A Novel Complex Mutant β-Lactamase, TEM-68, Identified in a Klebsiella pneumoniae Isolate from an Outbreak of Extended-Spectrum β-Lactamase-Producing Klebsiellae

Abstract

Twenty-two Klebsiella pneumoniae and two K. oxytoca extended-spectrum β-lactamase (ESBL)-producing isolates were collected in 1996 from patients in two pediatric wards of the University Hospital in Wrocław, Poland. Molecular typing has revealed that the K. pneumoniae isolates represented four different epidemic strains. Three kinds of enzymes with ESBL activity (pI values of 5.7, 6.0, and 8.2) were identified. The pI 6.0 β-lactamases belonged to the TEM family, and sequencing of the bla_TEM genes amplified from representative isolates revealed that these enzymes were TEM-47, previously identified in K. pneumoniae isolates from pediatric hospitals in Łódź and Warsaw. One of the TEM-47-producing strains from Wrocław was very closely related to the isolates from the other cities, and this indicated countrywide spread of the epidemic strain. The pI 5.7 β-lactamase was produced by a single K. pneumoniae isolate for which, apart from oxyimino-β-lactams, the MICs of β-lactam–inhibitor combinations were also remarkably high. Sequencing revealed that this was a novel TEM β-lactamase variant, TEM-68, specified by the following combination of mutations: Gly238Ser, Glu240Lys, Thr265Met, and Arg275Leu. The new enzyme has most probably evolved from TEM-47 by acquiring the single substitution of Arg275, which before was identified only twice in enzymes with inhibitor resistance (IR) activity. TEM-68 was shown to be a novel complex mutant TEM β-lactamase (CMT-2) which combines strong ESBL activity with relatively weak IR activity and, when expressed in K. pneumoniae, is able to confer high-level resistance to a wide variety of β-lactams, including inhibitor combinations. This data confirms the role of the Arg275Leu mutation in determining IR activity and documents the first isolation of K. pneumoniae producing the complex mutant enzyme.

Isolated since the mid-1980s, extended-spectrum β-lactamases (ESBLs) are usually encoded by plasmid-located genes and confer resistance to penicillins, cephalosporins (except cephamycins), and monobactams. β-Lactamase inhibitors (clavulanic acid, sulbactam, and tazobactam) block the activity of ESBLs, and this often causes ESBL-producing organisms to appear susceptible to some β-lactam–inhibitor combinations (2, 24, 27). ESBLs are derivatives of broad-spectrum penicillinases, such as TEM-1 or -2 or SHV-1, and ESBL activity is determined or enhanced by mutations at several positions, i.e., 104, 164, 237, 238, and 240, within their amino acid sequences (21, 27). Extensive use of newer-generation cephalosporins has been a strong factor selecting for ESBL variants formed de novo in a given environment (8, 32, 34), promoting their further evolution (7, 15), spread in bacterial populations by means of plasmid transmission (20, 22, 31), and clonal dissemination of producer strains (16, 30, 36), including their exportation to other health care institutions (7, 14, 38).

Another group of β-lactamases demonstrating inhibitor resistance (IR) activity has been isolated since the beginning of 1990s (44, 45). These enzymes confer resistance to penicillins and their combinations with β-lactamase inhibitors (24, 27, 29). The majority of IR β-lactamases known to date are derivatives of TEM-1 and -2 penicillinases (IRT variants), and mutations at several amino acid positions of these, i.e., 69, 130, 244, 275, and 276, were revealed or postulated to play a role in determining IR activity (11, 21, 33). A combination of ESBL- and IR-specific mutations within a single β-lactamase results in the formation of a so-called complex mutant enzyme (40). Two natural variants of such β-lactamases, TEM-50/CMT-1 and SHV-10, have been studied to date and were found to express either both of the activities at a moderate level or only one of these (33, 40). Here we report a novel enzyme of this kind, TEM-68/CMT-2, presented in a context of the epidemiological study of ESBL-producing klebsiellae in a hospital in Wrocław, Poland.

MATERIALS AND METHODS

Bacterial strains.

Twenty-four klebsiella clinical isolates (22 of Klebsiella pneumoniae and 2 of Klebsiella oxytoca) were collected in 1996 from different patients in the neonatal ward and in the Pediatric Intensive Care Unit (PICU) of the University Hospital in Wrocław. The isolates were cultured from various specimen types, mostly from bronchial-exudate, urine, and blood samples. Clinical data concerning the isolates is presented in Table 1. Species identification (using the ID32E ATB test; bioMerieux, Charbonnieres-les-Bains, France) and preliminary susceptibility testing were performed in the hospital microbiology laboratory. All of the isolates were identified as putative ESBL producers by the double-disk test (19).

TABLE 1.

Clinical data, RAPD patterns, PFGE types, plasmid fingerprints, IEF of β-lactamases, and ESBLs identified in klebsiella isolates and E. coli transconjugants

Isolate	Date of isolation (mo/yr)	Ward	RAPD-7 pattern	RAPD-1283 pattern	PFGE type	Isoelectric point(s) of β-lactamase	ESBL identified by bioassay (CAZ and CTX)a	ESBL type identified by PCR	ESBL identified by sequencingb	Plasmid fingerprint
K. pneumoniae 3144	1/96	Neonatology	A	A	a	6.0c, 7.6	6.0	TEM	TEM-47	A1
K. pneumoniae 3145	1/96	Neonatology	A	A	a	6.0, 7.6	6.0	TEM		A1
K. pneumoniae 3146	2/96	Neonatology	A	A	a	6.0, 7.6	6.0	TEM		A1
K. pneumoniae 3147	4/96	Neonatology	A	A	a	6.0, 7.6	6.0	TEM		A1
K. pneumoniae 3148	4/96	Neonatology	A	A	a	6.0, 7.6	6.0	TEM		A1
K. pneumoniae 3149	5/96	Neonatology	A	A	a	6.0, 7.6	6.0	TEM		A1
K. pneumoniae 3151	6/96	Neonatology	A	A	a	5.7, 7.6	5.7	TEM	TEM-68	A4
K. pneumoniae 3152	9/96	Neonatology	A	A	a	6.0, 7.6	6.0	TEM		A1
K. pneumoniae 3153	10/96	Neonatology	A	A	a	6.0, 7.6	6.0	TEM		A1
K. pneumoniae 3155	12/96	Neonatology	B	B	b1	7.6, 8.2	8.2	SHV		B1
K. pneumoniae 3156	1/96	PICU	C	C1	c	7.6, 8.2	8.2	SHV		B2+d
K. pneumoniae 3157	1/96	PICU	B	B	b2	6.0, 7.6	6.0	TEM		A2
K. pneumoniae 3158	3/96	PICU	C	C2	c	7.6, 8.2	8.2	SHV		B2+
K. pneumoniae 3159	4/96	PICU	B	B	b2	6.0, 7.6	6.0	TEM	TEM-47	A2
K. pneumoniae 3160	4/96	PICU	B	B	b2	7.6, 8.2	8.2	SHV		B1+
K. pneumoniae 3161	5/96	PICU	B	B	b1	7.6, 8.2	8.2	SHV	SHV-5	B1+
K. pneumoniae 3162	6/96	PICU	D1	D	d	6.0, 7.6	6.0	TEM	TEM-47	A3
K. pneumoniae 3163	6/96	PICU	D2	D	d	6.0, 7.6	6.0	TEM		A3
K. pneumoniae 3164	7/96	PICU	D1	D	d	6.0, 7.6	6.0	TEM		A3
K. pneumoniae 3165	9/96	PICU	B	B	b1	7.6, 8.2	8.2	SHV		B1
K. pneumoniae 3166	10/96	PICU	B	B	b1	7.6, 8.2	8.2	SHV		B1+
K. pneumoniae 3167	11/96	PICU	B	B	b1	7.6, 8.2	8.2	SHV		B1+
K. oxytoca 3150	5/96	Neonatology	a	a1	A	8.2	8.2	SHV		B1+
K. oxytoca 3154	12/96	Neonatology	a	a2	A	8.2	8.2	SHV		B1+
K. pneumoniae L-267e	1/95	Łódź	B	B	b2	6.0, 7.6	6.0	TEM	TEM-47	A5
K. pneumoniae 1027f	3/96	Warsaw	B	B	b3	6.0, 7.6	6.0	TEM	TEM-47	A6
K. pneumoniae 1099f	4/96	Warsaw	B	B	b4	6.0, 7.6	6.0	TEM	TEM-47	A6
K. pneumoniae 1592f	5/96	Warsaw	B	B	b5	6.0, 7.6	6.0	TEM	TEM-47	A7

^aBoth ceftazidime (CAZ)- and cefotaxime (CTX)-hydrolyzing activities of β-lactamases were detected in the bioassay. 

^bEmpty fields represent isolates for which no sequencing was performed. 

^cUnderlined pI values refer to β-lactamases, which were also produced by transconjugants. 

^dA plus sign in fingerprint designations represents additional plasmids present in the plasmid profile. 

^eClinical data, β-lactamase content, ESBL identification, and typing in the context of other isolates have already been published (14, 15). 

^fClinical data, β-lactamase content, ESBL identification, and typing in the context of other isolates have already been published (14). 

The set of four previously characterized TEM-47 ESBL-producing K. pneumoniae clinical isolates was used for comparative typing. The L-267 strain was isolated in 1995 in the Polish Mother Memorial Hospital in Łódź (15), whereas strains 1027/96, 1099/96, and 1592/96 were recovered in 1996 in the University Children's Hospital in Warsaw (14).

Escherichia coli A15 R⁻ (resistant to rifampin) was used as the recipient strain in resistance transfer experiments. DNA cloning was performed with the use of E. coli DH5α as the host strain. E. coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853 were used as reference strains for antimicrobial susceptibility evaluation.

RAPD typing.

Genomic DNAs of the isolates were purified with the Genomic DNA Prep Plus kit (A & A Biotechnology, Gdańsk, Poland). Randomly amplified polymorphic DNA (RAPD) analysis was performed using the RAPD-7 and RAPD-1283 (35) oligonucleotides as primers. PCRs were run as described previously (15).

PFGE typing.

For pulsed-field gel electrophoresis (PFGE) typing, total DNA preparations embedded in 1% agarose plugs (InCert Agarose; FMC Bioproducts, Rockland, Maine) were digested with XbaI restrictase (MBI Fermentas, Vilnius, Lithuania) and separated in 1% agarose gel (Pulsed Field-Certified; Bio-Rad, Hercules, Calif.) using a CHEF DRII PFGE system (Bio-Rad). The procedure was performed as described by Struelens et al. (41), and results were interpreted in accordance with the criteria proposed by Tenover et al. (43).

Plasmid DNA fingerprinting.

Plasmid DNA was purified from bacterial cells by the alkaline lysis method (6) using the QIAGEN Plasmid Midi Kit (QIAGEN, Hilden, Germany) as previously described (4). For fingerprinting analysis, plasmid DNA was digested with the restriction enzyme PstI (MBI Fermentas) and electrophoresed in 1% agarose gels (Sigma Chemical Company, St. Louis, Mo.).

Antimicrobial susceptibility testing.

The MICs of various antibiotics were determined by the agar dilution method in accordance with National Committee for Clinical Laboratory Standards (NCCLS) guidelines (28). The antibiotics used were ampicillin, cefotaxime, and gentamicin from Polfa, Tarchomin, Poland; amikacin and aztreonam from Bristol-Myers Squibb, New Brunswick, N.J.; cefoxitin from Sigma Chemical Company; ceftazidime from Glaxo Wellcome, Stevenage, United Kingdom; lithium clavulanate from SmithKline Beecham Pharmaceuticals, Betchworth, United Kingdom; imipenem from Merck, Sharp & Dohme Research, Rahway, N.J.; and piperacillin and tazobactam from Wyeth Ayerst Laboratories and Lederle Laboratories, respectively, Pearl River, N.Y. In all β-lactam–inhibitor combinations, the constant concentrations of clavulanate and tazobactam were 2 and 4 μg/ml, respectively.

The MICs of piperacillin in combination with inhibitors at various concentrations were evaluated to characterize selected strains. The analysis was performed as described above. The concentrations of inhibitors used doubled from 0.0075 to 4 μg/ml in the case of clavulanate and from 0.5 to 4 μg/ml in the case of tazobactam.

Resistance transfer.

Ceftazidime resistance transfer was carried out as previously described (15). Transconjugants were selected on MacConkey agar (Oxoid, Basingstoke, United Kingdom) containing rifampin (128 μg/ml; Polfa) and ceftazidime (2 μg/ml).

IEF of β-lactamases.

Supernatants of bacterial sonicates (3) were subjected to isoelectric focusing (IEF) in accordance with the procedure by Matthew et al. (26) with modifications (3), using a model 111 Mini IEF Cell (Bio-Rad). Following IEF, β-lactamase bands were visualized by staining gels with nitrocefin (Oxoid).

Bioassays for detection of ESBL activity.

Following IEF, the ceftazidime- and cefotaxime-hydrolyzing activities were assigned to particular β-lactamase bands by the bioassay approach described by Bauernfeind et al. (3). The concentration of both cephalosporins used in the experiment was 2 μg/ml.

PCR detection of bla_TEM and bla_SHV genes.

Total DNA extracted from the isolates was used in specific PCRs for the detection of bla_TEM and bla_SHV genes. Primers TEM-A and TEM-B (25) were used for amplification of entire bla_TEM genes; primers SHV-A and SHV-C (15, 31) were used for partial amplification of bla_SHV genes. Primer SHV-C was designed to specifically amplify genes encoding SHV β-lactamases with the Gly238Ser and Glu240Lys substitutions (31). PCRs were run as described previously (15).

Sequencing of bla_TEM- and bla_SHV-specific PCR products.

PCR products containing the amplified bla_TEM and bla_SHV genes were purified with a QIAquick PCR Purification Kit (QIAGEN) and subjected to direct sequencing reactions (37) using an ABI PRISM 310 automatic sequencer (PE Biosystems, Foster City, Calif.). Primers TEM-A, TEM-B, TEM-C, TEM-D, and TEM-E (25) were used for sequencing of bla_TEM genes. The complete bla_SHV gene was amplified with primers SHV-D (5′-CTCAAGGATGTATTG-3′) and SHV-H (5′-TTAGCGTTGCCAGTGC-3′) and plasmid purified from a transconjugant strain as the template. Primers SHV-A (15), SHV-D, SHV-F (5′-TCTGGTGGACTACTC-3′), SHV-G (5′-GTTGTCGCCCATCTG-3′), and SHV-H were used for sequencing.

Cloning of bla_TEM-47 and bla_TEM-68 genes.

The bla_TEM-47 and bla_TEM-68 genes were amplified together with their promoter regions using primers TEM-A/EcoRI and TEM-B/BamHI. These primers are modified versions of primers TEM-A and TEM-B (25) with the respective restriction sites added on their 5′ ends. The resulting products were cut with EcoRI and BamHI and cloned into vector pGB2, which is a low-copy plasmid containing the spectinomycin-streptomycin resistance gene as a transformation marker (13). E. coli DH5α transformants were selected on tryptic soy agar (Oxoid) supplemented with streptomycin (Polfa) at 30 μg/ml and ceftazidime at 2 μg/ml. The resulting pGB2 derivatives containing the bla_TEM-47 and bla_TEM-68 genes were designated pGBT-47 and pGBT-68, respectively.

Determination of IC₅₀s.

Supernatants of sonicates (3) of the E. coli DH5α transformants producing TEM-47 and TEM-68 were used for comparative determination of the inhibitor concentrations that reduced β-lactamase activity by 50% (IC₅₀s). IC₅₀ evaluation was performed as described by Bush et al. (10). Aliquots of extracts containing about 60 μg of protein were used in reactions run in a volume of 575 μl at room temperature in a DU 640 spectrophotometer (Beckman Instruments, Fullerton, Calif.).

Nucleotide sequence accession number.

The nucleotide sequence of the bla_TEM-68 gene will appear in the EMBL database under accession number AJ239002.

RESULTS

Typing of ESBL-producing klebsiellae from the hospital in Wrocław.

Results of PFGE and RAPD typing are shown in Table 1. PFGE analysis distinguished four different types of K. pneumoniae isolates with the most prevalent PFGE types, a and b, grouping nine and eight isolates, respectively. The remaining PFGE types, c and d, represented two and three isolates, respectively. Only the type b isolates could be further classified into subtypes b1 and b2 by a single DNA band difference in their PFGE patterns. The two K. oxytoca isolates were found to be indistinguishable in the PFGE analysis. Results of RAPD typing were very similar, with isolates producing identical or nearly identical banding patterns forming clusters, which correlated well with the distribution of PFGE types.

Resistance transfer.

All of the clinical isolates were subjected to the ceftazidime-resistance transfer experiment. The majority produced transconjugants, with the exception of the K. pneumoniae isolates of PFGE subtype b2 and type c and the single isolate (3151/98) of PFGE type a.

β-Lactamase contents of clinical isolates and their transconjugants and identification of ESBLs.

Protein extracts of all of the clinical isolates and transconjugants were separated by IEF in order to reveal their β-lactamase content. Results of the analysis are shown in Table 1. Each isolate of K. pneumoniae was found to express a β-lactamase with a pI of 7.6 together with another enzyme with a pI of 5.7, 6.0, or 8.2. The pI 5.7 β-lactamase was produced by the single isolate, 3151/98, belonging to PFGE type a. The pI 6.0 enzymes were expressed by all of the remaining type a isolates, two isolates of PFGE subtype b2, and all isolates of type d. The pI 8.2 β-lactamases were predominant among isolates of PFGE type b (subtypes b1 and b2) and were also produced by the PFGE type c isolates and both isolates of K. oxytoca. Enzymes with a pI of 6.0 or 8.2 were identified in extracts of all of the corresponding transconjugants (Table 1). The bioassay experiment revealed that only the β-lactamases with a pI of 5.7, 6.0, or 8.2 were able to hydrolyze ceftazidime and cefotaxime under the conditions used (Table 1), and so these enzymes demonstrated ESBL activity.

Plasmid fingerprinting.

Results of plasmid fingerprinting analysis are presented in Table 1. All isolates producing the pI 5.7 or 6.0 ESBLs contained high-molecular-weight (high-MW) plasmids with similar PstI fingerprints, designated A1 to A4. Fingerprints A1, A2, and A3 characterized the pI 6.0 ESBLs producers, and their distribution correlated fully with the distribution of PFGE types a, b2, and d, respectively. The plasmid with fingerprint A4 was found in K. pneumoniae isolate 3151/98, which expressed the pI 5.7 enzyme and was smaller than other type A molecules. The PFGE type b (b1 and b2) K. pneumoniae and the K. oxytoca isolates producing the pI 8.2 ESBLs carried large plasmids of the same fingerprint, B1, which in several cases were copurified with molecules with lower MWs. The remaining two pI 8.2 ESBL-producing K. pneumoniae isolates of PFGE type c contained high-MW plasmids with a very similar fingerprint, B2.

PCR detection and sequencing of ESBL-encoding genes.

Specific PCRs were run in order to identify the ESBL types produced by the clinical isolates studied. Results of the analysis are presented in Table 1. Total DNA preparations of the isolates expressing the pI 5.7 or 6.0 ESBLs were tested for the presence of bla_TEM genes. Amplification products of the expected size of about 1 kb were obtained with primers TEM-A and TEM-B (25) for all of these isolates. Detection of bla_SHV genes encoding SHV β-lactamases that contain the Gly238Ser and Glu240Lys ESBL-specific substitutions was carried out on DNAs purified from isolates producing the pI 8.2 enzymes. PCR products of the expected size of about 220 bp were identified for all of the isolates in this group.

The bla_TEM PCR products obtained for pI 6.0 ESBL-expressing K. pneumoniae isolates 3144/98 (PFGE type a), 3159/98 (PFGE subtype b2), and 3162/98 (PFGE type d) and for the 3151/98 strain producing the pI 5.7 enzyme were selected for DNA sequencing. Sequences of the entire PCR products encompassing protein-coding frames, together with their 5′-adjacent regions, were determined and compared with the sequence of the bla_TEM1a gene (42). Amino acid sequences of the β-lactamases were deduced and compared with other enzymes of the TEM family (G. Jacoby and K. Bush, http: //www.lahey.org/studies/webt.htm). Results are shown in Tables 1 and 2. The sequences of PCR products specific for isolates expressing the pI 6.0 ESBLs were found to be identical to each other, and these contained genes encoding the TEM-47 β-lactamase identified previously in K. pneumoniae isolates from pediatric hospitals in Łódź (15) and Warsaw (14). The DNA sequence of the PCR product specific for the 3151/98 isolate producing the pI 5.7 enzyme was identical to the bla_TEM-47-containing amplicons, except for a single mutation, G1020→T, located within the coding region and specifying the additional amino acid substitution Arg275Leu. This novel sequence variant of a TEM β-lactamase was designated TEM-68. (Numbering of nucleotide positions is in accordance with that of Sutcliffe [42], and that of amino acid residues is in accordance with that of Ambler et al. [1].) A PCR product containing the entire bla_SHV gene was obtained from plasmid DNA purified from the E. coli transconjugant of K. pneumoniae isolate 3161/98. Sequencing has revealed that the product encompassed the bla_SHV-5 gene of the identical sequence with the one reported by Billot-Klein et al. (5).

TABLE 2.

Nucleotide sequence of the bla_TEM-68 coding region compared with those of the bla_TEM-1a, bla_TEM-1b, bla_TEM-2, bla_TEM-47, bla_TEM-48, and bla_TEM-49 genes

Nucleotidea (amino acidbc) position	Nucleotide (amino acid)d
	blaTEM-1ae	blaTEM-1bf	blaTEM-2f	blaTEM-48g	blaTEM-49g	blaTEM-47g	blaTEM-68
226	C	T				T	T
263 (21)	C (Leu)			T (Phe)	T (Phe)
317 (39)	C (Gln)		A (Lys)
346	A		G	G	G
436	C	T	T	T	T	T	T
512 (104)	G (Glu)
604	G	T
682	T		C	C	C	C	C
692 (164)	C (Arg)
914 (238)	G (Gly)			A (Ser)	A (Ser)	A (Ser)	A (Ser)
917 (240)	G (Glu)			A (Lys)	A (Lys)	A (Lys)	A (Lys)
925	G		A	A	A	A	A
990 (265)	C (Thr)			T (Met)	T (Met)	T (Met)	T (Met)
998 (268)	A (Ser)				G (Gly)
1020 (275)	G (Arg)						T (Leu)

^aNucleotide positions are numbered as described by Sutcliffe (42). 

^bAmino acid positions are numbered as described by Ambler et al. (1). 

^cAmino acid numbers are indicated only for positions at which substitutions are observed. 

^dOnly mutational changes with respect to the TEM-1 β-lactamase and/or bla_TEM-1a gene sequences are listed. 

^eReference 42. 

^fReferences 12 and 17. 

^gReference 15. 

Comparative typing and plasmid fingerprinting of TEM-47-producing K. pneumoniae isolates from different hospitals.

The group of representative K. pneumoniae TEM-47-producing isolates from the hospitals in Łódź (15) and Warsaw (14) were typed along with the TEM-47 producers from the Wrocław hospital. Results of PFGE and RAPD analyses are shown in Table 1. Wrocław isolates of PFGE subtype b2 were found to be indistinguishable by both approaches from the L-267 strain from Łódź, in which the TEM-47 enzyme was originally identified, and closely related to the set of TEM-47-expressing isolates from Warsaw. Plasmids purified from the Łódź and Warsaw isolates produced PstI fingerprints very similar to the type A molecules carrying the bla_TEM-47 genes in isolates from Wrocław (Table 1).

Antimicrobial susceptibility testing of clinical isolates and transconjugants strains.

Table 3 shows the antimicrobial susceptibility data presented with regard to the ESBL types and the PFGE types of the isolates. Increased MICs of the majority of β-lactams tested characterized all of the clinical isolates; however, K. pneumoniae isolates demonstrated a remarkably higher level of resistance than K. oxytoca isolates. MICs of ceftazidime and aztreonam were significantly higher than those of cefotaxime. For some of the K. pneumoniae isolates, the MICs of cefoxitin were elevated (16 to 32 μg/ml) and all of the isolates were fully susceptible to imipenem. For TEM-68-producing K. pneumoniae strain 3151/98, the MICs of inhibitor combinations (e.g., piperacillin-tazobactam, >512 μg/ml; ceftazidime-clavulanate, 128 μg/ml) were very high; in all other cases, β-lactamase inhibitors efficiently restored the activity of β-lactam antibiotics. Susceptibility patterns of transconjugants correlated with the data obtained for the corresponding clinical isolates.

TABLE 3.

Antimicrobial susceptibilities of clinical isolates and transconjugants

Strain (PFGE type, pI of ESBL)	No. of isolates	MIC(s) (μg/ml) or range of MICs
		AMPa	PIP	PIP-TAZ	CTX	CTX-CLAV	CAZ	CAZ-CLAV	ATM	ATM-CLAV	FOX	IPM	AK	GN
K. pneumoniae (a, 5.7)	1	>512	>512	>512	32	8	256	128	256	64	32	0.25	2	128
K. pneumoniae (a, 6.0)	8	>512	>512	2–4	8–16	≤0.03–0.06	128	0.25–0.5	128	0.125	2–4	0.125	1–2	32, 128
K. pneumoniae (d, 6.0)	3	>512	>512	32	16	0.125–0.25	128	1	256	0.25	8	0.125	1	32
K. pneumoniae (b2, 6.0)	2	>512	>512	4–8	16	0.06–0.125	64–128	0.5–1	128–256	0.125–0.25	4–8	0.125	1	32–64
K. pneumoniae (b2, 8.2)	1	>512	128	4	8	0.125	128	0.5	256	0.125	4	0.125	16	4
K. pneumoniae (b1, 8.2)	5	>512	256	4–8	8–16	0.125–0.25	128	0.5–1	256	0.125–0.25	4–8	0.125	8–16	4–8
K. pneumoniae (c, 8.2)	2	>512	256	16	8	0.25	128–256	0.5–1	256–512	0.125	16	0.125–0.25	8	4
K. oxytoca (8.2)	2	512–>512	64	0.5–1	1–2	≤0.03	8	0.125	8–16	≤0.03	1	0.125	16	4–8
K. pneumoniae R+b (6.0)	11	>512	64–128	0.5	2–4	≤0.03	8	0.06–0.125	8–16	≤0.03	1–2	0.125	0.5	16–32
K. pneumoniae, K. oxytoca R+ (8.2)	7	256	16–32	0.5	1	≤0.03	8	0.125	16	≤0.03	1–2	0.125	4–8	2
E. coli A15 R−		2	≤0.5	≤0.5	≤0.03	≤0.03	0.125	0.125	0.06	≤0.03	1	0.125	0.5	0.25
E. coli ATCC 25922		4	2	2	0.06	≤0.03	0.25	0.125	0.06	0.06	4	0.125	1	0.5

^aAMP, ampicillin; PIP, piperacillin; CTX, cefotaxime; CAZ, ceftazidime; ATM, aztreonam; FOX, cefoxitin; IPM, imipenem; TAZ, tazobactam (4 μg/ml); CLAV, clavulanate (2 μg/ml); AK, amikacin; GN, gentamicin. 

^bR⁺, transconjugants. 

Comparative analysis of TEM-47 and TEM-68 activities.

In order to compare activities of TEM-47 and TEM-68, genes coding for both β-lactamases were cloned and expressed in an isogenic system. The cloned DNA fragments containing the bla_TEM-47 and bla_TEM-68 genes together with their promoters differed from each other only by a single mutation specifying the Arg275Leu substitution in the TEM-68 enzyme. The pGB2 plasmid, which was used as a vector, does not contain any other β-lactamase-encoding gene.

Table 4 presents MICs for E. coli DH5α strains transformed with pGBT-47 (TEM-47) and pGBT-68 (TEM-68) constructs together with MICs for the K. pneumoniae 3144/98 (TEM-47) and 3151/98 (TEM-68) isolates, which were the sources of the bla_TEM genes. In contrast to clinical isolates, for the E. coli transformants producing TEM-47 and TEM-68, there were no differences in the MICs of inhibitor combinations, in which the constant concentrations of tazobactam and clavulanate were 4 and 2 μg/ml, respectively. The same set of strains was used in the evaluation of MICs of piperacillin in combinations with various concentrations of tazobactam and clavulanate. Results of the analysis are shown in Table 5. A significant difference in piperacillin MICs between the pGBT-47 and pGBT-68 transformants was observed starting with a tazobactam concentration of 1 μg/ml (MICs, 1 and 64 μg/ml, respectively) and a clavulanate concentration of 0.03 μg/ml (MICs, 4 and 32 μg/ml, respectively).

TABLE 4.

Antimicrobial susceptibilities of clinical isolates and E. coli transformants producing TEM-47 and TEM-68

Strain (ESBL type)	MIC (μg/ml)
	AMPa	PIP	PIP-TAZ	CTX	CTX-CLAV	CAZ	CAZ-CLAV	ATM	ATM-CLAV	FOX	IPM	AK	GN
K. pneumoniae 3144/98 (TEM-47)	>512	>512	2	8	≤0.03	128	0.25	128	0.125	4	0.125	1	32
K. pneumoniae 3151/98 (TEM-68)	>512	>512	>512	32	8	256	128	256	64	32	0.25	2	128
E. coli DH5α/pGBT-47 (TEM-47)	>512	256	1	8	≤0.03	64	0.25	64	0.125	8	0.125	0.5	0.25
E. coli DH5α/pGBT-68 (TEM-68)	>512	256	1	8	≤0.03	64	0.25	64	0.125	8	0.125	0.5	0.25
E. coli DH5α	2	≤0.5	1	0.06	≤0.03	0.25	0.125	0.06	0.125	8	0.125	0.5	0.25

^aAMP, ampicillin; PIP, piperacillin; CTX, cefotaxime; CAZ, ceftazidime; ATM, aztreonam; FOX, cefoxitin; IPM, imipenem; TAZ, tazobactam (4 μg/ml); CLAV, clavulanate (2 μg/ml); AK, amikacin; GN, gentamicin. 

TABLE 5.

MICs of piperacillin combined with various concentrations of clavulanate and tazobactam for TEM-47- and TEM-68-producing klebsiella clinical isolates and E. coli transformants

Strain (ESBL type)	MIC (μg/ml) of piperacillin–clavulanate/tazobactam at inhibitor concn (μg/ml) of:
	0	0.0075	0.015	0.03	0.06	0.125	0.25	0.5	1	2	4
K. pneumoniae 3144 (TEM-47)	>512/>512	512/—a	256/—	64/—	16/—	8/—	4/—	4/256	4/4	4/2	2/2
K. pneumoniae 3151 (TEM-68)	>512/>512	>512/—	>512/—	>512/—	>512/—	>512/—	>512/—	>512/>512	>512/>512	512/>512	8/>512
E. coli DH5α/pGBT-47 (TEM-47)	256/256	64/—	32/—	4/—	2/—	1/—	1/—	1/2	1/1	1/1	0.5/1
E. coli DH5α/pGBT-68 (TEM-68)	256/256	128/—	128/—	32/—	4/—	2/—	1/—	1/128	1/64	1/2	0.5/1
E. coli DH5α	0.5/1	0.5/—	0.5/—	0.5/—	0.5/—	0.5/—	0.5/—	0.5/1	0.5/1	0.5/1	0.5/1
E. coli ATCC 25922	2/2	2/—	2/—	2/—	2/—	2/—	2/—	2/2	2/2	2/2	2/2

^a—, not determined. 

Protein extracts of the pGBT-47 and pGBT-68 transformants were used for comparative evaluation of IC₅₀s. The tazobactam IC₅₀ for TEM-68, 0.400 μM, was about 10 times higher than that for the TEM-47 enzyme, 0.045 μM. The difference in clavulanate IC₅₀s was less marked, with values of 0.150 and 0.033 μM, respectively.

DISCUSSION

Twenty-four ESBL-producing klebsiella isolates were collected in 1996 from patients in two wards of the University Hospital in Wrocław. Molecular typing has revealed a remarkable clonal diversity within the group, with four distinct clusters of related K. pneumoniae isolates and one of K. oxytoca. The most prevalent were K. pneumoniae isolates of PFGE types a and b, and these most probably represented two strains with relatively high epidemic potential, clonally spread in the neonatal ward and the PICU, respectively.

At least three different ESBLs of the TEM (pI, 5.7 or 6.0) and SHV (pI, 8.2) families were produced by the isolates. The pI 6.0 β-lactamases were identified in 13 K. pneumoniae isolates of three different PFGE types (a, b2, and d), all of which contained plasmids with very similar restriction patterns (type A plasmids). Sequencing of bla_TEM genes amplified from representative isolates revealed that these were bla_TEM-47 genes with the same sequence. An analogous situation was observed in the case of isolates producing the pI 8.2 SHV ESBLs. These enzymes were identified in eight K. pneumoniae isolates differing by PFGE (b1, b2, and c) and in the isolates of K. oxytoca, all with large plasmids of the same or very similar fingerprints (type B plasmids). PCR with the use of specifically designed primers revealed that all of the SHV enzymes contained the Gly238Ser and Glu240Lys ESBL-specific substitutions; sequencing of the bla_SHV gene from a single isolate demonstrated that it codes for SHV-5. It is likely that all of the pI 6.0 and 8.2 ESBLs analyzed were TEM-47 and SHV-5, respectively, and that genes coding for both kinds of enzymes were disseminated among nonrelated klebsiellae by parallel plasmid transfer events.

Identification of TEM-47-producing K. pneumoniae isolates in the hospital in Wrocław raised the question of their possible relatedness to the K. pneumoniae clinical isolates expressing the same enzyme and recovered in pediatric hospitals in Łódź in 1995 (15) and in Warsaw in 1996 (14). The comparative typing of isolates representing all of the identified PFGE types and subtypes of TEM-47 producers from the three institutions revealed that the fraction of isolates from Wrocław (subtype b2) was very closely related to isolates from the other cities. This data indicated a very probable transfer of the epidemic K. pneumoniae strain between the pediatric centers, leading to its spread over the country.

A novel variant of a TEM β-lactamase, TEM-68 (pI 5.7), was identified in the single K. pneumoniae isolate 3151/98 recovered from a patient in the neonatal ward. This isolate was indistinguishable by PFGE and RAPD from the PFGE type a TEM-47-producing isolates. The bla_TEM-68 gene, together with its promoter region, differs by only a single point mutation from bla_TEM-47 and was carried by the nonconjugative plasmid (fingerprint A4), which very likely has emerged by recombination from the transferable type A1 molecule. All of this data indicated that the bla_TEM-68 gene has most probably evolved from bla_TEM-47 in the genetic background of the PFGE type a K. pneumoniae strain. This finding has extended our view of TEM-2-related ESBL evolution in Poland (Table 2). In the previous study, it was postulated that genes coding for TEM-47 and TEM-49 have emerged independently by single genetic events from the TEM-48-encoding sequence (15).

TEM-68 represents a new variant of complex mutant β-lactamases which combine ESBL-specific mutations with those determining IR activity. Two β-lactamases of this kind have been identified in clinical isolates to date, and these were TEM-50 (40) and SHV-10 (33). Two ESBL-specific mutations, Glu104Lys and Gly238Ser, and two IR-type substitutions, Met69Leu and Asn276Asp, are known to characterize the TEM-50 enzyme. The SHV-10 β-lactamase combines the Gly238Ser and Glu240Lys (both ESBL-type) substitutions and the Ser130Gly (IR-specific) substitution. Activities of the complex mutant β-lactamases were compared with those of the corresponding ESBL and IR enzymes. It was shown that the SHV-10 β-lactamase manifested increased resistance to inhibitors and no ESBL activity compared to SHV-9 (33). On the other hand, the TEM-50 enzyme retained both activities; however, the ESBL activity was found to be weaker than that of the TEM-15 β-lactamase and the IR activity was reduced compared with that of TEM-35/IRT-4 (40).

TEM-68 contains the Gly238Ser and Glu240Lys ESBL-type substitutions and the Arg275Leu mutation, which up to now has been identified once, in the TEM-38/IRT-9 β-lactamase (18). Another substitution at this site, the Arg275Gln mutation, was found in the TEM-45/IRT-14 enzyme (11). Both TEM-38 and TEM-45 are enzymes with strong IR activity; however, the Arg275 mutations are accompanied by well-characterized IR-type Met69 substitutions (21) within their sequences. The Arg275 substitutions have never been observed separately until now, and so their contribution to the IR phenotype, even if postulated, was not clear (11, 18). Comparative analysis of the TEM-47 and TEM-68 β-lactamases, which differ only by the Arg275Leu substitution and are expressed at the same level in an isogenic background, has provided an opportunity to study the influence of this substitution on β-lactamase activity.

Compared with the PFGE type a TEM-47 producers (e.g., isolate 3144/98), for the TEM-68-expressing K. pneumoniae 3151/98 isolate, the MICs of penicillins, ceftazidime, and aztreonam were similarly high and the MICs of the inhibitor combinations studied were much increased (Table 4). For the TEM-68-producing isolate, the MIC of cefoxitin was also significantly higher than for the TEM-47-expressing isolate. Since both of these isolates belonged to the same PFGE and RAPD types, it could be suggested that the differences in the MICs of β-lactams were due mostly to the diverse activities of their β-lactamases. In order to check this hypothesis, the bla_TEM-47 and bla_TEM-68 genes were cloned together with their original, identical promoters in the E. coli laboratory strain. The resulting transformants were characterized by MIC evaluation, and protein extracts of the recombinant strains were used for determination of IC₅₀s for the TEM-47 and TEM-68 enzymes. Surprisingly, for the TEM-47- and TEM-68-producing E. coli transformants, the MICs of all of the β-lactams tested were identical, including inhibitor combinations in which clavulanate and tazobactam were at fixed routine concentrations of 2 and 4 μg/ml, respectively (Table 4). The inhibitors efficiently reduced the MICs of piperacillin, cefotaxime, ceftazidime, and aztreonam. Nevertheless, when inhibitor concentrations were reduced, remarkable differences between TEM-47 and TEM-68 recombinant producers were revealed, demonstrating the IR activity of TEM-68, which affected the inhibition by tazobactam more than that by clavulanate (Table 5). Both effects were reflected by kinetic data; the IC₅₀s of the inhibitors were significantly higher for TEM-68 than for TEM-47, and this difference was more explicit in the case of tazobactam (about 10 times) than in that of clavulanate (about 5 times). All of the results indicated that TEM-68 is a complex mutant β-lactamase which, similar to the TEM-50 enzyme, combines ESBL and IR activities. In contrast to TEM-50, its ESBL activity is high and fully comparable to that of TEM-47, the corresponding ESBL variant. The IR activity of TEM-68 is determined by the Arg275Leu mutation alone and is relatively weak since this could be demonstrated only at reduced concentrations of inhibitors in the E. coli laboratory strain. However, the effect of the IR activity was strongly enhanced when the enzyme was produced by the original wild-type strain of K. pneumoniae, which was most probably of lower permeability for antibiotics than the E. coli laboratory strain. It is possible that the permeability of the TEM-68-producing isolate, 3151/98, was additionally reduced by a mutation, as suggested by the raised MIC of cefoxitin. The ESBL and IR activities of TEM-68, when expressed in the context of the relatively low permeability of K. pneumoniae cells, together conferred on the clinical isolate a remarkably high level of resistance to a wide spectrum of β-lactam antibiotics, including combinations of piperacillin with tazobactam and oxyimino β-lactams with clavulanate. TEM-68-producing isolate 3151/98 is the first reported strain of K. pneumoniae expressing a complex mutant β-lactamase and one of the very few examples of non-E. coli isolates producing a class A β-lactamase with IR activity (9, 23, 39).

This work allowed us to document several concurrent epidemiological phenomena concerning ESBL-mediated resistance in a single medical center. Five different ESBL-expressing klebsiella strains were clonally disseminated at the same time. They produced at least three different ESBL variants, and two of these were spread due to plasmid transfer among the nonrelated strains. ESBL-producing strains must have been transmitted between the two wards and were undergoing evolutionary diversification. Two different plasmids carrying different ESBL genes occurred alternatively in cells of a single epidemic strain, and one of these strain variants was spread to other hospitals located in distant cities. The ongoing evolution of ESBLs has led to a new complex mutant enzyme conferring resistance to a very wide variety of β-lactam antibiotics, including inhibitor combinations.