PCA-MutPred: Prediction of Binding Free Energy Change Upon Missense Mutation in Protein-carbohydrate Complexes

Area of Science:

• Biochemistry and Molecular Biology
• Structural Biology
• Computational Biology

Background:

• Protein-carbohydrate interactions are vital in biological processes.
• Mutations in carbohydrate-binding proteins can alter binding affinity, leading to dysfunction and disease.
• Predicting the impact of mutations on binding affinity (ΔΔG) is complex.

Purpose of the Study:

• To identify sequence and structural factors influencing binding affinity changes in protein-carbohydrate complexes upon mutation.
• To develop a predictive model for ΔΔG in mutated protein-carbohydrate complexes.
• To create a publicly accessible web server for predicting these affinity changes.

Main Methods:

• Collected experimental binding affinity change data for 318 unique mutants.
• Analyzed sequence and structural features such as accessible surface area, secondary structure, mutation preference, conservation score, hydrophobicity, and contact energies.
• Developed multiple regression equations for prediction and validated using cross-validation and an independent test set.

Main Results:

• Identified accessible surface area, secondary structure, mutation preference, conservation score, hydrophobicity, and contact energies as key factors.
• Achieved an average correlation of 0.74 and MAE of 0.70 kcal/mol on 10-fold cross-validation.
• Validated the model on an independent dataset with a correlation of 0.79 and MAE of 0.56 kcal/mol.

Conclusions:

• Developed PCA-MutPred, a web server for predicting binding affinity changes in protein-carbohydrate complexes.
• The predictor demonstrates high accuracy in forecasting ΔΔG.
• PCA-MutPred can serve as a valuable tool for designing protein-carbohydrate complexes with tailored affinities.