Salmonella Typhimurium strain ATCC14028 requires H2-hydrogenases for growth in the gut, but not at systemic sites

Abstract

Salmonella enterica is a common cause of diarrhea. For eliciting disease, the pathogen has to colonize the gut lumen, a site colonized by the microbiota. This process/initial stage is incompletely understood. Recent work established that one particular strain, Salmonella enterica subspecies 1 serovar Typhimurium strain SL1344, employs the hyb H2-hydrogenase for consuming microbiota-derived H2 to support gut luminal pathogen growth: Protons from the H2-splitting reaction contribute to the proton gradient across the outer bacterial membrane which can be harvested for ATP production or for import of carbon sources. However, it remained unclear, if other Salmonella strains would use the same strategy. In particular, earlier work had left unanswered if strain ATCC14028 might use H2 for growth at systemic sites. To clarify the role of the hydrogenases, it seems important to establish if H2 is used at systemic sites or in the gut and if Salmonella strains may differ with respect to the host sites where they require H2 in vivo. In order to resolve this, we constructed a strain lacking all three H2-hydrogenases of ATCC14028 (14028hyd3) and performed competitive infection experiments. Upon intragastric inoculation, 14028hyd3 was present at 100-fold lower numbers than 14028WT in the stool and at systemic sites. In contrast, i.v. inoculation led to equivalent systemic loads of 14028hyd3 and the wild type strain. However, the pathogen population spreading to the gut lumen featured again up to 100-fold attenuation of 14028hyd3. Therefore, ATCC14028 requires H2-hydrogenases for growth in the gut lumen and not at systemic sites. This extends previous work on ATCC14028 and supports the notion that H2-utilization might be a general feature of S. Typhimurium gut colonization.

Figure 1. 14028^hyd3 is impaired in early gut ecosystem invasion.

Gut colonization was monitored in two different mouse models: streptomycin-pretreated conventional mice (sm-SPF) and low complexity microbiota (LCM) mice. Mice were infected with a 1∶1 mixture (5×10⁷ cfu by gavage) of 14028^hyd3 and the isogenic background strain (14028^WT). Fecal loads of both strains were determined by selective plating. (A) Competitive infection indices were determined over 4 days. Ns  =  not significant (P≥0.05), ** P<0.01, Mann-Whitney U test. (B) Bacterial loads of both competing strains (14028^WT and 14028^hyd3) are depicted Ns  =  not significant (P≥0.05), * P<0.05; one-tailed Wilcoxon matched pairs signed rank test on paired data (dashed lines).

Figure 2. Oral infection experiments revealed that in ATCC14028, hydrogenases fuel pathogen growth in the intestine.

(A) LCM mice from Figure 1 were sacrificed at day 4 post infection and competitive indices in the cecum and at systemic sites were determined. ** P<0.01, Mann-Whitney U test. (B) Bacterial loads in the cecum and at systemic sites of both competing strains are plotted. Ns  =  not significant (P≥0.05), * P<0.05; one-tailed Wilcoxon matched pairs signed rank test on paired data (dashed lines). (C) Cecal tissue sections were HE-stained and scored for intestinal inflammation.

Figure 3. Intravenous infection experiments verified that in ATCC14028, hydrogenases are not required for growth at systemic sites.

(A) LCM mice were intravenously infected with a 1∶1 mixture of the 14028^hyd3 and the isogenic background strain (14028^WT) (5×10³ cfu). Animals were sacrificed at day 3 p.i. and competitive indices in the cecum and at systemic sites were determined. Ns  =  not significant (P≥0.05), Mann-Whitney U test. (B) Bacterial loads in the cecum and at systemic sites of both competing strains are plotted. Ns  =  not significant (P≥0.05); one-tailed Wilcoxon matched pairs signed rank test on paired data (dashed lines). (C) Cecal tissue sections were HE-stained and scored for intestinal inflammation.