Supplemental Material for Phillips et al., 2026

The rapid emergence of antimicrobial resistance in bacterial pathogens threatens the efficacy of nearly all available antibiotics. One evolution-informed strategy to limit the emergence of resistance is the exploitation of collateral sensitivity, whereby resistance to one compound results in increased sensitivity to another. Here, we investigate the collateral sensitivity relationship between the cephalosporin antibiotic ceftazidime and the natural product borrelidin A in Escherichia coli. Previously, we evolved replicate populations of E. coli under four drug regimens: ceftazidime alone, ceftazidime with norfloxacin, and ceftazidime with either 32 or 128µM borrelidin A, and found that borrelidin A prevented the evolution of clinical ceftazidime resistance. Here, we isolated 36 evolved strains, nine per treatment, measuring their resistances and susceptibilities to additional β-lactam and quinolone antibiotics, their relative fitness, and performed whole-genome sequencing. Co-dosing with 128µM borrelidin A significantly reduced the evolution of ceftazidime resistance, while preserving fitness in the absence of drug. Overall co-dosed strains had reduced resistance and cross-resistance to all tested antibiotics. Whole-genome sequencing revealed that co-dosing suppressed the accumulation of mutations in known resistance-associated genes; however, in some cases, selected for a possible novel borrelidin A resistance mutation. Our results show that ceftazidime-borrelidin A co-dosing limits both resistance and cross-resistance through selection against costly resistance mutations. While borrelidin A itself is cytotoxic, our findings highlight the promise of targeting bacteria-specific vulnerabilities to curb the emergence of multidrug resistance. These findings contribute to the growing body of evidence supporting collateral sensitivity-informed approaches as practical strategies to mitigate antimicrobial resistance in bacterial populations.