First Detection of Plasmid-Encoded bla_OXY β-Lactamase

Abstract

Three Klebsiella oxytoca isolates and one Klebsiella pneumoniae isolate from three children admitted to the Hematology Unit of Hospital Vall d'Hebron (Barcelona, Spain) exhibited a susceptibility pattern suggesting OXY β-lactamase hyperproduction. All the isolates contained a 95-kb plasmid that harbored bla_OXY-1, which was transferred by electrotransformation but could not be self-transferred by conjugation. A qnrS1 gene was also harbored in the bla_OXY-1-carrying plasmid. This is the first report of a plasmid-encoded OXY β-lactamase.

Klebsiella oxytoca is a member of the Enterobacteriaceae with broad environmental distribution; it also colonizes the human gut and is able to produce severe opportunistic infections, particularly in neonates (2, 9). This species possesses a chromosomally encoded class A β-lactamase (bla_OXY) which is constitutively expressed at low levels, conferring resistance to aminopenicillins and carboxypenicillins. However, such strains remain susceptible to other β-lactams such as cephalosporins, monobactams, and the β-lactamase inhibitor combination agents. However, hyperproduction of OXY, which generally is due to mutations in the promoter region of the gene, confers resistance to all penicillins (except temocillin) and to the combination of amoxicillin (amoxicilline) plus clavulanic acid, narrow- and expanded-spectrum cephalosporins, and aztreonam, plus a variable level of reduced susceptibility to cefotaxime and ceftriaxone, without affecting susceptibility to ceftazidime (1, 6, 7, 13).

Six bla_OXY groups (bla_OXY-1 to bla_OXY-6) have been described on the basis of the gene nucleotide sequence (5). The different bla_OXY groups have similar profiles of activity against β-lactams; however, three bla_OXY-2 derivatives have been described which confer resistance to β-lactamase inhibitor combination agents or to ceftazidime by point amino acid substitutions in different positions (14, 18, 20).

In January 2008, a Klebsiella oxytoca isolate (KO279) was obtained from the urine of a 3-year-old boy who was undergoing bone marrow transplantation. The isolate had a phenotype of resistance to β-lactams compatible with the presence of bla_OXY hyperproduction (Table 1). Simultaneously, a Klebsiella pneumoniae isolate (KP278) with a pattern of susceptibility to β-lactams similar to that of K. oxytoca KO279 was obtained from the urine of a 4-year-old girl, admitted in the same nursing unit also for bone marrow transplantation.

TABLE 1.

MICs of antimicrobial agents for the studied strains

Straina	Plasmid-encoded β-lactamase(s)	MIC (μg/ml) of drugb:
		AMP	AMC	FOX	CAZ	CTX	FEP	ATM	IPM	NOR	CIP
K. oxytoca
KO279	OXY-1	≥256	48	3	4	1.5	1.5	≥256	0.25	1.5	0.38
KO280	OXY-1	≥256	48	3	4	1.5	1.5	≥256	0.25	1	0.25
KO281	OXY-1 + CMY-2	≥256	≥256	≥256	≥256	16	1.5	≥256	0.25	1.5	0.38
K. pneumoniae
KP101A	CMY-2	≥256	≥256	≥256	64	6	0.5	6	0.38	0.125	0.047
KP278	OXY-1	≥256	64	4	1.5	1	1	32	0.25	1	0.38
KP278sp		96	2	4	1	0.047	0.19	0.064	0.25	0.125	0.047
E. coli
TF-KO279	OXY-1	≥256	32	2	0.75	0.25	0.125	64	0.19	0.5	0.125
TF-KO281	OXY-1	≥256	32	2	1	0.38	0.38	128	0.19	0.5	0.125
TF-KP278	OXY-1	≥256	32	2	0.75	0.25	0.125	64	0.19	0.38	0.125
DH5α		4	4	2	0.25	0.023	0.032	0.023	0.19	0.094	0.032

^aK. oxytoca KO279, KO280, and KO281, clinical isolates; K. pneumoniae KP101A and KP278, clinical isolates; K. pneumoniae KP278sp, K. pneumoniae KP278 derivative cured of the bla_OXY plasmid; E. coli DH5α, recipient strain in transformation experiments; E. coli TF-KO279, E. coli TF-KO281, and E. coli TF-KP278, transformant isolates carrying bla_OXY plasmids from K. oxytoca KO279, K. oxytoca KO281, and K. pneumoniae KP278, respectively.

^bAMP, ampicillin; AMC, amoxicillin-clavulanate; CIP, ciprofloxacin; CTX, cefotaxime; FOX, cefoxitin; CAZ, ceftazidime; FEP, cefepime; IPM, imipenem; NOR, norfloxacin; ATM, aztreonam.

A search for patients on this unit colonized with enterobacterial isolates with resistance phenotypes similar to those of the above strains was performed by plating stool samples on MacConkey agar supplemented with 2 μg/ml of cefotaxime and MacConkey agar supplemented with 10 μg/ml of aztreonam. This process identified a K. oxytoca isolate (KO280) in the feces of a different 3-year-old male bone marrow transplant recipient. In addition, a K. pneumoniae isolate (KP101A) and a K. oxytoca isolate (KO281), each carrying a bla_CMY-2 β-lactamase, were isolated from the feces of the 4-year-old girl whose urine yielded K. pneumoniae KP278. Of the latter two fecal isolates, K. pneumoniae KP101A was resistant to cefoxitin and ceftazidime and susceptible to aztreonam. In contrast, K. oxytoca KO281 was resistant to cefoxitin, ceftazidime, and aztreonam (Table 1). Strain identification was performed using the API 20E system (BioMérieux, Marcy l'Etoile, France) and sequencing of 16S rRNA. In addition, the K. pneumoniae isolates were confirmed to possess the bla_SHV gene, and the K. oxytoca isolates were confirmed to possess the polygalacturonase pehX gene, by PCR using previously described primers (10, 12). MICs of selected antibiotics were determined by Etest (AB Biodisk, Solna, Sweden) in duplicate (Table 1).

Screening for the presence of the bla_OXY β-lactamase gene was performed by PCR using primers OXY-E and OXY-G, as previously described (5). As expected, all K. oxytoca strains yielded amplicons for bla_OXY. Surprisingly, K. pneumoniae KP278 also gave a positive result. K. pneumoniae KP101A did not amplify for bla_OXY. All positive results were confirmed by direct sequencing of both strands. All coding region sequences showed 100% identity to bla_OXY-1 (accession number AY077482), indicating that the β-lactamase belonged to the OXY-1 group. Additionally, an adenine was located in the fifth base of the −10 region of the promoter sequence, a mutation which previously was shown to increase the level of β-lactamase expression (7).

Strain typing by pulsed-field gel electrophoresis (PFGE) was performed with a CHEF DRII system (Bio-Rad, Richmond, CA) after restriction of total DNA with XbaI, as previously described (8). All the K. oxytoca isolates exhibited indistinguishable PFGE profiles, as did all the K. pneumoniae isolates (Fig. 1).

FIG. 1.

(a and b) PFGE analysis of the studied strains digested with S1 endonuclease (a) and Southern blot hybridization with a bla_OXY-1 probe (b). (c and d) PFGE analysis of the studied strains digested with XbaI (c) and Southern blot hybridization with a bla_OXY-1 probe (d). Hybridization bands at approximately 95 kb in panels b and d correspond with the plasmid harboring bla_OXY-1. Hybridization bands at the compression zone in panel b and at approximately 330 kb in panel d may correspond with the chromosomally carried bla_OXY-1 gene in K. oxytoca strains. Lanes MK, molecular size markers (numbers at left in kb).

Based on these results, we hypothesized the presence of a plasmid-encoded bla_OXY gene in K. pneumoniae KP278. To test this hypothesis, the plasmid location of the bla_OXY gene and the size of the corresponding plasmid were determined for all bla_OXY-positive isolates by S1 nuclease digestion as previously described (8). S1-digested total DNA from PFGE gels was transferred to positively charged nylon membranes and hybridized with specific probes for bla_OXY. This showed that the gene was located on a plasmid of approximately 95 kb in K. pneumoniae KP278 and all the K. oxytoca strains (Fig. 1). Despite K. oxytoca possessing a chromosomally encoded OXY β-lactamase, no double sequences of the bla_OXY gene were observed in any of these strains, suggesting that the plasmid and chromosomally carried genes exhibited 100% identity. In order to determine if the plasmid harboring bla_OXY in K. oxytoca KO281 also contained the bla_CMY-2 gene, the same membrane was hybridized with a specific probe for bla_CMY-2. This showed that bla_CMY-2 was located on a different plasmid, of approximately 60 kb (data not shown).

Plasmid curing was performed for KP278 and KO280 by using acridine orange. Strains were grown at 42°C in brain heart infusion broth with 80 μg/ml acridine orange (Sigma-Aldrich Inc., Steinheim, Germany) for 24 h with shaking. Cells which had lost the plasmid containing the bla_OXY gene were selected by negative selection-replica plating on LB agar supplemented with aztreonam (20 μg/ml). K. pneumoniae KP278 could be cured (KP278sp) of the plasmid harboring bla_OXY, with reversion to the wild-type resistance phenotype of K. pneumoniae, including susceptibility to amoxicillin-clavulanate and aztreonam (Table 1). A plasmid-cured derivative of K. oxytoca KO280 could not be selected by this method, presumably due to the chromosomally hyperproduced bla_OXY gene. No other antibiotic resistance marker was observed to be cotransferred on the same plasmid that could be used for negative selection.

Conjugation experiments involving the bla_OXY-positive isolates were performed by using a liquid mating assay as previously described (8), with a rifampin (rifampicin)-resistant derivative of Escherichia coli HB101 as the recipient, and selection on LB agar supplemented with 100 μg/ml of rifampin and 50 μg/ml of ampicillin. No transconjugants were obtained after several attempts, suggesting that the plasmid was not self-transferable.

Plasmid mobilization by electrotransformation of the bla_OXY-carrying plasmid of K. pneumoniae KP278, K. oxytoca KO279, and K. oxytoca KO281 was performed into the nalidixic acid-resistant derivative E. coli DH5α. Transformants (TF-KP278, TF-KO279, and TF-KO281, respectively) were selected on LB agar plates supplemented with 10 μg/ml of aztreonam and confirmed by PCR and sequencing. Both transformants showed high MICs to ampicillin, amoxicillin-clavulanate, and aztreonam (Table 1). Conjugation experiments using E. coli transformants as donors also gave negative results, providing additional evidence that the plasmid was not self-transferable.

Susceptibility to amikacin, gentamicin, kanamycin, streptomycin, tobramycin, chloramphenicol, tetracycline, ciprofloxacin, and sulfonamides was determined by disk diffusion for the transformant E. coli strains in order to assess resistance cotransferred with the bla_OXY-carrying plasmid. No coresistance to the studied antimicrobials was observed except to ciprofloxacin, for which reduced susceptibility was detected and was confirmed by Etest (Table 1). Based on this finding, the studied strains were screened for the qnrA, qnrB, qnrS, aac(6′)-Ib-cr, and qepA genes by PCR as previously described (3, 12, 16). PCR and sequencing showed that qnrS1 was present in all the strains harboring the bla_OXY-carrying plasmid, and Southern blot hybridization revealed that the two genes were located in the same plasmid (data not shown).

A PCR-based search for the mobile elements ISCR1 and ISEcp1, which are responsible for mobilizations of other bla genes, such as those of the CTX-M-type and AmpC-type β-lactamases, was performed in the transformants by using previously described primers (11, 19). This yielded negative results, suggesting the involvement of a different mobilization mechanism.

This report describes for the first time a plasmid location of the OXY β-lactamase bla_OXY in two different species of Enterobacteriaceae, K. oxytoca and K. pneumoniae. Several plasmid-encoded β-lactamases, such as the AmpC type, SHV, and the CTX-M type, have been shown to derive from chromosomally encoded β-lactamases (4, 15, 17). The plasmid location of these antimicrobial resistance genes has facilitated their spread. As a result, species that previously lacked chromosomally encoded β-lactamases have become resistant, and species that already possess a chromosomally encoded narrow-spectrum β-lactamase have broadened their resistance phenotype by acquisition of extended-spectrum β-lactamases. The occurrence of OXY β-lactamase on a plasmid could facilitate its horizontal transmission among different bacteria, as demonstrated in this work. Early detection in order to institute isolation measures could avoid the spread of such strains. However, detection of plasmid-located bla_OXY β-lactamase in K. oxytoca strains that are already bla_OXY hyperproducers could be difficult in clinical laboratories if molecular techniques are not used.