Molecular analysis as an aid to assess the public health risk of non-O157 Shiga toxin-producing Escherichia coli strains

Abstract

Shiga toxin-producing Escherichia coli (STEC) strains are commensal bacteria in cattle with high potential for environmental and zoonotic transmission to humans. Although O157:H7 is the most common STEC serotype, there is growing concern over the emergence of more than 200 highly virulent non-O157 STEC serotypes that are globally distributed, several of which are associated with outbreaks and/or severe human illness such as hemolytic-uremic syndrome (HUS) and hemorrhagic colitis. At present, the underlying genetic basis of virulence potential in non-O157 STEC is unknown, although horizontal gene transfer and the acquisition of new pathogenicity islands are an expected origin. We used seropathotype classification as a framework to identify genetic elements that distinguish non-O157 STEC strains posing a serious risk to humans from STEC strains that are not associated with severe and epidemic disease. We report the identification of three genomic islands encoding non-LEE effector (nle) genes and 14 individual nle genes in non-O157 STEC strains that correlate independently with outbreak and HUS potential in humans. The implications for transmissible zoonotic spread and public health are discussed. These results and methods offer a molecular risk assessment strategy to rapidly recognize and respond to non-O157 STEC strains from environmental and animal sources that might pose serious public health risks to humans.

FIG. 1.

Genetic organization of disease-associated type III effector nle genes in non-O157 STEC. (A) Graphic representation of the O-Island PAIs containing the nle effector genes examined in the present study in O157:H7 STEC strain EDL933. The direction of transcription for each open reading frame is indicated by an arrowhead. The type III effector genes examined are color-coded to each O-Island. nleG9 is adjacent to OI-71 but separated by multiple transposase elements and was therefore not included as part of OI-71. The annotation of open reading frames within each O-Island was modified from Perna et al. (26). (B) Pairwise association of nle genes from O-Islands 122, 57, 36, and 71 in non-O157:H7 STEC strains linked to outbreaks. The data are P values (Fisher exact test) for the co-occurrence of each disease-associated nle gene within the linked O-Island. nle genes that did not associate with outbreaks or HUS by MRA (nleB2 and nleD) were not included in pairwise comparisons. *, not-significant co-occurrence in non-HUS-associated strains (data not shown).

FIG. 2.

Type III effector-gene dosage correlates with STEC virulence. (A) Distribution of nle gene content in O157 strains (n = 15) and non-O157 STEC strains (n = 57) (see Table S1 in the supplemental material for strain details). All O157 strains examined (red) contain all of the nle genes presented in Table 1. Non-O157 strains linked to human disease (yellow, green, and blue) contain a variable repertoire of disease-associated nle genes. Seropathotype E STEC strains not associated with human disease (purple) contain fewer nle genes relative to strains that are associated with outbreaks and HUS (seropathotype B), HUS only (seropathotype C), and diarrhea (seropathotype D). The data shown are the percentage of strains positive for the number of nle genes within each seropathotype. (B) Prevalence of nle genes in O157 and non-O157 STEC strains. Epidemic non-O157 strains have a significantly higher nle gene content compared to strains not linked to epidemics. Similarly, HUS-associated non-O157 strains have a significantly higher nle gene content compared to non-HUS-associated strains. O157 STEC contains all disease-associated nle genes identified. The data are means with standard errors.