Bacterial isolates and species identification
For this study, 48 clinical isolates of the Acinetobacter species that harboured the MBL gene were identified in the Kyoto-Shiga region. Twenty-five isolates were identified as A. pittii, nine isolates were identified as A. bereziniae, five isolates were identified as A. nosocomialis, and three isolates were unclassified Acinetobacter species. Single isolates of A. baumannii, Acinetobacter guillouiae (formerly Acinetobacter genomic species 11), Acinetobacter genomic species 15TU, Acinetobacter genomic species 16, A. junii and Acinetobacter ursingii were also identified. Fig. 2 illustrates the phylogenetic relationships between non-ACB complex isolates and Acinetobacter reference strains. A total of three Acinetobacter species isolates exhibited c. 96% identity to the rpoB gene in Acinetobacter gyllenbergii and belonged to the same species with high bootstrap values in the phylogenetic tree. However, these three isolates had a low bootstrap value (49.2%) for the corresponding branch of A. gyllenbergii. Therefore, we defined these isolates as ‘unclassified’Acinetobacter species. The presence of unclassified species has been previously reported for the identification of Acinetobacter species using the partial rpoB gene sequence . A. pittii isolates were identified from four hospitals (A, B, D and E), A. bereziniae isolates were identified from three hospitals (A, B and C), A. nosocomialis isolates were identified from three hospitals (A, B and D) and unclassified Acinetobacter species were identified from two hospitals (A and B) (Table 1). The origins of A. pittii, A. bereziniae, A. nosocomialis and unclassified Acinetobacter species isolates are shown in Fig. 3. The origins of other isolates were obtained as follows: one A. baumannii (strain A12) was recovered from a nasopharyngeal swab; one A. guillouiae isolate (A13) was recovered from bile; single isolates of A. genomic species 16 (A5) and A. genomic species 15TU (A9) were recovered from ascites; a single A. junii isolate (E7) was recovered from sputum; and a single A. ursingii isolate (A7) was recovered from blood.
Figure 2. Neighbour-joining tree based on the partial rpoB gene (zone 1) sequence from isolates containing the non-ACB complex and reference strains of the Acinetobacter species. The cluster analysis was conducted using the CLUSTAL X program. The GenBank accession numbers are shown in parentheses. The numbers at the branch points indicate the bootstrap values (%, based on 1000 resamplings). Bar, 1% sequence divergence.
Download figure to PowerPoint
Table 1. Numbers of isolated Acinetobacter strains
| A. pittii (ST119)|| ||1||5||7||3||2||18||A, B, D, E|
| A. pittii (non-ST119a)||4|| || ||2||1|| ||7||A, B|
| A. bereziniae ||1||3||4||1|| || ||9||A, B, C|
| A. nosocomialis ||3||1||1|| || || ||5||A, B, D|
|Unclassified Acinetobacter species|| ||1|| || ||2|| ||3||A, B|
|Othersb|| ||2||1||2|| ||1||6||A, E|
Figure 3. The results of the PFGE analysis with ApaI, MLST, and the profiles of carbapenemase-encoding genes. The results shown in panel (a) were obtained from 25 A. pittii isolates, the results in panel (b) were obtained from nine A. bereziniae isolates, the results in panel (c) were obtained from five A. nosocomialis isolates, and the results shown in panel (d) were obtained from three unclassified Acinetobacter species isolates. The STs were based on a Pasteur Institute scheme. The MLST was performed using only the A. calcoaceticus-A. baumannii complex. The MBL genes and CHDL gene columns indicate the types of carbapenemase-encoding genes harboured. *Swab: nasopharyngeal swab. ST, sequence type; MBL, metallo-β-lactamase; CHDL, carbapenem-hydrolysing class D β-lactamase.
Download figure to PowerPoint
The genus Acinetobacter currently comprises 27 validly named species and several other genomic species with provisional designations [21–23]. Using PCR-based species-level identification [12,14], we identified Acinetobacter species harbouring MBL genes that had not previously been reported (e.g. A. guillouiae, genomic species 15TU and A. ursingii). These identification and molecular typing methods are necessary for developing a detailed epidemiology of MBL genes among non-A. baumannii Acinetobacter species.
Fig. 3 shows the results of molecular typing using PFGE and MLST. The 25 A. pittii isolates were divided into eight PFGE types and five sequence types (STs) using MLST (ST63, ST64, ST93, ST119 and ST121) (Fig. 3a). ST119 was identified in closely related A. pittii isolates from four hospitals (A, B, D and E) (Table 1). The A. pittii non-ST119 isolates were identified from two hospitals (A and B), but only one clonal PFGE type was identified in the ST63 isolates. The other isolates were sporadically identified (Fig. 3a). The nine A. bereziniae isolates were primarily identified from two hospitals (A and B) and could be divided into five PFGE types; the PFGE analysis revealed that five of the nine isolates were closely related (Fig. 3b). The five A. nosocomialis isolates were divided into two PFGE types and two STs (ST68 and ST71) (Fig. 3c). The three unclassified Acinetobacter species isolates were divided into two PFGE types (Fig. 3d). PFGE was not performed on the other Acinetobacter species because only one isolate of each species was positive for the MBL gene. An A. baumannii isolate belonging to ST120 was not related to European clones I, II or III.
The molecular epidemiologies of the Acinetobacter isolates harbouring MBL genes are lacking. Although the molecular epidemiologies of A. pittii are currently insufficient for comparison with A. baumannii, the Pasteur Institute MLST scheme might be more applicable than the Bartual scheme for the number of sequence types submitted (http://pubmlst.org/abaumannii). Thus, we conducted MLST using the Pasteur Institute scheme.
Carbapenemase-encoding genes and the genetic environment
All of the isolates harboured the class 1 integrase gene (intI1). Table 2 outlines the distribution of MBL genes among the Acinetobacter species. The blaIMP-19 gene was predominantly detected in Acinetobacter isolates, including 23 isolates of A. pittii (all 18 isolates of ST119) and all isolates of A. bereziniae, A. nosocomialis, A. baumannii, A. guillouiae, genomic species 16 and A. junii. The blaIMP-1 gene was detected in four isolates (A. pittii, unclassified Acinetobacter species, genomic species 15TU and A. ursingii) and blaIMP-11 was detected in three isolates (A. pittii and two unclassified Acinetobacter species isolates).
CHDL genes primarily contribute to carbapenem non-susceptibility among Acinetobacter species [2,24]. Even in Japan, a recent study showed that 94.5% of the carbapenem-resistant Acinetobacter isolates harboured CHDL genes, and only one A. pittii isolate harboured an IMP-type MBL gene . However, in another study conducted in Asia, A. pittii was reported to harbour MBL genes, including blaIMP-1, -4, and -8 and blaVIM-2 [6,7,9].
All of the MBL genes detected in this study were embedded in a class 1 integron. The blaIMP-19 gene was identified in a gene cassette array of blaIMP-19-aac(6′)-31-blaOXA-21-aadA1, the blaIMP-1 gene was in an array of catB8-like/aacA4-blaIMP-1, and blaIMP-11 was the only component of another gene cassette array (Fig. 1). The same class 1 integron carrying blaIMP-19 was detected in various Acinetobacter species. While this finding was reported in our previous study, the identification of this integron in A. nosocomialis, A. bereziniae and A. guillouiae was a recent discovery. These findings provide strong evidence for the inter- and intraspecies dissemination of this class 1 integron among Acinetobacter species. In contrast, no evidence of the intraspecies dissemination of a class 1 integron carrying blaIMP-1 and blaIMP-11 was demonstrated, although the regional spread and interspecies dissemination of these two MBL genes was detected.
One isolate of A. baumannii ST120 harboured blaOXA-51-like and ISAba2 in addition to blaIMP-19, but ISAba2 was not identified up- or downstream of blaOXA-51-like. ISAba3 flanked both sides of the blaOXA-58 gene, which was detected in the single isolate of A. pittii ST93 harbouring blaIMP-1 and three isolates (one isolate of A. pittii ST63 and two isolates of unclassified Acinetobacter species) harbouring blaIMP-11 (Fig. 3). Other CHDL genes were not detected in the isolates included in this study. This finding is consistent with previous reports [7,26]. Among the non-A. baumannii Acinetobacter species, two ISAba3 genes flanked the blaOXA-58-like gene, which is widely disseminated in Eastern Asian countries.
Notably, there was one major limitation of this study. We screened the Acinetobacter isolates harbouring MBL genes collected between 2001 and 2006 and could not directly determine whether the MBL genes described in this study were still disseminated among Acinetobacter species. From 2007 to 2010, 10 Acinetobacter isolates harboured exactly the same class 1 integron containing blaIMP-19 as identified in our institute . These data suggest the possibility of continued integron-borne MBL gene dissemination among the species in this region.
In conclusion, this is the first report that demonstrates the regional and molecular epidemiology of Acinetobacter species harbouring integron-borne MBL genes and provides details of their species-level identification. The IMP-type MBL genes were detected in various Acinetobacter species. In particular, the blaIMP-19 gene was predominantly identified among many Acinetobacter species, particularly A. pittii ST119. The CHDL genes were indeed major contributors to carbapenem-non-susceptibility worldwide; however, the continued investigation of MBL genes in Acinetobacter species is required.