Prevalence of plasmid-mediated quinolone resistance genes among ciprofloxacin-nonsusceptible Escherichia coli and Klebsiella pneumoniae isolated from blood cultures in Korea.

OBJECTIVES: To analyze the prevalence of plasmid-mediated quinolone resistance (PMQR) determinants in ciprofloxacin-nonsusceptible Escherichia coli and Klebsiella pneumoniae isolated from patients at a tertiary care hospital in Korea.
METHODS: A total of 102 nonduplicate isolates of ciprofloxacin-intermediate or ciprofloxacin-resistant E coli (n=80) and K pneumoniae (n=22) from blood cultures were obtained. The qnr (qnrA, qnrB, qnrS), aac(6')-Ib-cr, qepA and oqxAB genes were detected using polymerase chain reaction (PCR) and confirmed using direct sequencing. To determine whether the PMQR-positive plasmid was horizontally transferable, conjugation experiments were performed.
RESULTS: Of the 102 isolates, 81 (79.4%) had one or more PMQR genes; these consisted of 59 (73.8%) E coli and 22 (100%) K pneumoniae isolates. The qnr genes were present in 15 isolates (14.7%): qnrB4 was detected in 10.8% and qnrS1 was detected in 3.9%. The aac(6')-Ib-cr, qepA and oqxAB genes were detected in 77.5%, 3.9% and 10.8%, respectively. In conjugation experiments, PMQR genes were successfully transferred from seven (8.6%) isolates. The range of minimum inhibitory concentrations of ciprofloxacin for these seven transconjugants increased to 0.5 mg/L to 1 mg/L, which was 16- to 33-fold that of the recipient E coli J53 bacteria.
CONCLUSIONS: PMQR genes were highly prevalent among ciprofloxacin-nonsusceptible E coli and K pneumoniae from blood cultures in the authors' hospital. Therefore, it is necessary to monitor for the spread of PMQR genes of clinical isolates and to ensure careful antibiotic use in a hospital setting.

The quinolone class of antibiotics was introduced into clinical use in the 1960s (1) and has since been important for the treatment of bacterial infections. In the late 1980s, more systemically active drugs (eg, fluoroquinolone) became clinically available (2). Over the decades since the introduction of fluoroquinolones, resistance to these agents in Enterobacteriaceae has become common and widespread.

The main mechanisms of quinolone resistance arise from chromosomal mutations in genes encoding DNA gyrase and topoisomerase IV (3). Upregulation of efflux pumps and/or decreased expression of outer membrane porins are also classically described mechanisms resulting from chromosomal mutations (4,5). Recently, however, plasmid-mediated quinolone resistance (PMQR) genes have been detected in Enterobacteriaceae (6). Since the first PMQR determinant, termed Qnr (now known as QnrA1), was reported in a Klebsiella pneumoniae isolate in 1998 (6), two mechanisms of PMQR have been reported including the quinolone modification with a piperazinyl substituent by the acetyltransferase AAC(6′)-Ib-cr and active efflux by QepA and OqxAB, which are pumps related to major facilitator superfamily transporters (7–10). The PMQR genes confer low-level quinolone resistance and supplement the level of resistance caused by other resistance mechanisms.

There are very few reports investigating these four different PMQR determinants (Qnr, AAC(6′)-Ib-cr, QepA and OqxAB), especially OqxAB, from blood cultures in Korea. Therefore, in the present study, we determined the prevalence of PMQR determinants in ciprofloxacin-nonsusceptible Escherichia coli and Klebsiella pneumoniae isolated from patient blood cultures in Korea.

METHODS

Bacterial isolates

A total of 102 nonduplicate clinical isolates of ciprofloxacin-intermediate or ciprofloxacin-resistant E coli (n=80) and K pneumoniae (n=22) were obtained from blood cultures collected between January 2005 and December 2010 at the Kyung Hee Medical Center (Seoul, Republic of Korea). Bacterial identification and antimicrobial susceptibilities were determined according to routine laboratory protocols using conventional biochemical tests and the MicroScan WalkAway 96 (Dade Behring, USA), following the Clinical and Laboratory Standards Institute guidelines: ciprofloxacin susceptible, minimum inhibitory concentration (MIC) ≤1 μg/mL; intermediate, MIC 2 μg/mL; and resistant, MIC ≥4 μg/mL (11). Each isolate was obtained from an individual patient.

Polymerase chain reaction amplification and sequencing for detection of PMQR genes

Amplification of PMQR genes (qnrA, qnrB, qnrS, aac(6′)-Ib, qepA, oqxA and oqxB) was performed using primers as described previously (12–15). Plasmid DNA was extracted from each isolate using a plasmid purification kit (SolGent Co, Daejeon, Korea) according to the manufacturer’s instructions. All qnr (qnrA, qnrB and qnrS) genes were detected using multiplex polymerase chain reaction (PCR), and aac(6′)-Ib, qepA, oqxA and oqxB were detected using PCR. Positive and negative controls were included for quality control. For the qnr PCR, 2 μL plasmid DNA was added to 50 μL reaction mixture containing 5 μL PCR buffer (15 mM MgCl2) (JMR Holdings, United Kingdom), 2.5 mM dNTPs (GeneACT Inc, Japan), 20 pM/μL of each primer and 1.5 U Taq polymerase. PCR conditions using the Gene AmpPCR system 9600 (Perkin-Elmer Centus Corp, USA) were: 5 min at 95°C; 35 cycles of amplification consisting of 60 s at 95°C, 60 s at 54°C and 60 s at 72°C; and 10 min at 72°C for the final extension. For aac(6′)-Ib PCR, 1 μL plasmid DNA was added to 20 μL reaction mixture containing 2.0 μL PCR buffer, 2.5 mM dNTPs, 10 pM/μL primer and 0.4 U Taq polymerase. PCR conditions were: 12 min at 95°C; 35 cycles of amplification consisting of 45 s at 94°C, 60 s at 53°C and 60 s at 72°C; and 5 min at 72°C for the final extension. For qepA PCR, 3 μL plasmid DNA was added to 16 μL reaction mixture containing 10 pM/μL primer and 2× multiplex PCR premix (SolGent, Korea). PCR conditions were: 12 min at 95°C; 35 cycles of amplification consisting of 60 s at 96°C, 60 s at 60°C and 60 s at 72°C; and 5 min at 72°C for the final extension. For oqxA and oqxB PCRs, 3 μL plasmid DNA was added to 16 μL reaction mixture containing 10 pM/μL primer and 2× multiplex PCR premix. PCR conditions were: 12 min at 95°C; 32 cycles of amplification consisting of 45 s at 94°C, 45 s at 64°C and 60 s at 72°C; and 5 min at 72°C for the final extension. The PCR products were analyzed using electrophoresis in a 2% agarose gel containing 0.5 μg/mL ethidium bromide at 130 V for 30 min. Positive and negative controls were included for quality control. Direct sequencing of the PCR products was used to confirm qnr, aac(6′)-Ib and qepA positivity for PMQR genes. To identify aac(6′)-Ibcr, aac(6′)-Ib-positive PCR products were confirmed by direct sequencing using a 3130XL DNA genetic analyzer (Applied Biosystems, USA). Isolates positive for both oqxA and oqxB were regarded as oqxAB-positive because the OqxAB protein is encoded by oqxA and oqxB genes located within the same operon. Nucleotide sequences were analyzed using the BLAST online service provided by the National Center for Biotechnology Information website (www.ncbi.nlm.nih.gov/BLAST).

Conjugation experiments to determine PMQR transferability

To determine whether quinolone resistance was transferable from the bacterial strains with plasmids carrying PMQR determinants, conjugation experiments were performed with azide-resistant E coli J53 as the recipient. Each clinical strain was inoculated along with the recipient strain into tryptic soy broth and incubated at 37°C for 3 h. Transconjugants were selected on MacConkey agar containing sodium azide (100 mg/L) and ciprofloxacin (0.06 mg/L). To determine the presence of PMQR determinants, colonies were picked from the selection agar and analyzed by PCR.

Antimicrobial susceptibility test

MICs of various antibiotics (amikacin, gentamicin, tobramycin, nalidixic acid, ciprofloxacin, levofloxacin and olaquindox) were determined for the PMQR gene-positive donors and the recipient transconjugants using the broth microdilution method according to Clinical Laboratory Standards Institute guidelines (11) and using E coli ATCC 25922 as a control.

Statistical analysis

Statistical analysis of species-specific distributions of PMQR genes was performed using Fisher’s exact test; P<0.05 was considered to be statistically significant. MedCalc version 10.4.5 (MedCalc Software, Belgium) was used for calculations.

RESULTS

Prevalence of PMQR genes

Among the 102 total ciprofloxacin-intermediate or ciprofloxacin-resistant isolates, 81 (79.4%) were positive for at least one PMQR gene. PMQR genes were detected in 59 of 80 (73.8%) E coli and all 22 (100%) K pneumoniae isolates (Table 1).

TABLE 1

Annual distribution of plasmid-mediated quinolone resistance (PMQR) genes of Escherichia coli and Klebsiella pneumoniae isolates from 2005 to 2010

Year of isolate	PMQR-positive isolates total isolates, n/n (%)		Isolates with any PMQR genes, n (%)
	E coli	K pneumoniae
2005	1/10 (10.0)	0 (0)	1 (10.0)
2006	1/5 (20.0)	0 (0)	1 (20.0)
2007	3/7 (42.9)	0 (0)	3 (42.9)
2008	14/17 (82.4)	3/3 (100.0)	17 (85.0)
2009	19/20 (95.0)	8/8 (100.0)	27 (96.4)
2010	21/21 (100.0)	11/11 (100.0)	32 (100.0)
Total	59/80 (73.8)	22/22 (100.0)	81 (79.4)

Of the PMQR genes, qnr genes were present in 15 (14.7%) isolates. The qnrA gene was not detected in any isolate; however, qnrB was detected in 11 (50.0%) K pneumoniae isolates and qnrS was detected in two (2.5%) E coli and two (9.1%) K pneumoniae isolates. The sequences of qnrB and qnrS were identical to those of qnrB4 and qnrS1, respectively. Eighty-two of the 102 (80.4%) isolates were positive for aac(6′)-Ib, and 79 of 102 (77.5%) isolates were positive for aac(6′)-Ibcr. The aac(6′)-Ib-cr gene was detected in 59 of 80 (73.8%) E coli and 20 of 22 (90.9%) K pneumoniae isolates. The qepA gene was present in four of 102 isolates (3.9%), all of which were E coli strains. Eleven of the 102 (10.8%) isolates were positive for both oqxA and oqxB. The oqxAB gene was not found in any E coli isolate; all 11 oqxAB-positive isolates were K pneumoniae strains (Table 2).

TABLE 2

Prevalence of plasmid-mediated quinolone resistance genes in Escherichia coli and Klebsiella pneumoniae isolates

Species	Isolates, n (%)
	qnrB4	qnrS1	aac(6′)-Ib-cr	qepA	oqxAB
E coli (n=80)	0 (0)	2 (2.5)	59 (73.8)	4 (5.0)	0 (0)
K pneumoniae (n=22)	11 (50.0)	2 (9.1)	20 (90.9)	0 (0)	11 (50.0)
Total (n=102)	11 (10.8)	4 (3.9)	79 (77.5)	4 (3.9)	11 (10.8)

Among the 102 isolates, 13 (12.7%) had two PMQR genes. Two E coli isolates contained both qnrS1 and aac(6′)-Ib-cr genes, and four were positive for both aac(6′)-Ib-cr and qepA (Table 3). Of the K pneumoniae isolates, one contained both qnrS1 and aac(6′)-Ib-cr genes, three contained both qnrB4 and aac(6′)-Ib-cr genes, and two contained both qnrB4 and oqxAB genes. Seven (6.9%) isolates, all of which were K pneumoniae strains, had three PMQR genes; one of these possessed qnrS1, aac(6′)-Ib-cr and oqxAB genes, and six contained qnrB4, aac(6′)-Ib-cr and oqxAB genes (Table 4).

TABLE 3

Plasmid-mediated quinolone resistance (PMQR) genes and minimum inhibitory concentrations of antimicrobial agents for donors and their transconjugants in Escherichia coli isolates

Isolate	PMQR determinant	Minimum inhibitory concentration, mg/L
		AMK	GEN	TOB	NAL	CIP	LEX	OLQ
Ec 7	aac(6′)-Ib-cr	8	4	4	>256	64	32	32
Ec 13	aac(6′)-Ib-cr	16	2	4	>256	32	32	32
Ec 18	aac(6′)-Ib-cr	8	128	4	>256	64	32	32
Ec 19	aac(6′)-Ib-cr	16	256	32	>256	64	32	32
Ec 20	qnrS1, aac(6′)-Ib-cr	8	2	2	256	4	16	32
Tc Ec 20	qnrS1	1	0.5	0.5	8	1	1	32
Ec 23	aac(6′)-Ib-cr	8	4	4	>256	64	32	32
Ec 24	aac(6′)-Ib-cr	32	8	8	>256	64	16	32
Ec 25	aac(6′)-Ib-cr	8	2	2	>256	128	32	32
Ec 26	aac(6′)-Ib-cr	16	4	4	>256	128	64	16
Ec 30	aac(6′)-Ib-cr	8	2	2	>256	128	64	64
Ec 31	aac(6′)-Ib-cr	8	4	4	>256	128	64	32
Ec 32	aac(6′)-Ib-cr	16	2	32	>256	256	16	16
Ec 33	aac(6′)-Ib-cr	4	128	16	>256	128	32	32
Ec 34	aac(6′)-Ib-cr	16	128	61	>256	>256	32	32
Ec 35	aac(6′)-Ib-cr, qepA	16	64	32	>256	>256	64	32
Ec 36	aac(6′)-Ib-cr	8	64	16	>256	128	32	32
Ec 37	aac(6′)-Ib-cr	8	4	4	>256	>256	128	16
Ec 38	aac(6′)-Ib-cr	8	256	16	>256	128	32	32
Ec 39	aac(6′)-Ib-cr, qepA	16	8	8	>256	256	32	32
Ec 40	aac(6′)-Ib-cr	8	4	4	>256	64	16	32
Ec 41	aac(6′)-Ib-cr	2	4	4	>256	64	32	16
Ec 42	aac(6′)-Ib-cr, qepA	1	2	2	>256	16	4	32
Ec 43	aac(6′)-Ib-cr	4	2	2	>256	16	16	64
Ec 44	aac(6′)-Ib-cr	4	64	16	>256	>256	>256	32
Ec 45	aac(6′)-Ib-cr	16	2	32	>256	>256	64	32
Tc Ec 45	aac(6′)-Ib-cr	1	0.5	1	128	0.5	2	32
Ec 46	aac(6′)-Ib-cr	16	128	16	>256	>256	128	32
Tc Ec 46	aac(6′)-Ib-cr	1	0.5	1	64	1	0.5	32
Ec 47	aac(6′)-Ib-cr	8	2	4	>256	>256	64	32
Ec 48	aac(6′)-Ib-cr	16	4	4	>256	>256	64	64
Ec 49	aac(6′)-Ib-cr, qepA	8	>256	16	>256	128	32	16
Ec 50	aac(6′)-Ib-cr	16	2	4	>256	128	16	32
Ec 52	aac(6′)-Ib-cr	8	2	2	>256	128	16	32
Ec 53	aac(6′)-Ib-cr	8	128	8	>256	256	32	32
Ec 54	aac(6′)-Ib-cr	4	64	4	>256	128	32	16
Ec 55	aac(6′)-Ib-cr	8	4	4	>256	128	32	32
Ec 56	aac(6′)-Ib-cr	8	4	32	>256	>256	32	32
Ec 57	aac(6′)-Ib-cr	4	128	16	>256	32	16	32
Ec 58	aac(6′)-Ib-cr	16	4	64	>256	>256	32	16
Ec 59	aac(6′)-Ib-cr	8	2	4	>256	128	64	32
Ec 60	aac(6′)-Ib-cr	8	2	4	>256	64	32	32
Ec 61	aac(6′)-Ib-cr	>256	2	4	>256	64	16	256
Ec 62	aac(6′)-Ib-cr	8	32	4	>256	128	32	32
Ec 63	aac(6′)-Ib-cr	16	32	16	>256	128	32	16
Ec 64	aac(6′)-Ib-cr	8	4	4	>256	>256	32	16
Ec 65	aac(6′)-Ib-cr	8	2	4	>256	128	32	32
Ec 66	aac(6′)-Ib-cr	8	4	4	>256	128	32	32
Ec 67	aac(6′)-Ib-cr	32	32	64	>256	>256	32	32
Ec 68	aac(6′)-Ib-cr	16	>256	64	>256	128	32	32
Ec 69	aac(6′)-Ib-cr	8	2	4	>256	2	2	32
Ec 70	qnrS1, aac(6′)-Ib-cr	8	128	16	>256	4	4	32
Ec 71	aac(6′)-Ib-cr	16	2	4	>256	128	64	32
Ec 72	aac(6′)-Ib-cr	16	2	4	>256	64	32	32
Ec 73	aac(6′)-Ib-cr	16	4	4	>256	128	32	32
Ec 74	aac(6′)-Ib-cr	4	2	4	>256	256	128	16
Ec 75	aac(6′)-Ib-cr	16	4	4	>256	>256	64	32
Ec 76	aac(6′)-Ib-cr	32	128	16	>256	256	32	32
Ec 77	aac(6′)-Ib-cr	4	4	4	>256	128	64	64
Ec 78	aac(6′)-Ib-cr	8	2	4	>256	128	32	32
Ec 79	aac(6′)-Ib-cr	8	128	16	>256	256	32	32
Ec 80	aac(6′)-Ib-cr	32	256	64	>256	>256	64	64
Recipient
Ec J53	None	1	0.5	1	2	0.03	0.06	16

AMK Amikacin; CIP Ciprofloxacin; Ec E coli; GEN Gentamicin; LEX Levofloxacin; OLQ Olaquindox; NAL Nalidixic acid; Tc Transconjugant; TOB Tobramycin

TABLE 4

Plasmid-mediated quinolone resistance (PMQR) genes and minimum inhibitory concentrations of antimicrobial agents for donors and their transconjugants in Klebsiella pneumoniae isolates

Isolate	PMQR determinant	Minimum inhibitory concentration, mg/L
		AMK	GEN	TOB	NAL	CIP	LEX	OLQ
Kp 1	qnrB4, aac(6′)-Ib-cr	4	>256	>256	>256	16	256	>256
Tc Kp 1	qnrB4, aac(6′)-Ib-cr	2	0.5	1	4	0.5	0.5	32
Kp 2	aac(6′)-Ib-cr	4	2	2	>256	>256	16	16
Kp 3	aac(6′)-Ib-cr	1	1	2	>256	>256	128	>256
Kp 4	qnrB4, aac(6′)-Ib-cr	>256	>256	>256	>256	128	16	>256
Kp 5	qnrS1, aac(6′)-Ib-cr, oqxAB	2	64	8	>256	>256	128	>256
Kp 6	qnrB4, aac(6′)-Ib-cr, oqxAB	>256	>256	>256	>256	>256	256	>256
Kp 7	qnrS1, aac(6′)-Ib-cr	2	1	1	>256	8	16	256
Tc Kp 7	qnrS1, aac(6′)-Ib-cr	1	0.5	0.5	4	1	0.5	16
Kp 8	qnrB4, oqxAB	>256	256	32	>256	>256	256	>256
Kp 9	qnrB4, aac(6′)-Ib-cr, oqxAB	>256	1	2	>256	>256	256	>256
Kp 10	qnrB4, aac(6′)-Ib-cr, oqxAB	>256	>256	>256	>256	128	256	>256
Kp 11	qnrB4, aac(6′)-Ib-cr	>256	>256	>256	>256	8	32	256
Kp 12	aac(6′)-Ib-cr	2	1	2	>256	8	8	63
Kp 13	qnrB4, aac(6′)-Ib-cr, oqxAB	>256	>256	>256	>256	256	128	>256
Kp 14	qnrB4, oqxAB	>256	>256	>256	>256	>256	128	>256
Kp 15	aac(6′)-Ib-cr	2	0.5	2	>256	16	32	>256
Kp 16	aac(6′)-Ib-cr, oqxAB	2	16	4	>256	64	128	128
Tc Kp 16	aac(6′)-Ib-cr	1	0.5	1	4	0.5	0.5	32
Kp 17	qnrB4, aac(6′)-Ib-cr, oqxAB	>256	>256	>256	>256	256	128	>256
Kp 18	aac(6′)-Ib-cr	16	256	16	>256	64	64	256
Kp 19	aac(6′)-Ib-cr, oqxAB	>256	1	1	>256	16	128	128
Kp 20	qnrB4, aac(6′)-Ib-cr, oqxAB	>256	>256	>256	>256	256	128	>256
Tc Kp 20	aac(6′)-Ib-cr, oqxAB	16	0.5	1	>256	0.5	128	256
Kp 21	aac(6′)-Ib-cr	8	32	4	>256	128	64	>256
Kp 22	aac(6′)-Ib-cr	16	2	16	>256	32	32	256
Recipient
Ec J53	None	1	0.5	1	2	0.03	0.06	16

AMK Amikacin; CIP Ciprofloxacin; Ec Escherichia coli; GEN Gentamicin; Kp K pneumoniae ; LEX Levofloxacin; NAL Nalidixic acid; OLQ Olaquindox; Tc Transconjugant; TOB Tobramycin

Conjugation experiment

Seven transconjugants were successfully obtained from the 81 PMQR-positive isolates used as donors in conjugation experiments. The qnr gene was successfully transferred in three of the 15 qnr-positive isolates (two were qnrS1 and one was qnrB4). The aac(6′)-Ib-cr gene was transferred in six of 79 isolates and the oqxAB gene was transferred in one of 11 isolates; transconjugation produced no qepA-positive isolates.

Transconjugants were obtained from three of 59 (5.1%) PMQR-positive E coli isolates and four of 22 (18.2%) PMQR-positive K pneumoniae isolates. Of the three transconjugants with E coli donors, the transfer of aac(6′)-Ib-cr occurred in two and the transfer of qnrS1 occurred in one. Of the four transconjugants with K pneumoniae donors, transfer of the aac(6′)-Ib-cr gene occurred in one, and cotransfer of qnrB4 and aac(6′)-Ib-cr, qnrS1 and aac(6′)-Ib-cr, or aac(6′)-Ib-cr and oqxAB occurred from different donors. Transferability was highest for qnrS1 (two of four [50.0%]), followed by qnrB4 (one of 11 [9.1%]) and oqxAB (one of 11 [9.1%]), and aac(6′)-Ib-cr (six of 79 [7.6%]) (Tables 3 and 4).

Among the 81 PMQR-positive isolates, the MIC of ciprofloxacin ranged from 2 mg/L to >256 mg/L. The resistance rates of PMQR-positive isolates to nalidixic acid, levofloxacin, amikacin, gentamicin and tobramycin were 100% (81 of 81), 96.3% (78 of 81), 14.8% (12 of 81), 43.2% (35 of 81) and 40.7% (33 of 81), respectively.

The MIC of ciprofloxacin for the seven transconjugants ranged from 0.5 mg/L to 1 mg/L, or 16- to 33-fold higher than that for the E coli J53 recipient bacteria (MIC 0.03 mg/L). All three qnr-containing transconjugants conferred decreased susceptibility to ciprofloxacin (MIC range 0.5 mg/L to 1 mg/L), nalidixic acid (MIC range 4 mg/L to 8 mg/L) and levofloxacin (MIC range 0.5 mg/L to 1 mg/L); these MICs are 16- to 33-fold, two- to fourfold and eight- to 16-fold the MICs for the preconjugated recipient E coli J53 bacteria (0.03 mg/L, 2 mg/L and 0.0625 mg/L, respectively). The MIC of ciprofloxacin for six aac(6′)-Ibcr-containing transconjugants ranged from 0.5 mg/L to 1 mg/L, or 16- to 33-fold the MIC for the preconjugated recipient. All aac(6′)-Ib-cr-containing transconjugants exhibited decreased susceptibility to nalidixic acid and levofloxacin. The two transconjugants with qnr and aac(6′)-Ibcr exhibited increased MICs for ciprofloxacin (range 0.5 mg/L to 1 mg/L), which were 16- to 33-fold higher than the MIC for the preconjugated recipient. For one transconjugant with both aac(6′)-Ib-cr and oqxAB, the MIC to ciprofloxacin was 0.5 mg/L, or 16-fold the MIC of the preconjugated recipient (Tables 3 and 4).

DISCUSSION

We evaluated the incidence of qnr, aac(6′)-Ib-cr, qepA and oqxAB genes in ciprofloxacin-nonsusceptible E coli and K pneumoniae strains isolated from patient blood cultures in Korea.

The qnr genes encode proteins that protect DNA gyrase and topoisomerase IV from inhibition by quinolones (16,17), and have recently been identified worldwide. The prevalence of the qnr genes in bacterial isolates may range from <1% to >50% (18–21), depending on the selection criteria and study period for bacterial isolates. Among ciprofloxacin-resistant E coli and K pneumoniae isolates, the incidences of qnr in China are 7.5% and 11.9%, respectively. qnrA, qnrB and qnrS were detected either alone or in combination in 3.8%, 4.7% and 3.8% of these isolates, respectively (18). In Korea, Shin et al (20) reported that 5.6% of E coli and 55.9% of K pneumoniae ciprofloxacin-resistant isolates contained only qnrB (qnrB2, qnrB4 and/or qnrB6). Jeong et al (19) reported that the prevalence of qnrA in Korea was 0.8% in E coli isolates (ciprofloxacin susceptible and resistant) between 2001 and 2003. Kim et al (21) determined that 0.5% of E coli and 5.9% of K pneumoniae (ciprofloxacin susceptible and resistant) isolates in Korea contained qnr (qnrB or qnrS). Of the qnr variants, we did not detect qnrA; qnrB4 was the most common, followed by qnrS1. Epidemiological investigations, including the present study, have shown that qnrB (especially qnrB4) (22) is common, while qnrA and qnrS are present in Korea at relatively low prevalences (19–21). In our study, the prevalence of qnrB in K pneumoniae (50%) was significantly higher than that in E coli (0%) (Fisher’s exact test, P<0.001), as noted previously (18,20).

The aac(6′)-Ib-cr gene, a variant of the gene encoding AAC(6′)-Ib, was first described in 2006 (7). The AAC(6′)-Ib-cr enzyme reduces only ciprofloxacin and norfloxacin activity by acetylation (7). Quinolones without piperazinyl nitrogen were not affected by aac(6′)-Ib-cr (23). However, transconjugants containing only aac(6′)-Ib-cr also exhibited reduced susceptibilities to levofloxacin in the present study, suggesting it contributes to antimicrobial resistance through additional mechanisms. The prevalence of aac(6′)-Ib-cr was higher in our study (77.9%) than in previous studies (7,15,24–26). Among clinical E coli isolates collected in China, 51% had aac(6′)-Ib-cr (7). In the United States, aac(6′)-Ib-cr was detected in 32% of E coli and 16% of K pneumoniae isolates (15). In Korea, aac(6′)-Ib-cr was detected in 3.4% of Enterobacteriaceae (24) and in 34.1% of extended-spectrum β-lactamase (ESBL)-producing E coli and K pneumoniae (26). In some reports, the presence of aac(6′)-Ib-cr was prevalent among qnr-positive isolates compared with qnr-negative isolates, suggesting a genetic assocication of quinolone resistance with aminoglycoside resistance (25,26). We also found that the prevalence of aac(6′)-Ib-cr in qnr-positive isolates (13 of 15 [86.7%]) was slightly higher than in qnr-negative isolates (66 of 87 [75.9%]).

The qepA gene encodes a novel efflux pump that resembles a 14-transmembrane-segment putative efflux pump belonging to the major facilitator superfamily (8). In 2007, qepA was first reported in clinical E coli isolates from Japan (8) and Belgium (27). According to recent studies, qepA has a low prevalence (<1% in Korea [24,28]). In the present study, the prevalence of qepA among the 80 ciprofloxacinnonsusceptible E coli isolates (5%) was higher than that in previous studies (24,28). Another plasmid-mediated efflux pump gene belonging to the resistance-nodulation-cell division family, oqxAB, confers reduced susceptibility to multiple agents including olaquindox (a growth promoter in pigs), quinolones and fluoroquinolones (29,30). OqxAB is encoded by the oqxA and oqxB genes, which are located in the same operon. The oqxAB genes are chromosomally located in K pneumoniae. Thus, the plasmid containing oqxAB appears to be the result of the capture of a chromosomal cassette from Klebsiella species (30). Also, Rodriguez-Martinez et al (31) found simultaneous oqxA and oqxB signals in both chromosomal and large plasmid locations. The prevalence of the oqxAB gene was 74% to 100% in other studies; thus, the detected prevalence of 50% among K pneumoniae isolates in the present study was a relatively low value (12,32). However, we obtained only plasmid DNA using a plasmid purification kit; other studies obtained the chromosomal and/or plasmid DNA for detection of oqxAB gene. Plasmid-mediated OqxAB was first detected in a human clinical isolate of E coli from Korea (12). However, none of the E coli isolates in the present study possessed oqxAB. In previous studies, oqxAB-positive K pneumoniae isolates yielded no transconjugants. However, one transconjugant with a K pneumoniae donor obtained the oqxAB gene, which conferred decreased susceptibility to ciprofloxacin and olaquindox. There is still a lack of epidemiological information about oqxAB gene in humans, and this requires further study.

Park et al (33) reported that the prevalence of qnr determinants or aac(6′)-Ib-cr was 97.4% in isolates with ciprofloxacin MICs of 1 mg/L, but 6.7% in isolates with ciprofloxacin MICs of 0.25 mg/L among ciprofloxacin-susceptible isolates of K pneumoniae in Korea. In this study, the prevalence of qnr determinants or aac(6′)-Ib-cr was 100% in ciprofloxacin-nonsusceptible isolates of K pneumoniae; PMQR genes were remarkably high in isolates with ciproxacin MICs >1 mg/L (33).

Nam et al (34) studied mutations in the DNA gyrase and topoisomerase IV gene in the same isolates as included in the present study, and the mutation of the gyrA and parC genes were 98.0% and 91.1%, respectively, in these ciprofloxacin-nonsusceptible E coli and K pneumoniae. Of these, two K pneumoniae exhibited no mutations in the DNA gyrase and topoisomerase IV genes, but both had PMQR genes.

Conjugation experiments demonstrated that PMQR was transferable. The MICs of ciprofloxacin for seven transconjugants were 16- to 33-fold higher than the MIC for the unconjugated recipient E coli J53 strain, and the MICs of ciprofloxacin for three transconjugants carrying multiple PMQR genes (qnr and aac[6′]-Ib-cr, or aac[6′]-Ib-cr and oqxAB) were 16-to 33-fold higher than the MIC for the unconjugated recipient. The MICs of ciprofloxacin for transconjugants carrying aac(6′)-Ib-cr in combination with qnr or oqxAB were not significantly higher than those for transconjugants carrying aac(6′)-Ib-cr only, suggesting the presence of additional mechanisms contributing to fluoroquinolone resistance. These PMQR determinants confer low-level fluoroquinolone resistance and may facilitate higher-level resistance under selective pressure from antimicrobial agents at therapeutic levels (35,36). PMQR has been closely associated with ESBL, AmpC-type β-lactamase and aminoglycoside resistance mechanisms (5). In our study, the prevalence of ESBL-producing isolates in PMQR-positive isolates (28 of 81 [34.6%]) was higher than in PMQR-negative isolates (three of 21 [14.3%]), but the difference was not statistically significant (Fisher’s exact test, P=0.109). Cotransfer of PMQR genes may contribute to the spread of multidrug resistance. Clinicians should be careful in prescribing quinolone and fluoroquinolone to prevent the spread of multidrug resistance.

In the present study, we investigated a variety of PMQR genes in E coli and K pneumoniae and provided additional information about the actively investigated qepA and oqxAB genes. Analysis of the genes over several years made it possible to predict the presence of PMQR genes, and offers important information for antimicrobial selection and infection control.

It is important to note that the present study had several limitations. It was conducted at a single hospital and did not analyze the clonal relationships among PMQR-positive isolates. Also, it is necessary to confirm the colocalization of the qnr gene and other PMQR genes by PCR or Southern blot hybridization with both DNA probes of a single plasmid. Further nationwide epidemiological surveys and additional molecular studies for the possibility of horizontal transmission are required to support our results.

CONCLUSION

We identified PMQR genes in 79.4% (81 of 102) of ciprofloxacinnonsusceptible E coli and K pneumoniae isolated from a tertiary-care hospital in Korea. The prevalent PMQR gene was aac(6′)-Ib-cr, followed by qnrB4 and oqxAB, and qnrS1 and qepA. PMQR genes were highly prevalent among ciprofloxacin-nonsusceptible E coli and K pneumoniae isolated from blood cultures in our hospital. Therefore, it is necessary to monitor for spread of PMQR genes of clinical isolates and to ensure careful antibiotic use in a hospital setting.