Abstract
Two resistancenodulationcell division (RND) type active efflux systems, AdeABC and AdeDE, have been described in Acinetobacter (Chau et al., 2004; Magnet et al., 2001). Both systems confer upon the host a reduced susceptibility to aminoglycosides, fluoroquinolones, erythromycin, tetracycline and chloramphenicol. AdeABC differs from AdeDE by protecting the host from cefotaxime, whilst AdeDE increases the host resistance to ceftazidime and rifampicin (Chau et al., 2004; Magnet et al., 2001). The three genes encoding the tripartite AdeABC were found as an array but AdeDE has only the membrane fusion protein (MFP) gene (adeD) and the RND transporter gene (adeE) clustered together. The outer-membrane protein (OMP) for AdeDE has not been identified.
The clinical significance of differences in antimicrobial-susceptibility pattern among different Acinetobacter genomic DNA groups (GDGs) has been demonstrated (Houang et al., 2003), and the importance of efflux systems in multidrug resistance in Acinetobacter is increasingly appreciated. Therefore, we applied specific PCR to detect the transporter genes adeB and adeE in our collection of Acinetobacter strains in an attempt to study the distribution of these two efflux systems among different GDGs.
Unexpectedly, a PCR designed to screen for adeE detected a variant in an Acinetobacter isolate belonging to GDG 3. Genome walking utilizing the commercial kit GenomeWalker (BD Biosciences Clontech) was employed to clone the flanking regions of the amplicon as described previously (Chau et al., 2004). Alignment of 15 contigs generated after 12 rounds of walking yielded 3 consecutive ORFs of 1248, 3177 and 1452 bp (GenBank accession no. DQ223769). These three ORFs, theoretically encoding three peptides resembling an MFP (415 residues), an RND transporter (1058 residues) and an OMP (483 residues), were putatively designated adeX, adeY and adeZ, respectively. When they were compared with other known RND efflux systems, AdeX had highest identity (46 %) to MexA in Pseudomonas aeruginosa (GenBank accession no. L11616), AdeY (60 %) to SmeE in Stenotrophomonas maltophilia (GenBank accession no. AJ746241) and AdeZ (41 %) to OprM in P. aeruginosa (GenBank accession no. AB011381), as determined by the software ClustalW (EBI). Conversely, AdeX had only 35 and 38 % peptide sequence identity to AdeA and AdeD, respectively; whereas AdeY had 45 and 51 % identity to AdeB and AdeE, respectively, and AdeZ had a relatively high identity, 67 %, to AdeC. Moreover, when the DNA sequence of the entire adeXYZ gene cluster was used to search against the genome of the strain ADP1, formally identified as Acinetobacter calcoaceticus GDG 1 (Barke et al., 2004), a homologous gene cluster (2 872 149 to 2 878 057 bp) was found with an identity of 77 %; the identity of the theoretical peptides to AdeX, AdeY and AdeZ were 80, 89 and 87, respectively.
Twelve attempts were made to disrupt adeY by targeted integration, as previously described (Chau et al., 2004), in order to assess the involvement of adeY in antimicrobial resistance. However, no stable transformant was obtained and the reason behind the repeated failures was not known. Although the role of AdeXYZ in antimicrobial resistance has not been proven, given its increased similarity in loci organization and peptide sequences to known RND drug efflux systems, the possibility of its involvement cannot be ruled out. Furthermore, hydrophobicity analysis of the theoretical peptide sequence using the software TopPred (Pasteur Institut) demonstrated that AdeY has a topology highly compatible with it being an RND transporter (data not shown) (Tseng et al., 1999).
Based on the new sequence information, new primer pairs AdeE-245A (5'-GTA GTA GTT CGG CAG GAC AA-3') and AdeE-620B (5'-GCG GTT CTA ACA TCT GAT GG-3'), and AdeY-343A (5'-CAA TCT GCA ACT GCG CTT-3') and AdeY-920B (5'-TCA ACA GCT TCT GCG GTA-3'), were designed to detect the presence of adeE (expected amplicon size 376 bp) and adeY (expected amplicon size 587 bp), respectively. Primers O3 and O4, described by Magnet et al. (2001), were used to screen for adeB. A total of 207 Acinetobacter blood culture isolates obtained between 1997 and 2000 in our hospital were screened for the presence of adeB, adeE and adeY. The GDGs of all the isolates were determined by amplified rDNA restriction analysis as described elsewhere (Vaneechoutte et al., 1995).
Of the 59 GDG 2 (Acinetobacter baumannii) isolates tested, 39 (70 %) were positive with adeB PCR. The adeB sequence was not detected in any of the other GDGs, suggesting that adeB is an active efflux system specific to GDG 2 (Table 1). For the adeE and adeY PCR, 70 (70 %) and 75 (90 %) out of 83 GDG 3 isolates were positive, respectively, whilst all of the GDG 2 isolates tested were negative. Furthermore, adeE and adeY were found not only closely associated with GDG 3, but also coexisting in many isolates, including one GDG 13TU isolate and one GDG 17 isolate (Table 1). Of the 83 GDG 3 isolates tested, 68 (82 %) possessed both adeE and adeY sequences, whilst just 2 (2 %) had adeE only and 7 (8 %) had adeY only. The identity of all the amplicons was confirmed by RFLP analysis with restriction enzymes, HhaI, HinfI, HincII and MseI for adeX, and HhaI, HinfI, HincII, ApaLI and RsaI for adeE (data not shown). The identity of 10 % of all the amplicons was reconfirmed by DNA sequencing.