Selective effects of transition metal cations (MR+) on biological activities of nerve growth factor (NGF) have recently been described. It has been suggested that four residues in NGF (His(4), His(8), His(84'), and Asp(105')) form a distorted square base pyramidal coordination complex [M(N . His)(3)((-O2Cgamma). Asp)], thereby inducing a conformational transition within the NGF amino terminus (residues Ser(1)-Phe(12)), which constitutes a critical part of the receptor binding determinant. In this report, we provide theoretical and experimental data validating this structure and suggest a model for the selectivity of the M(II)-NGF interaction. The structures of the model complexes [M(NH3)(3)(-O2CCH3)] and [M(HNCH2)(3)(-O2CCH3)] (mimicking the M(II)-NGF coordination site) with first- and second-row divalent transition metal cations Co(II), Ni(II), Cu(II), Zn(II), Rh(II), Pd(II), and Cd(II) were studied by fully optimized ab initio molecular orbital calculations. Regardless of the chemical nature of the neutral ligands, these cations split into three groups: (i) Ni(II), Cu(II), and Pd(II) (d(8) and d(9) metals), which prefer a square pyramidal coordination; (ii) Co(II), Rh(II) (d(7) metals), and Zn(II) (the d(10) first-row transition metal), which prefer a triangular bipyramidal environment; and (iii) Cd(II) (the d(10) second-row transition metal), which has no intrinsic stereochemical preference. It should be noted, however, that stereochemical preferences of Cu(II) and Zn(II) are minor. Molecular mechanics calculations demonstrate that particular geometric features of the M(II)-NGF coordination sire are most suitable for metal cations of intermediate sizes. Taken together with the intrinsic stereochemical preference of transition metal cations, three ions (Zn(II), Cu(II) and Pd(II)) are expected to be specific NGF antagonists, which is consistent with the effects of these ions on the conformation and biological activities of NGF.