New G-protein-coupled receptor crystal structures: insights and limitations

Transcription

1 Opinion New G-protein-coupled receptor crystal structures: insights and limitations Brian Kobilka 1 and Gebhard F.X. Schertler 2 1 Departments of Molecular and Cellular Physiology and Medicine, Stanford University School of Medicine, 279 Campus Drive, Stanford, Palo Alto, CA 94305, USA 2 MRC Laboratory of Molecular Biology, Cambridge, CB2 0QT, UK G-protein-coupled receptors (GPCRs) constitute a large family of structurally similar proteins that respond to a chemically diverse array of physiological and environmental stimulants. Until recently, high-resolution structural information was limited to rhodopsin, a naturally abundant GPCR that is highly specialized for the detection of light. Non-rhodopsin GPCRs for diffusible hormones and neurotransmitters have proven more resistant to crystallography approaches, possibly because of their inherent structural flexibility and the need for recombinant expression. Recently, crystal structures of the human b 2 adrenoceptor have been obtained using two different approaches to stabilize receptor protein and increase polar surface area. These structures, together with the existing structures of rhodopsin, represent an important first step in understanding how GPCRs work at a molecular level. Much more high-resolution information is needed for this important family of membrane proteins, however: for example, the structures of GPCRs from different families, structures bound to ligands having different efficacies, and structures of GPCRs in complex with G proteins and other signaling molecules. Methods to characterize the dynamic aspects of the GPCR architecture at high resolution will also be important. Introduction Rhodopsin has been an extraordinarily valuable system for understanding the structure and mechanism of activation of G-protein-coupled receptors (GPCRs) [1 3]. Our first insights came from two- and three-dimensional crystals of rhodopsin [4 14], and recently a structure of a thermally stabilized mutant of rhodopsin was obtained using expressed protein [15]. Rhodopsin is highly specialized for the detection of light, however, and it exhibits functional and biochemical characteristics that distinguish it from GPCRs for diffusible hormones and neurotransmitters. Crystal structures have recently been determined for the human b 2 adrenoceptor (b 2 AR) [16 18], a receptor for adrenalin and noradrenalin that is involved in the regulation of cardiovascular and pulmonary function by the sympathetic nervous system. b 2 AR was the first nonrhodopsin GPCR to be cloned and is one of the most Corresponding author: Kobilka, B. (kobilka@stanford.edu). extensively studied members of this family [19]. These structures provide a long-awaited comparison and contrast with rhodopsin. Here, we will briefly review the technical obstacles that were involved in obtaining these structures, and the potential application of these methods to other GPCRs. We will assess what the structures have taught us about receptor activation, and what impact they might have in drug discovery. Finally, we will discuss the limitations of these structures, as well as future directions and opportunities in GPCR research. Technical challenges of GPCR crystallography Successful crystallization of the human b 2 AR involved overcoming several technical barriers: the production of large quantities of functional protein, the stabilization of flexible domains of the receptor structure through protein engineering or binding to an antibody fragment, providing a more native environment using lipid-based crystallography screens and obtaining X-ray data from very small crystals using microdiffraction technology. Expression and purification of large quantities of functional protein is the foundation of any crystallography effort [20]. Identifying the initial crystallization conditions and refining these conditions to produce diffraction-quality crystals for data collection required production and purification of over 200 mg of functional receptor from insect cells. The ability to produce protein was not by itself sufficient, however, because exhaustive trials failed to produce diffractionquality crystals of the wild-type b 2 AR alone. The inability to obtain crystals of native b 2 AR might be attributed in part to the dynamic nature of many non-rhodoposin GPCRs [21]. The third intracellular loop (ICL3) linking the cytoplasmic ends of transmembrane segments (TM) 5 and 6 is one of the most dynamic regions of the b 2 AR and many other GCPRs. This region, which varies considerably in length among family A receptors, is known to be important for the specificity of receptor-g-protein interactions and for G-protein activation. The dynamic nature of ICL3 might contribute to conformational heterogeneity and structural instability, two properties that interfere with the formation of diffraction-quality protein crystals. Crystals of wild-type b 2 AR were obtained by stabilizing the cytoplasmic ends of TM5 and TM6 linked by ICL3 using an antibody fragment (Fab5) that recognizes a three /$ see front matter ß 2007 Elsevier Ltd. All rights reserved. doi: /j.tips Available online 14 January

2 to facilitate crystallization of other GPCRs following exchange of the ICL3 with that of the b 2 AR. Figure 1. Structures of the b 2 AR. (a) Wild-type b 2 AR in complex with Fab5. (b) Engineered b 2 AR-T4-lysozyme fusion protein.tm, transmembrane segment; ICL2, second intracellular loop. Fab5 and T4 lysozyme serve similar functions in these two crystal structures. They both stabilize interactions between TM5 and TM6, and provide additional polar surface area for crystal lattice contacts. dimensional epitope comprising the N- and C-terminal ends of the native ICL3 [16,22] (Figure 1a). Antibody fragments have previously been shown to facilitate the crystallization of membrane proteins [23,24]. We also obtained crystals by protein engineering, replacing unstructured ICL3 sequence from Q230 to S262 with T4 lysozyme (T4L), a well-folded soluble protein (Figure 1b) [17,18]. These approaches also increased the polar surface area available for crystal lattice contacts. Both b 2 AR crystal forms were obtained in a lipidic environment. The b 2 AR-Fab5 crystals were grown in DMPC/CHAPSO bicelles [16], and b 2 AR-T4L crystals were grown in a cholesterol-doped monolein lipidic cubic phase [25,26]. It is likely that these lipid environments contribute to the stability of the receptor by maintaining it in a more native conformation. Finally, the development of microdiffraction technology [27,28] was critical for obtaining these b 2 AR structures, for both screening and data collection. Both b 2 AR-Fab5 and b 2 AR-T4L crystals were too small and radiation sensitive to collect data on conventional synchrotron beamlines where the width of the X-ray beams is typically greater than 50 microns. Data for b 2 AR crystals were obtained at the microfocus beamlines ID-13 and ID 23.2 at the European Synchrotron Radiation Facility (ESRF) in Grenoble and beamline GM/CA-CAT at the Advanced Photon Source (APS) in Argonne, where high-intensity X-ray beams smaller than 10 microns are produced. The value of this work will be even greater if some of the methods used to obtain b 2 AR crystals can be applied to other GPCRs. Stabilization of helices linked by ICL3 might facilitate crystal formation for other GPCRs. It should be possible to generate similar T4L fusions for many family A receptors. Although Fab5 is specific for the b 2 AR, methods used to obtain the antibody could be applied to other GPCRs [22], and it is possible that Fab5 could be used Structural insights into receptor function High-resolution structural information is essential for understanding molecular mechanisms of protein function; however, some limitations of crystallography must be recognized. In forming a crystal, a protein becomes locked in a single conformational state. This is a significant drawback when one considers the body of functional and biophysical evidence that GPCRs are conformationally complex and dynamic proteins [21]. They do not behave as simple bimodal switches but adopt conformations that are specific for the bound ligand and the associated signaling partner (e.g. G proteins, arrestins). In the reported crystal structures, the conformation of the b 2 AR bound to carazolol is close to an inactive state. Carazolol is an inverse agonist, but suppresses 50% of basal activity in the b 2 AR [16]. Therefore, the b 2 AR structure might not represent a fully inactive receptor and could differ significantly from the unliganded receptor or one of the potential active states. The structures of the b 2 AR do provide new insights into process of activation in several ways. There is a relatively large body of mutagenesis and biophysical data for the b 2 AR and closely related receptors. We now have a structural framework for interpreting these studies and for generating testable hypothesis for future studies. As an example, there are several amino acids where mutations lead to elevated basal activity (constitutive activity mutants [CAMs]). In the crystal structure, these amino acids form packing interactions with sidechains of one or more adjacent TM segments, thereby stabilizing an inactive state. Several of these CAMs are linked through packing interactions with amino acids that are essential for receptor activation [18]. These observations suggest that the process of activation involves sidechain rearrangement of many amino acids on adjacent TMs extending from the ligand binding site to cytoplasmic domains. Activation of G proteins and other cytosolic signaling proteins could be mediated by these sidechain rearrangements and/or by larger-scale movement of TM segments that might result from these rearrangements. Perhaps the greatest insights into structure/function relationships can be gained from comparing b 2 AR and rhodopsin structures. Rhodopsin is a GPCR that evolved for the efficient detection of light. It is present in only one organ and serves only one purpose. In the dark state it has virtually no activity toward its G protein transducin, but a single photon can photoisomerize its covalently bound ligand, retinal, and convert it from an inverse agonist to a full agonist in a matter of femptoseconds. By contrast, the b 2 AR, like many other GPCRs, has a broader range of signaling behavior: coupling to more than one G protein and to G protein independent pathways, and responding to a spectrum of diffusible ligands [29]. Like many other GPCRs, the b 2 AR exhibits basal, agonist-independent activity towards its preferred G protein Gs. A comparison of the structures provides clues to the structural basis for these functional differences. In rhodopsin, the binding pocket for covalently bound retinal is 80

3 Figure 2. A comparison of the extracellular domains of the b 2 AR and rhodopsin. (a) In rhodopsin, extracellular loop 2 (ECL2) and the N terminus form a lid over the retinal (blue) binding pocket. (b) In the b 2 AR, ECL2 is displaced away from the carazolol (yellow) binding pocket by two disulfide bonds (gold). TM, transmembrane segment; ECL2, second extracellular loop. covered by a b sheet lid formed by the second extracellular loop (ECL2), and a cap-like microdomain formed by the N terminus, protecting cis-retinal from hydrolysis (Figure 2a). This structure would limit access for diffusible agonists and is not present in the b 2 AR (Figure 2b). By contrast, in the b 2 AR ECL2 forms a helix that is constrained by two disulfide bonds such that there is open access to the ligand-binding pocket. As noted above, the b 2 AR exhibits basal, agonist-independent signaling, even when bound to the inverse agonist carazolol, whereas rhodopsin has virtually no basal activity. An alignment of b 2 AR and rhodopsin reveals structural differences around the highly conserved E/ DRY sequence at the cytoplasmic end of TM3 that might contribute to functional differences in basal activity and/or the specificity of G-protein coupling. In rhodopsin, the second intracellular loop (ICL2) is displaced outward away from the TM segments. By contrast, in the b 2 AR, residues in ICL2 interact with residues in TM segments 2, 3, 5 and 6. Of particular interest, the ionic interaction (the ionic lock [30]) between the R135 of ERY (TM3) and E247 at the end of TM6 of rhodopsin is absent in both crystal structures of the b 2 AR [16,18]. As a result, the distance between cytoplasmic ends of TM3 and TM6 is more than 1.5Å greater for the b 2 AR. In contrast to these differences, evidence from biophysical [31,32] and functional studies [33,34] suggest that the b 2 AR and rhodopsin undergo similar conformational changes in the process of activation. In the crystal structure of rhodopsin, a network of hydrogen bonds involving a cluster of ordered water molecules extends from the retinal binding pocket to the cytoplasmic end of helices involved in G-protein coupling. This network, which includes several of the most highly conserved residues in family A GPCRs, has been proposed to be involved in propagating activating conformational changes in rhodopsin and other GPCRs [14,35]. This hypothesis is supported by the existence of a similar network in b 2 AR [18]. Implications for drug discovery Over the past decade, crystallography has become a validated tool in drug discovery for soluble protein targets [36]. Crystal structures can be used for in silico screening to identify lead compounds or to prescreen a complex chemical library for the most promising candidates. Structures have also been used in optimizing leads identified by highthroughput screening. A relatively new and potentially powerful approach, fragment-based drug discovery [37,38], involves screening comparatively small libraries (typically around 1000 compounds) of small organic molecules ( Da) using NMR or protein crystallography to identify ligand fragments that bind with relatively low affinity (high micromolar range) to non-overlapping sites in the binding pocket. These ligand fragments can then be joined to form higher-affinity leads. This approach has the advantage of covering a larger fraction of chemical space with smaller libraries of compounds [37]. Fragment-based screening has been most widely applied using NMR spectroscopy; however, successful applications using protein crystallography has also been reported [37]. With crystal structures available, it will now be possible to apply some structure-based drug discovery tools to GPCRs. However, it is unlikely that GPCR crystallography can be applied to methods such as fragment-based drug discovery where crystals are soaked in libraries of compounds and hits are identified from diffraction data. In the case of the b 2 AR, the current crystals are too fragile to soak in different ligands. Moreover, to use this approach, it will probably be necessary to start with receptor protein crystallized in the absence of ligand. This would be more challenging because of the absence of the structural stabilization afforded by a bound ligand. The most immediate application of new GPCR structures to drug discovery will be the use of ligand docking programs for in silico screening and lead optimization. To date, these methods have been limited to rhodopsin-based homology models. Although the structures of b 2 AR and rhodopsin are highly homologous, the differences are large enough that rhodopsin-based homology models of family A GPCRs might not be sufficiently accurate for ligand-docking experiments. The b 2 AR structure might prove to be a better starting point than rhodopsin for homology-based models, particularly for the closely related members of the monoamine subfamily of GPCRs; however, these models 81

4 might still be inadequate, especially in screening for agonists. As discussed above, the current structure of the b 2 AR is an inactive state, and may therefore only be useful for identifying ligands having an efficacy similar to that of carazolol. Structures of the b 2 AR bound to ligands having different efficacies (neutral antagonist, partial agonists and full agonists) will be needed to understand the extent to which the ligand-binding pocket is altered by the bound ligand. As discussed above, these structures probably cannot be obtained by simply soaking receptor crystals in solutions of different ligands. As such, it will be necessary to obtain new crystals for each ligand. Because the receptor might have a slightly different structure for each ligand, the conditions for crystallization may differ, and it might not be possible to obtain diffraction-quality crystals with all ligands. In the case of b 2 AR agonists, biophysical studies provide evidence for intermediate conformational states such that at equilibrium the agonist bound b 2 AR exists in at least two different conformational states [31,39]. This conformational heterogeneity could prevent crystal formation, and if crystals form, it might not be possible to determine if the conformation represents the most active state or an intermediate state. For the b 2 AR, and probably most GPCRs, a well-defined fully-active state might only exist when the receptor is bound to an agonist and complexed with a G protein. Future directions The crystal structures of the b 2 AR represent an important milestone, yet we are still in the infancy of GPCR structural biology. To understand the structural basis for the functional versatility of GPCRs and to fully enable structure-based drug discovery, we will need high-resolution structures of many more GPCRs, structures of GPCRs bound to ligands having different efficacies, and structures of GPCRs in complex with G proteins and other signaling partners. As discussed above, crystallography provides a static view of a very dynamic signaling molecule. We will also need to apply non-crystallographic biophysical tools that can characterize the dynamic nature of these conformationally complex membrane proteins at a comparable resolution. NMR spectroscopy is a versatile tool that has the potential to provide high-resolution structural information about receptor-drug interactions and about the dynamic aspects of GPCR structure. To do this, it is not necessary to solve the three-dimensional solution structure using NMR (a task that currently might be impossible because of GPCR size). Much structural information is accessible, however, if sequence-specific resonance assignments can be made for the polypeptide backbone. This is currently possible for proteins up to 150 kda if they are uniformly labeled with the NMR active isotopes 15 N and 13 C [40]. The b 2 AR is a 50 kda protein; when the detergent micelle is included, the apparent mass is closer to 90 kda. This is well within the size range for obtaining resonance assignments. Once it is feasible to obtain and assign high-resolution NMR spectra, it is possible to map chemical shift changes induced by binding of small-molecule ligands (including fragment-based drug discovery) or protein binding partners. Such chemical-shift mapping experiments could yield two types of valuable information: (a) the identity of residues involved in ligand binding for ligands having a spectrum of efficacies; and (b) the identity of protein domains undergoing allosteric conformational changes. The former has clear implications for understanding drug-binding mechanisms and for drug discovery; the latter for understanding fundamental mechanisms of GPCR signaling. Constraints derived from NMR experiments will be particular powerful when combined with GPCR crystal structures to generate more accurate models of conformational changes. NMR is applied routinely to soluble proteins for structure determination, dynamical studies, and drug discovery; however, the application of NMR to GPCRs is currently not practicable. This is primarily owing to limitations in sample preparation (expression, isotope labeling, and purification) rather than limitations in NMR instrumentation and methodology. Success in solving these technical problems, together with continued progress in NMR instrumentation and methodology, will see this biophysical tool playing a more prominent role in the structural biology of GPCRs and other membrane proteins. Finally, GPCR structural biology is a high-risk, expensive endeavor. Projects may require many years of work with relatively few publishable results before success is achieved. Current mechanisms for funding academic research often consider such projects too risky, or do not provide adequate funds. Future progress in GPCR structural biology will only occur if this funding problem can be addressed. Acknowledgements B.K.K. was supported by grants from the National Institutes of Health (R01NS028471, R21GM075811) and gifts from the Mather Charitable Foundations and Lundbeck. G.F.X.S. was supported by a Human Frontier Science Project (HFSP) program grant (RG/0052), a European Commission FP6 specific targeted research project (LSH ) and an ESRF long term proposal. References 1 Sakmar, T.P. et al. (2002) Rhodopsin: insights from recent structural studies. Annu. Rev. Biophys. Biomol. Struct. 31, Hubbell, W.L. et al. (2003) Rhodopsin structure, dynamics, and activation: a perspective from crystallography, site-directed spin labeling, sulfhydryl reactivity, and disulfide cross-linking. Adv. Protein Chem. 63, Palczewski, K. (2006) G protein-coupled receptor rhodopsin. Annu. Rev. Biochem. 75, Krebs, A. et al. (2003) The three-dimensional structure of bovine rhodopsin determined by electron cryomicroscopy. J. Biol. Chem. 278, Krebs, A. et al. (1998) Characterisation of an improved twodimensional p22121 crystal from bovine rhodopsin. J. Mol. Biol. 282, Schertler, G.F. (1998) Structure of rhodopsin. Eye 12, Unger, V.M. et al. (1997) Arrangement of rhodopsin transmembrane alpha-helices. Nature 389, Schertler, G.F. et al. (1993) Projection structure of rhodopsin. Nature 362, Okada, T. et al. (2000) X-Ray diffraction analysis of three-dimensional crystals of bovine rhodopsin obtained from mixed micelles. J. Struct. Biol. 130, Palczewski, K. et al. (2000) Crystal structure of rhodopsin: A G proteincoupled receptor. Science 289,

5 11 Okada, T. et al. (2002) Functional role of internal water molecules in rhodopsin revealed by X-ray crystallography. Proc. Natl. Acad. Sci. U. S. A. 99, Teller, D.C. et al. (2001) Advances in determination of a high-resolution three-dimensional structure of rhodopsin, a model of G-protein-coupled receptors (GPCRs). Biochemistry 40, Okada, T. et al. (2004) The retinal conformation and its environment in rhodopsin in light of a new 2.2 A crystal structure. J. Mol. Biol. 342, Li, J. et al. (2004) Structure of bovine rhodopsin in a trigonal crystal form. J. Mol. Biol. 343, Standfuss, J. et al. (2007) Crystal structure of a thermally stable rhodopsin mutant. J. Mol. Biol. 372, Rasmussen, S.G. et al. (2007) Crystal structure of the human beta(2) adrenergic G-protein-coupled receptor. Nature 450, Cherezov, V. et al. (2007) High-resolution crystal structure of an engineered human {beta}2-adrenergic G protein coupled receptor. Science 318, Rosenbaum, D.M. et al. (2007) GPCR Engineering yields highresolution structural insights into {beta}2 adrenergic receptor function. Science 318, Lefkowitz, R.J. (2000) The superfamily of heptahelical receptors. Nat. Cell Biol. 2, E133 E Kobilka, B.K. (1995) Amino and carboxyl terminal modifications to facilitate the production and purification of a G protein-coupled receptor. Anal. Biochem. 231, Kobilka, B.K. and Deupi, X. (2007) Conformational complexity of G- protein-coupled receptors. Trends Pharmacol. Sci. 28, Day, P.W. et al. (2007) A monoclonal antibody for G protein-coupled receptor crystallography. Nat. Methods 4, Iwata, S. et al. (1995) Structure at 2.8 A resolution of cytochrome c oxidase from Paracoccus denitrificans. Nature 376, Zhou, Y. et al. (2001) Chemistry of ion coordination and hydration revealed by a K+ channel-fab complex at 2.0 A resolution. Nature 414, Landau, E.M. and Rosenbusch, J.P. (1996) Lipidic cubic phases: a novel concept for the crystallization of membrane proteins. Proc. Natl. Acad. Sci. U. S. A. 93, Caffrey, M. (2003) Membrane protein crystallization. J. Struct. Biol. 142, Cusack, S. et al. (1998) Small is beautiful: protein microcrystallography. Nat. Struct. Biol. 5 (Suppl), Riekel, C. et al. (2005) Protein crystallography microdiffraction. Curr. Opin. Struct. Biol. 15, Lefkowitz, R.J. and Shenoy, S.K. (2005) Transduction of receptor signals by beta-arrestins. Science 308, Ballesteros, J.A. et al. (2001) Activation of the beta 2-adrenergic receptor involves disruption of an ionic lock between the cytoplasmic ends of transmembrane segments 3 and 6. J. Biol. Chem. 276, Yao, X. et al. (2006) Coupling ligand structure to specific conformational switches in the beta2-adrenoceptor. Nat. Chem. Biol. 2, Farrens, D.L. et al. (1996) Requirement of rigid-body motion of transmembrane helices for light activation of rhodopsin. Science 274, Kim, J.M. et al. (2005) Light-driven activation of beta 2-adrenergic receptor signaling by a chimeric rhodopsin containing the beta 2-adrenergic receptor cytoplasmic loops. Biochemistry 44, Cerione, R.A. et al. (1985) Specificity of the functional interactions of the b-adrenergic receptor and rhodopsin with guanine nucleotide regulatory proteins reconstituted in phospholipid vesicles. J. Biol. Chem. 260, Pardo, L. et al. (2007) The role of internal water molecules in the structure and function of the rhodopsin family of G protein-coupled receptors. Chem BioChem 8, Blundell, T.L. et al. (2006) Structural biology and bioinformatics in drug design: opportunities and challenges for target identification and lead discovery. Philos. Trans. R. Soc. Lond. B Biol. Sci. 361, Carr, R.A. et al. (2005) Fragment-based lead discovery: leads by design. Drug Discov. Today 10, Hajduk, P.J. and Greer, J. (2007) A decade of fragment-based drug design: strategic advances and lessons learned. Nat. Rev. Drug Discov. 6, Ghanouni, P. et al. (2001) Functionally different agonists induce distinct conformations in the G protein coupling domain of the beta 2 adrenergic receptor. J. Biol. Chem. 276, Pellecchia, M. et al. (2002) NMR in drug discovery. Nat. Rev. Drug Discov. 1,