Opinion
The apparent cooperativity of some GPCRs does not necessarily imply dimerization

https://doi.org/10.1016/j.tips.2009.01.003Get rights and content

When the binding of one ligand to its receptor is influenced by a second ligand acting on a different receptor, one might assume that the receptors dimerize, enabling allosteric interactions between ligands. This reasoning is frequently used to explain the complex binding curves of ligands of class A G-protein-coupled receptors (GPCRs). Here, we argue that in classical in vitro experiments the lack of GTP makes ligand-binding properties dependent on the available pool of G protein. Under such conditions a 1:1 GPCR–G-protein complex is stabilized, in which the G protein lacks a nucleotide and ligand binding is of high affinity. In vivo, this complex, a key intermediate of G-protein activation, never accumulates because of fast and irreversible GTP binding. In vitro, this complex creates interference in ligand binding when two monomeric GPCRs compete for the same G protein. Interestingly, this competition explains some in vivo effects of orphan GPCRs.

Introduction

G-protein-coupled receptors (GPCRs) are classified into different subtypes 1, 2. The simplest are the rhodopsin-like class A receptors, which are characterized by a binding pocket for small ligands in the seven-helices transmembrane region. The most complex are the class C GPCRs, such as the metabotropic glutamate receptor, which are characterized by a large extracellular domain that associates in a constitutive homo- or heterodimer [3]. Class A GPCRs were originally considered to be monomeric, but subsequent functional biophysical, biochemical and pharmacological studies suggested that they could dimerize 4, 5. The relevance of these observations has, however, been questioned 6, 7, 8, 9. Moreover, the use of well-defined detergent micelles or nanodiscs has demonstrated that a single rhodopsin or a single β2-adrenoceptor molecule is capable of highly efficient coupling to a G protein 10, 11, 12.

Recently, Gurevich and Gurevich have thoroughly addressed the issue of GPCR dimerization from the structural point of view [13]. Here, we wish to complement this critical analysis by reassessing the evidence for dimerization that is based on pharmacological studies. In membrane preparations from cells expressing two closely related class A GPCRs, it is frequently observed that the binding of one ligand specific for the first receptor is affected by the addition of a second ligand acting on the second receptor. Depending on the nature of the ligands (i.e. agonist, inverse agonist), the apparent cooperativity can be negative or positive. Nevertheless, interference in ligand binding is usually considered to be a clue for the formation of GPCR heterodimers in which the two ligand-binding sites are coupled in an allosteric manner.

Before detailing several cases of apparent cooperativity, it is important to remember that, in membrane preparations, even the binding of a single agonist to a GPCR does not follow a simple bimolecular scheme. Pharmacologists are familiar with the GTP shift: a drop in the affinity of GPCRs for agonists induced by guanine nucleotides. This effect, which is still observed in preparation where GPCRs are unambiguously in a monomeric state (i.e. enwrapped in a nanodisc) [11], is the pharmacological indication that agonist-liganded GPCRs control the GDP/GTP cycle of G proteins (Figure 1). In part, the popular ‘ternary complex model’ accounts for the effect of G proteins on agonist affinity and is briefly summarized 14, 15. However, because this model does not detail the role of guanine nucleotides in the G-protein activation process and ignores the irreversible aspects of some reactions, we detail the catalytic model that has successfully accounted for the kinetics and stoichiometry of G-protein activation by class A GPCRs such as rhodopsin and the muscarinic receptor 16, 17. We argue that this model can explain the apparent ligand-binding cooperativity of class A GPCRs without implying any receptor dimer.

Section snippets

The ternary complex equilibrium model

The classical ‘ternary complex model’ [14] proposes that the receptor is in equilibrium between two states: R, the inactive conformation, and R*, the active conformation that activates the G protein. In the absence of ligand, R predominates. Agonists (A) bind only to R* and stabilize this active conformation, whereas inverse agonists bind only to R and stabilize the inactive conformation. It is a strict lock-and-key mechanism. Importantly, the model does not specify a mechanism for the

The ‘catalytic out-of-equilibrium’ kinetic model

A catalytic scheme was proposed in 1981 to account for the fast kinetics and high stoichiometry of transducin activation by rhodopsin 20, 21. This model was further developed with the identification of intermediate steps in the catalytic process and the observation that affinities were inadequate to describe the reaction flow 17, 22, 23. Later, a more elaborate form was proposed by Waelbroeck to describe the kinetics of muscarinic-receptor–G-protein coupling [16]. The catalytic model is

Apparent negative cooperativity might imply competition for the G protein

Class A GPCR pairs have often been assumed to form heterodimers on the basis of their apparent ligand-binding cooperativity. Two well-documented cases are that of the CCR2 and CCR5 chemokine receptors [28] and the μ- and δ-opioid receptors 29, 30, 31.

Let us consider two closely related class A GPCRs, say R1 and R2, that respond to specific agonists, A1 and A2, but are both active on the same type of G protein. The typical observation is that on membrane preparations in the absence of GTP the

How could the G-protein pool be limiting?

The relative stoechiometry between GPCRs and G proteins is difficult to address. Should we know the total number of G protein and receptors in a cell, we could not be certain that all potentially interact because subcompartmentalization should favor some interactions and exclude other ones. Obviously, the pool of G protein can be limiting when receptors have been overexpressed. Yet, even without overexpression, the G-protein pool seems usually to be smaller than that of their cognate receptors,

Apparent positive cooperativity of ligand binding to GPCRs

Some cases of positive cooperativity in ligand binding might also be explained by the catalytic model. If one receptor, say R1, has substantial constitutive activity, an inverse agonist acting on R1 can liberate the pool of G protein pre-coupled to R1, which becomes available to R2. This will increase the high-affinity-binding component of R2. Such an effect has been observed in the case of the μ- and δ-opioid receptors: the binding of an agonist of the μ-opioid receptor is of greater amplitude

Functional cooperativity in vivo: the case of an orphan receptor

Functional cooperativity of GPCR pairs in vivo should be difficult to explain by our model because the presence of GTP precludes the formation of nucleotide-free receptor–G-protein complexes. An interesting case, however, is that of GPR50, an orphan GPCR that has been suggested to inhibit the MT1 melatonin receptor through heterodimerization [34]. GPR50 and MT1 are closely related class A GPCRs and both are coupled to Gi. Co-expression of GPR50 with MT1 is shown to antagonize MT1 signaling in a

Conclusion

We have discussed three cases of apparent cooperativity between class A GPCRs that had been formerly presented as evidence for the existence of GPCR dimers. We argue that it is difficult to get a definitive proof of the existence of dimers from ligand-binding studies because an alternative explanation based on the formation of GPCR–G-protein complexes can be proposed. In this view, cooperativity is not due to an allosteric coupling of the two ligand-binding sites in a GPCR dimer, but rather to

Acknowledgements

We thank Cathy Jackson for comments on the manuscript.

References (39)

  • C.N. Davis

    Chemokine receptor binding and signal transduction in native cells of the central nervous system

    Methods

    (2003)
  • W.P. Hausdorff

    A mutation of the β2-adrenergic receptor impairs agonist activation of adenylyl cyclase without affecting high affinity agonist binding. Distinct molecular determinants of the receptor are involved in physical coupling to and functional activation of Gs

    J. Biol. Chem.

    (1990)
  • V.V. Gurevich

    Agonist-receptor-arrestin, an alternative ternary complex with high agonist affinity

    J. Biol. Chem.

    (1997)
  • M. Kosloff

    Electrostatic and lipid anchor contributions to the interaction of transducin with membranes: mechanistic implications for activation and translocation

    J. Biol. Chem.

    (2008)
  • J. Bockaert et al.

    Molecular tinkering of G protein-coupled receptors: an evolutionary success

    EMBO J.

    (1999)
  • R. Fredriksson

    The G-protein-coupled receptors in the human genome form five main families. Phylogenetic analysis, paralogon groups, and fingerprints

    Mol. Pharmacol.

    (2003)
  • R. Maggio

    Coexpression studies with mutant muscarinic/adrenergic receptors provide evidence for intermolecular ‘cross-talk’ between G-protein-linked receptors

    Proc. Natl. Acad. Sci. U. S. A.

    (1993)
  • D. Fotiadis

    Atomic-force microscopy: rhodopsin dimers in native disc membranes

    Nature

    (2003)
  • M. Chabre

    Biophysics: is rhodopsin dimeric in native retinal rods?

    Nature

    (2003)
  • Cited by (0)

    View full text