ABSTRACT
We report multifocal detection (four different cities in northern Italy) of Proteus mirabilis isolates resistant to both oxyimino- and 7-α-methoxy-cephalosporins and producing a novel acquired AmpC-like β-lactamase. The enzyme, named CMY-16, is a variant of the CMY/LAT lineage, which differs from the closest homologues, CMY-4 and CMY-12, by a single amino acid substitution (A171S or N363S, respectively) and from CMY-2 by two substitutions (A171S and W221R). Expression of the cloned blaCMY-16 gene in Escherichia coli decreased susceptibility to penicillins, cephalosporins, and aztreonam. Tazobactam was more effective than clavulanate at antagonizing the enzyme activity. Genotyping, by random amplification of polymorphic DNA and pulsed-field gel electrophoresis of genomic DNA digested with SfiI, showed that isolates were clonally related to each other, although not identical. The blaCMY-16 gene was not transferable to E. coli by conjugation or transformation. In all isolates, it was chromosomally located and inserted in a conserved genetic environment. PCR mapping experiments revealed that the blaCMY-16 was flanked by ISEcp1 and the blc gene, similar to other genes of this lineage from plasmids of Salmonella enterica, Klebsiella spp., and E. coli. Overall, these results revealed multifocal spreading of a CMY-16-producing P. mirabilis clone in northern Italy. This finding represents the first report of an acquired AmpC-like β-lactamase in Proteus mirabilis from Italy and underscores the emergence of similar resistance determinants in the European setting.
AmpC-type β-lactamases (CBLs), belonging to Ambler's molecular class C and group 1 of the Bush-Jacoby-Medeiros functional classification (6), are a large group of enzymes of broad substrate specificity. They can degrade penicillins and most cephalosporins (including 7-α-methoxy derivatives), usually being poorly inhibited by serine β-lactamase inactivators. CBLs are usually not active or very poorly active on aztreonam, the zwitterionic oxyimino-cephalosporins (such as cefepime and cefpirome), and imipenem but can contribute to decreased susceptibility or even resistance to these compounds, especially when enzyme overproduction is associated with permeability defects or drug efflux (13, 15).
A number of CBLs are encoded by chromosomal genes resident in some gram-negative pathogens (e.g., Pseudomonas aeruginosa, Acinetobacter spp., and several members of the family Enterobacteriaceae), while others are encoded by genes associated with mobile DNA elements that can be acquired by horizontal gene transfer (15, 23). These acquired CBLs, which are usually plasmid mediated and belong to at least five different lineages, are emerging worldwide in various species of Enterobacteriaceae, such as Klebsiella pneumoniae, Klebsiella oxytoca, Escherichia coli, Salmonella spp., and Proteus mirabilis (2, 14, 16-19, 23, 34). Although the prevalence of the acquired CBLs is not really known, they have become an important cause of resistance to expanded-spectrum β-lactams in some settings (2, 14, 17).
P. mirabilis, which lacks resident chromosomal β-lactamase genes, is entirely dependent upon acquisition of heterologous β-lactamase genes to express a β-lactamase-mediated resistance phenotype (15). A broad repertoire of acquired β-lactamases have been reported in this species, including broad- and extended-spectrum enzymes of molecular class A/functional group 2 (4, 15, 20, 22, 25, 26) and CBLs (8, 14, 23).
Acquired CBLs have been reported in P. mirabilis isolates from several geographic areas (France, Tunisia, Poland, and Korea) and include enzymes of the ACC, CMY/LAT, and DHA/MOR lineages, although members of the CMY/LAT lineage (derived from the chromosomal AmpC enzyme of Citrobacter freundii) are most frequently reported (3, 8, 12, 14, 23, 24, 34). Production of such enzymes is usually responsible for a broad-spectrum β-lactam resistance phenotype, including penicillins, narrow-spectrum cephalosporins, and expanded-spectrum cephalosporins (except the zwitterionic oxyimino-cephalosporins) (23).
In this report, we describe the identification and characterization of a new acquired AmpC-type β-lactamase of the CMY/LAT lineage, CMY-16, in clinical isolates of P. mirabilis from different cities of northern Italy.
MATERIALS AND METHODS
Clinical isolates.The P. mirabilis isolates investigated in this work were isolated at the clinical microbiology laboratories of four hospitals located in different cities of northern Italy (Bergamo, Milan, Novara, and Varese) during the period August 2003 to June 2004. Identification of isolates was carried out by the systems routinely used in each laboratory and confirmed using the Phoenix automated system (BD Diagnostic Systems, Sparks, Md.).
In vitro susceptibility testing.MICs for clinical isolates of P. mirabilis and for E. coli DH5α(pBC-CMY16) were determined using the E test (AB Biodisk, Solna, Sweden). Results were interpreted according to the criteria of the Clinical and Laboratory Standards Institute (7). E. coli ATCC 25922 was used as a reference strain for quality control of in vitro susceptibility testing.
β-Lactamase assays.The double-disk diffusion test for extended-spectrum β-lactamase (ESBL) detection was carried out as described previously (21) by placing clavulanate- or tazobactam-containing disks at a distance of 25 mm (center to center) from disks containing the expanded-spectrum β-lactams. Analytical isoelectric focusing (IEF) of crude cell extracts, detection of β-lactamase bands by nitrocefin, and detection of the activities of the β-lactamase bands separated by IEF against β-lactams by a substrate-overlaying procedure were carried out as described previously (21). Reference strains producing TEM-1, TEM-2, TEM-12, SHV-1, SHV-5, and MIR-1 were used as controls, as described previously (21).
DNA analysis and manipulation methodology.PCR was always carried out in a 50-μl volume, with 30 pmol of each primer, 200 μM deoxynucleoside triphosphates, 1.5 mM MgCl2, and 0.5 U of the Expand High-Fidelity PCR system (Roche Biochemicals, Mannheim, Germany) in the reaction buffer provided by the enzyme manufacturer. The PCR primers used in this study for amplification of plasmid-mediated CBL genes and mapping of flanking regions, and the cycling conditions, are reported in Table 1. PCR for amplification of blaTEM alleles was carried out as described previously (21). Nucleotide sequences were determined on both strands directly on PCR amplification products at an external sequencing facility (Macrogen Inc., Seoul, South Korea). Plasmid pBC-CMY16 was constructed by amplification of the blaCMY-16 coding sequence with primers CMY-Exp_Fw and CMY-Exp_Rev (Table 1) and by cloning the amplification product, digested with XbaI and BamHI, into the E. coli plasmid vector pBC-SK(+) (Stratagene, Inc., La Jolla, Calif.) digested with the same enzymes. The authenticity of the cloned fragment was confirmed by sequencing. The E. coli DH5α strain (27) was used as the host for this plasmid. Total DNA from P. mirabilis was extracted as described previously (27). Southern blot hybridizations were carried out on dried gels as described previously (30), using as a probe an amplification product containing the blaCMY-2 gene generated with primers CMY/F and CMY/R (Table 1) or a 16S rRNA probe obtained by PCR using primers EubA and EubB as described previously (11). The probes were labeled with 32P by the random-priming technique (Rediprime II DNA Labeling system; Amersham Biosciences).
Oligonucleotide primers used in this work
Genotyping methodology.Random amplification of polymorphic DNA (RAPD) fingerprinting was carried out using primer 1254 (5′-CCGCAGCCAA) (1) or AP12h (5′-CGGCCCCTGT) (32). Reactions were carried out in a 25-μl volume with 40 pmol of primer 1254 or AP12h, 240 μM deoxynucleoside triphosphates, 1.5 mM MgCl2, and 0.5 U of the Expand High-Fidelity PCR system in the reaction buffer provided by the enzyme manufacturer. Cycling parameters were as follows: 5 min at 94°C; 35 cycles of 1 min at 94°C, 1 min at 37°C, and 1 min at 72°C; and a final extension step of 10 min at 72°C. Pulsed-field gel electrophoresis (PFGE) profiles of genomic DNA were analyzed by means of the Gene Path procedure (Bio-Rad Laboratories, Richmond, Calif.) using the no. 5 pathogen group reagent kit and the restriction enzyme SfiI (Bio-Rad) or I-CeuI (New England Biolabs, Hertfordshire, United Kingdom). DNA fragments were electrophoresed in 1% agarose gels in 0.5× Tris-borate-EDTA buffer at 14°C and 6 V/cm for 20 h, with pulse times ranging from 5 to 50 s (following SfiI digestion), or in 1% agarose gels in 1× Tris-borate-EDTA buffer at 14°C and 4 V/cm for 48 h, with pulse times ranging from 90 to 325 s (following I-CeuI digestion), using the Gene Path system (Bio-Rad). Bacteriophage λ concatemers (Bio-Rad) or Yeast Chromosome PFG (New England Biolabs) were used as DNA size markers. Clonal relationships based on PFGE patterns were interpreted according to the criteria proposed by Tenover et al. (29).
Gene transfer experiments.Conjugation experiments were performed on a solid medium (Mueller-Hinton agar; Difco Laboratories, Detroit, Mich.) using E. coli J53 (F− met pro Rifr) as the recipient strain. The initial donor/recipient ratio was 0.1. Mating plates were incubated at 37°C for 12 h. Transconjugants were selected on Mueller-Hinton agar containing ampicillin (50 μg/ml) plus rifampin (300 μg/ml). The detection sensitivity of the assay was ≤1 × 10−10 transconjugants/recipient. Electroporation of total-DNA preparations of P. mirabilis in E. coli DH5α was carried out using a Gene Pulser apparatus (Bio-Rad) according to the manufacturer's instructions. Transformants were selected on LB agar containing ampicillin (100 mg/liter).
Nucleotide sequence accession number.The GenBank accession number for blaCMY-16 is AJ781421.
RESULTS AND DISCUSSION
Multifocal detection of P. mirabilis isolates resistant to oxyimino- and 7-α-methoxy-cephalosporins.During the period August 2003 to June 2004, eight nonreplicate clinical isolates of P. mirabilis resistant to oxyimino-cephalosporins (including ceftriaxone, cefotaxime, and ceftazidime) and cefoxitin were detected at the clinical microbiology laboratories of four hospitals located in four different cities of northern Italy. Most isolates were from inpatients in long-term care facilities (LTCFs) or in geriatric or rehabilitation wards, but one isolate (BG-073/03) was from an outpatient (Table 2). Most isolates were from urinary tract infections (Table 2).
Antimicrobial susceptibility to β-lactams and epidemiological features of the eight P. mirabilis isolates
The resistance profiles of the eight isolates, including ampicillin, amoxicillin-clavulanate, cephalothin, ceftriaxone, cefotaxime, ceftazidime, cefoxitin, and fluoroquinolones, were overall similar. All isolates were intermediate to piperacillin and susceptible to piperacillin-tazobactam, aztreonam, cefepime, carbapenems, and aminoglycosides (except one that was resistant to gentamicin) (Table 2). A double-disk synergy test did not reveal any detectable synergy between clavulanate and cefotaxime, ceftazidime, or aztreonam. However, a synergy pattern was detectable between tazobactam or clavulanate and cefepime (data not shown).
Analysis of the β-lactam resistance determinants.Analytical IEF revealed, in all isolates, the presence of two β-lactamase bands, of pI 5.4 and >8.4, respectively (data not shown). In a substrate overlay assay, the pI 5.4 band was unable to hydrolyze cefoxitin, cefotaxime, and ceftazidime, suggesting the presence of a non-ESBL TEM-like enzyme. On the other hand, the alkaline pI band was able to hydrolyze cefoxitin, cefotaxime, and ceftazidime, suggesting the presence of an acquired CBL.
Molecular analysis confirmed the presence of a blaTEM-1b allele in all isolates. Multiplex PCR for detection of plasmid-mediated CBL genes using the AmpC/I-IV set of primers (Table 1) yielded, from all isolates, an amplicon whose size (760 bp) was consistent with the presence of a gene of the blaCMY/LAT lineage. Sequencing the entire gene, amplified with the CMY/F and CMY/R primers (Table 1), revealed the presence in all isolates of identical alleles, which encoded a new CMY variant named CMY-16. Compared with either CMY-4 (31) or CMY-12 (8), which are the closest homologues, CMY-16 has a single amino acid substitution: A171S or N363S, respectively, according to the numbering scheme used by Decre et al. (8). Compared with CMY-2 (CAA62957.1), CMY-16 has two amino acid substitutions: A171S and W221R. At the nucleotide sequence level, blaCMY-16 exhibits two point mutations compared with blaCMY-4 (T511G and G1140A), blaCMY-12 (G1088A and G1140A), or blaCMY-2 (T511G and C661T). In the first two cases, the G1140A transition is silent.
PCR mapping using primers ISEcp1_Fw and blc-Rev (Table 1), designed based on the flanking regions of other blaCMY-2-like genes, yielded amplification products of about 1.7 kb from all isolates, suggesting that in all of them the blaCMY-16 gene was located between ISEcp1 and the blc gene, i.e., in the same genetic context reported for other genes of this lineage from plasmids of Salmonella enterica (10), Klebsiella spp. (33), and E. coli (9).
Functional characterization of CMY-16.The functional properties of CMY-16 were investigated by expression of the enzyme in E. coli DH5α and by testing susceptibility to several β-lactams. Analytical IEF of DH5α(pBC-CMY16) revealed the presence of a β-lactamase band with a pI of >8.4 (data not shown), confirming the expression of the CMY-16 enzyme in this strain. Compared to DH5α, DH5α(pBC-CMY16) exhibited decreased susceptibility to a broad array of β-lactams (Table 3). The impact was higher for penicillins, narrow-spectrum cephalosporins, cephamycins, oxyimino-cephalosporins (except cefepime), and aztreonam. However, some reduction of susceptibility was also observed for cefepime. Serine β-lactamase inhibitors showed a variable effect, with tazobactam being significantly more active than clavulanate at antagonizing the enzyme activity.
β-Lactam susceptibilities of E. coli DH5α(pBC-CMY16) producing the CMY-16 enzyme and DH5α(pBC-SK)
Clonal relatedness of the P. mirabilis isolates.Genotyping by RAPD fingerprinting yielded identical or almost identical profiles with all isolates (Fig. 1). PFGE profiles of genomic DNAs digested with SfiI were identical or clearly related (Fig. 2). Overall, these results indicated that the eight isolates were clonally related to each other.
RAPD fingerprinting of isolates investigated in this work, carried out with primer 1254 (A) or AP12h (B). Lanes 1, BG-073/03; lanes 2, NO-080/03; lanes 3, VA-1017/03; lanes 4, NO-051/03; lanes 5, PM207RED; lanes 6, PM208RED; lanes 7, PM209RED; lanes 8, PM210RED. DNA size standards are indicated in bp on the left.
PFGE profiles of genomic DNAs of the eight P. mirabilis isolates producing CMY-16 after digestion with SfiI. Lane 1, PM207RED; lane 2, PM208RED; lane 3, PM209RED; lane 4, PM210RED; lane 5, NO-051/03; lane 6, VA-1017/03; lane 7, NO-080/03; lane 8, BG-073/03. DNA size standards, expressed in kbp, are indicated on the left.
Clonal relatedness was not unexpected for isolates from the same ward of the same hospital (e.g., NO-051/03 and NO-080/03 or the four isolates from the LTCF), while it could be more surprising for isolates from different settings. However, the four cities are located at relatively small distances (65 to 90 km) from each other, and subsequent admissions of the same patient in different hospitals, which are not uncommon (especially for elderly patients attending LTCFs), might facilitate interhospital dissemination of resistant clones.
Transferability and genetic support of the blaCMY-16 gene.Transfer of β-lactam resistance to E. coli was not detected with any of the P. mirabilis isolates, either in conjugation assays with E. coli J53 as a recipient or following electroporation of total-DNA preparations into E. coli DH5α. Since ampicillin was used for the selection of transconjugants and transformants, neither the blaTEM-1b nor the blaCMY-16 determinant carried by the P. mirabilis isolates was apparently transferable to E. coli.
Agarose gel electrophoresis of total-DNA preparations of the eight P. mirabilis isolates revealed the presence of plasmid DNA bands in one isolate (BG-073/03) (data not shown), but a Southern blot analysis using a blaCMY-2 probe revealed a single hybridization signal corresponding to the band of chromosomal DNA with all isolates (Fig. 3 and data not shown). An identical hybridization pattern, resulting in a single band of approximately 15 kb with all isolates, was observed in a Southern blot analysis of total DNAs of the P. mirabilis isolates digested with XbaI, using the blaCMY-2 probe (Fig. 3).
Lanes 1 to 9, results of Southern blot hybridization of genomic DNAs of the eight P. mirabilis isolates producing CMY-16 digested with XbaI and hybridized to the blaCMY-2 probe. Lane 1, PM207RED; lane 2, PM208RED; lane 3, PM209RED; lane 4, PM210RED; lane 5, BG-073/03; lane 6, NO-080/03; lane 7, NO-051/03; lane 8, VA-1017/03. DNA size standards, expressed in kbp, are indicated on the left. The hybridization signal obtained with undigested genomic DNA of one isolate (VA-1017/03) and the corresponding agarose gel electrophoresis images are shown in lanes 9 and 10.
Southern blot analysis of genomic DNA restricted with I-CeuI and separated by PFGE, using either a 16S rRNA probe or a blaCMY-16 probe, showed a hybridization signal on the same band (of approximately 280 kb) in all isolates (Fig. 4).
PFGE profiles of genomic DNAs of the eight P. mirabilis isolates after digestion with I-CeuI (left) and Southern blot analysis of the same gel following hybridization with the blaCMY-16 probe (right). All bands visible in the gel and an additional band of smaller size (indicated by the asterisk) were recognized by the 16S rRNA gene probe. Lanes 1, PM207RED; lanes 2, PM208RED; lanes 3, PM209RED; lanes 4, PM210RED; lanes 5, NO-051/03; lanes 6, NO-080/03; lanes 7, BG-073/03; lanes 8, VA-1017/03. DNA size standards, expressed in kbp, are indicated on the left.
Altogether, these findings showed that the blaCMY-16 gene was inserted in the chromosome and that its genetic environment was conserved in all isolates, supporting the view that all the CMY-16-producing P. mirabilis isolates were derived from an original ancestor that had acquired blaCMY-16.
A chromosomal location was also previously reported for P. mirabilis isolates carrying blaCMY-3, blaCMY-4, blaCMY-12, blaCMY-14, and blaCMY-15 (5, 8, 14). A similar arrangement could be due to the fact that plasmids carrying these CBL genes and circulating among other Enterobacteriaceae are unable to replicate in P. mirabilis, and upon transfer to the species, this could result in recombination of β-lactamase genes in the chromosome of the new host.
Concluding remarks.Resistance of P. mirabilis to expanded-spectrum cephalosporins is an increasing problem in several epidemiological settings (8, 14, 21, 28). A similar resistance phenotype is usually mediated by the production of class A/group 2be ESBLs. In that case, resistance is usually reversible by clavulanate and other β-lactamase inhibitors, while susceptibility to cephamycins is not affected. The CBLs are overall less common than class A ESBLs as acquired resistance determinants, but emergence of these enzymes in P. mirabilis has been reported in several areas (3, 8, 12, 14, 23, 24, 34). To our knowledge, this is the first report of an acquired CBL in P. mirabilis from Italy.
As previously observed elsewhere (14), even in this case, the population of CBL-producing P. mirabilis isolates appeared to be monophyletic, likely reflecting the spread of a single clone after the event of acquisition of the β-lactamase determinant. It will be interesting to compare the Italian isolates with those from other countries to ascertain their potential clonal relationships.
CMY-16 is a new variant of the CMY/LAT lineage that might have evolved from either CMY-4 or CMY-12. Based on expression studies, its substrate profile and susceptibility to inhibitors seem to be overall similar to those of other members of the lineage, including CMY-2, CMY-4, CMY-12, CMY-14, and CMY-15 (14, 23). Comparative kinetic analyses of purified enzymes of this group to ascertain potential functional roles of the variable amino acid residues that characterize the different variants will be interesting.
ACKNOWLEDGMENTS
This work was supported by grants from the European Commission (LSHM-CT-2003-503335, COBRA Specific Targeted Research Project).
FOOTNOTES
- Received 21 July 2005.
- Returned for modification 7 September 2005.
- Accepted 17 November 2005.
- American Society for Microbiology
