Computational Medicinal Chemistry Approaches for GPCR Structure-Based Drug Discovery
By Chris de Graaf | Nov 1, 2019
About the Conference
GPCR Pharmacology: Activation, signalling and drug design
ERNEST FIRST MEETING / COST CA18133
28 – 30 October 2019
Queen’s University Belfast, Northern Ireland, UK
Dr. Chris De Graaf, Head of Computational Chemistry at Sosei Heptares, recently attended the first conference of the European Research Network on Signal Transduction (ERNEST, COST Action CA18133) named “GPCR Pharmacology: Activation, signalling and drug design”, at the Queen’s University Belfast in Northern Ireland on October 28-30, 2019. As an invited speaker, he addressed several important repercussions and learnings from the analysis of GPCR structures for ligand design that should be transferable and relevant for many targets.
Novel crystal structures of GPCR-ligand complexes solved at Sosei Heptares and elsewhere continue to reveal a diversity of potential ligand binding sites. Emerging sets of GPCR crystal structures of multiple diverse ligands bound to closely related receptors furthermore finally enable a protein-structure based view of how different ligands bind this major drug target class.
This presentation will address several important repercussions and learnings from the analysis of GPCR structures for ligand design that should be transferable and relevant for many targets, including:
- Caveats in using pharmacophore-based similarity principles for modelling receptor ligand complexes different ligand chemotypes
- The important roles of lipophilic hot spots and water networks as drivers of GPCR druggability, ligand binding, and selectivity
- Binding mode diversity of (chemically similar) ligands across the structural GPCRome
This presentation will show how the breakthroughs in GPCR structural biology can be complemented by computational and experimental studies for a more accurate description and prediction of molecular and structural determinants of ligand-receptor binding affinity, kinetics, potency, and selectivity.
Integrated cheminformatics workflows will be described that combine structural, pharmacological, and chemical data to explore receptor-ligand interaction space and steer structure-based virtual ligand screening and ligand design. Novel cheminformatics driven molecule design approaches will be discussed, combining retrosynthetic analysis, library enumeration approaches (e.g. matched molecular pairs analysis), and AI driven (e.g. Recurrent Neural Network combined with Reinforcement Learning) molecule generation.
We will discuss how the accumulated GPCR structural data can be used to develop customized structure-based virtual screening and drug design approaches. We will discuss the systematic evaluation of different docking and structural interaction fingerprint scoring methodologies across class A GPCRs.
Orthogonal physics-based (Molecular Dynamics, e.g. Free Energy Perturbation FEP+ from Schrödinger) and empirical (e.g. GRID, WaterFLAP and BioGPS from Molecular Discovery) structure-based drug design methods will be presented to target lipophilic hotspots, water networks, and cryptic ligand binding pockets for a variety of GPCR subfamilies. We will describe prospective Free Energy Perturbation applications to guide GPCR ligand optimisation for a variety of GPCR ligand binding modes and binding sites, including deeply buried binding pockets, flexible, shallow solvent exposed binding sites, and extrahelical GPCR binding sites at the receptor-membrane interface. We will discuss the essential associated customised solutions to enable FEP for SBDD on these challenging systems, including careful consideration and sampling of ligand ring conformations, residue sidechain rotamers, protein tautomers and protonation states, and binding site solvation networks.