The gut commensal Bacteroides thetaiotaomicron exacerbates enteric infection through modification of the metabolic landscape

Abstract

The enteric pathogen enterohemorrhagic Escherichia coli (EHEC) causes severe diarrhea, but the influence of the gut microbiota on EHEC infection is largely unknown. A predominant member of the microbiota, Bacteroides thetaiotaomicron (Bt), is resident at EHEC attachment sites. We show that Bt enhances EHEC virulence gene expression through the transcription factor Cra, which is functionally sensitive to sugar concentrations. This enhanced virulence accompanies increased formation of attaching and effacing (AE) lesions requisite for EHEC colonization. Infection with Citrobacter rodentium, a natural mouse pathogen homologous to EHEC, in Bt-reconstituted mice results in increased gut permeability along with exacerbated host pathology and mortality compared to mice deplete of microflora. Bt modifies the metabolite environment at infection sites, increasing metabolites involved in gluconeogenesis, with stark increases in succinate, which can be sensed by Cra. Our findings suggest that microbiota composition affects disease outcome and may explain links between microbiota composition and disease susceptibility.

Figure 1. B. thetaiotaomicron (Bt) increases expression of one-fifth of the E. coli array probe sets

A, Summary of microarray results comparing EHEC grown with Bt to EHEC grown alone (EHEC+Bt/EHEC only). B, C, Pie graph categorizing the E. coli genes by function b, that increased ≥ 4-fold or c, that decreased ≥ 2-fold in the presence of Bt D, E, Growth curves (of 6 biological samples, experiments were repeated 2 times, each with 3 independent biological samples) of in vitro growth of D, EHEC or E, Bt. The error bars indicate the standard deviation of the mean. Doubling time for EHEC grown alone and in the presence of Bt is 50 min/gen and 53 min/gen, respectively. Doubling time for Bt grown alone and in the presence of EHEC is 282 min/gen and 97 min/gen, respectively.

Figure 2. Bt augments EHEC virulence via the catabolite repressor/activator protein

A, qRT-PCR of LEE, stcE and stx2A genes in EHEC grown alone (−) Bt or in the presence of Bt (+) Bt (n= 9; error bars, s.d.; ***P < 0.001, **P < 0.01, *P < 0.05). B, Fluorescent actin staining assay of HeLa cells infected with EHEC alone or in the presence of Bt, stained with fluorescein isothiocyanate-phalloidin (actin, green) and propidium iodide (bacterial and HeLa DNA, red). Original magnification, 63X. C, Quantification of fluorescent actin staining assay of the percentage of HeLa cells infected, as defined by pedestal formation by EHEC. D, Number of pedestals/infected cell. (n = 350 cells; error bars, s.d.; *** P<0.001). E, qRT-PCR of ler in WT, ΔqseC, ΔqseE, Δcra, ΔkdpE, ΔΔcrakdpE, and ΔfusK grown alone or in the presence of Bt (n = 9–15; error bars, s.d.; ***P < 0.001, **P < 0.01, P > 0.05 = ns). Each mutant has been normalized to 1 to show the fold-increase of the mutant when grown in the presence of Bt (+) Bt F, Representation of LEE pathogenicity island regulation.

Figure 3. E. faecalis, a member of the Firmicutes phylum, augments EHEC virulence gene expression

qRT-PCR of LEE genes in EHEC grown alone (−) E. faecalis or in the presence of E. faecalis (+) E.faecalis (n= 9; error bars, s.d.; ***P < 0.001).

Figure 4. C. rodentium as an EHEC infection model

A, B, Growth curves (of 6 biological samples, experiments were repeated 2 times, each with 3 independent biological samples) of in vitro growth of A, C. rodentium or B, Bt. The error bars indicate the standard deviation of the mean. Doubling time for C. rodentium grown alone and in the presence of Bt is 69 min/gen and 66 min/gen, respectively. Doubling time for Bt grown alone and in the presence of C. rodentium is 282 min/gen and 91 min/gen, respectively. C, qRT-PCR of LEE, nleA and stx genes in C. rodentium grown alone (−) Bt or in the presence of Bt (+) Bt (n = 6; error bars, s.d.; **P < 0.01, * P < 0.05).

Figure 5. Bt mediates its pro-virulence effect on C. rodentium by enhancing expression of the T3SS, not by a bloom in the C. rodentium population

C3H/HeJ mice were treated for 5 days with an antibiotics regimen to deplete gut microbiota. Half of the mice were reconstituted with Bt (+Bt) while the remainder of the mice were left deplete of gut microbiota. Mice were mock-infected (PBS, Bt only) or infected with C. rodentium (Citro, Stx+), C. rodentiumΔstx (CitroΔstx, Stx-), or C. rodentiumΔescN (CitroΔescN). A, Weight loss or gain from baseline (weight at day 0) over the course of infection (blue: mock-infected, reconstituted with Bt; red: Citro-infected, deplete of microbiota; green: Citro-infected, reconstituted with Bt; purple: CitroΔstx-infected, deplete of microbiota; turquoise: CitroΔstx-infected, reconstituted with Bt; orange: CitroΔescN-infected, deplete of microbiota; light blue: CitroΔescN-infected, reconstituted with Bt). B, Survival after infection (n = 7–22 mice/group; error bars, s.d.; ***P < 0.001, **P < 0.01: comparison of Citro vs. Citro+Bt; ^##P < 0.01, ^#P < 0.05: comparison of CitroΔstx vs. CitroΔstx+Bt;⁺⁺⁺P < 0.001, ⁺⁺P < 0.01: comparison of Citro+Bt vs. CitroΔstx+Bt). C, qRT-PCR analysis of ler, espA, eae, and nleA from mRNA isolated from fecal pellets of infected animals. Significance is indicated as follows: one asterisk P≤0.05, two asterisks P≤0.01; three asterisks P≤0.001. D, Ultrastructure of the distal colon harvested five days post-infection from mock-infected (PBS, Bt) or C. rodentium-infected mice either deplete of gut microbiota or reconstituted with Bt, 2500×. Microvilli destruction and C. rodentium forming attaching and effacing (AE) lesions on the colonic epithelium can be observed. Original magnification, 2500X (TEM). E, qRT-PCR of 16s rRNA from the major phylogenetic groups (green: Proteobacteria, blue: Bacteroidetes, purple: Firmicutes) from feces collected on day 1 and day 4 post-infection.

Figure 6. Bt contributes to the accelerated loss of a protective mucosal layer during C. rodentium infection

A, B, Antibiotics-treated C3H/HeJ mice either deplete of gut microbiota or reconstituted with Bt (+Bt) were mock-infected (PBS, Bt only) or infected with C. rodentium (Citro, Stx+), C. rodentiumΔstx (CitroΔstx, Stx-), or C. rodentiumΔescN (CitroΔescN). The histological changes in the colon and cecum were analyzed on day 5 post-infection based on the following scoring system: edema, 0 is no edema and 5 has the highest edema in the submucosa; crypt integrity, 1 = normal, 2 = irregular crypts, 3 = mild crypt loss, 4 = severe crypt loss, 5 = complete crypt loss; neutrophil, neutrophilic infiltration in the wall; apoptosis, number of apoptotic cells per 600× field (n = five fields); bacteria attachment, bacteria associated to the epithelial surface; vasculitis, 0 is no evidence of vasculitis and 5 is the most severe vasculitis. Scoring was performed blindly, and the scores for each parameter are an average of the cecum and distal colon, taken from two independent experiments with 3 mice/experiment. C, Cecum harvested on day 5 post-infection was stained with Muc2 (green) and DAPI (DNA, blue). Original magnification, 10X. D, E, qRTPCR of D, Tff3 (trefoil factor 3) and D, Muc2. RNA was isolated from colonic tissue harvested from mice on day 5 post-infection. (n = 6; error bars, s.d.). F, qRT-PCR of p1411 from feces collected on day 2 post-infection. G, Levels of FITC-Dextran in the serum of mice on day 5 post-infection. Mice were fasted 4 hrs prior to administration of FITC-Dextran via oral gavage. FITC-Dextran levels were determined by measuring fluorescence at excitation 490 nm, emission 525 nm (n = 6; error bars, s.d. ***P < 0.001, **P < 0.01, *P < 0.05, P > 0.05 = ns).

Figure 7. Comparison of metabolic profiles of Bt-reconstituted and microflora-deplete animals

A, The expression profiling of metabolites present in the cecum on day 2 post-infection (n = 5). The data are presented as ratios to show fold-changes of the specified metabolite in the individual groups. B, The normalized data were analyzed by Principal Component Analysis (PCA) with unit variance scaling in order to obtain an unbiased overview of the data and to visualize clustering, trends, and outlier metabolites among all groups on the score plots. PCA score plots showed clear separation among the four groups in different colors (yellow: PBS, red: Citro+Bt, blue: Citro, green: Bt), with no outliers detected. The amount of variance in the X matrix explained by PC1 (R²) was 0.732, and estimate of the predictive ability of the model (Q²) (cumulative) was 0.446. C, Absolute levels of succinate, fumarate, and pyruvate present in the cecum on day 2 post-infection. (A.U. Absolute Units; n = 5; error bars, s.d. **P < 0.01, *P < 0.05). D, Protein level of the T3SS translocon protein EspA, encoded within LEE4, secreted into the supernatant of wild-type or Δcra EHEC cultures grown in the presence of 0.1% glucose, plus increasing levels of succinate. RpoA protein levels in the cell lysate serve as the loading control.