Phosphorylation of Nrf2 at Ser40 by Protein Kinase C Regulates Antioxidant Response Element-mediated Transcription

Together these results suggest that one critical step in the signaling cascade towards activation may be the phosphorylation of Nrf2 by PKC, which promotes the nuclear translocation of this transcription factor in response to oxidative stress. The present study is a continuation of our investigation into the involvement of PKC in regulating the ARE. We sought to identify the site of phosphorylation in Nrf2 by PKC, and to characterize the mechanistic significance of Nrf2 phosphorylation.


INTRODUCTION
The antioxidant response element (ARE) 1 is a regulatory sequence involved in the coordinated transcriptional activation of genes coding for a number of antioxidant enzymes and phase II detoxifying enzymes (1)(2)(3)(4)(5)(6). Reactive oxygen species and electrophiles are potent activators of genes containing an ARE, mediated by the basic leucine zipper (bZIP) transcription factor Nrf2 (NF-E2-related factor 2) (7)(8)(9). Accumulated evidence from studies of nrf2-null mice has established that Nrf2 is an essential ARE-binding factor involved in both constitutive and inducible gene expression via the ARE (9)(10)(11). An important regulatory step leading to ARE activation is the oxidative stress-induced nuclear translocation of Nrf2, which normally appears to be sequestered in the cytoplasm by the cytoskeleton-binding Keap1 protein (12)(13)(14). However, the precise mechanism by which ARE-activating signals reach Nrf2 and cause dissociation of the putative inhibitory Nrf2-Keap1 complex remains unclear.
Several protein kinase pathways have been implicated in transducing oxidative stress signals to gene expression mediated through the ARE. A number of reports have addressed a possible role for extracellular signal-regulated kinase (ERK1/2) in ARE activation. The findings have however remained controversial: ERK1/2 has been found to regulate the ARE positively in certain hepatoma cells (15)(16)(17) but negatively in others (18). Similarly, p38 MAP (mitogenactivated protein) kinase has also been shown to affect ARE activity, either positively (17,19,20) or negatively (16,21). More recently, phosphatidylinositol 3-kinase (PI3-kinase) and its downstream target Akt/PKB have been linked to activation of the ARE in hepatoma (18,19) and neuroblastoma (22) cell lines. However, none of the known cellular components involved in ARE regulation have been shown to be targets of any of these kinases.
Recently, we reported several findings that indicate an important role for protein kinase C (PKC) in the ARE-mediated gene expression (23). 1) Phorbol 12-myristate 13-acetate (PMA), a potent PKC-activating phorbol ester, stimulates ARE-driven transcription, which is blocked by selective PKC inhibitors. 2) Nuclear translocation of Nrf2 is induced by PMA but arrested by PKC inhibitors. 3

A Peptide Inhibitor of Nrf2 Phosphorylation by PKC
In prior studies we showed that Nrf2 is phosphorylated in HepG2 cells (23 1A). All were readily soluble in 20% DMSO except peptide 36-43, whose solubility was greatly improved when it was extended at both ends by one residue to a 10mer (EVFDFSQRQK). In vitro kinase assays were performed as described previously (23), using commercially obtained catalytic subunits of rat brain PKC and purified rat Nrf2 as substrate. Although Nrf2 phosphorylation by PKC appeared unaffected or even enhanced by several of the peptides, it was reduced by more than 90% in the presence of 5 mM peptide 35-44 ( Fig. 1B). PKC activity against a standard substrate was not suppressed by this peptide (not shown), indicating that its effect was not on the enzyme itself. These findings suggest that Ser40 of Nrf2 is an authentic site of phosphorylation by PKC.

A Nrf2 Mutant Defective for PKC Phosphorylation
To confirm these peptidomimetics data a site-directed mutagenesis approach was utilized. Rat Nrf2 gene bearing a AGT->GCT (Ser-to-Ala) mutation at amino acid position 40 ( nrf2-S40A) was cloned into a high-level expression plasmid containing a His 6 -tag N-terminal to the insert. Escherichia coli-expressed Nrf2-S40A protein was purified to near homogeneity by metal chelate affinity chromatography ( Fig. 2). The prominent band ~90 kD was confirmed to be Nrf2 by Western blot using an antibody against Nrf2 (not shown). Nrf2-S40A was then used in parallel with wild-type Nrf2 as substrates in in vitro PKC assays. As shown in Fig. 3, the single amino acid change from Ser to Ala at position 40 completely abolished PKC phosphorylation of Nrf2. The lack of residual phosphorylation in this mutant indicates that Ser40 is the only PKC site, consistent with the peptide competition data.
It should be noted that Ser40 is also one of four potential PKC sites in human Nrf2 (GenBank accession number Q16236), which is highly homologous to the rat Nrf2. It is likely that PKC phosphorylates human Nrf2 at the same site, and that the phosphorylation of Nrf2 observed in human hepatoma HepG2 cells (23)  These complexes were supershifted upon incubation with an antibody against Nrf2 (Fig. 4A).
The intensity and mobility of the wild-type and S40A mutant complexes are virtually indistinguishable. Furthermore, formation of both types of complexes was completely blocked by excess unlabeled QR ARE but not by random oligonucleotides (Fig. 4B). Therefore the S40A mutation did not alter the specific high-affinity interaction between Nrf2/MafK and the ARE.
Indeed, PKC-phosphorylated Nrf2 bound to the ARE in a similar manner as non-phosphorylated wild-type Nrf2 or the S40A mutant defective for PKC phosphorylation (Fig. 4C). Thus the formation of ARE-binding transcriptional complex containing Nrf2 and small Maf proteins does not appear to be regulated by the phosphorylation of Nrf2 by PKC.

Keap1 Is Involved in the Impaired ARE Activation by Nrf2-S40A
To determine whether Since Keap1 has been shown to repress Nrf2 activity by sequestering it in the cytoplasm (12), we asked if any functional defect of the Nrf2-S40A mutant might involve Keap1. As expected, co-transfection of Keap1 with wild-type Nrf2 decreased Nrf2-dependent ARE activation. Interestingly, overexpression of Keap1 resulted in a partial impairment of ARE activation by the Nrf2-S40A mutant that was not seen with endogenous levels of Keap1 in HepG2 cells. ARE activation by Nrf2-S40A was reduced to a level less than 50% of that achieved with wild-type Nrf2, from an approximately 6-fold activation to 2.5-fold (Fig. 5). In these transfected cells, Nrf2 wild-type and S40A proteins were expressed to comparable levels as verified by Western blot using an anti-Nrf2 antibody (not shown). These findings indicate a role for Keap1 in the apparent transactivation defect exhibited by the Nrf2-S40A mutant.

Phosphorylation of Nrf2 by PKC Promotes its Dissociation from Keap1
Studies from Itoh et al. (12)  quantitatively co-precipitated with Nrf2 by the anti-Nrf2 antibody (Fig. 6A).
We then investigated the interaction between Keap1 and the Nrf2-S40A mutant. Similar amounts of Keap1 co-precipitated with both wild-type and S40A Nrf2 (Fig. 6B). However, when immunoprecipitation was carried out after incubation of these components in the presence of PKC, the amount of Keap1 associated with wild-type Nrf2 was reduced by about 50%. By contrast, the S40A mutant interacted with Keap1 to a similar extent with or without PKC (Fig.   6B). Furthermore, the dissociation of Keap1 from wild-type Nrf2 was abolished when PKC was preincubated in the presence of 10 nM staurosporine, a potent inhibitor of PKC (not shown). We therefore conclude that phosphorylation of Nrf2 by PKC at Ser40 plays a critical role in facilitating the release of Nrf2 from Keap1. Consequently, the impaired ability of the Nrf2-S40A mutant to activate ARE-mediated transcription is most likely due to a defect in the dissociation of Nrf2 from its cytoplasmic inhibitor Keap1.

DISCUSSION
We recently reported that phosphorylation of Nrf2 by PKC induces nuclear translocation of this transcription factor and activation of the ARE in response to oxidative stress (23 Reports from several laboratories have indicated the involvement of other kinase pathways in ARE-mediated transcription (17)(18)(19)(20)(21)(22). Transfection studies using wild-type and dominant-negative MAPK and Nrf2 have suggested a positive role for MAPK in ARE regulation in HepG2 cells (15,16). However, inhibition of ERK in H4IIE cells resulted in increased expression of the GSTA2 gene, suggesting a negative role for ERK in ARE-mediated activity (18). The exact role for the p38 kinase in ARE regulation is also confusing from existing literature, as several laboratories have reported a positive effect (17,19,20), while another showed that a p38 inhibitor caused the activation of ARE-dependent reporter gene (21).
Recently, PI3 kinase and its downstream Akt kinase have been shown by inhibitor studies to be positive regulators of ARE activity in H4IIE hepatoma (18) and IMR-32 neuroblastoma cells (22). Further elucidation of the precise roles of these kinases will be facilitated by the identification of specific cellular targets known to be involved in ARE regulation. It should be noted that in IMR-32 cells, ARE-mediated transcription has been reported to be PKCindependent (33). However, the same experimental system was also shown to be oxidative-stress independent, in contrast to the many studies from hepatocytes, where oxidative stress-induced ARE-mediated gene expression has been firmly established.  (36). ATF4 has also been demonstrated to interact with Nrf2 in regulating ARE-driven heme oxygenase-1 gene expression (37).
It should be noted that while the nuclear translocation of Nrf2 has been shown to be a major mechanism for ARE activation in all cell types examined, Nrf2 also has a documented role in constitutive ARE-mediated gene expression. We have shown that Nrf2 is present in the nucleus without tBHQ stimulation, and its essential role in mediating the basal activity of the ARE has been reported in a cell-free system (24) and in nrf2-null mice (11). Future investigations should reveal molecular details that comprise the coordinated transcriptional activation of antioxidant enzymes, of which phosphorylation of Nrf2 at Ser40 by PKC is but one of many critical regulatory steps.