Phylogenetic classification of Escherichia coli O157:H7 strains of human and bovine origin using a novel set of nucleotide polymorphisms

Abstract

Background: Cattle are a reservoir of Shiga toxin-producing Escherichia coli O157:H7 (STEC O157), and are known to harbor subtypes not typically found in clinically ill humans. Consequently, nucleotide polymorphisms previously discovered via strains originating from human outbreaks may be restricted in their ability to distinguish STEC O157 genetic subtypes present in cattle. The objectives of this study were firstly to identify nucleotide polymorphisms in a diverse sampling of human and bovine STEC O157 strains, secondly to classify strains of either bovine or human origin by polymorphism-derived genotypes, and finally to compare the genotype diversity with pulsed-field gel electrophoresis (PFGE), a method currently used for assessing STEC O157 diversity.

Results: High-throughput 454 sequencing of pooled STEC O157 strain DNAs from human clinical cases (n = 91) and cattle (n = 102) identified 16,218 putative polymorphisms. From those, 178 were selected primarily within genomic regions conserved across E. coli serotypes and genotyped in 261 STEC O157 strains. Forty-two unique genotypes were observed that are tagged by a minimal set of 32 polymorphisms. Phylogenetic trees of the genotypes are divided into clades that represent strains of cattle origin, or cattle and human origin. Although PFGE diversity surpassed genotype diversity overall, ten PFGE patterns each occurred with multiple strains having different genotypes.

Conclusions: Deep sequencing of pooled STEC O157 DNAs proved highly effective in polymorphism discovery. A polymorphism set has been identified that characterizes genetic diversity within STEC O157 strains of bovine origin, and a subset observed in human strains. The set may complement current techniques used to classify strains implicated in disease outbreaks.

Figure 1

Regions of the STEC O157 genome (Sakai reference strain) targeted for polymorphism validation. Black rectangles represent known phage or phage remnant integrations (Sakai prophages; Sp1-18, [4]) that were not queried for polymorphism validation. All other regions, totaling 87.8% of the genome, were included for polymorphism validation. The triangle points to nucleotide 1 of the Sakai genome sequence [GenBank:NC_002695].

Figure 2

Frequencies of 42 polymorphism-derived genotypes in STEC O157 strains of human and cattle origin.

Figure 3

Neighbor-joining tree of full-length polymorphism-derived genotypes. The triplicate sets of numbers on the tree represent bootstrap values from neighbor-joining, parsimony, and maximum-likelihood algorithms, respectively. Outer taxonomic unit genotype numbers correspond with genotype sequences recorded in Additional data file 3. The outer taxonomic units are color coded by genotype for the tir 255 T>A polymorphism and host origin. Roman numerals depict two subclades that account for 92% of the human STEC O157 strains genotyped in this study. The scale bar represents substitutions per site.

Figure 4

Median-joining network of polymorphism-derived genotypes tagged with a minimal set of 32 polymorphisms. Taxonomic unit genotype numbers correspond with genotype sequences recorded in Additional data file 4 and the units are color coded by genotype for the tir 255 T>A polymorphism and host origin.

Figure 5

Number of strains and PFGE patterns per genotype. Each stacked bar represents a total of the number of strains per genotype and the number of different PFGE patterns observed per genotype. The black portion of the bars represents the number of strains per genotype. The white speckled portions of the bars represent the number of different PFGE patterns observed per genotype.

Figure 6

Neighbor-joining tree placement of ten PFGE profiles onto corresponding polymorphism-derived genotypes. Each of the ten profiles was observed with more than one STEC O157 strain. The PFGE profile numbers match with those in Additional data file 1. Identical PFGE profiles that occurred with distantly related STEC O157 strains as determined with polymorphism-derived genotypes are highlighted in black.