Population bottlenecks shape laboratory evolution of piperacillin-tazobactam resistance in Klebsiella grimontii and reveal a shared within-patient evolutionary trajectory

Abstract

Laboratory-based experimental evolution is widely used to investigate how antimicrobial resistance (AMR) emerges and to identify resistance-associated trade-offs that could inform treatment strategies. However, there is limited understanding of how in vitro AMR evolution reflects the complexity of resistance evolution within the human host, where selective pressures, and therefore evolutionary pathways, are more variable. Here, we investigated the effect of population bottleneck size and growth environment on the evolution of piperacillin-tazobactam (TZP) resistance in Klebsiella grimontii and compared this to resistance evolution observed during a recurrent bloodstream infection. Three clonal K. grimontii isolates cultured from one patient over four months included a TZP-susceptible ancestor and a within-patient evolved TZP-resistant isolate. The susceptible ancestor was evolved under TZP selection using either a small 0.1% bottleneck or a larger 5% bottleneck, and under a second environment, LB supplemented with 5% sheep blood, using a 0.1% bottleneck. Evolved isolates were assessed for TZP susceptibility, {beta}-lactamase activity, fitness, and genomic changes. A single nucleotide polymorphism (SNP) in the promoter region of the chromosomally located {beta}-lactamase gene blaOXY-6-4 was identified in the within-patient evolved isolate and was replicated in all 0.1% bottleneck lineages across both environments. In contrast, the larger 5% bottleneck lineages exhibited greater phenotypic variation and genetic diversity, including multiple blaOXY-6-4 promoter variants and variable TZP MICs. These findings show that laboratory evolution can reproduce key within-patient resistance mechanisms, but that bottleneck size strongly shapes the resistance phenotypes and mutational landscapes observed in vitro. ImportanceAdaptive laboratory evolution is increasingly used to predict how antimicrobial resistance emerges and to identify trade-offs associated with resistance acquisition that could inform future treatment strategies. Here, we directly compared piperacillin-tazobactam resistance evolution in the laboratory with resistance that emerged within a patient during a recurrent bloodstream infection. We show that a small population bottleneck reproducibly selected the same blaOXY-6-4 promoter mutation observed in the patient, whereas a larger bottleneck produced more diverse evolutionary outcomes. These findings build on previous work showing that experimental conditions shape laboratory evolution outcomes and highlight population bottleneck size as an important experimental parameter when designing laboratory evolution studies that intend to model clinically relevant resistance evolution.