Cavity Versus Ligand Shape Descriptors: Application to Urokinase Binding Pockets.

Area of Science:

• Structural bioinformatics
• Computational chemistry
• Drug discovery

Background:

• Understanding protein-ligand interactions is vital for drug development.
• The urokinase plasminogen activator (uPA) catalytic domain is a key target in various diseases.
• Accurate characterization of binding pockets is essential for predicting drug efficacy and polypharmacology.

Purpose of the Study:

• To analyze the geometric properties of 78 human uPA catalytic domain binding pockets.
• To develop and apply novel computational descriptors for characterizing these binding pockets.
• To investigate the relationship between binding pocket geometry and ligand properties for polypharmacology prediction.

Main Methods:

• Analysis of 78 binding pockets from crystallized uPA-ligand complexes.
• Application of an original geometric method for pocket computation, avoiding arbitrary parameters.
• Measurement of pocket shape deviation from convexity using the pocket convexity index (PCI).
• Definition and computation of the distributional sphericity coefficient (DISC) using freeware PCI.

Main Results:

• Protein atoms lining uPA binding pockets are located on or near the surface of their convex hull.
• A high correlation (r=0.9) was observed between the radii of uPA binding pockets and their ligands.
• Novel pocket descriptors (PCI and DISC) were computed and shown to correspond well with ligand descriptors.

Conclusions:

• The study introduces a parameter-free geometric method for analyzing protein binding pockets.
• The developed descriptors (PCI and DISC) provide quantitative measures of pocket shape and sphericity.
• The findings highlight the geometric relationship between uPA binding pockets and their ligands, crucial for polypharmacology prediction.