WW Domain–Containing Proteins, WWOX and YAP, Compete for Interaction with ErbB-4 and Modulate Its Transcriptional Function

Introduction

Gene expression can be regulated through a variety of mechanisms including the modulation of transcription factor localization ( 1). Transcription factors may associate with different cofactors that may repress or enhance their transcriptional activity. Modulation of subcellular localization of transcription factors in the nucleus or cytoplasm plays an important role in the integration of signal transduction and control of transcriptional machinery ( 1). Recently, we showed that WW domain–containing oxidoreductase (WWOX) functions as a regulator of subcellular localization of p73 and AP2γ transcription factors ( 2, 3). WWOX inhibits nuclear translocation of p73 and AP2γ hence suppressing their transactivating ability.

WWOX was originally cloned as a putative tumor suppressor gene that spans one of the most common active fragile sites in the human genome, FRA16D ( 4– 6). WWOX is located on chromosome 16q23.3 and exhibits genomic alterations in several cancer types ( 7) and a recent study showed a possible involvement of methylation in the regulation of WWOX expression ( 8). WWOX expression is frequently altered in tumor tissues ( 9– 13) and introduction of WWOX into WWOX-negative tumor cells resulted in tumor suppression and apoptosis ( 10, 14). The tandem WW domains of WWOX ( Fig. 1A ) play an important role in WWOX function. Both WW domains of WWOX contain the central core of two consecutive aromatic amino acids and therefore belong to the class I specificity of domains, which recognize ligands with the PPxY consensus motif (where P is proline, Y is tyrosine, and x is any amino acid; refs. 2, 15– 19). Recent mapping of the WW domain in the human proteome identified a repertoire of PPxY-containing ligands that bind to individual domains of WWOX ( 20). The first WW of WWOX bound 18 and the second WW bound 16 ligands, all with PPxY consensus. The mapping data clearly documented that although the second WW domain of WWOX contains a tyrosine in the place of the second conserved tryptophan, the signature residue directly involved in ligand binding ( 21), the specificity of the “WY domain” toward PPxY core was not changed ( 20).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Schematic illustration of WWOX and ErbB-4 proteins. A, structure of full-length of human WWOX. WWOX contains two WW domains and short dehydrogenase domain (SDR). Myc-WWOX, Myc-WWOXY33R, and WWOX-GFP constructs were used to express WWOX protein. B, structure of human YAP. There are two YAP isoforms. YAP2 contains two WW domains compared with one WW domain in YAP1 but both share the same Akt phosphorylation site and transcription activation domain. C, structure of human ErbB-4. All isoforms of ErbB-4 contain cysteine-rich domains, domains I to IV in the ectodomain, transmembrane region (TM), and the tyrosine kinase domain. In the CYT-2 isoform, the PI3K-binding motif is deleted. There are three PPxY motifs that are possible binding sequences for the class I WW domains. The second PPxY motif is included in the PI3K binding domain site that is deleted in the CYT-2 isoform. The COOH-terminal fragment (residues, 676-1292) that is known to translocate to the nucleus is indicated as CTF. CTF-ΔK (residues, 988-1292) lacks the tyrosine kinase domain but still have better transactivation ability than CTF.

Yes-associated protein (YAP) was originally cloned as a partner of Yes protein tyrosine kinase, which formed a complex with the SH3 domain of Yes ( 22). Detailed characterization of YAP lead to the identification of the WW domain as a modular protein domain with predilection toward proline-rich ligands ( 15, 23). Two WW domains of YAP ( Fig. 1B) also belong to the class I of domain specificity that recognizes proteins containing PPxY motifs. YAP has been characterized as a transcriptional coactivator of several transcription factors including Runx factors and p73 ( 24– 26). Two major isoforms of YAP were identified ( 27); YAP1 contains one WW domain, whereas YAP2 harbors two WW domains and is a more potent transcriptional coactivator when compared with YAP1 ( 28). Intracellular localization of YAP is regulated by Akt kinase phosphorylation at Ser¹²⁷ and subsequent binding of phosphorylated YAP to 14-3-3 protein ( 29).

Recently, YAP was shown to interact with ErbB-4 receptor acting as a transcriptional coactivator of the COOH terminus domain of ErbB-4 ( 28, 30). The HER4/ERBB4 gene is a member of type I receptor tyrosine kinase subfamily that includes epidermal growth factor receptor ERBB1, ERBB2, and ERBB3 (reviewed in ref. 31). ErbB-4 regulates cell proliferation and differentiation ( 32). After binding its ligand heregulin or following the activation of protein kinase C by 12-O-tetradecanoylphorbol-13-acetate (TPA), the ErbB-4 ectodomain is cleaved by a metalloprotease tumor necrosis factor-α–converting enzyme followed by γ-secretase cleavage ( 33). The cleavage by γ-secretase facilitates the translocation of the COOH-terminal fragment of ErbB-4 (CTF) to the nucleus where it may affect the transcription of target genes ( 34). CTF by itself does not possesses transactivating activity; however, in the presence of YAP, its transcriptional activity is revealed ( 28, 30). Different isoforms of ErbB-4 were previously described ( Fig. 1C). JM-a isoform is sensitive to cleavage, whereas the JM-b isoform is insensitive to the cleavage because of the sequence difference in the juxtamembrane region ( 35). CYT-1 and CYT-2 are two other isoforms of ErbB-4. CYT-1 contains a binding motif for the phosphatidylinositol 3-kinase (PI3K), whereas CYT-2 lacks this sequence ( 36, 37).

Because WWOX and YAP have the similar tandems of WW domains and interact with common protein targets but have different effects on transactivation, we decided to investigate whether WWOX and YAP expression have opposite effects on their target proteins. Here we report that WWOX physically associate with the cytoplasmic region of ErbB-4 via its first WW domain and suppresses transactivation activity of ErbB-4 mediated by YAP.

Materials and Methods

Cell culture and transient transfection. Human HEK293, HeLa, and MCF-7 cells were grown in RPMI supplemented with 10% fetal bovine serum (Invitrogen, Carlsbad, CA) and Gentamicin. Cells were maintained at 37°C in a water saturated atmosphere of 5% CO₂ in air. Overexpression of proteins was achieved by transient transfections using FuGENE 6 transfection reagent, according to the manufacturer's instructions (Roche Applied Science, Indianapolis, IN).

Plasmid construction. The mammalian expression plasmids encoding MYC epitope-tagged WWOX and pCMV-MYC-WWOXY33R were previously described (ref. 2; Fig. 1A). An additional Myc Tag was added in 3′ terminus of both WWOX constructs. Full-length WWOX cDNA was cloned into a pEGFP-N1 vector (Clontech, Palo Alto, CA). Full-length cDNA of YAP2 were cloned into a pCDNA4-HisMaxB vector (Omni-YAP2; Invitrogen) using standard protocols ( Fig. 1B). Full-length ErbB-4 with COOH-terminal HA epitope tag, ErbB-4 CTF, and CTF lacking the kinase domain (CTF-ΔK) were previously described (ref. 28; Fig. 1C). Full-length ErbB-4PY1-HA, ErbB-4PY3-HA and ErbB-4Py1,3-HA mutants were generated using Site Directed Mutagenesis kit (Stratagene, La Jolla, CA) according to manufacturer's instructions. The reporter plasmid pFR-luc expressing Firefly luciferase and contains a synthetic promoter with five tandem repeats of the yeast GAL4-binding sites and pTK-RL expressing Renilla luciferase as an internal control were purchased from Stratagene and Promega (Madison, WI), respectively. The reporter p53RE-Luc (Stratagene) carrying the p53-responsive element derived from p53 promoter was used to assay p73 transactivation.

Immunoprecipitation and immunoblot analysis. Cells were lysed and immunoprecipitations were carried out as previously described ( 2). Antibodies used were mouse monoclonal anti-HA (Covance, Princeton, NJ), mouse monoclonal anti-Omni (Santa Cruz Biotechnology), mouse monoclonal anti-Myc (Zymed, South San Francisco, CA) and mouse monoclonal anti-WWOX ( 10). Antibodies used for immunoblot were anti-HA-horseradish peroxidase (HRP, Roche Diagnostics), anti-Myc-HRP, and mouse monoclonal anti-Omni (Santa Cruz Biotechnology).

Immunofluorescence. Cells were seeded on fibronectin-covered cell culture slides (Becton Dickinson, San Jose, CA). HeLa cells were transfected with expression vectors encoding WWOX-GFP, Omni-YAP2, and ErbB-4-HA using FuGENE 6 (Roche Diagnostics). After 24 hours, cells were treated with TPA (100 ng/mL) or DMSO for 90 minutes and then cells were prepared as previously described ( 2). MCF-7 cells were treated with TPA and prepared as above. Antibodies used were mouse monoclonal anti-HA (Covance), rabbit polyclonal anti-ErbB-4 (Santa Cruz Biotechnology), mouse monoclonal anti- WWOX, anti-Rabbit Flouro-conjugated antibody and anti-mouse Texas red–conjugated antibody (Invitrogen). Cells were visualized using Zeiss LCM 510 confocal microscope.

Luciferase assays. In these experiments, we mainly used the CTF isoform that lacks kinase domain (CTF-ΔK) because it shows a better transactivation activity when compared with the entire CTF ( 28, 33). We also used pFR-luc reporter construct (Stratagene) containing five copies of Gal4 DNA binding sequences. 293 cells in 6-well dishes were transfected with the plasmids indicated in the figure legends using FuGENE 6 transfection reagent. Cells were collected 24 hours later and Firefly and Renilla luciferase activities were assayed with the Dual Luciferase Assay Reporter System (Promega). All experiments were done at least thrice.

Results

WWOX associates physically with ErbB-4. Because the WW domain–containing protein YAP associates physically and functionally with ErbB-4, we decided to investigate whether WWOX also interacts with ErbB-4 through its WW domains. We did coimmunoprecipitation experiments to assess this interaction. In all immunoprecipitation experiments, we used HA-tagged CYT-2 isoform of the full-length ErbB-4 that lacks the PI3K-binding motif ( Fig. 1C). We transiently cotransfected 293 cells with constructs expressing ErbB-4-HA and Myc-WWOX and cell lysates were immunoprecipitated with anti-HA or anti-Myc antibody followed by immunoblotting with HRP-conjugated anti-HA or anti-Myc antibodies. As shown in Fig. 2A , we were able to detect immune complexes between ErbB-4 and WWOX in vivo (lane 2, top and lane 4, bottom). As a control, there were no detectable complexes in anti-IgG immunoprecipitates ( Fig. 2A, lanes 3 and 6). To determine whether this interaction is mediated by the first WW domain of WWOX, we generated a mutated form of WWOX in which the conserved Tyr³³ (Y33) was substituted to arginine (R), WWOXY33R. The Y33 is located within the central core of consecutive aromatics of the WW domain and was shown to be a site of phosphorylation that regulates signaling of WWOX ( 2, 6). This Y-R mutation abolished interaction between WWOX and p73 or AP2γ proteins ( 2, 3) and at the molecular level it could be explained as a substitution that compromises the “aromatic cradle” structure that is essential for the formation of WW domain-ligand complexes ( 38). Similarly, the WW domain mutation in WWOX completely inhibited its binding to ErbB-4 ( Fig. 2A, lanes 5, top and lane 7, bottom). These results indicate that the first WW domain of WWOX is primarily responsible for the interaction with ErbB-4.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

WWOX binds ErbB-4 via its first WW domain. A, 293 cells were transiently transfected with the expression plasmids encoding ErbB-4-HA and Myc-WWOX or Myc-WWOXY33R. Twenty-four hours after transfection, whole cell lysates were immunoprecipitated (IP) with anti-Myc (lanes 2 and 5), anti-IgG (lanes 3 and 6), or anti-HA (lanes 4 and 7) antibodies. The immunoprecipitates were analyzed by immunoblotting (IB) with anti-HA (top) or anti-Myc (bottom)-HRP–conjugated antibodies. Lanes 1 and 8, lysate expression of ErbB-4 (top) and WWOX (bottom). B, 293 cells were transiently cotransfected with the expression plasmids encoding Myc-WWOX and ErbB-4-PY1-HA or ErbB-4-PY3-HA. Immunoprecipitation were as follows: anti-HA (lanes 2 and 5), anti-IgG (lanes 3 and 6), or anti-Myc (lanes 4 and 7) and immunoblotting were as in (A). Lanes 1 and 8, expression of ErbB-4 (top) and WWOX (bottom). C, 293 cells were transiently cotransfected with the expression plasmids encoding Myc-WWOX and ErbB-4-HA or ErbB-4-PY1,3-HA. Immunoprecipitation and immunoblotting were as in (B).

The cytoplasmic region of ErbB-4 CYT-2 isoform contains two PPxY motifs at positions 1031 to 1040 (NIPPPIYTSR; PY1) and 1280 to 1288 (LPPPPYRHR; PY3). The PY3 motif in ErbB-4 has been previously shown to mediate interaction with YAPs ( 28). To prove that WWOX binds to PPxY motif of ErbB-4, we replaced the tyrosine residues in the PPxY motif in the cytoplasmic region of ErbB-4 by alanine generating ErbB-4-PY1-HA and ErbB-4-PY3-HA. To determine whether WWOX coimmunoprecipitates ErbB-4 mutants, we cotransfected 293 cells with Myc-WWOX and either ErbB-4-HA-PY1 or ErbB-4-HA-PY3 and extracts were immunoprecipitated with anti HA or anti-Myc antibodies and detected using anti-HA (top) or anti-Myc (bottom) antibodies. Figure 2B shows that WWOX coimmunoprecipitates with both ErbB-4 mutants at a level comparable with its coimmunoprecipitation with wild-type ErbB-4 ( Fig. 2B, lanes 4 and 7, top and lanes 2 and 6, bottom). We then examined whether the WW domain of WWOX interacts with both the PPxY motifs of ErbB-4 and generated a double-mutated form of ErbB-4 that harbor both mutations of PY1 and PY3, named ErbB-4-PY1,3-HA. Toward this end, 293 cells were transfected with Myc-WWOX and either wild-type ErbB-4-HA or ErbB-4-PY1,3-HA and the cell lysates were immunoprecipitated with anti-HA or anti-Myc antibodies and probed using anti-HA (top) and anti-Myc (bottom) antibodies. As shown in Fig. 2C, mutation of both PPxY motifs in ErbB-4 inhibited interaction with WWOX ( Fig. 2C, lane 7, top and lane 5, bottom). These results imply that WWOX interacts with both of the proline/tyrosine-containing motifs in ErbB-4. ⁴

WWOX and ErbB-4 colocalize in the cytoplasm. We next examined the effect of WWOX-ErbB-4 interaction on the subcellular localization of WWOX and ErbB-4. To address this, we studied the localization of both proteins with the aid of confocal microscopy. ErbB-4 receptor has a cytoplasmic domain that translocates to the nucleus following stimulation with TPA ( 33). To examine the effect of WWOX-ErbB-4 association on this step of signaling in our experimental system, ErbB-4-HA alone or with WWOX-GFP was transiently expressed in HeLa cells. Localization of the HA-tagged proteins was then determined by immunofluorescent staining, as described in Materials and Methods. As shown in Fig. 3A , ErbB-4 alone localizes in the cell membrane and cytoplasm, whereas WWOX is mainly in the cytoplasm ( Fig. 3A, b and a). Coexpression of WWOX-GFP and ErbB-4-HA resulted in partial colocalization of WWOX and ErbB-4 in the cytoplasm ( Fig. 3A, c). Figure 3A (bottom) shows translocation of CTF into the nucleus following treatment with TPA; however, localization of WWOX-GFP was not affected ( Fig. 3A, e and d). By contrast, cells coexpressing WWOX and ErbB-4 and treated with TPA showed a cytoplasmic colocalization of both proteins ( Fig. 3A, f). These results indicate that expression of WWOX causes retention of CTF in the cytoplasm and prevents its translocation to nucleus. To show this effect on endogenous proteins, we used MCF-7 breast carcinoma cells that express high levels of endogenous WWOX and moderate levels of endogenous ErbB-4. We used confocal microscopy and localization of endogenous proteins was then determined by immunofluorescent staining using polyclonal antibody against ErbB-4 (Santa Cruz Biotechnology) and our monoclonal antibody against WWOX ( 10). As expected, MCF-7 cells express both ErbB-4 and WWOX in the cytoplasm ( Fig. 3B, a-b). Treatment of MCF-7 cells with TPA resulted in partial translocation of CTF into the nucleus ( Fig. 3B, c-d). Interestingly, cells displaying nuclear CTF have significantly reduced expression of WWOX ( Fig. 3B, d), which might explain this partial sequestration of endogenous ErbB-4. However, we cannot exclude that there might be other factor(s) regulating ErbB-4 translocation into nucleus. Altogether, our results suggest that WWOX binds ErBb-4 in the cytoplasm and prevents CTF translocation to the nucleus.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Subcellular localization of WWOX and ErbB-4. A, HeLa cells were transfected with the indicated plasmids. Twenty-four hours later, cells (d-f) were treated with TPA (100 ng/mL) for 90 minutes. Cells then were fixed and permeabilized and immunostained with anti-HA antibodies followed by Texas red–conjugated anti-mouse IgG (red, CTF). Colocalization of WWOX and CTF is in yellow as a result of colors merging. B, MCF-7 cells were treated with TPA (100 ng/mL) for 90 minutes (c-d). Cells were prepared as in (A). Antibodies used were polyclonal anti-ErbB-4 (green) and monoclonal anti-WWOX (red). All cells were visualized by confocal microscopy using ×63 objective lens.

WWOX antagonizes coactivation function of YAP. Stimulation of cells expressing ErbB-4 with TPA causes the translocation of CTF to the nucleus where it may affect transcription of target genes. Although CTF by itself has no transactivation activity, Komuro et al. recently showed that YAP can act as a transcriptional coactivator of CTF in the Gal4 transactivation system ( 28). Therefore, we used the same system to determine the effect of WWOX expression on this transactivation. YAP2 and CTF transiently expressed in 293 cells as expected showed ∼35-fold activation compared with Gal4 control ( Fig. 4A, column 8 ). In contrast, coexpression of WWOX and CTF did not show any transactivation ( Fig. 4A, columns 2-6). Interestingly, WWOX significantly suppressed YAP2-CTF-mediated transactivation in a dose-dependent manner ( Fig. 4A, columns 9-11). Furthermore, WWOXY33R mutant lacking interaction with ErbB-4 had no effect on YAP2 transactivation ( Fig. 4A, columns 12 and 13). In addition, a YAP2 mutant harboring an inactivation mutation in the first WW domain and lacking interaction with CTF ( 28) did not show any transactivation of the reporter ( Fig. 4A, columns 7 and 14). These data indicate that transactivation caused by YAP-CTF interaction is suppressed by WWOX, suggesting that WWOX may antagonize the function of YAP.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

A, association of WWOX-ErbB-4 suppresses transactivation of CTF mediated by YAP. 293 cells were cotransfected with the Gal4DNA binding domain (Gal4BD, residues 1-142) fused to the CTF-ΔK fragment (residues 988-1292) and with pFR-luc (0.3 μg), Omni-YAP2 (0.2 μg), Omni-YAP2-1stWW* [mYAP2] (0.2 μg), Myc-WWOX (0.1-0.3 μg), and Myc-WWOXY33R [mWWOX] (0.2-0.3 μg) as indicated. Empty vector was cotransfected to normalize DNA amount where required. A reporter plasmid encoding Renilla luciferase was also cotransfected to normalize transfection efficiencies. Twenty-four hours posttransfection, cells were lysed and luciferase activity was determined. Results are shown as fold activation of the luciferase activity compared with control cells transfected with empty vector alone. Columns, averages of three experiments. B, HeLa cells were transfected with constructs expressing CTF-ΔK alone or together with WWOX-GFP. Cells were immunostained with polyclonal anti-ErbB-4 (red) and visualized as in Fig. 3. C, WWOX suppresses transactivation of p73 mediated through p73-YAP association. 293 cells were cotransfected with vectors expressing p53RE-Luc (0.1 μg), HA-p73β (10 ng), Omni-YAP2 (0.2 μg), Myc-WWOX (0.5 μg), and Nuc-WWOX-Myc (0.5 μg) as indicated. Cells were then treated as in (A). D, HeLa cells were transfected with pCMV-NUC-WWOX-MYC plasmid. Twenty-four hours later, cells were incubated with MitoTracker Red (red, mitochondria) and immunostained with anti-Myc antibodies followed by FITC-conjugated anti-mouse IgG (green, WWOX). Cells were visualized as in Fig. 3.

To confirm interaction between WWOX and CTF in the transactivation assay, we cotransfected HeLa cells with WWOX-GFP and CTF-ΔK. As shown in Fig. 4B, CTF-ΔK when expressed alone showed clear localization in the nucleus ( Fig. 4B, a). Coexpression of WWOX and CTF-ΔK caused partial colocalization of WWOX and CTF-ΔK in the cytoplasm ( Fig. 4B, b) suggesting that WWOX sequesters CTF in the cytoplasm and prevents its nuclear translocation leading to its reduced transactivation ability.

Because it was shown that YAP could serve as a coactivator of p73 ( 25), we decided to check the effect of WWOX expression on this transactivation. We previously showed that WWOX suppresses p73-mediated transactivation ( 2). To assess the antagonizing effect of WWOX on YAP function further, we used the previously described construct containing the luciferase gene driven by a p53 cis-element from the p53 promoter (p53RE-Luc) that can be used in p73β transactivation studies ( 2). As expected, p73β showed significant transactivation of luciferase reporter and expression of YAP2 increased this effect ( Fig. 4C, columns 2 and 6). By contrast, expression of p73β and WWOX significantly suppressed the transactivation function of p73 ( Fig. 4C, column 7). In addition, coexpression of YAP2, p73β, and WWOX resulted in the suppression of p73 transactivation function ( Fig. 4C, column 8) indicating that WWOX indeed antagonizes YAP function.

We previously showed that WWOX sequesters p73 in the cytoplasm explaining WWOX suppressive effect of p73 transactivation ( 2). To further confirm that the effect of WWOX is due to the interaction with p73 and not solely to WWOX-p73 colocalization in the cytoplasm, we constructed a myc-tagged-nuclear WWOX expression vector, named Nuc-WWOX-Myc, expressing WWOX exclusively in the nucleus ( Fig. 4D). Coexpression of Nuc-WWOX and p73β had the same suppressive effect on p73 transactivation as wild-type WWOX ( Fig. 4C, column 9). Furthermore, expression of YAP2 did not affect this suppression ( Fig. 4C, column 10) suggesting that the effect of WWOX expression is superior to that of YAP. Similar results were seen for the CTF-YAP2 and Nuc-WWOX complex (data not shown). This suggests that even when WWOX is in the nucleus together with p73 or CTF, it still inhibits its association with YAP and thus prevents its coactivation.

WWOX and YAP compete for interaction with ErbB-4. Our results may also suggest that WWOX and YAP compete for interaction with ErbB-4 and modulate its transactivation activity. To check whether YAP and WWOX compete for interaction with ErbB-4, we cotransfected 293 cells with fixed amounts of vectors expressing ErbB-4-HA and Myc-WWOX and variable amounts of Omni-YAP2. Cell lysates were then immunoprecipitated with anti-WWOX or anti-Omni antibodies and immunoblotted with anti-HA antibody. Our results show that in the presence of reduced amount of YAP2, ErbB-4 interacts exclusively with WWOX ( Fig. 5A, lane 4 versus 5 ). On the other hand, increased expression of YAP2 causes some ErbB-4-YAP2 association, although ErbB-4-WWOX interaction was still the predominant complex ( Fig. 5A, lane 9 versus 8). To show that WWOX expression indeed interfered with the formation of YAP-ErbB-4 complex, we cotransfected 293 cells with ErbB-4-HA, Omni-YAP2, and either Myc-WWOX or Myc-WWOXY33R. In this context, expression of Myc-WWOXY33R resulted in significant rescue of ErbB-4-YAP2 interaction ( Fig. 5B, lane 9 versus 8). Altogether, our results from the Gal4 transactivation assay ( Fig. 4A) and immunoprecipitation experiments ( Fig. 5) indicate that WWOX and YAP may compete for the association with ErbB-4 and that WWOX-ErbB-4 interaction is prevalent.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

WWOX and YAP compete for interaction with ErbB-4. A, 293 cells were transiently transfected with the expression plasmids encoding ErbB-4-HA (6.0 μg), Myc-WWOX (7.0 μg), and Omni-YAP2 (1.0 or 2.0 μg). Twenty-four hours after transfection, whole cell lysates were immunoprecipitated (IP) with anti-HA (lanes 2 and 6), anti-IgG (lanes 3 and 7), anti-WWOX (lanes 4 and 8), or anti-Omni (lanes 5 and 9) antibodies. The immunoprecipitates were analyzed by immunoblotting (IB) with anti-HA antibody (top). Lanes 1 and 10, expression of ErbB-4; bottom, expression of Myc-WWOX (left) and Omni-YAP2 (right). B, 293 cells were transiently transfected with the expression plasmids encoding ErbB-4−HA (6.0 μg), Omni-YAP2 (1.0 μg), and Myc-WWOX (7.0 μg) or Myc-WWOXY33R (7.0 μg). Cells were then treated as (A).

Discussion

WW domain–containing proteins are widely expressed in all eukaryotes and play a vital role in the regulation of a wide variety of cellular functions including regulation of gene expression ( 39, 40). Here we show that WWOX binds ErbB-4 and suppresses transactivation of CTF mediated by YAP coactivation. We also report that WWOX sequesters ErbB-4 in the cytoplasm and inhibits its nuclear translocation following treatment with TPA. Our results show that the association of WWOX and ErbB-4 in the cytoplasm might explain the suppressed transactivation ability of CTF, although a nuclear-targeted WWOX still has the suppressive effect on CTF function.

YAP is another WW domain–containing protein that functions as a modulator of several transcription factors. YAP has been characterized as a cotransactivator for several transcription factors including the Runx family proteins, the TEAD/TEF family of transcription factors, and p73 ( 24, 25, 41). YAP activity typically involves either of the two WW domains that associate directly with PPxY motifs within the transcription factors, although other mechanisms of interaction may occur. Although YAP can modulate the activity of multiple transcription factors, several reports suggest that it localizes predominantly to the cytoplasm. In the cytoplasm YAP interacts with EBp50 (ezrin/radixin/moesin-binding phosphoprotein of 50 kDa), 14-3-3 and Yes protein tyrosine kinase ( 22, 29, 42). Most recently, it was shown that 14-3-3 interacts with YAP phosphorylated at Ser¹²⁷ and that it prevents YAP from translocation into the nucleus ( 29). In our study, we used the alternatively spliced CYT-2 isoform of ErbB-4 that lacks the PI3K binding motif; therefore, Akt is not activated to phosphorylate YAP at Ser¹²⁷; thus, YAP is retained in the cytoplasm. The subcellular localization of YAP and WWOX in the cytoplasm makes them accessible to bind the PPxY motif of ErbB-4. Our results imply that WWOX binds ErbB-4 in the cytoplasm and makes it inaccessible for interaction with YAP. Therefore, it is possible that WWOX binds CTF and prevents its interaction with YAP; thus, CTF cannot translocate to the nucleus where it may transactivate target genes. A mutated form of WWOX, WWOXY33R, does not bind ErbB-4; therefore, when coexpressed with YAP, it does not prevent YAP from association with ErbB-4 and transcriptional coactivation of the nucleus translocated ErbB4-YAP complex. Interestingly, WWOX binds ErbB-4 with better affinity than YAP, perhaps due to the interaction of WWOX with both the PPxY motifs of ErbB-4 rather than only with the most COOH-terminal PPxY sequence in the case of YAP-ErbB-4 association ( 28).

The functional importance of CTF translocation to the nuclear compartment is not yet clear. However, the fact that CTF was detected in the nucleus in different tissues and expression of nuclear CTF was reported in a number of tumors may indicate an important role of this localization. Both the Notch receptor and Alzheimer's precursor protein, are similarly cleaved by γ-secretase and their intracellular fragments translocate to the nucleus and directly regulate transcription of different genes (reviewed in ref. 43). Moreover, the founding member of the epidermal growth factor receptor family, ErbB-1 has been shown to bind the cyclin D1 promoter and possibly to transactivate genes essential for proliferation ( 44). In addition, ErbB-1 has been proposed to function as a carrier for signal transducers and activators of transcription 3 to the nuclear compartment ( 45). It is thus quite possible that ErbB4-CTF can transactivate target genes in the nucleus or that it functions as a shuttle protein for different targets destined for the nucleus (e.g., YAP; refs. 28, 31). Our findings show that in the presence of WWOX, CTF is retained in the cytoplasm; thus, CTF cannot transactivate target genes or shuttle proteins into the nucleus.

Recently, we showed that WWOX binds p73 transcription factor and suppresses its transcriptional function ( 2). Interestingly, YAP, via its WW domain also associates with p73 and this interaction causes the enhanced transcriptional activity of p73 ( 25). When coexpressed with YAP and p73, WWOX was able to reverse this effect. These results suggest that WWOX acts as a transcriptional corepressor antagonizing the function of YAP. A proposed model is shown in Fig. 6 , where both WW domain–containing proteins, YAP and WWOX, in the cytoplasm are competing for interaction with PPxY-containing target proteins. Whereas WWOX inhibits nuclear translocation of target proteins, YAP translocates with these targets into the nucleus thus determining the transcriptional outcome of these target proteins.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

A proposed model of WWOX and YAP function competing for interaction with ErbB-4 and other target proteins. In this model, ErbB-4-CTF is cleaved following binding of its ligand heregulin or activation of PKC with TPA. CTF by itself or together with YAP translocates to the nucleus and transactivate target genes in the nucleus. In the presence of WWOX, CTF is sequestered in the cytoplasm and transactivation is suppressed. WWOX also interact with other transcription factors such as p73 and AP-2γ and inhibit their translocation to the nucleus hence suppressing their transactivation activity.

ErbB-4 regulates cell proliferation and differentiation ( 32); therefore, the impairment of its proper function may contribute to cancer initiation and/or progression. ErbB-4 belongs to the type I receptor tyrosine kinase family which include ErbB-1 and ErbB-2 that are known proto-oncogenes in breast cancer and other carcinomas ( 46). Indeed, ErbB-4 expression is altered in several tumors including breast, prostate, ovary, and other carcinomas. Srinivasan et al. reported significant percentage of ErbB-4 nuclear staining in invasive breast cancer tumors ( 47). Interestingly, WWOX is inactivated in invasive breast carcinoma ( 12) and YAP was documented to be differentially expressed with malignant transformation and metastatic tumor progression of squamous cell carcinoma ( 48). Moreover, the most recent work from the laboratory of Elenius has documented a clear correlation between a clinical outcome for estrogen receptor–positive breast cancer patients and the presence of cleavable and nucleus-localized ErbB4 isoforms ( 49). This may suggest that in WWOX-negative cells, CTF of ErbB-4 can translocate to the nucleus and transactivate target genes in the presence of YAP, estrogen receptor, or other coactivators. Suppression of transcription factors such as ErbB-4-CTF may have a great effect because these genes have been shown to be amplified and overexpressed in number of cancers.