GSK-3β Acts Upstream of Fyn Kinase in Regulation of Nuclear Export and Degradation of NF-E2 Related Factor 2*

NF-E2-related factor 2 (Nrf2) regulates expression and coordinated induction of a battery of chemoprotective genes in response to oxidative and electrophilic stress. This leads to protection against oxidative stress and neoplastic diseases. Nuclear import and export of Nrf2 play a significant role in control of nuclear levels of Nrf2 and thus the expression of Nrf2 down-stream genes. Tyrosine kinase Fyn phosphorylates tyrosine 568 of Nrf2 that leads to the nuclear export of Nrf2. In this study, we investigated the upstream factor(s) in regulation of Fyn and Fyn-mediated nuclear export of Nrf2. The investigations shed light on a novel mechanism of Nrf2 regulation in response to oxidative stress. We demonstrate that GSK-3β acts upstream of Fyn kinase in control of nuclear export of Nrf2. Chemical and short interfering RNA-mediated inhibition of GSK-3β led to nuclear accumulation of Nrf2 and transcriptional activation of the Nrf2 downstream gene nqo1. Chemical and short interfering RNA inhibition of GSK-3β and Fyn individually and in combination revealed that both kinases follow the same pathway to regulate nuclear export of Nrf2. We further demonstrate that hydrogen peroxide phosphorylates tyrosine 216 of GSK-3β. This leads to activation of GSK-3β. The activated GSK-3β phosphorylates Fyn at threonine residue(s). Phosphorylated Fyn accumulates in the nucleus and phosphorylates Nrf2 at tyrosine 568. This leads to nuclear export, ubiquitination, and degradation of Nrf2.

NF-E2-related factor (Nrf2) 2 is a nuclear transcription factor that binds to antioxidant-response element (ARE) and regulates expression and coordinated induction of a battery of chemoprotective genes in response to antioxidants, oxidants, and radiations (1). This induction involves a mechanism essential for cellular protection against oxidative and electrophilic stress and neoplastic diseases (1). Nrf2-null mice are born normal and viable, indicating that Nrf2 is not required for development and growth of mice (2). Nrf2-null mice express significantly lower levels and no induction of chemoprotective proteins that include NAD(P)H:quinone oxidoreductase 1 (NQO1), glutathione S-transferase Ya subunit (GST-Ya), ␥-glutamylcysteinyl synthetase, and heme oxygenase (HO-1) (1). These mice demonstrate slower wound healing and pulmonary emphysema in response to exposure to tobacco smoke (3,4).
A cytosolic inhibitor of Nrf2, INrf2 (inhibitor of Nrf2) or Keap1, retains Nrf2 in the cytoplasm (5,6). The INrf2-Nrf2 complex serves as cellular sensor of oxidative and electrophilic stress (1). The cellular exposure to oxidants, antioxidants, and radiations antagonizes this interaction and leads to the release of Nrf2 from INrf2 (1). Nrf2 translocates into the nucleus and induces the expression of chemoprotective proteins. Electrophilic adduction of selected cysteines in INrf2 and PKC phosphorylation of Nrf2 are known to contribute to the release of Nrf2 from INrf2 (1). Disruption of INrf2 in mice leads to postnatal death, probably from malnutrition resulting from hyperkeratosis in the esophagus and forestomach, presumably because of excessive nuclear accumulation of Nrf2 (7). Therefore, sustained nuclear accumulation of Nrf2 is lethal to the cells. Indeed, longer retention of Nrf2 in the nucleus leads to apoptotic cell death (8).
The abundance of Nrf2 inside the nucleus is tightly regulated by positive and negative factors that control nuclear import, binding to ARE, export, and degradation of Nrf2 under normal and stress (chemical/radiation) conditions (9 -12). The early response to stress leads to nuclear import of Nrf2 resulting in coordinated activation of chemoprotective genes. The delayed response to stress is Fyn-mediated phosphorylation of Nrf2Y568 inside the nucleus (13). Tyrosine 568 phosphorylation leads to nuclear export of Nrf2 (13). However, upstream events that control signal transduction from oxidative and electrophilic stress to Fyn that phosphorylates Nrf2Y568 leading to nuclear export of Nrf2 remains unknown. In this study, we investigated the upstream factors that regulate Fyn and nuclear export and degradation of Nrf2.

MATERIALS AND METHODS
Construction of Plasmids-The construction of pGL2B-NQO1-ARE and pcDNA-Nrf2 has been described previously (14). The construction of pcDNA-Nrf2-V5, pcDNA-Nrf2Y568A-V5, and pCMV-FLAG-mINrf2 was also described previously (11). Modified pCMV vectors were used to clone the FLAGtagged Fyn protein. Mouse Fyn cDNA was amplified from the IMAGE clone obtained from ATCC using the following primers: forward 5Ј-GCGCTCTAGAGAATTCGTCGAGACCAT-GGGCTGTGTG-3Ј and reverse 5Ј-CGCGGATCCGATATCC-AGGTTTTCACCAGGTTGGTA-3Ј. The PCR-amplified DNA contained XbaI and BamHI restriction sites at the 5Ј and 3Ј end, * This work was supported by National Institutes of Health Grants RO1 GM47466 and RO1 ES012265. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1  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, Valencia, CA). Cells were co-transfected with 0.2 g of reporter construct (human NQO1-ARE-Luc) and 10 times less quantities of firefly Renilla luciferase encoded by plasmid pRL-TK. Renilla luciferase was used as the internal control in each transfection. To analyze the effect of GSK-3␤ inhibitors on NQO1-ARE activity, the transfected cells were treated for 12 h with the indicated inhibitor (lithium chloride or TDZD-8 or PP2) in the concentrations as indicated in the figures. Lithium chloride was purchased from Sigma, and TDZD-8 and PP2 were purchased from Calbiochem and were the highest purity available. After the treatment for the specified times, the cells were washed with 1ϫ phosphate-buffered saline 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 (15).
siRNA Transfection-Mouse GSK-3␤ siRNA, Fyn siRNA, and Lamin A/C siRNA (control) were purchased from Dharmacon and were transfected using Lipofectamine transfection reagent (Invitrogen) following the manufacturer's protocol. For luciferase assay, mouse Hepa-1 cells were co-transfected with 0.2 g of reporter construct (human NQO1-ARE-Luc) and 10 times less quantities of firefly Renilla luciferase encoded by plasmid pRL-TK as described above along with the indicated amounts of GSK-3␤ or Fyn siRNA. The luciferase analysis was done as described above. For Western analysis, Hepa-1 cells in 100-mm plates were transfected with siRNA in different doses, and subcellular fractionation and immunoblotting were done as described below.
Subcellular Fractionation and Western Blotting-HepG2/Hepa1 cells, seeded in 100-mm plates and treated/transfected as displayed in the figures, were washed twice with ice-cold phosphate-buffered saline, scraped in phosphate-buffered saline 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. For making total cell lysate, the cells were lysed 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 mM sodium vanadate supplemented with protease inhibitor mixture (Roche Applied Science), and phosphatase inhibitor mixtures I and II (Sigma)). The protein concentration was determined using the protein assay reagent (Bio-Rad). 100 g of total cell lysate or cytosolic or nuclear fractions were resolved on a 10% SDS-polyacrylamide gel, Western-blotted, and probed with anti-Nrf2 antibody, anti-Fyn antibody (both from Santa Cruz Biotechnology), anti-GSK-3␤ antibody (Cell Signaling), anti-Tyr(P)-GSK-3 antibody (BIOSOURCE), anti-FLAG-HRP antibody (Sigma), or anti-NQO1 antibody (11). 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 A-C, siRNA-mediated inhibition of GSK-3␤. A and B, Western analysis. Hepa-1 cells were transfected with GSK-3␤ siRNA in concentrations as indicated. The transfected cells were harvested, subjected to subcellular fractionation to prepare cytosolic and nuclear extracts (Nuc Ext), or lysed to obtain total cell lysate. The cytosolic and nuclear extracts (A) and total cell lysate (B) were analyzed by Western blotting and probing with anti-GSK-3␤ and anti-Nrf2 antibody. The blots were also probed with anti-Lamin B (nuclear specific), anti-LDH (cytosol-specific), and ␤-actin (equal loading) antibodies. GSK-3␤ and Nrf2 levels were normalized to ␤-actin levels by using QuantityOne Image software, and the fold amount of protein is plotted versus the amount of GSK-3␤ siRNA. The densitometry results are presented as Ϯ S.E. of three independent experiments, and a representative blot is shown. C, ARE-luciferase assay. Hepa-1 cells co-transfected with NQO1-ARE luciferase reporter, firefly Renilla luciferase, and increasing concentrations of either control siRNA (left panel) or GSK-3␤ siRNA (right panel) were harvested and analyzed for luciferase activity. D, Western analysis. Hepa1 cells transfected with GSK-3␤ siRNA were lysed in RIPA buffer as in B and analyzed for NQO1 and GSK-3␤ protein levels by immunoblotting. E, RT-PCR. In a similar experiment, total RNA was prepared from Hepa1 cells transfected as in D, and RT-PCR was done using the primers as indicated under "Materials and Methods." (Santa Cruz Biotechnology). 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.
Phosphorylation Analysis-HepG2 cells were seeded in 100-mm plates and transfected with 0.5 g of pCMV-FLAG-mFyn plasmid. 24 h after transfection, the cells were pretreated with LiCl/PP2 for 8 h followed by treatment with hydrogen peroxide (H 2 O 2 ) Ϯ LiCl/PP2 for 1 and 4 h. At the end of treatment, the cells were harvested, and biochemical fractionation was done following the procedures described above. To analyze tyrosine phosphorylation of GSK-3␤, 100 g of total cell lysate was resolved on 10% SDS-PAGE and immunoblotted with phosphospecific (Tyr(P))-GSK-3 antibody (BIOSOURCE). The same membrane was reprobed with anti-GSK-3␤ and anti-␤-actin antibodies. To analyze the Fyn phosphorylation, 1 mg of nuclear extract was used to immunoprecipitate either with anti-Thr(P) antibody (Cell Signaling) or anti-FLAGM2 beads (Sigma). Briefly, nuclear extract supplemented with 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium vanadate, Ser/Thr phosphatase inhibitor mixture (Sigma), and protease inhibitor mixture (Roche Applied Science) was incubated with 2.5 g of antibody overnight at 4°C with shaking. 40 l of washed protein A beads (Santa Cruz Biotechnology) were added and 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 to the beads and boiled, and immunoprecipitates were resolved on 10% SDS-polyacrylamide gel followed by immunoblotting with anti-Ser(P) or anti-FLAG-HRP antibody.
In Vitro Kinase Assay-Purified GSK-3␤, phospho-GSK-3 peptide, and Fyn kinase were obtained from Upstate Biotechnology, Inc. His-Nrf2 was purified from bacterial lysates using His-Trap columns (GE Healthcare). 100 ng of GSK-3 peptide, Fyn kinase, or His-Nrf2 were incubated with active GSK-3␤ in the kinase assay buffer (20 mM Tris, pH 7.5, 10 mM MgCl 2 , 5 mM dithiothreitol, 200 M ATP, and 1.5 Ci of [␥-32 P]ATP) for 30 min at 30°C. Phospho-GSK-3 peptide (3 kDa) was used as a positive control for in vitro GSK-3␤ kinase reaction. Reaction was stopped by adding 2ϫ SDS-gel loading dye followed by boiling for 5 min. The samples were then resolved on a 12% SDS-polyacrylamide gel followed by autoradiography.
Ubiquitination of Nrf2-HepG2 cells were seeded in 100-mm plates and co-transfected either with pcDNA-Nrf2-V5 (1.0 g), pCMV-HA-Ub (0.5 g), or pCMV-FLAG-mINrf2 (0.25 g) in different combinations as indicated in the figure. The transfected cells were treated with either Me 2 SO or 50 mM GSK-3␤ inhibitor LiCl or 20 M proteasomal inhibitor MG132 for 5 h. The cells were lysed in RIPA buffer, and 100 g of protein were analyzed by SDS-PAGE, Western blotting, and probing with anti-V5, anti-FLAG, and ␤-actin antibodies. In a similar experiment, 1 mg of protein was immunoprecipitated with anti-V5 antibody and immunoblotted with anti-HA antibody. The cells were harvested, subjected to subcellular fractionation to prepare cytosolic and nuclear extracts (Nuc Ext), or lysed to prepare total cell lysate. The cytosolic and nuclear extracts (A and B, left panels) and total cell lysate (A and B, middle panels) were analyzed by Western blotting and probing with anti-Nrf2 antibody. The blots were also probed with anti-Lamin B (nuclear specific), anti-LDH (cytosol-specific) and anti-actin (equal loading) antibodies. A and B, right panels, ARE-luciferase assay. HepG2 cells were co-transfected with NQO1-ARE luciferase reporter and firefly Renilla luciferase as described under "Materials and Methods." Thirty six h after transfection, the cells were treated with either TDZD-8 (10, 25 or 50 M in top panel) or lithium chloride (10, 25, or 50 mM, bottom panel) for 8 h. Cells were harvested, lysed, and analyzed for the luciferase activity. The results are presented as Ϯ S.E. of three independent experiments, and each experiment was done in triplicate. C, Western analysis. Hepa1 cells were treated with GSK-3␤ inhibitor; LiCl were lysed in RIPA buffer as in B and analyzed for NQO1 protein levels by immunoblotting. D, RT-PCR. In a similar experiment, total RNA was prepared from Hepa1 cells treated as in C, and RT-PCR was done as described in Fig. 1.

RESULTS
After a series of experiments, we found that siRNA against GSK-3␤ inhibited GSK-3␤ and led to nuclear accumulation of Nrf2 as analyzed by immunoblotting and probing with anti-GSK3␤ and anti-Nrf2 antibodies (Fig. 1A). The blots were reprobed with anti-LDH (cytosol-specific) and anti-Lamin B (nuclear specific) antibodies to confirm the purity of cytosolic and nuclear fractions. GSK-3␤ is a serine/threonine kinase that regulates vital processes of cell division and apoptosis (16). Interestingly, siRNA-mediated inhibition of GSK-3␤ also led to stabilization of Nrf2 (Fig. 1B) and induction of the Nrf2 downstream gene nqo1 ARE-luciferase expression (Fig. 1C, right panel). The nuclear accumulation and stabilization of Nrf2 was GSK-3␤ siRNA concentration-dependent. Nuclear accumulation of Nrf2 after GSK-3␤ siRNA transfection resulted in higher NQO1 protein levels (Fig. 1D, left panel). This increase in NQO1 protein was because of Nrf2-mediated increased NQO1 transcription as analyzed by RT-PCR analysis (Fig. 1E). The increased nuclear accumulation and stability of Nrf2 was not because of increased transcription of Nrf2 as Nrf2 transcript levels did not change after GSK-3␤ siRNA transfection (Fig.  1E). In a related set of experiments, the treatment of cells with chemical based GSK-3␤ inhibitors, lithium chloride (LiCl) and TDZD-8, showed similar results as observed with GSK-3␤ siRNA (Fig. 2). Both inhibitors enhanced nuclear accumulation and stabilization of Nrf2 in total cell lysates and increased ARE-luciferase activity (Fig. 2, A and B). Lithium chloridemediated inhibition of GSK-3␤ enzyme activity also led to increased NQO1 protein (Fig. 2C) and gene expression without affecting Nrf2 transcription (Fig. 2D). These results suggested that a decrease in GSK-3␤ protein or inhibition of GSK-3␤ activity led to nuclear accumulation and stabilization of Nrf2 without affecting its transcription. The nuclear accumulation of Nrf2 in response to GSK-3␤ inhibitors is because of blocking of nuclear export of Nrf2 as reported previously (17).
Earlier, we have shown that Fyn kinase-mediated phosphorylation of Nrf2Y568 is essential for the nuclear export of Nrf2 (13). Therefore, the treatment of human hepatoblastoma (HepG2) cells with the Fyn inhibitor PP2 led to nuclear accumulation and stabilization of Nrf2 due to blocking of nuclear export of Nrf2 (13). The inhibition of GSK-3␤ also led to nuclear accumulation and stabilization of Nrf2 due to the inhibition of nuclear export of Nrf2 (Fig. 1). This raised questions if GSK-3␤ is upstream to Fyn because Fyn directly phosphorylated Nrf2Y568 or GSK-3␤ and Fyn regulated nuclear export of Nrf2 by two independent mechanisms? We performed experiments to determine whether GSK-3␤ inhibitor LiCl and Fyn inhibitor PP2 act independent of each other or in concert leading to nuclear accumulation of Nrf2. HepG2 cells were treated with different doses of either LiCl or PP2 or both in combination; cytosol and nuclear fractions were prepared and analyzed by SDS-PAGE, immunoblotting, and probing with Nrf2 antibody (Fig. 3A). The blot was stripped and reprobed with anti-LDH (cytosolic marker) and anti-Lamin B (nuclear marker) antibodies. The cytosolic fractions had undetectable amounts of Nrf2, therefore not shown. The results demonstrated that Nrf2 accumulated in the nucleus in a PP2 dose-dependent manner (Fig.  3A, lanes 1-3). Low and high dose LiCl treatment also led to dose-dependent nuclear accumulation of Nrf2 (Fig. 3A, compare lanes 1, 4, and 7). However, when LiCl was included with a low dose of PP2, Nrf2 followed PP2 dose-dependent nuclear accumulation (Fig. 3A, lanes 4 -6). Finally, a combination of higher doses of PP2 (1 M) with different doses of LiCl could not result in increased nuclear accumulation of Nrf2 as compared with PP2 alone (Fig. 3A, lanes 7-9). These results indicate that inclusion of LiCl with PP2 did not increase further the nuclear accumulation of Nrf2 over that observed with PP2 alone. In a related experiment, HepG2 cells were transfected with the NQO1 gene ARE-luciferase. The transfected cells were treated with different doses of LiCl or PP2 alone or in combinations and analyzed for luciferase activity (Fig. 3B). The results showed similar increase in luciferase activity with PP2 alone or in combination with LiCl. In other words, the LiCl had no effect on PP2 induction of ARE-luciferase gene expression. To further prove these observations, we used siRNA against both Fyn and GSK-3␤ alone or in combinations. In accordance with our earlier published results (13), ARE-luciferase activity increased after siRNA-mediated inhibition of Fyn in a dose-dependent manner (Fig. 3C). Similar results were obtained with GSK-3␤ siRNA. However, when both the siRNAs were used in combination, the AREluciferase activity followed the pattern as observed by inhibition of Fyn alone indicating that both these kinases affect Nrf2 by a similar pathway (Fig. 3C). Also, inhibition of both the kinases together did not show a further increase in Nrf2 protein stabilization as seen by Fyn siRNA alone (Fig. 3D). These results suggested that inhibition of GSK-3␤ and Fyn had a concerted and not an additive effect on nuclear accumulation of Nrf2. In other words, GSK-3␤ is upstream to Fyn in the control of nuclear export of Nrf2.
Next, we determined the effect of LiCl inhibition of GSK-3␤ on hydrogen peroxide-induced nuclear import and export of Nrf2 and Fyn (Fig. 4). The untransfected and Nrf2-V5-or FLAG-Fyn-transfected HepG2 cells were treated with hydrogen peroxide in the absence and presence of LiCl and were subcellularly fractionated. The distribution of Nrf2 and Fyn in nuclear and cytosolic fractions was analyzed by immunoblotting. Hydrogen peroxide treatment led to import of Nrf2 in the nucleus at 1 h and export of Nrf2 at 4 h of treatment (Fig. 4A). The nuclear accumulation of Nrf2 is an early response of the Nrf2-INrf2 complex to hydrogen peroxide and was expected as reported earlier (11,13). Hydrogen peroxide treatment at 1 h released Nrf2 from INrf2. Nrf2 translocated to the nucleus resulting in increased nuclear Nrf2. Nrf2 was exported out of the nucleus at 4 h after hydrogen peroxide treatment as a result of delayed response to oxidative stress (13). Inclusion of LiCl, an inhibitor of GSK-3␤, blocked the nuclear loss of Nrf2 at 4 h of hydrogen peroxide treatment. Interestingly, Fyn response to hydrogen peroxide was opposite that of Nrf2. Endogenous Fyn protein levels were reduced in the nucleus at 1 h but increased at 4 h after hydrogen peroxide treatment of HepG2 cells (Fig. 4A). The LiCl-mediated inhibition of GSK-3␤ reduced the nuclear gain of Fyn at 4 h after hydrogen peroxide treatment (Fig. 4A). The exogenously expressed Nrf2-V5 protein showed similar localization patterns as observed with endogenous Nrf2 (compare Fig. 4B and Fig. 2B, Nuc. Ext panels). Nrf2-V5 accumulated in the nucleus in a LiCl dose-dependent manner (Fig. 4B, left panel). However, the nuclear export-deficient mutant Nrf2Y568A-V5 showed accumulation in the nucleus even in the absence of LiCl. This is because of loss of tyrosine 568 phosphorylation and the absence of nuclear export of mutant protein (13). Treatment with LiCl had no effect on nuclear levels of Nrf2Y568A-V5 (Fig. 4B, right  panel). Hydrogen peroxide treatment in the absence and presence of LiCl also showed similar results for exogenously expressed FLAG-Fyn as observed for endogenous Fyn (compare Fig. 4, C with A, nuclear panels). FLAG-Fyn depleted from the nucleus at 1 h post hydrogen peroxide treatment which was followed by accumulation in the nucleus at 4 h of hydrogen peroxide exposure (Fig. 4C). The hydrogen peroxide mediated nuclear accumulation of FLAG-Fyn at 4 h was blocked in presence of LiCl (Fig. 4C, Nuc. Ext panel, compare lanes 1-3 and 4  and 5). These results revealed that the treatment of HepG2 cells with hydrogen peroxide leads to nuclear loss of Fyn and nuclear import of Nrf2 as an early response for activation of genes encoding chemoprotective proteins. Delayed/late response to hydrogen peroxide is to accumulate Fyn in the nucleus to phosphorylate Nrf2Y568 and activate the nuclear export of Nrf2. The results also revealed that inhibition of GSK-3␤ with LiCl led to the loss of Fyn accumulation in the nucleus and activated nuclear accumulation of Nrf2. Taken together, these results suggested that GSK-3␤ is upstream to Fyn as accumulation of Fyn in the nucleus was dependent on GSK-3␤ activity. The hydrogen peroxide-induced nuclear accumulation of Fyn at 4 h after treatment is likely because of nuclear translocation of Fyn after phosphorylation by GSK-3␤.
The above results raised an interesting question regarding the mechanism of GSK-3␤ regulation of Fyn? Additional experiments were performed to address this question. Phosphorylation of GSK-3␤Y216 is known to activate GSK-3␤ (16). We performed experiments to test if hydrogen peroxide activated GSK-3␤? HepG2 cells were treated with hydrogen peroxide and analyzed by immunoblotting with anti-phospho-GSK-3␤-pY216 antibody (Fig. 5A, left panel). Hydrogen peroxide induced GSK-3␤Y216 phosphorylation within 1 h of the treatment. The analysis showed that hydrogen peroxide also phosphorylated GSK3␣ at tyrosine 279 (Fig. 5A, left  panel). The anti-phospho-GSK3␤-pY216 antibody is known to cross-react with phospho-GSK3␣-pY279 (16). LiCl and PP2 failed to inhibit hydrogen peroxide-induced tyrosine phosphorylation of GSK-3␤ and GSK-3␣ (Fig. 5A, left, panel). However, AG18 and genistein, two general tyrosine kinase inhibitors, significantly reduced phosphorylation of GSK-3␤-pY216 (Fig. 5A,  right panel, compare lanes 2, 4,  and 6). The inhibition of GSK-3␤ activation in the presence of general tyrosine kinase inhibitors also revealed that Fyn did not phosphorylate GSK-3␤ in response to hydrogen peroxide because PP2, an inhibitor of Fyn, failed to block hydrogen peroxide-induced GSK-3␤Y216 phosphorylation. The results did suggest that unknown tyrosine kinase(s) phosphorylate GSK-3␤-Y216 in response to hydrogen peroxide that led to activation of GSK-3␤. Further experiments were performed to confirm that GSK-3␤ is upstream to Fyn and regulated Fyn in response to hydrogen peroxide. We tested the hypothesis that hydrogen peroxide activated GSK-3␤ which then phosphorylated Fyn resulting in its nuclear localization. Inside the nucleus, Fyn phosphorylates Nrf2Y568 and results in nuclear export of Nrf2. HepG2 cells were transfected with FLAG-Fyn, treated with hydrogen peroxide in the absence and presence of GSK3␤ inhibitor LiCl or Fyn inhibitor PP2. Nuclear extract was prepared and used for analyzing the phosphorylation status of Fyn by immunoprecipitation with anti-Thr(P) antibody followed by immunoblotting with anti-FLAG antibody (Fig. 5B, top left panel). In a similar experiment, the nuclear extract was analyzed by immunoblotting and probing with anti-FLAG antibody (Fig.  5B, right panel). The results revealed that hydrogen peroxide indeed induced time-dependent phosphorylation of Fyn at threonine residue(s) leading to nuclear accumulation of Fyn (Fig. 5, left and right panels). We also checked the phosphorylation of Fyn at serine residues by immunoprecipitating with anti-FLAG and immunoblotting with anti-Ser(P) antibody (Fig. 5B, bottom left panel). The results indicate that Fyn is not phosphorylated at serine residues in response to hydrogen peroxide. Untreated cells also showed a low amount of phosphorylated Fyn in the nucleus. This is presumably because of endogenous cellular stresses that might have led to Fyn phosphorylation and nuclear accumulation. The results also revealed that almost all of the Fyn translocated in the nucleus is phosphorylated. The treatment with LiCl blocked hydrogen peroxide-induced phosphorylation and nuclear accumulation of FLAG-Fyn, which was in agreement with GSK-3␤ being upstream to Fyn. However, PP2 had no effect on phosphorylation status or nuclear accumu- In a similar experiment, the cells were treated with H 2 O 2 in the presence or absence of AG18 or genistein for 1 h (right panel). In both the cases, cells were lysed and total cell lysate was analyzed by Western blotting (WB) and probing with anti-(Tyr(P))-GSK3, anti-GSK-3␤, and anti-␤-actin antibodies. B and C, phosphorylation and localization of FLAG-Fyn. HepG2 cells in 100-mm plates were transfected with 1.0 g of FLAG-Fyn. 24 h after transfection cells were treated as in A, left panel. Cells were harvested, lysed, or subjected to subcellular fractionation to prepare nuclear extracts (Nuc Ext). For analyzing phosphorylation of FLAG-Fyn, 1 mg of nuclear extract was used to immunoprecipitate (IP) with either anti-Thr(P) or anti-FLAG antibody, and the immune complexes were immunoblotted with anti-FLAG-HRP or anti-Ser(P) antibodies, respectively. 50 g of nuclear extract (B, right panel) or total cell lysate (C) were analyzed by Western blotting and probing with anti-FLAG-HRP and anti-LaminB or anti-␤-actin antibodies.
lation of Fyn (Fig. 5B, compare ϩLiCl with ϩPP2). In a related experiment, immunoblot analysis of total cell lysate revealed similar expression of FLAG-Fyn in untreated and LiCl-treated cells (Fig. 5C). Increased nuclear FLAG-Fyn in ϩPP2 panels is not because of more nuclear accumulation but because of more overall FLAG-Fyn expression. The total cell lysates showed more FLAG-Fyn expression in the ϩPP2 panels (Fig. 5C). This indicated that the alteration in phosphorylation and nuclear localization of Fyn is not because of alterations in the expression of FLAG-Fyn plasmid in HepG2 cells. The combined results revealed that hydrogen peroxide induced GSK-3␤Y216 phosphorylation leading to activation of GSK-3␤. The activated GSK-3␤ phosphorylated Fyn at threonine residue(s). Thr-phosphorylated Fyn accumulated in the nucleus and phosphorylated Nrf2Y568 leading to nuclear export of Nrf2 (Figs. 5 and 4A). To further analyze GSK-3␤-mediated Fyn and Nrf2 phosphorylation, we employed in vitro kinase assays using purified Nrf2, GSK-3␤, and Fyn proteins (Fig. 6). In the initial experiment, GSK-3␤ was incubated with a 3-kDa GSK3 peptide, and in vitro kinase reaction was performed to test the activity of GSK-3␤. GSK-3␤ successfully phosphorylated GSK peptide in vitro in a kinase dose-dependent manner (Fig. 6A). Bacterially purified Nrf2 was then incubated with active GSK-3␤ in the presence or absence of LiCl (Fig. 6B, lanes 3 and 4). The Nrf2 phosphorylation by GSK-3␤ was not detected (Fig. 6B, lanes 3 and  4, upper panel). This was despite the presence of both Nrf2 and GSK-3␤ proteins in the test reaction (Fig. 6B, middle and lower  panels, lanes 3 and 4). In the same experiment, phosphorylated Nrf2 was detected when purified Nrf2 was incubated with active Fyn kinase (Fig. 6B, lane 5). It is noteworthy that in in vitro kinase reactions GSK-3␤ and Fyn are autophosphorylated (Fig. 6B, upper  panel). In related experiments, GSK-3␤ and Fyn kinases were incubated individually and together in an assay reaction to test the GSK-3␤ phosphorylation of Fyn. Both GSK-3␤ and Fyn were autophosphorylated. The amount of Fyn autophosphorylation was in significantly high proportions so that addition of GSK-3␤ could not give any conclusive results.
We also performed experiments to investigate the role of nuclear export of Nrf2, i.e. to determine the fate of Nrf2 exported from the nucleus. We tested the hypothesis that Nrf2 exported out of the nucleus binds to INrf2, ubiquitinates, and degrades in the cytosol.
INrf2 is known to function as ubiquitin-protein isopeptide ligase and in association with Cul3 leads to ubiquitination and degradation of Nrf2 (18,19). Ubiquitination analysis of Nrf2 was performed in HepG2 cells transfected with Nrf2-V5 or the nuclear export-deficient mutant Nrf2Y568A-V5, HA-Ub, and FLAG-INrf2 in combinations as shown in Fig. 7. The transfected cells were treated with vehicle control, Me 2 SO or GSK-3␤ inhibitor, LiCl or proteasomal inhibitor, MG132. The cells were lysed and immunoblotted with either anti-V5 to detect Nrf2-V5 or Nrf2Y568A-V5, anti-FLAG antibody to detect FLAG-INrf2, or anti-␤-actin antibody to demonstrate equal loading. The lysates were immunoprecipitated with anti-V5 antibody and immunoblotted with anti-HA antibody to detect ubiquitination of Nrf2-V5 and Nrf2Y568A-V5. The results are shown in Fig. 7. The co-transfection of Nrf2-V5 with HA-Ub showed ubiquitination and degradation of Nrf2-V5 (Fig. 7, left panel). Inclusion of INrf2 in the transfection increased ubiquitination and degradation of Nrf2-V5 (Fig. 7, left panel, Me 2 SO lanes). The pretreatment of transfected cells with LiCl significantly reduced ubiquitination and degradation of Nrf2-V5 suggesting the role of the nuclear export in degradation of Nrf2. The treatment with the proteasome inhibitor MG132 increased ubiquitination without degradation of Nrf2-V5. Interestingly, the nuclear export-deficient mutant Nrf2Y568A-V5 showed very little ubiquitination in the absence or presence of GSK3␤ inhibitor LiCl (Fig. 7B, right panel). In other words, LiCl-mediated inhibition of GSK3␤ activity had no effect on ubiquitination of Nrf2Y568A-V5. However, the nuclear export-deficient mutant Nrf2Y568A protein degraded efficiently in the presence of INrf2. The treatment of cells with the proteasome inhibitor MG132 showed ubiquitination of Nrf2Y568A without degradation. The combined results suggested that the Fyn-phosphorylated Nrf2 protein is exported out of nucleus to bind to INrf2, ubiquitinate, and degrade. The nuclear export-deficient mutant Nrf2Y568 protein binds to INrf2, ubiquitinates, and degrades inside the nucleus. INrf2 is predominantly shown to be present in the cytosol (5,6). However, recent studies have shown that some of the INrf2 is also present in the nucleus (20,21).

DISCUSSION
The regulation of Nrf2, especially its abundance in the nucleus, is extremely important for controlling the normal and induced expression of a battery of chemoprotective genes (1). Exposure of cells to chemical stress leads to the release of Nrf2 from its cytosolic inhibitor INrf2 as an early cellular response to chemical stress (Fig.  8). The release of Nrf2 from INrf2 is mediated by PKC and/or cysteine modification of INrf2 (1). A bipartite nuclear localization signal directs Nrf2 to the nucleus (11). Nrf2 after translocation in the nucleus heterodimerizes with known (Jun and small Maf) and unknown nuclear factors, binds to ARE, and activates gene transcription. The increase in expression of chemoprotective genes neutralizes the chemical stress. Because persistent increase in chemoprotective genes expression threatens cell survival, Nrf2 is exported out of the nucleus and degraded. The nuclear export of Nrf2 is delayed/late response of cells to oxidative/electrophilic stress. A leucine-rich nuclear export signal at the C terminus of Nrf2 has been characterized (11). However, the nuclear export signal in Nrf2 is activated only after Fyn is accumulated inside the nucleus that phosphorylates tyrosine 568 of Nrf2 (13). The phosphorylated Nrf2Y568 binds to Crm1 and is exported out of the nucleus (13). The studies in the present report demonstrate that GSK-3␤ is upstream to Fyn in regulation of nuclear export of Nrf2. Phosphorylation status of GSK-3␤ regulates its activity (16). GSK-3␤ phosphorylated at a serine 9 residue via PKC or other similar enzymes is inactive. Activation of GSK-3␤ is mediated by phosphorylation at tyrosine 216 residue and/or de-phosphorylation of serine 9 (16). The hydrogen peroxide in our studies induced tyrosine 216 phosphorylation of GSK-3␤ resulting in its activation. The activated GSK-3␤ phosphorylated Fyn at threonine residue(s) leading to nuclear localization of Fyn. Fyn phosphorylates tyrosine 568 of Nrf2 (13). Phosphorylated Nrf2Y568 is exported out of the nucleus, ubiquitinated, and degraded. Recently, GSK3␤ was shown to phosphorylate Nrf2 at FIGURE 7. Ubiquitination and degradation of Nrf2. HepG2 cells were co-transfected either with pcDNA-Nrf2-V5/pcDNA-Nrf2Y568A-V5, pCMV-HA-Ub, or pCMV-FLAG-mINrf2 in combinations as shown. The transfected cells were treated with either Me 2 SO or GSK3␤ inhibitor LiCl or proteasomal inhibitor MG132 for 5 h. The cells were lysed in RIPA buffer, and 100 g of protein was analyzed by SDS-PAGE and immunoblotted with anti-V5, anti-FLAG, and ␤-actin antibodies. 500 g of protein was immunoprecipitated (IP) with anti-V5 antibody and immunoblotted (WB) with anti-HA antibody to detect ubiquitinated Nrf2 or Nrf2Y568A-V5. unknown residues with implications in nuclear export of Nrf2 (17). However, our in vitro kinase assay data did not show Nrf2 phosphorylation by GSK-3␤. Also, our data reveal that GSK3␤ is upstream to Fyn, which phosphorylates Tyr-568 of Nrf2 that regulated nuclear export of Nrf2. The reasons and significance of this difference remain unknown.
It is noteworthy that PKC is also known to inactivate GSK-3␤ (16). The decrease in nuclear Fyn at 1 h after hydrogen peroxide treatment in Fig. 4A is presumably due to PKCmediated inactivation of GSK-3␤; however, this remains to be determined. Therefore, the inactivation of GSK-3␤ and activation of Nrf2 via serine phosphorylation are both regulated by PKC. Together these mechanisms might work in harmonization leading to a synergistic action to early response to oxidative stress resulting in nuclear accumulation of Nrf2 both by activating import and blocking the export of Nrf2. The signaling events between hydrogen peroxide and tyrosine 216 phosphorylation of GSK-3␤ remain unknown. In addition, the tyrosine kinase(s) that phosphorylate GSK-3␤Y216 also remains unknown. The signaling events between hydrogen peroxide and phosphorylation of GSK-3␤ might involve phosphatidylinositol 3-kinase, AKT, PP2A, and PKC that could inactivate or activate GSK-3␤ during early and delayed/later events (22).
In conclusion, we have investigated the Fyn upstream signaling in the regulation of nuclear export of Nrf2 and the regulation of expression of chemoprotective proteins in response to oxidative/electrophilic stress. We demonstrate that GSK-3␤ is upstream to Fyn. The chemical stress induced phosphorylation of GSK-3␤Y216 and activated GSK-3␤. The activated GSK-3␤ phosphorylated Fyn. The phosphorylation of Fyn leads to nuclear accumulation of Fyn and phosphorylation of Nrf2 resulting in nuclear export, ubiquitination, and degradation of Nrf2. This study dissects the mechanism involved in Nrf2 regulation via Fyn kinase as a result of delayed response to oxidative stress.