Rho Gtpase-Dependent Transformation by G Protein-Coupled Receptors

Ian P Whitehead*,1, Irene E Zohn2 and Channing J Der3

1Department of Microbiology and Molecular Genetics, UMDNJ-New Jersey Medical School, Newark, New Jersey, NJ 07103- 2714, USA; 2The Laboratory of Molecular Embryology, The Rockefeller University, 1230 York Avenue, Box 32, New York, NY 10021, USA; 3University of North Carolina at Chapel Hill, Lineberger Comprehensive Cancer Center, Department of Pharmacology, Chapel Hill, North Carolina, NC 27599-7295, USA

G protein coupled receptors (GPCRs) constitute the amino terminus, and are often reversibly palmitylated largest family of cell surface receptors, with more than on a conserved carboxy terminal cysteine. 1000 members, and are responsible for converting a Ligands that have been shown to interact with diverse array of extracellular stimuli into intracellular members of the GPCR family include neurotransmit- signaling events. Most members of the family have ters, hormones, phospholipids, photons, odorants and de®ned roles in intermediary metabolism and generally purine nucleotides (Dohlman et al., 1987). At least four perform these functions in well-dierentiated cells. distinct mechanisms have been described through However, there is an increasing awareness that some which ligands interact with their corresponding GPCRs can also regulate proliferative signaling path- GPCRs. Biogenic amines bind within hydrophilic ways and that chronic stimulation or mutational pockets that are formed by coordinated interactions activation of receptors can lead to oncogenic transforma- between the extracellular loops of their receptors. tion. Activating mutations in GPCRs are associated with Peptide hormones also bind within such pockets but several types of human tumors and some receptors require additional interactions with extracellular resi- exhibit potent oncogenic activity due to agonist over- dues. Ligands such as glycoproteins, hormones or expression. Additionally, expression screening analyses glutamate bind directly to the amino terminus of the for novel oncogenes identi®ed GPCRs whose expression receptor, and then the ligand receptor complex folds causes the oncogenic transformation of NIH3T3 mouse into the hydrophilic pocket. Finally, thrombin cleaves ®broblasts. These include Mas, G2A, and the PAR-1 the amino terminus of its receptor, Par-1, thus thrombin receptor. In this review we summarize the releasing a tethered ligand that causes receptor signaling and transforming properties of these GPCR activation. Regardless of the mechanism of ligand oncoproteins. What has emerged from these studies is the stimulation, engagement of a GPCR by its ligand delineation of a GTPase cascade where transforming generally leads to activation of an associated member GPCRs cause aberrant growth regulation via activation of the heterotrimeric G protein family. of Rho family small GTPases. Oncogene (2001) 20, GPCRs associate with members of the heterotrimeric 1547 ± 1555. G protein family that are anchored to the intracellular surface of the plasma membrane (Bourne, 1997). These Keywords: gene transfer assays; Dbl family proteins G proteins consist of an a subunit (Ga) that binds guanine nucleotides, and a dimer that consists of a b and g subunit (Gbg). The intracellular signaling events GPCRs and heterotrimeric G proteins that are triggered by the activation of heterotrimeric G proteins is determined by the particular combination of Members of the GPCR family are characterized by a Ga and Gbg subunits that are present in the complex, tertiary structure that consists of seven transmembrane and the availability of speci®c eectors and regulatory domains that are connected by a series of extracellular proteins. The GPCR interacts with the G protein and intracellular loops (Dohlman et al., 1987). The through multiple contacts, with the most critical amino terminus of the receptor is extracellular and residues being located in the third cytoplasmic loop often contributes to ligand recognition and binding, and in the membrane proximal portion of the while the intracellular carboxy terminus contributes to cytoplasmic tail. eector binding and to the propagation of signaling In the absence of ligand, the GPCR associates with events. The transmembrane regions, which represent an inactive GDP-bound complex. Upon ligand engage- the most highly conserved domains within the ment, the GPCR assumes an active con®guration and structure, consist of 20 ± 25 amino acid stretches that functions as a guanine nucleotide exchange factor are predicted to form alpha helices. Typically, GPCRs (GEF), catalyzing the exchange of GDP for GTP on contain at least one N-linked glycosylation site in the the Ga subunit. Upon conversion to the active GTP- bound form, the Ga subunit will dissociate from the Gbg subunit and each component is free to elicit a distinct array of eector functions. For example, *Correspondence: IP Whitehead activated Ga subunits have been shown to be both G protein-couple receptor oncoproteins and Rho GTPases IP Whitehead et al 1548 activators and inhibitors of adenylyl cyclases, or have ®broblasts (Gupta et al., 1992; Pace et al., 1991). The been shown to directly activate phospholipase C. Gbg wild type Ga12 subunit (gep oncogene) was isolated in subunits have also been shown to activate adenylyl a screen of a soft tissue sarcoma library for proteins cyclases or phospholipase C, but there is also evidence whose expression cause focus formation in NIH3T3 that they can activate Ras-mediated signaling pathways cells (Chan et al., 1993). However, subsequent analysis (Crespo et al., 1994; Koch et al., 1994), although the has failed to establish that Ga12 overexpression is nature of this activating event remains somewhat responsible for the original malignancy. obscure. Experimental mutation of Ga subunits that render them GTPase-de®cient (with Q to L mutations analogous to the oncogenic Q61L mutation of Ras Oncogenic potential of GPCRs and heterotrimeric G proteins) has also revealed the growth-promoting proteins activity of Ga12,Ga13,Gai2,Gaq, and Gao (reviewed in Dhanasekaran et al. (1998)). Of these, the QL Although GPCRs have been best characterized for the mutants of the Ga12/Ga13 family have exhibited the speci®c role that many play in regulating diverse greatest transforming potency (Jiang et al., 1993; Vara metabolic processes, it is becoming increasingly clear Prasad et al., 1994; Voyno-Yasenetskaya et al., 1994; that members of this family can regulate proliferative Xu et al., 1993). Taken together, these examples signaling pathways (Coso et al., 1995a; Gutkind, 1998; provide a persuasive body of evidence linking the van Biesen et al., 1996). These include the serotonin 1C deregulated activation of GPCRs, and their corre- receptors (Julius et al., 1989) and the M1,M3 and M5 sponding G proteins, to human malignancies. subtypes of the muscarinic acetylcholine receptors (Gutkind et al., 1991) that cause agonist-dependent transformation of NIH3T3 cells. Point mutations Expression cloning for novel oncogenes: identi®cation of render the a1B-adrenergic receptor activated and Mas, G2A, and PAR-1 transforming in a ligand-independent fashion (Allen et al., 1991). Activating mutations in GPCRs are also Adding to the body of evidence that members of the known to be associated with several types of human GPCR family may represent important oncoproteins tumors. Mutations in the thyroid-stimulating hormone has been the identi®cation of GPCRs in biological receptor have been identi®ed in 30% of hyperfunction- screens for proteins whose expression can cause the ing thyroid adenomas and similar mutations in the oncogenic transformation of murine ®broblasts. We luteinizing hormone receptor are associated with male describe the identi®cation and characterization of three precocious puberty (Parma et al., 1993; Shenker et al., transforming GPCRs, Mas, G2A, and Par-1, that were 1993). Activation of both of these receptor types leads identi®ed in these gene transfer analyses. to increased cellular concentrations of cAMP, and they The identi®cation of the mas oncogene by Wigler are transforming in cell types in which cAMP functions and colleagues provided the ®rst direct evidence for the as a mitogen. oncogenic potential of GPCRs (Young et al., 1986). There is also evidence that GPCRs may contribute The mas oncogene was originally isolated from a to tumorigenesis via paracrine and autocrine mechan- human epidermoid carcinoma genomic DNA library isms. For example, small cell lung carcinomas secrete and identi®ed by its ability to render NIH3T3 cells high levels of gastrin-reducing peptide (GRP) and GRP tumorigenic in nude mice. The mas gene isolated in this receptor antagonists can block the growth of these cells screen contained rearrangements in its 5' noncoding both in vitro and in vivo (Cuttitta et al., 1985; sequences, but not in its coding sequence. Thus, mas Mahmoud et al., 1991). Gastric cancers also secrete transforming activity is mediated by overexpression GPCR ligands such as GRP and gastrin, colon and and not mutational activation. Mas transcripts and ovarian cancer patients exhibit elevated levels of protein are expressed primarily in the central nervous gastrin and lysophosphatidic acid (LPA) in their serum system. High levels of mas transcripts were detected in (Seitz et al., 1991; Xu et al., 1995). the cerebral cortex, hippocampus and cerebellum The identi®cation of tumor-associated activating (Young et al., 1988). mutations in Ga subunits provides further evidence The mas gene encodes a 325 amino acid protein with for aberrant GPCR function in growth transformation. seven potential transmembrane segments and other Activating mutations in the Ga subunits of several features characteristic of GPCRs. Mas was originally heterotrimeric G proteins have been shown to be thought to be a receptor for angiotensin II, but recent oncogenic. The gsp oncogene has been identi®ed in evidence argues against this possibility and the ligand human thyroid and pituitary tumors and is associated for Mas is currently unknown. Instead, Mas appears to with the McCune ± Albright syndrome (Landis et al., de®ne a new family of orphan GPCRs that includes 1989). gsp encodes a constitutively activated derivative RTA and MRG (Monnot et al., 1991; Ross et al., of Gas. The gip2 oncogene, which was identi®ed in 1990). both human adrenal cortical tumors and ovarian sex A second transforming GPCR was identi®ed by Kay cord stromal tumors, encodes a constitutively activat- and colleagues in a retrovirus cDNA expression library ing mutation in the Gai subunit (Lyons et al., 1990). screen for oncogenes that were expressed in the T28 gip2 is fully transforming when overexpressed in Rat-1 murine T hybridoma cell line. This gene encoded a

Oncogene G protein-couple receptor oncoproteins and Rho GTPases IP Whitehead et al 1549 novel GPCR that could induce the loss of density-, tion was due to receptor overexpression. Introduction anchorage-, and growth factor-dependent growth of of a mutation in PAR-1 that is predicted to impair NIH3T3 cells (Zohn et al., 2000). The transforming receptor cleavage abolished its transforming activity gene was found to be identical in sequence to the G2A suggesting that transformation by PAR-1 may occur in gene that was identi®ed independently by Witte and a ligand-dependent manner. However, high levels of colleagues (Weng et al., 1998). They identi®ed G2A by thrombin inhibitors failed to block PAR-mediated using representational dierence analysis to isolate transformation suggesting that transformation may be genes deregulated in expression by the transforming mediated by additional proteases that target PAR-1. Bcr ± Abl oncoprotein and not by a transformation- Consistent with their potential role in oncogenic impaired mutant of Bcr ± Abl when expressed in transformation is an increasing body of evidence that murine pre-B cells. From this screen they identi®ed some members of the GPCR family can regulate signal G2A as a Bcr ± Abl-inducible gene. Paradoxically, their transduction pathways that have demonstrated roles in studies showed that ectopic expression of G2A in the regulation of cell growth and proliferation. Recent NIH3T3 cells caused the accumulation of cells in the reviews have described the involvement of various G2/M transition point in the cell cycle. Hence, they cytoplasmic signaling pathways, such as the ERK named this gene G2A (G2 accumulation). Ectopic mitogen-activated protein kinase cascade, in mediating overexpression of G2A was also shown to antagonize GPCR growth regulation (Gutkind, 1998). GPCRs are the growth-promoting activity of Bcr-Abl in Rat-1 also known to elicit a number of cellular responses via ®broblasts and mouse bone marrow cells. Thus, G2A activation of members of the Rho family of small may exhibit a growth-promoting or growth-antagoniz- GTPases. Although the best described cellular function ing function that may be dictated by the presence of of Rho proteins concerns their ability to regulate other signals or by cell type dierences. various aspects of actin cytoskeletal architecture, more G2A mRNA expression was found to be highest in recent studies have established roles for several family hematopoietic tissue and cell lines. Bcr-Abl, as well as members in the regulation of proliferative signaling exposure to various DNA-damaging agents, was found pathways. In the following sections, we provide an to cause upregulation of G2A mRNA (Weng et al., overview of the function of Rho family GTPases and 1998). The encoded protein is a member of a subclass we then discuss the evidence that Mas, G2A, and of GPCRs with no de®ned function (Strader et al., PAR-1 cause growth transformation by activation of 1994, 1995). The G2A sequence identi®ed in the these small GTPases. transformation screen was identical to that isolated in the Bcr-Abl study, indicating that G2A transforming activity is not caused by mutational activation. Since, Rho family small GTPases and oncogenesis as is the case with Mas, the ligand for G2A is not known, whether G2A transformation is ligand-inde- Members of the Rho branch of the Ras superfamily of pendent or relies on a ligand present in calf serum or is proteins have attracted considerable attention over the expressed by NIH3T3 cells is not known. past eight years for three primary reasons (reviewed in More recently, we have identi®ed cDNAs that Symons, 1995; Van Aelst and D'Souza-Schorey, 1997; encode the thrombin receptor PAR-1 in two indepen- Zohn et al., 1998a). First, it is well established that dent screens for proteins whose expression causes focus Rho family proteins are regulators of extracellular formation in NIH3T3 cells (Martin et al., 2001; stimuli-mediated signaling pathways that control actin Whitehead et al., 1995). Retrovirus cDNA expression cytoskeletal organization, gene expression and cell libraries derived from mRNA from 32D and B6SutA1 cycle progression. Second, the aberrant activation of mouse myeloid cell lines identi®ed Par-1 as a Rho family proteins promotes the uncontrolled pro- transforming gene in NIH3T3 focus formation assays. liferation and invasive and metastatic properties of Overexpression of wild type human PAR-1 was found tumor cells. Third, Ras and other oncoproteins require to exhibit a similar transforming activity. In addition Rho family protein activation to mediate their to its potent focus forming activity, constitutive transforming activities. Finally, the discovery of new overexpression of PAR-1 in NIH3T3 cells promoted members of the Dbl family oncoproteins continues at the loss of anchorage- and serum-dependent growth. an amazing rate and their ability to cause transforma- PAR-1 is one of a family of receptors that is tion and promote tumor cell invasion is due to activated by the serine protease thrombin (Coughlin, activation of Rho family proteins. Taken together, 1993). Although the best described function of these properties of Rho family proteins have implicated thrombin concerns its role in blood coagulation, it them as important regulators of cellular processes that has also been shown to be a powerful agonist that can may contribute to the development of human cancers. elicit a variety of cellular responses including mitogen- Rho family proteins are members of a major branch esis (Brass et al., 1998; Van Obberghen-Schilling and of the Ras superfamily of small GTPases and share PouysseÂ gur, 1993). Additionally, PAR-1 overexpression approximately 30% amino acid identity with Ras has been implicated in the invasive properties of (Chardin, 1993). Like Ras and Ga subunits, Rho various human cancers (Even-Ram et al., 1998; family proteins function as GTP/GDP-regulated Henrikson et al., 1999). The isolated sequences encoded switches that cycle between active GTP- and inactive for the wild type protein, indicating that transforma- GDP-bound forms (Ridley, 1996). This cycle is

Oncogene G protein-couple receptor oncoproteins and Rho GTPases IP Whitehead et al 1550 regulated by two main classes of regulatory proteins. eector action involves their regulation of the First, Rho guanine nucleotide exchange factors (GEFs; organization of the actin cytoskeleton, which in turn also referred to as Dbl family proteins) serve as in¯uences cell shape, cell motility, cell-substratum and activators and stimulate the replacement of GDP by cell-cell interactions (reviewed in Ridley, 1996; Symons, GTP (Cerione and Zheng, 1996; Whitehead et al., 1996). Whereas Rac1 promotes the induction of 1997). All Dbl family proteins share a Dbl homology lamellipodia and membrane ruing, RhoA stimulates (DH) domain catalytic domain and a ¯anking the formation of stress ®bers and focal adhesions, and pleckstrin homology (PH) domain that regulates DH Cdc42 induces the formation of ®lopodia, which are domain function. Second, Rho GTPase activating ®nger-like cytoplasmic extensions that may be involved proteins (GAPs) have also been identi®ed that serve in the recognition of the extracellular environment. as negative regulators of Rho family protein function A second function of Rho GTPases is their and stimulate their intrinsic GTPase activities (Cerione regulation of transcription factor activity and gene and Zheng, 1996; Van Aelst and D'Souza-Schorey, expression (reviewed in Van Aelst and D'Souza- 1997). Schorey, 1997; Zohn et al., 1998a). Transcription Like Ras, Rho family proteins also serve to relay factor targets of Rho GTPase-mediated signaling extracellular ligand-mediated signals to cytoplasmic pathways include the c-Fos and c-Jun oncoproteins, signaling pathways that regulate diverse cellular the serum response factor (SRF), which cooperates processes and cell behavior (reviewed in Van Aelst with ternary complex factors (TCFs) and the serum and D'Souza-Schorey, 1997; Zohn et al., 1998a). Rho response DNA elements found in certain promoters proteins are activated transiently in response to ligands such as the c-fos promoter, and NF-kB, a regulator of that stimulate cell surface receptors that include cell survival. Finally, Rho family proteins (RhoA, Rac1 GPCRs. Presently, at least 16 mammalian Rho family and Cdc42) are also required for progression through proteins have been identi®ed that share signi®cant the G1 phase of the cell cycle (Yamamoto et al., 1993; amino acid identity with each other (ranging from 50 ± Olson et al., 1995). 90%): RhoA, RhoB, RhoC, RhoD, RhoE/Rnd3, A strong involvement of Rho family proteins in Rnd1, Rnd2, RhoG, Rac1, Rac2, Cdc42, TC10, TTF oncogenesis has been established. First, the transform- (Zohn et al., 1998a), Chp (Aronheim et al., 1998), TCL ing action of Ras and other oncoproteins (e.g., Vav, (Vignal et al., 2000), and Rif (Ellis and Mellor, 2000). Abl, Met) is dependent on Rho family protein function To date, RhoA, Rac1 and Cdc42 have been the most (Khosravi-Far et al., 1995; Prendergast et al., 1995; intensively studied. Qiu et al., 1995a,b, 1997). Second, constitutively The activated GTP-bound Rho GTPases interact activated mutants of Rac1, RhoA, RhoB or Cdc42 with a broad spectrum of functionally diverse eectors have been shown to cause tumorigenic transformation to mediate their functions (Bishop and Hall, 2000) of rodent ®broblasts (Khosravi-Far et al., 1995; (Figure 1). For example, over 11 eectors of RhoA Prendergast et al., 1995; Qiu et al., 1995a,b, 1997). function have been identi®ed, including a number of Third, Rho family proteins promote tumor cell serine/threonine kinases. One major consequence of invasion and metastasis. For example, Collard and

Figure 1 Rho GTPases interact with multiple downstream eector targets. The active, GTP-bound forms of RhoA, Rac1, and Cdc42 have been found to associate with a diverse spectrum of eectors. Additionally, some eectors can interact with multiple Rho GTPases

Oncogene G protein-couple receptor oncoproteins and Rho GTPases IP Whitehead et al 1551 colleagues determined that Rac1, as well as its Dbl Elk-1 activation, it does regulate SRF-mediated gene family protein activator Tiam1, induced T lymphoma expression in a RhoA-dependent manner. In contrast cell invasion (Habets et al., 1994; Michiels et al., 1995). to GPCRs and Rho proteins, Ras and Raf are potent Activated Rac1 and Cdc42 promoted human breast activators of ERK MAPK and Elk-1 in NIH3T3 cells, carcinoma cell invasion (Keely et al., 1997). Lacal and and this activation is required for Ras-mediated colleagues showed that three constitutive-activated transforming activity. RhoA GEFs (Dbl, Vav, and Ost) caused metastatic Mas transformation appears to be mediated by transformation of rodent ®broblasts (del Peso et al., activation of Rac, whereas G2A and PAR-1 transfor- 1998). Hynes and colleagues recently showed that mation involves the activation of RhoA. Mas-trans- RhoC activation promoted melanoma cell metastastic formed NIH3T3 cells retain stress ®bers and focal growth (Clark et al., 2000). adhesions and exhibit increased membrane ruing (Zohn et al., 1998b). These eects are mimicked by a constitutively activated derivative of Rac1. Stable PAR-1, G2A and Mas transform by activation of Rho expression of G2A in NIH3T3 cells causes the and Rac proteins formation of stress ®bers, but not membrane rues, thus suggesting RhoA activation (Kabarowski et al., It has now been solidly established within the literature 2000; Zohn et al., 1998b). In contrast, stable expression that several of the cellular behaviors that are elicited by of an activated derivative of Ras results in a dramatic GPCRs are dependent upon the activation of Rho loss of stress ®bers and focal adhesions, but induction GTPase-controlled signaling pathways. This was ®rst of membrane ruing is observed. Similar observations demonstrated through the striking eects that GPCR were made when expression plasmids for PAR-1, G2A agonists such as lysophosphatidic acid (LPA), bombe- and Mas were microinjected into porcine aortic sin and bradykinin have on cell morphology (Ridley endothelial (PAE) cells. Both PAR-1 and G2A induce and Hall, 1992; Ridley et al., 1992a). In Swiss 3T3 stress ®ber formation suggesting that they are mouse ®broblasts, LPA application causes Rho- activators of RhoA, while Mas induces lamellipodium dependent stress ®ber formation, bombesin causes formation similar to Rac1. Finally, expression of Rac-dependent membrane ruing and lamellipodia dominant inhibitory versions of Rho family proteins formation, and bradykinin causes Cdc42-mediated is sucient to block the transforming activity of G2A, ®lopodia formation. It has been demonstrated subse- PAR-1 and Mas in NIH3T3 cells. Both Mas and G2A quently that Rho proteins and GPCRs activate similar transforming activity is blocked by dominant inhibitory signal transduction pathways, and in some cases derivatives of RhoA and Rac1 (Cdc42 was not tested) activation of these pathways by GPCRs is dependent while PAR-1 transforming activity is blocked by RhoA upon Rho function (reviewed in Seasholtz et al., 1999). (Racl and Cdc42 have not been tested). Finally, using a Since there is clear evidence that Rho GTPases and pulldown assay that can be used to directly assess the GPCRs regulate similar cellular processes it is not GTP-loading of Rho proteins in vivo, both PAR-1 and surprising that they have been linked in a number of G2A have been shown to activate RhoA in COS7 and cellular behaviors. Evidence now points to the Swiss 3T3 cells, respectively. involvement of Rho proteins in the regulation of several GPCR mediated cellular activities including phospholipid metabolism, smooth muscle contraction, Linking GPCR transformation to Rho GTPases cell migration, cell growth and survival responses. There are several lines of evidence that Mas, G2A, The mechanisms by which Mas, G2A, and PAR-1 and PAR-1 transform NIH3T3 cells by activation of cause Rho GTPase activation has been best delineated speci®c Rho family members (Kabarowski et al., 2000; for G2A. The minimum requirement for GPCR Martin et al., 2001; Weng et al., 1998; Zohn et al., activation of a Rho GTPase would involve hetero- 1998b, 2000). First, we have observed that NIH3T3 trimeric G protein activation of a Dbl family GEF. cells that are transformed by these GPCRs, or by First, what Ga or Gbg subunits can promote activation activated derivatives of Rho family members, form of RhoA? Since constitutively activated mutants of similar foci of transformed cells that are comprised of Ga12,Ga13,andGaq can cause RhoA-dependent rounded, piled-up non-refractile cells. This phenotype formation of actin stress ®bers (Buhl et al., 1995; is distinct from that seen when cells are transformed Zohn et al., 2000), they are the likely candidates to with unrelated oncogenes such as Ras, Raf or Src, but mediate G2A activation of RhoA. The demonstration similar to foci induced by other Rho GTPase- that G2A can eciently induce stress ®ber formation in dependent transforming proteins (e.g., Lsc, Ga12). mouse ®broblast cells that lack Ga12 or Gaq/11, but not Second, the signaling pro®le of GPCRs in NIH3T3 Ga13, suggests that Ga13 is a key regulator of G2A cells is highly reminiscent of Rho activation. Like the activation of RhoA (Kabarowski et al., 2000). Rho proteins, Mas and PAR-1 are good activators of Second, what RhoA-speci®c GEF may function NF-kB-and SRF, but are poor activators of Elk-1. Elk- downstream of Ga13? Is the connection direct or is it 1 is a ternary complex factor that is transcriptionally indirect and involves additional signaling components? activated when phosphorylated by ERK MAPKs. Human p115 RhoGEF, and its mouse counterpart, Although G2A has not been tested for NF-kBor Lsc, is a RhoA-speci®c GEF in vitro (Glaven et al.,

Oncogene G protein-couple receptor oncoproteins and Rho GTPases IP Whitehead et al 1552 1996; Hart et al., 1996) and its failure to activate JNK may activate RhoA by another Dbl family protein in vivo argues that it is not an activator of Rac or (Seasholtz et al., 1999). Cdc42 (Westwick et al., 1998). Like G2A, Lsc was also A second Rho exchange factor, PDZ-RhoGEF, has identi®ed in an expression library screen for transform- also been shown to bind to Ga12 and Ga13, and to ing proteins (Whitehead et al, 1996). Furthermore, like regulate LPA-mediated activation of SRE (Fukuhara G2A, p115 RhoGEF/Lsc transcripts are most abun- et al., 1999). However, whether the RGS domain of dant in hematopoietic tissues and cell lines. Recently, PDZ-RhoGEF can act as a GAP for these Ga Sternweis, Bollag and colleagues identi®ed a regulator subunits, and whether Ga12 and/or Ga13 binding of G protein signaling (RGS) domain in the amino promotes the GEF activation, have not been deter- terminus of p115 RhoGEF/Lsc (Kozasa et al., 1998). mined. A highly related protein, identi®ed by Caligiuri Like other RGS proteins, p115 RhoGEF serves as a and colleagues as a leukemia-associated Rho GEF GAP for both Ga12 and Ga13. This activity is highly (LARG), may also link GPCRs with Rho GTPases speci®c since it is not observed with Gas,Gai or Gaq (Kourlas et al., 2000). Like PDZ-RhoGEF, LARG family members. Co-expression of the isolated RGS also contains an amino terminal PDZ and RGS domain of Lsc inhibited G2A, as well as Ga12 and domains and carboxyl terminal DH and PH domains. Ga13, transformation of NIH 3T3 cells, implicating Gutkind and colleagues showed recently that PDZ- their importance in G2A function. Finally, Ga13, but RhoGEF and LARG can cause activation of RhoA in not Ga12, stimulated the activity of p115 RhoGEF in vivo, suggesting that both are Rho-speci®c GEFs vitro and in vivo (Hart et al., 1998; Mao et al., 1998). (Fukuhara et al., 2000). Since Ga12 and Gaq can also Taken together, these observations support a surpris- cause RhoA activation and stress ®ber formation ingly simple model where G2A-mediated activation (Zohn et al., 2000), it will be interesting to determine causes the activation of Ga13, which in turn directly whether PDZ-RhoGEF or LARG provide the links to associates with and stimulates the GEF activity of p115 RhoA for these Ga subunits or if additional RGS- RhoGEF/Lsc, leading to the activation of RhoA containing RhoA GEFs remain to be identi®ed. PDZ- (Figure 2). Additionally, since p115 RhoGEF is also RhoGEF and LARG are expressed ubiquitously and a GAP for Ga13, this may allow rapid activation- hence may link GPCRs with RhoA in a wide range of deactivation of the Ga:Rho GEF complex. cell types. PAR-1 activation of RhoA is also likely to be How Mas activates Rac remains incompletely under- mediated by a similar mechanism as G2A. Over- stood. One logical candidate for the heterotrimeric G expression of the Lsc RGS domain also blocked the protein component is Gaq and this possibility is focus-forming activity of PAR-1 in NIH3T3 cells, thus supported by several observations. First, Mas has been implicating Ga12 and/or Ga13 in PAR-1 transformation shown to be an activator of phospholipase C (Poyner (Martin et al., 2001). However, thrombin-mediated et al., 1990), which is an eector of Gaq (van Biesen et stress ®ber formation appears to be mediated primarily al., 1996). Second, bombesin-mediated GPCR activa- through Ga12 rather that Ga13, suggesting that PAR-1 tion causes activation of phospholipase C and Rac-

Figure 2 The p115 RhoGEF/Lsc Dbl family protein may couple G2A to RhoA. G2A activation causes activation of Ga13. Activated Ga13 binds to the RGS domain of p115 RhoGEF and stimulates the GEF activation of p115 Rho GEF, leading to formation of RhoA-GTP. Overexpression or mutational activation of each component of this pathway alone can cause transformation of NIH3T3 cells. Activated RhoA can associate with a multitude of eectors (E) to mediate growth transformation. However, whether one eector alone, or multiple eectors, are required to promote RhoA transformation, invasion and metastasis is not known. DH, Dbl homology domain; PH, pleckstrin homology domain; RGS, regulator of G protein signaling domain

Oncogene G protein-couple receptor oncoproteins and Rho GTPases IP Whitehead et al 1553 dependent induction of lamellipodia (Oermanns et al., trimeric G proteins and Rho GTPases adds to an ever- 1994; Ridley et al., 1992b). Third, the stimulation of expanding number of GTPase cascades (Chant and GPCRs that couple to Gaq, or activated Gaq itself, Stowers, 1995). The delineation of a remarkably also cause activation of JNK, a Rac/Cdc42-mediated straightforward mechanism by which G2A and other MAPK (Coso et al., 1995b; Heasley et al., 1996; Zohn Ga13-coupled GPCRs cause activation of RhoA et al., 1995). However, we found that microinjection of identi®es the best-characterized involvement of a Dbl activated Gaq caused stress ®ber, rather than lamelli- family protein in relaying extracellular signals to Rho podia, induction (Zohn et al., 2000). Furthermore, GTPases. The number of Dbl family proteins continues activated Gaq (QL) alone did not cause the same focus- to increase, with over 40 mammalian members. Aside forming activity as Mas. However, we did ®nd that co- from the conserved DH/PH domain structure shared expression of activated Gal (QL) and Gaq (QL) did by all Dbl family proteins, these Rho GEFs diverge cause the appearance of Mas-like foci (Zohn et al., signi®cantly in the sequences that ¯ank this catalytic unpublished). Additionally, another GPCR that caused module. In addition to the RGS domains found in coordinate activation of Gaq and Gal caused activation p115 RhoGEF, PDZ-RhoGEF, and LARG, other of Rac and transformation of NIH3T3 cells (Zohn et protein-protein or lipid-protein interaction elements al., unpublished). We did ®nd that Mas transforming of Dbl family proteins include PDZ, Src homology 2 activity was impaired by pertusis toxin treatment, and 3, C2, EH, and zinc ®nger domains. Dbl family indicating that Mas was coupled to Gal (Zohn et al., proteins can also be activated by phosphorylation or 2000). Perhaps it is the coordinate activation of these by association with phosphoinositides, thus implicating two Ga subunits by Mas that is required to cause the protein or lipid kinases as possible intermediates that activation of Rac. Alternatively, it is also possible that promote Rho GTPase activation. Hence, these distinct activated Gbg may, via interaction with the PH domain regulatory elements of Dbl family proteins may of a Rac-speci®c Dbl family protein, that promotes provide additional mechanisms by which GPCR GPCR activation of Rac (Inglese et al., 1995). Finally, activation can regulate Rho GTPases. In summary, while a number of Rac-speci®c GEFs have been how GPCRs cause activation of Rac and Cdc42, identi®ed (Cerione and Zheng, 1996; Whitehead et whether additional RGS-containing Dbl family pro- al., 1997), whether any of these are involved in teins remain to be identi®ed, and what the precise mediating Mas activation of Rac remains to be contributions of Rho GTPases in GPCR function are evaluated. questions for future studies.

Conclusions and future directions Acknowledgments We thank Misha Rand for excellent assistance in prepara- The identi®cation of GPCRs in biological screens for tion of the text, references and ®gures. Our studies were transforming proteins provides further evidence for the supported by Public Health Service grants CA42978, contribution of aberrant GPCR activation in human CA55008, CA63071 (CJ Der.) and CA77493 (IP White- oncogenesis. Additionally, the linkage between hetero- head) from the National Cancer Institute.

References

Allen LF, Lefkowitz RJ, Caron MG and Cotecchia S. (1991). Coso OA, Chiariello M, Kalinec G, Kyriakis JM, Woodgett Proc. Natl. Acad. Sci. USA., 88, 11354 ± 11358. J and Gutkind JS. (1995b). J. Biol. Chem., 270, 5620 ± Aronheim A, Broder YC, Cohen A, Fritsch A, Belisle B and 5624. Abo A. (1998). Curr. Biol., 8, 1125 ± 1128. Coso OA, Chiariello M, Yu J-C, Teramoto H, Crespo P, Xu Bishop AL and Hall A. (2000). Biochem. J., 348, 241 ± 255. N, Miki T and Gutkind JS. (1995a). Cell, 81, 1137 ± 1146. Bourne HR. (1997). Curr. Opin. Cell Biol., 9, 134 ± 142. Coughlin SR. (1993). Thrombosis and Haemostasis, 66, 184 ± Brass LF, Ahuja M, Belmonte E, Pizarro S, Tarver A and 187. Hoxie JA. (1998). The human platelet thrombin receptor. Crespo P, Xu N, Simonds WF and Gutkind JS. (1994). In Annals New York Academy of Sciences. Nature, 379, 557 ± 560. Buhl AM, Johnson NL, Dhanasekaran N and Johnson GL. Cuttitta F, Carney DN, Mulshine J, Moody TW, Fedorko J, (1995). J. Biol. Chem., 270, 24631 ± 24634. Fischler A and Minna JD. (1985). Nature, 316, 823 ± 826. Cerione RA and Zheng Y. (1996). Curr. Opin. Cell Biol., 8, delPesoL,HernaÂ ndez-Alcoceba R, Embade N, Carnero A, 216 ± 222. Esteve P, Paje C and Lacal JC. (1998). Oncogene, 15, ChanAML,FlemingTP,McGovernES,ChedidM,MikiT 3047 ± 3057. and Aaronson SA. (1993). Mol. Cell. Biol., 13, 762 ± 768. Dhanasekaran N, Tsim ST, Dermott JM and Onesime D. Chant J and Stowers L. (1995). Cell, 81, 1±4. (1998). Oncogene, 17, 1383 ± 1394. Chardin P. (1993). Structural conservation of ras-related Dohlman HG, Caron MG and Lefkowitz RJ. (1987). proteins and its functional implications. In: GTPases in Biochemistry, 26, 2657 ± 2664. Biology I, Chapter 12. Dickey BF and Birnbaumer L. Ellis S and Mellor H. (2000). Curr. Biol., 10, 1387 ± 1390. (eds). Springer-Verlag: Berlin Heidelberg, Chapter 12. Even-Ram S, Uziely B, Cohen P, Grisaru-Granovsky S, Clark EA, Golub TR, Lander ES and Hynes RO. (2000). Maoz M, Ginzburg Y, Reich R, Vlodavsky I and Bar- Nature, 406, 532 ± 535. Shavit R. (1998). Nat. Med., 4, 909 ± 914.

Oncogene G protein-couple receptor oncoproteins and Rho GTPases IP Whitehead et al 1554 Fukuhara S, Chikumi H and Gutkind JS. (2000). FEBS Lett, Olson MF, Ashworth A and Hall A. (1995). Science, 269, 485, 183 ± 188. 1270 ± 1272. Fukuhara S, Murga C, Zohar M, Igishi T and Gutkind JS. Pace AM, Wong YH and Bourne HR. (1991). Proc. Natl. (1999). J. Biol. Chem., 274, 5868 ± 5879. Acad.Sci.USA,88, 7031 ± 7035. Glaven JA, Whitehead IP, Nomanbhoy T, Kay R and Parma J, Duprey L, Van Sande J, Cochaux P, Gervy C, Cerione RA. (1996). J. Biol. Chem., 271, 27374 ± 27381. Mochel J, Dumont J and Vassart G. (1993). Nature, 365, Gupta SK, Gallego C, Lowndes JM, Pleiman CM, Sable C, 649 ± 651. Eisfelder BJ and Johnson GL. (1992). Mol. Cell. Biol., 12, Poyner DR, Hawkins PT, Benton HP and Hanley MR. 190 ± 197. (1990). Biochem. J., 271, 605 ± 611. Gutkind JS. (1998). J. Biol. Chem., 273, 1839 ± 1842. Prendergast GC, Khosravi-Far R, Solski PA, Kurzawa H, Gutkind JS, Novotny EA, Brann MR and Robbins KC. Lebowitz PF and Der CJ. (1995). Oncogene, 10, 2289 ± (1991). Proc. Natl. Acad. Sci. USA, 88, 4703 ± 4707. 2296. Habets GGM, Scholtes EHM, Zuydgeest D, van der Qiu R-G, Abo A, McCormick F and Symons M. (1997). Mol. Kammen RA, Stam JC, Berns A and Collard JG. (1994). Cell. Biol., 17, 3449 ± 3458. Cell, 77, 537 ± 549. QiuR-G,ChenJ,KirnD,McCormickFandSymonsM. Hart MJ, Jiang X, Kozasa T, Roscoe W, Singer WD, Gilman (1995a). Nature, 374, 457 ± 459. AG, Sternweis PC and Bollag G. (1998). Science, 280, Qiu R-G, Chen J, McCormick F and Symons M. (1995b). 2112 ± 2114. Proc. Natl. Acad. Sci. USA, 92, 11781 ± 11785. HartMJ,SharmaS,elMasryN,QuiR-G,McCabeP, Ridley AJ. (1996). Curr. Biol., 6, 1256 ± 1264. Polakis P and Bollag G. (1996). J. Biol. Chem., 271, Ridley AJ and Hall A. (1992). Cell, 70, 389 ± 399. 25452 ± 25458. Ridley AJ, Paterson HF, Johnston CL, Diekmann D and Heasley LE, Storey B, Fanger GR, Butter®eld L, Zamarripa Hall A. (1992b). Cell, 70, 401 ± 410. J, Blumberg D and Maue RA. (1996). Mol. Cell. Biol., 16, Ridley AJ, Paterson HF, Johnston CL, Diekmann D and 648 ± 656. Hall A. (1992a). Cell, 70, 401 ± 410. Henrikson KP, Salazar SL, Fenton JW and Pentecost BT. Ross PC, Figler RA, Corjay MH, Barber CM, Adam N, (1999). Br.J.Cancer,79, 401 ± 406. Harcus DR and Lynch KR. (1990). Proc. Natl. Acad. Sci. Inglese J, Koch WJ, Touhara K and Lefkowitz RJ. (1995). USA, 87, 3052 ± 3056. Trends Biochem. Sci., 20, 151 ± 161. Seasholtz TM, Majumdar M and Brown JH. (1999). Mol. Jiang H, Wu D and Simon MI. (1993). FEBS Lett., 330, Pharmacol., 55, 949 ± 956. 319 ± 322. Seitz JF, Giovannini M, Gouvernet J and Gauthier M. Julius D, Livelli TJ, Jessell TM and Axel R. (1989). Science, (1991). J. Clin. Gastroenterol, 13, 1541 ± 1545. 244, 1057 ± 1062. Shenker A, Lane L, Kosugi S, Merendino JJ, Minegiski T Kabarowski J, Feramisco J, Le LQ, Gu J, Luoh S, Simon M and Cutler GB. (1993). Nature, 365, 652 ± 654. and Witte ON. (2000). Proc. Natl. Acad. Sci. USA, 97, Strader CD, Fong TM, Graziano MP and Tota MR. (1995). 12109 ± 12114. FASEB J, 9, 745 ± 754. Keely PJ, Westwick JK, Whitehead IP, Der CJ and Parise Strader CD, Fong TM, Tota MR, Underwood D and Dixon LV. (1997). Nature, 390, 632 ± 636. RAF. (1994). Annu.Rev.Biochem.,63, 101 ± 132. Khosravi-Far R, Solski PA, Kinch MS, Burridge K and Der Symons M. (1995). Curr. Opin. Biotechnol., 6, 668 ± 674. CJ. (1995). Mol. Cell. Biol., 15, 6443 ± 6453. Symons M. (1996). Trends Biochem. Sci., 21, 178 ± 181. Koch WJ, Hawes BE, Allen LF and Lefkowitz RJ. (1994). Van Aelst L and D'Souza-Schorey C. (1997). Genes Dev., 11, Proc. Natl. Acad. Sci. USA, 91, 12706 ± 12710. 2295 ± 2322. KourlasPJ,StroutMP,BecknellB,VeroneseML,Croce van Biesen T, Luttrell LM, Hawes BE and Lefkowitz RJ. CM, Theil KS, Krahe R, Ruutu T, Knuutila S, Bloom®eld (1996). Endocrine Rev., 17, 698 ± 714. CD and Caligiuri MA. (2000). Proc. Natl. Acad. Sci. USA, Van Obberghen-Schilling E and PouysseÂ gur J. (1993). 97, 2145 ± 2150. Thrombosis and Haemostasis, 70, 163 ± 167. Kozasa T, Jiang X, Hart MJ, Sternweis PM, Singer WD, Vara Prasad MVVS, Shore SK and Dhanasekaran N. (1994). Gilman AG, Bollag G and Sternweis PC. (1998). Science, Oncogene, 9, 2425 ± 2429. 280, 2109 ± 2111. Vignal E, De Toledo M, Comunale F, Ladopoulou A, Landis CA, Masters SB, Spada A, Pace AM, Bourne HR and Gauthier-Rouviere C, Blangy A and Fort P. (2000). J. Vallar L. (1989). Nature, 340, 692 ± 693. Biol. Chem., 275, 36457 ± 36464. Lyons J, Landis CA, Harsh G, Vallar L, Grunewarld K, Voyno-Yasenetskaya TA, Pace AM and Bourne HR. (1994). Feichtinger H, Duh Q-Y, Clark OH, Kawasaki E, Bourne Oncogene, 9, 2559 ± 2565. HR and McCormick F. (1990). Science, 249, 655 ± 699. Weng Z, Fluckiger AC, Nisitani S, Wahl MI, Le LQ, Hunter Mahmoud S, Staley J, Taylor J, Bogden A, Moreau JPI, Coy CA, Fernal AA, Le Beau MM and Witte ON. (1998). Proc. D, Avis I, Cuttitta F, Mulshine JL and Moody TW. Natl. Acad. Sci. USA, 95, 12334 ± 12339. (1991). Cancer Res., 51, 1798 ± 1802. WestwickJK,LeeRJ,LambertQT,SymonsM,PestellRG, MaoJ,YuanH,XieWandWuD.(1998).Proc. Natl. Acad. Der CJ and Whitehead IP. (1998). J. Biol. Chem., 273, Sci. USA, 95, 12973 ± 12976. 16739 ± 16747. Martin CB, Mahon GM, Klinger MB, Kay RJ, Symons M, Whitehead IP, Campbell S, Rossman KL and Der CJ. (1997). Der CJ and Whitehead IP. (2001). Oncogene, in press. Biochem. Biophys. Acta, 1332, F1 ± F23. Michiels F, Habets GG, Stam JC, van der Kammen RA and Whitehead IP, Khosravi-Far R, Kirk H, Trigo-Gonzalez G, Collard JG. (1995). Nature, 375, 338 ± 340. Der CJ and Kay R. (1996). J. Biol. Chem., 271, 18643 ± Monnot C, Weber V, Stinnakre J, Bihoreau C, Teutsch B, 18650. Corvol P and Clauser E. (1991). Mol. Endocrinol., 5, Whitehead IP, Kirk H and Kay R. (1995). Mol. Cell. Biol., 1477 ± 1487. 15, 704 ± 710. Oermanns S, Heiler E, Spicher K and Schultz G. (1994). Xu N, Bradley L, Ambdukar I and Gutkind JS. (1993). Proc. FEBS Lett., 349, 201 ± 204. Natl. Acad. Sci. USA, 90, 6741 ± 6745.

Oncogene G protein-couple receptor oncoproteins and Rho GTPases IP Whitehead et al 1555 Xu Y, Fang X-J, Casey G and Mills GB. (1995). Biochem. J., ZohnIE,KlingerM,KarpX,KirkH,SymonsM, 309, 933 ± 940. Chrzanowska-Wodnicka M, Der CJ and Kay RJ. (2000). YamamotoM,MaruiN,SakaiT,MoriiN,KozakiS,IkaiK, Oncogene, 19, 3866 ± 3877. Imamura S and Narumiya S. (1993). Oncogene, 8, 1449 ± Zohn IE, Symons M, Chrzanowska-Wodnicka M, Westwick 1455. JK and Der CJ. (1998b). Mol. Cell Biol., 18, 1225 ± 1235. Young D, O'Neill K, Jessell T and Wigler M. (1988). Proc. Zohn IE, Yu H, Li X, Cox AD and Earp HS. (1995). Mol. Natl. Acad. Sci. USA, 85, 5339 ± 5342. Cell Biol., 15, 6160 ± 6168. Young D, Waitches G, Birchmeier C, Fasano O and Wigler Zohn IE, Campbell SL, Khosravi-Far R, Rossman KL and M. (1986). Cell, 45, 711 ± 719. Der CJ. (1998a). Oncogene, 17, 1415 ± 1438.

Oncogene