Polarizable Force Field for Proteins and Lipids

点击翻译按钮获取中文摘要

PROJECT SUMMARY from the Parent Award Molecular dynamics (MD) simulations based on atomistic models play an increasingly important role in understanding the fundamental physical forces driving the structure and dynamics of biological membranes. However, to have meaningful simulations, accurate and empirical force fields (FFs) are necessary. Although nonpolarizable additive FFs are useful approximations, polarizable models of biological membranes are needed to account for their complex molecular nature. The present efforts to develop and optimize a polarizable FF for lipids will allow fundamental simulation studies of a broad range of processes associated with biological membranes. During the last funding period we made significant progress with the development of a polarizable FF based on the classical Drude oscillator model. The Drude FF has been implemented in CHARMM, NAMD, ChemShell QM/MM, GROMACS and the OpenMM GPU suite, allowing for unbiased simulations on the microsecond time scale as well as simulations exploiting a range of enhanced sampling technologies. We can already model water, ions, proteins, nucleic acids, carbohydrates, and a few neutral phospholipids. At this point there is a critical need to expand the type of phospholipids covered by the Drude FF to enable the modeling of a wider range of biological membrane systems, as over one third of all MD simulations of biological systems involve bilayer membranes. Polarizable models of biological membranes are necessary to account for their complex molecular nature, where a variety of strong electrostatic factors compete with one another over microscopic length-scales. The plan is to perform a global optimization and validation of the Drude FF for lipids using structural, dynamical, and mechanical information as physical target membrane data for the optimization, and improve the treatment of nonbonded interactions via a lattice sum of the long-range van der Waals dispersion (LJ-PME), and simultaneously extend the Drude FF to cover charged lipids with a special attention to the strong interactions of ions with the polar headgroup (Aim 1). We will then use the refined Drude FF to study fundamental aspects biomembrane electrostatics and elucidate the contribution of electronic polarization on the fundamental electrostatics features of biological membranes (Aim 2) with a study of the classical concept of ζ-potential and the properties of membrane-bound voltage-sensitive dyes with polarizable models for the ground and excited state. Lastly, we will develop accurate polarizable models of phosphatidylinositol-4,5-bisphosphate (PIP2), with a special attention to the interactions with monovalent and divalent cations to study ion-induced domain formation and lateral clustering of PIP2 (Aim 3).