Modeling the hydration layer around proteins and nucleic acids.

摘要

We develop a theory, termed HyPred, for describing the hydration shells of biological macromolecules and use the model to calculate accurate small/wide angle X-ray scattering (SWAXS), study protein-RNA binding, and calculate electrostatic energies of hydration. HyPred is developed using the approximation that a solute molecule's hydration is determined by local interactions and that long range interactions may be neglected. Thus, by examining the hydration of several different biological macromolecules, the hydration of other similar molecules (e.g., proteins, RNA, DNA) can be predicted since each class has similar surface chemical properties. More specifically all-atom explicit solvent molecular dynamics simulations are performed for several proteins and nucleic acids, and proximal radial distribution functions (pRDFs) are calculated by finding the average solvent density in each cube of a discretized representation of the simulation boxes and then evaluating the average density of cubes as a function of distance from the nearest solute atom and a function of the atom type of the nearest solute atom. This model is validated by reproducing the solvent densities found in the MD simulations and comparing the locations of predicted crystallographic water molecules to experiment. The HyPred model is then further validated by comparing calculated SWAXS patterns of proteins, hydrated by HyPred, to experimental data. The HyPred model is improved by the addition of the dependence of the pRDFs on the second nearest neighbor, the angle formed by the cube, the atom, and the atom to which the nearest solute atom is bonded to, and the local geometry of the solute. HyPred is then extended to model the hydration layer of nucleic acids enabling the study of protein-RNA complexes using HyPred. Finally, HyPred is extended to predict the charge density and the electrostatic energies of hydration and electrostatic potential energy maps