A-Raf and Raf-1 Work Together to Influence Transient ERK Phosphorylation and Gl/S Cell Cycle Progression

A-Raf and Raf-1 work together to inﬂuence transient ERK phosphorylation and Gl/S cell cycle progression

Kathryn Mercer1, Susan Giblett1, Anthony Oakden2, Jane Brown2, Richard Marais3 and Catrin Pritchard*,1

1Department of Biochemistry, University of Leicester, Adrian Building, University Road, Leicester LEI 7RH, UK; 2Division of Biomedical Services, University of Leicester, University Road, Leicester LEI 7RH, UK; 3Institute of Cancer Research, 237 Fulham Road, London SW3 6JB, UK

The Raf/MEK/ERK (extracellular regulated kinase) Kerkhoff and Rapp, 1997, 1998; Woods et al., 1997; signal transduction pathway controls the ability of cells Marshall, 1999; Squires et al., 2002). This function to respond to proliferative, apoptotic, migratory and is mediated by the ability of ERKs to translocate to the differentiation signals. We have investigated the combined nucleus where they are able to phosphorylate transcrip- contribution of A-Raf and Raf-1 isotypes to signalling tion factors, particularly the AP-1 complex that through this pathway by generating mice with knockout comprises various heterodimers of c-fos (c-fos, FosB, mutations of both A-raf and raf-1 genes. Double knockout Fra-1, Fra-2) and c-jun (c-Jun, JunB, JunD) family (DKO) mice have a more severe phenotype than single null members (Whitmarsh and Davis, 1996; Balmanno mutations of either gene, dying in embryogenesis at E10.5. and Cook, 1999; Cook et al., 1999; Shaulian and The DKO embryos show no changes in apoptosis, but Karin, 2002). This transcription factor complex is staining for Ki67 indicates a generalized reduction in involved in promoting D-type cyclin expression, thus proliferation. DKO mouse embryonic fibroblasts (MEFs) allowing the cyclin-dependent kinases CDK4 exhibit a delayed ability to enter S phase of the cell cycle. and CDK6 to phosphorylate and de-repress the This is associated with a reduction in levels of transiently retinoblastoma family members (Brown et al., 1998; induced MEK and ERK phosphorylation and reduced Kerkhoff and Rapp, 1998; Marshall, 1999; Shaulian and expression of c-Fos and cyclin Dl. Levels of sustained Karin, 2002). The consequent release of the E2F ERK phosphorylation are not significantly altered. Thus, transcription factors from retinoblastoma binding al- Raf-1 and A-Raf have a combined role in controlling lows expression of genes required for progression physiological transient ERK activation and in mainte- through the G0–G1–S transitions of the cell cycle nance of cell cycle progression at its usual rate. (Stevaux and Dyson, 2002). Oncogene (2005) 24, 5207–5217. doi:10.1038/sj.onc.1208707; There are three Raf family members in mammals; published online 25 April 2005 A-Raf, Raf-1 and B-Raf (Mercer and Pritchard, 2003). All three are able to induce MEK/ERK activation Keywords: A-Raf; Raf-1; knockout; ERK activation; (Pritchard et al., 1995), but they differ in their strength cell cycle; c-Fos; cyclin Dl of activation of ERK. In kinase cascade assays, immunoprecipitated B-Raf has a far stronger ability to activate MEK/ERK than Raf-1 and Raf-1, in turn, is far stronger than A-Raf (Huser et al., 2001). In B-rafÀ/À Introduction mouse embryonic fibroblasts (MEFs), ERK phosphorylation and activation are reduced to E30% that in Extracellular regulated kinases 1 and 2 (ERK1/2) are wild-type cells (Wojnowski et al., 2000; Pritchard et al., highly conserved cellular enzymes that play crucial roles 2004) whereas there is no noticeable decrease in ERK in the control of multiple cellular processes including cell phosphorylation or activation in raf-1À/À MEFs or proliferation (Lewis et al., 1998; Chen et al., 2001). A-rafÀ/Y MEFs (Huser et al., 2001; Mikula et al., 2001; Treatment of mammalian cells with mitogens leads to an Mercer et al., 2002). The ability of B-Raf to induce increase in intracellular levels of Ras.GTPfollowed by strong ERK activity has been highlighted by the the sequential activation of Raf, MEK1/2 and ulti- discovery of activating mutations of the BRAF gene in mately ERK1/2 (Marais and Marshall, 1996; Lewis human cancers (Davies et al., 2002; Kimura et al., 2003; et al., 1998). It has been shown that activation of ERK Mercer and Pritchard, 2003). The most common in this way is an important step in G0–Gl–S phase mutation, a valine to glutamic acid change at residue progression of the cell cycle (Samuels et al., 1993; 599, is thought to contribute to tumorigenesis by stimulating constitutive ERK phosphorylation. A simi- *Correspondence: C Pritchard; E-mail: [email protected] lar mutation in RAF1 does not induce such high levels of Received 11 October 2004; revised 25 January 2005; accepted 16 March ERK activation and is not detected in human cancers 2005; published online 25 April 2005 (Davies et al., 2002). Influence of A-Raf and Raf-1 on ERK phosphorylation and Gl/S cell cycle progression K Mercer et al 5208 Table 1 Genotyping data of embryos obtained from A-rafÀ/Àraf-1+/À Â A-raf/Y raf-1+/À intercrosses Genotype Age P1 Total Expected

E9.5E10.5Ell.5E12.5E13.5E14.5

A-raf+/+raf-1+/+ a 3235 1 562522 A-raf+/Àraf-1+/+ 6576 1 383622 A-rafÀ/Àraf-1+/+ a 7 6 13 9 5 2 6 48 44 A-raf+/+raf-1+/À a 5 2 20 7 1 3 12 50 44 A-raf+/Àraf-1+/À 6 6 11 16 6 1 17 63 44 A-rafÀ/Àraf-1+/À a 18 5 26 22 5 1 9 86 88 A-raf+/+raf-1À/À a 1044 1 011122 A-raf+/Àraf-1À/À 4235 0 001422 A-rafÀ/Àraf-1À/À a 5543 1 001844

aMale and female embryos indistinguishable

The duration and magnitude of ERK activation has a Results significant impact on whether cells enter the cell cycle or not (Marshall, 1995). In fibroblasts, transient and Generation of DKO mice sustained ERK activation are observed following mitogen treatment (Vouret-Craviari et al., 1993; Cook and To generate the DKO mice, two rounds of breeding À/À À/À McCormick, 1996; Weber et al., 1997; Balmanno and were set up with A-raf and raf-1 single mutant þ /À Cook, 1999). The sustained phase of ERK activation mice. In the first round, A-raf female mice were þ /À correlates with the induction of expression of a subset of crossed to raf-1 male mice. Wild-type (WT), single and double heterozygote mice were obtained at the AP-1 components (Fra-1, Fra-2, c-Jun and JunB), which /Y in turn are required for sustained cyclin Dl expression expected Mendelian frequency, while A-raf and A-raf IY þ lÀ and progression through the cell cycle (Kovary and raf-1 adult mice were obtained at a reduced Bravo, 1991, 1992; Lavoie et al., 1996; Balmanno and frequency and displayed a phenotype typical of that Cook, 1999; Cook et al., 1999). The transient activation previously described (data not shown; Pritchard et al., þ /À þ /À of ERK correlates with transient c-Fos expression and is 1996). In the second round of breeding, A-raf raf-1 /Y þ /À not associated with the induction of cyclin Dl expression female animals were crossed to A-raf raf-1 male or cell cycle progression in several cell systems (Vouret- mice. No DKO animals survived postnatally and so Craviari et al., 1993; Cook and McCormick, 1996; embryos were harvested from matings timed at between Weber et al., 1997). However, other studies have E9.5 to birth (Table 1). Genotypes of each embryo were indicated an important role for c-Fos in controlling confirmed by PCR analysis of yolk sac DNA (Figure 1a) cyclin Dl expression and cell cycle progression (Kovary and Western blot analysis was used to confirm the Raf-1 À/À and Bravo, 1991; Won et al., 1992; Miao and Curran, and/or A-Raf protein deficiencies (Figure 1b). A-raf 1994; Brown et al., 1998). In particular, mice with embryos were obtained at the expected frequency, knockout mutations of both the c-Fos and FosB genes indicating that loss of A-Raf alone has no detrimental À/À have defects in proliferation that result at least in part effect on embryogenesis (Table 1). Raf-1 embryos from a failure to induce cyclin Dl expression (Brown were obtained at the expected frequencies up to E13.5 et al., 1998). but lethality was observed from this point onwards and In this report, we have used gene targeting in mice very few survived to birth (o3%). DKO embryos were to investigate the roles of A-Raf and Raf-1 in regulating obtained at the expected frequency at E9.5 but they MEK/ERK activation and downstream cellular re- appeared at a reduced frequency at E10.5 onwards and sponses. Despite the fact that single knockout mutations none survived to birth (Table 1). The expression of either gene do not lead to noticeable changes (Figure 1c) and activity (Figure 1d) of the remaining in ERKl/2 activation (Huser et al., 2001; Mikula et al., Raf isotype, B-Raf, were not noticeably altered in DKO 2001; Mercer et al., 2002), the combined knockout embryos compared to WT embryos. mutation of both genes leads to reduced ERKl/2 phosphorylation. These results indicate that A-Raf Phenotype analysis of single knockout and DKO mice and Raf-1 play compensatory roles in ERKl/2 activation that cannot be rescued by B-Raf However, Embryos were harvested at E10.5. and photographed only transient ERK phosphorylation is significantly (Figure 2a–d). A-rafÀ/À embryos exhibited no abnorm- reduced in the double knockout (DKO) MEFs, whereas alities and were identical in size and developmental age the sustained phase of ERK phosphorylation is to WT embryos (Figure 2a, b). Raf-1À/À embryos not significantly affected. This reduction in appeared relatively normal although they were consis- transient ERK phosphorylation is associated with a tently small in size and were developmentally retarded reduction in transient c-Fos expression and cyclin Dl by approximately 0.5 days of gestation (Figure 2c). By expression, and delayed progression through the cell contrast, the DKO embryos were grossly abnormal cycle. (Figure 2d). They were extremely small, demonstrated

Oncogene Inﬂuence of A-Raf and Raf-1 on ERK phosphorylation and Gl/S cell cycle progression K Mercer et al 5209 a * antibody for the S phase marker Ki67 (Figure 2m–t).

1 2 3 4 5 6 7 Consistent with previous observations, the A-rafÀ/À

→ 450 bp embryos showed no differences in levels of apoptosis raf-1 allele → 355 bp or proliferation compared to WT embryos (compare

Figure 2j, n, r with Figure 2i, m, q). The raf-1À/À WT A-raf allele → 230 bp embryos also showed no increase in apoptosis

(Figure 2k) or changes in proliferation (Figure 2o, s). → The DKO embryos showed no increase in TUNEL Mutant 342 bp A-raf allele staining (Figure 2l). However, the number of cells staining with Ki67 was noticeably reduced in these embryos (Figure 2p, t), indicating a generalized reduc- b **

1 2 3 4 5 6 7 tion in the percentage of cells in S phase. → Raf-1

Growth, apoptosis and proliferation analysis of single

→ A-Raf knockout and DKO MEFs → actin Primary MEFs were derived from the DKO, A-rafÀ/À, raf-1À/À and sibling WT embryos by standard procedures. Growth rates were determined by counting cells c

1 2 3 4 in triplicate over 9 days in culture following immediate → B-Raf isolation from the embryo (Figure 3a). A-rafÀ/À MEFs

had growth profiles similar to that of the control MEFs. → actin By day 7, the growth of raf-1À/À MEFs was slightly reduced compared to the WT cells. However, the growth d 1.6 of the DKO MEFs was significantly compromised 1.4 compared to MEFs of all other genotypes (Figure 3a). Apoptosis was induced by treatment of the primary 1.2 MEFs with a-CD95 antibody and cell death was 1.0 assessed by annexin V staining (Huser et al., 2001; 0.8 Mercer et al., 2002). Treatment of cells with a-CD95 0.6 WT antibody led to a small but significant increase in DKO 0.4 apoptosis in all samples. No significant difference was À/À

B-Raf activation (fold) 0.2 observed in the level of apoptosis induced in the A-raf cells compared to WT cells, while the raf-1À/À and DKO 0 0 10 20 30 40 50 60 cells appeared to be more protected from a-CD95 Time + EGF antibody-induced apoptosis than WT cells (Figure 3c). Figure 1 PCR and protein analysis of embryos. (a)PCR These results show that increases in apoptosis cannot genotyping of embryos. Results are shown for a typical litter account for the greatly reduced growth of the DKO cells arising from an A-raf þ /Àraf-1 þ /À intercross. PCR genotyping or the slightly reduced growth of the raf-1À/À cells results are shown for the WT and mutant raf-1 allele (upper (Figure 3a). panel), WT A-raf allele (middle panel) and mutant A-raf allele FACS analysis of propidium iodide-stained cells (lower panel). DKO embryo is indicated by * in lane 7. (b) À/À À/À Expression of A-Raf and Raf-1. a-Raf-1 antibody (upper panel) showed that A-raf and raf-1 MEFs had a similar and an a-A-Raf antibody (middle panel) were used to detect the proportion of cells in each phase of the cell cycle presence/absence of Raf-1 and A-Raf protein in each embryo. As a compared to control cells (Figure 3c). Our recent control for protein loading, the blots were analysed with an analysis has shown that the raf-1À/À cells consistently antibody for actin (lower panel). DKO embryos are indicated by * in lanes 2 and 7. (c) Expression of B-Raf. An a-B-Raf antibody grow to a lower saturation density (KM, unpublished (upper panel) was used to detect the expression level of B-Raf in data) that may account for their slightly reduced ability each embryo. As a control for protein loading, the blots were to grow (Figure 3a). However, the DKO cells had a analysed with an antibody for actin (lower panel). Lanes: 1 – A-raf significantly reduced percentage of cells in S phase KO, 2 – A-raf/raf-1 double heterozygote, 3 – DKO, 4 – raf-1 KO. (Figure 3c); 5.7% of the DKO MEFs compared with (d) Activity of B-Raf. DKO and WT MEFs were stimulated with EGF over a time course of 0–60 min. B-Raf activity in each sample 8.3% of the WT MEFs were in S phase (n ¼ 8; 95% CI was determined by performing the immunoprecipitation kinase for difference 0.89% to 4.48%, P ¼ 0.006). Bromodeox- cascade assay yuridine (BrdU) incorporation assays showed that, following serum stimulation for 8 h, only 12% of DKO cells had entered S phase compared to 23% of tail truncations and the vast majority died before they WT cells (Figure 3d). However, by 16 h, the proportion reached E10.5. Embryos at E10.5 were fixed, embedded of cells in S phase was approximately the same for the in paraffin and sectioned for histological analysis. DKO (38%) and WT (42%) cells. These results suggest Sections were stained either with haematoxylin and that the reduction in growth of the DKO cells is eosin (Figure 2e–h), subjected to TUNEL analysis to associated with a delayed ability to pass through the detect apoptotic cells (Figure 2i–l) or stained with an G0–G1–S transitions of the cell cycle.

Oncogene Inﬂuence of A-Raf and Raf-1 on ERK phosphorylation and Gl/S cell cycle progression K Mercer et al 5210 WT A-raf -/- raf-1-/- DKO a b c d

e f g h

i jkl

m nop

q r s t

Figure 2 Phenotype analysis of embryos. WT embryos are shown in (a,e,i,m,q); A-rafÀ/À embryos are shown in (b,f,j,n,r); raf-1À/À embryos are shown in (c,g,k,o,s); DKO embryos are shown in (d,h,l,p,t). In (a–d), embryos were extracted from yolk sacs and photographed. In (e–t), embryos were ﬁxed, sectioned and stained with haematoxylin and eosin (e–h), subjected to TUNEL analysis (i– l) or stained with an antibody for Ki67 (m–t). q–t shows higher magniﬁcation of Ki67 staining of the somites and adjacent mesenchymal tissue. Scale bars: a–d, 200 mm. q–t,25mm

Assessment of MEK/ERK phosphorylation in DKO phoERKl/2. In the A-rafÀ/À, raf-1À/À and WT control MEFs MEFs, the levels of phosphoMEK increased following 2 min of serum treatment and continued to increase up MEFs of each genotype were serum starved and treated to 10 min of treatment but started to decline after this with serum for various lengths of time, protein lysates point (Figure 4a). For the DKO MEFs, the levels of were prepared and Western blots were incubated with phosphoMEK increased slightly following 2–10 min of antibodies speciﬁc for phosphoMEKl/2 or phos- stimulation but they were signiﬁcantly reduced in

Oncogene Inﬂuence of A-Raf and Raf-1 on ERK phosphorylation and Gl/S cell cycle progression K Mercer et al 5211 a 9 30 b WT -/- ) 8 WT

4 A-raf -/- 25 7 A-raf DKO DKO raf-1-/- 6 raf-1-/- 20 5 15 4 10 3 nnexin positive %a

Number of cells (x10 2 5 1 0 0 Untreated + α-CD95 antibody 2345678 Days

c 80 14 35 70 12 30 60 10 25 50 40 8 20 30 6 15 20 4 10 % of cells in S

% of cells in G1 10 2 5 0 0 % of cells in G2/M 0 – – – – – – –/ –/ –/ –/ –/ –/ WT DKO WT DKO WT DKO A-raf raf-1 A-raf raf-1 A-raf raf-1

d 70 DKO 60 WT 50 40 30 20 10 % of BrdU positive cells % of BrdU positive 0 2h 4h 8h 16h Hours + serum Figure 3 Growth, apoptosis and proliferation analysis of primary MEFs. (a) Growth analysis of DKO MEFs (closed circles) compared to raf-1À/À (closed squares), A-rafÀ/À (open squares) and WT MEFs (open circles) over 8 days in culture. (b) Apoptosis analysis of primary MEFs. Cells were either untreated or treated with a-CD95 antibody for 20 h. The percentage of cells undergoing apoptosis was quantified by FACS analysis of annexin V staining. Each experiment was performed three times and the data show mean values7standard deviation. (c) Analysis of progression through the cell cycle. Propidium iodide staining followed by FACS analysis was performed to assess DNA content of asynchronously growing MEFs. The percentages of cells in the Gl (left panel), S (middle panel) and G2/M (right panel) phases of the cell cycle were determined. Each experiment was performed eight times and the data show mean values7standard deviation. (d) Analysis of entry into S phase. BrdU proliferation assays were performed to compare the percentage of DKO and WT cells entering S phase over a time course of stimulation with serum from 2 to 16 h comparison to control MEFs (Figure 4a). Similar Throughout these experiments, some variations were profiles of phosphorylation were observed for phos- observed in the profiles of ERK phosphorylation phoERK (Figure 4b). Levels of phosphoERK in cells between cells of the same genotype. For this reason, stimulated with serum for 2–5 min were significantly the experiments were repeated multiple times. The levels lower in the DKO cells compared to the control cells of phosphoERK induced at each time point in each (Figure 4b). At later time points from 10 min onwards, experiment were quantified using NIH Image software, the levels of phosphoERK reached similar levels to pooled and analysed using a paired t-test (Figure 4f). those in control MEFs. The profiles of phosphoERK The results of this analysis confirm that at 2 min (n ¼ 6; stimulation over a time course of EGF treatment were P ¼ 0.015) and at 5 min (n ¼ 6; P ¼ 0.043) of serum similar to that observed with serum (Figure 4c). The treatment the levels of phosphoERK are significantly sustained activation of phosphoERK following serum reduced in the DKO cells compared to WT cells. stimulation was also assessed (Figure 4d). At time points However, at all other time points, although there is a over 1 h of stimulation, there was no significant trend towards a lower level of phosphoERK in the difference in the levels of phosphoERK observed in DKO samples, this is not statistically significant at any DKO cells compared to WT cells (Figure 4d). Total time point. protein lysates were also generated from DKO embryos as well as from littermate control embryos. The total Analysis of expression of c-Fos in DKO MEFs level of phosphoERK in the DKO embryo lysates was reduced compared to that in control embryo lysates The AP1 transcription factor complex is an important (Figure 4e). element in mediating the effects of the ERK pathway on

Oncogene Inﬂuence of A-Raf and Raf-1 on ERK phosphorylation and Gl/S cell cycle progression K Mercer et al 5212

a A-raf -/- DKO raf-1 -/- WT 0 2 5 10 30 60 3h 0 2 5 10 30 60 3h 0 2 5 10 30 60 3h 0 2 5 10 30 60 3h Min/hr + serum PP-MEK

actin

b A-raf -/- DKO raf-1 -/- WT 0 2 5 10 20 40 60 0 2 5 10 20 40 60 0 2 5 10 20 40 60 0 2 5 10 20 40 60 Min + serum

PP-ERK

Total ERK

c A-raf -/- DKO raf-1 -/- WT 0 2 5 10 20 40 60 0 2 5 10 20 40 60 0 2 5 10 20 40 60 0 2 5 10 20 40 60 Min + EGF PP-ERK1/2

Total ERK

- d e +/+ +/ DKO WT - raf-1 - raf-1 0 10 20 1 4 8 20 0 10 20 1 4 8 20 Min/Hr + serum -/- +/ +/

PP-ERK1/2 A-raf A-raf A-raf DKO WT PP-ERK1/2 Total ERK

actin

f 2 1.8 WT DKO 1.6 p = 0.043 1.4 p = 0.015 1.2 1 0.8

phosphorylation 0.6

Fold changes in ERK1/2 0.4 0.2 0 025102030604h8h Time + serum (Min/hr) Figure 4 MEK and ERK phosphorylation (a) MEK phosphorylation following serum stimulation. Primary MEFs of each genotype were treated with serum over the indicated time course and protein cell lysates were harvested. Western blots were prepared and analysed with an antibody against phosphoMEK (upper panels) and actin (lower panels). (b–d) ERK phosphorylation following serum (b, d) and EGF (c) stimulation. Western blots were prepared and analysed with an antibody against phosphoERKl/2 (upper panels) and total ERK (lower panels). (e) ERK phosphorylation in embryo lysates. Protein lysates were harvested from individual embryos of the genotypes indicated and Western blots were analysed with an antibody against phosphoERKl/2 (upper panel) or against actin (lower panel). (f) Quantitation of ERKl/2 phosphorylation following serum stimulation. Quantitation was achieved by scanning Western blots from multiple experiments using NIH Image Software. Fold changes in ERKl/2 phosphorylation over basal in WT samples were calculated and pooled at each time point. Each experiment was performed six times and the data show mean values7standard error

cell proliferation (Herber et al., 1994; Albanese et al., ERK activation (Balmanno and Cook, 1999; Cook 1995; Shaulian and Karin, 2002). Of the AP1 compo- et al., 1999). Therefore, we examined whether the nents, only the induction of c-Fos expression has been expression of c-Fos was disrupted in the DKO cells speciﬁcally associated with the transient stimulation of (Figure 5a). In WT cells, c-Fos expression was induced

Oncogene Inﬂuence of A-Raf and Raf-1 on ERK phosphorylation and Gl/S cell cycle progression K Mercer et al 5213 a DKO WT reduced in cycling DKO cells and in whole embryo 0 0.2 0.5 0.75 1 4 8 0 0.2 0.5 0.75 1 4 8 Hrs + serum lysates (Figure 5b and c). Levels of Cdk4 were slightly c-Fos reduced in the DKO MEFs, whereas levels of cyclin D3 expression were unchanged (Figure 5b). Levels of cyclin Dl were assessed over a time course of stimulation with serum (Figure 5d). In WT cells, serum stimulation gave rise to an induction of expression of cyclin Dl at 1 h and

-/- b c -/- -/- this continued to rise up to 20 h. In the DKO cells, there

WT A-raf DKO DKO A-raf raf-1 WT was no detectable cyclin Dl expression until after 4 h of cyclinD1→ serum stimulation. Therefore, consistent with the pro-

+/ → liferation data, induction of cyclin Dl expression is - raf-1 CDK4

-/- +/ delayed in the DKO cells compared to the WT cells DKO →

A-raf WT A-raf actin → cyclin D1 Rescue of DKO phenotype by overexpression of Raf-1 -/- -/- -/-

-1 → actin DKO A-raf raf A-raf

WT We transfected the DKO cells with vectors expressing → cyclinD3 Raf isotypes in an attempt to rescue the phenotype.

Coexpression of A-Raf and Raf-1 together proved → actin technically difﬁcult. Therefore, we overexpressed Raf-1 alone. DKO cells were transfected with vectors expres- DKO WT d Hr + sing either myc-tagged human Raf-1 or with a control

0 0.5 0.75 1 4 8 20 0 0.5 0.75 1 4 8 20 serum vector expressing GFPusing the Amaxa ‘Nucleofector’. → cyclinD1 Transfection of the DKO cells with the Raf-1 vector

clearly led to the overexpression of Raf-1 (Figure 6a). → actin This was associated with a significant increase in the levels of phospho ERK (Figure 6a) and cyclin Dl Figure 5 Analysis of the expression of cell cycle control proteins. (Figure 6b) as well rescue of the percentage of cells (a) Expression of c-Fos in MEFs. WT and DKO MEFs were serum entering S phase (Figure 6c). starved for 20 h and then stimulated with serum for up to 8 h. Cell lysates were harvested, Western blots were prepared and analysed with an antibody against c-Fos (upper panel). A nonspecific band detected by the c-Fos antibody that remains unchanged during the Discussion time course provided a control for protein loading (lower panel). (b) Expression of cyclin Dl in embryos. Cell lysates were harvested Mice containing knockout mutations of the raf-1 and A- from E10.5 embryos and Western blots were analysed with an raf genes have been previously documented (Pritchard antibody against cyclin Dl. An antibody against actin was used to confirm protein loading. (c) Expression of cyclin Dl, cdk4 and et al., 1996; Huser et al., 2001; Mercer et al., 2002). In cyclin D3 in primary MEFs. Cell lysates were harvested and contrast to the present study, the analysis of these single Western blots were analysed with antibodies against cyclin Dl, mutant mice provided no evidence for a role of A-Raf cyclin D3 and Cdk4. An antibody against actin was used to confirm and Raf-1 in ERK1/2 activation or cell proliferation, protein loading. (d) Expression of cyclin Dl in MEFs over a time course of stimulation with serum. WT and DKO MEFs were serum but suggested an MEK kinase-independent role of Raf-1 starved for 20 h and then stimulated with serum for 0–20 h as in the prevention of apoptosis (Huser et al., 2001; indicated. Cell lysates were harvested, Western blots were prepared Jesenberger et al., 2001; Mikula et al., 2001; Hindley and and analysed with an antibody against cyclin Dl (upper panel). An Kolch, 2002). In this report, we intercrossed the A-raf/Y antibody against actin was used to confirm protein loading (lower and raf-T’ mice on the MF1 background to generate panel) mice carrying knockout mutations of both genes and following 45 min of treatment with serum and phos- show that A-Raf and Raf-1 work together to influence phorylated forms of c-Fos were observed at later time ERK1/2 activation, c-Fos expression, cyclin Dl expres- points, particularly after 4 h. By contrast, the level of sion and G0–G1–S cell cycle progression. Thus, the role expression of c-Fos in the DKO cells was reduced at of Raf-1 as an MEK kinase involved in proliferation is 45 min of serum treatment compared to the levels in WT dependent on A-Raf. Our results show that A-Raf and cells. Although phosphorylated c-Fos isoforms were Raf-1 compensate for one another in the single mutant observed at the later time points in the DKO cells, their mice and that B-Raf cannot compensate for the level of expression was reduced (Figure 5a). MEFs with combined loss of A-Raf and Raf-1. These data are a single knockout mutation of Raf-1 or A-Raf did not consistent with recent results from one of our labora- have noticeable changes in c-Fos expression or phos- tories showing that the downregulation of A-Raf and phorylation (data not shown). Raf-1 by siRNA in B-Raf-transformed melanoma cells is associated with a reduction in DNA synthesis in both Analysis of expression of cell cycle proteins in DKO cases (Karasarides et al., 2004). The downregulation of MEFs B-Raf in these cells by siRNA also led to a reduction in DNA synthesis. Therefore, both A-Raf/Raf-1 and B- Consistent with the reduction in G0–G1–S cell cycle Raf are involved in activating pools of MEK/ERK, progression, levels of expression of cyclin Dl were both of which have an important influence on cell

Oncogene Inﬂuence of A-Raf and Raf-1 on ERK phosphorylation and Gl/S cell cycle progression K Mercer et al 5214 a b The combined disruption of A-Raf and Raf-1 leads to a signiﬁcant decrease in transient ERK activation following growth factor stimulation (Figure 4). The fact

DKO DKO WT DKO+GFP DKO+Raf-1 WT DKO+GFPDKO+Raf-1 that transient MEK phosphorylation is also disrupted in

-+ - + - +- ++Serum the DKO cells indicates that the changes observed in

→ cyclin D1 → phosphoERK are more likely to be due to changes in the

Raf-1 activities of Raf/MEK rather than changes in the

→ actin activities of MEK or MAPK phosphatases.

→ Previous data indicated that the three Rafs have → PP-ERK tissue-speciﬁc patterns of gene expression (Storm et al., 1990). However, through the availability of good antibodies for each Raf protein, it has now become

→ actin clear that all three have ubiquitous patterns of gene expression (Mercer and Pritchard, 2003). Therefore, the c different role of each Raf isotype in regulating 30 endogenous MEK/ERK activation is more likely to be 20 due to different mechanisms in the regulation of their kinase activities rather than differences in their expres- cells 10 sion patterns. Indeed, the proﬁles of activation of 0 endogenous Raf-1 and A-Raf in MEFs following % BrdU positive BrdU %

WT growth factor stimulation, as measured by the immu- DKO noprecipitation MEK/ERK kinase cascade assay, cor-

DKO+GFP relate well with the transient peak of MEK/ERK DKO+Raf-1 activation they are involved in stimulating (Huser Figure 6 Rescue of the DKO phenotype by overexpression of et al., 2001). Raf-1 and A-Raf activities are stimulated Raf-1. (a) Rescue of phosphoERK levels. DKO cells were transiently transfected with vectors expressing either GFPor five-fold and 1.5-fold, respectively, following growth human Raf-1. Cell lysates were prepared from WT and DKO cells, factor treatment and maximal activities are observed as well as transfected cells that were either unstimulated cells or after 5 min of treatment. Their activities drop signifi- stimulated with serum for 10 min. Western blots were analysed with cantly after this peak and return to basal levels by antibodies for Raf-1 (top), phosphoERK (middle) and actin as a loading control (bottom). (b) Rescue of cyclin Dl levels. Cell lysates 60 min. By contrast, the profile of B-Raf activation in were prepared from WT and DKO cells, as well as transfected cells MEFs correlates with sustained ERK activation. B-Raf grown in complete medium. Western blots were analysed with has a high level of basal activity in unstimulated cells antibodies for cyclin Dl (top) and actin as a loading control and its activity is further stimulated 1.5-fold by growth (bottom). (c) Rescue of proliferation. WT, DKO and transfected factors and reaches a maximum after 5 min of treatment. cells were serum starved for 20 h and then stimulated with serum for 4 h. The percentage of cells entering S phase was determined by However, its activity is sustained for far longer than that performing a BrdU proliferation assay (see Materials and methods) of Raf-1 or A-Raf as it remains active at later time points (Huser et al., 2001). The reasons for the differences in profiles of activation/deactivation of the three Rafs are not entirely clear proliferation. To further explore the role of Raf-1 and but must reflect biochemical similarities between A-Raf B-Raf in regulating cell proliferation via MEK/ERK and Raf-1 and important biochemical differences activation, we are currently generating knockin mice between these two Raf kinases and B-Raf. All three with kinase inactive versions of Raf-1 and B-Raf. Raf kinases are located in the cytosol bound to 14-3-3 in Previous studies of raf-1À/À mice showed that the lack inactive states and translocate to the plasma membrane of Raf-1 is associated with an increase in apoptosis in the presence of active Ras (Kolch, 2000; Marais et al., (Huser et al., 2001; Jesenberger et al., 2001; Mikula 1995; Marais et al., 1997; Avruch et al., 2001; Dhillon et al., 2001; Hindley and Kolch, 2002) whereas the raf- and Kolch, 2002; Ory et al., 2003). All three Rafs require 1À/À embryos and MEFs analysed here do not show an phosphorylation/dephosphorylation for full activation apoptotic phenotype (Figures 2 and 3). The difference at the membrane but the patterns of phosphorylation between these studies is the genetic background of the differ somewhat (Dhillon and Kolch, 2002; Chong et al., mice investigated. The previous studies analysed mice on 2003). Both Raf-1 and B-Raf require dephosphorylation either the 129Ola/C57BL6 or 129Ola/129Sv mixed of Akt consensus sequences at serine 259 in Raf-1 and at inbred genetic backgrounds, whereas mice on the the equivalent serine 364 in B-Raf (Chong et al., 2001; outbred MF-1 background were studied here. This Zhang et al., 2001; Dhillon et al., 2002) as well as would suggest that the role of Raf-1 in apoptosis is phosphorylation of threonine and serine residues in the complex and is highly dependent on the genetic origin of activation loop for full activation (Zhang and Guan, the cells under study and indeed may explain some 2000; Chong et al., 2001). In B-Raf, the purpose of these conflicting data in the literature that has defined Raf-1 activation loop phosphorylations appear to be to both as a promoter (Blagosklonny et al., 1996, 1997; disrupt the P-loop–DFG interaction stabilizing the Kauffmann-Zeh et al., 1997; Basu et al., 1998) and inactive conformation of the kinase (Wan et al., 2004). inhibitor of apoptosis (Baccarini, 2002). However, Raf-1 also requires phosphorylation of

Oncogene Influence of A-Raf and Raf-1 on ERK phosphorylation and Gl/S cell cycle progression K Mercer et al 5215 tyrosine 341 and serine 338 residues in the N region of has shown that, when ERK activation is transient, c-Fos the kinase domain to achieve full activation, whereas B- expression is induced but the protein is unstable and is Raf does not require phosphorylation in this region as it degraded (Murphy et al., 2002). However, under is constitutively phosphorylated at serine 445, the conditions when ERK activation is sustained, the c- equivalent residue to serine 338 in Raf-1, and possesses Fos induced during the immediate early phase is a phosphomimetic aspartic acid residue at 448, the phosphorylated on multiple sites by ERK and p90RSK equivalent residue to tyrosine 341 in Raf-1 (Marais and is thus stabilized and contributes to the induction of et al., 1995, 1997; Mason et al., 1999; Dhillon and expression of cyclin Dl. Our data support this important Kolch, 2002; Dhillon et al., 2002). A-Raf appears to be functional role of c-Fos as a sensor for the duration of regulated in a similar way to Raf-1 (Marais et al., 1997). Raf/ERK signalling. The process of Raf deactivation has been little studied to date but must involve loss of Ras binding due to the conversion of Ras.GTPto Ras. GDPby RasGAPs, Materials and methods phosphorylation/dephosphorylation events leading to the reformation of the P-loop–DFG interaction, and the Derivation of mice and culture ofMEFs formation of new protein–protein interactions in the Mice containing a homozygous knockout mutation of the A- cytosol, particularly with 14-3-3. The sustained activa- raf gene and the raf-1 gene have been described previously tion of B-Raf following growth factor treatment would (Pritchard et al., 1996; Huser et al., 2001). These mice were suggest that this deactivation process is more prolonged backcrossed onto the MF1 genetic background. Embryos were for B-Raf than it is for Raf-1 or A-Raf. collected at E10.5–E14.5, homogenized and fibroblasts were derived by standard procedures (Huser et al., 2001). PCR An important step for cell cycle re-entry and genotyping was performed using the primers and conditions progression through Gl is expression of the D-type described previously (Huser et al., 2001; Mercer et al., 2002). cyclins and consequent activation of cyclinD : cdk4/6 Each primary MEF culture was isolated from a single embryo. complexes (Jiang et al., 1993; Resnitzky et al., 1994; Fibroblasts were cultured in high glucose (4.5 g/1) Dulbecco’s Stacey, 2003). A key mechanism by which signalling modified Eagle’s medium (DMEM; Life Technologies) con- through the Raf/MEK/ERK can induce cell cycle taining 10% (v/v) foetal calf serum (FCS; SeraQ) and 100 U/ progression is through stimulation of cyclin Dl mRNA ml penicillin/streptomycin (Life Technologies) in a 10% CO2 expression (Kerkhoff and Rapp, 1997; Woods et al., humidified incubator at 371C. 1997; Marshall, 1999). This is achieved by the ability of ERKs to translocate to the nucleus and phosphorylate Histology Ets and AP-1 family members that transactivate the For sectioning, embryos were fixed in methacarn (10% (v/v) cyclin Dl promoter (Herber et al., 1994; Albanese et al., glacial acetic acid, 30% (v/v) chloroform, 60% (v/v) metha- 1995; Lavoie et al., 1996; Whitmarsh and Davis, 1996; nol), and paraffin embedded. Paraffin sections at 5 mm Balmanno and Cook, 1999; Cook et al., 1999; Shaulian thickness were mounted onto microscope slides pretreated and Karin, 2002). In fibroblasts, several studies have with silane. For Ki67 staining, after blocking in 6% (v/v) shown that the sustained activation of ERKs induced by hydrogen peroxide in methanol for 30 min and 1 : 25 dilution various mitogens is associated with cell cycle progres- of swine serum for 15 min, sections were incubated with a sion, whereas transient peaks of ERK activation are not 1 : 200 dilution of a rabbit polyclonal Ki67 antibody (Vector Laboratories, Burlingame, CA, USA) at room temperature (Vouret-Craviari et al., 1993; Cook and McCormick, overnight. Antigen–antibody complexes were visualized with 1996; Weber et al., 1997; Balmanno and Cook, 1999). biotinylated secondary antibody and with the avidin–biotin– An extensive study of the AP-1 components expressed peroxidase complex (Vector Laboratories, Burlingame, CA, during these different phases showed that while many USA). For TUNEL assays, the ApoAlertt DNA fragmenta- AP-1 components (c-Fos, Fra-1, Fra-1, c-Jun and JunB) tion assay kit (BD Biosciences) was used following the are expressed at immediate early time points following manufacturer’s instructions. mitogen treatment of cells, the expression of c-Fos is not observed at sustained time points whereas the expression Cell stimulations, B-Raf kinase assays and immunoblotting of the other AP-1 components persists at later time Primary MEFs were placed in serum-free media for 20 h and points (Kovary and Bravo, 1992; Balmanno and Cook, then stimulated with either 10 ng/ml EGF or with 10% (v/v) 1999; Cook et al., 1999). The role of the immediate early FCS over a time course of up to 20 h. Protein lysates and induction of c-Fos expression in cell cycle re-entry was Western blots were carried out as described previously therefore brought into question (Balmanno and Cook, (Luckett et al., 2000). B-Raf assays were performed using the 1999; Cook et al., 1999). immunoprecipitation kinase cascade assay (Marais et al., In this report, our results show that transiently 1997). Primary antibodies were a 1 : 1000 dilution of a rabbit induced ERK phosphorylation is significantly reduced polyclonal antibody for A-Raf (Santa Cruz Biotechnology in the DKO cells and the cell cycle is not maintained at Inc.), a 1 : 1000 dilution of a mouse monoclonal antibody for its usual rate. This may be mediated in part by transient Raf-1 (Transduction Laboratories), a 1 : 1000 dilution of a mouse monoclonal antibody for B-Raf (Santa Cruz Biotech- phosphoERK induction of c-Fos expression and con- nology Inc.), a 1 : 1000 dilution of an antibody for actin sequent induction of cyclin Dl mRNA expression (Sigma), a 1 : 1000 dilution of a mouse monoclonal antibody (Figure 5), although transient phosphoERK may also against Thr202/Tyr204 phospho-p44/42 ERK1/2 (Cell Signal- influence other parts of the cell cycle machinery that ling Tech.), a 1 : 1000 dilution of a rabbit polyclonal antibody have yet been subjected to investigation. A recent study for ERK2 (Zymed Laboratories Inc.), a 1 : 1000 dilution of a

Oncogene Influence of A-Raf and Raf-1 on ERK phosphorylation and Gl/S cell cycle progression K Mercer et al 5216 rabbit polyclonal antibody against Ser217/Ser221 phospho- cycle analysis, primary cells at 80% confluency on 6 cm dishes MEKl/2 (Cell Signalling Tech.), a 1 : 2000 dilution of a mouse were collected and fixed in 70% (v/v) ethanol at 41C for monoclonal antibody against cyclin Dl, a 1 : 2000 dilution of a 30 min. Fixed cells were then resuspended in PBS containing mouse monoclonal antibody against cdk4, a 1 : 2000 dilution of 10 mg/ml propidium iodide and 100 mg/ml RNase at room temp a mouse monoclonal antibody against cdk6, a 1 : 2000 dilution for 1 h. FACS analysis was performed using a Becton of a mouse monoclonal antibody against cyclin D3 (all from Dickinson flow cytometer. BrdU incorporation assays for Cell Signalling Tech.) and a 1 : 1000 dilution of a rabbit proliferation were performed by using the BrdU labelling and polyclonal antibody for c-Fos (Santa Cruz Biotechnologies). detection kit provided by Roche. The percentage of BrdU- positive cells was visualized by fluorescence microscopy using a Transfection of MEFs Zeiss Axiophot microscope. To induce apoptosis, primary cells at 80% confluency on 6 cm dishes were treated with 50 ng/ml DKO MEFs were transfected with vectors expressing either anti-CD95 antibody with 0.5 mM cycloheximide for 20 h in a myc-tagged full-length human Raf-1 or GFPusing a 371C humidifying incubator. Annexin V staining and FACS Nucleofector under the MEF conditions recommended by analysis were performed as described previously (Huser et al., the manufacturer (Amaxa Biosystems, Germany). At 24 h 2001). following transfection, cells were placed in serum-free media for 20 h and then either left unstimulated or stimulated with 10% (v/v) FCS for 10 min. Protein lysates and western blots Acknowledgements were carried out as described above. We are extremely grateful to The Wellcome Trust and Cancer Research UK for providing financial support for this project. Proliferation and apoptosis assays We thank Mabel Iwobi and Vicky Aldridge for help during the For growth curves, 2 Â 104 cells were plated and counted at initial stages of this project and Simon Cook for advice on 24 h intervals in triplicate using a haemocytometer. For cell detecting c-Fos.

References

Albanese C, Johnson J, Watanabe G, Eklund N, Vu D, Arnold Dhillon AS, Meikle S, Yazici Z, Eulitz M and Kolch W. A and Pestell RG. (1995). J. Biol. Chem., 270, 23589–23597. (2002). EMBO J., 21, 64–71. Avruch J, Khokhlatchev A, Kyriakis JM, Luo Z, Tzivion G, Herber B, Truss M, Beato M and Muller R. (1994). Oncogene, Vavvas D and Zhang XF. (2001). Recent Prog. Horm. Res., 9, 1295–1304. 56, 127–155. Hindley A and Kolch W. (2002). J. Cell. Sci., 115, 1575–1581. Baccarini M. (2002). Cell Death Differ., 9, 783–785. Huser M, Luckett J, Chiloeches A, Mercer K, Iwobi M, Balmanno K and Cook SJ. (1999). Oncogene, 18, 3085–3097. Giblett S, Sun XM, Brown J, Marais R and Pritchard C. Basu S, Bayoumy S, Zhang Y, Lozano J and Kolesnick R. (2001). EMBO J., 20, 1940–1951. (1998). J. Biol. Chem., 273, 30419–30426. Jesenberger V, Procyk KJ, Ruth J, Schreiber M, Theussl HC, Blagosklonny MV, Giannakakou P, el-Deiry WS, Kingston Wagner EF and Baccarini M. (2001). J. Exp. Med., 193, DG, Higgs PI, Neckers L and Fojo T. (1997). Cancer Res., 353–364. 57, 130–135. Jiang W, Kahn SM, Zhou P, Zhang YJ, Cacace AM, Infante Blagosklonny MV, Schulte T, Nguyen P, Trepel J and Neckers AS, Doi S, Santella RM and Weinstein IB. (1993). LM. (1996). Cancer Res., 56, 1851–1854. Oncogene, 8, 3447–3457. Brown JR, Nigh E, Lee RJ, Ye H, Thompson MA, Saudou F, Karasarides M, Chiloeches A, Hayward R, Niculescu-Duvaz Pestell RG and Greenberg ME. (1998). Mol. Cell. Biol., 18, D, Scanlon I, Friedlos F, Ogilvie L, Hedley D, Martin J, 5609–5619. Marshall CJ, Springer CJ and Marais R. (2004). Oncogene, Chen Z, Gibson TB, Robinson F, Silvestro L, Pearson G, Xu 21, 21. B, Wright A, Vanderbilt C and Cobb MH. (2001). Chem. Kauffmann-Zeh A, Rodriguez-Viciana P, Ulrich E, Gilbert C Rev., 101, 2449–2476. Coffer P, Downward J and Evan G. (1997). Nature, 385, Chong H, Lee J and Guan KL. (2001). EMBO J., 20, 544–548. 3716–3727. Kerkhoff E and Rapp UR. (1997). Mol. Cell. Biol., 17, Chong H, Vikis HG and Guan KL. (2003). Cell Signal., 15, 2576–2586. 463–469. Kerkhoff E and Rapp UR. (1998). Oncogene, 17, 1457–1462. Cook SJ, Aziz N and McMahon M. (1999). Mol. Cell. Biol., Kimura ET, Nikiforova MN, Zhu Z, Knauf JA, Nikiforov YE 19, 330–341. and Fagin JA. (2003). Cancer Res, 63, 1454–1457. Cook SJ and McCormick F. (1996). Biochem. J., 320, 237–245. Kolch W. (2000). Biochem. J., 351, 289–305. Davies H, Bignell GR, Cox C, Stephens P, Edkins S, Clegg S, Kovary K and Bravo R. (1991). Mol. Cell. Biol., 11, 4466–4472. Teague J, Woffendin H, Garnett MJ, Bottomley W, Davis N, Kovary K and Bravo R. (1992). Mol. Cell. Biol., 12, Dicks E, Ewing R, Floyd Y, Gray K, Hall S, Hawes R, 5015–5023. Hughes J, Kosmidou V, Menzies A, Mould C, Parker A, Lavoie JN, L’Allemain G, Brunet A, Muller R and Pouyssegur Stevens C, Watt S, Hooper S, Wilson R, Jayatilake H, J. (1996). J. Biol. Chem., 111, 20608–20616. Gusterson BA, Cooper C, Shipley J, Hargrave D, Pritchard- Lewis TS, Shapiro PS and Ahn NG. (1998). Adv. Cancer Res., Jones K, Maitland N, Chenevix-Trench G, Riggins GJ, 74, 49–139. Bigner DD, Palmieri G, Cossu A, Flanagan A, Nicholson A, Luckett JC, Huser MB, Giagtzoglou N, Brown JE and Ho JW, Leung SY, Yuen ST, Weber BL, Seigler HF, Darrow Pritchard CA. (2000). Cell Growth Differ., 11, 163–171. TL, Paterson H, Marais R, Marshall CJ, Wooster R, Marais R, Light Y, Paterson HF and Marshall CJ. (1995). Stratton MR and Futreal PA. (2002). Nature, 417, 949–954. EMBO J., 14, 3136–3145. Dhillon AS and Kolch W. (2002). Arch. Biochem. Biophys., Marais R, Light Y, Paterson HF, Mason CS and Marshall CJ. 404, 3–9. (1997). J. Biol. Chem., 111, 4378–4383.

Oncogene Inﬂuence of A-Raf and Raf-1 on ERK phosphorylation and Gl/S cell cycle progression K Mercer et al 5217 Marais R and Marshall CJ. (1996). Cancer Surv., 27, 101–125. Shaulian E and Karin M. (2002). Nat. Cell. Biol., 4, Marshall C. (1999). Curr. Opin. Cell. Biol., 11, 732–736. E131–E136. Marshall CJ. (1995). Cell, 80, 179–185. Squires MS, Nixon PM and Cook SJ. (2002). Biochem. J., 366, Mason CS, Springer CJ, Cooper RG, Superti-Furga G, 673–680. Marshall CJ and Marais R. (1999). EMBO J., 18, Stacey DW. (2003). Curr. Opin. Cell. Biol., 15, 158–163. 2137–2148. Stevaux O and Dyson NJ. (2002). Curr. Opin. Cell. Biol., 14, Mercer K, Chiloeches A, Huser M, Kiernan M, Marais R and 684–691. Pritchard C. (2002). Oncogene, 21, 347–355. Storm SM, Cleveland JL and Rapp UR. (1990). Oncogene, 5, Mercer KE and Pritchard CA. (2003). Biochim. Biophys. Acta., 345–351. 1653, 25–40. Vouret-Craviari V, Van Obberghen-Schilling E, Scimeca JC, Miao GG and Curran T. (1994). Mol. Cell. Biol., 14, Van Obberghen E and Pouyssegur J. (1993). Biochem. J., 4295–4310. 289, 209–214. Mikula M, Schreiber M, Husak Z, Kucerova L, Ruth J, Wan PT, Garnett MJ, Roe SM, Lee S, Niculescu-Duvaz D, Wieser R, Zatloukal K, Beug H, Wagner EF and Baccarini Good VM, Jones CM, Marshall CJ, Springer CJ, Barford D M. (2001). EMBO J., 20, 1952–1962. and Marais R. (2004). Cell, 116, 855–867. Murphy LO, Smith S, Chen RH, Fingar DC and Blenis J. Weber JD, Raben DM, Phillips PJ and Baldassare JJ. (1997). (2002). Nat. Cell Biol., 4, 556–564. Biochem. J., 326, 61–68. Ory S, Zhou M, Conrads TP, Veenstra TD and Morrison DK. Whitmarsh AJ and Davis RJ. (1996). J. Mol. Med., 74, (2003). Curr. Biol., 13, 1356–1364. 589–607. Pritchard CA, Bolin L, Slattery R, Murray R and McMahon Wojnowski L, Stancato LF, Larner AC, Rapp UR and M. (1996). Curr Biol, 6, 614–617. Zimmer A. (2000). Mech. Dev., 91, 97–104. Pritchard CA, Hayes L, Wojnowski L, Zimmer A, Marais RM Won KA, Xiong Y, Beach D and Gilman MZ. (1992). Proc. and Norman JC. (2004). Mol. Cell. Biol., 24, 5937–5952. Natl. Acad. Sci. USA, 89, 9910–9914. Pritchard CA, Samuels ML, Bosch E and McMahon M. Woods D, Parry D, Cherwinski H, Bosch E, Lees E and (1995). Mol. Cell. Biol., 15, 6430–6442. McMahon M. (1997). Mol. Cell. Biol., 17, 5598–5611. Resnitzky D, Gossen M, Bujard H and Reed SI. (1994). Mol. Zhang BH and Guan KL. (2000). EMBO J., 19, 5429–5439. Cell. Biol., 14, 1669–1679. Zhang BH, Tang ED, Zhu T, Greenberg ME, Vojtek AB and Samuels ML, Weber MJ, Bishop JM and McMahon M. Guan KL. (2001). J. Biol. Chem., 276, 31620–31626, [E-pub (1993). Mol. Cell. Biol., 13, 6241–6252. 2001 Jun 15].

Oncogene