A density functional theory investigation of Fe-N-O bonding in heme proteins and model systems

We report the results of a series of density functional theory (DFT) calculations of the Mössbauer quadrupole splittings and isomer shifts in NO heme model compounds, together with the results of calculations of the Mössbauer quadrupole splittings, isomer shifts, and electron paramagnetic resonance hyperfine coupling constants in a model Fe(II)(NO)(imidazole) complex as a function of Fe-NO bond length and Fe-N-O bond angle. The results of the Mössbauer quadrupole splitting and isomer shift calculations on the NO heme model compounds show good accord between theory and experiment, with the largest errors being observed for structures having the largest crystallographic R(1) values. The results of the property surface calculations were then used to calculate Fe-NO bond length and Fe-N-O bond angle probability surfaces (Z-surfaces) for a nitrosyl hemoglobin, using, in addition, an energy filter. The results obtained yielded a most probable Fe-NO bond length (r) of 1.79 A and an Fe-N-O bond angle (beta) of 136 degrees -137 degrees. This bond length is somewhat longer than those observed in most model compounds but may be due, at least in part, to hydrogen bond formation with the distal His residue. Bond elongation was also observed in a geometry optimized Fe(II)(NO)(imidazole) complex hydrogen bonded to an imidazole residue, in which we find r = 1.76-1.78 A and beta = 137 degrees -138 degrees. The computed bond angles are close to the canonical approximately 140 degrees value found in most model systems. Highly bent Fe-N-O bond angles or very long Fe-NO bond lengths seem unlikely to occur in proteins, due to their high energies. We also investigated the molecular orbitals and spin densities in each of the six coordinate systems investigated and found the orbitals and spin densities to be generally similar those described previously for five coordinate systems. Taken together, these results show that Mössbauer quadrupole splittings and isomer shifts, in addition to electron paramagnetic resonance hyperfine coupling constants, can now be calculated for nitrosyl heme systems with relatively good accuracy and that the results so obtained can be used to determine Fe-N-O geometries in metalloproteins. The Z-surface approach is thus applicable to both diamagnetic (CO) and paramagnetic (NO) heme proteins with in both cases the metal-ligand binding geometries found in the proteins being very close to those seen in model systems.