Computed Protein-Protein Enthalpy Signatures as a Tool for Identifying Conformation Sampling Problems.

Area of Science:

• Computational Chemistry
• Biophysics
• Structural Biology

Background:

• Protein-peptide binding thermodynamics are crucial but difficult to compute.
• Existing computational methods face challenges due to molecular complexity and degrees of freedom.

Purpose of the Study:

• To assess the accuracy of computational methods for predicting protein-peptide binding enthalpy.
• To develop a benchmark dataset for validating computational calorimetry approaches.

Main Methods:

• Curated a dataset of 11 protein-peptide complexes with structural and isothermal titration calorimetry (ITC) data.
• Computed absolute enthalpy of binding using computational methods.
• Iteratively refined simulations by addressing identified sampling issues and parameterization.

Main Results:

• Initial calculations showed modest agreement with experimental ITC data.
• Refined simulations achieved a strong correlation with experimental ITC (R^2=0.88, RMSE=1.48 kcal/mol).
• Demonstrated that experimental ITC can guide computational simulations to ensure representative ensembles.

Conclusions:

• Computational calorimetry is increasingly feasible for protein-peptide systems.
• Combining computed and experimental ITC data can validate simulation accuracy and interpret dynamics.
• The curated dataset serves as a valuable benchmark for advancing computational calorimetry.