Novel reusable animal model for comparative evaluation of in vivo growth and protein-expression of Escherichia coli O157 strains in the bovine rumen

Abstract

Previously, we had demonstrated that Escherichia coli O157:H7 (O157) strain 86-24 expresses proteins involved in survival rather than virulence in vitro in rumen fluid from dairy cattle limit fed a maintenance diet. Here, we verified if this observation would be true for different O157 strains grown in vitro in rumen fluid from, and in vivo in the rumen of, animals on contrasting maintenance (high fiber) and lactation (high energy-protein) diets usually limit fed to dairy cattle. For the in vivo studies, an economical, novel, reusable and non-terminal rumen-fistulated animal model permitting simultaneous evaluation of multiple bacterial strains in the bovine rumen was developed. All experiments were conducted in duplicate using different animals to account for host-related variations. The O157 strains included, 86-24, EDL933 and the super shed SS-17. E. coli NalR (#5735), derived from a bovine intestinal commensal E. coli, was included as a control. As expected, diet influenced ruminal pH and volatile fatty acid (VFA) composition. The pH ranged from 6.2-7.0 and total VFA concentrations from 109-141 μM/ml, in animals fed the maintenance diet. In comparison, animals fed the lactation diet had a ruminal pH ranging between 5.18-6.0, and total VFA of 125-219 μM/ml. Strain dependent differences in O157 recovery from the rumen fluid of cattle fed either diet was observed, both in vitro and in vivo, with O157 strains 86-24 and EDL933 demonstrating similar survival patterns. Analysis of the O157 proteomes expressed in the rumen fluid/rumen verified previous observations of adaptive responses. Any difference in the adaptive response was mainly influenced by the animal's diet and growth conditions (in vitro and in vivo) and not the O157 strain. These new insights into the O157 responses could help formulate modalities to control O157 across strains in cattle at all stages of husbandry.

Fig 1. Materials and method used to introduce and retrieve bacteria from the rumen of a fistulated cow allowing reuse of the animal.

Fig 2

Graphs shown represent survival characteristics of three O157 strains in comparison to E. coli Nal^R (#5735), in vitro (A) and in vivo (B), in MRF. Bacterial survival characteristics depicted are following anaerobic incubation for 48 h, in vitro in flasks with MRF or in vivo in the rumen of animals fed the maintenance diet. Viable counts in colony forming units [cfu]/ml, with the standard error of means, are shown in both graphs. Key for each bacterial strain tested is also shown. Inserted tables show pH and VFA composition of MRF used in the in vitro (1A) and in vivo (1B) experiments.

Fig 3

Graphs shown represent survival characteristics of three O157 strains in comparison to E. coli Nal^R (#5735), in vitro (A) and in vivo (B), in LRF. Bacterial survival characteristics depicted are following anaerobic incubation for 48 h, in vitro in flasks with LRF or in vivo in the rumen of animals fed the lactation diet. Viable counts in colony forming units [cfu]/ml, with the standard error of means, are shown in both graphs. Inserted tables show pH and VFA composition of LRF used in the in vitro (2A) and in vivo (2B) experiments.

Fig 4. Analysis of O157 proteome as expressed in MRF and LRF, under different diet and growth conditions, using iBAQ and reference-based iTRAQ analysis #1.

iBAQ: A. Venn diagram showing the number of O157 proteins differentially expressed (L2FC >1) between MRF and LRF. B. Venn diagram showing the number of O157 proteins uniquely expressed in MRF and LRF. iTRAQ analysis #1 (reference-based): A. Venn diagram showing number of differentially expressed proteins in the in vitro and in vivo growth conditions, relative to the 86–24 in-vitro proteome, in MRF. B. Venn diagram showing number of differentially expressed proteins in the in vitro and in vivo growth conditions, relative to the 86–24 in-vitro proteome, in LRF.

Fig 5. Multidimensional scaling plots generated with data from iTRAQ analysis #2 (reference-free).

The plots depicting the similarity of global protein expression between O157 strains tested in (A) MRF or (B) LRF, in vitro and in vivo. This is based on Bray-Curtis dissimilarities generated from Scaffold normalized iTRAQ reporter intensities generated in reference-free iTRAQ analysis #2.