Ras-stimulated Extracellular Signal-related Kinase 1 and RhoA Activities Coordinate Platelet-derived Growth Factor-induced G1 Progression through the Independent Regulation of Cyclin D1 and p27KIP1 *

Platelet-derived growth factor (PDGF)-induced Ras activation is required for G1 progression in Chinese hamster embryo fibroblasts (IIC9 cells). Ras stimulates both extracellular signal-related kinase (ERK) activation and RhoA activation in response to PDGF stimulation. Inhibition of either of these Ras-stimulated pathways results in growth arrest. We have shown previously that Ras-stimulated ERK activation is essential for the induction and continued G1 expression of cyclin D1. In this study we examine the role of Ras-induced RhoA activity in G1 progression. Unstimulated IIC9 cells expressed high levels of the G1 cyclin-dependent kinase inhibitor p27KIP1. Stimulation with PDGF resulted in a dramatic decrease in p27KIP1 protein expression. This decrease was attributed to increased p27KIP1 protein degradation. Overexpression of dominant-negative forms of Ras or RhoA completely blocked PDGF-induced p27KIP1 degradation, but only dominant-negative Ras inhibited cyclin D1 protein expression. C3 transferase also inhibited PDGF-induced p27KIP1degradation, thus further implicating RhoA in p27KIP1regulation. Overexpression of dominant-negative ERK resulted in inhibition of PDGF-induced cyclin D1 expression but had no effect on PDGF-induced p27KIP1 degradation. These data suggest that Ras coordinates the independent regulation of cyclin D1 and p27KIP1 expression by the respective activation of ERK and RhoA and that these pathways converge to determine the activation state of complexes of cyclin D1 and cyclin-dependent kinase in response to mitogen.

Progression through the G 1 phase of the mammalian cell cycle is mediated in part through the early induction of D-type cyclins by mitogenic stimulation (1)(2)(3). Cell cycle progression is orchestrated by distinct families of cyclin-dependent kinases (CDKs) 1 whose activities depend upon cyclin binding, positive and negative phosphorylation, and association with inhibitory polypeptides (10). Progression through the G 1 phase of the cell cycle is controlled by one of three D-type cyclins (D1, D2, or D3), which assemble with their catalytic partner CDK4 or CDK6, and cyclin E, which assembles with its catalytic partner CDK2 (1)(2)(3)(4)(5)(6)(7)(8)(9). D-and E-type CDKs are required for G 1 progression, and both contribute to the phosphorylation and inactivation of the retinoblastoma (Rb) protein thus canceling its growth-inhibitory properties (1, 2, 5, 7, 10 -17). The activation of CDK4/ CDK6 following association with cyclin D is critical for G 1 progression. Inhibition of cyclin D1 expression through antisense cDNA or microinjection of antibodies specific to cyclin D results in G 1 growth arrest (18,19). D-type cyclins have been referred to as G 1 mitogenic sensors because their induction requires mitogen, and removal of mitogen in G 1 results in their rapid degradation and subsequent growth arrest (1)(2)(3).
The Ras/MAP kinase (ERK) pathway has been implicated in transducing mitogenic signals from growth factor receptors to the cell cycle machinery. Inhibition of the Ras/ERK pathway blocks mitogen-induced expression of cyclin D1 in Chinese hamster fibroblasts, demonstrating the importance of this pathway in mediating the mitogenic signals responsible for cyclin D1 induction (20 -22). We have shown recently that PDGF induces the sustained activation of ERK and that this sustained activation is required for the continued accumulation of cyclin D1, implicating ERK activation in the regulation of cyclin D1 expression (21).
Concomitant with increased G 1 cyclin D expression, cyclin D⅐CDK-associated activity increases in G 1 (1-9, 20 -22). The increase in cyclin D⅐CDK activity is a result of both an increase in cyclin D and a decrease in G 1 cyclin-dependent kinase inhibitor expression (1,2,7). Although several cyclin-dependent kinase inhibitors have been identified as potent inhibitors of cyclin⅐CDK complexes, p27 KIP1 is the only cyclin-dependent kinase inhibitor whose protein expression decreases as mitogen-induced cells enter the cell cycle (7,(23)(24)(25). The decrease in p27 KIP1 expression occurs through protein degradation via the ubiquitin-proteasome pathway (26). The retention of inhibitory levels of p27 KIP1 appears to be involved in the growth-inhibitory properties of transforming growth factor-␤, rapamycin, and cyclic AMP (27)(28)(29). In contrast, overexpression of p27 KIP1 antisense cDNA results in mitogen-independent G 1 progression, demonstrating the importance of p27 KIP1 in maintaining cell quiescence (30,31). The mitogenic signals responsible for p27 KIP1 degradation have not been defined clearly.
PDGF stimulation causes the rapid activation of Ras and the subsequent downstream activation of ERK (21,22). In addition, Ras also stimulates the downstream activation of RhoA presumably to induce changes in cytoskeleton structure associated with growth (32)(33)(34)(35)(36). However, RhoA activation has not been linked directly to the regulation of the cell cycle. In this study we demonstrate that Ras coordinates G 1 progression through two independent pathways: ERK regulation of cyclin D1 expression and RhoA regulation of p27 KIP1 degradation to ensure the proper activation state of cyclin D1⅐CDK complexes following mitogenic stimulation.

EXPERIMENTAL PROCEDURES
Cell Culture and Reagents-IIC9 cells, a subclone of Chinese hamster embryo fibroblasts (37), were grown and maintained in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) containing 10% fetal calf serum and 2 mM L-glutamine (Sigma). Subconfluent (60 -70%) were growth arrested by washing once with fresh Dulbecco's modified Eagle's medium and reculturing in serum-free Dulbecco's modified Eagle's medium for 48 h. Human recombinant PDGF-BB (Calbiochem) was added to cultures at 10 ng/ml in all experiments. Growth-arrested IIC9 cells were preincubated with 10 M PD98059 (New England Biolabs, Beverly, MA) before the addition of PDGF. Dominant-negative ERK2 (dnERK Ϫ ) was a generous gift from Dr. Jacques Pouyssegur (University of Nice, France). Dominant-negative Ras (dnRas Ϫ ) and RhoA (dnRhoA 19 ) and constitutively active RhoA (RhoA 63 ) were constructed as described previously through site-directed mutagenesis of Thr to Asn at codon 17 and 19 or Gln to Leu at codon 63, respectively, with the Transformer TM site-directed mutagenesis kit (CLONTECH) (38). Transient transfection of IIC9 cells (50 -60% confluence) using Lipo-fectAMINE (Life Technologies, Inc.) as recommended by the manufacturer consistently resulted in Ͼ90% expression efficiency as visualized by ␤-galactosidase staining.
Thymidine Incorporation-[ 3 H]Thymidine incorporation into IIC9 cells was measured as described previously (21,22). Briefly, growtharrested IIC9 cells were stimulated with PDGF (10 ng/ml) for 20 h. Approximately 17 h after the addition of PDGF, 1 Ci of [ 3 H]thymidine (NEN Life Science Products) was added, and the cells were incubated for an additional 3 h. Cells were washed twice with cold 1 ϫ PBS and incubated for an additional 30 min with 5% trichloroacetic acid. Trichloroacetic acid-precipitated DNA was washed with cold 5% trichloroacetic acid and solubilized with 2% sodium bicarbonate and 0.1 N NaOH. After neutralization with 5% trichloroacetic acid, precipitated [ 3 H]DNA was quantitated by scintillation counting.

RESULTS
PDGF Induces the Loss of p27 KIP1 Protein-Protein levels of p27 KIP1 are increased in contact-inhibited or serum-deprived cells and decrease when cells are stimulated by mitogen to enter the cell cycle (7,(23)(24)(25). Various mitogens including epidermal growth factor, PDGF, and serum are capable of stimulating cell cycle entry and p27 KIP1 degradation (23)(24)(25). However, the mechanism by which these mitogens stimulate p27 KIP1 degradation remains unclear.
We have shown previously that PDGF is a potent mitogen for IIC9 cells, and addition of PDGF to quiescent IIC9 cells resulted in up-regulation of cyclin D1 protein expression and D-type cyclin-dependent kinase activity (21,22). Stimulation with PDGF also resulted in the time-dependent degradation of p27 KIP1 protein ( Fig. 1). 2 h after PDGF stimulation, p27 KIP1 protein levels decreased approximately 50%, and by 24 h they were nearly undetectable ( Fig. 1). Levels of CDK4, which we have shown previously do not increase with PDGF stimulation (21), were measured to ensure equal protein loading ( Fig. 1). Previous studies have shown that loss of p27 KIP1 protein occurs via a ubiquitin-mediated degradation pathway (26). In agreement with these observations, incubation of IIC9 cells with a calpain I inhibitor resulted in the appearance of polyubiquitinated forms of p27 KIP1 (data not shown).
Ras but Not ERK Activity Is Required for PDGF-induced Loss of p27 KIP1 -We have shown previously that PDGF-induced G 1 progression requires the sustained activation of ERK in a MAP kinase/ERK kinase 1 (MEK1)-dependent manner (21). The sustained activation of ERK following PDGF stimulation was responsible for the continued accumulation of cyclin D1, and inhibition of this activity resulted in the loss of cyclin D1 protein expression (21). To determine whether PDGF-induced ERK activation also contributed to the degradation of p27 KIP1 , we overexpressed a dnERK Ϫ in IIC9 cells. Although dnERK Ϫ inhibits PDGF-induced G 1 progression (21), it did not inhibit the PDGF-induced loss of p27 KIP1 (Fig. 2A). IIC9 cells preincubated with an inhibitor of MEK1 activation, PD98059, displayed normal PDGF-induced p27 KIP1 protein degradation with p27 KIP1 protein levels being reduced to 10% maximal levels by 16 h (Fig. 2B). These data suggest that downstream effectors of MEK1 and ERK are not responsible for the degradation of p27 KIP1 . Lysates/proteins (10 g) were electrophoresed on 12% SDS-polyacrylamide gels and immunoblotted with a polyclonal p27 KIP1 or CDK4 antibody.
We next looked at Ras, an upstream activator of the MAP kinase pathway, which we have shown previously is activated rapidly by PDGF (22). The addition of PDGF to IIC9 cells overexpressing dnRas Ϫ did not affect p27 KIP1 protein levels, demonstrating the requirement of Ras activation for PDGFinduced p27 KIP1 degradation. These data demonstrate clearly that mitogen-regulated destruction of p27 KIP1 is downstream of Ras.
RhoA Regulates p27 KIP1 Degradation-It has become apparent that both MAP kinase and Rho pathways are important in the control of cell proliferation (20 -22, 32-36, 39, 40). Whereas the role of the MAP kinase cascade has been shown clearly to regulate cyclin D1 expression (20 -22, 40, 42), the role of the Rho cascade in cell cycle progression is unknown. To investigate the importance of PDGF-induced RhoA activity, we transfected IIC9 cells with dnRhoA 19 and examined the effect of dnRhoA 19 expression on several proteins that control progression through G 1 . Overexpression of dnRhoA 19 inhibited PDGFinduced reduction of p27 KIP1 protein levels in IIC9 cells (Fig.  3A) similar to that seen in dnRas Ϫ -transfected cells ( Fig. 2A), suggesting that RhoA is a downstream Ras-dependent signaling molecule required for PDGF-induced p27 KIP1 degradation. Incubation with C3 transferase, an inhibitor of RhoA activity, also resulted in the inhibition of PDGF-induced p27 KIP1 degradation, further implicating RhoA activation in p27 KIP1 destruction (Fig. 4A). Overexpression of a constitutively active RhoA mutant, RhoA 63 , resulted in the mitogen-independent decrease in p27 KIP1 protein expression (Fig. 3B) identical to that of PDGF-stimulated IIC9 cells. These data demonstrate that activated RhoA alone is sufficient for loss of p27 KIP1 . The requirement of RhoA for PDGF-induced p27 KIP1 degradation and the ability of RhoA 63 mutant to stimulate p27 KIP1 degradation independently show clearly that RhoA activation has an important role in G 1 progression and provide further evidence of the separate and distinct properties of the Ras/ERK and Ras/ RhoA pathways in cell cycle regulation.
Ras/ERK but Not Ras/Rho Pathway Regulates Cyclin D1 Expression-We and others have demonstrated previously the importance of mitogen-stimulated Ras/ERK activation on cyclin D1 induction (20 -22, 40, 42). The regulation of cyclin D1 induction and its continued G 1 expression may be attributed to the ability of mitogens to stimulate the sustained activation of ERK (20,21,39). Overexpression of a dnRas mutant resulted in the inhibition of PDGF-stimulated cyclin D1 induction (Fig.  5A). In agreement with previous reports, overexpression of a dnERK mutant resulted in a similar inhibition in PDGF-stimulated cyclin D1 induction (Fig. 5B). However, it has not yet been determined whether other Ras-stimulated pathways are important for cyclin D1 induction. Overexpression of dnRhoA 19 , which resulted in an inhibition of PDGF-induced p27 KIP1 degradation, did not affect PDGF-stimulated ERK activity (Fig. 6). We hypothesized that the separation of these pathways would allow for their independent regulation of different G 1 gene products: RhoA for p27 KIP1 degradation and ERK for cyclin D1 induction. In agreement with this hypothesis, overexpression of dnRhoA 19 did not affect the induction and accumulation of cyclin D1 protein following PDGF stimulation (Fig. 5C), suggesting RhoA is not required for cyclin D1 protein expression.
Constitutively Active Ras Requires ERK1 or RhoA Activity for the Downstream Regulation of Cyclin D1 or p27 KIP1 -Ras has many downstream effectors of which two, ERK1 and RhoA, reside in separate and distinct growth-promoting pathways.

FIG. 4. C3 transferase inhibits PDGF-induced p27 KIP1 degradation and cyclin D1⅐CDK activity.
Growth-arrested IIC9 cells were preincubated for 2 h with C3 transferase (40 g/ml) and harvested 0 and 24 h after the addition of PDGF (10 ng/ml) by scraping in cold 1 ϫ PBS and lysed. Panel A, lysates/proteins (15 g) were electrophoresed on 12% SDS-polyacrylamide gels and immunoblotted with a polyclonal p27 KIP1 or CDK4 antibody. Panel B, conversely, lysates (100 g) were incubated for 1-2 h at 4°C with a monoclonal cyclin D1 antibody. Cyclin D1 immune complexes were precipitated with protein G-Sepharose and assayed for their ability to phosphorylate soluble GST-Rb fusion protein in vitro as described under "Experimental Procedures." We and others have provided evidence previously which demonstrates the requirement of ERK1 for cyclin D1 up-regulation and active cyclin D1⅐CDK complexes following mitogenic stimulation (20,21). We have also provided data in this study which strongly implicate RhoA activation in the regulation of p27 KIP1 degradation. Constitutively active Ras mutants result in cellular transformation (35,36), and in IIC9 cells a constitutively active Ras mutant (Ras 12 ) resulted in ERK1 and RhoA activity independent of mitogen (data not shown) in agreement with several previous studies. We hypothesized that mitogen-independent regulation of cyclin D1 and p27 KIP1 by Ras 12 required ERK and RhoA activity, respectively. Ras 12 stimulated cyclin D1 up-regulation as well as p27 KIP1 degradation in the absence of mitogen (Fig. 7, A and B). In agreement with this hypothesis, IIC9 cells overexpressing Ras 12 (IIC9-Ras 12 ) required ERK1 activation by MEK1 to increase cyclin D1 expression in the absence of mitogen. IIC9-Ras 12 cells incubated with PD98059 displayed reduced (6 -8-fold) cyclin D1 protein expression levels (Fig. 7A), indicating a downstream requirement of ERK1 activity. Similarly, IIC9-Ras 12 cells transfected with dnRhoA 19 failed to induce the loss of p27 KIP1 protein (Fig. 7B), demonstrating further the requirement of Ras-stimulated RhoA activity in p27 KIP1 degradation. These data also provide evidence for the necessity of ERK and RhoA activities in the regulation of critical G 1 events and suggest that other Ras-stimulated pathways are unable to compensate for the loss of either activity to regulate cyclin D1 and p27 KIP1 protein expression.
PDGF-induced RhoA Activation Is Required for Active Cyclin D1⅐CDK Complexes and Subsequent G 1 Progression-Active cyclin D1⅐CDK complexes in concert with other G 1 cyclin⅐CDKs are responsible for progression into S phase in part through their ability to phosphorylate and inactivate the Rb protein (1, 2, 10 -17). Stimulation of growth-arrested IIC9 cells resulted in a 6 -7-fold increase in cyclin D1⅐CDK activity (Figs. 4B and 8A). Although overexpression of dnRhoA 19 did not affect cyclin D1 levels (Fig. 5C), overexpression of dnRhoA 19 resulted in the complete inhibition of PDGF-induced cyclin D1⅐CDK activity (Fig. 8A). Incubation with C3 transferase, a specific inhibitor of RhoA activity, also resulted in the complete inhibition of PDGF-induced cyclin D1⅐CDK activity (Fig. 4B), further implicating RhoA in the downstream determination of the cyclin D1⅐CDK activation state. Concomitant with its ability to inhibit cyclin D1⅐CDK activity, dnRhoA 19 inhibited PDGF-stimulated G 1 progression (Fig. 8B), demonstrating further the importance of RhoA in mediating events important in G 1 progression. DISCUSSION Ras/ERK are critical mediators of mitogen-dependent cyclin D1 expression (20 -22, 40, 42). Inhibition of mitogen-induced MEK1 or ERK1 activation resulted in the inhibition of cyclin D1 induction (20,21). Furthermore, the sustained activation of MEK1/ERK1 was required for the continued presence of cyclin D1, demonstrating the importance of ERK1 in regulating cyclin D1 expression positively (21). However, the ERK pathway does not appear to control p27 KIP1 degradation. Expression of constitutively active ERK does not result in p27 KIP1 degradation in the absence of mitogen, implicating an independent mitogenic pathway in the regulated destruction of p27 KIP1 (31). We have focused our study on the Ras/RhoA mitogenic pathway in regulating p27 KIP1 degradation. PDGF or serum stimulates a rapid induction of Ras activity followed by the sustained activation of ERK1 (20 -22). We show for the first time that RhoA regulates mitogen-induced p27 KIP1 degradation. Overexpression of dnRas Ϫ inhibited the PDGF-induced loss of p27 KIP1 ( Fig. 2A). In contrast, inhibition of ERK activity or MEK1 activity did not affect the PDGF-induced degradation of p27 KIP1 (Fig. 2, A and B).
Recent data indicate that Ras activates two independent pathways that are important for Ras-transforming ability (35,36,38). Whereas overexpression of constitutively active Ras in NIH 3T3 cells resulted in transformation, expression of constitutively active ERK or RhoA alone was ineffective (38). However, expression of both constitutively active ERK and RhoA was as effective as overexpression of constitutively active Ras. Our findings are consistent with these results and demonstrate for the first time that RhoA signaling regulates p27 KIP1 , a protein important in the regulation of G 1 progression. Expression of dnRhoA 19 inhibited the PDGF-induced degradation of p27 KIP1 (Fig. 3A), and RhoA 63 stimulated a loss of p27 KIP1 in the absence of mitogen, demonstrating the requirement of RhoA activity for Ras-dependent p27 KIP1 degradation. C3 transferase also inhibited PDGF-induced p27 KIP1 degradation (Fig. 4A), further implicating RhoA activation in the stimulation of this process. In contrast, cyclin D1 induction was not affected by dnRhoA 19 expression after PDGF stimulation (Fig.  5C), suggesting separate pathways for Ras-dependent cyclin D1 and p27 KIP1 regulation. PDGF-stimulated induction of cyclin D1 protein was not sufficient for progression through G 1 because C3 transferase or dnRhoA 19 blocked PDGF-induced cyclin D1⅐CDK activity and subsequent G 1 progression (Figs. 4B and 8A). These data demonstrate the coordinated signaling between ERK and RhoA required for G 1 progression. We cannot rule out the possibility that overexpression of dnRhoA 19 may affect or inhibit other G 1 events necessary for cell growth. However, it is clear that RhoA activity is required for PDGFinduced cyclin D1⅐CDK activity and that the phosphorylation of Rb by these activated complexes is an integral component of G 1 progression.
Anti-mitogens (transforming growth factor-␤) promote growth arrest through their ability to maintain high levels of p27 KIP1 (27)(28)(29). High levels of p27 KIP1 stoichiometrically inhibit cyclin D⅐CDK complexes (7,29), and a loss of p27 KIP1 through mitogenic stimulation or antisense cDNA expression promotes G 1 /S transit (7, 23-25, 30, 31). We were able to disrupt the normal PDGF-induced degradation of p27 KIP1 in cycling cells by overexpressing dnRhoA 19 (Fig. 3A) or inhibiting RhoA activity with C3 transferase (Fig. 4A). The imbalance of p27 KIP1 protein levels was sufficient to arrest PDGF-stimu- lated cells (Fig. 8B). Growth arrest was not caused by downregulation of cyclin D1 protein expression because IIC9 cells overexpressing dnRhoA 19 expressed wild type levels of cyclin D1 after PDGF stimulation (Fig. 5C). This result is consistent with Kato et al. (29) who showed that macrophages treated with 8-bromo-cAMP, dibutyryl cAMP, prostaglandin E 2 plus isobutylmethylxanthine, or rapamycin displayed high levels of p27 KIP1 and normal induction of cyclin D1 protein following colony-stimulating factor 1 stimulation. Together, these data suggest that mitogen-induced sustained activation of ERK is sufficient to induce cyclin D1 protein expression and that p27 KIP1 protein levels do not affect this up-regulation negatively as suggested previously (31).
Our data provide further evidence that p27 KIP1 acts downstream of cyclin D1 induction. Kato et al. (29) demonstrated the ability of p27 KIP1 to inhibit the activation of cyclin D⅐CDK4 complexes. Overexpression of dnRhoA 19 inhibited PDGF-induced cyclin D1-CDK activity (Fig. 8A) likely because of the maintenance of high levels of p27 KIP1 and thus inhibited G 1 /S transit (Fig. 8B). These results clearly demonstrate the impor-tance of the RhoA signaling for G 1 progression.
The ERK cascade is largely responsible for the immediateearly induction and/or activation of mitogen-induced transcription factors including, but not limited to, c-myc, Elk-1, c-fos, and c-jun (41). Recent studies have focused on the role of these transcription factors in promoting cyclin D1 transcription (20,40,42). We have shown previously the requirement of sustained ERK activation for continued cyclin D1 mRNA and protein expression following PDGF stimulation, demonstrating the importance of continued growth factor signaling in G 1 progression (21). It is not clear whether the mitogen-induced p27 KIP1 degradation requires sustained growth factor signaling, but evidence showing p27 KIP1 up-regulation following mitogen removal suggests that such regulation may exist. 2 We have shown that Ras/RhoA activities are required for PDGFstimulated p27 KIP1 degradation (Figs. [2][3][4] and that constitutively active RhoA stimulates a loss of p27 KIP1 independent of mitogen. It is unclear, however, whether active RhoA acts directly or indirectly on p27 KIP1 to target p27 KIP1 for ubiquitinmediated degradation. p27 KIP1 is primarily a nuclear protein, which raises the possibility that active RhoA or downstream effectors of RhoA may translocate to the nucleus where they promote the degradation of p27 KIP1 . Although few proteins have been identified as downstream RhoA effectors, investigation into whether they may be downstream nuclear effectors involved in p27 KIP1 destruction seems warranted. However, the identification of RhoA as a necessary mediator of p27 KIP1 degradation clearly implicates RhoA as a signaling pathway for gene products important for G 1 progression. By disrupting the expression of newly formed cyclin D1⅐CDK complexes and p27 KIP1 by altering protein levels of cyclin D1 and p27 KIP1 , we were able to inhibit the mitogenic properties of PDGF. Overexpression of dnRhoA 19 was able to inhibit PDGFinduced cyclin D1⅐CDK complexes (Fig. 8A) overriding PDGFstimulated Ras/ERK signals. Inhibition of ERK activation blocked cyclin D1 induction but had no effect on PDGF-induced p27 KIP1 degradation, and inhibition of RhoA activity blocked p27 KIP1 degradation but had no effect on PDGF-induced cyclin D1 expression. These data provide evidence of the separate actions of these Ras-stimulated pathways. Although these pathways appear to determine the fate of distinct cell cycle proteins independently, this is the first study to show that they converge downstream to determine the activation state of cyclin D1⅐CDK complexes and subsequently coordinate mitogeninduced G 1 progression. FIG. 8. PDGF-induced RhoA activity is required for active cyclin D1⅐CDK complexes and subsequent G 1 /S transit. Panel A, growth-arrested IIC9 cells (WT) or IIC9 cells transfected with dnRhoA 19 were harvested at 0 and 24 h after the addition of PDGF (10 ng/ml) by scraping in cold 1 ϫ PBS and lysed. Lysates were incubated for 1-2 h at 4°C with a monoclonal cyclin D1 antibody. Cyclin D1 immune complexes were precipitated with protein G-Sepharose and assayed for their ability to phosphorylate soluble GST-Rb fusion protein in vitro as described under "Experimental Procedures." Panel B, growth-arrested wild type IIC9 cells (open bar) or dnRhoA 19 transfected IIC9 cells were stimulated with PDGF (10 ng/ml) (solid bar and lined bar, respectively) for 20 h and then assayed for [ 3 H]thymidine incorporation as described under "Experimental Procedures." The data indicate the mean Ϯ S.D. (n ϭ 3).