Direct observation of the interconversion of normal and toxic forms of α-synuclein

Abstract

Here, we use single-molecule techniques to study the aggregation of α-synuclein, the protein whose misfolding and deposition is associated with Parkinson's disease. We identify a conformational change from the initially formed oligomers to stable, more compact proteinase-K-resistant oligomers as the key step that leads ultimately to fibril formation. The oligomers formed as a result of the structural conversion generate much higher levels of oxidative stress in rat primary neurons than do the oligomers formed initially, showing that they are more damaging to cells. The structural conversion is remarkably slow, indicating a high kinetic barrier for the conversion and suggesting that there is a significant period of time for the cellular protective machinery to operate and potentially for therapeutic intervention, prior to the onset of cellular damage. In the absence of added soluble protein, the assembly process is reversed and fibrils disaggregate to form stable oligomers, hence acting as a source of cytotoxic species.

Graphical abstract

Figure 1

Experimental Protocol (A) Example of the kinetics of amyloid fibril formation, including hypothetical snapshots of the ensemble of αS species present at different phases of the aggregation process, wherein the oligomeric species present in the sample are highlighted with red circles. (B) Schematic representation of the instrument used for smFRET measurements. (C) Schematic description of the experimental protocol for aggregation experiments. Bursts of fluorescence coincident in both channels indicate the presence of FRET-positive oligomeric species (marked as asterisks). Noncoincident bursts can be attributed to monomers and are normally much less bright than those corresponding to oligomers. (D) Comparison between the kinetics of oligomer formation under bulk conditions obtained by quantitative SEC analysis (red circles; see also Figure S1G) and from smFRET experiments, after extrapolating the concentrations from sm to bulk conditions (gray squares). The data reported correspond to the mean and standard errors of two repetitions in the case of the SEC data and five repetitions for the sm data. See Figure S1 for a detailed characterization of the effect of the fluorophores on αS aggregation.

Figure 2

αS Oligomers Present during Fibril Formation Aggregation of αS was followed by smFRET at a range of time points during incubation. (A) 2D plots corresponding to the relative mass distribution of oligomers at different incubation times as a function of apparent oligomer size (x axis) and FRET efficiency (y axis); for a clearer visualization of the presence of large oligomers, the mass rather than the number distribution is represented. (B) A representative 2D plot of the number distribution of oligomers after 60 hr of incubation illustrates the size-dependence of the FRET efficiency distributions (vertical lines corresponding to apparent oligomer sizes of 5- and 15-mer have been added as a visual guide), where two main FRET oligomer populations can be identified. The FRET-derived distributions based on small (red bars), medium (green bars), and large (blue bars) classification are plotted on the graph shown below. (C) Representative global fits of the four size-derived FRET oligomer distributions to Gaussian functions as the incubation time varies. See also Figure S2.

Figure 3

Characterization of the Different Oligomeric Species Formed during αS Aggregation and Fibril Formation (A) The time dependence of the mass fraction of the four oligomeric distributions A_small (red squares), A_medium (orange circles), B_medium (green triangles), and B_large (blue triangles). The data shown correspond to the average and standard error of five different experiments. (B) Proteinase K degradation curves of the different protein species (monomer in red, type A oligomer in orange, type B oligomer in blue, and fibrils in black). The data shown here correspond to the average and standard error of three different experiments. See also Figure S3.

Figure 4

αS Fibril Disaggregation (A) 2D plots of the relative mass distribution of oligomers according to apparent size and FRET efficiency, as explained in Figure 2A. (B) Representative global fits of the size-derived FRET oligomer distributions as a function of the incubation time. See also Figure S4 and Table S1.

Figure 5

Kinetic Analysis of αS Oligomerization (A) Scheme for the minimalistic kinetic model used to fit the early stages of αS aggregation. (B) Results of the global fitting (continuous lines) of the kinetics of formation of the two types of oligomeric species estimated under bulk conditions from smFRET experiments. Data for type A oligomers and type B oligomers are shown as orange circles and blue squares, respectively (average and standard error of five different experiments). The vertical dashed line is at 30 hr, corresponding to the lag phase of fibril formation estimated under bulk conditions, up to which time our model accounts well for the different microscopic processes governing the aggregation reaction. (C) Cartoon showing the conversion of an 8-mer of αS from a collapsed to an ordered proteinase-K-resistant form. Residues of each monomer are colored according to their location in the amino acid sequence. The average distance between fluorophores, represented as green and red spheres, is different for each type of oligomer and hence gives rise to different average FRET efficiencies.

Figure 6

Evaluation of the Pathophysiological Effects of the αS Species on Rat Primary Neuronal Cultures (A) Representative traces showing the rate of uptake of the different types of αS species into the cell body. The fluorescence of AF488-labeled αS was used to track monomeric αS uptake, whereas the fluorescence emission of AF647-labeled αS was used to follow the uptake of oligomeric and fibrillar forms of the protein (see also Figure S5), as only these species are able to produce FRET signals. (B) Bar chart displaying cytoplasmic (HEt) and mitochondrial (MitoSOX) ROS production induced in cells after exposure to each αS sample (see “PK-Resistant Oligomers Induce Higher Aberrant Levels of ROS in Cells than Do PK-Sensitive Oligomers” section in the main text for a detailed explanation of the composition of each sample). ^∗p < 0.05; ^∗∗p < 0.005. (C) Representative cellular traces showing the slower production of ROS induced by type B αS oligomeric samples after cellular treatment with ABSEF. Error bars in (B) and (C) represent SEM.

Figure S1

Characterization of Labeled A90C αS at Bulk Conditions, Related to Figure 1 (A) αS aminoacidic sequence, highlighting the three main regions: N-terminal (residues 1-60), NAC (61-95) and C-terminal (95-140); and the five regions proposed to form the strands of the beta-sheet sandwich core of the fibrillar structure (Vilar et al., 2008): strand β1 comprising residues 37-43; strand β2, 52-59; strand β3, 62-66; strand β4, 68-77 and strand β5, 90-95. The fluorescence dyes were covalently attached to position 90 in the sequence by cysteine chemistry. (B) The fold of αS fibrillar structure proposed by Vilar et al. (Vilar et al., 2008) is shown. The fluorescence dyes would be positioned parallel at the periphery of the fibril core according to this model. (C) Dynamic light scattering derived size distribution of unincubated AF488 A90C αS (red line) compared with unlabeled WT protein (black line). Representative size distribution of 30 μM protein concentration in Tris 25 mM, pH 7.4, 0.1 M NaCl at 25°C measured on the Zetasizer Nano ZS instrument at 633nm. The distribution represents the average of 4 measurements and the intensity was normalized and weighted by the particle size. The influence of AF647 could not be measured with our DLS instrument due to the direct absorption of the laser light (633 nm) by the fluorophore, although very similar results are expected. (D) The kinetics of fibril formation for the unlabeled protein was independently analyzed by the addition of Thioflavin T (ThT) to the reaction sample at different incubation times (open circles: the signal was normalized and 1-(normalized signal) was plotted to compared with the kinetics of the depletion of the monomeric protein analyzed by SEC). The decrease in monomeric protein during incubation was also estimated by quantitative SEC analysis after centrifuging the samples to remove the insoluble material (closed circles) or after ultracentrifugation to remove big soluble oligomers, but no differences were observed. The lag time and apparent kinetic rate were obtained. Error bars represent SEM. (E) For the case of labeled protein, analysis using ThT cannot be applied, since the fluorescence increase of ThT molecules upon binding to the amyloid fibrils is significantly reduced, probably due to fluorescence quenching and/or FRET between ThT and the Alexa fluorophores. For this reason, the kinetics of fibril formation was followed by quantitative SEC (closed circles; see Figure S1G) and SDS-PAGE gel (open circles), where the soluble protein material was quantified by the spectroscopic properties of the Alexa fluorophores. The aggregation of mixed AF488 and AF647-labeled protein is shown as blue circles, and the aggregation corresponding to AF488-labeled protein incubated alone is shown in red. Error bars represent SEM. (F) TEM images show that the morphology of the labeled fibrils formed (image on the right) is very similar to those formed with unmodified protein (image on the left). Amorphous aggregates were not observed in any case. (G) Oligomers were detected and analyzed by quantitative SEC as a function of incubation time. Three peaks were observed in the chromatogram: a peak eluting at 7 ml, which corresponds to the column void volume, and then to oligomeric species, a large peak eluting at 9.3 ml which corresponds to the monomeric protein, and a third peak at eluting volumes bigger than 15 ml, corresponding to some fragments of the protein generated upon incubation. The concentration of both monomeric and oligomeric fractions of the protein at the different incubation times recorded were estimated from the area of the peaks, taking into account the known initial concentration of protein and that at time zero, all the protein remains monomeric.

Figure S2

Aggregation Kinetics of αS, Related to Figure 2 (A) Some examples of the raw data obtained in smFRET experiments: the fluorescence intensity recorded for AF488-labeled molecules is in blue, the fluorescence intensity recorded for AF647-labeled molecules in red, and the coincident events in both channels are highlighted in green. In smFRET experiments, AF488-labeled molecules are directly excited by a 488nm-laser so both monomeric and oligomeric species containing this fluorophore are detected in the blue channel (non-coincident and coincident event, respectively). However, the AF647-labeled molecules in the oligomeric species are indirectly excited by FRET from excited AF488-labeled molecules, and therefore only oligomeric species are detected in the red channel. (B) Aggregation of αS at physiologically relevant concentrations. The same type of experiment and analysis was carried out at 6 μM protein concentration, a value close to the proposed physiological concentration of the protein in cells, but below the critical concentration for fibrillization found in vitro. With this concentration, after incubating the protein sample for more than 150 hr, no fibrils were detectable by TEM nor reduction in the amount of soluble protein according to SDS-PAGE gel analysis, although oligomeric species were detected by smFRET experiments. In the top panels, 2D plots corresponding to the mass distribution of oligomers at different incubation times as a function of apparent oligomer size (x axis) and FRET efficiency (y axis) are represented. In the bottom panels, FRET-derived distributions of the three size classes of oligomers: small (red bars), medium (green bars) and large (blue bars) are plotted. The FRET distributions at different incubation times were globally fitted to Gaussian functions (continuous lines) to estimate the mass fraction of each class of oligomers as a function of incubation time.

Figure S3

Analysis of the Stability of the Different Protein Species against Proteinase K Degradation, Related to Figure 3 (A) Incubations of an aliquot of an aggregation sample at late incubation times (around 100 h) with different concentrations of proteinase K were analyzed by single-molecule fluorescence to determine the stability of the different oligomeric species against proteinase K degradation. The pair of A_small and A_med distributions shows similar dependence on proteinase K concentration, as well as for the pair of B_med and B_large distributions, although this latter pair has much higher resistance to proteinase K. (B) The stability to proteinase K degradation of monomeric and fibrillar forms of αS was estimated from the disappearance of the band corresponding to full-length monomeric in a SDS-PAGE gel. Different concentrations of proteinase K were used as indicated in the figure. All samples from monomeric, fibrillar and oligomeric forms of αS were treated exactly in the same way for direct comparison.

Figure S4

Fibril Disaggregation Experiments, Related to Figure 4 Both the kinetics of the fraction of oligomers (A) and the monomer concentration of the solution (B) as a function of the incubation time are qualitatively well reproduced between smFRET experiments of fibril disaggregation reactions (the three sets of data correspond to three different fibril disaggregation experiments carried out in the same conditions). (C) Single- molecule Dual-Color Total Internal Reflectance Fluorescence (TIRF) images of αS oligomers obtained from a fibril disaggregation experiments (the three images on the left). The oligomers often displayed uneven fluorescence along their length (i) and bent/branched structures (ii). Others appeared punctate, either indicating a size smaller than the diffraction limit or oligomers standing upright in the TIRF field. The size and shape of large oligomers observed by TIRF are very similar with those seen by TEM (images on the right). Scale bar = 100 nm. Zoomed images showed branched/bent structure and aggregates with uneven width along their length, which could correspond to fluorescent structures shown in the panels on the left. (D) Example of processed data for fibril dissolution experiment with the red channel data on the left half of the image and blue channel data on the right. Non-coincident red and blue events are marked with pink and blue circles respectively. Events, which are coincident within 500 nm, are marked with yellow circles. (E) Histogram of FRET efficiencies for early fibril dissolution (FD) and late aggregation samples (LA). Values are normalized to the total number of detected objects: FD = 1888, LA = 938. F) Apparent oligomer size distribution showing greater size of oligomers for fibril dissolution (FD) sample compared to late aggregation (LA) sample.

Figure S5

Cell Uptake of αS Species, Related to Figure 6 (A) Monomeric sample, (B) oligomeric A sample, (C) oligomeric B sample, and (D) fibrillar sample. (i) Phase contrast images of rat primary neurons for each protein sample. Boxed areas show the cytoplasmic regions of cells utilized to record αS uptake. Cytoplasmic regions were specifically selected to ensure species uptake and not plasma membrane absorption was being measured. (ii) Fluorescence signal of AF488, exciting at 488nm after the addition of each αS species sample. (iii) Fluorescence signal of AF647, exciting the sample at 488nm to identify uptake of oligomeric and fibrillar species by FRET (monomers are unable to undergo FRET). (iv) Magnified merged image of boxed area in corresponding images ii and iii. (v) Profile of AF488 and AF647 fluorescence emission intensities along the red arrow in image iv. Simultaneous increase of both AF488 and AF648 signals, indicated by arrows, identify FRET signal within cell body. Images iii – v are absent for monomeric αS sample (A) because the monomeric species do not FRET.