Molecular dynamic simulation of multi-component lipid bilayers using multi-scale computing techniques

Abstract

Cholesterol is an indispensable component of animal cell membranes. The basic properties of biomembranes are strongly related to the interaction between cholesterol and other biomembrane molecules. Molecular Dynamics (MD) simulation was used in this study to investigate the interactions between cholesterol, phospholipids, and proteins. The Umbrella model is a leading conceptual model of cholesterol-lipid interaction. One prediction of Umbrella model is that at high cholesterol concentrations, cholesterol molecules tends to line up one by one forming the so-called "maze patterns". In this study, the coarse-grained (CG) MD simulation techniques was used to simulate DOPC, POPC and DPPC bilayers containing 66 mol % of cholesterol. The results strongly support the Umbrella model which predicted the maze pattern. The results in three PC/66%CHOL systems were compared with the simple MC simulation, and we characterized the maze pattern at molecular scale. This study sets a good foundation for future experimental measurement. In my second research project, gramicidin A, which forms cation channels in biological membranes was investigated in liquid-ordered (Lo) and liquid disordered (Ld) membrane domains using MD simulation techniques. Simulation results explained the experimental finding that gramicidin partition favorably into the Ld domains. Local lateral distribution of lipids around gramicidin A was investigated. The results indicate that gramicidin A recruits a shell of short DOPC lipids surrounding each protein and keeps cholesterol and taller DSPC away from the protein-bilayer interface. The result also indicates that membrane proteins are capable of inducing non-uniform distributions of lipids and creating a local bilayer environment, which favors protein functions.