Phosphorylation and Dephosphorylation of Tyrosine 141 Regulate Stability and Degradation of INrf2

INrf2-Nrf2 proteins are sensors of chemical/radiation stress. Nrf2, in response to stresses, is released from INrf2. Nrf2 is translocated into the nucleus where it binds to the antioxidant response element and coordinately activates the expression of a battery of genes that protect cells against oxidative and electrophilic stress. An autoregulatory loop between INrf2 and Nrf2 regulates their cellular abundance. Nrf2 activates INrf2 gene expression, and INrf2 serves as an adapter for degradation of Nrf2. In this report, we demonstrate that mutation of tyrosine 141 in bric-a-bric, tramtrack, broad complex domain to alanine rendered INrf2 unstable and nonfunctional. INrf2Y141A mutant degraded rapidly as compared with wild type INrf2, although it could dimerize and bind Nrf2. De novo synthesized INrf2 protein was phosphorylated at tyrosine 141. Tyrosine 141-phosphorylated INrf2 was highly stable. Treatment with hydrogen peroxide, which is an oxidizing agent, led to dephosphorylation of INrf2Y141, resulting in rapid degradation of INrf2. This resulted in stabilization of Nrf2 and activation of ARE-mediated gene expression. These results demonstrate that stress-induced dephosphorylation of tyrosine 141 is a novel mechanism in Nrf2 activation and cellular protection.

tive stress, GSK-3␤ acts upstream to Fyn kinase in regulating the nuclear export of Nrf2 to abrogate the induction of AREcontaining genes (14).
In addition to post-translational modification in Nrf2 resulting in ARE induction, several crucial residues in INrf2 have also been proposed to be important for this activation. Homodimerization of INrf2 is crucial for retaining Nrf2 in the cytoplasm (15). Serine 104 of INrf2 is essential for dimerization, suggesting that disruption of the INrf2 dimer is associated with release of Nrf2 (15). Thus, stress-mediated modification of this serine residue might be one important mechanism of ARE gene induction. The high cysteine content of INrf2 and the ability of phase II inducers to modify sulfhydryl groups by alkylation, oxidation, and reduction suggest that INrf2 would be an excellent candidate as an oxidative stress sensor. Studies based on the electrophile-mediated modification, location, and mutational analyses revealed that three cysteine residues, Cys 151 , Cys 273 , and Cys 288 , are crucial for INrf2 activity (16). INrf2 itself undergoes ubiquitination by the Cul3 complex, which was markedly increased in response to phase II inducers such as t-butylhydroquinone (17). It has been suggested that normally INrf2 targets Nrf2 for ubiquitine-mediated degradation, but electrophiles may trigger a switch of Cul3-dependent ubiquitination from Nrf2 to INrf2, resulting in ARE gene induction. Recently, INrf2 is also proposed to shuttle between cytoplasm and nucleus, but the reason for the presence of INrf2 in the nucleus is not yet known (18,19).
In this study, we demonstrate a novel mechanism of regulation of stability/degradation of INrf2 and release of Nrf2. We show that INrf2 protein phosphorylated at tyrosine 141 is highly stable. Tyrosine 141-phosphorylated INrf2 formed homodimers, bound to Nrf2, and actively participated in ubiquitination and degradation of Nrf2. Mutation of tyrosine 141 to alanine led to rapid degradation of INrf2 and up-regulation of Nrf2 downstream gene expression. Treatment of cells with hydrogen peroxide led to a significant decrease in phosphorylation of Tyr 141 and rapid degradation of INrf2. This resulted in stabilization of Nrf2 and up-regulation of Nrf2 downstream gene expression. These results led to the conclusion that phosphorylation and dephosphorylation of tyrosine 141 control the stability/degradation of INrf2. In addition, the results reveal that oxidant-induced dephosphorylation of INrf2 tyrosine 141 leads to dephosphorylation and degradation of INrf2. This results in stabilization of Nrf2 and activation of ARE-mediated gene expression.
Cell Culture, Co-transfection of Expression Plasmids, Luciferase Reporter Assay, and NQO1 Gene Expression Analysis-Hepa1 and Hep-G2 cells were grown in 6-well plates and 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 along with 0.5 g of plasmids encoding either pcDNA or INrf2-V5 or  INrf2Y141A-V5 or INrf2Y208A-V5 or INrf2Y141A-Y208A-V5 double mutant. Renilla luciferase was used as the internal control in each transfection. Transfections were done using the Effectene transfection reagent (Qiagen). 36 h after transfection, the cells were washed with 1ϫ PBS and lysed in 1ϫ passive lysis buffer from a dual luciferase reporter assay system kit (Promega, Madison, WI). The luciferase activity was measured using the procedures described previously (12).
In related experiment, Hepa-1 cells were transfected with 0.5 g of pcDNA or INrf2 or INrf2Y141A-V5 or INrf2Y208A-V5  or INrf2Y141A-Y208A-V5. 0.05 g of Renilla luciferase was included as a control of transfection efficiency. The cells were harvested 36 h after transfection and analyzed by SDS-PAGE, Western blotting, and probing with anti-NQO1 antibody. Immunoprecipitation and Phosphorylation of Endogenous INrf2-The cells either transfected or treated for appropriate times were washed twice with ice-cold PBS and harvested. The total cell lysates were prepared in RIPA buffer as described above. Five hundred micrograms of extract was used for immunoprecipitation. Briefly, lysate was incubated with either anti-V5 antibody (Invitrogen) or anti-phosphotyrosine (anti-Tyr(P)) antibody (clone 4G10; Upstate Biotechnology, Lake Placid, NY) in RIPA binding buffer 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 Inc., Piscataway, NJ) were added and incubated for 1 h at 4°C with shaking. The slurry was centrifuged at 10,000 rpm for 30 s, and supernatant was discarded. The beads were washed twice with RIPA buffer. 25 l of SDS sample dye was added and boiled, and immunoprecipitates were resolved on a 10% SDS-PAGE followed by immunoblotting with anti-V5HRP, anti-Tyr(P), or anti-FLAG-HRP antibodies. FLAG immunoprecipitation was done using the anti-FLAG-M2 agarose beads (Sigma).
In related experiments, 1.0 mg of Hepa-1 cell total lysate was immunoprecipitated with either IgG (control) or mouse anti-Nrf2 antibody (Santa Cruz, CA). The immunoprecipitate was analyzed by Western blotting and probing with anti-phosphotyrosine antibody. In reverse IP, Hepa-1 total cell lysate was immunoprecipitated with anti-phosphotyrosine antibody followed by Western analysis with anti-INrf2 antibody.
INrf2 Phosphorylation Analysis in Transfected Cells-Hepa-1 cells transfected with INrf2-V5 or INrf2Y1418A-V5 or with INrf2Y208A-V5 or INrf2Y141A-Y208A-V5 double mutant or treated with cycloheximide ϩ H 2 O 2 Ϯ genestein were lysed in RIPA buffer supplemented with tyrosine phosphatase inhibitor mixture and protease inhibitor mixture. 1 mg of total cell lysate was used to immunoprecipitate with anti-V5 or anti-Tyr(P) 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.
Degradation of INrf2-Hepa-1 cells were grown in 6-well tissue culture plates and were transfected with 0.5 g of either pcDNA-INrf2-V5 or pcDNA-INrf2Y141A-V5 plasmids. Twenty-four h after transfection, the cells were pretreated with either Me 2 SO or MG132 (20 M) for 8 h. The cells were washed twice with medium and treated with 30 g/ml cycloheximide for different time points (1, 3, 6, or 8 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 or fractionated to obtain cytosol and nuclear extracts. Biochemical fractionation of cells was done using a nuclear extract kit (Active Motif, Carlsbad, CA) following the manufacturer's protocol. 50 g of total cell lysate or cytosolic/nuclear extracts were resolved on a 10% SDS-PAGE, Western blotted, and probed with anti-V5-HRP, anti-lactate dehydrogenase, anti-LaminB, and anti-␤-actin antibodies.
Pulse-Chase Assay-Hepa-1 cells were transfected with INrf2-V5 or INrf2Y141A-V5. 24 h after transfection cells were incubated with methionine-deficient Dulbecco's modified Eagle's medium (Sigma) for 30 min. The cells were then labeled with methionine-deficient Dulbecco's modified Eagle's medium 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 (Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum), the cells were chased by normal culture medium supplemented with 100 g/ml L-methionine for 0, 1, 3, 6, and 8 h. MG132 (20 M) was added wherever indicated. The 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.
In Vitro Binding-The in vitro transcription/translation of the plasmids encoding INrf2-V5, INrf2Y141A-V5, and FLAG-Nrf2 were performed using the TNT-coupled rabbit reticulocyte lysate system (Promega). Redivue L-[ 35 S]methionine (Amersham Biosciences) substituted 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 and autoradiography. All of the in vitro transcribed/translated proteins gave the expected size bands. An in vitro binding assay was performed as described earlier (13). Briefly, 5 l of each in vitro translated protein (INrf2-V5ϩFLAG-Nrf2 or INrf2Y141A-V5ϩFLAG-Nrf2) 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 the addition of 2.5 g of anti-V5 antibody and sufficient protein binding buffer to make the volume 100 l and incubated the mixture overnight at 4°C with shaking. After incubation, 40 l of washed protein A beads (Santa Cruz Biotechnology, Santa Cruz, CA) 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 the protein binding buffer. In a similar binding experiment, the protein mixtures were immunoprecipitated with anti-FLAG-M2 beads (Sigma). Finally, the beads were boiled in SDS sample dye and analyzed by SDS-PAGE as described above.
In Vitro Dimerization-The in vitro transcription/translation of the plasmids encoding INrf2-V5 and INrf2Y141A-V5 was performed as described above. 5 l of each in vitro translated protein was incubated with 0.005, 0.01, or 0.02% glutaraldehyde for 30 min at room temperature. The reaction was terminated by adding SDS sample dye. The samples were resolved on SDS-PAGE and autoradiographed.
In Vivo Dimerization-Hepa-1 cells were cultured in 6-well plates and transfected with 0.5 g of INrf2-V5 or INrf2Y141A-V5. 24 h after transfection the cells were washed and incubated with fresh medium containing 0.005, 0.01, or 0.02% glutaraldehyde at room temperature for 30 min. The cells were washed twice with PBS and lysed with RIPA buffer as described above. 50 g of total cell lysate was immunoblotted with anti-V5-HRP and anti-actin antibodies.

Tyrosine 141 Mutation Renders INrf2 Protein Unstable-
Mouse INrf2 protein consists of five discrete domains (Fig. 1A). These include the N-terminal region, the bric-a-bric, tramtrack, broad complex (BTB), the intervening region, the diglycine repeats, and the C-terminal regions. Several cysteine residues have been identified in the BTB domain that are crucial for dimerization and function of INrf2. The amino acid sequence of mouse INrf2 protein was aligned with human, rat, Xenopus, and zebrafish INrf2 sequences to identify conserved domains of putative significance. Two putative tyrosine phos-phorylation sites (Tyr 141 and Tyr 208 ) that are conserved among the various species were identified and are shown in Fig. 1B.
Recombinant plasmids were generated that upon transfection in Hepa-1 cells expressed V5-tagged INrf2 and mutants   INrf2Y141A, INrf2Y208A, and double mutant INrf2Y141A-Y208A.  Western analysis of Hepa-1 cells  transfected with INrf2-V5 and  mutant-V5  plasmids  revealed  expression of comparable amounts  of INrf2-V5 and mutant INrf2-Y208A-V5 proteins ( Fig. 2A, lanes 2  and 4). Interestingly, the transfection of mutants INrf2Y141A-V5 and INrf2DM-V5 plasmids showed significantly lower amounts of proteins (Fig. 2, lanes 3 and 5). These results indicated that mutation of tyrosine 141 causes INrf2 protein to degrade because the protein stabilized after inhibition of protein degradation by MG132. Next, Hepa-1 cells were transfected with varying concentrations of plasmids encoding wild type INrf2 and mutant INrf2Y141A to determine the fold difference in the stability of wild type and mutant INrf2 proteins. The analysis revealed that 0.05 g of wild type INrf2 plasmid produced an equivalent amount of protein to that observed in cells transfected with 0.5 g of mutant INrf2Y141A plasmid (Fig. 2B). This indicated a 10-fold difference in the stability of two proteins (Fig. 2B). The plasmids encoding wild type and mutant proteins were in vitro transcribed and translated in presence of protease inhibitors (Fig. 2C). All four plasmids transcribed and translated equivalent amounts of wild type and mutant proteins (Fig.  2C). In other words, wild type and mutant plasmids upon transcription and translation in an in vitro cell-free system produced similar amounts of the respective proteins (Fig. 2C). This indicated that all four plasmids encoding wild type and mutant INrf2 proteins are Hepa-1 total cell lysate was immunoprecipitated with mouse anti-INrf2 antibody followed by Western analysis with antiphosphotyrosine antibody. In reverse IP, Hepa-1 lysate was immunoprecipitated with anti-phosphotyrosine antibody and probed with anti-INrf2 antibody. Both immunoprecipitation with anti-INrf2 antibody and reverse IP analysis revealed that endogenous INrf2 protein is phosphorylated at tyrosine residue (Fig. 2D). In a related experiment, Hepa-1 cells expressing V5-tagged wild type and mutant INrf2 proteins were lysed and immunoprecipitated with either anti-V5 followed by Western analysis with anti-V5 and anti-Tyr(P) antibody. Reverse immunoprecipitation with anti-Tyr(P) antibody was also performed followed by Western blot with anti-V5 antibody. The results revealed that wild type and mutant INrf2Y208A proteins were phosphorylated at tyrosine residues (Fig. 2E). However, tyrosine phosphorylation was absent in mutant INrf2Y141A and INrf2DM bearing Y141A mutation (Fig. 2E). These results suggested that phosphorylation of tyrosine 141 stabilizes INrf2 protein.
In related experiment, we analyzed the effects of wild type and mutant INrf2 proteins on ARE-luciferase expression in Hepa-1 and HepG2 cells (Fig. 2F) 3A). Similar results were also obtained in a pulse-chase experiment involving radiolabeled methionine (Fig. 3B). Contrary to the stability of wild type INrf2, a majority of mutant INrf2Y141A degraded within 8 h of cycloheximide treatment. These results indicated that mutation of tyrosine 141 causes the INrf2 to degrade rapidly. Interestingly, further experiments revealed rapid degradation of mutant INrf2Y141A in both cytosolic and nuclear compartments (Fig. 3C). But the wild type INrf2 protein was stable in both cytosolic and nuclear compartments. These results together indicated that INrf2Y141A degrades much faster than wild type INrf2 irrespective of subcellular distribution.   (Fig. 5A, middle panel). Similarly, in reverse immunoprecipitation experiment, anti-FLAG antibody immunoprecipitated FLAG-Nrf2, and INrf2-V5 and INrf2Y141A-V5 were co-precipitated in the complex (Fig. 5A, right panel). In related experiment, Hepa-1 cells were co-transfected with INrf2-V5 or mutant INrf2Y141A-V5 with FLAG-Nrf2, and total cell lysates were immunoprecipitated with anti-V5 or anti-FLAG antibodies in separate experiments (Fig. 5B). Next, we compared the capacities of wild type INrf2 and mutant INrf2Y141A to degrade Nrf2 (Fig. 5C). Hepa-1 cells were co-transfected with 0.5 g of FLAG-Nrf2 and increasing amounts of INrf2-V5 (Fig. 5C, left panel) or INrf2Y141A-V5 (Fig. 5C, right panel) (Fig. 6A,  left panel) or hydrogen peroxide and cycloheximide in the absence (Fig.  6A, middle panel) or presence (Fig.  6A, right panel) of the tyrosine kinase inhibitor genestein for different time intervals. The amount of INrf2-V5 was analyzed by Western blotting of total cell lysates obtained from these cells as shown in Fig. 6. INrf2-V5 protein more or less did not degrade during 8 h of cycloheximide treatment, indicating that INrf2-V5 is a stable protein (Fig. 6A,  left panel). Interestingly, the treatment with hydrogen peroxide resulted in degradation of INrf2 (Fig. 6A, middle panel). More than 50% of INrf2 protein was degraded at 6 h after cycloheximide and hydrogen peroxide treatment, as compared with less than 10% degradation in cells treated with cyclo-    Fig. 6A with Fig. 3A). In a related experiment, tyrosine phosphorylation of INrf2 remained more or less unchanged during 8 h of cycloheximide treatment (Fig. 6B, left panel). Interestingly, the tyrosine phosphorylation of INrf2 was reduced in the presence of hydrogen peroxide (Fig. 6B, right panel). In the same experiment, endogenous INrf2 showed similar results on hydrogen peroxide-induced dephosphorylation of INrf2 (data not shown). INrf2 modification under oxidative stress is very rapid because most of the INrf2 was dephosphorylated within 3 h of hydrogen peroxide treatment (Fig. 6B, right panel). In a similar experiment, INrf2Y141A mutant failed to demonstrate tyrosine phosphorylation in the absence or presence of hydrogen peroxide (data not shown). The above results suggested that mutation of Tyr 141 and inhibition of INrf2 phosphorylation both render the INrf2 protein susceptible to degradation. The rapid degradation of INrf2 in presence of hydrogen peroxide implies involvement of an unknown tyrosine phosphatase that can cause activation of Nrf2 by destabilizing INrf2 during the initial response to oxidative stress.
teines modification or modification of cysteine is followed by dephosphorylation of INrf2Y141. In either case, the INrf2 modification(s) leads to degradation of INrf2 and stability of Nrf2. The PKC modification of Nrf2 might take place at the same time as tyrosine dephosphorylation or following tyrosine dephosphorylation and remains to be determined. The above events result in nuclear translocation of Nrf2 and activation of ARE-mediated gene expression. The tyrosine kinase that phosphorylates tyrosine 141 of INrf2 is not known. The protein phosphatase that can act on phospho-INrf2 to remove the phospho group also remains unknown. Based on our data, it is suggested that the phosphatase involved in this pathway might be the one that can respond to oxidative/electrophilic stress. A model depicting phosphorylation/dephosphorylation control of INrf2 is proposed in Fig. 7. Under normal conditions, de novo synthesized INrf2 is phosphorylated at tyrosine 141 by an unknown kinase and forms homodimers at the BTB domain. INrf2-INrf2 homodimers then hold Nrf2 in cytoplasm and by interacting with Cul3 ubiquitine ligase initiates ubiquitination and degradation of Nrf2. In response to oxidative/electrophilic stress, INrf2 and Nrf2 are modified in either independently or sequentially in a single mechanism. INrf2 modifications include electrophiles binding with cysteines and tyrosine 141 dephosphorylation, and Nrf2 modification is phosphorylation of serine 40 by PKC. The modification of cysteines presumably initiates structural alterations in INrf2 followed by tyrosine 141 modifications. Tyrosine 141 dephosphorylation leads to degradation of INrf2. PKC-mediated serine 40 phosphorylation in Nrf2 releases Nrf2 from INrf2. These events stabilize Nrf2. Nrf2 translocates inside the nucleus, binds to the ARE, and up-regulates the expression of chemoprotective and defensive genes.
In conclusion, we demonstrate that phosphorylation of tyrosine 141 of INrf2 regulates its stability and function. Tyrosine 141-phosphorylated INrf2 is stable and can dimerize and unbiquitinate and degrade Nrf2. Oxidative/electrophilic stressmediated elimination of the phosphate group from tyrosine 141 in INrf2 results in instability and loss of function of INrf2. The resultant dephosphorylated INrf2 can dimerize but is highly unstable and nonfunctional. Taken together, these results demonstrate the presence of a unique model of Nrf2 activation via de-activation of Nrf2 inhibitor, INrf2 by post-translational modifications.