p204 Protein Is a Novel Modulator of Ras Activity*

The murine p200 family protein, p204, modulates cell proliferation and tissue differentiation. Many of its activities are exerted in the nucleus. However, in cardiac myocytes, p204 accumulated in the cytoplasm. A yeast two-hybrid assay revealed a p204-cytoplasmic Ras protein interaction. This was confirmed (i) by coimmunoprecipitation of p204 with Ras in mouse heart extract and with endogenous or ectopic H-Ras and K-Ras in cell lysates as well as (ii) by binding of purified H-Ras-GTP to purified p204 in vitro. p204 inhibited (i) the cleavage of RasGTP to RasGDP by RasGAP; (ii) the binding to RasGTP of Raf-1, phosphatidylinositol 3-kinase, and Ral-GDS, effectors of Ras signaling; and (iii) activation by the Ras pathway of the phosphorylation and thus activation of downstream targets (e.g. MEK, Akt, and p38MAPK). Oncogenic Ras expression triggered the phosphorylation and translocation of p204 from the nucleus to the cytoplasm. This is expected to increase the interaction between the two proteins. Translocation triggered by Ras oncoprotein was blocked by the LY294002 inhibitor of phosphatidylinositol 3-kinase. Ras did not promote phosphorylation or translocation to the cytoplasm of mutated p204 in which serine 179 was replaced by alanine. p204 overexpression inhibited the anchorage-independent proliferation of cells expressing RasQ61L oncoprotein. Ras oncoprotein triggered in MEF3T3 cells the rearrangement of the actin cytoskeleton and the enhancement of cell migration through a membrane. Overexpression of p204 inhibited both. Ras oncoprotein or activated, wild-type Ras was described to increase Egr-1 transcription factor expression. We report that a sequence in the gene encoding p204 bound Egr-1, and Egr-1 activated p204 expression. Ras oncoprotein or activated wild-type Ras increased the expression in 3T3 cells of p204 together with that of Egr-1. Furthermore, the activation of expression of a single copy of K-ras oncogene in cultured murine embryonic cells induced the expression of a high level of p204 as well as its distribution between the nuclei and the cytoplasm. Thus, p204 may serve as a negative feedback inhibitor of Ras activity.

The interferons are vertebrate cytokines with antimicrobial, immunomodulatory, and cell growth and differentiation regulatory activities (1)(2)(3). They function by modulating the expression of many genes, including those of the gene 200 cluster (4). In mice, this cluster consists of at least 10 genes that encode the p200 family proteins. The human counterpart of the cluster consists apparently of four genes (MNDA, IFI16, AIM2, and IFIX) that encode the Hin200 family proteins (5,6).
Among the best characterized members of the murine p200 family proteins is p202a (which was designated earlier as p202) (7). p202a modulates transcription, cell proliferation, and apoptosis, and its overexpression was correlated with symptoms of lupus erythematosus (8,9). It functions primarily by binding numerous sequence-specific transcription factors and transcription modulators, including pRb and p53, and inhibiting their activity generally, but not exclusively, by binding them and inhibiting their sequence-specific binding to DNA (8 -10).
A second p200 family protein, p205, was originally designated as D3. p205 can also bind pRb and p53. It can inhibit cell proliferation by inducing pRb and increasing the level of p21 in a p53-dependent manner but can inhibit it also independently of pRb and p53 (11).
A third much studied p200 family protein is p204, encoded by the Ifi204 gene. p204 is structurally related to p202a (7). p204 is involved in the modulation of transcription, protein degradation, cell proliferation, and differentiation of numerous tissues. The tissues include, among others, skeletal muscle myotubes (12,13), heart muscle myocytes (14,15), bone osteoblasts (16,17), and macrophages (18). p204 modulates a series of biochemical mechanisms to promote the differentiation of different tissues. Moreover, the expression of p204 in different tissues can be promoted by distinct cytokines, promoters, and tissue-specific transcription factors. Depending on the cell type and stage of differentiation, p204 can be nucleolar, nucleoplasmic, and/or cytoplasmic. In performing its various functions, p204 binds and affects the activities of numerous proteins. These include, among others, UBF, the factor involved in ribosomal RNA synthesis (19), the pocket proteins (pRb, p107, and p130) (19,20) involved in the control of cell proliferation and differentiation, and the Id (inhibitor of differentiation) proteins (13,15). p204 also occurs in multiprotein complexes (e.g. in the p204-pRb-Cbfa1 complex involved in osteoblast differentiation). In this complex, p204 acts as a transcriptional coactivator (17). p204 was also reported to be required for the replication of cytomegalovirus (21).
Among 10 adult mouse tissues tested, the level of p204 was highest in cardiac myocytes (12). Moreover, in the course of differentiation of the myocytes from embryonal stem cells, the bulk of p204 was translocated from the nucleus to the cyto-plasm (12,14). These facts prompted us to search for cytoplasmic proteins in the cardiac myocytes to which p204 binds. Here we report that cytoplasmic H and K-Ras proteins (which are encoded by the H-and K-ras genes) were bound by p204 and that p204 inhibited the various activities of these multifunctional proteins. Particular mutants of the ras genes (designated as ras oncogenes) play a causal role in more than a quarter of human cancers (22)(23)(24)(25)(26). Wild-type Ras proteins control signaling pathways. They are GTP-binding proteins, which, when acted upon by specific factors, cycle between an activated and an inactivated form, RasGTP and RasGDP, respectively (27). Ras proteins can be localized on the inner cell membrane or in the Golgi or endosomal compartments (28,29). They are coupling growth factor receptors to downstream signaling pathways regulating transcription, translation, cell growth, proliferation, cell shape, apoptosis, senescence, and malignant transformation (26,30). ras oncogenes encode Ras oncoproteins, which are persistently activated, since (unlike wild-type Ras proteins) they are unable to promote the cleavage of the bound GTP to GDP (31)(32)(33). This cleavage is promoted in the case of wildtype Ras proteins by RasGAP binding to RasGTP.
Here, we report that p204 bound to H-RasGTP and K-Ras-GTP and inhibited their cleavage by RasGAP and also the binding to H-RasGTP of the effectors Raf, PI3K, and RAL-GDS as well as the activation by phosphorylation of some downstream targets (e.g. MEK, Akt, and p38 MAPK ). Oncogenic Ras expression triggered the phosphorylation and translocation of p204 from the nucleus to the cytoplasm. This was blocked by inhibiting PI3K activity. p204 inhibited the anchorage-independent proliferation of cells expressing Ras oncoprotein and also blocked the rearrangement of the actin cytoskeleton and the enhancement of cell migration through a membrane by Ras oncoprotein.
As described by others, Ras oncoprotein or wild-type Ras-GTP increased the expression of the Egr-1 transcription factor (42). We found that ectopic Ras oncoprotein or wild-type Ras-GTP increased the level of p204, together with that of Egr-1, in consequence of the boost of the transcription of p204 by Egr-1. Moreover, the activation of expression of a single copy of the K-ras oncogene (43) in cultured murine embryonic cells induced the expression of high levels of p204 as well as its distribution between the nuclei and the cytoplasm. Thus, p204 might also serve as a negative feedback inhibitor of Ras activity.

EXPERIMENTAL PROCEDURES
Plasmids pBI-H-Ras G12V and pBI-p204 were generated by inserting the appropriate H-Ras G12V PCR product or 204 PCR product into the PstI/SalI sites or MuII/NheI sites, respectively, of the pBI plasmid (BD Clontech). The ecdysone-inducible plasmid pIND204 was constructed as described (19). pCGN-H-Ras Q61L (44) was obtained from H. P. Shao. An MBP-p204 expression plasmid was constructed by inserting 204 cDNA between the EcoRI and HindIII sites of the pMAL plasmid (New England Biolabs). pGEX-H-Ras G12V was obtained from H. Koide. pcDNA3P110␣-HA and pCMV6-P85␣-FLAG were donated by J. Luo. The pBabepro-Myc-Ral-GDS plasmid was from L. A. Feig. GST-Raf-1 Ras-binding domain (RBD) was from Upstate Biotechnology, Inc. (Lake Placid, NY). EGFP-H-Ras Q61L and pIRESpuro3-H-Ras Q61L were generated by inserting H-Ras Q61L cDNA into the EcoRI and BamHI sites of the EGFP-C1 and the pIRESpuro3 vectors (BD Clontech), respectively. pCGNHA-K-Ras4B G12V and three other expression plasmids encoding the same protein, except with single amino acid substitutions in the effector segment, were generous gifts from G. J. Clark (45). pCMV-Egr-1 (46) was from L. Nagy. The pGL-Egr-1 reporter construct was generated by inserting a PCR product extending from nucleotide Ϫ6927 to Ϫ6474 in the Ifi204 gene 5Ј-flanking region from the BAC225 clone (4) between the KpnI and BglII sites of the pGL vector (Promega). The mutant Egr-1 reporter construct (pGL-mEgr-1) was generated using the QuikChange kit (Stratagene). pCMV204 was described (15). The expression plasmid pCMV204S179A was generated from pCMV204 using the same kit. This plasmid encoded a mutant p204 in which Ser at position 179 was replaced by Ala.

Fusion Proteins
A plasmid encoding the MBP-p204 fusion protein was expressed in Escherichia coli BL21-Gold(DE3)pLysS codon plus (kindly provided by D. Soll). MBP-p204 was purified using amylose beads (New England Biolabs). Its purity was verified by 4 -20% PAGE and Coomassie Brilliant Blue staining according to the New England Biolabs manual. The purity of His-RasGAP (a kind gift of T. Koleske) was also verified by PAGE. GST-H-Ras expressed in E. coli DH5␣ was purified on a GST column (see Fig. S2).

Interferon
A modified human ␣-interferon active in murine cells was a kind gift of C. Weissmann (47).

Generation of Stable Cell Lines
A stable cell line with a low level of constitutive H-Ras expression was generated by transfection of the pIRESpuro3-H-Ras Q61L plasmid with Lipofectamine 2000 into an AKR-2B line in which p204 was inducible by ecdysone or ponasterone A (19) and selection with 6 g/ml puromycin in DMEM, 10% FCS for 2 weeks. A clone expressing upon induction a low level of H-Ras Q61L was picked. A control cell line was generated by transfection of the vector. The doxycycline-inducible lines MEF3T3-Tet-off-H-Ras G12V , MEF3T3-Tet-off-p204, and MEF3T3-Tet-off-H-Ras G12V -p204 were obtained by transfection into the MEF3T3-Tet-off line (Clontech) of pBI-H-Ras G12V and/or pBI-204, together with a hygromycin resistance plasmid, followed by selection with hygromycin. Fig. 2

, B-D)
FT9 cells (i.e. NIH3T3 cells expressing an oncogenic H-Ras mutant) (49) were maintained in DMEM, 10% FCS medium and were harvested in MLB lysis buffer (25 mM HEPES (pH 7.5), 150 mM NaCl, 1% Igepal CA-630 (Sigma), 0.25% sodium deoxycholate, 10% glycerol, 25 mM NaF, 10 mM MgCl 2 , 1 mM EDTA, 1 mM sodium orthovanadate, 10 g/ml aprotinin, 10 g/ml leupeptin). Lysate aliquots with 500 g of total protein per 0.5 ml were incubated with protein G beads with conjugated anti-H-Ras (Upstate Biotechnology, Inc.) or anti-p204 antibodies at 4°C overnight. The beads were washed and boiled in SDS loading buffer, and the eluate was subjected to 4 -20% SDS-PAGE for Western blotting, using antibodies against H-Ras, K-Ras, or p204. Similar procedures were used in the coimmunoprecipitation experiments in Fig. 2, C and D, which involved different cell lines from that used in the experiments shown in Fig. 2B.

Assay of RasGTP/GDP Binding to p204 (for Fig. 2F)
Cells of the MEF3T3-Tet-off-p204 line were plated in 20 10-cm dishes and, for inducing p204, were incubated first with 1 g/ml doxycycline for 24 h. Then the doxycycline was removed, and the cultures were incubated for a further 48 h. The cultures were lysed in MLB lysis buffer, and aliquots from the lysate were incubated with GTP␥S or GDP at 30°C for 30 min. Ras protein was pulled down, using protein A beads conjugated with anti-p204 and, as a positive control, beads to which GST-Raf-1 RBD was conjugated.
For Fig. 3, pcDNA3-204 or pcDNA3-204 segment expression plasmids were generated by inserting 204 cDNA or its various segments (prepared by PCR) between the EcoRI and BamHI sites of the vector. The plasmids were expressed in a reticulocyte lysate (Promega) in the presence of [ 35 S]methionine. 10-l aliquots from each 35 S-labeled segment were added to 0.5-ml aliquots containing 500 g of protein from a lysate of cultured HEK293 cells that had been transfected with the expression plasmid pCGN-HA-H-Ras Q61L and preincubated with GTP␥S. The reaction mixture was incubated at 4°C overnight. The labeled p204 or its various segments bound to HA-H-Ras Q61L -GTP␥S were pulled down with protein G/A beads loaded with anti-HA antibodies and assayed by SDS-PAGE and autoradiography.

Assays of the Inhibition by p204 of the Binding of H-Ras to Raf-1, PI3K, and RalGDS (for Fig. 5)
Raf-1-MBP and MBP-p204 fusion protein were purified using amylose beads (New England Biolabs) and were quantified by the Bio-Rad protein assay (see Fig. S2). NIH3T3 cells were incubated in DMEM, 0.5% FCS for 48 h, and, if so indicated, the level of RasGTP was increased in them by adding 100 ng/ml EGF to the medium for the last 10 min. MBP-p204, MBP, or neither was incubated with GST-Raf-1 RBD (Upstate Biotechnology) immobilized on agarose beads at 4°C for 30 min and then added to the lysates of the NIH3T3 cells and further incubated at 4°C for 30 min. The beads were washed and eluted with boiling SDS loading buffer, and the H-Ras eluted was assayed by Western blotting, following the Upstate Biotechnology protocol.
PI3K-HEK293 cells were transfected with pIRES-H-Ras Q61L or with pcDNA3P110␣-HA and pCMV6-P85␣-FLAG, using Lipofectamine 2000, and incubated in DMEM, 10% FCS medium in 10% CO 2 for 48 h. The cells were harvested and lysed in MLB buffer. Total protein levels were assayed and adjusted to 500 g of protein/0.5 ml of lysate. If so indicated, 400 nmol of MBP or 400 of nmol MBP-204 were incubated with 0.5 ml of lysate of cells expressing H-Ras Q61L at 4°C for 30 min and were then supplemented with 0.5 ml of a lysate from cells expressing P110␣-HA and P85␣-FLAG and incubated with protein G beads loaded with anti-HA antibodies or mouse IgG control at 4°C overnight. The beads were washed and eluted with boiling SDS loading buffer. 30-l samples were analyzed by Western blotting, using anti-H-Ras antibodies.
RalGDS-A culture of HEK293 cells was infected with retrovirus-packaged pBabepro-Myc-RalGDS. A second culture of HEK293 cells was transfected with pCGN-HA-H-Ras Q61L . After incubation, cell lysates were prepared. The lysate from the culture transfected with pCGN-HA-H-Ras Q61L was incubated, as indicated, without or with MBP or MBP204. These incubated lysates were mixed with the lysate from the culture infected with retrovirus-packaged pBabepro-Myc-RalGDS and processed by pull-down with immobilized Myc antibodies or mouse IgG and Western blotting, using anti-HA.

Assay of the Inhibition by p204 of the Phosphorylation of MEK, Akt, and p38 MAPK
For Fig. 6A, an AKR-2B line, in which p204 expression was induced by ponasterone A (AKR-2B-Ind-p204) and a control line (AKR-2B-Ind-Con) were established from AKR-2B cells by transfection of constructs encoding ponasterone A receptors and selectable markers from the ecdysone-inducible expression kit (Invitrogen) as well as inducible p204, followed by selection (12). To induce p204, the cells were incubated with 1 M ponasterone A in DMEM, 0.5% FCS for 48 h. EGF (200 ng/ml) was added 30 min before harvesting the cells in SDS loading buffer. 25-l aliquots of the lysates were subjected to Western blotting with p204 antibodies. Phosphorylation of MEK triggered by EGF (via the Ras/Raf-1 pathway) as well as the amount of total MEK protein were assayed by Western blotting, using antiphospho-MEK and anti-MEK antibodies. The levels of phospho-Akt, total Akt, phospho-p38 MAPK , total p38 MAPK , and p204 proteins were similarly assayed after stripping the membranes and applying appropriate antibodies.
For Fig. 6B, one NB508 culture was transfected with a pCMV204AS expression plasmid, a second was transfected with the pCMV vector, and both cultures were also transfected with a hygromycin resistance expression plasmid. After selection with hygromycin, resistant clones were selected. These were incubated in DMEM, 10% fetal bovine serum for about 48 h. When reaching 80% confluence, the culture was washed with cold phosphate-buffered saline containing 1 mM NaF and 1 mM Na 3 VO 4 and lysed in SDS loading buffer. Aliquots containing 60 g of protein were subjected to Western blotting.

Assay of the Growth Inhibition by p204 in Monolayer Cultures (for Fig. 7B)
The Con, Ind-p204, and H-Ras Q61L -Ind-p204 lines were plated at 3 ϫ 10 5 cells/ml in 60-mm tissue culture dishes in growth medium, and, if so indicated, 1 M ponasterone A was added. The cultures were incubated for 72 h. After digestion with trypsin-EDTA, the cells were counted, using a microscope.

Assays of Anchorage-independent Growth in Soft Agar Were Performed as Described (52) (for Fig. 7, C and D)
In brief, the bottom agar layer contained 0.51% agar (Bacto-Agar; Difco) in DMEM (Invitrogen), 10% fetal bovine serum. The upper agar layer contained 0.34% agar in DMEM, 13.3% fetal bovine serum. If indicated, 4 M ponasterone A was added to the upper layer. AKR-2B-Ind-p204 or AKR-2B-H-Ras Q61L -Ind-p204 were plated in the upper layer at 10 4 cells/60-mm dish. Triplicate cultures were prepared for each cell type and treatment. Cells were cultured in 5% CO 2 in a well humidified incubator at 37°C for 2 weeks and photographed. Fig. 9) MEF3T3-Tet-off-H-Ras G12V , MEF3T3-Tet-off-p204, or MEF3T3-Tet-off-H-Ras G12V -p204 cell lines were incubated in the absence or presence of 20 M PD98059 or LY294002 dissolved in Me 2 SO or of Me 2 SO (serving as a control) in the absence of doxycycline for 36 -48 h. MEF3T3-Tet-off-H-Ras G12V cells were transfected with pCMV204 or pCMV204S179A and incubated in the absence of doxycycline for 48 h to induce H-Ras G12V . The cells were fixed and processed, using 3.7% formaldehyde with Alexa Fluor 555 phalloidin (Invitrogen), anti-p204 antibodies, and fluorescein isothiocyanate-or Texas Red-conjugated anti-rabbit IgG, as indicated in the figures. Fig. 11A (

b))
The chromatin was cross-linked with 3% formaldehyde. The pGL-Egr-1 reporter gene (in Fig. 11A (d)) was constructed by inserting a 453-bp PCR product (extending from nucleotide Ϫ6927 to Ϫ6474 in the Ifi204 gene 5Ј-flanking region) into the pGL3 vector. A control reporter gene with a mutated Egr-1 binding sequence (pGL-mEgr-1) was constructed, using the QuikChange kit. 3T3 cells were cotransfected with the plasmids specified. Luciferase activity was assayed after a 36-h incubation.

Mouse Embryonic Fibroblasts (MEFs) (for Fig. 12)
MEFs were prepared from LSL-KrasG12D mice (43) and littermate controls. In brief, 13.5-day-old embryos were isolated, minced with razor blades, trypsinized, and cultured as described previously (48). Cre recombinase was introduced into MEFs via retrovirus-mediated gene transfer of Cre-GFP. Infected cells were isolated by fluorescence-activated cell sorting. Cre-mediated deletion of the STOP cassette in LSL-KrasG12D was confirmed by PCR.

RESULTS
The Bulk of p204 Is Cytoplasmic in the Heart; Cytoplasmic Proteins Binding to p204-In mouse embryos, the level of p204 was low in the heart on embryonic day 10.5, and it greatly increased until birth (12). This high p204 level was maintained for at least 5 months (Fig. 1A). In tissue slices FIGURE 1. The level of p204 protein is similarly high in the hearts of 1-day to 5-month-old C129 mice. The bulk of p204 is cytoplasmic. A, hearts were obtained from mice of the ages specified. Total protein extracts were subjected to Western blotting using p204 antiserum. ␤-Tubulin was used as an internal control. B, tissue slices prepared from hearts of 3-month-old mice were stained with rabbit p204 antiserum and fluorescein isothiocyanate-tagged anti-rabbit IgG. Nuclei were stained with DAPI. For further details, see "Experimental Procedures." p204 Protein Is a Novel Modulator of Ras Activity from hearts of 3-month-old mice, the bulk of p204 was cytoplasmic (Fig. 1B). The translocation of p204 from the nucleus to the cytoplasm depends on a nuclear export signal (NES) in p204, and it takes place in the course of the differentiation of embryonal carcinoma stem cells to cardiac myocytes (14).
The proteins known to be bound by p204 at present are primarily nuclear. Wishing to identify cytoplasmic proteins interacting with p204, we performed a yeast two-hybrid assay using a cDNA library from hearts of 8 -12-week-old male mice. The assay revealed numerous cDNA clones encoding proteins interacting with p204. Many of these proteins were related to the cytoplasmic H-Ras protein (not shown).
Binding of H-Ras and K-Ras to p204-Testing the validity of the results of the two-hybrid assay, we performed a series of coimmunoprecipitation and pull-down assays (Fig. 2). Immunoprecipitation of a heart extract from 3-month-old mice with anti-p204 resulted in the coimmunoprecipitation of the endogenous H-Ras, together with the endogenous p204 ( Fig. 2A). Immunoprecipitation with anti-H-Ras coprecipitated p204, and immunoprecipitation with anti-p204 coprecipitated H-Ras from a lysate of FT9 cells (Fig. 2B). These cells were generated by transfection of approximately six copies of the H-ras oncogene into NIH3T3 cells (49). The experiments in Fig. 2C were performed using two murine primary adenocarcinoma cell lines (NB490 and NB508), which were established from freshly isolated tumor specimens (53,54). The generation of the tumors in mice involved the activation of expression at the endogenous level of a latent K-Ras G12D knock-in allele (LSL-K-Ras) by Cre-mediated excision of a transcriptional stopper element (43). As revealed by microassay-based comparative genomic hybridization, both lines should have nearly physiological levels of K-Ras expression. Line NB490 appears to have a single extra FIGURE 2. Binding of endogenous and ectopic p204 to endogenous and ectopic murine wild type and mutant activated Ras proteins. Coimmunoprecipitation and pull-down assays are shown. A, coimmunoprecipitation of H-Ras and p204 from the extracts of hearts of 3-month-old C129 mice. Hearts were washed, their ventricles were excised and disintegrated in MLB lysis buffer, and the debris was sedimented by centrifugation. Anti-p204 antiserum was used for immunoprecipitation from the supernatant fraction (5 mg of protein), and anti H-Ras antibodies were used for Western blotting. Preimmune IgG was used as a negative control. B, coimmunoprecipitation of endogenous p204 with endogenous H-Ras in lysates of cells from the murine FT9 line that was generated by transfection of NIH3T3 cells with a plasmid encoding activated H-Ras G12V . Top, coimmunoprecipitation with anti-H-Ras antibodies immobilized on protein A and Western blotting with anti-p204 antibodies. Bottom, coimmunoprecipitation with anti-p204 antibodies immobilized on protein A and Western blotting with anti-H-Ras antibodies. In both panels, preimmune IgG was used as a negative control, and the amounts of input p204 and H-Ras are indicated. C, the NB490 and NB508 mouse tumor cell lines (kindly provided by N. Bardeesy) were cultured in DMEM, 10% FCS. If so indicated (ϩ) 1000 units/ml ␣-interferon was added for 48 h. Coimmunoprecipitation from the culture lysates was performed with anti-K-Ras antibodies conjugated to protein G beads and Western blotting with anti-p204. As a negative control, preimmune IgG was used. As determined by microarraybased comparative genomic hybridization, the NB490 cells appear to have a single extra gene copy of either the K-Ras G12D allele or of wild-type K-Ras, and NB 508 has the wild-type gene copy number. 4 D, binding of p204 to K-Ras G12V and to three different mutants of K-Ras G12V in the K-Ras effector binding region (kindly provided by G. J. Clark). Plasmids encoding K-Ras G12V and the three mutants were transfected into different NIH3T3 cultures. Coimmunoprecipitation from culture lysates with anti K-Ras antibodies conjugated to protein G beads and Western blotting with anti-p204 antibodies. E, binding of purified MBP-p204 to purified GST-H-Ras G12V -GTP␥S. The pull-down assay used immobilized anti-MBP antibodies. GST-H-Ras G12V and MBP-p204, both from bacterial lysates, were purified on GST-beads and MBP-beads, respectively. GST-H-Ras G12Vconjugated beads were first incubated with GTP␥S and subsequently with MBP-p204 (which was eluted and concentrated from MBP beads) at 4°C overnight. The pulled down MBP-p204 was assayed by Western blotting with anti-MBP antibodies. As negative controls, MBP fused to Paramyosin, MBP, and MBP-p204 incubated with GST-beads were used. F, binding of H-RasGTP␥S to induced endogenous p204. Shown is a pull-down assay with immobilized anti-p204 antibodies and Western blotting with anti-H-Ras antibodies. Bottom, p204 was induced in a MEF3T3-Tet-off-p204 culture by incubation in DMEM without doxycycline. Western blotting is shown. Top, cultures were incubated without doxycycline to induce p204 and lysed, and aliquots from the lysate were incubated with GTP␥S or GDP. For the pull-down assay, aliquots were incubated as specified with immobilized anti-p204 antibodies (or as a negative control with immobilized mouse IgG). As positive controls for the pull-down assay, immobilized Raf-1 RBD was used. The pulled down H-Ras was detected by Western blotting with anti-H-Ras antibodies. For further details, see "Experimental Procedures." p204 Protein Is a Novel Modulator of Ras Activity FEBRUARY 29, 2008 • VOLUME 283 • NUMBER 9 copy of either the K-Ras G12D allele or of wild-type K-Ras, and line NB508 has the wild-type gene copy number. 4 Immunoprecipitation with anti-K-Ras (but not with preimmune IgG) of lysates from NB508 and NB490 cells coimmunoprecipitated p204 (Fig. 2C). Treatment with ␣-interferon increased the level of endogenous p204 only in the NB490 line. This resulted in an increase of the amount of p204 coimmunoprecipitated by anti-K-Ras. Anti-K-Ras also coimmunoprecipitated p204 from a lysate of NIH3T3 cells expressing ectopic K-Ras G12V oncoprotein (Fig. 2D).
Purified p204 (actually MBP-p204) could be pulled down by immobilized, purified H-Ras (actually GST-H-Ras G12V ) (Fig.  2E). This established that the binding of p204 to H-Ras did not require the involvement of further proteins.
p204 was bound to H-RasGTP (the activated form of H-Ras) but not to H-RasGDP (Fig. 2F, top). In the MEF3T3 cells, serving as the source of p204, endogenous p204 was induced by removal of doxycycline from the medium (Fig. 2F, bottom). As a positive control, the Ras-binding domain of the Raf-1 protein (Raf-1 RBD) was used (Fig. 2F, top). Raf-1 RBD is known to bind to RasGTP but not to RasGDP.
Mapping the Regions in p204 That Bind H-Ras-Aliquots of HA-H-Ras Q61L loaded with GTP␥S (a GTP derivative that is resistant to cleavage by RasGAP) were incubated with p204 ( Fig. 3A) or its various segments (Fig. 3, B-F) labeled with [ 35 S]methionine. Thereafter, the HA-H-Ras Q61L GTP was pulled down by immobilized anti-HA antibodies, and the associated labeled p204 or its segments were assayed by SDS-PAGE and autoradiography. The band patterns (Fig. 3, bottom) reveal that the N-terminal (N) p204 segment bound to Ras at least as well as complete p204. Linkage of N to the isolated a-segment of p204 alone impaired the binding (probably in consequence of altered conformation). Linkage of N to other segments of p204 had no inhibitory effect. These results suggested that the major H-Ras binding domain of p204 was in the N segment.
p204 Inhibited RasGAP Activity-The binding of p204 to RasGTP but not to RasGDP made it conceivable that p204 may have RasGAP activity (i.e. it could cleave RasGTP to RasGDP). Our assays revealed that this was not the case (data not shown). However, as shown in the autoradiograph in Fig. 4  p204 Inhibited the Binding of the Ras Effectors Raf-1, PI3K, and RalGDS to RasGTP-The binding of p204 to RasGTP, but not to RasGDP, suggested that the p204-binding domain(s) in  Ras might include the "effector loop" (55). The conformation of this loop is different in RasGTP from that in RasGDP.
The various downstream signaling pathways activated by RasGTP are initiated by "effector" proteins. The earliest identified of these proteins are Raf-1-kinase, PI3K, and Ral-GDS (37)(38)(39)(40)(41). The effector proteins bind to RasGTP but not to RasGDP. The regions in RasGTP to which the effector proteins bind include the effector loop (34). These considerations prompted us to test whether p204 (shown in Fig. 2F to also bind RasGTP) inhibited the binding of the effector proteins to RasGTP.
As expected, the binding of wild-type H-Ras to immobilized Raf-1 RBD was increased in a lysate from serum-starved cells if the cells were incubated with EGF prior to their lysis (Western blotting in Fig. 5A). This incubation with EGF is known to promote the replacement of the GDP bound to Ras by GTP (56). Adding purified p204 (actually MBP-p204) strongly decreased the amount of H-Ras bound to GST-Raf-1 RBD in a lysate from cells exposed to EGF, whereas the addition of a purified MBP control had no such effect. These results indicated that p204 inhibited the binding of the Raf-1 RBD domain to RasGTP (Fig. 5A).
In the experiments in Fig. 5, B and C, permanently activated Ras oncoprotein, (i.e. H-Ras Q61L ) was used. This oncoprotein is incapable of cleaving the GTP it binds to GDP. The inhibition by p204 of the binding to RasGTP of a second Ras effector, PI3K, was tested in Fig. 5B. This effector consists of a catalytic subunit (P110␣ forming a heterodimer with a regulatory subunit (P85␣)). We used an HA-tagged version of P110␣, allowing its immunoprecipitation with anti-HA antibodies. The data obtained revealed that purified MBP-p204 (but not MBP) inhibited the binding of H-RasGTP to PI3K.
The data in Fig. 5C revealed that the binding of HA-tagged, activated H-Ras Q61L GTP to a third Ras effector (i.e. Ral-GDS) (tagged with Myc) was also inhibited by MBP-p204. As a negative control for the specific antibodies used, we established that mouse IgG did not pull down PI3K-H-RasGTP (Fig. 5B), or Ral-GDS-H-RasGTP (Fig. 5C). These data revealed that the binding of p204 to RasGTP resulted in the inhibition of the binding of each of the above tested three Ras effectors to RasGTP.
As noted earlier, the regions in RasGTP to which the effector proteins bind include the effector loop. We tested by coimmunoprecipitation with anti-K-Ras the binding of p204 to three mutants of the activated K-Ras G12V protein (45). These mutants contain in their Ras effector binding segment (effector loop) single amino acid substitutions. The substitution in mutant (I) is E27G, in mutant (II) T35S, and in mutant (III) Y40C. Among these three K-Ras G12V mutants, (I) was not bound by the PI3K and Raf effector proteins, (II) was not bound by the PI3K and RalGDS effector proteins, and (III) was not bound by the Raf and RalGDS effector proteins (57). As revealed in Fig. 2D, p204 bound each of these K-Ras G12V mutants, similarly to its binding of (wild-type) K-Ras G12V . These findings suggest that the binding of p204 to activated K-Ras G12V is distinct from that of Raf-1, PI3K, and Ral-GDS, the three Ras effector proteins tested. (See also Fig. S1.) At least one H-Ras mutant, H-Ras (S17N), was found to bind p204 much more weakly than wild-type H-Ras (Fig. S1). H-Ras (S17N) is a dominant negative mutant known to inhibit the activity of wild-type H-Ras (58). p204 Protein Is a Novel Modulator of Ras Activity FEBRUARY 29, 2008 • VOLUME 283 • NUMBER 9 The weaker binding of p204 to this mutant than to wild-type H-Ras is probably due in part to the fact that the mutant's affinity to GTP is weaker than that of wild-type H-Ras (and as shown in Fig. 2F, p204 binds to RasGTP).
An Increase in p204 Level Inhibited, whereas a Decrease in Endogenous p204 Level Enhanced the Phosphorylation of the MEK, Akt, and p38 MAPK Mediators of Ras Signaling-We explored whether the inhibition by p204 of the binding of the Ras effectors to RasGTP (Fig. 5) inhibited the signaling through the pathways initiated by the effectors tested. The results in Fig.  6 revealed that this was the case. The experiments involved cells from a line (AKR-2B Ind-p204) in which the level of endoge-nous p204 could be induced by ponasterone A (Fig. 6, left). The level of p204 induced by ponasterone A was physiological (i.e. not higher than that obtained by induction with the physiological p204 inducer ␣-interferon) (Fig. 6A, two left bottom panels of Western blotting and the bottom right diagram). As a control, we used a cell line (AKR-2B Ind-Con cells) in which the level of p204 was unaffected by ponasterone A (Fig. 6, right). To increase the RasGTP level in the cells, they were serum-starved and then exposed to EGF for 10 min prior to testing the levels of the various proteins by Western blotting.
As shown in Fig. 5A, p204 inhibited the binding to RasGTP of the Ras effector Raf-1 kinase RBD.
When not bound to RasGTP, Raf-1 kinase is cytosolic and inactive. Ras-GTP is attached to the cell membrane, thus, by binding to it, Raf-1 kinase also becomes membrane-attached. This results in its phosphorylation and the activation of its kinase. The target of the activated Raf-1 kinase that it phosphorylates is MEK, which is converted to phospho-MEK (dual specificity protein kinase). The two top panels of Western blots and the top diagram of Fig.  6A revealed that the inhibition of the binding of Raf-1 kinase to Ras-GTP by p204 decreased the extent of phospho-MEK formation by ϳ40% without affecting the level of total MEK protein.
The Western blots also revealed that in cells not exposed to EGF no formation of phospho-MEK was detected. This was presumably in consequence of the very low level of RasGTP in the serum-starved cells (compare the leftmost and the middle panels in the Western blots assayed with anti-phospho-MEK).
The second Ras effector, whose binding to RasGTP was shown to be inhibited by p204 in Fig. 5, is PI3K (39). When activated by binding to RasGTP, PI3K phosphorylates a phospholipid (i.e. a membrane-embedded phosphatidylinositoldiphosphate). The product of this phosphorylation, the membrane-associated phosphatidylinositoltriphosphate, attracts a serine-threonine kinase designated as Akt (or PKB). The binding of Akt to the phosphatidylinositoltriphosphate moiety results in the phosphorylation of Akt (i.e. phospho-Akt) (59). As shown in Fig. 6A, the formation of phospho-Akt decreased in  1, 2, 3, and 4 in the anti-p204 panel). Cells from both lines were serum-starved and, if so indicated, were exposed to ponasterone A for 48 h, and to EGF for 30 min before lysing the cells in SDS loading buffer and determining the levels of MEK, Akt, p38 MAPK , phospho-MEK, phospho-Akt, phospho-p38 MAPK , and p204 by Western blotting with appropriate antibodies. The level of p204 was also assayed as above in an AKR-2B Ind-p204 culture exposed to ␣-interferon. Bottom left panels, the Western blots were scanned and graphed using ImageQuant 5.2 Excel software and are shown in the right diagrams. The bar numbers in the right diagrams correspond to the numbers below the columns of the Western blots. The two lowest panels compare the levels of p204 in AKR-2B Ind-p204 cells exposed to ponasterone A (2) with that in the same type of cells but exposed to 1000 units/ml ␣-interferon (5); the latter level was taken as 1. B, NB508 murine primary adenocarcinoma cells contain a single copy of K-Ras G12D allele and a single copy of wild-type K-Ras allele. One NB508 culture (204AS) was transfected with a pCMV204AS plasmid (encoding 204 antisense RNA), whereas a control NB508 culture (Con) was transfected with a pCMV vector. Stably transfected cell clones were generated after selection with hygromycin. After incubation, cells were harvested, and the levels of p204, phospho-MEK, total MEK, phospho-Akt, and total Akt were determined in the cell lysates by Western blotting with appropriate antibodies (left). The Western blots were scanned and graphed, and the diagrams are shown (right). S.D. values are indicated. For further details, see "Experimental Procedures."

p204 Protein Is a Novel Modulator of Ras Activity
consequence of the inhibition of the binding of PI3K to RasGTP by p204 by about 40% with no decrease in the level of total Akt protein. Fig. 6A revealed that the induction of p204 also inhibited the phosphorylation of a further protein, p38 MAPK , by about 50% without affecting the total level of p38 MAPK protein. p38 MAPK is also among the mediators of ras oncogene action. It was reported to be phosphorylated and activated in H-ras oncogene-transformed MCF10A human breast epithelial cell lines in consequence of the cooperative functioning of the PI3K and Rac pathways (60). p38 MAPK can, however, also be activated by tumor necrosis factor in endothelial cells by a complex pathway involving, among others, AIP (a RasGAP family member) (61).
The Western blots in Fig. 6B reveal that the expression of 204AS (i.e. 204 antisense RNA) decreased the level of endogenous p204 in transfected NB508 cells (top). This decrease in turn was correlated with an increase in the phosphorylation of two mediators of Ras signaling, MEK and Akt, but with no change in the levels of total MEK and total Akt proteins (bottom four panels of Western blots). The extents of the decrease in p204 level and of the increase in phospho-MEK and phospho-Akt levels are shown in the diagrams on the right. Thus, the results in Fig. 6 revealed that an increase in p204 level inhibits, whereas a decrease in p204 level enhances, the phosphorylation (i.e. the activation) of Ras signaling by at least two mediators.
p204 Inhibited the Anchorage-independent Growth of Rastransformed Cells-The inhibition of various Ras signaling pathways by p204 prompted us to explore the effects of p204 on the proliferation of H-ras Q61L -transformed cells in different growth conditions. The experiments involved the use of AKR-2B control cells (Con), AKR-2B Ind-p204 cells in which the expression of p204 was inducible by ponasterone A (Ind-p204), and derivatives of this cell line expressing ectopic H-Ras Q61L (H-Ras Q61L -Ind-p204) (Fig. 7A). Fig. 7B shows that the induction of p204 by ponasterone A resulted in an ϳ27% decrease in the rate of proliferation in monolayer culture of Ind-p204 cells and also of the H-Ras Q61L -Ind-p204 cells. Thus, the extent of inhibition by p204 of the proliferation in monolayer cultures was not significantly affected in our conditions by the expression of ectopic H-Ras Q61L .
When grown in suspension culture, the expression of H-Ras Q61L oncoprotein greatly increased colony formation (Fig. 7C, compare lower panels 1 and 3) (57, 62). In the absence of H-Ras Q61L , only microcolonies were formed, whereas, in its presence, numerous colonies with a diameter over 1.5 m were formed. This was expected, since, for proliferation, most types of untransformed cells require "anchorage," (i.e. the attachment of their cell surface integrin receptors to the extracellular matrix). This attachment was found to generate signals that activate the Ras pathway. It is in line with this finding that the expression of the constitutively activated H-Ras oncoprotein allowed anchorage-independent cell proliferation.
The induction of p204 strongly (by ϳ66%) inhibited the anchorage-independent growth of the H-Ras Q61L -Ind-p204 cells (Fig. 7C, compare panels 3 and 4; see also Fig. 7D). This inhibition is in accord with the inhibition by p204 of the signaling triggered by RasGTP (Fig. 6). Fig. 8) using monkey COS-7 cells expressing ectopic wild type or mutated murine p204 proteins. Sequences with 3 or more basic residues that might serve as an NLS in p204 are shown in boldface type in Fig. 8A. The arginine and lysine residues in three of these sequences were individually replaced by glutamine residues, and the effects of the replacements on the subcellular location of p204 are shown in Fig. 8, B, C, and D, right. The results established that the NLS determining the nuclear localization is the KKSKAAK sequence extending in p204 from amino acid 188 to 194 (Fig. 8D).

Ras Oncoprotein Promoted a PI3K-dependent Translocation of p204 from the Nucleus to the Cytoplasm-The identity of a nuclear localization signal (NLS) (63) in p204 was established in experiments (
The translocation of p204 from the nucleus to the cytoplasm during the differentiation of various cell types (12,14) and the finding that p204 can modulate the signaling by the cytoplasmic Ras protein (this study) prompted us to study whether the expression of activated Ras can affect the subcellular location of p204.
For the study, we used three cell lines derived from MEF3T3. The removal of doxycycline (Tet) from the medium induced (i) in the MEF3T3-Tet-off-p204 line, the expression of p204, (ii) in the MEF3T3-Tet-off-Ras G12V line, the expression of activated H-Ras, and (iii) in the MEF3T3-Tet-off-Ras G12V -p204 line, the expression of both activated H-Ras and p204 (Fig. 9A).
The selective induction of p204 (in MEF3T3-Tet-off-p204) resulted in its expression in the nucleus (Fig. 9B); the induction of both p204 and activated H-Ras (in MEF3T3-Tet-off-Ras G12V -p204) resulted in p204 distributed between the nucleus and the cytoplasm (Fig. 9C, top). The induction of H-Ras and p204 (in MEF3T3-Tet-off-Ras G12V -p204), in the presence of a selective inhibitor of the PI3K pathway (LY294002) (64), however, resulted in the exclusively nuclear localization of p204 (middle). A different inhibitor (PD98059), blocking the signaling by the Raf1-MEK-ERK signaling pathway, had no effect on the distribution of p204 between the nucleus and the cytoplasm when induced together with Ras in the MEF3T3-Tet-off-Ras G12V -p204 line (bottom).
These results indicate that activated Ras affected the subcellular distribution of p204 by triggering the appearance of a portion of p204 in the cytoplasm. Moreover, this effect of activated Ras required signaling by the PI3K pathway.
To generate more quantitative data on the effect of H-Ras G12D induction on the distribution of p204 among different subcellular locations than obtained by visualization based on immunofluorescence (Fig. 9), we used Western blotting (Fig.  S4). We fractionated cell lysates from MEF3T3-Tet-off-Ras G12V cultures transfected with pCMV204 without, or after, inducing Ras G12V expression (by removal of doxycycline) into nuclear, cytoplasmic, and crude membrane fractions and performed Western blotting using anti-p204. The results in Fig. S4 confirm and extend the conclusions in Fig. 9. They reveal that (i) prior to induction of Ras G12V , the bulk of p204 is in the nuclear fraction, its presence in the cytoplasmic fraction is barely detectable, and none is detected in the crude membrane fraction; (ii) after induction of Ras G12V , the majority of p204 is in the cytoplasmic fraction, and a lesser but significant amount of p204 is in the crude membrane fraction. FEBRUARY 29, 2008 • VOLUME 283 • NUMBER 9

p204 Protein Is a Novel Modulator of Ras Activity
In the case of the translocation of p204 from the nucleus to the cytoplasm in the course of skeletal muscle myoblast and cardiac myocyte differentiation, the translocation was accompanied (possibly triggered) by the phosphorylation of p204 (12,15). As shown in Fig. S5, the translocation elicited by Ras also appears to be accompanied by posttranslational modification, including phosphorylation of p204. We did not attempt to identify the amino acid residue(s) phosphorylated in p204 in the course of the activated Ras-mediated translocation. We established, however (in experiments involving the expression of ectopic, wild-type, and mutant p204), that the replacement of serine at position 179 of p204 by alanine prevented the translocation from the nucleus to the cytoplasm, as triggered by activated Ras (Fig.  9D). The sequences at the 5Ј and 3Ј sides of serine 179 in p204 indicate that it is a potential target for phosphorylation by several kinases, including Akt (which is activated by PI3K, whose activity, as noted, is a prerequisite for the translocation) as well as calcium/calmodulindependent protein kinase (CaMK) and protein kinase A (65).  Fig. 10A, a and b). The induction of p204 had little, if any, effect on either (i) or (ii) (cf. c and d). However, the coinduction of p204 with H-Ras G12V blocked (i) as well as (ii) (i.e. both effects of H-Ras G12V (Fig. 10B, a and b). The p204 (coinduced with H-Ras G12V ) was distributed between the nuclei and the cytoplasm (a), and its location in the cytoplasm generally overlapped with that of the actin cytoskeleton (c).

Induction of Ras Oncoproteins in Cells Promoted an Alteration of the Actin Cytoskeleton and an
The effects of H-Ras G12V on the actin cytoskeleton in Fig.  10A were as reported in numerous earlier publications (e.g. see Refs. 66 -69). The inhibition by p204 of the effect of H-Ras G12V on the organization of the cytoskeleton is in accord with the binding of p204 to H-Ras G12V (Fig. 2E) as well as the inhibition by p204 of the signaling by H-Ras oncoprotein through several signaling pathways (e.g. Ras-Raf, Ras-PI3K, and Ras-RalGDS) (Figs. 5 and 6).
Ras and various components of its signaling pathways were reported to promote cell motility (70 -72). Fig. 10C (a and b) demonstrates a cell motility test based on the assaying of the number of cells passing through a membrane. It revealed that the induction of the H-Ras G12V oncoprotein (in MEF3T3-Tetoff-H-Ras G12V cells) increased the number of transmembrane cells ϳ5-fold (cf. a and b; see also the bottom diagram). The simultaneous induction of p204 with H-Ras G12V decreased by ϳ70% the H-Ras G12V -promoted increase in the number of transmembrane cells (cf. b and d; see also the bottom diagram). This inhibition by p204 of the H-Ras G12V -promoted cell motility is also in accord with the inhibition by p204 of H-Ras signaling (Figs. 5 and 6).
The Egr-1 Transcription Factor (Whose Expression Is Known to Be Promoted by Activated Ras) Promoted the Expression of p204-Ras oncoprotein or wild-type Ras protein, when activated in cells exposed to growth factors, was reported to  Left, sequences with three or more basic residues that are potential NLS-s are shown in boldface letters, above the schematic diagrams of p204 protein in A-D. The NES and the NLS identified are shown. The mutated amino acid residues (in which Lys residues were substituted by Gln residues) are printed in gray letters. (The significance of the KRKSMR sequence for subcellular localization is examined in Fig. 9D). Right, subcellular localization of wild type p204 (A) and of the mutated p204 protein with the amino acid substitutions indicated in the left panels (B-D). Shown is staining with antibodies to p204 and Tex-conjugated IgG (red) and of nuclei with DAPI (blue). For further details, see "Experimental Procedures." p204 Protein Is a Novel Modulator of Ras Activity FEBRUARY 29, 2008 • VOLUME 283 • NUMBER 9 JOURNAL OF BIOLOGICAL CHEMISTRY 5841 increase the expression of the Egr-1 transcription factor (42,73). We noted in the Ifi204 gene a binding sequence for Egr-1 (designated as S2 in Fig. 11A, a). A chromatin immunoprecipitation assay with anti-Egr-1 antibodies (but not preimmune IgG) revealed the coprecipitation of a chromatin segment from Ifi204 that was shown by PCR to include the region with the Egr-1 binding S2 sequence (Fig.  11A, b).
The transfection of an Egr-1 expression plasmid (pCMV-Egr-1) into a 3T3 culture (74) resulted in a strong increase in the level of endogenous p204 protein (Fig. 11A, d, top Western blot). The expression of a reporter construct (pGL-Egr-1) driving Luciferase and including the Egr-1-binding S2 sequence from the Ifi204 gene (but not that of a control reporter with a mutated Egr-1 binding sequence, pGLm-Egr-1) was strongly increased by the transfection of pCMV-Egr-1 into the 3T3 culture (Fig. 11A, d, right and left  diagrams).
The presence of a functional Egr-1 binding sequence in the Ifi204 gene, whose expression was increased by ectopic Egr-1, together with the report (42) that activated H-Ras can increase the expression of Egr-1, prompted us to test whether the introduction into the 3T3 culture of activated H-Ras increases the level of endogenous p204. The data in Fig. 11B revealed that it did. Transfection into the 3T3 culture of the pCGN-H-Ras Q61L expression plasmid (but not of the pCGN vector) increased the level of p204 (as well as of H-Ras) proteins (Fig. 11B, Western blot and  diagram).
The transient activation of wildtype Ras (i.e. the conversion of Ras-GDP to RasGTP) in a serum-starved 3T3 culture by PDGF (Fig. 11C, top left, GST-Raf-1-RBD pull down assay followed by Western blotting with anti-H-Ras) resulted in an increase first in the level of Egr-1 protein and, after a delay, of p204 protein (left three Western blots and right diagrams).

p204 Protein Is a Novel Modulator of Ras Activity
All of the above observations are in accord with the following interpretation: the increase in the level of activated Ras protein triggered (e.g. by PDGF) resulted in an increase in the level of the Egr-1 transcription factor, which, in turn, increased the level of p204 expression. It is possible that besides Egr-1 also further transcription factors or other agents may be involved in mediating the increase in p204 level promoted by activated Ras.
The above findings prompted us to test whether the induction of endogenous p204 (by removing doxycycline from the medium of a culture of MEF3T3-Tet-off p204 cells) affects the level of H-Ras. The data in Fig. 11D reveal no such effect. It is consistent with this result that p204 (which was shown to promote the proteasomic degradation of Id proteins in consequence of promoting their ubiquitination (15)) did not promote the degradation of H-Ras (not shown).
Physiological Levels of Activated K-Ras G12D Induced the Expression of p204 and Its Partial Translocation from the Nucleus to the Cytoplasm-The data in Fig. 11B revealed that expression of ectopic H-Ras Q61D (in cells transfected with an H-Ras Q61D expression plasmid) induced p204 expression. We wished to establish whether a physiological level of Ras oncoprotein present in cells carrying only a single copy of the ras oncogene would also result in an increased p204 level. For these studies, N. Bardeesy kindly provided MEF cultures. The MEFs were prepared from mice with the latent LSL-K-ras G12D allele (43) and from littermate controls. The LSL-K-ras G12D allele is not normally expressed due to the presence of a floxed stop cassette but can be activated by expression of Cre recombinase in these cells. Wild type control cells (11.1) infected with a Cre-GFP retrovirus (11.1wtϩCre) and LSL-K-ras G12D cells without Cre (11.7-Cre) expressed comparably low p204 levels, and exposure to 1000 units/ml interferon strongly increased this level. In contrast, LSL-K-ras G12D cells induced to express endogenous levels of K-Ras G12D (11.7ϩCre) (by infection with a Cre-GFP retrovirus, resulting in the deletion of the floxed stop cassette) exhibited greatly increased p204 levels; exposure of the culture to 1000 units/ml interferon increased this level only slightly (Fig. 12, six bottom panels). A quantitative comparison p204 Protein Is a Novel Modulator of Ras Activity FEBRUARY 29, 2008 • VOLUME 283 • NUMBER 9 of the p204 levels in the cultures are shown in the Western blots and the diagram of Fig. 12. These revealed an ϳ5.5-fold increase in the p204 level in consequence of the expression of the single copy of the K-ras G12D oncogene. This increase somewhat exceeded that triggered by 1000 units/ml interferon in the 11.1wtϩCre and the 11.7-Cre cultures. These results established that the presence of a single copy of the K-ras G12D oncogene in a cell was sufficient to strongly boost p204 expression.

DISCUSSION
p204 was first found in the nucleolar and nucleoplasmic compartments of various cultured cells (75). In the nucleolus, it inhibits ribosomal RNA synthesis by blocking the activity of the UBF1 ribosomal RNA transcription factor (19). Its effects in the nucleoplasm include, for example, the enhancement of transcription by the Cbfa1 transcription factor by forming a ternary complex with pRb and Cbfa1 (17). During the differentiation of cultured myoblasts to myotubes and of cultured P19 embryonal carcinoma stem cells to cardiac-type myocytes, part of p204 is translocated from the nucleus to the cytoplasm (12,14). Moreover, the bulk of p204 is cytoplasmic in isolated cardiac myocytes (12) and in heart slices from mice (Fig. 1B). A yeast two-hybrid assay (not shown) identified Ras, among other cytoplasmic proteins, that binds to p204. Pull-down assays (Fig. 2E) established that the binding between purified p204 and purified Ras was direct. The binding of endogenous p204 to endogenous Ras in mouse heart extract ( Fig. 2A) and two murine tumor cell lines (Fig. 2C) revealed the physiological relevance of the binding. p204 bound both H-Ras (Fig. 2, A and B) and K-Ras (Fig. 2, C and D).
The fact that p204 bound to H-RasGTP (but not to H-Ras-GDP) (Fig. 2F) supported the assumption that, similarly to RasGAP and various Ras effectors, p204 might also bind to the Ras effector domain (although not necessarily exclusively). The validity of this assumption was strengthened by the finding that p204 inhibited the activity of RasGAP (Figs. 4B and S3) and also inhibited the binding to H-Ras of the Ras effectors (i) Raf-1 (actually, Raf-1 RBD was tested, since Raf-1 also binds a second domain in H-Ras) (76), (ii) PI3K, and (iii) RalGDS (Fig. 5, A-C).
Induction of p204 also inhibited signaling (as assayed by testing for target phosphorylation) by some Ras effectors including (i) phosphorylation of MEK promoted by Raf-1, (ii) phospho-p204 Protein Is a Novel Modulator of Ras Activity rylation of Akt promoted indirectly by PI3K, and (iii) phosphorylation of p38 MAPK as promoted directly or indirectly by several of the Ras signaling proteins Tiam, Rac, PI3K, and MKK3/6 ( Fig. 6A). A decrease in the level of endogenous p204, however, resulted in an increase of the phosphorylation of MEK and Akt (Fig. 6B). In addition to Raf-1, PI3K, and RalGDS, several further Ras effectors were identified (35,36). It remains to be established which of these are prone to inhibition by p204.
A comparison of the binding to H-RasGTP of p204 and its various segments revealed that both p204 and its N segment bind to H-RasGTP. The N segment includes among other regions two characteristic domains: (i) an ϳ95-amino acid-long sequence of the PAAD/DAPIN/Pyrin domain type (77) (such domains occur at the N termini of several proteins involved in apoptosis and inflammatory signaling pathways) and (ii) a segment consisting of eight imperfect repeats of a 7-amino acid sequence (7). It remains to be seen whether (i) or (ii) are involved in the binding of p204 to RasGTP.
The induction of p204 inhibited the anchorage-independent growth of H-Ras Q61L -expressing cells in suspension culture much more strongly than their growth in monolayer culture (Fig.  7, B-D). This is in accord with (i) the strong dependence of anchorage-independent growth (but not growth in monolayer culture) on the constitutive expression of Ras oncoproteins (57,62) and (ii) the inhibition of Ras signaling by p204 (Figs. 5 and 6).
The condition-dependent nuclear localization of p204 in various cell types prompted us to identify its NLS. This was found to be the KKSKAAK sequence extending from amino acid 188 to 194 in the p204 protein (Fig. 8). This sequence fits the conclusions of a structural study concerning the key requirements for a monopartite NLS; it contains Lys residues in its positions 1, 2, and 4 (of these, Lys in positions 2 and 4 could be substituted by Arg) (78,79). We did not rule out the possibility, however, that the KRKS sequence extending from amino acid 176 to 179 in p204 may be part of a bipartite NLS sequence whose second part might be the above discussed KKSKAAK sequence.
The induction of constitutively active Ras oncoprotein in MEF-3T3-Tet-off-Ras G12V cells resulted in the translocation of the bulk of p204 from the nucleus to the cytoplasm (Fig. 9C). (A more quantitative demonstration of the redistribution of p204 from the nuclear to the cytoplasmic and crude membrane frac-tions is shown in the Western blots of Fig. S4.) The translocation of p204 from the nucleus did not take place (i) in the presence of an inhibitor of PI3K (Fig. 9C) or (ii) if Ser 179 in p204 was replaced by Ala (Fig. 9D). As noted earlier, the Ser 179 is part of a KRKSMR sequence. It is a potential target of phosphorylation by cyclic AMP-activated protein kinase, by CaMK, and by Akt (that is activated by PI3K) (65). p204 contains several additional sequences that are potential targets of the above kinases.
It may be relevant that the deletion from histone deacetylase 5 of two consensus sequences of the above type that are targets of phosphorylation by, among others, CaMK, inhibited the CaMK-mediated nuclear export of histone deacetylase 5 (80).
The data in Fig. S5 reveal that expression of ectopic H-Ras G12V triggered the phosphorylation of p204. We have made no attempt to identify the site(s) phosphorylated in p204, the kinase(s) responsible, or other posttranslational modifications (if any) triggered in p204 by activated Ras. We established, however, that the induction of the catalytic subunit of cyclic AMP-activated protein kinase in NIH3T3 cells or the transfection of CaMK into cultured C2C12 myoblasts resulted in the translocation of p204 from the nucleus to the cytoplasm (data not shown).
This translocation of p204 occurs also in the course of the differentiation of C2C12 myoblasts (12) to myotubes and of P19 embryonal carcinoma stem cells to beating cardiac-type myocytes (14). p204 contains a classical NES (whose position is indicated in Fig. 8) known to be required for the translocation of p204 from the nucleus to the cytoplasm. Ectopic (wild-type) p204 promoted the differentiation of P19 cells to cardiac-type myocytes, whereas p204 from which the NES was deleted or 3 of the Leu residues in NES were replaced by Ala residues did not promote it (14). Moreover, p204 isolated from undifferentiated myoblasts was unphosphorylated, whereas p204 from the cytoplasmic fraction (but not from the nuclear fraction) of differentiated myotubes was phosphorylated (12). p204 was also found to promote the translocation from the nucleus to the cytoplasm of Id proteins (15). p204 binds the Id proteins and accelerates their degradation by promoting their ubiquitination (15). It is conceivable that the inhibition of Ras activity by p204 in the FIGURE 11. The Ifi204 gene encoding p204 contains an Egr-1 transcription factor binding sequence, and Egr-1 promoted p204 expression. Ectopic Ras oncoprotein, known to increase the expression of Egr-1, also boosted the expression of p204. By activating Ras and increasing Egr-1 expression, also PDGF promoted the expression of p204. Induction of endogenous p204 had little or no effect on the level of H-Ras. A (a), schematic structure of the Ifi204 gene with an Egr-1 binding sequence (S2) and a control sequence (not binding Egr-1) (S1); b, binding of Egr-1 to an Ifi204 gene segment including S2. Shown is a chromatin immunoprecipitation assay. Wild-type 3T3 cells were transfected with pCMV-Egr-1 and cross-linked with formaldehyde. c, purified genomic DNA was sonicated to generate 200 -1000 base pair segments. b, chromatin immunoprecipitation was performed using anti-Egr-1 antibodies or preimmune IgG. The immunoprecipitates specifically pulled down by anti-EGR-1 antibodies were assayed using PCR to amplify the Ifi204 S2 region with the Egr-1 binding site (shown in a). d, bottom right, the reporter gene (pGL-Egr-1) driving luciferase expression included the Egr-1-binding S2 segment from the Ifi204 gene; a control reporter gene (pGLmEgr-1) contained a mutated version of the Egr-1 binding sequence. Bottom left, ectopic Egr-1 strongly increased luciferase activity encoded by the pGL-Egr-1 reporter but not by the mutated pGL-mEgr-1 reporter. d (top), ectopic Egr-1 (encoded by pCMV-Egr-1) promoted the expression of endogenous p204 in 3T3 cells. Western blotting using anti-p204 antibodies. B (left), transfection of a pCGN-H-Ras Q61L plasmid (but not of the pCGN vector) into 3T3 cells increased the expression of p204 (and H-Ras) proteins in a dosage-dependent manner. As an internal control, ␤-tubulin was used. Western blotting with appropriate antibodies is shown. Right, diagram showing the plasmid-dosage dependent increase in the H-Ras and p204 protein levels based on the scanning of the Western blots. C, in a serum-starved 3T3 culture, PDGF increased the H-RasGTP level transiently. Shown is an assay by GST-Raf-1 RBD pull down (left top). In the same conditions, PDGF also increased the levels of the Egr-1 and p204 proteins. Levels were assayed hourly for 5 h by Western blotting with appropriate antibodies, using ␤-tubulin as an internal control (left bottom). The time course of the increase in the levels of the Egr-1 and p204 proteins is shown on the right, as obtained by scanning of the left bottom panels. D, two cultures of MEF3T3-Tet-off-p204 cells were incubated in the presence of doxycycline until reaching 75% confluence. At this time, one of the cultures was further incubated in fresh medium with doxycycline (Tet-on), control culture, and the second culture also in fresh medium, but without doxycycline (Tet-off) to induce p204. After 48 h, the cultures were lysed, and the levels of p204, H-Ras, and (as an internal control protein) ␤-actin were assayed by Western blotting. For further details, see "Experimental Procedures." cytoplasm may contribute to the promotion of differentiation by p204.
The induction of H-Ras G12V in MEF3T3 cells resulted in a profound change in cell morphology (Fig. 10A, a and b). This was primarily in consequence of the rearrangement of the actin cytoskeleton triggered by the signaling of activated Ras (66, 68). Ras signaling via the Raf pathway promotes the activation of ERK1/2; Ras signaling via the PI3K-phosphatidylinositoltriphosphate pathway promotes the activation, among others, of RhoGEF; and Ras signaling via the RalGEF pathway promotes the activation of the Rac and Cdc2 proteins (70,72).
The ERK1/2, Rho, Rac, and Cdc2 family GTPases, among others, in turn, promote changes in the actin cytoskeleton and in the adhesion to the extracellular matrix (62,71). They also elicit extensions of fingerlike filopodia and broad lamellipodiatype ruffles from the plasma membrane (66, 67). All of the above effects of signaling by Ras contribute to the alterations of cell morphology (cf. Fig. 9A, a and b).
Simultaneous induction of p204 with H-Ras G12V inhibited the above changes (Fig. 10B). This inhibitory effect of p204 is consistent with its inhibition of signaling by Ras via several pathways (Figs. 5 and 6).
Ras oncoprotein is known to enhance cell motility, which promotes, among other activities, the invasiveness and metastatic activity of tumor cells (70,71). Cell motility requires constant changes in the cell morphology (i.e. constant restructuring of the actin cytoskeleton in different parts of the cell, together with constant making and breaking of the attachments of different portions of the cell to the extracellular membrane). Consequently, the same proteins that are activated by Ras signaling pathways and modulate the cell morphology are also involved in enhancing cell motility.
Oncogenic Ras G12V increased the motility of MEF3T3 cells, as determined in a transmembrane migration assay, ϳ5-fold, and the induction of p204 diminished this increase by ϳ70% (Fig. 10C). This effect of p204 was in accord with its inhibition of Ras signaling.
The finding of a binding sequence for the Egr-1 transcription factor (81) in the Ifi204 gene prompted us to verify (i) the activity of this sequence when inserted into a reporter construct driven by Egr-1 and (ii) the induction of endogenous p204 by ectopic Egr-1. The verification of (i) and (ii) (Fig. 11A (d)) and the fact that Egr-1 expression can be activated by Ras oncoprotein and also by PDGF (primarily via the Ras-Raf-MEK-ERK pathway) (42,73,82) prompted us to test whether activated Ras Q61L or PDGF does induce the expression of p204. The results in Fig. 10, B and C, established that both did. An increase in the level of p202 (7), a sister protein of p204, was reported in H-ras oncogene-transformed NIH3T3 cells (83).
Egr-1 is known to induce the expression of its corepressor Nab2 (84). This induction can be accounted for by the presence of a series of Egr-1 binding sequences in the promoter of the Nab2 gene. Nab2 protein expression follows that of Egr-1 by several h. Nab2 binds and inhibits the activity of Egr-1, thus making the expression of the Egr-1 target genes transient.
The findings that (i) activated Ras promoted the expression of p204 (via the Ras-Egr-1-p204 pathway) and (ii) p204 inhibited Ras signaling (this study) suggest that p204 might be considered as one of the negative feedback inhibitors of Ras. It should be noted that the induction of p204 had little or no effect on the level of H-Ras (Fig. 11D).
Egr-1 plays a master regulatory role in multiple cardiovascular pathological processes (85). It remains to be established whether Egr-1 activity contributes to the high level of p204 in heart tissue. FIGURE 12. Activation of the expression of a single copy of the K-ras G12D oncogene triggered an increase in the level of p204 as well as the translocation of part of p204 from the nucleus to the cytoplasm in cultured murine embryonic cells. Shown is an assay by immunofluorescence and Western blotting. The following three cultures of murine embryo fibroblasts were generated and kindly provided by Nabeel Bardeesy and Satish Sankaran (MGH Cancer Center, Harvard Medical School). The cultures were prepared from day 13.5 embryos from the FVBn strain. 11.1wt fibroblasts were from an embryo carrying only wild-type K-ras genes. 11.7-Cre cells were from an embryo in which one copy of the K-ras G12D oncogene was present. In the 11.7-Cre fibroblasts, the coding region of the K-ras G12D gene included a transcriptional stopper element (between two lox sequences), and thus the K-Ras G12D protein was not expressed. Infection of 11.7-Cre fibroblasts (but not of 11.1wt fibroblasts) with a viral Cre expression plasmid resulted in the deletion of the stopper element and thus gave rise to the 11.7ϩCre fibroblasts in which the K-Ras G12D oncoprotein was expressed. The 11.1wtϩCre, 11.7-Cre, and 11.7ϩCre fibroblasts were cultured without (Ϫ) or in the presence of 1000 units/ml ␣-interferon (ϩ) for 48 h and were fixed with Ϫ20°C methanol, acetone (1:1). The fibroblasts were incubated with anti-p204 antiserum diluted 1:200 and stained with fluorescein isothiocyanate-anti-rabbit IgG diluted 1:200 or stained with DAPI (top). For Western blotting, the three types of fibroblasts were cultured as above and were lysed in MLB lysis buffer. 40 g of total protein samples were subjected to SDS-PAGE and Western blotting using anti-p204 antiserum and, as an internal control, anti-␤-actin (bottom left). The diagram (bottom right) shows the levels of p204 in the three types of fibroblasts without or after exposure to ␣-interferon. The S.D. values are indicated. For further details, see "Experimental Procedures."

p204 Protein Is a Novel Modulator of Ras Activity
The triggering by the ras oncogene of (i) the translocation of p204 from the nucleus to the cytoplasm ( Fig. 9) and (ii) the induction of p204 expression (Fig. 11) were tested by using induced (Fig. 9) or ectopic ras oncogene (Fig. 11). To extend the relevance of these observations to a case analogous to that in which a single copy of the ras oncogene arises in a cell (in vivo), we performed the following experiments (Fig. 12). In these, the effect of the removal (by Cre) of a transcriptional stopper from a single copy of K-ras oncogene in murine embryonic cells was tested on p204 expression and subcellular localization. This removal (which allowed the expression of the oncogene) triggered the expression of p204 at a level similar to that of the exposure of the culture to 1000 units/ml ␣-interferon as well as the redistribution of p204 between the nucleus and the cytoplasm.
These results suggest that "physiological" levels of ras oncogene can trigger the expression of high levels of p204, and thus, the mutational activation in vivo of a single copy of a ras gene may generate enough p204 to serve as its negative feedback inhibitor. Thus, it will be of interest to explore the effects of the cell-specific conditional inactivation in vivo of the gene encoding p204 on the tumorigenic activity of cells with a single copy of the ras oncogene.
A pioneering study, using MEFs, in which only a single copy of the K-ras G12D oncogene was expressed, unexpectedly revealed an attenuation of Ras signaling (compared with the signaling in wild-type cells) via the Raf and PI3K pathways (43). However, the protein levels of Raf, MEK, Erk, PI3K, and Akt were unaffected by the presence of the ras oncogene. No mechanism for the attenuation of Ras signaling was proposed.
The results presented in this study may provide such a mechanism (Fig. 13). This can be based on the findings that (i) activated wild-type or oncogenic Ras did induce p204 synthesis (mediated by Egr-1, whose expression was increased in consequence of Ras signaling) (42,73) (Fig. 11), (ii) a single copy of K-ras G12D in a cell induced a high level of p204 (Fig. 12), (iii) p204 bound RasGTP and inhibited Ras signaling via the Raf and PI3K (and other) pathways. Decreasing the level of endogenous p204 in cells expressing a single copy of K-ras G12D by ectopic 204 antisense RNA, however, resulted in the increase in Ras signaling via the Raf and PI3K pathways (Fig. 6B). It is conceiv-able that, besides p204, other agents might be induced by Ras that also cause or contribute to the attenuation of Ras signaling.
Types of proteins were described that, similarly to p204, bind Ras proteins and (unlike Ras effectors) inhibit Ras signaling (86). Carabin (an endogenous inhibitor of the phosphatase activity of calcineurin in T cells (87)) exhibits RasGAP activity. Members of the Ras association domain family (45,88) can interact directly with RasGTP. The biological characteristics of Ras association domain family proteins suggest that they are tumor suppressors. Sprouty (Spry) proteins function as suppressors of K-Ras-mediated tumorigenesis (89). There are no published data on a direct interaction between Ras and Spry proteins.
The recently described human Sin1 (stress-activated protein kinase-interacting protein) contains a Ras binding domain, it binds activated H-and K-Ras protein in vivo and in vitro, and its overexpression inhibits the activation of the ERK, Akt, and Jun terminal kinase signaling pathways (90). Thus, among the Rasbinding proteins described, the Ras inhibitory activity of the murine p204 protein resembles that of the human Sin1 protein the most.
The results of this study established that activated Ras can induce the synthesis of p204 and, by binding to RasGTP p204, can inhibit the signaling by activated Ras. It is conceivable that, in addition to inhibiting the signaling by numerous Ras effectors, p204 might also initiate as yet unidentified signals if triggered by binding RasGTP.