INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

The production of -lactamases makes a major contribution to -lactam resistance in gram-negative bacteria (13). During therapy with one of the newer -lactams, resistance due to derepression of an inducible chromosomally encoded class C -lactamase (also named AmpC) appears to emerge in 10 to 50% of patients infected with Citrobacter freundii, Enterobacter cloacae, Serratia marcescens, and Pseudomonas aeruginosa (46). The derepressed -lactamase production results from mutations in ampD (16, 22, 26, 30, 51). Plasmid-encoded class C -lactamases have also been described in clinical isolates of Klebsiella pneumoniae, Klebsiella oxytoca, Escherichia coli, and Salmonella senftenberg and are a significant cause of -lactam resistance (8, 10, 13, 27, 29, 37). This increase in the clinical importance of class C -lactamases and their capacity to confer resistance to expanded-spectrum -lactam antibiotics such as cefotaxime and ceftazidime and to -methoxy--lactams such as cefoxitin and cefotetan have led to considerable interest in the understanding of their mechanisms of action.

According to a molecular classification scheme (2, 3, 24, 28), the class C -lactamases, together with class A and D -lactamases, are active-site serine enzymes that catalyze, via a serine-bound acyl intermediate, the hydrolysis of the -lactam to an inactive acid. The class C -lactamases have been placed in the functional group 1 enzymes, which are described as cephalosporinases that are not inhibited by clavulanic acid (13). In addition to the three functional and structural elements (SerXaaXaaLys, TyrXaaAsn, and LysThrGly) that are conserved throughout the class C serine -lactamases (19), the amino acid sequences of these enzymes exhibit more than 36% identity, and several conserved motifs have been identified (52a).

In the work described here, the class C -lactamase from E. cloacae P99, for which the three-dimensional structure has been established by X-ray crystallography at 2-Å resolution (36), was used as a model to study the hydrolysis mechanism of class C enzymes. Amino acid residue Ser-289 is located in a critical region at the top of the E. cloacae -lactamase active site (as viewed in Fig. 2), with its side chain near the C-3 substituent and the C-4 carboxylic acid group of cephalosporins. To elucidate the role of this amino acid residue in the catalytic activities of class C -lactamases, we undertook an investigation of the effects that mutagenesis had on the kinetics of hydrolysis, and we describe our results in relation to the known crystallographic structure of the wild-type enzyme.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Chemicals. Kanamycin, ampicillin, penicillin G, cephalothin, cephaloridine, cefotaxime, cefazolin, cefoperazone, cefaclor, phenylmethylsulfonyl fluoride (PMSF), and DNase I were purchased from Sigma (St. Louis, Mo.). Bovine serum albumin (BSA) was obtained from Pierce (Rockford, Ill.). Ceftazidime was obtained from Glaxo Canada Inc. (Montréal, Québec, Canada). Clavulanic acid, sulbactam, and tazobactam were gifts from SmithKline Beecham Pharma Inc. (Oakville, Ontario, Canada), Pfizer (Groton, Conn.), and Lederle (Carolina, Puerto Rico), respectively. Restriction endonucleases were obtained from New England Biolabs (Mississauga, Ontario, Canada). Nitrocefin was purchased from Oxoid (Nepean, Ontario, Canada).

Bacterial strains and plasmids. E. coli SNO3 (ampA1 ampC8 pyrB recA rpsL) (41) was obtained from Staffan Normark (Karolinska Institutet, Stockholm, Sweden). E. coli CJ236 [dut ung thi relA; pCJ105 (Cm^r)] (Bio-Rad Laboratories Ltd., Mississauga, Ontario, Canada) (33) was used to prepare single-stranded uracil DNA templates. E. coli MV1190 [(lac-proAB) thi supE (sr1-recA)306::Tn10(Tet^r) (F'::traD36 proAB lacI^qZ M15)] (Bio-Rad) (39) was used to eliminate the wild-type uracil DNA template and to produce single-stranded DNA. Plasmid pBGS19⁺ [(Km^r) f1 Ori lacPOZ] was used as a cloning vector (50). Single-stranded DNA production was performed with phage M13K07 (53). All transformed bacteria were grown in tryptic soy broth (TSB; Difco Laboratories, Detroit, Mich.) and on tryptic soy agar plates (Difco) containing appropriate antibiotics when necessary for plasmid selection.

Production of mutant -lactamases. A 1.2-kb DNA fragment containing the ampC gene encoding the class C -lactamase of E. cloacae P99 was amplified by PCR with Vent DNA polymerase (New England Biolabs) and a lysate of this bacterial strain, the latter of which was prepared by the freezing and thawing method (52). The oligonucleotide primers used for amplification (Ecampc3 [5'-CGCGGGATCCACATCCCCTTGACTCGC-3'] and Ecampc4 [5'-CGCGAAGCTTCAATGTTTTACTGTAGC-3']) contained a BamHI restriction site and an HindIII restriction site at their respective 5' ends and were derived from the published E. cloacae P99 ampC sequence (20). The PCR amplification product was cloned into the pBGS19⁺ vector and was identified as pHUL3-4. Molecular biology techniques were performed as described by Sambrook et al. (45). Site-directed mutagenesis was done by the uracil template protocol (32, 33) by using the Muta-Gene Kit (Bio-Rad) as described by Huletsky et al. (23). Six mutant -lactamases were constructed by replacing Ser-289 by Ala, Thr, Gln, Arg, Lys, or Cys (see Table 1). Clones containing mutant genes were selected by DNA sequencing by using the dideoxy method (47) and the T7 Sequencing Kit from Pharmacia Biotech (Baie d'Urfé, Québec, Canada). Mutant genes were sequenced entirely.

Phenotypic characterization of mutant -lactamases. Wild-type and mutant genes were transformed in E. coli SNO3. MICs were determined by a broth dilution method in TSB with 96-well plates (see Table 1). To verify the inhibitory profiles of the mutant enzymes, clavulanic acid, sulbactam, and tazobactam were used in combination with cephaloridine as a substrate (see Table 2).

-Lactamase purification. All -lactamases were purified to >95% homogeneity as follows. The E. coli SNO3 cells harboring the plasmids coding for the wild-type or mutant AmpC -lactamases were grown overnight in TSB at 37°C, collected by centrifugation, and resuspended in TEAA buffer (20 mM triethanolamine, 0.5 M NaCl [pH 7.0]). DNase I (100 µg/ml) and 1 mM PMSF were added to the suspension, and the cells were disrupted by three passages through a French pressure cell (18,000 lb/in²). Insoluble material was removed by centrifugation at 229,000 × g for 1 h at 4°C, and the clarified supernatant was loaded on an immobilized phenylboronate gel (MoBiTec, Göttingen, Germany) equilibrated with TEAA buffer (14). The column was then washed with TEAA buffer and the -lactamases were eluted in borate buffer (0.5 M boric acid, 0.5 M NaCl [pH 7.0]). Fractions containing the -lactamases were pooled and concentrated by ultrafiltration with a Centriplus-10 device (Amicon, Beverly, Mass.) and were dialyzed overnight against 10 mM potassium phosphate buffer (pH 6.4). The enzymes were then placed on an Econo-Pac High S strong cation-exchange cartridge (Bio-Rad) equilibrated in the same buffer. The enzymes were eluted with a gradient of 0 to 300 mM NaCl in the same buffer. Fractions containing -lactamase activity were pooled and concentrated as described above. The protein concentration was measured by the Bradford dye-binding procedure (9) by the Bio-Rad protein assay with BSA as a standard. A total of 750 ng of purified -lactamase was loaded onto a sodium dodecyl sulfate (SDS)-12% polyacrylamide gel. Enzyme homogeneity was demonstrated by the presence of a single band on the silver-stained gel (Bio-Rad). Purified -lactamases were stored at 20°C in 50% glycerol and 1 mg of BSA per ml.

Evaluation of kinetic parameters. Hydrolysis of -lactam antibiotics was determined at 30°C in 50 mM phosphate buffer (pH 7.0) on a Varian Cary 1 spectrophotometer. Dilution of the -lactamases was performed with phosphate buffer solution containing 0.1 mg of BSA per ml and 5% glycerol. Substrate hydrolysis was monitored by determining the loss of absorbance at 260 nm for cephalothin, 270 nm for cefaclor, 275 nm for cefoperazone, and 295 nm for cephaloridine and by determining the increase in absorbance at 320 nm for cefazolin. The changes in extinction coefficients (s) were as follows: for cephalothin, ₂₆₀ = 7,300 M¹ · cm¹; for cefaclor, ₂₇₀ = 6,510 M¹ · cm¹; for cefoperazone, ₂₇₅ = 8,640 M¹ · cm¹; for cephaloridine, ₂₉₅ = 889 M¹ · cm¹; and for cefazolin, ₃₂₀ = 1067 M¹ · cm¹. The steady-state kinetic parameters (K_m, k_cat, and k_cat/K_m) were determined from measurements of the initial rates (at least in triplicate) by fitting the Michaelis-Menten equation with the program Leonora (15).

Molecular structures. Atomic coordinates of the crystallographic structure of the AmpC -lactamase from E. cloacae P99 at a 2-Å resolution (36) are available from the Protein Data Bank (entry no. 2BLT). The crystallographic structure of a covalent seryl-phosphonate complex of the P99 -lactamase (35) (entry no. 1BLS) was also used.

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

Description of mutant -lactamases. Six mutant -lactamases were constructed by site-directed mutagenesis of Ser-289 of the class C -lactamase of E. cloacae P99. DNA sequencing identified recombinant plasmids with changes at position 289. The wild-type and mutant -lactamases studied are summarized in Table 1.

View this table: [in this window] [in a new window]   TABLE 1.   MICs for E. coli SNO3 containing bases encoding wild-type and mutant -lactamases at position 289

Microbiological activity. The susceptibilities of plasmid-transformed E. coli SNO3 strains producing wild-type and mutant -lactamases to the selected group of antibiotics were determined by measuring the MICs. As seen in Table 1, production of the wild-type E. cloacae P99 -lactamase in E. coli SNO3 increased the level of resistance to all -lactams tested. The susceptibilities of the organisms producing the S289A, S289T, or S289Q mutant to all these -lactams was not affected compared to that of the wild type-producing organism. The cells with the S289R mutant also had about the same levels of resistance to all -lactams with the exception of cefoperazone, the MIC of which was decreased fourfold. However, production of the last two mutants, S289K and S289C, in E. coli cells led to increased susceptibilities to many -lactams. For the organism producing the S289K mutant, the MIC of cephaloridine was reduced eightfold and the MICs of cephalothin, cefazolin, and cefoperazone were reduced fourfold. The same level of decrease in the MIC was observed for the organism producing the S289C mutant for the narrow-spectrum -lactams cephaloridine and cephalothin. However, the MICs of cefoperazone and cefazolin were further decreased (16-fold), and a 4-fold increased susceptibility to two other -lactams, cefaclor and penicillin G, was observed. Table 2 presents the MICs of cephaloridine in combination with clavulanic acid, sulbactam, and tazobactam. The wild-type enzyme was not inhibited by clavulanic acid but was well inhibited by sulbactam and tazobactam. The mutant enzymes exhibited the same inhibitory profile as that of wild-type enzyme for the three inhibitors.

View this table: [in this window] [in a new window]   TABLE 2.   MICsa for E. coli SNO3 containing bases encoding wild-type and mutant -lactamases at position 289

Electrophoretic mobilities of purified -lactamases by SDS-polyacrylamide gel electrophoresis. To elucidate the role of Ser-289 in the activity of the E. cloacae P99 -lactamase, the wild-type and mutant enzymes were purified to near homogeneity (95% pure). Figure 1 shows the electrophoretic mobilities of the purified wild-type P99 -lactamase and the six mutants on an SDS-polyacrylamide gel. A single protein band was observed for the seven -lactamases, and this indicates their purity. The molecular mass was estimated to be 39 kDa for the wild-type P99 -lactamase and for most of the mutants. However, two mutants, S289R and S289K, exhibited slightly different electrophoretic mobilities and their calculated molecular mass was 36 kDa. Nevertheless, a mass spectrum study (in collaboration with Steve Withers from the University of British Columbia) with wild-type P99 -lactamase and the S289R mutant confirmed that their molecular masses correspond to their respective amino acid sequences (wild-type P99 -lactamase, 39,238.9 ± 2.8 Da; S289R -lactamase, 39,307.6 ± 2.6 Da).

View larger version (34K): [in this window] [in a new window]   FIG. 1.   Silver-stained SDS-polyacrylamide gel of the purified wild-type and mutant -lactamases from E. cloacae P99. Lane M, low-molecular-mass marker (in kilodaltons).

Kinetic parameters. The steady-state parameters for the wild-type and mutant enzymes for cephaloridine, cephalothin, cefoperazone, cefaclor, and cefazolin were determined (Tables 3 to 7, respectively). The results indicated that the S289A, S289T, S289C, and S289Q mutants exhibit catalytic efficiencies almost similar to that of the wild-type P99 enzyme toward most cephalosporins studied with the exception of that of the mutant with the S289C substitution for cefoperazone, which was decreased by more than fivefold. This ratio resulted from a twofold decrease in k_cat combined with a twofold increase in K_m. The catalytic efficiencies of the S289K and S289R mutants were reduced for most of the cephalosporins studied, especially for cephalothin, cefoperazone, and cephaloridine. The Lys at position 289 decreased by two-, three-, and fivefold the catalytic efficiencies of these three antibiotics, respectively. This effect was mostly due to a decrease in k_cat. With Arg at position 289, the catalytic efficiencies for cephalothin and cefoperazone were decreased three- and fourfold, respectively, and these decreases also resulted from a lower k_cat. For this last mutant, the highest decrease in k_cat/K_m (sevenfold) was associated with cephaloridine and was due to both a lowering of the k_cat and an increase in the K_m.

View this table: [in this window] [in a new window]   TABLE 3.   Kinetic parameters of wild-type and mutant -lactamases for cephaloridine

View this table: [in this window] [in a new window]   TABLE 4.   Kinetic parameters of wild-type and mutant -lactamases for cephalothin

View this table: [in this window] [in a new window]   TABLE 5.   Kinetic parameters of wild-type and mutant -lactamases for cefoperazone

View this table: [in this window] [in a new window]   TABLE 6.   Kinetic parameters of wild-type and mutant -lactamases for cefaclor

View this table: [in this window] [in a new window]   TABLE 7.   Kinetic parameters of wild-type and mutant -lactamases for cefazolin

Structural analysis. Examination of the crystallographic structure of the P99 -lactamase shows that Ser-289 is on a loop near the -lactam binding site between helix H10 and -strand B2g (as shown in Fig. 2). The side chain is exposed and is directed toward the binding site, and it would be able to make contact with the C-3 substituent of a cephalosporin. The carboxylic acid group at C-4 is about 8 Å from the side chain at position 289 and is partly shielded from it by the C-3 substituent. However, in the serine-bound acyl intermediate (Fig. 3), ring opening is known to permit a rotation of the dihydrothiazine ring (34) which would bring the C-4 carboxylic acid group 3 to 4 Å closer to the side chain at position 289. The side chain of a long amino acid such as Lys or Arg could easily form an electrostatic bond with the carboxylic acid group of the acyl intermediate. In the more slowly hydrolyzed cephalosporins such as cefoperazone, the C-3 group may depart from the intermediate, further opening the carboxylic acid group to interaction with the positive side chain (38).

View larger version (60K): [in this window] [in a new window]   FIG. 2.   MOLSCRIPT drawing (31) of the crystallographic structure of the E. cloacae P99 -lactamase (36). Ser-289 is located on an exposed loop above the binding site. The positioning of cefoperazone adjacent to the reactive Ser-64 is based on crystallographic observations of aztreonam (43) and phosphonate (35) intermediates at the binding site.

View larger version (64K): [in this window] [in a new window]   FIG. 3.   A conformation of the ring-opened serine-bound cefoperazone intermediate based on crystallographic structures of the seryl complexes of aztreonam, cefotaxime, and cephalothin (see Materials and Methods). After rotation of dihydrothiazine ring, Arg-289 is able to form a salt linkage to the carboxylic acid group.

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

In this study, the role of Ser-289 in the -lactam hydrolysis mechanism of the class C -lactamase from E. cloacae P99 was investigated by site-directed mutagenesis. Each of the mutants was purified to homogeneity, and precise kinetic analysis was performed.

An alignment of 25 chromosomal and plasmid-mediated class C -lactamases (52a) revealed that most of these enzymes contain a small and hydrophilic residue at position 289 (Ser, Thr, or Asn) (5, 6, 7, 8, 10, 11, 13, 18, 20, 29, 37, 40, 42, 48, 54). In E. cloacae P99 AmpC, Ser-289 lies at the top of the binding site, where it can interact with the C-3 substituent of cephalosporins (Fig. 2 and 3) (36). The native Ser-289 hydroxyl group could donate to or, in some cases, accept hydrogen bonds from the rather large C-3 substituents of most of the substrates listed here. A similar bonding is possible with threonine, and accordingly, little change in catalytic efficiency is seen in the S289T mutant. This hydrogen bonding would also be possible with a cysteine sulfhydryl group, but the bonding would be much weaker than that of serine or threonine, and therefore, greater changes in catalytic efficiency are seen with the S289C mutant. This consideration of hydrogen bonding fails to explain why the mutant with the alanine mutation exhibits little change in catalytic efficiency. It is worth mentioning that in seven known class C -lactamases, an apolar Ala or Pro is also found at position 289 (6, 8, 13, 37, 40, 54). Therefore, hydrogen bonding between residue 289 and the C-3 substituent of cephalosporins or with the acyl intermediate would not be essential in the catalytic mechanism of class C -lactamases.

When examining changes in the kinetic parameters of the substrates, one generally considers K_m to reflect molecular interactions in the Michaelis complex and k_cat to reflect interactions during the formation and breakdown of the serine-bound acyl intermediate. The ratio of the two, the catalytic efficiency, incorporates both contributions to the overall kinetics. In earlier kinetic studies of the wild-type P99 enzyme, Mazzella and Pratt (38) found that deacylation is the rate-determining step and therefore that k_cat and k_cat/K_m describe the deacylation and acylation steps, respectively. Three types of mutants with mutations at position 289, cysteine, lysine, and arginine, consistently show decreases in k_cat/K_m for all the cephalosporins studied. For example, cefoperazone hydrolysis by these three mutants exhibited two- to threefold decreases in k_cat values, perhaps meaning that interactions of the side chain at position 289 with the acyl intermediate are more important than precatalytic Michaelis interactions.

Given the somewhat greater effects seen in the mutants with lysine and arginine mutations, the positive charge of the side chain must be important. Furthermore, the longer lengths of these basic side chains make possible a strong electrostatic linkage (salt bridge) to the C-4 carboxylic acid group of the dihydrothiazine ring. This linkage is more likely to occur in the acyl intermediate than in the Michaelis complex for two reasons. First, the rotation of the dihydrothiazine ring after acylation (34) moves the carboxylic acid group toward the side chain at position 289. Second, in slowly hydrolyzed broad-spectrum cephalosporins the likely elimination of the C-3 substituent (38) will permit an even more direct interaction with the carboxylic acid group (Fig. 3). A strong linkage could cause a reduction in k_cat either by hindering the approach of the hydrolytic water molecule to the seryl ester bond (12) or by slowing the release of the hydrolysis product from the binding site.

Clavulanic acid (44), sulbactam (17), and tazobactam (4, 21) are commonly used inhibitors which are very effective against most bacterial strains that produce class A -lactamases. The inability of clavulanic acid to inhibit the class C -lactamase (Table 2) has been discussed in relation to the architecture of the binding site and its differences with respect to class A enzymes (36). On the basis of the earlier modeling of clavulanate binding to the wild-type P99 -lactamase (36), the relative insensitivities of all the mutants is possibly due to the small size of clavulanate and the absence of strong interactions with the side chain at position 289. The different behaviors of both sulbactam and tazobactam relative to that of clavulanate toward class A -lactamases have been analyzed (25) and may apply here as well.

These results suggest that Ser-289 is not essential in the binding or hydrolysis mechanism of AmpC -lactamase and that many amino acid substitutions are possible at this position. However, changes for positively charged amino acids reduce catalytic efficiency and have not been selected during evolution. These findings are supported by the study of Siemers et al. (49), who showed, by using random mutagenesis in the region from positions 286 to 289 of the AmpC -lactamase from E. cloacae P99, that Ser-289 was not strictly conserved in their active mutants and that two of the four inactive mutants identified in their library contained Arg or Lys at position 289.