Metal-Catalyzed Oxidation of Alpha Synuclein : Helping to Define the Relationship between Oligomers , Protofibrils and Filaments

Genetic Diseases Research Branch, National Human Genome Research Institute, National Institutes of Health, Bethesda, MD 20892 Division of Extramural Activities, National Institute of Neurological Disorders and Stroke, NIH Division of Bioengineering and Physical Science, Office of Research Services, NIH Laboratory of Biochemistry, National Heart, Lung, and Blood Institute, NIH Running title: Synuclein oxidation inhibits filament formation ‡ To whom correspondence should be addressed: 49 Convent Drive/Room 4B67; Genetic Disease Research Branch, National Human Genome Research Institute, NIH, Bethesda, Maryland 20892, USA; Ph: (301) 402-2038; FAX: (301) 402-2170; email: ncole@mail.nih.gov

Oxidative stress is implicated in a number of neurodegenerative diseases, and is associated with the selective loss of dopaminergic neurons of the substantia nigra in Parkinson's disease.The role of alpha synuclein as a potential target of intracellular oxidants has been demonstrated by identification of posttranslational modifications of synuclein within intracellular aggregates that accumulate in PD brains, as well as the ability of a number of oxidative insults to induce synuclein oligomerization.The relationship between these relatively small soluble oligomers, potentially neurotoxic synuclein protofibrils, and synuclein filaments remains unclear.We have found that metal-catalyzed oxidation of alpha synuclein inhibited formation of synuclein filaments, with a concomitant accumulation of beta-sheet rich oligomers that may represent synuclein protofibrils.Similar results with a number of oxidative and enzymatic treatments suggest that covalent association of synuclein into higher molecular mass oligomers/protofibrils represents an alternate pathway from filament formation, and renders synuclein less prone to proteasomal degradation.
The role of oxidative stress in the etiology of Parkinson's disease (PD) 1 has received extensive attention, yet little is understood regarding the primacy of damaging events that ultimately lead to cell loss and brain dysfunction.The detection of biochemical landmarks of oxidative stress in postmortem tissues yields clues as to the initiating event(s) (1,2), although it is difficult to distinguish a causative role for oxidative stress in disease initiation and progression versus damage as a product of general metabolic failure (3).The major clinical symptoms in PD, which involve the selective demise of neurons of the substantia nigra pars compacta, have been attributed largely to the unique sensitivity of these dopamine-containing neurons to impairments in mitochondrial complex I activity (4).Pesticides and other environmental toxins that affect complex I, such as 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (5,6), paraquat (7,8), and rotenone (9,10) replicate many of the features of sporadic PD, including the selective degeneration of dopamine neurons and formation of intracellular aggregates containing alphasynuclein.One common feature of these toxins is their ability to induce oxidative damage via the generation of reactive oxygen species (ROS) produced by mitochondrial respiratory chain dysfunction (4).
Inherent in the toxicity due to generation of ROS is the fundamental role of transition metals, principally iron and copper, in catalyzing ROS formation via their ability to undergo interconversion between oxidized and reduced states (11)(12)(13)(14)(15)(16).In fact, considerable evidence suggests that much of the toxicity of enzymatically-generated superoxide radical anion and hydrogen peroxide is due to their reaction with reduced iron or copper ions to generate the highly reactive hydroxyl radical via iron-catalyzed Haber-Weiss and Fenton reactions (15,16).Similarly, transition metals stimulate the formation of alkoxyl and peroxyl radicals from the decomposition of polyunsaturated fatty acids (PUFAs) (17,18) and semiquinones derived from oxidized catecholamines (19).
Alpha synuclein is a 140 amino acid, principally unstructured cytosolic protein, an unknown fraction of which reversibly associates with membranes via a series of amphipathic helices, located in its N-terminal region (20).A central, relatively hydrophobic domain (NAC) is responsible for alpha synuclein's capacity to form beta sheet rich amyloid filaments (21), which is followed by an acidic tail region which blocks rapid synuclein filament assembly (22) and allows synuclein to function as a molecular chaperone (23,24).Both wild type and mutant forms of alpha synuclein have been linked genetically to PD (25)(26)(27)(28).In addition, biochemical and histological analyses have identified alpha synuclein as a major component of the intraneuronal inclusions known as Lewy bodies and Lewy neurites associated with a number of diseases, including PD, collectively termed synucleinopathies (29)(30)(31).Although the filamentous nature of these inclusions can be mimicked structurally by the in vitro polymerization of recombinant alpha synuclein (32)(33)(34), the identification of both SDSsoluble and insoluble oligomers from synucleinopathy brains (31,(35)(36)(37)(38) indicates that changes in synuclein's solubility and/or the formation of covalently-associated synuclein oligomers, rather than the exclusive non-covalent association of synuclein into filaments, may underlie its role in pathogenesis.Indeed, a number of treatments induce the formation of SDSresistant synuclein oligomers in vitro, many of which require oxidizing conditions.These include cross-linking in the presence of transition metals (39)(40)(41)(42), peroxynitrite (43,44), polyunsaturated fatty acids (PUFAs) (45,46), and tissue transglutaminase (tTGase) (47,48).Larger oligomerized forms of synuclein termed "protofibrils" 2 , operationally defined from gel filtration experiments, have also been described that may function as intermediates in the pathway leading to filament formation, and have been proposed to play a role as toxic intermediates in the etiology of PD (49)(50)(51).Although the relationship between the formation of smaller soluble SDS-resistant oligomers, larger soluble synuclein protofibrils, and insoluble synuclein filaments in pathogenesis remains unclear (51), abnormally accumulated or modified synuclein may compromise the ability of the proteasome to efficiently degrade cellular proteins (52)(53)(54)(55), resulting in proteolytic stress that leads to neuronal dysfunction and degeneration (56).Whether metal catalyzed oxidative processes are causative in Lewy body formation or accumulate subsequent to defects in protein and lipid turnover remains unknown.

Metal catalyzed oxidation (MCO) of alpha
synuclein-Recombinant alpha synuclein was incubated at the indicated concentrations in 20 mM sodium phosphate buffer (pH 7.4) for 16-18 h at 37 o C, unless otherwise indicated.Where indicated, incubations included iron (III) chloride or copper (II) sulfate, hydrogen peroxide, DTT, and metal chelators at the indicated concentrations.Variability was observed in the effectiveness of different lots of DTT to result in synuclein oxidation/oligomerization, which was likely due to the presence of contaminating metals present in the various solutions.Addition of neutralized solutions derived from low pH elution of Chelex 100-treated DTT stock solutions demonstrated that the likely primary source of contaminating trace metal was the DTT itself.Anaerobic conditions were created by exposing all solutions to an anoxic atmosphere containing 95% N 2 /5% H 2 for 3 hr at room temperature prior to incubation.Dr. Tracey Rouault (Cell Biology and Metabolism Branch, NICHD, NIH) generously provided the anoxic chamber.After the reactions, SDS sample buffer was added to the samples, which were then boiled for 5 min then loaded onto precast 10-20% Tris-tricine gels for SDS-PAGE.No differences were detected whether the synuclein samples were boiled or not, or whether the SDS sample buffer contained 100 mM DTT or not (except when using the thiol-reducible crosslinker DSP, where DTT was omitted).Synuclein was detected by immunoblotting after transfer to nitrocellulose membranes (Protran, 0.2 um, Schleicher & Schuell BioScience, Keene, NH), or by silver staining, according to the manufacturer's instruction (Silver Stain Plus, Bio-Rad).
In vitro assembly of synuclein filamentsultrastructural and biochemical analysis-Purified recombinant synuclein was concentrated to ~ 10 mg/ml (Centriprep 10,000 MWCO) and dialyzed overnight against phosphate-buffered saline [PBS (Invitrogen); 1 mM K 2 HPO 4 , 155 mM NaCl, 3 mM Na 2 HPO 4 , pH 7.4], and incubated at a concentration of 3 mg/mL (200 µM) for up to 72 hr at 37 o C with shaking (1400 rpm; Eppendorf Thermomixer R). 100 µM Fe(III)/10 mM DTT or 10 mM deferoxamine were included, where indicated.For incubations using preformed synuclein "seeds", synuclein filaments generated from 72 h incubations were briefly sonicated (5 sec) and added to incubations at a 1:100 dilution (based on total synuclein concentration).We found that sonication dramatically reduced the lag time required to generate new synuclein filaments over unsonicated preformed filament seeds (62), and under these conditions incubations were routinely performed for 16-20 h.For electron microscopy, aliquots (1 µL) were taken, diluted two-fold in water, placed on 1% pioloform and copper-coated grids and negatively stained with 1% uranyl acetate.Filaments and other nonfilamentous structures were observed on a JEOL-100 transmission electron microscope at an accelerating voltage of 80 keV.For Thioflavin T fluorescence measurements, 5 µL aliquots from the synuclein incubations were added to 1 mL of a solution containing 10 uM Thioflavin T (prepared from a 100 µM filtered stock solution in water) in 90 mM glycine-NaOH, pH 8.5 (The resulting synuclein concentration was 1 µM, and under oxidizing conditions, the final Fe(III) concentration was 0.5 µ M).Fluorescence measurements were performed in a 1-cm cuvette (Uvette, 220-1600 nm, Eppendorf) with excitation at 440 nm (slit width 4 nm) and emission scan 450-600 nm (slit width 4 nm), and analyzed with Felix fluorescence analysis software (Photon Technology International), courtesy of Paul Smith and Alec Eidsath, Instrumentation Research and Development Resource (OD), NIH.The graphs are representative of at least three independent experiments from at least two separate synuclein preparations.
For analysis of synuclein protofibrils, insoluble fibrils were removed by centrifugation (16,000 x g for 10 min), and the supernatants (200 µ L volume) subjected to size exclusion chromatography using a Superdex HR200 column (Amersham Pharmacia Biotech AB) with PBS containing 0.02% sodium azide as the mobile phase, and a flow rate of 0.4 mL/min.The column eluate was monitored at 280 nm (Waters 2487).An on-line multi-angle light scattering detection system (Wyatt Technology Dawn EOS) was used to determine the masses of the eluting species.Wyatt Technology Astra software was used for the data analysis.
Purification of synuclein oligomers-Synuclein oligomers generated by treating 3 mg/ml synuclein at 37 o C overnight with 100 µM Fe(III)/10 mM DTT were loaded (0.2 ml) onto a Superose 6 HR 10/30 gel filtration column (Amersham Pharmacia Biotech, Piscataway, NJ) using 50 mM potassium phosphate, pH 7.0, 100 mM KCl as the mobile phase and a flow rate of 0.5 mL/min.We found this column gave a slightly better separation of oligomeric from monomeric synuclein compared with Superdex 200.Fractions containing synuclein oligomers were concentrated by Centricon (Amicon) centrifugal filter devices or by reverse osmosis with Slide-A-Lyzer dialysis cassettes (10K MWCO) and concentrating solution (Pierce), which resulted in less loss of material, presumably due to binding of oligomers to the filters.
Circular dichroism (CD) spectroscopy-Far-UV CD spectra of purified monomeric (70 µM) and oligomeric synuclein (20 µM) were collected at 20 o C on a Jasco J-720 spectropolarimeter equipped with a temperature controller using a 0.02 cm path length cell.Data were collected at 1 nm intervals with a response time of 1 s per measurement at a scan speed of 10-20 nm/min.The final spectrum was obtained by subtracting the background from the mean of four individual scans.
Analysis of synuclein fibrillization by centrifugal sedimentation-Untreated synuclein or seeded assembly reactions were centrifuged at 16,000 x g for 10 min.Supernatants and pellets were separated, pellets were washed with 0.5 ml PBS, repelleted and diluted to the same volume as the supernatants.SDS sample buffer was added to equal sample volumes, and the samples were heated to 100°C for 5 min.Synuclein proteins were resolved by SDS-PAGE and immunoblotted with anti-synuclein antibodies.For seeded reactions, up to 10 µg total protein was loaded per lane in order to visualize the "seed", which was added at a 1/100 dilution (100 ng).
20S Proteasome degradation assay-20S catalytic core was purified from rat liver, as described (63).To assay for synuclein degradation, 1 µ M synuclein, either control or oxidized by DTT/iron, was incubated at 37 o C for 1 h with 50 nM purified 20S proteasomes in 50 mM HEPES buffer, pH 7.8, in the absence or presence of 50 µM MG132.The reaction was stopped after 1 h by addition of 0.05% trifluoroacetic acid (TFA).SDS sample buffer was added followed by SDS-PAGE and immunoblotting with anti-synuclein antibodies.
Chymotrypsin-like, trypsin-like, and peptidylglutamyl-peptide hydrolase (glutamyl-like) activities of the 20S proteasome were assayed with fluorogenic peptides AAF-MCA, LSTR-MCA, and LLE-MCA, respectively, according to (64).A typical assay was carried out with 50 nM (final concentration) 20S in a total volume of 0.2 ml in 50 mM Hepes, pH 7.8 supplemented with the appropriate peptide substrate (100 µM, final concentration) in a 96-well plate.Following incubation at 37°C for 60 min the reaction was quenched by addition of acid.The samples in each well of the plate were read using either a SpectraMax Gemini EM (Molecular Devices) or a CytoFluor series 4000 (PerSeptive Biosystems) multi-well plate reader.The excitation and emission wavelength were 360 nm and 460 nm, respectively.A standard curve of the fluorescence for the pure product was used to calculate the concentration of liberated aminomethylcoumarin product in the assays.Statistical analyses used the non-parametric Wilcoxon rank sum test for significance that makes no assumption about whether the data are normally distributed.
Synuclein cross-linking in the presence of tissue transglutaminase (tTGase), lipid hydroperoxides (LHP), DSP and dopamine-tTGase catalyzes the formation of an ε-(γ-glutamyl)lysine isopeptide bond between glutamine and lysine residues and requires calcium for maximal activity (65,66).Synuclein incubations in the presence of tTGase (50 nM) were performed in PBS supplemented with 5 mM CaCl 2 .Under control conditions in the absence of tTGase, we found that calcium alone dramatically enhanced synuclein filament formation in PBS (see Fig. 7A; data not shown).This was likely due to a seeding effect of calcium phosphate crystals, since addition of 1-5% (by volume) of preformed and washed calcium phosphate crystals significantly reduced the lag time for synuclein filament formation.However, we found that the oligomerization of synuclein in the presence of tTGase was enhanced in PBS versus buffers such as Tris or HEPES.DTT did not appear to enhance tTGase activity, and so was omitted to prevent reduction of contaminating endogenous metals.
The peroxides of arachidonic acid (20:4) and docosahexaenoic acid (22:6) were prepared according to Armstrong et al. (67).Briefly, 10 mM fatty acid (final concentration) was incubated in 1 ml 0.1 M sodium borate, pH 9.0 in the presence of 0.1 mg lipoxygenase for 2 h at 4 o C. Samples were extracted three times with an equal volume of diethyl ether and the collected upper phases were evaporated with nitrogen.0.1 ml 0.1 M sodium borate, pH 9.0 was added, and samples were vortexed and evaporated for 5 min to remove residual ether and sonicated briefly (5 sec) in an ice bath.Lipid peroxide concentrations were calculated from the absorbance at 234 nm using a molar extinction coefficient of 27,000 M -1 cm -1 .Docosahexaenoic acid and arachidonic acid and their respective peroxides were used at 200 µM.The lysine-specific cross-linker DSP was used at 500 µ M, and dopamine at 200 µ M. The concentrations of these reagents were chosen to reflect the optimal production of oligomers, as assessed by immunoblotting with anti-synuclein antibodies, when synuclein was used at a concentration of 200 µM.Proportionally less of the cross-linking reagents were necessary at lower synuclein concentrations.

Enhanced metal catalyzed oxidation of synuclein in the presence of dithiothreitol (DTT)-
Although traditionally used as antioxidants in biochemical systems, reducing agents, such as ascorbate and DTT, are extremely effective in generating ROS in the presence of trace amounts of transition metals, principally iron and copper (68)(69)(70).We wished to compare the effectiveness of supplying reducing equivalents in the form of DTT with reagents used previously to study the metal-catalyzed oxidation of alpha synuclein (39)(40)(41)(42).Purified recombinant synuclein was incubated overnight at 37 o C with 10 µM of exogenous Cu(II) or Fe(III) in the presence of 300 µ M hydrogen peroxide.Lower concentrations of metals were used compared with other reports (39)(40)(41)(42) to test the sensitivity of synuclein to the different MCO systems, and because these levels are likely to be more physiologically relevant.Under these conditions, minimal oxidation of synuclein was observed, as reflected by the formation of low levels of synuclein oligomers (dimers, trimers, etc) (Fig. 1A, lanes 3, 4).No enhancement of oligomerization was observed if the incubations contained reduced form of the metals (Cu(I) or Fe(II)), nor was there a significant increase in the production of synuclein oligomers observed up to 50 µM metal, the maximum level tested (data not shown; although see below).In the presence of DTT, however, a substantial increase of synuclein oligomers were generated in the presence of copper (lane 7), and especially iron (lane 8).The oligomers were resistant to dissociation with 6 M urea, as well as SDS, thus demonstrating their covalent association (data not shown).Fe(II) and Fe(III) enhanced synuclein oligomerization equally in the presence of DTT, but were without effect alone (data not shown).This demonstrates that DTT effectively recycled oxidized iron to its reactive ferrous form, and at similar metal concentrations, was more effective in enhancing MCO of synuclein when compared to other systems (39)(40)(41)(42).Oligomerization was largely inhibited by the iron chelators deferoxamine (Df) and diethylenetriaminepentaacetic acid (DTPA) (Fig. 1A, lanes 9, 10), and clioquinol (not shown), but less so by the copper chelator bathocuproine disulfonic acid (Fig. 1A, lane 11).Even with DTT alone, a substantial increase in generation of synuclein oligomers was observed (Fig. 1A, lanes 5, 6).The effect of adding DTT alone raises the likelihood that redox cycling of trace metals present in the buffered solutions was sufficient to induce formation of synuclein oligomers.Consistent with this is the lack of oligomerization in the presence of DTT when trace iron in the solution is complexed by deferoxamine (Fig. 1A, lane 12) or when divalent metals are depleted by prior treatment with Chelex 100 beads (data not shown).A requirement for oxygen under these conditions was addressed by incubating synuclein solutions anaerobically.As can be seen in Fig. 1B, oligomerization of synuclein in the presence of DTT alone, as well as with DTT and added iron, was virtually eliminated in the absence of oxygen.These results demonstrate a critical role for oxygen in synuclein oligomerization, and provide evidence that the continued regeneration of reduced forms of transition metals, especially iron, can significantly enhance the metal-catalyzed oxidation of synuclein.Due to the complex nature of iron chemistry (15,18,71) and the observation that the site-specific "caged" generation of ROS may be unavailable to general free radical scavengers (15,69), we have not yet attempted to identify the actual oxidizing species generated under these conditions.

MCO prevents synuclein filament formation-
Since the role of synuclein oxidation during fibrillization has not been clearly defined (39,(42)(43)(44)72,73), the effect of MCO on the formation of synuclein filaments in vitro was examined.Treatment of 200 µM synuclein at 37 o C for 72 h with shaking under control conditions (i.e. in the absence of DTT or iron) resulted in the production of relatively uniform, mostly unbranched, but often twisted, 6-10 nm wide filaments up to several micrometers long (Fig 2A).Little in the way of nonfilamentous background was observed, or was the background increased when DTT/iron were added to monomeric synuclein or preformed synuclein filaments just prior to their deposition onto EM grids (data not shown).When DTT/iron were present during the incubation, however, a large number of primarily small 50-100 nm short curvilinear structures were generated, as well as a number of apparently anastomosing 100-250 nm networks (Fig. 2B) that may be derived from coalescence of the smaller structures, and which occasionally contained loops within the networks (see Fig. 2C, arrowhead).Although rare, 20 nm diameter annular structures were also observed (Fig 2D , arrow).These results demonstrate that in the presence of an enhanced MCO system, synuclein filaments fail to be generated and are replaced by an assortment of structures of varying morphologies.
To examine whether these oxidized assemblies contained a crossed-beta-sheet structure common to amyloid fibrils, the Thioflavin T dye binding characteristics of control and iron-treated synuclein solutions was compared.As shown in Fig. 3A, the nonincubated synuclein monomer did not bind dye, as expected from its natively unfolded conformation.
Synuclein incubated under control conditions (i.e.no additions) resulted in production of filaments, reflected by an increase in Thioflavin T fluorescence.Synuclein incubated in the presence of DTT alone generally prevented filament formation, although this effect was variable, and could be prevented by treating the DTT solutions with Chelex-100 beads to remove contaminating metals (see Experimental Procedures).Addition of exogenous iron in the presence of DTT (whether Chelex-treated or not) completely inhibited formation of beta-sheet containing structures, as shown by a lack of Thioflavin T binding.This was not due to quenching of the Thioflavin T signal by iron, since addition of a one hundred fold excess of exogenous iron (to 50 µ M Fe(III), see Experimental Procedures) to preformed synuclein filaments had no effect on fluorescence emission (data not shown).The Thioflavin T signal was quenched, however, by 50% at 250 µM Fe(III) and no signal was detected at 500 µM Fe(III) (data not shown).Conversely, addition of the iron chelator deferoxamine to the DTT/iron-containing solution restored filament formation to control levels, and enhanced filament formation when added to control incubations (Fig. 3A).No difference in the morphology of the synuclein filaments was observed in the presence of deferoxamine; however filaments began forming more rapidly, often appearing in less than 24 h (data not shown).These results suggest that synuclein oxidation/oligomerization, at least under the conditions described here, and synuclein filament formation, are divergent processes.This is supported by experiments examining filament formation under anaerobic conditions (Fig 3B ), where oxygen depletion significantly enhanced filament formation over control incubations performed in the presence of oxygen and, even in the presence of DTT/iron restored filament formation to control levels.

The role of synuclein oxidation in "protofibril" formation-
The soluble forms of oligomeric synuclein that elute in the void volume from Superdex 200 gel filtration columns after synuclein filaments have been removed by centrifugation have been defined as representing synuclein protofibrils, the possible toxic intermediates in the etiology of PD (49)(50)(51).Since the MCO system described here prevented the assembly of synuclein into filaments, we wished to address whether formation of synuclein protofibrils was similarly affected.Synuclein was incubated under control or oxidizing conditions as described above, insoluble material (including filaments) was removed by centrifugation, and the resulting supernatants were loaded onto a Superdex 200 column (Fig. 4).On-line light scattering detection, together with SDS-PAGE of isolated fractions, indicated that the untreated synuclein monomer eluted at 19 min with little or no material eluting in the void volume (Fig. 4A).Supernatants from control incubations without DTT/iron contained monomeric synuclein, as well as larger synuclein species that eluted in the void volume peak (Fig. 4B), as previously described (50).These larger synuclein species routinely represented less that 3% of the amount of monomeric synuclein.Smaller synuclein fragments were also generated that could be detected by silver staining and by immunoblot analysis using antibodies directed toward the Nterminus, but not the extreme C-terminus (data not shown), suggesting that C-terminal cleavage of synuclein was occurring during the incubations.The mechanism by which these synuclein fragments are generated is currently under investigation.The molecular mass of the void volume oligomers was approximately 5.4 x 10 5 kDa, as determined by on-line light scattering, representing about 37 synuclein subunits, similar to what has been previously reported using gel filtration and scanning transmission electron microscopy (74).
Incubation with DTT/iron resulted in an approximately five to twenty-fold enhancement in production of synuclein protofibrils, as well as generation of a range of species that vary from small oligomers (dimers, etc.) up to larger oligomers that elute just after the void volume (Fig. 4C, see below).Conversely, incubation in the presence of deferoxamine almost completely prevented protofibril formation, as well as generation of smaller synuclein oligomers (Fig. 4D).The amount of monomeric synuclein remaining in the supernatant was reduced approximately two-fold in the presence of deferoxamine when compared with control incubations, although the production of truncated synuclein fragments appeared not to be affected (compare Fig. 4D with 4C).This is consistent with the results from Figure 3, in which filament assembly was enhanced when oxidation was prevented.These results demonstrate that synuclein oxidation enhances the production of synuclein protofibrils, and raise the possibility that oxidation/covalent cross-linking of synuclein may actually be required for their formation.The demonstration that the in vitro treatment of synuclein with dopamine and other catecholamines result in enhanced protofibril formation that was inhibited when oxidation was prevented (72), is consistent with this view (see below).

Characterization of synuclein oligomers-
To further characterize the oligomers generated by DTT/iron treatment, oligomers were separated from the oxidized synuclein monomer by Superose 6 gel filtration chromatography (Fig 5A).Although some monomeric synuclein eluted with the higher molecular mass oligomers (fractions 20-28), most of the monomeric synuclein eluted later from the column (fractions [36][37][38].Given the effectiveness of the separation, it is likely that the low levels of monomeric synuclein observed in fractions 20-26 represents molecules dissociated from higher molecular mass oligomers by SDS rather than the failure of the column to separate unincorporated monomer from the oligomers.Similar results were observed when purified synuclein protofibrils were isolated by Superdex 200 gel filtration chromatography (from Fig. 3C) and analyzed by SDS-PAGE/immunoblotting (data not shown).Thus, higher molecular mass oligomers likely consist of both SDS-sensitive and SDS-resistant synuclein (54).Secondary structure analysis of purified oligomers by far-UV CD showed a transition from the unfolded conformation of native alpha synuclein, characterized by a maximum negative ellipticity at ca. 200 nm (75), to a likely beta-sheet structure with a minimum around 218 nm (Fig. 5B).This spectrum is similar to that previously reported for purified protofibrils (50).Significant light scattering prevented us from more accurately determining the percentage beta-sheet content within these structures (76).These CD changes were not associated with binding of Thioflavin T (as expected from Fig. 3) or with binding to the hydrophobic dye bis-ANS (or ANS), nor did oligomerization result from the formation of appreciable dityrosine cross-links, as monitored by changes in intrinsic fluorescence (data not shown).Thus, although the synuclein oligomers generated by oxidative cross-linking contain beta-sheet, they lack the amyloid structure characteristic of synuclein filaments (77).It has not yet been determined the minimal synuclein species that contributes to the beta-sheet conformation (small o l i g o m e r s , o r l a r g e r s y n u c l e i n oligomers/protofibrils).
Given that treatment of synuclein with DTT/iron inhibited filament formation, it is unlikely that the oligomers produced this treatment would function as seeds to promote synuclein fibrillization, as is the case with seeds from preformed filaments (62).Indeed, inclusion of 1% purified oligomers (based on total synuclein concentration) to 200 µM untreated monomeric synuclein failed to stimulate filament formation after 18 h incubation, compared with significant fibril formation when 1% sonicated control seeds were added (Fig. 5C).However, the oligomeric seeds did not inhibit subsequent filament formation during more extended incubations (Fig. 5C, 48 h).Sedimentation analysis (Fig. 5D) demonstrated that the soluble oligomer seeds became incorporated into pelletable material after 48 h at 37 o C (compare lanes 5, 6 with 7, 8).Thus, at these low concentrations, synuclein oligomers do not function as crystal poisons that inhibit nucleation-dependent fibrillization, (78), but can assemble with unmodified monomeric synuclein into fibrils.Whether higher proportions (>1%) of purified synuclein oligomers would inhibit fibrillization of unmodified synuclein, similar to results with hydrogen peroxide-treated (73) or nitrated synuclein (79), has not been tested.Also shown in Fig. 5C is the observation that the oxidized synuclein monomer that had been purified by gel filtration (fractions 36-38, Fig. 5A) was incapable of assembling into fibrils after 48 h, or even after more extended incubations (up to 5 days tested; data not shown).These results are similar to those of Fink and colleagues, who demonstrated that progressive oxidation of the four methionines of synuclein to methionine sulfoxide by hydrogen peroxide prevented fibril formation in the absence of oligomer formation (73).The nature of the modification(s) to monomeric synuclein under our oxidizing conditions has not been determined.

Inhibition of synuclein filament formation by cross-linked synuclein oligomers-
Given the above results, we wished to address the general issue of whether treatments previously shown to generate SDS-resistant synuclein crosslinks would also inhibit synuclein filament formation.These include: DTT/iron, as described above; tissue transglutaminase (tTGase) (48), which catalyzes the calcium-dependent formation of an e-(g-glutamyl)lysine isopeptide bond between glutamine and lysine residues; PUFAs, such as arachidonic acid (45) and docosahexaenoic acid (46), and their respective lipid peroxides; synthetic cross-linkers, such as DSP (59); and the pro-oxidant dopamine (72).Filament formation was monitored by both Thioflavin T fluorescence and by sedimentation analysis.To eliminate the effects of non-specific synuclein oxidation during long-term (72 h) incubations, "seeded" reactions were used, which resulted in efficient filament formation in as little as 12 h.Compared with control reactions (Fig. 6A, and 6B, lanes 1, 2), treatment with DTT/iron inhibited formation of synuclein filaments (as previously shown in Fig. 3A), even when incubated with control seeds, and resulted in the production of oligomers that remained soluble during centrifugation (Fig 6B, lanes 3, 4).In the absence of Ca 2+ , tTGase had no effect on filament assembly or in the amount of synuclein pelleted by centrifugation, with only minor amounts of dimeric synuclein generated during the incubation, similar to control conditions (Fig. 6B, compare lanes 1, 2 with 5, 6).However, in the presence of 5 mM Ca 2+ , fibril formation was blocked, and oligomeric forms of synuclein were generated, but remained soluble (Fig. 6B, lanes 7, 8).Interestingly, Ca 2+ alone stimulated fibril formation over control conditions (Fig. 6A), presumably due to an enhanced seeding effect of calcium phosphate crystals (see Experimental Procedures).Incubation in the presence of docosahexaenoic acid resulted in a moderate decrease in fibril assembly and a concomitant increase in formation of soluble oligomers (Fig. 6B, lanes 9, 10).Enzymatically generated docosahexaenoic acid lipid peroxides enhanced this effect, completely disrupting filament formation (Fig. 7A) while inducing formation of almost exclusively soluble oligomers (Fig. 6B, lanes 11, 12).Similar results were observed with arachidonic acid, although higher concentrations of lipid peroxide were required (data not shown).Thus, oxidation of PUFAs dramatically enhance synuclein cross-linking, which is likely due to the formation of reactive aldehydes derived from lipid peroxide decomposition (80) (see Discussion).The lysine-directed cross-linking agent DSP and the oxidant dopamine both inhibited fibril formation and resulted in the formation of mostly soluble oligomers (Fig. 6B, lanes 13, 14 and 15, 16, respectively).Thus, both enzymatic (e.g.tTGase) and oxidative (DTT/iron, lipid peroxidation, dopamine) processes, as well as synthetic crosslinking reagents (e.g.DSP) inhibit synuclein filament formation by forming covalently associated soluble synuclein oligomers.
Finally, it has been reported that while both nitrative and oxidative stresses resulted in production of synuclein oligomers, only nitrative stress induced by peroxynitrite treatment prevented filament formation (42).Consistent with this, we also observed that peroxynitrite treatment induced the formation of soluble synuclein oligomers and prevented filament formation (data not shown).However, in contrast to previous results (42), we found that treatment with 0.5 mM Cu(II) and 0.3 mM hydrogen peroxide prevented formation of synuclein filaments (Fig. 6C) and resulted in generation of soluble soluble oligomers (Fig. 6D).However, treatment with the same concentrations of Fe(III) and hydrogen peroxide failed to prevent filaments from forming and did not result in synuclein oligomerization (Fig. 6C,  D).

Effects of oxidized synuclein on the 20S proteasome-
It has previously been demonstrated that nonubiquitinylated monomeric synuclein is effectively degraded by purified 20S proteasomal particles (81).The lack of a requirement for the unfolding activity of the 19S cap is likely a consequence of synuclein's already natively unfolded structure (75).More recently, using fluorogenic peptides as substrates, it has been reported that high concentrations of monomeric as well as significantly lower concentration of amyloidcontaining synuclein filaments and oligomers inhibit purified 20S proteasomes in vitro (54,55).Given the role of DTT/iron in enhancing the oligomerization of synuclein, we asked if these oxidized synuclein species would be efficiently degraded by the proteasome.As shown in Fig. 7, monomeric synuclein was effectively degraded in a proteasome-dependent manner, as reported ( 81), which could be blocked in the presence of the inhibitor MG132.In contrast, degradation of oxidized oligomeric synuclein was almost completely prevented.Interestingly, degradation of the monomeric component of this oxidized mixture was also blocked (Fig. 7).These results are consistent with previous reports that oligomeric synuclein inhibits the activity of the proteasome (54,55).However, when the activities of the 20S proteasome were assayed by hydrolysis of fluorogenic peptides, no significant differences were observed between untreated and oxidized synuclein, except for a minor, increase in the trypsin-like activity in the presence of oxidized synuclein (Table I).We did find, however, that both monomeric and oxidized synuclein inhibited the glutamyl-peptidase hydrolyzing activity by ~45%.Therefore, the failure to specifically degrade oxidized synuclein is not likely a result of active-site inhibition, but may reflect structural alterations that render oxidized/oligomerized synuclein inaccessible to the 20S catalytic subunits (see Discussion).

DISCUSSION
The role of oxidation in the pathogenesis of PD, and more specifically in the relationship of alpha synuclein oxidation to pathology remains unclear.In this paper, we sought to improve our understanding of the biochemical transitions of alpha synuclein under oxidizing conditions, that a more defined a role for synuclein oxidation in neuronal degeneration may be developed.We find that in the presence of the reducing agent DTT, the metal catalyzed oxidation (MCO) and oligomerization of synuclein is dramatically stimulated when compared with previously used MCO systems.This is thought to occur by recycling of oxidized metal ions into their redox active forms with subsequent catalytic activation of molecular oxygen (57).Although DTT itself is not physiologically relevant, abundant intracellular reductants (e.g.ascorbate, reduced flavin and pyridine nucleotides, and low molecular weight thiol-containing compounds, including metabolites of glutathione) have the capacity to reduce iron to its ferrous form under control and pathological conditions (82)(83)(84)(85)(86).As such, reduced forms of transition metals play an integral role in the generation of highly reactive "oxyradicals" capable of oxidizing biomolecules such as proteins, lipids, and DNA (69,71).
In the presence of iron and DTT, synuclein filaments are not generated.Instead, small, curvilinear as well as larger interconnected networks are formed, with occasional annular structures observed, similar to those described by Lansbury and colleagues (74).These oxidized structures do not effectively bind Thioflavin T (Fig. 3), and as such are not amyloid, yet likely contain beta-sheet structure, as determined by far-UV CD spectroscopy.Bis-ANS (or ANS) binding is also not observed, demonstrating that extensive hydrophobic patches are not exposed during oxidation/oligomerization.This is consistent with previous results in which only weak binding of ANS to synuclein aggregates could be detected at neutral pH (87), although see (88).Although synuclein filaments fail to be generated under oxidizing conditions, formation of both SDSresistant synuclein oligomers (as assessed by immunoblotting after SDS-PAGE) and synuclein protofibrils (as defined by gel filtration chromatography) are enhanced (Fig. 4).Conversely, iron chelation or oxygen depletion prevents the generation of both oligomers and protofibrils while enhancing formation of synuclein fibrils.These results raise the possibility that rather than functioning as a precursor in synuclein filament formation, as has been suggested (89), extensive formation of synuclein protofibrils represents an off pathway, with synuclein oxidation/oligomerization actually required for protofibril formation.Indeed, the results of Conway and colleagues ( 72) are consistent with this view.Treatment of synuclein with dopamine or other catecholamines enhanced protofibril formation, including the formation of low levels of dopamine-synuclein adducts, and led to a reduction in the production of synuclein filaments.These effects required oxidizing conditions and could be inhibited by iron chelators and antioxidants (72).Thus, synuclein protofibrils may simply represent soluble nonamyloid oligomeric forms of synuclein, in which a significant proportion is covalently associated (see below).
The relationship between synuclein oxidation/oligomerization as it relates to filament formation is unclear.Our results indicate there is an inverse relationship between formation of soluble SDS-resistant synuclein oligomers and filament formation.It is not clear, however, whether the inhibition of filament formation we observe requires covalent cross-linking per se, or rather results from specific modification(s) of amino acid side chains of the synuclein monomer, with oligomerization serving only as a consequence of these modification(s).It is reasonable to assume that covalent cross-linking of synuclein is not required to inhibit fibrillogenesis.For example, oxidation of methionine residues with hydrogen peroxide (73) or nitration of tyrosine residues with tetranitromethane (79) inhibit synuclein fibril formation in the absence of covalent association, and proportionally inhibit filament formation when added to unmodified synuclein.This is thought to occur via formation of hetero-oligomers between modified and unmodified synucleins, in the absence of covalent cross-linking (73,79).Thus, the effects of synuclein oligomerization on filament assembly may be related not only to the types of modifications generated, but also to the extent to which the synuclein monomer population is modified.
On the other hand, it has been suggested that covalent cross-linking of synuclein may actually promote fibril formation (44).This hypothesis is based primarily on an observed increase in lag time for fibril formation when synuclein is incubated for extended periods of time (15-20 days) in the presence of EDTA and the anti-oxidant methionine.This delay in fibril formation was proposed to be due to the failure to form dityrosine-linked dimers that served to nucleate fibril assembly (44).Our observations, however, that filament formation is accelerated and enhanced, not delayed (44) under anti-oxidant conditions (chelating endogenous metals or oxygen depletion, see Fig. 3) are inconsistent with this view, unless undetectable levels of dityrosinelinked synuclein are generated under chelation and anoxic conditions that lead to the enhanced filament assembly.
It has been reported that addition of relatively high levels of exogenous copper (0.5 mM) to hydrogen peroxide (0.3 mM) resulted in formation of SDS-resistant synuclein oligomers, yet no effect on fibril assembly was observed (42).This contrasts with our observations that production of soluble synuclein oligomers under these conditions correlates with an inhibition of filament formation (Fig. 6C, D).At present, it is unclear as to the nature of this difference, as different detection methods were used to monitor the effectiveness of synuclein oligomerization.In contrast to the effects of copper, we did observe that relatively high concentrations of iron in the presence of hydrogen peroxide do not result in the significant production of synuclein oligomers nor prevent fibril formation (Fig. 6, C and D).This contrasts with the complete inhibition of fibril assembly with significantly lower concentrations of iron in the presence of DTT (Fig 7A).Conversely, copper is much less effective than iron in inducing synuclein oligomerization in the DTT system (see Fig. 1A; data not shown).Thus, it appears that under the different oxidizing systems, copper and iron affect synuclein oligomerization and filament assembly differently.Whether this reflects distinct binding sites for the different metals (90) that result in differences in the site-specific generation of ROS (15,69), is unclear.That different metals may target distinct regions of the synuclein molecule is supported by studies indicating that oligomerization of synuclein in the presence of copper and hydrogen peroxide generated dityrosine cross-links, whereas no dityrosine formation could be observed in the presence of iron (41).
It must be emphasized that free radicalmediated oxidation of proteins can lead to a number of modifications of amino acid side chain residues, including the formation of reactive carbonyl groups (ketones and aldehydes) that render proteins susceptible to intra-and intermolecular cross-linking (91) (see below).Thus, the covalent association of proteins under oxidizing conditions is not restricted to dityrosine formation.Indeed, tyrosine-deficient synuclein retains the capacity to form oligomers under oxidizing conditions (42), and in the absence of peroxynitrite, we were unable to detect appreciable dityrosine formation under any of our oxidizing conditions (data not shown).Rather, each of the treatments that generate synuclein cross-links in our system are likely to occur through modifications of one or more of the fifteen lysine residues present in the N-terminal twothirds of synuclein.One of the main products of MCO of proteins is derivatization of lysine residues to aminoadipic semialdehyde (92), and tTGase (65,66), lipid peroxide decomposition products (80), and the cross-linker DSP all have the capacity to introduce cross-links into proteins through lysine modifications.In contrast, out of the four tyrosine residues that may be susceptible to dityrosine formation, three are located in the unstructured C-terminal tail of synuclein.Thus, the possibility remains that cross-linking via the N-terminal and central regions of synuclein abrogate fibril formation, whereas cross-linking via the C-terminus of synuclein does not.A careful examination of the nature of these crosslinks and amino acid side chain modification(s) generated under different oxidizing conditions is necessary before deciding on a mechanistic role for synuclein oxidation in oligomer formation and possible PD pathogenesis.
Based on the available evidence and the results presented here, a potential model relating synuclein oxidation/oligomerization to fibril assembly can be envisioned (Fig. 8).The normally unmodified synuclein monomer (1) is susceptible to oxidative modifications (2), especially during long-term incubations in vitro.These include oxidation of histidinyl residues to oxo-histidine and its degradation products and conversion of methionine to methionine sulfoxide (92).Carbonyl groups, introduced principally by MCO of specific amino acid side chains, include glutamic and aminoadipic semialdehydes resulting from oxidation of prolyl and lysyl residues, respectively (91,92), and in the presence of PUFAs can also be introduced by adduction of aldehydic lipid peroxide degradation products [e.g.malondialdehyde (MDA) and 4-hydroxy-2nonenal (HNE)] that also target histidinyl and lysyl residues (80).Indeed, purified recombinant synuclein can be cross-linked in the presence of both HNE and MDA (data not shown).Modification(s) of the synuclein monomer that do not result in species capable of subsequent intraor inter-molecular cross-linking (e.g. by hydrogen peroxide or tetranitromethane) become incapable of assembling into filaments, in the absence of extrinsic modifiers (3) (73,79,93).Modification(s) resulting in reactive carbonyl formation (e.g.MCO, lipid peroxidation) or through enzymatic processes (e.g.tTGase) are capable of generating intra-and inter-molecular cross-links (4) that can progress from dimers up to large soluble synuclein conglomerates that may represent beta-sheet rich but nonamyloid synuclein "protofibrils" (5) that, when accumulated, inhibit filament formation.Conversely, when oxidation/oligomerization is prevented, protofibrils do not accumulate and assembly of synuclein into amyloid filaments becomes more efficient (6).
Fink and colleagues have suggested that synuclein undergoes a transition from the natively unfolded monomer into a more compact partially folded intermediate that ultimately leads to fibril formation (94,95).For example, fully reversible structural changes are induced by brief incubations at elevated temperatures (95).Extended incubations (with heating), however, lead to the irreversible formation of cross-linked synuclein that was proposed to stabilize an intermediate in the fibril assembly pathway (88).Rather, we favor the idea, that if cross-linking becomes extensive, and involves a large fraction of the monomer population, then these oligomers are removed from the pathway leading to filaments (7).This two-pathway model (i.e.soluble oligomers versus filaments) is similar to that suggested by Fink and colleagues (96), although we have no information on the formation of a partially folded monomeric i n t e r m e d i a t e d u r i n g s y n u c l e i n oxidation/oligomerization, nor do we have evidence of whether the covalently associated synuclein oligomers generated under our oxidizing conditions can reenter the fibril assembly pathway in the presence of modifiers, such as Zn(II) (96) or low pH (79).
Complicating the above model is the observed transient nature of the protofibrils described by Lansbury and colleagues (49,89) and cross-linked dimers by Krishnan and colleagues (44).During incubations, these structures appear then disappear as filaments are produced.These results suggest that during the mildly oxidizing conditions encountered during long-term incubations and in the absence of exogenous oxidizing species, low levels of synuclein oligomers may be generated that subsequently reenter the fibril pathway (8), possibly functioning as nucleating centers, as proposed by Krishnan and colleagues (44).Indeed, our observation that purified synuclein oligomers become incorporated into synuclein filaments when supplied with a large excess of unmodified monomer supports the idea that when oxidation is limited, oligomers are capable of becoming assembled into filaments, acting neither as seeds nor poisons.However, extensive oxidation, in which a significant proportion of the synuclein population becomes modified, irrespective of whether these modifications ultimately result in oligomerization, renders synuclein incompetent to assemble into fibrils.
Oxidized synuclein inhibits the 20S proteasome-Although it has not been determined whether the generation of synuclein oligomers within diseased tissue represents a cause or a consequence of the pathology, it is likely that their accumulation is reflected by insufficient degradation by cellular proteases.The discovery of gene defects in familial PD linked to protein turnover by the proteasome (e.g.parkin, UCH-L1), as well as evidence for impairment of the ubiquitinproteasome system (UPS) in sporadic PD, has led to the proposal that dysregulation of protein turnover by the proteasome represents a major aspect in the pathogenesis of PD (3).Although the normal cellular mechanism for synuclein turnover has been debated (97), purified recombinant synuclein is efficiently degraded by isolated 20S proteasomes without participation of the 19S regulatory cap or the need for ubiquitinylation (81).Previously, it has been reported that monomeric synuclein has little effect on the activity of purified erythrocyte 20S proteasomes, except at high concentrations (54,55).We found, however, using 20S proteasomes purified from rat liver, that at 1 µM (representing a twenty-fold molar excess over the 20S core particle), monomeric synuclein inhibit the glutamylpeptidase activity by ~45%, whereas the trypsinlike and chymotrypsin-like activities are relatively unaffected.Whether the decrease in glutamylpeptidase activity is through substrate competition or an allosteric effect mediated through binding of monomeric synuclein to noncatalytic site(s) on the 20S particle (see (98,99)), is unknown.Similar mechanism(s) may play a role in the increase in trypsin-like activity observed with oxidized synuclein (see Table I).
Previously, oligomeric and fibrillar forms of synuclein were shown to effectively inhibit the chymotrypsin-like activity of purified 20S proteasomes, when assayed by proteolysis of small fluorogenic peptides (54,55).This was demonstrated to occur through a non-competitive mechanism, and thus likely occurred through direct binding to the proteasomal subunits themselves (55).By immunoblotting, we observe that oxidized synuclein is poorly degraded by 20 S proteasomes.Yet, this appears not to be due to inhibition of the various proteasomal activities, since no additional effects are detected using fluorogenic peptides when compared to unmodified synuclein (see Table I).Rather, this inhibition may be due to a failure of cross-linked synuclein to act as an efficient 20S proteasomal substrate, for example, through structural alterations that occur during synuclein oxidation (cross-links, etc.).Failure to degrade the monomeric portion of synuclein within the oxidized sample may occur through steric inhibition by synuclein oligomers, thereby preventing access of monomers to the 20S active site(s), or by the stable association of synuclein monomers with cross-liked species.Lindersson and colleagues have shown that inhibition of the chymotrypsin-like activity of the proteasome required an amyloid-like structure of synuclein oligomers (55).Since the oligomers generated under our oxidizing conditions do not contain amyloid (see Figure 3), and are not recognized by anti-oligomer antibodies (61) (data not shown), the inhibition we observe is likely mechanistically dissimilar to that reported previously (55).Thus, discounting possible differences due to the source of proteasomes, various oligomeric forms of synuclein (amyloid/non-covalently associated versus non-amyloid/covalently-cross-linked) can have different net effects on proteasomal activity and substrate turnover.Nevertheless, our results suggest that failure to clear oxidatively modified synuclein assembled into higher molecular mass oligomers may ultimately result in a cascade of events leading to cell degeneration in PD and other synucleinopathies (52)(53)(54)(55).Determining the chemical nature of these modifications that lead to synuclein cross-linking and ultimately to synuclein deposition within insoluble aggregates (31,(35)(36)(37) will be an important aspect of future PD research. 1The abbreviations used are: PD, Parkinson's disease; MCO, metal catalyzed oxidation; ROS, reactive oxygen species; PUFA, polyunsaturated fatty acid; DTT, dithiothreitol; DSP, dithiobis [succinimidyl propionate; bis-ANS, 4-4'-dianilino-1-1'-binaphthyl-5-5'-disulfonic acid.
2 Footnote: The terms "protofibril" and "protofilament" have been used to represent distinct onpathway intermediates that forms during protein filament assembly, often containing filamentlike (100,101) or curved (74,102,103) morphologies.To avoid confusion, we will use the term synuclein "protofilbril" to represent nonfibrillar, nonfluorogenic (or weakly fluorogenic) synuclein oligomers that elute in the void volume from Superdex 200 or Superose 6 gel-filtration columns, as originally described by Lansbury and colleagues (74,104).1. Dithiothreitol (DTT) enhances metal catalyzed oxidation (MCO) of human alpha synuclein.A. 0.1 mg/ml purified recombinant human alpha synuclein was treated under different oxidizing conditions in 20 mM sodium phosphate buffer, pH 7.4 at 37 o C for 16 hr.SDS sample buffer was added, the samples were boiled for 5 min and 500 ng of synuclein was loaded on a 10-20% Tris-tricine gel.Detection was with an antibody (202) that recognizes synuclein's extreme C-terminus.Molecular mass markers are on the left.Unit synuclein oligomers are designated n=1 (monomer), n=2 (dimer), n=3 (trimer), etc. Deferoxamine (Df); bathocuproine disulfonic acid (BCS) B. Stock solutions (synuclein, DTT, iron, buffer) were exposed separately to an anoxic atmosphere for 3 hr at room temperature before being diluted and incubated as in (A).Incubations were in the presence (left panel) or absence (right panel) of oxygen.The apparent increase in the amount of synuclein (monomeric and oligomeric) under oxidizing conditions (i.e. with DTT/iron) in (A) and (B) may be due to conformational changes that accompany synuclein oxidation, and result in an increase in detection with anti-synuclein antibodies.These effects were also observed by silver staining (not shown).With higher levels of iron, increased oligomerization does result in progressive loss of monomeric synuclein into higher molecular mass oligomers (see Fig 5A).Excitation was at 440 nm, and emission scans at 450-600 nm.The left panel (A) shows the effects of iron treatment and iron chelation on synuclein filament formation under oxygen containing conditions, while the right panel (B) compares the effect of oxygen containing (-ox) or oxygen depleted (-anoxic) solutions on filament formation.Monomeric synuclein (nonincubated) and synuclein incubated under control conditions in the absence of oxidants (no additions) are indicated.
Figure 4. Oxidation with DTT/iron stimulates synuclein protofibril formation.200 µM synuclein solutions were incubated for 72 h as in Figure 3.After incubation, the solutions were centrifuged at 16,000 x g for 10 min to remove filaments and other large aggregates, and the supernatants loaded onto a Superdex HR200 gel filtration column.Detection of the eluates was monitored at 280 nm (UV, solid line) and by on-line light scattering to determine the masses of the eluting species (LS, dotted line).At 12-14 min for all samples, the sensitivity of the UV detector was decreased from 0.1 mV/au (absorbance unit) to 0.5 mV/au to prevent saturation from the monomer signal.A, untreated synuclein monomer, 350 ug; B, control incubation; C, incubation with DTT/iron; and D, control incubation with deferoxamine.Note the scale differences in the ordinates under the different conditions.Figure 5. Purification and structural analysis of synuclein oligomers.200 µM synuclein was oxidized in PBS containing 10 mM DTT/100 µM Fe(III) overnight at 37 o C. A. Oligomers were separated from the monomer by Superose 6 gel filtration chromatography. 1 mL fractions were collected, and oxidized oligomers (fractions 20-28) and monomer (fractions 36-38) were concentrated and dialyzed against PBS overnight.B. Far-UV CD spectroscopy of purified oligomers and untreated synuclein monomer.C. Thioflavin T assay.18 or 48 h incubations with 200 µM unseeded (control) synuclein or synuclein containing 1% preformed filament seeds or 1% purified oligomer seeds.At 18 h synuclein filaments were generated in the presence of filament seeds (dark blue broken line), but not in their absence (solid red line).Oligomer seeds have no enhancing effect at 18 h (green broken line), but by 48 h have reached control levels (compare red and black dotted lines).Also note that the oxidized monomer fails to generate fibrils.D. Sedimentation assay.Incubations from 48 h treatments in (C) were pelleted, and equal volumes of supernatant (s) and pellet (p) fractions were analyzed by SDS-PAGE and immunoblotting with the anti-synuclein antibody 202.Untreated synuclein is shown in lanes 1 and 2. In order to detect the distribution of the "seeds", which were added at 1% of monomeric synuclein, up to 10 ug total synuclein was loaded in each lane.Purified oligomeric synuclein used at the same ratio (1%) in the seeded reactions (100 ng) is shown in lane 9.A three-fold excess of purified oligomers (300 ng) are shown in lane 10 to illustrate that low levels of synuclein monomer dissociate from the oligomers in the presence of SDS.A mixture of monomeric synuclein and oligomeric seeds, but maintained at 4 o C is shown in lanes 7 and 8.Note that the oligomers remain in the supernatant after centrifugation under these conditions.Figure 7. Oxidation of synuclein prevents its degradation by purified 20S proteasomes.Control (untreated) or DTT/iron treated synuclein was incubated with purified 20S proteasomal particles at 37 o C for 1 hr, in the presence or absence of MG132, as indicated.Synuclein was detected with the monoclonal anti-synuclein antibody, LB 509.Note the relative resistance of oxidized synuclein (monomer as well as oligomers) to proteasomal degradation.The natively unfolded synuclein monomer (1) can undergo oxidative modifications (2) resulting in molecules refractory (3, stars) or sensitive (4, lightning bolts) to subsequent intra-and intermolecular cross-linking.Depending on the level of oxidation and extent of oligomerization, betasheet rich soluble oligomers form (4) that may progress to synuclein protofibrils (5).Partially folded intermediates that may normally function as early intermediates during filament formation, can enter the oxidative pathway when oxidation/oligomerization becomes extensive (6).At high levels of oxidation, the entire population of synuclein molecules are likely to be affected, rendering synuclein incapable of forming filaments, as well as becoming resistant to degradation by the proteasome.At low levels of oxidation, where protofibrils (or dityrosine cross-linked dimers) represent a small proportion of the available synuclein pool, reentry into the pathway leading to filaments may occur, with these structures possibly serving as nucleating centers (7).In general, oxidation reduces the capacity to form filaments, whereas when oxidation is prevented fibril formation is enhanced (8).

Figure
Figure1.Dithiothreitol (DTT) enhances metal catalyzed oxidation (MCO) of human alpha synuclein.A. 0.1 mg/ml purified recombinant human alpha synuclein was treated under different oxidizing conditions in 20 mM sodium phosphate buffer, pH 7.4 at 37 o C for 16 hr.SDS sample buffer was added, the samples were boiled for 5 min and 500 ng of synuclein was loaded on a 10-20% Tris-tricine gel.Detection was with an antibody (202) that recognizes synuclein's extreme C-terminus.Molecular mass markers are on the left.Unit synuclein oligomers are designated n=1 (monomer), n=2 (dimer), n=3 (trimer), etc. Deferoxamine (Df); bathocuproine disulfonic acid (BCS) B. Stock solutions (synuclein, DTT, iron, buffer) were exposed separately to an anoxic atmosphere for 3 hr at room temperature before being diluted and incubated as in (A).Incubations were in the presence (left panel) or absence (right panel) of oxygen.The apparent increase in the amount of synuclein (monomeric and oligomeric) under oxidizing conditions (i.e. with DTT/iron) in (A) and (B) may be due to conformational changes that accompany synuclein oxidation, and result in an increase in detection with anti-synuclein antibodies.These effects were also observed by silver staining (not shown).With higher levels of iron, increased oligomerization does result in progressive loss of monomeric synuclein into higher molecular mass oligomers (see Fig5A).

Figure 2 .
Figure 2. Oxidation with DTT/iron inhibits synuclein filament formation and generates curvilinear and branched networks.Negative stain electron microscopy, under control (A) or oxidizing (B-D) conditions.200 µM synuclein was incubated at 37 o C for 72 hr with shaking in PBS, without (A) or with (B-D) 10 mM DTT/100 µM Fe(III).Scale bar equals 100 nm (A-C) and 20 nm (D).The oxidized assemblies in (B) often contained loop-like (C, arrowhead), and occasional annular (D, arrow) structures.

Figure 6 .
Figure 6.Synuclein cross-linking under various conditions.A, B. Seeded reactions were incubated for 18 h at 37 o C in PBS with no additions or in the presence of: 10 mM DTT/50 µM Fe(III); tTGase (50 nM) without or with 5 mM Ca 2+ ; 5 mM Ca 2+ alone; docosahexaenoic acid (DHA; 200 µM); docosahexaenoic acid peroxide (DHA perox; 200 µM); DSP (500 µM); and dopamine (200 µM). A. Filament formation as detected by Thioflavin T fluorescence, as in Fig. 3. B. Sedimentation analysis from samples in A. C, D. Seeded reactions were incubated for 18 h at 37 o C in PBS with no additions or in the presence of: 300 µM H 2 O 2 alone or with 500 µM

Figure 8 .
Figure 8. Model of synuclein oxidation/oligomerization as it relates to filament formation.The natively unfolded synuclein monomer (1) can undergo oxidative modifications (2) resulting in molecules refractory(3, stars)  or sensitive (4, lightning bolts) to subsequent intra-and intermolecular cross-linking.Depending on the level of oxidation and extent of oligomerization, betasheet rich soluble oligomers form (4) that may progress to synuclein protofibrils(5).Partially folded intermediates that may normally function as early intermediates during filament formation, can enter the oxidative pathway when oxidation/oligomerization becomes extensive(6).At high levels of oxidation, the entire population of synuclein molecules are likely to be affected, rendering synuclein incapable of forming filaments, as well as becoming resistant to degradation by the proteasome.At low levels of oxidation, where protofibrils (or dityrosine cross-linked dimers) represent a small proportion of the available synuclein pool, reentry into the pathway leading to filaments may occur, with these structures possibly serving as nucleating centers(7).In general, oxidation reduces the capacity to form filaments, whereas when oxidation is prevented fibril formation is enhanced(8).