Epidemic Plasmid Carrying bla CTX-M-15 in Klebsiella penumoniae in China

Objective To investigate the local epidemiology of Klebsiella penumoniae carrying bla CTX-M-15 in southern China and to characterize the genetic environment of bla CTX-M-15. Methods PCR and DNA sequencing were used to detect and characterize the genetic contexts of bla CTX-M-15. The clonal relatedness of isolates carrying bla CTX-M-15 was determined by pulse-field gel electrophoresis. Conjugative plasmids carrying bla CTX-M-15 were obtained by mating and were further subject to restriction analysis and replicon typing. Results A total of 47CTX-M-15 ESBL-producing isolates of K. pneumoniae were collected from nine hospitals in China from October 2007 to October 2008. Isolates were clustered into various clonal groups. The local spread of bla CTX-M-15 was mainly mediated by one major conjugative plasmid as determined by S1-PFGE and restriction analysis. A 90-kb plasmid belonging to incompatible group FII was the major carrier of bla CTX-M-15 in K. pneumoniae. Except bla TEM-1, the resistance genes such as bla SHV, bla DHA-1, bla OXA-1, qnrB, qnrS, aac(3)-II, and aac(6′)-Ib were not found in the plasmid. In the comparing of conjugative gene sequence, it is 100% identical with the plasmid pKF3–94, which was found in K. pneumonia from Zhejiang province of china previously. Conclusions bla CTX-M-15 was prevalent in K. pneumonia of southern China. The dissemination of bla CTX-M-15 appeared to be due to the horizontal transfer of a 90-kb epidemic plasmid.


Introduction
CTX-M enzymes have emerged as the predominant type of extended-spectrum b-lactamases (ESBL) produced by clinical isolates of Enterobacteriaceae in the world [1]. At present, more than 90 CTX-M variants have been designated (http://www.lahey. org/Studies/other.asp), of which CTX-M-15 is the most prevalent variant globally [2]. The global spread of bla CTX-M-15 is largely due to Escherichia coli of sequence type (ST) 131 and IncFII plasmids [3,4]. In China, bla CTX-M-14 has been the dominant bla CTX-M variant, especially in E. coli [5,6]. A recent small-scale investigation shown that 64.7% (11/17) E. coli isolates and 27.3% (3/11) K. pneumoniae isolates carried bla CTX-M-14 [7]. Interestingly, 64.3% of the bla CTX-M-14 genes were mainly located on an IncF plasmid in the 14 bla CTX-M-14 -positive isolates, which indicate the IncF plasmids plays a main role in dissemination of bla CTX-M-14 among the E. coli isolates in China [7].
The resistant rate of ceftazidime has increased sharply in both E. coli and K. pneumoniae recently in China [8]. However, the data about bla CTX-M- 15 in Klebsiella pneumoniae is scarce, especially the clonal relatedness of the plasmids in K. pneumoniae isolates carrying bla CTX-M-15 in China. Owing to CTX-M-15 enzyme hydrolyse ceftazidime at a higher rate than cefotaxime, it was speculated that bla CTX-M-15 gene may be emerging and spreading in China. Our previous study has also shown that bla CTX-M-15 had emerged as the more common type of bla CTX-M genes in K. pneumoniae in Guangzhou during 2007 to 2008, and revealed that 28.3% and 26.1% ESBL-producing K. pneumoniae isolates carried bla CTX-M- 14 and bla CTX-M-15, respectively [9]. The mechanisms responsible for the high prevalence of bla CTX-M-15 in K. pneumoniae occurred in the period remained undetermined. Therefore, all of the 47 K. pneumoniae isolates carrying bla CTX-M-15 from our previous study were studied for clonal relatedness and were also subjected to plasmid analysis including replicon typing and restriction analysis. This study show that bla CTX-M-15 was prevalent in K. pneumoniae of southern China and the dissemination of bla CTX-M-15 appeared to be due to the spread of a 90-kb epidemic plasmid.

Clinical Isolates
Clinical K. pneumoniae isolates carrying bla CTX-M-15 identified in our pervious study [9] were collected from nine hospitals, as part of the MOH National Antimicrobial Resistant Investigation net (Mohnarin) Program from October 2007 to October 2008. Briefly, Nine hospitals are scattered geographically over the three cities in southern China, Hospital A, B, C, D, E are located in the city I, and hospital F, G, H in the city II, and hospital K in the city III. Hospital A, B, C are close to each other, and the nearest distance between two hospitals is about 1 km, the city II and city III are far from the city I, and the distance is about 100 km and 300 km, respectively. Sometimes, the same patients may walk through the hospitals in one city for treatment. All the 181 isolates K. pneumoniae producing ESBLs were screened by PCR with special primers for all known CTX-M type encoding gene [9]. Purified PCR products were directly sequenced from both ends or cloned in pMD18-T and then sequenced. The DNA sequences and deduced amino acid sequences were compared to genes in GenBank or the blactamases classification system (www.lahey.org/studies/webt. html) to conform the subtypes of b-lactamase genes. The species identification for all isolates was performed using the VITEK-2 (bioMérieux, Marcy l'Etoile, France) automated microbiological analyzing system.

Antimicrobial Susceptibility Testing
Antimicrobial susceptibility testing of isolates and their transconjugants carrying bla CTX-M-15 was performed using the microdilution method according to Clinical and Laboratory Standards Institute (CLSI) guidelines [10]. Antimicrobial agents tested were ampicillin, ampicillin-sulbactam, piperacillin, piperacillin-tazobactam, cefoperazone-sulbactam, cefoxitin, cefotaxime, ceftazidime, cefepime, imipenem, aztreonam, amikacin and ciprofloxacin. All of the agents were provided by the Chinese National Institute for the Control of Pharmaceutical and Biological Products, except for cefoperazone-sulbactam, which was obtained from Pfizer (New York, NY). E. coli ATCC 25922 was used as the control strain. Isolates were classified as susceptible or resistant according to the interpretative standards recommended by CLSI.

Clonal Relatedness of Isolates Carrying bal CTX-M-15
Pulsed-field gel electrophoresis (PFGE) analysis of XbaI-digested genomic DNA was performed to determine the genetic relatedness of CTX-M-producing K. pneumoniae isolates using a CHEF-Mapper XA System (Bio-Rad Laboratories, Hercules, CA, USA) as described by Seifert [11]. The interpreting criteria were described by Tenover [12] combining UPGMA (unweighted pair group method with hierarchic averages) method, Isolates were assigned the same pulsetype if the value of Dice coefficient of similarity was .80% [13].

Conjugation
Conjugation experiments were performed in broth as described previously [14]. K. pneumoniae strains of different clones carrying bla CTX-M-15 were used as donor strains, while a rifampicin-resistant variant of E. coli strain C600, C600 (Rif r ), was used as the recipient. Transconjugants were selected on MH agar (Oxoid, Basingstoke, UK) supplemented with ceftazidime (2 mg ml 21 ) plus rifampicin (300 mg ml 21 ).

Plasmid Analysis
Plasmids from transconjugants were prepared using a modified alkaline lysis method [15]. Plasmid DNA was linearized with the S1 enzyme followed by PFGE [16]. The estimation of the plasmid size was compared with the molecular weight marker, Salmonella braenderup H9812. Plasmids were restricted with EcoRI. Plasmid replicons were determined using the PCR-based replicon typing scheme(PBRT) by using eighteen pairs of primers to perform PCR which recognized F,FIA, FIB, FIC, HI1, HI2, I1-Ic, L/M, N, P, W, T, A/C, K, B/O, X, Y and FII replicons as described by Carattoli [17]. All the primers and targets were listed in Table S2. To further determine the backbone of the plasmid, the genes traF, traH, traN, traU, traW and trbC, specific to the backbone of F system and ought to demonstrate the relationship of plasmid [18], were employed to analyze the backbone of the 90-kb plasmid by PCR and sequenced. (Primers listed in Table 1).

Antimicrobial Susceptibility Testing
A total of 47 out of 125 bla CTX-M postive strains of K. pneumoniae carrying bla CTX-M-15 , were originated from sputum (n = 38), blood (n = 6), abscess (n = 3). Results of the microdilution method showed that both types of ESBL-producing strains had low resistance to imipenem, piperacillin/Tazobactam and cefoperazone/sulbactam, with the resistance rate of 2.1%, 10.6% and 14.9%, respectively. The third-generation cephalosporin and cefepime had an antimicrobial resistance rate ranging from 57.4% to 89.5% (Table S1).

Clonal Relatedness of Isolates Carrying bla CTX-M-15
According to the patterns of PFGE isolates, 30 pulsotypes were designated for 47 bla CTX-M-15-producing K. pneumoniae isolates (Fig.  S4). Besides 25 individual pulsotypes, the remaining 22 isolates collected from three hospitals were classified into 5 pulsotypes, designated types A, B, C, D and E (Fig. 1). Five isolates from hospital A belonged to type A, 3 isolates from hospital A and 3 isolates from hospital B belonged to type B, 4 isolates from hospital B belonged to type C, 2 isolates from hospital B and 2 isolates from hospital C belonged to type D, 3 isolates from hospital C belonged to type E. In summary, the small clonal dissemination just occurred in each hospital or near hospital. There is no a predominant clone carrying bla CTX-M-15 found in this study.

Antibiotic Resistance Profile of 30 K. pneumoniae Isolates and the Transconjugants
For isolates of the same pulsotype, only one was chosen as a representative for further study. Therefore, a total of 30 non-clonal isolates carrying bla CTX-M-15 were subject to plasmid analyzing. The conjugative plasmids were obtained in all the 30 isolates, and 21 transconjugants containing a single 90-kb conjugative plasmid carrying bla CTX-M-15, while 9 transconjugants containing three different size of plasmid (60-kb,90-kb,160-kb) (Fig. S1). The susceptibility characteristics of all transconjugants are shown in Table 2. All the transconjugants were susceptible to imipenem and ciprofloxacin, while almost of them were resistant to ampicillin, piperacillin, aztreonam and cefotaxime. For these above antimicrobial agents, transconjugants shared similar MICs with their corresponding parent strains. In addition, the MIC of ceftazidime for transconjugants, were ranged from 4 to 64 mg/l, with 2-8 folds decreased than the parent strains, respectively. We also observed that 30.0% (9/30) isolates and 83.3% (25/30) transcon-jugants were susceptible to ceftazidime, according to CLSI breakpoint.

Plasmids Carrying bla CTX-M-15
To determine the role of plasmid in dissemination of bla CTX-M-15, the single 90-kb conjugative plasmids were subject to analyzed. Firstly, all the 90-kb plasmids were confirmed to be the IncFII group by PCR-based replicon typing (PBRT). Restriction analysis revealed that the twenty-one 90-kb plasmids restricted with EcoRI had highly similar patterns ( Figure 2). The sequences of traF, traH, traN, traU, traW and trbC in all the 90-kb plasmids were 100% identical to the plasmid pKF3-94 which is an epidemic plasmid carried bla CTX-M-15 in Zhejiang province of China [18]. Meanwhile, these sequences were found to have low identities to those of plasmid pC15-1a though both were in similar size. Besides, we also found that the length of EcoRI-EcoRI fragment of pKF3-94 mated with some restriction fingerprints of the epidemic plasmid theoretically, as shown in Table S3, there are 13 restriction fragments digested by EcoRI in pKF3-94 with the similar length of fragments of the restricted plasmids.
To demonstrate that the bla CTX-M-15 is located on the 90-kb plasmid of the remaining 9 transconjugants which contains multiplasmid (60-kb, 90-kb, 160-kb) (Fig. S2), the 90-kb plasmid DNA was acquired by plasmid electrophoresis and gel extraction, and then it was used as a template, bla CTX-M-15 was amplified by PCR and sequenced. Together, these results indicated that all the 90-kb plasmid harbored bla CTX-M-15 .   [19][20]. However, PCR amplification with primers specific for the tnpA genes of IS26    was negative for all 47 CTX-M-15 groups. In addition, bla TEM-1 was detected in all the 90-kb plasmid of transconjugants by sequenced. Other resistance genes such as bla SHV , bla DHA-1 , bla OXA-1 , qnrB, qnrS, aac(3)-II, and aac(69)-Ib were not found in the 90-kb plasmids consistent with the MIC data.

Discussion
Here, we show that K. pneumoniae carrying bla CTX-M-15 have emerged in southern China. The local spread of bla CTX-M-15 in K. pneumoniae was not mainly due to one or several dominant clones as various pulsotypes were identified in this study, though there were some medical tourism might occurred among the nine hospitals in the study period, unlike the situation seen in E. coli ST131 [21].
In contrast to the diversity of clonal relationships, many local isolates harbored a 90-kb IncFII plasmid carrying bla CTX-M-15 , suggesting that this plasmid appeared to be a major vehicle mediating the local dissemination of bla CTX-M-15 in K. pneumoniae. Indeed, previous reports [1] suggested that plasmid is one major factor responsible for the worldwide spread of bla CTX-M-15 . For example, in E. coli, many plasmids carrying bla CTX-M-15 found in France, Tunis, Bangui and India [22][23], shared common features with pC15-1a from Canada [3]. Furthermore, emergence of K. pneumoniae isolates producing CTX-M-15 were also found in European countries and bla CTX-M-15 transfer were mediated by IncFII-related plasmids with different sizes among part of them [24]. In this study, non-clonal isolates of K. pneumoniae from different hospitals had the same plasmid carrying bla CTX-M-15 . To our best knowledge, this is the first evidence of an epidemic plasmid carrying bla CTX-M-15 in K. pneumoniae.
IncFII-related plasmids are narrow-host range plasmids that are frequently involved in the worldwide dissemination of the bla CTX-M-15 gene [1]. However, unlike plasmids belonging to other incompatibilty groups, IncFII-related plasmids possess a great versatility of intracellular adaptation by the rapid evolution of the regulatory sequences of the replicons. Therefore, the backbones of IncFII-related plasmids exhibit a significant heterogeneity in terms of the size and number of replicons [25]. A genetic comparison of two widely distributed bla CTX-M-15 -carrying IncFII-related plasmids pC15-1a and pEK516 revealed three genetic events potentially accounted for all of the differences [26] though a 60kb region is higher homologous between the two plasmid which is originated from the non-R-determinant region of plasmid R100 [4]. For the 90-kb IncFII-related plasmid identified in this study, it appeared to have genetic components in both the backbone and the resistance region with different origins from those of plasmid R100. In contrast, this 90-kb plasmid appears to be closer to pKF3-94 as both plasmids originated from K. pneumoniae, belonged to the IncFII group, carried bla TEM and bla CTX-M-15 , and had the same traF, traH, traN, traU, traW and trbC sequences. The close relatedness between the 90-kb plasmid and pKF3-94 was also evidenced by their almost identical EcoRI restriction patterns, suggesting that the two plasmids may contain a common backbone. Interestingly, this 90-kb plasmid was not detected to harbor certain resistance genes such as bla OXA-1 , qnrB, qnrS, and aac(69)-Ib, which were usually found in other IncFII-related plasmid carrying bla CTX-M-15. The phenomenon is also found in the resistance region of pKF3-94 as only the ISEcp1-bla CTX-M-15 genetic stucture (positions 88110-88985) and bla TEM-1 (positions 89767-90627) but no IS26 found on pKF3-94.
IncFII-related plasmids are complex and sophisticated vehicles mediating the dissemination of antimicrobial resistance genes.
IncFII-related plasmids may co-exist with other Inc type plasmids and even are able to co-exist each other due to the differences on the inc sequence, which controls incompatibility in the same hosts [27]. The co-existence of plasmids were evidenced that three different sizes of plasmids existed in 9 out of 30 transconjugants containing 90-kb IncFII plasmid. The extensive recombination between different plasmids or plasmid fusions could facilitate IncFII-related plasmids to acquire various resistance genes and therefore generate new vehicles encoding multiple resistances. It could be predicted that the local 90-kb IncFII-related plasmid or pKF3-94 plasmid may evolve into plasmids with new phenotypes by acquiring multiple resistant genes during their further spread.
In summary, this study reveals that a high prevalence of bla CTX-M-15 genes in K. pneumoniae was contributed to a 90-kb IncFII-type plasmid which has different backbone structure with the epidemic plasmid found in Europe and other countries. It looks like that the epidemic IncFII plasmid carrying bla CTX-M-15 is not restricted in southern China but might have been spread the whole country.