Genomic epidemiology reveals geographical clustering of multidrug-resistant Escherichia coli sequence type (ST)131 associated with bacteraemia in Wales, United Kingdom

Abstract

Increasing resistance to third-generation cephalosporins (3GCs) threatens public health, as these antimicrobials are prescribed as empirical therapies for systemic infections caused by Gram-negative bacteria. Resistance to 3GCs in urinary tract infections (UTIs) and bacteraemia is associated with the globally disseminated, multidrug-resistant, uropathogenic Escherichia coli sequence type (ST)131. This study combines the epidemiology of E.coli blood culture surveillance with whole-genome sequencing (WGS) to investigate ST131 associated with bacteraemia in Wales between 2013 and 2014. This population-based prospective genomic analysis investigated temporal, geographic, and genomic risk factors. To identify spatial clusters and lineage diversity, we contextualised 142 genomes collected from twenty hospitals, against a global ST131 population (n=181). All three major ST131 clades are represented across Wales, with clade C/H30 predominant (n=102/142, 71.8%). Consistent with global findings, Welsh strains of clade C/H30 contain β-lactamase genes from the blaCTX-M-1 group (n=65/102, 63.7%), which confers resistance to 3GCs. In Wales, the majority of clade C/H30 strains belonged to sub-clade C2/H30Rx (n=88/151, 58.3%), whereas sub-clade C1/H30R strains were less common (n=14/67, 20.9%). A sub-lineage unique to Wales was identified within the C2/H30Rx sub-clade (named GB-WLS.C2/H30Rx) and is defined by six non-recombinogenic single-nucleotide polymorphisms (SNPs), including a missense variant in febE (ferric enterobactin transport protein) and fryC (fructose-like permease IIC component), and the loss of the capsular biosynthesis genes encoding the K5 antigen. Bayesian analysis predicted that GB-WLS.C2/H30Rx diverged from a common ancestor (CA) most closely related to a Canadian strain between 1998 and 1999. Further, our analysis suggests a descendent of GB-WLS.C2/H30Rx arrived through an introduction to North Wales circa 2002, spread and persists in the geographic region, causing a cluster of cases (CA emerged circa 2009) with a maximum pair-wise distance of 30 non-recombinogenic SNPs. This limited genomic diversity likely depicts local transmission within the community in North Wales. This investigation emphasises the value of genomic epidemiology, allowing detection of suspected transmission clusters and the spread of genetically similar/identical strains in local areas. These analyses will enable targeted and timely public health interventions.