Apoptosis Signal-regulating Kinase 1 (ASK1)-p38 Pathway-dependent Cytoplasmic Translocation of the Orphan Nuclear Receptor NR4A2 Is Required for Oxidative Stress-induced Necrosis*

Background: The molecular mechanisms of p38 MAPK-mediated necrosis currently have not been well elucidated. Results: During oxidative stress, NR4A2 is phosphorylated and translocated into the cytosol in an ASK1-p38-dependent manner, which ultimately leads to the promotion of necrosis. Conclusion: ASK1-p38 MAPK pathway-dependent phosphorylation and subsequent cytoplasmic translocation of NR4A2 promote oxidative stress-induced necrosis. Significance: We found a novel intracellular signaling pathway that regulates oxidative stress-induced and p38-mediated necrosis. p38 mitogen-activated protein kinases (MAPKs) play important roles in various cellular stress responses, including cell death, which is roughly categorized into apoptosis and necrosis. Although p38 signaling has been extensively studied, the molecular mechanisms of p38-mediated cell death are unclear. ASK1 is a stress-responsive MAP3K that acts as an upstream kinase of p38 and is activated by various stresses, such as oxidative stress. Here, we show that NR4A2, a member of the NR4A nuclear receptor family, acts as a necrosis promoter downstream of ASK1-p38 pathway during oxidative stress. Although NR4A2 is well known as a nucleus-localized transcription factor, we found that it is translocated into the cytosol after phosphorylation by p38. Because the phosphorylation site mutants of NR4A2 cannot rescue the cell death-promoting activity, ASK1-p38 pathway-dependent phosphorylation and subsequent cytoplasmic translocation of NR4A2 may be required for oxidative stress-induced cell death. In addition, NR4A2-mediated cell death does not depend on caspases and receptor-interacting protein 1 (RIP1)-RIP3 complex, suggesting that NR4A2 promotes an RIP kinase-independent necrotic type of cell death. Our findings may enable a more precise understanding of molecular mechanisms that regulate oxidative stress-induced and p38-mediated necrosis.

Recently, we identified NR4A2, which is a member of the NR4A orphan nuclear receptor family, as a target molecule of p38 (6). The NR4A family consists of three members, NR4A1 (also called Nur77), NR4A2 (also called Nurr1), and NR4A3 (also called NOR-1). These proteins function as transcription factors that play important roles in various cellular functions, such as differentiation and cytokine production (7,8).
In addition to their well known functions as transcription factors, NR4As possess cell death-promoting activities. In particular, the proapoptotic function of NR4A1 has been extensively studied in T cells (9,10). Although initial studies argued that the proapoptotic activity of NR4A1 is dependent on its transcriptional activity (11), recent studies have indicated that the nuclear export of NR4A1 and NR4A3 also contributes to apoptosis (12)(13)(14). In addition, NR4A1 has been reported to be involved in necrosis, caspase-independent cell death, induced by the combined treatment of cells with LPS and pancaspase inhibitor, Z-VAD-fmk (15). Nevertheless, the cell death-promoting activity of NR4A2 in apoptosis or necrosis has not been studied.
In this study, we elucidated the roles of ASK1-p38 pathwaydependent phosphorylation of NR4A2 in oxidative stress-in-duced necrosis. Although we previously demonstrated that p38 directly phosphorylates NR4A2 and potentiates its transcriptional activity (6), here we identified that NR4A2 is translocated from the nucleus into the cytosol during oxidative stress after it is phosphorylated downstream of the ASK1-p38 pathway. The alanine substitution of the oxidative stress-induced phosphorylation sites of NR4A2 inhibits the cytoplasmic translocation and pro-necrosis activity of NR4A2 induced by oxidative stress, suggesting that p38-dependent phosphorylation of NR4A2 is a crucial regulation to confer the necrosis-promoting activity to NR4A2.
It has been reported that pretreatment with p38 inhibitor attenuates ischemia-reperfusion injury, which is a typical in vivo model of oxidative stress-induced necrosis (16), suggesting that p38 regulates not only apoptosis but also oxidative stressinduced necrosis. However, the substrates of p38 in the context of necrosis and, thus, the molecular mechanisms by which p38 promotes necrosis are largely unknown. Our findings further the understanding of the molecular mechanisms by which p38 exerts its necrosis-promoting activity during oxidative stress.
Lactate Dehydrogenase (LDH) Assay-H 2 O 2 -induced necrotic cell death was monitored using the LDH-cytotoxic test (Wako) according to the manufacturer's protocols. Released LDH activity into the culture media was quantified as a percentage of the total LDH activity.
Measurement of Caspase-3 Activity-The activities of caspase-3 were measured using a CPP32/caspase-3 fluorometric protease assay kit (Medical and Biological Laboratories) according to the manufacturer's protocols.
Immunostaining Analysis and Cell Image Analyzer-After fixation in 4% formaldehyde for 10 min, the cells were permeabilized with 1% Triton X-100 in phosphate-buffered saline for 15 min and blocked with TBST solution containing 5% skim ASK1-p38-NR4A2 Axis in H 2 O 2 -induced Necrosis milk for 30 min at room temperature. Anti-FLAG antibody (1E6, Wako) and anti-active caspase-3 antibody (ab13847, Abcam) were used as primary antibodies and visualized using Alexa Fluor 555 and 488 secondary antibodies, respectively (Molecular Probes). Hoechst 33342 solution (Dojindo) was used for nuclear staining. Images were acquired using confocal microscopy (TCS SP5, Leica). For imaging analysis using Array-Scan VTI (Thermo Scientific Cellomics), samples were immunostained as described above and analyzed. After cells were identified using the Hoechst signal of their nuclei, the localization of FLAG-NR4A2 was defined from 200 cells, using the Cellomics Cell Health Profiling V3 software.

RESULTS
The ASK1-p38 Pathway Is Required for H 2 O 2 -induced Caspase-independent Cell Death-H 2 O 2 is one of the most potent stresses to activate ASK1 (2). We previously reported that ASK1 was required for H 2 O 2 -induced cell death in various types of cells (3)(4)(5). Because it was shown that H 2 O 2 induces apoptosis and necrosis depending on its concentration (18), we began to examine which type of cell death ASK1 regulates in response to H 2 O 2 . First, we characterized the type of cell death that was induced by various concentrations of H 2 O 2 in HeLa cells (Fig. 1, A-C). As shown in Fig. 1A, immunocytochemical analysis using an active caspase-3 antibody revealed that caspase-3 activity was detected when cells were treated with 1 mM, but not 3 mM, H 2 O 2 . To further analyze the dose-dependent H 2 O 2 -induced cell death features in HeLa cells, we utilized two different quantitative cell death assays. To detect apoptosis, caspase-3 activity was measured. As a necrosis marker, we measured the levels of LDH released into the culture media, which represents disruption of the plasma membrane. As shown in Fig. 1B, caspase-3 activity was significantly increased by treatment with 1 mM, but not 3 mM, H 2 O 2 . By contrast, LDH release was induced by 3 mM, but not 1 mM, H 2 O 2 (Fig. 1C). Although we further examined caspase-3 activation every hour after 3 mM H 2 O 2 treatment, caspase-3 activation was not observed at any time point (Fig. 1D). Moreover, treatment with the pancaspase inhibitor Z-VAD-fmk did not suppress LDH release induced by 3 mM H 2 O 2 (Fig. 1E). These data indicate that, in HeLa cells, treatment with 3 mM H 2 O 2 induces caspase-independent cell death.
Next, we investigated the type of cell death in which ASK1 and its downstream targets, p38 and JNK, were involved. As shown in Fig. 1, F and G, neither siRNA treatment with ASK1 nor pretreatment with p38 or JNK inhibitors suppressed the caspase-3 activation that was induced by 1 mM H 2 O 2 . Conversely, 3 mM H 2 O 2 -induced LDH release was significantly suppressed by ASK1 knockdown (Fig. 1H). In addition, pretreatment with p38 inhibitors, but not a JNK inhibitor, suppressed 3 mM H 2 O 2 -induced LDH release (Fig. 1I). We further examined the requirement of p38 for H 2 O 2 -induced HeLa cell death by using another p38 inhibitor, PH797804, whose high potency and selectivity has been established (19). As shown in Fig. 1J 1K). Although the suppressive effects of these p38 inhibitors on LDH release were less striking compared with that of ASK1 knockdown (Fig. 1H), this might be due to different experimental conditions to suppress the activity of proteins (one is the small compound-mediated enzymatic inhibition, and the other is the siRNA-based suppression of protein expression). It would also be possible that substrates of ASK1 other than p38 may be involved in H 2 O 2 -induced LDH release.
Next, we examined whether 3 mM H 2 O 2 -induced p38 activation is ASK1-dependent. As shown in Fig. 1L, two independent siRNAs that target different regions of the ASK1 gene suppressed p38 activation, and their effects on JNK activation were different. The signal of phospho-JNK was even intensified by ASK1 siRNA 1 (Fig. 1L). These results suggest that ASK1 is required for H 2 O 2 -induced p38 activation, although the contribution of ASK1 to the 3 mM H 2 O 2 -induced JNK activation in HeLa cells is ambiguous. From these results, we concluded that the ASK1-p38 pathway is required for 3 mM H 2 O 2 -induced caspase-independent cell death, but not for 1 mM H 2 O 2 -induced apoptosis, in HeLa cells.
The ASK1-p38 Pathway Is Required for H 2 O 2 -induced RIP1-RIP3-independent Necrosis-Although it has been reported that necroptosis is not induced in HeLa cells due to low expression of RIP3 (20), we examined the effect of a RIP1 kinase inhibitor, necrostatin-1 (Nec-1), on 3 mM H 2 O 2 -induced HeLa cell death to confirm that this cell death is independent of RIP1-RIP3-dependent necrosis. As expected, treatment with Nec-1 did not inhibit 3 mM H 2 O 2 -induced LDH release ( Fig. 2A). We also address this point by using other cells in which expression of endogenous RIP3 has been confirmed and, thus, reported to be sensitive to RIP kinase-induced necroptosis (20,21). We tested the sensitivity of MEFs and a human lung adenocarcinoma cell line (A549) to H 2 O 2 -induced cell death. As a result, in both of these cells, H 2 O 2 -induced LDH release was observed even under the pretreatment with Nec-1 and Z-VAD-fmk ( Fig.  2, B and C), suggesting the existence of RIP1-RIP3-independent necrosis pathways in MEFs and A549 cells.
We next examined the potential involvement of ASK1-p38 pathway in H 2 O 2 -induced necrosis in MEFs and A549. Although p38 inhibitor did not attenuate H 2 O 2 -induced LDH release in A549 (Fig. 2E), the requirement of ASK1 and p38 for H 2 O 2 -induced LDH release was clearly observed in MEFs ( Fig.  2, D, F, and G). These results suggest that the ASK1-p38 pathway is required for oxidative stress-induced necrosis independently of RIP kinases at least in HeLa cells and MEFs but not in A549. The mechanism behind the cell type-specific requirement of the ASK1-p38 pathway for H 2 O 2 -induced cell death should be elucidated.
NR4A2 Is Required for H 2 O 2 -induced RIP1-RIP3-independent Necrosis-Next, we attempted to identify the downstream targets of p38 in the context of H 2 O 2 -induced necrosis. We previously reported that ASK1-activated p38 directly phosphorylates NR4A2, which is a member of the NR4A nuclear receptor family (6). Several recent studies have suggested that other NR4A family members, such as NR4A1 and NR4A3, exert cell death-promoting activity in addition to their canonical roles in transcription (13). Therefore, we examined the potential involvement of NR4A2 in H 2 O 2 -induced necrosis. We knocked down NR4A2 using two independent siRNAs that target different regions of the Nr4a2 gene. We confirmed the knockdown efficiency of these siRNAs using real-time PCR analysis. Treatment with NR4A2 siRNAs specifically suppressed the expression of NR4A2 without apparent effects on the expression of other NR4A family members in HeLa cells (Fig. 3A). We also confirmed the knockdown efficiency of these siRNAs on NR4A2 protein by immunoblotting in both HeLa cells and MEFs (Fig. 3, B and C). Using these siRNAs, we examined the effect of NR4A2 knockdown on 3 mM H 2 O 2 -induced LDH release. As shown in Fig. 3D, we observed that 3 mM H 2 O 2induced LDH release was significantly attenuated by NR4A2   (Fig. 2, D, F, and G), the knockdown of NR4A2 in MEFs also attenuated the H 2 O 2 -induced caspase-and RIP1-RIP3-independent cell death (Fig. 3E).
Phosphorylation of NR4A2 Is Mediated by the ASK1-p38 Pathway-Because we observed a retarded migration of the NR4A2 band by immunoblotting after H 2 O 2 , we examined the possibility that this H 2 O 2 -induced band shift of NR4A2 can be ascribed to phosphorylation by using broadly specific -protein phosphatase. As expected, H 2 O 2 -induced retarded migration of the NR4A2 band was reversed by incubation with -protein phosphatase (PPase; Fig. 4A), suggesting that NR4A2 is phosphorylated in response to H 2 O 2 . We also observed that the band shift of NR4A2 was similarly inhibited, at least in part, by pretreatment with ASK1 and p38 inhibitors (Fig. 4B). These results suggest that the H 2 O 2 -induced phosphorylation of NR4A2 is mediated by the ASK1-p38 pathway in HeLa cells.
We attempted to identify the H 2 O 2 -induced phosphorylation sites of NR4A2. There are 15 consensus phosphorylation sites (Ser-Pro or Thr-Pro) targeted by p38 in both human and mouse NR4A2. We previously divided these sites into four clusters that we designated cluster I (Ser 11 , Ser 16 , and Thr 36 ), II (Ser 126 , Thr 129 , Thr 132 , Ser 140 , Thr 168 , Ser 181 , and Thr 185 ), III (Ser 250 , and Ser 256 ), and IV (Ser 351 , Ser 356 , and Ser 359 ), as depicted in the diagram in Fig. 4C (6). To determine the phosphorylation site(s) by H 2 O 2 -activated p38, we utilized combinational alanine substitution mutants of each cluster of NR4A2. We have previously shown that the retarded migration of NR4A2 that is induced by co-expression of ASK1⌬N (a constitutively active form of ASK1 that lacks its N-terminal inhibitory domain) was suppressed in the cluster II mutant (6). We also found that the H 2 O 2 -induced, retarded migration of NR4A2 was suppressed in the cluster II mutant (Fig. 4D). Seven potential phosphorylation sites are located in cluster II, as shown in the sequence of NR4A2 in Fig. 4E. We therefore examined the effect of each alanine substitution on the H 2 O 2 -induced band shift of NR4A2. None of these single mutants showed clear suppression of the band shift by immunoblotting (Fig. 4F). To narrow down the phosphorylation sites, we tested the NR4A2 mutants by substituting each potential phosphorylation site in cluster II with Ala beginning at the N terminus of cluster II. We named these mutants as follows: 6A (Ser 126 , Thr 129 , Thr 132 , Ser 140 , Thr 168 , and Ser 181 to Ala), 5A (Ser 126 , Thr 129 , Thr 132 , Ser 140 , and Thr 168 to Ala), 4A (Ser 126 , Thr 129 , Thr 132 , and Ser 140 to Ala), and 3A (Ser 126 , Thr 129 , and Thr 132 to Ala). The H 2 O 2induced retarded migration of NR4A2 was suppressed in the 3A mutant to the similar extent with the cluster II mutant (Fig.  4G). Moreover, double alanine mutation (Ser 126 /Thr 129 , Thr 129 /Thr 132 , and Ser 126 /Thr 132 ) did not show clear effects on the band shift of NR4A2 (Fig. 4H), suggesting that Ser 126 , Thr 129 , and Thr 132 all have the potential to be phosphorylated by H 2 O 2 -induced p38, and the phosphorylation of one of the three phosphorylation sites is sufficient for the retarded migration of NR4A2.
In order to confirm the p38-dependent phosphorylation sites of NR4A2, we generated an anti-phospho-NR4A2 antibody (p-NR4A2 S126/T129/T132) raised against a phosphopeptide, KPSpSPPpTPTpTPGFQ (where pS and pT represent phosphoserine and phosphothreoine, respectively). As shown in Fig. 4I, under the H 2 O 2 -treated condition, the band of wild-type NR4A2 but not of 3A mutant was clearly detected by immunoblottingusingthisantibody,indicatingthatH 2 O 2 -inducedphosphorylation of Ser 126 , Thr 129 , and Thr 132 of NR4A2 can be successfully monitored by this antibody. Moreover, we observed that the H 2 O 2 -induced enhanced signal of the phospho-NR4A2 antibody was similarly attenuated by the pretreatment with ASK1 and p38 inhibitors but not by the pretreatment with JNK inhibitor (Fig. 4, J and K), suggesting that H 2 O 2 -induced phosphorylation of NR4A2 at Ser 126 , Thr 129 , and Thr 132 is mediated by the ASK1-p38 pathway.
H 2 O 2 -induced Cytoplasmic Translocation of NR4A2 Is Regulated by the ASK1-p38 Pathway-We investigated whether the role of NR4A2 in H 2 O 2 -induced cell death is dependent on its function as a transcriptional factor by using actinomycin D, a general inhibitor of mRNA transcription. H 2 O 2 -induced necrosis in HeLa cells is not affected by the pretreatment with actinomycin D (Fig. 5A). Therefore, we concluded that transcriptional activity of NR4A2 is not required for H 2 O 2 -induced necrosis in HeLa cells.
It has been previously reported that, in some cases, the nuclear export of NR4A1 and NR4A3 is involved in their cell death-promoting activity (13). Moreover, a recent report has suggested that NR4A2 is exported from the nucleus in response to oxidative stress (22). Because transcriptional activity was not required for H 2 O 2 -induced necrosis in HeLa cells, as shown in Fig. 5A, we speculated that cytoplasmic translocation of NR4A2 might be involved in H 2 O 2 -induced necrosis independently of its transcriptional activity. We found that NR4A2 was translocated to the cytosol after H 2 O 2 treatment (Fig. 5, B and C). Moreover, we also found that the ASK1-p38 pathway is involved in this H 2 O 2 -induced cytoplasmic translocation of NR4A2. Although some NR4A2 is still translocated to cytosol in response to H 2 O 2 , even by the treatment with ASK1 siRNA or p38 inhibitor, NR4A2 remaining in the nucleus was clearly increased under these conditions (Fig. 5, B and C). To verify the H 2 O 2 -induced cytoplasmic translocation of NR4A2 and its dependence on the ASK1-p38 pathway more clearly, we evalu-

ASK1-p38-NR4A2 Axis in H 2 O 2 -induced Necrosis
ated cytoplasmic and nucleic NR4A2 quantitatively by using a cell image analyzer (Array Scan VTI, Thermo Scientific). As shown in Fig. 5D, in the algorithm that we set up, the cell image analyzer correctly recognizes nucleic and cytoplasmic FLAG-NR4A2 (green and red, respectively). By using this algorithm, we measured the cytoplasmic and total intensity of FLAG-NR4A2 from 200 cells in each well and calculated the ratio of these parameters. As a result, the ratio of the cytoplasmic intensity of NR4A2 to the total intensity of NR4A2 was increased after H 2 O 2 treatment (Fig. 5, E and F). In addition, consistent with immunocytochemistry data (Fig. 5, B and C), ASK1 knockdown and pretreatment with p38 inhibitor significantly reduced the ratio of these parameters (Fig. 5, E and F). We also confirmed the H 2 O 2 -induced cytoplasmic translocation of NR4A2 and its dependence on ASK1-p38 pathway biochemically. As shown in Fig. 5G, cytoplasmic fraction of NR4A2 was increased after H 2 O 2 treatment and reduced by the pretreatment with p38 and ASK1 inhibitor.
From these results, we concluded that the ASK1-p38 pathway is required for efficient cytoplasmic translocation of  APRIL

ASK1-p38-NR4A2 Axis in H 2 O 2 -induced Necrosis
NR4A2 in response to H 2 O 2 . As for the partial suppressive effects of ASK1 siRNA or p38 inhibitor on cytoplasmic translocation of NR4A2, it is conceivable that the efficiency of siRNA or inhibitors may not be sufficient or that certain regulators other than the ASK1-p38 pathway may also contribute to NR4A2 nuclear export. Because it has been reported that NR4A2 has two NESs and is exported from nucleus to cytosol in a chromosome region maintenance 1 (CRM1/exportin 1)-dependent manner (22), we examined whether H 2 O 2 -induced cytoplasmic translocation of NR4A2 is CRM1-dependent by using leptomycin B, a CRM1 inhibitor (Fig. 5H), or by mutating the NES of NR4A2 (Fig. 5I). The pretreatment with leptomycin B reduced the cytoplasmic translocation of p38-regulated/activated kinase (PRAK/MAP-KAPK5), which has been known to be exported from nucleus to cytoplasm in a CRM1-dependent manner (Fig. 5H). However, H 2 O 2 -induced cytoplasmic translocation of NR4A2 was not suppressed (Fig. 5H). In addition, NES-disrupted NR4A2 mutant was also translocated from nucleus to cytoplasm after H 2 O 2 treatment similarly to NR4A2 WT in both cellular fractionation and immunocytochemistry analyses (Fig. 5, J and K, respectively). These results suggest that H 2 O 2 -induced cytoplasmic translocation of NR4A2 occurs in a CRM1-independent and unknown nuclear export mechanism.
Alanine Substitution of Ser 126 , Thr 129 , and Thr 132 Reduces H 2 O 2 -induced Cytoplasmic Translocation of NR4A2-As shown in Fig. 4, J and K, we identified Ser 126 , Thr 129 , and Thr 132 ofNR4A2asH 2 O 2 -inducedASK1-p38pathway-dependentphosphorylation sites, which are successfully monitored by the p-NR4A2 S126/T129/T132 antibody using immunoblotting. Consistent with these results, we observed that the signal of the p-NR4A2 S126/T129/T132 antibody was increased by H 2 O 2 treatment also in immunocytochemistry analysis (Fig. 6A, top  panels). The specificity of this antibody to phospho-NR4A2 in immunocytochemistry analysis was confirmed, because the signal of this antibody was suppressed by alanine substitution at Ser 126 , Thr 129 , and Thr 132 (Fig. 6A, bottom panels). Interestingly, the signal of this antibody was enhanced not only in the nucleus but also in the cytosol in response to H 2 O 2 treatment (Fig. 6A, top panels), suggesting that NR4A2 is phosphorylated by the ASK1-p38 pathway under oxidative stress conditions and is subsequently translocated from the nucleus to the cytosol.
In order to further confirm the requirement of ASK1-p38 pathway-dependent phosphorylation of NR4A2 for H 2 O 2 -in-duced cytoplasmic translocation of NR4A2, we examined whether cytoplasmic translocation of NR4A2 could be suppressed by the alanine substitution of ASK1-p38 pathway-dependent phosphorylation sites. As we mentioned above, the ASK1-p38 pathway was required for efficient cytoplasmic translocation of NR4A2 in response to H 2 O 2 (Fig. 5, B, C, E, F,  and G). Consistent with these data, H 2 O 2 -induced cytoplasmic translocation of NR4A2 was suppressed by alanine substitution of ASK1-p38 pathway-dependent phosphorylation sites, although a small population of these mutants still translocated into the cytosol (Fig. 6B). To confirm the suppression of cytoplasmic translocation of NR4A2 by alanine substitution more clearly, we measured the cytoplasmic and nucleic NR4A2 quantitatively using a cell image analyzer in the same way that is shown in Fig. 5, D-F. The H 2 O 2 -induced cytoplasmic translocation of NR4A2 was significantly suppressed by the cluster II mutant but not by the cluster I, III, or IV mutant (Fig. 6C). In addition, the 3A mutant was also refractory to H 2 O 2 -induced cytoplasmic translocation (Fig. 6D). These findings suggest that ASK1-p38 pathway-dependent phosphorylation of NR4A2 at Ser 126 , Thr 129 , and Thr 132 is required for the efficient cytoplasmic translocation of NR4A2 in response to oxidative stress.
Phosphorylation of NR4A2 at Ser 126 , Thr 129 , and Thr 132 Is Required for H 2 O 2 -induced Necrosis-Finally, we examined whether ASK1-p38 pathway-dependent phosphorylation of NR4A2 is required for the H 2 O 2 -induced necrosis. We performed rescue experiments to examine whether the inhibitory effect of NR4A2 knockdown on H 2 O 2 -induced necrosis could be reversed by expressing a siRNA-resistant construct of NR4A2 and its alanine mutants (cluster II mutant and 3A mutant). These constructs were successfully expressed, even when cells were treated with NR4A2 siRNAs (Fig. 6E). H 2 O 2induced LDH release was significantly rescued by the expression of wild-type NR4A2 but not by the cluster II mutant (Fig.  6F). Furthermore, the expression of the 3A mutant did not rescue H 2 O 2 -induced LDH release (Fig. 6G). These findings suggest that ASK1-p38 pathway-dependent phosphorylation of NR4A2 at Ser 126 , Thr 129 , and Thr 132 is also required for H 2 O 2induced necrosis as well as H 2 O 2 -induced efficient cytoplasmic translocation.

DISCUSSION
In this study, we observed that NR4A2 is translocated from the nucleus to the cytoplasm in an ASK1-p38 pathway-depen-   dent manner in response to oxidative stress. Because the ala-ninesubstitutionofASK1-p38pathway-dependentNR4A2phosphorylation sites reduces the cytoplasmic translocation of NR4A2, the ASK1-p38 pathway-dependent phosphorylation of NR4A2 appears to be required for the efficient translocation of NR4A2 into the cytosol (Fig. 6, B-D). Moreover, the alanine substitution of these phosphorylation sites did not rescue the necrosis-promoting activity of NR4A2 under oxidative stress, suggesting that the ASK1-p38 pathway-dependent phosphorylation of NR4A2 is also required for the induction of necrosis (Fig. 6, F and G). Taking these results together, we proposed the model of ASK1-p38-NR4A2 axis-regulated H 2 O 2 -induced necrosis as shown in Fig. 6H. Oxidative stress-induced phosphorylation of NR4A2 induces the efficient cytoplasmic translocation of NR4A2 in an ASK1-p38 pathway-dependent manner, and this localization change of NR4A2 may play important roles in necrosis induction. Because this cell death was not inhibited by RIP kinase inhibitor, ASK1-p38-NR4A2 axis-regulated H 2 O 2 -induced necrosis is RIP1-RIP3-independent.
Although the involvement of p38 in necrosis has been implicated (16), the molecular mechanisms by which p38 promotes necrosis have been largely unknown. Our findings may provide insights into the molecular mechanisms of p38-mediated necrosis.
NR4A2 is a member of the NR4A orphan nuclear receptor family (23), whose well known function is transcriptional regulation. Because NR4A family members lack intrinsic ligand binding and classical trans-activation domains that are common to other nuclear receptor families, NR4A family members appear to activate gene expression in a constitutive, ligand-independent manner (24). Therefore, it has been considered that post-translational modifications, such as phosphorylation, may be important for the transcriptional activity of NR4As. We previously reported that p38 directly phosphorylates unidentified sites in cluster II of NR4A2 and activates its transcriptional activity (6). In this study, we identified Ser 126 , Thr 129 , and Thr 132 as phosphorylation sites that are required for its efficient cytoplasmic translocation (Fig. 6, B and D) and subsequent necrosis induction (Fig. 6G). Interestingly, these sites also resided in cluster II (Fig. 4E). Therefore, it is conceivable that the transcriptional activity of NR4A2 is up-regulated by p38dependent phosphorylation under oxidative stress conditions. However, pretreatment with a transcription inhibitor, such as actinomycin D, did not inhibit H 2 O 2 -induced necrosis (Fig.  5A). Although the reason for p38-dependent phosphorylation within the same region in NR4A2 having different roles remains to be elucidated, the above mentioned result suggests that NR4A2 induces necrosis at least in a transcription-independent manner.
Although our experiments using the NR4A2 alanine mutants suggest that phosphorylation of Ser 126 , Thr 129 , and Thr 132 is required for efficient cytoplasmic translocation of NR4A2 (Fig.  6, B-D), the overexpression of NR4A2 aspartic acid mutant (NR4A2 3D) showed neither cytoplasmic localization nor necrosis-promoting activity under steady-state conditions (data not shown). However, we cannot exclude the possibility that the NR4A2 3D mutant does not act as a phosphomimetic mutant in this case. For a long time, necrosis has been regarded as an accidental and uncontrollable cell death that does not involve intracellular signaling, unlike apoptosis. However, it has now been recognized that necrosis can also occur in a regulated manner (25). Increasing evidence has suggested that H 2 O 2 -induced necrosis is also a regulated process. For example, cyclophilin D and its interaction partner p53 mediate H 2 O 2 -induced necrosis by regulating mitochondrial permeability transition pore opening (26 -28). In another report, poly(ADP-ribose) polymerase 1 overactivation and apoptosis-inducing factor nuclear translocation were found to be important processes for the induction of necrosis under oxidative stress (29). Because the suppression of NR4A2 did not completely attenuate the H 2 O 2 -induced necrosis, it is possible that not only NR4A2 but also the other regulators mentioned above might be involved downstream of ASK-p38 pathways. Moreover, to understand the mechanisms by which cytoplasmic NR4A2 is exported from the nucleus in response to H 2 O 2 -induced necrosis, it will also be important to investigate the relationship between cytoplasmic NR4A2 and the above mentioned necrosis regulators.
Recently, cytoplasmic localization of NR4A2 has been observed in some pathologic conditions, such as in a brain ischemic reperfusion injury model (30) and in bladder cancer tissue (31). Although the involvement of the necrosis-promoting activity of cytoplasmic NR4A2 in these models is not known, we hope that our findings will lead to the discovery of new therapeutic targets for these disease conditions.