Effects of Acids, Bases, and Heteroatoms on Proximal Radial Distribution Functions for Proteins

Abstract

The proximal distribution of water around proteins is a convenient method of quantifying solvation. We consider the effect of charged and sulfur-containing amino acid side-chain atoms on the proximal radial distribution function (pRDF) of water molecules around proteins using side-chain analogs. The pRDF represents the relative probability of finding any solvent molecule at a distance from the closest or surface perpendicular protein atom. We consider the near-neighbor distribution. Previously, pRDFs were shown to be universal descriptors of the water molecules around C, N, and O atom types across hundreds of globular proteins. Using averaged pRDFs, a solvent density around any globular protein can be reconstructed with controllable relative error. Solvent reconstruction using the additional information from charged amino acid side-chain atom types from both small models and protein averages reveals the effects of surface charge distribution on solvent density and improves the reconstruction errors relative to simulation. Solvent density reconstructions from the small-molecule models are as effective and less computationally demanding than reconstructions from full macromolecular models in reproducing preferred hydration sites and solvent density fluctuations.

Solute–Solvent Energetics Based on Proximal Distribution Functions

“…28,33,35,39 Essentially, the richer the set of atom types we include, the higher the accuracy in solvent densities we can expect during the reconstructions, especially when we consider the error within the solute molecule’s first shell. 34 …”

“…As discussed in the previous section, the energetic estimations essentially depend on the accuracy of the reconstructed density, particularly in the region near each solute atom. Although the relative error within the neighboring hydration shell of the solutes may not be dominant in global solvent densities reconstructions, 34 the steep repulsion in proximity to the solute atoms can numerically amplify these small errors. We demonstrate this by plotting the accumulated $U_{\mathrm{O}-\text{OT}}^{\text{vdW}}$ as a function of distance from the carbonyl oxygen (as shown in Figure 6a).…”

Solvation and cavity occupation in biomolecules

“…Our 3D IE method with exact Coulomb interactions [27] and proximal Radial Distribution Function reconstruction [46] are compared to all-atom molecular dynamics simulations. The solutions for both the IE and pRDF techniques are easily and efficiently determined for small and large molecules and comparison to all-atom MD simulations provide a determination of how well they predict the solvent distribution around and inside myoglobin.…”

“…We also find that reconstructions of water density from the side chain analogs-water pair radial distribution functions more closely resemble simulated solvent density distribution. [46] This most recent method is used in the current work to reconstruct the solvent density distribution within different confined regions of myoglobin to scan for internal hydration sites.…”

Nonpolar Solvation Free Energy from Proximal Distribution Functions

“…48 That is, pDFs constructed for small molecules can be used to predict the solvent structures around complex biological macromolecules that are composed of chemically similar components. 49–51 Average solute–solvent interaction energies, and subsequently solvation free energies, can then be calculated from these reconstructed solvent structures.…”