Crystallizing Thinking about the 2-Adrenergic Receptor

Two recently determined crystal structures of the human 2adrenergic receptor ( 2AR) provide a long-awaited advance in the field of G protein-coupled receptor research. The 2AR is only the second member of this, the largest family of receptors encoded in the human genome, whose structure has been solved. It follows structures of rhodopsin that were determined previously. Here we set these developments in historical context, discuss the daunting challenges that have been overcome, and appraise what has and has not been learned. For many years, the study of two prototypic members of the huge family of GPCRs, the visual light “receptor” rhodopsin and the 2AR for adrenaline and noradrenaline, has guided research in this field (Lefkowitz, 2007). Now, the recent publication of two crystal structures of the human 2AR once again permits new insights to be gained from comparison of the properties of these two model seven transmembranespanning receptors (7TMRs) (Cherezov et al., 2007; Rasmussen et al., 2007; Rosenbaum et al., 2007). The remarkable and unique abundance of rhodopsin in rod outer segments (it constitutes 90% of the protein in rod outer segment membranes) and its stability led to the determination of its complete amino acid sequence in 1982 by classic Edman degradation and to the appreciation of its seven transmembrane organization (Ovchinnikov, 1982; Hargrave et al., 1983). By comparison, the rarity of the 2AR and essentially all other GPCRs (they need to be purified several hundred thousandfold from naturally occurring sources to obtain homogeneous preparations) greatly hindered its biochemical study. Nonetheless, sufficient protein was purified so that, based on small stretches of amino acid sequence obtained from the receptor, its gene and cDNA were successfully cloned in 1986 (Dixon et al., 1986). It is remarkable that, in retrospect, it was only then that its close structural relationship with rhodopsin was first appreciated. This, despite the general understanding at the time that both rhodopsin and the 2AR signaled by activation of G proteins (transducin and Gs, respectively). The discovery of the homology between the 2AR and rhodopsin followed rapidly by the cloning of additional adrenergic receptors and others in turn triggered the rapid realization that all GPCRs share a conserved seven transmembrane organization and are members of the same gene family (Dohlman et al., 1991). The first crystal structure of rhodopsin in its inactive state was reported in 2000 (Palczewski et al., 2000), and the new 2AR structures are the first of any other GPCR to appear. Given the close parallels and centrality of research on these two model 7TMRs, it seems somehow fitting that a comparison of their molecular structures should once again be in the spotlight of molecular pharmacologists. Why Did It Take So Long? Crystallization of membrane proteins, especially eukaryotic ones, remains a very difficult and time-consuming process. In contrast to several thousand Protein Data Bank entries for soluble proteins, only 148 unique structures of membrane proteins have been determined, of which only 40 are eukaryotic membrane proteins, and only 4 of these are of human origin (available at http://www.blanco.biomol.uci.edu/ Membrane_Proteins_xtal). There are multiple problems asDr. Robert J. Lefkowitz is an Investigator with the Howard Hughes Medical Institute. Article, publication date, and citation information can be found at http://molpharm.aspetjournals.org. doi:10.1124/mol.108.045849. ABBREVIATIONS: GPCR, G protein-coupled receptor; 2AR, 2-adrenergic receptor; TM, transmembrane; 7TMR, seven transmembranespanning receptor. 0026-895X/08/7305-1333–1338$20.00 MOLECULAR PHARMACOLOGY Vol. 73, No. 5 Copyright © 2008 The American Society for Pharmacology and Experimental Therapeutics 45849/3326447 Mol Pharmacol 73:1333–1338, 2008 Printed in U.S.A. 1333 at A PE T Jornals on Jne 3, 2017 m oharm .aspeurnals.org D ow nladed from sociated with the crystallization of membrane proteins and GPCRs in particular. As noted, unlike rhodopsin, which is present in abundant amounts in rod outer segments and which can be purified easily, other GPCRs, including the 2AR, are expressed in only tiny amounts in tissues. Therefore, isolation of sufficient amounts for crystallization purposes requires heterologous overexpression of recombinant receptors, and even then, substantial purification is required. Another problem associated with crystallization of these receptors is their intrinsic conformational flexibility. To interact with a set of diverse ligands and transmit signals through multiple signaling pathways, 7TMRs adopt ensembles of different conformations (or different active and inactive states), which leads to conformational heterogeneity. Moreover, approximately 50% of the residues in these receptors are buried in the membrane bilayer, which limits the polar surface area available for crystal contacts. As discussed below, the successful addressing of these problems, in addition to recent advances in crystallography, were the keys to obtaining diffraction-quality crystals, which finally led to the structure determination of the 2AR. What Made It Possible? In addition to high-level heterologous expression of the 2AR in Sf9 cells to produce large amounts of functional receptors, two additional technical advances, one at the protein engineering level and the other at the crystallographic level, finally yielded well-ordered crystals that diffracted to high resolution. Based on prior biophysical studies, the third intracellular loop and the C terminus of the 2AR seem to be the most flexible regions (Granier et al., 2007). In addition to truncating part of the flexible C terminus of the 2AR, two parallel approaches were taken to address this issue. First, an antibody fragment (Fab) was generated against a threedimensional epitope corresponding to the third intracellular loop (Fig. 1A) using hybridoma technology (Day et al., 2007). This approach of cocrystallizing membrane proteins with an antibody fragment was first developed and used successfully in the cytochrome c oxidase structure determination and since then has been used for several other membrane proteins (Hunte and Michel, 2002). The second approach used to reduce the flexibility of the receptor and increase its polar surface area was to replace the third intracellular loop with T4 lysozyme, a highly crystallizable soluble protein (Fig. 1B). In addition to these protein engineering and antibody approaches, recent developments in membrane protein crystallography were also crucial to the success. In contrast to traditional detergent crystallization of membrane proteins, the 2AR was crystallized in either DMPC/CHAPSO bicelles or monolein lipidic cubic phase with cholesterol as the additive. Both of these methods, which essentially rely on the use of different lipids to present a more native environment to the protein, have been reported to yield well-diffracting crystals for several membrane proteins such as bacteriorhodopsin and the photosynthetic reaction center (Faham and Bowie, 2002). Moreover, because the crystals of the 2AR Antibody fragment (Fab) T4 lysozyme Extracellular