Core Flipping in Lead Optimization: Rank Ordering Using λ-Dynamics.

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Accurate ligand-receptor structural models are crucial for structure-activity relationship (SAR) studies in drug discovery.
• Molecular docking can generate multiple plausible ligand poses, necessitating methods to distinguish the most likely binding mode.
• Substitutions in congeneric series can lead to diverse binding orientations, complicating SAR analysis.

Purpose of the Study:

• To develop and validate a computational methodology for determining relative binding affinities between alternative ligand poses.
• To provide a reliable method for ranking ligand poses in structure-based drug discovery.
• To assess the predictive power of the proposed approach using pharmaceutically relevant targets.

Main Methods:

• Relative binding free energy calculations utilizing the λ-dynamics method.
• A dual-topology approach incorporating distance-restraining schemes.
• A novel one-step perturbation strategy to quantify the contribution of applied restraints.

Main Results:

• The proposed methodology correctly ranked known X-ray poses as more favorable than alternative flipped poses for two pharmaceutically relevant targets.
• The approach demonstrated validity and predictive power using eight small-molecule inhibitors.
• The one-step perturbation strategy was effective, though FEP/MBAR showed convergence issues.

Conclusions:

• The developed methodology offers a robust way to discriminate between alternative ligand binding poses.
• This approach can significantly aid in establishing accurate structure-activity relationships for congeneric series.
• The methodology is extensible to ranking multiple poses and evaluating amino acid rotamers.