Canonical Single Nucleotide Polymorphisms (SNPs) for High-Resolution Subtyping of Shiga-Toxin Producing Escherichia coli (STEC) O157:H7

Abstract

The objective of this study was to develop a canonical, parsimoniously-informative SNP panel for subtyping Shiga-toxin producing Escherichia coli (STEC) O157:H7 that would be consistent with epidemiological, PFGE, and MLVA clustering of human specimens. Our group had previously identified 906 putative discriminatory SNPs, which were pared down to 391 SNPs based on their prevalence in a test set. The 391 SNPs were screened using a high-throughput form of TaqMan PCR against a set of clinical isolates that represent the most diverse collection of O157:H7 isolates from outbreaks and sporadic cases examined to date. Another 30 SNPs identified by others were also screened using the same method. Two additional targets were tested using standard TaqMan PCR endpoint analysis. These 423 SNPs were reduced to a 32 SNP panel with the almost the same discriminatory value. While the panel partitioned our diverse set of isolates in a manner that was consistent with epidemiological data and PFGE and MLVA phylogenies, it resulted in fewer subtypes than either existing method and insufficient epidemiological resolution in 10 of 47 clusters. Therefore, another round of SNP discovery was undertaken using comparative genomic resequencing of pooled DNA from the 10 clusters with insufficient resolution. This process identified 4,040 potential SNPs and suggested one of the ten clusters was incorrectly grouped. After its removal, there were 2,878 SNPs, of which only 63 were previously identified and 438 occurred across multiple clusters. Among highly clonal bacteria like STEC O157:H7, linkage disequilibrium greatly limits the number of parsimoniously informative SNPs. Therefore, it is perhaps unsurprising that our panel accounted for the potential discriminatory value of numerous other SNPs reported in the literature. We concluded published O157:H7 SNPs are insufficient for effective epidemiological subtyping. However, the 438 multi-cluster SNPs we identified may provide the additional information required.

Enterohemorrhagic Escherichia coli O157 subclade 8b strains in Chiba Prefecture, Japan, produced larger amounts of Shiga toxin 2 than strains in subclade 8a and other clades

“…Since a SNP set for classifying clade 8 strains into subclades 8a and 8b was not available [ 20 ], a SNP set for this study was constructed as follows. Griffing et al [ 28 ] developed a SNP set for epidemiological molecular subtyping of O157 strains using SNPs at 32 loci. This typing divided clade 8 strains into two subgroups.…”

“…This typing divided clade 8 strains into two subgroups. Several SNPs were selected from 32 SNPs used by Griffing et al [ 28 ] to construct the SNP set for this study. To select these several SNPs, the SNPs at 32 loci used by Griffing et al [ 28 ] were investigated using clade 8 strains with various genetic features (i.e., the various distribution of IS 629 ).…”

“…Several SNPs were selected from 32 SNPs used by Griffing et al [ 28 ] to construct the SNP set for this study. To select these several SNPs, the SNPs at 32 loci used by Griffing et al [ 28 ] were investigated using clade 8 strains with various genetic features (i.e., the various distribution of IS 629 ). In this study, a minimum spanning tree (MST) was reconstructed from the distribution of IS 629 in the epidemiologically unlinked clade 8 strains ( S2 Fig ).…”

“…Whole-genome sequence (WGS) analyses were conducted on these six strains, and the SNPs at the 32 loci used by Griffing et al [ 28 ] were determined as previously described [ 29 , 30 ]. Briefly, sequence libraries were prepared with Nextera XT DNA Sample Prep Kits (Illumina, Inc., San Diego, CA, USA), and 100 cycles of dual-index paired-end sequencing were carried out using the Illumina HiSeq2500 System (Illumina, Inc.).…”

Whole Genome Sequencing for Genomics-Guided Investigations of Escherichia coli O157:H7 Outbreaks

“…No specific virulence pattern distinguished the outbreak strains from sporadic strains (Berenger et al, 2015 ). Recent studies have expanded WGS typing to globally distributed strains and identified geographical genomic structuring based on distribution of stx -converting phage integration sites and SNPs (Mellor et al, 2015 ; Strachan et al, 2015 ) and provided a more detailed subtyping of E. coli O157:H7 (Griffing et al, 2015 ). However, clarity can also be gained by comparing closely related isolates to each other, rather than to reference strains (Leopold et al, 2009 ; Turabelidze et al, 2013 ).…”