Heterotrimeric G proteins are key players in transmembrane signaling by coupling a huge variety of receptors to channel proteins, enzymes, and other effector molecules. Multiple subforms of G proteins together with receptors, effectors, and various regulatory proteins represent the components of a highly versatile signal transduction system. G protein-mediated signaling is employed by virtually all cells in the mammalian organism and is centrally involved in diverse physiological functions such as perception of sensory information, modulation of synaptic transmission, hormone release and actions, regulation of cell contraction and migration, or cell growth and differentiation. In this review, some of the functions of heterotrimeric G proteins in defined cells and tissues are described.
All cells possess transmembrane signaling systems that allow them to receive information from extracellular stimuli like hormones, neurotransmitters, or sensory stimuli. This fundamental process allows cells to communicate with each other. All transmembrane signaling systems share two basic elements, a receptor which is able to recognize an extracellular stimulus as well as an effector which is controlled by the receptor and which can generate an intracellular signal. Many transmembrane signaling systems like receptor tyrosine kinases incorporate these two elements in one molecule. In contrast, the G protein-mediated signaling system is relatively complex consisting of a receptor, a heterotrimeric G protein, and an effector. This modular design of the G protein-mediated signaling system allows convergence and divergence at the interfaces of receptor and G protein as well as of G protein and effector. In addition, each component, the receptor, the G protein as well as the effector can be regulated independently by additional proteins, soluble mediators, or on the transcriptional level. The relatively complex organization of the G protein-mediated transmembrane signaling system provides the basis for a huge variety of transmembrane signaling pathways that are tailored to serve particular functions in distinct cell types. It is probably this versatility of the G protein-mediated signaling system that has made it by far the most often employed transmembrane signaling mechanism. In this review we summarize some of the biological roles of G protein-mediated signaling processes in the mammalian organism which are based on their cell type-specific function. Although we have tried to cover a wide variety of cellular systems and functions, the plethora of available data forced us to restrict this review. Particular emphasis is placed on cellular G protein functions that have been studied in primary cells or in the context of the whole organism using genetic approaches.
A. Basic Principles of G Protein-Mediated Signaling
More than 1,000 G protein-coupled receptors (GPCRs) are encoded in mammalian genomes. While most of them code for sensory receptors like taste or olfactory receptors, ∼400–500 of them recognize nonsensory ligands like hormones, neurotransmitters, or paracrine factors (53, 185, 519, 534, 649). For more than 200 GPCRs, the physiological ligands are known (Table 1). GPCRs for which no endogenous ligand has been found are “orphan” GPCRs (376, 389, 688).
Upon activation of a receptor by, e.g., its endogenous ligand, coupling of the activated receptor to the heterotrimeric G protein is facilitated. Multiple site-directed mutagenesis experiments have been performed on G protein-coupled receptors, and they have revealed various cytoplasmic domains of the receptors that are involved in the specific interaction between the receptors and the G protein. However, despite the determination of the structure of rhodopsin at atomic resolution (504), it is still not clear how specificity of the receptor-G protein interaction is achieved and how a ligand-induced conformational change in the receptor molecule results in G protein activation (177, 212, 213, 565, 674).
The heterotrimeric G protein consists of an α-subunit that binds and hydrolyzes GTP as well as of a β- and a γ-subunit that form an undissociable complex (233, 255, 475). Several subtypes of α-, β-, and γ-subunits have been described (Table 2). To dynamically couple activated receptors to effectors, the heterotrimeric G protein undergoes an activation-inactivation cycle (Fig. 1). In the basal state, the βγ-complex and the GDP-bound α-subunit are associated, and the heterotrimer can be recognized by an appropriate activated receptor. Coupling of the activated receptor to the heterotrimer promotes the exchange of GDP for GTP on the G protein α-subunit. The GTP-bound α-subunit dissociates from the activated receptor as well as from the βγ-complex, and both the α-subunit and the βγ-complex are now free to modulate the activity of a variety of effectors like ion channels or enzymes. Signaling is terminated by the hydrolysis of GTP by the GTPase activity, which is inherent to the G protein α-subunit. The resulting GDP-bound α-subunit reassociates with the βγ-complex to enter a new cycle if activated receptors are present. For recent excellent reviews on basic structural and functional aspects of G proteins, see References 49, 83, 361, and 526.
While the kinetics of G protein activation through GPCRs has been well described for quite a while, only recently has the regulation of the deactivation process been understood in more detail. Based on the observation that the GTPase activity of isolated G proteins is much lower than that observed under physiological conditions, the existence of mechanisms that accelerate the GTPase activity had been postulated. Various effectors have indeed been found to enhance GTPase activity of the G protein α-subunit, thereby contributing to the deactivation and allowing for rapid modulation of G protein-mediated signaling (23, 45, 348, 571). More recently, a family of proteins called “regulators of G protein signaling” (RGS proteins) has been identified, which is also able to increase the GTPase activity of G protein α-subunits (272, 481, 550). There are ∼30 RGS proteins currently known, which have selectivities for G protein α-subfamilies. The physiological role of RGS proteins is currently under investigation. Besides their role in the modulation of G protein-mediated signaling kinetics, they also influence the specificity of the signaling process and in some cases may have effector functions.
The interaction of G proteins with the inner side of the plasma membrane is facilitated by lipid modifications of both the α-subunit as well as of the γ-subunit of the βγ-complex (97, 448, 589, 728). Recent data provide evidence that heterotrimeric G proteins of the Gi family are also involved in receptor-independent processes (52, 413), which appear to be critically involved in the positioning of the mitotic spindle and the attachment of microtubules to the cell cortex (234). These processes also involve a group of proteins that carry a so-called GoLoco motif which functions as a guanine nucleotide dissociation inhibitor (684).
B. G Protein α-Subunits and βγ-Complexes
The functional versatility of the G protein-mediated signaling system is based on its modular architecture and on the fact that there are numerous subtypes of G proteins. The α-subunits that define the basic properties of a heterotrimeric G protein can be divided into four families, Gαs, Gαi/Gαo, Gαq/Gα11, and Gα12/Gα13 (Table 2). Each family consists of various members that often show very specific expression patterns. Members of one family are structurally similar and often share some of their functional properties. The βγ-complex of mammalian G proteins is assembled from a repertoire of 5 G protein β-subunits and 12 γ-subunits (Table 2). While β1- to β4-subunits form a tight complex with γ-subunits which can only be separated under denaturing conditions, the β5-subunit interaction with γ-subunits is comparably weak (347, 543). The β5-subunit is an exception in that it can also be found in a complex with a subgroup of RGS proteins (689). The βγ-complex was initially regarded as a more passive partner of the G protein α-subunit. However, it has become clear that βγ-complexes freed from the G protein α-subunit can regulate various effectors (112). These βγ-mediated signaling events include the regulation of ion channels (488), of particular isoforms of adenylyl cyclase and phospholipase C (169, 615), as well as of phosphoinositide-3-kinase isoforms (641). With a few exceptions, the ability of different βγ-combinations to regulate effector functions does not dramatically differ (112).
Most receptors are able to activate more than one G protein subtype. The activation of a G protein-coupled receptor therefore usually results in the activation of several signal transduction cascades via G protein α-subunits as well as through the freed βγ-complex. The pattern of G proteins activated by a given receptor determines the cellular and biological response, and activated receptors that lead to functionally similar or identical cellular effects usually activate the same G protein subtypes. The G protein receptor interaction in general does not occur in an absolutely specific or in a completely promiscuous manner. Some receptors appear to interact only with certain G protein subforms, and in some cellular systems, the compositions of defined G protein-mediated signaling pathways can be very specific. However, there are some characteristic patterns of receptor-G protein coupling that have been described for the majority of receptors (Fig. 2).
The G proteins of the Gi/Go family are widely expressed and especially the α-subunits of Gi1, Gi2 and Gi3 have been shown to mediate receptor-dependent inhibition of various types of adenylyl cyclases (615). Because the expression levels of Gi and Go are relatively high, their receptor-dependent activation results in the release of relatively high amounts of βγ-complexes. Activation of Gi/Go is therefore believed to be the major coupling mechanism that results in the activation of βγ-mediated signaling processes (112, 543). The function of members of the Gi/Go family has often been studied using a toxin from Clostridium botulinum (pertussis toxin; PTX) which is able to ADP-ribosylate most of the members of the Gαi/Gαo family close to their COOH termini. COOH-terminally ADP-ribosylated Gαi/Gαo is unable to interact with the receptor. Thus PTX treatment results in the uncoupling of the receptor and Gi/Go. The structural similarity between the 3 Gαi subforms suggests that they may have partially overlapping functions. In contrast to other G proteins, the effects of Go, which is particularly abundant in the nervous system, appears to be primarily mediated by its βγ-complex. Whether Gαo can regulate effectors directly is currently not clear. A less widely expressed member of the Gαi/Gαo family is Gαz (438), which in contrast to Gi and Go is not a substrate for PTX. Gαz is expressed in various tissues including the nervous system and platelets. It shares some functional similarities with Gi-type G proteins but has recently been shown to interact specifically with various other proteins including Rap1GAP and certain RGS proteins (438). Several α-subunits like gustducin and transducins belong to the Gαi/Gαo family and are involved in specific sensory functions (24, 126).
The Gq/G11 family of G proteins couples receptors to β-isoforms of phospholipase C (169, 538). The α-subunits of Gq and G11 are almost ubiquitously expressed while the other members of this family like Gα14 and Gα15/16 (Gα15 being the murine, Gα16 the human ortholog) show a rather restricted expression pattern. Receptors that are able to couple to the Gq/G11 family do not appear to discriminate between Gq and G11 (490, 660, 696, 705). Similarly, there is obviously no difference between the abilities of both G protein α-subunits to regulate phospholipase C β-isoforms. While Gαq and Gα11 both are good activators of β1-, β3-, and β4-isoforms of phospholipase C (PLC), the PLC β2-isoform is a poor effector for both (538). The biological significance of the diversity among the Gαq gene family is currently not clear. While the importance of Gq and G11 in various biological processes has been well established, the roles of Gα14 and Gα15/16, which show very specific expression patterns, are not clear. Mice carrying inactivating mutations of the Gα14 and Gα15 genes have no or very minor phenotypical changes (132; H. Jiang and M. I. Simon, personal communication). In contrast, mice lacking Gαq or both Gαq and Gα11 have multiple defects (489, 494, 495) (see below).
The G proteins G12 and G13, which are often activated by receptors coupling to Gq/G11, constitute the G12/G13 family and are expressed ubiquitously (139, 607). The analysis of cellular signaling processes regulated through G12 and G13 has been difficult since specific inhibitors of these G proteins are not available. In addition, G12/G13-coupled receptors usually also activate other G proteins. Most information on the cellular functions regulated by G12/G13 therefore came from indirect experiments employing constitutively active mutants of Gα12/Gα13. These studies showed that G12/G13 can induce a variety of signaling pathways leading to the activation of various downstream effectors including phospholipase A2, Na+/H+ exchanger, or c-jun NH2-terminal kinase (139, 193, 276, 523, 607). Another important cellular function of G12/G13 is their ability to regulate the formation of actomyosin-based structures and to modulate their contractility by increasing the activity of the small GTPase RhoA (79). Activation of RhoA by Gα12 and Gα13 is mediated by a subgroup of guanine nucleotide exchange factors (GEFs) for Rho which include p115-RhoGEF, PDZ-RhoGEF, and LARG (194, 236, 618). While the RhoGEF activity of PDZ-RhoGEF and LARG appears to be activated by both Gα12 and Gα13, p115-RhoGEF activity is stimulated only by Gα13. Recently, an interesting link between G12/G13 and cadherin-mediated signaling was described, both Gα12 and Gα13 interact with the cytoplasmic domain of some type I and type II class cadherins, causing the release of β-catenin from cadherins (434, 435). Various other proteins including Bruton's tyrosine kinase, the Ras GTPase-activating protein Gap1m, radixin, heat shock protein 90, AKAP110, protein phosphatase type 5, or Hax-1 have also been shown to interact with Gα12 and/or Gα13 (309, 359, 485, 532, 638, 709).
The ubiquitously expressed G protein Gs couples many receptors to adenylyl cyclase and mediates receptor-dependent adenylyl cyclase activation resulting in increases in the intracellular cAMP concentration. The α-subunit of Gs, Gαs, is encoded by GNAS, a complex imprinted gene that gives rise to several gene products due to the presence of various promoters and splice variants (Fig. 3). In addition to Gαs, two transcripts encoding XLαs and Nesp55 are generated by promoters upstream of the Gαs promoter. While the chromogranin-like protein Nesp55 is structurally and functionally not related to Gαs, XLαs is structurally identical to Gαs but has an extra long NH2-terminal extension that is encoded by a specific first exon (329). In contrast to Gαs, XLαs has a limited expression pattern being mainly expressed in the adrenal gland, heart, pancreatic islets, brain, and the pars intermedia of the pituitary (509). However, XLαs shares with Gαs the ability to bind to βγ-subunits and to mediate receptor-dependent stimulation of cAMP production (33, 339). Interestingly, the first exon of the Gnasxl gene encodes another protein termed ALEX (338), which is able to interact with the XL domain of XLαs and to inhibit its activity (188, 338). Interestingly, Nesp55 and XLαs are differentially imprinted. While the promoter of Nesp55 is DNA-methylated on the paternally inherited allele resulting in the expression only from the maternally inherited allele, the promoter driving XLαs expression is methylated on the maternal allele, and XLαs is only expressed from the paternal allele (244, 245, 515). Several other transcripts like Nespas of which some are believed to be untranslated show ubiquitous expression and are derived from the paternal allele due to differentially methylated promoter regions (243, 294, 396, 619). In contrast, the promoter driving the expression of Gαs has been shown to be biallelically active and to lack differential methylation (85, 244, 245, 733). However, in a few tissues such as the renal proximal tubules, the thyroid, pituitary, and ovaries, the paternal Gαs expression is silenced by an as yet undefined mechanism (209, 242, 394, 414, 723).
II. CARDIOVASCULAR SYSTEM
A. Autonomic Control of Heart Function
Cardiac regulation by the sympathetic system is mediated by β-adrenergic receptors that are coupled primarily to Gs (Fig. 4). cAMP produced in response to Gs activation directly modulates the gating of hyperpolarization-activated, cyclic nucleotide-gated channels and activates protein kinase A (PKA) which in turn phosphorylates several proteins involved in excitation-contraction coupling including L-type Ca2+ channels, phospholamban, or troponin I (44). These cellular changes are believed to underlie the well-known effects of sympathetic cardiac activation including positive chronotropic, dromotropic, lusitropic, and inotropic effects (545). Transgenic overexpression of the short form of Gαs (Gαs-S) in the murine heart had no effect on the basal cardiac function but resulted in an enhanced efficacy of β-adrenoceptor Gs signaling, and chronotropic and inotropic responses to catecholamines were increased (299). Once Gαs-overexpressing mice become older, they develop clinical and pathological signs of cardiomyopathy (300). These pathological processes are accompanied by a lack of normal heart rate variability as well as of protective desensitization mechanisms (635, 650). The development of cardiomyopathy after prolonged overexpression of Gαs is in line with the current concept of the pathophysiological mechanisms underlying the development of chronic heart failure. The insufficient cardiac output characteristic for heart failure typically goes along with an increased sympathetic tone resulting in chronic catecholamine stimulation of cardiomyocytes, which is believed to be deleterious (71). Although the β1-adrenoceptor is the predominant subtype expressed in cardiomyocytes, also β2-adrenoceptors are expressed in the heart (545). Interestingly, there is increasing evidence that β1- and β2-adrenoceptors play different roles in catecholamine-induced cardiomyopathy. Mice overexpressing human β2-adrenoceptors have only slightly altered cardiac function and appear to have normal life expectancy (259, 260, 443, 544) while mice overexpressing β1-adrenoceptors develop severe hypertrophy and die of heart failure (162). The β1- and β2-adrenoceptors also differ with regard to their signal transduction. While β1-adrenergic receptors are Gs coupled, β2-adrenoceptors are also able to couple to Gi-type G proteins (700, 701) (Fig. 4). The additional activation of Gi via β2-adrenergic receptors may explain the observed differences in signaling induced via β1- and β2-adrenergic receptors (116, 125, 232, 405, 725, 736). This led to the hypothesis that β2-adrenoceptor stimulation exerts some sort of protection against cardiac hypertrophy and failure, especially under conditions of chronic activation of the β-adrenergic system and that this is due to signaling via Gi. The well-documented upregulation of Gi in human heart failure (174, 482) may be a mechanism to counteract deleterious Gs-mediated signaling. The potential cardioprotective role of Gi is also supported by studies in mice. While the overexpression of β2-adrenergic receptors in normal cardiomyocytes is well tolerated, mice which lack in addition the major Gi α-subunit, Gαi2, die within a few days after birth (180). Mice that overexpress β2-adrenoceptors in cardiomyocytes and which carry only one intact Gαi2 gene allele develop more pronounced cardiac hypertrophy and earlier heart failure compared with β2-adrenoceptor transgenic animals with normal Gαi2 levels.
The muscarinic acetylcholine (M2) receptor that is coupled to Gi/Go G proteins mediates the parasympathetic regulation of the heart (Fig. 4). The negative chronotropic and dromotropic effects of the parasympathetic system are believed to result from the Gi-mediated inhibition of adenylyl cyclase, resulting in an inhibition of the cAMP production as well as by the activation of G protein-regulated inward rectifier potassium channels (GIRK) by βγ-subunits released from activated Gi/Go (601). The atrial GIRK consists of Kir3.1 and Kir3.4 subunits. Mice lacking either of the two channel subunits have normal basal heart rates but show reduced vagal and adenosine-mediated slowing of heart rate and markedly reduced heart rate variability, which is thought to be determined by the vagal tone (47, 680). The involvement of Gβγ complexes in regulation of GIRK channels has been well established using electrophysiological and biochemical approaches (349, 398, 679). Mice in which the amount of functional Gβγ protein was reduced by more than 50% in cardiomyocytes also show an impaired parasympathetic heart rate control (207). The central role of Gαi in inhibitory regulation of heart rate and atrioventricular conductance has led to attempts to treat cardiac arrhythmias by atrioventricular nodal gene transfer of Gαi2 in a model of persistent atrial fibrillation in swine (146). While wild-type Gαi2 did not change basal heart rate, a constitutively active mutant of Gαi2 resulted in a significant decrease in heart rate. When tested for their effects in a model for tachycardia-induced cardiomyopathy, the condition was significantly improved by wild-type Gαi2 and even more by constitutively active Gαi2 (37). In addition to the stimulatory regulation of potassium channels, muscarinic regulation of heart function also involves inhibition of voltage-dependent L-type Ca2+ channels via an unknown mechanism. In mice lacking the α-subunit of Go, inhibitory muscarinic regulation of cardiac L-type Ca2+ channels was abrogated, although Gαo represents only a minor fraction of all G proteins in the heart (639). Interestingly, mice which lack the α-subunit of Gi2 (Gαi2) also show a severely affected inhibitory regulation of L-type Ca2+ channels via muscarinic M2 receptors (101, 468). This suggests that both G proteins, Go and Gi2, are involved in the regulation of cardiac L-type Ca2+ channels.
B. Myocardial Hypertrophy
Myocardial hypertrophy is the chronic adaptive response of the heart to injury or increased hemodynamic load. It is characterized by increased cardiomyocyte size and protein content, as well as altered gene expression, recapitulating an embryonic phenotype (109, 301). Such pathological myocardial hypertrophy was shown to be associated with increased cardiac mortality (191, 285, 535), raising the question whether prevention of pathological hypertrophy is beneficial or not (191). Several mechanosensitive mechanisms involving stretch-activated ion channels, integrins or Z-disc proteins were suggested to mediate myocardial hypertrophy in response to pressure overload (191, 285, 535). In addition, GPCR agonists like norepinephrine/phenylephrine, angiotensin II, or endothelin-1 were shown to induce a hypertrophic phenotype in cultured rat embryonic cardiomyocytes (4, 341, 560, 581). These ligands are known to activate Gq/G11-coupled receptors, such as the α1-adrenergic receptor, the angiotensin AT1 receptor, or the endothelin ETA receptor (362, 561, 592). Activation of Gαq by Pasteurella multocida toxin (559) or expression of wild-type Gαq (5, 362) induces the hypertrophic phenotype in cultured cardiomyocytes, while inhibition of Gq/G11 by the RGS domain of GRK2 inhibited agonist-induced hypertrophy (423). In vivo, cardiac-restricted expression of wild-type (128) or constitutively active Gαq (437) results in cardiac hypertrophy. In addition, in vivo overexpression of typically Gq/G11-coupled receptors (444, 474) or their downstream effectors (65, 454, 656) induces hypertrophy. Conversely, in vivo inhibition of Gq/G11 by overexpression of RGS4, a GTPase-activating G protein for Gq/G11 and Gi/Go (547), or by overexpression of the COOH terminus of Gαq (10) results in a reduced hypertrophic response, and cardiomyocyte-specific inactivation of the genes encoding Gαq/Gα11 completely abrogates the hypertrophic response elicited by pressure overload (677). Interestingly, an impaired hypertrophic response due to inhibition of Gq/G11-mediated signaling does not negatively influence long-term cardiac function (166), suggesting that hypertrophy in response to pressure overload is not necessarily required to maintain cardiac function. In addition to pressure overload-induced myocardial hypertrophy, the Gq/G11-mediated signaling pathway was also implicated in the pathogenesis of diabetic cardiomyopathy. Gαq levels and PKC activity were shown to be enhanced in the streptozotocin-induced diabetic rat heart (714), and heart specific overexpression of RGS4 protected mice against different models of diabetic cardiomyopathy. In contrast, heart-specific expression of a RGS-resistant Gαq caused sensitization towards diabetic cardiomyopathy (235). The downstream signaling processes in Gq/G11-mediated hypertrophy are complex and not fully understood (Fig. 5). Intracellular Ca2+ mobilization in response to activation of Gq/G11-coupled receptors promotes Ca2+/calmodulin (CaM)-dependent activation of calcineurin, which in turn mediates dephosphorylation and nuclear translocation of transcription factors of the NFAT (nuclear factor of activated T cells) family. Although activation of the calcineurin/NFAT signaling pathway is clearly sufficient to induce myocardial hypertrophy, it is not completely clear whether inhibition of this signaling pathway prevents hypertrophy (for review, see Refs. 190, 191). In addition, a variety of other effectors have been implicated in myocardial hypertrophy, such as protein kinase C (PKC) isoforms, mitogen-activated protein (MAP) kinases, the phosphatidylinositol (PI) 3-kinase/Akt/GSK-3 pathway or small GTPases (for review, see Refs. 148, 191, 285, 535).
GPCRs known to mediate myocardial hypertrophy can also activate G12/G13 family G proteins, resulting in the activation of RhoA (215, 257). RhoA was suggested to be involved in the hypertrophic responses to phenylephrine, endothelin, chronic hypertension, or overexpression of Gαq (128, 264, 278, 360, 562, 567). Expression of inhibitory COOH-terminal peptides of Gα12 and Gα13, as well as expression of the G12/G13 specific RGS domain of p115RhoGEF, inhibited phenylephrine-mediated JNK activation in neonatal cardiomyocytes (423). In addition, overexpression of a constitutively active mutant of Gα13 induced a hypertrophic response in neonatal cardiomyocytes, with increased expression of the hypertrophy-associated embryonic gene program (179). However, no in vivo data on the role of G12/G13 in myocardial hypertrophy are available.
C. Smooth Muscle Tone
Smooth muscle tone is controlled by the phosphorylation state of the regulatory light chain (MLC20) of myosin II (for review, see Refs. 269, 593, 594). MLC20 is phosphorylated by the Ca2+/CaM-dependent myosin light chain kinase (MLCK), leading to enhanced velocity and force of actomyosin cross-bridging. Dephosphorylation of MLC20 is mediated by myosin phosphatase, an enzyme that is negatively regulated by the Rho/Rho-kinase pathway. Thus increased contractility can be achieved through Ca2+-mediated MLCK activation and through Rho-dependent inhibition of MLC20 dephosphorylation. A variety of transmitters and hormones regulate smooth muscle tone through GPCRs (Fig. 6). Typical vasoconstrictor receptors, such as the angiotensin AT1 receptor, the endothelin ETA receptor, or the α1-adrenergic receptor, act on Gq/G11-coupled receptors (159, 215, 724) to enhance intracellular Ca2+ concentration, leading to MLCK activation. Increased intracellular Ca2+ levels are not only due to IP3-mediated Ca2+ release from the sarcoplasmic reticulum, but also to Ca2+ influx through cation channels or voltage-gated Ca2+ channels (for review, see Refs. 269, 593). In addition, many Gq/G11-coupled receptors have been shown to activate RhoA, thereby contributing to Ca2+-independent smooth muscle contraction (593, 594). Smooth muscle specific overexpression of a COOH-terminal Gαq peptide, which is believed to inhibit the receptor/G protein interaction, ameliorates hypertension induced by long-term treatment with phenylephrine, serotonin, or angiotensin II (331). Mice lacking RGS2, a GTPase activating G protein which accelerates the inactivation of Gq/G11, suffer from hypertension (261). Interestingly, it was recently shown that the nitric oxide/cGMP cascade, which constitutes the main relaxant pathway in smooth muscle cells, negatively regulates Gq/G11 signaling by cGMP kinase-mediated phosphorylation and activation of RGS2 (625). However, in addition to this peripheral vascular mechanism, an increased sympathetic tone might contribute to elevated arterial blood pressure in RGS2-deficient mice (227).
In vitro, most Gq/G11-coupled vasoconstrictor receptors also activate G12/G13 family G proteins, like the receptors for endothelin-1, vasopressin, angiotensin II (215, 257), thrombin (411), or thromboxane A2 (491, 492). Constitutively active forms of Gα12 and Gα13 induced a pronounced, RhoA-dependent contraction in cultured vascular smooth muscle cells, and receptor-mediated contractions were strongly inhibited by dominant negative forms of Gα12 and Gα13 (215). These data suggest that also the G12/G13-mediated signaling pathway is involved in the regulation of smooth muscle tone, most likely by modulating the activity of myosin phosphatase via Rho/Rho-kinase. In accordance with this, inhibition of Rho-kinase was shown to normalize blood pressure in humans and experimental animals (426, 636). The relative contribution of Gq/11/Ca2+-mediated and G12/13/Rho/Rho-kinase-mediated signaling to regulation of vascular smooth muscle tone is not clear. However, data obtained in visceral smooth muscle suggested that Gq/G11 conveys a fast, transient response, while G12/G13 mediates a sustained, tonic contraction (257).
Vascular smooth muscle relaxation is mediated by a variety of mechanisms, one of them being the activation of Gs-coupled receptors like the adenosine A2 receptors, β2-adrenergic receptors, or prostaglandin receptor subtypes IP, DP, and EP2. How the subsequent increase in cAMP levels reduces smooth muscle tone is not understood. In vitro data suggest that the relaxant effect is partially due to a PKA-mediated MLCK phosphorylation, which decreases the enzyme's affinity for the Ca2+/CaM complex, but the physiological relevance of this signaling pathway is unclear. Possible other substrates for PKA are heat shock protein 20, RhoA, or myosin phosphatase (for review, see Ref. 269). cAMP has also been suggested to cross-activate cGMP kinase I in vascular or airway smooth muscle (32, 390), but this hypothesis has been questioned by the finding that vessels from cGKI-deficient mice relax normally in response to cAMP (518). In addition, cAMP-independent mechanisms of Gs-mediated relaxation involving large-conductance, Ca2+-activated K+ (MaxiK, BK) channels have been proposed (624).
The role of Gi-mediated signaling in vascular smooth muscle tone seems to differ between different vessel types. An inhibitory effect of PTX on norepinephrine-induced contractility was reported in rat tail artery (516, 600) but was absent in aorta (517). High blood pressure in spontaneously hypertensive rats is preceded by increased expression of Gi proteins (18, 19), and PTX treatment delayed the onset of hypertension (387), suggesting that decreased cAMP levels play a role in the pathogenesis of this model of hypertension. Enhanced signaling via a PTX-sensitive G protein was reported in immortalized B lymphoblasts from patients with essential hypertension (520), and this was attributed to a C825T polymorphism in the gene coding for Gβ3, a constituent of the Gi heterotrimer (586). The C825T polymorphism was suggested to be associated with an increased risk of hypertension, obesity, and arteriosclerosis in some (for review, see Ref. 585) but not in all studies (283, 616, 617). However, the significance of genetic association studies in general remains controversial (20, 199, 291).
Very similar to vascular smooth muscle, also airway smooth muscle tone is mainly regulated by Gq/G11 family G proteins, which mediate bronchoconstriction, and Gs family G proteins, which mediate bronchorelaxation. Acetylcholine released from postganglionic parasympathetic nerves controls resting tone mainly via the Gq/G11-coupled M3 receptor subtype (90), but also other Gq/G11-coupled receptors are expressed in airway smooth muscles, like the H1 histamine receptor (133, 222), the leukotriene CysLT1 receptor (314), the B2 bradykinin receptor (421, 630), the ETB endothelin receptor (216, 241, 441), and others. Airway hyperreactivity in the A/J mouse strain was suggested to be due to enhanced agonist affinity and increased G protein coupling efficiency of the M3 muscarinic receptor (205), and Gq protein was shown to be upregulated in antigen-induced airway hyperresponsive rats (106). Mice lacking the α-subunit of Gq showed impaired metacholine-induced airway responses and lacked the typical increase in metacholine sensitivity after allergen sensitization and reexposition (55). Not much is known about the role of Gα12 and Gα13 in airway smooth muscle tone regulation. The fact that repetitive antigen challenge significantly increases the expression of these proteins in airway smooth muscle suggests a role in allergic asthma (105, 108), but direct evidence for an involvement of G12/G13 is still lacking. Gs-coupled receptors play an important role in the relaxation of contracted airway smooth muscle, most prominently the β2-adrenergic receptor, but also the prostaglandin E2 receptor EP2 (512) or the prostacyclin IP receptor (41) (for review, see Ref. 628). The Gi family of G proteins contributes to the regulation of airway smooth muscle contractility mainly by inhibiting the relaxant effects of Gs. The inhibitory effect of PTX on acetylcholine-induced bronchoconstriction is negligible in normal rats, but significant in rats suffering from antigen-induced airway hyperresponsiveness (107). In these mice, Gαi3 protein is upregulated in bronchial smooth muscle cells, suggesting that the relative contribution of Gi-mediated constriction is increased in antigen-challenged airway smooth muscle (107).
D. Platelet Activation
Platelets are small cell fragments that circulate in the blood and adhere at places of vascular injury to the vessel wall where they become activated resulting in the formation of a platelet plug that is responsible for primary hemostasis. Platelets can also become activated under pathological conditions, e.g., on ruptured atherosclerotic plaques leading to arterial thrombosis. Platelet adhesion and activation is initiated by their interaction with adhesive macromolecules like collagen and von Willebrand factor (vWF) at the subendothelial surface (303, 554). While collagen is able to induce firm adhesion of platelets to the subendothelium (666), the recruitment of additional platelets to the growing platelet plaque requires the local accumulation of diffusible mediators that are produced or released once platelet adhesion has been initiated, and some level of activation through platelet adhesion receptors has occurred (3). These mediators include ADP/ATP and thromboxane A2 (TxA2), which are secreted or released from activated platelets as well as thrombin, which is produced on the surface of activated platelets. These platelet stimuli have in common their action through G protein-coupled receptors. While ADP induces the activation of Gq and Gi via P2Y1 and P2Y12 receptors (197, 354), the activated TxA2 receptor (TP) couples to Gq and G12/G13 (337, 492) (Fig. 7). G protein-coupled protease-activated receptors (PARs) that are activated by thrombin are functionally coupled to Gq, G12/G13, and in some cases to Gi (121). In response to these secondary mediators of platelet activation, platelets immediately undergo a shape change reaction during which they become spherical and extrude pseudopodia-like structures. In addition, the glycoprotein IIb/IIIa (integrin αIIbβ3) undergoes a conformational change resulting in binding of fibrinogen/vWF and subsequent platelet aggregation. Finally, the formation and release of TxA2, thrombin, and ADP is further stimulated. Thus secondary mediators increase through G protein-coupled receptors their own formation resulting in an amplification of their effects, and eventually all G protein-mediated signaling pathways induced via these receptors become activated. The multiple positive feedback mechanisms operating during platelet activation have obscured the exact analysis of the roles individual G protein-mediated signaling pathways play during the platelet activation process. Progress has recently been made using genetic mouse models in understanding the role of individual G protein-mediated signaling pathways during platelet activation.
The requirement of Gq-mediated signaling for agonist-induced platelet activation has been demonstrated by the phenotype of Gαq-deficient platelets, which fail to aggregate and to secrete in response to thrombin, ADP, and TxA2 due to a lack of agonist-induced phospholipase C activation. This dramatic phenotype found in Gαq-deficient platelets is due to the fact that platelets lack Gα11 (313), which is in most other cells coexpressed with Gαq and can compensate Gαq deficiency. Mice lacking Gαq have increased bleeding times and are protected against collagen/epinephrine-induced thromboembolism (494). Although Gq-mediated signaling appears to be absolutely required for platelet activation, there is clear evidence that also Gi type G proteins need to be activated to induce full activation of integrin αIIbβ3. In mice lacking the α-subunit of Gi2, the response of ADP that acts through the Gi-coupled P2Y12 receptor is reduced (304). However, also the effects of mediators like thrombin and TxA2, which primarily signal through Gq and G12/G13 were found to be inhibited in platelets lacking Gαi2 (304, 712). This supports the view that platelet activation by thrombin and thromboxane A2 requires in part the action of secondary mediators like ADP, which are released after activation of Gq-mediated signaling pathways through TxA2 and thrombin receptors. An important role of the Gi-mediated signaling pathway in platelet activation is also suggested by studies in platelets lacking the Gq-coupled P2Y1 receptor or after pharmacological blockade of P2Y1 (170, 251, 310, 379, 568). These platelets do not aggregate in response to low and intermediate concentrations of ADP unless Gq-mediated signaling is induced via activation of another receptor. Similarly, platelets lacking P2Y12 or in which P2Y12 was pharmacologically blocked did not aggregate in response to ADP unless the Gi-mediated pathway was activated via a different receptor (181, 568). Thus there is clear evidence that Gq and Gi synergize to induce platelet activation. It is currently not clear how Gi contributes to integrin αIIbβ3 activation in platelets, but a decrease in cAMP levels is unlikely to be involved (129, 529, 569, 712). Another member of the Gi family of heterotrimeric G proteins, Gz, has been implicated in platelet activation induced by epinephrine acting on α2-adrenergic receptors. In contrast to ADP, TxA2, and thrombin, epinephrine is alone not able to fully activate mouse platelets. However, it is able to potentiate the effect of other platelet stimuli. In Gαz-deficient platelets, the inhibitory effect of epinephrine on adenylyl cyclase and epinephrine-potentiating effects were strongly impaired while the effects of other platelet activators appear to be unaffected (713).
Despite the central role of Gq in platelet activation, it was recently demonstrated that induction of Gi- and G12/G13-mediated signaling pathways is sufficient to induce integrin αIIbβ3 activation (149, 483). Interestingly, in Gα13-deficient platelets, but not in Gα12-deficient platelets, the potency of various stimuli including TxA2, thrombin, and collagen to induce platelet shape change and aggregation is markedly reduced (455). These defects are accompanied by a defect in the activation of RhoA and a delayed phosphorylation of the myosin light chain as well as by an inability to form stable platelet thrombi under high sheer stress conditions (455). In addition, mice carrying platelets that lack Gα13 have an increased bleeding time and are protected against the formation of arterial thrombi induced in a carotid artery thrombosis model (455). These data indicate that in addition to Gq and Gi also G13 is crucially involved in the signaling processes mediating platelet activation via G protein-coupled receptors both in hemostasis and thrombosis. These findings also indicated that G13-mediated signaling is not only involved in the response of platelets to relatively low stimulus concentrations that induce platelet shape change but is also required for normal responsiveness of platelets at higher stimulus concentrations. A reduced potency of platelet activators in the absence of G13-mediated signaling becomes in particular limiting under high flow conditions that lead to a rapid clearance of soluble stimuli from the site of platelet activation and formation of mediators. In addition, the defective activation of RhoA-mediated signaling in the absence of G13 appears to contribute to the observed defect in the stabilization of platelet aggregates under high sheer stress ex vivo as well as in vivo. In fact, RhoA-mediated signaling has been suggested to be required for platelet aggregation under high sheer conditions as well as for the irreversible aggregation of platelets in suspension (450, 570).
These studies have clearly shown that three G proteins are major mediators of platelet activation via G protein-coupled receptors: Gq, Gi2, and G13. However, even in the absence of either Gq, Gi2, or G13 some platelet activation can still be induced, while in the absence of both Gαq and Gα13, platelets are unresponsive to thrombin, TxA2, or ADP. This indicates that the activation of Gi-mediated signaling alone is not sufficient to induce any platelet activation (456). The optimal activation of platelets under physiological and pathological conditions obviously requires the parallel signaling through several heterotrimeric G proteins.
III. ENDOCRINE SYSTEM AND METABOLISM
The endocrine system consists of a variety of glands and other structures that produce, store, and secrete hormones directly into the systemic circulation, thereby controlling electrolyte and water homeostasis, metabolism, growth, reproduction, etc. GPCRs contribute to endocrine functions in a twofold way: 1) by mediating hormonal end organ effects and 2) by controlling hormone secretion itself. Hormone secretion, as well as secretion from neuronal or exocrine cells, typically involves elevation of cytosolic Ca2+ and/or cAMP (for review, see Refs. 160, 738). In most secretory cells, Ca2+ influx through voltage-operated Ca2+ channels is the dominant mode of regulation, like in adrenal chromaffin cells (160), while in other cells, such as anterior pituitary gonadotropes, Ca2+ mobilization from internal stores is the critical step (633). In yet another endocrine cell, such as lactotroph cells, both increased intracellular Ca2+ levels and cAMP production contribute to secretion (130, 186, 620). Accordingly, with rare exceptions, activation of Gs and/or Gq/G11 family G proteins enhances secretion regardless of the endocrine cell type involved.
A. Hypothalamo-Pituitary System
Hormone release from the anterior pituitary is tightly controlled by hypothalamic releasing hormones and release inhibiting factors, all of which act through GPCRs. The receptors for corticotropin-releasing hormone (263) and growth hormone-releasing hormone (GHRH) (588) primarily act through Gs, while receptors for gonadotropin-releasing hormone (445), thyrotropin-releasing hormone (211, 720), and the many prolactin-releasing factors (186) mainly act through Gq/G11 family G proteins, and only partly through Gs. In addition to their secretagogue effects, hypothalamic releasing hormones regulate hormone synthesis and cell proliferation (81, 430, 531, 583, 587, 647). Anterior pituitary secretion and proliferation is not only stimulated by the classical hypothalamic releasing hormones, but also by a variety of other factors, such as the gastrointestinal peptide hormone ghrelin, which enhances growth hormone (GH) secretion via the predominantly Gq/G11-coupled growth hormone secretagogue receptor GHS-R (343, 588), or members of the pituitary adenylate cyclase-activating polypeptide (PACAP)/glucagon superfamily, which exert secretagogue effects on a variety of pituitary cell types via their Gs-coupled receptors (579). The in vivo relevance of Gs family G proteins in anterior pituitary function was studied in mice and in patients with inactivating or activating Gs mutants. Somatotroph-specific overexpression of cholera toxin, which irreversibly activates Gs by ADP ribosylation, caused somatotroph hyperplasia, increased GH levels and gigantism in mice (82). In humans, activating mutations of GNAS can be found in ∼40% of GH producing pituitary tumors (363, 406), as well as in 10% of nonfunctioning pituitary adenomas (406, 632, 685). These activating mutations of GNAS encode substitutions of either Arg-201 or Gln-227, two residues that are critical for the GTPase reaction (187, 223, 363, 406). In GH-secreting tumors, the mutation is almost always in the maternal allele, presumably because Gαs is mainly expressed from the maternal allele (“paternally imprinted”) in pituitary cells (242). Activating GNAS mutations were also, though rarely, found in corticotroph (541, 686), but not in thyrotroph tumors (147, 406). Such activating somatic GNAS mutations are not necessarily restricted to the pituitary, but are often part of the McCune-Albright syndrome, which is defined by the trias fibrous dysplasia of bone, café-au-lait skin pigmentation, and endocrine hyperfunctions of variable degree (for review, see Refs. 599, 669). Endocrine hyperfunction is due to constitutive activation of Gs signaling in other endocrine glands, leading to adrenal hyperplasia with Cushing syndrome (60, 182), precocious puberty (138, 574), or hyperthyroidism (see sect. iiiC). In melanocytes, increased Gs activity mimics the activity of melanocyte stimulating hormone, leading to typical café-au-lait hyperpigmentation (334).
Heterozygous inactivating GNAS mutations result in Albright hereditary osteodystrophy (AHO), a congenital disorder characterized by obesity, short stature, brachydactyly, subcutaneous ossifications, and neurobehavioral deficits of variable severity (for review, see Refs. 13, 366, 599, 669). In addition to these defects, patients with maternally inherited mutations show multihormone resistance (termed pseudohypoparathyroidism type Ia, PHP1a) in tissues with a paternally imprinted GNAS allele, such as proximal tubules of the kidney, thyroid, or ovaries (209, 242, 414, 723). In these tissues, the effects of Gs-coupled hormone receptors, like those for parathyroid hormone, thyroid stimulating hormones, or the gonadotropins, are impaired. Clinically, this results in variable degrees of hypocalcemia and hyperphosphatemia, hypothyroidism (see also sect. iiiC), and delayed or incomplete sexual development and reproductive dysfunction in women (13, 366, 599, 669). These abnormalities of the reproductive system are easily explained by malfunction of receptors for follicle-stimulating hormone and luteinizing hormone. In addition, at least in mice, the Gs-coupled orphan receptor GPR3 is crucially involved in the maintenance of meiotic arrest in oocytes (432, 433).
The phenotype of humans heterozygous for an inactivating GNAS mutation is partly reproduced in mice carrying a targeted disruption of Gnas exon 2. In these animals, PTH resistance was only found if the mutation was maternally inherited, and only these animals showed reduced Gαs expression in the renal cortex (723). In humans, renal PTH resistance without Albright hereditary osteodystrophy (PHPIb) can also be due to other GNAS mutations, such as a mutant which results in a biallelic paternal imprinting phenotype (395), or a mutant unable to interact with the PTH receptor (697). Yet another GNAS mutation causes impaired signaling via the PTH and TSH receptors, but enhanced signaling via the likewise Gs-coupled receptor for luteinizing hormone, leading to enhanced testosterone production. This paradoxical combination of gain and loss of function is explained by the fact that the underlying GNAS mutation results in a constitutively active form of Gαs which, however, is temperature sensitive. The mutant is stable only at the relatively low temperature in the testis, but rapidly degraded at 37°C, leading to Gαs deficiency (287). With respect to pituitary function, patients with inactivating GNAS mutations show variable degrees of GHRH resistance (415), GH deficiency (210), or hypoprolactinemia (88). In accordance with the important role of Gs family G proteins in lactotrophs and somatotrophs, hypothalamic inhibiting hormones, like dopamine or somatostatin, act through Gi-coupled receptors (311, 536).
Releasing hormone secretion itself is influenced by GPCRs, and several former orphan receptors were recently shown to positively regulate releasing hormone secretion. Kisspeptins for example, a family of peptides derived from the metastasis suppressor gene Kiss-1, were shown to enhance hypothalamic gonadotropin-releasing hormone secretion via the GPR54 receptor (137, 220, 472), and genetic inactivation of GPR54 in mice or mutation in humans causes hypogonadotropic hypogonadism (137, 196, 220, 575). The peptide hormone ghrelin induces pituitary growth hormone release not only directly via activation of GHS-R on somatotroph cells, but also acts as a releasing factor for hypothalamic GHRH (343). Both GPR54 and GHS-R are known to activate Gq/G11 family G proteins (343, 345), suggesting that releasing hormone release is controlled by the same mechanisms as pituitary hormone release. In line with this notion, mice lacking both Gαq alleles and one Gα11 allele selectively in the nervous system show severe somatotroph hypoplasia with dwarfism due to reduced hypothalamic GHRH production, which is probably secondary to impaired GHS-R signaling (676).
B. Pancreatic β-Cells
The tight regulation of blood glucose levels is mainly achieved by the on-demand release of insulin from pancreatic β-cells. High glucose levels result in enhanced intracellular glucose metabolism with ATP accumulation and consecutive closure of ATP-sensitive K+ channels, leading to the opening of voltage-operated Ca2+ channels and Ca2+-mediated insulin exocytosis (27, 119). In addition to the ATP-dependent mechanism of insulin release, several GPCRs have been shown to either amplify or to inhibit glucose-induced insulin release (for review, see Refs. 161, 364, 558), and these receptors and their respective ligands play an important role in the regulation of islet function by, e.g., the autonomous system (for review, see Ref. 7). Neuropeptides and hormones that potentiate insulin secretion mainly act though Gs-coupled receptors, like glucose-dependent insulinotropic polypeptide, secretin, cholecystokinin, PACAP, glucagon, vasoactive intestinal polypeptide, or glucagon-like peptide-1 (GLP-1) (for review, see Refs. 161, 418, 542). The potentiating effect of Gs on glucose-induced insulin release (576, 608, 609) might either be mediated by phosphorylation of voltage-operated Ca2+ channels (295) or through the opening of nonselective cation channels (275). Transgenic expression of a constitutively active Gαs mutant in mouse β-cells caused increased islet cAMP production and insulin secretion, but these changes were only detectable in the presence of phosphodiesterase inhibitors, suggesting that increased Gαs activity is normally compensated by upregulation of cAMP degrading enzymes like phosphodiesterases (408). Conversely, activation of receptors coupled to Gi or Go, like the α2-adrenergic receptor or receptors for somatostatin, neuropeptide Y, prostaglandin E2, or galanin, inhibits insulin secretion in a PTX-sensitive manner (319, 328, 365, 514, 542).
Not only Gs family members, but also Gq/G11 family G proteins, can mediate potentiation of glucose-induced insulin release. Acetylcholine released from postganglionic parasympathetic nerves or muscarinic agonists act through the Gq/G11-coupled M3 receptor (58, 157) to enhance insulin release during the cephalic phase of insulin secretion (8, 447, 691). This effect was shown to depend on PLC activation and consecutive inositol 1,4,5-trisphosphate (IP3)-mediated intracellular Ca2+ elevation (46, 469, 726) and PKC activation (22). The exact pathways leading to increased insulin secretion are not clear, but activation of L-type Ca2+ channels (58), modulation of ATP-sensitive K+ channels (469), activation of CaM-kinase II (427, 537), or enhanced plasma membrane Na+ permeability (254) have been suggested. In addition to acetylcholine, a variety of other local mediators act through Gq/G11-coupled receptors to enhance insulin release, like cholecystokinin via the CCK1 receptor (652), bombesin via the BB2 receptor (521, 646), arginine vasopressin via the V1b receptor (375, 501, 539), or endothelin via the ETA receptor (224). Fatty acids such as palmitate potentiate insulin secretion at high glucose levels independently of ATP formation (527, 663), and Ca2+ influx via voltage-operated Ca2+ channels or intracellular Ca2+ mobilization was suggested to mediate these effects. The former orphan GPCR GPR40 was shown to mediate the effects of saturated and unsaturated fatty acids (C>6) on intracellular Ca2+ mobilization in pancreatic β-cells (70, 296, 346). These effects were not PTX sensitive, suggesting that Gq/G11-mediated intracellular Ca2+ mobilization was involved (70, 296). Another recently deorphanized GPCR probably coupled to Gq/G11 is GPR120, which was suggested to be involved in fatty acid-induced release of GLP-1 from intestinal cells, thereby contributing to GLP-1-mediated insulin release (265).
C. Thyroid Gland/Parathyroid Gland
Thyroid stimulating hormone (TSH) regulates thyroid cell proliferation as well as thyroid hormone synthesis and release through the G protein-coupled TSH receptor (508). In vitro, the TSH receptor was shown to couple to all four G protein families (15, 16, 368), but the major signal transduction pathway in vivo seems to be the Gs/cAMP cascade, which was shown to activate iodide organification, thyroid hormone production, secretion, and thyroid cell mitogenesis (153–155, 546). In TSH receptor-deficient mice, direct stimulation of adenylyl cyclase restores the ability to concentrate and organify iodide, suggesting that expression of the sodium-iodide symporter is controlled by Gαs (417). In vitro, overexpression of constitutive active Gαs in a thyroid cell line (462) or activation of Gs by transgenic expression of cholera toxin in the mouse thyroid (727) caused hyperplasia and increased hormone secretion. In line with this, spontaneous activating mutations of GNAS can cause thyroid cell hyperfunction in humans, leading to hyperthyroidism, goiter, and benign adenoma (175, 406, 424, 502). Malignant transformation of thyroid cells has also been observed (165, 611, 711), but seems to require additional mutational or epigenetic events (115). Much more frequent than activating GNAS mutations are the activating mutations of the TSH receptor itself (508), which mainly lead to constitutive activation of the Gs/cAMP cascade, or, in some cases, to activation of both Gs- and Gq/G11-coupled pathways (48, 643). Since the paternal GNAS allele is partly imprinted in thyroid cells (209, 394), inactivating GNAS mutants inherited from the maternal side cause TSH resistance with moderately elevated TSH levels and low thyroid hormone levels (171, 381, 382, 670, 719). Mild TSH resistance can also be observed in patients with PHP1B due to a GNAS mutation with paternal specific epigenotype of both exon 1A regions (34, 394, 395). In addition to adenylyl cyclase activation, TSH stimulates PLC activity (344, 369, 644), but the TSH concentrations needed for PLC stimulation are 100-fold higher than those needed for adenylyl cyclase stimulation (643).
The parathyroid gland controls Ca2+ homeostasis through parathyroid hormone (PTH), which enhances Ca2+ (re)absorption in gut and kidney, as well as Ca2+ release from bone. High extracellular Ca2+ concentrations activate the G protein-coupled extracellular Ca2+ sensing receptor (CaR), leading to inhibition of PTH production and secretion in parathyroid cells (for review, see Refs. 74, 268, 661). Activation of the CaR causes a PTX-insensitive (240) stimulation of PLC-β isoforms, with consecutive increments in inositol phosphates, DAG, and intracellular Ca2+ levels (73, 75, 478). In addition, an activation of phospholipases A2 and D (333) as well as a PTX-sensitive suppression of cAMP formation was observed (98). These studies suggested that the CaR couples both to Gq/G11 and Gi/Go family G proteins, and this notion was supported by the finding that CaR activation induced incorporation of radiolabeled GTP into Gαq and Gαi in Madin-Darby kidney (MDCK) cells, which endogenously express low levels of the CaR (25). The fact that increased intracellular Ca2+ levels result in decreased, not increased, hormone release is quite exceptional, and parathyroid cells are, besides renin-secreting juxtaglomerular cells (572), the only endocrine cells showing such inverse coupling. However, the molecular mechanism underlying Ca2+-mediated inhibition of PTH secretion is not understood (for review, see Refs. 74, 86, 268, 661).
D. Regulation of Carbohydrate and Lipid Metabolism
Normal blood glucose levels are maintained both by regulating the activity of enzymes involved in carbohydrate metabolism and by controlling glucose uptake into peripheral tissues. Insulin is a major regulator of both processes, but also a variety of GPCR agonists, like catecholamines and glucagon, contribute to glucose homeostasis. In hepatocytes, activation of Gs-coupled receptors like the β2-adrenergic receptor or glucagon receptors causes PKA-mediated phosphorylation of key enzymes which regulate glycogen synthesis, glycogen breakdown, glycolysis, or gluconeogenesis (167, 168). Together, these changes lead to enhanced hepatic glucose release. Comparable changes can be induced by activation of Gq/G11-coupled receptors like the α1-adrenergic receptor or receptors for vasopressin and angiotensin, and these effects are probably mediated by Ca2+/calmodulin-dependent alteration of enzymatic activity (167, 168). In the periphery, glucose uptake into skeletal muscle and adipocytes is mediated by translocation of glucose transporter subtype 4 (GLUT4) from intracellular vesicles to the plasma membrane (665). A variety of GPCRs have been demonstrated to modify insulin-induced GLUT4 translocation, and even a direct interaction between the insulin receptor and heterotrimeric G proteins was suggested (412). Heterozygous GNAS-deficient mice show, in addition to other metabolic abnormalities (for effects on lipolysis, see below), an increased sensitivity towards insulin, which was attributed to enhanced insulin-dependent glucose uptake into the skeletal muscle (104, 721). In line with this, transgenic expression of a constitutively active mutant of Gαs (GαsQ227L) in fat, liver, and skeletal muscle decreased glucose tolerance (284), suggesting that Gαs-mediated signaling negatively regulates insulin-induced GLUT4 translocation. In contrast, Gi family G proteins seem to facilitate insulin effects. Pretreatment of isolated adipocytes and soleus muscle with PTX results in reduced insulin-stimulated glucose uptake (111, 325), and mice in which Gαi2 was downregulated in liver and adipose tissue using an antisense RNA approach show insulin resistance with hyperinsulinemia, decreased glucose tolerance, and insulin resistance (459–461). In the latter mice, both insulin-induced GLUT4 translocation to the plasma membrane and activation of glycogen synthase and antilipolytic mediators were impaired (461). The decrease in Gαi2 levels was accompanied by an increase in protein tyrosine phosphatase-1B (PTP-1B) activity, an enzyme known to dephosphorylate phosphorylated tyrosine residues on the insulin receptor and on insulin receptor substrate-1. This suggests that the inhibition of insulin signaling in mice with reduced Gαi2 levels is due to disinhibition of PTP-1B (461). Mice expressing a constitutively active mutant of Gαi2 (Gαi2Q205L) in fat, liver, and skeletal muscle displayed reduced fasting blood glucose levels and increased glucose tolerance (103). Adipocytes from these mice showed enhanced insulin-induced glucose uptake and GLUT4 translocation, as well as increased PI 3-kinase and Akt activities (596). In addition, PTP-1B is suppressed in Gαi2 overexpressing mice (626), and streptozotocin-induced diabetic changes are ameliorated (732). Up to now it is not clear at which levels the insulin receptor-mediated pathway interacts with the Gi-mediated pathway. It was suggested that Gi/Go family G proteins physically interact and are phosphorylated by the activated insulin receptor (for review, see Ref. 510). In addition, Gi/Go might be involved in insulin-mediated autophosphorylation of the insulin receptor (351).
Data from 3T3 L1 adipocytes strongly point to an involvement of Gq/G11 family G proteins in basal and insulin-induced GLUT4 translocation. Overexpression of wild-type or constitutively active Gαq increased basal GLUT4 translocation (290, 326), while microinjection of Gαq/Gα11 antibodies or RGS2 protein inhibited insulin-induced GLUT4 translocation (290, 326). In line with these findings, inhibition of GRK2, a negative regulator of Gq/G11 signaling, increased insulin-stimulated GLUT4 translocation, while adenovirus-mediated overexpression of GRK2 reduced translocation as well as 2-deoxyglucose uptake (637). The exact mechanisms underlying Gq/G11-mediated GLUT4 translocation are not fully understood. Several Gq/G11-coupled receptors are able to stimulate glucose uptake via GLUT4 translocation, such as receptors for endothelin-1 (ET-1), norepinephrine, platelet-activating factor, or bradykinin (335, 336, 698). Microinjection of an anti-Gαq/Gα11 antibody or of RGS2 protein causes inhibition of ET-1-induced GLUT4 translocation (289). Of note, chronic ET-1 treatment inhibits insulin-stimulated glucose uptake and GLUT4 translocation in 3T3-L1 adipocytes, and decreased tyrosine phosphorylation of insulin receptor substrates was suggested to mediate this heterologous desensitization (292). The Gq/G11-mediated effect on GLUT4 translocation was shown to be PI 3-kinase dependent in some (289, 290) but not in all studies (59, 326). Recently, a role for the ADP ribosylation factor 6 and the Ca2+-activated tyrosine kinase Pyk2 was suggested (59, 371, 506).
Other examples for G protein-regulated metabolic processes are adipocyte lipolysis and lipogenesis. Activation of Gs-coupled receptors like those for catecholamines, ACTH, glucagons, TSH, or PTH enhances lipolysis via PKA-dependent phosphorylation of hormone-sensitive lipase (274, 401, 606). In adipocytes from patients heterozygous for an inactivating GNAS mutation, the lipolytic effect in response to epinephrine is impaired, and this was suggested to contribute to obesity observed in AHO patients (87). Heterozygous disruption of Gnas exon 2 in mice causes not only enhanced insulin sensitivity (721), but also a variety of other metabolic defects that depend on the parental origin of the inherited mutation (722). While mice with a maternally inherited inactivating mutation of the Gnas gene are obese and hypometabolic, paternally inherited Gnas mutations cause abnormal leanness with decreased serum, liver, and muscle triglycerides, and lipid oxidation in adipocytes is enhanced (104, 722). Since these mice have increased urine norepinephrine excretion, it was suggested that increased sympathetic stimulation of adipocytes caused the changes in lipid metabolism (104, 722). The reason for the restriction to paternally inherited mutations is not completely clear. Deletion of Gnas exon 2 does not only affect the expression of Gαs, but also of alternative gene products like XLαs (Fig. 3) (245, 667). An involvement of XLαs in the hypermetabolic phenotype seems likely for three reasons. First, XLαs is only expressed from the paternal allele (104) and is therefore well suited to explain a paternally inherited phenotype. Second, XLαs-deficient mice show a phenotype similar to paternally exon 2-deficient mice (522). Third, mice with paternal inheritance of a Gnas exon 1 deletion, which does not affect XLαs expression, do not show the respective phenotype (668). Interestingly, studies in XLαs-deficient mice suggest that Gαs and XLαs can exert opposing effects on whole body metabolism (522, 668). In addition to Gs family G proteins, also the Gq/G11 family was implicated in the regulation of lipolysis. Expression of antisense RNA to Gαq in liver and white fat caused hyperadiposity, which was suggested to be attributed to an impaired lipolytic response towards α1-adrenergic agonists (198).
Inhibition of adipocyte lipolysis is mediated by the insulin receptor or by activation of Gi-coupled receptors like the α2-adrenergic receptor, receptors for adenosine or prostaglandin (401) or the recently deorphanized receptor for nicotinic acid, GPR109A (HM74a/PUMA-G) (634). Gαi was suggested to be directly involved in the antilipolytic effect of insulin, since insulin-dependent inhibition of lipolysis and activation of glucose oxidation in adipocytes were shown to be PTX sensitive (219).
IV. IMMUNE SYSTEM
A. Leukocyte Migration/Homing
Directed cell movement in response to an increased concentration of chemoattractant underlies the correct targeting of leukocytes to lymphatic organs during antigen surveillance and also allows them to migrate to sites of infection and/or inflammation (for review, see Refs. 653, 694). Known lymphocyte chemoattractants either belong to the large family of chemokines (355, 500) or to the lysophospholipid family, like sphingosine-1-phosphate (S1P) or lysophosphatidic acid (LPA) (123, 221, 282). Both chemokine and lysophospholipid receptors are GPCRs. While chemokine receptors primarily act through Gi family G proteins (355), lysophospholipid receptors activate Gi, G12/G13, and Gq/G11 family G proteins depending on type and activation state of the respective cell (377, 584, 612).
Inactivation of Gi family G proteins by PTX pretreatment strongly impairs lymphocyte migration in vitro (28, 598) and causes defective homing to spleen, lymph nodes, and Peyer's patches in vivo (31, 124, 597, 662), suggesting that Gi family G proteins are involved in these processes. In line with this, inactivation of Gi by transgenic expression of the S1 subunit of PTX in murine thymocytes resulted in accumulation of mature T cells in the thymus, with greatly reduced levels of T cells in peripheral lymphatic organs (91, 92). To investigate the role of Gi-mediated signaling in more detail, mouse lines carrying inactivating mutations of Gαi subtypes were generated. Both Gαi2 and Gαi3 are expressed in the murine thymus (91), but only inactivation of Gαi2 mimicked the phenotype of PTX expressing mice with respect to thymic accumulation of mature T cells. Neither Gαi2- nor Gαi3-deficient mice showed defects in T cell homing to the periphery (552), suggesting that the homing defects induced by PTX probably result from the combined inactivation of Gi2 and Gi3.
Despite the predominant role of Gi signaling in chemokine-induced lymphocyte migration, an involvement of Gq/G11 family G proteins was suggested by a variety of in vitro studies (11, 21, 353, 590, 716). In contrast to other tissues, hematopoietic cells do not only express the α-subunits Gαq and Gα11, but also Gα15, which corresponds to human Gα16 (17, 683). Mice deficient for Gα11, Gα15, or both Gαq and Gα15 were normal under basal conditions and after antigenic challenge, and only a minor signaling defect in response to complement C5a was found in Gα15-deficient macrophages (132).
The G12/G13 effector RhoA was repeatedly shown to be involved in the regulation of lymphocyte adhesion and migration (367, 631), but direct evidence for an involvement of the G12/G13 family is still lacking. Indirect evidence for a role of G13 in lymphocyte migration comes from studies in Lsc-deficient mice. The murine Rho-specific guanine nucleotide exchange factor Lsc (p115RhoGEF in humans) is expressed exclusively in hematopoietic cells and couples Gα13 to the activation of RhoA (236). Lsc-deficient mice show impaired agonist-induced actin polymerization and motility, as well as abnormal B-cell homing and altered T- and B-cell proliferation (214). Defective migration was also observed in mice lacking the proton-sensing receptor G2A (465), which was shown to couple to Gα13 (316). G2A-deficient macrophages (659) and T cells (533) showed reduced migration towards lysophosphatidylcholine (LPC), while G2A overexpression in a macrophage cell line enhanced migration towards LPC (533, 715). How LPC effects are affected by the absence of G2A is currently not clear (690). In the latter cells, LPC-induced chemotaxis was inhibited by overexpression of dominant negative mutants of Gαq/Gα11 or Gα12/Gα13, as well as expression of RGS domains or GRK2 (specific for Gq/G11) or of p115RhoGEF (specific for G12/G13), while PTX treatment was without effect (715).
Also neutrophils respond to a variety of chemoattractants, such as N-formyl-Met-Leu-Phe (fMLP), C5a, platelet activating factor (PAF), or the chemokine interleukin (IL)-8, with polarization and directed migration. This effect was shown to be PTX sensitive (43, 217, 577, 598) and to be mainly mediated via βγ-subunits (479, 480) (Fig. 8). Lentiviral-mediated knockdown of different G protein subunits in a macrophage cell line revealed that complement C5a-induced migration critically depends on Gβ2, but not on Gβ1, Gαi2, or Gαi3 (286). Intracellular effectors of Gβγ in neutrophils are the γ-isoform of PI 3-kinase (PI3Kγ) and the β2- and β3-isoforms of phospholipase C (PLC-β2, -β3). While neutrophils from mice lacking PLC-β2 and -β3 show normal, or even enhanced, chemotactic responses (388), migration of PI3Kγ-deficient neutrophils was severely impaired (266, 388, 566). Moreover, PI3Kγ-deficient neutrophils failed to accumulate at sites of inflammation in a septic peritonitis model (266), suggesting that PI3Kγ mediates chemotactic responses both in vitro and in vivo. In addition to PI3Kγ, RhoA activity seems to be required for fMLP-induced neutrophil chemokinesis and chemotaxis (review in Ref. 484), and especially the deadhesion of the uropod was attributed to RhoA (12, 397). Interestingly, fMLP receptor-mediated chemotaxis of HL60 cells was recently suggested to involve not only Gi, but also G12/G13 (702). In these cells, activation of Gi defined “frontness” by activating PI3K and Rac, whereas activation of G12/G13 defined “backness” by inducing RhoA-dependent actomyosin interaction (702).
B. Immune Cell Effector Functions
Activation of immune cells by antigen contact initiates complex signaling cascades leading to proliferation, differentiation, or initiation of effector functions like cytokine production, mediator release, phagocytosis, etc. Although these functions are not directly G protein mediated, G proteins might have important modulatory roles (312, 400).
In neutrophils, chemoattractant-induced O2− formation is mediated via Gi-coupled receptors (39) and is strongly impaired both in neutrophils from mice lacking PLC-β2 and -β3 (388, 695, 699) and in PI3Kγ-deficient neutrophils (266, 388, 566). PI3K-mediated formation of phosphatidylinositol 3,4,5-trisphosphate (PIP3) was shown to activate the small GTPase Rac, which contributes to neutrophil migration by actin polymer formation in the leading edge (540, 653) and to O2− formation by NADPH oxidase activation (144). Recently, a new Rac GEF termed P-Rex1 was shown to mediate Rac activation in response to PIP3 and Gβγ in neutrophils (672).
In lymphocytes, the balance between the TH1-driven cellular immunity and TH2-driven humoral response is mainly shaped by the cytokine pattern present during T helper cell activation. PTX has long been known to promote TH1 responses (246, 270, 464), and this was suggested to be due to a negative regulation of IL-12 production by Gi, especially in dendritic cells (246, 279). Mice lacking Gαi2 develop a diffuse inflammatory colitis resembling ulcerative colitis in humans (277, 552), and adenocarcinomas secondary to chronic inflammation were often observed (552). In these mice, levels of proinflammatory TH1-type cytokines and of IL-12 were increased (277), and these changes clearly preceded the onset of bowel inflammation (497, 498). In addition, antigen presenting cells like CD8a+ dendritic cells showed a highly increased basal production of IL-12 (246), suggesting that colitis in these mice is a TH1-driven disease and that production of proinflammatory TH1 cytokines is constitutively suppressed through a Gi-mediated pathway. Accordingly, activation of Gs-mediated signaling opposes these effects. Cholera toxin, which activates Gs by ADP-ribosylation, can be used as an adjuvant to promote TH2 responses (419, 687, 707), and systemic administration of cholera toxin during induction of TH1-based autoimmune diseases can shift the immune response to the nonpathogenic TH2 phenotype (610). In general, the Gs family was suggested to attenuate proinflammatory signaling, but direct evidence is lacking. In vitro, application of cAMP (62, 321, 671) or activation of typically Gs-coupled receptors, like those for vasoactive intestinal polypeptide or PACAP (135, 201), was shown to inhibit immune cell functions. Immune cells from mice lacking the Gs-coupled adenosine A2A receptor respond with enhanced cytokine transcription and NFκB activation to activation of Toll-like receptors (404), and activation of Gs by cholera toxin inhibits signaling in T cells (476, 595) or natural killer cells (678).
The relevance of Gq/G11 and G12/G13 family G proteins in lymphocyte activation and proliferation is unclear. In vitro studies suggested an involvement of both families in the activation of Bruton's tyrosine kinase, a protein that is required for normal B-cell development and activation (42, 309, 402). The Gq/G11 family, especially human Gα16, has been implicated in T-cell activation in a variety of in vitro studies (392, 603, 734), but the physiological relevance of these findings is unclear. In vivo, no obvious immunological defects were observed in Gα11−/−, Gα15−/−, or Gα15−/−;Gαq−/− mice (132). However, after antigenic challenge, Gαq-deficient mice showed impaired eosinophil recruitment to the lung, probably due to an impaired production of granulocyte macrophage colony stimulating factor GM-CSF by resident airway leukocytes (56). Indirect evidence for a role of Gq/G11 and/or G12/G13 in T-cell activation comes from RGS2-deficient mice, which show impaired T-cell proliferation and IL-2 production after T-cell receptor stimulation, as well as defective antiviral immunity in vivo (499). Inactivation of the murine TxA2 receptor, which is typically coupled to Gq/G11 and/or G12/G13 family G proteins (337, 492), caused a lymphoproliferative syndrome due to prolonged T-cell/DC interaction (317). Genetic inactivation of the proton-sensitive G2A receptor (465), which was shown to activate G12/G13, causes a late-onset autoimmune syndrome in mice (372), suggesting that these G proteins may be involved in the regulation of lymphocyte activation.
V. NERVOUS SYSTEM
G proteins play multiple roles in the nervous system as most neurotransmitters also activate G protein-coupled metabotropic receptors to modulate neuronal activity. In contrast to ionotropic receptors, GPCRs that are present pre- and postsynaptically mediate comparatively slow responses. Only a few well-studied examples are described below.
A. Inhibitory Modulation of Synaptic Transmission
The regulation of neurotransmitter release at presynaptic terminals is an important mechanism underlying the modulation of synaptic transmission in the nervous system. Inhibitory regulation of neurotransmitter release is mediated by various GPCRs like α2-adrenoceptors, μ- and δ-opioid receptors, GABAB receptors, adenosine A1, or endocannabinoid CB1-receptors. These receptors have in common that they couple to G proteins of the Gi/Go family. A major mechanism by which these G proteins mediate the inhibition of transmitter release is the inhibitory modulation of the action potential-evoked Ca2+ entry to the presynaptic terminal which is required to trigger neurotransmitter release. N- and P/Q-type calcium channels that are concentrated at nerve terminals as well as R-type calcium channels have been shown to be inhibited via Gi/Go-coupled receptors (145). This inhibition is due to the interaction of the G protein βγ-complex with the α1-subunit of Cav2.1–2.3 (89, 420). The β-subunit of the channel as well as strong depolarization can reduce this inhibition (61, 439). Most Gβγ combinations are similarly effective in inhibiting channels of the Cav2 family (202, 288, 557). There is some evidence that βγ-mediated inhibition of Cav2 channels is not the only mechanism through which GPCRs mediate inhibition of neurotransmitter release (449). Also, G protein-coupled inwardly rectifying K+ channels (GIRKs) are localized at presynaptic terminals; however, their physiological role in regulation of transmitter release from presynaptic terminals is less clear (156, 524). GIRKs are well-established effectors for G protein βγ-subunits (113, 356, 398, 718), and most Gβγ combinations appear to be similarly effective in activating GIRKs (681, 708). There is also increasing evidence that the Gβγ-mediated inhibition of neurotransmitter release at the presynaptic terminus involves mechanisms downstream of the regulation of Ca2+ (51). This may involve the direct interaction between Gβγ and the core vesicle fusion machinery as suggested by the observation that Gβγ can directly bind to SNARE proteins like syntaxin and SNAP25 as well as cysteine string protein (CSP) (50, 51, 305, 410). Evidence exists that presynaptic Ca2+ channels, syntaxin1, and the α-subunit of Go are components of a functional complex at the presynaptic nerve terminal release site (384, 602).
Both Gi-type G proteins as well as Go are highly abundant in the nervous system. Mice lacking the α-subunit of Go are smaller and weaker than their littermates and have a greatly reduced life expectancy (308, 639). In addition, these animals have tremors and occasional seizures, and they show an increased motor activity with an extreme turning behavior. Gαo-deficient mice have also been shown to be hyperalgesic (308). When neuronal cells from Gαo-deficient mice were analyzed by electrophysiological methods for the regulation of GIRKs and voltage-dependent Ca2+ channels through GPCRs, it was found that the recovery kinetics after agonist washout were much slower in the absence of Gαo. However, current modulation via various receptors was as effectively as in wild-type cells (225, 308). This indicates that other G proteins especially Gi-type G proteins that are activated by the same receptors can compensate for Go deficiency.
B. Modulation of Synaptic Transmission by the Gq/G11-Mediated Signaling Pathway
In the nervous system, the G proteins Gq and G11 are widely expressed (622), and they are involved in multiple pathways that modulate neuronal function. The modulation of synaptic transmission has best been described at the parallel fiber (PF)-Purkinje cell (PC) synapse in the cerebellum. At the PF-PC synapse, Gq/G11 mediate the effects of metabotropic glutamate group 1 receptors (mGluR1). The Gq/G11-mediated IP3-dependent transient increase in the dendritic Ca2+ concentration, which does not require changes in the membrane potential (178, 621), is important for the induction of long-term depression (LTD) at the PF-PC synapse (452). LTD in turn is believed to be one of the cellular mechanisms underlying cerebellar motor learning (66). Interestingly, while Gαq-deficient mice develop an ataxia with clear signs of motor coordination deficits, Gα11-deficient animals show only very subtle defects in motor coordination (237, 489). A detailed comparison of both mouse lines showed that Ca2+ responses to mGluR1 activation were absent in Gαq-deficient mice, whereas they are indistinguishable between wild-type and Gα11-deficient animals (237). However, synaptically evoked LTD was decreased in Gα11-deficient animals, but to a lesser extent than in Gαq-deficient mice. The predominant role of Gq in Purkinje cells can be explained by the higher expression compared with Gαq (489). In fact, quantitative single-cell RT-PCR analysis showed that Purkinje cells express at least 10-fold more Gαq than Gα11 (237).
Lack of Gαq also results in a defect in the regression of supernumerary climbing fibers innervating Purkinje cells in the third postnatal week. This process is believed to be due to a defect in the modulation of the PF-PC synapse. Similar cerebellar phenotypes as in Gαq-deficient mice have been described in mice lacking the mGluR1 (9, 323) as well as in mice lacking the β4-isoform of PLC, which is predominantly expressed in the rostral cerebellum (324, 453). Interestingly mGluR1, Gαq, and PLC-β4 can be found colocalized in dendritic spines of PCs (324, 622, 664), suggesting that they are components of a signaling cascade involved in the modulation of the PF-PC synapse.
Interestingly, hippocampal synaptic plasticity is equally impaired in Gαq- and Gα11-deficient mice (451). However, mGluR-dependent long-term depression in the hippocampal CA1-region was absent in Gαq-deficient mice but appeared to be unaffected in mice lacking Gα11 (340). Similarly, suppression of slow afterhyperpolarizations via muscarinic and metabotropic glutamate receptors was nearly abolished in Gαq-deficient mice but was unchanged in mice lacking Gα11 (350).
Besides the voltage-dependent, G protein-mediated inhibition of calcium channels via Gβγ (see above), functional studies in calyx-type nerve terminals and in sympathetic neurons have identified a voltage-insensitive inhibition of calcium currents via metabotropic receptors that is believed to involve Gq/G11 (322, 449). How Gq/G11 activation can lead to inhibition of voltage-dependent calcium channels is not clear. However, recent evidence suggests that the depletion of phosphatidylinositol 4,5-bisphosphate (PIP2), the substrate of PLC plays a role (200).
There is also evidence that activation of Gq/G11-mediated signaling on the postsynaptic site is involved in the induction of retrograde signaling via the endocannabinoid system. Gq/G11-coupled receptors like group I mGluRs as well as muscarinic M1 receptors can lead to the activation of the retrogradely acting cannabinoid system by stimulating the formation of endocannabinoids (136, 189, 380, 409).
C. Roles of Gz and Golf in the Nervous System
Gs is the ubiquitous G protein that couples receptors in a stimulatory fashion to adenylyl cyclases. However, in a few tissues the G protein Golf is the predominant G protein that couples receptors to adenylyl cyclase. Apart from the olfactory system (see below), Gαolf expression levels exceed those of Gαs also in a few other defined brain regions like the nucleus accumbens, the olfactory tubercle, and the striatum (40, 120, 737). In the striatum, Golf appears to be critically involved in dopamine (D1) and adenosine (A2) receptor-mediated effects (258, 737). Data from various laboratories show that in the striatum, the dopamine D1 receptor, Gαolf and adenylyl cyclase type V as well as the G protein γ7-subunit are coexpressed. Strikingly, mice lacking either of these signaling components show clear signs of motor abnormalities (40, 298, 573, 703). Mice lacking adenylyl cyclase V show an attenuated D1-receptor/Golf-mediated adenylyl cyclase activation in the striatum (298). Gγ7-deficient mice have strongly reduced levels of striatal Gαolf, and activation of adenylyl cyclase via dopamine receptors is abolished in the striatal cells (573). Thus, in rodents, striatal D1 receptors appear to activate adenylyl cyclase V through a G protein containing Gαolf and γ7. In humans, it has recently been shown that Gαolf expression is markedly diminished in the putamen of patients with Huntington disease, while the putamen of patients with Parkinson disease showed significantly increased levels of Gαolf and Gγ7 (120).
The G protein Gz is a member of the Gi/Go family. It shares with other Gi/Go family members the ability to inhibit adenylyl cyclase (176, 267). It is expressed primarily in brain, retina, adrenal medulla, and platelets. Gαz-deficient mice are viable and have no major phenotypical abnormalities. However, Gαz deficiency results in an abnormal response to certain psychoactive drugs (253, 713). This includes a considerably pronounced cocaine-induced increase in locomotor activity as well as a reduction in the short-term antinociceptive effects of morphine (713). In addition, Gαz-deficient animals develop significantly increased tolerance to morphine (374). Interestingly, the behavioral effects of catecholamine reuptake inhibitors that are used as antidepressant drugs were abolished in mice lacking Gαz (713). While mice lacking Gαz clearly indicate that Gz plays a role in the nervous system and is involved in various drug responses, the function of Gz on a cellular level in the nervous system remains unclear.
VI. SENSORY SYSTEMS
G protein-mediated signal transduction processes mediate the perception of many sensory stimuli. Odors, light, and especially sweet and bitter taste substances act directly on GPCRs that modulate the activity of primary sensory cells via often very specialized heterotrimeric G proteins.
A. Visual System
The human retina contains two types of photoreceptor cells, rods and cones. While rods mediate mainly achromatic night vision and contain one type of GPCR, rhodopsin, cones are responsible for chromatic day vision and contain three GPCRs (opsins), which are sensitive to different parts of the visible light spectrum. Rhodopsin and opsins are coupled to different but closely related heterotrimeric G proteins. Rod-transducin (Gt-r) and cone-transducin (Gt-c), which mediate the effects of rhodopsins and opsins, respectively, couple these receptors in a stimulatory manner to cGMP-phosphodiesterase (PDE) by binding and sequestering the inhibitory γ-subunit of the retinal type 6 PDE (PDE6). Activation of PDE lowers cytosolic cGMP levels leading to a decreased open probability of cGMP-regulated cation channels in the plasma membrane, which eventually results in hyperpolarization of the photoreceptor cells (24). In mice lacking Gαt-r, the majority of retinal rods does not respond to light any more, and these animals develop mild retinal degeneration with increasing age (84). To ensure an adequate time resolution of the light signal, the transducin-mediated signaling process initiated by light-induced receptor activation requires an efficient termination mechanism. At least three proteins contribute to the rapid deactivation of transducin by increasing its GTPase activity: RGS9 and Gβ5L, which form a complex, and the γ-subunit of the transducin effector cGMP-PDE (24). In mice lacking Gβ5, the levels of RGS9 and other G protein γ-like (GGL) domain RGS proteins in various tissues including retina are drastically reduced, suggesting that the formation of the RGS9 Gβ5 complex is required for expression and function of RGS9 (100). Animals lacking Gβ5 or RGS9 show a strongly impaired termination of transducin-mediated signaling (99, 352, 407).
The light-induced hyperpolarization of photoreceptor cells results in a decreased release of glutamate. Glutamate released from photoreceptor cells acts on two classes of second-order neurons in the retina, one that depolarizes in response to glutamate most likely via ionotropic glutamate receptors (OFF bipolar cells) and another that hyperpolarizes (ON bipolar cells). ON bipolar cells are in the absence of light inhibited by glutamate released from rods and cones, and this effect is mediated by the metabotropic glutamate receptor mGluR6. Light-induced hyperpolarization of rods results in decreased glutamate release and disinhibition of ON bipolar cells due to a decreased activation of mGluR6. In mice lacking mGluR6 or the Gαo splice variant Gαo1, the modulation of ON bipolar cells in response to light is abolished (142, 143, 425), indicating that Go is the principle mediator of glutamate-induced inhibition of ON bipolar cells, which occurs especially in the absence of light. The retina-specific RGS protein Ret-RGS1 is localized in the dendritic tips of ON bipolar cells together with mGluR6 and Gαo1 and may play a critical role in increasing the deactivation of Go1 resulting in an acceleration in the rising phase of the light response of the ON bipolar cells. This mechanism has been suggested to match the kinetics of ON bipolar cell activation to that of the OFF bipolar cells that arises directly from ligand-gated channel activation by glutamate (141).
B. Olfactory/Pheromone System
Chemosensation by the olfactory system is based on the expression of a huge variety of GPCRs specifically in the olfactory epithelium. Individual olfactory sensory neurons appear to often express only one olfactory receptor, and olfactory sensory neurons expressing one receptor type converge to the same subgroup of glomeruli in the olfactory bulb (457, 548). Rodents have ∼1,000 different olfactory receptors, whereas the number in humans is considerably smaller and has been estimated to be ∼350 (457). Despite this remarkable diversity on the level of the receptor, the olfactory GPCRs appear to employ the same G protein-mediated signal transduction pathway in olfactory sensory neurons. The G protein Golf is centrally involved in signaling by olfactory receptors in response to odorant stimuli, and Gαolf-deficient mice are anosmic exhibiting dramatically reduced electrophysiological responses to all odors tested (40). Golf, a G protein related to Gs, couples olfactory receptors in the cilia to adenylyl cyclase III, resulting in the increased formation of cAMP. cAMP then activates a cyclic nucleotide-gated (CNG) cation channel consisting of three different subunits, CNGA2, CNGA4, and CNGB1. The increase in cellular Ca2+ activates a Ca2+-activated Cl− channel that further depolarizes the cell membrane (457, 548). Most Gαolf-deficient pups die a few days after birth due to insufficient feeding. Rare surviving animals exhibit inadequate maternal behavior resulting in the death of all pups born to Gαolf-deficient mothers. This indicates that normal nursing and mothering behavior in rodents is greatly dependent on an intact olfactory system (40). The fundamental role of this signaling pathway is underscored by the anosmic phenotype found not only in mice lacking Gαolf, but also adenylyl cyclase (693) as well as in mice lacking the subunits of the olfactory CNG channel (30, 76, 731).
A second olfactory system called the accessory olfactory system or the vomeronasal system exists in most mammals (152, 330). The peripheral sensory structure of this system, the vomeronasal organ, is localized at the bottom of the nasal cavity. The vomeronasal system responds to pheromones that mediate defined effects on individuals of the same species and modulate social, aggressive, reproductive, and sexual behaviors. In the vomeronasal organ, two families of GPCRs, which in mice consist of ∼150 members each, have been identified. The V1 receptor family is expressed in vomeronasal sensory neurons together with the G protein Gi2, whereas the V2 receptor family is expressed in a different population of neurons that coexpresses the G protein Go (428). In humans, no intact genes coding for V2 family receptors have been found, and most V1 receptor genes are pseudogenes, suggesting that the function of the vomeronasal organ of rodents is not fully preserved in humans and higher primates. Similar to the olfactory receptors, individual members of the V1 and V2 receptor families appear to be specifically expressed in individual sensory neurons that project to a small group of glomeruli in defined regions of the accessory olfactory bulb. The striking coexpression of V1 receptor family members and Gi2 and of V2 receptor family members and Go suggests that these G proteins mediate the cellular effects induced by activation of the respective pheromone receptors. Consistent with this, Gαi2-deficient mice have a reduced number of vomeronasal sensory neurons that normally express Gαi2 and show alterations in the behaviors for which an intact vomeronasal organ is believed to be required like maternal aggressive behavior as well as aggressive behavior in males exposed to an intruder (486). In mice lacking Gαo apoptotic death of vomeronasal sensory cells which usually express Gαo has been observed (623). Despite these behavioral and anatomical abnormalities in Gαi2- and Gαo-deficient mice, a direct involvement of Gαi2 in signaling of V1 receptor-expressing cells and of Gαo in V2 receptor-expressing cells has not been demonstrated so far. The transient receptor potential channel TRP2, which is expressed in vomeronasal sensory neurons, has been identified as a critical downstream mediator of the signal transduction pathway in vomeronasal sensory neurons (383, 605).
C. Gustatory System
Unlike the olfactory system, chemosensation by the gustatory system involves only in part G protein-mediated signal transduction mechanisms. The taste qualities sweet, bitter, and amino acid (umami) signal through GPCRs, whereas salty and sour tastants act directly on ion channels (391). During recent years, two families of candidate mammalian taste receptors, T1 receptors and T2 receptors, have been implicated in sweet, umami, and bitter detection. The T2 receptors are a group of ∼30 GPCRs that are specifically expressed in taste buds of the tongue and that are linked to bitter taste in mice and humans (6, 78, 94, 429). The T1 receptor family consists of only three GPCRs and has been shown to form heterodimers. T1R1 and T1R3 form a receptor that responds to amino acids carrying the umami taste (477), and T1R2 forms heterodimers with T1R3 that functions as a sweet receptor (386, 477). Studies in knockout mice support a role of T1R2/T1R3 as a sweet receptor and of T1R1/T1R3 as an umami receptor. In addition, T1R3 may form homodimers that can also mediate some sweet tastes, and there is evidence that also a splice variant of the mGluR4 glutamate receptor may be involved in umami sensation (95, 96, 127, 730). The signal transduction mechanisms used by different G protein-coupled taste receptors are less clear. Gustducin, a G protein mainly expressed in taste cells, is believed to be able to couple receptors to phosphodiesterase resulting in a decrease of cyclic nucleotide levels. Mice lacking the α-subunit of gustducin show impaired responses to bitter, sweet, as well as umami tastes (247, 555, 692). However, the residual bitter and sweet taste responses in Gαgust-deficient mice and the finding that expression of a dominant negative mutant of α-gustducin reduces this responsiveness even further indicates that α-gustducin is not the only α-subunit involved in sweet, bitter, and umami signal transduction (416, 556). In addition, α-gustducin-deficient mice expressing the α-subunit of rod-transducin as a transgene driven by the α-gustducin promoter partially recovered responses to sweet and bitter compounds (247). However, in mice lacking the α-subunit of rod-transducin, responses to bitter and sweet compounds were normal, whereas the responses to various umami compounds were impaired (249). Thus gustducin is involved in sweet, bitter, and umami taste detection, whereas rod-transducin mediates part of umami taste sensation. Both gustducin and transducin are able to activate phosphodiesterase, thereby leading to decreased cGMP levels, disinhibition of cyclic nucleotide-inhibited channels, and calcium influx. In addition, bitter tastants have also been shown to activate PLC-β2 through Gβγ subunits (probably Gβ3γ13) released from gustducin or other G proteins (416), and PLCβ2 has recently been shown to be involved in the activation of transient receptor potential M5 (TRPM5) cation channels, which cause depolarization of the cells (729). Interestingly, mice lacking PLC-β2 or TRPM5 are completely insensitive not only to bitter tastants but also to sweet and umami (729), indicating that the G protein/PLC-β2/TRPM5 signaling cascade is centrally involved in the sensation of these three taste qualities. However, other signal transduction pathways have been suggested, and it is not clear how gustducin or transducin functions are related to those of PLCβ2 and TRPM5.
Although heterotrimeric G proteins are expressed throughout the prenatal development of the mammalian organism, only a limited number of studies have so far addressed the role of G proteins in development. Most insights into the developmental role of G protein-mediated signaling pathways came from studies on mouse mutants lacking individual G protein α-subunits. Some null mutations like those of the genes encoding Gαs, Gα13, or Gαq/Gα11 are embryonic lethal (493, 495, 723) and clearly indicate an essential role during development.
A. G13-Mediated Signaling in Embryonic Angiogenesis
Both G12 and G13 have been shown to induce cytoskeletal rearrangements in a Rho-dependent manner (79, 563). Lack of Gα13 in mice results in embryonic lethality at midgestation. At this stage, mouse embryos express both Gα12 and Gα13. Gα13-deficient mouse embryos show a defective organization of the vascular system, which is most prominent in the yolk sac and in the head mesenchyme (493). Vasculogenic blood vessel formation through the differentiation of progenitor cells into endothelial cells was not affected by the loss of Gα13. However, angiogenesis which includes sprouting, growth, migration, and remodeling of existing endothelial cells, was severely disturbed, and the maintenance of the integrity of newly developed vessels appeared to be defective in Gα13-deficient embryos. Chemokinetic effects of thrombin, which acts through protease-activated receptors (PARs), were completely abrogated in fibroblasts lacking Gα13, indicating that Gα13 is required for full migratory responses of cells to certain stimuli. Interestingly, approximately one-half of the embryos that lack the protease-activated receptor 1 (PAR-1) also die at midgestation with bleeding from multiple sites (118). This phenotype of embryos lacking PAR-1 which is expressed in endothelial cells, can be rescued by a PAR-1 transgene whose expression is driven by an endothelial-specific promoter (226). This clearly indicates that PAR-1 function is required for proper vascular development. The more severe embryonic defect of Gα13 compared with PAR-1-deficient embryos suggests that Gα13 function is not restricted to protease-activated receptor signaling.
The defects observed in Gα13-deficient embryos and cells occurred in the presence of Gα12, and loss of Gα12 did not result in any obvious defects during development. Interestingly, Gα12-deficient mice that carry only one intact Gα13 allele also die in utero (228). This genetic evidence indicates that Gα13 and its closest relative, Gα12, fulfill at least partially nonoverlapping cellular and biologic functions, which are required for proper development.
B. Gq/G11-Mediated Signaling During Embryonic Myocardial Growth
The Gαq/Gα11-mediated signaling pathway appears to play a pivotal role in the regulation of the physiological myocardial growth during embryogenesis. This is demonstrated by the phenotype of mice lacking both Gαq and Gα11. These mice die at embryonic day 11 due to a severe thinning of the myocardial layer of the heart (495). Both the trabecular ventricular myocardium as well as the subepicardial layer appeared to be underdeveloped. There are several Gq/G11-coupled receptors that may be involved in the regulation of cardiac growth at midgestation. Inactivation of the gene encoding the Gq/G11-coupled serotonin 5-HT2B receptors in mice resulted in cardiomyopathy with a loss of ventricular mass due to a reduction in the number and size of cardiomyocytes (473), and lack of both endothelin A (ETA) and endothelin B (ETB) receptors, which can signal through Gq/G11, resulted in midgestational cardiac failure (710). Most likely there is some degree of signaling redundancy with several inputs into the Gq/G11-mediated signaling pathway, and only deletion of both the Gαq and the Gα11 gene results in severe phenotypic defects during early heart development. Interestingly, one intact allele of the Gαq or the Gα11 gene was sufficient to overcome the early developmental block in heart development. However, newborn mice that have only one intact Gαq or Gα11 allele show an increased incidence of cardiac defects ranging from septal defects to univentricular hearts (495).
C. Neural Crest Development
Signaling through Gq/G11 has been implicated in the proliferation and/or migration of neural crest cells. ET-1 and the Gq/G11-coupled endothelin A (ETA) receptor are essential for normal function of craniofacial and cardiac neural crest. ET-1 and ETA receptor-deficient mice die shortly after birth due to respiratory failure (114, 357, 358). Severe skeletal abnormalities could be observed in their craniofacial region, including a homeotic transformation of mandibular arch-derived structures into maxillary-like structures as well as absence of auditory ossicles and tympanic ring (503, 553). A hypomorphic phenotype similar to, but less severe than ETA or ET-1 null mice could be observed in Gαq (−/−);Gα11 (−/+) mice (495). In Gαq/Gα11 double-deficient mice studied at embryonic day 9.5, a expression of ET-1-dependent transcription factors like Dlx3, Dlx6, dHAND, and eHAND was ablated (297), a finding similar to those made in mice lacking ETA or ET-1 (114, 629). This suggests that in the neural crest-derived pharyngeal arch mesenchyme, a signaling pathway involving ET-1, ETA receptors, and Gαq/Gα11 is operating. ET-1, which is expressed in the pharyngeal endoderm binds to ETA receptors in neural crest cells of the pharyngeal archs of embryonic day 9.5 embryos (114). Gq/G11 then couples ETA receptors via phospholipase C-β activation to the activation of the expression of genes encoding transcription factors including Dlx3, Dlx6, dHAND and eHAND (297). In contrast, Gαq (−/+);Gα11 (−/−) mice did not show any craniofacial abnormalities indicating that ETA receptor-mediated neural crest development requires a certain amount of Gαq/Gα11. It is also possible that Gα11 is expressed at lower levels than Gαq in the neural crest-derived mesenchyme of the pharyngeal archs, or that the ETA receptor couples preferentially to Gαq.
The ET-3 and ETB receptor system has been shown to be involved in the development of neural crest cells taking part in the formation of epidermal melanocytes as well as the myenteric ganglia of the distal colon. In mice lacking ET-3 or the ETB receptor, this results in white-spotted hair and skin color as well as a dilation of the proximal colon (38, 710). These defects are very similar to those present in humans suffering from multigenic Hirschsprung disease, which in various cases has been shown to be caused by mutations in ETB or ET-3 genes (158, 204). Whether Gq/G11-mediated signaling is involved in ET-3-induced differentiation of cutaneous and enteric neural crest cells has not been studied directly so far. However, several hypermorphic alleles of the Gαq and Gα11 genes have recently been found in mouse mutants with an aberrant accumulation of pigment-producing melanocytes (302, 642). Genetic evidence was provided that the action of Gαq/Gα11 depended on the ETB receptor.
VIII. CELL GROWTH AND TRANSFORMATION
During recent years it has become obvious that GPCRs and heterotrimeric G proteins play important roles in the regulation of cell growth and that some G proteins can induce cellular transformation (140, 229, 549). Many G protein α-subunits are able to transform cells under certain conditions when rendered constitutively active by inducing GTPase inactivating mutations. In some cases, activated G protein α-subunits have been found in human tumors.
A. Constitutively Active Mutants of Gαq/Gα11 Family Members
Transforming mutants of Gαq/Gα11 have not been found in human tumors. However, their potential oncogenic function has been studied in cell lines by expression of constitutively active forms like GαqQ209L or Gα11Q209L. These studies indicated that activated Gαq/11 can lead to transformation of fibroblasts when expressed at low levels (318) but induces growth inhibition and apoptosis when expressed at higher levels (140). Interestingly, various Gq/G11-coupled receptors have been shown to possess highly transforming activity. This has very early been shown for the 5-HT2C serotonin receptor as well as for the M1, M3, and M5 muscarinic receptors, which are able to transform fibroblasts in the presence of a receptor agonist (231, 315). It is well known that various Gq/G11-coupled neuropeptide receptors like those for gastrin-releasing peptides, galanin, neuromedin B, vasopressin, and others are involved in the autocrine and paracrine stimulation of proliferation in small cell lung cancer cells (250). This suggests that endogenous Gq/G11-coupled receptors can be tumorigenic in the presence of excess ligands. In vitro experiments indicate that constitutively active Gq/G11-coupled receptors can enhance mitogenesis and tumorigenicity (14), and the virally encoded KSHV-G protein-coupled receptor, which is a constitutively active Gq/G11-coupled receptor, has been shown to behave as a viral oncogene and angiogenesis activator and appears to be involved in Kaposi sarcoma progression (26, 29, 458).
B. The Oncogenic Potential of Gαs
Gs-mediated increases in cAMP can induce growth inhibition in some tissues while in others it can have a transforming potential. The gsp oncogene encodes a GTPase-deficient mutant of Gαs and has been found in up to 30% of thyroid toxic adenomas and in a few thyroid carcinomas. A constitutively active mutant of Gαs has also been found in GH-secreting pituitary adenomas as well as in the McCune-Albright syndrome (172, 599). Interestingly, responses to constitutively active Gαs are cell type specific. In some cells, an increase in cAMP and consecutive activation of PKA can inhibit Raf kinase and prevent transformation of cells (102). In contrast, in various neuronal and neuroendocrine cells, Gαs-mediated cAMP formation activates cell growth (164, 551). The transforming potential of Gs obviously depends on the cellular context that links Gs-mediated cAMP formation to inhibition or stimulation of ERK (604). While various kinases upstream of ERK have been involved in cAMP/PKA-induced ERK regulation, the determinants of the net proliferative effect of cAMP remain elusive (604). A proposed role of the cAMP-activated Epac/Rap1 pathway in cAMP-mediated ERK activation has been questioned (163).
The potential of Gs-mediated signaling to induce tumor formation in endocrine cells is also underlined by the fact that constitutively active mutants of various Gs-coupled receptors have been detected in human tumors. Up to 80% of hyperfunctioning human thyroid adenomas and a minority of differentiated thyroid carcinomas have been shown to contain constitutively active TSH receptors (507, 508). Similarly, mutations resulting in constitutive activation of the luteinizing hormone receptor have been shown to cause hyperplastic growth of Leydig cells in a form of familial male precautious puberty (578) as well as in Leydig cell tumors (393).
C. Gi-Mediated Cell Transformation
The gip2 oncogene that encodes an active mutant of Gαi2 has been found in adrenal cortical tumors as well as in human ovarian sex cord stromal tumors (406). It is a matter of debate how frequent these Gαi2 mutations occur in these tumors. Expression of constitutively active Gαi2 in fibroblast cell lines leads to cell transformation, an effect which has not been completely understood yet, but may result from the derepression of the Ras/ERK pathway in response to a decrease in cellular cAMP levels (446, 640). Depending on the cellular system, Gi-mediated release of free βγ-subunits may also be involved in the growth-promoting effect of Gαi2 (122).
D. Cellular Growth Induced by Gα12/Gα13
The G protein α-subunit Gα12 was identified as an oncogene present in soft tissue sarcoma (93). This oncogene, termed gep, encoded the wild-type form of Gα12. At the same time, the potent transforming ability of constitutively active mutants of Gα12 and Gα13 could be shown in various systems (307, 645, 655, 703, 704). While no activating mutation of Gα12 or Gα13 has been found in human tumors, increased expression levels of both proteins have been detected in various human cancers (230). Both Gα12 and Gα13 can induce activation of the small GTPase RhoA via regulation of a subgroup of RhoGEF proteins (195, 236). Interestingly, RhoA has been found to be overexpressed in various tumor types (2, 192, 320), and Rho GTPases have been involved in carcinogenesis and cancer progression (385, 564). Recently, an interesting link between G12/G13- and cadherin-mediated signaling was described (327, 434, 435). Active Gα12 and Gα13 can interact with the cytoplasmic domain of some type I and type II class cadherins, like E-cadherin, N-cadherin, or cadherin-14, causing the release of β-catenin from cadherins and its relocalization to the cytoplasm and nucleus, where it is involved in transcriptional activation. In addition, activated Gα12 can block cadherin-mediated cell adhesion in various cells including breast cancer cells (434). This downregulation of cadherin-mediated cell adhesion and the release of β-catenin from cadherins may be a mechanism by which Gα12/Gα13 induces cellular transformation and metastatic tumor progression.
IX. CONCLUDING REMARKS
The field of heterotrimeric G proteins was launched more than 30 years ago when the involvement of guanine nucleotides in the hormonal stimulation of adenylyl cyclase was discovered. Since then, the field has enormously expanded, and G protein-mediated signal transduction has turned out to be the most widely used transmembrane signaling system in higher organisms. It consists of hundreds of receptors that signal to dozens of G proteins and effectors. The modular nature of this signal transduction system with multiple components enables cells to assemble vast, combinatorially complex signaling units that allow different cells in different contexts to respond adequately to extracellular signals. There is probably not a single cell in a mammalian organism that does not employ several G protein-mediated signaling pathways. Often these pathways integrate the information conveyed by several receptors recognizing different ligands. Only in recent years did we gain a more systematic insight into the role of individual G proteins on a cellular and supracellular level. However, many aspects of G protein-mediated signaling, especially under in vivo conditions, still remain to be elucidated. While many components of the system have been identified, new approaches are required to determine the exact composition of individual signaling units and to define their exact cellular localization. This will probably lead to the identification of even more proteins and nonprotein factors that modulate G protein-mediated signaling. An increasing use of in vivo models is necessary to test the significance of signaling mechanisms described in vitro under normal conditions as well as in disease states. The parallel application of genetic, genomic, and of new proteomic approaches will be required to continue to define how the G protein-mediated signaling system works on a molecular, cellular, and systemic level. Such an integrated view will provide the basis for a full understanding of the physiological and pathophysiological role of G protein-mediated signaling and will allow the full exploitation of this multifaceted signaling system as a target for pharmacological interventions.
Address for reprint requests and other correspondence: S. Offermanns, Institute of Pharmacology, Univ. of Heidelberg, Im Neuenheimer Feld 366, D-69120 Heidelberg, Germany (E-mail:).
- Copyright © 2005 the American Physiological Society