Experimental evolution under colistin selection increases colistin minimum inhibitory concentration (MIC) without detectable collateral MIC shifts in Escherichia coli ATCC 25922

Abstract

Abstract Colistin is a last-resort antimicrobial in human and veterinary medicine, yet it remains unclear whether prolonged exposure can drive collateral resistance beyond polymyxins. Here, we used a spatial experimental-evolution platform, the Microbial Evolution and Growth Arena (MEGA)-plate, to examine adaptation of Escherichia coli ATCC 25922 to stepwise colistin selection in three independent runs. Endpoint isolates from all three runs ( n  = 3 isolates per exposure zone) were phenotyped, whereas one randomly selected representative isolate per exposure zone (0× , 1× , 10× , 100× , and 1000×) was sequenced because MIC profiles were identical across runs. Colistin MIC increased from 0.5 to 64 μg/mL across the selection gradient, while MICs for the other tested agents remained unchanged. Across the sequenced representatives, no plasmid-mediated colistin resistance determinants were detected and the overall set of CARD strict hits remained unchanged. Comparative genomics identified only limited coding changes with potential relevance to envelope-associated adaptation, notably an ftsI missense variant in the 100× and 1000× representatives and an ompC synonymous variant of uncertain functional significance in the 1× representative. Relative to the untreated representative, exposed isolates differed by only a small number of baseline-unique high-confidence annotated coding variants. These data support a compound-restricted phenotypic response in this single-strain model under the conditions tested. However, because genomic sampling was limited to one endpoint representative per exposure zone and no transcriptomic or functional assays were performed, broader regulatory adaptation or efflux-related responses cannot be excluded.