Molecular simulations of G protein-coupled receptors : A journey into structure-based ligand design and receptor function

Sammanfattning: The superfamily of G protein-coupled receptors (GPCRs) contains a large number of important drug targets. These cell surface receptors recognize extracellular signaling molecules, which stimulates intracellular pathways that play major roles in human physiology. Breakthroughs in structural biology have led to an exponentially increasing number of atomic resolution GPCR structures, which have provided insights into the molecular basis of ligand binding and receptor activation. However, in order to use these structures in rational drug design, computational methods able to predict ligand binding modes and affinities are required. In the first part of this thesis, molecular simulations were used to explore the potential of using structure-based approaches to discover and optimize GPCR ligands. In paper I, molecular dynamics (MD) simulations in combination with free energy perturbation (FEP) guided improvements of binding affinities for fragment-like ligands of the A2A adenosine receptor (A2AAR), which is a target for Parkinson’s disease and cancer. Two computational approaches were then explored to design selective GPCR ligands. MD/FEP was first used to guide the optimization of a weak fragment ligand for subtype selectivity. Simulations of the A1- and A2AARs led to the discovery of high affinity and selective A1AR antagonists (paper II). In the second approach, a molecular docking screen of millions of molecules was carried out against AR crystal structures with the goal to identify A1AR ligands. Structure-based optimization of two hits resulted in the discovery of potent and selective A1AR antagonists (paper III). In paper IV, the role of a binding site water in agonist binding to the A2AAR was probed by modifying the endogenous agonist adenosine. MD simulations highlighted the complexity of ligand binding and the benefits of using FEP calculations to guide ligand optimization. In the second part of the thesis, MD simulations were used to study the activation mechanism of class A GPCRs and the function of class F receptors. The allosteric communication between the orthosteric and G protein binding sites of the β2 adrenergic receptor was investigated, which revealed the roles of structural motifs in receptor activation (paper V). Finally, MD simulations of a homology model of the Frizzled 4 receptor, which is a target for the development of anticancer drugs, led to the identification of a conserved structural motif that is important for receptor signaling (paper VI). The results of the thesis show that computer simulations can be valuable tools in structure-based drug discovery and studies of GPCR function.