Oncogene (2001) 20, 1635 ± 1642 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

In vivo functions of heterotrimeric G-proteins: studies in Ga-de®cient mice

Stefan O€ermanns*,1

1Pharmakologisches Institut, UniversitaÈt Heidelberg, Im Neuenheimer Feld 366, 69120 Heidelberg, Germany

Heterotrimeric guanine nucleotide binding proteins (G- heterotrimeric G-proteins are not simple signal trans- proteins) mediate the e€ects of numerous hormones, ductors but that they represent a central part of a neurotransmitters or sensory stimuli by coupling their sophisticated molecular machine which is able to transmembranous receptors to various e€ectors like receive, to integrate and to process information carried enzymes and ion channels. Changes in the activity of by extracellular signals. (For recent reviews on these e€ector molecules eventually lead to the regulation heterotrimeric G-proteins see Neer, 1995; Gudermann of multiple cellular functions ranging from short term et al., 1996; Fields and Casey, 1997; Hamm, 1998; regulatory processes like the control of secretion rates, Knall and Johnson, 1998; Wilkie, 2000). muscle tonus or metabolic processes to long term e€ects G-protein-mediated signaling is involved in many like regulation of growth and di€erentiation. Hetero- physiological and pathological processes. The function trimeric G-proteins play a pivotal role in this transmem- of G-proteins has mainly been studied in cell-free brane signaling process as they take part in processing systems or on the cellular level while little is known and sorting of incoming signals as well as in adjusting about their role on a more complex level. To analyse the sensitivity of the system. This review describes some G-protein functions in the context and under the of the new insights into the biological role of G-protein condition of the intact organism, almost all known mediated signaling processes provided by the analysis of genes encoding G-protein a-subunits have been mice genetically engineered to lack distinct G-protein a- inactivated by gene targeting in mice with often subunits. Oncogene (2001) 20, 1635 ± 1642. unexpected results (Table 1). So far, no mouse line was reported carrying targeted mutations of Gb-or Keywords: G-protein; knock-out; transgenic mouse; Gg-genes. This review summarizes recent ®ndings in signal transduction Ga-de®cient mice by focusing on those systems and functions which have been reported to be mainly a€ected by inactivating mutations of murine Ga Introduction genes.

All cells of the mammalian organism express G- protein-coupled receptors as well as several types of Central nervous system heterotrimeric G-proteins and e€ectors. G-proteins consist of a-, b-andg-subunits and couple activated Most neurotransmitters of the central nervous system receptors to e€ector proteins. The main properties of (CNS) act on G-protein coupled receptors to modulate individual G-proteins appear to be primarily deter- neuronal activity. The receptors are found pre- and mined by the identity of the a-subunit. More than 20 postsynaptically and mediate relatively slow responses. G-protein a-subunits have been described in the Inhibitory modulation is mostly mediated by coupling mammalian system, and they can be devided into four of receptors to members of the Gi/o family whereas Gq- subfamilies based on structural and functional homo- and Gs-family members are primarily involved in logies (Table 1; Simon et al., 1991). Five G-protein b- excitatory responses. subunits and 11 g-subunits have been found so far in The G-protein Go is highly abundant in the the mammalian system. The bg-complex has long been mammalian nervous system where it constitutes up to regarded as a more passive partner of the G-protein a- 0.5% of membrane proteins. Expression has also been subunit. During the last years it has become clear that shown in neuroendocrine cells as well as at low levels bg-complexes also play a very important role in the in the heart. On a cellular level, Go has been regulation of various e€ectors (Clapham and Neer, demonstrated to mediate inhibition of neuronal (N-, 1997). It is not clear how the structural and functional P/Q-, R-type) Ca2+ channels through the interaction of complexity of the G-protein-signaling system leads to a its bg complex with the a1 subunit of the channel speci®c and adequate response to extracellular stimuli (Zamponi and Snutch, 1998). This process has been on the cellular level. It has, however, become implicated in the negative feedback regulation via increasingly obvious during the last years that presynaptic neuronal Ca2+ channels. In growth cone membranes, Go makes up 10% of the membrane protein (Strittmatter et al., 1990). This and the fact that *Correspondence: S O€ermanns GAP43, another protein enriched in growth cones and Studies in Ga-deficient mice S Offermanns 1636 Table 1 Mammalian G-protein a-subunits and phenotypes resulting from homozygous inactivating mutations in Ga genes Family Subtype/Gene Expression Effector(s) Phenotype Reference

a Gas Gas /Gnas ubiquitous AC: Embryonic lethal, parent of origin specific defects [1,2] in heterozygotes Gaolf/Gnal brain, olf. epith. AC: Anosmia, hyperactivity, altered dopamine D1 [3,4] receptor mediated striatal effects Gai/o Gai1/Gnai-1 widely AC (I,V,VI); No obvious phenotype seen so far [*] Gai2/Gnai-2 ubiquitous AC (I,V,VI); Inflammatory bowel disease, mild platelet [5,6,7] activation defect Gai3/Gnai-3 widely AC (I,V,VI); No obvious phenotype seen so far [*] b 2+ 2+ Gao /Gnao neuronal, neuroendocrine Ca -channel (N, P/Q); Functional CNS defects, defective Ca -ch. [8,9] regulation in DRG and heart Gaz/Gnaz neuronal, platelets AC; Impaired platelet activation by epinephrine, altered [10,11] responses to psychoactive drugs Gagust/Gnag taste cells, brush cells ? Impaired bitter and sweet sensation [12] Gat-r/Gnat-1 retinal rods, taste cells cGMP-PDE: Mild retinal degeneration [18] Gat-c/Gnat-2 retinal cones cGMP-PDE: No mouse mutant available - Gaq Gaq/Gnaq ubiquitous PLC-b: Cerebella ataxia, defective platelet activation [13,14] Ga11/Gna-11 almost ubiquitous PLC-b: No obvious phenotype seen so far [15] Ga14/Gna-14 kidney, lung, spleen, testis PLC-b: No obvious phenotype seen so far [$] c Ga15 /Gna-15 hematopoietic cells PLC-b: Discrete signaling defects in macrophages [16] Ga12 Ga12/Gna-12 ubiquitous ? No obvious phenotype seen so far [#] Ga13/Gna-13 ubiquitous ? Defective embryonic angiogenesis (lethal) [17]

a b c 1 2 3 Several splice variants, 2 splice variants, mouse form (Ga16, human counterpart); Yu et al., 1998; Yu et al., 2000; Belluscio et al., 1998; 4Zhuang et al., 2000; 5Rudolph et al., 1995; 6HoÈ rnquist et al., 1997; 7Jantzen et al., 1999; 8Valenzuela et al., 1997; 9Jiang et al., 1998; 10Yang et al., 2000; 11Hendry et al., 2000; 12Wong et al., 1996; 13O€ermanns et al., 1997b; 14O€ermanns et al., 1997c; 15O€ermanns et al., 1998; 16Davignon et al., 2000; 17O€ermanns et al., 1997a; 18Calvert et al., 2000; *L Birnbaumer, M Jiang, G Boulay, K Spicher (personal communication); $H Jiang and MI Simon (personal communication); #SMuÈ ller, SO, MI Simon (unpublished data). AC, adenylyl cyclase; Ca2+-ch., Ca2+-channel; cGMP PDE, cGMP-phosphodiesterase, PLC, phospholipase C; DRG, dorsal root ganglia; CNS, central nervous system. For details, see text

involved in neuronal path®nding, can regulate the Gz, a member of the Gi/o-family of G-proteins shares activity of Gao (Strittmatter et al., 1990, 1995) has led with Gi1,Gi2 and Gi3 the ability to inhibit adenylyl to the hypothesis that Gao plays a role in neuronal cyclases (Wong et al., 1992; Kozasa and Gilman, 1995). guidance. Surprisingly, Gao (7/7) mice showed no In cotransfection experiments, it was shown that gross morphological abnormalities, and neuronal several of the inhibitory receptors which couple to Go path®nding as well as growth cone collapse appeared and Gi-type G-proteins in the central nervous system, to be normal in these animals (Valenzuela et al., 1997). like opioid-, serotonin 5-HT1A, a2-adrenergic or Gao-de®cient mice, however, are clearly impaired being dopamine D2 receptors, also couple to Gz (Ho and smaller and weaker than their littermates and showing Wong, 1998). In contrast to Gi-type G-proteins, Gz is greatly reduced postnatal survival rates (Valenzuela et not a substrate for pertussis toxin and has a rather al., 1997; Jiang et al., 1998). Gao-de®cient mice su€er limited pattern of expression, being found in brain, from tremors and have occasional seizures. A severely adrenal medulla and platelets (Fields and Casey, 1997; abnormal motor behavior can be observed in Gao- Ho and Wong, 1998). In addition, Gaz has quite de®cient mice which show an elevated level of motoric unusual biochemical properties which distinguish it activity and an extreme turning behavior. In addition, from Gai and other G-protein a-subunits. It has a very Gao-de®cient mice appear to be hyperalgesic when low basal GDP dissociation and GTP hydrolysis rate tested in the hot plate assay (Jiang et al., 1998). Opioid (Casey et al., 1990), and interacts with RGSZ1 as well receptor-mediated inhibition of Ca2+ currents in dorsal as with Rap1GAP (Glick et al., 1998; Wang et al., root ganglia (DRG) from Gao-de®cient animals was 1998; Meng et al., 1999; Mochizuki et al., 1999). Mice reduced by about 30% compared to those in wild-type which lack Gaz are viable and do not show any DRGs (Jiang et al., 1998) indicating that inhibition of obvious neurological defects. However, Gaz-de®cient this current is not exclusively mediated by Go. mice exhibit altered responses to a variety of However, the reduction of opioid-induced Ca2+ psychoactive drugs. Cocaine-induced increases in channel inhibition in DRGs from Gao (7/7) mice locomotor activity were much more pronounced in may contribute to the hyperalgesia seen in the hot plate these animals compared to wild-type mice and short test as analgesic e€ects of opioids on the spinal level term antinociceptive e€ects of morphine were found to are believed to involve presynaptic Ca2+ channel be slightly reduced (Yang et al., 2000). However, inhibition in the a€erent pathway. The neurological Hendry et al. (2000) using a di€erent strain and a defects observed in Gao-de®cient mice indicate that di€erent experimental protocol did not observe an Gao plays an important functional role in the central alteration in the acute e€ect of morphine while nervous system while it is obviously not crucially tolerance to the antinociceptive e€ects after chronic involved in the morphogenesis of the CNS. morphine administration was increased. Most strik-

Oncogene Studies in Ga-deficient mice S Offermanns 1637 ingly, it was found that behavioral e€ects of defect in regression of supernumerary CFs in the third catecholamine reuptake inhibitors like reboxetene and postnatal week which most likely resulted from a desipramine were abolished in Gaz de®cient mice functional defect at the PF-PC-synapse. A defect in the (Yang et al., 2000). The receptors involved in the modulation of the PF-PC synapse in mice lacking Gaq various e€ects of psychoactive drugs are not always is also suggested by the fact that long term depression clearly de®ned. Nevertheless, the results of these studies (LTD) of the PF-PC synapse was de®cient in Gaq (7/ clearly indicate that Gz is involved in signaling 7) mice (M Kano et al., unpublished). Very similar processes regulated by various neurotransmitters. phenotypes have been described in mice lacking the Gaolf, a member of the Gas family, has been shown metabotropic glutamate type 1 receptor (GluR1) (Aiba to be expressed in olfactory sensory neurons as well as et al., 1994; Kano et al., 1997) which couples to Gaq/11 in the basal ganglia, olfactory tubercle, the hippocam- as well as in mice de®cient in the b4 isoform of pus and the Purkinje cells of the cerebellar cortex phospholipase C (Kano et al., 1998) which shows (Herve et al., 1993; Zhuang et al., 2000). Most of these predominant expression in cerebellar Purkinje cells brain regions also express Gas. However, Gaolf (Roustan et al., 1995). The mGluR1, Gaq and PLCb4 expression levels clearly exceed those of Gas in the are colocalized in dendritic spines of PCs (Watanabe et nucleus accumbens, the olfactory tubercle and in the al., 1998; Kano et al., 1998; Tanaka et al., 2000), striatum (Belluscio et al., 1998; Zhuang et al., 2000). suggesting that a de®ned signaling cascade in the Gaolf-de®cient mice exhibit dramatically reduced elec- postsynaptic membrane of PF-PC synapses is involved trophysiological responses to all odors tested (Belluscio in the postnatal elimination of multiple CF innervation et al., 1998) supporting the idea that Golf mediates the as well as in cerebellar LTD. Similar signaling path- activation of seven transmembrane domain receptors ways involving Gq are likely to exist in di€erent areas for odorants in the olfactory cilia (Schild and Restrepo, of the brain. It is currently not clear whether the levels 1998). Since nursing and mothering behavior in rodents of Ga11 are too low to compensate for the loss of Gaq is mediated to a great deal by the olfactory system, or whether both proteins serve at least in part non- most Gaolf-de®cient pups die a few days after birth due interchangeable functions. to insucient feeding, and rare surviving mothers exhibit inadequate maternal behavior resulting in the death of all pups born to Gaolf-de®cient mothers Development (Belluscio et al., 1998). Apart from these olfactory defects, surviving Gaolf-de®cient mice exhibit clear The phenotype of various Ga mutant mice have motoric abnormalities like hypermotoric behavior pointed to the involvement of G-protein mediated (Belluscio et al., 1998). Similar phenotypical changes signaling pathways in certain developmental processes, have been observed in mice lacking the dopamine D1 an aspect of G-protein function not fully appreciated receptor (Xu et al., 1994) which has been found to be so far. G12 and G13 constitute the G12 family and coexpressed with Gaolf in striatal neurons (Herve et al., appear to be expressed ubiquitously (Strathmann and 1995), and recent data further support the existence of Simon, 1990). Both G-proteins have been shown to a pathway consisting of D1 receptors and Golf. induce cytoskeletal rearrangements in a Rho-dependent Functional data as well as regional coexpression manner (Buhl et al., 1995; Sah et al., 2000). Lack of suggest the existence of a signaling cascade consisting Ga13 in mice results in embryonic lethality at about of the D1 receptors, Golf and adenylyl cyclase type 5 in midgestation. At this stage, mouse embryos express the nigrostriatal and in parts of the mesolimbic both, Ga12 and Ga13. Analysis of Ga13-de®cient mouse dopaminergic pathway (Zhuang et al., 2000). Thus, embryos revealed that loss of Ga13 leads to a defective the function of Golf is not restricted to olfactory organization of the vascular system which is most sensory cells but obviously plays a role also in other prominent in the yolk sac and in the head mesenchym areas of the central nervous system. (O€ermanns et al., 1997a). Vasculogenic blood vessel The two main members of the Gq-family, Gq and formation through the di€erentiation of progenitor G11, are widely expressed in the central nervous system cells into endothelial cells was not a€ected by the loss and couple numerous receptors to b-isoforms of of Ga13. However, angiogenesis which include sprout- phospholipase C. Usually, levels of Gaq exceed those ing, growth, migration and remodeling of existing of Ga11 several fold in the CNS. Mice lacking Gaq endothelial cells (Risau, 1997) was severely disturbed in develop an ataxia with clear signs of motor coordina- Ga13 (7/7) embryos. Chemokinetic e€ects of throm- tion de®cits. The main neuronal connections in the bin were completely abrogated in ®broblasts lacking cerebellar cortex are believed to play an important role Ga13, indicating that Ga13 is required for full migratory in motor coordination and motor learning (Kim and responses of cells to certain stimuli. Angiogenesis Thompson, 1997), and functional defects could be requires endothelial cell interaction with extracellular observed in the cerebellar cortex of Gaq de®cient mice surfaces, local cell movement and remodeling, pro- (O€ermanns et al., 1997c). While excitatory synaptic cesses which require local cell movement and orienta- transmission from parallel ®bers (PFs) to cerebellar tion in response to speci®c extracellular stimuli. The Purkinje cells (PCs) and from climbing ®bers (CFs) to defects observed in Ga13-de®cient embryos and cells PCs was functional, about 40% of adult Gaq-de®cient occurred in the presence of Ga12, and loss of Ga12 did PCs remained multiply innervated by CFs due to a not result in any obvious defects. Interestingly, Ga12-

Oncogene Studies in Ga-deficient mice S Offermanns 1638 de®cient mice which carry only one intact Ga13 allele humans carrying a mutation in one of the Gas alleles also die in utero (S MuÈ ller, S O€ermanns, MI Simon, (Weinstein and Yu, 1999). unpublished data). This genetic evidence indicates that Ga13 and its closest relative, Ga12, ful®ll at least partially non-overlapping cellular and biological func- Immune system tions. Apart from the role of Gq/11 in heart development Gai2-de®cient mice develop a lethal, di€use in¯amma- signaling through Gq class members has also been tory bowel disease which resembles in many aspects implicated in the proliferation and/or migration of ulcerative colitis in humans (Rudolph et al., 1995). The neural crest cells (Bronner-Fraser, 1995). Endothelin-1 penetrance of this phenotype was greatly a€ected by and the Gq/11-coupled endothelin A (ETA) receptor are the genetic background of the mice homozygous for essential for normal function of craniofacial and the mutation. While 100% of Gai2-de®cient mice on an cardiac neural crest. Endothelin-1 and ETA receptor- inbred 129/Sv background developed in¯ammatory de®cient mice die shortly after birth due to respiratory bowel disease homozygotes on a 129/Sv6C57BL/6 failure (Kurihara et al., 1994, 1995; Clouthier et al., outbred background showed greatly reduced occur- 1998). Severe skeletal abnormalities could be observed rence of the disease. Gai2-de®cient mice show increased in their craniofacial region, including retarded man- numbers of CD4+CD87 and CD47CD8+ thymocytes dibular bones, aberrant zygomatic and temporal with increased staining of CD3. A variety of other bones, and absence of auditory ossicles and tympanic abnormalities were observed in Gai2-de®cient mice, and ring. A milder form of the endothelin-1/ETA-receptor in subsequent studies, dramatic changes in the (7/7) craniofacial phenotype was observed in phenotype and function of intestinal lymphocytes and Gaq(7/7); Ga11(7/+) mice (O€ermanns et al., epithelial cells have been described which are likely to 1998). In contrast, Gaq(7/+); Ga11(7/7) mice did be due to defective lymphocyte homing in enteric not show craniofacial abnormalities suggesting that epithelia (HoÈ rnquist et al., 1997). Dysregulation of the ETA receptor-mediated neural crest development immune system in the intestinal mucosa clearly involves primarily Gaq. It is also possible that a precedes the histopathological and clinical onset of certain amount of Gaq/Ga11 is required for endothelin- bowel in¯ammation further supporting a role of 1 dependent craniofacial development and that this is immunological abnormalities in the pathogenesis of È only provided by one intact allele of the Gaq gene but colitis in the Gai2-de®cient mice (Ohman et al., 2000). not of the Ga11 gene. On a cellular level, Gi2 may be involved in the Stimulatory regulation of adenylyl cyclases through regulation of T cell function and tracking, processes G-protein-coupled receptors involves G-proteins of the which can be controlled through Gi-coupled chemokine Gs-family of which two main members are known, Gs receptors (Baggiolini, 1998; Moser et al., 1998). This is and Golf. The ubiquitously expressed Gas gene gives rise in line with a variety of reports demonstrating that to several splice variants. Gas is the only member of its inactivation of Gi-type G-proteins by pertussis toxin subfamily expressed in most if not all mammalian cells, pretreatment of lymphocytes results in impaired and the complete loss of Gas in mice homozygous for an homing to peripheral lymph nodes and Peyer's patches inactivating Gas mutation leads to embryonic lethality (Spangrude et al., 1984; Chan et al., 1990; Chan before embryonic day 10 (Yu et al., 1998). Hetero- and Perlmutter, 1991, Bargatze and Butcher, 1993; zygotes which inherited the intact allele from their Warnock et al., 1998). In addition, Gi-type G-proteins fathers [Gas(m7/p+)] have increased birth weights have been shown to mediate chemotactic responses of with subcutaneous edema. Later on, they exhibit a lymphocytes and neutrophils to various chemoattrac- delayed development of thymus, kidney and cerebellum tants (Spangrude et al., 1985; Goldmann et al., 1985). as well as neurological signs like ataxia and tremors. While various mechanisms linking loss of Gai2 to the Most of these animals die within the ®rst postnatal development of in¯ammatory bowel disease are weeks. Surviving animals develop obesity and are conceivable, the exact pathogenesis is still obscure. In resistant to parathyroid hormone (PTH). Heterozygous addition to the colitis, many Gai2-de®cient mice mice which have inherited one intact Gas-allele from develop colonic adenocarcinomas. Since Gai2 is ex- their mothers [Gas(m+/p7)] are born hypotrophic and pressed in in¯ammatory cells as well as in colonic the majority die within one day after birth. In contrast epithelial cells dysplasia may be secondary to the to the maternally inherited allele, the paternal allele was in¯ammatory process or may result from an intrinsic not expressed in white and brown adipose tissue as well defect in colonic epithelial cells. Cytogenetic examina- as in the proximal tubulus of the kidney (Yu et al., 1998, tion of normal non-in¯amed mucosa and in¯amed 2000) suggesting that the Gas gene is paternally mucosa in Gai2-de®cient mice suggest that hyperplasia imprinted (i.e. only the maternal allele is expressed) in and dysplasia were secondary to colonic in¯ammation a tissue speci®c manner. This explains the phenotype (Broaddus et al., 1998). Interestingly, mice lacking the seen in [Gas(m7/p+)] mice. It is, however, not clear catalytic domain of phosphatidylinositol-3-kinase g whether defects found in the presence of the maternally (PI3Kg), an e€ector of Gi-type G-proteins, also inherited allele are due to genetic haploinsuciency develop colorectal carcinomas (Sasaki et al., 2000b). and/or to a tissue-speci®c imprinting of the maternal PI3Kg has been shown to be involved in macrophage Gas allele. Similar phenotypes have been observed in and neutrophil chemotactic responses as well as in T

Oncogene Studies in Ga-deficient mice S Offermanns 1639 cell activation (Hirsch et al., 2000; Sasaki et al., 2000a; ventricular myocardium as well as the subepicardial Li et al., 2000), however, no signs of in¯ammatory layer appeared to be underdeveloped. The Gq/G11- bowel disease have been reported in these animals so coupled receptors involved in the regulation of cardiac far. growth at mid-gestation are currently unknown. The a-subunits of the murine G-protein G15 and its Interestingly, pharmacological blockade of 5-HT2B human counterpart G16 belong to the Gaq family and receptors in mouse embryos resulted in abnormal are exclusively expressed in hematopoietic cells (Ama- organization of the subepicardial layer and the absence truda et al., 1991; Wilkie et al., 1991). Expression of the trabecular cell layer of the ventricular appears to be highly regulated during di€erent stages myocardium (Choi et al., 1997), and lack of both, of bone marrow cell maturation suggesting that G15/ endothelin A (ETA) and B (ETB) receptors resulted in G16 play a role in hematopoiesis. Surprisingly, Ga15 midgestational cardiac failure (Yanagisawa et al., de®cient mice show normal hematopoiesis of various 1998). It is possible that there is signaling redundancy lineages, and the morphology and function of the with several inputs into the Gq/11 pathway and that major immunological tissues appeared to be normal only deletion of both the Gaq and the Ga11 genes (Davignon et al., 2000). Apart from a reduced e€ect of results in severe phenotypic defects during early heart C5a in macrophages no cellular defects were found. development due to the ablation of the Gq/G11- Thus, other G-proteins of the Gi- and/or Gq-families phospholipase-C-b-pathway. Interestingly, one intact may compensate for the loss of Ga15, and its allele of the Gaq or Ga11 gene was obviously sucient physiological function remains obscure. to overcome the early developmental block in heart development resulting from the complete absence of Gaq/Ga11. However, Gaq(7/7); Ga11(7/+) and to a Heart lesser degree in Gaq(7/+); Ga11(7/7) pups showed an increased incidence of cardiac defects ranging from Adult cardiomyocytes are terminally di€erentiated post septal defects to univentricular hearts (O€ermanns et mitotic cells which respond to stimulatory signals with al., 1998) (Table 2). There is increasing evidence that cell growth rather than proliferation. By contrast, the Gq/G11-mediated pathway is also involved in embryonic cardiomyocytes retain the capacity to myocardial hypertrophy in the adult heart following proliferate. The Gaq/Ga11-mediated signaling pathway mechanical stress. In line with this, transgenic appears to play a pivotal role in the regulation of expression of wild-type Gaq or of a constitutively physiological myocardial growth during embryogen- active mutant of Gaq in the heart (D'Angelo et al., esis. This is demonstrated by the phenotype of Gaq/ 1997; Mende et al., 1998) results in cardiac hyper- Ga11-double de®cient mice which die at embryonic day trophy. In a reversed approach to study the role of 11 due to a severe thinning of the myocardial layer of Gq/G11 in the induction of cardiac hypertrophy, a the heart (O€ermanns et al., 1998). Both the trabecular short fragment of the Gaq-C-terminus was transgeni- cally expressed in mouse hearts (Akhter et al., 1998). Cardiac expression of this minigene, which was shown Table 2 Phenotype abnormalities in mice carrying inactivating to interfere with receptor-Gq/G11 interaction, reduced mutations in more than one Ga gene ventricular hypertrophy due to aortic constriction. Genotype Phenotype Reference This supports the concept that Gq/G11-mediated phospholipase C activation is critically involved in Gaq Ga11 [1] 7/+ 7/+ no obvious abnormalities the development of mechanical stress-induced cardiac 7/+ 7/7 cardiac malformations; postnatal mortality hypertrophy by coupling receptors of various para- 7/77/+ cardiac malformations; craniofacial defects; and autocrine factors to the induction of a genetic postnatal mortality program which results in the growth of cardiomyo- 7/77/7 myocardial hypoplasia; embryonic lethal e11 cytes. In contrast to its abundance in the central nervous Gaq Ga15 [2] 7/77/7 similar to Gaq (7/7) phenotype system, Gao represents only a minor fraction of all G- proteins in the heart. Nevertheless, it was shown that Ga12 Ga13 [*] the inhibitory muscarinic regulation of cardiac L-type 7/+ 7/+ no obvious abnormalities 2+ 7/+ 7/7 embryonic lethal e9.5 Ca channels in the heart was completely abrogated 7/77/+ embryonic lethal in the Gao (7/7) mice (Valenzuela et al., 1997) 7/77/7 embryonic lethal e8.5 suggesting that Go plays a crucial role in this process. In a study using nodal/atrial-like cardiocytes di€er- Ga Ga [}] i2 i3 entiated from Ga-de®cient mouse embryonic stem cells 7/77/7 embryonic or perinatal lethal an important role for Gao in muscarinic regulation of 2+ Gai1 Gai3 [$] L-type Ca -channels could be con®rmed (Ye et al., 7/77/7 no obvious phenotype seen so far 1999). However, this study also points to an

1 2 involvement of Gi-type G-proteins in this inhibitory e11, embryonic day 11. O€ermanns et al., 1998; Davignon et al., regulation which occurs through an unknown indirect 2000; *SMuÈ ller, SO, MI, Simon (unpublished data); }K Spicher and L Birnbaumer (personal communication); $L Birnbaumer, M Jiang, mechanism. The physiological signi®cance of Go- G Boulay, K Spicher (personal communication); For details, see text mediated Ca2+ channel inhibition in the heart remains

Oncogene Studies in Ga-deficient mice S Offermanns 1640 unknown as heart function of Gao-de®cient mice al., 1996). Platelets from Gaq-de®cient mice did not under non-challenged conditions appeared to be aggregate and secrete their granule contents in response basically intact. to thromboxane A2, ADP, thrombin as well as to low concentrations of collagen (O€ermanns et al., 1997b). Similarly, thromboxane A2, ADP and thrombin failed Sensory systems to induce production of inositol-1,4,5-trisphosphate and transient increases in the free cytosolic Ca2+- Signal transduction of most sensory stimuli involves concentration in Gaq-de®cient platelets, indicating that heterotrimeric G-proteins. Odors, light and tastants act Gaq-mediated activation of phospholipase C is the directly on G-protein coupled receptors. The G-protein central pathway through which various physiological Golf is centrally involved in the transduction of odorant platelet activators signal in order to induce full stimuli while rod-transducin (Gt-r) and conetransducin activation of mouse platelets. G-protein bg-subunits (Gt-c) play well established roles in the phototransduc- released from other G-proteins are obviously not able tion cascade in the outer segments of retinal rods and to compensate the loss of Gaq. Lack of Gaq-mediated cones where they couple light receptors to cGMP- phospholipase C activation did not interfere with the phosphodiesterase (Yar®tz and Hurley, 1994). While ability of thromboxane A2 and thrombin to induce mice heterozygous for an inactivating mutation of the platelet shape change. Thus, induction of platelet shape Gt-r a-subunit appear to be normal, homozygotes show change through receptors of di€erent platelet stimuli is defective light responses and develop mild retinal mediated by G-proteins other than Gq. Studies degeneration with age (Calvert et al., 2000). employing Gaq-de®cient platelets indicate that the G- Unlike the perception of odorants and light gustatory proteins G12/G13 but not Gi are critically involved in stimuli are only in part transduced through G-protein the receptor-mediated shape change response in mediated mechanisms. Among the four taste qualities platelets and that the G12/G13-mediated shape change sweet, bitter, sour and salty, bitter and sweet tastes involves a Rho/Rho-kinase-mediated pathway resulting appear to signal through heterotrimeric G-proteins. in the phosphorylation of the myosin light chain Gustducin is a G-protein mainly expressed in taste cells (Klages et al., 1999). The defective activation of Gaq- (McLaughlin et al., 1992). It has high homology to de®cient platelets results in a primary hemostasis defect transducins and has been proposed to couple receptors as demonstrated by the dramatically increased bleeding for bitter tastants to speci®c taste cell phosphodies- times of adult Gaq-de®cient mice while coagulation terases. This concept is supported by studies on Gagust- times appeared to be normal. In addition, Gaq (7/7) de®cient mice which show impaired electrophysiological mice are protected against platelet-dependent throm- and behavioral responses to bitter agents while responses boembolism as they survived an intravenous injection of gustducin (7/7) mice to sour and salty stimuli were of collagen which was usually lethal in wild-type mice. indistinguishable from those of wild-type mice (Wong et The role of G-proteins of the Gi/o family in platelet al., 1996). Surprisingly, studies in these mice suggest that activation has recently been elucidated. Platelets gustducin is also involved in tasting sweet substances as contain at least three members of this class, Gi2,Gi3 behavioral and neuronal taste sensitivity to sweet and Gz. ADP, which is released from activated tastants was impaired (Lindemann, 1996; Kinnamon platelets and functions as a positive feedback mediator and Margolski, 1996). during platelet activation induces full platelet activa- tion through a Gq-coupled P2Y1 receptor as well as through a Gi-coupled purinergic receptor (Gachet et Platelets al., 1997). The importance of the Gi-mediated pathway is indicated by the fact that responses to ADP were Platelets are discoid cell fragments which under markedly reduced in platelets lacking Gai2 (Jantzen et physiological conditions become activated at sites of al., 1999). In contrast to ADP or thrombin, epinephr- vascular injury. Activated platelets immediately under- ine is not a full platelet activator per se in murine go a shape-change-reaction during which they become platelets. However, it is able to potentiate the e€ect of spherical and extrude pseudopodia-like structures. Full other platelet stimuli. In platelets from Gaz de®cient platelet activation includes secretion of granule con- mice, inhibition of adenylyl cyclase by epinephrine as tents as well as inside-out activation of the ®brinogen well as epinephrine's potentiating e€ects were clearly receptor, integrin aIIbb3, resulting in platelet aggrega- impaired, while the e€ects of other platelet activators tion. Most of the physiological activators of platelets appeared to be una€ected by the lack of Gaz (Yang et like thrombin, thromboxane A2 or ADP act through al., 2000). Thus, Gq-, G12- as well as Gi/o-family G-protein-coupled receptors which in turn activate Gi, members are involved in processes leading to full Gq,G12 and G13 (Shenker et al., 1991; O€ermanns et platelet activation. al., 1994). The main signaling pathway leading to full platelet activation is believed to involve b-isoforms of phospholipase C which can be regulated through G- Conclusions protein bg-subunits or a-subunits of the Gq family. While most cells in mammalian organisms express both Gene targeting techniques have been used to delete Gaq and Ga11, platelets only contain Gaq (Johnson et almost all known genes encoding G-protein a-subunits

Oncogene Studies in Ga-deficient mice S Offermanns 1641 in mice. In some cases, models for complex human obstacles, researchers have begun to cross individual diseases have been generated by targeted gene altera- mutants and to generate mouse lines which allow for tion, in other cases, potential new target-structures or the conditional inactivation of genes in a time- and -processes for new drugs could be de®ned. These tissue-speci®c manner. These approaches will soon studies have led to considerable advances in the allow further insights into the biological functions of understanding of G-protein-mediated signaling pro- G-protein-mediated signaling pathways. cesses. However, only a minor fraction of the wide- ranged biological functions of G-protein mediated signaling processes have probably been elucidated by targeted inactivation of Ga genes so far. Many functions remain unrecognized because of functional Acknowledgments redundancy of closely related G-proteins or because of The authors own research was supported by the Deutsche early death of the de®cient mice. To circumvent these Forschungsgemeinschaft.

References

Aiba A, Kano M, Chen C, Stanton ME, Fox GD, Herrup K, Gudermann T, Kalkbrenner F and Schultz G. (1996). Annu. Zwingman TA and Tonegawa S. (1994). Cell, 79, 377 ± Rev. Pharmacol. Toxicol., 36, 429 ± 459. 388. Hamm HE. (1998). J. Biol. Chem., 273, 669 ± 672. Akhter SA, Luttrell LM, Rockman HA, Iaccarino G, Hendry IA, Kelleher KL, Bartlett SE, Leck KJ, Reynolds Lefkowitz RJ and Koch WJ. (1998). Science, 280, 574 ± AJ, Heydon K, Mellick A, Megirian D and Matthaei KI. 577. (2000). Brain Res., 870, 10 ± 19. Amatruda III TT, Steele DA, Slepak VZ and Simon MI. Herve D, Levi-Strauss M, Marey-Semper I, Verney C, Tassin (1991). Proc. Natl. Acad. Sci. USA, 88, 5587 ± 5591. JP, Glowinski J and Girault JA. (1993). J. Neurosci., 13, Baggiolini M. (1998). Nature, 392, 565 ± 568. 2237 ± 2248. Bargatze RF and Butcher EC. (1993). J. Exp. Med., 178, Herve D, Rogard M and Levi-Strauss M. (1995). Brain Res. 367 ± 372. Mol. Brain Res., 32, 125 ± 134. Belluscio L, Gold GH, Nemes A and Axel R. (1998). Neuron, Hirsch E, Katanaev VL, Garlanda C, Azzolino O, Pirola L, 20, 69 ± 81. Silengo L, Sozzani S, Mantovani A, Altruda F and Broaddus R, Dinh M and Finegold M. (1998). FASEB J., 12, Wymann MP. (2000). Science, 287, 1049 ± 1053. A736. Ho MK and Wong YH. (1998). Biol. Signals Recept., 7, 80 ± Bronner-Fraser M. (1995). Exp. Cell Res., 218, 405 ± 417. 89. Buhl AM, Johnson ML, Dhanasekaran N and Johnson G. HoÈ rnquist CE, Lu X, Rogers-Fani PM, Rudolph U, Shappell (1995). J. Biol. Chem., 270, 24631 ± 24634. S, Birnbaumer L and Harriman GR. (1997). J. Immunol., Calvert PD, Krasnoperova NV, Lyubarsky AL, Isayama T, 158, 1068 ± 1077. Nicolo M, Kosaras B, Wong G, Gannon KS, Margolskee Jantzen H-M, Milstone DS, Gousset L, Conley PB and RF, Sidman RL, Pugh Jr EN, Makino CL and Lem J. Mortensen R. (1999). Blood, 94, Suppl. 1, 618a ± 619a. (2000). Proc. Natl. Acad. Sci. USA, 97, 13913 ± 13918. Jiang M, Gold MS, Boulay G, Spicher K, Peyton M, Brabet Casey PJ, Fong HK, Simon MI and Gilman AG. (1990). J. P, Srinivasan Y, Rudolph U, Ellison G and Birnbaumer L. Biol. Chem., 265, 2383 ± 2390. (1998). Proc. Natl. Acad. Sci. USA, 95, 3269 ± 3274. Chan KE, Beals CR, Wilkie TM, Forbush KA, Simon MI Johnson GJ, Leis LA and Dunlop PC. (1996). Biochem. J., and Perlmutter RM. (1990). EMBO J., 9, 3821 ± 3829. 318, 1023 ± 1031. Chan KE and Perlmutter RM. (1991). Eur. J. Immunol., 21, Kano M, Hashimoto K, Kurihara H, Watanabe M, Inoue Y, 2565 ± 2573. Aiba A and Tonegawa S. (1997). Neuron, 18, 71 ± 79. Choi DS, Ward SJ, Messaddeq N, Launay JM and Kano M, Hashimoto K, Watanabe M, Kurihara H, Maroteaux L. (1997). Development, 124, 1745 ± 1755. O€ermanns S, Jiang H, Wu Y, Jun K, Shin HS, Inoue Clapham DE and Neer EJ. (1997). Annu.Rev.Pharmacol. Y, Simon MI and Wu D. (1998). Proc. Natl. Acad. Sci. Toxicol., 37, 167 ± 203. USA, 95, 15724 ± 15729. Clouthier DE, Hosoda K, Richardson JA, Williams SC, Kim JJ and Thompson RF. (1997). Trends Neurosci., 20, Yanagisawa H, Kuwaki T, Kumada M, Hammer RE and 177 ± 181. Yanagisawa M. (1998). Development, 125, 813 ± 824. Kinnamon SC and Margolskee RF. (1996). Opin. Neurobiol., D'Angelo DD, Sakata Y, Lorenz JN, Boivin GP, Walsh RA, 6, 506 ± 513. Ligget SB and Dorn GW. (1997). Proc. Natl. Acad. Sci. Klages B, Brandt U, Simon MI, Schultz G and O€ermanns S. USA, 94, 8121 ± 8126. (1999). J. Cell Biol., 144, 745 ± 754. Davignon I, Catalina MD, Smith D, Montgomery J, Knall C and Johnson GL. (1998). J. Cell Biochem., 30 ± 31, Swantek J, Croy J, Siegelman M and Wilkie TM. (2000). 137 ± 146. Mol. Cell. Biol., 20, 797 ± 804. Kozasa T and Gilman AG. (1995). J. Biol. Chem., 270, Fields TA and Casey PJ. (1997). Biochem. J., 321, 561 ± 571. 1734 ± 1741. GachetC,HechlerB,LeonC,VialC,LerayC,OhlmannP Kurihara Y, Kurihara H, Suzuki H, Kodama T, Maemura and Cazenave JP. (1997). Thromb. Haemost., 78, 271 ± 275. K, Nagai R, Oda H, Kuwaki T, Cao WH, Kamada N, Glick JL, Meigs TE, Miron A and Casey PJ. (1998). J. Biol. Jishage N, Ouchi Y, Azuma S, Toyoda Y, Ishikawa T, Chem., 273, 26008 ± 26013. Kumada M and Yazaki Y. (1994). Nature, 368, 703 ± 710. Goldman DW, Chang FH, Gi€ord LA, Goetzl EJ and Bourne HR. (1985). J. Exp. Med., 162, 145 ± 156.

Oncogene Studies in Ga-deficient mice S Offermanns 1642 Kurihara Y, Kurihara H, Oda H, Maemura K, Nagai R, Simon MI, Strathmann MP and Gautam N. (1991). Science, Ishikawa T and Yazaki Y. (1995). J. Clin. Invest., 96, 293 ± 252, 802 ± 808. 300. Spangrude GJ, Braaten BA and Daynes RA. (1984). J. LemJ,SidmanR,KosarasB,CalvertP,MakinoC,Nicolo Immunol., 132, 354 ± 362. M, Margolskee R, Wong G, Gannon K and Krasnoper- Spangrude GJ, Sacchi F, Hill HR, Van Epps DE and Daynes aova N. (1998). IOVS, 39, S644. RA. (1985). J. Immunol., 135, 4135 ± 4143. Li Z, Jiang H, Xie W, Zhang Z, Smrcka AV and Wu D. Strathmann M and Simon MI. (1990). Proc. Natl. Acad. Sci. (2000). Science, 287, 1046 ± 1049. USA, 87, 9113 ± 9117. Lindemann B. (1996). Curr. Biol., 6, 1234 ± 1237. Strittmatter SM, Valenzuela D, Kennedy TE, Neer EJ and McLaughlin SK, McKinnon PJ and Margolskee RF. (1992). Fishman MC. (1990). Nature, 344, 836 ± 841. Nature, 357, 563 ± 569. Strittmatter SM, Fankhauser C, Huang PL, Mashimo H and Mende U, Kagen A, Cohen A, Aramburu J, Schoen FJ and Fishman MC. (1995). Cell, 80, 445 ± 452. Neer EJ. (1998). Proc. Natl. Acad. Sci. USA, 95, 13893 ± Tanaka J, Nakagawa S, Kushiya E, Yamasaki M, Fukaya 13898. M,IwanagaT,SimonMI,SakimuraK,KanoMand Meng J, Glick JL, Polakis P and Casey PJ. (1999). J. Biol. Watanabe M. (2000). Eur. J. Neurosci., 12, 781 ± 792. Chem., 274, 36663 ± 36669. Valenzuela D, Han X, Mende U, Fankhauser C, Mashimo Mochizuki N, Ohba Y, Kiyokawa E, Kurata T, Murakami H, Huang P, Pfe€er J, Neer EJ and Fishman MC. (1997). T, Ozaki T, Kitabatake A, Nagashima K and Matsuda M. Proc. Natl. Acad. Sci. USA, 94, 1727 ± 1732. (1999). Nature, 400, 891 ± 894. Wang J, Ducret A, Tu Y, Kozasa T, Aebersold R and Ross Moser B, Loetscher M, Piali L and Loetscher P. (1998). Int. EM. (1998). J. Biol. Chem., 273, 26014 ± 26025. Rev. Immunol., 16, 323 ± 344. Warnock RA, Askari S, Butcher EC and von Andrian UH. Neer EJ. (1995). Cell, 80, 249 ± 257. (1998). J. Exp. Med., 187, 205 ± 216. O€ermanns S, Mancino V, Revel JP and Simon MI. (1997a). Watanabe M, Nakamura M, Sato K, Kano M, Simon MI Science, 275, 533 ± 536. and Inoue Y. (1998). Eur. J. Neurosci., 10, 2016 ± 2025. O€ermanns S, Toombs CF, Hu YH and Simon MI. (1997b). Weinstein LS and Yu S. (1999). Trends Endocr. Sci., 10, 81 ± Nature, 389, 183 ± 186. 85. O€ermanns S, Hashimoto K, Watanabe M, Sun W, Wilkie TM, Scherle PA, Strathmann MP, Slepak VZ and Kurihara H, Thompson RF, Inoue Y, Kano M and Simon Simon MI. (1991). Proc. Natl. Acad. Sci. USA, 88, 10049 ± MI. (1997c). Proc.Natl.Acad.Sci.USA,94, 14089 ± 10053. 14094. Wilkie TM. (2000). Annu. Rev. Biochem., 69, 795 ± 827. O€ermanns S, Zhao L-P, Gohla A, Sarosi I, Simon MI and Wong YH, Conklin BR and Bourne HR. (1992). Science, Wilkie TM. (1998). EMBO J., 17, 4304 ± 4312. 255, 339 ± 342. OÈ hman L, Franzen L, Rudolph U, Harriman GR and Wong GT, Gannon KS and Margolskee RF. (1996). Nature, Hultgren Hornquist E. (2000). Scand. J. Immunol., 52, 381, 796 ± 800. 80 ± 90. Xu M, Moratalla R, Gold LH, Hiroi N, Koob GF, Graybiel Risau W. (1997). Nature, 386, 671 ± 674. AM and Tonegawa S. (1994). Cell, 79, 729 ± 742. Roustan P, Abitbol M, Me nini C, Ribeaudeau F, Ge rard M, Yanagisawa H, Yanagisawa M, Kapur RP, Richardson JA, Vekemans M, Mallet J and Du®er J-L. (1995). NeuroRe- Williams SC, Clouthier DE, deWit D, Emoto N and port, 6, 1837 ± 1841. Hammer RE. (1998). Development, 125, 825 ± 836. Rudolph U, Finegold MJ, Rich SS, Harriman GR, Yang J, Wu J, Kowalska MA, Dalvi A, Prevost N, O'Brien Srinivasan Y, Brabet P, Boulay G, Bradley A and PJ, Manning D, Poncz M, Lucki I, Blendy JA and Brass Birnbaumer L. (1995). Nat. Genet., 10, 143 ± 150. LF. (2000). Proc. Natl. Acad. Sci. USA, 97, 9984 ± 9989. Sah VP, Seasholtz TM, Sagi SA and Brown JH. (2000). Annu. Yar®tz S and Hurley JB. (1994). J. Biol. Chem., 269, 14329 ± Rev. Pharmacol. Toxicol., 40, 459 ± 489. 14332. Sasaki T, Irie-Sasaki J, Jones RG, Oliveira-dos-Santos AJ, Ye C, Sowell MO, Vassilev PM, Milstone DS and Mortensen Stanford WL, Bolon B, Wakeham A, Itie A, Bouchard D, RM. (1999). J. Mol. Cell Cardiol., 31, 1771 ± 1781. KozieradzkiI,JozaN,MakTW,OhashiPS,SuzukiAand Yu S, Yu D, Lee E, Eckhaus M, Lee R, Corria Z, Accili D, Penninger JM. (2000a). Science, 287, 1040 ± 1046. Westphal H and Weinstein LS. (1998). Proc. Natl. Acad. Sasaki T, Irie-Sasaki J, Horie Y, Bachmaier K, Fata JE, Li Sci. USA, 95, 8715 ± 8720. M, Suzuki A, Bouchard D, Ho A, Redston M, Gallinger S, Yu S, Gavrilova O, Chen H, Lee R, Liu J, Pacak K, Parlow Khokha R, Mak TW, Hawkins PT, Stephens L, Scherer AF, Quon MJ, Reitman ML and Weinstein LD. (2000). J. SW, Tsao M and Penninger JM. (2000b). Nature, 406, Clin. Invest., 105, 615 ± 623. 897 ± 902. Zamponi GW and Snutch TP. (1998). Curr. Opin. Neurobiol., Schild D and Restrepo D. (1998). Physiol. Rev., 78, 429 ± 466. 8, 351 ± 356. Shenker A, Goldsmith P, Unson CG and Spiegel AM. Zhuang X, Belluscio L and Hen R. (2000). J. Neurosci., 20, (1991). J. Biol. Chem., 266, 9309 ± 9313. RC91 (1 ± 5).

Oncogene