Biochemical and Genetic Characterization of Carbapenem-Hydrolyzing β-Lactamase OXA-229 from Acinetobacter bereziniae

Abstract

Acinetobacter bereziniae (formerly Acinetobacter genomospecies 10) isolate Nec was recovered from a skin sample of a patient hospitalized in Paris, France. It was resistant to penicillins, penicillin-inhibitor combinations, and carbapenems. Cloning and expression in Escherichia coli identified the carbapenem-hydrolyzing class D β-lactamase OXA-229, which is weakly related to other oxacillinases (66% amino acid identity with the closest oxacillinase, OXA-58). It hydrolyzed penicillins, oxacillin, and imipenem but not expanded-spectrum cephalosporins. Sequencing of the genetic context of the bla_OXA-229 gene did not identify an insertion sequence but did identify mutations in the promoter sequences in comparison to the fully susceptible A. bereziniae reference strain. The overexpression of bla_OXA-229 in A. bereziniae Nec as a source of carbapenem resistance was identified by quantitative real-time PCR.

INTRODUCTION

Acinetobacter bereziniae, formely Acinetobacter genomospecies 10, was described as a new species of the genus Acinetobacter on the basis of DNA-DNA hybridization in 1986 by Bouvet et Grimont (1). A. bereziniae is a Gram-negative, strictly aerobic, oxidase-negative, and nonmotile coccobacillus (15). A. bereziniae was found to be responsible for health care-associated infections, including severe sepsis in an immunocompromised patient (12). Previous reports of β-lactamase-mediated resistance to carbapenems in A. bereziniae were related to the production of metallo-β-lactamases, namely, IMP-1, SIM-1, and VIM-2 (14, 16).

Some bacterial species naturally carry class D β-lactamase genes within their genome (22). In Acinetobacter species, intrinsically chromosomal genes encoding carbapenem-hydrolyzing class D β-lactamases (CHDLs) have been identified. Acinetobacter species A. baumannii naturally harbors bla_OXA-51-like genes, A. radioresistens harbors bla_OXA-23-like genes, and A. lwoffii bla_OXA-134-like genes, and recently A. johnsonii, A. calcoaceticus, and A. haemolyticus have been shown to harbor bla_OXA-211-like, bla_OXA-213-like, and bla_OXA-214-like genes, respectively (4, 6, 9, 19) (Fig. 1). Characterization of those naturally occurring CHDL-encoding genes provides additional tools for identification of Acinetobacter species.

Fig 1.

Dendrogram obtained for 13 class D β-lactamases by neighbor-joining analysis. The alignment used for tree calculation was performed with ClustalX. Branch lengths are drawn to scale and are proportional to the number of amino acid changes. The distance along the vertical axis has no significance. The class D β-lactamases, which are considered to be naturally occurring, are indicated together with the names of the corresponding species.

The present study was initiated by the isolation of two imipenem-nonsusceptible and non-baumannii Acinetobacter species.

MATERIALS AND METHODS

Bacterial strains and plasmids.

Identification of A. bereziniae isolates Nec and Baz was performed by using the API 32GN system (bioMérieux, Marcy l'Etoile, France) and was confirmed by 16S rRNA and partial rpoB gene sequencing, as described previously (7, 11). Reference strain Escherichia coli TOP10 was used as a host for clonings. A. bereziniae strain CIP70.12, isolated from a human wound in 1970, was used as the reference strain (15).

MIC determinations.

The antibiotic susceptibility profiles of the clinical and reference strains were determined by the agar dilution method, and the results were interpreted according to CLSI guidelines (2). MICs of cephalosporins and carbapenems were determined by using Etest strips (bioMérieux, Marcy-l'Étoile, France).

PCR and cloning experiments.

Whole-cell DNA of Acinetobacter spp. was extracted as described previously (20). PCRs were performed under standard conditions using Taq DNA polymerase (Applied Biosystems, Courtaboeuf, France), as described previously (19). Screening for the CHDL-encoding genes was performed by PCR using degenerate internal primers OXA-CHDL A (5′-CCHGCHTCDACHTTYAARAT-3′) and OXA-CHDL B (5′-KYHAYABCCMWKSCCCADCC-3′) with H corresponding to A, T, or C; D to G, A, or T; Y to C or T; R to A or G; K to G or T; S to G or C; B to G, T, or C; and M to A or C, as described previously (4). This DNA was used as a template under standard PCR conditions with a series of primers designed for the detection of acquired (bla_OXA-23, bla_OXA-40, bla_OXA-58, and bla_OXA-143) or naturally occurring class D β-lactamase genes (bla_OXA-51, bla_OXA-134, bla_OXA-211, bla_OXA-213, and bla_OXA-214) (4, 6, 8–10, 21). The genetic environment of the bla_OXA-229 gene was determined by a shotgun cloning procedure using plasmid pBKCMV as described previously (21). For that purpose, BamHI-, HindIII-, or SalI-restricted DNAs were ligated into pBKCMV and introduced into E. coli TOP10 by electroporation as described previously (23). Recombinant plasmids were selected on tryptic soy (TS) agar plates containing ticarcillin (50 μg/ml) and kanamycin (30 μg/ml). All enzymes for DNA manipulations were used according to the recommendations of the supplier (GE Healthcare, Orsay, France). Recombinant plasmids were sequenced by using combinations of universal T3 and T7 primers and specific gene primers, designed on the basis of sequences obtained on an Applied Biosystems sequencer (ABI 3130).

Determination of the transcription initiation sites by 5′-RACE.

Total RNA was isolated from the E. coli (pOXA-229) recombinant strain with an RNeasy Midi kit (Qiagen, Courtaboeuf, France) using RNAprotect (Qiagen) following the recommendations of the manufacturer. The 5′ rapid amplification of cDNA ends (5′-RACE) reactions were performed with 5 μg of total RNA of E. coli (pOXA-229) and a 5′-RACE system kit (version 2.0; Invitrogen, Cergy-Pontoise, France), following the recommendations of the manufacturer. The first-strand synthesis was primed with the specific primer OXA-229-5′ext.1 (5′-GCTTTTGCATTTTGCAAGCC-3′), and amplification of the target cDNA was performed with the dC-tailed cDNAs as templates by using the newly described primers OXA-229-5′ext.2 (5′-ATATTCAGTCTTTGCTCGGC-3′) and OXA-229-5′ext.3 (5′-TCTCGCTCTGATATGACTGG-3′).

Expression of the bla_OXA-228-like gene.

Real-time reverse transcription-PCR (RT-PCR) was performed to measure mRNA expression levels of bla_OXA-228, bla_OXA-229, bla_OXA-230, and the chromosomal 16S rRNA-encoding genes. One-step RT-PCR was performed using a Rotor-Gene (Qiagen) with the Rotor-Gene SYBR green RT-PCR kit (Qiagen) with 100 ng of total RNA and 1 μmol/liter of primer in a total volume of 25 μl. Relative transcript levels were calculated using the 2^−ΔΔCT method (24) (ratio = 2^−ΔΔCT, where ΔΔCT = ΔCT_reference − ΔCT_target). Levels of 16S rRNA transcription were used as internal controls to normalize the data. At least three independent RNA samples isolated from three separate cultures were used to determine average transcript levels for each strain.

β-Lactamase purification.

Cultures of E. coli TOP10 harboring the recombinant plasmid pBK-OXA-229 were grown overnight at 37°C in 4 liters of TS broth containing amoxicillin (50 μg/ml) and kanamycin (30 μg/ml). β-Lactamase OXA-229 was purified by ion-exchange chromatography. Briefly, the bacterial suspension was pelleted, resuspended in 40 ml of 100 mM sodium phosphate buffer (pH 7.0), sonicated on ice, cleared by ultracentrifugation, and treated with DNase. The extract was then dialyzed against 20 mM bis-Tris-H₂SO₄ {[bis(2-hydroxyethyl-imino]tris(hydroxymethyl)methane} (pH 6.4) and loaded onto a preequilibrated Q-Sepharose column with the same buffer. The resulting enzyme extract was recovered in the flowthrough and dialyzed against 20 mM Tris-H₂SO₄ (pH 9) overnight at 4°C. This extract was then loaded onto a preequilibrated (20 mM Tris-H₂SO₄ [pH 9]) Q-Sepharose column, and the proteins were eluted with a linear K₂SO₄ gradient (0 to 0.5 M). Finally, fractions containing the highest β-lactamase activities were pooled and subsequently dialyzed overnight against 100 mM phosphate buffer (pH 7.0). The β-lactamase activity was determined qualitatively using nitrocefin hydrolysis (Oxoid, Dardilly, France). The protein content was measured using the Bio-Rad DC protein assay.

Kinetic studies.

Kinetic measurements (k_cat and K_m) of purified β-lactamase OXA-229 were performed as described previously (18). The 50% inhibitory concentration (IC₅₀) was determined for OXA-229 as the concentration of clavulanate or tazobactam that reduced the hydrolysis rate of 100 μM benzylpenicillin by 50%, under conditions in which OXA-229 was preincubated with various concentrations of inhibitor for 10 min at 30°C, before the substrate was added.

Nucleotide sequence accession numbers.

The nucleotide sequences of genes bla_OXA-228 from A. bereziniae CIP 70.12, bla_OXA-229 from A. bereziniae clinical isolate Nec, and bla_OXA-230 from clinical isolate Baz reported in this article are available in the GenBank nucleotide database under the following accession numbers: JQ422053 for OXA-228, JQ422052 for OXA-229, and JQ422054 for OXA-230.

RESULTS AND DISCUSSION

Identification of clinical isolates and susceptibility testing.

A. bereziniae Nec was isolated from a skin sample of a patient hospitalized at the Necker-Enfants-Malades university hospital, Paris, France, in August 2011. This isolate was first identified as Acinetobacter junii using the API32GN method. Partial sequencing of 16S rRNA and rpoB genes revealed that it actually belonged to the A. bereziniae species. It was resistant to penicillins, penicillin-inhibitor combinations, and carbapenems as observed for many CHDL-producing Acinetobacter species isolates (Table 1). A second isolate, named A. bereziniae Baz, was isolated from the urine culture of a patient hospitalized in another French hospital in September 2011. It was also identified at the species level using molecular methods, although the phenotypic identification technique identified the isolate as Acinetobacter lwoffii. This isolate exhibited a resistance pattern similar to that of A. bereziniae Nec, but with a lower level of resistance to β-lactams. For comparison, the susceptibility pattern of A. bereziniae reference strain CIP70.12 was evaluated, showing that it was fully susceptible to all β-lactams (Table 1).

Table 1.

MICs of β-lactams for A. bereziniae isolate Nec, A. bereziniae isolate Baz, E. coli (pBK-OXA-229) recombinant strain, A. bereziniae CIP70.12, and E. coli TOP10 reference strains

β-Lactam	MIC (μg/ml)
	A. bereziniae isolate Nec	A. bereziniae isolate Baz	A. bereziniae isolate CIP 70.12	E. coli TOP10 (pBK-OXA-229)	E. coli TOP10
Amoxicillin	>256	96	12	>256	2
Amoxicillin + CLAa	>256	96	12	128	2
Ticarcillin	128	32	4	256	2
Ticarcillin + CLA	128	32	4	128	2
Cephalothin	>256	256	256	16	4
Cefotaxime	2	0.5	0.5	0.06	0.06
Ceftazidime	3	2	2	0.12	0.12
Cefepime	0.5	0.5	0.5	0.06	0.06
Aztreonam	32	24	16	0.03	0.03
Meropenem	12	4	0.5	0.06	0.03
Doripenem	8	2	0.5	0.06	0.03
Imipenem	12	2	0.38	0.12	0.06

^aCLA, clavulanic acid (4 μg/ml).

Identification of β-lactamase genes.

PCR for detection of acquired CHDLs in the three A. bereziniae strains failed. Moreover, PCR to detect naturally occurring CHDLs also failed. PCR using degenerated primers OXA-CHDL-A and OXA-CHDL-B gave a 485-bp amplicon. Further sequencing of the amplicons obtained from A. bereziniae identified a gene encoding a novel class D β-lactamase, named OXA-229 (www.lahey.org/studies).

Cloning and sequencing of CHDL-encoding genes.

Cloning of HindIII-restricted whole-cell DNA of A. bereziniae Nec into pBK-CMV, followed by expression in E. coli TOP10, gave E. coli TOP10 (pOXA-229), which was resistant to penicillins and exhibited reduced susceptibility to imipenem (Table 1). Sequencing of the insert of pOXA-229 identified the bla_OXA-229 gene, encoding class D β-lactamase OXA-229, possessing 276 amino acids. Within the deduced protein encoded by this open reading frame (ORF), a serine-threonine-phenylalanine-lysine tetrad (STFK) was found at positions 70 to 73 according to the numbering of class D β-lactamases (3). A KSG element (positions 216 to 218) was found, as observed in the carbapenem-hydrolyzing β-lactamases OXA-40, -51, -134, and -143, whereas a KTG motif is present in the carbapenem-hydrolyzing oxacillinases OXA-23, OXA-48, OXA-54, and OXA-55 and in most of the class D β-lactamases without carbapenemase activity (3, 13, 22) (see Fig. S1 in the supplemental material). The class D structural element YGN at positions DBL144 to 146 was replaced in OXA-229 by an FGN motif. OXA-229 shared 66% amino acid identity with OXA-58 and was weakly related to other CHDLs, showing 56%, 56%, and 60% amino acid identity with OXA-23, OXA-40, and OXA-143, respectively (see Fig. S1 in the supplemental material).

Primers specific for bla_OXA-229 were designed and used for PCR amplification of DNAs of A. bereziniae Baz and CIP70.12. It identified bla_OXA-228 and bla_OXA-230, respectively encoding variants OXA-228 and OXA-230, exhibiting four and six amino acid substitutions compared to OXA-229.

Biochemical properties.

After purification from extracts of E. coli TOP10 (pOXA-229), the specific activity of OXA-229 against benzylpenicillin was found to be 12.5 mU per mg of protein, and its purity was estimated to be 95% by sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis (data not shown). OXA-229 exhibited a narrow-spectrum hydrolysis profile, including mostly penicillins (Table 2). Hydrolysis of imipenem was low, whereas hydrolysis of meropenem was undetectable, as observed for OXA-58 (21). However, determination of the K_i for meropenem revealed a strong affinity (0.03 μM) of OXA-229 for this substrate. In general, catalytic activities of OXA-229 were similar to those of OXA-58, which was taken as the reference for an oxacillinase with carbapenem-hydrolyzing activity (Table 2). Activity against ceftazidime, cefotaxime, and cefepime remained undetectable (Table 2). Studies of activity inhibition, as measured by determination of IC₅₀, showed that OXA-229 was weakly inhibited by clavulanic acid (350 μM) and tazobactam (9 μM), as found for most of the oxacillinases (21).

Table 2.

Kinetic parameters for β-lactamase OXA-229^a

Substrate	Km (μM)		kcat (s−1)		kcat/Km (mM−1/s−1)
	OXA-229	OXA-58b	OXA-229	OXA-58b	OXA-229	OXA-58b
Benzylpenicillin	70	50	11	5.5	150	110
Amoxicillin	405	NA	15.2	NA	38	NA
Ampicillin	127	130	1	1	8	8
Ticarcillin	343	240	1.7	1	5	4
Piperacillin	29	50	2.2	2.5	80	50
Cephalothin	165	150	0.2	0.1	1	1
Cefotaxime	NH	NH	ND	ND
Ceftazidime	NH	NH	ND	ND
Cefepime	NH	NH	ND	ND
Oxacillin	1850	70	4.2	1.5	2	2
Aztreonam	NH	NH	ND	ND
Imipenem	30	7.5	0.09	0.1	3	14
Meropenem	0.03c	0.075c	<0.01	<0.01	<0.1	<0.1

^aData are the means of three independent experiments. Standard deviations were within 10% of the geometrical means. ND, not determinable; NH, no detectable hydrolysis; NA, not available.

^bData from Poirel et al. (21).

^cK_m was obtained as a K_i value.

Genetic context and expression of the bla_OXA-229 gene.

Sequencing of the entire insert of pOXA-229 did not reveal any insertion sequence located upstream or downstream of the bla_OXA-229 gene. Upstream of the bla_OXA-229 gene, a gene encoding a hypothetical outer membrane protein was identified, whereas a gene encoding a GGDEF family protein of unknown function was identified downstream.

By using the 5′-RACE PCR technique, the sites of initiation of transcription of the bla_OXA-229 gene were mapped in A. bereziniae Nec. The +1 transcription start was located 10 bp upstream of the start codon of the bla_OXA-229 gene in A. bereziniae Nec. Upstream of this transcriptional start site, a −35 sequence (TTGAAT) separated with an optimal spacing of 17 bp from a −10 sequence (TAGTAT) constituted a putative promoter (Fig. 2). In A. bereziniae Baz and CIP70.12, the promoter sequences were identical, being TTCAAT and TGGTAT for the −35 and −10 sequences, respectively, but two substitutions occurred in comparison with the promoter sequences of A. bereziniae isolate Nec, suggesting a putative stronger promoter than that observed in isolates Baz and CIP70.12 (Fig. 2). To verify this hypothesis, quantitative reverse RT-PCR was performed to measure the bla_OXA-229 expression in comparison with expression of the bla_OXA-228 and bla_OXA-230 genes. The bla_OXA-229 transcript levels were compared to those of the 16S RNA gene taken as a standard, and analyses were repeated in triplicate. Transcriptional profile analysis indicated a 23-fold-increased expression of bla_OXA-229 in isolate Nec compared with that of isolates Baz and CIP 70.12 (mean ± standard deviation, 23 ± 3.9). The transcript level of the bla_OXA-230 gene from A. bereziniae isolate Baz was also evaluated and was not different from that of bla_OXA-228 from reference strain A. bereziniae CIP 70.12, as expected, since the promoter sequences were identical to that of A. bereziniae CIP 70.12 (Fig. 2). However, MICs for β-lactams of clinical isolate Baz were higher than those of the A. bereziniae CIP 70.12 reference strain. This is probably due to nonenzymatic resistance mechanisms such as active efflux or impermeability, since the MIC for ticarcillin was low.

Fig 2.

Promoter sequences of the bla_OXA-228-like genes in A. bereziniae Nec, Baz, and CIP70.12 strains. The −35 and −10 boxes and the + 1 area of transcription are underlined. Mutations in the promoter are indicated in bold.

Conclusion.

The identification of bla_OXA-228-like genes could be a useful and rapid molecular tool for correct identification of Acinetobacter bereziniae and can be confirmed by sequencing of the 16S rRNA gene. Characterization of this novel oxacillinase further underlines the diversity of CHDL genes in Acinetobacter spp. as the main identified reservoirs of these genes. An increased expression of the bla_OXA-228-like genes was observed in a single isolate, mirroring the β-lactam resistance pattern observed and being mediated by promoter changes and not by strong promoters mediated by insertion sequences as shown for the bla_OXA-51 gene in A. baumannii (5, 22). That finding is different from those related to the overexpression of the intrinsic class D β-lactamase genes in A. baumannii, with insertion sequences ISAba1 and ISAba9 located upstream of the bla_OXA-51-like genes and providing strong promoters (5, 17).