Potential energy and electron paramagnetic resonance (EPR) g tensor surfaces of model five- and six-coordinated porphyrins were examined. For both types of complexes, the NO ligand is preferably coordinated end-on, with a Fe-N-O bond angle of approximately 140 degrees. In the free five-coordinated structure, NO undergoes free rotation around the axial Fe-N(NO) bond. This motion is strongly coupled to the saddle-type distortion of the porphyrin ligand. Coordination by the second axial ligand (imidazole) raises the calculated barrier for NO rotation to about 1 kcal/mol, which is further increased by displacements of imidazole from the ideal axial position. The potential energy surface for the dissociation of the weakly coordinated imidazole ligand is exceptionally flat, with variation of the Fe-N(Im) bond length between 2.1 and 2.5 A changing the energy by less than 1 kcal/mol. Experimental orientations of both axial ligands, as well as the Fe-N(Im) bond length, are therefore likely to be determined by the environment of the complex. In contrast to the total energy, calculated EPR g-tensors are sensitive to the orientation of the NO ligand and to the Fe-N(Im) bond length. Contrary to a common assumption, the g tensor component closest to the free-electron value does not coincide with the direction of the Fe-N(NO) bond. From comparison of the calculated and experimental g-tensor components for a range of structures, the rhombic ("type I") EPR signal is assigned to a static structure with NO oriented toward the meso-C atom of the prophyrin ring, and RFe-N(Im) approximately 2.1 A (calcd g1 = 1.95, g2 = 2.00, g3 = 2.04; exptl g1 = 1.96-1.98, g2 = 2.00, g3 = 2.06-2.08). The axial ("type II") EPR signal cannot correspond to any of the static structures studied presently. It is tentatively assigned to a partially dissociated six-coordinated complex (RFe-N(Im) > 2.5 A), with a freely rotating NO ligand (calcd g parallel = 2.00, g perpendicular = 2.03; exptl g parallel = 1.99-2.00, g perpendicular = 2.02-2.03).