Distribution of Solvent Molecules Around Apolar Side-chains in Protein Crystals

Abstract

We have analysed the distribution of solvent sites within 5·0 Å of the apolar side-chains alanine, valine, leucine, isoleucine and phenylalanine based on experimental data from 24 high-resolution protein structures. Clustering of solvent molecules into specific regions can be seen superimposed on a broad background of sites. The non-random nature of these distributions is confirmed by quantitative analysis of the solvent sites according to a spherical polar (r, θ and φ) co-ordinate system with the apolar atom of interest at the centre. One of the general features of these solvent sites is that they peak at around 4·0 Å from an apolar protein carbon atom. Preferences in orientation (θ and φ) are also seen in the solvent distributions especially around the alanine CB atom and the phenylalanine ring. Most (around 75%) of the solvent sites around apolar groups are also within hydrogen bonding distance of protein (main chain) polar groups which leads to a distribution dependent on the local secondary structure. The remaining 25% of solvent sites are referred to as "non-polar" water molecules and their crystallographic temperature factors are higher than average by between 15% to 28%. For alanine and phenylalanine there are enough data to show that water molecules not within hydrogen bonding distance of protein polar atoms also cluster into specific regions. However, the main conclusion appears to be that the hydrophobic hydration in protein crystals is correlated with hydration of polar groups and thus depends on the local environment as well as on the stereochemistry of the apolar atoms.