Cellular Oligomerization of α-Synuclein Is Determined by the Interaction of Oxidized Catechols with a C-terminal Sequence*

The mechanisms that govern the formation of α-synuclein (α-syn) aggregates are not well understood but are considered a central event in the pathogenesis of Parkinson's disease (PD). A critically important modulator of α-syn aggregation in vitro is dopamine and other catechols, which can prevent the formation of α-syn aggregates in cell-free and cellular model systems. Despite the profound importance of this interaction for the pathogenesis of PD, the processes by which catechols alter α-syn aggregation are unclear. Molecular and biochemical approaches were employed to evaluate the mechanism of catechol-α-syn interactions and the effect on inclusion formation. The data show that the intracellular inhibition of α-syn aggregation requires the oxidation of catechols and the specific noncovalent interaction of the oxidized catechols with residues 125YEMPS129 in the C-terminal region of the protein. Cell-free studies using novel near infrared fluorescence methodology for the detection of covalent protein-ortho-quinone adducts showed that although covalent modification of α-syn occurs, this does not affect α-syn fibril formation. In addition, oxidized catechols are unable to prevent both thermal and acid-induced protein aggregation as well as fibrils formed from a protein that lacks a YEMPS amino acid sequence, suggesting a specific effect for α-syn. These results suggest that inappropriate C-terminal cleavage of α-syn, which is known to occur in vivo in PD brain or a decline of intracellular catechol levels might affect disease progression, resulting in accelerated α-syn inclusion formation and dopaminergic neurodegeneration.

␣-Synuclein (␣-syn) 2 (NACP, synelfin), a small, neuron-specific protein, was first linked to PD by genetic analysis of fami-lies with autosomal dominant inheritance of the disease. Genetic analysis discovered a point mutation in the ␣-syn gene (SNCA), resulting in an amino acid conversion of Ala 53 to Thr (1). Subsequently, ␣-syn protein was detected in Lewy bodies within the dopaminergic neurons of the substantia nigra pars compacta (2), the intracellular proteinaceous inclusions characteristic of PD and related disorders. Since this discovery, the process of ␣-syn aggregation has been proposed to underlie dopaminergic degeneration that occurs in PD. Therefore, delineating the mechanisms of ␣-syn aggregation and its pathophysiological role in neurodegeneration has been the focus of many investigations.
Although the in vivo factors that regulate ␣-syn aggregation are not well understood, mechanisms involving genetic and environmental factors have been proposed. Genetic analysis has uncovered two additional missense mutations in SNCA (A30P and E46K) (3,4) as well as triplication of the SNCA genomic region (5). Mutations of ␣-syn or gene triplication may increase the rate of ␣-syn aggregation (6,7), impair cellular degradation (8), or increase the amount of cytosolic ␣-syn beyond the critical concentration required to initiate polymerization (9). Most PD cases are sporadic and involve aggregation of wild type ␣-syn, indicating that other factors may influence disease progression. For example, slow, chronic administration of the selective dopaminergic toxin, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine can induce the formation of ␣-syn aggregates in mice, putatively through mitochondrial damage and subsequent oxidative stress (10). Additionally, the pesticides rotenone and paraquat have been shown to induce or alter ␣-syn pathology in rodent models through complex I inhibition and oxidative injury (11,12). Consistent with these findings, oxidative and nitrative modifications to ␣-syn have been documented in human PD brain (13) and can alter the aggregation rate of the protein through direct modifications or by impaired protein degradation (14,15). Although ␣-syn has been clearly linked to PD pathogenesis, the mechanism by which a ubiquitously expressed neuronal protein causes accelerated cell death in circumscribed brain nuclei remains unknown.
Catechols, such as dopamine, can be oxidized at physiological pH to generate reactive ortho-quinone (o-quinone) and aminochrome species. Electron-deficient o-quinones can directly modify protein structure and function through covalent attachment to cysteine or other nucleophilic residues (16,17). In relation to PD pathophysiology, cell-free in vitro analysis has shown that oxidized catechols can inhibit the formation of ␣-syn aggregates by stabilizing soluble oli-gomeric intermediates (18 -20). However, several critical questions relating to the mechanisms of dopamine-induced alterations in ␣-syn aggregation remain unanswered. Therefore, we employed a previously characterized cellular model (21) to explore the biochemical mechanisms that govern the interaction of ␣-syn with catechols. We utilized SH-SY5Y cells expressing a five-amino acid-mutated ␣-syn in the C-terminal region: Y125F,E126A,M127A,P128F,S129A, designated ␣-syn 125m. This pentapeptide motif has been suggested to be an important putative catechol interaction domain responsible for inhibiting ␣-syn aggregation as determined by cell-free in vitro studies (19). Here, we have manipulated intracellular catechol levels in ␣-syn 125m-expressing cells to determine the effect of this motif on catechol-induced inhibition of cellular ␣-syn aggregation. Further, we utilized cell-free in vitro systems to gain insight into the chemical nature of the catechol-␣-syn interaction.
Generation of Stably Transfected ␣-Syn Cell Lines-SH-SY5Y cells (American Type Culture Collection) were incubated at 37°C, 5% CO 2 and cultured in Dulbecco's modified Eagle's medium/F-12 medium (Invitrogen) containing 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin (complete medium). ␣-Syn-containing plasmids were transfected in 10-cm plates seeded at 6 ϫ 10 4 cells/cm 2 with Lipofectamine 2000 according to the manufacturer's protocol (Invitrogen) using 24 g of plasmid per 10-cm plate. Stable transfected cell lines were selected by culturing in 300 g/ml Geneticin (G418; Invitrogen) for 1 month. Monoclonal stable cell lines were selected, propagated in 300 g/ml Geneticin, and initially screened for ␣-syn protein levels by Western blot analysis (see below). Cell lines were selected for experiments based on cellular morphology, responsiveness to differentiation by retinoic acid, and ␣-syn protein levels.
Cell lines were differentiated into a neuron-like phenotype by the addition of 20 M trans-retinoic acid (Sigma) into the complete medium, which was changed every 48 h. In some experiments (Fig. 4a), n-acetylcysteine (NAC; Sigma) was dissolved in sterile water at a stock concentration of 200 mM and diluted in complete medium containing trans-retinoic acid (RA) to a final concentration of 200 M. Fresh NAC stocks were made for each addition and replenished in the medium every 24 h.
Lentiviral Infection of ␣-Syn-expressing SH-SY5Y Cell Lines-Generation of lentiviral vectors expressing TH RR-EE and preparation of virus have been described previously (21). For each experiment, cells were seeded at 6 ϫ 10 4 cells/cm 2 in culture dishes, allowed to recover for 2 days, infected with lentivirus containing empty vector or TH expression plasmids, and differentiated with 20 M RA for 8 days.
Size Exclusion Chromatography-␣-Syn was extracted as described above, and 0.5-1 mg of total Triton-soluble lysate was resolved on a Superdex 200 HR10/30 column (GE Healthcare) using 25 mM HEPES, 150 mM NaCl as the mobile phase at a flow rate of 0.3 ml/min. Fractions were collected (0.5 ml), concentrated with 5-kDa cut-off filters (Millipore), and loaded onto 12% SDS-polyacrylamide gels for Western blot analysis as described above and previously (21).
Quantification of Intracellular Catechols by High Pressure Liquid Chromatography-electrochemical detection (ECD)-After viral transduction and RA differentiation, cells (5 ϫ 10 5 cells/well) were washed in D-PBS and harvested in 100 l of 0.1 M perchloric acid containing 100 nM dihyroxybenzylamine and lysed by sonication. Lysates were centrifuged at 16,000 ϫ g for 10 min at 4°C, and 50 l of supernatant was injected on the high pressure liquid chromatograph for analysis as described previously (21).
Purification of Recombinant WT and Mutant ␣-Syn-BL21-CodonPlus (DE3)-RIL-competent cells (Stratagene) were transformed with expression plasmid prk172 WT or mutant ␣-syn and cultured to an optical density of 0.6 -0.8 at 37°C with shaking. Protein expression was induced by the addition of 1 mM isopropyl-␤-D-thiogalactopyranoside for 2 h. Bacterial cell pellets were homogenized with a pestle in high salt buffer (0.75 M NaCl, 10 mM Tris-Cl, pH 7.0, 1 mM EDTA) containing a mixture of protease inhibitors and 0.2 mM phenylmethanesulfonyl fluoride, heated to 100°C for 10 min, and centrifuged at 20,000 ϫ g for 30 min. Supernatant was dialyzed against 10 mM Tris containing 1 mM EDTA and 0.2 mM phenylmethanesulfonyl fluoride, pH 7.5, and applied to a Superdex 200 HR10/30 gel filtration column (GE Healthcare) in a mobile phase of 10 mM Tris, 150 mM NaCl, pH 7.5, at a flow rate of 0.4 ml/min. Fractions containing ␣-syn were collected, concentrated, and applied to a Resource Q column (GE Healthcare) and eluted with a 0 -0.5 M NaCl gradient. Collected fractions containing ␣-syn were purified once more by gel filtration in D-PBS. Protein concentration was determined by the bicinchoninic acid protein assay (Pierce) using bovine serum albumin as a standard.
In Vitro Analysis of ␣-Syn Aggregation-␣-Syn was incubated in conditions previously described for promoting fibril forma-tion (6) at 346 M (5 mg/ml) in D-PBS for 48 h with constant shaking at 1000 rpm, 37°C with a mineral oil overlay. For sedimentation analysis, the reaction mixture was centrifuged at 100,000 ϫ g for 30 min, 1ϫ sample buffer (20 mM Tris, 1% (v/v) glycerol, 180 mM ␤-mercaptoethanol, 0.003% (w/v) bromphenol blue, and 2% (w/v) SDS, pH 6.8) was added to both supernatant and pellets, and samples were boiled for 5 min. Samples were loaded onto 12% SDS-polyacrylamide gels, and protein was visualized by colloidal blue stain. In some experiments, NAC was dissolved in D-PBS at a stock concentration of 50 mM and then added to the reaction at a final concentration of 5 mM. For thioflavin T analysis, 5 l of sample (25 g) was removed after the incubation period and diluted into 10 M thioflavin T made in 100 mM glycine, pH 8.5, in a final volume of 100 l and incubated for 5 min at room temperature. Samples were loaded into black Maxisorp microplates (Nunc, Rochester, NY), and fluorescence at 490 nm was measured (excitation, 450 nm; bandwidth, 9 nm; emission, 490 nm; cut-off, 475 nm). For incubations of ␣-syn with catechols, L-DOPA, DA, and dihydroxyphenylacetic acid (DOPAC) (Sigma) were dissolved at a stock concentration of 10 mM in D-PBS and then immediately diluted in the reaction mixture to a final concentration of 0.5 M (1:1.5 protein/catechol ratio).
Detection of Catechol-modified Proteins Using Near Infrared Fluorescence (nIRF) and Redox-cycling Staining-This method utilizes the specific fluorescent properties of quinone-containing compounds and therefore is unable to detect unoxidized catechols or polymerization products of oxidized catechols, such as melanin. ␣-Syn was incubated with catechols in D-PBS (1:1.5 molar ratio) at 37°C for 48 h. Samples were boiled in 1ϫ sample buffer (20 mM Tris, 1% glycerol, 180 mM ␤-mercaptoethanol, 0.003% bromphenol blue, and 2% SDS, pH 6.8) and analyzed by SDS-PAGE (30 -50 g of protein/lane). Gels were immediately scanned with an Odyssey infrared imager after the run (700-nm channel, intensity ϭ 10) (Li-Cor, Lincoln, NE), and then the same gel was stained with colloidal blue to visualize total protein. Densitometric analysis by integrating protein band intensities with Odyssey infrared imaging system software, version 1.2, was utilized to quantify the amount of catechol-modified ␣-syn. Redox-cycling staining was performed as described (23). Briefly, after SDS-PAGE analysis, gels were transferred to nitrocellulose membranes and stained in a 2 M potassium glycine solution, pH 10, containing 0.24 mM nitro blue tetrazolium for 40 min and scanned with a Canon Canonscan 8400F color scanner using Adobe Photoshop software, version 6.0.
Mass Spectrometry of Catechol-modified ␣-Syn-Recombinant ␣-syn was incubated with constant shaking in D-PBS, 37°C for 48 h as described above. Following incubation, the sample was centrifuged at 100,000 ϫ g at 4°C for 30 min, and 2 g of total soluble protein was injected into the mass spectrometer. Mass spectrometry was performed on an Agilent 1100 series quadrupole mass spectrometer equipped with an electrospray ion source in a mobile phase of 0.1% (v/v) formic acid. The spectrometer was operated in positive ion mode.
In Vitro Aggregation of Fibrinogen, Alcohol Dehydrogenase, and ␥-Crystallin-Aggregation for all three proteins was determined by light scattering at 360 nm and followed as a function of time in seconds, using a SpectraMax 250 microplate spectrophotometer system (Molecular Devices, Sunnyvale CA). Aggregation was assessed in the presence of a 1:2 ratio (protein/catechol) of freshly made 10 mM dopamine or DOPAC (Sigma) in D-PBS or autoxidized dopamine or DOPAC (incubated at 25°C for 1 month in D-PBS). Human fibrinogen (1 mg/ml; American Diagnostica, Inc., Greenwich, CT) aggregation was initiated by the addition of 0.1 units/ml human thrombin (American Diagnostica) in Tris-buffered saline (50 mM Tris base, 150 mM sodium chloride, pH 7.4, at 25°C as previously described (24). Horse liver alcohol dehydrogenase (1 mg/ml; Sigma) thermalinduced aggregation was performed as described previously (25) in 100 mM phosphate buffer, pH 7.0, 45°C. Bovine ␥-crystallin (0.1 mg/ml; Sigma) was subjected to acid denaturation in 100 mM sodium citrate, pH 3.0, at 20°C.
Statistical Analysis-Data were analyzed using Sigmastat software version 2.03 (SPSS Inc., Chicago, IL) and are expressed as the mean Ϯ S.E. One-way analysis of variance followed by Tukey's post hoc test was utilized to determine if groups were statistically different, with p values of Ͻ0.05 considered significant.

RESULTS
To determine the cellular mechanism of catechol-␣-syn interactions and the effect on inclusion formation, stably trans-fected ␣-syn 125m SH-SY5Y cell lines were generated. Because expression of ␣-syn 125m did not result in a substantial amount of cells containing inclusions (Fig. S1,  a and b), the A53T mutation was introduced into the 125m construct to generate ␣-syn A53T-125m. Several cell lines were generated and analyzed for A53T-125m expression levels by Western blot analysis (Fig. 1, a and b). Based on this analysis, line 34 was selected for subsequent experiments. Immunofluorescence analysis of RA-differentiated A53T-125m cells revealed the presence of cytoplasmic, punctated aggregates as well as larger, juxtanuclear inclusions that co-localized with thioflavin S, indicating the presence of amyloid inclusions (Fig. 1, c and  d). The number of A53T-125m cells containing punctated and juxtanuclear inclusions that co-localized with thioflavin S was comparable with cells expressing A53T ␣-syn. As a control, cells expressing ⌬71-82 ␣-syn, a mutant ␣-syn protein that is incapable of polymerizing in vitro (22) and in vivo (26), were used. Immunofluorescence analysis showed a diffuse, cytoplasmic staining pattern with only a small percentage of cells (Ͻ1%) that contained small punctated ␣-syn aggregates (Fig. 1, c and d).
The transfected SH-SY5Y cell lines used for these experiments do not contain detectable amounts of catechols or the rate-limiting synthesizing enzyme, tyrosine hydroxylase (TH) (vector condition; Fig. 2, a and b). Therefore, to determine the effect of increasing catechol levels on ␣-syn aggregation in A53T-125m cells, TH was expressed by transduction with a lentiviral vector encoding for human TH-1. To achieve elevated levels of cytosolic catechols, a mutant form of TH was utilized, TH RR-EE, which contains an altered catecholamine feedback inhibition site (residues 37 and 38 converted from Arg to Glu), allowing the enzyme to produce higher catechol levels compared with wild type TH (21,27). After RA differentiation, the TH RR-EE-transduced A53T-125m cells contained higher levels of catechols compared with empty vector-transduced cells (Fig. 2b). As shown previously, commensurate to increasing catechol levels, the number of cells containing ␣-syn aggregates in A53T expressing cells declined (Fig. 3, a and b). However, the same treatment in A53T-125m cells had no effect on ␣-syn cytoplasmic distribution (Fig. 3, a and b). Co-staining for ␣-syn and TH showed that a significantly higher amount of TH-positive A53T-125m cells contained ␣-syn aggregates compared with A53T cells (Fig. 3, a and b). These findings were further verified by immunostaining A53T-125m cells with polyclonal Ab SNL-4, which recognizes an N-terminal region of ␣-syn, and monoclonal Ab Syn 514, which preferentially detects mature, juxtanuclear ␣-syn inclusions. Both SNL-4 and Syn 514 revealed a higher proportion of TH-positive A53T-125m cells containing ␣-syn inclusions compared with TH-positive A53T cells (Figs. 3, a and b, and S2, a and b). These results indicated that the 125 YEMPS 129 region of ␣-syn is required for catecholmediated inhibition of ␣-syn aggregation in SH-SY5Y cells.
Previous analysis revealed that cytosolic catechols inhibit the formation of Triton-insoluble ␣-syn with a concomitant increase of soluble oligomers in A53T ␣-syn-expressing cell lines (21). To determine if this process requires catechol oxidation, cells were transduced to express TH RR-EE and then differentiated with RA in the presence of NAC. Formation of Triton-soluble ␣-syn oligomers was assessed by size exclusion chromatography, followed by ␣-syn Western blot analysis of collected fractions. Consistent with previous observations (21), monomeric ␣-syn eluted off at a fraction corresponding to a 32-Å protein in empty vector-transduced A53T cells (Fig. 4a). The proportion of 100 -63-Å-sized ␣-syn oligomers was increased in cells with elevated cathechol levels (Fig. 4a). The elution profile of TH RR-EE-expressing A53T cells was reversed by treatment with NAC, showing that formation of soluble oligomers by catechols is an oxidation-dependent process in SH-SY5Y cells (Fig. 4a). Moreover, consistent with the data in Fig. 3, no change in the elution profile of A53T-125m ␣-syn was observed upon transduction with lentivirus contain-ing TH RR-EE (Fig. 4b), validating that the 125 YEMPS 129 C-terminal region is required for catechols to arrest ␣-syn oligomers from being incorporated into mature inclusions. Additionally, ⌬71-82 ␣-syn-expressing cells were analyzed to determine if oxidized catechols have the ability to interact with monomeric ␣-syn to induce the formation of soluble ␣-syn oligomers. Increasing catechols did not the change the elution profile of ⌬71-82 ␣-syn (Fig. 4c). Taken together, these results imply that oxidized catechols interact with the 125 YEMPS 129 motif of ␣-syn oligomers that are in the process of forming polymers.
To gain further insight into the mechanism of catechol-induced inhibition of ␣-syn aggregation, cell-free in vitro systems were utilized. Purified recombinant ␣-syn was incubated in the presence of the deaminated metabolite of DA, DOPAC, for 48 h (1:1.5 molar ratio of protein/catechol), and aggregation was assessed by centrifugal sedimentation analysis followed by SDS-PAGE, which detects aggregated ␣-syn in the pellet fraction of the reaction mixture. Similar to the previously described effects of L-DOPA and DA, DOPAC also inhibited the formation of pelletable ␣-syn (Fig. 5a). The soluble protein was composed of the monomer (18 kDa) and SDS/heat-stable dimers (36 kDa) and trimers (54 kDa). As anticipated, this effect was reversed by the addition of the antioxidant NAC, indicating a requirement for catechol oxidation to form o-quinone (Fig. 5a). These results were further corroborated by quantification of thioflavin T binding (Fig. 5b). L-DOPA, DA, and DOPAC effectively inhibited ␣-syn aggregation (percentage inhibition: 91 Ϯ 1, 90 Ϯ 1, and 85 Ϯ 1, respectively). Because DOPAC lacks a reactive amine required to form the cyclic aminochrome structure, these data suggest that inhibition of ␣-syn aggregation by catechols occurs by a general mechanism, through interactions with o-quinone-containing compounds.
Although electron-deficient o-quinone-containing compounds have been shown to covalently attach to proteins by reacting with nucleophilic amino acid residues, previous analysis determined that incubation of ␣-syn with DA does not form significant amounts of covalent protein adducts (19,20). This effect was presumed to be due to both the absence of cysteine residues in ␣-syn and the relative rapid in vitro polymerization rate of oxidized DA, which consumes the reactive o-quinone into cyclic aminochrome structures. However, we reasoned that incubation with DOPAC would probably result in relatively higher amounts of catechol-protein adducts, due to the extended half-life of the reactive DOPAC-o-quinone, which is not consumed into polymers as rapidly as DA under physiological conditions. Moreover, previous studies in rodent models have shown that intrastriatal DA injections result in nearly 6-fold higher protein-bound DOPAC compared with DA adducts (28), indicating that intracellular DOPAC-o-quinone reacts more readily with proteins compared with DA. Therefore, we developed a new method for in-gel detection of quinone-modified proteins based on nIRF, which takes advantage of the specific excitation-emission spectra of oxidized catechols. nIRF analysis revealed the presence of o-quinone-containing ␣-syn monomers (18 kDa) and dimers (36 kDa) in DOPAC-incubated conditions (Fig. 5c) but not with L-DOPA or DA (Fig. S3a). Detection of a DOPAC-␣-syn adduct was corroborated with a well established method for detecting qui-  OCTOBER 26, 2007 • VOLUME 282 • NUMBER 43 none-modified proteins using redox-cycling staining to reduce tetrazolium to formazan (23) (Fig. 5c). Finally, to verify the presence of the ␣-syn-DOPAC adduct, electrospray ionization-mass spectrometry was utilized. After a 48-h incubation, mass shifts (multiples of 16 atomic mass units) were observed in both control and DA-treated samples, probably indicating the pres- . Increasing intracellular catechols decreases the number of cells containing A53T ␣-syn aggregates but does not affect the aggregation of A53T-125m ␣-syn. a, ␣-Syn-expressing cell lines were infected with either empty vector or TH RR-EE-containing lentiviral vectors to increase intracellular catechol levels, differentiated for 8 days with RA, and then analyzed by immunofluorescence using polyclonal Ab SNL-1 (left, red, ϫ30 magnification). Monoclonal Ab Syn 514 (right, red) was also used, since it preferentially detects juxtanuclear (JN) ␣-syn inclusions. Inclusion bodies are marked with white arrowheads. Due to the low endogenous levels of TH in these cells, lentivirus-transduced cells were easily identified by TH immunofluorescence (green). b, quantification of cells containing total ␣-syn aggregates with SNL-1 (small and large juxtanuclear). Juxtanuclear aggregates alone are also quantified with Syn 514 and SNL4 (see Fig. S2) (n ϭ 3 for each antibody; values are the mean Ϯ S.E.; *, p Ͻ 0.05). ence of oxidized methionine residues; however, no evidence of DA-modified protein was observed (Fig. S3b). DOPAC-treated samples also showed mass shifts corresponding to oxidized methionine but showed an additional mass shift of 170 atomic mass units, indicating the presence of a covalent ␣-syn-DOPAC adduct (Fig. S3b). Taken together, these results suggest that DOPAC reacts more readily with ␣-syn to form a covalent adduct compared with L-DOPA or DA, presumably due to the relatively slow polymerization of DOPAC-o-quinone.

Role of C Terminus in ␣-Synuclein-Catechol Interaction
To further explore the effect of covalent modification by catechols on ␣-syn fibril formation, recombinant purified ␣-syn 125m protein was incubated with DOPAC under the same conditions as WT ␣-syn, and amyloid formation was assessed by thioflavin T fluorescence. Although DOPAC inhibited amyloid formation of WT ␣-syn by 80%, it inhibited ␣-syn 125m amyloid formation by only 20% (Fig. 6a). Additionally, nIRF analysis revealed that DOPAC formed a covalent adduct with ␣-syn 125m to the same extent as WT ␣-syn (Fig. 6, b and c). Covalent modification of ␣-syn 125m by DOPAC was verified by mass detection (not shown). Therefore, DOPAC forms a covalent adduct to amino acids other than 125 YEMPS 129 , and this modification does not alter the rate of fibril formation.
To further explore if oxidized catechols specifically interact with a YEMPS motif during the process of fibril formation, the effect of catechols on the aggregation of other proteins was assessed in cellfree systems. Similar to ␣-syn, the polymerization of human fibrinogen involves a transition through protofibrillar intermediates to form fibrils (29). The reaction is initiated by incubation with thrombin, which sequentially cleaves two fibrinopeptides from the A and B chains. Incubation of fibrinogen with a 1:2 molar excess of fresh DA, autoxidized DA, DOPAC, or autoxidized DOPAC did not prevent fibril formation (Fig.  S4c). The lack of inhibition may relate to the relatively large size of fibrinogen molecules as compared with ␣-syn or the lack of the interactive motif YEMPS in fibrinogen. Similarly, a 1:2 molar excess of fresh DA, oxidized DA, DOPAC, or oxidized DOPAC did not prevent the thermal or acid-induced aggregation of alcohol dehydrogenase or bovine lens ␥-crystallin (Fig. S4, a  and b). Bovine lens ␥-crystallin is a 175-amino acid protein with an apparent molecular mass of 20,000 Da and contains a YEMPS motif (residues 135-139) in the C terminus. Unlike ␣-syn polymerization, thermal or acid protein aggregation of these proteins does not involve the formation of fibrils (30). Therefore, these results suggest that catechols and oxidized derivatives are not capable of preventing protein aggregation that does not undergo organized conformational changes to protofibrils and fibrils, even when a YEMPS sequence is present.

DISCUSSION
The discovery of insoluble, ␣-syn fibrils in Lewy body inclusions of PD brain has suggested that aberrant conformational changes that induce ␣-syn aggregation are an important pathological event (2). Although the in vivo factors governing the formation of ␣-syn aggregates are not entirely understood, recent studies have suggested that oxidative modifications by dopamine may be critically involved. Cell-free in vitro studies revealed that dopamine and other catecholamines have the ability to inhibit the formation of ␣-syn fibrils through stabilization of soluble ␣-syn oligomers (18), providing a possible explanation into how accelerated dopaminergic cell death might occur in PD. The cell-free data have been further vali-dated in human neuroblastoma cell lines, which employed molecular and biochemical approaches to show that cytosolic catechols influence ␣-syn aggregation (21). Despite these advances, the biochemical mechanisms that mediate the intracellular effects of catechols on ␣-syn inclusion formation remained unclear. The current work revealed the importance of the 125 YEMPS 129 C-terminal region of ␣-syn in the intracellular interaction with catechols by showing that inclusion formation in A53T-125m cells is not affected by the increase in intracellular catechol levels as compared with A53T ␣-syn. The data also showed that intracellular oxidation of dopamine and the noncovalent interaction of oxidized catechols with the 125 YEMPS 129 prevented the formation of ␣-syn inclusions. Collectively, these data suggest that dopamine and other catechols have the ability to modulate the progression of PD through C-terminal interactions with ␣-syn. Aberrant ␣-syn C-terminal processing may lead to selective vulnerability of dopaminergic neurons by accelerating the formation of insoluble ␣-syn inclusions.
Previous studies have shown that the negatively charged C terminus of ␣-syn is an important regulator of fibril formation in vitro, since removal of this region accelerates the rate of ␣-syn polymerization (31)(32)(33). Enrichment of C-terminal truncated ␣-syn has been documented in regions of PD brain, including the substantia nigra, implicating the forma-FIGURE 5. Inhibition of ␣-syn aggregation arises from the interaction of protein with o-quinones. a, recombinant purified ␣-syn (5 mg/ml) was incubated for 48 h in the presence or absence of catechols (0.5 mM) under conditions that promote aggregation (1:1.5 ratio of protein/catechol). Assembly was monitored by centrifugal sedimentation. Supernatant (S) and pellet (P) fractions were analyzed by SDS-PAGE, and proteins were visualized by colloidal blue. b, ␣-syn was incubated as described in a, and fibril formation was assessed by thioflavin T (ThT) binding (n ϭ 3; values are the mean Ϯ S.D.; *, p Ͻ 0.05 compared with WT ␣-syn alone; ‡, p Ͻ 0.05 compared with DA and L-DOPA). c, detection of the ␣-syn DOPAC covalent adduct. Proteins were incubated in the presence of DOPAC or DOPAC plus NAC, boiled in 2% SDS, and analyzed by SDS-PAGE. Total protein (30 g/lane) was visualized by colloidal blue, whereas quinone-modified protein was detected with nIRF, as described under "Experimental Procedures," and verified by redox cycling staining. NBT, nitro blue tetrazolium. The assay was repeated three times with the same results. tion of this species in the pathological mechanisms of neurodegeneration (31,34). The data presented here underscore the importance of the C-terminal region of ␣-syn and how modifications to this region might relate to accelerated dopaminergic degeneration in PD brain. For example, truncation to remove the 125 YEMPS 129 region renders ␣-syn incapable of interacting with catechols, potentially causing a selective aggregation in dopaminergic neurons. This suggestion is supported by recent studies in transgenic mice. Although mice expressing full-length human WT or A53T ␣-syn are devoid of inclusions in dopaminergic brain regions (35,36), expression of truncated ␣-syn-(1-120) induces the formation of inclusions in these regions (37). Therefore, in human PD, C-terminal cleavage may act synergistically with a decline in intracellular catechol levels, eventually leading to the formation of insoluble ␣-syn inclusions.
Data from cell-free in vitro studies provided further insights into the nature of the catechol-␣-syn interaction. We show that oxidation of DOPAC, the deaminated metabolite of DA, also has the ability to inhibit ␣-syn aggregation. This suggests that o-quinone structures, in addition to cyclized aminochromes, interact with ␣-syn. Other analysis revealed that covalent modification of ␣-syn by o-quinones does not mediate ␣-syn aggregation. This conclusion was based on the following observations: 1) although L-DOPA, DA, and DOPAC are almost equally effective at inhibiting fibril formation, only DOPAC has the ability to covalently modify ␣-syn, as shown by nIRF analysis and electrospray ionization-mass spectrometry; 2) DOPAC o-quinone was found to covalently modify 125m ␣-syn to the same extent as WT ␣-syn, irrespective of the effect on fibril formation. Collectively, these data indicate that noncovalent interaction of oxidized catechols with the C-terminal motif of ␣-syn oligomers governs the ability to prevent the incorporation into fibrils.
Although implications for the involvement of ␣-syn in neurodegeneration are clear, the specific form of ␣-syn (i.e. monomer, oligomer, or fibril) responsible for toxicity has not been determined. Studies in animal models have shown that elevated soluble forms of ␣-syn may cause toxicity without the formation of insoluble aggregates (38,39), whereas others have shown a strong correlation between inclusion formation, disease onset, and neurodegeneration (35)(36)(37). In SH-SY5Y cells, although elevated levels of A53T-␣-syn induced toxicity, variation of intracellular catechol levels did not affect cellular degeneration despite dramatic effects on altering ␣-syn aggregation (21). Similarly, increasing catechol levels in A53T-125mexpressing cells had no effect on cell viability (data not shown). Whereas oxidation of cytosolic catechols is relatively well tolerated, as evident by the presence of neuromelanin in apparently healthy nigral neurons, excessive cytosolic oxidation of catechols can be neurotoxic, as documented in cell culture and rodent models (28,40,41). Overall, the data indicate that oxidized catechols influence ␣-syn polymerization, which in turn may control the toxicity of ␣-synuclein.