Plasmid-Mediated qnrB2 and Carbapenemase Gene bla_KPC-2 Carried on the Same Plasmid in Carbapenem-Resistant Ciprofloxacin-Susceptible Enterobacter cloacae Isolates

Abstract

Fourteen out of 16 carbapenem-resistant quinolone-susceptible Enterobacter cloacae isolates were found to carry qnrB2 and bla_KPC-2 genes encoded on the same plasmid. One isolate also carried the aac(6′)-Ib-cr gene. Coexistence of quinolone resistance determinants and bla_KPC-2 on the same plasmid in quinolone-susceptible E. cloacae isolates may have important clinical implications.

Quinolone resistance in gram-negative pathogens is commonly mediated by chromosomal mutations in DNA gyrase and topoisomerase IV by decreasing intracellular drug accumulation either due to decreased membrane permeability and/or due to chromosomally located efflux pumps (2, 21). In addition, quinolone resistance may be conferred by plasmid-mediated determinants, including the qnr genes qnrA, qnrB, and qnrS (4, 5, 7, 10), aac(6′)-Ib-cr (18), and the recently identified transporter qepA (15, 26). The qnr gene products lead to increased MICs of quinolones, which in many cases are still below the CLSI breakpoints, and require the coexistence of an additional mechanism to lead to frank resistance. qnr genes may, however, facilitate selection of resistance-conferring chromosomal mutations (6, 10, 16, 20), and the penetration of qnr into the population of Klebsiella pneumoniae and Enterobacter spp. coincided with a rapid increase in fluoroquinolone resistance (23). qnr genes were found to be cocarried with various extended-spectrum or AmpC-type beta-lactamases on the same plasmid (6, 13, 17), as well as with metallo-beta-lactamase (25).

Recent carbapenem-resistant Enterobacter cloacae isolates from our hospital showed susceptibility to quinolones (9) according to the CLSI breakpoints (MIC ≤ 1 μg/ml). We aimed to describe the potential for emergence of quinolone resistance among these strains by determining the occurrence of the qnr and aac(6′)-Ib-cr genes.

(This work was presented in part at the 47th Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, IL, 17 to 20 September 2007.)

Sixteen unique patient Enterobacter cloacae isolates, resistant to at least one carbapenem (imipenem, meropenem, or ertapenem), collected in Tel Aviv Medical Center, a 1,200-bed tertiary care teaching hospital between January 2004 and May 2006, were included in the study. Antibiotic susceptibilities were determined by Vitek-2 (bioMerieux Inc., Marcy l'Etoile, France). MICs of fluoroquinolones (ciprofloxacin [CIP] and levofloxacin [LEV]) and carbapenems were determined by Etest (AB Biodisk, Solna, Sweden), and MICs of nalidixic acid were determined by agar dilution. Genetic relatedness was analyzed by pulsed-field gel electrophoresis (PFGE). Bacterial DNA was cleaved with 20 U SpeI endonuclease (New England BioLabs) and electrophoresed in a CHEF-DR III apparatus (Bio-Rad Laboratories, Inc., Hercules, CA) (12, 22). DNA macrorestriction patterns were visually compared and interpreted (24). PCR using primers specific for bla_KPC genes (3) and for bla_SHV genes, (22) followed by sequencing was performed. Screening for qnrA, qnrB, and qnrS was carried out using multiplex PCR with primers described by Robicsek et al. (19). Bacterial cell lysates were used as DNA templates (22). Escherichia coli strain J53 carrying various plasmids, plasmids pMG252 (qnrA1), pMG298 (qnrB1), and pMG306 (qnrS1), were used as control strains. Amplification of the complete qnrB2 gene was performed with primer F (ATG GCT CTG GCA CTC GTT GG) and primer R (CTA GCC AAT AAT CGC GAT GC), followed by sequencing. Amplifications were performed using Hot-StarTaq DNA polymerase (Qiagen, Hilden, Germany). PCR conditions were as follows: (i) 15 min at 95°C and (ii) 35 cycles, with 1 cycle consisting of 1 min at 94°C, 1 min at 64°C, 1 min at 72°C, and (iii) 10 min at 72°C. The presence of aac(6′)-Ib-cr was determined by PCR followed by digestion with BtsCI (New England BioLabs, Beverly, MA) (14).

Transconjugation experiments were performed using the filter mating method with E. coli strain HB101 as a recipient. Transconjugants were selected on Muller-Hinton plates containing 0.5 mg/ml streptomycin and 2 μg/ml imipenem or 0.5 mg/ml streptomycin and 100 μg/ml ampicillin.

Plasmid DNA from clinical strains and their transformants were isolated using NucleoBond PC 100 midi kit (Macherey-Nagel, Germany). Transformation experiments were carried out by electroporation (electroporator 2510; Eppendorf, Hamburg, Germany) into an E. coli strain GeneHogs (recA1 mutant) (Invitrogen, Carlsbad, CA). Transformants were selected on LB agar plates containing 100 μg/ml ampicillin. Transformed colonies were screened by PCR for the presence of bla_KPC, bla_SHV, and qnr genes. Plasmid size was estimated using BAC-Tracker supercoiled DNA ladder (Epicenter Biotechnologies, Madison, WI). Further size estimation of the transforming plasmids was performed by digestion with S1 nuclease, followed by PFGE. DNA was prepared as described previously (12, 22) and digested with S1 nuclease (Promega, Madison, WI) (1). Electrophoresis was carried out as described previously (8). Lambda ladder PFG marker (New England Biolabs) was used as a molecular size marker. For Southern blot analysis, plasmid DNA from transformants that was digested with EcoRI (New England Biolabs) or not digested was electrophoresed, transferred to a Hybond N⁺ membrane (Amersham Biosciences, Buckinghamshire, United Kingdom), and cross-linked with UV light. A 469-bp PCR amplicon (nucleotides 138 to 607) was used to generate a probe for qnrB. The complete nucleotide sequence of bla_KPC-2 (892 bp) was used as a probe for KPC-2. Labeling of probes was performed using random primer DNA labeling mixture (Biological Industries, Beit Haemek, Israel).

The 16 studied carbapenem-resistant E. cloacae isolates belonged to three different genetic clones, clones A, B, and C according to the criteria established by Tenover et al. (24) (Table 1). All isolates were resistant to aztreonam and gentamicin. Fifteen isolates were resistant to ceftazidime, and isolate 151 had a MIC for ceftazidime of 16 μg/ml. Resistance to cefepime varied (data not shown). All isolates were susceptible to quinolones and amikacin, except for isolate 837 that had a MIC for amikacin of 32 μg/ml and carried aac(6′)-Ib.

TABLE 1.

PFGE genetic clones, resistance genes, and antibiotic susceptibilities of E. cloacae clinical isolates and their transformants

Bacteria	PFGE genetic clone	Resistance genesa	MIC (μg/ml) for antibioticb:
			CIP	LEV	NAL	GEN	IMP	MP	ERT	CAZ	ATM	AMK	PIP	TZP
E. cloacae clinical isolates
1	A	qnrB2 blaKPC-2blaSHV-12	0.19	0.38	16	≥16	16	2	32	≥64	≥64	≤2	≥128	≥128
18	A	qnrB2 blaKPC-2blaSHV-12	0.25	0.5	16	≥16	8	>32	4	≥64	≥64	≤2	≥128	≥128
140	A	blaKPC-2blaSHV-12	0.032	0.094	4	≥16	>32	4	>32	≥64	≥64	≤2	≥128	≥128
151	A	qnrB2 blaKPC-2	0.25	0.5	16	≥16	12	>32	>32	16	≥64	≤2	≥128	≥128
185	A	qnrB2 blaKPC-2blaSHV-12	0.25	0.5	16	≥16	8	24	4	≥64	≥64	≤2	≥128	≥128
293	A	qnrB2 aac(6′)-Ib-cr blaKPC-2blaSHV-12	1.5	0.75	16	≥16	>32	8	>32	≥64	≥64	16	≥128	≥128
533	A	qnrB2 blaKPC-2blaSHV-12	0.19	0.38	16	≥16	4	2	24	≥64	≥64	≤2	≥128	≥128
622	A	qnrB2 blaKPC-2blaSHV-12	0.25	0.5	16	≥16	24	8	32	≥64	≥64	≤2	≥128	≥128
625	A	qnrB2 blaKPC-2blaSHV-12	0.38	0.5	16	≥16	16	6	24	≥64	≥64	≤2	≥128	≥128
656	A	qnrB2 blaKPC-2blaSHV-12	0.25	0.5	16	≥16	32	6	16	≥64	≥64	≤2	≥128	≥128
354	B	blaKPC-2blaSHV-12	0.094	0.19	16	≥16	>32	>32	>32	≥64	≥64	≤2	≥128	≥128
649	B	qnrB2 blaKPC-2blaSHV-12	0.38	0.75	16	≥16	32	4	16	≥64	≥64	≤2	≥128	≥128
658	B	qnrB2 blaKPC-2blaSHV-12	0.38	1	16	≥16	>32	24	12	≥64	≥64	≤2	≥128	≥128
781	B	qnrB2 blaKPC-2blaSHV-12	0.25	0.5	16	≥16	>32	>32	>32	≥64	≥64	≤2	≥128	≥128
837	B	qnrB2 blaKPC-2blaSHV-12	0.25	0.75	16	≥16	6	6	8	≥64	≥64	32	≥128	≥128
4469	C	qnrB2 blaKPC-2blaSHV-12	0.25	0.5	16	≥16	24	8	>32	≥64	≥64	≤2	≥128	64
Transformants
T-622		qnrB2 blaKPC-2blaSHV-12	0.023	0.023	2	≥16	4	3	8	32	≥64	≤2	≥128	64
T-625		qnrB2 blaKPC-2blaSHV-12	0.047	0.094	4	≥16	8	3	8	32	≥64	≤2	≥128	≥128
T-781		qnrB2 blaKPC-2blaSHV-12	0.023	0.047	2	≥16	3	2	12	32	32	≤2	≥128	64
GeneHogs recipient strain			0.008	0.012	2	≤1	0.125	0.016	0.012	≤1	≤1	≤2	≤4	≤4

^aResistance genes from positive PCRs and sequencing for bla_KPC-2, bla_SHV-12, and qnrB2 genes.

^bCIP, ciprofloxacin; LEV, levofloxacin; NAL, nalidixic acid; GEN, gentamicin; IMP, imipenem; MP, meropenem; ERT, ertapenem; CAZ, ceftazidime; ATM, aztreonam; AMK, amikacin; PIP, piperacillin; TZP, piperacillin-tazobactam.

Molecular analysis of antibiotic resistance genes revealed the presence of bla_KPC-2 in all isolates, and 14 of these isolates also carried the qnrB gene. The two isolates negative for qnrB (isolates 140 and 354) had the lowest MICs for CIP and LEV (Table 1). Fifteen of the isolates also possessed bla_SHV-12. The complete amplified qnrB gene was sequenced and proved to be qnrB2. Screening for the presence of aac(6′)-Ib-cr, encoding the modified aminoglycoside acetyltransferase, was positive in one isolate (isolate 293) that had the highest MIC for CIP (1.5 μg/ml [Table 1]). The plasmid patterns of all strains belonging to genetic clones A and B were similar and revealed several plasmids ranging in size from ca. 8 kb to larger than 164 kb. The genetic clone C isolate contained two additional plasmids of about 28 and 38 kb (data not shown). Attempts to transfer qnrB2- and bla_KPC-2-carrying plasmids from eight isolates (belonging to clones A, B, and C) to E. coli strain HB101 failed, suggesting that these plasmids are not conjugable.

Transformation experiments succeeded with genetic clones A and B and failed with clone C. Transformants T-622, T-625, and T-781 were further characterized (Table 1). PCR analysis of this plasmid revealed the presence of qnrB2 accompanied by bla_KPC-2 and bla_SHV-12. All transformants showed increased MICs to quinolones upon acquisition of this plasmid; transformants T-622 and T-781 exhibited threefold- and twofold-higher MICs for CIP and LEV, respectively, whereas T-625 exhibited sixfold- and fourfold-higher MICs for CIP and LEV, respectively (Table 1). MICs of carbapenems increased in all transformants, and resistance to gentamicin, ceftazidime, piperacillin, and piperacillin-tazobactam was transferred in all of the transformants (Table 1). All transformed strains contained a single plasmid (Fig. 1A and B, panel I). Plasmids from transformants T-622 and T-781 had similar sizes and migrated with an apparent size of lambda bands of 291 to 339.5 kbp. Plasmid from transformant T-625 migrated in the same size range but seemed to be larger (Fig. 1A, lane 2). These findings revealed that the two E. cloacae isolates that belonged to clones A and B possessed plasmids of similar size, whereas the two clone A isolates (622 and 625) possessed different-size plasmids. To further confirm the coexistence of bla_KPC-2 and qnrB2 on the same plasmid and to examine the similarity between the plasmids carrying them, Southern blot analysis was performed. Hybridization of uncut plasmids from transformed strains T-622, T-625, and T-781 confirmed the presence of bla_KPC-2 and qnrB2 on the transforming plasmids (Fig. 1B, panels II and III). The hybridization patterns of transformants digested with EcoRI and using qnrB2 or bla_KPC-2 as a probe were similar (Fig. 1C, panels I and II).

FIG. 1.

(A) PFGE of transformants T-622 (lane 1), T-625 (lane 2), and T-781 (lane 3) digested with S1 nuclease. Transformants T-622 and T-625 originated from E. cloacae clone A isolates, and T-781 originated from a clone B isolate. Lane M, lambda ladder PFG marker (New England Biolabs). Plasmid DNAs are indicated by white arrows. (B) Panel I shows electrophoresis of the plasmids that transformed the bacterial strain (lanes 1 to 3). Plasmid DNA (p) and chromosomal DNA (c) are indicated by arrows. Lane M, BAC-Tracker supercoiled DNA ladder (Epicenter). Panels II and III show the results of Southern blot analysis of undigested transforming plasmids hybridized with a 469-bp fragment of qnrB2 (II) and with bla_KPC-2 (KPC-2) (III). The positions of molecular size standards are indicated to the left of the gels in panels A and B. (C) Panels I and II show plasmid DNA digested with EcoRI and hybridized with qnrB2 and bla_KPC-2 (KPC-2). The positions of molecular size markers (1-kb DNA ladder; Fermentas)are indicated to the left of the gel.

All bla_KPC-2 carbapenem-resistant E. cloacae strains included in this study were susceptible to quinolones. Fourteen out of 16 strains were found to carry the qnrB2 gene, and one strain also carried aac(6′)-Ib-cr. This isolate had a higher level of resistance to CIP than the isolates carrying only qnrB2, similar to a previous report on qnrA (18). Transformations studied by PCR and Southern blot analysis proved the coexistence of qnrB2 and bla_KPC-2 on the same plasmid.

qnr determinants alone may not confer resistance to quinolones (10), but they can supplement other quinolone resistance mechanisms (11, 16). In our study, 14 isolates carrying qnrB2 were classified as ciprofloxacin susceptible by CLSI criteria. Thus, treatment with a fluoroquinolone might easily select for resistant strains. Our finding that 14 of 16 carbapenem-resistant E. cloacae strains, isolated between 2004 and 2006, carried qnrB2 is similar to a previous report from Israel (23); qnrA was more prevalent until 2001, but since then, qnrB has predominated (23).

qnr determinants are usually associated with additional antimicrobial-resistant determinants (6, 13, 23, 25). This is the first report of the association of qnrB2 with bla_KPC-2 carbapenemase on a single plasmid in two genetically unrelated clones of E. cloacae. In addition to the potential of these strains to develop high-level resistance to quinolones, the coexistence of multiple resistant genes on the same plasmid, which may confer resistance to both carbapenems and quinolones, poses a serious epidemiological, clinical, and public health threat.