Ligand binding characteristics for orphan receptor GPR139

Dnr:

SNIC 2016/2-32

Type:

SNAC Small

Principal Investigator:

Willem Jespers

Affiliation:

Uppsala universitet

Start Date:

2016-11-23

End Date:

2017-12-01

Primary Classification:

10601: Strukturbiologi

Webpage:

Allocation

Abstract

The G-protein coupled receptor (GPCR) GPR139 is a class A GPCR for which the endogenous ligand has not been characterised and can therefore be considered an orphan receptor. Furthermore, no experimental structural data like crystal structures are available. The low sequence similarity as compared to other class A GPCRs (20-25%) make it a challenging target for homology modelling. (Gloriam, BBA, 2005, 3, 235) GPR139 is mainly expressed in the brain and has a potential role in food intake, making it a possible candidate for obesity treatments. Recent efforts towards the synthesis and characterisation of high affinity (radioligand) agonists undertaken by JNJ and Lundbeck in collaboration with academia (Copenhagen University) show the growing interest for this receptor. Extensive SAR on two of these ligand series provide an excellent starting point for the (computational) characterisation of the binding mode if these ligands. (Dvorak, ACS Med. Chem. Lett., 2015, 6 (9), pp 1015–1018) (Shi, ACS Med. Chem. Lett., 2011, 2 (4), pp 303–306) In our current study we proposed several residues involved in ligand binding using mutagenesis studies. The mutations were based on initially generated homology models. Both experimental values and calculations for these residue mutations showed large discrepancies. This finding led us to believe that the generated homology model did not sufficiently describe the ligand binding mode and could benefit from optimisation using molecular dynamics (MD) simulations. We will use our recently published protocol for computational characterisation of (binding) free energies of mutant receptors using the free energy perturbation (FEP) methodology (Keränen, Chem. Commun., 2015, 51, 3522) in combination with our webserver GPCR-modsim (Esguerra, NAR, 2016, 44, 455), which models and minimises GPCRs in their native (trans)membrane environment. Here, we will analyse several homology models and several docking poses of two high affinity compounds from the JNJ and Lundbeck programs in an iterative fashion to obtain the model best correlating with experimental (mutagenesis) data. Finally we will test our top models for both ligands by using the FEP methodology on several ligands from the two series to further understand the available SAR of these compounds. The model can thereafter be used in the proposition of new high affinity compounds for GPR139.