G Protein-coupled Receptors (GPCRs) are transmembranal proteins in charge of signal transduction across cellular membranes. The mature stage of the research in this field have been recognized with the Nobel prize in Chemistry (2012), and the biomedical relevance is clear with 40% of the marketed drugs target a GPCR.
Our research group has a longstanding experience on the structural biology and ligand design of GPCRs. We are currently running several projects, alone or in collaboration with experimental groups, to understand the structural basis of GPCR structure-function relationship and to assist in the ligand design process for several receptors. In the present proposal, we want to expand our computational exploration of GPCRs (project code SNIC 2016/1-417) as follows:
1. MD studies of GPCR-ternary complexes: dimerization and G-protein coupling and activation. We are characterizing the complexes between the GPCRs (homo and heterodimerization) as well as GPCR in complex with the signaling protein (G-protein). Fore this, we are using the recently updated PyMemDyn module (Esguerra et al, Nucl Acid Res, 2016, 44:W455), which is installed as a standalone python wrapper for the GROMACS MD package in Triolith. PyMemDyn is also implemented in our webserver for the 3D structural modeling of GPCRs GPCR-ModSim (http://open.gpcr-modsim.org). Molecular dynamics calculations are combined with protein-protein docking protocols to build and equilibrate GPCR dimers (collaboration with Juan Fernandez-Recio, Barcelona Supercomputing Center), and the new implementation on GPCR-ModSim of this module will be part of the novel version of the server, which we intend to submit to the NAR special web server issue (2018). The A3-A3 homodimer is a current case study for ligand design of design of homobivalent ligands in collaboration with Prof. Eddy Sotelo (see below on adenosine receptors). Finally, the characterization of A2A / G-protein binding will be the project of a new MSc student (coming from Leiden University on an Erasmus program). The project will consist on: i) building the initial complex (based on crystal structure of the A2A in complex with a G-protein fragment) ii) MD simulations of the complex with GROMACS, and iii) Free energy simulations of mutant variants performed with our MD software Q to design mutational studies, carried on at Leiden University. The mechanism of G-protein binding and selectivity will be the matter of this study.
2. Design of peptidic agonists of the neuropeptide Y (NPY) type-2 (Y2) receptor and assessment of the binding modes of its endogenous ligands. This multidisciplinary project is in collaboration with Prof D. Larhammar (Uppsala University) and the pharmaceutical company Novo Nordisk (Denmark). The scope of this project is: i) to identify the molecular determinants for enhanced agonistic activity of NPY and PYY family of peptides and ii) to assess the chemical modifications that enhance this activity. The binding mode of PYY in Y2 was determined (Xu et al., Biochemistry, 2013, 52:7987) and refined with long MD simulations (using resources SNIC 2015/1-304, SNIC 2016/1-417). This model was the starting point to evaluate the effect of point mutations on agonist binding, via free energy perturbation (FEP) calculations performed with our Molecular Dynamics software Q-software installed in Triolith (Keränen, Chem. Commun., 2015, 51, 3522). These calculations were used to design and interpretat new experiments from our collaborators, i.e. site-directed mutagenesis on the Y2 receptor (Larhammar) and chemical modifications of selected positions of the (3-36)-PYY peptide (Novo Nordisk), the results communicated in a manuscript undergoing final revision stage in the journal Molecular Pharmacology. In this application, we plan to undertake the following actions:
a) The crystal structure of the homologous Y1 receptor will be released by the time this application is under evaluation (Stevens and Larhammar, personal communication). We will rebuild our Y2 model based on this structure, and perform selectivity studies via FEP simulations using a thermodynamic cycle that compares the binding of selective/non-selective ligands to the two proteins (see workflow in Fig 2 of Jespers et al., Molecules, 2017, 22:1945).
b) The project with our pharmaceutical collaborators enters into the ligand-design stage project: we will examine the effect of substitutions on short peptide analogues synthesized and evaluated by Novo Nordisk, by applying our FEP protocols.
3. Automation, and application of our MD protocol for free energy perturbations on amino acids and ligand series. Our recently published tool for FEP calculations on amino acids (Keränen, Chem. Commun., 2015, 51, 3522) in combination with our GPCR-ModSim webserver (Esguerra, NAR, 2016, 44, 455) is currently used in several projects with different collaborators.
a) De-orphanization of GPCRs in collaboration with D. Gloriam (Copenhagen University), see Nøhr et al., Sci. Rep. (2017) 7:1128.
b) Development of new ligands for the adenosine receptor with Prof. E. Sotelo (University of Santiago de Compostela), see recent publications in Azuaje, J. Med. Chem. 2017 60:7502; Jespers et al., Molecules, 2017, 22:1945.
c) Characterization of selectivity in adenosine receptors: with the new structure of A1 adenosine receptor, the A1/A2A selectivity will be characterized using our FEP protocols Jespers et al., Molecules, 2017, 22:1945. We here collaborate with Prof. Ad IJzerman (Leiden University).
d) A2A / G-protein binding. The project of a new MSc student (coming from Leiden University on an Erasmus program) will be on the
4. Structure-based design of selective agonist ligands for the angiotensin II receptor type 2 (AT2). This project is in internal collaboration with Uppsala University groups of Prof. Anders Hallberg (Medicinal Chemistry) and Mathias Hallberg (Pharmaceutical Sciences). After a low activity phase, due to maternity leave in our group member dedicated to this project, we have produced some preliminary data that we intend to expand along the present year using the SNIC resources. The latest developments include the availability of crystal structures of AT1 and AT2 receptors (respectively, inactive and active-like) were published recently and, together with models that were produced in the group during the last years (Sallander et al., Bioorg Med Chem Lett, 2016, 26:1355), provide the structural grounds for this project. In addition, we have selected series of AT2 selective and not selective agonists with a wide range of affinities from the group of Prof. Hallberg (Hallmerg et al., Med Res Rev DOI: 10.1002/med.21449). The compounds have been docked on the AT2 X-ray structure, and pharmacophore and 3D-QSAR models generated. From here, we propose to move on to the next phase of the project, as follows:
• To develop structure- and energy-based SAR models based on Linear Interaction Energy calculations of the different series of molecules, for which we need to run MD sampling in Q software of each complex generated, and analyze with the corresponding Q-gui LIE module
• Perform FEP simulations between selected pairs of ligands where small variations of affinity lead to significant changes in affinity, together with changes in efficacy (i.e.,m stabilization of active or inactive conformation of the receptor, see Fig 2 in Jespers et al., Molecules, 2017, 22:1945). Also with Q software.