Automated method for predicting enzyme functional surfaces and locating key residues with accuracy and specificity

Area of Science:

• Biochemistry
• Structural Biology
• Computational Biology

Background:

• Identifying enzyme active sites and catalytic residues is crucial for understanding enzyme mechanisms.
• Current methods often rely on structural similarity or known active site databases, limiting their applicability.

Purpose of the Study:

• To present novel, fold-independent methods for predicting and characterizing enzyme catalytic sites at the atomic level.
• To develop a computational approach for identifying functionally important protein surfaces and key catalytic residues.

Main Methods:

• Extraction of atomic patterns from catalytic residues within geometrically computed surface pockets.
• Development of a library of atomic patterns from approximately 700 protein structures.
• Integration of secondary structure propensities and residue occurrence data for prediction.

Main Results:

• A method to identify functionally important protein surfaces and locate key residues was developed.
• The method was applied to various enzyme classes including amylase, dioxygenase, deaminase, dehalogenase, and hydratase.
• Large-scale cross-validation demonstrated high sensitivity and specificity of the prediction method.

Conclusions:

• The presented methods offer a robust, fold-independent approach for predicting enzyme catalytic sites.
• This technique enhances the ability to study enzyme functions by accurately identifying critical residues and surfaces.
• The findings have broad implications for enzyme engineering and drug discovery.