The primary goal of this project is to develop methods to predict structures of membrane proteins from their sequences and available experimental data and to use these methods to develop structural models of specific membrane proteins. Most of our effort has been devoted to developing methods of combining computer graphics techniques with energy calculations by the program CHARMM. Structural models we, developed for a series of peptides (melittin, delta lysin, magainin, PGLa, and pardaxin) that interact with membranes to lyse cells and/or form membrane channels. We modeled these systems to study how these monomers aggregate on one surface of a membrane, how they insert across the membrane, and how they form channels. All of these peptides have sequences consistent with their forming amphipathic alpha helices. Melittin and delta lysin are lytic toxins from bee venom and staphylococcus. Magainin, PGLa, and XPF are potent antimicrobial agents secreted from Xeneopus granular glands. They resemble melittin in that they can form positively charged amphipathic alpha helices; however, they are not hemolytic at concentrations that kill microbes. Magainin induces anion selective channels in artificial membranes. PGLa and XPF are homologous peptides that have no homology to Magainin but have very similar antimicrobial effects. Pardaxin is a potent shark repellent made by the Red Sea Moses Sole. It forms channels in artificial membranes that, unlike most other alpha helical peptide channels, appear to have only one size. Pardaxin channels allow inorganic cations and anions but not Tris base or large organic anions to pass.
