Comparative binding energy analysis for binding affinity and target selectivity prediction.

Area of Science:

• Computational chemistry
• Drug discovery
• Molecular modeling

Background:

• Drug design faces challenges in achieving receptor selectivity, leading to potential side effects.
• Structurally related proteins, like thrombin, trypsin, and uPA, often bind similar ligands, complicating selective inhibition.

Purpose of the Study:

• To investigate the efficacy of Comparative Binding Energy (COMBINE) analysis, Protein Interaction Property Similarity Analysis (PIPSA), and ligand docking for predicting drug target selectivity.
• To develop and evaluate target-specific scoring functions for discriminating between inhibitors of thrombin, trypsin, and urokinase-type plasminogen activator (uPA).

Main Methods:

• Generated target-specific COMBINE models using known inhibitors and complex structures.
• Employed PIPSA and ligand docking (GOLD) for virtual screening and re-ranking of compounds.
• Validated scoring functions against experimental receptor selectivity data.

Main Results:

• COMBINE models identified key interactions driving binding specificity for each serine protease.
• The integrated approach showed promise in predicting ligand selectivity between thrombin, trypsin, and uPA.
• Limitations in docking accuracy and scoring affected overall predictive power.

Conclusions:

• Target-specific COMBINE scoring functions can predict ligand selectivity for related serine proteases.
• This computational strategy aids in designing selective inhibitors and reducing off-target effects in drug discovery.