A nosocomial outbreak of KPC-2-producing Klebsiella pneumoniae in a Chinese hospital: dissemination of ST11 and emergence of ST37, ST392 and ST395


Corresponding authors: D. Shen and L. Han, Department of Microbiology, Chinese PLA General Hospital, 28# Fuxing Street, 100853, Beijing, China and Centre for Hospital Infection Control, Chinese PLA Institute for Disease Control and Prevention, 20# Dongda Street, 100071, Beijing, China

E-mails: shendx301@yahoo.com; hanli@bmi.ac.cn


In China, Klebsiella pneumoniae carbapenemase (KPC) -producing K. pneumoniae isolates have been identified. However, little is known about the spread and outbreak of KPC-producing enterobacterial pathogens. In this study, 48 non-duplicated KPC-producing isolates were analysed for genetic relatedness by pulsed-field gel electrophoresis (PFGE), antimicrobial susceptibility by E-test, and sequence type (ST) by multilocus sequence typing. S1-PFGE and Southern blot were used for plasmid profiling, and PCR and subsequent sequencing were performed to determine the effects of genetic background on the blaKPC gene. From December 2011 to June 2012, an outbreak of the KPC-2-producing K. pneumoniae was observed. The 48 isolates of K. pneumoniae are categorized into eight PFGE types (A1, A2, A3, A4, B, C, D and E). The predominant pathogens of the outbreak were strains with PFGE types A1, A2 and A3, which all belong to ST11. Furthermore, ST37, ST392 and ST395 KPC-2-producing K. pneumoniae isolates have also been sporadically identified. The blaKPC-2-carrying plasmids vary in size from 30 to 220 kb. The genetic environments of the blaKPC-2 gene for most strains were consistent with the genetic structure of blaKPC-2 on the plasmid pKP048. In conclusion, the dissemination and outbreak of KPC-2-producing K. pneumoniae isolates in this study appeared to be clonal, and ST11 K. pneumoniae was the predominant clone attributed to the outbreak. This is the first study to report the emergence and spread of KPC-producing K. pneumoniae ST392 and ST395 worldwide. Our findings suggest that horizontal transfer of Tn3-based transposons might mediate the spread of blaKPC-2 gene between different K. pneumoniae clones in China.


The emergence and dissemination of carbapenem-resistant Enterobacteriaceae posed an important growing public health threat [1-5]. Among various determinants sustaining resistance to carbapenems, variants of Klebsiella pneumoniae carbapenemases (KPCs) have comprised the largest and most prevalent group of class A carbapenem-hydrolysing enzymes in Enterobacteriaceae. The blaKPC genes have been detected among various bacterial species, and KPC-producing pathogens have been detected worldwide after identification of the first KPC variant (KPC-2) in 1996 [1, 2]. In China, KPC-producing K. pneumoniae isolates have been identified [6-15]. An outbreak of blaKPC-carrying K. pneumoniae has been observed in the Zhejiang province, involving 28 patients in an intensive care unit (ICU) over a 6-month period [15]. Among these KPC-producing K. pneumoniae isolates, the single variant identified was blaKPC-2, whereas other blaKPC alleles were not detected. The blaKPC-2 gene is located on bacterial plasmids with size of about 40–180 kb, and coexist with other resistance determinants in clinical K. pneumoniae isolates from China, such as blaTEM, blaSHV, blaCTX-M, blaDHA, blaIMP, qnr, aac(6')-Ib-cr and rmtB [6, 7, 9-15].

As founders of clonal complex 258 (CC258), K. pneumoniae sequence type 11 (ST11) and its single-locus (tonB) variant, ST258, have spread globally and are significantly associated with the worldwide dissemination of KPC-producing K. pneumoniae [16]. ST258 is the predominant clone observed in European countries and the USA [16-18], while ST11 is the prevalent clone associated with the spread of KPC-producing K. pneumoniae in China [12]. Meanwhile, other blaKPC-carrying K. pneumoniae clones—including ST15, ST23, ST349, ST 351, ST438, ST439 and ST476—have been identified in China [9, 12, 13].

Genetic structures surrounding the blaKPC gene are mostly associated with two transposon-related structures. One is a Tn3-based transposon, Tn4401, which is about 10 kb in size and is composed of a transposase gene, a resolvase gene, the blaKPC gene and two insertion sequences (ISKpn6 and ISKpn7) [19]. Currently, five isoforms of Tn4401 have been identified which differ by various upstream deletions of the blaKPC gene [20]. Another genetic structure related to the blaKPC gene is also a Tn3-based transposon, which contains an integration structure of Tn3-based transposon and partial Tn4401 segment, with a substitution of ISKpn8 for ISKpn7 [8]. This structure was first identified on the plasmid pKP048 (GenBank No. FJ628167) from a clinical K. pneumoniae isolate. Several variants have been found among Enterobacteriaceae in China that contain various fragment insertions (most are truncated blaTEM) between the ISKpn8 and the blaKPC coding region [8, 12, 13, 21-23].

In this study, we report a nosocomial outbreak of KPC-2-producing K. pneumoniae isolates at our hospital, which involved 44 patients and covered three ICUs, and one surgical and medical ward. The molecular and epidemiological features of KPC-2-producing K. pneumoniae isolates and of the blaKPC-2 gene were also analysed.

Materials and Methods

Bacterial strains

Since 2009, all clinical K. pneumoniae strains isolated from a 4000-bed tertiary-care hospital were collected. All clinical isolates were identified using VITEK 2 GN ID cards (bioMérieux, Inc., Hazelwood, MO, USA) and VITEK® MS (bioMérieux SA, Marcy-l'Etoile, France). The isolates that exhibited non-susceptibility to carbapenems were screened for blaKPC gene by PCR amplification. Escherichia coli ATCC 25922 was used as the quality control strain for antimicrobial susceptibility testing. Salmonella ser. Braenderup strain (H9812) was used as a reference standard of pulsed-field gel electrophoresis (PFGE). The sodium azide-resistant E. coli J53 strain was used as the recipient for conjugation testing.

Antimicrobial susceptibility testing

The MICs of cefotaxime, cefepime, piperacillin–tazobactam, imipenem, meropenem, ertapenem, amikacin and levofloxacin were measured by E-test (AB bioMérieux, Solna, Sweden). All protocols were performed according to the manufacturer's instructions. The modified Hodge test was performed for suspected KPC production in K. pneumoniae isolates. All susceptibility results were interpreted according to the 2012 CLSI performance standards [24].

Molecular detection of resistance genes

β-lactamase genes including blaKPC, blaTEM, blaSHV, blaCTX-M, blaPER, blaVEB, blaGES and plasmid-mediated quinolone resistance genes, including qnrA, qnrB, qnrC, qnrD, qnrS, qepA, oqxAB and aac(6')-Ib-cr, and 16S rRNA methylase genes, including armA, rmtA, rmtB, rmtC and rmtD were screened by several multiplex PCR methods as previously described [25].

PFGE analysis

PFGE typing of KPC-producing isolates was performed as described by the US Centers for Disease Control and Prevention PulseNet program [26]. DNA fingerprints were obtained from PFGE profiles of genomic DNA digested with XbaI (New England Biolabs, Ipswich, MA, USA). The PFGE patterns were analysed by BioNumerics software (Applied Maths NV, Sint-Martens-Latem, Belgium) using the dice similarity coefficient. Strains were considered as the same clone (type) if they possessed ≥85% genetic similarity or fewer than four fragment differences in PFGE profiles [27]. Clusters were defined as DNA patterns sharing ≥70% similarity.

Plasmid analysis and Southern blot

A blaKPC probe was constructed by labelling a blaKPC PCR product through the PCR DIG Probe Synthesis Kit (Roche Applied Sciences, Mannheim, Germany). The S1-PFGE and Southern blot were performed as previously described [28].

Conjugation testing

Conjugation experiments were performed in Luria–Bertani broth. Cultures of donor and recipient cells in the logarithmic phase (0.5 mL of each) were added to 4 mL of fresh Luria–Bertani broth and incubated overnight without shaking. Transconjugants were selected on China blue agar plates containing meropenem (1 mg/L) and sodium azide (200 mg/L; Sigma-Aldrich Co., St Louis, MO, USA).

Multilocus sequence typing

Multilocus sequence typing was carried out for all PFGE types of K. pneumoniae isolates by amplifying and sequencing of seven housekeeping genes according to protocols provided on the multilocus sequence typing website for K. pneumoniae (http://www.pasteur.fr/recherche/genopole/PF8/mlst/Kpneumoniae.html). All amplification products were sequenced on a DNA analyser (3730xl; Applied Biosystems, Life Technologies, Foster City, CA, USA).

Genetic environment analysis of blaKPC gene

The genetic background of the blaKPC gene was investigated with a PCR mapping approach. Amplifications were performed using specific primers targeting the context of the blaKPC gene [8]. All amplification products were sequenced in a DNA analyser (3730xl; Applied Biosystems, Life Technologies).


Emergence and outbreak of KPC-2-producing strains

In this study, the strains that were isolated from different patients or different sites of the same patient were defined as non-duplicated strains. In total, 48 non-duplicated KPC-2-producing K. pneumoniae strains were isolated from 44 patients, which covered three ICUs, and one surgical and medical ward (Fig. 2). All of them were positive for Hodge test. Among these isolates, 25 (52%) were isolated from sputum samples, while two, three, three, seven and eight strains were isolated from bile, venous cannula, drainage fluid, urine sample and blood, respectively. The KPC-2-producing K. pneumoniae isolates were detected between February 2009 and November 2011. During this period, 1636 non-duplicated K. pneumoniae strains were isolated from various clinical specimens. In general, KPC-positive K. pneumoniae was presented at low prevalence (2.9%; 48/1636) at our hospital. An outbreak was observed in the next 7 months (December 2011 to June 2012) (Figs 1 and 2). A relatively higher prevalence of KPC-producing K. pneumoniae was observed during the outbreak, where 11.7% (41/351) of K. pneumoniae isolates were KPC-positive. According to the medical diagnosis on death, 12 of 44 (27%) patients with KPC-2-producing K. pneumoniae isolates died, and infection caused by KPC-2-producing K. pneumoniae strains continues to be a leading cause of death, including two urinary tract, one bloodstream, one biliary and eight pulmonary infections. Thirty patients (68%) showed improvement and two patients were discharged. For the eight patients with bloodstream infections caused by KPC-2-producing K. pneumoniae isolates, one died and the others showed improvement or were discharged.

Figure 1.

The distribution of the KPC-2-producing Klebsiella pneumoniae isolates.

Figure 2.

The distribution of the KPC-2-producing Klebsiella pneumoniae isolates with different pulsed field gel electrophoresis type among clinical wards. The date, frequency and location of emergence of different KPC-2-producing K. pneumoniae strains were shown in the Table. The head shows the date on which KPC-2-producing K. pneumoniae strains were isolated. The serial numbers of wards are listed in the first column. Different types of clinical isolates are indicated by the following symbols: ■:type A1; : type A2; : type A3; : type A4; ●: type B; 〇: type C; ⊙: type D; ◎: type E. The outbreak period is shown between the grey rows.

PFGE and multilocus sequence typing analysis

The 48 isolates of K. pneumoniae are categorized into five PFGE clusters (A to E). Cluster A contains four types (A1 to A4) while only one type was categorized among the other clusters. The distribution of KPC-2-producing isolates with different PFGE types is shown in Fig. 2. Cluster A comprises the majority of KPC-2-producing K. pneumoniae strains associated with the outbreak (variant types A1 to A4). Type A1 was the most prevalent strain that disseminated across all five wards, followed by types A2, A3 and D strain. Before November 2011, types A3 and D strains were scattered and identified in two wards, while a type E isolate was detected twice in ward II and has not emerged since. In a period of 7 months (December 2011 to June 2012), types A1 (23 isolates), A2 (five isolates) and A3 (eight isolates) strains were detected in an outbreak at two medical ICUs (ward I and II) and one surgical ward (ward III). Types A4 (one isolate), B (three isolates), C (one isolate) D (five isolates) and E (two isolates) strains were found in three wards (II, III and IV). Besides the type B strain, which was isolated from three specimens of the same patient, only type A1 was detected in wards IV and V (Fig. 2). Four sequence types (STs) were identified among K. pneumoniae isolates carrying blaKPC-2. Types A1, A2, A3, A4 and type C strains belonged to ST11, while type B was categorized with ST37. Type D and E were also identified as ST395 and ST392, respectively (Table 1).

Table 1. Phenotypic and genotypic characteristics of KPC-2-producing Klebsiella pneumoniae isolates
PFGE typeMLST typeNumber of isolateSize of plasmids carrying blaKPC-2Genetic environment of blaKPC-2Resistance determinantsMinimal inhibitory concentration (mg/L)
  1. MLST, multilocus sequence typing; CTX, cefotaxime; FEP, cefepime; TZP, piperacillin–tazobactam; IMP, imipenem; MEM, meropenem; ETP, ertapenem; AMK, amikacin; LEV, levofloxacin.

  2. a

    GenBank accession number: FJ628167 [5].

A1ST 1123~220 kbConsistent with pKP048ablaSHV, blaCTX-M-group 9, qnrS>256>256>256>32>32>320.75>32
A2ST 115~140 kbConsistent with pKP048ablaSHV, blaTEM, blaCTX-M-group 1, armA>256>256>256>32>32>32>256>32
A3ST 118~120 kbConsistent with pKP048ablaSHV, blaTEM, blaCTX-M-group 1, blaCTX-M-group 9, qnrS, rmtB>256>256>256>32>32>32>256>32
A4ST 111~120 kbConsistent with pKP048ablaSHV, qnrB>256>256>256>32>32>322>32
BST 373~40 kbVariant 1blaSHV, blaOXA-1-like, blaCTX-M-group 1, qnrB, aac(6')-Ib-cr, oqxAB>25632>2561283280.5
CST 111~130 kbConsistent with pKP048ablaSHV, blaTEM, blaOXA-1-like, qnrB, oqxAB, armA>256>256>256>32>32>32>256>32
DST 3955~50 kb and ~120 kbConsistent with pKP048ablaSHV, blaCTX-M-group 9, qnrB, aac(6')-Ib-cr, oqxAB>256>256>256>32>32>324>32
EST 3922~30 kbVariant 2blaSHV, blaTEM, blaCTX-M-group 1, aac(6')-Ib-cr, oqxAB>256>256>256>32>32>324>32

Antimicrobial susceptibilities

The MICs of common antimicrobial agents tested against all types of KPC-2-producing K. pneumoniae isolates are shown in Table 1. All isolates were resistant to oxyimino-cephalosporins (cefotaxime and cefepime) and carbapenems including imipenem, meropenem and ertapenem. Type A2, A3 and C isolates exhibited high-level amikacin resistance, while most isolates (except type B) presented resistance to levofloxacin (Table 1).

Prevalence of resistant determinants

Table 1 display the prevalence of common resistance determinants among each type of isolates. Type A1 strain did not produce any 16S rRNA methylase and was sensitive to amikacin, whereas type A2 and A3 strains were resistant to amikacin, but carried armA and rmtB, respectively. Furthermore, although most strains produced CTX-M, significant differences were exhibited in the prevalence of various resistant determinants. All isolates were negative for blaPER, blaVEB, blaGES, qnrA, qnrC, qnrD, rmtA, rmtC, rmtD and npmA.

Conjugation of plasmids

Only IR1301 (type A1) and IR131 (type B) transferred the blaKPC-2-carrying plasmids to the recipient by conjugation, and no transconjugants were collected for the other types of K. pneumoniae isolates. The MICs of the transconjugants of IR1301 and IR131 are as follow: cefotaxime (>256 and 12 mg/L), cefepime (6 and 2 mg/L), piperacillin–tazobactam (>256 and 128 mg/L), imipenem (4 and 4 mg/L), meropenem (1.5 and 1 mg/L), ertapenem (2 and 1.5 mg/L), amikacin (1 and 1 mg/L) and levofloxacin (0.023 and 0.023 mg/L).

Plasmid analysis

S1-PFGE and Southern blot analysis showed that all types of clinical K. pneumoniae isolates carried plasmids of different size. Furthermore, the blaKPC-2 gene was located on various plasmids, with size ranging from approximately 30 kb to 220 kb. Strains IR1304 (type A3), IR1215 (type A4) and IR1205 (type D) contained blaKPC-2-carrying plasmids with similar sizes (about 120 kb). Interestingly, besides the 120 kb-size of plasmid, another blaKPC-2-carrying plasmid with size of 50 kb coexisted in the strain IR1205.

Genetic environment of blaKPC gene

Most plasmids encode for structural components, which was consistent with the background of the blaKPC-2 gene in the plasmid pKP048 [8]. Inserted fragments of 321 and 259 bp in truncated blaTEM were detected upstream of the blaKPC-2 gene for strain IR1314 (type B) and IR1201 (type E), respectively. (Fig. 3 and Table 1).

Figure 3.

Schematic representation of the genetic structures surrounding the blaKPC-2 gene. Various genes and their directions of transcription are presented as broad arrows. The area delimited by two dotted lines indicates identical regions. * Pulsed field gel electrophoresis types.


KPC-producing K. pneumoniae isolates have been sporadically reported across China [6-14]. In this study, 2.9% of clinical K. pneumoniae isolates produced KPC-2, so KPC-producing K. pneumoniae exhibited relatively low prevalence in China compared with some European countries and the USA [3-5]. In this study, 12 of 44 (27%) patients died, with KPC-2-producing K. pneumoniae being a possible cause of death. However, most patients suffered from a serious underlying disease and presented with severe conditions. It is difficult to determine whether patients will succumb to infections associated with KPC-producing K. pneumoniae. Other studies have reported a higher mortality rate associated with bloodstream infections and no data are shown correlating patient survival with carbapenem dosage [29-31]. Therefore, it seems that mortality was more associated with the clinical conditions of the patients upon infection.

In this study, type A1 was the most prevalent strain that disseminated across all five wards. The multilocus sequence typing analysis demonstrated that the type A1 strain was ST11 (Table 1). Therefore, the dissemination of KPC-producing K. pneumoniae isolates in this study appeared to be clonal. This is consistent with a previous report showing that ST11 is the most prevalent clone in China [12]. Recently, KPC-producing K. pneumoniae ST11 was also identified in other parts of Asia, including Taiwan, Singapore and Korea, as well as European and American countries, including Greece, Brazil, the UK and Poland. ST11 belongs to the clone complex 258 (CC258) and possesses a single-locus variance (tonB) from ST258 [16]. As members of a complex, ST258 has been implicated in over 70% of all outbreaks of KPC-producing K. pneumoniae in the USA [18], whereas ST11 contributed significantly to the dissemination of the KPC-2-producing K. pneumoniae in Asia. Furthermore, ST11 was associated with the spread and outbreak of K. pneumoniae producing OXA-48 and NDM-1 [32, 33]. It has been speculated that CC258 might spread rapidly worldwide [12]. The K. pneumoniae CC258 also produces proteins involved in cell motility, secretion and DNA repair and modification, which might contribute to its rapid spread worldwide [34]. As for other KPC-producing K. pneumoniae STs, ST37 and ST392 did not spread in our hospital. ST395 (PFGE type D) was isolated from five patients and had a limited spread in ward II (Fig. 2). Recently, KPC-producing K. pneumoniae ST37 has been detected sporadically in other areas of the world [35, 36], and ST392 and ST395 have been reported to have contributed to the dissemination of blaOXA-48 in Europe [37]. To date, this is the first study to report about the emergence and spread of KPC-producing K. pneumoniae ST392 and ST395 worldwide.

Generally, KPCs alone do not confer resistance but are able to reduce the susceptibility to carbapenems, and impaired outer-membrane permeability contributes to the ineffectiveness of carbapenems for most KPCs [2]. In this study, most of the clinical strains exhibited a high-level of resistance to carbapenems (Table 1), whereas the two transconjugants exhibited low-level resistance to IPM and ETP, but were susceptible to MEM. These observations indicated that the carbapenem-resistant phenotype is attributed to the combination of KPCs and other mechanisms. In this study, the clinical strains exhibited different phenotypic and genotypic characterization (Table 1). Furthermore, the blaKPC-2-carrying plasmids vary dramatically in size among the different strains. The different plasmid profiles suggested that the spread of blaKPC-2 gene was not mediated by plasmid exchange between different strains. Altogether, the independent evolutionary processes and the clonal transmission of KPC-2-producing K. pneumoniae isolates may exist in China.

In this study, the blaKPC-2-carrying plasmids vary dramatically in size among the different strains. The different plasmid profiles suggested that the spread of the blaKPC-2 gene was not mediated by plasmid exchange between strains of different type, but by the clonal transmission of KPC-2-producing K. pneumoniae isolates in China. The genetic environment of the blaKPC gene has been characterized as a transposon-associated element, named Tn4401 [19]. Until recently, five isoforms of Tn4401, which differ by polymorphisms located upstream of the blaKPC gene, have been identified [20]. Meanwhile, a distinct genetic background of the blaKPC-2 gene located on the plasmid pKP048 has been identified in China, which contains an integration structure of a Tn3-based transposon and partial Tn4401 segment, with the gene order Tn3-transposase, Tn3-resolvase, ISKpn8, the blaKPC-2 gene and the ISKpn6-like element [8]. For most strains analysed in this study, the genetic environment of the blaKPC-2 is consistent with pKP048. Likewise, for most KPC-2-producing Enterobacteriaceae isolated from China, the genetic environments of the blaKPC-2 were also Tn3-based transposons similar to pKP048, with or without an inserted fragment of truncated blaTEM located upstream of the blaKPC gene [8, 12, 13, 21-23]. Although the region upstream of the blaKPC gene is variable, Tn4401 possesses genes that encode for transposase (tnpA) and resolvase (tnpR) activity, and have been characterized as an active transposon that is able to mobilize the blaKPC genes at high frequency without target specificity [38]. For K. pneumoniae and other common enterobacterial pathogens isolated from China, the Tn3-based transposons surrounding blaKPC-2 gene have accordant transposal structure compared with Tn4401. This suggests that the spread of the blaKPC-2 gene between different K. pneumoniae strains in China might be mediated by horizontal transfer of Tn3-based transposons but not by transmission of blaKPC-2-carrying plasmids.

The present study has several limitations. First, limited patient information was available for discriminating between infection with KPC-producing K. pneumoniae and colonization or transient carrier. Second, the outer-membrane permeabilities of KPC-producing K. pneumoniae isolates have not been analysed. Third, the type of blaKPC-carrying plasmids has not been characterized. Fourth, except for blaKPC, the types of other resistance genes were not identified.


The study was supported by the Programme of Development and Scientific Research of Beijing (No. 2009-1023).

Author Contributions

JY performed the molecular genetic analysis, plasmid profiling, multilocus sequence typing analysis and drafted the manuscript. LY, LG, QZ, RC and YL participated in the collection of strains and clinical details. YC, ST and JZ participated in sequencing and sequence analysis of PCR amplicons. DS and LH participated in the design and organization of the study. All authors read and approved the final manuscript.

Transparency Declaration

The authors have no conflicts of interest to declare.