点击显示 收起
Abstract |
---|
Top Abstract Introduction Broad Evidence for GPCR... Contributions of GPCR... Minimal Model for GPCR... Functional Consequences of GPCR... Importance for Cardiovascular... References |
---|
Key Words: G protein–coupled receptors oligomerization heterodimerization signal transduction G protein activation
Introduction |
---|
Top Abstract Introduction Broad Evidence for GPCR... Contributions of GPCR... Minimal Model for GPCR... Functional Consequences of GPCR... Importance for Cardiovascular... References |
---|
There is now a large and diverse body of evidence (recounted in many recent reviews2–7) that suggests that GPCRs do indeed function as dimers (or higher order oligomers), and that dimerization can occur among identical GPCRs, close family members, or GPCRs that are in distinct families. GPCR dimerization has documented effects on ligand binding, receptor activation, desensitization and trafficking, as well as receptor signaling. Mechanistic interpretation of these changes in GPCR function requires refinement of models for GPCR–G protein interactions. Although the genome provides a rich diversity of GPCRs, the potential for GPCR heterodimerization greatly increases the number of phenotypes that can be obtained from this already vast family of receptors, having implications for both cardiovascular physiology and the specificity of pharmacological interventions.
Broad Evidence for GPCR Oligomerization |
---|
Top Abstract Introduction Broad Evidence for GPCR... Cntributions of GPCR... Minimal Model for GPCR... Functional Consequences of GPCR... Importance for Cardiovascular... References |
---|
Contributions of GPCR Oligomerization to GPCR–G Protein Interactions |
---|
Top Abstract Introduction Broad Evidence for GPCR... Contributions of GPCR... Minimal Model for GPCR... Functional Consequences of GPCR... Importance for Cardiovascular... References |
---|
The traditional view of GPCR/G protein coupling incorporates a monomeric GPCR interacting through specific intracellular domains (typically i2, i3, and/or proximal carboxyl terminal regions) with a single heterotrimeric G protein; agonist-induced conformational changes in the receptor, propagated through the transmembrane domain, ultimately result in exposure of critical residues at the GPCR/G protein interface that promote interaction with and activation of G proteins.10,11 This "minimal model" is supported by countless mutagenesis studies of GPCRs in which the critical intracellular determinants for G protein activation have been empirically identified. Recent bioinformatics approaches have used this wealth of experimental data to attempt predictions of G protein coupling specificities from GPCR sequences.12 From both experimental studies and bioinformatics, it is clear that the GPCR–G protein interaction surface is complex, and incorporates distinct regions of the GPCR, including the i2 loop, the N- and C-terminal portions of the i3 loop, and/or residues within the proximal portion of the carboxyl terminus (the so-called i4 loop anchored by lipid-modified cysteine residues). The potential for GPCR dimerization/oligomerization raises critical questions that impact on this minimal model for GPCR–G protein interactions, including (1) what is the nature of the GPCR dimer interface; (2) do allosteric interactions occur between partner GPCRs; (3) does agonist-mediated dimerization occur, and does it regulate the interaction of GPCRs with G proteins; and finally, (4) what is the stoichiometry of GPCR/G protein interactions?
What Is the Nature of the GPCR Dimer Interface?
The nature of the dimer interface specifies not only which GPCRs can exhibit productive interactions, but influences models for potential allosteric interactions between dimer partners. Although many GPCRs have been shown to participate in homodimerization, heterodimerization is more variable, with some GPCRs exhibiting broad promiscuity,13 and others exhibiting a high degree of selectivity.14 Evidence to date suggests that there are distinctive dimerization interfaces or domains both within the transmembrane helices15–18 as well as at either the extracellular amino terminus15,19–21 or the intracellular carboxyl terminus,22 depending on the GPCR. Regardless of the domain responsible for initiation of dimerization, proximity of the transmembrane domains of the GPCR partners is assured. Family C GPCRs, CaR19 and metabotropic glutamate receptors mGluRs,23 are stabilized by interdimer disulfide bonds localized at the extracellular, amino-terminal domain, whereas GABABRs interact via a coiled-coil domain at the carboxyl terminus.24 m3 muscarinic receptors are also stabilized by disulfide bonds.25 The majority of GPCRs, however, are stabilized as dimers/oligomers via noncovalent interactions. Two modes of interaction have been described (illustrated schematically in ), namely (1) contact dimerization, in which the relevant helix (helices) from one monomer contact(s) partner(s) in the other monomer, stabilizing the dimer pair; and (2) domain swapping, in which several helices from each receptor are "swapped" in the dimer, such that the functional monomer within the dimer contains helices contributed by both receptors (indicated by the distinct colors in ). Select examples of both types of interactions are available among GPCRs (see later), although the nature of the dimer interface for most GPCRs is unknown, and therefore the relative prevalence of the two modes remains to be established.
|
Contact site(s) have been identified for a very few GPCRs, and to date, only for homodimers. Experimental approaches include either (1) the use of synthetic peptides corresponding to various transmembrane helices to determine effects on receptor dimerization and/or activation, or (2) disulfide "trapping" in which cysteine residues are incorporated into suspected partner helices, and the propensity for disulfide bond formation is assessed. Synthetic peptides related to TM6 block both dimerization and activation of the ß2AR,26 whereas TM4 has been shown to mediate homodimerization of D2 receptors by cysteine scanning mutagenesis and chemical crosslinking27 and C5a receptors by disulfide trapping.18 TMs 1 and 2 were shown to be involved in yeast -factor receptor dimerization by FRET analysis.15 A computational subtractive correlation method (based on the rhodopsin crystal structure, solvent accessibility, and location of residues on outer faces of helix bundles) has been applied to opioid receptor homodimers and heterodimers, and results (not yet experimentally verified) indicate a high degree of variability in interaction domains.28,29 With respect to homodimers, TM4:TM4, TM5:TM5, or TM4:TM5 were predicted to be likely interfaces for OR, TM1:TM1 was most likely for µOR, whereas TM5:TM5 was most likely for OR.29 The model predicted that OR–µOR heterodimers were most likely stabilized by association of TM 4, 5, or 6 of the OR with TM1 of µOR, and accurately predicted no interaction(s) between µOR and OR.28 Evolutionary trace30 and lipid-facing correlation models31 have also identified the most likely interfaces for representative GPCRs, but so far there has been little experimental data to verify the predictions on a GPCR family-wide scale. As additional GPCR pairings are established by experimental means, these models can be put to the test and refined. Because dimers as well as higher order oligomers have been characterized among GPCRs, it is highly likely that multiple forms of interactions will contribute to the stabilization of the final functional complex. In this regard, it is interesting to note the recent demonstration that mGluR1 receptors, which are constitutive disulfide-linked homodimers,23 can be specifically coimmunoprecipitated with adenosine A1 receptors from both HEK-293 cells and rat brain.32
The domain-swapping model was first suggested by the functional rescue upon coexpression of m3/2c and 2c/m3 chimeric receptors, because binding of both m3 and 2c agonists was restored.33 Although not established experimentally, domain swapping has also been invoked as an explanation for the altered ligand binding specificity of coexpressed and opioid receptors28,34 because altered affinities (when compared with homodimers) can result when a functional binding unit comprises helices from two distinct receptor types (as illustrated by the color coding in ). It is theoretically possible to test domain swapping as the explanation for the altered agonist affinities of – opioid receptor heterodimers by generating chimeras between these two receptors containing helices 1 to 5 from one subtype and 6 to 7 from the other. Coexpression of these chimeras [(1–5)/(6–7) and (1–5)/(6–7)] should generate agonist affinities comparable to those seen with expressed – or – homodimers, because domain swapping should generate (1–5)/(6–7) plus (1–5)/(6–7) binding units.
"Rescue" of receptor activity on coexpression of complementary inactivating mutants has also been suggested to result from domain swapping; in this case, one of the two "swapped" receptors contains the defective portions of both receptor mutants (and is nonfunctional), whereas the partner receptor is "reconstituted" with wild-type helices from each contributing receptor. For example, functional cooperation between partners in the ß2AR homodimer has been inferred because nonpalmitylated, constitutively desensitized mutant ß2AR (C341G–ß2AR) can be rescued by coexpression with wild-type receptor.35 The mutant and wild-type ß2ARs assemble into functional heterodimers (as demonstrated by coimmunoprecipitation of epitope-tagged receptors), and have an activity and desensitization rate equivalent to wild-type homodimers.35 Functional rescue was also observed when wild-type ß2AR was coexpressed with a mutant lacking protein kinase A phosphorylation sites.35
Although there is clear evidence for domain swapping in specific cases (see review30), systematic study of several other GPCRs has revealed no evidence for this mechanism.36,37 Thus, it is likely that domain swapping may represent only one of several mechanisms utilized by particular GPCRs in dimer/oligomer formation and stabilization. It has been suggested that contact dimers and domain-swapped dimers have equivalent abilities to signal to G proteins, and the ability of a particular GPCR pair to form a functional dimer via either interaction depends on the relative energetics of the two possible pairings.30
Do Allosteric Interactions Occur Between Partner GPCRs?
Binding of agonist to a GPCR causes conformational changes within the core of the helical transmembrane domain that are transmitted to the intracellular loops, resulting in G protein activation.10,38 The presence of GPCRs in dimeric or oligomeric complexes makes allosteric interactions between the monomer partners within the dimer possible. The most straightforward example of allosteric interactions between dimer partners is the GABAB receptor obligate heterodimer: agonist binding occurs only at the GABAB1R amino terminal binding domain,39 and G protein coupling has been mapped to the GABAB2R intracellular domain.40 Mutations in the GABAB2R agonist binding domain do not affect receptor signaling39 nor do mutations within the intracellular loops of the GABAB1R40–42inhibit G protein activation. This segregation of agonist binding on one monomer and critical G protein interactions on the other requires allosteric interactions between monomers at some level, either between agonist binding domains or between transmembrane helical domains. Interestingly, a chimera containing the functional domains of both receptors, ie, GABAB1R extracellular domain plus GABAB2R helical plus carboxyl terminal domains is not functional, whereas coexpression of the GABAB1R/GABAB2R plus GABAB2R/GABAB1R chimeras restores activity.43 The requirement for allosteric coupling between subunits of the functional dimer is further supported by the ability of the GABAB2R to increase the affinity of the GABAB1R subunit for agonist, and to stabilize the closed (active) state of its agonist binding domain.39,43 Likewise, although GABAB1R does not productively interact with G proteins, the presence of the transmembrane plus intracellular domains of GABAB1R is necessary for efficient GABAB2R-mediated G protein activation.40,43
GABABR heterodimers also provide a provisional answer to a fundamental question regarding whether both receptors in a dimer must bind agonist for transduction, ie, G protein activation, to occur. With respect to GABABRs, only one receptor in the pair is competent to bind agonist39 (see review44) and therefore undergoes a conformational change to the closed state of the venus fly-trap module (extracellular agonist binding domain of family C members). Similar results have been derived from a crystal structure analysis of the dimerized extracellular venus fly trap module (ligand binding domain) of a related family C receptor, the metabotropic glutamate receptor, mGluR1.21,23 Only one of the two binding modules "closes" on interaction with its agonist glutamate, triggering a global conformational change in both disulfide bond-linked partners,21,23,44 and presumably causing propagated conformational changes to the respective transmembrane domains. Allosteric interactions can also be mediated by domains other than the transmembrane helices, as has recently been demonstrated for the membrane-proximal portion of the carboxyl terminus of calcium sensing receptors,45,46 which can mediate cooperativity with respect to G protein activation.
As described above, allosteric interactions have been identified for the major members of family C GPCRs, ie, GABABRs, mGluRs, and CaR. Does this reflect a general property of GPCRs in dimers/oligomers, or does this reflect a specific property of a highly restrictive group whose members function as obligate dimers? To answer this question, we need look no further than opioid receptors, members of family A (rhodopsin-like) GPCRs, which have been extensively characterized with respect to both homo- and heterodimerization. All three opioid receptors (, , and µORs) have been shown to undergo homodimerization, and both – and –µ heterodimers have been demonstrated by coimmunoprecipitation34,47 or BRET,48 whereas –µ heterodimers have not been observed.34,49
Heterodimerization of opioid receptor subtypes has been evoked to explain the discrepancy between the number of known opioid receptors genes (, , and µ) and the pharmacologically distinct receptor subtypes characterized in vivo (1, 2, 1, 2, 3, µ1, and µ2).49,50 As discussed in the previous sections, both contact28,29 and domain-swapping28,34 models have been proposed to explain the changes in receptor function within the heterodimers, although neither model has been experimentally verified. Regardless of which mechanism is responsible for the interactions between monomers, the heterodimers formed by and opioid receptors have pharmacological properties resembling the 2 receptor subtype characterized in vivo, ie, distinct from homodimers of either subunit. Not only does the – heterodimer exhibit decreased affinities for either - or -selective agonists and antagonists, but the heterodimer synergistically binds certain partially selective agonists with high affinity. Synergy with respect to agonist binding is also reflected in the activation of signaling pathways, both inhibition of adenylyl cyclase and phosphorylation of MAPK.34 Similarly, µ– heterodimers exhibit reduced affinities for either µ- or -selective agonists, but synergistic interactions when both µ- and -selective agonists are applied simultaneously51 including enhanced affinities for endomorphin-1 and Leu-enkephalin,47 suggesting alterations in the agonist binding site(s). Signaling via µ– heterodimers is insensitive to pertussis toxin, suggesting a shift in G protein preference from Gi to pertussis toxin–resistant subtypes.47 Although all of these heterodimer effects on agonist binding and/or signaling can be interpreted as a result of domain swapping, they can also potentially reflect allosteric interactions between receptors within the homo- and heterodimers that are dependent on the location and/or specificity of the dimer interface.
The occurrence of allosteric interactions between partner GPCRs may depend critically on the identity of the receptors. Here again, opioid receptors represent a well-studied case, because heterodimerization among opioid receptor subtypes evokes a variety of changes in agonist binding and/or G protein specificity that can be taken as a reflection of allosteric interactions. In sharp contrast, heterodimerization between opioid receptor subtypes and a variety of other family A GPCRs has also been noted, although none of the binding or signaling properties of the opioid receptor are altered, rather desensitization and/or internalization are affected. For example, both and µ opioid receptors are capable of forming heterodimers with ß2AR. OR–ß2AR heterodimers exhibit unaltered ligand binding properties and signaling, but activating either partner results in desensitization and internalization of the heterodimer complex.52 OR homodimers do not undergo agonist-promoted endocytosis, and this phenotype is also dominant in OR–ß2AR heterodimers, which do not endocytose in response to either agonist, suggesting that heterodimerization in this case obscures domains critical to the process of ß2AR internalization. Furthermore, ß2AR-mediated phosphorylation of MAPK was reduced in cells expressing the OR–ß2AR heterodimer.52 µOR–sst2A heterodimers displayed no alterations in ligand binding or signaling, but exhibited crossphosphorylation and crossdesensitization in response to agonists for either receptor.53 Given the potential for distinct dimer interfaces between homo- and heterodimers of opioid receptors and their various partners, it is likely that some associations between GPCRs may represent a form of specific clustering or aggregation, which permits participants to be coimmunoprecipitated and cotrafficked (desensitized and/or internalized), but does not permit allosteric interactions between partners that could potentially alter agonist specificities or G protein signaling. A more intimate form of interaction between receptor partners, ie, what has been referred to as dimerization, may be required for allosteric interactions between receptor monomers. It should be stated that the dimer stoichiometry is a minimal one, and that higher order oligomeric complexes with distinct associations may occur, ie, (µOR:µOR):(sst2a:sst2a) tetramers in which distinct GPCR homodimers are associated in a cluster or aggregate would exhibit the same apparent stoichiometry via immunoprecipitation as the dimer µOR:sst2a, although the functional consequences of the two types of association may be different. In this regard, it is interesting that recent studies have shown that GPCRs may associate with and traffic in oligomeric complexes with non-GPCR proteins. D1 dopamine receptors colocalize and interact (by BRET criteria) with NMDA NR1 subunits, and this interaction alters D1 dopamine receptor targeting to the plasma membrane as well as dopamine-induced sequestration.54 Another study demonstrated colocalization and direct interactions between mGluR1 and voltage-sensitive calcium channels (Cav1.2) mediated by carboxyl terminal sites on the two proteins, resulting in enhanced regulation of calcium signaling in dendrites.55 It seems likely that GPCRs will be found to undergo not only true dimerization (with attendant allosteric interactions), but also oligomerization, which alters targeting and/or provides the enhanced signaling and/or specificity normally thought to be promoted exclusively by scaffold proteins.
Does Agonist-Mediated Dimerization/Oligomerization Occur, and Does It Regulate the Interaction of GPCRs With G Proteins?
Agonist binding to some GPCRs induces or stabilizes the dimerized state, which is the form of the receptor that productively interacts with G proteins. For other receptors, notably but not restricted to family C members, dimerization is constitutive and occurs in the endoplasmic reticulum. Examples of both types of associations are available, but the prevalence of each type of association within the GPCR superfamily is not yet known.
ßARs have been extensively characterized both biochemically and pharmacologically, and since the early 1980s, have been shown to purify as monomers or dimers (see review56). ß2AR dimerization has been studied by both coimmunoprecipitation of differentially epitope-tagged receptors26 and by BRET.57,58 By functional criteria, agonists have been shown to stabilize the dimeric state, whereas inverse agonists promote or stabilize the monomer.26 The identity of transmembrane helix 6 as the dimerization interface and the importance of dimerization in ß2AR signaling was underscored by inhibition of both dimerization and stimulation of adenylyl cyclase by addition of a peptide derived from transmembrane helix 6 (residues 276 to 296).26 Thus, dimerization is required for efficient ß2AR signaling, and can be promoted or inhibited by ligands. Agonist-induced oligomerization has also been reported for somatostatin receptors,59,60 thyrotropin-releasing hormone receptors and gonadotropin-releasing hormone receptors,14 dopamine receptors,61 and heterodimers of adenosine A1 and P2Y1 receptors,62 whereas agonist-induced monomerization has been reported for cholecystokinin receptors63 and OR.64 Agonist-promoted changes in dimerization state may not be universal, and there are many examples of constitutive dimerization, including representative members of family C GPCRs (CaR, mGluRs, and GABABRs) (see review44), vasopressin V1a and V2a,65 oxytocin,65 m3 muscarinic receptors,25 melatonin receptors,66 or neuropeptide Y receptors.67 There are, however, divergent results for a number of GPCRs depending on the method used to determine dimerization. For example, by functional assay, ß2 receptor dimerization has been shown to depend on agonist,26 whereas via BRET, ß1 or ß2 homodimers and ß1–ß2 heterodimers have been shown to be constitutive.57,58 Similar divergent results have been shown for OR (function and coimmunoprecipitation34 versus BRET,48,68 discussed in Levac et al49), and CCR5 receptors (divergent results by coimmunoprecipitation69,70 and BRET71). There are several likely sources of discrepancy between the two methods. First, the relative sensitivity of the two methods to the affinities between partner GPCRs, with coimmunoprecipitation requiring generally higher affinity interactions between partners to survive isolation procedures, whereas BRET requires affinities sufficient to stabilize proximity in vivo. Second, FRET/BRET depends on the distance between the two fluorescent tags within the complex, which may change as a result of ligand or regulator binding. Whether a difference in the strength of association between monomer partners is observed on agonist binding may depend on the magnitude of the conformational changes that the receptor monomers undergo, and the attendant changes in the distance between fluorescent tags. Differences in BRET resulting from agonist-induced conformational changes within constitutive complexes14,66,72 have been described, whereas for other receptors, no change in BRET on agonist binding was observed.71 As more GPCRs are characterized with the explicit goal of understanding the contribution(s) of dimerization/oligomerization to their function, criteria for establishing constitutive versus regulated interactions among partners will of necessity be established.
What Is the Stoichiometry of GPCR–G Protein Interactions?
A critical question in GPCR signaling is the stoichiometry of GPCR–G protein interactions, although this question has been difficult to address because G protein–mediated signaling incorporates a catalytic component, ie, a single activated receptor molecule may serially activate a number of G proteins before agonist unbinding. Speculation that receptor dimers interacted with a single G protein heterotrimer began when the first crystal structure of a heterotrimeric G protein was solved: functional studies have mapped the many receptor interaction sites on G protein and ß subunits, and the data were not consistent with models incorporating the structure of a monomer of rhodopsin as the prototypical GPCR and the known structure of the G protein heterotrimer.73–75 There are now several independent lines of evidence that suggest that G protein activation may require a dimeric GPCR interacting with a single heterotrimeric G protein.
Recent studies on mammalian rhodopsins provides the most compelling physical evidence for an R2:G(ß) stoichiometry. First, the x-ray crystal structure of bovine rhodopsin indicates that the cytoplasmic "footprint" of the transmembrane domain may be no more than 40 Å across,76 which cannot account for interactions of the same monomer with both the transducin subunit, and parts of the subunit (for discussion, see Marshall77). Second, a recent study utilizing atomic force microscopy found dimers to be the predominant form of mouse photoreceptor rhodopsin in isolated rod outer segment disk membranes, with the dimer interface at helices 4 and 5.78 Finally, a number of functional studies have characterized cooperativity with respect to binding of transducin to rhodopsin, which could be explained by an R2:G(ß) stoichiometry.79,80
Strong biochemical evidence for an R2:G(ß) stoichiometry comes from a recent study on leukotriene B4 receptors.81,82Optimization of methods for overexpression, recovery, and refolding of LTB4R from bacterial inclusion bodies produced a soluble receptor that was appropriately refolded and bound both LTB4 and structurally related antagonists with native affinities.81 The stoichiometry of LTB4R:G(ß) protein interactions was determined by two independent methods; both chemical crosslinking in the presence of receptor, heterotrimeric G protein (Gi2ß12), and agonist or solution phase neutron scattering indicated a pentameric stoichiometry, ie, R2:G(ß).82 Conditions predicted to lead to low affinity of the receptor for G protein (GTPS or antagonist) promoted the appearance of the monomeric species of receptor, leading to the minimal model82 (depicted generically in ), ie, agonist binding to a monomeric receptor promotes receptor dimerization, followed by interaction with heterotrimeric G protein (GDP-bound), ultimately resulting in GDP release, GTP binding, and dissociation of activated G protein subunits from the dimeric receptor.
|
A different type of evidence supporting GPCR dimers as the minimal functional unit comes from studies with membrane-tethered peptides, termed pepducins.83,84 Peptides corresponding to the i3 loop of protease-activated receptors or melanocortin receptors that were membrane tethered (by either hydrophobic residues at the amino and carboxyl termini or by addition of a palmitate group) activated receptors in the absence of agonist.84 Results from the studies were consistent with the tethered peptide and the monomeric receptor interacting to reconstitute a complete "dimeric" binding site for G proteins, presumably in the active conformation, permitting signaling in the absence of agonist.83 Shortened peptides lacking the ability to induce receptor activation acted as antagonists, preventing agonist-mediated receptor activation.84
Finally, heterodimerization of a variety of GPCRs alters the G protein specificity of signaling (). Although a shift in G protein specificity can be explained by a number of mechanisms, examination of suggests that interaction of a single G protein with a receptor dimer can account for this data. That is, if the G protein interacts with the i2 or i3 loop(s), the composite heterodimer G protein binding site contains contributions from both GPCR partners, potentially altering the electrostatic binding surface and therefore G protein specificity.11,85
|
Minimal Model for GPCR Dimer–Mediated Activation of G Proteins |
---|
Top Abstract Introduction Broad Evidence for GPCR... Contributions of GPCR... Minimal Model for GPCR... Functional Consequences of GPCR... Importance for Cardiovascular... References |
---|
Functional Consequences of GPCR Oligomerization |
---|
Top Abstract Introduction Broad Evidence for GPCR... Contributions of GPCR... Minimal Model for GPCR... Functional Consequences of GPCR... Importance for Cardiovascular... References |
---|
Effects on GPCR Pharmacology
Alterations in the agonist binding sites of GPCRs may accompany dimerization as a result of allosteric interactions or domain swapping between partners ). The general strategy of using a cell line stably expressing the GPCR of interest in screening for potent/specific pharmacophores can result in identification of compounds that are not as effective in vivo. Given the propensity of many GPCRs to heterodimerize, with attendant changes in binding site affinities, the promiscuity of some GPCRs may generate a heterogeneous receptor population that is cell-type specific. Iterative strategies in which likely GPCR partners in the tissue of interest are coexpressed with the target GPCR may be required to identify physiologically significant ligands. The ability of a partner to modify the binding affinity of a GPCR can potentially be used to clinical advantage, allowing development of agents specific for a given heterodimer pair to regulate a signaling pathway present in only a subset of tissues. For GPCRs that undergo agonist-mediated dimerization, stabilization of the dimer by bifunctional agonists could lead to greater specificity and efficacy of signaling.88
Effects on GPCR Signaling
GPCRs have been extensively characterized with respect to both ligand binding and activation of various signaling pathways. Widespread demonstrations of GPCR oligomerization raise interesting questions regarding which aspects of already well-characterized GPCR-mediated functions are potentially due to heterodimerization in vivo. Heterodimerization potentially alters G protein specificity, coupling to signaling pathways, and may also attenuate signaling (). It has been suggested that some GPCR heterodimers act as "coincidence detectors," synergistically increasing signaling when both agonists are present.3 Conversely, partners within heterodimers can negatively interact, attenuating signaling relative to the respective homodimers.89 ß-Adrenergic receptors are a well-studied example relevant to the cardiovascular system. When expressed in HEK-293 cells, ß1–ß2AR heterodimers display hybrid signaling compared with the respective homodimers. ß1–ß2AR heterodimers stimulate adenylyl cyclase, to levels achieved by either ß1 or ß2AR homodimers.90 In contrast, ß2AR homodimers robustly stimulate ERK1/2 phosphorylation, whereas ß1AR homodimers have no effect on this signaling pathway. ß1–ß2AR heterodimers are not able to activate the MAPK signaling cascade.93 ß2AR internalization is a prerequisite for activation of the MAPK cascade,93 and heterodimerization with ß1AR limits ß2AR interactions with arrestin and subsequent internalization. Cardiac myocytes express both ß1- and ß2AR, and selective up- and/or downregulation of receptor subtype expression accompanies a number of disease states. For example, during development of heart failure, ß1AR expression is reduced whereas ß2AR expression is increased, allowing selective increases in ERK1/2 signaling (discussed in Lavoie et al90). In contrast to HEK-293 cells where coexpression of ß1AR and ß2AR results in attenuation of ERK1/2 signaling,90 neonatal rat cardiomyocytes exhibit ß1AR and ß2AR signaling through adenylyl cyclase, as well as activation of ERK1/2,91 suggesting that in vivo segregation of receptors to distinct microdomains, eg, caveolae, may limit the extent of heterodimerization.92 Heterodimerization of ßARs has also been demonstrated with opioid receptor subtypes ( and )52 and 2A receptors.93 A particular GPCR may thus exhibit variable coupling to signaling pathways as a result of heterodimerization with diverse GPCR partners and/or as a result of subcellular segregation of receptors, which may also be specified by the partner GPCR.
Contributions to GPCR Crosstalk
It has become increasingly clear that GPCR signaling does not result from sequential activation of a linear pathway of proteins/enzymes, but rather results from the complex interactions of multiple, branched signaling pathways, ie, signaling networks. Positive and negative feedback in such pathways, activated by multiple GPCRs, is termed crosstalk (see reviews94–96). Unrecognized heterodimerization between distinct GPCRs can generate signaling phenotypes that may be interpreted as signaling pathway crosstalk. Crosstalk between signaling pathways regulating the cardiovascular system is often observed, and in general, is thought to result from the intersection of signaling pathways at multiple levels within signaling networks (see comprehensive review97). It is highly likely that some well characterized examples of receptor crosstalk in the cardiovascular system will be the result of direct heterodimerization between the GPCRs involved, eg, the well-characterized crosstalk between OR and ß1AR, which results in attenuation of myocardial responses to stress,98,99 may very well be the result of receptor heterodimerization.
A broad illustration of the potential contributions of GPCR heterodimerization in receptor crosstalk is available from recent studies on the angiotensin receptor. Two subtypes of angiotensin receptors have been identified, type 1 (AT1) and type 2 (AT2), each with distinct signaling properties.100 Negative crosstalk has been observed when both AT1 and AT2 receptors are expressed in vascular smooth muscle cells, with AT2 receptors inhibiting AT1-mediated cell growth.96,100 Most of the physiological effects of angiotensin II are mediated by AT1 receptors (AT1A and AT1B). AT1 receptors are noncovalent dimers101; allosteric interactions between the monomers within the homodimer has been demonstrated by coexpression studies with deficient mutants, which restores a normal binding site and signaling.102 Whereas G protein coupling and signal transduction via AT1 receptors is well characterized, less is known about AT2 receptors, although coexpression with AT1 receptors often attenuates AT1 responses.96,100 Recent studies have demonstrated heterodimerization of AT1 and AT2 receptors both in vivo103–105 and in vitro,104,105with attendant inhibition of AT1 receptor signaling, suggesting that AT2 receptors act as a dominant-negative or antagonist of AT1 receptor function,104 ie, the negative crosstalk defined in vivo may result directly from heterodimerization of AT1 and AT2 receptors. Amino acids within the AT2R third intracellular loop inhibit AT1R-mediated IP3 generation, and mutations of the AT2R residues to their AT1R counterparts are sufficient to permit AT2 receptor–mediated generation of IP3.106 Both AT1 and AT2 receptors are expressed in many of the same tissues, and thus AT1R signal attenuation may depend on relative AT1/AT2 receptor expression levels.
AT1 receptors also exhibit positive crosstalk in vivo as a result of heterodimerization, in this case with bradykinin B2 receptors. Heterodimerization between AT1 and B2 receptors potentiates AT1 receptor signaling and sequestration after activation.107 Angiotensin II is a more potent activator of AT1 receptors within AT1:B2 heterodimers, whereas the potency and efficacy of bradykinin is reduced. The enhancement of AT1R signaling within the heterodimer does not require activation of the B2 receptor, but does require the ability of the B2 receptor to couple to G proteins.107 AT1 and B2 receptors are coexpressed in smooth muscle,108 kidney,109 platelets, and omental vessels,105 and thus AT1:B2 heterodimers may contribute to the normal functioning of the renin-angiotensin system, as suggested by alterations observed in preeclampsia. A consistent feature of preeclampsia, a multifactorial disorder characterized by hypertension and proteinuria during pregnancy,110 is an increased vascular responsiveness to angiotensin II, which is not the result of increased circulating levels of the hormone nor its receptor AT1.111 A partial explanation for the increased responsiveness to angiotensin II is the 4- to 5-fold increase in B2 receptor expression in both platelets and omental vessels from women with preeclampsia, but not from normotensive pregnant women.105 The increase in B2 receptor expression in platelets and omental vessels leads to an increase in AT1-B2 receptor heterodimerization; platelets from preeclamptic women displayed enhanced signaling in response to angiotensin II, despite levels of AT1 receptor comparable to normotensive women. Interestingly, AT1-B2 heterodimers were insensitive to reactive oxygen species, which inactivate AT1 receptor homodimers. The increased responsiveness to angiotensin II in women with preeclampsia results from both increases in AT1-B2 heterodimer formation and from the resistance to inactivation by H2O2 of AT1 receptors within heterodimers.105 Preeclampsia is thus the first physiological disorder linked to altered levels of heterodimerization of G protein–coupled receptors, and a prime example of crosstalk at the level of the receptors themselves.
Importance for Cardiovascular Physiology |
---|
Top Abstract Introduction Broad Evidence for GPCR... Contributions of GPCR... Minimal Model for GPCR... Functional Consequences of GPCR... Importance for Cardiovascular... References |
---|
Acknowledgments |
---|
References |
---|
Top Abstract Introduction Broad Evidence for GPCR... Contributions of GPCR... Minimal Model for GPCR... Functional Consequences of GPCR... Importance for Cardiovascular... References |
---|