Abstract
This study used a diverse collection of epidemiologically unrelated Acinetobacter baumannii isolates to compare the robustness of a multilocus sequence typing (MLST) scheme, based on conserved regions of seven housekeeping genes, gltA, gdhB, recA, cpn60, rpoD, gyrB, and gpi, with that of sequence-based typing of bla(OXA-51-like) genes (SBT-bla(OXA-51-like) genes). The data obtained by analysis of MLST and SBT-bla(OXA-51-like) genes were compared to the data generated by pulsed-field gel electrophoresis (PFGE). The topologies of the phylogenetic trees generated for the gyrB and gpi genes showed evidence of recombination and were inconsistent with those of the trees generated for the other five genes. MLST identified 24 sequence types (STs), of which 19 were novel, and 5 novel alleles. Clonality was demonstrated by eBURST analysis and standardized index of association values of > 1 (P < 0.001). MLST data revealed that all isolates harboring the major bla(OXA-51-like) alleles OXA-66, OXA-69, and OXA-71 fell within the three major European clonal lineages. However, the MLST data were not always in concordance with the PFGE data, and some isolates containing the same bla(OXA-51-like) allele demonstrated <50% relatedness by PFGE. It was concluded that the gyrB and gpi genes are not good candidates for use in MLST analysis and that a SBT-bla(OXA-51-like) gene scheme produced results comparable to those produced by MLST for the identification of the major epidemic lineages, with the advantage of having a significantly reduced sequencing cost and time. It is proposed that studies of A. baumannii epidemiology could involve initial screening of bla(OXA-51-like) alleles to identify isolates belonging to major epidemic lineages, followed by MLST analysis to categorize isolates from common lineages, with PFGE being reserved for fine-scale typing.