An in vivo demonstration of functional G protein-coupled receptor dimers.

W ith the demonstration that two inactive lutropin receptor mutants can complement each other functionally, in vivo, the study by I. Huhtaniemi’s group in this issue of PNAS (1) constitutes a landmark in the long-lasting saga about di/oligomerization of rhodopsin-like G protein-coupled receptors (class A GPCRs). In an early “dimer period,” arguments for dimerization were found in the negative cooperativity displayed by some GPCRs (2) and in the disproportionately large molecular radius of some of them, as obtained by radiation inactivation (3). Thereafter, a “monomer period” ensued, when cloning of hundreds of molecules displaying a stereotypical serpentine domain with seven transmembrane segments made it reasonable to consider monomolecular GPCRs as being already “polymers”. . . of their transmembrane helices. Over the past 15 years, arguments favoring the existence of di/oligomers of rhodopsinlike GPCRs accumulated from a wide variety of experimental approaches. These include plain Western blotting, coimmunoprecipitation, functional complementation of mutants in transfected cells, FRET and BRET experiments, and atomic force microscopy (4, 5). By analogy, the functional or structural demonstrations that class C GPCRs (GABAB receptor, metabotropic glutamate receptors, and sweet-umami receptors) function as exclusive homoor heterodimers (6), added arguments to the di/ oligomeric nature of class A GPCRs. In a revival of the early “cooperativity period,” a number of studies endowing putative hetero-di/oligomers with properties different from those of putative homo-di/ oligomers (7), presented interaction between protomers as the basis of the apparent allosteric behavior of GPCRs (8). The case for di/oligomerization of rhodopsin-like GPCRs is not definitively set, however. Convincing studies have demonstrated that rhodopsin or the beta2 adrenergic receptor can signal to their respective G protein as monomeric units (9, 10). Also, the cooperative behavior displayed by some GPCRs in native or transfected cells has been proposed to reflect postreceptor regulatory effects rather that di/oligomerization (11). With the exception of the very peculiar case of rhodopsin in the retina (5), all experiments demonstrating a physical interaction between class A GPCRs have up to now been performed ex vivo, in transfected cells, which implies nonphysiological cellular environments and, in most cases, unrealistic concentrations of receptors in the membranes. In this context the in vivo study by Rivero-Müller et al. (1) constitutes a major contribution to the field. The luteinizing hormone receptor (LHR), together with the follitropin and thyroid stimulating hormone (TSH) receptors, belongs to the glycoprotein hormone receptors, a subfamily of rhodopsin-like GPCRs characterized by a large ectodomain responsible for hormone binding (12). Huhtaniemi’s group built on the observation that some loss-of-function mutants of the LHR displayed functional complementation when transfected in vitro (13). The first mutant they chose is deficient for hormone binding as it bears an amino acid substitution in the ectodomain. The second mutant harbors a deletion of transmembrane helices VI and VII. Although it is competent for hormone binding, it is completely deficient for signal transduction. Following careful characterization of the phenotype of each mutant in transfected HEK293 cells, and after demonstration by immunoprecipitation and Western blotting that tagged mutants were capable of specific interaction in transiently transfected cells, Rivero-Müller et al. shifted from the culture dish to the living transgenic animal (for an illustration of the strategy, see Fig. 1). They used as recipient animals a mouse line generated previously by homologous recombination, which is totally deficient for the LHR and, as a consequence, displays profound hypogonadism and complete infertility (14). Two transgenic lines were established by microinjection in fertilized oocytes of bacterial artificial chromosomes bearing the individual loss-of-function LHR mutants in the genomic environment of the normal LHR gene. This ensures (and it was verified by the authors) that, in the two mouse lines, the mutant LHR genes are expressed qualitatively (i.e., in the normal cell types) and quantitatively normally. As expected, when expressed on the LHR knockout background by an adequate breeding strategy, both mutants were well expressed in the testes, but only homogenates of the signaling-deficient mutant displayed efficient binding of IhCG. As also expected, the hypogonadism phenotype of the LHR knockout animals was not modified by addition of either the binding-deficient or the signaling-deficient transgene alone (Fig. 1). The key experiment, generation of mice coexpressing the Fig. 1. Schematic representation of the strategy followed by Rivero-Müller et al. to demonstrate in vivo dimerization of the lutropin receptor. (Upper) The various genotypes are represented. (Lower) Corresponding phenotypes. The binding-deficient or signaling-deficient receptors are schematized with mutations (yellow dot) affecting the ectodomain or the serpentine domain, respectively.