Ethical StatementBrain samples from four patients were used in this paper for RT-QuIC analysis, consisting of 3 male (healthy control, PD, and MSA) and 1 female (DLB). The were provided by the UK Brain bank and the ethical approval was obtained from the Cambridgeshire 2 Research Ethics Committee.Purification of α-synucleinα-Synuclein (WT or N122C variant) was overexpressed in Escherichia coli BL21 cells. The cells were centrifuged (20 min, 3985 × g, 4 °C; JLA-8.1000 rotor, Beckman Avanti J25 centrifuge (Beckman Coulter)), and the pellet resuspended in buffer (10 mM tris, 1 mM EDTA, protease inhibitor) prior to lysis by sonication on ice. Debris was removed by centrifugation (20 min, 39,121 × g, 4 °C; JLA-25.5 rotor), and the supernatant incubated (20 min, 95 °C). Heat-sensitive proteins were removed by centrifugation (15 min, 39,121 × g, 4 °C; JLA-25.5 rotor). Subsequent incubation with streptomycin sulfate (10 mg/mL, 15 min, 4 °C) precipitated out DNA. α-Synuclein was extracted from the supernatant (15 min, 39,121 × g, 4 °C; JLA-25.5 rotor) by the gradual addition of ammonium sulfate (361 mg/mL). The α-synuclein-containing pellet was collected by centrifugation (15 min, 39,121 × g, 4 °C; JLA-25.5 rotor) and resuspended in buffer (25 mM tris, pH 7.4). Dialysis was used for complete buffer exchange, and the resulting mixture run on a HiLoadTM 26/10 Q Sepharose high performance column (GE Healthcare), at room temperature. Under a gradient of 0–1.5 M NaCl over 600 mL, α-Synuclein was eluted at ~350 mM. Selected fractions were fractionated at room temperature on a Superdex 75 26/600 (GE Healthcare) and eluted in PBS (pH 7.4). For the N122C variant, 3 mM dithiothreitol (DTT) was added to all buffers to prevent dimerization. The concentration of α-synuclein was determined by absorbance at 275 nm, using a molar extinction coefficient of 5600 M-1 cm-1. Aliquots were then flash-frozen in liquid nitrogen and stored at −80 °C.Labelling of α-synucleinThe N122C variant of α-synuclein was fluorescently labeled with AlexaFluor-488 dye. DTT was removed from purified α-synuclein by buffer exchange into PBS using P10 desalting columns containing Sephadex G25 matrix (GE Healthcare). The DTT-free protein was incubated (overnight, 4 °C, rolling system) with a 1.5x molar excess of AlexaFluor-488 dye functionalized with a maleimide moiety. Excess unbound dye and α-synuclein dimers were removed by eluting the mixture over P10 desalting columns containing Sephadex G25 matrix (GE Healthcare). The resulting α-synuclein concentration was estimated by the dye absorbance at 495 nm, using a molar extinction coefficient of 72,000 M−1 cm−1. Aliquots were flash-frozen in liquid nitrogen and stored at -80 °C for up to 3 weeks prior to experiments.Aggregation of α-synucleinAggregation of α-synuclein was carried out in non-binding 96-well plates (Corning) at 37 °C in a FLUOstar Omega microplate reader (BMG Labtech). Each well contained 100 μL of reaction mixture and a glass bead. The buffer used was Dulbecco’s PBS (pH 7.4) with 0.01% (w/v) sodium azide. Interwell areas and empty wells were filled with PBS prior to sealing the plate with a foil cover. For experiments under shaking conditions, plates were shaken for 355 s at 200 rpm between each reading cycle; quiescent reactions were read at the same rate, but in the absence of all shaking. WT reactions were followed by the addition of 50 μM thioflavin T, whereas labelled N122C reactions (100% labelled protein) were monitored by AlexaFluor-488 fluorescence (Supplementary Fig. S7). Fibrils from unseeded reactions of 100 μM monomer under the same buffer and shaking conditions were used directly as seeds for seeding reactions, with no sonication (Supplementary Figs. S9 and Supplementary Fig. S10). For CSF-seeded aggregation reactions, CSF was added to a total of 4% volume in 100 μM α-synuclein under the same quiescent conditions as described above, and the sample subjected to shaking (10 min, 200 rpm) prior to measurement by FFE. CSF biospecimens used in the analyses presented in this article were obtained from The Michael J. Fox Foundation for Parkinson’s Research.RT-QuICBrain samples from UK brain bank and with confirmed autoptic diagnosis of synucleinopathies were employed in the present study. The samples included: Parkinson’s Disease (PD) cingulate cortex, Multiple System Atrophy (MSA) occipital cortex, Dementia with Lewy bodies (DLB) occipital cortex and healthy control occipital cortex. Brain homogenates (BH; 10% w/v) were prepared by homogenizing the tissue in 40 mM PBS containing 1% protease inhibitor (Halt, Thermo Scientific – 1860932) and 0.1% phenylmethylsulfonyl fluoride (PMSF), using a Bead Beater (Biospec Products; 11079110z) for 2 minutes at maximum speed. The homogenate was then spun at 3000 × g for 5 minutes at 4 °C and the supernatant was transferred to a new 0.5 ml lowBind Eppendorf tube and stored at −80 °C until RT-QuIC analysis. RT-QuIC reactions were performed in black 96-well plates with a clear bottom (Nalgene Nunc International). Plates were preloaded with 3 silica beads (1 mm diameter, BioSpec Products) per well. For BH-seeded reactions, 4 μL of the indicated BH was added to wells containing 95 μL of the reaction buffer to give final concentrations of 40 mM phosphate buffer (pH 8.0), 170 mM NaCl, 100 μM of monomeric AlexaFluor-488-labelled N122C α-synuclein (filtered through a 100 kDa MWCO filter immediately prior to use). The plate was then sealed with a plate sealer film (Nalgene Nunc Inter- national) and incubated at 37.5 °C in a BMG FLUOstar Omega plate reader with cycles of 1 min shaking (500 rpm double orbital) and 15 minutes rest throughout the indicated incubation time. Fluorescence measurements (490 ± 5-nm excitation and 520 ± 5-nm emission; bottom read) were taken every 15 min. Each sample was run in three technical replicates.Analysis of bulk kinetic dataThe signal obtained during aggregation kinetics (ThT fluorescence or AlexaFluor-488 fluorescence) was taken to be proportional to the fibril mass present in the sample. The data were then fitted using the AmyloFit Platform and following the protocol in Meisl et al. to a model including primary nucleation, elongation and secondary nucleation with reaction order 030. This model was able to describe the data well across concentrations. To produce misfits, the same data were fitted with a model including only primary nucleation and elongation.Measurement of fibril length distributionsAt certain timepoints in the plateau phase of the aggregation reaction, 1 μL of the reaction was withdrawn and the plate returned to the platereader. The reaction sample was mixed with 9 μL PBS (pH 7.4), and applied to a transmission electron microscopy (TEM) grid (continuous carbon film on 300 mesh Cu). Following adsorption, the sample was washed with milliQ water (2 × 10 μL). Samples were negatively stained with uranyl acetate (2% w/v, 10 μL, 2 min) and washed with milliQ water (2 × 10 μL). TEM grids were glow discharged using a Quorum Technologies GloQube instrument at a current of 25mA for 60s. TEM images were obtained using a Thermo Scientific (FEI Company, Hillsboro, OR) Talos F200X G2 microscope operated at 200 kV. TEM images were recorded using a Ceta 4k × 4k CMOS camera. The lengths of imaged fibrils were manually determined with ImageJ (example images in Figure S9). The lengths of between 650 and 1550 individual fibrils were measured for each sample.Analysis of fibril length distributionsThe fibril length in the plateau phase of the reaction, during which the fibril mass concentration is constant, can be modeled by the following equations. By definition, in the plateau phase, the aggregate mass concentration is constant, i.e., M(t) = M∞. While nucleation processes become negligible when the monomer is depleted, fragmentation still takes place, thus the number concentration of fibrils, P(t), is given by$$\frac{dP}{dt}={k}_{-}{M}_{\infty },$$
(1)
which can be solved to yield$$P(t)={k}_{-}{M}_{\infty }t+{P}_{plateau},$$
(2)
where k− is the fragmentation rate constant, t is the time since the plateau was first attained and Pplateau is the number concentration of fibrils at t = 0.The mean length, L(t), at the plateau is thus given by:$$L(t)=\frac{M(t)}{P(t)}=\frac{{M}_{\infty }}{{k}_{-}{M}_{\infty }t+{P}_{plateau}}=\frac{1}{{k}_{-}t+\frac{1}{{L}_{plateau}}},$$
(3)
where Lplateau is the average length when the plateau is first attained. Using the steady-state expression for the average length during an aggregation reaction derived e.g., in Cohen et al.60,61,62 as an estimate for Lplateau, the rate of fibril formation due to fragmentation is approximately given by κ(frag) = Lplateauk− = 0.01 h−1. This is thus an estimate of the rate of fibril formation based purely on measurements of fibril lengths which can then be compared with kinetic measurements of the actual rate of fibril accumulation κ to see if this is consistent with a purely fragmentation-driven mechanism.Fabrication of microfluidic free-flow electrophoresis devicesMicrofluidic free-flow electrophoresis (μFFE) devices were designed and fabricated as follows. Briefly, acetate masks were used to produce SU-8 molds of devices by photolithography, the heights of which were measured with a profilometer (Dektak, Bruker, Billerica, MA). Polydimethylsiloxane (PDMS; 1:10 mixture of primer and base, Dow Corning) was applied to the mold and baked (65 °C, 1.5 h). μFFE devices were then excised and biopsy punches used to create holes for tubing and electrode connections, with diameters of 0.75 mm and 1.5 mm, respectively. Following sonication in isopropanol (5 min), devices were bonded to glass coverslips (#1.5) by activation with oxygen plasma. Immediately prior to use, prolonged oxygen plasma treatment was used to hydrophilize device surfaces.
μFFE device operationThe μFFE device design used contains liquid electrodes (3 M KCl solution containing 1 nM Atto-488 dye) to connect the electrophoresis chamber to the external electric circuit21,63. These liquid electrodes were connected to the circuit via hollow metal electrodes made from bent syringe tips, which also constituted the outlets for the liquid electrodes. Samples were flowed into the device at controlled flow rates by the use of syringe pumps (Cetoni neMESYS, Korbussen, Germany), connected to polytetrafluoroethylene (PTFE) tubing (0.012″ inner diameter × 0.030″ outer diameter, Cole-Parmer, St. Neots, UK). The flow rates used were 1000, 200, 140, and 10 μL h−1 for the auxiliary buffer (15× diluted PBS in milliQ water), electrolyte, desalting milliQ water, and sample, respectively. The electric field was applied by a benchtop power supply (Elektro-Automatik EA-PS 9500-06, Viersen, Germany) connected to the metal electrode outlets.Acquisition of μFFE dataμFFE data were acquired using laser confocal fluorescence microscopy; a 488 nm wavelength laser beam (Cobolt 06-MLD 488 nm 200 mW diode laser, Cobolt, Stockholm, Sweden) was coupled into single-mode optical fibre (P3-488PM-FC01, Thorlabs, Newton, NJ) and collimated (60FC-L-4-M100S-26, Schäfter und Kirchhoff, Hamburg, Germany) before being directed into the back aperture of an inverted microscope body (Applied Scientific Instrumentation Imaging, Eugene, OR). The laser beam was then reflected by a dichroic mirror (Di03-R488/561, Semrock, Rochester, NY) and focused to a concentric diffraction-limited spot in the microfluidic channel through a high-numerical-aperture water-immersion objective (CFI Plan Apochromat WT 60x, NA 1.2, Nikon, Tokyo, Japan). Photons arising thorugh fluorescence emission were detected using the same objective. Fluorescence was then passed through the dichroic mirror and imaged onto a 30 μm pinhole (Thorlabs), removing out of focus light. The signal was then filtered through a bandpass filter (FF01-520/35-25, Semrock), and focused onto a single-photon counting avalanche diode (APD, SPCM-14, PerkinElmer Optoelectronics, Waltham, MA). Photons were recorded using a time-correlated single photon counting (TCSPC) module (TimeHarp 260 PICO, PicoQuant, Berlin, Germany) with 25 ps time resolution. Single-photon counting recordings were obtained using custom-written Python code.Aggregation samples (100 μL) were withdrawn from the plate at various times during the aggregation reaction and centrifuged (21,130 × g, 10 min, 20 °C). The top 70 μL was carefully withdrawn without disturbing the pellet containing large aggregates. An aliquot of the supernatant was then diluted in PBS to ~5 μM total monomer mass concentration and injected into the device. An electric field of 300 V was applied and photon count timetraces obtained at 5–10 positions laterally distributed across the field direction for a total of at least 1 min per position.Analysis of μFFE dataA detailed account of the analysis of μFFE data is provided in the SI.Fitting of oligomer dynamicsThe coarse-grained rate equations governing oligomers (concentration S(t)) formed by primary nucleation during an amyloid fibril formation reaction are:$$\frac{dS}{dt}={k}_{{{{\rm{o1}}}}}m{(t)}^{{n}_{{{{\rm{o1}}}}}}-{k}_{e1}S(t),$$
(4)
where m(t) is the concentration of monomeric protein, ko1 is the rate constant for oligomer formation from primary nucleation with reaction order no1, and ke1 is the rate constant for dissociation of oligomers to monomers. In a system dominated by secondary nucleation of oligomers, the rate equations are instead well-approximated by:$$\frac{dS}{dt}={k}_{{{{\rm{o2}}}}}m{(t)}^{{n}_{{{{\rm{o2}}}}}}M(t)-{k}_{e2}S(t)M(t),$$
(5)
where M(t) is the mass concentration of amyloid fibrils and ko2 is the rate constant for oligomer formation from fibril-dependent processes, with reaction order no2. For simplicity, we modelled M(t) using the analytical expressions given in ref. 51 with rate parameters chosen as the values determined by fitting the bulk data on fibril formation in the main text (Fig. 2 and Table S2). Eqs (4) and (5) were then fitted numerically to the experimental data on oligomer concentration using python. Eq. (5) provided the superior fit, supporting the conclusion that oligomers are formed predominantly by secondary processes in this assay.Reporting summaryFurther information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
α-Synuclein oligomers form by secondary nucleation
