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

Plasmid encoded ^bla CTX-M enzymes represent an important sub-group of class-A β-lactamases which hydrolyse broad-spectrum β-lactam antibiotics causing an extended spectrum β-lactamase (ESBL) phenotype, which is increasingly found in enterobacterial species including Klebsiella [1, 2]. To date, over 60 different CTX-M-type β-lactamases have been described [3] and divided into five different clusters that reflect similarity at the amino-acid sequence level, namely ^bla CTX-M-1, ^bla CTX-M-2, ^bla CTX-M-8, ^bla CTX-M-9, ^bla CTX-M-25 [2], respectively. More recently, it has been suggested that ^bla CTX-M-45 forms a new, separate cluster [4].

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.

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 [27‐31]. 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 β-lactamase 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].

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.