Carbohydrate utilization by enterohaemorrhagic Escherichia coli O157:H7 in bovine intestinal content

Abstract

The bovine gastrointestinal (GI) tract is the main reservoir for enterohaemorrhagic Escherichia coli (EHEC) responsible for food-borne infections. Characterization of nutrients preferentially used by EHEC in the bovine intestine would help to develop ecological strategies to reduce EHEC carriage. However, the carbon sources that support the growth of EHEC in the bovine intestine are poorly documented. In this study, a very low concentration of glucose, the most abundant monomer included in the cattle dietary polysaccharides, was detected in bovine small intestine contents (BSIC) collected from healthy cows at the slaughterhouse. Six carbohydrates reported to be included in the mucus layer covering the enterocytes [galactose, N-acetyl-glucosamine (GlcNAc), N-acetyl- galactosamine (GalNAc), fucose, mannose and N-acetyl neuraminic acid (Neu5Ac)] have been quantified for the first time in BSIC and accounted for a total concentration of 4.2 mM carbohydrates. The genes required for enzymatic degradation of the six mucus-derived carbohydrates are highly expressed during the exponential growth of the EHEC strain O157:H7 EDL933 in BSIC and are more strongly induced in EHEC than in bovine commensal E. coli. In addition, EDL933 consumed the free monosaccharides present in the BSIC more rapidly than the resident microbiota and commensal E. coli, indicating a competitive ability of EHEC to catabolize mucus-derived carbohydrates in the bovine gut. Mutations of EDL933 genes required for the catabolism of each of these sugars have been constructed, and growth competitions of the mutants with the wild-type strain clearly demonstrated that mannose, GlcNAc, Neu5Ac and galactose catabolism confers a high competitive growth advantage to EHEC in BSIC and probably represents an ecological niche for EHEC strains in the bovine small intestine. The utilization of these mucus-derived monosaccharides by EDL933 is apparently required for rapid growth of EHEC in BSIC, and for maintaining a competitive growth rate as compared with that of commensal E. coli. The results suggest a strategy for O157:H7 E. coli survival in the bovine intestine, whereby EHEC rapidly consumes mucus-derived carbohydrates that are poorly consumed by bacteria belonging to the resident intestinal microbiota, including commensal E. coli.

Fig. 1

Growth curves of EDL933 incubated in minimal media. M9 minimal medium was supplemented with 10 mM of each carbon source. Cultures were incubated at 37°C with aeration. Each time point is the mean of three independent experiments.

Fig. 2

Relative expression levels of genes required for the catabolism of mucus-derived carbohydrates during incubation of EDL933 in BSIC compared with M9-Glc. The ratio of mRNA level of each gene was measured in EDL933 incubated in filtered BSIC in comparison to cells grown in M9-Glc. RNA samples were collected during the exponential growth phase (grey), when the bacteria entered into the stationary phase (white) and during the stationary phase (black). Values are the mean ± 1 SEM of three independent experiments.

Fig. 3

Disappearance rate of mucus-derived carbohydrates. A. Concentration of total mucus-derived carbohydrates was monitored during incubation of BSIC-LEM samples (○) and BSIC-LEM samples inoculated with the EHEC strain EDL933 (•) or with the commensal E. coli strain BG1 (□). The concentration of each carbohydrate was quantified individually. B. Concentration of each carbohydrate was quantified in bacterial supernatants of BSIC-LEM samples (white) and BSIC-LEM samples inoculated with BG1 (black) or EDL933 (grey) after 3 h of incubation. Bars represent the SEM of three independent experiments. ***P < 0.01 and **P < 0.05 vs BSIC-LEM samples as determined by the Student t-test for independent samples.

Fig. 4

Concentration of each carbohydrate during incubation of EDL933 in BSIC-LEM samples. Bars represent the SEM of three independent experiments.

Fig. 5

Growth competition assays between EDL933 and its isogenic mutants. The BSIC-LEM samples were inoculated with a 1:1 mixture of the two strains. The mutant strains tested were defective for the pathway required for the catabolism of GalNAc (EDL933ΔagaF), fucose (EDL933ΔfucAO), galactose (EDL933ΔgalK), GlcNAc (EDL933ΔnagE), mannose (EDL933ΔmanA) or Neu5Ac (EDL933ΔnanAT). Bars represent the SEM of three independent experiments. The double asterisk denotes statistical significance, P < 0.05 and the triple asterisk denotes statistical significance, P < 0.01 as determined by the Student t-test for paired samples.

Fig. 6

Fold-change comparison of carbohydrate catabolism gene expression between the EHEC strain EDL933 and the commensal E. coli strain BG1. The ratio of mRNA level of each gene was measured in the E. coli strains incubated in filtered BSIC. RNA samples were collected during the exponential growth phase (grey), when the bacteria entered into the stationary phase (white) and during the stationary phase (black). Values are the mean ± 1 SEM of three independent experiments.