Derivation of Escherichia coli O157:H7 from its O55:H7 precursor

Abstract

There are 29 E. coli genome sequences available, mostly related to studies of species diversity or mode of pathogenicity, including two genomes of the well-known O157:H7 clone. However, there have been no genome studies of closely related clones aimed at exposing the details of evolutionary change. Here we sequenced the genome of an O55:H7 strain, closely related to the major pathogenic O157:H7 clone, with published genome sequences, and undertook comparative genomic and proteomic analysis. We were able to allocate most differences between the genomes to individual mutations, recombination events, or lateral gene transfer events, in specific lineages. Major differences include a type II secretion system present only in the O55:H7 chromosome, fewer type III secretion system effectors in O55:H7, and 19 phage genomes or phagelike elements in O55:H7 compared to 23 in O157:H7, with only three common to both. Many other changes were found in both O55:H7 and O157:H7 lineages, but in general there has been more change in the O157:H7 lineages. For example, we found 50% more synonymous mutational substitutions in O157:H7 compared to O55:H7. The two strains also diverged at the proteomic level. Mutational synonymous SNPs were used to estimate a divergence time of 400 years using a new clock rate, in contrast to 14,000 to 70,000 years using the traditional clock rates. The same approaches were applied to three closely related extraintestinal pathogenic E. coli genomes, and similar levels of mutation and recombination were found. This study revealed for the first time the full range of events involved in the evolution of the O157:H7 clone from its O55:H7 ancestor, and suggested that O157:H7 arose quite recently. Our findings also suggest that E. coli has a much lower frequency of recombination relative to mutation than was observed in a comparable study of a Vibrio cholerae lineage.

Figure 1. Alignment of the genomes of CB9615, Sakai and EDL933.

Scales are Mbp. The grey and yellow shaded regions represent segments present in all strains, with inverted segments shaded in yellow. Purple and cyan boxes within the grey or yellow regions represent the indels of phages or phage-like elements (insertions and deletions respectively) defined in Table S9. Red and green boxes represent other major indels (insertions and deletions respectively) that involve the changes of the gene numbers in Table S10, with the indel numbers also shown as defined in Table S9. The orange boxes indicate the E. coli O157:H7 O-antigen segment gained by recombination. Figure S1 is a greatly expanded version of this figure showing individual genes.

Figure 2. Tree showing the relationships of CB9615 and 2 O157:H7 strains.

The tree topography is taken from the alignment of 26 completed genomes (Figure 3) and Whittam . For each lineage the number of mutations (including small indels), recombination events and insertion or deletion events (large indels) are shown in a grid, as specified with the key. Mutations are shown as intergenic, other non-coding, non-synonymous or synonymous SNPs (igSNPs, ncSNPs, nsSNP, sSNPs), small insertions and small deletions or indels if not differentiable. Large indels are separated into insertions or deletions where possible. Events allocated to the divergence between CB9615 and O157:H7, or between Sakai and EDL933, respectively, but not to either lineage, are shown in the grids between the two lineages. The branch point estimates for group B including strain 493-89), G5101 and F6141, and clusters 1 (Strains 14359 and 87-14) and 2 (86-24) are marked with dotted lines on the O157:H7 lineage, and TB182A on the O55:H7 lineage. The distribution of SNPs along that lineage is based on reanalysis of data from Zhang et al. and Leopold et al. .

Figure 3. Maximum likelihood phylogenetic tree of 26 Escherichia coli and Shigella strains.

The phylogenetic tree of the Escherichia core genome genes was constructed from the concatenated alignments of the 2034 genes in the core genome of the E. coli/Shigella genomes. The closely related species, E. fergusonii (CU928158), was chosen to root the tree.

Figure 4. Tree showing the relationships of UTI89, S88 and APEC 01 ExPEC strains.

The tree topography is for mutational SNPs as allocated by the virtual outgroup analysis (Tables S7 and S8). For each lineage the number of mutations and recombination events is shown in grids as for Figure 1.