Escherichia coli Pathogenicity Island-Like Domains

In their molecular epidemiological analysis of Escherichia coli urosepsis isolates, Bingen-Bidois et al. interpret the finding of a pathogenicity island (PAI) II_J96-like domain in multiple ribotypes as conflicting with previous reports, which ostensibly describe this element as being confined to a single clone (1). This is potentially misleading, since the two references cited actually pertain to papG allele I, which is a hallmark of a different PAI from that of strain J96, i.e., PAI I. Indeed, papG allele I, which was absent from the population analyzed by Bingen-Bidois et al. (rather than being found in multiple ribotypes), is quite rare, having been found almost exclusively in strains closely related to strain J96 (4). In contrast, many previous reports have documented the co-occurrence of papG allele III, hly, and cnf1 (which implies the presence of a PAI II_J96-like domain, as discussed by Bingen-Bidois et al.) in multiple group B2-derived lineages (3). These include clonal groups 1 and 2 of Cherifi et al. (2) (which equate, respectively, electrophoretic types [ETs] 1 and 21 of Whittam et al. [7] and ETs 47 and 42 of Maslow et al. [6], the latter clonal group including archetypal strain 536), the O2:K5/K7:H1 clonal group (ET 39 of Maslow et al.), and the O18:K1:H7 clonal group that includes archetypal strains RS218 (meningitis) and NU14 (cystitis) (3, 5).