University of Southern California Dissertations and Theses

Structure and function of the orphan G protein-coupled receptor 6

(USC Thesis Other)

Structure and function of the orphan G protein-coupled receptor 6

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Structure and Function of the Orphan G Protein-Coupled Receptor 6 by Mahta Barekatain A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (CHEMISTRY) August 2022 Copyright 2022 Mahta Barekatain ii Acknowledgements To Vadim – I am utterly grateful for the opportunity you gave me to grow and learn, and for your continuous support over the course of my Ph.D. program. I absolutely appreciate everything you have taught me. The science world is vast, and the amount of unknown is overwhelming; thank you for providing the resources and guidance for me to study and understand bits of it. To Ray – I want to extend my gratitude for your support and encouragement over the years, for inspiring me, and introducing me to ideas and people who continue to inspire me in different ways. Thanks for mentoring and helping me diversify my skills, personally and professionally. To Scott, Seva, Kate, and Linda – I was fortunate enough to work with every single one of you closely and absolutely appreciate all the time and effort you invested to help me advance my projects, explore adjacent fields of research, and grow in various areas beyond the bench. To my collaborators – This project and my research could have not been concluded without your significant contribution, thanks for providing your support and expertise to tackle challenging problems that can only be addressed through collaborative work. To my colleagues and friends – Thanks for elevating my Ph.D. experience with your endless support, positivity, and kindness. Jeffrey, Gye Won, Kelly, and Chris – I am grateful to have worked with you and learned so much from you. Thank you for working to support my research and encouraging my perseverance, enthusiasm, and curiosity at all stages of my studies. iii To my family – I could have never achieved so much without your unconditional love and support. Mana, thank you for always encouraging me to pursue my goals and recognizing all my minor achievements along the way. Collin, you have been my best friend, partner, and supporter through some of the more challenging times of my graduate years. I am beyond grateful for having shared parts of this transforming personal and professional journey with you. Mom and dad, words cannot express how lucky and blessed I feel like I am for having you two by my side. Thanks for always believing in me, cheering me up, and celebrating me. I could not have asked for a better family. iv Table of Contents Acknowledgements …………………………………………………………………………........ ii List of Tables …………………………………………………………………..……….…..…... vii List of Figures ………………………………………………………….………..…….…..…..… ix List of Terms and Abbreviations …………………………………..…….……………..…….… xii Abstract ………………………………………………………………………………………… xv Chapter 1 | Introduction to G protein-coupled receptors ……………………………………....… 1 1.1 | Ligand-induced GPCR activation and signal transduction …………….…………… 3 1.2 | Constitutive activity of GPCRs …………………………………………….…….… 10 1.3 | Orphan GPCRs ………………………………………………………………...……11 1.4 | GPR6 activity and implications in neurodegenerative diseases …………………… 13 1.5 | GPR6 modulation by endogenous ligands ……………………………………….... 17 Chapter 2 | GPR6 optimizations for structural studies ………………………………….….…… 20 2.1 | Construct evolution ………………………………………………………….…..… 23 2.2 | Ligand screening ……………………………………………………….……..…… 25 2.3 | Initial crystallization trials ………………………………..…….……….………… 27 2.4 | Fusion protein screening ……………………………………………..….………… 28 2.5 | Point mutation screening ………………………………………………….…..…… 31 2.6 | Crystallization ligand ……………………………………………………….…...… 34 2.7 | Crystallization and data collection at synchrotron ……………………….……....... 37 2.7.1 | Co-crystallization with inverse agonist ……………………………......… 40 2.7.2 | Ligand screening in crystallization trials …………………………….....…47 v 2.7.3 | Crystallization of apo GPR6 ……………………………………......…… 50 2.7.4 | Co-crystallization with S1P …………………………………...........…… 53 2.7.5 | Crystallization of apo GPR6 expressed in mammalian cells ………....…. 54 2.7.6 | Structure solution of GPR6-CVN424 …………………………….……... 55 2.8 | Conclusion ………………………………………………………………........…… 57 Chapter 3 | Structure of GPR6 …………………………………………………………….…..... 60 3.1 | Structure of inactive GPR6 in complex with the inverse agonist 1485 ……….…… 60 3.2 | Structure of inactive GPR6 in complex with the inverse agonist CVN424 …..….…63 3.3 | Structure of active GPR6 in complex with a putative endogenous ligand …...…..…67 3.4 | Structure comparison of GPR6-pseudoapo and other active conformation structures.74 3.4.1 | GPR6-pseudoapo comparison with GPR6-S1P …………………..………74 3.4.2 | GPR6-pseudoapo comparison with GPR6-HEK …………………..….… 76 3.5 | Mechanism of GPR6 activation ………………………………………………….... 78 3.6 | Molecular mechanisms of inverse agonism in GPR6 ……………………………… 82 3.7 | Structure comparison of GPR6 with other lipid receptors of the MECA cluster …. 83 3.8 | Conclusion ………………………………………………………………..…......… 88 Chapter 4 | GPR6 structure-informed studies …………………………………………...….……89 4.1 | In silico ligand docking ……………………………………………….……….……89 4.2 | MD simulations …………………………………………………………..……...… 91 4.3 | Radioligand binding assays …………………………………………………...….... 96 4.4 | Functional cell-based assays ……………………………………………….......… 100 4.5 | Mass spectrometry …………………………………………………………..…… 105 4.6 | Conclusion …………………………………………………………………..…… 109 vi Chapter 5 | Conclusion and future direction …………………………………………………… 111 References ……………………………………………………………………………….…… 115 vii List of Tables Table 1 | GPR6 inverse agonists used in ligand screening. Ligands were provided by Takeda ... 26 Table 2 | Melting temperature (Tm) of 8652 stabilized by various ligands …………………….. 27 Table 3 | Amino Acid Sequences of Fusion Proteins Screened for GPR6 Constructs …………. 29 Table 4 | Summary of construct yield and Tm from fusion protein and tag screening …………. 31 Table 5 | Summary of construct yield and Tm from mutation/truncation/junction screening ….. 32 Table 6 | Comparison of constructs 8652 and 10164 …………………………………………… 34 Table 7 | GPR6 inverse agonists derivatives of ligand 1485 used in ligand screening ………… 35 Table 8 | GPR6 consortium ligands used in ligand screening ……………………….…….…… 35 Table 9 | Crystallization conditions with potential 10164-1485 crystal hits …………….……... 40 Table 10 | Protein yield and concentration for initial 10164-1485 crystallization experiments … 41 Table 11 | Crystallization conditions of initial 10164-1485 crystals …………………………… 41 Table 12 | Initial optimization of crystallization conditions for 10164-1485 crystals ………….. 43 Table 13 | Protein yield and concentration for the first successful 10164-1485 crystallization … 44 Table 14 | Optimized crystallization condition of 10164-1485 crystals ………………………… 44 Table 15 | Protein yield and concentration for the final 10567-1485 crystallization …………… 45 Table 16 | Optimized crystallization condition of 10567-1485 crystals ………………………… 45 Table 17 | Protein yield and concentration for 10567-792 crystallization ……………………… 48 Table 18 | Best crystallization conditions of 10567-792 crystals ……………………………….. 49 Table 19 | Protein yield and concentration for 10567-apo crystallization …………………......... 50 Table 20 | Crystallization conditions of 10567-apo crystals …………………………………….. 50 Table 21 | Crystallization conditions of 10567-S1P crystals ……………………………………. 54 Table 22 | Crystallization conditions of 10917-apo crystals……………………………………... 55 viii Table 23 | Crystallization conditions of 10567-CVN424 crystals …………………………......... 56 Table 24 | Data collection and refinement statistics related to GPR6-IAG structures . ………… 58 Table 25 | Data collection and refinement statistics - GPR6 active conformation structures .….. 59 Table 26 | GPR6 residues of the orthosteric pocket at 4 Å distance of the ligand …………….... 66 Table 27 | Comparison of GPR6 residues within 4 Å of CVN424 with GPR3 and GPR12 ……. 67 Table 28 | Comparison of GPR6 residues within 4 Å of Lig-OLA and/or 1485 ………………... 73 Table 29 | Similarity matrix of the human lipid receptors of the MECA cluster ….….…….…... 84 Table 30 | Conservation of GPR6 residues within 4 Å of Lig-OLA in the MECA cluster …...... 87 Table 31 | Top GPR6 hit compounds from in silico screening …………….…………………..... 90 Table 32 | Summary of Radioligand Kd for different GPR6 mutants …………………………… 98 ix List of Figures Figure 1 | Five major GPCR families and their common topology of 7TM …………………..…. 2 Figure 2 | Efficacy of ligands …………………..……………………..………………………..… 4 Figure 3 | Molecular mechanisms underlying activation pathway of class A GPCRs …..………..6 Figure 4 | Overview of GPCR signaling through G protein-dependent and -independent pathways.7 Figure 5 | G protein signaling pathways …………………..……………………..………………. 9 Figure 6 | Ligand binding can alter the equilibrium between conformational states…………….. 11 Figure 7 | Phylogenetic tree of the MECA cluster family of class A GPCRs ………………….. 14 Figure 8 | Snake plot representation of GPR6 …………………………………………….....… 15 Figure 9 | GPR6 human mRNA expression levels ……...……………………………………… 16 Figure 10 | Evolution of GPR6 constructs from October 2015 to February 2016 …….….…….. 24 Figure 11 | Construct 8652 stabilized by a ligand 3887 and its characterization ………………. 25 Figure 12 | Construct 8652 thermostability profile in the presence of various ligands …………. 26 Figure 13 | Potential 8652-3887 microcrystal hits imaged in the cross-polarized mode ….….… 28 Figure 14 | Summary of fusion protein and his tag site screening ……………………………… 30 Figure 15 | Snake plot of construct 10164 and its comparison with 8652 ……………………..... 33 Figure 16 | Temperature ramping-SEC of apo and ligand-stabilized receptor ………………….. 34 Figure 17 | Temperature incubation-SEC of 1485 derivative ligand stabilized receptor ………. 36 Figure 18 | Temperature ramping-SEC of consortium ligand-stabilized receptor ……………… 37 Figure 19 | First successful crystallization of 10164-1485 ………………………………………42 Figure 20 | Successful crystallization of 10164-1485 in optimized conditions …………………. 44 Figure 21 | Successful crystallization of 10567-1485 in optimized conditions ……..……….……46 x Figure 22 | GPR6 construct 10164 co-crystalized with 3 different consortium ligands ………... 47 Figure 23 | Crystallization of GPR6 construct 10567-792 ……………………………………… 49 Figure 24 | Crystallization of GPR6 construct 10567-apo ………………………………………. 51 Figure 25 | Density of the lipid-like ligand in the orthosteric pocket of the GPR6 active structure.52 Figure 26 | Crystal structure of GPR6 10567-1485 ……………………………………………... 61 Figure 27 | Architecture of the orthosteric binding pocket in 10567-1485 ……………………... 62 Figure 28 | Crystal structure of GPR6 10567-CVN424 …………………………………………. 64 Figure 29 | Architecture of the orthosteric binding pocket in 10567-CVN424 …………………. 65 Figure 30 | Global conformational changes upon receptor activation ………………………….. 68 Figure 31 | Crystal structure of GPR6 pseudo-apo …………………………………………….... 70 Figure 32 | Architecture of the orthosteric binding pocket in GPR6-psuedoapo ……………….. 72 Figure 33 | Chemical structure of OLA and S1P ………………………………………………....75 Figure 34 | Electron density difference map for GPR6-pseudoapo and GPR6-S1P ……………..75 Figure 35 | Hydrolysis of Monoolein ……………………………………………………………. 76 Figure 36 | Electron density difference map for GPR6-pseudoapo and GPR6-HEKpseudoapo ... 77 Figure 37 | Conformational changes of extra-cellular loops upon receptor activation …..……….79 Figure 38 | Molecular microswitches involved in receptor activation ……………………….….. 81 Figure 39 | Structure comparison of lipid receptors of the MECA cluster ……………………… 86 Figure 40 | Trajectories of ligand interactions with GPR6 binding pocket residues ……………. 92 Figure 41 | Overlay of simulation trajectories for GPR6-S1P model ……………………............ 94 Figure 42 | Radioligand saturation binding to GPR6 mutants of the binding pocket residues .... 100 Figure 43 | GPR6 basal signaling in cAMP HTRF assay ……………………………………… 105 Figure 44 | Lysolipids identified in GPR6 lipidome and tested for binding in MS experiment…108 xi Figure 45 | Native MS identifies lyso-PC bound GPR6 species after incubation with the lipid .. 109 xii List of Terms and Abbreviations ANOVA Analysis of Variance APS Advanced Photon Source CC Crystallization Construct CHS Cholesterol Hemisuccinate CNS Central Nervous System CPM 7-Diethylamino-3-(4'-Maleimidylphenyl)-4-Methylcoumarin DDM n-Dodecyl-β-D-Maltopyranoside DMSO Dimethyl Sulfoxide ECL Extracellular Loop FDA Food and Drug Administration GDP Guanosine Diphosphate GPCR G Protein-Coupled Receptor GRK G Protein-Coupled Kinase GTP Guanosine Triphosphate HEK Human Embryonic Kidney Cells HTRF Homogeneous Time Resolved Fluorescence IAG Inverse Agonist ICL Intracellular Loop IMAC Immobilized Metal Affinity Chromatography LCP Lipidic Cubic Phase LPA Lyso-Phosphatidic Acids xiii LPC Lyso-Phosphatidyl Choline MD Molecular Dynamics MECA Melanocortin/EDG/Cannabinoid/Adenosine MR Molecular Replacement MS Mass Spectrometry MSM Markov State Model MW Molecular Weight OLA Oleic Acid OLC Monoolein PCR Polymerase Chain Reaction PD Parkinson’s Disease PDB Protein Data Bank PEG Polyethylene Glycol PI Protease Inhibitor PPG Poly Propylene Glycol RL Radio Ligand S1P Sphingosine-1-Phosphate SD Standard Deviation SDS Sodium Dodecyl Sulfate SDS-PAGE Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis SEC Size Exclusion Chromatography TLS Translation/Libration/Screw TM Transmembrane Helix xiv WT Wild Type XFEL X-ray Free Electron Laser xv Abstract G Protein-Coupled Receptors (GPCR) comprise the largest superfamily of membrane proteins in the human genome with over 800 members, each featuring a unique sequence. GPCRs mediate many of our physiological responses to hormones, neurotransmitters, and environmental stimulants (ions, lipids, etc.) and therefore are implicated in various pathophysiological conditions. Hence, in addition to their endogenous ligands, GPCRs are the target of over 30% of the drugs approved by the United States Food and Drug Administration (FDA) which highlights the tremendous therapeutic potential for this class of membrane proteins. Growing number of high- resolution structures of GPCRs over the past decades has provided insight into their function and enabled structure-based drug design for several GPCR targets. However, the structure and function of many GPCRs with high therapeutic potentials remains elusive. In this study, we utilized x-ray crystallography to determine the structure of the orphan G Protein- Coupled Receptor 6 (GPR6) in various conformational states of the receptor. Our structural data are supported and complemented by insights from computational biology, pharmacology, cellular biology, and protein biochemistry. Collectively, our data elucidate the structure of GPR6 and shed light on ligand binding, signal propagation, and activation of this receptor. The work described in this dissertation was done by me, with the exception of the contributions listed below. xvi Linda Johansson performed the initial construct screenings, as specified in chapter 2, and suggested homology-based stabilizing point mutations. Vsevolod Katritch and Petr Popov suggested stabilizing point mutations predicted by CompoMug. Vadim Cherezov assisted with crystal harvesting and data processing. Gye Won Han assisted with final stages of refinement and structure quality control. Kelly Villers and Chris Hanson carried out expression of GPR6 constructs. Jeffry Velasquez provided molecular biology support. I have also included some of the work done by our collaborators, in chapter 4, for the purpose of completeness. All GPR6 inverse agonist ligands were provided by Takeda Pharmaceuticals and the GPCR consortium. Anastasiia Sadybekov performed ligand docking. Ho Ming Lam performed molecular dynamics simulations. Scientists from Takeda Pharmaceuticals performed radioligand and functional assays. Idlir Liko and Corinne Lutomski performed native Mass Spectrometry. 1 Chapter 1 | Introduction to G protein-coupled receptors G protein-coupled receptors (GPCRs) comprise the largest superfamily of membrane proteins in the mammalian genome. While GPCRs are diverse in their primary structure, they can be categorized according to phylogenic criteria, including amino acid sequences and functional similarities. The classification groups over eight hundred members of GPCRs into five distinct families (Figure 1). The largest group is known as the rhodopsin-like family or class A, and accounts for around 80% of all GPCRs. The adhesion family is the second largest family and together with the secretin family make up the class B. Metabotropic glutamate family, GABA receptors, calcium-sensing receptors, and taste receptors are members of the class C, and Frizzled/Taste2 compose the class F GPCRs 1,2 . The overall architecture of GPCRs features common topology of seven transmembrane helices (7TM), along with three intra- and three extra-cellular connecting loops (ICLs and ECLs). The majority of existing structures also display an additional amphipathic α-helix 8 that follows the 7TM bundle. The N- and C-termini of GPCRs can significantly vary in their length and may have important functional implications 3 . Despite the common topology, the five major GPCR classes share very little sequence identity, accounting for substantial structural and, hence, ligand diversity. Further categorizing members of each family into subgroups generates subfamilies that share higher sequence similarity and often common ligand selectivity 2 . However, specific details of ligand selectivity and activation for many receptors remain elusive and awaits further structural and functional insights. 2 Figure 1 | Five major GPCR families and their common topology of 7TM Different classes of GPCR share similar structure of their transmembrane region, have distinct structural features at the extra-cellular region. Members of each family of receptors respond to a set of endogenous stimuli. 4 Reprinted from reference [4] with permission. 3 1.1 | Ligand-induced GPCR activation and signal transduction GPCRs respond to a variety of external stimuli including photons, odorants, hormones, proteins, lipids, neurotransmitters, natural products, and intermediary metabolites 5 . The signal transduction by these transmembrane receptors regulates myriad of physiological processes including vision, smell, and taste, as well as neurotransmission, cardiovascular, endocrine, and reproductive functions. Versatile functionality of GPCRs is tied to their disease association which suits this family of receptors as very attractive drug targets 2 . Most GPCRs alternate between active and inactive states within an equilibrium and are stabilized in a signaling state in the presence of a signaling partner like G proteins or arrestins. Endogenous or exogenous ligands can shift this equilibrium towards the active or inactive states based on their pharmacology 5,6 . The efficacy of a ligand reflects its ability to activate a particular signaling pathway. A commonly accepted pharmacological classification of ligands with varying efficacies recognizes four different types of ligands (Figure 2). 1) Full agonists are ligands that induce maximal intracellular signal generated by their cognate receptor. Endogenous ligands like neurotransmitters and hormones typically fall into this category of ligand efficacy. 2) Partial agonists can induce receptor’s activity but only elicit responses below the maximal level. 3) Inverse agonists are ligands that inhibit the receptor’s basal activity by shifting the conformational equilibrium towards the inactive states. 4) Neutral antagonists are able to bind the receptor’s orthosteric pocket but do not alter the basal activity of the receptor. While these ligands do not directly induce a response at their cognate receptor, they compete with other pharmacologically active ligands for the receptor’s orthosteric pocket 5 . 4 Figure 2 | Efficacy of ligands. Depending on their pharmacology, endogenous and exogenous ligands are able to elicit a range of various responses at GPCRs by shifting the conformational equilibrium of their cognate receptor towards a more inactive or active-like state. Reprinted from reference [5] with permission. Ligand-induced activation of GPCRs has been extensively studied 7,8 . While the intricate molecular mechanisms underlying ligand recognition and binding are receptor-specific, a general receptor activation mechanism can be deduced for a given class of receptors (e.g. rhodopsin or secretin) from the well-studied systems within that class. In the case of rhodopsin family or class A GPCRs, common molecular switches that are involved in several GPCR activation pathways have been mapped and described as a network of intertwined conformational changes that translate into a conformational change at the intracellular region of the receptor; one that accommodates the engagement of signaling partners inside the cell 7 . Receptor activation is typically induced by extracellular binding of an agonist. Ligand binding initiates a series of receptor-specific conformational changes, leading to reorganization of an intra- helical contact between residues of TM6 along with inter-helical contacts between residues of TM3, 5 and 6, resulting in contraction of TM3-5-6 interface. This reorganization involving 5 conserved class A molecular motifs CWxP and PIF, and collapse of Na + pocket, initializes the rotation of the cytoplasmic end of TM6, followed by the movement of TM7 toward TM3 7,9 . Further opening of the hydrophobic lock between TM3 and TM6 loosens their packing and facilitates the outward kinking of the cytoplasmic end of TM6 10-12 . This outward movement later allows for engagement of the signaling partner. Subsequent rewiring of the NPxxY motif reinforces the packing of TM3-TM7 13 . These reorganizations ultimately result in freeing the Arginine residue of the DRY motif 14 , which allows for its engagement (together with other residues of TM3, 5, and 6) with the G protein (Figure 3). 6 Figure 3 | Molecular mechanisms underlying activation pathway of class A GPCRs Common class A receptor activation is described in four layers of signal initiation (1), signal propagation (2), microswitches rewiring (3) and G-protein coupling (4) 7 . Residues are numbered according to Ballesteros and Weinstein (B.W.) scheme 15 . Reprinted from reference [7], licensed under the Creative Commons Attribution 4.0 International. Once a GPCR is activated, it can regulate various G protein-dependent or independent signaling pathways (Figure 4). There are four major G-protein families (Gs, Gi/o, Gq/11, and G12/13), and each of these G protein heterotrimeric complexes is composed of three subunits, Gα, Gβ and Gγ. In its inactive state, G protein’s guanosine diphosphate (GDP) -bound Gα subunit is associated with Gβγ. Most GPCRs can engage their cognate G protein inactive heterotrimer upon their activation. This engagement then initiates the release of GDP from Gα in exchange for guanosine triphosphate 7 (GTP). The latter is what results in dissociation of Gα subunit from Gβγ subunits. These subunits (Gα and Gβγ) can then each initiate cascades of downstream signaling. Once GTP is hydrolyzed back to GDP, the subunits reassociate and awaits further activation. The reassembled G protein heterotrimeric complex could then be subjected to activation by the same or another activated GPCR 16,17 . Figure 4 | Overview of GPCR signaling through G protein-dependent and -independent pathways. Agonist binding stabilizes the receptor in the active conformation. This in turn mediates downstream signaling by facilitating the engagement of G proteins and/or GRKs. The latter results in phosphorylation of the receptor and subsequent recruitment of arrestins which in turn stops G protein signaling and initiates signaling cascades separate from those originated from the G protein activation 18 . Reprinted from reference [18] with permission. 8 There are 16 human Gα subunits which can be divided into the four subfamilies of Gαi, Gαs, Gα12, and Gαq 19 . The activated Gα subunit, depending on its type, differentially modulates several second messenger systems (e.g. adenylyl cyclase and phospholipases), guanine nucleotide exchange factors (GEFs), and Rho and Ras GTPases. These downstream effector systems ultimately stimulate various phosphorylation events modulated by different kinases including mitogen-activated protein kinases (MAPK) (e.g. AKT and mTOR) as well as second messenger- regulated kinases (e.g. PKA, PKC, and CaMK). The numerous consequences of these signaling cascades in the cytosol and nucleus collectively govern and regulate gene expression, cell metabolism, migration, proliferation, and cell survival. Gβγ dimers also set out to activate a series of signaling events independent from the Gα subunit 19 . For example, they can regulate ion channels by inhibiting Ca 2+ channels or activating G protein–activated inwardly rectifying K channels (GIRKs) 20 . The Gβγ subunits can also stimulate enzymes like phospholipases C beta (PLC-β) and adenylyl cyclases as well as kinases like PI3Ks and GEFs. These regulatory pathways ultimately contribute to the prosurvival and migratory activity of many GPCRs 19 (Figure 5). Alternate intracellular signaling by GPCRs may involve regulatory and scaffolding proteins like arrestins. Activated receptor may be phosphorylated by G protein-coupled receptor kinase (GRK), and subsequently associate with arrestins. GPCR phosphorylation by GRKs and arrestin binding is the common fate of activated GPCRs and marks the end of G protein-mediated signaling from the plasma membrane. This coupling mediates the receptor’s endocytosis and various G protein- independent signaling pathways 16,21 . 9 Figure 5 | G protein signaling pathways. Activated heterotrimeric G protein dissociates into Gα and Gβγ subunits. Each of these subunits is able to regulate various downstream signaling cascades. This figure summarizes different signaling pathways that are modulated by the subtypes of Gα subunit, as well as those regulated by the Gβγ subunit. Reprinted with permission from reference [19], licensed under the Creative Commons Attribution 4.0 International. 10 1.2 | Constitutive activity of GPCRs Consistent with their inherent dynamic behavior in the absence of a ligand, GPCRs typically exist in an equilibrium between various states ranging from inactive to more active-like conformations. The free energy associated with each intermediate conformation is what governs the proportion of the receptor populations across various functional states (Figure 6) 6,18 . A considerable population of unliganded receptors is energetically inclined to remain in their inactive state, stabilized by intramolecular interactions 22 . Nevertheless, many GPCRs have been reported to show some levels of constitutive activity 23,24 , with rare cases where the majority of the receptor population maintains an almost fully active basal state in the equilibrium of the unliganded conformations 23 . Constitutive activity of a receptor is its intrinsic ability to engage and activate signaling partners in the absence of an agonist. This characteristic has been attributed to the dynamic behavior and structural flexibility of these receptors towards their intracellular region which is accentuated in some wild type receptors as well as several mutated GPCRs 25-27 . However, the validity of the studies depends on the established knowledge of the receptor’s endogenous ligand as well as availability of pharmacodynamically varied ligands (agonist, neutral antagonist, and inverse agonist), which is limited in the cases where the endogenous ligand of the receptor is not known 27 . 11 Figure 6 | Ligand binding can alter the equilibrium between conformational states. Hypothetical histograms in each plot reflect the relative populations of different conformational states in an equilibrium. Hypothetical free energy landscapes (gray) are inversely related to the populations of the conformational states. a) Unliganded GPCRs mostly remain in the inactive state. b) Agonist-bound GPCRs tend to sample intermediate and fully active conformations more frequently. c) Agonist-bound and G protein-bound GPCRs primarily populate fully active conformations. In the case of a highly constitutively active receptor, the basal free energy landscape of the unliganded receptor is closer to b and c, rather than a. Reprinted with permission from reference [6] under ACS AuthorChoice/Editors’ Choice usage agreement. 1.3 | Orphan GPCRs The discovery of 7TM topology of opsins and β2-adrenergic receptor gave rise to the concept that these receptors may belong to the same superfamily of proteins, later called GPCRs 28 . From there on, many new GPCRs were discovered using low-stringency hybridization 29 , polymerase chain reaction (PCR) 30 , and whole-genome sequencing 31 . With many newly discovered GPCRs, the challenge was then to link these receptors to their endogenous ligands. The receptors for which an endogenous ligand was not validated were referred to as ‘orphan’ receptors and the process of assigning a highly selective endogenous ligand to these receptors was therefore called ‘deorphanization’. Deorphanizing GPCRs was initially tackled by reverse pharmacology as well 12 as assaying calcium flux, GTPγ binding, and modulation of cyclic adenosine-3',5'-monophosphate (cAMP) levels 32 . Concurrent with ongoing search for endogenous ligands of many GPCRs that remained orphan, new attempts were made to identify novel transmitters 33-36 . These attempts included screening the extracts of tissues that contain potential ligands 33,35 and using GPCRs as probes to screen the extracellular matrix component 37,38 . The quest to identify endogenous ligands of orphan GPCRs led to deorphanization of over three hundred GPCRs since their discovery 32 . However, deorphanization has been lagging over the past decade. The challenges in deorphanizing the remaining orphan GPCRs have been extensively reviewed and discussed 32,39,40 . Ligand availability is the foremost limiting factor as many assays in reverse pharmacology rely on screening extracts of the tissue in search for the endogenous ligand of a given receptor. In many cases, these extracts are obtained from regions of the body where the orphan receptor is highly expressed. However, in cases where the site of the receptor differs from the source of the ligand, the screening might not be fruitful. Other hurdles involve functionality of the receptor, especially when the monomeric orphan GPCR of interest may depend on dimerization to gain functionality 39 . Moreover, many of the unsuccessful deorphanization attempts proved to be inconclusive and controversial, partly due to insufficient screening and control experiments. Hence, several of the receptor ligand pairs that were suggested by multiple groups over the years ultimately did not actually deorphanize the receptor 40 . Lastly, constitutive activity of some orphan receptors poses a new challenge in the quest to deorphanize them. While these receptors might be genuine orphans whose activity is not induced by a ligand 41 , screening of their potential endogenous agonists is merely impractical in the absence of a functional antagonist or inverse agonist. 13 1.4 | GPR6 activity and implications in neurodegenerative diseases G protein-coupled receptor 6 (GPR6) is an orphan receptor in the MECA (Melanocortin/Endothelial differentiation/Cannabinoid/Adenosine) cluster (Figure 7) of the class A GPCRs 42 . The human gene encoding GPR6 was first cloned in 1995 by two independent groups 43,44 , after its mouse 45 and rat 46 homologues were cloned in 1993 and 1994, respectively. This human gene encodes a protein of 362 amino acids (Figure 8) which shares 93.9% amino acid identity with its rodent (mouse and rat) homologues 47 . GPR6, along with its two paralogs GPR3 and GPR12, forms a subgroup of class A orphan receptors that are phylogenetically closest to the cannabinoid receptors. The three receptors share over 60% sequence identity amongst themselves, over 45% sequence similarity with lysophosphatidic acid receptor LPA1, sphingosine-1-phosphate (S1P) receptors S1P1 and S1P5, and over 40% homology with the cannabinoid receptors CB1 and CB2 48 . 14 Figure 7 | Phylogenetic tree of the MECA cluster family of class A GPCRs. The phylogeny of MECA cluster members was determined using the full sequence of the receptors in neighbor- joining distance calculation mode on the GPCRdb.com 49 . 15 Figure 8 | Snake plot representation of GPR6. GPR6 sequence features a 71 amino acid N- terminus, a short C-terminus, and the key conserved motifs among class A GPCRs including NPxxY, DRY, SWxP motif (CWxP in most Class A GPCRs). The receptor also carries two cysteine residues in the extra-cellular loop 2 (ECL2) involved in the internal disulfide bridge characteristic of the receptors of the MECA cluster. The snake plot diagram was obtained from GPCRdb.com 49 . It’s been shown that GPR6 transcripts are mainly localized in human central nervous system (CNS), most notably in the putamen and to a lesser extent, in the frontal cortex, hippocampus, and hypothalamus 43 (Figure 9). The human putamen is part of the striatum which forms the basal ganglia and is involved in learning and motor control. Accordingly, putaminal irregularities are Hydrophobic Hydrophobic aliphatic Hydrophobic aromatic Hydrogen binding Charged positive Charged negative α-helix flexibility α-helix kink Disulfide-forming Residue properties Disulfide bridge 16 observed in both motor and cognitive dysfunctions 50,51 . Additionally, relatively lower GPR6 mRNA expression levels have been reported in the pituitary gland of the endocrine tissue 52 . Interestingly, unlike most GPCRs which reside in the plasma membrane, researchers have reported GPR6 to be mainly localized in the intracellular compartments 53 . Figure 9 | GPR6 human mRNA expression levels. The human atlas reports GPR6 mRNA expression levels to be highest in the basal ganglia of the human brain, specifically in the caudate nucleus, nucleus accumbens, and putamen. Image credit: Human Protein Atlas, licensed under the Creative Commons Attribution-ShareAlike 3.0 International License 54 . GPR6 is the first (indirect pathway) striatopallidal neuron–specific genetic regulator of instrumental conditioning in mammals 55 , and has been shown to enhance neurite outgrowth 56 . This receptor has been reported to constitutively signal through Gs 24,53,56,57 and β-arrestin 58 . However, reports on its signaling through Gi/o 24,57,59 have been contradictory. GPR6 expression profile and constitutive activity have led scientist to speculate its regulatory function in neuronal activities. 17 Hence, its role in neurodegenerative diseases has been reviewed in several separate studies 48 . First indication of GPR6 involvement in Huntington disease (HD) was reported in 2006 where GPR6 decreased mRNA level of GPR6 in human caudate samples was amongst most significant observed changes compared to control samples 60 . One year later, scientists reported enhanced instrumental performance of GPR6-knock-out mice 55 . Another study has revealed the upregulation of GPR6 gene expression in the complement protein C1q-mediated neuroprotection in Alzheimer disease (AD) mouse models 61 . Moreover, GPR6 has been shown to be involved in Schizophrenia 62 and Parkinson’s disease (PD) 62,63 . The motor control circuits of the brain that are regulated by dopamine work in direct and indirect pathways which are hypo- and hyper-active, respectively, in the PD patients 64 . Studies of the mice model of PD have revealed that GPR6 deficiency causes alteration in the striatal cAMP and dopamine levels, and results in higher locomotor activity as well as reduced abnormal involuntary movements 62 . 1.5 | GPR6 modulation by endogenous ligands Despite attempts of different research groups to identify the endogenous ligand of GPR6, there’s yet no consensus on a putative ligand. Reports of ligand screening on GPR6 and its two paralogs, GPR3 and GPR12, have identified S1P and two of its derivatives as potential agonists of this subfamily of receptors 57,59 . S1P is a known regulator of the S1P subfamily of GPCRs and differentially agonizes their signaling through Gi, Gq, G12/13 and Rho to multiple intracellular effector systems 65,66 . Studies have revealed a boost in GPR6-mediated intracellular calcium (Ca 2+ ) mobilization induced by S1P. This signaling is completely abolished in the presence of pertussis toxin that is known to hamper Gi but not Gq coupling to GPCRs 67 which suggests receptor’s 18 signaling through Gi 55,57,59 . However, follow-up screenings which used β-arrestin recruitment to verify GPR6 activation by S1P did not substantiate these claims 68,69 . Hence, International Union of Basic and Clinical Pharmacology (IUPHAR) continues to recognize these three receptors as orphans 70 . Several research groups have published reports on various endogenous and exogenous ligands that have shown to elicit inverse agonism on this receptor. Screenings of phyto-, endo-, and synthetic cannabinoids for their potential regulatory role on GPR6 has identified several novel inverse agonists of this receptor 71 , some of which exert their effects at micromolar concentrations. These molecules selectively inhibit receptor recruitment of β-arrestin and include cannabidiol (CBD) and cannabidavarin (CBDV), two non-psychoactive components of marijuana, WIN55,212-2, a prototypical aminoalkylindole cannabinoid agonist, as well as SR141716A and SR144528 which are CB1 and CB2 receptor antagonists, respectively 58,72 . Moreover, N-acyl dopamines have been identified as inverse agonists for GPR6. These compounds which are dopamine N-acyl condensation products include N-arachidonoyl dopamine (NADA), N-docosahexaenoyl dopamine (DHDA), N-oleoyl dopamine (OLDA) and N- palmitoyl dopamine (PALDA). Much like the cannabinoid candidates, N-acyl dopamines show potencies for GPR6 in the micromolar range and are functionally selective inverse agonists for β-arrestin recruitment 73 . However, unlike the cannabinoid molecules mentioned above, the N-acyl dopamines are naturally found in the striatum of human brain 74,75 , yet their molecular targets are not fully characterized. 19 Inhibition of GPR6 constitutive activity via inverse agonists has been of interest ever since the first indications of GPR6 involvement in neurodegenerative diseases where observed. Accordingly, several pharmaceutical companies have filed for patents on GPR6 inverse agonists over the past two decades. Arena Pharmaceuticals as the first company to pursue compounds that elicit GPR6 inhibition developed two small molecules with potencies in the nanomolar range 76 . Over a decade later, Envoy Therapeutics claimed the use of pyrazine derivatives as inverse agonists of GPR6 in treatment of several neurovegetative diseases 77 . This was later accompanied by several other patents claiming the use of similar compounds, under the name of Takeda Pharmaceutical Company 78-80 that acquired Envoy Therapeutics in 2012. Investigating molecules of pyrazine scaffold by Takeda Pharmaceutical Company later led to development of a potent and selective GPR6 inverse agonists, by Cerevance, to treat motor symptoms in Parkinson’s disease 81 . Despite indications of GPR6 involvement in several neurodegenerative disorders and growing interest in its therapeutic potentials, the receptor’s structure, its constitutive activity, and endogenous ligand remain largely understudied and elusive. Additionally, it has not been substantiated whether the constitutive activity is the inherent characteristic of this receptor, or a mere effect of a ubiquitous yet unknown endogenous ligand. In this dissertation, I describe my work in solving the structure of this receptor in both active and inactive conformations. Further I detail the structural analysis of GPR6 and comparison of its active and inactive conformations. Lastly, I elaborate on how these structures inform the identity of GPR6 endogenous ligand and the steps we have taken towards deorphanization of this orphan receptor. 20 Chapter 2 | GPR6 optimizations for structural studies Structural studies of membrane proteins using x-ray crystallography require highly stable protein constructs that favor crystallization. Conventional protein engineering strategies have enabled structure solution of many GPCRs 82 . Modifications made to the protein include N- and/or C- terminal truncations as well as mutations that are primarily designed to increase protein stability and enhance expression or to reduce heterogeneity due to post translational modifications such as glycosylation and palmitoylation 83 . Additionally, crystallographers have benefitted from engineered fusion proteins that provide higher proportional polar surface area which contributes to strong directional intermolecular contacts required for forming a crystal lattice 84 . The selected fusion proteins feature minimal internal flexibility and improve the overall crystallization properties of the membrane protein 85 . We benefitted from these protein engineering methodologies in designing a stable and highly expressed GPR6 construct suitable for crystallographic studies. All cell culture and protein expression for this work was done by the Culture Core facility at the Bridge Institute. We used Bac-to-Bac Baculovirus expression system for heterologous expression of GPR6 in insect cells (Spodoptera frugiperda, sf9), purchased from ATCC, CRL-1711, and authenticated by the supplier. Even though mammalian cell lines would be more relevant for studies of human proteins, sf9 cells have proven successful for expressing milligrams of stable GPCRs and have been used in majority of structural studies of human GPCRs since crystallization of first non-rhodopsin GPCR expressed using the same system 82,86 . For each construct, the GPR6 gene of interest was cloned into a pFastBac1 vector (Invitrogen) downstream of an N-terminal hemagglutinin signal peptide used to increase GPCR surface expression by enabling insertion of the receptor’s N-terminal through the endoplasmic reticulum (ER) membrane into its lumen and 21 further trafficking to the plasma membrane 87 . The pFastBac1 vector carrying the gene of interest was then transformed into DH10Bac cells (Gibco) containing a baculovirus shuttle vector (bacmid) to generate the recombinant bacmid. Following the culture of DH10Bac (E. Coli strain) cells containing the recombinant bacmid, the bacmid DNA were purified from cell cultures and transfected into insect cells using X-tremeGENE HP DNA transfection reagent (Roche) to generate a recombinant baculovirus used for expression. The baculoviral stock was then used to infect sf9 cells at a density of ~2x10 6 cells/mL with multiplicity of infection of 5. Cells were then incubated at 27 °C and harvested by centrifugation 48 h after infection and (typically) stored at - 80 °C until use. Frozen cell pellets were thawed and lysed in a hypotonic buffer containing 10 mM HEPES (pH 7.5), 10 mM MgCl2, 20 mM KCl, and 1000 x dilution of a homemade protease inhibitor (PI) cocktail (0.5 M AEBSF, 1 mM E-64, 1.13 mM leupeptin, and 151.36 μM aprotinin) to prevent the proteolytic cleavage of the receptor (used at all wash and solubilization steps with 1000 × dilution). Extensive washing of the raw membranes was performed at 4º C by repeated dounce homogenization and ultracentrifugation at 200,000 x g for 20 minutes (once in hypotonic buffer, and twice in high osmotic buffer containing 1.0 M NaCl, 10 mM HEPES (pH 7.5), 10 mM MgCl2, 20 mM KCl, and homemade protease inhibitor cocktail), thereby separating membrane fraction from soluble. 10-100 mM stocks of ligands were typically dissolved in dimethyl sulfoxide (DMSO) and supplemented in the subsequent steps whenever ligands were used in sample preparation. Washed membranes were resuspended into a buffer containing 100 mM HEPES (pH 7.5) and 800 mM NaCl, supplemented with 2 mg/mL iodoacetamide, as well as 50 μM of the ligand whenever a ligand was used, and incubated at 4 °C for 30 min before solubilization. The same buffer supplemented with detergent was then added to the samples at 1:1 volume ratio and the membranes were solubilized in a buffer with final concentrations of 100 mM HEPES (pH 7.5), 22 800 mM NaCl, 1% (w/v) n-dodecyl-β-D-maltopyranoside (DDM, Anatrace), 0.2% (w/v) cholesteryl hemisuccinate (CHS, Sigma-Aldrich), and 25 μM of the ligand (whenever used) at 4 °C for 2.5 h. Typically, a total of 100 mL solubilization buffer was used per liter of biomass. The supernatant was isolated by centrifugation at 400,000×g for 30 min and incubated in 100 mM HEPES (pH 7.5), 800 mM NaCl, and 20 mM imidazole with Talon (immobilized metal affinity chromatography IMAC) resin (Clontech) (2 mL per 1 L of biomass) overnight at 4°C. For further purification, 25-50 μM of ligand was added to all purification buffers whenever a ligand was used in the experiment. After overnight incubation with TALON, the resin was washed with ten column volumes of wash buffer 1 (50 mM HEPES, pH 7.5; 400 mM NaCl; 10% v/v glycerol; 0.1% w/v DDM; 0.02% w/v CHS; 10 mM imidazole), followed by five column volumes of wash buffer 2 (20 mM HEPES, pH 7.5; 200 mM NaCl; 10% v/v glycerol; 0.05% w/v DDM; 0.01% w/v CHS; 50 mM imidazole). The protein was then eluted in minimal volumes of elution buffer (20 mM HEPES, pH 7.5; 100 mM NaCl; 10% v/v glycerol; 0.02% w/v DDM; 0.01% w/v CHS; 300 mM imidazole) and concentrated to 30-40 mg ml -1 with a 100 kDa molecular mass cut-off Vivaspin centrifuge concentrator (Sartorius). Various biochemical and biophysical methods were used for analyzing the protein throughout the experiments including analytical size-exclusion chromatography (aSEC), sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE), and thermostability profiling using the thiol-specific fluorochrome N-[4-(7-diethylamino-4-methyl-3-coumarinyl) phenyl] maleimide (CPM) 88 . For aSEC, typically 20-30 μL of purified was injected and analyzed on a Nanofilm SEC- 250 column (Sepax: 201250-4630) and run for 20 minutes using a 1260 / 1290 Infinity Series HPLC (Agilent Technologies) with the SEC buffer (20 mM HEPES, pH 7.5; 100 mM NaCl; 10% 23 v/v glycerol; 0.02% w/v DDM; 0.01% w/v CHS) at 0.3 mL/min flow rate. The absorption of sample at 280 nm was measured and analyzed as a measure of homogeneity (monodispersity of the aSEC peak) and protein yield (peak height) which was calculated based on a control sample (BSA) of known concentration. For SDS-PAGE, typically ~5 μg of purified protein was mixed with 4× NuPAGE LDS Sample Buffer (Invitrogen: NP0007) and a final of ~16 μL sample was loaded on a NuPAGE 4 to 12%, Bis-Tris protein gel (Invitrogen: NP0321BOX) alongside 5 μL of Spectra Multicolor Broad Range Protein Ladder (Thermo Scientific: 26623). The proteins would then be separated on the gel during a 50-minute run with ~200 V at RT using MOPS running buffer. The gels were then removed from the cassette and stained with Coomassie G-250 dye to visualize protein bands. For CPM assays, ~10 μg of purified protein was incubated with 5-10 μg/mL of CPM dye in a total sample volume of 100 μL and incubated for 20 minutes at 4ºC before being subjected to temperature ramping (2ºC/min) between 25ºC and 95ºC and fluorescence readout in the Rotor-Gene Q (Qiagen). CPM data were analyzed within the Rotor-Gene Q software and the melting temperature (Tm) for each sample was extracted from the derivative of the fluorescence signal plotted against temperature. 2.1 | Construct evolution Three constructs based on the human GPR6 gene (UniProt ID P46095) — wild type (WT), N- terminal bRIL fusion 83 (N-bRIL), and ICL3 bRIL fusion (ICL3-bRIL) — were obtained from iHuman Institute in Shanghai in October 2015 89 , all with HA signal sequence followed by a FLAG tag at the N-terminus as well as an HRV-3C protease cleavage site followed by a 10× His tag at the C-terminus. Of the two constructs with fusion protein, construct 6393 (ICL3-bRIL) with fusion 24 protein inserted between residues 256 and 269, with 1-43 and 355-362 truncations at the N- and C-termini, respectively, showed a relatively higher surface expression (~43% vs ~17% for N- bRIL), as determined by flow cytometry using the fluorescence signal from the monoclonal ANTI- FLAG ® M2-FITC (Sigma-Aldrich: F4049) 89 . Extensive construct screening for stabilizing effects and increasing yield was done by Linda Johansson, PhD. The ~50 constructs that were screened included point mutations, insertion and/or moving of fusion proteins and tags, and deletion and/or addition of fusion protein junction residues or receptor termini. Among all point mutations, S291 6.47 C of a microswitch based on homology across other receptors of the same branch, GPR3 and GPR12, was found to stabilize the receptor. This mutation together with the junction of the bRIL fusion between residues 256 and 270 (with the SA linker at the C-terminus of the bRIL), 1- 47 N-terminal residue truncation, and an intact C-terminus were combined in construct 6969. Later, supplementing an inverse agonist ligand (#287, IC50GPR6= 31.6 nM) helped stabilize the same construct 6969 even further (Figure 10). Figure 10 | Evolution of GPR6 constructs from October 2015 to February 2016. The human GPR6 gene was expressed, purified, and analyzed by aSEC. The uniform aSEC peak on the right 25 reflects protein monodispersity and stability. All data represented in this figure were acquired by Linda Johansson, PhD. A subsequent round of construct screening (~100 constructs) was followed, covering predicted mutations based on machine learning 90 and homology modeling as well as junction walk of the fusion protein, reaching the construct 8652 (Figure 11). Figure 11 | Construct 8652 stabilized by a ligand 3887 and its characterization. (a) This construct carries S291 6.47 C and Y320 7.44 L mutations and bRIL fusion protein with a C-terminal alanine linker inserted between residues 256 and 270. (b) monodispersity of 8652-3887 analyzed by SEC (acquired by Linda Johansson, PhD). (c) purity of 8652-3887 analyzed by SDS-PAGE. 2.2 | Ligand screening A series of experiments were followed to identify the most stabilizing ligands (Table 1) for subsequent construct screenings and crystallization setups (Figure 12). One of the top stabilizing compounds 3887 which was also abundantly available was chosen for subsequent experiments. 26 Compound 1485 was identified as the most stabilizing ligand but was only used later in crystallization trials due to the limited quantity available (Table 2). Table 1 | GPR6 inverse agonists used in ligand screening. Ligands were provided by Takeda. Figure 12 | Construct 8652 thermostability profile in the presence of various ligands. Construct 8652-apo was screened alongside the construct stabilized by various ligands at 25 μM to compare their thermostability profiles. Protein was solubilized and purified with ligand- supplemented buffers for each sample. 27 Table 2 | Melting temperature (Tm) of 8652 stabilized by various ligands. 2.3 | Initial crystallization trials Given the apparent stability of construct 8652 and its reasonable yield (~ 1.5 mg/mL), attempts were made to crystalize the receptor in lipidic cubic phase (LCP). Protein was purified in the presence of 50 μM ligand 3887 and concentrated to ~15 mg/mL for crystallization screening. The potential crystal hits observed in two crystallization conditions in these experiments were tested at synchrotron facilities, but no obvious diffraction was observed (Figure 13). Subsequent attempts to crystalize 8652-3887 were not successful, which led to a new round of construct and ligand screening. 28 Figure 13 | Potential 8652-3887 microcrystal hits imaged in the cross-polarized mode. Bright birefringent objects were observed in two crystallization conditions: 100 mM sodium hepes, pH 7.0; 30% v/v polyethylene glycol (PEG) 400; and 100 mM a) magnesium nitrate hexahydrate b) magnesium acetate tetrahydrate. 2.4 | Fusion protein screening Next, several different fusion proteins (Table 3) were screened for their potential to increase stability. Additionally, the insertion site of his tag was moved to the receptor’s N-terminus to compare its potential stabilizing effects with constructs bearing the C-terminal his tag. Even though construct 8652 was not included as a control in these experiments, overall no substantial improvement in stability or monodispersity (Figure 14 and Table 4) was observed in the new constructs (compared to previous 8652 data), thus the succeeding construct screenings were built on 8652 bearing ICL3-bRIL fusion and C-terminal his tag. 29 Table 3 | Amino Acid Sequences of Fusion Proteins Screened for GPR6 Constructs 30 Figure 14 | Summary of fusion protein and his tag site screening. Construct 8652 was used to screen the stabilizing effect of various fusion proteins engineered at the N-terminus or ICL3 of the receptor. All experiments were done in the presence of 25 μM ligand 3887. Tm and yield calculated for each construct are summarized in Table 4. 31 Table 4 | Summary of construct yield and Tm from fusion protein and tag screening. 2.5 | Point mutation screening Next, a series of point mutations, predicted through machine learning/homology modeling or based on general GPCR physiology, as well as several N- or C-terminus truncations and fusion protein junction walks were tested (Table 5). All constructs were designed using construct 8652 as the template and all experiments were done in the presence of 25 μM ligand 3887. Point mutations that resulted in either increased thermostability or yield were combined to generate the final construct to be moved to crystallization trials. These point mutations were screened in constructs 10101, 10107, 10112, and 10143; the combined constructive effect of mutations in the first three constructs was observed in construct 10172, and the mutation from the last construct was then added for improved yield of protein, generating construct 10164 (Figure 15 and Table 6). 32 Table 5 | Summary of construct yield and Tm from mutation/truncation/junction screening. 33 Figure 15 | Snake plot of construct 10164 and its comparison with 8652. (a) Four mutations from point mutation screenings were selected to be combined for a final construct to be used in crystallization trials. The point mutations were shown to either have stabilizing effects or improve the yield considerably. (b) 10164 (red) shows monodispersity comparable to 8652 (blue). (c) Thermostability of 10164 is considerably higher than 8652. 34 Table 6 | Comparison of constructs 8652 and 10164. 2.6 | Crystallization ligand Ligand 1485 was previously found to stabilize GPR6 (Table 2). To confirm similar effects of the ligand on the optimized construct, a temperature ramping-SEC experiment was done on apo and ligand-stabilized receptor (Figure 16). Tm of the apo receptor was found to be 45-55ºC while the stabilized receptor-ligand complex showed a Tm of 61-63ºC and was chosen for crystallization. Figure 16 | Temperature ramping-SEC of apo and ligand-stabilized receptor. SEC profiles of samples are shown for a)10164-apo and b)10164-1485. Legend shows temperature (°C) at which sample was incubated for 5 minutes prior to analyzing for monodispersity/aggregation using SEC. 35 Two other ligand series were screened later, one including very close derivatives of ligand 1485 (Table 7), and the other including GPCR consortium ligands (Table 8) in comparison with 1485 and apo construct. Table 7 | GPR6 inverse agonists derivatives of ligand 1485 used in ligand screening. Table 8 | GPR6 consortium ligands used in ligand screening. 36 To assess the stabilizing effect of the ligands, construct 10164 was solubilized in 1% DDM and 0.2% CHS and purified as apo, followed by incubation with 50 μM ligand for 1.5 hours at 4ºC. For 1485 derivatives, subsequent incubation was done at 60ºC or 64ºC for 5 minutes. Samples were then analyzed by their degree of aggregation as indicated in the SEC traces (Figure 17). Figure 17 | Temperature incubation-SEC of 1485 derivative ligand stabilized receptor. SEC traces of construct 10164 are shown after incubation with ligands and heating at a) 60ºC or b) 64ºC for 5 minutes. Legend shows the number-coded ligands for each plot. All experiments included apo, DMSO- and 1485-supplemented samples as controls. Screening of the GPCR consortium ligands was done in a similar manner, but following 1.5-hour incubation with the ligand, samples were incubated only at 60ºC for 5 minutes. Samples were then analyzed by their degree of aggregation as indicated in the SEC traces (Figure 18). 37 Figure 18 | Temperature ramping-SEC of consortium ligand-stabilized receptor. SEC traces of construct 10164 are shown after incubation with ligands and heating at 60ºC for 5 minutes. Legend shows the number-coded ligands for each plot. All experiments included apo, DMSO- and 1485-supplemented samples as controls As for the consortium ligands, five ligands (791, 792, 254, 309, and 296) were observed to have exerted stabilities similar to that of 1485. Derivatives of 1485 showed similar stabilizing effect except for 7420 whose addition to the receptor was found to have no positive effect, thus 1485 was continued to be used as the main ligand in crystallization experiments, while other 10 ligands from both series were tested in crystallization screens of the apo receptor later. 2.7 | Crystallization and Data collection at synchrotron Using construct 10164 along with the stabilizing inverse agonist ligand 1485, we attempted to crystalize GPR6 for structural studies using x-ray crystallography. In each experiment, about 20 μL of protein at a final concentration of 10-40 mg/mL is purified from 1 liter culture of sf9 cells 38 infected with the viral vector carrying GPR6 gene. A 9:1 (w/w) monoolein (Sigma):cholesterol (Sigma) lipid mixture was used to reconstitute this membrane protein in a lipidic cubic phase (LCP) that mimics the protein’s natural environment in the plasma membrane 91,92 . Concentrated protein sample was pipetted into a 100 μL gas-tight syringe (Hamilton) from its plunger side and the plunger was replaced to push the protein towards the tip of the syringe while removing bubbles from the sample. Depending on the volume of the protein sample, the volume of the lipid mixture needed was calculated so the two can be mixed at a 2:3 volume ratio of protein:lipid. The lipid mixture was melted at 40°C prior to mixing with protein and the desired volume was pipetted into a second gas-tight syringe. Mixing was done by connecting the two syringes with a syringe coupler and back and forth strokes of the two plungers, ~ 200 times, to make a homogeneous LCP. We sought to identify crystallization conditions that favored crystal nucleation as a basis for our future optimizations aimed to maximize crystal growth. Initial crystallization experiments of GPR6 receptor were setup with high throughput NT8-LCP crystallization robot (Formulatrix) in 96-well glass sandwich plates (glass plate and cover glass: Paul Marienfeld GmBH & Co, KG) setup with a 150 μm-thick spacer (Saunders, a division of R.S. Hughes Co., Inc.). For each crystallization plate, the NT8 robot was programmed to dispense 40 nL of LCP in 8 wells of each column on the 96-well plate, followed by dispensing 800 nL of precipitant solutions in those 8 wells, and the cycle was repeated for all 12 columns of the plate. Each plate was then sealed with a glass slide, stored at 20°C and imaged using a Rock Imager 1000 (RI 1000, Formulatrix) to monitor crystal growth. This allowed high-throughput screening of hundreds of different crystallization conditions in a single experiment. The microcrystals generated in the initial (and subsequent) successful rounds of crystallization experiments were harvested using micromount 39 loops (MiTeGen), flash frozen in liquid nitrogen, and stored in cryogenic condition to be later screened at synchrotron facilities. All x-ray diffraction and data collection were completed at the GM/CA beamline 23ID-B of the Advanced Photon Source (APS) in the Argonne National Laboratory using an Eiger-16m detector (at 300- or 350-mm distance) with the x-ray wavelength of 1.0332 Å (12 keV). Whenever the crystals were small and hence not easily detectable by the beamline camera, typically for the initial experiments, a rastering strategy 93 was used to test their diffraction. For this purpose, typically the entire loop containing crystals and LCP was outlined and divided to a grid with 10x10 μm cell size. Each cell was then subjected to 1-2 s exposure of unattenuated mini beam (10 μm in diameter) with 0.0-0.2° oscillation per frame. Crystal diffraction of the x-ray beam was then visually analyzed in real time which informed the subsequent optimization of crystallization conditions. For data collection, typically a 5x attenuated beam was used in rastering to minimize radiation damage prior to data collection. This enabled spotting the microcrystals in the loop and centering the beam on them for diffraction data collection. Depending on the size of the crystal, a 5 or 10 μm (in diameter) minibeam was used for data collection and each frame was subjected to 0.2 s exposure of the unattenuated beam with 0.2º oscillation. Data were typically collected in 20-60º wedges, depending on the size of the crystal, until diffraction resolution would start decreasing. Additional data would then be collected on either a new site of the same crystal, or a new crystal. The final rounds of data collection at APS yielded high resolution structures of GPR6 in various conformations (Table 24-25). 40 2.7.1 | Co-crystallization with inverse agonist First crystallization screenings of 10164-1485 were done after cleavage of the C-terminal his-tag (at the HRV-3C protease site) and did not yield promising results. Some potential hits were initially observed (Table 9) but could not be improved or even reproduced in the subsequent experiments. In order to raise the final concentration of the receptor, by minimizing sample loss, and improve sample homogeneity/monodispersity, his-tag cleavage was eliminated from subsequent sample preparations. In a following prep (Table 10), several crystallization screens were designed and used in the experiment in the search for optimal crystallization conditions. Multiple conditions were identified to facilitate receptor crystallization, yielding needle shaped crystals of 20-40 μm in length that appeared within 1-4 days of setting up crystallization plates and grew to the final size within a week of nucleation. Crystals grown in conditions with sodium chloride and ammonium chloride as the salt (Table 11) were harvested and tested for diffraction at APS, and those from the sodium chloride condition showed diffraction of the x-ray beam to beyond 7.7 Å (Table 10-11 and Figure 19). Crystals from conditions with ammonium formate and ammonium nitrate as the salt were also harvested to be tested at later APS beam times but did not diffract as well as those from sodium chloride conditions despite their comparable crystal size (Table 11). Table 9 | Crystallization conditions with potential 10164-1485 crystal hits. 41 Table 10 | Protein yield and concentration for initial 10164-1485 crystallization experiments Table 11 | Crystallization conditions of initial 10164-1485 crystals. 42 Figure 19 | First successful crystallization of 10164-1485. a) Needle shaped crystals of 30-40 μm in length were grown in LCP. b) Crystals obtain in 100 mM sodium chloride at pH 4.0 buffered by 100 mM sodium acetate with 30% (v/v) of PEG400 diffracted to beyond 7.7 Å in rastering at synchrotron. Resolution ring at ~ 7.7 Å is shown. Next, crystallization conditions were optimized based on observed diffraction of crystals in sodium chloride and ammonium chloride salt conditions. Crystals obtained using protein at 19.3 mg/mL concentration in optimized conditions containing sodium chloride as the salt showed very weak diffraction spots up to 3.4 Å, however, no complete dataset could be collected at the time which concluded the Spring 2019 data collection. Diffraction from crystals grown in similar pH (~4) but with ammonium chloride salt was limited to ~ 7 Å and was not improved further (Table 12). 43 Table 12 | Initial optimization of crystallization conditions for 10164-1485 crystals. Subsequent crystallization condition optimizations were expanded to include additive (Hampton Additive Screen: HR2-428) and buffer (Hampton Slice pH Screen: HR2-070) screens as well as quadrantal grid screens to optimize for salt and PEG concentrations. This round of optimization yielded much more orderly packed crystals with larger sizes in all three dimensions. Attempts were made to collect complete diffraction datasets of crystals grown in 13 different conditions. One of these datasets (Table 13-14 and Figure 20) was what led to a structure at an anisotropic resolution of 3.7 Å, 3.7 Å, 3.0 Å along the principal crystal axes. To solve the GPR6-bRIL(10164)-1485 structure, diffraction frames were successfully indexed and integrated using HKL2000 94 and the space group was found to be P3221. LPA1 (PDB ID: 4Z35) structure with all residues replaced with alanine and loops omitted as well as bRIL (PDB ID: 1M6T) were used for a molecular replacement (MR) solution with Phenix Phaser-MR 95 . The structure was subjected to several rounds of model building with Phenix AutoBuild and refinement with Phenix.refine followed by manual refinement in Coot 96 . Although, missing densities of the N- and C-terminal parts as well as ambiguous density around ECL2 and ligand binding pocket made it challenging to reach a complete final structure, all the other loops and helices were modeled with some uncertainty due to a noisy electron density map. 44 Table 13 | Protein yield and concentration for the first successful 10164-1485 crystallization. Table 14 | Optimized crystallization condition of 10164-1485 crystals. Figure 20 | Successful crystallization of 10164-1485 in optimized conditions. a) Hexagonal rod crystals of 40-60 μm in length were grown in LCP at 36.2 mg/mL concentration of the receptor. b) Apparent diffraction up to 3.6 Å was observed in diffraction snapshots of crystals grown in 100 mM sodium chloride at pH 4.1 buffered by 100 mM sodium acetate with 30% (v/v) of PEG400 and in the presence of 50 mM potassium chloride. 45 The incomplete structure necessitated a higher resolution dataset. Thus, crystallization conditions were further optimized around precipitant solutions containing sodium chloride and potassium chloride salts. This was accompanied by some modifications to the receptor to remove tags with the intention of minimizing flexible parts of the construct. All constructs 10164, 10563 (missing HRV 3C cleavage site), and 10567 (missing both HRV 3C cleavage site and FLAG tag) were tested in crystallization experiments with successful crystal hits to be further analyzed at APS. Of all the samples tested for diffraction, thousands of frames were collected on a single large hexagonal crystal from 10567-1485 which ultimately yielded the final structure with 2.6 Å (anisotropic) resolution (Table 15-16 and Figure 21). Table 15 | Protein yield and concentration for the final 10567-1485 crystallization. Table 16 | Optimized crystallization condition of 10567-1485 crystals. 46 Figure 21 | Successful crystallization of 10567-1485 in optimized conditions. a) 10567-1485 crystallization in LCP at 24.5 mg/mL concentration of the receptor. The insert shows the same crystal harvested on a micromount loop for data collection at APS. b) Apparent diffraction up to 2.6 Å was observed in diffraction snapshots of crystals grown in 100 mM sodium chloride, 100 mM sodium acetate pH 4.1, 30 %v/v PEG400, 2 %v/v PPG P400. Best datasets were successfully indexed and integrated using HKL2000. The resolution was cut to 3.0 Å (a*), 3.0 Å (b*), and 2.6 Å (c*) using STARANISO 97 after anisotropic correction. Using this new dataset and starting from the previous model of GPR6 (from 3Å dataset), we were able to complete the unmodeled loops and improve the structure by modeling more of the amino acid side chains that were not well resolved before. The data was found to be ~40% twinned, thus the twinning operator (-h, -k, l) was used for twinning refinement in phenix.refine. Following iterative rounds of manual model building in coot and subsequent refinement with phenix.refine, the final structure of GPR6-1485 was refined to R work/R free of 24.4 % and 26.5 %, respectively (Table 24). 47 2.7.2 | Ligand screening in crystallization trials In a parallel effort for structure solution of GPR6, ten stabilizing ligands were screened in the crystallization trials. These screenings included all derivatives of 1485, except for 7420, (Figure 17) and the five most stabilizing consortium ligands (791, 792, 254, 309, and 296) (Figure 18). Construct 10164 was solubilized and purified as apo and concentrated to 69.4 mg/mL for crystallization. Screens were designed based on previous successful crystallization of this receptor and included conditions of varying concentrations of sodium and potassium chloride with sodium acetate as the buffering agent at pH 4.2. Ligands were added to the screens at a final concentration of 50 μM and were tested in duplicates for reproducibility, while similar conditions lacking any ligand served as negative controls. All stabilizing derivatives of ligand 1485 yielded crystals while the negative controls remained free of any signs of crystal nucleation. Three out of five consortium ligands (792, 791, and 309) enabled successful crystallization of the receptor (Figure 22). Figure 22 | GPR6 construct 10164 co-crystalized with 3 different consortium ligands. Successful co-crystallization of 10164 with ligands a) 792, b) 791, and c) 309. Crystals of 10164- 309 appeared after 1 day, while crystal nucleation for 10164-791 and 10164-792 took over 1 week. 48 Crystals obtained in the presence of ligands 792 and 791 were harvested due to their larger size and tested at APS (June 2019), showing weak and low-resolution diffraction; these spots were extending to 6 Å in case of 791. Thus, a subsequent experiment was conducted using this ligand for co-crystallization with the receptor. Given the limited availability of ligand 791, construct 10164 was solubilized and purified as apo. Ligand 791 was added to the elution buffer and supplemented in the buffers there on. 10164-791 was concentrated to 33.1 mg/mL and used for crystallization screening in over six hundred conditions yet did not yield any crystals. Though puzzling at the time, later experiments and observations proved that ligand substitution is only possible when the receptor is in its natural environment of membrane lipid bilayer or reconstituted in the lipidic cubic phase, but not when it’s solubilized in the detergent micelles. Due to limited availability of this ligand (791), further optimization of crystallization trials with 791 was not pursued. The following experiment were conducted with 792 as the stabilizing small molecule ligand which was relatively more abundantly available. The final crystallization construct 10567 was solubilized and purified in the presence of ligand 792. 10567-792 was concentrated (Table 17) and used for crystallization in over a thousand different precipitant conditions. Needle-shaped crystals from two of these conditions (Table 18 and Figure 23) were harvested and screened for diffraction at synchrotron but were not able to diffract the beam to high resolution (September 2019). Table 17 | Protein yield and concentration for 10567-792 crystallization 49 Table 18 | Best crystallization conditions of 10567-792 crystals. Crystals of 10567-792 were later produced and used for screening of diffraction at X-ray Free- Electron Laser (XFEL) sources (Korean PAL-XFEL) (September 2019). Although high resolution diffraction of the beam was observed for these samples, not enough crystals had been prepared to collect a complete dataset for structure solution. Due to limited availability of the ligands, further experiments with the consortium ligands were not pursued. Figure 23 | Crystallization of GPR6 construct 10567-792. Cross-polarized images of 10567-792 crystals obtained in a) sodium citrate tribasic dihydrate and b) sodium chloride. Complete composition of the precipitant solutions is listed in Table 18. 50 2.7.3 | Crystallization of apo GPR6 In order to gain a deeper insight into the structure and function of GPR6, the architecture of its binding pocket, and ligand entry, we sought to crystalize the apo receptor. Construct 10567 was purified as apo (Table 19) and used in crystallization screenings. Reported values for yield and final concentration are for two separate experiments. Several conditions were found to accommodate crystallization of apo receptor (Table 20), and crystals from these experiments were used for synchrotron data collection and structure determination (Figure 24). Table 19 | Protein yield and concentration for 10567-apo crystallization. Table 20 | Crystallization conditions of 10567-apo crystals. 51 Figure 24 | Crystallization of GPR6 construct 10567-apo. a) representative crystals and b) representative diffraction pattern. Various crystals had diffractions to resolutions better than 2.5 Å during APS data collection. Thousands of diffraction frames were collected on eleven 10567-apo crystals grown in different conditions including salts potassium sodium tartrate tetrahydrate, sodium tartrate dibasic dihydrate, sodium formate, and potassium formate (Table 20). 10 μm (in diameter) beam was used for data collection and each frame was subjected to 0.2 s exposure of the unattenuated beam with 0.2º oscillation, with the detector set at 300 mm distance. Data were collected in 20-60º wedges, depending on the size of the crystal, until diffraction resolution would start decreasing. Additional data would then be collected on either a new site of the same crystal, or a new crystal. Several of these datasets collected on four crystals from two crystallization conditions with salts potassium sodium tartrate tetrahydrate and sodium tartrate dibasic dihydrate (Table 20) were successfully indexed, integrated, scaled and merged using HKL2000 94 to yield a final reflection file (space 52 group P21212) with diffraction data up to 2.1 Å resolution. Previously solved structure of GPR6 was used for MR. The structure was subjected to iterative rounds of manual model building in coot and refinement in Phenix. Translation/Libration/Screw (TLS) refinement was used in the later stages of refinement with twelve TLS groups. The final structure was refined to Rwork/Rfree of 0.19% and 0.23%, respectively (Table 25). Upon structure solution of GPR6-apo and its comparison with GPR6-1485 as well as other GPCRs’ structure in active and inactive conformations, GPR6-apo structure was determined to be in active conformation. Our structure of 10567-apo in the active conformation revealed the presence of a lipid-like molecule in the orthosteric pocket despite the absence of exogenous ligands throughout the crystallographic experiments (Figure 25). This observation informed several of future experiments to validate our structural findings (discussed below and in chapters 3 and 4 of this dissertation). Figure 25 | Density of the lipid-like ligand in the orthosteric pocket of the GPR6 active structure. 2Fo-Fc map is shown for GPR6 transmembrane helices (grey), and the lipid-like molecule observed in the orthosteric pocket (green). Map is contoured at 1.0 σ. 53 2.7.4 | Co-crystallization with S1P Following several claims of S1P modulating GPR6 signaling, we speculated that the ligand co- purified and -crystalized with GPR6 could be S1P 55,57,59 , although the density of the headgroup of the co-crystalized ligand did not seem to accommodate the bulky phosphate head group of S1P. We decided to repeat the crystallization experiments, this time supplementing S1P in the lipid mixture for LCP formation. We theorized that if S1P is in fact a strong binder of GPR6, its presence in the lipid mixture should allow its replacing of the ligand co-purified with the receptor given the co-purified ligand is most likely not as abundantly available as S1P to rebind the receptor after dissociation and will instead be replaced by S1P in the orthosteric binding pocket. GPR6 (construct 10567) was purified apo and mixed with a lipid mixture containing S1P (Cayman Chemical) (w/w MO:Chol:S1P ~ 89:10:1) to form the LCP (with ~15 mM S1P) for high throughput crystallization screening in plates. Crystals were obtained in several different conditions (Table 21) and harvested to screen for diffraction at APS. Complete datasets were successfully collected on 25 crystals with diffractions beyond 2.5 Å. Data collected on a single crystal from the sodium formate salt condition (Table 21) was indexed, integrated, scaled, and merged in HKL2000 94 and ultimately yielded a reflection file containing data with 2.3 Å resolution (Table 25). The structure was solved via MR using the previously solved pseudo-apo GPR6 model. The overall structure and the density of the ligand in the orthosteric pocket were found to be very similar to our previous pseudo-apo structure (discussed in the next chapter). Hence, this crystallographic experiment did not validate binding of S1P to GPR6 and its role as a functional agonist, stabilizing the receptor in its active conformation. 54 Table 21 | Crystallization conditions of 10567-S1P crystals. 2.7.5 | Crystallization of apo GPR6 expressed in mammalian cells We speculated that the ligand of the orthosteric pocket may have been carried on with the receptor from the cells, if not introduced to the sample from the LCP. Given the ongoing debate about GPR6 endogenous agonist, we sought to investigate whether a similar lipid-like molecule is co- purified with the receptor from a mammalian expression system. Confirming the presence of a co- purified lipid from mammalian cells would support the relevance of our observations to shed light on the activation mechanism of human GPR6. We used the codon-optimized gene of GPR6 crystallization construct 10567 (10917) for expression in the Human Embryonic Kidney cells (HEK293). Construct 10917 was cloned in a pcDNA3.1(-) vector and used for transfection (with 293fectin transfection reagent) of 1L HEK293 (Thermo Fisher Scientific) suspension cell culture (1.0 μg of plasmid DNA per mL of culture volume). Cells were shaken at 125 rpm at 37 °C with a humidified atmosphere of 5% CO2 in air for 48 hours and harvested by centrifugation, then 55 frozen until further use. The receptor was purified apo as before and setup for high throughput crystallization. Crystals were grown in conditions similar to those facilitating pseudo-apo crystallization (Table 22). Using data collection strategy as before, collecting diffraction data on two crystals from the same condition (sodium tartrate dibasic dihydrate buffered at pH 5.9 by DL- Malic acid, Table 22) at APS, and further data processing with HKL2000 yielded a 2.6 Å dataset (Table 24). Performing MR using previously solved pseudo-apo GPR6 model enabled structure solution and showed the receptor in an active conformation with a lipid-like ligand stabilizing the receptor, much similar to the pseudo-apo structure obtained from the receptor expressed in sf9 cells. This observation suggested that a naturally occurring ligand, present in both sf9 and HEK293 cells, might be the endogenous agonist of GPR6 and copurify with the receptor, unless the ligand is introduced to the sample during crystallization from the LCP. Table 22 | Crystallization conditions of 10917-apo crystals 2.7.6 | Structure solution of GPR6-CVN424 Our structural studies of GPR6 at this time was concurrent with advances in phase 2 clinical trials 98 , led by Cerevance, of a small molecule inverse agonist that modulates GPR6 activity. This 56 small molecule CVN424 has been shown to restore motor activity in patients of PD by inhibiting GPR6 constitutive signaling in the basal ganglia. CVN424 is the only drug against GPR6 with proven effectiveness in preclinical models and currently in clinical screening for potential use for patients of PD 81,99 . In a collaboration with Cerevance, we attempted to obtain the structure of GPR6 co-crystalized with CVN424 (Table 23), with several successful hits in all the screens. Using data collection strategies previously described, a complete dataset was collected on a single crystal from the condition with 150 mM sodium chloride, buffered at pH 4.2 by sodium acetate (Table 23). Diffraction data was processed in HKL2000 as before and a successful MR solution was generated using previously solved model of inactive GPR6. The 3.5 Å resolution structure of the GPR6- CVN424 was subjected to iterative rounds of manual refinement in coot followed by refine.phenix, and the final structure was refined to Rwork/Rfree of 0.295 and 0.320, respectively (Table 24). The structure revealed the receptor in its inactive conformation, as expected, and was stabilized with CVN424 in the orthosteric binding pocket. Detailed analysis of this structure alongside the previously described structures will be included in the chapter 3 of this dissertation. Table 23 | Crystallization conditions of 10567-CVN424 crystals 57 2.8 | Conclusion This chapter outlined series of efforts towards structural studies of GPR6. Overall, 6 complete datasets were collected on different samples of GPR6 each of which enabled the structure solution of the receptor in active or inactive conformations. The two structures of GPR6 in its inactive state were solved by co-crystallization with small molecule inverse agonists 1485 and CVN424 provided by Takeda Pharmaceuticals and Cerevance, respectively. The three structures of GPR6 in its active conformation were solved by crystalizing the receptor either without addition of any exogenous ligand or addition of S1P as the potential agonist. All three structures were solved with diffraction data collected APS. Altogether, the three structures share very similar structural features and univocally reveal the presence of a lipid-like molecule that stabilizes the active conformation of the receptor from the orthosteric pocket. The density associated with this ligand seems to be identical regardless of whether or not S1P was exogenously introduced to the sample during the experiments and whether the receptor was expressed in sf9 or HEK293 cells. Although the electron density of this lipid-like molecule did not suffice to confidently identify the ligand, it did inform a series of follow-up experiments that were designed to help discover this ligand and pave the path towards GPR6 deorphanization. The crystallography data and refinement statistics for the finalized structures highlighted in this chapter are summarized at the end of the chapter (Tables 24-25). Structural analysis and insights from these experiments will be discussed in depth in the third chapter of this manuscript. 58 Table 24 | Data collection and refinement statistics related to GPR6-IAG structures. 59 Table 25 | Data collection and refinement statistics - GPR6 active conformation structures. 60 Chapter 3 | Structure of GPR6 Our crystallographic experiments, as thoroughly discussed in chapter 2, enabled structured solution of GPR6 in both active and inactive conformations. In this chapter, structural features of GPR6 in the two main conformations will be described. The structure analyses will then extend to compare different conformations of the receptor as well as similar conformations obtained through various experimental procedures. 3.1 | Structure of inactive GPR6 in complex with the inverse agonist 1485 The first structure of GPR6 was obtained by co-crystallization with the small molecule inverse agonist 1485. This structure revealed the overall topology of GPR6 as well as the architecture of its orthosteric binding pocket. GPR6 displays the canonical seven transmembrane α-helical arrangement with three intracellular loops (ICL1-ICL3; where bRIL replaces ICL3), three extracellular loops (ECL1-ECL3), and an amphipathic helix 8. Receptor’s extracellular N terminus and the short intracellular C-terminus were not resolved in our structures due to flexibility of these regions (Figure 26). An intra-ECL2 disulfide bond, as depicted in GPR6 snake plot (Figure 6), preserves the structural integrity of this receptor, and is evident in the structure (Figure 26), similar to other lipid receptors of the MECA cluster 100,101 . Well-resolved density of the small molecule ligand in the orthosteric binding pocket allowed for modeling of the inverse agonist 1485 and revealed the overall architecture of the binding pocket, consisting of primarily hydrophobic and aromatic residues (Figure 27). Further, surface representation of the receptor indicates top opening of the binding pocket, making it accessible from the extracellular milieu (Figure 27). 61 Figure 26 | Crystal structure of GPR6 10567-1485. The structure depicts GPR6 in its inactive conformation. bRIL fusion between TMs V and VI is replaced with a dashed line and not shown in the figure. Approximate plasma membrane boundaries are shown in dotted discs; membrane boundaries for LPA1 (PDB 7TD0) were obtained from OPM 102 and approximated for GPR6 by aligning GPR6 with LPA1. Ligand 1485 is shown in lemon spheres. (a) Intra-ECL2 loop disulfide bond which preserves the structural integrity of GPR6 is established between C209 and C216. (b) Small molecule inverse agonist 1485 modeled according to its clear electron density in the orthosteric pocket. 2Fo-Fc electron density (blue mesh) is shown around the ligand and contoured at 1σ. (c) Surface representation of the receptor at 90º shows the opening of the orthosteric pocket to the extra cellular milieu. 62 Figure 27 | Architecture of the orthosteric binding pocket in 10567-1485. The binding pocket of GPR6 consists of primarily hydrophobic and aromatic residues, predominantly leucine and valine. Key residues that maintain stabilizing π-stacking interactions with the inverse agonist 1485 include H128 2.60 , F152 3.36 , F295 6.51 , and W386 6.48 . (a) Residues that interact with 1485 within 4 Å are labeled and their sidechains are shown in stick representation. (b) 2D plot of the residues that line GPR6 pocket and interact with 1485 and their relative location around the ligand. π- stacking interaction are shown in lemon dotted lines. 63 3.2 | Structure of inactive GPR6 in complex with the inverse agonist CVN424 Co-crystallization of GPR6 10567 with the receptor’s potent inverse agonist CVN424, currently in clinical trials, allowed for structure solution of the receptor at 3.5 Å resolution which depicted the receptor in its inactive conformation, similar to 10567-1485 with the root mean square deviation (RMSDCα) of 0.41 Å (after alignment of 218 Cɑ atoms from either of the models). The clear electron density of CVN424 in the binding pocket allowed for confident modeling of this small molecule and investigating its interactions, although the electron density was less resolved as compared to the density of 1485, consistent with the overall lower resolution of 10567-CVN424 (3.5 Å) compared to 10567-1485 (2.6 Å) (Figure 28). Binding pose of CVN424 in the receptor’s orthosteric pocket was found to be similar to 1485, stabilized by several primarily hydrophobic and π-stacking interactions with the surrounding residues (Figure 29). Residues lining the pocket and interacting with the small molecules 1485 and CVN424 are summarized in Table 26. 64 Figure 28 | Crystal structure of GPR6 10567-CVN424. Structure of GPR6 is stabilized in the inactive conformation in the presence of the small molecule inverse agonist CVN424, shown in cyan spheres. Approximate membrane boundaries are shown in dotted disc; membrane boundaries for LPA1 (PDB 7TD0) were obtained from OPM 102 and approximated for GPR6 by aligning GPR6 with LPA1. (a) Clear density of the ligand enabled modeling and refinement of CVN424 in the receptor’s orthosteric pocket. 2Fo-Fc electron density (blue mesh) is shown around the ligand and contoured at 1 σ. (b) Surface representation of the receptor from 90º shows accessibility of the orthosteric pocket from the extracellular milieu. 65 Figure 29 | Architecture of the orthosteric binding pocket in 10567-CVN424. Similar hydrophobic residues that stabilize 1485 were found to interact with CVN424. A hydrogen bond between CVN424 and H128 2.60 seems to replace the π-stacking interaction observed between 1485 and H128 2.60 in the 10567-1485 structure. (a) Residues that interact with CVN424 are labeled and their sidechains are shown in stick representation. (b) 2D plot of the residues that line GPR6 pocket and interact with CVN424 within 4 Å and their relative location around the ligand. π-stacking interactions and hydrogen-bonding are shown in dashed purple and lemon lines, respectively. 66 Table 26 | GPR6 residues of the orthosteric pocket at 4 Å distance of the ligand. CVN424 is a highly potent and selective inverse agonist of GPR6 with binding affinity (Ki) of 9.4 nM and functional effective concentration (EC50) of 55.1 nM 81,99 . This same compound has significantly lower potency for GPR3 (35-fold) and GPR12 (126-fold) with EC50 values of 1921 nM and 6965 nM, respectively 99 . Comparison of the binding pocket residues between the three receptors (Table 27) indicates higher sequence identity between GPR6 and GPR3 (73.3%) than GPR12 (66.7%), which is consistent with the observed potency of CVN424 for these different receptors. These insights along with our structures of GPR6 in its inactive state can inform future drug designs of selective inverse agonists for GPR6, as well as its paralogs GPR3 and GPR12. 67 Table 27 | Comparison of GPR6 residues within 4 Å of CVN424 with GPR3 and GPR12. 3.3 | Structure of active GPR6 in complex with a putative endogenous ligand Crystallization of apo GPR6 led to structure solution of the receptor in complex with a lipid like molecule that stabilizes the receptor in its active conformation (GPR6-pseudoapo). Superimposition of the active-state GPR6 onto its inactive-state structure (10567-1485) with RMSDCα of 1.32 Å (205 to 205 atoms) highlights global conformational changes between the two states. Outward movement of TM6 by 8.2 Å and TM5 by 5.3 Å as well as slight inward movement of TM7 by 3.2 Å are clear indications of receptor’s transition towards an active conformation (Figure 30). 68 Figure 30 | Global conformational changes upon receptor activation. Superimposition of inactive (gray) GPR6 (10567-1485) onto its active conformation (blue) highlights receptor’s conformational changes upon activation. (a) shows a side view of the superimposed structures, depicting distinct global conformational changes including kinking of TM5, outward movement of TM6 as well as contraction of ECL2. (b) shows the cytoplasmic view of the receptor with notable helix shifts measured for TMs 5, 6, and 7 as 5.5, 8.2 and 3.7 Å, respectively. The distances were measured between the Cɑ of residues A256 5.69 (TM5), H270 6.26 (TM6), and Y329 7.53 (TM7). Although the identity of the ligand in our pseudo-apo structure remains elusive, we were able to model a simple lipid named oleic acid in this density and refine the structure to investigate potential ligand interactions and further analyze the receptor’s activation mechanism. Oleic acid (OLA) is an 18-carbon-chain monounsaturated fatty acid with a carboxyl headgroup that fits this density 69 well (Figure 31). Additionally, in the active state structure of GPR6, a cholesterol molecule, amongst other 18 annular lipids, is identified which mainly interacts with W193 4.50 , at a GPCR cholesterol site first observed on the surface of the β2-adrenergic receptor 103 . Additional hydrophobic interactions with V109 2.41 , G112 2.44 , S113 2.45 , T116 2.48 , F147 3.31 , A150 3.34 , S151 3.35 , and L197 4.54 stabilizes cholesterol in the groove of TMs 2, 3 and 4 (Figure 31). High resolution of the map also allowed the placement of several solvent molecules, including those residing in a hydrophilic channel (occupied with an elongated molecule modeled as tetra-ethylene glycol; PG4) as well as the stable waters intermediating interactions between the helices. Inspection of the binding pocket and residues surrounding the modeled OLA (from here-on referred to as Lig-OLA) revealed a scheme of interacting residues similar to that of the inverse agonists (Figure 32). Comparing these residues with those interacting with the inverse agonist 1485 identified several residues in ECL2 and TMs 5-7 with specific interactions with Lig-OLA (and not 1485) (Table 28). Four residues of ECL2 (V218, V219, R220, and L222) establish interactions with Lig-OLA, however, the conformation of this region of ECL2 is fairly similar in both active and inactive conformations. 70 Figure 31 | Crystal structure of GPR6 pseudo-apo . High resolution of the electron density map allowed for modeling of several annular lipids (grey) including OLA and 1-oleoyl-R-glycerol (OLC), as well as a cholesterol molecule (purple) that surround the receptor in its active state. Additionally, a PG4 molecule that is modeled in the hydrophilic channel towards the cytoplasmic side of the receptor is shown in orange. (a) Lig-OLA and an 11-carbon saturated aliphatic chain are tentatively modeled in the ligand density of the orthosteric pocket and the disconnected density 71 that follows Lig-OLA from the lipid bilayer into GPR6 binding pocket, respectively. 2Fo-Fc electron density (blue mesh) is shown around the ligand and contoured at 1 σ. (b) Stable binding of cholesterol (purple) in the groove of TMs 2, 3 and 4 is stabilized by several hydrophobic interactions with the surrounding residues as well as hydrogen bonding with the backbone of V109 2.41 (dashed yellow). (c) Cross section of GPR6 binding pocket in surface representation shows the extension of this pocket through a channel that connects the pocket to the membrane. (d) 90º view of the receptor shows the opening of the ligand-binding pocket to the extracellular milieu, similar to the inactive conformation. (e) 130º rotation along the y axis shows the opening in the transmembrane domain of the receptor from the ligand-binding pocket to the membrane in the active conformation which is absent in the inactive conformation of GPR6. 72 Figure 32 | Architecture of the orthosteric binding pocket in GPR6-pseudoapo. Primarily- hydrophobic residues lining GPR6 binding pocket interact with the acyl chain of Lig-OLA. Residues of ECL2 maintain hydrogen bonding and ionic interaction with the headgroup of the ligand. (a) Top view of GPR6 binding pocket in GPR6-pseudoapo structure. Interacting residues are labeled and their sidechains are shown in the stick representation. (b) 2D plot Lig-OLA 73 interactions with residues of GPR6 binding pocket and their relative position to Lig-OLA is represented. Hydrogen bonds with the head group of Lig-OLA are shown in lemon dashed lines, and the salt bridge between R220 ECL2 and Lig-OLA is shown with blue dashed line. Table 28 | Comparison of GPR6 residues within 4 Å of Lig-OLA and/or 1485. 74 3.4 | Structure comparison of GPR6-pseudoapo and other active-conformation structures The subsequent structures of GPR6 in its active conformation served to provide more insight into the identity and origin of the orthosteric pocket ligand that was tentatively modeled as OLA. Attempts to co-crystallize GPR6 with S1P based on previous claims that S1P is an agonist of this receptor 57,59 were made to investigate the potential difference in the density of orthosteric ligand upon structure determination. Further, expressing GPR6 in HEK293 cells and its crystallization following the same protocol used for GPR6-pseudoapo structure served to clarify whether the same ligand is carried on with the receptor from a mammalian cell line. The results of these experiments and their interpretation are described below. 3.4.1 | GPR6-pseudoapo comparison with GPR6-S1P Structure solution of GPR6-S1P revealed the receptor in its active conformation with a seemingly identical ligand density to that observed in GPR6-pseudoapo structure. The electron density difference map for these two structures was calculated to investigate whether the electron density for co-crystallized ligand is the same or different between these two structures (Figure 33). The difference electron density map confirmed the ligand densities of the two structures to be similar, with no significant extra density of the ligand headgroup in GPR6-S1P that could accommodate the ligand’s phosphate group (Figure 34). 75 Figure 33 | Chemical structure of OLA and S1P. Both OLA and S1P shown here have an 18- carbon fatty acyl chain that fits in the long electron density of the lipid tail in our structures. Comparison of the two ligands identifies the bigger headgroup of S1P ligand that cannot be accommodated in either of GPR6-pseudoapo or GPR6-S1P structures. Figure 34 | Electron density difference map for GPR6-pseudoapo and GPR6-S1P. Electron density of the difference map is visualized within 10 Å radius of Lig-OLA modeled in the GPR6- pseudoapo structure. The meshed Fo-Fo electron density is contoured at +/-3.0 σ (green/red). 76 Three possible conclusions could be drawn from our observations. 1) S1P is in fact the co- crystalized ligand in the GPR6-pseudoapo structure, but the lack of strong and stable interactions of the S1P headgroup with the surrounding residues results in relative flexibility of this headgroup as compared to the fatty acyl chain, which leads to relatively weak average electron density observed for the headgroup. 2) The observed electron density in the orthosteric ligand pocket of the GPR6-pseudoapo structure, in fact, belongs to OLA that is abundantly present in the crystallization experiments as a hydrolyzed form of monoolein (Figure 35). In this case, it would be possible that S1P, even if bound to GPR6 in the purified receptor, could be replaced with OLA during receptor crystallization. 3) An unknown lipid-like ligand bound in the orthosteric pocket is co-purified with GPR6, and S1P is unable to replace it. Figure 35 | Hydrolysis of Monoolein. Hydrolysis of monoolein which is the primary lipid used in our crystallization setups yields OLA and glycerol. 3.4.2 | GPR6-pseudoapo comparison with GPR6-HEK Structure solution of GPR6-HEK-pseudoapo revealed the receptor in its active conformation as expected. The overall electron density map as well as the binding pocket’s local electron density were identical to that obtained from GPR6-pseudoapo. The electron density difference map for these two structures (Figure 36) indicates the orthosteric ligand is likely the same in both. Two possible conclusions from this observation are 1) the same lipid exists in both sf9 and HEK293 77 cells and is co- purified and -crystalized with GPR6 resulting in an identical electron density of the ligand in the receptor’s binding pocket, or 2) a different lipid may have been co-purified from HEK293 cells but replaced with OLA during crystallization. Figure 36 | Electron density difference map for GPR6-pseudoapo and GPR6-HEK- pseudoapo. Electron density of the difference map is visualized within 10 Å radius of Lig-OLA modeled in the GPR6-pseudoapo structure. The meshed Fo-Fo electron density is contoured at +/- 3.0 σ (green/red). 78 3.5 | Mechanism of GPR6 activation Our highest resolution structures for both active and inactive GPR6 (GPR6-pseudoapo and GPR6- 1485, respectively) were juxtaposed to reveal molecular mechanisms underlying GPR6 activation. Comparison of the two states highlights notable conformational changes on the extracellular side of the receptor, throughout the transmembrane region, and all the way towards the cytoplasmic side of GPR6. On the extracellular side, movement of residues 210-215 of ECL2 allow for new interactions between the backbone of 4 residues in this region. These interactions, established between E212 ECL2 and A215 ECL2 as well as R213 ECL2 and C216 ECL2 , seem to stabilize the inward hinging of the middle part of ECL2 towards the orthosteric pocket (Figure 37). The ultimate contraction of ECL2 and ECL3 is stabilized by hydrogen bonding of R220 ECL2 side chain with the backbone of H304 ECL3 . Additionally, hydrophobic contacts of H304 ECL3 with A214 ECL2 , C216 ECL2 , and S217 ECL2 further conserve this contraction. The sidechain of H304 ECL3 in the receptor’s inactive conformation is held back by stabilizing polar interactions with the side chain of E305 ECL3 . R220 ECL2 also maintains interaction, through its backbone, with the headgroup of the Lig-OLA, and, through its sidechain, with the backbone of S217 ECL2 (Figure 39). The intra- and inter-loop interactions of ECL2 and ECL3 contribute to the overall conformational stability of these loops and enable stable interactions of ECL2 residues with Lig-OLA. 79 Figure 37 | Conformational changes of extra-cellular loops upon receptor activation. Intra- and inter-loop interactions contribute to conformational changes observed in the extracellular loops 2 and 3. Active and inactive states of the receptor are colored blue and grey, respectively, in all the panels. (a) Hydrogen-bonding of the backbones of E212 ECL2 and A215 ECL2 as well as R213 ECL2 and C216 ECL2 enable the inward hinging of ECL2 towards the orthosteric pocket. (b) Interactions of the sidechain of R220 ECL2 with the backbone of S217 ECL2 and the backbone of H304 ECL3 seems to enable the contraction of these two loops. Residues T311 7.35 and Y312 7.36 maintain hydrophobic interactions with Lig-OLA and contribute to shifting of the extracellular tip of TM7 toward the ligand binding pocket (2.9 Å, measured between Cɑ of A308 7.32 ) which is likely one of the triggering events of activation. This inward shift extends throughout the TM7 and on the cytoplasmic side, is stabilized with new interactions that form in the active state of the receptor (discussed later). Towards the bottom of the ligand-binding pocket, the !-stacking interactions of F152 3.36 with F234 5.47 and W292 6.48 (toggle switch) from CWxP motif, are broken by not Lig-OLA but a second lipid that protrudes its tail into the binding pocket (Figure 38). This allows for a downward shift of W292 6.48 as well as a rotation of F234 5.47 side chain in the active state while both residues maintain stabilizing interactions with Lig-OLA. 80 The downward shift of W292 6.48 is accompanied by rewiring of other hydrophobic and !-stacking interactions between TMs 3, 5 and 6, including rearrangements in residues of the P 5.50 I 4.40 F 6.44 (V 5.50 and V 4.40 in GPR6) motif. In the active state, the hydrophobic lock between F288 6.44 and TM3 residues S155 3.39 (Na + pocket) and V156 3.40 breaks and allows for outward movement of TMs 5 and 6, distancing from TM3. F288 6.44 of the active state instead maintains stabilizing !- stacking interactions with F238 5.51 , allowing outward movement of TMs 5 and 6 (Figure 30). Additionally, while we do not see a strong density of Na + ion in the Na + pocket of inactive GPR6, in the active conformation, rotation of S155 3.39 towards the Na + pocket together with the inward shifting of N325 7.49 (NPxxY motif), both stabilized by strong polar interactions with D118 2.50 (Na + pocket), seems to result in the collapse of this pocket (Figure 38). Conformational rearrangements of residues and TMs extend to the cytoplasmic end of the receptor and are reinforced by disrupted interactions between TMs 3 and 6 in this region, replaced by stabilizing interactions between TMs 3 and 7. Accordingly, polar interaction of R166 3.50 (DRY motif) with the backbone of G277 6.33 in the inactive GPR6 is broken and further facilitates outward movement of TM6, distancing from TM3. R166 3.50 of the active conformation is instead stabilized by hydrogen bonding with Y329 7.53 from the NPxxY motif, enabled by shifting of TM7 towards the receptor core (Figure 38). Collective of these molecular rearrangements ultimately open the cytoplasmic end of the receptor and enable its interaction with the G proteins. 81 Figure 38 | Molecular microswitches involved in receptor activation. Agonist binding and its interactions with residues of the binding pocket induce conformational changes along the transmembrane region of the receptor. These signals are propagated towards the cytoplasmic side of the receptor, with new interactions forming to stabilize the active state. Active and inactive states of the receptor are colored blue and grey, respectively, in all the panels. Water molecules in the active state are shown as red spheres. Hydrogen bonds in the active and inactive states are shown with dashed yellow and magenta lines, respectively. (a) F152 3.36 is one of the major 82 molecular microswitches with 137-degree inward rotation that opens up the orthosteric pocket to the transmembrane environment through TMs 3-5. (b) !-stacking stabilizes the active and inactive conformations. Residues interacting with the residue F288 6.44 in both active and inactive conformations are shown. In the active state, interaction of F288 6.44 with S155 3.39 and V156 3.40 (present in the inactive state) breaks. F288 6.44 of the active state maintains stabilizing !-stacking interactions with F238 5.51 . (c) Rearrangements of Na + pocket residues including S155 3.39 and N325 7.49 result in collapse of this pocket in the active conformation of GPR6. (d) Stable interaction between R166 3.50 of the DRY motif and G277 6.33 in the inactive state is replaced by interaction of R166 3.50 with Y329 7.53 from the NPxxY motif that stabilizes the active state. 3.6 | Molecular mechanisms of inverse agonism in GPR6 The opening between TMs 3 and 5 towards the tail of the cognate lipid (Figure 31) suggests the entry of this ligand from the membrane which is compatible with the inherent hydrophobic properties of the lipid molecules. Comparison of the active and inactive conformations of GPR6 identifies F152 3.36 as the gate keeper of the agonist entry from the membrane (Figure 38). In the inactive structures of GPR6, binding of both inverse agonists 1485 and CVN424 induces conformational changes of F152 3.36 . Inspection of the active and inactive structures suggest that steric clashes of isopropyl-amine and tetrahydrofuran-amine groups on the bicyclic ring of 1485 and CVN424, respectively, with the side chain of F152 3.36 in its active position trigger the outward swing of this side chain towards TMs 5 and 6. This rotation enables !-stacking interactions of F152 3.36 with F234 5.47 and W292 6.48 , blocks the membrane entry to the binding pocket and prevents the outward shifting of TMs 5 and 6 as observed in the active conformation of GPR6 (and other 83 class A GPCRs). This further supports the interference of inverse-agonist binding with the agonist entry through the membrane opening and thus antagonizing the receptor activity. 3.7 | Structure comparison of GPR6 with other lipid receptors of the MECA cluster GPR6 shares close homology with other members of the MECA cluster including LPA1 100,104 , S1P1 100,105 and the CB1 106,107 and CB2 101,108 receptors, all of which are lipid-binding receptors (Table 29). Superimposition of GPR6-pseudoapo structure and the active structure of members of LPA, S1P, and Cannabinoid receptor families with highest homology (LPA1 100 , S1P1 100 , and CB2 101 , respectively) (Figure 39) indicates that the overall fold of GPR6 active state is most similar to the LPA1 (PDB 7TD0) with an RMSDCα of 1.6 Å (177 to 177 atoms), whereas the RMSDCα for S1P1 (PDB 7TD3) and CB2 (PDB 6PT0) is 2.5 Å (200 to 200 atoms) and 2.0 Å (182 to 182 atoms), respectively. However, superimposition of GPR6-1485 structure and the inactive structure of the same receptors indicated that the overall fold of GPR6 inactive state is most similar to CB2 (PDB 5ZTY) with an RMSDCα of 1.8 Å (179 to 179 atoms) while the RMSDCα for GPR6-1485 with LPA1 (PDB 4Z36) and S1P1 (PDB 3V2Y) is 1.9 Å (182 to 182 atoms) and 2.2 Å (194 to 194 atoms), respectively. 84 Table 29 | Similarity matrix of the human lipid receptors of the MECA cluster 109 . Despite the overall similarity, several structural differences were observed between GPR6 and the other lipid-binding receptors of the MECA branch. For instance, in GPR6, the ECL2 is tilted inwards over the binding pocket compared to LPA1 and S1P1 (Figure 39). On the contrary, the short ECL1 loop in GPR6 (and CB2) adopts a coil conformation allowing it to tilt outwards. The helical conformation of this loop in S1P1 and LPA1 allows for inter helical interactions with the N-terminal helix, stabilizing the N-terminal helix over the orthosteric pocket in both receptors. In LPA1 and S1P1 100 , K39 (K34 in S1P1) and Y34 (Y29 in S1P1) are key N-terminal residues which interact with LPA and S1P. These residues are not conserved in GPR6. In the case of GPR6, the N-terminus consists of poly-Alanine and poly-Valine stretches suggesting that it does not form a helix. Residues 1-47 were therefore truncated for crystallization experiments. In our structures, the 85 remaining N-terminal residues 48-68 are not resolved which most likely is due to structural flexibility of this region and lack of stable interaction with the ligand or extracellular loops of GPR6, so we speculate that N-terminal residues do not contribute to agonist stabilizing interactions significantly. Instead, interactions by residues of ECL2 seem to stabilize the ligand in GPR6 (and CB2) (Figure 39). Our structures suggest that GPR6 endogenous ligand might enter from the transmembrane opening framed by TMs 3-5. This opening appears to be gated by a single residues F152 3.36 whose side chain rotation is one of the main molecular switches between the active and inactive conformations. This transmembrane entry is closed in active structures of other MECA cluster lipid receptors including S1P1, LPA1 100 and CB2 101 . In CB2, F117 3.36 takes on a similar orientation to F152 3.36 of GPR6 active conformation and instead, F197 5.46 blocks this entry. In S1P1 and LPA1 F 3.36 is not conserved and instead, residues F133 3.41 and F213 5.46 block the transmembrane entry, respectively (Figure 39). This analysis could explain the high basal activity of GPR6 compared to LPA1 and S1P1. Lastly, the Lig-OLA binding pose in GPR6 is closest to S1P binding in S1P1, however, analysis of the residues that line GPR6 binding pocket does not suggest significant similarity to S1P1 or any other lipid receptor of the MECA cluster with a known endogenous agonist (Table 30). This is while high homology of GPR6 binding pocket with GPR3 and GPR12 (86%) suggests that insight into GPR6 endogenous agonist could pave the path towards deorphanization of these two other orphan receptors of the MECA cluster as well (Table 30). 86 Figure 39 | Structure comparison of lipid receptors of the MECA cluster. LPA1 with its agonist LPA (7TD0, cyan), S1P1 with its agonist S1P (7TD3, magenta), and CB2 with its agonist WIN 55,212-2 (6PT0, yellow) were superimposed on GPR6-pseudoapo with its tentatively-modeled- agonist OLA (blue). (a) ECL3 of GPR6 tilts inward and closer to ECL2 compared to other lipids receptors that were analyzed. (b) Top part of GPR6 ECL2 hinges inward over the bottom part of ECL2 and towards the binding pocket. This also enables contraction of ECL3. (c) Short coil of ECL1 in GPR6 compared to the helix of ECL1 in LPA1 and S1P1. This helix stabilizes the N- terminal helix over the binding pocket. (d) Two N-terminal residues from the N-terminal helix in LPA1 and S1P1 stabilize the lipid ligand in the binding pocket. Ligand interactions in GPR6 and CB2 are mediated by residues of ECL2 instead. These interacting residues in all four models are labeled and shown in wire representation. Ligands are shown in stick representation. (e) Different 87 phenylalanine residues seem to block the transmembrane ligand entry point that is open in the GPR6 active conformation. These residues are labeled and shown in wire representation. Table 30 | Conservation of GPR6 residues within 4 Å of Lig-OLA in the MECA cluster. 88 3.8 | Conclusion Our structures of GPR6 with two potent inverse agonists reveal the architecture of the binding pocket and highlight key residues that are involved in ligand binding. Our findings can inform and facilitate future drug discovery efforts targeted at this receptor. Our structures of GPR6 in its active conformation allowed us to investigate mechanism of activation in this receptor and identify key molecular switches that propagate the signal from the ligand binding pocket to the cytoplasmic side of the receptor. Moreover, our structures revealed the presence of a lipid-like molecule that seems to stabilize GPR6 in its active conformation, potentially contributing to its high basal activity. Tentatively modeling of OLA in this pocket allowed for thorough analysis of residues lining the pocket and interacting with the ligand. The follow-up crystallographic studies to elucidate the identity and origin of the orthosteric ligand were not conclusive and motivated us to investigate this question using complimentary experimental and computational approaches that are described in the chapter four of this dissertation. 89 Chapter 4 | GPR6 structure-informed studies Structural studies of GPR6 in its active and inactive conformations provided novel insights into receptor’s architecture of the orthosteric pocket, molecular mechanism of activation and its potential induced activation by a lipid molecule that occupies GPR6 orthosteric pocket in its active state. Although invaluable, structural studies of a GPCR only provide a static snapshot of its conformation in a given state, and in cases where structures of multiple states of the receptor are available, mechanistic insights into the receptor’s ligand binding, inhibition/activation and signaling can be inferred. These preliminary findings can further inform a series of subsequent interdisciplinary studies to compliment the structural insights and enable thorough investigation of the target using different experimental or computational tools. In this chapter, the five complimentary approaches that were pursued by a team of multidisciplinary scientists to further elucidate structure and function of GPR6 will be described. These disciplines/experiments include in silico ligand docking, molecular dynamics (MD) simulations, radioligand binding studies, cell- based functional assays, and mass spectrometry. 4.1 | In silico ligand docking Once the structure of a GPCR (or other proteins) is solved, in silico ligand screening using virtual libraries of relevant compounds can identify ligands of computationally-predicted high affinity, judged by their binding score (measure of how well a ligand binds in the pocket) 110 . Therefore, the structural model not only provides information about interactions between the receptor and the co- crystalized ligand, but also serves to identify new hits. This screening is usually restricted to the 90 orthosteric pocket of the receptor but can be extended to search for binders of alternate sites (e.g. allosteric modulators), albeit computationally exhaustive and expensive. For GPR6, given its status as an orphan receptor, gaining additional information about its endogenous binder is of outmost interest. In collaboration with Anastasiia Sadybekov, Ph.D. from Professor Vsevolod Katritch laboratory, we were able to use the active conformation model of GPR6 (from GPR6- pseudoapo structure) to virtually screen about 43 thousand compounds (100<MW<800) from the library of human metabolome database 111 for potential binders of the receptor, using ICM-Pro Molecular Modeling Software (Molsoft) 112 . This search identified fourteen top hit compounds (Table 31) to be tested for GPR6 stimulation in vitro. However, none of these lipids were able to induce GPR6 signaling through Gs beyond its basal levels and hence, the search for GPR6 endogenous agonist continued. Table 31 | Top GPR6 hit compounds from in silico screening. 91 4.2 | MD simulations MD is another computational technique that is often utilized to investigate GPCR signaling mechanisms, including ligand binding and receptor (de)activation. Using the static model of a GPCR, in our case solved via x-ray crystallography, this technique enables simulation of its detailed structural dynamics over time based on an atomic force field. Information obtained from these simulations can further explain dynamics of ligand binding and ligand-induced GPCR (de)activation. In collaboration with Ho Ming Lam from the laboratory of Professor Vsevolod Katritch, we were able to investigate molecular dynamics of GPR6 to primarily shed light on its mechanism of ligand binding and activation and also potentially gain insight into the receptor’s structural features governing its constitutive activity. We first used the ligand-bound structures of GPR6 to investigate conformational dynamics of the bound ligands. Crystal structures of GPR6-1485 and GPR6-pseudoapo (with Lig-OLA modeled) as well as a GPR6-S1P (with the ICM-docked S1P) were used to compare dynamics of inverse agonist and lipid bound in the orthosteric pocket of GPR6. Trajectories of ligand interactions with neighboring residues of the binding pocket were recorded and analyzed over 1200 ns, and results indicated stable interactions, primarily, of hydrophobic nature, for the tightly bound small molecule inverse agonist 1485 (Figure 40). This is while interactions observed for the headgroup of the two tested lipids (OLA and S1P) appeared to be highly dynamic (Figure 40). Overlay of several snapshots of GPR6-pseudoapo structure over the course of this simulation further depicts the highly dynamic headgroup of both OLA and S1P in the pocket, while their fatty acid tail stays 92 relatively more stable (Figure 41), which could also explain the well resolved density of the latter in our structures. Figure 40 | Trajectories of ligand interactions with GPR6 binding pocket residues. Distance of closest heavy atoms between GPR6 residues and ligands were recorded over 1200 ns. (a) The interactions for 1485 were found to be highly stable and of hydrophobic nature, primarily maintaining less than 5Å distance of the heavy atoms. (b-c) Interactions of the headgroups of OLA and S1P are overall more dynamic than interactions of 1485, while there are indications of polar 93 or ionic interactions. The two dotted light grey lines mark 2.5 Å and 5 Å of the closest heavy atoms from the interacting pair. MD simulation trajectories confirm our identification of several residues that seems to stably interact with 1485 in GPR6-1485 crystal structure. Effect of several of these residues on ligand binding were later investigated in radioligand binding studies (discussed later in this chapter). The simulations further restate ionic interactions of residue R220 ECL2 with the headgroup of the putative ligand in the GPR6-pseudoapo crystal structure, as well as interactions with H128 2.60 and Y312 7.36 . The simulations additionally suggest ionic interactions of S1P with D76 1.36 , however the headgroup of the ligand in our crystal structure is oriented away from D76 1.36 and does not seem to interact with it. This potential interaction is also visually obvious in the overlay of simulation trajectories (Figure 41). 94 Figure 41 | Overlay of simulation trajectories for GPR6-S1P model. Highly dynamic headgroup of S1P may explain poorly resolved electron density of this region obtained from x-ray diffraction and prevent confident identification and modeling of the ligand. Fatty acyl chain of S1P is relatively less dynamic but adopts two different binding poses at the bottom of the pocket. MD simulations of GPR6-S1P model additionally show two binding poses for the fatty acid chain of S1P molecule. The primary pose is similar to OLA in GPR6-Lig-OLA model, and the alternate pose is one where this chain extends through the membrane opening (Figure 41), partially occupying the space that, in the GPR6-Lig-OLA model, is occupied by the second lipid fraction which follows Lig-OLA’s tail into the pocket from the membrane opening (Figure 31). Other evidence from the simulations suggests that this opening is also transiently occupied by lipid tails of POPCs from both sides of the bilayer. These observations, together with our structural insights informed by the GPR6-pseudoapo model propose that the fatty acid chains of either the ligand of 95 orthosteric pocket or other ligands of the membrane bilayer may interchangeably enter into GPR6 membrane opening of the binding pocket and contribute to the stability of its active conformation. To further investigate GPR6 mechanism of activation and understand its high basal activity, we employed a Markov State Model (MSM) 113 for the analysis of the MD simulations. An MSM is a discrete-time model that comprises metastable conformations of the receptor and serves as an unbiased way to analyze dynamics of the receptor without the restrains of having to overcome high energy barriers associated with GPCRs activation in transitioning between various conformational states. 114,115 For GPR6, our MSM consisted of three active-like and three inactive meta stable conformations of the receptor, all of which were in the apo state since the primary goal of this simulation was to investigate the stationary distribution of these meta stable conformations in the absence of a ligand and understand whether high basal signaling of the receptor comes from a true constitutive activity of the receptor, and not induced activity by an endogenous agonist. Analysis of the free energy landscape indicated higher stability of the receptor in the inactive conformations with the stationary distributions for metastable inactive states totaling 83% and the metastable inactive states amounting up to only 17% of all conformational states of the receptor. This observation suggests that high basal activity of the receptor may not be an intrinsic characteristic of the receptor and rather regulated by a ubiquitous agonist that constantly induces this activation. The insights from MD simulations were crucial in understanding some of our structural observations. These computational results served to validate our structural insights concerning inverse agonist binding and key residues involved in maintaining stabilizing interactions with both the inverse agonist and lipid-like molecule in our inactive and active conformation structures. 96 Further, they offer potential explanation regarding poorly resolved electron density of lipid headgroup in our active conformation structure. Lastly, the simulations provide novel insights into the receptor’s mechanism of activation. GPR6 MSM analysis suggests activation of the receptor may in fact be induced by an endogenous agonist, which further validates our structural data. However, the identity of this ligand still remains elusive. 4.3 | Radioligand binding assays Once the structure of a GPCR (or other proteins) is solved, radioligand binding assays can be utilized to gain quantitative information about the affinity of a ligand for different mutants of the GPCR. GPR6 structure with the small molecule inverse agonist 1485 and follow-up MD simulations identified key receptor residues that maintain stable interactions with the ligand. In order to further investigate the role of some of these residues in ligand binding, we engineered point mutations of these residues to either interfere with binding by posing steric clashes of the mutated residues and the ligand or simply eliminate stabilizing interactions with a given residue. We generated mutant variants of GPR6 for some of the key residues that are shown to maintain stable interactions with the inverse agonist 1485 from both our crystal structure and MD simulations. Mutant gene of interest was cloned in a pcDNA3.1(-) vector and used for transfection of 30 mL culture of HEK293 suspension cells, as described before. Three biological replicas of the biomass were produced for each mutant receptor. For each sample, frozen biomass was homogenized in a hypotonic buffer containing 10 mM HEPES (pH 7.5), 10 mM MgCl2, 20 mM KCl, and 1000 x dilution of PI cocktail. Sample was then ultracentrifuged at 200,00 x g for 20 minutes. Membranes (pellet) were resuspended in a tris buffer (50 mM tris pH 7.4, 1 mM 97 EDTA) with 1000 x PI cocktail and dounce-homogenized followed by syringe homogenization. Membranes were flash frozen, stored at -80º C, and later shipped to Takeda Pharmaceuticals on dry ice for radioligand binding assays. Mutations V219 ECL2 P, F152 3.36 A, F295 6.51 A, and L3157 7.39 A all seemed to increase the dissociation constant (Kd) of the radioligand [ 3 H]8983 (RL) to over 160 nM (Figure 42), while comparing ligand binding between the crystallization construct and the wild type (WT) receptor showed comparable results (Kd, RL = ~ 17 nM, Table 32) (Figure 42). We also observed notable effects of L148 3.32 mutations on ligand binding in the two mutants L148 3.32 A and L148 3.32 F (Figure 42). Ligand binding was negatively affected in L148 3.32 A, increasing the Kd by over 4 folds while the L148 3.32 F resulted in stronger binding of RL to GPR6, lowering the Kd by over 6 folds. Although we originally expected L148 3.32 F mutant to interfere with ligand binding through steric clashes, upon investigating the RL chemical structure and its difference from 1485, we concluded that strong π-stacking of phenylalanine with the ligand could be stabilizing its binding in the pocket further. Additionally, we tested the effect of mutants that may not directly interfere with ligand binding but might disrupt interactions that stabilize the receptor in its inactive conformation, hence indirectly negatively affect inverse agonist binding. We generated the mutants S151 3.35 F and V237 5.50 W (PIF motif) and observed complete loss of RL binding in both variants (Figure 42). Our structures suggest the phenylalanine residue of S151 3.35 F mutant could disrupt inter-helical contacts between TMs 2 and 3 in both active and inactive conformations, but less so in the latter. Furthermore, PIF domain has been proposed as one of the main signal transmission and activation switches in other class A GPCRs, and its conformational re- arrangements upon receptor activation have been observed in different cases 116-118 . Accordingly, 98 mutation of V237 5.50 to W could induce conformational rearrangements of PIF and favor a specific global conformation of the receptor and affect its downstream signaling, discussed in the next section of this chapter. Table 32 | Summary of Radioligand Kd for different GPR6 mutants. 99 100 Figure 42 | Radioligand saturation binding to GPR6 mutants of the binding pocket residues. (a) Chemical structure of radioligand [ 3 H]8983 (RL). (b) Superimposition of docked RL (cyan) and 1485 (lemon) in GPR6 pocket. Residues interacting with 1485 within 4 Å are shown in stick representation. (c-f) Radioligand binding results are expressed as mean specific binding ± S.E.M in count per minute (CPM) at each concentration of the RL, calculated from three independent experiments, and the x axis is in logarithmic scale. (c) RL binding to wild-type (WT) receptor, crystallization construct (CC: six point-mutations, 1-47 N-terminal truncation, ICL3-bRIL fusion, and C-terminal his tag), and crystallization construct without the bRIL fusion protein (CC-bRIL). (d) RL binding to GPR6 mutants of ligand binding residues which increased the Kd of the RL to above 160 nM. (e) Mutations of L148 3.32 and their differential effect on ligand binding depending on the mutant, compared to WT. (f) Mutations of V237 5.50 from the PIF motif and S151 3.35 adjacent to GPR6 activation switch F152 3.36 are deleterious to RL binding. The insights obtained from RL binding assays, performed by Takeda Pharmaceuticals Inc., further support our structural insights and analyses, and highlight the key role of specific residues of GPR6 orthosteric pocket in stable ligand binding which can inform the future attempts at designing potent drugs targeting GPR6. 4.4 | Functional cell-based assays In order to gain further insight into GPR6 constitutive activity and effect of different mutants on receptor coupling to Gs protein we collaborated with Takeda Pharmaceuticals to perform functional cell-based assays on various mutants of the receptor to determine the receptor’s basal 101 activity by indirectly measuring the cAMP signaling. Mutant gene of each construct of interest was cloned into pcDNA3.1(-) vector. Plasmids were sent to Takeda Pharmaceuticals for transfection of CHO cells for functional assays. The results verified the key role of Lig-OLA (or other lipids) interacting residues of GPR6 binding pocket. Most notably, mutations F152 3.36 A and V219 ECL2 P decreased GPR6 basal signaling by 1.8 and 1.5 folds, respectively (Figure 43). In our structure of the GPR6 active state we observed stabilizing π-stacking interactions of F152 3.36 with W292 6.48 of the toggle switch. Given these key interactions and F152 3.36 apparent role in signal transduction and propagation during activation of GPR6, we suspect that its mutation to Alanine disrupts the activation mechanism of GPR6. Also, V219 ECL2 seems to maintain polar interactions with the headgroup of GPR6 putative agonist through its amino group, hence mutating it to proline that is not able to engage in such interactions would lower the overall stability of the bound ligand. Furthermore, we expect this mutation to disrupt the conformation of ECL2 beyond just V219 ECL2 and therefore can explain the disruptive nature of V219 ECL2 P mutation. The two polar residues H128 2.60 and Q132 2.64 that are shown to make polar interactions with other potential lipid binders in docking studies also show significant decrease of GPR6 signaling upon their mutation to alanine, while mutation R220 ECL2 A does not seems to lower signaling significantly. While docking studies suggest interactions of R220 ECL2 side chain with the polar headgroup of potential lipid binders, our crystal structures indicate R220 ECL2 ’s interaction with the putative endogenous ligand is mediated through both a salt bridge with residue’s side chain and strong hydrogen bonding with the residues back bone amino group, and the latter remains unaffected upon mutation to Alanine. The observed effect of R220 ECL2 A mutant on GPR6 activation (1.3 folds) could be partially caused by loss of transient interactions with the lipid headgroup as well as loss of stabilizing interactions between the side chain of R220 ECL2 and H304 ECL3 and S217 ECL2 that stabilize the contracted 102 conformation of ECL2-3 in the active conformation. Much like in radioligand binding assays, we see the differential effect of L148 3.32 mutations in the functional assays. We observed L148 3.32 A lowering RL binding affinity while L148 3.32 F improved it. If this were to be true for the endogenous agonist as well, the mutations would result in a decrease and increase of basal activity, respectively, which is expectedly what we note in the functional assays. This observation reinforces the role of L148 3.32 in stabilizing the ligand without directly favoring the active or inactive conformations of the receptor. Additionally, mutations of T311 7.35 A and Y312 7.36 F markedly increased GPR6 basal signaling (1.7 and 2.1 folds, respectively). In our GPR6- pseudoapo structure, T311 7.35 maintains hydrogen bonding with the backbone of V218 ECL2 . V218 ECL2 also hydrogen bonds with the headgroup of the putative lipid agonist, therefore mutating T311 7.35 to A can free V218 ECL2 to more stably hydrogen bond with the agonist lipid headgroup and hence further stabilize the active conformation. Y312 7.36 in both active and inactive structures primarily stabilizes TMs 1 and 7 by hydrogen bonding with D76 1.36 and establishing π-stacking interactions with the nearby tryptophan W75 1.35 . Our MD simulations (Figure 40) indicated D76 1.36 as one of the residues interacting with the head group of the putative agonist, therefore mutating Y312 7.36 to F would free residue D76 1.36 to form strong ionic interactions with the agonist head group. This is while F can still maintain π-stacking interactions with the neighboring tryptophan of TM1 (W75 1.35 ) and preserve the structural integrity of the receptor. Moreover, the purpose of A153 3.37 M mutants was to create steric hinderance for the lipid binder of GPR6 orthosteric pocket entry from the membrane. This mutation is predicted to also interfere with binding of any long fatty-acid chain lipid (d ≥ 20) in the orthosteric pocket, but its effects on the basal activity of GPR6 are negligible, suggesting the endogenous ligand has a shorter fatty-acid chain, similar to that observed in the crystal structure. We further observed that CC construct significantly reduces the 103 receptor activity. This effect is expected because the bRIL fusion protein inserted in the ICL3 of CC construct which will interfere with binding of the effector (Gs) protein. Accordingly, we removed the bRIL fusion from the CC construct and restored the ICL3. However, this new construct (CC_ΔbRIL) appeared to have low basal activity, although higher than CC. Therefore, we decided to test the effect of individual modifications (mutation/truncation) of CC on receptor’s activity. The most pronounced effect for the point mutations of the crystallization construct was observed in A173 ICL2 P variant (1.3 folds) which is expected given the role of this loop in engaging the effector protein in GPCRs. Further, deletion of residues 1-47 from the N-terminus of GPR6 did not affect receptor’s activation significantly. Hence, the role of GPR6 N-terminal domain in the receptor’s ligand binding and activation remains elusive. Lastly, mutation of V237 5.50 W (PIF motif) only slightly improved GPR6 signaling, unlike its deleterious effect on RL binding. On the contrary, S151 3.35 F mutant significantly decrease GPR6 basal signaling which suggests that the mutation destabilizes both active and inactive conformations of the receptor. 104 105 Figure 43 | GPR6 basal signaling in cAMP HTRF assay. TR-FRET signal from CHO-K1 cells is measured at 655 nm and the readout is plotted as % response of CHO-K1 cells. Results are shown in mean ± SD (N=20; 3 biologically independent samples were measured in 4, 8, and 8 replicas). Data is analyzed using one-way ANOVA with Dunnett multiple comparison test to determine significance compared to GPR6 WT response. The purple dotted line along all plots indicates GPR6 WT basal signaling level. (a) Most mutants of the Lig-OLA interacting residues significantly reduce GPR6 basal signaling with P values = <0.0001 for F152 3.36 A and V219 ECL2 P, but Y312 7.36 F seems to increase GPR6 activity by 2 folds (P = <0.0001). (b) Modifications of GPR6 construct for crystallization seem to have a cumulative negative effect on receptor’s signaling through Gs but the individual mutations contribute to this effect non-significantly. (CC: crystallization construct, CC-ΔbRIL: CC with intact ICL3 and no bRIL fusion, Δ1-47: N-terminal truncation of WT receptor). (c) Significant effect of S151 3.35 F on GPR6 activation while V237 5.50 W of the PIF domain increases GPR6 Gs activation by only 1.25 folds (P = 0.5112). P values are indicated for each mutant: P <0.0001 (****), P <0.0002 (***), P <0.0021 (**), P <0.0332 (*) and P <0.1234 (ns). 4.5 | Mass spectrometry While different computational and in vitro assays provided additional insights into GPR6 ligand binding and receptor activation, none of them were able to directly identify the putative endogenous ligand of GPR6 orthosteric pocket in the receptor’s active conformation. We next sought to perform Native Mass Spectrometry (Native-MS) experiments as a mean to directly identify the ligand in GPR6 purified protein samples 119,120 . In these experiments, the goal is to 106 subject the sample to high-enough collision energy to strip off molecules of the detergent micelles from surrounding the protein as well as the loosely-bound co-purified lipids, but not too high as to activate ions of tightly-bound species of interest (e.g. lipid or other ligands in the orthosteric pocket) which would result in them being stripped off of the protein like the detergent micelle. In an ideally fine-tuned experiment with a well-behaved sample, signal from both ligand-bound and -free protein can be picked up in the mass analyzer the identity of the bound ligand can be inferred from the difference between the m/z of the two signals 119,120 . In collaboration with Corrinne Lutomski and Idlir Liko from the laboratory of Professor Carol Robinson (University of Oxford), we were able to perform native-MS on GPR6 purified protein samples. For sample preparation, CC was expressed in HEK293 cells like before and the same purification procedure as followed for crystallization was used. After elution of the protein from the TALON resin, protein was typically concentrated and then desalted, either over a desalting column or during a preparative SEC run. The final buffer used in desalting contained 20 mM Hepes pH 7.5, 100 mM NaCl, 5% glycerol, and 0.015/0.003% w/v DDM/CHS. Sample was typically concentrated to ~1 mg/mL and flash frozen for shipment to Oxford. Initial experiments yielding complex mass spectra of the sample proved identification of the orthosteric pocket-bound lipid to be extremely challenging. After fine-tuning the experimental procedure, we were ultimately able to obtain a relatively less complex mass spectrum of the sample but could not identify any obvious signals from a potential lipid-bound population of GPR6 (compared to the signal from GPR6-apo). This could be caused by either natural liberation of the bound lipid from majority of the receptor’s population through the course of the experiment and before the species reach the mass analyzer (resulting in very low S/N of the lipid-bound receptor population), or by activation of the lipid ion and its subsequent depletion from the receptor. Although the signal for the activated ion of this freed lipid might be 107 picked up by the mass analyzer, associating it back to the specific binding site of the receptor is not possible. We next sought to analyze the lipidome of GPR6 and detected several classes of abundant lysolipids in the sample. The 16-carbon fatty acid chain subtype of each class of identified lysolipids was chosen for subsequent experiments (Figure 44). This decision was informed by the approximate overall MW we expect for this lipid based on its electron density observed in our structural studies. We exogenously added each of these lipids to purified GPR6 at the final molar ratio of 1:10 for protein:lipid and incubated the samples for thirty minutes at room temperature before subjecting the samples to native MS. We were able to detect strong signal for LysoPC-bound species (Figure 45) while negligible to no signal was observed for any of the other species. Our molecular docking studies previously identified several classes of lysolipids as potential binders of GPR6 (Table 31). The longer chain (22:0) LysoPC hit was not able to induce activation in GPR6 functional assay. However, molecular docking revealed the potential binding pose for this class of lipids in the orthosteric pocket of GPR6, suggesting the stabilizing interactions of oxygens from the LysoPC phosphate group with H128 2.60 and Q132 2.64 as well as interactions of the LysoPC hydroxy group with the backbone of residues of GPR6 ECL2, including V218, V219, and R220. Further, the trimethylamine head of LysoPC lipids can fit in the space towards the extracellular side of the orthosteric pocket, and the fatty acid tail, depending on the length, may extend into the lipid bilayer through the membrane opening at bottom of the orthosteric pocket. While MS experiments suggest that LysoPC might bind to GPR6 and potentially induce its activation and docking studies identify the potential binding pose, further binding and functional assays need to be done to confirm our observations and claims. 108 Figure 44 | Lysolipids identified in GPR6 lipidome and tested for binding in MS experiment. Different classes of lysolipids detected from GPR6 lipidomics include lysosphingolipid (S1P), lysophosphocholine (LysoPC), lysophosphatidylethanolamine (LysoPE), lysophosphatidylserine (LysoPS), lysophosphatidic acid (LysoPA). 109 Figure 45 | Native MS identifies lyso-PC bound GPR6 species after incubation with the lipid. Titration of 2.5 μM purified GPR6 with 0-500 μM LysoPC shows clear indication (right spectrum) of ligand binding at 25 μM while incubation with 500 μM lipid results in it nonspecific binding to the receptor. 4.6 | Conclusion Following structural studies of GPR6, combination of complimentary interdisciplinary approached enabled thorough investigation of the receptor mechanism of ligand binding and activation. Collective of our observations (particularly MD) suggests the activation of GPR6 is induced by a ubiquitous lipid ligand that stabilizes the receptor in its active conformation. This is supported by both our structural insights and functional assays where mutations of key ligand-binding residues lowered GPR6 basal signaling, potentially caused by destabilized binding of the endogenous 110 ligand. MS experiments were further utilized to shed light on the identity of GPR6 endogenous ligand, paving the path to its deorphanization. Identification of LysoPC as a potential agonist of GPR6 needs to be verified by orthogonal approaches (e.g. binding and functional assays) and could guide receptor’s deorphanization for both GPR6 and its two paralogs GPR3 and GPR12. 111 Chapter 5 | Conclusions and future direction In this project, I set out to utilize advanced structural biology techniques to study the structure and function of an orphan GPCR with unique properties which make it an attractive target both from a basic biology standpoint and for pharmaceutical purposes. Given its high homology with lipid receptors of the MECA cluster of class A GPCRs, GPR6 is believed to be stimulated by a lipid molecule 42 . However, despite its high basal signaling through Gs, no known ligand to this date has been confirmed to induce activation of GPR6 70 . The high basal activity and lack of an efficacious inverse agonist, until recent years, have further complicated efforts in deorphanizing this receptor, to the point that even culture of GPR6 expressing cells in a charcoal-stripped media, that is free of lipophilic materials, did not completely abolish GPR6 signaling to facilitate the search for its endogenous agonist through screening of a lipid library 57 . With all the evidence in hand, it is still not clear whether GPR6 has true constitutive activity that is intrinsic to the receptor or that its continuously activated by a ubiquitous ligand. This ambiguity poses an exciting opportunity for discovery structural biology research which is further encouraged by the GPR6 therapeutic potentials. This receptor has high expression levels in the mammalian brain 43 and is suggested to be highly exclusively localized to D2R expressing neurons of the basal ganglia 55 . High basal activity of GPR6 in these neurons and the receptor’s opposing activity to D2R which primarily signals through Gi to lower cAMP production 121 makes GPR6 a very attractive target of inverse agonists that produce a net effect of downregulating cAMP production in the indirect pathway of the basal ganglia which is precisely what is pursued in treatments of Parkinson’s disease to alleviate motor symptoms in the patients. We were able to pair our scientific curiosity with a drive 112 for drug discovery and benefited from state-of-the-art structural biology techniques to enable structural studies of GPR6. Our structures of GPR6 in its both active and inactive states shed light on receptors mechanism of ligand binding, inverse agonism, and activation. Our active-state structure of GPR6 suggests the receptor’s activation is potentially induced by a putative lipid-like agonist whose identity has yet to be confidently determined. Our findings further laid the foundation of several collaborations with a multidisciplinary team of scientist with expertise spanning computational biology, pharmacology, cellular biology, and protein biochemistry. Each of these disciplines served to both validate our structural findings and provide further insights into GPR6 structure and function. Collectively, our computational and experimental observations confirm GPR6 ability to bind lipophilic ligands and highlight key residues involved in ligand binding, signal propagation, and activation of this receptor. Our findings further suggest potential role of lipids of plasma membrane in activation of GPR6. However, whether this role is more of an allosteric modulation or direct activation of the receptor is still not clear. Our insights further nominate LPCs as potential binders of GPR6 which is pending further validation from orthogonal experiments. This open question of GPR6 endogenous agonist ID and other avenues of research with potentials to provide additional insight into GPR6 structure and function are briefly described below. One explanation for our structural observations of a lipid molecule in GPR6 orthosteric pocket is the introduction of this lipid from the LCP matrix to the receptor. Cryo-electron microscopy (cryo- EM) is an alternative structural biology approach that has enabled structure solution of many GPCRs in the recent years 122 . Exploiting cryo-EM for high resolution structure solution of GPR6 113 in complex with its effector Gs protein, without the need to introduce LCP to the sample, could serve to shed light on the identity of the putative endogenous ligand of the orthosteric pocket. One challenge that is expected from MD simulations is the highly dynamic behaviors of the headgroup of the putative lipid in GPR6 binding pocket which could complicate confident identification of this molecule through both x-ray crystallography and cryo-EM. Using GPR6 antibody as a chaperone in cryo-EM sample preparation could potentially further stabilize the endogenous lipid in the pocket and minimize its dynamics allowing for successful resolution of its headgroup electron density and therefore high-confidence identification of the lipid. Alternatively, various N- terminal fusion proteins for crystallization may result in differential packing of the monomers in the crystal lattice with potentials to minimize dynamics of the extracellular side of the receptor. Lastly, rational engineering of point mutations in the binding pocket to confine the lipid head group may serve the same purpose. Another interesting avenue to investigate is structure solution with other small molecules of differential pharmacology at GPR6, like neutral antagonists. Given the constitutive activity of GPR6 is not fully explained yet, co-crystal structure of GPR6 with its neutral antagonist could potentially give insight into receptor’s intermediate states and provide additional insight into its mechanism of activation. Moreover, previous studies suggest the role of N-acyl dopamines in GPR6-mediated β-arrestin recruitment 73 . These studies show the biased signaling of GPR6 induced by N-acyl dopamines which is dependent on both the dopamine headgroup and the fatty acid chain length. We also observed effects of LPC molecules’ fatty acid chain length on the cation distribution of GPR6 populations (results not discussed in this manuscript). Structural studies of GPR6 with varied chain length of putative ligands can serve as an alternate route to investigate its 114 activation mechanism further. The fatty acid chain length effect has also been shown to cause biased signaling in other lipid receptors 123 , hence elucidating how it affects GPR6 signaling would not only further explain signaling mechanisms in GPR6 but also shed light on regulation and signaling of its close paralogues GPR3 and GPR12. Lastly, allosteric modulation of GPCRs by ions has been demonstrated in several cases 124-126 . Mass spectrometry of GPR6 revealed a differential distribution of bivalent ions Zn 2+ and Ca 2+ in GPR6- IAG and GPR6-pseudoapo samples which suggests a potential regulatory role for these ions in different states of the receptor. Investigation of GPR6 regulation by ions is another avenue of research worth pursuing which could further elucidate the receptor’s mechanism of high basal activity, potentially regulated by ions. In conclusion, our structural and follow-up interdisciplinary studies greatly advanced understanding of GPR6 structure and function and opened up new avenues for studies of this receptor and its two orphan paralogs GPR3 and GPR12. Studies of GPR6 structure and function were primarily led by me, but only made possible to be completed at the depth and scope described by major contributions from members of laboratories of Professors Vadim Cherezov, Vsevolod Katritch, and Carol Robinson, most importantly Drs./Messrs./Mmes. Linda Johansson, Gye Won Han, Anastasiia Sadybekov, Ho Ming Lam, Idlir Liko, and Corinne Lutomski along with significant support received from Takeda Pharmaceuticals and our main collaborators within the company, Drs. Hans Schiffer, Kumar Singh Saikatendu, and Holger Monenschein. 115 References 1 Lagerström, M. C. & Schiöth, H. B. Structural diversity of G protein-coupled receptors and significance for drug discovery. Nature Reviews Drug Discovery 7, 339-357, doi:10.1038/nrd2518 (2008). 2 Katritch, V., Cherezov, V. & Stevens, R. C. Structure-function of the G protein-coupled receptor superfamily. Annu Rev Pharmacol Toxicol 53, 531-556, doi:10.1146/annurev- pharmtox-032112-135923 (2013). 3 Dijkman Patricia, M. et al. Conformational dynamics of a G protein–coupled receptor helix 8 in lipid membranes. Science Advances 6, eaav8207, doi:10.1126/sciadv.aav8207. 4 Tesmer, J. J. G. Hitchhiking on the heptahelical highway: structure and function of 7TM receptor complexes. Nature Reviews Molecular Cell Biology 17, 439-450, doi:10.1038/nrm.2016.36 (2016). 5 Wacker, D., Stevens, R. C. & Roth, B. L. How Ligands Illuminate GPCR Molecular Pharmacology. Cell 170, 414-427, doi:10.1016/j.cell.2017.07.009 (2017). 6 Latorraca, N. R., Venkatakrishnan, A. J. & Dror, R. O. GPCR Dynamics: Structures in Motion. Chemical Reviews 117, 139-155, doi:10.1021/acs.chemrev.6b00177 (2017). 7 Zhou, Q. et al. Common activation mechanism of class A GPCRs. eLife 8, e50279, doi:10.7554/eLife.50279 (2019). 8 Venkatakrishnan, A. J. et al. Molecular signatures of G-protein-coupled receptors. Nature 494, 185-194, doi:10.1038/nature11896 (2013). 116 9 Eddy, M. T. et al. Allosteric Coupling of Drug Binding and Intracellular Signaling in the A2A Adenosine Receptor. Cell 172, 68-80.e12, doi:https://doi.org/10.1016/j.cell.2017.12.004 (2018). 10 Tehan, B. G., Bortolato, A., Blaney, F. E., Weir, M. P. & Mason, J. S. Unifying Family A GPCR Theories of Activation. Pharmacology & Therapeutics 143, 51-60, doi:https://doi.org/10.1016/j.pharmthera.2014.02.004 (2014). 11 Wescott, M. P. et al. Signal transmission through the CXC chemokine receptor 4 (CXCR4) transmembrane helices. Proceedings of the National Academy of Sciences 113, 9928, doi:10.1073/pnas.1601278113 (2016). 12 Martí-Solano, M., Sanz, F., Pastor, M. & Selent, J. A Dynamic View of Molecular Switch Behavior at Serotonin Receptors: Implications for Functional Selectivity. PLOS ONE 9, e109312, doi:10.1371/journal.pone.0109312 (2014). 13 Venkatakrishnan, A. J. et al. Diverse activation pathways in class A GPCRs converge near the G-protein-coupling region. Nature 536, 484-487, doi:10.1038/nature19107 (2016). 14 Valentin-Hansen, L. et al. The Arginine of the DRY Motif in Transmembrane Segment III Functions as a Balancing Micro-switch in the Activation of the β2-Adrenergic Receptor. Journal of Biological Chemistry 287, 31973-31982, doi:https://doi.org/10.1074/jbc.M112.348565 (2012). 15 Ballesteros, J. A. & Weinstein, H. in Methods in Neurosciences Vol. 25 (ed Stuart C. Sealfon) 366-428 (Academic Press, 1995). 16 Hilger, D., Masureel, M. & Kobilka, B. K. Structure and dynamics of GPCR signaling complexes. Nature Structural & Molecular Biology 25, 4-12, doi:10.1038/s41594-017- 0011-7 (2018). 117 17 Wootten, D., Christopoulos, A., Marti-Solano, M., Babu, M. M. & Sexton, P. M. Mechanisms of signalling and biased agonism in G protein-coupled receptors. Nature Reviews Molecular Cell Biology 19, 638-653, doi:10.1038/s41580-018-0049-3 (2018). 18 Weis, W. I. & Kobilka, B. K. The Molecular Basis of G Protein–Coupled Receptor Activation. Annual Review of Biochemistry 87, 897-919, doi:10.1146/annurev-biochem- 060614-033910 (2014). 19 Wu, V. et al. Illuminating the Onco-GPCRome: Novel G protein–coupled receptor-driven oncocrine networks and targets for cancer immunotherapy. Journal of Biological Chemistry 294, 11062-11086, doi:10.1074/jbc.REV119.005601 (2019). 20 Dascal, N. & Kahanovitch, U. in International Review of Neurobiology Vol. 123 (eds Paul A. Slesinger & Kevin Wickman) 27-85 (Academic Press, 2015). 21 Pfleger, J., Gresham, K. & Koch, W. J. G protein-coupled receptor kinases as therapeutic targets in the heart. Nature Reviews Cardiology 16, 612-622, doi:10.1038/s41569-019- 0220-3 (2019). 22 Gether, U. & Kobilka, B. K. G Protein-coupled Receptors: II. MECHANISM OF AGONIST ACTIVATION*. Journal of Biological Chemistry 273, 17979-17982, doi:https://doi.org/10.1074/jbc.273.29.17979 (1998). 23 Martin, A. L., Steurer, M. A. & Aronstam, R. S. Constitutive Activity among Orphan Class-A G Protein Coupled Receptors. PLOS ONE 10, e0138463, doi:10.1371/journal.pone.0138463 (2015). 24 Schihada, H., Shekhani, R. & Schulte, G. Quantitative assessment of constitutive G protein–coupled receptor activity with BRET-based G protein biosensors. Science Signaling 14, eabf1653, doi:10.1126/scisignal.abf1653 (2021). 118 25 Gether, U. et al. Structural Instability of a Constitutively Active G Protein-coupled Receptor: AGONIST-INDEPENDENT ACTIVATION DUE TO CONFORMATIONAL FLEXIBILITY*. Journal of Biological Chemistry 272, 2587-2590, doi:https://doi.org/10.1074/jbc.272.5.2587 (1997). 26 Samama, P., Cotecchia, S., Costa, T. & Lefkowitz, R. J. A mutation-induced activated state of the beta 2-adrenergic receptor. Extending the ternary complex model. Journal of Biological Chemistry 268, 4625-4636, doi:https://doi.org/10.1016/S0021-9258(18)53442- 6 (1993). 27 Seifert, R. & Wenzel-Seifert, K. Constitutive activity of G-protein-coupled receptors: cause of disease and common property of wild-type receptors. Naunyn-Schmiedeberg's Archives of Pharmacology 366, 381-416, doi:10.1007/s00210-002-0588-0 (2002). 28 Dixon, R. A. et al. Cloning of the gene and cDNA for mammalian beta-adrenergic receptor and homology with rhodopsin. Nature 321, 75-79, doi:10.1038/321075a0 (1986). 29 Bunzow, J. R. et al. Cloning and expression of a rat D2 dopamine receptor cDNA. Nature 336, 783-787, doi:10.1038/336783a0 (1988). 30 Libert, F. et al. Selective amplification and cloning of four new members of the G protein- coupled receptor family. Science 244, 569-572, doi:10.1126/science.2541503 (1989). 31 Vassilatis, D. K. et al. The G protein-coupled receptor repertoires of human and mouse. Proc Natl Acad Sci U S A 100, 4903-4908, doi:10.1073/pnas.0230374100 (2003). 32 Chung, S., Funakoshi, T. & Civelli, O. Orphan GPCR research. Br J Pharmacol 153 Suppl 1, S339-S346, doi:10.1038/sj.bjp.0707606 (2008). 119 33 Sakurai, T. et al. Orexins and orexin receptors: a family of hypothalamic neuropeptides and G protein-coupled receptors that regulate feeding behavior. Cell 92, 573-585, doi:10.1016/s0092-8674(00)80949-6 (1998). 34 Meunier, J. C. et al. Isolation and structure of the endogenous agonist of opioid receptor- like ORL1 receptor. Nature 377, 532-535, doi:10.1038/377532a0 (1995). 35 Tatemoto, K. et al. Isolation and characterization of a novel endogenous peptide ligand for the human APJ receptor. Biochem Biophys Res Commun 251, 471-476, doi:10.1006/bbrc.1998.9489 (1998). 36 Civelli, O., Saito, Y., Wang, Z., Nothacker, H. P. & Reinscheid, R. K. Orphan GPCRs and their ligands. Pharmacol Ther 110, 525-532, doi:10.1016/j.pharmthera.2005.10.001 (2006). 37 Lin, H. H. et al. Multivalent protein probes for the identification and characterization of cognate cellular ligands for myeloid cell surface receptors. Methods Mol Biol 531, 89-101, doi:10.1007/978-1-59745-396-7_7 (2009). 38 Tang, X.-l., Wang, Y., Li, D.-l., Luo, J. & Liu, M.-y. Orphan G protein-coupled receptors (GPCRs): biological functions and potential drug targets. Acta Pharmacologica Sinica 33, 363-371, doi:10.1038/aps.2011.210 (2012). 39 Civelli, O. et al. G Protein–Coupled Receptor Deorphanizations. Annu Rev Pharmacol Toxicol 53, 127-146, doi:10.1146/annurev-pharmtox-010611-134548 (2013). 40 Laschet, C., Dupuis, N. & Hanson, J. The G protein-coupled receptors deorphanization landscape. Biochem Pharmacol 153, 62-74, doi:10.1016/j.bcp.2018.02.016 (2018). 41 Tang, H. et al. Constitutively active BRS3 is a genuinely orphan GPCR in placental mammals. PLOS Biology 17, e3000175, doi:10.1371/journal.pbio.3000175 (2019). 120 42 Fredriksson, R., Lagerström, M. C., Lundin, L.-G. & Schiöth, H. B. The G-Protein- Coupled Receptors in the Human Genome Form Five Main Families. Phylogenetic Analysis, Paralogon Groups, and Fingerprints. Molecular Pharmacology 63, 1256, doi:10.1124/mol.63.6.1256 (2003). 43 Heiber, M. et al. Isolation of Three Novel Human Genes Encoding G Protein-Coupled Receptors. DNA and Cell Biology 14, 25-35, doi:10.1089/dna.1995.14.25 (1995). 44 Song, Z.-H., Modi, W. & Bonner, T. I. Molecular Cloning and Chromosomal Localization of Human Genes Encoding Three Closely Related G Protein-Coupled Receptors. Genomics 28, 347-349, doi:https://doi.org/10.1006/geno.1995.1154 (1995). 45 Saeki, Y. et al. Molecular cloning of a novel putative G protein-coupled receptor (GPCR21) which is expressed predominantly in mouse central nervous system. FEBS Letters 336, 317-322, doi:https://doi.org/10.1016/0014-5793(93)80828-I (1993). 46 Song, Z. H., Young, W. S., Brownstein, M. J. & Bonner, T. I. Molecular cloning of a novel candidate G protein-coupled receptor from rat brain. FEBS Letters 351, 375-379, doi:https://doi.org/10.1016/0014-5793(94)00888-4 (1994). 47 Pundir, S., Martin, M. J. & O'Donovan, C. UniProt Tools. Curr Protoc Bioinformatics 53, 1.29.21-21.29.15, doi:10.1002/0471250953.bi0129s53 (2016). 48 Morales, P., Isawi, I. & Reggio, P. H. Towards a better understanding of the cannabinoid- related orphan receptors GPR3, GPR6, and GPR12. Drug Metab Rev 50, 74-93, doi:10.1080/03602532.2018.1428616 (2018). 49 Kooistra, A. J. et al. GPCRdb in 2021: integrating GPCR sequence, structure and function. Nucleic Acids Research 49, D335-D343, doi:10.1093/nar/gkaa1080 (2021). 121 50 Koikkalainen, J. et al. Shape variability of the human striatum—Effects of age and gender. NeuroImage 34, 85-93, doi:https://doi.org/10.1016/j.neuroimage.2006.08.039 (2007). 51 Haber, S. N. Corticostriatal circuitry. Dialogues Clin Neurosci 18, 7-21, doi:10.31887/DCNS.2016.18.1/shaber (2016). 52 Uhlén, M. et al. Tissue-based map of the human proteome. Science 347, 1260419, doi:10.1126/science.1260419 (2015). 53 Padmanabhan, S., Myers, A. G. & Prasad, B. M. Constitutively active GPR6 is located in the intracellular compartments. FEBS Letters 583, 107-112, doi:https://doi.org/10.1016/j.febslet.2008.11.033 (2009). 54 GPR6 Human Brain RNA Expression. Human Protein Atlas (2021). 55 Lobo, M. K., Cui, Y., Ostlund, S. B., Balleine, B. W. & William Yang, X. Genetic control of instrumental conditioning by striatopallidal neuron–specific S1P receptor Gpr6. Nature Neuroscience 10, 1395-1397, doi:10.1038/nn1987 (2007). 56 Tanaka, S., Ishii, K., Kasai, K., Yoon, S. O. & Saeki, Y. Neural Expression of G Protein- coupled Receptors GPR3, GPR6, and GPR12 Up-regulates Cyclic AMP Levels and Promotes Neurite Outgrowth*. Journal of Biological Chemistry 282, 10506-10515, doi:https://doi.org/10.1074/jbc.M700911200 (2007). 57 Uhlenbrock, K., Gassenhuber, H. & Kostenis, E. Sphingosine 1-phosphate is a ligand of the human gpr3, gpr6 and gpr12 family of constitutively active G protein-coupled receptors. Cellular Signalling 14, 941-953, doi:https://doi.org/10.1016/S0898- 6568(02)00041-4 (2002). 58 Laun, A. S., Shrader, S. H. & Song, Z.-H. Novel inverse agonists for the orphan G protein- coupled receptor 6. Heliyon 4, e00933-e00933, doi:10.1016/j.heliyon.2018.e00933 (2018). 122 59 Ignatov, A., Lintzel, J., Kreienkamp, H. J. & Schaller, H. C. Sphingosine-1-phosphate is a high-affinity ligand for the G protein-coupled receptor GPR6 from mouse and induces intracellular Ca2+ release by activating the sphingosine-kinase pathway. Biochem Biophys Res Commun 311, 329-336, doi:10.1016/j.bbrc.2003.10.006 (2003). 60 Hodges, A. et al. Regional and cellular gene expression changes in human Huntington's disease brain. Human Molecular Genetics 15, 965-977, doi:10.1093/hmg/ddl013 (2006). 61 Benoit, M. E. et al. C1q-induced LRP1B and GPR6 Proteins Expressed Early in Alzheimer Disease Mouse Models, Are Essential for the C1q-mediated Protection against Amyloid-β Neurotoxicity*. Journal of Biological Chemistry 288, 654-665, doi:https://doi.org/10.1074/jbc.M112.400168 (2013). 62 Oeckl, P., Hengerer, B. & Ferger, B. G-protein coupled receptor 6 deficiency alters striatal dopamine and cAMP concentrations and reduces dyskinesia in a mouse model of Parkinson's disease. Exp Neurol 257, 1-9, doi:10.1016/j.expneurol.2014.04.010 (2014). 63 Oeckl, P. & Ferger, B. Increased susceptibility of G-protein coupled receptor 6 deficient mice to MPTP neurotoxicity. Neuroscience 337, 218-223, doi:10.1016/j.neuroscience.2016.09.021 (2016). 64 McGregor, M. M. & Nelson, A. B. Circuit Mechanisms of Parkinson’s Disease. Neuron 101, 1042-1056, doi:https://doi.org/10.1016/j.neuron.2019.03.004 (2019). 65 Pyne, S. & Pyne, N. J. Sphingosine 1-phosphate signalling in mammalian cells. Biochemical Journal 349, 385-402, doi:10.1042/bj3490385 (2000). 66 Spiegel, S. & Milstien, S. Sphingosine-1-phosphate: signaling inside and out. FEBS Letters 476, 55-57, doi:https://doi.org/10.1016/S0014-5793(00)01670-7 (2000). 123 67 Mangmool, S. & Kurose, H. G(i/o) protein-dependent and -independent actions of Pertussis Toxin (PTX). Toxins (Basel) 3, 884-899, doi:10.3390/toxins3070884 (2011). 68 Southern, C. et al. Screening β-Arrestin Recruitment for the Identification of Natural Ligands for Orphan G-Protein–Coupled Receptors. Journal of Biomolecular Screening 18, 599-609, doi:10.1177/1087057113475480 (2013). 69 Yin, H. et al. Lipid G protein-coupled receptor ligand identification using beta-arrestin PathHunter assay. J Biol Chem 284, 12328-12338, doi:10.1074/jbc.M806516200 (2009). 70 Alexander, S. P. et al. THE CONCISE GUIDE TO PHARMACOLOGY 2017/18: G protein-coupled receptors. Br J Pharmacol 174 Suppl 1, S17-S129, doi:10.1111/bph.13878 (2017). 71 Laun, A. S., Shrader, S. H., Brown, K. J. & Song, Z.-H. GPR3, GPR6, and GPR12 as novel molecular targets: their biological functions and interaction with cannabidiol. Acta Pharmacologica Sinica 40, 300-308, doi:10.1038/s41401-018-0031-9 (2019). 72 Laun, A. S. & Song, Z.-H. GPR3 and GPR6, novel molecular targets for cannabidiol. Biochemical and Biophysical Research Communications 490, 17-21, doi:https://doi.org/10.1016/j.bbrc.2017.05.165 (2017). 73 Shrader, S. H. & Song, Z.-H. Discovery of endogenous inverse agonists for G protein- coupled receptor 6. Biochemical and biophysical research communications 522, 1041- 1045, doi:10.1016/j.bbrc.2019.12.004 (2020). 74 Huang, S. M. et al. An endogenous capsaicin-like substance with high potency at recombinant and native vanilloid VR1 receptors. Proc Natl Acad Sci U S A 99, 8400-8405, doi:10.1073/pnas.122196999 (2002). 124 75 Chu, C. J. et al. N-oleoyldopamine, a novel endogenous capsaicin-like lipid that produces hyperalgesia. J Biol Chem 278, 13633-13639, doi:10.1074/jbc.M211231200 (2003). 76 Beeley NRA, B. D., Chalmers DT, Menzaghi F, Strah-Pleynet S Arena Pharmaceuticals, Inc. USA. WO2001062765A2 Small molecule modulators of G protein-coupled receptor six. (2001). 77 Hitchcock S, M. H., Reichard H, Sun H, Kikuchi S, Macklin T, Hopkins M Envoy Therapeutics, Inc. USA. . WO2014028479A1 Quinoxaline derivatives as GPR6 modulators. (2014). 78 Adams ME, B. J., Hitchcock S, Kikuchi S, Lam B, Monenschein H, Reichard H, Sun H Takeda Pharmaceutical Company Limited. Japan WO2015123533A1 Pyrazines as modulators of GPR6. (2015). 79 Brown J, H. S., Hopkins M, Kikuchi S, Monenschein H, Reichard H, Schleicher K, Sun H, Macklin T Takeda Pharmaceutical Company Limited. Japan. WO2015095728A1 Pyridopyrazines modulators of GPR6. (2015). 80 Hitchcock S, H. M., Kikuchi S, Monenschein H, Reichard H, Sun H, Macklin T Takeda Pharmaceutical Company Limited. Japan. . WO2015123505A1 Pyridopyrazines modulators of GPR6. (2015). 81 Sun, H. et al. First-Time Disclosure of CVN424, a Potent and Selective GPR6 Inverse Agonist for the Treatment of Parkinson’s Disease: Discovery, Pharmacological Validation, and Identification of a Clinical Candidate. Journal of Medicinal Chemistry 64, 9875-9890, doi:10.1021/acs.jmedchem.0c02081 (2021). 82 Munk, C. et al. An online resource for GPCR structure determination and analysis. Nature Methods 16, 151-162, doi:10.1038/s41592-018-0302-x (2019). 125 83 Chun, E. et al. Fusion partner toolchest for the stabilization and crystallization of G protein- coupled receptors. Structure 20, 967-976, doi:10.1016/j.str.2012.04.010 (2012). 84 Privé, G. G. et al. Fusion proteins as tools for crystallization: the lactose permease from Escherichia coli. Acta Crystallogr D Biol Crystallogr 50, 375-379, doi:10.1107/s0907444993014301 (1994). 85 Engel, C. K., Chen, L. & Privé, G. G. Insertion of carrier proteins into hydrophilic loops of the Escherichia coli lactose permease. Biochim Biophys Acta 1564, 38-46, doi:10.1016/s0005-2736(02)00398-x (2002). 86 Cherezov, V. et al. High-resolution crystal structure of an engineered human beta2- adrenergic G protein-coupled receptor. Science 318, 1258-1265, doi:10.1126/science.1150577 (2007). 87 Guan, X. M., Kobilka, T. S. & Kobilka, B. K. Enhancement of membrane insertion and function in a type IIIb membrane protein following introduction of a cleavable signal peptide. Journal of Biological Chemistry 267, 21995-21998, doi:https://doi.org/10.1016/S0021-9258(18)41623-7 (1992). 88 Alexandrov, A. I., Mileni, M., Chien, E. Y., Hanson, M. A. & Stevens, R. C. Microscale fluorescent thermal stability assay for membrane proteins. Structure 16, 351-359, doi:10.1016/j.str.2008.02.004 (2008). 89 Lv, X. et al. In vitro expression and analysis of the 826 human G protein-coupled receptors. Protein Cell 7, 325-337, doi:10.1007/s13238-016-0263-8 (2016). 90 Popov, P. et al. Computational design of thermostabilizing point mutations for G protein- coupled receptors. eLife 7, e34729, doi:10.7554/eLife.34729 (2018). 126 91 Landau, E. M. & Rosenbusch, J. P. Lipidic cubic phases: a novel concept for the crystallization of membrane proteins. Proc Natl Acad Sci U S A 93, 14532-14535, doi:10.1073/pnas.93.25.14532 (1996). 92 Caffrey, M. & Cherezov, V. Crystallizing membrane proteins using lipidic mesophases. Nature Protocols 4, 706-731, doi:10.1038/nprot.2009.31 (2009). 93 Cherezov, V. et al. Rastering strategy for screening and centring of microcrystal samples of human membrane proteins with a sub-10 microm size X-ray synchrotron beam. J R Soc Interface 6 Suppl 5, S587-597, doi:10.1098/rsif.2009.0142.focus (2009). 94 Otwinowski, Z. M., W., . Processing of X-ray diffraction data collected in oscillation mode. . Methods Enzymol 276, 307-326 (1997). 95 McCoy, A. J. et al. Phaser crystallographic software. J Appl Crystallogr 40, 658-674, doi:10.1107/S0021889807021206 (2007). 96 Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr D Biol Crystallogr 66, 486-501, doi:10.1107/S0907444910007493 (2010). 97 Tickle, I. J., Flensburg, C., Keller, P., Paciorek, W., Sharff, A., Vonrhein, C., Bricogne, G. , . STARANISO. Cambridge, United Kingdom: Global Phasing Ltd ( 2018). 98 Cerevance Beta, I. Study of CVN424 in Parkinson's Disease Patients With Motor Fluctuations. doi:NCT04191577 (2022). 99 Brice, N. L. et al. Development of CVN424: A Selective and Novel GPR6 Inverse Agonist Effective in Models of Parkinson Disease. J Pharmacol Exp Ther 377, 407-416, doi:10.1124/jpet.120.000438 (2021). 127 100 Liu, S. et al. Differential activation mechanisms of lipid GPCRs by lysophosphatidic acid and sphingosine 1-phosphate. Nature Communications 13, 731, doi:10.1038/s41467-022- 28417-2 (2022). 101 Xing, C. et al. Cryo-EM Structure of the Human Cannabinoid Receptor CB2-Gi Signaling Complex. Cell 180, 645-654.e613, doi:https://doi.org/10.1016/j.cell.2020.01.007 (2020). 102 Lomize, M. A., Lomize, A. L., Pogozheva, I. D. & Mosberg, H. I. OPM: Orientations of Proteins in Membranes database. Bioinformatics 22, 623-625, doi:10.1093/bioinformatics/btk023 (2006). 103 Hanson, M. A. et al. A Specific Cholesterol Binding Site Is Established by the 2.8 Å Structure of the Human β2-Adrenergic Receptor. Structure 16, 897-905, doi:https://doi.org/10.1016/j.str.2008.05.001 (2008). 104 Chrencik, J. E. et al. Crystal Structure of Antagonist Bound Human Lysophosphatidic Acid Receptor 1. Cell 161, 1633-1643, doi:10.1016/j.cell.2015.06.002 (2015). 105 Hanson, M. A. et al. Crystal structure of a lipid G protein-coupled receptor. Science 335, 851-855, doi:10.1126/science.1215904 (2012). 106 Hua, T. et al. Crystal Structure of the Human Cannabinoid Receptor CB1. Cell 167, 750- 762 e714, doi:10.1016/j.cell.2016.10.004 (2016). 107 Shao, Z. et al. High-resolution crystal structure of the human CB1 cannabinoid receptor. Nature 540, 602-606, doi:10.1038/nature20613 (2016). 108 Li, X. et al. Crystal Structure of the Human Cannabinoid Receptor CB2. Cell 176, 459-467 e413, doi:10.1016/j.cell.2018.12.011 (2019). 109 Isberg, V. et al. GPCRdb: an information system for G protein-coupled receptors. Nucleic Acids Research 44, D356-D364, doi:10.1093/nar/gkv1178 (2016). 128 110 Shoichet, B. K. & Kobilka, B. K. Structure-based drug screening for G-protein-coupled receptors. Trends in Pharmacological Sciences 33, 268-272, doi:10.1016/j.tips.2012.03.007 (2012). 111 Wishart, D. S. et al. HMDB 4.0: the human metabolome database for 2018. Nucleic Acids Res 46, D608-d617, doi:10.1093/nar/gkx1089 (2018). 112 Abagyan, R. A., Orry, A., Raush, E., Budagyan, L. & Totrov, M. ICM User’s Guide and Reference Manual v.3.9, MolSoft. (2021). 113 Lane, T. J., Bowman, G. R., Beauchamp, K., Voelz, V. A. & Pande, V. S. Markov State Model Reveals Folding and Functional Dynamics in Ultra-Long MD Trajectories. Journal of the American Chemical Society 133, 18413-18419, doi:10.1021/ja207470h (2011). 114 Lovera, S., Cuzzolin, A., Kelm, S., De Fabritiis, G. & Sands, Z. A. Reconstruction of apo A2A receptor activation pathways reveal ligand-competent intermediates and state- dependent cholesterol hotspots. Scientific Reports 9, 14199, doi:10.1038/s41598-019- 50752-6 (2019). 115 Taylor Bryn, C., Lee Christopher, T. & Amaro Rommie, E. Structural basis for ligand modulation of the CCR2 conformational landscape. Proceedings of the National Academy of Sciences 116, 8131-8136, doi:10.1073/pnas.1814131116 (2019). 116 Huang, W. et al. Structural insights into µ-opioid receptor activation. Nature 524, 315-321, doi:10.1038/nature14886 (2015). 117 Rasmussen, S. G. et al. Structure of a nanobody-stabilized active state of the β(2) adrenoceptor. Nature 469, 175-180, doi:10.1038/nature09648 (2011). 118 Wacker, D. et al. Structural features for functional selectivity at serotonin receptors. Science 340, 615-619, doi:10.1126/science.1232808 (2013). 129 119 Gault, J. et al. High-resolution mass spectrometry of small molecules bound to membrane proteins. Nature Methods 13, 333-336, doi:10.1038/nmeth.3771 (2016). 120 Gupta, K. et al. Identifying key membrane protein lipid interactions using mass spectrometry. Nature Protocols 13, 1106-1120, doi:10.1038/nprot.2018.014 (2018). 121 Sibley, D. R. & Neve, K. in xPharm: The Comprehensive Pharmacology Reference (eds S. J. Enna & David B. Bylund) 1-13 (Elsevier, 2007). 122 Danev, R. et al. Routine sub-2.5 Å cryo-EM structure determination of GPCRs. Nature Communications 12, 4333, doi:10.1038/s41467-021-24650-3 (2021). 123 Maeda, S. et al. Endogenous agonist–bound S1PR3 structure reveals determinants of G protein–subtype bias. Science Advances 7, eabf5325, doi:10.1126/sciadv.abf5325. 124 Katritch, V. et al. Allosteric sodium in class A GPCR signaling. Trends Biochem Sci 39, 233-244, doi:10.1016/j.tibs.2014.03.002 (2014). 125 Yanamala, N. & Klein-Seetharaman, J. Allosteric Modulation of G Protein Coupled Receptors by Cytoplasmic, Transmembrane and Extracellular Ligands. Pharmaceuticals (Basel) 3, 3324-3342, doi:10.3390/ph3103324 (2010). 126 Yu, J. et al. Determination of the melanocortin-4 receptor structure identifies Ca2+ as a cofactor for ligand binding. Science 368, 428-433, doi:10.1126/science.aaz8995 (2020).

Abstract (if available)

Abstract G Protein-Coupled Receptors (GPCR) comprise the largest superfamily of membrane proteins in the human genome with over 800 members, each featuring a unique sequence. GPCRs mediate many of our physiological responses to hormones, neurotransmitters, and environmental stimulants (ions, lipids, etc.) and therefore are implicated in various pathophysiological conditions. Hence, in addition to their endogenous ligands, GPCRs are the target of over 30% of the drugs approved by the United States Food and Drug Administration (FDA) which highlights the tremendous therapeutic potential for this class of membrane proteins. Growing number of high- resolution structures of GPCRs over the past decades has provided insight into their function and enabled structure-based drug design for several GPCR targets. However, the structure and function of many GPCRs with high therapeutic potentials remains elusive.
In this study, we utilized x-ray crystallography to determine the structure of the orphan G Protein- Coupled Receptor 6 (GPR6) in various conformational states of the receptor. Our structural data are supported and complemented by insights from computational biology, pharmacology, cellular biology, and protein biochemistry. Collectively, our data elucidate the structure of GPR6 and shed light on ligand binding, signal propagation, and activation of this receptor.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Exploring the structure of G protein-coupled receptors

PDF

Structural studies of the human leukotriene B4 receptor 1

PDF

Structure – dynamics – function analysis of class A GPCRs and exploration of chemical space using integrative computational approaches

PDF

Biochemical regulation of hunger by the melanocortin 4 receptor

PDF

Structural studies of nicotinic acetylcholine receptors and their regulatory complexes

PDF

Developing peptide and antibody-mimetic ligands for the cell surface receptors β2AR and DC-SIGN

PDF

Cone arrestin 4 contributes to vision, cone health, and desensitization of the dopamine receptor D4

PDF

Functional analyses of androgen receptor structure pertaining to prostate cancer

PDF

Physical principles of membrane mechanics, membrane domain formation, and cellular signal transduction

PDF

The folding, misfolding and refolding of membrane proteins and the design of a phosphatidylserine-specific membrane sensor

PDF

Antibody discovery and structural studies of the viral oncogene ORF74

PDF

Computer aided inhibitor design for fungal bromodomains

PDF

Molecular basis of CD33 receptor function, and effects of a destabilizing transmembrane motif on αXβ2 integrin

PDF

Unlocking tools in chemistry to facilitate progress in drug discovery

PDF

Modeling biochemical components and systems through data integration and digital arts approaches

PDF

Engineering artificial cells to elucidate GPCR 5-HT1AR activity and plasma membrane compositional dependence

PDF

X-ray structural studies on DNA-dependent protein kinase catalytic subunit:DNA co-crystals

PDF

The symmetry of membrane protein polyhedral nanoparticles

PDF

Non-invasive live-cell imaging for monitoring and evaluating pancreatic islet and beta cell metabolism

PDF

Role of translocator protein (18 kDa) in cell function and metabolism

Asset Metadata

Creator Barekatain, Mahta (author)

Core Title Structure and function of the orphan G protein-coupled receptor 6

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Chemistry

Degree Conferral Date 2022-08

Publication Date 07/21/2024

Defense Date 05/05/2022

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag basal activity,constitutive activity,G protein-coupled receptor,GPCRs,GPR6,OAI-PMH Harvest,orphan receptors,protein engineering,signaling,structural biology,structure and function,X-ray crystallography

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Cherezov, Vadim (committee chair), Fraser, Scott E. (committee member), Katritch, Vsevolod (committee member), White, Kate L (committee member)

Creator Email mahtabarekatain@gmail.com,mbarekat@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC111373693

Unique identifier UC111373693

Legacy Identifier etd-Barekatain-10891

Document Type Dissertation

Format application/pdf (imt)

Rights Barekatain, Mahta

Type texts

Source 20220722-usctheses-batch-959 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu