Binding energy molecular system simulations

A review is given of the application of Molecular Dynamics (MD) computer simulation to complex molecular systems. Three topics are treated in particular the computation of free energy from simulations, applied to the prediction of the binding constant of an inhibitor to the enzyme dihydrofolate reductase the use of MD simulations in structural refinements based on two-dimensional high-resolution nuclear magnetic resonance data, applied to the lac repressor headpiece the simulation of a hydrated lipid bilayer in atomic detail. The latter shows a rather diffuse structure of the hydrophilic head group layer with considerable local compensation of charge density. 

Abstract. The physical nature of nonadditivity in many-particle systems and the methods of calculations of many-body forces are discussed. The special attention is devoted to the electron correlation contributions to many-body forces and their role in the Be r and Li r cluster formation. The procedure is described for founding a model potential for metal clusters with parameters fitted to ab initio energetic surfaces. The proposed potential comprises two-body, three-body, and four body interation energies each one consisting of exchange and dispersion terms. Such kind of ab initio model potentials can be used in the molecular dynamics simulation studies and in the cinalysis of binding in small metal clusters. 

Quantum chemical methods are valuable tools for studying atmospheric nucle-ation phenomena. Molecular geometries and binding energies computed using electronic structure methods can be used to determine potential parameters for classical molecular dynamic simulations, which in turn provide information on the dynamics and qualitative energetics of nucleation processes. Quantum chemistry calculations can also be used to obtain accurate and reliable information on the fundamental chemical and physical properties of molecular systems relevant to nucleation. Successful atmospheric applications include investigations on the hydration of sulfuric acid and the role of ammonia, sulfur trioxide and/or ions... 

The explicit/implicit solvent approach just described requires doing at least one simulation for each protein-ligand complex. Therefore, it is still difficult to examine the binding of a large number of compounds to a receptor. However, if one focuses on a small subset of chemical space around a lead compound, one can adopt the same approximations as described earlier in free energy calculations so that simulations on the reference systems alone can be used to predict the effects of making many modifications on a lead compound. In these calculations, no molecular dynamics simulation needs to be performed on the derivatives of a lead compound. Instead, snapshots of the reference simulations are modified... 

---