# AMOEBAE AS HOST MODELS TO STUDY THE INTERACTION WITH PATHOGENS

EDITED BY : Sascha Thewes, Thierry Soldati and Ludwig Eichinger 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 978-2-88945-935-3 DOI 10.3389/978-2-88945-935-3

#### 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

# AMOEBAE AS HOST MODELS TO STUDY THE INTERACTION WITH PATHOGENS

Topic Editors:

Sascha Thewes, Freie Universität Berlin, Germany Thierry Soldati, University of Geneva, Switzerland Ludwig Eichinger, University Hospital Cologne, Germany

*Dictyostelium discoideum* cells phagocytosing TRITC-labelled yeast cells. IImage: Maria Stumpf. Image processing: Ludwig Eichinger.

The human body is constantly faced with microorganisms. Most of these bacteria, fungi, and viruses are harmless, many of them are beneficial, and a small fraction is pathogenic. For humans, infection with pathogenic microorganisms can be very serious or even fatal, ranging from mild transient or chronic infections to death. The first line of defence against pathogens is our innate immune system. Beside chemical and physical defence mechanisms of the innate immune system, phagocytic cells such as macrophages play a crucial role in the fight against pathogenic microorganisms. However, phagocytic cells and pathogens are in a constant evolutionary arms race, inventing new strategies to successfully kill pathogens and learning how to resist phagocytosis and intracellular killing, respectively. If pathogens are not obligatory adapted to the human body or other animals, they also have to face environmental phagocytes in the form of amoebae. Many aspects of phagocytosis and intracellular killing are surprisingly well conserved between amoebae and macrophages. Therefore, pathogens that have evolved with environmental amoebae as their "training grounds" can also be successful during infection of macrophages and other animal phagocytic cells.

In this Research Topic, we provide the latest knowledge about the potential of using amoebae as host models to study the interaction with pathogens. The Research Topic covers the interaction of amoebae with bacteria, fungi, and viruses and also illustrates the similarities and differences between amoebae and macrophages. Investigation of evolutionary conserved pathways of amoebae and macrophages furthers our understanding of the biology of host-pathogen interactions and helps to develop new anti-infection therapies.

Citation: Thewes, S., Soldati, T., Eichinger, L., eds. (2019). Amoebae as Host Models to Study the Interaction with Pathogens. Lausanne: Frontiers Media. doi: 10.3389/978-2-88945-935-3

# Table of Contents

*06 Editorial: Amoebae as Host Models to Study the Interaction With Pathogens*

Sascha Thewes, Thierry Soldati and Ludwig Eichinger

#### 1. AMOEBAE AND THEIR INTERACTION WITH BACTERIA 1.1. AMOEBAE AND BACTERIA IN THE ENVIRONMENT

*09 Diversity of Free-Living Environmental Bacteria and Their Interactions With a Bactivorous Amoeba*

Debra A. Brock, Tamara S. Haselkorn, Justine R. Garcia, Usman Bashir, Tracy E. Douglas, Jesse Galloway, Fisher Brodie, David C. Queller and Joan E. Strassmann

*21 Environmental* Mycobacterium avium *subsp.* paratuberculosis *Hosted by Free-Living Amoebae*

Ascel Samba-Louaka, Etienne Robino, Thierry Cochard, Maxime Branger, Vincent Delafont, Willy Aucher, Wilfrid Wambeke, John P. Bannantine, Franck Biet and Yann Héchard

*29 Evolution of Bordetellae From Environmental Microbes to Human Respiratory Pathogens: Amoebae as a Missing Link*

Dawn L. Taylor-Mulneix, Illiassou Hamidou Soumana, Bodo Linz and Eric T. Harvill

#### 1.2. AMOEBAE AS ESTABLISHED HOST MODELS FOR BACTERIAL INFECTIONS

*36* Acanthamoeba *and* Dictyostelium *as Cellular Models for* Legionella *Infection*

A. Leoni Swart, Christopher F. Harrison, Ludwig Eichinger, Michael Steinert and Hubert Hilbi

*53 When Dicty Met Myco, a (Not So) Romantic Story About One Amoeba and its Intracellular Pathogen*

Elena Cardenal-Muñoz, Caroline Barisch, Louise H. Lefrançois, Ana T. López-Jiménez and Thierry Soldati

#### 1.3. THE IMPACT OF MICRONUTRIENTS FOR THE INFECTION PROCESS


#### 1.4. METHODS TO STUDY THE INTERACTION OF AMOEBAE AND BACTERIA

*111 Quantification of Live Bacterial Sensing for Chemotaxis and Phagocytosis and of Macropinocytosis*

Netra P. Meena and Alan R. Kimmel

#### 2. AMOEBAE AND THEIR INTERACTION WITH FUNGI


#### 3. AMOEBAE AND THEIR INTERACTION WITH VIRUSES

*139 Amoebae, Giant Viruses, and Virophages Make up a Complex, Multilayered Threesome*

Jan Diesend, Janis Kruse, Monica Hagedorn and Christian Hammann

#### 4. AMOEBAE AS MODEL ORGANISMS FOR IMMUNE DEFENSE MECHANISMS


Ranjani Dhakshinamoorthy, Moritz Bitzhenner, Pierre Cosson, Thierry Soldati and Matthias Leippe

# Editorial: Amoebae as Host Models to Study the Interaction With Pathogens

Sascha Thewes <sup>1</sup> \*, Thierry Soldati <sup>2</sup> \* and Ludwig Eichinger <sup>3</sup> \*

<sup>1</sup> Department of Biology, Chemistry, Pharmacy, Institute for Biology – Microbiology, Freie Universität Berlin, Berlin, Germany, <sup>2</sup> Department of Biochemistry, Faculty of Science, University of Geneva, Sciences II, Geneva, Switzerland, <sup>3</sup> Medical Faculty, Center for Biochemistry, University Hospital Cologne, Cologne, Germany

Keywords: amoebae, host-pathogen-interaction, evolution, phagocytes, immune system

**Editorial on the Research Topic**

#### **Amoebae as Host Models to Study the Interaction With Pathogens**

Edited and reviewed by:

Thomas Rudel, Universität Würzburg, Germany

\*Correspondence:

Sascha Thewes sascha.thewes@fu-berlin.de Thierry Soldati thierry.soldati@unige.ch Ludwig Eichinger ludwig.eichinger@uni-koeln.de

#### Specialty section:

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

Received: 13 January 2019 Accepted: 13 February 2019 Published: 19 March 2019

#### Citation:

Thewes S, Soldati T and Eichinger L (2019) Editorial: Amoebae as Host Models to Study the Interaction With Pathogens. Front. Cell. Infect. Microbiol. 9:47. doi: 10.3389/fcimb.2019.00047

Amoebae are eukaryotic microorganisms of great diversity. They do not form a single taxonomic group and are found among the protozoa, fungi, and algae. However, all amoebae are characterized by the amoeboid life style—the ability to change cell shape by extending and retracting pseudopods—and most amoebae can be considered as professional phagocytes, which feed on bacteria and other microorganisms by phagocytosis. But despite their impressive phagocytic activity, amoebae cannot degrade all microorganisms. Some resist the digestion and some even use amoebae as host cells for their own replication. Such microbial resistance is also observed in the complex relationship between phagocytes of the mammalian immune system and pathogenic bacteria. For example, Legionella pneumophila usually infects fresh-water amoebae, but can also use human alveolar macrophages to further its own replication. This similarity has led to the concept of "amoebae as training grounds for (intracellular) pathogenic bacteria," namely that pathogenic microorganisms can establish, select and "train" their virulence traits in free-living amoebae before being faced with phagocytic immune cells of animals (Molmeret et al., 2005). Recent studies using amoebae as host models for pathogens confirmed the broad plausibility of this concept. The amoebae used as host models predominantly belong to the phylum Amoebozoa, the closest phylum to fungi and animals, with their most prominent representatives from the genera Acanthamoeba and Dictyostelium. Both have been demonstrated to be useful host cells for the study of the complex interactions with bacterial pathogens such as Legionella, Mycobacterium, Salmonella, Francisella, and others (for reviews see Clarke, 2010; Bozzaro and Eichinger, 2011). Additionally, amoebae can also be used as model hosts for pathogenic fungi like Cryptococcus, Aspergillus, and Candida (Steenbergen et al., 2003; Chrisman et al., 2011; Hillmann et al., 2015; Mattern et al., 2015; Koller et al., 2016; Maisonneuve et al., 2016). These examples highlight the potential of amoebae as model hosts to study the interaction with a wide range of pathogenic microorganisms. In this Frontiers Research Topic a large range of aspects concerning amoebae as host cells for pathogens is covered.

### WHICH AMOEBAE CAN BE USED AS HOSTS?

The currently most used and best-investigated amoebae hosts are Acanthamoeba castellanii and Dictyostelium discoideum. This is highlighted in a review by Swart et al. The authors summarize comprehensively the current knowledge on the interaction of L. pneumophila with its natural host A. castellanii and the versatile model host D. discoideum. The potential of D. discoideum as a host model to study different cellular aspects of the interaction with (pathogenic) bacteria is also presented in a method-oriented article by Meena and Kimmel. The authors share a range of protocols using D. discoideum to study chemotaxis, phagocytosis, and macropinocytosis.

Further amoebae are used to study specific aspects of the interaction with various pathogens. Diesend et al. summarize the current knowledge about the infection of free-living amoebae with giant viruses. Here, the best-studied amoeba is Acanthamoeba polyphaga, which can be infected by a mimivirus. For the study of the interaction of amoebae with pathogenic fungi, the amoebal repertoire is currently the largest. In a review by Novohradská et al. the authors describe the amoebal species that have been used to study fungal interactions. Beside A. castellanii and D. discoideum, amoebae such as the common water contaminant Vermamoeba vermiformis, Protostelium mycophagum, and other mycophagous soil amoebae from the genera Thecamoeba, Arachnula, and Vampyrella have been used to study the interaction with human-, plant-, and entomopathogenic fungi.

#### WHICH PATHOGENS CAN BE STUDIED USING AMOEBAE?

The range of microorganisms, which can be studied using different amoebae, is not limited to a specific group. Amoebae are already well-established model systems to study the interaction with bacteria. The most prominent bacteria studied using A. castellanii and D. discoideum are currently from the genera Legionella and Mycobacterium (see above and reviews by Swart et al.; Cardenal-Muñoz et al.). Two original research articles also highlight this fact. Buracco et al. show that iron depletion or overload of D. discoideum cells impacts on the intracellular growth of Legionella, whereas changes in zinc and copper concentrations do not have any effect. Concerning mycobacteria, Samba-Louaka et al.show that free-living amoebae might serve as a reservoir for Mycobacterium avium subsp. paratuberculosis. The authors found that on farms, where cattle were infected with M. avium subsp. paratuberculosis, the bacteria were also found in environmental samples containing amoebae of the poorly described Rosculus genus. Salmonella enterica serovar Typhimurium is a further example of bacteria that can be studied using amoebae as hosts. Varas et al. show that inorganic polyphosphate (polyP) is crucial for the virulence of these bacteria. Mutation of the polyP kinase (ppK) gene in S. Typhimurium affects the ability of intracellular replication. Finally, the spectrum of pathogenic bacteria is constantly expanding. Recently, it has been shown that Bordetella bronchiseptica can use amoebae as an environmental niche and transmission vector (Taylor-Mulneix et al., 2017). In a perspective article Taylor-Mulneix et al. now ask whether amoebae are the missing link for the evolution of Bordetellae from environmental microbes to human respiratory pathogens. In this article the concept of amoebae as "training ground" for human pathogens is picked up again. Brock et al. underline this ecological relevance for amoebae as a reservoir for (pathogenic) bacteria in an original research article. The authors isolated environmental samples of D. discoideum and found that one third of the wild isolates carried one to six bacterial species per fruiting body. The majority of the isolated bacteria were from the genus Proteobacteria but also Actinobacteria, Bacteriodetes, and Firmicutes have been identified. Many of the bacterial genera isolated in this study include species that are implicated in causing diseases, such as Brucella, Nocardia, pathogenic E. coli, Shigella, and Staphylococcus, suggesting that freeliving amoebae can serve as reservoirs for animal and human pathogenic bacteria.

In recent years the use of amoebae as host cells has been expanded to study interactions with different fungi. Novohradská et al. sum up the current knowledge regarding filamentous fungi (human-, plant-, and entomopathogenic). Watkins et al. show in their original research article that the environmental pathogenic yeast Cryptococcus neoformans escapes from D. discoideum cells after phagocytosis by two different pathways: canonical exocytosis and non-lytic release by vomocytosis, which are the same mechanisms for C. neoformans to escape from macrophages.

Last but not least, amoebae as host cells are not restricted to bacteria or fungi. They can also be used to study the interaction with viruses such as the giant mimivirus (Diesend et al.).

All these examples show that different amoebae can be used with a plethora of microorganisms to learn more about the relationship and interaction from an ecological, evolutionary, and medical point of view.

### WHICH ASPECTS OF PHAGOCYTOSIS AND KILLING BY THE AMOEBAE ARE IN COMMON WITH MAMMALIAN PHAGOCYTES?

Although amoebae are used as host model systems to study interactions with pathogenic microorganisms, one major drawback is that most amoebae grow at environmental temperatures and are usually not able to grow at elevated temperatures such as the human body temperature. However, mammalian phagocytes and amoebae share many conserved cellular and molecular processes concerning phagocytosis and intracellular killing of pathogens. In a mini-review Mori et al. compare and discuss the roles for mammalian and D. discoideum coronins in the trafficking and survival of intracellular pathogens. Although many aspects of the regulation and function of coronins are still elusive, a conserved role for coronins in D. discoideum and mammals can be delineated. Similarly, mechanisms of host defense processes and intracellular virulence of Legionella and Mycobacterium are surprisingly highly evolutionarily conserved between animal macrophages and amoebae, as exhaustively reviewed by Swart et al. and Cardenal-Muñoz et al. Additionally, Watkins et al. show that vomocytosis of Cryptococcus cells is mechanistically conserved between amoebae and mammalian macrophages. Further, Novohradská et al. nicely compare the parallel events in the phagocytic processing of Aspergillus conidiae in mammalian macrophages and amoebae, again supporting the concept of amoebae as training ground for pathogens.

Amoebae are also interesting models to study the early evolution of cell-autonomous, innate immunity (Leippe, 1999), which is reflected in an original research article by Dhakshinamoorthy et al. who investigated the pore-forming saposin-like proteins in the defense against bacterial infection. Such studies show that the limit of amoebae as host models systems is not reached yet.

#### CONCLUSIONS

It is undisputable that phagocytosis and intracellular processing of pathogenic microorganisms is not identical in every aspect in animal phagocytes and amoebae. However, this research topic and abundant recent publications have illustrated that numerous aspects of these processes are evolutionarily conserved between amoebae and animal phagocytes. Further, the advantages of amoebae as host cells are obvious. Amoebae can be cultivated easily in the lab and—especially in the case of D. discoideum can be easily genetically manipulated. Additionally, many other tools for molecular and cell biology have been established for the use with amoebae.

#### REFERENCES


Such powerful systems will allow to further our understanding of the biology and evolution of host-pathogen interactions. This will constitute the basis for the development of novel anti-infection therapies as it has been shown already for Mycobacterium and Legionella (Kicka et al., 2014; Harrison et al., 2015; Ouertatani-Sakouhi et al., 2017; Diop et al., 2018; Trofimov et al., 2018). In summary, the future for amoebae as model systems to unravel host-pathogen interactions has just begun.

#### AUTHOR CONTRIBUTIONS

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

#### FUNDING

ST received funding from the Freie Universität Berlin. The TS lab is supported by grants from the Swiss National Science Foundation (310030\_149390 and 310030\_169386) and the SystemsX.ch initiative grant HostPathX. LE acknowledges support by the Deutsche Forschungsgemeinschaft (CRC670, TP01) and by Köln Fortune.

comparison with other phagocytic cells. Parasitol. Res. 115, 4097–4105. doi: 10.1007/s00436-016-5182-3


**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 Thewes, Soldati and Eichinger. 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.

# Diversity of Free-Living Environmental Bacteria and Their Interactions With a Bactivorous Amoeba

Debra A. Brock <sup>1</sup> \*, Tamara S. Haselkorn<sup>1</sup> , Justine R. Garcia<sup>1</sup> , Usman Bashir <sup>1</sup> , Tracy E. Douglas <sup>1</sup> , Jesse Galloway <sup>2</sup> , Fisher Brodie<sup>2</sup> , David C. Queller <sup>1</sup> and Joan E. Strassmann<sup>1</sup>

#### *Edited by:*

Sascha Thewes, Freie Universität Berlin, Germany

#### *Reviewed by:*

Frank C. Gibson III, University of Florida, United States Eric D. Cambronne, University of Texas at Austin, United States Pauline Schaap, University of Dundee, United Kingdom

> *\*Correspondence:* Debra A. Brock dbrock@wustl.edu

#### *Specialty section:*

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

*Received:* 01 August 2018 *Accepted:* 05 November 2018 *Published:* 23 November 2018

#### *Citation:*

Brock DA, Haselkorn TS, Garcia JR, Bashir U, Douglas TE, Galloway J, Brodie F, Queller DC and Strassmann JE (2018) Diversity of Free-Living Environmental Bacteria and Their Interactions With a Bactivorous Amoeba. Front. Cell. Infect. Microbiol. 8:411. doi: 10.3389/fcimb.2018.00411

<sup>1</sup> Queller/Strassmann Laboratory, Washington University in St. Louis, Department of Biology, St. Louis, MO, United States, <sup>2</sup> Mountain Lake Biological Laboratory, University of Virginia, Mountain Lake, VA, United States

A small subset of bacteria in soil interact directly with eukaryotes. Which ones do so can reveal what is important to a eukaryote and how eukaryote defenses might be breached. Soil amoebae are simple eukaryotic organisms and as such could be particularly good for understanding how eukaryote microbiomes originate and are maintained. One such amoeba, Dictyostelium discoideum, has both permanent and temporary associations with bacteria. Here we focus on culturable bacterial associates in order to interrogate their relationship with D. discoideum. To do this, we isolated over 250 D. discoideum fruiting body samples from soil and deer feces at Mountain Lake Biological Station. In one-third of the wild D. discoideum we tested, one to six bacterial species were found per fruiting body sorus (spore mass) for a total of 174 bacterial isolates. The remaining two-thirds of D. discoideum fruiting body samples did not contain culturable bacteria, as is thought to be the norm. A majority (71.4%) of the unique bacterial haplotypes are in Proteobacteria. The rest are in either Actinobacteria, Bacteriodetes, or Firmicutes. The highest bacterial diversity was found in D. discoideum fruiting bodies originating from deer feces (27 OTUs), greater than either of those originating in shallow (11 OTUs) or in deep soil (4 OTUs). Rarefaction curves and the Chao1 estimator for species richness indicated the diversity in any substrate was not fully sampled, but for soil it came close. A majority of the D. discoideum-associated bacteria were edible by D. discoideum and supported its growth (75.2% for feces and 81.8% for soil habitats). However, we found several bacteria genera were able to evade phagocytosis and persist in D. discoideum cells through one or more social cycles. This study focuses not on the entire D. discoideum microbiome, but on the culturable subset of bacteria that have important eukaryote interactions as prey, symbionts, or pathogens. These eukaryote and bacteria interactions may provide fertile ground for investigations of bacteria using amoebas to gain an initial foothold in eukaryotes and of the origins of symbiosis and simple microbiomes.

Keywords: *Dictyostelium*, bacteria, microbiome, symbiosis, amoebae, sociality, persistence, predation

### INTRODUCTION

Eukaryotes evolved in the context of a world already fully populated by bacteria and archaea (McFall-Ngai et al., 2013). Understanding the consequences and extent of this fairly recent realization is an enterprise still in its infancy. There are several different major approaches that include studies of the importance of gut microbiomes for food acquisition, plant root microbiomes for nutrient up-take, and the obligate symbioses of sap feeding insects and their bacteria (Moran et al., 2005; Alcock et al., 2014; Lareen et al., 2016). These studies and others are changing the way we view the place of eukaryotes in the living landscape. Important and paradigm changing as these studies are, they do not let us go back over two billion years to when eukaryotes were first adapting to the microbial world. Indeed, such a study would not be possible. However, we can make a closer approach to this goal by studying extant lineages from lineages more basal than animals and plants, though of course even their associations with bacteria still reflect modern adaptations and patterns.

Here, we use Dictyostelium discoideum, a single-celled bactivorous soil amoeba from a basal eukaryotic lineage, to investigate the range of interactions a micro-eukaryote has with bacteria in nature. The soil context is important because it is an environment where different organisms are in close contact with an abundance of potential bacterial prey. As such, our study organism exhibits a wide range of interactions with bacteria. First discovered as a predator of soil bacteria (Raper, 1937), D. discoideum has been employed as a model in many host-pathogen studies (Solomon et al., 2000; Alibaud et al., 2008; Hagedorn et al., 2009; Jia et al., 2009; Hasselbring et al., 2011), and has recently been found to engage in symbiotic associations with bacteria throughout their life history (Brock et al., 2011; DiSalvo et al., 2015). However, little is known about what bacteria D. discoideum associate with or prey upon in nature. The natural habitat of wild D. discoideum is deciduous forest soil and leaf litter and it is most commonly isolated from the acidic soils of eastern North America Appalachian forests (Cavender and Raper, 1965a,b; Raper, 1984; Landolt and Stephenson, 1986). Historically, Dictyostelids were first isolated from the dung of various herbivores (Raper, 1984) most notably from deer feces in North American forests (Stephenson and Landolt, 1992; Gilbert et al., 2007). Early work by Kenneth Raper indicated NC-4, the type-clone of D. discoideum isolated in the forests of North Carolina, could feed and develop on both gram-negative and gram positive non-pathogenic bacteria (Raper, 1937). In a later report, Raper and Smith also tested the growth of NC-4 against pathogenic bacteria species that may similarly occur in the soil, ascertaining that D. discoideum has a wide range of prey bacteria (Raper and Smith, 1939). More recently it has become clear that D. discoideum has different mechanisms for dealing with different prey species (Nasser et al., 2013).

D. discoideum can be thought of as a primitive macrophage capable of ingesting, killing and digesting at least one bacterium per minute as long as prey bacteria are available (Cosson and Soldati, 2008). Yet, upon depletion of available food, singlecelled starving amoebae aggregate by the thousands using cAMP signals to form a multicellular organism (Kessin, 2001). During this transformation into a multicellular organism, D. discoideum slugs have some cells capable of immune-like functions (Chen et al., 2007; Brock et al., 2016a). These cells known as sentinel cells can remove bacterial pathogens as well as detoxify the slug to protect the presumptive spore population from harm. Slugs are motile and can migrate toward more favorable locations using light and heat as cues before terminally differentiating into a fruiting body. The fruiting body is composed a spherical sorus containing spores held aloft by a stalk to allow for spore dispersal. For decades, fruiting bodies containing the next generation of Dictyostelium spores were thought to be free from bacteria (Raper, 1937). However, some lineages of wild amoebae that we call farmers have stable interactions with different bacterial partners that are both food and weapons (Brock et al., 2011; Stallforth et al., 2013). Farmers have reduced sentinel cell numbers compared to D. discoideum clones without bacteria, suggesting a relaxation of innate immunity to allow carriage of bacterial associates (Brock et al., 2016a). We have reported two major clades of inedible Burkholderia as the most prominent bacterial partners of D. discoideum farmers (DiSalvo et al., 2015). These stable Burkholderia partners can colonize naïve Dictyostelium hosts and are the drivers of the proto-farming phenotype by allowing co-carriage of food bacteria through multiple social cycles.

To more completely explore the full nature and extent of the associations of a micro-eukaryote with bacteria in natural environments, we conducted a large-scale survey of soil and deer feces at Mountain Lake Biological Station in Virginia. We isolated wild D. discoideum to determine what, if any, culturable bacteria remain associated after D. discoideum amoebae differentiate into fruiting bodies. We focused on culturable bacteria in this survey to perform edibility and persistence assays to interrogate the type of interaction each bacterium had with wild D. discoideum. We found many different bacteria genera, which group into 32-species level OTUs (Operational Taxonomic Unit), transiently associated with the wild D. discoideum recovered from this survey. Here we detail the diversity of bacteria acquired by and maintained in D. discoideum throughout the social stage, and how bacterial diversity differs among different D. discoideum habitats. We show what these bacteria mean to their eukaryote host in the most ancient of relationships: predation and resistance of predation. The data presented here highlight the rich opportunities available to gain greater understanding of eukaryote-bacteria associations using a simple single-celled model organism.

### MATERIALS AND METHODS

#### Deer Feces and Soil Collection

We collected samples of feces from white-tailed deer (Odocoileus virginianus) and of soil on 23 and 30 July 2014 from mixed deciduous forests at Mountain Lake Biological Station, Virginia (37.375654◦ −80.522140◦ ) at Salt Pond Mountain on the eastern continental divide in the Appalachian Mountains (see **Figure 1** for collection locations and **Supplementary Table 1** for GPS coordinates and dates). On each collection date, we first identified and collected ten separate piles of fresh deer feces. In addition to the feces, we collected two soil samples about 5 meters from each feces pile. One soil sample was shallow, just under the leaf litter about one to five centimeters deep. We collected the second soil sample about fifteen centimeters deep, adjacent to the shallow soil sample. We sterilized collection implements using ethanol wipes before and after each use to avoid cross contamination between sampling. This design gave us three samples at each location, one feces, one shallow soil, and one deep soil for a total of thirty samples for each date. From each location, we had four experimental plating categories: deep soil, shallow soil, feces slurry, and feces ball. We divided feces into slurry and ball to examine differences between the interior and surface of feces in regards to our survey method. See below for feces slurry preparation.

### Culture of Dictyostelids From Soil and Feces

We plated all samples within two days of collection on hay agar plates. We prepared hay agar plates following Douglas et al. (2013). For food bacteria, we used overnight cultures of single colonies of Klebsiella pneumoniae grown in sterile Luria Broth (10 gm tryptone, 5 gm Oxoid yeast extract, and 10 gm NaCl per liter DDH2O). K. pneumoniae was obtained from the Dictyostelium Stock Center (http://dictybase.org/StockCenter/ StockCenter.html). To prepare the soil and feces for plating, we weighed five grams of soil or feces from each sample and placed it in a sterile 50 ml Falcon tube with sterile deionized water up to 50 ml. We dispersed the soil and feces in water by vortexing, then pipetted 300 µl of soil or feces slurry, per hay agar plate. We plated ten hay agar plates for each sample: five with 200 µl added overnight culture of food bacteria and five with no added food bacteria but with 200 µl Luria Broth added as control. Plating with and without added food bacteria allows us to germinate as many spores as possible present in the soil and feces samples in case suitable food bacteria for D. discoideum was not present or did not multiply under our growth conditions. After the liquid on each plate was absorbed (leaving the plates open in a biological safety cabinet for about 2–5 min), we added 6–12 activated charcoal pieces (Mars Fishcare Inc.) weighing about 100 milligrams. Activated charcoal can absorb a gaseous repellent released by culminating fruiting bodies aiding in their collection (Bonner and Dodd, 1962). Additionally, we placed one feces ball on each of five hay plates for each site. The feces ball was placed in a small indentation in the agar in the center of the hay plate with no additional bacteria added. We left the hay plates at room temperature (21◦C) and checked them for growth of D. discoideum using a dissecting microscope for 6–10 days after plating. D. discoideum are readily and easily identifiable by their unique morphology compared to other Dictyostelids. We counted the number of distinct groups of fruiting bodies as areas positive for D. discoideum for each type of soil and feces (**Supplementary Table 2**). However, D. discoideum spores from the same soil sample could have come from the same fruiting body during the original collection process. Therefore, if more than one positive area is present on a test plate from the same soil sample, each area has the potential to be genetically different but may also be genetically the same.

#### Bacterial Presence or Absence in *D. discoideum* Fruiting Bodies

We found one to three distinct groupings of D. discoideum fruiting bodies on the positive hay plates, and we tested the sorus from a single fruiting body from each grouping to see if they contained bacteria. To do this, we used a sterile pipette tip to collect the sorus from a single fruiting body while viewing under a dissecting microscope. We placed the sorus in a 1.5 ml Eppendorf tube containing 50 µl of sterile non-nutrient buffer (2.25 g KH2PO<sup>4</sup> and 0.67 g K2HPO<sup>4</sup> per liter DDH2O). Next, we vortexed the Eppendorf tube to disperse the fruiting body sorus contents. We then plated 10 µl of the sorus suspension on each of two SM/5 nutrient agar plates (2 g glucose, 2 g Oxoid bactopeptone, 2 g Oxoid yeast extract, 0.2 g MgSO4, 1.9 g KH2PO4, 1 g K2HPO4, and 15.5 g agar per liter DDH2O) either in association with K. pneumonia or in non-nutrient buffer at room temperature (21◦C). This technique will reveal aerobic bacteria carried in D. discoideum sori that can be cultured, but not unculturable or anaerobic bacteria. We then sequenced 198 16S rRNA bacterial isolates from 95 fruiting body sori after streaking them to pick up a single clonal isolate (**Supplementary Table 3**). Twenty-four of the sequenced bacteria isolates had the same 16S rRNA identity as our food bacteria K. pneumoniae. We removed these bacteria sequences to eliminate any chance they were introduced during our isolation procedure and were not present in the soil or feces samples. We used the remaining 174 sequences in all subsequent calculations.

### PCR Amplification, Bacterial Identification, and Diversity Analyses

We used the procedure outlined in "Identifying Unknown Bacteria using Biochemical and Molecular Methods" (www. nslc.wustl.edu/elgin/genomics/Bio3055/IdUnknBacteria06.pdf) to prepare template bacterial DNA for sequencing with one modification to the procedure. Instead of collecting bacteria grown in overnight liquid cultures, we collected a small amount of stationary phase bacteria clonally grown on a nutrient agar plate and prepared a suspension in water. The PCR amplification was done using a Gene Amp kit from Applied Biosystems (Roche). We used forward sequence 5′ -CGG CCC AGA CTC CTA CGG GAG GCA GCA G-3′ and reverse sequence 5′ -GCG TGG ACT ACC AGG GTA TCT AAT CC-3′ as primers to amplify about 460 bp of the 16S ribosomal RNA gene. The PCR fragments we generated were Sanger sequenced at GeneWiz (www.genewiz.com; South Plainfield New Jersey). Sequences were cleaned using Geneious 6.0 (http://www.geneious.com; Kearse et al., 2012), then aligned and clustered into OTUs (97% similarity) and identified with mothur (Schloss et al., 2009). We aligned sequences and taxonomically classified them using release 128 of the SILVA rRNA database (https://www.arbsilva.de/documentation/release-128/). We generated rarefaction curves and species richness estimates following the guidelines of Gotelli and Colwell (2011), and ran AMOVA (Analysis of

Molecular Variance) analysis in mothur. We pooled the deer feces slurry and ball samples for these analyses because very few bacteria associated with D. discoideum were isolated from the ball samples.

### Phylogeny Construction

We aligned unique haplotypes of our cleaned sequences in Geneious 6.0, along with representative taxa of each major bacterial clade. We reconstructed the phylogeny using a maximum likelihood analysis in Mega 6 (Tamura et al., 2013). We used a general time-reversible model of sequence evolution and rooted the tree at the midpoint. Statistical support was generated using 1,000 bootstrap replicates. We display the phylogeny as a cladogram with edibility and sampling data mapped on using ITOL (Letunic and Bork, 2006). The full phylogeny with bootstrap values is included in the **Supplementary Materials**.

#### Edibility Assay

We tested bacterial isolates for edibility against four wild D. discoideum strains collected from this survey that were unassociated with bacteria. For this assay, we define edible bacteria as those bacteria with the capacity to support growth of D. discoideum. We used 159 bacteria isolates rather than all 174 bacteria species from our survey because fifteen grew too poorly on plates to assay. We chose two D. discoideum strains isolated from soil (4S 6.1 and 7S 4.1) and two D. discoideum strains isolated from feces (14P 2.2 and 18P 7.1). For the assay, we grew these four D. discoideum strains from spores on SM/5 nutrient agar plates supplemented with K. pneumoniae as our food bacteria at room temperature (21◦C). We also grew each test bacterial isolate on SM/5 nutrient agar plates under the same conditions. Using a sterile inoculating loop, we first gathered a small amount of stationary phase test bacteria and then gathered spores from one of the test D. discoideum using the same loop. We streaked the bacteria/spore mixture densely across one quadrant of a subdivided SM/5 nutrient agar plate. We repeated the same procedure with different D. discoideum isolates across the other three quadrants. See **Supplementary Figure 1** for a cartoon schematic of the assay. After 1 week, we scored and tabulated the plates for edibility using the following qualitative categories: excellent (no bacteria present and abundant fruiting bodies), good (few bacteria present and many fruiting bodies), poor (many bacteria present and few fruiting bodies), inedible (abundant bacteria present and no fruiting bodies). See **Figure 7** for cartoon and representative images of edibility assay. We used Pearson's Chi-square goodness of fit to determine if our sample data are consistent with a hypothesized distribution of equal proportions as our null hypothesis, or of dissimilar distributions between feces and soil environments. We used software available from http://quantpsy.org/chisq/chisq.htm to calculate the chisquare test (Preacher, 2001).

#### Persistence Assay

We plated spores from one naïve clone unassociated with bacteria (QS9) individually with eighteen different bacteria genera to test for persistence through multiple D. discoideum social cycle rounds without additional bacteria. Seventeen genera were isolated from this screen. The bacteria genera tested were Achromobacter, Ancylobacter, Burkholderia, Comamonas, Escherichia, Flavobacterium, Oxalicibacterium, Paenibacillus, Pandoraea, Pseudomonas, Rahnella, Rhizobium, Serratia, Shinella, Staphylococcus, Stenotrophomonas, Variovorax. We used the lab food bacteria Klebsiella pneumoniae as a negative (no persistence) control for a total of eighteen genera. We streaked the bacteria onto SM/5 plates from frozen stocks and grew to stationary phase at 21◦C. We prepared bacterial suspensions in non-nutrient buffer at an OD<sup>600</sup> of 1.5. To set up the assay, we collected spores from fruiting bodies under a dissecting scope by touching the top of a sorus using a sterile pipette tip. The spores were placed into sterile non-nutrient buffer. We determined spore density using a hemacytometer and a light microscope. We plated 2 × 10<sup>5</sup> spores and 200 µl of each of the eighteen test bacteria onto SM/5 nutrient agar plates. This is the initial plating. After the initial plating, all subsequent transfers were done bacteria-free meaning spores were collected from each test plate in non-nutrient buffer and plated onto nutrient agar as above without adding additional test bacteria. Bacteria growth and the eventual formation of fruiting bodies would only be able to occur in the test rounds if bacteria persisted in the spores from the initial plating with the test bacteria genera. We grew all rounds at room temperature (21◦C) until fruiting bodies formed. Next, for each round we placed ten individual spore masses (sori) from randomly selected fruiting bodies onto SM/5 nutrient agar (spot test). We recorded bacterial growth and fruiting body formation in the individual test spots of spore masses after 5 days at room temperature (21◦C).

#### RESULTS

We found bacteria associated with 95 of the 254 D. discoideum sori isolated from this survey. The sorus is the spore-containing area located at the top of the fruiting body. The other 161 inspected sori contained no culturable bacteria. We isolated D. discoideum at all twenty locations from at least one type of soil or feces sample using either added or no added food bacteria (**Figure 1**; **Supplementary Table 2**). In total, we clonally isolated 174 culturable bacteria isolates (**Supplementary Table 3**). We found that one to six genetically distinct bacteria isolates can transiently persist through a social cycle in a single sorus isolated from wild D. discoideum (**Figure 2**). Although a fruiting body sorus may contain up to six bacteria isolates, we found the majority (82.1%) contain one or two bacteria isolates per fruiting body sorus.

We constructed a phylogeny of bacteria using unique haplotypes from 16S rRNA sequences (**Figure 3**). These haplotypes are located in four bacteria phyla: Actinobacteria, Bacteriodetes, Firmicutes, and Proteobacteria, the latter represented by three classes (Alpha, Beta, and Gamma). We found the majority (71.4%) of these unique bacterial haplotypes are Proteobacteria and these haplotypes are almost all edible (84.4%). Additionally, the bacteria haplotypes from Actinobacteria, Alphaproteobacteria, and Bacteriodetes were isolated only from D. discoideum originating from feces

environments. There were about twice as many culturable bacteria genera associated with D. discoideum isolated from feces compared to soil environments. Seven genera overlap between the soil and feces environments (**Figure 4**), and these account for 63.8% of bacterial isolates from this survey.

above line graphs. The inset line graph quantifies the total number of bacteria isolates per fruiting body we isolated from all environments combined.

When 16S rRNA sequences were clustered into OTUs at 97% similarity, we found the diversity of bacteria was highest in deer feces (27 OTUs), and lower in both the shallow (11 OTUs) and deep soil (4 OTUs) samples. Our experimental design included adding food bacteria (K. pneumoniae) to half of the selection plates. We found the addition of K. pneumoniae had no effect on the diversity of bacteria we collected from D. discoideum (AMOVA, F (1,172) = 1.04, p = 0.346; with Kp = 25 OTUs, without Kp = 22 OTUs). Since each sample type varied in the number of bacteria we were able to culture, we used rarefaction curves and species richness estimates to compare the diversity between the sample types. None of the rarefaction curves reached saturation, indicating we did not fully sample the diversity in any sample type (**Figure 5A**). However, the Chao1 species richness estimates suggest that the total number of OTUs associated with D. discoideum at Mountain Lake is low and could be fully sampled with moderately more effort. The deer feces had a Chao1 estimated richness of 46.5 bacterial associates (95% CI of 32.2– 99.9) while the shallow and deep soil were estimated to have 13.5 (11.3–28) and 5 (4.1–17.3) bacterial associates, respectively. We also assessed how well the number of sites we sampled captured the bacterial diversity since individual-based rarefaction can overestimate species richness when species distribution is patchy (Gotelli and Colwell, 2011), which is likely for Dictyostelium and its associated bacteria. Rarefaction curves at different taxonomic levels indicate we did not fully sample bacterial diversity except

end of the branches represent the bacteria we isolated during this survey. See Supplementary Table 3 for further details. On the outside edge of the figure, we used circles to represent bacteria from our two environments: feces (brown) and soil (green). We used squares to represent edibility: edible (blue) and inedible (black) for each unique haplotype. Edibility for some bacteria genera had both inedible and edible bacteria and the blue or black edibility symbols square shown here represent a single unique bacterium in that genus.

at the phylum level (**Figure 5B**). As indicated by the sample type rarefaction, most of this undiscovered diversity is likely to be in deer feces.

We asked if the distribution of our four edibility types is the same in both feces and soil environments. D. discoideum amoebae generally lyse and digest phagocytized bacteria very quickly (Clarke and Maddera, 2006). Yet we found numerous bacteria persisting through the multicellular stage, which is triggered by starvation. Are these bacteria persisting because they were inedible and could not be broken down? We scored

recommend by Schloss and Handelsman, 2005. The "unique" category counts every non-identical sequence as a single OTU.

and tabulated edibility using four categories: excellent, good, poor, and inedible (**Figure 6A**). We found 75.2% and 81.8% of D. discoideum-associated bacteria from feces and soil habitats, respectively are edible with a range from excellent to poor. Additionally, we asked if there were any differences between soilcollected and feces-collected bacteria in the degree of edibility. Our null hypothesis is that no relationship exists between the distributions of edibility categorical types from the feces and soil sets in the population. We performed Pearson's chisquare test and rejected this hypothesis. The distributions of bacterial edibility types that associate with D. discoideum differ significantly from expected between soil and feces habitats though not strongly (chi-square = 9.1713; df = 3; p = 0.027; **Figure 6B**). The main driver is the larger proportions of excellent edible type and inedible type bacteria found in feces compared to soil.

discoideum: excellent (no bacteria present and abundant fruiting bodies), good (few bacteria present and many fruiting bodies), poor (many bacteria present and few fruiting bodies), inedible (abundant bacteria present and no fruiting bodies). Bottom panel includes representative images of bacteria and fruiting bodies on nutrient agar plates corresponding to each category from the assay. (B) Chi-square contingency table. The contingency table provides the following information for the edibility data: the top line is the observed total for each edibility type, the middle line in parentheses is the expected total for each edibility type, and the bottom line in brackets is the chi-square statistic for each edibility type.

Avoidance of digestion by bacteria after phagocytosis is thought to have first evolved as a method to resist predatory amoeba (Cosson and Soldati, 2008). We tested seventeen bacteria genera from our screen with the lab food bacteria K. pneumoniae as control to determine if these edible bacteria were able to persist through multiple social cycles (**Figure 7**). Each tested bacteria was the only food source for the D. discoideum amoebae so if persistence occurred the bacteria would be partially evading phagocytosis. We found nine out of 18 bacteria tested persisted through the first round and seven out of 18 through three social cycle rounds (**Figure 7**).

## DISCUSSION

The extent of natural variation in D. discoideum relationships with soil bacteria has been largely unexplored. In this study, we investigated the range of culturable aerobic bacterial associations with D. discoideum fruiting bodies in a natural environment by conducting a survey of forest soil and deer feces at Mountain Lake Biological Station. D. discoideum are predators of bacteria and were previously thought to be bacteria-free during all life stages (Raper, 1937) until we reported some D. discoideum maintain a symbiotic relationship with a handful of Burkholderia species (Brock et al., 2011; DiSalvo et al., 2015). In this report, we demonstrate that newly isolated wild D. discoideum fruiting bodies are far from bacteria-free and are able to transiently carry a broad variety of bacteria throughout their life history. We also establish that many of the bacteria are able to evade phagocytosis and persist through multiple social cycle rounds lab conditions.

Overall, our collection strategy yielded similar numbers of D. discoideum clones from feces compared to soil environments. We isolated numerous bacteria from about one-third of these D. discoideum clones after terminal differentiation when fruiting bodies are formed. Edible bacteria make up the majority of the bacteria we collected (85% of feces and 81% of soil) and the high degree of edibility is in agreement with previous early reports that the type-clone of D. discoideum could feed on a wide range of bacterial species (Raper, 1937; Raper and Smith, 1939). The soil samples seem to be fairly poor sources of D. discoideumassociated bacteria based on the lower diversity of culturable bacteria in shallow soil (11 OTUs) and deep soil (4 OTUs) compared to deer feces (27 OTUs). Despite being extremely

bacteria-rich (Fierer, 2017), soil may have a scarcity of bacteria that serve as prey or are able to persist with D. discoideum. Deer feces, another environment densely colonized by bacteria, harbored more bacteria that could associate with D. discoideum, suggesting the type of habitat D. discoideum grows in likely affects which bacteria are available for uptake. In addition, the relatively few bacteria (Chao1 estimates of 5–47 culturable species) found within D. discoideum living in such bacteria-rich environments suggests that the association may be selective rather than being either promiscuous or haphazard. It is also important to note that we focused here on culturable aerobic bacteria in order to test the edibility and persistence of the bacteria collected. However, there are likely to be other bacteria associated with D. discoideum that cannot grow on the media we used, and a culture-free survey is necessary to get a complete picture of the D. discoideum microbiome.

Klebsiella, Paenibacillus, Rahnella, Shinella, Staphylococcus, and Variovorax.

The many bacteria we isolated associated with our wild D. discoideum isolates may or may not be residing inside the spores. These bacteria could be associating with D. discoideum in some other fashion. We have previously established that some wild D. discoideum known as farmers facultatively associate with certain bacteria, that these bacteria are located inside spores, and that they can be vertically transmitted to the next generation (Brock et al., 2011; DiSalvo et al., 2015). In this study we found several bacteria genera were able to persist through multiple social cycle rounds suggesting wherever these bacteria are located they are able to be transmitted to the next generation in some fashion. However, spot test percentages were lower in the persistence assay than we have previously shown with symbiotic Burkholderia sp. nov. in D. discoideum (DiSalvo et al., 2015), so carriage in the spores or fidelity of transmission of the tested bacteria genera may be less than we found for our symbiotic Burkholderia. Where bacteria from this survey are located within the D. discoideum fruiting bodies will need to be determined by future research, but it seems likely that at least a portion of them could be persisting inside the spores.

Non-digestion (persistence) of bacteria in hosts can be aided by manipulation of the host immune response, and bacteria have developed an arsenal of diverse mechanisms to do this (Sansonetti and Di Santo, 2007; Ribet and Cossart, 2015). Secretion systems are commonly used by bacteria to secrete effectors into a host cell after engulfment (Green and Mecsas, 2016). These secreted effectors facilitate escape of the bacteria from the phagosome or block phagosome fusion to the lysosome preventing bacterial cell death thereby creating a new niche in the host. Proteobacteria are particularly rich in secretion systems (T1SS, T2SS, T3SS, T4SS, T5aSS, T5bSS, T5cSS, and T6SS) having more different types of secretion systems than all other phyla (Abby et al., 2016). The majority of bacteria we isolated in total were in the Alpha, Beta, and Gamma classes of Proteobacteria and this was also true of the type bacteria that persisted through multiple social rounds of growth. A similar abundance of Proteobacteria classes was found among digestionresistant bacteria in surveys of marine and fresh water ciliates, another group of unicellular eukaryotes (Pucciarelli et al., 2015; Gong et al., 2016). Gong et al. (2016) suggested bacteria with type IV and VI secretion systems (T4SS and T6SS) as having a possible role in promoting these marine protist/bacteria associations. Both type IV and VI secretion systems transport proteins and effector molecules into eukaryotic cells and have been implicated in suppressing immunity and exporting virulence factors (Cascales and Christie, 2003; Bingle et al., 2008). Type VI secretion systems have also been shown to have nonpathogenic roles mediating symbiotic relationships between bacteria and eukaryotes as well (Parsons and Heffron, 2005; Chow and Mazmanian, 2010; Jani and Cotter, 2010). However, Proteobacteria are often abundant in forest soil (Janssen, 2006), so we are not certain how over-represented they might be in our samples.

Secretion systems are also one of the mechanisms bacteria use to secrete effectors that directly target components of the innate immune system such as Toll-like receptors and Nodlike receptors (Reddick and Alto, 2014). When multicellular, D. discoideum contain a cell type known as sentinel cells that provides immune-like and detoxification functions (Chen et al., 2007; Brock et al., 2016a). Additionally, Chen et al. (2007) found a Toll/Interleukin-1 Receptor (TIR) domain protein that is necessary for sentinel cells to effectively target bacteria to allow clearance of the bacteria from the slug. We have previously shown that wild D. discoideum farmers have fewer sentinel cells compared to D. discoideum with no associated bacteria suggesting a relaxed immune response promoting symbiotic interactions through a reduction in the clearance of carried bacteria (Brock et al., 2016a). The bacteria we found associated in the sori of wild D. discoideum fruiting bodies may be persisting through a similar mechanism.

Amoebae are proposed as reservoir hosts for pathogenic bacteria (Molmeret et al., 2005; Scheid, 2014; Strassmann and Shu, 2017). In support, Benavides-Montaño and Vadyvaloo recently reported Yersinia pestis can survive in Acanthamoeba castellanii amoebae for prolonged periods by using the type three secretion system to inhibit phagocytosis (Benavides-Montaño and Vadyvaloo, 2017). Bacteria can increase in virulence even after short associations with eukaryotes leading to environmental and health consequences. One example reports enhanced virulence of Mycobacterium avium in macrophage and mouse models after growth in A. castellanii (Cirillo et al., 1997). Another recent example describes increased dispersal ability and amplified virulence of Bordetella bronchiseptica after growth in D. discoideum along with macrophage and mouse models (Taylor-Mulneix et al., 2017). Many of the bacterial genera we isolated in this study are known to have members implicated in causing disease, such as Brucella (Brucellosis), Escherichia/Shigella (urinary tract infections/diarrhea), Nocardia (Nocardiosis), and Staphylococcus (urinary tract infections), suggesting similar possibilities for health consequences.

Digestion of bacteria by amoebae should occur quickly after engulfment (Clarke and Maddera, 2006). However, our ability to isolate culturable bacteria from wild fruiting body spore masses formed after starvation suggests the possibility of some level of persistence and host manipulation. Amoebae/bacteria relationships are important because they can inform us about the interdependent relationships of bacteria to all eukaryotes (Gilbert et al., 2012). The vast range of eukaryote-bacteria interactions can affect development, immunity, and evolution (Rosenberg and Zilber-Rosenberg, 2016). Based on the data presented here, we find a more multifaceted relationship between D. discoideum and bacteria than just that of predator/prey. D. discoideum has already been utilized extensively as a model system to study host/pathogen relationships as detailed in a recent review (Bozzaro and Eichinger, 2011), and more recently to study symbiotic relationships (Brock et al., 2011, 2013, 2016b; DiSalvo et al., 2015). This system will provide an excellent platform to discover if bacteria persistence in wild D. discoideum amoebae can initiate either increased infectivity in other hosts or, alternatively, symbiosis. Our study shows that the range of bacterial associates, whether pathogenic, symbiotic, or haphazard, is much larger than has been appreciated.

## DATA ACCESSIBILITY

The bacterial 16S rRNA gene sequences from this study have been deposited in the Genbank numbers are MK177283–MK177456.

## AUTHOR CONTRIBUTIONS

DB, DQ, and JS designed the experiments. FB and JG mapped and collected the samples. DB, UB, and TD performed the experiments. DB, JRG, TH, DQ, and JS analyzed the results and wrote the manuscript. All authors approved the final manuscript.

## FUNDING

This material is based upon work supported by the National Science Foundation under Grants No. NSF IOS-1256416, NSF IOS-1656756, and the John Templeton Foundation grant 43667. The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

## ACKNOWLEDGMENTS

We thank the Queller/Strassmann lab group for useful feedback and discussion, and especially Alicia Hubert for final touches on bacteria sequencing and figures. Many thanks to William E. Winston for creating the GIS location map of our experimental site.

## SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00411/full#supplementary-material

#### REFERENCES


Brock, D. A., Read, S., Bozhchenko, A., Queller, D. C., and Strassmann, J. E. (2013). Social amoeba farmers carry defensive symbionts to protect and privatize their crops. Nat. Commun. 4:2385. doi: 10.1038/ncomms3385


genetic analysis of host-pathogen interactions. Infect. Immun. 68, 2939–2947. doi: 10.1128/IAI.68.5.2939-2947.2000


**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 Brock, Haselkorn, Garcia, Bashir, Douglas, Galloway, Brodie, Queller and Strassmann. 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.

# Environmental *Mycobacterium avium* subsp. *paratuberculosis* Hosted by Free-Living Amoebae

Ascel Samba-Louaka<sup>1</sup> , Etienne Robino<sup>1</sup> , Thierry Cochard<sup>2</sup> , Maxime Branger <sup>2</sup> , Vincent Delafont <sup>1</sup> , Willy Aucher <sup>1</sup> , Wilfrid Wambeke<sup>2</sup> , John P. Bannantine<sup>3</sup> , Franck Biet <sup>2</sup> \* and Yann Héchard<sup>1</sup> \*

<sup>1</sup> Université de Poitiers, Laboratoire Ecologie et Biologie des Interactions, UMR Centre National de la Recherche Scientifique 7267, Equipe Microbiologie de l'Eau, Poitiers, France, <sup>2</sup> Institut National de la Recherche Agronomique, Université de Tours, UMR1282, Infectiologie et Santé Publique, Nouzilly, France, <sup>3</sup> National Animal Disease Center, Agricultural Research Service, United States Department of Agriculture, Ames, IA, United States

Mycobacterium avium subsp. paratuberculosis is responsible for paratuberculosis in animals. This disease, leading to an inflammation of the gastrointestinal tract, has a high impact on animal health and an important economic burden. The environmental life cycle of M. avium subsp. paratuberculosis is poorly understood and several studies suggest that free-living amoebae (FLA) might be a potential environmental host. FLA are protozoa found in water and soil that are described as reservoirs of pathogenic and nonpathogenic bacteria in the environment. Indeed, bacteria able to survive within these amoebae would survive phagocytosis from immune cells. In this study, we assessed the in vitro interactions between several strains of M. avium subsp. paratuberculosis and Acanthamoeba castellanii. The results indicate that the bacteria were able to grow within the amoeba and that they can survive for several days within their host. To explore the presence of M. avium subsp. paratuberculosis in environmental amoebae, we sampled water from farms positive for paratuberculosis. A M. avium subsp. paratuberculosis strain was detected within an environmental amoeba identified as related to the poorly described Rosculus genus. The bacterial strain was genotyped, showing that it was similar to previous infectious strains isolated from cattle. In conclusion, we described that various M. avium subsp. paratuberculosis strains were able to grow within amoebae and that these bacteria could be found on farm within amoebae isolated from the cattle environment. It validates that infected amoebae might be a reservoir and vector for the transmission of M. avium subsp. paratuberculosis.

#### Keywords: paratuberculosis, amoebae, water, infection, *Mycobacterium*, *Rosculus*

## INTRODUCTION

Mycobacterium avium subsp. paratuberculosis (Map), the etiologic agent of paratuberculosis also called Johne's disease, induces extensive inflammation in the gastrointestinal tract of ruminants. This inflammation prevents absorption of nutrients, which leads to a chronic and progressive weight loss despite eating habits. The hallmark of advance stage disease in cattle is copious diarrhea, appearance of ribs and other skeletal bones. Paratuberculosis is responsible for economic losses up

#### *Edited by:*

Thierry Soldati, Université de Genève, Switzerland

#### *Reviewed by:*

Marina Santic', University of Rijeka, Croatia Otmane Lamrabet, Université de Genève, Switzerland Fabienne Misguich, Versailles Saint-Quentin-en-Yvelines University, France

*\*Correspondence:*

Franck Biet franck.biet@inra.fr Yann Héchard yann.hechard@univ-poitiers.fr

*Received:* 31 October 2017 *Accepted:* 23 January 2018 *Published:* 09 February 2018

#### *Citation:*

Samba-Louaka A, Robino E, Cochard T, Branger M, Delafont V, Aucher W, Wambeke W, Bannantine JP, Biet F and Héchard Y (2018) Environmental Mycobacterium avium subsp. paratuberculosis Hosted by Free-Living Amoebae. Front. Cell. Infect. Microbiol. 8:28. doi: 10.3389/fcimb.2018.00028 to \$1.5 billion annually to the dairy industries on the five continents (Ott et al., 1999; Sweeney et al., 2012). The management of this epidemic, is particularly difficult because efficient prophylactic tools are lacking. There is no treatment and the vaccine available is not very effective and not widely used because it compromises the diagnosis of bovine tuberculosis. Serological diagnostic tools lack specificity and are not adapted to early diagnosis of an infection, although this is beginning to change with the discovery of new antigens (Li et al., 2017). Another factor complicating the management of this epidemic involves the shedding of the bacteria in the environment where it is able to survive. Map is a genetically homogenous subspecies of M. avium, especially among bovine, human and wildlife isolates. However, two prominent lineages of Map have emerged following molecular strain typing and comparative genomic analysis—ovine (Map-S) and bovine (Map-C) strains (Biet et al., 2012). Initially, the Map-S and Map-C type strains were distinguished based on their molecular fingerprints using IS1311 polymorphism, MLSSR typing and hsp65 sequencing. However, beyond genetic typing, there are observed phenotypic differences including differences in their pigmentation, growth characteristics and lipopeptide structures in their cell walls (Bannantine et al., 2017). The environmental life cycle of Map is still poorly understood and at least one recent publication suggests that free-living amoebae (FLA) might be a potential environmental host (Salgado et al., 2015).

Free-living amoebae (FLA) are protozoa found in the same environmental niches as Map, including water and soil (Rodríguez-Zaragoza, 1994). FLA mainly feed on bacteria in the environment and digest them by phagocytosis. They thus play, along with other grazing protists, a key role in shaping the bacterial community composition in the environment (Jürgens and Matz, 2002). However, it has been repeatedly shown over the last decades that some bacteria may resist this digestion and persist or even grow within amoebae (Greub and Raoult, 2004). Importantly, these resistant bacteria are generally also more resistant to phagocytic immune cells, such as macrophages. This is well documented for Legionella pneumophila and for various Mycobacterium species in interaction with Acanthamoeba castellanii. The latter is the most common species in water and the best known FLA. As a consequence, FLA are considered as a training ground for pathogenic bacteria, including some Mycobacterium species (Molmeret et al., 2005; Salah et al., 2009). The fate of internalized mycobacteria can drastically vary according to the species, ranging from digestion (e.g., M. bovis BCG) and survival without replication (e.g., M. tuberculosis) to intra amoebal multiplication (e.g., M. abscessus or M. chelonae) (Drancourt, 2014). However, as described in this paper, most of the interactions were studied in vitro and only few studies have reported mycobacteria detected inside amoebae isolated from the environment (White et al., 2010; Amissah et al., 2014). Our group has previously described such interaction in drinking water (Delafont et al., 2014). Importantly, intra-amoebal growth of Mycobacterium would lead to improve its virulence (Cirillo et al., 1997).

The interaction between Map and amoebae has been poorly described. It was first reported in 2006 through a study showing in vitro that a strain of Map was ingested by Acanthamoeba and resisted digestion for at least 24 days (Whan et al., 2006). Four years later a short communication described the cooccurrence of Map and FLA in the soil (White et al., 2010). Amoebae and Map were found in the same soil samples and an amoeba strain was isolated harboring an intracellular acidfast stained bacterium. Mura et al. (2006) described in vitro that Map persisted for up to four years in presence of Acanthamoeba (Mura et al., 2006). Recently, Map were found within amoebae isolated from soil after application of cattle slurry spiked with Map (Salgado et al., 2015). However, these studies used a very limited range of strains and no study has demonstrated clearly the presence of Map in environmental FLA.

Our study was aimed to assess the infectious potential of several Map strains including the two main genetic lineages of Map, C and S types toward A. castellanii. We also investigated whether Map-infected FLA might be found in the environment surrounding infected cattle.

#### MATERIALS AND METHODS

#### Cultivation of Microorganisms

Bacterial strains used in this study are listed in **Table 1**. Mycobacterial strains were grown at 37◦C in Sauton medium or Middlebrook 7H9 broth (Difco Laboratories, Detroit, MI) with 0.2% glycerol and albumin-dextrose-catalase enrichment medium (ADC, Becton Dickinson, Le Pont de claix, France). M. avium subsp. paratuberculosis cultures were supplemented with 2 mg l−<sup>1</sup> mycobactin J (Allied Monitor). Bacteria were harvested at mid-log phase and kept frozen (−80◦C) in aliquots

TABLE 1 | Bacterial and amoebal strains.


<sup>a</sup>Profiles described by Thibault et al. (2007) and available on http://mac-inmv.tours.inra.fr/

until use. Recombinant M. avium subsp. paratuberculosis, strain K-10 expressing the Green fluorescent protein (GFP) - and luciferase described in Lefrançois et al. (2011) was grown in the Middlebrook 7H10 medium supplemented with ADC, 0.2 mg ml-1 mycobactin and 50 mg ml−<sup>1</sup> of hygromycin. Escherichia coli strain K12 was grown in LB liquid medium (Lysogeny Broth: 10 g/L NaCl, 5 g/L yeast extract, 10 g/L tryptone), and incubated for 24 h at 37◦C, under agitation (180 rpm).

The free-living amoeba A. castellanii ATCC 30010 was grown in PYG liquid medium (ATCC 712; 20 g/L proteose peptone, 1 g/L yeast extract, 1 g/L sodium citrate, 0.1 M D-glucose, 0.4 mM CaCl2, 4 mM MgSO4, 2.5 mM Na2HPO4, 2.5 mM KH2PO4, 50µM Fe(NH4)2(SO4)2, pH 6.5) in 25 or 75 cm<sup>2</sup> flask. A. castellanii cultures were incubated at 30◦C for 3–4 days, until a cell confluency of 80–90% was reached.

#### Isolation of Environmental Amoebae

Environmental water samples were collected in March 2016 from two farms located in "Indre et Loire" and reported to be contaminated with Map since more than 5 years. Two 1-L samples were taken for each herd from the drinking troughs. The samples were filtered through a 5µm nitrocellulose membrane (N3771-100EA, Sigma-Aldrich). The membranes were placed onto NNA plates (Non Nutrient Agar, 15 g/L) seeded with live E. coli strain K12. The plates were incubated at 30◦C and examined daily under phase-contrast microscopy. The presence of amoebal cells is characterized by the formation of a migration front, representing the amoebal movement upon agar plates, seeking for food. Amoebae were recovered by scrapping the agar plate at the migration front for collection or DNA extraction.

#### *A. castellanii* Infection by Different Genotypes of Map

A. castellanii cells from fully grown flasks were rinsed once using PAS buffer [Page's Amoeba Saline; 1 g/L sodium citrate, 0.4 mM CaCl2, 4 mM MgSO4, 2.5 mM Na2HPO4, 2.5 mM KH2PO4, 50µM Fe(NH4)2(SO4)2, pH 6.5], and resuspended in a mixture of PAS/PYG (1:1) supplemented with 0.005% triton X-100, to avoid mycobacteria clumping in the medium. Trophozoites cells were harvested and detached by vigorously tapping the flask. The cell suspension was adjusted to a concentration of 1.25 × 10<sup>5</sup> cells/mL, and 2 mL of the suspension distributed in each well of a 12 well plates. Mycobacteria cells were dispersed by 3 passages through a 18 Gauge needle, followed by a centrifugation step at 200 g for 5 min. The suspensions were subsequently quantified by optical density (OD) at 600 nm, using an equivalence of 6 × 10<sup>8</sup> mycobacteria per mL for 1 OD unit. Mycobacteria suspensions were used to infect A. castellanii at an MOI (multiplicity of infection) of 10, i.e., 10 bacteria per amoeba cell. The infected cultures were then centrifuged at 500 g for 10 min to facilitate the infection, and incubated for 1 h at 30◦C. The supernatant was then discarded and directly replaced with a fresh mixture of PAS/PYG (1:1), 0.005% triton X-100, supplemented with amikacin at 10µM. Timepoints at 1, 24, 48, and 72 h post infection at 30◦C were selected to estimate presence or growth of Mycobacteria by qPCR. To estimate the growth of A. castellanii, cells were harvested at different time points and counted using plastic counting slides FastRead 102 (Biosigma).

The localization of bacteria within amoeba was visualized using a recombinant rGFP-Map strain K-10 (**Table 1**) previously described (Lefrançois et al., 2011). Infected cultures were analyzed by fluorescence microscopy (BX41, Olympus).

#### Total Genomic DNA Extraction

Cells from infection experiments as well as from environmental amoebae isolated as described above, were used for DNA extraction. Amoeba cells (4 × 10<sup>4</sup> - 2 × 10<sup>6</sup> ) and associated mycobacteria were harvested and resuspended in 500 µL of PAS buffer. Cells were lysed by beadbeating in tubes containing 500 mg of small diameter glass beads (100µm) and 4 glass beads of 2 mm diameter (Sigma) using Fastprep apparatus for 30 s (speed 5 m/s). The suspension was then processed for DNA extraction using NucleoSpin Microbial DNA (Macherey-Nagel), following manufacturer recommendations for bacterial DNA extraction.

### Detection and Quantification of Mycobacteria by Quantitative PCR (qPCR)

The presence of bacterial genomic DNA was assessed by qPCR performed in a final volume of 10 µL, including 2 µL of extracted DNA, 2 µL of 5x Master Mix LightCycler <sup>R</sup> FastStart DNA Masterplus SYBR Green I (Roche), a set of primer pairs (**Table 2**), at a concentration of 0.5µM each, and water in sufficient quantity for 10 µL. Experiments were carried out using a LightCycler 1.5 thermocycler (Roche), consisting in an initial denaturation step at 95◦C for 10 min, 45 cycles of denaturation at 95◦C for 10 s, annealing for 10 s at a temperature according to the primer pairs used, and elongation at 72◦C for 20 s. Fusion curves were collected for each experiment. Quantification of DNA was performed by calculating fold changes using the 2 <sup>−</sup>1Ct method. 1Ct corresponds to the difference in cycle threshold (Ct) between a specific time point and the condition at 1 h.

#### TABLE 2 | Oligonucleotide primers.


### Multiple-Locus Variable Number Tandem Repeat Analysis (MLVA) Typing of Environmental Mycobacteria, Using MIRU-VNTR Markers

The method has been described previously (Thibault et al., 2007). The PCR mixture was composed as follows using the Go Taq Flexi DNA polymerase (Promega). Five microliters from DNA solution were added to a final volume of 25 µL containing 0.1 µL of Go Taq Flexi DNA polymerase (5 U/µL), 5 µl of betaine (Sigma), or 1 µl of dimethyl sulfoxide (Sigma); 0.2 mM (each) dATP, dCTP, dGTP, and dTTP (Promega); 5 µL of 5 × PCR buffer supplied by the manufacturer; 1µM of primers see (Thibault et al., 2007); and 1.5 mM of MgCl2. The reactions were carried out using a TC-520 thermal cycler (Techne). PCR conditions were as follows: 1 cycle of 5 min at 94◦C; 30 cycles of 30 s at 94◦C, 30 s at 55, 58, 60, 64◦C according to MIRU-VNTR searched, and 30 s at 72◦C; 1 cycle of 7 min at 72◦C. To detect differences in repeat numbers, the PCR products were analyzed by electrophoresis using 1.5% agarose gels (agarose electrophoresis grade; Invitrogen).

#### Amoebal Identification

The identification of the amoeba isolated from the water troughs was performed through sequencing of a 18S rRNA gene amplicon. From the total genomic DNA extract, a PCR was performed using the universal F566 and R1200 eukaryotic primers, targeting the 18S rRNA gene (Hadziavdic et al., 2014). The amplicon was sequenced and the sequence compared to the nucleotide nr database using BLASTn.

### Evaluation of Phagosomal Acidification

Mycobacteria suspension prepared as described previously were pre-labeled using pHrodoTM Red succinimidyl ester (ThermoFischer Scientific) following manufacturer recommendations, but excluding the methanol washing step. Briefly, bacterial suspensions were incubated for 1 h in 0.1 M sodium bicarbonate buffer containing 20µM pHrodoTM Red succinimidyl ester, in the dark. Bacteria were then pelleted by centrifugation at 12,000 g for 5 min, and washed twice in PAS buffer supplemented with 0.005% triton X-100. A. castellanii monolayers were infected by labeled mycobacteria at a MOI of 10. After 4 h of incubation, trophozoites were detached, fixed with 2% paraformaldehyde for 10 min in the dark. Cells were pelleted by centrifugation at 800 g for 10 min, and resuspended in a 30 µL of SlowFade Diamonds antifade mountant (ThermoFischer Scientific) with Hoechst at a concentration of 400 ng/mL. Samples were examined using an epifluorescence microscope (BX41, Olympus).

#### Statistics

Results were statistically analyzed using GraphPad Prism software. Multiple comparison was performed using Kruskal– Wallis analysis followed by a Dunn's post-test. Differences with p < 0.05 were considered statistically significant.

### RESULTS

## Map Strains Grow within *A. castellanii*

In order to see whether Map strains would behave differently within amoeba, we investigated infection of A. castellanii by different genotype and genetic lineage of Map strains. A. castellanii were co-cultured with the reference strain (K10), two cattle strains (205, 7912), and two sheep strains (397, 6796). The fate of Map internalized by A. castellanii was determined by a qPCR based bacterial quantification assay, at 1, 24, 48, and 72 h post infection. We found that all Map persisted and even grew within A. castellanii (**Figure 1A**). There were no significant differences in growth rate between the strains (p = 0.4441) even though the K10 strain tended to be slightly slower. The fate of infected A. castellanii was also monitored throughout the infection, indicating that the host proliferation was not significantly modified (p = 0.8939) by the presence of Map in comparison with an uninfected A. castellanii culture (**Figure 1B**). Taken together, these results suggest that Map can successfully infect and grow within A. castellanii without any deleterious effect on its host.

FIGURE 1 | A. castellanii is permissive to both ovine (S) and bovine (C) Map. (A) A. castellanii were infected with ovine (397, 6796) and bovine (K10, 7912, 205) strains of Map at a MOI of 10. Presence of Mycobacterium spp. was quantified (Mycobacterium genome unit) through qPCR amplification of the atpE gene, 1, 24, 48, and 72 h post-infection. Results represents the mean (±SEM) of four independent experiments (p = 0.4441). Results are normalized on the condition 1 h. (B) Uninfected and infected A. castellanii were counted 1, 24, 48, and 72 h post-infection. Results represents the mean (±SEM) of four independent experiments (p = 0.8939).

### Map Resides within Phagosomal Compartments

To localize Map within A. castellanii, amoebae were infected with a Map K10 strain expressing a green fluorescent protein (GFP). Microscopy analysis of infected A. castellanii showed colocalization between the Map K10-GFP and A. castellanii 7 days after infection (**Figure 2A**). This figure also suggests that the bacteria were found inside A. castellanii vesicles. Later observations allowed to visualize A. castellanii cysts but no GFP signal from the K10-GFP strain was detected, suggesting that in our conditions Map K10 was not present within the cysts. We have also assessed the acidification of Map-containing intracellular compartments in the first hours of infection. Before infection, the Map K10-GFP strain was stained with the pHrodo dye, which fluoresces in a pH-dependent manner. The results show that a fraction of intracellular Map was labeled in red, indicating that they were found within acidified vacuoles, likely phagosomes, while other Map were not labeled (**Figure 2B**). These observations suggest that Map are phagocytised by A. castellanii and resides, at least transiently, within acidified vacuoles.

#### Map Is Found within an Environmental Amoeba in Infected Farm

As no study had isolated environmental FLA infected by Map so far, we investigated the potential Map-amoeba association in water from the drinking troughs in herds naturally infected by Map. From the two farms enrolled in this study, only one sample was positive for the presence of cultivable amoebae. Total genomic DNA extracted from the amoebal culture was used to

acidification of bacteria through GFP and red signals.

detect the presence of Map DNA using a Map-specific qPCR assay, targeting the IS900 and F57 genes (**Table 3**). The successful amplification of the specific genomic target highlighted the presence of Map in association with the isolated amoebal culture. The Map strains K10, 397, and 7912 were used as positive control and M. llatzerenze as negative control for all PCR-based assays (**Table 3**). We have previously established a database containing isolated and genotyped Map strains from feces of infected cattle (Holbert et al., 2015). Therefore, we genotyped the Map-positive DNA from the amoeba culture, using MLVA genotyping and compare it to our database. The result of the MLVA genotyping based on the eight classical loci MIRU VNTR (Thibault et al., 2007) gave a profile INMV 2 identical to that of the 453 strain isolated from feces of cattle (**Table 4**).

Furthermore, in order to identify the amoeba isolate, we tried to subculture the strain but no subculture was observed. Thus, we performed a PCR targeting a portion of 18S rRNA coding gene. Sequencing of purified amplicon (1265 nucleotides, GenBank MG859939), indicated that the most closely related sequence to the isolate was Rosculus ithacus, a cercozoan amoeba, although the sequence identity was moderate (78%). Therefore, this amoebal isolate is likely to represent a newly described genus among the sainouroid clade.

#### DISCUSSION

A growing number of studies underlines the important role of amoebae in the carrying of pathogenic and non-pathogenic


<sup>a</sup>AtpE primer set is specific of Mycobacterium spp. IS900 and F57 primer sets are specific of Map.

mycobacteria, as well as in providing an environmental source of mycobacterial infection. On a microbiological level, the question of on farm disease transmission is particularly relevant because of the characteristics of Map biology. Indeed the disease transmission happens through ingestion of Map then during the disease progression, animals excrete bacilli in their environment, ranging from a few and discontinuous bacterial excretions in asymptomatic phases to very high levels and continuously shedding at clinical stage of the disease (Magombedze et al., 2017). This means that even in the silent phases of the disease the animals can excrete Map in the environment and in particular in the water points of the herds. In this context, it was interesting to investigate whether Map can be hosted and survive in amoeba and if water points of infected herds could be a source of contamination of Map mediated by FLA. Some other pathogenic mycobacteria were able to survive in soil and this environment might be a source of infection (Ghodbane et al., 2014).

In this study, the five selected strains (C- and S-type) not only survived but also showed a growth trend within A. castellanii for at least 72 h. In previous studies the growth of Map was not so high in the first days post-infection. This difference could be due to the Acanthamoeba strains and the infection protocols that were used in each study (Mura et al., 2006; Whan et al., 2006). For example, the co-cultures were centrifuged to increase efficiency and better synchronize the infection process. These differences might also explain the higher levels of bacteria per amoeba we observed. Moreover, we demonstrated that Map can persist for at least 7 days within A. castellanii. Within Map there is little genetic variability nevertheless the evolutionary pattern deciphered by Turenne et al. (2006) and Alexander et al. (2009) showed that the Map sub-species evolved from a common ancestor M. avium subsp. hominissuis into two different genetic lineages associated with a host preference. We speak of a lineage of S-type for sheep and C-type for cattle. In this study the strains were selected for their difference in terms of genotype and to include representative S and C-type. Altogether these results suggest that amoeba may serve as a significant phagocytic model to further investigate comparable or distinct phenotypic traits obtained by each lineage S and C-type of Map.

We also observed that the growth of A. castellanii was not adversely affected by infection with any of the Map strains. This indicates that Map had no deleterious effects


<sup>a</sup>MLVA analysis and profile identification according to web database application http://mac-inmv.tours.inra.fr/

TABLE 4 | MLVA genotyping of Map strains.

on amoebal viability. These results suggest that interactions between these two types of microorganisms could very likely happen in the environment. This stability could indeed favor Map resilience in the environment, providing a shelter against harsh conditions, as well as a potential substrate source for mycobacterial multiplication (Caire-Brändli et al., 2014; Barisch et al., 2015; Delafont et al., 2017).

Microscopy studies indicate that Map was localized within phagosomes and that some bacteria were found in acidified environment while others were not. This suggests that Map could block, to a certain extent, the normal acidification process of A. castellanii phagosomes. Inhibition of phagosomal acidification has been observed with other Mycobacterium species both in amoebae and macrophages (Kuehnel et al., 2001; Rohde et al., 2007; Delafont et al., 2017). Although much of the cell biology has been done on mycobacteria-macrophage interactions, further investigations would be required to better understand the fate of Map after its internalization in amoeba.

Amoeba-Map association in the environment has been poorly documented. To our knowledge, only one short communication has reported that amoeba isolated from soil were positive for Map-specific PCR but neither the Map nor the amoebae strains were characterized (White et al., 2010). In our study, two water troughs were selected for sampling, as those were providing the water source to distinct herds previously declared as infected by Map. We were able to recover cultivable FLA from one of the water troughs, confirming their presence in this environment. From the cultivated amoeba, the presence of Map DNA was detected using two specific primer sets by qPCR. This finding represents a strong argument in favor of interactions between amoebae and Map in environmental conditions. MLVA genotyping confirmed the identification of the species Map. The MLVA profile of the DNA was identical to that of an isolated strain from feces of a Map naturally infected cattle in the same farm (Holbert et al., 2015). These results not only confirm the presence of FLA harboring Map in the environment of animals on farm, but the identical genotypes recovered may also suggest a link between residence in amoeba and ability to infect cattle. Further investigations could be done to isolate the Map strains and improve the discrimination of genotype by combination of genotyping methods to ascertain the strains isolated from amoebae and animals are strictly identical. These results indicate that the strains might persist and circulate in the herd. They further suggest that fresh drinking water sources on farm may

#### REFERENCES


be important for healthy cows and that the problem of Johne's disease may be difficult to solve using herd management practices alone.

In addition, the identification of the amoebal isolate indicated that its closest sequence corresponds to R. ithacus. The latter is a coprophilic amoeba belonging to the sainouroid clade, as part of the Cercozoan radiation. Interestingly, protozoa from this clade were also identified in cow and sheep dung in a previous study (Bass et al., 2016). It is therefore not surprising to have identified such an amoeba from a drinking through providing cattle. However, we were not able to subculture this amoebal strain for some reason even after several try. It suggests that the culture method that was used is not fully efficient to subculture any FLA.

In conclusion, our study showed that various Map strains were able to persist and grow within A. castellanii, without observable deleterious effects on the host. A screening of environment surrounding Map-infected herds allowed to recover a free-living amoeba, representative of the sainouroid clade, from a drinking trough. This environmental amoeba harbored a Map strain that was genotyped. This study is bringing additional pieces of evidence reinforcing the role of amoebae in the persistence, and potential transmission, of pathogenic mycobacteria, including Map. This work should stimulate further studies focused on the characterization of environmental interactions between amoebae and pathogenic mycobacteria.

#### AUTHOR CONTRIBUTIONS

All the authors have substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; drafting the work or revising it critically for important intellectual content and final approval of the version to be published.

#### ACKNOWLEDGMENTS

We would like to thank Manon Duforestel for her technical help and Laure Malherbe Duluc, GDS veterinarian responsible for the epidemiological surveillance of infected herds. YH laboratory is partly financed by the European Union and the region of Nouvelle Aquitaine through the "Habisan program" (CPER-FEDER).

prediction to exact structure of a nonribosomal peptide. Mol. Microbiol. 105, 525–539. doi: 10.1111/mmi.13717


contribute to a large diversity of faecal-associated protists. Environ. Microbiol. 18, 1604–1619. doi: 10.1111/1462-2920.13235


**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 Samba-Louaka, Robino, Cochard, Branger, Delafont, Aucher, Wambeke, Bannantine, Biet and Héchard. 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.

# Evolution of Bordetellae from Environmental Microbes to Human Respiratory Pathogens: Amoebae as a Missing Link

#### Dawn L. Taylor-Mulneix\* † , Illiassou Hamidou Soumana† , Bodo Linz † and Eric T. Harvill

Department of Infectious Diseases, Center for Vaccines and Immunology, College of Veterinary Medicine, University of Georgia, Athens, GA, United States

The genus Bordetella comprises several bacterial species that colonize the respiratory tract of mammals. It includes B. pertussis, a human-restricted pathogen that is the causative agent of Whooping Cough. In contrast, the closely related species B. bronchiseptica colonizes a broad range of animals as well as immunocompromised humans. Recent metagenomic studies have identified known and novel bordetellae isolated from different environmental sources, providing a new perspective on their natural history. Using phylogenetic analysis, we have shown that human and animal pathogenic bordetellae have most likely evolved from ancestors that originated from soil and water. Our recent study found that B. bronchiseptica can evade amoebic predation and utilize Dictyostelium discoideum as an expansion and transmission vector, which suggests that the evolutionary pressure to evade the amoebic predator enabled the rise of bordetellae as respiratory pathogens. Interactions with amoeba may represent the starting point for bacterial adaptation to eukaryotic cells. However, as bacteria evolve and adapt to a novel host, they can become specialized and restricted to a specific host. B. pertussis is known to colonize and cause infection only in humans, and this specialization to a closed human-to-human lifecycle has involved genome reduction and the loss of ability to utilize amoeba as an environmental reservoir. The discoveries from studying the interaction of Bordetella species with amoeba will elicit a better understanding of the evolutionary history of these and other important human pathogens.

Keywords: Bordetella, Amoeba, Dictyostelium discoideum, environmental microbes, respiratory pathogens

## INTRODUCTION

The genus Bordetella comprises several bacterial species, which are pathogenic to animals and humans. The most clinically relevant species is B. pertussis, the causative agent of Pertussis disease, or Whooping Cough. This acute respiratory disease, known for the characteristic symptoms of paroxysmal cough, whooping, and post-tussive vomiting, is particularly serious and sometimes fatal in infants and elderly people. From a veterinary perspective B. bronchiseptica and B. avium are important. B. bronchiseptica, a respiratory pathogen of a wide range of mammals, causes a variety of pathologies ranging from chronic and often asymptomatic infection to more severe and acute

#### Edited by:

Thierry Soldati, Université de Genève, Switzerland

#### Reviewed by:

Joan Strassmann, Washington University in St. Louis, United States Falk Hillmann, Leibniz-Institut für Naturstoff-Forschung und Infektionsbiologie, Hans Knöll Institut, Germany

\*Correspondence:

Dawn L. Taylor-Mulneix dlt4ylor@gmail.com

† These authors have contributed equally to this work.

Received: 13 October 2017 Accepted: 27 November 2017 Published: 11 December 2017

#### Citation:

Taylor-Mulneix DL, Hamidou Soumana I, Linz B and Harvill ET (2017) Evolution of Bordetellae from Environmental Microbes to Human Respiratory Pathogens: Amoebae as a Missing Link. Front. Cell. Infect. Microbiol. 7:510. doi: 10.3389/fcimb.2017.00510 diseases, including bronchopneumonia and atrophic rhinitis in pigs, bronchitis in cats, snuffles in rabbits, and acute tracheobronchitis ("Kennel Cough") in dogs (Goodnow, 1980; Mattoo and Cherry, 2005). B. avium infects the respiratory tract of wild and domesticated birds, particularly turkeys, causing a respiratory disease with the symptoms known as bordetellosis or coryza in turkey chicks (Panigrahy et al., 1981; Kersters et al., 1984; Raffel et al., 2002). Recently, we identified that the bordetellae likely originated evolutionarily from soil and water environments and showed that the animal pathogen, B. bronchiseptica, can utilize amoeba, such as Dictyostelium discoideum, as environmental reservoirs and transmission vectors. We hypothesize that evolving the ability to evade amoebic predation and utilize amoebae as an environmental niche allowed bordetellae the transition from survival in soil and water to being respiratory pathogens.

#### ENVIRONMENTAL ORIGIN OF BORDETELLA

Bacteria of the genus Bordetella are known as colonizers of human and animal respiratory tracts. B. pertussis, B. bronchiseptica, and B. parapertussis form an evolutionary monophyletic group commonly referred to as "classical" bordetellae. Over the past decades, more distantly related Bordetella species, the "non-classical" species, have been described that colonize the respiratory tract and cause disease in birds (B. avium and B. hinzii), mice (B. pseudohinzii) and humans (B. holmesii, B. bronchialis, B. flabilis, and B. sputigena) (Kersters et al., 1984; Vandamme et al., 1995, 2015; Weyant et al., 1995; Ivanov et al., 2015, 2016), or were isolated from infected wounds of immunocompromised patients (B. trematum and B. ansorpii) (Vandamme et al., 1996; Ko et al., 2005). Although species associated with humans and animals have attracted the most attention, recent studies have revealed Bordetellalike bacteria in the environment. B. petrii, the first Bordetella species identified from a non-animal source, was initially isolated from an anaerobic bioreactor culture enriched by river sediment (von Wintzingerode et al., 2001), and has been subsequently found in marine sponges (Sfanos et al., 2005) and grass root consortia (Chowdhury et al., 2007). B. petrii has also been isolated from immunocompromised patients with ear infections and with pulmonary disease (Fry et al., 2005; Biederman et al., 2015; Nagata et al., 2015), suggesting that this species could also be an opportunistic pathogen in humans and animals, as has been observed for Pseudomonas spp. (de Bentzmann and Plesiat, 2011). In a recent study, three novel environmental Bordetella species were described, B. muralis, B. tumulicola, and B. tumbae, that have been isolated from a plaster wall surface of 1,300-year-old mural paintings in Japan (Tazato et al., 2015). With the vastly improved ability to study bacterial communities inhabiting diverse environmental niches, including soil, rocks, water, air, ice, plants, and animals (Schuster, 2008; Blottiere et al., 2013; Cowan et al., 2015; Anantharaman et al., 2016; Gionfriddo et al., 2016; Makhalanyane et al., 2016), and the ability to identify sequences from Bordetella-like bacteria in metagenomics datasets, we expect to see an increase in the number of described environmental Bordetella species.

We recently identified numerous 16S rRNA sequences of bacteria isolated from various environmental sources, including soil, water, plants and sediment, that displayed high sequence similarity with the 16S rRNA gene sequence of Bordetella (Hamidou Soumana et al., 2017). Based on 48 full-length 16S RNA sequences of strains recovered from these environments, we performed a phylogenetic analysis to determine the relatedness between human- and animal-associated Bordetella species and those isolated from environmental samples. The phylogenetic tree provided evidence for an environmental origin of bordetellae, as sequences from environmental samples possessed a significantly higher genetic diversity than those from human- and animal-associated samples. Sequences from environmental samples were present in all 10 sequence clades, including sequence clades at the root of the phylogenetic tree. In contrast, sequences from animal-associated species were found in only four sequence clades at the top of the tree. Together, the order of branching events within the phylogenetic tree suggested that Bordetella species, including human-restricted pathogens, arose from environmental ancestors (Hamidou Soumana et al., 2017). The evolution and adaptation to human and animal hosts most likely occurred after acquisition of virulence factors that enabled them to respond to new hosts.

The ability of micro-organisms to adapt to different environments and hosts requires a high genomic plasticity coupled with the capacity to sense and respond to environmental changes (Aujoulat et al., 2012). Under laboratory conditions, classical bordetellae such as B. bronchiseptica can respond to different environmental stimuli by switching between two distinct life styles. When cultured at 37◦C, which mimics mammalian host temperatures, expression of genes associated with virulence and colonization in mammalian hosts is upregulated. However, B. bronchiseptica adopts a second life style when cultured at 25◦C or lower, during which expression of virulence-associated genes is down-regulated while expression of a large, alternative set of genes, such as those involved in motility and growth in dilute nutrients, is up-regulated. Expression of the latter genes has been hypothesized to be important under extra-host growth conditions or in an environmental niche (Taylor-Mulneix et al., 2017). This global gene regulation is under the control of the two-component system BvgAS, consisting of a sensor protein, BvgS, a transcriptional activator, BvgA, and a transcriptional repressor, BvgR (**Figure 1**). Upon phosphorylation by BvgS, BvgA binds to the promoter regions of the Bvg-activated genes and activates transcription. The last gene, bvgR, is responsible for the regulation of the Bvg-repressed genes (Merkel et al., 1998). Virulence-associated factors are expressed in the Bvg positive (Bvg+) phase, and the alternative set of genes in the Bvg negative (Bvg−) phase (**Figure 1**). The ability to switch between life styles seems to be conserved amongst bordetellae as bvgA and bvgS gene homologs have been found in the genomes of animal-associated species as well as the environmental B. petrii (Gerlach et al., 2004; Gross et al., 2010; Linz et al., 2016). In addition, B. petrii, B. bronchiseptica, and B. hinzii, all known to associate with mammalian hosts, were shown to grow efficiently in soil at 25◦C (Hamidou Soumana et al., 2017). These observations indicate that even though B. bronchiseptica has adapted to mammals, it has conserved the ability to survive under environment conditions, and respond to changes such as temperature fluctuations (Coote, 2001).

#### AMOEBA AS AN ENVIRONMENTAL RESERVOIR OF BORDETELLAE

Recently, we have reported the novel ability of B. bronchiseptica strain RB50 to utilize amoebae as an environmental niche and transmission vector (Taylor-Mulneix et al., 2017). B. bronchiseptica was chosen for this study because it is believed to resemble the progenitor of the classical bordetellae which are most highly associated with human disease (Diavatopoulos et al., 2005). While B. pertussis and B. parapertussis have undergone genome reduction in their adaption to human, B. bronchiseptica retains the largest genome, the widest range of animal hosts, and characteristics like motility and nutrient scavenging that are associated with the ability to survive in some environmental reservoir. Indeed, our data found that B. bronchiseptica establishes a commensal relationship with amoeba. While, as of yet, there is no data indicating whether B. bronchiseptica results in any benefit or harm to the amoeba, there is clear evidence that it benefits from its interactions with the amoeba Dictyostelium discoideum (Taylor-Mulneix et al., 2017).

Amoeba predation involves chemotaxis toward their bacterial food source, internalization, phagocytosis, killing, and metabolizing bacteria. B. bronchiseptica is amongst the few bacteria that can survive intracellularly within amoeba for extended periods of time (Hägele et al., 2000; Abd et al., 2003; Greub and Raoult, 2004). This intracellular survival was observed both in D. discoideum and in Acanthamoeba castellanii (Taylor-Mulneix et al., 2017). Furthermore, we showed that B. bronchiseptica can localize to the amoeba sori and associate with the amoeba through multiple passages while utilizing another bacterial food source. These data were quite striking as association with the amoeba was maintained through multiple passages despite the observation that B. bronchiseptica localized to the sori are not intracellular within the D. discoideum spores. However, B. bronchiseptica within the amoeba sori up-regulate expression of factors associated with cell adherence. These data suggest that bordetellae have evolved mechanisms that allow them to survive in long-term association with amoeba, including surviving intracellularly within these phagocytic cells. These mechanisms may also allow them to evade phagocytosis and killing by mammalian immune cells (Mattoo et al., 2000).

As the BvgAS two-component system appears to regulate genes associated with the different life styles of bordetellae (Bvg<sup>+</sup> in animal hosts and Bvg<sup>−</sup> in a putative environmental niche), we hypothesized that the Bvg<sup>−</sup> phase may be important during the bacterial interaction with amoebae. Indeed, B. bronchiseptica that were mutated to constantly express Bvg<sup>+</sup> associated virulence factors had a significantly lower recovery from the amoeba sori in comparison to wildtype or a Bvg<sup>−</sup> mutant. These observations indicated an advantage for the Bvg<sup>−</sup> phase and represents the

FIGURE 1 | The BvgAS phosphorelay. (A) The master regulatory system of bordetellae, Bordetella Virulence Genes (BVG), is expressed by bvgS and bvgA. (B,C) BvgS is a transmembrane sensor protein consisting of a periplasmid domain (PP) connected to the histidine phosphotransfer domains (HP) in the cytosol through a linker domain (L). (B) BvgS is inactive and un-phosphorylated when bacteria grow at temperatures below 25◦C. (D) Bvg<sup>−</sup> phase genes are transcribed when the BvgAS system is inactive. (C) Upon receiving inducing signals such as 37◦C, BvgS autophosphorylates and initiates a phosphor-relay that leads to phosphorylation and activation of BvgA. (E) When the BvgAS system is active, Bvg<sup>+</sup> phase-associated genes are transcribed, including bvgR. BvgR represses expression of Bvg<sup>−</sup> phase associated genes.

first report of a role for the Bvg<sup>−</sup> phase in association with any host. Therefore, the presence of the Bvg<sup>−</sup> phase across the Bordetella genus may be a clue to this genus' origin and evolution (Taylor-Mulneix et al., 2017).

#### AMOEBA AS A "TRAINING GROUND" FOR HUMAN PATHOGENS

The evolution of Bordetella species from environmental microbes to animal and human pathogens suggests the existence of an intermediate stage, or commensal host, that has facilitated the adaptation. Free-living amoeba such as Acanthamoeba spp. have been shown to serve as an environmental niche for several opportunistic bacterial pathogens such as Burkholderia spp., Pseudomonas aeruginosa, Listeria monocytogenes, Legionella pneumophila, and Mycobacterium (Cirillo et al., 1997; Hägele et al., 2000; Greub and Raoult, 2004; Drancourt, 2014; DiSalvo et al., 2015; Jose Maschio et al., 2015). L. pneumophila is an example of a bacterial pathogen that utilizes a similar strategy of invasion and life style inside both amoeba and macrophages (Hägele et al., 2000; Greub and Raoult, 2004). Our research has now added B. bronchiseptica to the list of bacteria that can survive amoebic predation and utilize amoeba as an environmental reservoir. Interestingly, the ability of B. bronchiseptica to survive internalization and resist digestion by eukaryotic cells is not limited to amoebae. Upon infection of a mammalian host, B. bronchiseptica and B. pertussis can survive inside macrophages, thereby enabling the bacteria to evade host immunity (Siciliano et al., 2006; Lamberti et al., 2010, 2013). The strategy to hide inside host immune cells such as macrophages provides a means of persistence in the host and facilitates the spread to other tissues and organs.

Single-celled amoeba and macrophages are fairly similar physiologically, therefore, these observations suggest that adaptation of environmental bacteria to amoeba may have served as an intermediate step to become animal- or humanassociated pathogens (Molmeret et al., 2005). Similar to other species, the long-term bacteria-amoeba association may be the key that has allowed bordetellae to evolve the ability to associate with vertebrate hosts.

#### GENOME REDUCTION RESULTS IN FAILURE TO EMPLOY AMOEBA AS A HOST

B. pertussis and B. parapertussis evolved independently from a B. bronchiseptica-like ancestor (Parkhill et al., 2003; Diavatopoulos et al., 2005; Park et al., 2012; Linz et al., 2016). These species possess ≥98% nucleotide identity on the genome level and share many important virulence-associated factors, including toxins and adhesins, suggesting a relatively recent divergence. Indeed, a global phylogeny based on whole genome sequences of a worldwide collection of B. pertussis isolates showed two deep branches that coalesce to a last common ancestor about 2,300 (range 1,428–3,340) years ago. A subsequent expansion and diversification of one of the branches, that contains over 98% of all isolates, occurred about 500 years ago (Bart et al., 2014), which correlates with the first historic descriptions of Whooping Cough disease in Persia (Aslanabadi et al., 2015) and Europe (Bart et al., 2014). The extremely low sequence diversity in B. pertussis (Linz et al., 2016) indicates that the speciation and specialization to the human host was associated with a strong genetic bottleneck that drastically reduced the sequence diversity. It appears that this process was accompanied by the acquisition and massive expansion of insertion sequence elements (ISE), particularly of IS481. ISE's were inserted at hundreds of genomic locations, and homologous recombination between these identical DNA repeats resulted in a mosaic-like structure of the chromosome, in which short blocks of perfect collinearity are broken up by nearly 150 individual rearrangements, including chromosomal inversions. In fact, 88% of the genomic rearrangements in the genome of B. pertussis strain Tohama\_I, in comparison to the genome of B. bronchiseptica strain RB50, are bordered by IS elements, mostly IS481 (Parkhill et al., 2003). Recombination between ISE's also caused a large amount of deletion in the chromosome, which is reflected in the much smaller genome of B. pertussis (about 4.1 Mb with 3,816 CDS's) compared to that of B. bronchiseptica (about 5.3 Mb with 5,007 CDS's). B. pertussis lost over 1,000 genes during the evolution from a B. bronchiseptica-like ancestor, most of which encode transcriptional regulators, proteins involved in transport and metabolism of a wide range of compounds, as well as proteins of unknown function (Linz et al., 2016).

Similar to B. pertussis, B. parapertussis has also undergone host specialization; one lineage causes pertussis-like disease in humans (hereafter referred to as Bpphu), while the other causes pneumonia in sheep (Bppov). Both B. parapertussis lineages have undergone genome reduction via ISE's, but much less drastic than B. pertussis, resulting in genome sizes of 4.7 Mb in Bpphu strain 12822 and of 4.8 Mb in Bppov strain Bpp5 (Parkhill et al., 2003; Park et al., 2012). Genome comparisons of Bpphu strain 12822 and of Bppov strain Bpp5 to B. bronchiseptica strain RB50 revealed a large degree of collinearity between the genomes, with a limited number of IS1001-flanked genomic breakpoints. However, the evolution of both B. parapertussis lineages proceeded independently as evidenced by different genomic rearrangements and different gene content (Parkhill et al., 2003; Park et al., 2012; Linz et al., 2016). As a result, 3,592 genes are shared between the B. parapertussis genomes while 829 genes are specific to the genome of Bpphu strain 12822 and 592 are specific to the genome of Bppov strain Bpp5.

Another human-restricted pathogen, B. holmesii, which has been co-isolated with B. pertussis during outbreaks of Whooping Cough, also appears to have undergone substantial genome reduction. The genome of this emerging non-classical species (3.7 Mb) is about 1.2 Mb smaller than that of its closest relative B. hinzii (4.9 Mb). The B. holmesii genome contains multiple copies of three different ISE's (IS407, IS481, and ISL3). Similar to B. pertussis, many of the putative genomic breakpoints that disrupt the genome synteny are flanked by ISE's, suggesting genome reduction and rearrangements through homologous recombination between ISE's (Linz et al., 2016). Thus, specialization of several Bordetella species to an exclusively human host was associated with acquisition and expansion

of different classes of ISE's, and subsequent recombination between those perfect DNA repeats resulted in chromosomal rearrangements and extensive genome reduction in each of those lineages.

During the assessment of the ability of classical bordetellae to utilize D. discoideum, an interesting dichotomy has arisen. The ability to survive intracellularly and localize to the amoeba sori is conserved amongst B. bronchiseptica, but is absent in B. pertussis (manuscript in preparation). These data suggested that as B. pertussis adapted to the closed-cycle of using only humans as a host it underwent genome reduction and lost the ability to utilize D. discoideum as an environmental reservoir (**Figure 2**). In support of this theory, the Bppov strain Bpp5 retained the ability to survive intracellularly and localize to the sori, while the human isolated Bpphu strain 12822 did not (manuscript in preparation). The observed difference between the two B. parapertussis lineages may be associated with a substantially different gene content in their genomes. All in all, the data suggest that while the ancestor-like B. bronchiseptica strains can utilize the amoeba as an environmental reservoir, genome reduction associated with host specialization has led to the inability of human-adapted bordetellae to evade amoebic predation (**Figure 2**). While the ability of the human-associated B. holmesii to employ the amoebic life cycle remains to be elucidated, we predict that its severe genome reduction likely involved loss of genes necessary for survival in and utilization of amoebae. Thus, during the evolution of bordetellae from soil microbes to human pathogens, the interaction with and role of amoeba is an important step that may be overlooked if one were to only compare B. pertussis with soil isolated bordetellae. All in all, it remains important to consider the whole bordetellae genus and the ability of each individual species to interact with amoeba. The association of genome data with the ecology of amoebic-bacterial interaction will provide important clues that will ultimately reveal key steps in the evolution of Bordetella from environmental microbes to human pathogens. Along this line, we have begun to investigate the interaction of multiple Bordetella species with the social amoeba D. discoideum to expand this fascinating field of research beyond the classical bordetellae.

### CONCLUSION

The human immune system is so complex that it is able to efficiently resist invasion and/or resolve infection with little damage to the host. Yet, many human pathogens have arisen from progenitors found in soil and water environments. Therefore, it is important to understand the evolutionary pressures, which allow bacteria adapted to success in other environments to emerge as important human pathogens that cause major disease burden. Herein, we describe that the interaction of bordetellae with predatory amoeba in the environment may have favored the evolution of tools that prepared them for their interactions with mammalian phagocytes, contributing to their emergence as important human pathogens. As these bacteria established a chain of transmission in mammals and further adapted to these hosts, their specialization, and genome reduction, may come with a cost of losing the environmental reservoir. Further work in studying these and other human pathogens and their interaction with amoebae will be important for understanding the mechanisms by which pathogens can evolve and develop mechanisms to evade host immune systems.

#### AUTHOR CONTRIBUTIONS

DT-M, IH, and BL contributed equally to this manuscript in regards to concept, writing, and editing. These authors should all

#### REFERENCES


be considered equally contributing first authors and we wish that the publication would reflect this. EH contributed to editing and obtaining funding.

### FUNDING

NIH (grant numbers: AI116186, GM113681, AI107016) awarded to EH.


of Dictyostelium discoideum as an amplifying transmission vector. PLoS Biol. 15:e2000420. doi: 10.1371/journal.pbio.2000420


**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 © 2017 Taylor-Mulneix, Hamidou Soumana, Linz and Harvill. 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) or licensor 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.

# *Acanthamoeba* and *Dictyostelium* as Cellular Models for *Legionella* Infection

A. Leoni Swart <sup>1</sup> , Christopher F. Harrison<sup>2</sup> , Ludwig Eichinger <sup>3</sup> \*, Michael Steinert <sup>4</sup> \* and Hubert Hilbi <sup>1</sup> \*

1 Institute of Medical Microbiology, Medical Faculty, University of Zurich, Zurich, Switzerland, <sup>2</sup> Max von Pettenkofer Institute, Medical Faculty, Ludwig-Maximilians University Munich, Munich, Germany, <sup>3</sup> Institute for Biochemistry I, Medical Faculty, University Hospital Cologne, Cologne, Germany, <sup>4</sup> Department of Life Sciences, Institute of Microbiology, Technical University of Braunschweig, Braunschweig, Germany

#### *Edited by:*

Matthew S. Francis, Umeå University, Sweden

#### *Reviewed by:*

Hayley J. Newton, University of Melbourne, Australia Eric D. Cambronne, Oregon Health and Science University, United States Jason King, University of Sheffield, United Kingdom

#### *\*Correspondence:*

Ludwig Eichinger ludwig.eichinger@uni-koeln.de Michael Steinert steinert@tu-bs.de Hubert Hilbi hilbi@imm.uzh.ch

*Received:* 23 November 2017 *Accepted:* 13 February 2018 *Published:* 02 March 2018

#### *Citation:*

Swart AL, Harrison CF, Eichinger L, Steinert M and Hilbi H (2018) Acanthamoeba and Dictyostelium as Cellular Models for Legionella Infection. Front. Cell. Infect. Microbiol. 8:61. doi: 10.3389/fcimb.2018.00061 Environmental bacteria of the genus Legionella naturally parasitize free-living amoebae. Upon inhalation of bacteria-laden aerosols, the opportunistic pathogens grow intracellularly in alveolar macrophages and can cause a life-threatening pneumonia termed Legionnaires' disease. Intracellular replication in amoebae and macrophages takes place in a unique membrane-bound compartment, the Legionella-containing vacuole (LCV). LCV formation requires the bacterial Icm/Dot type IV secretion system, which translocates literally hundreds of "effector" proteins into host cells, where they modulate crucial cellular processes for the pathogen's benefit. The mechanism of LCV formation appears to be evolutionarily conserved, and therefore, amoebae are not only ecologically significant niches for Legionella spp., but also useful cellular models for eukaryotic phagocytes. In particular, Acanthamoeba castellanii and Dictyostelium discoideum emerged over the last years as versatile and powerful models. Using genetic, biochemical and cell biological approaches, molecular interactions between amoebae and Legionella pneumophila have recently been investigated in detail with a focus on the role of phosphoinositide lipids, small and large GTPases, autophagy components and the retromer complex, as well as on bacterial effectors targeting these host factors.

Keywords: amoebae, effector protein, GTPase, host-pathogen interaction, pathogen vacuole, phosphoinositide lipid, retrograde transport, type IV secretion

#### *LEGIONELLA* SPP.—ENVIRONMENTAL BACTERIA AND OPPORTUNISTIC PATHOGENS

Legionella spp. are the causative agents of a potentially fatal pneumonia termed Legionnaires' disease. Overall, Legionella pneumophila is the clinically most relevant species, followed by Legionella longbeachae, which is associated with outbreaks of Legionnaires' disease particularly in Australia and New Zealand (Newton et al., 2010). The Gram-negative bacteria are ubiquitously found in the environment, where they inhabit natural and manmade freshwater systems. Upon inhalation of Legionella-laden aerosols, the opportunistic pathogenic bacteria enter the human lung, where they infect alveolar macrophages and might cause a fulminant pneumonia (**Figure 1**). Legionella spp. are usually regarded to only accidentally infect humans after transmission by technical vectors such as cooling towers, fountains, or showers (Benin et al., 2002; Newton et al., 2010; Hilbi et al., 2011; Yamaguchi et al., 2017). However, recently the first probable transmission of a highly virulent L. pneumophila strain from person-to-person was reported (Correia et al., 2016).

L. pneumophila replicates intracellularly within a unique membrane compartment, which is called the Legionellacontaining vacuole (LCV). This compartment avoids acidification and fusion with lysosomes, but extensively communicates with multiple vesicle trafficking pathways, including the endosomal, secretory and retrograde route, and after a series of maturation steps associates with the endoplasmic reticulum (ER) (Isberg et al., 2009; Asrat et al., 2014; Finsel and Hilbi, 2015; Personnic et al., 2016; Bärlocher et al., 2017b). LCV formation is mechanistically similar in amoebae and macrophages and requires the Icm/Dot (Intracellular multiplication/Defective organelle transport) type IV secretion system (T4SS) (Segal et al., 1998; Vogel et al., 1998; Segal and Shuman, 1999; Hägele et al., 2000; Solomon et al., 2000). This T4SS of L. pneumophila translocates ∼300 different "effector" proteins into host cells, wherein these virulence factors subvert pivotal processes such as signal transduction, cytoskeleton dynamics, and membrane trafficking (Isberg et al., 2009; Hubber and Roy, 2010; Hilbi and Haas, 2012; Haneburger and Hilbi, 2013; Sherwood and Roy, 2013; Finsel and Hilbi, 2015; Qiu and Luo, 2017). The Icm/Dot T4SS (but not the "effectorome") is conserved in the genus Legionella, including L. longbeachae, where it is also essential for virulence in amoebae and mice (Cazalet et al., 2004, 2010; Chien et al., 2004). Of note, the bioinformatics analysis of the Legionella genome sequences revealed a number of genes encoding eukaryotic-like proteins or motifs, which likely subvert host cell functions in order to allow the pathogen to survive and replicate intracellularly (Cazalet et al., 2004, 2010; Chien et al., 2004).

#### PROTOZOA AS ENVIRONMENTAL NICHES OF *LEGIONELLA* SPP.

Environmental microorganisms commonly colonize and form communities called biofilms, which comprise large numbers of prokaryotic and eukaryotic cells associating via adhesion molecules and secreted compounds. Biofilms are ubiquitously found in the environment and contain a plethora of bacterial species, many of which communicate via low molecular weight signaling molecules in a process termed quorum sensing (Hall-Stoodley et al., 2004; Hochstrasser and Hilbi, 2017). These biofilms are attacked ("grazed") by predatory amoebae and ciliates, resulting in a reduction in the bacterial population. In contrast to most other bacterial genera, Legionella spp. resist degradation by amoebae and other protozoa (Newton et al., 2010; Hilbi et al., 2011).

In its aquatic habitats L. pneumophila replicates intracellularly in various free-living protozoa, including amoebae such as Acanthamoeba, Vermamoeba (formerly Hartmannella), Naegleria, Echinamoeba, and Vahlkampfia, as well as in ciliates like Tetrahymena (Fields, 1996; Greub and Raoult, 2004; Hilbi et al., 2011; Hsu et al., 2015). From an evolutionary point of view, interactions of L. pneumophila with protozoa shaped the relationship and likely led to adaptive responses, which enable the pathogen to also infect mammalian phagocytes such as human alveolar macrophages (Greub and Raoult, 2004; Molmeret et al., 2005; Jäger et al., 2014). In addition to their role as natural reservoir and selection niche of virulence traits, protozoa can also enhance the transmission of L. pneumophila either as intact host cell or as expelled vesicles filled with bacteria (Rowbotham, 1986; Brieland et al., 1997; Amaro et al., 2015).

A number of studies indicated that L. pneumophila is highly adapted and manages to infect a wide range of protozoan and metazoan hosts (Hilbi et al., 2011; Bergmann and Steinert, 2015; Hsu et al., 2015). This suggests that many protozoan host species and other interaction partners of L. pneumophila in the environment remain to be discovered. Recently, we reported that L. pneumophila, protozoa and aquatic nematodes thrive in close association within biofilms (Rasch et al., 2016). Microscopic inspection of biofilms and inoculation experiments with mCherry-labeled L. pneumophila identified the ciliates Oxytricha bifaria, Stylonychia mytilus, and Ciliophrya sp. as potential new protozoan hosts of L. pneumophila. While several metazoan organisms were identified in the biofilm (Copepoda, Nauplius larvae, Rotifera, and nematodes), only the nematodes accumulated L. pneumophila within their intestines. This correlates with the finding that the nematode Caenorhabditis elegans can be infected with L. pneumophila under laboratory conditions (Brassinga et al., 2010; Komura et al., 2010).

To study the Legionella-protozoa-nematode interactions, we infected axenically grown Acanthamoeba castellanii and C. elegans with L. pneumophila. This model system indicated that nematode larvae rupture the infected amoebae cells and thus are exposed to L. pneumophila (Rasch et al., 2016). The interaction may be different for Legionellafilled amoebae cysts and spores. D. discoideum spores resist destruction by the pharyngeal grinders of nematodes such as C. elegans and seem to be disseminated by the nematode (Kessin et al., 1996). Thus, if infected with Legionella, cysts and spores might disseminate the pathogen along this route.

L. pneumophila grown in amoebae shows distinct features compared to bacteria grown in broth (**Figure 1**). Accordingly, upon growth in Acanthamoeba polyphaga the bacteria are more tolerant to antibiotics (Barker et al., 1995), while grown in A. castellanii the bacteria are more virulent compared to growth on agar plates (Cirillo et al., 1994). Furthermore, the amoeba species within which replication occurs seems to also play a role,

**Abbreviations:** Icm/Dot, intracellular multiplication/defective organelle trafficking; LCV, Legionella-containing vacuole; OCRL, oculocerebrorenal syndrome of Lowe; PI, phosphoinositide; T4SS, type IV secretion system; TGN, trans-Golgi network.

such that resistance to chlorination is higher in bacteria released from Hartmanella vermiformis than from A. castellanii (Chang et al., 2009).

### EVOLUTION AND GENE REGULATION OF *LEGIONELLA* IN AMOEBAE

Legionella spp. are known to undergo horizontal gene transfer (Miyamoto et al., 2003). Gene transfer between bacteria has been shown to be significantly enhanced within the digestive vacuole of protozoa, presenting the possibility that they may act as a "trading post" to allow the acquisition of virulence genes (Schlimme et al., 1997). Perhaps reflecting this process, L. pneumophila contains a number of genes believed to be derived from eukaryotic, bacterial, or even viral sources (Lurie-Weinberger et al., 2010). These findings suggest that amoebae can be considered as an evolutionary niche, wherein the selective pressure and the potential for Legionella to acquire new virulence-related genes is higher than in the extracellular environment (Greub and Raoult, 2004; Molmeret et al., 2005).

An analysis of the L. pneumophila transcriptome under different growth conditions revealed that expression of almost half the genome is altered upon shifting from replicative to transmissive phase (Brüggemann et al., 2006). This was observed for bacteria grown in either broth cultures or in A. castellanii. The transmissive phase, which occurs following completion of intracellular replication, is characterized by upregulation of virulence and invasion genes, including substrates of the Icm/Dot T4SS and motility genes (Brüggemann et al., 2006). The gene regulation of L. pneumophila is further complicated by the existence of non-coding (nc) RNAs, which have been shown to play a role in regulation of virulence traits (Romby et al., 2006). L. pneumophila contains a number of these regulatory ncRNAs, which likely are implicated in virulence, since their expression changes during the bacterial biphasic life cycle or upon infection of A. castellanii (Weissenmayer et al., 2011).

### *ACANTHAMOEBA* AS A NATURAL AND MODEL HOST OF *LEGIONELLA*

The distinct features of the protozoan genera Acanthamoeba and Dictyostelium have been particularly useful in examining the ecology and cellular host-interactions of L. pneumophila. Given the similarity of the infection process in amoebae and macrophages, the amoebae are powerful models to study bacteria-macrophage interactions (Hilbi et al., 2007; Escoll et al., 2013; Hoffmann et al., 2013; Bergmann and Steinert, 2015). Acanthamoeba has often been found in Legionella-positive habitats and has an extremely wide distribution. The amoebae have been isolated from diverse aquatic environments, including soil (Sawyer, 1989), roadside puddles (Sakamoto et al., 2009), fresh water lakes and rivers (John and Howard, 1995), frozen lakes (Brown and Cursons, 1977), the atmosphere (Rodriguez-Zaragoza et al., 1993), and even the Antarctic (Brown et al., 1982).

The majority of Acanthamoeba isolates from these aquatic environments harbor endosymbionts, such as viruses, bacteria, yeast, and protists (Greub and Raoult, 2004). However, for laboratory studies of host-pathogen interactions Acanthamoeba strains adapted to axenic growth are mostly used (Shevchuk et al., 2014; Eisenreich and Heuner, 2016). Like many other free-living amoebae, Acanthamoeba adopts a bi-phasic life cycle, comprising a vegetative trophozoite stage and a dormant cyst stage, which at least in part might explain the wide environmental range of these protozoa (Chávez-Munguía et al., 2005). L. pneumophila utilizes the trophozoites for replication and the double-walled cysts as shelter to escape harsh environmental conditions (Fields, 1996).

Acanthamoeba replicates by binary fission and, under culturing conditions, has a doubling time reported to range from 8 to 18 h (Berk and Garduño, 2013). The amoebae are highly motile and follow a random walk pattern with a movement speed of 0.1–0.2 µm/s. In the presence of chemical signals produced by bacteria, Acanthamoeba will use chemotaxis to "hunt," actively moving toward the source of these signals (Preston and King, 1984; Schuster and Levandowsky, 1996). Moreover, Acanthamoeba spp. are characterized by contractile vacuoles, which expel water for osmotic regulation (Bowers and Korn, 1973), and they contain glycogen storage vacuoles, lysosomes and digestive vacuoles, wherein L. pneumophila seems to replicate (Ulsamer et al., 1971; Bowers and Korn, 1973; Eisenreich and Heuner, 2016). Actin microfilaments adjacent to the plasma membrane of A. castellanii are responsible for forming protrusions (Pollard et al., 1970), and trophozoites produce short, spine-like projections at the edge of the cell, known as acanthapodia. As the amoebae are very motile, these projections are short-lived but constantly replaced, normally collapsing after less than a minute during active movement (Preston and King, 1984). A. castellanii also employs a diverse repertoire of putative pattern recognition receptors (PRRs), many of which with postulated orthologous functions in the innate immune systems of higher organisms (Clarke et al., 2013).

The genome of A. castellanii is polyploid and harbors ∼15,500 compact intron-rich genes, a number of which are predicted to have been acquired through inter-kingdom lateral gene transfer (Clarke et al., 2013). The genomic complexity of A. castellanii remains one of the barriers to its broad utilization as a model organism. Transfection has been predominantly unsuccessful, with only one report of success (Peng et al., 2005). Furthermore, the polyploid genome organization prevents simple knockout of genes of interest, although approaches using RNA interference to knock down gene expression have been successfully used (Lorenzo-Morales et al., 2005). Overall, however, the known sequence of the A. castellanii genome and the increasing number of available molecular tools will strongly augment ecological and cellular research with this authentic host of L. pneumophila.

### *DICTYOSTELIUM DISCOIDEUM*: A VERSATILE CELLULAR MODEL FOR *LEGIONELLA* INFECTION

In contrast to Acanthamoeba spp., Dictyostelium discoideum is not a prevalent natural host for L. pneumophila in the environment. This social amoeba, which naturally lives in forest soil and has a narrow temperature tolerance of 20– 25◦C, undergoes a complex developmental program from a single cell to become a multicellular organism (Taylor et al., 2009; Bretschneider et al., 2016). D. discoideum replicates by mitotic division of single amoebae. These amoeba cells live of bacteria, which are taken up by phagocytosis. Laboratory strains, e.g., Ax2, can also thrive in liquid medium through macropinocytosis. Exhaustion of the nutrient supply triggers an intricate multicellular cooperativity. In response to cAMP signals, solitary cells aggregate by chemotaxis, and ∼100,000 cells form a migrating slug, which responds thermo-tactically and photo-tactically. At the final stage of morphogenesis a fruiting body is formed, consisting of a basal disc, a stalk, and spores. Accordingly, the genome of D. discoideum encodes morphogenetic traits such as cell-type determination, spatial patterning, cell death, and other features that are essential in multicellular but not unicellular organisms (Steinert and Heuner, 2005; Steinert, 2011; Schilde et al., 2016). The genome of D. discoideum is about 34 Mb and comprises six chromosomes, which range from 3.5 to 8.6 Mb in size. Most genes contain introns with conserved splice junctions. About 100 times smaller than the human genome, the D. discoideum genome encodes ∼12,500 predicted proteins, including a large number of mammalian orthologs (Eichinger et al., 2005). In addition, D. discoideum is increasingly used as a model for the study of genes that in mutant form cause disease in humans (Müller-Taubenberger et al., 2013).

The completed genome sequence of D. discoideum, a remarkable repertoire of molecular genetic tools, and the intrinsic biological features of the amoebae allow to study many fundamental cellular processes (Bergmann and Steinert, 2015; Bretschneider et al., 2016; Hochstrasser and Hilbi,

2017). In the haploid amoeba D. discoideum non-essential genes can be easily disrupted or replaced by homologous recombination. Further genetic manipulation strategies include random insertion mutagenesis (restriction-enzyme-mediated integration, REMI), multiple gene deletions, Cre/LoxP-mediated recombination, RNA interference techniques, and ectopic expression of endogenous or foreign genes (Kuspa and Loomis, 1992; Chen et al., 1994; Faix et al., 2004; Kuhlmann et al., 2006; Al-Quadan and Kwaik, 2011). The large mutant collection of the Dictyostelium stock center (http://www.dictybase.org) (Gaudet et al., 2011), easy cultivation, amenability to diverse biochemical and biological approaches, such as in vivo expression of fluorescence-tagged proteins, also contributed to the present strength and versatility of the D. discoideum model. Recent highlights resulting from using this model host of L. pneumophila will be discussed below.

#### INTRACELLULAR REPLICATION AND COMPETITION OF *L. PNEUMOPHILA* IN AMOEBAE

The amoeba plate test is a relatively simple assay to determine the ability of L. pneumophila mutants to form colonies in the presence of A. castellanii (Albers et al., 2005) (**Figure 2A**). A suspension of A. castellanii or media alone is spread onto charcoal yeast extract (CYE) agar plates, and allowed to dry. Serial dilutions of L. pneumophila are spotted onto the plates and incubated at 30 or 37◦C for 3 days. This method allows comparing the growth of multiple strains on a single plate, in both the presence and absence of amoebae. In this manner, L. pneumophila mutants defective for amoebae resistance and intracellular replication can be determined in a semi-quantitative manner, alongside bacterial strains capable of suppressing these phenotypes (Albers et al., 2005). A variant of this method has been employed to screen clonal libraries of >20,000 L. pneumophila mutants for virulence defects (Aurass et al., 2009). In this approach, mutagenized bacteria are used to infect A. castellanii prior to plating.

While the amoeba plate test allows screening of strong replication phenotypes, the resulting output is not easily quantified. A more detailed analysis can be performed by measuring the number of bacteria present within or released by amoebae at given time points by determining bacterial colony forming units (CFU). A. castellanii or D. discoideum is infected with L. pneumophila in Ac medium or MB medium, respectively. In these media L. pneumophila is unable to replicate, which allows detection of intracellular replication exclusively. The host cells are then lysed by shear forces or mild detergent treatment, followed by plating on CYE agar to determine CFU. This assay has been used to assess the infection and replication of L. pneumophila, L. dumoffii, and L. feelii in A. castellanii (Moffat and Tompkins, 1992). Upon uptake of L. pneumophila, bacterial CFU drop significantly in the first few hours. Thereafter, the pathogen replicates intracellularly, leading to a 100–1,000 fold increase in bacterial numbers, and at the end of the cycle, the bacteria escape and disseminate. Bacterial amoebae resistance and intracellular replication can also be applied to screening of water samples. In this approach, samples negative for Legionella by normal culture techniques are incubated with amoebae, allowing intracellular growth of the bacteria to easily detectable numbers (Sanden et al., 1992).

Faster and more sensitive methods to determine intracellular replication of L. pneumophila in amoebae utilize an automated microtiter plate reader. This approach allows to measure the increase in optical density at 600 nm (Coil et al., 2008) or bacteria-produced fluorescence, e.g., GFP (Harrison et al., 2013, 2015a,b) (**Figure 2B**). To monitor intracellular growth by fluorescence increase over time, A. castellanii or D. discoideum is infected at 30/37◦C or maximum at 25◦C, respectively, with either GFP- or mCherry-producing L. pneumophila strains. To this end, we preferentially use LoFlo medium, a commercially available, complex D. discoideum medium that exhibits low auto-fluorescence and does not allow extracellular growth of L. pneumophila. The fluorescent signal for each sample is measured by microtiter plate reader at various time points post infection (p.i.). Since the fluorescence intensity is proportional to the number of bacteria present in a sample, the quantification of fluorescence increase reflects the rate of intracellular bacterial replication. Using the CFU or fluorescence assays, a number of L. pneumophila regulators or effector proteins, including LqsR (Tiaden et al., 2007), RidL (Finsel et al., 2013b), LppA (Weber et al., 2014a), and LegG1 (Rothmeier et al., 2013), have been found to promote intracellular replication in amoebae.

In contrast, the deletion of some other regulators (LqsA), effectors (SidC, SidM), or unknown proteins (HdeD) did not result in an intracellular growth defect of L. pneumophila. This might be due to robust intracellular replication of L. pneumophila involving redundant effectors, or simply because a gene is not required under the conditions tested. To discover weaker phenotypes of L. pneumophila mutant strains, we established the more sensitive amoebae competition test (Kessler et al., 2013). In this assay, A. castellanii is co-infected with the L. pneumophila parental and deletion mutant strain at a 1:1 ratio (multiplicity of infection of 0.01 each) and grown at 37◦C for 15–21 days. Every third day the supernatant and lysed amoebae are diluted 1:1000, fresh amoebae are infected, and CFU are determined on agar plates containing kanamycin (to select for presence of the resistance cassette in the mutant strains) or not. Using this approach, L. pneumophila mutant strains lacking individual Lqs components (Kessler et al., 2013) or single effectors (LegG1, RidL, LppA, SidC) were outcompeted by the parental strain (Finsel et al., 2013b; Rothmeier et al., 2013; Dolinsky et al., 2014; Weber et al., 2014a). Thus, in the course of successive rounds of infection the fitter strain (usually the parental strain) will come to dominate the bacterial pool.

Using these assays, in addition to L. pneumophila mutant strains, the effect of host proteins on intracellular bacterial replication can also be assessed. Upon pharmacological inhibition of phosphoinositide (PI) 3-kinases (PI3Ks) by wortmannin or LY294002, or deletion of the kinases in D. discoideum, L. pneumophila replicated more efficiently, and therefore the PI3Ks restrict intracellular growth (Weber et al., 2006). In contrast, overproduction of the atlastin (Atl) homolog

barely (1icmS) grow. Growth defect of the mutants is complemented by plasmid-borne expression of the corresponding genes (1icmT+icmT, 1icmS+icmS). (B) Intracellular replication quantified by fluorescence from GFP-producing L. pneumophila. Virulent bacteria show an increase in fluorescence over time, while 1icmT mutant bacteria do not grow. (C) Fluorescence microscopy images of L. pneumophila-infected, dying amoebae that round up, losing the characteristic spiky morphology, and take up the fluorescent dye propidium iodide. (D) Live-cell confocal fluorescence microscopy of D. discoideum strain Ax3 producing P4C-mCherry (red) and GFP-Sey1 (green), infected (MOI 10, 2 h) with mCerulean-producing L. pneumophila JR32 (white). Scale bar: 2µm. (E) Flow cytometry of A. castellanii infected with L. pneumophila wild-type or 1icmT and stained with propidium iodide. Dye uptake and changes in morphology are quantified by fluorescence and light scattering, respectively. (F) Flow cytometry gating on infected A. castellanii reveals replication of GFP-producing intracellular L. pneumophila wild-type but not 1icmT mutant bacteria. (G) LCV isolation by immuno-magnetic separation and density gradient centrifugation from lysates of D. discoideum Ax3, infected (MOI 10, 2 h) with DsRed-producing L. pneumophila JR32 (red). The preparation was immuno-stained with an anti-calnexin antibody, followed by a FITC-coupled secondary antibody (green). Scale bar: 10µm, zoom: 2µm. Image (A) was reproduced with permission from Albers et al. (2005) and images (C,E,F) from Tiaden et al. (2007).

Sey1 in D. discoideum promoted intracellular growth, while production of the catalytically inactive, dominant negative mutant protein Sey1\_K154A or depletion by RNAi of Atl3 in mammalian cells inhibited intracellular replication of L. pneumophila (Steiner et al., 2017a).

### THE TRANSCRIPTIONAL RESPONSE OF *D. DISCOIDEUM* UPON INFECTION WITH *L. PNEUMOPHILA*

In several infection models, including D. discoideum and human macrophages, it has been shown that the quantitative and qualitative modulation of diverse host cell functions is crucial for efficient intracellular replication of L. pneumophila (Bozzaro and Eichinger, 2011; Hochstrasser and Hilbi, 2017). Accordingly, the analysis of the transcriptional changes that occur in D. discoideum upon infection with L. pneumophila revealed important aspects of the complex host-pathogen crosstalk (Farbrother et al., 2006). The experiment was performed with cDNA-microarrays, covering approximately half of the Dictyostelium genome, and led to the identification of 731 differentially regulated host genes in a 48 h infection with the virulent L. pneumophila Philadelphia-1 strain JR32. In addition, a detailed analysis of the 24 h time point p.i. of L. pneumophila JR32, an avirulent L. pneumophila 1dotA mutant, or Legionella hackeliae (reduced virulence), uncovered 131 differentially expressed genes common to all three strains compared with uninfected amoebae. The corresponding gene products are either involved in host-specific defense mechanisms or required for successful infection and proliferation of L. pneumophila (Farbrother et al., 2006).

Functional annotation of the differentially regulated genes of the 48 h and 24 h time points revealed that in addition to triggering a stress response, L. pneumophila not only interferes with intracellular vesicle trafficking, but also subverts and exploits the host metabolism. A considerable enrichment of differentially regulated genes involved in translation, proteolysis, nucleotide metabolism, and lipid modification was observed. Moreover, L. pneumophila induces the expression of multiple amino-acyl t-RNA synthetases (aaRSs). Interestingly, besides the canonical coupling of correct amino acids to t-RNA molecules, aaRSs were described to play an unexpected role in splicing, apoptosis, and regulation of transcription and translation (Hausmann and Ibba, 2008; Yannay-Cohen et al., 2009). Thus, the upregulation of D. discoideum aaRSs upon L. pneumophila infection may serve a similar function in different signaling pathways. The transcriptome results also showed that infection with L. pneumophila results in differential expression of host genes that are involved in bacterial degradation and autophagy. Additionally, genes encoding cytoskeleton proteins were enriched ∼7-fold during the early stage of infection. These findings likely reflect the uptake of bacteria and transport of the nascent phagosome inside the host cell (Farbrother et al., 2006).

In a different study it was found that infection of D. discoideum by L. pneumophila caused a decrease in mitochondrial messenger RNAs and a cleavage of the mitochondrial large subunit ribosomal RNA (LSU rRNA). The specific cleavage of the LSU rRNA in D. discoideum required functional L. pneumophila type II and type IV secretion systems. LSU rRNA cleavage was, however, not observed upon infection of A. castellanii or human U937 macrophages, suggesting that L. pneumophila might use distinct mechanisms to interrupt the mitochondrial metabolism in different hosts (Zhang and Kuspa, 2009).

In summary, the transcriptional changes in L. pneumophilainfected D. discoideum suggest complex and still poorly understood regulatory interactions between pathogen and host. In the course of establishing its replicative niche, L. pneumophila not only interferes with signaling processes, mitochondrial function, and intracellular vesicle trafficking, but also profoundly influences the metabolism of its host.

### ANALYSIS OF *LEGIONELLA* INFECTION IN DEFINED *D. DISCOIDEUM* MUTANT STRAINS

Investigations of D. discoideum mutant strains have significantly contributed to our understanding of Legionella infection. Host cell factors implicated in host-pathogen interactions are either involved in the cellular defense or exploited by the pathogen in the course of infection. Consequently, the inactivation of distinct host factors may either favor or impede Legionella infection. To identify host genes crucial for Legionella infection, candidate gene approaches as well as untargeted mutational screens have been carried out in D. discoideum (Bozzaro et al., 2008; Li et al., 2009). In the candidate gene approach, either available D. discoideum mutants are used, which in many cases can be ordered from the Dictyostelium stock center (Fey et al., 2013), or novel genes of interest are disrupted, silenced, or the corresponding proteins overproduced as fluorescently tagged fusions (Bozzaro et al., 2008; Steinert, 2011; Steiner et al., 2017b). In order to dissect the complex cross-talk between D. discoideum and L. pneumophila, classical infection assays were performed, and a versatile methodology has been developed based on fluorescently tagged marker proteins, heterologously produced bacterial effectors and different visualization techniques (Eichinger and Rivero, 2006; Steinert, 2011; Steiner et al., 2017b). As select examples, we will discuss the role of the host cytoskeleton, autophagy machinery, and phospholipid metabolism in the infection process.

Most D. discoideum knock-out mutants of cytoskeletal proteins caused a decrease in the uptake and in the proliferation of L. pneumophila, indicating that an intact cytoskeleton is important (Bozzaro et al., 2013). Exceptions to this rule were the coronin A and B genes. Coronins comprise a large family of proteins that function mainly in actin cytoskeletonassociated processes. The proteins harbor a WD (Trp-Asp) repeat domain containing seven repeats that form a sevenbladed β-propeller, adjacent to a unique region, and a Cterminal coiled coil region which mediates trimerization (Shina and Noegel, 2008). We studied uptake and replication of L. pneumophila in D. discoideum mutants where corA (encoding the conventional coronin),corB (encoding coronin 7 (CRN7)), or vilA (encoding villidin a member of the coronin 4 family) were disrupted. Absence of conventional coronin reduced the uptake of L. pneumophila and enhanced intracellular growth. Villidindeficient cells showed an even higher reduction in uptake, and intracellular growth was strongly reduced (Fajardo et al., 2004). In contrast, uptake of L. pneumophila was enhanced in CRN7 deficient cells but intracellular replication was not affected. Overexpression of CRN7 on the other hand caused reduced internalization and an increased replication of L. pneumophila, suggesting that CRN7 negatively regulates the internalization of the bacteria. A double mutant wherein the corA and corB genes were inactivated behaved like the corA deletion strain (Shina et al., 2010, 2011) (**Table 1**). CRN7 harbors a Cdc42- and Racinteractive binding (CRIB) domain in each of its two β-propeller domains with a preference for GDP-loaded Rac. Thus, CRN7 might keep Rac GTPases in their GDP-bound form and locally prevent the activation of the downstream targets SCAR and WASP. Loss of CRN7 hyper-activates WASP, which promotes Factin assembly leading to increased phagocytosis (Swaminathan et al., 2015).

In recent years, autophagy has been identified as an ancient and important cellular defense mechanism that targets intracellular pathogens for destruction (Sherwood and Roy, 2016). A number of pathogens are able to subvert this cellular defense, and also L. pneumophila interferes with the autophagy process of the host. In early studies with D. discoideum mutants deficient for either one of five different core autophagy genes autophagy appeared to be dispensable for intracellular replication of L. pneumophila at later stages of infection (Otto et al., 2004). However, infection assays with atg9, atg16, and atg9/16 deletion mutants revealed a reduced uptake of L. pneumophila in these strains in comparison to Ax2 wild-type cells (Tung et al., 2010; Xiong et al., 2015). For the atg9 deletion mutant it was further shown that those bacteria that entered the host were initially cleared less efficiently and at later stages of the infection multiplied more efficiently compared to the parental amoebae (Tung et al., 2010) (**Table 1**). It must be kept in mind that apart from its role in autophagy ATG9 likely has additional roles in trafficking, as it is localized to a number of cytoplasmic vesicular structures (Noda, 2017). Noteworthy, atg9 and atg16 were oppositely regulated in parental D. discoideum in response to infection (Farbrother et al., 2006; Tung et al., 2010).

The L. pneumophila Icm/Dot substrate RavZ is a cysteine protease that irreversibly cleaves lipid-conjugated Atg8 proteins on the membranes of nascent autophagic structures (Choy et al., 2012). Furthermore, the L. pneumophila sphingosine-1 phosphate lyase (LpSpl) targets host sphingolipid metabolism and inhibits autophagy in mouse macrophages (Rolando et al., 2016). These findings suggest that both, the host cell and the pathogen, modulate autophagy for their own benefit.

Another cellular process that is part of the defense reaction of the host and exploited by L. pneumophila is the PI lipid metabolism (Weber et al., 2009b; Haneburger and Hilbi, 2013). PI lipids are important for vesicle trafficking and organelle identity, since the cellular compartments are defined in part by their PI composition. The lipid head group can be phosphorylated or dephosphorylated at the 3′ , 4′ , and/or 5′ positions by specific PI kinases and PI phosphatases, resulting in an array of seven specific PIs (Di Paolo and De Camilli, 2006). Deletion or pharmacological inhibition of PI3Ks class I promotes intracellular replication of L. pneumophila and impairs the transition from tight to spacious LCVs (Weber et al., 2006). Disruption of D. discoideum PTEN (phosphatase and tensin homolog), a PI phosphatase antagonizing PI3Ks, reduces the uptake but does not affect proliferation of L. pneumophila (Peracino et al., 2010). D. discoideum PLC (phospholipase C) is involved in the PI metabolism through hydrolysis of PI(4,5)P<sup>2</sup> to diacylglycerol (DAG) and inositol 1,4,5-triphosphate (IP3). Its inhibition dramatically reduced the engulfment of L. pneumophila, but had no effect on bacterial replication (Peracino et al., 2010).

Furthermore, in D. discoideum cells lacking the PI 5-phosphatase Dd5P4, a homolog of human OCRL (oculocerebrorenal syndrome of Lowe), bacterial replication and LCV formation occurred more efficiently. Dd5P4 localizes to LCVs via its N-terminus and might interact with the L. pneumophila PI(3)P-binding virulence factor LpnE (Weber et al., 2009a). The PI 5-phosphatase is catalytically active on the LCV, thereby increasing the PI(4)P available for binding of the L. pneumophila effector proteins SidC or SidM. An additional host protein, which plays a role in the PI metabolism is RpkA, an unusual seven-helix trans-membrane receptor with a predicted intracellular PI(4)P 5-kinase activity. RpkA is specific for lower eukaryotes and is recruited to nascent LCVs in D. discoideum. It disturbs the PI balance and thereby plays a role in the defense against L. pneumophila. RpkA interacts with the V-ATPase complex, but whether it can actively recruit the V-ATPase to the LCV is not known (Riyahi et al., 2011) (**Table 1**).

These examples illustrate the power of the D. discoideum system for the investigation of the host side during infection. The biological properties of this amoeba in combination with the straight-forward generation and analysis of mutant strains will continue to provide a potent and versatile model for the dissection of the complex cross-talk with L. pneumophila.

#### FLUORESCENCE MICROSCOPY OF *LEGIONELLA*-INFECTED LABELED *D. DISCOIDEUM*

The difficulty of performing genetic manipulations in A. castellanii implies that modifications of host cell factors, for example the production of GFP-labeled proteins, are not feasible. One of the few fluorescence assays that can be performed is the determination of Legionella-triggered cytotoxicity. To this end, the number of infected A. castellanii cells permeable to fluorescent dyes such as propidium iodide is counted, as has been done to assess the cytotoxicity of L. pneumophila deletion mutants (Albers et al., 2007) (**Figure 2C**). Given these constraints, genetically tractable amoebae such as D. discoideum, are more useful for cell biology and pathogen-host interaction studies, which examine the localization of specific markers to the LCV.

A plethora of genetic tools is available to study hostpathogen interactions using D. discoideum. In addition to DNA microarrays and targeted deletions or random mutations of



L. pneumophila uptake and intracellular growth in D. discoideum mutants was assessed by different methods. Effects ("up" or "down") on bacterial uptake or intracellular growth are indicated relative to the parental amoeba strain or untreated amoebae, respectively. ATG, autophagy; PLC, phospholipase C; PI3K, phosphoinositide 3-kinase class I; PTEN, phosphatase and tensin homolog; Dd5P4, D. discoideum inositol 5-phosphatase 4; OCRL, oculocerebrorenal syndrome of Lowe; RpkA, receptor phosphatidylinositol kinase A; DN, dominant negative; WT, wild-type; n.t.: not tested.

genes, an extensive library of plasmids is available allowing constitutive or inducible production of fluorescently labeled probes in D. discoideum (http://wiki.dictybase.org/dictywiki/ index.php/Vectors). Specifically, various expression vectors for N- or C-terminal fusions with green- or red-fluorescent proteins are available. These include extrachromosomal (Levi et al., 2000), extrachromosomal inducible expression (Veltman and van Haastert, 2013), or integrating plasmids (Manstein et al., 1995), as well as vectors for efficient generation of gene knockouts using one-step cloning (Wiegand et al., 2011). For the production of fluorescent probes, D. discoideum is transformed with one or multiple plasmids by electroporation (Weber et al., 2014b). Transformants are isolated by antibiotic selection and form microcolonies within 7–12 days after transformation.

A GFP-fusion protein commonly used in D. discoideum is calnexin-GFP, an ER-specific type I transmembrane protein and well-established LCV marker (Fajardo et al., 2004; Weber et al., 2006; Ragaz et al., 2008; Dolinsky et al., 2014). Within 2 h after uptake, calnexin accumulates on the membrane of LCVs containing virulent L. pneumophila, but not avirulent 1icmT mutant bacteria, and to a significantly smaller extent on vacuoles harboring 1sidC mutants (Ragaz et al., 2008; Dolinsky et al., 2014; Weber et al., 2014b). Using D. discoideum cells producing fluorescently labeled proteins, numerous other host proteins were found to localize to the LCV. These include small GTPases of the Rab family (Urwyler et al., 2009; Hoffmann et al., 2014), Rap1 (Schmölders et al., 2017), and Ran (Rothmeier et al., 2013), the ER tubule-resident large GTPase Atl3/Sey1 (Steiner et al., 2017a) (**Figure 2D**), as well as components implicated in the retrograde vesicle trafficking pathway, such as the D. discoideum OCRL homolog Dd5P4 (Weber et al., 2009a) and subunits of the retromer complex (Finsel et al., 2013a).

A caveat of microscopy methods using fixed samples is that the fixation techniques might alter the morphology of cellular structures such as membranes, vesicles, or microtubules. Timelapse imaging of living cells on the other hand is optimal to follow cellular dynamics over time. A well-studied process is the PI conversion on the LCV membrane during infection (Weber and Hilbi, 2014; Weber et al., 2014b). L. pneumophilainfected D. discoideum producing various PI probes, including PHCRAC-GFP for PI(3,4,5)P<sup>3</sup> and PI(3,4)P2, 2×FYVE-GFP for PI(3)P or GFP-P4CSidC for PI(4)P, were imaged live for time-lapse analysis (Weber et al., 2014b). Immediately upon bacterial uptake, the phagosome containing L. pneumophila transiently accumulates PI(3,4,5)P<sup>3</sup> for up to 1 min, and within that time-span, PI(3)P is acquired in a Icm/Dot independent manner. During the following 2 h, LCVs containing virulent L. pneumophila gradually replace PI(3)P with PI(4)P, whereas 1icmT vacuoles remain PI(3)P positive (Weber et al., 2014b). Using dually labeled D. discoideum, the P4C-mRFPmars signal was spatially separated from calnexin-GFP-positive membranes and maintained for at least 8 h. Manipulation of the LCV PI pattern is of great importance for the bacterium. The PI 5 phosphatase Dd5P4 as well as PI3Ks and PI4KIIIβ have been implicated in this process (Brombacher et al., 2009; Weber et al., 2009a,b).

Live-cell imaging was also employed to assess the LCV and microtubule dynamics in infected D. discoideum producing calnexin-GFP (Rothmeier et al., 2013) or GFP-α-tubulin (Simon et al., 2014). The D. discoideum strains were infected with DsRed-producing L. pneumophila, wild-type, 1legG1, or the complemented mutant. LegG1 is an Icm/Dot-translocated RCC1-like-repeat effector protein, which activates the small GTPase Ran and increases cellular RanGTP levels (Rothmeier et al., 2013). Ran is a member of the Ras superfamily of small GTPases and controls various cellular processes, including nucleo-cytoplasmic transport (Stewart, 2007), assembly of the mitotic spindle apparatus (Goodman and Zheng, 2006; Clarke and Zhang, 2008), and the dynamics of non-centrosomal microtubules (Yudin and Fainzilber, 2009). LCV motility as well as microtubule stability was significantly reduced in D. discoideum infected with 1legG1 mutant bacteria.

### QUANTITATIVE ASSESSMENT OF *LEGIONELLA* INFECTION BY (IMAGING) FLOW CYTOMETRY

Flow cytometry is a high-throughput technique that allows a simultaneous multi-parameter analysis of single cells. Large data sets, high resolution, accuracy, reproducibility, and sensitivity are major advantages of this approach to study uptake, cytotoxicity, and intracellular replication of L. pneumophila (Tiaden et al., 2013). The physiological status and integrity of a cell is reflected in the forward scatter/sideward scatter pattern (corresponding to cell size and "granularity," respectively), and membrane integrity can be tested by staining cellular DNA with the fluorescent dye propidium iodide, which is recorded as "cell-associated fluorescence."

Flow cytometry can be used to study in a quantitative manner Acanthamoeba phagocytosis. To this end, the rate and degree of uptake of fluorescent beads are assessed, i.e., the ratio of fluorescent vs. non-fluorescent amoebae (Avery et al., 1995), or, analogously, the infection by fluorescently labeled L. pneumophila is determined (Tiaden et al., 2013). Uptake of L. pneumophila is scored immediately after infection by determining an "uptake index," which is calculated from the signal-strength and percentage of amoebae infected with GFPproducing bacteria (Harf et al., 1997; Tiaden et al., 2013). Hence, cytotoxicity of L. pneumophila can be assessed as a change in size and shape of A. castellanii, measured by forward/sideward light scattering, as well as by loss of membrane integrity, as determined by the increase in permeability to propidium iodide (Tiaden et al., 2007, 2008) (**Figure 2E**).

Moreover, using flow cytometry and GFP-producing bacteria, not only the uptake efficiency can be determined, but also intracellular replication in A. castellanii or D. discoideum can be followed over several days. This approach has been utilized to examine the uptake and replication defects of L. pneumophila deletion mutants lacking components of the Legionella quorum sensing (Lqs) system: LqsR (Tiaden et al., 2007), LqsA or LqsS (Tiaden et al., 2010), LqsT or LqsS-LqsT (Kessler et al., 2013), as well as the entire lqs gene cluster (Tiaden et al., 2008) (**Figure 2F**). Analogously, flow cytometry can be employed to assess uptake and replication phenotypes of D. discoideum deletion mutant strains lacking PI3Ks (Weber et al., 2006) or Dd5P4 (Weber et al., 2009a). Finally, in addition to its use as a research tool, flow cytometry has also been developed as a diagnostic method to rapidly screen for Legionella contamination of bathing facilities and other water sources (Taguri et al., 2011).

In a technique termed imaging flow cytometry (IFC), flow cytometry is combined with fluorescence microscopy, enabling fast quantification of microscopic images. Using this approach, the high-throughput capacity and information about cell size, volume, and shape provided by the flow cytometer is combined with high-resolution spatial localization of fluorescent proteins of interest provided by a microscopic image, resulting in nonbiased and quantitative, large datasets (Barteneva et al., 2012; Johansson et al., 2015). By means of specialized software, the acquired cellular images can be analyzed using bi-variant plots and histograms, in parallel with a sequential gating strategy.

Recently, we employed IFC and D. discoideum ectopically producing GFP- and mCherry-labeled fusion proteins to assess the accumulation of specific host proteins on LCVs in whole cells (Schmölders et al., 2017; Steiner et al., 2017a) or homogenates (Bärlocher et al., 2017a). Thus, active small GTPase Rap1 was found to localize to a higher extent to LCVs containing the parental L. pneumophila strain Lp02 as compared to LCVs harboring the "pentuple" mutant strain (Schmölders et al., 2017), which lacks five gene clusters encoding 31% of the effector proteins (O'Connor et al., 2011). The method was also utilized to quantify the recruitment of calnexin-GFP or GFP-Sey1 to PI(4)P-positive LCVs labeled with P4C-mCherry. Dually labeled D. discoideum amoebae producing P4C-mCherry and GFP-Sey1 or GFP-Sey1\_K154A were infected, and colocalization of PI(4)P with L. pneumophila was analyzed for 10,000 events per time point p.i. IFC confirmed and quantified the fluorescence microscopy finding that PI(4)P accumulates at the LCV in a Sey1-independent manner. Finally, the Rab7 GAP TBC1D5 was scored on LCVs. However, in this case lysates of infected D. discoideum had to be prepared to reduce background and increase sensitivity (Bärlocher et al., 2017a). To this end, D. discoideum amoebae producing in tandem GFP-TBC1D5 and the endosomal marker AmtA-mCherry were infected with mPlum-producing L. pneumophila, and the localization of TBC1D5 was analyzed by IFC in cell homogenates. For each condition, over 3,000 cells were analyzed, and TBC1D5 was found to localize to wild-type-, but not 1icmT-containing LCVs.

### CHEMOTACTIC MIGRATION OF *LEGIONELLA*-INFECTED *D. DISCOIDEUM*

In a process termed chemotaxis, phagocytes migrate in a directed manner toward an attractant. Thus, D. discoideum, macrophages or neutrophils sense and respond to cAMP, the chemokines CCL5, and tumor necrosis factor (TNF)-α, or the formyl-methionyl-leucyl-phenylalanine (fMLP) peptide, respectively. The migration of eukaryotic cells critically depends on microtubule polarization and dynamics (Etienne-Manneville, 2013), which in turn is controlled by the small GTPase Ran.

Given the impact of the Icm/Dot substrate LegG1 on Ran GTPase activation and microtubule dynamics (Rothmeier et al., 2013), we analyzed the role of the L. pneumophila T4SS and LegG1 on host cell motility (Simon et al., 2014). Using D. discoideum amoebae or mammalian immune cells in underagarose or Boyden chamber migration assays, L. pneumophila was found to inhibit cell migration in an Icm/Dot-dependent manner. D. discoideum infected with virulent L. pneumophila showed a substantially reduced migration when compared to amoebae infected with 1icmT mutant bacteria. The migration inhibition observed was not due to an uptake defect or cytotoxicity. Interestingly, D. discoideum infected with the 1legG1 mutant strain was hyper-inhibited for directed migration in the under-agarose assay, which was reverted upon overproduction of LegG1 to an extent comparable to amoebae infected with wild-type bacteria. Single cell tracking indicated that the directionality as well as the velocity of D. discoideum infected with the parental strain or 1legG1 is impaired. Thus, the Ran activator LegG1 promotes cell motility by stabilizing microtubules and thereby might antagonize the possibly deleterious impact of other L. pneumophila Icm/Dot substrates on the host cytoskeleton.

Recently, the effect of the small signaling molecule LAI-1 (Legionella autoinducer-1, 3-hydroxypentadecane-4-one) on chemotactic migration of D. discoideum was analyzed (Simon et al., 2015). This study was based on the observations that L. pneumophila lacking the autoinducer synthase LqsA no longer impaired the migration of infected amoebae, and a 1icmT mutant strain overproducing LqsA inhibited cell migration, obviously in an Icm/Dot-independent manner. Synthetic LAI-1 dose-dependently inhibited the migration of D. discoideum in the micromolar range. RNA interference with epithelial cells subjected to a scratch ("wound closure") assay revealed that LAI-1 signaling requires the scaffold protein IQGAP1, the small GTPase Cdc42 as well as the Cdc42-specific guanine nucleotide exchange factor ARHGEF9, but not other modulators of Cdc42, or RhoA, Rac1 or Ran GTPase. Thus, the L. pneumophila signaling molecule LAI-1 produced either in the LCV or added exogenously promotes inter-kingdom signaling.

In summary, L. pneumophila modulates the motility of eukaryotic cells not only by means of Icm/Dot substrates, but also through the Lqs system and the small signaling molecule LAI-1 (Personnic et al., 2017). At present, it is unclear what the benefit of host cell migration inhibition for L. pneumophila might be. The phenomenon might simply be an indirect effect of perturbation of trafficking pathways and cytoskeletal dynamics. More intriguingly, however, it might also reflect the fact that macropinocytosis competes with migration of D. discoideum, two processes balanced by PI(3,4,5)P<sup>3</sup> (Veltman, 2015). Hence, by inhibiting cell migration, L. pneumophila might promote its macropinocytic uptake by the amoebae.

#### CELL BIOLOGICAL, BIOCHEMICAL, AND PROTEOMICS ANALYSIS OF PURIFIED INTACT LCVS

Intact LCVs from D. discoideum can be purified by a straight-forward two-step protocol involving immuno-affinity enrichment using an antibody against the Icm/Dot substrate SidC, which selectively decorates the pathogen vacuole, and a secondary antibody coupled to magnetic beads. This first step is followed by conventional Histodenz gradient density centrifugation, yielding intact LCVs in high purity and yield (Urwyler et al., 2010; Hoffmann et al., 2012, 2013) (**Figure 2G**). Using D. discoideum producing the LCV marker calnexin-GFP, infected with red fluorescent L. pneumophila, the LCV purification procedure can be easily followed by fluorescence microscopy.

Proteomics analysis of these purified LCVs revealed the presence of 560–1150 host proteins (Urwyler et al., 2009; Hoffmann et al., 2014; Schmölders et al., 2017). Amoebae proteins identified include small Rab GTPases, Rap1, Ran, and its effector Ran binding protein 1 (RanBP1). Moreover, large GTPases, components of the endosomal and late secretory trafficking pathways, as well as protein or lipid kinases and phosphatases were identified, many of which had not been implicated in LCV formation before. The accumulation of several of these host proteins on the LCV was confirmed by fluorescence microscopy using D. discoideum strains producing the corresponding GFP-fusion proteins (Urwyler et al., 2009; Finsel et al., 2013a; Rothmeier et al., 2013; Hoffmann et al., 2014; Schmölders et al., 2017; Steiner et al., 2017a). While some of these host factors accumulate on LCVs in an Icm/Dot-dependent manner (e.g., Rab1,−7,−8,−14, Rap1, Ran, RanBP1, Sey1), others localize to the pathogen vacuole Icm/Dot-independently (e.g., retromer subunits).

Preparations of intact LCVs can also be utilized for biochemical experiments (Steiner et al., 2017a). To this end, D. discoideum producing calnexin-GFP, GFP-Sey1, or catalytically inactive GFP-Sey1\_K154A were infected with mCherry-producing virulent L. pneumophila. LCVs were purified according to the two-step protocol described above, and subsequently, the effect of GTP, GDP or the non-hydrolysable GTP analog Gpp(NH)p was analyzed by fluorescence microscopy. Interestingly, the purified LCVs aggregated and expanded their size about 2-fold in a Sey1 and GTP-dependent manner. This experiment revealed in a reductionist system the importance of Sey1 activity for LCV aggregation and circumferential ER remodeling.

#### SCREEN FOR ANTIVIRULENCE COMPOUNDS IN *LEGIONELLA*-INFECTED AMOEBAE

Antibiotic resistance of human pathogens is an ever-rising and global problem. The approach to develop new bactericidal or bacteriostatic antibiotics is compromised by a high selection pressure on pathogens to become resistant against these compounds. A more efficient method is to target the virulence and the ability of bacteria to infect the host, rather than targeting directly the capacity to multiply. Such an "anti-virulence" strategy aims at specific mechanisms that promote bacterial pathogenicity, including metabolic pathways, binding to and uptake by host cells, signaling, and vesicle trafficking pathways, or the secretion and mode of action of toxins. There is the hope that anti-virulence compounds reduce the evolutionary pressure to develop drug resistance, and by selectively targeting virulent bacteria, spare the normal host microbiota.

Another major obstacle in developing novel antibiotic or antivirulence compounds is the limited compound accessibility of intracellular bacteria, reflected in compromised bioavailability of a compound. To efficiently address these problems, and given the similarities between amoebae and human immune phagocytes such as macrophages or neutrophils (Hilbi et al., 2007; Escoll et al., 2013; Hoffmann et al., 2013), we established A. castellanii as a robust and high-throughput compatible model system to screen for anti-virulence compounds. To this end, the amoebae were seeded into 96-well plates and infected with GFP-producing L. pneumophila. The progress of intracellular growth correlates with an increase in fluorescence (Harrison et al., 2013, 2015b). The medium chosen minimizes auto-fluorescence and growth of the amoebae, as well as extracellular growth of the bacteria. Under these conditions, bacterial replication typically comprised a lag and a replicative phase, after which all amoebae were dead and bacterial fluorescence remained constant. Upon addition of low molecular weight compounds to infected amoebae, we identified palmostatin M—a β-lactone inhibitor of eukaryotic Ras depalmitoylases and Ras signaling (Hedberg et al., 2011; Rusch et al., 2011)—as an inhibitor of intracellular growth of L. pneumophila as well as of Mycobacterium marinum and Mycobacterium tuberculosis (Harrison et al., 2013; Kicka et al., 2014).

Moreover, using this A. castellanii—L. pneumophila infection assay, we screened a highly diverse, pathway-based chemical library, referred to as the Sinergia library (Harrison et al., 2015a). Structure-activity relationship (SAR) studies using variants of a hit compound thus identified revealed essential structural features, namely a triple-ring scaffold with a central triazine moiety, substituted by two piperidine or pyrrolidine rings in positions 3 and 5, and by an amine group bearing a single aliphatic chain moiety in position 1. The most effective compound, ZINC00615682, inhibited intracellular growth of L. pneumophila with an IC<sup>50</sup> of ∼20 nM in A. castellanii and somewhat less efficiently in D. discoideum or macrophages (Harrison et al., 2015a). These results validate the amoebae-based screen and demonstrate that a corresponding SAR analysis allows the identification of novel inhibitors of intracellularly growing L. pneumophila and other bacteria.

### CONCLUSIONS AND OUTLOOK

Intense research over the last years revealed that amoebae not only represent an important ecological niche of the environmental bacterium L. pneumophila, but are also versatile and powerful models for the molecular and cellular analysis of host-pathogen interactions. The natural host A. castellanii is permissive for L. pneumophila over a range of different temperatures, and this robust system is useful to assess bacterial mutant phenotypes, as well as to screen for anti-virulence compounds. The social amoeba D. discoideum is genetically tractable and allows analysis of host mutant strains. Dually labeled D. discoideum strains are ideally suited for (live-cell) fluorescence microscopy, imaging flow cytometry, cell migration assays, or LCV isolation. Purified intact LCVs can be further analyzed by proteomics and cell biological or biochemical tests. Obvious limitations of the amoebae models are the apparent lack of key components of cell-autonomous innate immunity pathways, such as caspase family proteases and the transcription factor NF-kB (Eichinger et al., 2005; Clarke et al., 2013). On the other hand, amoebae models are useful for the analysis of host-pathogen interactions beyond Legionella infection. Accordingly, Acanthamoeba spp. and D. discoideum can serve to assess various aspects of cellular virulence for a large number of pathogenic bacteria including Mycobacterium spp., Burkholderia spp., and Vibrio cholerae (Hilbi et al., 2007; Cosson and Soldati, 2008). Hence, amoebae and other protozoa will continue to provide valuable insights into important aspects of the ecology and virulence of amoeba-resistant environmental bacterial pathogens.

#### AUTHOR CONTRIBUTIONS

ALS, CFH, LE, MS, and HH wrote the manuscript.

#### ACKNOWLEDGMENTS

We would like to thank Stephen Weber for providing **Figure 2D**. Research in the laboratory of HH was supported by the Swiss National Science Foundation (SNF; 31003A\_153200), the University of Zürich, the Novartis Foundation for Medical-Biological Research, the OPO foundation, and the German Ministry of Education and Research (BMBF) in the context of the EU Infect-ERA initiative (project EUGENPATH; 031A410A). LE acknowledges support of this work by the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG: CRC670, TP01) and by Köln Fortune. Research in the laboratory of MS was supported by DFG STE 838/8-1. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

#### REFERENCES

Albers, U., Reus, K., Shuman, H. A., and Hilbi, H. (2005). The amoebae plate test implicates a paralogue of lpxB in the interaction of Legionella pneumophila with Acanthamoeba castellanii. Microbiology 151, 167–182. doi: 10.1099/mic.0.27563-0

Albers, U., Tiaden, A., Spirig, T., Al Alam, D., Goyert, S. M., Gangloff, S. C., et al. (2007). Expression of Legionella pneumophila paralogous lipid A biosynthesis genes under different growth conditions. Microbiology 153, 3817–3829. doi: 10.1099/mic.0.2007/009829-0


major phosphatidylinositol 4-phosphate-binding effector protein of Legionella pneumophila. J. Biol. Chem. 284, 4846–4856. doi: 10.1074/jbc.M807505200


tower water using a portable microfluidic system. Sci. Rep. 7:3092. doi: 10.1038/s41598-017-03293-9


**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 Swart, Harrison, Eichinger, Steinert and Hilbi. 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 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.

# When Dicty Met Myco, a (Not So) Romantic Story about One Amoeba and Its Intracellular Pathogen

Elena Cardenal-Muñoz\*, Caroline Barisch, Louise H. Lefrançois, Ana T. López-Jiménez and Thierry Soldati

Department of Biochemistry, Sciences II, Faculty of Sciences, University of Geneva, Geneva, Switzerland

In recent years, Dictyostelium discoideum has become an important model organism to study the cell biology of professional phagocytes. This amoeba not only shares many molecular features with mammalian macrophages, but most of its fundamental signal transduction pathways are conserved in humans. The broad range of existing genetic and biochemical tools, together with its suitability for cell culture and live microscopy, make D. discoideum an ideal and versatile laboratory organism. In this review, we focus on the use of D. discoideum as a phagocyte model for the study of mycobacterial infections, in particular Mycobacterium marinum. We look in detail at the intracellular cycle of M. marinum, from its uptake by D. discoideum to its active or passive egress into the extracellular medium. In addition, we describe the molecular mechanisms that both the mycobacterial invader and the amoeboid host have developed to fight against each other, and compare and contrast with those developed by mammalian phagocytes. Finally, we introduce the methods and specific tools that have been used so far to monitor the D. discoideum—M. marinum interaction.

#### *Edited by:*

Ludwig Eichinger, University of Cologne, Germany

#### *Reviewed by:*

Salvatore Bozzaro, Università degli Studi di Torino, Italy Hubert Hilbi, University of Zurich, Switzerland

> *\*Correspondence:* Elena Cardenal-Muñoz elena.cardenal@unige.ch

*Received:* 03 November 2017 *Accepted:* 18 December 2017 *Published:* 09 January 2018

#### *Citation:*

Cardenal-Muñoz E, Barisch C, Lefrançois LH, López-Jiménez AT and Soldati T (2018) When Dicty Met Myco, a (Not So) Romantic Story about One Amoeba and Its Intracellular Pathogen. Front. Cell. Infect. Microbiol. 7:529. doi: 10.3389/fcimb.2017.00529 Keywords: *Dictyostelium discoideum*, *Mycobacterium marinum*, host-pathogen interactions, model organisms, phagocytosis, infection, methods

### INTRODUCTION

### The Amoeba *Dictyostelium discoideum*

Dictyostelium discoideum is a soil amoeba that feeds on bacteria by phagocytosis. The Amoebozoa branch split from the common lineage leading to fungi and animals, shortly after these crown organisms split from the Plant lineage (Eichinger et al., 2005). D. discoideum has a very ATrich (77.75%) haploid genome, organized into 6 chromosomes, which has been sequenced and annotated (dictyBase, 2004; Eichinger et al., 2005). The isolation in the 70's of axenic strains that could feed on liquid medium by pinocytosis (Watts and Ashworth, 1970) enabled the easy laboratory culture of D. discoideum. In addition, this amoeba is genetically tractable, being relatively easy to transform for gene expression from extrachromosomal plasmids, generation of knockins, and knockouts by homologous recombination (Veltman et al., 2009a,b; Wiegand et al., 2011; Mukai et al., 2016), and mutagenesis by Restriction Enzyme-Mediated Integration (REMI) (Kuspa, 2006). Moreover, D. discoideum can be used in multiple cell biology, biochemistry, and imaging assays, as well as for high-throughput screens and RNAseq analysis (Eichinger and Rivero, 2013; Kicka et al., 2014; Rosengarten et al., 2015). For these reasons, D. discoideum has been extensively used as a model system to study very diverse biological processes such as motility, chemotaxis, vesicular trafficking, gene expression, facultative multicellularity, and hostpathogen interactions (Barry and Bretscher, 2010; Maniak, 2011; Stevense et al., 2011; Loomis, 2014; Tosetti et al., 2014; Nichols et al., 2015).

#### Central Role of Autophagy in Development and Host-Microbe Interaction

In conditions of nutrient repletion, D. discoideum lives as a vegetative unicellular organism. However, when resources become scarce, the amoeba triggers a developmental program for the aggregation of hundreds of thousands of cells into a true multicellular organism (Raper, 1935; Dormann et al., 2000; Williams, 2006). This process is known as the "developmental cycle" of D. discoideum and has been extensively studied as one of the evolutionary origins of multicellularity. Aggregation and morphogenesis during development require a high level of cellular activity. To undertake this metabolic demand in the absence of nutrients, D. discoideum relays on macroautophagy (hereafter referred as autophagy), a major catabolic pathway in eukaryotes (Otto et al., 2004). Non-selective autophagy consists of the engulfment of bulk cytosolic material in a doublemembrane compartment called the autophagosome. Upon fusion with lysosomes, autolysosomes degrade, and recycle their content, enabling cell survival under starvation. One of the major regulators of autophagy is the target of rapamycin complex 1 (TORC1). Under nutrient repletion, TORC1 downregulates autophagy by repressing the expression of autophagy genes and/or by phosphorylation and inhibition of proteins involved in autophagosome formation (reviewed in Nakatogawa, 2015). In D. discoideum, the TORC1 complex consists of the TOR serine/threonine kinase and two TOR activators, Lst8, and Raptor (Wullschleger et al., 2006; Rosel et al., 2012). Like in mammals, starvation decreases TORC1 activity, leading to induction of autophagy (King et al., 2011; Rosel et al., 2012).

Since many of the well-studied proteins known to participate in mammalian autophagy are conserved in this amoeba, autophagy can be easily monitored in D. discoideum. For instance, like in mammals, the activity of D. discoideum TORC1 can be analyzed by monitoring the phosphorylation state of Raptor and the TORC1 effector protein 4E-BP1 by immunoblotting (Rosel et al., 2012; Cardenal-Muñoz et al., 2017). In addition, in D. discoideum also the changes in autophagosome formation can be distinguished from changes in autophagic degradation by the differential response of Atg18 [WD-repeat protein interacting with phosphoinositides (WIPI) in mammals] and Atg8a (LC3/GABARAP family proteins in mammals) (Calvo-Garrido et al., 2010). Accordingly, during induction of autophagosome formation Atg8 and Atg18 translocate from the cytosol to membranes of nascent and elongating phagophores (the yet unclosed autophagic double membrane compartment). Immediately after closure of the autophagosome, Atg8 stays in both inner and outer membranes, while Atg18 dissociates. Upon fusion of the autophagosome with lysosomes, the hydrolytic enzymes delivered from the lysosome not only degrade the autophagic cargo but also the inner membrane of the autolysosome and its associated Atg8, while the Atg8 on the outside is recycled to the cytosol (reviewed in Dominguez-Martin et al., 2017).

In recent years, D. discoideum has also become an interesting model to study the molecular mechanisms regulating xenophagy (reviewed in Mesquita et al., 2016). Xenophagy is a selective autophagy pathway specifically recognizing and digesting intracellular pathogens. It relies on selective receptors that recruit the cargo to be degraded [i.e., bacteria or remnants of their phagosome decorated with "eat-me" signals such as ubiquitin (Ub) or galectins in mammals] to autophagic membranes (Thurston et al., 2009, 2016; Boyle and Randow, 2013; Noad et al., 2017). Several xenophagy receptors have been described in mammalian cells, but only one, p62, has been identified or studied so far in D. discoideum (Calvo-Garrido and Escalante, 2010; Gerstenmaier et al., 2015; Lampe et al., 2015; Cardenal-Muñoz et al., 2017). In addition, although D. discoideum lacks galectins, which recognize bacterial or host membrane glycans, another family of cytosolic lectins in this amoeba, the discoidins (Dsc), share many molecular and biological characteristics with galectins. Dsc are highly expressed upon starvation (Rosen et al., 1973), bind to self-glycans during development (Cooper and Barondes, 1984; Eitle et al., 1993) and can be secreted to the extracellular medium (Barondes et al., 1984). Interestingly, Dsc bind to sonicated extracts of Escherichia coli and Klebsiella aerogenes (Cooper et al., 1983), as well as to glutaraldehyde-fixed K. aerogenes bacteria (Madley and Hames, 1981). Recent studies have revealed that Dsc are enriched on compartments containing virulent Legionella pneumophila strains (i.e., Corby and JR32) but not the less virulent L. hackelieae, suggesting a role in innate immunity (Shevchuk et al., 2009; Urwyler et al., 2009). Moreover, one ortholog of the human TRIM37, a member of the Tripartite Motif protein superfamily (TRIMs) of autophagy receptors and regulators that often interact with galectins (Mandell et al., 2014; Kimura et al., 2015; Chauhan et al., 2016), has been identified in D. discoideum (Dunn et al., 2018). It will be interesting to uncover whether or not Dsc, TRIM-like protein and p62 cooperate in the immunity of D. discoideum against pathogenic bacteria.

### The Bacterial Pathogen *Mycobacterium marinum*

Mycobacterium is the only genus of the Mycobacteriaceae family. It comprises more than 150 different species, among them two of the most important human pathogens, M. leprae and M. tuberculosis (Mtb), causing leprosy and tuberculosis, respectively. The latter represents the most severe bacterial diseases, responsible for 1.8 million deaths worldwide in 2015 (World Health Organization, 2015, http://www.who.int/ tb/publications/global\_report). One of the important features of Mtb is its capacity to persist inside its host in a nonreplicating state with low metabolic activity. This phenomenon is called dormancy (or latency) and affects 2–3 billion people (i.e., one third of the world population). Latent tuberculosis is non-transmissible and presents no clinical manifestation. However, the disease reactivates in 5–10% of the cases usually as a consequence of immunosuppression, becoming contagious. Therefore, providing treatments to prevent the reactivation of latent tuberculosis is an important concern in the control of this disease (World Health Organization, 2015).

Mtb belongs to the group of genetically related mycobacteria, the Mtb complex, that cause tuberculosis in human or other animals. However, another group of human pathogens emerging as a major public health problem is the Non-Tuberculous Mycobacteria (NTM) or Mycobacteria Other Than Tubercle bacilli (MOTT), which includes among others the species M. ulcerans, M. abscessus, M. avium-intracellulare complex (MAC) and M. marinum (Ryu et al., 2016). In contrast to Mtb, with no environmental reservoir demonstrated so far, M. marinum is an intracellular pathogenic mycobacterium with ubiquitous distribution. It can be found in aquatic environments, and it affects a wide range of freshwater and marine vertebrates including fish, amphibians, and turtles (Decostere et al., 2004). During infection of these animals, M. marinum forms granulomatous lesions highly similar to the ones produced by Mtb in humans. In addition, M. marinum also develops a latent disease in the zebrafish model (Parikka et al., 2012). Very importantly, and reflecting their genetic proximity, M. marinum is also an opportunistic pathogen of humans, but due to its optimal growth at 30◦C, M. marinum infections are mostly localized and restricted to the skin and extremities. These lesions are usually painless and can be treated with anti-Mtb antibiotics when healing is not spontaneous (Aubry et al., 2017). At the cellular level, the infection course of M. marinum is also very similar to that of Mtb (reviewed in Ramakrishnan, 2004; Tobin and Ramakrishnan, 2008). Both bacteria avoid degradation within the host cells by arresting the maturation of their phagosomes and, contrary to the non-tuberculous M. avium, they escape to the host cytosol prior to dissemination (Stamm et al., 2003; Hagedorn et al., 2009; Simeone et al., 2015; Jamwal et al., 2016).

The genome of the M. marinum strain M was sequenced and annotated in 2008 by Stinear et al. (2008). Whole genome comparisons revealed that M. marinum is the ancestor of M. ulcerans and that both species are the closest relatives of the Mtb complex (Stinear et al., 2000, 2008). Among these species, M. marinum has the largest genome with 6.6 Mb, while the genome size of M. ulcerans and Mtb is 5.8 and 4.4 Mb, respectively. The genome of M. marinum shares 85% nucleotide identity with Mtb and 97% with M. ulcerans. Although phylogenetically and genetically close, Mtb and M. marinum present some differences such as the cell wall composition (Daffe et al., 1991; Tonjum et al., 1998), the abundance of prolineglutamic acid polymorphic guanine-cytosine-rich sequence (PE-PGRS) and proline-proline-glutamic acid (PPE) proteins (Cole et al., 1998; Bottai et al., 2012; Ates et al., 2015), and the synthesis of carotenoid pigments, which only occurs in M. marinum (Stinear et al., 2008).

Nevertheless, and more importantly, many virulence factors are conserved between M. marinum and Mtb. One example are the genes involved in dormancy (Dos-regulon) (Gerasimova et al., 2011). In addition, M. marinum and Mtb secrete virulence effector proteins through their complex cell walls thanks to very sophisticated secretion systems. In this context, the two mycobacteria possess five type VII secretion systems (T7SS) called ESX1-5 (reviewed in Groschel et al., 2016), in reference to the first secreted factor identified ESAT6 (Sorensen et al., 1995). The T7SS are not exclusive to mycobacteria and have been found in other bacterial genera such as Streptomyces and Bacillus (Gey Van Pittius et al., 2001; Pallen, 2002; Unnikrishnan et al., 2017). Despite their diversity, the T7SS share two common characteristics: the presence of FtsK/SpoIIIE transmembrane proteins and the secretion of small peptides of around 100 amino acids with the conserved motif Trp-X-Gly (WXG) (Pallen, 2002). The WXG motif is involved in the generation of a helixturn-helix structure and is present in two of the most studied T7SS secreted peptides, the 6 kDa early secretory antigenic target ESAT6 and the 10 kDa culture filtrate protein CFP10 (Renshaw et al., 2005). Each ESX system is encoded in its respective ESX cluster or locus, which also contains the genes for secreted peptides and other accessory proteins (Bitter et al., 2009). The ESX loci do not complement each other (reviewed in Abdallah et al., 2007), which suggests that they are involved in different functions. Only three of them have been shown to play a role in pathogenesis: ESX-1, 3, and 5. Whereas ESX-3 is required for iron and zinc acquisition (Serafini et al., 2009, 2013; Siegrist et al., 2009, 2014), ESX-5 is involved in nutrient intake (Ates et al., 2015), cell wall integrity and secretion of PE and PPE proteins. The ESX-1 locus is the most studied, since it was shown to be very important for mycobacterial virulence (Pym et al., 2002). The attenuated M. microti and M. bovis BCG bacille Calmette–Guérin (BCG), used for vaccines, lack a partially overlapping region of the ESX-1 locus, which was called Region of Difference 1 (RD1). This region comprises two genes that encode the above mentioned secreted peptides ESAT6 and CFP10. Early studies using transposon libraries and purified proteins identified ESX-1, and in particular ESAT6 and CFP10, as membranolytic factors in both Mtb and M. marinum (Hsu et al., 2003; Gao et al., 2004). Importantly, the deletion of M. marinum RD1 can be complemented with homologous genes from Mtb, thus confirming the functional similarity of the RD1 regions in both species (Gao et al., 2004). Even though a 3.7 kb portion of M. marinum RD1 is partially duplicated within its genome (Stinear et al., 2008), deletion of RD1 prevents M. marinum escape from its phagosome as occurs with Mtb, suggesting that the duplicated fragment of M. marinum is not functional (Simeone et al., 2015; Cardenal-Muñoz et al., 2017).

These pathologic, phylogenetic, and genomic aspects make of M. marinum an interesting model to study mycobacterial infections. Importantly, M. marinum also has advantages for laboratory work. Despite the fact that this species is considered a slow growing mycobacterium, M. marinum grows faster than Mtb, with a doubling time of 6–8 h vs. the 20–24 h of Mtb. Furthermore, M marinum is a level 2 biosafety organism, and numerous molecular tools are available and transposable from Mtb to M. marinum, and vice versa.

#### The *D. discoideum*—*M. marinum* Infection Model

D. discoideum has been used to study conserved mechanisms of bacterial pathogenicity and host defenses involved in many human diseases induced by diverse bacteria genera such as Francisella, Legionella, Salmonella, or Mycobacterium (Bozzaro and Eichinger, 2011; Weber and Hilbi, 2014; Brenz et al., 2017; Cardenal-Muñoz et al., 2017; Steiner et al., 2017). In both D. discoideum and human phagocytes, the main cellautonomous defenses rely on the highly conserved microbicidal machinery that arms the phagosome (Boulais et al., 2010). This includes a number of genes and pathways involved in bacteria recognition, phagocytic uptake, actin dynamics, phagosome identity, acidification and maturation, lysosomal degradation, production of Reactive Oxygen Species (ROS), metal poisoning and nutrition immunity (recently reviewed in Dunn et al., 2018).

The ability of M. marinum to replicate within D. discoideum was first described in 2003 (Solomon et al., 2003). Solomon and collaborators monitored monolayers of D. discoideum infected with GFP-producing M. marinum at 25.5◦C, a temperature optimal for the joint growth of host and bacterium. For these long-term infection assays, and to avoid rapid host cell death, D. discoideum needs to be infected at low multiplicity of infection (MOI ≤1–10). After an initial incubation without antibiotics, to allow the phagocytosis of M. marinum, streptomycin was added to the infection sample to impede the growth of extracellular bacteria. In this first approach, the increased growth of M. marinum was assessed by direct observation, but other quantitative methods such as luminescence recording, fluorescence-activated cell sorting (FACS) and colony-forming units (CFUs) counting have confirmed the intracellular proliferation of this bacterium within D. discoideum, both in static and shaking cultures (Solomon et al., 2003; Hagedorn and Soldati, 2007; Arafah et al., 2013). Other antibiotics such as amikacin or a mix of penicillin and streptomycin can be used to kill the extracellular M. marinum when performing infection assays in D. discoideum (Arafah et al., 2013). The presence of extracellular bacteria in the infection sample can be tested by incubating live cells with an anti-M. marinum serum (anti-mar) and performing immunofluorescence assays (Supplementary Table 1); Only the extracellular bacteria, and not those inside the amoeba, are accessible to the serum (Hagedorn et al., 2009).

### *M. MARINUM* INFECTION COURSE IN *D. DISCOIDEUM*

The infection cycle of M. marinum in D. discoideum starts with the phagocytosis of the bacterium by the amoeba, and finishes with the bacterial egress from the cell (**Figure 1** and **Table 1**). Three different growth stages of M. marinum can be distinguished during the infection cycle: (i) an initial phase during the first 12 h post-infection (hpi), when M. marinum actively manipulates the D. discoideum phagocytic pathway to convert the phagosome where it resides into a replicationpermissive niche [mycobacteria-containing vacuole (MCV)], (ii) an enhanced proliferation phase between 12 and 37 hpi, and (iii) a phase of arrested proliferation and/or decreased bacterial load due to bacterial death or release after 37 hpi (Hagedorn and Soldati, 2007). This infection cycle might be impacted by the number of bacteria initially taken up. When a single amoeba internalizes a clump of bacteria, the M. marinum clump either cannot initiate its intracellular growth, remaining in a dormant or undigested state as occurs in zebrafish, or causes cytotoxicity to D. discoideum (Solomon et al., 2003; Parikka et al., 2012).

### Early *M. marinum* Infection Phase (∼0–12 hpi)

During uptake, M. marinum is surrounded by a circular ruffle and ingested into a newly produced phagosome (**Figure 2**). The engulfment of the bacterium by the tight phagocytic cup has been observed by microscopy of live and fixed cells, using markers of the actin cytoskeleton of D. discoideum (Hagedorn et al., 2009). In addition, phagocytosis of GFP-producing or Alexa 488 hydrazide-labeled M. marinum has been quantitatively monitored by flow cytometry (Sattler et al., 2013). It is important to notice that the labeling of surface amines with Alexa Fluor 488 tetrafluorophenyl ester (Alexa Fluor 488-TFP), normally used for other bacteria such as Klebsiella pneumoniae or Salmonella enterica, is not suitable for M. marinum, presumably due to its extremely hydrophobic and poor-in-proteins cell wall (Sattler et al., 2013).

Ten minutes after phagocytosis, the early endosomal marker Rab5a is withdrawn from the M. marinum phagosome, compared to 3 min for the non-pathogenic M. smegmatis (Barisch et al., 2015a). Subsequently, the late endosome and lysosome marker Rab7a is detected at the early MCV (Cardenal-Muñoz et al., 2017), as occurs in macrophages (Lerena and Colombo, 2011). However, M. marinum rapidly bypasses the phagocytic pathway and blocks the maturation of its compartment. Thus, late maturation markers such as vacuolin and cathepsin D are absent from the MCV from 0.3 to 6 hpi, and the vacuolar H+- ATPase (vATPase) is almost undetectable on the MCV at the end of this infection phase (12 hpi, **Figure 2**) (Hagedorn and Soldati, 2007). In this context, and contrary to M. smegmatis, only a small percentage of the phagocytosed M. marinum are killed by D. discoideum within the first hours (Hagedorn and Soldati, 2007), even in conditions of mild starvation that stimulate autophagy (i.e., phosphate buffer supplemented with 5% nutrient-rich medium) (Delince et al., 2016).

It has been shown that actin polymerization is required for the efficient maturation arrest of the M. marinum MCV during early infection of D. discoideum (Kolonko et al., 2014). Thus, Arp2/3 and WASH (Wiskott-Aldrich syndrome protein and Scar Homolog), two complexes involved in actin nucleation and polymerization on endosomal membranes (Duleh and Welch, 2010), can be observed in increasing association with the MCV from 0 to 6 hpi (**Figure 2**). This is concomitant with the presence of a patchy F-actin coat covering the MCV and with the absence of VatA, a peripheral subunit of the vATPase, on the same MCV at 6 hpi (Kolonko et al., 2014). On the contrary, amoebae either treated with the inhibitor of actin polymerization latrunculin A (LatA), producing a fusion construct between VacA and the actin-depolymerizing factor cofilin (VMC), or lacking the WASH complex (wshA-) harbor actin-negative

Recapture into lysosome-like compartments may precede late exocytosis. However, this has not been shown yet in D. discoideum. Early exocytosis (9.3) can be

MCVs which accumulate VatA and VatM (a transmembrane subunit of the vATPase), while losing the endocytic marker p80. Interestingly, p80 is not lost from the phagosome containing M. smegmatis and the avirulent mutant M. marinum L1D (Kolonko et al., 2014) (Supplementary Table 2). In agreement with the LatA-dependent accumulation of the vATPase on MCVs, LatA treatment also leads to acidification of the bacterial compartment, resulting in decreased bacteria viability (Kolonko et al., 2014). Interestingly, actin polymerization also prevents the acidification of the M. marinum and Mtb phagosomes in murine phagocytes (Kolonko et al., 2014). In addition to Arp2/3 and WASH, the D. discoideum homolog of mammalian flotillin, vacuolin B (VacB), is also essential for the establishment of the MCV, assisting M. marinum in the retrieval of the vATPase from its niche. On the contrary, the small GTPase RacH of D. discoideum contributes to the maturation of the MCV into phagolysosomes, reducing the intracellular bacterial load (Hagedorn and Soldati, 2007) (Supplementary Table 3).

induced upon starvation. Intercellular dissemination occurs after M. marinum release from the amoeba (10).

It has to be noticed that M. marinum does not interfere with the contractile vacuole of D. discoideum (Kolonko et al., 2014). This osmoregulatory organelle consists of interconnected vesicles, cisternae and tubules that fuse with the plasma membrane to expel the excess of water ingested during macropinocytosis and maintain cell volume (Patterson, 1980). The contractile vacuole accumulates Rab11 GTPase, vATPase, and calmodulin (Du et al., 2008), but the latter does not localize to the MCV (Kolonko et al., 2014; Cardenal-Muñoz et al., 2017). In addition, tetramethylrhodamine isothiocyanate (TRITC)-dextran, which labels the endolysosomal system but not the contractile vacuole (Clarke et al., 2002), can be found inside the MCV (Gerstenmaier et al., 2015).

### *M. marinum* Proliferation Phase (∼12–37 hpi)

After establishment as a replication-permissive niche, the MCV becomes more spacious (**Figures 1**, **2** and **Table 1**). It still harbors Rab7a and, as occurs in human monocytes, it accumulates the post-lysosomal marker vacuolin/flotillin (Hagedorn and Soldati, 2007; Barisch et al., 2015b). On the contrary, VatA association with MCVs significantly decreases (Kolonko et al., 2014). Intriguingly, WASH contributes to M. marinum growth, and the WASH and Arp2/3 complexes can be observed at the MCV at 24 hpi (Kolonko et al., 2014). However, the F-actin coat of the MCV is almost completely lost and treatment with LatA does not alter the association of p80 or VatA with the MCV (Kolonko et al., 2014). Kil2, a D. discoideum P-type ATPase involved in intra-phagosomal killing of Klebsiella, does not play any role in the establishment of the MCV nor in M. marinum intracellular replication (Lelong et al., 2011). In addition, the D. discoideum homolog of coronin, a mammalian protein involved in the organization of the actin cytoskeleton that positively modulates intracellular mycobacterial growth in macrophages (Jayachandran et al., 2007), appears heterogeneously localized at the MCV but



TABLE

1


Continued

seems to negatively regulate the uptake or early survival and proliferation of M. marinum (Solomon et al., 2003). These data suggest that other host factors must be responsible for the support of the M. marinum intracellular replication in this amoeba.

One of the bacterial effectors required for the successful proliferation of M. marinum within D. discoideum is MAG24-1 (Solomon et al., 2003; Hagedorn and Soldati, 2007), a PE-PGRS protein also essential for replication in mammalian macrophages, frogs and flies (Ramakrishnan et al., 2000; Dionne et al., 2003). Actually, due to its role in intracellular growth, MAG24-1 can be used as a tool to monitor the proliferation of M. marinum mutant strains in D. discoideum by simply following the expression of GFP under the control of the mag24-1 promoter in bacteria carrying the map24::GFP vector (Hagedorn and Soldati, 2007) (Supplementary Table 1). Another factor essential for optimal M. marinum intracellular growth is the ESX-1 secretion system, responsible for the rupture of the phagosomes containing M. marinum and its release into D. discoideum cytosol at the end of this infection phase (21–37 hpi) (Hagedorn and Soldati, 2007; Hagedorn et al., 2009). Mtb and M. avium can be seen in D. discoideum within spacious compartments that accumulate vacuolin but, contrary to Mtb, the ESX-1-deficient M. avium (Houben et al., 2014) does not escape its MCV (Hagedorn et al., 2009). A functional ESX-1 is important for optimal M. marinum proliferation (Hagedorn et al., 2009; Cardenal-Muñoz et al., 2017), and the exogenous expression of M. marinum ESAT-6 in D. discoideum rescues the inefficient intracellular replication of 1RD1 bacteria (Hagedorn et al., 2009). The role of ESX-1 in phagosomal escape and intracellular proliferation of M. marinum and Mtb has also been described in mammalian cells (Stamm et al., 2003; Gao et al., 2004; Volkman et al., 2004; Tan et al., 2006; Simeone et al., 2015; Zhang et al., 2016). In addition to ESX-1, other mycobacterial factors have been shown to contribute to phagosomal rupture. The mycobacterial cell wall components phthiocerol dimycocerosates (PDIMs) have been revealed essential for ESAT-6 activity, suggesting some sort of cooperativity not yet understood (Augenstreich et al., 2017).

The D. discoideum ammonium transporter AmtA is a useful marker to determine whether or not mycobacteria are inside a compartment, since its fluorescent versions ubiquitously localize to all the endosomes and the MCV (Barisch et al., 2015b; Barisch and Soldati, 2017b) (Supplementary Table 4). Methods based on fluorescence resonance energy transfer (FRET), such as FRETbased microscopy or flow cytometry, have been performed in macrophages infected with Mtb and M. marinum to detect the mycobacteria-mediated phagosome disruption (Simeone et al., 2012, 2015; Acosta et al., 2014).

### Late *M. marinum* Infection Phase (∼37–72 hpi)

After 37 hpi, when M. marinum has escaped its MCV, the intracellular bacterial load stabilizes and even decreases (Hagedorn and Soldati, 2007). This might be the result of the convergence between recapture and killing of cytosolic bacteria in lytic compartments, and the release of the bacteria to the extracellular medium or neighboring cells (**Figure 1**). Events of recapture have not been reported so far in D. discoideum, but in macrophages cytosolic M. marinum is ubiquitinated and recaptured into non-autophagic double membrane compartments (LAMP-1-positive, LC3/Atg8-negative) (Collins et al., 2009). It was proposed that ESCRT, a protein machinery that sorts ubiquitinated cargo to degradation into intraluminal vesicles (Schmidt and Teis, 2012), would deliver these bacteria to multivesicular bodies (MVB) and finally to lysosomes, but the bactericidal activity of these putative recapture compartments was not established (Collins et al., 2009).

Regarding release, and as a consequence of the wide range of strategies used by bacteria to survive and replicate within their hosts, multiple egress strategies have emerged in the course of evolution (Friedrich et al., 2012) (**Figure 1** and **Table 1**). In some cases, egress implies the disruption of one or several host membranes, which may lead to the host cell death. Mycobacteria modulate the death of their host cells, inducing apoptotic or necrotic processes with different biological implications. These processes vary depending on the mycobacterial species, host cell types, and infection stage (reviewed in Aguilo et al., 2013; Srinivasan et al., 2014). In this context, many studies have associated ESX-1 activity and mycobacterial vacuolar escape to an increased host cell cytotoxicity and death (Hsu et al., 2003; Derrick and Morris, 2007; Kaku et al., 2007; Kinhikar et al., 2010; Simeone et al., 2012; Augenstreich et al., 2017). Thus, incubation with purified ESAT-6, as well as infection with wildtype (wt) Mtb and M. marinum but not with 1RD1 mutant bacteria, lyse epithelial and red blood cells and macrophages (Hsu et al., 2003; Gao et al., 2004). This is of special relevance since mycobacterial egress from its host might influence spread of the infection in a tissue. Similarly, by using a microfluidic device (the InfectChip) to microscopically monitor single-cell long-term infection in D. discoideum, it has been shown that a quarter of the phagocytosed M. marinum induces death and lysis of the amoebae, contributing to bacteria release (Delince et al., 2016).

Alternative non-lytic mechanisms of bacterial egress also occur during mycobacterial infection in D. discoideum. For instance, release of M. marinum within the first hours of infection, when most of the bacilli are still intravacuolar, thereby suggesting an exocytic process, can be induced by mild starvation of D. discoideum (Delince et al., 2016). This phenomenon is likely akin to the known massive exocytosis of lysosomal enzymes that D. discoideum undergoes upon starvation (Smith et al., 2010). In addition, bacterial ejection occurs at later time points (or earlier if the MOI is increased to 50–100, Gerstenmaier et al., 2015) once mycobacteria have escaped the MCV and accessed the host cytosol (Hagedorn et al., 2009). During ejection, M. marinum and Mtb, but not the ESX-1-deficient M. avium, use the actin cytoskeleton of D. discoideum to form an actin-based short barrel structure formed at the plasma membrane, called the ejectosome, through which they exit the cell (Hagedorn et al., 2009). In this context, some D. discoideum actin-binding proteins such as coronin and myosin IB are enriched at the ejectosome (Hagedorn et al., 2009) (Supplementary Table 4). In addition, ejection cannot occur in amoebae lacking the actin polymerization complex WASH (Kolonko et al., 2014). Strikingly, other markers of the actin cytoskeleton such as the actin nucleator complex Arp2/3 and myosin II are not enriched at the ejectosome, suggesting that this structure may assemble de novo during ejection (Hagedorn et al., 2009). It has been proposed that the D. discoideum plasma membrane reseals at the posterior of the bacterium after ejection, preventing the amoebal lysis or leakage (Gerstenmaier et al., 2015) (section D. discoideum Plasma Membrane Damage and Sealing at Later Time Points). Exclusion of DAPI (a membraneimpermeant DNA binding dye) from the nuclei of live amoebae during ejection demonstrates that M. marnium does not kill or induce leakage of its host while egressing by this mean (Hagedorn et al., 2009).

Exocytosis and ejection can be discriminated by fluorescence microscopy with various D. discoideum markers. For instance, only during exocytosis the intracellular pole of the egressing bacterium is positive for endosomal and MCV markers such as p80 and vacuolins. In addition, the extracellular part of the exocytosed bacterium is free of any plasma membrane marker (Hagedorn et al., 2009; Gerstenmaier et al., 2015). On the contrary, upon ejection, extracellular Mtb and M. marinum are wrapped in plasma membrane remnants positive for p80 or PM4C4 (Supplementary Table 4). Clumps of bacteria can be exocytosed, but ejectosomes, which are often observed to form clusters, only release individual bacteria (Hagedorn et al., 2009).

Once bacteria exit the cell, neighboring cells might become new hosts. Within macrophages, a subpopulation of cytosolic M. marinum sheds its ubiquitinated cell wall possibly to avoid recapture into lysosomes. These bacteria induce the polymerization of actin tails and spread via actin-dense filopodia that can be caught by other cells (Stamm et al., 2003; Collins et al., 2009; Hagedorn et al., 2009). However, and presumably due to the high rates of actin turnover, cytosolic M. marinum does not form persistent actin tails in D. discoideum, but events of cell-to-cell transmission synchronized with actin-based ejection are observed (Hagedorn et al., 2009). D. discoideum RacH, mentioned before to contribute to the maturation of the MCV, also assists M. marinum in its release from the amoeba and the subsequent intercellular dissemination. Accordingly, ejection of M. marinum is not observed in racH-cells (Hagedorn et al., 2009), and the population of infected racH- amoebae almost fully retain all bacteria until 37 hpi (Hagedorn and Soldati, 2007; Hagedorn et al., 2009) (Supplementary Table 3). In agreement with these results, quantitative dissemination assays have confirmed the positive role of RacH in M. marinum cell-tocell spread. In these fluorescence microscopy assays, an unlabeled donor D. discoideum strain (wt or racH-) was infected with red fluorescent M. marinum, and a wt acceptor strain expressing GFP was added to the infection after 12 h. Transmission from racHinfected cells to wt cells was reduced over eight-fold (Hagedorn et al., 2009).

#### *M. MARINUM* INDUCES MEMBRANE DAMAGE THAT IS COUNTERACTED BY *D. DISCOIDEUM*

Rupture of host membranes by bacterial pathogens is achieved by the well-orchestrated secretion of lipases, proteases and poreforming toxins. This is the case of vacuolar bacteria such as Mycobacterium, Shigella, Listeria, or Salmonella, which escape to the host cytosol via the disruption of their vacuole. Although the membranolytic activity of the mycobacteria ESX-1 secretion system is widely accepted, it has been proposed that ESX-1 dependent membrane disruptions do not require ESAT-6 activity in physiological conditions in macrophages (Conrad et al., 2017). Instead, the perforations would occur upon direct contact of Mtb and M. marinum ESX-1 with host MCV membranes (Conrad et al., 2017), which would confine the membranolytic activity to the bacterial poles, where the ESX-1 is enriched (Carlsson et al., 2009). Incompatible with this proposed model, M. marinum mutants specifically lacking only ESAT-6 and CFP10 (1CE) in an otherwise intact RD1 locus are defective in hemolysis, cytolysis, and cytotoxicity (Gao et al., 2004). Similarly, the virulent Mtb H37Rv strain with a 12 amino acids deletion in ESAT-6 shows impaired membranolytic activity in the THP-1 human macrophages (Houben et al., 2012). Regarding CFP10, it has been shown that the protein dissociates from ESAT-6 in acidic conditions similar to those encountered within the phagosome, and it does not disrupt artificial liposomes on its own (de Jonge et al., 2007). As a consequence, CFP10 has been proposed to act as a chaperone of ESAT-6, conferring stability or protection against proteases. Despite the fact that the structures of ESAT-6 and CFP10 have been solved (Renshaw et al., 2005), it is not yet well-understood how ESAT-6 induces membrane disruptions.

### Perforation and Repair of the MCV Membrane at Early Time Points

M. marinum resides inside a phagosome that shields it against the host intracellular immune responses but limits nutrient access. Maybe as a reaction, the bacterium creates a porous vacuole which might appear ideal to obtain unrestricted access to nutrients. Indeed, very early during infection of D. discoideum, M. marinum induces ESX-1-dependent damages in the MCV membrane, which results in the decoration of cytosol-exposed bacteria with Ub already at 1.5 hpi (Cardenal-Muñoz et al., 2017). This MCV membrane damage leads to a well-orchestrated early autophagic response consisting in: (i) transient downregulation of TORC1 and upregulation of transcription of autophagy genes such as atg1, the two paralogs atg8a and atg8b, and p62, (ii) enhanced formation of autophagosomes, and (iii) recruitment of the xenophagy machinery (Ub, GFP-p62, and Atg8a) to the damaged MCV. These responses to MCV damage are significantly reduced or even abolished when M. marinum lacks the RD1 locus and, more specifically, ESAT-6 and CFP-10 (Cardenal-Muñoz et al., 2017).

However, contrary to the increased bacterial killing expected from an enhanced xenophagic response, M. marinum survives and proliferates within D. discoideum by blocking the autophagic flux and its subsequent degradation within autolysosomes (Cardenal-Muñoz et al., 2017). Once again, the main player involved in this blockage is the mycobacterial ESX-1 secretion system, also shown to block the autophagic flux in human cells infected with Mtb (Romagnoli et al., 2012). Along these lines, even though 40–50% of the early M. marinum MCVs are positive for the autophagosomal marker Atg8a and for markers of acidic compartments [vATPase LysoSensor Green and neutral red (NR) (Hagedorn and Soldati, 2007; Kolonko et al., 2014; Cardenal-Muñoz et al., 2017)], and assuming that the presence of Rab7 on the MCV (Cardenal-Muñoz et al., 2017) should ensure fusion with late endosomes and lysosomes (Zhang et al., 2009), the MCVs containing wt bacteria are devoid of lysosomal enzymes and degradative activity, as demonstrated by the lack of cathepsin D and proteolysis-dependent DQ-Green BSA fluorescence (Hagedorn and Soldati, 2007; Cardenal-Muñoz et al., 2017). This suggests that the ESX-1-induced membrane damage make MCVs leaky for hydrolytic enzymes. This MCV leakage, together with the retrieval of the vATPase, and possibly hydrolases, by WASH- and Arp2/3-dependent polymerization of actin[see section Early M. marinum Infection Phase (∼0–12 hpi) and **Figure 2**], would affect both autophagic flux and killing of mycobacteria. Interestingly, depletion of hydrolases from the acidic Salmonella-containing compartment (SCV) has already been described, although the mechanism is different. Salmonella subverts the Rab9-dependent retrograde trafficking of mannose-6-phosphate receptors in human cells, which results in fusion of the SCV with lysosomes that are reduced in hydrolytic enzymes (McGourty et al., 2012).

The membrane damage inflicted by Salmonella induces an amino acid starvation response in the host cell that inhibits TOR and induces autophagy and membrane repair, leading to the normalization of the amino acid levels and the reactivation of TOR at the surface of SCVs (Tattoli et al., 2012a,b; Kreibich et al., 2015). Interestingly, it has been recently shown that mammalian TRIM16, a member of the TRIM family, recognizes the endomembrane perturbations induced by Mtb ESX-1 and affects TORC1 activity, inducing autophagy to protect macrophages from damage (Chauhan et al., 2016). The role of autophagy in mediating membrane repair during D. discoideum infection with M. marinum has also been recently proposed (Cardenal-Muñoz et al., 2017). After downregulation of TORC1 and induction of the autophagic response mentioned above, D. discoideum Lst8 (a TORC1 component) and two TORC1 activators, the Rheb GTPase and Lamtor1 [a member of the guanine nucleotide exchange factor (GEF) Ragulator complex] localize to damaged MCVs coinciding with reactivation of TORC1 at 7 hpi (Cardenal-Muñoz et al., 2017) (**Figure 2**). In addition, EM inspection of infected amoebae lacking the major autophagy initiator Atg1 (ULK in humans) (Mesquita et al., 2015) reveals that most bacteria escape very early to the cytosol (1 hpi), followed by a striking increased load of highly ubiquitinated cytosolic bacteria that cannot be degraded by xenophagy. This suggests that the autophagy pathway of D. discoideum is required for the repair of the MCV membranes, preventing the escape of mycobacteria to the cytosol. It remains unclear why M. marinum accelerates its growth only after 24 hpi in the atg1-amoebae (Cardenal-Muñoz et al., 2017). One plausible explanation is that autophagy would not only recognize damage and repair the MCV, but it could also provide M. marinum with cytoplasmic nutrients such as membranes and lipid droplets (LDs) by the fusion of autophagosomes with the MCV (see section M. marinum Exploits D. discoideum Lipids).

#### Cardenal-Muñoz et al. Dictyostelium and Mycobacterium

### *D. discoideum* Plasma Membrane Damage and Sealing at Later Time Points

Despite the function of autophagy in membrane repair during early infection, most M. marinum bacilli are able to break completely their MCV and finally access the D. discoideum cytosol (Hagedorn and Soldati, 2007). As previously described, M. marinum perforates the plasma membrane of D. discoideum to egress through an actin-based ejectosome (Hagedorn et al., 2009). The ESX1 secretion system, and more specifically secreted ESAT-6, are required for ejection of the bacteria (Hagedorn et al., 2009). This is demonstrated by restoration of ejection of the 1RD1 mutant by ectopic expression of M. marinum ESAT-6 in D. discoideum (Hagedorn et al., 2009). Despite this perforation, the integrity of the D. discoideum plasma membrane is continuously maintained during and after ejection by the autophagy machinery (**Figure 1**). This is necessary to avoid lytic cell death and to improve cell-to-cell transmission of the bacteria (Gerstenmaier et al., 2015). Accordingly, the autophagosomal markers Atg8, GFP-p62, and Ub are present at a pocket formed around the intracellular distal pole of ejecting M. marinum (Gerstenmaier et al., 2015). On the contrary, GFP-Atg18 and GFP-2xFYVE, two reporters of expanding phagophores (Calvo-Garrido et al., 2014; Mesquita et al., 2016), are less or rarely present, respectively (Gerstenmaier et al., 2015). This, together with the fact that the transcription levels of autophagy genes remain unchanged at these late stages of infection (Cardenal-Muñoz et al., 2017), suggest that M. marinum exploits preformed autophagosomes during its ejection.

The core autophagy machinery of D. discoideum (Atg1, Atg5, Atg6, and Atg7) is required for the recruitment of Atg8 to the distal pole of ejecting M. marinum (Supplementary Table 3). Strikingly, atg1- cells can form the actin-based structure of ejectosomes, but these are not functional for cell-to-cell transmission (Gerstenmaier et al., 2015). In addition, the selective receptor p62 is not required for the recruitment of Atg8 to the ejecting bacterium (Gerstenmaier et al., 2015). This, together with the fact that M. marinum grows normally within p62-cells (Cardenal-Muñoz et al., 2017), suggest that alternative autophagy receptors might exist in D. discoideum. Interesingly, the ESX-1 secretion system, which induces the recruitment of autophagic markers to bacteria escaping the MCV (Cardenal-Muñoz et al., 2017), does not play a role in the localization of Atg8 to bacteria during ejection (Gerstenmaier et al., 2015). This was shown by co-infecting cells with wt and either M. marinum 1RD1 or the non-pathogenic M. smegmatis (whose ESAT-6 is inactive in membrane disruption). In these cells both bacteria form ejectosomes and recruit Atg8 to their distal pole (Gerstenmaier et al., 2015).

### *M. MARINUM* EXPLOITS *D. DISCOIDEUM* LIPIDS

Inside their hosts, pathogens are restricted to a limited supply of nutrients. For instance, to drive proliferation, intracellular bacteria need to exploit a suitable energy source (Abu Kwaik, 2015; Abu Kwaik and Bumann, 2015). In recent years, it has become evident that Mycobacterium spp. access host lipids (Peyron et al., 2008; Caire-Brandli et al., 2014) to gain energy via β-oxidation and the glyoxylate shunt (McKinney et al., 2000; Munoz-Elias and McKinney, 2005). The importance of lipids for Mtb, but also for other Mycobacterium spp., is supported by the high number of genes (more than 250) harbored in the Mtb genome and required for lipid and fatty acid (FA) metabolism (Cole et al., 1998). The encoded proteins have diverse roles in cell wall synthesis, formation of intracytosolic lipid inclusions (ILIs) and production of energy (reviewed in Dumas and Haanappel, 2017). Strikingly, 100 of these genes exclusively encode proteins involved in the five reactions of β-oxidation (Cole et al., 1998). Interestingly, the chromosome of M. marinum is 50% larger than the one of Mtb (6.6 million base pairs compared to 4.4 million), which is also reflected by a higher number of genes involved in lipid metabolism (Stinear et al., 2008). For instance, the genome of M. marinum encodes 32 fadD acyl-CoA synthase paralog, compared with 23 in Mtb. This larger genome likely reflects the capacity of M. marinum to infect and produce a tuberculosis-like disease in a larger range of environmental hosts compared to Mtb (Stinear et al., 2000).

### Lipid Droplets Are Recruited to the MCV during the Early Infection Phase

Despite the increasing evidence pointing to a central role of lipids during Mtb infection, the basic cellular mechanisms by which pathogenic mycobacteria exploit host lipids are still poorly understood. One of the major characteristics of tuberculosis is the appearance of foamy macrophages during chronic infection. The accumulation of LDs that leads to the foamy characteristics of the respective host cells was observed during Mtb infection (reviewed in Russell et al., 2009), but also in infections with M. bovis (D'Avila et al., 2006) and M. leprae (de Mattos et al., 2012). By using D. discoideum as a model for foamy macrophages, it was recently shown that host LDs move to the MCV within minutes after phagocytosis (**Figure 3**, step 1; Barisch et al., 2015b). The clustering of host LDs at the MCV can be monitored by live microscopy using Bodipy 493/503 as a neutral lipid marker for LDs that were induced by feeding cells with exogenous FAs prior to infection. We propose that host LDs might be translocated into the MCV by a process similar to lipophagy (**Figure 3**, step 1b), a selective autophagy pathway for LDs (reviewed in Schulze et al., 2017), or by a fusion-like process similar to that observed in a macrophage/M. avium infection model system (Caire-Brandli et al., 2014) (**Figure 3**, step 2). In analogy to LD-bilayer or LD-LD fusion (Thiam et al., 2013), the phospholipid monolayer of an LD might first fuse with the outer leaflet of the MCV membrane leading to the release of the hydrophobic core lipids into the membrane bilayer of the MCV. In a second step, a new LDlike structure might bud into the lumen of the MCV (for more detailed information, please see (Barisch and Soldati, 2017a).

### Host Lipid Distribution at the *M. marinum* Proliferation Phase

At later infection stages (19 hpi), LD-like structures disappear, but neutral lipids and sterols, which can be detected by Bodipy

493/503 and filipin staining, are homogenously distributed within the MCV (**Figure 3**, step 3; Barisch et al., 2015b). Interestingly, these observations are in agreement with the fact that mycobacteria are one of the rare species able to catabolize host sterols and, more precisely, cholesterol (Pandey and Sassetti, 2008; Senaratne et al., 2008). By labeling D. discoideum with Topfluor-Lysophosphatidylcholine (Topfluor-LysoPC), a fluorescent phospholipid precursor, it was observed by live microscopy that host phospholipids accumulate at the membranes of the MCV, most likely by standard membrane trafficking (**Figure 3**, step 4; Barisch and Soldati, 2017b).

Strikingly, intravacuolar bacteria accumulate considerably more ILIs when D. discoideum cells are treated with exogenous FAs (**Figure 3**, step 5). Not only the FAs released from host triacylglycerols (TAGs) but also those from phospholipids are used by M. marinum for ILI formation (**Figure 3**, step 5; Barisch and Soldati, 2017b). This was shown by thin layer chromatography (TLC) using a D. discoideum dgat1/2 double mutant, which is depleted in both diacyglycerol Oacyltransferases Dgat1 and Dgat2. This mutant shuttles excessive lipids into phospholipids, and does not synthesize TAGs and generate LDs (Du et al., 2013; Barisch and Soldati, 2017b). Interestingly, by measuring bacteria growth with the help of high-content microscopy and a luminescence-based assay, it was observed that M. marinum growth was unaffected in the dgat1/2 double mutant (Barisch and Soldati, 2017b) leading to the conclusion that host phospholipids can successfully substitute for host TAGs. The transfer of host lipids to the intracellular pathogen can also be monitored by live-cell microscopy. To this end, host cells are labeled with Topfluor-LysoPC prior to infection. Strikingly, fluorescently-labeled host phospholipids become first enriched at the membrane of the MCV 10 min postinfection, and inside the bacteria at 21 hpi. Consequently, it was proposed that host lipids are transferred to the MCV by membrane trafficking and then further processed by bacterial or host phospholipases to facilitate uptake of FAs into the bacteria (Barisch and Soldati, 2017b).

Recently, the question whether Mtb has enzymes to successfully cleave/hydrolyse host lipids was addressed (Singh et al., 2017). The authors propose that host lipids are cleaved by Msh1, which is upregulated upon hypoxic conditions, and secreted into the host cytosol, where it hydrolyses TAGs. For M. marinum, it has been shown that the formation of ILIs depends on its type VII secretion system ESX-5, presumably thanks to the absorption of nutrients such as FAs through membrane porins or channels that are formed into the outer membrane of M. marinum by this ESX-5 apparatus (Ates et al., 2015). In the case of Mtb, FAs are transported into the bacteria via the mammalian cell entry 1 (Mce1) complex and LucA that interacts with subunits of the complex and coordinates its activity, and which is required for full virulence of Mtb in vivo (Nazarova et al., 2017).

The reason why intracellular mycobacteria accumulate ILIs during infection is so far poorly understood. One possibility is that these bacteria use ILIs as an energy source. For instance, it has been shown that in macrophages infected with M. avium and without any additional fat source, LDs are rapidly depleted from the macrophages, immediately followed by ILIs depletion from the pathogens (Caire-Brandli et al., 2014). Moreover, with the help of an in vitro dormancy model, it was proposed that M. bovis BCG uses the TAGs released from ILIs as an energy source during reactivation from dormancy (Low et al., 2009). In addition, when the synthesis of TAGs is inhibited in Drosophila melanogaster, the accumulation of host LDs induced by M. marinum infection is also prevented, and the mycobacteria-LDs association, as well as the number of intracellular viable M. marinum, are reduced (Pean et al., 2017). Furthermore, ILI-rich mycobacteria have been shown to be more tolerant to rifampicin, isoniazid, ethambutol, and ciprofloxacin (Hammond et al., 2015). Finally, an antioxidant role of LDs (or ILIs) was proposed recently (Bailey et al., 2015).

Importantly, host lipids might not only be used by these bacteria as an energy source but also as building blocks for cell wall lipids. The core of the mycobacterial cell wall consists of mycolic acids that are esterified to arabinogalactan (AGs) polysaccharides, which are in turn covalently attached to a peptidoglycan backbone (reviewed in Brennan, 2003; Jackson, 2014). This core structure serves as anchor for extractible "free lipids" such as PDIMs and phenolic glycolipids (PGLs). One unique feature of pathogenic Mycobacterium spp. is their antibiotic resistance that is conferred by the "mycomembrane," the impermeable lipid bilayer that is formed by the mycolic acids and the free lipids (Jackson, 2014). During infection, PDIMs are involved in Mtb uptake (Astarie-Dequeker et al., 2009), block of phagosomal maturation and acidification arrest (Pethe et al., 2004; Stewart et al., 2005), in escape to the cytosol and egress from the host cells (Quigley et al., 2017). In addition, PDIMs are essential for the multiplication and persistence of Mtb in the lungs of infected mice (Cox et al., 1999), and Mtb strains inhibited in PDIM synthesis are more susceptible to killing by an early innate host response (Day et al., 2014). Corroborating this, it has been demonstrated that FAs released from host LDs are used by Mtb to synthesize PDIMs (Lee et al., 2013).

PDIMs are also implicated in the virulence of M. marinum. Like Mtb in macrophages, M. marinum synthesizes PDIMs (and TAGs) from FAs released from D. discoideum LDs and phospholipids (Barisch and Soldati, 2017b). To monitor lipid transfer from the host to the pathogen, D. discoideum lipids can be labeled by using fluorescent FAs such as Bodipy 558/568 C12 or Topfluor-LysoPC prior to infection (Barisch and Soldati, 2017b). The incorporation of the fluorescent label into D. discoideum and M. marinum lipids can be detected by TLC and fluorescent lipid species can be identified with the help of fluorescent standard lipids.

A transposon insertion in the M. marinum tesA gene, encoding a putative type II thioesterase, leads to an altered cell wall without PDIMs and PGLs and the subsequent lack of the permeability barrier (Alibaud et al., 2011). As a consequence, the M. marinum tesA mutant is attenuated in D. discoideum and in zebrafish embryos (Alibaud et al., 2011). Inhibitor and gene mutations altering the PDIMs and PGLs of M. marinum affect its virulence in D. discoideum and adult zebrafish, respectively (Yu et al., 2012; Ouertatani-Sakouhi et al., 2017). In addition, Cambier et al., elegantly described how Mtb and M. marinum preferentially recruit and infect permissive macrophages while microbicidal ones are evaded (Cambier et al., 2014). This bypass of the innate immune system was proposed to be the result of cell-surface-associated PDIMs hiding underlying pathogenassociated molecular patterns (PAMPs).

### LD Proteins Target Cytosolic *M. marinum* at Late Infection Phases

At later infection stages (42 hpi), LD proteins such as the D. discoideum homolog of perilipin as well as Dgat2 localize to cytosolic M. marinum (**Figure 3**, steps 6 and 7; Barisch et al., 2015b; Barisch and Soldati, 2017b). Perilipin belongs to the class II of LD proteins, which bind the LD surface through hydrophobic and amphipatic domains, and circulates between LDs and the cytosol. During infection, perilipin is proposed to partition into the waxy mycobacterial cell wall in a similar way, via its amphipathic and hydrophobic regions (Barisch and Soldati, 2017a). Importantly, by using a luminescencebased assay to measure intracellular M. marinum growth, it became evident that in a D. discoideum perilipin (plnA) mutant, bacteria growth is inhibited starting from 20 hpi, the time at which M. marinum reaches the cytosol (Barisch et al., 2015b). Consequently, it was proposed that perilipin might exert a barrier function, protecting the M. marinum cell wall during the cytosolic phase of the infection from degradative host processes such as neutral lipases.

Dgat2 belongs to the class I of LD proteins, which bind the LD surface through an hydrophobic hairpin motif, and is exclusively localized at LDs. Consequently, Dgat2 is transferred to the waxy cell wall of M. marinum by coalescence with LDs and lateral diffusion (Barisch and Soldati, 2017b).

#### THE END OF THE ROMANCE: USING *D. DISCOIDEUM* TO UNCOVER THERAPEUTIC TARGETS IN *M. MARINUM*

### *D. discoideum* Plaque Assays

As for animals, M. marinum is pathogenic for D. discoideum. Contrary to the food bacterium Klebsiella, the growth of D. discoideum is restricted when plated on M. marinum lawns (Ouertatani-Sakouhi et al., 2017). Researchers have made use of this growth incapacity to perform genetic and chemical screens for the search of virulence genes and anti-mycobacterial compounds. For instance, M. marinum TesA, a putative type II thioesterase required for the synthesis of the cell wall lipids PDIMs and PGLs, was identified in a D. discoideum plaque assay where a M. marinum transposon mutant library was screened for bacterial attenuation. The role of TesA in virulence was further confirmed during infection of zebrafish embryos, and the tesA gene was proposed as a possible genetic target for disruption in human mycobacterial pathogens (Alibaud et al., 2011). In addition, another plaque assay screen of M. marinum transposon mutants allowed Chen and collaborators to discover that two genes, mmar\_2318 and mmar\_2319, involved in the biosynthesis of lipooligosaccharide (LOS) contribute to M. marinum virulence in D. discoideum and limit bacterial entry in human macrophages (Chen et al., 2015) (Supplementary Table 2). Chemical screens can also be performed in amoebae (Ouertatani-Sakouhi et al., 2017; Trofimov et al., in press). For instance, drugs targeting the M. marinum cell wall have been recently identified to block/decrease M. marinum virulence by inhibiting bacterial aggregation and permeability, sliding motility and/or intracellular replication within D. discoideum (Ouertatani-Sakouhi et al., 2017). All these findings highlight D. discoideum as an alternative host for the screening of Mycobacterium virulence factors and potential antibacterial compounds.

#### High throughput Screening

Since the advent of next-generation sequencing (NGS) technologies, numerous methods to investigate the hostpathogen interaction at different levels (i.e., RNA, DNA, and proteins) have been developed. Among them, RNA sequencing (RNA-Seq) is becoming the standard to analyze the transcripts of each actor involved in a bacterial infection, either independently (bacteria or host) or simultaneously (bacteria and host) by dual RNA-Seq (Saliba et al., 2017; Westermann et al., 2017). This method has been used to describe the gene expression profile of M. marinum during exponential and stationary growth (Wang et al., 2013), as well as under stress conditions (Pettersson et al., 2015). In addition, it has allowed to determine the contribution of specific cells (neutrophils and macrophages) to the innate immune response against M. marinum, Mtb, and M. bovis (Schnappinger et al., 2003; Nalpas et al., 2015; Kenyon et al., 2017) and to identify genes such as whiB4, which encodes a protein regulating the PE/PPE genes and is required for M. marinum and Mtb virulence (Chawla et al., 2012; Wu et al., 2017). Moreover, a recent metabolomics study coupled to dual RNA-Seq allowed to reconstruct the dietary map of Mtb during macrophage infection (Zimmermann et al., 2017). Since the RNA-Seq methodology has already been established in D. discoideum (Miranda et al., 2013), it will be of high interest to perform dual RNA-seq in the D. discoideum—M. marinum infection model.

Another NGS method is transposon sequencing (Tn-Seq), which allows the analysis of fitness and genetic interactions in microorganisms by using random transposon insertion libraries in bacteria (van Opijnen et al., 2009; DeJesus and Ioerger, 2013; DeJesus et al., 2013, 2017a,b; van Opijnen and Camilli, 2013; Nambi et al., 2015) Thanks to this method, various researchers have identified essential genes for Mtb growth and cholesterol metabolism (Griffin et al., 2011; Zhang et al., 2012; DeJesus and Ioerger, 2013) and have investigated the Mtb virulence factors required for survival in human dendritic cells (Mendum et al., 2015). Concerning M. marinum, only one study has been published so far (Weerdenburg et al., 2015), in which the authors identified ∼300 genes essential for survival of the bacteria in vitro. These genes are mostly shared with Mtb and correspond to 6% of the total coding sequences. The authors investigated the strategies used by M. marinum to exploit different phagocytic cells, from protozoan and vertebrate origin. The behavior of the M. marinum transposon mutant pool was highly similar between the two mammalian cell lines (human THP-1 monocytes and mouse RAW264.7 macrophages) and between the two amoebae D. discoidum and A. castellani, while an intermediate behavior was observed in cells derived from fish (carp CLC leukocytes), probably due to their higher similarity with natural M. marinum hosts. Thus, the authors concluded that M. marinum has both conserved and host-specific virulence determinants. Regarding the conserved virulence factors, transposon interruption of PDIMs, ESX-1, or Mce1/4 leads to high attenuation of M. marinum virulence in multiple hosts. In addition, a new virulence factor of M. marinum was highlighted and validated in vivo, the cpsA gene. This gene encodes a protein of the LytR family of transcriptional regulators and has been recently demonstrated to be required for M. marinum cell wall integrity and virulence in the zebrafish model (Wang et al., 2015; Weerdenburg et al., 2015). Moreover, amoebae-specific pathways have been pointed out by the authors. For instance, transposon insertions in multiple genes involved in biosynthesis of vitamin B<sup>12</sup> highly increased fitness of M. marinum in Acanthamoeba, while they had no or little impact on the bacteria within the other host cells (Weerdenburg et al., 2015). To conclude, M. marinum uses specific mechanisms to adapt to different intracellular environments. Some specific requirements for infection of mammalian cells are not applicable for protozoa infection, and vice versa.

Finally, REMI sequencing (REMI-Seq) has recently been developed for D. discoideum by the group of C. Thompson (https://thethompsonlab.wordpress.com/mutant-library/). The REMI method was first described in Saccharomyces (Schiestl and Petes, 1991) and then adapted to D. discoideum (Kuspa and Loomis, 1992; Kuspa, 2006). The technique comprises the integration of linear DNA into eukaryotic genomes for insertional mutagenesis. Combining REMI with NGS, a genomewide collection of tagged mutants can be created to investigate host genes. However, REMI-Seq has not yet been used in the context of mycobacterial infection. Therefore, this tool, combined with other methods described above, are promising and innovative approaches that could help identify genes and pathways involved in mycobacteria-host interaction.

#### CONCLUDING REMARKS

In this review, we have described the so far known processes involved in the intracellular cycle of M. marinum within the amoeba D. discoideum. The comparison with studies carried out in human cells with Mtb revealed that the infection cycles of both bacteria share numerous similarities. However, many of the virulence mechanisms used by both pathogens to subvert the immune response of their hosts remain to be elucidated. We propose the D. discoideum—M. marinum system as a suitable and powerful model to study infection by tuberculous mycobacteria.

It needs to be noticed that not all the markers listed here were monitored in all the listed strains, and vice versa. Please see main text and supplementary tables for more details. IFA, immunofluorescence assay; TEM, transmission electron microscopy; qPCR, quantitative polymerase chain reaction; SEM, scanning electron microscopy; CLEM, correlative light and electron microscopy.

#### AUTHOR CONTRIBUTIONS

EC-M, CB, LL, and AL-J wrote the original draft. EC-M and CB drew the figures and prepared the tables. EC-M, CB, LL, and TS reviewed and edited the draft. TS supervised this work.

#### REFERENCES


#### FUNDING

The TS laboratory is supported by multiple grants from the Swiss National Science Foundation, and TS is a member of iGE3 (www. ige3.unige.ch) as well as of the COST Actions BM1203 EU-ROS and CA15138 TRANSAUTOPHAGY.

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2017.00529/full#supplementary-material


from transposon insertion sequencing data. BMC Bioinformatics 14:303. doi: 10.1186/1471-2105-14-303


M. tuberculosis via human lung epithelial cells. Mol. Microbiol. 75, 92–106. doi: 10.1111/j.1365-2958.2009.06959.x


triggers an innate host defense program. Cell Host Microbe 11, 563–575. doi: 10.1016/j.chom.2012.04.012


for survival and replication in different hosts. Infect. Immun. 83, 1778–1788. doi: 10.1128/IAI.03050-14


**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 Cardenal-Muñoz, Barisch, Lefrançois, López-Jiménez and Soldati. 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) or licensor 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.

# Differential Effects of Iron, Zinc, and Copper on Dictyostelium discoideum Cell Growth and Resistance to Legionella pneumophila

Simona Buracco<sup>1</sup> , Barbara Peracino<sup>1</sup> , Claudia Andreini <sup>2</sup> , Enrico Bracco<sup>3</sup> and Salvatore Bozzaro<sup>1</sup> \*

<sup>1</sup> Department of Clinical and Biological Sciences, University of Torino, Turin, Italy, <sup>2</sup> Magnetic Resonance Center (CERM), University of Florence, Florence, Italy, <sup>3</sup> Department of Oncology, University of Torino, Turin, Italy

Iron, zinc, and copper play fundamental roles in eucaryotes and procaryotes, and their bioavailability regulates host-pathogen interactions. For intracellular pathogens, the source of metals is the cytoplasm of the host, which in turn manipulates intracellular metal traffic following pathogen recognition. It is established that iron is withheld from the pathogen-containing vacuole, whereas for copper and zinc the evidence is unclear. Most infection studies in mammals have concentrated on effects of metal deficiency/overloading at organismal level. Thus, zinc deficiency or supplementation correlate with high risk of respiratory tract infection or recovery from severe infection, respectively. Iron, zinc, and copper deficiency or overload affects lymphocyte proliferation/maturation, and thus the adaptive immune response. Whether they regulate innate immunity at macrophage level is open, except for iron. The early identification in a mouse mutant susceptible to mycobacterial infection of the iron transporter Nramp1 allowed dissecting Nramp1 role in phagocytes, from the social amoeba Dictyostelium to macrophages. Nramp1 regulates iron efflux from the phagosomes, thus starving pathogenic bacteria for iron. Similar studies for zinc or copper are scant, due to the large number of copper and zinc transporters. In Dictyostelium, zinc and copper transporters include 11 and 6 members, respectively. To assess the role of zinc or copper in Dictyostelium, cells were grown under conditions of metal depletion or excess and tested for resistance to Legionella pneumophila infection. Iron shortage or overload inhibited Dictyostelium cell growth within few generations. Surprisingly, zinc or copper depletion failed to affect growth. Zinc or copper overloading inhibited cell growth at, respectively, 50- or 500-fold the physiological concentration, suggesting very efficient control of their homeostasis, as confirmed by Inductively Coupled Plasma Mass Spectrometry quantification of cellular metals. Legionella infection was inhibited or enhanced in cells grown under iron shortage or overload, respectively, confirming a major role for iron in controlling resistance to pathogens. In contrast, zinc and copper

#### Edited by:

Thierry Soldati, Université de Genève, Switzerland

#### Reviewed by:

Olivier Neyrolles, Centre National de la Recherche Scientifique (CNRS), France Hubert Hilbi, University of Zurich, Switzerland

> \*Correspondence: Salvatore Bozzaro salvatore.bozzaro@unito.it

Received: 31 October 2017 Accepted: 26 December 2017 Published: 11 January 2018

#### Citation:

Buracco S, Peracino B, Andreini C, Bracco E and Bozzaro S (2018) Differential Effects of Iron, Zinc, and Copper on Dictyostelium discoideum Cell Growth and Resistance to Legionella pneumophila. Front. Cell. Infect. Microbiol. 7:536. doi: 10.3389/fcimb.2017.00536 depletion or excess during growth did not affect Legionella infection. Using Zinpyr-1 as fluorescent sensor, we show that zinc accumulates in endo-lysosomal vesicles, including phagosomes, and the contractile vacuole. Furthermore, we provide evidence for permeabilization of the Legionella-containing vacuole during bacterial proliferation.

Keywords: Dictyostelium discoideum, Legionella pneumophila, host-pathogen interactions, phagocytosis, Nramp1, NrampB, divalent metals, Zinpyr-1

#### INTRODUCTION

Transition metals, such as iron, zinc, copper, or manganese, play fundamental roles in many biological processes in both procaryotes and eucaryotes. Thanks to their ability to easily shift between different oxidation states, they act as co-factors of enzymes, and it is estimated that 30–45% of known enzymes are metalloproteins containing one of these metals (Andreini et al., 2008; Waldron et al., 2009). Among the transition metals, iron is the most abundant in living organisms. By shifting between the ferric and the ferrous forms, iron catalyzes redox reactions that are essential for processes such as respiration, oxygen transport, metabolic energy production, and gene regulation (Andreini et al., 2008; Pantopoulos et al., 2012). Whereas, iron is mostly present in the environment as insoluble ferric form, copper is easily bioavailable in the soluble Cu2<sup>+</sup> cupric form (Festa and Thiele, 2011). Similarly to iron, copper is involved in redox reactions, regulating the activity of enzymes, such as the respiratory chain cytochrome c oxidase and the Cu-Zn superoxide dismutase (Tapiero and Tew, 2003; Rubino and Franz, 2012). Zinc is the second most abundant transition metal in living organisms after iron, is incorporated in about 10% of human proteins and necessary for over 300 enzymes (Andreini et al., 2011). In contrast to iron and copper, zinc is redox-inert, but has many structural and catalytic roles, stabilizing negative charges of the substrates or organizing protein subdomains in zinc motifs (Tapiero and Tew, 2003; Cerasi et al., 2013; Maret, 2013).

Excess iron, copper, or zinc are toxic, as they perturb the redox potential, producing highly reactive hydroxyl radicals (copper and iron), bind to sulfide and thiol groups, thus destabilizing iron-sulfur clusters (copper and zinc), or interfere with the metabolism of other ions, displacing them from their binding proteins (Letelier et al., 2005; Dupont et al., 2011; Imlay, 2014). Thus, cells have developed complex mechanisms to regulate import, export, and storage of transition metals (Stafford et al., 2013; Bird, 2015; Weiss and Carver, 2017).

Tight regulation of transition metal bioavailability is also a vital part of host-pathogen interactions. In bacteria, transition metals are involved in metabolism and regulation of virulence as a mechanism of host invasion, and many opportunistic pathogens, such as pathogenic strains of Escherichia or Klebsiella, Mycobacteria, Salmonella, or Legionella, have developed sophisticated sensing, uptake, and export mechanisms to accumulate transition metals according to their physiological needs (Graham et al., 2009; Dupont et al., 2011; Rowland and Niederweis, 2012; Porcheron et al., 2013; Braymer and Giedroc, 2014; Neyrolles et al., 2015; Skaar and Raffatellu, 2015; Capdevila et al., 2016).

The source of transition metals for intracellular pathogens is the cytoplasm of the host cell, which in turn can manipulate metal uptake and intracellular traffic following pathogen recognition. Several lines of evidence support the notion that iron and manganese are withheld from the phagosome, or the pathogencontaining vacuole, to prevent reconstruction by the engulfed bacteria of Fe-S clusters and the use of Mn2<sup>+</sup> as a protectant against reactive oxygen species (Weinberg, 2000; Kehres and Maguire, 2003; Nairz et al., 2010; Lisher and Giedroc, 2013). Iron efflux from phagosomes by the Nramp1 transporter, both in mammalian macrophage and the lower professional phagocyte Dictyostelium discoideum, is such a paradigmatic case of starvation strategy, or so-called "nutritional immunity" (Appelberg, 2006; Cellier, 2012; Bozzaro et al., 2013a). In contrast to iron, copper, and zinc are presumably used as a mechanism to poison bacterial pathogens (Wagner et al., 2005; White et al., 2009; Botella et al., 2012; Soldati and Neyrolles, 2012; Djoko et al., 2015), though zinc sequestration as antibacterial weapon has also been described (Stafford et al., 2013; Vignesh et al., 2013; Djoko et al., 2015; Besold et al., 2016). Copper withdrawal has been shown to play a role for resistance to fungal pathogens, raising the possibility that it may be used also for bacteria (Besold et al., 2016). Except for iron, most studies with mammals provide only indirect evidence of the role of copper or zinc in immune defense, while an understanding of the underlying processes at macrophage level is lacking or very limited (White et al., 2009). Indeed, most infection studies have concentrated on microbial mechanisms of defense against metal poisoning or on the effects of metal deficiency or overload at organismal level, both in laboratory animals and in epidemiological studies with humans (Prohaska and Lukasewycz, 1981; Chaturvedi and Henderson, 2014; Prasad, 2014; Pasricha and Drakesmith, 2016). These studies have shown the importance of zinc or copper for several physiological processes, including proper development of the immune system, lymphocyte differentiation, and adaptive immunity (Bonaventura et al., 2015; Weiss and Carver, 2017). Their role in cell-autonomous defense mechanisms, i.e., innate immunity, at the level of monocytes or macrophages, is less clear. It has been shown that both free zinc and copper accumulate in endosomal vesicles, including phagosomes, in response to inflammatory signals and mycobacterial infection (Wagner et al., 2005; Botella et al., 2011), and that silencing of the Cu-ATPase ATP7A impairs macrophage bactericidal activity (White et al., 2009). Enhanced expression of a mycobacterial P-type ATPase, CtpC, which favors zinc efflux, has been also reported, with a ctpc-null mutant growing poorly in macrophages (Botella et al., 2011).

For assessing the role of zinc or copper, macrophage cell lines or explanted monocytes could be grown in defined serum-free media, which contain, however, traces of transition metals in addition to proprietary components. Addition of extracellular or intracellular metal chelators to circumvent this limitation often fails, as they are not specific for a given metal and may have other unknown effects, stripping metals from exposed cellular proteins (Kay, 2004). A valid alternative is the use of singlecelled amoebae or protozoa, which are the natural hosts of many bacterial pathogens that occasionally infect animals or humans. It is generally recognized that the ability of many bacterial pathogens to grow in macrophages and cause animal or human diseases is a consequence of their adaptation and survival in the normally hostile amoeboid niche, which is the training ground for evolution of virulence traits (Greub and Raoult, 2004; Casadevall, 2008; Cosson and Soldati, 2008; Salah et al., 2009; Erken et al., 2013; Tosetti et al., 2014). An established amoeboid model organism for phagocytosis and host-pathogen interactions is the social amoeba D. discoideum. Dictyostelium cells are freeliving soil amoebae that grow by engulfing and digesting bacteria, and as such they are potential hosts of pathogens (Bozzaro et al., 2008, 2013a; Cosson and Soldati, 2008). Being haploid and amenable to molecular genetic techniques, Dictyostelium offers many advantages for identifying and characterizing host genes involved in resistance to pathogens (Bozzaro and Eichinger, 2011). Studies in the last decade have shown that Dictyostelium cells share with mammalian macrophages not only the basic phagocytic machinery, but also many mechanisms of innate and nutritional immunity (Bozzaro et al., 2008, 2013a; Cosson and Soldati, 2008; Soldati and Neyrolles, 2012; Nasser et al., 2013; Gaudet et al., 2016). Concerning transition metals, Dictyostelium cells share with macrophages the expression of the Nramp1 iron transporter in the phagolysosome, which is essential for proton-driven iron efflux from the phagosome, thus potentially starving bacteria for iron and manganese (Forbes and Gros, 2003; Courville et al., 2006; Peracino et al., 2006; Buracco et al., 2015). In agreement with this function, nramp1 KO mutants display increased susceptibility to infection by L. pneumophila and M. avium (Peracino et al., 2006). L. pneumophila was also shown to hinder H<sup>+</sup> V-ATPase, but not Nramp1, recruitment to the Legionella-containing macropinosome, thus manipulating Nramp1 activity (Peracino et al., 2010). Dictyostelium is also unique among amoebae and protozoa, for encoding in the genome a second Nramp protein, NrampB (formerly Nramp2), belonging to the prototypical Nramp family (Courville et al., 2006; Peracino et al., 2013). NrampB, is expressed in the membrane of the contractile vacuole, and, together with Nramp1, appears to regulate iron homeostasis by transporting iron across the membrane of the contractile vacuole. Mutants defective in NrampB display also increased susceptibility to Legionella, suggesting that perturbation of iron homeostasis affects hostpathogen interactions (Bozzaro et al., 2013a).

The Dictyostelium genome encodes three SLC31 (CTR) copper transporters, and three P-type Cu-ATPases, one of which is a homolog of the human ATP7A P-type ATPase (The Dictyostelium webpage: http://www.dictybase.org). Both ATP7A and the CTR protein p80 are localized in the plasma membrane and transitorily in phagosomes (Ravanel et al., 2001; Burlando et al., 2002; Hagedorn and Soldati, 2007). ATP7A activity in the plasma membrane is apparently responsible for the refractoriness of Dictyostelium cells to high copper concentrations in medium (Burlando et al., 2002; Balbo and Bozzaro, 2008), whereas its transient recruitment to the phagosomal membrane points to a potential involvement in pumping copper in the phagosomal lumen, favoring a potential toxic effect of this metal on bacteria (Hao et al., 2016). The p80 copper transporter could, instead, be involved in copper efflux from the phagosome, but no functional studies have been done in this regard. The zinc transporter family includes 11 members, with seven ZIP and four ZNT family members (Sunaga et al., 2008; The Dictyostelium webpage: http://www. dictybase.org), but no data are available on their localization in phagosome and their potential involvement in host-pathogen interactions.

To assess a role for zinc or copper in Dictyostelium phagocytosis and defense mechanisms against bacterial pathogens, and given the large number of transporters for these metals, we have followed in this paper a holistic approach, based on cultivation of wild type cells or Nramp1 knockout mutant in a minimal medium depleted of, or overloaded with either zinc, copper, or iron. The rationale is that extensive growth in media deprived of or with high content of a given metal, should results in either metal deficiency or overload in cells, thus potentially altering their resistance to pathogens. We show here that iron, but not copper or zinc deficiency affects cell growth. Metal excess results in inhibition of cell growth, with the cells being much more sensitive to iron than to copper or zinc overloading. Cells grown under iron depleting or overloading conditions, show, respectively, resistance to or exacerbation of L. pneumophila infection, whereas Legionella intracellular growth is unaltered by depletion or overloading with copper or zinc. Thus, it appears that zinc and copper, in contrast to iron, play a minor role in Dictyostelium resistance against Legionella. We provide also evidence for very tight regulation of copper and zinc homeostasis in Dictyostelium, which can in part explain these results, and for accumulation of free zinc ions in vesicles of the endo-lysosomal pathway and in the contractile vacuole.

#### MATERIALS AND METHODS

#### Cell and Bacterial Strains and Culture Methods

Dictyostelium discoideum parental strain AX2 and the Nramp1- KO mutant were used. The Nramp1-KO mutant or cells producing Nramp1-RFP and NrampB-RFP were generated previously in the lab (Peracino et al., 2006; Buracco et al., 2015). AX2 cells producing calnexin-GFP (Müller-Taubenberger, 2001) and the 389-2 vector for the expression of mRFPmars (Fischer et al., 2004) were provided by Annette Müller-Taubenberger. For generating AX2 cells producing CshA-RFP, the cshA gene was cloned into the EcoR1 site of the 389-2 vector (C-terminal mRFPmars). AX2 cells producing CshA-RFP or RFP alone were generated by electroporation (Pang et al., 1999), and transformants were selected on plates in nutrient medium containing 10µg/ml G418. All strains were cultured axenically in AX2 axenic medium (AX2M) (Watts and Ashworth, 1970) or in the growth media listed in **Table 1** (Franke and Kessin, 1977). In all cases, cells were grown in Erlenmayer flasks under shaking at 150 rpm at 22◦C on a gyratory shaker in a climatic cabinet (Kueh ner, Basel, Switzerland) as previously described (Peracino et al., 2013). Blasticidin at the concentration of 10µg/ml was added to the Nramp1-KO mutant. Cells producing GFP- or RFP-fused proteins were cultured in the presence of 10–30µg/ml G418.

Escherichia coli B2 strain, Klebsiella aerogenes, and Salmonella typhimurium (ATCC number 14028) were used. Pre-cultures of bacteria were maintained in LB agar.

Legionella pneumophila Corby producing GFP or Ds-red Express were provided by Michael Steinert and grown on buffered charcoal yeast extract agar (BCYE; 10 g/l ACES, 10 g/l yeast extract, 2g/l activated charcoal powder, 15 g/l agar, 3.3 mM L-cysteine, 0.6 mM Fe(NO3)<sup>3</sup> (Feeley et al., 1979) supplemented with 5µg/ml chloramphenicol. Bacteria were incubated for 72 h at 37◦C and 5% CO<sup>2</sup> (Fajardo et al., 2004; Peracino et al., 2013).

#### Cell Growth Assays

Exponentially growing AX2 cells were washed twice in 0.017 M Soerensen Na/K phosphate buffer (pH 6.0) and resuspended at a final concentration of 5 × 10<sup>4</sup> cells/ml in the different media listed in **Table 1**. The required metal concentration was added from a stock solution of the salt dissolved in water and filtered. For iron, the stock solution was freshly prepared to avoid formation of precipitates. Cells were grown under shaking and growth was assessed daily over a period of 1 to 3 weeks, when required by subculturing the cells in the same medium shortly before they reached the stationary phase. Cell number was counted using a hemocytometer Bürker chamber. For some experiments, growth was also assessed in M2 and M3 medium with addition of 50, 100, or 500µM zinc extracellular chelator 2- {[Bis(2-pyridinylmethyl)amino]ethylamino}benzenesulfonic

acid hydrate sodium salt ZX1 (Strem Chemicals Inc.,



\*As indicated in the recipe of the FM medium defined in Franke and Kessin (1977). M1 to M3: medium composition as in FM, except for the metal indicated.

Newburyport, USA) (Pan et al., 2011) or 50, 100, or 500µM of copper extracellular chelator trientine hydrochloride (Sigma-Aldrich, St Louis, MO, USA).

### Cell Development on Non-nutrient Agar Plates

AX2 cells were grown for 24 h in M1 ± Fe medium or for 480 h in M2 ± Zn or M3 ± Cu medium. Cells were then harvested from culture media, washed twice in Soerensen phosphate buffer and resuspended in the same buffer a concentration of 1 × 10<sup>7</sup> cells/ml. Aliquots of 50 µl were plated on agar plates buffered with Soerensen phosphate buffer at a density of about 3 × 10<sup>5</sup> cells per cm<sup>2</sup> , and incubated at 23◦C (Peracino et al., 2013). Images were acquired using a Wild M3Z Stereomicroscope (Wild Heerbrugg, Switzerland) supplied with a MicrOcular 3.0 MP Electronic Eyepiece (Bresser GmbH, Rhede, Germany).

### Cell Growth on Bacteria

A single colony of bacteria was picked from plate and bacteria were grown in LB medium overnight with shaking. The next day, 50 µl of the bacterial culture were spread in each well of a 24-well N-agar (1 g of peptone, 1 g of glucose, 15 g of agar in 1 l of Soerensen phosphate buffer) for E.coli B/2 and S. typhimurium, or SM-agar (10 g of bactopeptone, 1 g of yeast extract, 10 g of glucose, 1 g of MgSO<sup>4</sup> 7(H2O), 2.5 g of KH2PO4, 1 g of Na2HPO<sup>4</sup> 2(H2O), 18 g of agar in 1 l of water) for K. aerogenes, and further grown overnight at 22◦C. AX2 cells were grown as previously described in the minimal media listed in **Table 1**. Cells were then washed twice in Soerensen phosphate buffer and serial dilutions were plated on the dry bacterial lawn, as described in Froquet et al. (2008). Pictures of the appearance and widening of growth plaques were acquired daily with a scanner.

#### Legionella pneumophila Uptake and Infection Assays

AX2 or Nramp1-KO cells grown in the media listed in **Table 1** were washed twice in Soerensen phosphate buffer containing 50µM CaCl2, resuspended in the corresponding culture medium without glucose containing 5µg/ml chloramphenicol, and 1.0 × 10<sup>5</sup> cells were plated in 96-well tissue culture plates. For every condition, cells were tested in triplicates for a series of time points ranging from 0 to 96 hpi. As control, AX2 cells grown in axenic medium and resuspended in low fluorescence medium (LoFlo Medium supplemented with yeast extract, Formedium, Norfolk, UK) containing 5µg/ml chloramphenicol were used. Freshly collected L. pneumophila Corby producing GFP were added to Dictyostelium cells at a MOI (Multiplicity Of Infection) of 1:1 and immediately centrifuged at 600 g for 10 min at room temperature to synchronize infection. Plates were incubated at 25◦C and samples were collected every 24 h (Hägele et al., 2000; Skriwan et al., 2002). Uninfected cells were used as a control. Flow cytometry analysis was performed according to Tiaden et al. (2013), using a CyAn ADP Analyzer (Beckman Coulter, Brea, CA, USA). Cells were identified based on forward and side scatter parameters. GFP fluorescence was measured in the FL1 channel (excitation wavelength: 488 nm; emission: 530–540 nm), and the mean fluorescence was determined for at least 10,000 cells. For some experiments, 25µM membrane-permeable zinc chelator N,N,N′ ,N′ -Tetrakis(2-pyridylmethyl)ethylenediamine (TPEN) (Sigma-Aldrich, St Louis, MO, USA) was added to the infection medium.

For the uptake assay, cells were grown as described and infected with L. pneumophila Corby producing GFP at a MOI of 10:1. After centrifugation, cells were detached and analyzed by flow cytometry at 0 and 40 min post-infection. For every time, duplicates were tested and the percentage of fluorescent cells was recorded.

Data analysis was carried out using FlowJo software (FlowJo LLC, Oregon USA).

### In Vivo Microscopy and Fluorescence Imaging

Cells infected with L. pneumophila producing GFP or Ds-red Express were plated in black 96-well tissue culture µ-plates (ibidi GmbH, Planegg/Martinsried, Germany) and incubated at 25◦C. Confocal series images were taken daily on an inverted Zeiss LSM800 with AiryScan (Carl Zeiss, Inc., Oberkochen, Germany) equipped with a Plan-Apochromat 63x/1.40 DIC Oilimmersion objective. For excitation, a 488 nm Diode laser was used and its emission collected with a 495–519 nm filter for GFP and a 592–614 nm filter for RFP. Phase contrast was recorded simultaneously.

For monitoring zinc intracellular localization, cells producing Nramp1 or NrampB-RFP or AX2 cells mixed with TRITClabeled E. coli (MOI 100:1) were used. Cells were washed and resuspended in Soerensen phosphate buffer and incubated for 30 min with 5µM Zinpyr-1 fluorescent probe (Abcam, Cambridge, UK). Non-incorporated probe was removed by washing with Soerensen buffer and the cells were plated on glass coverslips. Images were acquired as previously described.

### Metal Analysis by Inductively Coupled Plasma Mass Spectrometry (ICP-MS)

AX2 cells were grown for 24 h in M1 ± Fe medium or for 480 h in M2 ± Zn or M3 ± Cu media. A total of 10<sup>8</sup> cells were washed five times in Soerensen phosphate buffer and pelleted in Eppendorf tubes. Digestion was performed by adding 1 ml of concentrated HNO<sup>3</sup> (70%) to each sample. After complete dissolution, samples were further digested by applying microwave heating (Milestone MicroSYNTH, Microwave labstation equipped with an optical fiber temperature control and HPR-1000/6M six position highpressure reactor, Bergamo, Italy). After digestion, the volume of each sample was brought to 3 ml with ultrapure water, filtered with 0.45 mm filter and analyzed by ICP-MS, using a Thermo Scientific ELEMENT 2 ICP-MS (Finnigan, Rodano, Italy) (Fiorito et al., 2012). The same analysis was also performed on 800 µl of M2 or M3 medium. The quantification was obtained through a calibration curve measured by using six Fe/Cu/Zn absorption standard solutions (Sigma-Aldrich) in the range 0.0025–0.3µg/ml. Samples with metals concentrations higher than the upper limit were diluted opportunely. Sample digestion and metal quantifications were carried out by the facility of the Molecular Imaging Center (Department of Chemistry, University of Torino, Italy).

### RESULTS

### Effects of Transition Metal Depletion or Overloading on Growth of Wild Type AX2 Cells

We first investigated how alterations in the availability of iron, copper, or zinc would affect Dictyostelium cell growth. This was achieved by evaluating AX2 wild type cell growth in a defined FM minimal medium (Franke and Kessin, 1977), in which the concentrations of single metals was modified. The FM minimal medium contains 100µM FeCl3, 8µM ZnSO4, and 0.6µM CuSO<sup>4</sup> (Franke and Kessin, 1977). We omitted or added to the original recipe one of these three metals at varying concentrations, as summarized in **Table 1**. AX2 cells exponentially growing in AX2 medium (AX2M) were diluted in these minimal media to an initial concentration of 5 × 10<sup>4</sup> cells/ml, and growth was evaluated under shaking for 1 to 3 weeks. As previously reported, cell growth is optimal at an iron concentration of 100µM, but inhibited after a few generations if iron is omitted or the amount increased to 200µM (**Figure 1A** and Peracino et al., 2010). In sharp contrast, cell growth up to 10<sup>7</sup> /ml was unaffected in M2 or M3 medium or with up to 20 or 200-fold excess of zinc or copper, respectively, compared to control FM (**Figures 1B,C**). Higher concentrations of copper led to a delay in the rate of cell duplication and total inhibition with 1,000-fold excess (**Figure 1C**). In the case of zinc, a toxic effect was observed with 50 or 100-fold excess (**Figure 1B**).

To assess whether cell growth would be affected by prolonged incubation in medium without added zinc or copper or with high, but non-toxic, concentrations of zinc or copper, namely 20- or 200-fold excess, respectively, cells were re-diluted in the same medium, shortly before reaching the stationary phase, and cell growth measured over 480 h. As shown in **Figure 1D**, no significant changes in the rate of cell doubling time was observed. Thus, Dictyostelium cells appear to be very efficient in controlling high concentrations of copper or zinc. More surprising is that their growth rate is unaffected by depletion of either metal. A preliminary analysis of the metal proteome, by using a modified metal predator as bioinformatic tool (Valasatava et al., 2016), shows that the Dictyostelium genome encodes 840 zinc- and 250 iron-binding proteins, corresponding to 6.6 and 2%, respectively, of the entire proteome. There are fewer copperbinding proteins: about 40. Especially for zinc, one would expect that zinc shortage in the medium would have detrimental effects for the activity of the zinc-binding proteins, many of which are involved in basic cellular processes, such as transcription, translation, cell replication, and protein turnover. We thus tested whether incubation for 480 h, i.e., about 24 generations, in M2 or M3 medium led to a shrinkage in cell volume, but no significant changes in size with control cells were found (M2: 4.8 ± 0.9 µl; M3: 5.2 ± 1.5 µl; FM: 4.7 ± 1.8 µl, in all cases per 10<sup>7</sup> cells). Addition to M2 or M3 medium of extracellular zinc chelator

ZX1 or copper chelator trientine hydrochloride at concentrations varying between 50 and 500µM also failed to affect cell growth.

#### Copper and Zinc Concentration in Cells Growing in Medium Depleted or Overloaded with Metals

Presuming there are no other sources of these metals in the medium, at any cell doubling the intracellular amount of iron, zinc, or copper should halve. Initially, cells contain the metals engulfed with the axenic medium and might undergo some duplication rounds, as observed for iron (**Figure 1A**), but this does not explain the ability of the cells to sustain more than 20 generations in the absence of external zinc or copper (**Figures 1B,C**). However, even in minimal medium without added metals, metal contamination cannot be excluded, particularly due to amino acids. Indeed most amino acids contain up to 0.5–5 ppm iron, zinc or copper, based on manufacturer sheets. In addition, metals can leach from the glassware and, depending on their composition, from some plastics used for cell culture and treatment (Kay, 2004).

To get an unbiased measure of the amount of zinc or copper in medium and in cells, we measured their concentrations by using Inductively Coupled Plasma Mass Spectrometry (ICP-MS). In the nominally metal-free M2 and M3 medium, 0.58 and 0.028µM zinc or copper, respectively, were detected, amounting to 7.26 and 4.67% the concentration of zinc or copper, respectively, present in the standard FM medium (**Table 2**). When cells subcultured for 3 weeks in M2 or M3 medium were analyzed by ICP-MS, we found that 10<sup>7</sup> cells still contained about 84 or 3.1 ng of zinc

#### TABLE 2 | ICP-MS analysis of media.


\*The concentration of metals added to the FM medium are reported. The nominally metalfree media M2 and M3 were subjected to ICP-MS analysis as described in Material and Methods.

or copper, respectively, i.e., about 87 and 51% of the amount present in cells growing in standard minimal medium or AX2 medium. In cells incubated with a 20 or 200-fold excess of zinc or copper, respectively, the cellular concentration increased of only 2.7- and 2.4-fold, respectively, compared to FM or AX2 medium (**Table 3**). Normalized for cell volume, the intracellular copper concentration in Dictyostelium is 6 ± 0.7 and 30.6 ± 0.3µM for cells grown in, respectively, nominally copper-free medium or medium containing 120µM Cu. The intracellular zinc concentration amounts to 243 ± 12.3 to 912 ± 37.7µM in cells grown in, respectively, nominally zinc-free medium or in presence of 160µM zinc. The intracellular values are, at least for zinc, comparable to those found in many eukaryotic cells (Krezel and Maret, 2006; Colvin et al., 2008; Vignesh et al., 2013; Kambe et al., 2015). While incubation in M2 or M3 medium slightly affected the intracellular amount of zinc or copper after 3 weeks of culture, just 24 h of growth in M1 medium (corresponding to 1 cell doubling) were sufficient to halve the iron cellular content (**Table 3**). ICP-MS analysis in cells incubated with excess of iron could not be performed for the persistence in the sample of iron precipitates that would bias the measure.

These results indicate that, unlike iron, Dictyostelium cells are extremely efficient in acquiring and storing copper and zinc, maintaining intracellular levels sufficient to support growth, even when the concentration of these metals, particularly zinc, in the medium is strongly reduced. Similarly, their high tolerance toward both zinc and copper toxicity seems to derive from their high efficiency in extruding excessive amounts of these metals from the cell.

#### Effects of Iron, Zinc, and Copper Shortage or Overload on Growth on Bacteria and Development

Cells previously grown under deficiency or overloading of iron, zinc, or copper were tested for development on agar. For iron, we employed the same conditions used for the growth assay (**Figure 1A**): cells were incubated in nominally iron-free M1 medium or with iron supplementation for 24 h, washed in Soerensen phosphate buffer and then analyzed for development. Cells grown in M1 developed normally. Cells grown in M1 plus 200µM iron developed with normal timing (**Figure 2A**), but formed very large streams, which underwent fragmentation into multiple mounds that developed to fruiting TABLE 3 | Cellular metal content (ICP-MS).


Total concentration of Zn or Cu measured by ICP-MS in lysates from cells grown for 24 h in M1 ± Fe or for 3 weeks in M2 ± Zn or M3 ± Cu. Cells growing in AX2M and FM were tested after 24 h for Fe, and 3 weeks for Zn and Cu. Data are means ± SD of one experiment in duplicate. The cell lysate concentrations are based on 10<sup>8</sup> cells in ∼130 µl.

bodies (**Figure 2B**). In the case of copper or zinc overloading, we tested concentrations that produced no toxic effects on cell growth, and the cells were grown for 3 weeks (**Figure 1D**), before being washed and assayed for development. Cells were able to develop and form fruiting bodies similarly to control cells, except for cells grown in M2 medium which failed to complete development and were arrested at the mound stage (**Figure 2A**). Cells grown in a 200-fold excess of copper formed loose mounds, which elongated to small fruiting bodies, though many cells failed to aggregate (**Figure 2A**).

To assess whether previous growth under metal deficiency or overloading affected phagocytosis and growth on nonpathogenic bacteria, cells grown as just described were also plated on agar with E. coli B/2, K. aerogenes, or S. thyphimurium, and the increase in size of the colonies scored every 24 h. The growth rate was similar for cells grown under iron, zinc, or copper deficiency or overload on all three bacterial lawns, indicating that phagocytosis and bacterial digestion are unaffected by the previous growth conditions of the cells. In all cases, fruiting bodies were also formed in the middle of the plaque (**Figure S1**).

#### Dynamics of Legionella Infection Analyzed by Flow Cytometry and Live-Cell Imaging

The effects of metal deficiency or overload on Dictyostelium resistance to pathogenic bacteria, was studied with Legionella pneumophila. L. pneumophila infection was monitored, as schematized in **Figure 3**, by confocal microscopy and by measuring in flow cytometry cellular fluorescence due to intracellular proliferation of GFP-producing bacteria, and cytotoxicity as changes in cellular physical parameters. In control cells grown in AX2 medium and incubated with Legionella, the intracellular fluorescence increased over time, reaching a peak 72 h post-infection (hpi), and decreasing thereafter (**Figure 4A**). Concomitantly, the percentage of dead cells (identified as a smaller and granulous population) incremented, exceeding 80% at 72–96 hpi (**Figure 4B**). Indeed, as observed by confocal microscopy, Legionella started to replicate inside the infected cells soon after its entry, with about 5–15 bacteria/cell 24 hpi, continuing to proliferate until filling almost completely

beginning of cell starvation. Scale bar: 1 mm.

the cell. The host cells lost the amoeboid form, rounding up and detaching from the bottom of the well by around 24 hpi. From 72 hpi, but strongly at 96 hpi, we noticed the presence of cellular debris and an increase in the number of extracellular bacteria, indicating cell lysis (**Figure 4C**). Under the conditions used, Legionella cannot grow extracellularly, thus an increase in the number of extracellular bacteria can only be the result of leakage from lysed cells. This explains also the reduction of cellular mean fluorescence measured at 96 hpi (**Figure 4A**).

After its entry, L. pneumophila forms a replication-permissive compartment inside the host cell, known as the Legionellacontaining vacuole (LCV) (Horwitz, 1983; Lu and Clarke, 2005; Steiner et al., 2017). We observed that, even at later stages of infection, the bacteria appeared to be packed together, as if contained in a cellular compartment (**Figure 4C** and **Movie S1**). We infected Dictyostelium cells producing calnexin-GFP with a RFP-producing Legionella strain. Calnexin is an ER-specific protein that was previously demonstrated to be recruited to LCVs (Fajardo et al., 2004; Ragaz et al., 2008), where it remains for at least 8–14 hpi (Urwyler et al., 2009; Weber et al., 2014). Using this marker, we evidenced the presence of a membrane surrounding the large mass of replicating bacteria even at 48 hpi (**Figure 5A**), indicating that the LCV continues to tightly interact with the ER also at advanced stages of infection. The membrane integrity of the LCV was then evaluated by infecting cells producing free RFP or CshA-RFP. The citrate synthase CshA is a cytosolic protein, composed of 492 amino acid residues and containing a peroxisomal targeting PTS2 sequence, which by interacting with Pex7 leads to its transport to peroxisomes (Huang et al., 2004). In CshA-RFP producing cells peroxisomes are labeled, in addition to a diffused labeling of the cytosol, due to the synthesized, but not yet transported, protein. Cellular compartments whose membrane is not permeable appear as "black holes" in RFP or CshA-RFP producing cells. At 24 hpi the replicating legionellae were clearly contained within a RFP or CshA-RFP impermeable membrane (**Figures 5C,D**). In some cells the RFP or CshA-RFP labeling diffused between bacteria, suggesting that the LCV membrane had become permeable to the fluorescent proteins (**Figures 5C,D**). As infection progressed, the proportion of cells showing the latter condition increased. Since RFP is smaller than CshA-RFP (27 kDa vs. at least 75 kDa for the CshA-RFP noncomplexed with Pex7), we calculated the percentage of LCVs showing co-localization of with either RFP or CshA-RFP at each time point. As shown in **Figure 5B**, at 24 hpi only 44% of LCVs were positive for CshA-RFP against 88% for RFP, with a gradual increase to 94 or 90% at 72 hpi for RFP or CshA-RFP, respectively. It is worth mentioning that no peroxisomes were found mixed with legionellae within the LCV, even at 72 hpi, but only free RFP or CshA-RFP.

fluorescence, FSC/SSC forward/sideward scatter channel.

From these data we conclude that the LCV continues to be surrounded by a membrane even at later stages of infection, with the membrane becoming gradually permeable to proteins of increasing molecular size between 24 and 72 h after its formation, with no egression of bacteria in the cytosol, but dispersal into the extracellular milieu following cell lysis.

### Legionella Uptake and Infection in AX2 or Nramp1-KO Cells Grown in Minimal Medium with or without Iron Supplementation

We followed Legionella infection in AX2 cells grown for 24 h in minimal medium containing 0, 100, or 200µM FeCl3. In cells deprived of iron during growth, Legionella intracellular replication was strongly reduced (**Figure 6A**) and the percentage of dead cells did not increase significantly over time (**Figure 6B**). These data were confirmed by live microscopy: at 72 hpi, only a very small number of cells were infected with bacteria (**Figure 6C**).

When cells previously grown in the presence of iron were tested, Legionella intracellular growth was significantly higher compared to the condition of iron shortage, with the number of intracellular bacteria rapidly increasing during the first 48 hpi, and starting to decrease concomitantly with a boost in the percentage of dead cells (**Figures 6A,B**). Indeed, by confocal microscopy we observed already at 48 hpi and strongly at 72 hpi an increase in the number of extracellular bacteria and of lysed cells (**Figure 6C**). No differences were detected when the amount of iron in the growth medium was doubled to 200µM (**Figures 6A,B**). Despite the inhibition of growth induced by the latter concentration of iron (**Figure 1A**), no increase in the percentage of dead cells was observed in the uninfected control (4.8% of dead cells compared to 77.3% in infected cells at 72 hpi). Therefore, the cytotoxic effect observed in infected cells is not due to the metal, but to the rapid proliferation of the bacteria. Adding iron during the infection assay to cells previously grown in the presence of iron facilitated Legionella intracellular growth, with a maximal value already reached at 24 hpi (**Figure 6A**).

We also tested whether iron could affect Legionella uptake. AX2 cells were grown for 24 h in M1 ± Fe, as previously described. Thereafter, they were washed, resuspended in nominally iron-free medium, co-centrifuged with GFPproducing Legionella (MOI 10:1) and incubated for 40 min at 25◦C. Uptake was measured as increase of the percentage of fluorescent (i.e., harboring bacteria) cells. When cells grown in M1 were assayed, only about 23% of them were fluorescent, rising to 38% 40 min post-infection (**Figure 7**). In cells grown in M1 plus 100 or 200µM iron, the percentage of fluorescent cells was already significantly higher at time 0 (47–50%), and further increased over 66% after 40 min, with no differences between 100 and 200µM Fe (**Figure 7**). It is worth mentioning that though the number of infected cells increased, the mean number of Legionella per cell did not increase significantly, as confirmed by confocal microscopy. Thus, growing cells in the presence or absence of iron affects both uptake and infection of Legionella.

We previously showed that Dictyostelium Nramp1 is rapidly recruited from the Golgi to phagosomes and macropinosomes. Here it mediates iron efflux to the cytosol and, by doing so, it confers resistance to infection by intracellular pathogens, such as Legionella and Mycobacterium avium (Peracino et al., 2006, 2013; Buracco et al., 2015). Similarly to AX2 cells, when Nramp1- KO cells were grown for 24 h in M1 medium, infection was significantly reduced in both parameters, Legionella intracellular replication and bacteria-triggered cytotoxicity (**Figure 8**). In cells grown in the presence of iron and infected in a medium containing the same concentration of the metal, Legionella intracellular growth reached again a peak at 24 hpi and then started to decrease over time (**Figure 8A**). At the same time, the percentage of dead cells increased steadily, exceeding 80% at 72 hpi (**Figure 8B**). Thus, iron deficiency or overload has similar effects on Legionella infection in cells lacking Nramp1 and parental cells, though in the Nramp1-KO mutant intracellular growth of Legionella was more rapid, consistent with its increased susceptibility to infection (Peracino et al., 2006).

### Effects of Altering Zinc or Copper Availability on Legionella Infection

We tested whether zinc or copper exert similar effects as iron. AX2 cells grown for 168 h in M2 ± Zn or M3 ± Cu were infected and analyzed by flow cytometry. Under all conditions tested, neither bacterial intracellular proliferation nor cytotoxicity were significantly altered compared to the control incubated in standard FM medium (**Figure 9**). As

cytometry. The intracellular growth of Legionella was assessed by measuring the mean fluorescence of GFP-producing L. pneumophila within AX2 cells and L. pneumophila-triggered cytotoxicity by determining the percentage of dead cells. For ICG, the trend in different experiments was similar, although the absolute values varied strongly between replicates, due to changes in cell autofluorescence. Thus, Legionella replication was expressed as mean fold-increase in fluorescence intensity relative to the autofluorescence of the corresponding non-infected cells. All measurements are mean values of seven different experiments with error bars (±SD). (C) Confocal fluorescence images of infected living cells at different times post-infection. (Top) GFP-producing Legionella and (Bottom) merge with corresponding phase-contrast. In intact cells, the proliferating bacteria are packed together, except for extracellular bacteria in lysed cells (96 hpi). Scale bars: 5µm.

shown in **Table 3**, omitting zinc or copper in the culture medium affects only minimally the cellular concentration of these metals, particularly for zinc, compared to the standard FM medium, with no inhibitory effect on cell growth, though in the case of zinc postaggregative development is blocked. In some experiments the intracellular zinc chelator TPEN was added during Legionella infection to cells previously grown in M2 medium, but no effects on Legionella growth or cytotoxicity were observed compared to the M2 medium (**Figure 9**). The absence of significant differences in Legionella infection between cells grown in nominally-free metal medium, supplemented in the case of zinc with TPEN during infection, or at the highest concentrations of metal tested suggests to us that zinc and copper play a minor role in Dictyostelium resistance to Legionella infection.

### Intracellular Distribution of Free Zinc Ions in Dictyostelium Cells

To better understand if free zinc ions could play a role in predation and infection, we investigated their intracellular localization using the cell-permeable fluorogenic Zn2<sup>+</sup> reporter Zinpyr-1, which fluoresces in the pH range 3.5 to 8 upon binding Zn2<sup>+</sup> and is highly selective for the labile Zn2<sup>+</sup> pool (Burdette et al., 2001; Figueroa et al., 2014). When the probe was administered to AX2 cells producing the iron transporter Nramp1-RFP, the fluorescent signal accumulated in vesicles coated with Nramp1-RFP (**Figure 10A**). Nramp1 traffics between trans-Golgi and endosomes, including macropinosomes, phagosomes, and endolysosomes, but is not found in postlysosomal vesicles (Peracino et al., 2006). In cells producing the iron transporter NrampB-RFP, we

and expressed as percentage of the total. For all confocal images, red and green fluorescence merge is shown. The diffusion of free-RFP or CshA-RFP fluorescence inside the LCV allows evaluating the integrity of the LCV membrane. Scale bars: 5µm.

observed also transient labeling of the contractile vacuole (CV) membrane, with a faint labeling inside the vacuole, most evident under continuous Zinpyr-1 loading (**Figure 10B**). NrampB is localized exclusively in the membrane of the CV tubule network (Peracino et al., 2013). Since Zinpyr-1 under continuous loading failed to label other membranes, including the plasma membrane, the fluorescence in the CV membrane cannot be due to some unspecific effect of Zinpyr-1 diffusing across the membrane bilayer. As additional control, we incubated cells treated with the intracellular membrane-permeable chelator TPEN with Zinpyr-1 under continuous loading. As shown in **Figure 10C**, the Zinpyr-1 fluorescence is undetectable both at the level of the contractile vacuole and endosomal vesicles. We conclude that the Zinpyr-1 fluorescence in the CV is due to chelation of free zinc ions being transported across the CV membrane. No differences in Zinpyr-1 cellular localization were observed in cells grown in M2 medium ± Zn. In cells incubated with TRITC-labeled E. coli, we observed co-localization of

the fluorescent probe also with the bacteria in phagosomes (**Figure 10D**).

Thus, in Dictyostelium cells, free zinc ions accumulate in vesicles of the endo-lysosomal pathway, including phagosomes. The transient labeling of the CV membrane and the faint labeling inside the CV suggests that excess free zinc is transported across the CV membrane, where it could be stored or secreted extracellularly. Similar experiments with copper could not be done, as fluorescent reporters for copper are not commercially available.

#### DISCUSSION

The high number of copper and zinc transporters in Dictyostelium makes difficult a targeted genetic approach to decipher the role of these metals in host-pathogen interactions, in contrast to what has been possible with the Nramp iron transporters (Peracino et al., 2006, 2013). In this paper we have followed an indirect approach, by growing cells under conditions of metal deficiency or overload and then testing the cells for resistance to infection by L. pneumophila, which is a very

well-studied pathogen for these cells under laboratory conditions (Hägele et al., 2000; Solomon and Isberg, 2000; Peracino et al., 2006; Hilbi et al., 2007; Steinert, 2011; Bozzaro et al., 2013b; Hoffmann et al., 2014).

assuming unequal variance) vs. the corresponding time of the condition in M1.

We used a minimal culture medium (FM Medium) made of amino acids, glucose, a few selected vitamins and trace elements (Franke and Kessin, 1977), whose concentration could be modified. As contamination of pico- to nano-molar traces of metals could derive from reagents (in particular amino acids), glass or other materials (Kay, 2004), the culture medium could not be totally depleted of iron, copper, or zinc, as confirmed by the ICP-MS analysis. The nominally zinc- or copper-free medium still contained, respectively, 7.26 and 4.6% of metal, compared to the standard FM medium. What is more important, the metal concentration in cells growing for more than 20 generations in these nominally zinc- or copper-free media was only slightly reduced compared to cells growing in FM medium, namely of about 15 and 50% for zinc and copper, respectively. Cell growth in media containing 276- or 4,285 fold zinc or copper, respectively, compared to the concentration in the nominally metal-free media, resulted in an increase of only 3- and about 5-fold cellular zinc or copper, respectively, after 24 cell doublings. These data indicate that Dictyostelium cells possess very tight regulatory mechanisms of uptake, storing and release of these metals, which can explain why the cell growth rate was absolutely unaffected in media within such a broad range of external zinc or copper concentrations. Only at higher concentrations of copper and zinc, inhibitory and toxic effects on cell growth were detectable, confirming previous results (Burlando et al., 2002; Balbo and Bozzaro, 2008). The high tolerance, at least for copper, could be linked to the activity of one or more MNK Cu-ATPases, whose expression in the plasma membrane is enhanced by addition of salts (Burlando et al., 2002; Hao et al., 2016). For zinc, nothing is known on the cellular localization and activity of the transporters (Sunaga et al., 2008).

In the case of iron, the cellular concentration was reduced to a quarter after 24 h incubation, and about two duplications, in nominally iron-free medium compared to FM medium. We have confirmed that iron is essential for cell growth, with the growth rate being slowed down and cell duplication inhibited within few generations in both nominally iron-free medium and iron overload (Peracino et al., 2013). Iron overload led to formation of very large aggregating streams, which underwent fragmentation. A similar phenotype has been already described for the Nramp1/NrampB double knockout mutant (Peracino et al., 2010), and recalls the streamer mutants, some of which are defective in cGMP phosphodiesterase (Ross and Newell, 1979; Newell and Liu, 1992), suggesting a possible regulation of chemotaxis by iron homeostasis that needs further investigation.

Interestingly, the intracellular growth of Legionella was strongly inhibited in cells grown in nominally iron-free medium, but enhanced in cells exposed during growth to 100 or 200µM iron. Addition of iron to the medium during infection further stimulated Legionella growth in iron overloaded cells. The uptake of Legionella was also affected by iron, in that the number of cells ingesting Legionella was higher in the population grown under iron overload, though the mean number of Legionella per cell was constant. Whether this is due to iron being secreted by the iron-overloaded cells, thus stimulating bacterial attachment to the cell surface, is unclear. Increased Legionella uptake due to extracellular iron has been recently described in macrophages (O'Connor et al., 2016).

Similar results were obtained with the Nramp1 knockout mutant. The iron transporter Nramp1 mediates iron efflux from the phagosome, and its inactivation increases susceptibility to L. pneumophila, M. avium, and F. tularensis (Peracino et al., 2006; Buracco et al., 2015; Brenz et al., 2017). Growing the mutant cells in nominally iron-free medium for 24 h strongly inhibited Legionella intracellular growth, whereas cell growth under iron overload enhanced its susceptibility to infection to a higher level than in the parental cells, consistent with a higher accumulation of iron in the LCV, due to inactivation of Nramp1 (Buracco et al., 2015). Remarkably, Legionella intracellular proliferation in iron overloaded cells was so rapid that already at 48 hpi almost all cells underwent lysis with dispersal of the bacteria in the extracellular milieu. Taken together, these results indicate that a condition of iron starvation is protective for Dictyostelium cells against Legionella infection whereas high iron cellular content predisposes to rapid infection.

Iron is essential for Legionella growth and expression of virulence genes (Robey and Cianciotto, 2002; Cianciotto, 2015; Portier et al., 2015). In Dictyostelium, the source of iron for Legionella is the macropinosome, where the bacterium is

FIGURE 8 | Iron effects on L. pneumophila intracellular replication and cytotoxicity in Nramp1-KO mutant. Infection was evaluated in Nramp1-KO cells as previously described for AX2 cells in Figure 5, in the presence of iron in the infection medium at the same concentrations used for cell growth. Flow cytometry analysis of ICM (A) and cytotoxicity (B) are shown. In cells lacking Nramp1, growth in M1 affected bacterial replication and cytotoxicity similarly to AX2, whereas growth in the presence of iron stimulated Legionella infection to a higher degree than in AX2 (see Figure 5). Mean values of at least two experiments in triplicate with error bars (±SD). \*P < 0.05 or \*\*P < 0.01 (two-tailed Student's t-test, assuming unequal variance) vs. the condition of 100µM FeCl3.

taken up (Peracino et al., 2010). By pumping iron outside the macropinosome via Nramp1, Dictyostelium attempts to starve Legionella for iron, but Legionella can manipulate Nramp1 activity, by avoiding recruitment to the macropinosome of the V-H<sup>+</sup> ATPase (Peracino et al., 2010), and inhibiting its activity via the Dot/Icm substrate SidK (Xu et al., 2010). The V-H<sup>+</sup> ATPase is essential for Nramp1-mediated iron efflux (Buracco et al., 2015), thus iron accumulation transforms the macropinosome in a replication-permissive LCV (Steiner et al., 2017). If so, then a condition of iron starvation during cell growth, resulting in less iron being available in the macropinosome, independently of Nramp1 activity, will result in increased resistance of the host cell to Legionella infection, while iron overload will increase susceptibility to infection. Recently, an inhibitory effect of iron depletion on Legionella infection has been shown in macrophages treated with the iron chelator deferoxamine mesylate during infection or in Acanthamoeba castellanii infected in iron-free buffer (O'Connor et al., 2016). In the case of A. castellanii, the experimental conditions were different from those used in our experiments, as A. castellanii was grown in iron-containing medium and only the infection assay was done in iron-free buffer, which could explain the rather mild effect reported by the authors.

We have shown that Legionella proliferation occurs inside the LCV up to cell lysis, with the LCV in close contact to, if not fused with the endoplasmic reticulum (ER) membrane even up to later stages of infection. Association with the ER has been already described at early stages of infection, presumably induced by Legionella for acquiring new membrane for the LCV (Fajardo et al., 2004; Lu and Clarke, 2005; Urwyler et al., 2009; Weber et al., 2014; Steiner et al., 2017). Our data suggest that this process sets forth up to later stages.

Interestingly, the LCV membrane becomes gradually permeable between 24 and 48 hpi, allowing diffusion in the vacuole of cytosolic proteins first of lower (27 kDa) and later of higher molecular mass (about 76 kDa), but the bacteria remain packed together and do not egress into the cytosol, suggesting that, though permeable to proteins, the LCV membrane does not allow exit of the bacteria before cell lysis. Vacuole disruption shortly before cell lysis has been reported (Creasey and Isberg, 2012), but the present data suggest that Legionella induces permeability of the vacuole in the proliferation phase, possibly to get easy access to nutrients of the host. The dynamics of permeability by using fluorescent probes of increasing molecular size will be reported elsewhere. We failed to detect in the extracellular space clumps of bacteria, which could suggest non-lytic release in LCV's, as described in A. castellanii (Berk et al., 1998; Chen et al., 2004; Lau and Ashbolt, 2009).

In sharp contrast to iron, variation in cellular zinc or copper content over a range of, respectively, 3- to 5-fold did not significantly affect timing or extent of Legionella intracellular growth. For both, the infection rate was slower, though statistically not significant or borderline, in cells grown in nominally zinc or copper-free medium compared to cells with 5-fold more intracellular copper. If any, this would suggest that a higher intracellular zinc or copper concentration is not toxic for Legionella. In the case of zinc, Dictyostelium cells with the lowest zinc concentration were blocked at mound stage, suggesting that a suboptimal zinc concentration is deleterious for development. Zinc ions have been shown to enhance prestalk to stalk differentiation (Kubohara and Okamoto, 1994) and a family of three ZNT zinc transporters has been shown to be expressed at postaggregative stage in pre-stalk cells (Sunaga et al., 2008).

FIGURE 9 | Alterations of zinc or copper availability do not affect Legionella infection in Dictyostelium cells. AX2 cells were grown for 3 weeks (24 generations) in (A,B) M2 ± Zn or (C,D) M3 ± Cu medium and then infected with GFP-producing L. pneumophila. ICG and cytotoxicity were evaluated as previously described in Figure 4, with the first expressed as mean fold-increase in fluorescence intensity relative to the autofluorescence of the corresponding non-infected cells. Omitting or adding zinc or copper in cell growth medium did not significantly affect Legionella intracellular growth and induced cytotoxicity compared to FM medium. In the case of zinc, adding the intracellular chelator TPEN during infection failed to have any effect. Infection in cells grown in M3 + 6µM Cu was comparable to the other conditions and, for simplicity, it was removed from the graphics. Mean values of three different experiments with error bars (±SD). (C) P = 0.054, at 24 hpi for M3 compared to M3 + 60µM. The numbers indicate metal and chelator concentration in µM.

By using the iron fluorescent reporter calcein, we showed previously that iron is rapidly depleted from endo-lysosomal vesicles via Nramp1 (Buracco et al., 2015). By using the membrane-permeable zinc fluorescent reporter Zinpyr-1, we have now shown that free zinc ions accumulate in vesicles of the endo-lysosomal pathway, including phagosomes, and in the contractile vacuole. Free zinc in other cellular compartments or in the cytosol was below detection with this fluorescent reporter. Zinc transport across the membrane of the contractile vacuole is consistent with the proposed CV role as store or sink for metals (Bozzaro et al., 2013a; Peracino et al., 2013). By using bacteria expressing a copper and a reactive oxygen species biosensor, Hao et al. (2016) showed that copper and reactive oxygen species accumulate in Dictyostelium phagosomes upon bacterial phagocytosis. Thus, both zinc and copper are found in Dictyostelium phagosomes, consistent with a potential intoxicating role against pathogens, as suggested for macrophages (Babu and Failla, 1990; Wagner et al., 2005; White et al., 2009; Ward et al., 2010; Botella et al., 2011; Chaturvedi and Henderson, 2014; Djoko et al., 2015), though the evidence, at least for zinc, is less clear, and zinc sequestration, likely as it occurs for iron, has been also proposed as defense mechanism (Kehl-Fie and Skaar, 2010; Vignesh et al., 2013).

We have no evidence for phagocytosis inducing a burst of zinc in endocytic vesicles, as shown in macrophages with nonpathogenic bacteria (Botella et al., 2011). We used TRITC-labeled E. coli, which are not alive, whereas Botella and coworkers used living bacteria. If this would be the reason, then it would not be phagocytosis per se that triggers zinc accumulation, but a cellular response to the ingested living bacteria. Alternatively, the different results can be explained with the different zinc reporters used, Zinpyr-1 in this paper and FluoZin-3 in the Botella et al. (2011) paper. It is worth mentioning that Zinpyr-1 is more specific and selective for free zinc than FluoZin-3 (Figueroa et al., 2014). Further experiments are required to distinguish between these possibilities.

Our data suggest that, in contrast to iron, changes in cellular zinc or copper play a minor role in Dictyostelium resistance to pathogens, at least with L. pneumophila. Whether this holds true for other pathogens, such as mycobacteria, is under investigation.

A shortcoming of our experimental approach is the apparently very high efficiency of Dictyostelium cells to cope with conditions of zinc or copper starvation or overloading. To which extent this is typical of Dictyostelium cells or is a common feature of other cells, including macrophages, is unknown. The experimental approach described here offers now the opportunity to identify, in a transcriptomic study, the genes responsible for this high homeostatic adaptation to zinc and copper deprivation or intoxication. From the site of the pathogen, the same approach can be used to identify potential mutants in metal dependent virulence genes.

There is an analogy between the conditions used in this paper and those encountered with humans affected by transition metal deficiency or overloading, due to nutritional or environmental problems. In particular, iron or zinc deficiency is a global nutritional problem, with negative health consequences, including anemia, impaired immune function and neurological diseases. Dietary iron supplementation to prevent micronutrient deficiency in areas with a high burden of infectious diseases is a double–edged weapon, as it can exacerbate morbidity and mortality from infectious diseases (Oppenheimer, 2001; Soofi et al., 2013; Pasricha and Drakesmith, 2016). Similarly, in patients undergoing multiple blood transfusion, the risk of infections is higher, and iron therapy in patients at high risk of infection has been questioned (Ozment and Turi, 2009; Weiss and Carver, 2017). The previous (Peracino et al., 2006; Buracco et al., 2015) and present results with the professional phagocyte Dictyostelium support this caution, offering a mechanistic explanation for the iron effects on infection that holds true also for macrophages. Concerning copper, but in particular zinc, further studies are required with macrophages isolated from animals exposed to metal deficiency or overloading. Identifying the genes responsible in Dictyostelium for the tight regulation of zinc and copper homeostasis and for their transport in phagosomes can shed some light on cellular zinc and copper store and mobilization, and the genetic basis of antimicrobial metal transport, also in macrophages.

#### AUTHOR CONTRIBUTIONS

SBu and SBo designed the experiments; SBu conducted the experiments, acquired, and analyzed the data; BP generated

#### REFERENCES


the Dictyostelium KO mutants and the plasmids for RFP-fused proteins and helped with some experiments; CA was responsible for the bioinformatic analysis of the metalloproteome. EB cloned the cshA gene, generated, and characterized cells expressing CshA-RFP. SBu and SBo wrote the manuscript; all approved the paper.

#### ACKNOWLEDGMENTS

We thank Annette Mueller Taubenberger for AX2 cells producing GFP and for the 389-2 vector (C-terminal mRFPmars), Michael Steinert for the L. pneumophila Corby producing GFP and Ds-red Express. We are grateful to the Dictybase team (www.dictybase.org) for maintenance of the Dictyostelid genome database, and Alessandra Viale of the Molecular Imaging Center (Department of Chemistry, University of Torino, Italy) for ICP-MS analysis. This work was supported by a research grant of the Compagnia San Paolo/Università di Torino (12-CSP-C03-065).

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2017.00536/full#supplementary-material

Movie S1 | 3-D reconstruction of AX2 cells infected with GFP-producing L. pneumophila 48 hpi. 3-D projections of a single AX2 cell infected with GFP-tagged L. pneumophila were generated and animated with the confocal microscope LSM800 software (Zen Software, Carl Zeiss, Inc., Oberkochen, Germany) in.avi format. The movie shows bacteria contained in high number inside the infected cell. Legionellae are not spread in the cell, but are packed together in a "ring"-like structure, presumably around the nucleus.

Figure S1 | Plaque formation by AX2 cells previously grown in media depleted or overloaded with iron, zinc, or copper. Ax2 cells were grown in M1 ± Fe for 24 h or in M2 ± Zn or M3 ± Cu for 3 weeks. Cells were then washed, resuspended in Soerensen buffer, and serial dilutions were plated on a lawn of E. coli B2, S. typhimurium, or K. aerogenes, as described in Materials and Methods. Pictures of the appearance and widening of growth plaques were acquired daily with a scanner. Images of the lowest dilution used (10 cells per well) are shown that were acquired 3 or 4 days after plating. No significant differences in growth were observed. Numbers indicate metal concentration in µM (Top) or time in days (on the right).


pathways: emerging roles for zinc and copper. Biosci. Rep. 33:e00049. doi: 10.1042/BSR20130014


**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 Buracco, Peracino, Andreini, Bracco and Bozzaro. 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) or licensor 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.

# Inorganic Polyphosphate Is Essential for *Salmonella* Typhimurium Virulence and Survival in *Dictyostelium discoideum*

Macarena A. Varas <sup>1</sup> , Sebastián Riquelme-Barrios <sup>2</sup> , Camila Valenzuela<sup>2</sup> , Andrés E. Marcoleta<sup>3</sup> , Camilo Berríos-Pastén<sup>3</sup> , Carlos A. Santiviago<sup>2</sup> \* and Francisco P. Chávez <sup>1</sup> \*

<sup>1</sup> Laboratorio de Microbiología de Sistemas, Departamento de Biología, Facultad de Ciencias, Universidad de Chile, Santiago, Chile, <sup>2</sup> Laboratorio de Microbiología, Departamento de Bioquímica y Biología Molecular, Facultad de Ciencias Químicas y Farmacéuticas, Universidad de Chile, Santiago, Chile, <sup>3</sup> Laboratorio de Biología Estructural y Molecular, Departamento de Biología, Facultad de Ciencias, Universidad de Chile, Santiago, Chile

#### *Edited by:*

Thierry Soldati, Université de Genève, Switzerland

#### *Reviewed by:*

Francisco Ramos-Morales, Universidad de Sevilla, Spain Salvatore Bozzaro, Università degli Studi di Torino, Italy Richard H. Gomer, Texas A&M University College Station, United States

#### *\*Correspondence:*

Carlos A. Santiviago csantiviago@ciq.uchile.cl Francisco P. Chávez fpchavez@uchile.cl

*Received:* 03 November 2017 *Accepted:* 09 January 2018 *Published:* 30 January 2018

#### *Citation:*

Varas MA, Riquelme-Barrios S, Valenzuela C, Marcoleta AE, Berríos-Pastén C, Santiviago CA and Chávez FP (2018) Inorganic Polyphosphate Is Essential for Salmonella Typhimurium Virulence and Survival in Dictyostelium discoideum. Front. Cell. Infect. Microbiol. 8:8. doi: 10.3389/fcimb.2018.00008 Inorganic polyphosphate (polyP) deficiency in enteric bacterial pathogens reduces their ability to invade and establish systemic infections in different hosts. For instance, inactivation of the polyP kinase gene (ppk) encoding the enzyme responsible for polyP biosynthesis reduces invasiveness and intracellular survival of Salmonella enterica serovar Typhimurium (S. Typhimurium) in epithelial cells and macrophages in vitro. In addition, the virulence in vivo of a S. Typhimurium 1ppk mutant is significantly reduced in a murine infection model. In spite of these observations, the role played by polyP during the Salmonella-host interaction is not well understood. The social amoeba Dictyostelium discoideum has proven to be a useful model for studying relevant aspects of the host-pathogen interaction. In fact, many intracellular pathogens can survive within D. discoideum cells using molecular mechanisms also required to survive within macrophages. Recently, we established that S. Typhimurium is able to survive intracellularly in D. discoideum and identified relevant genes linked to virulence that are crucial for this process. The aim of this study was to determine the effect of a polyP deficiency in S. Typhimurium during its interaction with D. discoideum. To do this, we evaluated the intracellular survival of wild-type and 1ppk strains of S. Typhimurium in D. discoideum and the ability of these strains to delay the social development of the amoeba. In contrast to the wild-type strain, the 1ppk mutant was unable to survive intracellularly in D. discoideum and enabled the social development of the amoeba. Both phenotypes were complemented using a plasmid carrying a copy of the ppk gene. Next, we simultaneously evaluated the proteomic response of both S. Typhimurium and D. discoideum during host-pathogen interaction via global proteomic profiling. The analysis of our results allowed the identification of novel molecular signatures that give insight into Salmonella-Dictyostelium interaction. Altogether, our results indicate that inorganic polyP is essential for S. Typhimurium virulence and survival in D. discoideum. In addition, we have validated the use of global proteomic analyses to simultaneously evaluate the host-pathogen interaction of S. Typhimurium and D. discoideum. Furthermore, our infection assays using these organisms can be exploited to screen for novel anti-virulence molecules targeting inorganic polyP biosynthesis.

Keywords: *Salmonella*, *Dictyostelium*, polyphosphate, *ppk*, virulence, intracellular survival, proteomics

#### INTRODUCTION

The ability of Dictyostelium discoideum cells to feed on bacteria has prompted the development of virulence assays for identifying host defense mechanisms and deciphering bacterial virulence factors (Cosson et al., 2002; Froquet et al., 2009). Basic cellular processes such as phagocytosis, phagosomal development and autophagy, are evolutionarily well conserved between Dictyostelium and macrophages (Hägele et al., 2000; Bozzaro and Eichinger, 2011; Dunn et al., 2018). Consequently, D. discoideum has been established as a model to study hostpathogen interaction in a wide range of pathogenic bacteria such as Legionella, Salmonella, Francisella, Mycobacterium, and Pseudomonas, among others (Pukatzki et al., 2002; Hagedorn and Soldati, 2007; Weber et al., 2014; Lampe et al., 2015; Bravo-Toncio et al., 2016; Riquelme et al., 2016; Cardenal-Muñoz et al., 2017). Unlike mammalian phagocytes, D. discoideum is amenable to a diverse array of genetic manipulations facilitating the in vivo identification of host susceptibility determinants and pathogen virulence factors (Carilla-Latorre et al., 2008; Hasselbring et al., 2011; Pan et al., 2011; Tosetti et al., 2014; Zhang et al., 2016). However, in vivo host-pathogen interaction during bacterial infection in D. discoideum remains poorly understood.

Salmonella enterica serovar Typhimurium (S. Typhimurium) is a foodborne pathogen causative of gastroenteritis in a variety of warm-blooded animals that largely relies on its ability to survive inside host cells. Relevant genes required for this process are located in pathogenicity islands such as SPI-1 and SPI-2, which encode two independent type III secretion systems (T3SS-1 and T3SS-2, respectively) that inject effector proteins into host cells and are critical during different stages of infection (reviewed in Haraga et al., 2008). In a previous study, we showed that S. Typhimurium genes linked to virulence are required to survive in D. discoideum, including those encoding factors involved in the biosynthesis of aromatic compounds, the production of a lipopolysaccharide containing a complete O-antigen, T3SS-1, T3SS-2, the type VI secretion system (T6SS) encoded in SPI-6 and the PhoP/PhoQ two-component system (Riquelme et al., 2016). Hence, S. Typhimurium exploits a common set of molecular mechanisms to survive within amoeba and animal host cells, supporting the use of D. discoideum as a model for host-pathogen interactions and to study the cellular processes that are affected during infection.

We are particularly interested in inorganic polyphosphate (polyP) metabolism because this biopolymer is important for D. discoideum development and predation, and for virulence in many bacterial pathogens (Zhang et al., 2005; Brown and Kornberg, 2008). In fact, we have demonstrated that polyP biosynthesis is essential for P. aeruginosa PAO1 virulence toward this amoeba (Bravo-Toncio et al., 2016). Inorganic polyP is an abundant and ubiquitous biopolymer that has been conserved in every cell in nature. In the last decades, an increasing number of physiological functions have been reported for polyP in bacteria (Brown and Kornberg, 2008). Due to their phosphoanhydride bonds similar to those in ATP and their properties as polyanions, polyP serve as microbial phosphagen in a variety of biochemical reactions, as a buffer against alkalis, and as a metal storage and metal-chelating agent. In addition, recent studies have revealed the importance of polyP metabolism in signaling and regulatory processes, cell viability and proliferation, and as modulator of microbial stress response (Gray and Jakob, 2015). In numerous pathogenic bacteria, inactivation of the polyP kinase gene (ppk) encoding the enzyme responsible for polyP biosynthesis causes defects in biofilm formation, quorum sensing, motility, general stress and stringent responses, and production of virulence factors (Rao et al., 1998; Rashid and Kornberg, 2000; Rashid et al., 2000a,b; Brown and Kornberg, 2008; Varela et al., 2010; Varas et al., 2017). In S. Typhimurium, inorganic polyP is essential for long-term survival and virulence factors production (Kim et al., 2002). However, the exact mechanism that links polyP metabolism and Salmonella virulence remains to be elucidated.

In this study, we used D. discoideum as a host model to study the link between polyP biosynthesis and virulence in S. Typhimurium. To this end, we assessed the intracellular survival of S. Typhimurium wild-type and 1ppk strains in the amoeba, and the effect of these strains in the social development of the host. Our results indicate that inorganic polyP is essential during S. Typhimurium infection of D. discoideum. Also, we used global proteomic profiling to get a global view of host cellular responses toward infection that gave insight into Salmonella-Dictyostelium interaction.

#### MATERIALS AND METHODS

#### Bacterial Strains and Culture Conditions

The bacterial strains used in this study are listed in **Table 1**. All S. Typhimurium strains are derivatives of the wild-type, virulent strain 14028s (Fields et al., 1986). Bacteria were routinely grown at 37◦C with agitation in Luria-Bertani (LB) medium (10 g/L tryptone, 5 g/L yeast extract, 5 g/L NaCl). When required, LB medium was supplemented with ampicillin (Amp, 100 mg/L), chloramphenicol (Cam, 20 mg/L) or kanamycin (Kan, 75 mg/L). LB medium was solidified by the addition of agar (15 g/L). All procedures involving the use of pathogenic organisms were conducted following the guidelines in the Biosafety Manual of the National Commission of Scientific and Technological Research


(CONICYT), and were approved by the Institutional Biosafety Committee of Universidad de Chile, Campus Norte.

#### Construction of Mutant Strains

All S. Typhimurium mutants were generated by the Lambda Red recombination method (Datsenko and Wanner, 2000) with modifications (Santiviago et al., 2009), using plasmid pCLF4 (Kan<sup>R</sup> , GenBank accession number EU629214) or pCLF2 (Cam<sup>R</sup> , GenBank accession number HM047089) as template. Correct allelic replacement in these mutants was confirmed by PCR amplification using primers flanking the substitution site. All primers for PCR amplifications are listed in **Table 2**.

### Construction of Complementing Plasmid pPPK

A DNA fragment containing the ppk gene (including its promoter region) was amplified from the genome of S. Typhimurium strain 14028s using Taq DNA polymerase (Invitrogen) and primers ppk\_Out5 and ppk\_Out3 (**Table 2**). The PCR product was purified from 1% agarose gels using the "QIAquick Gel Extraction Kit" (QIAGEN) and cloned into pBAD-TOPO using the "pBAD-TOPO TA Expression Kit" (Invitrogen). The presence and orientation of the insert in the recombinant plasmid generated (pPPK) was confirmed by PCR amplification using combinations of primers ppk\_Out5, ppk\_Out3, pBAD\_Forward and pBAD\_Reverse (**Table 2**). Finally, S. Typhimurium 1ppk was transformed by electroporation with plasmid pPPK for complementation assays.

#### *Dictyostelium* Culture Conditions

D. discoideum strain AX4 was obtained from Dicty Stock Center (Kreppel et al., 2004; Basu et al., 2013; Fey et al., 2013), and cultured according to standard protocols (Fey et al., 2007). Briefly, amoebae were maintained at 22◦C in SM medium (10 g/L glucose, 10 g/L peptone, 1 g/L yeast extract, 1 g/L MgSO<sup>4</sup> × 7H2O, 1.9 g/L KH2PO4, 0.6 g/L K2HPO4, 20 g/L agar), growing on a confluent lawn of Escherichia coli B/r. Before development and intracellular survival assays, amoebae were grown in the absence of bacteria (axenic cultures) at 22◦C with agitation (180 rpm) in liquid HL5 medium (14 g/L tryptone, 7 g/L yeast extract, 0.35 g/L Na2HPO4, 1.2 g/L KH2PO4, 14 g/L glucose, pH 6.3).



Underlined sequences correspond to the region that anneals to the 5′ or 3′ end of the antibiotic-resistance cassette in template vectors pCLF2 and pCLF4.

When required, media were supplemented with streptomycin (Stp; 300 mg/L) and Amp (100 mg/L). Amoebae were harvested in early exponential phase (1–1.5 × 10<sup>6</sup> cells/mL) and centrifuged at 500 × g for 10 min at 4◦C. The supernatant was discarded and the pellet was adjusted to 1 × 10<sup>6</sup> cells/mL in HL5 medium for development assays or washed three times using Soerensen buffer (2 g/L KH2PO4, 0.36 g/L Na2HPO<sup>4</sup> × 2H2O, pH 6.0) for intracellular survival assays. The population of viable amoebae was evaluated by Trypan blue exclusion and counting in a Neubauer chamber.

#### Development Assay

Individual wells of a 24-well plate containing N agar (Soerensen buffer supplemented with 1 g/L peptone, 1 g/L glucose and 20 g/L agar) were inoculated with 30 µL of a stationaryphase culture from each bacterial strain to be evaluated. The plate was incubated overnight at 22◦C to generate bacterial lawns. The next day, 10 µL of a cellular suspension containing 1 × 10<sup>4</sup> D. discoideum AX4 cells in HL5 was spotted in the middle of each well and the plate was further incubated at 22◦C for 6 days. Amoebae were monitored daily and the developmental phase reached ("aggregation," "elevation," and "culmination") was scored. A score of "1" was assigned when amoebae aggregated forming a phagocytosis plaque. A score of "2" was assigned when elevated structures, such as "slugs" or "fingers", were observed all across the surface of the agar in the well. A score of "3" was assigned when fruiting bodies were formed all across the surface of the agar in the well. Intermediate states among two developmental phases were recorded as average values of the corresponding scores. In addition, representative images of D. discoideum development were obtained at days 2 and 4 using an Olympus MVX10 stereomicroscope.

#### Intracellular Survival Assay

Intracellular survival assays were performed as described previously (Riquelme et al., 2016). Briefly, each bacterial strain to be assessed was grown to stationary phase, harvested and washed twice with Soerensen buffer. Next, ∼2 × 10<sup>7</sup> D. discoideum AX4 cells were mixed with each bacterial strain until reaching a multiplicity of infection (MOI) of 100 bacteria/amoeba in 10 mL of Soerensen buffer. After 1 h of co-incubation at 22◦C with agitation (180 rpm), the extracellular bacteria were removed by three sequential washing steps using Soerensen buffer. The infected amoebae were suspended in 10 mL of Soerensen buffer (t = 0) and further incubated at 22◦C with agitation (180 rpm). Aliquots were obtained at 0, 1, 3, 4.5, and 6 h post infection. The population of viable amoebae were determined at each time point. In parallel, infected amoebae recovered at each time point were lysed with 0.2% Triton X-100 and loads of intracellular bacteria were estimated by serial dilutions and plating on LB agar. Statistical significance was determined using a one-way ANOVA and two-way ANOVA with Fisher's LSD post-test.

### Global Proteomic Profiling Using Q-Exactive Mass Spectrometry

For proteomic analyses, ∼1 × 10<sup>6</sup> amoeba cells were obtained from individual intracellular survival assays after 6 h of infection with the wild-type strain or the 1ppk mutant. Uninfected amoebae were used as control condition. Amoebae from each experimental condition were concentrated by centrifugation at 500 × g for 10 min, quick-frozen, and kept at −80◦C until further use. Global proteomic profiles from samples representing the different experimental conditions were obtained from Bioproximity, LLC (USA). In each case, a unique proteomic analysis was performed using a pool of cells from three independent assays. Protein denaturation, digestion and desalting of samples were prepared using the filter-assisted sample preparation (FASP) method (Wi´sniewski et al., 2009). Briefly, the samples were digested using trypsin, and each digestion mixture was analyzed by ultra-high pressure liquid chromatography (UHPLC-MS/MS) coupled to a high resolution, high mass accuracy quadrupole-Orbitrap mass spectrometer (Q-Exactive, Thermo Fisher). For protein quantification, intensity measurements derived from the area-under-the-curve of the MS/MS scan for each peptide ion identification were summed for each sample. These values were averaged across all samples and a normalization factor determined and applied for each sample (Zhang et al., 2010). MS/MS data were compared with the most recent protein sequence libraries available from UniProtKB. Proteins were required to have one or more unique peptides detected across the analyzed samples with an e-value ≤ 0.0001.

Proteomes from each experimental condition were compared using an online tool that generates Venn diagrams and lists of proteins detected in any given condition (http://bioinformatics. psb.ugent.be/webtools/Venn/). The proteins detected in 2 experimental conditions were analyzed by calculating log<sup>2</sup> values of condition\_1/condition\_2 detection ratios. Enrichment for a protein in a particular condition was considered when the corresponding calculated value was >0.6 (∼1.5 fold enrichment).

## Analysis of *D. discoideum* Proteins Detected by Global Proteomic Profiling

D. discoideum proteins were identified from Q-proteomics data using library ID: 44689 (PubMed Taxonomy database), which include proteomes of strains AX2, AX3 y AX4. Proteins were classified according to the predicted functions annotated in the Clusters of Orthologous Groups of Proteins (COGs) database (Tatusov et al., 2000). The UniProtKB ID of each protein was mapped to the COGs database of the Social Amoeba Comparative Genome Browser (SACGB, http://sacgb.leibniz-fli.de/cgi/cog.pl? ssi=free), and the EggNOG database of orthologous groups and functional annotation (Jensen et al., 2008). To do this, both databases were downloaded, and mappings were performed using custom Python scripts. Using this approach we were able to assign COG categories for ∼80% of all proteins detected in each experimental condition.

In addition, overrepresentation analyses were performed using the Protein Annotation Through Evolutionary Relationship (PANTHER) tool (http://www.pantherdb.org/), a comprehensive system that combines gene function, ontology, pathways and statistical analysis tools that enable the analysis of largescale, genome-wide data from sequencing, proteomics or gene expression experiments (Mi et al., 2013, 2016). To do this, the UniProtKB IDs from the different sets of detected protein were used as input in PANTHER. This tool was able to map the UniProtKB ID for ∼96% of all proteins detected in each experimental condition. The results of overrepresentation analyses were filtered according to three annotation data sets: "Biological process", "Cellular component", and "Reactome pathways". The cut-off for the analyses was set to P < 0.05.

### Analysis of Bacterial Proteins Detected by Global Proteomic Profiling

S. Typhimurium proteins were identified from Q-proteomics data using the libraries ID: 588858 and ID: 99287 (PubMed Taxonomy database), which include proteomes of strains 14028s and LT2. Virulence-related proteins were assigned following inspections of databases PATRIC\_VF, VFDB and VICTORS, using tools available at the Pathosystems Resource Integration Center web site (PATRIC; www.patricbrc.org). Also, proteins encoded in genes located in close proximity to known virulence determinants, as well as genes located within Salmonella pathogenicity islands (SPIs) or prophages, were identified and analyzed using Artemis V.14, IslandViewer 3, Pathogenicity Island DataBase (PAIDB v2.0), PHAST and Islander (Rutherford et al., 2000; Zhou et al., 2011; Dhillon et al., 2015; Hudson et al., 2015; Yoon et al., 2015).

## RESULTS AND DISCUSSION

### Inorganic Polyphosphate Is Essential for *S.* Typhimurium Virulence in *D. discoideum*

To evaluate the role played by polyP biosynthesis in the virulence of S. Typhimurium in D. discoideum, we constructed a 1ppk derivative of the wild-type, virulent strain 14028s. This mutant is impaired in the synthesis of polyP. In addition, the mutant was transformed with a plasmid harboring a wild-type version of ppk (i.e., pPPK) to confirm the specificity of the effects attributed to the inactivation of this gene in our assays.

Previous reports indicate that virulent pathogenic bacteria delay the social development of D. discoideum, while attenuated or non-pathogenic bacteria allow its rapid progression (within 3–4 days in our experimental conditions). Thus, assessing the effect of a given bacterial strain on the social development of D. discoideum can be used to evaluate its virulence (Sillo et al., 2011; Bravo-Toncio et al., 2016; Ouertatani-Sakouhi et al., 2017). Notably, this host-pathogen interaction model has been recently used to identify compounds that inhibit bacterial virulence (Bravo-Toncio et al., 2016; Ouertatani-Sakouhi et al., 2017). Therefore, we compared the effect of feeding D. discoideum with the wild-type strain or its 1ppk derivative on the social development of the amoeba. In addition, a 1aroA derivative of strain 14028s (known to be attenuated in murine infection models Sebkova et al., 2008), and Escherichia coli B/r (routinely used to feed D. discoideum during growth under laboratory conditions Fey et al., 2007) were included in our assay as attenuated and non-pathogenic controls, respectively.

The different bacterial strains were inoculated on N agar and incubated overnight to generate bacterial lawns. Then, D. discoideum cells were deposited on top of each bacterial lawn and the plates were monitored for 6 days to follow the progression of D. discoideum development, which mainly involves three sequential stages: aggregation, elevation, and culmination (**Figure 1**). E. coli B/r allowed the rapid progression of the development, culminating within 3 to 4 days, where mature fruiting bodies were predominant (**Figure 1**). A similar phenotype was observed in the case of the 1aroA mutant, indicating that this strain is not virulent for D. discoideum. This observation is in agreement with previous reports indicating that an aroA mutant of S. Typhimurium is unable to survive intracellularly in this amoeba (Riquelme et al., 2016). On the contrary, the wild-type strain produced a delay in the development of the amoeba where only the aggregation phase was reached after 6 days of co-incubation (**Figure 1**). This phenotype was not observed in the case of the 1ppk mutant, which allowed the development of the amoebae until reaching the elevation phase at 6 days of co-incubation. Of note, the 1ppk mutant harboring plasmid pPPK showed a wild-type phenotype

development in presence of each bacterial strain. The development stages reached at days 2 and 4 are indicated with red and blue arrows, respectively. Scale bars, 100µm. (B) Development progress evaluated using a numerical scale defined according to the developmental phase reached at each time point (see Materials and Methods section). Graph shows mean values ± SD from 10 independent assays.

(**Figure 1**). These results indicate that polyP synthesis is essential for S. Typhimurium virulence in D. discoideum.

Our observations on the interaction of S. Typhimurium with D. discoideum are consistent with a previous study that evaluated the virulence of the pathogen by assessing its effect on the social development of the amoeba. The authors reported that Dictyostelium growth on a lawn of wild-type S. Typhimurium causes the aberrant development of the amoeba, or even resulted in cell death depending on nutrient conditions of the medium used for the assay. Both phenotypes required a functional T3SS-2 (Sillo et al., 2011). Additionally, the effect observed for a polyP deficiency on the social development of the amoeba has been also reported in the case of P. aeruginosa (Zhang et al., 2005; Bravo-Toncio et al., 2016) highlighting the essential role played by polyP biosynthesis in bacterial virulence using this model organism.

### Inorganic Polyphosphate Is Essential for *S.* Typhimurium Survival in *D. discoideum*

Recently, we reported that wild-type S. Typhimurium can survive within D. discoideum and requires relevant genes linked to virulence for this process (Riquelme et al., 2016). To evaluate the role played by polyP in the intracellular survival of S. Typhimurium in D. discoideum, we performed infection assays where vegetative amoebae were co-incubated with the wild-type strain or its 1ppk derivative. At different times post infection, intracellular bacteria were recovered from infected amoebae and titrated. The attenuated 1aroA mutant was included as a control in our assays.

First, we evaluated the internalization of each mutant strain after 1 h of co-incubation with the amoebae, and observed that 1ppk and 1aroA mutants were internalized at higher levels than the wild-type strain. In contrast, the 1ppk mutant harboring plasmid pPPK was internalized at wild-type levels (**Figure 2A**). Then, we evaluated the intracellular survival of each strain at different times post infection and observed that the wild-type strain was able to survive and replicate in the amoebae. On the contrary, the 1ppk mutant was defective for intracellular survival at all time points evaluated. The same phenotype was observed in the case of the attenuated 1aroA mutant. The intracellular survival of the 1ppk mutant harboring plasmid pPPK was comparable to that shown by the wild-type strain (**Figure 2B**). It is worth mentioning that no effect in amoeba viability was observed during the course of these experiments (**Figure 2C**), indicating that the differences observed in the titers of intracellular bacteria are not attributable to changes in the number of viable amoebae.

Overall, our results indicate that polyP synthesis is essential for S. Typhimurium to survive intracellularly in D. discoideum. These observations are in line with a previous study indicating that a 1ppk mutant of S. Typhimurium is deficient for intracellular survival in RAW 264.7 murine macrophages (Kim et al., 2002). In addition, several studied indicate that S. Typhimurium ppk mutants present a variety of phenotypes, including defective long-term survival in vitro, defective responses to oxidative stress and starvation, sensitivity to

polymyxin, intolerance to acid and heat, impaired invasiveness in HEp-2 epithelial cells, and loss of swimming motility, all of which strongly influence virulence (Kim et al., 2002; McMeechan et al., 2007; Cheng and Sun, 2009). Accordingly, it has been reported that a ppk mutant of S. Typhimurium was attenuated in orally-infected Rhode Island Red chickens and BALB/c mice (McMeechan et al., 2007).

### Global Proteomic Profiling of *Dictyostelium*-*Salmonella* Interaction

In order to determine the global response of D. discoideum to infections with S. Typhimurium wild type or its 1ppk mutant derivative, we performed a global proteomic profiling of such interactions. To achieve this, the amoebae were co-incubated with each bacterial strain until reaching 6 h of infection. This time was chosen because the intracellular survival of both strains in D. discoideum showed the highest differences (**Figure 2B**). Non-infected amoebae were used as control condition. Next, shotgun proteomic profiling of infected and control amoebae were performed by UHPLC-MS/MS (Q-proteomics). Thus, 1779, 1950, and 1850 proteins were detected in samples of amoebae infected with the wild-type strain, amoebae infected with the 1ppk mutant, and uninfected amoebae, respectively (**Table S1**).

A total of 258, 250, and 336 proteins were exclusively detected in non-infected amoebae, in amoebae infected with the wildtype strain, or in amoebae infected with the 1ppk mutant, respectively. Additionally, 1277 proteins were detected in all three experimental conditions tested (**Figure 3A**). Considering that these proteins could be interesting if they show significant differences in expression levels, we carried out a comparative analysis of proteins detected in pairs of experimental conditions

FIGURE 3 | COG functional categorization of D. discoideum proteins detected during infection with S. Typhimurium strains. (A) Venn diagram of D. discoideum proteins detected in uninfected amoebae or in amoebae infected with S. Typhimurium wild-type or its 1ppk derivative. (B) Graph showing number of proteins detected in each experimental condition and classified according to COG functional categories (see Materials and Methods section). COG categories were further grouped in three main classes: "Cellular processes and signaling", "Information storage and processing", and "Metabolism".

(i.e., non-infected amoebae vs. infected with the wild-type strain; non infected amoebae vs. infected with the 1ppk mutant; and amoebae infected with the wild-type strain vs. infected with the 1ppk mutant) to determine enrichment in a particular condition (cut-off: ∼1.5 fold change). Proteins found exclusively or enriched in a given experimental condition were classified according to COG categories, which in turn were grouped in three main classes: "Cellular processes and signaling", "Information storage and processing", and "Metabolism" (**Table S2**). For most COG categories, the total number of proteins detected from non-infected amoebae was similar to those detected from amoebae infected with the 1ppk mutant, in contrast to proteins detected in amoebae infected with the wild-type strain (**Figure 3B**). This was particularly evident in the case of COG categories "Posttranslational modification, protein turnover, chaperones", "Signal transduction mechanisms", "Intracellular trafficking, secretion, and vesicular transport" (associated with "Cellular processes and signaling"), "Chromatin structure and dynamics", "Translation, ribosomal structure, and biogenesis", "Mobilome, prophages, and transposons" (associated with "Information storage and processing"), "Energy production and conversion", "Nucleotide transport and metabolism", "Carbohydrate transport and metabolism", "Coenzyme transport and metabolism", "Lipid transport an metabolism", and "Secondary metabolites biosynthesis, transport and catabolism" (associated with "Metabolism"; **Figure 3B**).

Additionally, proteins that were exclusive or significantly enriched in each experimental condition were identified and used to perform overrepresentation analyses using the PANTHER tool (**Table S3**). The results were filtered according to three annotation data sets: "Biological process", "Cellular component", and "Reactome pathways" (**Table 3**). Overrepresented groups of proteins detected in D. discoideum infected with either S. Typhimurium strain include those involved in endomembrane trafficking, actin cytoskeleton organization, social development, chemotaxis and response to cAMP, immune system, response to bacteria, ubiquitination and proteasome degradation.

Regarding endomembrane trafficking, common overrepresented proteins include those involved in Secretion of lysosomal enzymes (GO:0033299), Exocytosis (GO:0006887), Phagocytosis (GO:0006909), Early phagosome (GO:0032009), Endosome (GO:0005768), Signaling by Rho GTPases (R-DDI-194315), Rho GTPases Activate WASPs and WAVEs (R-DDI-5663213), and Fc gamma receptor (FCGR) dependent phagocytosis (R-DDI-2029480). Overrepresented proteins in amoebae infected with the wild-type strain include those involved in Rab protein signal transduction (GO:0032482), Positive regulation of guanyl-nucleotide exchange factor activity (GO:1905099), Regulation of vacuole fusion, non-autophagic (GO:0032889), Lysosomal lumen acidification (GO:0007042), Protein localization to lysosome (GO:0061462), Vesicle transport along microtubule (GO:0047496), Phosphatidylinositol phosphorylation (GO:0046854), Phosphatidylinositol-mediated signaling (GO:0048015), G beta:gamma signaling through PI3Kgamma (R-DDI-392451), and PI3K Cascade (R-DDI-109704). On the other hand, overrepresented proteins in amoebae infected with the 1ppk mutant include those linked to Phosphatidylinositol 3-kinase signaling (GO:0014065), Phosphatidylinositol 3-kinase complex (GO:0005942), Regulation of exocytosis (GO:0017157), Vacuolar acidification (GO:0007035), Late endosome to vacuole transport (GO:0045324), Establishment of protein localization to vacuole (GO:0072666), Post-lysosomal vacuole (GO:0032195), Late endosome (GO:00057770), and Lysosome (GO:0005764). It is well known that S. Typhimurium delivers T3SS-2 effector proteins that interfere with the maturation of the endocytic route in eukaryotic host cells in order to avoid phagolysosomal fusion. This process results in a unique vacuolar compartment referred to as the Salmonella-containing vacuole (SCV), where this pathogen resides (Haraga et al., 2008; LaRock et al., 2015). We have described that inactivation of T3SS-2 abolishes the intracellular survival of S. Typhimurium in D. discoideum (Riquelme et al., 2016). Accordingly, our analysis suggest that wild-type S. Typhimurium resides in an intracellular compartment of D. discoideum comparable to an early endosome, while the 1ppk mutant resides in a compartment that ultimately fuses with the lysosome, explaining the defective intracellular survival phenotype shown by this strain in the amoeba (**Figure 2**).

Most overrepresented proteins associated with the response to bacterial infection were detected in amoeba infected with either strain of S. Typhimurium. These proteins include those linked to Immune System (R-DDI-168256), Innate Immune System (R-DDI-168249), Adaptive Immune System (R-DDI-1280218), Cross-presentation of soluble exogenous antigens (endosomes) (R-DDI-1236978), Antigen processing: Ubiquitination & Proteasome degradation (R-DDI-983168), and ROS, RNS production in response to bacteria (R-DDI-1222556). No overrepresented proteins were exclusively detected in infections with the wild-type strain, while overrepresented proteins only detected during infections with the 1ppk mutant include those associated with Metabolism of nitric oxide (R-DDI-202131) and eNOS activation and regulation (R-DDI-203765). These proteins (SprA/Q54GP3, PtsA/Q1ZXI0, and GchA/Q94465) are required for the de novo biosynthesis of tetrahydrobiopterin, an essential co-factor for the aromatic amino acid hydroxylases and nitric oxide synthases in mammals (Thöny et al., 2000; Choi et al., 2006; Vásquez-Vivar, 2009). These results indicate that D. discoideum infected with either wildtype or 1ppk S. Typhimurium generates a robust response that includes production of ROS and RNS in order to eliminate the pathogen.

Proteins exclusively detected in D. discoideum infected with the wild-type strain indicate that the pathogen induces DNA damage in the host. This is revealed by a number of proteins involved in Positive regulation of single strand break repair (GO:1903518), Positive regulation of DNA repair (GO:0045739), DNA ligation involved in DNA repair (GO:0051103), Site of double-strand break (GO:0035861), and Cell death signaling via NRAGE, NRIF, and NADE (R-DDI-204998), the latter being a process associated with apoptotic cell death. It is tempting to speculate that this DNA damage is the result of excessive ROS/RNS production in response to the pathogen. TABLE 3 | Overrepresentation analysis of D. discoideum proteins detected during infections with S. Typhimurium wild type and 1ppk.

#### A. *D. discoideum* infected with *S*. Typhimurium wild type


(Continued)

#### TABLE 3 | Continued


#### B. *D. discoideum* infected with *S*. Typhimurium 1*ppk*




(Continued)

#### TABLE 3 | Continued


Overrepresentation analyses were performed using PANTHER (http://www.pantherdb.org/). UniProtKB IDs from exclusive and significantly-enriched proteins of D. discoideum detected during infections with the wild-type strain or the ∆ppk mutant were used as input. The results were filtered according to 3 annotation data sets: "Biological process," "Cellular component," and "Reactome pathways." The cut-off for the analyses was set to p < 0.05.

Furthermore, this DNA damage needs to be repaired by the amoeba in order to resume growth and development. This is in agreement with the ability of S. Typhimurium to delay the social development of D. discoideum according to our development assay (**Figure 1**). In addition, two proteins linked to Protein sulfation (GO:0006477) (i.e., Kil1/Q556K8 and Phg1a/Q55FP0) were only detected in amoeba infected with the wild-type strain. Both proteins have been implicated in D. discoideum killing of intracellular K. pneumoniae (Benghezal et al., 2006; Cosson and Soldati, 2008; Le Coadic et al., 2013). Of note, protein sulfation has been implicated in decreased bacterial adherence to eukaryotic cells and reduced T3SS-dependent cytotoxicity (Blondel et al., 2016). Thus, variations in sulfation levels on the surface of D. discoideum generated during S. Typhimurium infection may explain differences in internalization observed between the wild-type and attenuated strains 1ppk and 1aroA in our infection assays (**Figure 2A**).

Intracellular pathogens need to cope with a hostile environment inside the host during infection. Thus, our global proteomic profiles of D. discoideum proteins suggest that the amoeba recognizes wild-type S. Typhimurium and elicits a strong response that includes production of toxic ROS and RNS. Nevertheless, the pathogen manipulates the endocytic pathway of the host to generate an intracellular replicative niche to survive this cell-autonomous defense response. On the contrary, after recognition the 1ppk mutant appears to be unable to subvert the amoebal autonomous defense mechanism and to survive the unfavorable conditions within the host. This idea is in agreement with the phenotypes reported for null mutants of ppk in several bacterial pathogens (Rao et al., 1998; Rashid and Kornberg, 2000; Rashid et al., 2000a,b; Kim et al., 2002; McMeechan et al., 2007; Brown and Kornberg, 2008; Varela et al., 2010; Gray and Jakob, 2015; Varas et al., 2017). In addition the mutant seems to be incapable to modify the endocytic pathway, ending in a degradative intracellular compartment. Consequently, the 1ppk mutant is unable to survive within the amoeba and to subvert the social development of this host.

In addition to D. discoideum proteins, we attempted to identify S. Typhimurium proteins expressed during the infection of the amoeba with the wild-type strain or the 1ppk mutant. Using our Q-proteomics approach we were able to detect a limited number of bacterial proteins, most probably due to their low relative abundance in each sample in comparison to D. discoideum proteins. Thus, a total of 54 and 34 proteins were identified in infections of amoeba cells with the wild-type strain and the 1ppk mutant, respectively. A group of seven proteins was detected in amoebae infected with either strain (**Figure 4A**). The identities of all S. Typhimurium proteins detected are listed in **Table 4**. The location of genes encoding all these proteins in the genome of S. Typhimurium 14028s is shown in **Figure 4B**. The list of detected proteins was compared with a list of 469 classic virulence-related proteins included in databases PATRIC\_VF, VFDB and VICTORS using tools implemented in PATRIC (www. patricbrc.org). We identified 11 S. Typhimurium proteins linked to virulence in amoebae infected with the wild-type strain, and 4 of these proteins in amoebae infected with the 1ppk mutant, respectively (proteins highlighted in bold type in **Table 4**).

The limited amount of S. Typhimurium proteins detected in our global proteomic profiling impeded us conducting an insightful comparative analysis in order to understand differences between infections of D. discoideum with the wildtype strain and the 1ppk mutant. However, it is worth mentioning a particular group of S. Typhimurium proteins detected during amoebae infections that includes AroA, SprB, STM14\_0329, H-NS, Fis, ArcB, NuoG, EntC, IbpB, CsgB, StiC, and FliK.

AroA is a 3-enolpyruvylshikimate-5-phosphate synthetase that is crucial for biosynthesis of aromatic compounds. As a result, Salmonella aroA null-mutants are highly attenuated in different infection models (Hoiseth and Stocker, 1981; Stocker et al., 1983; Cooper et al., 1990) and present strong defects in survival within macrophages (Fields et al., 1986; Lowe et al., 1999) and in D. discoideum (Riquelme et al., 2016). This protein was detected in amoebae infected with both S. Typhimurium wild-type and 1ppk strains.

SprB is a transcription factor from the LuxR/UhaP family that is encoded in SPI-1. This protein regulates the coordinate expression of SPI-1 and SPI-4 genes during Salmonella infection (Saini and Rao, 2010). STM14\_0329 (also known as SciO and TssK) is one of the 13 core components of the T6SS encoded

during infection of D. discoideum with the wild-type strain or its 1ppk derivative. (B) Circular representation of S. Typhimurium 14028s chromosome. Boxes in magenta and gray indicate regions corresponding to known Salmonella pathogenicity islands (SPIs) and prophages, respectively. Blue and green lines indicate the location of genes encoding proteins detected during the infection with the wild-type strain or its 1ppk derivative, respectively. Yellow lines indicate the location of genes encoding virulence-associated proteins listed in databases PATRIC\_VF, VFDB, and VICTORS (see Materials and Methods section).

in SPI-6 (Blondel et al., 2009; Journet and Cascales, 2016). T6SS are versatile weapons exploited by numerous bacterial pathogens to target either eukaryotic host cells or competitor bacteria (Cianfanelli et al., 2016; Hachani et al., 2016; Journet and Cascales, 2016). Noteworthy, we have recently described that null mutations in S. Typhimurium genes encoding essential components of T3SS-1 and SPI-6 T6SS cause intracellular survival defects in D. discoideum (Riquelme et al., 2016).

The nucleoid-associated protein H-NS selectively silences horizontally-acquired genes by direct binding to DNA sequences with high TA content in the genome, including all major SPIs in S. Typhimurium (Lucchini et al., 2006; Navarre et al., 2006). Accordingly, hns mutations are highly pleiotropic in S. Typhimurium (Hinton et al., 1992) and produce attenuated strains (Harrison et al., 1994). A recent study indicates that the fitness defects presented by hns mutants of S. Typhimurium are mainly due to a misregulation of SPI-1 genes (Ali et al., 2014). The factor for inversion stimulation (Fis) is a nucleoid-associated protein that influences the topological state of DNA in the cell by direct binding to DNA and by modulating DNA gyrase and topoisomerase I gene expression. As in the case of H-NS, Fis acts as a key regulator of virulence in S. Typhimurium mainly by controlling the coordinate expression of genes located in several SPIs, as well as genes involved in motility (reviewed in Duprey et al., 2014). Therefore, fis mutants of S. Typhimurium are defective for intracellular survival in macrophages (O Cróinín et al., 2006; Wang et al., 2013). ArcB is the sensor component of the master regulatory two-component system ArcA/ArcB, that controls the expression of several genes and operons encoding proteins linked to the metabolic shift from anaerobic to aerobic conditions, and the enzymatic defenses of bacteria against ROS (Evans et al., 2011). It is worth mentioning that H-NS, Fis and ArcB were only detected in amoebae infected with the 1ppk mutant, perhaps reflecting adjustments required to cope with the pleiotropic phenotypes presented by this kind of mutant.

NuoG is a subunit of the NADH dehydrogenase I complex. S. Gallinarum 1nuoG mutants are attenuated in chicken and show reduced survival and multiplication in the reticuloendothelial system of this host (Zhang-Barber et al., 1998; Turner et al., 2003). Noteworthy, it has been reported that NuoG is involved in detoxification of ROS produced by macrophages during M. tuberculosis infection (Miller et al., 2010). Thus, it is tempting to speculate that NuoG can play a similar role during S. Typhimurium infection of D. discoideum.

EntC is an isochorismate synthase involved in the biosynthesis of catecholate siderophores enterobactin and salmochelin, produced by Salmonella (and other bacterial pathogens) to capture iron from the host during infection (reviewed in Fischbach et al., 2006). It has been established that production of salmochelin is essential for full virulence of S. Typhimurium in mice (Crouch et al., 2008). Furthermore, the production of enterobactin and salmochelin is required for S. Typhimurium to survive in macrophages at early stages of infection, and these siderophores protect the pathogen against reactive oxygen species produced by macrophages during the infective process (Achard et al., 2013). Our results suggest that S. Typhimurium produces catecholate siderophores inside D. discoideum, perhaps contributing to the intracellular survival of this pathogen.

IbpB is a small heat-shock protein (sHSP) being member of a widely conserved family of ATP-independent molecular chaperones that bind to misfolded proteins and protect them from irreversible aggregation (Laskowska et al., 1996; Lee et al., 1997). It has been reported that E. coli chaperones IbpB and IbpA are substrates for the ATP-dependent Lon protease (Bissonnette TABLE 4 | S. Typhimurium proteins detected during infections of D. discoideum with the wild-type strain or the 1ppk mutant.


(Continued)


Bold type indicates proteins detected during infection with either the wild-type strain or the 1ppk mutant. Red indicates classic virulence-related proteins included in databases PATRIC\_VF, VFDB, and VICTORS that were identified using tools implemented in PATRIC (www.patricbrc.org).

et al., 2010). In addition, polyP forms a complex with Lon and stimulates the degradation of selected proteins (Kuroda et al., 2001, 2006; Nomura et al., 2004; Kuroda, 2006). Thus, there is a functional link between IbpB, Lon and the biosynthesis of polyP during the bacterial stress response.

Among the bacterial proteins detected in amoebae infected with wild-type S. Typhimurium we found CsgB, StiC, and FliK, which are associated to the assembly of proteinaceous structures such as fimbriae and flagellum, respectively. This is noteworthy because it is generally accepted that this kind of surface structures are repressed upon invasion of host cells. CsgB participates in the assembly of the curli fimbriae, favoring the polymerization of its major component CsgA (reviewed in Evans and Chapman, 2014). Curli fimbriae are amyloid fibers that act as scaffolding agents in biofilms of E. coli and Salmonella, providing increased resistance to desiccation and to sodium hypochlorite, and being extremely resistant to proteolysis and chemical denaturation (Chapman et al., 2002; White et al., 2006). These fimbriae have been linked to cell-cell contacts, and to adherence to various eukaryotic cells, tissues, and abiotic surfaces, thus promoting community behavior and host colonization, playing an important role in the initial stages of the infection process (Barnhart and Chapman, 2006). To our knowledge, there are no previous reports on the expression of CsgB when Salmonella resides inside host cells. StiC is an usher protein encoded in one of the 11 chaperone-usher fimbrial gene clusters in the genome of S. Typhimurium (Jarvik et al., 2010). Chaperone-usher fimbriae normally have one or more structural subunits, which are exported and assembled on the bacterial surface by cognate periplasmic chaperone proteins and an outer-membrane usher protein. Some of them have demonstrated roles in binding to different receptors, persisting in specific niches, promoting infections, or forming biofilms (Weening et al., 2005; Clayton et al., 2008; Yue et al., 2012). The stiABCH gene cluster encodes a class γ-fimbriae, characterized for harboring subunits comprising the domains PFAM00419 or COG3539 (Nuccio and Bäumler, 2007). To our knowledge, no previous studies have addressed the specific contribution of this fimbrial operon to S. Typhimurium virulence. However, a previous study showed that the chaperone-usher SEF14 fimbriae are essential for an efficient uptake and survival of S. Enteritidis in murine macrophages, suggesting that they may be required at stages beyond the initial host colonization (Edwards et al., 2000). FliK is the protein that controls the length of the flagellar hook, and is encoded in one of the 17 operons composing the flagellar regulon of S. Typhimurium (Chilcott and Hughes, 2000). The flagellum is required for bacterial access to the intestinal epithelium, adherence to several tissues, and immune modulation (Rossez et al., 2015). Although flagellum assembly is normally prevented inside host cells, it was previously reported that intracellular Salmonella triggers swelling of macrophages (referred to as "oncotic macrophages") in a process where flagellated bacilli intermittently escape from infected host cells (Sano et al., 2007).

#### CONCLUSIONS

Overall, our results indicate that polyP biosynthesis is crucial for virulence and intracellular survival of S. Typhimurium in D. discoideum. In addition, we have validated the use of global proteomic analyses to gain insight into the host-pathogen interaction of D. discoideum and S. Typhimurium. The analysis of host proteins related to endocytic pathway, immune response,

### REFERENCES


cell death, cytoskeleton dynamics, and developmental process revealed mechanisms that may explain the phenotypes shown by a S. Typhimurium strain lacking polyP during D. discoideum infection. Thus, our work demonstrates that unbiased highthroughput proteomics can be used as a powerful approach to provide new perspectives on host-pathogen interactions. Furthermore, our infection and development assays using these organisms can be exploited to screen for novel anti-virulence molecules targeting inorganic polyP biosynthesis.

### AUTHOR CONTRIBUTIONS

MV, SR-B, CV, CS, and FC: Conceived and designed the experiments; MV, SR-B, CV, and AM: Performed the experiments; MV, SR-B, CV, AM, CB-P, CS, and FC: Analyzed the data; MV, CB-P, CS, and FC: Contributed with reagents/animals/materials/analysis tools; MV, AM, CS, and FC: Wrote the paper. All authors read and approved the final manuscript.

### FUNDING

This work was supported in part by FONDECYT grants 1120209 (to FC), 1140754 and 1171844 (to CS). MV was supported by CONICYT fellowship 21120431 and FONDECYT grant 3170449. SR-B and CV were supported by CONICYT fellowships 21160818 and 21140615, respectively.

#### ACKNOWLEDGMENTS

We are indebted to Nicole Molina for technical support.

### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00008/full#supplementary-material

Table S1 | List of all proteins detected in uninfected amoebae, and in amoebae infected with S. Typhimurium wild type or 1ppk, and their corresponding PANTHER annotation.

Table S2 | Proteins found exclusively or enriched in a given experimental condition, classified according to COG categories.

Table S3 | Overrepresentation analysis for proteins found exclusively or enriched in amoebae infected with S. Typhimurium wild type or 1ppk.


secretion systems differentially distributed in Salmonella serotypes. BMC Genomics 10:354. doi: 10.1186/1471-2164-10-354


and Dublin in chickens and mice. Infect. Immun. 71, 3392–3401. doi: 10.1128/IAI.71.6.3392-3401.2003


**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 Varas, Riquelme-Barrios, Valenzuela, Marcoleta, Berríos-Pastén, Santiviago and Chávez. 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.

# Quantification of Live Bacterial Sensing for Chemotaxis and Phagocytosis and of Macropinocytosis

#### Netra P. Meena and Alan R. Kimmel\*

Laboratory of Cellular and Developmental Biology, National Institute of Diabetes and Digestive and Kidney Diseases, The National Institutes of Health, Bethesda, MD, United States

Initial immunological defense mechanisms to pathogen invasion rely on innate pathways of chemotaxis and phagocytosis, original to ancient phagocytes. Although chemotaxis has been well-studied in mammalian and model systems using purified chemoattractants in defined conditions, directed movement toward live bacteria has been more difficult to assess. Dictyostelium discoideum is a professional phagocyte that chemotaxes toward bacteria during growth-phase in a process to locate nutrient sources. Using Dictyostelium as a model, we have developed a system that is able to quantify chemotaxis to very high sensitivity. Here, Dictyostelium can detect various chemoattractants at concentrations <1 nM. Given this exceedingly sensitive signal response, Dictyostelium will migrate directionally toward live gram positive and gram negative bacteria, in a highly quantifiable manner, and dependent upon bacterially-secreted chemoattractants. Additionally, we have developed a real-time, quantitative assay for phagocytosis of live gram positive and gram negative bacteria. To extend the analyses of endocytic functions, we further modified the system to quantify cellular uptake via macropinocytosis of smaller (<100 kDa) molecules. These various approaches provide novel means to dissect potential for identification of novel chemoattractants and mechanistic factors that are essential for chemotaxis, phagocytosis, and/or macropinocytosis and for more detailed understanding in host-pathogen interactive defenses.

#### Edited by:

Thierry Soldati, Université de Genève, Switzerland

#### Reviewed by:

Robert Roger Kay, Medical Research Council, United Kingdom Salvatore Bozzaro, Università degli Studi di Torino, Italy

#### \*Correspondence: Alan R. Kimmel alank@helix.nih.gov

Received: 31 October 2017

Accepted: 13 February 2018 Published: 02 March 2018

#### Citation:

Meena NP and Kimmel AR (2018) Quantification of Live Bacterial Sensing for Chemotaxis and Phagocytosis and of Macropinocytosis. Front. Cell. Infect. Microbiol. 8:62. doi: 10.3389/fcimb.2018.00062 Keywords: cell migration, Bacillus, Pseudomonas, E. coli, pterin, cAMP, innate immunity, Dictyostelium

## 1. INTRODUCTION

Host–pathogen interactions are crucial aspects to the immune response. Innate immunity comprises generalized defense to invading microorganisms and involves cellular functions, including chemotaxis and phagocytosis, that can be traced at least 1 billion years (Cosson and Soldati, 2008; Jin et al., 2008; Bozzaro and Eichinger, 2011), where ancient professional phagocytes utilized the pathways in a search for and an uptake of new nutrient sources. Thus, similar molecular processes, including receptor signaling and directed cytoskeletal mobilization, are essential for both destruction of pathogenic bacteria by macrophage and neutrophils, and for nutrient capture by Dictyostelium discoideum (Cosson and Soldati, 2008; Jin et al., 2008; Bozzaro and Eichinger, 2011). Significantly, the ease in genetic manipulation of Dictyostelium (Basu et al., 2013) has made it an ideal model for mechanistic studies of these processes.

Dictyostelium can rapidly sense multiple chemoattractants in very broad and dynamic concentration gradients (Jin, 2013; Artemenko et al., 2014; Nichols et al., 2015). We established a new quantitative system for high-throughput chemoattractant gradient perception at very low concentration (<1 nM) sensitivity and then extended conditions for use of live bacteria as the source for secreted chemoattractants (Meena and Kimmel, 2017). This further permitted analyses to understand simultaneous response to multiple chemoattractants and to identify bacterially secreted factors. Beyond, we have modified the system to examine functions secondary to chemotaxis that are essential for target removal through phagocytosis and macropinocytosis.

Procedural details are, thus, presented for quantification of chemotaxis in real-time to both purified ligands, individually, or in multiple combinations, or to live bacteria. We also outline specific approaches for quantifying cellular uptake of large (>1µm) particles by phagocytosis or of molecules (<100 kDa) by macropinocytosis, again in a real-time and high-throughput mode.

### 2. MATERIALS AND EQUIPMENT

#### 2.1. Solutions


#### 2.2. Supplies


### 3. PROCEDURES AND RESULTS

#### 3.1. Chemotaxis by Growth Phase Dictyostelium to Chemoattractants, Folate, Methopterin, Pterin, and cAMP

#### 3.1.1. Growth and Preparation of Dictyostelium


#### 3.1.2. Quantification of Dictyostelium Chemotaxis to Chemical Ligands


#### 3.1.3. Results

Dictyostelium grow as single cells that migrate to bacterially secreted pterin-derived folate moieties and to cAMP (Meena and Kimmel, 2017), in a search for nutrient sources. Assaying chemotaxis of growth-phase Dictyostelium had been a great challenge (Veltman et al., 2014). However, we developed a highly sensitive, quantitative, and reproducible method for chemotaxis of growth-phase Dictyostelium that is easily modified and optimized for most single-cell types (Meena and Kimmel, 2017).

The approach adapts the IncuCyte <sup>R</sup> live-cell analysis system from Essen Bioscience for real-time imaging of chemotaxis in a 96-well format. IncuCyte <sup>R</sup> uses high definition phase contrast imaging and ClearView Cell Migration Plates. The scan intervals, digital zoom, and imaging channels are set manually. But following, image acquisition is fully automated, as are integrated real-time data analyses and full time-course metrics, with application of the IncuCyte <sup>R</sup> Chemotaxis Cell Migration Software Module. The specialized 96-well chemotaxis dish is comprised of three sections: a transparent lid, an upper well plate (which will contain cells to be assayed), and a bottom well plate (which will have chemoattractants). A membrane of uniformly-spaced 8µm pores separates each aligned upper and bottom chamber well.

Chemoattractant gradients are established by chemical diffusion from the bottom wells, through the membrane pores, into the upper wells. Cell migration results from the movement of cells out of the upper well, through the membrane toward the bottom well, and chemotaxis is quantified over time, by imaging cells that had migrated from the upper well base to the under surface of the membrane within the bottom chamber. The assays are highly reproducible, with generally <10% non-specific directed migration to phosphate buffer controls [**Figure 1** (Meena and Kimmel, 2017)]. In addition, by comparing migration toward buffer controls, using cells incubated in the presence or absence of chemoattractants, we can assess a potential chemokinetic contribution to random

FIGURE 1 | Chemotactic dose-responses to different chemoattractants. (A) Time-course quantification of Dictyostelium migration to various dose concentrations of folate. Relative chemotaxis is normalized to 500 nM folate at 1.5 h; no additional chemotaxis is observed with folate increased to 5,000 nM (see Figure 2). Standard deviations are shown based upon three replicates. The structure of folate is shown. (B) Time-course quantification of Dictyostelium migration to various dose concentrations of methopterin. Relative chemotaxis is normalized to 500 nM methopterin at 1.5 h. Standard deviations are shown based upon three replicates. The structure of methopterin is shown. (C) Time-course quantification of Dictyostelium migration to various dose concentrations of pterin. Relative chemotaxis is normalized to 5,000 nM pterin at 1.5 h. Standard deviations are shown based upon three replicates. The structure of pterin is shown. (D) Time-course quantification of Dictyostelium migration to various dose concentrations of cAMP. Relative chemotaxis is normalized to 500 nM cAMP at 1.5 h; no additional chemotaxis is observed with cAMP increased to 5,000 nM (see Figure 2). Standard deviations are shown based upon three replicates. The structure of cAMP is shown.

migration (Meena and Kimmel, 2017). Chemokinesis may increase operative backgrounds to ∼15%, for fully responsive cells (Meena and Kimmel, 2017).

**Figure 1** illustrates a typical chemotactic dose-response, timecourse for migration of growth-phase Dictyostelium toward four different chemoattractants. Folate, methopterin, and pterin represent 3 structurally related moieties that interact with the same chemoattractant receptor (NPM and ARK, in preparation), but at differing activating affinities, in the order of methopterin>folate>pterin (NPM and ARK, in preparation); the EC<sup>50</sup> for chemotaxis of each is consistent with these affinity differences (**Figure 2**). Separately, we also show high chemosensitivity of growth-phase Dictyostelium to the unrelated chemoattractant cAMP (**Figures 1D**, **2**), which interacts with a distinct chemoattractant receptor (Liao et al., 2013; Meena and Kimmel, 2016, 2017).

### 3.2. Chemotaxis of Growth Phase Dictyostelium to Live Bacteria

#### 3.2.1. Growth and Preparation of Growth Phase Dictyostelium

Follow all steps in **3.1.1.**

#### 3.2.2. Growth of Bacteria


#### 3.2.3. Preparation of Bacteria for Chemotaxis


#### 3.2.4. Quantification of Chemotaxis Using Bacteria as Signaling Agents


#### 3.2.5. Results

The response sensitivity of growth-phase Dictyostelium to <1 nM concentrations of various chemoattractants suggested that the assay might be sufficiently effective to detect directed migration to live bacteria secreting endogenous compounds, which act as chemoattractants to Dictyostelium. Indeed, we had previously demonstrated that Dictyostelium would chemotax to both Klebsiela planticola and Bacillus subtilis (Meena and Kimmel, 2017). Relative chemotaxis was directly correlated to cell number, indicating an excellent dose-dependent quantifiable response. We also showed that chemotaxis relied on folate- and/or cAMPdependent signaling pathways and that secreted cAMP could be detected by bioassay (Meena and Kimmel, 2017).

Here, we extend the system to define conditions for quantifiable chemotaxis to other live bacterial species. Dictyostelium show strong chemotactic response to both gram positive (Bacillus cereus, Micrococcus luteus) and gram negative (Escherichia coli, Pseudomonas fluorescens) bacteria (**Figure 3**), in a manner similar to that of purified chemoattractants (see **Figure 1**). Physio-chemical analyses confirm the secretion of cAMP and/or folate, by the various bacteria.

#### 3.3. Phagocytosis of Live Bacteria by Growth Phase Dictyostelium

#### 3.3.1. Growth and Preparation of Growth Phase Dictyostelium

Follow all steps in **3.1.1.**

3.3.2. Growth and Preparation of Bacteria Follow all steps in **3.2.2.** and **3.2.3.**

#### 3.3.3. Preparation of pHrodo®-Labeled Live Bacteria

**3.3.3.1.** pHrodoTM is a fluorogenic dye that exhibits a dramatic increase in fluorescence at an acidified pH. Thus, pHrodo <sup>R</sup> -labeled bacteria show minimal basal fluorescence, but upon phagocytosis by Dictyostelium and transport to lysosomes, detected fluorescence increases. ∼2 x 10<sup>9</sup> bacteria from **3.2.3.** were centrifuged

cereus, (B) gram positive (Gr+) Micrococcus luteus, (C) gram negative (Gr -) Pseudomonas fluorescens, and (D) gram negative (Gr–) Escherichia coli; buffer controls are included for each. Relative chemotaxis is normalized to values at 1.5 h. Standard deviations are shown based upon three replicates.

for 2 min at 12,000 rpm and resuspended in ∼50 µL of Component C (ThermoFisher Scientific, pHrodoTM Phagocytosis particle labeling kit). Succinimidyl ester dye (10 mM) was added to 0.5 mM and incubated for 45 min at room temperature. To prevent cell lethality during pHrodo <sup>R</sup> -labeling, toxic compounds (e.g., methanol) were eliminated from the labeling.

**3.3.3.2.** To remove non-complexed dye, bacteria were washed three times at 12,000 rpm with Component C. All pHrodo <sup>R</sup> -labeled bacteria species were viable, as observed by cell division and motility, when re-cultured in growth medium.

#### 3.3.4. Quantification of Phagocytosis of Live Bacteria


(nm) 560/585]), a desired time (e.g., 1.5–4 h), at 15 min intervals. Export the data into Microsoft Excel for processing.

#### 3.3.5. Phagocytosis of pHrodo-Labeled Bacteria Using Immunofluorescence Microscopy


#### 3.3.6. Results

In addition to chemotaxis, professional phagocytes are defined by an ability to engulf large (>1µM) particles, for nutrient uptake, directed killing of invading unicellular organisms, and/or removal of cell debris (Lim et al., 2017). Essential to phagocytic processes is the acidification of phagosomes through a vesicle trafficking pathway leading to foreign particle destruction (Levin et al., 2016). We have taken advantage of the rhodamine probe pHrodo **<sup>R</sup>** that has minimal fluorescence in a pH neutral milieu, but which displays enhanced (∼10x) fluorescence at decreasing pH (Kapellos et al., 2016). We developed non-toxic conditions to surface label various bacterial species with pHrodo **<sup>R</sup>** , thus enabling a quantifiable mode for the dynamic determination of bacterial internalization. Further, the pHrodo **<sup>R</sup>** -labeled bacteria remain fully viable and continue to serve as functional sources for chemoattraction of Dicytostelium (Meena and Kimmel, 2017).

**Figure 4** shows a time-dependent increase in fluorescence upon phagocytosis of pHrodo **<sup>R</sup>** -labeled gram positive and gram negative bacteria by Dictyostelium and their transport into acidified intracellular vesicles; in the absence of Dictyostelium, fluorescence remains comparatively low throughout the timecourse. The functional difference in fluorescent intensity is well visualized by confocal microscopy (**Figure 4**). Extracellular bacteria have very low level fluorescence, whereas, following phagocytosis, intracellular vesicles of Dictyostelium become progressively, highly, and specifically fluorescent (Meena and Kimmel, 2017).

#### 3.4. Macropinocytosis of pHrodo®Red Dextran by Growth Phase Dictyostelium 3.4.1. Macropinocytosis of pHrodo®Red Dextran

## Using Immunofluorescence Microscopy


#### 3.4.2. Quantification of Macropinocytosis of pHrodo-Red Dextran


intervals. Export the data into Microsoft Excel for processing.

#### 3.4.3. Results

Certain cells of the immune system (e.g., macrophage and dendritic) utilize macropinocytosis in a central role for antigen uptake, processing, and presentation during pathogen defense (Lim and Gleeson, 2011; Marques et al., 2017). In Dictyostelium, macropinocytosis plays a role for nutrient uptake (Bloomfield et al., 2015). Macropinocytosis, however, is a distinct pathway to phagocytosis (Lim and Gleeson, 2011; Levin et al., 2016; Marques et al., 2017); the former functions for cellular uptake of fluid-phase elements, whereas the latter is focused to large particle recognition. Still, the macropinosome and phagosome systems have certain commonality with the endosomal/lysosomal pathway and vesicle acidification (Lim and Gleeson, 2011; Levin et al., 2016; Marques et al., 2017). Therefore, we utilized the pH-sensitive probe pHrodo **<sup>R</sup>** dextran 10K that has minimal fluorescence at neutral pH, but displays dramatic enhancement of fluorescence as the pH decreases during macropinocytosis, with endosome maturation, and fusion with lysosomes.

The principle of the macropinocytosis assay perfectly parallels that of phagocytosis and is, thus, readily observed by confocal fluorescence microscopy (**Figure 5A**) or by high-throughput measurement, where, following a lag, there is a time-dependent increase in fluorescence upon macropinocytosis of pHrodo <sup>R</sup> labeled dextran by Dictyostelium and transport into acidified intracellular vesicles is highly quantifiable (**Figure 5B**). In the absence of Dictyostelium, fluorescence of pHrodo <sup>R</sup> -labeled dextran remains unchanged.

The observed kinetics do differ from that generally observed using FITC-dextran and shaking assays, where uptake is linear for ∼120 min, which then plateaus. Several physical parameter distinctions underlie these kinetic differences. With FITC-dextran, intracellular fluorescence can be detected upon immediate uptake, whereas pHrodo <sup>R</sup> -labeled dextran must first traffic through acidic vesicles to become fluorescent. In suspension assays, Dictyostelium are exposed to continuous and non-varying concentrations of FITC-dextran, from 0 time. In the IncuCyte system, Dictyostelium are first adhered at 0 time, before pHrodo <sup>R</sup> -labeled dextran is added; chemical diffusion and cell movement can lead to a further time-lag for maximal dextran-cell interactions. Collectively, these factors contribute to non-linear (lag phase) curves during initial stages, and relative kinetic delays throughout the time-course. Following the lag, we see linearity for ∼120 min, which begins to plateau for the lower pHrodo <sup>R</sup> -labeled dextran input ratios.

### 4. DISCUSSION

We have applied a high-throughput platform for quantified time-course assays of chemotaxis, phagocytosis, and macropinocytosis. Although we developed these for Dictyostelium, they are easily adaptable to phagocytic cells of any species. The 96-well format permits pharmaceutical or small molecule evaluations for each, with sufficient scale for multiple controls and replicates. As each process is unique,

their combined and distinct set of quantitative assay parameters must be considered when comparing to other approaches. Thus, quantifications rely on collective cellular response, and cannot simply separate the contribution of each individual component. While we have had success with continuous assay in excess of 6 h, we generally limit data collection to <2 h, where physiological differences between cells at their start and end points are minimized. We have confirmed utility for genetic screens in network pathway testing (Meena and Kimmel, 2017) and, thus, identified specific receptors,

G proteins, etc. as essential or non-essential components in these pathways.

Although the chemotaxis data provide direct quality proof for high sensitivity at <1 nM chemoattractant concentrations, the data have significant complexity to interpretation, which differ for purified ligands or live bacteria as the chemoattractant source. Whereas the system does not provide direct quantification of migratory speed, directionality, or cell-shape change with comparison of WT and mutant cells as do single-cell, timecourse chemoattractant response assays, it presents an effective complementary alterative that is highly quantifiable, in a highthroughput manner. It is, for example, highly effective to analyze response to live bacteria, receptor affinity response, essential pathway genes, novel chemoattractants, and small molecule targets, where multiple independent replicates and controls require precise temporal parallel evaluations.

After careful evaluation of diverse experiments and algorithms involving multiple replicates and controls, we conclude that precise normalizations provide the most consistent and reproducible data for interpretation, in comparison to scoring absolute numbers of migratory cells. Values for relative chemotaxis reach a plateau with time, independent of chemoattractant concentration, or bacterial numbers used [see **Figures 1**, **3** (Meena and Kimmel, 2017)]. Nonetheless, this is not simply because there are no migratory cells remaining in the upper well chamber at the time of the plateau. Several other factors are contributory. First, for chemotaxis, we image the number of cells associated with the bottom of the transmembrane at a given time point. However, this association is transitory, and eventually cells will drop off the membrane into the bottom well and be lost to the chemotaxis cell count. Even as new cells continuously reach the membrane, they are in balance with the cell numbers that have become dissociated, thus contributing to a chemotactic plateau appearance. Second, Dictyostelium continuously secrete enzymes that inactivate chemoattractants folate or cAMP; a deaminase (Pan and Wurster, 1978) oxidizes the essential amino group at position 2 of the pyrimidine ring (see **Figure 1**), while a phosphodiesterase converts cAMP into 5 ′ -AMP (Franke and Kessin, 1992; Brzostowski and Kimmel, 2006). As these inactivating enzymes accumulate, they can create a buffer area around migrating cells that reduces chemoattractant gradient strength stimulation of naïve cells, and, thus, reduce their chemotactic response (Tweedy et al., 2016; Meena and Kimmel, 2017).

The chemoattractant gradients established are neither linear nor static, with variations occurring in both signal strength and gradient steepness. Initially, for ligand chemotaxis, concentrations detected by cells in the upper well chamber are relatively low, but with a very steep chemoattractant gradient. However, with time, as the ligand concentration in the upper well rises progressively, the steepness of the gradient diminishes in parallel. For bacteria as chemoattractant sources, it is even more complex. The available levels of chemoattractant is ever increasing, with their continued secretion by bacteria and accumulation in the wells of the lower chamber altering the shape of the gradient.

Apart from dissecting mechanistic pathways for chemotaxis, we are also interested to identify the chemical composition of bacterially secreted chemoattractants. We have applied several approaches to this. The mixing of saturating levels of any specific chemoattractant with cells plated in the upper well chamber will desensitize cells for chemotactic response to that same chemoattractant, but will not compete with unrelated chemoattractants in the bottom chamber wells (Meena and Kimmel, 2017). Alternatively, we have used Dictyostelium lacking components of known pathways to determine the presence or absence of known chemoattractants. Finally, we applied physiochemical assays of bacterially secreted compounds for potential to identify novel molecules. Other factors need to be considered for cellular migration to bacteria that secrete more than a single type chemoattractant. Although folate and cAMP bind and activate different G protein coupled receptors, their downstream intracellular paths converge (Liao et al., 2013; Meena and Kimmel, 2017). Thus, although low sub-saturating levels (e.g., 1 nM) of folate and cAMP elicit greater chemotactic response when provided in combination than with either alone (Meena and Kimmel, 2017), no differences are seen between a single or mixed ligand preparation when either or both are at signal saturation (e.g., 1µM). Accordingly, if a bacterial population can secrete multiple chemoattractants to saturating levels overall cell migration might remain unaltered, even if signal response were fully suppressed for one ligand.

Chemotaxis orients cell movement toward component targets for directed removal and destruction. Since phagocytosis serves as a primary process for this latter cellular function, we developed an engulfment assay of live bacteria that would precisely parallel chemotaxis and allow co-analyses to define common and unique pathways. We labeled bacteria with the rhodamine based probe pHrodo <sup>R</sup> that exhibits enhanced fluorescence with decreasing pH, and quantified phagocytic rates as a function of an increase in fluorescence, upon the internalization of pHrodo <sup>R</sup> -bacteria and trafficking into acidified intracellular vesicles. The phagocytosis assay is highly quantitative in real-time, and the 96-well format facilitates the dissection of molecular pathways, and evaluation of sensitivity to pharmaceutical agents.

Our labeling conditions do not compromise bacterial viability and pHrodo <sup>R</sup> -bacteria remain strongly chemoattractant for Dictyostelium (Meena and Kimmel, 2017). Moreover, we have demonstrated that Dictyostelium lacking chemosensitivity to pHrodo <sup>R</sup> -bacteria, through the loss of chemoattractant receptor function, retain full phagocytic capacity, thus, separating essential dependency of these cellular processes (Meena and Kimmel, 2017).

We have further demonstrated cell-viable labeling and quantitative phagocytosis of multiple gram negative and gram positive bacterial species and suggest that these procedures are adaptable to virtually any bacteria, providing an excellent extension to existing tools for dynamic studies of host-pathogen interactions during chemotaxis and phagocytosis.

Macropinocytosis is another highly regulated pathway for endocytosis, which, like phagocytosis, is essential in pathogen defense (Lim and Gleeson, 2011; Marques et al., 2017). Although both phagocytosis and macropinocytosis are actindependent pathways, they separately involve many distinct molecular elements (Lim and Gleeson, 2011; Bloomfield et al., 2015; Levin et al., 2016; Lim et al., 2017; Marques et al., 2017). Macropinocytosis is well-coordinated to signaling response for cytoskeletal remodeling for fluid phase uptake and, although many components and regulators are known, their precise sequence in event activations is still poorly defined. Since most macropinocytosis assays have limited real-time and/or high-throughput outputs, we consequently used pHrodo <sup>R</sup> -dextran as a fluorescent read-out, as per phagocytosis. Our results have demonstrated defined conditions for highly quantitative measurements of macropinocytosis in Dictyostelium in a physiological, but high-throughput, setting.

We anticipate that the multiple assays described here will have significant utility in application for functional host-pathogen studies.

### 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 by the Intramural Research Program of the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health.

## ACKNOWLEDGMENTS

We thank our colleagues who have contributed to defining the excellence of Dictyostelium as a model for the study of chemotaxis, phagocytosis, and macropinocytosis and, especially, dictyBase (http://dictybase.org/).

## 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 Meena and Kimmel. 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.

# Exploring Virulence Determinants of Filamentous Fungal Pathogens through Interactions with Soil Amoebae

#### Silvia Novohradská1, 2, Iuliia Ferling1, 2 and Falk Hillmann<sup>1</sup> \*

<sup>1</sup> Evolution of Microbial Interactions, Leibniz Institute for Natural Product Research and Infection Biology-Hans Knöll Institute, Jena, Germany, <sup>2</sup> Institute of Microbiology, Friedrich Schiller University Jena, Jena, Germany

Infections with filamentous fungi are common to all animals, but attention is rising especially due to the increasing incidence and high mortality rates observed in immunocompromised human individuals. Here, Aspergillus fumigatus and other members of its genus are the leading causative agents. Attributes like their saprophytic life-style in various ecological niches coupled with nutritional flexibility and a broad host range have fostered the hypothesis that environmental predators could have been the actual target for some of their virulence determinants. In this mini review, we have merged the recent findings focused on the potential dual-use of fungal defense strategies against innate immune cells and soil amoebae as natural phagocytes. Well-established virulence attributes like the melanized surface of fungal conidia or their capacity to produce toxic secondary metabolites have also been found to be protective against the model amoeba Dictyostelium discoideum. Some of the recent advances during interaction studies with human cells have further promoted the adaptation of other amoeba infection models, including the wide-spread generalist Acanthamoeba castellanii, or less prominent representatives like Vermamoeba vermiformis. We further highlight prospects and limits of these natural phagocyte models with regard to the infection biology of filamentous fungi and in comparison to the phagocytes of the innate immune system.

#### *Edited by:*

Sascha Thewes, Freie Universität Berlin, Germany

#### *Reviewed by:*

Sven Krappmann, University of Erlangen-Nuremberg, Germany Marie-Helene Rodier, CHU/Faculté de Médecine, France

#### *\*Correspondence:*

Falk Hillmann falk.hillmann@leibniz-hki.de

*Received:* 14 September 2017 *Accepted:* 20 November 2017 *Published:* 05 December 2017

#### *Citation:*

Novohradská S, Ferling I and Hillmann F (2017) Exploring Virulence Determinants of Filamentous Fungal Pathogens through Interactions with Soil Amoebae. Front. Cell. Infect. Microbiol. 7:497. doi: 10.3389/fcimb.2017.00497 Keywords: dictyostelium, acanthamoeba, aspergillus, phagocytosis, macrophages, amoebae

## ENVIRONMENTALLY ACQUIRED FUNGAL PATHOGENS

Fungi are ubiquitous in nature, inhabiting various ecological niches. Even among those which thrive as saprophytes and do not exhibit any host requirement for survival, there are pathogens which cause devastating diseases in humans and animals resulting in thousands of deaths every year (Brown et al., 2012). Classical examples include filamentous fungi like Aspergillus fumigatus and Fusarium sp., but also several dimorphic fungi such as Blastomyces dermatitidis or Histoplasma capsulatum, and the yeast Cryptococcus neoformans, have environmental reservoirs. One of the most prevalent groups of fungi in the environment is represented by the aspergilli (Shelton et al., 2002). With several hundred species, only a few of them have a considerable impact on human health: A. fumigatus, A. flavus, A. terreus, A. nidulans, and A. niger.

Aspergillus fumigatus is one of the most important air-borne fungal pathogens, living ubiquitously in terrestrial environments. This fungus disseminates by releasing thousands of asexual spores (conidia) from each conidiophore which, upon inhalation, pass through the nasal cavity and reach the alveoli. Most of them are expelled by mucocilliary clearance while the residual ones are eliminated by macrophages and neutrophils of an immunocompetent host. In the case of an immunodeficient host, conidia swell and grow into a mycelium. Once the fungus overcomes the natural immune barrier, it can cause asthma-associated allergies, sinusitis, allergic bronchopulmonary aspergillosis (ABPA) and, in the worst case, life-threatening invasive aspergillosis (IA), occasionally reaching mortality rates even beyond 50% due to rapid progression and misdiagnosis (Brown et al., 2012). Aspergilli can also infect wild and domestic animals nearly encompassing all major phyla including corals, honey bees, reptiles, and warm-blooded animals such as birds, mammals, and non-human primates (reviewed by Seyedmousavi et al., 2015). Another group of environmental filamentous fungi represent exclusively entomogenous pathogens, such as Metarhizium anisopliae or Beauveria bassiana. Microscopic examination and phagocytosis assays with these fungi suggested that their ability to adhere to the insect cuticle, penetrate through the haemocoel using hydrolyzing enzymes, and ultimately survive phagocytic haemocytes may be a consequence of adaptations that have been acquired early in evolution to avoid predation by soil amoebae (Bidochka et al., 2010).

Unlike Aspergillus, C. neoformans is not ubiquitous in the soil; rather it has been isolated from areas frequented by pigeons, chickens, turkeys, and other avian species. After inhalation of infectious particles, Cryptococcus resides in the lung alveoli where it can persist and replicate while a thick polysaccharide capsule surrounding the yeast cell helps to avoid its killing by macrophages. Other studies have shown that Cryptococcus is also able to survive intracellularly, even a few hour after phagocytosis (Feldmesser et al., 2000). Dissemination to the brain results in severe meningoencephalitis, especially in immunocompromised patients. On the other hand, C. gatii which has been isolated from trees, mainly causes pulmonary infections in an immunocompetent host (García-Rodas and Zaragoza, 2012; Kwon-Chung et al., 2014).

Among other environmentally acquired pathogenic fungi, thermally dimorphic fungi such as Histoplasma capsulatum, Blastomyces dermatitidis, and Coccidioides immitis are especially clinically relevant and classified as Biosafety Level 3 (BSL3) organisms. Despite their divergent phylogeny, they all share similar patterns of existence: temperature-dependent morphological dimorphism, pulmonary infectivity, and endemism. After the inhalation of conidia, transformation into a yeast-form is crucial to promote pathogenicity through escape from phagocytosis, modulation of the cytotoxic environment of the phagolysosome or enhanced degradation of reactive oxygen species (Boyce and Andrianopoulos, 2015).

Considering the diversity of environmental niches and strategies to survive and replicate within a variety of mammalian hosts, the aforementioned virulence attributes may confer a dual-use capability to defend against phagocytes in both animal hosts and the environment. Moreover, such parallels gave rise to the idea that selective pressures in the environment have led to the emergence and maintenance of these traits that have later supported virulence in higher eukaryotes. Although the filamentous life style of aspergilli suggests little need for any specific attributes to avoid or withstand any phagocytes, their infectious and reproductive stage is formed by small, unicellular conidia which are easily ingested by such cells. A number of profound studies over the last years have uncovered a multitude of mechanisms which aid in the escape and defense of these fungi against their opponents of human or environmental origin (**Table 1**).

### FUNGAL RESISTANCE TO INNATE IMMUNE CELLS AND SOIL AMOEBAE

Macrophages and neutrophil granulocytes are the most prominent representatives among innate immune cells which counteract fungal pathogens. Alveolar macrophages are the first line of defense against IA by killing inhaled conidia and initiating the pro-inflammatory response that recruits neutrophils to the site of infection (reviewed by Brakhage et al., 2010). As a consequence, patients with reduced numbers of macrophages have long been known to be at a higher risk to develop IA (Brakhage, 2005). A number of in-vitro studies suggest that even macrophages from immunocompetent individuals show variable killing efficiencies of ingested fungal conidia ranging from 10 to 90% as summarized in Philippe et al. (2003). The same study demonstrated that fungal conidia are especially resistant to killing when remaining in a dormant state or when the confronting macrophages are derived from immunosuppressed donors. Overall, most of these results are based on counts of colony forming units (CFUs) and certain methodological issues, such as the time of co-incubation, conidial aggregation or the incomplete experimental removal of non-ingested conidia, could have also contributed to the heterogeneous killing rates that have been reported. A similar assay with Dictyostelium discoideum revealed that conidia were readily ingested but remained viable over more than 24 h based on the number of CFUs (Hillmann et al., 2015). Even the comparably robust pathogen Acanthamoeba castellanii ingested conidia within the first hour of their interaction, but no signs of digestion were observed at any stage (Van Waeyenberghe et al., 2013). Along this line, co-incubation of A. fumigatus or Fusarium oxysporum conidia with the common water contaminant Vermamoeba vermiformis did not result in any reduction of viable conidia, but instead phagocytic uptake promoted filamentation and growth of the fungus (Cateau et al., 2014; Maisonneuve et al., 2016).

#### RECOGNITION AND PROCESSING OF FUNGAL CONIDIA BY HUMAN AND ENVIRONMENTAL PHAGOCYTES

Given their role as a first line of defense, studies on the recognition and phagocytic processing of fungal conidia by macrophages have a long-standing history, mainly with regard to aspergilli. The surface of their conidia represents the immediate interface and pathogen-associated molecular patterns (PAMPs) include cell wall constituents like α- and β-glucans, chitins, galactomannans, and other polysaccharides. However, these are TABLE 1 | Fungal virulence determinants studied using amoeba model systems.


usually masked by a proteinaceous, hydrophobic rodlet layer which is immunologically inert and diminishes the recognition by immune cells (Aimanianda et al., 2009). The green-gray dihydroxynapthalene (DHN)-melanin pigment coating dormant conidia is another surface component assumed to play a similar role in A. fumigatus (Jahn et al., 1997; Tsai et al., 1998; Chai et al., 2010). These protective layers are lost during swelling and subsequent germination of the conidia (**Figure 1**), exposing the PAMPs and allowing recognition, as demonstrated for strains of A. fumigatus lacking the DHN melanin pigment which were ingested by macrophages at higher rates than wild type strains (Luther et al., 2007). Interestingly, nearly identical ratios were observed with D. discoideum. Conidia of the wild-type, covered by green DHN-melanin, were taken up, but at least threefold less efficiently than the white conidia of the melanin deficient mutant (Hillmann et al., 2015). These data suggest that hiding "prey"-associated molecular patterns could be an asset to escape also from environmental predators and hence, well suited to be studied in an amoeba model.

Swollen conidia of A. fumigatus that have lost their DHNmelanin cover expose cell wall β-glucans which activate the Dectin-1/Syk kinase/NADPH signaling cascade in macrophages (Luther et al., 2007; Ma et al., 2012). After recognition and internalization, phagolysosome maturation is initiated through assembly of the NADPH oxidase and LC3-associated phagocytosis (LAP) (Akoumianaki et al., 2016). Melanized conidia of A. fumigatus are able to inhibit the crucial process of phagolysosomal maturation at the acidification step which contributes to their significantly reduced killing rates relative to melanin deficient conidia (Jahn et al., 2002; Thywißen et al., 2011, **Figure 1**). In sharp contrast, melanized conidia were shown to inhibit the apoptosis of macrophages by activating the PI3 kinase/Akt signaling pathway (Volling et al., 2011).

The chemically distinct Asp-melanin of Aspergillus terreus does not inhibit phagolysosomal acidification in macrophages or D. discoideum (**Figure 1**), indicating that even closely related fungi, such as A. fumigatus and A. terreus, apply different persistence and propagation strategies inside the harsh phagolysosomes of macrophages and some environmental phagocytes (Slesiona et al., 2012; Geib et al., 2016).

#### PHAGOCYTE-ESCAPE MECHANISMS OF *A. FUMIGATUS*

Apart from conidial killing, macrophage encounters with A. fumigatus may also result in disruption of the host cell by the elongation of the hyphae (**Figure 1**). Here, Aspergillus fumigatus conidia swell up and initiate growth despite the nutrient-scarce environment of the phagolysosome thereby exploiting their glyoxylate cycle and siderophore machinery to overcome the limited supply of carbon and iron, respectively (Behnsen et al., 2007; Schrettl et al., 2010). It seems likely that these pathways could also be active during conidial processing in environmental phagocytes as this escape strategy was well documented during the interaction with A. castellanii as well as D. discoideum, resulting in host cell lysis for both amoebae (Van Waeyenberghe et al., 2013; Hillmann et al., 2015). Some of the consequences

of intracellular fungal germination on the host cell were only recently resolved for human alveolar macrophages. Orchestrated by calcineurin, the infected host cell can complete exocytosis of the fungus containing endosome, followed by the lateral transfer to neighboring cells as part of programmed necrosis (Shah et al., 2016). In perspective of the well-studied social behavior of D. discoideum, this amoeba model seems to be especially promising for future studies on such intercellular or altruistic processes. It has become obvious that filamentous fungi can be tough opponents of innate immune cells, but questions on the origin of such anti-immune properties of filamentous fungi remain open. Amoeba models could help to close the knowledge gaps which cannot be addressed from in vitro studies with human cells.

### AMOEBAE PREDATION AS AN EVOLUTIONARY TRAINING GROUND FOR FILAMENTOUS FUNGAL PATHOGENS

Free living amoebae (FLA) are ubiquitous unicellular protozoa, distributed worldwide in various environments such as soil, water, or air. Mycophagous species, either opportunists or specialists, are widespread and have been isolated over the past 60 years (Olive and Stoianovitch, 1960; Old and Darbyshire, 1978; Chakraborty et al., 1983; Old et al., 1985; Chakraborty and Old, 1986). One of the first mycophagous amoebae ever described was Protostelium mycophagum (Olive and Stoianovitch, 1960). It has been isolated alongside the pink-pigmented yeast Rhodotorula mucilaginosa, but it can also feed on Phoma sp., Ustilago violacea, Sporobolomyces sp., and Cryptococcus laurentii (Spiegel et al., 2006). Other mycophagous soil amoebae from the genera Thecamoeba, Arachnula, and Vampyrella have been shown to suppress the growth of the plant pathogenic fungus Gaeumannomyces graminis var. tritici and therefore significantly contribute to the reduction of "take-all" wheat crops disease caused by this fungus (Chakraborty et al., 1983).

In environments where fungi have encountered constant protozoal predation and competition for nutrients, they must have developed strategies to counteract phagocytic uptake or intracellular passage. Consequently, the same determinants that were effective against amoeba predation, could have later promoted the survival of fungi in a human host. The coincidental evolution hypothesis suggests that virulence factors have evolved as a response to more ordinary selection pressures than for virulence per se (Erken et al., 2013). Several studies using freeliving amoeba A. castellanii and more recently D. discoideum have supported this hypothesis for yeast-like fungi and have underlined the suitability of these models to study basic fungal virulence determinants (Steenbergen et al., 2001, 2004; Chrisman et al., 2010; Derengowski et al., 2013; Koller et al., 2016; Rizzo et al., 2017, see **Table 1** for an overview).

Aspergillus fumigatus requires no obvious residence inside the mammalian host to survive and replicate; it appears to lack classical virulence factors and its pathogenicity primarily depends on host impairment. Therefore, it is conceivable that the ability to counteract the compromised immune system was partially tuned from the long-term interplay between fungi and their predators in their natural environment. Such fungivorous organisms are by no means limited to amoebae, but also include higher animals like nematodes, mites, or insects. For instance, a primary study on A. nidulans demonstrated that the multitude of fungal lowmolecular-mass compounds known as secondary metabolites could present a selective advantage against predation by the fungivorous springtail Folsomia candida (Rohlfs et al., 2007).

Such defensive actions could also provide protection against microscopic predators, as active components have been shown to diffuse from the non-germinating spores and inhibit certain functions of phagocytosis (Slight et al., 1996). The anti-amoebae effects of diffusates from clinical and environmental isolates of A. fumigatus and A. terreus have been described on Naegleria gruberi, proposing it as a primary function of such metabolites (Hobson, 2000). Even at the early stages of a direct interaction, mycotoxins are encountered immediately by the ingesting phagocytes, as it has recently been shown for the amoebicidal polyketide trypacidin which resides primarily on the surface of the A. fumigatus spore (Gauthier et al., 2012; Mattern et al., 2015, **Figure 1**). Following germination and escape, further potent, soluble toxins are synthesized.

Among them, the sesquiterpene fumagillin was one of the first for which amoebacidal properties were observed and has initially been used for the treatment of infections caused by Entamoeba histolytica (Killough et al., 1952). Fumagillin (**Figure 1**) and its synthetic analogs thereby irreversibly inhibit the methionine aminopeptidase-2 (MetAP2), making them promising therapeutic candidates against malaria parasites, trypanosomes, or other amoebae (Arico-Muendel et al., 2009). When using D. discoideum as a model, however, cytotoxic effects on the phagocytes could largely be attributed to the nonribosomal peptide gliotoxin (Hillmann et al., 2015). The toxic and immunosuppressive properties of gliotoxin, the prototype of the epidithiodioxopiperazine (ETP)-type mycotoxins, are directed toward the host's immune effector cells via the activity of its unusual intramolecular disulfide bridge (**Figure 1**). Several target molecules for gliotoxin have been well described, including the NADPH oxidase of polymorphonuclear leukocytes or central regulatory hubs like the phosphatidylinositol 3,4,5-trisphosphate metabolism and the transcription factor NFκB (Pahl et al., 1996; Tsunawaki et al., 2004; Schlam et al., 2016). Among these studies, Schlam and colleagues have shown that gliotoxin further prevents integrin activation in immortalized and primary macrophages and interferes with actin dynamics. As both of these are essential instruments during phagocytosis and membrane ruffling, such pathways may be attractive targets in the defense against FLA. Previously it was thought that gliotoxin production is restricted only to clinical isolates of A. fumigatus; however, it was demonstrated recently that the vast majority (>96%) of both environmental and clinical isolates of aspergilli are able to produce this mycotoxin (Kupfahl et al., 2008; Scharf et al., 2012). Consequently, it is only plausible to suspect that fungi have maintained their whole repertoire of active secondary metabolites to counteract not only their numerous competitors, but also predators in their natural environment whose numbers and diversity have long been underestimated. A recent study supports this conclusion by demonstrating that mycophagous protists are abundant, taxonomically widespread, and central ecological players in the soil food web (Geisen et al., 2016).

#### PERSPECTIVES

Both Dictyostelium discoideum and Acanthamoeba castellanii have been extensively studied as model organisms in terms of phagocytic interactions, mainly due to the similarity with human macrophages (Tosetti et al., 2014). As unicellular eukaryotes, with a compartmentalized cytoplasm, relatively small size, and active chemotactic movement they exemplify an ideal non-mammalian model for host-pathogen interactions and microbial infections. Both amoebae are easy to cultivate (either on bacteria or axenically), giving high cell yields with defined identity. Besides, working with these model organisms is highly advantageous in terms of genetic malleability combined with a profound knowledge of its phagocytic pathways (Eichinger et al., 2005; Siddiqui and Khan, 2012). A major advantage of D. discoideum over A. castellanii is the well annotated genome, the high number of molecular tools and protocols, and the wide availability of targeted mutants that are easily accessible at www.dictybase. org (Eichinger et al., 2005; Bozzaro and Eichinger, 2011; Fey et al., 2013). However, the fact that prolonged subcultivations and axenization of native amoeba cultures may have led to the accumulation of undesired mutations and loss of original phagocytic abilities has to be taken into consideration. Another drawback of D. discoideum is its limited maximal survival temperature of roughly 27◦C, which may be unfavorable for some fungi that may express their full virulence potential only at temperatures around 37◦C. However, at present there is little support for the idea that filamentous fungi like A. fumigatus tend to regulate their general virulence attributes strictly in response to higher temperatures. In fact, a recent study revealed that 11 of 37 clusters encoding for the biosynthesis of secondary metabolites were activated by a temperature shift from 37 to 30◦C, including those coding for DHN-melanin, gliotoxin, and trypacidin (Lind et al., 2016). Nevertheless, when using amoebae as a tool to study the origin of fungal virulence, it is important to keep in mind that throughout their evolution, fungi have also encountered many other soil-dwelling predators that may have contributed to the emergence of fungal virulence determinants. Certainly not all virulence traits are equally beneficial to protect fungi in different hosts, giving rise to a number of different invertebrate animal models which have been used to study fungal pathogenesis (Desalermos et al., 2012). Very little is known about the interaction of soil amoebae and fungi in their natural environment. It is plausible, that not every amoeba is amenable to infection by every fungal pathogen. Furthermore, bacteria are the preferential prey for D. discoideum, while A. castellanii was originally discovered as a contaminant in yeast cultures (Castellani, 1930). Even if fungal infection biologists have successfully exploited these two established amoeba models, further studies on the evolution of fungal virulence may call for an inclusion of natural fungivorous predators from the Amoebozoa phylum.

### AUTHOR CONTRIBUTIONS

SN, IF, and FH wrote the initial version of the manuscript. IF and SN prepared illustrations. FH revised the manuscript. All authors read and approved the final version.

#### REFERENCES


#### ACKNOWLEDGMENTS

Research in the authors lab is supported by a fellowship of the DFG funded excellence graduate school "Jena School of Microbial Communication"—JSMC (to SN), a grant of the European Social Fund ESF "Europe for Thuringia" (2015 FGR 0097) and DFG grant HI 1574/2-1.


the interaction between phagocytes and yeasts. Front. Microbiol. 7:1665. doi: 10.3389/fmicb.2016.01665


**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 © 2017 Novohradská, Ferling and Hillmann. 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) or licensor 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.

# *Cryptococcus neoformans* Escape From *Dictyostelium* Amoeba by Both WASH-Mediated Constitutive Exocytosis and Vomocytosis

Rhys A. Watkins 1,2†, Alexandre Andrews 1,2†, Charlotte Wynn1,2, Caroline Barisch<sup>3</sup> , Jason S. King1,4 \* and Simon A. Johnston1,2 \*

<sup>1</sup> Bateson Centre, University of Sheffield, Sheffield, United Kingdom, <sup>2</sup> Department of Infection Immunity and Cardiovascular Disease, Medical School, University of Sheffield, Sheffield, United Kingdom, <sup>3</sup> Department of Biochemistry, Faculty of Science, University of Geneva, Geneva, Switzerland, <sup>4</sup> Department of Biomedical Sciences, University of Sheffield, Sheffield, United Kingdom

#### *Edited by:*

Sascha Thewes, Freie Universität Berlin, Germany

#### *Reviewed by:*

Charley Christian Staats, Federal University of Rio Grande do Sul (UFRGS), Brazil Sascha Brunke, Leibniz-Institut für Naturstoff-Forschung und Infektionsbiologie and Hans Knöll Institut, Germany

#### *\*Correspondence:*

Jason S. King jason.king@sheffield.ac.uk Simon A. Johnston s.a.johnston@sheffield.ac.uk

†These authors have contributed equally to this work.

*Received:* 24 November 2017 *Accepted:* 19 March 2018 *Published:* 09 April 2018

#### *Citation:*

Watkins RA, Andrews A, Wynn C, Barisch C, King JS and Johnston SA (2018) Cryptococcus neoformans Escape From Dictyostelium Amoeba by Both WASH-Mediated Constitutive Exocytosis and Vomocytosis. Front. Cell. Infect. Microbiol. 8:108. doi: 10.3389/fcimb.2018.00108 Cryptococcus neoformans is an environmental yeast that can cause opportunistic infections in humans. As infecting animals does not form part of its normal life-cycle, it has been proposed that the virulence traits that allow cryptococci to resist immune cells were selected through interactions with environmental phagocytes such as amoebae. Here, we investigate the interactions between C. neoformans and the social amoeba Dictyostelium discoideum. We show that like macrophages, D. discoideum is unable to kill C. neoformans upon phagocytosis. Despite this, we find that the yeast pass through the amoebae with an apparently normal phagocytic transit and are released alive by constitutive exocytosis after ∼80 min. This is the canonical pathway in amoebae, used to dispose of indigestible material after nutrient extraction. Surprisingly however, we show that upon either genetic or pharmacological blockage of constitutive exocytosis, C. neoformans still escape from D. discoideum by a secondary mechanism. We demonstrate that constitutive exocytosis-independent egress is stochastic and actin-independent. This strongly resembles the non-lytic release of cryptococci by vomocytosis from macrophages, which do not perform constitutive exocytosis and normally retain phagocytosed material. Our data indicate that vomocytosis is functionally redundant for escape from amoebae, which thus may not be the primary driver for its evolutionary selection. Nonetheless, we show that vomocytosis of C. neoformans is mechanistically conserved in hosts ranging from amoebae to man, providing new avenues to understand this poorly-understood but important virulence mechanism.

Keywords: cryptococcus, *Dictyostelium*, amoeba, pathogen, exocytosis, cryptococcosis, vomocytosis, WASH

#### INTRODUCTION

Cryptococcus neoformans is a basidiomycete yeast found globally in a wide variety of natural environments. Unusually for an environmental yeast, C. neoformans is also a pathogen of animals. Most significant is the fatal infection of the severely immunocompromised, with cryptococcal meningitis caused by C. neoformans responsible for 15% of AIDS-related deaths (Rajasingham et al., 2017). The interaction of C. neoformans with its host is highly complex, and what differentiates normal immunity from the development of life threatening cryptococcal meningitis is well defined (Tenforde et al., 2017).

Macrophages have been repeatedly demonstrated to be critical for protection against C. neoformans infection. However, macrophages may also have a role in pathogenesis in the immunocompromised as cryptococci are able to grow and survive within macrophages and may use macrophages as a Trojan horse to disseminate from the lung. In vitro, almost every aspect of macrophage antimicrobial activity is either avoided or manipulated by C. neoformans, which are able to survive and replicate intracellularly, following uptake by phagocytosis (Johnston and May, 2013; Ballou and Johnston, 2017). Survival traits include the generation of a characteristic polysaccharide capsule, which is both anti-phagocytic and helps protect the yeast from the host antimicrobial machinery if it is engulfed, as well as melanin production which serves as a potent antioxidant—protecting the yeast from the phagosomal oxidative attack and the immunomodulatory activity of cell wall chitin (Casadevall et al., 2000; Nosanchuk and Casadevall, 2006; Wiesner et al., 2015).

An additional pathogenic mechanism is the remarkable ability of C. neoformans to promote its non-lytic expulsion from host cells in a process known as vomocytosis (Alvarez and Casadevall, 2006; Ma et al., 2006). This enables the yeast to escape whilst leaving the host phagocyte intact, thus preventing immune stimulation and promoting dissemination. Whilst it has been shown that vomocytosis is suppressed by host actin polymerization (Johnston and May, 2010) and can be modulated by host Annexin A2 and Mitogen Activated Protein kinase (ERK5) activity (Stukes et al., 2016; Gilbert et al., 2017) little is known of the underlying molecular mechanisms underlying expulsion. Nonetheless, vomocytosis has been observed in both cell culture and in vivo models and is thought to significantly contribute to C. neoformans virulence (Alvarez and Casadevall, 2006; Ma et al., 2006; Bojarczuk et al., 2016; Johnston et al., 2016; Gilbert et al., 2017).

As with other opportunistic pathogens, it is unlikely that interactions with mammalian macrophages have been the evolutionary drivers of C. neoformans virulence. Cryptococci are free-living fungi with a life cycle that is not dependent on infecting an animal host. It has therefore been proposed that the mechanisms that allow C. neoformans to survive and grow in macrophages have primarily evolved to avoid predation by phagocytes in its natural environment, such as amoebae (Steenbergen et al., 2001; Casadevall, 2012; Watkins et al., 2017).

Like leukocytes, amoebae are professional phagocytes, using their chemotactic and phagocytic abilities to capture and kill environmental microbes for food. Despite the large evolutionary distance between them, much of the machinery and mechanisms for phagocytosis and phagosome maturation are highly conserved between amoebae and mammalian immune cells (Boulais et al., 2010). Traits that have evolved to help yeast and bacteria avoid being killed by amoebae in the environment are therefore likely to have similar effects when they encounter mammalian immune cells.

Previous studies have demonstrated similarities in the interactions between C. neoformans with amoebae and macrophages. C. neoformans is able to both survive phagocytosis and replicate intracellularly within Acanthamoeba castellanii, ultimately being released alive without causing lysis of the host amoeba (Steenbergen et al., 2001; Chrisman et al., 2010). Due to its amenability to genetic manipulation, the social amoeba Dictyostelium discoideum has been used a model host for a number of human pathogens and is also susceptible to C. neoformans infection (Steenbergen et al., 2003). Importantly, passage through D. discoideum caused a stimulation in C. neoformans capsule expansion and melanization together with a corresponding increase in subsequent virulence in mice (Steenbergen et al., 2003). Interactions with amoebae can therefore directly influence interactions between C. neoformans and mammalian immune cells.

The fate of internalized material in animal cells is variable and complex. There are examples of the expulsion of internalized material from a variety of cell types, particularly in the context of antigen presentation (Chen and Jondal, 2004; Peters et al., 2006; Griffiths et al., 2012; Le Roux et al., 2012; Turner et al., 2016). However, animal macrophages (notably tissue resident cells, such as alveolar macrophage) have the ability to retain particulate matter that may otherwise be damaging (Bai et al., 2015). In contrast, the constitutive exocytosis of phagocytosed material by amoeba has been demonstrated in diverse species including D. discoideum, Amoeba proteus, Entamoeba histolytica and A. castellanii (Weisman and Korn, 1967; Ravdin et al., 1988; Christofidou-Solomidou and Stockem, 1992; Clarke et al., 2010) and thus appears to be a necessary and general feature of freeliving amoebae. Therefore macrophages and amoebae differ in their retention of phagocytosed material.

Recently it was shown that constitutive exocytosis in D. discoideum is dependent on the activity of the WASH (WASP And SCAR Homolog) complex (Carnell et al., 2011). WASH is a direct activator of the ARP2/3 (Actin Related Protein 2/3) complex, causing the polymerization of actin on the surface of vesicles and driving membrane protein sorting and recycling (Derivery et al., 2009; Gomez and Billadeau, 2009; Zech et al., 2011; Seaman et al., 2013). Whilst an early phase of WASH activity drives the retrieval of cell surface proteins from phagosomes (Buckley et al., 2016), a second phase of activity occurs after 40–60 min of digestion, driving the removal of the vacuolar (V)-ATPase and phagosomal neutralization. This facilitates hydrolase retrieval (King et al., 2013) and is a prerequisite for exocytosis. Consequently D. discoideum cells lacking WASH have a complete block in constitutive exocytosis (Carnell et al., 2011).

As constitutive exocytosis represents the normal mechanism of non-lytic release of phagosomal contents, we hypothesized that this may be involved in the escape of live C. neoformans from amoebae. We show that C. neoformans survive phagocytosis by D. discoideum, but follow an apparently normal phagosomal transit and are normally released alive by WASH-dependent constitutive exocytosis. However, when constitutive exocytosis is blocked, C. neoformans still escape in a WASH and actinindependent manner reminiscent of vomocytosis. This demonstrates redundant, mechanistically different egress mechanisms with implications for the understanding of the evolutionary drivers of cryptococcal virulence.

#### MATERIALS AND METHODS

#### Strains and Cell Culture

For all experiments the Ax2 axenic strain of D. discoideum was used, both as "wild type" as well as the parental of the previously published WASH-mutant strain (Carnell et al., 2011). D. discoideum were cultured in filter-sterilized HL-5 medium (Formedium, Norfolk, UK) at 22◦C. Cells expressing GFP fused to the vatM subunit of the vacuolar-ATPase were generated using plasmid pMJC25 (Carnell et al., 2011).

Unless otherwise stated, C. neoformans var. grubii (serotype A) strain H99α stably expressing mCherry was used (Gibson et al., in review). The previously published plb1 and cap59 mutants were generated from the alternative wild-type parent H99, which was used as control when appropriate (Chen et al., 2000; Voelz et al., 2010). C. neoformans were grown in YPD medium at 28◦C prior to experiments, but washed and resuspended in HL5 medium before infecting D. discoideum. C. neoformans were heat-killed by incubation at 65◦C for 30 min, before washing in PBS. UV-killing was performed by exposure of C. neoformans cultures to 4 J of UV using a UVIlink CL-5087 cross-linker illuminator. Killing was always confirmed by plating samples on YPD agar to check absence of growth.

#### Infections and Microscopy

Prior to imaging, 2 × 10<sup>6</sup> D. discoideum cells were seeded in 2 ml HL5 medium in glass-bottom 35 mm dishes (Mat-Tek), left to adhere for 30 min. Then 2 × 10<sup>5</sup> C. neoformans cells, 4.5µm green fluorescent YG unmodified beads (Polysciences Inc., Pennsylvania, USA), or TRITC-labeled heat killed S. cerevisiae (kind gift from Thierry Soldati; University of Geneva) were added prior to imaging. This differed for latrunculin treatment experiments: amoebae were mixed with yeast at a ratio of 1:1 and left for 1 h for phagocytosis to occur before addition of 5µm Latrunculin A (Cayman chemical Co.).

Long term time-lapse movies were recorded using a Nikon TI- E with a CFI Plan Apochromat λ 20x N.A.0.75 at 22◦C. Images were captured on a large-format Andor Neo 5.5 s CMOS camera for 12 h, imaging 4 fields of view every 30 s. Transit times were determined as the time from phagocytosis to release. Events where cells could not be tracked for a minimum of 500 min post-phagocytosis were excluded from the analysis.

Spinning disc microscopy was performed on a Perkin-Elmer Ultraview VoX spinning disk confocal microscope running on an Olympus 1 × 81 body with an UplanSApo 60x oil immersion objective (NA 1.4). Images were captured on a Hamamatsu C9100-50 EM-CCD camera.

#### Statistical Analysis

Mann Whitney test was used to test significance between continuous data and Fisher's exact test for categorical data. P-values below 0.05 (with modification for multiplicity of testing) were considered statistically significant. All statistical tests were performed with GraphPad Prism version 7.

#### RESULTS

#### *Cryptococcus neoformans* Is Constitutively Exocytosed by Amoebae

To determine whether C. neoformans utilizes a pathogen-specific mechanism of release or follows the normal phagocytic pathway of engulfment and constitutive exocytosis in amoeba we first compared the transit of C. neoformans with that of inert particles.

Using time lapse microscopy, we followed the phagocytosis and release of C. neoformans compared to heat killed nonpathogenic yeast Saccharomyces cerevisiae and 4.0µm latex beads by D. discoideum. All three cargoes were of a similar size (between 3 and 5µm) and all were exocytosed from amoebae (**Figures 1A–C**). However, unlike the vomocytosis of C. neoformans from macrophages which happens stochastically and inefficiently over a period of many hours (Ma et al., 2006; Johnston and May, 2010), 100% of phagosomes fused with the plasma membrane, exocytosing their contents within 4 h of engulfment in all cases.

Surprisingly, whilst we attempted to use killed C. neoformans as controls, both heat- or UV-killing of the yeast reduced phagocytosis by D. discoideum by 90% (**Figure 1H**). This effect was further compounded in WASH-null amoebae, in which phagocytosis of UV-killed C. neoformans was never observed in 3 independent experiments, consistent with previous reports of phagocytosis defects in these mutants due to reduced surface levels of phagocytic receptors (Buckley et al., 2016). The reason why killed C. neoformans resist phagocytosis is unclear, but it was not possible to observe sufficient events for analysis and therefore heat-killed S. cerevisiae were used as non-pathogenic controls in subsequent experiments.

From the time-lapse movies, we were able to define precise transit times for phagosomes containing the different cargoes, tracking individual particles from engulfment to release (**Figures 1D–F**). Comparison of the transit time demonstrated that both fungal cells took slightly longer to complete the phagocytic cycle than latex beads however, there was no significant difference between heat killed S. cerevisiae and C. neoformans (median transit time latex beads = 61 min, heat killed S. cerevisiae = 82 min and C. neoformans = 84 min, **Figure 1G**). Importantly the C. neoformans were able to resist killing by the amoebae and were exocytosed alive, as indicated by subsequent budding and dividing after egress consistent with previous studies (Steenbergen et al., 2003). Therefore, despite their ability to survive phagocytosis by D. discoideum, transit time of C. neoformans is indistinguishable from normal constitutive exocytosis.

#### The V-ATPase Is Rapidly Recruited to Phagosomes Containing *C. neoformans* and Removed Prior to Exocytosis

Unlike the vomocytosis of cryptococci, a number of molecular requirements for constitutive exocytosis in amoebae have

FIGURE 1 | Cryptococcus neoformans is constitutively exocytosed from Dictyostelium discoideum amoeba. (A–C) Example exocytosis of (A) 4.5 um green fluorescent latex beads (images start at 40 min after phagocytosis) (B) heat killed Saccharomyces cerevisiae (images start at 60 min after phagocytosis) and (C) C. neoformans strain Kn99mCherry from wild type D. discoideum strain Ax2 (images start at 60 min after phagocytosis). Time 0 s indicates point of exocytosis. Scale bars 5µm. (D–F) Frequency histograms of combined transit times measured from three independent 12 h time lapses. (D) Latex beads 126 transit times. (E) Heat killed S. cerevisiae 66 transit times. (F) C. neoformans 57 transit times. (G) Comparison of transit times for latex beads, heat killed S. cerevisiae and C. neoformans. (H) Quantification of phagocytic events for live, heat killed and UV killed cryptococci in Ax2 and WASH null cells. P-values are Mann-Whitney test.

been identified. We therefore next asked whether passage of C. neoformans through D. discoideum followed the normal path of maturation and constitutive exocytosis.

During normal transit, phagosomes rapidly accumulate the V-ATPase and acidify within 2–3 min of engulfment; the V-ATPase is subsequently retained for ∼45 min to allow digestion before retrieval and phagolysosomal neutralization prior to constitutive exocytosis (Clarke et al., 2002, 2010). Using cells expressing GFP-fused to the VatM subunit of the V-ATPase, we monitored recruitment to phagosomes following engulfment of cryptococci. V-ATPase was present, on average, within 114 ± 44 s (**Figure 2A**, S.D., n = 6) of phagocytosis consistent with published data for inert phagosomes (Clarke et al., 2002; Buckley et al., 2016). To test if V-ATPase recruitment was maintained, we also measured the proportion of C. neoformans-containing phagosomes positive for GFP-VatM after 20 min incubation of cryptococci with amoebae—before any post-lysosomal transitions should have occurred (**Figure 1F**). V-ATPase was clearly visible on 90.0% of phagosomes (68/76 from three independent experiments). However, when we looked after 1 h, we were able to observe recycling of the V-ATPase from live (budding) C. neoformanscontaining phagosomes prior to exocytosis (**Figure 2B**). Both phagosomal transit time and V-ATPase dynamics are therefore unaffected by pathogenic C. neoformans indicating that the normal mode of release is through canonical constitutive exocytosis.

### *C. neoformans* Can Escape From Amoebae by Both WASH-Dependent and -Independent Mechanisms

To test the hypothesis that Cryptococci-containing phagosomes are normally expelled through the constitutive exocytosis pathway, we investigated exocytosis in WASH-null cells. In D. discoideum, WASH is essential for V-ATPase recycling and constitutive exocytosis, allowing us to specifically genetically ablate this pathway (Carnell et al., 2011).

In agreement with previous studies we found that phagosomes containing heat killed S. cerevisiae were never released from WASH-null amoebae within our 12 h period of observation (**Figure 3A**). In contrast, phagosomes containing C. neoformans were still exocytosed but with significantly altered dynamics (**Figure 3B**). Whilst >90% of C. neoformans-containing phagosomes are exocytosed within 2 h in wild type amoebae (**Figure 1F**), release from WASH-null cells was much more variable with between 20 and 60% escaping over 12 h (**Figure 3C**). Cryptococcus-containing phagosome transit was much slower through WASH-null cells with very little overlap with the exocytosis from wild-type amoebae (ca. **Figures 1F**, **3C**) and was much less synchronous, appearing to occur stochastically any time from 3 to >10 h (**Figure 3D**). Notably, both this variation and timing is comparable to that reported for vomocytosis of cryptococci from animal cells (Johnston and May, 2010).

The actin cytoskeleton is a negative regulator of vomocytosis in mammals, but essential for constitutive phagosome exocytosis in D. discoideum (Ma et al., 2006; Carnell et al., 2011). Therefore, we predicted that inhibition of actin polymerization would copy the phenotype of WASH-null Dictyostelium for both heat killed S. cerevisiae and C. neoformans containing phagosomes. When we measured the percentage exocytosis of heat killed S. cerevisiae containing phagosomes we found that while 100% of phagosomes were exocytosed by untreated amoebae we only observed a single exocytosis event out of 90 phagosomes analyzed over 12 h when actin polymerization was blocked by latrunculin A treatment after phagocytosis (**Figure 3E**). In contrast, ∼20% of C. neoformans-containing phagosomes in latrunculin A-treated amoebae were released over the same period (**Figure 3F**). Transit times were again significantly longer than with untreated amoebae, and were indistinguishable to the phenotype observed with WASH-null amoebae (**Figure 3G**). Thus, whilst cryptococci are normally released by constitutive exocytosis from D. discoideum, they can also escape by a mechanistically different route upon either pharmacological or genetic blockade of the constitutive pathway, highly reminiscent of vomocytosis.

#### *C. neoformans* Mutants *cap59* and *plb1* Do Not Exhibit Defects in Vomocytosis in Amoebae

We next tested whether cryptococcal virulence factors that affect vomocytosis from mammalian cells play conserved roles in egress from D. discoideum. C. neoformans mutants with defects in polysaccharide capsule formation or deletion of the phospholipase PLB1 both exhibit reduced rates of vomocytosis and pathogenicity in animal cells (Cox et al., 2001; Noverr et al., 2003; Chayakulkeeree et al., 2011; Evans et al., 2015). However, when we measured the rates of release of acapsular cap59 and plb1 mutant strains from wild-type D. discoideum we found that release of both mutants was unaffected: there were no significant differences in either frequency or transit time compared to the parental C. neoformans strain (H99) both with and without blockage of the constitutive pathway with latrunculin A (**Figures 4A,B**). This was an intriguing finding, suggesting differences in the signaling pathways to vomocytosis between amoebae and macrophages but a conservation of molecular mechanism.

### In the Absence of Constitutive Exocytosis *C. neoformans* Can Persist and Grow Intracellularly

Constitutive exocytosis of indigestible phagosomal material is critical for organisms that rely on phagocytosis for nutrition. Furthermore, amoebae may ingest microbes such as cryptococci that are able to actively resist phagosomal killing. Therefore, we investigated if, in the absence of constitutive exocytosis, cryptococci posed a greater threat to amoebae.

We observed no cell lysis of either wild-type or WASH-null cells infected with C. neoformans as indicated by the continued motility of the amoebae throughout the phagocytic cycle and after fungi egress. Low levels of amoeba lysis were observed upon latrunculin A treatment, most likely due to the severe effects of complete actin depolymerization. Killing and digestion

of yeast is indicated by the transformation of the phagosome to a granular and irregular shape, whereas growth can be inferred from yeast budding. When constitutive exocytosis was blocked with latrunculin A on average, more than 80% of cryptococci persisted within Dictyostelium phagosomes without signs of death and digestion, with 13% actively budding over the 12 h of the experiment (**Figures 5A,B**). In this respect the plb1 mutant behaved similarly to wild type cryptococci. The acapsular strain cap59 was phagocytosed twice as efficiently as the wild type cryptococcus strain (11 internalized cap59 cryptococci per 100

cell. Inset box is a magnification of exocytosed cryptococcal cell demonstrating absence of V-ATPase prior to exocytosis. Scale bars 10µm.

FIGURE 3 | Exocytosis of Cryptococcus neoformans from Dictyostelium is dependent on WASH and the actin cytoskeleton. (A) Heat killed S. cerevisiae are not exocytosed from WASH null D. discoideum. Example from 12 h time lapse imaging of heat killed S. cerevisiae in WASH null D. discoideum representative of 60 phagosomes containing heat killed S. cerevisiae from three independent experiments (images start at 500 min after phagocytosis). (B) C. neoformans are exocytosed from WASH null D. discoideum. Example from 12 h time lapse imaging representative of 62 phagosomes containing C. neoformans from three independent experiments (images start at 200 min after phagocytosis). Scale bars 5µm. (C) Quantification of the percentage of exocytosis of C. neoformans and heat killed S. cerevisiae from WASH null D. discoideum from three independent 12 h time lapses. (D) Frequency histogram of combined 35 transit times measured from three independent 12 h time lapses. (E) Exocytosis of heat killed S. cerevisiae from wild type Ax2 but not WASH null or latrunculin A treated Ax2 D. discoideum. Quantification of the percentage of exocytosis from three independent 12 h time lapses. Total of 60 phagosomes were analyzed from each condition. P-values are Fishers test. (F) Exocytosis of C. neoformans from wild type Ax2, WASH null, and latrunculin A treated D. discoideum. Quantification of the percentage of exocytosis from three independent 12 h time lapses. Total of 60 phagosomes were analyzed from each condition. P-values are Fishers test. (G) Transit times of C. neoformans through WASH null and latrunculin A treated Ax2 D. discoideum are not significantly different. P-values are Mann-Whitney test.

amoeba vs. 5.3 cryptococci per 100 amoeba, 300 amoeba analyzed from three independent experiments). However, the acapsular cap59 strain appeared to be growth-arrested within D. discoideum phagosomes, as budding was never observed (0/90 phagosomes, P = 0.0011 compared to H99, Fishers test; **Figure 5B**). Whilst this is consistent with other studies (Feldmesser et al., 2000; Steenbergen et al., 2001) surprisingly, we also never observed the collapse of cap59-containing phagosomes within the 12 h

indicating yeast death. This implies that either acapsular cells survive under growth arrest, or that death of cap59 cells occurs over a longer period than we are able to observe in this assay. Nonetheless, in the absence of constitutive exocytosis, yeast are able to replicate intracellularly within the retained phagosomes.

C. neoformans H99 wild type, mutant cap59 and plb1. Phagosomes analyzed the same as (A). P-values for between cryptococcal strain significance tests

are non-significant in all cases. P-values are Mann-Whitney test.

### DISCUSSION

In this work we have investigated the interactions between the environmental yeast C. neoformans and an environmental amoeba, D. discoideum. As C. neoformans can evade human immune cells and cause opportunistic infection in immunocompromised individuals we sought to test the hypothesis that the yeast virulence mechanisms had evolved to protect against amoebae in the environment.

In agreement with previous reports (Steenbergen et al., 2003), we found that C. neoformans were almost completely impervious to predation by D. discoideum; we consistently observed lower rates of C. neoformans phagocytosis compared to other particles, and even when the yeast were engulfed the amoebae were unable to kill them. This is similar to the interactions with mammalian macrophages, indicating that C. neoformans employs similar phagocyte evasion strategies. Chief amongst these is the characteristic cryptococcal capsule

which is highly anti-phagocytic and also provides protection from oxidative attack. Consistent with this both we, and others found capsule deficient yeasts were unable to grow within amoebae (Steenbergen et al., 2001, 2003) and we found that the acapsular strain was taken up twice as frequently as the wild type.

Whilst C. neoformans typically persist and proliferate in macrophages for many hours before escaping by vomocytosis, we found they passed through D. discoideum via an apparently normal phagosomal transit in just ∼80 min. This short time period provides limited opportunity for intracellular growth of fungal cells or the lysis of amoebae. Whilst others have shown that C. neoformans are able to grow (and therefore extract nutrients) when co-incubated with amoebae over several days (Steenbergen et al., 2001, 2003). The relatively short amount of time spent inside the amoeba implies that much of this growth is extracellular and intracellular growth has not been directly demonstrated by high resolution time lapse imaging (Steinman et al., 1976; Derengowski et al., 2013; Fu and Casadevall, 2018). We were unable to identify intracellular proliferation in the absence of the constitutive exocytosis inhibition. It therefore seems unlikely that C. neoformans has successfully evolved mechanisms to support intracellular replication within amoebae that exhibit rapid phagosomal transit, as demonstrated for D. discoideum, Amoeba proteus, Entamoeba histolytica and A. castellanii (Weisman and Korn, 1967; Ravdin et al., 1988; Christofidou-Solomidou and Stockem, 1992; Clarke et al., 2010).

Although D. discoideum are unable to kill phagocytosed C. neoformans the transit time is identical to that of heatkilled non-pathogenic yeast. This suggests that there is no major subversion of normal phagosome maturation. Consistent with this we find that the V-ATPase is both recruited, and retrieved from phagosomes with normal dynamics. This is in contrast to a recent study in macrophages, demonstrating that C. neoformans is able to disrupt phagosome maturation to inhibit acidification and proteolysis to permit intracellular proliferation (Smith et al., 2015). As the crypotococcal-containing phagosome in macrophages is permeabilized shortly after phagocytosis by macrophages (Tucker and Casadevall, 2002) V-ATPase delivery may be intact in both systems, but ineffective due to proton leakage.

Surprisingly, although C. neoformans are released alive from D. discoideum by canonical, WASH-dependent constitutive exocytosis, we found they were still expelled when the constitutive pathway was blocked. This second pathway strongly resembles vomocytosis from macrophages, being non-lytic, stochastic and inhibited by actin (Alvarez and Casadevall, 2006; Ma et al., 2006; Johnston and May, 2010). Vomocytosis remains mechanistically poorly understood and defined, but it seems highly likely that the WASH-independent egress of C. neoformans from D. discoideum is an analogous process.

Whilst laboratory strains have mutations that facilitate macropinocytosis and phagocytosis of large particles, wild-type isolates of D. discoideum are bacterivores and cannot engulf yeasts (Bloomfield et al., 2015). Therefore although D. discoideum provides a genetically tractable model for amoebae in general, it

#### REFERENCES


is highly unlikely to be an environmental host for cryptococci. The high phagocytic throughput of amoebae necessitates a mechanism to dispose of indigestible material, and unlike macrophages, there is no advantage in retaining phagosomes indefinitely to restrict an inflammatory response. As constitutive exocytosis appears to be sufficient for successful escape of C. neoformans on its own, the evolutionary drivers of a secondary, redundant escape mechanism are unclear. Whether other, more environmentally-relevant amoeba behave differently, or another environmental interaction altogether selects for this virulence trait requires further study.

Whatever the evolutionary basis, we have shown that vomocytosis-like egress is a mechanistically distinct process from constitutive exocytosis and is conserved in hosts from D. discoideum to man. The genetic tractability of Dictyostelium amoebae present an unparalleled opportunity to study the molecular cell biology of host cryptococcal interactions, and the differences in the environmental niche of C. neoformans and infection of humans.

#### AUTHOR CONTRIBUTIONS

RW, AA, CW, and CB performed experiments and data analysis. Initial concept, funding, experimental design were undertaken by JK and SJ. All authors prepared and edited the manuscript.

#### ACKNOWLEDGMENTS

We would like to thank Thierry Soldati for assistance and support developing this project. This work was supported by a Royal Society University Research Fellowship UF140624 to JK, and Medical Research Council and Department for International Development Career Development Award Fellowship MR/J009156/1 to SJ. We are grateful for a Microbiology Society Harry Smith Vacation studentship to support CW. RW was supported by a Bateson Centre BMedSci scholarship. Microscopy studies were supported by an MRC grant (G0700091) and a Wellcome Trust grant (GR077544AIA).


for secretion of phospholipase B1 and pathogenicity of Cryptococcus neoformans. Mol. Microbiol. 80, 1088–1101. doi: 10.1111/j.1365-2958.2011. 07632.x


growth but not intracellular survival in macrophages. Sci. Rep. 6:20977. doi: 10.1038/srep20977


Cryptococcus species. Microorganisms. 5:E65. doi: 10.3390/microorganisms5 040065


**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 Watkins, Andrews, Wynn, Barisch, King and Johnston. 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.

# Amoebae, Giant Viruses, and Virophages Make Up a Complex, Multilayered Threesome

Jan Diesend, Janis Kruse, Monica Hagedorn and Christian Hammann\*

Ribogenetics Biochemistry Lab, Department of Life Sciences and Chemistry, Jacobs University Bremen, Bremen, Germany

Viral infection had not been observed for amoebae, until the Acanthamoeba polyphaga mimivirus (APMV) was discovered in 2003. APMV belongs to the nucleocytoplasmatic large DNA virus (NCLDV) family and infects not only A. polyphaga, but also other professional phagocytes. Here, we review the Megavirales to give an overview of the current members of the Mimi- and Marseilleviridae families and their structural features during amoebal infection. We summarize the different steps of their infection cycle in A. polyphaga and Acanthamoeba castellani. Furthermore, we dive into the emerging field of virophages, which parasitize upon viral factories of the Megavirales family. The discovery of virophages in 2008 and research in recent years revealed an increasingly complex network of interactions between cell, giant virus, and virophage. Virophages seem to be highly abundant in the environment and occupy the same niches as the Mimiviridae and their hosts. Establishment of metagenomic and co-culture approaches rapidly increased the number of detected virophages over the recent years. Genetic interaction of cell and virophage might constitute a potent defense machinery against giant viruses and seems to be important for survival of the infected cell during mimivirus infections. Nonetheless, the molecular events during co-infection and the interactions of cell, giant virus, and virophage have not been elucidated, yet. However, the genetic interactions of these three, suggest an intricate, multilayered network during amoebal (co-)infections. Understanding these interactions could elucidate molecular events essential for proper viral factory activity and could implicate new ways of treating viruses that form viral factories.

#### Edited by:

Thierry Soldati, Université de Genève, Switzerland

#### Reviewed by:

Guruprasad R. Medigeshi, Translational Health Science and Technology Institute, India Bernard La Scola, Aix-Marseille University, France

#### \*Correspondence:

Christian Hammann c.hammann@jacobs-university.de

Received: 15 October 2017 Accepted: 13 December 2017 Published: 11 January 2018

#### Citation:

Diesend J, Kruse J, Hagedorn M and Hammann C (2018) Amoebae, Giant Viruses, and Virophages Make Up a Complex, Multilayered Threesome. Front. Cell. Infect. Microbiol. 7:527. doi: 10.3389/fcimb.2017.00527

Keywords: Acanthamoeba polyphaga mimivirus (APMV), virophage, nucleocytoplasmatic large DNA virus (NCLDV), mimivirus, pathogen defense

## INTRODUCTION TO GIANT VIRUSES

The discovery of giant viruses in the early 2000s led to a mind shift in the field of virology with respect to the potential origins of viruses (La Scola et al., 2003; Raoult et al., 2004). Originally, viruses were thought of as submicroscopic particles with a self-evident denial that viruses might exist, whose size would be large enough to be resolved with a simple light microscope (Lwoff, 1957; Raoult, 2013). Due to this mindset, the large, gram-positive particles in an Acanthamoeba polyphaga population were at first erroneously classified as bacteria (Birtles et al., 1997; La Scola et al., 2003; Raoult et al., 2007). Only the absence of ribosomal DNA in the presumed bacterium, led to the discovery and definition of the A. polyphaga mimivirus (APMV) in 2003 (La Scola et al., 2003). The acronym mimivirus (for mimicking microbe) reflects the resemblance to bacteria upon gram staining. At the same time, the discovery of APMV was the first ever report of a virus infecting amoebae. Amongst other features that are detailed below, APMV is unusual as it contains a large genome of 1.14 Mbp, thereby even surpassing the genome size of some bacterial species (Raoult et al., 2004). APMV particles are characterized by an up to 700 nm large capsid (**Figure 1A**), which is well above the resolution of a simple light microscope. Once it was established that giant DNA viruses of amoebae exist, many more such viruses, belonging to the nucleoplasmatic large DNA viruses (NCLDV) were found in the environment, as well as within a wide range of host organisms from humans, monkeys, and oysters (Boughalmi et al., 2013a; Dornas et al., 2014; Andrade et al., 2015). Ex vivo studies of human cell lines revealed that APMV is capable of infecting myeloid and mononuclear blood cells and interferes with the type I Interferon system (Silva et al., 2014). In addition, a distantly APMV-related NCLDV family member has been shown to productively infect T-lymphocytes under laboratory conditions (Popgeorgiev et al., 2013). In 2008, a small particle called Sputnik 1 (La Scola et al., 2008) was discovered in A. polyphaga, which parasitizes viral factories of giant viruses. Due to the functional similarity to bacteriophages in mediating lateral gene transfer, Sputnik was classified as a virophage (La Scola et al., 2008). Here, we will review the expanding family of virophages and discuss the implications for giant virus reproduction inside amoebae.

#### THE DIVERSE FAMILIES OF GIANT VIRUSES THAT INFECT AMOEBAE

The discovery of APMV sparked the interest in giant viruses and spawned a contemporary research field of its own (La Scola et al., 2003). Up until today, two giant virus families belonging to the NCLDV have been described that primarily infect amoebae: the Mimiviridae and the Marseilleviridae (**Figure 1B**). The latter has the A. polyphaga marseillevirus (APMaV) as founding member, which was discovered in 2009 (Boyer et al., 2009; Colson et al., 2013). In the last decade, nine additional viruses have been associated with the Marseilleviridae group (Colson et al., 2017). The Acanthamoeba castellani lausannevirus (ACLaV) was discovered by incubating water from the Seine river in France with A. castellani, a close relative of A. polyphaga (Thomas et al., 2011). ACLaV is the first known giant virus to encode histonelike proteins, which could point towards a DNA packaging mechanism similar to eukaryotes (Thomas et al., 2011). The Cannes 8 virus (Ca8V) (La Scola et al., 2010) and the Senegal virus (SNGV) (Lagier et al., 2012) have been isolated using similar co-culture methods and are grouped with the Marseilleviridae. The icosahedral capsid of the Marseilleviridae is between 190 and 250 nm in diameter (Colson et al., 2013). Like the genome of the Mimiviridae, the 370,000 bp dsDNA genome is encased in a lipid bilayer and encodes about 450 proteins (Boyer et al., 2009; La Scola et al., 2010; Thomas et al., 2011; Lagier et al., 2012). Both, Mimiviridae and Marseilleviridae, share only nine core genes with all NCLDVs (**Figure 1C**) and 180 genes are shared with at least two of the NCLDV families (Yutin et al., 2009; Yutin and Koonin, 2012). Based on the discovery of APMV and its complex genome, it was suggested to incorporate viruses into the tree of life by defining them as capsid-encoding organisms contrary to the ribosome-encoding organisms, which are represented by eukarya, bacteria, and archaea (Raoult and Forterre, 2008).

#### APMV—THE BEST STUDIED GIANT VIRUS OF AMOEBAE

APMV was the first giant virus to be discovered (La Scola et al., 2008) and confronted the scientific community with features never observed in a virus before. Its capsid size and genetic complexity with many genes usually found in eukaryotic and prokaryotic cells challenged the Lwoff's characteristics of a virus (Raoult et al., 2004; Raoult and Forterre, 2008). The AT-rich 1.14 Mbp APMV genome features an impressive number of 979 protein-encoding genes in a dense arrangement (Raoult et al., 2004; Legendre et al., 2011). Several of its genes are only found in giant viruses of amoebae and code for virus-atypical proteins involved in DNA repair, protein folding, tRNA synthesis and translation, and more (Raoult et al., 2004). In addition, the APMV genome displays some plasticity and encodes self-splicing introns, inteins, and a specific set of mobile genetic elements called transpovirons (Desnues et al., 2012). Furthermore, the genome contains many genes likely acquired via horizontal gene transfer, paralogous genes, and so called ORFans, genes that encode proteins with unknown function (Suhre, 2005; Filée et al., 2007; Moreira and Brochier-Armanet, 2008; Forterre, 2010). Many of these genes are shared with the poxviruses, phycodnaviruses, and other NCLDVs (Filée et al., 2007). ORFans represent roughly 50% of genes and about 40% of the APMV proteome, which results in a high number of factors with unknown functions that might act during viral replication and morphogenesis (Renesto et al., 2006). Alike "classical" viruses, APMV genes are partly under the control of early and late stage-specific promoters (Raoult et al., 2004; Suhre et al., 2005).

The APMV particles possess remarkable structural features, separating them from the classical structures of viruses (**Figure 1A**). In its center, the viral DNA, mRNAs and proteins are packed into the core compartment (Xiao et al., 2009; Kuznetsov et al., 2013) and enclosed by a lipid membrane. Among the pre-packed proteins are 12 enzymes involved in transcription, five in DNA repair, two in RNA modification, and five in protein modification (Renesto et al., 2006). The central compartment is surrounded by an approximately 340 nm-large lipid bilayer and a secondary bilayer directly underneath an icosahedral capsid. This is comprised of major capsid proteins and features a five-branch proteinaceous structure, the "stargate," at one vortex (Kuznetsov et al., 2013). The capsid itself is covered by a compact layer of about 120–140 nm long, heavily glycosylated fibrils, which potentially facilitate the attachment of APMV to its host cells (Rodrigues et al., 2015).

As of now, only four fiber associated proteins (FAP1-4) have been functionally associated with either fibril biosynthesis or as components of the fibrils (Sobhy et al., 2015). FAP1 is an aryl

alcohol oxidase, which catalyzes the degradation of lignin or lignin derivatives. This suggests that APMV might also be able to infect lignin-containing algae (Klose et al., 2015; Rodrigues et al., 2015). However, the fibrils and associated proteins are not essential for the productive infection of amoebae: during longterm intraamoebal culture (150 generations), the responsible genes are lost (Boyer et al., 2011; Rodrigues et al., 2015). This indicates that the genomic complexity of APMV might be maintained to allow for a broad host range. If so, only a subset of its diverse molecular tools would come in use to enter and infect individual hosts.

#### INFECTION CYCLES OF GIANT VIRUSES IN AMOEBAE

Even though the replication cycle of most giant viruses differ in aspects like nuclear involvement, duration, assembly, and release of the viral progeny, key steps in the infection appear to be conserved, as summarized recently (Colson et al., 2017). For example, all known giant viruses enter the host cell by phagocytosis and release their DNA into the cytosol in a similar manner (Ghigo et al., 2008). Furthermore, viral replication takes place in specialized endoplasmatic reticulum (ER)-derived compartments that are found in the cytosol and are called viral factories (Xiao et al., 2009; Mutsafi et al., 2010; Kuznetsov et al., 2013).

After uptake, the virus resides in a de-novo phagosome. Subsequently, the phagosomal and viral membranes fuse, which allows the release of the viral core, that contains the genome, proteins, and mRNAs into the cytosol (Zauberman et al., 2008; Mutsafi et al., 2010). Alike the well-described poxvirus (Broyles, 2003), the structural integrity of the viral core seems to be retained until viral factories arise (Claverie et al., 2009; Mutsafi et al., 2010). Intriguingly, recent experiments suggest that viral transcription might be initiated already before the release of the viral core (Mutsafi et al., 2014). Once in the cytosol, replication of the viral genome begins immediately and the expression of early stage genes leads to the formation of early viral factories (Suzan-Monti et al., 2007; Mutsafi et al., 2013, 2014). The replication cycle is confined to the cytosol, again a trait shared with the poxvirus (La Scola et al., 2003; Claverie et al., 2009). This also suggests that giant viruses (like the poxvirus) must carry transcription complexes to initiate transcription immediately after infection (Resch et al., 2007; Claverie et al., 2009). In later stages of infection, these viral factories merge into one large cytosolic compartment for replication and capsid assembly (Suzan-Monti et al., 2007; Mutsafi et al., 2014). It should be noted that viral factories are not chaotic, but rather appear to feature distinct assembly lines for their progeny. The viral factory is made up of functional regions playing discrete roles in replication, capsid assembly, DNA packaging, and attachment of fibrils (Suzan-Monti et al., 2007; Mutsafi et al., 2014). In the outermost layer of the viral factory, the internal membrane layers of APMV are assembled from host-derived membrane vesicles, which are thought to rupture, thereby forming open single-layer membrane sheets (Mutsafi et al., 2013). Capsid assembly occurs around these membrane sheets and is scaffolded by the major capsid protein L425 (Mutsafi et al., 2013). Upon capsid formation, the genome is deposited into the empty viral particle through a transient interstice distal from the "stargate" structure (Zauberman et al., 2008). There is little evidence for a nuclear stage of giant viruses. However, the nuclei of A. polyphaga and A. castellani exhibit transient changes in their morphology during the early stages of infection with members of the Marseilleviridae family (Arantes et al., 2016). This indicates that nuclear host factors might play a role in the APMV replication, a notion that is supported by a two-fold decrease of the nuclear size in infected A. polyphaga cells (Colson et al., 2017). This might be due to a substantial redistribution of nuclear factors for viral replication, transcription or other processes (Colson et al., 2017). Albeit indirectly, this scenario is supported by data on the cytoplasmic replication of the Vaccinia virus (a poxvirus), to which mimivirus replication bears similarities (Mutsafi et al., 2010) and for which the involvement of nuclear enzymes has been demonstrated (Oh and Broyles, 2005).

### VIROPHAGES AS PARASITES OF THE MEGAVIRALES

The description of Megavirales infection of amoebae was followed by the discovery of the fascinating virophage Sputnik in 2008 (La Scola et al., 2008). Sputnik was found infecting the viral factories of the mamavirus, a close relative of APMV (La Scola et al., 2008). Replication of the Sputnik virophages inside APMVinfected A. castellani cells is deleterious to APMV replication and results in abortive DNA replication and disruption of capsid biogenesis (La Scola et al., 2008). There is an ongoing discussion on the classification of virophages, that are denoted in several articles as satellite viruses (Krupovic and Cvirkaite-Krupovic, 2011; Blanc et al., 2015; Koonin and Krupovic, 2017). Satellite viruses are characterized by their dependency on factors of a helper virus. However, the Sputnik genomes itself encodes factors involved in viral replication (La Scola et al., 2008), suggesting that Sputnik can be classified as a virus, rather than a defective viral particle or sub-viral agent (Fischer, 2011; Desnues and Raoult, 2012).

All known members of the virophage family parasitizing on giant viruses are categorized into the large virus-dependent or -associated (Lavida-)viridae family that is divided into the Sputnikvirus and Mavirus genera (Krupovic et al., 2016). At the species level, the Sputnikvirus genus can be differentiated into the APMV-dependent Sputnik virophage and the APMV-dependent Zamilon virophage (**Table 1**), while Mavirus genus contains only the Cafeteria roenbergensis virus (CroV)-dependent mavirus (Krupovic et al., 2016).

Virophage replication has been extensively studied in particular for Sputnik, Zamilon and mavirus. Studies on amoebae infected with different mimiviruses revealed that Sputnik virophages can parasitize mimiviruses from all Mimiviridae lineages but apparently not the Marseilleviridae lineages (Gaia et al., 2013). Sputnik replicates inside mamavirus-infected A. castellani cells within the viral factories, nonetheless, with different kinetics as the mamavirus and at multiple hot spots inside the factory (La Scola et al., 2008). In APMVs viral factories, Sputnik infection results in the emergence of newly generated particles 6 h post infection with a concomitant decrease of infective APMV particles (Ogata and Claverie, 2008). The 18,343 kilobase circular dsDNA genome of Sputnik possesses 21 partly overlapping open-reading frames (ORFs) encoding for several factors involved in DNA replication (La Scola et al., 2008). Interestingly, four of the ORFs are strongly homologous to APMV-encoded genes (La Scola et al., 2008; Gaia et al., 2013). Since Sputnik virophages encodes a lambda-type integrase, the molecular tools for genomic integration are present (La Scola et al., 2008). Indeed, an integration of the Sputnik genome into the genome of the Lentille virus, a relative of APMV, could be observed experimentally (Desnues et al., 2012). There is no indication of Sputnik genome integration into the host cell genome, in line with the lack of indications for a nuclear phase.

The Zamilon virophage (belonging to the Sputnikvirus genus) was discovered together with the Mont1 mimivirus in soil samples from Tunisia (Boughalmi et al., 2013a; Gaia et al., 2014). The 60 nm-wide, spherical virophage carries a 17,276 bp dsDNA genome encoding 20 genes. Although Zamilon shares 76% of its genomic sequence with Sputnik, Zamilon can only infect lineages B and C (Gaia et al., 2014). Furthermore, the tv\_L8 protein, encoded in the transpovirons of the Monve mimivirus, shares significant homology with the ORF8-encoded protein of Zamilon (Gaia et al., 2014). This suggests that an exchange of genetic material can in principle occur between the giant virus and the Zamilon virophage within co-infected amoebae, although this has not been observed experimentally so far.

The Maverick-related virus (mavirus), lonely member of the Mavirus genus, parasitizes the viral factories of CroV that infects the marine heterotrophic nanoflagellate C. roenbergensis (Fischer et al., 2010; Fischer and Suttle, 2011). Although this review is predominantly concerned with infection of amoebae, mavirus is included here for its unique features for a virophage. Its 19,063 bp circular genome possesses 20 ORFs including a retroviral integrase, an unsual, protein-primed DNA polymerase, plus four additional proteins, all of which are also found conserved in Maverick/Polinton (MP) retroelements (Fischer and Suttle, 2011; Krupovic et al., 2014, 2016). Additionally, the termini of the mavirus genome consist of long terminal repeats similar to those found in MP retroelements (Yutin et al., 2013; Krupovic et al., 2016). Both findings suggest that these retroelements might have originated from mavirus genome integration events in mavirus co-infected cells (Fischer and Suttle, 2011; Krupovic et al., 2016). Nonetheless, this hypothesis for the origins of MP retroelements remains to be tested experimentally. Fischer and Hackl (2016) succeeded to monitor the integration of mavirus into the C. roenbergensis genome by co-infection with a low multiplicity of infection of CroV. Intriguingly, genes in the


(Continued)

**143**

Frontiers in Cellular and Infection Microbiology | www.frontiersin.org January 2018 | Volume 7 | Article 527


aLineage.

mavirus genome possess promoter sequences similar to the late stage promoter of CroV (Fischer and Hackl, 2016). As a consequence, re-infection of C. roenbergensis carrying the integrated mavirus genome with CroV resulted in inhibition of CroV DNA replication, concomitantly with an increased survival of C. roenbergensis (Fischer and Hackl, 2016).

Other virophages have been discovered by metagenomic analysis of water samples [e.g., the Organic Lake virophage (Yau et al., 2011), the Yellowstone Lake virophages (Zhou et al., 2013, 2015)]. However, the viral and cellular host for these remain to be determined (Krupovic et al., 2016), unlike the situation of the Rio Negro virophage that has the Samba virus as viral host (Campos et al., 2014).

#### OUTLOOK

Since the discovery of its first member APMV in 2003, new giant viruses are discovered continuously in samples

#### REFERENCES


from all over the world and added to the Megavirales family. The addition of virophages as parasites of giant viruses, their high abundance in the environment, and the genetic interactions between cell, giant virus, and virophage, suggest an intricate, multilayered network during amoebal co- and super-infections. Future studies of these dynamic interactions could elucidate the inner mechanics of viral factories.

#### AUTHOR CONTRIBUTIONS

All authors designed the mini review. JD: generated the figures and drafted the text; MH and CH: wrote the manuscript.

#### FUNDING

This work is supported by a grant of the Tönjes-Vagt-Stiftung Bremen, Germany.


**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 Diesend, Kruse, Hagedorn and Hammann. 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) or licensor 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.

**147**

# From Phagocytes to Immune Defense: Roles for Coronin Proteins in *Dictyostelium* and Mammalian Immunity

#### Mayumi Mori, Ravindra Mode and Jean Pieters\*

Biozentrum, University of Basel, Basel, Switzerland

Microbes have interacted with eukaryotic cells for as long as they have been co-existing. While many of these interactions are beneficial for both the microbe as well as the eukaryotic cell, several microbes have evolved into pathogenic species. For some of these pathogens, host cell invasion results in irreparable damage and thus host cell destruction, whereas others use the host to avoid immune detection and elimination. One of the latter pathogens is Mycobacterium tuberculosis, arguably one of the most notorious pathogens on earth. In mammalian macrophages, M. tuberculosis manages to survive within infected macrophages by avoiding intracellular degradation in lysosomes using a number of different strategies. One of these is based on the recruitment and phagosomal retention of the host protein coronin 1, that is a member of the coronin protein family and a mammalian homolog of coronin A, a protein identified in Dictyostelium. Besides mediating mycobacterial survival in macrophages, coronin 1 is also an important regulator of naïve T cell homeostasis. How, exactly, coronin 1 mediates its activity in immune cells remains unclear. While in lower eukaryotes coronins are involved in cytoskeletal regulation, the functions of the seven coronin members in mammals are less clear. Dictyostelium coronins may have maintained multiple functions, whereas the mammalian coronins may have evolved from regulators of the cytoskeleton to modulators of signal transduction. In this minireview, we will discuss the different studies that have contributed to understand the molecular and cellular functions of coronin proteins in mammals and Dictyostelium.

#### *Edited by:*

Sascha Thewes, Freie Universität Berlin, Germany

#### *Reviewed by:*

Maria Isabel Colombo, Universidad Nacional de Cuyo, Argentina Barbara Walzog, Ludwig-Maximilians-Universität München, Germany

#### *\*Correspondence:*

Jean Pieters jean.pieters@unibas.ch

*Received:* 13 October 2017 *Accepted:* 27 February 2018 *Published:* 22 March 2018

#### *Citation:*

Mori M, Mode R and Pieters J (2018) From Phagocytes to Immune Defense: Roles for Coronin Proteins in Dictyostelium and Mammalian Immunity. Front. Cell. Infect. Microbiol. 8:77. doi: 10.3389/fcimb.2018.00077 Keywords: coronins, *Dictyostelium discoideum*, Mycobacterium, phagocytes, immune cells

### INTRODUCTION

All eukaryotes are surrounded by microorganisms that on the one hand fulfill important roles in providing symbiotic support for eukaryotic life, and at the same time can pose a threat in the form of virulent bacteria, causing infections, disease and death.

For both lower eukaryotes such as the amoeba Dictyostelium discoideum, as well as mammalian macrophages, the first encounter with microbes, especially bacteria, results in the activation of phagocytic processes leading to engulfment of the bacteria within phagosomes, followed by lysosomal digestion (Flannagan et al., 2012). When bacteria serve as nutrients, as is the case for Dictyostelium, lysosomal degradation will allow the availability of amino acids, lipids, and other molecules that serve the need for Dictyostelium growth (Allen and Aderem, 1996). In case of bacteria being phagocytosed by mammalian cells such as dendritic cells and macrophages, the end point is usually inactivation of the pathogen, as well as presentation of pathogen fragments to the immune system in order to activate effector lymphocytes (T cells) for the generation of adaptive immune responses (Pieters, 2000; Blum et al., 2013).

Phagocytosis is an ancient and evolutionary conserved mechanism (Aderem and Underhill, 1999). Indeed, the basic mechanisms regulating the formation of a phagocytic cup, internalization of the bacteria and the transfer from phagosomes to lysosomes are conserved from lower eukaryotes such as Dictyostelium to mammalian phagocytes, including macrophages, neutrophils and dendritic cells. Phagocytosis involves cell surface recognition through different plasma membrane receptors that transmit signals through a variety of pathways to the cytoskeleton in order to allow plasma membrane deformation to accommodate the incoming particles/bacteria (Flannagan et al., 2012).

### Virulence Strategies Employed by Pathogenic Microbes

Whereas the encounter of mammalian phagocytes with bacteria most often results in its destruction and activation of specific immunity, many pathogenic bacteria have evolved a diverse array of strategies to circumvent phagocytosis and lysosomal destruction. For example, several pathogens, including Neisseria spp., Pseudomonas spp., Streptococcus spp., and Salmonella spp. are known to avoid phagocytic uptake (Sarantis and Grinstein, 2012). Also, certain pathogens, including Listeria spp. and Shigella spp. are engulfed by phagocytic cells but rapidly transfer to the cytosol where they can proliferate and spread (Hamon et al., 2012; Mellouk and Enninga, 2016). Yet other bacilli such as Brucella spp. and Salmonella spp. enter the phagocyte through phagocytosis but divert to the endoplasmic reticulum (Brucella spp.) or the Golgi (Salmonella spp.) instead of being delivered to lysosomes (Celli and Gorvel, 2004; Escoll et al., 2016). Furthermore, some pathogens are phagocytosed and delivered to lysosomes and subsequently withstand the hostile environment of the lysosomal pathway through the release of neutralizing factors (Voth and Heinzen, 2007).

Several pathogens use more than one of these strategies to circumvent host cell destruction; in fact, one of the most notorious pathogens, Mycobacterium tuberculosis, has evolved multiple mechanisms to avoid killing by immune cells as well as recognition by the adaptive immune system. For example, M. tuberculosis manipulates the host ability to detect and internalize the bacilli through pathogen-associated molecular patterns (PAMPs) (Stamm et al., 2015). Also, once inside the host cell, M. tuberculosis uses a range of strategies to avoid destruction, including manipulation of V-ATPase levels to avoid lysosomal acidification (Rohde et al., 2007), expression of genes that allow the bacilli to withstand the low pH of the endosomal/lysosomal pathway (Vandal et al., 2008), neutralizing reactive nitrogen and oxygen species (Shiloh and Nathan, 2000), resisting delivery to autophagosomes (Deretic, 2005; Romagnoli et al., 2012), counter-balancing iron depletion encountered upon phagocytosis (Weiss and Schaible, 2015) and interfering with lysosomal delivery following phagocytosis (Pieters, 2008; Gengenbacher and Kaufmann, 2012).

Inhibition of lysosomal delivery is an important trait of pathogenic mycobacteria, and the bacilli devote considerable efforts toward achieving that goal (Rohde et al., 2007; Pieters, 2008). This is achieved through the release of molecules, both signaling molecules, lipids and proteins (Cowley et al., 2004; Walburger et al., 2004; Hmama et al., 2015; Lovewell et al., 2016), as well as through the recruitment of host factors allowing mycobacterial escape from lysosomal degradation.

### Role for Coronin 1 in the Interaction of *M. tuberculosis* With Macrophages

One of the host proteins co-opted by M. tuberculosis to avoid lysosomal killing is coronin 1 (encoded by the coro1a gene; for a discussion on the coronin nomenclature see; Pieters et al., 2013), a ∼51 kD protein that is located in the macrophage cytosol and cell cortex. Coronin 1, also known as P57 or TACO, for Tryptophan Aspartate containing Coat protein, was identified in a search for host components potentially involved in the intracellular survival of mycobacteria within macrophages (Ferrari et al., 1999). Coronin 1 is a member of the highly conserved family of coronin proteins whose members are expressed across the eukaryotic kingdom and are characterized by the presence of a central WD (tryptophan-aspartate) 40 repeat that in coronin 1-folds in a 7-bladed beta propeller (Suzuki et al., 1995; Okumura et al., 1998; Gatfield et al., 2005; Appleton et al., 2006). Coronin 1 is highly expressed in all hematopoietic cells as well as to a lower degree in neurons (Ferrari et al., 1999; Nal et al., 2004; Jayachandran et al., 2014). Upon entry of pathogenic mycobacteria into macrophages, coronin 1 is retained on phagosomes containing viable, but not killed mycobacteria and its retention on phagosomes prevents intracellular killing of M. tuberculosis through activation of the Ca2+/calcineurin pathway (Jayachandran et al., 2007). Apart from M. tuberculosis, M. leprae as well as virulent H. pylori recruit coronin 1 to their intracellular niche, although the exact consequences of coronin 1 retention for pathogen survival in these latter cases remain unclear (Zheng and Jones, 2003; Suzuki et al., 2006).

### CORONINS IN DICTYOSTELIUM

Coronin was initially isolated as a 55 kD actin/myosin-interacting molecule from Dictyostelium discoideum lysates (de Hostos et al., 1991). Subsequently, it was realized that Dictyostelium also expresses a so-called "tandem" coronin containing a duplicated tryptophan-aspartate repeat-containing region, now referred to as coronin B (Shina et al., 2010). Coronin B assembles with actin and deletion of this coronin in Dictyostelium was shown to both enhance as well as reduce phagocytosis, dependent on the particles to be internalized (Shina et al., 2010). Coronin B was proposed to act upstream of the suppressor of cAMP receptor (SCAR) and Wiskott-Aldrich syndrome protein

(WASp) (Swaminathan et al., 2015) which connect signals from G protein-coupled receptors and cell surface tyrosine receptors to the actin cytoskeleton, respectively (Bear et al., 1998; Pollitt and Insall, 2009).

In Dictyostelium, coronin A is involved in a diverse array of activities, including cell motility, cAMP-mediated chemotaxis, and cytokinesis (de Hostos et al., 1993; Nagasaki et al., 2002). Given the initial isolation of coronin A with actin/myosin, the roles for coronin A in the above-mentioned activities have been attributed to the capacity of Dictyostelium coronin A in modulating the F-actin cytoskeleton. Coronin A is localized within regions of actin turnover (Maniak et al., 1995; Heinrich et al., 2008), leading to the conclusion that Dictyostelium coronin A is a regulator of the F-actin cytoskeleton, thereby modulating chemotaxis, cell motility and cytokinesis; it should however be noted that Dictyostelium lacking coronin A do not show an obvious alteration in the assembly or localization of actin filaments (de Hostos et al., 1993). Separate work using yeast coronin (Crn1) has shown that while deletion of yeast Crn1 does not result in an aberrant F-actin cytoskeleton (Heil-Chapdelaine et al., 1998; Goode et al., 1999), Crn1 binds to and bundles F-actin in vitro (Goode et al., 1999) as well as can modulate F-actin polymerization either positively (Liu et al., 2011) or negatively (Humphries et al., 2002) in an actin-related protein 2/3 (ARP2/3) complex-dependent manner. Indeed, yeast Crn1 possesses a number of regions of homology with actinand tubulin-binding proteins, including microtubule- and Factin/ARP2/3-interacting domains (Liu et al., 2011), that are lacking in most other coronins (Eckert et al., 2011).

More recent work that revisited the role for coronin A in Dictyostelium found that coronin A is important for the initiation of multicellular differentiation following deprivation of nutrients (Vinet et al., 2014); in Dictyostelium, nutrient starvation induces the aggregation of individual amoebae into a multicellular structure ultimately forming a fruiting body. Such aggregation is mediated by the second messenger cAMP, that functions both as a chemoattractant as well as an intracellular signal mediating gene transcription. Aggregation is initiated by cell density and nutrient-sensing factors released by the starving culture that induce the expression of genes involved in multicellular aggregation (Devreotes, 1989; Loomis, 2014). It was found that coronin A plays an essential role in the initiation of this developmental program by being involved in the response to factors secreted during the transition from growth to development of the cells. Since application of cAMP to coronin A-deficient cells is sufficient to restore chemotaxis and multicellular aggregation, coronin A appears to be dispensable for the cAMP relay as well as for processes downstream of cAMP. Also, the fact that folate chemotaxis occurs normally in the absence of coronin A argues against an exclusive role for coronin A in cytoskeletal remodeling (Vinet et al., 2014). Furthermore, consistent with earlier results showing that Factin rearrangement is not required for the initiation of cAMP signaling (Parent et al., 1998; Kriebel et al., 2008), coronin Adependent induction of genes required for development such as aca and carA does not require F-actin-based rearrangement (Vinet et al., 2014).

These data suggests that instead of modulating F-actin, coronin A is responsible for the sensing of factors secreted in the conditioned medium (Vinet et al., 2014). Coronin A may function downstream of conditioned medium factor (CMF) (Jain et al., 1992; Yuen et al., 1995), or possibly other, as yet undefined factors that are essential for the initiation of cAMPdependent chemotaxis and aggregation. Whether the motility and cytokinesis defects observed upon coronin A deletion (de Hostos et al., 1993; Vinet et al., 2014) are related to a possible function of coronin A in the modulation of the cAMP pathway or linked to a role for coronin A in F-actin-mediated processes remains to be clarified. In this context, it is interesting to note that in Dictyostelium, myosin-independent cytokinesis (one of the two types of cytokinesis, the other one being myosin-IIdependent, see Nagasaki et al., 2002; Li, 2007) has been linked to both coronin A as well as AmiA (also known as Pianissimo, a target of rapamycin complex (TORC) 2-associated protein) that in Dictyostelium is a cytosolic regulator of adenylate cyclase (Pergolizzi et al., 2002). Furthermore, recent work implicated coronin A in regulating the availability of GTP-Rac through a domain with homology to a Cdc42- and Rac-interactive binding (CRIB) motif for activation of downstream effectors, thereby being responsible for myosin II disassembly (Swaminathan et al., 2014). Interestingly, this CRIB-like domain was found to be dispensable for the role of coronin A in cytokinesis, since expression of a mutant coronin A lacking Rac binding activity rescued the cytokinesis defect of coronin A-deficient cells (Swaminathan et al., 2014). This suggest that coronin A domains distinct from the CRIB-like motif are involved in the regulation of cytokinesis.

The precise role for Dictyostelium coronins in the modulation of bacterial uptake and survival is less clear. A role for coronin A in phagocytosis appears to be dependent on the type of cargo that is internalized; upon coronin A deletion, phagocytosis of E. coli and yeast particles is reduced, while uptake of Mycobacterium marinum is enhanced (Maniak et al., 1995; Solomon et al., 2003). Also, it should be noted that yeast phagocytosis is increased upon coronin B deletion (Shina et al., 2011). Furthermore, while coronin A becomes enriched on newly formed phagosomes, it is rapidly dissociated once the bacteria have been fully internalized (Maniak et al., 1995; Rauchenberger et al., 1997; Lu and Clarke, 2005; Hagedorn and Soldati, 2007). For both M. marinum as well as Legionella pneumonia, deletion of coronin A renders Dictyostelium more permissive for intracellular bacterial growth (Solomon et al., 2000, 2003). These data suggest that in Dictyostelium coronin A may play a protective role for the host, although the mechanisms involved remain unclear.

### CORONIN 1 FUNCTION IN MAMMALIAN LEUKOCYTES

As described above, in resting macrophages, the only role for coronin 1 appears to be the modulation of the intracellular trafficking and survival of M. tuberculosis (Jayachandran et al., 2007, 2008). Given the proposed role for Dictyostelium coronin A in the modulation of F-actin-dependent processes such as chemotaxis and cytokinesis, it was initially anticipated that its mammalian homolog in macrophages, coronin 1, also modulates F-actin. However, in macrophages depleted for coronin 1 either by short interference (si)RNA or gene ablation, actindependent functions appear to be unperturbed as judged by the analysis of cell motility, macropinocytosis and membrane ruffling (Jayachandran et al., 2007, 2008). In contrast to studies using macrophages from coronin 1-deficient mice (Jayachandran et al., 2007) or J774 macrophages depleted for coronin 1 by siRNA (Jayachandran et al., 2008), TAT-mediated transduction of the WD repeat domain of coronin 1 in RAW 264.7 macrophages and neutrophils was shown to affect early phagocytosis (Yan et al., 2005, 2007); this may reflect differences in the experimental setup (such as time to allow uptake) or differential requirements for coronin 1 in phagocytosis in different macrophage and/or cell types; alternatively, especially given the propensity of coronin 1 mutants to cause aggregation, the introduction of misfolded protein domains may compromise cellular functions such as chemotaxis and phagocytosis (Gatfield et al., 2005). Whereas coronin 1 does not appear to be required for the functionality of resting macrophages, during an inflammatory stimulus coronin 1 is responsible for reprogramming the uptake pathway from phagocytosis to macropinocytosis in order to rapidly eliminate pathogens, a function that is dependent on activation of phosphoinositol (PI)-3-kinase (Bosedasgupta and Pieters, 2014). Furthermore, coronin 1 is largely dispensable for the functioning of B cells, mast cells, dendritic cells and natural killer cells, although the latter have been described to be affected by coronin 1 mutation in human (Moshous et al., 2013; Mace and Orange, 2014; Jayachandran and Pieters, 2015; Tchang et al., 2017). Also, coronin 1 was found to be dispensable for neutrophil function and recruitment in an in vivo model of liver injury and concanavanin A-induced hepatitis (Combaluzier and Pieters, 2009; Siegmund et al., 2013); in humans, coronin 1 has been associated with neutrophil survival and recent work has implicated coronin 1 in integrin-mediated functioning (Moriceau et al., 2009; Pick et al., 2017).

The in-depth analysis of coronin 1-deficient mice revealed that besides protecting intracellular mycobacteria from degradation within macrophages, a major function of coronin 1 is to regulate peripheral naïve T cell homeostasis; upon depletion of coronin 1, naïve T cells are virtually absent despite a normal development and selection in the thymus (Föger et al., 2006; Haraldsson et al., 2008; Mueller et al., 2008; Shiow et al., 2008; Lang et al., 2017). The role for coronin 1 in maintaining peripheral naïve T cells is conserved in humans: deletion or mutation of the coro1a gene has been reported to result in a selective depletion of naïve T cells (Moshous et al., 2013; Yee et al., 2016), or, when coro1a deletion or mutation is combined with other genetic aberrations, in more complex phenotypes including B and NK cell deficits besides the naïve T cell depletion (Shiow et al., 2009; Mace and Orange, 2014; Stray-Pedersen et al., 2014; Punwani et al., 2015; Yee et al., 2016). The mechanism underlying coronin 1-dependent naïve T cell survival remains controversial; in one study, coronin 1 was suggested to modulate the F-actin cytoskeleton, thereby regulating T cell survival (Föger et al., 2006); however, separate studies showed that altered F-actin levels do not correlate with T cell viability nor are other actin-dependent leukocyte functions affected by coronin 1 deletion (Jayachandran et al., 2007; Mueller et al., 2011). Instead, coronin 1-deficient T cells were shown to be unable to respond to a range of T cell stimuli and the defect was narrowed down to be at the level of activation of the Ca2+/calcineurin pathway (Haraldsson et al., 2008; Mueller et al., 2008).

Further analysis of mice lacking coronin 1 revealed an important function of this coronin family member in neurons, where it regulates Ca2+- and cAMP-dependent signaling thereby modulating various neuronal activities, including cognition and behavior as well as target innervation (Jayachandran et al., 2014; Suo et al., 2014). It is also interesting to note that M. tuberculosis is known to subvert host cAMP signaling (Agarwal et al., 2009), and retention of coronin 1 at the phagosomal membrane may be part of this strategy. Whether and how Ca2<sup>+</sup> and cAMP signaling are interconnected and whether coronin 1-dependent cAMP signaling plays a role in T cell homeostasis and mycobacterial survival within macrophages remains to be analyzed.

### A CONCERTED ROLE FOR CORONINS IN *DICTYOSTELIUM* AND MAMMALS?

The available information on the role for coronins in Dictyostelium and mammals suggests that these proteins play diverse roles in a number of physiological processes. The hallmark of all coronin protein family members is their central WD40 repeat, that folds into a beta propeller structure. Beta-propellers, also known as beta-transducin repeats, form structural domains that are involved in protein-protein interaction (Smith, 2008). Both in Dictyostelium as well as in mammalian cells, several coronin family members colocalize with and are associated with actin (de Hostos et al., 1991; Shina et al., 2010; Pieters et al., 2013) whereas for a number of mammalian coronin proteins (coronin 2, 5, and 7) actin binding remains unclear (see e.g., Rybakin et al., 2004; Cai et al., 2007). It of course remains possible, especially in mammals with up to 7 coronin molecules being expressed, that the role for coronins in actin rearrangement is redundant and therefore single deletions may not result in an actin-dependent phenotype. On the other hand, it could be that the interaction of coronin molecules with the actin cytoskeleton ensures a local source of coronin molecules to allow conversion of extracellular signals into local changes in the cortical actin cytoskeleton (Wang et al., 1998; Eichinger et al., 1999; Gatfield et al., 2005). Such a role for coronins is consistent not only with their sequence as well as structural homology with the beta subunit of trimeric G proteins that function downstream of G protein-coupled receptor molecules (de Hostos et al., 1991; Gatfield et al., 2005; Appleton et al., 2006), but also with the activities of Dictyostelium coronin A and mammalian coronin 1 in the modulation of Ca2+- and cAMP-dependent signal transduction pathways (Jayachandran et al., 2014; Suo et al., 2014; Vinet et al., 2014). Also, recent work linking coronins to the activation of small GTP binding proteins, that are well known regulators of the actin cytoskeleton (Berzat and Hall, 2010; Castro-Castro et al., 2011; Swaminathan et al., 2014), suggests that coronins may be placed upstream of F-actin reorganization. Interestingly, in both Dictyostelium as well as mammalian cells, the role for coronin in the activation of Ca2+/cAMP signaling could be separated from a potential involvement in F-actin rearrangement (Mueller et al., 2007, 2011; Jayachandran et al., 2014; Vinet et al., 2014). How, exactly, coronin molecules are being regulated is unknown. Also, the molecules upstream of coronin 1 possibly involved in the sensing of extracellular signals remain to be identified. In light of the here described roles for mammalian and Dictyostelium coronins in the trafficking and survival of intracellular pathogens, elucidation of these upstream receptor molecules may also shed light on the intricate relationship of pathogenic microbes and their eukaryotic hosts.

#### REFERENCES


#### AUTHOR CONTRIBUTIONS

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

#### ACKNOWLEDGMENTS

We thank Rajesh Jayachadran and Tohnyui Ndinyanka Fabrice for critical comments on the manuscript. Work in the Pieters Laboratory is supported by the Swiss National Science Foundation, the Swiss Multiple Sclerosis Society, the Novartis Foundation for Medical-Biological Research, the Gebert Ruef Foundation and the Canton of Basel.


mediates functions of coronin. Proc. Natl. Acad. Sci. U.S.A. 111, E25–E33. doi: 10.1073/pnas.1315368111


**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, Mode and Pieters. 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 Saposin-Like Protein AplD Displays Pore-Forming Activity and Participates in Defense Against Bacterial Infection During a Multicellular Stage of *Dictyostelium discoideum*

#### Ranjani Dhakshinamoorthy 1†, Moritz Bitzhenner <sup>1</sup> , Pierre Cosson<sup>2</sup> , Thierry Soldati <sup>3</sup> and Matthias Leippe<sup>1</sup> \*

#### *Edited by:*

Patricia Ann Champion, University of Notre Dame, United States

#### *Reviewed by:*

Falk Hillmann, Leibniz-Institut für Naturstoff-Forschung und Infektionsbiologie, Hans Knöll Institut, Germany Serge Ankri, Technion – Israel Institute of Technology, Israel

> *\*Correspondence:* Matthias Leippe mleippe@zoologie.uni-kiel.de

#### †*Present Address:*

Ranjani Dhakshinamoorthy, Department of Biotechnology, Bhupat and Jyoti Mehta School of Biosciences, Indian Institute of Technology Madras, Chennai, India

> *Received:* 22 November 2017 *Accepted:* 27 February 2018 *Published:* 15 March 2018

#### *Citation:*

Dhakshinamoorthy R, Bitzhenner M, Cosson P, Soldati T and Leippe M (2018) The Saposin-Like Protein AplD Displays Pore-Forming Activity and Participates in Defense Against Bacterial Infection During a Multicellular Stage of Dictyostelium discoideum. Front. Cell. Infect. Microbiol. 8:73. doi: 10.3389/fcimb.2018.00073

<sup>1</sup> Zoological Institute, Comparative Immunobiology, University of Kiel, Kiel, Germany, <sup>2</sup> Department of Cell Physiology and Metabolism, Faculty of Medicine, University of Geneva, Geneva, Switzerland, <sup>3</sup> Department of Biochemistry, Faculty of Science, University of Geneva, Geneva, Switzerland

Due to their archaic life style and microbivor behavior, amoebae may represent a source of antimicrobial peptides and proteins. The amoebic protozoon Dictyostelium discoideum has been a model organism in cell biology for decades and has recently also been used for research on host-pathogen interactions and the evolution of innate immunity. In the genome of D. discoideum, genes can be identified that potentially allow the synthesis of a variety of antimicrobial proteins. However, at the protein level only very few antimicrobial proteins have been characterized that may interact directly with bacteria and help in fighting infection of D. discoideum with potential pathogens. Here, we focus on a large group of gene products that structurally belong to the saposin-like protein (SAPLIP) family and which members we named provisionally Apls (amoebapore-like peptides) according to their similarity to a comprehensively studied antimicrobial and cytotoxic pore-forming protein of the protozoan parasite Entamoeba histolytica. We focused on AplD because it is the only Apl gene that is reported to be primarily transcribed further during the multicellular stages such as the mobile slug stage. Upon knock-out (KO) of the gene, aplD<sup>−</sup> slugs became highly vulnerable to virulent Klebsiella pneumoniae. AplD<sup>−</sup> slugs harbored bacterial clumps in their interior and were unable to slough off the pathogen in their slime sheath. Re-expression of AplD in aplD<sup>−</sup> slugs rescued the susceptibility toward K. pneumoniae. The purified recombinant protein rAplD formed pores in liposomes and was also capable of permeabilizing the membrane of live Bacillus megaterium. We propose that the multifarious Apl family of D. discoideum comprises antimicrobial effector polypeptides that are instrumental to interact with bacteria and their phospholipid membranes. The variety of its members would allow a complementary and synergistic action against a variety of microbes, which the amoeba encounters in its environment.

Keywords: amoebapore, antimicrobial peptides, *Dictyostelium discoideum*, host-pathogen interactions, saposinlike proteins, slugs, Sentinel cells, innate immunity

## INTRODUCTION

Amoebozoa are interesting models to study the early evolution of innate immunity (Leippe, 1999). The social amoeba Dictyostelium discoideum, a genetically tractable model for the study of cell biology, has recently become a powerful model organism in infection biology. In particular, it has been employed as a surrogate host for human pathogens such as Legionella (Farbrother et al., 2006), Mycobacteria (Hagedorn and Soldati, 2007; Hagedorn et al., 2009), and Pseudomonas (Cosson et al., 2002; Alibaud et al., 2008). In its amoebic stage, D. discoideum phagocytoses microbes and thereby resembles phagocytes of the innate immune system (Cosson and Soldati, 2008). Upon starvation, the amoebae aggregate and undergo a programmed differentiation and morphogenesis. One of the specific stages is the so-called slug, formed by the aggregation of about 100,000 amoebae, which will eventually differentiate into a fruiting body. Among slug cells, Sentinel cells represent a simple and efficient immune system. These phagocytes patrol the slug body to capture toxic compounds and invading bacteria. The Sentinel cells are continuously shed behind in the slime sheath of migrating slugs (Chen et al., 2007). These immune-like phagocytes have recently been reported to possess the capacity to produce extracellular DNA traps around the pathogen/foreign body (Zhang et al., 2016) in a way similar to phagocytes of vertebrates and invertebrates (Brinkmann et al., 2004; Robb et al., 2014). These features make of D. discoideum an even more attractive model to trace back the conserved functions of the innate immune system across evolution from protozoans to metazoans (Chen et al., 2007; Hagedorn et al., 2009; Zhang and Soldati, 2016).

Despite these promising results, the knowledge about the arsenal that D. discoideum employs to kill phagocytosed microbes and to combat potential pathogens is still scarce. The only example of an antimicrobial effector characterized at the protein level is the AlyA lysozyme, which had been isolated from amoebic extracts and found to be able to degrade bacterial cell walls (Müller et al., 2005).

According to the information derived from the genome project, D. discoideum possesses at least 15 genes potentially coding for lysozymes belonging to several different classes of these hydrolytic enzymes (Eichinger and Noegel, 2005; Müller et al., 2005). Beside the family of Alys, putative Entamoeba-type lysozymes, phage-type lysozymes, and C-type lysozymes (LyCs) can be identified in databases (Müller et al., 2005).

Another multifarious gene family of D. discoideum, which products may target bacterial membranes and thereby kill bacteria directly, is the one coding for saposin-like proteins (SAPLIPs). Structurally, SAPLIPs are characterized by four or five compactly packed alpha-helices, and are typically stabilized by three disulfide bonds built by a conserved array of six cysteine residues (Liepinsh et al., 1997). Functionally, SAPLIPs fulfill various biological functions, but the members of this family have in common that they interact with lipids and membranes (Munford et al., 1995; Bruhn, 2005; Kolter et al., 2005). SAPLIPS with antimicrobial activity can be found in phylogenetically diverse organisms ranging from protozoans to mammals (Leippe et al., 1991; Andersson et al., 1995; Peña et al., 1997).

In pathogenic amoebae, saposin-like proteins are well-known as pore-forming proteins that permeabilize the membranes of bacteria and human host cells (Leippe, 2014). The most comprehensively studied member of amoebic SAPLIPs is amoebapore A from E. histolytica, the tertiary structure revealed the characteristic SAPLIP fold (Hecht et al., 2004). We therefore provisionally termed the putative SAPLIPs of which D. discoideum Apls for amoebapore-like peptides.

Dictyostelium discoideum possesses 17 Apl genes that potentially can give rise to 33 SAPLIP peptides given that larger precursor proteins containing more than one SAPLIP domain (also termed saposin B domain) might be processed to release several mature SAPLIPs, as exemplified for the name-giving saposins (O'Brien and Kishimoto, 1991) and for naegleriapores of the free-living amoeba Naegleria fowleri (Herbst et al., 2004). Such an enormous variety of SAPLIPs in one species is only known so far from C. elegans, a nematode that also feeds on microbes (Roeder et al., 2010).

In D. discoideum, one may speculate that these proteins act complementarily and synergistically and constitute an important part of the antimicrobial armamentarium during its unicellular and multicellular stages. Nonetheless, functions other than killing of bacteria by membrane permeabilization have been reported for amoebic SAPLIPs (Michalek and Leippe, 2015).

In the present study, we have chosen AplD among the plethora of potential D. discoideum SAPLIPs for a more detailed functional study because it became apparent that: (i) the primary translation product comprises a single SAPLIP domain preceded by a putative signal peptide as known for amoebapores; (ii) the gene has been reported to be differentially expressed upon bacterial challenge; and (iii) most importantly with respect to immunity, aplD is reported to be primarily transcribed during the multicellular stages (Dicty express: https://dictyexpress.research. bcm.edu/bcm/#/all?genes=DDB\_G0293010).

At the protein level, we characterized the ability of AplD to permeabilize the membranes of live bacteria and liposomes by monitoring quantitatively the activity of a recombinantly expressed protein (rAplD). In vivo, we phenotypically analyzed the effect of ablation of aplD on amoebic growth on various bacterial lawns, on bacterial killing in the amoebic stage, and on the slug's capacity to fight a bacterial infection.

#### MATERIALS AND METHODS

#### Bacterial Cultures

Various bacterial strains were used in the study (Supplementary Table 1). The bacterial cultures were grown at 220 rpm at 37 ◦C for 12 to 14 h in Luria-Bertani medium.

#### *Dictyostelium discoideum* Cultures

Ax2 obtained from the Dicty stock center was used for generating the apl and lyC KO mutants. Ax2 amoebae were grown in maltose-HL5 medium and blasticidin (8 µg/ml) containing maltose-HL5 medium was used to select the KOs. The D. discoideum cells transfected with prestalk reporter (ecmAO-RFP), prespore reporter (pSA-RFP) plasmids, and aplD−/[act6]:aplD.FLAG cells [rescue strain; mentioned as aplD−(+)] were cultured in maltose-HL5 medium complemented with G418 (10 µg/ml) and blasticidin (8 µg/ml).

### Creation of KO Vectors

Short gene fragments from the 5′ and 3′ regions of apl and lyC genes were amplified by PCR and ligated appropriately at the 5 ′ and 3′ flanking ends of the blasticidin resistance (Bsr) gene present in the pLPBLP gene disruption vector as described in Faix et al. (2004). Before transfection, the KO vectors were subjected to DNA sequencing (Eurofins MWG operon, Germany) to verify vector orientations and mutations.

### Plasmids

pLPBLP, ecmAO-RFP, pSA-RFP, and pDneo2a-3xFLAG plasmids were received from the Dicty stock center.

### Targeted Gene Ablations in *D. discoideum*

Gene disruptions were achieved by homologous recombination between the respective apls and lyCs with their corresponding KO vectors. Ax2 cells (2 × 10<sup>7</sup> ) were sedimented at 560 × g at RT for 5 min. The pellet was resuspended in 1 ml ice-cold electroporation (EP) buffer (10 mM K2HPO4, 10 mM KH2PO4, 50 mM sucrose, pH 6.2.), washed at 10,000 × g at RT for 2 min, and briefly incubated on ice. The respective linearized KO vector was resolved in 100 µl ice-cold EP++ buffer (EP buffer containing 1 mM MgSO4, 1 mM NaHCO3, 1µM CaCl2, and 1 mM ATP, pH 6.2) and subsequently mixed with Ax2. Immediately, this cell mixture was transferred to 2-mm gap electroporation cuvette (BTX electroporation cuvettes, Harvard apparatus) and electroporation was carried out at 300 V, 2 ms time constant with five square pulses, including 5-s intervals inbetween the pulses, using a BTX 830 electroporator (Harvard apparatus). After 10 min incubation on ice, the cell suspension was transferred to Petri dishes that contain 10 ml maltose-HL5 medium and incubated at 22◦C. Blasticidin (8 µg/ml) selection was introduced 24 h later. Blasticidin-resistant clones had appeared after a week and were aspirated from the Petri dishes and seeded in 24-well plates (both from Sarstedt, Germany) for PCR analyses.

### Identification of KO Clones

Cell lysates were prepared from the blasticidin-resistant clones following the method described in Charette and Cosson (2004). As described in Faix et al. (2004) three different PCR analyses were performed to verify the apl and lyC KO vectors insertion in 5′–3′ orientation at the apl and lyC loci. PCR screening results for aplD KO clones are shown in the Supplementary Figure 1. KO vectors (vector control, VC) and Ax2 genomic DNA (wild-type, WT) were used as respective controls for PCR analyses. The KO clones were confirmed for gene ablation at the respective gene locus by performing Southern hybridization (Supplementary Figure 2), using the DIG-High prime DNA labeling and detection starter kit II (Roche), following manufacturer's instructions. A probe was generated for the Bsr-resistance gene sequence, which spans for 249 base pairs and dioxygenin (DIG) was used as the labeling agent.

#### Construction of an *aplD* Rescue Strain

The cDNA sequence of aplD was amplified by PCR and cloned under the Actin 6 promoter (act6) of pDneo2a-3xFLAG vector (Dubin and Nellen, 2010). The 3′ end of aplD cDNA was fused to the FLAG gene. Finally, the pDneo2a-3xFLAG[act6/aplD] plasmid was transfected into aplD<sup>−</sup> cells. AplD-FLAG fusion protein expression in the rescue strain [aplD−(+)] was confirmed by Western blot analysis (Supplementary Figure 3). After separation by SDS-PAGE, proteins were transferred onto a polyvynilidene fluoride membrane and the blot was incubated with a mouse anti-FLAG M2 monoclonal antibody (Sigma Life Science) and subsequently developed using a goat antimouse IgG (H+L) conjugated to alkaline phosphatase (Jackson ImmunoResearch).

### *AplD* Transcriptional Profiling

To analyse transcription in axenic cultures, RNA was isolated from axenically grown D. discoideum amoebae (5 × 10<sup>6</sup> cells) and the first strand cDNA was synthesized using the RNA template (cDNA synthesis kit, Invitrogen). Quantitative RT-PCR (qRT-PCR) was performed using the cDNA template as described by the manufacturer (qRT-PCR mix, TAKARA and Light cycler, Roche). For analysis of xenic cultures, Ax2 cells were mixed with bacteria (ratio 1:10) and the amoebae-bacteria cell suspensions were incubated at 140 rpm and 22◦C for 8 h. Subsequently, the xenic Ax2 cells were washed three times at 425 × g at 4◦C for 2 min to remove excess bacteria. cDNA was synthesized and qRT-PCR analyses were performed as described above. AplD transcription in axenic Ax2 amoebae was considered as reference to determine the aplD expression in xenic Ax2. Each symbol denotes individual experiment. The house keeping genes tested were Ig7 (rnlA) and Glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The bacterial strains tested include KpLM21, B. subtilis, and PT531 (see Supplementary Table 1). For analysis during development, axenic Ax2 cells were washed with Sörensen's buffer (SB) at 510 × g and 22◦C for 7 min. The cell pellets were resuspended in 100 µl SB and deposited on SB agar (1%) plates. Subsequently, the plates were incubated in a dark, moist chamber at 22◦C and RNA was isolated from the starving amoebae (0 h), streaming cells (8 h), mounds (12 h), slugs (16 h), and fruiting bodies (24 h) and qRT-PCR was performed in duplicates. AplD expression in the starving amoebae (0 h) was considered as reference to quantify the aplD expression at other stages of development (Sillo et al., 2008). Three independent development experiments were performed and the bars represent standard deviation.

### Growth of *Dictyostelium* on Bacterial Lawns

Bacterial growth assays were performed on routinely used SM agar medium, which was devoid of glucose. SM agar plates were prepared as described in Froquet et al. (2009). Varying number (10<sup>4</sup> , 10<sup>3</sup> , 10<sup>2</sup> , and 10<sup>1</sup> ) of D. discoideum amoebae were deposited on the bacterial lawn layered on SM agar plates and incubated in the dark at 22◦C. Five days later, the KOs were scored for their plaque forming abilities by having Ax2 as comparative controls. Three independent experiments were performed with triplicates (see Supplementary Table 1 for details on the bacterial strains tested).

### Intracellular Killing of Bacteria by *Dictyostelium*

Dictyostelium discoideum amoebae were mixed with bacteria (100:1) and incubated at 140 rpm, 22◦C. The total number of viable bacteria was measured at the indicated time points by following the method detailed in Benghezal et al. (2006). The rate of killing of K. pneumoniae bacteria by amoebae was quantified by counting the number of colony forming units (CFU) at each time point. Three independent experiments were performed.

### *Dictyostelium discoideum* Development

Dictyostelium discoideum cells were washed with SB and deposited on SB agar (1%) plates at a cell density 5 × 10<sup>5</sup> cells/cm<sup>2</sup> . Subsequently, plates were incubated in a dark, moist chamber at 22◦C. The morphogenetic stages were imaged under the stereomicroscope (Olympus ULWCD 0.30) at time points mentioned (Supplementary Figure 4). Dictyostelium discoideum development on SB agar was monitored in three independent experiments.

### Prestalk (pst) and Prespore (psp) Slug Patterns

Slug pattern analyses were performed as described in Parkinson et al. (2009). For prestalk reporter (ecmAO-RFP) examination, D. discoideum amoebae carrying the ecmAO-RFP plasmid were mixed with D. discoideum amoebae at 20:80 ratio and seeded on a SB agar (1%) plate at a cell density 5 × 10<sup>5</sup> cells/cm<sup>2</sup> . For prespore reporter (pSA-RFP) examination, D. discoideum amoebae (20%) were mixed with D. discoideum amoebae marked with pSA-RFP (80%) and deposited on SB agar plates. The plates were incubated in a dark, moist chamber at 22◦C until the migrating slugs were formed and were imaged under the stereomicroscope (Olympus ULWCD 0.30) (Supplementary Figure 4). Two independent experiments were performed in duplicates and at least ten slugs were imaged per plate.

#### Slugs Infection

Dictyostelium discoideum cells were harvested from Petri dishes, 2.5 × 10<sup>7</sup> cells were sedimented, and the pellets were resuspended in 50 µl SB. This high density cell suspension was deposited on SB agar (1%) plates and incubated in a dark, moist chamber with an unidirectional light source until migrating slugs were formed (∼18 h). Slug infection experiments were performed following Chen et al. (2007) with some modifications. The slugs were injured with a sterile needle (23G × 1 1/4 , B|BRAUN Injekt F) and a dense bacterial suspension, which was prepared from an overnight culture, was layered on the injured slugs. K. pneumoniae expressing GFP reporter (KpGFP) and E. coli expressing DsRed reporter were used for infecting the injured slugs (see Supplementary Table 1). The slugs infected with Ec DsRed were imaged under the stereo microscope (Olympus ULWCD 0.30) 8 h post infection and those slugs infected with KpGFP were imaged 20 h and 24 h post infection. KpGFP was grown with ampicillin (1 mg/ml) for 12 h at 37◦C. Ec DsRed was grown with isopropyl β-D-1-thiogalactopyranoside (IPTG, 100 mM) and kanamycin (150 µg/ml). Three independent experiments were performed and at least ten slugs were imaged per experiment for each strain [Ax2, aplD−, and aplD−(+)].

### Sentinel Cells Examination

Dictyostelium discoideum cells were allowed to form slugs on SB agar (1%) plates containing ethidium bromide (EtBr, 3 µg/ml) as described in Chen et al. (2007). After 3 h, Sentinel cells present in the migrating slugs were visualized under the stereomicroscope (Olympus ULWCD 0.30). Three independent experiments were performed and at least ten slugs were imaged per experiment for each strain (Ax2 and aplD−).

#### Recombinant Protein Production and Purification

The nucleotide sequence encoding putatively mature AplD (DDB0216216) was codon optimized for bacterial expression and synthesized by GeneArt (Regensburg, Germany). The cDNA was ligated into the pET-32a (+) (Novagen) expression vector containing an ampicillin resistance gene using KpnI and XhoI cleavage sites. The resulting plasmid encodes a fusion protein, which comprised an N-terminal thioredoxin followed by a hexahistidine (His6) tag and contained thrombin and tobacco etch virus (TEV) protease cleavage sites preceding the primary structure of AplD. TEV cleavage yielded a product with the amino acid sequence as follows: GEIDNNQCQICELLVKDIIE GLTANQSVEVIEHGLNLICDHIPLHVRVCKQFVDSNFQKIV QFIENHDDPQEICEKCGVC. The protein was recombinantly expressed in E. coli C 43 at 37◦C after induction with 0.5 mM IPTG for 6 h and extracted by sonication on ice (Sonoplus HD 2200 sonicator, MS-73 titanium microtip, Bandelin electronic GmbH, Germany). The fusion protein in the soluble fraction was purified by immobilized-metal affinity chromatography (IMAC) using a TALON resin (Clontech, Saint-Germain-en-Laye, France) and subsequent anion-exchange chromatography (1-ml Resource Q column, GE Healthcare) using an Äkta purifier system (model P-900; GE Healthcare) and 20 mM Tris-HCl, pH 7.0 with a continous linear gradient of 0-1 M NaCl for elution. The N-terminal tag was removed by cleavage with TEV protease (ProTEV Plus protease, Promega) at 30◦C overnight. The cleaved product was subjected again to IMAC and anion-exchange chromatography (Mini Q column 4.6/50, GE Healthcare, Solingen, Germany) to remove the fusion partner and the TEV from recombinant AplD. After final purification, the apparently homogeneous protein fraction was lyophilized, redissolved, and dialyzed against 10 mM sodium phosphate, pH 6.8.

Purity was proven by tricine-SDS polyacrylamid gel (SDS PAGE) using 13% separation gels (Schägger and von Jagow, 1987) and SeeBluePlus2 (Invitrogen) as standard protein marker for molecular masses. Protein concentration was determined with the BCA assay (Pierce, Thermo Scientific, Bonn, Germany). The molecular mass of recombinant AplD was determined by mass spectrometry in linear mode using a 4700 Proteomic Analyzer MALDI-TOF/TOF mass spectrometer (Life Technologies, Darmstadt, Germany).

### Assays for Membrane-Permeabilizing Activities

Pore-forming activity using the liposome depolarisation assay (Leippe et al., 1991) and permeabilization of cytoplasmic membranes of live bacteria using the fluorescent dye SYTOX Green (Herbst et al., 2002) were measured as described previously. Briefly, in the liposome-depolarization assay a valinomycin-induced diffusion potential across the membrane of liposomes prepared from asolectin, a crude mixture of soy bean phospholipids, resulted in quenching of the enclosed fluorescent dye. Application of a pore-forming protein disrupts the membrane potential and the fluorescence intensity increases by the release of the dye. An increase of 5% within 1 min after addition of the pore-forming agent at 25◦C is defined as one activity unit. In the membrane-permeabilization assay the fluorescent dye SYTOX Green (Invitrogen) is added to Bacillus megaterium, E. coli, Klebsiella aerogenes, K. pneumonia KpLM21, or K. pneumoniae K<sup>−</sup> as viable target giving a signal while intercalating in DNA. Accordingly, only when the bacterial plasma membrane become compromised, e.g., by an antimicrobial protein, fluorescence appears. The values were expressed as the mean of three independent experiments. Kinetics of the membrane permeabilization of B. megaterium were monitored in that fluorescence was measured for different doses at various time points and presented as a representative example from four replicates with similar results. The synthetic peptides alamethicin and cecropin P1 served as positive controls in the activity assays (Sigma, Taufkirchen, Germany).

#### RESULTS

#### The Multifarious Apl Family

In the genome of D. discoideum, we detected 17 genes that may code for Apl proteins containing a single or several SAPLIP domains (**Figure 1**). The sequences of apl A-R and the corresponding putative proteins can be retrieved from the D. discoideum database (www.dictybase.org) using apl<sup>∗</sup> as a query. Their primary translation products each contain a predicted signal peptide that may allow trafficking of the precursor proteins to lysosomal compartments. Proteolytic processing of larger precursors can result in 33 mature peptides with one SAPLIP domain. An N-terminal glycosylation site can be predicted for nearly all of them.

### *Apl* Mutants Were Defective for Growth on Virulent LM21 *K. pneumoniae*

The permissiveness of various bacteria for amoebic growth is routinely assessed by testing the ability of D. discoideum mutants to generate growth plaques on bacteria lawns. This strategy has led to the identification of several bacterial virulence genes (Cosson et al., 2002; Pukatzki et al., 2002; Benghezal et al., 2007; Alibaud et al., 2008). Here, we adapted the method to test the growth abilities of D. discoideum Ax2 strain in which genes potentially coding for antimicrobial proteins (apls and lyCs) have been disrupted. To quantitate the degree of bacterial virulence, varying number of D. discoideum cells were spotted on bacterial lawns and the growth plaques

were monitored after 5 days. It became apparent that aplD<sup>−</sup> was defective for growth on KpLM21, a clinical isolate and two apl mutants, aplD<sup>−</sup> and aplP−, were non-permissive for growth on K <sup>−</sup>, a capsule defective Kp strain (**Figure 2A**). In the case of Ax2, even 10 amoebae were sufficient to create growth plaques on K <sup>−</sup> and KpLM21, whereas at least 10,000 amoebae of aplD<sup>−</sup> and at least thousand amoebae of aplP<sup>−</sup> were required to generate visible plaques on K −. Likewise, KpLM21 allowed plaque formation only when 10,000 aplD<sup>−</sup> amoebae were deposited (**Figure 2A**). We also found that D. discoideum with a KO of the gene encoding the C-type lysozyme 2, lyC2−, was defective for growth on K − (Supplementary Figure 5). The growth abilities of Apls and lyCs mutants on bacteria such as B. subtilis, E. coli Br, non-pathogenic K. pneumoniae, and P. aeruginosa are summarized in a scheme (**Figure 2B**).

### Vegetative *aplD*<sup>−</sup> Amoebae Efficiently Kill *K. pneumoniae Kp*LM21

As described in Benghezal et al. (2006), we performed a specific assay to investigate whether aplD<sup>−</sup> amoebae are defective in phagocytosis and/or killing of KpLM21. These assays measure the total number of live bacteria remaining at indicated time points. We found that aplD<sup>−</sup> amoebae were able to phagocytose and kill Klebsiella aerogenes and KpLM21 as efficiently as the wildtype Ax2 (**Figure 3**). This indicates that AplD is not essential for intracellular killing of K. pneumoniae bacteria in free-living Dictyostelium amoebae.

### *AplD* Upregulation Upon Exposure to *Kp*LM21

To examine whether the expression of aplD is regulated by exposure to various bacteria, we measured the level of the aplD mRNA by qRT-PCR. Contact with KpLM21 increased aplD at least three-fold, but contact with Bacillus subtilis did not induce a significant change. In contrast, exposure to Pseudomonas aeruginosa PT531 downregulated aplD on average about 6-fold (**Figure 4A**). Analyses of aplD regulation during D. discoideum development on non-nutrient KK2 agar showed that aplD was weakly expressed in axenic conditions, but was strongly upregulated during late development stages, starting from the mound stage [12 h] and slug stage [16 h], and peaking at fruiting body stage [24 h] (**Figure 4B**). This profile is comparable to the aplD transcriptional regulation observed in Ax4 grown on bacteria and subjected to development on SB agar (Dicty express, Supplementary Figure 6). The transcriptional regulation of antimicrobial genes during the growth phase of D. discoideum might be essential to adapt to various food sources, but also combat various pathogens. It has been reported that amoebae stop feeding during development. Therefore, the main purpose for the aplD upregulation might be an extracellular role in cellular defenses inside the multicellular

structures or inside phagocytic Sentinel cells. Interestingly, upregulation of aplD during development, in particular at late stages, is several folds higher than during exposure to KpLM21.

### *AplD*<sup>−</sup> Slugs Are Vulnerable to *K. pneumoniae*

To test the idea that AplD is primarily instrumental in multicellular stages of D. discoideum, we infected slugs derived from aplD<sup>−</sup> amoebae with a K. pneumoniae (Kp) strain that expressed GFP. Unlike wild type slugs, at 20 h post Kp infection, aplDslugs showed bacterial deposits both on their surface and in their body. At 24 h post Kp infection, aplD<sup>−</sup> slugs showed Kp clumps mainly inside. Additionally, aplD<sup>−</sup> slugs were not successful in sloughing off Kp in their slime sheath (**Figure 5**). By contrast, aplD<sup>−</sup> slugs were able to shed E. coli in their slime sheath as early as 8 h post infection (**Figure 5**). As the major defense mechanism reported so far in D. discoideum slugs is based on the phagocytic nature of Sentinel cells, we examined whether aplD<sup>−</sup> slugs possess a functional population of these cells. As evidenced by the uptake of the fluid-phase tracer EtBr, aplDslugs were capable of generating Sentinel cells that internalized the toxic dye and were sloughed off during migration (**Figure 6**) and aplD<sup>−</sup> did not depict any prominent developmental defect (Supplementary Figure 4).

## *Klebsiella Pneumoniae* Susceptibility Was Rescued in *aplD*−(+) Amoebae and Slugs

As described above, aplD<sup>−</sup> amoebae were restricted for growth on K. pneumoniae, K <sup>−</sup>, and KpLM21, and aplDslugs were vulnerable to Kp infections. We tested whether restoration of aplD expression in aplD<sup>−</sup> would rescue the phenotypic defects. Constitutive expression of aplD in aplD<sup>−</sup> amoebae, termed aplD−(+), under the control of the act6 promoter of an extrachromosomal vector rescued the formation of plaques on K − but showed only partial rescue on KpLM21 lawns (**Figure 7**). Bacterial clearing was observed on KpLM21 lawns when 10<sup>4</sup> and 10<sup>3</sup> amoebae were deposited, but not with 10<sup>2</sup> and 10<sup>1</sup> amoebae. This observation might indicate that spatio-temporal regulation of aplD expression is crucial for full functionality. AplD−(+) slugs infected with Kp were able to clear off Kp by sloughing them off in their slime sheath. It appeared that the interior of aplD−(+) slugs was virtually free of Kp deposits and clumps (**Figure 5**).

### *In-Vitro* Functional Analysis of the Protein AplD

AplD was heterologously synthesized in E. coli as a fusion protein and subsequently the thioredoxin-His<sup>6</sup> tag was removed by proteolytic cleavage. The final product was purified to apparent homogeneity by repeated steps of IMAC and anionexchange chromatography (**Figure 8A**). Analysis by MALDI-TOF mass spectrometry confirmed the molecular identity of rAplD. It revealed an experimental average mass of 9,070.8 Da, which is in good agreement with the calculated molecular mass (9,070.3 Da) provided that six cysteine residues are involved in disulfide bonds. The recombinant protein was tested for its pore-forming and bacterial-membrane permeabilizing activities in vitro. AplD formed pores in liposomes composed of asolectin. At mildly acidic pH (5.2), AplD depolarized liposomal membranes with similar activity as the prototype of a poreforming peptide, alamethicin (**Figure 8B**). AplD acted in a pH-dependent manner. The protein displayed the highest poreforming activity at pH 4.4 and gave decreasing values with increasing pH (**Figure 8C**). In the Sytox-Green assay, performed exemplarily with B. megaterium, it became apparent that AplD is capable of permeabilizing the cytoplasmic membranes of

live bacteria. Membrane-permeabilizing activity increased with time. After 2 min, virtually no activity was detectable, but already after 5 min bacteria with compromised membranes appeared, particularly at higher protein concentrations. After 1 h, approximately 50% of the bacteria were permeabilized by AplD at 2.5 µM (**Figure 8D**). We could not detect bacterial membrane permeabilization with up to 5 µM AplD in the same assay when we used Gram-negative bacteria, K. pneumoniae LM21, K. pneumoniae K−, K. aerogenes, or E. coli, as target cells (data not shown) indicating that the outer membrane of these species constitutes an additional barrier for AplD for reaching the bacterial cytoplasmic membrane.

### DISCUSSION

Dictyostelium discoideum is a non-pathogenic, unicellular host used for bacterial and fungal risk assessment studies (Cosson et al., 2002; Alibaud et al., 2008; Koller et al., 2016) and the outcomes from such studies remain surprisingly similar to the observations made in animal hosts (Benghezal et al., 2006; Hagedorn et al., 2009). Although D. discoideum serves as a versatile system to explore bacterial virulence genes and their mechanisms of action, not much is known about the effector molecules of D. discoideum and their mode of action to defend against pathogens. In D. discoideum, host-pathogen interaction studies can be attempted both at the single-cell phagocyte stage and during the multicellular stages of its life cycle. Interestingly, the single-cell and multicellular phases are clearly separated and involve differential regulation of gene expression. During the growth phase, D. discoideum internalizes microbes by phagocytosis and the resulting phagosome fuses with lysosome, a compartment that houses hydrolytic enzymes and antimicrobial peptides, leading to efficient killing and degradation of the microbes to fulfill the nutritional need of the growing amoeba.

However, several intracellular pathogens have evolved mechanisms to block phago-lysosome fusion and to exploit the phagosomal compartment as their replication hub (Hagedorn et al., 2009; Shevchuk et al., 2014). During the multicellular stages of development, the antibacterial defense mechanisms are even more sophisticated. For example, in the slug formed mainly of non-phagocytic prespore and prestalk cells, about 1% of the cells, called the Sentinel cells retain their phagocytic ability. The foreign bodies and poisonous compounds that enter the slug are swallowed by these cells and finally get sloughed off the migrating slug and remain in the slime sheath, where they are entangled within extracellular DNA traps (Zhang and Soldati, 2016; Zhang et al., 2016). Mechanistically, this helps to restrict microbial infection and/or the accumulation of toxins. The present study reveals for the first time the specific involvement of the AplD gene product in antimicrobial defense during multicellular stages of development. In addition, aplD expression is differentially regulated by exposure to different bacteria, e.g., induced in the presence of K. pneumoniae and repressed in the presence of P. aeruginosa.

These in vivo evidences strongly suggest that Apl peptides are involved in antimicrobial defense, but a direct membranelysing activity had not been reported so far. Here, we show that purified, recombinant AplD displayed pore-forming activity in the minimalistic system of liposomes and also readily permeabilized the membranes of the live Gram-positive bacterium B. megaterium as monitored by the Sytox green assay. However, an activity against Gram-negative bacteria was not detected at the concentrations employed. One might speculate that AplD can also permeabilize Klebsiella, provided that the outer membrane does not hamper the access to the primary target, the bacterial plasma cell membrane. The antimicrobial efficacy of the relatively negatively charged AplD might be substantially enhanced by the synergistic action of other antimicrobial proteins, e.g., Apls with a more positive net charge,

aplD<sup>−</sup> slugs (E,G) at 20 and 24 h post infection are represented. KpGFP infections in Ax2 (B,D) and aplD<sup>−</sup> slugs (F,H) at 20 and 24 h post infection are shown under green fluorescence filter. KpGFP infection in Ax2 slug imaged 24 h post KpGFP infection (D) is also magnified. KpGFP infection on slug surface (arrowheads) and slug interior (arrow) and in the slime sheath of aplD<sup>−</sup> slug at 20 h post infection are magnified in the insets below. KpGFP clumps inside aplD<sup>−</sup> slug 24 h post infection (H) is (Continued)

FIGURE 5 | magnified in the adjacent panel. AplD<sup>−</sup> slugs sloughed off E. coli in their slime sheaths. DIC images of Ax2 and aplD<sup>−</sup> slugs (I,K) 8 h post infection with E. coli DsRed are represented. E. coli DsRed clumps (arrowheads) in Ax2 (J) and aplD<sup>−</sup> (L) slime sheaths (boxes) are shown under red fluorescence filter and are also magnified in the respective insets. AplD- (+) slugs sloughed off KpGFP in their slime sheaths. AplD−(+) slugs were infected with KpGFP as described earlier and were imaged 20 h post KpGFP infection. DIC images of aplD−(+) slugs are shown in (M,O,Q). K. pneumoniae infections in aplD−(+) slugs are shown under green fluorescence filter in (N,P,R). AplD−(+) slime sheaths showing KpGFP deposits (boxes) and clumps in the slug interior (arrows) at 20 h post infection are depicted. Scale bars, 100 µm.

or when other Dictyostelium factors that are not bactericidal per se but are capable of perturbing the outer membrane structure act in concert with AplD.

There are several examples of SAPLIPs including the saposins themselves that have been found to be glycosylated in their natural form. The vast majority of the potential peptides released from Apl precursors bear a potential N-glycosylation site. This also holds true for AplD. In amoebae, we have previously compared the pore-forming activity toward liposomes of glycosylated and deglycosylated naegleriapores and could not detect a substantial negative effect upon glycan removal (Herbst et al., 2002). However, we cannot exclude that a natural AplD, glycosylated specifically by D. discoideum, might have a different activity spectrum against natural targets than the recombinant unglycosylated version. A single glycosylation motif may determine a particular oligomeric structure, e.g., dimerization, to generate a more active protein, it may mediate an interaction with a partner molecule that helps to overcome the outer membrane of Gram-negative bacteria such as Klebsiella, or it may simply facilitate binding to bacterial targets, e.g., by shielding the negative charge of particular regions of the protein.

Apart from Apls, there are two other proteins that harbor a SAPLIP domain, i.e., acyloxyacyl hydrolase (AOAH) and countin (Ctn). The enzyme AOAH deacylates lipopolysaccharides in vertebrates, but has not been characterized in amoebae so far (Munford et al., 2009). Countin is a well-studied major component of a protein complex called counting factor, known to play a crucial role in aggregate size determination during development (Brock and Gomer, 1999). When recombinant countin was monitored for pore-forming activity in a liposomedepolarization assay, it exhibited only marginal activity (Gao et al., 2002). In our study, we found that aplD<sup>−</sup> amoebae showed mild defects at early development, such as stream breaks and developmental delay, but finally were able to form slugs and fruiting bodies, although they were smaller than from wild-type Ax2 cells.

We found that aplD<sup>−</sup> amoebae are specifically impaired for growth on one virulent strain of K. pneumoniae. Although aplD is poorly transcribed in the amoebic stage under axenic growth conditions, it appears as if its gene product is nevertheless instrumental to allow growth in the presence of some K. pneumoniae strains. The plethora of Klebsiella bacteria present over a longer period in the growth assay, different from the short exposure to K. pneumoniae in a killing assay, may change gene expression of several genes including aplD. More impressively, we found that aplD<sup>−</sup> slugs are immune-compromised specifically for K. pneumoniae. During that multicellular stage, we observed a dramatic increase of aplD transcript abundance. Notably, complementation of aplD in aplD<sup>−</sup> rescued the capacity of slugs to effectively capture and slough off K. pneumoniae in their slime sheath in a way similar to wild-type Ax2.

With respect to the evolution of innate immunity, the antimicrobial capacities of D. discoideum are still a relatively unexplored area of research. On the one hand, the amoeba is an interesting model because it resembles in its unicellular stage mammalian macrophages in several aspects (Bozzaro et al., 2008; Cosson and Soldati, 2008; Dunn et al., 2018). On the other hand, the developmental cycle of the social amoeba represents an intriguing example in biology of appearance of multicellularity before the advent of true metazoans. In particular, the slug, which is composed of about 100,000 prespore and prestalk cells, hosts about 1% of Sentinel cells. These phagocytic cells reminiscent of patrolling neutrophils and tissue macrophages of animals, represent an extraordinarily appealing model system for comparative immunologists, as they defend the slug against infection using bactericidal phagocytosis and DNA extracellular traps (Chen et al., 2007; Zhang et al., 2016). Here, we reveal the involvement of an amoebic SAPLIP in the struggle of the slug stage against virulent bacteria, as an early example of a molecular defense implemented at the transition from unicellular to multicellular organisms.

FIGURE 7 | AplD−(+) strain growth on K. pneumoniae. Growth abilities of aplD−(+) cells were tested on K. pneumoniae KpLM21 and <sup>K</sup> <sup>−</sup> (A) and on K. pneumoniae Kp52145 and Ka <sup>13883</sup> (B). Growth abilities of Ax2 and aplD<sup>−</sup> cells are also represented for comparison. Scale bar, 1.5 cm.

FIGURE 8 | Purification and functional in-vitro analysis of recombinant AplD. (A) Purification steps of AplD after recombinant synthesis evidenced by SDS PAGE and Coomassie staining of the gel: Extracts of E. coli C43 transformed with aplD in pET-32a (+) in TBS, pH 7.0 (start), enrichment with IMAC, further purification with anion-exchange chromatography (Resource Q), proteolytic cleavage with TEV, and final purification via a Mini Q column. Purity of AplD after lyophylisation and dialysis is also shown at the right. All samples were reduced and alkylated before separation. At the very left, marker proteins (M) were separated and the molecular masses of these are depicted in kDa. (B) Time course of pore formation induced by AplD. The dissipation of a valinomycin-induced diffusion potential in vesicles of asolectin after addition of AplD (80 pmol) (trace 1), control peptide alamethicin (100 pmol) (trace 2), and the peptide solvent (trace 3) were recorded at pH 5.2. Pore-forming activity is reflected by the increase of fluorescence as a function of time. (C) pH dependence of pore-forming activity of AplD. Activity of AplD in comparison with that of alamethicin was measured in six independent experiments in Tris-maleate buffer (50 mM Tris-maleate/50 mM sodium sulfate/0.5 mM EDTA/0.02 % sodium azide) adjusted to various pH. Mean and standard deviation are shown. (D) Kinetics of membrane permeabilization by AplD. Membrane permeabilization of viable B. megaterium was measured as an increase in fluorescence of the DNA intercalating dye SYTOX Green at pH 5.2 induced by AplD and monitored after different time points at various concentrations.

## AUTHOR CONTRIBUTIONS

Conceived the study: ML. Designed the experiments: RD, PC, TS, and ML. Performed the experiments: RD and MB. Analyzed the data: RD, MB, PC, TS, and ML. Wrote the paper: RD, PC, TS, and ML.

#### ACKNOWLEDGMENTS

This study was supported by the German Research Foundation (DFG); CRC 1182 Function and Origin of Metaorganisms (project A1). TS is a member of iGE3 (http://www.ige3.unige.ch); is supported by an RTD grant from SystemsX.ch, and by multiple grants from the Swiss National Science Foundation. The authors thank the Dicty

### REFERENCES


stock center for providing the plasmids/strains used in this study. RD is supported by DST, INDIA. RD thanks Baskar Ramamurthy for providing laboratory space to conduct the rescue experiments. The authors thank Heidrun Lieβegang for technical assistance in the activity assays with the recombinant protein, Judith Bossen for performing Southern hybridization, Christoph Gelhaus for mass spectrometry measurements and Rosa Herbst for creating **Figure 1**.

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb. 2018.00073/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 Dhakshinamoorthy, Bitzhenner, Cosson, Soldati and Leippe. 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.

digital media

of impactful research

article's readership