The purpose of this project is to investigate the physical mechanisms by which the lipid membrane influences the protein functions that underlie most biological processes. A typical project in the lab identifies a hypothesis for a particular mechanism in conceptual terms, forms a mathematical or physical model for the process, then tests and refines the model using a molecular simulation. Next the project is developed to make predictions that can be tested in the laboratory. The projects use the NIH Biowulf computing cluster to run the simulations and models. Molecular dynamics software (such as NAMD and CHARMM) are used to conduct molecular simulations. In-house software development for eventual public distribution is a key element of the lab. A number of projects have been developed this year. 1) Interpretation of Small Angle Neutron Scattering (SANS) experimental data on lipid membrane systems by simulation. In a typical lipid SANS experiment, a coherent (phase-aligned) neutron beam is aimed at a vesicle sample. Neutrons are scattered by the sample and arrive at the detector with a quantum mechanical phase shift, resulting in an interference pattern that implies the molecular structure of the sample. However, this interference pattern is collapsed onto a single coordinate, the angle of scattering. From one point of view, molecular simulation can assist in the interpretation of the SANS experiment by explicitly showing what structures give rise to the interference pattern. From the reverse point of view, SANS experiments can validate observations in simulations. The software we have developed allows for the calculation of the SANS interference pattern from a molecular simulation. With new experiments by our collaborators, we have applied it to validate our previous observations of lateral structure in complex lipid membranes. 2) Modeling the effect of lipid redistribution in complex membranes to explain membrane softening experiments. We have used our newly developed arbitrary topology membrane modeling software to examine the effect of diffusing additives (lipids or proteins) on membranes. Given a modest effect of the additive on membrane mechanical properties, thermal redistribution will result in apparent softening of the membrane. However, this softening is not real -- it will not trivially lower the energy of membrane structures. Rather, our model must be applied to examine the effect of the additive on a particular structure, e.g., a fusion pore or membrane pit. 3) Simulating the free energy of Mucin-1 dimerization at the plasma membrane. We have conducted two interesting free energy simulations of Mucin-1 dimerization, using umbrella sampling and free energy perturbation theory. The results are in qualitative agreement with our collaborators experiments on the same system, suggesting the modeling and methodological framework can be applied to various single-pass dimerization processes relevant to cellular signaling.
