Physical Association with WWOX Suppresses c-Jun Transcriptional Activity

WWOX encodes a WW domain–containing oxidoreductase tumor suppressor that spans the second most active fragile site in the human genome, FRA16D (1, 2). WWOX expression is reduced or lost in a wide range of human tumors including lung, gastric, and cutaneous squamous cell carcinoma (3–5). Initially, WWOX has been cloned as a candidate tumor suppressor because the gene is located in a chromosomal region with high incidence of loss of heterozygosity in preinvasive breast carcinoma and of homozygous deletions in adenocarcinomas (1, 2). Indeed, ectopic expression of WWOX in cancer cells lacking WWOX expression induces apoptosis in vitro and suppresses tumorigenicity in athymic nude mice (6).

WWOX protein harbors two WW domains and a short-chain dehydrogenase/reductase domain (1, 2). We have previously shown that the first WW domain of WWOX mediates protein-protein interaction with proline-tyrosine (PY)–rich motifs class I, PPxY (7). WW domains are present in numerous and unrelated proteins such as Yes-associated protein (YAP), Itch, Nedd4, and others (8). It is interesting to note that the PY motif is present in the sequence of many transcription factors. This observation may suggest that the PY motif plays a role in mediating transcription regulation. Indeed, YAP has been shown to function as transcriptional coactivator of several target genes such as p73 and ErbB4 (9, 10). On the other hand, the E3 ligase Itch, which contains four WW domains, regulates the stability of transcription factors, such as c-Jun and p63 (11, 12), through regulation of their protein turnover. Similarly, WWOX may regulate transcriptional activity of target genes through its WW domain-PY interaction. We have recently reported that WWOX physically interacts with and suppresses the transactivation ability of p73, AP2γ, and ErbB4 (7, 13, 14). Furthermore, we have also shown that WWOX competes with the WW domain–containing protein YAP for binding target proteins and thus regulates their transactivation function (14). These findings may suggest that WW domain–containing proteins may compete to regulate multiple cellular processes, such as transcription and protein turnover, through their ability to interact with target proteins.

Here, we describe a physical and functional interaction between the WWOX tumor suppressor and the activating protein-1 (AP-1) transcription factor component c-Jun. c-jun proto-oncogene is a transcription factor that regulates proliferation, differentiation, and apoptosis in multiple cell types (reviewed in ref. 15). c-Jun is highly induced following exposure to UV radiation due to its rapid phosphorylation and activation by c-Jun NH₂-terminal kinase (JNK), which in turn elicits its UV response through regulation of transcription of downstream targets such p53 and p21 (15). WWOX, via its first WW domain, associates with c-Jun following UV radiation and is able to repress its transcription, thus suggesting that WWOX might be involved in regulating c-Jun transactivation activity.

Plasmids. Myc-WWOX plasmids were previously described (7). Glutathione S-transferase (GST)-WWOX was obtained by subcloning the cDNA of human WWOX into BamHI and NotI sites of pEBG (New England Biolabs, Ipswich, MA). Human cDNA of c-Jun was cloned into pCMV-hemagglutinin (pCMV-HA) vector (Clontech, San Diego, CA) using EcoRI and BamHI sites. pFC-MEKK was purchased from Stratagene (La Jolla, CA). GST-WWOXY33R, Myc-WWOXW44, P47L, and HA-c-JunY170A were obtained using the QuikChange II Site-Directed Mutagenesis Kits (Stratagene).

Cell culture and transfection. Human embryonic kidney cells (HEK293), HeLa, and immortalized human keratinocytes HaCaT (gifts from Dr. Eleonora Candi, IDI-IRCCS, c/o University of Rome ‘Tor Vergata’, Rome, Italy) were grown in DMEM (Sigma, St. Louis, MO). All cell lines were grown at 37°C in a humidified atmosphere of 5% (v/v) CO₂ in air. Transient transfections were done with Fugene6 reagent (Roche Diagnostics, Indianapolis, IN) according to the manufacturer's instructions.

Immunoprecipitation, GST pull-down, and Western blot. As described (7), HEK293 cells were transiently transfected with the indicated mammalian expression plasmids and harvested 24 to 30 hours after transfection. Following preclearing for 1 hour at 4°C, we did immunoprecipitation by incubating 1 mg of whole-cell extracts with the indicated antibodies with rocking overnight at 4°C. GST pull-down was done by incubating 1 mg of whole-cell extracts with GST-beads with rocking overnight at 4°C. Antibodies used for immunoprecipitation were monoclonal anti-WWOX (7), monoclonal anti-HA (Covance, Princeton, NJ), and monoclonal anti-Myc (Santa Cruz Biotechnology, Santa Cruz, CA). Antibodies used for immunoblot were as follows: c-Jun, phospho-c-Jun (Santa Cruz Biotechnology), polyclonal rabbit anti-GST-WWOX (gift from Dr. Kay Huebner), HA-horseradish peroxidase (HRP; Roche), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Calbiochem, San Diego, CA).

Immunofluorescence. As described in ref. 7, HeLa or HaCaT cells were blocked with 5% goat serum in PBS and then incubated with mouse anti-WWOX and anti-c-Jun antibodies. After washing thrice in PBS, sections were incubated for 1 hour with secondary antiserum (goat anti-mouse Alexa Fluor 488 or goat anti-rabbit Alexa Fluor 568) and the nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Sections were then mounted using Vectashield antifade reagent (Vector Labs, Burlingame, CA) and analyzed with fluorescent microscope (Zeiss, Thornwood, NY).

Luciferase assay. HEK293 cells seeded in 12-well plates were cotransfected with indicated plasmids together with pAP-1-Luc reporter and pRL-TK Renilla reporter constructs (Stratagene). Cells were collected 24 hours later and Firefly and Renilla luciferase activities were assayed with Dual-Luciferase Assay System (Promega, Madison, WI) and firefly luciferase activity was normalized to Renilla luciferase activity. All experiments were done at least thrice.

Physical interaction between WWOX and c-Jun. We have recently shown that the WW domain–containing oxidoreductase WWOX associates with PY motifs of p73 and AP2γ and suppresses their transcriptional ability (7, 13). Because c-Jun also contains a PPVY motif in its sequence and has been shown to interact with WW domain–containing proteins (11), we decided to examine whether WWOX physically associates with c-Jun and regulates its transactivation activity. To examine whether WWOX and c-Jun physically interact, HEK293 cells were transiently cotransfected with HA-c-Jun and Myc-WWOX constructs. Cell lysates were immunoprecipitated with anti-HA or anti-Myc antibodies followed by immunoblotting with HRP-conjugated antibody to HA or Myc. Our results show that WWOX interacts with c-Jun as determined by immunoprecipitation with anti-Myc and immunoblotting with anti-HA antibody (Fig. 1A,, lane 1, top). As a control, there were no detectable complexes in anti–immunoglobulin G immunoprecipitates (Fig. 1A,, lane 2). The same complexes were detected in reverse by immunoprecipitation with anti-HA and immunoblotting with anti-Myc antibody (Fig. 1A , lane 3, bottom).

Because mitogen-activated protein kinase kinase kinase 1 (Mekk1) activates JNK, which subsequently causes phosphorylation and activation of c-Jun (15), we investigated the effect of c-Jun phosphorylation on the interaction with WWOX. To this end, HEK293 cells were cotransfected with a mammalian GST-WWOX expression vector, HA-c-Jun, and MEKK1, followed by GST-pull down and immunoblotting with anti-HA. We observed that overexpression of MEKK1, which mediates c-Jun phosphorylation, significantly enhances WWOX-c-Jun complex formation (Fig. 1B , lane 4 versus lane 2).

To verify the physiologic relevance of WWOX-c-Jun interaction, we examined the association of the endogenous proteins. As mentioned above, exposure to UV radiation results in rapid JNK activation that eventually leads to phosphorylation of c-Jun and enhancement of its transcriptional activity (15). Therefore, we examined association of WWOX-c-Jun in the human HaCaT keratinocyte cell line following exposure to UV radiation. HaCaT cells were treated with increased levels of UVC and, 2 hours later, protein levels were assessed by Western blot. As shown in Fig. 1C, increased expression of c-Jun was observed following irradiation in a dose-dependent manner. In addition, we detected a slightly higher band of c-Jun, which most likely represents the phosphorylated form of c-Jun (Fig. 1C). No changes were observed in endogenous WWOX expression levels (Fig. 1C). To determine whether endogenous WWOX interacts with endogenous c-Jun following UV irradiation, we carried out immunoprecipitation experiment using monoclonal anti-WWOX antibody. Immunoblot with anti-Jun antibody revealed that WWOX interacts with c-Jun following UV irradiation in a dose dependent manner (Fig. 1D). This result confirms a WWOX-c-Jun interaction in a physiologic setting.

We next proceeded to map regions responsible for WWOX-c-Jun association. We have previously shown that a point mutation in tyrosine 33 (Y33), Y33R, in the first WW domain of WWOX is sufficient to abolish interaction of WWOX and its partners (7, 13, 14). Therefore, we tested whether this mutant form of WWOX (WWOXY33R) can bind c-Jun. HEK293 cells were cotransfected with expression vectors encoding HA-c-Jun, Mekk1, and GST-WWOX or GST-WWOXY33R and extracts were pulled down with GST beads and subsequently blotted with anti-HA antibodies. We found that Y33R point mutation in the first WW domain of WWOX significantly reduced WWOX association with c-Jun (Fig. 2A,, top, lane 4 versus lane 3). The same decrease in interaction was shown when extracts were blotted with anti–phospho-Jun antibody (Fig. 2A , middle, lane 4 versus lane 3).

In addition, we generated a point mutation in the terminal tyrosine (Y) of the PPVY motif of c-Jun and determined the ability of this mutant to bind WWOX. The c-Jun proline-containing motif Y170 was mutated to alanine (A), producing the expression vector HA-c-JunY170A. To determine whether WWOX interacts with c-JunY170A, we cotransfected HEK293 cells with both expression vectors in the presence of Mekk1. Figure 2B shows that point mutation in the proline-containing motif of c-Jun completely abolishes WWOX-c-Jun association (Fig. 2B , lane 4 versus lane 6).

WWOX sequesters c-Jun in the cytoplasm. c-Jun is a transcription factor that localizes in the nucleus where it forms AP-1, together with c-Fos, and regulates gene expression (15, 16). On the other hand, WWOX is a cytoplasmic protein (7, 13, 14). We therefore examined the effect of WWOX-c-Jun interaction on the subcellular localization of WWOX and c-Jun using immunofluorescence. HA-c-Jun alone or with MYC-WWOX, in the presence or absence of Mekk1, was transiently expressed in HeLa cells. Localization of the HA- or Myc-tagged proteins was then determined by immunofluorescent staining using the appropriate antibodies, as described in Materials and Methods. Although the majority (∼90%) of c-Jun localizes to the nucleus (Fig. 3A), some cells also showed weak cytoplasmic staining when MEKK1 was coexpressed in HeLa cells (Fig. 3B). In contrast, exogenous WWOX is mainly detected in the cytoplasm (Fig. 3A and B). In cells cotransfected with Myc-WWOX and HA-c-Jun, 20% to 30% of cells showed cytoplasmic staining of c-Jun (Fig. 3A). When Mekk1 was coexpressed, >50% of cells displayed cytoplasmic staining of c-Jun where it colocalized with WWOX (Fig. 3B).

We further studied the subcellular localization of endogenous proteins in HaCaT keratinocytes after UV radiation. HaCaT keratinocytes were treated with UVC (100 J/m²) and, 2 hours later, cells were immunostained. As shown in Fig. 3C, following irradiation, some cells display cytoplasmic c-Jun that colocalizes with WWOX staining. Altogether, these results indicate that WWOX-c-Jun interaction takes place mainly in the cytoplasm.

WWOX suppresses the transcriptional activity of c-Jun mediated through JNK-activation. Because WWOX-c-Jun association alters the nuclear localization of c-Jun, we next examined whether this interaction affects the transcriptional activity of c-Jun. Mekk1 activates JNK and subsequently causes the activation of c-Jun, which complexes with c-Fos, forming the activator protein-1 (AP-1) that regulates various cellular activity (15, 16). To examine the effect of WWOX-c-Jun association on c-Jun transactivation function, we used the AP-1-luciferase reporter (AP-1-Luc). When HEK293 cells were cotransfected with AP-1-Luc and empty vector, endogenous c-Jun caused very low activation of luciferase reporter, most likely due to the lack of JNK activation (Fig. 4A). Expression of AP-1-Luc and MYC-WWOX constructs resulted in decreased reporter activity (Fig. 4A), although fold activation was extremely low. Because JNK-induced phosphorylation turns c-Jun into a strong transcriptional activator, we repeated our analysis using Mekk1 in cotransfection. Indeed, following transfection of Mekk1, endogenous c-Jun increases luciferase activity by 68-fold compared with that in the absence of Mekk1 (Fig. 4B). In these experiments, expression of WWOX significantly suppressed the transactivation function of c-Jun (Fig. 4B). In contrast, expression of two WWOX mutants, which do not interact with c-Jun (Fig. 2A and data not shown), has no effect on c-Jun transactivation function (Fig. 4B), indicating that physical interaction with WWOX is responsible for transcriptional repression. These results indicate that WWOX can suppress c-Jun transactivation activity mediated by JNK activation.

WW domain class I recognizes and binds to the short PY oligopeptide motif PPxY to mediate protein-protein interactions (8). The PY motif is present in the transcription activation domains of a wide range of transcription factors including c-Jun, EGR-1, CCAAT/enhancer binding protein-α, and polyomavirus enhancer binding protein 2/core binding factor, suggesting that it plays an important role in transcriptional regulation. We have recently shown that WWOX suppresses transcriptional activity of target proteins (7, 13, 14). Our results of WWOX-c-Jun association further confirm these observations. WWOX expression is down-regulated following UV radiation, perhaps due to WWOX colocalization within a fragile site (17, 18). In addition, clinical evidence has suggested that WWOX expression is altered in cutaneous squamous cell carcinoma, which is highly associated with long-term exposure to sunlight (5). In our experimental setting, we did not observe a decrease in WWOX expression because we mainly used short-time exposure to UVC, which might not be enough to affect WWOX expression. Nevertheless, our data show that following stress conditions, such as UV exposure, WWOX may participate in the immediate response and modulate gene transcription.

One of the most potent stimuli that induce AP-1 dependent gene expression is short-wavelength UV light, which elicits a biological process called the UV response (15). Exposure to UV radiation results in rapid JNK activation, which eventually leads to c-Jun and activating transcription factor-2 phosphorylation, and hence enhances their transcriptional capacity (15). Regulation of c-Jun transcription is then determined through an autoregulatory mechanism and ubiquitination (15, 16). JNK-dependent phosphorylation has been shown to accelerate c-Jun degradation by allowing its recognition by the E3 ligase Fbw7-containing Skp/Cullin/F-box protein complex (SCF^FBW7; ref. 19). In addition, Gao et al. (11) have shown that c-Jun turnover is controlled through JNK-dependent phosphorylation of the E3 ligase Itch. Our findings suggest that c-Jun transcriptional activity can also be modulated by c-Jun association with WWOX. Following exposure to UV radiation, WWOX-Jun interaction is enhanced, and this association results in suppression of c-Jun transactivation ability. Along with other WW domain–containing proteins, WWOX might participate to modulate c-Jun transcriptional function.

The oncogenic function of c-Jun is evident through transcriptional regulation of many genes during cancer transformation (reviewed in ref. 20). c-Jun activity affects multiple cellular processes such as cell cycle, apoptosis, and growth (15, 20). Our data suggest that c-Jun oncogenic function might be suppressed through its association with the WWOX tumor suppressor.

Taken together, these data suggest an important role for the WWOX/c-Jun functional interaction and support the hypothesis that WWOX tumor suppressor acts as a transcriptional repressor of transactivator genes.

Grant support: Kimmel Scholar Award (R.I. Aqeilan), National Cancer Institute grants (C.M. Croce), and CLL Global Research Foundation grant (Y. Pekarsky).

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.

We thank Dr. John Hagan for the critical reading of the manuscript.