Clonal origin, virulence factors, and virulence.

Abstract

The recent report from Picard et al. regarding extraintestinal Escherichia coli infections provides new insights into the relationships between clonal origin, virulence factor repertoire, and virulence (1). That clonal origin was only secondarily associated with virulence in this study was suggested by the finding that although mouse lethality was more frequent overall among strains of phylogenetic group B2 than among other strains, lethality was proportional to the number of virulence traits present (other than aer and afa) and was similar among B2 and non-B2 strains after stratification according to virulence factor numbers (Table ​(Table1).1). TABLE 1 Lethality for mice according to B2 phylogenetic group status and number of virulence factorsa Stepwise logistic regression analysis of the data in Table ​Table11 of Picard et al. (1) showed that, when competing with the individual virulence factors, phylogenetic group B2 status was never the strongest predictor of lethality for any of the three lethality endpoints analyzed (data not shown). When the presence or absence of two or more non-aer non-afa virulence factors was entered as a single dichotomous variable into logistic regression analysis along with B2 status, the presence of multiple virulence factors accounted for almost all the observed lethality, with B2 status exhibiting a weak and marginally significant residual association with lethality for only one of the three lethality endpoints (Table ​(Table22). TABLE 2 Logistic regression analysis of multiple virulence factors and B2 status as predictors of mouse lethalitya These observations suggest that the group B2 genomic background is not required for virulence, whereas specific virulence properties are. They further suggest that the association of phylogenetic group B2 with extraintestinal virulence is attributable to the abundance of virulence factors in this lineage rather than to some undefined quality of the B2 genomic background per se. This concept is consistent with the traditional “virulent clone” hypothesis (4). What remains unclear is why extraintestinal virulence factors are so selectively concentrated within group B2. The major competing hypotheses to explain this phenomenon include (i) the existence of a special compatibility between the B2 genome and virulence genes and (ii) chance and timing. As proposed by Picard et al. (1), preexisting features of the B2 genome may have made this lineage preferentially receptive to the initial acquisition of exogenous virulence genes and/or may have promoted retention of such genes once acquired. Or, compensatory (adaptive) mutations which now help maintain virulence genes in group B2 may have appeared subsequent to the entry of these genes into the lineage, which initially may not have been particularly receptive to the foreign genes. Alternatively, there may be no special affinity between group B2 and virulence genes. Virulence genes may have been acquired by chance by a group B2 ancestor soon enough after the group's emergence as to be inherited vertically by most members of the group as the group underwent clonal expansion. This could account for the broad prevalence of virulence genes within group B2. Furthermore, because of the comparative “youth” of group B2, the initial virulence gene acquisition event(s) may have occurred so recently that there simply has not been sufficient evolutionary time for these newcomer sequences to diffuse outward from group B2 through the rest of the species despite their comparatively high rates of horizontal transfer relative to other components of the genome. Future studies designed to select between these competing hypotheses will refine our understanding of the origins of virulent extraintestinal E. coli and perhaps suggest new preventive strategies. It should be noted that knockout studies are required to confirm that the suspected virulence traits themselves (rather than associated bacterial properties) are actually responsible for enhanced virulence (2). Furthermore, knockout studies can identify bacterial traits which, although widely prevalent among commensal as well as pathogenic E. coli (e.g., guaA and argC), are necessary (albeit not sufficient) for virulence and hence may constitute potential targets for preventive or therapeutic interventions despite having no epidemiological associations with disease (3).