Vibrio parahaemolyticus Chromosomal qnr Homologue VPA0095: Demonstration by Transformation with a Mutated Gene of Its Potential To Reduce Quinolone Susceptibility in Escherichia coli

The gene qnr (qnrA [G. A. Jacoby, K. Walsh, D. Mills, V. Walker, A. Robicsek, H. Oh, and D. C. Hooper, Abstr. 44th Intersci. Conf. Antimicrob. Agents Chemother., abstr. C2-1898a, 2004]; encoding QnrA) was first identified as a transferable quinolone resistance determinant on a multidrug-resistant plasmid in Klebsiella pneumoniae (6, 9) and has been detected in K. pneumoniae, Escherichia coli, Enterobacter spp., and Citrobacter freundii isolates in various countries (3, 4, 8, 10, 11). Furthermore, the qnrA homologues qnrB (encoding QnrB, 40% identity with QnrA [Jacoby et al., 44th ICAAC]) and qnrS (encoding QnrS, 59% identity with QnrA [2]) were discovered on plasmids in K. pneumoniae and Enterobacter spp. and in Shigella flexneri, respectively. Thus, several qnr genes have been found on enterobacterial plasmids, but their origin or ancestry is unknown.

Hata et al. recently reported that QnrS exhibited significant identity with a Photobacterium profundum protein (CAG22829; 64%) and a Vibrio vulnificus protein (AAO07889; 53%) (2), which also showed 66 and 60% identity with QnrA, respectively. In addition, by a homology search using the DNA Data Bank of Japan (http://www.ddbj.nig.ac.jp/search/blast-j.html ), we had noticed that another P. profundum protein (CAG21998) and a Vibrio parahaemolyticus protein (BAC61438) both showed 58% identity with QnrA (52 and 56% with QnrS, respectively). Taken together, three species of the Vibrionaceae family were revealed to chromosomally possess putative qnr homologues, raising the hypothesis that qnr genes on enterobacterial plasmids had derived from one of them or their relatives. These chromosomal homologues are not associated with any integron-like structure. Meanwhile, qnrA is present in an integron (9, 11), but qnrS is not (2).

To gain a functional insight, we examined the V. parahaemolyticus qnr homologue, VPA0095, for qnr-like potential. A DNA fragment corresponding to the gene (part of the sequence, BA000032) was amplified from V. parahaemolyticus strain 8611 (isolated at Tohoku University Hospital, Sendai, Japan) by PCR with the primers 5′-TTAAAAACCGATCTCATTTTTGAACGAG and 5′-ACTTCCTCGTCGAcGTTATTCGGTAAGTC (the small character indicates an introduced artificial base exchange; the SalI site was newly established as underlined). The fragment, 701 base pairs from the second codon to 53 bases downstream of the stop codon, was digested with SalI after blunting and kination using a TaKaRa BKL kit (Takara Bio Inc., Otsu, Japan). It was next ligated into a plasmid vector, pTV118N (pUC118 derivative; Takara Bio Inc.), which was previously treated with NcoI, KOD DNA polymerase (TOYOBO, Osaka, Japan) for blunting, and then SalI, followed by introduction into E. coli strain MC1061 (5). Plasmids were extracted from some transformants of the strain, and their partial sequences encompassing the ligated fragment were determined. We consequently obtained a plasmid, pVPQNR8, carrying VPA0095 connected to a start codon derived from lacZ following the lac promoter and Shine-Dalgarno sequence, and a plasmid, pVPQNR2, carrying a gene with a single mutation (TGT to TAT) at the 115th codon of VPA0095, which probably occurred as a PCR error and was accompanied by an amino acid change from cysteine to tyrosine (C115Y). E. coli strain KL16 (1) was transformed with each plasmid and, as a control, with pTV118N and its derivative carrying qnrA, pKPQNR (constructed with qnrA from a K. pneumoniae isolate in Japan: unpublished). The transformants were then subjected to drug susceptibility tests, in which MICs of ciprofloxacin, levofloxacin, nalidixic acid (all synthesized at Daiichi Pharmaceutical Co., Ltd., Tokyo, Japan), and minocycline (Wyeth Japan, Tokyo, Japan) were determined by the agar dilution method (7). The results are shown in Table 1. pVPQNR8, compared to the negative control, pTV118N, did not result in significant changes of the MICs. Interestingly, on the other hand, 8- to 16-fold increases in quinolone MICs were provided by pVPQNR2 carrying a mutated gene, comparable to pKPQNR. The MIC of minocycline for strain KL16 was not changed by the plasmids, indicating the specific effect of introduced genes on quinolone susceptibility.

To confirm the expression of qnr-like potential by a single mutation, we next performed site-directed mutagenesis experiments. Single base exchanges at the 115th codon of the gene on plasmids were induced using a QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, Calif.) and primers 5′-CACCATGTTTTTCTaTTCTGCTTTTATTACTGGC and5′-GCCAGTAATAAAAGCAGAAtAGAAAAACATGGTGfor the exchange in pVPQNR8 (TGT to TAT [resulting in pVPQNR8C115Y, the mutant type]) or 5′-CACCATGTTTTTCTgTTCTGCTTTTATTACTGGC and 5′-GCCAGTAATAAAAGCAGAAcAGAAAAACATGGTG for the exchange in pVPQNR2 (TAT to TGT [resulting in pVPQNR2Y115C, the native type]). Lowercase letters in the primer sequences indicate an introduced artificial base exchange. Acquired plasmids were introduced into E. coli strain KL16. As shown in Table 1, pVPQNR8C115Y clearly increased the quinolone MICs for the strain to the same level as pVPQNR2, while pVPQNR2Y115C was basically identical to pVPQNR8, indicating the verified effect of the mutation conferring C115Y in VPA0095.

Thus, one of the Vibrionaceae chromosomal qnr homologues was experimentally confirmed to possess the potential to reduce quinolone susceptibility in E. coli, suggesting that the homologues should relate closely to the origin or ancestry of qnr genes carried on enterobacterial plasmids. In addition, the potential enhancement of VPA0095 by a single mutation suggests, although its mechanism remains to be elucidated, that qnr genes and/or its homologues could be comparatively easily converted to or selected as a higher quinolone resistance determinant.