The landmark paper of Warshel and Levitt (J. Mol. Bio. 103 p227, 1976) paved the way for quantitative studies of enzymatic reactions. This work introduced the hybrid Quantum mechanical/Molecular mechanics (QM/MM) method and a microscopic dielectric model that have represented entire enzyme-substrate complex in solution. This facilitated the first consistent modeling of the catalytic effect of an enzyme. Our group continued to lead the field by further developing QM/MM methods, including the powerful Empirical Valence Bond (EVB) approach, simulating the dynamics of enzymatic reactions and introducing the use of free energy perturbation methods to such reactions. These studies clarified the relationship between reactions in solution and enzymes and established the catalytic role of preorganized active sites. The book on Computer Modeling of Chemical Reactions in Enzymes and Solutions is at present the only book that gives a detailed description of the field. Our current studies continue in pusing the frontiers of the field focusing on (i) developing rigorous yet effective ways for evaluation of activation free energies of enzymatic reactions. This includes the use of ab initio potential surfaces with the corresponding EVB surfaces as reference states in free energy perturbation studies, (ii) modeling classes of enzymes of major biological importance, (iii)modeling quantum mechanical tunneling processes in proteins, (iv)studies of entropic effects in catalysis and binding. Our projects include:
The first molecular dynamics simulation of a biological process in general and of a photochemical process in particular was reported by Warshel in 1976 (Nature, 260, 679). This early work predicted correctly that the primary event in the visual process takes around 100 femtoseconds (this process was assumed at that time to take ~6 picoseconds). Our group was also the first to introduce computer simulation approaches to electron transfer reactions in proteins and to perform the first realistic simulation of the primary light induced charge separation in photosynthetic reaction centers. This simulation study predicted that the primary event involves a stepwise mechanism (Biochem. 27, 774 (1988)), at the time where most workers supported the superexchange mechanism.
Our current effort in this field is focused on providing more detailed description of photobiological process. This includes the (i) studies of conical intersections and non-adiabatic surface crossing in bacteriorhodopsin and rhodopsin, (ii) developing density matrix approaches for simulating excited state dynamics, (iii) studies of new photobiological systems, and (iv) studies of light-induced proton translocations in proteins. The systems studied currently include:
Microscopic simulations of chemical reactions in solution were pioneered in Warshel's work ( J. Phys. Chem., 1982) . This included the use of molecular dynamics to evaluate the activation free energies of chemical reactions in solution, developing the microscopic equivalent of Marcus' parabolas for electron transfer reactions (J. Phys. Chem., 1982; J. Am. Chem. Soc., 1987), and introducing the practical microscopic simulations of Dispersed Polaron approach for simulating nuclear tunneling effects in electron transfer reactions. This Dispersed Polaron approach is sometime known as the Spin-Boson model. We were also the first that implemented path integral centroid methods in simulating chemical reactions in solutions and enzymes (see J. Phys. Chem. 1991 and J. Am. Chem. Soc. 1996). Our "work horse" in the studies of chemical reactions in solution has been the EVB method. This method, which is now used widely by other research groups, provided for the first time a correct coupling between the solute Hamiltonian and the solvent field and allows us to evaluate nonequilibrium solvation effets in a consistent way. The EVB also provides what is at present the most consistent way of transferring ab initio gas phase potential surfaces to solution. We also made significant advances in pioneering QM/MM studies of chemical reactions in solution (J. Phys. Chem. 1979). Our progress in this direction is now focused on approaches that couple the Langevin Dipole (LD) model with ab initio methods (J. Phys. Chem. 1987) and in using the EVB model as a reference potential for ab initio QM/MM simulations (J. Phys. Chem. 1998) and on EVB driven ab initio simulations of all-atom solvent models.
Our early studies (J. Mol. Bio. 1976) paved the way for quantitative studies of electrostatic energies in macromolecule. Our studies involved the development of the first physically complete treatments of electrostatic energies in proteins, ranging from the first simplified microscopic treatment of the energy of charges in solvated proteins to the first free energy perturbation study of such systems. Our electrostatic models are included in the computer program POLARIS. Our current efforts in studies of electrostatic effects in macromolecules include:
The simplified model for protein folding introduced by Levitt and Warshel (Nature 253 p694, 1975) is now emerging as the method of choice for studying protein folding. Our recent studies in this direction focus on the development of effective ways for using the simplified model in rigorous evaluation of the free energies of the corresponding all-atom model (Theor. Chem. Acct. 103, p77, 1999).
Warshel in collaboration with Levitt and Lifson has developed the consistent Force Field (CFF) method and the corresponding computer programs which are the basis of most current molecular modeling methods (CHARMM, AMBER, GROMOS etc). We are focusing currently on refining parameters for chemical reactions in solution.