Antioxidant-induced modification of INrf2 cysteine 151 and PKC-δ-mediated phosphorylation of Nrf2 serine 40 are both required for stabilization and nuclear translocation of Nrf2 and increased drug resistance

Nuclear factor erythroid 2-related factor 2 (Nrf2) binds to antioxidant response element (ARE) and regulates expression and coordinated induction of a battery of genes encoding defensive proteins (reviewed by Jaiswal, 2004). This induction is a mechanism essential for cellular protection against oxidative stress and neoplasia. Nrf2-null mice are viable and live to adulthood, showing that Nrf2 is not required for erythropoiesis, development or growth (Chan et al., 1996). Nrf2-null mice express significantly lower levels and no induction of chemopreventative proteins, and demonstrate slower wound healing and emphysema in response to tobacco smoke (Braun et al., 2002; Rangasamy et al., 2005). Nrf2-null mice are more sensitive to chemical-induced carcinogenesis (Hu et al., 2006; Ramos-Gomez et al., 2001; Ma et al., 2006). In addition, the accumulation of Nrf2 in the nucleus provides protection to transforming cells and assists in development of lung cancer (Padmanabhan et al., 2006).

Nrf2 is retained in the cytoplasm by its inhibitor, INrf2 [inhibitor of Nrf2; also known as KEAP1 (Kelch-like ECH-associated protein 1)] (Dhakshinamoorthy and Jaiswal, 2001; Itoh et al., 1999). INrf2 controls the ubiquitylation of Nrf2 through the Cullin 3 (Cul3)-dependent ubiquitin ligase (E3) (Kobayashi et al., 2004; Cullinan et al., 2004; Zhang et al., 2004). The antioxidants and xenobiotics lead to dissociation of Nrf2 from INrf2 resulting in stabilization and nuclear translocation of Nrf2. Nrf2 binds to ARE and induces the expression of defensive genes. Several proteins, including Bach1 and Fyn/GSK3β, are thought to shut down the activation of defensive genes once the stress signal subsides as a result of late response to antioxidants (Dhakshinamoorthy et al., 2005; Karapetian et al., 2005; Velichkova and Hasson, 2005; Jain et al., 2005; Jain and Jaiswal, 2006; Jain and Jaiswal, 2007; Niture and Jaiswal, 2009).

Controversy exists as to whether it is stress-induced INrf2 modification or Nrf2 phosphorylation that controls release of Nrf2 from INrf2, the stabilization and nuclear translocation of Nrf2 and activation of gene expression (Jaiswal, 2004). Several studies have pointed towards INrf2 being the oxidative stress sensor. Modification of the cysteines in INrf2 is postulated to directly cause ARE activation (Dinkova-Kostova et al., 2001). It has also been proposed, as a result of mutational analysis, that C273 and C288 of INrf2 are required for its Nrf2 ubiquitylation activity and might act as a sensor to release Nrf2 from the complex (Zhang et al., 2003). More recently, it has been suggested that C151 of INrf2 is a highly reactive cysteine that can be easily oxidized or modified by electrophilic agents such as biotinylated iodoacetamide (BIA) (Eggler et al., 2007). However, this chemical modification is not sufficient for dissociation of Nrf2 from INrf2, but it induces the confirmational changes in the BTB domain of INrf2 that leads to weakening dimerization and destabilization of INrf2 (Eggler et al., 2005). By contrast, several reports demonstrated that Nrf2 is a substrate for modifications in response to oxidative stress. Two independent studies have demonstrated that antioxidant-induced PKC-mediated phosphorylation of serine 40 in Nrf2 leads to dissociation of Nrf2 from INrf2 (Bloom and Jaiswal, 2003; Huang et al., 2002).

Protein kinase C (PKC) has eleven isoforms (α, βI, βII, γ, δ, ϵ, η, θ, μ, ξ and ι), which are serine/threonine kinases (Newton, 1995; Newton, 2003). A growing body of evidence indicates that PKCs play divergent roles in controlling cell growth, differentiation, apoptosis and carcinogenesis (Nishizuka, 1995; Niwa et al., 2002; Dhakshinamoorthy and Jaiswal, 2001). The specific role of these isoenzymes in antioxidant induction of Nrf2 activation is not yet known.

Here, we demonstrate that PKC-δ is the major PKC isoform that phosphorylates Nrf2S40, which leads to stabilization and nuclear localization of Nrf2 and induction of cytoprotective/detoxifying phase2 genes. In addition, we demonstrate that antioxidant modification of INrf2C151 followed by phosphorylation of Nrf2S40 both mediate Nrf2 release from INrf2 and the stabilization and nuclear translocation of Nrf2. The accumulation of Nrf2 in the nucleus leads to induction of detoxifying/cytoprotective proteins, reduction of drug induced apoptosis, drug resistance and cell survival.

We used isoform-specific PKC inhibitors, Go6976, which inhibits α and βI isoforms, Go6850, which inhibits α, βI, βII and γ isoforms, and Go6983, which inhibits α, β and γ isoforms (IC₅₀ 7-10 nM) to identify the role of specific PKC isoform(s) in ARE activation. The treatment of Hep-G2 cells with the antioxidant tertiary butylhydroquinone (tBHQ) alone stabilized Nrf2, as expected (Fig. 1A). However, pre-treatment of Hep-G2 cells with the three PKC inhibitors failed to block tBHQ-mediated stabilization of Nrf2 (Fig. 1A). This suggested that, PKC-α, βI, βII and γ isoforms have no role in tBHQ stabilization of Nrf2. Pretreatments of cells with staurosporine (non-specific inhibitor of PKC) and rottlerin (a specific inhibitor of PKC-δ) in separate experiments inhibited tBHQ-mediated stabilization of Nrf2 (Fig. 1B). Staurosporine and rottlerin also inhibited tBHQ induction of human NQO1 ARE-mediated luciferase gene expression (Fig. 1C). We used dominant-negative mutants for each PKC isoform in ARE-luciferase assays, to determine the isoform-specific effect of PKC on Nrf2 transcriptional activity. The construction of PKC dominant-negative mutants, their expression in mammalian cells and effectiveness in inhibiting specific PKC isoforms have been previously described (Soh and Weinstein, 2003). We transiently transfected HA-tagged PKC dominant-negative mutants of PKC isoforms (α, βI, βII, γ, δ, ϵ, ξ, η, ι) and protein expression was analyzed by immunoblotting with anti-HA antibody. All dominant-negative mutants were expressed in Hep-G2 cells at comparable levels (Fig. 1D, upper panel). To test the effect of the dominant-negative mutant of each PKC isoform on human NQO1-ARE-luciferase activity, Hep-G2 cells were co-transfected with the reporter plasmids and plasmids expressing dominant-negative mutants of each PKC isoform. The transfected cells were then treated with tBHQ and cell extracts were analyzed for luciferase activity (Fig. 1D, lower panel). The results of the luciferase analysis revealed that the dominant-negative mutant against PKC-δ was the only one that significantly inhibited tBHQ-induced ARE-mediated luciferase activity (Fig. 1D, lower panel). Dominant-negative mutants against all other PKC isoforms had no significant effect on basal and tBHQ-induced ARE-mediated luciferase gene expression. These results suggested that PKC-δ mediates tBHQ stabilization of Nrf2.

Bacterially purified Nrf2 was used as a substrate for phosphorylation by the commercially available purified PKC isoforms in the presence of [³²P]ATP (Fig. 2A). In one set of reactions, we used only the PKC isoforms in a kinase reaction to show that the enzymes are active as they can be auto-phosphorylated (Fig. 2A lower panel). The reactions were then carried out in the absence or presence of Nrf2 and autoradiography was done to detect phosphorylated Nrf2. PKC isoforms βI, βII, ϵ and ξ did not phosphorylate Nrf2 (Fig. 2A upper panels). By contrast, PKC-α, γ and θ phosphorylated Nrf2 weakly. In the same experiment, PKC-δ and PKC-η demonstrated strong phosphorylation of Nrf2 (Fig. 2A). These results indicate that PKC-δ and PKC-η isoforms might be involved in phosphorylation of Nrf2.

The results from reporter assays demonstrated that dominant negative mutant against PKC-δ but not against PKC-η significantly inhibited basal and tBHQ-induced ARE-luciferase gene expression (Fig. 2B). Furthermore, we used siRNAs against PKC-δ and PKC-η to analyze their effect on ARE-luciferase expression and induction (Fig. 2C). Lamin-A/C siRNA was used as a control. Lamin-A/C siRNA decreased Lamin-A/C cellular protein but had no effect on ARE-luciferase gene expression and induction (Fig. 2C). siRNA against PKC-η also knocks down PKC-η protein levels, whereas, decrease in PKC-η had no significant effect on ARE expression and induction (Fig. 2C). By contrast, siRNA against PKC-δ resulted in a significant reduction in PKC-δ protein and inhibition of both basal and tBHQ-induced ARE-mediated expression (Fig. 2C). These results indicate that PKC-δ phosphorylates Nrf2 and mediates antioxidant induction of gene expression. Although, there was a strong phosphorylation of Nrf2 by PKC-η in vitro, there was no effect of PKC-η on Nrf2 activity in vivo. To examine, the role of PKC-δ and tBHQ in the stabilization and activation of Nrf2, we performed immunoprecipitation assays to analyze the interaction between Nrf2 and INrf2. Immunoprecipitation assays revealed that exposure of the cells to tBHQ completely dissociated Nrf2 from INrf2 (Fig. 2, lane 2), whereas, Nrf2 immunoprecipitated with INrf2 when cells were treated with DMSO or transfected with PKC-δ siRNA (Fig. 2, lanes 3 and 4). Interestingly, transfection with PKC-δ siRNA also interfered with tBHQ-induced Nrf2 release from INrf2 since a significant amount of Nrf2 was immunoprecipitated with INrf2 (Fig. 2, compare lane 2 with lane 5). These results demonstrated that tBHQ and PKC-δ are both required for releasing Nrf2 from INrf2 and for Nrf2 stabilization.

Human breast cancer cell line MCF-7 constitutively expressing PKC-δ (PKC-δ^+/+) and BT549 cells deficient in PKC-δ (PKC-δ^–/–) have been described previously (Jackson et al., 2005). PKC-δ^+/+ and PKC-δ^–/– cells were treated with either DMSO or tBHQ. In the same experiments cells were also pretreated with the PKC-δ-specific inhibitor rottlerin and then treated with tBHQ. The cells were lysed and analyzed for enodogenous PKC-δ, Nrf2, INrf2, HO1 and NQO1 protein levels by immunoblotting and probing with the respective antibodies. The results are shown in Fig. 3. PKC-δ was detected in MCF-7 PKC-δ^+/+ but not in BT549 PKC-δ^–/– cells (Fig. 3A). Interestingly, tBHQ treatment significantly increased PKC-δ in PKC-δ^+/+ cells and upregulated the genes for INrf2, HO1 and NQO1, which are downstream of the Nrf2 gene, in PKC-δ^+/+ cells (Fig. 3A). However, tBHQ failed to stabilize Nrf2 in PKC-δ^+/+ cells pretreated with the PKC-δ inhibitor rottlerin and in PKC-δ-deficient BT549 cells (Fig. 3A). In a related experiment, tBHQ also failed to increase ARE-luciferase gene expression in PKC-δ^–/– cells that was observed in PKC-δ^+/+ cells (Fig. 3B). An analysis of nuclear localization demonstrated that tBHQ induced nuclear localization of Nrf2 in PKC-δ^+/+ but not in PKC-δ^–/– cells (Fig. 3C). Immunocytolocalization of endogenous Nrf2 in PKC-δ^+/+ and PKC-δ^–/– cells treated with DMSO, tBHQ and tBHQ plus rottlerin are shown in Fig. 3D. tBHQ induced nuclear localization of Nrf2 in PKC-δ^+/+ but not in PKC-δ^–/– cells. Pretreatment with PKC-δ inhibitor rottlerin also abrogated tBHQ induced nuclear localization of Nrf2 (Fig. 3D). All the above analysis provided strong support for antioxidant-induced PKC-δ-mediated stabilization and nuclear translocation of Nrf2.

The transfection of the HA-PKC-δ plasmid resulted in the expression of HA-tagged PKC-δ in both cell types (Fig. 4A, lanes 3, 4, 7, 8). tBHQ, as observed earlier, stabilized Nrf2 and upregulated NQO1 in PKC-δ^+/+ but not in PKC-δ^–/– cells (Fig. 4A, compare lanes 1 and 2 with 5 and 6). Intriguingly, overexpression of HA-PKC-δ in PKC-δ^+/+ cells and expression of HA-PKC-δ in PKC-δ^–/– cells failed to stabilize Nrf2 (Fig. 4A, lanes 3 and 7). However, the treatment of PKC-δ^+/+ cells overexpressing HA-PKC-δ, and PKC-δ^–/– cells expressing HA-PKC-δ with tBHQ led to stabilization of Nrf2 (Fig. 4A, lanes 4 and 8). NQO1 upregulation followed a similar pattern to Nrf2 stabilization. These results led to two significant conclusions. First, the results confirm the role of PKC-δ in tBHQ-induced Nrf2 stabilization and induction of NQO1 gene expression since expression of HA-PKC-δ in PKC-δ-deficient cells led to tBHQ stabilization of Nrf2. Second, more intriguingly, PKC-δ alone was insufficient to stabilize Nrf2 and upregulate NQO1 gene expression. Both tBHQ and PKC-δ are essentially required for stabilization of Nrf2 and activation of NQO1 gene expression. These results were supported by tBHQ induction of ARE-luciferase gene expression in transfected cells (Fig. 4B) and nuclear localization of stabilized Nrf2 (Fig. 4C).

Hepa-1, PKC-δ^+/+, PKC-δ^–/– cells were treated with DMSO or tBHQ and endogenous PKC-δ protein levels analyzed by immunoblotting (Fig. 5A). tBHQ induced PKC-δ protein in Hepa-1 and PKC-δ^+/+ cells, whereas PKC-δ was not detected in PKC-δ^–/– cells (Fig. 5A). Next we analyzed serine phosphorylation of Nrf2 in these three cell lines after tBHQ treatment. Imunoprecipitation assays showed low base level of Nrf2 serine phosphorylation in Hepa-1 and PKC-δ^+/+ cells (Fig. 5B). Treatment with tBHQ increased time-dependent Nrf2 serine phosphorylation. Nrf2 serine phosphorylation was not detected in PKC-δ^–/– cells, even after tBHQ treatments (Fig. 5B, third blot from top). To test involvement of PKC-δ in Nrf2 serine phosphorylation, we analyzed the interaction of PKC-δ with Nrf2 by immunoprecipitation in PKC-δ^+/+ and PKC-δ^–/– cell lines. PKC-δ immunoprecipitated with Nrf2 in PKC-δ^+/+ cells (Fig. 5C). Treatment with tBHQ increased interaction between Nrf2 and PKC-δ compared with DMSO-treated cells. In addition, pretreatment with rottlerin significantly reduced interaction between Nrf2 and PKC-δ (Fig. 5C). Since PKC-δ is not present in PKC-δ^–/– cells, no interaction of Nrf2 and PKC-δ was found in PKC-δ^–/– cells (Fig. 5C).

We also used Nrf2S40 phospho-specific antibody from Abcam (ab76026) to demonstrate that PKC-δ phosphorylated Nrf2S40, which led to stabilization and nuclear localization of Nrf2. First, we characterized the specificity of Nrf2S40 phospho-specific antibody (supplementary material Fig. S1). Hepa-1 cells transfected with FLAG-Nrf2 or mutant Nrf2S40A-GFP were treated with either DMSO or tBHQ and immunoprecipitated with IgG (control) or anti-FLAG or anti-GFP and immunoblotted with anti-FLAG, anti-GFP and anti-phosphoserine40-Nrf2 antibody (supplementary material Fig. S1). Reverse immunoprecipitations were also performed. The Nrf2S40 phospho-specific antibody detected a single band of phosphorylated FLAG-Nrf2 in cells transfected with FLAG-Nrf2 but not in cells transfected with mutant Nrf2S40A-GFP. This demonstrated the specificity of the phospho-specific Nrf2S40 antibody in the detection of phosphorylated Nrf2S40 protein. PKC-δ^+/+ and PKC-δ^–/– cells treated with DMSO or tBHQ were lysed and immunoblotted with anti-Nrf2S40 phospho-specific antibody (Fig. 5D, upper panel). tBHQ induced Nrf2S40 phosphorylation in PKC-δ^+/+ but not PKC-δ^–/– cells. In related experiments, the majority of the Nrf2S40-phosphorylated protein was found localized in the nucleus (Fig. 5D, lower panel).

The cells were co-transfected with INrf2-V5 and FLAG-Nrf2 plasmids, immunoprecipitated and INrf2:Nrf2 interaction analyzed by immunoblotting. Overexpression of INrf2 in PKC-δ^+/+ cells led to degradation of Nrf2 (∼80%) as expected (Fig. 5E, lane1). The tBHQ treatment stabilized Nrf2 (Fig. 5E, lane 2). However, overexpression of INrf2 failed to degrade Nrf2 in PKC-δ^–/– cells even after DMSO or tBHQ treatment (Fig. 5E, lanes 3 and 4). The immunoprecipitation assay showed that Nrf2 interacts with INrf2 in DMSO-treated PKC-δ^+/+ cells and completely dissociates from INrf2 after treatment with tBHQ (Fig. 5F, lanes 2 and 3). However, Nrf2 is not released from INrf2 in PKC-δ^–/– cells even after DMSO or tBHQ treatment (Fig. 5F, lanes 4 and 5), suggesting that both antioxidant (tBHQ) treatment and PKC-δ are required for Nrf2-INrf2 dissociation.

To test whether serine 40 phosphorylation of Nrf2 is required for localization of Nrf2 in the nucleus, localization of wild-type Nrf2 and mutant Nrf2S40A was analyzed. Hepa-1 cells were transfected with FLAG-Nrf2 and Nrf2S40A-GFP plasmids, and cytosolic and nuclear extracts were prepared and immunoblotted with anti-FLAG and anti-GFP antibodies. Wild-type Nrf2 pre-dominantly localized in the nucleus (Fig. 6A). By contrast, Nrf2S40A mutant failed to localize in the nucleus (Fig. 6A). We also analyzed the time-dependent tBHQ-induced nuclear phosphorylation of serine 40 and tyrosine 568 of Nrf2. It has been shown earlier that tBHQ-induced Fyn kinase phosphorylates tyrosine 568 of Nrf2 that leads to nuclear export and degradation of Nrf2 (Jain and Jaiswal, 2006). Hepa-1 cells were co-transfected with FLAG-Nrf2, mutant Nrf2S40A-GFP or Nrf2Y568A-V5 plasmids and treated with DMSO or tBHQ as shown in Fig. 6B. Nuclear Nrf2 was immunoprecipitated and immunoblotted with anti-phosphoserine and anti-phosphotyrosine antibodies. Treatment with tBHQ for 4-8 hours increased nuclear Nrf2 serine phosphorylation by twofold compared with that in DMSO-treated cells. The phosphorylation of Nrf2 serine was halved when cells were incubated with tBHQ-free medium for a further 8 hours (Fig. 6B, top blot). In cells treated with tBHQ, nuclear phosphorylated tyrosine was not detected after 4 hours, but increased at 8 hours after tBHQ treatment and significantly increased (five- to sixfold) in cells treated with medium without tBHQ for a further 8 hours. The levels of nuclear phosphorylation of Nrf2 serine and tyrosine in response to antioxidant was quantified from three independent experiments and plotted (Fig. 6B right panel). Since Nrf2S40A mutant did not localize in the nucleus, we did not find phosphorylated serine Nrf2 in the nucleus (Fig. 6C, upper two blots). In addition, Nrf2Y568A mutant protein localized in the nucleus but did not show phospho-tyrosine adducts in response to DMSO or tBHQ treatments (Fig. 6C, lower two blots). These results suggested that Nrf2 undergoes PKC-δ-mediated serine 40 phosphorylation in response to tBHQ. Nrf2 is imported into the nucleus and binds with ARE. Then Nrf2 protein presumably undergoes serine 40 dephosphorylation by an unknown phosphatase and is exported from the nucleus by Fyn-mediated tyrosine phosphorylation as previously reported (Jain and Jaiswal, 2007).

We also examined the above mechanism in PKC-δ^+/+ and PKC-δ^–/– cells. PKC-δ^+/+ and PKC-δ^–/– cells were transfected with FLAG-Nrf2 or Nrf2S40A-GFP plasmids and treated with DMSO or tBHQ. Endogenous, transfected and Nrf2S40A mutant Nrf2 was immunoprecipitated from whole, cytosolic and nuclear extracts and immunoblotted with anti-FLAG, anti-GFP and anti-phosphoserine antibodies. tBHQ induced Nrf2 serine phosphorylation compared with DMSO in PKC-δ^+/+ but not in PKC-δ^–/– cells (Fig. 6D). Similarly, serine phosphorylation of endogenous Nrf2 and transfected FLAG-Nrf2 was found mainly in the nuclear compartments of PKC-δ^+/+ cells after tBHQ treatments, but was not detected in PKC-δ^–/– cells (Fig. 6E upper four blots). No serine phosphorylation was found in cytosolic or nuclear compartments of PKC-δ^+/+ and PKC-δ^–/– cells transfected with Nrf2S40A mutant (Fig. 6E lower two blots) suggesting that antioxidant treatment and PKC-δ are both required for phosphorylation of serine 40 in Nrf2.

We investigated the possible role of INrf2C151 in regulation of Nrf2. INrf2C151 was mutated to INrf2C151A (INrf2 cysteine 151 mutated to analine). Hepa-1 cells were co-transfected with FLAG-Nrf2, INrf2-V5, INrf2C151A-V5 and Nrf2S40A-GFP plasmids in combinations as shown in Fig. 7 and treated with DMSO or tBHQ for 2 hours and immunoblotted to show their expression in transfected cells (Fig. 7A, inputs). INrf2-V5 and mutant INrf2C151A-V5 were immunoprecipitated with V5 antibody and probed with anti-V5 and FLAG antibody (Fig. 7B). Immunoprecipitation of INrf2-V5 pulled down FLAG-Nrf2 in DMSO-treated cells but not in tBHQ-treated cells because of release of FLAG-Nrf2 from INrf2-V5 (Fig. 7B-left panel). In the same experiment, V5-antibody-immunoprecipitated mutant INrf2C151A-V5 and FLAG-Nrf2 was pulled down from both DMSO- and tBHQ-treated cells. In other words, tBHQ failed to release FLAG-Nrf2 from mutant INrf2C151A-V5 (Fig. 7B, left panel). Replacement of FLAG-Nrf2 with mutant Nrf2S40A-GFP in similar experiments failed to release mutant Nrf2S40A-GFP from both INrf2 and mutant INrf2C151A (Fig. 7B, right panel). Replacement of Hepa-1 cells with PKC-δ^+/+ cells showed similar results (Fig. 7C, left panel). tBHQ failed to release Nrf2 from mutant INrf2C151A. Similar experiments in PKC-δ^–/– cells demonstrated that Nrf2 binds to INrf2 and mutant INrf2C151A. However, tBHQ failed to release Nrf2 from either INrf2 or mutant INrf2C151A (Fig. 7C, right panel). Localization data demonstrated that, tBHQ treatment completely dissociated FLAG-Nrf2 from INrf2-V5 in the cytosol and Nrf2 localized in the nucleus (Fig. 7D, left panel). INrf2C151A-V5 retained more than 50% FLAG-Nrf2 in the cytosol and reduced its nuclear localization (Fig. 7D, left panel) after tBHQ treatments. Interestingly, Nrf2S40A-GFP protein did not localize in the nucleus therefore it accumulated in the cytosol even after tBHQ treatments (Fig. 7D, right panel). The above results together suggest that INrf2C151 is required for tBHQ-mediated release of Nrf2 from INrf2. However, the results also demonstrate that this is not sufficient since Nrf2S40A was not released from INrf2 in response to tBHQ treatment, and also Nrf2 was not released from INrf2 in PKC-δ-deficient cells.

Hepa-1 cells were transfected with INrf2-V5 and INrf2A151A-V5 plasmids and treated with cyclohexamide to stop the protein synthesis, treated with DMSO or tBHQ, lysed and immunoblotted to show the rate of degradation of INrf2 and mutant INrf2 (Fig. 8A). Immunoblotting analysis showed that INrf2 protein degrades by more than 60% within 4 to 6 hours after tBHQ treatment compared with DMSO treatment (Fig. 8A). Interestingly, INrf2C151A degrades approximately four times faster than wild-type INrf2 during tBHQ treatments. The pattern of degradation of INrf2-V5 and INrf2C151A was quantified and plotted (Fig. 8A, lower panel). These results indicate that tBHQ induces degradation of INrf2 protein. In a related experiment, treatment with the protease inhibitor MG132 blocked the degradation of INrf2 and mutant INrf2C151A (Fig. 8B).

We analyzed tBHQ-mediated destabilization and/or ubiquitylation of INrf2 in PKC-δ^+/+ and PKC-δ^–/– cells to demonstrate that both cells have more or less similar capacity to ubiquitylate after co-transfection with HA-Ub, INrf2-V5 and FLAG-Nrf2 plasmids, and degradation of INrf2 was visualized by immunoblotting (Fig. 8C). tBHQ treatment degrades INrf2 faster than DMSO treatment in both PKC-δ^+/+ and PKC-δ^–/– cells (Fig. 8C), and tBHQ also triggers ubiquitylation of INrf2 similarly in PKC-δ^+/+ and PKC-δ^–/– cells (Fig. 8D). This data suggested that destabilization and ubiquitylation of INrf2 in response to antioxidant might be an early event in stabilization of Nrf2 and that this mechanism works independently in PKC-δ^+/+ and PKC-δ^–/– cells.

We analyzed the role of Nrf2 in drug resistance and cell survival to investigate the biological significance of antioxidant-induced PKC-δ-mediated stabilization and nuclear translocation of Nrf2. PKC-δ^+/+ and PKC-δ^–/– cells were treated with optimal concentration of etoposide (30 μM), a potent apoptosis enhancing drug, for 72 hours and further treated with DMSO or tBHQ for an additional 48 hours in the presence of the drug. Cells were harvested and cellular Nrf2 levels were analyzed by western blotting (Fig. 9A). Treatment with etoposide alone stabilized Nrf2 in PKC-δ^+/+ cells. The treatment of PKC-δ^+/+ cells with a combination of etoposide and tBHQ further increased cellular level of Nrf2 (∼1.8-fold) compared with etoposide alone. In the same experiment, PKC-δ^–/– cells failed to show stabilization of Nrf2 either with etoposide or a combination of etoposide and tBHQ. The NQO1 gene, downstream of the Nrf2 gene, showed a similar pattern to that of Nrf2 in PKC-δ^+/+ cells. NQO1 expression was not detected in PKC-δ^–/– cells. We also used etoposide to analyze the role of PKC-δ and Nrf2 in caspase 3 activation and apoptosis. PKC-δ^+/+ (MCF-7) cells were transfected with control or PKC-δ siRNA and after 24 hours, cells were post-treated with etoposide for an additional 48 hours, cells were harvested and cellular caspase 3 (native and cleaved), PKC-δ, Nrf2 and actin levels were analyzed by western blotting (Fig. 9B). PKC-δ knock down downregulated Nrf2 and resulted in etoposide-mediated activation of caspase 3 (Fig. 9B). The level of cleaved caspase 3 was significantly increased in PKC-δ knock down cells after etoposide treatment indicating cellular apoptosis. Furthermore, we analyzed the etoposide-mediated DNA fragmentation pattern in PKC-δ^+/+ and PKC-δ^–/– cells. Etoposide treatment of PKC-δ^+/+ and PKC-δ^–/– cells increased DNA fragmentation by 45% and 75%, respectively, compared with DMSO treatment (Fig. 9B). Interestingly, tBHQ reduced etoposide-mediated DNA fragmentation by (∼25%) in PKC-δ^+/+ cells. However, tBHQ failed to reduce DNA fragmentation in PKC-δ^–/– cells (Fig. 9B). The cell survival was similar to the DNA fragmentation. Exposure of PKC-δ^+/+ and PKC-δ^–/– cells to etoposide reduced the cell survival by 43% and 70%, respectively, compared with DMSO treatment (Fig. 9C). Interestingly, inclusion of tBHQ significantly enhanced cell survival by (∼38%) in PKC-δ^+/+ cells. In the same experiment, tBHQ failed to increase cell survival in PKC-δ^–/– cells (Fig. 9C). These results suggested that PKC-δ is a key factor in regulation of antioxidant-induced Nrf2 stabilization and nuclear localization with significant impact on drug resistance and cell survival.

The present study investigated the role of the various isoforms of PKC in phosphorylation of Nrf2 and release from INrf2. The results revealed that PKC-δ isoform plays a major role in phosphorylation of Nrf2S40 that leads to the release and stabilization of Nrf2. Rottlerin, a specific inhibitor of PKC-δ significantly inhibited the basal and antioxidant-mediated induction of ARE gene expression to a similar magnitude as staurosporine, a non-specific inhibitor of all PKC isoforms. Rottlerin also inhibited stabilization and nuclear localization of Nrf2 in response to antioxidant. In similar experiments, chemical inhibitors of other PKC forms failed to inhibit stabilization of Nrf2 and antioxidant induction of localization of Nrf2. In addition, PKC-δ phosphorylated bacterially purified Nrf2 in an in vitro kinase reaction. PKC isoform-specific dominant-negative mutants and siRNA experiments provided further evidence of PKC-δ being the master regulator of Nrf2. Dominant negative mutants against PKC-δ but not other PKC isoforms interfered with antioxidant-induced stabilization and activation of Nrf2. Similarly, PKC-δ-specific siRNA decreased PKC-δ protein that led to repression of antioxidant induction of ARE-mediated gene expression. The role of PKC-δ in ARE-mediated gene expression is also supported by an earlier report showing rottlerin inhibition of HO-1 gene expression (Zhang et al., 2008). PKC-η also phosphorylated bacterially purified Nrf2. However, unlike PKC-δ, PKC-η failed to demonstrate a role in Nrf2 regulation of ARE-mediated gene expression. This suggested that although PKC-η phosphorylated bacterial Nrf2, its role in ARE-mediated gene expression remains unknown.

Controversies exist as to the role of chemical modification of INrf2 or phosphorylation of PKC in release of Nrf2 from INrf2. Several experiments in the present studies demonstrated that both modification of INrf2C151 and PKC phosphorylation of Nrf2S40 are required for the release of Nrf2 from INrf2, leading to the stabilization and nuclear localization of Nrf2. These included: (1) Nrf2 bound to both INrf2 and mutant INrf2C151A. However, antioxidant treatment released Nrf2 from INrf2 but failed to release Nrf2 from mutant INrf2C151A. This indicated that INrf2C151 is required for antioxidant-induced release of Nrf2 from INrf2. (2) Antioxidant release of Nrf2 but not mutant Nrf2S40A from INrf2. (3) Antioxidant released Nrf2 from INrf2 in PKC^+/+ cells but failed to release Nrf2 from INrf2 in PKC-δ^–/– cells. (4) Both PKC-δ^+/+ and PKC-δ^–/– cells showed more or less equal capacity to ubiquitylate INrf2 in response to antioxidant. However, Nrf2 failed to dissociate from INrf2 in PKC^–/– cells. (5) Expression of PKC-δ alone in PKC-δ^–/– cells failed to stabilize Nrf2. Both expression of PKC-δ and antioxidant were required for the release of Nrf2 from INrf2 and stabilization of Nrf2.

The results also suggested that antioxidant-induced modification of INrf2C151 leads to conformational change in INrf2 thus exposing buried Nrf2S40 for phosphorylation by PKC-δ. This leads to the release of Nrf2 from INrf2. It is known that PKC is activated in response to oxidative stress (Ha et al., 2001; Otieno and Kensler, 2000) and it is also known to play an important role in cell proliferation and apoptosis (Niwa et al., 2002). Tyrosine phosphorylation of PKC-δ regulates nuclear translocation of PKC-δ leading to apoptosis (Humphries et al., 2008). Most of the PKC-δ is present in the cytosolic compartment. The results in this study demonstrate that antioxidant induced PKC-δ expression that in turn phosphorylated Nrf2S40, leading to release and stabilization of Nrf2. The results also demonstrated that once the stress is gone, tyrosine phosphorylation of Nrf2 is increased, presumably to export Nrf2 out of the nucleus for degradation.

Our data suggest that oxidation of INrf2C151 followed by PKC-δ-mediated serine 40 phosphorylation are the key events for stabilization of Nrf2 during oxidative stress. The accumulation of Nrf2 in the nucleus led to increased expression of Nrf2 downstream genes that reduced etoposide-mediated DNA fragmentation in PKC-δ^+/+ cells but not in PKC-δ^–/– cells. This led to drug resistance and increased cell survival in PKC-δ^+/+ cells. PKC-δ^–/– cells failed to demonstrate drug resistance and increase in cell survival because of the absence of PKC-δ and Nrf2 stabilization. The role of Nrf2 in drug resistance is known and Nrf2 is being targeted to reduce drug resistance (Singh et al., 2006). The present report also highlights PKC-δ as another target of drug resistance. This is evident from the observation that antioxidant treatment induced PKC-δ stabilized Nrf2 and reduced etoposide resistance.

It is noteworthy that a large number of endogenous and exogenous agents, drugs and radiation dissociate Nrf2 from INrf2 leading to stabilization of Nrf2 and activation of defensive genes expression (Jaiswal, 2004). All these agents might stabilize Nrf2 by a single mechanism involving modification of INrf2C151 followed by PKC-δ phosphorylation of Nrf2S40, as shown in the current study, or some agents might be involved with additional, as yet unknown, mechanism(s) to dissociate Nrf2 from INrf2 resulting in activation of defensive genes expression.

In summary, the present studies led to the following conclusions, as also shown in a model in Fig. 10. First, PKC-δ phosphorylates Nrf2S40 leading to release of Nrf2 from INrf2. Second, antioxidant-induced modification of INrf2C151 followed by PKC-δ phosphorylation of Nrf2S40 is required for the release of Nrf2 from INrf2 and activation of ARE-mediated gene expression. Third, antioxidant induction of PKC-δ leads to release and stabilization of Nrf2, increased expression of Nrf2 downstream genes, reduced apoptosis and increased drug resistance. Lastly, PKC-δ like Nrf2 is an excellent target for controlling drug resistance.

The construction of pGL2B-NQO1-ARE, pcDNA-FLAG-Nrf2, pcDNA-Nrf2-V5, pcDNAY568ANrf2-V5, pcDNA-INrf2-V5, pcDNA-FLAG-INrf2, HA-UB and pcDNA-Nrf2S40A-GFP has been described previously (Jain and Jaiswal, 2006; Jain and Jaiswal, 2007; Niture and Jaiswal, 2009). The cysteine151 residue of INrf2 was mutated to alanine by using a site directed mutagenesis kit (Invitrogen) and forward primer: 5′-TCCATCTCCGTGGGCGAGAAGGCTGTCCTGCACGT-3′ and reverse primer: 5′-CTTCTCGCCCACGGAGATGGAGGCCGTGTAGGCGA-3′. HA-tagged PKC-δ plasmid was obtained from Addgene (plasmid-16386).

Human hepatoblastoma (Hep-G2), mouse hepatoma (Hepa-1), MCF7-PKC-δ^+/+ and BT549-PKC-δ^–/– cells were grown in MEM (Hep-G2), DMEM (Hepa-1 and MCF-7) and RPMI (BT549) supplemented with 10% fetal bovine serum, penicillin (40 IU/ml), and streptomycin (40 μg/ml) in an incubator at 37°C in 95% air and 5% CO₂. Transient transfections were performed in cells grown to ∼50% confluence using the Effectene Transfection reagent (Qiagen). Cells were co-transfected with 0.2 μg of reporter construct (human NQO1-ARE-Luc) and 0.02 μg firefly-Renilla luciferase plasmid pRL-TK. Renilla luciferase was used as the internal control in each transfection. Plasmid expressing dominant negative isoforms of PKC was co-transfected with the reporter plasmid in the concentrations indicated in the figures. The transfected cells were pretreated for 8 hours with the indicated specific inhibitor (staurosporine, Go6850, Go6976, Go6983 or rottlerin; Calbiochem) at the concentrations given in the figures. After inhibitor treatments, cells were again treated with DMSO or with tBHQ (50 μM) for 16 hours in the medium containing the indicated kinase inhibitors. The cells were washed with 1× PBS and lysed in 1× passive lysis buffer from the Dual-Luciferase Reporter Assay System Kit (Promega). The effect of tBHQ on NQO1-ARE-luciferase activity in MCF7-PKC-δ^+/+ and BT549-PKC-δ^–/– cells was measured as described above and after co-transfection of 0.1 μg NQO1-ARE-Luc and 10 ng of firefly-Renilla luciferase encoded by plasmid pRL-TK.

For an in vitro kinase assay, 0.5 μg bacterially purified Nrf2 protein (Jain and Jaiswal, 2006) was used as the substrate. The purified protein was incubated with the PKC enzyme(s) and [γ-³²P]ATP in PKC kinase assay buffer [20 mM Hepes (pH 7.4), 1 mM dithiothreitol, 10 mM MgCl₂, 1.7 mM CaCl₂ and 0.1 mg/ml phosphatidylserine] for 1 hour at 30°C. The proteins were then resolved by SDS-PAGE and visualized by autoradiography.

To prepare whole cell lysates, the cells were lysed in RIPA buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 0.2 mM EDTA, 1% Nonidet P-40, 0.5% deoxycholic acid, 1 mM phenylmethylsulfonyl fluoride and 1 mM sodium orthovanadate) supplemented with protease inhibitor mixture (Roche Applied Science). Cytoplasmic and nuclear lysates were separated by using the Active Motif nuclear extract kit (Active Motif, Carlsbad, CA) by following the manufacturer's protocol. The protein concentration was determined using the protein assay reagent (Bio-Rad). 60-80 μg of proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes. The membranes were blocked with 3% non-fat dry milk and incubated with anti-INrf2 (E-20; 1:1000), anti-Nrf2 (H-300; 1:500), anti-heme oxygenase 1 (HO1; 1:1000), anti PKC-δ (C-20; 1:1000) antibodies, all purchased from Santa Cruz Biotechnology. Anti-FLAG-HRP (1:10,000), anti-HA-HRP (1:10,000) and anti-actin (1:10,000) antibodies were obtained from Sigma. Anti-GFP and anti-V5 HRP antibodies were obtained from Invitrogen and anti-caspase 3 antibodies from Promega. The membranes were washed three times with TBST and immunoreactive bands were visualized using a chemiluminescence system ECL (Amersham). The intensity of protein bands was quantified by using QuantityOne 4.6.3 Image software (ChemiDoc XRS; Bio-Rad) and normalized against proper loading controls. To confirm the purity of nuclear-cytoplasmic fractionation, the membranes were reprobed with cytoplasm-specific, anti-lactate dehydrogenase (Chemicon) and nuclear-specific, anti-lamin B antibodies (Santa Cruz Biotechnology).

Hepa-1, MCF7 and BT549 cells were transfected with the indicated plasmids for 36 hours. The transfected cells were lysed in RIPA buffer containing 1 mM PMSF, 1 mM sodium vanadate and serine/therione phosphatase inhibitor cocktail (Sigma) and protease inhibitors (Roche Applied Science). For immunoprecipitation, 1 mg of whole cell lysates or nuclear and/or cytoplasmic extracts was pre-cleaned by protein-AG plus agarose (Santa Cruz Biotechnology) and then immunoprecipitated with the respective antibodies (1 μg) at 4°C overnight. Immune complexes were analyzed by immunoblotting using previously described procedures (Niture and Jaiswal, 2009). The immunoblots were blocked in 2% BSA in TBST and probed with anti-Nrf2 antibody, anti-Nrf2phosphoserine40 antibody (Abcam) or anti-phosphoserine-HRP antibody and anti-phosphotyrosine antibodies (Stressgen), as described in the figures. Immunoreactive bands were visualized using a chemiluminescence system ECL (Amersham).

siRNA against human PKC-δ, PKC-η and lamin A/C (control) were obtained from Dharmacon (Lafayette, CO). Cells were co-transfected with hNQO1-ARE-luciferase reporter plasmid, Renilla luciferase plasmid and different PKC-isoform-specific siRNA using Lipofectamine (Invitrogen) following the manufacturer's instructions. 48 hours after transfection cells were treated with DMSO or 50 μM tBHQ for 16 hours and luciferase activity determined as described above. 100 μg of cell lysate from siRNA transfected and DMSO treated cells were resolved on 10% SDS-PAGE and immunoblotted with anti-PKC-δ, anti-PKC-η, anti-lamin A/C (all from Santa Cruz Biotechnology) and anti-β-actin antibody (Sigma). The effect of PKC-δ siRNA on interaction of INrf2-Nrf2 in Hapa-1 cells was examined by transfection of control or PKC-δ siRNA for 10 hours followed by co-transfection of INrf2-V5 and FLAG-Nrf2 constructs. One mg of cell lysates was immunoprecipitated with anti-V5 antibody and immunoblotted with anti-FLAG-HRP antibody.

MCF7-PKC-δ^+/+ and BT549-PKC-δ^–/– cells were grown on coverslips as described (Niture and Jaiswal, 2009). To determine the location of endogenous Nrf2, the cells were pre-treated with 25 mM rottlerin for 2 hours in some cases, followed by treatment with 50 mM tBHQ with or without rottlerin for 2 hours. Cells were then fixed and probed with Nrf2 antibody. FITC-conjugated anti-rabbit antibody (Chemicon International, Temecula, CA) was used as a secondary antibody as previously described (Niture and Jaiswal, 2009). To visualize the nuclei, the cells were stained with DAPI (Vector laboratories). After immunostaining, cells were observed under a Nikon fluorescence microscope and photographed.

Hepa-1 cells were transfected with INrf2-V5 and C151AINrf2-V5. After 24 hours cells were pretreated with 30 μg/ml cyclohexamide for 90 minutes and further treated with either DMSO or tBHQ (50 μM) for various time periods (1-8 hours). 60 μg of cell extracts were immunoblotted with anti-V5 antibody and anti-actin antibody. To test the effect of proteasome inhibitor MG132 on the stability of wild-type INrf2 and C151AINrf2 protein, Hepa-1 cells were transfected with INrf2-V5 and C151AINrf2-V5. After 24 hours, cells were treated with 10 μM MG132 for 12 hours. 60 μg of cell extracts were immunoblotted with anti-V5 antibody and anti-actin antibody. For ubiquitylation assay, Hepa-1 cells were seeded in 100-mm plates and co-transfected with INrf2-V5 (1 μg), HA-Ub (0.2 μg) and FLAG-Nrf2 (1.0 μg) for 24 hours, and treated with either DMSO or tBHQ for an additional 8 hours. Cells were lysed in RIPA buffer and 1 mg cell lysates were immunoprecipitated with anti-V5 antibody and western blotted with anti-HA antibody. To check the effect of tBHQ on ubiquitylation-dependent degradation of INrf2, 60 μg of the cell lysates were immunoblotted with anti-FLAG, anti-V5 and anti-actin antibodies.

PKC-δ^+/+ and PKC-δ^–/– cells were plated, allowed to recover for 12 hours, and then exposed to etoposide (30 μM) for 72 hours. After 72 hours cells were treated with DMSO or tBHQ for 48 hours. A photometric enzyme immunoassay was performed for the quantitative in vitro determination of cytoplasmic histone-associated DNA fragments (mono and oligonucleosomes) after etoposide exposure to cells using a Cell Death Detection ELISA kit (Roche) following the manufacturer's instructions. In a related experiment the cells were incubated with fresh MTT solution (200 μl/well; stock 5 mg/ml in PBS) for 2 hours, and absorbance at 490 nm was measured. For each drug concentration eight replicates were tested, and the experiment was repeated three times. Each data point represents a mean ± s.d. and was normalized to the value of the corresponding control cells.