ELECTRON-TRANSFER REACTIONS IN PROTEINS - AN ARTIFICIAL-INTELLIGENCE APPROACH TO ELECTRONIC COUPLING

The electronic interactions which are responsible for electron transfer in proteins are treated using an artificial intelligence (AI) approach and a quantum mechanical formulation of superexchange. An AI search procedure is devised to select the most important amino acid residues which mediate long-range electron transfer. All the valence orbitals of these amino acid residues are used in a diagonalization of the ''bridge'' orbitals. The electronic coupling matrix element is then calculated by using second-order perturbation theory to couple the bridge orbitals to the donor and acceptor orbitals. The relative values of the electronic coupling elements obtained with this model are shown to be in good agreement with experimental results for cytochrome c derivatives, without use of adjustable parameters. Further, an optimum path calculation in which the path consists of several amino acids is also presented and compared with the many amino acid calculation. The various results show that not merely the separation distance but also the nature of the protein medium separating the redox centers is an important factor in controlling the rate of these electron-transfer reactions.