TWO MUTATIONS IN THE ACTIVE SITE OF ALCOHOL DEHYDROGENASE PERTURB THE CATALYTIC GEOMETRY Thomas D. Colby, Brian J Bahnson, Jodie K. Chin, Judith P. Klinman, Barry M. Goldstein, Dept. of Biophysics, University of Rochester Medical Center, Rochester, NY 14642, Dept. of Chemistry, University of California, Berkeley, CA 94720.

Two structures of ternary complexes of active site mutants of horse liver alcohol dehydrogenase with NAD and substrate analog have been solved, revealing perturbations that change the relative orientation of the substrate and the cofactor ring. The two mutations are F93->W (in the substrate site) and V203->A (at the nicotinamide end of the cofactor site). Both mutants were designed in order to study the kinetic contribution of quantum mechanical tunneling to the hydride transfer step of alcohol oxidation. Structural results are consistent with kinetic measurements.

Both mutants were crystallized in the presence of NAD and the substrate inhibitor trifluoroethanol.The V203A structure is unusual, having four unique monomers in the asymmetric unit. Complexes were solved by molecular replacement, and refined to 2.0Å (F93W) and to 2.5Å (V203A). Both complexes adopt the catalytically competent "closed form" of the enzyme, characterized by a narrowing of the inter-domain active-site cleft . In both mutants, hydrogen bonds between the carboxamide group of NAD and mainchain atoms from two domains maintain the closed conformation. However, significant differences in cofactor-substrate geometries are observed between the two structures.

In the F93W mutant, the substrate is positioned very close to the nicotinamide ring of NAD, with ~3.1Å between hydride donor and acceptor carbons.This geometry is stabilized by the bulky Trp 93 substitution. In the second mutant, Val is replaced by the reduced bulk of Ala at position 203.

The neighboring nicotinamide ring of NAD rotates toward the resulting pocket, away from the substrate. In each of the four monomers, the cofactor -substrate distance is increased to ~3.5Å, compromising the catalytic geometry.

The F93W mutant was designed to increase the off-rate of bulky alcohol substrates in order to make the hydride-transfer step rate-limiting. Kinetic isotope effects for this mutant suggest an enhanced tunneling contribution. The V->A mutation reduces the kcat/km of the enzyme dramatically, and displayes no increase in tunneling.Structural perturbations in the active site geomeries are consistent with these observations.