Phosphorylation of Tyrosine 568 Controls Nuclear Export of Nrf2*

Nuclear factor Nrf 2, under normal conditions, is retained in the cytosol by INrf 2. Antioxidants and oxidants antagonize this interaction, resulting in the release of Nrf 2. Nrf 2 translocates to the nucleus binds to ARE and activates a battery of chemopreventive genes. Once this is achieved, Nrf 2 is exported out of the nucleus, binds with INrf 2, and degrades. Nrf 2 contains well defined signals that control nuclear import and export of Nrf 2. The present studies demonstrate that phosphorylation of tyrosine 568 is required for Crm1-mediated nuclear export and degradation of Nrf 2. Mutation of tyrosine 568 to alanine and phenylalanine resulted in the loss of interaction with Crm1 and abrogation of nuclear export of Nrf 2. Nrf 2Y568A is deficient in nuclear export and displays delayed degradation compared with wild-type Nrf 2. In addition, Src inhibitor PP2 caused nuclear accumulation of Nrf 2 in normal and hydrogen peroxide-treated cells but had no effect on localization of mutant Nrf 2Y568A. Further experiments with small interfering RNA revealed that Fyn phosphorylated Nrf 2Y568 leading to nuclear export and degradation of Nrf 2.

NF-E2-related factor2 (Nrf2) 2 belongs to the family of the leucine zipper/cap'n'collar-containing nuclear factor proteins (1)(2)(3). Nrf2 binds to antioxidant response element (ARE) and regulates expression and coordinated induction of a battery of genes encoding chemopreventive proteins, including detoxifying enzymes NAD(P)H:quinone oxidoreductases (NQO1 and NQO2), glutathione S-transferase Ya subunit, ␥-glutamylcysteinyl synthetase, and heme oxygenase-1 (4). Nrf2-regulated expression and induction comprise a mechanism essential for cellular protection against oxidative stress and neoplasia (4). Nrf2Ϫ/Ϫ mice are viable and live to adulthood showing that Nrf2 is not required for erythropoiesis, development, or growth (5). Nrf2Ϫ/Ϫ mice express significantly lower levels and no induction of chemopreventive proteins, and demonstrate slower wound healing and emphysema in response to tobacco smoke (6 -7).
A cytosolic inhibitor of Nrf2, INrf2 (inhibitor of Nrf2) or KEAP1 (Kelchlike ECH-associated protein1), was reported (8 -9). INrf2 retains Nrf2 in the cytoplasm. The INrf2-Nrf2 complex serves as cellular sensor of oxidative and electrophilic stress generated from endogenous reactions and exogenous chemicals, xenobiotics, drugs, UV, and ionizing radia-tions (4). The treatment of cells with antioxidants and xenobiotics leads to the release of Nrf2 from INrf2. Nrf2 translocates into the nucleus and induces the expression of chemopreventive genes. Reports showed that Nrf2 binding to INrf2 leads to degradation of Nrf2 (4). The antioxidants and xenobiotics lead to dissociation of Nrf2 from INrf2 resulting in stabilization of Nrf2 followed by nuclear translocation. Disruption of INrf2 in mice leads to postnatal death, probably from malnutrition resulting from hyperkeratosis in the esophagus and forestomach, presumably due to nuclear accumulation of Nrf2 (10). The ARE-mediated gene expression was found de-repressed in INrf2 knock-out mice (10). Interestingly, combined knock-out of INrf2 and Nrf2 survive to adulthood but demonstrate compromised expression of chemopreventive proteins and increased sensitivity to oxidative stress (11). This suggested that Nrf2 accumulation in nucleus because of disruption of INrf2 leads to adverse effects on cell growth and survival. Indeed, a recent study demonstrated that Nrf2 accumulation in nucleus for longer time periods leads to apoptotic cell death (12). Therefore, Nrf2 abundance in nucleus is tightly regulated by positive and negative factors (4). Based on these studies, it is suggested that the Nrf2 up-regulation of ARE-mediated gene expression is an early response to antioxidants (4). Late response of antioxidants appears to recruit negative factors, including Bach1:MafG; MafG/K/F:MafG/K/F, c-Jun:c-Fos, and c-Jun-Fra1, presumably to rapidly bring down the induced ARE-mediated gene expression to normal levels. Recently, studies have demonstrated that INrf2 is also localized in the nucleus presumably to degrade Nrf2 (13,14). The mechanism of INrf2 presence inside the nucleus is not understood (14). It is believed that the presence of INrf2 inside the nucleus has something to do with the control of Nrf2 level inside nucleus by controlling degradation of Nrf2 (14).
Several cytosolic kinases that include PKC, mitogen-activated protein kinase, and phosphatidylinositol 3-kinase have been shown to modify Nrf2 and participate in the mechanism of signal transduction from antioxidants and xenobiotics to the ARE (4). Among these, two independent studies have demonstrated antioxidant-induced PKC phosphorylation of serine 40 in Nrf2 leading to dissociation of Nrf2 from INrf2 (15,16). Recently, it was shown that redox modulation of cysteines in INrf2 is capable of releasing Nrf2 from INrf2 (4). It is possible that this mechanism is redundant to the phosphorylation of Nrf2 by PKC, or that the two mechanisms work in concert.
More recently, we identified and characterized a bipartite nuclear localization signal and a leucine-rich nuclear export signal, which regulate Nrf2 shuttling in and out of nucleus and contribute immensely to regulation of nuclear abundance of Nrf2 (17). The study also revealed that increase in oxidative stress first leads to nuclear import of Nrf2 followed by nuclear export of Nrf2. However, the mechanism and function of nuclear export of Nrf2 remain unknown.
Current studies demonstrate that phosphorylation of Nrf2 at tyrosine 568 is essential for nuclear export of Nrf2. Mutants Nrf2Y568A and Nrf2Y568F fail to phosphorylate and thus accumulated in the nucleus. The accumulation of mutants in the nucleus is due to the loss of inter-action of mutant Nrf2 with Crm1 and abrogation of nuclear export. Hydrogen peroxide treatment initially increased nuclear import of Nrf2, presumably to increase the ARE-mediated gene expression to prevent oxidative/electrophilic stress. This was followed by increase in phosphorylation and nuclear export of Nrf2. Further studies with siRNA revealed that Fyn kinase phosphorylated Nrf2Y568, which facilitated nuclear export of Nrf2 for binding to INrf2 and degradation.

MATERIALS AND METHODS
Construction of Plasmids-The construction of pGL2B-NQO1-ARE and pcDNA-Nrf2 has been previously described (18). The pcDNA-Nrf2 was used as a template to construct Nrf2 mutants. The forward 5Ј-GCAGGACATGGATTTGATTGACATCC-3Ј and reverse 5Ј-GTT-TTTCTTTGTATCTGGCTTCTTG-3Ј primers were used to amplify Nrf2 coding region without stop codon. The PCR-amplified product was TA-cloned in pcDNA3.1/V5-His TOPO to generate plasmid pcDNA-Nrf2-V5. GeneTailor TM site-directed mutagenesis kit and protocol (Invitrogen) were used to mutate Nrf2568 tyrosine to alanine. Briefly, the wild-type Nrf2 plasmid was methylated using DNA methylase provided with the kit. A mutagenesis PCR reaction was performed on the methylated plasmid using the primers containing the mutation. Primers, forward, 5Ј-GTGATGAGGATGGAAAGCCTGCCTCTC-CCAGTGAA-3Ј; reverse, 5Ј-AGGCTTTCCATCCTCATCACGTAA-CATGCT-3Ј, were used to amplify the fragment corresponding to Nrf2Y568A. The PCR fragment containing the desired mutation was gelpurified and transformed into maximum efficiency DH5␣ chemically competent cells. A similar strategy was used to generate the Nrf2Y568F mutant. The plasmids were confirmed by sequencing.
Modified pCMV vectors were used to clone the FLAG-tagged INrf2 and Crm1 proteins. A modified polylinker containing a Kozak start sequence, two FLAG epitopes placed adjacent to each other, and stop codons in each frame was cloned into the pCMV vector to generate FLAG-vector. Mouse INrf2 was amplified from the pcDNA-INrf2 plasmid using the primers, forward (5Ј-GCGCTCTAGAGAATTCCG-GAACCCCATGCAGCCCGAA-3Ј) and reverse (5Ј-CGCGGATCCG-ATATCGCAGGTACAGTTTTGTTGATC-3Ј). Mouse Crm1 cDNA was amplified from the IMAGE clone obtained from ATCC using the primers, forward (5Ј-GCGCTCTAGAGAATTCAGTGGCGCAATG-CATGAAGAG-3Ј) and reverse (5Ј-CGCGGATCCGATATCATC-ACACATTTCTTCTGGAAT-3Ј). The PCR-amplified DNA contained XbaI and BamHI restriction sites at 5Ј-and 3Ј-ends, respectively. The amplified DNA was digested with XbaI and BamHI and subcloned into the FLAG vector digested with similar enzymes. The resultant plasmids were designated as pCMV-FLAG-mINrf2 and pCMV-FLAG-mCrm1.
Cell Culture, Co-transfection of Expression Plasmids, and Luciferase Reporter Assay-Human hepatoma (HepG2) cells were grown in monolayer cultures in 6-well plates in minimum essential medium-␣ supplemented with 10% fetal bovine serum. Transient transfections were done in cells grown to ϳ50% confluence using the Effectene Transfection reagent (Qiagen). Cells were co-transfected with 0.2 g of reporter construct (human NQO1-ARE-Luc) and ten times less quantities of firefly Renilla luciferase encoded by plasmid pRL-TK. Renilla luciferase was used as the internal control in each transfection. Wherever indicated, the cells were also co-transfected with 0.5 g of pcDNA expression plasmids encoding wild-type Nrf2 or mutant Nrf2Y568A. To analyze the effect of tyrosine kinase inhibitors on NQO1-ARE activity, the transfected cells were pretreated for 8 h with the indicated kinase inhibitor (Genistein or AG18 or PP2) in the concentrations mentioned in the figures. All the inhibitors were purchased from Calbiochem and were of the highest purity available. Cells were then treated with Me 2 SO or induced with t-BHQ (50 M) for 16 h in the media containing the indicated kinase inhibitors. After the treatment for specified time, the cells were washed with 1ϫ PBS and lysed in 1ϫ Passive lysis buffer from the Dual-Luciferase Reporter Assay System Kit (Promega, Madison, WI). The luciferase activity was measured using the procedures described previously (18). Pre-designed siRNA against mouse Fyn protein and control scrambled siRNA were purchased from Ambion and transfected in Hepa-1 cells using the Effectene transfection reagent following the manufacturer's suggested protocol.
Subcellular Fractionation, Western Blotting, and NQO1 Activity-Hepa-1 cells were transfected with 2.0 g of either pcDNA-Nrf2-V5 or pcDNA-Nrf2Y568A-V5 plasmids using the Effectene transfection reagent as described above. 24 h after transfection, the cells were treated with Me 2 SO, t-BHQ, or PP2 as indicated in the figures. To analyze the localization of endogenous Nrf2, HepG2 cells were seeded in 100-mm plates and treated with Src inhibitor PP2 with or without Me 2 SO or antioxidant t-BHQ, H 2 O 2 with or without nuclear export inhibitor leptomycin B (LMB), or PP2 as indicated in the figures. At the end of treatment, cells were washed twice with ice-cold PBS, scraped in PBS using a rubber policeman, and centrifuged at 500 rpm for 5 min. Biochemical fractionation of the cells was done using the Nuclear Extract Kit (Active Motif, Carlsbad, CA) following the manufacturer's protocol. The protein concentration was determined using the Protein Assay reagent (Bio-Rad, Hercules, CA). 100 g of the cytosolic and 50 g of nuclear fractions were resolved on a 10% SDS-PAGE, Western blotted, and probed with anti-Nrf2 antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA), anti-V5-HRP antibody (Invitrogen), and ␤-actin antibody (Sigma). To confirm the purity of subcellular fractionations, the extracts were Western blotted with cytoplasm-specific anti-lactate dehydrogenase (LDH) antibody (Chemicon International, Temecula, CA) and nuclear specific anti-lamin B antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA). The levels of protein on a Western blot were quantitated by using QuantityOne image software (ChemiDoc XRS, Bio-Rad) and normalized against proper loading controls. The NQO1 activity was determined by previously described procedures (19).
In Vitro Binding-The in vitro transcription/translation of the plasmids encoding Nrf2-V5, Nrf2Y568A-V5, and INrf2 were performed using the TNT-coupled rabbit reticulocyte lysate system (Promega). Redivue L-[ 35 S]methionine (Amersham Biosciences) was substituted for methionine in the reactions to radiolabel the translated proteins. The plasmid encoding luciferase provided in the kit was used as a control for the transcription translation reaction. After the coupled transcription/ translation, the proteins were checked for their correct size by SDS-PAGE, autoradiography, and Western analysis with V5-HRP antibody. V5 antibody was used to detect the V5-tagged wild-type Nrf2 and Nrf2Y568A proteins. All of the in vitro transcribed/translated proteins gave expected size bands. Binding assay: 5 l of each in vitro translated protein (Nrf2-V5ϩINrf2 or Nrf2Y568A-V5ϩINrf2) in protein binding buffer (1 M Tris, pH 7.5, 2 M NaCl, 10% glycerol, 10% Nonidet P-40, 1 M sodium vanadate supplemented with protease inhibitors) were mixed and incubated at 37°C for 30 min. This was followed by addition of 2.5 g of anti-V5 antibody and sufficient protein binding buffer to make the volume to 100 l, and the mixture was incubated overnight at 4°C with shaking. After incubation, 40 l of washed Protein A beads (Santa Cruz Biotechnology) were added, and the mixture was incubated for 1 h at 4°C with shaking. The slurry was centrifuged at 10,000 rpm for 30 s, and the supernatant was discarded. The beads were washed twice with the protein binding buffer. Finally, the beads were boiled in SDS sample dye and analyzed by SDS-PAGE as described above.
Degradation of Nrf2-Hepa-1 cells were grown in 100-mm tissue culture plates and were co-transfected with 2.0 g of either pcDNA-Nrf2-V5 or pcDNA-Nrf2Y568A-V5 plasmids. 24 h after transfection, the cells were pretreated with either Me 2 SO or MG132 (20 M) for 8 h. Cells were washed twice with media and treated with 30 g/ml cycloheximide for different time points (0.5, 1, 2, or 3 h). One set of the cells was left treated with MG132 alone. After treatment for the indicated time points, the cells were washed twice with ice-cold 1ϫ PBS and lysed in RIPA buffer. 100 g of lysate was resolved on a 10% SDS-PAGE, Western blotted, and probed with anti-V5, anti-LDH, anti-lamin B, and anti-␤-actin antibodies.
Immunoprecipitation-Cells either transfected or treated for appropriate times were washed two times with ice-cold PBS and harvested. Cytosolic, nuclear, or whole cell fractions were prepared. Five hundred micrograms of extract was used for immunoprecipitation. Briefly, extract was incubated with either mouse IgG, anti-V5 antibody (Invitrogen) or anti-phospho-tyrosine (anti-pTyr) antibody (clone 4G10, Upstate Biotechnology, Waltham, MA). The immunoprecipitation reaction was performed in RIPA buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholic acid, 0.1% SDS, 1 mM phenylmethylsulfonyl fluoride, and 1 M sodium vanadate supplemented with tyrosine phosphatase inhibitor mixture (Sigma) and protease inhibitors). The extract was incubated with 2.5 g of antibody overnight at 4°C with shaking. 40 l of washed Protein A beads (Santa Cruz Biotechnology) was added, and the mixture was incubated for 1 h at 4°C with shaking. The slurry was centrifuged at 10,000 rpm for 30 s, and the supernatant was discarded. The beads were washed twice with RIPA buffer. 25 l of SDS-sample dye was added, the mixture was boiled, and immunoprecipitates were resolved on a 10% SDS-PAGE followed by immunoblotting with anti-V5HRP, anti-pTyr, anti-Nrf2, or anti-FLAG antibodies. FLAG immunoprecipitation was done using the FLAG-agarose beads (Sigma).
Phosphorylation Analysis-Hepa-1 cells transfected with Nrf2-V5 or Nrf2Y568A-V5 were lysed in RIPA buffer supplemented with tyrosine phosphatase inhibitor mixture and protease inhibitor mixture. Hepa-1 cells co-transfected with Fyn siRNA were also lysed in similar manner. HepG2 cells treated with H 2 O 2 and PP2 were also lysed in RIPA buffer to determine the phosphorylation status of endogenous Nrf2. 500 g of total cell lysate was used to immunoprecipitate with mouse IgG, anti-V5, or anti-pTyr antibodies as described above. The input and immunoprecipitates were boiled in SDS-sample dye and resolved on 10% SDS-PAGE and immunoblotted with respective antibodies.
Pulse-Chase Assay-Hepa-1 cells were transfected with Nrf2-V5 or Nrf2Y568A-V5. The cells 24 h after transfection were incubated with methionine-deficient DMEM (Sigma) for 30 min. The cells were then labeled with methionine-deficient DMEM containing ϳ200 Ci of [S 35 ]methionine mixture (Expre 35 S 35 S, PerkinElmer Life Sciences), for 1 h at 37°C (Pulse). After rinsing with normal culture medium (DMEM supplemented with 10% fetal bovine serum), the cells were chased by normal culture medium supplemented with 100 g/ml L-methionine for 0, 0.5, 1, 2, and 4 h. MG132 (20 M) was added wherever indicated. Cells were rinsed once with PBS and lysed in RIPA on ice for 30 min. Insoluble cellular debris was cleared by centrifugation at 10,000 rpm for 5 min at 4°C. After centrifugation, the supernatants were used for immunoprecipitation with anti-V5 antibody as described earlier.
Immunoprecipitates were boiled in 1ϫ SDS buffer and resolved on 10% SDS gel. The gel was treated with Amplify solution to enhance the 35 S signal, dried, and autoradiographed. The band intensities were quantified using the QuantityOne image software, and the percent Nrf2 or Nrf2Y568A remaining was plotted against time.
Immunofluorescence-Hepa-1 cells were grown in Lab-Tek II chamber slides in DMEM supplemented with 10% fetal bovine serum. Cells were transfected with either pcDNA-Nrf2-V5 or pcDNA-Nrf2Y568A-V5 using the procedures described above. The transfected cells were treated with either Me 2 SO or PP2 (1 M) or LMB (20 ng/ml). For studying the localization of endogenous Nrf2, the untransfected Hepa-1 cells were treated with either Me 2 SO or 50 M Genistein for 2 h. After treatment, the cells were fixed in formalin (Polysciences, Inc., Warrington, PA) and permeabilized with cold acetone (Fisher Scientific, Fair Lawn, NJ). Anti-V5fluorescein isothiocyanate antibody (Invitrogen) was used to probe the cells expressing V5-tagged protein. Similarly, Nrf2 antibody (Santa Cruz Biotechnology Inc.) was used to probe endogenous Nrf2 protein. In this case, the fluorescein isothiocyanate-conjugated anti-rabbit antibody (Chemicon International, Temecula, CA) was used as secondary antibody by procedures as previously described (17). To visualize the nuclei, the cells were stained with Hoechst stain (Bio-Rad). The fluorescent images were captured using appropriate filters in a Nikon eclipse TE 2000-U fluorescence microscope fitted with a Photometrics CoolSnap CF camera.

Effect of Tyrosine Kinase Inhibitors on ARE-mediated Gene Expression and Induction and Phosphorylation of Nrf2Y568-Human hepatoblastoma
Hep-G2 cells were transfected with NQO1 gene ARE-luciferase reporter plasmid, treated with tyrosine kinase inhibitors Genistein or AG18 alone or with Me 2 SO (control) or antioxidant t-BHQ, and analyzed for luciferase activity. Both Genistein and AG18 treatment resulted in a dose-dependent increase in basal and antioxidant-induced ARE-mediated luciferase activity (Fig. 1, A and B). Similar results were also observed with mouse hepatoma Hepa-1 cells (data not shown). We examined the subcellular distribution of endogenous Nrf2 in Me 2 SO and Genistein-treated Hepa-1 cells by immunofluorescence (Fig. 1C). Interestingly, Genistein treatment caused nuclear accumulation of Nrf2 as ϳ80% of cells displayed Nrf2 to be accumulated in the nucleus (Fig.  1C, the graph in the right panel). These results suggested that Nrf2 might be phosphorylated by tyrosine kinase(s) and that inhibition of tyrosine phosphorylation resulted in increased concentration of Nrf2 in the nucleus. Amino acid sequence analysis of mouse Nrf2 revealed the presence of a single putative tyrosine kinase phosphorylation site in the C terminus of Nrf2 protein at residue Tyr-568 (Fig. 1D). This tyrosine residue was found conserved across the species (human, rat, mouse, and chicken). We mutated Nrf2Y568 to Nrf2Y568A and Nrf2Y568F to investigate the role of tyrosine phosphorylation in Nrf2 signaling and ARE-mediated gene expression. To analyze the phosphorylation of Nrf2, Hepa-1 cells were transfected with either Nrf2-V5 or mutant Nrf2Y568A-V5 or Nrf2Y568F-V5, lysed in RIPA buffer, and immunoprecipitated with either anti-V5 or anti-phospho-tyrosine antibodies. The immunoprecipitated proteins were analyzed by SDS-PAGE, Western blotting, and probing with anti-phospho-tyrosine and anti-V5 antibodies to determine phosphorylation of Tyr-568 residue in Nrf2. The results demonstrated that Nrf2 but not mutant Nrf2Y568A and Nrf2Y568F are phosphorylated (Fig. 1, E and F). In other words, mutation of tyrosine 568 in Nrf2 to alanine or phenylalanine resulted in the loss of tyrosine 568 phosphorylation of Nrf2. This was evident from: 1) Western analysis of V5 antibody immunoprecipitates with antiphospho-tyrosine antibody showed immunoprecipitation of Nrf2-V5 (Fig.  1E, lower panel) and tyrosine phosphorylation of Nrf2-V5 (Fig. 1E, upper  panel); 2) in the same experiment, mutant Nrf2Y568A-V5 was immunoprecipitated but did not demonstrate tyrosine phosphorylation (Fig. 1E); and 3) anti-phospho-tyrosine antibody immunoprecipitated Nrf2-V5 but failed to immunoprecipitate tyrosine to alanine Nrf2Y568A-V5 or tyrosine to phenylalanine Nrf2Y568F-V5 mutants of Nrf2 (Fig. 1F). Genistein-induced nuclear accumulation of Nrf2 raised questions regarding enhanced nuclear import because of altered interaction of mutant Nrf2Y568A with INrf2 and/or loss of nuclear export of Nrf2.
Interaction of Nrf2 and Mutant Nrf2Y568 with INrf2-We performed in vitro and in vivo experiments to investigate Nrf2 and mutant Nrf2Y568 interaction with INrf2 (Fig. 2). Luciferase (Luc control), Nrf2-V5, Nrf2Y568A-V5, and INrf2 all were successfully in vitro transcribed and translated ( Fig. 2A, lanes 1-4). [ 35 S]Methionine was included during translation to radiolabel the proteins. The mutant Nrf2Y568A-V5 protein migrated slightly faster than wild-type Nrf2-V5 protein ( Fig.  2A). The translated and 35 S-labeled proteins were mixed in combinations as shown in Fig. 2A (lanes 5-12), immunoprecipitated with mouse IgG (control) and anti-V5 antibody, and analyzed by SDS-PAGE and autoradiography or Western blotting and probing with anti-V5 antibody. The results showed that both Nrf2-V5 and mutant Nrf2Y568A-V5 interacted with INrf2, because anti-V5 antibody immunoprecipitated Nrf2 and mutant Nrf2Y568A and INrf2 was co-precipitated with them ( Fig. 2A,  lanes 11 and 12). In related experiment, Hepa-1 cells were co-transfected with Nrf2-V5 or mutant Nrf2Y568A-V5 with FLAG-INrf2, fractionated into cytosolic and nuclear fractions, and immunoprecipitated with anti-FLAG or anti-V5 antibodies in separate experiments. The immunoprecipitates were analyzed by SDS-PAGE, Western blotting, and probing with anti-FLAG and anti-V5 antibodies. The inputs from the subcellular fractions were probed with cytosol and nuclear-specific antibodies to confirm the purity of fractionation (Fig. 2B). The results demonstrated that anti-FLAG antibody successfully immunoprecipitated FLAG-INrf2 from both cytosolic and nuclear fractions (Fig. 2, C and D, upper  panels). Interestingly, Nrf2-V5 and Nrf2Y568A-V5 were co-precipitated along with FLAG-INrf2 (Fig. 2, C and D, upper panels). Similarly, anti-V5 antibody successfully immunoprecipitated Nrf2-V5 and Nrf2Y568A-V5 from cytosolic and nuclear compartments of transfected cells (Fig. 2, C and D, lower panels). The FLAG-INrf2 was coprecipitated along with Nrf2-V5 and Nrf2Y568A-V5 because of its interaction with these proteins in transfected cells (Fig. 2, C and D, lower

Phosphorylation Controls Nuclear Export of Nrf2
APRIL 28, 2006 • VOLUME 281 • NUMBER 17 panels). Therefore, the results from in vitro translated and overexpressed proteins in transfected cells clearly demonstrate that both mutant Nrf2Y568A and Nrf2 proteins interacted with INrf2 in both cytosol and nuclear fractions. Therefore, Genistein-induced nuclear accumulation is not due to the loss of interaction of mutant Nrf2Y568A with INrf2. The results also demonstrated that INrf2 exists in the nucleus and binds to Nrf2 in similar fashion as in the cytosol.
Nrf2 and Nrf2Y568A Interaction with Crm1 and Nuclear Export of Nrf2-The studies were extended to determine the role of phosphorylation of Nrf2Y568 in nuclear export of Nrf2. Immunohistochemistry and Western assays were performed to investigate subcellular localization of Nrf2-V5 and mutant Nrf2Y568A-V5 in transfected Hepa-1 cells (Fig. 3, A and B). Both assays demonstrated distribution of Nrf2-V5 between cytosol and nucleus. Treatment with nuclear export inhibitor leptomycin B (LMB) led to inhibition of nuclear export and accumula-tion of Nrf2-V5 in the nucleus (Fig. 3, A and B). On the contrary, the mutant Nrf2Y568A localized predominately in the nucleus. LMB had no effect on nuclear localization of Nrf2Y568A-V5. A small amount of Nrf2Y568A-V5 observed in cytosolic fractions (Fig. 3B) is due to cytosolic retention of mutant protein by INrf2. These results combined with results in Fig. 1 led to the conclusion that tyrosine 568 phosphorylation is required for nuclear export of Nrf2. Next, we determined the mechanism of the role of tyrosine 568 phosphorylation in nuclear export of Nrf2. Crm1, also known as exportin 1, is known to bind to several proteins and export them out of the nucleus (20). Immunoprecipitation followed by immunoblotting was used to analyze the interaction of Nrf2-V5 and Nrf2Y568A-V5 with FLAG-Crm1 in transiently transfected Hepa-1 cells (Fig. 3C). Because it is known that antioxidant t-BHQ induces nuclear export of Nrf2 at 4 h of treatment (17), we used t-BHQ to enhance interaction between Nrf2 and Crm1. Immunopre-  1-4), resolved on a 10% SDS-PAGE, treated with Amplify solution to enhance the 35 S signal, dried, and autoradiographed (upper panel) or Western blotted and probed with anti-V5 antibodies (lower panel). For in vitro binding assay, equal amounts of in vitro translated proteins were mixed in binding buffer in combinations as displayed (lanes 5-12), incubated at 37°C for 30 min, and immunoprecipitated with either mouse IgG or anti-V5 antibody. The immunoprecipitates along with input controls were analyzed for 35 S signal by autoradiography (upper panel) and by SDS-PAGE, Western blotting, and probing with anti-V5 antibodies (lower panel). B-D, co-immunoprecipitation assay. Hepa-1 cells were co-transfected with Nrf2-V5 or Nrf2Y568A-V5 along with FLAG-INrf2 in a 4:1 ratio. The cells were harvested 24 h after transfection and cytosol, and nuclear extracts were prepared by standard procedures. 50 g of cytosol/nuclear extracts (input) were immunoblotted with anti-V5 and anti-FLAG antibodies, the same blot was also reprobed with anti-lamin B and anti-LDH antibodies to confirm the purity of fractionation. 500 g of extracts was immunoprecipitated with IgG, anti-V5, or anti-FLAG antibody. The input (one of five) and immunoprecipitates were resolved on a 10% SDS-PAGE and Western blotted with anti-V5 and anti-FLAG antibodies, respectively. IP, immunoprecipitation; WB, Western blotting.  cipitation was performed with anti-V5 or anti-FLAG antibodies, and Western blots were probed with both anti-V5 and anti-FLAG antibodies (Fig. 3C). Anti-V5 antibodies immunoprecipitated Nrf2-V5 and FLAG-Crm1 was co-precipitated in the immunoprecipitate (Fig. 3C, left  panel, lanes 6 and 8). However, in same experiment, FLAG-Crm1 failed to co-precipitate with mutant Nrf2Y568A-V5 immunoprecipitated with anti-V5 antibodies (Fig. 3C, left panel, lanes 7 and 9). In the reverse experiments, the anti-FLAG antibodies successfully immunoprecipitated FLAG-Crm1. Nrf2-V5 but not mutant Nrf2Y568A-V5 co-precipitated with FLAG-Crm1 (Fig. 3C, right panel compare lanes 6 -9). These observations suggest that tyrosine phosphorylation of Nrf2Y568 is required for interaction with Crm1 and nuclear export of Nrf2.
Hydrogen Peroxide-mediated Induction of Tyrosine 568 Phosphorylation and Nuclear Export of Nrf2-Hydrogen peroxide, an oxidizing agent, was used to study in vivo relevance of phosphorylation of tyrosine 568 in Nrf2 export. Hep-G2 cells were treated with hydrogen peroxide for different time intervals, and the subcellular localization of endogenous Nrf2 was followed by immunoblotting.
Hydrogen peroxide treatment of Hep-G2 cells led to nuclear import of Nrf2 within 30 min of treatment, presumably to increase expression of Nrf2 downstream genes to provide cellular protection against hydrogen peroxide-induced oxidative stress (Fig. 4A, ϪLMB). Once this is achieved, Nrf2 started exiting the nucleus 2 h after hydrogen peroxide treatment and was reduced to lower than control (untreated) levels by 4 h after treatment (Fig. 4A, ϪLMB). The nuclear export of Nrf2 was LMB-sensitive, as there was little or no clearance of Nrf2 from the nucleus when treated with H 2 O 2 in the presence of LMB (Fig. 4A, right panel, ϩLMB). The nuclear levels of Nrf2 were quantitated by densitometry and displayed as a graph, which demonstrates a significant reduction in the amount of Nrf2 from the nucleus at 4 h of H 2 O 2 treatment but not in the presence of LMB (Fig. 4A, right panel, last columns, p Ͼ 0.005). Interestingly, the Src family of tyrosine kinase inhibitor PP2 significantly blocked the hydrogen peroxide-induced nuclear export of Nrf2 beginning at 2 h (p Ͻ 0.005) (Fig. 4B, left panel and graph). This observation with PP2 was similar to that observed with nuclear export inhibitor LMB (Fig.   FIGURE 5. Nrf2-V5 degraded at faster rate than mutant Nrf2Y568A. A, Hepa-1 cells were transfected with Nrf2-V5 or Nrf2Y568A-V5. The transfected cells after 24 h were pretreated without (ϪLMB) or with leptomycin B (20 ng/ml; ϩLMB) followed by treatment with either Me 2 SO or MG132 (20 M) for 6 h. MG132-treated cells were then treated with 30 g/ml cycloheximide (CHX) for different time points (0.5, 1, 2, or 3 h). Cells were harvested, lysed, and immunoblotted with anti-V5 and anti-actin antibodies. Nrf2 levels were normalized to ␤-actin levels by using QuantityOne image software, and the percent Nrf2 remaining is plotted against time (left panel, Nrf2-V5; right panel, Nrf2Y568-V5). B, pulse-chase assay. Hepa-1 cells transfected with Nrf2-V5 or Nrf2Y568A-V5 were metabolically labeled with [ 35 S]methionine (pulse). 45 min after pulse the medium was replaced with complete medium containing cold methionine, and cells were harvested at different time points (chase). MG132 was added to the media wherever indicated. Cell lysate was prepared using RIPA buffer; 1 mg of cell lysate was immunoprecipitated with 2 g of anti-V5 antibody. The immunoprecipitates were resolved on 10% SDS-PAGE and autoradiographed for 35 S signal. Percent Nrf2-V5 or Nrf2Y568A-V5 remaining was quantitated and plotted against time. A with B). In addition, PP2 also blocked the tyrosine phosphorylation of Nrf2 at 2 h of hydrogen peroxide treatment (Fig.  4C). These results suggested that hydrogen peroxide-induced nuclear export of Nrf2 is mediated via tyrosine phosphorylation of Nrf2 by Src kinase(s).

4, compare
Nuclear Export and Degradation of Nrf2-The experiments in Fig. 2 indicated that both Nrf2 and Nrf2Y568A interacted with INrf2. INrf2 is known to function as cul3-based E3 ligase in Nrf2 degradation (21,22). Experiments were performed to compare the rate of degradation of Nrf2 and mutant Nrf2Y568A so as to determine the role of phosphorylation of tyrosine 568 in Nrf2 degradation. Based on the assumptions that Nrf2 is degraded in the cytoplasm, we hypothesized that Nrf2Y568A, which is deficient in export, will be more stable than Nrf2. The results of the degradation experiment showed that the rate of degradation of mutant Nrf2Y568A was significantly slower than wild-type Nrf2 (Fig.  5A, ϪLMB compare top lanes of left and right panels). Blocking of the nuclear export of Nrf2 with LMB showed similar slower rate of degradation as mutant Nrf2Y568A (Fig. 5A, ϩLMB, left panel). In the same experiment, LMB had no effect on rate of degradation of mutant Nrf2Y568A with compromised nuclear export (Fig. 5A, right panel). The results are also displayed as graphs plotted for percent Nrf2 remaining in the cells against time (Fig. 5A). The results clearly demonstrate that nuclear export of Nrf2 is required for degradation of Nrf2. Therefore, both the mutation and LMB treatment reduced the rate of degradation of Nrf2 due to blocking of nuclear export. To further elucidate these observations, we performed pulse-chase analysis in Hepa-1 cells after transient transfection with Nrf2-V5 and Nrf2Y568A-V5 (Fig. 5B). The results from pulse-chase analysis were in agreement with the in vivo degradation analysis. Nrf2 degraded much faster compared with Nrf2Y568A (Fig. 5B). This suggested that Nrf2Y568 phosphorylation leads to nuclear export of Nrf2 that binds to INrf2 and degrade.
PP2-mediated Inhibition of Nuclear Export of Nrf2 and Activation of Nrf2 Downstream Genes-Western and immunohistochemistry assays revealed that treatment of transfected Hepa-1 cells with PP2 led to time-dependent nuclear accumulation of Nrf2-V5 (Fig. 6A, left panel Cells were harvested, and cytosol and nucleus fractions were prepared and analyzed by immunoblotting with anti-Nrf2 and nuclear specific anti-lamin B antibody. Nrf2 levels were normalized to lamin B levels and the -fold induction in the amount of Nrf2 in the nucleus in Me 2 SO/t-BHQ with or without PP2-treated cells is shown (D). E and F, in a similar experiment the cytosolic extracts were probed against anti-NQO1 antibody to determine NQO1 protein levels (E) and measured cytosolic NQO1 activity (F). One unit of NQO1 activity is the amount of activity that reduced 1 mol of 2,6-dichlorophenolindophenol in 1 min. The values represent mean Ϯ S.E. of three independent transfection experiments. *, nonspecific band. and Fig. 6B). However, PP2 did not affect the nuclear localization of mutant Nrf2Y568A-V5 (Fig. 6A, right panel and Fig. 6B). In related experiments, PP2 treatment increased ARE-mediated luciferase activity in Nrf2-V5-overexpressing cells but not in cells overexpressing mutant Nrf2Y568A-V5 (Fig. 6C). In further experiments, PP2 treatment induced nuclear accumulation of endogenous Nrf2 in Me 2 SO-and t-BHQ-treated cells (Fig. 6, D and graph). The PP2-induced nuclear accumulation of Nrf2 was followed by increased expression of Nrf2 downstream gene NQO1 (Western analysis in Fig. 6E) and NQO1 activity (Fig. 6F). PP2 is a known specific inhibitor of Src tyrosine kinase family of enzymes (23). PP2-mediated inhibition of Nrf2Y568 phosphorylation (Fig. 4) and nuclear export of Nrf2 (Fig. 6) indicated that one or more members of the Src family of tyrosine kinases might have a role in phosphorylating Nrf2Y568.
Fyn siRNA-mediated Inhibition of Nuclear Export and Degradation of Nrf2 and Activation of ARE-luciferase-Four members of the Src family, namely Fyn, Src, Lyn, and Yes, are ubiquitously expressed as Nrf2 (4,23). Among the Src kinases, Fyn kinase is interesting, because it is known to be phosphorylated in response to UV followed by localization of phosphorylated Fyn in the nucleus (24). Therefore, we studied the role of Fyn in phosphorylation of Nrf2Y568. Hepa-1 cells were transfected with mouse Fyn siRNA or control siRNA and analyzed for Fyn expression, Nrf2Y568 phosphorylation, nuclear localization, and stability of Nrf2 and ARE-mediated luciferase gene expression and induction in response to antioxidant t-BHQ (Fig. 7). Western analysis revealed that Fyn siRNA transfection caused dose-dependent inhibition of Fyn, which was unaffected by control siRNA (Fig. 7A). Fyn siRNA also catalyzed dose-dependent inhibition of phosphorylation of Nrf2-V5 (Fig. 7B, blot and the graph in the lower panel). In addition, Fyn siRNA led to dosedependent stabilization (Fig. 7C) and nuclear accumulation of Nrf2 ( Fig.  7D) but had no effect on stability and localization of mutant Nrf2Y568A (Fig. 7, C and D). Furthermore, Fyn siRNA but not control siRNA showed dose-dependent increase in ARE-mediated luciferase gene expression in mock and Nrf2-V5-transfected cells (Fig. 7E). The Fyn siRNA-mediated increase in ARE-mediated luciferase gene expression and induction was absent in cells overexpressing mutant Nrf2Y568A-V5 (Fig. 7E). The slight increase in luciferase activity observed was expected because of endogenous Nrf2. These results indicated that Fyn, a member of the Src family of tyrosine kinases, phosphorylated Nrf2Y568 and this phosphorylation is required for nuclear export and degradation of Nrf2.

DISCUSSION
The studies have shown that INrf2-Nrf2 complex serves as an oxidative sensor generated from chemicals, xenobiotics, drugs, UV, and radiation (4). This leads to dissociation of Nrf2 from INrf2. Nrf2 moves to the nucleus and binds to ARE. This results in coordinated activation of a battery of greater than 100 chemopreventive genes essential for protection against oxidative stress, cellular transformation, neoplasia, and other adverse effects. Nrf2 contains well defined signals that control its B, immunoprecipitation/Western analysis. Fyn siRNA inhibited phosphorylation of Nrf2-V5. Hepa-1 cells co-transfected with Nrf2-V5 or Nrf2Y568A-V5 with Fyn siRNA were analyzed by immunoprecipitation with anti-phosphotyrosine antibody and immunoblotted with anti-V5 antibody. Nrf2 levels were quantitated by using QuantityOne image software, and the percent phosphorylated Nrf2 in presence of Fyn siRNA is shown. The densitometry results are presented as Ϯ S.E. of three independent experiments and representative blot is shown (*, p Ͻ0.005). C and D, subcellular fractionation and Western analysis. Fyn siRNA stabilized Nrf2 and led to nuclear accumulation of Nrf2. Hepa-1 cells co-transfected with Nrf2-V5 or Nrf2Y568A-V5 with Fyn siRNA were subcellular fractionated to prepare cytosolic and nuclear extracts. Total lysate (C), cytosolic and nuclear extracts (D) were analyzed by Western blotting and probing with anti-V5 antibody. The blots were also probed with anti-actin (equal loading), anti-lamin B (nuclear-specific), and anti-LDH (cytosol-specific) antibodies. E, ARE-Luciferase assay. Hepa-1 cells co-transfected with Nrf2-V5 or Nrf2Y568A-V5 and increasing concentration of Fyn or control siRNA and NQO1 gene ARE-luciferase plasmids were treated with Me 2 SO or t-BHQ (50 M for 16 h) and analyzed for luciferase activity.
nuclear import and export (17). PKC-mediated phosphorylation of Nrf2S40 is known to dissociate Nrf2 from INrf2 leading to its nuclear import and activation of ARE-mediated gene expression (15,16). However, what regulates nuclear export of Nrf2 is unknown.
The present studies demonstrate that tyrosine 568 in Nrf2 is phosphorylated, and this phosphorylation is essential for Nrf2 binding with Crm1 and nuclear export. Mutation of tyrosine 568 to alanine or phenylalanine resulted in the loss of phosphorylation and interaction of Nrf2 with Crm1 and abrogation of nuclear export of Nrf2. The wild-type Nrf2 and mutant Nrf2Y568A both interacted with INrf2 and were released/imported in the nucleus in response to endogenous cellular stressors. The mutant Nrf2Y568A lacking the tyrosine phosphorylation accumulated in the nucleus due to the loss of nuclear export of mutant protein. This was clearly evident from the observations that accumulation of mutant protein inside nucleus was insensitive to nuclear export inhibitor leptomycin B (LMB) and was similar to nuclear accumulation of wild-type Nrf2 protein in response to leptomycin B. The studies also indicated that hydrogen peroxide initially led to nuclear accumulation of Nrf2, presumably to activate chemoprotective genes, and later induced phosphorylation of tyrosine 568 for enhanced nuclear export of Nrf2. The mechanism of phosphorylated Nrf2 interaction with Crm1 remains unknown. It is expected that phosphorylation of Nrf2Y568 leads to structural changes that expose the leucine-rich nuclear export signal region (amino acid 545-554) for interaction with Crm1. The results further indicated that exported Nrf2 binds to INrf2 and degrades. The mutant Nrf2Y568A failed to exit the nucleus and degraded at a more reduced rate than wild-type Nrf2. The regulation of nuclear export and degradation of Nrf2 is one mechanism of high significance that controls Nrf2 abundance inside the nucleus. This is especially important because cells face constant everyday challenge of oxidative stress that continuously leads to import of Nrf2 that after use has to be exported out and degrade. The accumulation of Nrf2 inside the nucleus for longer time is lethal to cell survival (12). 3 It is well established that INrf2 binds to Nrf2 and retains it in the cytosol until the signal is received to release Nrf2. INrf2 is also known to function as Cul3-based E3 ligase in proteasomal degradation of Nrf2 (21,22). The results from present studies demonstrate that tyrosine 568 phosphorylation is not essential for binding of Nrf2 with INrf2, because both Nrf2 and mutant Nrf2Y568A interacted with INrf2. This observation leads to an interesting hypothesis that INrf2 could recognize and differentiate between de novo synthesized unphosphorylated Nrf2 and the nuclear exported tyrosine-phosphorylated Nrf2. This recognition might allow INrf2 to target the nuclear exported phosphorylated Nrf2 for degradation and hold de novo synthesized unphosphorylated Nrf2 until a signal is received for its release. The viability of this reasonable hypothesis remains to be determined by experiments. The only result that requires explaining to fit in this hypothesis is the observation of INrf2-mediated degradation of mutant Nrf2Y568A inside the nucleus. It is possible that INrf2 recognized the mutant Nrf2Y568A the same as phosphorylated Nrf2 and targeted it for degradation.
The present studies suggest that a member of the Src family of tyrosine kinases, Fyn, phosphorylates Nrf2Y568 and regulates nuclear export and degradation of Nrf2. Fyn siRNA inhibited Fyn kinase expres-3 A. K. Jain and A. K. Jaiswal, unpublished observation.

Phosphorylation Controls Nuclear Export of Nrf2
APRIL 28, 2006 • VOLUME 281 • NUMBER 17 sion, blocked phosphorylation of Nrf2, and led to stability and nuclear accumulation of Nrf2 due to abrogation of nuclear export of Nrf2. It is possible but unknown if other Src kinases, including Src, Lyn, and Yes are also capable of phosphorylating tyrosine 568 in Nrf2. Fyn-mediated phosphorylation of Nrf2 also raises questions regarding mechanism by which Fyn kinase receives signals from hydrogen peroxide and other chemicals leading to activation and nuclear export of Nrf2.
In conclusion, we demonstrated that Fyn kinase-mediated phosphorylation of tyrosine 568 regulates Nrf2 interaction with Crm1 and nuclear export. The phosphorylated Nrf2 is degraded in the cytosol by binding to INrf2. A model is described in Fig. 8. PKC-mediated phosphorylation of Nrf2S40 leads to the release and nuclear translocation of Nrf2. Nrf2 binds with ARE and activates gene expression. Once this is done, Nrf2Y568 is phosphorylated by Fyn and presumably by other Src kinases leading to nuclear export of Nrf2 that binds to INrf2 and degrades. INrf2 could also translocate inside the nucleus, bind to Nrf2, and degrade Nrf2 inside the nucleus, and/or INrf2-Nrf2 complex is transported out for degradation of Nrf2 in the cytosol. The mutant Nrf2Y568A accumulates in the nucleus due to lack of nuclear export. INrf2 binds and degrades mutant Nrf2Y568A inside the nucleus or might bring it out in the cytosol and degrade it.