Interactions of solvent with the heme region of methemoglobin and fluoro-methemoglobin.

It is now more than 20 years since Davidson and collaborators (1957, Biochim. Biophys, Acta. 26:370-373; J. Mol. Biol. 1:190-191) applied the theoretical ideas of Bloembergen et al. (1948. Phys. Rev. 73:679-712) on outer sphere magnetic relaxation of solvent protons to studies of solutions of methemoglobin. From then on, there has been debate regarding the relative contributions to paramagnetic solvent proton relaxation by inner sphere (ligand-exchange) effects and by outer sphere (diffusional) effects in methemoglobin solutions. Gupta and Mildvan (1975. J. Biol. Chem 250:146-253) extended the early measurements, attributed the relatively small paramagnetic effects to exchange with solvent of the water ligand of the heme-Fe3+ ion, and interpreted their data to indicate cooperativity and an alkaline Bohr effect in the presence of inositol hexaphosphate. They neglected the earlier discussions entirely, and made no reference to outer sphere effects. We have measured the relaxation rate of solvent protons as a function of magnetic field for solutions of methemoglobin, under a variety of conditions of pH and temperature, and have given careful consideration to the relatively large diamagnetic corrections that are necessary by making analogous measurements on oxyhemoglobin, carbonmonoxyhemoglobin, and cyano- and azide-methemoglobin. (The latter two, because of their short electronic relaxation times, behave as though diamagnetic). We show that the paramagnetic contribution to solvent relaxation can be dominated by outer sphere effects, a result implying that many conclusions, including those of Gupta and Mildvan, require reexamination. Finally, we present data for fluoro-methemoglobin, which relaxes solvent protons an order of magnitude better than does methemoglobin. Here one has a startling breakdown of the dogma that has been the basis for interpreting many ligand-replacement studies; in contrast to the prevailing view that replacement of a water ligand of a protein-bound paramagnetic ion by another ligand should decrease relaxation rates, replacement of H2O by F- increases the relaxation rate drastically. The data can all be reconciled, however, with what is anticipated from knowledge of ligand interactions in the heme region.