Abstract
Antimicrobial resistance has renewed interest in bacteriophage therapy, yet bacterial evolution frequently undermines treatment efficacy. Combination phage therapy is commonly implemented as simultaneous phage cocktails, but whether this is optimal remains in question. Here, we experimentally compared simultaneous versus sequential administration of two phages, an evolved λ called 'λtrn' and T2, on Escherichia coli K-12 under controlled laboratory conditions. Across replicated experiments, treatment outcome depended strongly on delivery strategy, dosing order, and timing. Contrary to expectations, sequential delivery consistently achieved greater and more sustained bacterial suppression than simultaneous cocktails, although only when T2 initiated the sequence. Phenotypic assays revealed that treatment differences were driven by the accessibility and timing of cross-resistance evolution. λ-first treatments rapidly selected for cross-resistant bacteria prior to exposure to the second phage, rendering subsequent treatment ineffective. In contrast, T2-first sequential treatments delayed or limited cross-resistance and frequently produced single-phage resistance or collateral sensitivity. Cocktail treatments showed intermediate dynamics, with cross-resistance evolving more slowly but consistently. Whole genome sequencing identified distinct genetic routes to cross-resistance, including regulatory mutations in envZ affecting expression of the phage receptor OmpF, as well as envelope-modifying, mucoidy-associated mutations. Engineering envZ mutations into unevolved backgrounds confirmed the mutation's sufficiency to confer low-cost cross-resistance. Together, these results demonstrated that phage therapy efficacy depended not only on phage composition but on how selection pressures were ordered in time, highlighting evolutionary steering as a powerful principle for multi-phage therapy design.