List of bacterial strains used in the study.

Pathways of intrinsic resistance in bacteria are promising targets for novel antibiotics and resistance breakers. Here, we used a genome-wide screen to identify single gene knockouts of Escherichia coli that were hypersusceptible to trimethoprim and chloramphenicol, two chemically diverse broad-spectrum antibiotics. Among the hits from our screen, knockouts of acrB, an efflux pump, and rfaG or lpxM, both involved in cell envelope biogenesis, were hypersensitive to multiple antimicrobials and could sensitize genetically resistant E. coli strains to antibiotics. Using experimental evolution under trimethoprim pressure, we show that high drug selection regimes drove these knockouts to extinction more frequently than wild type. Among them, ΔacrB was most compromised in its ability to evolve resistance, establishing it as a promising target for “resistance proofing.” At a sub-inhibitory trimethoprim concentration, however, all three knockouts adapted to the antibiotic and consequently recovered from hypersensitivity, albeit to different extents. This recovery was driven by mutations in drug-specific resistance pathways, rather than compensatory evolution, frequently involving upregulation of the drug target. Notably, resistance-conferring mutations could by-pass defects in cell wall biosynthesis more effectively than efflux even though resistant mutations did not directly engage either pathway. Since inhibiting drug-efflux emerged as a better strategy, we tested the ability of chlorpromazine, an efflux pump inhibitor (EPI), to resistance proof E. coli against trimethoprim. While qualitatively similar in the short term, genetic and pharmacological inhibition differed dramatically on an evolutionary time scale due to evolution of resistance to the EPI. Further, adaptation to the EPI-antibiotic pair also led to multidrug adaptation. The lack of concordance between genetic and pharmacological inhibition revealed a crucial lacuna in our understanding of the mutational repertoires that facilitate adaptation to antibiotics in bacteria. We propose that while intrinsic resistance mechanisms are effective targets for antibiotic sensitization, rapid evolutionary recovery may significantly limit their utility.