Depletion of Butyrate-Producing Clostridia from the Gut Microbiota Drives an Aerobic Luminal Expansion of Salmonella

Abstract

The mammalian intestine is host to a microbial community that prevents pathogen expansion through unknown mechanisms, while antibiotic treatment can increase susceptibility to enteric pathogens. Here we show that streptomycin treatment depleted commensal, butyrate-producing Clostridia from the mouse intestinal lumen, leading to decreased butyrate levels, increased epithelial oxygenation, and aerobic expansion of Salmonella enterica serovar Typhimurium. Epithelial hypoxia and Salmonella restriction could be restored by tributyrin treatment. Clostridia depletion and aerobic Salmonella expansion were also observed in the absence of streptomycin treatment in genetically resistant mice but proceeded with slower kinetics and required the presence of functional Salmonella type III secretion systems. The Salmonella cytochrome bd-II oxidase synergized with nitrate reductases to drive luminal expansion, and both were required for fecal-oral transmission. We conclude that Salmonella virulence factors and antibiotic treatment promote pathogen expansion through the same mechanism: depletion of butyrate-producing Clostridia to elevate epithelial oxygenation, allowing aerobic Salmonella growth.

Figure 1. Cytochrome bd-II oxidase contributes to post-antibiotic pathogen expansion

(A and B) Minimal medium was inoculated with the indicated strains and the competitive index (CI) determined after 24 hours incubation in the presence of 8%, 0.8% or 0% oxygen. (C) Groups of C57BL/6 mice (N = 4) were infected intraperitoneally with the indicated strain mixtures and the CI determined four days after infection. (D–I) Streptomycin (Strep)-treated or mock-treated C57BL/6 mice were mock-infected or infected intragastrically with the indicated S. Typhimurium strains or strain mixtures (N is indicated in panels F and H or in Fig. S1A). (D, E) The CI in colon contents was determined four days after infection. (E) One day after infection, some mice were inoculated intragastrically with a culture of 17 human Clostridia isolates (C17) or received tributyrin (TB) supplementation daily. (E and I) One day after infection, some mice received chloroform-treated cecal contents (CTCC) of treatment naïve mice intragastrically. (F) CFU recovered from colon contents four days after infection with individual S. Typhimurium strains. (F and H) Each circle represents data from an individual animal. (G and I) Clostridia 16S rRNA gene copy numbers present in 20 ng of total bacterial DNA were determined at the indicated time points. (H) The concentration of butyrate was measured in cecal contents at the indicated time points after streptomycin treatment (see also Fig. S1B and S1C). Bars represent geometric means ± standard error. *, P < 0.05; **, P < 0.01; ***, P < 0.005; ****, P < 0.001; ns, not statistically significantly different; nd, none detected; wt, S. Typhimurium wild type.

Figure 2. Tributyrin treatment restores physiologic hypoxia of colonocytes and prevents cytochrome bd-II oxidase-dependent post-antibiotic pathogen expansion

(A–C) Groups of C57BL/6 mice (N = 4) were mock-treated (A), treated with streptomycin (Strep) (B) or streptomycin and tributyrin (TB) (C) and the colon collected one day later. Binding of pimonidazole (red fluorescence) was detected in sections of the colon counter stained with DAPI nuclear stain (blue fluorescence). Representative images are shown. (D–F) Groups of C57BL/6 mice (N = 5) were mock-treated, treated with streptomycin or streptomycin and TB. Concentrations of butyrate (D), propionate (E) and acetate (F) in cecal contents were determined 8 hours after TB supplementation. (G–H) Groups of C57BL/6 mice (N is shown in Fig. S1A except for TB supplementation where N = 4) were mock-treated or treated with streptomycin and infected one day later with the indicated strain mixtures. Some mice received mock supplementation or were supplemented with tributyrin 8 hours prior to pimonidazole injection. One day after infection, the competitive index (CI) in colon contents was determined. Bars represent geometric means ± standard error. *, P < 0.05; ***, P < 0.005; ****, P < 0.001; ns, not statistically significantly different; wt, S. Typhimurium wild type.

Figure 3. Virulence factors drive a cytochrome bd-II oxidase-dependent expansion of S. Typhimurium at later time points after infection

(A–C) Groups of C57BL/6 mice (N is shown in Fig. S1A) were mock-treated or treated with streptomycin (Strep) and infected one day later with the indicated strain mixtures. (A) At one and four days after infection (Days p.i.) the competitive index (CI) in colon contents was determined. (B) Cecal histopathology was scored using tissue collected four days after infection from four mice per group using criteria listed in Table S1 (see also Fig. S2). Each bar represents data from one individual animal. (C) Representative images of H/E-stained cecal sections scored in panel B. All images were taken at the same magnification. (D–F) Groups of C57BL/6 mice were treated with streptomycin or mock-treated. One day after streptomycin treatment organs were collected for analysis or mice were mock infected or infected with the indicated S. Typhimurium strain mixtures. Concentrations of butyrate (D), acetate (E) and propionate (F) in colon contents were determined at the indicated time points. (A–C and D–F) Bars represent geometric means ± standard error. *, P < 0.05; **, P < 0.01; ***, P < 0.005; ****, P < 0.001; ns, not statistically significantly different; wt, S. Typhimurium wild type.

Figure 4. S. Typhimurium virulence factors induce intestinal inflammation and are required for Clostridia depletion in the absence of antibiotic treatment

Groups of CBA mice (N = 6) were mock-infected or infected intragastrically with 1 × 10⁹ CFU/animal of the virulent S. Typhimurium wild type (wt) or the avirulent invA spiB mutant (invA spiB). (A) Fecal lipocalin-2 levels were determined by ELISA (see also Fig. S3). (B) S. Typhimurium CFU were determined in the feces over time. (C) Clostridia 16S rRNA gene copy numbers were determined by quantitative real-time PCR. (A–C) Data points represent geometric means ± standard error. (D) Average relative abundances of phylogenetic groupings at the class level determined by 16S profiling of the microbial community present in colon contents on day 10 after infection. (E) Normalized abundances of members of the class Clostridia in colon contents at the indicated days after infection (days p.i.). Boxes in whisker plots represent the second and third quartiles, while lines indicate the first and fourth quartiles. (F) Weighted principal coordinate analysis of 16S profiling data in which all reads identified as Enterobacteriaceae were excluded from analysis. For additional analysis of 16S profiling data see Fig. S4–S7. Each dot represents data from one animal. *, P < 0.05; **, P < 0.01; ****, P < 0.001.

Figure 5. Cytochrome bd-II oxidase and nitrate reductases contribute to a luminal S. Typhimurium expansion in the absence of antibiotic treatment

(A) Groups of CBA mice were infected intragastrically with a 1:1 mixture of the S. Typhimurium wild type (black circles) and the respiration-deficient napA narZ narG cyxA mutant (blue circles) and samples collected at the indicated days after infection (days p.i.). Dotted red lines connect strains recovered from the same animal. CI, competitive index; ****, P < 0.001; ns, not statistically significantly different; nd, not determined. (B) Groups of CBA mice (N = 4) were infected intragastrically with 1 × 10⁸ CFU/animal of one of the indicated S. Typhimurium strains and samples collected at the indicated days after infection.

Figure 6. Depletion of Clostridia increases colonocyte oxygenation and drives a cytochrome bd-II oxidase-dependent expansion of S. Typhimurium

Groups of CBA mice were mock-infected or infected intragastrically with a 1:1 mixture of the S. Typhimurium wild type (wt) and a cyxA mutant. At 5, 7 and 10 days after infection, mice were mock-inoculated or inoculated with 17 human Clostridia isolates (C17). (A) The competitive index (CI) in feces or colon contents was determined at the indicated time points. (B–D) Concentrations of butyrate (B), acetate (C) and propionate (D) in cecal contents are shown. (A–D) Bars represent geometric means ± standard error. (E–G) Binding of pimonidazole (red fluorescence) was detected in colonic sections counter stained with DAPI nuclear stain (blue fluorescence). Representative images are shown. *, P < 0.05; ***, P < 0.005; ****, P < 0.001; ns, not statistically significantly different.

Figure 7. Respiration is required for fecal-oral transmission

Groups of CBA mice (N = 6) were infected intragastrically with 1 × 10⁸ CFU/animal of either the S. Typhimurium wild type, a cyxA mutant or a napA narZ narG cyxA mutant. 10 days after infection, two infected mice (donors, black circles) were co-housed with two naïve mice (recipients, red circles) per cage and feces (A) or colon contents (B) were collected after 7 (A) or 18 days of co-housing (B).