We have performed extensive ab initio and classical molecular dynamics (MD) simulations of benzene in water in order to examine the unique solvation structures that are formed. Qualitative differences between classical and ab initio MD simulations are found and the importance of various technical simulation parameters is examined. Our comparison indicates that nonpolarizable classical models are not capable of describing the solute-water interface correctly if local interactions become energetically comparable to water hydrogen bonds. In addition, a comparison is made between a rigid water model and fully flexible water within ab initio MD simulations which shows that both models agree qualitatively for this challenging system.

The optimized cluster expansion methods developed in the first article of this series (I) are generalized to apply to molecular fluids. These methods make use of summations of ring and chain cluster diagrams. The summations are performed explicitly for certain classes of molecular models. The molecules in these classes contain several ``interaction sites,'' and the total interaction between two molecules is a sum of site-site potentials that depend on the scalar distances between sites on the two molecules. The principal results of this work are computationally simple techniques for calculating the thermo-dynamic properties and pair correlation functions of molecular fluids in which the intermolecular interactions are highly angular dependent. The techniques should be reliable since they arise from the same approximations that have been shown to be very accurate when applied to simple fluids

We propose to improve the existing free energy expressions obtained within the framework of the reference interaction site model (RISM) combined with the hypernetted closure. The proposed expression is based on the partial wave expression [S. Ten-no, J. Chem. Phys. 115 (2001) 3724] but includes semiempirical corrections to account properly for excluded volume and hydrogen bonding effects. Testing several free energy expressions for various polar and hydrophobic solutes, we have found that such empirical parameterization of the partial wave expression can provide accurate estimates of hydration energies for different hydrophobic and polar solutes. The pro- posed correction allows one to reduce the discrepancy between the experimental and the calculated data down to 0.7 kcal/mol.

Simulations of solvated macromolecules often use periodic lattices to account for long-range electrostatics and to approximate the surface effects of bulk solvent. The large percentage of solvent molecules in such models (compared to macromolecular atoms) makes these procedures computationally expensive. The cost can be reduced by using periodic cells containing an optimized number of solvent molecules (subject to a minimal distance between the solute and the periodic images). We introduce an easy-to-use program “PBCAID” to initialize and optimize a periodic lattice specified as one of several known space-filling polyhedra. PBCAID reduces the volume of the periodic cell by finding the solute rotation that yields the smallest periodic cell dimensions. The algorithm examines rotations by using only a subset of surface atoms to measure solute/image distances, and by optimizing the distance between the solute and the periodic cell surface. Once the cell dimension is optimized, PBCAID incorporates a procedure for solvating the domain with water by filling the cell with a water lattice derived from an ice structure scaled to the bulk density of water. Results show that PBCAID can optimize system volumes by 20 to 70% and lead to computational savings in the nonbonded computations from reduced solvent sizes.

The dynamic coupling between a polarizable protein force field and a particle-based implicit solvent model is described. The polarizable force field, TCPEp, developed recently to simulate protein systems, is characterized by a reduced number of polarizable sites, with a substantial gain in efficiency for an equal chemical accuracy. The Polarizable Pseudo-Particle (PPP) solvent model represents the macroscopic solvent polarization by induced dipoles placed on mobile Lennard-Jones pseudo-particles. The solvent-induced dipoles are sensitive to the solute electric field, but not to each other, so that the computational cost of solvent–solvent interactions is basically negligible. The solute and solvent induced dipoles are determined self-consistently and the equations of motion are solved using an efficient iterative multiple time step procedure. The solvation cost with respect to vacuum simulations is shown to decrease with solute size: the estimated multiplicative factor is 2.5 for a protein containing about 1000 atoms, and as low as 1.15 for 8000 atoms. The model is tested for six 20 ns molecular dynamics trajectories of a traditional benchmark system: the hydrated Bovine Pancreatic Trypsin Inhibitor (BPTI). Even though the TCPEp parameters have not been refined to be used with the solvent PPP model, we observe a good conservation of the BPTI structure along the trajectories. Moreover, our approach is able to provide a description of the protein solvation thermodynamic at the same accuracy as the standard Poisson-Boltzman continuum methods. It provides in addition a good description of the microscopic structural aspects concerning the solute/solvent interaction.