Molecular Basis of Peptide Recognition in Metallopeptidase Dug1p from Saccharomyces cerevisiae Appu Kumar Singh, Mirage Singh, Vaibhav Kumar Pandya, Balasubramani G L, Vijay Singh, Mary Krishna Ekka, Monica Mittal, and S. Kumaran* G. N. Ramachandran Protein Centre, Institute of Microbial Technology, Council of Scientific and Industrial Research (CSIR), Sector 39A, Chandigarh 160036, India * S Supporting Information ABSTRACT: Dug1p, a M20 family metallopeptidase and human orthologue of carnosinase, hydrolyzes Cys-Gly dipeptide, the last step of glutathione (GSH) degradation in Saccharomyces cerevisiae. Molecular bases of peptide recog- nition by Dug1p and other M20 family peptidases remain unclear in the absence of structural information about enzyme-peptide complexes. We report the crystal structure of Dug1p at 2.55 Å resolution in complex with a Gly-Cys dipeptide and two Zn 2+ ions. The dipeptide is trapped in the tunnel-like active site; its C-terminus is held by residues at the S1′ binding pocket, whereas the S1 pocket coordinates Zn 2+ ions and the N-terminus of the peptide. Superposition with the carnosinase structure shows that peptide mimics the inhibitor bestatin, but active site features are altered upon peptide binding. The space occupied by the N-terminus of bestatin is left unoccupied in the Dug1p structure, suggesting that tripeptides could bind. Modeling of tripeptides into the Dug1p active site showed tripeptides fit well. Guided by the structure and modeling, we examined the ability of Dug1p to hydrolyze tripeptides, and results show that Dug1p hydrolyzes tripeptides selectively. Point mutations of catalytic residues do not abolish the peptide binding but abolish the hydrolytic activity, suggesting a noncooperative mode in peptide recognition. In summary, results reveal that peptides are recognized primarily through their amino and carboxyl termini, but hydrolysis depends on the properties of peptide substrates, dictated by their respective sequences. Structural similarity between the Dug1p-peptide complex and the bestatin-bound complex of CN2 suggests that the Dug1p-peptide structure can be used as a template for designing natural peptide inhibitors. D ug1p, encoded by the DUG1 gene (defective in utilization of glutathione), was shown to be essential for glutathione (GSH, L-γ-glutamyl-L-cysteinyl-glycine) metab- olism in Saccharomyces cerevisae. 1 Degradation of the GSH into its constituent amino acids begins with the action of γ-glutamyl transpeptidase, which cleaves the γ-glutamyl moiety from the GSH and generates Cys-Gly dipeptides as intermediates. Cys- Gly dipeptide is reported to be toxic to cells, and hence, an efficient scavenging system is required for its removal. Dug1p, identified as a part of the multiprotein Dug1p-Dug2p-Dug3p GSH degradosomal assembly, catalyzes the last step in GSH degradation in S. cerevisiae by degrading the Cys-Gly intermediate with high specificity. 2 To improve our under- standing of the peptide recognition properties of Dug1p, we studied the binding of a variety of peptides to this enzyme and found that the substrate specificity of Dug1p may be encoded at two different levels. 3 We showed that Dug1p binds a broad range of small peptides but cleaves Cys-Gly peptide with specificity, suggesting that catalytic specificity is achieved after binding. We proposed that Dug1p may employ an induced fit- like mechanism for discriminating substrates from nonsubstrate peptides. However, in the absence of structural information, molecular determinants of peptide recognition and catalysis remain unknown. Dug1p and its two human orthologues, carnosinases CN1 and CN2, belong to the M20 metallopeptidase family. 4 Both CN1 and CN2 cleave carnosine, a bioactive dipeptide (β- alanine-L-histidine) preferably and utilize a common catalytic mechanism in which two divalent metal ions in the active site activate the catalytic water molecule for commencing nucleophilic attack on the incoming peptide bond. 5 CN2 shares a high degree of sequence identity (∼52%) with Dug1p, and the crystal structure of mouse CN2 in complex with bestatin [2-(3-amino-2-hydroxy-4-phenylbutylamino)-4-meth- ylpentanoic acid] has been reported. 6 CN2 functions as a homodimer, and each subunit consists of one large catalytic domain and a lid domain. The catalytic domain consists of residues essential for substrate and metal binding, whereas the Received: October 7, 2014 Revised: November 21, 2014 Published: November 26, 2014 Article pubs.acs.org/biochemistry © 2014 American Chemical Society 7870 dx.doi.org/10.1021/bi501263u | Biochemistry 2014, 53, 7870-7883