Nearly two decades after the cloning of the first GPCRs, there are still many open questions relating to the mechanisms of GPCR/G protein interaction and the molecular elements determining G protein coupling specificity. Over the past few years, an increasing number of GPCRs with a broad G protein coupling profile has been identified (Gudermann etal. 1997). Structural elements determining signalling specificity are located in both the G protein and the receptor. Numerous in vitro mutagenesis studies have been performed with G protein a subunits to understand how coupling selectivity is achieved (see Chapter 4). The C terminus as well as the N terminus of the a subunit make important contributions to appropriate receptor/G protein recognition (Wess 1998).
The exact nature of G protein interaction sites within the receptor is currently unknown and may vary between the different GPCRs and G proteins. It is assumed that not only the intracellular loops but also the cytoplasmic sides of the TMDs participate in GPCR/G protein coupling. Indeed, peptides derived from the i3 loop/TMD6 junction can activate G proteins (Abell and Segaloff 1997; Varrault et al. 1994). Likewise, site-directed mutagenesis studies with GPCRs examined the structural elements that participate in G protein interactions and that determine the coupling profile of a given receptor. For example, the i1 loop of the formyl peptide receptor (Amatruda et al. 1995) and the cholecystokinin CCKa receptor (Wu et al. 1997), the i2 loop of the V1a vasopressin and GABAb receptors (Liu and Wess 1996; Robbins et al. 2001), and the i3 loop of the endothelin ETb receptor (Takagi et al. 1995) have been demonstrated to participate in G protein activation. Taking advantage of chimeric GPCRs designed between structurally related receptor subtypes that are clearly distinguishable with regard to their signalling abilities, studies on muscarinic (Blüml et al. 1994; Blin et al. 1995) and vasopressin receptors (Erlenbach and Wess 1998) disclosed the importance of several distinct residues for selective G protein recognition. However, the established view of the importance of the intracellular loops in G protein interaction and specificity is challenged by recent studies with the TSHR and LHR. Large deletions or alanine replacement of most amino acid residues in the i3 loops did not abolish signal transduction, excluding a substantial participation of the i3 loop in G protein recognition at least in glycoprotein hormone receptors (Wonerow et al. 1998; Schulz et al. 1999, 2000 b). In the m3 muscarinic receptor, a segment of 112 amino acids (central portion of the i3 loop) can be deleted without loss of receptor function (Schöneberg et al. 1995). Similarly, systematic reduction of the length of the i3 loop in the tachykinin NK-1 receptor revealed that most of the loop sequence can be substituted or even deleted without affecting ligand affinity or signal transduction (Nielsen et al. 1998). Further support for the notion that most of the i3 loop sequences are dispensable for G protein coupling comes from structural comparison of GPCR loop sequences showing that especially the i3 loop varies extremely in length and that there is no obvious sequence homology between GPCRs of similar coupling profiles.
Studies with receptor peptides that are able to activate G proteins directly indicate that the cytoplasmic extensions of the TMDs probably provide the surface for G protein interaction rather than the loops themselves (Abell and Segaloff 1997). Besides from serving simply as connectors between the TMDs, the intracellular loops may be involved in additional GPCR functions that have been discovered recently.
The ability of a GPCR to couple to more than one G protein subfamily can be conceived as a loss of specificity due to the absence of inhibitory determinants or as a gain of specific contact sites within the receptor molecule. Mutational analyses of GPCRs interacting with more than one G protein family provide evidence for both concepts. It has been demonstrated that point mutations can selectively abolish receptor coupling to one G protein subfamily (Surprenant et al. 1992; Biebermann et al. 1998). On the other hand, the coupling profile of a GPCR can be extended by mutational changes. Concomitantly with Gs activation, the LHR mediates fairly modest agonist-induced phosphoinositide breakdown via Gi recruitment. It was observed that several LHR mutations at the very N-terminal end of TMD6 profoundly enhanced agonist-induced IP accumulation, most likely via Gq/11 activation (Schulz et al. 1999). Two general mechanisms explaining multiple coupling events have been suggested—a parallel and a sequential G protein activation model. Lefkowitz and colleagues offered experimental evidence for a sequential G protein activation mechanism (Daaka et al. 1997). Agonist-induced activation of the ^-adrenergic receptor results in cAMP formation via the Gs/adenylyl cyclase pathway followed by cAMP-dependent protein kinase-mediated receptor phosphorylation. Receptor phosphorylation represents a crucial molecular switch mechanism to allow for Gi-mediated ERK activation. These findings are of interest since several primarily Gs-coupled receptors are also capable of activating Gj. However, in the case of the LHR both signal transduction pathways, cAMP and IP formation, can be functionally separated by point mutations as shown for K583 in the e3 loop. Whereas ligand binding and agonist-induced phosphoinositide breakdown remained unaltered, cAMP formation was found to be abrogated (Gilchrist et al. 1996). These results show that ligand-independent adenylyl cyclase stimulation does not always represent a prerequisite for efficient coupling to the phospholipase C signalling pathway. Both signalling events appear to be independent, favouring a model of parallel activation of G proteins. These findings are consistent with the concept that GPCRs can exist in at least two distinct active conformations, R* and R**, which differ in their G protein coupling pattern. Based on this notion, it is also conceivable that different agonists can stabilize distinct ternary complexes. One may speculate that such pathway-selective agonists may represent the protagonists of a new class of therapeutic agents.
Was this article helpful?