Escherichia coli killing by epidemiologically successful sublineages of Shigella sonnei is mediated by colicins

Shigella sp. are enteric pathogens which causes >125 million cases of shigellosis annually. S. sonnei accounts for about a quarter of those cases and is increasingly prevalent in industrialising nations. Being an enteric pathogen, S. sonnei benefits from outcompeting gut commensals such as Escherichia coli to establish itself and cause disease. There are numerous mechanisms that bacterial pathogens use to outcompete its rivals including molecules called colicins. A Type 6 Secretion System (T6SS) was recently described as contributing to E. coli killing in S. sonnei.

We used Bulk Phenotyping of Epidemiological Replicates (BPER) which combined bacterial Genome Wide Association Studies (bGWAS) and high throughput phenotyping on a collection of S. sonnei surveillance isolates to identify the genetic features associated with E. coli killing and explore their relationship with epidemiological behaviour. We further explored the presence of colicins and T6SS components in the isolates using genomics, laboratory experimentation, and proteomics.

Our bGWAS analysis returned known and novel colicin and colicin related genes as significantly associated with E. coli killing. In silico analyses identified key colicin clusters responsible for the killing phenotype associated with epidemiologically successful sub-lineages. The killing phenotype was not associated with the presence of a T6SS. Laboratory analyses confirmed the presence of the key colicin clusters and that killing was contact-independent.

Colicins are responsible for E. coli killing by S. sonnei, not a T6SS. This phenotype contributes to shaping the observed epidemiology of S. sonnei and may contribute to its increasing prevalence globally. BPER is an epidemiologically relevant approach to phenotypic testing that enables the rapid identification of genetic drivers of phenotypic changes, and assessment of their relevance to epidemiology in natural settings.