Pathological structural conversion of α-synuclein at the mitochondria induces neuronal toxicity

Measuring oligomerization using a FRET biosensor

For visualization of α-Syn, fluorescent labels were introduced at an alanine-to-cysteine mutation at residue 90 (A90C), which is located at the periphery of the structure proposed to be highly organized in the fibrillar form and has negligible effects on both the aggregation and toxicity of α-Syn10,18. Two populations of fluorescently tagged α-Syn, one labeled with Alexa Fluor 488 (α-Syn-AF488) and the second with Alexa Fluor 594 (α-Syn-AF594) (Fig. 1ai,aii), were mixed and added to neurons, and the intracellular accumulation of α-Syn was visualized by measuring the intensity of AF594 α-Syn within the cells using direct excitation with 594-nm irradiation (Fig. 1bi,ci and Extended Data Fig. 1ai,aii). This direct excitation signal gives a measure of the total α-Syn, regardless of its aggregation state. The formation of oligomers was visualized via the presence of signal from the acceptor fluorophore (AF549) after excitation of the donor fluorophore with 488-nm irradiation, which occurs as energy is non-radiatively transferred from AF488 to AF594. We refer to this as the FRET signal (Extended Data Fig. 1b), and it can only occur when the fluorophores are in close proximity (<10 nm), as is the case within aggregates.

Fig. 1: FRET sensor detects rapid intracellular oligomerization of A53T α-Syn.figure 1

ai, Schematic illustration showing how FRET sensor detects aggregation. aii, AF488-α-Syn and AF594-α-Syn monomers are applied to cells, and the FRET signal is detected. bi, Representative bright-field (BF) FRET images after 72-hour incubation with oligomers. bii, Application of 500 nM WT oligomeric α-Syn exhibits detectable FRET, which increases over time (n = 3 independent experiments). ci, Representative FRET images after 72-hour incubation with monomers. cii, Application of 500 nM monomeric α-Syn exhibits low FRET signal initially, followed by an increase in FRET over time (n = 3 independent experiments). di,dii, A53T monomer exhibits the highest intracellular accumulation of α-Syn and the highest intracellular FRET intensity over time (n = 3 or 4 independent experiments). eieiii, FRET efficiency was calculated and binned into histograms that were fit to two Gaussian distributions. After 3 hours, only a low-FRET-efficiency population (centered at E = 0.24) was present. After 3 days, a second higher-FRET-efficiency population appeared (E = 0.48), and the fraction of this increased over time. f, Fraction of the high-FRET-efficiency population (out of total FRET events) increases over time for the WT and all mutants. Fitting error is shown in Extended Data Fig. 2c (n = 3 or 4 independent experiments). gi, Single-molecule confocal microscopy under conditions of fast flow used to analyze cell lysates. gii, 2D contour plots of approximate oligomer size and FRET efficiency after application of the monomers and oligomers. Both the number of events and the size of the oligomers increase over time in all cases. giii, Number and type of oligomeric events in cell lysates from the monomer/oligomer-treated cells. Data are represented as data ± s.d., as fraction of coincident events (n = 2 independent samples). Fitting error is shown in Extended Data Fig. 3a. hi, Photobleaching step analysis for A53T-oligomer-treated cells. hii, Step-fit example of a single A53T oligomer (24 hours) intensity trace (intensity is plotted as analogue-to-digital units (ADU)). hiii, Each step indicates photobleaching of a single fluorophore, from which oligomer size can be estimated. Note: Data are represented as data ± s.e.m. (box) unless otherwise mentioned. Detailed statistical information is shown in Supplementary Table 1. See also Extended Data Figs. 14. a.u., arbitrary units.

Source data

We first applied 500 nM oligomeric α-Syn (containing ~1% oligomer and ~99% monomer) to primary neurons: uptake of α-Syn resulted in a direct excitation signal (total α-Syn) and a FRET signal (oligomer formation), which both increased in intensity over time, reflecting the continual uptake of α-Syn and subsequent aggregation inside cells (Fig. 1bi,bii).

The kinetics of aggregation are dependent on α-Syn concentration: application of a range of initial concentrations of 5–50 nM oligomeric α-Syn induced oligomerization in a time-dependent and concentration-dependent manner (Extended Data Fig. 1ci). To verify that the intracellular FRET signal is a biosensor of aggregate formation, we confirmed that the FRET signal co-localized with an ATTO425-labeled aptamer specific to β-sheet-rich aggregates (Extended Data Fig. 1di,diii)19 and with a conformation-specific antibody (Extended Data Fig. 1dii,diii)12.

Next, we tested whether the application of α-Syn monomers alone (in the absence of oligomers) resulted in self-assembly and oligomer formation. After an initial lag phase, the FRET signal intensity increased (Fig. 1ci,cii). Oligomerization of the monomer occurred in a time-dependent and concentration-dependent manner (Extended Data Fig. 1cii). Within 3 hours, 500 nM monomer induced a FRET signal (oligomer formation), whereas 50 nM induced a FRET signal by 48 hours. We treated cells with 500 nM A53T, A30P, E46K or wild-type (WT) α-Syn monomers and measured both the total α-Syn and oligomer formation at different timepoints. Total α-Syn uptake and the FRET signal increased for WT and all mutants, showing a time-dependent increase in aggregate formation (Fig. 1di,dii and Extended Data Fig. 2ai,aii). Of note, A53T showed the greatest increase in both direct excitation signal and intracellular FRET signal over time, indicating increased uptake and oligomerization compared to the other mutations (Fig. 1di,dii). Oligomerization is concentration dependent for A53T α-Syn (Extended Data Fig. 2b). The formation of aggregates was further confirmed using an α-Syn oligomer-specific ELISA (Extended Data Fig. 2c).

Structural conversion measured by FRET efficiency

smFRET measurements can identify the in vitro structural conversion from less toxic, loosely associated ‘Type-A’ oligomers into toxic, proteinase-K-resistant, β-sheet-rich ‘Type-B’ oligomers18. Taking advantage of the intracellular FRET signal detected here, we determined whether distinct FRET populations of oligomers could form within neurons by calculating the FRET efficiency (E) of the aggregates within the cells (Equation 1 and Methods).

FRET efficiency histograms were generated from the intracellular aggregates after a short 3-hour incubation with WT 500 nM α-Syn monomer. The histograms were globally fit to two Gaussian distributions and integrated to obtain the number of converted oligomers in each sample (Fig. 1ei–eiii). The FRET efficiency increased over time for WT α-Syn, showing that the protein can oligomerize and structurally convert inside cells. At early timepoints (3 hours and 3 days), only the population of low-FRET-efficiency (centered at E = 0.24) oligomers were present. After 3 days, a second population of higher-FRET-efficiency oligomers (centered at E = 0.48) appeared. Assembly from monomer to oligomer occurred rapidly with a short lag phase (<3 hours), whereas conversion to a high-FRET oligomer occurred over days. Short incubation with A53T, A30P and E46K mutants (Fig. 1f and Extended Data Fig. 2d) revealed a similar structural conversion in cells.

A53T shows accelerated oligomerization and reduced lag phase

To measure the FRET efficiencies of individual oligomers formed in cells, we used single-molecule confocal microscopy on cell lysates (Fig. 1gi)18. We detected both low-FRET-efficiency and high-FRET-efficiency oligomers with a range of different sizes as shown in the two-dimensional (2D) contour plots (Fig. 1gii). Fitting of the resultant FRET efficiency histograms (Extended Data Fig. 3a) allowed the small (dimeric), Type-A and Type-B oligomers to be quantified (Fig. 1giii; the fitting errors are shown in Extended Data Fig. 3b). To determine the systematic error in separating two different FRET populations using this approach, we analyzed mixtures of two different dual-labeled DNA duplexes that had different FRET efficiencies (see section in Methods and Extended Data Fig. 2ei,eii).

A53T-oligomer-treated cells (1% oligomer and 99% monomer) exhibited rapid assembly to form oligomeric species of increasing size over 24 hours. A53T-monomer-treated cells also exhibited rapid assembly into small oligomeric species at 0–30 minutes, which increased in size over 24 hours. Self-assembly and the early steps of aggregation were observed in A30P-treated cells, although less than A53T-treated cells (Fig. 1gii,giii and Extended Data Fig. 3a,b). We confirmed that the rapid assembly of oligomeric species is formed inside cells after intracellular uptake, as aggregate levels were negligible in the media and high in the cell lysates as measured by the fraction of coincident events (Extended Data Fig. 3ci,cii) and by ELISA (Extended Data Fig. 3di,dii).

We used total internal reflection fluorescence (TIRF) microscopy to image the oligomers (Fig. 1hi and Extended Data Fig. 3e). As each α-Syn monomer carries a single dye molecule, there is a stepwise decrease in intensity as each one photobleaches upon irradiation. By counting the number of these photobleaching steps, the number of monomers per oligomer can be determined (Fig. 1hi,hii) in lysates after the addition of A53T or A30P monomer and oligomer. Histograms showing the number of monomers per oligomer were generated (Fig. 1hiii and Extended Data Fig. 3ei–eiii). These size distributions (Fig. 1hiii) correlated well with those measured using single-molecule confocal microscopy; approximately half of the population of oligomers contained two monomers, whereas the rest contained 4–10 monomers. Due to simultaneous photobleaching of multiple fluorophores, larger oligomers could not be distinguished using this method.

Structural conversion from Type-A to Type-B oligomers is associated with increased resistance to proteinase K degradation. We tested the effect of increasing concentrations of proteinase K on A53T aggregates in vitro and showed that the Type-A oligomers formed at early timepoints are sensitive to low concentrations of proteinase K, whereas the Type-B and larger oligomers are more resistant to proteinase K degradation, requiring higher concentrations (Extended Data Fig. 3fi,fii).

In addition to the added labeled α-Syn, cells also contain unlabeled, endogenous α-Syn. To test the effect of this on our FRET sensor, we aggregated 70 μM labeled (1:1 equimolar ratio of AF488 and AF594) α-Syn either in the absence or presence of 7 μM unlabeled WT α-Syn and characterized the aggregates using both single-molecule confocal and TIRF microscopy. Both the FRET and size distribution were unaffected by the presence of unlabeled α-Syn (Extended Data Fig. 4ai–4iii). We subsequently investigated whether different concentrations of endogenous (unlabeled) α-Syn would affect the FRET biosensor in human cells by applying AF488-labeled and AF594-labeled α-Syn to an isogenic series of iPSC lines with a range of SNCA alleles: SNCA null, SNCA 2 alleles and SNCA 4 alleles (Extended Data Fig. 4bi,bii). smFRET analysis of the size and FRET distribution of the labeled α-Syn aggregates in lysates, as well as the FRET in cell microscopy (Extended Data Fig. 4c,d), confirmed that there was a negligible effect of the endogenous α-Syn on the FRET signal.

Taken together, intracellular FRET, smFRET and TIRF microscopy can detect the initial stages of self-assembly, oligomer formation and structural conversion to proteinase-K-resistant Type-B oligomers inside cells. Aggregation inside cells is concentration and time dependent, and there is a reduced lag phase inside cells. Finally, there is an increased rate of oligomerization for A53T compared to WT and other mutants due, in part, to enhanced uptake and increased intracellular concentration of A53T.

Oligomer formation occurs in ‘hotspots’ at varied cell locations

To study the ultrastructural location of A53T oligomer formation in cells, we combined FRET imaging of labeled α-Syn with serial section electron microscopy (EM) and with focused ion beam milling combined with scanning electron microscopy (FIB-SEM) at 5-nm voxel resolution in human induced pluripotent stem cell (hiPSC)–derived neurons (Fig. 2a). We applied AF488-A53T α-Syn and AF594-A53T α-Syn monomer and oligomer to neurons, generated FRET signal heat intensity maps of aggregate formation (Fig. 2bi,bii and Extended Data Fig. 5a–c) and then tracked the same cells for EM.

Fig. 2: Oligomer formation occurs in multiple cell ‘hotspots’ at heterogeneous locations.figure 2

a, Schematic workflow of the experimental steps for 3D-CLEM. The light imaging was performed at 0.4-μm intervals; the radial point spread function (PSF) (as given by \(\lambda /2NA\), where λ is the excitation wavelength and NA is the numerical aperture of the objective lens) is 212 nm, and the axial PSF is approximately 500 nm. The EM section thickness is 5 nm, and the CLEM precision errors are shown in Extended Data Fig. 6. bi, FRET intensity heat maps showing aggregate formation with high FRET signal intensity at the core and low FRET intensity in the surrounding rim. bii, Intracellular FRET intensity increases after application of 500 nM A53T monomers (n = 3 independent experiments). c,d, Application of equimolar concentration of AF488-A53T α-Syn and AF-594-A53T α-Syn (total 500 nM); images were obtained at three different timepoints: 3 hours (c), 24 hours (d) and 7 days (e). Each panel is composed of a confocal image of CLEM (EM + FRET heat map) and zoom of EM alone. Colored arrows indicate aggregates at mitochondria (red), nucleus (white), membrane (yellow), Golgi apparatus (blue) and vesicles (orange). f,g, CLEM alignment using genetically engineered construct mitoGFP to label mitochondria (f: SEM and g: TEM). The red arrows indicate α-Syn detection within mitochondria. Note: Error maps for all images used for this figure are presented in Extended Data Fig. 5. Data are represented as data ± s.e.m. (box). *P < 0.05 and **P < 0.005. Detailed statistical information is shown in Supplementary Table 1. See also Extended Data Figs. 5 and 6. a.u., arbitrary units; PSF, point spread function.

Source data

Integrated spatial information was obtained by aligning using a nuclear marker only. At 3-hour and 24-hour incubation, FRET-FIB-SEM shows that oligomer formation occurs in a range of localizations in the cell that encompass nucleus, mitochondria and cytosol (Fig. 2c,d and Extended Data Fig. 5a). After 7 days, the FRET signal, combined with serial section EM, showed that the maturing aggregates (7 days; Fig. 2e) expand to occupy regions of the cell that encompass several different organelles as well as the cytoplasm in between organelles. Error maps using nucleus alignment confirm an overall error of 20–141 nm (Extended Data Fig. 6a–c)20. To improve confidence in the localization, we adopted organellar alignment of light and EM images using a genetically modified construct to visualize mitochondria (mitoGFP) and integrated the mito-GFP and AF-594 α-Syn fluorescence with the EM images. We show co-localization of α-Syn and mitoGFP fluorescence in mitochondria and degradative pathways for mitochondrial components—for example, autophagosome-like structures—using both FIB-SEM (Fig. 2f) and transmission electron microscopy (TEM) (Fig. 2g). Error maps for mitochondria-aligned CLEM reveal the accuracy of 20–134 nm (Extended Data Fig. 6d,e).

Aggregation hotspots form in cellular spaces crowded with organelles and then mature in a stereotypical manner into an aggregate that contains highly ordered aggregates at the center and loosely ordered protein in the rim. Early aggregation of A53T occurs in multiple locations within the cell, which act as a ‘seed’ for aggregation, including nuclear membrane, mitochondria and vesicles. Multiple seeding events in human cells may contribute to the reduced lag phase of aggregation.

Cardiolipin accelerates the oligomerization of A53T

α-Syn was observed in the FRET-CLEM data in association with mitochondria. Cardiolipin (CL) constitutes 10–20% of the mitochondrial membrane21,22, and so we investigated the interaction between α-Syn and CL-containing 100-nm-sized lipid vesicles (Extended Data Fig. 7ai,aii). Circular dichroism (CD) spectroscopy showed that α-Syn first adopts an α-helical conformation in the presence of CL-containing vesicles, as previously reported (Fig. 3ai)23. This change depends on the CL content in the vesicles and the lipid:protein molar ratio. With higher CL content and a higher lipid:protein molar ratio, we detected an increase in the α-helical structure (Fig. 3ai). Over time, CD spectroscopy (40% CL liposomes; 8:1 lipid:protein ratio) revealed a transition from α-helix to a mixture of protein structures (Fig. 3aii). Centrifugation and CD analysis of the pellet showed significant β-sheet content, as evidenced by a minimum at 210–220 nm (ref. 24). These data confirm that α-Syn interacts with CL and forms secondary structures (α-helical) initially but, over time, can acquire β-sheet-rich content.

Fig. 3: CL triggers and accelerates the aggregation of A53T α-Syn.figure 3

a, Effect of CL on far-UV CD spectra of α-Syn. ai, A53T monomer (10 µM) in the presence of 15% and 40% CL at 1:8, 1:16 and 1:40 lipid:protein ratios. aii, A53T monomer in the absence of liposomes (black curve), in the presence of 40% CL before incubation (green curve), at plateau phase (red curve) and insoluble fraction (blue curve) after incubation at 37 °C in the presence of 40% CL and 60% PC liposomes with 1:8 ratio. The minima between 210 nm and 220 nm for the insoluble fraction shows the presence of amyloid structures in the sample. bibiv, 50 µM of the A53T monomer led to a substantially fast increase in ThT fluorescence in the presence of 40% or 100% CL compared to WT monomer. c, Time-dependent SAVE images of A53T monomers incubated with 15% or 40% CL over 0–10 days show an increase in the number of aggregates over time. ci, Representative images. Red arrows indicate amyloid fibrils. cii,ciii, Quantification of the TIRF microscopy images. di,dii, TIRF microscopy analysis shows co-localization between CL and α-Syn fibrils (ThT positive). ei, TEM images show that, in the presence of CL, fibrils of α-Syn have different morphology. eii, Quantitative histogram of fibril width shows the large distribution of width in the presence of CL (100% CL), which is expected for a hierarchical self-assembly model of amyloid formation. A total of 200 fibrils were analyzed for each group using an Image-J plugin54. Note: Data are represented as mean ± s.d. Detailed statistical information is shown in Supplementary Table 1. See also Extended Data Fig. 7. a.u., arbitrary units; DMPC, dimyristoylphosphatidylcholine.

Next, we followed the formation of amyloid structure using the amyloid-binding dye thioflavin-T (ThT). A53T monomer exhibited rapid aggregation, with an absence of the lag phase observed as a sharp increase in the ThT fluorescence (Fig. 3bi,bii) in the presence of CL liposomes (40–100% CL, with phosphatidylcholine (PC), which alone does not induce aggregation (Extended Data Fig. 7b,c; 8:1 lipid:protein ratio)). A long and variable lag phase was observed in 40–100% CL and WT α-Syn (Fig. 3biii,biv). Single aggregate visualization by enhancement (SAVE) imaging with ThT (Fig. 3ci) showed that, in 15% or 40% CL (8:1 lipid:protein), aggregates formed over 10 days of incubation (Fig. 3cii,ciii). Forty percent CL led to the formation of large fibrillar aggregates after 10 days. Using biotinylated CL (100% CL, 8:1 lipid:protein ratio), we simultaneously visualized CL using Alexa Fluor 647 (AF647)–tagged streptavidin and aggregates using ThT, showing that amyloid fibrils contained CL, the percentage coincidence between ThT and CL using A53T monomer was 84% (95% confidence interval (CI), 73%, 106%), and using control (non-biotinylated lipid) it was 0.6% (95% CI, 0.3%, 3.5%) (Fig. 3di,dii). Our SAVE imaging data show that the interaction of A53T α-Syn and lipid vesicles results not only in extraction of CL but also in the incorporation of CL into the amyloid fibrils. We then studied the morphology of the fibrils formed in the presence of CL (8:1 lipid:protein ratio) by TEM and found that fibrils generated are markedly different from those formed in the absence of CL, showing helical periodicity along the length of fibrils and a greater number of protofilaments (Fig. 3ei,eii). Thus, CL vesicles may promote the lateral association of α-Syn protofilaments, which is consistent with the hierarchical self-assembly model of amyloid fibrils25.

Inside cells, we visualized endogenous CL using a fluorescent probe, nonyl acridine orange (NAO; Biotium, 70012), which co-localized with another mitochondrial indicator, tetramethylrhodamine methyl ester (TMRM) (Extended Data Fig. 7d). Control iPSC-derived neurons were treated with 1 uM AF488-A53T α-Syn or AF488-A30P α-Syn monomers for 48 hours. AF594 A53T α-Syn co-localizes with endogenous CL increasingly over time (Fig. 4a) and to a lesser degree than AF594 A30P a-Syn (Fig. 4b). We used super-resolution microscopy (aptamer DNA-PAINT to detect α-Syn aggregates and direct stochastic optical reconstruction microscopy (dSTORM)26) to show the presence of aggregates at the mitochondria (Extended Data Fig. 7e). To capture aggregate formation by both the externally applied α-Syn and the endogenous α-Syn, we used the dye amytracker, which binds specifically to β-sheet structures of amyloid aggregates. iPSC-derived neurons treated with WT oligomer exhibited increased aggregate formation (labeled with amytracker) in SNCA-A53T cells compared to isogenic control (iso-CTRL) and increased co-localization of those aggregates with CL (Extended Data Fig. 7fi–fiii).

Fig. 4: A53T α-Syn contacts CL as it aggregates.figure 4

ai,aii, Visualization of α-Syn contacts with CL in hiPSC-derived neurons using NAO. hiPSC-derived neurons were treated with 1 μM AF488-A53T α-Syn monomer, and contacts were measured at three timepoints: time 0, day 1 and day 5 (n = 5 fields imaged). bi,bii, Total α-Syn co-localisation with CL was higher in the cells treated with A53T monomers than with A30P monomers (n = 4 or 6 fields imaged). c, SMLM images show mitochondria labeled with Tomm20 (dSTORM) and aggregates labeled with a DNA-based aptamer (aptamer DNA PAINT). Panels on the right show a higher magnification. Note: Data are represented as data ± s.e.m. (box). **P < 0.005 and ***P < 0.0005. Detailed statistical information is shown in Supplementary Table 1. See also Extended Data Fig. 7.

Source data

Taken together, these data suggest that CL can rapidly trigger the aggregation of A53T, and CL is then incorporated into the aggregating structure of α-Syn, which may further accelerate the aggregation process.

A53T impairs mitochondrial bioenergetics and induces mitochondrial dysfunction

We investigated the functional consequence of A53T and WT α-Syn monomers applied exogenously to primary rodent neurons. Autofluorescence of complex I substrate NADH was used to measure cellular redox state and complex I function. We found that 500 nM A53T induced an increase in NADH fluorescence (119 ± 2.48%; Fig. 5ai and Extended Data Fig. 8ai), suggesting an inhibition of complex I function similar to the effect of oligomeric α-Syn27. A53T-induced complex I inhibition was associated with mitochondrial depolarization (120.4 ± 2.6%; Fig. 5bi and Extended Data Fig. 8aii), measured using rhodamine 123 fluorescence. NADH fluorescence could be partially restored by pre-incubation of cells with substrates for complex I (5 mM pyruvate) or complex II (membrane-permeable analog of succinate-dimethyl succinate (5 mM DMsuccinate)), showing that respiratory chain function could be rescued (108.9 ± 3.48% by pyruvate, 110 ± 0.5%; Fig. 5aii and Extended Data Fig. 8ai) and that improvements in the respiratory chain function also restore the mitochondrial membrane potential (Δψm) (107 ± 0.88% by DMsuccinate; Fig. 5bii and Extended Data Fig. 8aii). A53T reduced Δψm, measured by TMRM, after 30 mininutes (76.1 ± 3.1%), whereas Δψm was unchanged in WT monomer-treated cells (108.25 ± 6.30%; Fig. 5c). We investigated the decreased Δψm by testing its sensitivity to complex I and V inhibition. In healthy mitochondria in WT-treated cells, Δψm is maintained predominantly through the action of complex I–dependent respiration (reduction by 59.8 ± 4.04% in Δψm after rotenone but only 0.43 ± 0.25% reduction by the complex V inhibitor oligomycin). However, A53T-treated cells exhibited only 22.4 ± 2.94% reduction of Δψm after complex I inhibition but 40.3 ± 6.85% reduction by complex V (Fig. 5di–diii). Therefore, in cells exposed to the A53T mutant, the Δψm cannot be maintained sufficiently through respiration and must use complex V (in reverse mode as an ATPase) to maintain it. A53T-treated cells exhibited significantly reduced ATP production than WT-treated cells (Fig. 5eiii and Extended Data Fig. 8b).

Fig. 5: A53T α-Syn impairs mitochondrial bioenergetics and induces mitochondrial dysfunction.figure 5

ai, Increase in NADH autofluorescence after application of 500 nM A53T α-Syn (normalized to 1). aii, The increase in NADH is prevented by pre-incubation with pyruvate and succinate. bi, A53T monomer depolarizes Δψm as measured by an increase in rhodamine 123 (Rh123) fluorescence. bii, The decreased Δψm is also reversed by pre-application of pyruvate and succinate. ci,cii, Images showing reduction in Δψm after 30-minute incubation with A53T compared to WT and the quantitative histogram (n = 4 independent experiments). didiii, Response of Δψm to complex V inhibitor (oligomycin: 2.4 μg ml−1), complex I inhibitor (rotenone (ROT): 5 μM) and mitochondrial uncoupler (FCCP: 1 μM). The basal fluorescence intensity was reset at 1,500–2,500 a.u. (n = 3 independent experiments). eieiii, A53T reduces the total ATP production measured by FRET-ATP sensor compared to WT-treated or untreated cells (n = 3 or 4 independent experiments). fi,fii, Superoxide was increased after application of A53T but not WT α-Syn (n = 8 independent experiments). fiii, Inhibition of A53T-induced ROS by different inhibitors (n = 3 independent experiments). gigiii, mROS production was increased by A53T (n = 8 independent experiments). Note: 500 nM α-Syn monomer was applied for each experiment unless otherwise mentioned. Note: Data are represented as data ± s.e.m. (box). *P < 0.05, **P < 0.005 and ***P < 0.0005. Detailed statistical information is shown in Supplementary Table 1. See also Extended Data Fig. 8. a.u., arbitrary units.

Source data

We found that 500 nM A53T increased the rate of superoxide production in contrast to WT (Fig. 5fi,fii). The generation of cytosolic ROS was dependent on the concentration of applied monomer (Extended Data Fig. 8c). We investigated the source of the A53T-induced ROS using a range of inhibitors: mito-TEMPO or mitoQ (mitochondria-targeted antioxidants); Trolox (a water-soluble analog of vitamin E); and diphenyleneiodonium chloride (DPI) or 4-(2-aminoethyl)-benzolsulfonylfluorid-hydrochloride (AEBSF) (inhibitors of NADPH oxidase (NOX)). Mitochondrial ROS (mROS) scavengers effectively blocked A53T-induced ROS, suggesting that mROS may be a major source of excess ROS (Fig. 5fiii), with additional activation of NADPH oxidase. A53T monomer induced higher levels of mROS (176.6 ± 9.3% of basal) compared to WT (113.8 ± 4.4% of basal) (Fig. 5gi–giii). Time-lapse imaging of mitochondria within cells reveals the kinetics of uptake and mitochondrial dysfunction induced by A53T monomer application (Extended Data Fig. 8di,dii).

A53T α-Syn induces mitochondrial permeability transition pore opening

We previously showed that early opening of mitochondrial permeability transition pore (mPTP) mediates α-Syn oligomer-induced cell toxicity12. To test whether A53T α-Syn affects mPTP opening, cells were loaded with TMRM and the cytosolic calcium dye Fluo-4, followed by stepwise application of ferutinin, an inducer of mPTP opening by mitochondrial calcium overload28 (reviewed in ref. 29). A53T α-Syn lowered the threshold of mPTP opening (Fig. 6aiv). The latency to mPTP opening (rapid loss of TMRM fluorescence) after a high concentration of ferutinin was also reduced (Fig. 6bii–biv), and we observed that caspase 3–dependent apoptosis was induced in all cells that exhibited mPTP opening.

Fig. 6: A53T α-Syn induces mPTP opening, and mROS accelerates oligomerization and cell death.figure 6

ai, Representative time course images showing that Δψm reduction (TMRM) is followed by an increase of cytoplasmic calcium level (Fluo-4) at the point of mPTP opening. aii,aiii, Representative traces from the cells treated with 500 nM of WT or A53T α-Syn, respectively. aiv, A53T-treated cells require lower concentrations of ferutinin to open the mPTP than WT-treated cells (n = 4 or 6 independent experiments). bi, Representative time course images showing that apoptosis (NucView) is induced after a substantial loss of Δψm after ferutinin-induced PTP opening. bii,biii, Representative traces and WT-treated or A53T-treated cells. biv, A53T α-Syn treatment induces earlier PTP opening than WT α-Syn (n = 9 or 19 cells over two independent experiments). c, mPTP opening in isolated mitochondria from permeabilized cells. ci, Representative time course images of mPTP opening after applying AF-488-A53T α-Syn. cii,ciii, The mitochondrial area (ROI 1 area) exhibited a rapid loss of Δψm, whereas the extra-mitochondrial area (ROI 2) exhibited increased intensity of Rhod-5N after mPTP opening. civ, Quantitative histogram showing that PTP opening occurs earlier in A53T-treated than WT-treated mitochondria (n = 10 or 13 cells over two independent experiments). d, FRET intensity and FRET efficiency of A53T are reduced by treatment with mito-TEMPO. di, The representative images. dii,diii, Mito-TEMPO-treated cells show reduced A53T FRET intensity (dii; n = 3 or 4 independent experiments) and efficiency (diii; n = 20 or 15 cells over three independent experiments, and error bars represent 95% CIs). div, Application of Trolox to cells reduced FRET intensity signal by reducing uptake of donor (diidiv; n = 3 or 4 independent experiments). ei,eii, Cell death was induced by 48-hour incubation of A53T but not by WT or A30P/E46K (n = 3 independent experiments). fi,fii, A53T-induced cell death was rescued by treatment with mito-TEMPO (n = 4 or 5 independent experiments). Note: 100 μM Trolox and 0.5 μM mito-TEMPO (MitoT) were pre-treated 30 minutes before α-Syn application. Note: Data are represented as data ± s.e.m. (box). *P < 0.05, **P < 0.005 and ***P < 0.0005. Detailed statistical information is shown in Supplementary Table 1. a.u., arbitrary units.

Source data

We measured mPTP opening in isolated mitochondria of permeabilized cells. Cells were loaded with Rhod-5N and permeabilized with 40 µM digitonin in pseudo-intracellular solution30. Application of A53T monomers induced a rapid loss of Rhod-5N as the PTP opened and the dye left the mitochondria (ROI 1; Fig.

留言 (0)

沒有登入
gif