Population structure analysis and laboratory monitoring of Shigella by core-genome multilocus sequence typing

Iman Yassine,François-Xavier Weill,Maria Pardos De La Gandara,Nicholas R Thomson,Fouad Dabboussi,Dominique Clermont,Monique Lejay-Collin,Isabelle Carle,Rayane Rafei,Sophie Lefèvre,Elisabeth E Hansen,Laëtitia Fabre,Corinne Ruckly

doi:10.1038/s41467-022-28121-1

Abstract

The laboratory surveillance of bacillary dysentery is based on a standardised Shigella typing scheme that classifies Shigella strains into four serogroups and more than 50 serotypes on the basis of biochemical tests and lipopolysaccharide O-antigen serotyping. Real-time genomic surveillance of Shigella infections has been implemented in several countries, but without the use of a standardised typing scheme. Here, we study over 4000 reference strains and clinical isolates of Shigella, covering all serotypes, with both the current serotyping scheme and the standardised EnteroBase core-genome multilocus sequence typing scheme (cgMLST). The Shigella genomes are grouped into eight phylogenetically distinct clusters, within the E. coli species. The cgMLST hierarchical clustering (HC) analysis at different levels of resolution (HC2000 to HC400) recognises the natural population structure of Shigella. By contrast, the serotyping scheme is affected by horizontal gene transfer, leading to a conflation of genetically unrelated Shigella strains and a separation of genetically related strains. The use of this cgMLST scheme will facilitate the transition from traditional phenotypic typing to routine whole-genome sequencing for the laboratory surveillance of Shigella infections.

Highlights

The laboratory surveillance of bacillary dysentery is based on a standardised Shigella typing scheme that classifies Shigella strains into four serogroups and more than 50 serotypes on the basis of biochemical tests and lipopolysaccharide O-antigen serotyping
Modern population genetics methods based on bacterial DNA sequences, such as multilocus sequence typing (MLST) analysis, and, more recently, core-genome single-nucleotide variant analysis, have shown that shigellae form distinct lineages within the species E. coli, from which they emerged following the acquisition of a large virulence plasmid (VP) enabling the bacterium to invade intestinal cells[8–11]
The size of these rfb clusters ranges from 9 kb to 17 kb[23]. Consistent with their phylogenetic relationships, the rfb clusters of Shigella serotypes are identical, highly similar to, or adapted from E. coli O-antigen gene clusters, accounting for the many cross-agglutinations observed with the serotyping scheme[23]. We extended this analysis to populations not previously considered (Fig. 6) and showed that similarity between E. coli and Shigella rfb DNA sequences was identified in S. dysenteriae 8 and E. coli a S3 HC2000_192

Summary

Introduction

The laboratory surveillance of bacillary dysentery is based on a standardised Shigella typing scheme that classifies Shigella strains into four serogroups and more than 50 serotypes on the basis of biochemical tests and lipopolysaccharide O-antigen serotyping. Modern population genetics methods based on bacterial DNA sequences, such as multilocus sequence typing (MLST) analysis (which analyses the allelic profiles of — generally seven — housekeeping genes), and, more recently, core-genome single-nucleotide variant (cgSNV) analysis, have shown that shigellae form distinct lineages within the species E. coli, from which they emerged following the acquisition of a large virulence plasmid (VP) enabling the bacterium to invade intestinal cells[8–11] In parallel, these host-restricted pathogens converged independently on the Shigella phenotype (non-motility, no decarboxylation of lysine, no use of citrate and malonate, and other characteristics, as reported by Pupo and coworkers). A set of 22 accessory genome genes was recently used to reassign Shigella serotypes to eight clusters[16]

Methods

Results

Conclusion