Continued development of an empirical function for predicting and rationalizing protein-protein binding affinities

Area of Science:

• Biochemistry
• Computational Biology
• Structural Biology

Background:

• Protein-protein and protein-peptide interactions are crucial for biological processes.
• Accurate prediction of binding affinities is essential for drug discovery and understanding molecular mechanisms.
• Existing methods for binding affinity prediction can be computationally intensive or lack accuracy.

Purpose of the Study:

• To develop and validate a fast, simple empirical equation for predicting protein-protein and protein-peptide binding affinities.
• To structurally rationalize binding affinity predictions based on physical descriptors.
• To identify limitations of the current model and suggest avenues for improvement.

Main Methods:

• Developed an empirical equation using six regression-weighted physical descriptors.
• Assumed rigid-body association and negligible contributions from unmodeled factors (approx. 0 kcal).
• Tested predictions against experimental binding free energy data for 197 interface alanine mutants.

Main Results:

• Achieved excellent agreement between predicted and experimental binding affinities for 79% of alanine mutations.
• The empirical equation demonstrated good predictive power within the framework of rigid-body association.
• Failed predictions were often linked to mutation-induced violations of the simplifying assumptions.

Conclusions:

• The developed empirical equation shows promise for fast and accurate prediction of binding affinities.
• Explicit consideration of electrostatic interactions, particularly charge and aromatic side chains, is necessary for a more universally applicable physics-based approach.
• Further refinement of the model to include specific interaction types will enhance its predictive capabilities.