Metallo--lactamases (MLs) represent one of the main mechanisms of bacterial resistance against -lactam antibiotics. The elucidation of their mechanism has been limited mostly by the structural diversity among their active sites. All MLs structurally characterized so far present a Cys or a Ser residue at position 221, which is critical for catalysis. GOB lactamases stand as an exception within this picture, possessing a Met residue in this location. We studied different mutants in this position, and we show that Met221 is essential for protein stability, most likely due to its involvement in a hydrophobic core. In contrast to other known MLs, residue 221 is not involved in metal binding or in catalysis in GOB enzymes, further highlighting the structural diversity of MLs. We also demonstrate the usefulness of protein periplasmic profiles to assess the contribution of protein stability to antibiotic resistance.
The expression of -lactam-degrading enzymes (-lactamases) is the most prevalent mechanism of antibiotic resistance in bacteria (13, 31). Metallo--lactamases (MLs) employ Zn(II) as a catalytic cofactor to cleave the -lactam ring, thus inactivating these antibiotics (2, 6-8, 13-15, 18, 31, 35). These enzymes are particularly worrisome in the clinical setting, in that they can hydrolyze a broad spectrum of -lactam substrates (including the latest-generation carbapenems, such as meropenem and imipenem) and are resistant to most clinically employed inhibitors (2, 6-8, 15, 35). These facts, together with the worldwide dissemination of ML-encoding genes (21), raise a concerning clinical problem. In addition, the design of an efficient pan-ML inhibitor has been limited by the diversity of their active-site structures, catalytic profiles, and metal ion requirements for activity among different members of this family of enzymes (3,4,8,12,16,17).MLs have been classified into subclasses B1, B2, and B3 based on their primary structures (14). Molecular structures solved by X-ray crystallography of MLs from the three subclasses have revealed a common ␣/␣ sandwich fold (3,4,10,12,16,17,33). However, many differences exist regarding zinc coordination environments and metal site occupancies (Fig. 1). MLs bind up to two metal ions in their active sites. In the broad-spectrum B1 and B3 enzymes, Zn1 is tetrahedrally coordinated to three histidine ligands (His116, His118, and His196) and a water/OH Ϫ molecule (3H site) (3,12,17,33). On the other hand, the coordination polyhedron of Zn2 in B1 enzymes is provided by Asp120, Cys221, His263, and one or two water molecules (DCH site) (4, 12). Notably, this site conforms the active species in mono-Zn(II) B2 enzymes (which hydrolyze only carbapenems) (16). Instead, two mutations (Cys221Ser and Arg121His) affect the Zn2 coordination geometry in B3 MLs, where the metal ion is bound to Asp120, His121, His263 (DHH site), and one or two water molecules, while Ser221 is no longer a metal ligand (10,17,33). A remarkable exception is the B3 ML GOB from Elizabethkingia menin...