Abstract
G protein–coupled receptors have revealed themselves to be complex information-processing units that may be exploited for subtle therapeutic signaling effects. Thus, ligands may not only turn receptors on and off, but may also select from their repertoire of signaling effects to further refine drug response.
In the early days of receptor pharmacology, life was seemingly simple. There were agonists and antagonists, and pharmacologists could discover representatives of each to either turn on or turn off various cellular responses. However, with the advent of large-scale screening for synthetic ligands in functional receptor assays has come the realization that some molecules also have the ability to modify signals, not just initiate or block them; allosteric molecules that can modify signals are more prevalent than perhaps first realized. Similarly, with the advent of the ability to study receptors in contrived recombinant systems has come evidence to show that receptors are considerably more complex than mere on-off switches with variable volume control; that is, receptors may be more like microprocessors than rheostats.
The recent study by Tateyama and Kubo (1) illustrates a number of these new ideas related to receptors as microprocessors. Specifically, the authors show that metabotropic glutamate receptors 1α bind two types of ligand, typified by glutamate and Gd3+, and that these ligands activate two distinct cellular pathways involving Ca2+ [mediated through the Gq family of heterotrimeric guanine nucleotide–binding proteins (G proteins)] and adenosine 3′,5′-monophosphate (cAMP, mediated through Gs proteins). The receptor conformations produced by these ligands most likely differ, given that glutamate activates both pathways and induces receptor dimerization, whereas Gd3+ activates only the Ca2+ pathway and has a biphasic effect on dimerization. The metabotropic glutamate receptor 1α has been shown, from x-ray crystallographic studies, to adopt distinct structural "open" and "closed" conformations, and the experimental data of Tateyama and Kubo (1) are consistent with the authors’ suggestion that glutamate induces a receptor conformation referred to as closed-open/active, whereas Gd3+ stabilizes one referred to as closed-closed/active. It is worth considering the data obtained with this receptor in the context of what is generally known about the complex signaling properties of G protein–coupled receptors (GPCRs).
Although all proteins adopt a number of conformations, different physiological consequences can result when proteins in different conformations differentially interact with multiple components of cells. In the case of GPCRs, one type of observation that can be used to discern distinct receptor conformation is through the receptor interactions with different G proteins. Numerous receptors, including the metabotropic glutamate receptor 1α, are pleiotropic with respect to their interaction with G proteins. This fact, when coupled to the idea that different regions of the intracellular loops of GPCRs interact with different G proteins, compels the notion that different receptor conformations most likely reveal these intracellular regions differentially; that is, it is unlikely that two different "activated" receptor conformations will involve identical changes in two separate regions of the G protein–binding domains. Thus, it follows that different conformations will cause differential interaction with multiple G proteins. The distinct activation of G proteins that increase intracellular cAMP or Ca2+ concentration has been used experimentally in studies on the human calcitonin receptor to reveal distinctive signaling patterns (2). Specifically, the human calcitonin receptor type 2 was transfected into both normal human embryonic kidney (HEK) cells and into HEK cells cotransfected with Gs protein (cells enriched in Gs compared to wild-type cells), and the relative potencies of various calcitonin agonists in activating both responses were measured. In these studies, it was shown that certain agonists, such as porcine calcitonin, preferentially activated the cAMP pathway rather than the Ca2+ pathway, in comparison to agonists such as rat amylin. These data are consistent with the notion that porcine calcitonin stabilizes a receptor active state that preferentially interacts with the Gs protein (rather than the Gi protein). This is just one line of experimental investigation that has been used to detect the preferential activation of different responses by various agonists acting on the same receptor; this selective activation of cellular pathways has been termed "stimulus trafficking" [for reviews, see (3, 4)]. The therapeutic implications of such mechanisms are widespread in that the selection of particular signaling pathways through production of particular receptor active states adds another potential level of agonist selectivity beyond the reliance on receptor subtypes (Fig. 1). In addition to selecting signaling pathways, ligands also can be chosen to select for various other properties such as the ability to induce receptor desensitization (5) and internalization (6, 7).
The next level of selectivity: receptor active states. Although receptor signaling can be obtained by activation of receptor subtypes, there is evidence to suggest that different agonists may stabilize different receptor active states, further restricting the signaling pathways that are activated by that agonist. For a review of these effects, see (3).
The assignment of different physiological properties to different receptor conformations (that is, the different "efficacies" for the various receptor states) opens the possibility of matching the pharmacological properties of the ligand with the receptor active state, in essence choosing the efficacy therapeutically required. Thus, the efficacy of a ligand acting on a given receptor could be considered "collateral" (that is, the ability of a ligand to select a subset of receptor behaviors rather than the production of all of the receptor behaviors in a given cell) and not necessarily linear. In this latter instance, it is assumed that all physiological actions of the receptor necessarily follow upon activation by agonist, in essence a "linear" efficacy. There are now numerous instances where this has been shown not to be the case. For example, although parathyroid hormone (PTH) is known to produce activation of the PTH receptor and subsequent receptor internalization, amino-truncated peptide analogs [antagonists PTH(7–34) and PTH(7–84)] produce receptor internalization with receptor binding without causing receptor activation (8). If developed, collateral agonists could offer unique therapeutic opportunities such as opioid analgesia with no subsequent tolerance or internalization of CCR5 receptors to prevent HIV-1 entry with no subsequent activation of chemokine pathways and possible inflammation (9).
GPCRs are natural allosteric proteins in that ligands bind to one domain and alter receptor conformation in another intracellular domain to affect receptor–G protein interaction. The stabilization of any of the various conformations adopted by a given receptor could lead to a bias toward that conformation so that receptor behavior would predispose the production of physiological effects mediated by those conformations. In the study by Tateyama and Kubo (1), the separation of pathway activation is concomitant with separation of ligand binding; that is, glutamate and Gd3+ bind to separate regions of the receptor. It should be noted that such separate geography of binding is not necessary for the separate activation of physiological pathways. Indeed, the basis for the concept of agonist efficacy, as described by Stephenson (10), is that binding of a very similar series of alkyltrimethylammonium compounds to the muscarinic receptor results in a different maximal pharmacological response: The compounds bind to the same site but have differing efficacy for the production of response. However, the different binding domains for glutamate and Gd3+ illustrate the independence of allosteric effects on binding locale. Theoretically, any site on the GPCR surface that binds a molecule can stabilize a given receptor conformation and induce a particular pharmacological behavior—that is, physiological activity need not be linked to interaction at the GPCR endogenous agonist binding site. From this standpoint, screening for ligands with functional assays is optimal for defining such activity and is clearly superior to binding assays, in which only interactions that affect the binding of a radioligand with the endogenous GPCR binding site will be detected. For example, the allosteric agonist alcuronium produces activation of the muscarinic receptor that is insensitive to the orthosteric muscarinic receptor antagonist quinuclidinyl benzylate (QNB) (11); this agonist would not have been detected in a 3H-QNB binding assay.
There are increasing reports of ligand-selective GPCR effects, and the most simple hypothesis to account for such behavior is the stabilization of heterogeneous receptor conformations that then select for particular cellular pharmacological machinery. The key to exploiting this complex behavior is to have the appropriate assay systems to detect it. Although G-protein activation is a time-honored method of detecting interactions with GPCRs, new technology such as that used by Tateyama and Kubo (1) (fluorescent resonance energy transfer, or FRET) offers different eyes with which to view GPCR behavior. Assuming that particular behaviors are linked with specific receptor conformational changes, any means to detect changes in receptor state with ligand binding offers the potential to uncover activity of possible therapeutic utility. From this standpoint, GPCRs can be viewed as microprocessors with as yet untapped therapeutic potential.
