Constitutive JNK Activation in NIH 3T3 Fibroblasts Induces a Partially Transformed Phenotype*

The c-Jun N-terminal kinases (JNKs) (also known as stress-activated protein kinases or SAPKs), members of the mitogen-activated protein kinase (MAPK) family, regulate gene expression in response to a variety of physiological and unphysiological stimuli. Gene knockout experiments and the use of dominant interfering mutants have pointed to a role for JNKs in the processes of cell differentiation and survival as well as oncogenic transformation. Direct analysis of the transforming potential of JNKs has been hampered so far by the lack of constitutively active forms of these kinases. Recently, such mutants have become available by fusion of the MAPK with its direct upstream activator kinase. We have generated a constitutively active SAPKβ-MKK7 hybrid protein and, using this constitutively active kinase, we are able to demonstrate the transforming potential of activated JNK, which is weaker than that of classical oncogenes such as Ras or Raf. The inducible expression of SAPKβ-MKK7 caused morphological transformation of NIH 3T3 fibroblasts. Additionally, these cells formed small foci of transformed cells and grew anchorage-independent in soft agar. Furthermore, similar to oncogenic Ras and Raf, the expression of activated SAPKβ resulted in the disassembly of F-actin stress fibers. Our data suggest that constitutive JNK activation elicits major aspects of cellular transformation but is unable to induce the complete set of changes which are required to establish the fully transformed phenotype.

Mitogen-activated protein (MAP) 1 kinases (MAPKs) are essential signaling molecules that translate extracellular stimuli into nuclear responses through the phosphorylation of transcription factors. Three major subfamilies of MAPKs have been isolated to date: the extracellular signal-regulated kinase (ERK), the c-Jun N-terminal kinase (JNK, also known as stress-activated protein kinase or SAPK), and the p38 MAP kinase (1,2). MAP kinase pathways are evolutionarily conserved and respond differentially to multiple physiological and non-physiological stimuli (3). Signals for ERK activation, including growth and differentiation factors (hormones as well as tumor promoters) (4), are relayed from the cell surface to the nucleus through a conserved signal transduction pathway, which includes the small G-protein Ras and a downstream kinase cascade consisting of Raf, MEK (MAPK/ERK kinase), and ERK (5). Two components of this pathway, Ras and Raf, have initially been identified as oncogenes, stressing the role of this pathway in deregulated growth (6). Moreover, the Raf 3 MEK 3 ERK cascade is essential for transformation by various classes of oncogenes (7)(8)(9).
In contrast to ERKs, SAPK/JNKs and p38 are poorly activated by mitogens but strongly stimulated in response to stress inducers like UV-light, ionizing radiation, osmotic or redox stress, heat shock, and inflammatory cytokines (e.g. tumor necrosis factor ␣ and interleukin-1) (10). They have been implicated in embryonic development, immune response, DNA repair, cell proliferation, cell survival, and apoptosis (11,12). More recently, a role for JNK/SAPK in oncogenic transformation and tumor development has been postulated (13)(14)(15). Moreover, overexpression of the JNK substrate c-Jun transforms chicken embryo fibroblasts (16), and c-Jun is required for the transformation of mammalian or rodent fibroblasts by different oncogenes (17)(18)(19). Although these data suggest that SAPK/JNK may be a universal mediator of cellular transformation, an evaluation of the contribution of SAPK/JNK signaling to this process has been prevented by the lack of constitutively active mutants of this kinase. Until now, attempts to generate structure-based, site-directed mutants of MAP kinases with constitutive activity have failed. Only recently did the direct enzyme-substrate fusion yield a constitutively active form of the MAPK ERK (20). In an analogous approach, we fused the MAP kinase SAPK␤ in-frame with its direct upstream activator MKK7 to generate a constitutively active SAPK␤-MKK7 (Ref. 21 and this paper). Employing this tool, we were able to show that SAPK␤-MKK7 behaves as a moderately transforming oncogene. Inducible SAPK␤-MKK7 expression in NIH 3T3 fibroblasts resulted in a morphologically transformed phenotype. In these cells the F(ilamentous)-actin cytoskeleton was disassembled, and cells expressing SAPK␤-MKK7 were able to grow anchorage-independent in soft agar and form small foci of transformed cells in monolayers.

MATERIALS AND METHODS
Cell Lines-Mouse NIH 3T3 fibroblasts were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum (heat inactivated at 56°C for 45 min), 2 mM L-glutamine, and 100 units/ml penicillin/streptomycin at 37°C in humidified air with 5% CO 2 . Generation of the tetracycline-regulated NIH 3T3 tet-off cell line (NIH 5.15) * This work was supported by a grant from the Deutsche Forschungsgemeinschaft (special research fields 465 and 581) and by the Mildred Scheel Stiftung. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

DNA Constructs Generation of pcDNA3-Myc-SAPK␤-MKK7
A full-length SAPK␤ cDNA (GenBank TM accession number L27128; provided by Dr. J. Woodgett, Ontario Cancer Institute, Toronto, Canada) was cloned into the XhoI/XbaI sites of the mammalian expression vector pcDNA3neo (Invitrogen) following PCR amplification with the following forward and reverse primers: 5Ј-CCG CTC GAG AGC AAA AGC AAG GTA GAT AAC CAG TTC TAC-3Ј and 5Ј-TGC TCT AGA CCT GCA ACA ACC CAG CGG TCC CGC CGA-3Ј. Behind the stop codonless SAPK␤, a PCR-generated fragment of the coding sequence of mouse MKK7 (GenBank TM accession number AF0026216; provided by Dr. J. M. Penninger, Ontario Cancer Institute, Toronto Canada) was cloned in-frame into the XbaI sites using the following primers: 5Ј-TGC TCT AGA GCT AGC GTT AAC GAC CAG AAG CTG CAG GAG ATC ATG AAG-3Ј and 5Ј-TGC TCT AGA CTA CCT GAA GAA GGG CAG ATG GTG CTG ACT CAG GAC-3Ј. Between the two cDNAs, an alternating copolymer of 10 amino acids of glutamic acid and glycine was inserted in-frame into the NheI/HpaI sites. This linker was synthesized by in vitro annealing of the following forward and reverse primers: 5Ј-CTA GCT AGC GAA GGT GAG GGC GAA GGA GAA GGT GAG GGC GTT AAC ATC GAT GG-3Ј and 5Ј-CCA TCG ATG TTA ACG CCC TCA CCT TCT CCT TCG CCC TCA CCT TCG CTA GCT AG-3Ј. At the 5Ј-end of the fusion construct, a sequence encoding the Myc tag epitope was inserted by BamHI/XhoI. pcDNA3-Myc-SAPK␤-MKK7 KD, a kinaseinactive mutant form of this hybrid protein, has been described previously (21).
Immunoblotting-Cells were washed twice in phosphate-buffered saline (PBS) and lysed directly by the addition of Laemmli buffer (60 mM Tris-HCl, pH 6.8, 10% (w/v) glycerine, 3% (w/v) SDS, 5% (w/v) ␤-mercaptoethanol, and 0.05% (w/v) bromphenol blue) and boiling at 100°C for 5 min. Total protein extracts were separated on SDS-polyacrylamide gels and then transferred to nitrocellulose membranes in a buffer containing 25 mM Tris (pH 8), 190 mM glycine, and 0.1% (w/v) SDS using a tank blot procedure. Blotting was performed at 800 mA for 1 h. The membrane was blocked for 1 h at room temperature on a platform shaker in PBS supplemented with 5% (w/v) nonfat dried milk (Carnation) and 0.05% (v/v) Tween 20 (PBS-T). Incubation with the primary antibody appropriately diluted in PBS supplemented with 0.05% (v/v) Tween 20 and 5% (w/v) milk powder was carried out overnight at 4°C. Afterward, the blot was washed three times for 10 min with PBS supplemented with 0.05% (v/v) Tween 20 and then incubated with anti-mouse or anti-rabbit horseradish peroxidase-linked F(abЈ) 2 fragments (Amersham Biosciences) diluted 1:3000 in PBS-T for 45 min. Following three washes, the bands were visualized with the ECL detection system (Amersham Biosciences). Bound antibodies were removed from the membrane with stripping buffer (100 mM ␤-mercaptoethanol, 2% (w/v) SDS, and 62.5 mM Tris-HCl, pH 6.7) at 50°C for 30 min followed by three washing steps in PBS-T and incubation in blocking buffer. The following primary antibodies were used: Myc tag, 9E10 (mouse monoclonal; MSZ); HA tag, 12CA5 (mouse monoclonal; MSZ); phospho-JNK, G-7 (mouse monoclonal; Santa Cruz Biotechnology catalog no. sc-6254); JNK1 FL (rabbit polyclonal; Santa Cruz Biotechnol-ogy catalog no. sc-571); phospho-c-Jun; KM-1 (mouse monoclonal; Santa Cruz Biotechnology catalog no. sc-822); ERK1 (rabbit polyclonal; Santa Cruz Biotechnology catalog no. sc-94). SAPK/JNK control cell extracts were commercially available from New England Biolabs.
Immune Complex Kinase Assay-Cells were washed twice in PBS and lysed at 4°C in ICKA lysis buffer (50 mM HEPES, pH 7.5, 150 mM NaCl, 1 mM EDTA, 2.5 mM EGTA, 1 mM dithiothreitol, 1% Triton X-100, 10% glycerol, 5 M 4-(2-aminoethyl)benzenesulfonyl fluoride, 1.5 nM aprotinin; 10 nM E-64; 5 nM EDTA; 10 nM leupeptin). Following centrifugation for 10 min at 12,000 ϫ g, 10 l of mouse monoclonal anti-Myc (9E10) were added to the precleared lysates and agitated for 2h at 4°C. After the addition of 20 l of protein A-agarose (Roche Molecular Biochemicals), the suspension was mixed again for 2 h at 4°C. Immunoprecipitates were washed three times in ICKA lysis buffer and once with a 50 mM HEPES, 1 mM dithiothreitol buffer. For the kinase reaction, 30 l of a mix containing kinase buffer (50 mM HEPES, pH 7.5, 10 mM magnesium chloride, 1 mM dithiothreitol, 1 mM sodium fluoride, 0.1 mM sodium orthovanadate, 2.5 mM EGTA, 20 M cold ATP, and 10 Ci of [␥-32 P]ATP) supplemented with bacterially expressed and purified glutathione S-transferase GST-c-Jun N-terminal protein (amino acids 1-135) (24) was added. The reaction was carried out for 30 min at 37°C under constant agitation and terminated by adding 10 l of 5ϫ Laemmli buffer (for 1ϫ concentration, see above) and boiling the suspension for 5 min at 100°C. Proteins were separated by SDS-PAGE and transferred to a nitrocellulose membrane. Phosphorylation of the substrate protein was visualized by autoradiography. The amount of immunoprecipitated protein was controlled by immunoblotting.
Immunostaining/Indirect Immunofluorescence-NIH 3T3 fibroblasts were seeded on glass cover slides. 36 h post-transfection the cells were washed three times with PBS and fixed with 3.7% paraformaldehyde at 4°C for 20 min. After washing three times in PBS, cells were permeabilized in PBS supplemented with 0.2% (v/v) Triton X-100 for 3-4 min at room temperature. For immunodetection the PBS-washed cells were incubated for 1 h in the presence of 10 l of anti-Myc (9E10) antibody diluted in 40 l of PBS followed by three PBS washing steps. Cells were then incubated for 1 h with a mix of Cy3-labeled goat anti-mouse antibody (1:200 dilution; Jackson ImmunoResearch Laboratories) and fluorescein isothiocyanate-conjugated phalloidin (1:15 dilution; Molecular Probes). After washing with PBS, cells were mounted with Mowiol and viewed with a fluorescence microscope (Leitz DMRB).
Soft Agar Assay-6-cm tissue culture plates were covered with 5 ml of 0.5% soft agar (SeaPlaque). Cells were suspended in 2 ml of 0.3% soft agar and added to each plate. Cells were plated in triplicate at a density of 10 5 , 10 4 , and 10 3 cells/dish and incubated for 18 days. Dishes were kept in humidified chambers. For tetracycline treatment, plates were overlaid every fourth day with 2 ml of complete Dulbecco's modified Eagle's medium containing 2 g/ml tetracycline. Colonies that had undergone several cell divisions (approximately more than 12 cells) were counted as positive colonies. A total of 200 cells/plate within randomly selected fields of view were analyzed, and mean values and standard deviations were calculated.
Focus Formation Assay-Either the inducible NIH 3T3 tet-off cell line or transiently transfected NIH 3T3 cells were used for this assay. In the latter case, NIH 3T3 fibroblasts were transfected with 0.75 g of total DNA. 1 day post-transfection four-fifths of the transfected cells were seeded on 9-cm tissue culture plates. Cells were refed twice weekly. 15-18 days postseeding they were fixed with methanol and stained with Giemsa to visualize the foci of transformed cells. Data shown mean values obtained from three independent plates for each transfection. One-fifth of the transfected cells were cultivated for 2 days, and then total protein lysates were prepared to control the expression of each transfected construct.

Generation of a Constitutively Active SAPK␤ Mutant-Until
now, structure-based, site-directed mutagenesis failed to produce constitutively active forms of MAP kinases. As an alternative approach, the direct enzyme-substrate fusion was first successfully applied to generate a constitutively active ERK2 kinase (20) and most recently also to obtain activated JNK (25).
Here we generated an activated version of SAPK/JNK through the fusion of SAPK␤ in-frame to its specific upstream activator, MKK7. A decapeptide linker (EG) 5 providing a flexible region was inserted between the two coding sequences to allow for proper folding of the domains (Fig. 1A). In addition. we incorporated a Myc tag at the N-terminal end to facilitate detection of the fusion protein (21). Western blot analysis of transiently transfected NIH 3T3 cells showed the expression of the expected 90-kDa SAPK␤-MKK7 fusion protein and also its kinase dead form, SAPK␤-MKK7 KD, in which the critical lysine residues in the ATP-binding sites of SAPK␤ (K55A and K56A) and MKK7 (K76E) had been replaced by nonphosphorylatable amino acids (Fig. 1B, lanes 2 and 3).
SAPK␤-MKK7 Fusion Protein Is Constitutively Active and Does Not Stimulate Endogenous SAPK/JNKs-MKK7 activates SAPK/JNK by phosphorylating a threonine-proline-tyrosine (TPY) motif in the central part of the protein. The presence of phosphorylation at these sites is indicative of the activated form of the kinase. To analyze whether SAPK␤-MKK7 is constitutively active in mammalian cells, the construct was transiently transfected into NIH 3T3 fibroblasts. Before the preparation of lysates, the cells were cultured for 24 h in the presence of 0.05% serum to shut down serum growth factor signaling. As shown in Fig. 2, the inactive SAPK␤-MKK7 KD fusion protein (lane 2) was present as an unphosphorylated protein in cells, whereas SAPK␤-MKK7 was constitutively phosphorylated (lane 1). Furthermore, the expression of the constitutively active SAPK␤-MKK7 fusion protein did not affect the activation status of the endogenous SAPK/JNKs. Comparable basal levels of both p54 and p46 SAPK phosphorylation were detectable in the cell lysates of SAPK␤-MKK7 and SAPK␤-MKK7 KD or in serum-starved NIH 3T3 cells (Fig. 2,  lanes 1, 2, and 3). An increase in phosphorylation could only be observed when cells were treated with the known JNK activators arsenite (lane 4) or anisomycin (lane 5) but not after the overexpression of SAPK␤-MKK7 (lane 1). To confirm the identity of the p46/p54 SAPK/JNK proteins, their migratory and immunoblot behavior was compared with SAPK/JNKs present in commercially available control lysates and found to be iden-tical (data not included in Fig. 2). An analysis of MAPK signaling has revealed a high degree of cross-talk between these related signaling pathways. Using the same approach as above, we therefore tested these lysates for a potential activation of p38 or ERK1/2. Consistent with a previous report (25), we failed to detect any effect of SAPK␤-MKK7 on the activity of these MAPKs (data not shown).
SAPK␤-MKK7 Possesses c-Jun Kinase Activity in Vivo-To test whether the constitutively active SAPK␤-MKK7 fusion protein is able to phosphorylate the SAPK/JNK substrate c-Jun in vivo, SAPK␤-MKK7 and c-Jun were co-expressed in NIH 3T3 cells. An immunoblot was performed using an antibody that recognizes unphosphorylated c-Jun as well as the slower migrating phosphorylated and transcriptionally active form of c-Jun (Fig. 3). Whereas the kinase dead SAPK␤-MKK7 fusion protein caused no activation of c-Jun in vivo (Fig. 3, lane 1), the expression of constitutively active SAPK␤-MKK7 lead to significant phosphorylation of c-Jun as indicated by the electrophoretic mobility shift (Fig. 3, lane 2 5 g each). Before protein analysis, cells were grown in 0.05% serum for 24 h. As controls, empty vector-transfected, serum-starved NIH 3T3 cells were treated with the potent SAPK/JNK activators arsenite (500 M) or anisomycin (10 g/ml) for 45 min. Following SDS-PAGE, all samples were analyzed by immunoblotting with an antiphospho JNK antibody that recognizes the activated forms of p46 and p54 SAPK/JNKs. The same membrane was stripped and reprobed with an anti-JNK1 antibody to detect the expression of endogenous p46 and p54 SAPK/JNKs. Afterward, the blot was stripped again, and the upper part was incubated with an anti-Myc antibody to detect the expression of SAPK␤-MKK7 and SAPK␤-MKK7 KD. The lower part of the blot confirms equal loading to incubation with an anti-ERK1 antibody to visualize endogenous ERK1 (p44) and ERK2 (p42).
to be able to phosphorylate both serine residues because a third intermediate c-Jun band could be observed most likely representing the singly phosphorylated form of c-Jun. The activation of c-Jun by SAPK␤-MKK7 could be completely inhibited when the SAPK/JNK binding domain of JIP-1 is coexpressed (Fig. 3,  lane 3). JIP-1 has been described as a scaffold protein that interacts with different components of its SAPK/JNK signaling pathway, and overexpression of the SAPK/JNK binding domain blocked signaling by JNK/SAPK, presumably by the cytoplasmic retention of the kinase (26). These results are in support of our in vitro data where SAPK␤-MKK7 phosphorylated recombinant GST-c-Jun (amino acids 1-135) in an immune complex kinase assay (21).
Subcellular Localization of SAPK␤-MKK7-Upon stimulation, SAPK/JNK kinases translocate from the cytoplasm to the nucleus where they phosphorylate a variety of transcription factors including members of the Jun family ATF2 (activating transcription factor 2) and Elk1 but also p53, c-Myc, and NFAT4 (nuclear factor of activated T cells 4) (27)(28)(29)(30)(31). In stimulated cells, a small fraction of activated SAPK/JNK also remains in the cytoplasm to potentially interact with or phosphorylate proteins that are present in this compartment, for instance 3pK (24) or MAP-2 (microtubules-associated protein 2) (32). To analyze the distribution of SAPK␤-MKK7, NIH 3T3 cells transiently transfected with SAPK␤-MKK7 or the kinase dead mutant SAPK␤-MKK7 KD were stained with an antibody directed against the Myc tag epitope (9E10) present in the fusion protein. SAPK␤-MKK7 was found predominantly in the nucleus (Fig. 4). Also, in cells expressing lower amounts of the protein the majority of SAPK␤-MKK7 was always detected in the nucleus, whereas its inactive mutant form (Fig. 4) was in all cases totally excluded from the nucleus. These findings are in good agreement with the data of Zheng et al. (25).
Tetracycline-regulated SAPK␤-MKK7 Expression Results in c-Jun Phosphorylation in Vitro and in Vivo-To further elucidate the physiological function of SAPK/JNK kinase in mammalian cells, we generated an inducible NIH 3T3 tet-off cell line expressing SAPK␤-MKK7. For this purpose SAPK␤-MKK7 or its inactive form, SAPK␤-MKK7 KD, were cloned into the tet-off-inducible expression vector pREV-TRE and transfected into the previously established NIH 3T3 fibroblast cell line NIH 5.15 (22). Following selection, we identified several clones in which the expression of SAPK␤-MKK7 or SAPK␤-MKK7 KD was induced upon the removal of tetracycline. In comparison to transiently transfected cells, all inducible NIH3T3 tet-off clones expressed SAPK␤-MKK7 at lower amounts. Over time the extent of SAPK␤-MKK7 expression gradually declined (data not shown), suggesting that cells cannot tolerate higher amounts of constitutively active SAPK␤-MKK7 for longer time periods.
To test whether SAPK␤-MKK7 is able to phosphorylate c-Jun in vitro, we performed an immune complex kinase assay using GST-c-Jun (amino acids 1-135) as the substrate. Three independent NIH 3T3 tet-off clones were grown in the absence of tetracycline to induce the expression of SAPK␤-MKK7. In parallel, cells from the same clones were maintained in the presence of tetracycline to shut down protein synthesis. SAPK␤-MKK7 was immunoprecipitated using the anti-Myc antibody (9E10) and assayed for kinase activity (Fig. 5). Immunoprecipitates from all three clones were able to phospho- rylate c-Jun in vitro (Fig. 5). Reprobing of the same blot with the 9E10 antibody confirmed the presence of the SAPK␤-MKK7 protein (Fig. 5).
We also analyzed the phosphorylation of c-Jun under in vivo conditions. Different NIH 3T3 tet-off cell line clones harboring SAPK␤-MKK7 were transiently transfected with c-Jun and cultivated for 24 h in 0.3% serum (Fig. 6). Again, the induction of SAPK␤-MKK7 expression upon tetracycline removal induced an electrophoretic mobility shift of c-Jun.

Constitutively Active SAPK␤-MKK7 Induces Morphological Transformation and Supports Growth in Soft Agar-Having
established that the SAPK␤-MKK7 fusion functions as an activated JNK/SAPK mutant, we used this tool to analyze the effect of constitutive JNK/SAPK signaling on cellular transformation. As early as 4 days after tetracycline removal many cells (derived from clone 3) growing under subconfluent conditions in 10% serum medium displayed a transformed phenotype (Fig. 7, clone 3). In contrast with the flat phenotype normally observed with NIH 3T3 fibroblasts (Fig. 7A, ϩ tet), these cells were spindle shaped (Fig. 7A, Ϫ tet). The phenotype was reversed when cells were switched to tetracycline-containing medium (data not shown). The same results were obtained with two independently isolated NIH 3T3 tet-off cell clones (clones 5 and 33). Following up parallel dishes for a longer time period, we could detect the formation of foci (Fig. 7B, Ϫ tet 115.0 Ϯ 21.6 N ϩ tet 77.5 Ϯ 13.2 foci per dish), which, in contrast to v-Rafinduced foci (Fig. 7B), remained relatively small in size. However, a transient transfection of SAPK␤-MKK7 into NIH 3T3 cells did not result in foci formation. Western blot analysis 18 days after transfection of these cells failed to detect any expression of SAPK␤-MKK7, suggesting that appropriate selection conditions would be required.
To further confirm the transforming potential of constitu-tively active SAPK␤-MKK7, we analyzed SAPK␤-MKK7 expressing cells for their potential to grow in soft agar. In a first attempt the inducible NIH 3T3 tet-off cell line expressing SAPK␤-MKK7 was used. 1000 cells (clone 3) were plated in soft agar in the presence or absence of tetracycline (Fig. 7C, Ϯ tet). Approximately 12 days later, the formation of small cell clusters became clearly visible (Fig. 7C, Ϫ tet). On average, ϳ19.9 Ϯ 4.6% of the cells seeded yielded colonies in the absence of tetracycline in comparison to 5.4 Ϯ 0.6% when grown in the presence of tetracycline. Additionally, we transiently transfected NIH 3T3 fibroblasts with an expression vector encoding SAPK␤-MKK7 or the empty vector (Fig. 7D). After 1 week selections with G418, cells were seeded in soft agar. Under these experimental conditions SAPK␤-MKK7-expressing cells also showed growth in soft agar (SAPK␤-MKK7 13.3% Ϯ 2.6) in contrast with the vector control (pcDNA3 6.1% Ϯ 2.8) or the inactive fusion protein (SAPK␤-MKK7 KD 3.0% Ϯ 1.3) (Fig. 7D).
Constitutively Active SAPK␤-MKK7 Causes Disassembly of the F-actin Cytoskeleton in NIH 3T3 Fibroblasts-Morphological changes seen in transformed cells are linked to changes in the cytoskeleton (33). Small GTPases of the Rho family are critical in this process. Interestingly, Rac and Cdc42, which regulate the formation of filopodia and lamellipodia and are also important for cell migration, have been identified as upstream activators of SAPK/JNK kinases (34). Therefore, we performed analysis to see whether active SAPK␤-MKK7 could lead to alterations in the cytoskeleton. Cells of the inducible NIH 3T3 tet-off cell line (clone 3) were grown for 48 h in the presence or absence of tetracycline. Afterward, the F-actin of the cytoskeleton was stained with phalloidin (Fig. 8A). Cells depicted in Fig. 8 expressed SAPK␤-MKK7 at relatively high amounts. In the presence of tetracycline, which blocks SAPK␤-MKK7 expression, cells showed an intact F-actin cytoskeleton (Fig. 8A, ϩ tet). In contrast to this, the expression of SAPK␤-MKK7 caused a partial or in some cases even a complete loss of stress fibers (Fig. 8A, Ϫ tet). To control for SAPK␤-MKK7, expression cells were stained in parallel with the anti-Myc antibody (9E10) (Fig. 8A, Ϯ tet). Similar effects of activated JNK were also seen in cells after selection for growth in soft agar (data not shown). Additionally, the transient expression of SAPK␤-MKK7 resulted in a dramatic reduction to a complete loss of actin stress fibers, whereas the inactive form, SAPK␤-MKK7 KD, had no such effect (Fig. 8B). In our experiments we also included a constitutively active MKK7 SE3, which was only a weak activator of endogenous SAPK/JNK (data not shown). The majority of cells expressing MKK7 SE3 had their F-actin cytoskeletons preserved, and only a few high expressing cells were seen that had a completely remodeled cytoskeleton. In contrast, when MKK7 (SE3) was coexpressed with SAPK␤ the disassembly of the F-actin stress fibers was pronounced (data not shown). Identical results were obtained when cotransfecting NIH 3T3 cells with the upstream activators of SAPK/JNKs, MLK3 or Cot (35-37), together with SAPK␤ (data not shown). Following expression of a kinase dead version of MKK7 SA3, all inspected cells demonstrated an intact cytoskeleton (Fig. 8B).
Analysis of the subcellular location of SAPK␤-MKK7 had shown that the kinase was predominantly located in the nucleus with only a fraction of the protein remaining in the cytosol. We thus tried to determine whether the cytosolic fraction of SAPK␤-MKK7 was responsible for the observed effects on the cytoskeleton. As seen in the bottom panel of Fig. 8B, the coexpression of SAPK␤-MKK7 together with the scaffold protein JIP-1 prevented the translocation of the constitutive active fusion protein into the nucleus. Inspection of these cells, which excluded SAPK␤-MKK7 from the nucleus, revealed no detectable remodeling of the cytoskeleton. This finding could suggest that perhaps the nuclear fraction of SAPK␤-MKK7 is required for the observed effect. However, we can not exclude the possibility that JIP-1 sequesters SAPK␤-MKK7 away from its cytoskeletal target(s) and therefore prevents disassembly of the actin stress fibers. Apart from modulating cellular localization, the effects of JIP-1 may also be the direct result of its inhibition of SAPK␤-MKK7 activity (62). 2

DISCUSSION
In this report we fully characterized a constitutively active version of SAPK␤ and demonstrated that the expression of this mutant caused many aspects of morphological transformation in NIH 3T3 cells usually associated with oncogenes. The fusion of SAPK␤ to its specific upstream activator, MKK7, resulted in a constitutively active JNK/SAPK (21) (Figs. 1 and 2). SAPK␤-MKK7 phosphorylated the c-Jun protein in vitro and also induced an electrophoretic mobility shift of c-Jun when expressed in NIH 3T3 cells, which is indicative of an in vivo phosphorylation (Figs. 3 and 5). The inducible expression of SAPK␤-MKK7 in the NIH 3T3 fibroblasts caused morphological transformation and allowed cells to grow in soft agar (Fig. 7). Additionally, the presence of activated JNK resulted in the disassembly of actin stress fibers in NIH 3T3 cells (Fig. 8).
Thus our experiments revealed a transforming potential for constitutively active SAPK␤-MKK7 that is, however, much weaker than that of dominant oncogenes such as ras or raf. Oncogenic transformation is characterized by enhanced growth rates but also by morphological changes resulting from alterations in the actin cytoskeleton and adhesive interactions (38). Small GTPases of the Ras superfamily play key roles in these processes and also have been shown to activate SAPK/JNK; members of the Ras subfamily regulate cell proliferation and differentiation, whereas members of the Rho subfamily (RhoA, RhoB, RhoC, Rac1, Rac2, and Cdc42) were first identified as regulators of the actin cytoskeleton but also affect gene expression and proliferation (39). RhoA controls the assembly of actin stress fibers and focal adhesion complexes (40,41), and Rac regulates actin filament accumulation at the plasma membrane to produce lamellipodia and membrane ruffles, whereas Cdc42 stimulates the formation of filopodia (34,42,43). It has been reported that the transient expression of activated Ras, Rac, and RhoA in cells results in the formation of stress fibers (44 -46), whereas chronic stimulation, as seen in Ras-transformed cells, leads to the inhibition of stress fiber formation (47). The potential mechanism responsible for the absence of stress fibers in cells transformed by Ras and Raf has been dissected recently (63). Sustained ERK activation down-regulates p160 Rho-associated coiled-coil-containing protein kinase (ROCK) and Lin11, Isl-1, and Mec-3 domain (LIM) kinase, two Rho effectors required for actin stress fiber formation (39). In FIG. 7. SAPK␤-MKK7 demonstrate properties of growth transformation. A, constitutively active SAPK␤-MKK7 induces morphological transformation of NIH 3T3 tet-off cells. Cells from clone 3 were cultivated for 4 days in 10% serum in the presence (ϩ) or absence (Ϫ) of tetracycline. B, SAPK␤-MKK7 induces foci formation in the NIH 3T3 tet-off cell line. Parallel dishes from the experiment described in A were grown for 24 days in the presence (ϩ) or absence (Ϫ) of tetracycline (2 g/ml) in 10% serum. Plates were fixed with methanol and stained with Giemsa. For the focus formation assay, the same number of cells was plated in triplicate onto 90-mm dishes. The number of foci on each dish was counted. 115.0 Ϯ 21.6 foci were formed by SAPK␤-MKK7-expressing cells (Ϫtet) in comparison to 77.5 Ϯ 13.2 foci in the presence of tetracycline (ϩ tet). NIH 3T3 cells stably expressing oncogenic v-Raf were used as a positive control (shown only as a section). C, SAPK␤-MKK7 expression induces anchorage-independent growth. 1000 cells from the inducible NIH 3T3 tet-off cell line expressing SAPK␤-MKK7 (clone 3) were seeded into 0.3% soft agar over a 0.6% agar bottom layer. Colonies were photographed on day 18. In the absence of tetracycline, cells expressed SAPK␤-MKK7 and formed little colonies (Ϫ tet 19.9% Ϯ 4.6 N ϩ tet 5.4% Ϯ 0.6; n ϭ 3). D, SAPK␤-MKK7 transfected into NIH 3T3 cells results in anchorage-independent growth. 2 ϫ 10 5 cells were transiently transfected with 1.5 g of the following plasmids: pcDNA3neo encoding SAPK␤-MKK7 (13.3% Ϯ 2.6), SAPK␤-MKK7 KD (3.0% Ϯ 1.3), or the empty expression vector (6.1% Ϯ 2.8). One week after selection with G418 (450 g/ml), cells were pooled and 5000 cells per dish were seeded in soft agar (n ϭ 3). Stably transfected NIH 3T3 cells expressing oncogenic v-Raf were used as a positive control. All pictures were taken after 18 days in soft agar addition, oncogenic Raf blocks upstream activators of Rac and thus impairs Rac-mediated adhesion of cells to the extracellular matrix (63). Interference with Rac activation may involve the transcriptional down-regulation of Tiam1, an activator of Rac (48).
MAP kinases function as effectors of the Ras and the Rho subfamily of small GTPases. An essential role for the classical mitogenic cascade Ras 3 Raf 3 MEK 3 ERK in oncogenesis has been established. MEK/ERK signaling is required for transformation by most oncogenes (36, 49 -51), and the expression of constitutively active mutants of ERK or MEK is sufficient for the establishment of the transformed phenotype (52). Recently published data suggest a functional role for SAPK/ JNK signaling in oncogenic transformation based on studies with dominant negative mutants that block transformation by oncogenes like ras (53), met (13), and bcr-abl (14,15). Additional evidence for a functional role of SAPK/JNK in this process has been obtained by the use of a dominant negative mutant of c-Jun that efficiently blocked transformation by various classes of oncogenes (14,19,54,55). A role for SAPK/JNK in cellular transformation is further substantiated by the demonstration of the albeit limited oncogenic potential of kinases, whose expression results in the activation of SAPK/JNKs such as MLK3. The overexpression of MLK3 in NIH 3T3 fibroblasts leads to a similar phenotype as seen with SAPK␤-MKK7 in our experiments (36). The expression of MLK3 also results in a low level activation of MEK/ERK, which was required to achieve the fully transformed phenotype because the treatment of MLK3-transformed NIH 3T3 cells with synthetic MEK inhibitors resulted in a partial reversion of the transformed phenotype (36). In the meantime, we have also begun to directly test the tumorigenic potential of SAPK␤-MKK7 in the nude mouse model. In agreement with the in vitro data presented in this paper, the injection of SAPK␤-MKK7-expressing fibroblasts resulted in the establishment of a well defined fibrosarcoma, although with a longer latency than in the case of v-raf transformed cells (17 days for v-raf versus 36 days for SAPK␤-MKK7. 3 Tumor cells readily could be taken into culture, and, using a PCR approach, immunoblot analysis, and direct staining of cells with the Myc antibody 9E10, we could confirm the expression of SAPK␤-MKK7. Thus, the expression of activated JNK is sufficient to initiate tumor development in vivo. However, an analysis of tumors induced by the ERK-and JNKactivating kinase Cot demonstrated the lack of activation of SAPK/JNK. 4 Thus, although the expression of activated SAPK/ JNK induces certain aspects of morphological transformation in vitro, SAPK/JNK signaling may be redundant or deleterious in later stages of tumor growth in vivo. A direct link between actin organization and SAPK/JNK has not been definitively established. The Drosophila JNK (DJNK) homologue Basket plays a role in the regulation of cell shape changes and actin reorganization in the process of dorsal closure during embryogenesis (56,57). Recently, p150-Spir was found in a yeast two-hybrid screen as a new substrate of DJNK (21). It belongs to the Wiscott-Aldrich syndrome protein (WASP) homology domain (WH2) family of proteins involved in actin reorganization (58,59). N-terminal sequences of Spir can interact with Rho family GTPases (60). p150-Spir, which colocalizes with F-actin in NIH 3T3 cells, acts as an initiator of the actin polymerization (21). Currently, it is not known whether SAPK/JNK could function as a positive or negative regulator of p150-Spir. Additionally, DJNK is known to directly phosphorylate and modify cytoskeletal components involved in dorsal closure such as Zipper (nonmuscle myosin), Coracle (Drosophila homologue of the vertebrate band 4.1 cytoskeletal protein), Inflated, or Myospheroid (integrins involved in cell adhesion). Alternatively, DJNK (or mammalian JNK/SAPK) could also modify the activity of transcription factors that are necessary for the process of dorsal closure (61).
The microinjection of activated Rac and Cdc42 into cells causes changes of the actin structures within minutes before de novo protein synthesis starts (2,3). Contrary to this, our data obtained after artificial retention of activated SAPK␤ in the cytoplasm by the overexpression of JIP-1 (Fig. 8) suggest that perhaps nuclear translocation, resulting most likely in the regulation of gene transcription, is required for the observed effects on the cytoskeleton. The differences may be explained in part by the fact that multiple effectors have been described for Rac/Cdc42 that may independently affect the process of cellular transformation, whereas in our studies we analyzed an isolated pathway. However, it is also possible that JIP-1 merely sequesters SAPK␤-MKK7 away from its cytoskeletal target(s). Furthermore, an inhibitory effect of JIP-1 on SAPK/JNK activity can not be excluded (62). 2 Preliminary experiments using the inhibitors actinomycin D (0.1 g/ml for 16 h) and cycloheximide (1 g/ml for 16 h) support the notion that at least de novo protein synthesis is not required for SAPK␤-MKK7-induced disassembly of actin stress fibers (data not shown).
In summary, the data presented here provide direct evidence that constitutively active SAPK␤-MKK7 leads to a morpholog-ically transformed phenotype when expressed in NIH 3T3 fibroblasts. Its transforming potential is weak in comparison to classical oncogenes like ras and raf. This may be due to the fact that SAPK␤-MKK7 expression only recapitulates some aspects of transformation such as the disassembly of stress fibers but is unable to induce the whole set of events required for establishing a transformed phenotype, e.g. enforced cell cycle progression. Despite these limited in vitro effects, our preliminary in vivo data suggest that the expression of activated JNK is able to trigger tumor formation in vivo.