In their recent paper, Kilic et al. (2) reported on a hospital outbreak of bloodstream infections caused by a non-glucose-acidifying Acinetobacter strain. The authors showed that this strain (AK001) was phenotypically most similar to Acinetobacter ursingii but had two properties atypical of this species: i.e., the ability to hemolyse sheep blood and the inability to grow on citrate. Additional taxonomic methods to clarify the taxonomic status of AK001 included 16S rRNA and rpoB gene sequence analysis and DNA-DNA reassociation. Although the results of the sequence analyses supported the close relationship of the strain to A. ursingii, the value of DNA-DNA reassociation between AK001 and the A. ursingii type strain was found to be 64.7 to 68.7%, which is slightly less than 70%, the recommended threshold for species delineation (6). The authors concluded that the strain represents a new species, for which the name “Acinetobacter septicus” was proposed.
As, in our opinion, the data provided by Kilic et al. (2) do not convincingly substantiate the proposal of a novel species, we analyzed strain AK001 using methods that had previously been used to resolve Acinetobacter taxonomy (3). AK001 was obtained by the Culture Collection, University of Goteborg (CCUG) from Y.-W. Tang and was designated CCUG 56015. The strain was shown to have properties consistent with the original description of A. ursingii, except for the inability to utilize citrate (4). It is of note, however, that there are other citrate-negative A. ursingii strains in our current collection, which indicates the intraspecies variation of this property. Importantly, the hemolytic activity of AK001 reported by Kilic et al. (2) could not be confirmed, although sheep blood agar plates obtained from different suppliers were used at different cultivation temperatures and with several control hemolytic Acinetobacter strains.
In the study by Kilic et al. (2), the 16S rRNA gene sequences of AK001 were 99.5% similar to that of the A. ursingii type strain, which is the value that can be found in strains within one species (4). To obtain more comprehensive information on the rpoB-based position of AK001 within the genus Acinetobacter, we determined the rpoB sequence of AK001 and compared it to that of eight A. ursingii strains and all hitherto-described species. The similarity values between AK001 (GenBank no. {"type":"entrez-nucleotide","attrs":{"text":"EU742582","term_id":"190688713","term_text":"EU742582"}}EU742582) and the A. ursingii strains ({"type":"entrez-nucleotide","attrs":{"text":"EU477105","term_id":"526150767","term_text":"EU477105"}}EU477105 and {"type":"entrez-nucleotide-range","attrs":{"text":"EU742575 to EU742581","start_term":"EU742575","end_term":"EU742581","start_term_id":"190688699","end_term_id":"190688711"}}EU742575 to EU742581) ranged from 96.9% to 98.0%, which partially overlaps with the range of intraspecies values found for Acinetobacter beijerinckii (3). In contrast, the rpoB similarity values between AK001 and the other species were between 78.8% and 84.5%.
The proposal of “A. septicus” was based mainly on the results of DNA-DNA reassociation. However, DNA-DNA reassociation values close to the proposed value of 70% have to be treated with caution as incongruent results for the same strains can be obtained in different laboratories and/or by different methods, which may have taxonomic implications (5). Concerning also other limitations of DNA-DNA reassociation (1), no conclusions based on these intermediate values should be made without further evidence of genetic isolation between strains. To address this problem, the relatedness of AK001 to hitherto-described species was investigated by AFLP (Keygene NV) analysis, a whole-genome fingerprinting method. Previous studies using this method have shown that strains of the same (genomic) species cluster together at a ≥50% similarity level (4). AFLP analysis showed that AK001 and 25 reference A. ursingii strains were linked in a distinct cluster at a 59% similarity level, while AK001 was separated from other species at 34% similarity, which strongly supports the assignment of AK001 to A. ursingii.
Based on the above arguments, strain AK2001 is considered as belonging to A. ursingii and not representing a novel species.
