Worldwide Diversity of Klebsiella pneumoniae That Produce β-Lactamase blaKPC-2 Gene

TOC summary: Clones harboring different plasmids with identical genetic structure could be the origin of worldwide spread.

Outside the United States, KPC-producing K. pneumoniae are also being reported more often. The fi rst case of KPC-producing K. pneumoniae infection was reported in 2005 in France and had a US origin (11). The fi rst outbreak of KPC-producing K. pneumoniae outside the United States was in Israel (12). In South America, dissemination of KPC-producing K. pneumoniae was initially reported in 2006 in Colombia (13) and then in Brazil and Argentina (14,15). KPC enzymes have also been identifi ed in P. aeruginosa isolates from Colombia (16). In the People's Republic of China, KPC enzymes in several enterobacterial species are being increasingly reported (17). Finally, in Europe a few cases of KPC-producing K. pneumoniae infection have been described, but in Greece, outbreaks have occurred (18). In Europe, different variants of KPCs (KPC-2 and KPC-3) have been described; some patients carrying KPC-positive isolates had been transferred from the United States, Israel, or Greece (19)(20)(21).
Reports of this β-lactamase being found in novel locations are increasing worldwide, probably signaling active spread. The genetic element carrying the bla KPC-2 gene, Tn4401, was recently elucidated (22). Three isoforms of this Tn3-like transposon (a, b, and c) are known. Several other genetic environments of bla KPC gene have been described; other insertion sequences have been found upstream of the bla KPC gene (23,24). Nevertheless, the downstream sequences of the bla KPC gene matched perfectly with Tn4401, which suggests that these insertion sequences have been inserted into Tn4401.
Insertion sequences may play major roles in the evolution of Tn4401, but little information is available about the bacterial strains and the plasmids that may explain this rapid spread. Our goal, therefore, was to characterize the genetic background of several bla KPC-2 -harboring K. pneumoniae isolates from various geographic origins.

Bacterial Strains
K. pneumoniae isolates used in this study and their origin are listed in Table 1 (11,13,16,21,25). Electrocompetent Escherichia coli DH10B (Invitrogen, Eragny, France) was used as a recipient in electroporation experiments. E. coli J53Az R , which is resistant to sodium azide, was used for conjugation experiments. E. coli 50192 was used as a reference strain for plasmid extraction (22).

Antibiograms and MIC Determinations
Antibiograms were created by using the disk-diffusion method on Mueller-Hinton agar (Bio-Rad Laboratories, Marnes-La-Coquette, France), and susceptibility break points were determined as previously described and interpreted as recommended by the Clinical and Laboratory Standards Institute (22,26). All plates were incubated at 37°C for 18 h. MICs of β-lactams were determined by using the Etest technique (bioMérieux, Marcy l'Etoile, France).

Electroporation and Plasmid Extraction
Direct transfer of resistance into azide-resistant E. coli J53 was attempted as reported (22). Plasmids were introduced by electroporation into E. coli DH10B (22) by using a Gene Pulser II (Bio-Rad Laboratories).
Plasmid DNA was extracted by using a QIAGEN Plasmid Maxi Kit (QIAGEN, Courtaboeuf, France) and analyzed by agarose gel electrophoresis (Invitrogen, Paris, France). Natural plasmids were extracted by using the Kieser extraction method (27) and subsequently analyzed by electrophoresis on a 0.7% agarose gel.

Hybridization
DNA-DNA hybridization was performed as described by Sambrook et al. (28) with Southern transfer of an agarose gel containing Kieser method-extracted total DNA. The probe consisted of a 796-bp PCR-generated fragment from recombinant plasmid pRYC-1 (22) and was internal to the bla KPC-2 gene. Labeling of the probe and detection of signal were conducted by using an ECL nonradioactive labeling and detection kit according to the manufacturer's instructions (Amersham Biosciences, Orsay, France).

PCR Amplifi cation and Sequencing
The bla CTX-M -, bla SHV -, bla TEM -, and bla OXA-1/9 -like genes were searched for and characterized as described (21). PCR experiments were performed on an ABI 2700 thermocycler (Applied Biosystems, Les Ulis, France) by using laboratory-designed primers ( Table 2). PCR products were then analyzed on agarose gel and sequenced.
Both strands of the PCR products were sequenced by using laboratory-designed primers with an automated sequencer (ABI PRISM 3100; Applied Biosystems). The nucleotide and the deduced protein sequences were analyzed by using software from the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov).

Multilocus Sequence Typing
Multilocus sequence typing (MLST) with 7 housekeeping genes (rpoB, gapA, mdh, pgi, phoE, infB, and tonB) was performed according to Diancourt et al. (31). Allele sequences and STs were verifi ed at http://pubmlst.org/ kpneumoniae. A different allele number was given to each distinct sequence within a locus, and a distinct ST number was attributed to each distinct combination of alleles.

Replicon and Transposon Typing
PCR-based replicon typing of the main plasmid incompatibility groups reported for Enterobacteriaceae was performed as described (32). Genetic structures surrounding the bla KPC-2 gene were determined according to the Tn4401 PCR-mapping scheme as described (22).

Pulsotypes
Molecular typing by PFGE identifi ed 9 major pulsotypes among the isolates ( Table 3). The fi rst pulsotype (pulsotype A) corresponded to the strains from the United States and Greece. We found 4 different pulsotypes (B-E) among strains from Colombia, which suggested polyclonal diffusion inside this country. We also identifi ed 2 different clones among strains from Brazil (pulsotypes F and G) and from Israel (pulsotypes H and I). These results indicate much heterogeneity among KPC-producing isolates from various geographic regions.

Antimicrobial Drug Susceptibility
All isolates were resistant to penicillins and cephalosporins but showed varying levels of susceptibility to carbapenems (Table 4). Resistance to other drug classes varied among the isolates. For aminoglycosides, 2 clones (A and I) were susceptible to gentamicin only, 1 clone (H) was susceptible to amikacin only, and 3 clones (C, D, and G) were resistant to all tested aminoglycosides. Six clones (A, C, D, F, G, and I) showed resistance to fl uoroquinolones. Percentages of nonsusceptible isolates to the non-βlactam drugs were as follows: gentamicin, 75%; amikacin, 81.3%; ciprofl oxacin, 81.5%; trimethoprim/sulfamethoxazole, 81.5%; and tetracycline, 87.5%. Two isolates were also resistant to colistin (K. pneumoniae GR and K. pneumoniae K271); each was from Greece, where this drug is often used (33).
To investigate the fl anking sequences of Tn4401, we used PCR primers located in the Tn4401 structure and in the fl anking sequences derived from K. pneumoniae YC (22). PCR products of expected size were obtained for K. pneumoniae GR and K. pneumoniae K271 isolates only. For all other strains, no PCR product could be obtained, suggesting that the Tn4401 insertion site might differ from that found in K. pneumoniae YC.

Genetic Support for bla KPC in the Isolates
The carbapenem-resistant K. pneumoniae isolates contained several plasmids of different sizes, ranging from <5 kb to >170 kb ( Figure 2, left panel). At least 1 plasmid hybridized with an internal probe for bla KPC-2 gene in each isolate, ranging from 13 kb to 80 kb (Figure 2, right panel; Table 3). We observed 2 hybridization signals (35 kb and 75 kb) for K. pneumoniae KN2303, as described (22). Plasmid location of the bla KPC genes was confi rmed by electroporation of these plasmids into E. coli DH10B, but no transformant could be obtained for K. pneumoniae 2020532. The E. coli transformants had a β-lactam resistance pattern that corresponded to the expression of a bla KPC -like gene. Electroporation of 4 plasmids harboring the bla KPC -like gene into E. coli DH10B conferred resistance to at least an aminoglycoside molecule; pINC-H1521-6, pA33504, and p588 conferred resistance to all aminoglycosides except gentamicin, and electroporation of p475 into E. coli DH10B led to resistance to all aminoglycosides tested. No other antimicrobial drug resistance marker was cotransferred; the transformants remained susceptible to nalidixic acid, levofl oxacin, ciprofl oxacin, rifampin, tetracycline, trimethoprim/sulfamethoxazole, and colistin.
Mating-out assays showed that the ≈75-80-kb plasmids harboring bla KPC-2 from K. pneumoniae YC, GR, K271, and KN2303 were self-transferable to E. coli. The smaller plasmid from K. pneumoniae KN633 was not transferred to E. coli.

Origin of Replication
PCR-based replicon typing of the major plasmid incompatibility groups showed that the bla KPC-2 -positive plasmids belonged to at least 3 incompatibility groups (In-cFIIAS, IncN, and IncL/M) ( Table 3). The plasmids of K. pneumoniae KN633, HPTU-2020532 from Colombia and K. pneumoniae A33504 from Brazil gave negative results with the Inc primers tested and could not be classifi ed into a major plasmid incompatibility group.

Discussion
Rapid spread of KPC-producing K. pneumoniae is a major clinical and public health concern. These broadspectrum β-lactamases are increasing in new locations worldwide, indicating an ongoing process. Recently, a novel Tn3-based transposon, Tn4401, was identifi ed in nonclonally related KPC-producing K. pneumoniae and P. aeruginosa isolates (22). This transposon is in most recently described isolates (20,35,36), although a recently characterized novel variant from China had another insertion sequence inserted upstream of bla KPC gene (24). Identifi cation of Tn4401 inserted at different loci, on different plasmids, and fl anked by different 5-bp target site duplications indicates a frequent and dynamic process of transposition.  Table  2 are shown below, with results of PCRs for each isolate. B) PCR results with primers 7 and 8 ( Table 2) It has been suggested that this novel transposon is at the origin of bla KPC -like gene acquisition and dissemination (22). Sixteen K. pneumoniae isolates that express the bla KPC gene from 5 countries were characterized here.
PFGE and MLST showed that several clones are currently spreading in different geographic locations. In Colombia, 3 pulsotypes could be identifi ed. Overall, among the 16 isolates, 1 major ST (258) and its derivative ST 11 seemed to predominate. In a recent study that gathered isolates from 10 US states, ST 258 accounted for 70% of isolates, according to a database of KPC-producing K. pneumoniae PFGE results maintained by the Centers for Disease Control and Prevention (8). This ST has also been identifi ed for KPC-producing K. pneumoniae in Sweden (in isolates imported from Greece and Israel) and more recently in Poland (36,37). These fi ndings suggest possible international dissemination of KPC-producing ST 258. Apparently, the K. pneumoniae clone that contains the extended-spectrum β-lactamase (ESBL) determinant CTX-M-15 belongs to ST 11 (38).
KPC-producing K. pneumoniae contained diverse β-lactamases. All except 2 isolates harbored at least another β-lactamase; bla TEM-1 and a bla CTX-M -type ESBLs were expressed by >80% and 62.5% of isolates, respectively. KPC producers have already been associated with other β-lactamase genes, such as the widespread ESBL gene bla CTX-M-15 (17). SHV ESBLs have been found among isolates, as has been described for strains from the United States (39) and Norway (36). These additional β-lactamases are likely to complicate phenotype-based identifi cation of KPC producers. Three isolates harbored the chromosomeencoded bla OKP-A/B genes and belonged to phylogenetic group KpII, which accounts for <10% of K. pneumoniae strains (34). Coexpression of OKP enzymes and ESBLs has rarely been reported.
Isolates also demonstrated diversity in their molecular features. In this study, the KPC-2 genes were encoded on a broad variety of plasmids, as shown by previous studies (22,35). These plasmids differed in size and incompatibility groups. Similar plasmids were observed among isolates with the same ST, whereas different plasmids were also associated with similar STs. Therefore, epidemiologic investigation of KPC producers should be performed at different molecular levels.
Tn4401 was present in all tested strains. The overall structure of Tn4401 seemed to be conserved, except for the 100-bp to 200-bp deletion. Of the 16 isolates, 11 encoded the full-length Tn4401b isoform, 3 encoded the Tn4401a isoform containing a 100-bp deletion (ST 258), and 2 encoded the Tn4401c isoform containing a 200-bp deletion upstream of the bla KPC gene. These types of transposons tend to evolve by capturing various insertion sequences, as illustrated for the vanA-containing Tn1546 transposon (40). For Tn4401, three descriptions have been published in which different insertion sequences were present upstream of bla KPC-2 (22)(23)(24). None of these atypical structures were found in our strains. Observation of Tn4401 on different plasmids further supports the hypothesis that this transposon contributes to the mobilization and dissemination of the bla KPC genes. Our analysis of several K. pneumoniae isolates from 5 geographic origins indicates the spread of different clones that were harboring different plasmids but with an identical genetic structure, Tn4401, that sustained a bla KPC gene acquisition, which could likely be at the origin of the worldwide spread of this emerging resistance gene. Finally, taken together, our fi ndings and those of recent studies report a major KPC-producing clone with ST 258, even if novel ST types could also be evidenced, especially from Colombia. Our data suggest that KPC genes benefi t all molecular ingredients (transposon location, self-transferable plasmids, effi cient STs) by facilitating their rapid spread to K. pneumoniae and other bacterial species.