Dityrosine cross-linking promotes formation of stable α-synuclein polymers

Intracellular proteinaceous aggregates are hallmarks of many common neurodegenerative disorders, and recent studies have shown that alpha-synuclein is a major component of several pathological intracellular inclusions, including Lewy bodies in Parkinson's disease (PD) and glial cell inclusions in multiple system atrophy. However, the molecular mechanisms underlying alpha-synuclein aggregation into filamentous inclusions remain unknown. Since oxidative and nitrative stresses are potential pathogenic mediators of PD and other neurodegenerative diseases, we asked if oxidative and/or nitrative events alter alpha-synuclein and induce it to aggregate. Here we show that exposure of human recombinant alpha-synuclein to nitrating agents (peroxynitrite/CO(2) or myeloperoxidase/H(2)O(2)/nitrite) induces formation of nitrated alpha-synuclein oligomers that are highly stabilized due to covalent cross-linking via the oxidation of tyrosine to form o,o'-dityrosine. We also demonstrate that oxidation and nitration of pre-assembled alpha-synuclein filaments stabilize these filaments to withstand denaturing conditions and enhance formation of SDS-insoluble, heat-stable high molecular mass aggregates. Thus, these data suggest that oxidative and nitrative stresses are involved in mechanisms underlying the pathogenesis of Lewy bodies and glial cell inclusions in PD and multiple system atrophy, respectively, as well as alpha-synuclein pathologies in other synucleinopathies.

Oxidative stress has been implicated in pathogenic mechanisms of PD and many other neurodegenerative diseases (25)(26)(27). Both analytical and immunological methodologies have revealed the presence of oxidized and nitrated proteins in a number of neurodegenerative disorders (27-32), supporting a role for oxidants and nitrating agents in the development of these diseases. Moreover, ␣-synuclein inclusions in brains from patients with PD (28), dementia with LBs, and multiple system atrophy are strongly labeled by antibodies specific to 3-nitrotyrosine as well as by novel antibodies that specifically detect nitrated synuclein. 2 Thus, available evidence supports the hypothesis that ␣-synuclein is a target for nitration. The generation of 3-nitrotyrosine is a post-translational modification of proteins that results from the reaction of nitrating agents with proteins. Nitrating agents are generated by the interaction of oxygen and nitric oxide-derived reactive species (26,27). As such, the detection and quantification of nitrotyrosine residues in proteins provide indirect evidence for the formation of nitrating agents and their possible involvement in pathological processes (27).
␣-Synuclein is a small protein of 140 amino acids with four tyrosine residues (33,34) that is relatively unstructured in aqueous solution (35,36), raising the possibility that the tyrosine residues may be readily exposed and accessible for modification by nitration agents. Indeed, the data reported here show that exposure of recombinant ␣-synuclein to nitrating agents leads to tyrosine nitration as well as tyrosine oxidation to form o,oЈ-dityrosine, resulting in cross-linking of ␣-synuclein to form stable high molecular mass ␣-synuclein aggregates, whereas preformed ␣-synuclein filaments are stabilized by nitrating agents in vitro.

MATERIALS AND METHODS
Expression and Purification of ␣and ␤-Synucleins-Human ␣and ␤-synuclein cDNAs subcloned into the bacterial expression vector pRK172 were expressed in Escherichia coli BL21 and purified as described previously (17). Protein concentration was determined using the bicinchoninic acid protein assay (Pierce) and bovine serum albumin as a standard. Monoclonal antibodies SYN-1 (Transduction Laboratories) and 208 (11) against ␣-synuclein and affinity-purified polyclonal antibodies against 3-nitrotyrosine (26, 27) were used in this study.
Exposure of Synucleins to Nitrating Agents-Peroxynitrite was synthesized from sodium nitrite and acidified H 2 O 2 as described previously (26). Excess H 2 O 2 was removed by reacting stock peroxynitrite with manganese dioxide. The concentration of peroxynitrite was determined spectrophotometrically at 302 nm in 1 M NaOH (⑀ 302 ϭ 1700 M Ϫ1 cm Ϫ1 ) and H 2 O 2 from the extinction coefficient The reaction of nitrating agents with ␣and ␤-synucleins as well as RNase A was carried out as described previously (37). Briefly, protein (5 mg/ml) was dissolved in nitration buffer (100 mM potassium phosphate, 25 mM sodium bicarbonate, pH 7.4, and 0.1 mM diethylenetriaminepentaacetic acid), and peroxynitrite was added in two boluses to the desired final concentration. For control experiments, peroxynitrite was decayed for 5 min in nitration buffer before being added to the protein. As an alternative mean of oxidizing ␣-synuclein, it was independently reacted with 200 M H 2 O 2 and 100 M NaNO 2 in the presence of 0.25 M human myeloperoxidase for 1 h at room temperature. To stop the reaction and to remove the excess of salts after the treatments, the proteins were eluted from a fast flow Sephadex G-25 gel filtration column (PD-10, Amersham Pharmacia Biotech). In some experiments, nitrated ␣-synuclein (1 mg/ml) was reduced with 1.5 mM sodium hydrosulfite (Na 2 S 2 O 4 ) in 25 mM Tris-HCl, pH 8.0, for 30 min at room temperature or treated with 8 M urea or 8 M guanidine HCl. After treatment, the protein was dialyzed and analyzed by standard SDS-12% polyacrylamide gel electrophoresis, followed by electrophoretic transfer to nitrocellulose membranes.
Mapping of Tyrosine Nitration in ␣-Synuclein Monomers-After nitration, ␣-synuclein monomers were exhaustively dialyzed with 100 mM ammonium bicarbonate, pH 8.0. The protein was then digested with sequencing grade-modified trypsin (Promega) at a ratio of 1:100 (w/w) at 37°C for 16 h. The peptides were eluted from a reverse-phase column, and their molecular masses were determined by mass spectrometry at the Wistar Institute Mass Spectrometry Facility, University of Pennsylvania.
Detection of o,oЈ-Dityrosine-␣-Synuclein oligomers were purified by gel filtration using a Superdex 75 HR pre-packed column (Amersham Pharmacia Biotech). The oligomers were eluted from the column with 0.15 M NaCl and 50 mM potassium phosphate, pH 7.4. The oligomerized forms were concentrated and subjected to acid hydrolysis (6 N HCl under anaerobic condition for 20 h at 110°C). The samples were then lyophilized and resuspended in 100 l of ultrapure water. The hydrolysates were separated on an octadecyl silica gel reverse-phase column (5 m, 4.6 ϫ 250 mm; Jupiter, Phenomenex Inc.) and a Hewlett-Packard HPLC system equipped with a diode array and fluorescence detector. Solvent A was 0.1% trifluoroacetic acid in ultrapure water, and solvent B was acetonitrile. Tyrosine, 3-nitrotyrosine, and o,oЈdityrosine were eluted using an increasing linear gradient of solvent B from 0 to 60% in 30 min with a flow rate of 1 ml/min. The HPLC detector was set at 210, 275, and 365 nm, and the fluorescence detection was set at 283 nm excitation and 412 nm emission. Identification and quantification of tyrosine, 3-nitrotyrosine, and o,oЈ-dityrosine were performed using external standards. o,oЈ-Dityrosine standard was synthesized by reaction of tyrosine with horseradish peroxidase and hydrogen peroxide (38).

Nitrating Agents Specifically Induce Nitration and
Oligomerization of ␣-Synuclein-Purified human recombinant ␣-synuclein was exposed to peroxynitrite in the presence of CO 2 as well as to myeloperoxidase in the presence of H 2 O 2 and nitrite (myeloperoxidase/H 2 O 2 /NO 2 Ϫ ), two putative in vivo ni-trating agents (27, 39,40). Fig. 1A shows that exposure of ␣-synuclein to peroxynitrite/CO 2 induced the formation of distinct ␣-synuclein dimers, trimers, and oligomers. Increasing the ratio of nitrating agent to protein resulted in increased nitration of ␣-synuclein dimers and higher order oligomers (Fig. 1B). Even when ␣-synuclein was reacted with substoichiometric concentrations of peroxynitrite/CO 2 , dimers of ␣-synuclein were detected (Fig. 1, A and B, compare lanes 2-4).
Nitrating conditions induce nitration and oligomerization of ␣-synuclein. Purified human recombinant ␣-synuclein protein that was either unreacted (lane 1) or reacted with molar ratios of peroxynitrite/CO 2 to protein of 0.1, 0.28, 1, 5, and 10 (lanes 2-6, respectively) was loaded in each lane of 12% polyacrylamide gels, followed by electrophoretic transfer to nitrocellulose membranes and Western blot analysis with an anti-␣-synuclein monoclonal antibody (A) or an anti-3-nitrotyrosine antibody (Anti-3-NT; B). Purified human recombinant ␣-synuclein was reacted with myeloperoxidase/H 2 O 2 (C, lanes 1 and 3) or with myeloperoxidase/H 2 O 2 plus NO 2 Ϫ (lanes 2 and 4), and the membranes were probed with an anti-3-nitrotyrosine antibody (lanes 1 and 2) or with an anti-␣-synuclein monoclonal antibody (lanes 3 and 4). An equal amount of protein was loaded in each lane.
␣-synuclein dimers to be nitrated and augmented the formation of SDS-stable oligomers, whereas myeloperoxidase/H 2 O 2 alone induced the formation of dimers without nitrating ␣-synuclein (Fig. 1C). The formation of SDS-and heat-stable oligomers appears to be a specific property of ␣-synuclein since ␤-synuclein and RNase A did not form aggregates after exposure to peroxynitrite/CO 2 under experimental conditions identical to those described above for ␣-synuclein (Fig. 2). Thus, although ␤-synuclein can be nitrated in vitro, it does not significantly oligomerize, which may indicate a difference in the structure of this protein relative to ␣-synuclein despite the high overall sequence homology and the presence of three identically placed tyrosine residues within the conserved C terminus of both of these synucleins (33,34).
To identify the site(s) of nitration in ␣-synuclein, purified human nitrated wild-type ␣-synuclein monomers were digested with trypsin, and the tryptic fragments were separated by HPLC. Mass spectrometry of these HPLC-purified tryptic fragments revealed that two peptides with retention times of 34 and 46 min showed increases in mass of 45 Da (fragment Glu 35 -Lys 43 ) and 137 Da (fragment Asn 103 -Ala 140 ), respectively. Based on the increase in mass and the amino acid sequences of the fragments, Glu 35 -Lys 43 must be nitrated at tyrosine 39 and Asn 103 -Ala 140 at tyrosines 125, 133, and 136 (Fig. 3).

␣-Synuclein
Aggregates Are Crossed-linked through the Formation of Dityrosine Bonds-To characterize the stability of the ␣-synuclein aggregates further, nitrated oligomers were reduced with sodium hydrosulfite (dithionite). Dithionite reduced 3-nitrotyrosine to 3-aminotyrosine and eliminated the immunoreactivity with the anti-nitrotyrosine antibody (Fig.  4A, lane 3), but it did not disrupt the high molecular mass oligomers (lane 6). In addition, the higher order oligomers were extremely stable since they were resistant to denaturation by SDS, 8 M urea, 6 M guanidinium chloride, and boiling (Fig. 4B). Given the stability of the interaction, it is likely that the oligomerization takes place through formation of a covalent bond, most likely o,oЈ-dityrosine. Oligomerized ␣-synuclein was isolated by gel exclusion chromatography, and the oligomers were subjected to acid hydrolysis, followed by HPLC. A distinct o,oЈ-dityrosine peak was detected both by UV and by fluorescence (Fig. 5, A and B). Quantification of o,oЈ-dityrosine in the ␣-synuclein oligomers revealed that 30% of the tyrosine residues were oxidized to form o,oЈ-dityrosine and 60% were nitrated to form 3-nitrotyrosine. Therefore, cross-linking through dityrosine appears to be responsible for the oligomerization of ␣-synuclein after exposure to the nitrating agent.
Cross-linking of Preformed ␣-Synuclein Filaments Is Augmented and Stabilized by Exposure to Peroxynitrite-To assess  1 and 3) or reacted with peroxynitrite in the presence of CO 2 (protein/nitrating agent ratio of 1:10) (lanes 2 and 4). In lanes 1 and 2, the proteins were detected by Western blot analysis using an antibody to 3-nitrotyrosine. In lanes 3 and 4, the proteins resolved by SDS-polyacrylamide gel electrophoresis were visualized by staining the gels with Coomassie Blue R-250.

FIG. 3. Mass determination of nitrated tryptic fragments from nitrated human ␣-synuclein digest.
Human recombinant ␣-synuclein was exposed to peroxynitrite in the presence of CO 2 ; the nitrated protein was digested with trypsin, and the tryptic fragments were separated by reverse-phase HPLC. The differences in masses of the two major peptides (retention times (Rt) of 34 and 46 min) from the control protein are shown in the inset. the effects of nitration on preformed filaments of human recombinant ␣-synuclein, we generated ␣-synuclein filaments in vitro as described previously (17) and then exposed these filaments to peroxynitrite in the presence of CO 2 . Similar to the unassembled protein, assembled ␣-synuclein filaments formed SDS-and heat-stable oligomers after exposure to nitrating/ oxidizing conditions. However, exposure of filamentous ␣-synuclein to these conditions resulted in a significant increase in higher molecular mass species compared with the unassembled proteins (Fig. 6A, lanes 3 and 6). Moreover, nitration/oxidation of assembled ␣-synuclein filaments stabilized these polymers since nearly 50% of the assembled protein was recovered in the 100,000 ϫ g pellet after 4 M urea denaturation (Fig. 6B). Finally, exposure of unmodified ␣-synuclein polymers to 4 M urea was sufficient to induce near complete disassembly. DISCUSSION Recent studies have shown that the majority of LBs contain 3-nitrotyrosine immunoreactivity (28) and that novel monoclonal antibodies to nitrated synuclein bind specifically to LBs and GCIs in a variety of synucleinopathies. 2 Thus, insights into the oxidation/nitration of ␣-synuclein are critical for understanding the pathogenesis of these inclusions in neurodegenerative synucleinopathies, including PD and multiple system atrophy. Significantly, our data demonstrate that ␣-synuclein is a substrate for tyrosine nitration and that it is readily crosslinked through covalent o,oЈ-dityrosine bonds. Indeed, the ability of ␣-synuclein to form oligomers in the presence of oxidizing/ nitrating agents provides a plausible explanation for the formation of SDS-resistant forms of wild-type as well as truncated and higher molecular mass species of ␣-synuclein in purified LBs (4). These and other observations support the hypothesis that nitration or other modifications associated with oxidative and nitrative stress-mediated post-translation modifications are involved in altering the biophysical properties of ␣-synuclein and promoting the aggregation of this otherwise soluble protein.
The high efficiency of ␣-synuclein nitration and oxidation is likely due to the unstructured conformation of the protein in aqueous solution since all tyrosine residues could be exposed to the solvent phase, thereby increasing the probability for reaction with nitrating agents. The efficiency of ␣-synuclein nitration is in sharp contrast with the nitration of other proteins with a similar size and number of tyrosine residues as ␣-synuclein, but with a distinctly different structural organization that appears to restrict nitration to only one or two tyrosine residues after exposure to the same nitrating agent and conditions used here. For example, under the same conditions, RNase, which contains six tyrosine residues and has a molecular mass of 14 kDa, was nitrated mostly at tyrosine 115 (37). Other critical factors that may promote nitration of ␣-synuclein include the local environment of the tyrosine residues, the absence of steric hindrance and reactive thiols, and the presence of negatively charges in the vicinity of the tyrosine residue(s) (37). Specifically, in human ␣-synuclein, there are no cysteine residues, and several glutamate residues are near tyrosines 125, 133, and 136 in the carboxyl-terminal domain of ␣-synuclein.
Exposure to peroxynitrite also led to the formation of SDSand heat-stable ␣-synuclein aggregates that resulted from the formation of o,oЈ-dityrosine cross-linking. Although nitration of tyrosine residues by peroxynitrite has been reported for several Oligomerized forms of ␣-synuclein were purified by gel filtration and hydrolyzed. The hydrolysis products were analyzed by reverse-phase HPLC using UV absorbance at 210 nm (A) and fluorescence detection of dityrosine ( ex ϭ 275 nm; em ϭ 410 nm) (B). Peak 1 corresponds to tyrosine, peak 2 to o,oЈ-dityrosine, and peak 3 to 3-nitrotyrosine. Peaks were confirmed by elution of standards and UV/fluorescence spectrum. The percentage of tyrosine, 3-nitrotyrosine, and o,oЈ-dityrosine in ␣-synuclein oligomers was determined with standards and is indicated in the inset. AFU, arbitrary fluorescence units.
proteins, formation of o,oЈ-dityrosine has been detected only in a few proteins such as fatty acid-free bovine serum albumin, human manganese-superoxide dismutase, and possibly epidermal growth factor receptor (38 -40). Exposure to myeloperoxidase/H 2 O 2 /NO 2 Ϫ , which is known to oxidize the tyrosine residues and is expected to cause the formation of o,oЈ-dityrosine (41)(42)(43), also yielded similar stable ␣-synuclein aggregates. The formation of a tyrosyl radical and its reaction with a tyrosyl radical on another ␣-synuclein molecule are the basis for o,oЈ-dityrosine covalent cross-linking (41). Similarly, the reported formation of ␣-synuclein dimers after exposure of the protein to oxidants generated by H 2 O 2 and ferric iron or H 2 O 2 plus cytochrome c also may be due to the formation of o,oЈdityrosine (44,45). Analysis of covalently oligomerized ␣-synuclein after exposure to nitrative conditions revealed that one out of four tyrosine residues was oxidized to dityrosine, whereas at least two of the three remaining residues were nitrated in the oligomerized ␣-synuclein, resulting in protein molecules that were nitrated and covalently linked via dityrosine. It is unlikely that the same tyrosine residue is nitrated and covalently linked because the addition of a nitro group in the ortho-position of tyrosine renders the aromatic ring less reactive toward further modifications. Moreover, the formation of dimerized tyrosine will impede nitration at these tyrosine residues since dimerization also deactivates the aromatic ring. Formation of o,oЈ-dityrosine most likely involves or at the very least requires tyrosine residues located in the C terminus since exposure of ␣-synuclein fragment-(2-97) (generated by digestion with endoproteinase Asp-N) to nitrating agents results in nitration of tyrosine 39, but not in the formation of aggregates (data not shown). Although ␤-synuclein contains similarly placed tyrosine residues in the C terminus and it is nitrated after exposure to the nitrative conditions used here for ␣-synuclein, it is much less prone to form stably cross-linked oligomers. These results demonstrate that the propensity of ␣-synuclein to form o,oЈ-dityrosine cross-links is a property of this molecule, which may explain, at least in part, its aggregation in pathological lesions, whereas other synuclein proteins such as ␤-synuclein are not found in LBs and GCIs (3,4,6,7,22,24).
The mechanism(s) whereby ␣-synuclein, a highly soluble protein, polymerizes in living cells to form filamentous inclusions remains incompletely understood. However, a number of recent FIG. 7. Proposed model for the role of oxidative/nitrative stress in the formation of ␣-synuclein oligomers. Under normal conditions, ␣-synuclein in the cytosol predominantly assumes a random conformation. However, ␣-synuclein can loosely associate with lipid bilayer, where it preferentially assumes ␣-helical confirmation (step A). By contrast, a small proportion of ␣-synuclein molecules in the cytosol may acquire a ␤-sheet structure (step B), but the formation of this species may normally be prevented by a chaperone or other activities. Under pathological conditions, ␣-synuclein assuming ␤-pleated sheet conformation can oligomerize (step C), but this process is probably not favorable and requires seeding. Under oxidative/nitrative conditions, short oligomers of ␣-synuclein may be covalently cross-linked by dityrosine formation (step D), which makes the polymerization process irreversible. ␣-Synuclein polymers can further elongate, and further crosslinking may occur (step E). Alternatively, short polymers can elongate into long filaments, which are subsequently stabilized by dityrosine cross-linking (step F). q ϭ ␣-helix; f ϭ random-coiled conformation; ¬ ϭ ␤-pleated sheet; ϭ o,oЈ-dityrosine-cross-linked.
FIG. 6. Stabilization of ␣-synuclein polymers. Human recombinant ␣-synuclein (5 mg/ml) was polymerized by continuous shaking in 100 mM sodium acetate, pH 6.9, at 37°C for 48 h as described previously (17). Assembled protein was recovered by sedimentation at 100,000 ϫ g for 30 min. Unassembled (lanes 1-3) or polymerized (lanes 4 -6) ␣-synuclein was untreated (lanes 1 and 4), reacted in reverse order of addition with the nitrating agent (decayed reagent) (lanes 2 and 5), or reacted with peroxynitrite/CO 2 (lanes 3 and 6). Following the reaction, SDS sample buffer was added, and samples were heated at 100°C for 10 min and resolved on an SDS-12% polyacrylamide gel. Proteins were visualized by staining with Coomassie Blue R-250 (A). Unassembled (Un) or polymerized (Poly) ␣-synuclein was treated either with decayed peroxynitrite/CO 2 (Ϫ) or with peroxynitrite/CO 2 (ϩ) at room temperature. 4 M urea was added, and following incubation at room temperature for 10 min, the samples were pelleted at 100,000 ϫ g for 30 min. SDS sample buffer was added to pellets (P) and supernatants (S), which were heated to 100°C. Samples were loaded on an SDS-12% polyacrylamide gel and visualized by Coomassie Blue R-250 (B). An equal amount of protein was loaded in each lane of the gels. The arrows indicate the formation of high molecular mass ␣-synuclein aggregates. studies and the work presented here suggest a mechanistic model based on the specific and unusual properties of ␣-synuclein (Fig. 7). Spectroscopic techniques including CD spectra revealed that wild-type ␣-synuclein contains as little as 3% ␣-helices and 23% ␤-sheets, whereas the rest of the protein assumes a random conformation (35,36). However, upon binding to small unilamellar vesicles of various compositions of acidic phospholipids, the protein folds largely in an ␣-helical conformation (63-71% of the protein) (35). Moreover, recombinant ␣-synuclein can readily assemble into filaments, and this process is accompanied by a change to an anti-parallel ␤-sheet confirmation (16,18). Thus, it is likely that ␣-synuclein exists in multiple conformations in a living cell and that the formation of ␤-sheet conformation may be restricted by a chaperone or other factors. However, under certain pathological conditions, a threshold concentration of ␣-synuclein with a ␤-sheet conformation may be attained, which enables ␣-synuclein to assemble into short filamentous polymers; and these short oligomers may "seed" the assembly of longer polymers that form ␣-synuclein filaments. The formation of these short oligomers may be the rate-limiting step in the polymerization of elongated fibrils since assembly in vitro proceeds through a "seeding" mechanism (17,46). The data presented here indicate that ␣-synuclein can undergo oxidative cross-linking via the formation of o,oЈ-dityrosine and that this modification may play a role in the fibrillogenesis by stabilizing short oligomers and/or elongated ␣-synuclein polymers (Fig. 7). Cross-linking may also contribute to the bundling of fibrils into intracellular proteinaceous aggregates and cause the covalent linking of normal, non-amyloidogenic ␣-synuclein to these aggregates. Thus, the oxidative cross-linking of ␣-synuclein may act at different steps in the formation of large intracellular lesions. This model is consistent with the elevated oxidative environment of dopaminergic neurons, which are the neurons that are primarily affected in PD (25), with the presence of oxidized metabolites in dementia with LBs (47) 2 as well as the neurons that are burdened with iron deposition in Hallervoden-Spatz disease (48). On the other hand, oxidative stress has not been demonstrated to occur in multiple system atrophy; however, the observation that GCIs contain nitrated proteins 2 suggests that oxidative stress and ␣-synuclein aggregation are common pathological features in these neurodegenerative diseases and synucleinopathies. Taken together, our data presented here and data from several previous reports provide increasing evidence that oxidative/nitrative stress plays a mechanistic role in the onset/progression of PD and related synucleinopathies.