Ablation of Goα Overrides G1Restriction Point Control through Ras/ERK/Cyclin D1-CDK Activities*

We have generated stable IIC9 cell lines, Goa1 and Goa2, that overexpress full-length antisense Goα RNA. As shown previously, expression of antisense Goα RNA ablated the α subunit of the heterotrimeric G protein, Go, resulting in growth in the absence of mitogen. To better understand this change in IIC9 phenotype, we have characterized the signaling pathway and cell cycle events previously shown to be important in control of IIC9 G1/S phase progression. In this paper we clearly demonstrate that ablation of Goα results in growth, constitutively active Ras/ERK, elevated expression of cyclin D1, and constitutively active cyclin D1-CDK complexes, all in the absence of mitogen. Furthermore, these characteristics were abolished by the transient overexpression of the transducin heterotrimeric G protein α subunit strongly suggesting the transformation of Goα-ablated cells involves Goβγ subunits. This is the first study to implicate a heterotrimeric G protein in tumor suppression.

We have generated stable IIC9 cell lines, Goa1 and Goa2, that overexpress full-length antisense G o ␣ RNA. As shown previously, expression of antisense G o ␣ RNA ablated the ␣ subunit of the heterotrimeric G protein, G o , resulting in growth in the absence of mitogen. To better understand this change in IIC9 phenotype, we have characterized the signaling pathway and cell cycle events previously shown to be important in control of IIC9 G 1 /S phase progression. In this paper we clearly demonstrate that ablation of G o ␣ results in growth, constitutively active Ras/ERK, elevated expression of cyclin D1, and constitutively active cyclin D1-CDK complexes, all in the absence of mitogen. Furthermore, these characteristics were abolished by the transient overexpression of the transducin heterotrimeric G protein ␣ subunit strongly suggesting the transformation of G o ␣ablated cells involves G o ␤␥ subunits. This is the first study to implicate a heterotrimeric G protein in tumor suppression.
In IIC9 cells, a subclone of Chinese hamster embryo fibroblasts, platelet-derived growth factor (PDGF) 1 is a potent mitogen (1). PDGF stimulates an increase in cyclin D1 expression concomitant with an increase in cyclin D1-CDK activity (1). Cyclin D1 is an important G 1 protein in that its delayed early induction in response to mitogen is required for G 1 progression (2). Microinjection of cyclin D1 antibodies or antisense cyclin D1 plasmids into normal fibroblasts arrests them in G 1 but has no effect on cells already beyond the G 1 /S boundary (3,4). Thus, accumulation of cyclin D1 in G 1 in response to mitogen is required for progression through the restriction point and entrance into S phase (2,5,6). Cyclin D1 preferentially binds to its catalytic partner, cyclin-dependent kinase 4 (CDK4), to form a holoenzyme (7)(8)(9)(10). The activated cyclin D1-CDK4 complex preferentially binds to and phosphorylates the retinoblastoma gene product (pRb) (11)(12)(13)(14)(15)(16). Hence, the mitogen-induced activation of cyclin D1-CDK4 complexes allows for progression through the restriction point in vivo presumably through the hyperphorylation and inactivation of pRb in concert with other G 1 cyclin-CDK complexes.
We and others previously have shown that PDGF stimulates ERK1 activity in IIC9 and CCL39 cells and that inhibition of this activity is sufficient to cause the loss of PDGF-induced cyclin D1 expression, as well as a loss of cyclin D1-CDK activity (1). The loss of PDGF-induced cyclin D1-CDK activity was correlated with G 1 growth arrest. Attention has focused on the mitogen-induced signals involved in cell growth and more recently those signals regulating mitogen-dependent induction cyclin D1 (1,6). Expression of constitutively active mutant Ras has been shown to transform several cell types and elevate cyclin D1 expression (3,(17)(18)(19). The role of Ras proteins in the mitogenesis and transformation of cells is mediated, in part, by a downstream cascade of serine-threonine kinases that terminates with the activation of p42 and p44 MAPKs (ERK1, ERK2). Activated ERKs phosphorylate several nuclear factors that control gene expression. Evidence for the role of the Rasmediated MAPK cascade is well documented with kinase-deficient mutants of Raf-1, MEK and MAPKs inhibiting Ras transformation.
Recent evidence has shown that G o ␣ activates ERK (20). G o ␣ activation of ERK is mediated by a novel protein kinase C-dependent mitogenic signaling pathway which is independent of Ras activation. The studies reported here examine the effects of ablation of G o ␣ by overexpression of antisense G o ␣ RNA in IIC9 cells on the Ras/ERK pathway and cell cycle activities known to be important in mitogen-induced growth. Currently, there is evidence suggesting that ␤␥ subunits from pertussis toxinsensitive heterotrimeric G proteins are capable of ERK activation through a Ras-dependent signaling pathway (20 -23). The ␤␥-mediated ERK activation is blocked by the expression of dominant negative Ras (22). Our data demonstrate that loss of G o ␣ expression in IIC9 cells results in unregulated growth by constitutively activating the Ras/ERK pathway, a pathway we and others have shown positively regulates cyclin D1-CDK activity through the increased expression of cyclin D1.

EXPERIMENTAL PROCEDURES
Cells and Cell Culture-IIC9 cells, a subclone of Chinese hamster embryo fibroblasts (46), were grown and maintained in DMEM (Life Technologies, Inc., Grand Island, NY) containing 10% fetal calf serum and 2 mM L-glutamine (all chemicals were obtained from Sigma, unless specified otherwise). Subconfluent cultures were growth-arrested by washing once with serum-free DMEM and cultured for 48 -60 h in serum-free media. PDGF was obtained from Calbiochem (La Jolla, CA). Stable G o ␣ antisense transfectants were produced and maintained as described previously (48). Briefly, pcDNAI containing G o ␣ cDNA in an antisense orientation to the cytomegalovirus promoter was transfected into IIC9 cells using LipofectAmine protocol (Life Technologies, Inc.). Following an 18-h transfection period, cells were cultured for 48 h in DMEM containing 10% fetal calf serum to allow for the expression of neomycin-resistant gene products. Transfected IIC9 cells were grown for several weeks in selection media containing G418 (500 g/ml). G418-resistant clones were isolated and subcultured in DMEM containing 10% fetal calf serum and 250 g/ml G418. Several G o ␣-ablated clones were isolated of which two, Goa1 and Goa2 (described in the previous paper (48)), were used for the studies in this paper. The experiments in this paper utilized both clones even though the data presented is only that of the Goa1 clone.
Transient transfection of IIC9 cells using LipofectAmine resulted in 85-90% expression efficiency as visualized by ␤-galactosidase staining. IIC9 cells were also transiently transfected with G o ␣ antisense cDNA (3 g/ml) for 18 h. Transfected cells were stimulated with serum for 20 -24 h and cultured in serum-free media for 48 h for subsequent experiments. Similarly, Goa1 cells were transiently transfected with G t ␣ cDNA (3 g/ml) to achieve the phenotype of the Goa1/G t ␣ cell type. However, Goa1/G t ␣ cells were cultured in serum-free media for only 24 h to ensure the continued expression of G t ␣ protein in the absence of mitogen.
IIC9 and Goa1 cells were incubated with aphidicolin (5 g/ml) for 12-16 h to achieve growth arrest. Arrested cells were released by washing twice with DMEM and incubating in serum-free DMEM. Cells were then harvested and lysed 8 h after release from aphidicolin arrest as described above. Protein concentrations of lysates were determined by Bio-Rad Protein Assay (Bio-Rad) as recommended by the manufacturer. Lysates/proteins (15 g) were electrophoresed on 10% SDS-polyacrylamide gels. Separated proteins were then transferred to polyvinylidene difluoride membranes (Millipore, Boston, MA). Membranes were probed with a cyclin D1 polyclonal antibody (Santa Cruz Biotechnology). Goat anti-rabbit IgG (HϩL) horseradish peroxidase conjugate (Bio-Rad) was added as the secondary antibody and specific protein bands were visualized using ECL (Amersham) as recommended by the manufacturer.
Northern Blots-Total RNA was isolated from IIC9 and Goa1 cells (4 -8 ϫ 10 6 ) cultured on 100-mm dishes with Trizol Reagent (Life Technologies, Inc.) using the manufacturer's protocol. RNA was electrophoresed on 2% agarose/formaldehyde gels. Formaldehyde was removed by washing gels in 0.5 M ammonium acetate. RNA was transferred onto Hybond N ϩ nylon membranes (Amersham) using the Turboblotter system (Schleicher & Schuell, Keene, NH). RNA was cross-linked onto membranes with an Ultraviolet Crosslinker (Amersham) using the manufacturers protocol. Transferred RNA was visualized using a methylene blue/sodium acetate stain. Randomly-labeled [␣-32 P]dCTP cDNA probes (murine cDNA for cyclin D1 was a generous gift from Dr. Charles Sherr) were made using the Random Primed DNA Labeling Kit (Boehringer Mannheim, Germany). Blots were probed simultaneously with cyclin D1 and glyceraldehyde-3-phosphate dehydrogenase probes for 2 h at 65°C using Rapid-hyb buffer (Amersham) and washed once at room temperature with 5 ϫ SSPE (20 mM EDTA, 1 M NaCl, 50 mM NaH 2 PO 4 -H 2 O), 0.1% SDS and subjected to either autoradiography or direct quantitation with a PhosphorImager (Molecular Dynamics). More stringent washes were done at 65°C with 0.5% SDS when necessary.
Thymidine Incorporation-Thymidine incorporation was performed as described previously (1) with minor modifications. Briefly, growtharrested IIC9 cells were stimulated with PDGF (10 ng/ml) for approximately 20 h. Goa1 cells were serum-starved by washing twice with DMEM and incubated for 48 h in serum-free DMEM supplemented with 2 mM L-glutamine. To growth-arrest Goa1 cells, aphidicolin (5 g/ml) was added to serum-deprived cells as described above. Following a 24-h incubation, aphidicolin-arrested cells were released from arrest by washing twice with DMEM. Following primary incubation, cells were incubated for 3 h with 1 Ci of [ 3 H]thymidine/ml (NEN Life Sciences Products). 3 H-Labeled cells were washed twice with cold 1 ϫ PBS and the DNA was precipitated by incubating the cells for 30 min with cold 5% trichloroacetic acid. The trichloroacetic acid-precipitated DNA was washed twice with cold 5% trichloroacetic acid and solubilized with 2% sodium bicarbonate, 0.1 N NaOH. The solution was neutralized by addition of one-fifth volume of 5% trichloroacetic acid and the trichloroacetic acid-precipitated [ 3 H]DNA was quantitated by scintillation counting.
Cyclin D1/CDK Kinase Assay-Cyclin D1/CDK activity was assayed as described previously (8) with modifications. Briefly, growth-arrested IIC9 cells were stimulated with PDGF (10 ng/ml) and harvested at 0 and 24 h after stimulation by scraping in cold 1 ϫ PBS and lysed in 50 l of IP buffer (50 mM Hepes, 150 mM NaCl, 0.1 mM sodium vanadate, 1 mM EDTA, 2.5 mM EGTA, 1 mM dithiothreitol, 10 mM ␤-glycerophosphate, 1 mM sodium fluoride, 0.1% Tween 20, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, and 10 g/ml pepstatin). Goa1 cells were grown in serum, serumdeprived for 48 h, aphidicolin-arrested, or released from aphidicolin arrest for 8 h and harvested by scraping in cold 1 ϫ PBS. Lysates were sonicated briefly and insoluble material was pelleted by centrifugation at 10,000 ϫ g for 10 min. Cyclin D1 monoclonal antibody (2 g) was added to supernatants and incubated at 4°C. After 1-2 h cyclin D1 complexes were precipitated for 2-3 h with protein G-Sepharose. Cyclin D1 immune complexes were washed 4 times with 1 ml of cold IP buffer and 2 times with 1 ml of cold wash buffer (50 mM Hepes, 10 mM MgCl 2 , and 1 mM dithiothreitol). Cyclin D1 immune complexes were pelleted and resuspended in 30 l of reaction buffer (50 mM Hepes, 10 mM MgCl 2 , 1 mM dithiothreitol, 2.5 mM EGTA, 10 mM ␤-glycerophosphate, 0.1 mM sodium vanadate, and 20 M ATP). The immune complexes were incubated for 30 min at 30°C with 2 g/ml soluble GST-Rb fusion protein (GST-Rb cDNA was a generous gift from Dr. Mark Ewen) and 5 Ci of [␥-32 P]ATP. The reaction was stopped by addition of 15 l of 2 ϫ Laemmli sample buffer. Samples were boiled for 5 min and subjected to SDS-polyacrylamide electrophoresis. Gels were dried and subjected to autoradiography.
Analysis of GDP and GTP-bound Ras-The assay is a modification of the protocol by Downward et al. (47). Serum-deprived IIC9 cells or Goa1 cells were labeled for 18 h with [ 32 P]P i at 0.2 mCi/100 mM dish in phosphate-free DMEM. Cells were washed twice with phosphate-free media and with a saline buffer (50 mM Tris-HCl, pH 7.5, and 150 mM NaCl). Cells were incubated with PDGF (10 ng/ml) for 15 min and washed twice with cold 1 ϫ PBS. Cells were harvested and lysed by scraping in 500 l of IP buffer (50 mM Tris-HCl, pH 7.5, 20 mM MgCl 2 , 150 mM NaCl, 1% Triton X-100, 2 mM p-nitrophenyl phosphate, 10 g/ml pepstatin, 10 g/ml aprotinin, and 10 g/ml leupeptin). After a 10-min incubation the homogenate was centrifuged at 750 ϫ g for 5 min. The supernatant was treated with 100 l of bovine serum albumin-coated charcoal for 5 min at 4°C and then centrifuged at 750 ϫ g to remove charcoal. A monoclonal p21 ras antibody (Oncogene) was added to supernatants and Ras immune complexes were precipitated with protein G plus agarose (Oncogene) overnight at 4°C. The precipitate was washed twice with IP buffer and 3 times with wash buffer (Tris-HCl, pH 7.5, 20 mM MgCl 2 , and 150 mM NaCl). Final pellets were drained and bound nucleotides were eluted in 20 l of elution buffer (20 mM Tris-HCl, pH 7.5, 20 mM EDTA, 2% SDS, 0.5 mM GDP, and 0.5 mM GTP). Eluants were heated at 65°C for 5 min and centrifuged. The supernatant was spotted onto a polyethyleneimine-cellulose thin layer plate (Merck) and developed with 0.75 M KH 2 PO 4 (pH 3.4). GDP and GTP 32 P-labeled fractions were quantified by scintillation counting.

Goa1 Cells Do Not Exit the Cell Cycle upon Removal of
Mitogen-We previously generated a panel of stable IIC9 clones (including Goa1 and Goa2 clones) overexpressing fulllength G o ␣ RNA (48). We examined the phenotype of two of these clones, Goa1 and Goa2. In contrast to IIC9 cells, Goa1 and Goa2 cells do not express G o ␣ protein. Goa1 and Goa2 cells formed multiple foci in monolayer cultures and anchorage independent colonies in soft agar (48).
To understand the mechanism for the transformed phenotype of these cells, we examined whether Goa1 cells growth arrested upon removal of mitogen. Goa1 cells did not growth arrest upon removal of serum (Fig. 1A) and flow cytometry showed Goa1 cells randomly distributed throughout the cell cycle with a slight majority of cells in S phase (data not shown). To arrest and synchronize populations of mitogen-independent Goa1 cells, aphidicolin, a novel DNA polymerase ␣ inhibitor, was utilized. Addition of aphidicolin in the absence of serum resulted in growth arrest within 12-16 h after treatment (Fig.  1B). Previous studies have shown that removal of aphidicolin by washing with serum-free media is sufficient to allow cells to enter S phase in the absence of mitogen within 4 -8 h (5). However, aphidicolin-released cells arrest when they reach the subsequent G 1 /S restriction point of the next cycle. Goa1 cells resumed cell cycle progression after release from aphidicolin in serum-free media and continued through subsequent cycles in the absence of mitogen while IIC9 cells arrested at the proceeding G 1 /S restriction point (Fig. 1B) further demonstrating the mitogen-independent growth of Goa1 cells. Transient overexpression of G o ␣ antisense RNA also resulted in mitogen-independent growth suggesting that the transformed phenotype of the stable Goa1 cell type is a result of G o ␣ ablation and not a result of a deletion of another gene due to homologous recombinantion (data not shown).
Ablation of G o ␣ Results in the Mitogen-independent Activation of Ras-IIC9 cells overexpressing constitutively activated Ras form multiple foci when grown in soft agar and do not growth arrest when serum-depleted (data not shown). To examine whether the Ras pathway was involved in the transformation of Goa1 cells we first examined Ras activation. In growth-arrested IIC9 cells, levels of activated Ras as determined by the ratio GTP/(GTP ϩ GDP) associated with Ras were quite low (Fig. 2) and increased 6-fold within 5 min after the addition of PDGF (Fig. 2) or several other growth factors (data not shown). In contrast, Goa1 cells exhibited high levels of activated Ras in the absence of mitogen similar to levels found in IIC9 cells treated with PDGF (Fig. 2). Addition of PDGF did not increase the level of activated Ras above the levels seen in serum-depleted Goa1 cells. Similar results were observed with a second clonal G o ␣-ablated cell line (Goa2) (data not shown) indicating that ablation of G o ␣ in IIC9 cells results in significant activation of Ras. These results are consistent with the observed inability of these cells to growth arrest with the removal of mitogen and suggest that in certain cell types loss of G o ␣ could result in neoplastic transformation.
Goa1 Cells Exhibit Constitutively Active ERK-Recent data from several laboratories has suggested the importance of ERK activation in Ras-dependent growth. We have previously demonstrated that suppression of PDGF-induced ERK activation blocked G 1 progression in IIC9 cells (1). To determine whether ablation of G o ␣ resulted in constitutive activation of the ERK pathway, we next examined the endogenous activity of ERK in Goa1 cells. Addition of PDGF (Fig. 3) and several other growth factors (data not shown) to growth-arrested IIC9 cells increased ERK activity approximately 7-8-fold within 15 min. As previously found in CCL39 cells (24) and IIC9 cells (1) addition of thrombin or PDGF to growth-arrested cells induced a biphasic increase in ERK activity. A rapid 8 -12-fold increase of ERK activity within 5-10 min is followed by a sustained 4 -6-fold increase in ERK activity. Asynchronous Goa1 cells grown in serum-free media express levels of ERK activity (7-9-fold) ( Fig.  3) similar to the levels of ERK activity of IIC9 cells stimulated with PDGF (Fig. 3).
Elevated Expression of Cyclin D1 in the Absence of Mitogen-An ERK-responsive region has recently been identified in the cyclin D1 promoter (25) and we have previously shown that addition of PDGF to IIC9 cells induces cyclin D1 mRNA and protein expression (1). In addition, we and others have shown that inhibition of mitogen-induced ERK activity blocks expression of cyclin D1 and progression of IIC9 cells through G 1 (1,6). Many tumor cell lines express elevated levels of oncogenic Ras and cyclin D1 (26 -28). To identify a possible downstream target of the Goa1 constitutively active Ras/ERK pathway, we measured levels of cyclin D1. G o ␣ ablation conferred increased cyclin D1 protein expression by approximately 3-4-fold (Fig.  4A). The levels of cyclin D1 in Goa1 cells remained constitutively elevated in the absence of mitogen.
We next investigated the effect of aphidicolin arrest and release on cyclin D1 expression. Treatment of Goa1 cells with aphidicolin for 12 h resulted in sustained levels of cyclin D1 protein in the absence of mitogen (Fig. 4A). Matsushime et al. (5) previously showed that Bac1.2F5A macrophages released from aphidicolin arrest required the presence of growth factor (CSF-1) to sustain the expression of cyclin D1. Release of IIC9 cells from aphidicolin arrest in the absence of PDGF resulted in the rapid (within 5 h) decrease in the levels of cyclin D1 protein (Fig. 4A). However, cyclin D1 protein levels in Goa1 cells did not decrease significantly when released from aphidicolin in the absence of PDGF suggesting a significant difference in the requirement of sustained presence of growth factor for cyclin D1 expression. It is clear that the constitutive activation of the Ras/ERK pathway (Figs. 2 and 3) provides the sustained mitogenic signals required for the continued expression of cyclin D1. Cyclin D1 protein expression remained high through the next round of replication (approximately 24 h after aphidicolin release) (Fig. 4A). In the absence of mitogen, cyclin D1 mRNA levels in Goa1 cells were similar to the levels found in IIC9 cells treated with PDGF (Fig. 4B). Aphidicolin-treated and released

FIG. 4. Goa1 cells display mitogen-independent expression of cyclin D1.
A, IIC9 and Goa1 cells were grown and serum-deprived for 48 h as described previously. IIC9 cells were then incubated for 24 h with no serum (SS) or with 10 ng/ml PDGF. Goa1 cells were incubated with no serum (SS) or 10% serum (Serum) for 24 h. Conversely, serumdeprived IIC9 and Goa1 cells were supplemented with 5 g/ml aphidicolin (APH) for 12 h. Aphidicolin was removed by washing with fresh media and both IIC9 and Goa1 cells were incubated in serum-free media for an additional 5 h (aphidicolin release; AR). IIC9 and Goa1 cells were lysed as described previously and all lysates were immunoblotted with a polyclonal mouse cyclin D1 antibody as described under "Experimental Procedures." B, serum-deprived IIC9 cells were incubated in the absence (SS) or presence of 10 ng/ml PDGF (PDGF) for 24 h. Serum-deprived Goa1 cells were unsupplemented (SS) or supplemented with 5 g/ml aphidicolin for 12 h or aphidicolin was removed by washing with fresh media and cells incubated in serum-free media for an additional 5 h (AR). IIC9 and Goa1 cells were lysed and RNA extracted using Trizol as described under "Experimental Procedures." RNA was electrophoresed on 2% agarose/formaldehyde gels and transferred onto Hybond N ϩ membranes. Membranes were probed simultaneously with [ 32 P]CTP-labeled murine cyclin D1 and glyceraldehyde-3phosphate dehydrogenase (GAPDH) cDNA as described under "Experimental Procedures." C, serum-deprived IIC9 cells were incubated in the absence (SS) or presence of 10 ng/ml PDGF (PDGF) for 24 h. Serum-deprived Goa1 cells were unsupplemented (SS) or supplemented with 5 g/ml aphidicolin (APH) for 12 h or aphidicolin was removed by washing with fresh media and cells incubated in serum-free media for an additional 5 h (AR). IIC9 and Goa1 cells were lysed and lysates incubated with a cyclin D1 monoclonal Ab for 1-2 h at 4°C. Cyclin D1 immune complexes were precipitated with protein G-Sepharose and immunoprecipitated complexes assayed for their ability to phosphorylate soluble GST-Rb fusion protein in vitro as described under "Experimental Procedures." Goa1 cells exhibited a 1.6-fold decrease in cyclin D1 mRNA expression although these levels were still 3-fold higher than serum-deprived IIC9 cells (Fig. 4B) providing further evidence for the positive role of the Ras/ERK pathway in sustaining cyclin D1 expression in Goa1 cells in the absence of mitogen.
Cyclin D1-CDK Complexes Are Constitutively Active-Phosphorylation of Rb by active cyclin D1-CDK complexes has been demonstrated to be required for progression through G 1 in several cell types (12)(13)(14)(15)(16). Although several transformed tumor cells express abnormally high levels of cyclin D1, the role of cyclin D1 in tumor formation is unclear. However, cyclin D1-CDK activity is thought to play an important role in mitogeninduced progression of cells through G 1 . In the absence of mitogen, IIC9 cells contain low levels of cyclin D1-CDK activity (Fig. 4C). PDGF treatment induced a 6-fold increase in cyclin D1-CDK activity within 4 h (data not shown) and sustained this level of activity through 24 h (Fig. 4C). In contrast to IIC9 cells, Goa1 cells display significant cyclin D1-CDK activity in the absence of mitogen (Fig. 4C). In addition, treatment of Goa1 cells with PDGF did not increase further cyclin D1-CDK activ- or presence (solid bars) of PDGF (10 ng/ml) for 24 h, or serum-deprived IIC9, Goa1, and Goa1/G t ␣ cells were supplemented with aphidicolin (5 g/ml) for 12 h. Gray bars, in those cells treated with aphidicolin, aphidicolin was removed by washing with fresh media and the cells incubated in serum-free media for an additional 5 h. Cells were lysed and RNA extracted using Trizol as described previously. RNA was electrophoresed on 2% agarose/formaldehyde gels and transferred onto Hybond N ϩ membranes. Membranes were probed simultaneously with [ 32 P]CTP-labeled murine cyclin D1 and glyceraldehyde-3-phosphate dehydrogenase cDNA as described under "Experimental Procedures." Results are reported as the mean Ϯ S.D. (n ϭ 3) of cyclin D1 normalized to glyceraldehyde-3-phosphate dehydrogenase as quantitated by a PhosphorImager . D, serum-deprived IIC9, Goa1, and Goa1/G t ␣ cells were incubated in the absence (SS) or presence of PDGF (10 ng/ml) for 24 h and assayed for cyclin D1-CDK activity as described under "Experimental Procedures." Gels were dried and scanned with a PhosphorImager and viewed using Lview Pro 1.D © Software (MMedia Research Corp.). ity (data not shown). Treatment of Goa1 cells with aphidicolin as well as release from aphidicolin arrest in serum-free media did not result in a decrease in cyclin D1-CDK activity (Fig. 4C) further correlating the mitogen independence of Goa1 cells with the activation of the cyclin D1-CDK complexes in the absence of mitogen. Preliminary results from our laboratory also suggest that high levels of cyclin D1-CDK activity are required to retain the transformed phenotype of Goa1 cells. These data further suggest that in certain cell types Ras transformation involves the up-regulation of cyclin D1-CDK activity.
A Possible Role for G o ␤␥ Subunits in Goa1 Transformation-The ablation of G o ␣ from IIC9 cells provides interesting insights into the mechanisms of G 1 progression as well as cellular transformation. Ablation of G o ␣ from IIC9 cells results in growth and high levels of Ras activity in the absence of mitogen (Fig. 2). In addition, Goa1 cells contain constitutive levels of ERK activity (Fig. 3) as well as cyclin D1-CDK activity (Fig.  4C). We and others have shown that ERK activity is essential for mitogen-induced expression of cyclin D1 and subsequent cyclin D1-CDK activity. Inhibition of ERK activity not only resulted in a concomitant loss of cyclin D1-CDK activity but also resulted in growth arrest (1). These data show that Ras transformation through the ablation of G o ␣ results in sustained cyclin D1-CDK activity and suggest that in certain cell types Ras transformation may involve up-regulation of cyclin D1-CDK activity.
Recent evidence demonstrating that the G o ␣ subunit stimulates ERK activity through a Ras-independent mechanism suggested a role for G o ␣ in ERK activation (20). Ablation of G o ␣ from IIC9 cells resulted in constitutive elevation of Ras and ERK activation in the absence of mitogenic stimulation. Several studies, however, have provided new insights into G protein-coupled signaling (29). Overexpression of ␤␥ subunits also activates the Ras/ERK pathway (21,22). In addition, several studies present findings suggesting that G␤␥ activates the Ras/ERK pathway through the recruitment of cytosolic or membrane-associated factors containing pleckstrin homology domains or through direct activation of G␤␥ with Ras (29). Transfection of a bARKct G␤␥ antagonist which contains a ␤␥-binding region significantly reduces Ras activation (23). Furthermore, transfection of dominant negative Ras eliminates ERK activation suggesting that G␤␥-mediated ERK activation involves Ras (22).
To determine whether the transformation of Goa1 cells was mediated by G␤␥ activation of the Ras/ERK pathway in the absence of G o ␣, we transiently transfected Goa1 cells with G t ␣ (Goa1/G t ␣) which has previously been used to sequester G␤␥ subunits (21,22). Transient overexpression of G t ␣ did not result in the expression of G o ␣ protein, demonstrating continued ablation of G o ␣ in the presence of G t ␣ (data not shown). As observed in IIC9 cells, Goa1/G t ␣ cells growth arrested when cultured in the absence of mitogen (Fig. 5A). Stimulation with PDGF resulted in progression through S phase and this progression was inhibited by addition of aphidicolin as described previously (Fig. 1). In contrast to Goa1 cells, stimulation of Goa1/G t ␣ cells with PDGF resulted in a significant increase in [ 3 H]thymidine incorporation as compared with mitogen-deprived cells (Fig. 5A). Goa1/G t ␣ cells also required the presence of PDGF after release from aphidicolin arrest for entrance into S phase 24 h after release (data not shown).
Goa1/G t ␣ cells displayed a 60 -70% decrease in ERK activity compared with unstimulated Goa1 cells (Fig. 5B). Addition of PDGF to Goa1/G t ␣ cells resulted in a marked stimulation of ERK activity demonstrating that these cells were able to respond to PDGF (Fig. 5B). These data demonstrate that transfection of G t ␣ significantly reduces ERK activity in unstimu-lated Goa1/G t ␣ cells but does not inhibit the ability of PDGF to stimulate ERK activity in Goa1/G t ␣ cells. Cyclin D1 expression in IIC9 cells as well as macrophages requires mitogen after release from aphidicolin arrest (5). Absence of mitogen results in a rapid decrease in cyclin D1 (Fig. 4A). However, withdrawal of PDGF from Goa1 cells did not decrease cyclin D1 levels (Fig.  4, A and B). This regulation is restored in Goa1/G t ␣ cells. Release of Goa1/G t ␣ cells from aphidicolin arrest results in a 2.5-fold decrease of cyclin D1 mRNA expression (Fig. 5C). Concomittant with a decrease in cyclin D1 protein and mRNA expression, cyclin D1-CDK activity was significantly reduced in Goa1/G t ␣ cells in the absence of PDGF (Fig. 5D). It is important to note that the decrease in cyclin D1-CDK activity is due to cyclin D1 down-regulation since the levels of inhibitors of this complex (p27 KIP1 and p16 INK4 ) are relatively high in cycling Goa1 cells. 2 These data suggest that sequestration of G o ␤␥ by G t ␣ results in a reversal of the Goa1 phenotype rendering the cells sensitive to the machinery involved in the regulation of cell cycle progression. This results in the restoration of mitogen-induced regulation of cyclin D1 expression and its role in G 1 progression. Previous results from our laboratory (1) and others (6) show that mitogen-induced ERK activation is required for growth and the positive regulation of cyclin D1 expression. Preliminary results from our laboratory now suggest that ERK activation absolutely is required to maintain the transformed phenotype of Goa1 cells independent of G o ␤␥-mediated signaling. 2 This is the first study to implicate a heterotrimeric G protein, G o ␣, in tumor suppression. We have shown that ablation of G o ␣ from IIC9 cells results in cellular transformation (48) and that this involves the constitutive activation of the Ras/ERK pathway and the mitogen-independent activation of cyclin D1-CDK complexes. Transient ablation of G q , G i2 , and G i3 in IIC9 cells did not result in an increase in basal ERK activity in the absence of mitogenic stimulation suggesting that only the specific ␤␥ associated with G o ␣ activate the Ras/ERK pathway. 3 The possibility that G o ␣ is a tumor suppressor becomes more intriguing by its genomic location. The G o ␣ gene (GNAO1) is located on chromosome 16 at position 16q13 and is the only heterotrimeric G protein gene on chromosome 16 (30 -33). The instability of chromosome 16 has been well documented and has led to the discovery of several break points in its lower (q) arm (34 -38). These chromosomal abnormalities (translocations and deletions) are now thought to be associated with a wide array of human cancers including breast, prostate, colorectal, pituitary, alveolar, Ewing sarcomas, Wilms' tumors, and turban tumors all of which map near or at the 16q13 region (34, 36 -38, 39 -45). Our data is the first to implicate G o ␣ in tumor suppression and warrants further investigation into the apparent tumor suppressive properties of G o ␣.