The selective pressures leading to the evolution and maintenance of virulence in the case of facultative pathogens are quite unclear. For example, Escherichia coli, a commensal of the gut of warm-blooded animals and humans, can cause severe extraintestinal diseases, such as septicemia and meningitis, which represent evolutionary dead ends for the pathogen as they are associated to rapid host death and poor interhost transmission. Such infectious process has been linked to the presence of so-called "virulence genes." To understand the evolutionary forces that select and maintain these genes, we focused our study on E. coli B2 phylogenetic group strains that encompass both commensal and pathogenic (extra- and intraintestinal) strains. Multilocus sequence typing (MLST), comparative genomic hybridization of the B2 flexible gene pool, and quantification of extraintestinal virulence using a mouse model of septicemia were performed on a panel of 60 B2 strains chosen for their genetic and ecologic diversity. The phylogenetic history of the strains reconstructed from the MLST data indicates the emergence of at least 9 subgroups of strains. A high polymorphism is observed in the B2 flexible gene pool among the strains with a good correlation between the MLST-inferred phylogenetic history of the strains and the presence/absence of specific genomic regions, indicating coevolution between the chromosomal background and the flexible gene pool. Virulence in the mouse model is a highly prevalent and widespread character present in all subgroups except one. Association studies reveal that extraintestinal virulence is a multigenic process with a common set of "virulence determinants" encompassing genes involved in transcriptional regulation, iron metabolism, adhesion, lipopolysaccharide (LPS) biosynthesis, and the recently reported peptide polyketide hybrid synthesis system. Interestingly, these determinants can also be viewed as intestinal colonization and survival factors linked to commensalism as they can increase the fitness of the strains within the normal gut environment. Altogether, these data argue for an ancestral emergence of the extraintestinal virulence character that is a coincidental by-product of commensalism. Furthermore, the phenotypic and genotypic markers identified in this work will allow further epidemiological studies devoted to test the niche specialization hypothesis for the B2 phylogenetic subgroups.