Strain-specific galactose utilization by commensal E. coli mitigates Salmonella establishment in the gut

Abstract

Salmonella enterica serovar Typhimurium (S. Tm) is a major cause of gastrointestinal diseases worldwide. To date, options for prevention or curative therapy remain limited. The gut microbiota plays a protective role against enteric diseases, particularly in preventing establishment and proliferation of S. Tm. While most research has focused on microbiota-mediated pathogen exclusion during the later, inflammation-dominated stages of infection, little is known about how microbiota members mitigate S. Tm early gut colonization. To address this gap, we conducted 24 h in vivo competitive experiments using S. Tm and different commensal E. coli strains. We observed a significant reduction in pathogen load, which was strain-specific and particularly evident with E. coli 8178. To investigate the underlying molecular mechanisms, we performed an in vivo screen using a rationally designed S. Tm library-which includes a wide range of carbohydrate utilization mutants-both in the absence and presence of E. coli strains. Our findings revealed that E. coli 8178-mediated S. Tm competition was driven by the exploitation of galactose during the early stage of infection. Identifying galactose as a key metabolite in pathogen exclusion by gut microbiota members enhances our mechanistic understanding of microbiota-mediated protection and opens new avenues for developing microbiota- and dietary-based strategies to better control intestinal infections.

Fig 1. E. coli 8178-mediated S. Tm early growth restriction is independent of interfering competition interactions.

(A) Experimental scheme. Streptomycin-pretreated specific pathogen-free (SPF) 129S6/SvEvTac mice were infected with either S. Tm alone or a 1:1 mixture of S. Tm and the indicated E. coli strain. The E. coli and S. Tm loads were determined by selective plating from faecal samples collected 24 h post-infection (p.i.) prior to euthanasia. Created in BioRender. Cherrak, Y. (2025) https://BioRender.com/n43x499. (B-C) E. coli mitigates the early growth of S. Tm in a strain-dependent manner. The S. Tm loads at 24 h p.i. are plotted and compared between S. Tm mono-infected and S. Tm + E. coli co-infected mice. E. coli 8178 (B) and Z1331 (C) were tested as competitors. (D) E. coli loads at 24 h p.i from Fig 1B and 1C are shown. (E) E. coli 8178 restricts S. Tm establishment independently of interference mechanisms. The faecal load of S. Tm collected 24 h p.i. is plotted and compared between S. Tm mono-infected and S. Tm + E. coli 8178 co-infected mice. The E. coli 8178 mutant strains tested are listed and categorized based on the nature of the interference system disrupted. (F) Inflammation does not trigger E. coli 8178-mediated S. Tm competition 24 h p.i. E. coli 8178 competitiveness was tested against an avirulent S. Tm mutant (ΔinvG/ΔssaV). (B-F) All competitive experiments are presented in a box-and-whiskers plot, showing the minimum to maximum values, median, and interquartile range (25^th to 75^th percentiles). The bar plots show the median. Two-tailed Mann–Whitney U tests to compare 2 groups in each panel. P ≥ 0.05 not significant (ns), p < 0.05 (*), p < 0.01 (**). A minimum of 5 mice (n ≥ 5) were used for each experimental group. CFU: colony forming units.

Fig 2. Evaluation of S. Tm metabolic requirement in presence of E. coli strains.

(A) Schematic representation of the S. Tm metabolic mutant pool, illustrating the targeted carbohydrate utilization pathways, adapted from [8] (https://www.nature.com/articles/s41467-025-56890-y#Fig1). The mutants affect key enzymatic steps, including dehydratases, kinases, and isomerases, positioned between transport reactions and central metabolic intermediates in glycolysis, the Entner-Doudoroff pathway, and the pentose phosphate pathway. Each circle represents an enzymatic step. Abbreviations are listed in S4 Table PG, peptidoglycan. (B) Experimental scheme. A WISH-barcoded S. Tm mutant pool library comprising 7 SL1344 wild type (WT), 2 control mutants (Δfrd and ΔdcuABC) and 35 deficient strains in carbohydrate utilization was pooled and used in this study. Streptomycin-pretreated SPF 129S6/SvEvTac mice were infected with the S. Tm pool alone or in combination with an E. coli strain (8178 or Z1331). The fitness of each S. Tm mutant was assessed by WISH-barcode sequencing from faecal samples. Created in BioRender. Cherrak, Y. (2025) https://BioRender.com/c18e944. (C) Effect of E. coli strains on S. Tm carbohydrate mutant fitness in vivo. The competitive index for each S. Tm mutant (listed on the left) is calculated relative to the WT and depicted as a heat map, across 3 different conditions: without E. coli (/), in presence of E. coli 8178 (+ E. coli 8178) in presence of E. coli Z1331 (+ E. coli Z1331). Median values from a minimum of 6 mice (n ≥ 6) are shown. Two-tailed Mann–Whitney U tests to compare 2 groups (S. Tm library alone vs S. Tm library + E. coli 8178 or S. Tm library alone vs S. Tm library + E. coli Z1331). P ≥ 0.05 not significant (ns), p < 0.05 (*), p < 0.01 (**), p < 0.005 (***), p < 0.001 (****).

Fig 3. Galactose utilization by E. coli 8178 is strain-specific and facilitates the early inhibition of S. Tm growth in the gut.

(A) Fitness of the galK-deficient E. coli strain in vivo. Streptomycin-pretreated 129S6/SvEvTac mice were infected with both the E. coli WT and ΔgalK strains, either in the presence or absence of the WT S. Tm (+ S. Tm). The competitive index of the ΔgalK mutant in E. coli 8178 (left) and Z1331 (right) at 24 h post-infection (p.i.) are plotted. (B) Effect of E. coli 8178 on the fitness of galK-deficient S. Tm in vivo. Streptomycin-pretreated 129S6/SvEvTac mice were infected with both the S. Tm WT and ΔgalK strains, in the presence of either the WT or galK-deficient E. coli 8178. The competitive indexes at 24 h p.i. of the S. Tm ΔgalK mutant are plotted across these two conditions. (C) Competitiveness of the E. coli 8178 ΔgalK mutant against S. Tm. The S. Tm load at 24 h p.i. is plotted and compared between S. Tm mono-infected and S. Tm + E. coli co-infected mice. The strains tested are indicated. (A-B) Dotted line: Fitness expected for a fitness-neutral mutation. (A-C) All experiments are presented in a box-and-whiskers plot, showing the minimum to maximum values, median, and interquartile range (25^th to 75^th percentiles). The bar plots show the median. Two-tailed Mann–Whitney U tests to compare 2 groups in each panel. P ≥ 0.05 not significant (ns), p < 0.05 (*), p < 0.01 (**). A minimum of 5 mice (n ≥ 5) were used for each experimental group. CFU: colony forming units.