ABSTRACT
The twin-arginine translocation (Tat) system is involved in not only a wide array of cellular processes but also pathogenesis in many bacterial pathogens; thus, this system is expected to become a novel therapeutic target to treat infections. To the best of our knowledge, involvement of the Tat system has not been reported in the gut infection caused by Citrobacter rodentium. Here, we studied the role of Tat in C. rodentium gut infection, which resembles human infection with enterohemorrhagic Escherichia coli (EHEC) and enteropathogenic E. coli (EPEC). A C. rodentium Tat loss-of-function mutant displayed prolonged gut colonization, which was explained by reduced inflammatory responses and, particularly, neutrophil infiltration. Further, the Tat mutant had colonization defects upon coinfection with the wild-type strain of C. rodentium. The Tat mutant also became hypersensitive to bile acids, and an increase in fecal bile acids fostered C. rodentium clearance from the gut lumen. Finally, we show that the chain form of C. rodentium cells, induced by a Tat-dependent cell division defect, exhibits impaired resistance to bile acids. Our findings indicate that the Tat system is involved in gut colonization by C. rodentium, which is associated with neutrophil infiltration and resistance to bile acids. Interventions that target the Tat system, as well as luminal bile acids, might thus be promising therapeutic strategies to treat human EHEC and EPEC infections.
INTRODUCTION
Enterohemorrhagic Escherichia coli (EHEC) and enteropathogenic E. coli (EPEC) are enteric pathogens that colonize the colon and cause diarrhea in humans. Inflammatory bowel diseases inflicted by these pathogens result in mortality and morbidity worldwide. Citrobacter rodentium is an enteric murine pathogen that infects mice through infectious strategies that are similar to those employed by EHEC and EPEC (1, 2). Thus, C. rodentium is widely used to study human diarrheal infection caused by EHEC and EPEC (1, 3). In C57BL/6 mice, C. rodentium infection is self-limiting, and pathogen clearance occurs without antibiotic treatment. Further, colonic colonization by C. rodentium and shedding peak at 5 to 12 days postinfection (dpi). At the later stages of infection, bacterial strains are finally eliminated from the gut, which is mainly mediated by robust inflammatory responses, including the recruitment of immune cells such as neutrophils, Th22 and Th17 CD4+ T cells, and B cells (4, 5). Furthermore, immunity involves both innate and adaptive lymphocytes, which is characterized by collaboration between the epithelial barrier and the aforementioned immune cells (6). However, a complete understanding of the molecular mechanisms that govern pathogen elimination from the gut lumen is still lacking.
A specialized export system is required for proteins that are localized partially or completely outside the bacterial cytoplasm. Two representative export mechanisms, the secretory (Sec) and twin-arginine translocation (Tat) systems, transport these proteins outside the cytoplasm. The former is the major export system that transports unfolded proteins, whereas the latter is an additional system that can export proteins in a folded conformation. Indeed, mutations in sec confer a conditional lethal phenotype due to severe protein export defects (7–9). In contrast, Tat is also involved in a variety of cellular processes, especially anaerobic respiration, iron uptake, copper resistance, and cell wall metabolism (10). In E. coli and Salmonella spp., TatA, TatB, TatC, and TatE comprise the Tat export apparatus. Each protein consists of an inner membrane-spanning component, forming an oligomeric complex through which Tat substrates can cross the membrane. In particular, TatC is essential for protein transport via the Tat system (11). Many bacterial pathogens possess a Tat system, and remarkably, it has been shown that Tat is required for the full virulence of EHEC (12), Legionella pneumophila (13, 14), Pseudomonas syringae (15), Mycobacterium tuberculosis (16), Yersinia pseudotuberculosis (17), Vibrio cholerae (18), Campylobacter jejuni (19), and Salmonella enterica (20–23). Notably, since it is not found in humans or any known mammalian cells, the Tat system could be a promising target for the development of novel agents to regulate bacterial infections.
Although few reports have indicated that the Tat system is involved in gut infections caused by enteropathogenic bacteria (18, 20, 22, 23), its involvement in intestinal infection caused by C. rodentium remains unknown. This paper describes the role of the Tat system in C. rodentium infection. Here, we found that it is involved in murine gut colonization, which is associated with inflammatory responses and resistance to bile acids. The Tat system might thus be a promising therapeutic target to treat human infections with EHEC and EPEC.
RESULTS
The Tat export system is functional in C. rodentium, and mutation of the tatC gene confers the export defect.To study the role of the Tat system in C. rodentium virulence, an isogenic deletion mutant of the tatC gene, encoding the Tat export apparatus, was constructed as described in Materials and Methods. To investigate Tat functionality, the C. rodentium strain DBS100 was transformed with plasmid pZ300 (Table 1), encoding a green fluorescent protein (GFP) reporter fused with the Tat signal peptide from TorA of Salmonella enterica serovar Typhimurium. In agreement with a previous report (24), cytoplasmic expression and the specific periplasmic export of the TorA::GFP fusion protein could be observed (Fig. 1A). Next, to confirm that the deletion of the tatC gene confers a Tat export defect, the plasmid pZ300 was introduced into the tatC mutant strain of C. rodentium. In tatC mutant cells, GFP fluorescence was localized to the cytoperiplasmic space but was not observed at the periphery of cells (Fig. 1A). These results suggest that C. rodentium indeed possesses a functional Tat system and that the TatC protein encoded by the tatC gene is essential for its function.
Strains and plasmids used in this study
Tat system is functional in C. rodentium. (A) Tat-dependent export of TorA::GFP fusion protein in C. rodentium. Fluorescence images of the C. rodentium wild-type (wt) strain and tatC mutant harboring plasmid pZ300 encoding the signal peptide from TorA in S. Typhimurium fused to GFP. Bottom images show a zoomed-in view of the boxed areas in the panels on the top row. Scale bar = 5 μm. (B) Light micrographic images of indicated C. rodentium strains. Scale bar = 5 μm. (C) Quantitative analyses of the experiment shown in panel B. Bars represent the mean ± standard deviation (SD) from three independent experiments. ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001; unpaired Student's t test. (D) Growth curves of the indicated C. rodentium strains in M9 minimum medium and DMEM. Data points represent the mean ± SD from the results from three independent experiments. OD, optical density.
It is known that Tat mutant Gram-negative bacteria, such as E. coli, Salmonella enterica serovar Enteritidis, and S. Typhimurium, form long chains due to cell division defects (20, 22, 23, 25). Similar to that, in these bacteria, tatC mutant C. rodentium cells generated chains (Fig. 1A and B). Almost all cells of the wild-type strain were single or paired, and no chains were observed (Fig. 1C), whereas approximately 60% of the tatC mutant cells formed chains (Fig. 1C). This increased proportion was restored by the introduction of a plasmid carrying wild-type tatC but not by the parent plasmid. These results indicate that the Tat system is involved in cell division in C. rodentium.
Next, we checked whether the cell division defect could interfere with bacterial growth and colony formation on agar plates. Growth curves based on optical density and CFU per milliliter with wild-type, tatC mutant, and complementation strains suggested that a mutation in the tatC gene does not affect bacterial growth and colony formation (Fig. 1D). This indicated that direct plating on media in this study was suitable to determine and compare C. rodentium numbers between wild-type and tatC mutant strains.
A mutation in tatC prolongs C. rodentium gut colonization.To assess whether Tat of C. rodentium is required for virulence, we next compared bacterial fitness in the gut between the wild-type and tatC mutant strains. C57BL/6 mice were infected intragastrically with 1 × 109 CFU of the wild-type strain harboring plasmid pJB861 or the tatC mutant. At the indicated days postinfection, bacterial loads in feces were determined. At 5 dpi, tatC mutant loads were lower than those of the wild-type strain, whereas at 7 dpi, similar colonization levels were observed (Fig. 2A). After 9 dpi, bacterial loads in the feces of both groups were decreased by bacterial shedding. Notably, at 9 to 15 dpi, the tatC mutant displayed higher levels of colonization than seen with the wild-type strain (Fig. 2A), indicating that bacterial shedding of the tatC mutant is delayed. Furthermore, we evaluated the levels of gut inflammation by determining the amount of fecal lipocalin-2. Lipocalin-2 levels in both groups peaked at 9 to 12 dpi, after which they decreased to basal levels (Fig. 2B). These results suggest that Tat is dispensable for C. rodentium infectivity but is involved in bacterial fitness in the gut. Notably, it is unclear where the tatC mutant is present during the course of the gut infection, namely, the cecum, colonic lumen, or surface of colonic epithelial cells.
C. rodentium tatC mutant displays prolonged colonization. C57BL/6 mice (n = 6 per group) were infected orally with 1 × 109 CFU of C. rodentium wild-type strain (wt, T307) or tatC mutant (tatC, T329). (A) Fecal C. rodentium CFU over the indicated time. (B) Fecal lipocalin-2 ELISA. The horizontal bars indicate the median values. ns, not significant; *, P < 0.05; Mann-Whitney U test.
The induction of neutrophil recruitment is impaired with the tatC mutant.To determine the reason for prolonged colonization with the tatC mutant, we next examined host responses at 12 dpi, at which time inflammation levels seemed to peak during C. rodentium infection. Similar to the previous infection experiment (Fig. 2), mice were infected via the oral route with the wild-type or tatC mutant strain for the next 12 days. Similar to the results shown in Fig. 2A, prolonged colonization was observed with the tatC mutant (Fig. 3A). Colonic histopathological analysis at 12 dpi showed that levels of mucosal inflammation in mice infected with the tatC mutant were lower than those in animals infected with the wild-type strain (Fig. 3B and C). For further analysis, we investigated inflammatory responses in colonic tissue based on the expression of inflammatory cytokine (TNF-α and IL-6) and chemokine (Mip2 and Kc) genes. Infection with both wild-type and tatC mutant strains induced the expression of inflammatory TNF-α and Mip2 but not IL-6 and Kc (Fig. 4A). Notably, the expression levels of IL-6 and Mip2 were significantly lower in mice infected with the tatC mutant than in those administered the wild-type strain. Infiltrating neutrophils have been shown to play a critical role in C. rodentium gut infection (26–28). Therefore, we next compared levels of neutrophil recruitment to colon tissue. To detect infiltrating neutrophils, colonic sections from mice infected with wild-type and tatC mutant strains were stained with antibodies against Ly6G, a marker of neutrophils. The number of infiltrating neutrophils in colon tissue was decreased in mice infected with the tatC mutant compared to that with the wild-type strain (Fig. 4B and C). Collectively, these results suggest that the tatC mutant is impaired in its ability to induce inflammatory responses, and, especially, diminished Mip2 chemokine-induced neutrophil recruitment in colon tissue is likely to contribute to prolonged colonization.
Attenuated inflammation in mice infected with C. rodentium tatC mutant. C57BL/6 mice (n = 7 or 9) were infected orally with 1 × 109 CFU of C. rodentium wild-type strain (wt, T307) or tatC mutant (tatC, T329). (A) Fecal C. rodentium CFU over time. The horizontal bars indicate the median values. ns, not significant; *, P < 0.05; **, P < 0.01; Mann-Whitney U test. (B) Representative H&E section of colonic tissue at 12 days postinfection (dpi) with C. rodentium wt (T307) or tatC mutant (T329). Scale bar =100 μm. (C) Colonic pathologies in H&E-stained colonic tissue section at 12 dpi were scored separately and plotted as stacked vertical bars. **, P < 0.01; Mann-Whitney U test.
Reduced neutrophil infiltration in mice infected with C. rodentium tatC mutant. C57BL/6 mice (n = 7 or 9) were infected orally with 1 × 109 CFU of C. rodentium wild-type strain (wt, T307) or tatC mutant (tatC, T329). As a control, uninfected mice (n = 8) were prepared. (A) Transcript levels of inflammatory cytokines (TNF-α and IL-6) and chemokines (Mip2 and Kc) in colonic tissue at 12 dpi. The horizontal bars indicate the median values. ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001; Mann-Whitney U test. (B) Photomicrographs of neutrophil immunostained colonic sections at 12 dpi. Scale bar = 50 μm. (C) Quantitative analyses of the experiment shown in panel B. Respective plot is mean of 5 fields per each sample. The horizontal bars indicate the median values. **, P < 0.01; Mann-Whitney U test.
Attenuated inflammation in mice infected with tatC mutants is not mediated by the type III secretion system.The activities of the type III secretion system (T3SS) are essential for C. rodentium colonization, which is accompanied by inflammation (2, 29). Thus, we examined whether attenuated inflammation in mice infected with the tatC mutant is associated with T3SS activities. We first examined T3SS gene expression based on the quantitative PCR (qPCR) analysis of ler and tir gene expression during incubation in Dulbecco’s modified Eagle’s medium (DMEM), in which T3SS-encoding genes are induced. The expression levels of both genes were not significantly changed in either the wild-type or the tatC mutant strain, indicating that the tatC mutation has no effect on the expression of T3SS genes (Fig. 5A).
Type III secretion system (T3SS) is functional in a C. rodentium tatC mutant. C. rodentium wild-type strain (wt, T307) or tatC mutant (tatC, T329) was grown in DMEM, which is a T3SS-inducible condition. (A) Transcript levels of type III genes (ler and tir) relative to rpoD, determined by qPCR. Bars represent the mean ± SD of the results from three independent experiments. ns, not significant; unpaired Student's t test. (B) Coomassie blue stain of secreted proteins (supernatant) and whole-cell proteins (whole cell). Arrowheads to the right of the gel indicate two major T3SS-secreted proteins, EspA and EspB. (C) Western blot analyses of secreted proteins (supernatant) and whole-cell proteins (whole cell), using EspB and DnaK antibodies.
We next tested the secretion profile of T3SS substrates under the same conditions. Coomassie blue staining of supernatant and whole-cell fractions from wild-type and tatC mutant strains grown in DMEM showed identical patterns in the two fractions (Fig. 5B). Notably, bands corresponding to two major T3SS substrates, EspB and EspA, were found in the supernatant fraction from each strain. Furthermore, Western blot analysis using an antibody against the type III-secreted protein EspB revealed that the ability of the tatC mutant to secrete this protein was equivalent to that of the wild-type strain (Fig. 5C). In addition, we used an antibody against the cytoplasmic protein DnaK to check protein leakage from the cytoplasmic compartment to the supernatant fraction. DnaK proteins were exclusively detected in the whole-cell fraction but not the supernatant, and these levels were equivalent in the two strains (Fig. 5C). These results suggest that neither protein leakage nor bacterial lysis occurred and together clarify that the attenuated inflammation observed with the tatC mutant is not based on altered T3SS activity.
The tatC mutant is rapidly eliminated from the gut lumen upon coinfection with wild-type C. rodentium.Previous data suggested a causal link between impaired inflammatory responses, such as neutrophil chemotaxis, and prolonged colonization of the tatC mutant. To further verify this, we next performed mouse infection experiments with a mixture of wild-type and tatC mutant strains. C57BL/6 mice were orally infected with a 1:1 bacterial mixture of each strain, and bacterial loads in feces were monitored for the next 21 days. In the feces, the tatC mutant exhibited a pronounced colonization defect by 5 dpi (Fig. 6A and B). During the next 7 days, the wild-type strain displayed high levels of colonization (107 to 108 CFU/g), whereas the colonization levels of the tatC mutant were profoundly reduced (Fig. 6B). These data might support the hypothesis that robust inflammatory responses induced by the wild-type strain can diminish the sustained colonization of the tatC mutant.
Colonization of C. rodentium tatC mutant is impaired in mixed infection with wild-type strain. C57BL/6 mice (n = 5) were infected with 1:1 mixture (1 × 109 CFU by gavage) of C. rodentium wild-type strain (wt, T307) and tatC mutant (tatC, T329). The fecal loads of both strains were determined by selective plating. (A) Competitive infection (CI) indices were determined over the indicated time. dpi, days postinfection. The horizontal bars indicate the median values. ns, not significant; **, P < 0.01; Mann-Whitney U test. (B) Bacterial loads of both competing strains (wt and tatC mutant) are plotted. ns, not significant; *, P < 0.05; one-tailed Wilcoxon matched-pairs signed rank test on paired data (dashed lines).
The tatC mutant exhibits impaired resistance to bile acids.Antimicrobials such as antimicrobial peptides and bile acids can affect bacterial fitness in the intestinal tract (23, 30, 31). Furthermore, previous studies have shown that tatC mutants of E. coli or S. Typhimurium are impaired in resistance to detergents such as sodium dodecyl sulfate and bile acids (22, 23, 25, 32). These facts led us to hypothesize that antimicrobials can contribute to C. rodentium clearance from the gut lumen. Thus, we next tested resistance to antimicrobials by determining the MICs of magainin-2 and deoxycholate (DOC) (Table 2). The MICs of magainin-2 for all tested strains were identical. However, of note, the MICs of DOC for the tatC mutant were reduced by approximately 10-fold compared to those with the wild-type and complementation strains. These data suggest that the tatC mutant is impaired in resistance to bile acids.
MICs of magainin-2 and deoxycholate toward C. rodentiuma
Bile acids contribute to C. rodentium clearance from the gut lumen.Next, we investigated whether bile acids contribute to C. rodentium clearance from the gut lumen. To control levels of bile acids in the intestinal tract, we fed mice with rodent chow containing colestimide resin, which sequesters bile acids in the intestinal tract, resulting in the increased excretion of bile acids in the colon, as confirmed by increased levels of bile acids in feces (23, 33, 34). We monitored gut colonization levels of the C. rodentium wild-type strain for the next 12 days in mice fed chow containing colestimide resin or control chow. Lower levels of colonization at 5, 7, 9, and 12 dpi were found in mice fed chow containing colestimide resin than in mice fed control chow (Fig. 7A). Concentrations of fecal bile acids in the mice fed chow containing colestimide resin were gradually increased and were significantly higher than those in the mice fed control chow (Fig. 7B). Further, increased levels of bile acids in the colon were also confirmed based on the elevated expression of Fabp6, a bile acid-bound nuclear receptor FXR-targeted gene (Fig. 7C). In addition, concentrations of bile acids in the colon were correlated with colonization levels of C. rodentium (Fig. 7D; P = 0.0345, Pearson’s correlation coefficient). Furthermore, we investigated the expression levels of the TNF-α, IL-6, Mip2, Kc, and IL-1β genes. The expression levels of TNF-α in the mice fed chow with colestimide resin were reduced compared to those in the mice fed control chow, whereas the expression levels of IL-6, Mip2, Kc, and IL-1β were similar between the two conditions (Fig. 7E), suggesting that the two groups of mice had similar extents of gut inflammation.
Excreted bile acids in the colon foster C. rodentium clearance. C57BL/6 mice (n = 10 per group) on chow containing colestimide resin (+ colestimide) or control chow (− colestimide) were infected orally with 1 × 109 CFU of C. rodentium wild-type strain (T307). (A) Fecal C. rodentium CFU over the indicated time. The horizontal bars indicate the median values. ns, not significant; *, P < 0.05; ***, P < 0.001; Mann-Whitney U test. (B) Concentrations of total bile acids in feces. The horizontal bars indicate the median values. *, P < 0.05; **, P < 0.01; Mann-Whitney U test. (D) Correlation between C. rodentium colonization and concentrations of excreted bile acids in colon at 12 dpi. (C and E) Transcript levels of farnesoid X receptor-targeted (Fabp6), inflammatory cytokine (TNF-α, IL-6, and IL-1β), and chemokine (Mip2 and Kc) genes in colonic tissue at 12 dpi. The horizontal bars indicate the median values. ns, not significant; **, P < 0.01; Mann-Whitney U test.
Chenodeoxycholic acid (CDCA), a component of bile acids, has been shown to enhance the mucosal barrier by upregulating the expression of intestinal antimicrobial factors such as RegIII lectin and Muc2 (35). Thus, we next investigated whether the increase in excreted bile acids in mice administered colestimide resin affects the expression of RegIIIβ and Muc2. qPCR analysis revealed that the increase in bile acids had no effect on the levels of either gene (Fig. S1).
Finally, to investigate whether the colestimide resin affects the DOC-mediated killing of C. rodentium, MICs for DOC were determined in the presence of this bile sequestrant (Table S1). In addition to C. rodentium strains, S. Typhimurium strains also were used as a control because DOC-susceptible S. Typhimurium strains have been shown to become resistant in the presence of cholestyramine resin, a similar bile sequestrant (36). Similarly to the previous report, the MIC of the S. Typhimurium DOC-susceptible strain (tatC mutant) was identical to that of the wild-type strain in the presence of colestimide resin (Table S1). In contrast, the C. rodentium tatC mutant remained DOC susceptible in the presence of colestimide resin. Thus, we next performed the mixed infection experiment with the mice fed chow containing colestimide resin. The tatC mutant tended to have more of a colonization defect; however, the competitive index (CI) values were not significantly different between mice fed chow with colestimide resin and those fed control chow (Fig. S2). Collectively, these results provide evidence that resistance to excreted bile acids in the colon plays a critical role in C. rodentium fitness in the gut lumen. Further, the enhanced clearance mediated by bile acids might be due to their bactericidal activities and not those of other antimicrobials.
Paired and chained forms of C. rodentium confer susceptibility to bile acids.Earlier work showed that chains of S. Typhimurium cells caused by cell division defects are associated with impaired resistance to DOC (23). Therefore, we next investigated whether the chain form of C. rodentium is also susceptible to this detergent. GFP-expressing wild-type and tatC mutant strains of C. rodentium were incubated with sublethal concentrations of DOC (0.5%), which was followed by incubation with propidium iodide (PI), an indicator of membrane integrity, to evaluate the cytotoxic effect. DOC-killed bacteria would not express GFP in the cytoplasm due to membrane damage, resulting in propidium iodide (PI)-stained cells (Fig. 8A). Consistent with the previous results based on MICs (Table 2), approximately 65% of the tatC mutant cells were killed, whereas the wild-type strain was resistant to DOC (38% relative killing) (Fig. 8B). In individual cells, approximately 60% of wild-type single and paired cells expressed GFP, indicating that these cells were not killed by DOC. Notably, DOC-mediated cytotoxicity in paired wild-type cells was significantly increased compared to that in single cells (Fig. 8C). Similar results were observed with tatC mutant single and paired cells. Furthermore, almost all tatC mutant chains comprised PI-stained cells, and the proportion of PI-stained chains was higher than that with single or paired cells (Fig. 8C), indicating that the paired and chained forms of C. rodentium strains exhibit attenuated resistance to DOC-mediated bactericidal effects.
Chain of C. rodentium by cell division defect confers attenuated resistance to deoxycholate. GFP-expressing C. rodentium wild-type strain (wt, T307) or tatC mutant (tatC, T329) grown in LB medium was incubated with 0.5% deoxycholate (DOC), followed by treatment with propidium iodide (PI) solution to evaluate the DOC-mediated killing. The mixture was placed on a 1.5% agarose pad, sealed under a glass coverslip, and observed by fluorescence microscopy. (A) Representative fluorescence microscopy images. Scale bar = 20 μm. (B) Microscopy quantification of PI-stained cells. Three independent experiments were performed. The horizontal bars indicate the median values. ***, P < 0.001; Mann-Whitney U test.
DISCUSSION
Robust inflammation is necessary for C. rodentium clearance from the gut lumen, which is mediated by an array of innate and adaptive immune responses (6). Notably, earlier studies have shown that CXCR2-dependent neutrophil infiltration is essential for the elimination of C. rodentium from the inflamed gut lumen (28, 37, 38). Moreover, it was recently shown that the RIPK2/RegIIIβ axis-dependent recruitment of neutrophils results in pathogen clearance as a main source of interleukin 22 (IL-22) production (39). Therefore, we conclude that lower levels of infiltrated neutrophils are linked to the prolonged gut colonization of C. rodentium in mice infected with the Tat mutant. At present, it is unclear why the ability of the Tat mutant to induce neutrophil influx is attenuated. Earlier work showed that Tir-induced actin remodeling determines neutrophil influx levels in mice infected with C. rodentium (40), leading us to hypothesize that interactions between the tatC mutant and murine colonic cells could be insufficient to induce inflammatory responses. In contrast, our results showed that the Tat mutant possesses normal T3SS gene expression and secretion compared to the wild-type strain. Future work will be needed to reveal the role of Tat in the interaction between host cells and C. rodentium in the inflamed gut.
Based on our data, we propose that excreted bile acids in the colon comprise a host factor that controls C. rodentium gut colonization. Bile acids are cholesterol-derived surfactants that emulsify dietary lipids and thereby facilitate absorption. In addition, this property also confers bactericidal activity to bile acids. Notably, bile acids can affect microbial communities in the gut. For example, dietary fat content can alter the microbiota composition by facilitating the excretion of bile acids in the colon (41, 42). Gram-negative enteric bacteria, such as Salmonella spp. (43, 44), E. coli (45), V. cholerae (46, 47), and C. jejuni (48), are relatively resistant to the cytotoxic effects of bile acids. Basically, resistance to bile acids involves lipopolysaccharide (LPS) and efflux pumps. Especially, since rough LPS mutants are highly sensitive to bile acids (43–45, 49), the robust barrier of the outer membrane probably confers this resistance by inhibiting interactions between bile acids and the bacterial cell. Our data show that paired and chained cells of C. rodentium are more sensitive to deoxycholate-mediated killing than are single cells. Therefore, it is tempting to speculate that bile acids can more easily interact with these susceptible cells than with single cells. Furthermore, this could be explained by the fact that LPS biosynthesis is immature at the septum of dividing and chained cells, and thus, LPS cannot act as a barrier to prevent bile acids from accessing bacterial cells in this case. Indeed, it was previously shown that certain peptidoglycan recognition proteins target bacterial cells at the site of daughter cell separation during cell division (50). Furthermore, LPS has been shown to inhibit the access of the antimicrobial lectin RegIIIβ to bacterial cells through steric hindrance (51, 52). Thus, defining the selective targeting mechanisms of bile acids and analyzing their cytotoxic effects represent promising topics for future work.
Besides their bactericidal effect, bile acids can modulate mucosal immunity in the intestinal tract (53). It is notable that bile acids have been shown to trigger an inflammatory response by activating the NLRP3 inflammasome (54) and that activation of NLRP3 and NLRP4 inflammasomes is required to limit C. rodentium infection (55). Based on these facts, excreted bile acids in the colon could also protect mice from C. rodentium infection by activating the NLRP3 inflammasome. Earlier work showed that bile acids, particularly CDCA and deoxycholic acid, act as a danger-associated molecular pattern for the NLRP3 inflammasome, resulting in elevated expression of IL-1β (54). However, our results showed that increased bile acids in mice fed a diet containing colestimide resin had no effect on the expression levels of IL-1β, suggesting that the bile acid-dependent C. rodentium clearance might not be due to activation of the NLRP3 inflammasome. A detailed analysis of this is thus required in future work.
As another role in mucosal immunity, certain bile acids such as CDCA fortify the mucosal membrane barrier by elevating expression levels of antimicrobials, including antimicrobial peptides, RegIII lectins, and mucins (35). However, given that increased bile acids in mice fed a diet containing colestimide resin had no effect on the expression of RegIIIβ or Muc2, the amount of excreted CDCA might be insufficient to induce antimicrobials. Indeed, it is also notable that CDCA accounts for only 2% of the total bile acids in murine feces (56). It remains unclear how CDCA activates the expression of antimicrobials. Regardless, it seems reasonable to conclude that C. rodentium clearance by excreted bile acids can be attributed to its bactericidal activity but not the CDCA-mediated activation of an antimicrobial program.
Bile acids also have an anti-inflammatory effect, as they can bind and subsequently activate the farnesoid X receptor (FXR) and G protein-coupled bile acid receptor 1 (GPBAR1 or Takeda G-protein receptor 5) (56, 57). Notably, both receptors have been shown to limit activation of the NLRP3 inflammasome (54, 58), suggesting that repression of NLRP3-dependent inflammation by the activation of bile acids bound to these receptors could contribute to C. rodentium colonization. Given that our results showed that an increase in excreted bile acids in mice fed a diet containing colestimide resin could induce the expression of Fabp6, a target gene of FXR, increased bile acids could activate these receptors. The use of FXR or GPBAR1 knockout mice in future studies would provide better insights into the link between C. rodentium colonization and these receptor activation-mediated immune responses.
Animal models are extremely helpful for studying bacterial gut infection and the host factors associated with the infection. However, the results obtained from animal models may not be applicable to humans due to genetic and epigenetic differences. Furthermore, Shiga toxin is an additional virulence factor of EHEC, which may account for the differences observed in the gut infections caused by C. rodentium and EHEC. Nevertheless, the data obtained from the animal models used in this study will help in the design of better clinical studies to address the unresolved issues.
In summary, we demonstrate that the Tat system of C. rodentium plays a key role in gut infection. This is further associated with the ability of C. rodentium to induce neutrophil recruitment. We also demonstrate that excreted bile acids in the colon can define the colonization levels of C. rodentium. Our findings suggest that controlling Tat activity and bile acids in the colon could be promising therapeutic strategies to treat human EHEC and EPEC infections.
MATERIALS AND METHODS
Bacterial strains and plasmids.The bacterial strains and plasmids used in this study are listed in Table 1. C. rodentium strains are derivative of a mouse virulent C. rodentium wild-type strain DBS100 (a kind gift from A. Kuwae and A. Abe, Kitasato University, Tokyo, Japan) (59). pJB861 was obtained from the National Institute of Genetics (NIG) in Japan. In the mouse infection experiments, C. rodentium strain T307, a DBS100 strain harboring the plasmid pJB861, was used as mouse virulent strain. The plasmid stability of pJB861 in vivo was verified by a comparison of CFU grown on LB supplemented with nalidixic acid or kanamycin with strain T690, a spontaneous nalidixic acid-resistant DBS100 harboring the plasmid pJB861 (Fig. S3).
Construction of C. rodentium mutants.C. rodentium mutant strains harboring chromosomal in-frame deletions were created using a lambda/red homologous recombination system (60). The primers used for construction of the mutant strains are listed in Table S2.
Tat assay by using TorA::GFP reporter system to analyze Tat functionality.The Tat reporter plasmid pZ300 expresses signal sequence of TorA (trimethylamine N-oxide reductase) of Salmonella enterica serovar Typhimurium fused to green fluorescent protein (GFP). Wild-type or tatC mutant strains of C. rodentium were transformed by electroporation. The transformants were grown in liquid culture containing arabinose to induce the expression of the TorA::GFP fusion proteins, and localization of the fusion proteins was determined by fluorescence microscopy.
Construction of complementary plasmid.Complementary plasmids were constructed using DNA fragments containing the tatC gene, which was amplified by PCR with the primer set CR tatC-KpnI-FW and CR tatC-SalI-RV, and using C. rodentium strain DBS100 chromosomal DNA as the template, which were digested with KpnI and SphI and then ligated between the same sites of pMW118, yielding pMW-tatC. The primers used for construction of complementary plasmid are listed in Table S2.
Bacterial morphology analysis.C. rodentium strains were grown in LB broth, placed on a 1.5% agarose pad, and sealed under a glass coverslip. The samples were observed and imaged using the Zeiss Axio Vert.A1 microscope.
Animals.Six- to 8-week-old female C57BL/6 mice were purchased from CLEA Japan. All animal experiments were performed according to protocols approved by the Kitasato University Institutional Animal Care and Use Committee (permit numbers 17-52, 17-54, and 17-55).
Mouse infection experiment.C. rodentium strains were grown in LB medium with appropriate antibiotic(s) with shaking at 160 rpm at 37°C for 12 h, diluted 1:20, and subcultured for 4 h in the same medium without supplementation of antibiotics. After washing, the inocula were prepared in sterile phosphate-buffered saline (PBS). Mice were infected orally with 1 × 108 CFU C. rodentium strains by gavage. For bacterial quantification, fecal pellets were collected in sterile PBS at the time indicated and subjected to bead beating; serial dilutions were plated on LB plates containing the appropriate antibiotics (50 μg/ml kanamycin and 10 μg/ml chloramphenicol). A competitive index (CI) was determined by dividing the population sizes of wild-type strains of C. rodentium by those of its derivative mutants. To control the luminal levels of bile acids, mice were fed a normal rodent chow containing colestimide resin, an anion exchange resin (1.5%; Mitsubishi Tanabe Pharma).
Fecal lipocalin-2 ELISA.Fecal pellets from mice were dissolved in 0.5 ml of sterile PBS and diluted (1:20) in PBS. Fecal lipocalin-2 content was measured using the mouse lipocalin-2 enzyme-linked immunosorbent assay (ELISA) DuoSet (R&D Systems), according to the manufacturer’s instructions. The optical density at 405 nm (OD405) was measured using an iMark microplate absorbance reader (Bio-Rad), and data were analyzed using the GraphPad Prism program.
Histological analysis.Parts of the distal colon were fixed in 4% formaldehyde and embedded in paraffin. Sections were cut and stained with hematoxylin and eosin (H&E). For immunostaining of infiltrated neutrophils, sections were immunostained with rat anti-Ly-6G antibody (1A8; BioLegend). To determine the degree of inflammation, sections were scored as previously described (61), evaluating epithelial hyperplasia, epithelial integrity, inflammatory cell infiltrate, loss of goblet cells, and submucosal edema.
Mouse gene expression of distal colon tissue.The distal colon tissue was excised and stored in the RLT buffer (Qiagen) at −80°C until further analysis. Total RNA was extracted from homogenized colon tissue using the RNeasy minikit (Qiagen). The concentration of RNA was measured using a NanoDrop spectrophotometer (Thermo Fisher Scientific). Using 500 μg of RNA, cDNA was prepared using TaqMan reverse transcription reagents (Thermo Fisher Scientific), following the manufacturer’s protocol. Quantitative PCR (qPCR) was conducted using a CFX96 real-time PCR detection system (Bio-Rad), using SYBR Fast qPCR master mix (Kapa Biosystems), according to the manufacturer’s instructions. Relative mRNA expression was calculated using the ΔCT method with GAPDH as a reference gene (62). The primers used in the analysis are shown in Table S2.
Analysis of type III secretion system activity.To determine the activities of the T3SS, gene expression and secretion were analyzed. C. rodentium strains grown overnight in LB broth in a shaker at 37°C and 160 rpm were diluted 1:50 into prewarmed DMEM and grown for 6 h to an optical density at 600 nm (OD600) of 0.6 to 0.7 in a tissue culture incubator containing 5% CO2 with the tubes standing. The cultures were centrifuged for 5 min at 16,000 relative centrifugal force (RCF) and filtered through a 0.22-μm-pore-size filter (Millipore), yielding the cell-free supernatant and bacterial cells, respectively. For analyzing gene expression, total RNA from bacterial pellet was isolated by using an RNeasy minikit (Qiagen), samples were treated with the Turbo DNA-free kit (Thermo Fisher Scientific), and reverse transcription was performed using TaqMan reverse transcription reagents (Thermo Fisher Scientific). The targeted genes were amplified with specific primer pairs listed in Table S2, using a CFX96 real-time PCR detection system (Bio-Rad) under standard cycle conditions for SYBR Fast qPCR master mix (Kapa Biosystems). Relative transcript levels were normalized to the rpoD gene and calculated by using the 2−ΔCT method (62). For analyzing secretion, the filtered supernatant was precipitated with 10% (vol/vol) trichloroacetic acid (TCA) to concentrate the secreted proteins, and the precipitates were dissolved in sodium dodecyl sulfate-PAGE (SDS-PAGE) sample buffer, boiled, and separated on 10% or 12% SDS-PAGE gels. The secreted and whole-cell lysate proteins were visualized by Coomassie blue staining. In addition, these proteins also were separated on 10% or 12% SDS-PAGE gels and transferred onto polyvinylidene difluoride (PVDF) membranes, and T3S substrate EspB and cytoplasmic protein DnaK were detected by Western blotting using anti-EspB (63) and anti-DnaK (Abcam) antibodies.
Measurement of MICs for antimicrobials.MICs were determined as previously described (23). Briefly, C. rodentium strains were diluted to 1 × 106 CFU per ml with different concentrations of magainin-2 (LKT Laboratories, Inc.) or deoxycholate (Nacalai Tesque) in sterile LB broth and incubated at 37°C. After incubation, the OD595 was measured, and MICs were determined as the lowest concentrations of antimicrobials that were shown to prevent bacterial growth by more than 50% in comparison with the growth of the positive control, which contains no antimicrobials.
Measurement of bile acids in feces.Total bile acids in feces were quantified as previously described (23). Briefly, hot ethanol-extracted bile acids were analyzed using the total bile acids test (Fujifilm Wako Pure Chemical), according to the manufacturer’s instructions.
Deoxycholate killing analysis by propidium iodide staining.C. rodentium expressing GFP was mixed with 0.5% deoxycholate and incubated at 37°C for 20 min. Subsequently, propidium iodide (PI) solution (Dojindo) was added and the mixture was incubated further for 5 min at room temperature. GFP-expressing and PI-stained C. rodentium cells were counted by fluorescence microscopy.
Statistical analysis.Statistical tests were performed using Prism (version 5) for Mac OSX (GraphPad Software). Statistical significance (P < 0.05) was determined by a Mann-Whitney U test, Student's t test, a one-sided Wilcoxon matched-pairs signed rank test, or Pearson’s correlation coefficient (*, P < 0.05; **, P < 0.01; ***, P < 0.001).
ACKNOWLEDGMENTS
We thank Takuro Shiga for technical assistance.
This work was supported in part by Japan Society for the Promotion of Science (JSPS) KAKENHI grants JP16K08783 (to T.M.), JP19K07543 (to T.H.), and JP18K07119 (to N.O.).
The funders had no role in the study design, data collection and interpretation, or the decision to submit the work for publication.
FOOTNOTES
- Received 27 November 2019.
- Accepted 27 November 2019.
- Accepted manuscript posted online 9 December 2019.
Supplemental material is available online only.
- Copyright © 2020 American Society for Microbiology.
