Low resolution structure prediction of transmembrane helical bundles

PHYS 24

Kirill Speranskiy, kirill@cmu.edu and Maria Kurnikova, kurnikova@cmu.edu. Department of Chemistry, Carnegie Mellon University, 4400 Fifth Ave., Pittsburgh, PA 15213
Understanding the mechanisms of ion channel functioning is often hindered by the lack of knowledge about their spatial organization. Knowing of even a low resolution structure is vital for developing models of ion channels, e.g. recent advances in understanding the nicotinicoid receptor family were facilitated by development of homological models of the ligand binding domain and a cryo microscopic structure of the transmembrane domain of one member of the family. In this work we present a new coarse grained knowledge-based potential for prediction of the structure of membrane proteins formed by helical bundles, often found in transmembrane domains of ion channels. A set of pairwise residue interaction potentials was derived using information collected from the interior of water soluble proteins. The solvation effects are based on optimization of lipid accessible surface area of the protein. The model potential was tested using structures of well characterized proteins.