Binding of substrate in a ternary complex of horse liver alcohol dehydrogenase

Horse liver alcohol dehydrogenase was crystallized from an equilibrium mixture containing predominantly NAD+ and p-bromobenzyl alcohol. X-ray diffractometer data to a resolution of 2.9 A were collected and used to compute electron density maps with phases calculated from the isomorphous enzyme . NADH . dimethyl sulfoxide complex, which has been refined to an R value of 25.6%. The electron density maps were readily interpreted in a graphics display system. Both subunits of the dimer bind coenzyme and alcohol in essentially the same manner; there is no evidence of asymmetry between subunits. The bromophenyl group is accommodated in a large hydrophobic pocket that has the side chain of Leu-116 rotated into a different position than in the complex with dimethyl sulfoxide. The alcohol oxygen is directly ligated to the catalytic zinc atom. The zinc is tetracoordinate and there is no room for a water molecule to make the zinc pentacoordinate. A hydrogen-bonded system formed with the hydroxyl groups of the alcohol, Ser-48 and nicotinamide ribose (2'), and the imidazole of His-51 may provide a proton relay system that links the buried alcohol to solvent. The insertion of the coenzyme's hydroxyl group into this system appears to install the catalytically active species. The observed structure has the pro-R hydrogen on C1 of the alcohol pointing away from C4 of the nicotinamide ring. This is probably a nonproductive complex that easily becomes productive by a rapid rotation of the alcohol to put the pro-R hydrogen within 3 A of C4 of the nicotinamide ring and in position for a direct transfer of hydrogen. A model of the productive complex readily explains the stereospecificity of hydride transfer observed for ethanol.