Getting Phosphorylated: Is it Necessary to be Solvent Accessible ?

Abstract

The solvent accessibility of the Ser/Thr/Tyr containing peptide stretches in the substrate proteins of various kinases are likely to play a major role in substrate recognition by various kinases. Even though some of the computational tools make use of the solvent accessibility probabilities for prediction of phosphosites, no systematic analysis has been carried to investigate the solvent accessibilities of known phosphorylation sites. In this study, we have systematically analyzed the solvent accessibilities of serine, threonine and tyrosine containing phosphorylation sites in known substrate proteins by mapping them onto PDB structures, and compared with the accessibilities of sites which are not phosphorylated. The average relative solvent accessible area of phosphorylation site residues was found to be significantly more than their onphosphorylated counterparts. The difference between phospho and non-phospho residues was statistically significant as judged by Wilcoxon test p-values of 2.20 × 10–16, 5.07 × 10–6 and 2.34 × 10–8 for serine, threonine and tyrosine containing sites respectively. Nonetheless, there are several known hosphorylation sites whose relative accessible surface areas are lower than 10Å2, thus these sites are not accessible on the surface of the substrate protein. MD simulations on representative protein structures suggest that, thermal fluctuations at 300K can significantly enhance the accessibilities of such buried Ser/ Thr/ Tyr containing peptides, thus making them available for phosphorylation. Thus our analysis highlights the fact that, it would be in general difficult to set a deterministic criterion based on accessibility values for identifying potential phosphorylation sites.