RhoA Stimulates p27Kip Degradation through Its Regulation of Cyclin E/CDK2 Activity*

RhoA has been identified as an important regulator of cell proliferation. We recently showed that the Ras/RhoA pathway regulates the degradation of p27Kip and the progression of Chinese hamster embryo fibroblasts (IIC9 cells) through G1 into S phase (Weber, J. D., Hu, W., Jefcoat, S. C., Raben, D. M., and Baldassare, J. J. (1997) J. Biol. Chem.272, 32966–32971). In this report, we have demonstrated that, in IIC9 cells, RhoA regulates cyclin E/CDK2 activity, which is required for p27Kip degradation. As previously shown in several fibroblasts cell lines, expression of dominant-negative CDK2 in IIC9 cells blocked serum-induced cyclin E/CDK2 activity and p27Kip degradation. In the absence of serum, expression of constitutively active RhoA(63) resulted in significant stimulation of cyclin E/CDK2 activity and degradation of p27Kip. Cotransfection of dominant-negative CDK2 and RhoA(63) inhibited RhoA(63)-induced cyclin E/CDK2 activity and p27Kipdegradation. In addition, expression of dominant-negative RhoA blocked serum-induced cyclin E/CDK2 activity and p27Kipdegradation. Finally, expression of catalytically active cyclin E/CDK2 rescued the effect of expression of dominant-negative RhoA. Taken together, these data show that RhoA regulates p27Kipdegradation through its regulation of cyclin E/CDK2 activity.

Cyclin-dependent kinases and their inhibitors are important for ordered progression through the cell cycle (1). Progression through G 1 into S phase requires activation of both cyclin D/CDK4 and cyclin E/CDK2 activities (1). In addition to the cyclins, specific inhibitory proteins (CKI proteins) associate with distinct cyclin kinase complexes and regulate their activities. These CKI proteins are divided into two families on the basis of sequence homology: the Cip/Kip and Ink4 families. The Cip/Kip family, which includes p21 Cip (2), p27 Kip1 (3,4), and p57 Kip2 (5), inhibits CDK4, CDK6, and CDK2 activities. The Ink4 family, which includes p16, p15, p18, and p19, specifically inhibits CDK4 and CDK6 activities (6 -8).
p27 Kip is expressed in most cells and is thought to be an important negative regulator of mitogen-induced progression through G 1 (9 -11). Expression of p27 Kip results in G 1 arrest in a variety of cell types (12). The level of p27 Kip increases when cells are growth-arrested by contact inhibition, mitogen withdrawal, or other anti-proliferative signals (9, 10, 14 -18). Consistent with the notion that p27 Kip is important in G 1 progression, mitogens stimulate elimination of p27 Kip during late G 1 or as cells enter G 1 /S. Interestingly, decreases in p27 Kip protein levels are not a result of any changes in mRNA levels (18,19), but mostly a result of decreased translation concomitant with increased ubiquitin-directed degradation (18 -20). Recently, considerable effort has been directed at understanding the pathway important for ubiquitin-directed degradation and the upstreamsignalingpathwaysthatregulatetargetingtoubiquitindependent proteolysis (21)(22)(23).
Mitogens stimulate p27 Kip degradation by inducing cyclin E/CDK2 phosphorylation of p27 Kip at Thr-187 (21)(22)(23). Expression of dominant-negative CDK2 in NIH-3T3 cells blocks serum-induced phosphorylation of p27 Kip at Thr-187 and p27 Kip degradation (21). Mutation of Thr-187 to alanine results in a stable form of p27 Kip , which is resistant to phosphorylation by cyclin E/CDK2 and degradation (21,22) and, when overexpressed, induces G 1 arrest. Taken together, these data strongly indicate that inhibition of p27 Kip degradation alone results in elevated levels of p27 Kip and inhibition of G 1 progression.
RhoA is a member of the Ras-like small G protein superfamily (24). In addition to its role in the formation of stress fibers and focal adhesion (25), RhoA also regulates Ras-dependent cell cycle progression and cell transformation (26,27). Activation of RhoA is required for G 1 cell cycle progression in Swiss 3T3 cells (26,27), and expression of geranylgeranylated RhoA facilitates the S phase entry in rat thyroid FRTL-5 cells (28). Moreover, Ras-mediated transformation in NIH-3T3 cells requires activation of both RhoA and Raf (26, 29 -31). We previously demonstrated that overexpression of dominant-negative Ras or dominant-negative RhoA inhibits PDGF 1 -induced DNA synthesis of IIC9 cells (32).
Recent data suggest that RhoA regulates IIC9 cell cycle progression by stimulating p27 Kip degradation. Expression of geranylgeranylated RhoA promotes the degradation of p27 Kip in rat thyroid FRTL-5 cells (28). In IIC9 cells, we found that expression of dominant-negative RhoA inhibits p27 Kip degradation (32). In the present study, we show that RhoA-induced activation of cyclin E/CDK2 precedes decreases in p27 Kip protein amounts. This is the first study to demonstrate a requirement for RhoA in the activation of cyclin E/CDK2 activity and clearly identifies a role for RhoA in growth and transformation.

EXPERIMENTAL PROCEDURES
Cells and Cell Culture-IIC9 cells, a subculture of Chinese hamster embryo fibroblasts, were grown and maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% (v/v) fetal calf serum. Subconfluent cultures were growth-arrested by incubation for 2 days in serum-free DMEM. Serum-arrested cells were washed twice and equilibrated in basal medium for 30 min prior to the addition of serum. Dominant-negative (dnCDK2) and wild-type CDK2 were the generous gifts Dr. Ed Harlow. Cyclin E was a generous gift from the laboratory of Dr. James Roberts. Dominant-negative RhoA (dnRhoA) and constitutively active RhoA (RhoA(63)) were constructed as described previously by site-directed mutagenesis of Thr to Asn at codon 19 or Gln to Leu at codon 63, respectively (32).
Transient Transfection-The cDNAs encoding wild-type CDK2, dnCDK2, wild-type cyclin E, dnRhoA, and RhoA(63) were transfected into subconfluent IIC9 cells using LipofectAMINE TM according to the manufacturer's protocol (Life Technologies, Inc.). Following expression of the cDNAs by incubation for 24 h in DMEM plus 10% serum, the cells were serum-arrested by incubation in DMEM for 48 h before agonist stimulation. Transient transfection using LipofectAMINE TM resulted in 80 -90% expression efficiency as visualized by ␤-galactosidase staining.
Cyclin E/CDK2 Kinase Assay-Subconfluent (80%) growth-arrested IIC9 cells were stimulated with 10% fetal calf serum. Cells were harvested at the indicated times after serum addition by scraping into 150 l of ice-cold lysis buffer (50 mM Hepes, 150 mM NaCl, 10% glycerol, 0.1% Tween 20, 1 mM NaF, 0.1 mM sodium vanadate, 50 mM ␤-glycerophosphate, 10 g/ml aprotinin, 10 g/ml pepstatin, and 10 g/ml leupeptin). The lysates were briefly sonicated, and insoluble material was pelleted by centrifugation at 14,000 ϫ g for 5 min at 4°C. Protein concentrations were determined using the Bio-Rad protein assay according to the manufacturer's specifications. Cyclin E-CDK2 immune complexes were isolated by incubation with anti-cyclin E antibody (Santa Cruz Biotechnology, Santa Cruz, CA) plus protein G. After several washes, the cyclin E-CDK2 complexes were incubated for 45 min at 30°C with histone and 0.5 Ci of [␥-32 P]ATP. The reactions were terminated by the addition of 5 l of 4ϫ Laemmli sample buffer. The samples were boiled for 5 min and subjected to SDS-polyacrylamide gel electrophoresis. The gels were dried, and cyclin E/CDK2 activities were quantified by PhosphorImager analysis (Molecular Dynamics, Inc.).
Western Blot Analysis-Serum-deprived cell cultures were incubated in the absence or presence of serum or PDGF. At the indicated times, the samples were scraped into 150 l of cold phosphate-buffered saline, centrifuged at 4°C for 5 min at 14,000 ϫ rpm, and then lysed on ice in 30 l of solubilization buffer (25 mM Hepes, 300 mM NaCl, 2 mM EDTA, 1.5 mM MgCl 2 , 1% Triton X-100, 0.5 mM phenylmethylsulfonyl fluoride, 0.1 mM Na 3 VO 4 , 10 g/ml leupeptin, 10 g/ml aprotinin, and 10 g/ml pepstatin). The samples were subjected to brief sonication and centrifuged for 5 min at 4°C. Protein concentrations of the supernatants were determined by the Bio-Rad protein assay as recommended by the manufacturer. Typically, 50 g of lysate protein were resolved by 12% SDS-polyacrylamide gel electrophoresis, and the separated proteins were transferred to polyvinylidene difluoride membranes (Immobilon, Millipore Corp., Bedford, MA). The membranes were probed with the appropriate antibody. Immunoreactive bands were visualized by chemiluminescence using the ECL Western blotting system (Amersham Pharmacia) as recommended by the manufacturer.

Cyclin E/CDK2 Activation Precedes Degradation of p27 Kip -
Recent studies have demonstrated that phosphorylation of p27 Kip at Thr-187 by cyclin E/CDK2 is required for ubiquitindependent degradation of p27 Kip (21)(22)(23). To determine whether cyclin E/CDK2 activity precedes serum-induced p27 Kip degradation, we examined the time course of p27 Kip degradation and cyclin E/CDK2 activity. Serum stimulated a rapid increase in cyclin E/CDK2 activity (Fig. 1A). Increased activity was detectable within 30 min, was maximal at 17 h, and remained elevated for 24 h (Fig. 1A). In contrast to cyclin E/CDK2 activity, a change in p27 Kip protein was not detectable at 30 min (Fig. 1B). A significant decrease in p27 Kip could be detected at 4 h, and p27 Kip was barely detectable at 24 h (Fig.  1B). These data demonstrate that cyclin E/CDK2 activity precedes any detectable decrease in p27 KIP .
The rapid increase in cyclin E/CDK2 activity was unexpected since, in most cell types, this activity does not increase until late in G 1 subsequent to activation of cyclin D/CDK4/CDK6. Previously, we showed that cyclin D/CDK4 activity is not detectable until 1-2 h after the addition of PDGF or serum (33), yet cyclin E/CDK4 activity increased within 30 min (Fig. 1A), indicating that, in IIC9 cells, cyclin E/CDK2 increases prior to a detectable increase in cyclin D/CDK4 activity. The timing of an increase in cyclin E/CDK2 in most cell types in late G 1 is likely a result of the dependence of cyclin E expression on cyclin D/CDK4 activity (34). To understand the early activation of cyclin E/CDK2 observed with IIC9 cells, we examined whether expression of cyclin E in IIC9 cells was independent of cyclin D/CDK4 (Fig. 1C). Significant amounts of cyclin E were detectable in quiescent IIC9 cells (Fig. 1C). A modest increase in cyclin E protein was detectable within 4 h, was maximal (1.5-2-fold) at 17 h, and remained elevated for 24 h (Fig. 1C). Cyclin E is therefore detectable in quiescent IIC9 cells and remains elevated throughout G 1 into S phase. The high level of cyclin E, which is expressed in quiescent IIC9 cells, is markedly different from the level of cyclin D1 protein (32,33). Cyclin D1 protein is very low but detectable in quiescent IIC9 cells and increases at least 5-fold with PDGF (32,33). Significant levels of cyclin E are present in quiescent cells, and serum stimulates only small increases. These data suggest that there are sufficient amounts of cyclin E in quiescent cells and indicate that the increases in this cyclin E/CDK2 activity occur independent of increased expression of cyclin E.
RhoA Regulates the Activity of Cyclin E/CDK2 and the Degradation of p27 Kip -Data obtained from overexpression of dominant-negative mutants of both RhoA and CDK2 indicate that mitogen-induced degradation of p27 Kip requires activation of RhoA (32) and cyclin E/CDK2 (21)(22)(23). To determine whether RhoA regulates cyclin E/CDK2 activity, we next examined the effect of blocking RhoA activation on cyclin E/CDK2 activity and p27 Kip protein (Fig. 2). Expression of dnRhoA inhibited Growth-arrested IIC9 cells were stimulated with 10% fetal calf serum, and cell lysates were prepared as described under "Experimental Procedures." Cyclin E immune complexes were immunoprecipitated from lysates containing 50 g of protein and assayed for their ability to phosphorylate histone protein (Histone-P) as described under "Experimental Procedures" (A). Similar results were obtained in four separate experiments. Lysates containing 40 g of protein were electrophoresed on 12% SDS-polyacrylamide gels and immunoblotted with anti-p27 Kip1 antibody (Santa Cruz Biotechnology) (B) or anti-cyclin E antibody (C) . Similar results were obtained in four separate experiments. serum-stimulated cyclin E/CDK2 activity ( Fig. 2A) as well as p27 Kip degradation (Fig. 2B). To further demonstrate that RhoA regulates cyclin E/CDK2 activity, IIC9 cells were treated with botulinum C3 exoenzyme, which specifically ADP-ribosylates RhoA and blocks its function (35). This treatment significantly inhibited serum activation of cyclin E/CDK2 activity and the serum-induced decrease in p27 Kip (data not shown).
Our data show that cyclin E levels were significantly elevated in quiescent IIC9 cells (Fig. 1C). Furthermore, expression of dnRhoA inhibited serum-induced cyclin E/CDK2 activity ( Fig. 2A) and p27 Kip degradation (Fig. 2B). Previously, we found that PDGF stimulates p27 Kip degradation in the absence of cyclin D expression and cyclin D/CDK4 activity (32), suggesting that, in IIC9 cells, RhoA itself regulates the levels of p27 Kip . We therefore decided to determine whether expression of a constitutively active mutant of RhoA (RhoA(63)) stimulates cyclin E/CDK2 activity in the absence of agonist (Fig. 3A). Expression of RhoA(63) induced a significant increase in cyclin E/CDK2 activity in the absence of serum (Fig. 3A). This activity was ϳ80% of the maximal activity seen with serum-stimulated IIC9 cells. Although expression of RhoA(63) could induce cyclin E/CDK2 activity in quiescent IIC9 cells, serum stimulated a further increase (ϳ1.4-fold) in the activity of cyclin E/CDK2 in IIC9 cells expressing RhoA(63) (Fig. 3A). This increase in activity is consistent with the ability of serum to induce a 2-fold increase in cyclin E (Fig. 1C). In addition to increasing cyclin E/CDK2 activity and consistent with the notion that cyclin E/CDK2 activity is necessary for p27 Kip degradation, expres-sion of RhoA(63) reduced p27 Kip protein to levels comparable to those seen with serum stimulation (Fig. 3B). These data clearly demonstrate that expression of constitutively active RhoA alone can markedly increase cyclin E/CDK2 activity and induce p27 Kip degradation. To further examine the ability of RhoA(63) to mediate the loss of p27 Kip , we next determined whether coexpression of dnCDK2 with RhoA(63) blocks p27 Kip breakdown. In contrast to expression of RhoA(63) alone, coexpression of dnCDK2 and RhoA(63) did not result in increased activity (data not shown) or a decrease in p27 Kip protein in quiescent IIC9 cells (Fig. 4). These data are in agreement with the results demonstrating that serum-induced degradation of p27 Kip is dependent on cyclin E/CDK2 activity and suggest that cyclin E/CDK2 activity is downstream of RhoA.
Expression of Cyclin E/CDK2 Rescues dnRhoA-dependent Inhibition of p27 Kip Degradation-We next examined whether expression of cyclin E/CDK2 induces the degradation of p27 Kip protein in cells expressing dnRhoA. Expression of cyclin E/CDK2 results in significant cyclin E/CDK2 activity in quiescent IIC9 cells (Fig. 5A) and loss of p27 Kip (Fig. 5B). These changes occurred in the absence of cyclin D/CDK4 activity (data not shown). Therefore, up-regulation of cyclin E/CDK2 activity itself induced p27 Kip degradation. To determine whether up-regulation of cyclin E/CDK2 activity rescues the effects of dnRhoA, we examined cyclin E/CDK2 activity and p27 Kip levels in cells cotransfected with dnRhoA and cyclin E/CDK2 (Fig. 6). Cyclin E/CDK2 activity was markedly increased in cells overexpressing dnRhoA and cyclin E/CDK2 (Fig. 6A). Furthermore, p27 Kip levels were reduced in these cells (Fig. 6B). Taken together, these data strongly suggest that

FIG. 3. Expression of RhoA(63) induces cyclin E/CDK2 activity and p27 Kip degradation in quiescent IIC9 cells. Subconfluent IIC9
cells were transfected with 1 g of RhoA(63) and growth-arrested as described for Fig. 2. A, cyclin E immune complexes were immunoprecipitated from lysates (50 g of protein) of untransfected cells. The immune complexes were assayed for their ability to phosphorylate histone protein (Histone-P) as described under "Experimental Procedures." Similar results were obtained in four separate experiments. B, lysates from samples described in A containing 40 g of protein were electrophoresed on 12% SDS-polyacrylamide gels and immunoblotted with anti-p27 Kip1 antibody. Similar results were obtained in three separate experiments. WT, wild type.

RhoA is upstream of cyclin E/CDK2.
Expression of Cyclin E/CDK2 Rescues dnRhoA-induced Growth Arrest-Inhibition of RhoA by expression of dnRhoA or treatment with botulinum C3 exoenzyme resulted in significant inhibition of growth (Fig. 7A). This is consistent with the observation that RhoA is important in mammalian cell growth and transformation (26,27). Our data indicate that RhoA regulates cell cycle progression at least partially through its regulation of cyclin E/CDK2 activity and p27 Kip levels. If cyclin E/CDK2 activity is the major cell cycle downstream effector, cotransfection of cyclin E/CDK2 should rescue the arrest observed with expression of dnRhoA. Indeed, serum stimulated significant growth in IIC9 cells expressing dnRhoA and cyclin E/CDK2 (Fig. 7A). In IIC9 cells coexpressing dnRhoA and cyclin E/CDK2, serum was less mitogenic than in wild-type IIC9 cells (Fig. 7A). However, growth under these conditions was virtually identical to growth in the presence of a potent mitogen, PDGF (Fig. 7B), indicating that cyclin E/CDK2 activation is a target for RhoA-dependent regulation of cell cycle progression through G 1 into S phase. Whereas expression of dnRhoA could block PDGF-stimulated growth, overexpression of cyclin E/CDK2 rescued growth to near identical values seen in cells treated with PDGF alone (Fig. 7B). These data demonstrate that cyclin E/CDK2 activity is a major target of RhoA.

DISCUSSION
Observations from several independent studies suggest that oncogenic Ras may mediate its action through both Raf-dependent and RhoA-dependent pathways (29,31). Expression of constitutively active Ras results in transformation in NIH-3T3 cells; expression of constitutively active Raf or RhoA alone only weakly transforms NIH-3T3 cells (29). Furthermore, cotransfection with either dominant-negative Rac or RhoA blocks Rasinduced transformation (29). Our previous studies have shown that the Ras/Raf/mitogen-dependent protein kinase pathway is important for the induction of cyclin D and that RhoA regulates p27 Kip degradation (32,33). In this study, we have demon- FIG. 4. Expression of dnCDK2 blocks RhoA(63)-induced p27 Kip degradation. Subconfluent IIC9 cells were transfected with 1 g of RhoA(63) plus 1 g of dnCDK2 and growth-arrested as described for Fig. 2. Lysates containing 40 g of protein were electrophoresed on 12% SDS-polyacrylamide gels and immunoblotted with anti-p27 Kip 1 polyclonal antibody. Similar results were obtained in four separate experiments.

FIG. 5. Expression of cyclin E/CDK2 induces cyclin E/CDK2
activity and p27 Kip degradation in quiescent IIC9 cells. Subconfluent IIC9 cells were transfected with 1 g of cyclin E plus 1 g of CDK2 and growth-arrested as described for Fig. 2. Lysates containing 50 g of protein were prepared from growth-arrested IIC9 cells or from cells stimulated for 17 h with 10% fetal calf serum as described under "Experimental Procedures." A, cyclin E immune complexes were prepared from lysates containing 50 g of protein and assayed for their ability to phosphorylate histone protein (Histone-P) as described under "Experimental Procedures." Similar results were obtained in four separate experiments. B, lysates (40 g of protein) were electrophoresed on 12% SDS-polyacrylamide gels and immunoblotted with anti-p27 Kip 1 antibody. Similar results were obtained in three separate experiments. WT, wild type.
FIG. 6. Expression of cyclin E/CDK2 rescues dnRhoA-dependent inhibition of p27 Kip degradation. Subconfluent IIC9 cells were transfected with dnRhoA alone or dnRhoA plus cyclin E/CDK2 and then growth-arrested as described for Fig. 2. Lysates were prepared from quiescent cells or from cells stimulated with 10% fetal calf serum for 17 h. A, cyclin E immune complexes were isolated from lysates containing 50 g of protein and assayed for their ability to phosphorylate histone protein (Histone-P) as described under "Experimental Procedures." Similar results were obtained in four separate experiments. B, lysates (40 g of protein) were electrophoresed on 12% SDS-polyacrylamide gels and immunoblotted with anti-p27 Kip1 antibody. Similar results were obtained in four separate experiments. strated for the first time that RhoA induces p27 Kip degradation through regulation of cyclin E/CDK2 activity. Our data therefore indicate that the Ras/Erk pathway regulates cyclin D/CDK4/CDK6 activity and that the Ras/RhoA pathway regulates cyclin E/CDK2 activity. Both these activities are crucial for passage through G 1 into S phase (Fig. 8).
Several recent reports (21)(22)(23) and the data in this paper clearly demonstrate that mitogen-induced p27 Kip degradation is dependent on cyclin E/CDK2 activity. Previously, we found that, in IIC9 cells, PDGF stimulates p27 Kip degradation in the absence of cyclin D/CDK4/CDK6 activity (32). This is likely a result of the high level of cyclin E expressed in quiescent IIC9 cells (Fig. 1C). Interestingly, progression through G 1 in these cells still requires cyclin D expression, suggesting that, in cells containing normal levels of retinoblastoma protein, 2 progression through G 1 requires both cyclin D-and cyclin E-dependent kinases. These results are in agreement with data that indicate that retinoblastoma protein must be phosphorylated by cyclin D-dependent kinases prior to cyclin E/CDK2 (37).
How RhoA regulates cyclin E/CDK2 activity is still an open question. Olson et al. (36) found that RhoA blocks the induction of p21 Cip1 , which is an inhibitor of cyclin/CDK activity and an adapter protein for assembly of D cyclins with their respective CDK proteins (37). Microinjection of constitutively active Ras into NIH-3T3 cells in the absence of active RhoA induces p21 Cip1 expression and does not stimulate DNA synthesis (36). However, in the presence of active RhoA, p21 Cip1 expression is repressed, and cells enter S phase and synthesize DNA. More important, in cells from transgenic mice nullizygous for the gene encoding p21 Cip1 , active RhoA is not required for Rasinduced DNA synthesis, suggesting that the suppression of p21 Cip1 is the only role of RhoA in Ras-induced DNA synthesis. Although this is an intriguing suggestion for the role of RhoA, it does not likely account for our results and therefore is not the only function for RhoA in progression through G 1 . Ras-induced p21 Cip1 induction is likely a result of activation of the Ras/Erk pathway (39). This suggests that Ras stimulates cyclin D and p21 Cip1 expression. Previously, we showed that the Ras/Rho (but not the Ras/Erk) pathway is required for PDGF-induced p27 Kip degradation (32). The data in this paper clearly show that RhoA stimulates p27 Kip degradation by activation of cyclin E/CDK2. This occurs in the absence of Erk activation 2 and likely in the absence of p21 Cip1 expression. These distinct roles of RhoA in G 1 progression may possibly be determined by careful analysis of RhoA protein with mutations of the effectorbinding sites and therefore await further investigation.
Oddly, p27 Kip is an inhibitor and substrate of cyclin E/CDK2. Phosphorylation of p27 Kip by cyclin E/CDK2 is required for its degradation in G 1 (21). In vitro kinetic analysis of the association of p27 Kip with cyclin E/CDK2 suggests that p27 Kip binds to cyclin E/CDK2 with high and low affinities (21). Initially, it binds with low affinity, and then slowly, the binding shifts to high affinity (21). Furthermore, when it binds with low affinity, it is a substrate, and at high affinity, an inhibitor (21). At 2 W. Hu, C. J. Bellone, and J. J. Baldassare, unpublished data. FIG. 7. Expression of cyclin E/CDK2 rescues dnRhoA-induced growth arrest. Subconfluent IIC9 cells were transfected with empty vector, 1 g of dnRhoA alone, or 1 g of dnRhoA plus 1 g of cyclin and 1 g of CDK2 and then growth-arrested as described for Fig. 2. A, growth-arrested cells were incubated for 17 h with DMEM alone or with DMEM plus 10% fetal calf serum. In those samples treated with botulinum C3 transferase, growth-arrested IIC9 cells were preincubated for 2 h with 40 g/ml botulinum C3 transferase and then treated for 17 h with 10% fetal calf serum. Following the 17-h incubation, the cells were assayed for [ 3  equilibrium, p27 Kip inhibits the activity. It is possible that, in vivo, certain conditions favor high or low affinity binding and thus determine whether it is a substrate or inhibitor. RhoA may be an important regulator of this association. The present data clearly demonstrate that RhoA regulates cyclin E/CDK2 activation and identify a role for RhoA in progression through G 1 of the cell cycle.