Examination of the source and extended virulence genotypes of Escherichia coli contaminating retail poultry meat

Abstract

Extraintestinal pathogenic Escherichia coli (ExPEC) are major players in human urinary tract infections, neonatal bacterial meningitis, and sepsis. Recently, it has been suggested that there might be a zoonotic component to these infections. To determine whether the E. coli contaminating retail poultry are possible extraintestinal pathogens, and to ascertain the source of these contaminants, they were assessed for their genetic similarities to E. coli incriminated in colibacillosis (avian pathogenic E. coli [APEC]), E. coli isolated from multiple locations of apparently healthy birds at slaughter, and human ExPEC. It was anticipated that the retail poultry isolates would most closely resemble avian fecal E. coli since only apparently healthy birds are slaughtered, and fecal contamination of carcasses is the presumed source of meat contamination. Surprisingly, this supposition proved incorrect, as the retail poultry isolates exhibited gene profiles more similar to APEC than to fecal isolates. These isolates contained a number of ExPEC-associated genes, including those associated with ColV virulence plasmids, and many belonged to the B2 phylogenetic group, known to be virulent in human hosts. Additionally, E. coli isolated from the crops and gizzards of apparently healthy birds at slaughter also contained a higher proportion of ExPEC-associated genes than did the avian fecal isolates examined. Such similarities suggest that the widely held beliefs about the sources of poultry contamination may need to be reassessed. Also, the presence of ExPEC-like clones on retail poultry meat means that we cannot yet rule out poultry as a source of ExPEC human disease.

FIG. 1.

Two-way cluster analysis of individual PCR genotyping results for every avian-source E. coli isolate examined in this study. Also included for comparison purposes were human NMEC (n = 91) and human UPEC (n = 531) from a previous study (Johnson et al., 2008b). For each of the 1671 isolates (top to bottom), PCR results are depicted as positive (black) or negative (lime green) from left to right. On the left, a source bar was included to reference the sources of individual isolates. On the right, a Phylotype bar was included to illustrate E. coli phylogenetic group of individual isolates (Clermont et al., 2000). Isolates are clustered (top to bottom) according to their overall similarity in gene/trait possession, not including Source or Phylotype. Also, genes/traits correlated with one another are clustered (left to right). The dashed line over the dendrogram on the right depicts the cut-off for the generation of the nine clusters described in the text. NMEC, neonatal meningitis E. coli; UPEC, uropathogenic E. coli.

FIG. 2.

Two-way cluster analysis of average gene prevalence data for the four groups of avian-source E. coli. Also included for comparison purposes were human NMEC (n = 91) and human UPEC (n = 531) from a previous study (Johnson et al., 2008b). From this analysis a heat map was constructed to illustrate the relationships among the groups (Y-axis) with regard to traits examined (X-axis). Colors indicate gene prevalence, ranging from dark blue (least prevalent) to gray (of intermediate prevalence) to dark red (most prevalent). E. coli groups are clustered (top to bottom) according to their overall similarity in average gene prevalence. Also, traits correlated with one another are clustered (left to right).