Almost two thousand genes have been identified, on the basis of their predicted protein structure, that encode for proteins characterized as G-protein-coupled receptors (GPCRs). It is likely that there are actually several thousand. GPCRs are currently classified into 100 or so subfamilies, based on their structure, ligand specificity, and biological effects. These molecules have a characteristic sequence that predicts seven transmembrane-spanning segments, with a cytoplasmic tail of varying lengths. GPCRs are known to interact with a variety of hormones, neuropeptides, chemokines, biogenic amines, nucleosides, eicosanoids, phospholipids, growth factors, and aromatic compounds that act as olfactants or odorants. Some examples include thyrotropin, interleukin 8, epinephrine, and bombesin. Moreover, scores of "orphan" receptors have been identified by cDNA cloning and are still in search of ligands.
1. Receptor Structure
The wide variety of ligands for GPCRs dictate that several structural variations might allow for the variety of their regulatory features. One common feature of all of these receptors is seven stretches of 20 to 25 hydrophobic amino acid residues, which are thought to form a-helices that span the membrane (see a-Helix). Thus, the overall structures of these receptors resemble that of bacteriorhodopsin (1). The amino-terminal sequences lie outside the cell, and they vary in size among the family from 7 to 600 amino acid residues. In the case of some receptors, especially those for larger peptide hormones, such as thyrotropin or thrombin, the amino-terminal sequences are critically involved in ligand binding. In most cases, however, ligand binding is thought to occur primarily within key residues in the membrane-spanning regions. A series of mutational studies with the b-adrenergic receptor and other GPCRs (2) have suggested that the ligand may associate with the a-helix near the outer surface, binding in a plane parallel with the membrane. Another conserved structural feature is the loop between the fifth and sixth a-helices, the domain most critical for interactions with G proteins (see GTP-Binding Proteins). It is likely that agonist binding induces a conformational change in the receptor that permits interactions of this domain with G proteins (3). The carboxyl tail of the receptor is cytoplasmically oriented and varies in length among receptors. Mounting evidence suggests that this domain may play a crucial role in receptor cross-talk and desensitization.
Identification of the regions in GPCRs involved in ligand binding has emerged mainly from studies employing point mutations and receptor chimeras. Advances in this area have been led by studies on b-adrenergic receptors (4). Mutant forms of the receptor have been constructed and expressed in cells, to evaluate the binding of both agonists and antagonists. Studies such as these have led to models in which the ligand is thought to bind at the outer surface of the receptor, among several of the transmembrane domains. In the case of the catecholamine agonist isoproterenol, two serine residues in the fifth transmembrane helix are thought to form hydrogen bonds with the hydroxyl groups on the catechol ring (2). In addition, a phenylalanine residue in a-helix 6 participates in binding via hydrophobic interactions with the catechol ring itself. Similar types of interactions have been elucidated for nucleoside and nucleotide receptors.
Agonist binding to GPCRs is thought to induce movements among the transmembrane domains, a conformational change that is translated into G-protein interactions. Although the molecular dynamics of this process remain poorly understood, a number of models have been proposed to explain signal generation within the receptor (3). One aspect of receptor activation that may play an important role is receptor dimerization, a process that is well characterized in activation of other receptor subtypes. Recently, it has been proposed that the protonation of key transmembrane residues are required for activation. Additionally, it has been reported that intramolecular interactions within certain intracellular loops maintain the receptor in the inactive state prior to ligand binding, which is subsequently disrupted by the interaction with activating ligand.
The specificities of receptors, both for ligands and for G proteins, have been explored by construction of chimeras. Although the overall structural features of these receptors are generally conserved, there is little sequence similarity among the superfamily. Indeed, even among the three known subtypes of the b-receptor, there is only 50% identity. Thus, chimeric receptors have revealed domains of the proteins that are necessary for binding specificity and that dictate which G proteins might be coupled to each receptor (4).
2. G Proteins
The proteins that serve as tranducers for the GPCRs are known as GTP-binding proteins, or G proteins. These molecular linkers exhibit a heterotrimeric structure, consisting of a, b,and g subunits. The heterotrimeric G proteins belong to a superfamily of Gtpases, which share a common structural core. In general, the functions of G proteins are dictated by the receptors with which they interact, an interaction mediated by the asubunit. Thus far, over 30 different a subunits have been identified, along with six b and g subunits. The a subunits fall into five different classes, each with a characteristic impact on its effector.
The asubunit of G proteins interacts with guanyl nucleotides and in general controls the activity of the complex (5, 6). When no ligand is bound to the receptor, a is maintained in the GDP-bound state. In this form, a is complexed to the b and g subunits, which prevents it from interacting with effectors. b and g subunits together form a tight complex. Upon ligand binding, the receptor interacts with the asubunit, leading to the displacement of GDP and the subsequent binding of GTP. In its GTP-bound state, Ga then dissociates from the bgcomplex and is activated, interacting with effectors such as adenylate cyclase. This activation is temporal and is terminated by the intrinsic GTPase activity of Ga, which hydrolyzes the GTP, leaving GDP bound to the protein.
The GTPase activity of Gafamily members is subject to different types of regulation. The proteins can be frozen in their activated state by adding nonhydrolyzable analogues of GTP, such as GTPgS, effectively blocking the GTPase activity of the protein. Additionally, certain bacterial toxins irreversibly modify G-protein a subunits, catalyzing the covalent addition of ADP-ribose from intracellular NAD. ADP-ribosylated Ga cannot hydrolyze GTP, thus leaving the protein in a persistently active state. In the case of cholera toxin, this activation results in a modification of Gas, causing the persistent activation of adenylyl cyclase, and leading to massive diarrhea (5, 6). Ga proteins are also regulated by endogenous proteins called for regulators of G proteins (RGs), (7). This family of regulatory proteins now comprises over 20 members, and they seem to function as long-term regulators of G protein signaling.
Although they were originally thought to function only as inhibitors of a, bg subunits in some cases can interact directly with effectors upon their release from Ga subunits, causing either stimulation or inhibition of activity. For example, in certain neuronal systems, cells may contain different forms of adenylyl cyclase that are differentially regulated by G proteins. bg subunits can stimulate the activity of the type III form of cyclase, but have no effect, or can even inhibit, the type II form of the enzyme. Additionally, bg subunits can directly stimulate phospholipase C-b in some cells (8).
3. Effectors
In most cases, G-protein a subunits interact directly with one or more effector proteins in the plasma membrane, to induce the release of soluble second messengers that in turn amplify the hormonal signal into the cell. While the number of effectors continues to grow, four distinct types have been studied in detail: adenylate cyclases, phospholipases, ion channels, and phosphodiesterases.
