Pathogenomes and virulence profiles of representative big six non-O157 serogroup Shiga toxin-producing Escherichia coli

Abstract

Shiga toxin (Stx)-producing Escherichia coli (STEC) of non-O157:H7 serotypes are responsible for global and widespread human food-borne disease. Among these serogroups, O26, O45, O103, O111, O121, and O145 account for the majority of clinical infections and are colloquially referred to as the "Big Six." The "Big Six" strain panel we sequenced and analyzed in this study are reference type cultures comprised of six strains representing each of the non-O157 STEC serogroups curated and distributed by the American Type Culture Collection (ATCC) as a resource to the research community under panel number ATCC MP-9. The application of long- and short-read hybrid sequencing yielded closed chromosomes and a total of 14 plasmids of diverse functions. Through high-resolution comparative phylogenomics, we cataloged the shared and strain-specific virulence and resistance gene content and established the close relationship of serogroup O26 and O103 strains featuring flagellar H-type 11. Virulence phenotyping revealed statistically significant differences in the Stx-production capabilities that we found to be correlated to the strain's individual stx-status. Among the carried Stx_1a, Stx_2a, and Stx_2d phages, the Stx_2a phage is by far the most responsive upon RecA-mediated phage mobilization, and in consequence, stx_2a + isolates produced the highest-level of toxin in this panel. The availability of high-quality closed genomes for this "Big Six" reference set, including carried plasmids, along with the recorded genomic virulence profiles and Stx-production phenotypes will provide a valuable foundation to further explore the plasticity in evolutionary trajectories in these emerging non-O157 STEC lineages, which are major culprits of human food-borne disease.

Keywords: Shiga toxin (Stx)-producing Escherichia coli (STEC); non-O157 big six serogroups; phylogenomics; virulence phenotyping; whole genome sequencing and typing (WGST).

Figure 1

Comparison of ATCC MP-9 panel genomes BRIG comparison of six sequenced strains, along with Escherichia coli strains O157:H7 EC4115 and strain K-12 substrain MG1665, referenced to the 5,840,137 bp chromosome of BAA-2196 (O26:H11). CDS are presented on the +/−strands as blue arrows and functional annotations for virulence genes and other loci of interest are highlighted as shown in the legend. Query genomes are color-coded, and the order plotted in the circle reflects the inferred phylogenomic relationships.

Figure 2

Comparison of a shared colicin plasmid BRIG comparison of a shared colicinogenic plasmid l present in serotypes O26:H11, O103:H11, and O111:H8 and referenced to the 6,673 bp plasmids of BAA-2196 pCol-O26-3. The plasmids are differentiated by a total of 30 SNPs and InDels. CDS are presented on the +/− strands as blue arrows. Query plasmids are plotted according to the strain’s inferred phylogenomic relationships.

Figure 3

Phylogenomic position of ATCC-MP9 strains The relatedness of panel strains, including O157:H7 strain EC4115, was determined using MLST in Ridom SeqSphere+: (A) targeted seven-gene MLST accessed in EnteroBase. Numbers on connecting branches indicate the number of genes with differing allele status, and (B) cgMLST-based phylogeny using the closed chromosome of E. coli strain K12 subst. MG1655 as seed. The shared gene inventory was determined at 4,304 genes, according to the inclusion/exclusion criteria of the SeqSphere+ Target Definer and is comprised of 3,148 core and 908 accessory loci. Colors denote ST-classifications established in (A).

Figure 4

Prevalence and distribution of plasmid-borne virulence determinants Percentage identities of virulence and antimicrobial resistance genes identified in VFDB and ResFinder are visualized in a heatmap. The Plasmid incompatibility group and predicted mobility were determined with MobTyper. The shared colicinogenic plasmid pCol present in the serotype O26:H11, O103:H11, and O111:H8 strains is indicated with a star.

Figure 5

Prevalence and distribution of chromosomal virulence determinants Percentage identities for each virulence gene identified in VFDB are visualized in a heatmap. The panel strains encode a total of 149 distinct virulence genes of which 113 are shared. Among these are toxin suballeles stx_1a, stx_2a, and stx_2d, the LEE genomic island, and siderophores, among others. The strain order reflects the inferred phylogenomic relationships. The hierarchical clustering of virulence genes based on their pair-wise distance is shown on the left.

Figure 6

Comparison of Stx-prophages BLASTn-based comparison of architectures and content of carried ΦStx_2a-, ΦStx_1a-, and ΦStx_2d-prophages. For uniform annotation, we inferred the annotation from the curated ΦStx_2a-prophage genome of strain BAA-2196 (O26:H11).

Figure 7

Comparison of LEE islands The complete LEE islands were extracted and compared in GeneSpy. Genes are colored according to their nucleotide homologies. The organization of the LEE1 to 5 operons is indicated above. The order of strains reflects their inferred phylogenomic position and is mirrored in the LEE island organization.

Figure 8

Variability in Stx-production The concentration of Stx₁ and Stx₂ produced under non-induced and MMC-induced phage mobilizing conditions was quantified by ELISA. Differences between the non-induced and MMC-induced phage mobilizing conditions for each strain were assessed using a two-way ANOVA, followed by Sidak’s multiple comparisons. Statistical significance is denoted as *p < 0.05; **p < 0.005; ***p < 0.0005; and ****p < 0.00005. For strain–strain comparison under MMC-induced conditions, differences in toxin concentration are indicated by letters (a–d), with “a” denoting the highest concentration group, in a descending order determined by a one-way ANOVA with Tukey’s multiple comparisons test.