JNK1 Physically Interacts with WW Domain-containing Oxidoreductase (WOX1) and Inhibits WOX1-mediated Apoptosis
2003; Elsevier BV; Volume: 278; Issue: 11 Linguagem: Inglês
10.1074/jbc.m208373200
ISSN1083-351X
AutoresNan‐Shan Chang, Joan Doherty, Amy Ensign,
Tópico(s)Genetics and Neurodevelopmental Disorders
ResumoTransient activation of c-Jun N-terminal kinase (JNK) promotes cell survival, whereas persistent JNK activation induces apoptosis. Bovine testicular hyaluronidase PH-20 activates JNK1 and protects L929 fibroblasts from staurosporine-mediated cell death. PH-20 also induces the expression of a p53-interacting WW domain-containing oxidoreductase (WOX1, also known as WWOX or FOR) in these cells. WOX1 enhances the cytotoxic function of tumor necrosis factor and mediates apoptosis synergistically with p53. Thus, the activated JNK1 is likely to counteract WOX1 in mediating apoptosis. Here it is demonstrated that ectopic JNK1 inhibited WOX1-mediated apoptosis of L929 fibroblasts, monocytic U937 cells, and other cell types. Also, JNK1 blocked WOX1 prevention of cell cycle progression. By stimulating cells with anisomycin or UV light, JNK1 became activated, and WOX1 was phosphorylated at Tyr33. The activated JNK1 physically interacted with the phosphorylated WOX1, as determined by co-immunoprecipitation. Alteration of Tyr33 to Arg33 in WOX1 abrogated its binding interaction with JNK1 and its activity in mediating cell death, indicating that Tyr33 phosphorylation is needed to activate WOX1. A dominant negative WOX1 was developed and shown to block p53-mediated apoptosis and anisomycin-mediated WOX1 phosphorylation but could not inhibit JNK1 activation. This mutant protein bound p53 but could not interact with JNK1, as determined in yeast two-hybrid analysis. Taken together, phosphorylation of JNK1 and WOX1 is necessary for their physical interaction and functional antagonism. Transient activation of c-Jun N-terminal kinase (JNK) promotes cell survival, whereas persistent JNK activation induces apoptosis. Bovine testicular hyaluronidase PH-20 activates JNK1 and protects L929 fibroblasts from staurosporine-mediated cell death. PH-20 also induces the expression of a p53-interacting WW domain-containing oxidoreductase (WOX1, also known as WWOX or FOR) in these cells. WOX1 enhances the cytotoxic function of tumor necrosis factor and mediates apoptosis synergistically with p53. Thus, the activated JNK1 is likely to counteract WOX1 in mediating apoptosis. Here it is demonstrated that ectopic JNK1 inhibited WOX1-mediated apoptosis of L929 fibroblasts, monocytic U937 cells, and other cell types. Also, JNK1 blocked WOX1 prevention of cell cycle progression. By stimulating cells with anisomycin or UV light, JNK1 became activated, and WOX1 was phosphorylated at Tyr33. The activated JNK1 physically interacted with the phosphorylated WOX1, as determined by co-immunoprecipitation. Alteration of Tyr33 to Arg33 in WOX1 abrogated its binding interaction with JNK1 and its activity in mediating cell death, indicating that Tyr33 phosphorylation is needed to activate WOX1. A dominant negative WOX1 was developed and shown to block p53-mediated apoptosis and anisomycin-mediated WOX1 phosphorylation but could not inhibit JNK1 activation. This mutant protein bound p53 but could not interact with JNK1, as determined in yeast two-hybrid analysis. Taken together, phosphorylation of JNK1 and WOX1 is necessary for their physical interaction and functional antagonism. WW domain-containing oxidoreductase (WOX1, 1The abbreviations used are: WOX1, WW domain-containing oxidoreductase; JNK, c-Jun N-terminal kinase; TNF, tumor necrosis factor; NLS, nuclear localization sequence; ADH, short-chain alcohol dehydrogenase; GFP, green fluorescent protein 1The abbreviations used are: WOX1, WW domain-containing oxidoreductase; JNK, c-Jun N-terminal kinase; TNF, tumor necrosis factor; NLS, nuclear localization sequence; ADH, short-chain alcohol dehydrogenase; GFP, green fluorescent protein also known as WWOX or FOR) is a proapoptotic protein and is known to enhance the cytotoxic function of tumor necrosis factor (TNF) as well as overexpressed TNF receptor-associated death domain protein (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar, 2Chang N.-S. Int. J. Mol. Med. 2002; 9: 19-24Google Scholar, 3Richards R.I. Trends Genet. 2001; 17: 339-345Google Scholar). Enhancement of TNF cytotoxicity by WOX1 is due, in part, to its significant down-regulation of the apoptosis inhibitors Bcl-2 and Bcl-xL but up-regulation of proapoptotic p53 (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar, 2Chang N.-S. Int. J. Mol. Med. 2002; 9: 19-24Google Scholar). WOX1 physically interacts with p53, and these proteins mediate apoptosis synergistically (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar, 2Chang N.-S. Int. J. Mol. Med. 2002; 9: 19-24Google Scholar). Blocking of WOX1 expression by antisense mRNA abolishes p53 apoptosis, suggesting that WOX1 is a potential partner of p53 in apoptosis (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar). Interestingly, hyaluronidases PH-20, Hyal-1, and Hyal-2 induce WOX1 expression and enhance TNF cytotoxicity (4Chang N.-S. Am. J. Physiol. 1997; 273: C1987-C1994Google Scholar, 5Chang N.-S. BMC Cell Biol. 2002; 3: 8Google Scholar). The WWOX gene coding for human WOX1 is mapped to a fragile site on the chromosome ch16q (for reviews, see Refs. 2Chang N.-S. Int. J. Mol. Med. 2002; 9: 19-24Google Scholar and 3Richards R.I. Trends Genet. 2001; 17: 339-345Google Scholar). High frequency of loss of heterozygosity of this chromosomal region has been shown in breast (6Driouch K. Dorion-Bonnet F. Briffod M. Champeme M.H. Longy M. Lidereau R. Genes Chromosomes Cancer. 1997; 19: 185-191Google Scholar) and prostate cancers (7Suzuki H. Komiya A. Emi M. Kuramochi H. Shiraishi T. Yatani R. Shimazaki J. Genes Chromosomes Cancer. 1996; 17: 225-233Google Scholar, 8Li C. Berx G. Larsson C. Auer G. Aspenblad U. Pan Y. Sundelin B. Ekman P. Nordenskjold M. van Roy F. Bergerheim U.S. Genes Chromosomes Cancer. 1999; 24: 175-182Google Scholar). This chromosome region appears to confer suppression of the metastatic activity in a prostatic cancer cell line (9Mashimo T. Watabe M. Cuthbert A.P. Newbold R.F. Rinker-Schaeffer C.W. Helfer E. Watabe K. Cancer Res. 1998; 58: 4572-4576Google Scholar). Both in vitro and in vivoexperiments showed that WOX1 could act as a tumor suppressor (10Bednarek A.K. Keck-Waggoner C.L. Daniel R.L. Laflin K.J. Bergsagel P.L. Kiguchi K. Brenner A.J. Aldaz C.M. Cancer Res. 2001; 61: 8068-8073Google Scholar). WOX1 (46 kDa) possesses two N-terminal WW domains (containing conserved tryptophan residues), a nuclear localization sequence (NLS) and a C-terminal short-chain alcohol dehydrogenase (ADH) domain (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar). The human WWOX gene has nine exons. At least seven alternatively spliced human WOX protein variants have been found. The ADH domain region, which is encoded by exons 4–8, is frequently deleted or alternatively spliced in the variants. In contrast, the WW domain region is rarely deleted, and its amino acid sequence is highly conserved among humans, mice, and rats. However, inDrosophila, the sequence of the N-terminal WW domain region is altered (11Paige A.J. Taylor K.J. Taylor C. Hillier S.G. Farrington S. Scott D. Porteous D.J. Smyth J.F. Gabra H. Watson J.E. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 11417-11422Google Scholar). A portion of cytosolic WOX1 is located in the mitochondria, and the mitochondrial targeting sequence was mapped within the ADH domain (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar). Induction of mitochondrial permeability transition by TNF, staurosporine, and atractyloside results in WOX1 release from the mitochondria and subsequent nuclear translocation (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar, 2Chang N.-S. Int. J. Mol. Med. 2002; 9: 19-24Google Scholar). c-Jun N-terminal kinase (JNK) is involved in cellular stress response and has been implicated in both cell growth and apoptosis (for a review, see Ref. 12Davis R.J. Cell. 2000; 103: 239-252Google Scholar). JNK1 is involved in signaling pathways that initiate cell cycle checkpoints and cell cycle progression (13MacCorkle-Chosnek R.A. VanHooser A. Goodrich D.W. Brinkley B.R. Tan T.H. Biochem. Biophys. Res. Commun. 2001; 289: 173-180Google Scholar). In contrast, inhibition of the JNK signaling pathway prevents death of trophic factor-deprived neurons (14Harris C.A. Deshmukh M. Tsui-Pierchala B. Maroney A.C. Johnson Jr., E.M. J. Neurosci. 2002; 22: 103-113Google Scholar). Persistent activation of JNK1 causes apoptosis (15Kobayashi K. Tsukamoto I. Biochim. Biophys. Acta. 2001; 1537: 79-88Google Scholar). However, early activation of JNK1 is essential for protecting cells against TNF-mediated apoptosis (16Roulston A. Reinhard C. Amiri P. Williams L.T. J. Biol. Chem. 1998; 273: 10232-10239Google Scholar, 17Chang N.-S. Biochem. Biophys. Res. Commun. 1999; 263: 107-112Google Scholar). Transient activation of JNK by growth factors promotes survival of human thyrocytes (18Shklyaev S.S. Namba H. Mitsutake N. Alipov G. Nagayama Y. Maeda S. Ohtsuru A. Tsubouchi H. Yamashita S. Thyroid. 2001; 11: 629-636Google Scholar). Notably, hyaluronidase activation of JNK1 and JNK2 protects L929 fibroblasts from staurosporine-mediated cell death (19Chang N.-S. Biochem. Biophys. Res. Commun. 2001; 283: 278-286Google Scholar). In this study, we demonstrated that anisomycin and UV light activated JNK1 and induced WOX1 phosphorylation in Tyr33, and both proteins physically interacted. Ectopic JNK1 inhibited WOX1-mediated apoptosis of L929 fibroblasts, monocytic U937 cells, and other tested cells. Also, WOX1 prevention of cell cycle progression was abolished by JNK1. The mechanisms of JNK1 and WOX1 binding interaction and the functional significance of this interaction were investigated. Cell lines used in these studies were murine L929 fibroblasts, human monocytic U937 cells, human prostate Du145 cells, and human neuroblastoma SK-N-SH cells. These cells were grown according to the instructions of the ATCC (Manassas, VA). Expression constructs for the murine wild type Wox1 cDNA (sense and antisense directions) and a nuclear localization mutant Wox1mtNLS were established as previously described (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar). These constructs were made in pEGFP-C1 vector (Clontech, Palo Alto, CA). The expressed proteins were tagged with an N-terminal enhanced green fluorescent protein. The FLAG-tagged wild type JNK1 (pcDNA3-Flag-JNK1) and dominant negative JNK1 (DN-JNK1 or pcDNA3-Flag-JNK1apf) expression constructs were kind gifts of Dr. Roger Davis (University of Massachusetts Medical Center, Worcester, MA). The HA-tagged wild type JNK2 and JNK3 (peVRFO-HA-JNK2 and peVRFO-HA-SAPKβ) and dominant negative JNK2 and JNK3 (peVRFO-HA-JNK2-K55R and peVRFO-HA-SAPKβ-K55R or DN-JNK2 and DN-JNK3) were kind gifts of Dr. Michael Kracht (Medical School Hannover, Hannover, Germany). Where indicated, L929 cells were electroporated with the wild type or DN-JNK constructs (20 μg of DNA/3 × 106 cells; 200 V and 50 ms; BTX ECM 830 Square Wave Electroporator; Genetronics, San Diego, CA). Electroporation caused immediate cell death (or necrosis) by ∼20–40%. These electroporated cells were seeded and cultured on 60-mm Petri dishes (Corning Glass) overnight. The culture supernatants were discarded to remove dead cells. The cells were fed with fresh medium (RPMI 1640 plus 10% fetal bovine serum) and then treated with or without bovine testicular hyaluronidase (100 units/ml; Sigma) for 24 h. The cell layers, reaching 80–90% confluence in 48 h, were gently washed once with phosphate-buffered saline, treated with 70% ethanol, harvested by repeat pipetting, precipitated by centrifugation (500 × g), and then stained with propidium iodide. The cellular DNA content was analyzed by flow cytometry (Becton Dickinson, Franklin Lakes, NJ). In some experiments, L929 and other cells were electroporated with WOX1 and/or JNK1 constructs and cultured for 24 h (in eight replicates). The extent of cell death was examined by crystal violet staining (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar) or by Promega's MTS proliferation assay (Promega, Madison, WI). Antibody against the N-terminal amino acid sequence of WOX1 was generated as previously described (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar). Commercial antibodies used in the Western blotting were against JNK1 and a phospho-JNK1 peptide (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Co-immunoprecipitation was performed as previously described (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar,20Chang N.-S. J. Biol. Chem. 2002; 277: 10323-10331Google Scholar). Briefly, U937, L929, or SK-N-SH cells were treated with anisomycin (100 μm) for various indicated times to activate JNK1. The cytosolic and nuclear fractions of these cells were prepared using the NE-PER nuclear and cytoplasmic extraction reagents (Pierce). Endogenous JNK1 was precipitated by the specific anti-JNK1 antibody (Santa Cruz Biotechnology) and protein A-agarose beads (Pierce), followed by separation in SDS-PAGE and detection by immunostaining. Co-presence of WOX1 in the precipitates was examined similarly by immunostaining. Alternatively, immunoprecipitation was performed using anti-WOX1 or p53 (Santa Cruz Biotechnology) antibody. Generation of antibody against synthetic peptides in rabbits was performed as described (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar). A synthetic peptide NH2-CKDGWVpYYANHTEEKT-COOH, with tyrosine 33 phosphorylation (pY), was made (Genemed Synthesis, South San Francisco, CA). A control peptide without phosphorylation was made also. The N-terminal cysteine was added to allow the peptide to covalently conjugate to keyhole limpet hemocyanin for immunization or to SulfoLink beads for antibody purification (Pierce). The peptide sequence is located at the first WW domain of murine WOX1 (amino acids 28–38). Antibody generation was performed using the EZ antibody production and purification kit (Pierce). The antiserum was preadsorbed with a column covalently conjugated with the control peptide, and then the antibody was purified by affinity chromatography using a column immobilized with the synthetic phosphopeptide. The Ras rescue-based CytoTrap yeast two-hybrid system (Stratagene, La Jolla, CA) for protein binding interaction was used (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar, 20Chang N.-S. J. Biol. Chem. 2002; 277: 10323-10331Google Scholar). Briefly, binding of a cytosolic Sos-tagged bait protein to a cell membrane-anchored target protein (tagged with a myristoylation signal) results in activation of the Ras signaling pathway in yeast. Activation of the Ras signaling pathway allows mutant yeast Cdc25H to grow in 37 °C using a selective agarose plate containing galactose. Without binding, yeast cells fail to grow at 37 °C. The human JNK1 cDNA was constructed in a pMyr vector and utilized as target. The bait constructs in a pSos vector for binding interactions with JNK1 were murine Wox1 (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar), human WOX3 (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar), and the first WW domain of murine Wox1 (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar). Additional constructs for control experiments were human p53 in pMyr, MafB in both pMyr and pSos, collagenase in pSos, and lamin C in pMyr (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar, 20Chang N.-S. J. Biol. Chem. 2002; 277: 10323-10331Google Scholar). Site-directed mutagenesis was performed using the QuikChange site-directed mutagenesis kit (Stratagene). Tyr33 in WOX3 was mutated to Arg33 (designated Y33R) using the following primers by PCR: forward, 5′-AAGGACGGCTGGGTTCGATACGCCAATCACACC; reverse, 5′-GGTGTGATTGGCGTATCGAACCCAGCCGTCCTT. Lys28 and Asp29 in Wox1 were mutated to Thr28and Val29 (designated K28T/D29V): forward, 5′-AGAGAACCACCACGGTCGGCTGGGTGTACT; reverse, 5′-AGTACACCCAGCCGACCGTGGTGGTTCTCT. Where indicated, Student's t tests were performed to analyze the experimental data (using the Microsoft Excel program). All experiments indicated above were performed 2–5 times. In most cases, a representative set of data is shown. Bovine testicular hyaluronidase PH-20 activates JNK1 in L929 fibroblasts (19Chang N.-S. Biochem. Biophys. Res. Commun. 2001; 283: 278-286Google Scholar). PH-20 promotes the proliferation of L929 cells (4Chang N.-S. Am. J. Physiol. 1997; 273: C1987-C1994Google Scholar). By measuring cellular DNA content using flow cytometry, PH-20 was shown to increase cell cycle progression toward the S and G2/M phases during 24-h exposure (Fig.1). No PH-20-mediated apoptosis or cell death was observed (Sub G0/G1) (Fig. 1). To examine the possible role of JNK1 in the PH-20-mediated growth regulation, L929 cells were electroporated with a wild type human JNK1 or a dominant negative JNK1 construct. The cells were cultured overnight, exposed to PH-20 for 24 h, and then subjected to cell cycle analysis by flow cytometry. In controls, the cells were electroporated with nothing (or medium only) and treated similarly. The results showed that ectopic wild type JNK1 did not induce apoptosis or death of L929 cells (Fig. 2). Exposure of these JNK1-expressing cells to PH-20 could not induce apoptosis (Fig.2). In contrast, dominant-negative JNK1 significantly induced L929 cell death (56.4%), and PH-20 did not significantly increase the extent of cell death (60.9%) (Fig. 2). In comparison, ectopic human wild-type JNK2 and JNK3 mediated apoptosis or death of L929 cells by ∼30%, and PH-20 limitedly increased the cell death (∼40%) (Fig. 2). In contrast, dominant negative JNK2 and JNK3 significantly mediated cell death up to 60–70%, and PH-20 slightly increased the cell death (Fig. 2). PH-20 induces the expression of proapoptotic p53 and WOX1 in L929 cells, thereby increasing cellular sensitivity to the cytotoxic effect of TNF (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar). Nonetheless, PH-20 also activates JNK1 in these cells (19Chang N.-S. Biochem. Biophys. Res. Commun. 2001; 283: 278-286Google Scholar). Here, we examined whether JNK1 inhibited WOX1-mediated apoptosis of L929 cells. In agreement with our previous observations (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar), electroporation of L929 cells with a wild-type WOX1 expression construct (tagged with enhanced green fluorescent protein) resulted in cell death (Fig.3). The extent of cell death was determined by the MTS assay. Similar results were observed by staining the cells with crystal violet (data not shown). The extent of WOX1 protein expression was ∼60–75% in L929 cells, as determined by counting cells under fluorescence microscopy. When L929 cells were cotransfected with both wild type JNK1 and WOX1 constructs, and JNK1 blocked WOX1-mediated cell death (Fig. 3). In contrast, WOX1mtNLS, which possesses a mutated sequence at the NLS, had a significantly reduced activity in mediating cell death (Fig. 3). When in combination, JNK1 and WOX1mtNLS could not induce cell death (Fig.3). The mutation in WOX1mtNLS prevents nuclear translocation of this protein (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar). In agreement with our previous observations (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar), expression of antisense WOX1 mRNA in L929 cells failed to result in cell death, and JNK1 had no effect in mediating cell death during cotransfection (Fig. 3). Together, WOX1 mediates apoptosis in the nucleus, and JNK1 blocks the WOX1-mediated cell death at the nuclear level. To further verify the above observations, we examined the ability of wild type JNK1 in blocking apoptosis of monocytic U937 cells by WOX1. U937 cells were electroporated with various amounts of WOX1 in the presence or absence of wild type JNK1. The cells were cultured 48 h and subjected to cell cycle analysis by flow cytometry. The results showed that WOX1-mediated apoptosis was suppressed by JNK1 (see the sub-G0/G1 phase) (Fig. 4). Additionally, WOX1-mediated reduction of cellular distribution in the G0/G1, S, and G2/M phases was reversed by JNK1 (Fig. 4). Thus, our data strongly support the protective role of JNK1 in blocking WOX1-mediated cell death. We investigated whether JNK1 physically interacts with WOX1, thereby blocking WOX1-mediated apoptosis. As determined by yeast two-hybrid analysis (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar, 20Chang N.-S. J. Biol. Chem. 2002; 277: 10323-10331Google Scholar), the wild type human JNK1 interacted with murine WOX1 and human WOX3in vivo, as evidenced by the growth of mutant Cdc25H yeast at 37 °C (Fig. 5). Interestingly, the WOX3/JNK1 binding interaction promoted a strong yeast cell growth, compared with that of the WOX1/JNK1 interaction (Fig. 5). Human WOX3 has a deletion of its ADH domain at the C-terminus, whereas itsN-terminal WW domain region is highly homologous to that of murine WOX1 (Fig. 5) (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar). The first WW domain also interacted with JNK1 (Fig. 5). These results suggest that the first WW domain is responsible for binding to JNK1. A conserved phosphorylation site at the first WW domain, Tyr33, was then altered to arginine in WOX3. This mutant, designated WOX3(Y33R), could not interact with JNK1 (Fig. 5). The observations further suggest that phosphorylation at Tyr33 is essential for WOX1 and JNK1 binding interaction. In positive control experiments, the binding interactions between p53 and WOX1 (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar) and the self-binding interaction of MafB were demonstrated (Fig. 5). In negative controls, no binding interactions were observed when testing empty pSos vector versus empty pMyr vector, or collagenase versus lamin C (Fig. 5). To further verify the above observations, U937 cells were treated with anisomycin to activate JNK1. Anisomycin is a potent activator of JNK1 (17Chang N.-S. Biochem. Biophys. Res. Commun. 1999; 263: 107-112Google Scholar). In resting cells, no JNK1 and WOX1 binding interaction was observed, as determined by co-immunoprecipitation using specific anti-JNK1 antibody (Fig. 6 A). Upon stimulation of cells with anisomycin for 30 min, expression of JNK1 was increased, and a portion of JNK1 interacted with WOX1 (Fig. 6 A). Similarly, exposure of SK-N-SH cells to anisomycin for 30 min also resulted in an increased binding interaction between JNK1 and WOX1, as determined using anti-WOX1 antibody in co-immunoprecipitation (Fig.6 B). Similar results were observed when L929 cells were treated with anisomycin, or the above cells were exposed to UV light (data not shown). To further verify Tyr33 phosphorylation in WOX1, we synthesized a peptide containing phospho-Tyr33 for immunization in rabbits. Anisomycin rapidly increased Tyr33phosphorylation in WOX1, as well as JNK1 phosphorylation, in SK-N-SH cells (Fig. 7 A). Similar results were observed using L929 and U937 cells (data not shown). Exposure of L929 cells to UV light, followed by growing for 1 h, resulted in increased p53 expression (Fig. 7 B). p53 is known to interact with WOX1 (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar) and JNK1 (21Fuchs S.Y. Adler V. Pincus M.R. Ronai Z. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 10541-10546Google Scholar). Co-immunoprecipitation of p53 with a specific antibody resulted in the presence of p53, JNK1, and phosphorylated WOX1 in the precipitates (Fig. 7 B). We developed a dominant-negative WOX1. Lys28 and Asp29 inWox1 were mutated to Thr28 and Val29, respectively. These residues are located at the first WW domain. In agreement with our previous observations (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar), p53 and WOX1 mediated U937 cell death in a synergistic manner, and antisense WOX1 abolished p53-mediated cell death (Fig.8 A). The mutated full-length WOX1 (WOX1mt) abolished p53 apoptosis (Fig. 8 A). Similarly, the mutated WW domain region (WOX1wwmt) also abolished p53 apoptosis (Fig. 8 A). These data support the dominant negative effect of the generated WOX1 mutant. Anisomycin-mediated WOX1 phosphorylation in Tyr33 was abolished by the dominant negative WOX1 (Fig. 8 B). SK-N-SH cells were transfected with a GFP-tagged WOX1wwmt or a control GFP construct by electroporation. The cells were cultured for 48 h and then treated with anisomycin for 30 min. The dominant negative WOX1wwmt abolished anisomycin-mediated WOX1 phosphorylation in Tyr33(Fig. 8 B). However, WOX1wwmt failed to inhibit anisomycin-induced JNK1 activation (Fig. 8 B). By yeast two-hybrid analysis, both WOX1mt and WOX1wwmt physically interacted with p53, whereas WOX1wwmt could not bind JNK1 (Fig.8 C). Apparently, these dominant negative WOX1 proteins have an altered conformation, and inhibition of p53 apoptosis by the dominant negatives is due to a direct binding interaction. In contrast, these dominant negatives failed to bind and prevent JNK1 activation. Overexpression of the WW domain region of WOX1, designated WOX1ww, induces cell death (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar). Alteration of Tyr33 to Arg33 was performed in WOX1ww, and the resulting mutant, WOX1ww(Y33R), had a significantly reduced activity (∼50% decrease) in mediating death of L929 and Du145 cells (Fig.9). The findings were also observed using U937, SK-N-SH, and other types of cells (data not shown). JNK1 significantly blocked WOX1ww-mediated death of L929 and Du145 cells by 30–50% (Fig. 9). Interestingly, JNK1 increased the apoptotic function of WOX1ww(Y33R) mutant in L929 but not in Du145 cells (Fig. 9). In this study, we demonstrated that stress stimuli such as anisomycin and UV light mediated WOX1 phosphorylation at Tyr4. The phosphorylation is essential for functional activation of WOX1 in mediating apoptosis. Alteration of Tyr33 to Arg33 reduced the WOX1 activity. In agreement with our previous observations (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar), WOX1-mediated apoptosis occurs at the nuclear level. Mutation of the nuclear localization signal reduced the WOX1 apoptotic function. We determined that ectopic JNK1 counteracted WOX1-mediated apoptosis, as well as cell cycle progression. By stimulating cells with anisomycin, JNK1 became activated and physically interacted with the Tyr33-phosphorylated WOX1, as determined by co-immunoprecipitation. A dominant negative WOX1 blocked p53-mediated apoptosis and anisomycin-mediated WOX1 phosphorylation but failed to inhibit JNK1 activation. This dominant negative WOX1 bound p53 but could not interact with JNK1, as determined in yeast two-hybrid analysis. Together, these observations indicate that a physical contact between activated or phosphorylated WOX1 and JNK1 is essential for their functional antagonism. How JNK1 inhibits WOX1 function remains to be determined. The upstream kinase that phosphorylates WOX1 is unknown. Nonetheless, we do not exclude the possibility that activated JNK1 binds WOX1 and then phosphorylates WOX1. Alternatively, an unknown kinase phosphorylates WOX1 in response to stress stimuli. The phosphorylated WOX1 then interacts with JNK1. The dominant negative WOX1 bound p53 and blocked p53 apoptosis. Also, this mutant protein inhibited WOX1 phosphorylation, suggesting its possible interaction with the wild type WOX1. This potential binding interaction remains to be determined by co-immunoprecipitation and yeast two-hybrid analysis. An additional conserved phosphorylation site in WOX1 is Tyr61. Alteration of this residue to Arg61results in abrogation of the binding interaction between WOX1 and p53 in yeast. 2N.-S. Chang, J. Doherty, and A. Ensign, unpublished results. Thus, it is likely that both Tyr33 and Tyr61 in WOX1 are phosphorylated when cells are exposed to stress stimuli. Functional significance of Tyr61 phosphorylation in WOX1 is being determined in this laboratory. The functional property of WOX3 in vivo and in vitro is unknown. Whether WOX3 expression is increased during stress response is unknown. By RT-PCR and Northern blot, Driouch demonstrated that about 50% of breast tumors have overexpressed WOX1 and WOX3 (22Driouch K. Prydz H. Monese R. Johansen H. Lidereau R. Frengen E. Oncogene. 2002; 21: 1832-1840Google Scholar). WOX1 appears to have a tumor suppressor function (10Bednarek A.K. Keck-Waggoner C.L. Daniel R.L. Laflin K.J. Bergsagel P.L. Kiguchi K. Brenner A.J. Aldaz C.M. Cancer Res. 2001; 61: 8068-8073Google Scholar,11Paige A.J. Taylor K.J. Taylor C. Hillier S.G. Farrington S. Scott D. Porteous D.J. Smyth J.F. Gabra H. Watson J.E. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 11417-11422Google Scholar). Presumably, the overexpressed WOX3 enhances cancer growth by neutralizing the WOX1 function. In addition to activate JNK1, anisomycin inhibits protein synthesis and mediates apoptosis (23Kageyama A. Kusano I. Tamura T. Oda T. Muramatsu T. Biosci. Biotechnol. Biochem. 2002; 66: 835-839Google Scholar, 24Stadheim T.A. Kucera G.L. Leuk. Res. 2002; 26: 55-65Google Scholar). Anisomycin-mediated apoptosis of leukemia HL-60 cells depends upon JNK1 (24Stadheim T.A. Kucera G.L. Leuk. Res. 2002; 26: 55-65Google Scholar). Most interestingly, ansiomycin-activated JNK1 phosphorylates Bcl-2, thereby increasing cell survival (25Deng X. Xiao L. Lang W. Gao F. Ruvolo P. May Jr., W.S. J. Biol. Chem. 2001; 276: 23681-23688Google Scholar). Previously, we found that when L929 cells were exposed to anisomycin for less than 1 h, these cells became resistant to TNF cytotoxicity (17Chang N.-S. Biochem. Biophys. Res. Commun. 1999; 263: 107-112Google Scholar). We determined that IκBα, the inhibitor of NF-κB, maintains the basal level of JNK1 activation and regulates JNK1-mediated TNF resistance (17Chang N.-S. Biochem. Biophys. Res. Commun. 1999; 263: 107-112Google Scholar). However, whether this protective event is related to JNK1-mediated Bcl-2 phosphorylation remains to be determined. Hyaluronidase PH-20 activates JNK1, and the activation is necessary to prevent L929 cell death by various anticancer drugs such as staurosporine, daunorubicin, and doxorubicin (19Chang N.-S. Biochem. Biophys. Res. Commun. 2001; 283: 278-286Google Scholar). In contrast, PH-20 increases L929 cell sensitivity to TNF cytotoxicity by increasing the expression of WOX1 (1Chang N.-S. Pratt N. Heath J. Schultz L. Sleve D. Carey G.B. Zevotek N. J. Biol. Chem. 2001; 276: 3361-3370Google Scholar). Apparently, different mechanisms are involved in the regulation of cellular sensitivity to TNF and anticancer drugs. Indeed, we determined that ectopic JNK1 and exogenous PH-20 act synergistically in blocking L929 cell death by WOX1 in the presence of staurosporine, daunorubicin, or doxorubicin.2 We believe that PH-20 activates both endogenous and ectopic JNK1 that block the cell death. Whether JNK1 mediates Bcl-2 phosphorylation in these tested cells is unknown. The hyaluronidase HYAL1 gene, also known asLUCA1, is a candidate tumor suppressor (for a review, see Ref. 26Csoka A.B. Frost G.I. Stern R. Matrix Biol. 2001; 20: 499-508Google Scholar). Inactivation of the HYAL1 gene by aberrant splicing of pre-mRNA has been shown in head and neck squamous cell carcinomas (27Frost G.I. Mohapatra G. Wong T.M. Csoka A.B. Gray J.W. Stern R. Oncogene. 2000; 19: 870-877Google Scholar). Nonetheless, overexpressed hyaluronidases are frequently associated with metastatic cancers (28Lokeshwar V.B. Rubinowicz D. Schroeder G.L. Forgacs E. Minna J.D. Block N.L. Nadji M. Lokeshwar B.L. J. Biol. Chem. 2001; 276: 11922-11932Google Scholar, 29Patel S. Turner P.R. Stubberfield C. Barry E. Rohlff C.R. Stamps A. Tyson K. Terrett J. Box G. Eccles S. Page M.J. Int. J. Cancer. 2002; 97: 416-424Google Scholar, 30Beech D.J. Madan A.K. Deng N. J. Surg. Res. 2002; 103: 203-207Google Scholar). Also, plasma HYAL1 is capable of promoting tumor cell cycling (31Lin G. Stern R. Cancer Lett. 2001; 163: 95-101Google Scholar). PH-20 enhances cell proliferation (4Chang N.-S. Am. J. Physiol. 1997; 273: C1987-C1994Google Scholar). Here, we demonstrated that PH-20 increased cell cycle progression in L929 cells. PH-20-activated JNK1 is essential for cell survival, since dominant negative JNK1 mediated apoptosis in L929 cells. UV light up-regulates and activates p53 (32Lakin N.D. Jackson S.P. Oncogene. 1999; 18: 7644-7655Google Scholar) and also activates JNK1 (12Davis R.J. Cell. 2000; 103: 239-252Google Scholar). Co-immunoprecipitation of the anisomycin-treated cells with anti-p53 resulted in the presence of p53, JNK1, and WOX1 in the precipitate. The observation suggests the presence of a p53·WOX1·JNK1 complex during stress response. Most recently, we determined that a portion of cytosolic IκBα binds p53 in resting cells and the complex dissociates in response to apoptotic stress, hypoxia, DNA damage, and tumor growth factor-β1-mediated growth suppression (20Chang N.-S. J. Biol. Chem. 2002; 277: 10323-10331Google Scholar). The dissociation allows p53 nuclear translocation. Alternatively, when dissociated from IκBα, p53 may complex with WOX1 and JNK1. Although transient JNK1 activation provides a signal for cell survival (13MacCorkle-Chosnek R.A. VanHooser A. Goodrich D.W. Brinkley B.R. Tan T.H. Biochem. Biophys. Res. Commun. 2001; 289: 173-180Google Scholar, 16Roulston A. Reinhard C. Amiri P. Williams L.T. J. Biol. Chem. 1998; 273: 10232-10239Google Scholar, 17Chang N.-S. Biochem. Biophys. Res. Commun. 1999; 263: 107-112Google Scholar, 18Shklyaev S.S. Namba H. Mitsutake N. Alipov G. Nagayama Y. Maeda S. Ohtsuru A. Tsubouchi H. Yamashita S. Thyroid. 2001; 11: 629-636Google Scholar, 19Chang N.-S. Biochem. Biophys. Res. Commun. 2001; 283: 278-286Google Scholar), a persistent JNK1 activation can mediate cell death (14Harris C.A. Deshmukh M. Tsui-Pierchala B. Maroney A.C. Johnson Jr., E.M. J. Neurosci. 2002; 22: 103-113Google Scholar,15Kobayashi K. Tsukamoto I. Biochim. Biophys. Acta. 2001; 1537: 79-88Google Scholar). How the antiapoptotic JNK1 turns into apoptotic is not known. It may be related to the status of JNK1 phosphorylation, or it may be associated with JNK1-regulated gene expression at different stages of cell differentiation or stress. We believe that at a certain stage of apoptosis, the apoptotic JNK1 may act synergistically with WOX1 in facilitating cell death. In transient expression, both wild type JNK2 and JNK3 mediated L929 cell death, yet the dominant negatives greatly increased the cell death. The likely explanation is that the kinetics of activation of JNK isoforms is different, and their functional roles may be different also. For example, JNK1 and JNK2 have distinct roles in antiviral immunity, in which JNK1 is involved in survival of activated T cells during immune responses, and JNK2 plays a role in control of CD8(+) T cell expansion in vivo (33Arbour N. Naniche D. Homann D. Davis R.J. Flavell R.A. Oldstone M.B. J. Exp. Med. 2002; 195: 801-810Google Scholar). Similarly, JNK1 and JNK2 play distinct roles in controlling the activation of CD8(+) T cells (34Conze D. Krahl T. Kennedy N. Weiss L. Lumsden J. Hess P. Flavell R.A. Le Gros G. Davis R.J. Rincon M. J. Exp. Med. 2002; 195: 811-823Google Scholar). Also, JNK1-deficient cells resist UV-mediated apoptosis, whereas JNK2-deficient cells have increased sensitivity to UV irradiation (35Hochedlinger K. Wagner E.F. Sabapathy K. Oncogene. 2002; 21: 2441-2445Google Scholar). JNKs are widely distributed in the nervous system (12Davis R.J. Cell. 2000; 103: 239-252Google Scholar), and JNK3 is involved not only in neuronal development but also apoptosis of neuronal cells (36Waetzig V. Herdegen T. J. Biol. Chem. 2003; 278: 567-572Google Scholar). L929 cells express both JNK1 and JNK2 (17Chang N.-S. Biochem. Biophys. Res. Commun. 1999; 263: 107-112Google Scholar, 19Chang N.-S. Biochem. Biophys. Res. Commun. 2001; 283: 278-286Google Scholar). Whether JNK3 is expressed in L929 cells is unknown. Dominant negative JNK1 mediated L929 cell apoptosis, suggesting that JNK1 is essential for the survival of these cells. Inhibition of JNK2 and JNK3 functions by dominant negatives induced cell death, suggesting that L929 cells also require a certain level of JNK2 and JNK3 for survival. In contrast to the wild type JNK1, transiently expressed wild type JNK2 and JNK3 induced L929 cell death, indicating that the kinetics of activation of JNK isoforms is different. Taken together, phosphorylation of Tyr33 and probably Tyr61 is important for activating WOX1. Our data support the possibility that WOX1 is a signaling protein in the p53 and JNK1 stress pathways. Whether activated JNK1 phosphorylates WOX1 remains to be determined. We thank Terri Zimmer for antibody production in rabbits, John Heath and Lori Schultz for technical assistance, and Drs. Roger Davis and Michael Kracht for JNK constructs.
Referência(s)