Heterotrimeric G-Protein Bg-Dimers in Growth and Di€Erentiation

Home , GNB1, GNB2, GNB3, GNB4, GNB5, GNG11

Heterotrimeric G-protein bg-dimers in growth and dierentiation

William F Schwindinger1 and Janet D Robishaw*,1

1Weis Center for Research, Geisinger Clinic, 100 North Academy Avenue, Danville, Pennsylvania, PA 17822, USA

Heterotrimeric G-proteins are components of the signal ways in normal growth and development will facilitate transduction pathways for the soluble and cell-contact eorts to elucidate the pathogenesis of cancer. signals that regulate normal growth and dierentiation. The evidence that heterotrimeric G-proteins are a There is now a greater appreciation of the role of the relevant signal transduction pathway has been the Gbg-dimer in the regulation of a variety of intracellular subject of numerous recent reviews (Dhanaskaran et eectors, including ion channels, adenylyl cyclase, and al., 1998; Schwartz and Baron, 1999; Luttrell et al., phospholipase Cb. In many cases, Gbg-dimers are 1999a; Murga et al., 1999; Naor et al., 2000; Gutkind, required for the activation of mitogen activated protein 2000). Here we suggest that in many cases it is the kinase (MAPK) pathways that promote cellular pro- Gbg-dimers rather than the Ga-subunits that transmit liferation, although the underlying mechanisms have yet the signals for proliferation and survival. We review to be fully elucidated. Activation of phosphotidylinositol- the evidence that Gbg-dimers stimulate proliferation 3-kinase (PI3K) is a critical step in the intracellular via mitogen activated protein kinase (MAPK) path- transduction of survival signals. Gbg-dimers directly ways and promote cell survival by the activation of activate PI3Kg as well as the more widely distributed phosphotidylinositol-3-kinase (PI3K). We propose that PI3Kb. The activation of PI3Kg by Gbg-dimers likely speci®c Gbg-dimers are likely to have unique roles in involves direct binding of speci®c Gbg-dimers to both the regulation of normal growth and development, and subunits of PI3Kg. Thus, Gbg-dimers transmit signals consequently in the pathogenesis of cancer. from numerous receptors to a variety of intracellular eectors in distinct cellular contexts. Five distinct Gb- subunits and 12 distinct Gg-subunits have been identi®ed. Heterotrimeric G-proteins and the GTPase cycle New experimental approaches are needed to elucidate the speci®c roles of individual Gbg-dimers in the pathways Heterotrimeric G-proteins consist of a Ga-subunit and that transduce signals for proliferation and survival. aGbg-dimer. The Ga-subunit binds guanine nucleo- Oncogene (2001) 20, 1653 ± 1660. tides and possesses an intrinsic GTPase activity with a low turnover number. The Gbg-dimer does not Keywords: MAPK; PI3K dissociate under physiologic conditions. The heterotrimer is held at the inner surface of the plasma membrane through an isoprenyl group covalently Introduction attached to the Gg-subunit and by lipid groups attached to the Ga-subunit. Agonist occupied G- At the cellular level, cancer is characterized by protein coupled receptors (GPCR) catalyze the release proliferation (due to loss of cell cycle control or failure of GDP from the active site of the Ga-subunit. The of programmed cell death), by loss of the dierentiated Ga-subunit then binds GTP, resulting in a conforma- phenotype, and ultimately by invasiveness and metas- tional change that reveals eector interacting surfaces tasis. Cancer is caused by the stepwise accumulation of of both the Ga-subunit and the Gbg-dimer. Ga- mutations in proto-oncogenes and tumor-suppressor subunits and Gbg-dimers both initiate intracellular genes that subvert the normal cellular processes of signal transduction pathways through a variety of growth, survival, dierentiation and attachment. In eector molecules. Intracellular signaling is terminated normal growth and development, these cellular pro- by hydrolysis of GTP. The GTPase activity of the Ga- cesses are regulated by soluble signals and by contact subunit is accelerated by certain eectors and by RGS with adjacent cells or with components of the proteins (Wilkie, 2000). Signaling through the Gbg- extracellular matrix. Thus, mutations in cell surface dimer is limited by binding of the Gbg-dimer to receptors and in intracellular signal transduction path- phosducin (Craft et al., 1998) or pleckstrin (Abrams ways contribute to the development of malignancy. An et al., 1996). Agonist occupied receptor is down- understanding of the relevant signal transduction regulated by phosphorylation of the receptor by a G- pathways and of the interactions between these path- protein coupled receptor kinase (GRK), a process which requires Gbg-dimer (Pitcher et al., 1992). The accepted model of G-protein activation assumes that the Ga-subunit dissociates from the Gbg-dimer *Correspondence: JD Robishaw following binding of GTP (Gilman, 1987). A modi®ca- Gbg in cancer WF Schwindinger and JD Robishaw 1654 tion of this model suggests that the Ga-subunit and acid identity. The Gb5-subunit diers from the other Gbg-dimer remain associated throughout the GTPase Gb-subunits in gene structure and shares only 51 ± 53% cycle, but on binding GTP the heterotrimer adopts a amino acid identity. As a group, the Gg-subunits are more open conformation revealing eector interacting more diverse than the Gb-subunits, sharing only 27 ± surfaces (Rebois et al., 1997). This model is supported 76% amino acid identity. The Gg-subunits can be by chemical (Yi et al., 1991) and molecular (Klein et divided into families based on sequence homology and al., 2000) cross-linking studies which show that on dierences in post-translational modi®cations. Gg1, heterotrimers that are physically tethered are still Gg11, and Gg14 share 62 ± 73% amino acid identity, functional. Inability of Ga-subunits and Gbg-dimers and appear to form a distinct subclass. The recently to dissociate would have two important implications described Gg13 protein is the most distinct Gg-subunit, for G-protein mediated signaling. First, a mutation sharing only 24 ± 32% amino acid identity with the that constitutively activates a Ga-subunit would other Gg-subunits (Huang et al., 1999). simultaneously reveal eector interacting surfaces of the tethered Gbg-subunit, and thereby initiate Gbg- mediated signal transduction pathways. Second, there Post-translational modi®cations would be no pool of free Ga-subunits or free Gbg- dimers in the cell, which would increase the speci®city Gg-subunits undergo several post-translational mod- of heterotrimer interactions. i®cations. The best characterized of which involves Disturbances in the GTPase cycle lead to cellular isoprenylation of an invariant Cys residue in a transformation. In some cases, constitutively activating conserved CAAX motif at the carboxyl terminus. Most mutations of Ga-subunits lead to direct activation of Gg-subunits have a Leu at the carboxyl terminus which eectors which promote cellular proliferation. In this directs the addition of a 20-carbon geranylgeranyl regard, GTPase de®cient mutants of certain Ga- group, however three Gg-subunits (Gg1,Gg11,and subunits have been identi®ed in human tumors and Gg14) have a Ser at the carboxyl terminus and shown to transform cell lines in vitro, including gsp consequently a 15-carbon farnesyl group is added. (Gas), gip2 (Gai2) and gep (Ga12) (Dhanasekaran et al., The isoprenyl group is involved in anchoring the Gbg- 1998). However, this mechanism seems to be operative dimer to the membrane, and aects the interaction of in a minority of cases. A more common mechanism the Gbg-dimer with other proteins, including Ga- may involve prolonged release of Gbg-dimers. As we subunit, receptor, eector and GRK. For example, will discuss, over-expression of Gbg-dimers, but not of Gbg-dimers with the shorter farnesyl group are less constitutively active Ga-subunits, activates MAPK able than those with a geranylgeranyl group to pathways in most cell types (Crespo et al., 1994; Ito stimulate type II adenylyl cyclase and phospholipase- et al., 1995; Coso et al., 1996; Yamauchi et al., 1997). Gb (Myung et al., 1999). Consequently, Gg1,Gg11,and Although overexpression of Gbg-dimers has not been Gg14 form a functional subclass that is likely to have shown to stimulate cell proliferation or cause malig- distinct interactions with membranes, receptors and nant transformation (Dhanasekaran et al., 1998), eectors; while expression of Gg1 is restricted to the sequestration of Gbg-dimers prevents the stimulation retina and brain, Gg11 and Gg14 are more widely of DNA synthesis and cell proliferation in several distributed (Balcueva et al., 2000). experimental systems (Iaccarino et al., 1999; Ghahre- Isoprenylation of the Gg-subunit is followed by mani et al., 2000). cleavage of the last three amino acids and methylation of prenylcysteine. Gg5 is the only Gg-subunit that has an aromatic residue in the CAAX motif; this prevents Diversity of Gbg-dimers the cleavage of the last three amino acids in a fraction of processed Gg5 (Cook et al., 1998). Cleavage of the As shown in Table 1, ®ve distinct Gb-subunits and 12 last three amino acids is important for intracellular distinct Gg-subunits have been identi®ed in human localization of isoprenylated proteins. De®ciency of the cDNA, EST, or genomic libraries (Hurowitz et al., endopeptidase Rce1 in mice prevents processing of 2000). Homologs of each of these proteins have also both farnesylated (Ha-Ras, N-Ras, Gg1) and geranyl- been identi®ed in one or more additional mammalian geranylated (Ki-Ras, Rap1B) proteins, resulting in species including mouse, rat, cow, and dog. Individual mislocalization of at least one isoprenylated protein Gb-subunits and Gg-subunits share nearly 100% (N-Ras) within the cell (Kim et al., 1999). amino acid identity between mammalian species, One potential role for Gg-subunits in the treatment suggesting conserved functions for each individual of cancer is as targets for farnesyltransferase inhibitors subunit. Moreover, the tissue speci®c expression of (FTIs). Peptide mimetic compounds of the CAAX individual Gg-subunits indicates unique roles in speci®c motif function as inhibitors of farnesyltransferase (FT) signal transduction pathways. For example, the and geranylgeranyltransferase type I (GGT-I). FTIs expression of Gg1 is restricted to retinal rod cells, selectively inhibit FT, and block farnesylation of Ras, and Gbg1-dimers interact better with rhodopsin than Rho and presumably other farnesylated proteins with other receptors (Kisselev and Gautam, 1993). (Lebowitz and Prendergast, 1998). In models of Ras The Gb-subunits are highly homologous to each transformation, FTIs suppress anchorage-independent other, with Gb1 through Gb4 sharing 78 ± 88% amino growth in cell culture and slow tumor growth in nude

Oncogene Gbg in cancer WF Schwindinger and JD Robishaw 1655 Table 1 Human G-protein b and g-subunits Gene Protein Unigene or Cytogenetic cDNA Sources name name (s) Genbank position (UniGene ESTs)

GNB1 b1 Hs.215595 1p36.21 ± 36.33 WIDE+Adipose, adrenal gland, cervix, esophagus, eye, liver, small intestine, stomach, thymus, thyroid GNB2 b2 Hs.273457 7q21.3 ± q22.1 (WIDE7Aorta, Bone, Skin)+adrenal gland, esophagus, larynx, stomach, synovial membrane, thymus GNB3 b3 Hs.71642 12p13 Brain, colon, eye, germ cell, heart, lung, pancreas (GNB4) b4 Hs.172654 3q26 ± q27 Brain, breast, heart, kidney, liver, lung, lymph, muscle, ovary, pancreas, placenta, uterus GNB5 b5 Hs.275353 15q21 Brain, eye, germ cell, lung, lymph, ovary GNGT1 g1, grod Hs.73112 7q21.3 Brain, eye GNGT2 g14, gcone, g8 Hs.181781 17q21 Eye, colon, heart, kidney, lung, lymph, skin, tonsil, uterus GNG2 g2, g6 Hs.23767 14q21 (WIDE7Parathyroid)+Adrenal gland, cervix, ear, eye, liver, stomach, tongue GNG3 g3 Hs.179915 11p11 Blood, brain, germ cell GNG4 g4 Hs.32976 1 Brain, breast, colon, germ cell, kidney, lung, prostate GNG5 g5 Hs.5322 1p22 (WIDE7Germ Cell)+adipose, ear, esophagus, skin GNG7 g7 Hs.127828 19p13.3 Brain, colon, eye, germ cell, kidney, lung, thymus, testis (GNG8) g8, g9 AF188179 19q13.2 ± q13.3 GNG10 g10 Hs.79126 9q31 ± q32 (WIDE7Blood, Muscle)+Bladder, stomach, thyroid GNG11 g11 Hs.83381 7q21.3 Aorta, brain, esophagus, foreskin, heart, kidney, lung, ovary, pancreas, placenta, prostate, spleen, stomach, testis, thymus, uterus (GNG12) g12 AF188181 1p31 ± p33 (GNG13) g13 AL031033 16p13.3

(WIDE=Aorta, Blood, Bone, Brain, Breast, Colon, Foreskin, Germ Cell, Heart, Kidney, Lung, Lymph, Muscle, Ovary, Pancreas, Parathyroid, Placenta, Prostate, Skin, Testis, Tonsil, and Uterus). Gene names in parentheses are not ocial mice. Recent work has shown that FTIs suppress proliferation of cancer cells through a mechanism other than inhibition of Ras isoprenylation (Prendergast, 2000). Several Gg-subunits, Gg1,Gg11 and Gg14, are farnesylated, and are potential targets for the action of FTIs. Consequently, future studies ought to examine the role of Gg-subunits in the antineoplastic actions of FTIs, and consider the potential toxicity of inhibiting farnesylation of Gg-subunits.

Gbg-regulated eectors

As shown in Figure 1, Gbg-dimers mediate signal transduction by interaction with numerous other proteins in the cell, including Ga-subunits, GPCRs, Figure 1 Interactions of Gbg-dimers in the GTPase cycle and GRKs, phosducin, pleckstrin, and various eectors signal transduction pathways (Clapham and Neer, 1997). Gbg-dimers participate in receptor activation, the GTPase cycle, receptor desen- sitization, and eector activation. Gbg-dimers bind dimers also activate the MAPK kinase pathway and directly to GPCRs and enhance the binding of Ga- stimulate both PI3Kg and PI3Kb. subunits to GPCRs. The interaction of the Gbg-dimer A more thorough examination of the regulation by with the Ga-subunit covers the eector interacting Gbg-dimers of adenylyl cyclase and PLCb is appro- surfaces of both proteins. Gbg-dimers function as GDP priate in light of the involvement of these eectors in release inhibiting factors for Ga-subunits. Gbg-dimers the regulation of growth and dierentiation. Adenylyl are required for phosphorylation of receptors by some cyclase produces the intracellular second messenger GRKs. Phosducin, pleckstrin, and possibly other cAMP, which in turn regulates the activity of protein proteins with pleckstrin homology domains bind Gbg- kinase A. All nine isoforms of adenylyl cyclase are dimers, and inhibit Gbg-mediated signaling. Various stimulated by Gas, and most isoforms are inhibited by eectors have been shown to be regulated by Gbg- a-subunits of the Gi family, with the exception of Type dimers including adenylyl cyclase, phospholipase Gb II (Simonds, 1999). Types II and IV adenylyl cyclase (PLCb), inward recti®er G-protein gated potassium are directly activated by Gbg-dimers, Type I adenylyl channels (GIRK), and N and P/Q type calcium cyclase is inhibited, and Type III is unaected channels. As discussed in the following sections, Gbg- (Simonds, 1999). Activation of PLC leads to the

Oncogene Gbg in cancer WF Schwindinger and JD Robishaw 1656 production of the intracellular second messengers Gat (Faure et al., 1994; Crespo et al., 1994), Gao (Ito et inositol-1,4,5-trisphosphate and diacylglycerol, which al., 1995) and a carboxyl terminal fragment of the b- regulate the release of intracellular calcium and the adrenergic receptor kinase (bARK-CT) (Koch et al., activity of protein kinase C (PKC). PLCg isoforms are 1994). Similar studies have shown that Gbg-dimers also activated by tyrosine phosphorylation and by interac- couple GPCR to the stimulation of JNK (Coso et al., tion with receptor tyrosine kinases (RTKs) through 1996) and p38 MAPK (Yamauchi et al., 1997). their SH2 domains. PLCb isoforms are activated Ga- There appears to be substantial overlap in the subunits of the Gaq family and by Gbg-dimers. Four pathways from Gbg-dimer or RTK to the activation isoforms of PLCb have been cloned, and the G-protein of MAPK. First, Gbg-dependent pathways and RTK- activation of PLCb has been studied in detail (Singer et dependent pathways converge with the activation of a al., 1997). Direct comparisons have shown that PLCb1 small G-protein (Gutkind, 2000). Activation of ERK2 and PLCb3 are more responsive to activation by Gaq by Gbg-dimer is blocked by co-transfection with a than is PLCb2. PLCbs are activated by Gbg-dimers dominant negative mutant of Ras, N17-Ras (Koch et with a rank order of PLCb34PLCb24PLCb1, while al., 1994). Activation of JNK by Gbg-dimer is blocked PLCb4 is not activated by Gbg-dimers. Only PLCb3is by co-transfection with either N17-Ras or N17-Rac further activated by Gaq in the presence of Gbg-dimer. (Coso et al., 1995). Second, the pathway by which There are unique sites on PLCb for interaction with Gbg-dimers stimulate MAPK may include a non- Gaq and Gbg-dimer (Smrcka and Sternweis, 1993). As receptor tyrosine kinase. For example, overexpression we will see below, there are many parallels in the of Gbg-dimers stimulates autophosphorylation of Src activation of PLC and PI3K by Gbg-dimers and (Luttrell et al., 1996). Other non-receptor tyrosine RTKs. kinases may replace Src in certain cell types. In lymphoid cells, Csk and Lyn are required for Gq coupled stimulation of MAPK, while Btk is required Stimulation of MAPK pathways by Gbg-dimers for Gi coupled stimulation (Wan et al., 1997). Finally, in certain cells, the activation of MAPK by Gbg-dimers Two classes of growth factor receptors cooperate to may also require trans-activation of RTKs or assembly regulate responses to mitogenic stimuli: receptor of focal adhesions as a scaold for assembly of the Ras tyrosine kinases and G-protein coupled receptors activation complex (Della Rocca et al., 1999). (Schwartz and Baron, 1999). Evidence for proliferative Several intracellular eectors have been proposed as signaling and activation of MAPK by GPCR has the ®rst step towards activation of MAPK by Gbg- recently been reviewed (Dhanaskaran et al., 1998; dimers: (i) PI3K was found to be an early intermediate Gutkind, 2000). However, the mechanisms whereby in Gbg-stimulated MAPK activity (Hawes et al., 1996). GPCR activate MAPK cascades are not well under- Direct activation of PI3K by Gbg-dimers could stood (Luttrell et al., 1999a; Murga et al., 1999; Naor increase the activity of Src family tyrosine kinases et al., 2000, Gutkind, 2000). It must be recognized that and activate MAPK (Bondeva et al., 1998); (ii) Gbg- there are several MAPK pathways, including pathways dimers may activate non-receptor tyrosine kinases that terminate with extracellular signal-regulated directly; (iii) Activation of PLCb by Gbg-dimers may kinases (ERK1, ERK2), jun N-terminal kinases lead to activation of MAPK pathway, as does (JNK1, JNK2, JNK3), p38 MAPKs (p38a, p38b, activation of PLCb by Gaq. p38g, and p38d), and big MAPK (ERK5) (Garrington Gbg-dimers may serve to recruit a scaolding for the and Johnson, 1999). The mechanism of activation of assembly of the Ras activation complex. Binding of b- these distinct MAPK cascades by GPCR may dier. arrestin to a phosphorylated GPCR serves to down- Moreover, there appear to be diering cell type-, regulate signal transduction and initiate receptor GPCR-, and G-protein-dependent mechanisms of internalization. However, b-arrestin may also function activation of particular MAPK cascades (Rozengurt, as an adapter protein for the binding of Src and 1998). activation of the MAPK pathway (Luttrell et al., Early studies showed that Gbg-dimers couple GPCR 1999b). In this model, Gbg-dimers are required for to the stimulation of the ERK1/ERK2 pathway. Gbg- translocation of GRK2 to the membrane and phos- dimers stimulate MAPK under conditions where phorylation of GPCR (Pitcher et al., 1992). constitutively active mutants of Ga-subunits do not An alternative scaolding protein may be KSR-1, stimulate ERK2 (Crespo et al., 1994; Ito et al., 1995), which has binding sites for Raf-1, MEK and MAPK. and under conditions where constitutively active KSR-1 is a positive regulator of Ras-mediated mutants of Gas or Gaq, but not Gai, stimulate ERK1 signaling, initially identi®ed in genetic screens of (Faure et al., 1994). Stimulation of ERK1/ERK2 Drosphilia and C. elegans (Downward, 1995). In a requires a functional Gbg-dimer: Gb-subunits or Gg- yeast two hybrid screen KSR-1 was shown to interact subunits alone do not stimulate ERK1/ERK2 (Faure et with Gg2,Gg3, and Gg10. Moreover, KSR-1 was found al., 1994; Crespo et al., 1994; Ito et al., 1995), nor do to translocate to the plasma membrane in COS-7 cells isoprenylation de®cient mutants of Gg-subunits, even stimulated with LPA, an eect that was blocked by in combination with Gb-subunits (Crespo et al., 1994). pertussis toxin but not by N17-Ras, suggesting an Agents that bind to and sequester Gbg-dimers prevent interaction of Gbg-dimer with KSR-1 in vivo (Bell et the stimulation of ERK1/ERK2 by GPCR, including al., 1999). Although KSR-1 is a positive regulator of

Oncogene Gbg in cancer WF Schwindinger and JD Robishaw 1657 Ras, overexpression of KSR-1 inhibited Gb1g3 stimu- class IB enzyme (Stephens et al., 1994). In vitro, Class I lated MAPK activity in COS-7 cells (Bell et al., 1999). PI3Ks phosphorylate PtdIns, PtdIns-4-P, PtdIns-5-P However, the level of KSR-1 overexpression may be and PtdIns-4,5-P2 on the D3 position of the inositol critical in determining its eect. In Xenopus oocytes ring. In vivo, PtdIns-4,5-P2 is the predominant KSR-1 enhanced V12-Ras signaling at low levels of substrate for Class I PI3Ks. Class I PI3Ks also possess expression but blocked V12-Ras signaling at high levels a serine kinase activity, undergo autophosphorylation, of expression (Cacace et al., 1999). These results are and phosphorylate IRS-1 and other substrates (Fru- consistent with a model of MAPK activation in man et al., 1998). Class IA PIK3s, are activated by the mammalian cells that is similar to the model for S. small G-protein Ras and by association of the cerevisiae, and suggest that KSR-1 acts as a scaolding regulatory subunit with proteins containing phospho- protein similar to ste5 (Choi et al., 1994). tyrosyl residues (pY) within a speci®c sequence context In summary, there is ample evidence that Gbg- (pYMXM). The Class IB PI3K is activated by Gbg- dimers transmit signals from GPCR to activate MAPK dimers (Stephens et al., 1994) and also binds to Ras pathways in speci®c cellular contexts. However, the (Rubio et al., 1997). Class I PI3Ks are inhibited by the pathways by which Gbg-dimers activate MAPK fungal metabolite wortmannin (Stoyanova et al., 1997). cascades are not fully elucidated. Studies have shown The activity of p110g is modulated by p101, but the that various components of the pathway from RTK to function of p101 diers from that of the Class IA MAPK are activated by overexpression of Gbg-dimers, regulatory subunits. When expressed in SF9 cells or and that stimulation of MAPK by either RTK or Gbg- COS-7 cells, p110g has a signi®cant basal activity that dimer requires the activation of a small G-protein. The is only slightly stimulated by Gbg-dimers, but when initial step in the pathway from Gbg-dimer to MAPK p110g is co-expressed with its associated subunit, the activation is not de®ned; there is evidence to support p101/p110g heterodimer has a lower basal activity and activation of an intracellular eector (e.g. PI3K, PLCb, is stimulated to a much greater extent by Gbg-dimers or non-receptor tyrosine kinase) or assembly of a (Stephens et al., 1997). While puri®ed p110g has kinase scaold for Ras activation (e.g. GPCR or KSR-1). An activity towards several 3-OH-PtdIns, addition of p101 intriguing possibility is that each of these mechanisms increases speci®city for the PtdIns-4,5-P2 (Maier et al., may be operable for a particular Gbg-dimer in a 1999). All catalytically active fragments of p110g that particular cellular context. are capable of binding p101 are equally sensitive to stimulation by Gbg-dimers, but those incapable of binding p101 are insensitive to Gbg-stimulation Activation of PI3K by Gbg-dimers (Krugmann et al., 1999). Thus, the function of the p101-subunit is to increase substrate speci®city for Survival signals are mediated by three classes of PtdIns-4,5-P2, and to increase sensitivity to Gbg- transmembrane receptors: RTK (e.g. IGF I receptor), stimulation. receptors that are coupled to non-receptor tyrosine The mechanism whereby PI3Kg is activated by Gbg- kinases (e.g. integrins), and GPCR (e.g. LPA recep- dimers is not completely de®ned, but likely involves tors). Distinct intracellular signal transduction path- direct binding of Gbg-dimers to the p101/p110g- ways from each of these receptor classes converge upon heterodimer. Direct binding of Gbg-dimer to p110g- activation of PI3K. This leads to production of an subunit is evidenced by co-puri®cation of Gbg-dimers intracellular second messenger PtdIns-3,4,5-P3 and the with a p110g-GST fusion protein (Leopoldt et al., activation of PKB (also known as c-Akt). The central 1998). However, a deletion analysis of p110g-GST role of PtdIns-(3,4,5)-P3 in cell survival is supported by fusion protein appears to indicate that Gbg-dimer studies of tumor suppressor genes PTEN and SHIP. binds to multiple regions of the p110g-subunit. PTEN is a phophatase that is active on PtdIns-3-Ps, Interactions of multiple regions of p110g-subunit with while SHIP is active on PtdIns-5-Ps. In mice, de®ciency Gbg-dimers is further supported by the observation of PTEN is associated with failure of developmental that both N-and C-terminal fragments of p110g inhibit apoptosis and resistance to apoptotic stimuli. Similarly, Gbg-dimer stimulated PI3Kg activity (Leopoldt et al., de®ciency of SHIP causes excessive survival of myeloid 1998). Involvement of the p101-subunit in binding to lineages. PKB promotes survival at multiple levels by Gbg-dimer, is supported by a deletion analysis of p101 inactivating components of the apoptotic machinery, which indicates that Gbg-stimulated activity requires by inhibiting transcription of death genes, and by the N-terminus of p101 and correlates with the inducing expression of survival genes (Datta et al., tightness of binding of p101 to p110g (Krugmann et 1999). Here we will focus on the role of Gbg-dimers in al., 1999). the activation of PI3K, a central step in G-protein Speci®c Gbg-dimers may be involved in the activa- coupled survival signaling. tion of PI3Kg.Gbg-dimer isolated from retina is less Class I PI3Ks are heterodimeric proteins consisting eective than Gbg-dimers isolated from brain at of a catalytic subunit (p110) and an associated subunit activating PI3Kg (Leopoldt et al., 1998). This may (Fruman et al., 1998). Class IA PI3K include p110a, result from the dierences in post-translational mod- p110b and p110d, associated with various regulatory i®cation with retinal Gg1 being mostly farnesylated and subunits, p85a, p85b and p55g. While p110g with its brain Gg-subunits being mostly geranylgeranylated. associated subunit, p101, constitutes the only known The Gb-subunits also play a role in determining

Oncogene Gbg in cancer WF Schwindinger and JD Robishaw 1658 speci®city. Gb1g2,Gb2g2,Gb3g2 all stimulate PI3Kg chromosome 17 containing Gg7 has been observed in a with similar ecacy, but Gb5g2 does not activate variety of human cancers including gastrointestinal PI3Kg (Maier et al., 2000). tract cancers. Expression of Gg7 in a human esophageal The more widely expressed PI3Kb is also activated carcinoma cell, KYSE150, line was associated with by Gbg-dimers. Early studies found that PI3Kb was suppressed cell growth in con¯uent cultures, and stimulated by Gbg-dimer and pY-peptide synergisti- smaller tumor size in nude mice (Shibata et al., 1999). cally (Kurosu et al., 1997). When the various PI3K Two recent studies have demonstrated a decrease in isoforms are expressed in Sf9 insect cells, PI3Kg is only the levels of Gg5 during dierentiation. In the stimulated by Gbg-dimer, while PI3Ka and PI3Kd are developing rat brain, Gg5 levels are high in early only stimulated by pY-peptides, but PI3Kb is embryonic stages, fall progressively until birth, and stimulated by both, synergistically (Maier et al., remain low after birth. Gg5 is localized to the 1999). Co-transfection studies in NIH3T3 cells show undierentiated, proliferative region of the ventricular that both PI3Kg and PI3Kb, but not PI3Ka,will zone. Moreover, in P19 cells induced to dierentiate activate PKB in response to GPCR agonists. More- into neurons in vitro,Gg5 levels fall progressively over, activation of PKB by GPCR in cells co- (Morishita et al., 1999). Levels of Gg5 protein were transfected with PI3Kb, does not require phosphoryla- shown to decrease during the ®rst 2 days of in vitro tion of the regulatory subunit of PI3Kb, and is dierentiation of 3T3-L1 cells into adipocytes. More- mediated by Gbg-dimers rather than Ga-subunits over, a Gbg5-dimer was found to be associated with (Murga et al., 2000). AEBP1 in the nucleus of 3T3-L1 cells and to attenuate The importance of Gbg-dimers and PI3Kg in its transcriptional repression activity (Park et al., 1999). regulating growth has been explored through studies As detailed above, Gg5 may be uniquely suited to of transgenic animals. Mice that are homozygous for a translocate to the nucleus because of its distinct post- deletion of the PI3Kg gene are viable and develop translational modi®cations (Cook et al., 1998). normally, but begin to develop colorectal tumors at 12 weeks of age and by 6 months of age approximately 40% have visible tumors. The colonic tumors exhibit Conclusions cellular hyperproliferation in association with increased levels of Bcl-2 protein as well as several cell cycle Heterotrimeric G-proteins couple cell surface receptors proteins (Sasaki et al., 2000). In mice, transverse aortic for signals that regulate proliferation and survival to constriction results in a 48% increase in the ratio of intracellular eectors. Gbg-dimers function at many left ventricular weight to body weight over the course levels in the GTPase cycle of heterotrimeric G-proteins. of one week, which is associated with increased PI3Kg Disturbances in the GTPase cycle lead to cellular activity and increased PKB activity. In transgenic mice transformation. In many cases, Gbg-dimers rather than expressing a Gbg-sequestrant, there is no increase in Ga-subunits are required for the activation of eectors, PI3Kg activity nor in PKB activity one week after such as MAPK, that promote cellular proliferation. imposition of transverse aortic constriction, however The pathway(s) linking GPCRs to MAPKs have yet to there is still a similar 43% increase in the ratio of left be fully elucidated, but appear to converge with the ventricular weight to body weight (Naga Prasad et al., pathways from RTKs to MAPKs. Gbg-dimers may be 2000). Adenovirus mediated gene transfer of a Gbg- required for activation of an intracellular eector (e.g. sequestrant, bARK-CT, to vascular smooth muscle P13K, PLCb, Src) or assembly of a scaolding for cells of rat carotid arteries, substantially inhibited activation of a small G-protein (e.g. Ras). Dierent intimal hyperplasia following injury by balloon cells may employ dierent mechanisms, and conse- angioplasty (Iaccarino et al., 1999). quently dierent Gbg-dimers, in the activation of MAPKs. Activation of PI3K is a critical step in the intracellular transduction of survival signals, as well as Additional roles of Gbg-dimers in proliferation and a candidate for the initial step in the pathway from dierentiation GPCR to MAPKs. Gbg-dimers directly activate PI3Kg, and both Gbg-dimers and pY-peptides activate Although there have been no reports of mutations in the more widely distributed PI3Kb. The activation of Gb-orGg-subunits in cancer, a few groups have noted PI3Kg by Gbg-dimers likely involves direct binding of changes in the level of expression of Gg-subunits in the Gbg-dimer to both the catalytic and regulatory tumor tissues. One group has suggested a role for Gg7 subunits. Speci®c Gbg-dimers may be involved in the in the pathogenesis of human intestinal carcinoma. activation of PI3Kg, e.g. Gb5g2 does not activate Levels of Gg7 were found to be reduced in 12/12 PI3Kg. specimens of primary pancreatic cancer (Shibata et al., Recent studies have begun to de®ne functional 1998), and in 24/30 gastrointestinal tract cancers dierences between subtypes of Gb-subunits and Gg- (Shibata et al., 1999). The mechanism by which Gg7 subunits. However, more studies examining speci®c expression is reduced was not addressed in these studies. functions of the ®ve individual Gb-subunits and the 12 Examination of the data in The Cancer Chromosome individual Gg-subunits are needed. Experimental Aberation Project (www.ncbi.nlm.nih.gov/CCAP) techniques that sequester Gbg-dimers have demon- shows that loss of heterozygosity in the region of strated that Gbg-dimers are involved in signal

Oncogene Gbg in cancer WF Schwindinger and JD Robishaw 1659 transduction pathways that mediate survival and experimental approaches are needed to elucidate the proliferation, but these techniques cannot distinguish speci®c roles of individual Gbg-dimers. These may between Gbg-dimers with distinct subunit composi- include suppression of mRNA for individual Gb-or tions. Experimental designs involving overexpression or Gg-subunits by antisense RNA or by ribozymes, or reconstitution are not well suited to discover functional targeted disruption of individual Gb-orGg-subunits dierences between related proteins. High levels of genes. Development of transgenic or `knock-out' mice protein expression may force association of subunits with reduced or absent expression of particular Gb or that do not interact at physiologic concentrations; the Gg-subunits will likely elucidate the speci®c roles of use of detergents may result in mixing of subunits that individual subunits in the signal transduction pathways do not normally have access to each other. New that govern growth and dierentiation.

References

Abrams CS, Zhang J, Downes CP, Tang XW, Zhao W and Ito A, Satoh T, Kaziro Y and Itoh H. (1995). FEBS Letts, Rittenhouse SE. (1996). J. Biol. Chem., 271, 25192 ± 368, 183 ± 187. 25197. Kisselev O and Gautam N. (1993). J. Biol. Chem., 268, Balcueva EA, Wang Q, Hughes H, Kunsch C and Robishaw 24519 ± 24522. JD. (2000). Exp. Cell. Res., 257, 310 ± 319. Kim E, Ambroziak P, Otto JC, Taylor B, Ashby M, Shannon Bell B, Xing H, Yan K, Gautam N and Muslin AJ. (1999). K, Casey PJ and Young SG. (1999). J. Biol. Chem., 274, J. Biol. Chem., 274, 7982 ± 7986. 8383 ± 8390. Bondeva T, Pirola L, Bulgarelli-Leva G, Rubio I, Wetzker R Klein S, Reuveni H and Levitzki A. (2000). Proc. Natl. Acad. and Wymann MP. (1998). Science, 282, 293 ± 296. Sci. USA, 97, 3219 ± 3223. Cacace AM, Michaud NR, Therrien M, Mathes K, Copeland Koch WJ, Hawes BE, Allen LF and Lefkowitz RJ. (1994). T, Rubin GM and Morrison DK. (1999). Mol. Cell. Biol., Proc. Natl. Acad. Sci. USA, 91, 12706 ± 12710. 19, 229 ± 240. Krugmann S, Hawkins PT, Pryer N and Braselmann S. Choi KY, Satterberg B, Lyons DM and Elion EA. (1994). (1999). J. Biol. Chem., 274, 17152 ± 17158. Cell, 78, 499 ± 512. Kurosu H, Maehama T, Okada T, Yamamoto T, Hoshino S, Clapham DE and Neer EJ. (1997). Annu.Rev.Pharmacol. Fukui Y, Ui M, Hazeki O and Katada T. (1997). J. Biol. Toxicol., 37, 167 ± 203. Chem., 272, 24252 ± 24256. Cook LA, Schey KL, Wilcox MD, Dingus J and Hildebrandt Lebowitz PF and Prendergast GC. (1998). Oncogene, 17, JD. (1998). Biochemistry, 37, 12280 ± 12286. 1439 ± 1445. Coso OA, Teramoto H, Simonds WF and Gutkind JS. Leopoldt D, Hanck T, Exner T, Maier U, Wetzker R and (1996). J. Biol. Chem., 271, 3963 ± 3966. Nurnberg B. (1998). J. Biol. Chem., 273, 7024 ± 7029. Craft CM, Xu J, Slepak VZ, Zhan-Poe X, Brown B and Luttrell LM, Hawes BE, van Biesen T, Luttrell DK, Lansing Lolley RN. (1998). Biochemistry, 15758 ± 15772. TJ and Lefkowitz RJ. (1996). J. Biol. Chem., 271, 19443 ± Crespo P, Xu N, Simonds WF and Gutkind JS. (1994). 19450. Nature, 369, 418 ± 420. Luttrell LM, Daaka Y and Lefkowitz RJ. (1999a). Curr. Datta SR, Brunet A and Greenberg ME. (1999). Genes Dev., Opin. Cell Biol., 11, 177 ± 183. 13, 2905 ± 2927. Luttrell LM, Ferguson SSG, Daaka Y, Miller WE, Maudsley Della Rocca GJ, Maudsley S, Daaka Y, Lefkowitz RJ and S, DellaRocca GJ, Lin F-T, Kawakatsu H, Owada K, Luttrell LM. (1999). J. Biol. Chem., 274, 13978 ± 13984. Luttrell DK, Caron MG and Lefkowitz RJ. (1999b). Dhanasekaran N, Tsim S-T, Dermott JM and Onesime D. Science, 283, 655 ± 661. (1998). Oncogene, 17, 1383 ± 1394. MaierU,BabichA,MacrezN,LeopoldtD,GierschikP, Downward J. (1995). Cell, 83, 831 ± 834. Illenberger D and Nurnberg B. (2000). J. Biol. Chem., 275, Fruman DA, Meyers RE and Cantley LC. (1998). Annu. Rev. 13746 ± 13754. Biochem., 67, 481 ± 507. Maier U, Babich A and Nurnberg B. (1999). J. Biol. Chem., Faure M, Voyno-Yasenetskaya TA and Borne HR. (1994). J. 274, 29311 ± 29317. Biol. Chem., 269, 7851 ± 7854. Morishita R, Shinohara H, Ueda H, Kato K and Asano T. Garrington TP and Johnson GL. (1999). Curr. Opin. Cell (1999). J. Neurochem., 73, 2369 ± 2374. Biol., 11, 211 ± 218. Murga C, Fukuhara S and Gutkind JS. (1999). Trends. Ghahremani MH, Forget C and Albert PR. (2000). Mol. Endocrinol. Metab., 10, 122 ± 127. Cell. Biol., 20, 1497 ± 1506. Murga C, Fukuhara S and Gutkind JS. (2000). J. Biol. Gilman AG. (1987). Annu. Rev. Biochem., 56, 615 ± 649. Chem., 275, 12069 ± 12073. Gutkind JS. (2000). Science's STKE, www.stke.org/cgi/ Myung C-S, Yasuda H, Liu WW, Harden TK and Garrison content/full/OC sigtrans;2000/40/re1 JC. (1999). J. Biol. Chem., 274, 16595 ± 16603. Hawes BE, Luttrell LM, van Biesen TV and Lefkowitz RJ. Naga Prasad SV, Esposito G, Mao L, Koch WJ and (1996). J. Biol. Chem., 271, 12133 ± 12136. Rockman HA. (2000). J. Biol. Chem., 275, 4693 ± 4698. Huang L, Shanker YG, Dubauskaite J, Zheng JZ, Yan W, Naor Z, Benard O and Seger R. (2000). Trends. Endocrinol. Rosenzweig S, Spielman AI, Max M and Margolskee RF. Metab., 11, 91 ± 99. (1999). Nature Neurosci., 2, 1055 ± 1062. Park J-G, Muise A, He G-P, Kim S-W and Ro H-S. (1999). Hurowitz EH, Melnyk JM, Chen Y-J, Kouros-Mehr H, EMBO J., 18, 4004 ± 4012. Simon MI and Shizuya H. (2000). DNA. Res., 7, 111 ± 120. Pitcher JA, Inglese J, Higgins JB, Arriza JL, Casey PJ, Kim IaccarinoG,SmithwickLA,LefkowitzRJandKochWJ. C, Benovic JL, Kwatra MM, Caron MG and Lefkowitz (1999). Proc. Natl. Acad. Sci. USA, 96, 3945 ± 3950. RJ. (1992). Science, 257, 1264 ± 1267.

Oncogene Gbg in cancer WF Schwindinger and JD Robishaw 1660 Prendergast GC. (2000). Curr. Opin. Cell Biol., 12, 166 ± 173. Smrcka AV and Sternweis PC. (1993). J. Biol. Chem., 268, Rebois RV, Warner DR and Basi NS. (1997). Cell. Signal., 9, 9667 ± 9674. 141 ± 151. Stephens L, Smrcka A, Cooke FT, Jackson TR, Sternweis Rozengurt E. (1998). J. Cell. Physiol., 177, 507 ± 517. PC and Hawkins PT. (1994). Cell, 77, 83 ± 93. Rubio I, Rodriguez-Viciana P, Downward J and Wetzker R. Stephens LR, Eguinoa A, Erdjument-Bromage H, Lui M, (1997). Biochem. J., 326, 891 ± 895. Cooke F, Coadwell J, Smrcka AS, Thelen M, Cadwallader Sasaki T, Irie-Sasaki J, Horie Y, Bachmaier K, Fata JE, Li K, Tempst P and Hawkins PT. (1997). Cell, 89, 105 ± 114. M, Suzuki A, Bouchard D, Ho A, Redston M, Gallinger S, Stoyanova S, Bulgarelli-Leva G, Kirsch C, Hanck T, Klinger Khokha R, Mak TW, Hawkins PT, Stephens L, Scherer R, Wetzker R and Wymann MP. (1997). Biochem. J., 324, SW, Tsao M and Penninger JM. (2000). Nature, 406, 897 ± 489 ± 495. 902. Wan Y, Bence K, Hata A, Kurosaki T, Veillette A and Schwartz MA and Baron V. (1999). Curr. Opin. Cell Biol., 11, Huang X-Y. (1997). J. Biol. Chem., 272, 17209 ± 17215. 197 ± 202. Wilkie TM. (2000). Annu. Rev. Biochem., 69, 795 ± 827. Shibata K, Mori M, Tanaka S, Kitano S and Akiyoshi T. Yamauchi J, Nagao M, Kaziro Y and Itoh H. (1997). J. Biol. (1998). Biochem. Biophys. Res. Comm., 246, 205 ± 209. Chem., 272, 27771 ± 27777. Shibata K, Tanaka S, Shiraishi T, Kitano S and Mori M. Yi F, Denker BM and Neer EJ. (1991). J. Biol. Chem., 266, (1999). Cancer Res., 59, 1096 ± 1101. 3900 ± 3906. Simonds WF. (1999). Trends Pharmacol. Sci., 20, 66 ± 73. Singer WD, Brown HA and Sternweis PC. (1997). Ann. Rev. Biochem., 66, 475 ± 509.

Oncogene