Catechol-O-methyltransferase (COMT) degrades catecholamines, such as dopamine and epinephrine, by methylating them in the presence of a divalent metal cation (usually Mg(II)), and S-adenosyl-L-methionine. The enzymatic activity of COMT is known to be vitally dependent on the nature of the bound metal: replacement of Mg(II) with Ca(II) leads to a complete deactivation of COMT; Fe(II) is slightly less than potent Mg(II), and Fe(III) is again an inhibitor. Considering the fairly modest role that the metal plays in the catalyzed reaction, this dependence is puzzling, and to date remains an enigma. Using a quantum mechanical / molecular mechanical dynamics method for extensive sampling of protein structure, and first principle quantum mechanical calculations for the subsequent mechanistic study, we explicate the effect of metal substitution on the rate determining step in the catalytic cycle of COMT, the methyl transfer. In full accord with experimental data, Mg(II) bound to COMT is the most potent of the studied cations and it is closely followed by Fe(II), whereas Fe(III) is unable to promote catalysis. In the case of Ca(II), a repacking of the protein binding site is observed, leading to a significant increase in the activation barrier and higher energy of reaction. Importantly, the origin of the effect of metal substitution is different for different metals: for Fe(III) it is the electronic effect, whereas in the case of Ca(II) it is instead the effect of suboptimal protein structure.