Relating destabilizing regions to known functional sites in proteins

Area of Science:

• Structural biology
• Computational biophysics
• Bioinformatics

Background:

• Predicting protein functional sites often requires related protein data, limiting application to novel proteins.
• Existing prediction methods lack comprehensive, curated benchmarks for validation.
• Functional sites may be associated with residues that destabilize the native protein structure.

Purpose of the Study:

• To develop and validate a novel computational method for identifying protein functional sites based on residue destabilization.
• To create a large, hand-curated benchmark dataset of protein functional sites for method validation.
• To assess the utility of detecting destabilizing regions for predicting binding sites, especially for proteins with no known relatives.

Main Methods:

• Developed a procedure to identify clusters of destabilizing residues using atomic models and a validated force field.
• Applied the procedure to 63 unrelated protein apo-structures.
• Created a comprehensive benchmark by curating binding sites from structural data and literature.

Main Results:

• Identified destabilizing regions in protein structures.
• Compared destabilizing regions with annotated binding sites across 63 proteins, showing a statistically significant overlap.
• Observed varying overlap depending on ligand type: significant for polysaccharides and small ligands, but not for nucleic acids.

Conclusions:

• Destabilizing regions can provide valuable information for predicting functional sites, especially those binding polysaccharides and small ligands.
• This approach offers a potential method for functional site prediction in proteins lacking known relatives.
• The publicly available benchmark dataset will facilitate the development and validation of new prediction methods.