Genetic and phenotypic analysis of the virulence plasmid of a non-Shigatoxigenic enteroaggregative Escherichia coli O104:H4 outbreak strain

Abstract

Enteroaggregative Escherichia coli O104:H4 is best known for causing a worldwide outbreak in 2011 due to acquisition of a Shiga-like toxin alongside traditional enteroaggregative virulence traits; however, whilst the 2011 outbreak strain has been well-studied, the virulence plasmid of O104:H4 has been subjected to far less experimental analysis. In this paper, we analyse the genetic and phenotypic contribution of the pAA virulence plasmid to a non-Shigatoxigenic O104:H4 strain (1070/13) that was nonetheless implicated in a substantial UK outbreak in 2013. We find that pAA1070 is 99.95% identical across 88% of the plasmid sequence to pTY2 from the 2011 outbreak strain and has a copy number of approximately 2-3 plasmid molecules per chromosome. We demonstrate that pAA1070 carries a functional CcdAB plasmid addiction system that only marginally impacts its stability under the conditions tested and that none of the other toxin-antitoxin systems encoded by the plasmid appear to be functional, though we note a surprisingly high stability of the plasmid in vitro regardless. We demonstrate the expected contribution of pAA1070 to intestinal cell adhesion but find that it does not contribute to biofilm formation. When assessing the impact of pAA1070 on motility, we discovered a region of the O104:H4 chromosome that can be excised, abolishing motility via truncation of the fliR gene. Ultimately this work contributes to our knowledge of enteroaggregative E. coli as an important pathovar in its own right and demonstrates the complexity but necessity of experimentally characterising genuine outbreak strains rather than domesticated laboratory strains in order to understand virulence phenotypes. IMPACT STATEMENTEnteroaggregative Escherichia coli causes severe food-borne illness in humans. One serotype, O104:H4, caused a worldwide outbreak in 2011 that infected almost 4,000 people and killed dozens. The 2011 strain contained two key virulence factors: the Shiga toxin encoded by a bacteriophage and a plasmid that encodes aggregative adherence fimbriae for adhesion to host cell surfaces. Whilst the Shiga toxin has been studied extensively, the virulence plasmid has been far less well-characterised. In this study, we characterise the plasmid of a non-Shigatoxigenic O104:H4 outbreak strain and find that it contributes to host cell adherence, as expected, and that it encodes a functional CcdAB toxin-antitoxin stability system. We also demonstrate that part of the strains chromosome can be spontaneously lost during growth in the laboratory, converting the organism from highly-motile to non-motile, and suggesting a future evolutionary route for the bacterium, as has occurred in similar enteric pathogens such as Shigella. In summary, this work contributes to our knowledge of often-neglected enteroaggregative E. coli as an important pathovar and highlights the necessity of experimentally characterising genuine outbreak strains rather than "domesticated" laboratory strains. Repositories/Data SummaryThe authors confirm all supporting data, code and protocols have been provided within the article or through supplementary data files. Hybrid whole-genome sequence data is available for strains 1070/13 (accession numbers CP171474, CP171475 and CP171476) and 1070/13 pAA- (accession numbers CP171472 and CP171473). Sequencing data for 1070/13 produced by Dallman et al. (2014) is available from the European Nucleotide Archive: sample accession SAMN02730196.