Predicting the Binding of Fatty Acid Amide Hydrolase Inhibitors by Free Energy Perturbation.
Arjun SahaAmy Y ShihTaraneh MirzadeganMark J SeierstadPublished in: Journal of chemical theory and computation (2018)
Since a goal of most drug discovery projects in either academia or industry is to design molecules that selectively bind to the desired protein, determination of protein-ligand binding free energies is of utmost importance in computer aided drug design. With the help of significant improvements in computer power, enhanced sampling techniques and accuracy of force fields, FEP (free energy perturbation) is becoming an important tool to estimate binding free energies in many drug discovery projects both retrospectively and prospectively. We have evaluated the ability of Schrödinger's FEP+ to predict relative binding free energies of a congeneric series of noncovalent fatty acid amide hydrolase (FAAH) inhibitors using an in-house crystal structure. This study shows that although an impressively accurate correlation can be obtained with experimental IC50s considering small perturbations on the deeper side of the pocket, the same was not observed with small perturbations on the relatively more open-ended and solvent-accessible side of the pocket. To understand these observations, we thoroughly investigated several key factors including the sampling of asymmetrically substituted rings, different perturbation maps, impact of simultaneous perturbations at two different ends of the ligand, and selecting the perturbations in a "chemically sensible" way.
Keyphrases
- drug discovery
- fatty acid
- crystal structure
- binding protein
- density functional theory
- dna binding
- quality improvement
- protein protein
- amino acid
- emergency department
- deep learning
- molecular docking
- machine learning
- molecular dynamics
- atomic force microscopy
- solid phase extraction
- mass spectrometry
- transcription factor
- solar cells