Predicting the Effect of Amino Acid Single-Point Mutations on Protein Stability—Large-Scale Validation of MD-Based Relative Free Energy Calculations

摘要: The stability of folded proteins is critical to their biological function and for the efficacy protein therapeutics. Predicting energetic effects mutations can improve our fundamental understanding structural biology, molecular basis diseases, possible routes addressing those diseases with drugs. Identifying effect single amino acid point on thermodynamic equilibrium between unfolded states a pinpoint residues importance that should be avoided in process improving other properties (affinity, solubility, viscosity, etc.) suggest changes at positions increasing engineering. Multiple computational tools have been developed silico predictions recent years, ranging from sequence-based empirical approaches rigorous physics-based free energy methods. In this work, we show FEP+, which perturbation method based all-atom dynamics simulations, provide accurate thermal wide range biologically relevant systems. Significantly, FEP+ approach, while originally relative binding energies small molecules not specifically fitted calculations, performs well compared methods were predict stability. Here, present broadest validation energy-based approach applied reported date: 700+ single-point spanning 10 different targets. Across entire data set, correctly classify as stabilizing or destabilizing 84% cases, obtain statistically significant experiment [average error ~1.6kcal/mol coefficient determination (R2) 0.40]. This study demonstrates, first time large-scale validation, calculations used without parameterization system-specific customization, although further improvements additional sampling better representation state protein. describe summarize retrospective results, discuss limitations method, along future directions improvements.