Acquisition of plasmids from Shiga toxin-producingEscherichia colistrains had low or neutral fitness cost on commensalE. coli.

Abstract

Although it has been hypothesized that the acquisition of plasmids-especially those bearing virulence factors and antimicrobial resistance genes-increases the energetic burden and reduces the fitness of a bacterium in general, some results have challenged this view, showing little or no effect on fitness after plasmid acquisition, which may lead to change in the view that there are evolutionary barriers for a wide spread of such plasmids among bacteria. Here, to evaluate the fitness impact of plasmid-encoded antibiotic resistance and virulence genes, plasmids from O26:H11, O111:H8, and O118:H16 Shiga toxin-producing Escherichia coli (STEC) human and bovine isolates were transferred to the non-virulent E. coli HS and K-12 MG1655 strains. Sequencing and PCR were used to characterize plasmids, and to identify the presence of antimicrobial resistance and/or virulence genes. The fitness impact of plasmids encoding virulence and antimicrobial resistance upon bacterial hosts was determined by pairwise growth competition. Plasmid profile analysis showed that STEC strains carried one or more high and low molecular weight plasmids belonging to the B/O, F, I, K, P, Q, and/or X incompatibility groups encoding virulence genes (SPATE-encodinggenes) and/or antimicrobial resistance genes (aadA1, strAB, tetA, and/or tetB). Competition experiments demonstrated that the biological cost of carriage of these plasmids by the commensal E. coli strain HS or the laboratory strain E. coli K-12 MG1655 was low or non-existent, ranging from - 4.7 to 5.2% per generation. This suggests that there are few biological barriers-or, alternatively, it suggests that there are biological barriers that we were not able to measure in this competition model-against the spread of plasmid encoding virulence and resistance genes from STEC to other, less pathogenic E. coli strains. Thus, our results, in opposition to a common view, suggest that the acquisition of plasmids does not significantly affect the bacteria fitness and, therefore, the theorized plasmid burden would not be a significant barrier for plasmid spread.