Computational Biophysics

点击翻译按钮获取中文摘要

Our group simulates biological systems at the molecular level. The two themes of the proposal are:****VIRIAL EXPANSIONS FOR PROPERTIES OF DILUTE SOLUTIONS. Aqueous solutions are ubiquitous in biological systems. Thus a cell is usually bathed on both inside and outside by aqueous solutions of ions, nutrients, peptides, proteins, etc. The goal of theory and simulation is to explain the thermodynamic (and other) solution properties in terms of the intermolecular forces. In a classic but notoriously difficult paper, McMillan and Mayer presented a theory for the thermodynamic properties of dilute solutions. It involves two stages: (a) eliminating the solvent (water in our case) by averaging over its molecular motions, leaving the solute molecules with new, effective, interaction forces, and (b), expanding the osmotic pressure in a virial or solute concentration series, with coefficients (the virial coefficients) expressed in terms of the effective solute-solute intermolecular interactions. The effective solute interactions can be obtained by molecular simulation. We have greatly simplified the derivations, and extended the theory by deriving virial series for a number of other properties including the solution enthalpy, which can be measured accurately with modern calorimetry methods. By comparing virial coefficients obtained from experiment with those obtained by simulation, one learns about the effective solute-solute interactions in solution. We have applied the new theory to a classic test case, benzene in water, with good results when compared with osmotic pressure and calorimetry experiments. We propose to apply the theory to the biological system of antimicrobial peptides in solution, for which we have just done calorimetry experiments. **BIOMEMBRANE PERMEABILITY OF SMALL SOLUTES. The cell membrane is a selectively permeable barrier. Most molecules and ions require specific transporters to cross the membrane, but water, oxygen, other small solutes and most drugs cross it by simple diffusion. In the modern theory the permeabilty P of a molecule is predicted to depend on the effective interaction potential with the membrane w(z) at points z across the membrane, and on the local diffusion coefficient D(z). The quantities w(z) and D(z) must be obtained by simulation at the molecular level. Our group has developed new efficient and accurate methods to obtain w(z) and D(z) simultaneously from the same simulation. With these methods we have calculated P for water, oxygen and tyramine (a trace amine involved in neuroregulation), and where experimental data exist, agreement is good. We are proposing to extend the theory of membrane permeability to account for molecules which chemically react while permeating. This is in fact the case with tyramine, which interconverts between a neutral and a charged (protonated) form, and up to the present time we have neglected the interconversion.**