INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

Metallo--lactamases (class B of the molecular classification of Ambler [1] or group 3 according to the functional classification of Bush et al. [6]) constitute a very heterogeneous family. Although their primary structures exhibit low degrees of sequence isology (38) (generally less than 43%), their three-dimensional structures show high degrees of similarity (7, 8, 9, 12, 35). All class B -lactamases share five main characteristics (38): (i) they hydrolyze carbapenem compounds, (ii) they do not interact with monobactams, (iii) they are inhibited by chelating agents such as dipicolinic acid and 1,10-o-phenanthroline, (iv) they contain zinc ions as the naturally occurring cation, and (v) they exhibit two metal-binding sites.

On the basis of structural analyses, these enzymes cluster into three different groups: subclass B1 contains most known zinc -lactamases (for example, BcII from Bacillus cereus 569H [18], CcrA [also named CfiA] from Bacteroides fragilis [29], and the plasmid-encoded enzyme IMP-1 found in some isolates of Pseudomonas aeruginosa, Serratia marcescens, and other gram-negative bacteria [19, 24]), subclass B2 includes the Aeromonas enzymes (CphA [21], ImiS [37], and CphA2 [28]), and subclass B3 contains the tetrameric L1 enzyme produced by Stenotrophomonas maltophilia (33, 36) and the Chryseobacterium meningosepticum GOB-1 (2) and the Legionella (Fluoribacter) gormanii FEZ-1 (4) metallo--lactamases.

In the known crystal structures of metallo--lactamases, Zn-1 is tetrahedrally coordinated to three histidines and a water molecule. When present, Zn-2 interacts with three residues (a cysteine, an aspartic acid, and a histidine in the case of BcII, IMP-1, and CcrA or an aspartic acid and two histidines for L1) and two water molecules in a trigonal pyramid. In the di-zinc form, one water molecule is bridged between the metal ions.

The enzymes of subclass B1 are monomeric proteins. They possess a broad-spectrum activity profile (10, 13, 14, 20, 26, 40) and are inhibited by thiol compounds such as SB25566 (9, 16, 34). Interestingly, the mono- and di-zinc form of the BcII and CcrA enzymes are nearly equally active. Kinetic and spectroscopic studies indicated that for both forms a transient noncovalent intermediate is formed during the hydrolysis of the substrate.

To date, no structure of a subclass B2 enzyme is available. For these -lactamases, the optimal activity is observed with the mono-zinc form. The second zinc ion behaves as a noncompetitive inhibitor (17). Only carbapenems are efficiently hydrolyzed by these enzymes (13), while all other -lactams are poor substrates. In addition, cephamycins and oxacephems behave as poor inactivators of CphA (14, 27).

The enzymes belonging to subclass B3 can be either monomeric (GOB-1) or multimeric (L1). Detailed kinetic studies performed on the L1 and GOB-1 metallo--lactamases showed that the enzymes exhibit broad-spectrum activity profiles (2, 10). In FEZ-1, all the residues which interact with the zinc ions in L1 are conserved (4), while in GOB-1 the first histidine is replaced by a glutamine residue (2). Interestingly, preliminary biochemical studies of the L. gormanii FEZ-1 metallo--lactamase reveal a broad-spectrum activity profile, but with a striking preference for cephalosporin compounds (4, 15).

In the work described here we produced the FEZ-1 metallo--lactamase of L. gormanii ATCC 32197^T in Escherichia coli. The enzyme was purified to homogeneity, and a detailed kinetic study was performed. The effects of various chelating agents, pH, and zinc ion concentration on enzyme activity were tested. In addition, a molecular model of the enzyme structure was built by knowledge-based modeling methods.

(The results described here were presented in part at the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, Ontario, Canada, 17 to 20 September 2000.)

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Antibiotics and other chemicals. Choramphenicol, ampicillin (₂₃₅ = 820 M¹ cm¹), cephalothin (₂₆₀ = 6,500 M¹ cm¹), cephaloridine (₂₆₀ = 100,000 M¹ cm¹), cefoxitin (₂₆₀ = 7,000 M¹ cm¹), cefotaxime (₂₆₀ = 7,500 M¹ cm¹), methicillin (₂₆₀ = 100 M¹ cm¹), carbenicillin (₂₃₅ = 780 M¹ cm¹), cloxacillin (₂₆₀ = +140 M¹ cm¹), EDTA, pyridine-2,6-dicarboxylic acid (dipicolinic acid), and 1,10-o-phenanthroline were purchased from Sigma Chemical Co. (St. Louis, Mo.). Kanamycin was purchased from Merck (Darmstadt, Germany). Isopropyl--D-thiogalactopyranoside (IPTG) was purchased from Eurogentech (Liège, Belgium). Imipenem (₃₀₀ = 9,000 M¹ cm¹) was a gift from Merck Sharp & Dohme Research Laboratories (Rahway, N.J.). Meropenem (₂₉₈ = 6,500 M¹ cm¹) was a gift from ICI Pharmaceuticals (Macclesfield, England). Biapenem (₂₉₄ = 9,960 M¹ cm¹) and piperacillin (₂₃₅ = 820 M¹ cm¹) were gifts from Wyeth Lederle (Tokyo, Japan). Nitrocefin (₄₈₂ = +15,000 M¹ cm¹) was purchased from Unipath Oxoid (Basingstoke, United Kingdom). Benzylpenicillin (₂₃₅ = 775 M¹ cm¹) was a gift from Rhône-Poulenc (Paris, France). Cefuroxime (₂₆₀ = 7,600 M¹ cm¹) and ceftazidime (₂₆₀ = 9,000 M¹ cm¹) were from Glaxo Group Research (Greenford, United Kingdom). Temocillin (₂₃₅ = 660 M¹ cm¹), ticarcillin (₂₃₅ = 660 M¹ cm¹), 6-aminopenicillanic acid (6-APA; ₂₃₅ = 690 M¹ cm¹), and clavulanic acid were gifts from SmithKline Beecham Pharmaceuticals (Brentford, United Kingdom). Cefepime (₂₆₀ = 10,000 M¹ cm¹) and aztreonam (₃₂₀ = 700 M¹ cm¹) were gifts from S.A. Bristol-Meyers Squibb (Brussels, Belgium). Tazobactam (₂₃₅ = 1,970 M¹ cm¹) was a gift from Wyeth-Ayerst Laboratories (West Chester, Pa.). Moxalactam (₂₆₀ = 4,000 M¹ cm¹) and cefpodoxime (₂₆₀ = 10,000 M¹ cm¹) were gifts from Sankyo Pharmaceuticals (Tokyo, Japan).

Bacterial strains and vectors. Plamids pBLL/FEZ-1 and pET24/FEZ-1 have been described previously (4). E. coli XL1-Blue (Stratagene Inc., La Jolla, Calif.) was used as the host for recombinant plasmids during construction of the expression vectors. E. coli BL21(DE3)(pLysS) (Novagen Inc., Madison, Wis.) was used as the host for the T7-based expression vectors in overexpression experiments. Plasmid pCR 2.1 (TA Cloning Kit; Invitrogen BU, NV Leek, The Netherlands) was used to clone the PCR products. The expression vector pET26b(+) (Novagen Inc.) was used for the construction of the T7-based expression vector.

Construction of expression vector and preliminary expression experiments. The putative position of the signal peptidase cleavage site in the FEZ-1 pre--lactamase was calculated with the help of SignalP program (23), available under the server page of the Centre for Biological sequence analysis (http://www.cbs.dtu.dk).

Production and purification of the zinc -lactamase. The FEZ-1 enzyme was produced by E. coli BL21(DE3)(pLysS) carrying pDML1810 in SB medium containing kanamycin and choramphenicol as the selecting agents during growth of the bacteria at 28°C under orbital shaking. A total of 40 ml of an overnight preculture in SB was used to inoculate 1 liter of fresh SB supplemented as described above with kanamycin and choramphenicol. IPTG (final concentration, 0.5 mM) was added at an absorbance value of 0.6 at 600 nm, and the culture was continued for 6 h. Cells were harvested by centrifugation (5,000 × g for 10 min at 4°C), and the pellet was resuspended in 100 ml of 30 mM sodium cacodylate buffer (pH 6.0; buffer A). The bacteria were disrupted with the help of cell disrupter equipment (Basic Z model; Constant Systems Ltd., Warwick, United Kingdom). Cell debris was removed by centrifugation (30,000 × g for 45 min at 4°C), and the supernatant was dialyzed overnight against buffer A at 4°C. Thereafter, the crude extract was loaded onto an S-Sepharose FF column (2.6 by 34 cm; Pharmacia, Uppsala, Sweden) equilibrated in buffer A. The column was washed with buffer A until the A₂₈₀ of the effluent was <0.1, and the enzyme was eluted with a linear salt gradient (0 to 0.5 M) in five column volumes. The active fractions were collected and concentrated on a YM-10 membrane (Amicon, Beverly, Mass.) to a final volume of 5 ml. The sample was loaded in a molecular sieve Sephacryl-100 (1.5 by 56 cm) column equilibrated in 15 mM sodium cacodylate buffer (pH 6.0) containing 0.2 M NaCl. The fractions that exhibited -lactamase activity were collected and their specific activities were monitored by measuring the rate of hydrolysis of cefuroxime. The purity was also checked by SDS-polyacrylamide gel electrophoresis (PAGE). The fractions that exhibited constant specific activities were pooled (volume, 26 ml) and concentrated to a final concentration of 1 mg/ml. The enzyme preparation was stored at 20°C in 30 mM sodium cacodylate buffer (pH 6.0).

Determination of quaternary structure of native metallo--lactamase. A total of 100 µl of FEZ-1 (1 mg/ml) was loaded onto a molecular sieve (Superdex HR75; 1 by 30 cm; Pharmacia) column equilibrated in 30 mM sodium cacodylate buffer (pH 6.0)-0.2 M NaCl. The sample was eluted with the same buffer at a flow rate of 0.5 ml/min. The column was calibrated with lysozyme (14,400 Da), carbonic anhydrase (31,000 Da), ovalbumin (43,000 Da), and bovine serum albumin (BSA; 66,200 Da) as standard proteins. The volume of the fractions was 1 ml. The retention volumes were measured by monitoring the A₂₈₀, and the retention volumes for the sample containing -lactamase were measured by determining the enzyme activity.

Determination of N-terminal sequence and molecular mass. The N-terminal sequence was determined with the help of a gas-phase sequencer (Prosite 492 protein sequencer; Applied Biosystems, Foster City, Calif.) The M_r of the enzyme was estimated with an electrospray mass spectrometer (VG Bio-Q) upgraded with a Platform source (Micromass, Altrincham, United Kingdom). The samples (100 pmole) were suspended in 0.05% formic acid-50% acetonitrile in water and were injected into the source of the mass spectrometer with a syringe pump (Harvard Instruments, South Natick, Mass.) at a flow rate of 6 µl/min. The capillary was held at 2.7 kV, and the cone voltage was set at 40 V. Fifteen scans covering 600 to 1,500 amu were accumulated for 135 s and processed with the Masslynx software delivered with the instrument. Calibration was performed with horse heart myoglobin.

Determination of kinetic parameters. Hydrolysis of the antibiotics by FEZ-1 was followed by monitoring the variation in the absorbance of the -lactam solution in 15 mM sodium cacodylate buffer (pH 6.0). All the measurements were made on a Uvikon 940 spectrophotometer connected to a personal computer via an RS232C interface. The reactions were performed in a total volume of 500 µl at 30°C. BSA (20 µg/ml) was added to diluted solutions of -lactamase in order to prevent enzyme denaturation. The steady-state kinetic parameters (K_m and k_cat) were determined by analyzing the complete hydrolysis time courses as described by De Meester et al. (11) or by using the Hanes linearization of the Michaelis-Menten equation.

Inactivation by chelating agents. The loss of -lactamase activity was monitored in the presence of different concentrations of EDTA, dipicolinic acid, and 1,10-o-phenanthroline. The progressive inactivation of the enzyme was monitored by analyzing the complete hydrolysis time course of 100 µM cefuroxime in 15 mM sodium cacodylate (pH 6.0), used as a reporter substrate. The dependence of the pseudo-first-order inactivation rate constant (k_i) upon the chelating agent concentration was determined.

Zn²⁺ dependence of FEZ-1 metallo--lactamase activity. Enzyme activity was measured by monitoring the initial rates of hydrolysis of 100 µM cefuroxime, cefotaxime, imipenem, nitrocefin, benzylpenicillin, and moxalactam in 15 mM sodium cacodylate buffer (pH 6.0) containing 100 µM Zn²⁺. Kinetic parameters were compared with those obtained under the same conditions but without Zn²⁺.

pH dependence of FEZ-1 metallo--lactamase activity. The K_m and k_cat values were calculated for cefuroxime, a good substrate of FEZ-1, in the following buffers: 15 mM sodium acetate (pH 5.0), 15 mM sodium cacodylate (pH 6.0), 15 mM HEPES (pH 7.0), and 15 mM HEPES (pH 8.0).

Determination of Zn²⁺ content of FEZ-1 at different pH values. The pH dependence of the metal content of the metallo--lactamase was measured by inductively coupled mass spectrometry (ICPMS). One milliliter samples of FEZ-1 (0.88 mg/ml) were dialyzed overnight at 4°C against 1 liter of the following buffers: 15 mM sodium acetate (pH 5.0), 15 mM sodium cacodylate (pH 6.0), 15 mM HEPES (pH 7.0), and 15 mM HEPES (pH 8.0). The zinc concentration was determined in the protein sample and the dialyzing buffer by ICPMS, and the Zn/enzyme molar ratio was calculated.

Molecular modeling. The FEZ-1 -lactamase structural model was based on the three-dimensional structure of the L1 enzyme from S. maltophilia (35). The molecular model was built with the Homology module of the Insight program (Molecular Simulations, San Diego, Calif.) running on a Silicon Graphics workstation. After model building, energy minimization was achieved with the Discover module of the same package to avoid bad molecular contacts. Finally, the geometric features were analyzed with the Insight II program. The newly proposed BBL numbering of the class B -lactamases was used (15a).

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Overexpression of FEZ-1 -lactamase in E. coli. The production of FEZ-1 in E. coli BL21(DE3)(pLysS) was tested with three different expression vectors (pDML1809, pDML1810, and pDML1812). pET24/FEZ-1, which was constructed previously (4), was also included in these experiments for comparison.

1

1

View larger version (13K): [in this window] [in a new window]   FIG. 1.   Physical map of the different overproducing constructs. The double-headed arrow delineates the nucleotide sequence of the blaFEZ-1 gene. The nucleotide sequences corresponding to the FEZ-1 signal peptide, the pelB sequence, and the PMPNPFPP peptide are represented by boxes with light dashes, boxes with heavy dashes, and black boxes, respectively. mgP, milligrams of protein.

1

1

Purification of FEZ-1 -lactamase. FEZ-1 overproduced in E. coli BL21(DE3)(pLysS) was purified in two chromatography steps as described in Materials and Methods. In the first purification step, the enzyme was eluted from an S-Sepharose column at a salt concentration of 0.2 M. At that stage, the purification yield of FEZ-1 was about 66% (Table 1) and other proteins with a lower molecular masses were revealed by SDS-PAGE (Fig. 2). To remove these contaminants the enzyme preparation was loaded onto a Sephacryl-100 molecular sieve column equilibrated with 30 mM sodium cacodylate (pH 6.0) containing 0.2 M NaCl. The active fractions were pooled and concentrated on an Amicon YM-10 membrane to a final concentration of 1 mg/ml. Above this concentration, a decrease in the specific activity of the preparation was observed, a phenomenon that may be due to partial aggregation of the -lactamase. The final yield of the purification was 53%. Under these conditions it was not necessary to perform the additional cation-exchange chromatography step described by Fujii et al. (15). The ability to overproduce the zinc -lactamase in E. coli facilitated the purification process and allowed simplification of the protocol. The N-terminal sequence of the purified FEZ-1 enzyme was AYPMPNPFPPF, as expected. Mass spectrometry confirmed the homogeneity of the protein preparation (data not shown). The M_r value was estimated to be 29,447 ± 7, which is very close to that deduced from the amino acid sequence (M_r = 29,440).

View this table: [in this window] [in a new window]   TABLE 1.   Purification of the FEZ-1 metallo--lactamase of L. gormanii produced by a 1-liter culture of E. coli BL21(DE3)(pLysS)/pDML1810

View larger version (42K): [in this window] [in a new window]   FIG. 2.   SDS-PAGE after the different purification steps. Lane A, molecular mass standards; lane B, crude extract of E. coli BL21(DE3)(pLysS)/pDML1810; lane C, protein pattern after passage through an S-Sepharose column; lane D, protein pattern after passage through a Sephacryl-100 molecular sieve. The arrow indicates the Mr of FEZ-1.

The subclass B3 FEZ-1 -lactamase is a monomeric enzyme. Calibration of the Superdex HR75 column yielded the following retention volumes: lysozyme, 17.38 ml; carbonic anhydrase, 11.42 ml; ovalbumin, 10.08 ml; and BSA, 9.1 ml. The retention volume of FEZ-1 was 11.50 ml, close to that of carbonic anhydrase (Fig. 3). Its molecular mass, determined by molecular sieve chromatography on Superdex HR75, was thus approximately 30,000 Da, indicating that the native enzyme is a monomer, in agreement with data previously obtained with crude extracts (4). Therefore, metallo--lactamases of subclass B3 can be either monomeric (FEZ-1, GOB-1) or multimeric (L1), a phenomenon which was not observed for the metallo--lactamases of other subclasses.

View larger version (18K): [in this window] [in a new window]   FIG. 3.   Elution pattern of the FEZ-1 -lactamase and carbonic anhydrase on a Superdex HR75 molecular sieve column. The values on the y axis represent the A280.

Kinetic parameters of the FEZ-1 enzyme. The k_cat and K_m values were determined for a representative set of -lactam antibiotics (Table 2). The results showed that the -lactamase exhibited a broad-spectrum activity profile (Table 2), although with notable differences for different substrates. It did not significantly hydrolyze monobactams such as aztreonam and carumonam (k_cat/K_m <0.0001 µM¹ s¹). The activity of the -lactamase was not affected by a prolonged incubation (1 h at room temperature) of the enzyme in the presence of both monobactams at a concentration of 1 mM each. In addition, the rate of hydrolysis of nitrocefin was not modified by the presence of a high concentration (1 mM) of aztreonam. Among the suicide substrates designed against active-site serine -lactamase tested, clavulanic acid (final concentration, 1 mM) was not recognized and tazobactam behaved as a substrate (Table 2).

View this table: [in this window] [in a new window]   TABLE 2.   Kinetic parameters of the L. gormanii metallo--lactamase produced in E. coli BL21(DE3)(pLysS)

1

1

1

1

1

1

1

1

pH dependence of activity and zinc content of FEZ-1. The pH dependence of the zinc content of the metallo--lactamase of L. gormanii was measured by ICPMS the (Table 5). Other metallic ions such as Co(II), Cd(II), Ca(II), Cu(II), Mn(II), Ni(II), and Fe(II) and Fe(II) or Fe(III) were not found. In the absence of excess metal, between pH 6 and 8, two zinc ions were found per enzyme molecule. A similar zinc content was found for the L1 enzyme (10). No measurements of the zinc/enzyme molar ratios were made with the GOB-1 enzyme (2). The presence of a glutamine residue as a putative metal ligand may influence the zinc content of the latter enzyme. All the metallo--lactamases studied have contained two zinc-binding sites (8, 17, 20, 25). However, with the exception of metallo--lactamases isolated from Aeromonas species and when the mono-zinc species can be prepared, the presence of the second metal ion does not strongly modify the catalytic efficiency of the enzyme (17). The study of the interaction between nitrocefin and the L1 enzyme showed that the second zinc ion might participate in the stabilization of an intermediate (22). The same phenomenon was observed for CcrA (38). In the mono-zinc form, the zinc ion participates in the formation of the nucleophilic hydroxide and provides Lewis acid catalysis by polarization of the carbonyl of the -lactam ring (5). We can hypothesize that the mechanism of the FEZ-1 enzyme is similar to that of L1. Finally, the pH dependence of the steady-state kinetic parameters (k_cat and K_m) of FEZ-1 with cefuroxime in the absence of added zinc was measured. A time-dependent inactivation of the enzyme by cefuroxime was observed at pH 5. The k_i value was 3 × 10³ s¹ with 100 µM cefuroxime. The same phenomenon was observed even when the reaction was performed in the presence of 100 µM Zn²⁺. The K_m and k_cat values were not strongly modified between pH 6 and 8, with maximum activity detected at pH 6 (k_cat = 320 s¹; K_m = 48 µM). These data are in good agreement with the results obtained for other metallo--lactamases. For example, the maximum activities of the B. cereus 569H (4) and IMP-1 (20) enzymes were observed in the same pH range.

View this table: [in this window] [in a new window]   TABLE 5.   Metal/enzyme ratio and catalytic efficiencies of the FEZ-1 zinc -lactamase as a function of pHa

Inactivation by Zn-chelating agents. All the chelating agents tested behaved as strong inactivators. In all cases, a time-dependent k_i could be measured. The k_i values were found to be independent of the chelator concentration and were similar for all three compounds (Table 6). These data indicated that at pH 6 the chelators probably act by scavenging the free metal, and the k_i value might represent the rate of dissociation of the protein-zinc ion complex. Interestingly, and in contrast to all the subclass B1 and B2 enzymes (BlaB [31], IMP-1 [20], BcII [unpublished data], CfiA [unpublished data], and CphA [17]), in which inactivation occurs via the formation of a transient enzyme-metal-chelator ternary complex, a similar behavior is observed with the L1 subclass B3 enzyme (3) and the k_i values are strikingly similar (from 1.7 × 10² to 3.8 × 10² s¹). In addition, the concentration needed to completely inactivate FEZ-1 is far lower than the concentration needed to completely inactivate the other enzymes tested (17, 20). For example, EDTA and dipicolinic acid behaved as poor inactivators of the IMP-1 -lactamase. IMP-1 could be inactivated by behaved the enzyme in the presence of 10 mM EDTA or 300 µM dipicolinic acid (20). It would be interesting to further analyze the mechanisms of action of chelating agents against the other subclass B3 enzymes.

View this table: [in this window] [in a new window]   TABLE 6.   Inactivation of the L. gormanii metallo -lactamase by chelating agents

Structural features of the FEZ-1 enzyme. Molecular modeling of the FEZ-1 enzyme on the basis of the known structure of the L1 metallo--lactamase revealed some interesting features.

(i) Overall fold. As already pointed out by Boschi et al. (4), FEZ-1 and L1 could be aligned over the entire sequence without the introduction of major gaps (Fig. 4). The major differences between L1 and FEZ-1 are the absence of the N-terminal 3₁₀ helix region, a two-residue insertion in the loop between 3 and 8, and a six-residue insertion in the loop between 12 and the C-terminal 5 helix. The latter loop is already considerably longer in L1 than in the subclass B1 enzymes, and it is constrained by an intramolecular disulfide bridge that links the cysteine residues at positions 256 and 290 (35). Such a disulfide bridge between cysteine residues can be predicted in FEZ-1 (Fig. 4).

View larger version (33K): [in this window] [in a new window]   FIG. 4.   Comparison of the amino acid sequence of the mature zinc -lactamase of L. gormanii (FEZ-1) with that of the S. maltophilia zinc -lactamase (L1). The consensus sequence indicates the identical amino acid residues. Residues involved in zinc coordination are in boldface and are marked by asterisks. The residues in boldface and labeled with a plus sign are residues which may interact with the C-7 side chain of cefuroxime. Cysteine residues in boldface are presumably involved in a disulfide bridge in the FEZ-1 metallo--lactamase.

(ii) Zinc ligands. The six residues known to be involved in Zn²⁺ binding in the L1 -lactamase are conserved in the FEZ-1 enzyme, i.e., His-116, His-118, and His-196 for the first Zn²⁺ binding site and Asp-120, His-121, and His-263 for the second one. The orientations of the zinc ligands are maintained by an extensive hydrogen bond network which is also very similar in the two enzymes. However, the group that orients the first His is the Asp-220 side chain in L1 and the Ser-262 side chain via a water molecule in FEZ-1, as in most subclass B1 enzymes.

(iii) Substrate binding site. In the case of subclass B1 -lactamases (7, 8, 9), most investigators postulate that the carbonyl oxygen of the -lactam substrate lies in an oxyanion hole formed by Zn-1 and the side chain of a conserved Asn (Asn-233), while the substrate carboxylate moiety interacts with the ammonium group of a conserved lysine residue (Lys 224). In L1, the side chain of the Asn is 14 Å away from the substrate carbonyl oxygen, and it is proposed that in this case the oxyanion is stabilized by interaction with the side chain of Tyr-228 (35). This tyrosine is conserved and could play the same role in FEZ-1. In turn, the lysine that interacts with the -lactam carboxylate in subclass B1 -lactamases is replaced by Ser-223 in L1 and by a Gly residue in FEZ-1. However, just after this Gly, an Asn residue is inserted in the loop that connects 11 and 4. The side chain of this Asn-233 is ideally placed to interact with the substrate carboxylate, especially those of cephalosporins (Fig. 5). The hydrophobic  substituent of the -lactam substrate generally fits in a hydrophobic pocket formed by the "flap" that connects 3 and 4 and by the loop between 3 and 8, which is considerably longer in subclass B3 enzymes than in subclass B1 enzymes. In this pocket, the hydrophobic residues Phe-156 and Ile-162 in L1 are replaced by a Tyr and a Ser residue in FEZ-1, respectively. These substitutions should influence the substrate specificity, with a facilitated interaction between FEZ-1 and -lactams that bear a less hydrophobic  side chain.

View larger version (32K): [in this window] [in a new window]   FIG. 5.   Active site of the modeled structure of the FEZ-1 enzyme showing the side chains of the amino acids discussed in the text. A molecule of cefuroxime is docked in the catalytic cavity.

Conclusions. The study of the FEZ-1 metallo--lactamase underlines the heterogeneity of the biochemical properties of the class B -lactamases. The enzyme is monomeric and contains two zinc ions. As shown for other metalloproteins, the enzyme was inactivated by Zn chelators.