Probes of a role for remote binding interactions on hydrogen tunneling in the horse liver alcohol dehydrogenase reaction.

Abstract

A tunneling contribution to hydride transfer has been demonstrated previously in the oxidation of benzyl alcohol catalyzed by an active-site mutant (F93W) of horse liver alcohol dehydrogenase (LADH) [Bahnson, B. J., et al. (1993) Biochemistry 32, 5503-5507]. Mutation of a residue that lies directly behind the nicotinamide ring of the bound cofactor has further shown that side-chain bulk can contribute to catalytic efficiency and tunneling in a correlated fashion [Bahnson, B. J., et al. (1997) Proc. Natl. Acad. Sci. U.S.A. 94, 12797-12802]. Second site mutations of F93W have now been made at positions more remote from the active site. In particular, we have focused on an isoleucine residue that interacts with the adenine moiety of the NAD(+) cofactor, 20 A from the nicotinamide ring. Replacement of this remote residue with glycine (F93W:I224G), alanine (F93W:I224A), valine (F93W:I224V), and leucine (F93W:I224L) is concluded to destabilize the binding of NAD(+). All double mutants exhibited a K(M) for NAD(+) that is 2-25 times higher than that for the F93W enzyme. However, neither the catalytic efficiency for turnover of benzyl alcohol [k(cat)/K(M(benzyl alcohol))] nor the relationship between the secondary k(H)/k(T) and k(D)/k(T) isotope effects for benzyl alcohol oxidation was significantly affected. The lack of differences observed in the isotope effects indicates that these mutations have little effect on the extent of hydrogen tunneling in the reaction. The complete removal of the side chain at position 224 in the F93W:I224G enzyme resulted in a less than 5% decrease in the ratio of the secondary isotope effects, maintaining the ratio above the semiclassical limit for the indication of tunneling in the reaction. By contrast, K(i) for NAD(+) increased 60-fold for this mutant. The results obtained with F93W:I224G are consistent with remote interactions that affect the association and binding of cofactor in a reactive conformation. However, once this conformation is achieved, hydride transfer and its tunneling component proceed as with the single F93W mutant enzyme, uninfluenced by the remote mutation. Replacement of other side chains, with alpha-carbon positions from about 8 to over 20 A from the C4 position of the nicotinamide ring, demonstrated a similar insensitivity of k(cat)/K(M(benzyl alcohol)) to protein modification. Comparison to earlier studies with active-site mutants of LADH implicates a role for proximal, but not distal, side chains in the modulation of hydrogen tunneling for this enzyme.