MCR-3 and MCR-9 confer species-specific increases in colistin MICs.

For the clinical treatment of infections with Enterobacteriaceae, the development of resistance to last-resort antimicrobials like colistin is of concern, as it limits treatment options. Since 2015, ten families of mobile colistin resistance (mcr) genes have been discovered; however, our understanding of MCR remains limited because (i) mcr-1 is the most studied variant, (ii) mcr variants are rarely compared and (iii) mcr variants are primarily studied in Escherichia coli, even though studies have shown that mcr variants can confer significantly different phenotypes and that lipid A (i.e. the target of colistin and mcr-modification) structurally differs between bacterial species. To fill this gap, we examined how mcr-3 and mcr-9, as two less frequently studied mcr variants, impact the colistin resistance of laboratory-adapted and real-world (i.e. clinical or food) isolates of E. coli, Salmonella enterica and Klebsiella pneumoniae. We found that while mcr-3 consistently conferred colistin resistance to all tested strains, mcr-9 only conferred resistance to two strains. The fold-change in colistin minimum inhibitory concentrations was significantly impacted by both the bacterial species and the mcr variant but not by whether a strain was a laboratory-adapted or real-world isolate. Overall, our results suggest that while laboratory-adapted strains may provide a good estimate of mcr-mediated colistin resistance of real-world isolates, findings of mcr-mediated phenotypes in one bacterial species should not be extrapolated to another.