ã

Oncogene (2000) 19, 690 ± 699 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc

[Lys61]N-Ras is able to induce full activation and nuclear accumulation of Cdk4 in NIH3T3 cells Priam Villalonga1, EulaÁlia Rius1, Oriol Bachs1 and Neus Agell*,1 1

Departament de Biologia Cel.lular i Anatomia PatoloÁgica, Institut d'Investigacions BiomeÁdiques August Pi i Sunyer (IDIBAPS), Facultat de Medicina, Universitat de Barcelona, 08036 Barcelona, Spain

The elements of the cell cycle regulatory machinery activated by the oncogenic form of Ras, [Lys61]N-Ras, have been analysed in NIH3T3 cells. We demonstrate that [Lys61]N-Ras expression is able to induce full cdk4 activation. As already reported, oncogenic Ras expression was sucient to induce cyclin D1 and p21cip1 expression and their association with cdk4. Furthermore, serum-starved [Lys61]N-Ras NIH3T3 cells showed nuclear accumulation of cyclin D1 and cdk4 not observed in serum-starved NIH3T3 cells. This accumulation of cdk4 into the cell nucleus observed in serum-starved [Lys61]NRas NIH3T3 cells was inhibited by a microinjection of neutralizing anti-Ras antibodies. Thus, active [Lys61]NRas was a sucient signal to induce nuclear accumulation of cyclin D1/cdk4, leading to its full activation. Transfection of [Lys61]N-Ras NIH3T3 cells with an inactive form of MEK or their treatment with PD 98059, showed that nuclear translocation of cdk4 was MEK dependent. Interestingly, cells constitutively expressing [Lys61]N-Ras did not inactivate pRb and did not proliferate in the absence of serum. This may be due to the fact that although association of cdk2 with cyclin E and the translocation of those complexes to the nucleus were achieved, [Lys61]N-Ras expression was not sucient to induce cdk2 activation. The high levels of p27kip1 that were found in cyclin E/cdk2 complexes may be responsible for the inability of oncogenic Ras to activate this kinase. In consequence, oncogenic alterations that lead to a decrease in p27kip1 bound to cyclin E may cooperate with Ras to induce full cdk2 activation, pRb inactivation and thus cell proliferation. Oncogene (2000) 19, 690 ± 699. Keywords: Ras; cdk4; cdk2; cell cycle; cell nucleus Introduction Non transformed mammalian cells need growth factors and cell adhesion signals in order to enter the cell cycle and progress through G1 to the restriction point (Pardee, 1989). During this interval, growth stimulatory and growth inhibitory signals from the extracellular environment activate transduction pathways that converge on the cell cycle control machinery (Peeper and Bernards, 1997; Roussel, 1998; Sherr, 1994). This machinery is governed by a family of

*Correspondence: N Agell Received 2 February 1999; revised 22 October 1999; accepted 27 October 1999

serine/threonine protein kinases called cyclin-dependent kinases (cdks). These kinases are formed by two subunits, a catalytic subunit that is present throughout the cell cycle, and a regulatory subunit called cyclin that is present only at speci®c stages of the cell cycle. Cyclin D/cdk4, cyclin E/cdk2, cyclin A/cdk2, cyclin A/ cdk1 and cyclin B/cdk1 are sequentially activated during G1, G1/S transition, S phase, G2/M and mitosis, respectively (Morgan, 1997; Norbury and Nurse, 1992; Reed, 1992; Sherr, 1994). The major function of cyclin D/cdk4 during G1 is the phosphorylation of the proteins of the pocket family, which includes the retinoblastoma protein (pRb), p107 and p130 (Dong et al., 1998; Resnitzky and Reed, 1995). The hypophosphorylated form of these proteins binds the transcription factor E2F, and upon phosphorylation, E2F is released and can activate transcription of S phase-speci®c genes (Draetta, 1994; Lam and La Thangue, 1994; Mittnacht, 1998). Among the genes containing E2F-binding sites in their promoter regions there are those coding for enzymes and proteins that are essential for DNA replication such as DNA polymerase a, thymidin kinase, ribonucleotide reductase, cyclin A and cyclin E (Botz et al., 1996; Dalton, 1992; Dou et al., 1994; Geng et al., 1996; Zwicker et al., 1995). The activation of cyclin D/cdk4 and phosphorylation of pRb is thus a limiting step for G1 progression. Cyclin E/cdk2 is also essential for G1/ S transition. It also phosphorylates pRb, and is essential for its functional inactivation, and possibly other proteins essential for the onset of DNA replication (Lundberg and Weinberg, 1998; Reed, 1992). Cdk activity is regulated in response to extracellular and intracellular signals. Several mechanisms are involved in this regulation (Grana and Reddy, 1995; Morgan, 1995): cyclin binding, catalytic subunit phosphorylation (Jinno et al., 1994; Kato et al., 1994a) binding to speci®c inhibitory proteins (CKIs) (Kato et al., 1994b; Sherr and Roberts, 1995), and intracellular localization (Jin et al., 1998; Taules et al., 1998). To date, seven CKIs have been identi®ed in mammalian cells (Sherr and Roberts, 1995). They are divided into two major families: the INK4 and the CIP/KIP. The INK4 family includes p16INK4a, p15INK4b, p18INK4c and p19INK4d, which interact speci®cally with cdk4 and cdk6. Addition of p16INK4a to cyclin D/cdk4 complexes results in the dissociation of the kinase complex, and binding of p16INK4a to monomeric cdk4 prevents cyclin D association (Parry et al., 1995). The CIP/KIP family of inhibitors comprises three members: p21cip1, p27kip1 and p57kip2. They di€er from the INK4 family in that they can inhibit all the G1 cdk/cyclin complexes. Moreover, they have higher anity for the cyclin/cdk

Oncogenic Ras and cdks regulation P Villalonga et al

complexes than for the monomeric cdk subunit (Harper et al., 1993; Polyak et al., 1994). Most cancer cells have lost regulation of the cell cycle machinery by extracellular signals. This is a consequence not only of alterations in the cell cycle regulatory machinery but also in the transduction pathways that integrate those extracellular signals. Thus, it is of interest to analyse the link between extracellular signals and the cell cycle regulatory machinery. One of the central molecules in signal transduction pathways is the small guanine nucleotide-binding protein Ras. Ras signalling is involved in many cellular processes such as proliferation, di€erentiation, survival or even apoptosis (Downward, 1998). In response to mitogenic stimuli, including growth factors, G protein-coupled receptor agonists, and hemopoietic cytokines, cellular Ras becomes activated within minutes (McCormick, 1994). Active GTP-bound Ras then interacts with and activates a number of downstream e€ectors including Raf, the p110 catalytic subunit of phosphatidylinositol-3-kinase (Rodriguez-Viciana et al., 1994) and the guanine nucleotide exchange factor Ral-GDS (Katz and McCormick, 1997; Marshall, 1996). The serine and threonine protein kinase cascade consisting of Raf, MAP kinase kinase (MAPKK or MEK), and mitogen-activated protein kinase (MAPK or ERK) is one of the best characterized Ras e€ector systems (Davis, 1993; Lewis et al., 1998; Robinson and Cobb, 1997). The fact that Ras activation is involved in the development of many human cancers and that oncogenic Ras participates in the transformation of most immortalized cell lines (Barbacid, 1987; Bos, 1989), suggested that Ras activation would lead to the activation of the cell cycle machinery. At this moment the way in which Ras activation in¯uences the later events in the cell cycle is beginning to fall into place (Downward, 1997; Kerkho€ and Rapp, 1998; Lloyd, 1998). In fact microinjection of activated, but not wild-type forms of Ras can initiate DNA synthesis in quiescent ®broblasts (Feramisco et al., 1984), whereas neutralizing Ras antibodies block the DNA synthesis induced by mitogens (Dobrowolski et al., 1994; Mittnacht et al., 1997; Mulcahy et al., 1985; Peeper et al., 1997). The importance of pRb inactivation by Ras was demonstrated by the ®nding that Rb7/7 cells no longer require Ras activity to cycle (Peeper et al., 1997). Several reports implicate the Ras/ERK signalling pathway in the mitogendependent induction of cyclin D1 expression: oncogenic Ras proteins can induce expression of cyclin D1 (Albanese et al., 1995; Liu et al., 1995; Winston et al., 1996); in some cell lines the use of Ras negative mutants has shown that Ras is essential for cyclin D1 expression (Aktas et al., 1997; Weber et al., 1997); and, ERK1, 2 and MEK-1 activation also induces cyclin D1 expression (Cheng et al., 1998; Lavoie et al., 1996). Furthermore, Ras activation in Balb/c-3T3 ®broblasts (Winston et al., 1996) and MEK-1 activation in NIH3T3 cells (Cheng et al., 1998), induce assembly of cyclin D1 and cdk4 although not full activation of the complex. It is also established that Ras activation induces the expression of p21cip1 through the MAPK signalling pathway (Liu et al., 1996; Olson et al., 1998) and participates in the signalling of p27kip1 degradation (Aktas et al., 1997;

Takuwa and Takuwa, 1997; Winston et al., 1996). On the whole, an involvement of Ras in the activation of the two G1 kinases needed for Rb phosphorylation (cyclin D/cdk4 and cyclin E/cdk2) is well accepted. But the speci®c steps essential for the full activation of those kinases that can be induced with the single activation of Ras are not completely established. We have analysed the e€ect of oncogenic Ras on the expression of G1 cdks and cyclins, the composition of the induced complexes, their subcellular localization and their kinase activity. We show that although oncogenic [Lys61]N-Ras is able to induce full activation of cdk4 and its translocation to the cells nucleus, cdk2 activation is not accomplished due to the association of p27kip1 to the complex. In consequence, pRb is not inactivated and cells do not proceed to S phase.

691

Results Expression of oncogenic [Lys61]N-Ras does not induce proliferation in serum-starved NIH3T3 cells In order to analyse the cell cycle regulatory elements activated by Ras, NIH3T3 cells constitutively expressing [Lys61]N-Ras were used. To con®rm that [Lys61]NRas NIH3T3 cells expressed the activated mutant, we used RBD-Sepharose to determine the amount of GTP-bound Ras under serum deprivation conditions. As expected, [Lys61]N-Ras cells but not normal NIH3T3 cells had the Ras-GTP form after 2 days of serum starvation (0.5% FCS) (Figure 1a). Although Ras was present in the active form, cells were eciently arrested after serum starvation, as indicated by BrdU incorporation and FACS analysis (Figure 1b,c). Thus, in the absence of other stimuli, oncogenic [Lys61]N-Ras expressing cells did not enter S phase. Expression of oncogenic [Lys61]N-Ras does not lead to pRb inactivation in serum-starved NIH3T3 cells One essential process for G1 progression is the phosphorylation of pRb, which leads to its inactivation and in consequence to the expression of E2Fdependent genes. We analysed whether Ras activation by itself could induce pRb phosphorylation and the expression of genes with E2F binding sites in their promoter. Thus, pRb status in [Lys61]N-Ras NIH3T3 cells and in NIH3T3 control cells after serum starvation or after serum addition was analysed by Western blot. As shown in Figure 2a, at 0.5% FCS the activation of Ras was not enough to induce the gel mobility shift of pRb indicating that pRb was not hyperphosphorylated. Addition of FCS to 10% led to pRb phosphorylation 20 h later, as in control NIH3T3 cells. To con®rm that activation of Ras did not lead to pRb function inactivation, the expression of genes containing E2F-binding domains in their promoter was also analysed in serum-starved versus activated [Lys61]N-Ras NIH3T3 cells. As shown in Figure 2b, serum-starved cells did not express either p107, cdc2, cycA or PCNA (Proliferating Cell Nucleus Antigen), and those proteins were present 15 h after 10% FCS addition. Thus, oncogenic [Lys61]N-Ras cannot inactivate pRb in NIH3T3 cells. Oncogene

Oncogenic Ras and cdks regulation P Villalonga et al

692

Figure 1 [Lys61]N-Ras NIH3T3 cells have active Ras, but do not synthesize DNA in the absence of serum. (a) The presence of Ras-GTP was analysed in serum-starved NIH3T3 cells expressing (+) or not (7) [Lys61]N-Ras. Cells were lysed, active Ras was pulled-down with RBD-Sepharose and then detected by immunoblot as indicated in Materials and methods. (b) BrdU incorporation for 3 h was analysed by immunocytochemistry as indicated in Materials and methods in NIH3T3 and [Lys61]N-Ras NIH3T3 cells. Incorporation was measured in exponentially growing cells (FCS +) and in cells starved for 2 days (FCS 7). (c) [Lys61]NRas NIH3T3 cells were serum starved for 2 days and then activated to proliferate by FCS addition. Cells were collected just after the period of serum starvation (7FCS) or 16 h after FCS addition (+FCS 16 h) and FACS analysis performed as indicated in Materials and methods

Oncogenic [Lys61]N-Ras expression in NIH3T3 cells induces cdk4 but not cdk2 activation Because the growth suppressor function of pRb is cancelled by phosphorylation, the activity of cdk4 and cdk2, the two cdks that are major pRb-kinases, was analysed in oncogenic Ras-expressing cells versus control NIH3T3 cells. Serum-starved NIH3T3 cells did not show either cdk4 or cdk2 activity, and both activities were induced at 9 h after FCS addition (Figure 3). In serum-starved [Lys61]N-Ras cells, activation of cdk4 similar to that induced by FCS in control NIH3T3 cells was observed. In contrast, no activation of cdk2 was shown in the absence of FCS (Figure 3). Thus, the sole activation of Ras in NIH3T3 cells was able to induce the full activation of cdk4 but not of cdk2. Addition of FCS to [Lys61]N-Ras NIH3T3 cells led to activation of cdk2 as in control NIH3T3 cells (Figure 3). Oncogene

Figure 2 pRb is not inactivated in serum-starved [Lys61]N-Ras NIH3T3 cells. (a) Control NIH3T3 cells or [Lys61]N-Ras NIH3T3 cells were serum starved and then activated to proliferate by FCS addition. Cells were lysed after serum starvation (7) or at di€erent times after serum addition (9 h, 20 h) and Western blot performed using anti-pRb antibodies. The slowest migrating band corresponds to hyperphosphorylated pRb (pRb-P) and the fastest one to non-hyperphosphorylated pRb (pRb). (b) The levels of p107, cdc2, cyclin A and PCNA were analysed by Western blot in serum starved (7) or 16 h after serum addition (+) [Lys61]N-Ras NIH3T3 cells. The same amount of protein was loaded in each lane and as a control of loading cdk4 protein levels were analysed

Effect of oncogenic [Lys61]N-Ras expression in NIH3T3 cells on the levels of proteins involved in the regulation of cdk4 and cdk2 activity Many regulatory mechanisms are involved in the achievement of full cdk activity. First, we analysed the e€ect of Ras activation on the expression levels of di€erent proteins involved in the regulation of cdk4 and cdk2 activity. Protein levels of cdk4, cdk2, cdk7, cyclin D1, cyclin D3, cyclin E, cyclin H, p27kip1 and p21cip1 were compared in NIH3T3 and [Lys61]N-Ras NIH3T3 cells under serum starvation or 9 h after FCS addition. As shown in Figure 4, cdk4 and cdk2 were already present in serum-starved NIH3T3 cells and the levels did not increase after 10% FCS addition. The same behaviour was observed in [Lys61]N-Ras NIH3T3 cells. The e€ect of Ras activation on the regulatory subunit of those kinases was diverse. As already reported, cyclin D1 was not detected in the absence of serum, but Ras activation led to an increase in cyclin D1 levels even higher than that observed after 10% FCS addition in NIH3T3 cells (Figure 4). Levels of cyclin D1 were markedly decreased when [Lys61]NRas 3T3 cells were treated with MEK inhibitor PD98059 for the last 20 h of serum starvation, indicating that the increment of cyclin D1 observed in these cells was MEK dependent (data not shown). Cyclin D3 expression was almost the same in serumstarved and in FCS-activated cells, and no changes in

Oncogenic Ras and cdks regulation P Villalonga et al

Figure 3 Cdk4 and cdk2 activity in NIH3T3 and [Lys61]N-Ras NIH3T3 cells. NIH3T3 and [Lys61]N-Ras NIH3T3 cells were serum starved for 2 days (FCS7) and then activated to proliferate by FCS addition for 9 h (FCS+). Cells were lysed and 500 mg of proteins were immunoprecipitated with anti-cdk4, anti-cdk2 polyclonal antibodies or normal rabbit serum (NRS) as indicated in Materials and methods. Kinase activity in the immunoprecipitates was analysed as indicated using GST-pRb or H1 as substrates for cdk4 and cdk2 respectively. A representative experiment out of three repetitions is shown in the ®gure

Cdk7 was expressed at the same level in both control and oncogenic Ras expressing cells irrespective of the presence of mitogens, and cyclin H also showed the same pattern of expression. The levels of two of the CKIs, p21cip1 and p27kip1, were also analysed. As previously described, Ras activation induced p21cip1 expression: p21cip1 was almost absent in NIH3T3 in both serum-starved cells and 9 h after FCS stimulation. Nevertheless, in these cells, p21cip1 is induced upon mitogenic stimulation, peaking at 2 h, whereas its level decreases to basal levels at about 5 h after FCS addition (data not shown and (Bosch et al., 1998)). In [Lys61]N-Ras NIH3T3 cells, p21cip1 expression was induced in both serum-deprived and FCS-stimulated cells. The levels of p27kip1 were high in serum-starved NIH3T3 cells and decreased after serum addition. Ras activation in serum-starved cells did not induce a decrease in the amount of p27kip1, but a lower electrophoretical mobility band appeared in both serum-starved and non-starved [Lys61]N-Ras NIH3T3 cells. FCS addition to oncogenic Ras-expressing cells triggered p27kip1 downregulation as in NIH3T3 cells. Thus, oncogenic [Lys61]N-Ras, in the absence of mitogens, induced expression of cyclin D1 and p21cip1, but not p27kip1 degradation in NIH3T3 cells.

693

Effect of oncogenic [Lys61]N-Ras expression in NIH3T3 cells on the G1 cyclin/cdks complexes

Figure 4 Expression of G1 phase regulatory proteins in NIH3T3 and [Lys61]N-Ras NIH3T3 cells. NIH3T3 and [Lys61]N-Ras NIH3T3 cells were serum starved for 2 days (FCS7) and then activated to proliferate by FCS addition for 9 h (FCS+). Cells were lysed and 30 mg of protein loaded on di€erent SDS polyacrylamide gels and Western blots performed to visualize the proteins indicated in the ®gure. Exact protocol and antibodies used are indicated in Materials and methods

the protein levels were observed due to Ras activation. Cyclin E was expressed in serum-starved cells although an induction was observed after serum addition. Ras activation did not have any e€ect on cyclin E levels. The subunits of the kinase responsible for cdks phosphorylation (cyclin H/cdk7) were also analysed.

To better understand the molecular mechanisms responsible for the di€erent e€ect of oncogenic Ras expression on cdk4 and cdk2, the association between the catalytic subunits, regulatory subunits (cyclins) and inhibitors was analysed in NIH3T3 and [Lys61]N-Ras NIH3T3 cells. In serum-starved and FCS-stimulated (9 h) cells, the presence of cyclin D1, cyclin D3, p27kip1 and p21cip1 was analysed in anti-cdk4 immunoprecipitates by Western blot using speci®c antibodies (Figure 5a). Cyclin D3 was already complexed with cdk4 in serum-starved NIH3T3 cells irrespective of the presence of active Ras. Addition of FCS had no e€ect on the association of cyclin D3. No association of cyclin D1 with cdk4 was observed in serum-starved cells, consistent with the low levels of total cyclin D1 observed in these cells (Figure 4). After serum addition, cyclin D1 was found in cdk4 immunoprecipitates of these cells. In serum-starved [Lys61]N-Ras NIH3T3 cells the levels of cyclin D1 in the cdk4 immunoprecipitates were even higher than in FCS-stimulated NIH3T3 cells, indicating that Ras activation not only induced cyclin D1 expression (Figure 4) but also its association with cdk4. The same amount of p27kip1 was found associated with cdk4 in serum-starved or in FCSstimulated NIH3T3 cells. The presence of active Ras did not decrease the association of p27kip1 with cdk4 in either serum-starved or FCS-activated cells. No p21cip1 was found associated with cdk4 in serum-starved NIH3T3 cells but it was present after FCS addition. Active Ras induced, even in the absence of FCS stimulation, not only p21Cip1 expression (Figure 4) but also its association with cdk4 (Figure 5a). The presence of cyclin E, p27kip1 and p21cip1 was analysed in anti-cdk2 immunoprecipitates by Western blot using speci®c antibodies. In both serum-starved NIH3T3 and [Lys61]N-Ras cells, cyclin E was found already bound to cdk2, but in FCS-stimulated cells the Oncogene

Oncogenic Ras and cdks regulation P Villalonga et al

694

Figure 5 Analysis of cdk4 and cdk2 complexes in NIH3T3 and [Lys61]N-Ras NIH3T3 cells. NIH3T3 and [Lys61]N-Ras NIH3T3 cells were serum starved for 2 days (FCS7) and then activated to proliferate by FCS addition for 9 h (FCS+). Cells were lysed and proteins (800 mg each sample) were immunoprecipitated using anti-cdk4 polyclonal antibodies (a) or anti-cdk2 monoclonal antibodies (b). As controls NRS or a non-related anti-human monoclonal antibody were used respectively. Immunoprecipitated proteins were resolved on SDS ± PAGE, and Western blot performed to visualize the proteins indicated in the ®gure. The proteins were detected in the same membrane after stripping and reprobing. A representative experiment out of three repetitions is shown in the ®gure

amount of cyclin E bound to cdk2 was clearly increased (Figure 5b). In contrast, p27kip1 accumulated in serum-deprived NIH3T3 and [Lys61]N-Ras NIH3T3 cells, whereas in FCS-stimulated cells p27kip1 was almost completely absent from cdk2 immunocomplexes (Figure 5b), in correlation with the downregulation observed in FCS-stimulated cells by Western blot (Figure 4). The other CKI, p21cip1, was almost undetectable in cdk2 immunocomplexes. Thus, the high level of p27kip1, together with the low levels of cyclin E found in serum starved [Lys61]N-Ras NIH3T3 cdk2 immunocomplexes may account for the inability of [Lys61]N-Ras to induce H1 kinase activity. Expression of oncogenic [Lys61]N-Ras in NIH3T3 cells induces nuclear localization of cdk4, cyclin D1 and cdk2 Since de®nitive activation of cdks by cyclin H/cdk7 and phosphorylation of pocket proteins takes place in the cell nucleus, the e€ect of Ras activation on the intracellular location of cdk4 and cdk2 was also analysed using immuno¯uorescence techniques. Almost all serum-starved NIH3T3 cells (80 ± 90%) showed a clear cytoplasmatic localization of cdk4 and cdk2. At 9 h after serum addition 75 and 60% of the cells showed an intense nuclear labelling for cdk4 and cdk2 respectively, in parallel to a decrease in the cytoplasmatic staining (Figure 6). This indicated that FCS stimulation induced translocation of both kinases to Oncogene

the cell nucleus. In contrast, cdk4 and cdk2 were predominantly nuclear in serum-starved [Lys61]N-Ras NIH3T3 cells (85 and 75% of the cells with intense labelling for cdk4 and cdk2 respectively), and the intracellular localization of those kinases did not change upon serum addition. Thus, activation of Ras is enough to induce nuclear translocation of cdk4 and cdk2. Immunocytochemistry analysis of cyclin D1 was also performed (Figure 6). Cyclin D1 labelling was very low in serum-starved NIH3T3 cells, and 9 h after serum addition cyclin D1 staining increased, with a clear nuclear localization (80% of the cells). In both serumstarved and serum-stimulated [Lys61]N-Ras NIH3T3 cells, cyclin D1 staining was high and nuclear (90% of the cells), indicating that Ras activation is enough to induce not only an increment in the amount of cyclin D1 but also its translocation to the cell nucleus. To prove that nuclear accumulation of cdk4 in serum-starved [Lys61]N-Ras NIH3T3 cells was due to Ras activation and not to a secondary mutation of this particular cell clone, microinjection experiments were performed with neutralizing anti-Ras antibodies. As shown in Figure 7, cells microinjected with neutralizing anti-Ras antibodies (Y13 ± 259) showed a lower nuclear and a higher cytoplasmic staining for cdk4 compared with non-microinjected cells or with cells microinjected with control antibodies (Rat IgG). To analyse if nuclear accumulation of cdk4 induced by oncogenic Ras was due to the activation of Ras/Raf/MEK pathway, serum-starved [Lys61]N-Ras NIH3T3 cells were treated with PD 98059 (100 mM) for the last 20 h of serum starvation and then intracellular location of cdk4 analysed by immunocytochemistry. As shown in Figure 8a, while in control cells cdk4 was mainly nuclear, in cells treated with MEK inhibitor, cdk4 staining was dispersed through all the cell. Furthermore, transient transfections of [Lys61]N-Ras NIH3T3 cells were done with a plasmid expressing a dominantnegative MEK together with a plasmid expressing green ¯uorescent protein (GFP) in order to visualize transfected cells. As shown in Figure 8b, expression of the dominant negative form of MEK inhibited the nuclear accumulation of cdk4 in serum-starved [Lys61]N-Ras NIH3T3. A control was done transfecting the cells with a plasmid expressing GFP together with the empty vector used for MEK expression. Discussion In response to growth factors, cells are able to activate a variety of intracellular signalling pathways leading to the activation of the cell cycle machinery and progression through G1. Since oncogenic Ras is one of the most frequent alterations harboured by cancer cells, and deregulation of the cell cycle machinery is also a common feature of these cells, it is of major interest to investigate the contribution of Ras to the activation of the cyclin-dependent kinases that govern G1 progression to better understand how Ras drives cell transformation. Here we have identi®ed which elements of the cell cycle regulatory machinery can be activated by an oncogenic form of Ras, [Lys61]N-Ras. For this purpose, NIH3T3 cells constitutively expressing [Lys61]N-Ras were used. We found that expression

Oncogenic Ras and cdks regulation P Villalonga et al

695

Figure 6 Intracellular distribution of cdk4, cdk2 and cyclin D1 in NIH3T3 and [Lys61]N-Ras NIH3T3 cells. NIH3T3 and [Lys61]N-Ras NIH3T3 cells grown in coverslips were serum starved for 2 days (FCS7) and then activated to proliferate by FCS addition for 9 h (FCS+). Immunocytochemistry was performed to localize cdk4, cdk2 and cyclin D1 (Cyc D1)

Figure 7 Microinjection of neutralizing Ras antibody Y13-259 impairs nuclear accumulation of cdk4 in serum-starved [Lys61]NRas NIH3T3 cells. Serum-starved [Lys61]N-Ras NIH3T3 cells were microinjected with control rat IgGs or with neutralizing anti-Ras rat monoclonal antibody Y13 ± 259 (Oncogene Science, Inc., Cambridge, MA, USA). After microinjection, cells were recovered with DMEM 0.5% for 3 h and ®xed for immunocytochemistry to detect cdk4 and rat IgG as described in Materials and methods

of oncogenic Ras, in the absence of other mitogenic stimuli, can induce full activation of cdk4 but not cdk2 and the nuclear accumulation of both proteins. This activation of cdk4 is not sucient to induce pRb hyperphosphorylation and thus its inactivation and the

E2F-dependent transactivation of S phase genes such as cdc2, cyclin A or p107. Consequently, it is unable to induce cell proliferation. It has been reported that both G1 kinases, cdk4 and cdk2, are necessary to inactivate pRb (Lundberg and Weinberg, 1998). Furthermore, the fact that pRb was found in its faster-migrating, hypophosphorylated form, in the presence of active cdk4 is not surprising, as a similar result was reported (Vlach et al., 1996) in an experiment in which cdk2, but not cdk4, was inactivated in p27kip1-overexpressing Rat1 cells. We have analysed the di€erent regulatory elements of cdk4 and cdk2 in an attempt to understand how Ras activates cdk4 and why cdk2 is not active. Cdk4 is already present in serum-starved NIH3T3 cells, and no change in its amount is observed upon serum addition, or in the presence of oncogenic Ras. Consistent with previous reports (Albanese et al., 1995; Liu et al., 1995; Winston et al., 1996), we found that the sole expression of oncogenic Ras is sucient to induce the accumulation of cyclin D1 and also its association with cdk4. The expression levels of other regulators of cdks, such as cdk7 and cyclin H, components of the kinase responsible for cdk2 and cdk4 activation, are not altered by oncogenic Ras, because both are expressed in the absence of serum, irrespective of Ras activation. The e€ect of Ras on the levels and the binding of the two cdk inhibitors, p21cip1 and p27kip1, to cdk4 complexes was analysed. Alterations in p15INK4b and p16INK4a were not considered since NIH3T3 have a deletion of the genes coding for these inhibitors (Quelle et al., 1995). In serum-starved NIH3T3 cells, p21cip1 was poorly expressed, whereas its amount was slightly increased upon serum addition. In contrast, in oncogenic Ras expressing cells, p21cip1 was present at a higher level than in serum-treated NIH3T3. This is Oncogene

Oncogenic Ras and cdks regulation P Villalonga et al

696

Figure 8 MEK activity is essential for the Ras-dependent nuclear accumulation of cdk4 in [Lys61]N-Ras NIH3T3 cells. (a) [Lys61]N-Ras NIH3T3 cells were serum-starved for 2 days in DMEM 0.5% FCS, where indicated 100 mM of the MEK inhibitor PD 98059 during the last 20 h of serum starvation (PD 98059). Immunocytochemistry to detect cdk4 was performed as described in Materials and methods. (b) [Lys61]N-Ras NIH3T3 cells were co-transfected with 3 mg of empty vector (vector) or 3 mg of dominant-negative MEK (dnMEK) together with 3 mg of an expression plasmid for GFP to visualize transfected cells. After 2 days of serum-starvation, cells were ®xed and immunocytochemistry to detect cdk4 performed as described in Materials and methods

consistent with di€erent reports showing that the MAPK pathway drives transcription from the p21cip1 promoter (Liu et al., 1996). Although an increased binding of p21cip1 to cdk4 in cells expressing oncogenic Ras is observed, p21cip1 levels were not high enough to lead to an inhibition in cdk4 activity. We show that, although in NIH3T3 cells p27kip1 total levels decreased upon FCS addition as previously described (Winston et al., 1996), expression of Ras was not sucient to produce a detectable decrease in p27kip1 levels. Furthermore, although FCS addition to NIH3T3 cells induced a decrease in the total amount of p27kip1 and an activation of cdk4, the amount of p27kip1 found in cdk4 immunoprecipitates was not changed. We found that p27kip1 was also associated with cdk4 in cells expressing oncogenic Ras. An explanation for the fact that cdk4 is activated without a concomitant decrease in the association of p27kip1 could be the low availability of p27kip1 to inhibit cyclin D1/cdk4 complexes (Blain et al., 1997), or that cyclin D1 is increased and there are proportionally more cyclin D1/cdk4 complexes free of p27kip1. Our results Oncogene

are consistent with those of other authors (Winston et al., 1996), who demonstrated that conditional expression of [Val12]Ha-Ras in Balb/c-3T3 ®broblasts also failed to induce degradation of p27kip1. In this cell type only a partial activation of cdk4 is observed. Discrepancies about the ability of Ras to induce full cdk4 activation may be due to di€erent levels of expression of p27kip1 or cyclin D1 in the di€erent cell lines. They may also re¯ect signalling di€erences between Ras family members. Since the activating phosphorylation by cyclin H/cdk7 takes place in the nucleus, another regulatory step involved in cdk activation is the nuclear translocation of the kinase. We show here that [Lys61]N-Ras, in the absence of other mitogenic stimuli, is able to induce the translocation of cdk4 and cyclin D1 to the cell nucleus. Although the molecular events regulating cyclin D1/ cdk4 nuclear translocation are not completely understood, di€erent mechanisms have been implicated in this process: ®rst, association of cyclin D1/cdk4 with p21cip1 (or another member of the cip/kip family) (La Baer et al., 1997); second, a Calmodulin-dependent step (Taules et al., 1998); and third, that phosphorylation of cyclin D1 on Thr 286 by GSK3-b results in its nuclear export (Diehl et al., 1998). We show here that Ras is able to supply all the serum dependent signals that are essential for nuclear accumulation of cyclin D1/cdk4 complexes. This is consistent with the above reported data, since Ras induces assembly of cyclin D1/cdk4 with p21cip1 and impairs GSK3-b-mediated nuclear export, because this kinase is inhibited by the Ras/PI3-Kinase/PKB pathway (van-Weeren et al., 1998). In addition, Ras may activate the Ca2+/ Calmodulin dependent step shown to be involved in this process, although this could be a constitutively active mechanism essential for the nuclear import of cyclin D1/cdk4. Whatever the case, Ras is sucient to ful®ll the signalling inputs required to induce nuclear accumulation of cyclin D1/cdk4 and this is shown to be dependent on the MEK pathway. As mentioned, our results also show that expression of oncogenic Ras in the absence of other stimuli does not lead to the activation of cdk2 during G1. Cdk2 total levels are not decreased in serum-starved cells, and are independent of the presence of oncogenic Ras. Cyclin E is present in serum-starved NIH3T3 cells but the levels increase upon FCS addition. The presence of oncogenic Ras does not alter the amount of cyclin E either in absence or in presence of FCS. Since the cyclin E promoter contains E2F binding sites (Botz et al., 1996), the lack of induction of cyclin E expression by oncogenic Ras is consistent with the ®nding that Ras cannot inactivate pRb. Our results show that in serum-starved cells cdk2 is associated with cyclin E and that this association increases 9 h after serum stimulation. In contrast, the amount of p27kip1 associated with cdk2 is high in serum-starved cells and decreases upon serum stimulation, in correlation with the observed activity of this kinase. The presence of oncogenic Ras in serumstarved cells does not alter the amount of cyclin E or p27kip1 associated to cdk2, but it is able to induce translocation of cdk2 to the cell nucleus. We propose that oncogenic Ras is not able to induce cdk2 activation due to the low levels of cyclin E expression and to its incapacity to induce p27kip1 degradation. Thus, although cells expressing oncogenic Ras have active cdk4, they are

Oncogenic Ras and cdks regulation P Villalonga et al

not able to proliferate in the absence of other stimuli due to the lack of cdk2 activity. Since many Ras-transformed cells exhibit a wide range of serum requirements to proliferate, it is tempting to speculate that this is due to the relative levels of cyclin D1, p27kip1 and cyclin E, which would be a result of the strength of Ras signalling itself and the genetic background of each cell line. Whatever the case, we support a view in which Ras profoundly a€ects G1 cell cycle machinery, so upon Ras activation cells become sensitive to other oncogenic alterations that can ®nally lead to cell cycle activation independently of extracellular signals. Since we have shown that [Lys61]N-Ras can lead to cdk4 activation these cooperating alterations would help cdk2 activation. For instance, oncogenic myc expression, which cooperates with Ras to induce cell transformation, has been shown to be involved in p27kip1 inactivation (Leone et al., 1997; Vlach et al., 1996). In cells with both oncogenic Ras and myc, activation of cdk2 could be achieved because the ratio of cyclin E/cdk2 complexes without p27kip1 would increase. Apart from myc, other oncogenic insults such as overexpression of cyclin E, cyclin D1 and cdk4, which have been found in several tumours (Arber et al., 1997; Keyomarsi et al., 1994), could cooperate with Ras to induce cell proliferation. For instance, overexpression of cyclin E would increase the ratio of cdk2 not bound to p27kip1 and consequently, its activation. Overexpression of cyclin D1 or cdk4 would sequester the amount of free p27kip1 allowing cdk2 activation and thus progression to S phase (Cheng et al., 1998). In any case, only a slight activation of cdk2 may lead to cell proliferation in cooperation with Ras because: (i) cyclin E itself is regulated by E2F, so the activation of cdk2 leads to cyclin E accumulation, and (ii) cyclin E/cdk2 has been recently shown to be involved in triggering p27kip1 phosphorylation-dependent degradation (Vlach et al., 1997), both processes resulting in a positive-feedback that would ®nally lead to the full activation of cyclin E/cdk2, allowing the activation of replication origins (Ohtsubo et al., 1995). In conclusion, we show that in NIH3T3 cells, oncogenic [Lys61]N-Ras is able to induce cyclin D1 and p21cip1 expression, cdk4 association to cyclin D1 and p21cip1, and translocation of cdk4 and cyclin D1 to the nucleus. The cyclin D1/cdk4 complexes are active but no hyperphosphorylation of pRb is achieved, most probably due to the lack of cdk2 activation. Cdk2 is not active due to the inability of Ras to induce, without the cooperation of other stimuli, the production of enough cyclin E/cdk2 complexes free of p27kip1. Any oncogene that could increase the ratio of these complexes would cooperate with oncogenic Ras to ®nally induce cell proliferation independently of growth factors. Materials and methods Expression vectors pEGFP-C1, an expression plasmid for green ¯uorescent protein (GFP), was purchased from Clontech. PcDNAIIIMEKA, an expression vector for a dominant negative form of mitogen-activated protein kinase/extracellular-regulated kinase kinase (dnMEK), was kindly provided by Dr P

Crespo. The plasmids were puri®ed by double CsCl density gradient centrifugation.

697

Cell culture and transient transfection NIH3T3 cells (ATCC) or NIH3T3 cells constitutively expressing active mutant [Lys61]N-Ras under a CMV promoter ([Lys61]N-Ras NIH3T3 cells) (gift of Dr T Thompson, Barcelona) were made quiescent by culturing them in Dulbecco's minimum essential medium (DMEM) with 0.5% foetal calf serum (FCS) for 2 days. To allow them to reenter the cell cycle, 10% FCS was added directly to the media and cells were harvested at the time points indicated in the results. Where time points are not indicated, harvested cells were growing asynchronously in 10% FCS-containing medium. PD 98059 (Calbiochem, San Diego, CA, USA) was added directly to the media. For transfections with TfxTM50 (Promega, Madison, WI, USA), 2.16105 cells were plated into 35-mm diameter tissue culture dishes. On the following day, DNA-TfxTM50 complexes (6 mg of DNA and 18 ml of TfxTM50) were added to the cells in 800 ml of serum-free medium. After 1 h, cells were overlayed with 2 ml of DMEM 10% FCS. The day after transfection, the cells were serum deprived for 48 h and ®xed and cdk4 detected as described below (Immunocytochemistry). Transfected (GFP positives) cells were visualized at 509 nm. Microinjection 36105 cells were grown on glass coverslips in 60-mm tissue culture dishes and made quiescent as described above. Coverslips were transferred into DMEM HEPES Modi®cation (Sigma, St Louis, MO, USA) 0.5% FCS fresh medium just before injection. Neutralizing anti-Ras rat monoclonal antibody Y13 ± 259 (Oncogene Science, Inc., Cambridge, MA, USA) at 8 mg/ml in PBS was injected in the cytoplasm. Non speci®c rat immunoglobulin (IgG; ChromePure, Jackson ImmunoResearch, West Grove, Pennsylvania, USA) at the same concentration was injected as a control. Injections were carried out with the Automated Injection System AIS (Zeiss, Oberkochen, Germany), equipped with an Eppendorf 5246 transjector. After injection coverslips were placed into fresh DMEM 0.5% FCS medium and incubated for 3 h before ®xation and immunocytochemical detection of cdk4 as described below (Immunocytochemistry). Microinjected cells were visualized with ¯uorescein-conjugated anti-rat antibody (F-6258, Sigma, dilution 1/100). Immunocytochemistry Quiescent and 10% FCS stimulated cells were grown on glass coverslips. Cells were ®xed in cold methanol for 2 min, and the non-speci®c sites were blocked with 1% ovalbumin/PBS for 10 min at room temperature. In transfection experiments, cells grown in 35-mm diameter tissue culture dishes were washed and ®xed with 4% formalin in PBS for 10 min and treated with 0.1% Triton X-100 for 10 min at room temperature to permeabilize membranes, prior to blocking. Cells were then incubated for 1 h at 378C in a humidi®ed atmosphere, with the speci®c polyclonal antibodies: anti-cdk4 (sc-260-R, Santa Cruz, CA, USA 1 : 100 dilution), anti-cdk2 (sc-748, Santa Cruz, CA, USA 1 : 100 dilution) and anticyclin D1 (No. 06-194, UBI, 1 : 50 dilution) diluted in 1% ovoalbumin/PBS. Coverslips were then washed three times (5 min each) in PBS and incubated for 45 min at 378C with ¯uorescein-conjugated anti-rabbit antibody (F-1262, Sigma, dilution 1 : 100). After two washes in PBS, coverslips were mounted on glass slides with Mowiol (Calbiochem, San Diego, CA, USA). Cells were observed in a confocal microscope. In transfection and microinjection experiments, cyanine 3-conjugated anti-rabbit antibody (111-166-047, Jackson, dilution 1/400) was used to detect cdk4. Oncogene

Oncogenic Ras and cdks regulation P Villalonga et al

698

DNA synthesis assays Quiescent cultures or 10% FCS-supplemented cells were grown on coverslips, Bromo-deoxyUridine (BrdU) (3 mg/ml) was added to the medium for 3 h and cells were ®xed in ethanol/acetic acid (95 : 5) for 30 min. BrdU was detected with monoclonal anti BrdU antibody (Amersham Pharmacia Biotech, Uppsala, Sweden) according to the manfacturer procedures using peroxidase-conjugated anti-mouse antibody (dilution 1 : 50, Bio-Rad) as secondary antibody. Flow cytometry analysis Quiescent or 10% FCS stimulated synchronized cells were harvested at di€erent time points after serum addition. Trypsinized cells were washed in phosphate bu€ered saline (PBS). To determine the content of nuclear DNA, Cycle TESTTM PLUS DNA Reagent Kit (Becton Dickinson, San JoseÂ, CA, USA) was used. Samples were processed according to the manufacturer's instructions and analysed with a FACSCalibur (Becton Dickinson immunocytometry, San JoseÂ, CA, USA). At least 5000 cells were analysed for each sample. Gel electrophoresis and immunoblotting Cells were lysed in a bu€er containing 2% SDS, 67 mM TrisHCl pH 6.8 and 10 mM EDTA, and sonicated twice for 10 s. Protein content was measured by the Lowry procedure, using bovine serum albumin (BSA) as standard. The extracts were electrophoresed in SDS-polyacrylamide gels essentially as described by (Laemmli, 1970). After electrophoresis the proteins were transferred to Immobilon-P strips for 2 h at 60 V, for 20 h at 30 V in the case of pRb analysis, or for 45 min at 60 V in the case of Ras analysis. The sheets were preincubated in TBS (20 mM Tris-HCl pH 7.5, 150 mM NaCl), 0.05% Tween 20 and 5% BSA for 1 h at room temperature and then incubated for 1 h at room temperature in TBS, 0.05% Tween 20, 1% BSA and 0.5% defatted milk powder containing the following antibodies: anti-cdk2 (No. 06-505, UBI, 3 mg/ml), anti-cdc2 (No. 06-194, UBI, 3 mg/ml), anti-cdk4 (sc-260-R, Santa Cruz, CA, USA, 0.5 mg/ml), anticdk7 (sc-857, Santa Cruz, CA, USA, 1 mg/ml), anti-cyclin A (No. 06-138, UBI, 7 mg/ml), anti-cyclin E (sc-198, Santa Cruz, CA, USA, 1 mg/ml), anti-cyclin D1 (No. 06-450, UBI, 2 mg/ml), anti-cyclin D3 (sc-182, Santa Cruz, CA, USA, 1 mg/ ml), anti-cyclin H (sc-855, Santa Cruz, CA, USA, 1 mg/ml), anti-p107 (sc-318, Santa Cruz, CA, USA, 2 mg/ml), antip21cip1 (sc-397, Santa Cruz, CA, USA, 2 mg/ml) and antip27kip1 (1 : 500 dilution; gift from Dr MassagueÂ, Memorial Sloan-Kettering Cancer Center, New York, USA) polyclonal antibodies or anti-PCNA (1 170 406, Boehringer Mannheim, Mannheim, Germany, 10 mg/ml), pRb (14001A, Pharmingen, San Diego, CA, USA, 5 mg/ml) and pan-Ras (Ab-3) (OP-40, Oncogene Sciences Inc., Cambridge, MA, USA, 1 mg/ml monoclonal antibodies. After washing in TBS, 0.05% Tween 20 (three times, 10 min each), the sheets were incubated with either a peroxidase-coupled secondary antibody (1 : 2000 dilution) (Bio-Rad, Hercules, CA, USA) or an alkaline phosphatase-coupled secondary antibody (1 : 10000) (Promega) for 1 h at room temperature. After incubation, the sheets were washed twice in PBS, 0.05% Tween 20 and once in TBS. The reaction was visualized by ECL (Amersham) or with BCIP/NBT (Promega, Madison, USA). Immunoprecipitation and kinase assays To detect proteins associated with cdk4, cyclin D or cdk2, immunoprecipitations were performed as described (Harlow and Lane, 1988). Cells were lysed for 30 min at 48C in IP bu€er (50 mM Tris-HCl pH 7.4, 0.1% Triton X-100, 5 mM EDTA, 250 mM NaCl, 1 mM NaF, 0.1 mM Na3VO4, 1 mM phenylmethyl-sulfonyl ¯uoride, 1 mM dithiothreitol, 10 mM b-glycer-

Oncogene

ophosphate, 2 mg/ml aprotenin, and 10 mg/ml leupeptin). Lysates were sonicated twice for 10 s at 48C and clari®ed by centrifugation at 10 000 g for 10 min. Supernatant fraction protein content was measured using the Bradford method (Bradford, 1976), and 800 mg of protein from the lysates were incubated with 4 mg of anti-cdk4 (sc-260-R, Santa Cruz) or anti-cdk2 (sc-6248, Santa Cruz, CA, USA, monoclonal) antibodies or with 3 ml of normal rabbit serum (controls) or control non-related anti-human mouse antibody for 2 h at 48C. Protein immunocomplexes were then incubated with 20 ml protein A-Sepharose (Pierce) for 1 h at 48C or 20 ml protein GSepharose (Sigma, St Louis, MO, USA) in the case of cdk2, collected by centrifugation and washed four times in IP bu€er. Immunoprecipitated proteins were then analysed by electrophoresis and Western blotting. A lysate from NIH3T3 or [Lys61]N-Ras NIH3T3 cells was always loaded in the same gel as a control for the mobility of each protein. For kinase assays, immunoprecipitations were performed similarly, but 500 mg of protein from the lysates were incubated with 4 mg of anti-cdk4 (sc-260-R, Santa Cruz, CA, USA), 2 mg of anti-cdk2 (No. 06-505, UBI), or 3 ml of NRS and pulled down with 10 ml of protein A-sepharose (Pierce, Rockford, IL, USA). The immunoprecipitated complexes were washed four times in a bu€er containing 50 mM Tris-HCl pH 7.5, 150 mM NaCl, 50 mM NaF, 0.5% NP40, 0.1 mM Na3VO4, 10 mM b-Glycerphosphate and twice in kinase bu€er (HEPES-Na 50 mM pH 7.4 and 1 mM DTT) and then incubated in kinase bu€er containing 1 mCi [g-32P]ATP and 1 mg histone H1 (for cdk2) or 2 mCi [g-32P]ATP and 3 mg pGST-Rb (379-928) (gift of Dr Wang, San Diego, CA, USA) fusion protein (for cdk4 kinase) for 30 min at 308C in a ®nal volume of 30 ml. Then, the samples were boiled for 4 min and electrophoresed on SDS-polyacrylamide gels and the gels were stained with Coomassie blue, dried, and exposed to X-ray ®lms at 7808C. Measurement of Ras activation The capacity of Ras-GTP to bind to RBD (Ras-binding domain of Raf-1) was used in order to analyse the amount of active Ras (de-Rooij and Bos, 1997). Cells were lysed in the culture dish with 25 mM Tris-HCl, pH 7.5, 5 mM EGTA, 150 mM NaCl, 5 mM MgCl2, 1% Triton X-100, 1% N-Octyl glucoside, 1 mM PMSF, 1 mM aprotinin and 20 mM Leupeptin. Cleared (10 000 g) lysate (1 mg) was incubated with 30 mg of GST-RBD bound to glutathione-Sepharose beads for 2 h at 48C. Beads were washed four times in the lysis bu€er. Bound proteins were solubilized by the addition of 30 ml of Laemmli loading bu€er and run on 12.5% SDS ± PAGE gels. Proteins were then transferred and immunoblotted as described above using pan-Ras monoclonal antibody (Oncogene Sciences OP40, 1 : 100 dilution).

Acknowledgments We thank Dr Timothy Thompson (Barcelona) for the gift of the [Lys61]N-Ras NIH3T3 cell line, Dr Joan Massague (New York) for the gift of anti-p27 antibodies and Dr Piero Crespo for the gift of pcDNAIII-MEKA. We are also grateful to Dr FR McKenzie (Nice) for the gift of GST-RBD plasmid and the advice in the Ras activity analysis, and to Dr Albert Tauler and Silvia FernaÂndez (Barcelona) for helping in CsCl plasmid puri®cation. We also thank Anna Bosch (Serveis Cientõ ®co-TeÁcnics, Universitat de Barcelona, Campus Medicina, IDIBAPS) for the technical assistance in confocal microscopy and Dr Pablo Engel and Isabel SaÂnchez (Barcelona) for helping in FACS analysis. This work was supported by CICYT grants SAF95-0041-C02-02 and SAF97-0069. Priam Villalonga is a recipient of a pre-doctoral fellowship from the CIRIT.

Oncogenic Ras and cdks regulation P Villalonga et al

699

References Aktas H, Cai H and Cooper GM. (1997). Mol. Cell Biol., 17, 3850 ± 3857. Albanese C, Johnson J, Watanabe G, Eklund N, Vu D, Arnold A and Pestell RG. (1995). J. Biol. Chem., 270, 23589 ± 23597. Arber N, Doki Y, Han EK, Sgambato A, Zhou P, Kim NH, Delohery T, Klein MG, Holt PR and Weinstein IB. (1997). Cancer Res., 57, 1569 ± 1574. Barbacid M. (1987). Annu. Rev. Biochem., 56, 779 ± 827. Blain SW, Montalvo E and Massague J. (1997). J. Biol. Chem., 272, 25863 ± 25872. Bos JL. (1989). Cancer Res., 49, 4682 ± 4689. Bosch M, Gil J, Bachs O and Agell N. (1998). J. Biol. Chem., 273, 22145 ± 22150. Botz J, Zerfass-Thome K, Spitkovsky D, Delius H, Vogt B, Eilers M, Hatzigeorgion A and Jansen-Durr P. (1996). Mol. Cell. Biol., 16, 3401 ± 3409. Bradford MM. (1976). Anal. Biochem., 72, 248 ± 254. Cheng M, Sexl V, Sherr CJ and Roussel MF. (1998). Proc. Natl. Acad. Sci. USA, 95, 1091 ± 1096. Dalton S. (1992). EMBO J., 11, 1797 ± 1804. Davis RJ. (1993). J. Biol. Chem., 268, 14553 ± 14556. de Rooij J and Bos JL. (1997). Oncogene, 14, 623 ± 625. Diehl JA, Cheng MG, Roussel MF and Sherr CJ. (1998). Genes Devel., 12, 3499 ± 3511. Dobrowolski S, Harter M and Stacey DW. (1994). Mol. Cell Biol., 14, 5441 ± 5449. Dong F, Cress-WD J, Agrawal D and Pledger WJ. (1998). J. Biol. Chem., 273, 6190 ± 6195. Dou Q, Zhao S, Levin A, Wang J, Helin K and Pardee A. (1994). J. Biol. Chem., 269, 1306 ± 1313. Downward J. (1997). Curr. Biol., 7, R258 ± R260. Downward J. (1998). Curr. Opin. Genet. Dev., 8, 49 ± 54. Draetta G. (1994). Curr. Opin. Cell Biol., 6, 842 ± 846. Feramisco JR, Gross M, Kamata T, Rosenberg M and Sweet RW. (1984). Cell, 38, 109 ± 117. Geng Y, Eaton EN, Picon M, Roberts JM, Sardet C and Weinberg RA. (1996). Oncogene, 12, 1173 ± 1180. Grana X and Reddy EP. (1995). Oncogene, 11, 211 ± 219. Harlow E and Lane D. (1988). Cold Spring Laboratory Press. Harper JW, Adami G, Wei N, Keyomarsi K and Elledge SJ. (1993). Cell, 75, 805 ± 816. Jin P, Hardy S and Morgan DO. (1998). J. Cell Biol., 141, 875 ± 885. Jinno S, Suto K, Nagata A, Igarashi M, Kanaoka Y, Nojima H and Okayama H. (1994). EMBO J., 13, 1549 ± 1556. Kato J, Matsuoka M, Polyak K, Massague J and Sherr C. (1994b). Cell, 79, 487 ± 496. Kato J, Matsuoka M, Strom DK and Sherr CJ. (1994a). Mol. Cell. Biol., 14, 2713 ± 2721. Katz ME and McCormick F. (1997). Curr. Opin. Genet. Dev., 7, 75 ± 79. Kerkho€ E and Rapp UR. (1998). Oncogene, 17, 1457 ± 1462. Keyomarsi K, O'Leary N, Molnar G, Lees E, Fingert HJ and Pardee AB. (1994). Cancer Res., 54, 380 ± 385. La Baer J, Garrett MD, Stevenson LF, Slingerland JM, Sandhu CH, Chou HS, Fattaey A and Harlow E. (1997). Genes Devel., 11, 847 ± 862. Laemmli UK. (1970). Nature, 227, 680 ± 685. Lam E and La Thangue N. (1994). Curr. Opin. Cell Biol., 6, 859 ± 866. Lavoie JN, Lallemain G, Brunet A, Muller R and Pouyssegur J. (1996). J. Biol. Chem., 271, 20608 ± 20616. Leone G, DeGregori J, Sears R, Jakoi L and Nevins JR. (1997). Nature, 387, 422 ± 426.

Lewis TS, Shapiro PS and Ahn NG. (1998). Adv. Cancer Res., 74, 49 ± 139. Liu JJ, Chao JR, Jiang MC, Ng SY, Yen JJ and Yang YH. (1995). Mol. Cell Biol., 15, 3654 ± 3663. Liu Y, Martindale JL, Gorospe M and Holbrook NJ. (1996). Cancer Res., 56, 31 ± 35. Lloyd AC. (1998). Curr. Opin. Genet. Dev., 8, 43 ± 48. Lundberg AS and Weinberg RA. (1998). Mol. Cell Biol., 18, 753 ± 761. Marshall CJ. (1996). Curr. Opin. Cell Biol., 8, 197 ± 204. McCormick F. (1994). Curr. Opin. Genet. Dev., 4, 71 ± 76. Mittnacht S. (1998). Curr. Opin. Genet. Deve., 8, 21 ± 27. Mittnacht S, Paterson H, Olson MF and Marshall CJ. (1997). Curr. Biol., 7, 219 ± 221. Morgan DO. (1997). Ann. Rev. Cell Dev. Biol., 13, 261 ± 291. Morgan D. (1995). Nature, 374, 131 ± 134. Mulcahy LS, Smith MR and Stacey DW. (1985). Nature, 313, 241 ± 243. Norbury C and Nurse P. (1992). Annu. Rev. Biochem., 61, 441 ± 470. Ohtsubo M, Theodoras AM, Schumacher J, Roberts JM and Pagano M. (1995). Mol. Cell Biol., 15, 2612 ± 2624. Olson MF, Paterson HF and Marshall CJ. (1998). Nature, 394, 295 ± 299. Pardee AB. (1989). Science, 246, 447 ± 458. Parry SB, Bates S, Mann DJ and Peters G. (1995). EMBO J., 14, 503 ± 511. Peeper DS and Bernards R. (1997). FEBS Lett., 410, 11 ± 16. Peeper DS, Upton TM, Ladha MH, Neuman E, Zalvide J, Bernards R, DeCaprio JA and Ewen ME. (1997). Nature, 386, 177 ± 181. Polyak K, Lee M-H, Erdjument-Bromage H, Tempst P and Massague J. (1994). Cell, 78, 59 ± 66. Quelle DE, Ashmun RA, Hannon GJ, Rehberger PA, Trono D, Richter KH, Walker C, Beach D, Sherr CJ and Serrano M. (1995). Oncogene, 11, 635 ± 645. Reed SI. (1992). Annu. Rev. Cell Biol., 8, 529 ± 561. Resnitzky D and Reed DS. (1995). Mol. Cell. Biol., 15, 3463 ± 3469. Robinson MJ and Cobb MH. (1997). Curr. Opin. Cell Biol., 9, 180 ± 186. Rodriguez-Viciana P, Warne PH, Dhand R, Vanhaesebroeck B, Gout I, Fry MJ, Water®eld MD and Downward J. (1994). Nature, 370, 527 ± 532. Roussel MF. (1998). Adv. Cancer Res., 74, 1 ± 24. Sherr CJ. (1994). Cell, 79, 551 ± 555. Sherr CJ and Roberts J. (1995). Genes Devel., 9, 1149 ± 1163. Takuwa N and Takuwa Y. (1997). Mol. Cell Biol., 17, 5348 ± 5358. Taules M, Rius E, Talaya D, Lopez GA, Bachs O and Agell N. (1998). J. Biol. Chem., 273, 33279 ± 33286. van-Weeren PC, de-Bruyn KM, de-Vries-Smits AM, vanLint J and Burgering BM. (1998). J. Biol. Chem., 273, 13150 ± 13156. Vlach J, Hennecke S, Alevizopoulos K, Conti D and Amati B. (1996). EMBO J., 15, 6595 ± 6604. Vlach J, Hennecke S and Amati B. (1997). EMBO J., 16, 5334 ± 5344. Weber JD, Hu W, Jefcoat-SCJ, Raben DM and Baldassare JJ. (1997). J. Biol. Chem., 272, 32966 ± 32971. Winston JT, Coats SR, Wang YZ and Pledger WJ. (1996). Oncogene, 12, 127 ± 134. Zwicker J, Lucibello F, Wolfraim L, Gross C, Truss M, England K and Muller K. (1995). EMBO J., 14, 4514 ± 4522.

Oncogene