TITLE: Computational Approaches to Protein Folding and Molecular Recognition: Method and Applications

SPEAKER: Dr. Ray Luo,

TIME: Thursday Feb. 15, 2001 at 4 PM

PLACE: George P. Williams, Jr. Lecture Hall, (Olin 101)

Continuum solvent models have been widely accepted as an efficient treatment of solvation in computational studies of biochemical systems. However, there has been no satisfactory treatment of solvation in large biochemical systems unless the time-consuming Poisson-Boltzmann approach is used. Recently we implemented an efficient finite difference Poisson-Boltzmann treatment of solvation and long-range electrostatic interactions for macromolecular simulations. Molecular dynamics simulation with the Poisson-Boltzmann method has been demonstrated to be robust and efficient for many systems. The second part of the talk covers an application of continuum solvent models in protein folding simulations. Molecular dynamics simulation may provide very useful insights into protein-folding mechanisms and pathways. However, brute-force simulations by conventional molecular dynamics with explicit water molecules is too expensive to achieve this goal. In this study, we used the Poisson-Boltzmann/Surface Area method to accelerate dynamics simulations. To improve sampling during dynamics, we took advantage of the fast conformational transitions when a continuum solvent is used. In addition, "self-guided" forces, as proposed by Wu and Wang, were used to improve transition rate. Preliminary studies indicate that the current strategy is able to simulate folding of three small proteins ranging from 23 to 36 residues to native-like folds within two nanosecond's simulation time. The main-chain RMSD is in the range of 2 to 4 A for the three small proteins and protein domains. Finally I will address the study on molecular recognition with continuum solvent models. We developed a rigorous and robust free energy algorithm based on continuum solvent models to compute absolute binding free energy. The algorithm, termed "Mining Minima", directly computes free energy by numerically integrating partition functions based on the approximation of continuum solvents and predominant states. The new algorithm has been applied to a wide range of organic and simple biochemical systems.