Lack of structural information is a major stumbling block in the characterization of the majority of proteins and their naturally occurring variants. Modeling programs like Quanta, Insight and Look are useful in overcoming the above-mentioned problems, to a great extent, by providing tools to generate three-dimensional structures. We have initiated study on the structures of various membrane receptors and enzymes that are being investigated for genetic functional variation in the Laboratory of Neurogenetics. Structures of the transmembrane helical domains of various serotonin receptors are being generated using the electron density map of bovine rhodopsin as the template structure. Extra-cellular and intra-cellular loops having 16 or less amino acids will be added to this helical structure to arrive at a model for each of the serotonin receptors. Analyses of molecular dynamics, using the programs Charmm and Dock, will be carried out on these structures to understand ligand specificity and binding. In addition, homology modeling based on the crystal structure of yeast transketolase has been performed for the thiamin dependent enzymes, mammalian transketolase and the E1alpha and E1beta subunit of pyruvate dehydrogenase complex, whose activities are reportedly reduced in several neurodegenerative diseases. The structures of the various naturally occurring variants observed in these enzymes are being compared to those of the native counterparts to delineate the roles of certain variant amino acids in the binding of substrates and co-factors as well as enzyme catalysis. Furthermore, computer-aided modeling can be utilized in the characterization of other proteins that are being investigated by the NIAAA.
