Serine 129 Phosphorylation of α-Synuclein Induces Unfolded Protein Response-mediated Cell Death*

α-Synuclein is a major protein component deposited in Lewy bodies and Lewy neurites that is extensively phosphorylated at Ser129, although its role in neuronal degeneration is still elusive. In this study, several apoptotic pathways were examined in α-synuclein-overexpressing SH-SY5Y cells. Following the treatment with rotenone, a mitochondrial complex I inhibitor, wild type α-synuclein-overexpressing cells demonstrated intracellular aggregations, which shared a number of features with Lewy bodies, although cells overexpressing the S129A mutant, in which phosphorylation at Ser129 was blocked, showed few aggregations. In wild typeα-synuclein cells treated with rotenone, the proportion of phosphorylated α-synuclein was about 1.6 times higher than that of untreated cells. Moreover, induction of unfolded protein response (UPR) markers was evident several hours before the induction of mitochondrial disruption and caspase-3 activation. Eukaryotic initiation factor 2α, a member of the PERK pathway family, was remarkably activated at early phases. On the other hand, the S129A mutant failed to activate UPR. Casein kinase 2 inhibitor, which decreased α-synuclein phosphorylation, also reduced UPR activation. The α-synuclein aggregations were colocalized with a marker for the endoplasmic reticulum-Golgi intermediate compartment. Taken together, it seems plausible that α-synuclein toxicity is dependent on the phosphorylation at Ser129 that induces the UPRs, possibly triggered by the disturbed endoplasmic reticulum-Golgi trafficking.

ing neurons, although the mechanism that underlies LB biogenesis is poorly understood (2,3). ␣-Synuclein is a 140-amino acid protein physiologically localized in presynaptic terminals (4,5) and pathologically aggregated in hallmark inclusions, such as LB and Lewy neurites (6). Furthermore, point mutations (A53T, A30P, and E46K) or gene multiplications of ␣-synuclein were proved to be responsible for familial PD (7)(8)(9)(10). Thus, it is suggested that ␣-synuclein plays a key role in the neurodegenerative process of synucleinopathies (11). It has been demonstrated that ␣-synuclein undergoes several posttranslational modifications (5). Among them, serine 129 phosphorylation is thought to be one of the most important events (12)(13)(14)(15), because it has been reported that almost 90% of ␣-synuclein in LB is phosphorylated at serine 129 (12) and that the serine 129 phosphorylation is closely associated with aggregate formation in cellular models (16). However, it has not been clarified whether the serine 129 phosphorylation plays a critical role in the pathomechanisms of neuronal death.
Mitochondrial dysfunction has been implicated in the pathogenesis of PD and other neurodegenerative diseases (17,18). The administration of rotenone, an inhibitor of mitochondrial complex I, induced the major pathological and behavioral features of PD in animal and cellular models (19 -22). We established ␣-synuclein-overexpressing cell lines, which reproduced LB-like inclusions when exposed to ROS-inducing reagents including rotenone (21). It is still elusive whether aggregate formation is an adaptive response or is directly related to neuronal cell death (23). However, mounting evidence suggests that such protein aggregates may not directly trigger the cell death process in neurodegeneration (24,25). The aggregate formation is promoted by incorrect protein structures (26). As adaptive responses against the accumulation of misfolded or difficult to fold proteins, the unfolded protein response (UPR) regulates both protein translation and gene transcription to help the function. Thus, the UPR induction is cytoprotective, but if it fails to remedy the situation, the ER function is disrupted, and apoptosis is initiated (27). It was revealed that ␣-synuclein blocks ER-Golgi trafficking and induces UPR in a yeast model (28). Furthermore, mutations in the familial PDrelated gene, parkin, are also associated with ER-mediated UPR (29). These results suggest that the UPR pathway may constitute common pathomechanisms in nigral degeneration.
In this study using a cellular model overexpressing wild type (WT) ␣-synuclein, we showed that the appearance of UPR markers, especially activation of the PERK (PKR-like ER kinase) pathway, was evident several hours before disruption of the mitochondria. In contrast, the S129A ␣-synuclein-expressing cells showed much less UPR induction. Our data suggested that phosphorylation of ␣-synuclein at serine 129 plays a key role in ␣-synuclein-related cell death, which is triggered by UPR activation.

EXPERIMENTAL PROCEDURES
Expression Construct and Cell Culture-The WT ␣-synuclein cDNA was subcloned into the pUC18 vector at SalI and SphI sites and the S129A mutant was generated by site-directed mutagenesis (Takara LA PCR TM in vitro mutagenesis kit; Takara Biomedicals, Tokyo, Japan). The cDNAs were introduced into the eukaryotic episomal vector pCEP4 (Invitrogen). In a mock construct, the chloramphenicol acetyltransferase (CAT) gene was substituted for ␣-synuclein. The SH-SY5Y cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (Invitrogen) at 37°C in 5% of CO 2 and were transfected by these constructs using DOTAP (N-[1-(2,3-Dioleoyloxy)propyl]-N,N,Ntrimethylammonium methyl-sulfate) lipofection system (Roche Applied Science). To evaluate cell viability, transfected cells were plated at a density of 5 ϫ 10 4 cells/well in 96-well plates and allowed to grow in regular medium for 24 h, and then the cultures were exposed to 10 nM of rotenone (Sigma) for 24 to 120 h. Cell survival rates were evaluated using the 3-(4,5-dimethelthiazo-2yl)-2,5-diphenyltetrazolium bromide (MTT) assay.
Immunocytochemistry-Cells grown on poly-L-lysine coated coverslips were fixed with 4% paraformaldehyde for 20 min and permeabilized with 0.5% Triton X-100 in PBS for 10 min. After a brief wash with PBS, the cells were incubated in a blocking solution (1.5% normal goat serum in PBS) with 10 g/ml of RNase for 30 min and treated with the primary antibodies (described below) for 2 h at room temperature. After washing with PBS, the cells were incubated with fluorescent dye (Alexa 488 and/or Alexa 568)-conjugated goat anti-mouse or anti-rabbit antibodies (1:1000; Molecular Probes, Eugene, OR) for 1 h at room temperature. The nuclei were counterstained with TO-PRO3 (Molecular Probes). Fluorescent images were analyzed with a confocal laser scanning microscope (Fluoview FV300, Olympus, Tokyo, Japan). To evaluate the effect of ␣-synuclein toxicity on the mitochondrial membrane potential (⌬⌿ m ), we adopted a fluorescent indicator, JC-1 (10 mg/ml for 10min, Molecular Probes) (30). All of the images were scanned by laser scanning microscope in identical conditions of 512 ϫ 512, 12 bit/pixel resolution, the photo multiplier tube voltages of argon and krypton were 630, and the gain and the background offset were 0. At each time point, 10 -20 cells in six randomly chosen fields were analyzed to evaluate the 527:590 nm signal ratio by using the macro program in the Image-Pro Plus software (MediaCybernetics; Bethesda, MD). Intracellular ROS were detected by the 2,7-dichlorodihydrofluorescein (DCF) diacetate (Molecular Probes) as a fluorescent probe. Briefly, after washing with PBS, the cells were incubated with Dulbecco's modified Eagle's medium for 20 min. Following extensive washing in PBS, the intracellular levels of ROS were evaluated by laser scanning microscope. Four-micrometerthick, paraffin-embedded sections including the substantia nigra, locus ceruleus, and dorsal vagal nucleus from five patients with PD were subjected to immunohistochemical investigations using the avidin-biotin-peroxidase complex method with a Vectastain ABC kit (Vector, Burlingame, CA). Monoclonal antibody against ␤-COP (␤-coat protein, 1:100; Sigma) was used as primary antibody. The sections were pretreated by heating for 15 min at 121°C. Diaminobenzidine was used as the chromogen. The sections were counterstained with hematoxylin.
Immunoblot Analyses-After treatment with rotenone for the preparation of whole cell lysates, the cells were washed three times with PBS and sonicated in 100 mM Tris-HCl, 8% SDS, 1% Tween 20 containing the protease inhibitors, 1 mM phenylmethylsulfonyl fluoride, 5 g/ml of pepstain A, 5 g/ml of leupeptin, and 5 g/ml of aprotinin. Twenty microgram of total protein determined by the BCA assay (Pierce) were separated by SDS-PAGE and then transferred onto polyvinylidene difluoride membrane (Millipore Corp., Bedford, MA). The membranes were incubated with primary antibodies (described below) for 60 min. After vigorous washing, a rabbit anti-mouse IgG-horseradish peroxidase-conjugated antibody (1:2000; DAKO, Denmark) or goat anti-rabbit IgG-horseradish peroxidase-conjugated antibody (1:2000; Santa Cruz Biotechnology, Santa Cruz, CA) were used as secondary antibodies and developed by using an ECL detection kit (GE Healthcare) combined with an image analyzer (LAS-3000; Fuji Photo Film, Tokyo, Japan).
Immunoprecipitation and Casein Kinase Reaction-For immunoprecipitation, the cells were washed with TS buffer (50 mM Tris-HCl, 150 mM NaCl, pH 7.6) and harvested in radioimmune precipitation assay buffer (50 mM Tris-HCl, 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 1 mM phenylmethylsulfonyl fluoride, 5 g/ml of pepstain A, 5 g/ml of leupeptin, and 5 g/ml of aprotinin, pH 7.5). The lysates were sonicated and centrifuged 12,000 ϫ g for 5 min, and supernatants were isolated. Equal amounts of protein were incubated with 10 g of a monoclonal antibody against ␣-synuclein (Millipore; syn211) at 4°C overnight and then incubated with protein G plus agarose (GE Healthcare) at 4°C 1 h. For serine phosphorylation, immunoprecipitates were washed three times with kinase buffer (20 mM Tris-HCl, 50 mM KCl, 10 mM MgCl 2 , pH 7.5) and incubated with 1,000 units of casein kinase 2 (CK2; New England Biolabs) in the presence of 0.2 mM of ATP for 1 h at 30°C. The reaction was terminated by addition of SDS-PAGE sample buffer. The samples were boiled for 3 min and separated on a SDS-PAGE. Proteins were detected with either an anti-␣-synuclein antibody or an anti-phosphorylated ␣-synuclein antibody. For the experiment of casein kinase inhibition, the cells were incubated with rotenone in the presence of CK2 inhibitor (DMAT (2-Dimethylamino-4,5,6,7-tetrabromo-1Hbenzimidazole); calbiochem, Darmstadt, Germany, 1 M). To examine the effect of CK2 inhibitor on eIF2␣ phosphorylation, SH-SY5Y cell lines were incubated with 10 g/l of tunicamycin (Calbiochem) and 0.2-5 M of DMAT. Protein samples from these cells were analyzed by Western blots.
Real Time PCR-Total RNA samples were isolated by using the acid phenol method (RNA-bee; Tel-Test, Friendswood, TX) according to the manufacturer's instructions. Complementary DNA was synthesized from RNA samples using Super-ScriptIII (Invitrogen) containing oligo(dT) 20 . Quantitative real time PCR was run in LightCycler1.0 (Roche Applied Science).
Assay for Proteasome Activity-The proteasome activity was quantified using a 20 S proteasome assay kit (Affinity Research Product, Exeter, UK). Briefly, the cells were washed three times with PBS, resuspended into a buffer containing 50 mM Tris/ HCl, pH 7.5, 25 mM KCl, 10 mM NaCl, and 1 mM MgCl 2 and then lysed by a brief sonication. The lysates were incubated with the fluorogenic substrate, Suc-Leu-Leu-Val-Tyr-AMC, at 37°C for 30 min. The proteasome activity was detected by changes in fluorescence intensity at 355 nm of excitation and 460 nm of emission using an automatic multi-well fluorometer (Fluoroscan Ascent; Labsystems, Helsinki, Finland). The relative activity was standardized by the protein concentrations determined by the BCA method.
Statistical Analyses-The data were analyzed by one-way analyses of variance and posts hoc multiple comparison using Newman-Keul's multiple comparison test or two-way analyses of variance and posts hoc multiple comparison using Bonferroni's test on GraphPad Prism version 4.03 for Windows (Graph-Pad Software, San Diego, CA).

Low Dose and Long Time Exposure of Rotenone Preferentially Induced Intracellular Aggregates in ␣-Synuclein-overexpressing
Cells-Following the transfection of WT or S129A mutant ␣-synuclein constructs, the immunocytochemical examination demonstrated diffuse staining of ␣-synuclein in the entire cytoplasm. There were no obvious changes in the growth and morphology of these transfected cells as previously described (21). The Western blot revealed that there were no differences in the expression levels and molecular sizes between WT and S129Aexpressing cells, although the latter failed to show positive bands for anti-phosphorylated ␣-synuclein antibody (Fig. 1A). To confirm the loss of phosphorylation at serine 129 in the S129A mutant, the cells were stained by anti-phosphorylated ␣-synuclein monoclonal antibody (Pser#64) (12), and the signal intensities per cell were quantified. The signal intensities in WT ␣-synuclein-expressing cells were significantly higher than those in the other two cell lines (Fig. 1B). To evaluate amounts of phosphorylated ␣-synuclein after a short time of (6 h) exposure of rotenone, immunoprecipitation experiments were conducted using an antibody recognizing ␣-synuclein, and then a part of immunoprecipitates was incubated with CK2. The samples were analyzed on Western blots using antibody against ␣-synuclein or that of phosphorylated form. Increased amounts of phosphorylated ␣-synuclein were evident after rotenone exposure (Fig. 1C). When the amount of phosphorylated ␣-synuclein following CK2 treatment (Fig. 1C, lane 3) was defined as 100%, ϳ50% of ␣-synuclein was phosphorylated in untreated ␣-synuclein-overexpressing cells, and 80% was phosphorylated following rotenone treatment (Fig. 1D). The data were quantified by densitometry from four independent experiments of Western blots (Fig. 1D). After 120 h of treatment with rotenone (10 nM), intracellular aggregates were observed in WT ␣-synuclein-expressing cells ( Fig. 2A). The aggregates were immunopositive for anti-␣-synuclein, anti-phosphorylated ␣-synuclein, anti-ubiquitin, and anti-neurofilament antibodies (data not shown). The aggregates were also co-localized with ␤-COP and ERGIC-53, the marker for the ER-Golgi intermediate compartment (ERGIC) (31) but not with a marker for ER, lysosome, and Golgi ( Fig. 2A). Furthermore, some LBs in autopsied brain were also immunopositive for ␤-COP (Fig. 2C). After 24 h exposure, the cells expressing the S129A mutant showed much fewer aggregations compared with cells expressing WT (Fig. 2B), which was in good agreement with the previous reports (16,32). At 120 h of exposure, ϳ10% of WT cells were aggregate-positive but only 1% positive of S129A cells (Fig. 2B). The incidence of aggregates was roughly dependent on the period of rotenone exposure (Fig. 2B). S129A Mutation Ameliorates Cell Toxicity Induced by ␣-Synuclein Overexpression-Because of initial medium changes for rotenone treatments, values of MTT assay were transiently decreased during first 24 h in all cell lines. However, after that, the viability of WT ␣-synuclein-expressing cells was dramatically reduced compared with those of CAT-and S129A-expressing cells (Fig. 3A). To assess whether apoptotic processes played a role in this cellular model, cells overexpressing CAT, WT, and S129A were double-stained by anti-␣synuclein and anti-activated caspase-3 antibodies after the treatment with 10 nM of rotenone for 0 -120 h. In WT ␣-synuclein-overexpressing cells, activated caspase-3-positive cells were increased after the exposure to rotenone (Fig. 3B). The incidence of activated caspase-3-positive cells was significantly higher in the WT than in the CAT and S129A mutants after 72 h of exposure of rotenone (Fig. 3B). Thus, it is suggested that the cell death in this model was based on apoptotic processes.
Mitochondrial Dysfunction Following Rotenone Treatment-To evaluate the cell death process in this model, we investigated the mitochondrial membrane potential by the JC-1 staining method, in which red represents a normal mitochondrial membrane potential (⌬⌿ m ), and green represents decreased mitochondrial membrane potential. In each image, the green/red signal intensity ratio was calculated. In WT ␣-synuclein-expressing cells, the green/red score was much more increased after 24 h of exposure to rotenone (Fig. 4A). Consistent with the cytochemical data, immunoblots also indicated that activated caspase-3 expression was more increased in WT ␣-synuclein cells compared with CAT and S129A cells (Fig.  4B). Activated caspase-9 induction did not precede caspase-3 activation (Fig. 4C). On the other hand, DCF signal intensities were significantly elevated after 72 h of treatment of rotenone, but there were little difference in the ROS levels among the three cell lines (Fig. 4D). Only in WT ␣-synuclein-expressing cells, JNK inactivation was observed until 120 h after the rotenone exposure. However, in CAT and S129A mutant cells, the JNK activation was less prominent (Fig. 4E). The results were quantified by densitometry from three independent experiments (Fig. 4, B, C, and E).
Unfolded Protein Responses Precede Mitochondrial Dysfunction-The levels of transcripts for three target genes of the UPR were measured by real time PCR. The three target genes were MDG1/Erdj4, a specific target of the Xbp-1 pathway (33); BiP, a target shared by all UPR pathways (34,35); and CHOP, a downstream target of the PERK/activating transcriptional factor 4 pathway (34). In WT ␣-synuclein cells, the Erdj4 and CHOP mRNA levels were elevated (Fig. 5, A and B), but in CAT cells the Erdj4 mRNA levels were not elevated significantly (p ϭ 0.06) after 24 h of rotenone exposure (Fig. 5C). The induction of CHOP in WT ␣-synuclein cells was also observed in Western blot experiments (data not shown). Although the p-IRE1␣ and p-PERK expression showed few changes during the observed period (0 -120 h) (Fig. 5, D and E), higher expression of p-eIF2␣ was exhibited in the WT cells between 6 and 12 h of rote- Time (hour) Rate of viable cells FIGURE 3. Evaluation of cell viability and apoptosis. A, the SH-SY5Y cells overexpressing WT or S129A ␣-synuclein were treated with 10 nM of rotenone for 24 to 120 h. Cell viability was assessed using MTT assay, and the results were expressed as cell survival rates compared with those of no treatment controls (0 h). The values were represented as the means Ϯ S.E. (n ϭ 6). *, p Ͻ 0.001 versus both CAT and S129A. B, activated caspase-3 positive cells were counted following the rotenone treatment at 10 nM for 24 -120 h. For each sample, eight random fields were selected for counting. The results were analyzed by two-way analysis of variance test and were shown as the means Ϯ S.E. *, p Ͻ 0.05 versus both CAT and S129A.
Proteasome Activity Is Elevated in WT ␣-Synuclein-expressing Cells-Three cell samples were prepared at each time point to measure the proteasome activities. In WT ␣-synuclein cells, elevation of proteasome activities were moderate at 12 h and were significant at 24 h following 10 nM rotenone treatment (Fig. 6). On the other hand, in CAT and S129A ␣-synuclein cells, the proteasome activities showed few changes during the observed period (Fig. 6).
Casein Kinase 2 Inhibitor Blocked UPR Activation Caused by Rotenone Exposure-Cells overexpressing ␣-synuclein were cultured with rotenone and casein kinase inhibitors for 6 h.

DISCUSSION
In this study, we attempted to develop a cellular model of PD using ␣-synuclein-overexpressing cells combined with rote-none exposure at low doses (10 nM) and for a relatively longer time (up to 120 h). Under this condition, approximately half of the complex I activities were inhibited (36). High doses (100 nM) and short duration (within 24 h) exposure to rotenone induced only 2-3% aggregate-positive cells (data not shown), but low doses and long time exposure produced about 10% of aggregate-positive cells (Fig. 2B). Immunocytochemical experiments revealed that intracellular aggregates were immunopositive for ␣-synuclein, phosphorylated ␣-synuclein, ubiquitin, and neurofilament (data not shown). These results indicated that the aggregates shared a number of pathological features with LB (12,(37)(38)(39).
␣-Synuclein is modulated by several post-translational modifications (5). The serine 129 phosphorylation is thought to be one of the most important post-translational modifications (12)(13)(14). Several reports have described close relationships between serine 129 phosphorylation of ␣-synuclein and the PD pathogenesis (12)(13)(14). Several protein kinases, such as CK1, CK2, and a family of G-protein-coupled receptor kinases, have been proposed as candidates that phosphorylate ␣-synuclein (40,41). Moreover, it has been suggested that the haplotype of the G-protein-coupled receptor kinase 5 gene was related to sporadic PD (42). However, it is not clear whether serine 129 phosphorylation is an essential factor for forming LBs. Although it was reported that blocking of serine 129 phosphorylation increased inclusion formation in ␣-synuclein transgenic flies (32), there have been some claims against assessing synuclein aggregations in a Drosophila model. First, flies do not have a counterpart of ␣-synuclein because, phylogenically, the synuclein family is only seen in vertebrates. Second, the reported fly model was co-transfected with Gprk2 (40) for ␣-synuclein phosphorylation, but Gprk2 has not been proven to localize in LB (42,43). Thus, although Gprk2 can phosphorylate ␣-synuclein in vitro, it may not contribute to synuclein pathology in vivo. In our study using a mammalian cellular model, we clearly showed that the S129A mutant-expressing cells, in which the serine 129 and CHOP (C) were quantified by real time PCR and expressed as relative ratios compared with the levels of ␤-actin mRNA as controls. Three independent experiments were performed. The results were analyzed by two-way analysis of variance test and were shown as the means Ϯ S.E. *, p Ͻ 0.01 versus S129A; **, p Ͻ 0.001 versus S129A. D-F, Western blot analyses of marker proteins of ER stress were carried out using antibodies to p-IRE1␣ (D), p-PERK (E), and p-eIF2␣ (F). Protein amounts were quantified by using anti-␤-tubulin antibody as a loading control. The results were analyzed by two-way analysis of variance test and were shown as the means Ϯ S.E. (n ϭ 3). *, p Ͻ 0.01 versus CAT. **, p Ͻ 0.001 versus CAT. #, p Ͻ 0.05 versus S129A. ##, p Ͻ 0.01 versus S129A.
phosphorylation was blocked (Fig. 1, A and B), presented much fewer aggregations compared with WT cells (Fig. 2B). This result was in a good agreement with a previous study using mammalian cells (16). Moreover, we showed that the percentage of phosphorylated ␣-synuclein was increased from 50 to 80% following the rotenone treatment. It was reported in a mammalian cellular model that the expression of S129D mutant as a phosphorylated ␣-synuclein increased aggregates compared with WT cells (44). Taken together, phosphorylation of ␣-synuclein at serine 129 appears to facilitate aggregate formation in mammalian cells.
There has been little evidence showing a relationship between serine 129 phosphorylation and neuronal cell death. In this study, we demonstrated that blocking of serine 129 phosphorylation lead to decreased neuronal cell death induced by mitochondrial toxin exposure (Fig. 3, A and B). The treatment of rotenone at 10 nM was too low to cause apparent cellular death in SH-SY5Y cells, as indicated by the low activation of caspase-3 in controlled cells (Figs. 3B and 4B). Moreover, intracellular ROS elevation at the early phase was not evident in DCF staining (Fig. 4D). However, the addition of WT ␣-synuclein expression preferentially caused apoptosis even under these low dose conditions. Thus, these data suggested that the serine 129 phosphorylation was required not only for aggregate formation but also for the induction of ␣-synuclein toxicity. Because the toxic effect of ␣-synuclein was suggested to be associated with intermediate oligomer formation, so-called protofibrils (23,45), the phosphorylation of ␣-synuclein at serine 129 may be important for the formation of protofibrils. A recent study using NMR also suggested that serine 129 phosphorylation may destabilize the intramolecular interactions, converting ␣-synuclein into more unfolded forms that self-associated readily (46).
Furthermore, an animal model with chronic intravascular infusion of rotenone showed a parkinsonian phenotype and nigral degeneration pathology with cytoplasmic inclusions immunoreactive for ␣-synuclein and ubiquitin (19,22). Although the cells were treated with rotenone as an environmental insult in the present study, the dose of rotenone was not sufficient to cause cell death in mock cells. To study the intracellular ROS, DCF fluorescence and JNK activation were evaluated. Although the DCF signal intensities were not changed before 72 h of rotenone exposure (Fig. 4D), JNK activation was observed only in WT ␣-synuclein-expressing cells, which might reflect ROS generation induced by synuclein toxicity (Fig. 4E) (48). Then we evaluated the mitochondrial function using immunocytochemistry of JC-1 staining and Western blots using anti-activated caspase-9 antibodies. Although in WT ␣-synuclein-overexpressing cells decreased mitochondrial membrane potential by JC-1 staining was observed after 24 h of exposure to rotenone (Fig. 4A), activated caspase-9 expression did not precede caspase-3 activation. Thus, it is not plausible that mitochondrial dysfunction was a primary event triggering apoptosis in this cellular model.
To study other pathways of cell death, we then focused on ER stress. It was suggested that UPR is an important pathway in causing cellular death in nigral neurons of PD brain (29,49). The investigations of ER stress markers using real time PCR (Fig. 5, A-C) and Western blots (Fig. 5, D-F) revealed that UPR was activated within several hours following the rotenone exposure in WT ␣-synuclein-expressing cells. Especially, p-eIF2␣, a member of the PERK pathway family, showed remarkable induction in WT ␣-synuclein-expressing cells (Fig.  5F). This p-eIF2␣ expression was detected after 6 h of rotenone exposure (Fig. 5F) and preceded the mitochondrial disruption shown by JC-1 staining (Fig. 4A). Moreover, CHOP was also activated after the p-eIF2␣ activation (Fig. 5C), suggesting that ER stress was a trigger of apoptosis in which the PERK pathway played a key role. Recently, it was reported that p-eIF2␣-positive neurons were observed in substantia nigra in PD brain (50) and that CHOP was up-regulated in dopaminergic cells of rodent brain treated by 6-OHDA (51). Activation of eIF2␣ might be promoted by ␣-synuclein phosphorylation, because the proportion of phosphorylated ␣-synuclein was about 1.6 higher than that of untreated cells (Fig. 1D). Furthermore, we showed that CK2 inhibitor, which was known to block ␣-synuclein phosphorylation, decreased p-eIF2␣ expression induced by rotenone (Fig. 7). CK2 was thought to be major enzyme for the phosphorylation of human ␣-synuclein at serine 129 (52). Moreover, its ␤ subunits were co-localized with LB (53). The CK2 inhibition failed to suppress the induction of phosphorylated eIF2␣ by tunicamycin (Fig. 7B). Therefore, it was suggested that CK2 did not directly affect the eIF2␣ phosphorylation. Thus, it is possible that p-eIF2␣ may play an important role in the cellular pathogenesis of Parkinson disease.
The molecular mechanisms by which ␣-synuclein-induced UPR are still elusive. In the pathomechanisms of poly(Q) diseases, including Huntington's disease (54), it is generally believed that misfolding proteins trigger the UPR (55) by the robust attenuation of ER-associated degradation followed by decreased proteasome activities (27). In contrast, the proteasome activity was not decreased in the present study. Another possible explanation for the UPR demonstrated here is that the vesicular trafficking may be disturbed in this cellular model. It was reported that ␣-synuclein blocked ER-Golgi trafficking and might induce the accumulation of proteins in the ER to produce ER stress (28). The fact that intracellular aggregates were colocalized with both ␤-COP and ERGIC-53 ( Fig. 2A, panels a-c and d-f, respectively) strongly supported this hypothesis, because these marker proteins were essential molecules for the transport of proteins from ER to Golgi (56,57). Furthermore, we showed that some LBs in autopsied brain from patients with PD were immunopositive for ␤-COP (Fig. 2C). Although it has not been clarified yet how ␣-synuclein blocks ER-Golgi trafficking, serine 129 phosphorylation seemed to play a key role in promoting these processes.
It was reported that WT ␣-synuclein overexpression did not influence the proteasome activity in cellular and animal models (58 -60). However, we demonstrated here that the proteasome activities were increased in WT ␣-synuclein-overexpressing cells after exposure to a low dose of rotenone. It was possible that rotenone per se might increase the proteasome activity, because MPPϩ (1-methyl-4-phenylpyridinium), another complex I inhibitor, was shown to induce proteasome activity in cultured cells (61). However, the dose of rotenone in the present study was not sufficient to change proteasome activity, because no changes in the proteasome activities were observed in CAT and S129A ␣-synuclein-overexpressing cells under exposure to rotenone (Fig. 6). Thus, it appeared that the increased proteasome activity reflected cytoprotective responses because the proteasome activities were increased following the UPR induction (Figs. 5 and 6). A previous report using a cellular model with 6-OHDA treatment showed identical results and was in good agreement with our data (62). The in vivo proteasome activities have been evaluated in PD brain by several groups, but their results remain controversial (63,64). Because the nigral cell death of PD has already started several years before the disease onset (65), it may be difficult to assess transient events in the neurodegenerative process such as changes of proteasome activity by using post-mortem tissues from patients with PD, although some reports demonstrated no changes or decreased proteasome activity (63,64).
In summary, using ␣-synuclein-overexpressing cells exposed to a low dose of rotenone as an environmental toxin, we showed that phosphorylation of ␣-synuclein at serine 129 promoted intracellular aggregate-formation and induced ER stress that was followed by mitochondrial damage and apoptosis. These findings contribute to clarifying the pathomechanisms of PD and other related synucleinopathies, in which disturbed ER-Golgi trafficking might play a central role.