Predicting Binding Affinities for GPCR Ligands Using Free-Energy Perturbation

Area of Science:

• Computational chemistry and molecular modeling
• Structural biology and pharmacology
• Drug discovery and medicinal chemistry

Background:

• G-protein-coupled receptors (GPCRs) are crucial drug targets, with increasing structural data available.
• Predicting binding free energies for drug optimization has been challenging, limiting computational lead optimization.
• Accurate prediction of ligand binding is essential for efficient drug discovery.

Purpose of the Study:

• To systematically characterize the performance of free-energy perturbation (FEP) calculations for predicting relative binding free energies of ligands to GPCRs.
• To establish a consistent and parameter-free FEP protocol for GPCR targets.
• To guide the identification of potent molecules in drug discovery lead optimization.

Main Methods:

• Utilized the FEP+ package with a consistent protocol including a full lipid bilayer and explicit solvent.
• Validated the protocol by predicting binding affinities for 45 ligands across four GPCRs (adenosine A2AAR, β1 adrenergic, CXCR4, and δ opioid receptors).
• Applied the validated FEP+ protocol in a prospective study to predict affinities of novel adenosine A2A receptor antagonists.

Main Results:

• The FEP protocol demonstrated a highly predictive ranking correlation (average Spearman ρ = 0.55) and low root-mean-square error (0.80 kcal/mol) compared to experimental data.
• In a prospective study, four novel adenosine A2A receptor antagonists were synthesized, showing nanomolar affinities.
• The FEP+ predictions correctly identified the affinities of two novel and three previously reported ligands within 1 kcal/mol, including a tenfold affinity increase.

Conclusions:

• Established a systematic and reliable protocol for applying FEP+ calculations to GPCR targets.
• The FEP+ method provides accurate predictions of relative binding free energies, facilitating drug discovery.
• This work offers guidelines for using FEP+ to identify potent molecules in lead optimization projects for GPCRs.