Quercetin Prevents Escherichia coli O157:H7 Adhesion to Epithelial Cells via Suppressing Focal Adhesions

The attachment of Escherichia coli O157:H7 to intestinal epithelial cells is indispensable for its pathogenesis. Besides translocated-intimin receptor (Tir), E. coli O157:H7 interacts with host cell surface receptors to promote intimate adhesion. This study showed that integrin β1 was increased in Caco-2 cells upon E. coli O157:H7 infection, while Caco-2 cells subjected to integrin β1 antibody blocking or CRISPR/Cas9 knockout had reduced bacterial attachment. Infection of E. coli O157:H7 inactivated focal adhesion kinase (FAK) and paxillin, increased focal adhesion (FA) and actin polymerization, and decreased cell migration in Caco-2 cells, which were rescued by integrin β1 antibody blocking or knockout. Pre-treatment with quercetin, known for its anti-oxidant and anti-inflammatory activity, reduced bacterial infection to Caco-2 cells, which might be partially via interfering integrin β1 and FAK association augmented by E. coli O157:H7. In addition, quercetin decreased FA formation induced by bacterial infection and recovered host cell motility. Taken together, data showed that E. coli O157:H7 interacts with integrin β1 to facilitate its adhesion to host cells. Quercetin inhibits bacterial infection possibly by blocking the interaction between E. coli O157:H7 and integrin β1. Collectively, these data indicate that quercetin provides an alternative antimicrobial to mitigate and control E. coli O157:H7 intestinal infection, and suggest potential broad benefits of quercetin and related polyphenols in fighting other enteric pathogen infections.


INTRODUCTION
Formation intestinal attaching and effacing (A/E) lesions is of necessary for the pathogenesis of Escherichia coli O157:H7 (Kaper, 2005). After attachment to intestinal epithelial cells, E. coli O157:H7 induces actin rearrangement to form pedestals (Knutton et al., 1989). Through this tight association with the host cell surface, E. coli O157:H7 utilizes various strategies to manipulate host signaling, leading to enhanced bacterial colonization and persistence, and host tissue damage (Xue et al., 2017). The host extracellular matrix (ECM) is composed of multiple macromolecules, which mediate multiple biological functions including cell to cell adhesion, migration, proliferation, and death (Meredith et al., 1993). Integrin β1, the most abundant cell surface integrin, is a transmembrane glycoprotein receptor that interacts with ECM components such as fibronectin, laminin, and collagen. Through interactions with ECM components, integrin β1 induces multiple bidirectional signal exchanges (Schwartz et al., 1995;Burridge and Chrzanowska-Wodnicka, 1996). In addition, integrin β1 recruits intracellular proteins such as talin, paxillin, and α-actinin, leading to the formation of the focal adhesion (FA) complex.
To tightly associate with host cells, pathogens utilize integrin β1 as an adhesion factor. Yersinia pseudotuberculosis interacts with integrin β1 via adhesin YadA to promote tight binding to the host cells (Eitel et al., 2005). Neisseria gonorrhoeae attaches to ECM substrate with the assistance of host integrin β1 (Muenzner et al., 2005). In response to infection, the rapid turnover and exfoliation of epithelial cells are innate defense mechanisms against pathogens (Mulvey et al., 2000). However, many pathogenic bacteria can circumvent host exfoliation and colonize the epithelium efficiently. Shigella flexneri reduces adhesion complex turnover and suppresses the detachment of infected cells from the basement membrane to manipulate host exfoliation (Kim et al., 2009). Integrins transduce extracellular signals into the host cells through association with intracellular adaptor proteins and protein kinases such as focal adhesion kinase (FAK) (Dia and Gonzalez de Mejia, 2011) and integrinlinked kinase (ILK) (Gagne et al., 2010). FAK deficiency increases the recruitment of FAs and reduces cell motility (Ilic et al., 1995), indicating FAK is involved in FA formation during cell migration. Thus, pathogens may manipulate FAK and associated kinases, which stabilize the FAs and ultimately enable them to colonize the host cells.
Quercetin is a polyphenol widely found in vegetables and fruits. Our previous study demonstrated that quercetin had anti-inflammatory and anti-oxidative properties that prevented E. coli O157:H7-induced inflammasome activation (Xue et al., 2017). However, the antimicrobial mechanism of quercetin has not been elucidated. We hypothesized that E. coli O157:H7 attaches to host cells via interacting with host integrin β1 and stabilizing FAs formation; quercetin inhibits integrin β1 expression and FA formation thus preventing E. coli O157:H7 infection.

Cell Line, Media and Bacterial Strains
The human colonic epithelial cell line Caco-2 was obtained from the American Type Culture Collection (Manassas, VA, United States). Caco-2 cells were cultured in Dulbecco's Modified Eagle's medium (DMEM) (Sigma, St. Louis, MO, United States) supplemented with 10% fetal bovine serum (Sigma), 100 units/ml penicillin G, and 100 µg/ml of streptomycin (Sigma) at 37 • C with 5% CO 2 . The E. coli O157:H7 EDL933 wild type (EDL933) strain was obtained from the STEC center at Michigan State University. The E. coli O157:H7 EDL933 intimin ( eae) and tir ( tir) mutant strains were kindly provided by Dr. Carolyn H. Bohach's Lab at the University of Idaho. pEHEC tir plasmid was a generous gift from Dr. John M Leong at Tufts University (Campellone et al., 2002). EDL933 tir pEHEC tir strain was derived from E. coli O157:H7 EDL933 tir strain transformed with pEHEC tir plasmid. These strains were routinely grown in LB broth at 37 • C overnight with aeration.

Infection of E. coli O157:H7 to Colonic Epithelial Cells
Caco-2 cells were seeded in a 24-well plate at 5 × 10 5 cells/ml for 12 h. Then the growth medium was replaced with fresh DMEM complete medium without antibiotics and supplemented with or without 200 µM quercetin (Sigma) for 12 h. Quercetin at this concentration did not impact the viability and growth of E. coli O157:H7 EDL933 (Supplementary Figure S1), nor did it decrease cell viability of Caco-2 cells (Xue et al., 2017). For integrin β1 blocking assay, cell monolayers were pretreated with integrin β1 antibody (rat IgG1, monoclonal, 1:200 dilution, DSHB) for 1 h prior infection, followed by 3 washes with PBS (pH 7.4). Then the cells were challenged with E. coli O157:H7 EDL933 at multiplicity of infection (MOI) of 10 for 4 h at 37 • C with 5% CO 2 .

Quantitative Reverse Transcription PCR (qRT-PCR) Analysis
Total RNA was extracted from Caco-2 cells with TRI Reagent (Sigma) and reverse transcribed using an iScript kit (Bio-Rad, Hercules, California). cDNAs were used as templates for qRT-PCR analysis of selected genes using a CFX96 Real-Time PCR Detection System (Bio-Rad). SYBR green master mix (Bio-Rad) was used for all qRT-PCR reactions. β-actin was used as the housekeeping gene. Primers for qRT-PCR are listed in Supplementary Table S1. Amplification efficiency was 0.90 to 0.99 (Xue et al., 2017).
Adhesion of E. coli O157:H7 to Colonic Epithelial Cells Escherichia coli O157:H7 attachment to Caco-2 cells was conducted as previously reported (Xue et al., 2017). Briefly, Caco-2 cells were seeded at 5 × 10 5 cells/ml in a 24-well plate, cultured until 80∼90% confluence and treated with 0 or 200 µM quercetin for 12 h. The cell monolayers were next challenged with E. coli O157:H7 EDL933 strain (MOI = 10) and co-cultured at 37 • C with 5% CO 2 for 4 h, followed by 3 washes with ice cold PBS and lysed with 0.2% Triton X-100. Lysates were serially diluted and appropriate dilutions were plated on LB agar plates. (E) Co-immunoprecipitation analysis of integrin β1 and FAK interaction. Input, the whole cell lysate detected with β-actin; IP, lysates post integrin β1 immunoprecipitation were detected with integrin β1 or FAK antibody. Caco-2 cells were co-cultured with EDL933 strain for 4 h before respective analyses. Means ± SEM; n = 4. * * P ≤ 0.01.
The bacterial colonies were counted after 24 h incubation at 37 • C.

Immunofluorescent Staining
Cell culture, quercetin treatment, and infection procedure were conducted as described above. Post-infection, the cell monolayers were washed 3 times with ice cold PBS and fixed in fresh prepared 4% paraformaldehyde for 30 min at room temperature. The fixed cells were then permeabilized with 0.5% Triton X-100 for 10 min, washed with PBS, and blocked with 5% normal goat serum for 60 min at room temperature (RT). Then the cells were incubated with anti-integrin β1 antibody (rat monoclonal IgG1, DSHB), vinculin antibody (Santa Cruz) or phalloidin (Sigma) overnight at 4 • C. The cells were rinsed with PBS and stained with Alexa Fluor 555 goat anti-rat IgG or Alexa Fluor 488 goat anti-mouse IgG (Cell Signaling) for 60 min at RT. These stained cells were washed 3 times with PBS and mounted with Fluoro-gel with DAPI (Electron Microscopy Sciences, Hatfield, PA). Fluorescence signal was visualized with EVOS FL fluorescence microscope (Life Technologies, Grand Island, NY).

Co-immunoprecipitation
The post-infection cell monolayers were washed twice with ice-cold PBS and lysed in 200 µl IP buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% (v/v) Triton X-100, 0.1% (w/v) Na-deoxycholate, 1 mM EDTA, proteinase inhibitor cocktail) for 15 min on ice. The resulting cell lysates were transferred into pre-cooled 1.5 ml tubes, passed through a 29-gauge needle twice, and centrifuged for 10 min at 14,000 g, 4 • C. An aliquot of the supernatant was sampled for input protein content analysis. The remaining supernatants were pre-cleared with Protein G agarose beads (Thermo Scientific) with rotation for 30 min at 4 • C. The pre-cleared supernatants were incubated with anti-integrin β1 antibody (rat monoclonal IgG1, 1:100, DSHB) overnight with rotation at 4 • C. Then the Protein G magnetic agarose was added into the tubes and co-incubated overnight at 4 • C with rotation. The next day, tubes were placed on the magnetic stand to collect beads. The beads were washed with IP buffer 5 times, then resuspended in 100 µl of loading buffer and heated to 100 • C for 10 min to elute proteins. The supernatants collected were used for immunoblotting with anti-intimin-γ antibody (Gift from Dr. John M Leong) or anti-FAK antibody (Cell signaling), respectively.

Cell Migration Activity
Cell culture, quercetin treatment, and infection procedure were conducted as described above. A scratch was introduced to the Caco-2 cell monolayer using a pipet tip. Then the cells were washed with PBS and infected with EDL933 strain or left uninfected for 4 h. Cells were washed with PBS and replaced with DMEM complete medium. Cells were migrated into the wound at 37 • C for 24 h. The migration was assessed by counting the number of Caco-2 cells that crossed the wound border as published previously (Kung et al., 2008).

Statistical Analyses
Statistical analyses were conducted as previously described (Xue et al., 2017). Data were analyzed as a complete randomized design using GLM (General Linear Model of Statistical Analysis System, SAS, 2000). All data were analyzed by two-tailed Student's t-test. Means ± standard errors of mean (SEM) are reported. Statistical significance is considered as P ≤ 0.05.

RESULTS
Integrin β1 Was Involved in E. coli O157:H7 Attachment Integrin β1 was expressed higher in infected cells than in control cells ( Figure 1A). E. coli O157:H7 infection also increased surface level of integrin β1 ( Figure 1B) as well as integrin α5 mRNA expression (Supplementary Figure S2). Neutralizing integrin β1 with anti-integrin β1 antibody reduced bacterial adhesion to Caco-2 cells (Figure 1C). To further explore the role of integrin β1 in bacterial adhesion, integrin β1 was knocked out with ITGB1 CRISPR/Cas9 sgRNA plasmid, which significantly attenuated EDL933 adherence to Caco-2 cells ( Figure 1D). Integrin clustering is reported to be associated with FAK activation (Guan, 1997). Immunoprecipitation assay further showed FAK protein was associated with integrin β1 in Caco-2 cells infected with E. coli O157:H7, suggesting that infection induced FAK recruitment by integrin β1 (Figure 1E).
Integrin β1 Increased FA and Actin Polymerization in Response to E. coli O157:H7 Infection FA is responsible for cell adhesion and migration (Hu et al., 2014). Enhanced FA assembly reduces cell mobility (Wozniak et al., 2004). E. coli O157:H7 infection increased FA proteins including talin, vinculin, and α-actinin in Caco-2 cells. However, integrin β1 antibody blocking or KO reduced the levels of these proteins in infected cells (Figures 3A-D).
Immunofluorescence staining further showed that vinculin content was increased during E. coli O157:H7 infection, while integrin β1 KO impaired the accumulation of vinculin in response to infection ( Figure 3E). These data collectively showed that integrin β1 was an important factor that mediated host FAs recruitment and assembly in response to E. coli O157:H7 infection.
The assembly of integrins and FAs serve as a platform for the organization of actin filaments. E. coli O157:H7 attachment to host cells is typically associated with actin rearrangement.  When integrin β1 was KO or blocked by antibody, the actin polymerization induced by infection was subsided (Figure 4), showing that integrin β1 was also implicated in infection-induced actin polymerization.
Enhanced FA assembly and decreased FAK activation could lower the ability of cell migration (Sieg et al., 1999;Kim et al., 2009). Consistent with enhanced FA assembly, E. coli O157:H7 infection significantly inhibited cell migration during wound healing. This inhibition phenomenon was attenuated by integrin β1 KO or antibody blocking (Figure 5).

Intimin Is Involved in FAK Inhibition and FA Accumulation
Immunoprecipitation assay indicated that intimin was associated with integrin β1 (Figure 6A). To further understand the role of intimin in host FA formation, we infected Caco-2 cells with intimin mutant strain ( eae). Infection with eae strain did not suppress FAK and paxillin (Figures 6B-D), indicating a regulatory role of intimin in host signaling transduction. Consistently, FA proteins including talin, vinculin and α-actinin were not altered in cells infected with eae strain (Figures 6E-H). Immunofluorescent staining further showed that eae resulted in a lesser accumulation of vinculin as compared with EDL933 WT infected cells ( Figure 6I). Interestingly, our data also showed that the tir deletion mutant ( tir) strain was incapable of causing dephosphorylation of FAK and paxillin (Supplementary Figure S3). The cytoplasmic C and N-terminus of Tir bind to FA proteins such as talin, vinculin, and α-actinin (Freeman et al., 2000;Huang et al., 2002), which might interfere with FAK activity. The interaction between Tir and host FA may strengthen its association with host cell surface and facilitate colonization.

DISCUSSION
Integrin β1 Is a Potential Receptor for E. coli O157:H7 Adhesion Integrins are a large family of heterodimeric receptors that are associated with a wide range of cell-to-cell interactions (Hynes, 1992). Integrin α5β1 is the most expressed and best characterized integrin heterodimer and functions as a receptor for many bacteria, such as Shigella flexneri and Pseudomonas aeruginosa (Watarai et al., 1996;Roger et al., 1999). The adhesin protein, CagL of Helicobacter pylori binds to and activates integrin α5β1 receptor and induces intracellular signaling (Kwok et al., 2007). Notably, many pathogenic bacteria enhance the surface level of integrins. H. pylori-infected gastric epithelial cells have a higher expression of both integrin α5 and β1 (Cho et al., 2006), and S. flexneri infection increases integrin β1 in HeLa cells (Kim et al., 2009). Consistently, our data also showed that both integrin α5 and β1 were upregulated in E. coli O157:H7-infected cells as compared to non-infected cells. Integrin β1 KO or blocking by integrin β1 antibody decreased bacterial attachment, indicating that integrin β1 was involved in E. coli O157:H7 adhesion.

Inhibition of FAK May Strengthen Bacterial Colonization
Accumulating evidence shows that virulence factors of pathogens can utilize host kinases to manipulate host signaling. OspE, an effector of type III secretion system (T3SS) of Shigella (Miura et al., 2006), interacts ILK and subsequently reduces the phosphorylation of FAK and paxillin (Kim et al., 2009), resulting in stabilization of FAs and attenuated cell turnover (Miura et al., 2006). EspO1-1, a homolog of OspE in E. coli O157:H7 (Kim et al., 2009;Morita-Ishihara et al., 2013), similarly interacts with FAK to stabilize FA complex and inhibit the detachment of host cells from the ECM (Morita-Ishihara et al., 2013), indicating E. coli O157:H7 also has the ability to counteract the exfoliation of epithelial cells, which benefits its persistence. In our study, we found that intimin was co-immunoprecipitated with integrin β1, while intimin mutant strain was unable to induce FAK and paxillin dephosphorylation, suggesting that intimin mediates FAK and FA activity, and has ability to interact with integrin β1 to exploit host outside-in signaling.
Integrins transduce extracellular signals into the host cells through association with intracellular adaptor proteins and protein kinases such as FAK (Dia and Gonzalez de Mejia, 2011) and ILK (Gagne et al., 2010). These kinases serve as docking sites for recruitment of other kinases and FA components such as paxillin, talin, vinculin, and mediate cytoskeletal reorganization . FAK activation induces disassembly of FAs and correlates with enhanced cell turnover (Webb et al., 2004;Hamadi et al., 2010). In FAK −/− cells, the disassembly of FAs is significantly impaired with attenuated cell mobility (Webb et al., 2004). We found that infection enhanced interaction between FAK and integrin β1, which inhibits the phosphorylation of FAK and subsequently deactivates paxillin, thereby causing FA accumulation (Figure 9). As a result, cell migration was reduced in response to E. coli O157:H7 infection, which may inhibit host shedding and turnover. In support of our finding, FAK activation promotes migration of both endothelial cells and fibroblasts (Zhao and Guan, 2011), while FAK deficiency decreases cell migratory activity (Zhao and Guan, 2011) with an increased formation of FAs (Ilic et al., 1995).

Quercetin Decreases Bacterial Infection by Regulating Integrin β1
Quercetin decreases ECM components such as collagen III productions and assembly in human corneal fibroblasts (McKay et al., 2015), and decreases cell surface level of integrin β1 in different cell types (He et al., 2015;Doersch and Newell-Rogers, 2017). In this study, although quercetin did not alter integrin β1 expression in uninfected cells, quercetin prevented the increase of both integrin β1 and integrin α5 expressions, as well as FA protein assembly induced by infection. Mechanisms for such preventive effects are twofold. Quercetin could directly interfere with integrin signaling elicited by bacteria to suppress FA accumulation and bacterial attachment, or the reduced bacterial attachment due to quercetin proportionally weakened intracellular signaling in comparison to untreated cells with more bacterial attachment. These data collectively suggested that quercetin prevented E. coli O157:H7 adhesion to epithelial cells through attenuation of integrin β1 accessibility to bacteria and/or suppression of intracellular signaling. The resultant effect may contribute to the reduced FA assembly.
In summary, E. coli O157:H7 attached to epithelial cells partially through the interaction with host integrin β1, which inhibited FAK phosphorylation and stabilized FA formation.
Quercetin inhibits bacterial infection likely via attenuated association between integrin β1 and FAK. Given that antibiotics are not applicable for E. coli O157:H7 infection, these data provide a potential therapeutic application of quercetin for minimizing and eliminating E. coli O157:H7 infection. These data also suggest a broad application of polyphenolic compounds in the prevention of enteric pathogenic infection. However, additional in vivo studies to test the effects of quercetin on inhibiting E. coli O157:H7 infection will further strengthen our conclusions.

AUTHOR CONTRIBUTIONS
YX, MD, and M-JZ designed the study, analyzed the data, and reviewed the manuscript. YX conducted the experiments. YX drafted the manuscript. MD and M-JZ revised the manuscript.