The Active Site Cysteine of the Proapoptotic Protein Glyceraldehyde-3-phosphate Dehydrogenase Is Essential in Oxidative Stress-induced Aggregation and Cell Death*

Recent studies have revealed that the redox-sensitive glycolytic enzyme, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), is involved in neuronal cell death that is triggered by oxidative stress. GAPDH is locally deposited in disulfide-bonded aggregates at lesion sites in certain neurodegenerative diseases. In this study, we investigated the molecular mechanism that underlies oxidative stress-induced aggregation of GAPDH and the relationship between structural abnormalities in GAPDH and cell death. Under nonreducing in vitro conditions, oxidants induced oligomerization and insoluble aggregation of GAPDH via the formation of intermolecular disulfide bonds. Because GAPDH has four cysteine residues, including the active site Cys149, we prepared the cysteine-substituted mutants C149S, C153S, C244A, C281S, and C149S/C281S to identify which is responsible for disulfide-bonded aggregation. Whereas the aggregation levels of C281S were reduced compared with the wild-type enzyme, neither C149S nor C149S/C281S aggregated, suggesting that the active site cysteine plays an essential role. Oxidants also caused conformational changes in GAPDH concomitant with an increase in β-sheet content; these abnormal conformations specifically led to amyloid-like fibril formation via disulfide bonds, including Cys149. Additionally, continuous exposure of GAPDH-overexpressing HeLa cells to oxidants produced disulfide bonds in GAPDH leading to both detergent-insoluble and thioflavin-S-positive aggregates, which were associated with oxidative stress-induced cell death. Thus, oxidative stresses induce amyloid-like aggregation of GAPDH via aberrant disulfide bonds of the active site cysteine, and the formation of such abnormal aggregates promotes cell death.

In both prokaryotic and eukaryotic cells, glyceraldehyde-3phosphate dehydrogenase (GAPDH 2 ; EC 1.2.1.12) plays a central role in glycolysis, catalyzing the reversible conversion of glyceraldehyde-3-phosphate to 1,3-bisphosphoglycerate in a reaction that is accompanied by the reduction of NAD ϩ to NADH. Mammalian GAPDH is a homotetramer composed of four identical subunits. Recent studies show that mammalian GAPDH has diverse activities unrelated to its glycolytic function (1,2), including roles in membrane fusion, microtubule bundling, nuclear RNA transport (2), regulation of Ca 2ϩ homeostasis (3), and transcription (4). In addition to the various functions of GAPDH described above, particular attention is paid to its role in apoptosis (5)(6)(7). Although the proapoptotic role(s) of GAPDH seems to depend upon its accumulation in the particulate fractions, including the nucleus (5,7), the detailed mechanism is still unclear.
Recently, it has been suggested that a wide variety of neurodegenerative diseases are characterized by the accumulation of intracellular and extracellular protein aggregates (8,9). An initializing event in protein aggregation is thought to be the formation of an abnormal oligomer. For instance, ␤-amyloid and ␣-synuclein undergo conformational changes in Alzheimer disease and Parkinson disease, respectively, thereby acquiring a predominantly ␤-sheet structure that facilitates oligomer formation and subsequent amyloid fibril formation (10,11). Formation of abnormal oligomers and amyloid fibrils in brain is likely to be induced by a variety of causes, including oxidative stress (8,9,12). It is also reported that an increase of oxidative stress in brain is well correlated with the progression of neurodegenerative diseases (13)(14)(15). Both biochemical and immunohistochemical studies have revealed the presence of oxidized proteins in some neurodegenerative diseases (16,17), suggesting a role for oxidants in neurodegeneration. Therefore, the abnormal oligomer and the resultant fibrils formed by oxidative stress might be an important component in the pathogenesis of these diseases. Indeed, oxidative modification of ␣-synuclein by either hydrogen peroxide (H 2 O 2 ) or nitric oxide (NO) facilitated the formation of ␣-synuclein oligomers and aggregates, leading to neuronal cell death (18 -21).
Protein cysteine residues are highly susceptible to various types of oxidation (22). Most mammalian GAPDHs have either three or four cysteines, including one at the active site (Cys 149 ), which are readily oxidized to sulfenic acid or sulfonic acid (23)(24)(25)(26). Cysteine residues are also reported to play critical roles in protein aggregation, possibly initiated by the formation of stable cross-linked oligomers (27)(28)(29). These facts lead us to the hypothesis that oxidative stress may convert the normal conformation of GAPDH to an abnormal one via modification of the cysteines, resulting in aggregation. Indeed, GAPDH is deposited as insoluble aggregates in some neurodegenerative diseases, e.g. in senile plaques and neurofibrillary tangles in Alzheimer disease (30 -32) and inclusion bodies (so-called Lewy bodies) in Parkinson disease (33). More recently, disulfide-bonded aggregates of GAPDH were observed in detergentinsoluble extracts from Alzheimer disease subjects (34). Thus, in this study, the molecular mechanisms underlying the oligomerization and aggregation of GAPDH induced by oxidants are investigated using purified GAPDH and in vitro oxidative stress. In addition, we analyzed whether aggregates of GAPDH possessed amyloid-like properties. Furthermore, we prepared GAPDH-overexpressing HeLa cells to investigate whether oxidative stress also produces abnormal GAPDH aggregates in the cells and affects the cell death. Here we report that oxidative stresses induce oligomerization and aggregation of GAPDH through aberrant disulfide bonds of the active site cysteine, readily leading to the formation of insoluble and amyloid-like aggregates. The implications of these results on cell death are discussed.
Cloning of Rabbit and Human GAPDH-For the cloning of rabbit or human GAPDH, total RNA was isolated using an RNeasy TM mini kit from skeletal muscle of New Zealand White rabbit or human neuroblastoma cells SH-SY5Y. One microgram of total RNA was reverse-transcribed at 37°C for 1 h using Omniscript TM reverse transcriptase with oligo(dT) primers, to produce each of the cDNAs. The sequences of the primers for amplification of the full-length GAPDH cDNA (termed wild type) were as follows: for rabbit, 5Ј-GTCCCCGAGCTCATG-GTGAAGGTC-3Ј (forward) and 5Ј-GTGGGTTGGTACCTT-ACTCCTTGGA-3Ј (reverse); for human, 5Ј-GGCCGCGAGC-TCATGGGGAAGGTGAA-3Ј (forward) and 5Ј-CCTCTAGA-GGTACCTTACTCCTTGGAG-3Ј (reverse). The boldface letters in the sequences were designed for SacI and KpnI restriction sites, for the forward and reverse primers, respectively. The PCR products were digested, purified, and cloned into the pBAD-HisA bacterial expression vector with 5Ј-His 6 , according to the manufacturer's protocol (Invitrogen). Alternatively, GAPDH cDNAs were reamplified using the bacterial constructs as a template, and cloned into the pFLAG-CMV2 (for rabbit) or pEGFP-C1 (for human) mammalian expression vector according to the manufacturer's protocol. The sequences of the primers were as follows: for pFLAG-CMV2 construct, 5Ј-GCTCTCCAGCTGAATTCTATGGTGAAGG-TCGGA-3Ј (forward) and 5Ј-TATGATCAGTTAGTCGACT-TACTCCTTGGAGGC-3Ј (reverse); for pEGFP-C1 construct, 5Ј-GCTCAAGCTTCGAATTCATGGGGAAGGTGAAG-3Ј (forward) and 5Ј-CTGGTGGTCCAGTCGACTTACTCCTT-GGAG-3Ј (reverse). The boldface letters in the sequences were designed as EcoRI and SalI restriction sites for the forward and reverse primers, respectively. All constructs were transformed into Escherichia coli DH5␣ (TOYOBO, Tokyo, Japan), and then several colonies were grown in LB broth plus 100 g/ml ampicillin or 30 g/ml kanamycin. The plasmids were purified by a QIAfilter TM plasmid midi kit, and the complete sequences were confirmed using a 373A DNA sequencer (PerkinElmer Life Sci-ences). Comparisons with data from rabbit spleen GAPDH cDNA (GenBank TM accession number L23961) revealed replacements in the present GAPDH cDNA at two positions, 271 and 860. These replacements produced substitutions at E91Q and A286D. The sequence of the cloned cDNA of human GAPDH was completely identical to that reported (GenBank TM accession number M33197).
Expression and Purification of Recombinant GAPDH-For the expression of recombinant GAPDH proteins, expression vectors for the GAPDH gene were transformed into the gap(Ϫ) E. coli strain W3CG (generously provided from Prof. Pluckthun and Dr. Lindner at Zurich University) (36). The transformants were cultured for 2 h at 37°C in M63 minimal medium containing both 50 g/ml ampicillin and 15 g/ml tetracycline, and then 0.2% (w/v) L-(ϩ)arabinose was added to the medium. After 24 h, the cells expressing recombinant proteins were collected by centrifugation (3,000 ϫ g, 15 min at 4°C) and resuspended in lysis buffer containing 50 mM sodium phosphate (pH 8.0), 300 mM NaCl, 30 mM imidazole, 10% glycerol, and 2 mM 2-mercaptoethanol (2-ME). The suspensions were sonicated on ice and centrifuged at 15,000 ϫ g for 30 min (4°C). The supernatants were incubated with nickel-nitrilotriacetic acid-agarose resin (50% slurry) for 2 h at room temperature with rocking. The resin was washed with 50 mM phosphate buffer (pH 8.0) containing 300 mM NaCl, 50 mM imidazole, 10% glycerol, and 2 mM 2-ME. The proteins bound to the resin were eluted with 50 mM phosphate buffer (pH 8.0) containing 300 mM NaCl, 300 mM imidazole, 10% glycerol, and 2 mM 2-ME, and the eluates were immediately mixed with 1 mM NAD ϩ , 1 mM DTT, and 1 mM EDTA, followed by incubation at 4°C overnight. The reduced proteins were directly loaded onto a Hi-Load 16/60 Superdex prep-grade column equilibrated with buffer-G2Ј. The fractions containing GAPDH were pooled and concentrated using Amicon Ultra-15 (Millipore Japan, Tokyo, Japan). The protein concentrations were determined spectrophotometrically assuming a ⑀ 0.1% at 280 nm ϭ 1.0.
Treatment of GAPDH with Oxidative Stress or a Cross-linker-Native or recombinant GAPDH (0.6 mg/ml) was treated with various concentrations of NOR3, a NO donor, in buffer-G2Ј at 37°C for varying periods of time. NOR3 is stable in dimethyl sulfoxide (Me 2 SO) and decays with first-order kinetics to release NO in an aqueous buffer at physiological pH. To investigate the effects of other NO donors (SNGO and SNAP) or oxidizing reagents (SIN-1 and H 2 O 2 ) on GAPDH, these oxidants (10 M concentration) were evaluated in the same manner as NOR3. After treatment, the reactions were terminated by passing through a MicroSpin TM G-25 column (800 ϫ g, 2 min) equilibrated with buffer-G2Ј. The samples were subjected to the following experiments. Bis-maleimidohexane (BMH), an irreversible sulfhydryl to sulfhydryl cross-linking reagent, was freshly prepared as a 36 mM stock solution in Me 2 SO. BMH (1 mM) was added to 100 l of solution containing GAPDH (1 mg/ml), 20 mM sodium phosphate buffer (pH 7.4), 150 mM NaCl, and then the mixture was incubated at 4°C in the dark. After 40 min, the reaction was quenched by 20 mM DTT.
SDS-PAGE-To prevent unmodified thiols from oxidation during electrophoresis, the solutions of GAPDH were further incubated with 1 mM iodoacetamide for 10 min at room temperature in the dark. After the alkylation of thiols, the samples were mixed with an equal volume of SDS-sample buffer containing 250 mM Tris-HCl (pH 6.8), 2% SDS, 30% glycerol, 0.01% bromphenol blue in the presence (reduced) or the absence (nonreduced) of 100 mM DTT, and then heated at 100°C for 5 min. These samples were separated by 5-20% SDS-PAGE (DRC, Tokyo, Japan), and the gels were stained using Gel-Code Blue staining reagents.
In Vitro Turbidity of Solutions of GAPDH and Other Enzymes Containing the Reactive Cysteines-Native or recombinant GAPDH (0.6 mg/ml) was treated with NOR3 as described above, but the period of incubation with NOR3 was varied (1-48 h). To measure the turbidity of solutions derived from aggregated GAPDH, the absorbance at 405 nm was recorded using a VERSA Max microplate reader (Molecular Devices). To examine the effect of Congo red on the turbidity, preincubation with Congo red at a concentration of 200 M was performed. Alternatively, the reaction mixtures were centrifuged (22,000 ϫ g, 30 min), and the supernatants (soluble GAPDH) were collected. The pellets (insoluble GAPDH) were washed twice, resuspended in an equal volume of buffer-G2Ј, and sonicated on ice for 10 s. The samples were subjected to either reducing or nonreducing SDS-PAGE as described above. The turbidities of other enzymes (alcohol dehydrogenase, malate dehydrogenase, and aldolase) containing active site cysteines were also measured in the same manner.
Circular Dichroism-After GAPDH (0.6 mg/ml) was treated with a control or 100 M NOR3 for 1 h at 37°C, the reaction mixture was desalted with a NAP-5 column, and the eluate was centrifuged at 22,000 ϫ g for 30 min, providing cleared supernatant. The CD spectrum of GAPDH was measured with a spectropolarimeter model J-820 (Jasco, Tokyo, Japan). The temperature of the solutions in the cuvette was controlled at 37°C by circulating water. The path length of the optical quartz cuvette was 1.0 mm for far-UV CD measurements at 200 -250 nm and 10 mm for near-UV CD measurements at 250 -300 nm. GAPDH was dissolved in buffer-G2Ј at a concentration of 0.1 to 0.2 mg/ml. Spectra were obtained as the average of 10 succes-sive scans with a bandwidth of 2.0 nm. The data were expressed as molar residue ellipticity ().
Thioflavin-S Binding-dependent Fluorescence-This assay was carried out as reported previously (37). Thioflavin-S was dissolved in 20 mM Mops (pH 6.5) at a concentration of 3.2 mg/ml, filtered, and stored at Ϫ30°C. Stock solutions were diluted 50-fold with same buffer before use and kept in the dark. Fifty l of GAPDH sample (see above section) was mixed with 450 l of thioflavin-S solution, and the fluorescence intensity was measured at an excitation wavelength of 450 nm and an emission of 482 nm using a fluorescence spectrophotometer F-2000 (Hitachi, Tokyo, Japan).
Congo Red Birefringence-This assay was carried out as reported previously (38). Aliquots (100 l) of GAPDH (0.6 mg/ml), as is or treated with 100 M NOR3 for 24 h at 37°C, were added to 900 l of Congo red solution (25 g/ml in PBS). This mixture was incubated for 30 min at room temperature and then centrifuged at 15,000 ϫ g for 30 min. The pellet was resuspended in 100 l of milliQ water. The drops of this suspension were allowed to dry on a slide glass for 10 min. The birefringence was determined with an Eclipse LV100POL microscope equipped with a polarizing stage (Nikon, Tokyo, Japan).
Electron Microscopy-Aliquots of GAPDH solution were placed onto 200-mesh copper grids covered by carbon-stabilized Formvar film. After 10 min of incubation in a humid chamber at room temperature, the samples were negatively stained with 2% uranyl acetate. The specimens were observed using a JEOL 1200 CX transmission electron microscope (Tokyo, Japan) operated at 100 kV.
Cell Culture, Transfection, and Treatment with NOC18-HeLa cells (human carcinoma) were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and antibiotics/antimycotics at 37°C in 5% CO 2 . For transfection of GAPDH cDNA, ϳ5 ϫ 10 5 HeLa cells were seeded in a 6-cm dish. After 24 h, the cells were transiently transfected with 1 g of pFLAG-CMV2-GAPDH plasmid using Effectene (Qiagen) or HilyMax (Dojindo) transfection reagent according to the manufacturer's protocols. To investigate the effect of oxidative stress, the cDNA-transfected cells were treated with either a control (0.1 N NaOH) or NOC18 at the indicated concentrations and maintained for a further 48 h.
Cell Viability-HeLa cells were seeded at 2 ϫ 10 4 /well in a 96-well plate, transfected, and treated with NOC18 in the same manner as described above. The cell viability was assessed by an AQueous One Solution according to the manufacturer's protocol (Promega).
Subcellular Fractionation-After treatment with either a control or NOC18, cells were washed twice with PBS and then incubated for 5 min in ice-cold PBS containing 40 mM iodoacetamide to protect unmodified thiols from oxidation during fractionation. All the subsequent steps were performed at 4°C. Cells were scraped in 500 l of buffer-A containing 10 mM Tris-HCl (pH 7.5), 10 mM NaCl, 3 mM MgCl 2 , 0.05% Nonidet P-40, 0.5% Triton X-100, 40 mM iodoacetamide, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor mixture. After 10 min, the suspensions of cells were mixed vigorously for 15 s, and the aliquots were collected as a total cell lysate. The remain-ing lysates were centrifuged at 800 ϫ g for 5 min, and the pellets were collected. The pellets were then resuspended in 200 l of buffer-B containing 10 mM Hepes-KOH (pH 7.4), 25 mM NaCl, 3 mM MgCl 2 , 300 mM sucrose, 40 mM iodoacetamide, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor mixture, and washed three times by centrifugation (3,000 ϫ g, for 10 min) followed by suspension in buffer. Following addition of 100 l of buffer-B, the pellets were sonicated for 30 s, and finally obtained as particulate fractions (7). All samples were stored at Ϫ80°C until use. Protein concentrations of the samples were determined by the Bradford assay (Bio-Rad).
Immunofluorescent Microscopy-After treatment with either a control or NOC18, the transfected cells were washed twice with PBS and fixed with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for 10 min at room temperature. The cells were incubated with 5% bovine serum albumin (BSA) in PBST for 10 min at room temperature to block nonspecific binding and permeabilized with PBS containing 0.1% Triton X-100 for 5 min. The cells were then incubated with an anti-FLAG mAb (1:1000) in 1% BSA/PBST overnight at 4°C. After washing with PBST, the specific signal of FLAG tag was visualized by staining the cells with a Cy3-conjugated goat anti-mouse IgG antibody (1:5000) using a confocal scanning microscope (C1si-TE2000-E; Nikon, Tokyo, Japan). The fixed cells were also stained with 0.02% thioflavin-S as described previously (33). For nuclear staining, the cells were labeled with 4Ј,6-diamidino-2phenylindole dihydrochloride (1 g/ml) for 10 min. For quantification of cells with aggregates, six microscopic fields were selected at random, and the percentage of cells with aggregates was measured among at least 500 transfected cells.
Western Blotting-Both total cell lysates and fractionated proteins were mixed with an equal volume of SDS-sample buffer containing 0.25 M Tris-HCl (pH 6.8), 2% SDS, 30% glycerol, 0.01% bromphenol blue in the presence (reduced) or the absence (nonreduced) of 100 mM DTT, and then heated at 100°C for 5 min. These samples were separated by 5-20% SDS-PAGE and transferred to a nitrocellulose membrane (Bio-Rad). The membranes were incubated for 1 h with 5% BSA in PBST (0.05% Tween 20 and 0.02% NaN 3 in PBS) to block nonspecific binding. The membrane was then incubated for 2 h at room temperature with an anti-GAPDH mAb (1:300), an anti-FLAG mAb (1:490), or an anti-EGFP mAb (1:1000), followed by incubation for 1 h at room temperature with a peroxidase-conjugated affinity-purified secondary antibody (Zymed Laboratories Inc.). The membranes were also reprobed with an antihistone H2B pAb (1:5000). Detection was performed using ECL plus and HyperFilm, according to the manufacturer's protocol (Amersham Biosciences). The intensity of the bands was measured using Scion image software.

Oxidative Stress Induces Oligomerization of GAPDH through
Formation of Intermolecular Disulfide Bonds-GAPDH runs as a band of ϳ36 kDa upon nonreducing SDS-PAGE (Fig. 1A, lane  1). Treatment of GAPDH with NOR3, a NO donor, at a concentration of 10 M for 10 min at 37°C resulted in formation of three extra bands corresponding roughly to a dimer (ϳ66, 68, and 76 kDa) (Fig. 1A, lane 2). Upon SDS-PAGE under reducing conditions, the NOR3-induced dimer was not observed (Fig.  1A, lane 4). Treatment of GAPDH with 1 mM BMH, an irreversible sulfhydryl to sulfhydryl cross-linking reagent, resulted in formation of the dimer, despite the presence of DTT (Fig. 1A,  lane 5). These results suggest that NOR3 causes dimerization of GAPDH through formation of intermolecular disulfide bonds.
It is well known that NO donors such as SNGO and SNAP, peroxynitrite donors such as SIN-1, and hydroxyl radical donors such as H 2 O 2 inhibit GAPDH activity (23)(24)(25)(26). We further examined whether these oxidants affected the electrophoretic mobility of GAPDH under nonreduced conditions. At a concentration of 10 M, all of the oxidants induced dimerization (Fig. 1B). Treatment with either NO gas or sodium nitroprusside also induced the formation of moderate amounts of GAPDH dimer (data not shown). These results suggest that the disulfide bonding of GAPDH subunits can be induced not only by NO but also by other reactive oxygen species, i.e. oxidative stress. Thus, the effect of NOR3 as oxidative stress on GAPDH was further evaluated in the following experiments.
The concentration-dependent effect of NOR3 on the formation of oligomerization of GAPDH was examined next (Fig. 1C). Under nonreducing conditions, treatment of GAPDH with NOR3 at concentrations ranging between 1 and 10 M resulted in formation of the dimer (average mass ϭ 70 kDa) as noted above. In addition, higher concentrations (30 -1,000 M) of NOR3 induced the formation of trimers (118 kDa), tetramers (145 kDa), pentamers (188 kDa), and further multimers in a concentration-dependent manner (Fig. 1, C  and D). Thus, treatment with NOR3 induced the oligomerization of GAPDH. A time course study demonstrated that the oligomerization induced by NOR3 at 100 M was detectable at 5 min, and the oligomer was maintained until 60 min after treatment with NOR3 (Fig.  1E). Similar results were also obtained with the other oxidants described above (data not shown). Taken together, these results clearly indicate that oxidative stress (such as treatment with NOR3) formed distinct and stable oligomers of GAPDH through intermolecular disulfide bond formation in vitro.
Oxidative Stress-induced Oligomerization of GAPDH Leads to Its Insoluble Aggregation-Dimers and oligomers of diseaseassociated proteins such as ␣-synuclein have been reported to act as a seed of the assembly of longer polymers that form insoluble aggregates (19,21). To test the hypothesis that the oxidative stress-induced oligomerization of GAPDH leads to its insoluble aggregation, we measured the turbidity of GAPDH solutions treated with NOR3 at 37°C for 1-48 h in vitro. The NOR3 treatment at 100 M increased the turbidity of solutions in a time-dependent manner after a lag phase of 1 h (Fig. 2A). The increase in turbidity was also observed to be dependent on NOR3 concentration (Fig. 2B). After the NOR3 treatment, these solutions were centrifuged, providing the supernatants (soluble GAPDH) and the pellets (insoluble GAPDH). Upon nonreducing SDS-PAGE of the pellets, the monomer, the dimer, and the oligomer were detected (Fig. 2C, left panel). Notably, considerable amounts of the high molecular weight aggregates were found in the treatment of high concentrations of NOR3 (30 -300 M). In contrast, only the monomer and small amounts of the dimer were found in the supernatants, and the monomer amounts decreased in an NOR3-dependent man-ner (Fig. 2C, left panel). Upon reducing SDS-PAGE, both the oligomer and the aggregates in the pellets almost disappeared, but small amounts of the dimer were found in the pellets and rarely in the supernatants (Fig. 2C, right panel). Next, we investigated the effect of NOR3 treatment in the presence of DTT. DTT reduced NOR3-induced aggregation in a concentrationdependent manner (Fig. 2D). Thus, these results suggest that disulfidebonded oligomers of GAPDH act to seed aggregation in vitro.
Cys 149 , the Active Site Cysteine of GAPDH, Plays an Essential Role in the Aggregation Induced by Oxidative Stress-To identify the cysteine responsible for the oxidative stressinduced aggregation of GAPDH, we constructed cysteine-substituted mutants of GAPDH by site-directed mutagenesis. Rabbit GAPDH has four cysteines per monomer at amino acid positions 149 (the active site), 153, 244, and 281 ( Fig. 3A) (35). We substituted cysteine at these positions with serine or alanine, creating the single mutants C149S, C153S, C244A, C281S, and double mutant C149S/C281S. The purity of recombinant GAPDHs was assessed to be Ͼ98% by SDS-PAGE (Fig. 3B). We investigated NOR3-induced aggregation of each GAPDH mutant (Fig. 3C). In the wild type, NOR3 (24-h incubation) induced aggregation in a similar manner to native GAPDH as shown in Fig. 2B. In both C153S and C244A, aggregation could be detected at low concentrations of NOR3 (ϳ1 M) and was enhanced at higher concentrations. In contrast, the levels of aggregation in C281S were reduced to 45% of wild type at 100 M NOR3. Furthermore, treatment of C149S with NOR3 at 100 M led to low levels of aggregation (5% of wild type). Finally, C149S/C281S showed a complete absence of aggregation. Next, we analyzed insolubilities of these solutions by SDS-PAGE (Fig. 3D). As expected, the amount of insoluble GAPDH in each pellet was correlated with the aggregation level. These results demonstrate that both Cys 149 and Cys 281 are involved in NOR3-induced aggregation of GAPDH, and Cys 149 plays an essential role.
Effects of NOR3 on Far-UV and Near-UV CD Spectra of GAPDH-To elucidate the conformational changes of NOR3treated native GAPDH, CD spectra in both the far-UV (200 -250 nm) and near-UV (250 -300 nm) regions were measured. shown. GAPDH (0.6 mg/ml) was treated without or with 100 M NOR3 at 37°C for the indicated times (A) or treated without or with the indicated concentrations of NOR3 at 37°C for 24 h (B). C, the reaction mixtures treated without or with the indicated concentrations of NOR3 at 37°C for 24 h were centrifuged, providing the supernatants (soluble GAPDH) and the pellets (insoluble GAPDH). The pellets were resuspended in an equal volume of buffer-G2Ј and sonicated on ice. The samples were subjected to either reducing or nonreducing 5-20% SDS-PAGE by applying an equal volume of solution from either the supernatants or the pellets. D, concentration-dependent effects of DTT on NOR3-induced aggregation of GAPDH are shown. GAPDH (0.6 mg/ml) was treated with 100 M NOR3 at 37°C for 24 h in the absence or the presence of DTT at the indicated concentrations. The measurement of turbidity (left) and nonreducing 5-20% SDS-PAGE (right) were performed as described in A-C.
When GAPDH was treated with NOR3 for 1 h at 37°C, the absolute CD intensity increased between the wavelengths of 205-230 nm as compared with untreated GAPDH, suggesting augmentation of the secondary structure of GAPDH (Fig. 4A). The secondary structure content of GAPDH was calculated from the spectra by the method of Chen et al. (39). Untreated GAPDH was composed of 15% ␣-helix, 18% ␤-sheet, and 67% coil (Fig. 4A, lower panel). Upon treatment with NOR3, the ␤-sheet contents of GAPDH increased from 18 to 26%, whereas the coil content decreased from 67 to 56%, with a slight increase in the ␣-helical content from 15 to 18%. On the other hand, the near-UV CD spectrum of GAPDH exhibited cotton effects, with minima at 262, 268, 280, and 290 nm and maxima at 260, 265, 272, 288, and 296 nm, indicating an anisotropic environment of the aromatic side chains (Fig. 4B). However, these cot-ton effects were decreased by treatment with NOR3, indicating changes in tertiary structure. In the case of wild-type GAPDH, similar cotton effects were observed and also drastically decreased by treatment with NOR3 (Fig. 4C, panel a). In the case of C281S, however, the NOR3-induced decrease of cotton effects was smaller than that of wild-type GAPDH (Fig. 4C,  panel b). Moreover, both C149S and C149S/C281S displayed no NOR3-induced decrease of cotton effects (Fig. 4C, panels c  and d). Together, these results indicate that both the increase in ␤-sheet content and the conformational change of GAPDH are triggered by NOR3 treatment, and they suggest that these effects are mainly caused by a disulfide bond of Cys 149 .
NOR3-induced Aggregates of GAPDH Display Some Characteristics of Amyloid-like Fibrils-We subsequently investigated if GAPDH aggregates possessed any amyloid-like characteristics. We first measured the fluorescence intensity of thioflavin-S binding, which is one of the hallmarks of amyloid-like fibrils. In native rabbit GAPDH, NOR3 increased the fluorescence intensity of solutions in a time-dependent manner (Fig.  5A). Similar results were found in the wild-type GAPDH (Fig.  5A). NOR3 also increased the fluorescence intensity of C281S but to a lesser extent (42% of wild type at 24 h). On the other hand, NOR3 barely increased the fluorescence intensity of C149S (5% of wild type at 24 h). Finally, C149S/C281S showed no increase in the fluorescence intensity, similar to the results obtained from the turbidity study in Fig. 3. Next, we investigated the effect of another amyloid-binding dye, Congo red, on the aggregation of GAPDH. The azo-dye Congo red binds preferentially to ␤-sheet containing amyloid fibrils and also specifically inhibits the oligomerization of amyloidogenic proteins (40). Pretreatment with Congo red (200 M) significantly reduced the aggregation of native GAPDH treated with NOR3 to 20% of that of control at 24 h (Fig. 5B).
To further address whether aggregated GAPDH had formed amyloid-like fibrils, two morphological analyses were performed, Congo red birefringence and transmission electron microscopy. Birefringence upon staining with Congo red is used as a method for identification of amyloid in vitro and in vivo (10). Aggregated GAPDH exhibited Congo red binding under nonpolarized light (Fig. 5C, panel a). A portion of the sample also demonstrated typical orange-green birefringence under polarized light (Fig. 5C, panel b). The macromolecular structures formed by aggregated GAPDH were also examined using electron microscopy. In the case of NOR3-treated GAPDH, long helical fibrils were observed (Fig. 5D, panels a  and b). The fibrils were ϳ10 nm in width (Fig. 5D, panel b), and occasional broad ribbons were observed, which appeared to be composed of parallel aligned filaments (Fig. 5D, panel c). On the other hand, only amorphous structures and no distinguishable fiber-like structures were observed in the control GAPDH sample collected at 24 h after incubation without NOR3 (Fig. 5D,  panel d). Taken together, these results suggest that the insoluble aggregates of GAPDH formed by oxidants possess some properties of amyloid-like fibrils.
Effect of NOR3 on Aggregate Formation of Other Enzymes Containing Reactive Cysteines-Alcohol dehydrogenase, malate dehydrogenase, and aldolase also contain some cysteines in or near their active site, similar to GAPDH (25). Therefore, the effects of NOR3 on these enzymes were investigated to confirm whether the effects of NOR3 on GAPDH shown above are specific for GAPDH. NOR3 at a concentration of 100 M (which induces significant aggregation of GAPDH) did not increase the turbidity of alcohol dehydrogenase, malate dehydrogenase, or aldolase solutions (Fig. 6), suggesting that NOR3-induced aggregation is a specific characteristic of GAPDH.
Cys 149 Also Plays an Essential Role in the Intracellular Aggregation of GAPDH Induced by Oxidative Stress-To assess whether the aggregate-prone cysteines of GAPDH identified in vitro are also relevant under intracellular conditions, HeLa cells were transiently transfected with GAPDH cDNAs of wild type, C281S, C149S, and C149S/C281S. Fig. 7A shows the levels of GAPDH protein in the total cell lysates. N-terminal FLAGtagged GAPDHs were overexpressed at comparable amounts to endogenous GAPDH (Fig. 7A, upper panel). The efficacy of transfection of each mutant was also similar to that of the wild type (Fig. 7A, lower panel). We next determined the effective dose of NOC18, a continuous NO donor, on the cell viability. Treatment of mock-transfected HeLa cells with NOC18 for 48 h reduced the viabilities in a dose-dependent manner (Fig.  7B); the ED 50 of NOC18 was calculated as 1 mM. Next we used NOC18 at 1 mM to determine whether oxidative stress formed amyloid-like aggregates of GAPDH in the cells. To do this, immunostaining with the antibody against FLAG tag (red) and a thioflavin-S staining (green) were performed (Fig. 7C, panel  a). One day after transfection, cells were further exposed to either a control or 1 mM NOC18 for 48 h. Under the control-treated conditions, the wild type FLAG-GAPDH was diffusely distributed throughout the entire cytoplasm and thioflavin-S-negative. Treatment of the cells with NOC18 resulted in the formation of abnormal aggregates of FLAG-GAPDH in both the nucleus and the cytoplasm. Additionally, the aggregates were strongly stained with thioflavin-S, suggesting that oxidative stress induced the amyloid-like aggregation of GAPDH in cell as well as in vitro. The number of transfected cells with aggregates is summarized in Fig. 7C, panel b. Treatment of wild type-overexpressing cells with NOC18 produced a significant increase (10.5% of transfected cells) compared with that of control cells (0.9%). In C281S-overexpressing cells, NOC18 also induced the formation of the aggregates but to a lesser extent (6.5%). In contrast, in both C149S-and C149S/C281Soverexpressing cells treated with NOC18, cells with aggregates could not be detected (ϳ1%). To investigate the characteristics of the cellular aggregate, particulate fractions (including detergent-insoluble materials) were collected by centrifugation and analyzed by Western blotting using an anti-FLAG antibody (Fig. 7D,  panel a). Under nonreducing conditions (Fig. 7D, left panel), the wild type was highly aggregated by the NOC18 treatment. C281S also showed NOC18-induced aggregation, but the level was reduced compared with the wild type. There was no aggregation in both C149S and C149S/C281S. Under reducing conditions (Fig. 7D, right panel), as expected, the aggregates almost disappeared, but small amounts of the dimer in both wild type and C281S were observed. These results were largely identical to that obtained from in vitro studies. Thus, both Cys 149 and Cys 281 appear to be related to the formation of disulfidebonded amyloid-like aggregates of GAPDH induced by oxidative stress, and Cys 149 also plays an essential role in the cell. Despite NOC18 treatments, the relative amounts of Cys-substituted mutants in the particulate fractions decreased corresponding to the level of aggregation (Fig. 7D, panel b). These results also suggest that GAPDH aggregation leads to insolubility, and the protein accumulates into the particulate fraction.
Involvement of GAPDH Aggregation in Oxidative Stress-induced Cell Death-Finally, we asked whether oxidative stressinduced GAPDH aggregation is associated with cell death. Overexpression of each FLAG-GAPDH did not affect the viability of control-treated cells (Fig. 8A). When treated with NOC18, cell death in cells overexpressing wild type was significantly enhanced compared with mock-transfected cells (Fig. 8B). In C281S-expressing cells, NOC18 also induced significantly enhanced cell death, but to a lesser extent (Fig. 8B). The degree of cell death in both C149S-and C149S/ C281S-overexpressing cells was almost comparable with that in mock-transfected cells (Fig. 8B).
These results indicate that overexpression of GAPDH causes the enhancement of oxidative stress-induced cell death, the degree of which appears to be dependent on the potency of GAPDH aggregation.

DISCUSSION
In this study, we reconfirmed that oxidative stress induces the oligomerization and aggregation of GAPDH through formation of the intermolecular disulfide bonds in in vitro study (Figs. 1 and 2), as reported previously (34). Furthermore, we obtained several lines of new evidence. We first identified both Cys 149 and Cys 281 as the aggregate-prone cysteine residues of GAPDH (Fig. 3). The conformational changes in the GAPDH structure accompanied with the increase of ␤-sheet contents were found to be critical to the mechanism of oxidative stress-induced aggregation of GAPDH (Fig. 4). We also detected amyloid fibril-like features in the resultant aggregates ( Fig. 5) but not among other enzymes containing reactive cysteines (Fig. 6). Finally, we showed that exposure of GAPDH-overexpressing HeLa cells to oxidative stress induced the formation of aggregates of GAPDH in both the nucleus and the cytoplasm via disulfide bond formation, including Cys 149 and Cys 281 (Fig. 7), and that the degree of aggregation is correlated with that of the oxidative stress-induced cell death (Fig. 8). These results are compatible with previous studies in which ␤-amyloid induced the disulfide bonding and aggregation of GAPDH in either HT22 cells or rat primary cortical neurons (34) and in which overexpression of GAPDH in COS-7 cells promoted apoptosis concomitant with formation of Lewy body-like inclusions (33).
As shown in Fig. 3, the site-directed mutagenesis experiments revealed that the disulfide bonds via Cys 149 and Cys 281 played important roles in GAPDH aggregation induced by in vitro oxidative stress. Although the aggregation level of C281S was moderate compared with the wild type, both C149S and C149S/C281S displayed almost no aggregation under oxidative stress (Fig. 3). Moreover, similar results were obtained from the cell culture studies (Fig. 7). Therefore, we conclude that the  Cys 149 -Cys 149 disulfide, rather than Cys 281 -Cys 281 , plays an essential role in oxidative stress-induced GAPDH aggregation. Cys 149 is the active site cysteine and is reported to be completely conserved in all mammalian GAPDH sequences, including human (41). Indeed, we have found that oxidative stress increased the turbidity of solutions, including human GAPDH, and that the substitution of Cys 151 , the active site corresponding to Cys 149 in rabbit GAPDH, to alanine or serine (C151A and C151S) diminished the oxidative stress-induced aggregation both in vitro and in cell (see supplemental figure). Thus, oxida-tive stress-induced aggregation of GAPDH via the active site cysteine may be a general event in most mammalian cells.
It is well known that the native form of rabbit GAPDH is composed of four identical monomers (termed O, P, Q, and R) and has no disulfide bonds in the interface and the folded structure (Fig. 3A) (37). The distances (Å) between the sulfur atom of Cys 149 in each monomer were calculated using the analytical software CCP4mg to be: O-P, 29.9; O-R, 35.0; O-Q, 36.0; P-R, 36.1; P-Q, 35.0; and Q-R, 30.0. Therefore, in the native form each Cys 149 cannot form a disulfide bond because of distance constraints. In addition, Cys 149 is located at the bottom of the GAPDH active site ( Fig. 3A and ref. 37). These facts lead to a possibility that oxidative stress may cause a conformational change in GAPDH, allowing access to another Cys 149 , resulting in a paired Cys 149 disulfide bond. This concept is compatible with our near-UV CD spectra of NOR3-treated GAPDH (Fig. 4). However, the mechanism for oligomerization and aggregation of GAPDH induced by oxidative stress seems to be more complex. NOR3 treatment resulted in higher order multimeric oligomers of GAPDH (in addition to 1-4 subunit species), indicating that not only Cys 149 but also other cysteines could contribute to oligomer formation. In this regard, we propose a model of oxidative stress-induced oligomerization of GAPDH through formation of intermolecular disulfide bonds. First, oxidative stress causes a conformational change of GAPDH that allows a disulfide bond to form between Cys 149 residues of different subunits. Subsequently, the Cys 149 -Cys 149 disulfide bond induces further conformational changes, resulting in exposure of other cysteines. The newly exposed cysteines form the next disulfide bond, and the chain reaction accelerates formation of further multimeric oligomers.
It is reported that dimers and oligomers of disease-associated proteins may act as a seed of the assembly of longer polymers that form insoluble aggregates (19,21). Our results also reveal that the long term incubation of NOR3 with GAPDH converted the soluble oligomer to insoluble aggregates in vitro (Fig. 2). degree of aggregation. There are, however, some important questions regarding the aggregation and amyloid-like fibril formation of GAPDH induced by oxidative stress. First, what protein species are harmful to cells? It is well known that aggregated ␤-amyloid or ␣-synuclein and their amyloid fibrils are very toxic both extracellularly and intracellularly (8,9). On the other hand, there is increasing evidence that several kinds of partially folded intermediates in the aggregation process, namely the oligomers and so-called protofibrils, are possibly toxic, rather than amyloid-like fibrils (8,9). More recently, Nagai et al. (48) reported that the soluble ␤-sheet monomer of polyglutamine protein caused cytotoxicity, suggesting that the toxic conformational changes in the monomer are possibly responsible for neurodegeneration. In this regard, further investigations are required to determine the toxic protein species in the process of oxidative stress-induced aggregation of GAPDH.
A second question is as follows: how is the cell death mechanism of GAPDH aggregation associated with the NO-S-nitrosylation-GAPDH-Siah1 cascade, which is already proposed as a convincing mechanism of GAPDH-dependent cell death? Hara et al. (5) reported that S-nitrosylation of Cys 150 of GAPDH in HEK293 cells (corresponding to Cys 149 in rabbit) enhanced its binding with Siah1, leading to nuclear translocation and cell death. Therefore, this study could not exclude the possibility that the relief of NOC18-induced cell death in Cys 149 -substituted mutants resulted from the disappearance of GAPDH interactions with Siah1. However, we have observed a high correlation between GAPDH aggregation and cell death enhanced by oxidative stress (Fig. 8). Furthermore, we could not detect the S-nitrosylation of aggregated GAPDH treated by NOR3 in vitro (data not shown). Thus, not only the interactions of GAPDH with Siah-1 but also the GAPDH aggregations are associated with oxidative stress-induced cell death, and each cascade might operate independently.
In conclusion, our results indicate that oxidative stresses trigger a conformational change in GAPDH and induce aberrant disulfide bonds via the active site cysteine, resulting in the formation of amyloid-like fibrils and cell death. This study provides new insights into the roles of GAPDH in cell death, which has been implicated in the pathogenic mechanisms of neurodegenerative diseases.