Comment on: AbaR4 replaces AbaR3 in a carbapenem-resistant Acinetobacter baumannii isolate belonging to global clone 1 from an Australian hospital

Abstract

Sir, I read with great interest the valuable article by Hamidian and Hall regarding the AbaR4 genomic island in D36, an Acinetobacter baumannii isolate belonging to global clone (GC) 1, and proposing a new family of transposons. However, I would like to make some remarks on the authors’ statement that D36 was of a rare sequence type (ST), ST247, and represented a new GC 1 lineage. In their study, the ST of A. baumannii strain D36 was determined by the Oxford multilocus sequence typing (MLST) scheme, which employs gltA, gyrB, gdhB, recA, cpn6, gpi and rpoD as housekeeping genes. As a result, strain D36 showed the allelic profile 10-12-4-11-4-58-5 and was assigned to ST247, slightly different from ST109 (10-12-4-11-4-9-5), which includes the GC 1 reference strain RUH875 (A297). Therefore, the difference between ST247 and ST109 was in their gpi alleles. However, in a recent study it was demonstrated that the gene gpi resided in a part of the A. baumannii genome prone to recombination and showed significant variations compared with other housekeeping genes, and a similar observation was also described in another study. Thus it appears to be problematic to include gpi in the MLST system and therefore the conclusion that ST109 and ST247 are two distinct lineages is also doubtful. Besides AbaR4-associated features, such as lack of Tn6018 and the resistance genes catA1, blaTEM, tet(A) and aphA1, strain D36 also showed other characteristics (aadB gene cassette and sul2 gene) different from other GC 1 strains included in the study by Hamidian and Hall. It has been suggested that the accessory genome (genes present in some isolates, but not in others) can be used as a tool for epidemiological investigation, and that the distinct features of strain D36 might support its role as the representative of a new lineage. On the one hand, it is difficult to select epidemiological markers from the hundreds of genes in the accessory genome, which varies between isolates. On the other hand, comparative genomic analysis has demonstrated that there are frequent homologous recombination events across up to 20% of GC 2 genomes, and the recombination also took place across a large phylogenetic distance, including GC 1. The horizontal transfer of aminoglycoside resistance genes and gene cassettes representing the variable region of integrons between GC 1 and 2 isolates from the Czech Republic has also been described. Therefore, the usefulness of the accessory genome in epidemiological research remains to be determined. As a result, the taxonomic position of strain D36 should be determined by a new scheme, for example, the MStree analysis-based MLST system, and further evidence is required to support the conclusion that strain D36 belongs to a new GC 1 lineage.