GPCR assembly and oligomerization

The successful reconstitution of adrenergic receptors from two fragments demonstrated that the integrity of the GPCR polypeptide chain is not required for proper receptor function (Kobilka etal. 1988). Based on these findings it has been speculated that GPCRs are composed of two or more independent folding domains. To test the hypothesis rhodopsin and the m3 muscarinic receptor were split in all three intracellular and extracellular loops. It was shown that except for a construct containing only TMD1, a significant portion of all N- and C-terminal receptor fragments studied was found to be inserted into the plasma membrane in the correct orientation even when expressed alone (Ridge et al. 1995; Schöneberg et al. 1995). Co-expression of some complementary receptor polypeptide pairs, generated by splitting GPCRs in their intra- and extracellular loops, resulted in receptors which were able to bind ligands and to mediate agonist-induced signal transduction (reviewed in Gudermann et al. 1997). It is noteworthy that all attempts to assemble functional receptor proteins from solubilized receptor fragments in vitro were unsuccessful (Schöneberg et al. 1997). This indicates that molecular chaperones that are likely to assist folding of the wild-type receptor protein may also play a role in facilitating complex formation. It has been shown that chaperone-dependent mechanisms are essential for proper folding of rhodopsin (Baker et al. 1994; Ferreira et al. 1996) and gonadotropin receptors (Rozell etal. 1998). The identity of the chaperones and the molecular mechanisms required for correct folding of other GPCRs remains to be elucidated.

The ability of functional complementation from receptor fragments is consistent with reports showing or suggesting that GPCRs can form dimers and oligomers. For several non-GPCR receptor families, such as receptor tyrosine kinases and kinase-associated cytokine receptors, agonist-induced receptor dimerization is required for initiating a signal transduction cascade. Initial evidence for GPCR dimerization came from crosslinking and photoaffinity labelling experiments with the GnRH receptor, LHR and muscarinic receptors (Conn et al. 1982; Avissar et al. 1983; Podesta et al. 1983). Numerous studies describing similar findings followed, but most reports of GPCR di- and oligomerization were based on co-immunoprecipitation studies. It has been argued that biochemical evidence from co-immunoprecipitation and Western blot experiments supporting the existence of GPCR oligomers is questionable, since solubilization of integral transmembrane proteins can cause artificial aggregation. However, as shown for epitope-tagged ß2 adrenergic, muscarinic and vasopressin receptors, the association is highly specific for a given receptor subtype giving rise only to homodimers (Hebert et al. 1998; Schulz et al. 2000a). In addition to investigations in transient expression systems, in vivo studies with D2 and D3 dopamine receptors (Nimchinsky et al. 1997; Zawarynski et al. 1998), somantostatin receptor type 5

(SSTR5)/D2 dopamine receptors (Rocheville etal. 2000a), and rhodopsin (Colley etal. 1995) suggest the coexistence ofreceptor monomers and oligomeric complexes under physiological circumstances.

One question that arises from these studies is as to whether GPCR dimers are pre-formed or are induced in the presence of the appropriate ligand. Most co-immunoprecipitation data suggest the existence of oligomeric receptor complexes under basal conditions. Examining the biological relevance of GPCR homodimerization in vivo, Bouvier et al. 1995 used a bioluminescence resonance energy transfer (BRET) technique to study receptor-receptor interactions (Angers etal. 2000). It was shown that ^-adrenergic receptors form constitutive homodimers that are expressed at the cell surface where they interact with agonists. Constitutive receptor association appears to be a general phenomenon since the yeast a-mating factor receptor forms dimers under basal conditions, as shown by a fluorescence resonance energy transfer (FRET) approach (Overton and Blumer 2000). On the other hand, there is also experimental evidence for an agonist driven oligomerization mechanism. Thus, the B2 bradykinin receptor and the CXCR4 receptor undergo receptor dimerization after ligand binding (AbdAlla etal. 1999; Vila-Coro etal. 1999).

There is growing evidence that GPCR not only exist in homodimeric structures but also in complexes formed by different GPCRs. Expression of the recombinant GABAb1 receptor in COS cells resulted in a significantly lower agonist affinity when compared with native receptors. Interestingly, co-expression of the GABAb1 receptor and the GABAb2 receptor, a recently cloned novel GABAb receptor subtype, in Xenopus oocytes and HEK-293 cells led to efficient coupling to G protein-regulated inward rectifier K+ channels (GIRKs) with an agonist potency in the same range as for GABAb receptors in neurons (Jones et al. 1998; White et al. 1998; Kaupmann et al. 1998). Encouraged by these studies, an ever growing number of heterodimeric complexes has been identified. For example, there is biochemical and pharmacological evidence that the k and 8 opioid receptors as well as ^ and 8 opioid receptors associate with each other. The complexes exhibit ligand binding and functional properties that are distinct from those of either receptor (Jordan and Devi 1999). Hetero-dimer formation was also observed for other receptor subtypes such as 5-HT1b/5-HT1d receptors and SSTR1/SSTR5 (Xie etal. 1999; Rocheville etal. 2000b). In a very recent study, hetero-oligomerization between the D2 dopamine receptor and SSTR5 was demonstrated, resulting in a novel receptor with enhanced functional activity (Rocheville etal. 2000a). The ability of GPCRs to heterodimerize provides a new mechanism by which a cell can fine-tune its responsiveness to a neurotransmitter via co-expression of distinct GPCR subtypes.

The molecular mechanisms and structural requirements which are responsible for GPCR oligomerization are only poorly understood. In the case of the mGluR5 (Romano etal. 1996) and the calcium-sensing receptor (Bai et al. 1998; Ward et al. 1998), which are members of Family 3, disulfide bonds between the extracellular portions are of critical importance for receptor dimerization. In a recent crystallographic study, the homodimeric structure of the extracellular ligand-binding domain of the mGluR1 was resolved, providing direct evidence for a disulfide bond-stabilized dimer (Kunishima et al. 2000). In contrast, it was demonstrated that mutant calcium-sensing receptors without extracellular cysteines form dimers on the cell surface to a similar extent as observed for wild-type receptors (Zhang et al. 2001). Interestingly, the GABAb2 receptor was initially discovered by a yeast two hybrid approach using the C terminus of the GABAb1 receptor for screening a human brain cDNA library. Heterodimer formation was assumed to be mediated via a coiled-coil structure of the C termini of the two receptors (White et al. 1998). It was found later that a

C-terminal retention motif RXR(R) is masked by GABAb receptor dimerization allowing the plasma membrane expression of the assembled complexes (Margeta-Mitrovic et al. 2000). However, association of both GABAb receptors was demonstrated even in the absence of their cytoplasmic C termini (Calver etal. 2001; Pagano et al. 2001).

Most studies agree that homo- and heterodimers found for rhodopsin-like GPCRs represent non-covalent complexes. Thus, two structural models of dimer formation have been proposed for Family 1 GPCRs (Gouldson etal. 1998). In one dimeric structure, referred to as 'contact dimer', two tightly packed bundles of seven TMDs are positioned next to each other. The contact interface between the two monomeric receptors is assumed to be located between the lipid-orientated transmembrane receptor portions (Fig. 1.4). The so-called 'domain-swapped dimer' has been proposed to explain the reconstitution phenomenon observed with truncated and chimeric GPCRs (Maggio et al. 1993; Schulz et al. 2000a). In this dimer structure, the two receptor molecules fold around a hydrophilic interface by exchanging their N-terminal (TMDs1-5) and C-terminal (TMDs6-7) folding domains (Fig. 1.4). Attempting both hypothetical dimer structures data with the AVPR2 and D2 dopamine receptors strongly support an oligomeric structure in which Family 1 GPCRs form contact oligomers by lateral interaction rather than by a domain-swapping mechanism (Schulz et al. 2000a; Lee et al. 2000). High resolution X-ray structure determinations of two heptahelical membrane proteins, the bacteriorhodopsin and the halorhodopsin, clearly show that both proteins assemble to trimers (Luecke etal. 1999; Kolbe etal. 2000). The proton pump bacteriorhodopsin shares structural similarities with the GPCR family including the assembly from multiple independent folding units. In the trimeric structure found in bacteriorhodopsin and halorhodopsin crystals, TMDs2-4 of the three molecules face each other forming an inner circle of TMDs. Structural data did not provide any support for a domain-swapping mechanism of oligomer-ization. Similarly, other polytopic membrane-spanning proteins which homo-oligomerize

Fig. 1.4 Hypothetical structures of GPCR dimers. TMDs (numbered from 1 to 7) of GPCRs form a ring-like structure in a counter-clock wise fashion as viewed from extracellular. GPCRs are composed of at least two independent folding domains (TMDs1-5 and TMDs6-7) which are connected by the i3 loop. Accumulating evidence suggests that wild-type GPCRs can exist in dimeric complexes, and two structural models of dimer formation have been suggested (Gouldson etal. 1998). The contact interface of so-called 'swapped dimers' is recruited from the rearrangement of two independent folding domains of the individual receptor monomers. The ring-like TMD arrangement is still retained by the complementary exchange of the two folding domains. In contact dimers, a lateral interaction of the individual receptor molecules is assumed.

Fig. 1.4 Hypothetical structures of GPCR dimers. TMDs (numbered from 1 to 7) of GPCRs form a ring-like structure in a counter-clock wise fashion as viewed from extracellular. GPCRs are composed of at least two independent folding domains (TMDs1-5 and TMDs6-7) which are connected by the i3 loop. Accumulating evidence suggests that wild-type GPCRs can exist in dimeric complexes, and two structural models of dimer formation have been suggested (Gouldson etal. 1998). The contact interface of so-called 'swapped dimers' is recruited from the rearrangement of two independent folding domains of the individual receptor monomers. The ring-like TMD arrangement is still retained by the complementary exchange of the two folding domains. In contact dimers, a lateral interaction of the individual receptor molecules is assumed.

in order to build a functional complex, such as aquaporins, assemble via lateral interaction (Walz etal. 1997).

Taken together, current data strongly support oligomeric GPCR structures. Functional studies with mutant GPCRs provided evidence that oligomerization occurs by lateral interaction rather than by a domain-swapped mechanism. There is growing support for the idea that GPCR dimerization has consequences for physiologic receptor functions such as formation of receptor 'subtypes' with new ligand binding or signalling abilities.

0 0

Post a comment