Cooperative inhibition of SNARE-mediated vesicle fusion by α-synuclein monomers and oligomers

The primary hallmark of Parkinson's disease (PD) is the generation of Lewy bodies of which major component is α-synuclein (α-Syn). Because of increasing evidence of the fundamental roles of α-Syn oligomers in disease progression, α-Syn oligomers have become potential targets for therapeutic interventions for PD. One of the potential toxicities of α-Syn oligomers is their inhibition of SNARE-mediated vesicle fusion by specifically interacting with vesicle-SNARE protein synaptobrevin-2 (Syb2), which hampers dopamine release. Here, we show that α-Syn monomers and oligomers cooperatively inhibit neuronal SNARE-mediated vesicle fusion. α-Syn monomers at submicromolar concentrations increase the fusion inhibition by α-Syn oligomers. This cooperative pathological effect stems from the synergically enhanced vesicle clustering. Based on this cooperative inhibition mechanism, we reverse the fusion inhibitory effect of α-Syn oligomers using small peptide fragments. The small peptide fragments, derivatives of α-Syn, block the binding of α-Syn oligomers to Syb2 and dramatically reverse the toxicity of α-Syn oligomers in vesicle fusion. Our findings demonstrate a new strategy for therapeutic intervention in PD and related diseases based on this specific interaction of α-Syn.

. With the addition of α-Syn oligomers (10 nM), a significant reduction in fusion efficiency was observed (the black line). When α-Syn monomer was added to the T-V mixture together with oligomers, the FRET signal was lower than that of the α-Syn oligomer-treated vesicle fusion. Lipid mixing was confirmed to be driven by SNARE complex formation by the addition of the soluble domain of Syb2, which blocked lipid mixing (the light purple line). (D) Relative percentages of lipid mixing at 1800s from the FRET data in (C) (± SEM, N = 3; three measurements were performed in different days). **P < 0.01 by ANOVA test. The graphs shown were produced using Originlab 8.5 software (OriginLab Corporation, Northhampton, MA, USA) (https:// www. origi nlab. com/ index. aspx? go = PRODUCTS/Origin) and edited by Adobe Illustrator CS6 (Adobe). www.nature.com/scientificreports/ in the presynaptic terminals, investigating the inhibitory effect of α-Syn oligomers on vesicle fusion in the presence of α-Syn monomers is a prerequisite for improving our understanding of the toxicity of α-Syn oligomers.
Here, we show that the α-Syn monomer and oligomers cooperatively inhibit SNARE-mediated vesicle fusion. This cooperative effect results from the clustering of Syb2-carrying vesicles. Thus, the fusion inhibitory effect of α-Syn oligomers should be more severe in the presynaptic region where the α-Syn monomer is enriched. Based on this cooperative inhibition mechanism, we designed a strategy to reverse the fusion inhibitory effect of α-Syn oligomers. We reversed the fusion inhibition of α-Syn oligomers by modulating the interactions between α-Syn oligomers and Syb2 using small peptide fragments of α-Syn. The small peptide fragments compete with α-Syn oligomers for binding to Syb2 and dramatically restore SNARE-mediated vesicle fusion in the presence of α-Syn oligomers. Our findings suggest a new strategy for therapeutic intervention in PD and related diseases based on this specific protein-protein interaction of α-Syn.

Results
The α-Syn monomer enhances the toxicity of dopamine-induced α-Syn oligomers in inhibiting SNARE-mediated vesicle fusion. When α-Syn oligomers and monomers are added to the reaction mixture for vesicle fusion, three scenarios are possible ( Supplementary Fig. S1A): (1) α-Syn monomer binds to Syb2 more weakly than α-Syn oligomers do. Thus, the α-Syn monomer has no effect on the fusion inhibitory effect of α-Syn oligomers. (2) Both the α-Syn monomer and oligomers interact with Syb2, competing with each other for Syb2 binding. As a result, the addition of the monomer attenuates the fusion inhibitory effect of oligomers.
Then, we applied α-Syn oligomers to the T-V reaction mixture. When we used 10 nM α-Syn oligomers, the extent of lipid mixing was reduced by approximately 50% compared with that of the control (Fig. 1C, the black line), which is consistent with the results of our previous work 15 . Next, we incubated V with α-Syn monomer for 10 min, and then α-Syn oligomers and T were added to the premixture (Fig. 1B). To our surprise, more severe inhibition of vesicle fusion between T and V was observed as the concentration of α-Syn monomer increased (Fig. 1C). When 10 nM oligomers were added together with 850 nM monomer, the fusion efficiency was reduced to 25%; the 25% fusion efficiency was almost the same fusion efficiency when 30 nM oligomers were used ( Fig. 1D) 15 . As a control, α-Syn monomer at submicromolar concentrations (200-850 nM) was shown to have a negligible effect on lipid mixing ( Supplementary Fig. S3). This result suggests that even a small amount of α-Syn oligomers would cause severe toxicity by inhibiting synaptic vesicle fusion in presynaptic regions where the α-Syn monomer is abundant. α-Syn monomers and oligomers cooperatively induce the clustering of synaptic vesicle mimics. Next, we investigated how the α-Syn monomer enhances the fusion inhibitory effect of dopamine-induced α-Syn oligomers. We first tested whether the incubation of the α-Syn monomer with oligomers increases the amount of α-Syn oligomers. When we incubated the α-Syn monomer with preformed oligomers for 12 h without dopamine, no change in the amounts of oligomers and monomers was observed ( Supplementary Fig. S4). This is consistent with previous work showing that dopamine-modified α-Syn oligomers are not seeding competent 18 .
Then, we tested the effect on vesicle clustering. We have proposed that α-Syn oligomers inhibit vesicle fusion by clustering synaptic vesicle mimics in vitro, thereby limiting the number of vesicles available for docking 15 . A few micromolar α-Syn monomer is also known to induce vesicle clustering 27,29,30,38 . Thus, it is possible that the α-Syn monomer by itself cannot induce vesicle clustering at the low concentration (less than 1 μM) that we applied in this work but assists α-Syn oligomers in clustering vesicles. We obtained images of v-vesicles treated with α-Syn oligomers in the presence and absence of monomers using cryo-electron microscopy (cryo-EM) (Fig. 2). The molar ratio of lipid: α-Syn oligomers: α-Syn monomers was 200:1:20. When more than ten vesicles were attached together, we assigned them as a vesicle clustering. In the absence of α-Syn, v-vesicles appeared without clustering ( Fig. 2A). When we added the α-Syn monomer, no clustering of v-vesicles was observed (Fig. 2B). However, in the presence of α-Syn oligomers, v-vesicles formed compact clusters with irregular shapes (Fig. 2C). Cryo-EM could not resolve α-Syn oligomers because of its resolution limit. Since α-Syn oligomers are known to inhibit vesicle fusion through binding to the N-terminal domain of Syb2 15 , we tested if their specific interactions with Syb2 are responsible for clustering the vesicles. We reconstituted v-vesicles with Syb2 lacking the N-terminal region (nt-Syb2, amino acids 29-116) and incubated them with α-Syn oligomers. No clustering of v-vesicles was observed by cryo-EM ( Supplementary Fig. S5). In addition, protein-free vesicles were not  Fig. S5). These results show that α-Syn oligomers cluster vesicles using their specific interactions with Syb2. When v-vesicles were incubated with both α-Syn monomer and oligomers, they formed larger clusters (Fig. 2D) than those shown in Fig. 2C. While 22.7 ± 1.2 vesicles consisted of a cluster with α-Syn oligomers, 123.5 ± 11.5 vesicles clustered after the addition of α-Syn monomers together with oligomers. Interestingly, when the α-Syn monomer was added with oligomers to v-vesicles, locally flat and thickened adhesion regions between vesicles were often visualized (Fig. 2D, enlarged image). Although it is difficult to quantify the fraction of the flattened region, this structure was not observed in images of vesicles treated with α-Syn oligomers only (Fig. 2C, enlarged image). The addition of α-Syn monomer with α-Syn oligomers seemed to induce tighter adhesion between the vesicles. These results suggest that the combination of the α-Syn monomer and oligomers cooperatively clusters Syb2-carrying vesicles and that the α-Syn monomer may act as an adhesive between the vesicles.

Membrane-bound α-Syn monomers are required for cooperative fusion inhibition. Next, we
investigated the mechanism whereby α-Syn monomers and oligomers cooperatively cluster synaptic vesicle mimics and inhibit SNARE-mediated vesicle fusion at the molecular level. It is well known that the α-Syn monomer binds to negatively charged phospholipids 25,30,39 . Thus, we examined whether the interactions between the α-Syn monomer and the negatively charged membrane play a role in increasing the fusion inhibitory effect of α-Syn oligomers. We prepared the T44P/A89P mutant of α-Syn, which retains the ability to bind Syb2 but has a reduced lipid-binding affinity 39 . We measured the effect of this mutant on lipid mixing using the in vitro lipid-mixing assay (Fig. 3A). When we added both monomeric T44P/A89P mutant and α-Syn oligomers, the cooperative fusion inhibitory effect of α-Syn oligomers and wild-type α-Syn monomer was not observed. Sur- www.nature.com/scientificreports/ prisingly, in contrast to the case of the wild-type α-Syn monomer, vesicle fusion became more active as the concentration of the T44P/A89P mutant increased (Fig. 3A). As a control, when 850 nM T44P/A89P was added to the fusion mixture in the absence of α-Syn oligomers, the extent of lipid mixing was nearly invariant, but it slightly decreased at 2 μM T44P/A89P ( Supplementary Fig. S5). This result indicates that the binding of the α-Syn monomer to the lipid membrane is required for the cooperative inhibition of vesicle fusion by the α-Syn monomer and oligomers.
To confirm the role of membrane binding by the α-Syn monomer in cooperative fusion inhibition, we removed negatively charged 1,2-dioleoyl-sn-glycero3-(phospho-l-serine) (PS) from v-vesicles (Fig. 3B). We used the wild-type α-Syn monomer in this measurement. As shown in Fig. 3B, the cooperative fusion inhibitory effect disappeared again. The fusion inhibition by α-Syn oligomers was attenuated by the addition of the α-Syn . v-Vesicles (V) were incubated with the T44P/A89P monomer for 10 min, and then α-Syn oligomers and t-vesicles (T) were added to the premixture. T and V (20 μM lipid concentration) were mixed together at 37 °C without α-Syn (T-V, gray line). Fusion efficiency was significantly reduced by the addition of α-Syn oligomers (10 nM, the black line). When the T44P/A89P α-Syn monomer was added to the T-V mixture together with oligomers, the FRET signal increased (200 nM, yellow line; 500 nM, light green line; 850 nM, green line; 2 μM, dark green line). A control measurement using soluble Syb2 confirmed that lipid mixing was driven by SNARE complex formation (beige line). Bar graphs were obtained from the values at 2000s from the FRET measurements (± SEM, N = 3). *P < 0.05, **P < 0.01, ***P < 0.001 by Student's t test. (B) The effect of the negatively charged lipid (PS) on fusion inhibition. v-Vesicles (V) prepared with 0% PS were used in the measurement. v-Vesicles without PS were incubated with wild-type α-Syn monomer, and then t-vesicles (T) and oligomers were added to the premixture. The gray and black lines denote the T-V mixture and the T-V mixture with α-Syn oligomers, respectively. When wild-type α-Syn monomer was added to the T-V mixtures together with oligomers, the FRET signal increased (200 nM, cyan line; 500 nM, light blue line; 850 nM, navy line; 2 μM, purple line). A control measurement using soluble Syb2 confirmed that lipid mixing was driven by SNARE complex formation (the beige line). Bar graphs were obtained from the values at 2000s from the FRET measurements (± SEM, N = 3). *P < 0.05, **P < 0.01, ***P < 0.001 by Student's t test. All graphs were produced using OriginLab 8.5 software (OriginLab Corporation) and edited by Adobe Illustrator CS6 (Adobe). www.nature.com/scientificreports/ monomer (Fig. 3B), which was consistent with the results of adding the T44P/A89P mutant (Fig. 3A). Strikingly, the addition of 2 μΜ α-Syn monomer to the fusion mixture in the presence of α-Syn oligomers recovered almost 100% of the SNARE-mediated lipid mixing (Fig. 3B). These results clearly demonstrate that the binding of the α-Syn monomer to the membrane lipids is required for the cooperative fusion inhibition by α-Syn oligomers.
Lipid-binding-deficient α-Syn mutants reduce the clustering of synaptic vesicle mimics by α-Syn oligomers. We showed that when the α-Syn monomer was treated with α-Syn oligomers, the wildtype α-Syn monomer induced clustering of v-vesicles and further inhibited vesicle fusion (Fig. 2D). In contrast to the wild-type α-Syn monomer, the α-Syn mutant T44P/A89P, having less lipid binding affinity, showed the opposite effect: the mutant attenuated the fusion inhibitory effect of α-Syn oligomers (Fig. 3A). We speculated that T44P/A89P α-Syn might alter the clustering of v-vesicles induced by α-Syn oligomers. To investigate this speculation quantitatively, we performed fluorescence correlation spectroscopy (FCS) measurements (Supplementary Fig. S6A). FCS measures the diffusion time (t d ) of vesicles, which is reflected in an autocorrelation curve (Fig. 4A). The diffusion time is related to the diffusion coefficient, which is determined by the viscosity of the solvent and the hydrodynamic radius of the vesicles in solution 40 . When vesicles are clustered, their diffusion time increases (Supplementary Fig. S6B). The average diffusion time of v-vesicles was approximately 14 ms (Fig. 4A, black curve). The diffusion time of v-vesicles increased to 31 ms with the addition of α-Syn oligomers (Fig. 4A, red curve), which indicates that α-Syn oligomers induced clustering of v-vesicles. Since the diffusion time increased by approximately two-fold, the addition of α-Syn oligomers induces roughly ten times bigger vesicle clusters in average. When 500 nM and 850 nM monomeric T44P/A89P α-Syn were preincubated with v-vesicles and α-Syn oligomers were then added, the diffusion time of v-vesicles was reduced to 21 ms and 15 ms, respectively (Fig. 4A). These results indicate that T44P/A89P α-Syn attenuates the clustering of v-vesicles induced by α-Syn oligomers. The clustered vesicles have more chance to be deposited on the surface of the glass, while small vesicles are diffusing in solution and may be detected more frequently in FCS measurement. Thus, the diffusion times in Fig. 4A have to be used qualitatively for observing the effect of T44P/A89P α-Syn. To confirm the results of the FCS study in Fig. 4A, we obtained cyro-EM images of vesicles. The interpretation of Fig. 4A was further supported by cryo-EM images, showing that the addition of T44P/A89P α-Syn significantly reduces the clustering of v-vesicles (Fig. 4B) compared with the clustering in the α-Syn oligomer treatment in Fig. 2C. Taken together, these results clearly demonstrate that monomeric α-Syn binds to negatively charged lipids for helping the clustering of v-vesicles by α-Syn oligomers.

Small peptide fragments of α-Syn reverse the fusion inhibitory effect of α-Syn oligomers.
We showed that the T44P/A89P α-Syn monomer blocks the clustering of v-vesicles induced by α-Syn oligomers, which results in restoration of SNARE-mediated vesicle fusion. T44P/A89P α-Syn does not bind to membrane lipids but can interact with Syb2 39 . In the same manner, the dopamine-induced α-Syn oligomers do not interact with the lipid membrane but do interact with the N-terminal domain of Syb2, which results in the inhibition of SNARE-mediated lipid mixing. Considering that T44P/A89P α-Syn and α-Syn oligomers commonly interact www.nature.com/scientificreports/ with Syb2, it is possible that T44P/A89P α-Syn competes with α-Syn oligomers for binding to Syb2 and reduces the clustering of v-vesicles by α-Syn oligomers. This may restore the vesicle fusion in Fig. 3A. Similarly, when we used v-vesicles without PS (negatively charged lipid), wild-type α-Syn monomer could not bind to the vesicle membrane. Wild-type α-Syn monomer, however, could bind to Syb2, which restored vesicle fusion in the presence of α-Syn oligomers. These results strongly indicate that a competitor of α-Syn oligomer binding to Syb2 may reverse the fusion inhibitory effect of α-Syn oligomers.
To verify this, we set to reverse the fusion inhibitory effect of α-Syn oligomers using small peptide fragments that compete with the α-Syn oligomer for Syb2 binding and may be potential therapeutic agents (Fig. 5A). We prepared three 30-residue peptides (CFα1, CFα2, and CFα3) that were mainly derived from the C-terminal domain of α-Syn (Anygen, South Korea) ( Fig. 5A; Table 1) because the C-terminal domain of α-Syn (a.a. 96-140) is known to bind to the N-terminal domain of Syb2 26 . We confirmed the binding between the peptide and Syb2 using fluorescence anisotropy assay (Supplementary Fig. S8). Next, v-vesicles were incubated with various concentrations of each peptide for 10 min in the presence of α-Syn oligomers (10 nM), and then t-vesicles were added into the premixture. We tested the effect of each peptide at the concentrations from 100 to 500 nM The addition of the protein-free vesicles showed no effect on the reversal effect of the small peptides (± SEM, N = 3). All graphs shown were produced using OriginLab 8.5 software (OriginLab Corporation) and edited by Adobe Illustrator CS6 (Adobe).  Figs. S9, S10, and S11). CFα1 and CFα2 effectively reversed the fusion inhibitory effect of α-Syn oligomers: 200 nM CFα1 increased the fusion efficiency from 45 to 68% (blue bars in Fig. 5B; Supplementary  Fig. S9), and 200 nM CFα2 increased the fusion efficiency from 45 to 90% (orange bars in Fig. 5B; Supplementary Fig. S10). However, CFα3 had no effect on vesicle fusion in the presence of α-Syn oligomers (green bars in Fig. 5B; Supplementary Fig. S11). These results imply that the small peptide-induced reversal of fusion inhibition is sequence dependent, and thus, a specific interaction between small peptides and Syb2 is required for the reversal. The 96-125 C-terminal region of α-Syn seems to be used mostly for binding to Syb2, consistent with previous work suggesting that α-Syn 96-110 is the key domain for Syb2 binding 38 . For control experiments, no change in lipid mixing was observed with 250 nM peptides in the absence of α-Syn oligomers, while 500 nM peptides slightly reduced lipid mixing (Fig. 5C). In order to be used as clinical drugs, it needs to be tested whether the small peptides have any effect on the functions of α-Syn monomer by interfering the interactions between α-Syn monomer and Syb2 in animal models. At least, however, we showed that the small peptides do not interfere the SNARE-mediated vesicle fusion (Fig. 5C). In the presence of 500 nM α-Syn monomers, no change in the SNARE-mediated vesicle fusion was observed for in vitro FRET assays (Supplementary Fig. S12). Then, we tested whether the small peptides bind to the lipid membrane. When we added 20 μM proteinfree liposomes (F) to a 20 μM T-V mixture, no change in the reversal effect of CFα1 and CFα2 was observed (Fig. 5D). Dopamine-induced α-Syn oligomers do not bind to lipid membranes 15 . Thus, this result indicates that the CFα peptides, as expected, do not bind to lipid membranes (Fig. 5D). All these results indicate that the small peptides of CFα1 and CFα2 reverse the fusion inhibitory effect of α-Syn oligomers by disrupting the specific protein-protein interactions between α-Syn oligomers and Syb2.

Discussion
In this work, we investigated the role of the interactions between α-Syn and v-SNARE Syb2 in the mechanism underlying α-Syn oligomer toxicity in vesicle fusion. We propose a mechanistic model for the cooperative inhibition of SNARE-mediated vesicle fusion by α-Syn monomer and oligomers (Fig. 6A). The α-Syn monomer simultaneously binds to the membrane lipid using its amphipathic N-terminal region and to Syb2 using its C-terminal region 25,29,38 . Recently, Fusco et al. showed that α-Syn monomer clusters vesicles using its two lipid binding sites 41 . Thus, the α-Syn monomer may induce a low degree of vesicle clustering 29,30,38,41 , but this clustering does not interfere with vesicle fusion. α-Syn oligomers, however, induce severe clustering of vesicles 15 . In the presence of α-Syn oligomers, the membrane lipid-and Syb2-bound α-Syn monomer assists the formation of large vesicle clusters, which results in severe fusion inhibition (Fig. 6A). www.nature.com/scientificreports/ In our previous work, we reported that α-Syn oligomers effectively inhibit SNARE-mediated vesicle docking 15 . For the molecular mechanism of the inhibition, we proposed that α-Syn oligomers interfere with vesicle docking by clustering vesicles. In the present work, we clearly showed, by using cryo-EM, that α-Syn oligomers indeed cluster Syb2-carrying vesicles (Fig. 2). Moreover, we showed that the clustering of vesicles can be reversed by blocking the interactions between α-Syn oligomers and Syb2 (Figs. 5, 6B). This clearly demonstrates that α-Syn oligomers induce vesicle clustering using their specific interactions with Syb2 (Fig. 6A). We note that the soluble α-Syn oligomers in this work was generated by dopamine oxidation. Other α-Syn oligomers need to be investigated to generalize our finding.
Burre et al. reported that α-Syn multimers (or oligomers) generated upon membrane binding promote SNARE complex assembly 25 . Notably, these α-Syn multimers bound to the membrane are normal assemblies that become monomers when they are released to the cytosol. However, α-Syn oligomers generated by dopamine as in this work, which are abnormal aggregates, have a low binding affinity to membranes 15,35 and are very stable in the cytosol [35][36][37] . In addition to the in vitro work, Mor et al. recently showed that dopamine increases the level of α-Syn oligomers and induces progressive nigrostriatal degeneration in vivo 18 .
Nonaggregated α-Syn is enriched in presynaptic terminals; its cellular concentration is estimated to be 6 μM 42,43 . Nonaggregated α-Syn or its monomer uses its interaction with Syb2 to help SNARE complex assembly 26 and induces the clustering of vesicles or Syb2, which promotes SNARE complex formation 25,30 . Recently, a nonaggregated α-Syn monomer was reported to be beneficial for SNARE-dependent vesicle docking even at the excessive concentration 10 μΜ 42 . All the functions of the α-Syn monomer are beneficial for neurotransmitter release. However, our results show that the generation of α-Syn oligomers, although low in abundance, incapacitates the potential benefit of the α-Syn monomer at the physiologically active location: the toxic effect of α-Syn oligomers is enhanced by the presence of monomers. α-Syn oligomers generated by dopamine contain approximately 17 monomers 15 and may bind multiple Syb2 proteins. Thus, the generation of even a small concentration of α-Syn oligomers may severely inhibit vesicle fusion in presynaptic regions where nonaggregated α-Syn is abundant.
The pathogenesis of PD is heralded by synaptic dysfunction. Impaired vesicular trafficking is thought to occur at the early stages of PD 9,44,45 . A recent study showed that α-Syn oligomers decrease neurotransmitter release and subsequently degenerate nerve terminals prior to cell bodies, implying that PD may break out at the synapse 18 . Our work shows that the α-Syn oligomer interacts with Syb2 to inhibit vesicle fusion. In addition, α-Syn oligomers do not interact with lipid membranes, and thus, the lipid membranes surrounding a cell do not dilute the available α-Syn oligomers. Thus, the low concentration of α-Syn oligomers, which are even assisted by the abundant α-Syn monomer at the presynaptic terminal, can inhibit vesicle fusion very effectively. α-Syn oligomers surrounded by monomers lead to synaptic dysfunction, which may be directly linked to impaired exocytosis of neurotransmitters at the synapse at early stages of PD [46][47][48][49][50] . To reduce the amount of pathological α-Syn, many efforts have been focused on reducing its synthesis or increasing its clearance 19,[51][52][53][54] . However, α-Syn gene knockdown may hamper the as-yet-unknown physiological functions of α-Syn, and the degree of knockdown required for treatment is largely unexplored 19,51 . Although antibodies against α-Syn have been reported to bind to aggregated forms of α-Syn and exert neuroprotective effects 53,[55][56][57] , there are several hurdles to overcome: crossing the blood-brain barrier, efficient intracellular targeting, inflammation, autoimmune reactions and other side effects 19,51,58 . In addition to these approaches, targeting specific interactions of α-Syn oligomers, which are directly linked to disease progression, can be another therapeutic strategy 59,60 . Our finding that vesicle fusion inhibition can be reversed by small peptides shows that the interaction between Syb2 and α-Syn oligomers can be a potential target to reduce the toxicity of α-Syn oligomers.

Materials and methods
Preparation of α-Syn monomer and dopamine-induced oligomers. A gene of recombinant α-Syn fused with glutathione-S-transferase (GST) was clones into a pGEX-KG vector. Escherichia coli BL21 Rosetta (DE3) pLysS (Novagene) was used for protein expression. The detailed purification procedures have been described elsewhere 15 . Briefly, the cells were grown at 37 °C in Luria-Bertani (LB) medium with 100 μg/ mL ampicillin and induced to express overnight (0.5 mM isopropyl β-d-thiogalactopyranoside) at 16 °C. The harvested cell pellets were resuspended in lysis buffer (1% sarcosine and 2 mM 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride in 1 × PBS buffer). After sonication of the cells, the supernatant of the lysate was transferred to a column containing glutathione-agarose beads, which were then incubated for 1.5 h at 4 °C and washed several times with 1 × PBS buffer. A thrombin (30 U) reaction cleaved α-Syn from the resin. For the preparation of α-Syn oligomers, 15 μM α-Syn was incubated with 100 μM dopamine in 20 mM sodium phosphate buffer (pH 7.0) at 37 °C for 72 h. Using size exclusion chromatography using Superdex 200 10/300 GL (GE Healthcare), α-Syn oligomers were purified and concentrated again using an Ultracel 10 K membrane.

Cryo-electron microscopy (cryo-EM)
. v-Vesicles were incubated with α-Syn (molar ratio of lipid:α-Syn oligomer: monomer = 200:1:20). A 200-mesh carbon grid (Electron Microscopy Sciences) was glow-discharged using the PELCO easiGlow plasma cleaning system (Ted Pella, Inc.). Sample vitrification was performed using a vitrification robot (FEI Company) by plunging the samples in liquid ethane. A droplet of sample solution (3 μL) was placed on a grid, automatically blotted with filter paper, and plunge-frozen in liquid ethane. The grids were then mounted in dedicated cartridges and stored under liquid nitrogen until data collection. The vitrified specimens were examined using an FEI Talos L120C Cryo-EM instrument (FEI Company) operating at a 120-kV acceleration voltage in the Nanobioimaging center of Seoul National University.

Fluorescence correlation spectroscopy (FCS).
We incubated v-vesicles doped with 0.02% DiI with/ without 10 nM α-Syn oligomers for 10 min at room temperature. We also incubated the v-vesicles with α-Syn T44P/A89P for 10 min, and then α-Syn oligomers were added into the premixture. We performed FCS measurements for each sample for 5 min using a home-built microscope setup 62 , and autocorrelation curves were obtained using a homemade LabVIEW analysis program 32 . Autocorrelation curves were obtained at 100 μM lipid.
Fluorescence anisotropy measurement. Fluorescence anisotropy measurements were performed using a fluorometer with two linear polarizers (QM-4/2005SE, Photon Technology). 15 nM CFα2 labeled with Cy5 was mixed with various concentrations of the vesicles reconstituted with Syb2 or N-terminal truncated Syb2. Cy5-CFα2 was excited with a polarized light at 633 nm and the emitted light was detected at 650-690 nm through the horizontal and vertical polarizers. Dissociation-equilibrium constants (K d ) were estimated by a single-site binding equation 63 .