An Improved Free Energy Perturbation FEP+ Sampling Protocol for Flexible Ligand-Binding Domains

Area of Science:

• Computational chemistry
• Molecular modeling

Background:

• Free energy perturbation (FEP) calculations, particularly FEP+, are valuable for pharmaceutical lead optimization.
• Existing FEP+ protocols can be refined for enhanced accuracy in binding affinity predictions.

Purpose of the Study:

• To propose and validate a modified FEP/REST (replica exchange with solute tempering) sampling protocol.
• To improve the accuracy and precision of FEP+ binding affinity predictions.

Main Methods:

• Extended pre-REST sampling times (0.24 ns/λ to 5 ns/λ and 2×10 ns/λ).
• Increased REST simulation duration (5 ns to 8 ns) for convergence.
• Applied REST to the entire ligand and flexible protein residues (pREST region) within the ligand binding domain (LBD).
• Utilized preliminary molecular dynamics (MD) runs for accurate binding mode determination and alignment.

Main Results:

• Achieved more precise ∆∆G values, including correct transformation signs and reduced error.
• Demonstrated improved FEP+ predictions for flexible-loop motions and significant structural changes.
• Observed considerable improvements in FEP+ results across multiple targets by implementing the pREST region strategy.

Conclusions:

• The modified FEP/REST protocol significantly enhances the accuracy of binding affinity predictions.
• Extended sampling times and broader application of REST improve free energy calculations.
• This optimized protocol holds potential for further increasing FEP+ accuracy in drug discovery efforts.