4Metallo-␤-lactamases constitute the molecular class B of Ambler (1) and group 3 according to the Bush-Jacoby-Medeiros functional classification (6). In recent years, many new enzymes of this class have been described and the sequences of the corresponding genes have been determined. Their clinical importance is highlighted by the fact that they hydrolyze carbapenems, compounds which most often escape the activity of active-site serine ␤-lactamases. Moreover, most metallo-␤-lactamases are broad-spectrum enzymes which also hydrolyze a variety of penicillins and cephalosporins (13,21,22,26). On the basis of the known sequences, three different lineages, identified as subclasses B1, B2, and B3, can be characterized. Subclass B1 contains most known metallo-␤-lactamases, including the ␤-lactamase II (BcII) proteins from Bacillus cereus or other Bacillus spp. (15,16,19) (27,32), the GOB proteins from C. meningosepticum (2), the FEZ-1 enzyme from Legionella gormanii (5), and the THIN-B ␤-lactamase produced by Janthinobacterium lividum (25a).The three-dimensional structures of several B1 (BcII [7,9,12], CcrA [8,10], and IMP-1 [11]) enzymes and one B3 (L1 [31]) enzyme have been solved by X-ray crystallography. Despite a very low degree of sequence similarity between the two subclasses, the general structures and the relative positions of the secondary structure elements are similar. Surprisingly, the L1 enzyme is a tetramer (4, 31), whereas the B1, B2, and other B3 ( There are, however, no doubts that the proteins are homologous and the sequences of representatives of the three subclasses can be easily aligned. Indeed, in addition to the expected differences at the N and C termini, several insertions and deletions are necessary to allow the alignment of the few conserved residues acting, for instance, as ligands of the two zinc ions which can bind at the active site. Thus, homologous residues from the different class B sequences which are known to play a relevant role in the structure and function often differ in their numbering, even within each subclass.In order to facilitate the comparative analysis of the structures and of the catalytic mechanisms, we would like to propose a standard numbering scheme for the class B ␤-lactamases, the BBL numbering, by analogy with the ABL numbering which has been widely accepted for class A ␤-lactamases. For the class B enzymes, the task was complicated by insertions and deletions and by the generally low degree of similarity but facilitated by the availability of some three-dimensional structures, which allowed the identification of homologous secondary structure elements, even when the sequence similarity was not obvious. Figure 1 shows the proposed alignment and the derived numbering. The observed (B1 and B3) and expected (B2) secondary structure elements are indicated.The following comments can be made. (i) Not all the known sequences are shown. When variants of an enzyme are known and the amino acid alignment exhibits more than 80% sequence identity, only the first described seq...

beta-Lactamases are the main cause of bacterial resistance to penicillins, cephalosporins and related beta-lactam compounds. These enzymes inactivate the antibiotics by hydrolysing the amide bond of the beta-lactam ring. Class A beta-lactamases are the most widespread enzymes and are responsible for numerous failures in the treatment of infectious diseases. The introduction of new beta-lactam compounds, which are meant to be 'beta-lactamase-stable' or beta-lactamase inhibitors, is thus continuously challenged either by point mutations in the ubiquitous TEM and SHV plasmid-borne beta-lactamase genes or by the acquisition of new genes coding for beta-lactamases with different catalytic properties. On the basis of the X-ray crystallography structures of several class A beta-lactamases, including that of the clinically relevant TEM-1 enzyme, it has become possible to analyse how particular structural changes in the enzyme structures might modify their catalytic properties. However, despite the many available kinetic, structural and mutagenesis data, the factors explaining the diversity of the specificity profiles of class A beta-lactamases and their amazing catalytic efficiency have not been thoroughly elucidated. The detailed understanding of these phenomena constitutes the cornerstone for the design of future generations of antibiotics.

Optimization by energy minimization of stable complexes occurring along the pathway of hydrolysis of benzylpenicillin and cephalosporin C by the Streptomyces albus G beta-lactamase has highlighted a proton shuttle that may explain the catalytic mechanism of the beta-lactamases of class A. Five residues, S70, S130, N132, T235 and A237, are involved in ligand binding. The gamma-OH group of T235 and, in the case of benzylpenicillin, the gamma-OH group of S130 interact with the carboxylate group, on one side of the ligand molecule. The side-chain NH2 group of N132 and the carbonyl backbone of A237 interact with the exocyclic CONH amide bond, on the other side of the ligand. The backbone NH groups of S70 and A237 polarize the carbonyl group of the scissile beta-lactam amide bond. Four residues, S70, K73, S130 and E166, and two water molecules, W1 and W2, perform hydrolysis of the bound beta-lactam compound. E166, via W1, abstracts the proton from the gamma-OH group of S70. While losing its proton, the O-gamma atom of S70 attacks the carbonyl carbon atom of the beta-lactam ring and, concomitantly, the proton is delivered back to the adjacent nitrogen atom via W2, K73 and S130, thus achieving formation of the acyl-enzyme. Subsequently, E166 abstracts a proton from W1. While losing its proton, W1 attacks the carbonyl carbon atom of the S70 ester-linked acyl-enzyme and, concomitantly, re-entry of a water molecule W'1 replacing W1 allows E166 to deliver the proton back to the same carbonyl carbon atom, thus achieving hydrolysis of the beta-lactam compound and enzyme recovery. The model well explains the differences found in the kcat. values for hydrolysis of benzylpenicillin and cephalosporin C by the Streptomyces albus G beta-lactamase. It also explains the effects caused by site-directed mutagenesis of the Bacillus cereus beta-lactamase I [Gibson, Christensen & Waley (1990) Biochem J. 272, 613-619].

The VIM metallo-beta-lactamases are emerging resistance determinants, encoded by mobile genetic elements, that have recently been detected in multidrug-resistant nosocomial isolates of Pseudomonas aeruginosa and other Gram-negative pathogens. In this work a T7-based expression system for overproduction of the VIM-2 enzyme by Escherichia coli was developed, which yielded approximately 80 mg of protein per litre of culture. The enzyme was mostly released into the medium, from which it was recovered at >99% purity by an initial ammonium sulphate precipitation followed by two chromatography steps, with almost 80% efficiency. Determination of kinetic parameters of VIM-2 under the same experimental conditions previously used for VIM-1 (the first VIM-type enzyme detected in clinical isolates, which is 93% identical to VIM-2) revealed significant differences in K(m) values and/or turnover rates with several substrates, including penicillins, cephalosporins and carbapenems. Compared with VIM-1, VIM-2 is more susceptible to inactivation by chelators, indicating that the zinc ions of the latter are probably more loosely bound. These data indicated that at least some of the amino acid differences between the two proteins have functional significance. Molecular modelling of the two enzymes identified some amino acid substitutions, including those at positions 223, 224 and 228 (in the BBL numbering), that could be relevant to the changes in catalytic behaviour.

The role of the mobile loop comprising residues 60-66 in metallo-beta-lactamases has been studied by site-directed mutagenesis, determination of kinetic parameters for six substrates and two inhibitors, pre-steady-state characterization of the interaction with chromogenic nitrocefin, and molecular modeling. The W64A mutation was performed in IMP-1 and BcII (after replacement of the BcII 60-66 peptide by that of IMP-1) and always resulted in increased K(i) and K(m) and decreased k(cat)/K(m) values, an effect reinforced by complete deletion of the loop. k(cat) values were, by contrast, much more diversely affected, indicating that the loop does not systematically favor the best relative positioning of substrate and enzyme catalytic groups. The hydrophobic nature of the ligand is also crucial to strong interactions with the loop, since imipenem was almost insensitive to loop modifications.