Investigations of the myoglobin cavity mutant H93G with unnatural imidazole proximal ligands as a modular peroxide O-O bond cleavage model system.

A general inability to elucidate extensive variations in the electronic characteristics of proximal heme iron ligands in heme proteins has hampered efforts to obtain a clear understanding of the role of the proximal heme iron ligand in the activation of oxygen and peroxide. The disadvantage of the frequently applied site-directed mutagenesis technique is that it is limited by the range of natural ligands available within the genetic code. The myoglobin cavity mutant H93G [Barrick, D. (1994) Biochemistry 33, 6546-6554] has its proximal histidine ligand replaced with glycine, a mutation which leaves an open cavity capable of accommodating a variety of unnatural potential proximal ligands. We have carried out investigations of the effect of changing the electron donor characteristics of a variety of substituted imidazole proximal ligands on the rate of formation of myoglobin compound II and identified a correlation between the substituted imidazole N-3 pK(a) (which provides a measure of the electron donor ability of N-3) and the apparent rate of formation of compound II. A similar rate dependence correlation is not observed upon binding of azide. This finding indicates that O-O bond cleavage and not the preceding peroxide binding step is being influenced by the electron donor characteristics of the substituted imidazole ligands. The proximal ligand effects are clearly visible, but their overall magnitude is quite low (1.7-fold increase in the O-O bond cleavage rate per pK(a) unit). This appears to provide support for recent commentaries which concluded that the partial ionization of the proximal histidine ligand in typical heme peroxidases may not be enough of an influence to provide a mechanistically critical push effect [Poulos, T. L. (1996) JBIC, J. Biol. Inorg. Chem. 1, 356-359]. Further attempts were made to define the mechanism of the influence of N-3 pK(a) on O-O bond cleavage by using peracetic acid and cumene hydroperoxide as mechanistic probes. The observation of heme destruction in these reactions indicates that displacement of the proximal imidazole ligands by peracetic acid or cumene hydroperoxide has occurred. A combination mutation (H64D/H93G) was prepared with the objective of observing compound I of H64D/H93G with substituted imidazoles as proximal ligands upon reaction with H(2)O(2). This double mutant was found to simultaneously bind imidazole to both axial positions, an arrangement which prevents a reaction with H(2)O(2).