α-Synuclein Protofibrils Inhibit 26 S Proteasome-mediated Protein Degradation

The impaired ubiquitin-proteasome activity is believed to be one of the leading factors that contribute to Parkinson disease pathogenesis partially by causing α-synuclein aggregation. However, the relationship between α-synuclein aggregation and the impaired proteasome activity is yet unclear. In this study, we examined the effects of three soluble α-synuclein species (monomer, dimer, and protofibrils) on the degradation activity of the 26 S proteasome by reconstitution of proteasomal degradation using highly purified 26 S proteasomes and model substrates. We found that none of the three soluble α-synuclein species impaired the three distinct peptidase activities of the 26 S proteasome when using fluorogenic peptides as substrates. In striking contrast, α-synuclein protofibrils, but not monomer and dimer, markedly inhibited the ubiquitin-independent proteasomal degradation of unstructured proteins and ubiquitin-dependent degradation of folded proteins when present at 5-fold molar excess to the 26 S proteasome. Together these results indicate that α-synuclein protofibrils have a pronounced inhibitory effect on 26 S proteasome-mediated protein degradation. Because α-synuclein is a substrate of the proteasome, impaired proteasomal activity could further cause α-synuclein accumulation/aggregation, thus creating a vicious cycle and leading to Parkinson disease pathogenesis. Furthermore we found that α-synuclein protofibrils bound both the 26 S proteasome and substrates of the 26 S proteasome. Accordingly we propose that the inhibitory effect of α-synuclein protofibrils on 26 S proteasomal degradation might result from impairing substrate translocation by binding the proteasome or sequestrating proteasomal substrates by binding the substrates.

Parkinson disease (PD) 2 is a common age-associated neurodegenerative disorder affecting ϳ2% of the population aged 65 years or older (1). The pathological hallmarks of PD are the loss of dopaminergic neurons in substantia nigra pars compacta and the presence of cytoplasmic inclusions called Lewy bodies. Lewy bodies contain dozens of proteins and lipids with ␣-synuclein (␣Syn), which is deposited as fibrillar forms, as the predominant protein (2,3). ␣Syn is a 140-amino acid protein found enriched in the presynaptic terminus. Structurally ␣Syn adopts a random coil conformation in solution, whereas the N-terminal amphipathic domain turns into an ␣-helical structure when associated on anionic membrane surfaces (4). The possible physiological functions of ␣Syn include regulating synaptic plasticity, dopamine neurotransmission, endoplasmic reticulum/Golgi trafficking, and acting as a molecular chaperone (5). An abundance of evidence has implicated ␣Syn in PD pathogenesis. Genetically three point mutations in the ␣Syn gene (A30P, E46K, and A53T) are associated with rare autosomal dominant forms of PD (6 -8). In addition, duplication and triplication of the ␣Syn gene can cause early onset familial PD (9 -12), indicating that simply increasing ␣Syn concentration is enough for disease pathogenesis. Biochemically ␣Syn aggregation is a nucleation-dependent process, which is exacerbated by some disease-causing mutants and C-terminal truncations (4). Importantly, animals overexpressing ␣Syn, especially diseaseassociated mutants, develop movement disorder accompanying ␣Syn aggregation, recapitulating some features of PD and several other neurodegenerative diseases related to ␣Syn aggregation including dementia with Lewy bodies and multiple system atrophy (13).
The impaired mitochondrial function and ubiquitin-proteasome activity are the two leading factors contributing to PD pathogenesis at least in part through accelerating ␣Syn aggregation (14). The ubiquitin-proteasome pathway is responsible for the degradation of the majority of intracellular proteins, and it plays essential roles in maintaining almost every aspect of cellular activities including gene transcription, protein translation, cell development, signal transduction, and protein quality control (15). The 26 S proteasome is a 2.5-MDa complex consisting of the 20 S proteasome and the 19 S regulatory complex (called PA700 in mammals). The eukaryotic 20 S proteasome has a heptameric, four-ring stacked structure arranged as ␣ 1-7 ␤ 1-7 ␤ 1-7 ␣ 1-7 . Three of the ␤ subunits have distinct peptidase activities: ␤1 has a caspase-like activity, ␤2 has a trypsinlike activity, and ␤5 has a chymotrypsin-like activity. The six catalytic sites are housed inside the ␤ chamber, which is sequestered from intracellular proteins by ␣ chambers with sealed entrances (16). Interestingly the sealed entrance can be opened upon association with proteasomal regulatory complexes PA28, PA200, and PA700 (17). Only PA700, when associated on either or both ends of the 20 S proteasome that forms the 26 S proteasomes, has the ability to mediate the degradation of polyubiquitinated proteins. PA700 has 20 subunits, which possess activities to bind a polyubiquitin chain, bind denatured proteins, unfold protein substrates, catalyze substrate deubiquitination, and hydrolyze ATP (18,19). Coordinated actions of these activities are necessary to mediate the degradation of polyubiquitinated proteins (20). Although both the 20 and 26 S proteasome can degrade unstructured proteins without a polyubiquitin chain modification (21,22), the physiological role of the 20 S proteasome is still under debate.
Several lines of evidence indicate a direct link of the impaired proteasome function to PD pathogenesis. Genetic mutations in PARK genes (PARK 2 encoding Parkin, an E3 ubiquitin-protein ligase, and PARK 5 encoding ubiquitin C-terminal hydrolase L1) encoding two components of the ubiquitin-proteasome pathway are associated with familial PD (23,24). Histologically, immunohistological staining has revealed that Lewy bodies contain a great abundance of ubiquitin, ubiquitinated proteins, and proteasomal subunits (25,26). Biochemical studies have shown the loss of proteasomal subunits and the impaired proteasomal activities in the substantia nigra pars compacta in sporadic PD (27). Furthermore, studies on animal models corroborate the involvement of proteasome dysfunction in the neurodegenerative process where administration of proteasome inhibitors induces dopaminergic neuronal death and the formation of ␣Syn/ubiquitin-containing inclusions in the surviving neurons (28 -30) presumably because ␣Syn itself is a substrate of the proteasome (31,32). Interestingly, overexpressing ␣Syn inhibits proteasomal activities in several mammalian cell lines (33)(34)(35), yeast (36), and a mouse model (37), suggesting that the elevation of ␣Syn concentration impairs proteasomal activity. Consistently, two recent reports showed that ␣Syn or its insoluble aggregated forms directly inhibit proteasome activity. Lindersson et al. (38) reported that ␣Syn filaments and oligomers specifically inhibit the chymotrypsin-like activity of purified 20 S proteasome using a fluorogenic peptide as the substrate. Snyder et al. (34) found that insoluble aggregated ␣Syn inhibits both ubiquitin-dependent and ubiquitin-independent protein degradation using HEK293 cell lysates or rabbit reticulocyte lysates as sources of the proteasome. Also they found that the chymotrypsin-like activity of purified 20 S proteasome is inhibited by both ␣Syn monomer and insoluble ␣Syn aggregates (34).
An emerging concept in the neurodegenerative disease field is that the potential cytotoxic species are soluble oligomeric intermediates, such as protofibrils, but not the insoluble fibrils deposited in diseased brains. Much evidence supports this concept from ␣Syn aggregation related to PD pathogenesis. In vitro studies showed that both earlier onset familial PD-causing mutations, A30P and A53T, promote ␣Syn protofibril formation relative to wild-type protein, whereas the A30P mutation retards fibril formation (39), suggesting that the pathogenic form is ␣Syn protofibrils or other soluble oligomers. Consistently, ␣Syn protofibrils, not fibrils, promote cell death when introduced in cell culture media (40,41). Moreover, transgenic mice expressing human ␣Syn develop nonfibrillar intraneuronal inclusions, lose dopaminergic terminals, and show motor impairments (42). Introducing human ␣Syn gene into the substantia nigra of rats causes the formation of nonfibrillar inclusions and selective loss of dopaminergic neurons (43), supporting the idea that fibril formation is not necessary for neurodegeneration. Pathologically, recent studies demonstrate that the formation of soluble oligomers accompanies Alzheimer disease and PD pathology (40,44), indicating that the well characterized in vitro nucleation-dependent protein aggregation process might occur in vivo as well. Whether the potential cytotoxic ␣Syn protofibrils could directly impair proteasome function, especially the physiologically important 26 S proteasome, is unknown. Here using fluorogenic peptides, unstructured proteins, and a polyubiquitinated protein as substrates, we examined the effects of three soluble ␣Syn species (monomer, dimer, and protofibrils) on the protein degradation activity of highly purified 26 S proteasomes.
Proteasome Purification-PA700 and 26 S proteasomes were purified from bovine red blood cells as described previously (20,47). 4% native PAGE for assaying the in-gel peptidase activity of the 26 S proteasome was performed according to a previous report (20).
Preparation and Separation of ␣Syn Monomer, Dimer, and Protofibrils-␣Syn protofibrils were prepared by a method established by Lansbury and co-workers (48). To increase protofibril production, purified ␣Syn was repeatedly lyophilized and redissolved two to four times. Soluble samples were loaded onto a Superdex 200 gel filtration column on a fast performance liquid chromatography system. The concentration of protofibrils was quantitated by measuring the density of Coomassie-␣Syn Protofibrils Inhibit 26 S Proteasomal Degradation stained SDS-PAGE using bovine serum albumin as a standard and assuming an average molecular mass of 1,500 kDa. The concentrations of ␣Syn monomer and dimer were measured using ⑀ 280 ϭ 5,960 cm Ϫ1 M Ϫ1 and ⑀ 280 ϭ 11,920 cm Ϫ1 M Ϫ1 , respectively.
Circular Dichroism (CD) Spectroscopy-Far-UV CD spectra were collected at 22°C on a JASCO-810 spectropolarimeter using a cuvette with a 0.1-cm optical path length. The data were acquired at 0.2-nm intervals with an average of eight scans. Absorptions from individual buffer were subtracted as backgrounds.
Atomic Force Microscopy-Atomic force microscopy measurements were collected using a Nanoscope IIIa system (Digital Instruments Inc., Santa Barbara, CA) operating in the tapping mode. Tapping mode etched silicon tips with a cantilever length of 125 m and a characteristic frequency of 300 -330 kHz were used for image acquisition. Protein samples were prepared according to the following procedures: 1) dipping the silicon wafers into 10 mM poly(diallydimethylammonium) chloride (PDDA) solution for 10 min, 2) rinsing the PDDA-treated silicon wafers with Milli-Q water for 2 min, 3) dipping the washed silicon wafers into protein samples for 10 min, and 4) rinsing the protein-bound silicon wafers with Milli-Q water for 2 min. The advantages of using PDDA treatment are as follows. 1) PDDA is uniformly adsorbed on the silicon wafer, and the thickness of the PDDA layer is less than 0.3 nm so it does not have an obvious effect on the atomic force microscopy measurement of binding proteins. 2) The positive-charged PDDA layer allows the negative-charged proteins to be stably adsorbed on the silicon matrix due to the electrostatic attraction.
Assay of Proteasomal Peptidase Activities and Deubiquitination Activity-For the proteasomal peptidase assay, Suc-Leu-Leu-Val-Tyr-Amc, Z-Leu-Leu-Arg-Amc, and Z-Leu-Leu-Glu-Amc fluorogenic peptides were used to monitor the chymotrypsin-, trypsin-, and caspase-like activities of the 26 S proteasome, respectively. Appropriate concentrations of ␣Syn monomer, dimer, or protofibrils as indicated in the figure legend were incubated with 10 nM 26 S proteasome in buffer A (20 mM Tris, pH 7.2, 50 mM NaCl, 5 mM MgCl 2 , 1 mM ATP, and 1 mM dithiothreitol) for 5 min. 10-l mixtures were then added to 200 l of 50 M fluorogenic peptide substrates in 20 mM Tris, pH 7.2, and 2 mM ␤-mercaptoethanol. For the proteasomal deubiquitinase assay, 5 nM 26 S proteasome and 250 nM ␣Syn protofibrils in buffer A were preincubated for 5 min before adding 500 nM Ub-Amc. Fluorescence of released Amc was continuously monitored using a Synergy HT plate reader (BioTek) with an excitation/emission filter set at 360/40 and 460/40 nm, respectively. Activities were calculated by the slopes of the linear portions of the fluorescence curves. The results represented an average of three independent experiments.
Assaying Proteasomal Degradation of Non-ubiquitinated Proteins and Polyubiquitinated Protein in Vitro-Lys-48linked tetraubiquitin-conjugated UbcH10 (Ub 4 -UbcH10) was synthesized according to a recent report (20). Purified 26 S proteasomes were preincubated at 37°C for 30 min in buffer A with 1ϫ PA700 to promote the assembly of doubly capped 26 S proteasome. Degradation of ␣Syn monomer, dimer, and protofibrils; p21 cip1 ; securin; or Ub 4 -UbcH10 was performed in buffer A except that 125 mM NaCl was present in the buffer for the degradation of p21 cip1 and securin. Concentrations of the proteasome and substrates are indicated in the figure legends. To determine the effect of ␣Syn monomer, dimer, and protofibrils on the degradation of p21 cip1 , securin, and Ub 4 -UbcH10, 1-, 5-, 25-, or 125-fold molar excess amounts of ␣Syn species were preincubated with the 26 S proteasome for 5 min prior to the supplementation of each substrate. At each designated time point, samples were withdrawn, and reactions were stopped by adding 5ϫ SDS sample buffer. Degradation was evaluated by immunoblotting using appropriate antibodies.
Assaying ␣Syn Aggregation in Cultured Cells-HEK293 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 50 units/ml penicillin, and 250 ng/ml streptomycin in a 5% CO 2 , 95% air atmosphere. At 50 -60% confluence, the cells were transfected with empty pcDNA 3.1(ϩ) vector (Invitrogen) or the vector containing a myc-␣Syn-HA insert using ExGen 500 transfection reagent (Fermentas). 48 h post-transfection, cells in a 10-cm plate were lysed into buffer containing 20 mM Tris, pH 7.4, 150 mM NaCl, 1% Triton X-100, and Complete protease inhibitor mixtures (Roche Applied Science). Lysates were centrifuged at 200 ϫ g for 15 min at 4°C. The pellet was collected as P1. The supernatant (S1) was further centrifuged at 16,000 ϫ g for 30 min at 4°C to further separate the Triton X-100-soluble (S2) and -insoluble (P2) fractions. Both P1 and P2 were washed five times with the lysis buffer and resuspended in SDS sample buffer for immunoblotting using anti-␣Syn antibody.
Monitoring the Effect of ␣Syn Overexpression on Degradation of Endogenous Proteins in Cultured HEK293 Cells-48 h posttransfection of ␣Syn, cells in 35-mm plates were treated with 80 g/ml cycloheximide to inhibit protein synthesis. To determine IB␣ degradation, the cells were treated with 20 ng/ml tumor necrosis factor ␣ to induce its rapid degradation. The cells were harvested at designated time points and lysed with RIPA buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 1% SDS, 10 mM N-ethylmaleimide, 0.2 mM sodium orthovanadate, 100 mM sodium fluoride, 10 mM sodium pyrophosphate, and Complete protease inhibitor mixtures). The degradation of endogenous p21 cip1 , UbcH10, and IB␣ was evaluated by immunoblotting using appropriate antibodies.
Determining ␣Syn Protofibrils and PA700 Interaction-To assay ␣Syn protofibril binding to PA700, 100 g of PA700 was preincubated with 200 g of ␣Syn protofibrils at room temperature for 30 min and then centrifuged at 16,000 ϫ g for 5 min to pellet any insoluble aggregates prior to loading into a Superose 6 (16/60) gel filtration column equilibrated in 20 mM Tris, pH 7.2, 50 mM NaCl, and 2 mM ␤-mercaptoethanol. Protein separation was conducted by a fast performance liquid chromatography system at a flow rate of 0.75 ml/min and fraction size of 800 l. 20 l of each fraction was applied to SDS-PAGE followed by immunoblotting assays.
Size Exclusion Spin Column Assay-Micro Bio-Spin P-30 chromatography columns with an exclusion limit of 40 kDa (Bio-Rad) or homemade Sephadex G-75 spin columns with an exclusion limit of 80 kDa (matrix from Sigma and empty columns from Bio-Rad) were used to determine the interaction between ␣Syn species and the proteins of interest. Briefly, different molar ratios of ␣Syn monomer, dimer, or protofibrils were incubated with target proteins in 20 mM Tris, pH 7.2, and 1 mM dithiothreitol at room temperature for 10 min. 60 l of mixtures was then loaded into spin columns, and the centrifugation steps were followed according to the manufacturer's instructions for Bio-Spin P-30 columns. The flow-throughs were collected and mixed with 5ϫ SDS sample buffer followed by immunoblotting using appropriate antibodies. Comparable results were obtained in determining Ub 4 binding to ␣Syn protofibrils when using either Bio-Spin P-30 or homemade Sephadex G-75 columns.
8-Anilino-1-naphthalenesulfonic Acid Ammonium Salt (ANS) Fluorescence-5 M ANS was mixed with 20 M ␣Syn monomer, 10 M ␣Syn dimer, or 480 nM ␣Syn protofibrils in 20 mM Tris buffer, pH 7.2. Fluorescence measurements were recorded on a FluoroMax-3 spectrofluorometer (Horiba Jobin Yvon) using a cuvette with a 0.5-cm excitation optical path. Samples were excited at 350 nm, and emission spectra were recorded between 400 and 600 nm with an increment of 1 nm and an integration time of 1 s. Slit widths were 1 and 4 nm for excitation and emission, respectively.

RESULTS
Purified ␣Syn Protofibrils Are SDS-resistant, ␤-Strand-rich Soluble Oligomers-We prepared ␣Syn dimer and soluble high molecular weight oligomers (called protofibrils) by dissolving lyophilized ␣Syn at 5-7 mg/ml. Formed subpopulated ␣Syn protofibrils and dimer were separated from monomer by a Superdex 200 gel filtration column in which protofibrils eluted in the void volume (Fig. 1A). Atomic force microscopy images showed that ␣Syn protofibrils were predominantly spherical particles with an average size of approximately 2-4 nm, whereas the morphologies of ␣Syn monomer and dimer were not visible (Fig. 1B). Notably, purified ␣Syn protofibrils were found to be partially SDS-stable as evidenced by the fact that significant amounts of high molecular weight species remained in the stacking gel, but they were urea-soluble (Fig. 1C). In contrast, most purified ␣Syn dimer was dissociated into monomer in SDS, whereas a small amount remained stable even in the presence of urea. Western blots using an anti-␣Syn-specific antibody confirmed that it represented urea-resistant ␣Syn dimer, not a contaminant protein (Fig. 1C). The urea-resistant ␣Syn dimer may be formed through dityrosine cross-links during the heat preparation of recombinant ␣Syn (49). Interestingly, urea-resistant ␣Syn dimer was absent in ␣Syn protofibrils (Fig. 1C), indicating that it is not a key nucleator in the formation of protofibrils at least by the current method. Next we determined the secondary structures of purified ␣Syn monomer, dimer, and protofibrils using CD spectroscopy. CD spectra of both ␣Syn monomer and dimer had a minimal absorption at 198 nm indicative of a predominantly random coil conformation (Fig. 1D). In contrast, ␣Syn protofibrils adopted a conformation of mixed ␤-strand and ␣-helix structures as evidenced by the CD spectrum with two negative absorption peaks located at 217 and 208 nm (Fig. 1D). Together our results show that purified ␣Syn protofibrils are SDS-resistant but urea-soluble large oligomers with folded structures.
␣Syn Monomer, Dimer, and Protofibrils Are Not Efficiently Degraded by Purified 26 S Proteasomes-Previous studies from different groups have demonstrated that the 20 S proteasome is capable of degrading unstructured ␣Syn (22,32). In contrast to the sole ability to degrade unstructured proteins by the 20 S proteasome, the 26 S proteasome is the physiologically important form that has the capability to degrade polyubiquitinated proteins. Here we examined the ability of purified 26 S proteasomes to degrade ␣Syn monomer, dimer, and protofibrils. The intact 26 S proteasomes, which displayed characteristic subunit patterns on SDS-PAGE ( Fig. 2A), were purified from bovine red blood cells according to a recently published method (20). The purified 26 S proteasomes consisted of two bands on a 4% native polyacrylamide gel corresponding to the 26 S proteasomes capped with one or two copies of PA700 (Fig. 2B, upper panel). Both forms were catalytically active as demonstrated by an ingel overlay assay using a fluorogenic peptide, Suc-Leu-Leu-Val-Tyr-Amc, to detect their chymotrypsin-like activity (Fig. 2B, ␣Syn Protofibrils Inhibit 26 S Proteasomal Degradation JULY 18, 2008 • VOLUME 283 • NUMBER 29 lower panel). Despite that the 26 S proteasome was highly active to hydrolyze the short fluorogenic peptide, purified 26 S proteasomes could not efficiently degrade ␣Syn monomer, dimer, or ␣Syn protofibrils (Fig. 2C). Because purified 26 S proteasomes efficiently degraded two other in vitro unstructured proteins, p21 cip1 and securin (Fig. 2, D and E), and a polyubiquitinated protein (see Fig. 5B), the degradation resistance for the three ␣Syn species in Fig. 2C is not due to the inactivity of the 26 S proteasome per se. Accordingly, we concluded that ␣Syn monomer, dimer, and protofibrils are not efficiently degraded by purified intact 26 proteasomes.
␣Syn Monomer, Dimer, and Protofibrils Do Not Inhibit the Peptidase Activities of the 26 S Proteasome-Fluorogenic peptides have been developed to monitor individual hydrolytic activities of the proteasome. Next we examined whether the peptidase activities of the 26 S proteasome are directly affected by ␣Syn monomer, dimer, or protofibrils. In these assays, appropriate concentrations of tested ␣Syn species were preincubated with purified 26 S proteasome prior to being mixed with short peptide substrates. We found that none of the three ␣Syn species impaired the activity of 26 S proteasome to degrade these short peptide substrates (Fig. 3, A-C). Interestingly, concerted slight stimulations of all three peptidase activities were observed when high concentrations of ␣Syn monomer or dimer or even lower concentrations of ␣Syn protofibrils co-existed. These results demonstrate that ␣Syn monomer, dimer, and protofibrils do not block the translocation of short peptides into the degradation chamber of the 26 S proteasome nor do they directly compete with short peptides for any of the three catalytic activities.
␣Syn Protofibrils Inhibit Ubiquitin-independent Proteasomal Degradation of Unstructured Proteins-Recently we have demonstrated that the 26 S proteasome degrades some unstructured proteins in an ATP hydrolysis-independent manner (20). To examine whether ␣Syn monomer, dimer, or protofibrils affect the 26 S proteasomal degradation of unstructured proteins, we chose p21 cip1 and securin as two model substrates. Neither ␣Syn monomer nor dimer had an obvious effect on the 26 S proteasomal degradation of these two unstructured proteins even when they were present at a level of 125-fold molar excess over the 26 S proteasome (Fig. 4A). Strikingly ␣Syn protofibrils markedly inhibited the 26 S proteasome-mediated degradation of both p21 cip1 and securin at 25-and 5-fold molar excesses to the 26 S proteasome, respectively (Fig. 4A). These results indicate that ubiquitin-independent 26 S proteasomal degradation of unstructured proteins is impaired selectively by ␣Syn protofibrils but not by ␣Syn monomer or dimer.
Next we hypothesized that overexpression of ␣Syn in cells could impair the degradation of endogenous p21 cip1 , which has been found previously to be degraded by the proteasome with no need of polyubiquitination in cells (50 -52). It has been reported that overexpressing ␣Syn in cells causes ␣Syn aggregation (53). Consistently, we found that overexpression of ␣Syn in HEK293 cells led to accumulation of Triton X-100-insoluble ␣Syn aggregates that were pelleted from both 200 ϫ g (P1) and 16,000 ϫ g (P2) centrifugations, but these Triton X-100-insoluble aggregates were SDS-soluble as revealed by the fact that they were disrupted into the monomeric form in SDS-PAGE (Fig. 4B). Next we monitored endogenous p21 cip1 degradation by a cycloheximide-directed chase experiment. In mock-transfected cells, p21 cip1 was degraded with a half-life of about 1.94 h consistent with previous reports (54,55).
Remarkably degradation of p21 cip1 in ␣Syn-overexpressing HEK293  cells was inhibited (Fig. 4C). This result suggests that overexpression of ␣Syn impairs ubiquitin-independent degradation in cells.
␣Syn Protofibrils Inhibit Proteasomal Degradation of Polyubiquitinated Folded Proteins-In contrast to unstructured proteins, degradation of folded proteins requires polyubiquitination and coordinated actions including substrate engagement, substrate unfolding, substrate deubiquitination, substrate translocation, and ATP hydrolysis (20). To evaluate whether ␣Syn monomer, dimer, or protofibrils have an effect on the 26 S proteasome-mediated degradation of polyubiquitinated proteins, we chose UbcH10 as a model substrate. UbcH10 is a bona fide proteasomal substrate because the ubiquitination-dependent degradation of UbcH10 is critical for triggering the G 1 -S phase transition in cell cycle progression (56). We conjugated Lys-48-linked Ub 4 chain to UbcH10 using immunoprecipitated Xenopus anaphase-promoting complex/cyclosome as the E3 Ub ligase (57). After the polyubiquitination reaction, about 70% of UbcH10 was conjugated with a Ub 4 chain, whereas some UbcH10 remained non-ubiquitinated (Fig.  5A). Ub 4 -UbcH10 was less sensitive than UbcH10 toward an anti-UbcH10 antibody as revealed by immunoblotting (Fig. 5B) presumably because the epitope of UbcH10 was partially masked by the conjugated Ub 4 chain. We have showed previously that UbcH10 itself is not a substrate of the 26 S proteasome, whereas Ub 4 -UbcH10 is rapidly degraded (20). Degradation of Ub 4 -UbcH10 was judged by comparing the reactions without and with a proteasome inhibitor (Fig. 5B, lane  2 versus lane 3). Non-ubiquitinated UbcH10 accumulated when the 26 S proteasome was inhibited by adding epoxomicin because deubiquitination was still allowed (lane 3). In contrast, no obvious accumulation of UbcH10 was present in the reaction without epoxomicin (lane 2), indicating degradation. Notably, some Ub 4 -UbcH10 still remained when the proteasome was inhibited by epoxomicin (Fig. 5B, lane 3); this is consistent with our earlier report that proteasome inhibition can partially block deubiquitination (20).
We next examined the effects of ␣Syn monomer, dimer, or protofibrils on the 26 S proteasomal degradation of Ub 4 -UbcH10. Most strikingly, Ub 4 -UbcH10 degradation was completely abolished when ␣Syn protofibrils were added at 5-fold or higher molar excess relative to the 26 S proteasome (Fig. 5B, lanes 10 -12). In contrast, ␣Syn monomer and dimer did not obviously inhibit Ub 4 -UbcH10 degradation when present at 5-fold molar excess to the 26 S proteasome (Fig. 5B, lanes 4 and 7). However, partial accumulation of non-ubiquitinated UbcH10 was observed when ␣Syn monomer or dimer was added at 25-fold or higher molar excess to the 26 S proteasome (Fig. 5B, lanes 5, 6, 8, and 9), and slight accumulation of Ub 4 -UbcH10 was observed when the concentration of dimer was 125-fold more than the 26 S proteasome (Fig.  5B, lane 9). Obviously, when compared with the inhibited degradation in the presence of epoxomicin (Fig. 5B, lane 3), 125fold more ␣Syn monomer or dimer was not able to completely block the proteasomal degradation of Ub 4 -UbcH10 (Fig. 5B,  lanes 6 and 9). These results demonstrate that ␣Syn monomer and dimer have a mild inhibitory effect on 26 S proteasomal degradation when they are at 25-fold or higher molar excess to the 26 S proteasome, whereas ␣Syn protofibrils exhibit a dramatic inhibitory effect on proteasomal degradation of polyu- ␣Syn Protofibrils Inhibit 26 S Proteasomal Degradation JULY 18, 2008 • VOLUME 283 • NUMBER 29 biquitinated proteins when present at only 5-fold molar excess to the 26 S proteasome. Intriguingly, inhibition caused by ␣Syn protofibrils occurred at the deubiquitination step (Fig. 5B, lanes  10 -12).
Protein deubiquitination is tightly coupled with degradation, and inhibition of the proteasome-residing deubiquitination enzymes blocks the degradation of polyubiquitinated proteins (20,58,59). We next asked whether the deubiquitination activity of the 26 S proteasome per se is inhibited by ␣Syn protofibrils; this could directly contribute to the observed inhibitory effect of ␣Syn protofibrils on degradation of Ub 4 -UbcH10 as shown in Fig. 5B. Surprisingly, we found that purified 26 S proteasomes deubiquitinated Ub-Amc equally well in the absence and in the presence of 50-fold excess amounts of ␣Syn protofibrils (Fig. 5C). These results indicate that the overall deubiquitination activity of the 26 S proteasome is not inhibited by ␣Syn protofibrils. Certainly we cannot exclude the possibility that deubiquitination of polyubiquitinated proteins and Ub-Amc might be catalyzed differently by the 26 S proteasome.
Next we examined whether ubiquitin-dependent proteasomal degradation is affected by overexpressing ␣Syn in cells. To test this, we chose UbcH10 and IB␣ as substrates. In response to tumor necrosis factor ␣ stimulation, IB␣ is phosphorylated and then ubiquitinated for rapid degradation (60). We found that degradation of both UbcH10 and IB␣ was significantly inhibited in ␣Syn-overexpressing HEK293 cells as compared with mock-transfected cells (Fig. 5, D and E). Thus, ␣Syn overexpression impairs ubiquitin-dependent proteasomal degradation in vivo. ␣Syn Protofibrils Inhibit 26 S Proteasomal Degradation ␣Syn Protofibrils Directly Bind the 26 S Proteasome-␣Syn was reported to bind the S6Ј (also called Rpt5) ATPase subunit of the 26 S proteasome (34,61); therefore, it is interesting to examine whether ␣Syn monomer, dimer, or protofibrils bind the intact 26 S proteasome complex. To do this, we used gel filtration to monitor whether PA700 co-migrates with ␣Syn monomer, dimer, or protofibrils. We chose to use PA700 instead of the 26 S proteasome because the 2.5-MDa 26 S proteasome had a similar retention time with ␣Syn protofibrils, but the 700-kDa PA700 and ␣Syn protofibrils had different retention times on a 120-ml Superpose 6 column. As shown in Fig. 6, PA700 co-migrated with ␣Syn protofibrils and eluted about 10 fractions earlier than PA700 alone, indicating a direct interaction of ␣Syn protofibrils with the 26 S proteasome. Although ␣Syn was found to directly interact with the S6Ј ATPase sub-unit of the 26 S proteasome (34, 61), we did not detect comigration of ␣Syn monomer or dimer with the PA700 complex by using a similar gel filtration assay even when ␣Syn monomer or dimer was present at 10-fold molar excess to PA700 (data not shown). Presumably it reflects the difference between the PA700 complex and the individual S6Ј subunit. Nevertheless these results suggest that ␣Syn protofibrils have a higher binding affinity to PA700 or the 26 S proteasome than does ␣Syn monomer or dimer.
␣Syn Protofibrils Directly Bind Some Proteasomal Substrates-During ␣Syn fibrillization, formed protofibrils associate with ␣Syn monomer to develop fibrils. Therefore, it would be of interest to examine whether ␣Syn protofibrils bind other proteins as well. To do this, we used a size exclusion spin column assay to examine whether ␣Syn protofibrils bind proteasomal substrates. ␣Syn monomer (14 kDa) has an apparent molecular mass of about 50 kDa on a gel filtration column because of its unfolded structure. We detected that ␣Syn monomer was excluded by the matrix of the P-30 spin column, which has an exclusion limit of 40 kDa (data not shown). Similarly ␣Syn dimer and protofibrils were also excluded by the P-30 spin column (data not shown). In contrast, p21 cip1 or securin was trapped inside the column as evidenced by the fact that they were not detected in the flow-throughs after centrifugation ( Fig. 7A and data not shown). However, p21 cip1 was co-eluted with ␣Syn protofibrils in the flow-through when incubated with an equal molar ratio of ␣Syn protofibrils but not with 40-fold molar excess of ␣Syn monomer or 20-fold molar excess of ␣Syn dimer (Fig. 7A), demonstrating that p21 cip1 binds ␣Syn protofibrils. Using this size exclusion spin column assay and a gel filtration assay, we also detected ␣Syn protofibril binding to securin (supplemental Fig. S1A and data not shown). Lewy bodies contain significant amounts of ubiquitin; therefore, we next examined whether a Lys-48linked Ub 4 chain binds ␣Syn protofibrils. Using the same size exclusion spin column assay, we found that ␣Syn protofibrils, but not ␣Syn monomer or dimer even at a concentration of at least 20-fold higher than ␣Syn protofibrils, bind Ub 4 (Fig. 7A). Surprisingly, monoubiquitin did not bind ␣Syn protofibrils even when the ␣Syn protofibril concentration was increased to 8ϫ molar excess where binding of Ub 4 to ␣Syn protofibrils was saturated (Fig. 7B). This result indicates that ␣Syn protofibrils prefer to bind polyUb chains. Indeed we found that Ub 8 binds more efficiently than Ub 4 to ␣Syn protofibrils as judged by the fact that nearly 40% of Ub 8 FIGURE 6. ␣Syn protofibrils interact with PA700, the regulatory complex of the 26 S proteasome. 100 g of PA700 was preincubated with 200 g of ␣Syn protofibrils at room temperature for 30 min before injection into a Superose 6 (16/60) gel filtration column. Fractions were immunoblotted with anti-p31, a subunit of PA700, or anti-␣Syn antibody. PA700 or protofibrils alone were run as a control under the same condition.   7C) whereas only 5% of Ub 4 bound to ␣Syn protofibrils (Fig. 7B) when Ub 4 or Ub 8 was incubated at a 1:1 molar ratio with ␣Syn protofibrils. Interaction between polyUb chains and ␣Syn protofibrils was further confirmed by gel filtration (supplemental Fig. S1B). Interestingly, binding of Ub 8 to ␣Syn protofibrils was markedly increased when 0.5 M (NH 4 ) 2 SO 4 was added (Fig. 7C), suggesting that binding of polyUb chains to ␣Syn protofibrils is mediated by a hydrophobic-hydrophobic interaction. Indeed as determined by an ANS binding assay, ␣Syn protofibrils gained an exposed hydrophobic surface(s) when compared with ␣Syn monomer and dimer (Fig. 7D). Consistent with its increased capacity to bind Ub 8 in the presence of 0.5 M (NH 4 ) 2 SO 4 , ANS fluorescence of ␣Syn protofibrils increased 4-fold when 0.5 M (NH 4 ) 2 SO 4 was added in the mixture (Fig. 7D). Taken together, these results indicate that ␣Syn protofibrils bind both unstructured and polyubiquitinated substrates of the 26 S proteasome likely through a hydrophobichydrophobic interaction.

Inhibition of Proteasomal Degradation by ␣Syn Protofibrils-
The impaired proteasomal function is believed to be one of the main reasons for PD pathogenesis (14). In this report, we examined the effects of three soluble ␣Syn species (␣Syn monomer, dimer, and protofibrils) on the protein degradation activity of the 26 S proteasome. ␣Syn has been reported to be a substrate of both the proteasome (31,32) and the lysosome (62,63). Interestingly, we found that purified intact 26 S proteasomes degraded ␣Syn monomer and dimer at much slower rates as compared with unstructured p21 cip1 and securin (Fig. 2, C and  E). Mechanistically it is unclear why unstructured proteins have different access to the catalytic sites of the 26 S proteasome. The 26 S proteasome efficiently degrades some unstructured proteins without the requirement of ATP hydrolysis (20), indicating that translocation of these unstructured proteins (with no need for further unfolding) into the degradation chamber does not consume energy from ATP hydrolysis. However, the degradation distinctions between ␣Syn and p21 cip1 or securin suggest that the 26 S proteasome possesses a substrate recognition step for screening unstructured protein substrates.
The 26 S proteasome has three distinct peptidase activities located in the central ␤ chamber of the 20 S proteasome. We did not detect any inhibitory effect of ␣Syn monomer, dimer, and protofibrils on the peptidase activities of the 26 S proteasome using short fluorogenic peptides as substrates (Fig. 3), demonstrating that they do not block the translocation of small peptides nor directly compete for the catalytic sites. Because ␣Syn is a substrate of the 20 S proteasome (22,32), the inhibitory effect of ␣Syn on the 20 S proteasome found in two earlier studies (34, 38) using short fluorogenic peptides as substrates could simply be due to their competition for the catalytic sites. In striking contrast, ␣Syn protofibrils, but not the higher concentrations of monomer and dimer we examined, inhibited 26 S proteasomal degradation of p21 cip1 and securin (Fig. 4A). This inhibition cannot be due to a direct competition toward the catalytic sites because ␣Syn protofibrils with an average size of 2-4 nm cannot pass through the narrow substrate translocation channel of the proteasome (see the discussion below). Fur-thermore we found that ␣Syn protofibrils had a marked inhibitory effect on the degradation of Ub 4 -UbcH10 that occured at the deubiquitination level (Fig. 5B). In the presence of high concentrations of ␣Syn monomer and dimer (at 25-fold or higher molar excess to the proteasome), degradation of Ub 4 -UbcH10 by the 26 S proteasome was mildly inhibited as observed by the accumulation of UbcH10 (Fig. 5B). This result suggests that the inhibitory effect of high concentrations of ␣Syn monomer and dimer comes from impaired substrate translocation and/or unfolding. Consistent with our in vitro studies, we found that both ubiquitin-independent and ubiquitin-dependent proteasomal degradation were impaired in cells when ␣Syn was overexpressed (Figs. 4C and 5, D and E). Although the formation of ␣Syn aggregates was detected in ␣Syn-overexpressing cells (Fig. 4B), it is unknown whether the inhibited proteasomal degradation in ␣Syn-overexpressing cells results directly from the aggregated ␣Syn. A previous report showed that ␣Syn overexpression in PC12 cells has no effect on proteasome-mediated basal turnover of IB␣ (64). This discrepancy is currently unclear, but we monitored the signal-induced rapid IB␣ degradation.
Interestingly, ␣Syn protofibrils, but not monomer and dimer, had strong binding affinity to both the 26 S proteasome and proteasomal substrates (unstructured proteins and polyubiquitinated proteins) (Figs. 6 and 7). These results might provide mechanistic explanations for ␣Syn protofibrils to inhibit proteasomal degradation: by binding on PA700, presumably on the ATPase ring because it has a chaperone-like activity to bind denatured proteins with exposed hydrophobic surface(s) like ␣Syn protofibrils (Fig. 7), the large oligomeric ␣Syn protofibrils could block the narrow substrate translocation channel of the proteasome (the narrowest part is 13 Å), thus inhibiting degradation by impeding entry of protein substrates (but not short peptides) into the proteolytic chamber (Fig. 8). This scenario is supported by the evidence that the degradation of p21 cip1 was still completely inhibited when increasing the concentration of p21 cip1 while keeping ␣Syn protofibrils at 25-fold molar excess to the proteasome (data not shown). On the other hand, by ␣Syn Protofibrils Inhibit 26 S Proteasomal Degradation binding proteasomal substrates, ␣Syn protofibrils could sequester substrates from the proteasome, therefore blocking degradation (Fig. 8). This mechanism could, at least partially, account for the observed inhibitory effect for ␣Syn protofibrils on the degradation of Ub 4 -UbcH10 at the deubiquitination level. Because the deubiquitination activity of the 26 S proteasome per se was not impaired by ␣Syn protofibrils as revealed by the Ub-Amc deubiquitination assay (Fig. 5C), the inhibited Ub 4 -UbcH10 deubiquitination, as well as degradation, might come from substrate sequestration by ␣Syn protofibrils. Certainly these two possibilities for ␣Syn protofibrils to inhibit proteasomal degradation are not mutually excluded, but a certain model is favored depending on the affinity between protofibrils binding to the 26 S proteasome and binding to the proteasomal substrates.
A Possible Generality for Protofibrils in the Pathogenesis of Neurodegenerative Diseases by Inhibiting Proteasomal Activity-One of the potential cytotoxic mechanisms for protofibrils is that they form a porelike structure that enables them to penetrate the cell membrane, thus causing cell leaking (48,65,66). In the current study, our data suggest a novel potential mechanism for ␣Syn protofibrils to exert their cytotoxic effect: through interacting with other intracellular proteins ␣Syn protofibrils might disrupt important cellular pathways including the ubiquitin-proteasome pathway. Even worse, the formation of ␣Syn protofibrils and the impairment of proteasome activity might be mutually exacerbated processes. Loss of proteasomal activity could cause the accumulation of ␣Syn in cells (31,32,67), which may eventually result in ␣Syn aggregation and the production of protofibrils. Conversely, accumulating ␣Syn protofibrils impairs the proteasome function by directly binding to the proteasome or sequestering proteasomal substrates as found in this study. Interestingly, oligomeric forms of other disease-related proteins have been found to impair proteasomal activities. For example, soluble oligomeric amyloid ␤ peptide (A␤), but not monomer, inhibits proteasomal activity in vitro (68). In 3xTg-AD mice brains, the impaired proteasomal function occurs in parallel with the appearance of soluble amyloid ␤ peptide oligomers, and the impairment is relieved as soluble amyloid ␤ peptide oligomers are converted into insoluble aggregates (68). Inhibition of proteasomal activities by disease-associated prion protein (PrP sc ) is abolished by anti-oligomer-specific antibody consistent with the notion that oligomeric species mediate proteasome dysfunction (69). Therefore, we propose that a vicious cycle between the impaired proteasomal activity and the promoted protein aggregation (production of high molecular weight oligomers) might be a common feature involved in the pathogenesis of neurodegenerative diseases. Actually soluble oligomers formed from different amyloidogenic proteins, including ␣Syn, amyloid ␤ peptide, prion protein, and polyglutamine, share a common structure regardless of the protein sequences (40). Therapeutically small molecules and antibodies that inhibit oligomer formation or eliminate formed oligomers might be useful therapeutic reagents for these neurodegenerative diseases.