Escherichia coli is widely considered to not survive for extended periods outside the intestines of warm-blooded animals; however, recent studies demonstrated that E. coli strains maintain populations in soil and water without any known fecal contamination. The objective of this study was to investigate whether the niche partitioning of E. coli occurs between cattle and their pasture. We attempted to clarify whether E. coli from bovine feces differs phenotypically and genotypically from isolates maintaining a population in pasture soil over winter. Soil, bovine fecal, and run-off samples were collected before and after the introduction of cattle to the pasture. Isolates (363) were genotyped by uidA and mutS sequences and phylogrouping, and evaluated for curli formation (Rough, Dry, And Red, or RDAR). Three types of clusters emerged, viz. bovine-associated, clusters devoid of cattle isolates and representing isolates endemic to the pasture environment, and clusters with both. All isolates clustered with strains of E. coli sensu stricto, distinct from the cryptic species Clades I, III, IV, and V. Pasture soil endemic and bovine fecal populations had very different phylogroup distributions, indicating niche partitioning. The soil endemic population was largely comprised of phylogroup B1 and had a higher average RDAR score than other isolates. These results indicate the existence of environmental E. coli strains that are phylogenetically distinct from bovine fecal isolates, and that have the ability to maintain populations in the soil environment.

Bacterial species are commonly defined by applying a set of predetermined criteria, including DNA-DNA hybridization values, 16S rRNA gene sequence similarity, phenotypic data as well as genome-based criteria such as average nucleotide identity or digital DNA-DNA hybridization. These criteria mostly allow for the delimitation of taxa that resemble typical bacterial species. Their application is often complicated when the objective is to delineate new species that are characterized by significant population-level diversity or recent speciation. However, we believe that these complexities and limitations can be easily circumvented by recognizing that bacterial species represent unique and exclusive assemblages of diversity. Within such a framework, methods that account for the population processes involved in species evolution are used to infer species boundaries. A method such as genealogical concordance analysis is well suited to delineate a putative species. The existence of the new taxon is then interrogated using an array of traditional and genome-based characters. By making use of taxa in the genera Pantoea, Paraburkholderia and Escherichia we demonstrate in a step-wise process how genealogical concordance can be used to delimit a bacterial species. Genetic, phenotypic and biological criteria were used to provide independent lines of evidence for the existence of that taxon. Our six-step approach to species recognition is straightforward and applicable to bacterial species especially in the post-genomic era, with increased availability of whole genome sequences. In fact, our results indicated that a combined genome-based comparative and evolutionary approach would be the preferred alternative for delineating coherent bacterial taxa.