Quantification of western blot results revealed that the amount of claudin-1 and occludin increased around 2.5 times in MDCK-EhADH cells, in comparison to MDCK-3 cells (**Figure 10**). Meanwhile, ZO-1

lysates were used as positive control. Numbers at right: expected molecular weight of proteins. (C) Growth curves of MDCK cells. After 48 h of plasmid transfection into cells, growth medium was supplemented with 1 mg/ml G-418, whereas non-transfected cells were grown in normal medium. The number of cells were counted in a Neubauer haemocytometer. Each point represents the mean and standard error of three independent experiments. (D) Suspension, sparse and confluent transfected MDCK cells were fixed, permeabilized and processed for immunofluorescence assays using the mAbAdh antibody. Bottom panel: MDCK-EhADH cells treated only with the secondary antibody (FITC). All images represent xy-optical sections. Empty arrowheads: plasma membrane. Full arrowhead: perinuclear localization. phc: phase contrast. Bar = 20µm.

and ZO-2 showed a slight increase in MDCK-EhADH cells without significant differences when compared to MDCK-3 cells. Actin was used as loading control and all densitometric analysis were normalized regarding this protein. These results correlated with the immunofluorescence findings, confirming that EhADH expression provoked a significant increase of claudin-1 and occludin, which are mainly responsible for the TJ gate function.

# EhADH Expression in MDCK Cells Facilitates Epithelial Destruction by Trophozoites

One of the first signs of damage to MDCK cell monolayers during E. histolytica invasion is TEER dropping. After this, cells are detached from the substrate and destroyed by the parasite (Martínez-Palomo et al., 1985). To evaluate the effect of E. histolytica trophozoites over transfected MDCK cells, epithelial cells were grown in Transwell filters until confluence, and live trophozoites were added in the apical side of monolayers, then, TEER was monitored during 90 min. To compare the TEER behavior of transfected cells, we normalized each TEER value regarding the one registered at 0 min. Results showed a stable TEER performance by MDCK-EhADH and MDCK-3 cells non-incubated with parasites (**Figure 11A**). Meanwhile, a TEER drop was evident when transfected cells were incubated with trophozoites. However, TEER values were lower in MDCK-EhADH than in MDCK-3 cells, from 5 to 30 min incubation. After this time, the effect of trophozoites on TEER seemed indistinct between MDCK-3 and MDCK-EhADH cells, reaching a stable behavior from 45 to 90 min.

To visualize and determine epithelial destruction by live trophozoites on transfected MDCK cells, confluent monolayers were apically incubated with different number of parasites. Live trophozoites destroyed 33, 65, and 100% (using 50, 100, and 250 × 10<sup>3</sup> amoebas, respectively) more MDCK-EhADH than MDCK-3 cells monolayers (**Figure 11B**). Representatives images of methylene blue-stained monolayers evidenced the damage produced by parasites. Surprisingly, trophozoites displayed more

photographed. Representative images from 2 h incubation are shown. Bar = 10µm. (D) Adhered erythrocytes to each MDCK cell were counted. Values represent the mean and standard error from three independent experiments by triplicate. Differences found between groups were statistically significant (\*\*\*p < 0.001), according to

damage on MDCK cells expressing EhADH than MDCK-3 cells, suggesting that this protein facilitated the epithelial injury.

In summary, all results of this work suggested that EhADH altered TJ of the host epithelium, reaching the paracellular space, being internalized mainly by clathrin-coated vesicles, delocalizing and increasing TJ proteins, for eventually making epithelial cells more susceptible to other trophozoite effector proteins. Nevertheless, this hypothesis should be further addressed to get in depth on EhADH action mechanisms in this event.

# DISCUSSION

two-ways ANOVA test.

E. histolytica trophozoites use several molecules to invade the host epithelium. Among them, our group has been investigating the participation of the EhCPADH complex in the breaking of intercellular junctions. We have found that EhCPADH and also EhCP112 reach the paracellular pathway by opening intercellular junctions and eventually, both proteins penetrate epithelial cells. The damage begins by affecting proteins and functions of TJ, and then, it continues degrading AJ and DSM proteins (Betanzos et al., 2013; Cuellar et al., 2017; Hernández-Nava et al., 2017). Nevertheless, the EhADH contribution to epithelium impairment during trophozoite invasion, had not been yet elucidated. In this work, by using a recombinant protein we revealed that EhADH gets first into the intercellular space, and next, it enters into the epithelial cells mainly by clathrin-coated vesicles. To analyse the effect of EhADH inside host cells, we generated MDCK cells stably expressing this adhesin. The EhADH expression produced more adhesiveness to target cells and an enhanced TEER due to a higher amount of claudin-1 and occludin. Interestingly, the EhADH expression in MDCK cells facilitated the epithelial damage produced by

\*\*\*p < 0.001, according to two-tailed Student t-test.

live trophozoites, suggesting a putative role for EhADH in synergizing the effect of other E. histolytica molecules during host invasion.

EhCPADH and its components are secreted to the medium (García-Rivera et al., 1999; Ocádiz et al., 2005; Bolaños et al., 2016), and together with other parasite molecules (Lejeune et al., 2011; Cornick et al., 2016) constitute an efficient mechanism to reach and effectively damage the host epithelium. We previously demonstrated that the EhCPADH complex drops TEER and interacts with claudin-1 and occludin, to eventually degrade them along with ZO-1 and ZO-2 (Betanzos et al., 2013). Besides, we evaluated the participation of EhCP112, the proteolytic part of this complex, by producing a recombinant enzyme (rEhCP112) which reaches the apical side of epithelium and then invade it through intercellular junctions (Cuellar et al., 2017; Hernández-Nava et al., 2017). Here, we investigated if the other part of the EhCPADH complex, the EhADH adhesin, also extended into the paracellular pathway, by using a recombinant protein (rEhADH, **Figure 1**) and characterizing its effect on epithelial cells. When rEhADH was added to the apical side of MDCK cells, it was at cellular borders co-localizing with occludin and claudin-1 at TJ, and eventually penetrated the cell (**Figures 2**, **3**). These findings were consistent to that observed in Caco-2 cells (**Figure 4**), which have been used in other works for resembling the E. histolytica natural colonization site in host (Li et al., 1994; Ralston et al., 2014). Here, we preferred the MDCK cells model because this cell line is a friendly and easy system to study molecules and functions of TJ, and also to stably and efficiently transfect heterologous genes (Cereijido et al., 1978; Furuse et al., 2001; Dukes et al., 2011). In addition, they have been widely used as a model to characterize trophozoites damage and TEER dropping during epithelial invasion (Martínez-Palomo et al., 1985; Dolabella et al., 2012; Betanzos et al., 2013; Hernández-Nava et al., 2017).

To investigate the EhADH internalization mechanism, we followed this adhesin together with clathrin and caveolin-1, resulting in a higher co-localization with clathrin than with caveolin-1 (**Figure 5A**). The rEhADH endocytosis mediated by clathrin-coated vesicles was confirmed by using sucrose as an inhibitor (Mosso et al., 2008), which diminished these vesicles and rEhADH internalization, thus preventing occludin delocalization from the membrane (**Figure 5B**). The endocytosis mechanism mediated by clathrin-coated vesicles has been observed for other pathogens adhesins, such as AfaD of diarrhea-associated E. coli strains or HadA of Haemophilus influenzae (Jouve et al., 1997; Serruto et al., 2009). Moreover, similarly to EhADH, EhCP112 is transported by clathrin-coated vesicles, although this enzyme is also internalized by caveolae (Hernández-Nava et al., 2017).

In order to analyse the role of EhADH inside epithelial cells and also to eliminate the noise coming from other amoebic proteins that produce host cell damage, we performed the heterologous expression of this adhesin by transfecting MDCK cells with the pcDNA3-ehadh construct. MDCK-EhADH cells efficiently expressed the ehadh transcript and the corresponding protein with the expected 75 kDa molecular weight, maintaining a similar growing rate than control cells (**Figures 6A–C**). As well as in trophozoites (García-Rivera et al., 1999; Avalos-Padilla et al., 2015), EhADH was localized in cytoplasmic vesicles and at the plasma membrane of epithelial cells (**Figure 6D**).

EhADH possesses adhesive properties, facilitating trophozoites adhesion to target cells (García-Rivera et al., 1999; Bañuelos et al., 2012). Here, we demonstrated that MDCK-EhADH cells aggregated more among them, and adhered more erythrocytes than control cells (**Figure 7**). The observed adhesive features were specific, since the mAbAdh antibody directed against EhADH inhibited them (**Figure 7**). Our findings suggested that MDCK-EhADH cells acquired the adhesive

properties displayed by this trophozoite protein. A similar effect was also reported for the H. influenzae HadA adhesin during its heterologous expression in a non-invasive E. coli strain, which gained properties for aggregation and adherence to human epithelial cells and to extracellular matrix proteins (collagens I and III, fibronectin and laminin) (Serruto et al., 2009). On the other hand, it was not surprising that MDCK cells adhered erythrocytes, since epithelial cells are similar to erythroid cells and both contain similar membrane-cytoskeletal components (Bennett and Lorenzo, 2013). In addition, both cell types contain ankyrin and fodrin, involved in lateral membrane association upon cell-cell contact induction (Glenney and Glenney, 1983; Bennett and Lorenzo, 2013).

and xz-optical planes observed by confocal microscopy. Bar = 20µm.

The increased adhesiveness among MDCK-EhADH cells and the presence of EhADH at cellular borders, affected the TJ gate function evidenced as an augment of TEER (**Figure 8**). In agreement, macromolecules flux by the paracellular pathway remained low. Other amoeba as Acanthamoeba also produced a high TEER in MDCK cells, probably because claudin-4 was increased and re-targeted to TJ, while claudin-2 diminished and was removed from the cellular borders (Flores-Maldonado et al., 2017). TEER is a measurement of ions flux through the paracellular pathway, and this flux is mainly regulated by claudins in different fashions; for example, claudin-4 promotes a barrier function by forming an ion channel, whereas claudin-2, constitutively expressed in the most leaky-epithelia, makes paracellular cation and water channels (Barmeyer et al., 2017). Therefore, we evaluated the localization and expression of TJ integral (claudin-1 and occludin) and scaffold (ZO-1 and ZO-2) proteins. According to TEER results, by immunofluorescence and western blot assays in MDCK-EhADH cells we observed an increase of the analyszd proteins at cellular borders, co-localizing with EhADH at the TJ region (**Figure 9**), although the augment was only significant for claudin-1 and occludin (**Figure 10**), molecules responsible for the gate function (Liang and Weber, 2014). Thereby, the delicate nature of the molecular composition of junctions and their balance, could be influencing the

same membranes were stripped and blotted with the α-actin antibody (representative images). (B) The specific amount of TJ proteins was obtained as the ratio of TJ proteins between actin densitometric evaluations from (A). Data represent the mean values and standard error from three independent experiments. \*p < 0.05, according to two-tailed Student t-test.

FIGURE 11 | Effect of trophozoites on MDCK cells expressing EhADH. (A) Transfected MDCK cells were seeded in Transwell filters and when they reached confluence, trophozoites (10<sup>5</sup> /cm<sup>2</sup> ) were apically added. TEER was monitored for 2 h and data were normalized according to the initial value given by each Transwell. Means and standard errors represent each time point from three independent assays performed by triplicate. \*\*p < 0.01 and \*\*\*p < 0.001, according to two-tailed Student t-test. (B) Transfected confluent MDCK cells were incubated with 0, 50, 100, and 250 × 10<sup>3</sup> live trophozoites. Monolayers cell destruction was determined after 2 h incubation by methylene blue staining. Then, dye was eluted and quantified by spectrophotometry. Data represent the mean and standard error of three independent experiments by triplicate statistically analyzed by two-tailed Student t-test. \*p < 0.05, \*\*p < 0.01. Representative images of monolayers destruction are showed below the graph.

epithelium susceptibility for alterations due to pathogens and their molecules.

The effect produced by EhADH inside epithelial cells could be explained by its scaffold properties, characteristic of ALIX family protein members (Bissig and Gruenberg, 2014). TJ peripheral scaffold proteins link transmembrane and functional barrier proteins to the actomyosin-ring (Liang and Weber, 2014). In fact, in immortalized rat choroid plexus cells, Alix protein guarantees the proper assembly and actomyosin-ring positioning at the TJ region by interacting with F-actin, Par-3 and ZO-1, thus contributing to the maintenance of epithelial cell polarity and barrier function (Campos et al., 2016). Otherwise, from previous experiments, we demonstrated that an EhADH recombinant protein including the adherence epitope, interacts with a 97 kDa membrane protein of MDCK cells (Martinez-Lopez et al., 2004). We hypothesized this protein could correspond to a claudin-1 tetramer, since claudins are capable to form oligomers, which are responsible for junction selectivity (Coyne et al., 2003). Furthermore, EhCPADH binds to claudin-1, occludin, ZO-1 and ZO-2 in MDCK cells (Betanzos et al., 2013). Thus, our findings in this work suggested that EhADH could be acting as a peripheral scaffolding protein to reinforce the attachment of claudin-1 and occludin to the TJ region, probably tightening cell-cell contacts. This assumption can be reinforced by the fact that other amoeba proteins also act as intercellular junction molecules. Dictyostelium discoideum amoeba contains an adhesion molecule named Aardvark, which is similar to the AJ scaffold protein β-catenin, suggesting the presence of a rudimentary cell-cell adhesion during the formation of the fruiting body that develops in colonies deprived of nutrients (Grimson et al., 2000). Even more, E. histolytica expresses an occludin-like protein that can alter the colonic epithelial barrier (Goplen et al., 2013).

Interestingly, MDCK-EhADH cells incubated with live trophozoites were more susceptible to parasites damage, according to TEER and cytopathic experiments (**Figure 11**). Although, these cells developed higher TEER values than control cells, trophozoites dropped TEER in MDCK-EhADH cells more than in MDCK-3 cells (from 5 to 30 min amoeba incubation); in accordance, parasites caused more damage on monolayers expressing EhADH. In trophozoites, it has been reported that EhADH binds to diverse molecules such as EhCP112, EhVps32, EhNPC1, EhNPC2, LBPA and cholesterol (Avalos-Padilla et al., 2015; Bolaños et al., 2016; Castellanos-Castro et al., 2016; Cuellar et al., 2017), thus participating in distinct virulence events. Therefore, in MDCK-EhADH cells, we hypothesized that this adhesin could be modulating or associating to these molecules or others, to facilitate their participation on epithelial damage. Identity of these putative molecules and related mechanisms should be further investigated.

# REFERENCES


Summarizing our findings from previous reports and those obtained in this paper, we demonstrated that EhCPADH and its components by separate, could reach the intercellular space and affect the epithelial barrier function. In the case of EhADH, when it is secreted by trophozoites to the medium, extends toward the paracellular pathway and later, after its internalization, modulates the expression and localization of claudin-1 and occludin. Once inside, EhADH could prepare host cells for the action of other virulence factors, making the epithelium more susceptible to the trophozoite attack.

# AUTHOR CONTRIBUTIONS

AB and EO: Designed, performed and analyzed experiments, and wrote the manuscript. DZ: Carried out immunofluorescence assays. EH-N: Produced the recombinant protein. PC: Performed RT-PCR experiments. CB: Contributed to experiments analysis, discussion and manuscript writing.

# FUNDING

This work was financed by the Mexican National Council for Science and Technology (CONACYT, 284477 to AB and 220049 to EO).

# ACKNOWLEDGMENTS

We are grateful to Dr. Guillermina García-Rivera, Dr. Miriam Huerta-Pérez and Ricardo Ceja for their assistance with E. histolytica assays and Alejandrina Reyes for her technical support.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00340/full#supplementary-material

Figure S1 | Co-localization of rEhADH with claudin-1 on sparse MDCK cells. rEhADH coupled to Alexa 647 (green) was apically added to sparse MDCK cells and incubated for different times, then, cells were fixed and processed for immunofluorescence assays using the α-claudin-1 antibody (red). Nuclei were counterstained with DAPI (blue) and preparations were analyzed through a confocal microscope at xy- and zy-planes. Arrows: co-localization at cellular borders. Full arrowheads: co-localization at lateral membrane. Empty arrowheads: localization of EhADH (green) and claudin-1 (red) at cytoplasm. Bar = 20µm.

Figure S2 | rEhADH localization in MDCK cells where clathrin-transport was inhibited. MDCK cells were incubated for 1 h with DMEM medium (–) or 300 mM sucrose (+) at 37◦C and then rEhADH coupled to Alexa 647 (green) was apically added for 1 min. Preparations were analyzed through a confocal microscope at xy- and xz-planes. Arrows: rEhADH localization. Bar = 20µm.

vacuole-associated protein involved in pinocytosis and phagocytosis of Entamoeaba histolytica. PLoS Pathog. 11:e1005079. doi: 10.1371/journal.ppat. 1005079

Avalos-Padilla, Y., Knorr, R. L., Javier-Reyna, R., García-Rivera, G., Lipowsky, R., Dimova, R., et al. (2018). The Conserved ESCRT-III machinery participates in the phagocytosis of Entamoeba histolytica. Front. Cell. Infect. Microbiol. 8:53. doi: 10.3389/fcimb.2018.00053


degrades claudin-1 and claudin-2 at tight junctions of the intestinal epithelium. Front. Cell. Infect. Microbiol. 7:372. doi: 10.3389/fcimb.2017.00372


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Betanzos, Zanatta, Bañuelos, Hernández-Nava, Cuellar and Orozco. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# *Entamoeba histolytica* Cyclooxygenase-Like Protein Regulates Cysteine Protease Expression and Virulence

Preeti Shahi, France Moreau and Kris Chadee\*

Department of Microbiology, Immunology and Infectious Diseases, Snyder Institute for Chronic Diseases, University of Calgary, Calgary, AB, Canada

The intestinal protozoan parasite Entamoeba histolytica (Eh) causes amebiasis associated with severe diarrhea and/or liver abscess. Eh pathogenesis is multifactorial requiring both parasite virulent molecules and host-induced innate immune responses. Eh-induced host pro-inflammatory responses plays a critical role in disease pathogenesis by causing damage to tissues allowing parasites access to systemic sites. Eh cyclooxygenase (EhCox) derived prostaglandin E<sup>2</sup> stimulates the chemokine IL-8 from mucosal epithelial cells that recruits neutrophils to the site of infection to exacerbate disease. At present, it is not known how EhCox is regulated or whether it affects the expression of other proteins in Eh. In this study, we found that gene silencing of EhCox (EhCoxgs) markedly increased endogenous cysteine protease (CP) protein expression and virulence without altering CP gene transcripts. Live virulent Eh pretreated with arachidonic acid substrate to enhance PGE<sup>2</sup> production or aspirin to inhibit EhCox enzyme activity or addition of exogenous PGE<sup>2</sup> to Eh had no effect on EhCP activity. Increased CP enzyme activity in EhCoxgs was stable and significantly enhanced erythrophagocytosis, cytopathic effects on colonic epithelial cells and elicited pro-inflammatory cytokines in mice colonic loops. Acute infection with EhCoxgs in colonic loops increased inflammation associated with high levels of myeloperoxidase activity. This study has identified EhCox protein as one of the important endogenous regulators of cysteine protease activity. Alterations of CP activity in response to Cox gene silencing may be a negative feedback mechanism in Eh to limit proteolytic activity during colonization that can inadvertently trigger inflammation in the gut.

Keywords: *Entamoeba histolytica*, parasites, cox like protein, cysteine protease, actinin like protein, virulence, pro-inflammatory cytokines

# INTRODUCTION

Entamoeba histolytica (Eh) is an invasive extracellular protozoan parasite responsible for amebic colitis and liver abscess (World Health Organization, 1998). It is one of the major cause of severe diarrhea in areas of poor sanitation and nutrition particularly in tropical and developing countries. While the majority of Eh infection remains asymptomatic, about 10% of infection converts to an invasive phenotype where Eh invades the mucosal epithelium resulting in 100,000 death/year (World Health Organization, 1997; Stanley, 2003).

#### *Edited by:*

Mario Alberto Rodriguez, Centro de Investigación y de Estudios Avanzados (CINVESTAV), Mexico

#### *Reviewed by:*

Carlos Rosales, National Autonomous University of Mexico, Mexico Salil Kumar Ghosh, United States Food and Drug Administration, United States

> *\*Correspondence:* Kris Chadee kchadee@ucalgary.ca

#### *Specialty section:*

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

*Received:* 31 July 2018 *Accepted:* 17 December 2018 *Published:* 09 January 2019

#### *Citation:*

Shahi P, Moreau F and Chadee K (2019) Entamoeba histolytica Cyclooxygenase-Like Protein Regulates Cysteine Protease Expression and Virulence. Front. Cell. Infect. Microbiol. 8:447. doi: 10.3389/fcimb.2018.00447

The host innate immune status and parasite virulence factors play major roles in disease pathogenesis (Faust and Guillen, 2012; Verkerke et al., 2012; Marie and Petri, 2014; Nozaki and Bhattacharya, 2015). The host immune response during Eh invasion of the colonic mucosa is characterized by increased levels of pro-inflammatory cytokines that recruits inflammatory cells including macrophages and neutrophils (Seydel et al., 1997; Mortimer and Chadee, 2010; Nakada-Tsukui and Nozaki, 2016) to the site of infection. The major Eh virulent factors identified to date are the galactose/N-acetyl-D-galactosamine (Gal/GalNAc) lectin (Gal-lectin), cysteine proteinases, amoebapore, and prostaglandin E<sup>2</sup> (Moonah et al., 2013; Marie and Petri, 2014).

Prostaglandins are lipid mediators synthesize from arachidonic acid through cyclooxygenase and are associated with various diarrheal disease including bacterial and inflammatory bowel diseases (Ahrenstedt et al., 1994; Alcantara et al., 2001; Resta-Lenert and Barrett, 2002). We have shown that Eh synthesizes PGE<sup>2</sup> through a cyclooxygenase like enzyme as confirmed by gas chromatography/mass spectrometry analysis (Belley and Chadee, 2000). Surprisingly, EhPGE<sup>2</sup> was not immunosuppressive to reduce host defenses but rather, PGE<sup>2</sup> bound EP4 receptors on colonic epithelial cells and stimulated the potent neutrophil chemoattractant, IL-8 mRNA expression and protein production. In addition, EhPGE<sup>2</sup> also altered tight junction proteins and increased ion permeability that led to diarrhea in intestinal amebiasis (Lejeune et al., 2011). Even though Eh produces high levels of PGE<sup>2</sup> in the presence of arachidonic acid the parasite can also stimulate host cells such as macrophages and colonic epithelial cell to produce PGE<sup>2</sup> as part of the pro-inflammatory response elicited by the parasite (Stenson et al., 2001; Sanchez-Ramirez et al., 2004). EhCox encodes a functional cyclooxygenase enzyme as evidenced by Cox enzymatic assays. An unusual aspect is that EhCox is primitive and has little homology to other Cox proteins from mammalian cells and some eukaryotes. At present, the biological function of EhCox other than being an enzyme that catalyzes the production of PGE<sup>2</sup> in Eh is not known.

In this study, we made the seminal observation that silencing EhCox enhanced cysteine protease protein expression and enzyme activity independent of EhCox-induced PGE<sup>2</sup> production. These results suggest that increased cysteine protease activity in EhCoxgs is linked to increase parasite-induced inflammation and pathogenicity. These findings increase our understanding on the molecular basis of pathogenicity in Eh and how dissimilar enzymes can regulate their activity in the parasite.

# MATERIALS AND METHODS

# Reagent

E64, leupeptin, aprotinin, and Nonidet P-40 detergent were obtained from Sigma-Aldrich. Z-VVR-AMC substrate was purchased from Enzo Life Sciences. The Z-Arg-Arg-pNA.2 HCl substrate was purchased from Bachem. Mouse monoclonal antiactin clone C4 antibody was purchased from MP Biomedical, LLC. Antibodies to EhCP4 and the CP inhibitors WRR483 and WRR605 were a kind gift from Dr. Sharon Reed, University of California, San Diego. EhCP5 and EhCox1 like antibodies were generated in rabbits using recombinant proteins expressed in E. coli (Belley and Chadee, 2000). Ubiquitin antibody (P4D1) was from Cell Signaling Technology and cycloheximide from Sigma-Aldrich.

# Cultivation and Harvesting of *E. histolytica*

G3 Eh were grown axenically in TYI-S-33 medium at 37◦C. After 72 h, logarithmic-growth-phase Eh cultures were harvested by chilling on ice for 9 min, pelleted at 200 g, and washed two times with PBS. For the detection of proteins and enzymatic activity, Eh lysate was prepared by using three cycles of freeze-thawlysis and proteins quantified by the bicinchoninic acid protein assay, using bovine serum albumin as protein standard (#23225, Thermo Scientific). Eh secretory protein (SP) were prepared as described previously (Lidell et al., 2006). Briefly, secreted components were collected from mid-log phase Eh incubated in Hanks' balanced salt solution (Invitrogen) for 2 h at 37◦C at a final concentration of 2 × 10<sup>7</sup> Eh per ml. Following incubation, Eh was removed by centrifugation at 10,000 × g for 10 min. Secretory proteins were quantified by the bicinchoninic acid protein assay. To quantify the growth of control and EhCoxgs, 2.5 × 10<sup>5</sup> log phase Eh were inoculated in 14 ml TYI-S-33 medium and the number of parasites counted every 24 h using a hemocytometer.

# Cloning of the *Eh*Cox-Like Gene

EhCox-like gene (Acc No. AF208390) 500 bp long 5′ end of protein-coding region was amplified by PCR from cDNA using specific primer containing stu1 and sac1 restriction sites (**Table 1**). The PCR product was sub cloned using the pGEM-T Easy vector system (Promega) and then digested with the restriction enzymes, stu1 and sac1. The digested DNA insert was cloned into StuI- and SacI-digested pSAP2-gunma (kind gift from Dr. Tomoyoshi Nozaki, Tokyo, Japan). This construct was verified by sequencing. G3 Eh were harvested during midlog growth and transfected with a silencing plasmid (pSAP2 gunma-cox) using the Lipofectamine (Life technologies) and OPTIMEMI medium (Life Technologies) supplemented with 5 mg/ml L-cysteine and 1 mg/ml ascorbic acid (transfection medium) and pH 6.8 as previously described (Fisher et al., 2006) Transfected Eh were selected with G418 (Sigma) over a 3-week period, starting with 6µg/ml and ending with 24µg/ml. Silencing was assessed in these strains using quantitative reverse transcription-PCR (qRT-PCR) and western blot analysis. Once gene silencing was confirmed, the G418 selection was removed, and then silencing was again confirmed and quantified.

# Preparation of *E. histolytica* Nuclear Proteins (*Eh*NP)

Eh were washed twice in PBS and suspended in lysis buffer [100 mM Tris (pH 7.4), 1µg/ml E-64 (Sigma), 2µg/ml leupeptin, 7.4µg/ml aprotinin and 0.5% Nonidet P-40 detergent] on ice for 15 min. Nuclei were pelleted by centrifugation at 2,000 × g for 15 min, washed with lysis buffer, and suspended in 0.1 M sodium phosphate buffer (pH 7.0). The protein concentration was determined by the bicinchoninic acid method using bovine serum albumin as protein standard (Thermo Scientific).

#### TABLE 1 | Primers used in this study.


# Quantitative Real-Time (qRT) PCR-Based Analysis of Gene Expression

Total RNA was extracted from logarithmic-growth-phase Eh using a Trizol reagent method (Invitrogen; Life Technologies, Burlington, ON) and the yield and purity determined by the ratio of absorbance at 260/280 nm (NanoDrop, Thermo Scientific). DNase I-treated total RNA was used in the RT reaction using qScript cDNA Synthesis kit and PerfeCTa SYBR Green Supermix (Quantabio). Real-time qPCR was performed using a Rotor Gene 3000 real-time PCR system (Corbett Research). The PCR reaction mix (20 µl) comprised 1x SYBR Green, 25 ng cDNA and 1µM of primers. A complete list of the primer sequences and conditions used are listed in **Table 1**. Results were analyzed using the 2- 11CT methods and expressed as fold changes.

# Sample Preparation for Proteomics

Eh lysates were prepared as described before. After proteins precipitation with TCA, the supernatant was discarded and proteins were air dried and suspended in 200 µL of urea/HCl. The sample was then treated with 10 mM of dithiothreitol and incubated at 37◦C for 30 min and transferred into Microcon YM-30 (Millipore) and centrifuge at 14,000 × g at 20◦C for 15 min. The eluates were discarded, 200 µL of UA was pipetted into the filtration unit, and the units were centrifuged again. Then 100 µL of 0.05 M iodoacetamide in Urea/Tris was added to the filters, and samples were incubated in darkness for 20 min. Filters were washed twice with 100 µL of Urea/Tris and, with 100 µL of 0.05 M ammonium bicarbonate at 14,000 × g for 15 min. Proteins were digested in 40 µL of trypsin in 0.05 M ammonium bicarbonate at 37◦C overnight. The released peptides were collected by centrifugation at 14,000 × g for 10 min followed by two washes with 0.05 M, 40 µL ammonium bicarbonate and with 40 µL of 0.05 M NaCl. After isolation of the peptides, samples were transferred into new Eppendorf tube.

# LC-MS/MS Analysis

Total protein and peptides content were analyzed on an Orbitrap Fusion Lumos Tribrid mass spectrometer (Thermo Scientific) operated with Xcalibur (version 4.0.21.10) and coupled to a Thermo Scientific Easy-nLC (nanoflow Liquid Chromatography) 1,200 system. Isolated trypsin treated peptides (2 µL) were loaded onto an Easy Spray Column (ES803) at a maximum of 700 bars (2µm particle column). Further, peptides were eluted using a 120 min gradient from 5 to 40% (5 to 28% in 105 min followed by an increase to 40% B in 15 min) of 0.1% formic acid in 80% LC-MS grade acetonitrile at a flow rate of 0.3 µL/min and separated on a C18 analytical column (ES803). Then, peptides were electrosprayed using 2.0 kV voltages into the ion transfer tube (300◦C) of the Orbitrap Lumos operating in positive mode. A full MS scan was performed by Orbitrap at a resolution of 12,0000 FWHM to detect the precursor ion having a m/z between 375 and 1575 and a +2 to +7 charge. The Orbitrap AGC (Auto Gain control) and the maximum injection time were set at 4e5

and 50 ms, respectively. The Orbitrap was working with a 3 s cycle time for precursor selection and most intense precursor ions presenting a peptidic isotopic profile, having an intensity threshold of at least 5,000 were isolated using the quadrupole and fragmented with HCD (30% collision energy) in the ion routing multipole. The fragment ions (MS<sup>2</sup> ) were analyzed in the ion trap at a rapid scan rate. The AGC and the maximum injection time were set at 1e4 and 35 ms, respectively, for the ion trap. Dynamic exclusion was enabled for 45 s to avoid of the acquisition of same precursor ion having a similar m/z (plus or minus 10 ppm).

# Database Search

With the help Raw Converter (v1.1.0.18; The Scripps Research Institute), raw data files (<sup>∗</sup> .raw) were converted into Mascot Generic Format (MGF). Monoisotopic precursors having a charge state of +2 to +7 were selected for conversion. The mgf file was used to search a database by using Mascot algorithm (Matrix Sciences; version 2.4). Search parameters for MS data included trypsin as enzyme, a maximum number of missed cleavage of 1, a peptide charge equal to 2 or higher, cysteine carbamidomethylation as fixed modification, methionine oxidation as variable modification and a mass error tolerance of 10 ppm. A mass error tolerance of 0.6 Da was selected for the fragment ions. Only peptides identified with a score having a confidence higher than 95% were kept for further analysis. The Mascot dat files were imported into Scaffold (v4.3.4, Proteome Software Inc.) for comparison of different samples based on their mass spectral counting.

# Western Blots

Proteins (30 µg) were loaded on 12% SDS-PAGE gel followed by transfer onto polyvinylidene fluoride (PVDF) membrane and blocking in 5% skim milk. Blots were then probed with 1:5,000 EhCox antibody or 1:1,000 EhCP5 antibody/1:500 EhCP4 antibody/1:1,000 Actin antibody for 16 h at 4◦C. After incubation with one of the previously described primary antibodies, the blots were incubated with appropriate HRP-conjugated secondary antibody for 1 h at RT, and then developed using ChemiLucent ECL detection (EMD Millipore).

# *Eh*Cox Activity Assay

EhCOX activity assay was performed on nuclear fractions of Eh (EhNP). EhNP (100 µg) was incubated for 30 min with 1 mM aspirin (ASA) followed by 100µM arachidonic acid (AA) or vehicle for 1 h at 37◦C in sodium phosphate buffer containing 200µM tryptophan and 2µM hematin (Sigma) in 500 µl volume. After centrifugation, PGE<sup>2</sup> in the supernatant was extracted with Amprep C2 ethyl columns (Amersham Biosciences) following manufacturer's protocol. PGE<sup>2</sup> was quantified by using an enzyme immunoassay kit (Cayman Chemical, Ann Arbor, MI).

# *Eh* Cysteine Proteinase Activity

EhCP5 and EhCP4 activity in lysate and secretory protein was determined using known chromophoric substrate benzyloxycarbonyl-L-arginyl-L-arginine p-nitroanilide (Z-Arg-Arg-pNA) and fluorogenic substrate benzyloxycarbonyl-L-val-L-val- 7-amino-4-methylcoumarin (Z-VVR-AMC), respectively, as previously described (Leippe et al., 1995; He et al., 2010). Briefly, substrate was incubated for 0–20 min at 37◦C with either Eh lysate/secretory proteins (50 µg) alone or pretreated with CP inhibitor-E64 [L-trans-epoxysuccinyl-leucyl-amido- (4-guanidino) butane]/CP-1(WRR483) and CP-4 (WRR605) inhibitors. Cleavage of the chromophoric (Z-Arg-Arg-pNA) and fluorogenic (Z-VVR-AMC) substrate were detected at the 405 and 460 nm wavelength, respectively.

# Gelatinase Gel Substrate Gel Electrophoresis

For analysis of protease activity by gelatinase substrate gel electrophoresis, 12% SDS polyacrylamide was copolymerized with 0.1% gelatin as described previously (Hellberg et al., 2000). Briefly, Eh lysate were prepared by three freeze-thaw cycles in HBSS and centrifuged for 10 min at 10,000 × g. Supernatant was separated on 0.1% gelatin copolymerized 12% SDS PAGE. After separation of proteins, SDS was removed by two washings in 2.5% Triton X-100 for 1 h at room temperature. Gel was then incubated in developing buffer (20 mM DTT, 100 mM sodium acetate, pH 4.2, and 1% Triton X-100) at 37◦C for 3 h. The gel was stained with Coomassie blue. Clear band represent cysteine protease activity.

# *Eh* Erythrophagocytosis Assay

Fresh human erythrocytes were obtained in DPBS and stained with Phicoerythrin by using PKH26 Red Fluorescent cell linker kit (PKH26, Sigma-Aldrich). The erythrocytes were counted and used in 1:100 (Eh: erythrocyte) ratio. Eh and erythrocytes were incubated for 20 min at 37◦C in DPBS. After interaction, cells were washed twice with DPBS and centrifuged at 3,000 × g for 5 min at 4◦C. Lysis buffer (RBC lysis buffer, Sigma-Aldrich) was added for 1 min at RT followed by 0.5 ml addition of FBS and washed again with DPBS. The cells were fixed with 4% p-formaldehyde for 20 min at RT and washed with DPBS. To each sample a drop of fluoresave reagent was used for the slide preparation. The slides were observed under confocal microscope and fluorescent intensity was quantified by selecting region of interest (ROI) containing phagocytose erythrocyte and measuring the fluorescent intensity by using Image J software. The data was plotted as mean fluorescent intensity.

# *Eh* Cytopathic Assay

Caco-2 human colonic adenocarcinoma cells (ATCC, Manassas, VA) were grown to obtain confluent monolayers in DMEM medium (Invitrogen-Gibco) supplemented with 5% fetal bovine serum and 5 mg/ml penicillin-streptomycin under 5% CO<sup>2</sup> at 37◦C (Sigma-Aldrich). Eh disruption of a Caco-2 cell monolayer was determined using a previously described protocol (Belley et al., 1996). Briefly, Eh (10<sup>5</sup> /well) were incubated with Caco-2 cell monolayers in 24-wells tissue culture plates at 37◦C. The incubation was stopped by placing the plates on ice and Eh were removed by washing with cold PBS. The Caco-2 cells that remained attached to the plates were stained with methylene blue (0.1% in 0.1 M borate buffer, pH 8.7). The dye was extracted from stained cells with 0.1 M HCl and color intensity of the extracted dye was measured spectrophotometrically at OD 660.

# Stability of *Eh*CP-A4 and *Eh*CP-A5 Protein

Eh and EhCoxgs were inoculated at 2.5 × 10<sup>5</sup> and grown in TYI-S-33 media for 48 h before treatment with the protein synthesis inhibitor, cycloheximide at 100µg/ml for 6, 12, and 24 h. After treatment Eh were harvested, wash in PBS and cell lysates were prepared in RIPA buffer 100 mM Tris-HCl, 150 mM NaCl, 2 mM EDTA, 0.1% sodium dexycholate pH 7.4 containing a protease inhibitor cocktail and 2 nM PMSF (Cruz-Vera et al., 2003). EhCP protein remaining after treatment with cycloheximide was calculated as percentage protein remaining of levels at time zero (0 h). Protein half-life (t1/2) was calculated by linear regression analysis.

# Mice Colonic Loop Studies

Math1GFP mice were purchased from Jackson laboratory and bred in-house. Colonic loops studies were done by inoculating live Eh suspended in 100 µL PBS (1 × 10<sup>6</sup> ) into closed colonic loops as described previously (Belley and Chadee, 1999) This is a short-term infection model (3 h after infection). After 3 h, the colons were excised and tissue pro-inflammatory gene expression and myeloperoxidase activity was analyzed.

# Colonic Myeloperoxidase Activity Assay (MPO)

MPO activity was assayed in mouse colon samples as described previously (Kumar et al., 2017). Briefly, fresh frozen tissues were homogenized in 0.5% hexadecyltri-methylammonium bromide. Homogenized tissue was freeze-thawed three times, sonicated, and centrifuged at 10,000 g for 10 min at 4◦C. Clear supernatant was collected and the reaction was initiated by addition of 1 mg/ml dianisidine dihydrochloride (Sigma, St. Louis, MO) and 1% H2O2, and change in optical density was measured at 450 nm.

# Math1 Expression via Non-invasive Whole-Body Imaging *ex vivo*

After 3 h of Eh infection, colons of Math1GFP mice were surgically removed, imaged ex vivo using an in vivo Xtreme 4MP-imaging platform (Bruker, Billerica, MA, USA) to detect GFP expression. The imaging was performed in two steps. The first one is reflectance imaging (2 s exposure time) and the second one was with excitation at 470 nm and emission at 535 nm (5 s exposure time) i.e., fluorescent imaging. Images were acquired and analyzed from the in vivo Xtreme using Bruker molecular imaging software MI SE (version 7.1.3.20550). GFP expression in the colon was quantified by measuring the mean fluorescence (after background subtraction) in a constant ROI.

# Ethics Statements

All studies were carried out with the approval of the University of Calgary Animal Care Committee. Animal care committee have approved experimental procedure proposed and certifies that animal care was in accordance with recent policies by the Canadian council on Animal care.

# Statistical Analysis

Data was analyzed using Graphpad Prism 7 (Graph-Pad Software, San Diego, CA) for all statistical analysis. Student's t-test was used when two groups were compared. Statistical significance was assumed at P < 0.05.

# RESULTS

# Silencing of the *Eh* Cyclooxygenase Like Gene Responsible for PGE<sup>2</sup> Biosynthesis

Eh-derived PGE<sup>2</sup> not only induces pro-inflammatory IL-8 production but also disrupts colonic epithelial cell tight junction by coupling through EP4 receptors (Dey et al., 2003; Dey and Chadee, 2008; Lejeune et al., 2011). To analyze the biological functions of endogenous EhCox and EhPGE<sup>2</sup> mediated proinflammatory responses, we silenced the expression of the gene by small RNA-mediated transcriptional gene silencing in the G3 strain (Bracha et al., 2006). The specific gene repression was confirmed by reverse transcription PCR of corresponding cDNA and immunoblot analysis of proteins by using EhCox specific antibody. Complete silencing of EhCox was achieved in comparison to the Eh control (**Figures 1A,B**, **Supplementary Figure 1**). Immunoblot analysis detected the 72 and 66 Kda protein band in the lysate of control Eh as previously described (Dey et al., 2003). As predicted, EhCox enzymatic assay using nuclear fractions isolated from log-phase Eh showed almost no aspirin (ASA) inhibited PGE<sup>2</sup> release (Dey et al., 2003) by EhCoxgs in comparison to control Eh incubated with arachidonic acid (AA) substrate (**Figure 1C**). We used ASA inhibited PGE<sup>2</sup> release to accurately quantify PGE<sup>2</sup> levels as EhCoxgs showed modest non-specific binding by enzyme immunoassay (EIA) that was not inhibited with ASA.

# Growth Kinetics of *EhCoxgs*

To determine whether EhCoxgs had an effect on cell proliferation, the growth kinetics of EhCoxgs and control Eh were compared. We analyzed the growth of Eh over time and showed that during the first 24 h growth kinetics were similar, however at 48 and 72 h EhCoxgs proliferation was significantly slower than control Eh (**Figure 1D**). These results suggest that EhCox might be essential for cell growth and proliferation.

# *Eh*Cox Gene Silencing Caused Limited Proteome Change

To determine if EhCoxgs also affected the expression of other proteins, we took a proteomic approach and analyzed the proteome from EhCoxgs and control Eh. Only a limited number of proteins showed three-fold or higher expression (**Supplementary Table 1**). In EhCoxgs, 23 proteins were up regulated and 19 proteins were down regulated compared to control Eh (**Table 2**). EhCox protein (Q9U3Z8\_ENTHI) was not detected in EhCoxgs proteome that confirms complete absence of the EhCox protein and gene specific silencing. Among the proteins that were up regulated included those encoding for several uncharacterized proteins (M7X297\_ ENTHI, A0A175JHX8\_ENTHI, A0A175JJQ8\_ENTHI, A0a175JU71\_ENTHI, A0A175JQL9\_ENTHI, A0A175LQ0\_ ENTHI, A0A175JJP8\_ENTHI) and Rab family GTPase proteins (A0A175JL18\_ENTHI, A0A175JT01\_ENTHI, A0A175JSR0\_ ENTHI), UDP-glucose:glycoproteine glucosyltransferase (C4M0W6\_ENTHI), WD domain containing protein

FIGURE 1 | Silencing of the Cox-like protein in E. histolytica. (A) qPCR was used to monitor cox expression using cDNA from E. histolytica. The data indicate the changes in mRNA expression compared with controls. (B) Immunoblot blot analysis was performed on lysate prepared from control Eh and EhCoxgs. The proteins were separated on 8% SDS PAGE gels and analyzed with a polyclonal cox antibody or actin antibody. (C) PGE2 production was quantified using enzyme-linked immunosorbent assay kits following treatment of EhNP with or without AA or aspirin (ASA) for 1 h at 37◦C. The bars indicate the means and the error bars indicate the standard errors of the means for three different experiments. (D) Approximately 2.5 × 10<sup>5</sup> control Eh and EhCoxgs in the logarithmic growth phase were inoculated into 14 mL fresh culture medium and Eh were counted every 24 h. Data shown are the means and the error bars indicate the standard errors of the means for three different experiments. The asterisks indicate the results of comparisons with the controls. \*P < 0.05, \*\*P < 0.01, \*\*\*\*P < 0.001. Eh, E. histolytica.

(A0A175JMP8\_ENTHI), Galactose specific adhesin light subunit (A0A175JFH2\_ENTHI) and cysteine protease (A0A175JGF5\_ENTHI) suggesting that increased expression of these genes particularly Rab family GTAPase and WD domain containing proteins that regulate vesicular trafficking of cysteine protease, may be involved in EhCox-mediated cysteine protease upregulation. The proteomics data showed a slightly increase in other cysteine proteases including CP1 and CP2 (**Table 3**). However, we did not detect EhCP-A4 or EhCP-A5 specifically in the dataset (**Table 2**, **Supplementary Table 1**). These observations led us to hypothesize that the increased expression and activity of CP5 and CP4 was not limited to a specific protease but rather, all the major CPs were regulated by EhCox. Similar to up regulated proteins we found several down regulated uncharacterized proteins (A0A175JKF9\_ENTHI, C4LZW1\_ENTHI, A0A175JJ73\_ENTHI, A0A175JJ95\_ENTHI). Other down regulated proteins were DNAj family protein (A0A175JFG6\_ENTHI), Ras guanine nucleotide exchange factor (A0A175JYJ4\_ENTHI), Ph domain containing protein (A0A175JQH4\_ENTHI), and V-type proton ATPase subunit (A0A175JK72\_ENTHI).

# Gene Silencing of *Eh*Cox Increases Cysteine Protease Expression and Activity

Cysteine proteases play major roles in the pathogenesis of Eh. Specifically, EhCP-A5 has been shown to be involved in tissue invasion by disrupting the protective mucus layer and stimulating pro-inflammatory response in colonic cells (Moncada et al., 2003; Hou et al., 2010). To determine the functional significance of EhCox derived PGE<sup>2</sup> in Eh virulence, we analyzed the expression of EhCPs protein and mRNA in EhCoxgs. Cell lysates from EhCoxgs and control Eh were assayed for CP expression by western blotting using EhCP-A5 and EhCP-A4 specific antibodies. These proteases are highly expressed and released extracellular basally and during infection (He et al., 2010; Kissoon-Singh et al., 2011). An unexpected finding was that EhCoxgs showed a slight increased in EhCP-A5 and significantly up regulated EhCP-A4 protein expression compared to control Eh (**Figure 2A**, **Supplementary Figure 2**). Surprisingly, we did not detect a corresponding increase in CPs transcripts by qPCR (**Figure 2B**) using primer specific for CP1, CP2, CP4, and CP5 (**Table 1**). These results suggest that EhCox was regulating CPs at the translational level independent of transcription. To assess whether the increase in EhCPs protein expression resulted in a corresponding increase in EhCPs enzymatic activity, gelatin substrate gel electrophoresis was performed with Eh lysates that showed prominent bands of CPs activity in EhCoxgs as compared to control Eh lysate which was not present in E64 (CP inhibitor) treated Eh lysate (**Figure 2C**). To verify increased enzymatic activity, EhCoxgs and control Eh lysates were incubated with known substrates and proteinase activity was quantified by the liberation of chromogenic leaving group p-nitroanilide TABLE 2 | List of genes that are up or down regulated ≥3-fold upon EhCox gene silencing.


and the fluorescent leaving group 7-amino-4-methylcoumarin (AMC) from EhCP5 and EhCP4 peptide substrates, Z-Arg-ArgpNA and Z-VVR-AMC, respectively. Degradation of EhCPs substrate occurred in linear mode over time (**Figures 2D,E**) with significantly higher enzymatic activity in EhCoxgs as compared to control Eh lysates. Specificity for CPs enzymatic activity was confirmed using specific EhCP-A5 and EhCP-A4 inhibitors, WRR483 and WRR605, respectively (**Figures 2D,E**). EhCP-A5 and EhCP-A1 are unique as they both cleave the common substrate Z-Arg-Arg-pNA and here we used the EhCP-A1 specific inhibitor WRR483 to show that EhCP-A5 enzyme activity was specifically increased (St-Pierre et al., 2017). E64 inhibits all CPs in Eh. As EhCPs are secreted extracellular we also found significantly increased EhCP5/4 enzyme activity in EhCoxgs



compared to control Eh (**Figures 2F,G**). These results clearly show that EhCoxgs expressed higher levels and activity of CPs, which was regulated by EhCox protein.

# Effect of Arachidonic Acid (AA), Aspirin and Prostaglandin E<sup>2</sup> on Cysteine Protease Activity

From the studies above it was unclear if inhibition of Eh PGE<sup>2</sup> biosynthesis and/or Cox enzyme activity was regulating CP expression and enzyme activity in EhCoxgs. In live Eh, PGE<sup>2</sup> is produced in a time-dependent manner in the presence of AA substrate (Dey et al., 2003). AA is the most important ratelimiting step in EhCox driven biosynthesis of PGE<sup>2</sup> and aspirin is the only inhibitor known to inhibit EhCox enzymatic activity (Dey et al., 2003). To evaluate if PGE<sup>2</sup> played a role in regulating EhCP5 activity, control Eh was incubated with exogenous AA or aspirin and they had no effect on enzyme activity (**Figures 3A,B**). Similarly, the addition of exogenously prostaglandin (PGE2) to EhCoxgs showed no difference in EhCP5 activity (**Figure 3C**). These results strongly suggest that the increase in EhCP5/4 activity was not dependent on Cox enzymatic activity but rather appear to be a direct effect of the Cox protein on EhCP expression and activity.

# *EhCoxgs* Stabilizes *Eh*CP-A5 Protein Degradation

To address the mechanism whereby EhCoxgs increased protein expression and enzymatic activity, we hypothesize that the CPs in EhCoxgs were not degraded based on increased protein expression (**Figure 2A**). As several proteins were up/down regulated in EhCoxgs (**Table 2**), we were surprised that more proteins were ubiquitinated and destine for 26S proteasome degradation in EhCoxgs as compared to Eh (**Figure 4A**). Based on these findings we theorize that proteins critical in regulating the stability of EhCPs maybe degraded and quantified the half-lives of both EhCP-A4/5. To do this, Eh and EhCoxgs were treated with cycloheximide and the percentage protein remaining over 24 h determined. Surprisingly, EhCP-A4 protein was not degraded whereas the half-life for EhCP-A5 was 19.3 h in EhCoxgs as compared to 12.2 h for control Eh (**Figures 4B,C**). The increase in EhCP-A5 protein stability in EhCoxgs may account for increase protein accumulation and increase enzyme activity. In contrast, EhCP-A4 protein was very stable and turnover rate low. We did not treat Eh longer than 24 h with cyclohexamide (95% viable by trypan blue exclusion) as cells became rounded and detached from the glass tubes and we were concern about cell death.

# Increased Cysteine Protease Expression Results in Increased Erythrophagocytosis and Cytopathic Activity

Experimental evidence suggests a direct correlation between Eh virulence and the rate of phagocytosis and protease activity (Ankri et al., 1998, 1999; Okada et al., 2005; Hirata et al., 2007). Erythrophagocytosis is considered as one of the prominent marker of Eh virulence (Trissl et al., 1978; Orozco et al., 1983; Bhattacharya et al., 2002). Accordingly, we assayed the ability of EhCoxgs to phagocytose fluorescently labeled RBCs and measured the fluorescence intensity by confocal microscopy. EhCoxgs showed significantly higher RBCs uptake in comparison to control Eh (**Figure 5A**). EhCoxgs also significantly destroyed 27% of a Caco-2 monolayer after 2 h incubation compared to control Eh that destroyed 11% (**Figure 5B**). These results show that EhCoxgs is highly phagocytic with increased cytopathic activity.

# Differential Math1 Transcriptional Activity in Math1GFP Mice Exposed to *EhCoxgs*

Based on the results above, we then determined if there was a similar increase in EhCoxgs virulence using closed colonic loops in mice (Kissoon-Singh et al., 2013). We have previously shown that Eh induces hyper secretion of mucus by goblet cells and elicits an acute pro-inflammatory response in colonic loops (Dharmani et al., 2009). In particular, EhCP5 RGD motif has been shown to bind αvβ3 integrin on goblet cells to elicit mucin hyper secretion (Cornick et al., 2016). We hypothesized that increase CPs activity would result in a differential response toward mucin biosynthesis and secretion to EhCoxgs as compared to control Eh. To interrogate this, we used Math1GFP mice containing the green fluorescent protein (GFP) reporter for Math1-expressing goblet cells. In the colon, Math1 is expressed in epithelial cells to differentiate into Muc2-producing goblet cells lineage. Basally, Math1GFP activity was higher in the proximal colon in control mice. However, following Eh infection, Math1 activity was significantly decreased in the proximal colon, which was decreased further when mice were infected with EhCoxgs (**Figure 6**). We have previously shown that decrease in Math1GFP activity correlates with increase pro-inflammatory activity in response to DSS-induced colitis (Tawiah et al., 2018).

# Pro-inflammatory Responses Are Exacerbated in Math1 Mice Exposed to *EhCoxgs* Compared Control *Eh*

It is well-known that Eh infection in the gut elicits an acute proinflammatory response with the production of pro-inflammatory cytokines IL-1β, IL-8, IFN-γ, and TNF-α (Bansal et al., 2009; Galván-Moroyoqui et al., 2011). Based on the results above showing a marked decrease in Math1GFP activity in the proximal

FIGURE 2 | Gene silencing of EhCox increases cysteine protease expression and activity. (A) Immunoblot blot analysis was performed on lysate prepared from control Eh and EhCoxgs. The proteins were separated on 12% SDS PAGE gels and analyzed with CP5, CP4 or actin antibody. Quantifications of CP5/4 were performed by densitometric analysis from three independent experiments shown in right panel. (B) qPCR was used to monitor CPs expression using cDNA from Eh and EhCoxgs. Data indicate changes in mRNA expression compared with controls. (C) Gelatin substrate gel electrophoresis of lysate from control Eh and EhCoxgs treated/untreated (Continued)

FIGURE 2 | with E64. (D,E) CPs enzymatic activity was evaluated by incubating inhibitor-treated/non-treated Eh with EhCP-A5 or EhCP-A4 substrates for 10 and 20 min, respectively and calculated in µM/min/mg, shown in the bottom panel. (D) EhCP-A5 enzymatic activity in lysate with known substrates (Z-RR) and 20µM of inhibitor WRR483. (E) EhCP-A4 enzymatic activity in lysate with known substrates (Z-VVR) and 20µM of inhibitor WRR605 and 100µM E64. (F) EhCP-A5 and (G) EhCP-A4 enzymatic activity in secreted protein. The bars indicate the means and the error bars indicate the standard errors of the means for three different experiments. The asterisks indicate the results of comparisons with the controls. \*P < 0.05, \*\*P < 0.01, \*\*\*P < 0.002, \*\*\*\*P < 0.001.

FIGURE 4 | EhCoxgs stabilizes EhCP-A5 protein degradation. (A) Log phase Eh and EhCoxgs proteins (30 µg) were loaded on a 12% SDS-PAGE gel and immunoblotted with an ubiquitin antibody (P4D1). Actin was used as a loading control. (B,C) The half-life of EhCP-A4 and EhCP-A5 protein was determined by treating Eh with cyclohexamide (see details in Materials and Methods) and the remaining CPs quantified by Western blot over 24 h. Protein half-life (t½) was calculated using regression analysis.

colon, we hypothesize that enhanced pro-inflammatory cytokines were suppressing Math1 activity. Indeed, pro-inflammatory cytokines expression in colonic tissues after 3 h exposure with EhCoxgs showed increased expression of IL-1β and KC (human IL-8 homolog) but not TNF-α and IFN-γ as compared to Ehinoculated loops (**Figure 7**). Both IL-1β and KC are released by epithelial and immune cells recruited to the site of infection that can exacerbate tissue injury (Mortimer and Chadee, 2010). Specifically, KC is a potent neutrophil chemo attractant that recruits neutrophils to the site of infection but is ineffective in clearing the parasite. Pro-inflammatory cytokine expression correlated with increased myeloperoxidase (MPO) activity in response to EhCoxgs as compared to control Eh (**Figure 7**). These results reveal that EhCoxgs with increased CP enzyme expression and activity enhanced virulence in the mouse colon to elicit increased pro-inflammatory responses.

# DISCUSSION

In this study, we describe how Eh cysteine protease expression and activity is regulated by another enzyme, cyclooxygenase (Cox). Cox is the critical rate limiting step in the biosynthesis of PGE<sup>2</sup> (Smith and Marnett, 1991; Vane et al., 1998). Cox was identified in Eh that showed little homology to known Cox1/2 enzyme across different species with absence of conserved arachidonic acid binding domain and catalytic site that are present in other species. However, endogenous and recombinant EhCox showed Cox like enzymatic activity by using arachidonic acid as substrate that was inhibited only by aspirin but not by another Cox-1/2 inhibitors (Dey et al., 2003). In a later study, Cox like protein (Acc No. AF208390) was characterized as an actinin like protein in Eh as it has actin- and calcium-binding domains. The actin binding domain of Eh actinin like protein share only 30% identity with other actin binding proteins however, the protein showed 28% sequence identity to D. discoideum actinin protein. Based on some of the unusual domain architecture feature of Eh actinin like proteins it was proposed that this unusual protein might differ in function from known actinin proteins (Heike et al., 2005). Beside actin bundling, multiple cellular functions of the actinin protein have been proposed with putative interacting partners (Otey and Carpen, 2004; Cabello et al., 2007). For example, mammalian actinin regulates several receptor activity by linking the cytoskeleton to a variety of trans membrane proteins (Cabello et al., 2007). From these studies and our findings, we proposed that besides Cox like activity and actin polymerization function, this protein might have multifunctional roles in the regulation of several other proteins in Eh. Based on its molecular structure and proposed multiple functions we developed EhCoxgs to determine its biological function in the pathogenesis of amebiasis.

Cysteine proteases (CPs) are ubiquitous and are differentially regulated in virulent and non-virulent strains of Eh. Previous studies have showed that CPs are key virulent factors in the pathogenesis of Eh that are released during tissue invasion and

AU, arbitrary units. \*P < 0.05.

one of the most important protein involved in phagocytosis (Okada et al., 2005; Hirata et al., 2007). In the present study, we have shown that silencing the Cox gene increased cysteine protease expression and activity endogenously and extracellular without affecting CPs transcript. Increased cysteine protease activity led us to hypothesize that Cox derived PGE<sup>2</sup> was negatively regulating cysteine protease activity. However, the addition of exogenous arachidonic acid to live Eh to increase PGE<sup>2</sup> production or aspirin to inhibit PGE<sup>2</sup> or exogenous PGE<sup>2</sup> did not affect CP activity. These observations suggest that increase in CPs expression and activity was not the effect of Cox enzymatic activity/end product of biosynthesis but rather was regulated by the Cox protein itself. The EhCox protein acted as a negative regulator of CPs. Since expression of CPs was not regulated at the transcriptional level, there is the possibility of post-translational modification of CPs up regulating its expression and enzymatic activity. In support of this we have shown increased EhCP-A5 protein stability in EhCoxgs as compared to control Eh. While the mechanism of this posttranslational regulation is not known, our results suggest that proteins critically involved in regulating EhCP-A5 turnover were degraded in EhCoxgs by the appearance of more ubiquitinated proteins. In contrast, EhCP-A4 protein was stable and turnover rate low over 24 h that suggest different regulatory mechanisms between the CPs. No doubt uncovering the post-translational regulation of CPs enzyme will provide the basis to understand the mechanism of Cox mediated regulation and promote the development of more efficient therapeutic strategies of indirectly targeting CPs enzyme.

Proteomic analysis of the EhCoxgs revealed high expression of Rab family GTAPase and WD domain containing proteins as compared to control Eh. These proteins are involved in various cellular process including membrane trafficking, cytoskeletal assembly and cell proliferation (Saito-Nakano et al., 2005; Nakada-Tsukui and Nozaki, 2015). Based on this observation, we proposed that these proteins might be involved in up regulating CPs when EhCox is silenced. However, it needs to be further determined what functional advantages or constraints drive CPs up regulation. We also found iron-sulfur falvoprotein (A0A175JR31, M7W6A3; **Supplementary Table 1**) among the down regulated proteins in EhCoxgs which supports reduced growth of EhCoxgs as these proteins have been shown to be essential for the growth and survival of Eh under different condition (Nozaki et al., 1998; Shahi et al., 2016). However, the question that remains to be answered is the mechanism of how EhCox regulates Eh growth.

Phagocytosis is an active process in Eh and prominent marker of parasite pathogenicity. Phagocytosis involves several steps and activation of signaling pathways (Orozco et al., 1983; Hirata et al., 2007). A proteomic study of phagosome showed the involvement of several proteins in this process including cysteine protease and vesicular transport proteins (Okada et al., 2005). Since more cysteine protease (activity assay and expression) and several vesicular trafficking proteins (proteomics analysis) were observed in EhCoxgs, we analyzed the phagocytosis capacity of EhCoxgs as compared to control Eh. Unexpectedly, we found increased erythrophagocytosis and cytopathic activity in EhCoxgs as compared to control Eh. While it is difficult to explain how increased in Eh CPs can enhance phagocytosis, these data are consistent with previous studies that showed CPs are directly involved in destruction of colonic epithelial cells, tissue invasion, phagocytosis and degrade host antibodies and complement in immune evasion and disease pathogenesis (Hirata et al., 2007; Mortimer and Chadee, 2010; Nakada-Tsukui and Nozaki, 2016).

Several in vivo and ex vivo assays reported the recruitment of inflammatory cell in Eh-induced infection/lesion as a consequence of localized expression of chemokines and cytokines at the site of infection, leading to an inflammatory response (Bansal et al., 2009; Kissoon-Singh et al., 2013; Nakada-Tsukui and Nozaki, 2016). To address whether EhCoxgs were more virulent, we used closed colonic loop as short infection model to quantify acute inflammatory responses. Our study revealed that the Cox protein played a major role in Eh- induced fluid secretions and pro-inflammatory cytokine responses by regulating CPs activity. In particular, EhCoxgs elicited high levels of IL-1β and KC expression demonstrating cysteine protease dependent induction of IL-1β and IL-8 expression as observed in Eh infected SCID mouse-human intestinal xenograft (Seydel et al., 1997). Previous studies have shown CP5 to be important in enhancing mucosal inflammation by cleaving the released pro-IL-1β into its active form and inducing IL-8 expression in mast cell (Zhang et al., 2000; Lee et al., 2014). Both TNF-α and IFN-γ were up regulated in response to Eh and was not enhanced further by EhCoxgs. In EhCoxgs inoculated colonic loops we also found elevated MPO activity, a marker for increased flux of neutrophils that are responsible for exacerbating tissue injury.

Increased CP expression and virulence in EhCoxgs is intriguing which led us to propose the following hypotheses. First, EhCox is as endogenous inhibitor of cysteine proteases. While examples of identified cysteine proteinase inhibitors produced by parasites are rare they can be targeted to treat disease related to increased protease activity. For example, papain inhibitors were detected in parasitic protozoa including Leishmania, Trichomonas, and Trypanosoma, suggesting existence of these inhibitors are widespread (Irvine et al., 1992). In Schistosoma mansoni, a gene was identified which encode papain inhibitors (Cao et al., 1993). In Eh, few CP inhibitors have been characterized that negatively regulates CP secretion and thus, virulence of the parasite (Sato et al., 2006). Second, CP regulation by Cox is a negative feedback mechanism to reduce host inflammation. This negative feedback regulation may counteract excessive cysteine protease function at site of inflammation and thus, decrease the likelihood of tissue damage that led to amebic lesion/colitis. This phenomenon may furthered explain why most Eh infections are asymptomatic. Consistent with this hypothesis, a study in mice showed that ribosomal protein S19 interact with macrophage migration inhibitory factor and function as an extracellular inhibitory factor by attenuating its pro-inflammatory function (Filip et al., 2009). Another study in Arabidopsis thaliana proposed the formation of a complex network of cysteine protease-Serpin1 interaction controlling innate immunity during plant development (Rustgi et al., 2017). Based on these findings it would be interesting to determine whether these proteins can interact together; it is likely these interactions could regulate CPs activity either as a result of conformational changes of the enzyme or by impacting subcellular localization of CPs and thus affecting its interactions with Cox, to further modulate its activity. Furthermore, the effect of this interaction on activity and expression of cysteine protease or vice versa can be analyzed. This information can be used to develop chemotherapeutic target against Eh infection. Clearly further studies are needed to understand the underlying molecular mechanisms by which EhCoxgs increases CPs expression and activity.

# AUTHOR CONTRIBUTIONS

PS and KC conceived and design the experiments and wrote the manuscript. PS and FM performed the experiments. PS analyzed the data.

# FUNDING

This work was supported by an operating grant from the Canadian Institutes of Health Research (KC; MOP-142776).

# REFERENCES


# ACKNOWLEDGMENTS

We thank the staff of the SAMS Center at the University of Calgary for help with proteomic analysis of the data. We also thank Dr. Björn Petri from the Snyder Mouse Phenomics Resources Laboratory for acquiring the non-invasive wholebody imaging ex vivo and the Snyder Live Cell Imaging facility for technical support. PS was supported by an Eyes High Postdoctoral Scholar award from the University of Calgary.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00447/full#supplementary-material

Supplementary Table 1 | Entire proteomics profile of control Eh and EhCoxgs with mass spectrometry counts, accession number and protein ID.

Supplementary Figure 1 | Full scan of blots for Cox protein shown in Figure 1B.

Supplementary Figure 2 | Full scan blots of CP5/4/actin shown in Figure 2A.

receptor. J. Biol. Chem. 282, 12143–12153. doi: 10.1074/jbc.M6088 80200


Verkerke, H. P., Petri, W. A., and Marie, C. S. (2012). The dynamic interdependence of amebiasis, innate immunity, and undernutrition. Semin. Immunopathol. 34, 771–785. doi: 10.1007/s00281-012-0349-1

World Health Organization (1997). Amoebiasis. Wkly. Epidemiol. Rec. 72, 97–100. World Health Organization (1998). The World Health Report 1998: Life in the 21st

Century A Vision for All. World Heal. Organ. 226. Zhang, Z., Yan, L., Wang, L., Seydel, K. B., Li, E., Ankri, S., et al. (2000). Entamoeba histolytica cysteine proteinases with interleukin-1 beta converting enzyme (ICE) activity cause intetinal inflammation and tissue damage in amoebiasis. Mol. Microbiol. 37, 542–548. doi: 10.1046/j.1365-2958.2000. 02037.x

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 Shahi, Moreau and Chadee. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# *Entamoeba histolytica* Up-Regulates MicroRNA-643 to Promote Apoptosis by Targeting XIAP in Human Epithelial Colon Cells

Itzel López-Rosas <sup>1</sup> , César López-Camarillo<sup>2</sup> \*, Yarely M. Salinas-Vera<sup>2</sup> , Olga N. Hernández-de la Cruz <sup>2</sup> , Carlos Palma-Flores <sup>3</sup> , Bibiana Chávez-Munguía<sup>4</sup> , Osbaldo Resendis-Antonio<sup>5</sup> , Nancy Guillen<sup>6</sup> , Carlos Pérez-Plasencia7,8 , María Elizbeth Álvarez-Sánchez <sup>2</sup> , Esther Ramírez-Moreno<sup>9</sup> and Laurence A. Marchat <sup>9</sup> \*

### *Edited by:*

Serge Ankri, Technion Israel Institute of Technology, Israel

#### *Reviewed by:*

Carlo José Freire Oliveira, Universidade Federal do Triângulo Mineiro, Brazil Bellisa Freitas Barbosa, Federal University of Uberlandia, Brazil

#### *\*Correspondence:*

César López-Camarillo genomicas@yahoo.com.mx Laurence A. Marchat lmarchat@gmail.com

#### *Specialty section:*

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

*Received:* 19 August 2018 *Accepted:* 10 December 2018 *Published:* 08 January 2019

#### *Citation:*

López-Rosas I, López-Camarillo C, Salinas-Vera YM, Hernández-de la Cruz ON, Palma-Flores C, Chávez-Munguía B, Resendis-Antonio O, Guillen N, Pérez-Plasencia C, Álvarez-Sánchez ME, Ramírez-Moreno E and Marchat LA (2019) Entamoeba histolytica Up-Regulates MicroRNA-643 to Promote Apoptosis by Targeting XIAP in Human Epithelial Colon Cells. Front. Cell. Infect. Microbiol. 8:437. doi: 10.3389/fcimb.2018.00437 <sup>1</sup> Catedrática CONACYT, Laboratorio de Genómica Funcional y Biología Molecular, Colegio de Postgraduados Campus Campeche, Campeche, Mexico, <sup>2</sup> Posgrado en Ciencias Genómicas, Universidad Autónoma de la Ciudad de Mexico, Mexico City, Mexico, <sup>3</sup> Catedrático CONACYT, Instituto Politécnico Nacional, Mexico City, Mexico, <sup>4</sup> Departamento de Infectómica y Patogénesis Molecular, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico City, Mexico, <sup>5</sup> Instituto Nacional de Medicina Genómica y Coordinación de la Investigación Científica, Red de Apoyo a la Investigación, Universidad Nacional Autónoma de Mexico, Mexico City, Mexico, <sup>6</sup> Unidad de Análisis Cuantitativo de Imágenes, Instituto Pasteur, Paris, France, <sup>7</sup> Unidad de Biomedicina, Facultad de Estudios Superiores-Iztacala, Universidad Nacional Autónoma de México, Mexico City, Mexico, <sup>8</sup> Instituto Nacional de Cancerología, Mexico City, Mexico, <sup>9</sup> Programa en Biomedicina Molecular y Red de Biotecnología, Escuela Nacional de Medicina y Homeopatía, Instituto Politécnico Nacional, Mexico City, Mexico

MicroRNAs (miRNAs) are small non-coding RNAs that function as negative regulators of gene expression. Recent evidences suggested that host cells miRNAs are involved in the progression of infectious diseases, but its role in amoebiasis remains largely unknown. Here, we reported an unexplored role for miRNAs of human epithelial colon cells during the apoptosis induced by Entamoeba histolytica. We demonstrated for the first time that SW-480 colon cells change their miRNAs profile in response to parasite exposure. Our data showed that virulent E. histolytica trophozoites induced apoptosis of SW-480 colon cells after 45 min interaction, which was associated to caspases-3 and -9 activation. Comprehensive profiling of 667 miRNAs using Taqman Low-Density Arrays showed that 6 and 15 miRNAs were significantly (FC > 1.5; p < 0.05) modulated in SW-480 cells after 45 and 75 min interaction with parasites, respectively. Remarkably, no significant regulation of the 6-miRNAs signature (miR-526b-5p, miR-150, miR-643, miR-615-5p, miR-525, and miR-409-3p) was found when SW-480 cells were exposed to non-virulent Entamoeba dispar. Moreover, we confirmed that miR-150, miR-643, miR-615-5p, and miR-525 exhibited similar regulation in SW-480 and Caco2 colon cells after 45 min interaction with trophozoites. Exhaustive bioinformatic analysis of the six-miRNAs signature revealed intricate miRNAs-mRNAs co-regulation networks in which the anti-apoptotic XIAP, API5, BCL2, and AKT1 genes were the major targets of the set of six-miRNAs. Of these, we focused in the study of functional relationships between miR-643, upregulated at 45 min interaction, and its predicted target X-linked inhibitor of apoptosis protein (XIAP). Interestingly, interplay of amoeba with SW-480 cells resulted in downregulation of XIAP consistent with apoptosis activation. More importantly, loss of function studies using antagomiRs showed that forced inhibition of miR-643 leads to restoration of XIAP levels and suppression of both apoptosis and caspases-3 and -9 activation. Congruently, mechanistic studies using luciferase reporter assays confirmed that miR-643 exerts a postranscripcional negative regulation of XIAP by targeting its 3′ - UTR indicating that it's a downstream effector. In summary, we provide novel lines of evidence suggesting that early-branched eukaryote E. histolytica may promote apoptosis of human colon cells by modulating, in part, the host microRNome which highlight an unexpected role for miRNA-643/XIAP axis in the host cellular response to parasites infection.

Keywords: *Entamoeba histolytica*, apoptosis, microRNAs, SW480, XIAP

# INTRODUCTION

Entamoeba histolytica is the single-celled protozoan parasite causative of human amoebiasis that affects between 40 and 50 million people worldwide. About 10% of infected individuals are at risk for developing invasive amoebiasis, namely amoebic colitis and extra-intestinal disease, such as amoebic liver abscesses that can be fatal (Stanley, 2003). The parasite infection shown clinical variability associated to intestinal microbiota composition that may increase resistance to infection by decreasing the virulence properties and altering systemic immunity against parasites (Burgess et al., 2017). Indeed, specific gut microbiota patterns have been linked to colonization with parasitic protists. For instance, it was reported a differential fecal microbiota in subjects infected with Giardia duodenalis or Entamoeba spp./Blastocystis hominis (Iebba et al., 2016). Another study found that the Entamoeba is significantly correlated with microbiome composition and diversity, and that colonization can be predicted with 79% accuracy based on the composition of an individual's gut microbiota (Morton et al., 2015). Gilchrist et al. also reported that a high parasite burden coupled with increased levels of Prevotella copri was linked to symptomatic infection with E. histolytica in children (Gilchrist et al., 2016). In addition, dysbiosis induced by antibiotic treatment increased the severity of amebic colitis and delayed clearance of E. histolytica in an amoebic colitis mouse model (Watanabe et al., 2017). These data urge for a better understanding of the mechanisms underlying microbiota-mediated protection that may help explain clinical variability and help treat amoebiasis.

The main site of E. histolytica infection is the colon epithelium. Tissues damage resulting from adhesion, lysis, and phagocytosis of host cells is caused by the activity of several parasite proteins; however, the molecular mechanisms by which trophozoites cause epithelial damage are not fully understood. The activity of several parasite proteins including cysteine proteases (Sajid and McKerrow, 2002), the Gal/GalNAc lectin (Petri and Schnaar, 1995), and amoebapores (Leippe, 1997) among others, is important for disruption and invasion of colonic mucosa by E. histolytica trophozoites. Moreover, adherence of virulent amoebae to host cells results in cell death, mainly by apoptosis, both in vitro (Berninghausen and Leippe, 1997; Sim et al., 2007) and in vivo (Moncada et al., 2006), as well as in tissue inflammatory response (Seydel et al., 1997, 1998; Seydel and Stanley, 1998). These events are the result of the ability of parasites to alter gene expression in host cells. Several reports confirmed these assumptions, for instance genomewide transcriptional analyses of mouse liver cells revealed the impact of E. histolytica on transcription of infected cells which contributes to the activation of apoptosis, regenerative and inflammatory cellular pathways in host cells (Pelosof et al., 2006). Also, transcriptional response to adhesion of virulent parasites to liver sinusoidal endothelial cells leads to death and actin cytoskeleton disorganization of host cells (Faust et al., 2011). These data highlights the impact of E. histolytica on the gene expression programs of human cells during infection.

Over the last decade, microRNAs (miRNAs) have emerged as a new prominent class of negative regulators of gene expression. MiRNAs are evolutionary conserved small non-coding singlestranded RNAs of 21–25 nt length which function as guide molecules in posttranscriptional gene silencing by binding to the 3′ untranslated region (3′UTR) of target genes resulting in mRNA degradation or translational repression in P-bodies (Bartel, 2004). Notably, aberrant expression of microRNAs may greatly contribute to development of diverse infectious diseases. Interestingly, miRNAs have been investigated in the host-pathogen interactions including viral, bacterial, fungus, and parasitic infections where they mainly mediate inflammatory response and apoptosis in response to inflection (Drury et al., 2017). For instance, Toxoplasma gondii inhibits the apoptotic response of infected host cells through upregulation of miR-17-92 expression and downregulation of pro-apoptotic Bim in human macrophages challenged with parasites (Cai et al., 2014). In addition, infection of cholangiocytes with Cryptosporidium parvum, a protist causing intestinal and biliary disease, decreased the expression of primary let-7i and mature let-7 which in turn regulate TLR4 expression and contributes to epithelial immune responses against parasite infection (Chen et al., 2007). These findings suggested that miRNAs mediate post-transcriptional regulation of cellular pathways critical to host-cell regulatory responses to parasite infection. Importantly, miRNAs may also represent bonna fide biomarkers of infections, mainly in virus and bacterial infections, as its levels significantly differs in patients relative to healthy individuals. For instance, it was reported that miR-18a, miR-21, miR-29, miR-106b, and miR-122 were downregulated in serum of patients with Hepatitis B virus infection and liver cirrhosis relative to patients with chronic hepatitis B without liver cirrhosis. This set of miRNAs was able to distinguish both groups of patient's sensitivity of 85 and specificity of 70% (Jin et al., 2015). These data highlighted the potential of miRNAs as novel diagnostic tools in infectious diseases. However, no studies on miRNAs as potential amoebiasis biomarkers, neither clinical data about the relevance of miRNAs in outcome of patients with intestinal diseases have been reported yet. One mechanism by which pathogens may survive and disseminate in host tissues is by inducing cell death and apoptosis of cells. The effects of parasites on host miRNAs expression have been described in a few protozoan and nematode species, but nothing is known in E. histolytica. Moreover, the exact role of miRNAs in cell death of host cells is largely unknown in the context of amoebiasis. In this study, we aimed to determine whether the modulation of host cellular miRNAs is involved in the pathophysiological responses of epithelial colon cells to amoeba interactions. Our results revealed for the first time a miRNAs signature of human SW-480 and Caco2 epithelial colon cells in response to exposure to virulent E. histolytica trophozoites. Also, we showed a functional role for miR-643/XIAP axis in the apoptosis activation of host cells.

# MATERIALS AND METHODS

# Parasite and Cell Cultures

Virulent E. histolytica trophozoites (strain HM1:IMSS) were grown under axenic conditions in Diamond's TYI-S-33 medium at 37◦C (Diamond et al., 1978). Trophozoites in exponential phase of growth were used in all experiments. SW-480 (CCL-228) and Caco2 (HTB-37) human colorectal adenocarcinoma cell lines were purchased from ATCC collection and cultured in Dulbecco's modified Eagle's medium (DMEM; Invitrogen), supplemented with 10% fetal bovine serum, 100 U/ml penicillin and 100 mg/ml streptomycin. Cultures were maintained in a 5% CO<sup>2</sup> humidified atmosphere at 37◦C. For interaction assays, SW-480 and Caco2 cells (250,000) were cultured overnight in 6-well plates until reach 100% confluence, then amoeba was incubated with monolayers at 10:1 ratio (SW-480:amoeba) during 0, 15, 30, 45 and 75 min for downstream analysis.

# Transmission Electron Microscopy

After 15, 30, and 45 min interactions, SW-480/trophozoites cocultures were washed twice with PBS to remove any unattached amoeba and fixed for 60 min with 2.5 % (v/v) glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.2, and postfixed for 60 min with 1 % (w/v) osmium tetroxide in the same buffer. After dehydration, with increasing concentrations of ethanol and propylene oxide, samples were embedded in Polybed epoxy resins and polymerized at 60◦C for 24 h. Thin sections (60 nm) were obtained and stained with uranyl acetate and lead citrate for its examination in a Philips Morgagni 268 D electron microscope.

# miRNAs Profiling Using TaqMan Low Density Arrays

Profiling of microRNAs was performed as previously described with minor modifications (Flores-Pérez et al., 2016). Briefly, Total RNA was extracted from SW-480 cells, with or without interaction with E. histolytica trophozoites using the TRIzol reagent (Invitrogen). RNA integrity was assessed using capillary electrophoresis system Agilent 2100 Bioanalyzer with the eukaryotic nano-chip. Only samples with RNA integrity number (RIN) of six or higher were processed. Expression analysis of 667 human miRNAs was performed by reverse transcription and quantitative real-time polymerase chain reaction (RT-qPCR) using the Megaplex TaqMan Low-Density Arrays (TLDAs) v2.0 system (ThermoFisher), as described by the manufacturer. Briefly, 300 ng total RNA were retrotranscribed using stemloop primers. In order to detect low abundant microRNAs, a pre-amplification step was included. The pre-amplified product was loaded into the TLDA and amplification signal detection was performed in the 7900 FAST real time thermal cycler (ABI). Data were exported to the DataAssist software version 2.0 (Life Technologies) and normalized using the small-nucleolar RNA RNU44. Mean relative quantity (RQ) was calculated and microRNAs differentially expressed between groups (SW-480 cells interacting with trophozoites vs. SW-480 cells alone) were defined as those with fold change >1.5- and p-value < 0.05.

# Quantitative Real-Time Reverse Transcription-PCR (qRT-PCR)

The differential expression of six selected miRNAs (hsa-miR-409-3p, hsa-miR-526b-5p, hsa-miR-643, hsa-miR-150, hsa-miR-615-5p and hsa-miR-525) was verified using microRNA assays (Applied Biosystems, Foster City, CA, USA). Briefly, total RNA from SW-480 cells with and without interaction with E. histolytica, was reverse transcribed using a specific stemloop RT primer for each miRNA and the MultiScribeTM reverse transcriptase. Then, diluted retro-transcription reaction (1:15) was independently mixed with Universal PCR Master Mix, No AmpErase <sup>R</sup> UNG (2X), in presence of individual PCR probes for miR-409-3p, miR-526b-5p, miR-150, miR-643, miR-615-5p, and miR-525. The PCR reactions were done in a GeneAmp <sup>R</sup> PCR System 9700 (Applied Biosystems), using the following program: 95◦C for 10 min, and 40 cycles of 95◦C for 15 s and 60◦C for 1 min. The relative expression of microRNAs was measured by qRT-PCR using the comparative Ct (2−11CT) method. The snoRNA RNU44 was used as an internal control for data normalization.

# Prediction of miRNAs Targets

MiRNA targets were predicted by using four microRNA target prediction programs: DIANA-microT (http://diana.imis. athena-innovation.gr/DianaTools/index.php?r=microT\_CDS/ index) [44], TargetScan (http://genes.mit.edu/targetscan) [45], miRanda (http://www.microrna.org) [46] and PicTar (http:// pictar.mdc-berlin.de/). DIANA-microT identifies targets that are conserved in human and mouse; TargetScan reveals targets which are conserved in human, mouse, rat, chicken, and dog. miRanda detects targets which are conserved among human, mouse, and rat; and PicTar finds targets which are conserved in human, chimpanzee, mouse, rat, and dog. We selected all target-genes miRNAs pairs that were predicted by at least three programs. Biological pathways of predicted target-genes microRNAs were identified using mirPath (http://diana.imis. athena-innovation.gr/DianaTools/index.php?r=mirpath/index), miRWalk (http://www.umm.uni-heidelberg.de/apps/zmf/ mirwalk/predictedmirnapathway.php) and KEGG (http://www. genome.jp/kegg/pathway.html) databases.

# Apoptosis Assays

Apoptosis of SW-480 cells after 15, 30, 45, and 75 min interaction with E. histolytica trophozoites was quantified by flow cytometry. Annexin-V/propidium iodide (PI) double assay was performed using the annexin V-EGFP apoptosis detection kit (Roche). For interaction assays, SW-480 cells (250,000) were cultured overnight in 6-well plates until reach 100% confluence, then amoeba was incubated with monolayers at 10:1 ratio (SW-480:amoeba) during 0, 15, 30, 45, and 75 min. Subsequently, cells were washed twice with cold 1X PBS to detach amoebas. Then monolayers were seeded and resuspended in 400 µl binding buffer and stained with 5 µl annexin V-EGFP according to the manufacturer's recommendations. Then, 10 µl PI were added and cells were incubated for 5 min at 4◦C in the dark. Cells were analyzed in a FACS Calibur flow cytometer (BD Biosciences). The SW-480 cells cultured in DMEM or treated with 50µM cisplatin for 24 h were used as controls. Assay was repeated three times in triplicate and data were expressed as mean ± standard deviation. Statistical analyses were performed using the Student's t-test and a p-value < 0.05 was set as significant.

# Western Blots Assays

For interaction assays, SW-480 cells (250,000) were cultured overnight in 6-well plates until reach 100% confluence, then amoeba was incubated with monolayers at 10:1 ratio (SW-480:amoeba) during 0, 15, 30, and 45 min. Subsequently, cells were washed twice with cold 1X PBS to detach amoebas. Then remaining SW-480 cells were lysed in ice-cold cell lysis buffer (20 mM Tris, 250 mM NaCl, 2 mM EDTA, 1% Triton X-100, 0.4 mM PMSF), and 1X complete protease inhibitor cocktail (Roche). The protein concentration was determined with the Bradford protein assay kit (BioRad). Total protein extracts (30 µg) were separated through 15% SDS-PAGE and electrotransferred to a nitrocellulose membrane (BioRad) using standard protocols. After blocking with 5% dry skimmed milk, 0.2% Tween-20 in 1X PBS (pH 7.4), membranes were incubated with primary antibodies overnight at 4◦C.Then, they were washed twice in 1X PBS with 0.1% Tween-20 for 5 min and incubated for 1 h at room temperature with secondary antibodies conjugated to horseradish peroxidase (HRP) (Jackson). Finally, membranes were washed four times for 10 min in 1X PBS- 0.1% Tween-20 and developed with the ECL Western Blotting substrate (Amersham), according to the manufacturer's instructions. Caspase-3 and caspase-9 antibodies (1:1000 dilution, Cell Signaling) and XIAP antibody (1:3000 dilution, Cell Signaling) were used as primary antibodies. Bands were analyzed by densitometry (myImage Analysis, Thermo), and actin expression was used to normalize data.

# Inhibition of MIR-643 Expression Using Antagomirs

SW-480 cells were cultured in a 24-well plate (40,000 cells/well) and grown until reached a 100% confluence. Then cells were incubated with live E. histolytica trophozoites for 45 min. Subsequently, amoebas were removed by repeated washing of SW-480 monolayers with cold 1X PBS. Then, SW-480 cells were transfected with miR-643 inhibitor (at 30 and 50 nM) or scramble sequence (50 nM) using siPORT amine transfection agent (Ambion). Non-transfected SW-480 cells exposed to E. histolytica were used as control. After 24 h transfection, SW-480 cells were collected and total RNA was extracted using TRIzol reagent to evaluate miRNA-643 expression by stem-loop qRT-PCR and Western blot assays.

# Luciferase Gene Reporter Assays

A DNA fragment corresponding to the XIAP 3′UTR (which contains a miR-643 biding site) was synthesized and cloned into the XbaI site, downstream from the stop codon of the luciferase gene, in the pGL3 vector (GenScript) to obtain the pGL3- XIAP-3′UTR construct. SW-480 cells (40,000 cells/well) were transfected with the pGL3-XIAP-3′UTR plasmid (1 <sup>µ</sup>g) using lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. A pGL3-SIAH1-3′UTR construct that contains the 3 ′UTR of SIAH1 gene and lack of a miR-643 biding site was used as control. At 24 h postransfection SW-480 cells were incubated with E. histolytica trophozoites for 45 min, and then transfected with anti-miR-643 or scramble as described above. Luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega) at 24 h post transfection. Data were normalized against values corresponding to SW-480 cells transfected with pGL3-XIAP without interaction with E. histolytica. Each assay was repeated three times in duplicate, and data were expressed as mean ± standard deviation.

# Statistical Analysis

Statistical analyses were performed using the unpaired Student t-test to compare each condition with control cells and a p value<0.05 was considered as significant.

# RESULTS

# *E. histolytica* Alters miRNAs Expression During the Induced Apoptosis of Human SW-480 Colon Cells

Previous studies using in vitro and in vivo approaches demonstrated that E. histolytica induced apoptosis of human Caco2 and Jurkat cells (Seydel and Stanley, 1998; Huston et al., 2000, 2003; Kim et al., 2014; Lee et al., 2014). Here we asked about the contribution of non-coding miRNAs in apoptosis of SW-480 epithelial colon cells induced by E. histolytica live trophozoites as this issue remains completely understood. Firstly, we setup an interaction model of SW-480 cells with trophozoites and then evaluated the morphological changes induced during the

showing a well-spread cell with irregular microvilli, mitochondria distributed through the cytoplasm and enlarged nuclei. Scale bar 5µM. (B) Image shows transmission electron microscopy ultrastructure images of trophozoites in contact with colon cells after 15, (C) 30 and (D) 45 min of co-culture. (D) SW-480 cells showing typical early changes associated with apoptosis including a general cell-shrinkage, loss of microvilli and nuclear chromatin condensation denoted by arrows (close up in D). Scale bar 10µM. (E) Apoptosis assays. SW-480 cells after interaction with E. histolytica trophozoites for 0, 15, 30, 45, and 75 min were analyzed by flow cytometry using annexin V. SW-480 cells without exposure to trophozoites and cells treated with cisplatin (50µM) for 24 h were used as controls. Data are representative (Continued)

FIGURE 1 | of three independent experiments, performed in triplicate. Bars represent the mean of three independent experiments ± S.D. \*\*p < 0.01. \*\*\*p < 0.001. (F) Western blot assays showing cropped images for immunodetection of caspase-9 and−3. Whole protein extracts were obtained from SW-480 cells after interaction with trophozoites for 0, 15, 30, and 45 min and submitted to Western blot assays using caspase-3 and caspase-9 antibodies. Actin antibodies were used as loading control. Immunodetected bands intensity was quantified by densitometry and normalized against actin. Numbers between panels represent the fold change values relative to control (time 0).

interaction. Data showed that control SW-480 cells incubated without parasites form monolayers of well-spread cells with irregular microvilli (**Figure 1A**). At early times of interaction between host cells and amoeba (15 min), we hardly observed morphological changes indicative of apoptosis (**Figure 1B**). In contrast, after 30 and 45 min interaction with trophozoites, the SW-480 cells exhibited the typical early changes associated with apoptosis including a general cell-shrinkage, loss of microvilli, and nuclear chromatin condensation (**Figures 1C,B,D**). Finally, after 75 min interaction we found extensive SW-480 cell death and monolayers destruction (**Supplementary Figures 1, 2**) representative of late stages of apoptosis and cell killing, thus we did not considered longer time for downstream molecular analysis. These data indicate that 45 min of amoeba-colon cells interaction was able to induce morphological changes associated with apoptosis in agreement with previous reports in other human cells including colonic Caco-2 and T lymphocyte Jurkat cells (Huston et al., 2000, 2003; Kim et al., 2014; Lee et al., 2014).

To confirm whether E. histolytica parasites induce apoptosis in SW-480 cells we performed flow cytometry analysis using annexin V method. Data showed that in the absence of trophozoites a low percentage of SW-480 control cells (3.7 ± 0.4%) were in apoptotic stage. A similar proportion of apoptotic cells (3.9 ± 0.6%) were found at time 0 interaction, which corresponds to initial time when trophozoites and SW-480 cells were co-incubated (**Figure 1E**). The fraction of apoptotic SW-480 cells significantly increased in a time-dependent way from 19.5% at 15 min co-cultures to 25.1, 40, and 78% at 30, 45, and 75 min interaction, respectively (**Figure 1E**). In SW-480 cells exposed to apoptosis-inducer cisplatin (50µM) for 24 h, the proportion of apoptotic cells reached 59%. Next, we were wondering if apoptosis was related to intrinsic pathway activation, therefore we performed Western blot assays for caspase-3 and caspase-9 detection using whole protein extracts from SW-480 cells incubated at different times with virulent trophozoites. Results showed that both caspase-9 and caspase-3 were activated by proteolytic cleavage at 15, 30, and 45 min interaction with amoeba (**Figure 1F**). In contrast, we did not detected significant amounts of cleaved caspases in control cells (time 0). Actin used as control did not show significant changes during course of time. These data confirmed that E. histolytica trophozoites induced apoptosis via the intrinsic pathway in SW-480 colon cells.

MiRNAs are small non-coding RNAs that negatively regulate the expression of target genes involved in cell survival, migration, and apoptosis in diverse eukaryotic organisms (Bartel, 2004). We were asked whether apoptosis of SW-480 cells induced by trophozoites was mediated by changes in expression of miRNAs repertoire. To test this hypothesis, we performed a comprehensive profiling of 667 miRNAs using TaqMan Low Density Arrays (TLDAs) in SW-480 cells interacting or TABLE 1 | MicroRNAs modulated in human SW-480 epithelial colon cells after 45 and 75 min interaction with virulent E. histolytica trophozoites.


not with virulent parasites for 45 and 75 min. These time points were selected as represents the early and late stages of SW-480 cell destruction by trophozoites. Data showed that 6 miRNAs were differentially expressed (FC > 1.5; p < 0.05) after 45 min interaction. Particularly, 5 miRNAs denoted as miR-526b-5p, miR-150, miR-643, miR-615-5p, and miR-525 were significantly up-regulated, whereas miR-409-3p was down-regulated (**Table 1**). At 75 min interaction, we found 16 deregulated miRNAs (3 downregulated and 12 upregulated; **Table 1**). Intriguingly, none common miRNA was modulated at both times.

To validate these findings, we selected the six-miRNAs signature deregulated at 45 min and reexamined its expression by stem-loop qRT-PCR using individual TaqMan probes. In agreement with TLDAs data, the expression of miR-526b-5p, miR-643, miR-150, miR-615-5p, and miR-525 was found increased, while the abundance of miR-409-3p was decreased in SW-480 colon cells after 45 min interaction with trophozoites

(**Figure 2A**). To distinguish the possibility that a soluble factor released by trophozoites could be contributing in the induction of changes in miRNAs abundance, we carried out qRT-PCR analysis of the six-miRNAs signature in the presence of spent medium obtained from parasite's culture media growing without colon cells and compared with the miRNAs levels found after 45 min contact of trophozoites with SW-480 cells. Data showed that the expression of the six miRNAs was very low in the presence of spent culture media in comparison to incubation with trophozoites, indicating no significant regulation of miRNAs (**Figure 2A**). In addition, we also ruled out the possibility that virulent-deficient parasites could be inducing changes in the miRNAs abundance. Results showed no significant expression on the six-miRNAs signature after interaction of SW-480 cells with non-virulent Entamoeba dispar trophozoites (**Figure 2A**). We further confirmed the changes of the six-miRNAs levels using an additional Caco2 epithelial colon cell line. Results showed that miR-150, miR-643, miR-615-5p, and miR-525 exhibited similar levels in SW-480 and Caco2 cells after 45 min incubation with virulent trophozoites (**Figure 2A**). In contrast, miR-409-3p and miR-526b-5p were expressed at very low levels in Caco2 cells in comparison to SW-480 cell line. Then, we analyzed the 6 miRNAs expression in SW-480 and Caco2 cells in the presence of cisplatin used as cell death inductor. Data showed that cisplatin modulated the expression of miR-409-3p, miR-643, miR-615- 5p, and miR-525 in SW-480 and Caco2 cells in a similar way as observed after the 45 min interactions with virulent trophozoites (**Figure 2B**). Moreover, minimal differences in the 4-miRNAs levels in both cell types were observed after cisplatin treatment. In contrast, miR-526b-5p and miR-150 showed an inverse expression profile in cisplatin-treated SW-480 cells in TABLE 2 | Validated targets for miRNAs regulated at 45 min in SW-480 cells after interaction with E. histolytica according to miRTarBase (http://mirtarbase.mbc.nctu.edu. tw/php/index.php).


<sup>a</sup>Gene symbol. <sup>b</sup>Recommended protein name (UniProtKB).

comparison to the observed after SW-480/amoeba interactions without cisplatin.

# Modulated miRNAs Potentially Target Multiple Apoptosis-Related Genes

The identification of genes that could be targeted by miRNAs is a necessary step to understand its functions in host response to parasites. Several of the regulated miRNAs identified here have been previously described in various biological contexts and human diseases, and some gene targets have been functionally validated (**Table 2**). However, most of these studies have been performed in cancer cells from diverse types of human malignancies and until we know, there are no reports related to infectious diseases. To gain insights into the biological roles of cellular miRNAs that were modulated in SW-480 cells, we performed computational predictions to identify their gene targets using DIANA-microT, TargetScan, PicTar, and miRanda software's. We only focused in the bioinformatic analysis of potential gene targets of the set of miRNAs deregulated at 45 min as it represent the time in which apoptosis was clearly observed. Data showed that 120 predicted gene targets can be regulated by at least two miRNAs modulated at 45 min. In addition, DIANA-mirPath and miRWalk analyses revealed that 5 cellular pathways were mainly affected. These include mRNA surveillance pathway (16 genes), biosynthesis of unsaturated fatty acids (5 genes), ubiquitin mediated proteolysis (16 genes), PI3K/AKAT signaling pathway (5 genes), and apoptosis (8 genes). As others, here we also showed that incubation of amoebae with host cells resulted in apoptosis, therefore we decided to focus on this

cellular process. Interestingly, in silico analyses revealed a complex miRNAs-mRNAs coregulation network in which the anti-apoptotic XIAP, API5, BCL2, AKT3, and AKT1 genes were the major targets of deregulated miRNAs at both 45 and 75 min (**Figure 3**; **Table 3**).

# Apoptosis Inhibitor XIAP Is a Target of miR-643

To understand the biological role of cellular miRNAs in apoptosis during E. histolytica interplay, we initiated the functional characterization of miR-643 which was upregulated after 45 min interaction of trophozoites with SW-480 cells as it was predicted to modulate the expression of six genes related with apoptosis (**Table 1**, **Figure 3**). Bioinformatic analysis indicated that Xlinked inhibitor of apoptosis protein XIAP (also known as inhibitor of apoptosis protein 3, IAP3 and baculoviral IAP repeat-containing protein 4, BIRC), a member of the inhibitor of apoptosis family of proteins (IAPs) (Deveraux et al., 1997), it's a predicted target of miR-643. To demonstrate the role of miR-643/XIAP axis in apoptosis, we first evaluated whether interaction of trophozoites with SW-480 cells have a negative effect on expression of XIAP. Western blot results showed that the amount of XIAP protein in SW-480 cells was decreased in a time-dependent way in response to exposition with trophozoites. At 30 and 45 min we found a 10 and 30% reduction in XIAP levels, respectively, in comparison to time 0 (**Figure 4A**). Actin used as control did not show significant changes during course of time.

Examination of the XIAP gene sequence allowed us to identify a miR-643 binding site at the 3′ UTR (1654 to 1660 nucleotide position) suggesting that it may be a direct target (**Figure 4B**). Thus to explore whether miR-643 may negatively regulate XIAP, we inhibited it's expression using a specific antagomiR. SW-480 cells were or not exposed to trophozoites for 45 min and transfected with miR-643 inhibitor (30 and 50 nM) or with scramble sequence (50 nM) as negative control. Results from qRT-PCR assays indicated that in control SW-480 cells grown without trophozoites, the relative expression of miR-643 was low (**Figure 4C**). Interestingly, miR-643 expression was induced six-fold in non-transfected SW-480 cells after interaction with trophozoites in agreement with original TLDAs data (**Table 1**; **Figure 4C**). Moreover, we found that miR-643 amount was dramatically reduced in cells transfected with anti-miR-643 when compared with non-transfected and scramble-transfected SW-480 control cells interacting with E. histolytica, indicating that TABLE 3 | Modulated microRNAs in SW-480 cells interacting with E. histolytica for 45 min and predicted targets associated to apoptosis.


<sup>a</sup>Gene symbol, <sup>b</sup>Recommended protein name (UniProtKB).

both antagomiR doses (30 and 50 nM) were effective to suppress the expression of endogenous miR-643 (**Figure 4C**).

To corroborate whether miR-643 can exerts posttranscriptional repression of XIAP, we performed luciferase reporter gene assays. A DNA fragment corresponding to the 3 ′UTR sequence of XIAP which contains the putative miR-643 binding site was cloned downstream of the luciferase coding region into pGL3 vector (**Figure 4B**). Recombinant pGL3-XIAP plasmid was transfected alone or together miR-643 inhibitor into SW-480 cells and luciferase activity was analyzed after 24 h. Data showed that SW-480 cells transfected with pGL3- XIAP construct alone showed a high luciferase activity which was taken as 100% (**Figure 4D**). Interestingly, transfection of pGL3-XIAP into SW-480 cells co-incubated with trophozoites for 45 min resulted in a substantial reduction of the relative luciferase activity in comparison with SW-480 monoculture used as control (**Figure 4D**). A similar reduction (81%) was observed when cells were co-transfected with scramble control. These data suggested that during interaction of amoeba with colon cells endogenous miR-643 levels were increased allowing the repression of pGL3-XIAP mRNA product and luciferase activity. Remarkably, luciferase enzyme activity was restored when expression of miR-643 was inhibited using two different doses of antagomiR (30 and 50 nM). Importantly, in SW-480 cells transfected with 50 nM, luciferase activity almost reached similar levels that observed in cells that were not exposed with amoebas (**Figure 4D**). This effect was counteracted when cells were co-transfected with anti-miR-643 and an unrelated pGL3-SIAH1 construct (containing the 3′UTR of SIAH1 gene lacking of miR-643 binding sites) used as control and (**Figure 4D**). Furthermore, Western blot assays corroborated that expression of endogenous XIAP protein was restored (fold change 1.4 in comparison to control) in SW-480 cells exposed to trophozoites when miR-643 was inhibited (**Figure 4E**), strengthening the notion that miRNA-643 targets the XIAP gene during the interplay of SW-480:amoeba. Taken altogether, these data suggested that XIAP is targeted by miR-643 during interaction of E. histolytica trophozoites with human colon cells.

# Inhibition of miR-643 Impairs Apoptosis During Interaction of *E. histolytica* With Colon Cells

To obtain further evidence that upregulation of miR-643 is involved in apoptosis activation during the interplay of parasites with SW-480 cells, the expression of miR-643 was inhibited using antagomiRs and the effects in cell death were quantified using annexin-V assays. Again, results showed that in the absence of trophozoites a low percentage of SW-480 control cells were in apoptosis. In addition, data indicated that the fraction of apoptotic cells significantly decreased to 18% after 45 min interaction of miR-643-deficient SW-480 cells with parasites in comparison to cells transfected with scramble negative control and mock which display around 40% apoptosis (**Figure 5A**). In SW-480 cells exposed to cisplatin as control the proportion of apoptotic cells reached 61%. Then, we performed Western blot assays for the detection of caspases 3 and 9 using colon cells transfected with anti-miR-643 and controls. Results showed that activation of caspases 3 and 9 was decreased up to 47 and 49%, respectively, after 45 min interaction of antagomiR-643 transfected-SW-480 cells with amoeba (**Figures 5B,C**) in comparison to controls. Altogether these data suggested that E. histolytica induced apoptosis in SW-480 cells, at least in part, through induction of miR-643, which in turn downregulates XIAP.

# DISCUSSION

It has been largely described that E. histolytica trophozoites produce important changes in genetic programs of host cells during the invasive process. In addition, the impact of host noncoding RNAs in the fine tuning regulation of genes involved in apoptosis is poorly understood. In the current study, we identified for the first time a specific miRNA-signature of SW-480 colon cells that may be involved in the pathophysiological responses to amoeba infection. Our data revealed that exposure

luciferase gene (Luc+) of pGL3 Luciferase Reporter Vector. The miR-643 seed sequence is indicated in the colored box. (C) qRT-PCR assays for expression analysis of miR643. SW-480 cells were incubated with trophozoites for 45 min and transfected with miR-643 inhibitor at 30 nM and 50 nM or with scramble sequence (50 nM). (D) Luciferase reporter gene assays. SW-480 cells transfected with the pGL3-XIAP were incubated with trophozoites for 45 min and co-transfected with anti-miR-643 (30 and 50 nm) or scramble (50 nM). Luciferase activity was measured after 24 h. SW-480 cells transfected with pGL3-XIAP but not exposed to parasites were used as control. (E) Western blot assays showing cropped images for XIAP expression. SW-480 cells transfected with anti-miR-643, scramble or not transfected were incubated with trophozoites for 45 min and then submitted to immunoblotting with XIAP and actin antibodies. Bands intensity was quantified by densitometry and normalized to actin. In all cases, data are representative of three independent experiments by duplicate. Error bars represent S.D. The unpaired Student t-test was used to compare each condition with control cells that were not exposed to E. histolytica. \*p < 0.05; \*\*p < 0.01; \*\*\*p < 0.001.

to virulent E. histolytica trophozoites resulted in a significant deregulation of miRNAs repertoire in epithelial colon cells, as illustrated by the up-regulation of miR-526b-5p, miR-643, miR-150, miR-615-5p, and miR-525, and the down-regulation of miR-409-3p after 45 min interaction. Of these, 4 miRNAs genes are located at the genomic locus 19q13 that is a chromosomal region frequently altered in many human cancers (Kontos, 2014). The small number of modulated miRNAs suggests that amoeba has a limited, but at the same important, impact on the post-transcriptional control of host gene expression mediated by miRNAs, at least at 45 min interaction. Notably, at longer time of cell-cell interactions (75 min) we found an increased number of modulated miRNAs, but none was common at both times. These intriguingly findings showed and suggested that: (i) the apoptosis process could be differentially regulated by miRNAs at 45 min and 75 min. These data is in agreement with the many differences at morphological and molecular level that we evidenced here using multiple approaches. At 45 min

we observed about 40% apoptosis, whereas at 75 min most of colon cells showed apoptosis and presented severe damage and extensive cell death; thus is understood that differences at molecular level (miRNAs profile) also should be exist, (ii) our data also suggested that differential expression of miRNAs is not linear between 45 and 75 min, indicating that at late time of apoptosis a different set of miRNAs is needed to modulate the final stage of cell death, and; (iii) apoptosis is a complex event in which several gene regulation mechanisms are activated/deactivated at different stages of process. The different miRNAs modulated at both times revealed the different stages of apoptosis process and thus it's likely that they change during the time revealing a dynamic regulation of apoptosis by miRNAs. However, additional experimental evidences are needed to demonstrate these assumptions as well to understand the role of the deregulated miRNAs at late stage (75 min) of cell killing.

In related studies, the number of modulated miRNAs in response to parasite infection is variable, depending on the pathogen species and the use of an in vivo or in vitro model. For example, Eimeria papillata, another gastrointestinal protozoan, modulates only 4 miRNAs species in the mouse jejunum (Dkhil et al., 2011), while Plasmodium, Cryptococus, and Leishmania modulate a set of 19, 28, and 64 microRNAs in hepatocytes, cholangiocytes and fibroblasts, respectively (Chen et al., 2007; Delic et al., 2011; Lemaire et al., 2013 ´ ). Thus our datasets size is in the range of regulated miRNAs observed in different pathogen-host cells studies. Importantly, one should keep on mind that a single miRNA can potentially regulate as many as 1,000 different genes (Krek et al., 2005). Indeed, the 6-modulated miRNAs at 45 min were individually predicted to regulate from 365 (miR-615) to 1057 (miR-525). Altogether, they can potentially affect the expression of genes participating in at least five important biochemical pathways in SW-480 cells, including mRNA surveillance pathways, biosynthesis of unsaturated fatty acids, ubiquitin mediated proteolysis, PI3K-AKT signaling pathway and apoptosis. These data indicate that trophozoites interaction represents a stress condition that induces a physiological and adaptive response in host cells. The modulation of proteins involved in mRNA surveillance and ubiquitin mediated proteolysis points out that host cells modify their gene expression after E. histolytica interaction. Importantly, bioactive lipid molecules promote apoptosis by modulating mitochondrial membrane permeability and activating different enzymes including caspases (Vanhaesebroeck et al., 2010; Huang and Freter, 2015). Amoeba may be inducing apoptosis in host cells by altering, in part, miRNAs regulating genes involved in lipid metabolism, PI3K-AKT signaling pathway and apoptosis. Notably, the six modulated microRNAs potentially target genes related to apoptosis. Thus, the expression of BCL2L11, XIAP, AIFM3, BCL2L1, and BMF can potentially be regulated by at least two of these miRNAs. SW-480 cells were induced to apoptosis and caspase-3 and-9 were processed following interaction with E. histolytica trophozoites, suggesting the activation of the apoptotic intrinsic pathway. Several of the miRNAs identified here have been previously described in cancer cells with a role in the regulation of cell proliferation and apoptosis. In addition, a few targets related to apoptosis have been validated. For instance, miR-409 promotes epithelial-to-mesenchymal transition and

prostate tumorigenesis (Josson et al., 2015) while it suppresses tumor cell invasion and metastasis by directly targeting radixin in gastric cancer (Zheng et al., 2012). miR-150 promotes cell proliferation in gastric cancer and cancer epithelial cells (Zhou et al., 2008; Wu et al., 2010), while it functions as a tumor suppressor in human colorectal cancer by targeting c-Myb (Feng et al., 2014). miR-526b-5p induces cell cycle arrest and apoptosis in non-small cell lung cancer (Zhang et al., 2015). miR-615 functions as a tumor suppressor in pancreas and breast cancer (Bai et al., 2015; Sun et al., 2015).

Here we demonstrated that miR-643 target the anti-apoptotic XIAP gene in SW-480 cells. Previously, it was reported that miR-643 functions as a tumor suppressor gene in osteosarcoma cells. It was showed that miR-643 suppressed the cancer hallmarks and as a bona fide tumor suppressor it may also activates apoptosis (Wang et al., 2017). XIAP has a central regulatory role in programmed cell death by inhibiting the caspases cascade (Deveraux and Reed, 1999). In agreement, apoptosis of SW-480 cells was associated with XIAP repression mediated by upregulation of miR-643 in response to E. histolytica exposure. It is important to note that miR-643 and the other modulated miRNAs can also potentially regulate the expression of other target genes related to apoptosis, contributing to SW-480 cell death. Based on our results, we proposed a working model to describe the participation of cellular miRNAs in apoptosis driven by amoebas' interaction (**Figure 6**). In this model, after 45 min contact of SW-480 cells with parasites the miR-643 was upregulated. Then miR-643 target the XIAP 3′UTR inducing the downregulation of XIAP protein levels, which release the inhibition of caspase-3 and -9. In consequence the executioner caspase-3 triggers degradation of cellular components resulting in the apoptosis of SW-480 cells. Notably, miR-526-5p, miR-525, miR-150, and miR-615-5p could also contribute to XIAP inhibition, as they were upregulated during the SW480:amoeba interplay, and were also predicted to target XIAP, however additional experimental evidences are needed to support this assumption.

Limitations of the actual study include: (i) as in previous reports on apoptosis induced by E. histolytica in colonic Caco-2 and T lymphocyte Jurkat cells (Huston et al., 2000, 2003; Kim et al., 2014; Lee et al., 2014), the use of a particular cell line limits the extrapolation of findings to in vivo conditions; (ii) the miRNAs signature reported here are specific for SW-480 cell line; (iii) the parasite's processes that may trigger miRNAs changes during apoptosis of SW-480 cells remains unknown; (iv) the contribution of additional miRNAs in apoptosis and immune response evasion, and the mechanism by which amoeba contact regulates the transcription of selected host miRNAs remains to be elucidated. Our results showed that the knowledge of the impact of E. histolytica on host miRNAs may provide new insights into the relationships between this pathogen and human cells. This first miRNA profiling in SW-480 human colon cells exposed to amoeba revealed significant alterations in cellular miRNAs expression. The set of deregulated miRNAs represent a guide for further functional studies in apoptosis and immune response events. Thus, our findings represent novel data that contribute to our understanding of the cellular and molecular mechanisms activated by the host cells during E. histolytica contact.

# AUTHOR CONTRIBUTIONS

IL-R, YS-V, OH, CP-F, and BC-M conducted all the experiments. BC-M performed the transmission electron microscopy of cells. MÁ-S, ER-M, and LM performed experiments with amoeba. OR-A performed the in silico analysis. CL-C, NG, CP-P and LM contributed to experimental design, intellectual input, and interpreting data. CL-C and LM wrote the manuscript.

# ACKNOWLEDGMENTS

We thank Jacqueline Soto-Sanchez for helping with trophozoites cultures. We also acknowledge to Universidad Autónoma de la Ciudad de México and Instituto Politécnico Nacional for support.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00437/full#supplementary-material

# REFERENCES


B virus infection caused hepatitis and liver cirrhosis. Sci. Rep. 12:15026. doi: 10.1038/srep15026


pro-apoptotic purinergic P2X7 receptor by activation of instability sites at the 3′ -untranslated region of the gene that decrease steady-state levels of the transcript. J. Biol. Chem. 283, 28274–28286. doi: 10.1074/jbc.M802663200

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2019 López-Rosas, López-Camarillo, Salinas-Vera, Hernández-de la Cruz, Palma-Flores, Chávez-Munguía, Resendis-Antonio, Guillen, Pérez-Plasencia, Álvarez-Sánchez, Ramírez-Moreno and Marchat. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# *Entamoeba histolytica* Calreticulin Induces the Expression of Cytokines in Peripheral Blood Mononuclear Cells Isolated From Patients With Amebic Liver Abscess

#### *Edited by:*

Serge Ankri, Technion–Israel Institute of Technology, Israel

#### *Reviewed by:*

Jaishree Paul, Jawaharlal Nehru University, India Stefan Kappe, Center for Infectious Disease Research, United States

#### *\*Correspondence:*

Cecilia Ximénez cximenez@unam.mx

#### *Specialty section:*

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

*Received:* 23 March 2018 *Accepted:* 19 September 2018 *Published:* 18 October 2018

#### *Citation:*

Gonzalez Rivas E, Ximenez C, Nieves-Ramirez ME, Moran Silva P, Partida-Rodríguez O, Hernandez EH, Rojas Velázquez L, Serrano Vázquez A and Magaña Nuñez U (2018) Entamoeba histolytica Calreticulin Induces the Expression of Cytokines in Peripheral Blood Mononuclear Cells Isolated From Patients With Amebic Liver Abscess. Front. Cell. Infect. Microbiol. 8:358. doi: 10.3389/fcimb.2018.00358 Enrique Gonzalez Rivas, Cecilia Ximenez\*, Miriam Enriqueta Nieves-Ramirez, Patricia Moran Silva, Oswaldo Partida-Rodríguez, Eric Hernandez Hernandez, Liliana Rojas Velázquez, Angelica Serrano Vázquez and Ulises Magaña Nuñez

Laboratorio de Inmunología, Unidad de Investigación de Medicina Experimental, Facultad de Medicina, Universidad Nacional Autónoma de México (UNAM), Ciudad de Mexico, Mexico

Calreticulin (CRT) is a highly conserved protein in the endoplasmic reticulum that plays important roles in the regulation of key cellular functions. Little is known about the participation of E. histolytica CRT (EhCRT) in the processes of pathogenicity or in the modulation of the host immune response. The aim of this study was to evaluate the role of CRT in the proliferation and the cytokine profile in peripheral blood mononuclear cells (PBMCs) from patients with amebic liver abscess (ALA) during the acute phase (AP-ALA) of the disease compared to patients during the resolution phase (R-ALA). The PBMCs from each participant were cocultured with EhCRT and tested by the colorimetric method to evaluate their proliferation index (PI). The supernatants were subjected to an enzyme-linked immunosorbent assay (ELISA) to evaluate the concentration of cytokines. The mean values of all groups were compared using the independent t-test. When the PIs of individuals without diagnosis of liver abscess (NEG) were compared, there were no statistically significant differences in the proliferation of PBMCs between patients with AP-ALA and R-ALA when stimulated with EhCRT or concanavalin A (ConA). However, the levels of interleukins [IL-6, IL-10, granulocyte colony stimulating factor (GCSF), and transforming growth factor β1 (TGFβ1)] were higher in patients with AP-ALA, whereas in patients with R-ALA, higher levels of interferon gamma (IFNγ) were detected. These results suggest that EhCRT acts as a mitogen very similar to the activity of ConA. In addition, EhCRT is an excellent immunogen for the specific activation of PBMCs, inducing the differential expression of ILs depending on the outcome of disease, determining the type of immune response: a Th2 cytokine profile during the acute phase and a Th1 profile during the resolution phase.

Keywords: amebic liver abscess, calreticulin, *Entamoeba histolytica*, mitogen, proliferation index, interleukins

# INTRODUCTION

Infection with the enteric protozoan E. histolytica is one of the leading causes of death worldwide. The disease is a consequence of the parasite's abilities to invade the colon, causing amebic colitis. E. histolytica can disseminate to the liver via the portal venous system, resulting in amebic liver abscess (ALA). However, approximately 90% of infected people are asymptomatic cyst carriers (Haque et al., 2003). The molecular mechanisms by which this parasite causes invasive amebiasis are not fully understood. E. histolytica has adherence and cytotoxicity factors that are essential for its survival, but they are not directly responsible for ALA formation. It is known that the limitation and prevention of recurrent invasive amebiasis requires the development of an effective immune response. Thus, it is likely that the acute inflammatory response associated with E. histolytica infection is a key factor for the development of ALA (Chadee et al., 1985).

Parasite-specific immune responses are regulated by cytokines and chemokines that lead to the development of immunity, but these responses also contribute to infection, inducing pathogenesis and parasite persistence (Talvani et al., 2004). Little is known regarding the amebic signals that initiate an acute inflammatory response.

It has been reported that in mice infected with E. histolytica, host tissue damage is attributed primarily to the lectin activity of the galactose/N-acetyl-D-galactosamine (Gal/GalNAc) from E. histolytica, which promotes the accumulation of mononuclear cells, including neutrophils, inflammatory monocytes, and macrophages, at the site of infection (Blazquez et al., 2007). Intestinal epithelial cells infected with E. histolytica in vitro produce elevated levels of the cytokines, interleukin-8 (IL-8), growth-regulated protein alpha (GRO-α), granulocyte macrophage colony-stimulating factor (GMCSF), and IL-1 (Eckmann et al., 1995). Treatment of cultured human intestinal cells with the lectin Gal/GalNAc from pathogenic and nonpathogenic entamebas (E. histolytica and E. dispar) results in the secretion of chemoattractant and proinflammatory cytokines (Sharma et al., 2005), suggesting that these cells and cytokines also contribute to tissue damage, participating in the mechanisms of initiation, amplification, or limitation of the inflammatory processes during invasive amebiasis.

The identification of the mediators involved in leukocyte activation during infection by E. histolytica is of fundamental importance for understanding host responses in amebiasis. Cellular interactions and cytokines have been reported during amebic infections, and cytokines have been shown to be able to regulate monocyte function and increase the amebicidal activity of monocytes (Seydel et al., 2000; Lotter et al., 2013).

It is still not clear what other elements in the dynamics of host–parasite relationship define the outcome of infection, especially regarding the regulation of the immune response against E. histolytica.

To obtain further evidence about the relationships of immune cells with E. histolytica, other proteins have been studied to investigate the intracellular signals that promote the host immune response. Some examples include the cysteine proteinases 1 and 5 (CP1 and CP5) that breakdown IgA1 and IgA2 antibodies. These proteins cleave the Fc region that interacts with parasite surface receptors and mediates effector functions that can mask immunogenic surface molecules with inert Fab fragments, thus helping to prevent the parasite expulsion from the intestinal lumen (Garcia-Nieto et al., 2008).

The role of calreticulin (CRT) in host–parasite interactions has recently become a major area of research. The CRT genes from many parasites (Trypanosoma, Leishmania, Entamoeba, Onchocerca, Schistosoma, and Haemonchus) have been cloned and sequenced (Rokeach et al., 1994; Joshi et al., 1996; El-Gengehi et al., 2000; Marcelain et al., 2000; González et al., 2002; Suchitra and Joshi, 2005).

Although the functions of CRT are conserved in vertebrates, some CRT functions differ among parasites (Nakashi et al., 1998; Ferreira et al., 2004); parasite CRTs bind host C1q and inhibit C1q-dependent complement activation. The CRT of Haemonchus contortus binds host C-reactive protein and C1q (Naresha et al., 2009). The ecto-parasite Amblyomma americanum secretes CRT during feeding, suggesting that the anticoagulant ability of CRT may prevent blood clotting and allows the parasite to feed on the host and induce host antiparasite responses (Jaworski et al., 1995). The presence of CRT in the penetration gland cells of Schistosoma suggests that this molecule may be important for the host skin penetration (Khalife et al., 1994).

Previously, we reported the presence of CRT in E. histolytica (EhCRT) and that this protein induces an important immunogenic response in the human host. More than 90% of patients with ALA develop high levels of serum antibodies against EhCRT (González et al., 2002). We also reported the cloning of the CRT gene from E. histolytica, and the preparation of mono-specific antibodies against recombinant CRT. The immunohistochemical assays on trophozoites show that EhCRT is in cytoplasmic vesicles and in vesicles that are in close contact with the inner cytoplasmic membrane (González et al., 2011). In addition, it was demonstrated that the CRT from both pathogenic E. histolytica and nonpathogenic E. dispar species specifically interact with human C1q molecules and inhibit the activation of the classical complement pathway (Ximénez et al., 2014). This activity is consistent with that reported by Vaithilingam et al. (2012). The trophozoites activated by the presence of jurkat cells clearly show the binding of C1q to CRT on the surface of the phagocytic mouths during the process of erythrophagocytosis.

However, the activation of the host immune response and the cytokine profile induced by EhCRT have not yet been investigated.

In this study, we analyzed the proliferation and cytokine production of peripheral blood mononuclear cells (PBMCs) cultures isolated from ALA patients during the acute phase of the disease and from individuals who resolved ALA in order to characterize the cytokines profiles produced in response to EhCRT.

# METHODS

# Ethics Statement

The present work was designed according to the guidelines for the management of human samples for experimental purposes as indicated in the Official Regulation NOM-12SSA3-2007 included in the General Health Law of Mexican Health Ministry. In addition, the project was approved by the Scientific and Ethics Committee of the Faculty of Medicine from the National Autonomous University of Mexico. Patients were informed about the purposes of the project, the sampling, and the potential risks during procedure, and the patients were invited to voluntarily participate by signing an informed consent letter.

# Study Groups and Biological Samples

Patients with diagnosis of ALA admitted to the Internal Medicine, Gastroenterology, and Infectiology Services of the General Hospital of Mexico Dr. Eduaro Liciaga were recruited. These patients formed the group with acute phase amebic liver abscess (AP-ALA). The patients in the resolution phase of ALA (R-ALA) were recruited from a search of the archives of discharged patients from the General Hospital of Mexico, and they formed the ALA resolution group (R-ALA). Three fecal samples were collected from each patient from both groups at 1-week intervals after their hospitalization for microscopic examination for parasites. At the time of collection of the third sample, 10 ml of blood was drawn to evaluate serum antibodies against E. histolytica by enzyme-linked immunosorbent assay (ELISA), using an OD490nm ≥0.5 as an indicator of a positive result (Morán et al., 2007). The remaining sample was used to isolate PBMCs.

A third group, named the negative control (NEG), was formed with clinically healthy individuals from the Blood Bank of the General Hospital of Mexico, with ELISA serum levels of antiamebic antibodies below the OD threshold (OD<0.5). Using the same protocol for ALA groups, three fecal samples and 10 ml blood samples were taken for microscopic examination and PBMC isolation, respectively. Ten individuals for the group were included.

# Isolation and Culture of PBMC

The PBMCs were isolated from 10 ml peripheral blood samples collected from each participant in tubes with K2 ethylenediaminetetraacetic acid (EDTA) as anticoagulant (BD Vacutainer, Ref 368171). Cells were separated on a Ficoll-Hypaque gradient (Gibco, Life Technologies, Grand Island, NY, USA), and the PBMC pellet was separated and washed three times with phosphate-buffered saline PBS. After separation, PBMCs were centrifuged and resuspended in Roswell Park Memorial Institute medium RPMI culture medium supplemented with 10% fetal bovine serum. The cells (1 × 10<sup>6</sup> cells/ml) were incubated with rEhCRT (5µg/ml) at 37◦C with 5% CO2; concanavalin A (ConA) (Sigma Chemical St. Louis, MO USA) was used as a stimulating factor at 10µg/ml for different periods of time (24, 48, 72, and 96 h). Each experiment was performed in duplicate. At each time point, the cultures were centrifuged for 10 min at 1,000 × g, and the supernatant was reserved for cytokine analysis.

# Recombinant *Eh*CRT Production

Full-length rEhCRT protein was expressed and purified as previously described (González et al., 2011). Briefly, the plasmid pBluescript-KS+ (pbKS+) was used to clone and express the 1,178 bp Ehcrt gene (GB-EAL649855.1) (Loftus et al., 2005), which produces the full-length protein; it was subcloned into the prokaryotic expression vector pProEX HT-b (Gibco Life Technologies) to express the CRT protein with a six-histidine tag at the N-terminal end. Competent Escherichia coli BL21 cells were transformed with the recombinant plasmid. The expression of recombinant protein rEhCRT was induced with a final concentration of 1 mM isopropyl-β-D-thiogalactoside. The QIAexpressionist system (Qiagen, Valencia, CA, USA) was used to purify the recombinant protein, and, briefly, the cells were harvested by centrifugation at 3,000 × g for 12 min, and the bacterial pellet was resuspended in 5 ml lysis buffer (8 M urea, 0.1 M NaH2PO4, and 0.1 M Tris-HCl, pH 8.0). The lysate was added to a 50% suspension of Ni-NTA agarose (Qiagen,). The mixture was passed through a filtration column, and the recombinant protein was eluted with 8 M-urea buffer pH 4.5. The selected fractions were dialyzed against 19 mM PBS. Protein concentration was determined by the Bradford method, and the quality was evaluated by 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE). The reactivity grade of the recombinant protein was tested against sera from patients with ALA and for anti-E. coli lipopolysaccharide antibody (LPS) antibody (ab211144, Abcam) by Western blot.

# Proliferation Index (PI)

To obtain the PI at 24, 48, 72, and 96 h, 20 µl of 5µg/ml [3-(4,5 dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] (MTT) (Sigma) was added to each well containing PBMCs incubated with rEhCRT or ConA and incubated at 37◦C for 1 h in a humidified incubator with 5% CO2. After incubation, the plates were centrifuged at 1,000 × g for 10 min and the supernatant was discarded. The cells were resuspended in 300 µl dimethyl sulfoxide (Sigma) and the optical density (OD) was measured at 570 nm in an ELISA plate reader (EL × 800 BioTek). The proliferation index was calculated by [OD of test sample—OD of negative control/OD of negative control]. (Verma et al., 2010)

# Cytokine Detection by ELISA

The supernatants of the PBMC cultures, treated with rEhCRT, ConA, or without stimulus (RPMI-10% SFB), were tested for detection of interleukins IL-2, IL-4, IL-5, IL-6, IL-10, IL-12, IL-13, IL-17A, interferon gamma (IFNγ), tumor necrosis factor alpha (TNFα), granulocyte colony stimulating factor (GCSF), and transforming growth factor beta1 (TGFβ1) using a multianalytic ELISA array kit (MEH-003A, Qiagen), according to manufacturer's instructions. The concentrations of the cytokines are given in pg/ml and were calculated using the standard curve provided in the kit.

# Statistical Analysis

All values are expressed as the means ± S.D. The student's t test for unpaired results was used for the evaluation of differences

TABLE 1 | Demographic characteristics of individuals in the AP-ALA, R-ALA, and healthy control (NEG) groups.


The antibodies against E. histolytica antigens present in all belonging to different studied groups were evaluated by ELISA. The data are expressed as the average OD ± SD in each group.

between cytokine concentrations in each group. Differences were statistically significant when P ≤ 0.05.

# RESULTS

# Study Participants and *E. histolytica* Antibody Concentration

The demographic features of each group are shown in **Table 1**. Among all the individuals, the mean age was 39 ± 7 years old, and 70% were male and 30% were female. The microscopic examinations were negative in all samples. Results of ELISA assay are shown as OD 490nm values, considering negative results when values were under the cut value of 0.520. The mean value for the negative group was 0.15, the AP-ALA group was 0.87, and the R-ALA group was 1.0.

# Purification of r*Eh*CRT and PI

The purified rEhCRT from E. coli appeared as a single band at 60 kDa on 12% SDS-PAGE after bromophenol blue staining (**Figure 1A**). The results indicate that rEhCRT was a good immunogenic factor, and this was previously confirmed by the antibody production in animal models (rabbits and mice) (González et al., 2011). The rEhCRT conserved its reactivity when it was tested as an antigen using serum from ALA patients in ELISA assays (**Table 1**) and in Western blots, and no reactivity was observed with anti-E. coli LPS antibody (**Figure 1B**).

The role of rEhCRT as a costimulatory factor in the proliferation of PBMC was verified when cellular proliferation was measured by 3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide (MTT) assays, comparing the stimulatory capacity of rEhCRT vs. the mitogen activity of ConA. The differences in the PI values were not statistically significant for any time point when comparing between the NEG and ALA groups (AP-ALA, R-ALA) (**Figure 2**). However, at 72 h, we can see the highest PI (**Figure 3**) and the PI ≥ 1 is considered as positive proliferation (Verma et al., 2010).

# Detection of Cytokines in the Supernatants of PBMC Cultures

The concentrations of cytokines in PBMC samples obtained for sera from the ALA and NEG groups and treated with EhCRT were measured through ELISA only at 72 h (**Figure 4**). The

(9–10) reactivity against E. coli -LPS antibody.

pattern of pro-inflammatory cytokines from the AP-ALA and R-ALA groups were compared to the NEG group, who did not produce the mentioned cytokines in the absenceof EhCRT.

The trend of the data agrees with an overexpression of all interleukins analyzed in all groups.

However, in the ALA groups, the overexpression was, in general, larger than in the negative group, and significant differences were measured in the expression of the following interleukins; IL-2, IL-5, IL-6, IL-10, IL-17, IFNγ, TNFα, GCSF, and TGFβ1 (p = 0.0035, 0.0023, 0.019, 0.006, 0.0076, 0.0049, 0.0039, 0.0035, 0.0059, respectively).

The comparisons of differential expressions of the interleukins between groups of AP-ALA and R-ALA are shown in **Figure 4**. We found differences in the overexpression of interleukins IL-2, IL-5, IL-6, IL-10, GCSF, and TGFβ1 (p = 0.0067, 0.026, 0.0045, 0.0016, 0.0051, 0.0047, respectively), which were higher in the AP-ALA group. In the R-ALA group, we observed higher overexpression of the interleukins IL-17, INFγ, and TNFα; however, statistically significant differences were detected only for INFγ (p = 0.73, 0.0014, 0.096).

FIGURE 4 | Cytokine profile on PBMC cells stimulated with rEhCRT. The data shows the concentration of cytokines, obtained in the supernatant of PBMC cells of different studied groups (NEG, AP-ALA, and R-ALA) after 72 h of incubation and stimulated with rEhCRT or without stimulus (RPMI). The concentration of interleukins is expressed as pg/ml in a logarithmic scale. \*p ≤ 0.005, when compared groups of individuals (NEG) against the group of patients (ALA), \*\*when compared AP-ALA against R-ALA groups.

# DISCUSSION

The aim of this study was to examine the proliferation of PBMCs obtained from the blood of patients with AP-ALA and R-ALA stimulated in vitro with rEhCRT or ConA and determine the cytokine profiles induced in the different groups. The responses of the PBMC show that EhCRT is one of the many immunogenic proteins in E. histolytica that can induce activation and proliferation of PBMCs similarly to ConA. These results reinforce our previous observations that the EhCRT is highly immunogenic in humans, mice, and rabbits (González et al., 2002, 2011).

When comparing the cytokine profiles between the AP-ALA and R-ALA groups in contrast with negative or PBMC without stimulus, it is clear that EhCRT functions as a specific antigenic costimulator, inducing a different pattern of cytokines between different groups. This stimulatory action was specific because no reactivity with E. coli -LPS antibody was detected for the recombinant protein EhCRT (**Figure 1A**).

The expression levels of IL-2, IL-5, IL-6, IL-10, GCSF, and TGFβ1 were increased in patients with AP-ALA, while the expression levels of IL-17, INFγ, and TNFα were mainly upregulated in the R-ALA group. Nevertheless, there were statistically significant differences only for INFγ.

E. histolytica has different proteins that modulate the host immune response. The Gal/GalNAc-lectin induces the T cells to production of IL-2 and INFγ (Schain et al., 1992), whereas in macrophages, the amebicide activity of Gal/GalNAc-lectin induces the production of TNFα (Seguin et al., 1997). In dendritic cells, Gal/GalNAc-lectin favors a Th1 response in addition to inducing the production of major histocompatibility complex (MHC) class II molecules, and the costimulatory molecules CD80, CD86, and CD40 (Ivory and Chadde, 2007).

Another protein called monocyte locomotion inhibitory factor (MLIF) is produced by E. histolytica in axenic cultures that induces the production of pro-inflammatory cytokines (IL-1β, IL-2, IL-5, IL-6, IFN-γ) and anti-inflammatory cytokines such as IL-10, as well as the low expression of chemokines CCL1, CCL4, and the receptor CCR1 in human monocytes (Rico et al., 2003; Utrera-Barillas et al., 2003).

In the group of patients with AP-ALA, the cytokines IL2, IL5, Il6, IL17A, IFNγ, and TNFα displayed an increase in their concentration and demonstrated that the immune response had a pro-inflammatory profile. This response has been observed in stimulation assays using Gal/GalNAc-lectin in intestinal cell cultures (Sharma et al., 2005) and in other parasitic diseases such as malaria (Vazquez et al., 2015). It is important to highlight the effect on IL2, IL6, IL-17, IFNγ, and TNFα, whose levels were the highest and overexpressed in comparison with the profile observed in the negative group.

Guo et al. (2011), demonstrated in an Entamoeba histolytica vaccination model that IL-17 provides protection to mice vaccinated with the recombinant LecA fragment of the Gal/GalNAc-lectin. Interestingly, the major source of IL-17 in these mice was the CD8 T cells, whereas CD4 T cells express elevated levels of IFN-γ. The authors suggest that IL-17 may enhance the protective functions of Th1 immune response.

These results lead us to propose that treatment of PBMC in culture with EhCRT favors the production of IFNγ and increases the production of IL-17A, thus directing the cellular immune response to a Th1 Profile, in PMBCs obtained from R-ALA.

Our results agree with those published by Ghadirian and Denis (1992), who showed that IFNγ could activate mouse peritoneal macrophages, which, in turn, were able to eradicate the E. histolytica trophozoites from colon tissue in vitro. Studies in animal models (Seydel et al., 2000) and human infections (Haque et al., 2007) have established that amoeba-specific IFN-γ production is critically involved in the clearance of infection and in host protection. In addition, Meza et al. (2014) demonstrated that virgin T-cell differentiation into Th17 cells producing IL-17 occurred after the direct stimulation with other cytokines such as TGFβ, IL-6, and IL-1 in a murine model of infection with Trypanosoma cruzi. Moraes et al. (2015) reported that mononuclear cells collected from healthy individuals incubated with E. histolytica in culture induced the production of IFNγ and TGFβ, and that both had a beneficial effect on the modulation of the activity of these cells. Our results agree with data of Moraes regarding the effects on IFNγ. These cytokines are important in the control E. histolytica infection.

The R-ALA group showed an increase in the concentration of the cytokines IL-10 and TGFβ, which agrees with results of Bansal et al. (2005). These authors mention that, in addition to the production of these cytokines and the increased production of IL-4, a suppressive immune response was also induced in patients infected with E. histolytica, which, in turn, favored a symptomatic infection. The symptomatic group in Bansal et al. differs with our R-ALA group because the latter had no symptoms and they were all ALA with negative microscopic examinations and a resolved E. histolytica infection. In our opinion, this immunosuppressive effect is due to the direct stimulation of the PBMC by EhCRT, and through autocrine, paracrine, and pleiotropic effects on cytokine production, which favored the increase in other types of cytokines such as IL-5, IL-6, and IL-13 capable of generating a Th2 immune response.

In addition, we found that the increased cytokine IL-10 in the AP-ALA group in our study was in contrast with the results reported by Bansal. These results suggest that IL-10 is an immunomodulator resulting in proinflammatory cytokine profiles that could promote immunosuppression in the R-ALA individuals. The effect attributed to IL-10 differs in other parasitic diseases such as Leishmania donovani (Andreani et al., 2015), Trypanosoma cruzi (Longhi et al., 2014), and Giardia duodenalis (Babaei et al., 2016). In these reports, a decrease in IL-10 was observed that favored the spread of parasites into the hepatic tissue.

On the contrary, the IL-10 level was increased in dysenteric and ALA patients (Utrera-Barillas et al., 2003; Bansal et al., 2005). These studies indicate that invasion of the colon and liver by E. histolytica elicits an anti-inflammatory immune response and may successfully suppress immune reaction to the amebae.

In summary, the ameba needs to balance IL-10 and the proinflammatory cytokine to allow establishment of infection. In contrast, peritoneal monocytes and macrophages exposed to lipopeptidophosphoglycan (LPPG) secreted TNF-α, IL-6, IL-8, IL-12, and IL-10 via TLR2 (Maldonaldo-Bernal et al., 2005). Thus, LPPG-driven signaling may activate a negative feedback loop that attenuates inflammatory responses.

Host protective immunity involves participation of both humoral and cellular responses; however, the mechanism involved in the immune evasion of E. histolytica is not clear. One of these mechanisms could be associated with the ability of parasites to modulate cytokine expression in the inflammatory process, which is initiated by expression of proinflammatory cytokines. E. histolytica infections induce a state of transient suppression of cell-mediated immunity in early stages of inflammation in amebic hepatic abscess, and a complex signaling system of cytokines is triggered by pathogen invasion (Eckmann et al., 1993; Romagnani, 2000).

# CONCLUSIONS

The data obtained in this study confirmed that the EhCRT behaves like an amebic immunogenic protein for humans and suggest that the EhCRT participates in the specific stimulation of immune cells.

Our results suggest that the rEhCRT can stimulate human PBMC proliferation independently of the presence of E. histolytica trophozoites, acting as a specific costimulator of the immune response like that induced by ConA. In addition, these results underline EhCRT as a parasitic factor that can modulate the immune response, from the stimulation of proinflammatory cytokines to the immunosuppressive effects, depending on the progression of the ALA, thus inducing the development of a Th2 cytokine profile in the acute phase of disease and a Th1 profile once the individuals had resolved the ALA.

The functions of EhCRT and its role in the pathogenesis of ALA need further research, particularly on the interaction with the cells of immune system and the induction of chemokines and cytokines regulators that hopefully will allow a better understanding of the pathogenesis of ALA.

# AUTHOR CONTRIBUTIONS

CX conceptualized the manuscript. EG, MN-R, UM. PM, OP-R, EH, LR, and AS curated the data. CX, EG, MN-R, PM, OP-R, EH, LR, and AS performed the formal analysis. CX acquired the funding. EG, MN-R, and PM performed the investigation. MN-R, PM, AS, and UM provided the resources. EG, OP-R, and EH performed the software analysis. CX, EG, MN-R, and PM provided the supervision. CX, EG, MN-R, and PM executed the validation. CX, EG, MN-R, and PM wrote the original draft of the manuscript. CX, EG, MN-R, PM, OP-R, EH, UM, and AS wrote, reviewed and edited the manuscript.

# REFERENCES


# FUNDING

We are grateful for the support provided by the following institutions: Grants: IN218214 from PAPIIT (DGAPA), Universidad Nacional Autónoma de México (UNAM), and 272601 from the Consejo Nacional de Ciencia y Tecniología (CONACyT).

# ACKNOWLEDGMENTS

We thank Mario Nequiz-Avendaño for the axenic culture of E. histolytica (HM1:IMSS), and Marco Gudiño Zayas, for art design. LR-V, a doctoral student from the Programa de Doctorado en Ciencias Biomédicas, Universidad Nacional Autónoma de México (UNAM), is grateful for having received fellowship 239901 from CONACYT. We also appreciate the technical assistance of Martha Elena Zaragoza and Ma. de los Angeles Padilla, and the secretarial assistance of Mrs. Ma. Elena Ortiz.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Gonzalez Rivas, Ximenez, Nieves-Ramirez, Moran Silva, Partida-Rodríguez, Hernandez, Rojas Velázquez, Serrano Vázquez and Magaña Nuñez. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Entamoeba histolytica Induce Signaling via Raf/MEK/ERK for Neutrophil Extracellular Trap (NET) Formation

Zayda Fonseca<sup>1</sup> , César Díaz-Godínez <sup>1</sup> , Nancy Mora<sup>1</sup> , Omar R. Alemán<sup>1</sup> , Eileen Uribe-Querol <sup>2</sup> , Julio C. Carrero<sup>1</sup> \* and Carlos Rosales <sup>1</sup> \*

<sup>1</sup> Departamento de Inmunología, Instituto de Investigaciones Biomédicas, Universidad Nacional Autónoma de México, Mexico City, Mexico, <sup>2</sup> División de Estudios de Posgrado e Investigación, Facultad de Odontología, Universidad Nacional Autónoma de México, Mexico City, Mexico

### Edited by:

Serge Ankri, Technion–Israel Institute of Technology, Israel

#### Reviewed by:

Sinuhe Hahn, Universität Basel, Switzerland Deborah Schechtman, Universidade de São Paulo, Brazil Arturo Ortega, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional (CINVESTAV-IPN), Mexico

> \*Correspondence: Julio C. Carrero carrero@unam.mx Carlos Rosales carosal@unam.mx

Received: 25 March 2018 Accepted: 15 June 2018 Published: 04 July 2018

#### Citation:

Fonseca Z, Díaz-Godínez C, Mora N, Alemán OR, Uribe-Querol E, Carrero JC and Rosales C (2018) Entamoeba histolytica Induce Signaling via Raf/MEK/ERK for Neutrophil Extracellular Trap (NET) Formation. Front. Cell. Infect. Microbiol. 8:226. doi: 10.3389/fcimb.2018.00226 Amoebiasis, the disease caused by Entamoeba histolytica is the third leading cause of human deaths among parasite infections. E. histolytica was reported associated with around 100 million cases of amoebic dysentery, colitis and amoebic liver abscess that lead to almost 50,000 fatalities worldwide in 2010. E. histolytica infection is associated with the induction of inflammation characterized by a large number of infiltrating neutrophils. These neutrophils have been implicated in defense against this parasite, by mechanisms not completely described. The neutrophil antimicrobial mechanisms include phagocytosis, degranulation, and formation of neutrophil extracellular traps (NETs). Recently, our group reported that NETs are also produced in response to E. histolytica trophozoites. But, the mechanism for NETs induction remains unknown. In this report we explored the possibility that E. histolytica leads to NETs formation via a signaling pathway similar to the pathways activated by PMA or the Fc receptor FcγRIIIb. Neutrophils were stimulated by E. histolytica trophozoites and the effect of various pharmacological inhibitors on amoeba-induced NETs formation was assessed. Selective inhibitors of Raf, MEK, and NF-κB prevented E. histolytica-induced NET formation. In contrast, inhibitors of PKC, TAK1, and NADPH-oxidase did not block E. histolytica-induced NETs formation. E. histolytica induced phosphorylation of ERK in a Raf and MEK dependent manner. These data show that E. histolytica activates a signaling pathway to induce NETs formation, that involves Raf/MEK/ERK, but it is independent of PKC, TAK1, and reactive oxygen species (ROS). Thus, amoebas activate neutrophils via a different pathway from the pathways activated by PMA or the IgG receptor FcγRIIIb.

### Keywords: Entamoeba histolytica, neutrophil, NETosis, NETs, ROS, ERK, NF-κB

# INTRODUCTION

Entamoeba histolytica is a protozoan parasite with high prevalence in developing countries (Verkerke and Petri, 2012; Tellevik et al., 2015; Ghenghesh et al., 2016). Amoebiasis, the disease caused by E. histolytica affects the intestine and the liver, and is the third leading cause of human deaths among parasite infections (Walsh, 1986; Lozano et al., 2012). In this context, E. histolytica was found responsible for about 100 million cases of amoebiasis that led to some 50,000 global deaths in 2010 (Mortimer and Chadee, 2010). Although there is growing understanding of the immune response against amoebas, a full solution to amoebiasis is still needed (Moonah et al., 2013; Nakada-Tsukui and Nozaki, 2016; Cornick and Chadee, 2017). E. histolytica infection of the intestine or liver is associated with a strong inflammation characterized by a large number of infiltrating neutrophils (Prathap and Gilman, 1970; Tsutsumi et al., 1984; Tsutsumi and Martinez-Palomo, 1988; Espinosa-Cantellano and Martínez-Palomo, 2000). Usually, large numbers of neutrophil are seen surrounding trophozoites. Yet, amoebas do not seem to be damaged by this interaction. Neutrophils have been implicated in defense against this parasite playing a crucial protective role (Seydel et al., 1997; Velazquez et al., 1998; Jarillo-Luna et al., 2002; Asgharpour et al., 2005; Estrada-Figueroa et al., 2011). However, neutrophils and other leukocytes have also been reported as major inducers of tissue damage during intestinal and liver amoebiasis (Salata and Ravdin, 1986; Pérez-Tamayo et al., 1991, 2006; Seydel et al., 1998; Olivos-García et al., 2007; Dickson-Gonzalez et al., 2009). Therefore, the role of neutrophils in this parasitic infection remains controversial.

Neutrophils, the most abundant leucocytes in peripheral blood, migrate from the circulation to sites of inflammation. Typically, neutrophils are considered the first line of defense because they are the first cells to arrive at the infected site, and they present several antimicrobial functions (Deniset and Kubes, 2014; Mayadas et al., 2014). Among these functions, phagocytosis, degranulation, and formation of neutrophil extracellular traps (NETs) are the most important (Brinkmann et al., 2004; Yipp et al., 2012). NETs are formed by a process known as "NETosis" that involves activation in most cases of nicotinamide adenine dinucleotide phosphate (NADPH)-oxidase, which produces reactive oxygen species (ROS) (Fuchs et al., 2007; Bianchi et al., 2009; Remijsen et al., 2011). NETs are fibers of DNA decorated with histones (Neeli and Radic, 2012) and antimicrobial proteins, such as elastase, myeloperoxidase, lactoferrin, and metalloprotease 9 (Brinkmann et al., 2004; Fuchs et al., 2007). NETs can block the dissemination of microorganisms because they function as a physical barrier where pathogens get caught, and get also exposed to antimicrobial proteins. Consequently, NETs can eliminate pathogens extracellularly and independently of phagocytosis (Papayannopoulos and Zychlinsky, 2009). Several human protozoan parasites have been reported to induce the formation of NETs, including Leishmania amazonensis, L. major, L. chagasi, Leishmania donovani (Guimarães-Costa et al., 2009; Gabriel et al., 2010; Hurrell et al., 2015), Toxoplasma gondii (Abi Abdallah et al., 2012), and Trypanosoma cruzi (Sousa-Rocha et al., 2015). Recently, E. histolytica trophozoites were also demonstrated to induce NETs formation (Ávila et al., 2016; Ventura-Juarez et al., 2016). Yet, the mechanism of NETs induction by any of these parasites remains unknown. Although, many microorganisms can induce NETs, no single receptor for pathogen-associated molecular patterns (PAMPs) has been identified as responsible for inducing this neutrophil response. However, Toll-like receptors (TLRs) have been suggested to participate (Yipp et al., 2012). Only two receptors for antibody molecules are reported to be bona fide activators of NETs release from human neutrophils, the IgA receptor FcαR (Aleyd et al., 2014), and the IgG receptor FcγRIIIb (Behnen et al., 2014; Alemán et al., 2016a).

It was firstly published that signaling activated by phorbol 12 myristate 13-acetate (PMA) in neutrophils for NETs formation involves the Raf/ERK pathway (Hakkim et al., 2011) and requires ROS produced by the NADPH-oxidase (Almyroudis et al., 2013). In contrast, we previously found that signaling activated by the FcγRIIIb for NETs formation is different from the pathway activated by PMA (Alemán et al., 2016a,b). For this receptor, NETs formation is dependent on NADPH-oxidase, and extracellular signal-regulated kinase (ERK) activation (Alemán et al., 2016a), and requires signaling through the kinases spleen tyrosine kinase (Syk) and transforming growth factor-β-activated kinase 1 (TAK1) (Alemán et al., 2016b). These results emphasize the recent recognized fact that NETs formation is induced by different signaling pathways depending on diverse stimuli (Kenny et al., 2017). In the case of parasitic pathogens, such as E. histolytica, no receptor has been clearly identified as an inducer of NETs, and nothing is known about the signaling pathway activated by amoebas in neutrophils to induce NETs formation. Therefore, in this report we investigated whether E. histolytica leads to NETs formation via a signaling pathway that involves ERK activation. Neutrophils were stimulated by E. histolytica trophozoites and the effect of various pharmacological inhibitors on amoeba-induced NETs formation was assessed. Selective inhibitors of Raf, MEK, and NF-κB prevented E. histolyticainduced NET formation. In contrast, inhibitors of PKC and NADPH-oxidase, as previously reported, blocked PMA-induced (Hakkim et al., 2011), but not E. histolytica-induced NET formation. E. histolytica induced phosphorylation of ERK in a Raf and MEK dependent manner. Also, NF-κB phosphorylation was dependent on MEK. Our results indicate for the first time that E. histolytica triggers a signaling pathway to induce NETs formation, that involves Raf/MEK/ERK, but it is independent of PKC, ROS, Syk, and TAK1. Thus, amoebas activate neutrophils via a different pathway from the pathways activated by PMA or the IgG receptor FcγRIIIb.

# MATERIALS AND METHODS

# Neutrophils

Neutrophils (PMN) were purified from blood exactly as previously described (García-García et al., 2013). Adult healthy volunteers provided a written informed consent before donating blood. The Bioethics Committee at Instituto de Investigaciones Biomédicas—Universidad Nacional Autónoma de México (UNAM), approved the informed consent form, and all experimental procedures.

# Entamoeba histolytica

Entamoeba histolytica trophozoites (strain HM1:IMSS) were cultured axenically at 37◦C in TYIS-33 medium supplemented with 15% heat-inactivated adult bovine serum and Diamond vitamin Tween <sup>R</sup> 80 solution (Sigma-Aldrich; St. Louis, MO) (Diamond et al., 1978). The cultures were incubated for 72 h and trophozoites were collected after cooling on ice for 5 min and then centrifuging at 300 × g for 5 min at 4◦C. The pelleted trophozoites (amoebas) were resuspended in PBS.

# Reagents

Bovine serum albumin (BSA) was from F. Hoffmann-La Roche Ltd. (Mannheim, Germany). UO126, a specific MEK (ERK kinase) inhibitor was obtained from Promega (Madison, WI, USA). The antibiotic LLZ 1640-2 (also known as (5Z)-7- Oxozeaenol; cas 66018-38-0) (catalog no. sc-202055), a specific TAK1 inhibitor, was from Santa Cruz Biotechnology (Santa Cruz, CA). Wortmannin, a phosphatidylinositol 3-kinase (PI3K) inhibitor; Gö6976, a protein kinase C (PKC) inhibitor; Gö6983, another PKC inhibitor; SB 203580, a p38 MAP kinase inhibitor (catalog number 559389); 4′ ,6-diamino-2-fenilindol (DAPI), a cell-permeable DNA-binding dye (catalog no. 268298); and 3-(1-methyl-1H-indol-3-yl-methylene)-2-oxo-2,3-dihydro-

1H-indole-5-sulfonamide (iSyk), a Syk inhibitor (catalog no. 574711) were from Calbiochem/EMD Millipore (Billerica, MA). The cOmpleteTM protease inhibitor cocktail (catalog no. 11697498001) and PhosSTOPTM phosphatase inhibitor cocktail (catalog no. 04906845001) were from Roche Diagnostics (Basel, Switzerland). Dihydrorhodamine 123 (catalog no. AS-85711) was from AnaSpec, Inc. (Fremont, CA, USA), and Dihydroethidium (catalog no. 12013) was from Cayman Chemical (Ann Arbor, MI, USA). Diphenyleneiodonium chloride (DPI), an NADPH-oxidase inhibitor (catalog no. D2926); (E)-3-[4-methylphenylsulfonyl]-2-propenenitrile (BAY 117082), an NF-κB inhibitor (catalog no. B5556); 3-(3,5 dibromo-4-hydroxybenzyliden)-5-iodo-1,3-dihydroindol-2-one (GW5074), a cRaf1 kinase inhibitor (catalog no. G6416); phorbol 12-myristate 13-acetate (PMA) (catalog no. P8139), and all other chemicals were from Sigma-Aldrich (St. Louis, MO). The following antibodies were used: rabbit polyclonal anti-histone H4 (acetyl K12) antibody (catalog No. ab61238), and rabbit polyclonal anti-citrulline antibody (catalog No. ab100932) from Abcam, Inc. (Cambridge, MA). Monoclonal antibody IgG2a (IB4) anti-integrin β2 (catalog no. sc-65254), mouse monoclonal IgG2a anti-phospho-ERK 1/2 (pTyr204) (catalog no. sc-7383), mouse monoclonal IgG1 anti-NF-κB p50 subunit (catalog no. sc-8414), and rabbit polyclonal anti phospho-NF-κB p50 subunit (pSer337) (catalog no. sc-33022) from Santa Cruz Biotechnology (Santa Cruz, CA). Monoclonal antibody IgG1 (3G8) anti-human CD16 (catalog no. 556617), and R-phycoerythrin (PE)-conjugated monoclonal antibody IgG2a (clone CLB-gran1 1.5) anti-human CD16b (catalog No. 550868) were from BD PharmingenTM (San Diego, CA). Tetramethylrhodamine (TRITC) conjugate ZyMaxTM goat anti-rabbit IgG (catalog No. 81-6114) was from Thermo Fisher Scientific (Carlsbad, CA). Rabbit monoclonal IgG anti-ERK 1/2 (catalog no. 4695) was from Cell Signaling Technology, Inc. (Beverly, MA). HRP-conjugated F(ab')<sup>2</sup> goat anti-mouse IgG (catalog No. 0855572), and HRP-conjugated F(ab')<sup>2</sup> goat antirabbit IgG (catalog No. 0855686) were from MP Biomedicals (Santa Ana, CA).

# NET Formation Assay

Neutrophils (2.5 × 10<sup>5</sup> ) in 500 <sup>µ</sup>l RPMI-1640 medium (Gibco <sup>R</sup> ; Grand Island, NY) were added to each well of a 24-well plate (Costar <sup>R</sup> 3524; Corning Inc., Corning, NY), and incubated in a humidified incubator with 5% CO<sup>2</sup> at 37◦C for 30 min. Then 100 µl of 120 nM PMA in PBS, or 100 µl an E. histolytica suspension (1.25 × 10<sup>5</sup> cell/ml) were added to each well. The amoeba to neutrophil ratio was 1:20, as determined previously (Ávila et al., 2016). Plates were incubated in 5% CO<sup>2</sup> at 37◦<sup>C</sup> for 4 h. Next, 600 µl of 2% paraformaldehyde in PBS were gently added to each well, and the plates were incubated overnight in 5% CO<sup>2</sup> at 37◦C. The fixative was removed by very gentle aspiration at the side of the well, and then the cells were stained with 150 nM DAPI in PBS for 30 min at room temperature. Finally, the plates were observed with a fluorescence inverted microscope model IX-70 from Olympus (Center Valley, PA). Images were captured with an Evolution-VF Cooled Color camera from Media Cybernetics (Rockville, MD), and the computer program Q Capture pro 6.0 from QIMAGING Surrey (British Columbia, Canada). Images were processed with the computer program ImageJ 1.47v from The National Institutes of Health (Bethesda, MD).

In selected experiments, PMN were incubated on ice for 30 min before stimulation, with the inhibitors: Gö6983 (1µM), Gö6976 (1µM), GW5074 (100µM), UO126 (75µM), DPI (10µM), BAY 117082 (5µM), iSyk (1µM), antibody IB4 (10µg/ml), Wortmannin (50 nM), LLZ 1640-2 (10 nM), SB 203580 (200 nM), or the vehicle dimethyl sulfoxide (DMSO) alone.

# Immunofluorescence

For NETs staining, neutrophils (2 × 10<sup>5</sup> ) were incubated with E. histolytica trophozoites (1 × 10<sup>4</sup> ) in 200 µl RPMI-1640 medium (Gibco <sup>R</sup> ; Grand Island, NY) using Lab-TekTM chamber slides from Thermo Fisher Scientific (Rockford, IL). After 1 h, cultures were fixed with 4% formaldehyde for 10 min, then fixed cells were permeabilized with 0.2% Triton X-100 in PBS for 5 min, and washed three times with PBS. Cells were next blocked with 1% BSA, 0.3 M glycine, 0.1% Tween 20 in PBS for 30 min at 37◦C, and then incubated with anti-histone H4 or anti-citrulline antibodies diluted 1/100 in 1% BSA, 0.1% Tween 20 in PBS during 1 h at room temperature. Next, cells were gently washed with cold PBS and incubated in the dark with TRITC–conjugated goat anti-rabbit IgG antibody diluted 1/50 in the same solution for 1 h at room temperature. Cells were finally washed with PBS and stained with 150 nM DAPI. Coverslips were mounted with Fluoroshield before observation in a fluorescence Olympus BX51 microscope.

# Live Cell Imaging Quantification of NETs

NETs formation was quantitated by live imaging and an offline analysis similarly to previous reports (Hoffmann et al., 2016; van der Linden et al., 2017). Neutrophils (2.5 × 10<sup>5</sup> ) were incubated in 250 µl RPMI-1640 medium containing 500 nM SYTOX <sup>R</sup> Green in each well of a 48-well tissue culture plate (Costar <sup>R</sup> 3548; Corning Inc., Corning, NY). The plate was incubated for 20 min at 37◦C in the dark, and then E. histolytica trophozoites (1.25 × 10<sup>4</sup> ) were added in 50 µl to each well. The amoeba to neutrophil ratio was 1:20. NET release was monitored with an Olympus fluorescence inverted microscope during a period of 240 min. Images were captured every 5 min with an Evolution-VF Cooled Color camera from Media Cybernetics (Rockville, MD), and the computer program Q Capture pro 6.0 from QIMAGING Surrey (British Columbia, Canada). Image files were saved in tiff format and converted to 8-bit grayscale. Images were processed with the computer program Fiji (version 2.0.0-rc-65/1.52b) (Schindelin et al., 2012). This method for NET quantification is based on previously described protocols (Brinkmann et al., 2013; Hoffmann et al., 2016; van der Linden et al., 2017). Briefly, scale was set and threshold adjusted to define the area of fluorescent DNA. Then the area of NETs was measured using the tool "analyze particles." Total area of all fluorescent particles indicated the amount of NETs formation. Unstimulated neutrophils had an area of 62 ± 6.1 µm<sup>2</sup> . Thus only particles larger than 70 µm<sup>2</sup> were considered NETs.

# Spectrophotometric Quantification of NETs

NET formation was also quantified by detecting DNA release spectrophotometrically with the DNA-binding dye SYTOX <sup>R</sup> Green as previously described (Behnen et al., 2014; Gonzalez et al., 2014; Alemán et al., 2016a,b). Briefly, neutrophils were resuspended at 1 × 10<sup>6</sup> cell/ml in RPMI-1640 medium (Gibco <sup>R</sup> ; Grand Island, NY), containing 500 nM SYTOX <sup>R</sup> Green (Molecular Probes, Inc.; Eugene, OR). Then, 100 µl of this cell suspension (1 × 10<sup>5</sup> PMN) were added to each well of a 96-well plate (Costar <sup>R</sup> 3590; Corning Inc., Corning, NY). Next, the plate was incubated at 37◦C in a 5% CO<sup>2</sup> incubator for 20 min. Neutrophils were then stimulated by adding 20 µl of 120 nM PMA (20 nM final concentration), or 20 µl of an E. histolytica suspension (2.5 × 10<sup>5</sup> amoeba/ml) to each corresponding well. The amoeba to neutrophil ratio was 1:20. The plate was then incubated in a 35◦C pre-warmed microplate reader, model Synergy HT from BioTek Instruments (Winooski, VT), for up to 4 h. For this assay, cells were not fixed. The fluorescence from the bottom of the plate was read every 5 min, using the 485 nm excitation and 528 emission filters.

# Fluorescent Calcium Measurements

Neutrophils at 1 × 10<sup>7</sup> cell/ml in PBS with 1.5 mM Ca2<sup>+</sup> and 1.5 mM Mg2+, were loaded with Fura-2/AM (Calbiochem; San Diego, CA) and cytosolic calcium concentration calculated as previously described (Rosales and Brown, 1991, 1992; García-García et al., 2002). Briefly, 3 × 10<sup>6</sup> neutrophils in 1 ml PBS were transferred to a cuvette and then 1.5 × 10<sup>5</sup> amoebas were added in 80 µl PBS. Fluorescence changes were monitored with a Perkin Elmer (Waltham, MA) LS55 spectrofluorimeter, and calcium concentration calculated with the Perkin Elmer FL WinLab software, version 4.00.02.

# FACS

Fluoresce labeling of neutrophil surface receptors for flow cytometry analysis was completed exactly as described (García-García et al., 2007).

# Neutrophil Stimulation With Trophozoites and Protein Extraction

Neutrophils (1 × 10<sup>6</sup> ) in 500 µl PBS were placed in a 1.5 ml Eppendorf tube. Next, E. histolytica trophozoites (5 × 10<sup>4</sup> ) in 100 µl PBS were added. Cells were gently mixed and immediately incubated at 37◦C in a water bath for various periods of time as indicated. At the end of the corresponding time, 0.8 ml of cold PBS were added and cells centrifuged at 6,000 rpm in a microcentrifuge for 2 min. The supernatant was removed by aspiration and the cell pellets were then lysed in cold RIPA buffer (150 mM NaCl, 5 mM EDTA, 50 mM Hepes, 0.5% sodium deoxycholate, 1% Non-idet P-40, 2 mM Na3VO4, pH = 7.5), supplemented with 1X cOmpleteTM protease inhibitor cocktail and 1X PhosSTOPTM phosphatase inhibitor cocktail, which were added just before lysing the cells. Cell lysates were incubated on ice for 20 min, and then cleared by centrifugation. Cell lysates were immediately used for Western blotting.

# Western Blotting

Western blots were performed as previously described (Reyes-Reyes et al., 2001). Briefly, proteins in cell lysates were resolved on SDS 10% PAGE, and then electrotransfered onto polyvinylidine fluoride (PVDF) membranes (Immobilon-P; Millipore, Bedford, MA). Membranes were incubated in blocking buffer [1% BSA, 5% non-fat dry milk (Carnation; Nestle, Glendale, CA), and 0.1% Tween <sup>R</sup> 20 in Tris-buffered saline (TBS: 50 mM Tris-HCl, 150 mM NaCl, pH = 7.5)] for 2 h at room temperature. Membranes were subsequently probed with the corresponding antibody in blocking buffer, overnight at 4◦C, anti-phospho-ERK 1 (1/2,500 dilution), or anti-phospho NF-κB (1/2,500 dilution). Membranes were washed with TBS-Tween (TBS containing 0.1% Tween <sup>R</sup> 20) six times and incubated with a 1/3,000 dilution of the corresponding HRP-conjugated F(ab′ )<sup>2</sup> goat anti-mouse IgG or goat anti-rabbit IgG for 1 h at room temperature. After washing six more times with TBS-Tween, the membrane was developed with Immobilon Western chemiluminescent HRP substrate (catalog no. WBKLS0100) from EMD Millipore (Billerica, MA) according to the manufacturer's instructions. Afterwards, membranes were stripped with 0.2 M NaOH, reprobed with anti-ERK antibody (1/4,000 dilution), or anti-NFκB antibody (1/2,500 dilution), to assess protein loading in PAGE gels.

# Reactive Oxygen Species (ROS)

ROS production was assessed with three independent methods (Wojtala et al., 2014). First, the cell-permeative reagent 2′ ,7′ dichlorofluorescin diacetate (DCFDA) was used in a fluorescence spectroscopy assay. Second, Dihydrorhodamine 123 was used in a flow cytometry (FACS) assay. Third, Dihydroethidium was used in a fluorescence microscopy assay.

For ROS detection with DCFDA, the DCFDA-cellular ROS detection assay kit (catalog no. ab113851) from Abcam, Inc. (Cambridge, MA) was used following the manufacturer's instructions. Briefly, neutrophils were washed with 1X buffer and then incubated with 15µM DCFDA in 1X buffer for 30 min at 37◦C in the dark. After one wash in 1X buffer, neutrophils were resuspended at 1 × 10<sup>6</sup> cell/ml in 2X buffer. Fifty microliters of this neutrophil suspension (5 × 10<sup>4</sup> PMN) were added into each well of a 96-well clear-bottom black plate (Costar <sup>R</sup> 3614; Corning Inc., Corning, NY) and incubated for 20 min at 36◦C in a plate-reader, model Synergy HT from BioTek Instruments (Winooski, VT). Then, 50 µl of 40 nM PMA (20 nM final concentration), or 50 µl of an E. histolytica suspension (2,500 amoebas) were added. Fluorescence was read every 2 min for 2 h at excitation 485 nm and emission 535 nm.

For ROS detection with dihydrorhodamine 123, neutrophils (1 × 10<sup>6</sup> ) were resuspended in 100 µl of 15µM dihydrorhodamine 123 in PBS inside a 1.5 ml Eppendorf tube, and incubated for 30 min at 37◦C in the dark. Neutrophils were then centrifuged at 4,000 rpm for 1 min in a microcentrifuge model 5415C (Eppendorf; Mississauga, Ontario, Canada), and after removing the supernatant, they were washed in 500 µl PBS. Finally, neutrophils were resuspended in 500 µl PBS. In the same tube covered with aluminum foil, 100 µl of 120 nM PMA in PBS (final concentration 20 nM), or 100 µl of an E. histolytica suspension (5 × 10<sup>5</sup> amoebas/ml) were added for PMN stimulation. Cells were incubated for 60 min at 37◦C in a 5% CO<sup>2</sup> incubator, and then fixed by adding 600 µl of 2% paraformaldehyde. Finally, cells were analyzed by flow cytometry using a FACScalibur apparatus (Becton Dickinson; Franklin Lakes, NJ), with the 485 nm (excitation) and 520 nm (emission) filters. PMN were gated by dot-plot analysis and 10,000 cells were acquired per sample. Data analysis was performed using the Cellquest software (Becton Dickinson; Franklin Lakes, NJ).

For ROS detection with dihydroethidium, neutrophils (1 × 10<sup>6</sup> ) were resuspended in 100 µl of 15µM dihydroethidium in PBS, and were incubated for 30 min at 37◦C in the dark. Neutrophils were then centrifuged at 4,000 rpm (1,375 × g) for 1 min in a microcentrifuge and washed in 500 µl PBS. Next, neutrophils were resuspended in 500 µl PBS, and 250 µl of this cell suspension (5 × 10<sup>5</sup> PMN) were added to each well of a 48-well tissue culture plate (Costar <sup>R</sup> 3548; Corning Inc., Corning, NY). The plate was incubated for 20 min at 37◦C in the dark, and then 50 µl of 120 nM de PMA in PBS (20 nM final concentration), or 50 µl of a E. histolytica suspension (5 × 10<sup>5</sup> amoebas/ml) were added to each well. The plate was incubated for 60 min at 37◦C in a 5% CO<sup>2</sup> incubator. Next, 300 µl of 2% paraformaldehyde were added to each well for fixing the cells. After 30 min, the plates were observed with a fluorescence inverted microscope model IX-70 from Olympus (Center Valley, PA).

# Statistical Analysis

Quantitative data were expressed as mean ± standard error of mean (SEM). Single variable data were compared by paired-sample Student's t-tests using the computer program KaleidaGraph <sup>R</sup> version 3.6.2 for Mac (Synergy Software; Reading, PA). Differences were considered statistically different at a value p < 0.05.

# RESULTS

# Entamoeba histolytica Induce NET Formation

NETs formation has been mostly studied by using PMA, a potent activator of PKC as an inducer of NETosis (Brinkmann et al., 2004). The antibody receptor FcγRIIIb also induces a strong activation of NETs formation (Alemán et al., 2016a,b). In addition, we recently reported that E. histolytica also induce NETs formation (Ávila et al., 2016), but there are not reports on the mechanism of NETs induction by these parasites. When human neutrophils were stimulated by PMA, NETs are detected after 2.5 h of stimulation (Fuchs et al., 2007). Complete NETs were seen, as previously reported, by 4 h after stimulation (**Figure 1A**). Stimulation with E. histolytica trophozoites also induced NETosis (**Figure 1**). The extracellular DNA fibers co-localized with neutrophil elastase and myeloperoxidase (Díaz-Godínez et al., 2018), as well as with histone H4 and citrulline (**Figure 1B**), confirming that these fibers were bona fide NETs. Live cell imaging showed that by 30 min after incubation with amoebas, NETs were already visible (**Figure 2A**). By 2 h about half of the total amount of NETs had already been formed, and by 2.5 h NETs reached a maximum level (**Figure 2A**). NETs formation was induced only in neutrophils that were in direct contact with E. histolytica trophozoites (**Figure 2B**). The NETs were produced around the amoebas and progressively covered the parasites (**Figure 2B**). Neutrophils that were not in contact with amoebas did not release DNA fibers and never became SYTOX <sup>R</sup> Green-positive (**Figure 2B**), suggesting that the signaling for NETosis comes from a receptor that directly recognizes the parasite. Quantification of NETs from live cell imaging analysis showed that amoeba-induced NETs were formed with a much faster kinetics than PMA-induced NETs. After stimulation with amoebas, NETs could be easily detected by 30 min (**Figure 3A**). The amount of NETs progressively increased during the following 2 h, attaining a maximum level by 2.5 h, that was even higher than the one induced by PMA (**Figure 3A**). Since NETs were exclusively formed by neutrophils joining amoebas and the fluorescence staining of external DNA only reflected NETs, the formation of NETs was also quantitated spectrophotometrically as previously reported (Alemán et al., 2016a,b). In accordance with live cell imaging analysis, neutrophils treated with PMA showed NETs (extracellular DNA fibers) after 2 h of stimulation reaching a maximum level by 4 h (**Figure 3B**). Stimulation of neutrophils with E. histolytica trophozoites also induced NETosis that could be detected by 30 min and reached a maximum level around 2.5 h (**Figure 3B**). Neutrophils alone only presented background fluorescence that did not increase during the time of the experiment, thus confirming the imaging data showing that neutrophils did not lose membrane integrity. Therefore, both quantification methods are equivalent and allowed us to reach similar conclusions. The difference in kinetics for NETosis suggested that the signaling induced by E. histolytica was different from the one induced by PMA. Hence, we next explored whether the signaling molecules reported to be required for NETs formation after PMA or FcγRIIIb stimulation were also required for E. histolytica-induced NETosis.

# Entamoeba histolytica-Induced NETs Formation Is Dependent on Raf and MEK, but Not PKC

Because PMA is an activator of PKC, the involvement of this kinase in NET formation induced by E. histolytica was tested with two specific PKC inhibitors. PMN treated with PMA formed NETs as expected (**Figure 4**). However, when PMN were treated previously with Gö6983, an inhibitor of PKCα, PKCβ, and PKCγ isozymes (**Figure 4**), or with Gö6976, a conventional PKC inhibitor (**Figure 4**), NETs were not formed after PMA stimulation. In contrast, NETs formation after E. histolytica stimulation was not inhibited by these PKC inhibitors (**Figure 4**, Supplementary Figure 1). In addition, downstream of PKC, the Raf, MEK, ERK pathway has been reported to participate in NETs formation after PMA stimulation (Hakkim et al., 2011). When neutrophils were treated with GW5074, a specific Raf inhibitor, NETs were not formed after PMA stimulation (**Figure 4**), or after E. histolytica stimulation

(**Figure 4**, Supplementary Figure 2). In addition, when PMN were treated with UO126, a potent specific MEK inhibitor, NETs were not formed after PMA stimulation (**Figure 4**), or after E. histolytica stimulation (**Figure 4**, Supplementary Figure 3). These data suggested that E. histolytica stimulation led to NETs formation using Raf and MEK, but not through PKC activation.

# Extracellular Calcium Is Required for Entamoeba histolytica-Induced NETs Formation

The involvement of MEK in amoeba-induced NETs formation suggested that also ERK would be involved. Since calcium plays a key role in ERK activation, and calcium mobilization is important for NETosis induced by other stimuli (Gupta et al., 2014), we explored the role of intracellular or extracellular calcium pools in E. histolytica induced NETosis. Neutrophils were placed in PBS with the calcium chelator EGTA, and then stimulated with the N-formylated chemotactic peptide formylmethionyl-leucyl-phenylalanine (fMLF). As previously reported (Rosales and Brown, 1992), neutrophils showed an increase in cytosolic calcium concentration that comes from intracellular stores and is dependent on inositol 1,4,5-trisphosphate (IP3) (**Figure 5A**). Then, after adding an excess of calcium outside the cells a second peak of cytosolic calcium was observed, indicating that an extracellular calcium flux is also activated by fMLF (Rosales and Brown, 1991, 1992) (**Figure 5A**). In contrast, when neutrophils were stimulated with E. histolytica trophozoites, no increase in cytosolic calcium concentration was detected. After, addition of an excess of calcium outside the cells, a robust calcium mobilization was observed (**Figure 5B**). This indicates that amoebas induce a calcium flux in neutrophils that comes

FIGURE 4 | Entamoeba histolytica-induced NETs formation is dependent on Raf and MEK, but not PKC. Human neutrophils were left untreated (—), stimulated with 20 nM phorbol 12-myristate 13-acetate (PMA), or with E. histolytica trophozoites. PMN were previously treated with solvent alone (—) or (A) with the PKC inhibitors Gö6976 (1µM), or Gö6983 (1µM); or (B) with the Raf inhibitor GW5074 (100µM); or (C) with the MEK inhibitor UO126 (75µM). The relative amount of NETs was estimated from SYTOX® Green fluorescence in relative fluorescent units (RFU) at 4 h after stimulation. Data are mean ± SEM of 3 experiments. Asterisks denote conditions that are statistically different from control (p < 0.001).

SEM of three experiments done in triplicates.

Fonseca et al. E. hitolytica Induce NETosis via Raf/MEK/ERK

requires extracellular calcium, but this calcium does not activate the calcineurin pathway.

# Entamoeba histolytica Trophozoites Induced NETosis but Not Apoptosis

Live cell imaging revealed that only neutrophils in contact with amoebas released their DNA (**Figure 2**). In order to confirm that neutrophils were undergoing NETosis and no other forms of cell death in the presence of amoebas, the integrity of chromatin was analyzed. DNA from neutrophils treated with PMA, or with the calcium ionophore A23187, or exposed to amoebas did not show any fragmentation in agarose gels (Díaz-Godínez et al., 2018). In contrast, neutrophils exposed to 56◦C for 1 h, a well-known inducer of apoptosis, showed fragmented DNA (Díaz-Godínez et al., 2018). Moreover, heat-treated neutrophils, and some PMAtreated neutrophils, showed an increase in surface expression of phosphatidylserine; while neutrophils exposed to A23187 or to amoebas did not show phosphatidylserine surface expression (Díaz-Godínez et al., 2018). In addition, it has been reported that neutrophils undergoing apoptosis lose the expression of the antibody receptor FcγRIIIb (CD16b) (Sim et al., 2005). However, neutrophils exposed to E. histolytica trophozoites, did not show any difference in FcγRIIIb expression, as indicated by binding of two different monoclonal antibodies specific for this receptor (**Figure 6**). Together these results confirm that amoebas induce NETs formation and not apoptosis when they are in contact with neutrophils.

# Entamoeba histolytica-Induced NETs Formation Is Dependent on ERK

To confirm that ERK was activated after E. histolytica stimulation, neutrophils with or without the MEK inhibitor were incubated with amoebas and then ERK activation was detected by Western blotting. E. histolytica induced a rapid ERK phosphorylation, which reached a maximum at about 2 min after neutrophils and amoebas got in contact (**Figure 7A**). This phosphorylation then decreased with time and was barely detectable after 15 min. The antibodies used to detect ERK and phospho-ERK did not recognize any proteins from amoeba cell lysates (**Figure 7A**), thus ERK phosphorylation detected was only from neutrophils. E. histolytica-induced ERK phosphorylation was prevented by the MEK inhibitor UO126 (**Figure 7B**). Also, the Raf inhibitor GW5074 completely impeded E. histolyticainduced ERK phosphorylation (**Figure 7C**). These data suggested that E. histolytica induced, in neutrophils, the Raf, MEK, ERK signaling pathway to activate NETosis.

# The NADPH-Oxidase Inhibitor DPI Reduced Entamoeba histolytica-Induced NETs Formation

NETs formed after PMA stimulation require activation of NADPH-oxidase and formation of ROS (Patel et al., 2010; Almyroudis et al., 2013; Björnsdottir et al., 2015). Thus, we explored the involvement of these molecules in E. histolytica-induced NETs formation. Neutrophils treated with diphenyleneiodonium (DPI), a NADPH-oxidase inhibitor,

FIGURE 5 | Entamoeba histolytica trigger extracellular calcium mobilization in neutrophils. Human neutrophils were incubated with fura-2 for 30 min and then placed in PBS with EGTA. Neutrophils were stimulated (arrow) with (A) 10 nM formyl-methionyl-leucyl-phenylalanine (fMLF), or were stimulated with (B) E. histolytica trophozoites (amoeba). After about 250 s, an excess (4 mM) Ca2<sup>+</sup> was added to the buffer. Changes in cytosolic calcium concentration were assessed by measuring the variations in fluorescence of fura-2-loaded cells. Tracings are representative of three experiments. (C) The increase of cytosolic calcium concentration (<sup>1</sup> [Ca2+]) is shown from intracellular or extracellular calcium pools. Data are mean ± SEM of three experiments.

were not able to form NETs after PMA stimulation (**Figure 8**). Similarly, DPI-treated neutrophils did not efficiently form NETs after E. histolytica stimulation (**Figure 8**). However, because inhibition by DPI of E. histolytica-induced NETs formation was only to about half (**Figure 8**), we decided to use apocynin, a different inhibitor of NADPH-oxidase (Kim et al., 2012) and also a ROS scavenger (Heumüller et al., 2008). As expected, apocynin inhibited PMA-induced NETosis, indicating that this mechanism depends on ROS production (Díaz-Godínez et al., 2018). In contrast, apocynin did not decrease amoeba-induced NETs formation, suggesting that this type of NETosis is independent of ROS production by NADPH oxidases (Díaz-Godínez et al., 2018). In order to clarify the different effect of these two NADPH-oxidase inhibitors, we decided to directly assess ROS production after E. histolytica stimulation of neutrophils. When neutrophils were treated with PMA or with tert-butyl hydrogen peroxide (TBHP), a positive control for ROS production, ROS were generated in great amounts and could be easily detected with the 2′ ,7′ -dichlorofluorescin diacetate (DCFDA) method (**Figure 9A**). To our surprise, however, E. histolytica stimulation of neutrophils did not induce any ROS production (**Figure 9A**). Consequently, the effect of DPI did not seem to be related to inhibition of ROS production. Cell viability was then tested in the presence of DPI. Neutrophils remained viable in the presence of DPI for more than 2 h (**Figure 9B**). In contrast, E. histolytica trophozoites began losing viability around 45 min after treatment with DPI. By 90 min, only about 20% amoebas were still viable. Finally, by 2 h most amoebas were not alive (**Figure 9B**). Therefore, the effect of DPI on NETs production

then analyzed by flow cytometry. Histograms are representative of three independent determinations.

was not due to inhibition of ROS production, but due to a toxic effect on E. histolytica.

# Entamoeba histolytica Did Not Induce ROS Production

In order to confirm that E. histolytica did not induce ROS production, neutrophils were loaded with dihydrorhodamine 123 or with dihydroethidium, and ROS production assessed by flow cytometry and fluorescence microscopy respectively. When neutrophils were treated with PMA, a strong increase in dihydrorhodamine 123 fluorescence could be easily detected by flow cytometry (**Figure 10A**) indicating the presence of ROS. Similarly, PMA induced ROS could be easily detected by fluorescence microscopy (**Figure 10B**). In contrast, E. histolytica did not cause any rise in fluorescence from these ROS indicators (**Figure 10**). Thus, clearly E. histolytica did not induce any ROS production from neutrophils. These data suggest that amoebas can induce NETs production by a signaling pathway that is independent of ROS.

# Entamoeba histolytica-Induced NETs Formation Is Dependent on NF-κB

NETs formation after PMA stimulation requires activation of NFκB (Lapponi et al., 2013). Thus, we explored the involvement of this molecule in E. histolytica-induced NET formation. As previously reported (Alemán et al., 2016a), neutrophils treated with BAY117082, an NF-κB inhibitor, were not able to form NETs after PMA stimulation (**Figure 11**, Supplementary Figure 4). Similarly, neutrophils treated with BAY117082 did not

(100µM). Proteins were resolved by SDS-PAGE, and then Western blotted for phosphorylated-ERK (p-ERK) (upper panel) or for total ERK (lower panel) to show equal loading of proteins. Plots on the right show densitometric analysis for the ratio of pERK/ERK. Data are mean ± SEM of three experiments.

form NETs efficiently after E. histolytica stimulation (**Figure 11**, Supplementary Figure 4). To confirm that NF-κB was activated after E. histolytica stimulation, phosphorylation of NF-κB was detected by Western blotting. E. histolytica induced a rapid and transient NF-κB phosphorylation, which reached a maximum at about 1 min after neutrophils and amoebas got in contact (**Figure 12A**). This phosphorylation then decreased with time and was barely detectable after 10 min. The antibodies used to detect NF-κB and phospho- NF-κB did not recognize any proteins from amoeba cell lysates (**Figure 12A**). E. histolyticainduced NF-κB activation was completely blocked by the NFκB inhibitor BAY117082 (**Figure 12B**). In addition, the MEK inhibitor UO126 also blocked E. histolytica-induced NF-κB phosphorylation (**Figure 12C**), indicating that activation of NFκB is downstream from the ERK signaling pathway. Together these data suggested that amoebas could induce the formation of NETs independently of NADPH-oxidase activation, but with the involvement of NF-κB activation.

# Entamoeba histolytica-Induced NETs Formation Is Independent on Syk, TAK1, PI3K, p38 MAPK, and β2 Integrins

When neutrophils are stimulated through the FcγRIIIb, the kinases Syk and TAK1 are involved in a signaling pathway that leads to NETs formation (Alemán et al., 2016a,b). Similarly some reports suggest that β2 integrins are required for NETs formation (Raftery et al., 2014; Rossaint et al., 2014). Thus, we explored whether these signaling molecules were also involved in E. histolytica-induced NETs formation. Neutrophils pretreated with iSyk, a specific Syk inhibitor formed NETs efficiently after both PMA and E. histolytica stimulation (**Figure 13**). Similarly, inhibition of TAK1 with the antibiotic LLZ 1640- 2, or inhibition of β2 integrins with the blocking monoclonal antibody IB4 did not have any effect on NETs formation (**Figure 13**, Supplementary Figure 5). Also, inhibition of phosphatidylinositol 3-kinase (PI3K) with Wortmannin slightly

histolytica-induced NETs formation. Human neutrophils were not stimulated (—), or were stimulated with 20 nM phorbol 12-myristate 13-acetate (PMA), or with E. histolytica trophozoites. Some neutrophils were previously treated with 10µM diphenyleneiodonium (DPI), a NADPH-oxidase inhibitor (open symbols). The relative amount of NETs was estimated from SYTOX® Green fluorescence in relative fluorescent units (RFU) during 4 h after stimulation. Data are mean ± SEM of three experiments.

reduced, as previously reported (Alemán et al., 2016a), PMAinduced NETosis, but had no effect on E. histolytica-induced NETs formation (**Figure 13**). Finally, inhibition of p38 MAP kinase with the specific inhibitor SB 203580 did not have any effect on NETs formation (**Figure 13**, Supplementary Figure 5). Together these data indicate that these signaling molecules are not involved in the signal pathway activated by amoebas to induce NETs formation.

# DISCUSSION

Neutrophils present several antimicrobial defense mechanisms, including phagocytosis (Rosales and Uribe-Querol, 2017), respiratory burst, degranulation (Kolaczkowska and Kubes, 2013; Mayadas et al., 2014), and the formation of NETs (Yipp et al., 2012). Many pathogens, including virus, bacteria, fungi, and parasites are capable of inducing NETs formation (Papayannopoulos and Zychlinsky, 2009). Although the list of pathogens that induce NETs keeps growing every day, our knowledge about the molecular mechanisms that initiate this neutrophil function is very limited. Recently, we have reported that E. histolytica trophozoites were capable of inducing NETosis in human neutrophils (Ávila et al., 2016), but the role of this process in amoebiasis and the molecular mechanisms implicated in NETs formation were not clarified. In this report, we describe

(ROS) formation. (A) Human neutrophils (PMN) were previously incubated with the ROS-sensitive fluorescent compound DCFDA (15µM), and were not stimulated (—), or were stimulated with 20 nM phorbol 12-myristate 13-acetate (PMA), or with 200µM tert-butyl hydrogen peroxide (TBHP), or with E. histolytica trophozoites. Fluorescence was read in a plate-reader for 2 h at 36◦C. Data are mean <sup>±</sup> SEM of relative fluoresce units (RFU) from three experiments. (B) Neutrophils (PMN) or E. histolytica trophozoites were treated with solvent alone (black symbols) or with the NADPH-oxidase inhibitor diphenyleneiodonium (DPI) (10µM) (white symbols). Cell viability was estimated by Trypan Blue exclusion every 15 min.

for the first time the E. histolytica-induced signaling to activate NETs formation. This signaling pathway involves Raf/MEK/ERK, but it is independent of PKC, ROS, Syk, and TAK1 (**Figure 14**).

Neutrophils, the most abundant leukocytes in blood, are rapidly recruited to sites of infection, where they act as the first line of defense against invading pathogens (Kolaczkowska and Kubes, 2013). Neutrophil activation, through various membrane receptors (Mócsai et al., 2015), is important for initiation of the

various defense mechanisms of these cells. NETs are extracellular fibers formed by chromatin covered with histones (Neeli and Radic, 2012) and antimicrobial proteins derived from neutrophil granules (Brinkmann et al., 2004). NETs seem to act as a physical barrier for preventing pathogen disemination (Papayannopoulos and Zychlinsky, 2009). NETs also display antimicrobial activity that is independent of phagocytosis (Urban et al., 2006). Despite the fact that many pathogens, including virus, bacteria, fungi, and parasites (Papayannopoulos and Zychlinsky, 2009) have all been reported to induce NET formation, no particular receptor for PAMPs has been identified on the neutrophil membrane as responsible for inducing this neutrophil response. However, TLRs have been suggested to participate (Yipp et al., 2012). Only two receptors on the human neutrophil have been reported to be genuine activators of NETs release, the IgA receptor FcαR (Aleyd et al., 2014), and the IgG receptor FcγRIIIb (Behnen et al., 2014; Alemán et al., 2016a). In the case of human protozoan parasites, NETs formation has been described to occur in response to L. amazonensis, L. major, L. chagasi, L. donovani promastigotes (Guimarães-Costa et al., 2009; Gabriel et al., 2010; Hurrell et al., 2015), T. gondii (Abi Abdallah et al., 2012), T. cruzi (Sousa-Rocha et al., 2015), and

Neutrophils were previously treated with solvent alone (—) or with the NF-κB inhibitor BAY 117082 at 5µM. The relative amount of NETs was estimated from SYTOX® Green fluorescence in relative fluorescent units (RFU) at 4 h after stimulation. Data are mean ± SEM of four experiments. Asterisks denote conditions that are statistically different from control (p < 0.03).

E. histolytica (Ávila et al., 2016; Ventura-Juarez et al., 2016). Yet, the mechanism of NETs induction by any of these parasites remains unknown.

Most studies on NETs formation have been conducted with phorbol 12-myristate 13-acetate (PMA) stimulation (Brinkmann et al., 2004; Fuchs et al., 2007). PMA is a direct activator of protein kinase C (PKC), and therefore inhibition of PKC has been shown to block NETs formation (Neeli and Radic, 2013). In agreement with those reports, we found that two different inhibitors of PKC indeed blocked NETs formation after PMA stimulation. In contrast, E. histolytica-induced NETs formation was not affected by PKC inhibition. Since PMA directly activates PKC, any possible receptor involved is bypassed. Thus, in the case of amoeba, it seems that the receptor(s) involved can connect with downstream signaling molecules required for NETosis without the need for PKC. This result indicates that PKC is not always necessary for NETs formation. In contrast, the ERK signaling pathway seems to be a common denominator for NETs formation. In the case of PMA-induced NETosis, it was found that PKC leads to activation of the Raf/MEK/ERK pathway (Hakkim et al., 2011). In the case of FcγRIIIb stimulation, we also found that the ERK pathway is important for NETs formation (Alemán et al., 2016a,b). Now, we describe that E. histolytica also induces activation of the Raf/MEK/ERK pathway for NETs formation, but independently of PKC (**Figure 14**). Nevertheless, we could not identify a particular receptor that would recognize amoeba and activate the Raf/MEK/ERK signaling cascade. Several possible receptors on the neutrophil are candidates for amoeba recognition, including some TLRs. Our group continues exploring this line of research.

Whatever the neutrophil receptor for amoeba is, it clearly connects to Raf kinase and activates ERK signaling. At present, there is no information of how Raf can be activated after E. histolytica recognition by neutrophils. Since Raf is primarily activated by the small GTPase Ras (Lavoie and Therrien, 2015), it is possible that amoebas trigger Ras activation and in turn Raf signaling. But, Raf can also be activated by several other means, including PKC (Takahashi et al., 1999), other small GTPases (Mishra et al., 2005), and even independently of GTPases (Rouquette-Jazdanian et al., 2012). Receptors for growth factors that are usually receptor tyrosine kinases (RTK) activate Raf via the small GTPases Ras (Lavoie and Therrien, 2015) or Rap1 (Mishra et al., 2005). Yet, some RTK, such as the vascular endothelial growth factor (VEGF) receptor can activate Raf independently of Ras using PKC instead (Takahashi et al., 1999). Similarly, in lymphocytes the Tcell receptor uses Pak1 kinase to activate Raf-1 and MEK independently of Ras or PKC (Rouquette-Jazdanian et al., 2012). Thus, in the case of amoebas, Raf activation could be achieved via either a small GTPase or Pak1 kinase. However, another possibility for Raf activation seems more likely to be involved in amoeba-induced Raf activation. In keratinocytes, stimulation with extracellular calcium resulted in activation of Raf and ERK pathway, without the involvement of Ras (Schmidt et al., 2000). In addition, this Raf activation did not connect to the JNK or p38 pathways. In E. histolytica-induced NETs formation, we also found strong calcium mobilization (**Figure 5**), and ERK but not p38 activation (**Figure 13**). Therefore, the extracellular calcium flux into neutrophils that are in contact with amoebas may also serve to activate the Raf/MEK/ERK pathway. This possibility is actually being explored in our laboratory.

An important difference between PMA-induced and amoebainduced as well as FcγRIIIb-induced NETs formation was the time required for NETs release. As previously reported, release of NETs after PMA was detected 3–4 h after stimulation and was dependent on ROS, since the NADPH-oxidase inhibitor DPI abolished NETs release (Brinkmann et al., 2004; Fuchs et al., 2007). In contrast E. histolytica-induced NETs release was much rapid and stronger than the one induced by PMA (**Figure 3**). This response was similar to the rapid, oxidant-independent NETs release described after Staphylococcus aureus stimulation of neutrophils (Pilsczek et al., 2010). ROS are required for NETs formation in most cases (Brinkmann et al., 2004, 2010; Fuchs et al., 2007; Parker et al., 2012a), but ROS are not sufficient, since ROS production induced by phagocytosis cannot initiate NETs formation (Branzk and Papayannopoulos, 2013). ROS production was not detected when human neutrophils were mixed with E. histolytica trophozoites (Díaz-Godínez et al., 2018) and (**Figures 9**, **10**). Therefore, NETs formed after amoeba recognition by neutrophils seem to be independent of ROS. Because, generation of ROS has been reported during the interaction of neutrophils with E. histolytica trophozoites (Sim et al., 2005), our data suggest that the mechanism of NETs formation induced by amoebas is independent of ROS.

Also, it has been previously suggested that interaction of amoebas with neutrophils results in apoptosis (Sim et al., 2005). We did not find evidence for apoptosis of neutrophils interacting with E. histolytica trophozoites (Díaz-Godínez et al., 2018). We do not know exactly the reasons for the different results between that initial report and our present results. Apoptosis was evaluated by surface expression of phosphatidylserine and FcγRIIIb (CD16b) by FACS, and by cleavage of caspases in Western blots (Sim et al., 2005). Under conditions similar to those presented in that report (Sim et al., 2005), we did not detect any increase in phosphatidylserine expression (Díaz-Godínez et al., 2018), or any decrease in CD16b expression using two different antibodies anti-CD16 (**Figure 6**). In addition, no DNA degradation typical of apoptosis could be detected in our cells (Díaz-Godínez et al., 2018). Thus, we think neutrophils do not really undergo apoptosis from interacting with amoebas. Careful reading of the initial report reveals that neutrophils alone kept in culture at 37◦C for 1 h entered spontaneously into apoptosis, as suggested by having 27% of neutrophils positive for propidium iodide staining and 29% of neutrophils stained for annexin-V, a marker for phosphatidylserine (Sim et al., 2005). These percentages increased in the presence of E. histolytica trophozoites and were interpreted as amoebas inducing apoptosis of neutrophils (Sim et al., 2005). Also, the cleavage of caspases was very high in neutrophils alone (Sim et al., 2005), suggesting that neutrophils were already in apoptosis without interacting with amoebas. In addition, reduction of CD16 expression was used as a marker for apoptotic neutrophils. Although, apoptotic neutrophils show reduced expression of CD16 (Dransfield et al., 1994), the rapid loss of CD16 expression is better associated with the response of neutrophils to inflammatory signals (Moldovan et al., 1999). Moreover, neutrophils respond to inflammatory signals by producing ROS (Mayadas et al., 2014; El-Benna et al., 2016), and we did not find any evidence for ROS production in the presence of amoebas. Hence, it seems that, in the initial report, neutrophils were stimulated by other means,

and in consequence some of the cells underwent apoptosis, independently of amoebas.

We believe that a different scenario is taking place when E. histolytica trophozoites interact with neutrophils. Upon recognition of trophozoites, only the neutrophils in direct contact with the parasite release DNA fibers that can completely cover the amoeba (**Figure 2B**). Since, neutrophils cannot phagocytize large cells (Rosales and Uribe-Querol, 2017), they prefer to release NETs in those cases (Urban et al., 2006; Branzk et al., 2014) to prevent the pathogen from escaping. Thus, instead of the amoeba inducing neutrophil apoptosis, it is the neutrophil attacking the trophozoite by undergoing NETosis.

The exact role of NADPH oxidase-dependent ROS for NETs formation remains unclear. When neutrophils were stimulated by Candida albicans or by Group B Streptococcus (GBS), NETs were formed normally by healthy neutrophils when ROS were eliminated by the ROS scavenger pyrocatechol (Kenny et al., 2017), suggesting a ROS-independent via for NETosis. Yet, neutrophils from chronic granulomatous disease (CGD) patients were not able to form NETs in response to the same stimuli (Kenny et al., 2017), indicating a need for ROS in NETs formation. Possible explanations proposed by the authors are that in the case of GBS there was some residual ROS activity, and in the case of C. albicans the fungus itself produces low

kinase C (PKC), which in turn leads to activation of the Raf/MEK/ERK pathway. These kinases finally promote NETs formation. PKC is also required for NADPH-oxidase activation to form reactive oxygen species, which are required for NETs formation after PMA stimulation. In contrast, E. histolytica trophozoites are recognized by neutrophils via a, yet unknown, receptor, which connects to the Raf/MEK/ERK pathway. Also the nuclear factor kappa B (NF-κB) is activated to promote NETs formation. The antibody receptor FcγRIIIb also induces NETs formation via spleen tyrosine kinase (Syk) and transforming growth factor-β-activated kinase 1 (TAK1), which connects to MEK (Alemán et al., 2016b). Other signaling molecules (not shown), such as phosphatidylinositol 3-kinase (PI3K), and p38 MAP kinase are not involved in E. histolytica signaling to NETs formation.

levels of ROS that the neutrophil can use to activate NETosis (Kenny et al., 2017). In addition, other pathogens, such as L. amazonensis, have also been reported to induce NETs in the absence of ROS production (Rochael et al., 2015; DeSouza-Vieira et al., 2016). Together, these studies suggest that PMA absolutely depends on NADPH oxidase derived ROS for NETs formation, while C. albicans and GBS can elude this need to some degree, and parasites such as L. amazonensis and E. histolytica can induce NETosis in complete absence of ROS. In addition to these parasites, various other stimuli can also induce NETosis independently of NADPH oxidase activity, including nicotine, calcium ionophores, uric acid, and immune complexes (Parker et al., 2012b; Arai et al., 2014; Hosseinzadeh et al., 2016; Kraaij et al., 2016). Yet, other sources of ROS such as the mitochondrial respiratory chain or exogenous hydrogen peroxide produced by microorganisms have been considered key for NETosis induced by calcium ionophores (Douda et al., 2015) and by C. albicans (Kenny et al., 2017). Therefore, we cannot abandon the possibility that another ROS source, producing amounts that might not be detected with the methodology we used here, play a role in the amoeba-induced NETosis.

Although, the Raf/MEK/ERK pathway has a central role for NETs formation induced by both PMA- and E. histolytica, as shown by MEK inhibition blocking NETosis, the role of ERK in NETs formation remains unclear. Previously, it was reported that ERK is required for NADPH-oxidase activation (Hakkim et al., 2011), placing ERK upstream of ROS production. However, it has also been suggested that ROS are downstream of ERK activation (Keshari et al., 2013). Since, as discussed above, ROS are essential for PMA-induced NETosis, but they are not needed for amoebainduced NETosis, it seems that NADPH-oxidase activation for NET formation, may proceed not only through an ERK pathway, but also independently of ERK activation, depending on the stimulus (Neeli et al., 2009; Kenny et al., 2017). The actual targets downstream of ERK required for NETosis are still unknown.

One possible molecule activated downstream of ERK that has been implicated in PMA-induced NETosis is the nuclear factor kappa B (NF-κB) (Lapponi et al., 2013). Similarly, E. histolytica-induced NETosis was blocked when NF-κB activation was prevented (**Figures 11**, **12**). How NF-κB connects to NETs formation is a mystery. No clear function for this transcription factor has been reported. Originally, it was proposed that NF-κB would be required to increase the inflammatory response of neutrophils (Lapponi et al., 2013), but this idea has not been formally tested. In addition, it was reported that upon PMA stimulation, gene transcription does not have any role in NETs formation (Sollberger et al., 2016). Moreover, NF-κB is not always needed for NETosis. In the case of FcγRIIIb-induced NETs formation NF-κB was found not to be involved (Alemán et al., 2016b). Thus, the participation for this transcription factor in NETosis seems to depend on the type of stimuli used, and needs further exploration.

A possible explanation for the role of NF-κB, and other transcription factors, in NETosis has been provided recently. Through transcriptomics analyses of neutrophils, it was shown that the transcriptional activity reflects the degree of DNA decondensation occurring during NETs formation (Khan and Palaniyar, 2017). Interestingly, although both ROS-dependent and ROS-independent NETs formation require transcriptional activity, transcription starts at multiple loci in all chromosomes earlier in the rapid ROS-independent NETosis (induced by calcium ionophore A23187) than in the ROS-dependent NETosis (induced by PMA) (Khan and Palaniyar, 2017). Moreover, extensive citrullination of histones in multiple loci was found only during calcium-mediated NETosis, suggesting that citrullination of histone contributes to the rapid DNA decondensation seen in ROS-independent NETosis (Khan and Palaniyar, 2017). These data are in agreement with our findings that amoeba-induced NETs formation is rapid, requires calcium, is independent of ROS, and presents citrullination of histones. Therefore, the rapid activation of NF-κB (**Figure 12**) seems a reflection of the earlier transcriptional activity required for the rapid ROS-independent E. histolytica-induced NETs formation.

It is now generally recognized that there are several mechanisms of inducing NETs formation (Zawrotniak and Rapala-Kozik, 2013; Kenny et al., 2017; Papayannopoulos, 2018), but the particular signaling pathways involved remain confusing. Other signaling molecules that have been suggested to participate in NETosis initiated by the FcγRIIIb are Syk (Popa-Nita et al., 2009) and transforming growth factor-β-activated kinase 1 (TAK1) (Alemán et al., 2016b); in NETosis initiated by immune complexes is phosphatidylinositol 3-kinase (PI3K) (Behnen et al., 2014); in NETosis initiated by LPS (Neeli et al., 2009) and by yeast (Byrd et al., 2013) are p38 MAP kinase, and β2 integrins. Activation of Syk by PMA is dependent on PKC (Popa-Nita et al., 2009). However, inhibition of Syk with iSyk slightly reduced PMA-induced NETosis (Alemán et al., 2016a) and had no effect on E. histolytica-induced NETosis. In the case of FcγRIIIb, iSyk prevented TAK1 phosphorylation and NETs formation (Alemán et al., 2016a). This effect is interpreted as a result of Syk being activated by receptor engagement leading then to TAK1 activation (**Figure 14**). In the case of E. histolytica, the receptor used by neutrophils to recognize amoebas is still unknown. Thus, most likely this putative receptor does not use Syk to deliver a signal for NET formation. Similarly, inhibition of TAK1, PI3-K, or p38 MAP kinase had no effect on E. histolytica-induced NETosis (**Figure 14**). Thus, the putative receptor for E. histolytica recognition most likely connects to the Raf/MEK/ERK pathway independently of these signaling molecules (**Figure 14**). As discussed above, a possible connection for Raf activation might be the extracellular calcium flux.

Blocking β2 integrins with antibodies against both CD11b and CD18 chains prevented NET formation by LPS (Neeli et al., 2009), by β-glucan (Byrd et al., 2013), and by immobilized immune complexes (Behnen et al., 2014). However, integrin ligands are not sufficient to induce NETs formation in isolated neutrophils (Branzk and Papayannopoulos, 2013). Similarly, in our case selective crosslinking of β2 integrins with mAb IB4 also did not induce any NETs formation (Alemán et al., 2016a). Also, the mAb IB4 did not block FcγRIIIb-induced NETs formation (Alemán et al., 2016b). Similarly, blocking β2 integrins with mAb IB4 also did not inhibit E. histolytica-induced NETs formation (**Figure 13**). The involvement of β2 integrins in NET formation might be more related to the adhesion requirement of neutrophils to form NETs (Brinkmann et al., 2010) than to a signaling capacity of the integrin. Therefore, previous reports suggest that β2 integrins cooperate with other receptors to induce NETosis, but our data suggest that β2 integrins do not participate in NETs formation after E. histolytica engagement by neutrophils.

In conclusion, to our knowledge, we show for the first time that E. histolytica activates a signaling pathway for inducing NETs formation, that involves Raf/MEK/ERK, but it is independent of PKC, ROS, Syk, and TAK1. Hence, amoebas activate neutrophils to release NETs via a different pathway from the pathways activated by PMA or by the IgG receptor FcγRIIIb (**Figure 14**). Our results also support the idea that various stimuli promote NETs release via different signaling pathways.

# AUTHOR CONTRIBUTIONS

ZF performed most of the experiments and analyzed data. CD-G performed experiments and discussed data. NM performed Western blots. OA performed calcium experiments and discussed data. EU-Q performed live cell imaging and statistical analysis, discussed data, and prepared figures. JC designed the research, analyzed data, and contributed reagents. CR designed the research, performed statistical analysis, prepared figures, organized the references, and wrote the paper.

# FUNDING

Research in the authors' laboratories was supported in part by Grant 284830 (to JC) and Grant 254434 (to CR) from Consejo Nacional de Ciencia y Tecnología (CONACyT), Mexico (http:// conacyt.mx), by Grant IN206316 (to JC) from Dirección General de Asuntos del Personal Académico, Universidad Nacional Autónoma de México (http://dgapa.unam.mx), and by a special Grant for encouragement of medical research (to JC) from the Miguel Alemán Valdés Foundation, Mexico (https://www. miguelaleman.org).

# REFERENCES


# ACKNOWLEDGMENTS

Authors thank Mario Nézquiz Avendaño (Facultad de Medicina—UNAM) for culturing and providing E. histolytica trophozoites; José Pedraza Chaverri (Facultad de Química— UNAM) for providing dihydrorhodamine 123, dihydroethidium and for advice on ROS assays; and José Carlos Blanco-Camarillo (Instituto de Investigaciones Biomédicas—UNAM) for FACS assays.

# SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00226/full#supplementary-material

neutrophil extracellular traps in response to large pathogens. Nat. Immunol. 15, 1017–1025. doi: 10.1038/ni.2987


Clin. Microbiol. Rev. 13, 318–331. doi: 10.1128/CMR.13.2.318- 331.2000


neutrophils by augmenting free radical generation. Nitric Oxide 22, 226–234. doi: 10.1016/j.niox.2010.01.001


required for apoptosis of human neutrophils induced by Entamoeba histolytica. J. Immunol. 174, 4279–4288. doi: 10.4049/jimmunol.174.7.4279


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Fonseca, Díaz-Godínez, Mora, Alemán, Uribe-Querol, Carrero and Rosales. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# *Entamoeba histolytica* Trophozoites Induce a Rapid Non-classical NETosis Mechanism Independent of NOX2-Derived Reactive Oxygen Species and PAD4 Activity

César Díaz-Godínez <sup>1</sup> , Zayda Fonseca<sup>1</sup> , Mario Néquiz <sup>2</sup> , Juan P. Laclette<sup>1</sup> , Carlos Rosales <sup>1</sup> \* and Julio C. Carrero<sup>1</sup> \*

#### *Edited by:*

Mario Alberto Rodriguez, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Mexico

#### *Reviewed by:*

Humberto Lanz-Mendoza, Instituto Nacional de Salud Pública (INSP), Mexico Indira Neeli, University of Tennessee Health Science Center, United States

#### *\*Correspondence:*

Carlos Rosales carosal@biomedicas.unam.mx Julio C. Carrero carrero@biomedicas.unam.mx

> *Received:* 17 March 2018 *Accepted:* 14 May 2018 *Published:* 05 June 2018

#### *Citation:*

Díaz-Godínez C, Fonseca Z, Néquiz M, Laclette JP, Rosales C and Carrero JC (2018) Entamoeba histolytica Trophozoites Induce a Rapid Non-classical NETosis Mechanism Independent of NOX2-Derived Reactive Oxygen Species and PAD4 Activity. Front. Cell. Infect. Microbiol. 8:184. doi: 10.3389/fcimb.2018.00184

<sup>1</sup> Laboratory of Immunology, Department of Immunology, Instituto de Investigaciones Biomédicas, Universidad Nacional Autónoma de México, Mexico City, Mexico, <sup>2</sup> Laboratory of Immunopathology, Department of Experimental Medicine, Hospital General de México, Facultad de Medicina, Universidad Nacional Autónoma de México, Mexico City, Mexico

Neutrophil extracellular traps (NETs) are DNA fibers decorated with histones and antimicrobial proteins from cytoplasmic granules released into the extracellular space in a process denominated NETosis. The molecular pathways involved in NETosis have not been completely understood. Classical NETosis mechanisms involve the neutrophil elastase (NE) translocation to nucleus due to the generation of reactive oxygen species (ROS) by NADPH oxidase (NOX2) or the peptidyl arginine deiminase 4 (PAD4) activation in response to an increase in extracellular calcium influx; both mechanisms result in DNA decondensation. Previously, we reported that trophozoites and lipopeptidophosphoglycan from Entamoeba histolytica trigger NET release in human neutrophils. Here, we demonstrated in a quantitative manner that NETs were rapidly form upon treatment with amoebic trophozoites and involved both nuclear and mitochondrial DNA (mtDNA). NETs formation depended on amoeba viability as heat-inactivated or paraformaldehyde-fixed amoebas were not able to induce NETs. Interestingly, ROS were not detected in neutrophils during their interaction with amoebas, which could explain why NOX2 inhibition using apocynin did not affect this NETosis. Surprisingly, whereas calcium chelation reduced NET release induced by amoebas, PAD4 inhibition by GSK484 failed to block DNA extrusion but, as expected, abolished NETosis induced by the calcium ionophore A23187. Additionally, NE translocation to the nucleus and serine-protease activity were necessary for NET release caused by amoeba. These data support the idea that E. histolytica trophozoites trigger NETosis by a rapid non-classical mechanism and that different mechanisms of NETs release exist depending on the stimuli used.

Keywords: *Entamoeba Histolytica,* neutrophils, NETosis, NETs, ROS, NOX2, PAD4

# INTRODUCTION

Neutrophils are the most abundant leucocytes in the mammal peripheral blood. Neutrophils have different strategies to combat pathogens including phagocytosis, degranulation and production of neutrophil extracellular traps (NETs). NETs are DNA fibers decorated with histones and antimicrobial proteins from cytoplasmic granules (Papayannopoulos and Zychlinsky, 2009) that are combined in the cytosol and released into the extracellular space in a process denominated NETosis (Steinberg and Grinstein, 2007). NETosis has been described as a cellular death process different from apoptosis and necrosis as it is independent of caspase pathway and there is no phosphatidylserine exposition on cell surface. In addition, NETosis is culminated with DNA extrusion (Fuchs et al., 2007).

The molecular pathways involved in NETosis has not been completely understood. During phorbol miristate acetate (PMA) stimulation, NADPH oxidase (NOX2) assemble take place to produce reactive oxygen species (ROS) (Fuchs et al., 2007). ROS promote the translocation of neutrophil elastase (NE) from the cytoplasmic granules to the nucleus where this enzyme cleaves histones to promote DNA decondensation (Papayannopoulos et al., 2010). Nevertheless, calcium ionophores ionomycin and A23187 appear to induce NETosis independently of NOX2 derived ROS; however, they depend on peptidyl arginine deiminase 4 (PAD4) activity to promote DNA decondensation by changing positive charged arginine residues to none charged citrulline residues in histones (Wang et al., 2009). This mechanism occurs due to an increase in extracellular calcium influx that activates PAD4 (which has four sites for calcium binding), and theoretically any molecule with the capacity to form pores in the cytoplasmic membrane of neutrophils (e.g., bacterial toxins or complement) are able to induce NETosis through this pathway (Konig and Andrade, 2016). Additionally, mitochondrial DNA (mtDNA) NETosis (Yousefi et al., 2009) has been reported in neutrophils which were primed with granulocyte/macrophage-colony stimulatory factor (GM-CSF) and stimulated with lipopolysaccharide (LPS) or complement fraction C5a; this form of NETs release was named vital NETosis because only mtDNA is extruded without compromise cell viability of neutrophils, in contrast with the lethal nuclear NETosis.

Entamoeba histolytica is an intestinal parasite with high prevalence in developing countries (Tellevik et al., 2015; Ghenghesh et al., 2016). Neutrophils have been implicated in defense against this parasite playing a crucial protective role (Asgharpour et al., 2005); nevertheless, involvement of neutrophils and other leukocytes in tissue damage has also been reported (Olivos-García et al., 2007). Oxidative (ROS production) and non-oxidative mechanisms are proposed to be used by neutrophils to kill amoeba (Ghosh et al., 2010; Pacheco-Yépez et al., 2011; Campos-Rodríguez et al., 2016). Mechanisms triggering neutrophils by amoebas are unknown but they could involve innate immune receptors such as TLRs. In this sense, human monocytes TLR2 and TLR4 recognition of lipopeptidophosphoglycan (LPPG) present on cell surface of amoebic trophozoites results in IL-6 and TNF-α release, suggesting that at least amoebic LPPG can activate immune cells trough PRRs (Maldonado-Bernal et al., 2005). Previously, in vitro NETs production in response to E. histolytica trophozoites or the LPPG from this parasite was demonstrated by our group (Ávila et al., 2016). Interestingly, non-viable amoebas failed to induce NETosis and trophozoites treated with PMA-derived NETs did not decreased neither their viability nor the capacity to develop amoebic liver abscess in a hamster model (Ávila et al., 2016). The mechanism of NET induction by E. histolytica is still unknown and its characterization could contribute to our understanding of the still controversial role of neutrophils in amoebiasis.

Here, we demonstrated that E. histolytica trophozoites rapidly induced NETosis by a mechanism independent of NOX2-derived ROS and PAD4 activity; however, this mechanism was dependent on the presence of extracellular calcium and serine-protease activity. These data support the notion of the existence of different NETosis processes that are triggered depending on the stimuli used, the study of which may add to the understanding of the role of these innate immunity mechanisms in parasitic infections.

# MATERIALS AND METHODS

# *Entamoeba histolytica* Trophozoites

Entamoeba histolytica trophozoites (strain HM1:IMSS) were cultured axenically at 37◦C in TYIS-33 medium supplemented with 15% heat-inactivated adult bovine serum and Diamond vitamin tween solution (Sigma). The cultures were incubated for 72 h and trophozoites were harvested by chilling on ice for 5 min and centrifugation at 1,500 rpm for 5 min at 4 ◦C. The pelleted amoebas were resuspended in PBS pH 7.4 and counted. For some experiments, trophozoites were washed with PBS and formaldehyde-fixed or heat-inactivated at 56◦C during 10 and 30 min, respectively.

# Neutrophil Isolation

Neutrophils were isolated from peripheral blood of healthy volunteers using Ficoll-Paque gradient (GE Healthcare) and hypertonic shock to lyse erythrocytes, as previously described (García-García et al., 2013). Cells were resuspended in RPMI medium supplemented with 5% fetal bovine serum and kept until use at 4◦C. This study was carried out in accordance with the recommendations and approval of the Ethical Committee for Studies on Humans of the Instituto de Investigaciones Biomédicas, UNAM. All subjects gave written informed consent in accordance with the Declaration of Helsinki.

# NET Quantification Assay

Neutrophils (3 × 10<sup>5</sup> cells) were stimulated by co-culturing with viable or fixed or heat-inactivated non-viable E. histolytica trophozoites (1.5 × 10<sup>4</sup> cells) in 500 µL of RPMI during 4 h at 37◦C under 5% CO<sup>2</sup> atmosphere; 20 nM PMA was used as positive control of NETosis instead of trophozoites. In selected experiments 10µM of the calcium ionophore A23187 was used as NETosis inducer. After stimulation, supernatants were collected by centrifugation at 4,000 rpm for 2 min to eliminate cells and 50 µL of the supernatants were added to the wells of a 96 wellplate containing 50 µL of PBS pH 7.4 with 500 nM SYTOX Green (Thermo-Fisher). Fluorescence was read from the bottom of wells in a spectrofluorometer using 485 nm excitation and 528 nm emission filters.

To obtain fluorescence values during the first 1 h of stimulation, neutrophils (1 × 10<sup>5</sup> ) were co-cultured with viable E. histolytica trophozoites (5 × 10<sup>3</sup> ) in 100 µL RPMI added with 500 nM SYTOX Green. Fluorescence values were measured each 5 min during 1 h with the 485 nm excitation and 528 nm emission filters as above.

In parallel experiments, neutrophils were pre-incubated before each stimulation with either DMSO-dissolved inhibitors apocynin (200µM), PMSF (1 mM), or GSK484 (10µM) at 4◦<sup>C</sup> during 30 min. In the experiments with EGTA (20 mM), it was added 20 min before stimulation.

# NET Visualization

### Single DNA Staining

NETs staining was performed by co-culturing 2 × 10<sup>5</sup> neutrophils with 1 × 10<sup>4</sup> E. histolytica trophozoites (20:1 ratio) in 200 µL RPMI medium using Chamber SlideTM System (Lab-Tek) for the respective time-points at 37◦C under 5% CO<sup>2</sup> atmosphere. After indicated time-periods, cultures were fixed with 4% formaldehyde for 10 min, washed three times with PBS and stained with 5µg/mL DAPI. The coverslips were mounted with Fluoroshield (Sigma) before observation in a fluorescence microscope (Olympus BX51).

### Immunofluorescence

NETs were induced with E. histolytica trophozoites on the Chamber SlideTM System as described above, and the cells were fixed with 4% paraformaldehyde for 10 min. The fixed cells were then permeabilized by adding 0.2% Triton X-100 in PBS for 5 min. Detergent was washed out three times with cold PBS and unspecific protein binding was blocked with a solution of 1% BSA, 0.3 M glycine, 0.1% Tween 20 in PBS for 30 min at 37◦C. The samples were incubated with primary anti-NE (Santa Cruz Biotechnology), anti-myeloperoxidase (Abcam), anti-histone H4 (Abcam), or anti-citrulline (Abcam) antibodies diluted 1:100 in 1% BSA, 0.1%Tween 20 in PBS during 1 h at room temperature. The cells were gently washed with cold PBS and incubated with secondary anti-mouse IgG-FITC (Sigma) or anti-rabit IgG-TRITC (Zymax) antibodies diluted 1:50 in the same solution that primary antibody for 1 h at room temperature in darkness. The cells were then washed with PBS and stained with 5µg/mL DAPI. The coverslips were mounted with Fluoroshield (Sigma) before observation in a fluorescence microscope (Olympus BX51).

# Scanning Electron Microscopy

For scanning electron microscopy, neutrophils (3 × 10<sup>5</sup> ) were co-cultured with E. histolytica trophozoites (1.5 × 10<sup>4</sup> ) in RPMI medium on cover glass previously coated with poly-lysine. After 4 h of incubation, samples were fixed with 2.5% glutaraldehyde, washed with PBS and dehydrated by increasing concentrations of ethanol. Samples were air dried and coated with gold using Fine Coat Ion Sputter before analysis in an electron microscope Jeol JSM-7600F.

# Apoptosis and Necrosis Detection DNA Fragmentation Assay

DNA integrity assay was performed by incubating neutrophils under different conditions. Neutrophils were stimulated with 20 nM PMA or 10µM A23187 or E. histolytica trophozoites at 37◦C for 1 h to induce NETosis. Necrosis, and in a lesser extent apoptosis, were induced by culturing neutrophils at 56◦C for 1 h. The samples were centrifuged at 7,000 rpm during 5 min, the pellets were resuspended in 20 µL of lysis buffer (2 mM EDTA, 100 mM Tris-HCl, 0.8% SDS, pH 8.0) and 10 µL of 20 mg/mL proteinase K was added. Samples were incubated during 1.5 h at 56◦C and mixed with DNA loading buffer. Finally, the samples were run in 1.5% agarose gel and stained with ethidium bromide for visualization.

### Annexin V Assay

Phosphatidylserine exposure was tested by FITC Annexin V Apoptosis Detection Kit I (Beckton Dickinson). In brief, necrotic cells (and in a lesser extent apoptotic cells) obtained after heattreatment or NETotic cells obtained with PMA, A23187 or amoebas, as described above, were washed with PBS and then resuspended in binding buffer at a concentration of 1 × 10<sup>7</sup> cells/mL. A total of 100 µL from the cell suspension were transferred to a new tube, and added with 5 µL of FITC annexin V. After gently vortexing, samples were incubated in darkness for 15 min at room temperature and observed by fluorescence microscopy. A total of 300 cells were counted to determine the percentage of apoptotic cells in the samples.

# ROS Measurement

ROS generation was determined by flow cytometry using 2′ ,7′ dichlorodihydrofluorescein (H2DCFDA). In brief, 1 × 10<sup>6</sup> neutrophils were pre-incubated with 10µM H2DCFDA (Sigma) for 30 min at 37◦C and washed three times with PBS. Cells were resuspended in 1 mL of PBS, stimulated with 5 × 10<sup>4</sup> E. histolytica trophozoites or 20 nM PMA (positive control for ROS production) and incubated at 37◦C. Cultures were fixed with 4% formaldehyde at 0 or 1 h. Finally, the 2′ 7 ′ dichlorofluorescein (DCF) fluorescence was measured using a FACScalibur cytometer. At least 1 × 10<sup>4</sup> gated events of each sample were analyzed. Not stained and stained but notstimulated cells were used as autofluorescence and negative controls, respectively.

# Amplification of Nuclear and Mitochondrial Genes From NETs

### Obtaining of NET Solution

To obtain NETs solution, neutrophils (1 × 10<sup>6</sup> ) were co-cultured with E. histolytica trophozoites (5 × 10<sup>5</sup> ) in serum-free RPMI for 15 min. Supernatants were collected after centrifugation at 4,000 rpm for 2 min three times to eliminate residual cells. DNA concentration was measured with NanoDrop 2000.

## Neutrophil DNA Isolation

Neutrophils (4 × 10<sup>6</sup> ) were lysed in 400 µl lysis buffer containing 1% SDS, 50 mM Tris–HCl at pH 8, 100 mM EDTA and 200µg/ml proteinase K at 56 ◦C for 2 h. After addition of 125 <sup>µ</sup>l of 5M NaCl solution, DNA was precipitated adding ice-cold isopropanol, re-suspended in 100 µL nuclease-free water, and extracted with phenol-chloroform-isoamyl alcohol (25:24:1). DNA was precipitated from aqueous fraction adding 0.25 M sodium acetate (pH 5.2) and absolute ethanol. After a final wash with icecold 70% ethanol, DNA was resuspended in water (Cotter and Muruve, 2006) and the concentration determined in NanoDrop 2000 (Thermo-Fisher).

## Trophozoite DNA Isolation

DNA from E. histolytica trophozoites was obtained as described elsewhere (Bhattacharya et al., 1988). Trophozoites from 72 h cultures were harvested by centrifugation and resuspended in 5 mL of NET Buffer containing 100 mM NaCI, 10 mM EDTA, 10 mM Tris, pH 8, and lysed with 0.2% SDS. DNA from suspension was extracted with phenol–chloroform (25:24) and precipitated with absolute ethanol. The precipitated DNA was dissolved in TNE Buffer containing 10 mM NaC1, 1 mM EDTA, 10 mM Tris, pH 8, treated with proteinase K at 37◦C for 45 min, and extracted again with phenol-chloroform. After ethanol precipitation, the DNA was dissolved in TNE Buffer and its concentration was measured in NanoDrop 2000.

## PCR Conditions

PCR was perform using HotStarTaq <sup>R</sup> Plus Master Mix Kit (Qiagen) according to the manufacturer instruction. The primers used were as previously designed by Yousefi et al. (2009). For (I) amplification of mitochondrial genes: ATP synthase subunit 6 (atp6) (5′ -ATACACAACACTAAA GGACGAACCT-3′ and 5′ -GAGGCTTACTAGAAGTGT GAAAACG-3′ ), cytochrome oxidase c subunit 1 (co1) (5′ - GGAGTCCTAGGCACAGCTCTAA-3′ and 5′ -GGAGGGTAG ACTGTTCAACCTG-3′ ), NADH dehydrogenase subunit 1 (nd1) (5′ -GCATTCCTAATGCTTACCGAAC-3′ and 5′ -AAG GGTGGAGAGGTTAAAGGAG-3′ ), cytochrome oxidase b (cyb) (5′ -CTAGCAGCACTCCACCTCCTAT-3′ and 5′ -GTTGTCCTC CGATTCAGGTTAG-3′ ); (II) amplification of nuclear genes: glyceraldehyde phosphate dehydrogenase (gapdh) (5′ -CCCCTT CATTGACCTCAACTAC-3′ and 5′ -GAGTCCTTCCACGAT ACCAAAG-3′ ), <sup>β</sup>-actine (actb) (5′ -ATCTGGCACCACACC TTCTACAATGAGCTGCG-3′ and 5′ -CGTCATACTCCTGCT TGCTGATCCACATCTGC-3′ ), and FAS receptor (fas) (5′ -TCA CCACTATTGCTGGAGTCAT-3′ and 5′ -TAAACATCCTTG GAGGCAGAAT-3′ ). PCR protocol included 30 cycles with an alignment temperature of 55◦C for all primer pairs used. The PCR products were run in 1% agarose gels and stained with ethidium bromide for visualization.

# Statistics

Statistical significance was tested with paired two-tailed Student's t-test. Data are reported as mean ± SD. A p-value ≤ 0.05 was considered statistically significant.

# RESULTS

# Characterization of NETs Induced by *E. histolytica* Trophozoites Detection of NETs Components

DNA release during neutrophil/amoeba co-culturing has been reported previously (Ávila et al., 2016). Here, these structures were characterized to determine that they were effectively NETs. Neutrophils and amoebas were co-cultured in a proportion of 20:1, respectively, in RPMI medium supplemented with FBS 5% during 4 h. After this period, unstimulated neutrophils showed a cytoplasmic localization of NE and MPO, meanwhile, histone H4 co-localized with DNA as expected (**Figure 1A**). In contrast, in the neutrophil/amoeba co-culture, DNA extracellular fiber were detected by DAPI staining and these structures co-localized with proteins associated with NETs such as NE, MPO and histone H4 (**Figure 1B**). Similar structures were appreciated in neutrophils stimulated with 20 nM PMA, a well-characterized NETosis inductor (**Figure 1C**). Interestingly, while histone H4 showed a regular distribution along DNA filaments, NE and MPO were detected as immunostained spots only partially colocalized with DNA. Trophozoites were not stained with anti-NE or anti-MPO antibodies and the anti-histone H4 antibody showed a nuclear localization as expected (data not shown).

# Scanning Electron Microscopy of NETs

NETs were analyzed by scanning electron microscopy to observe interaction of trophozoites with extracellular DNA. Non-stimulated neutrophils were visualized as adherent spherical cells with an irregular surface (**Figure 2A**), while trophozoites were detected as pleomorphic cells (**Figure 2B**). After neutrophil–amoeba interaction, trophozoites appeared embedded in the extracellular material (NETs) forming clusters and surrounded by fiber-like structures in close contact with amoebas (**Figures 2C–D**). None-intact neutrophils were visualized in this condition suggesting that all neutrophils undergo NETosis.

# Trophozoites Induce NETosis and no Other Forms of Cell Death

In order to discard other forms of cell death during neutrophil– amoeba interaction, the nuclear morphology at the moments before DNA extrusion was analyzed. Non-stimulated neutrophils showed characteristic condensed-multilobular nuclei (**Figure 3A**), while PMA-treated neutrophils presented decondensed chromatin with loss of multilobular morphology associated to the NETosis process. During neutrophil–amoeba interaction, neutrophils also presented decondensed nuclei before DNA release suggesting NETosis process (**Figure 3A**). To confirm these observations, DNA from neutrophils treated with PMA, A23187, exposed to amoebas or subject to heat was run in agarose gels to verify its integrity. DNA from control neutrophils as well as neutrophils treated with PMA, A23187 or amoebas showed no fragmented DNA (**Figure 3B**), whereas smeared DNA was seen with heat-treated neutrophils. PMA-treated neutrophils, but above all, heat-treated neutrophils showed an increase in phosphatidylserine exposition respect to control

FIGURE 3 | Trophozoites induces NETosis without evidence of apoptosis or necrosis. (A) Neutrophils were incubated with 20 nM PMA or trophozoites (ratio 20:1). After 4 h of incubation, cells were fixed and stained with DAPI. Decondensed chromatin is shown (arrowheads). Images were taken at 100x magnification. Scale bar 10µm. (B) DNA from neutrophils (1 × 10<sup>6</sup> ) treated for 1 h with trophozoites (5 × 10<sup>4</sup> ) or 20 nM PMA or 10µM A23187 or heat (50◦C) were extracted and run in 1.8% agarose gel and the bands visualized by staining with ethidium bromide. (C) Phosphatidylserine (PS) exposition was assessed by fluorescence microscopy using FITC-annexin V. (D) Percentage of cells positive to PS was determined in a total of 300 stained neutrophils. Values are means ± SD of three independent experiments. \*p < 0.001.

(**Figures 3C,D**); meanwhile, A23187 and amoebas did not show phosphatidylserine exposition.

### NETosis Occurs Rapidly and Depends on the Viability of Amoebas

Previously, we demonstrated in a qualitative manner that E. histolytica trophozoites induce a rapid NETs release (Ávila et al., 2016); here, these results were verified quantitatively. We show that neutrophils extruded DNA evidently after 25 min of interaction with amoebas (**Figure 4A**) and the amounts of expelled DNA increased in the time-course of following 4 h (**Figure 4B**). PMA-treated neutrophils underwent NETosis after 1 h of stimulation and control-neutrophils did not release DNA in our experiments. These data were confirmed by fluorescence microscopy: control neutrophils were observed with multilobular and condensed nuclei (**Figure 4C**, panel i), PMA-treated neutrophils showed decondensed chromatin and extracellular DNA fibers (**Figure 4C**, panel ii), whereas in neutrophil-amoebas co-cultures, scarce intact neutrophils were observed and the trophozoites appeared surrounded by extracellular DNA fibers (**Figure 4C**, panels iii–vi).

Fixed amoebas have been reported to be unable to induce NETosis after 1 h of interaction with neutrophils (Ávila et al., 2016). Because some stimuli take more than 1 h to induce this phenomenon, NETs amounts were determined after 4 h of stimulation with either PMA, viable trophozoites or formaldehyde-fixed or heat-inactivated amoebas. As described before, PMA-treated neutrophils and neutrophils co-cultured with viable trophozoites extruded DNA (**Figures 5A,B**, panels iii and iv), but heat-inactivated or formaldehyde-fixed trophozoites failed to induce NETs after the same incubation period even when they were in close contact with neutrophils (**Figures 5A,B**, panels v and vi). These data suggest that the surface molecules present on the trophozoites are not the unique stimuli necessary to induce NETosis.

### NETs Contain Both Nuclear and Mitochondrial DNA

Previous studies show that DNA from NETs can have either nuclear or mitochondrial origin, or both, depending on stimuli used (Yousefi et al., 2009). Here, we assessed the origin of DNA from NETs amplifying three nuclear and four mitochondrial neutrophil genes. As expected, all amplicons were obtained from neutrophil DNA while no amplicons were obtained from purified E. histolytica DNA (**Figure 6A**) indicating that trophozoite DNA co-purified from NETs does not contribute to the amplification. The presence of both nuclear and mtDNA was detected from supernatant of amoebas-neutrophils co-cultured for 15 min (**Figure 6B**). No amplification was obtained using supernatant from PMN cultured alone for 4 h (**Figure 6C**).

# Dependence of NETosis Process of ROS Generation

### NETs Were Generated Independently of NOX2-Derived ROS

To establish the mechanisms involved in the induction of NETs by amoebas, we determined the dependency of this process on NOX2-derived ROS production using apocynin as inhibitor of NOX2 activity (the widely used NOX2 inhibitor DPI caused amoebas death; data not shown). As expected, apocynin (200µM) inhibited NETosis pretended with PMA treatment for 4 h showing that the mechanism of PMA-induced NETosis depends on ROS production. However, apocynin treatment did not decrease DNA release induced by trophozoites indicating that trophozoite-induced NETosis is independent of ROS production by NOX2 (**Figure 7A**). Apocynin treatment did not lead to morphological alterations in control neutrophils as well as did not affect NETs amount visualized by microscopy in neutrophils stimulated with trophozoites (**Figure 7B**).

## ROS Were Not Generated During Neutrophil–Amoeba Interaction

ROS generation has been previously linked to the defense against E. histolytica (Denis and Chadee, 1989). For this reason, assuming that NETosis induced by this parasite was independent of NOX2-derived ROS, the production of ROS from this source was investigated during neutrophil–amoeba interaction by flow cytometry. Population corresponded to neutrophils was selected by flow cytometry according to size and granularity features (**Figure 8A**). Negative region M1 was defined as covering the 98% events in the absence of stimuli; positive region M2 was defined as all values superior to M1.

Basal DCF fluorescence levels were detected in PMAstimulated neutrophils at 0 h but these increased significantly after 1 h of stimulation indicating ROS generation (**Figures 8B,D**). On the other hand, no significant differences were observed in neutrophils co-cultured with trophozoites after 1 h as compared to the starting time point suggesting that ROS production did not take place during interaction of neutrophils with the parasite (**Figures 8C,D**).

# Requirements for Neutrophil Elastase During NET Formation

### NE Is Translocated to Nucleus During NETosis

Nuclear translocation of NE is a step of NETosis that has been proposed to require ROS production by NOX2 (Papayannopoulos et al., 2010). This concept was explored in our model even though amoeba-induced NETosis resulted independent of NOX2-derived ROS. By immunofluorescence, control neutrophils showed multilobular nuclei and NE was localized to cytoplasm, as expected (**Figure 9**), while 20 nM PMA treatment caused NE migration to nucleus after 1 h of stimulation and before DNA decondensation (**Figure 9**). In both cases, a NE remnant was also detected in cytoplasm. Surprisingly, neutrophils co-cultured with trophozoites also presented co-localization of the enzyme with chromatin with patterns of irregular distribution in pre-NETotic nuclei after 15 min of interaction (**Figure 9**). These results indicate that NE translocation to the nucleus can take place independently of ROS generation by NOX2 during NETosis induced by trophozoites.

# Serine-Protease Inhibition Results in Reduction of NET Release

NE is a serine-protease that decondense chromatin after its translocation to nucleus in the process of NETosis

(Papayannopoulos et al., 2010). To study the role of NE in NETosis induced by E. histolytica trophozoites, 1 mM phenylmethylsulfonyl fluoride (PMSF) was used to inhibit its activity in neutrophils. Serine protease inhibition caused a statistically significant decrease in NET amounts released by neutrophils stimulated with 20 nM PMA or E. histolytica

taken at 60x magnification. Scale bar 50µm.

neutrophils (i), control amoebas (ii), neutrophils treated with PMA (iii), neutrophils co-cultured with viable (iv), fixed (v), or heat-inactivated amoebas (vi). Images were

trophozoites (**Figure 10A**). In these experiments, PMSF did not affect cell viability (data not shown). The result was confirmed under fluorescence microscopy showing that PMSF reduced the numbers of extracellular DNA fibers in PMAor trophozoites-treated neutrophils in respect to controls. Additionally, major numbers of non-NETotic cells were noted when neutrophils were pre-treated with PMSF (**Figure 10B**). These results suggest that a serine-protease, probably NE, is required to induce NET release in neutrophils co-cultured with trophozoites.

# Role of PAD4 in Netosis Induced by Amoebas

### Extracellular Calcium Is Required for NETosis

Extracellular calcium influx required for PAD4 activation was considered to be linked to NETosis processes independent of NOX2-derived ROS (Konig and Andrade, 2016). Therefore, herein calcium was chelated from the medium by adding 20 mM EGTA before neutrophil stimulation with 20 nM PMA or trophozoites. EGTA treatment did not affect cell viability of both neutrophils and amoebas (data not shown). EGTA addition reduced significantly DNA extrusion by neutrophils after stimulation with E. histolytica trophozoites; however, NETosis induced by PMA was also reduced (**Figure 11A**). This result was corroborated by fluorescence microscopy showing that extracellular DNA fibers were less abundant with respect to controls in PMA- or amoeba-treated neutrophils when calcium was previously chelated (**Figure 11B**). These data suggest that NETosis induced by amoeba is dependent on extracellular calcium, which appears to be required for NETosis process independently of requirements for PAD4 activity as PMA-induced NETosis do not depend on PAD4.

# PAD4 Inhibition Do Not Affect NET Release

Extracellular calcium dependent-NETosis has been associated to DNA decondensation by PAD4 activity (Douda et al., 2015). In order to explore the participation of PAD4 in NETosis induced by E. histolytica trophozoites, GSK484 was used to inhibit its activity. Neutrophils treated with 10µM GSK484 were able to form NETs after stimulation with PMA; on the other hand, neutrophils stimulated with the calcium ionophore A23187, a known

positive cells to DCF respect to the control. Values are means ± SD of three independent experiments. \*p < 0.001. NS, statistically non-significant.

inducer of PAD4-dependent NETosis, decreased NET release when neutrophils were pre-treated with GSK484 (**Figure 12A**). Surprisingly, neutrophils pre-treated with PAD4 inhibitor and co-cultured with amoebas retained their capacity to form NETs (**Figure 12A**). As seen by fluorescence microscopy, GSK484 did not affect the morphology of neutrophils in respect to controls, and no differences were detected in the release of NETs between neutrophils treated or not with the inhibitor after stimulation with PMA (**Figure 12B**). In a similar way, PAD4 inhibition did not affect NET release in neutrophils co-cultured with trophozoites; however, neutrophils pre-treated with GSK484 formed less NETs upon stimulation with A23187 (**Figure 12B**).

50µm.

FIGURE 10 | Serine protease activity is required for NETosis induce by E. histolytica trophozoites. (A) Neutrophils (3 × 10<sup>5</sup> ) were pre-incubated with 1 mM PMSF or DMSO for 30 min and then stimulated with 20 nM PMA or co-cultured with trophozoites (1.5 × 10<sup>4</sup> ) during 4 h. Supernatants were collected, mixed 1:1 with SYTOX Green (500 nM) and fluorescence was measured. NETs amount is expressed in fluorescence relative units (FRU). Values are means ±SD of three independent experiments in triplicate. \*p < 0.001. (B) Neutrophils were treated as describe previously, fixed and stained with DAPI. Images were taken at 60x magnification. Scale bar 50µm.

FIGURE 11 | NETosis induced by E. histolytica trophozoites require the presence of extracellular calcium. (A) Neutrophils (3 × 10<sup>5</sup> ) were culture in the presence or absence of 20 mM EGTA. EGTA-treated neutrophils were stimulated with 20 nM PMA or co-cultured with trophozoites (1.5 × 10<sup>4</sup> ) during 4 h. Supernatants were collected, mixed 1:1 with SYTOX Green (500 nM) and fluorescence was measured. NETs amount is expressed in fluorescence relative units (FRU). Values are means ± SD of three independent experiments in triplicate. \*p < 0.001. (B) Neutrophils were treated as describe previously, fixed and stained with DAPI. Images were taken at 60x magnification. Scale bar 50µm.

# Detection of Citrullinated Proteins

Citrullinated proteins are considered as markers of PAD4 mediated NETosis (Wang et al., 2009). Here we assayed if NETosis induced by trophozoites generated citrullination of neutrophil proteins even though PAD4 activity was not required for NETosis process. Control and PMA-treated neutrophils showed few citrullinated proteins, principally in cytoplasm, and were less frequent in the decondensed chromatin of PMA-induced neutrophils (**Figure 13**). Conversely, citrullinated proteins were abundant in NETs induced by A23187 or E. histolytica trophozoites, co-localizing with extruded DNA in both cases. Interestingly, citrullinated proteins were detected in a spot pattern similar to the observed during the detection of NE in NETs induced by trophozoites. These results suggest that protein citrullination take place in the NETosis triggered by amoeba.

# DISCUSSION

NETosis is a relatively new mechanism of innate immunity that has been linked to the defense against different pathogens including bacteria, fungi and protozoa by killing or inhibiting their growth, preventing their spread and contributing to the establishment of a protective immune response against pathogens (Brinkmann et al., 2004; Urban et al., 2006; Guimarães-Costa et al., 2009; Röhm et al., 2014; Halverson et al., 2015; Sousa-Rocha et al., 2015). However, contradictory reports have emerged since the formation of NET seems to depend on the pathogen size, with the large pathogens or aggregates of the small ones being responsible for triggering NETosis, while the small non-aggregated pathogens are targeted by phagocytosis (Branzk et al., 2014). In addition, several microorganisms evade the action of NETs by inhibiting their release, degrading the DNA with nucleases or resisting the anti-microbial effect of the NET-associated proteins through encapsulation (reviewed in Storisteanu et al., 2017). On the other hand, the excessive development of NETs has recently begun to be associated with autoimmune and vasculitic diseases, contributing in general to the pathology of some diseases associated with microbial infections (Yipp and Kubes, 2013). Therefore, the role of NETs in the outcome of most infectious diseases is still unknown and is a matter of intensive studies. Nevertheless, the mechanisms underlying the formation of NETs remain poorly understood.

E. histolytica is a parasite that causes intestinal amoebiasis and, in some cases, amoebic liver abscesses. Very early in both pathologies, neutrophils are rapidly recruited to the site of infection becoming the cell of the innate immunity more prevalent in the lesions. However, the role of neutrophils in the amoebic infection is still not well understood. Studies on the outcome of the interaction of neutrophils with amoeba in vitro have shown that priming of these cells with recombinant cytokines IFN-γ and TNF-α make them capable to kill 97% of E. histolytica trophozoites in a H2O2-dependent manner (Denis and Chadee, 1989). In contrast, non-primed neutrophils suffer apoptosis and/or are highly phagocyted when exposed to virulent E. histolytica trophozoites (Sim et al., 2004; Ávila et al., 2016). Likewise, in vivo, while some reports confer to neutrophils a protective activity during the infection (Asgharpour et al., 2005), evidence exists that immune cells, including neutrophils, are implicated in tissue damage (Olivos-García et al., 2007). More recently, we have reported that E. histolytica trophozoites were capable of inducing NETosis in human neutrophils (Ávila et al., 2016), but the role of this process in amoebiasis and the molecular mechanisms implicated in its formation were not clarified. Herein we study some important characteristics associated to the NETosis process induced by neutrophil–trophozoite interaction.

First, to verify the NETosis occurrence in neutrophils cocultures with amoebas, the presence of NETs components was tested. Like other reports, NE, MPO and histone H4 (Brinkmann et al., 2004; Kaplan and Radic, 2012) were found co-localized with extracellular DNA released by neutrophils after their interaction with trophozoites. By scanning electron microscopy, NETs were observed as extracellular fibers trapping clusters of amoebas in a similar way observed with other pathogens (Guimarães-Costa et al., 2009; Brinkmann and Zychlinsky, 2012; Della Coletta et al., 2015). Afterwards, we determined that during neutrophil–amoeba interaction the NETosis process took place and no other forms of cell death. As expected, the NETosis inducer PMA caused nuclear decondensation on neutrophils without affecting DNA integrity; however, an increase in the exposition of phosphatidylserine occurred in <20% of the treated neutrophils, which agrees with a report showing that PMA induces phosphatidylserine exposition in neutrophils (Saito et al., 2005). Necrotic and at lesser extent apoptotic neutrophils induced by treatment with heat showed DNA degradation and phosphatidylserine exposition as it has been reported in other cells (Krysko et al., 2004; Li and Zhou, 2016). Different reports indicate that E. histolytica trophozoites induce apoptosis in diverse cell types including neutrophils (Seydel and Stanley, 1998; Huston et al., 2000; Sim et al., 2004). However, we show that under our conditions, the interaction of trophozoites with human neutrophils resulted in rapid decondensation of the neutrophil nuclei without phosphatidylserine exposition in their external surfaces or genomic DNA breakdown, followed by DNA release to the extracellular space, indicating that neutrophils undergo NETosis and discarding both apoptosis and necrosis in the presence of amoebas. The result was similar to the effect obtained with the calcium ionophore A23187, another NETosis inductor, which did not affect the integrity of the neutrophils DNA and prevented phosphatidylserine exposition in the external surface. Together, the results suggest at this point that amoebas and the calcium ionophore could share similar NETosis pathways which are different to those of PMA.

The most important component of NETs is the DNA released from the neutrophils. Early studies of NETosis showed that traps were exclusively generated from nuclear DNA (nDNA) (Fuchs et al., 2007). Posteriorly, it was observed that under specific conditions, neutrophils were able to form NETs exclusively from mtDNA without compromise the cell viability (Yousefi et al., 2009). Here, we showed that NETs triggered by E. histolytica trophozoites contain both nDNA and mtDNA, as it has been reported using other stimuli as PMA and the nitric oxide donor DETA-NONOate (Keshari et al., 2012). The release of mtDNA during NETosis has been linked to autoimmune pathologies such as eritematous systemic lupus due to its proinflammatory properties (Lood et al., 2016). Thus, mtDNA of NETs could cause a positive feedback through TLR9 propitiating that more neutrophils enter to NETosis (Itagaki et al., 2015). In this regard, we can speculate that mtDNA of NETs induced by trophozoites could contribute to the pathogenesis of the

disease promoting inflammation, a well-known feature of the tissular damage caused by E. histolytica. However, we cannot rule out that mtDNA could come from neutrophils lysed during the interaction with amoebas. Therefore, further studies using specific mtDNA markers are necessary to confirm that NETs induced by amoebas effectively contain this type of DNA.

Initially, NETosis process was described as a slow mechanism when compared with phagocytosis and degranulation (Fuchs et al., 2007), taking ∼2 h since addition of stimulus until DNA release. In contrast, E. histolytica trophozoites induced NET release on few minutes after the contact with neutrophils, similarly to other stimuli as ionomicyn, Staphylococcus aureus or Leishmania amanzonensis promastigotes (Pilsczek et al., 2010; Douda et al., 2015; Rochael et al., 2015). It is worth to mention that these rapid stimuli were shown to induce NETosis through a distinct mechanism compared to PMA and recent evidence effectively suggest that the NETosis process varies depending on the stimuli used (Muth et al., 2017). In this regard, the mechanism underlying the formation of NETs induced by amoebas seems to be non-classical. Thus, ROS production was not detected when human neutrophils were co-cultured with E. histolytica trophozoites in our study and apocynin, a NOX2 inhibitor, failed to inhibit amoeba-induced NETosis. Since generation of ROS has been reported during the interaction of neutrophils with E. histolytica trophozoites at higher ratio of trophozoites that the ratio we used here (10:1 vs. 20:1; Sim et al., 2004, 2005), our results suggest that the mechanism of NETosis induced by amoebas is independent of high concentration of ROS, or at least, of ROS generated from NOX2. Like amoebas, different stimuli induce NETosis independently of NOX2 activity including calcium ionophores, uric acid or immune complexes (Parker et al., 2012; Arai et al., 2014; Kraaij et al., 2016). Nevertheless, we detected that amoebas induced NE translocation to nucleus of neutrophils, an important step during NOX2-dependent NETosis that requires ROS formation (Papayannopoulos et al., 2010; Metzler et al., 2014). Since ROS-independent NE translocation has not been described and evidence suggests that histone cleavage takes place in NOX2 independent NETosis (Muth et al., 2017), we speculate that NOX2-independent ROS could be participating in the NETosis triggered by amoebas promoting NE translocation to the nucleus. Recently, other sources of ROS such as the mitochondrial respiratory chain or exogenous hydrogen peroxide produced by microorganisms have been considered necessary for NETosis induced by calcium ionophores and Candida albicans hypha, respectively (Douda et al., 2015; Muth et al., 2017). Therefore, we cannot discard that another ROS source, producing amounts that might not be detected with the methodology we used here, play a crucial role in the amoeba-induced NETosis. On the other hand, the translocation of NE to the nucleus in combination with our result showing that the inhibition of its serine protease activity with PMSF reduced the amount of NETs induced by amoeba, confirms that NE translocation is an important step in this mechanism.

PAD4 is a peptidyl arginine deiminase that catalyzes the conversion of arginine into citrulline residues in histones and requires calcium for activation (Jones et al., 2009; Bicker and Thompson, 2013). PAD4 activity is relevant to decondense DNA in some NETosis processes independent of NOX2-derived ROS and in the NETosis induced by the non-regulated calcium influx caused by calcium ionophores or calcium channelforming proteins (Wang et al., 2009; Li et al., 2010; Konig and Andrade, 2016). Since E. histolytica possess amoebapores (Lynch et al., 1982), pore-forming peptides implicated in pathogenicity (Bracha et al., 2002; Zhang et al., 2004), we tested in our model the role of PAD4 in amoebic NETosis. Noteworthy, PAD4 inhibition did not affect amoebic NETosis but chelation of extracellular calcium does. This controversy can be explained by the fact that calcium is required for neutrophil adhesion, which in turn, is important for NETosis since neutrophils in suspension produce less NETs compared to adherent neutrophils (Fuchs et al., 2007). Is worth to mention that in contrast to amoebas, PAD4 inhibition reduced NETosis induced by calcium ionophore A23187, suggesting that at this point of nuclear decondensation, the NETosis processes triggered by amoebas and A23187 are different.

Because of its activity, the presence of citrullinated proteins, principally histones, is considered as a marker of PAD4-mediated NETosis (Brinkmann et al., 2016; Li et al., 2017). In accordance to other reports, our results showed that citrullinated proteins were scarce in control and PMA-treated neutrophils (Neeli and Radic, 2013; Konig and Andrade, 2016; Muth et al., 2017); however, abundant citrullinated proteins were seen in neutrophils treated with A23187 and amoebas. Considering that inhibition of PAD4 using GSK484 failed to inhibit NET release induced by amoebas, the results together suggest that protein citrullination take place but it could be no necessary in the NETosis induced by trophozoites. Further experiments are being carried out to elucidate the participation of citrullination in the process. In this regard, it was recently suggested that protein citrullination can be a mechanism that microorganisms employ to evade the immune response by inactivating the antimicrobial action of NETs proteins (Konig and Andrade, 2016). This notion may explain our previous observation that NETs failed to kill trophozoites in vitro and did not ever reduced their capacity to cause amoebic liver abscesses in hamsters (Ávila et al., 2016).

# REFERENCES


In conclusion, our results show that E. histolytica trophozoites trigger NETosis in human neutrophils by a non-classical pathway independent of NOX2-derived ROS and PAD4 activity but dependent of extracellular calcium and the serine protease activity of NE translocated to the nucleus. Our data contribute to understand how NETs are formed in the presence of E. histolytica trophozoites. However, more experiments are required to completely elucidate the mechanism of NETs formation triggered by this parasite and its role in protection or pathogenesis of amoebiasis.

# AUTHOR CONTRIBUTIONS

JC and CD-G conceived and designed the experiments. CD-G, ZF, and MN performed the experiments. CD-G, JC, and CR analyzed the data. JC, JL, and CR contributed reagents, materials, and analysis tools. JC and CD-G wrote the paper.

# FUNDING

This work was supported in part by Grant 284830 (to JC) and Grant 254434 (to CR) from Consejo Nacional de Ciencia y Tecnología (CONACyT), Mexico (http:// conacyt.mx), by Grant IN206316 (to JC) from Dirección General de Asuntos del Personal Académico, Universidad Nacional autónoma de México (http://dgapa.unam. mx), and by a special Grant for encouragement of medical research (to JC) from the Miguel Alemán Valdés Foundation, Mexico (https://www.miguelaleman.org). Funding agencies had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

# ACKNOWLEDGMENTS

CD-G is a Ph.D. student of the Programa de Maestría y Doctorado en Ciencias Bioquímicas, UNAM, and is recipient of a scholarship from Consejo Nacional de Ciencia y Tecnología (CONACyT), Mexico (432205/596731). We thank Patricia de la Torre for help in PCR standardization; Carlos Castellanos-Barba for flow cytometry assistance; Omar Rafael Alemán and Nancy Mora for technical assistance and Pavel Petrosyan for critical reading of the manuscript.


neutrophil extracellular traps in response to large pathogens. Nat. Immunol. 15, 1017–1025. doi: 10.1038/ni.2987


nuclear DNA and exhibit inflammatory potential. Cytometry A 81, 238–247. doi: 10.1002/cyto.a.21178


and early/rapid ROS-independent release of Neutrophil Extracellular Traps triggered by Leishmania parasites. Sci. Rep. 5:18302. doi: 10.1038/srep18302


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Díaz-Godínez, Fonseca, Néquiz, Laclette, Rosales and Carrero. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Reassessing the Role of *Entamoeba gingivalis* in Periodontitis

Mark Bonner <sup>1</sup> , Manuel Fresno2,3, Núria Gironès 2,3, Nancy Guillén4,5 and Julien Santi-Rocca<sup>6</sup> \*

1 International Institute of Periodontology, Victoriaville, QC, Canada, <sup>2</sup> Centro de Biología Molecular Severo Ochoa, Consejo Superior de Investigaciones Científicas, Universidad Autónoma de Madrid, Madrid, Spain, <sup>3</sup> Instituto de Investigación Sanitaria del Hospital Universitario de La Princesa, Madrid, Spain, <sup>4</sup> Institut Pasteur, Paris, France, <sup>5</sup> Centre National de la Recherche Scientifique, CNRS-ERL9195, Paris, France, <sup>6</sup> Science and Healthcare for Oral Welfare, Toulouse, France

The protozoan Entamoeba gingivalis resides in the oral cavity and is frequently observed in the periodontal pockets of humans and pets. This species of Entamoeba is closely related to the human pathogen Entamoeba histolytica, the agent of amoebiasis. Although E. gingivalis is highly enriched in people with periodontitis (a disease in which inflammation and bone loss correlate with changes in the microbial flora), the potential role of this protozoan in oral infectious diseases is not known. Periodontitis affects half the adult population in the world, eventually leads to edentulism, and has been linked to other pathologies, like diabetes and cardiovascular diseases. As aging is a risk factor for the disorder, it is considered an inevitable physiological process, even though it can be prevented and cured. However, the impact of periodontitis on the patient's health and quality of life, as well as its economic burden, are underestimated. Commonly accepted models explain the progression from health to gingivitis and then periodontitis by a gradual change in the identity and proportion of bacterial microorganisms in the gingival crevices. Though not pathognomonic, inflammation is always present in periodontitis. The recruitment of leukocytes to inflamed gums and their passage to the periodontal pocket lumen are speculated to fuel both tissue destruction and the development of the flora. The individual contribution to the disease of each bacterial species is difficult to establish and the eventual role of protozoa in the fate of this disease has been ignored. Following recent scientific findings, we discuss the relevance of these data and propose that the status of E. gingivalis be reconsidered as a potential pathogen contributing to periodontitis.

Keywords: *Entamoeba gingivalis*, periodontitis, gingivitis, inflammation, parasitic infection, infectious disease

# INTRODUCTION: A DISEASE WITH UNDERESTIMATED IMPACT

Periodontitis is a disease leading to alveolar bone destruction and eventually tooth loss. The prevalence of periodontitis is constant among the defined World Health Organization (WHO) regions, with around one person out of two between 35 and 44 years old (Petersen and Ogawa, 2005). This prevalence increases with age (Demmer and Papapanou, 2010). In the USA, between 2009 and 2012, 46% of adults aged 30 years or more suffer from periodontitis (Eke et al., 2015).

Periodontitis is a handicapping disease, for which the WHO calculated the loss of 3,518,002 DALYs (Disability-adjusted life years, a measure of disease burden as the loss of healthy

#### *Edited by:*

Serge Ankri, Technion–Israel Institute of Technology, Israel

#### *Reviewed by:*

Ashu Sharma, University at Buffalo, United States Sim K. Singhrao, University of Central Lancashire, United Kingdom

> *\*Correspondence:* Julien Santi-Rocca jsr@periodontitis.show

#### *Specialty section:*

This article was submitted to Parasite and Host, a section of the journal Frontiers in Cellular and Infection Microbiology

*Received:* 02 May 2018 *Accepted:* 08 October 2018 *Published:* 29 October 2018

#### *Citation:*

Bonner M, Fresno M, Gironès N, Guillén N and Santi-Rocca J (2018) Reassessing the Role of Entamoeba gingivalis in Periodontitis. Front. Cell. Infect. Microbiol. 8:379. doi: 10.3389/fcimb.2018.00379 life years) in 2015 in the world, meaning 0.132% of the worldwide DALYs (Organization, 2016), though the associated disability weight is low (0.007), reflecting only "minor bleeding of the gums from time to time, with mild discomfort" (Evaluation, 2016). Though periodontitis is linked to systemic diseases like diabetes (Nascimento et al., 2018) and ischemic stroke (Leira et al., 2017), the etiological link is difficult to demonstrate and the possible impact of periodontitis on other ailments is ignored for the calculation of DALYs.

The etiology of periodontitis is still unclear and it is classified by the WHO as a non-communicable disease. Some human genetic factors linked to periodontitis were demonstrated (Vieira and Albandar, 2014) and are still investigated to explain its prevalence in the global population, in conjunction with age (Demmer and Papapanou, 2010). Modifiable risk factors for the disease were also sought and some parameters have been identified: smoking (Eke et al., 2016), alcohol consumption (Wang et al., 2016), and poor oral hygiene (Lertpimonchai et al., 2017). Beyond pain from wounds, eventual edentulism, and defects in occlusion, the patients experience halitosis (Silva et al., 2017) and esthetic issues (Nieri et al., 2013). Altogether, these factors may account for their psychological and social distress (Lopez et al., 2012; Hsu et al., 2015; Dumitrescu, 2016; Reynolds and Duane, 2018).

Evolution toward gum disease goes through three stages (i) formation of dental plaque; (ii) gingivitis, which is an inflammation of the gums due to the dental plaque, and (iii) periodontitis, in which alveolar bone and fibers that hold the teeth in place are irreversibly damaged. The pathophysiology of the disease is harshly debated, but a consensus was reached about some key points. First, inflammation is compulsory and prior to bone loss, evidenced by pocket formation [reviewed in Van Dyke (2017)]. Second, the microbial flora in periodontal pockets differs from that observed in healthy sulci (Marsh, 1994). Last, plaque and calculi worsen prognosis (Löe et al., 1965). Consensual treatment in clinics is thus based on the mechanical and/or surgical removal of dental plaque, calculi, and damaged/inflamed tissues (Smiley et al., 2015). These paths lead to an inefficient solution dealing with late symptoms without considering the evoked causes of the disease. The keystone for the improvement of periodontitis management worldwide is a better knowledge of its pathophysiology.

Due to the important correlation of periodontitis with the presence of Entamoeba gingivalis in the oral cavity, here, we searched for the facts that can shed light on the question of whether E. gingivalis plays a role in the occurrence of the periodontitis. In this review, we summarize existing data on the biology of the amoeba Entamoeba gingivalis and on its potential role as an infectious agent in periodontitis. We aim at highlighting perspectives for new research on the pathophysiology and prophylaxis of this neglected disease.

# MICROBIOLOGY OF PERIODONTITIS: THE BACTERIAL PARADIGM

Though the saliva contains low nutrient concentrations and antimicrobial defense systems [reviewed in van 'T Hof et al. (2014)], the healthy oral cavity houses a commensal microbiota, composed of bacterial communities [about 1,000 species across humans, Consortium (2012)], whereas the contribution of viruses, parasites, archaea, and fungi is still to be characterized. Microorganisms and oral mucosae maintain a mutualistic, resilient symbiotic relationship (Rosier et al., 2018). The bacterial ecosystem of healthy sulci is intriguingly similar between individuals and it comprises immotile bacilli and cocci, as seen in microscopy (Listgarten, 1976), with bacterial species differing from those encountered on the tongue (Aas et al., 2005; Consortium, 2012; Rogers and Bruce, 2012). At the tooth surface, in particular in the dental sulcus, nutrients coming from food and cellular debris accumulate and support the survival of bacteria that adhere and colonize the dental enamel. Bacterial flagella, pili, and wall proteins can recruit other bacteria, by co-aggregation (Kolenbrander and Celesk, 1983; Gibbons et al., 1988). Furthermore, the secretion of polysaccharides initiates the formation and organization of a scaffold (Jakubovics, 2010), while intercellular signaling molecules regulate biofilm development, in particular through a quorum sensing mechanism mediated by different types of messengers, as cyclic di-guanosine monophosphate or LuxS [reviewed in Marsh et al. (2011)]. This intra- and inter-species communication leads to coordination of activities and increases the chances of genetic material transfer. The resulting dental plaque is an organized biofilm, whose formation is not pathologic (Gibbons and Van Houte, 1973), though it was thought to be responsible for gingivitis and periodontitis (Schultz-Haudt et al., 1954).

Some bacteria are associated with periodontitis and this led to the proposal of a specific plaque explanation for the disease (Loesche, 1979). These bacteria group in clusters associated with disease progression (Socransky et al., 1998), reflecting the sequential colonization of the periodontal sulcus and pocket (Li et al., 2004; Feres et al., 2016). The "periodontopathogenic" red complex is comprised of anaerobic bacteria (Aggregatibacter actinomycetemcomitans, Tannerella forsythia, and Porphyromonas gingivalis), supporting the hypothesis of a gradual modification of the environment prior to and necessary for colonization by other bacterium types (Marsh, 2003; Darveau, 2010). However, P. gingivalis is present in some healthy patients (Socransky et al., 1998) and is not abundant, even in periodontitis (Moore et al., 1982), while this "keystone pathogen" provokes environmental changes in the sulcus promoting inflammation (Hajishengallis et al., 2011). Thus, P. gingivalis cannot be considered an etiological agent for periodontitis by itself, at least with respect to Koch's postulates.

Koch's postulates are the extreme case of Hill's criteria for causation (Hill, 1965) in which infection by a single etiologic agent is the unique parameter influencing the occurrence of the disease (Inglis, 2007). Thus, the quest for a single pathogen explaining the etiology of periodontitis by itself, following Koch's postulates, may be in vain. Contrariwise, periodontitis, as a biofilm disease (Schaudinn et al., 2009), may result from the integration of various causative parameters. Bacteria are among these parameters and the composition of the microbial communities accurately correlates with clinical outcome (Feres et al., 2016; Hunter et al., 2016). Indeed, some species can be efficiently used as markers for diagnosis (Meuric et al., 2017), re-opening ways for considerations about the use of bacterial identification–though in a multi-variate fashion–for epidemiology or treatment follow-up.

Beside changes in its composition, the bacterial community can harbor changes in its functions, generating a new equilibrium (dysbiosis) that is possible in the new dental plaque environment (Hajishengallis and Lamont, 2012). Bacterial entities collaborate and functional changes, such as synergism, are evidenced at the transcriptional level (Kirst et al., 2015; Yost et al., 2015; Deng et al., 2018). Nevertheless, the abundance of some bacterium species is not synonymous for their activity (Mark Welch et al., 2016), a fact compatible with the "keystone pathogen" theory and the role of P. gingivalis in shaping the ecology of the periodontal pocket. This ecology is impacted by the dysbiotic communities, as evidenced in vitro with deregulated host inflammatory responses (Yost et al., 2017; Herrero et al., 2018). The consequences of dysbiosis were also evidenced in vivo at the systemic level, with metabolic changes linked to diabetes (Branchereau et al., 2016; Blasco-Baque et al., 2017).

# IMMUNO-PATHOLOGICAL PROCESSES TOWARD PERIODONTITIS

While the dental plaque stacks, mineralization leads to formation of tartar deposits, which can cause injury, as well as overhanging restorations or repetitive wounding. In parallel, the constant presence of bacterial components and the possible colonization by periodontopathogens can be sensed by the host and can cause chronic inflammation and an initial tissue lesion [reviewed in Kurgan and Kantarci (2018)]. The host takes part in fueling progression to disease by inflammation and active mediators of inflammation resolution improve the disease's outcome (Lee et al., 2016; Mizraji et al., 2018). Resident leukocytes and endothelial cells respond to bacterial biofilms: vascular permeability increases and interleukin 8 attracts neutrophils to the affected tissues (Tonetti et al., 1998). Neutrophils play a pivotal role in periodontitis (Ryder, 2010), producing reactive oxygen species (ROS), with probable impact at the systemic level [reviewed in Wang et al. (2017)]. While the early lesion progresses, other immune cells are recruited, including macrophages, lymphocytes, plasma cells, and mast cells. The inflammation can be visible at the microscopic level, where rete pegs–epithelial projections into the underlying connective tissue–form in the pocket epithelium and blood vessels proliferate [reviewed in Zoellner et al. (2002)], and at the macroscopic level, in particular by reddening and bleeding. Macrophages are predominant at this stage (Dennison and Van Dyke, 1997); they can derive from circulating monocytes produced in the spleen or be resident macrophages from embryonic origin, like Langerhans cells, with a possible different role during gum inflammation (Moughal et al., 1992). In established lesions, adaptive responses take place and lymphocytes are abundantly detected (Gemmell et al., 2007). Collagen fibers are increasingly altered, leading to a severe tissue remodeling and loosening of the pocket epithelium (Payne et al., 1975). Thus, greater amounts of dental plaque can accumulate in subgingival locations and aggravate gingivitis, which is considered reversible after elimination of the biofilm, in particular due to the absence of bone and periodontal ligament destruction (Ebersole et al., 2017).

Left untreated, gingivitis evolves to periodontitis, which is characterized by an inflammatory infiltrate composed of plasma cells, and by degradation of collagen fibers, loss of connective tissue, and bone destruction in an anaerobic environment (Mettraux et al., 1984). Periodontitis results in clinical attachment loss, i.e., the deeper positioning of the junction between the pocket epithelium and the cementum of the tooth root. The resulting volume below the gum forms the periodontal pocket, witnessing the breaking down of the initial epithelial attachment, the destruction of the connective tissues constituting the periodontal ligament, and the lysis of the alveolar bone. The sulcus depth–the distance between the free gingival margin and the epithelial attachment–is inferior to 3 mm in healthy or gingivitis sites. When superior to 3 mm, periodontitis is suspected and confirmed by inflammation (redness and swelling) and bleeding on probing.

During periodontitis, the balance between bone resorption and regeneration is displaced: Th17 cells induce osteoclastogenesis (Sato et al., 2006). Recently, it has been described that immunopathogenic Th17 lymphocytes [converted from Foxp3+ T cells; a recent review on T cell contribution to periodontitis in Kinane et al. (2017)] that cause bone damage in rheumatoid arthritis can also determine bone resorption and antimicrobial immunity in the oral cavity (Tsukasaki et al., 2018). In human periodontal lesions, Th17 lymphocytes are abundant (Hajishengallis, 2014) and the major source of IL-17. Foxp3+IL-17+ cells are found in the transition state. The generation of pathogenic exFoxp3+TH17 cells in the oral mucosa is dependent on IL-6, which is expressed by periodontal ligament fibroblasts in periodontitis and stimulated by bacterial PAMPs as LPS (Yamaji et al., 1995) through PRRs as TLR-2 and -4 (Sun et al., 2010; Makkawi et al., 2017). Furthermore, osteogenesis is impaired during periodontitis, while bone resorption by osteoclasts is promoted (Zhou et al., 2018), highlighting that not only immune mechanisms are involved in periodontitis pathophysiology. Beside the bacteria and human cells, some archaea, viruses, protozoa, and fungi are differentially present in healthy and diseased sites (Deng et al., 2017). The contributions of these different phenomena, as well as lysis by pathogens or other host immune cells, still need to be elucidated to solve the current paradigm of periodontitis physiopathology, in which only some of the players are visible in the game.

# *ENTAMOEBA GINGIVALIS* AND PERIODONTAL INFECTIONS

# Debate About the Presence of *Entamoeba gingivalis* During Periodontitis Discovery of Entamoeba gingivalis

Though periodontitis was described since antiquity (Langsjoen, 1998), its association with parasites has been evidenced only a century ago. The first description of Entamoeba gingivalis– then named "Amoebea gengivalis"–was laconically performed in 1849 from dental plaque samples, mentioning amoebic movement and the presence of internal vesicles (Gros, 1849). Though free amoebae were described since 1755 (Rösel Von Rosenhof, 1755), E. gingivalis is the first amoeba which was found in humans. The pathogenic association of amoebae with humans was first documented in 1875, validating Koch's postulates in an animal model for the pathogen Entamoeba histolytica, then named "Ameba coli" (Lösch, 1875). The pathogenicity of Entamoeba gingivalis was questioned early (Kartulis, 1893) and the first systematic study associating it with periodontitis was preliminarily published in 1914 (Barrett, 1914): amoebae were detected in the totality of the 46 cases of pyorrhea (periodontitis) that were enrolled in the study. The authors later included 7 healthy individuals from the same group of patients in the "Insane Department of the Philadelphia Hospital" and claimed they could not detect amoebae in "the detritus collected around the neck of the teeth" (Smith and Barrett, 1915a). Furthermore, administration of emetine caused the withdrawal of amoebae and was followed by the cure of pyorrhea in 13 patients (Barrett, 1914). As emetine was thought to be a specific amoebicidal alkaloid with poor bactericidal effect, the etiological link between Entamoeba gingivalis and periodontitis was extrapolated and led Smith and Barrett to rename the disease "amoebic pyorrhea" (Barrett, 1914) or "oral endamebiasis" (Smith and Barrett, 1915b).

# Rejection of Amoebic Etiology for Periodontitis

It is of epistemological importance to underline that systematic studies, some with a low number of participants, about the involvement of Entamoeba gingivalis in periodontitis were countered by specialist opinions without formal experimental proofs, as reported by Craig (Craig, 1916). This controversy led to an almost total abandon of the etiology- and emetinebased therapy, discrediting at the same time its original scientific background. Further statements about the non-permanent cure were made but were unsupported by scientific data, and incriminating relapses (Howitt, 1925). However, the possibility of re-infections was completely ignored and casts doubts about the understanding of both pathophysiology and epidemiology of the disease at that time. In parallel, a method for the culture of E. gingivalis was described, which allowed the study of the effects of emetine hydrochloride on the parasite: "emetine hydrochloride has, apparently, no very marked amoebicidal action in vitro against either of the strains of E. gingivalis used" (Howitt, 1925). However, in a single experiment were evaluated the minimum lethal concentration (MLC, around 116.5µM, 1:16,600 dilution) and the subjectively determined minimal inhibitory concentration (MIC, around 38.5µM, 1:50,000) were evaluated. It is noteworthy that the MLC obtained for the emetine-sensitive HM1:IMSS E. histolytica strain in a more recent study is higher than 100µM, the IC<sup>50</sup> is 29.9µM, and the MIC is lower than 1µM (Bansal et al., 2004). This underlines that the experimental methodology to study the effect of emetine on E. gingivalis was not accurate. The conclusions of this paper should thus be considered with caution because this in vitro study did not corroborate the results reported during the treatment of patients, as previously cited.

### Entamoeba gingivalis Infections in the Genetic Era

After this controversy, only a few studies based on the microscopic detection of E. gingivalis were published, but almost all of them revealed a prevalence of the parasite close to 100% in advanced periodontal pockets (Fisher, 1927; Hinshaw and Simonton, 1928; Wantland and Wantland, 1960; Wantland and Lauer, 1970; Gottlieb and Miller, 1971; Keyes and Rams, 1983; Lange, 1984; Linke et al., 1989). In the cited publications, the prevalence of E. gingivalis in healthy sulci–when studied–ranged from 0 to 26%, suggesting possible errors in the identification of the amoeba. Development of gene amplification by polymerase chain reaction (PCR) and the sequencing of a gene of E. gingivalis (Yamamoto et al., 1995) opened ways for the molecular identification of the parasite and accurate epidemiological studies. The first study– using a long amplicon (1.4 kb) and a sub-optimal DNA purification protocol–revealed a prevalence of 6.25% (2 sites out of 32, from 8 patients) in gingivitis or periodontitis sites, without precision of either their relative number or grade; no amplification was obtained from 20 healthy sites (Kikuta et al., 1996). In the second study, 69.2% of periodontitis sites were positive in real-time PCR for E. gingivalis, while none of the 12 healthy sites included in the study were (Trim et al., 2011). A third study showed a prevalence of 80.6% (58/72) in periodontitis sites and 33.3% (11/33) in healthy sites by conventional PCR, with controls of PCR inhibition and matrix degradation (Bonner et al., 2014). Recently, transcriptomics revealed that E. gingivalis 18s rRNA sequence was detected in all (4/4) periodontal pockets and was less abundant in 60% (6/10), or undetected in 40% (4/10) of healthy sites (Deng et al., 2017). Furthermore, genetic variants of E. gingivalis have been identified (Cembranelli et al., 2013; Garcia et al., 2018b) and different levels of virulence reflected at the transcriptomic levels in genetically identical parasites (Santi-Rocca et al., 2008) may account for discrepancies in their molecular detection as compared with microscopy or clinical diagnoses. The new clinical characterization of periodontitis (Tonetti et al., 2018) will avoid further confusion about the definition of health and various disease grades, that may also be responsible for variability between the studies.

Altogether, these data suggest that E. gingivalis may be asymptomatically present in some sulci and may be associated with the disease after environmental changes, reminiscent of the intestinal pathogen E. histolytica.

# Life Cycle of *E. gingivalis*

In most species of the genus Entamoeba, two cellular forms have been identified in nature: the cyst, which is the contaminant form found in the environment, and trophozoites, the vegetative cell able to divide, that initially derives from excystation of cysts ingested by the host. The survival of these Entamoeba species is ensured by their encystment in response to environmental changes (Mi-Ichi et al., 2016), permitting the survival in environments exposed to oxygen, like human stools, where identification of Entamoeba species is made by a simple morphological phenotyping that relies on the number of nuclei carried by the cyst. The sole E. gingivalis would not encyst, though cysts of E. gingivalis were reported in the literature at the beginning of the twentieth century (Chiavaro, 1914; Smith and Barrett, 1915b; Craig, 1916). However, it is now commonly accepted that E. gingivalis does not produce cysts, considering the absence of proof as a proof of absence. Nevertheless, the parasite E. gingivalis is essentially observed in periodontal pockets, suggesting that low oxygen levels are important for the survival of trophozoites, as in the case of E. histolytica and E. dispar [reviewed in Olivos-Garcia et al. (2016)]. Direct transmission of trophozoites to a new host would imply that they are resistant to oxygen, which raises questions about how E. gingivalis is transmitted in nature, and what ecological niche serves as a reservoir for this microorganism. Unfortunately, the complete life cycle of E. gingivalis is still missing and not addressed yet, hampering efficient prophylaxis.

In the closely-related specie E. histolytica, resistance to oxygen is modulated by interaction with bacteria (Varet et al., 2018), as well as virulence (Bracha and Mirelman, 1984; Galvan-Moroyoqui et al., 2008). The microbiota could be of major importance in switching from commensal to pathogenic forms and explain why only a minor part of E. histolytica intestinal infections are invasive and symptomatic. During periodontitis, bacterial virulence genes are strongly modulated (Deng et al., 2017) and the frequent and abundant detection of E. gingivalis in periodontitis pockets (Bonner et al., 2014) suggests and warns that interactions between constituents of the microbiota could be essential for their functions during the pathophysiology of the disease.

# Ingestion of Human Cells by *E. gingivalis*

Entamoeba gingivalis resembles E. histolytica in several aspects: trophozoites measure about 30µm, they are both endowed with mobility, and they ingest human cells and bacteria. The ability of E. histolytica to kill and phagocytose host cells correlates with parasite virulence and this amoeba is able to feed on human cells: erythrocytes, lymphocytes, and epithelial cells (Christy and Petri, 2011). Two mechanisms of cell killing and uptake have been discovered for E. histolytica: phagocytosis and trogocytosis (Ralston et al., 2014).

Phagocytosis is the phenomenon by which single cells ingest large volumes of material, like other cells or big particles; phagocytic cells include diverse unicellular entities as amoebae, but also macrophages and neutrophils that are cells from the immune system [a recent review in Niedergang and Grinstein (2018)]. In E. histolytica, phagocytosis is indispensable for its nutritional needs since this amoeba ingests bacteria in the intestinal lumen. Moreover, phagocytosis is correlated with virulence because E. histolytica kills human cells that are eventually phagocytosed. The current model is that E. histolytica first kills the host cells in a contact-dependent manner and then phagocytosis of dead cells takes place. However, live cells like bacteria and erythrocytes are also phagocytosed by E. histolytica.

In a recently-discovered second mechanism for cell damage, E. histolytica ingests fragments of live host cells in a nibblinglike process termed "trogocytosis" (Ralston et al., 2014). Though a specific AGC kinase1 was found exclusively involved in trogocytosis (Somlata and Nozaki, 2017), it seems premature to completely dissociate them from the phagocytic process. Furthermore, previous evidence has shown the existence of these structures during phagocytosis of epithelial and endothelial cells by E. histolytica [(Lejeune and Gicquaud, 1987; Nakada-Tsukui et al., 2009), Figure 2 in Faust et al. (2011)] and variability of phagocytic cup shape exist during phagocytosis (Tollis et al., 2010). It has been suggested that amoebic trogocytosis essentially concerns bits of live cells that are internalized, and phagocytosis is the process by which an entire cell is internalized (Ralston, 2015).

Phagocytosis of parts of human cells by E. gingivalis was reported almost a century ago (Child, 1926) but poorly studied since then. Some exceptions however existed and, thanks to video microscopy, the impressive process of cell ingestion by this amoeba has been highlighted (Lyons and Stanfield, 1989; Bonner et al., 2014). Entamoeba gingivalis is able to engulf one or more human cells at the time by a yet-undescribed mechanism (**Figure 1**). In the observed samples, cells around the amoebae present an altered cellular content, suggesting E. gingivalis can trigger signals leading to modification of human cells. As the only human cells observed in these samples were polymorphonuclear cells, and the literature mentions neutrophils are predominant in periodontal pockets, the target cells of amoebic phagocytosis may be the latter. The processes leading to the modifications in nuclear and cytoplasmic morphology in these cells remain to be defined and could be linked, for instance, to proteolytic activity of the amoebae and bacteria, or to a delayed/frustrated NETosis (Neutrophil extracellular traps). Whatever this process is, and whether the amoeba phagocytoses or trogocytosis, the cellular content of the neutrophils leads to one certain point: the first line of defense of cellular innate immunity against E. gingivalis and other organisms in the dysbiotic biofilm lacks its weapon (nuclei for NET formation and gene expression) and is thus unable to accomplish its functions.

# Hot Topics About *E. gingivalis* Infections

The parasite E. gingivalis was identified for its amoeboid movement (Gros, 1849) and characteristics of its locomotion have not been studied since, unlike the closely-related species E. histolytica (Aguilar-Rojas et al., 2016). While observations in wet-mounted slides are possible, culture of the parasite is still a bottleneck for the study of E. gingivalis biology. Division events have not recently been documented and the whole life cycle is yet to be described.

FIGURE 1 | Ingestion of material by an amoeba in a periodontal pocket. Pictures extracted every 5 s from a video-microscopy of saliva-mounted plaque from the deepest part of the periodontal pocket, at 1,000× magnification. In the first panel, the amoeba is pseudo colored in cyan; a black arrowhead indicates its nucleus with the typical peripheral chromatin, while the central karyosome is out of focus. The black arrows indicate food vacuoles. The white arrow designates the internal material (possibly a modified nucleus, perhaps with other subcellular structures) from a host cell (probably a leukocyte), whose ingestion has begun through a "channel," as already observed for trogocytosis and erythrophagocytosis in E. histolytica. After 30 s, a food vacuole begins to form at the extremity of the channel. It is noteworthy that the amoeba continues to emit pseudopods and to move during the process, and that it is surrounded by cells with nuclei of different shapes, or even lacking.

The genetic variability in the species E. gingivalis seems important (Garcia et al., 2018b) and the genetic distance between the ST1 and ST2 variants may indicate great differences in their biology, accounting for their probable association with different pathologies (Garcia et al., 2018a). As the genotyping of E. gingivalis is based on the only gene that has been sequenced, a greater genetic variability may be expected, and further epidemiological studies will precise if some subtypes colonize specific sulcus/pocket environments and, thus, correlate with different clinical outcomes.

The role of the microbiota cannot be ignored and the strong modulation of bacterial genes during periodontitis supports a dysbiotic environment that may impact and be impacted by E. gingivalis parasites. Transcriptomic studies will provide clues about these interactions and will rely on the identification of amoebic genetic material. This will be rendered possible by sequencing the E. gingivalis genome, probably after the axenic culture of the parasite.

The parasite E. gingivalis is more prevalent and more abundant in periodontal pockets, suggesting that this ecological niche is either propitious for its survival, or that the parasite induces changes leading to this environment. Further studies will have to take into consideration the physicochemical and biological characteristics of the periodontal pockets to allow relevant studies of the biology of the parasites, either in vitro or in animal models.

# CONCLUSION

The absence of reliable animal models for infection by E. gingivalis after its axenic culture impedes the ability to conclude about its etiological role in periodontitis, with respect to Koch's postulates. However, recent advances in the field of periodontitis have introduced moderation about this vision, considering that keystone pathogens can participate in changing the environment and consequently in causing dysbiosis, without being the exclusive etiological agent of the disease. Entamoeba gingivalis can be an important agent in the pathophysiology of periodontitis and, since its presence is documented and undoubted, it cannot be ignored.

Interestingly, the natural diversity in the human host has allowed the identification of various components of periodontal infections: some pathological traits are preferentially associated with human genetic variants (Offenbacher et al., 2016) and the P. gingivalis paradigm may be only one of the possibilities or one of the steps of periodontitis pathophysiology. Indeed, evidence about the kinetics of periodontitis setup is scarce and other possible agents must be considered. First, the ST2 "kamaktli" variant, with a high genetic divergence from E. gingivalis ST1 (Garcia et al., 2018b) reminds us that strains of parasites may not be equally virulent or may not have the same tropism (Garcia et al., 2018a). Second, some transcripts from archaea species and viruses are differently abundant between the healthy and the diseased (Deng et al., 2017). In particular, herpes viruses are associated with periodontitis (Slots, 2015; Zhu et al., 2015; Li et al., 2017). Finally, another protozoan is present in some cases of human periodontitis: Trichomonas tenax (Marty et al., 2017). Further studies are needed to decipher the ecosystem of the different stages of the periodontal pockets, assigning roles for all of the detected biological entities, which can be opportunistic, neutral, pathogenic, or mutualistic with the organisms within the pocket, and which can have the same type of interactions with the host tissues. Interestingly, the host directly participates in the pathogenic process: osteogenesis is impaired during periodontitis, while bone resorption by osteoclasts is promoted (Zhou et al., 2018). Furthermore, inflammation has an important role in the disease, since active mediators of inflammation resolution improve the disease's outcome (Lee et al., 2016; Mizraji et al., 2018). The contributions of these different phenomena, as well as lysis by pathogens and host immune cells, still need to be elucidated.

All species of Entamoeba are not necessarily pathogenic and some of them are commensals (e.g., E. dispar or

# REFERENCES


E. coli). The study of E. gingivalis biology during infection, in particular its virulence factors and pathogenic processes, will allow us to better understand the whole interactions in the ecosystem of the periodontal pockets and to determine the potential participation of E. gingivalis in the pathophysiogenesis of periodontitis. Further research should determine if it is taking part in the pathophysiology of periodontitis or just a bona fide marker of the disease. In addition, as the picture is getting more complex and the genetic susceptibility of patients shapes different microbiota (Offenbacher et al., 2016), what we call periodontitis might be a group of related diseases with comparable outcomes. Fundamental research must consider this variability to elucidate the pathophysiology of periodontal diseases and to implement efficient public health measures.

# AUTHOR CONTRIBUTIONS

All authors contributed equally to the redaction of the manuscript, conceived, and directed by JS-R.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Bonner, Fresno, Gironès, Guillén and Santi-Rocca. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

digital media

of impactful research

article's readership