Molecular identification of CTX-M and blaOXY/K1 β-lactamase genes in Enterobacteriaceae by sequencing of universal M13-sequence tagged PCR-amplicons

Background Plasmid encoded blaCTX-M enzymes represent an important sub-group of class A β-lactamases causing the ESBL phenotype which is increasingly found in Enterobacteriaceae including Klebsiella spp. Molecular typing of clinical ESBL-isolates has become more and more important for prevention of the dissemination of ESBL-producers among nosocomial environment. Methods Multiple displacement amplified DNA derived from 20 K. pneumoniae and 34 K. oxytoca clinical isolates with an ESBL-phenotype was used in a universal CTX-M PCR amplification assay. Identification and differentiation of blaCTX-M and blaOXY/K1 sequences was obtained by DNA sequencing of M13-sequence-tagged CTX-M PCR-amplicons using a M13-specific sequencing primer. Results Nine out of 20 K. pneumoniae clinical isolates had a blaCTX-M genotype. Interestingly, we found that the universal degenerated primers also amplified the chromosomally located K1-gene in all 34 K. oxytoca clinical isolates. Molecular identification and differentiation between blaCTX-M and blaOXY/K1-genes could only been achieved by sequencing of the PCR-amplicons. In silico analysis revealed that the universal degenerated CTX-M primer-pair used here might also amplify the chromosomally located blaOXY and K1-genes in Klebsiella spp. and K1-like genes in other Enterobacteriaceae. Conclusion The PCR-based molecular typing method described here enables a rapid and reliable molecular identification of blaCTX-M, and blaOXY/K1-genes. The principles used in this study could also be applied to any situation in which antimicrobial resistance genes would need to be sequenced.

Due to constitutive expression of a chromosomal class A β-lactamases [5], Klebsiella oxytoca was shown to have a high level resistance to ceftriaxone and cefotaxime [6]. Originally, this class A β-lactamase was named K1 [7] and later on referred to as KOXY [8] or bla OXY [6]. Sequence diversity of the chromosomally located K. oxytoca K1-gene and the existence of discrete groups of bla OXY-1 and bla OXY-2 genes has been described in detail [5,9].
Numerous PCR-based typing assays for the identification of bla CTX-M genes have been developed. Initially, detection of all members belonging to specific bla CTX-M groups was achieved by combining multiple PCR amplification assays [10,11]. To avoid multiple CTX-M PCR amplification steps, Boyd and co-workers [12] designed a pair of universal, degenerated CTX-M primers, allowing the amplification of hitherto all known bla CTX-M genes. However, identification of a bla CTX-M genotype at the nucleotide level often required cloning of the PCR-amplicons, followed by DNA sequencing. These methods are labour intensive, time-consuming, expensive and moreover, require a battery of amplicon specific sequencing primers.
In this study, we report on the use of a simple, accurate and universal CTX-M PCR amplification and sequencing assay well suited for high-throughput analysis.

Screening of Klebsiella spp. for cephalosporin resistance during 2001-spring 2007
At the Department of Clinical Microbiology, University Hospital Linköping, susceptibility testing was performed on all Klebsiella pneumoniae and K. oxytoca clinical isolates. Cefadroxil was used for the screening of cephalosporin resistance which was followed up by testing of resistant isolates with cefotaxime and ceftazidime or direct testing with cefotaxime and ceftazidime [14]. A biochemical panel for identification and differentiation of Klebsiella spp. was used. Indole-negative Klebsiella spp. clinical isolates were identified as K. pneumoniae and indole-positive clinical isolates as K. oxytoca.

Phenotypic ESBL-screening
All cefotaxime and/or ceftazidime resistant clinical isolates were phenotypically screened by Etest using ceftazidime and cefotaxime with and without clavulanic acid (bioMerieux Sverige AB, Askim, Sweden). A reduction of MIC by ≥3 twofold dilutions of the cephalosporin in the presence of clavulanic acid, i. e. a MIC ratio of ≥8 or the presence of phantom-or deformation zones was consid-ered indicative of an ESBL-phenotype. Clinical isolates were stored in glycerol containing Nutrient-broth No 2 (Lab M, Bury, UK) at -70°C until analysis.

Bacterial type and reference strains
Control strain K. oxytoca 1980K1 was kindly provided by Dr. D. Livermore, Health Protection Agency, Antibiotic Resistance Monitoring and Reference Laboratory, London, UK. Type strains were purchased from the American Type Culture Collection [15] or the Culture Collection University of Gothenburg [16]; K. pneumoniae ATCC 700603, K. pneumoniae CCUG 54718, and K. oxytoca CCUG 15717 T .

Multiple displacement amplification of bacterial DNA
To perform multiple genotyping analysis of our growing collection of CTX-M suspected K. pneumoniae and K. oxytoca of clinical origin and omitting tedious bacterial culturing, sufficient amounts of bacterial DNA were produced by means of multiple displacement amplification of bacterial DNA [17]. For that purpose, bacteria from frozen cultures (1 μl) and from the reference strains were added to a GenomiPhi-DNA V2 amplification-kit cocktail as recommended by the manufacturer (GE Healthcare Bio-Sciences AB, Uppsala, Sweden). PCR amplification conditions were as follows: initial denaturation step at 95°C for 15 min; 30 cycles of denaturation at 95°C for 30 s; annealing at 55°C for 30 s; extension at 72°C for 2 min, and a final extension step at 72°C for 10 min. Subsequently, PCR-amplicons were separated electrophoretically on a precast 2% agarose E-gel (Invitrogen, Carlsbad, CA, USA).

DNA sequence analysis
DNA sequence analysis of M13-sequence tagged CTX-M PCR-amplicons was carried out using a M13 uni (-21) primer by a customer DNA sequencing service (Eurofins MWG Operon GmbH, Martinsried, Germany). Prior to DNA sequencing, PCR-amplicons were treated with ExoSAP-IT to inactivate excess of oligonucleotide primers, following the supplier's protocol (USB Europe GmbH, Staufen, Germany). The PCR products were then lyophilised and sequenced. Generated DNA sequences were aligned, edited and compared with bla CTX-M DNA and bla CTX-M-like DNA sequences using the CLC bioinformatics freeware v.3.2.3 [18]. bla CTX-M, bla OXY, K1, and K1-like DNA sequences were retrieved from the NCBI Entrez Nucleotide database [19].  Fig. 1), were aligned using ClustalW [20]. A dendrogram consisting of all partial DNA sequences used for primer comparison in figure 1 and DNA sequences derived from the clinical isolates was constructed using the CLC bioinformatics freeware v.3.2.3 [18] and UPGMA clustering (Fig. 2). DNA sequences were edited to comprise the relevant DNA sequences between the two universal degenerated CTX-M primers (Fig. 1).

Distribution of Klebsiella spp
The number of K. pneumoniae and K. oxytoca clinical isolates of each year was in the range 457 to 864 (K. pneumo-niae) and 270 to 455 (K. oxytoca), respectively, comprising 99% of the genus Klebsiella collected and analysed at the Department of Clinical Microbiology, University Hospital, Linköping, Sweden.

Screening of cephalosporin resistance and ESBLphenotype
The cefadroxil resistance rate was in the range of 1.5% to 3.2% for K. pneumoniae and 3.7% to 6.4% for K. oxytoca. Twenty K. pneumoniae and 34 K. oxytoca clinical isolates were selected from a positive ESBL-phenotypic screening during 2001 to spring 2007. All K. oxytoca isolates (n = 34) revealed an ESBL-phenotype by screening with Etest using cefotaxime with and without clavulanic acid. All isolates were negative in the corresponding test with ceftazidime. Similarly, a majority of K. pneumoniae isolates (16 of 20) revealed an ESBL-phenotype in both ESBL Etests. Two of the K. pneumoniae isolates revealed an ESBL-phenotype only with cefotaxime and two isolates only with ceftazidime, respectively.

In silico DNA sequence comparison
The finding that the universal degenerated CTX-M primer-pair amplified the chromosomally located K1enzyme in K. oxytoca prompted us to perform a DNA sequence alignment of the universal CTX-M primer-pair with Enterobacteriaceae bla CTX-M, bla OXY, K1, and K1-like genes retrieved from the Entrez Nucleotide database (Methods). As illustrated in figure 1, the universal degenerated CTX-M primers revealed a high degree of DNA sequence similarity between the target sequences present in the E. coli, S. Typhimurium, C. freundii and C. amalonaticus bla CTX-M type-gene; K. oxytoca bla CTX-M-3 and bla CTX-M-35 genes, K. oxytoca K1 and bla OXY-1 to bla OXY-6 genes; the chromosomally encoded C. sedlakii Sed-1 and C. amalonaticus CdiA showed a lower degree of sequence similarity compared to C. koseri CKO, P. vulgaris K1, P. vulgaris CumA, and P. penneri HugA genes, respectively. With the exception of C. koseri CKO, P. vulgaris K1, P. vulgaris CumA and P. penneri HugA DNA sequences, most of the nucleotide variations are observed at 5'-Y, R, K-3' positions in primer CTX-M-U1.SE and 5'-R, R, R, S, Y-3' positions (were R stands for purine, Y stands for pyrimidine, S stands for G or C, and K stands for G or T) in primer CTX-M-U2.AS (Fig. 1). The GC-rich 3'-ends of the primers are highly conserved within the corresponding bla CTX-M, bla OXY, and K1 target DNA sequences. This may explain why the universal degenerated CTX-M primer-pair amplified bla CTX-M and K1 sequences.

Discussion
The increased prevalence of Enterobacteriaceae that produce bla CTX-M enzymes makes new demands on clinical routine microbiology laboratories to perform bla CTX-M typing. Due to the growing number of bla CTX-M enzymes, the traditional iso-electrofocusing appears not to be the method of choice for establishing an enterobacterial bla CTX-M genotype any longer. Molecular techniques for identification and classification of bla CTX-M genes in clinical isolates on a large scale have been described. Recently, a multiplex CTX-M PCR (MP-PCR) amplification assay was described which allows differentiation between different bla CTX-M subtype groups [21]. However, using this particular MP-PCR assay, we often observed non-specific PCR amplification in K. oxytoca isolates. Subsequent cloning and DNA sequencing analysis revealed that the unspecific PCR-amplicons represented K1-enzyme gene sequences. Thus, misinterpretation of strains as active bla CTX-M producers based on false positive PCR amplification might cause false reporting of bla CTX-M genes in K. oxytoca.
A different approach has been used by Galas and co-workers [22]. These authors described the use of a CTX-M-consensus primer-pair to establish a bla CTX-M-genotype in Enterobacteriaceae, including K. oxytoca [22]. However, in silico DNA sequence analysis reveals that this consensus CTX-M primer-pair [23] also targets bla OXY/K1-genes such as K. oxytoca bla OXY-2 [Genbank:Y17714] at positions 348-367 (MA1 primer) and 872-391 (MA2 primer), and K1 gene [GenBank:AY077482] at positions 369-378 (MA1 Primer) and 893-911 (MA2 primer), respectively. Thus, based on PCR amplification alone, it seems to be far from clear whether these K. oxytoca isolates would have a bla CTX-M, or a bla OXY/K1-genotype, respectively. This question can only be settled by DNA-sequencing of the PCR amplicons.
The use of M13-sequence tagged PCR-amplicons in combination with M13-specific sequencing primers was originally described for sequencing of Staphylococcus aureus protein A (Spa-typing) PCR-amplicons [24]. Our results employing the same technique for sequencing of β-lactamase PCR amplicons convincingly demonstrate that that the use of a M13-sequence tagged CTX-M.U1.SE primer allowed for an unequivocal discrimination of bla CTX-M and K1-genes. Moreover, our results support our previous findings indicating that bla CTX-M and K1-enzyme genes might have some degree of sequence similarity [13]. Extended in silico analysis furthermore revealed a high degree of sequence similarities between Enterobacteriaceae bla CTX-M, bla OXY 1-6, and K1 related sequences (Fig. 1), respectively. Thus, based on PCR amplification alone using universal degenerated CTX-M primer-pairs, it would be difficult, not to say impossible to distinguish between an Enterobacteriaceae bla CTX-M or K1/OXY genotype if sequencing had not been performed.
Fournier and co-workers [8] have described the existence of two discrete groups of K. oxytoca bla OXY-1 and bla OXY-2 enzymes. Later on, it was shown that bla OXY-1 and bla OXY-2 genes are expressed in two genetic K. oxytoca groups, namely K. oxytoca strain SG266 and SG271, respectively [25]. So far, six groups of OXY β-lactamases have been identified and characterised in K. oxytoca [26]. In our study we have found that all K. oxytoca clinical isolates form a distinct K. oxytoca K1/OXY-2 cluster group distinguishable from the K1/OXY-1 and the K. pneumoniae CTX-M15/28 cluster group, respectively. Moreover, the phylogenetic tree that was established implies the existence of a chromosomally located β-lactamase super-gene family in Enterobacteriaceae (Fig. 2). This is in agreement with previous reports describing that chromosomally encoded class A β-lactamases found in Klebsiella species are highly conserved at the amino-acid level compared to class A β-lactamases found in other Enterobacteriaceae [27][28][29][30][31]. The Citrobacter spp. DNA sequences included in the phylogenetic tree form separate clades. This is in agreement with earlier reports showing that C. koseri CKO-1 and C. amalonaticus CdiA isolates carry highly divergent β-lactamase genes despite the fact that they show a highly similar biochemical profile and 16S rDNA sequence similarity [32]. Biochemical methods may not always be adequate to identify Klebsiella spp. and their phylogenetic groups in clinical microbiology laboratories because several species share similar biochemical profiles [33,34]. Therefore, molecular techniques as applied in the present study may help to accomplish bacterial genotyping at reasonable costs and time.
MIC-value determination of piperacillin/tazobactam in this study also distinguished between K. oxytoca with K1 βlactamase from CTX-M producing K. pneumoniae, showing higher MIC-values. Furthermore, the MIC-values for cefotaxime and ceftazidime for K. oxytoca with K1-genes was lower than for CTX-M producing K. pneumoniae. Similar results have been reported by Potz and co-workers [35].

Conclusion
The PCR-based molecular typing method described here enables a rapid and reliable identification of CTX-M and OXY/K1-genes. The principles used in the present study can be applied to any situation in where antimicrobial resistance genes are to be sequenced. This is desirable because only sequencing of full-length reading frames will allow for an unequivocal discrimination between various subtypes of antimicrobial resistance genes such as bla CTX-M, bla SHV and bla TEM gene-families, respectively. Moreover, the use of M13-sequence tagged primers in PCR amplification assays facilitates amplicon sequencing since only one single sequencing primer (M13) is required.