| In order to understand the behavior of membrane-bound molecules—such as sidechain analogs and other small molecules, peptides, and proteins—we must develop a better understanding of the interactions between them and the membranes. One powerful tool for doing this is computer simulation. To use simulations, we must first decide how to represent the various components of the calculation. This work presents calculations using two different membrane representations: an all-atom environment, sampled by molecular dynamics, and a newly developed lattice dipole model. We ran molecular dynamics calculations of two tryptophan analogs embedded in lipid bilayers, in order to rationalize the placement of tryptophan residues in membrane proteins, and the experimentally observed behavior of those analogs. The simulations seem to indicate that, although the isolated analogs have free energy minima when packing in the acyl chains, tryptophan residues appear in the headgroup region because they make the most favorable interactions with that environment. The lattice dipole model is intended to calculate reasonably accurate solvation free energies for membrane-bound molecules without the computational cost of all-atom molecular dynamics. Once fully developed, this model could be used to investigate the effects of membrane width and headgroup type on membrane protein structure, predict membrane protein structure, or serve as a boundary for an aperiodic molecular dynamics simulation. The present work presents the theory, development, and testing of this model. Taken together, these projects indicate that computer simulations can be used to improve our understanding of membranes and membrane proteins. |
