A nuclear magnetic resonance study of cobalt II alcohol dehydrogenase: Substrate analog-metal interactions

Two models for the active site of liver alcohol dehydrogenase (EC 1.1.1.1) have been proposed. Results of X-ray diffraction studies (B.V. Plapp, H. Eklund, and C.-I. Brändén, 1978, J. Mol. Biol.122, 23–32) on the native enzyme indicate that substrates are directly coordinated to the active site zinc ion, while NMR studies (D. L. Sloan, J. M. Young, and A. S. Mildvan, Biochemistry14, 1998–2008) on the Co II enzyme indicate that substrates are not bound directly to the metal. It was unclear whether the basis for this difference was structural or technical. Therefore, this NMR study has been done with wellcharacterized zinc and cobalt enzymes, and to facilitate comparison with X-ray diffraction data, the substrate analogs chosen were dimethyl sulfoxide and trifluoroethanol. Binding of either analog to the zinc enzyme in the presence of the appropriate cofactor produced unique changes in the T₁ and T₂ relaxation rates of the ${}^1\text{H}$ and ${}^{19}\text{F}$ nuclei. Similar results were obtained when cobalt enzyme was used for T₁ measurements, but relaxation was more rapid due to the presence of the paramagnetic ion. From these data, the distances between the analog nuclei and the catalytic site cobalt ion were calculated to be 8.9 ± 0.9 and 10.5 ± 1.2 Å for the cobalt enzyme-NADH-dimethyl sulfoxide and the cobalt enzyme-NAD⁺-F₃CCH₂OH complexes, respectively. The distances are comparable and the magnitudes indicate that the functional groups are not directly coordinated to the active site cobalt ion. These values are in good agreement with those previously reported by Sloan et al. (1975) for the cobalt enzyme-NADH-isobutyramide complex, and are consistent with their model in which a metal water ligand forms a bridge between the substrate and the metal. Therefore, there must be a structural basis for the differences observed in magnetic resonance versus X-ray diffraction studies.