High-Level Expression of Chromosomally Encoded SHV-1 β-Lactamase and an Outer Membrane Protein Change Confer Resistance to Ceftazidime and Piperacillin- Tazobactam in a Clinical Isolate of Klebsiella pneumoniae

Abstract

We describe Klebsiella pneumoniae 15571, a clinical isolate resistant to ceftazidime MIC = 32 μg/ml) and piperacillin-tazobactam (MICs = 1,024 and 128 μg/ml). K. pneumoniae 15571 expresses a single β-lactamase with a pI of 7.6. However, when cloned in a high-copy-number vector in Escherichia coli, this bla_SHV-1 gene did not confer resistance to ceftazidime, a spectrum consistent with the nucleotide sequence, which was nearly identical to those of previously described bla_SHV-1 genes. Outer membrane protein (OMP) analysis of K. pneumoniae 15571 revealed a decrease in the quantity of a minor 45-kDa OMP in comparison to that in K. pneumoniae 44NR, a low-level ampicillin-resistant strain that also expresses a chromosomally determined bla_SHV-1. Crude β-lactamase enzyme extracts from K. pneumoniae 15571 produced roughly 200-fold more β-lactamase activity than K. pneumoniae 44NR. Northern hybridization analysis revealed that this difference was explainable by quantifiable differences in transcription of the bla_SHV-1 gene in the two strains. Primer extension analysis of bla_SHV-1 mRNA from K. pneumoniae 15571 and 44NR indicated that the transcriptional start sites were identical in the two strains. DNA sequencing of the promoter regions upstream of the of bla_SHV-1 open reading frames in the two K. pneumoniae strains revealed an A→C change in the second position of the −10 region in K. pneumoniae 44NR compared to that in 15571. Site-directed mutagenesis of the cloned K. pneumoniae 15571 bla_SHV-1, in which the A in the second position of the 15571 −10 region was changed to a C, resulted in a substantial lowering of the MIC of ampicillin. When the levels of β-lactamase enzyme expression in E. coli were compared, the bla_SHV-1 downstream of the altered −10 region produced 17-fold less β-lactamase enzyme. These results indicate that elevated levels of ceftazidime resistance can result from a combination of increased enzyme production and minor OMP changes and that levels of chromosomally encoded SHV-1 β-lactamase production can vary substantially with a single-base-pair change in promoter sequence.

Resistance to ceftazidime in Klebsiella pneumoniae is most commonly attributed to the production of plasmid-mediated extended-spectrum β-lactamases (ESBLs). These enzymes are frequently encoded by multidrug resistance conjugative plasmids and evolve from the more common TEM-1 and SHV-1 penicillinases through point mutations in regions important for β-lactam binding and/or hydrolysis (10). Conjugative dissemination of ESBL-encoding plasmids is thought to facilitate the spread of resistance in the clinical setting. In rare cases, resistance to ceftazidime has been attributed solely to increased production of SHV-1, although MICs of ceftazidime for these strains are only mildly increased (ca. 8 μg/ml) (15).

Resistance to β-lactam–β-lactamase inhibitor combinations occurs most commonly in gram-negative bacilli through the overproduction of common plasmid-mediated enzymes (20). Overproduction has been attributed to altered promoter sequences, gene duplication, or the presence of an ESBL gene on a multicopy plasmid (15, 20, 22). Rarely, a derivative of TEM-1 or SHV-1 is found that confers high-level resistance to both ceftazidime and β-lactam–β-lactamase inhibitor combinations (TEM-50) (21). This phenotype can be achieved by other means, including the overproduction of an ESBL, the coexistence of an ESBL with an SHV-1 enzyme produced in large quantities, or the acquisition of a plasmid-encoded ampC-type gene (3, 15, 17).

A single point mutation 177 bp upstream of the TEM initiation codon that results in the creation of dual overlapping promoters is the most commonly described promoter mutation resulting in overproduction of TEM-type enzymes (4). Transcription from these two new promoters results in TEM production that is roughly 10-fold higher than with the standard TEM-1 promoter sequence. For reasons that remain unclear, a high percentage of TEM-type ESBL genes have been identified downstream of this more active promoter (8).

Although SHV-1 was originally characterized as a plasmid-mediated β-lactamase, recent data suggest that its production may be intrinsic to K. pneumoniae, since it appears that many clinical K. pneumoniae strains (more than 90% of one series) encode SHV-1 β-lactamase production from their chromosomes (11). In most cases, β-lactamase production encoded by these chromosomal genes is at very low levels, resulting in relatively low levels of resistance to ampicillin (MIC, ca. 8 to 64 μg/ml). The product of the first sequenced chromosomal β-lactamase gene from K. pneumoniae was designated LEN-1 (1). The nucleotide sequence of the gene encoding LEN-1 is 85% identical to that of the gene encoding SHV-1, suggesting a close evolutionary relationship between the two enzymes (13). Recent data suggest that chromosomally encoded SHV-1 may be more prevalent than LEN-1 in clinical K. pneumoniae isolates (11).

Promoter mutations that are associated with increased production of SHV-1 are less well described than those for TEM-type genes. Original reports identified putative −35 and −10 promoter sequences upstream of the SHV-1 gene based on sequence similarity to Escherichia coli consensus promoter sequences (13). A more recent report identified the specific transcriptional start site of the SHV-1 gene as 50 bp upstream of the previously presumed transcriptional start site (16). In this report, a more active promoter sequence was identified upstream of an ESBL gene (bla_SHV-2a) resulting from the insertion of IS15 upstream of the gene (17). This promoter change was associated with a fivefold increase in the synthesis of SHV-2a mRNA compared to that observed with the traditional promoter. The extent to which extended-spectrum SHV-related β-lactamase genes are preceded by more active promoter sequences is unknown.

We report a K. pneumoniae strain that expresses in vitro resistance to both ceftazidime and β-lactam–β-lactamase inhibitor combinations. The sole enzyme expressed by this strain is a chromosomally encoded SHV-1 that is produced in large amounts, partially accounting for the increased level of ceftazidime resistance. This strain also exhibited decreased production of a minor 45-kDa outer membrane protein that may be involved in the transport of antimicrobial agents. We present evidence that this chromosomal bla_SHV-1 gene is encoded by an open reading frame downstream of what is considered to be the typical bla_SHV-1 promoter sequence. This promoter is associated with 200-fold greater SHV-1 production than an SHV-1-like enzyme produced by a second K. pneumoniae strain in which the promoter differs by only 1 nucleotide in the −10 region. We also present confirmatory evidence that this single-nucleotide change is responsible for most, if not all, of the increased β-lactamase production observed in this strain.

MATERIALS AND METHODS

Bacterial strains and plasmids.

The bacterial strains and recombinant plasmids used and analyzed in these studies are described in Table 1. K. pneumoniae 15571 was a clinical isolate from the urinary tract of an infected adult patient. K. pneumoniae 44NR and 44NRF have been described previously (18). In the prior publication, K. pneumoniae 44NR was described in the text as producing “no β-lactamase.” In Table 1 from that article, however, it is clear that the strain produced a small but detectable amount of β-lactamase. The present work confirms that this strain does in fact produce β-lactamase, but at very low levels.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Resistancea or genotype	Description (reference or source)
K. pneumoniae 15571	Apr Ceftazr P/Tr	Clinical isolate (this study)
K. pneumoniae 44NR	Nalr Rifr Apr	Nalidixic acid- and rifampin-resistant mutant derived from a clinical isolate (18)
K. pneumoniae 44NRF	Nalr Rifr Apr Foxr	Cefoxitin-resistant mutant of K. pneumoniae 44NR; cefoxitin resistance related to decreased expression of a 39-kDa outer membrane protein (18)
E. coli DH10B	F−mcrA Δ(mrr-hsdRMS-mcrBC) φ80dlacZΔM15 DlacX74 deoR recA1 araΔ139 Δ(ara,leu)7697 galU galK λ-rpsL endA1 nupG	Transformation-competent E. coli (BRL)
E. coli J53-2 (Rifr)	Rifr	Conjugation recipient (6)
Plasmids
pCWR338	Apr Cmr	3.6-kb blaSHV-1-containing BamHI fragment from 15571 cloned into pBC SK(−) (this study)
pCWR366	Apr Cmr	1.4-kb ClaI-ScaI subfragment of pCWR338 insert cloned into EcoRV-ClaI-digested pBCSK(−) (this study)
pCWR498	Apr Cmr	3.6-kb blaSHV-1-containing BamHI fragment from 44 cloned into pBCSK(−) (this study)
pBCSK(−)	Cmr	Cloning vector (Stratagene)

^aAp, ampicillin; Ceftaz, ceftazidime; Cm, chloramphenicol; Fox, cefoxitin; Nal, nalidixic acid; P/T, piperacillin-tazobactam; Rif, rifampin. 

Microbiological techniques.

MICs of different antibiotics were determined by standard agar dilution (14) techniques, except that Luria-Bertani (LB) agar was used instead of Mueller-Hinton agar and we used a fixed ratio of piperacillin-tazobactam (8:1) rather than a constant inhibitor concentration of 4 μg/ml.

Conjugation experiments.

Matings between K. pneumoniae 15571 and E. coli J53-2 (Rif^r) were carried out overnight at 37°C on sterile nitrocellulose filters as previously described (17).

β-Lactamase assays.

Crude β-lactamase extracts were prepared from 5-ml overnight cultures. K. pneumoniae and E. coli cells containing β-lactamase enzymes were grown to mid-log phase (optical density at 600 nm, 0.5) in LB broth containing 100 μg of ampicillin/ml. The cells were pelleted, washed, and resuspended in 500 μl of 50 mM Tris-HCl buffer, pH 7.4. A 40-mg/ml stock solution of freshly prepared lysozyme (Sigma, St. Louis, Mo.) in Tris-HCl buffer was added to a final concentration of 10 μg/ml, and bacterial cells were incubated for 15 min at room temperature. EDTA was added to a final concentration of 1 mM, and the mixture was gently shaken for 10 min. The cell suspension was clarified by centrifugation at 14,000 × g for 15 min, and the supernatant was collected. β-Lactamase activity in the supernatant was measured with the chromogenic cephalosporin nitrocefin (Becton Dickinson, Cockeysville, Md.). The amount of protein in each β-lactamase preparation was determined by the Bio-Rad (Hercules, Calif.) protein assay. The specific activity present in each sample was estimated by measuring the hydrolysis of 100 μM nitrocefin at 25°C in a Hewlett-Packard 8452 diode array spectrophotometer (λ = 482 nM; Δɛ = 17,400 M⁻¹ cm⁻¹). The initial (maximal) velocity was used to determine specific activity. Specific activity was defined as U/mg of protein. One unit was defined as the amount of nitrocefin hydrolyzed (micromolar) per minute.

aIEF.

Analytical isoelectric focusing studies (aIEF) were performed with a Multiphor isoelectric focusing apparatus (Pharmacia, Piscataway, N.J.) according to a modification of the method of Vecoli et al. (25). The crude β-lactamase preparations and prestained standards, pI 4.5 to 9.5, were directly applied to Ampholine PAGplates, pI 3.5 to 9.5 (Pharmacia). β-Lactamase bands were identified with an overlay of 0.5 mg of nitrocefin (100 μM dilute solution)/ml.

Outer membrane protein analysis.

Outer membrane proteins from three clinical strains of K. pneumoniae (15571, 44NR, and 44NRF) were isolated according to the method of Spratt (23) with modifications employed by Gutmann et al. (9).

Molecular techniques.

Plasmid DNA was extracted from clinical strains by a technique described by Takahashi and Nagano (24). Genomic DNA was extracted as previously described (19) and digested with restriction enzymes (Promega, Madison, Wis.) as recommended by the manufacturer. Digested DNA was separated on agarose gels and transferred to nylon membranes with a negative-pressure transfer apparatus (Pharmacia). An internal probe for the bla_SHV-1 gene was generated by PCR with primers designed to amplify a fragment internal to the bla_SHV-1 gene (SHV-238, 5′-CCGCGTAGGCATGATAGAAA-3′, and SHV-894, 5′-TCCCGCAGATAAA-3′). The template for the PCR was a cloned bla_SHV-1 gene (recombinant plasmid pCWR101) previously described (17). The probes were labeled with digoxigenin (Boehringer Mannheim) according to the specifications of the manufacturer. Hybridization occurred overnight at 68°C, after which the membranes were washed under conditions of high stringency. Labeled fragments were detected with an anti-digoxigenin-alkaline phosphatase conjugate and chromogenic enzyme substrate.

Cloning and sequencing techniques.

SHV-1-encoding BamHI fragments (3.6 kb) were identified in K. pneumoniae 15571 and K. pneumoniae 44NR by Southern hybridization with the PCR-generated internal bla_SHV-1 probe. Gel slices containing these fragments were excised and ligated to BamHI-digested pBC SK(−) (Stratagene, La Jolla, Calif.). Commercially purchased electrocompetent E. coli DH10B cells were transformed with the ligation mixture by electroporation, and transformed colonies were selected on LB agar plates containing chloramphenicol (20 μg/ml) and ampicillin (100 μg/ml). The resultant recombinant plasmids were designated pCWR338 (from 15571) and pCWR498 (from 44NR). pCWR338 was further subcloned with the restriction enzymes ScaI and ClaI. The complete nucleotide sequence of the cloned bla_SHV-1 gene from K. pneumoniae 15571 was determined on both strands with forward and reverse primers as well as commercially synthesized primers internal to the bla_SHV-1 open reading frame. Sequencing reactions were performed with the Cy-5 Thermosequenase dye terminator kit (Pharmacia) with unlabeled primers or primers were end-labeled with Cy-5 and sequencing was carried out with the Thermosequenase fluorescent-labeled primer cycle-sequencing kit (Pharmacia). Sequences were determined with the A.L.F. Express automated sequencer (Pharmacia), and analysis was carried out with either the MacDNAsis analysis program (Hitachi) or the Lasergene Navigator (DNAStar) analysis program.

Northern hybridization.

Total cellular RNA was extracted from the strains in this study with the Qiagen (Valencia, Calif.) RNeasy miniprep kit according to a protocol supplied by the manufacturer. RNA concentrations were measured spectrophotometrically, and equal amounts of RNA were separated on 1.2% agarose gels containing formaldehyde. The transfer of RNA from gels to nylon membranes was accomplished by capillary transfer. The digoxigenin-labeled internal bla_SHV-1 probe was generated as described above and used to hybridize the RNA fixed to the nylon membrane. The hybridized membranes were washed under high-stringency conditions, and hybridized RNA fragments were detected by a chemiluminescence technique with disodium 3-(4-methoxyspiro [1,2-dioxetane-3,2′-(5′-chloro)tricyclo[3.3.1.1^3–7]decan]-4-yl)phenyl phosphate (Boehringer Mannheim) as the reagent.

Primer extension analysis.

RNA was extracted and purified according to a commercial protocol (Rneasy Miniprep kit, Qiagen). Primer extension analysis was carried out with Cy-labeled primers designed to direct synthesis from the 3′ to the 5′ end of the open reading frame. The technique employed high temperatures to minimize complications resulting from secondary-structure formation, essentially as described by Yamada et al. (27).

Site-directed mutagenesis.

In vitro site-directed mutagenesis was performed with the QuikChange site-directed mutagenesis kit manufactured by Stratagene. Two complementary mutagenic oligonucleotide primers were constructed at the Case Western Reserve University Molecular Biology Core Laboratory with the following restrictions: a melting temperature higher than 78°C, a G+C content greater than 40%, and termination in a C or G nucleotide. To fulfill these requirements, we designed a 43-mer due to the high AT content of the promoter region. The primers were nonphosphorylated and high-performance liquid chromatography purified. Mutagenesis was performed with Pfu polymerase supplied with the kit. The plasmid pBC SK(−), prepared with the Promega Miniprep Wizard kit and containing bla_SHV-1, was used as the DNA template for the mutagenic PCR. The cycling parameters for mutagenesis were 95°C for 30 s, 55°C for 1 min, and 68°C for 10 min for a total of 16 cycles. DpnI was added after amplification to digest methylated (parental) DNA. Epicurean coli (E. coli) XL1-Blue supercompetent cells (Stratagene) were transformed with mutagenic DNA by heat pulse for 45 s at 42°C and then placed on ice for 2 min. The transformed cells were next incubated at 37°C for 1 h and plated, and single-colony transformants were selected the next day.

Nucleotide sequence accession number.

The nucleotide sequence for the bla_SHV-1 gene from K. pneumoniae 15571 has been entered in GenBank under accession no. AF124984.

RESULTS

Characterization of K. pneumoniae 15571.

The MICs for ceftazidime and piperacillin-tazobactam were 32 μg/ml and 1024 and 128 μg/ml, respectively, against K. pneumoniae 15571 (Table 2). Repeated conjugation experiments failed to result in the transfer of ampicillin or ceftazidime resistance to E. coli J53-2 when ampicillin (100 μg/ml) was used for selection (data not shown). In addition, plasmid analysis revealed several plasmids in this strain, none of which hybridized to the internal bla_SHV-1 probe (data not shown). Hybridization of K. pneumoniae 15571 genomic DNA digested individually with five different restriction enzymes revealed a single bla_SHV-1-hybridizing genomic band with each digestion (data not shown).

TABLE 2.

Susceptibilities and β-lactamase activities of different strains

Strain	MIC (μg/ml)
	Ampicillin	Piperacillin-tazobactam	Cefoxitin	β-Lactamase activity (U/mg of protein)
K. pneumoniae 15571	16,384	1024/128	16	1,280
K. pneumoniae 44NR	256	8/1	16–32	6
E. coli DH10B(pCWR366)	32,000	512/64	4	12,800
E. coli DH10B(pCWR366) (A→C mutant)	2,048	128/16	NDa	762
E. coli DH10B	4	<8/0	4	None

^aND, not determined. 

aIEF.

aIEF of crude enzyme extracts from K. pneumoniae 15571 indicated the production of a single β-lactamase with a pI of 7.6.

Outer membrane protein analysis.

Outer membrane protein analysis (Fig. 1) suggested a reduction in a minor (ca. 45-kDa) outer membrane protein in strain 15571 in comparison to that in K. pneumoniae 44NR (Fig. 1).

FIG. 1.

Outer membrane proteins were separated on a sodium dodecyl sulfate-containing polyacrylamide gel (acrylamide/bisacrylamide ratio, 29/1; 5% stacking gel with a 12% separating gel). Lane 1, molecular mass marker; lane 2, K. pneumoniae 15571; lane 3, K. pneumoniae 44NR; lane 4, K. pneumoniae 44NRF. Note the diminished quantity of 45-kDa protein in lane 2 versus lane 3 (arrow).

Northern hybridization.

In order to compare transcription of the bla_SHV-1 genes in K. pneumoniae 15571 and 44NR, we performed Northern hybridizations of equal amounts of mRNA extracted from the two strains, using the internal bla_SHV-1 fragment as a probe. The results are shown in Fig. 2. The hybridization signal from K. pneumoniae 15571 was dramatically greater than that in K. pneumoniae 44NR or its outer membrane protein-deficient, isogenic mutant, 44NRF. These data suggested that the differences in MICs were the result of marked differences in transcription of the genes in the two strains.

FIG. 2.

Northern hybridization of mRNAs extracted from K. pneumoniae 15571 and 44NR. Note the substantially greater levels of mRNA in K. pneumoniae 15571 compared to those in K. pneumoniae 44NR (44). K. pneumoniae 44NRFox (44 Fox) is a clinical strain with a missing outer membrane protein.

Cloning and characterization of the K. pneumoniae 15571 chromosomal β-lactamase gene.

The 3.6-kb bla_SHV-1-hybridizing BamHI restriction fragment from K. pneumoniae 15571 was ligated to pBC SK(−) and transformed into E. coli DH10B. The recombinant plasmid, designated pCWR338, was further digested with ScaI and ClaI, resulting in a reduction of the insert size to 1,342 nucleotides, and ligated to EcoRV-ClaI-digested pBC SK(−), resulting in recombinant plasmid pCWR366. E. coli DH10B(pCWR366) expressed high levels of ampicillin resistance (MIC, >16,000 μg/ml) but lower levels of ceftazidime resistance than the K. pneumoniae 15571 parent strain, implying a contribution of the outer membrane protein change to ceftazidime resistance. The entire bla_SHV-1 open reading frame was sequenced and found to be nearly identical to previously described bla_SHV-1 open reading frames (2). Specifically, no amino acid changes were noted when comparison was made to the previously reported mature enzyme. In addition, the area upstream of the open reading frame that was reported to contain the promoter region was sequenced and found to be identical to previously reported regions for bla_SHV-1 and bla_SHV-2. Importantly, sequence alterations previously reported to be associated with increased expression of bla_SHV-2a were not identified (16).

Promoter analysis.

In order to further analyze the effect of promoter changes on the expression of β-lactamase in K. pneumoniae 15571 in comparison to that in K. pneumoniae 44NR, we pursued a detailed comparison of the bla_SHV-1 promoter regions in the two strains. The MIC of ampicillin for K. pneumoniae 44NR was much lower than that for K. pneumoniae 15571, and 44NR expressed at least 200-fold less β-lactamase activity (Table 2). To ensure that this difference was not attributable to the SHV-1 protein itself, we cloned and sequenced the SHV-1 gene from the 44NR chromosome. A total of seven nucleotide changes were observed when the sequence of the 44NR β-lactamase gene was compared with that of 15571. Five of these changes were silent in that they conferred no change in the amino acid sequence. The other two led to amino acid changes. Using the adenine of the ATG start codon as nucleotide 1, we observed an A→G change at nucleotide 8 and a T→A change at nucleotide 9, which together resulted in a Tyr→Leu substitution at the third amino acid. The amino acid change at position 3 is removed by the signal peptidases and is not part of the mature enzyme. Therefore, it is extremely unlikely to confer significant changes to the enzymatic activity of the β-lactamase. The silent nucleotide changes occurred at nucleotides 324 (C→T), 402 (A→G), 633 (G→A), 705 (G→A), and 762 (T→C).

That the difference between the MICs for 15571 and 44NR was explainable by differences in transcription was confirmed by Northern hybridization of mRNAs from the two strains, indicating a substantial increase in bla_SHV-1 mRNA in K. pneumoniae 15571 in comparison to K. pneumoniae 44NR (Fig. 2). Analysis of the promoter region in this insert revealed two potentially important changes relative to the 15571 promoter (Fig. 3). The nucleotide that lies 4 bp upstream of the −35 region differs in K. pneumoniae 44 (T), K. pneumoniae 15571 and bla_SHV-2 (G), and bla_LEN-1 (A). In addition, the second nucleotide of the −10 region is a C in K. pneumoniae 44NR and bla_LEN-1 while it is an A in K. pneumoniae 15571 and bla_SHV-2. These differences suggest that the differences among the ampicillin MICs for K. pneumoniae LEN-1 (reported to be 7.8 μg/ml), K. pneumoniae 44 (256 μg/ml), and K. pneumoniae 15571 (>16,000 μg/ml) may be explainable, at least in part, by differences in the region upstream of the −35 region and internal to the −10 region.

FIG. 3.

Comparison of promoter sequences upstream of different SHV-type genes. K. p. LEN-1 denotes the sequence upstream of the chromosomal β-lactamase gene of K. pneumoniae LEN-1 (ampicillin MIC, 7.8 μg/ml). K. p. 44 denotes the sequence of the chromosomal β-lactamase gene of K. pneumoniae 44NR (ampicillin MIC, 128 to 256 μg/ml [see the text]). K. p. 15571 denotes the sequence upstream of the chromosomal β-lactamase gene of K. pneumoniae 15571 (ampicillin MIC, 32,000 μg/ml). The published sequences of the promoter regions of the bla_SHV-2 and bla_SHV-2a genes are also shown. The boxes mark the different −35 and −10 regions. The horizontal arrows indicate a direct repeat within the typical −10 region. The vertical arrow indicates the nucleotide upstream of the −35 region that differs in three of the four sequences. The asterisks indicate the transcriptional start sites as determined by primer extension analysis of K. pneumoniae 44NR and 15571.

In order to further explore this possibility, we first sought to confirm that the transcriptional start sites for the bla_SHV-1 message for K. pneumoniae 15571 and K. pneumoniae 44NR were the same and that these start sites were identical to those previously reported (16). Primer extension analysis confirmed that the transcriptional start site was identical to that previously reported. We next performed site-directed mutagenesis of the −10 region of the bla_SHV-1 gene in recombinant plasmid pCWR366, changing the A in the second position to a C. The mutant plasmid was transformed into E. coli DH10B, after which ampicillin MICs were determined and Northern hybridizations were performed. The A→C change was associated with a marked decrease in the ampicillin MIC from >16,000 to 2,000 μg/ml. This decrease in the MIC was associated with a decrease in the amount of bla_SHV-1 mRNA detectable by Northern hybridization (data not shown). These data suggest that the nucleotide located at the second position of the −10 region of the bla_SHV-1 promoter influences the expression of the enzyme and the ultimate level of ampicillin resistance.

DISCUSSION

Levels of antibiotic resistance expressed by β-lactamase-producing bacteria can be affected by several different factors, including the affinity of the enzyme for the antibiotic, the amount of β-lactamase produced, and the ease with which the antibiotic gains access to the periplasmic space. In this report, we describe a K. pneumoniae strain that expresses clinically important levels of resistance to both ceftazidime and the β-lactam–β-lactamase inhibitor combinations. Our data suggest this phenotype results from a combination of factors, most prominently increased production of SHV-1 enzyme and the reduction in a minor outer membrane protein that probably contributes to the entry of ceftazidime (and perhaps piperacillin and tazobactam) into the periplasmic space.

That increased production of the SHV-1 enzyme is not the sole explanation for ceftazidime resistance in K. pneumoniae 15571 is evidenced by the fact that the 15571 bla_SHV-1 gene, when cloned into a high-copy-number vector in E. coli, expresses levels of ceftazidime resistance of only 2 μg/ml. This level of resistance is expressed despite the fact that, by Northern analysis, E. coli DH10B(pCWR366) produces dramatically more bla_SHV-1 message than does 15571 (data not shown).

The loss of outer membrane proteins in clinical strains of K. pneumoniae has been repeatedly observed (3, 5, 12). Previously published analyses of the outer membrane proteins of K. pneumoniae have shown that the major proteins are the outer membrane doublet at ca. 35 to 40 kDa. A recent paper identified a third porin, OMPK37, which appears to be involved in resistance to β-lactam antibiotics (7). Based on size comparison, the diminished outer membrane protein seen in K. pneumoniae 15571 was not any of these other porins. Our gel demonstrates the loss of a minor protein also identified by others that migrates more slowly than the doublet (9). The presence of porin reduction and ceftazidime resistance has also been described for E. coli by Weber et al. (26).

One recently published study suggests that most K. pneumoniae clinical isolates produce a β-lactamase similar to SHV-1 and that in most cases these enzymes are chromosomally encoded (11). In the majority of cases, the level of enzyme production appears to be low. Our data suggest that this low level of expression is due to the activity of the promoter upstream of the gene. The observation that lower-level expression of β-lactamase in K. pneumoniae 44NR is due to small differences between the promoter sequences in the two isolates was confirmed by site-directed mutagenesis. We changed A to C in the second position of the −10 sequence and noted a marked lowering of the ampicillin MIC. Our data do not allow us to exclude the possibility that a repressor or activator interacts with the promoter sequence to inhibit or enhance transcription. Further studies are planned to specifically investigate the possibility of repressor activity, as well as to sequence promoter regions from a variety of clinical K. pneumoniae strains in an effort to correlate, if possible, the content of promoter sequences with the level of ampicillin resistance on a broader scale.