Accepted Manuscript GPCR Crystal Structures: Medicinal Chemistry in the Pocket Jeremy Shonberg, Ralf C. Kling, Peter Gmeiner, Stefan Löber PII: DOI: Reference:

S0968-0896(14)00882-7 http://dx.doi.org/10.1016/j.bmc.2014.12.034 BMC 11966

To appear in:

Bioorganic & Medicinal Chemistry

Received Date: Revised Date: Accepted Date:

26 September 2014 12 December 2014 16 December 2014

Please cite this article as: Shonberg, J., Kling, R.C., Gmeiner, P., Löber, S., GPCR Crystal Structures: Medicinal Chemistry in the Pocket, Bioorganic & Medicinal Chemistry (2014), doi: http://dx.doi.org/10.1016/j.bmc. 2014.12.034

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

GPCR Crystal Structures: Medicinal Chemistry in the Pocket Jeremy Shonberg#,*, Ralf C. Kling#, Peter Gmeiner, Stefan Löber Department of Chemistry and Pharmacy, Medicinal Chemistry, Emil Fischer Center, Friedrich Alexander University, Schuhstraße 19, 91052 Erlangen, Germany # Authors contributed equally * For correspondence: [email protected]

Abstract Recent breakthroughs in GPCR structural biology have significantly increased our understanding of drug action at these therapeutically relevant receptors, and this will undoubtedly lead to the design of better therapeutics. In recent years, crystal structures of GPCRs from classes A, B, C and F have been solved, unveiling a precise snapshot of ligandreceptor interactions. Furthermore, some receptors have been crystallized in different functional states in complex with antagonists, partial agonists, full agonists, biased agonists and allosteric modulators, providing further insight into the mechanisms of ligand-induced GPCR activation. It is now obvious that there is enormous diversity in the size, shape and position of the ligand binding pockets in GPCRs. In this review, we summarise the current state of solved GPCR structures, with a particular focus on ligand-receptor interactions in the binding pocket, and how this can contribute to the design of GPCR ligands with better affinity, subtype selectivity or efficacy.

Introduction G Protein-Coupled Receptors (GPCRs) are a large family of membrane-bound receptors that are encoded by more than 800 genes in the human genome.1 GPCRs function by the detection of a wide variety of chemicals, or other stimuli including photons, ions and proteins. For most GPCRs, the endogenous ligands are well known, and generally binding of the natural ligand results in a change of receptor conformation, enabling the coupling of effector proteins such as heterotrimeric G proteins or β-arrestins. This is followed by the activation of a cascade of intracellular signalling proteins, eventually resulting in a physiological response. An imbalance of chemicals in the body can therefore disrupt the normal regulation of a physiological response by the GPCR, and this makes GPCRs a highly attractive target for numerous disease therapies; including central nervous system disorders, inflammatory diseases, cancer, metabolic imbalances, cardiac diseases, monogenic diseases and many more.2 As a result, GPCRs are the target of approximately 30% of currently marketed drugs. However, despite our advances in understanding of GPCRs as drug targets, substantial challenges remain in our understanding of the mechanisms of drug action at these receptors.3 The family of GPCRs have been classified according to their pharmacological properties into four main sub-families: class A rhodopsin like (the largest family, and most well studied); class B secretin-like, class C metabotropic glutamate/pheromone, and class F frizzled receptors.4 Historically, drug design targeting GPCRs involved mimicking the endogenous ligand, resulting in synthetic activation of the GPCR signalling pathways. This is a particularly pertinent approach for a disease where biochemical synthesis of the endogenous ligand (for example dopamine in Parkinson’s disease) is reduced. Another classic approach to GPCR-based drug discovery is the design of a structurally related compound that competes with the natural ligand, but does not cause receptor activation, known as an antagonist. This method of blocking the signal of a receptor has shown utility in many disease states, for example the use of β-blockers in the treatment of cardiac disorders such as hypertension. Recent advances in the understanding of the structure and function of GPCRs have resulted in an explosion of possible new methods of targeting GPCRs. For example, many novel GPCR ligands have been shown to act in a topographically distinct binding site from that of the endogenous ligand.5 These ‘allosteric ligands’ can modulate the actions of the endogenous ligand to make it more potent (positive allosteric modulator), less potent (negative allosteric modulator), identical (neutral) or even have potency of its own (ago-allosteric). Other molecules can act simultaneously at the orthosteric and allosteric sites (bitopic ligands),6 or even bridge a dimeric construct of GPCRs to bind simultaneously at binding sites of adjacent receptors (bivalent ligands).7,8 Other novel mechanisms of drug action include biased agonism, where a ligand can stabilize a unique receptor conformation resulting in the biased activation of one signalling pathway over another.9 The power of drug discovery at GPCRs will be increasingly enhanced by progress in structural biology unveiling the precise structures of the receptors. The overall topology of GPCRs, with 7-transmembrane spanning regions, was first shown in 1993 from twodimensional crystals of rhodopsin,10 and seven years later the first high-resolution structure of rhodopsin was solved.11 For some time, rhodopsin served as the only atomistic scaffold for the structural investigation of GPCRs, and a significant amount of our current knowledge on the function and activation of GPCRs was originally derived from studies on rhodopsin.12,13 More recently, different crystal structures of rhodopsin have been solved, including dark-state rhodopsin, light-activated opsin (also coupled to the C-terminal fragment of transducin), and constitutively-active receptor-constructs. Whilst rhodopsin was crucial for the initial structural

understandings of GPCRs,14 it does not receive significant focus in this review due to its lack of druggability. The last 14 years, and in particular the last five years, have seen an explosion in the solving of GPCR crystal structures in multiple states. The result of this breakthrough in the understanding of GPCR structure and function has fostered an advance in the design of GPCR ligands through computational methods for better understanding ligand-receptor interactions. Homology models of related unsolved structures can also be generated, and other techniques include molecular dynamics simulations to understand the binding mode of ligands, and virtual ligand screening. This has resulted in a more detailed understanding of the binding pockets of GPCRs to small molecules has enhanced the ability for medicinal chemists to design novel compounds. In this review, we summarize the current status of GPCR crystal structures that have been solved, and present the information with a focus on the ligand binding sites, and key ligandreceptor interactions for the design of novel molecules. Furthermore, we focus on advances in bioorganic and medicinal chemistry that have been accomplished from these structures, and also focus on the history of the ligands that have been used in complex with the receptor for crystallography (see Table 1 for a comprehensive list of ligands co-crystallized with GPCRs) to gain an appreciation of ligand choice. With the current state of GPCR structural biology, we expect medicinal chemistry efforts targeting GPCRs in the future to be significantly enhanced for the synthesis of better, more selective, potent, drug-like compounds. We consider a structure to be in an “active-state” when the particular GPCR has been crystallized simultaneously in the presence of an agonist and an intracellular binding partner, such as a G protein or a G protein mimetic nanobody, which is capable of stabilizing the particular outward movement of TM6 known to accompany GPCR activation.15 In contrast, a crystal structure is defined to be in the “inactive state” when bound to an antagonist or inverse agonist, and is thus devoid of coupling to intracellular effectors. For the purpose of this review, we do not further distinguish between intermediate-states of receptor activation (for example, a GPCR structure crystallized in the presence of an agonist but in the absence of an intracellular binding partner) as this would be beyond the scope of this review.14

β-Adrenergic Receptors The adrenergic receptor subtypes β1-(β1AR) and β2- (β2AR) belong to the superfamily of G protein-coupled receptors (GPCRs). β1AR is mainly located in the heart and the kidney, where it is involved in physiological processes including the regulation of heart beat and blood pressure.16 Activation of β 2AR is known to result in the dilatation of smooth muscles of the lung, the uterus and blood vessels.16 β1AR and β2AR can be activated by the endogenous catecholamines norepinephrine and epinephrine, and the affinity of epinephrine was found to be higher at β2AR.17 Therapeutically, antagonists or inverse agonists for β1AR, so called βblockers, are applied to medicate diseases like hypertension and heart failure.18 In contrast, agonists at β2AR, known as β2-sympathomimetics, are most frequently used for the treatment of respiratory problems caused by asthma or chronic obstructive pulmonary disease (COPD).19 To avoid side effects caused by an undesired binding of drugs to both β1AR and β2AR, subtype-selective compounds are of great importance. Although this requirement is mainly fulfilled for the class of β2-sympathomimetics, recent studies demonstrate only a minor β1AR-selectivity for β-blockers.20,21

For some time now, β-adrenergic receptors have played an important role for the investigation of GPCRs. This is particularly true for β2AR, which has been frequently employed as a model system to study the function of hormone-activated GPCRs.15 β2AR was the first GPCR to be sequenced and cloned after rhodopsin, and amongst the first GPCRs for which radioligand binding assays had been developed.22 In addition, due to significant progress in enhancing expression levels, optimizing purification, and improving techniques for crystallization,15,23 βadrenergic receptors were the first GPCRs activated by diffusible ligands for which crystal structures could be solved (see Figure 4 for a list of all chemical structures). To date, 32 individual crystal structures for both β1AR and β2AR have been reported.24-42 In 2007, inactive-state crystal structures of human β2AR coupled to the antagonist carazolol were published,24-26 followed by the elucidation of the inactive-state crystal structure of turkey β1AR bound to the antagonist cyanopindolol in 2008.36 To account for their highly flexible nature, and to successfully capture them in a crystal, different methods were used to enhance the stability of β-adrenergic receptors, including removal or truncation of highly flexible Nand C-terminal domains.15 The stability of inactive-state β1AR was further enhanced by the introduction of several thermostabilizing point mutations.36,43 In addition, the presence of sodium ions was found to enhance the stability of inactive-state β1AR, without influencing the affinity of agonists or intracellular binding partners.42 However, the effect of sodium ions can vary depending on the particular GPCR investigated, and is summarized in more detail in the literature.44 At β2AR, the flexible intracellular loop 3 (IL3) was either stabilized by an additional antibody fragment (Fab5)26 or totally replaced by the well crystallizing protein T4lysozyme (T4L).24,25 Unfortunately, it was not possible to resolve the extracellular domains (including the binding pocket) of the β2AR-Fab5 complex.26 In contrast, these parts were well defined using β2AR-T4L, but the insertion of T4L was found to sterically prevent the activation of G proteins.24 As an alternative to the replacement of IL3, it has been successfully demonstrated that T4L can be attached to the truncated N-terminus of β2AR without affecting the global conformation, i.e. an inactive-state structure with the capacity of this particular receptor-construct to couple to G proteins in biological test systems.23,33

The cytoplasmic domains of β1AR and β2AR were found to form similar conformations of IL1, but exhibited differences at IL2 and the resolved parts of IL3, i.e. the intracellular tips of TM5 and TM6.24-26,36 In the case of IL3, these differences are thought to be of minor physiological relevance, as they were likely a product of introduction of stabilizing mutations and insertions.36 For IL2, a β-helical structure was observed at β1AR, whereas an extended loop conformation could be found at β2AR, which was originally proposed to have an impact on G protein activation.36 However, the absence of a helical structure for IL2 at β2AR was later suggested to be due to different crystal lattice contacts involving IL2.30 In contrast to these differences, at both β1AR and β2AR a non-interacting, open conformation of the socalled “ionic lock” was detected. Given that this ionic interaction between residue Arg3.50 (in this review we use the Ballesteros-Weinstein numbering system, which enumerates GPCR transmembrane residues in X.YY format in superscript, where X is the helix number and YY is the residue position relative to the most conserved residue in the helix, designated X.50)45 of the highly conserved DRY-motif at the bottom of TM3 and Glu6.30 at the bottom of TM6 was suggested to stabilize inactive-state GPCRs,46 a missing interaction was rather surprising. However, in the case of β1AR, additional crystal structures of thermostabilized β1AR constructs (co-crystallized with different antagonists) revealed both open and closed

conformations of the ionic lock, which may explain the high basal activity of β-adrenergic receptors.38 The dynamic nature of this interaction has been corroborated by MD simulations performed on inactive-state β2AR, which reported an equilibrium between open and closed ionic-lock conformations and that the presence of T4L leads to an increased probability to capture non-interacting conformations.47

For the first time since the rhodopsin structure was solved in 2000, the GPCR community was provided with fully atomistic structures of β-adrenergic receptors, which could be used to facilitate a structure-based drug design.41,48,49 Several examples have demonstrated the successful utility of β-adrenergic crystal structures as a template for structure-guided drug discovery, predominantly based on β1AR and β2AR.41,48 The crystal structures of β1AR and β2AR showed a highly similar overall architecture. In both cases, flexible extracellular loop (EL) regions grant access for the diffusible ligand to the orthosteric binding pocket, comprised of residues from TM3, TM5, TM6, TM7 and EL2.24,36 As part of their binding process, drugs associate with the extracellular receptor surface before they lose their hydration shell and enter the orthosteric binding pocket; a phenomenon visualized by long scale molecular dynamics simulations using adrenergic crystal structures as templates.50 Focusing on the binding modes of cyanopindolol and carazolol at β1AR and β2AR, respectively, the structures revealed common features of ligand binding at β-adrenergic receptors, including: a salt bridge between the positively charged ammonium moiety of the ligands and Asp3.32; hydrogen bonds between the aromatic head groups of the ligands and Ser5.42; and, hydrogen bonds between the aliphatic OH-group of the ligands and Asn7.39, the latter of which determine stereospecific recognition of β-adrenergic ligands.24,36 A set of additional crystal structures coupled to chemically diverse antagonists/inverse agonists (ICI118,55148 and alprenolol, see Figure 4), confirmed a well conserved binding mode, and it was demonstrated that the β2AR was capable of adapting to these ligands without significant conformational changes.28 Interestingly, most of the amino acids, which were identified to form interactions with carazolol and cyanopindolol, are completely conserved between β1AR and β2AR.24,51 Thus, these structures not only revealed similar binding modes for the co-crystallized ligands, but also highly conserved orthosteric binding pockets of the receptors. Given this pronounced sequence identity in the binding site, it is not surprising that the development of subtypeselective compounds represents a very challenging task.20,21 Nevertheless, the extracellular loop regions (EL1-3) were found to exhibit a higher variability within their amino-acid sequences.36 Thus, these differences offer medicinal chemists the opportunity for a targeted development of subtype-selective, allosteric modulators or so called bitopic (or dualsteric) compounds, which simultaneously address both the canonical orthosteric binding pocket and a less conserved allosteric cavity. The role of the extracellular surface in subtype-selective ligand binding was complemented by NMR experiments, which reported ligand-specific conformations for extracellular loop regions upon activation and thus a conformational cooperativity between these regions and the orthosteric binding pocket.32 Therefore, there is a possibility that compounds targeting the extracellular surface may be beneficial for subtypeselectivity at the highly related receptors.32

Inactive-state crystal structures of timolol-bound β2AR27, and of possible dimeric structures of β1AR40 have also been recently published. The solved crystal structure of the β2AR-timolol

complex revealed a structurally relevant binding site for cholesterol molecules between TM1, TM2, TM3 and TM4, which may attribute a possible role to cholesterol, and cholesterolderived compounds, in modulating the stability of inactive GPCRs.27 In the case of the β1AR structure, the receptor construct was purified in the presence of cyanopindolol, which had been removed (but with unknown efficiency) before starting the crystallization experiments.40 The solved structure of the dimeric β1AR revealed two distinct dimer interfaces, with one interface comprised of residues from TM1, TM2, helix 8 and EL1 and the second one involving TM4, TM5, IL2 and EL2.40 Additional cross-linking experiments confirmed that β1AR is capable of forming dimers via both interfaces under physiological conditions.40 Therefore, these dimer interfaces may serve as structural scaffolds to better understand the role of dimerization (and oligomerization) on GPCR signaling.40 In addition, a more detailed knowledge about specific dimeric conformations will facilitate the targeted development of bivalent ligands, which are compounds composed of two covalently linked binding pharmacophores capable of simultaneously binding at two receptor protomers. This strategy has shown previous utility in drug design targeting GPCRs.8

Efforts to obtain crystals of an agonist-bound receptor have generally been hindered by the low affinity of agonists (in contrast to antagonists/inverse agonists used for crystallization trials) and their relatively rapid on- and off-rates, which can result in incomplete occupancy of the receptor, and thus conformational heterogeneity.15 To overcome these drawbacks, the inactive-state structure of β2AR was used to design a tailor-made agonist, which is covalently attached to an H2.64A mutant of β2AR by a disulfide bridge.29 In 2011, the first agonist-bound crystal structure of human β2AR was solved in complex with a covalent agonist (see Figure 1 for ligand binding pocket comparison between the diffusible ligand carazolol, and the covalent ligand FAUC50).29 With this structure in hand, it was possible to demonstrate that an agonist alone is insufficient to fully stabilize the active-state of a receptor.52 Even in the presence of the non-dissociating agonist FAUC50, β2AR was found to crystallize in an inactive receptor conformation, suggesting that the absence of an intracellular binding partner prevents the generation of fully active GPCR crystal structures.29 However, the agonist-bound structure revealed subtle differences within the binding pocket of β2AR compared to the antagonist/inverse agonist-bound structures. Most notably, the formation of hydrogen bonds between Ser5.42 and Ser5.46 with the catechol-replacing ring of FAUC50, thereby altering the conformation of both serine residues.29 Ligand-receptor interactions with these serine residues have also been shown in other aminergic GPCRs to be crucial for driving ligand efficacy, and thus agonism.53 Therefore, this structure represented a significant breakthrough in the structural understanding of GPCR activation.

Figure 1. Comparison of the ligand binding poses of the inverse agonist carazolol (dark blue), and the non-dissociating agonist FAUC50 (light blue) which stabilized an inactive conformation of the receptor. Both ligands were found to bind in the same binding pocket, and make key interactions with Asp 3.32 and Ser5.42, but importantly only the agonist FAUC50 could make hydrogen bonding interactions with Ser5.46, which has been previously shown to drive agonism and efficacy. Refer to Figure 4 for chemical structures of carazolol and FAUC50.

Shortly after the agonist-bound structure of the human β2AR was solved, crystal structures of turkey β1AR in complex with various partial and full agonists were published, all of which representing an inactive-state receptor conformation.37 The agonist-occupied binding pockets were found to be contracted by 1 Å compared to the antagonist- or inverse agonist-bound equivalents, caused by a subtle inward movement of TM5. In agreement with observations made at the agonist-bound β2AR, the full agonists were found to stabilize the contracted binding pockets by simultaneously engaging the TM5 residues, Ser5.42 and Ser5.46 (Figure 2 (A) and (B)). In contrast, the binding pockets of the solved structures in complex with partial agonists were found to lack a hydrogen bonding interaction to Ser5.46, which is consistent with a weaker stabilization of the contracted binding pocket (Figure 2, (C) and (D)). These solved structures therefore demonstrated a structural explanation for ligand-directed control of efficacy at β-adrenergic receptors.37

Figure 2. Ligand-receptor interactions of full and partial agonists within the β1AR binding pocket. The full agonists (A) Isoprenaline and (B) Carmoterol engage the TM5 residues Ser5.42 and Ser5.46; whereas the partial agonists (C) Salbutamol and (D) Dobutamine were found to engage Ser5.42, but not Ser5.46, a potential explanation for ligand-directed control of efficacy at β-adrenergic receptors. Refer to Figure 4 for chemical structures of the ligands.

Adding to the advances in structural biology, thermo-stabilized avian β1AR was crystallized in complex with biased agonists, bucindolol and carvedilol.39 The term “biased (or functionally selective) ligand” refers to the capacity of a ligand to stabilize a unique conformation of the receptor, and consequently modulate the coupling of a GPCR with a certain subset of signal transducers, thereby preferentially activating some pathways in a biased manner.9,22 In functional assays at the β1AR, both bucindolol and carvedilol have been shown to preferentially stimulate non-G protein pathways such as β-arrestin recruitment, while acting as antagonists for G protein signaling.39 The head groups of bucindolol and carvedilol were found to adopt a similar conformation compared to previously obtained crystal structures of antagonist-bound β-adrenergic receptors. In contrast, both compounds form additional contacts to residues within an extended binding pocket comprised of TM2, TM3, TM7 and EL2, which has been suggested to specifically stimulate distinct receptor conformations responsible for biased signaling.39 A comparable observation has been made using 19F-NMR spectroscopy at β2AR, which reported certain pathway-specific conformational changes caused by biased agonists for the cytoplasmic domains of TM6 and TM7.54 However, crystal structures coupled to biased ligands and different intracellular binding partners, complemented by techniques which account for the dynamic nature of GPCRs (NMR-/fluorescence spectroscopy, MD simulations), seem to be important to fully understand the structural basis of functionally selective signaling.

In 2011, a significant milestone improving our understanding of structural consequences caused by agonist-induced receptor activation came with the elucidation of the crystal structure of active-state β2AR.30,31 As the dynamic nature of β2AR is enhanced for the agonistbound state,29,55,56 a key step towards capturing such transient active-state conformations was

implementation of an intracellular binding partner, which was either a G protein-mimicking nanobody (Nb80, the heavy chain of a camelid antibody)30 or the native, nucleotide-free G protein.31 In the latter case, removal of the nucleotides GDP and GTP was necessary for the formation of a stable receptor-G protein complex, given both GDP and GTP had previously been shown to promote dissociation of the G protein from the receptor.31 Both Nb80 and the G protein were shown experimentally to exert a comparable effect on agonist binding and the active-state receptor conformation.30 In addition, both the β2AR-Nb80 and the β2AR-G protein complexes were further stabilized by a high-affinity, slowly dissociating agonist (BI167,107).30,31 The crystallization process was facilitated by the insertion of T4L, either to replace IL3 (β2AR-Nb80)30 or as an attachment to the N-terminus of β2AR (β2AR-G protein).31 As the fully active-state of a GPCR was defined as that conformation, which “couples to and stabilizes a nucleotide-free G protein”,31 the β2AR-G protein complex represents an unprecedented, and detailed snapshot of a GPCR at the moment of signal transduction. However, both active-state structures can serve as invaluable scaffolds for the structure-based development of agonists57,58 or, in the case of the G protein-bound β2AR, as a structural template with which to study receptor-G protein selectivity on a molecular level.59,60

These ternary complexes consisting of an agonist-bound receptor which is further stabilized by an intracellular binding partner, revealed significant conformational changes of the intracellular surface upon receptor activation.30,31 An outward movement of TM6 and, to a minor extent, TM5, accompanied by an inward movement of TM3 and TM7, resulted in the exposure of a binding pocket for the respective intracellular binding partner. For the β2AR-G protein complex, the outward movement of TM6 is even more pronounced compared to the nanobody-derived structure.31 Otherwise, the overall conformations of active-state β2AR were virtually identical. In both complexes, several highly conserved GPCR-domains undergo conformational changes compared to inactive-state β2AR, including Arg3.50 of the DRY-motif, Tyr7.53 of the NPxxY-motif and Tyr5.58 at the bottom of TM5.30,31 Focusing on the β2AR-G protein complex, the G protein interacting domains were found to be comprised of residues at the bottom of TM3, TM5 and TM6 and residues from IL2, with Arg3.50 of TM3 representing the upper edge of the G protein binding site.31 Unlike the significant conformational changes observed at the cytoplasmic receptor surface, the extracellular domains (including the binding pocket) of inactive- and active-state β2AR exhibit less pronounced differences. Hydrogen bond interactions of the agonist BI167,107 to residues Ser5.42 and Ser5.46 lead to an inward movement of TM5 around Ser5.46, thereby altering the conformation of a hydrophobic switch comprised of Ile3.40, Pro 5.50 and Phe6.44, which culminates in the aforementioned rearrangements at the intracellular surface of β2AR.30 Long-term MD simulations employing the β2AR-Nb80 complex have suggested this hydrophobic switch to allosterically connect ligand-binding and conformational changes at the intracellular receptor surface.50 Specific interactions of the head group of the agonist and Asn6.55, which is simultaneously hydrogen bonded to Ser5.43 and Tyr7.35, enable Tyr7.35 to interact with Phe5.32, thereby adopting a conformation above the binding pocket, which was found to be more closed towards the extracellular surface.30 However, both active-state structures showed an unaltered conformation of Trp 6.48,30 which had been suggested to change its rotameric state upon receptor activation.61

In order to further improve the quality of active-state crystal structures, directed evolution of Nb80 has been used to enhance the affinity of this nanobody towards agonist-bound β2AR, while slowing down its dissociation rate from the receptor and maintaining its conformational selectivity for active-state β2AR.34 Incorporation of the developed antibody Nb6B9, yielded crystal structures of active-state β2AR coupled to the high-affinity agonists (see Figure 4) BI167,107 and hydroxybenzyl isoproterenol (HBI) and even the low-affinity, endogenous agonist epinephrine (adrenaline), all of which exhibited an improved resolution compared to the parent β2AR-Nb80 structure.34 In contrast to the previously obtained β2AR-Nb80 structure, an N-terminally attached T4L-β2AR construct was used.34 Owing to the higher resolution of the β2AR-Nb80 structures, several water molecules were identified, including one water molecule that mediates a hydrogen bond between Tyr5.58 of TM5 and Tyr7.53 of the NPxxY-motif at the bottom of TM7, which is only possible in the active state.34 On the extracellular side, water molecules were found to connect the conformation of residues on EL3 with those of TM6 from the binding pocket of β2AR.34 The β2AR-Nb6B9 structures revealed a comparable overall architecture and a similar binding mode for the co-crystallized agonists, suggestive of a conserved activation mode at β2AR.34 The most interesting exception was found to be a more contracted conformation of the extracellular surface around EL3 in the presence of catecholamine agonist. A slightly modified engagement of residue Asn6.55 due to the smaller head group of the catecholamine agonists (compared to the sterically more demanding hydroxychinolinone moiety of BI167,107) leads to a conformational rearrangement of the adjacent His6.58, connected to Asn6.55 via a structural water molecule, which in turn influences the conformation of EL3 and represents another example for the capacity of different ligands to modify the conformation of the extracellular surface of β2AR in a ligand-dependent manner.32,34

By converting the low-affinity endogenous agonist epinephrine into a covalently attaching probe (FAUC37) using a flexible linker based on the parent compound, FAUC50 (Figure 4),29 it was possible to increase the stability of the ternary epinephrine-β2AR-Nb6B9 complex.35 The obtained crystal structure showed a pronounced similarity to the previously crystallized epinephrine-β2AR-Nb6B9 complex (Figure 3 (A)), whereby the additionally introduced disulfide-linker adopts a conformation within an extended cavity, which entirely exploits the additional space of the epinephrine-structure and leaves the binding mode of the catecholamine moiety unchanged (Figure 3 (B)).35 In addition, it was successfully demonstrated that such a strategy can be applied to other aminergic GPCRs including pharmacologically important dopamine (D2R), serotonin (5-HT2A) and histamine (H1R) receptors.35

Figure 3. (A) Overall structure of the endogenous ligand adrenaline co-crystallized with the β2AR stabilized by Nb6B9, and zoomed cut-through section demonstrating the ligand binding pocket. (B) The overall structure of the adrenaline derivative FAUC37, covalently attached to β2ARH93C. The zoomed cut-through section demonstrating the binding pocket shows how the covalently attached linker occupies an extended cavity, without altering the conformation of the adrenaline moiety.

Figure 4. Chemical structures of adrenergic receptor agonists, partial agonists, antagonists and inverse agonists that have been crystallized in complex with an adrenergic receptor.

Adenosine A2A Receptor (A2AR) Adenosine is a purine nucleoside that plays an important role in numerous biochemical processes, and exerts its effects through activation of four guanine nucleotide-binding GPCR subtypes, A1, A2A, A2B and A3.62 Each of these receptors is linked to a variety of transduction mechanisms, where A1 (the most abundant in the brain) is coupled to activation of K+ channels and inhibition of Ca2+ channels; A2A and A2B are primarily linked to activation of adenylyl cyclase; and, A3 modulates the activity of other receptors.63 In response to adenosine, the four receptors play a critical role in the central nervous system for the regulation of sleep, arousal, neuroprotection and epilepsy.63 Caffeine, the most widely used psychoactive drug in the world, along with other naturally occurring methylxanthines, have been shown to block the actions of adenosine at A1, A2A and A2B receptors;64 and, evidence has suggested an inverse association of caffeine with the development of neurodegenerative diseases such as Parkinson’s disease.65 In particular, this effect has been linked to interaction with the A2AR, together with the D2R and mGluRs. The A2AR was first discovered when preparations from the striatum, as opposed to other parts of the brain, were stimulated by adenosine and resulted in activation of adenylyl cyclase.66 Further analysis on brain slices demonstrated two subclasses of A2 receptors, a high affinity receptor (A2AR) and a low affinity receptor (A2BR) for adenosine.64 Until the mid-1990’s, ligand discovery targeting the A2AR generally focussed on substituted adenosine derivatives, ribose/purine modified adenosine derivatives, and substituted 5′-Nalkylcarboxamidoadenosine (NECA) derivatives.67 The first, high affinity, selective, nonxanthine A2AR antagonist, ZM241385 (Figure 5), was discovered by Zeneca pharmaceuticals, and demonstrated improved solubility compared to previously developed selective xanthinebased ligands, which also allowed the study of A2AR antagonists in vivo.68 Unsurprisingly, given the therapeutic relevance of the A2AR, the discovery of ZM241385 spurred numerous efforts to synthesize selective antagonists and agonists, many of which have reached clinical trials, including the recently approved istradefylline (KW-6002).69

Figure 5. Chemical structure of the endogenous agonist, adenosine; the classical agonist NECA, and UK-432097; and antagonists for the A2AR including caffeine, ZM241385 and KW-6002.

The first crystal structure of the A2AR was solved in complex with ZM241385 by the Stevens group.70 The structure presented unique features that were distinct from other structures available at the time (rhodopsin, β1AR and β2AR), including a different organization of extracellular loops; an extended conformation of the ligand in the binding pocket towards extracellular space; and, a subtle divergence of helical positions which redefined the binding site.70 This demonstrated that the ligand binding pocket was located in a different position and orientation, relative to other structures available at the time, and that ligand selectivity could be achieved through targeting hydrophobic residues extending from the aromatic core of ZM241385.70 Prior to the publication of the A2AR crystal structure, the GPCR Dock 2008 competition assessed the status of GPCR structure prediction and ligand docking by challenging the computational chemistry community to predict the structure of the A2A-ZM241385 complex.71 Accurate prediction of the ligand-receptor complex in the binding site was found to be challenging; unsurprising given the low homology between the A2Areceptor and the structures of receptors available at the time, namely β-adrenergic receptors and rhodopsin. The release of the A2A-ZM241385 complex coordinates allowed significant advances in GPCR structure-based ligand discovery using virtual ligand screening (VLS). A study by Carlsson et al. demonstrated the use of VLS with the A2A-ZM241385 structure, and determined seven ligands that were structurally dissimilar to known adenosine ligands at the time (see Figure 6 for examples).72 A similar approach was taken by Katrich et al., who discovered at least 9 novel chemical scaffolds for A2AR antagonism using VLS based on the A2A-ZM241385 complex.73 Furthermore, extensive research has used the A2A-ZM241385 complex for ligand design, and structural understandings of ligand-receptor interactions for SAR studies, for example in the design of ligands to extend toward extracellular space for attachment of functionality, so called ‘functionalized congeners’.74,75

NH2 O HN

NH N H

F

Ki-A2A = 0.2 M

N H

N N O

N

N NH N

N

O

HN N

N H

Ki-A2A = 0.2 M

O Cl Ki-A2A = 0.03 M

Figure 6. Chemical structure of ligands discovered by VLS at the A2AR,72,73 which demonstrated high affinity, and significant structural novelty compared to classical A2AR ligands based on the adenosine structure.

Further advancements in structural biology have demonstrated the ability to discover novel A2AR antagonists using Structure Based Drug Design (SBDD), as illustrated by Heptares Therapeutics.76 These structures were achieved through incorporation of A2A- T4L fusion constructs, and thermostabilised receptor (StaR) constructs, which enabled the generation and solving of X-ray diffraction patters. Crystallization efforts of the A2AR have also resulted in significant advances to medicinal chemistry,77 with numerous efforts demonstrating successful Virtual Screening campaigns, as well as SBDD. Furthermore, crystallization efforts have been used to predict binding modes of clinical agents such as Lu AA47070,78 the design of A2AR agonists,79 and in the screening for other adenosine receptor subtypes.80 Since the publication of the first A2AR crystal structure, the active structures of the A2AR in complex with agonists UK-432097,81 adenosine and NECA,82 (Figure 5) were published in 2011 by the Stevens and Tate groups, respectively. The actions of agonists at the A2AR have been investigated for the treatment of cardiovascular mediation, ischemia-reperfusion injury, and most notably for systemic inflammation.83 Stimulation of A2AR on human neutrophil surfaces has been demonstrated to potently inhibit various neutrophil functions, which are important in the mediation of inflammation through, for example, superoxide anion radicals (O2-). Thus, selective agonism of the A2AR has shown clinical relevance with marketed A2AR specific agonists, such as regadenoson.84 Generally, the majority of A2AR agonists mimic the structure of adenosine, and require a bicyclic adenine core and a ribose ring for activation of the receptor, which differentiates them from corresponding antagonists, such as ZM241385. More recently, agonists without a ribose group (e.g. LUF5834) have been demonstrated to activate the receptor, and also interact with distinct residues compared to classic adenosine-derived agonists.85 The agonist UK-432097 was discovered at Pfizer in the search for selective A2AR agonists for the treatment of inflammation,86, and as an inhalant for treatment of COPD. In 2013, the compound was terminated from Phase II clinical trials following poor efficacy results. The agonist-bound structure of the A2AR in complex with UK-432907 was solved utilizing an engineered human A2AR-T4L-∆C complex, where T4 lysozyme was inserted into IL3, and the C-terminal was truncated by 96 residues. UK-432907 exhibited the highest thermal stability (melting temperature ~65 °C) of tested agonists screened for stabilization of the receptor.81 The ligand-binding pocket of UK-432097 demonstrated an extensive network of ligandreceptor interactions, including 11 hydrogen bonds, one aromatic stacking interaction, and numerous non-polar van der Waals interactions.81 Given the common bicyclic core of UK432097 and the majority of adenosine receptor ligands (including ZM241385), the moiety aligned well to the core of A2AR inactive crystal structures. Furthermore, the structure

extending from the bicyclic core was found to extend in a similar manner as ZM241385, and are involved in numerous important interactions that anchor the molecule into the binding pocket. The major defining feature was the presence of the ribose ring of UK-432097, which inserts deeply into a predominantly hydrophilic region of the binding cavity and forms numerous important hydrogen bonds which are important for high affinity binding of A2AR agonists.81 Compared to the antagonist-bound structure, the UK-432097 was found to trigger a series of modest local changes within the binding cavity that promoted large scale rearrangements in the 7TM bundle.81 The most prominent of these include interactions with the ribose ring, common to most A2AR agonists, which results in an inward movement of TM6 by 1.8 Å, an upward shift of the entire TM3 along the helical axis, and a seesaw-like movement of TM7.81 Furthermore, agonist binding resulted in coordinated movements of the intracellular parts of TM5 (inward) and TM6 (outward). The overall structure was found to resemble the structural rearrangements described for the agonist-bound structures of opsin and β2AR (both of which were available at the time of publication), despite each having different molecular triggers in the binding pocket to promote activation. At a similar time, the structures of adenosine and NECA bound to the A2AR were solved, and demonstrated similar structural rearrangements of the transmembrane bundle (r.m.s.d. 0.6 Å) in response to agonist binding.82 The main difference between these structures was in the extracellular surface due to the bulky extensions of UK-432907 interacting with extracellular loops.87 Together, the solved structures of agonist-bound A2AR revealed important structural modifications that occur upon agonist binding. Interestingly, despite the two unique strategies for crystallization (A2AR-T4L-∆C for UK-432097 structure, and A2AR-GL31 with four stabilizing point mutations for NECA and adenosine structures), and using different agonists, similar structural features were present in the agonist-bound receptors. This further enhances our understanding of GPCR activation by small molecules, and gives a significant enhancement to the design of selective agonists at the adenosine receptors in the future. Dopamine D3 Receptor (D3R) The essential neurotransmitter dopamine exerts its effects through activation of five distinct dopamine receptor subtypes, generally classified into two subtypes, D1-like (D1R and D5R) and D2-like (D2R, D3R and D4R).88 The D1-like receptors couple to the Gs G protein, and activate adenylyl cyclase, whereas D2-like receptors inhibit adenylyl cyclase through coupling to the Gi G protein, and activate potassium channels. Dopaminergic tone at these receptors is involved in a variety of critical functions within the central nervous system, including the control of voluntary movement, feeding, reward, and memory. It is therefore not surprising that numerous CNS disorders have been attributed to dopaminergic dysfunction, most notably Parkinson’s disease, schizophrenia and drug addiction.89 The D3R was first characterized in 1990,90 and was found to share a high sequence homology with the D2R within the transmembrane region (78%), particularly in the dopamine binding site. Consequently, most dopaminergic compounds are not subtype selective.91 D3 receptors are elevated in the mesolimbic system of schizophrenic patients, so inhibition of D3R binding sites by selective antagonists is expected to attenuate the positive symptoms of schizophrenia without creating the extrapyramidal side effects (EPS) of the classical D2R antagonists.

Because D3R antagonists also enhance D3-mediated acetylcholine release in the frontal cortex, beneficial effects on the modulation of attention, working memory and social memory (all of which are perturbed in the ‘negative symptoms’ of schizophrenia) might be expected.92 The D3R has also been postulated to be involved in drug dependence and addiction, however the lack of selective compounds has made the hypothesis difficult to prove.93 Some D3R-selective antagonists and partial agonists have been developed and have demonstrated attenuation of drug-seeking behaviour in animal models without associated motor effects.94,95 However, generally these compounds are highly lipophilic, therefore poorly bioavailable, and often toxic. Nonetheless, the discovery of highly selective D3R ligands has been challenging due to high sequence identity of the D2R and D3R orthosteric binding pockets. Recently, reports of very highly selective D3R antagonists (D2R/D3R = 1640) have achieved selectivity (see Figure 7 for general structure of D3R selective ligands)96 by demonstrating the critical role of an extended pharmacophore with a carboxamide linker attaching a lipophilic appendage, as in compound 8j (Figure 7).97 The structure of the D3R was solved in complex with a D2/D3 selective, potent antagonist, eticlopride, and published in 2010 by the Stevens group.98 The binding pocket was found to be similar to the β2AR pocket, as expected given the closely related catecholamine ligand structure, and predominantly consists of a salt bridge between the ionisable tertiary amine of eticlopride and Asp1103.32; and, a hydrophobic cavity formed around the substituted aromatic ring of eticlopride (see Figure 8). Of the 18 eticlopride contact residues in the D3R structure, 17 are identical in the D2R binding site. Accordingly, D3R selectivity has been demonstrated to require ligand extension toward the extracellular opening of the binding pocket (Figure 8 (C)), which can be achieved by molecules with extended flexible linkers connecting the tertiary amine to an extended aryl amide structure.93 Indeed docking studies, presented by the authors of the crystal structure at the D3R, with the D3-selective antagonist R-22, revealed that while the amine-containing aromatic substituent bound in the same binding pocket as eticlopride, the indole-2-carboxamide terminus oriented towards the extracellular end of the binding site consisting of EL2/EL1 and the junction of TM1, TM2 and TM7 to define an extracellular binding pocket.98 Whilst ligand-receptor interactions in this extracellular binding region have been associated with D3R selectivity, it has also been associated with various phenomena at the D2R, including biased agonism,99 and allosteric communication between D2R protomers in a dimer.100

Figure 7. A general scaffold for the design of D3R subtype selective ligands. Chemical structure of the endogenous ligand dopamine, and the small molecule antagonist eticlopride which was co-crystallized with the D3R. Compound 8j is a highly selective (>1000-fold for D3R over D2R) antagonist for the D3R, R-22 is a well-studied D3R selective antagonist. The crystal structure of the D3R has served as a platform for the development of many GPCR homology models. Like the CXCR4, the D3R was subject to the GPCR Dock Competition in 2010, and the coordinates of the solved structure were concealed until the modelling community submitted predicted binding modes of the complexed ligand.101 Unlike the CXCR4, the structure of the D3R-eticlopride complex was well estimated, and the top predictions sufficiently captured the key aspects of ligand recognition and binding.101 An extension of this experiment was subsequently published by the Shoichet group which examined the discovery of D3R ligands based on either homology models, or the crystal structure.102 Surprisingly, the hit rates against both the modelled and experimentally solved structures were essentially equivalent, and returned new scaffolds at a similar rate.102 Another example of novel ligand discovery applying the D3R crystal structure is a virtual ligand screening (VLS) study at the D3R by Lane et al., who identified novel D3R negative allosteric modulators with appreciable hit rates using optimized crystal-structure-based models of the D3R with an empty binding pocket, and in a dopamine complex.103 Given the high sequence homology of the D3R and D2R, and the therapeutic relevance of D2 R antagonism and partial/biased agonism in the treatment of schizophrenia,104 it is no surprise that one of the most common applications of the D3R crystal structure is as the template for construction of D2R homology models, and subsequent computational chemistry to understand ligand-receptor interactions at the D2R. This approach has been demonstrated in numerous attempts at discovering SAR, for example for biased agonists at the D2R,99,105-107 and useful information regarding the mechanisms of ligand-receptor interactions have thus been uncovered.

Figure 8. (A) Overall topology of the solved D3R (grey ribbons) structure, with the antagonist eticlopride (orange spheres) co-crystallized. (B) The D3R binding pocket as seen from extracellular space, with key residues displayed and labelled that are involved in the binding of eticlopride (orange sticks). (C) Surface representation of the D3R binding pocket from extracellular space, with the extended binding pocket shown in green.

Histamine H1 receptor: Histamine is an important hormone and neurotransmitter. It is known to participate in numerous physiological processes including gastric acid secretion, allergic and inflammatory reactions and neurotransmission in the central nervous system. Among the four known histamine receptor subtypes (H1-H4), the H1 receptor is the most relevant drug target. H1antagonists and inverse agonists are therapeutically used as antiallergic, antipruritic, antiemetic and antidepressive agents. However, therapeutics targeting the H1 receptor have long been plagued by adverse effects, and the first generation antihistamines such as doxepin and diphenhydramine (Figure 9) demonstrated poor receptor selectivity and significant bloodbrain barrier penetration, resulting in considerable side effects such as sedation and dry mouth. Second generation antihistamines, such as cetirizine, loratadine and fexofenadine are generally zwitterionic (Figure 9), making them too polar to cross the BBB, and are thus more selective for peripheral H1 receptors and have reduced sedative properties.108

First Generation H1 antagonists

N O

N

O

Diphenhydramine

Doxepin Second Generation H1 antagonists

N

N N

Cl

N

OH

O

Cl O

O O

Loratadine

Cetirizine

HO OH N Fexofenadine

O OH

Figure 9. Chemical structures of first generation H1 receptor antagonists; doxepin, the H1 antagonist crystallized in complex with the H1 receptor, and diphenhydramine, a nonprescription sleeping aid. Chemical structures of second generation H1 receptor antagonists including cetirizine, loratadine and fexofenadine, which are more specific H1 ligands, and too polar for BBB penetration.

In an effort to guide the design of more selective H1 receptor ligands, Shimamura et al. solved the structure of the H1 receptor in complex with the first-generation H1 antagonist doxepin.109 This ligand shows antagonist activity at H2, serotonin 5-HT2, α1-adrenergic, and muscarinic acetylcholine receptors in addition to the inhibition of the reuptake of serotonin and noradrenaline.109,110 The tricyclic moiety of doxepin adopts a deep binding position occupying a highly conserved hydrophobic pocket. Additionally, a unique anion binding site at the entrance region of the receptor was observed. This is occupied by a phosphate ion forming ionic interactions with the protonated tertiary amine. Docking experiments with H1-selective second-generation antihistaminergics featuring an additional carboxylic group revealed that the latter can address the non-conserved anion binding site. This finding elucidates the molecular basis of histamine receptor subtype selectivity. Leurs and coworkers used the H1 crystal structure for the development of a virtual fragment screening protocol resulting in novel “fragment-like” H1 ligands.111 Cordova-Sinjago et al. applied the H1 receptor structure to delineate the molecular determinants for ligand binding at this subtype.112 Moreover, several research groups have employed the H1 structure for homology modelling followed and subsequent docking experiments to investigate ligand selectivity at the histamine H3 and H4 receptors.113-117

Muscarinic receptors

The family of muscarinic acetylcholine receptors includes the five subtypes M1-M5, which are categorized according to their G protein coupling, where M1, M3 and M5 are Gq coupled, whereas M2 and M4 couple with the Gi/GO family of G proteins. Muscarinic acetylcholine receptors have a widespread tissue distribution, modulate a variety of physiological functions and are potential therapeutic agents in Alzheimer's disease, schizophrenia, Parkinson's disease and chronic obstructive pulmonary disease (COPD).118 The M2 receptor has a key role in modulating cardiac function and other important central processes including cognition and pain perception, and has long served as a model system in GPCR pharmacology. The M3 subtype mediates peripheral physiological functions like smooth muscle contraction and glandular secretion, and has a role in the central nervous system including the regulation of food intake as well as learning and memory functions. The discovery of highly subtype selective orthosteric ligands at M2 and M3 receptors has been challenged by a high sequence identity in the acetylcholine binding site. More recently, considerable progress has been made in the design of muscarinic ligands targeting a topographically distinct allosteric binding site, which are generally less well conserved than the orthosteric binding site.119 In 2012 the crystal structures of the T4L-modified M2 (human) and M3 (rat) receptors in complex with the antagonists 3-quinuclidinyl-benzilate (QNB) and tiotropium (Figure 10), respectively, were published.120,121 These structures provide insights in the subtle differences of the M2 and M3 binding sites (Figure 11), thereby creating the basis for the development of subtype selective muscarinic receptor ligands. The comparison of both structures shows that the receptors are highly homologous. In particular, the orthosteric binding pockets are virtually identical, and the only significant difference is a Phe181 (M2, Figure 11 (A)) versus Leu225 (M3, Figure 11 (B)) sequence in the second extracellular loop, which creates an enlarged binding pocket in the M3 receptor, and may facilitate the development of subtype selective ligands. Furthermore, the authors detected a significant difference in the interhelical distance of TM5 and TM6, particularly at the cyctoplasmic end of the domains. This finding may explain the different G protein binding selectivity of the M2 and M3 subtype. Additionally, molecular dynamics simulations indicated that tiotropium dissociates significantly slower from the M3 receptor when compared to M2, which can be attributed to the different flexibilities of the second extracellular loops of the two subtypes. Recent data based on site-directed mutagenesis experiments and molecular dynamics studies122 of the M3R–tiotropium complex support this finding, and thus explain tiotropium’s long duration of action observed in COPD therapy.

Figure 10. Chemical structures of muscarinic antagonists, QNB and Tiotropium, cocrystallized with the M2 and M3 receptors; M2 receptor agonist iperoxo, and the mustard derivative FAUC123 used for stabilization of muscarinic active states; muscarinic positive allosteric modulator, LY2119620, that was co-crystallized with the M2 receptor.

Figure 11. Comparison of (A) M2R binding pocket in complex with the antagonist QNB; and (B) M3R binding pocket in complex with the antagonist Tiotropium. The binding pocket sequences differ only by Phe181 (M2R) vs Leu225 (M3R), which causes the M3R to have a

larger pocket (5.0 Å compared to 3.7 Å in M2R), and therefore presents an opportunity for the design of more subtype selective compounds This example demonstrated that crystal structures of closely related and highly homologous receptors can be extremely valuable for drug design.123

In 2013, the Kobilka group published the structure of the M2 receptor in complex with the agonist iperoxo.124,125 This active structure represented the third GPCR active-state structure to be solved, along with the β2AR and rhodopsin. For this achievement, the selection of an appropriate Gprotein-mimetic camelid antibody to stabilize the receptor in the active state was found to be the crucial step. The llama nanobodies were displayed on yeast surface, which was then selectively stained with both agonist and inverse-agonist occupied M2 receptor populations, whereas different fluorophores were used for the active and inactive state receptors to allow fluorescence-activated cell sorting. To ensure the active state conformation of the tagged receptors, a mustard-type iperoxo derived ligand, FAUC 123, was developed to enable a covalent agonist binding mode. The cytoplasmic end of the iperoxo bound M2R showed the expected outward shift of TM6 in combination with an inward movement of TM7 when compared to its inactive structure, leading to the formation of the Gprotein-binding site. Additionally, conserved side chain assemblies like the DR(E)Y and the NPxxY motif reveal conformational changes similar to those seen in β2AR and in rhodopsin. Within the binding pocket the most pronounced change is observed at TM6. An inward movement of this helix allows the formation of a hydrogen bond between Asn4046.52 and iperoxo. This interaction is possibly involved in the M2 receptor activation. The authors also solved an X-ray crystal structure of the M2 receptor bound to iperoxo in combination with LY2119620 (Figure 12), a positive allosteric modulator. Although this structure is largely congruent with that of receptor and agonist without LY2119620, the allosteric modulator LY2119620 does induce a slight contraction around the allosteric ligand, and a change in conformation of Trp422 7.35 to allow aromatic stacking interactions with LY2119620. The authors assume that iperoxo binding pre-forms the allosteric binding site in the extracellular vestibule that can be selectively addressed by LY2119620. These structures of the muscarinic receptor have provided extensive insight into the binding of drug molecules to GPCRs, in particular for the understanding of receptor subtype selectivity, ligand-induced activation, and binding of ligands to an allosteric binding site.

Figure 12. (A) Top view on the allosteric binding pocket of the allosteric modulator LY2119620 (green) at active-state M2R (grey). For comparison, extracellular domains of inactive-state M2R (blue) are shown, demonstrating extracellular rearrangements of M2R upon activation to better stabilize LY2119620. (B) Close view of ligand-receptor interactions between LY2119620 (green) and extracellular domains of M2R (grey) are depicted. In addition, the orthosteric agonist Iperoxo is shown in orange.

Sphingosine 1-phosphate receptor Sphingosine 1-phosphate (S1P) is a zwitterionic lysophospholipid signalling molecule (Figure 13) that regulates important biological functions within the cardiovascular, immune, and nervous systems including angiogenesis, vascular maturation, heart development and immunity. S1P mediated signals are transduced via five GPCR subtypes named sphingosine 1-phosphate receptor 1-5. The five S1P receptors show a high degree of sequence homology featuring a conserved sphingolipid binding pocket. In the recent years the S1P receptor 1 (S1P1) has become a promising therapeutic target in chronic inflammatory diseases and cancer.126 Additionally, the nonselective S1P agonist prodrug fingolimod (FTY720, Figure 13) was recently approved for the treatment of multiple sclerosis.127 Using different X-ray diffraction data processing methods the Stevens group generated two crystal structures of the T4L-fused S1P1, at resolutions of 3.35 Å and 2.8 Å, respectively,128 in complex with the sphingolipid mimic antagonist (R)-3-amino-(3-hexylphenylamino)-4oxobutylphosphonic acid (ML056, Figure 13). This potent and selective phosphonatecontaining S1P1 antagonist demonstrated enhanced biological properties compared to other phosphate-containing S1P1 receptor ligands, which are limited by low potency, chemical liability, poor solubility and lack of in vivo activity.129 Furthermore, control of ML056 stereochemistry has been demonstrated to enhance the selectivity, stability and in vivo efficacy of the compound, whereby the R enantiomer is a potent competitive antagonist, whereas the S enantiomer was found to be significantly less potent.129 The crystal structure of the S1P1 features some significant differences to previously solved GPCR structures. Most notably, the N terminus, together with EL1, cap the entrance of the ligand binding pocket. The authors assume that the ligand enters the binding site from within the proximate cell membrane utilizing a gap between TM1 and TM2. The binding mode of

ML056 demonstrated strong ionic interactions of charged amino acid side chains with the zwitterionic head group. The lipophilic tail of the ligand is located in an aromatic pocket formed by residues of TM3-7. Previous SAR studies revealed that structural ligand modifications can induce an antagonist-agonist switch.130 The Stevens group performed mutagenesis and docking studies and could thereby elucidate the role of the hydrophobic binding site for receptor activation and subtype selectivity.

Figure 13. Chemical structures of the endogenous S1P; the marketed S1P agonist prodrug fingolimod; and the S1P1 receptor antagonist, ML056 which was co-crystallized with the S1P1 receptor.

Serotonin receptors: Serotonin (5-hydroxytryptamine, 5-HT) is one of the most important neurotransmitters in the human brain and is involved in the control of numerous brain functions. It is the endogenous ligand of a total of 15 receptors that are grouped into seven families (5-HT1-7). These GPCRs (except 5-HT3 which is an ion channel) are important drug targets for a wide variety of psychiatric and neurological disorders including depression, anxiety, migraine, nausea and vomiting. Additionally, serotonin regulates important peripheral processes like cardiovascular function, bowel motility and bladder control.131 Due to a high sequence identity amongst serotonin receptor subtypes, generation of subtype selective ligands targeting the 5-HT receptors can be a challenging task, and can result in severe side-effects associated with serotonergic drug treatment. For example, 5-HT1B receptor agonism is an important target in the treatment of migraine, but off-target stimulation of the closely related 5-HT2B subtype can result in cardiotoxic effects by thickening of heart valves.132

The crystal structures of the 5-HT1B and 5-HT2B receptors in complex with ergotamine (Figure 14) were published by the Stevens group in 2013 (Figure 15 (A)).133,134 To obtain a stabilized receptor suitable for crystallization without altering ligand interactions, both receptors were fused with the bacterial protein BRIL. Additionally, the 5-HT1B subtype was crystallized in complex with dihydroergotamine. Due to the high sequence identity in the orthosteric binding pockets of 5-HT1B and 5-HT2B receptors, the binding of the ergoline moiety of ergotamine in both receptor subtypes was found to be almost identical (Figure 15 (A)). The only substantial difference was the conformation of the benzyl group due to differences in the extended ligand binding pockets.

Figure 14. Chemical structure of ergotamine (with double bond as dashed line) and dihydroergotamine (where dashed line is absent), the ligands used to stabilize the 5-HT1B and 5-HT2B receptors. The authors could identify some key differences between the two receptors which may facilitate the design of more subtype selective ligands, including an outward shift of the extracellular part of TM5 in the 5-HT1B subtype by 3 Å relative to the 5-HT2B receptor, which creates a broadened accessory binding pocket. Furthermore, on close inspection of the binding pockets, the difference between Tyr2.64 (5-HT1B) vs Thr2.64 (5-HT2B) results in a larger cavity in 5-HT2B due to the smaller size of the residue (see Figure 15 (B) and (C) orange circles). This is another binding site difference that can be potentially exploited for the design of selective ligands. Additionally, differences in the intracellular receptor regions could be detected, which could explain the fact that ergolines show biased signalling in favour of the β-arrestin pathways at the 5-HT2B receptor, whereas signalling at 5-HT1B receptors appears to be non-biased.135 The ergotamine bound 5-HT1B structure features attributes of an agonist-induced active-like state, whereas the respective 5-HT2B receptor shows conformational characteristics of both the active and inactive state. The authors hypothesized that ergotamine stabilized a 5-HT2B receptor conformation that does not allow G protein signalling. These elusive structural insights can help to understand ligand-induced differential signalling and help to design novel functionally selective ligands.

Figure 15. (A) Overlay of ergotamine in complex with the 5-HT1B (blue) and 5-HT2B (green) receptors from a side-view, demonstrating the homologous binding mode of the ergoline moiety, the conformational change of the benzyl group, and the substantially inward position of TM5 in 5-HT2B. (B, C) Extracellular view of the binding sites demonstrating cavities (orange circles) for which the design of subtype selective ligands could utilize. In cooperation with the Stevens group, Cherezov and co-workers reported on a 5-HT2B receptor structure obtained by a different crystallographic technique.136 Using lipidic cubic phase derived microcrystals in combination with an x-ray free-electron laser (XFEL) with individual 50-femtosecond-duration x-ray pulses they were able to obtain a high-resolution room-temperature structure. The authors claim that this methodology is suitable for the generation of GPCR structures the represent the “real” receptor conformation more precisely.

Neurotensin receptor: Neurotensin (NT) is a 13-amino-acid neuromodulatory peptide (Figure 16) that exerts its function through activation of the GPCR subtypes NTS1 and NTS2, or through subtype 3 (NTS3) which belongs to the Vps10p family of sorting receptors. Besides its peripheral effects like regulation of digestive processes, neurotensin is known to modulate the activity of dopaminergic systems, opioid-independent analgesia, and the inhibition of food intake. The majority of the known NT promoted effects are mediated through the preferentially Gqcoupled NTS1 subtype,137 which is expressed in the CNS and periphery, and is associated with

the analgesic and antipsychotic actions of NT. The six residue carboxy-terminal fragment of NT, NT(8-13) (Figure 16) was found to be the portion responsible for agonism, with potent analgesic effects in mice.138 Mutagenesis and modelling studies have addressed some aspects of ligand binding at the NTS1, but many aspects of ligand binding remained relatively poorly understood at the molecular level.139 In 2012, Grisshammer and co-workers reported on the Rattus norvegicus crystal structure of neurotensin receptor 1 bound to the full agonist peptide NT(8–13), the C-terminal hexapeptide fragment of neurotensin, and this represents the only agonist peptide to be crystallized in complex with a GPCR.139 The structure was solved by engineering of a T4L-fusion receptor construct including six thermostabilizing mutations, which resulted in a crystal structure solved at 2.8 Å resolution.

Figure 16. Peptide sequence of neurotensin, and the fragment of the six residues NT(8-13) that, independently of the remaining sequence, potently activates NTS1. This fragment was crystallized in complex with the NTS1. The binding mode of NT(8-13) demonstrated significant differences in the ligand position compared to agonists in the β2ARand the A2AR. NT(8-13) does not deeply penetrate the receptor, and the binding cavity is located near the receptor surface. The peptidic agonist adopts an extended conformation nearly perpendicular to the membrane plane. This ligand conformation is in good agreement with that analysed by solid-state NMR using NT(8-13) bound to the wild-type NTS1.140 Comparison of the NTS1 structure with rhodopsin and the β2ARshowed that many features of an active state receptor are present. On the other hand, the authors demonstrated that the mutated thermostabilized NTS1 (without T4L modification) does not catalyze nucleotide exchange at Gαq in response to NT, which suggests that the presented structure does not represent the active conformation when looking at the intracellular surface of the receptor, but rather an active-like state. Plückthun and coworkers have recently reported on a directed evolution technology approach allowing the expression of thermostabilized NTS1 in E. coli as a robust host for GPCR overexpression system.141 The authors obtained three NTS1 structures in complex with neurotensin, with up to 2.75 Å resolution, using vapour diffusion crystallization experiments. Pharmacological characterization of the presented NTS1 included 11 point mutations and a shortened IL3 in order to confer sufficient expression levels, enhanced stability in detergent solution and good crystallization properties. However, the presented structures show an “inactive state type” conformation of the cytosolic ends of TM5 and TM6, which are reported to tilt outwards when adopting the active state. Human Chemokine Receptor 4 (CXCR4)

The first structure of the chemokine receptor family, the human chemokine receptor 4 (CXCR4), was published in 2010 by the Stevens/Wu group.142 The CXCR4 is broadly expressed in the immune and central nervous systems, and mediates migration of resting leukocytes and haematopoietic progenitors following activation by CXCL12, the natural ligand.143 The CXCR4 has also been demonstrated to be involved in neuronal cell migration and patterning,144 breast cancer metastasis,145 viral entry of human immunodeficiency virus (HIV),146,147 and inflammation.148 A range of ligands are known to antagonise the CXCR4,149 including the marketed antagonist Plerixafor (AMD3100, Figure 17), which disrupts adhesive tumor-stroma interactions and induces rapid mobilization of leukaemia cells from their protective marrow microenvironment,150 making them more accessible to conventional drugs. The non-peptidic AMD3100,151 is an example of mono- or bistetraazamacrocycle-AZT conjugates that have been developed to bind to the CXCR4 at the cell surface, driving antiviral AZT compounds into the cells. Other antagonists have been categorized, including: small peptide CXCR4 antagonists such as T140,152 anti-CXCR4 antibodies,153 and modified CXCL12 derivatives.154

Figure 17. Chemical structures of CXCR4 ligands, including the marketed Plerixafor, and IT1t, an antagonist used to stabilize the CXCR4 for crystallization. Despite this wealth of CXCR4 antagonists, many aspects of ligand binding and signalling were poorly understood prior to the publication of the crystal structure, in particular the understanding of the CXCR4 binding site to small molecules. The crystal structure of the CXCR4 was solved in complex with a cyclic peptide antagonist CVX15, and with the antagonist IT1t (Figure 17);142 a small drug-like, isothiourea derivative that was developed by Novartis as a highly potent, orally bioavailable, selective CXCR4 antagonist.155 The binding site of IT1t was found to be larger, more open, and located closer to the extracellular surface compared with other GPCR structures (Figure 19 (B)).142 Furthermore, key ligand-receptor interactions were observed, including: a salt bridge between the isothiourea nitrogens and Glu97 2.63, hydrophobic interactions between the cyclohexane rings and small subpockets of the receptor, and a salt bridge between the protonated imidazothiazole and Glu2887.39. Methylation of the isothiourea confirmed the importance of the salt bridge to Glu972.63, resulting in a 100-fold reduction of binding.155 The binding of the bulky 16-residue cyclic peptide CVX15 was found to fill most of the binding pocket volume, and overlap with the binding site of the small molecule IT1t, where IT1t binds at the deeper end of the pocket (Figure 19 (A)). However, CVX15 was found to induce major conformational differences at the base of the N-terminus, and a minor movement of extracellular tips of TM6 (~1.5 Å inwards), TM7 (~1.5 Å tangential) and TM5 (~0.5 Å outward).142

Interestingly, a comparison of the top of the binding pockets (Figure 19 (A, B)) demonstrates that compared to the binding of CVX15 which protrudes out of the receptor into extracellular space, the binding of IT1t is capped by residues in EL2 (demonstrated in PDB-ID 3ODU). This may explain the high potency of IT1t for the CXCR4. The GPCR Dock 2010 community-wide assessment involved a competition of thirty-five groups to submit receptor-ligand complex structures prior to the release of the crystallographic coordinates of the CXCR4 in complex with IT1t and CVX15, as well as the D3R-eticlopride complex.101 This competition demonstrated the value of a solved CXCR4 structure, which represented challenges for modellers using other previously solved GPCR structures as a template for CXCR4 homology models. For example, the crystal structure of the CXCR4 highlighted the importance of the proline-induced kink in TM2, the so-called TXP motif also present in other chemokine receptors,156 which was poorly modelled by competitors.101 Furthermore, the definition of the binding site was also poorly predicted for the CXCR4-IT1t structure (49% prediction of binding pocket definition) relative to, for example, the D3R-eticlopride structure (81% pocket definition). This was attributed to distant homology models, larger binding pocket, multiple polar residues critically involved in ligand binding, and lack of mutagenesis data for IT1t.101 These features highlight the significance of the crystal structure of the CXCR4 for drug design. Since that time, a number of papers have demonstrated the impact of the CXCR4 structure on docking-based virtual screening for druglike compounds.157 Indeed, a structure-based ligand discovery project targeting the proteinprotein interface of the CXCR4 demonstrated the value of the solved structure, which culminated in the discovery of four novel antagonists that were dissimilar to previously known scaffolds, and displayed specific CXCR4 activity.158 Further highlighting the importance of the solved CXCR4 structure, the same virtual ligand screen using homology models based on β1 and β2 adrenergic receptors, adenosine A2A receptor and rhodopsin, yielded only one antagonist which resembled known ligands and lacked specificity.

Human Chemokine Receptor 5 (CCR5) Adding to their advances in chemokine receptor structural biology, the group of Wu and Stevens published the CCR5 structure in 2013.159 The CCR5 was first characterized as a receptor for three endogenous chemokine agonists: macrophage inflammatory protein–1α (MIP-1α), MIP-1β and RANTES,160 and later monocyte chemotactic protein (MCP)-2.161 Both CCR5 162 and CXCR4 are required for HIV-1 entry,163 which requires the sequential interaction of the viral exterior envelope glycoprotein gp120, with the CD4 glycoprotein, and a chemokine receptor on the cell surface.164 However, the CCR5 serves as a principal coreceptor for non-syncytium-inducing, macrophage-tropic strains of HIV-1, which is believed to be the key pathogenic strain in vivo.163 Studies have shown that a 32-bp deletion in the CCR5 coding region generates a nonfunctional receptor (CCR5∆35) that does not support membrane fusion or infection by macrophage- and dual-tropic HIV-1 strains, and CCR5∆35 individuals are normal, but resistant to HIV-1 infection.165 Furthermore, studies have shown that most non-syncytiuminducing clinical isolates use only CCR5 for infection.166 The CCR5 has thus been seen as an attractive drug target for inhibition of HIV-1 replication.167 The natural ligands for the CCR5,168 and their derivatives,169 have been shown to block CCR5-dependent HIV-1 infection, and the first non-peptidic CCR5 antagonist, TAK-779 (Figure 18), was discovered in 1999 to be potent and selective for anti-HIV-1 activity.170 This

resulted in numerous efforts to generate non-peptidic, drug-like, orally available CCR5 antagonists,171,172 and culminated in the discovery of the marketed antiretroviral, CCR5 antagonist, Maraviroc (UK-427,857, Figure 18) from Pfizer’s high-throughput screening program, for the treatment of HIV infection.173 Maraviroc and TAK-779, have since been discovered to be inverse agonists of CCR5,174 that act through allosteric mechanisms to inhibit the actions of endogenous chemokines.

Figure 18. Chemical structure of the first non-peptidic CCR5 inverse agonist TAK779, and the marketed CCR5 inverse agonist Mariviroc. The 2.7 Å structure of CCR5 was solved in complex with maraviroc, and shows similar architecture with other solved class A GPCR structures.159 Compared to CXCR4-IT1t, the binding site for maraviroc in CCR5 is deeper and occupies a larger area of the pocket (see Figure 19 (C)). Other notable differences between CCR5 and CXCR4 binding pockets were pointed out by Tan et al. including a shift of the extracellular end of CCR5 TM7 away from the central axis by ~3 Å resulting in a shift of CCR5’s N-terminal, and the lack of a salt bridge in EL2 in CCR5, resulting in a 6 Å shift at the β-hairpin tip of EL2 of CXCR4 toward the ligand binding pocket. The entrance to the CXCR4 binding pocket is consequently partially covered by its N-terminus and EL2, whereas the CCR5 ligand binding pocket is more open159 (see Figure 19. (C)) A hydrophobic interaction between maraviroc and Trp248 6.48 was shown to prevent activation-related motion of the receptor, which is also manifested by the close packing of helices of the 7TM bundle at the intracellular side of the receptor, preventing G protein binding.

Figure 19. Side-view (left) and extracellular view (right) of the ligand-binding pockets of (A, B) CXCR4 (grey) in complex with (A) the cyclic peptide CVX15 (orange); and (B) the small antagonist IT1t (green, 3ODU); and (C) CCR5 (light blue) in complex with Maraviroc (dark blue). Compared to CXCR4, the binding pocket of CCR5 is more open to the extracellular surface. Opioid receptors including the nociceptin receptor

The opioid system controls pain as well as reward and addictive behaviours. Opioids exert their pharmacological actions through activation of the three opioid receptors, µ (MOP), δ (DOP) and κ (KOP), which show high sequence identity of approximately 75%.175 Opioid agonists are widely used in the treatment of pain, but these drugs are known to have a broad range of severe side effects including respiratory depression, reduced gastric motility, sedation and nausea.176 Additionally, chronic opioid use induces tolerance and dependence. At least some of these undesired effects could be possibly controlled by the design of ligands with tailor-made subtype selectivity, intrinsic activity and biased signalling profiles. To better accomplish this, the structures of the three opioid receptors were published in 2012 and could be highly useful for the design of better and safer analgesics. Manglik et al. solved the structure of the T4L modified Mus musculus µ-opioid receptor bound to the irreversible antagonist β-funaltrexamine (Figure 20) at 2.8 Å resolution.177 Interestingly, the receptor crystallized as intimately associated pairs, where the contact interface is mainly formed by TM5 and TM6. This type of parallel dimer had also been observed in the CXCR4 structure.

Figure 20. Chemical structures of the irreversible ligand β-funaltrexamine, κ receptor selective antagonist JDTic, highly selective delta opioid receptor antagonist naltrindole, and the peptide mimetic NOP antagonist compound-24. The irreversible ligand β-funaltrexamine is a morphine derivative with a fumaric amide side chain that forms a covalent bond with the amino group of Lys233 5.39 via a Michael addition. Interestingly, the MOR structure revealed that the ligand binding pocket is closely related to the carazolol binding pocket in the β2AR, more so than the binding pocket of the closer related CXCR4 binding pocket. Nine direct interactions with amino acid residues could be observed, whereas the respective positions are conserved within the opioid receptor family. The δ-selective ligand naltrindole does not fit into the MOR binding pocket due to a clash

with Trp3187.35 (see Figure 21, red arrow), whereas the leucine found in the equivalent position in DOR does not interfere. Naltrindole (Figure 20) served as the ligand for the crystal structure of the δ opioid receptor (Figure 21), which was published by Garnier et al. just a couple of weeks later in 2012.178 Again, the ligand binds deeply in the binding pocket, which – in contrast to the binding sites in typical aminergic GPCRs – is widely opened towards the extracellular side. This is caused by a β-strand fold in EL2 that is conserved among the opioid receptor subtypes. The observed key interactions of naltrindole within the DOR binding pocket are very similar to those seen for β-funaltrexamine in MOR. Based on the comparison of the opioid receptor structures, the authors conclude that the well-conserved lower portion of the binding pocket forms contacts with the core morphinane moiety, and results in changes to ligand efficacy. The upper region, including EL3, serves as an important subtype selectivity determinant due to lower sequence identity between the receptor subtypes in the upper region (see Figure 21). This finding mirrors the well described ‘message–address’ concept of opioid pharmacology.179

Figure 21. Overlay of ligand binding sites of DOR (green) and MOR (blue) highlighted with the DOR-Naltrindole (green carbons) complex. The high sequence identity between receptor subtypes in the deep end of the binding pocket is responsible for driving efficacy (‘message’), whereas the extracellular end of the pocket is responsible for selectivity (‘address’) due to lower sequence identity, indicated by the labelled residues for DOR (green) and MOR (blue). The red arrow indicates the likely clash of naltrindole with W7.35 of MOR, whereas the smaller L7.35 of DOR allows naltrindole binding.

Two months after the Mus musculus T4L-DOR structure was published, Fenalti et al. reported on the crystallization of the human δ receptor. Amino-terminal modification with a b 562RIL (BRIL) fusion protein led to a receptor-naltrindole complex at 1.8 Å resolution.180 This structure reveals the presence of a sodium ion in the allosteric binding site, confirming the previously described allosteric modulation of opioid receptors by sodium ions.181 The metal

ion is packed in a network of polar interactions and, thus, stabilizes the conformation of a reduced agonist affinity state. The authors could demonstrate that an interference of the polar interactions by site-directed mutagenesis leads to an increase of β-arrestin-mediated signalling, and that mutations of the Asp and Asn key residues into Ala transforms the antagonist naltrindole into a β-arrestin-biased agonist. Simultaneously to the MOR and DOR subtypes, Wu et al. reported on the structure of the T4 lysozyme modified κ receptor in complex with the selective antagonist JDTic.182 The crystal packing showed a parallel dimer with an interface formed through contacts among TM1, TM2 and helix 8. In similarity to MOR and DOR extracellular-facing domains, the κ subtype features a binding pocket that is wider than in the reported aminergic GPCRs, but at the same time more contracted and deeper when compared to the rather closely related CXCR4 receptor. The tetrahydroisoquinoline-3-carboxamide JDTic shows a tight fit at the bottom of the binding site, whereas both protonated amines form salt bridges with Asp1383.32. The authors hypothesize that this strong interaction fixes an observed V-shape conformation of the ligand. The outstanding subtype selectivity of JDTic183 can be explained by interactions with four residues (Val1082.53, Val1182.63, Ile2946.55 and Tyr3127.35) that are not present in the µ and δ subtypes. Additionally, the authors performed docking studies employing morphinane based KOR ligands and a salvinorin derived ultrapotent agonist RP-64 to further elucidate the possible binding modes of structurally diverse KOR ligands. In the 1990s the nociceptin/orphanin FQ (N/OFQ) peptide receptor (NOP) was discovered as the fourth member of the opioid receptor.175 The NOP system plays an important role in numerous central and peripheral functions including pain, anxiety, food intake, memory, locomotion, cardiovascular control, immune responses.184 The heptadecapeptide nociceptin, also termed orphanin FQ, is a neuropeptide featuring interesting structural homologies to the classical opioid peptide, particularly dynorphin A. Despite these similarities, nociceptin does not show distinct affinity for the opioid receptor subtypes µ, δ and κ.185 The Stevens group published the NOP crystal structure nearly simultaneously to the structures of the classical opioid receptors.186 The receptor was stabilized by BRIL modification,187 and was crystallized in complex with the peptide mimetic C-24.188 Interestingly, the NOP structure features a pronounced shift of the extracellular end of TM5 when compared to MOR and KOR. This leads to both a gap between TM4 and TM5 and an expansion of the orthosteric binding pocket. This rather large binding cavity reflects its ability to bind large endogenous peptides. The structure reveals some substantial differences in the geometry of the binding pocket and, thus, contributes to the molecular understanding of ligand selectivity of NOP versus the classical opioid receptor subtypes.

Protease-Activated Receptor (PAR) 1 Thrombin is a serine protease that regulates numerous cellular responses, including platelet aggregation, endothelial cell activation, haemostasis and inflammation, and plays a major role in myocardial infarction and other pathological processes.189 PAR1, a prototypical GPCR, is irreversibly activated by thrombin-induced cleavage of the N terminal extracellular domain at a specific site, resulting in the N terminus serving as a tethered ligand, binding

intramolecularly, and stimulating activation of G12/13, Gq and Gi proteins. Due to the irreversible binding of the N terminus, PAR signalling must be actively terminated, or prevented. Therefore, a PAR1 antagonist is expected to produce potent antiplatelet effects, and because thrombin-mediated fibrin generation is unaffected, a PAR1 antagonist would be expected to result in less bleeding liability than conventional antithrombotic agents.190 Early PAR1 antagonists (Figure 22) were designed as peptidomimetics of the tethered ligand sequence, and in 1999, the first non-peptidic PAR1 antagonists were reported with high affinity and promising antiplatelet activity based on the pyrroloquinazoline scaffold.191 Research that was later released by Schering-Plough in 2005, demonstrated the discovery of a novel scaffold for PAR1 antagonists based on the natural product himbacine, from a highthroughput screening program to give SCH205831.192 Subsequent optimization of the series resulted in the discovery of SCH530348 (vorapaxar), as a potent, orally available antiplatelet agent,193 which is a highly specific and functionally irreversible inhibitor of PAR1.194 In 2014, vorapaxar was approved by the FDA to reduce the risk of heart attacks and stroke in high risk patients, making it the first in a new class of PAR1 antagonists.195

Figure 22. Chemical structure of the first non-peptidic PAR1 antagonist, 4e, and PAR1 irreversible inhibitors based on the himbacine scaffold by Schering-Plough SCH205831 and Vorapaxar. The slow off rate, high affinity, PAR1 selectivity and therapeutic relevance made vorapaxar an ideal candidate for the determination of the PAR1 structure in complex with a small molecule antagonist.194 In the solved structure, vorapaxar was found to bind in an “unusual location”,194 close to the extracellular surface of PAR1, unlike most other ligands for class A GPCRs which generally penetrate much deeper into the transmembrane core. Interestingly, the solved crystal structure revealed little information regarding ligand entry to the binding pocket, because unlike other class A GPCRs which have a solvent exposed extracellular opening to the binding pocket, none of the three PAR1 openings was deemed large enough to accommodate passage of the ligand.194 Thus the authors speculated that vorapaxar may access the binding pocket through the lipid bilayer between TM6 and TM7, in a similar mode proposed for the binding of retinal to rhodopsin,11 and sphingosine-1-phosphate (S1P) to the S1P1 receptor.128 Another interesting observation was the identification of an allosteric sodium binding site, known to also exist for several family A GPCRs, including the µ- and δopioid receptors, A2AR, and D2R.

To date, few examples exist of studies involving utilization of the solved PAR1 complex for drug discovery purposes. A study from Merck Research Laboratories demonstrated the development of several C7-spirocyclic analogues of vorapaxar, which utilize ring closure of the C7-carbamate side chain of vorapaxar, a known site for metabolism, to give oxazolidinone derivatives.196 To determine ligand-receptor interactions for the further understanding of SAR, models of the compounds in complex with PAR1 were generated using the solved PAR1-vorapaxar complex, and utilized to determine appropriate aryl substituents for improved ligand-receptor interactions.

P2Y12 Receptor (P2Y12R) The P2Y purinergic GPCR family are stimulated by nucleosides such as adenosine diphosphate (ADP), and are present in almost all human tissues where they exert various physiological functions based on their coupling to G proteins.197 The platelet receptor P2Y12 receptor (also known as P2T receptor) plays a major role in platelet aggregation, and unlike the related P2Y1 receptor which is Gq coupled, the P2Y12 receptor (along with P2Y13) is coupled to Gi, inhibits adenylyl cylclase and regulates ion channels. Drugs that selectively target the P2Y12 receptor have been widely used as antiplatelet agents, even before cloning of the receptor,198 including the irreversible prodrug clopidogrel. Studies have also demonstrated the potentiating role for the P2Y12R in dense granule secretion, fibrinogen-receptor activation, and thrombus formation, thus indicating the P2Y12R as a central mediator of the hemostatic response.199 Drug discovery at the P2Y12R originated with the discovery of antagonists that act as mimics of ATP (Figure 23), the natural antagonist, with modified polyphosphate side chains to prevent breakdown to the agonist ADP, and adenine moiety substitutions to enhance affinity and selectivity.200 One such P2Y12R antagonist, AR-C69931MX (Cangrelor) has progressed to phase III clinical trials as an ultra-short acting, intravenously administered, antithrombotic agent for use in patients with coronary artery disease undergoing percutaneous coronary intervention.200 Other nucleoside-based P2Y12R antagonists have been developed, including AZD6140 (Ticagrelor, Figure 23) by AstraZeneca which is the first reversibly binding, orally active P2Y12R antagonist, and was approved by the FDA in 2011 as a platelet aggregation inhibitor.201 Other efforts for drug discovery at P2Y12R have focussed on non-nucleotide derived competitive antagonists, and some quinolone derivatives, such as PSB0702 (Figure 23), have been described as highly potent, competitive P2Y12R antagonists with selectivity over other P2 receptor subtypes.202 More recently, AstraZeneca revealed the discovery of a piperazinylpyridine based scaffold from high throughput screening of their compound collection for P2Y12R antagonists.203 Further development resulted in the discovery of AZD1283 (Figure 23) by replacement of the sulfonylurea linker with an acyl sulfonamide linker. AZD1283 demonstrated antithrombotic effects in vivo, as well as dose-dependent increases in blood flow and inhibition of ADP-induced platelet aggregation, and was recently progressed to human clinical trials.

Figure 23. Chemical structures of P2Y12R antagonists, including the ATP mimic, Cangrelor in clinical trials; Ticagrelor, another adenosine-based P2Y12R antagonist; the non-nucleotide derivative P2Y12R antagonist PSB0702l; and AZD1283, the P2Y12R antagonist cocrystallized with the receptor, and a new clinical candidate. The discovery of synthetic agonists (Figure 24) targeting the P2Y12R has proven significantly more challenging than the discovery of antagonists. The ligand properties of ATP binding to the human P2Y12 has been controversial, with some reporting ATP as an antagonist,201 but others as an agonist.204 Further confusing evidence has also suggested that diadenosine tetraphosphate (Ap4A) is a weak agonist of P2Y12R, but simultaneously a partial antagonist.205 Another example is AR-C66096, a non-cleavable triphosphate mimetic, which was developed by AstraZeneca as a potential antithrombotic drug, and demonstrated P2Y12R inverse agonism,206 and partial agonism.207 Nonetheless, recent findings have demonstrated that introduction of a 2ʹ-methylthio group to ATP and ADP (2MeSATD and 2MeSADP, respectively), has been shown to increase ligand potency (greater increase for 2MeSADP) at P2Y12R, and result in full agonism.208

Figure 24. Chemical structures of P2Y12 agonists, ADP, an endogenous ligand; and, the P2Y12 agonists 2MeSADP and 2MeSATP which were both co-crystallized with the P2Y12 receptor Two crystal structures of the P2Y12R were recently solved and published; one in complex with AZD1283209 (Figure 26), and one in complex with the agonists 2MeSADP and 2MeSADT.207 These structures represent the second solved structures of the δ group of class A GPCRs, but demonstrate important features that set it apart from another representative, PAR1. For example, TM5 in most class A GPCR structures is bulged and bent at a highly conserved Pro 5.50, however P2Y12R has Asn2015.50 resulting in a structure that lacks the corresponding helical bend. This is similar to the structure of the S1P1 receptor, which also lacks the corresponding helical bend at TM5. The binding pocket of AZD1283 in the inactive structure was found to be in a very distinct location and shape compared to other solved GPCR crystal structures, but shares a similar region as vorapaxar in the PAR1, which also belongs to the δ group of class A GPCRs. However, the residues comprising the P2Y12R binding pocket are likely to be more flexible and have a lesser role in ligand-receptor interactions compared to the PAR1 binding pocket. The extracellular ends of TM4, TM6 and TM7 in P2Y12R are shifted outwards compared with PAR1, resulting in a more open, deeper ligand binding pocket within the transmembrane domain. A barrier separating two binding pockets was also identified in the P2Y12R/AZD1283 crystal structure (Figure 26), with only one occupied by AZD1283. Pocket 2 is believed to be involved in the covalent binding of active metabolites of drugs such as Prasugrel and Clopidogrel (Figure 25) to the conserved Cys97 3.25 (red residues, Figure 26). Due to poor electron density of Cys97 3.25 in combination with mutational and size exclusion assays, the authors suggested the possibility of a labile or dynamic disulfide bond, a unique feature for GPCRs.

Figure 25. Chemical structures of the prodrugs, Clopidogrel and Prasugrel. The active metabolites are suggested to covalently bind to the Cys973.25 at the P2Y12 receptor.

Figure 26. Extracellular view of the P2Y12 receptor (grey) in complex with AZD1283 (orange) in ribbon depiction (left) and surface depiction (right). A bridge created by Tyr3.33 and Lys7.35 (green) separates the AZD1283 pocket from a secondary pocket (purple). In pocket 2, Cys3.25 is suggested as a covalent binding site for antiplatelet drugs such as clopidogrel.

The agonist-bound structure of the human P2Y12R was published in the same edition of Nature as the inactive structure,207 and demonstrated remarkable differences in receptor conformation, including large conformational changes in TM6 and TM7, where the extracellular part of TM6 shifts over 10 Å, and TM7 over 5 Å towards the axis of the 7TM helical bundle, compared to the antagonist-bound structure. Importantly, these shifts allow the formation of an extensive ionic and polar interaction network with the phosphate groups of 2MeSADP.207 Other key differences include the movement of the N terminus towards the axis of the helical bundle; and the position of the ELs, where a disulpfide bond stabilizing the conformation of EL2 was found to result in unwinding of the helical bulge, and a subsequent relocation of residues in the region compared with the AZD1283 complex. The result of these movements is a tighter receptor complex than seen for other class A GPCRs, where the agonist 2MeSADP is almost completely enclosed within the receptor. Due to the large scale movements of the helical bundle, the binding pocket of 2MeSADP was found to only partially overlap with the binding site of AZD1283 in the inactive structure, and although both ligands bind to the same pocket, their orientations are completely different. Numerous hydrophilic and positively charged residues dominate ligand-receptor interactions to the diphosphate group of 2MeSADP, and cysteine residues forming a conserved disulfide bond are also directly involved in the binding of 2MeSADP and subsequent agonist activity at the receptor. In terms of agonist drug discovery at the receptor, the negatively charged phosphate groups are of key importance for closing the ‘lid’ formed by the highly cationic ELs and the N terminus, which is unlike the antagonist-bound structure. However, agonist discovery at the receptor has proven highly challenging, and controversial, and confusion remains regarding the pharmacology of many P2Y12R ligands. The agonist-bound crystal structure is destined to significantly enhance the abilities for drug discovery of both agonists and antagonists at this therapeutically relevant receptor.

Human GPR40 (Free Fatty-Acid Receptor 1 (FFAR1)) The FFAR1, previously known as human GPR40, binds free fatty acids and is primarily localized in pancreatic β cells, and has thus been widely recognized as a potential target for the treatment of type 2 diabetes due to the involvement in regulation of insulin secretion,210 regulation of metabolic processes and glucose homeostasis.211 Therefore, it is thought that the development of FFAR1 agonists may mimic the effect of free fatty acids to enhance glucosestimulated insulin secretion, and thus serve as anti-diabetic drugs. One of the first structure-activity relationships for ligands targeting the FFAR1 was conducted by GlaxoSmithKline in 2006 based on a hit from high throughput screening of their chemical collection.212 From optimization of the hit compound they discovered highly potent FFAR1 agonists, including GW9508 (Figure 27). Further efforts by others to develop FFAR1 agonists resulted in the discovery of TUG-469213 which was demonstrated to be a promising candidate for further drug development based on in vitro and in vivo characterization.214 At a similar time, high throughput screening by Takeda Pharmaceuticals identified molecules containing phenylpropanoic acids, and optimization by ring fusing resulted in the discovery of the orally available TAK-875 as a highly potent and selective FFAR1 ago-allosteric modulator with insulintropic efficacy and excellent PK profile.215 TAK-875 (fasiglifam, Figure 26) was further examined in clinical trials for the treatment of type 2 diabetes, but development was terminated at the end of 2013 due to concerns about liver safety.216

Figure 27. Chemical structures of FFAR1 ligands, including one of the first discovered agonists GW9508; the optimized ligand, TUG-469, that progressed to the clinic as an FFAR1 agonist; and the highly selective FFAR1 agonist, TAK-875, which was co-crystallized with FFAR1, and progressed to Phase III clinical trials before development was recently terminated. Despite the failure of TAK-875, the efficacy of targeting the FFAR1 with agonists has proven the therapeutic value of the receptor in improving glycaemic control and low hypoglycaemic risk in diabetic patients.217 Takeda Pharmaceuticals have very recently published the high resolution structure of the FFAR1 bound to TAK-875,218 which revealed that TAK-875

uniquely binds to a non-canonical binding site created between TM3-TM5 and EL2, and is closer to the extracellular surface than most other GPCR ligands (see Figure 28). Despite the low sequence identity, FFAR1 was found to show structural similarities with peptide-binding GPCRs, such as the δ-opioid receptor, however the FFAR1-EL2 functions as a cap-like structure which covers the binding pocket from the central extracellular region. Entry of the ligand to the binding pocket was suggested to most probably occur through the lipid bilayer (Figure 28), a method previously proposed for the retinal loading of opsin, ligand entry of AM-841 and sn-2-arachidonylglycerol into the cannabinoid 2 receptor,219 and lipid entry for S1P1. Some of the key ligand-receptor interactions involve a complex charge network involving the carboxylate moiety of TAK-875 with Arg183 5.39 and Arg2587.35, which have been previously demonstrated by mutational studies to be important for agonist recognition and activation.220 However, the crystal structure also demonstrated the importance of direct ligand interactions with two tyrosine residues (Try913.37 and Tyr2406.51) for the stabilization of the carboxylate moiety within the pocket, indicating the presence of a complex charge network.

Figure 28. (A) Overall structure of the FFAR1 within a membrane, showing the cocrystallized TAK-875. (B) Zoomed cross-section of the binding mode of TAK-875, with the carboxylate moiety participating in a complex charge network, while the remaining ligand is located between TM3 and TM4.

The FFAR1 has also been suggested to contain at least one other binding site, and inspection of the crystal structure in complex with TAK-875 demonstrated the possible existence of two extra binding pockets in the receptor; one adjacent to TAK-875, allowing a ligand to pass between TM4 and TM5.218 The other possible binding site is located in proximity to the orthosteric sites observed in class A GPCRs between TM1 and TM7, and in close proximity to the binding site of LY2119620 in the M2 receptor. Given the novelty of the FFAR1 crystal structure, and the therapeutic importance of the receptor in treatment of an increasing global health problem, the FFAR1/TAK-875 is likely to be used in the design of future drugs targeting this, and related GPCRs.

Class B GPCRs The secretin family of GPCRs includes 15 receptors, which are classified into five subgroups. These class B GPCRs are promising drug targets in the field of diabetes, osteoporosis, migraine, depression and anxiety.221 The endogenous ligands for these receptors are, without exception, large peptide hormones like secretin or glucagon. Therefore, the development of potent non-peptidic molecules (Figure 29) has proven challenging, predominantly due to a poor understanding of non-peptide binding-sites of class B GPCRs, as no structural information about the transmembrane domain was available. In 2013, the structures of the first two members of the secretin family were published and have the potential of assisting the rational design of potent ligands. Heptares Therapeutics Ltd reported on the crystal structure of the transmembrane domain of the human corticotropin-releasing factor receptor type 1 (CRF1R) in complex with the nonpeptidic antagonist CP-376395 in 3.0 Å resolution, identifying a distinct binding location within the 7-TM region.222 CRF is an important regulating peptide in stress-response and is involved in appetite control, cardiovascular regulation, glucose metabolism, immune function and behavior. The CRF1R structure was obtained by a thermostabilization comprising twelve point mutations in combination with the fusion of T4L into IL2. Additionally, the large amino-terminal domain as well as the carboxy terminus including helix 8 were removed.

Figure 29. Chemical structure of CP376395, a CRF1R non-peptidic antagonist co-crystallized with the CRF1R. When compared to class A GPCRs, the CRF1R structure showed a distinct V-shape resulting in a large cavity, which presumably acts as the peptide-binding site. The most significant conformational differences can be observed in the orientation of TM7, which adopts a sharply kinked structure around a pivot near Gly356 7.50. The resulting structure demonstrates up to 10 Å difference in the extracellular position of the tip of TM7. While this significant kink was possibly caused by the truncation of helix 8, this modification did not affect ligand binding properties.

The structure also revealed an unanticipated antagonist-binding site. The ligand CP-376395 was found to fill a hydrophobic pocket in the cytoplasmic half of the receptor. This position is 13–23 Å away from the binding position of typical class A ligands. This surprising finding can now be utilized for the structure based design of novel small-molecule ligands. In the same year, the Stevens group published the structure of the glucagon receptor (GCGR), which is a promising drug target for type 2 diabetes.223 The receptor construct included the thermally stabilized E. coli apocytochrome b562RIL (BRIL), and lacked the extracellular domain and C terminus. Again, the cavity of the peptide-binding pocket is wider and deeper compared with class A GPCRs. The crystallization experiments were performed in the presence of the antagonist NNC0640, but data analysis did not show convincing electron density for the ligand. Additionally, site-directed mutagenesis experiments in combination with docking studies based on this crystal structure were performed to elucidate the binding mode of glucagon to the receptor. With these first two crystal structures of class B GPCRs, a very promising starting point for the structure-based design of novel ligands has been generated, and it can be expected that more receptor structures of the secretin family will be solved in the near future.

Metabotropic Glutamate Receptor 5 (mGluR5) Glutamate is a principle excitatory neurotransmitter in the CNS, and has shown to function in long term potentiation/depression, learning and memory.224 Glutamate receptors are membrane proteins that are categorized into ionotropic glutamate receptors (iGluRs), which internally contain ligand-gated ion channels; and, metabotropic glutamate receptors (mGluRs), which belong to the Class C GPCR family, and are expressed in neuronal and glial cells. The eight subtypes of mGluRs are classified into three subgroups according to sequence similarity, agonist selectivity and effector system differences; subgroup I (mGluR1 and -5), subgroup II (mGluR2 and -3) and subgroup III (mGluR4, -6, -7 and -8).225 Both receptors of mGluR subgroup I (mGluR1 and -5) are predominantly coupled to Gq/11, and activate phospholipase Cβ which hydrolyses phosphoinositides into inositol 1,4,5-trisphophate and diacylglycerol, thereby inducing intracellular calcium mobilization and protein kinase C activation. The overall mGluR structure has been classified to contain three distinct regions: the extracellular domain, the transmembrane domain, and the cytoplasmic domain. The extracellular domain of mGluRs, which is unique amongst GPCRs, consists of the ‘venus fly trap’ (VFT) which binds glutamate, and a cysteine-rich domain (CRD) linked to the transmembrane domain. The VFT of mGluR1 was solved by crystallography in 2000,226 and demonstrated the dimeric ligand-binding region of mGluR1 to adopt multiple conformations, stabilized by glutamate binding. In 2007, the crystal structure of the entire extracellular region of mGluR3 (group II) complexed with various agonists, and the VFT domain of mGluR7 (group III) were solved.227 Despite the rich information available for the extracellular domains, the transmembrane domain of mGluRs remained elusive until recently. This domain is of particular importance in

the mGluR5 subtype, where negative allosteric modulation of mGluR5 was postulated to demonstrate utility in the treatment of anxiety disorders by dampening activity in glutamatergic circuits, and positive allosteric modulators in the treatment of schizophrenia and disorders of cognitive function.119 To test these hypotheses required highly selective mGluR5 agents, and targeting of an allosteric GPCR binding site has demonstrated increased subtype selectivity due to lower subtype homology in the allosteric binding site.5 The first mGluR5 subtype selective, non-competitive antagonists (Figure 30) were (E)-2methyl-6-styryl-pyridine (SIB-1893) and the alkyne derivative, 2-methyl-6-(phenylethynyl)pyridine (MPEP),228 which were reported to be potent, and systemically active. However, despite the extended utility of MPEP as an in vitro and in vivo tool for studying the mGluR5, it was found not to be a suitable drug candidate due to blockade of N-methyl-D-aspartate (NMDA) receptors,229 off-target activity, and poor water solubility. Thus, numerous studies were initiated to discover non-competitive mGluR5 subtype selective antagonists. In 2004, the mGluR theory was proposed as an explanation for the characteristic phenotype observed in Fragile X Syndrome,230 and the subtype selective mGluR5 negative allosteric modulator mavoglurant (AFQ056)231,232 entered phase III clinical trials for the treatment of Fragile X Syndrome, but development has been recently discontinued. Basimglurant (RG7090) is another compound currently in Phase II clinical trials as an mGluR5 allosteric modulator for Fragile X Syndrome and treatment-resistant depression, and shares similar structural features such as the aromatic extended alkyne moiety.

Figure 30. Chemical structures of the first non-competitive mGluR5 antagonists, SIB1893 and MPEP; the discontinued mGluR5 negative allosteric modulator Mavoglurant that was cocrystallized with the mGlur5 transmembrane domain, and Basimglurant, which is currently in Phase II clinical trials as a mGluR5 allosteric modulator. In 2014, the crystal structure of the mGluR5 transmembrane domain in complex with mavoglurant was solved by Heptares Therapeutics,233 representing the first class C GPCR transmembrane domain to be structurally determined (Figure 31). The allosteric binding site of mavoglurant was found to be deeper than all class A receptor transmembrane binding sites determined, but not as deep as the CP-376395 binding site at the CRF1R from class B. The most striking difference between mGluR5, rhodopsin (from class A) and CRF1R was found to be the position of TM5, which was positioned further inwards by approximately 6 Å. This was found to contribute to the narrow entrance for the allosteric cavity determined by the crystal structure (see Figure 31 (A)). The structure of the allosteric binding pocket also explains the necessity for an alkyne linker present in many mGluR5 allosteric modulators,234 whereby the alkyne linker traverses a narrow channel leading the 3-methyl-substituted aromatic group towards a pocket at the cytoplasmic end of the pocket containing a water molecule in an extensive hydrogen bonding network (Figure 31 (B)). Interestingly, modification of the 3-methyl substituent of mavoglurant from methoxy to chloro to fluoro switches the ligand from a negative allosteric modulator, to a neutral binder, to a positive

allosteric modulator, respectively.235 Thus, perturbation of the hydrogen bonding network at the base of the allosteric binding pocket may be a potential activation switch.

Figure 31. Overall structure of the mGluR5 (grey TM bundle, blue EL2, and green Nterminus) in complex with Mavoglurant (orange). (A) Extracellular view of the receptor demonstrates the narrow entry to the binding pocket, which is occluded by the configuration of the helical bundle and EL2. (B) The binding pocket of Mavoglurant within the TM bundle, and the zoomed in section demonstrating key ligand-receptor interactions, as well as a water molecule at the bottom of the pocket stabilizing interhelical interactions. The solved mGluR5 crystal structure presents numerous opportunities for drug discovery given the therapeutic relevance of the receptor. Furthermore, the receptor structure increases the understanding of the mechanism of action of metabotropic receptors, and will provide a template for homology modelling of related receptors.

Metabotropic Glutamate Receptor 1 (mGluR1) At the same time as the mGluR5 was solved and published, the mGluR1 transmembrane domain was determined by the Stevens group in complex with the negative allosteric modulator 4-fluoro-N-(4-(6-(isopropylamino)pyrimidin-4-yl)thiazol-2-yl)-Nmethylbenzamide (FITM, Figure 32).236 Both mGluR1 and mGluR5 are sub-classified into subgroup I, however whereas mGluR5 is highly concentrated in forebrain and limbic structures, mGluR1 is widely distributed in the CNS, and modulates synaptic transmission, neuronal excitability and brain plasticity. The mGluR1 has demonstrated involvement in epilepsy, neurodegeneration, pain, and anxiety.237 However, a lack of selective mGluR1 compounds had previously prevented a better understanding of the precise role of mGluR1 in human physiology. The first identified negative allosteric modulator of the mGluR family was 7hydroxyiminocyclopropan[b]chromen-1a-carboxylic acid ethyl ester (CPCCOEt), which was found to be a highly potent antagonist of mGluR1 without affecting glutamate binding.238 Since this finding, allosteric modulators of the mGluR family have been well explored, and are arguably the most well studied GPCR family with respect to allosteric modulation.239 Numerous mGluR1 selective negative allosteric modulators (Figure 32) have been reported with systemic bioavailability and varying potency, including LY367385 with 8 µM IC50,240 YM298198 with low nanomolar affinity,241 and JNJ16259685 which selectively inhibits

mGluR1 at subnanomolar concentration.242 Additionally, numerous mGluR1 positive allosteric modulators have been identified by Hoffmann-La Roche,243 including Ro01-6128, Ro67-4853, and Ro67-7476, which were found to potentiate synaptically evoked mGluR1 responses in rat brain slices. The mGluR1 has also been shown to contain at least one other allosteric binding site, whereby N-(4-chloro-2-((1,3-dioxoisoindolin-2-yl)methyl)phenyl)-2hydroxybenzamide (CPPHA) has been demonstrated to allosterically activate the mGluR1 without binding at the MPEP binding site.244 N

O F

S N

OH

O HO

H N

N N

OH

O

N

O

O

NH2

O

FITM

LY367385

CPCCOEt

H2N

O N N

S

YM298198

O

N

O

N O JNJ16259685

Figure 32. Chemical structures of selective mGluR1 negative allosteric modulators, including FITM which was co-crystallized with the mGluR1, CPCCOEt the first discovered mGluR1 negative allosteric modulator, and other compounds pursued as mGluR1 negative allosteric modulators.

The solved crystal structure complex of the mGluR1-FITM transmembrane domain revealed important information about the FITM binding site, which was found to reside in a pocket defined by residues on EL2 and TM2, TM3, TM5, TM6 and TM7.236 Analogous to the mGluR5, the binding cavity of FITM was found to be occluded by EL2, leaving a narrow entrance for the ligand on the extracellular side. Unsurprisingly given the hydrophobicity and lack of heteroatoms on FITM, most of the ligand-receptor interactions were hydrophobic, with the exception of a polar contact to the pyrimidine-amine group with Thr8157.38. Based on docking studies of analogous compounds to FITM, with and without the equivalent pyrimidine-amine group, the authors demonstrated that the superior potency could be attributed to this polar interaction and the excellent fit of the ligand within the narrow binding pocket.236 The mGluR1 and mGluR5 transmembrane domain crystal structures together presented a missing link in terms of GPCR structural biology, and this will allow significant advances in drug design targeting the allosteric site of class C GPCRs.

Smoothened receptors

The class F family of GPCRs includes the Frizzled and the Smoothened receptors (SMO). The latter has been identified as a key player in the hedgehog (Hh) signalling pathway, which regulates embryonic development in animals. Dysregulation within this pathway results in severe malformations of the embryo and can cause cancer in adults.245 In 2013, Wang et al. published the structure of the human SMO.246 Crystals of a SMO-BRIL construct in complex with the antagonist LY2940680 (Figure 33) were obtained using a lipidic mesophase method. Although the sequence identity of the SMO receptor and class A GPCRs is less than 10%, the structure revealed the canonical 7TM bundle of this class F receptor. On the other hand, a series of distinct features that are specific to the SMO receptor could be identified. These structural differences are mainly located at TM5-TM7, where a particularly large number of glycine residues was observed. Interestingly, the receptor-ligand interactions are, for the most part, formed by the extracellular loops 2 and 3, which typically points outward from the ligand binding pocket in peptide receptors of the class A GPCR family. In a very recent study, Wang et al. reported on three more crystal structures of the human Smoothened receptor in complex with the antagonists SANT1 and Anta XV, and the agonist SAG1.5 (Figure 33).247 As had already been observed for the LY2940680 bound structure, the binding cavity was narrow and elongated, allowing different poses for the ligands. The antagonist SANT1 binds significantly deeper in this pocket compared to the other ligands. The agonist SAG1.5 induces a pronounced rearrangement of the residues Arg5.43, Asp6.54 and Glu 7.38 within the binding pocket. These conformational changes could possibly contribute to the activation of the Smoothened receptor. A fifth SMO structure was published by Weierstall et al. in 2014.248 Using a novel lipidic cubic phase technology the authors obtained a SMO structure in complex with the naturally occurring teratogen cyclopamine. The ligand binds close to the entrance region of the long and narrow cavity. While the hydroxy group of the steroid scaffold points into the binding pocket, the piperidine moiety is located outside the cavity. This finding supports previous SAR studies demonstrating that structural modifications are tolerated at the secondary amine whereas the 3,β-hydroxy group is sensitive to the attachment of substituents.249

Figure 33. Chemical structures of the SMO antagonists LY2940680, SANT1 and Anta XV, the naturally occurring teratogen cyclopamine, and the agonist SAG1.5.

Summary of Ligand-Receptor Interactions in Solved GPCR Crystal Structures

Figure 34. Overview of GPCR ligand binding pockets within the helical bundle, visualized for a representative subset of available crystal structures. This figure demonstrates the plasticity

and diversity of GPCRs and their ligand binding sites. The PDB-IDs of the structures used for this graphic are indicated.

Table 1. A summary of all currently available crystal structures of GPCRs that have been solved. Ligand function

Year

Res. [Å]

PDB-ID

Cyanopindolol Dobutamine Carmoterol (R)-Isoprenaline Salbutamol Dobutamine Carazolol Cyanopindolol Iodocyanopindolol Cyanopindolol Bucindolol Carvedilol ° (Cyanopindolol)

ANT AG AG AG AG AG ANT ANT ANT ANT ANT ANT ANT

2008 2011 2011 2011 2011 2011 2011 2011 2011 2011 2012 2012 2013

2.70 2.50 2.60 2.85 3.05 2.60 3.00 3.25 3.65 3.15 3.20 2.30 3.50

2VT4 2Y00 2Y02 2Y03 2Y04 2Y01 2YCW 2YCX 2YCZ 2YCY 4AMI 4AMJ 4GPO

turkey

4-Methyl-2-(piperazin-1yl)quinoline

ANT

2013

2.70

3ZPQ

β1AR

turkey

4-(Piperazin-1-yl)-1Hindole

ANT

2013

2.80

3ZPR

β1AR β2AR β2AR β2AR β2AR β2AR β2AR β2AR β2AR β2AR β2AR* β2AR* β2AR β2AR*

turkey human human human human human human human human human human human human human

ANT ANT ANT ANT ANT ANT ANT ANT ANT AG AG AG ANT AG

2014 2007 2007 2007 2008 2010 2010 2010 2010 2011 2011 2011 2012 2013

2.10 2.40 3.40 3.40 2.80 3.40 2.84 2.84 3.16 3.50 3.50 3.20 3.99 2.79

4BVN 2RH1 2R4R 2R4S 3D4S 3KJ6 3NY8 3NY9 3NYA 3PDS 3P0G 3SN6 4GBR 4LDE

β2AR*

human

AG

2013

3.10

4LDL

β2AR* β2AR* A2AAR A2AAR A2AAR A2AAR A2AAR A2AAR

human human human human human human human human

AG AG ANT AG AG AG ANT ANT

2013 2014 2008 2011 2011 2011 2011 2011

3.20 3.30 2.60 2.71 3.00 2.60 3.30 3.31

4LDO 4QKX 3EML 3QAK 2YDO 2YDV 3PWH 3REY

Class Receptor

Aα α

Species

Ligand

β1AR β1AR β1AR β1AR β1AR β1AR β1AR β1AR β1AR β1AR β1AR β1AR β1AR

turkey turkey turkey turkey turkey turkey turkey turkey turkey turkey turkey turkey turkey

β1AR

Cyanopindolol Carazolol (Carazolol)° (Carazolol)° Timolol (Carazolol)° ICI 118,551 Compound 2 Alprenolol FAUC50 BI167,107 BI167,107 Carazolol BI167,107 Hydroxybenzyl isoproterenol Adrenaline FAUC37 ZM241385 UK-432097 Adenosine NECA ZM241385 XAC

A, β

A, γ

A, δ

B C

F

A2AAR A2AAR A2AAR A2AAR A2AAR A2AAR D3R H1R M2R M2R* M2R* M3R S1P1 S1P1 5-HT1B 5-HT1B 5-HT2B 5-HT2B

human human human human human human human human human human human rat human human human human human human

Caffeine ZM241385 ZM241385 4e 4g ZM241385 Eticlopride Doxepin (E, Z) QNB Iperoxo Iperoxo, LY2119620 Tiotropium ML056 ML056 Dihydroergotamine Ergotamine Ergotamine Ergotamine

NTSR1 NTSR1 NTSR1 NTSR1 NTSR1 CXCR4 CXCR4 CXCR4 CXCR4 CXCR4 CCR5 δ-OR δ-OR κ-OR µ-OR N/OFQ-OR PAR1 P2Y12 P2Y12 P2Y12 FFAR1 CRF1R GCGR mGlu1 mGlu5 SMO SMO SMO SMO SMO

rat rat rat rat rat human human human human human human mouse human human mouse human human human human human human human human human human human human human human human

NT(8-13) NT(8-13) NT(8-13) NT(8-13) NT(8-13) CVX15 IT1t IT1t IT1t IT1t Maraviroc Naltrindole Naltrindole JDTic β-FNA Compound 24 Vorapaxar AZD1283 2MeSADP 2MeSATP TAK-875 CP-376395 (NNC0640)° FITM Mavoglurant LY2940680 SANT1 Anta XV SAG1.5 Cyclopamine

ANT ANT ANT ANT ANT ANT ANT ANT ANT AG AG, AM ANT ANT ANT AG AG AG AG

2011 2012 2012 2012 2012 2012 2010 2011 2012 2013 2013 2012 2012 2012 2013 2013 2013 2013

3.60 2.70 3.10 3.34 3.27 1.80 2.89 3.10 3.00 3.50 3.70 3.40 2.80 3.35 2.80 2.70 2.70 2.80

3RFM 3VG9 3VGA 3UZC 3UZA 4EIY 3PBL 3RZE 3UON 4MQS 4MQT 4DAJ 3V2Y 3V2W 4IAQ 4IAR 4IB4 4NC3

AG AG AG AG AG ANT ANT ANT ANT ANT ANT ANT ANT ANT ANT ANT ANT ANT AG AG AG ANT ANT AM AM ANT ANT ANT AG ANT

2012 2014 2014 2014 2014 2010 2010 2010 2010 2010 2013 2012 2013 2012 2012 2012 2012 2014 2014 2014 2014 2013 2013 2014 2014 2013 2014 2014 2014 2014

2.80 3.00 2.75 3.10 3.57 2.90 2.50 3.20 3.10 3.10 2.71 3.40 1.80 2.90 2.80 3.01 2.20 2.63 2.50 3.10 2.30 2.98 3.30 2.80 2.60 2.45 2.80 2.61 2.60 3.20

4GRV 3ZEV 4BUO 4BV0 4BWB 3OE0 3ODU 3OE6 3OE8 3OE9 4MBS 4EJ4 4N6H 4DJH 4DKL 4EA3 3VW7 4NTJ 4PXZ 4PY0 4PHU 4K5Y 4L6R 4OR2 4OO9 4JKV 4N4W 4QIM 4QIN 4O9R

Table 1 Footnote. The asterisk (*) refers to active-state conformations stabilized either by a G

protein or by a G protein-mimicking nanobody. The circlet (°) refers to structures, in which no electron-density was observed for the co-crystallized ligand. In the special case of 4GPO (PDB-ID), cyanopindolol was used for purification of the receptor, but was removed before starting the crystallization experiments. Abbreviations: ANT = antagonist/inverse agonist, AG = agonist/partial agonist, AM = allosteric modulator Outlook Advances in GPCR structural biology have, thus far, significantly contributed to the understanding of drug action, and the design of novel compounds at therapeutically relevant receptors. As the number of GPCR crystal structures rapidly increases, so do the opportunities for medicinal chemists in the design of compounds in the future. Already we have seen numerous advances in GPCR drug design through the use of structure based drug design, virtual ligand screening, and construction of homology models. Given the highly dynamic nature of GPCRs, and their binding sites, a more detailed understanding of GPCR binding sites has proven valuable in aiding more refined requirements for drug design. This review has focussed on the latest advances in GPCR structural biology, and in particular on the ligand-receptor interactions that have been solved. We have also focused on future drug design opportunities, and summarized the results of each crystal structure solved to date. We aimed to explore the question: what do we learn from the advances in GPCR structural information, and how can it be used for the design of novel compounds in the future? As more structures are solved in the future, we expect that this will translate to better design of GPCR ligands, with better potency and selectivity profiles than the past. We expect that this will be a major contributing factor to better methods of treating the many disease states associated with GPCR drug discovery.

Author Biographies Dr. Jeremy Shonberg Jeremy Shonberg was born in Melbourne, Australia in 1987. After a Bachelor of Medicinal Chemistry (Honours) in 2008, Jeremy completed his PhD in 2013 from the Department of Medicinal Chemistry at Monash University (Melbourne, Australia) under the supervision of Dr Ben Capuano, Prof. Peter J. Scammells and Dr J. Rob Lane. This work focussed on the design, synthesis and pharmacological analysis of novel ligands targeting the dopamine D2 receptor, including the design of bivalent, bitopic, allosteric and biased ligands. After completing his PhD, Jeremy stayed as a postdoctoral researcher for 9 months at Monash University, before moving to Germany to join the group of Prof. Peter Gmeiner as a postdoctoral researcher at Friedrich-Alexander University, Erlangen-Nuremberg. He is currently working on medicinal chemistry targeting adrenergic receptors and other GPCRs, and has recently been granted an Alexander von Humboldt Postdoctoral Research Fellowship to continue his research in Germany.

Ralf C. Kling Ralf C. Kling was born in 1985 in Arad, Romania and grew up in Nürnberg, Germany. He studied Pharmacy at the University of Erlangen-Nürnberg (2006-2009). After a practical year of internship in a pharmacy and in industry, he joined the group of Peter Gmeiner at the Department of Chemistry and Pharmacy in Erlangen in 2011. He works on the investigation of GPCR signaling using computational methods.

Prof. Peter Gmeiner Peter Gmeiner was born in 1959 in Vohenstrauß. He received his Ph.D. in 1986 from the University of Munich. From 1987 to 1988 he was a postdoc at the University of California in Berkeley, USA. Peter Gmeiner subsequently returned to Munich as a research associate. Upon receiving his habilitation in 1992, he was appointed as a Professor of Pharmaceutical Chemistry at the University of Bonn. Since October 1996, he has held the chair of Full Professor of Pharmaceutical/Medicinal Chemistry at the University of Erlangen-Nürnberg. Peter Gmeiner’s research spans the design, organic synthesis, and pharmacological investigation of bioactive molecules when class I G-protein coupled receptors (GPCRs) are addressed as allosteric target proteins. Peter Gmeiner is the spokesman of the DFG research training group “Medicinal Chemistry of Selective GPCR Ligands”.

Dr. Stefan Löber Stefan Löber was born in 1971 in Nürnberg, Germany. He studied Food Chemistry at the University of Erlangen-Nürnberg. After a research internship at the University Kaiserslautern,

he joined the group of Peter Gmeiner at the Department of Chemistry and Pharmacy in Erlangen. His Ph.D. received from the University of Erlangen-Nürnberg in 2000 concerned the synthesis and biological investigation of selective dopamine receptor ligands. He is currently holding a position as a senior scientist and works on the design of novel orthosteric, allosteric, bitopic and bivalent ligands of class A GPCRs. Stefan Löber is the coordinator of the DFG research training group “Medicinal Chemistry of Selective GPCR Ligands”.

Acknowledgements This work was supported by the DFG research training group (GRK1910) “Medicinal Chemistry of Selective GPCR Ligands”.

References

1. Fredriksson, R.; Lagerström, M. C.; Lundin, L.-G.; Schiöth, H. B. Mol. Pharmacol. 2003, 63, 1256. 2. Insel, P. A.; Tang, C.-M.; Hahntow, I.; Michel, M. C. Biochim. Biophys. Acta 2007, 1768, 994. 3. Jacoby, E.; Bouhelal, R.; Gerspacher, M.; Seuwen, K. ChemMedChem 2006, 1, 760. 4. Foord, S. M.; Bonner, T. I.; Neubig, R. R.; Rosser, E. M.; Pin, J.-P.; Davenport, A. P.; Spedding, M.; Harmar, A. J. Pharmacol. Rev. 2005, 57, 279. 5. Wootten, D.; Christopoulos, A.; Sexton, P. M. Nat. Rev. Drug. Discov. 2013, 12, 630. 6. Lane, J. R.; Sexton, P. M.; Christopoulos, A. Trends Pharmacol. Sci. 2013, 34, 59. 7. Hiller, C.; Kühhorn, J.; Gmeiner, P. J. Med. Chem. 2013. 8. Shonberg, J.; Scammells, P. J.; Capuano, B. ChemMedChem 2011, 6, 963. 9. Shonberg, J.; Lopez, L.; Scammells, P. J.; Christopoulos, A.; Capuano, B.; Lane, J. R. Med. Res. Rev. 2014, 34, 1286. 10. Schertler, G. F. X.; Villa, C.; Henderson, R. Nature 1993, 362, 770. 11. Palczewski, K.; Kumasaka, T.; Hori, T.; Behnke, C. A.; Motoshima, H.; Fox, B. A.; Trong, I. L.; Teller, D. C.; Okada, T.; Stenkamp, R. E.; Yamamoto, M.; Miyano, M. Science 2000, 289, 739. 12. Hofmann, K. P.; Scheerer, P.; Hildebrand, P. W.; Choe, H.-W.; Park, J. H.; Heck, M.; Ernst, O. P. Trends Biochem. Sci 2009, 34, 540. 13. Manglik, A.; Kobilka, B. Curr. Opin. Cell Biol. 2014, 27, 136. 14. Deupi, X. Biochimica et Biophysica Acta - Bioenergetics 2014, 1837, 674.

15. Kobilka, B. Angew. Chem. Int. Ed. Engl. 2013, 52, 6380. 16. Bylund, D. B.; Eikenberg, D. C.; Hieble, J. P.; Langer, S. Z.; Lefkowitz, R. J.; Minneman, K. P.; Molinoff, P. B.; Ruffolo, R. R., Jr.; Trendelenburg, U. Pharmacol. Rev. 1994, 46, 121. 17. Baker, J. G. PLoS One 2010, 5, e15487. 18. Prichard, B. N.; Cruickshank, J. M.; Graham, B. R. Blood Press. 2001, 10, 366. 19. Waldeck, B. Eur. J. Pharmacol. 2002, 445, 1. 20. Baker, J. G. Br. J. Pharmacol. 2005, 144, 317. 21. Baker, J. G. Br. J. Pharmacol. 2010, 160, 1048. 22. Lefkowitz, R. J. Angew. Chem. Int. Ed. Engl. 2013, 52, 6366. 23. Bertheleme, N.; Chae, P. S.; Singh, S.; Mossakowska, D.; Hann, M. M.; Smith, K. J.; Hubbard, J. A.; Dowell, S. J.; Byrne, B. Biochim. Biophys. Acta 2013, 1828, 2583. 24. Rosenbaum, D. M.; Cherezov, V.; Hanson, M. A.; Rasmussen, S. G. F.; Thian, F. S.; Kobilka, T. S.; Choi, H.-J.; Yao, X.-J.; Weis, W. I.; Stevens, R. C.; Kobilka, B. K. Science 2007, 318, 1266. 25. Cherezov, V.; Rosenbaum, D. M.; Hanson, M. A.; Rasmussen, S. G.; Thian, F. S.; Kobilka, T. S.; Choi, H. J.; Kuhn, P.; Weis, W. I.; Kobilka, B. K.; Stevens, R. C. Science 2007, 318, 1258. 26. Rasmussen, S. G. F.; Choi, H.-J.; Rosenbaum, D. M.; Kobilka, T. S.; Thian, F. S.; Edwards, P. C.; Burghammer, M.; Ratnala, V. R. P.; Sanishvili, R.; Fischetti, R. F.; Schertler, G. F. X.; Weis, W. I.; Kobilka, B. K. Nature 2007, 450, 383. 27. Hanson, M. A.; Cherezov, V.; Griffith, M. T.; Roth, C. B.; Jaakola, V.-P.; Chien, E. Y. T.; Velasquez, J.; Kuhn, P.; Stevens, R. C. Structure 2008, 16, 897. 28. Wacker, D.; Fenalti, G.; Brown, M. A.; Katritch, V.; Abagyan, R.; Cherezov, V.; Stevens, R. C. J. Am. Chem. Soc. 2010, 132, 11443. 29. Rosenbaum, D. M.; Zhang, C.; Lyons, J. A.; Holl, R.; Aragao, D.; Arlow, D. H.; Rasmussen, S. G. F.; Choi, H.-J.; DeVree, B. T.; Sunahara, R. K.; Chae, P. S.; Gellman, S. H.; Dror, R. O.; Shaw, D. E.; Weis, W. I.; Caffrey, M.; Gmeiner, P.; Kobilka, B. K. Nature 2011, 469, 236. 30. Rasmussen, S. G. F.; Choi, H.-J.; Fung, J. J.; Pardon, E.; Casarosa, P.; Chae, P. S.; DeVree, B. T.; Rosenbaum, D. M.; Thian, F. S.; Kobilka, T. S.; Schnapp, A.; Konetzki, I.; Sunahara, R. K.; Gellman, S. H.; Pautsch, A.; Steyaert, J.; Weis, W. I.; Kobilka, B. K. Nature 2011, 469, 175. 31. Rasmussen, S. G. F.; DeVree, B. T.; Zou, Y.; Kruse, A. C.; Chung, K. Y.; Kobilka, T. S.; Thian, F. S.; Chae, P. S.; Pardon, E.; Calinski, D.; Mathiesen, J. M.; Shah, S. T. A.; Lyons, J. A.; Caffrey, M.; Gellman, S. H.; Steyaert, J.; Skiniotis, G.; Weis, W. I.; Sunahara, R. K.; Kobilka, B. K. Nature 2011, 477, 549. 32. Bokoch, M. P.; Zou, Y.; Rasmussen, S. G.; Liu, C. W.; Nygaard, R.; Rosenbaum, D. M.; Fung, J. J.; Choi, H. J.; Thian, F. S.; Kobilka, T. S.; Puglisi, J. D.; Weis, W. I.; Pardo, L.; Prosser, R. S.; Mueller, L.; Kobilka, B. K. Nature 2010, 463, 108. 33. Zou, Y.; Weis, W. I.; Kobilka, B. K. PLoS ONE 2012, 7, e46039. 34. Ring, A. M.; Manglik, A.; Kruse, A. C.; Enos, M. D.; Weis, W. I.; Garcia, K. C.; Kobilka, B. K. Nature 2013, 502, 575. 35. Weichert, D.; Kruse, A. C.; Manglik, A.; Hiller, C.; Zhang, C.; Hubner, H.; Kobilka, B. K.; Gmeiner, P. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 10744. 36. Warne, T.; Serrano-Vega, M. J.; Baker, J. G.; Moukhametzianov, R.; Edwards, P. C.; Henderson, R.; Leslie, A. G. W.; Tate, C. G.; Schertler, G. F. X. Nature 2008, 454, 486. 37. Warne, T.; Moukhametzianov, R.; Baker, J. G.; Nehme, R.; Edwards, P. C.; Leslie, A. G. W.; Schertler, G. F. X.; Tate, C. G. Nature 2011, 469, 241. 38. Moukhametzianov, R.; Warne, T.; Edwards, P. C.; Serrano-Vega, M. J.; Leslie, A. G.; Tate, C. G.; Schertler, G. F. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 8228.

39. Warne, T.; Edwards, P. C.; Leslie, A. G.; Tate, C. G. Structure 2012, 20, 841. 40. Huang, J.; Chen, S.; Zhang, J. J.; Huang, X.-Y. Nat. Struct. Mol. Biol. 2013, 20, 419. 41. Christopher, J. A.; Brown, J.; Dore, A. S.; Errey, J. C.; Koglin, M.; Marshall, F. H.; Myszka, D. G.; Rich, R. L.; Tate, C. G.; Tehan, B.; Warne, T.; Congreve, M. J. Med. Chem. 2013, 56, 3446. 42. Miller-Gallacher, J. L.; Nehme, R.; Warne, T.; Edwards, P. C.; Schertler, G. F.; Leslie, A. G.; Tate, C. G. PLoS One 2014, 9, e92727. 43. Serrano-Vega, M. J.; Magnani, F.; Shibata, Y.; Tate, C. G. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 877. 44. Katritch, V.; Fenalti, G.; Abola, E. E.; Roth, B. L.; Cherezov, V.; Stevens, R. C. Trends Biochem. Sci 2014, 39, 233. 45. Ballesteros, J. A.; Weinstein, H. In Methods in Neurosciences; Stuart, C. S., Ed.; Academic Press, 1995; Vol. 25, pp. 366. 46. Ballesteros, J. A.; Jensen, A. D.; Liapakis, G.; Rasmussen, S. G.; Shi, L.; Gether, U.; Javitch, J. A. J. Biol. Chem. 2001, 276, 29171. 47. Dror, R. O.; Arlow, D. H.; Borhani, D. W.; Jensen, M. Ø.; Piana, S.; Shaw, D. E. Proc. Natl. Acad. Sci. USA 2009, 106, 4689. 48. Kolb, P.; Rosenbaum, D. M.; Irwin, J. J.; Fung, J. J.; Kobilka, B. K.; Shoichet, B. K. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 6843. 49. de Graaf, C.; Rognan, D. J. Med. Chem. 2008, 51, 4978. 50. Dror, R. O.; Arlow, D. H.; Maragakis, P.; Mildorf, T. J.; Pan, A. C.; Xu, H.; Borhani, D. W.; Shaw, D. E. Proc. Natl. Acad. Sci. USA 2011, 108, 18684. 51. Heyl, D. L.; Sefler, A. M.; He, J. X.; Sawyer, T. K.; Wustrow, D. J.; Akunne, H. C.; Davis, M. D.; Pugsley, T. A.; Heffner, T. G.; Corbin, A. E.; et al. Int. J. Pept. Protein Res. 1994, 44, 233. 52. Yao, X. J.; Velez Ruiz, G.; Whorton, M. R.; Rasmussen, S. G.; DeVree, B. T.; Deupi, X.; Sunahara, R. K.; Kobilka, B. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 9501. 53. Liapakis, G.; Ballesteros, J. A.; Papachristou, S.; Chan, W. C.; Chen, X.; Javitch, J. A. J. Biol. Chem. 2000, 275, 37779. 54. Liu, J. J.; Horst, R.; Katritch, V.; Stevens, R. C.; Wüthrich, K. Science 2012, 335, 1106. 55. Nygaard, R.; Zou, Y.; Dror, R. O.; Mildorf, T. J.; Arlow, D. H.; Manglik, A.; Pan, A. C.; Liu, C. W.; Fung, J. J.; Bokoch, M. P.; Thian, F. S.; Kobilka, T. S.; Shaw, D. E.; Mueller, L.; Prosser, R. S.; Kobilka, B. K. Cell 2013, 152, 532. 56. Ghanouni, P.; Gryczynski, Z.; Steenhuis, J. J.; Lee, T. W.; Farrens, D. L.; Lakowicz, J. R.; Kobilka, B. K. J. Biol. Chem. 2001, 276, 24433. 57. Costanzi, S.; Vilar, S. J. Comput. Chem. 2012, 33, 561. 58. Jacobson, K. A.; Costanzi, S. Mol. Pharmacol. 2012, 82, 361. 59. Kling, R. C.; Lanig, H.; Clark, T.; Gmeiner, P. PLoS ONE 2013, 8, e67244. 60. Rose, A. S.; Elgeti, M.; Zachariae, U.; Grubmuller, H.; Hofmann, K. P.; Scheerer, P.; Hildebrand, P. W. J. Am. Chem. Soc. 2014, 136, 11244. 61. Shi, L.; Liapakis, G.; Xu, R.; Guarnieri, F.; Ballesteros, J. A.; Javitch, J. A. J. Biol. Chem. 2002, 277, 40989. 62. Fredholm, B. B.; IJzerman, A. P.; Jacobson, K. A.; Klotz, K.-N.; Linden, J. Pharmacol. Rev. 2001, 53, 527. 63. Dunwiddie, T. V.; Masino, S. A. Annu. Rev. Neurosci. 2001, 24, 31. 64. Daly, J.; Butts-Lamb, P.; Padgett, W. Cell. Mol. Neurobiol. 1983, 3, 69. 65. Chen, J. F.; Xu, K.; Petzer, J. P.; Staal, R.; Xu, Y. H.; Beilstein, M.; Sonsalla, P. K.; Castagnoli, K.; Castagnoli Jr, N.; Schwarzschild, M. A. J. Neurosci. 2001, 21. 66. Premont, J.; Perez, M.; Bockaert, J. Mol. Pharmacol. 1977, 13, 662.

67. Cristalli, G.; Cacciari, B.; Dal Ben, D.; Lambertucci, C.; Moro, S.; Spalluto, G.; Volpini, R. ChemMedChem 2007, 2, 260. 68. Poucher, S. M.; Keddie, J. R.; Singh, P.; Stoggall, S. M.; Caulkett, P. W. R.; Jones, G.; Collis, M. G. Br. J. Pharmacol. 1995, 115, 1096. 69. Pinna, A. CNS Drugs 2014, 28, 455. 70. Jaakola, V. P.; Griffith, M. T.; Hanson, M. A.; Cherezov, V.; Chien, E. Y. T.; Lane, J. R.; Ijzerman, A. P.; Stevens, R. C. Science 2008, 322, 1211. 71. Michino, M.; Abola, E.; Brooks Iii, C. L.; Dixon, J. S.; Moult, J.; Stevens, R. C. Nat. Rev. Drug Discov. 2009, 8, 455. 72. Carlsson, J.; Yoo, L.; Gao, Z.-G.; Irwin, J. J.; Shoichet, B. K.; Jacobson, K. A. J. Med. Chem. 2010, 53, 3748. 73. Katritch, V.; Jaakola, V.-P.; Lane, J. R.; Lin, J.; Ijzerman, A. P.; Yeager, M.; Kufareva, I.; Stevens, R. C.; Abagyan, R. J. Med. Chem. 2010, 53, 1799. 74. Müller, C. E.; Jacobson, K. A. Biochim. Biophys. Acta 2011, 1808, 1290. 75. Jörg, M.; Shonberg, J.; Mak, F. S.; Miller, N. D.; Yuriev, E.; Scammells, P. J.; Capuano, B. Bioorg. Med. Chem. Lett. 2013, 23, 3427. 76. Congreve, M.; Andrews, S. P.; Doré, A. S.; Hollenstein, K.; Hurrell, E.; Langmead, C. J.; Mason, J. S.; Ng, I. W.; Tehan, B.; Zhukov, A.; Weir, M.; Marshall, F. H. J. Med. Chem. 2012, 55, 1898. 77. Andrews, S. P.; Brown, G. A.; Christopher, J. A. ChemMedChem 2014, 9, 256. 78. Sams, A. G.; Mikkelsen, G. K.; Larsen, M.; Langgård, M.; Howells, M. E.; Schrøder, T. J.; Brennum, L. T.; Torup, L.; Jørgensen, E. B.; Bundgaard, C.; Kreilgård, M.; BangAndersen, B. J. Med. Chem. 2011, 54, 751. 79. Tosh, D. K.; Phan, K.; Gao, Z.-G.; Gakh, A. A.; Xu, F.; Deflorian, F.; Abagyan, R.; Stevens, R. C.; Jacobson, K. A.; Katritch, V. J. Med. Chem. 2012, 55, 4297. 80. Areias, F. M.; Brea, J.; Gregori-Puigjané, E.; Zaki, M. E. A.; Carvalho, M. A.; Domínguez, E.; Gutiérrez-de-Terán, H.; Proença, M. F.; Loza, M. I.; Mestres, J. Biorg. Med. Chem. 2010, 18, 3043. 81. Xu, F.; Wu, H.; Katritch, V.; Han, G. W.; Jacobson, K. A.; Gao, Z.-G.; Cherezov, V.; Stevens, R. C. Science 2011, 332, 322. 82. Lebon, G.; Warne, T.; Edwards, P. C.; Bennett, K.; Langmead, C. J.; Leslie, A. G. W.; Tate, C. G. Nature 2011, 474, 521. 83. Ohta, A.; Sitkovsky, M. Nature 2001, 414, 916. 84. Chen, J.-F.; Eltzschig, H. K.; Fredholm, B. B. Nat. Rev. Drug. Discov. 2013, 12, 265. 85. Lane, J. R.; Klein Herenbrink, C.; van Westen, G. J. P.; Spoorendonk, J. A.; Hoffmann, C.; IJzerman, A. P. Mol. Pharmacol. 2012, 81, 475. 86. Mantell, S. J.; Stephenson, P. T. WO0194368 (A1) 2001. 87. Lebon, G.; Warne, T.; Tate, C. G. Curr. Opin. Struct. Biol. 2012, 22, 482. 88. Missale, C.; Russel Nash, S.; Robinson, S. W.; Jaber, M.; Caron, M. G. Physiol. Rev. 1998, 78, 189. 89. Beaulieu, J.-M.; Gainetdinov, R. R. Pharmacol. Rev. 2011, 63, 182. 90. Sokoloff, P.; Giros, B.; Martres, M.-P.; Bouthenet, M.-L.; Schwartz, J.-C. Nature 1990, 347, 146. 91. Löber, S.; Hübner, H.; Tschammer, N.; Gmeiner, P. Trends Pharmacol. Sci. 2011, 32, 148. 92. Lacroix, L. P.; Hows, M. E.; Shah, A. J.; Hagan, J. J.; Heidbreder, C. A. Neuropsychopharmacology 2003, 28, 839. 93. Heidbreder, C. A.; Newman, A. H. Ann. N.Y. Acad. Sci. 2010, 1187, 4. 94. Stemp, G.; Ashmeade, T.; Branch, C. L.; Hadley, M. S.; Hunter, A. J.; Johnson, C. N.; Nash, D. J.; Thewlis, K. M.; Vong, A. K. K.; Austin, N. E.; Jeffrey, P.; Avenell, K. Y.;

Boyfield, I.; Hagan, J. J.; Middlemiss, D. N.; Reavill, C.; Riley, G. J.; Routledge, C.; Wood, M. J. Med. Chem. 2000, 43, 1878. 95. Newman, A. H.; Grundt, P.; Cyriac, G.; Deschamps, J. R.; Taylor, M.; Kumar, R.; Ho, D.; Luedtke, R. R. J. Med. Chem. 2009, 52, 2559. 96. Micheli, F. ChemMedChem 2011, 6, 1152. 97. Banala, A. K.; Levy, B. A.; Khatri, S. S.; Furman, C. A.; Roof, R. A.; Mishra, Y.; Griffin, S. A.; Sibley, D. R.; Luedtke, R. R.; Newman, A. H. J. Med. Chem. 2011, 54, 3581. 98. Chien, E. Y. T.; Liu, W.; Zhao, Q.; Katritch, V.; Won Han, G.; Hanson, M. A.; Shi, L.; Newman, A. H.; Javitch, J. A.; Cherezov, V.; Stevens, R. C. Science 2010, 330, 1091. 99. Shonberg, J.; Herenbrink, C. K.; López, L.; Christopoulos, A.; Scammells, P. J.; Capuano, B.; Lane, J. R. J. Med. Chem. 2013, 56, 9199. 100. Lane, J. R.; Donthamsetti, P.; Shonberg, J.; Draper-Joyce, C. J.; Dentry, S.; Michino, M.; Shi, L.; López, L.; Scammells, P. J.; Capuano, B.; Sexton, P. M.; Javitch, J. A.; Christopoulos, A. Nat. Chem. Biol. 2014, 10, 745. 101. Kufareva, I.; Rueda, M.; Katritch, V.; Stevens, Raymond C.; Abagyan, R. Structure 2011, 19, 1108. 102. Carlsson, J.; Coleman, R. G.; Setola, V.; Irwin, J. J.; Fan, H.; Schlessinger, A.; Sali, A.; Roth, B. L.; Shoichet, B. K. Nat. Chem. Biol. 2011, 7, 769. 103. Lane, J. R.; Chubukov, P.; Liu, W.; Canals, M.; Cherezov, V.; Abagyan, R.; Stevens, R. C.; Katritch, V. Mol. Pharmacol. 2013, 84, 794. 104. Urban, J. D.; Vargas, G. A.; von Zastrow, M.; Mailman, R. B. Neuropsychopharmacology 2007, 32, 67. 105. Hiller, C.; Kling, R. C.; Heinemann, F. W.; Meyer, K.; Hübner, H.; Gmeiner, P. J. Med. Chem. 2013, 56, 5130. 106. Möller, D.; Kling, R. C.; Skultety, M.; Leuner, K.; Hübner, H.; Gmeiner, P. J. Med. Chem. 2014, 57, 4861. 107. Free, R. B.; Chun, L. S.; Moritz, A. E.; Miller, B. N.; Doyle, T. B.; Conroy, J. L.; Padron, A.; Meade, J. A.; Xiao, J.; Hu, X.; Dulcey, A. E.; Han, Y.; Duan, L.; Titus, S.; Bryant-Genevier, M.; Barnaeva, E.; Ferrer, M.; Javitch, J. A.; Beuming, T.; Shi, L.; Southall, N. T.; Marugan, J. J.; Sibley, D. R. Mol. Pharmacol. 2014, 86, 96. 108. Simons, F. E. R. New Engl. J. Med. 2004, 351, 2203. 109. Shimamura, T.; Shiroishi, M.; Weyand, S.; Tsujimoto, H.; Winter, G.; Katritch, V.; Abagyan, R.; Cherezov, V.; Liu, W.; Han, G. W.; Kobayashi, T.; Stevens, R. C.; Iwata, S. Nature 2011, 475, 65. 110. Cusack, B.; Nelson, A.; Richelson, E. Psychopharmacology 1994, 114, 559. 111. de Graaf, C.; Kooistra, A. J.; Vischer, H. F.; Katritch, V.; Kuijer, M.; Shiroishi, M.; Iwata, S.; Shimamura, T.; Stevens, R. C.; de Esch, I. J.; Leurs, R. J. Med. Chem. 2011, 54, 8195. 112. Cordova-Sintjago, T. C.; Fang, L.; Bruysters, M.; Leurs, R.; Booth, R. G. J. Chem. Pharm. Res. 2012, 4, 2937. 113. Lepailleur, A.; Freret, T.; Lemaitre, S.; Boulouard, M.; Dauphin, F.; Hinschberger, A.; Dulin, F.; Lesnard, A.; Bureau, R.; Rault, S. J. Chem. Inf. Model. 2014, 54, 1773. 114. Levoin, N.; Labeeuw, O.; Krief, S.; Calmels, T.; Poupardin-Olivier, O.; BerrebiBertrand, I.; Lecomte, J. M.; Schwartz, J. C.; Capet, M. Bioorg. Med. Chem. 2013, 21, 4526. 115. Feng, Z.; Hou, T.; Li, Y. J. Mol. Graph. Model. 2013, 39, 1. 116. Schultes, S.; Nijmeijer, S.; Engelhardt, H.; Kooistra, A. J.; Vischer, H. F.; de Esch, I. J. P.; Haaksma, E. E. J.; Leurs, R.; de Graaf, C. MedChemComm 2013, 4, 193. 117. Sirci, F.; Istyastono, E. P.; Vischer, H. F.; Kooistra, A. J.; Nijmeijer, S.; Kuijer, M.; Wijtmans, M.; Mannhold, R.; Leurs, R.; de Esch, I. J.; de Graaf, C. J. Chem. Inf. Model. 2012, 52, 3308.

118. Kruse, A. C.; Kobilka, B. K.; Gautam, D.; Sexton, P. M.; Christopoulos, A.; Wess, J. Nat. Rev. Drug. Discov. 2014, 13, 549. 119. Conn, P. J.; Christopoulos, A.; Lindsley, C. W. Nat. Rev. Drug Discov. 2009, 8, 41. 120. Haga, K.; Kruse, A. C.; Asada, H.; Yurugi-Kobayashi, T.; Shiroishi, M.; Zhang, C.; Weis, W. I.; Okada, T.; Kobilka, B. K.; Haga, T.; Kobayashi, T. Nature 2012, 482, 547. 121. Kruse, A. C.; Hu, J.; Pan, A. C.; Arlow, D. H.; Rosenbaum, D. M.; Rosemond, E.; Green, H. F.; Liu, T.; Chae, P. S.; Dror, R. O.; Shaw, D. E.; Weis, W. I.; Wess, J.; Kobilka, B. K. Nature 2012, 482, 552. 122. Tautermann, C. S.; Kiechle, T.; Seeliger, D.; Diehl, S.; Wex, E.; Banholzer, R.; Gantner, F.; Pieper, M. P.; Casarosa, P. J. Med. Chem. 2013, 56, 8746. 123. Kruse, A. C.; Hu, J.; Kobilka, B. K.; Wess, J. Curr. Opin. Pharmacol. 2014, 16, 24. 124. Kruse, A. C.; Ring, A. M.; Manglik, A.; Hu, J.; Hu, K.; Eitel, K.; Hubner, H.; Pardon, E.; Valant, C.; Sexton, P. M.; Christopoulos, A.; Felder, C. C.; Gmeiner, P.; Steyaert, J.; Weis, W. I.; Garcia, K. C.; Wess, J.; Kobilka, B. K. Nature 2013, 504, 101. 125. Kloeckner, J.; Schmitz, J.; Holzgrabe, U. Tetrahedron Lett. 2010, 51, 3470. 126. Blaho, V. A.; Hla, T. J. Lipid Res. 2014, 55, 1596. 127. Brinkmann, V.; Billich, A.; Baumruker, T.; Heining, P.; Schmouder, R.; Francis, G.; Aradhye, S.; Burtin, P. Nat. Rev. Drug. Discov. 2010, 9, 883. 128. Hanson, M. A.; Roth, C. B.; Jo, E.; Griffith, M. T.; Scott, F. L.; Reinhart, G.; Desale, H.; Clemons, B.; Cahalan, S. M.; Schuerer, S. C.; Sanna, M. G.; Han, G. W.; Kuhn, P.; Rosen, H.; Stevens, R. C. Science 2012, 335, 851. 129. Sanna, M. G.; Wang, S.-K.; Gonzalez-Cabrera, P. J.; Don, A.; Marsolais, D.; Matheu, M. P.; Wei, S. H.; Parker, I.; Jo, E.; Cheng, W.-C.; Cahalan, M. D.; Wong, C.-H.; Rosen, H. Nat. Chem. Biol. 2006, 2, 434. 130. Davis, M. D.; Clemens, J. J.; Macdonald, T. L.; Lynch, K. R. J. Biol. Chem. 2005, 280, 9833. 131. Berger, M.; Gray, J. A.; Roth, B. L. Annu. Rev. Med. 2009, 60, 355. 132. Roth, B. L. New Engl. J. Med. 2007, 356, 6. 133. Wang, C.; Jiang, Y.; Ma, J.; Wu, H.; Wacker, D.; Katritch, V.; Han, G. W.; Liu, W.; Huang, X.-P.; Vardy, E.; McCorvy, J. D.; Gao, X.; Zhou, X. E.; Melcher, K.; Zhang, C.; Bai, F.; Yang, H.; Yang, L.; Jiang, H.; Roth, B. L.; Cherezov, V.; Stevens, R. C.; Xu, H. E. Science 2013, 340, 610. 134. Wacker, D.; Wang, C.; Katritch, V.; Han, G. W.; Huang, X.-P.; Vardy, E.; McCorvy, J. D.; Jiang, Y.; Chu, M.; Siu, F. Y.; Liu, W.; Xu, H. E.; Cherezov, V.; Roth, B. L.; Stevens, R. C. Science 2013, 340, 615. 135. Huang, X. P.; Setola, V.; Yadav, P. N.; Allen, J. A.; Rogan, S. C.; Hanson, B. J.; Revankar, C.; Robers, M.; Doucette, C.; Roth, B. L. Mol. Pharmacol. 2009, 76, 710. 136. Liu, W.; Wacker, D.; Gati, C.; Han, G. W.; James, D.; Wang, D.; Nelson, G.; Weierstall, U.; Katritch, V.; Barty, A.; Zatsepin, N. A.; Li, D.; Messerschmidt, M.; Boutet, S.; Williams, G. J.; Koglin, J. E.; Seibert, M. M.; Wang, C.; Shah, S. T.; Basu, S.; Fromme, R.; Kupitz, C.; Rendek, K. N.; Grotjohann, I.; Fromme, P.; Kirian, R. A.; Beyerlein, K. R.; White, T. A.; Chapman, H. N.; Caffrey, M.; Spence, J. C.; Stevens, R. C.; Cherezov, V. Science 2013, 342, 1521. 137. Kitabgi, P. Current opinion in drug discovery & development 2002, 5, 764. 138. Dubuc, I.; Sarret, P.; Labbe-Jullie, C.; Botto, J. M.; Honore, E.; Bourdel, E.; Martinez, J.; Costentin, J.; Vincent, J. P.; Kitabgi, P.; Mazella, J. J. Neurosci. 1999, 19, 503. 139. White, J. F.; Noinaj, N.; Shibata, Y.; Love, J.; Kloss, B.; Xu, F.; GvozdenovicJeremic, J.; Shah, P.; Shiloach, J.; Tate, C. G.; Grisshammer, R. Nature 2012, 490, 508. 140. Luca, S.; White, J. F.; Sohal, A. K.; Filippov, D. V.; van Boom, J. H.; Grisshammer, R.; Baldus, M. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 10706.

141. Egloff, P.; Hillenbrand, M.; Klenk, C.; Batyuk, A.; Heine, P.; Balada, S.; Schlinkmann, K. M.; Scott, D. J.; Schutz, M.; Pluckthun, A. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, E655. 142. Wu, B.; Chien, E. Y. T.; Mol, C. D.; Fenalti, G.; Liu, W.; Katritch, V.; Abagyan, R.; Brooun, A.; Wells, P.; Bi, F. C.; Hamel, D. J.; Kuhn, P.; Handel, T. M.; Cherezov, V.; Stevens, R. C. Science 2010, 330, 1066. 143. Murphy, P. M. Cytokine Growth Factor Rev. 1996, 7, 47. 144. Zou, Y.-R.; Kottmann, A. H.; Kuroda, M.; Taniuchi, I.; Littman, D. R. Nature 1998, 393, 595. 145. Muller, A.; Homey, B.; Soto, H.; Ge, N.; Catron, D.; Buchanan, M. E.; McClanahan, T.; Murphy, E.; Yuan, W.; Wagner, S. N.; Barrera, J. L.; Mohar, A.; Verastegui, E.; Zlotnik, A. Nature 2001, 410, 50. 146. Berger, E. A.; Murphy, P. M.; Farber, J. M. Annu. Rev. Immunol. 1999, 17, 657. 147. Feng, Y.; Broder, C. C.; Kennedy, P. E.; Berger, E. A. Science 1996, 272, 872. 148. Charo, I. F.; Ransohoff, R. M. New Engl. J. Med. 2006, 354, 610. 149. Burger, J. A.; Peled, A. Leukemia 2008, 23, 43. 150. Broxmeyer, H. E.; Orschell, C. M.; Clapp, D. W.; Hangoc, G.; Cooper, S.; Plett, P. A.; Liles, W. C.; Li, X.; Graham-Evans, B.; Campbell, T. B.; Calandra, G.; Bridger, G.; Dale, D. C.; Srour, E. F. J. Exp. Med. 2005, 201, 1307. 151. Dessolin, J.; Galea, P.; Vlieghe, P.; Chermann, J.-C.; Kraus, J.-L. J. Med. Chem. 1999, 42, 229. 152. Tamamura, H.; Hori, A.; Kanzaki, N.; Hiramatsu, K.; Mizumoto, M.; Nakashima, H.; Yamamoto, N.; Otaka, A.; Fujii, N. FEBS Lett. 2003, 550, 79. 153. Endres, M. J.; Clapham, P. R.; Marsh, M.; Ahuja, M.; Turner, J. D.; McKnight, A.; Thomas, J. F.; Stoebenau-Haggarty, B.; Choe, S.; Vance, P. J.; Wells, T. N. C.; Power, C. A.; Sutterwala, S. S.; Doms, R. W.; Landau, N. R.; Hoxie, J. A. Cell 1996, 87, 745. 154. Kim, S.; Lee, C.; Midura, B.; Yeung, C.; Mendoza, A.; Hong, S.; Ren, L.; Wong, D.; Korz, W.; Merzouk, A.; Salari, H.; Zhang, H.; Hwang, S.; Khanna, C.; Helman, L. Clin. Exp. Metastasis 2008, 25, 201. 155. Thoma, G.; Streiff, M. B.; Kovarik, J.; Glickman, F.; Wagner, T.; Beerli, C.; Zerwes, H.-G. J. Med. Chem. 2008, 51, 7915. 156. Govaerts, C.; Blanpain, C.; Deupi, X.; Ballet, S.; Ballesteros, J. A.; Wodak, S. J.; Vassart, G.; Pardo, L.; Parmentier, M. J. Biol. Chem. 2001, 276, 13217. 157. Planesas, J. M.; Pérez-Nueno, V. I.; Borrell, J. I.; Teixidó, J. J. Mol. Graphics Modell. 2012, 38, 123. 158. Mysinger, M. M.; Weiss, D. R.; Ziarek, J. J.; Gravel, S.; Doak, A. K.; Karpiak, J.; Heveker, N.; Shoichet, B. K.; Volkman, B. F. Proc. Natl. Acad. Sci. USA 2012, 109, 5517. 159. Tan, Q.; Zhu, Y.; Li, J.; Chen, Z.; Han, G. W.; Kufareva, I.; Li, T.; Ma, L.; Fenalti, G.; Li, J.; Zhang, W.; Xie, X.; Yang, H.; Jiang, H.; Cherezov, V.; Liu, H.; Stevens, R. C.; Zhao, Q.; Wu, B. Science 2013, 341, 1387. 160. Samson, M.; Labbe, O.; Mollereau, C.; Vassart, G.; Parmentier, M. Biochemistry 1996, 35, 3362. 161. Ruffing, N.; Sullivan, N.; Sharmeen, L.; Sodroski, J.; Wu, L. Cell. Immunol. 1998, 189, 160. 162. Dragic, T.; Litwin, V.; Allaway, G. P.; Martin, S. R.; Huang, Y.; Nagashima, K. A.; Cayanan, C.; Maddon, P. J.; Koup, R. A.; Moore, J. P.; Paxton, W. A. Nature 1996, 381, 667. 163. Deng, H.; Liu, R.; Ellmeier, W.; Choe, S.; Unutmaz, D.; Burkhart, M.; Marzio, P. D.; Marmon, S.; Sutton, R. E.; Hill, C. M.; Davis, C. B.; Peiper, S. C.; Schall, T. J.; Littman, D. R.; Landau, N. R. Nature 1996, 381, 661. 164. Kwong, P. D.; Wyatt, R.; Robinson, J.; Sweet, R. W.; Sodroski, J.; Hendrickson, W. A. Nature 1998, 393, 648.

165. Samson, M.; Libert, F.; Doranz, B. J.; Rucker, J.; Liesnard, C.; Farber, C.-M.; Saragosti, S.; Lapoumeroulie, C.; Cognaux, J.; Forceille, C.; Muyldermans, G.; Verhofstede, C.; Burtonboy, G.; Georges, M.; Imai, T.; Rana, S.; Yi, Y.; Smyth, R. J.; Collman, R. G.; Doms, R. W.; Vassart, G.; Parmentier, M. Nature 1996, 382, 722. 166. Connor, R. I.; Sheridan, K. E.; Ceradini, D.; Choe, S.; Landau, N. R. J. Exp. Med. 1997, 185, 621. 167. De Clercq, E. J. Med. Chem. 2005, 48, 1297. 168. Cocchi, F.; DeVico, A. L.; Garzino-Demo, A.; Arya, S. K.; Gallo, R. C.; Lusso, P. Science 1995, 270, 1811. 169. Simmons, G.; Clapham, P. R.; Picard, L.; Offord, R. E.; Rosenkilde, M. M.; Schwartz, T. W.; Buser, R.; Wells, T. N. C.; Proudfoot, A. E. I. Science 1997, 276, 276. 170. Baba, M.; Nishimura, O.; Kanzaki, N.; Okamoto, M.; Sawada, H.; Iizawa, Y.; Shiraishi, M.; Aramaki, Y.; Okonogi, K.; Ogawa, Y.; Meguro, K.; Fujino, M. Proc. Natl. Acad. Sci. USA 1999, 96, 5698. 171. Strizki, J. M.; Xu, S.; Wagner, N. E.; Wojcik, L.; Liu, J.; Hou, Y.; Endres, M.; Palani, A.; Shapiro, S.; Clader, J. W.; Greenlee, W. J.; Tagat, J. R.; McCombie, S.; Cox, K.; Fawzi, A. B.; Chou, C.-C.; Pugliese-Sivo, C.; Davies, L.; Moreno, M. E.; Ho, D. D.; Trkola, A.; Stoddart, C. A.; Moore, J. P.; Reyes, G. R.; Baroudy, B. M. Proc. Natl. Acad. Sci. USA 2001, 98, 12718. 172. Tagat, J. R.; McCombie, S. W.; Nazareno, D.; Labroli, M. A.; Xiao, Y.; Steensma, R. W.; Strizki, J. M.; Baroudy, B. M.; Cox, K.; Lachowicz, J.; Varty, G.; Watkins, R. J. Med. Chem. 2004, 47, 2405. 173. Dorr, P.; Westby, M.; Dobbs, S.; Griffin, P.; Irvine, B.; Macartney, M.; Mori, J.; Rickett, G.; Smith-Burchnell, C.; Napier, C.; Webster, R.; Armour, D.; Price, D.; Stammen, B.; Wood, A.; Perros, M. Antimicrob. Agents Chemother. 2005, 49, 4721. 174. Garcia-Perez, J.; Rueda, P.; Staropoli, I.; Kellenberger, E.; Alcami, J.; ArenzanaSeisdedos, F.; Lagane, B. J. Biol. Chem. 2011, 286, 4978. 175. Mollereau, C.; Parmentier, M.; Mailleux, P.; Butour, J. L.; Moisand, C.; Chalon, P.; Caput, D.; Vassart, G.; Meunier, J. C. FEBS Lett. 1994, 341, 33. 176. Pathan, H.; Williams, J. British Journal of Pain 2012, 6, 11. 177. Manglik, A.; Kruse, A. C.; Kobilka, T. S.; Thian, F. S.; Mathiesen, J. M.; Sunahara, R. K.; Pardo, L.; Weis, W. I.; Kobilka, B. K.; Granier, S. Nature 2012, 485, 321. 178. Granier, S.; Manglik, A.; Kruse, A. C.; Kobilka, T. S.; Thian, F. S.; Weis, W. I.; Kobilka, B. K. Nature 2012, 485, 400. 179. Lipkowski, A. W.; Tam, S. W.; Portoghese, P. S. J. Med. Chem. 1986, 29, 1222. 180. Fenalti, G.; Giguere, P. M.; Katritch, V.; Huang, X. P.; Thompson, A. A.; Cherezov, V.; Roth, B. L.; Stevens, R. C. Nature 2014, 506, 191. 181. Costa, T.; Lang, J.; Gless, C.; Herz, A. Mol. Pharmacol. 1990, 37, 383. 182. Wu, H.; Wacker, D.; Mileni, M.; Katritch, V.; Han, G. W.; Vardy, E.; Liu, W.; Thompson, A. A.; Huang, X.-P.; Carroll, F. I.; Mascarella, S. W.; Westkaemper, R. B.; Mosier, P. D.; Roth, B. L.; Cherezov, V.; Stevens, R. C. Nature 2012, 485, 327. 183. Runyon, S. P.; Brieaddy, L. E.; Mascarella, S. W.; Thomas, J. B.; Navarro, H. A.; Howard, J. L.; Pollard, G. T.; Carroll, F. I. J. Med. Chem. 2010, 53, 5290. 184. Lambert, D. G. Nat. Rev. Drug. Discov. 2008, 7, 694. 185. Mogil, J. S.; Pasternak, G. W. Pharmacol. Rev. 2001, 53, 381. 186. Thompson, A. A.; Liu, W.; Chun, E.; Katritch, V.; Wu, H.; Vardy, E.; Huang, X. P.; Trapella, C.; Guerrini, R.; Calo, G.; Roth, B. L.; Cherezov, V.; Stevens, R. C. Nature 2012, 485, 395. 187. Chu, R.; Takei, J.; Knowlton, J. R.; Andrykovitch, M.; Pei, W.; Kajava, A. V.; Steinbach, P. J.; Ji, X.; Bai, Y. J. Mol. Biol. 2002, 323, 253.

188. Goto, Y.; Arai-Otsuki, S.; Tachibana, Y.; Ichikawa, D.; Ozaki, S.; Takahashi, H.; Iwasawa, Y.; Okamoto, O.; Okuda, S.; Ohta, H.; Sagara, T. J. Med. Chem. 2006, 49, 847. 189. Coughlin, S. R. Nature 2000, 407, 258. 190. Chackalamannil, S. J. Med. Chem. 2006, 49, 5389. 191. Ahn, H.-S.; Arik, L.; Boykow, G.; Burnett, D. A.; Caplen, M. A.; Czarniecki, M.; Domalski, M. S.; Foster, C.; Manna, M.; Stamford, A. W.; Wu, Y. Bioorg. Med. Chem. Lett. 1999, 9, 2073. 192. Chackalamannil, S.; Xia, Y.; Greenlee, W. J.; Clasby, M.; Doller, D.; Tsai, H.; Asberom, T.; Czarniecki, M.; Ahn, H.-S.; Boykow, G.; Foster, C.; Agans-Fantuzzi, J.; Bryant, M.; Lau, J.; Chintala, M. J. Med. Chem. 2005, 48, 5884. 193. Chackalamannil, S.; Wang, Y.; Greenlee, W. J.; Hu, Z.; Xia, Y.; Ahn, H.-S.; Boykow, G.; Hsieh, Y.; Palamanda, J.; Agans-Fantuzzi, J.; Kurowski, S.; Graziano, M.; Chintala, M. J. Med. Chem. 2008, 51, 3061. 194. Zhang, C.; Srinivasan, Y.; Arlow, D. H.; Fung, J. J.; Palmer, D.; Zheng, Y.; Green, H. F.; Pandey, A.; Dror, R. O.; Shaw, D. E.; Weis, W. I.; Coughlin, S. R.; Kobilka, B. K. Nature 2012, 492, 387. 195. ; U.S. Food and Drug Administration, 2014. 196. Chelliah, M. V.; Eagen, K.; Guo, Z.; Chackalamannil, S.; Xia, Y.; Tsai, H.; Greenlee, W. J.; Ahn, H.-S.; Kurowski, S.; Boykow, G.; Hsieh, Y.; Chintala, M. ACS Med. Chem. Lett. 2014, 5, 561. 197. Abbracchio, M. P.; Burnstock, G.; Boeynaems, J.-M.; Barnard, E. A.; Boyer, J. L.; Kennedy, C.; Knight, G. E.; Fumagalli, M.; Gachet, C.; Jacobson, K. A.; Weisman, G. A. Pharmacol. Rev. 2006, 58, 281. 198. Akbar, G. K. M.; Dasari, V. R.; Webb, T. E.; Ayyanathan, K.; Pillarisetti, K.; Sandhu, A. K.; Athwal, R. S.; Daniel, J. L.; Ashby, B.; Barnard, E. A.; Kunapuli, S. P. J. Biol. Chem. 1996, 271, 18363. 199. Dorsam, R. T.; Kunapuli, S. P. J. Clin. Invest. 2004, 113, 340. 200. Ingall, A. H.; Dixon, J.; Bailey, A.; Coombs, M. E.; Cox, D.; McInally, J. I.; Hunt, S. F.; Kindon, N. D.; Teobald, B. J.; Willis, P. A.; Humphries, R. G.; Leff, P.; Clegg, J. A.; Smith, J. A.; Tomlinson, W. J. Med. Chem. 1999, 42, 213. 201. Springthorpe, B.; Bailey, A.; Barton, P.; Birkinshaw, T. N.; Bonnert, R. V.; Brown, R. C.; Chapman, D.; Dixon, J.; Guile, S. D.; Humphries, R. G.; Hunt, S. F.; Ince, F.; Ingall, A. H.; Kirk, I. P.; Leeson, P. D.; Leff, P.; Lewis, R. J.; Martin, B. P.; McGinnity, D. F.; Mortimore, M. P.; Paine, S. W.; Pairaudeau, G.; Patel, A.; Rigby, A. J.; Riley, R. J.; Teobald, B. J.; Tomlinson, W.; Webborn, P. J. H.; Willis, P. A. Bioorg. Med. Chem. Lett. 2007, 17, 6013. 202. Baqi, Y.; Atzler, K.; Köse, M.; Glänzel, M.; Müller, C. E. J. Med. Chem. 2009, 52, 3784. 203. Bach, P.; Boström, J.; Brickmann, K.; van Giezen, J. J. J.; Hovland, R.; Petersson, A. U.; Ray, A.; Zetterberg, F. Bioorg. Med. Chem. Lett. 2011, 21, 2877. 204. Simon, J.; Filippov, A. K.; Göransson, S.; Wong, Y. H.; Frelin, C.; Michel, A. D.; Brown, D. A.; Barnard, E. A. J. Biol. Chem. 2002, 277, 31390. 205. Chang, H.; Yanachkov, I. B.; Michelson, A. D.; Li, Y.; Barnard, M. R.; Wright, G. E.; Frelinger Iii, A. L. Thromb. Res. 2010, 125, 159. 206. Ding, Z.; Kim, S.; Kunapuli, S. P. Mol. Pharmacol. 2006, 69, 338. 207. Zhang, J.; Zhang, K.; Gao, Z.-G.; Paoletta, S.; Zhang, D.; Han, G. W.; Li, T.; Ma, L.; Zhang, W.; Muller, C. E.; Yang, H.; Jiang, H.; Cherezov, V.; Katritch, V.; Jacobson, K. A.; Stevens, R. C.; Wu, B.; Zhao, Q. Nature 2014, 509, 119. 208. Schmidt, P.; Ritscher, L.; Dong, E. N.; Hermsdorf, T.; Cöster, M.; Wittkopf, D.; Meiler, J.; Schöneberg, T. Mol. Pharmacol. 2013, 83, 256.

209. Zhang, K.; Zhang, J.; Gao, Z.-G.; Zhang, D.; Zhu, L.; Han, G. W.; Moss, S. M.; Paoletta, S.; Kiselev, E.; Lu, W.; Fenalti, G.; Zhang, W.; Muller, C. E.; Yang, H.; Jiang, H.; Cherezov, V.; Katritch, V.; Jacobson, K. A.; Stevens, R. C.; Wu, B.; Zhao, Q. Nature 2014, 509, 115. 210. Itoh, Y.; Kawamata, Y.; Harada, M.; Kobayashi, M.; Fujii, R.; Fukusumi, S.; Ogi, K.; Hosoya, M.; Tanaka, Y.; Uejima, H.; Tanaka, H.; Maruyama, M.; Satoh, R.; Okubo, S.; Kizawa, H.; Komatsu, H.; Matsumura, F.; Noguchi, Y.; Shinohara, T.; Hinuma, S.; Fujisawa, Y.; Fujino, M. Nature 2003, 422, 173. 211. Rayasam, G. V.; Tulasi, V. K.; Davis, J. A.; Bansal, V. S. Expert Opin. Ther. Targets 2007, 11, 661. 212. Garrido, D. M.; Corbett, D. F.; Dwornik, K. A.; Goetz, A. S.; Littleton, T. R.; McKeown, S. C.; Mills, W. Y.; Smalley Jr, T. L.; Briscoe, C. P.; Peat, A. J. Bioorg. Med. Chem. Lett. 2006, 16, 1840. 213. Christiansen, E.; Due-Hansen, M. E.; Urban, C.; Merten, N.; Pfleiderer, M.; Karlsen, K. K.; Rasmussen, S. S.; Steensgaard, M.; Hamacher, A.; Schmidt, J.; Drewke, C.; Petersen, R. K.; Kristiansen, K.; Ullrich, S.; Kostenis, E.; Kassack, M. U.; Ulven, T. ACS Med. Chem. Lett. 2010, 1, 345. 214. Urban, C.; Hamacher, A.; Partke, H. J.; Roden, M.; Schinner, S.; Christiansen, E.; Due-Hansen, M. E.; Ulven, T.; Gohlke, H.; Kassack, M. U. Naunyn-Schmiedeberg's Arch. Pharmacol. 2013, 386, 1021. 215. Negoro, N.; Sasaki, S.; Mikami, S.; Ito, M.; Suzuki, M.; Tsujihata, Y.; Ito, R.; Harada, A.; Takeuchi, K.; Suzuki, N.; Miyazaki, J.; Santou, T.; Odani, T.; Kanzaki, N.; Funami, M.; Tanaka, T.; Kogame, A.; Matsunaga, S.; Yasuma, T.; Momose, Y. ACS Med. Chem. Lett. 2010, 1, 290. 216. Takeda-Pharmaceutical. 2013. 217. Burant, C. F.; Viswanathan, P.; Marcinak, J.; Cao, C.; Vakilynejad, M.; Xie, B.; Leifke, E. The Lancet379, 1403. 218. Srivastava, A.; Yano, J.; Hirozane, Y.; Kefala, G.; Gruswitz, F.; Snell, G.; Lane, W.; Ivetac, A.; Aertgeerts, K.; Nguyen, J.; Jennings, A.; Okada, K. Nature 2014, 513, 124. 219. Hurst, D. P.; Grossfield, A.; Lynch, D. L.; Feller, S.; Romo, T. D.; Gawrisch, K.; Pitman, M. C.; Reggio, P. H. J. Biol. Chem. 2010, 285, 17954. 220. Sum, C. S.; Tikhonova, I. G.; Neumann, S.; Engel, S.; Raaka, B. M.; Costanzi, S.; Gershengorn, M. C. J. Biol. Chem. 2007, 282, 29248. 221. Lagerstrom, M. C.; Schioth, H. B. Nat. Rev. Drug Discov. 2008, 7, 339. 222. Hollenstein, K.; Kean, J.; Bortolato, A.; Cheng, R. K. Y.; Dore, A. S.; Jazayeri, A.; Cooke, R. M.; Weir, M.; Marshall, F. H. Nature 2013, 499, 438. 223. Siu, F. Y.; He, M.; de Graaf, C.; Han, G. W.; Yang, D.; Zhang, Z.; Zhou, C.; Xu, Q.; Wacker, D.; Joseph, J. S.; Liu, W.; Lau, J.; Cherezov, V.; Katritch, V.; Wang, M.-W.; Stevens, R. C. Nature 2013, 499, 444. 224. Riedel, G.; Platt, B.; Micheau, J. Behav. Brain Res. 2003, 140, 1. 225. Nakanishi, S. Science 1992, 258, 597. 226. Kunishima, N.; Shimada, Y.; Tsuji, Y.; Sato, T.; Yamamoto, M.; Kumasaka, T.; Nakanishi, S.; Jingami, H.; Morikawa, K. Nature 2000, 407, 971. 227. Muto, T.; Tsuchiya, D.; Morikawa, K.; Jingami, H. Proc. Natl. Acad. Sci. USA 2007, 104, 3759. 228. Gasparini, F.; Lingenhöhl, K.; Stoehr, N.; Flor, P. J.; Heinrich, M.; Vranesic, I.; Biollaz, M.; Allgeier, H.; Heckendorn, R.; Urwyler, S.; Varney, M. A.; Johnson, E. C.; Hess, S. D.; Rao, S. P.; Sacaan, A. I.; Santori, E. M.; Veliçelebi, G.; Kuhn, R. Neuropharmacology 1999, 38, 1493. 229. O'Leary, D. M.; Movsesyan, V.; Vicini, S.; Faden, A. I. Br. J. Pharmacol. 2000, 131, 1429.

230. Bear, M. F.; Huber, K. M.; Warren, S. T. Trends Neurosci. 2004, 27, 370. 231. Vranesic, I.; Ofner, S.; Flor, P. J.; Bilbe, G.; Bouhelal, R.; Enz, A.; Desrayaud, S.; McAllister, K.; Kuhn, R.; Gasparini, F. Biorg. Med. Chem. 232. Levenga, J.; Hayashi, S.; de Vrij, F. M. S.; Koekkoek, S. K.; van der Linde, H. C.; Nieuwenhuizen, I.; Song, C.; Buijsen, R. A. M.; Pop, A. S.; GomezMancilla, B.; Nelson, D. L.; Willemsen, R.; Gasparini, F.; Oostra, B. A. Neurobiol. Dis. 2011, 42, 311. 233. Dore, A. S.; Okrasa, K.; Patel, J. C.; Serrano-Vega, M.; Bennett, K.; Cooke, R. M.; Errey, J. C.; Jazayeri, A.; Khan, S.; Tehan, B.; Weir, M.; Wiggin, G. R.; Marshall, F. H. Nature 2014, 511, 557. 234. Rocher, J.-P.; Bonnet, B.; Bolea, C.; Lutjens, R.; Le Poul, E.; Poli, S.; Epping-Jordan, M.; Bessis, A.-S.; Ludwig, B.; Mutel, V. Curr. Top. Med. Chem. 2011, 11, 680. 235. O'Brien, J. A.; Lemaire, W.; Chen, T.-B.; Chang, R. S. L.; Jacobson, M. A.; Ha, S. N.; Lindsley, C. W.; Schaffhauser, H. J.; Sur, C.; Pettibone, D. J.; Conn, P. J.; Williams, D. L. Mol. Pharmacol. 2003, 64, 731. 236. Wu, H.; Wang, C.; Gregory, K. J.; Han, G. W.; Cho, H. P.; Xia, Y.; Niswender, C. M.; Katritch, V.; Meiler, J.; Cherezov, V.; Conn, P. J.; Stevens, R. C. Science 2014, 344, 58. 237. Varney, M. A.; Gereau 4th, R. W. Curr. Drug Targets: CNS Neurol. Disorders 2002, 1, 283. 238. Litschig, S.; Gasparini, F.; Rueegg, D.; Stoehr, N.; Flor, P. J.; Vranesic, I.; Prézeau, L.; Pin, J.-P.; Thomsen, C.; Kuhn, R. Mol. Pharmacol. 1999, 55, 453. 239. Gregory, K. J.; Dong, E. N.; Meiler, J.; Conn, P. J. Neuropharmacology 2011, 60, 66. 240. Clark, B. P.; Baker, S. R.; Goldsworthy, J.; Harris, J. R.; Kingston, A. E. Bioorg. Med. Chem. Lett. 1997, 7, 2777. 241. Kohara, A.; Toya, T.; Tamura, S.; Watabiki, T.; Nagakura, Y.; Shitaka, Y.; Hayashibe, S.; Kawabata, S.; Okada, M. J. Pharmacol. Exp. Ther. 2005, 315, 163. 242. Mabire, D.; Coupa, S.; Adelinet, C.; Poncelet, A.; Simonnet, Y.; Venet, M.; Wouters, R.; Lesage, A. S. J.; Beijsterveldt, L. V.; Bischoff, F. J. Med. Chem. 2005, 48, 2134. 243. Knoflach, F.; Mutel, V.; Jolidon, S.; Kew, J. N. C.; Malherbe, P.; Vieira, E.; Wichmann, J.; Kemp, J. A. Proc. Natl. Acad. Sci. USA 2001, 98, 13402. 244. Chen, Y.; Goudet, C.; Pin, J.-P.; Conn, P. J. Mol. Pharmacol. 2008, 73, 909. 245. Ruat, M.; Hoch, L.; Faure, H.; Rognan, D. Trends Pharmacol. Sci. 2014, 35, 237. 246. Wang, C.; Wu, H.; Katritch, V.; Han, G. W.; Huang, X.-P.; Liu, W.; Siu, F. Y.; Roth, B. L.; Cherezov, V.; Stevens, R. C. Nature 2013, 497, 338. 247. Wang, C.; Wu, H.; Evron, T.; Vardy, E.; Han, G. W.; Huang, X.-P.; Hufeisen, S. J.; Mangano, T. J.; Urban, D. J.; Katritch, V.; Cherezov, V.; Caron, M. G.; Roth, B. L.; Stevens, R. C. Nat Commun 2014, 5. 248. Weierstall, U.; James, D.; Wang, C.; White, T. A.; Wang, D.; Liu, W.; Spence, J. C. H.; Bruce Doak, R.; Nelson, G.; Fromme, P.; Fromme, R.; Grotjohann, I.; Kupitz, C.; Zatsepin, N. A.; Liu, H.; Basu, S.; Wacker, D.; Won Han, G.; Katritch, V.; Boutet, S.; Messerschmidt, M.; Williams, G. J.; Koglin, J. E.; Marvin Seibert, M.; Klinker, M.; Gati, C.; Shoeman, R. L.; Barty, A.; Chapman, H. N.; Kirian, R. A.; Beyerlein, K. R.; Stevens, R. C.; Li, D.; Shah, S. T. A.; Howe, N.; Caffrey, M.; Cherezov, V. Nat Commun 2014, 5. 249. Beachy, P. A. USA Patent WO2001027135 A2; 2002.