Absolute binding free energy calculations for molecules binding to the GPCR GCGR at the protein/bilayer/water interface.

Abstract

Numerous structures have been solved of small molecules binding at the protein-lipid interface of membrane proteins, such as GPCRs and ion channels. Allosteric modulators that bind at these lipid-exposed sites provide a novel way to target these therapeutically important proteins with greater specificity. Computational tools play a major role in modern drug discovery pipelines, but these tools must be validated, extended, and developed before being applied prospectively to lipid-exposed sites. The creation of a small molecule with a high affinity for its target is a key step in the drug discovery process and the accurate prediction of affinity remains a central challenge in computational chemistry. Alchemical absolute binding free energy (ABFE) calculations have been shown to be highly accurate in predicting the binding affinity of small molecules to soluble proteins. However, ABFE calculations remain challenging, with little work assessing their application to membrane proteins. Our work has focused on extending ABFE calculations to lipid exposed sites. The binding free energy calculations for these sites are complicated by the fact the binding site is ‘solvated’ by lipids. We present a study on the GPCR GCGR in complex with the small molecule MK-0893. MK-0893 sits at the protein/bilayer/water interface. This, coupled with the size and charge of the molecule, makes ABFE calculations for this system particularly challenging. We discuss the adaptations we made to our ABFE workflow to overcome these challenges. Particularly, we make suggestions as to how to monitor the generation of a ‘correct’ apo state for the non-interacting ligand, with regards to the occupancy of the binding site and side-chain rotamers. These adaptations significantly improved our results and show a good agreement with experimental values from a publicly available SAR series.