Mutations in two-component signaling systems drive experimental evolution of tigecycline and colistin resistance in <i>Acinetobacter baumannii</i>

The treatment of infections by <i>Acinetobacter baumannii</i>, a clinically significant nosocomial gram-negative bacterial pathogen, is hampered by antibiotic resistance, which is exacerbated by its exceptional genetic plasticity. Knowledge of the dynamics and mechanisms underlying the acquisition of antibiotic resistance is essential for the proper stewardship of their utilization. Here, we used a continuous culture device (morbidostat) to characterize the evolutionary trajectories of two <i>A. baumannii</i> strains in response to the increasing pressure of last-resort drugs, tigecycline and colistin. This approach allows us to confidently and comprehensively map resistance-driving mutations while circumventing both the "driver vs passenger" uncertainty and "selection bottleneck" limitations characteristic of clinical isolate analysis and conventional laboratory evolution, respectively. Tigecycline resistance predominantly occurred through the combination of missense mutations in the <i>adeSR</i> two-component system and disruptive events in the S-adenosyl methionine (SAM)-dependent methyltransferase, <i>trm</i>, while colistin resistance predominantly occurred through missense mutations in the gene cluster responsible for lipid A phosphoethanolamine modification, <i>pmrCAB</i>. Mapping of these mutational events over numerous publicly available <i>A. baumannii</i> genomes identified a relatively low prevalence of resistance to these two drugs. This work represents an initial step toward predictive resistomics of <i>A. baumannii</i>, leveraging gene-level genomic variations in addition to the conventional approaches based on the presence or absence of antibiotic resistance genes.