Signal transduction within the nucleus by mitogen-activated protein kinase.

The nucleus is an important target of signal transduction by growth factor receptors that stimulate mitogen-activated protein (MAP) kinases. We tested the hypothesis that MAP kinases have a signaling role within the nucleus by examining the effect of the expression of a human MAP kinase isoform (p41mapk) in tissue culture cells. The expressed p41mapk was found to be localized in both the cytoplasmic and nuclear compartments of the cells. Significantly, the expression of p41mapk caused an increase in the phosphorylation of a nuclear substrate: Ser62 of c-Myc. Phosphorylation at Ser62 stimulated the activity of the NH2-terminal transactivation domain of c-Myc. Thus, p41mapk causes the phosphorylation and regulation of a physiologically significant nuclear target of signal transduction. These data establish that at least one MAP kinase isoform has a nuclear role during signal transduction.

The nucleus is an important target of signal transduction by growth factor receptors that stimulate mitogen-activated protein (MAP) kinases. We tested the hypothesis that MAP kinases have a signaling role within the nucleus by examining the effect of the expression of a human MAP kinase isoform (p41maPk) in tissue culture cells. The expressed ~4 1 ' " ' '~ was found to be localized in both the cytoplasmic and nuclear compartments of the cells. Significantly, the expression of p41mapk caused an increase in the phosphorylation of a nuclear substrate: Sere' of c-Myc. Phosphorylation at Sere' stimulated the activity of the NHz-terminal transactivation domain of c-Myc. Thus, p41mapk causes the phosphorylation and regulation of a physiologically significant nuclear target of signal transduction. These data establish that at least one MAP kinase isoform has a nuclear role during signal transduction.
The stimulation of cellular proliferation and differentiation by peptide growth factors involves the activation of signaling pathways that are initiated by specific receptors at the cell surface. An important target for these signaling pathways is the regulation of gene expression in the nucleus. The spatial separation between the site of signal initiation and the nuclear targets of signal transduction indicates that information must be transduced between different cellular compartments. An understanding of this process is required in order to obtain a description of the molecular mechanism of signal transduction by growth factor receptors.
Protein phosphorylation represents a major mechanism of signal transduction. There is a complex regulatory network of protein kinases and protein phosphatases that determines the state of phosphorylation and function of cellular proteins. An important role for mitogen-activated protein (MAP)' kinases in the integration of these signaling pathways has been proposed . These MAP kinases function within a protein kinase cascade (Sturgill et al., 1988;Ahn CACGTG (Blackwell et al., 1990;Halazonetis and Kandil, 1991;Kerkhoff et al., 1991;Prendergast and Ziff, 1991). Furthermore, a transactivation domain has been identified within the NH2-terminal region of c-Myc (Kato et al., 1990). It is therefore likely that c-Myc functions as a transcription factor that regulates the expression of specific genes (Eilers et al., 1991).
Treatment of cells with growth factors causes an increase in the expression of the c-Myc protein and mRNA (Kelly et al., 1983). The regulation of c-Myc expression by growth factors therefore contributes to the control of c-Myc function in cells. However, in growing cells, c-Myc is expressed constitutively Rabbitts et al., 1985;Thompson et al., 1985;Dean et al., 1986;Waters et al., 1991). Regulation of c-Myc function under conditions of constitutive expression may be caused by post-translational modification. It is known that c-Myc is a short-lived nuclear phosphoprotein that is phosphorylated by casein kinase I1 at several sites (Luscher et al., 1989). The role of casein kinase I1 phosphorylation of c-Myc is not understood. However, it is possible that casein kinase I1 may indirectly stimulate the DNA binding of c-Myc complexes in vivo by inhibiting DNA binding by Max homodimers (Berberich and Cole, 1992).
Recently, we have demonstrated that the c-Myc protein is a substrate for phosphorylation by MAP kinases at Ser6' . SeP2 is also a substrate for phosphorylation by ~3 4 '~" ' Seth et al., 1991) and by glycogen synthase kinase 3 (Saksela et al., 1992). The location of this phosphorylation site is significant because it is found within a proline-rich region of the NH2-terminal domain of c-Myc that is conserved among members of the Myc family and has been shown to function as a transcriptional activation domain (Kato et al., 1990;Alvarez et al., 1991;Seth et al., 1991). Interestingly, isolates of v-myc, and also c-myc in patients with Burkitt's lymphoma, exhibit a high frequency of mutation within this proline-rich region (Rabbitts et al., 1984;Westway et al., 1984;Showe et al., 1985;Symonds et al., 1989). The location of Ser6' within an important functional domain of the c-Myc protein suggests that it may be a regulatory site of phosphorylation. Consistent with this hypothesis is the observation of growth factor-stimulated phosphorylation at Ser6* . c-Myc may therefore be a direct target of signal transduction pathways by a mechanism involving phosphorylation at Ser6' by MAP kinases Seth et al., 1991). However, direct evidence supporting a nuclear function of MAP kinases has not been obtained.
The purpose of the experiments described in this report was to test the hypothesis that there is an important nuclear role for MAP kinases during signal transduction. We demonstrate that the human p4lWk isoform is functionally expressed in the nucleus of serum-treated cells. Expression of p41mapk caused an increase in the phosphorylation state of a nuclear substrate: the c-Myc transactivation domain a t S e P . Phosphorylation at Ser6' was associated with increased transactivation of reporter gene expression. Together, these data establish that a growth factor-stimulated MAP kinase has an important functional role during signal transduction within the nucleus.

Plasmids
The plasmids pCHllO and pUC13 were from Pharmacia LKB Biotechnology Inc. The plasmid pGEM-Luc was from Promega Biotech. The plasmid pRSV-Luc was provided by Dr. P. Dobner (University of Massachusetts Medical School). The plasmid pTK-Fos-Luc (Chen et al., 1987) was from Dr. M. Rosenfeld (University of California, San Diego). The plasmids pGAL4/Myc and pGAL4/ [Ala6']Myc were constructed using the vector pSG424 and have been described previously . The plasmids pGAL4/ VP16 (Sadowski et al., 1988) and pSG424 (Lillie and Green, 1989) were obtained from Dr. M. Green (University of Massachusetts Medical School). The expression vector pCMV5 (Anderson et al., 1989) was from Dr. D. Russell (University of Texas Southwestern Medical School). Plasmid expression vectors (pCMV5) encoding the a and @ subunits of human casein kinase I1 Heller-Harrison et al., 1989;Heller-Harrison and Czech, 1991) were obtained from Drs. H. Meisner and M. Czech (University of Massachusetts Medical School). The following plasmids were constructed.
Plasmid pCMV-p4lWk-The plasmid pCMV-p41Wk was constructed using a 1992-bp Espl restriction fragment of the human p41Wk (ERK2) cDNA (Gonzalez et al., 1992; GenBank accession number 211694) cloned as a blunt-ended fragment in the polylinker of the expression vector pCMV5 at the SmaI site. The protein sequence of human p41Wk has a calculated molecular mass of 41 kDa and is 98.6% identical to the rat ERK2 protein kinase (Boulton et al., 1991). Mutation of p41wk at the ATP binding site to inhibit protein kinase activity was performed by replacing LysS4 and LysS5 (sequence AAGAAA) with Ala residues (sequence GCGGCA) using the polymerase chain reaction-based procedure described by Theroux and Davis (1992) to create the plasmid pCMV-p41(Ala64Ala55)Wk.
Plasmid pCMV-raf-The plasmid pCMV-raf was prepared using the vector pCMV5 and a 2114-bp EcoRI-XbaI fragment of a human c-raf-1 cDNA (Bonner et al., 1986) that was isolated using the polymerase chain reaction from a human fibroblast library.' Plasmid pCMV-Luc-The plasmid pCMV-Luc was prepared using a 1745-bp SalI-Hind111 fragment of pGEM-Luc containing the firefly luciferase gene cloned into the polylinker of the expression vector pCMV5.
PlasmidpG5ElbLw-The plasmid pG5ElbLuc was prepared from pG5ElbCAT (Lillie and Green, 1989) by removing the CAT gene by restriction digestion (EcoRI and Styl) and inserting the firefly luciferase gene (from pGEM-Luc) as a blunt-ended 1745-bp StuI-Hind111 fragment.
Cell Culture and Transient Expression Assays COS-7 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 5% fetal calf serum (GIBCO/ BRL). Transfections were performed using the DEAE-dextran method as previously described . The cells were trypsinized 20 h post-transfection, seeded into duplicate 100-mm dishes, and incubated in DMEM containing 5% fetal calf serum for 28 h. Serum-starved cells were transferred to medium without serum during the last 18 h of incubation. Transactivation assays were performed using 0.02 pg of GAL4 activator plasmid (pGAL4/Myc, pGAL4/[Ala6']Myc or pGAL4/VP16) and 2 pg of a luciferase reporter plasmid (pG5ElBLuc). To normalize for transfection efficiency a control plasmid (1 pg of pCH110) that expresses &galactosidase under the control of a simian virus 40 promoter was employed. In control experiments, the reporter plasmid pG5ElbLuc was replaced with 0.2 pg of the plasmids pRSV-Luc, pTK-Fos-Luc, pCMV5-Luc, or pSG424-Luc. To investigate the effect of p4lmWk expression, the cells were transfected with 2 pg of pCMV5 (vector control), pCMV-p4lmPk, or pCMV-p41(Ala64-Ala55)Wk. The total DNA in all transfections was maintained at 10 pg using pUC13 as carrier DNA.

e-Myc Phosphorylation by MAP Kinase
Measurement of Luciferase and @-Galactosidase Activity COS-7 cells were harvested 48 h after transfection and cell extracts prepared as previously described . Luciferase activity was measured using 2.5 pl of the cell extract with an Analytical Luminescence Laboratory (San Diego, CA) model 2010 luminometer . @-Galactosidase activity was measured by mixing 15 pI of the cell extract with 3 pl of 0.1 M MgClz, 4.5 M 8-mercaptoethanol, 66 p1 of 4 mg/ml o-nitrophenyl-0-D-galactopyranoside (Sigma), and 216 pl of 0.1 M sodium phosphate (pH 7.5). The incubations were performed a t 37 "C for 5 min or until a faint yellow color appeared. The reactions were stopped by adding 500 pl of 1 M Na2C03 to each tube, and the optical density a t 420 nM was measured using a Pharmacia LKB spectrophotometer.

MAP K i m e Assays
MAP kinase activity in cell extracts was measured using a synthetic peptide substrate based on the sequence of the EGF receptor surrounding the ThrS9 phosphorylation site (Countaway et al., 1989;Takishima et al., 1991;Northwood et al., 1991) as described by Gonzalez et al. (1991). COS-7 cells were lysed in 25 mM Hepes (pH 7.4), 5 mM EDTA, 50 mM NaF, 100 p~ sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, and 10 pg/ml leupeptin. The lysate was cleared by centrifugation a t 100,000 X g for 20 min at 4 "C, and the supernatant containing the MAP kinase activity was collected. MAP kinase activity was measured a t 25 "C using 10 pl of cell extract, 25 pg of synthetic peptide (KRELVEPLT66'PSGEAPNQALLR), 50 p M [-pJ2P]ATP (10 Ci/mmol), 25 mM Hepes, pH 7.4, and 10 mM MgCIz in a final volume of 25 pl. Control incubations were performed without the synthetic peptide. The reaction was terminated after 20 min by the addition of 10 pl of 125 mM ATP dissolved in 45% (v/v) formic acid. The peptide was isolated by spotting onto phosphocellulose paper (P81, Whatman) and washing with 1 M acetic acid, 4 mM sodium pyrophosphate. The incorporation of [32P]phosphate was estimated by measuring the associated Cerenkov radiation with a Beckman liquid scintillation counter.

Western Blot Analysis
Cells were lysed in 25 mM Hepes, pH 7.5, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and 10 pg/ml leupeptin a t 4 "C. The extracts were then centrifuged a t 100,000 X g for 20 min a t 4 "C. The supernatant was subjected to SDS-PAGE, and the resolved proteins were electrophoretically transferred onto a polyvinylidene difluoride membrane (Immobilon-P, Millipore Corp.). The blots were probed with: 1) a mouse anti-MAP/ERK kinase monoclonal antibody (Zymed, San Francisco, CA) to detect ~41"'"~ and 2) a mouse antiphosphotyrosine monoclonal antibody (PY20; ICN, Irvine, CA) to detect proteins containing phosphotyrosine. The immune complexes were visualized by the enhanced chemiluminescence Western blotting procedure (Amersham International PLC).

Isolation of Metabolically Labeled Proteins by Immunoprecipitation
Metabolic labeling of transfected cells in 100-mm dishes was performed 40 h post-transfection by transferring the cells into culture medium (Flow Laboratories Inc.) supplemented with 0.1% fetal calf serum. The culture media used were: 1) methionine-free modified Eagle's medium containing 150 pCi/ml [?3]methionine (Amersham International PLC) or 2) phosphate-free modified Eagle's medium containing 2 mCi/ml [3'P]phosphate (Du Pont-New England Nuclear). After 4 h of additional incubation, the cells were lysed and the GAL4 fusion proteins were immunoprecipitated using anti-GAL4 antiserum and resolved by SDS-PAGE as described by Alvarez et al. (1991).

Phosphopeptide Mapping and Phosphoamino Acid Analysis
Phosphorylated proteins were digested with tosylphenylalanyl chloromethyl ketone-treated trypsin in 25 mM N-ethylmorpholine (pH 8.0) as described by Alvarez et al., (1991). Phosphopeptide mapping was performed by two-dimensional separation on 100-pm cellulose thin layer plates (Machery-Nagel, Duren, Germany). The first dimension was electrophoresis using 30% (v/v) formic acid a t 500 V for 2 h a t 4 "C. The second dimension was chromatography in the vertical direction using butan-1-ol/pyridine/acetic acid/water (75:5015:60). The origin is located in the lower left corner of each panel. Phosphoamino acid analysis was performed by partial acid hydrolysis (1 h a t 110 "C in 6 M HCl) and thin layer electrophoresis as described .

Immunocytochemistry
Cells were plated into 8-well tissue culture chambers mounted on a glass slide (Lab-tek) and incubated for 24 h in DMEM supplemented with 5% fetal bovine serum. The cells were washed and processed for immunofluorescence in KRH buffer (120 mM NaCl, 6 mM KC1, 1.2 mM MgC12, 1 mM CaC12, 25 mM Hepes, pH 7.4). The cells were fixed in KRH containing 3.7% formaldehyde (15 min, 22 "C), washed, and incubated in KRH containing 50 mM NH,Cl for 15 min at 22 "C. Next, they were permeabilized with 0.2% Triton X-100 in KRH (5 min, 22 "C), washed three times, and incubated without and with primary antibodies for 1 h in KRH containing 20% horse serum a t 22 "C. The primary antibodies and the dilutions used were: 1) a rabbit anti-GAL4 antiserum  diluted 1:200 and 2) a monoclonal anti-MAP/ERK kinase antibody (Zymed Laboratories, Inc., San Francisco, CA) diluted 1:500. After the incubation with primary antibodies, the cells were washed three times and incubated for 1 h a t 22 "C in KRH containing secondary antibodies and 20% horse serum. The secondary antibodies and the dilutions used were: 1) a fluorescein isothiocyanate-conjugated goat anti-rabbit Ig antibody (ICN Immunochemicals) diluted 1:2000, and 2) a rhodamineconjugated rabbit anti-mouse Ig antibody (Boehringer Mannheim) diluted 1:500. The cells were then washed with KRH and mounted for microscopy. The slides were examined using a Zeiss Axiovert microscope equipped for epifluorescence with an oil immersion lens (magnification, X 63; numerical aperture = 1.4). The specificity of the immunofluorescence observed was confirmed in control experiments performed 1) without and with the primary antibodies; 2) comparison of the staining of mock-transfected cells and cells transfected with GAL4/Myc and p41Wk.

Laser Scanning Confocal Microscopy
The confocal microscope, unlike standard fluorescence microscopes, provides an improved resolution along the z-axis by rejecting out-of-focus fluorescence (Wells et al., 1989). Thus, to identify the subcellular location of p41Wb and the GAL4/Myc proteins, we used fluorophore-labeled antibodies and a laser scanning confocal microscope (model MRC-600, Bio-Rad) equipped with a Nikon oil immersion lens (magnification, X 60; numerical aperture = 1.4). Image sets were obtained by acquiring 10 optical slices a t 0.5-pm intervals. Each pseudocolor image represents an optical slice that was acquired by accumulating four scans of the laser.

RESULTS
Expression of Functional p41mpk in the Nucleus-A transient transfection assay was employed to express p41mpk in COS-7 cells. Lysates prepared from the cells were examined by Western blot analysis. Blots probed with a MAP kinase antibody demonstrated a marked increase (>lOO-fold) in the level of expression of p4lWk (Fig. 1). The increased expression of p41mpk was associated with a higher level of MAP kinase activity detected in the lysates prepared from transfected cells compared with control cells (data not shown). Consistent with this observation, it was observed that immunoblots of the cell lysates probed with an anti-phosphotyrosine antibody indicated a high level of tyrosine phosphorylation of the expressed p41mpk (Fig. 1).
Both cell surface and nuclear proteins have been demon- strated to be in vitro substrates for purified MAP kinases . An example of a cell surface protein substrate is the EGF receptor (Countaway et al., 1989;Takishima et al., 1991;Northwood et al., 1991). Examples of nuclear proteins that are MAP kinase substrates include Myc family proteins  and c-Jun Pulverer et d., 1991). The in vivo function of different MAP kinase isoforms may depend on the intracellular location of these enzymes. We therefore investigated the cellular distribution of the expressed p41Wk by indirect immunofluorescence. Control experiments demonstrated the specificity of the immunofluorescence observed using antibodies to p41and GAL4/Myc. There was a marked staining of p41Wk located in the cytoplasm surrounding the nucleus (Fig. 2). In addition, significant nuclear staining of p4lWk was detected. To examine the possible intranuclear location of p41Wk, we performed confocal microscopy. Optical sectioning confirmed that the p41Wk was located within the nucleus and was excluded from specific intranuclear regions that may correspond to nucleoli (Fig. 3). The GAL4/Myc fusion protein was In previous studies, we have shown that the c-Myc transactivation domain is a substrate for MAP kinase at Ser6' in vitro and that phosphorylation of this site is regulated by growth factors in vivo . To examine the effect of p41Wk on Ser" phosphorylation, we constructed a fusion protein consisting of the transactivation domain of c-Myc and the DNA binding domain and nuclear localization signals of the yeast transcription factor GAL4. This fusion protein can be expressed and is correctly localized in the nucleus of transfected cells (Figs. 2 and 3). The state of phosphorylation of the GAL4IMyc fusion protein was examined in cells metabolically labeled by incubation in medium containing ["P]phosphate. It was observed that the expression of p41Wk caused a marked increase in the state of phosphorylation of the GAL4/Myc fusion protein (Fig. 4). Phosphoamino acid analysis demonstrated a marked increase in the level of [32P]phosphoserine (Fig. 5). A small increase in the level of [3'P]phosphothreonine was also detected (Fig.  5). The increase in the phosphorylation state of the GAL4/ Myc fusion protein was further characterized by tryptic phosphopeptide mapping. It was found that the stimulated phosphorylation was largely accounted for by an increase in the level of two tryptic [32P]phosphopeptides (Fig. 5). In previous studies, we have shown that these phosphopeptides are the result of incomplete tryptic digestion of the sequence surrounding the phosphorylation site Ser6' . To confirm the conclusion that Sere' was thy major site of p41Wk-stimulated phosphorylation, we exammed the effect of the replacement of Ser6' with an Ala residue. Fig. 4 shows that this mutation blocked the increase in phosphorylation of the GAL4/Myc fusion protein caused by p4lWk. Together, these data demonstrate that the expression of p41Wk causes an increase in the phosphorylation state of the c-Myc transactivation domain at Sera'.  (Fig. 6). Expression of p4lWk caused a marked increase in the level of GAL4/Myc-dependent luciferase activity (Fig. 6). Control experiments demonstrated that the effect of p41Wk on luciferase activity was not the result of an increase in the level of expression of the GAL41Myc fusion protein (Fig. 4). Together, these data suggest that p41mpk stimulates the function of the c-Myc transactivation domain.

c-Myc Phosphorylation by M A P Kinase
The effects of p41Wk on luciferase activity (Fig. 6) could be a specific action on the c-Myc transactivation domain or may be the result of a generalized action of this protein kinase on gene expression. We therefore performed control experiments to address this question. Several reporter plasmids were constructed containing different promoters and the firefly luciferase gene. Fig. 7 shows that no significant effects of p4lWk on luciferase activity were observed in experiments and by tryptic ["P]phosphopeptide mapping (panel B ) . Separate analyses were performed on the two forms of wild-type (Ser6*) GAIA/ Myc with different electrophoretic mobilities found during SDS-PAGE (Fig. 4). The origin is marked with a cross (+) at the lower left corner of each peptide map, and the electrophoretic (cathode at right) and chromatographic dimensions are illustrated. The exposure time for autoradiography was 10 days (panel A ) and 4 days (panel B ) at -80 "C using Kodak X-Omat AR film and a Du Pont Lightning Plus enhancing screen. Similar results were obtained in three separate experiments.
using reporter plasmids with a Rous sarcoma virus promoter, a cytomegalovirus promoter, a simian virus 40 promoter, and a synthetic promoter containing sequences derived from the c-fos and thymidine kinase genes. Thus, the effect of p41mopk on the function of the c-Myc transactivation domain is not accounted for by a generalized effect on gene expression (Fig.  7).
To further examine the specificity of the effect of p41mopk on the function of the c-Myc transactivation domain, we compared the effect of the expression of p41mpk with other protein serinelthreonine kinases, It was found that increased expression of casein kinase 11, c-Raf-1, and Ca2+/calmodulindependent protein kinase I1 failed to cause significant changes in the function of the c-Myc transactivation domain (Fig. 8). We therefore conclude that the effects of p41mopk expression + " Control --+ Inactive -+ -Wlld-type 1 . FIG. 6. ~4 1 ' " "~ increases the activity of the c-Myc transactivation domain. The effect of the expression of both the wild-type p41Wk and a mutated (kinase-inactive) p41Wk was investigated. Cells were transfected with 0.02 pg of the activator plasmid pGAL4/Myc, 2 pg of the reporter plasmid pG5ElBLuc and 1 pg of a @-galactosidase expression vector (pCH110) together with either 2 pg of pCMV5 (vector control), pCMV-p4lWk, or pCMV-p41(Ala"Ala65)mk. The cells were incubated in medium containing 5% fetal bovine serum, and extracts were prepared 48 h post-transfection. Luciferase activities (in light units) were normalized for differences in transfection efficiency by measurement of the P-galactosidase activity (OD value) and are presented as (light units)/(OD value). Each data point represents the mean * S.D. of determinations obtained from three independent transfections. Similar data were obtained in three separate experiments.

-
FIG. 7. Effect of ~4 1 " "~' on gene expression from plasmid promoter constructs. Reporter constructs containing the firefly luciferase gene (0.2 pg) were co-transfected together with 2 pg of pCMV-p41mk or pCMV5 (vector control) and 1 pg of pCHllO (@galactosidase control). The cells were incubated in medium containing 5% fetal bovine serum, and extracts were prepared 48 h post-transfection. The reporter plasmids used containing the TK-Fos, Rous sarcoma virus (RSV), cytomegalovirus (CMV), and simian virus 40 (SV40) promoters were pTK-Fos-Luc, pRSV-Luc, pCMV5-Luc, and pSG424-Luc, respectively. The luciferase activity is expressed as (light units)/(@-galactosidase activity) and is normalized to the value obtained for the vector control (100%). Each data point represents the mean * S.D. of data obtained from three independent transfections. Similar data were obtained in three separate experiments.
are not related to a generalized increase in protein serine/ threonine kinase activity. Instead, these data are consistent with the hypothesis that there is a specific action of p4lWk to regulate the function of the c-Myc transactivation domain.
Increased Transactivation of Gene Expression Caused by p41Wk Requires the c-Myc Phosphorylation Site Ser6' and Is Potentiated by Serum-Expression of p41Wk causes phosphorylation of the c-Myc transactivation domain at Ser6* (Figs. 4 and 5 ) and is associated with an increase in the transactivation of reporter gene expression (Fig. 6). We performed three experiments to test the hypothesis that the increased phosphorylation is mechanistically relevant to the L + .

FIG. 8. Comparison of the effect of Ser/Thr protein kinases on the activity of the c-Myc transactivation domain.
Cells were co-transfected with 2 pg of a reporter luciferase plasmid (pG5ElBLuc), 0.02 pg of an activator plasmid (pGAL4/Myc) and 1 pg of a @-galactosidase expression vector (pCH110). The effect of cotransfection with 2 pg of a plasmid encoding a protein serine/threonine kinase was investigated negative control (-; pUC13), vector control (pCMV5), p4lWk, Raf-1, Ca*+/calmodulin-dependent protein kinase I1 (CAMKZZ), and casein kinase I1 (CKZZ). The cells were incubated in medium containing 5% fetal bovine serum, and extracts were prepared 48 h post-transfection. Luciferase activities (light units) were normalized for differences in transfection efficiency by measurement of the @-galactosidase activity (OD value) and are presented as (light units)/(OD value). Each data point represents the mean * S.D. of determinations obtained from three independent transfections. Similar data were obtained in three separate experiments. stimulated activity of the c-Myc transactivation domain. 1) We tested the hypothesis that the effect of p41Wk on the transactivation of gene expression (Fig. 6) requires protein kinase activity. Site-directed mutagenesis was used to mutate the ATP binding site of p41Wk. Control experiments demonstrated that this mutated enzyme exhibited no detectable protein kinase activity (data not shown). Expression of the kinase-inactive p41Wk caused no significant change in the transactivation of reporter gene expression by the c-Myc transactivation domain (Fig. 6). This result demonstrates that the protein kinase activity of p41Wk is required for the regulation of gene expression (Fig. 6).
2) The protein kinase activity of p41Wk in tissue culture cells is increased by serum . We therefore investigated the effect of serum treatment on the function of the c-Myc transactivation domain. Fig. 9 shows that serum starvation caused a marked decrease in the p41Wk -stimulated activity of the c-Myc transactivation domain. Thus, the effect of p41Wk on the function of the c-Myc transactivation domain is serum-dependent. In contrast, no significant effect of serum or p4lWk on reporter gene expression was found in control experiments using the VP16 acidic transactivation domain (Fig. 9).
3) If phosphorylation at Ser6' is mechanistically relevant to the transactivation of reporter gene expression, it can be predicted that a mutation at this site should block the effects of expression of p4lmWk. We therefore investigated the effect of the replacement of the phosphorylation site Ser6* with an Ala residue. It was observed that this point mutation markedly reduced the effects of p41mpk on both the phosphorylation (Figs. 4 and 5 ) and function (Fig. 9) of the c-Myc transactivation domain.
Together, these data strongly support the hypothesis that increased phosphorylation at Ser6' accounts for the mechanism of p4lmnpk to stimulate the function of the c-Myc transactivation domain.

DISCUSSION
The binding of peptide growth factors to receptors expressed at the cell surface causes changes in gene expression that result in the proliferation or differentiation of cultured cells. Recently, progress has been made toward understanding the process of signal transduction by growth factor receptors (Ullrich and Schlessinger, 1990;Koch et al., 1991). However, the mechanism that couples signals initiated at the cell surface to the regulation of nuclear events is poorly understood. Several lines of evidence indicate that the regulation of pro-tion by M A P Kinase tein phosphorylation is an important mechanism of signal transduction. One example is represented by the nuclear localization of activated CAMP-dependent protein kinase (Nigg et al., 1985;Meinkoth et al., 1990;Adams et al., 1991), the phosphorylation of CREB at Ser'33, and the transactivation of gene expression (Gonzalez and Montminy, 1989). However, the role of other protein kinases that are activated during signal transduction is less well established.
A MAP Kinase Isoform (~4 1 "~) Is Functionally Expressed within the Nucleus-Recently, there has been considerable interest in the role of MAP kinases because these enzymes appear to integrate signaling pathways initiated by many cell surface receptors . I n uitro studies have demonstrated that nuclear proteins are substrates for MAP kinases, and it has been suggested that MAP kinases may have an important role within the nucleus Pulverer et al., 1991;Seth et al., 1991). Consistent with this proposal are observations that MAP kinases can be detected in the nucleus Chen et al., 1992).
However, a direct experimental demonstration of a nuclear function of MAP kinases has not been obtained.
The purpose of the study described here was to test the hypothesis that MAP kinases represent a significant regulatory mechanism for nuclear processes. The experimental strategy that we employed was to examine the effect of the expression of a human MAP kinase isoform (~41"~) in tissue culture cells. It was found that the expressed kinase was located within both the nuclear and cytoplasmic compartments of the cells (Figs. 2 and 3). Significantly, the expression of p4lWk increased phosphorylation of a nuclear substrate, the c-Myc transactivation domain at Ser@ (Fig. 4). The marked increase in nuclear MAP kinase substrate phosphorylation caused by the transient expression of p4lWk allows the direct experimental examination of MAP kinase signaling pathways within the nucleus. It is likely that this experimental strategy may be generally applicable to the investigation of MAP kinase function in the nucleus. Further studies to investigate the effects of p4lWk expression on nuclear function are therefore warranted. However, an important caveat that must be placed on the interpretation of data obtained from experiments using this method is that the very high levels of expression of MAP kinase may result in non-physiological responses.
Previously, i n uitro phosphorylation by MAP kinases has been demonstrated for proteins that are located within different compartments of intact cells: 1) membrane proteins including the EGF receptor (Countaway et al., 1989;Northwood et al., 1991;Takishima et al., 1991), 2) the protein kinases p 9 P k and Raf-1 (Sturgill et al., Lee et ab, 1992), and 3) the nuclear proteins c-Myc and c-Jun Pulverer et al., 1991). If MAP k' lnases account for the phosphorylation of all of these proteins i n uiuo, a significant problem of protein compartmentation exists. The MAP kinase family includes several isoforms with similar in uitro substrate specificity Gonzalez et al., 1991;Clark-Lewis et al., 1991), and it is possible that each isoform may have a distinct subcellular distribution. In previous studies of MAP kinase compartmentation, the relative distribution of individual isoforms has not been addressed Chen et al., 1992). In the present study, the data obtained indicate a significant nuclear accumulation of p4lWk in serum-treated cells (Fig. 2), but evidence about the location of other MAP kinase isoforms within these cells is lacking. Further studies of the compartmentation of specific MAP kinase isoforms are therefore required to resolve the roles of the different MAP kinases that have been identified .

The c-Myc Tramactivation Domain Is Regulated by Phosphorylation at Ser6'-We have previously reported that Ser6' is a site of c-Myc phosphorylation by MAP kinases in vitro
and that this site is also phosphorylated in vivo . Serfi2 is located within the NHz-terminal transactivation domain of c-Myc (Kato et al., 1990) and is conserved in all members of the Myc family . It is therefore likely that several members of the Myc family are phosphorylated at this site. Recently, this hypothesis has been confirmed by the demonstration that L-Myc is also phosphorylated at this site within the transactivation domain (Saksela et al., 1992).
In this study, we have investigated the functional significance of the phosphorylation of c-Myc at Ser6' by a MAP kinase. As this phosphorylation site is located within the NH2-terminal transactivation domain of c-Myc (Kato et al., 1990), we tested the hypothesis that phosphorylation at Ser6' regulates the transactivation of gene expression . The experimental strategy that we employed was to investigate the effect of transient expression of p41mpk on the function of the c-Myc transactivation domain. It was found that the expression of p41Wk caused both an increase in phosphorylation at Ser'j2 and an increase in the transactivation of gene expression. The p41mpk-induced level of transactivation was substantial (Fig. 9A) and corresponds to approximately 75% of that detected with the strong acidic activation domain of VP16 (Sadowski et al., 1988;Fig. 9C). Replacement of Ser'j' with Ala blocked these effects of p a W k expression. Together, these data strongly support the hypothesis that the c-Myc transactivation domain is regulated by phosphorylation at Ser'j'.
Recent studies have established a critical role for phosphorylation in the regulation of transcription factor function. The increase in c-Myc transactivation caused by phosphorylation at Ser'j' is similar to that reported for the regulation of c-Jun by phosphorylation at Ser63 and Ser73 (Bin6truy et al., 1991;Pulverer et al., 1991;Smeal et al., 1991). The regulation of c-Myc is also similar to the increased transactivation caused by phosphorylation of CREB at Ser'33 (Gonzalez and Montminy, 1989;Sheng et al., 1991;Dash et al. 1991). This form of regulation of transactivation domain function differs from the effects of phosphorylation to cause altered DNA binding activity: c-Myb (Luscher et al., 1990), c-Max (Berberich andCole, 1992), SRF (Manak et al., 1990;Janknecht et al., 1992), c-Jun (Boyle et al., 1991), E2F (Bagchi et al., 1989), and E4F (Raychaudhuri et al., 1989).
Physiological Regulation of the Phosphorylation State of c-Myc at Serfi2-The transfection studies described here demonstrate that p4lWk is expressed in both the nuclear and cytoplasmic compartments of cells. Significantly, the expression of p a W k causes increased phosphorylation of the c-Myc NHz-terminal domain at Ser6' and results in an increase in the transactivation of gene expression. A nuclear role for p41mapk is therefore established. However, the data obtained demonstrate only that p41Wk is a candidate protein kinase that could account for the phosphorylation of Ser'j' in vivo. A role for other MAP kinase isoforms is not excluded. Furthermore, there are related protein kinases (such as glycogen synthase kinase-3 and cyclin-dependent protein kinases) that may also regulate the phosphorylation state of Ser62 in vivo Seth et al., 1991;Saksela et al., 1992). The physiological regulation of c-Myc by phosphorylation may therefore be complex and involve more than one protein kinase that is active at different stages of the cell cycle.
Increased function of the c-Myc protein has previously been associated with oncogenesis (Luscher and Eisenman, 1990;Penn et al., 1990). c-Myc co-operates with Ras to cause transformation of rat embryo fibroblasts (Land et al., 1983). The molecular basis for the requirement of both Ras and c-Myc has not been established, but it is known that the NH2terminal region of c-Myc surrounding the phosphorylation site Ser6' is required for co-transformation (Stone et al., 1987).
Recently, it has been demonstrated that Ras expression causes an increase in MAP kinase activity (Leevers and Marshall, 1992;Nori et al., 1992;Thomas et al., 1992;Wood et al., 1992). The results of the present study suggest that this effect of Ras to stimulate MAP kinase may contribute to the activation of c-Myc function.