Lipid Dynamics and Phase Transition within α-Synuclein Amyloid Fibrils

The deposition of coassemblies made of the small presynaptic protein, α-synuclein, and lipids in the brains of patients is the hallmark of Parkinson’s disease. In this study, we used natural abundance 13C and 31P magic-angle spinning nuclear magnetic resonance spectroscopy together with cryo-electron microscopy and differential scanning calorimetry to characterize the fibrils formed by α-synuclein in the presence of vesicles made of 1,2-dimyristoyl-sn-glycero-3-phospho-L-serine or 1,2-dilauroyl-sn-glycero-3-phospho-L-serine. Our results show that these lipids coassemble with α-synuclein molecules to give thin and curly amyloid fibrils. The coassembly leads to slower and more isotropic reorientation of lipid molecular segments and a decrease in both the temperature and enthalpy of the lipid chain-melting compared with those in the protein-free lipid lamellar phase. These findings provide new insights into the properties of lipids within protein–lipid assemblies that can be associated with Parkinson’s disease.


Protein and lipid preparation
α-synuclein was expressed and purified using the same protocol as that reported previously. S1-S3 Lipid vesicles used for the aggregation experiment and differential scanning calorimetry measurement were prepared via extrusion and sonication, respectively. Briefly, lipids (DMPS or DLPS) in powder form were suspended in phosphate buffer and the solution was stirred for 4 h at 45 • C (a temperature above the melting temperature of both DLPS and DMPS). Then the lipid suspensions were sonicated on ice or extruded through 100nm pore membranes (Avanti Polar Lipids, Alabama, USA) at 45 • C.

Cryo-electron Microscopy (Cryo-EM)
50 µM α-synuclein was incubated in phosphate buffer in the presence of lipid vesicles made with DMPS or DLPS (100 µM) under quiescent conditions and at 30 • C for 4 d. The grids were then prepared as described previously. S4 Briefly, reaction mixtures containing α-synuclein proto-fibrils were deposited on lacey carbon filmed copper grids, which were then plunged into liquid ethane at -180 • C. The resulting grids were stored under liquid nitrogen until imaged using an electron microscope (Philips CM120 BioTWIN Cryo) equipped with a post-column energy filter (Gatan GIF100) (acceleration voltage: 120 kV). The images were recorded digitally with a CCD camera under low electron dose conditions.

Sample preparation
The proto-fibrils were prepared by incubating 100 µM α-synuclein and 2 mM DMPS or DLPS solubilised as vesicles in phosphate buffer (20 mM NaH 2 PO 4 /Na 2 HPO 4 , 0.01 % NaN 3 , pH 6.5) in Corning R 96-well half area black with clear flat bottom made of polystyrene treated with NBS TM (#3881, Corning Ltd, Corning, USA) under quiescent conditions at 30 • C for 4 d. Reaction mixtures were then collected from each well, gathered and centrifuged at 90 krpm for 1 h. The pellet was then inserted into NMR rotor inserts using the following procedure: the insert was attached to a 10 µL tip using parafilm and the sample was deposited at the larger end of the tip. The insert attached to the tip was then centrifuged for 30 s at 1,000 g to let the sample slide into the rotor.
Samples of protein-free lipid lamellar phase, referred to as "pure lipid system" throughout the text, were prepared by suspending 10 mg DMPS or DLPS powder in phosphate buffer for 2 h at maximum stirring and 50 • C. The samples were then centrifuged at 90 krpm for 1 h and the pellets transferred into the rotor, as described above. All experiments were performed with fully hydrated samples to avoid dehydration-induced changes in lipid self-assembly.

Data acquisition
A set of four spectra was acquired for each sample: 13 C MAS direct polarisation (DP), 13 C MAS cross-polarisation (CP), 13 C MAS Insensitive Nuclei Enhanced by Polarization Transfer (INEPT) and 31 P MAS NMR spectra. Measurements were performed using a Bruker Avance-II 500 spectrometer (Bruker, Karlsruhe, Germany) equipped with a 4 mm CP/MAS HX probe at a field of 11.7T, resulting in 1 H, 13 C and 31 P resonance frequencies of 500, 125 and 200 MHz, respectively. For temperatures below 20 • C, we used a BVT-2000 temperature control and a BCU-05 cooling box unit, with sample heating induced by MAS and the radiofrequency pulses being taken into account. When temperatures were set to S-3 values ranging from 25 to 60 • C, the same procedure without the cooling box unit was used.
Polarisation Transfer (PT) 13 C MAS NMR spectra were acquired using a spectral width of 200 ppm and an acquisition time of 50 ms, under 67 kHz TPPM 1 H decoupling. S5 For each 13 C spectrum, 3200 scans were accumulated with a recycle delay of 4 s. 13 C chemical shifts were externally referenced to solid α-glycine at 43.67 ppm. S6 1 H and 13 C hard pulses were applied at a ω H/C 1 2π = 80 kHz, CP was performed with t CP = 1 ms, ω C 1 2π = 80 kHz and ω H/C 1 2π linearly ramped from 72 to 88 kHz, covering the ± ω R matching conditions, and INEPT spectra were recorded with the delay times τ = 1.8 ms and τ ' = 1.2 ms. The experimental time-domain data were processed as described previously S4 with a Matlab in-house code partially derived from matNMR, S7 using line broadening of 20 Hz, and zero-filling of 8192 time-domain points.
DMPS and DLPS peak assignments were made based on DMPC spectra, as described in previous studies. S8 31 P spectra were acquired using a spectral width of 200 ppm, an acquisition time of 50 ms, 2048 scans and a recycle delay of 4s. The rate of spinning in the MAS experiments was 5000Hz for the 13 C spectra and 1250Hz for 31 P spectra.

P MAS NMR -Data analysis
Isotropic chemical shift values were set using the resonance of the phosphate buffer (δ = 0 ppm). 31 P chemical shift anisotropy values (∆σ) were determined using Herzfeld-Berger sideband analysis S9 as implemented in matNMR. S7 The anisotropic part of the ∆σ is defined as: where σ ii are the principal tensor components. When the lipid bilayer is in the liquid crystalline phase, the chemical shielding tensor is averaged to an effective tensor that is axially S-4 symmetric. S10 The anisotropic part of this time-average tensor has been defined as: where σ i is the isotropic chemical shift, σ is the low intensity shoulder (σ = σ 33 ), and σ ⊥ is the high intensity shoulder (σ ⊥ = σ 11 = σ 22 ) of the axially symmetric powder pattern. S10

Differential Scanning Calorimetry
Sample preparation DMPS-induced α-synuclein proto-fibrils were prepared as described above in the "MAS NMR -Sample preparation" section and DMPS vesicles were prepared as described above in the "Protein and lipid preparation" section.

Data acquisition
All the samples were degassed before the acquisition of the DSC thermograms. The thermograms were acquired using a Microcal VP-DSC calorimeter (Malvern Instruments) with a scanning rate of 1 • C.min −1 from 5 to 65 • C. All of the DSC thermograms reported in this article were corrected by subtracting the thermogram of the phosphate buffer and correspond to the first scan, unless otherwise stated.

Proteinase-K treatment
The proteinase-K digested fibrils were prepared by incubating DMPS-induced α-synuclein proto-fibrils (prepared as described in "MAS NMR -Sample preparation" section) in the presence of 3 µM proteinase-K (Ambion, Germany) under quiescent conditions for 3 h. The mixture was then placed in the calorimeter for measurements.
S-5 The proto-fibrils were formed by incubating 100 µM α-synuclein in the presence of 2mM DMPS or DLPS dispersed as vesicles in phosphate buffer at pH 6.5 and 30 • C, and incubating this mixture for 4 d under quiescent conditions. S-7