INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The most frequent mechanisms of resistance to extended-spectrum cephalosporins in Pseudomonas aeruginosa are derepression of the chromosomal AmpC -lactamase and increased efflux (1). Extended-spectrum -lactamases (ESBLs) are a greater problem in members of the family Enterobacteriaceae (19). Only one TEM-related ESBL has so far been described from P. aeruginosa (24); nevertheless, several other potent plasmid-mediated -lactamases have been described from the species. These include the IMP-1 zinc -lactamase (34), PER-1, which is a class A ESBL (25) that is widespread in Turkey (5, 26, 31); and several extended-spectrum mutants of class D -lactamases (4, 6, 10). The last group includes the OXA-11, -14, and -16 derivatives of OXA-10 -lactamase (4, 8, 10), the OXA-15 derivative of OXA-2 enzyme (6), and the OXA-18 enzyme, which is closely related to the OXA-12 and AmpS chromosomal enzymes of Aeromonas sobria (27, 29, 33).

In the present study we describe a further ESBL mutant of the OXA-10 enzyme, obtained from two P. aeruginosa isolates collected in Turkey.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. P. aeruginosa isolates 871 and 873 were isolated in February and March 1992, respectively, from patients treated for burns at Hacettepe University Hospital, Ankara, Turkey. They were believed to be isolates of the same strain, since they had identical enzyme DNA restriction patterns (5). Previous work showed that they produced PER-1 -lactamase together with -lactamases that had pIs of 6.1 and 7.7 and that they carried a gene which hybridized with a probe to bla_OXA-10 (5).

View this table: [in this window] [in a new window]   TABLE 1.   MICs of -lactams for P. aeruginosa isolates 871 and 873

Antibiotics and susceptibility tests. MICs were determined on DST agar (Unipath, Basingstoke, Hampshire, United Kingdom) with inocula of 10⁴ CFU per spot. Antibiotics were supplied by the following manufacturers: aztreonam and cefepime, Bristol Myers Squibb, Syracuse, N.Y.; cefsulodin, Novartis, Basel, Switzerland; ceftazidime, Glaxo-Wellcome, Stevenage, Hertfordshire, United Kingdom; ciprofloxacin, Bayer, Newbury, Berkshire, United Kingdom; piperacillin sodium and tazobactam, Wyeth, Taplow, Berkshire, United Kingdom; cephalothin and moxalactam, Lilly, Basingstoke, Hampshire, United Kingdom; imipenem, Merck Sharp and Dohme, Hoddesdon, Hertfordshire, United Kingdom; ceftriaxone, Roche, Welwyn Garden City, Hertfordshire, United Kingdom; cefotaxime and cefpirome, Roussel, Uxbridge, Middlesex, United Kingdom; benzylpenicillin, cephaloridine, cloxacillin, oxacillin, gentamicin, and rifampin, Sigma, St. Louis, Mo.; ampicillin sodium, carbenicillin disodium, and clavulanate lithium, SmithKline Beecham, Brentford, Middlesex, United Kingdom; and meropenem, Zeneca, Macclesfield, Cheshire, United Kingdom.

Plasmid transfer to P. aeruginosa PU21. Logarithmic-phase cells of P. aeruginosa 871 and 873 were mated overnight with P. aeruginosa PU21cip^r on DST agar, as described previously (7, 9). Transconjugants were selected on the same medium containing ciprofloxacin (30 µg/ml) plus either ceftazidime (25 or 50 µg/ml) or gentamicin (30 µg/ml).

Detection of -lactamases and their genes. -Lactamases were characterized by isoelectric focusing of ultrasonic extracts prepared from overnight nutrient agar cultures (22). For probing, which was undertaken as described previously (4, 10) total DNA was extracted and then digested with BamHI (when probing for bla_OXA-10) or HincII (when probing for bla_OXA-2) (Promega, Madison, Calif.). The probe for bla_OXA-10 was made by PCR amplification of a DNA fragment from pMLH51 with primers ABD1 and ABD4 (Fig. 1). The probe for bla_OXA-2 was made similarly from plasmid R46, with primers AH1 (GGAAAAGTCCATGGCAATCCGAATCTT) and AH2 (TTATCGCGCAGCGT-CCGAGTT [complementary strand]), corresponding to coordinates 118 to 144 and 955 to 935, respectively, in the sequence published by Danel et al. (6). The amplified fragments were labeled with digoxigenin with a DIG DNA labeling and detection kit (Boehringer, Lewes, East Sussex, United Kingdom).

View larger version (27K): [in this window] [in a new window]   FIG. 1.   Nucleotide and protein sequences of the coding region of OXA-17 -lactamase in comparison to those of OXA-10, -11, -14, and -16 enzymes. The nucleotide sequence shown in full corresponds to that determined in this study and was allocated the GenBank accession no. AF060206. Nucleotide differences in other blaOXA-10 family genes are marked above the blaOXA-17 sequence, and the deduced amino acid changes are indicated below. The signal peptide extends from amino acid residues 1 to 20, and the proposed cleavage site is shown by a vertical line. The underlined nucleotide sequences represent the primers ABD1, ABD2, ABD3, and ABD4 used for PCR amplification or sequencing. The sequences corresponding to the amplification primers have not been independently determined for OXA-17 and are shown in italics. The three conserved elements previously described in class D -lactamase are in boldface (14, 17).

Cloning -lactamase genes. Total DNA was extracted from the P. aeruginosa isolates 871 and 873 and transconjugant PU21(pMLH51) by a guanidium thiocyanate method (28). Two-microgram amounts of DNA from the test strain and of the vector pBC SK+ were then digested separately with restriction enzymes as follows: BamHI, HindIII, or PstI (Promega) for DNA from isolates 871 and 873 or BamHI only for DNA from strain PU21(pMLH51). The digestion conditions were as advised by the supplier of the enzymes. After digestion, the DNA was ligated and transformed into E. coli XL1-blue by electroporation, as described previously (6). Transformants were selected on Luria-Bertani agar (30) containing ampicillin (10 µg/ml).

Sequencing the -lactamase gene. The OXA-10-related genes from isolates 871 and 873 were amplified by PCR with ABD1 and 5' biotin-labeled ABD4 (Fig. 1) as primers. The products were sequenced by chain termination with ABD1, ABD2, and ABD3 primers, exactly as described by Hall et al. (10). The OXA-2-related gene was amplified with AH1 and 5' biotin-labeled AH2 and then sequenced with primers AH1 (see above); AH3, GCCTGCATCGACATTCAAGA (coordinates 434 to 453 in the sequence published by Danel et al. [6]); AH4, GCTCGGCGCTATTTGAAGAA (536 to 555); and AH5, TCGCAACTGGATACTGCGTG (833 to 851).

-Lactamase purification. An E. coli XL1-blue transformant with the cloned OXA-17 -lactamase was grown overnight, with shaking, at 37°C in 0.6 liters of Luria-Bertani broth and then diluted into 12 liters of fresh, warm, identical medium. After 5 h of incubation at 37°C, the bacteria were harvested by centrifugation at 5,000 × g for 15 min at 37°C and washed once in 20 mM triethanolamine buffer, pH 7.6. The pellet was resuspended in the same buffer and frozen and thawed six times. Further purification was undertaken as for OXA-16 enzyme, with one anion- and one cation-exchange chromatographic fractionation (8). OXA-10 -lactamase was purified from P. aeruginosa PU21(pMLH51) as described previously (7, 8). -Lactamase purity was estimated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis on the gel system of Lugtenberg et al. (21). Protein concentrations were determined by the micro-bicinchoninic acid method (Pierce, Rockford, Ill.).

-Lactamase spectrophotometric assays. The hydrolytic activities of -lactamases were assayed by UV spectrophotometry at 37°C in phosphate buffer at pH 7.0 (7, 10). The wavelengths were as listed by Danel et al. (4, 7), and kinetic parameters were determined from initial rate data, using the Enzfitter program (18).

Nucleotide sequence accession number. The GenBank database accession number for the sequence described in Fig. 1 is AF060206.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Susceptibilities and -lactamases of P. aeruginosa isolates 871 and 873. Extracts of isolates 871 and 873 yielded -lactamases that focused at pIs 5.3, 6.1, and 7.9. The enzyme with a pI of 5.3 corresponded to PER-1, as shown previously (5); the other two enzymes were characterized in this study.

Restriction patterns of DNA digests and hybridization with gene probes for OXA-10 and -2 -lactamases. DNA from isolates 871 and 873 hybridized with both the bla_OXA-2 and bla_OXA-10 probes, whereas DNA from PU21(pMG40) hybridized with only the bla_OXA-2 probe and DNA from PU21(pMLH51) bound only the bla_OXA-10 probe.

Sequencing bla_OXA-10 and bla_OXA-2 genes from isolates 871 and 873. The bla_OXA-10-related genes were amplified from both isolates by PCR, and base sequences corresponding to positions 179 to 912 of bla_OXA-10 (11) were determined (see Fig. 1). This allowed prediction of the protein sequence except for the first 16 amino acids, which are part of the signal peptide, and the last 5 residues. Both isolates had the same base substitution compared with bla_OXA-10, with guanine replacing adenine at position 192, predicting a change of asparagine to serine at position 73 of the protein. This new -lactamase was named OXA-17.

Cloning of bla_OXA-17 into E. coli XL1-blue. The presence of three -lactamases in isolates 871 and 873 complicated purification of the OXA-17 enzyme and determination of its contribution to resistance. Transconjugation experiments did not achieve transfer of any of the -lactamase genes, and cloning into E. coli XL1-blue was adopted to separate OXA-17 from the other -lactamases. Transformants resistant to ampicillin at 10 µg/ml were obtained from both isolates, and crude -lactamase extracts were prepared from representatives containing recombinant plasmids with inserts of different sizes. These extracts were subjected to isoelectric focusing (Table 2). Transformants with only the -lactamase that focused at pI 6.1 hybridized with the probe to bla_OXA-10 but not with that to bla_OXA-2; those with the -lactamase that focused at pI 7.9 showed the reverse hybridization pattern (Table 2). Transformants with each of these patterns were observed from each of the isolates 871 and 873. No producer of PER-1 -lactamase (pI 5.3) was detected by isoelectric focusing from any of the libraries.

Susceptibilities of the E. coli transformants with OXA-17 and -2 -lactamases. Susceptibility tests were performed on representative E. coli XL1-blue transformants with different inserts encoding OXA-17 and -2 -lactamases (Table 3). Transformants with OXA-17 -lactamase (pI 6.1) from the BamHI, PstI, and HindIII libraries possessed similar resistance profiles. Both the cloned OXA-17 and OXA-10 -lactamases increased the MICs of amino and carboxy penicillins and cefoperazone for E. coli XL1-blue by 64- to 128-fold and raised those of aztreonam and ceftriaxone by 8- to 16-fold. In addition, production of OXA-17 enzyme, but not OXA-10, increased the MICs of cefotaxime, cefsulodin, ceftazidime, latamoxef, and cefepime by at least fourfold. Neither the OXA-10 nor the OXA-17 enzyme protected against carbapenems.

Purification of OXA-17 -lactamase. An E. coli XL1-blue MRF' transformant from the HindIII library of P. aeruginosa 871 was used as a source of enzyme for purification to preclude contamination by any of the other -lactamases produced by the original P. aeruginosa isolates. OXA-10 -lactamase was purified from P. aeruginosa PU21(pMLH51). For both enzymes, the final purity exceeded 99%; 12 liters of culture yielded 4.4 mg of pure OXA-17 -lactamase and 3.75 mg of OXA-10 protein.

Kinetic parameters of the OXA-17 -lactamase. Unlike other OXA-10 derivatives (4, 8, 10), OXA-17 -lactamase gave linear kinetics for all of the substrates tested. The lowest K_m was for penicillin G (34 µM), whereas values between 153 and 300 µM were seen for oxacillin, ampicillin, carbenicillin, and cephalothin and values of 500 to 600 µM were seen for cloxacillin and ceftriaxone. The highest K_ms (>2,000 µM) were for cefotaxime and cephaloridine (Table 4). The k_cat exceeded 100 s¹ only for oxacillin; the values for all of the other -lactams were less than 30 s¹. In term of K_cat/K_m, the best substrate was oxacillin, followed by penicillin G and ampicillin, whereas k_cat/K_m, values for other substrates were at least 20-fold lower than for oxacillin. Hydrolysis of ceftazidime and carbapenems was not detected. For comparison, Table 4 also shows kinetic data for OXA-10 -lactamase, which had biphasic kinetics for many substrates. OXA-17 generally had higher K_m values than OXA-10 for cephalosporins and, except for cefotaxime, had lower k_cat rates for both penicillins and cephalosporins.

View this table: [in this window] [in a new window]   TABLE 4.   Kinetic parameters for OXA-17 -lactamase compared with OXA-10

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Isolates 871 and 873 were previously found to have PER-1 -lactamase together with two further -lactamases, one with a pI of 6.1 and the other with a pI of 7.9 (5). DNA from each isolate hybridized with a probe to bla_OXA-10, and the enzyme with a pI of 6.1 was surmised to be related to OXA-10. The -lactamase with a pI of 7.9 type was wrongly proposed to be an AmpC type (5). It appeared likely that one or both of these enzymes contributed to the resistance phenotype of the isolates, since their ceftazidime resistances were only partly reversed by clavulanate (Table 1) whereas strains with PER-1 -lactamase alone became highly sensitive to ceftazidime in the presence of this inhibitor.

The present study confirmed that the two isolates had an OXA-10-related enzyme, and sequencing revealed this to be a new variant, OXA-17, with serine replacing asparagine at position 73. In addition, the studies revealed that the enzyme with a pI of 7.9 was OXA-2, not AmpC. OXA-17 -lactamase, like OXA-10, increased the MICs of ampicillin, piperacillin, carbenicillin, ceftriaxone, cefoperazone, and aztreonam. Additionally, and unlike OXA-10, OXA-17 conferred protection against cefotaxime, latamoxef, cefsulodin, cefepime, and, marginally, ceftazidime. The small effect on the MIC of ceftazidime was in contrast to those of the previously described ESBL mutants of OXA-10, namely, OXA-11, -14, and -16, all of which give high-level ceftazidime resistance (MIC > 128 µg/ml). Unlike OXA-17, these enzymes all have glycine replaced by aspartate at position 157, and this may be critical to ceftazidime resistance. The only previous OXA-10 relative to have serine at position 73 was OXA-13, which gave no resistance to cefotaxime or ceftazidime in P. aeruginosa although it did reduce the susceptibility of E. coli, especially to cefotaxime (23). However, comparison is complicated by the fact that OXA-13 has nine other amino acid differences from OXA-10, although two of these lie in the signal peptide. However, seven of the nine differences in the mature OXA-13 protein, but not the serine at position 73, are present in OXA-7, which is not an ESBL.

The mutation in OXA-17 lies close to the first conserved element of class D enzymes (15, 17), which contains the active serine. This is in contrast to the mutations in previous OXA-10 ESBLs, which lie at positions 124, 143, and 157, and to the mutations that give ESBL activity in TEM and SHV enzymes, which affect critical residues that fold towards the active serine but are remote from it in the primary sequence (16).

OXA-17 was more active than OXA-10 against oxacillin, cloxacillin, and cefotaxime and had a higher k_cat for cefotaxime. Once purified, however, OXA-17 was less efficient (lower k_cat/K_m than OXA-10 against all the -lactams testedeven those to which it gave greater resistance than OXA-10. This reduced efficiency was largely because the k_cat values for OXA-17 were low, especially for penicillins and cephaloridine. A low k_cat can arise as an artifact if an enzyme is impure, but this seems unlikely because only a single band was detected by sodium dodecyl sulfate-polyacrylamide gel electrophoresis; furthermore, the elution peak from the final chromatography purification was symmetrical with its height (not shown). It may be that some of the purified enzyme was not active. In this context it should be noted that OXA-17 -lactamase, unlike other OXA-10 derivatives (4, 8), was purified from E. coli and not P. aeruginosa. Maybe much of the enzyme is not correctly folded in E. coli. This might also explain why cloned OXA-10 family -lactamases tend not to be stable in E. coli (unpublished observations). Alternatively, the OXA-17 enzyme may have been partially denatured during purification or may genuinely be less efficient than the other members of the family in hydrolyzing -lactams, though with equally good substrate binding.

The emergence in P. aeruginosa of an OXA-10 variant that gives greater resistance to cefotaxime than to other oxyimino-aminothiazolyl cephalosporins may seem surprising, because cefotaxime is not therapeutically useful against P. aeruginosa infections. It may be that the mutation emerged by chance, or it may have been selected by low-level ceftazidime exposure or the enzyme may have transferred from another species. The absence of transmissibility from isolates 871 and 873 does not preclude the last possibility: plasmids from other species often transfer into P. aeruginosa but then cannot replicate and become chromosomally associated (13). The contributions of PER-1, OXA-2, and OXA-17 enzymes to the resistance of the original P. aeruginosa isolates is unclear, but each enzyme appeared able to contribute to give protection against ceftazidime (Table 4). Finally, it should be added that the evolution of P. aeruginosa strains with such complex arrays of potent -lactamases is striking: the species has not, historically, been a major host for single secondary -lactamases, which have remained much rarer than in members of the family Enterobacteriaceae (19).