<i>In vitro</i> and <i>in vivo</i> selection and cost of bacteriophage resistance on natural <i>Escherichia coli</i>

Bacteriophages are a promising tool for treating bacterial infections, given the rise and spread of antibiotic resistances. However, phage-resistant bacteria can emerge during treatment, jeopardizing the success of therapy<i>. In vitro</i> studies with model organisms have provided valuable insights into the mechanisms by which phage resistance can evolve. However, the relevance of these findings often remains unclear. Here, we investigate the selection of phage-resistant variants and the cost of phage resistance <i>in vitro</i> and in the murine gut using a clinically relevant <i>Escherichia coli</i> K1 strain and a strain-specific phage cocktail. By performing experimental evolution studies in both settings, we obtained different phage-resistant <i>E. coli</i> mutants. Genome resequencing identified lipopolysaccharide (LPS) and the K1 capsule as bacterial surface structures altered in phage-resistant mutants. Targeted deletions of <i>waaO</i>, encoding an ɑ-1,3 glucosyltransferase, involved in the synthesis of the R core of LPS, a gene encoding a predicted O-antigen ligase and <i>emrR</i> involved in capsule gene regulation were generated and confirmed their role in phage-resistance. <i>Escherichia coli</i> mutants deficient in LPS or capsule showed a growth advantage <i>in vitro</i> when exposed to phages but LPS-deficient mutants exhibited severely attenuated growth in the murine gut, even in the presence of phages. Our observations add to the evidence that bacteria in the intestinal environment face a high cost of phage resistance conferred by cell surface alteration, which is not apparent in nutrient-rich culture media. Therefore, it is crucial to carefully consider the context in which phage cocktails are tested, particularly when studying phage efficacy and evolution of phage resistance.