First Description of Klebsiella pneumoniae Harboring CTX-M β-Lactamases (CTX-M-14 and CTX-M-3) in Taiwan

Abstract

Klebsiella pneumoniae isolates from Taiwan medical centers (50 strains; 1998 to 2000) with a CTX-M resistance phenotype (ceftazidime susceptible and ceftriaxone or cefotaxime nonsusceptible) were selected for initial isoelectric focusing analysis. β-Lactamases with pIs of 7.9 (n = 22) and 8.4 (n = 28) in addition to 5.4 and/or 7.6 were detected. DNA gene sequencing identified the β-lactamases with pIs of 7.9 and 8.4 as CTX-M-14 and CTX-M-3, respectively. Molecular typing suggested inter- and intrahospital clonal dissemination of these Taiwanese CTX-M-producing Klebsiella strains.

A new series of extended-spectrum β-lactamase (ESBL) enzymes, cefotaximases (CTX-M), resulting in higher MICs of cefotaxime and ceftriaxone than of ceftazidime, has been discovered in several members of the family Enterobacteriaceae and in various countries (10). A previous study in southern Taiwan (11) showed that CTX-M-3 was the predominant ESBL among clinical isolates of Escherichia coli. We have also discovered that 87 of 211 (41%) ESBL-producing Klebsiella pneumoniae isolates from Taiwan expressed a CTX-M phenotype (ceftriaxone MIC, ≥16 μg/ml; ceftazidime MIC, ≤8 μg/ml [unpublished data]). CTX-M enzymes have not been previously identified in klebsiellae from Taiwan. In the present study, we determined the CTX-M identification by nucleotide sequencing, the distribution of MICs of various broad-spectrum β-lactams, and the molecular epidemiology of these CTX-M-producing strains.

A total of 211 clinical isolates of K. pneumoniae producing probable ESBLs (6) were recovered from 24 hospitals in Taiwan between January 1998 and June 2000. MIC results for selected antimicrobials were determined by broth microdilution according to the methods described by the NCCLS (5, 6). The ESBL phenotype was confirmed by using the Etest (AB Biodisk, Solna, Sweden) ESBL screen strips with a reduction of ≥3 log₂ dilutions for either cefotaxime, ceftazidime, or cefepime in the presence of clavulanic acid (4 μg/ml) as a positive test according to criteria defined by the NCCLS (6).

Crude β-lactamase extracts were prepared by freeze-thaw lysis of bacteria grown to exponential phase in tryptic soy broth (2). Analytical isoelectric focusing (IEF) was performed on Ampholine-polyacrylamide gels (pH range, 3.5 to 9.5; Amersham Pharmacia Biotech, Piscataway, N.J.) with a Multiphore II electrophoresis system (4). Following IEF, β-lactamase bands were visualized with 0.5 mg of nitrocefin (Becton Dickinson, Cockeysville, Md.)/ml. The isoelectric point (pI) standards included TEM-1 (pI 5.4), TEM-4 (pI 5.9), SHV-3 (pI 7.0), SHV-1 (pI 7.6), SHV-4 (pI 7.8), and SHV-5 (pI 8.2). Unknown β-lactamase pIs were calculated using a regression analysis (Microsoft Excel 98 software).

Automated ribotyping was performed using the RiboPrinter Microbial Characterization System (Qualicon, Inc., Wilmington, Del.) as described by Pfaller et al. (7). Isolates were considered to have the same ribotype if their patterns yielded a similarity coefficient of ≥0.93. Pulsed-field gel electrophoresis (PFGE) with the enzyme SpeI was regarded as confirmatory molecular analysis for the isolates within each ribotype. Isolates differing by more than three bands were considered unique. All bands had to match exactly for isolates to be classified as indistinguishable (same type). Isolates differing by one to three bands were designated as a similar subtype (see Table 2).

TABLE 2.

Molecular epidemiology of K. pneumoniae isolates with CTX-M-type enzyme^a

pI patternb (enzyme; no. tested) and ribotype	PFGE type(s)c	No. of isolates
7.9, 5.4 (CTX-M-14; n = 22)
595.7	A2, A3, A4, A5	11
73.2	B2, B3, B4	7
255.3	E1, E2, E3	3
8.4, 5.4 (CTX-M-3; n = 28)
595.7	A2, A6	4
75.8	XV	3
746.6	F1, F3	2
79.3	XII	2
78.7	XIV	2
87.8	XVI	2
73.8	XVII	2
1012.5	XVIII	2

^aResults from strains having a single occurrence are not shown (10 strains).

^bStrains may contain an additional enzyme with a pI of 7.6.

^cThe letter indicates the PFGE type with interhospital clonal spread. The number after the letter represents the PFGE subtype. The roman numerals represent PFGE types with intrahospital clonal spread.

Conjugation experiments were performed using E. coli J53-2 (rifampin resistant) as the recipient strain. Transconjugants were selected on Luria-Bertani medium containing 2 and 100 μg of ceftriaxone and rifampin/ml, respectively (8).

PCR analysis was performed on total DNA as prepared using the cetyltrimethylammonium bromide protocol described by Barnaud et al. (1). Amplification was achieved with consensus primers for the bla genes encoding CTX-M β-lactamases: primers (5′-GCTTTATGCGCAGACGAGTG-3′ and 5′-TCATTGGTGGTGCCGTAGTC-3′) comprised the positions 75 to 94 and 743 to 724, respectively, based on the bla_CTX-M-14 sequence (GenBank accession no. AF252622). The primers for CTX-M-3 (5′-TGTTGTTAGGAAGTGTGCCGC-3′ and 5′-TCGTTGGTGGTGCCATAGTC-3′) comprised the positions 56 to 76 and 743 to 724, respectively, based on the bla_CTX-M-3 sequence (GenBank accession no. Y10278). PCR fragments were isolated using Qiaquick PCR columns (Qiagen, Valencia, Calif.). DNA sequence analysis was performed using Big Dye terminator cycle sequencing chemistry with AmpliTaq polymerase FS enzyme (Applied Biosystems, Foster City, Calif.). The reactions were performed and analyzed with an Applied Biosystems model 373A stretch fluorescent automated sequencer (University of Iowa, DNA Core Facility).

Among the 87 isolates having a CTX-M phenotype, 50 were randomly selected for IEF analysis to represent all institutions and resistant phenotypes. In addition to enzymes with pIs of 5.4 and/or 7.6, 22 isolates had an enzyme with a pI of 7.9 and 28 had a band with a pI of 8.4. The β-lactamases with pIs of 5.4 and 7.6 have been reported elsewhere to be TEM-1 and SHV-1, respectively, and documented in Taiwan hospitals (3, 9, 12). The distributions of MICs of six selected β-lactams for the strains containing the β-lactamases with pIs of 7.9 or 8.4 are shown in Table 1. The two groups had similar MIC patterns, characterized by high MICs of ceftriaxone (or cefotaxime) and aztreonam associated with low MICs of cefoxitin, ceftazidime, and piperacillin-tazobactam. The cefepime MICs were variable (4 to >16 μg/ml for isolates with a pI of 7.9 and 2 to >16 μg/ml for those with a pI of 8.4).

TABLE 1.

Distribution of results for MICs of six selected β-lactams tested against K. pneumoniae strains with CTX-M-type enzymes (50 strains from Taiwan)

Enzyme (pI; no. tested) and antimicrobial	No. of strains for which drug had MIC (μg/ml):
	≤0.5	1	2	4	8	16	>16	32	>32
CTX-M-14 (pI 7.9; n = 22)
Cefoxitin	0	0	6	12	2	0	—a	0	2
Ceftriaxone	0	0	0	0	0	0	—	1	21
Ceftazidime	0	1	13	6	2	0	0	—	—
Aztreonam	0	0	0	0	3	8	11	—	—
Cefepime	0	0	0	1	5	5	11	—	—
Piperacillin-tazobactam	0	0	5	5	9	2	—	0	1
CTX-M-3 (pI 8.4; n = 28)
Cefoxitin	0	3	11	11	2	0	—	0	1
Ceftriaxone	0	0	0	0	0	1	—	2	25
Ceftazidime	4	9	6	6	3	0	0	—	—
Aztreonam	0	0	0	1	8	8	11	—	—
Cefepime	0	0	3	5	13	1	6	—	—
Piperacillin-tazobactam	0	2	11	11	1	1	—	2	0

^a—, not included in the scale of microdilution tray panel.

Molecular epidemiology using ribotyping and PFGE revealed evidence of clonal spread involving 40 isolates of CTX-M-producing K. pneumoniae (Table 2). Different pI patterns could be observed within the same epidemic genotype (ribotype 595.7/PFGE-A2), indicating the possible presence of diverse plasmids among isolates within the same clone. Although Yan et al. (11) reported the isolation of an E. coli strain with production of CTX-M-3 (pI 8.4) in Taiwan, they did not offer evidence of clonal spread. Our present study demonstrated the inter- and intrahospital clonal dissemination of K. pneumoniae isolates producing CTX-M enzymes in Taiwan.

Two representative strains, K. pneumoniae 94 (pI 8.4, ribotype 87.8) and K. pneumoniae 103 (pI 7.9, ribotype 595.7), isolated from a hospital in central Taiwan, were selected for further analysis, including conjugation experiments, PCR with CTX-M primers, and subsequent DNA gene sequencing. K. pneumoniae strain 94 was obtained from cerebrospinal fluid culture, and K. pneumoniae strain 103 was isolated from a blood culture. MIC results of various β-lactam compounds against the donors and transconjugants are presented in Table 3. The two donors had a common enzyme with a pI of 5.4, which was the putative TEM-1 β-lactamase (3, 9, 12). The β-lactamases with pIs of 5.4, 7.9, and 8.4 were cotransferred by conjugation into the recipient E. coli strain J53-2. The MIC of cefotaxime (or ceftriaxone) for K. pneumoniae strain 103 containing a β-lactamase with a pI of 7.9 was at least 128 times greater than that of ceftazidime (MIC, 4 μg/ml). The transconjugants E. coli p94 and E. coli p103 presented a resistance pattern similar to that of the donor isolate. Clavulanate restored the susceptibility of the strains to cefotaxime (MIC, 0.125 to 1 μg/ml) and cefepime (MIC, 0.047 to 0.19 μg/ml).

TABLE 3.

MICs of 12 β-lactams for donor strains (K. pneumoniae 94 and 103), transconjugants (E. coli p94 and p103), and the recipient strain (E. coli J53-2)

Antimicrobial agent(s)	MIC (μg/ml) for strain:
	K. pneumoniae 94 (pI 5.4, 8.4)	E. coli p94 (pI 5.4, 8.4)	K. pneumoniae 103 (pI 5.4, 7.9)	E. coli p103 (pI 5.4, 7.9)	E. coli J53-2 (recipient)
Cefoxitin	2	8	2	8	8
Aztreonam	>16	8	16	16	0.5
Ceftriaxone	>32	>32	>32	>32	0.12
Cefotaxime	16	32	>256	>256	0.12
Cefotaxime plus clavulanatea	0.12	0.25	1	1	0.25
Ceftazidime	2	4	2	4	0.5
Ceftazidime plus clavulanatea	1	2	0.5	1	1
Cefepime	8	8	32	16	0.094
Cefepime plus clavulanatea	0.06	0.12	0.25	0.12	0.064
Piperacillin plus tazobactama	4	2	4	2	2
Imipenem	0.25	0.12	0.12	0.25	0.25
Meropenem	0.03	0.03	0.03	0.03	0.03

^aClavulanate or tazobactam was tested at a fixed concentration of 4 μg/ml (Etest).

From total DNA of transconjugants as a template, partial gene PCR amplification with the universal CTX-M primers and direct DNA sequencing revealed that the β-lactamases with pIs of 8.4 and 7.9 matched the CTX-M-3 and CTX-M-14 enzymes, respectively (600-bp fragment) (data not shown). A further PCR assay using specific primers for CTX-M-3 and CTX-M-14 identified the pI 8.4 β-lactamase as CTX-M-3 (100% identity within a 708-bp fragment) and the pI 7.9 β-lactamase as CTX-M-14 (100% homology within a 609-bp fragment).

The descriptive data for CTX-M-14 remain unavailable, and the first isolation of the enzyme was reported elsewhere for E. coli isolated in Guangzhou, China (GenBank accession no. AF252622). Since Taiwan is geographically close to Guangzhou (separated by the Taiwan Strait) and the present report suggested the clonal spread of K. pneumoniae harboring CTX-M-14 in Taiwan, these results suggest that CTX-M-14 may be more widely distributed in the Far East than previously thought.

SHV-5-type β-lactamases (SHV-5 or -12) have been previously regarded as the most prevalent type of ESBL among K. pneumoniae strains in Taiwan medical centers (3, 9, 12). This study identified another rapidly evolving group of CTX-M-family enzymes (CTX-M-3 and CTX-M-14) with clonal dissemination in Taiwan. The emergence of various ESBLs in K. pneumoniae strains highlights the importance of continuously monitoring resistance trends and enhancing the nosocomial infection control of this pathogen.