Abstract

We employed a stepwise selection model for investigating the dynamics of antibiotic-resistant variants in Escherichia coli K-12 treated with increasing concentrations of ciprofloxacin (CIP). Firstly, we used Sanger sequencing to screen the variations in the fluoquinolone target genes, then, employed Illumina NGS sequencing for amplicons targeted regions with variations. The results demonstrated that variations G81C in gyrA and K276N and K277L in parC are standing resistance variations (SRVs), while S83A and S83L in gyrA and G78C in parC were emerging resistance variations (ERVs). The variants containing SRVs and/or ERVs were selected successively based on their sensitivities to CIP. Variant strain 1, containing substitution G81C in gyrA, was immediately selected following ciprofloxacin exposure, with obvious increases in the parC SRV, and parC and gyrA ERV allele frequencies. Variant strain 2, containing the SRVs, then dominated the population following a 20× increase in ciprofloxacin concentration, with other associated allele frequencies also elevated. Variant strains 3 and 4, containing ERVs in gyrA and parC, respectively, were then selected at 40× and 160× antibiotic concentrations. Two variants, strains 5 and 6, generated in the selection procedure, were lost because of higher fitness costs or a lower level of resistance compared with variants 3 and 4. For the second induction, all variations/indels were already present as SRVs and selected out step by step at different passages. Whatever the first induction or second induction, our results confirmed the soft selective sweep hypothesis and provided critical information for guiding clinical treatment of pathogens containing SRVs.

• Keywords: soft sweeps; stepwise selection; Escherichia coli; ciprofloxacin; fitness cost