Defining the phylogenomics of Shigella species: a pathway to diagnostics

Abstract

Shigellae cause significant diarrheal disease and mortality in humans, as there are approximately 163 million episodes of shigellosis and 1.1 million deaths annually. While significant strides have been made in the understanding of the pathogenesis, few studies on the genomic content of the Shigella species have been completed. The goal of this study was to characterize the genomic diversity of Shigella species through sequencing of 55 isolates representing members of each of the four Shigella species: S. flexneri, S. sonnei, S. boydii, and S. dysenteriae. Phylogeny inferred from 336 available Shigella and Escherichia coli genomes defined exclusive clades of Shigella; conserved genomic markers that can identify each clade were then identified. PCR assays were developed for each clade-specific marker, which was combined with an amplicon for the conserved Shigella invasion antigen, IpaH3, into a multiplex PCR assay. This assay demonstrated high specificity, correctly identifying 218 of 221 presumptive Shigella isolates, and sensitivity, by not identifying any of 151 diverse E. coli isolates incorrectly as Shigella. This new phylogenomics-based PCR assay represents a valuable tool for rapid typing of uncharacterized Shigella isolates and provides a framework that can be utilized for the identification of novel genomic markers from genomic data.

FIG 1

A plot of unique gene clusters at different levels of identity. Coding regions for 69 Shigella genomes or 69 E. coli genomes were predicted with Prodigal (41). Coding regions were translated with BioPython (42) and concatenated. USEARCH (43) was then used to cluster all peptides at different levels of identity. The number of unique clusters at each identity threshold was plotted for both groups. The results demonstrate the smaller pan-genome size for Shigella compared to E. coli genomes.

FIG 2

Phylogenies inferred from a diverse set of E. coli and Shigella genomes (n = 336). (A) A phylogeny inferred from a concatenation of sequences from multilocus sequence typing markers (see Table S3 in the supplemental material) from the E. coli pubMLST system (34). Conserved sequences were extracted from BLAST (31) alignments and were aligned with MUSCLE (32). The phylogenetic tree was inferred with FastTree2 (33), with 1,000 bootstrap replicates. (B) A phylogeny was inferred with FastTree2 from a concatenation of sequence markers (see Table S2) used in a previous study of Shigella evolution (15). (C) A phylogenetic tree of E. coli and Shigella isolates using whole-genome sequence data. Conserved genomic fragments were first identified in a core set of 40 E. coli genomes aligned with Mugsy (29). Conserved genomic regions were extracted by BLASTN, aligned with MUSCLE, and concatenated. A tree was then inferred on this alignment with FastTree2, with 1,000 bootstrap replicates.

FIG 3

Whole-genome phylogeny. A whole-genome phylogenetic tree of 69 sequenced Shigella genomes, including 55 sequenced as part of this study, is shown. The tree was inferred with FastTree2 (33) on a Mugsy (29) whole-genome alignment, as has been done previously (17). Labels at branch nodes indicate the clade-naming convention developed in this study. Bootstrap support values from 100 replicates are shown at nodes. This tree demonstrates that Shigella genomes group into 5 monophyletic lineages (S1 to S5) and that there is a mixing of species, based on serology, in clades S1 and S3.

FIG 4

Shigella biomarker development. A gel electrophoresis image of amplicons from the 5 major clades identified in this study (lanes 1 to 5) is shown; two bands, one genus targeted (ipaH3) and one clade targeted (S1 to S5), indicate a positive reaction. Lane 6 shows a coinfection reaction with all 5 clades, plus the universally conserved ipaH3 marker. Numbers on the left represent numbers of base pairs in the DNA ladder.