Insights into the acquisition of the pks island and production of colibactin in the Escherichia coli population

Abstract

The <i>pks</i> island codes for the enzymes necessary for synthesis of the genotoxin colibactin, which contributes to the virulence of <i>Escherichia coli</i> strains and is suspected of promoting colorectal cancer. From a collection of 785 human and bovine <i>E. coli</i> isolates, we identified 109 strains carrying a highly conserved <i>pks</i> island, mostly from phylogroup B2, but also from phylogroups A, B1 and D. Different scenarios of <i>pks</i> acquisition were deduced from whole genome sequence and phylogenetic analysis. In the main scenario, <i>pks</i> was introduced and stabilized into certain sequence types (STs) of the B2 phylogroup, such as ST73 and ST95, at the <i>asnW</i> tRNA locus located in the vicinity of the yersiniabactin-encoding High Pathogenicity Island (HPI). In a few B2 strains, <i>pks</i> inserted at the <i>asnU</i> or <i>asnV</i> tRNA loci close to the HPI and occasionally was located next to the remnant of an integrative and conjugative element. In a last scenario specific to B1/A strains, <i>pks</i> was acquired, independently of the HPI, at a non-tRNA locus. All the <i>pks</i>-positive strains except 18 produced colibactin. Sixteen strains contained mutations in <i>clbB</i> or <i>clbD,</i> or a fusion of <i>clbJ</i> and <i>clbK</i> and were no longer genotoxic but most of them still produced low amounts of potentially active metabolites associated with the <i>pks</i> island. One strain was fully metabolically inactive without <i>pks</i> alteration, but colibactin production was restored by overexpressing the ClbR regulator. In conclusion, the <i>pks</i> island is not restricted to human pathogenic B2 strains and is more widely distributed in the <i>E. coli</i> population, while preserving its functionality.