Transition-state complex of the purine-specific nucleoside hydrolase of T. vivax: enzyme conformational changes and implications for catalysis.

Nucleoside hydrolases cleave the N-glycosidic bond of ribonucleosides. Crystal structures of the purine-specific nucleoside hydrolase from Trypanosoma vivax have previously been solved in complex with inhibitors or a substrate. All these structures show the dimeric T. vivax nucleoside hydrolase with an "open" active site with a highly flexible loop (loop 2) in its vicinity. Here, we present the crystal structures of the T. vivax nucleoside hydrolase with both soaked (TvNH-ImmH(soak)) and co-crystallised (TvNH-ImmH(co)) transition-state inhibitor immucillin H (ImmH or (1S)-1-(9-deazahypoxanthin-9-yl)-1,4-dideoxy-1,4-imino-D-ribitol) to 2.1 A and 2.2 A resolution, respectively. In the co-crystallised structure, loop 2 is ordered and folds over the active site, establishing previously unobserved enzyme-inhibitor interactions. As such this structure presents the first complete picture of a purine-specific NH active site, including leaving group interactions. In the closed active site, a water channel of highly ordered water molecules leads out from the N7 of the nucleoside toward bulk solvent, while Trp260 approaches the nucleobase in a tight parallel stacking interaction. Together with mutagenesis results, this structure rules out a mechanism of leaving group activation by general acid catalysis, as proposed for base-aspecific nucleoside hydrolases. Instead, the structure is consistent with the previously proposed mechanism of leaving group protonation in the T. vivax nucleoside hydrolase where aromatic stacking with Trp260 and an intramolecular O5'-H8C hydrogen bond increase the pKa of the N7 sufficiently to allow protonation by solvent. A mechanism that couples loop closure to the positioning of active site residues is proposed based on a comparison of the soaked structure with the co-crystallized structure. Interestingly, the dimer interface area increases by 40% upon closure of loop 2, with loop 1 of one subunit interacting with loop 2 of the other subunit, suggesting a relationship between the dimeric form of the enzyme and its catalytic activity.