Solution 1D and 2D NMR, together with limited isotope labeling, have led to the assignment of the heme, axial His, and numerous heme contact residues in sperm whale, horse, and human deoxy myoglobin. The paramagnetic relaxivity leads to increased line widths and shorter T(1)s with little compensation in increased dispersion due to dipolar shifts. Hence only limited, but crucial F helix standard backbone sequence specific assignment could be made for heme cavity residues. Numerous other residues with significant dipolar shifts could be assigned from the characteristic scalar connectivities and dipolar contacts to the heme predicted by the crystal structure. It is concluded that the complete and unambiguous assignment of the heme pyrrole substituent signals is not attainable by 2D NMR alone without either partial deuterium labeling of the heme or parallel assignment of key residues in dipolar contact with the heme; hence the present study revises some earlier assignments. The resulting dipolar shifts for nonligated residues, together with the crystal coordinates of deoxy myoglobin, were used to determine the orientation relative to the heme and the anisotropy of the paramagnetic susceptibility tenser. The significant anisotropy, \Delta chi\ similar to 1 x 10(-9) m(3)/mol, however, is shown to result in dipolar shift with reciprocal square, rather than just reciprocal, absolute temperature dependence, which is indicative of large zero field splitting rather than g-tensor anisotropy. The appropriate equation for a B-5(2) ground state allows an estimate of the zero-field splitting, D similar to-10 cm(-1), which is in good agreement with earlier results. The present NMR data favor a spin-only magnetic moment with S = 2 and D similar to-10 cm(-1) over a ground state with S < 2 and significant orbital contribution (Hendrich and Debrunner, 1989).