NRF2 Regulates PINK1 Expression under Oxidative Stress Conditions

Mutations of the PTEN-induced putative kinase 1 (PINK1) gene are a cause of autosomal recessive forms of Parkinson’s disease. Recent studies have revealed that PINK1 is an essential factor for controlling mitochondrial quality, and that it protects cells from oxidative stresses. Although there has been considerable progress in the elucidation of various aspects of PINK1 protein regulation such as activation, stability and degradation, the transcriptional regulation of PINK1 mRNA under stress conditions remains unclear. In this study, we found that nuclear factor (erythroid-derived 2)-like 2 (NRF2), an antioxidant transcription factor, regulates PINK1 expression under oxidative stress conditions. Damaged mitochondria arising from stress conditions induced NRF2-dependent transcription of the PINK1 gene through production of reactive oxygen species (ROS). Either an ROS scavenger or forced expression of KEAP1, a potent inhibitory partner to NRF2, restricted PINK1 expression induced by activated NRF2. Transcriptionally up-regulated PINK1 diminished oxidative stress-associated cell death. The results indicate that PINK1 expression is positively regulated by NRF2 and that the NRF2-PINK1 signaling axis is deeply involved in cell survival.


Cell transfection and luciferase reporter assay
For plasmid transfection, cells were transfected with the indicated plasmids using FuGEN-E-HD (Promega Biosciences) according to the manufacturer's instructions. The transfection was carried out at 40% cell density. For 12-well plates in transfection experiments, 1 μg of pDNA and 2 μl of FuGENE-HD were used.
Luciferase activity of transfected cells was measured using a luciferase reporter assay system (Perkin Elmer) and Fluoroskan Ascent FL (Thermo Scientific, Lafayette, CO, USA). The luciferase activity was normalized to co-transfected GFP fluorescence.

Real-time PCR
Total RNA was prepared using an SV Total RNA Isolation System (Promega Biosciences). First-strand cDNA synthesis was performed with total RNA using a SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen). Synthesized cDNA was used for PCR analysis using TaqMan Fast Uviversal PCR Master Mix (Applied Biosystems) with TaqMan Gene Expression Assay targeting PINK1 (Hs00260868_m1), NRF2 (NFE2L2) (Hs00975961_g1) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Hs02758991_g1) (Applied Biosystems). Relative expression levels were analyzed using StepOnePlus (Applied Biosystems) and they were calculated using the ΔCt method and normalized against GAPDH as an internal control.

Western blot analysis
Western blot analysis was performed under conventional conditions after lysing cells using SDS sample buffer with PhosphoSTOP (Roche Applied Science, Indianapolis, IN, USA). Ten μg of protein extracts was separated by SDS-polyacrylamide gel electrophoresis and electro-transferred onto an Immobilon membrane (Millipore). To detect the immunoreactive proteins, we used HRP-conjugated anti-mouse or anti-rabbit secondary antibodies (Cell Signaling Technologies) and Pierce Western Blotting Substrate Plus (Thermo Scientific).

Cell viability assay and ROS assay
The CellTiter 96 aqueous one solution cell proliferation assay (Promega Biosciences) was used for cell viability assessments. According to the manufacturer's instructions, cells were incubated with CellTiter 96 aqueous one solution reagent for 3 h. The absorbance of 450 nm was analyzed using an iMark microplate absorbance reader (Bio-Rad). ROS-Glo H 2 O 2 assay (Promega Biosciences) was used for an ROS assay. The cells were incubated with test compounds and H 2 O 2 substrate solution for 6 h and then ROS-Glo detection solution was added. The luminescence was observed using Fluoroskan Ascent FL.

Statistical analysis
Prior to statistical analysis, each experiment was repeated three times. The results are expressed as means +/-S.D. For comparison, analysis of variance (ANOVA) was used. If the ANOVA showed a significant difference, the Bonferroni procedure was used as a post hoc test. p values of less than 0.05 were considered statistically significant.

Transcriptional up-regulation of PINK1 under oxidative stress conditions
To examine the effect of oxidative stress on expression of PINK1, a reporter assay was carried out using the PINK1 promoter (3 kbp)-luciferase-reporter construct (PINK1 pro-luc). We used human neuroblastoma SH-SY5Y cells for this assay because the cells are often used as in vitro models to analyze neuronal function and Parkinson's disease. SH-SY5Y cells were transfected with PINK1 pro-luc for 24 h and then treated with a neurotoxic reagent, 6-hydroxydopamine (6-OHDA); a metal-stress inducer, CdCl 2 ; an inhibitor of mitochondrial respiratory complex I, rotenone; and an ROS inducer, paraquat for 48 h. As shown in Fig 1A-1D, these reagents increased the luciferase activity of PINK1 pro-luc in a dose-dependent manner. We next examined effects of the reagents on expression of endogenous PINK1 mRNA at regular time intervals. These reagents increased PINK1 mRNA levels at 12-48 h (Fig 1E-1H). These results indicate that transcriptional expression of PINK1 is positively regulated under oxidative stress conditions.

NRF2-mediated induction of PINK1 mRNA
We next attempted to identify a possible target transcription factor(s) that up-regulates PINK1 expression. FOXO3a has been reported to control PINK1 expression [11]; however, forced expression of FOXO3a did not show any increase in luciferase activity of the delivered PINK1 pro-luc in this experimental setting (Fig 2A). This prompted us to further search for a functional sequence(s) of the PINK1 promoter. Sequence analysis showed the existence of putative four antioxidant responsive elements (ARE) in the PINK1 promoter. These ARE candidates (ARE1, -2632 to -2624 bp; ARE2, -1742 to -1734 bp; ARE3, -660 to -652 bp; ARE4, -235 to -227 bp) were matched to the general consensus sequence (TGA----GC), which is the binding site of NRF2. PINK1 pro-luc was used to determine whether the expression of PINK1 is dependent on NRF2. As a result, we found that aberrant overexpression of wild-type NRF2 (NRF2-WT) greatly increased the luciferase activity (Fig 2A). The promoter activity induced by NRF2 was approximately 20-fold higher than that induced by FOXO3a. NRF2-mediated increase in activation of the PINK1 promoter was also observed in different cell lines (S1 Fig). In different cell lines such as HCT-15 and HCT116, FOXO3a also increased the luciferase activity. Based on the homology of cross-species orthologues, NRF2 has been divided into six domains, Neh1 to Neh6 [24], and Neh4 and Neh5 are transcription activation domains [27]. When a deletion mutant of NRF2 lacking the Neh4 and Neh5 domains (NRF2-Mut) was used to assess NRF2 function as a transcriptional regulator for the PINK1 promoter, NRF2-Mut showed only slight activity compared to that of the WT counterpart ( Fig 2B). We also confirmed that PINK1 mRNA level was increased by overexpression of NRF2-WT but not by overexpression of NRF2-Mut (Fig 2C). To determine whether lack of ARE sites for NRF2 binding decreases the promoter activity, the PINK1 promoter with mutation of the ARE sites (ARE1~4) were newly constructed. As shown in Fig 2D, promoter activity of all mutant constructs was decreased compared with WT-promoter even when NRF2-WT was overexpressed. Mutation of ARE1 reduced the promoter activity with the highest level. By mutating all ARE sites (ARE1~4), the promoter activity was completely diminished (Fig 2D). To examine physical binding of NRF2 to the ARE sites, electrophoretic mobility shift assay (EMSA) was employed. With the DNA probes matched to the ARE sites (ARE1~4) in the PINK1 promoter, we found that all ARE probes was able to form complex with proteins in nuclear extracts of SH-SY5Y cells (S2A Fig). The shifted bands were of the NRF2, because supershifted bands were detected by the NRF2 antibody. We also tried to detect intranuclear binding of NRF2 to the ARE sites in the PINK1 promoter. By ChIP analysis, we found that overexpression of NRF2-WT increased the binding of NRF2 to the ARE sites in cells (S2B Fig). These results indicate that NRF2 transcriptionally up-regulates PINK1 expression via binding to the ARE sites of the PINK1 promoter.

Involvement of ROS in the NRF2-PINK1 signaling axis
We next used tert-Butylhydroquinone (tBHQ), a strong inducer of NRF2 [28], to examine whether endogenous NRF2 regulates PINK1 expression. As shown in Fig 3A, tBHQ increased the luciferase activity of PINK1 pro-luc in a dose-dependent manner. Real-time PCR analysis showed that tBHQ increased PINK1 mRNA levels from 6 h to 48 h (Fig 3B). Almost the same results were obtained using stable cell lines expressing PINK1 pro-luc (S3A and S3B Fig). Induction of PINK1 mRNA by tBHQ was also confirmed in mature neurons derived from human iPS cells (Fig 3C). To evaluate whether the induction of PINK1 mRNA by tBHQ depends on NRF2 activity, we employed three types of siRNAs against NRF2. The siRNAs used were all confirmed to have suppressive effects on NRF2 expression in both mRNA and protein levels (Fig 3D). Using the validated NRF2 siRNAs, we found that the tBHQ-induced PINK1 mRNA level was reduced to almost the basal level (Fig 3E). In protein level, PINK1 is constitutively degraded by PARL, a mitochondrial proteinase, under normal conditions, while PINK1 escapes from PARL-processing and accumulates on mitochondria under mitochondrial depolarized conditions [29]. To examine possible linking between mRNA and protein levels of PINK1, SH-SY5Y cells were treated with tBHQ and CCCP, a strong inducer of mitochondrial Prior to statistical analysis, western blotting was repeated three times (Fig 3F and S3C Fig). *, p < 0.01; **, p < 0.05. doi:10.1371/journal.pone.0142438.g003 depolarization (Fig 3F). Single tBHQ-treatment caused slight increase in PINK1 protein. Single CCCP-treatment induced strong accumulation of PINK1 protein by depolarization of mitochondria. tBHQ + CCCP supplied much higher level of PINK1 protein than that supplied by single CCCP-treatment (Fig 3F). These results indicate that NRF2-derived PINK1 mRNA expression is important to regulate the cellular PINK1 level.
We also examined other NRF2 inducers for PINK1 expression. NRF2 inducers including sulforaphane [30], caffeine [31] and curcumin [32] induced increased activity of the PINK1 promoter (S4 Fig). Since these chemical substances including tBHQ are known to inactivate KEAP1, a negative regulator of NRF2, through oxidative stress-mediated ROS production in mitochondria [33], we further investigated an event triggering activation of NRF2 and subsequent PINK1 induction. Forced expression of aberrant KEAP1 efficiently suppressed the luciferase activity of PINK1 pro-luc induced by tBHQ (Fig 4A). Treatment of the cells with tBHQ increased production of H 2 O 2 , one of the ROS, and the anti-oxidant N-acetylcysteine (NAC) quenched the produced H 2 O 2 (Fig 4B). This NAC treatment effectively suppressed tBHQinduced PINK1 expression (Fig 4C). Interestingly, in other cases of stimulation with oxidative stress inducers, 6-OHDA and CdCl 2 , the chemical-enhanced production of H 2 O 2 and PINK1 was greatly suppressed by pretreatment with NAC ( Fig 4B, 4D and 4E). These results suggest that ROS production in damaged mitochondria is a critical step for activation of NRF2, which

Cytoprotective function of PINK1 induced by NRF2
We finally examined the physiological significance(s) of PINK1 induced by NRF2 in cells. To study the cytoprotective function of the NRF2-PINK1 axis against oxidative stress-stimulated cell death, NRF2 was down-regulated using three types of siRNAs in SH-SY5Y cells. Knockdown of NRF2 markedly reduced the ratio of cell survival with 6-OHDA treatment (Fig 5A). Over-expression of NRF2-WT, but not that of NRF2-Mut, rescued the reduction of cell survival caused by 6-OHDA treatment (Fig 5B). We then examined the involvement of PINK1 in this cytoprotective machinery. We used pDNAs expressing the wild type of PINK1 (PINK1-WT) or kinase-dead type of PINK1 (PINK1-KD) for an experiment on rescue of NRF2-knockdown. Over-expression of PINK1-WT, but not that of PINK1-KD, partially rescued the reduction of cell survival under the NRF2-knockdown condition (Fig 5C). These results indicate that PINK1 induced by NRF2 has a role in the protective machinery against oxidative stress-associated cell death.

Discussion
In this study, we demonstrated that NRF2 functions as a critical transcriptional up-regulator of the PINK1 gene in response to various stresses. PINK1 transcriptionally up-regulated by NRF2 has a protective effect on mitochondria, giving cells an advantage for cell survival.
Based on the results of differential gene expression analysis of metastatic cancer cells, we previously showed that the expression levels of PINK1 were much higher in cancer cell lines with higher metastatic potential than in cancer cell lines with lower metastatic potential [34]. NRF2 expression level also positively correlates with tumor grade. A high expression level of NRF2 stimulates metastatic potential and enhances survival of cancer cells [35,36]. The present findings regarding the functional link between NRF2 and PINK1, therefore, may also be relevant to cancer progression.
The induction rate of PINK1 expression by NRF2 differed depending on the cell line (Fig  2A and S1 Fig). It has been reported that NRF2 forms heterodimers with small Maf proteins (MafF, MafG, and MafK), binds to ARE, and activates transcription of target genes [15]. The different expression patterns of small Maf proteins may affect NRF2-derived PINK1 expression, since we found a very large individual variation in the expression of Maf proteins in different cell lines (data not shown). This possible link warrants further investigations. Mei et al. reported that FOXO3a regulates PINK1 expression in T-lymphocytes upon growth factor/ serum deprivation [11]; however, the level of PINK1 expression induced by FOXO3a in SH-SY5Y cells was much lower than that induced by NRF2. We confirmed up-regulation of PINK1 expression using tBHQ, an NRF2 inducer, in SH-SY5Y and human matured neurons. These results suggest that PINK1 transcription is differentially regulated in various cell lineages in which NRF2 has a significant role in PINK1 transcription, especially in neuronal cells under stress conditions. NRF2 is an antioxidant transcription factor that positively regulates basal and inducible expression of a large number of cytoprotective genes. When cells are in a non-stressed basal state, NRF2 is sequestered in the cytoplasm by KEAP1 through binding between NRF2 and KEAP1. Under an oxidative stress condition, oxidants and electrophiles oxidize cysteine residues in KEAP1, leading to release of NRF2, allowing it to translocate to the nucleus and induce the expression of its target genes. Indeed, we found that NRF2 functions as a novel transcriptional regulator of the PINK1 gene via activation of the ARE sequence in the PINK1 promoter. This NRF2-PINK1 signaling axis plays a significant role in cell survival, because induction of oxidative stress-stimulated cell death in NRF2-down-regulated cells was rescued by overexpression of PINK1 (Fig 5C). In addition, PINK1 is involved in the maintenance of mitochondrial homeostasis through various mechanisms such as removal of damaged mitochondria and reduction of ROS. It is hence reasonable to assume that NRF2, an antioxidant transcription factor, actually regulates PINK1 gene expression. Komatsu et al. reported that the selective autophagy substrate p62 activates NRF2 through inactivation of KEAP1 [37]. This mitophagyrelated pathway may help us to understand how the NRF2-PINK1 pathway is controlled in mitochondrial homeostasis.
Accumulating evidence indicates that mitochondrial dysfunction plays a central role in the pathogenesis of PD. Mitochondrial dysfunction leads to an increase in ROS production, which is closely linked to neuronal cell death. Therefore, it is important to remove dysfunctional mitochondria and then to synthesize fresh mitochondria for appropriate maintenance of mitochondrial homeostasis. For the renewal of mitochondria, the peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α) gene is important. PGC-1α is a master regulator of mitochondrial biogenesis [38,39]. We found that this gene also contains the ARE sequence of the NRF2-binding site in its promoter region. It is hence possible that NRF2 in turn regulates mitochondrial biogenesis through up-regulation of PGC-1α under stress conditions. This pathway may compensate the lack of mitochondria after mitophagy associated with the NRF2-PINK1 signaling axis.
The above-described molecular pathways to activate NRF2 may not function properly in PD, resulting in insufficient prevention of the progression of neurodegeneration. Thus, NRF2 and the NRF2-triggerd signaling axis could be targets for PD therapy. In fact, modulation of NRF2 has been studied as a therapeutic approach in PD. Treatment with Naringenin, a natural flavonoid compound that causes a dramatic increase in the protein level of NRF2, was shown to have a strong protective effect against 6-OHDA-induced nigrostriatal dopaminergic neurodegeneration [40]. Similarly, other chemical compounds that induced high levels of NRF2 were also reported to protect against neuronal damage in a neurotoxin-induced PD model [41,42]. Thus, NRF2 has become a potential therapeutic target to prevent the progression of neurodegeneration, and the NRF2-PINK1 pathway is also expected to be an emerging target to counteract mitochondrial dysfunction and its consequences in PD.