Second-site capsule-locus mutations, not Ade efflux disruption, determine phage susceptibility across an Acinetobacter baumannii transposon mutant panel

Resistance-Nodulation-Division (RND) efflux pumps have been implicated in phage adsorption in several Gram-negative bacteria, suggesting that AdeABC, AdeFGH and AdeIJK, the three characterised RND efflux systems of multidrug-resistant Acinetobacter baumannii, could play a similar role. We tested this hypothesis using five phages (Equilibrator, Taro, AT01, BB02, and AC03) and a panel of 13 A. baumannii strains, including nine AB5075 transposon mutants affecting AdeABC, AdeFGH, or AdeIJK. Host-range testing, efficiency-of-plating assays, capsule-associated density measurements, antimicrobial susceptibility testing, and whole-genome sequencing were performed to assess the relationship between genotype and phage infection. Overall, the data did not support a role for the Ade components as receptors. Phages Equilibrator, Taro and AT01 formed plaques on every strain, and the restricted host ranges of BB02 and AC03 did not correspond to disruption of a particular efflux system. Sequencing instead identified second-site mutations in the KL25 capsule locus in six of the ten AB5075 strains. These included a 1,998-bp deletion affecting wzc, substitutions in mnaA and mnaB, an ISAba13 insertion in a putative glycosyltransferase and a subpopulation-associated frameshift in wzx. AC03 formed plaques on all six strains with capsule-locus changes but on none of the five strains without them. Capsule dependence differed among the other phages: relative to the parental AB5075, infection by these phages, Equilibrator, Taro and BB02, was restricted on the capsule-deficient wzc::T26 strain, whereas AT01 efficiency was only slightly reduced. Insertion-site mapping also distinguished a near-terminal adeK insertion that retained the parental antimicrobial phenotype from an early adeK insertion associated with loss of AdeIJK function. Together, these findings identify capsule-locus variation as the strongest correlation with phage susceptibility in this panel and demonstrate the importance of verifying mutant genomes before assigning phenotypes to their annotated insertions.