New SNCA mutation and structures of α-synuclein filaments from juvenile-onset synucleinopathy

The case with JOS has been reported [37]. A 13-year-old individual developed rapidly progressive Parkinsonism, cognitive impairment, dysarthria and myoclonus, and died two years later. The number of α-synuclein deposits exceeded that of typical DLB.

Genomic DNA was extracted from frontal cortex. Coding exons of SNCA and flanking intronic sequences were amplified by polymerase chain reaction, screened by agarose gel electrophoresis and sequenced using the dideoxy method. Amplified exon 2 DNA from the JOS case was subcloned into pCR2.1 TA (Invitrogen). Recombinant plasmids were isolated from 48 clones and DNA from 5 of these clones was sequenced in both directions, as described [40].

Sarkosyl-insoluble material was extracted from frontal cortex of the individual with JOS, as described [38]. In brief, frozen tissues were thawed and homogenised in 20 vol (v/w) extraction buffer consisting of 10 mM Tris–HCl, pH 7.5, 0.8 M NaCl, 10% sucrose and 1 mM EGTA. Homogenates were brought to 2% sarkosyl and incubated for 30 min at 37 °C. Following a 10 min centrifugation at 10,000 g, the supernatants were spun at 100,000 g for 20 min. Pellets were resuspended in 500 µl/g extraction buffer and centrifuged at 3,000 g for 5 min. Supernatants were diluted three-fold in 50 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl, 10% sucrose and 0.2% sarkosyl, and spun at 166,000 g for 30 min. Sarkosyl-insoluble pellets were resuspended in 100 µl/g of 20 mM Tris–HCl, pH 7.4.

Immunogold negative-stain EM and immunoblotting were carried out as described [9]. Filaments were extracted from the frontal cortex of the individual with JOS. PER4, a rabbit polyclonal serum that was raised against a peptide corresponding to residues 116–131 of human α-synuclein [33], was used at 1:50. Images were acquired at 11,000 × with a Gatan Orius SC200B CCD detector on a Tecnai G2 Spirit at 120 kV. For immunoblotting, samples were resolved on 4–12% Bis–Tris gels (NuPage) and primary antibodies were diluted in PBS plus 0.1% Tween 20 and 5% non-fat dry milk. Before blocking, membranes were fixed with 1% paraformaldehyde for 30 min. Primary antibodies were: Syn303 [a mouse monoclonal antibody that recognises residues 1–5 of human α-synuclein [8]] (BioLegend) at 1:4000, Syn1 [a mouse monoclonal antibody that recognises residues 91–99 of human α-synuclein [25]] (BD Biosciences) at 1:4000 and PER4 at 1:4000. Histology and immunohistochemistry were carried out as described [31, 35]. Following deparaffinisation, the sections (8 µm) underwent heat-induced antigen retrieval in 10 mM citrate buffer, pH 6.0. The primary anti-α-synuclein antibodies were: Syn1 (1:250), 4B12 [a mouse monoclonal antibody that recognises residues 103–108 of human α-synuclein [14]] (BioLegend) at 1:250, MJFR14 [a rabbit monoclonal antibody that has been reported to recognise assembled human α-synuclein [17]] (Abcam) at 1:1000 and anti-pS129 [a rabbit monoclonal antibody specific for α-synuclein phosphorylated at S129] (Abcam) at 1:800. Anti-Iba1 (Antibodies.com) and anti-GFAP antibodies (Dako), which recognise microglia and astrocytes, respectively, were used at 1:200. For signal detection, we used ImmPress HRP anti-rabbit/mouse IgG and ImmPACT SG kits (Vector Laboratories) or Alexa fluorophore-conjugated secondary antibodies (Invitrogen) at 1:250. Luminescent conjugated oligothiophene pFTAA (pentameric form of formyl thiophene acetic acid) was used at 3 µM, as described [16]. To visualise inclusions, sections were also silver-impregnated according to Gallyas-Braak [3, 7]. Some sections were counterstained with nuclear fast red or haematoxylin.

Wild-type and mutant human α-synucleins were expressed and purified using modifications of a published protocol [20, 23]. Constructs were made via in-fusion PCR cloning using human α-synuclein in pRK172. Expression was carried out in E. coli BL21(DE3) cells that were induced with 1 mM IPTG at an OD₆₀₀ of 0.8 for 3–4 h and centrifuged at 2000 g at 4 °C for 20 min. They were resuspended in cold buffer (10 mM Tris–HCl, pH 7.4, 5 mM EDTA, 0.1 mM AEBSF, supplemented with cOmplete EDTA-free protease inhibitor cocktail) and lysed by sonication using a Sonics VCX-750 Vibra Cell Ultrasonic processor, followed by centrifugation at 20,000 g at 4 °C for 45 min. The pellets were discarded and the pH of the supernatants was lowered to 3.5 with HCl, stirred for 30 min at room temperature and centrifuged at 50,000 g at 4 °C for 1 h. The pH was then increased to 7.4 with NaOH and the supernatants loaded onto an ion exchange HiTrap Q HP column and eluted with a 0–1 M NaCl gradient. Fractions containing α-synuclein were precipitated with ammonium sulphate (0.3 g/ml) for 30 min at 4 °C and centrifuged at 20,000 g at 4 °C for 30 min. The pellets were resuspended in PBS and loaded onto a HiLoad 16/60 Superdex (GE Healthcare) column equilibrated in PBS. The purity of α-synuclein was analysed by SDS-PAGE and protein concentrations were determined spectrophotometrically using an extinction coefficient of 5600 M⁻¹ cm⁻¹. The protein-containing fractions were pooled and concentrated to 3–6 mg/ml. To separate wild-type from mutant α-synuclein, the samples were run on 16% Tricine gels (Invitrogen) and immunoblotted with Syn1. Before assembly, to remove aggregates, wild-type α-synuclein, the insertion mutant and their 1:1 mixture were centrifuged at 2000 g for 10 min at 4 °C and the supernatants used. For assembly, 130 µM recombinant α-synucleins were incubated in microplates (PerkinElmer, Viewplate-384F), with 200 rpm shaking (Fluostar Omega; BMGLabtech) for 1 min every 2 min in 30 µl phosphate-buffered saline (PBS) at 37 °C for 4 days. Prior to further processing, the presence of filaments was verified by negative-stain EM.

Extracted filaments and assembled fractions were centrifuged at 3000 g for 3 min and applied to glow-discharged holey carbon gold grids (Quantifoil Au R1.2/1.3, 300 mesh), which were glow-discharged with an Edwards (S150B) sputter coater at 30 mA for 30 s. Aliquots of 3 µl were applied to the grids and blotted for approximately 3–5 s with filter paper (Whatman) at 100% humidity and 4 °C, using a Vitrobot Mark IV (Thermo Fisher). Datasets were acquired on Titan Krios G2, G3 and G4 microscopes (Thermo Fisher Scientific) operated at 300 kV. Images of α-synuclein filaments extracted from frontal cortex of the JOS case were obtained using a Falcon-4i detector and a Selectris-X energy filter (Thermo Fisher Scientific) with a slit width of 10 eV to remove inelastically scattered electrons. Images of α-synuclein filaments assembled from recombinant proteins were obtained using a Falcon-4 detector without energy filter or a Gatan K3 detector in super-resolution counting mode, using a Bioquantum energy filter with a slid width of 20 eV. Images were recorded with a total dose of 40 electrons per Ų.

Movie frames were gain-corrected, aligned, dose-weighted and summed into a single micrograph using RELION’s own motion correction program [44]. Contrast transfer function (CTF) parameters were estimated using CTFFIND-4.1 [27]. Subsequent image-processing steps were performed using helical reconstruction methods in RELION [12]. Filaments were picked manually for the JOS dataset; auto-picking was used for the in vitro assembly datasets. Reference-free 2D classification was performed to select suitable segments for generating initial models and for further processing. Initial models were generated de novo from 2D class averages using relion helix inimodel2d [29]. Helical twist and rise were optimised during 3D autorefinement. Bayesian polishing and CTF refinement were used to further improve the resolution of reconstructions [45]. Final maps were sharpened using standard post-processing procedures in RELION and their resolutions calculated based on the Fourier shell correlation (FSC) between two independently refined half-maps at 0.143 [28]. Helical symmetry was imposed on the post-processed maps using the relion helix toolbox program.

Atomic models were built in Coot [4] using shared substructures of MSA Type I and Type II filaments (PDB:6XYO; PDB:6XYQ) as templates. Coordinate refinements were performed using Servalcat [41]. Final models were obtained using refinement of the asymmetric unit against the unsharpened half-maps with helical symmetry in Servalcat. Model statistics are given in Supplementary Tables 1 and 2.

PCR amplification of control exon 2 DNA of SNCA produces a fragment of 377 nucleotides. Gel electrophoresis of amplified exon 2 DNA from the JOS patient revealed the presence of two fragments, the wild-type band and an additional band of 398 nucleotides (Fig. 1a). DNA sequencing showed a 21-nucleotide duplication (TGGCTGCTGCTGAGAAAACCA) in one allele of SNCA (c.47_67dup). In α-synuclein, this translates into the insertion of 7 residues (MAAAEKT) after residue 22, resulting in a protein of 147 amino acids. The PCR products of exon 2 amplification from JOS were subcloned into pCR2.12 and DNA sequencing confirmed the presence of both wild-type and mutant alleles. The mutation probably occurred de novo, since the proband’s parents did not exhibit symptoms of Parkinsonism or cognitive impairment.

Recombinant wild-type (140 amino acids) and mutant (147 amino acids) α-synucleins ran as discrete bands of approximately 15 and 16 kDa on 16% Tricine gels. When mixed together, they gave two separate bands on immunoblots with Syn1. Immunoblots of sarkosyl-insoluble material from the frontal cortex of the JOS individual gave two separate bands of similar intensities that aligned with the bands of recombinant wild-type and mutant α-synucleins (Fig. 1b). Mass spectrometry of the same material confirmed the presence of wild-type and mutant α-synucleins (Figure S1).

In cerebral cortex and several subcortical nuclei of JOS, we previously reported the presence of abundant intracytoplasmic α-synuclein inclusions in nerve cell bodies and processes [37]. These findings were confirmed by immunostaining of frontal cortex with anti-α-synuclein antibodies Syn1, 4B12, MJFR14 and pS129 (Figure S2a). α-Synuclein inclusions were also present in a few cells with a smaller diameter, consistent with their presence in some glial cells. The α-synuclein inclusions of JOS were labelled by luminescent conjugated oligothiophene pFTAA (Figure S2b) and were weakly Gallyas-Braak silver-positive (Figure S2c). Triple-labelling immunofluorescence using anti α-synuclein antibody Syn1, as well as anti-Iba1 and anti-GFAP antibodies, showed prominent gliosis in areas with α-synuclein inclusions (Figure S2d). By immunoelectron microscopy of sarkosyl-insoluble material, abundant α-synuclein filaments with a diameter of 10 nm were present (Figure S3a). By immunoblotting with antibodies Syn303, Syn1 and PER4 (Figure S3b), high-molecular weight material, as well as full-length α-synuclein were the predominant species. Truncated α-synuclein was less abundant. None of the antibodies distinguished between wild-type and mutant α-synucleins.

The vast majority of α-synuclein filaments from the frontal cortex of the individual with JOS (83%) comprise a single protofilament. The structure of these filaments was determined at 2.0 Å resolution and revealed a new fold that we call the JOS fold. In the structure of the remaining α-synuclein filaments (17%), determined at 2.3 Å resolution, there are two identical protofilaments with the JOS fold that pack against each other with C2 symmetry (Figs. 2, 3). Both singlet and doublet filaments have a left-handed helical twist, as established from the densities for backbone oxygen atoms in the cryo-EM maps (Figure S4).

The JOS fold consists of a compact core, spanning residues 36–100 (in the numbering of wild-type α-synuclein), and two disconnected density islands (A and B). There is also a large non-proteinaceous density between the core and island A that may correspond to a bound cofactor and is similar in size and environment to the unidentified densities of Lewy and MSA folds [30, 42]. An additional density, next to the side chain of H50, may represent a post-translational modification. Other densities probably correspond to solvent molecules.

The core sequence is unaffected by the mutation and has virtually the same conformation for residues K43-Q99 as the corresponding segment of MSA protofilament IIB2 (PDB:6XYQ) [30]. The conformation of this segment, which is stabilised in part by a characteristic salt bridge between residues E46 and K80, is also similar to the conformations of wild-type α-synuclein filaments formed following seeded aggregation using MSA brain seeds (PDB:7NCK) [20] and filaments assembled in vitro from H50Q α-synuclein (PDB:6PEO) [1] or A53T α-synuclein (PDB:6LRQ) [36], with minor differences in the inversion of side chain orientations of residues K58 and T59 (Figure S5a).

In the core, most side chain densities are fully resolved in the 2.0 Å resolution map of JOS singlet filaments. In the islands, however, the side chain densities are not well resolved, precluding unambiguous sequence assignment. Since their main chain densities are fully resolved, we attribute the lack of distinct side chain densities to the islands consisting of a mixture of different sequences. These sequences probably come mainly from the amino- and/or carboxy-terminal regions of the same α-synuclein chain, but the incorporation of exogenous peptide fragments cannot be excluded.

In the structure of island A, there are at least three clues that point to which sequences may be incorporated. Island A makes a curved β-strand, the amino-terminal half of which is tightly packed against residues A53 and V55 of the core, whereas its carboxy-terminal half bends away opposite residue H50 and faces the non-proteinaceous density, separating it from residues K43 and K45. Its interface with the core resembles part of the protofilament interface of MSA filaments between the N-terminal arm of protofilaments A and the C-terminal body of protofilaments B. The MSA protofilament interface harbours a non-proteinaceous density that is surrounded by four lysine residues in approximately the same location, suggesting that the residues in island A that face a similar density in the JOS fold are also lysines. A residue with a small side chain, which is packed between A53 and V55, is likely alanine. Side chain densities for Cβ atoms suggest that island A comprises ten consecutive non-glycine residues. Only two α-synuclein sequences satisfy all three clues: ₁₄GVVAAAEKTKQG₂₅ in the 140-residue wild-type protein (Fig. 2c–d) and ₂₁KTMAAAEKTKQG₃₂ in the 147-residue insertion mutant protein (Fig. 2f–g).

There are insufficient clues to suggest which sequences the shorter island B may be made of. However, our assignment of island A with the insertion mutant leaves an N-terminal segment that is long enough to reach and fill up the island B density. A model of the JOS fold made only of mutant protein, in which island B is assigned the sequence ₆KGLSKAKE₁₃ and is connected to ₂₁KTMAAAEKTKQG₃₂ of island A, is shown in Figure S6. In this model, residues 13–20, which connect islands B and A, mimic the conformation of the N-terminal arm of MSA protofilaments A. The absence of density for these residues in the cryo-EM map indicates that they are not ordered in the JOS filaments. An alternative explanation for the density of island B is that it can be reached from the last ordered residue of the core (L100) and that it consists of sequences from the C-terminal region of α-synuclein.

The cryo-EM reconstructions of JOS doublet filaments show that island A is essential for dimerisation, with the islands from both protofilaments facing each other across the filament axis. The 2.3 Å resolution map of the doublet filament further supports our sequence assignments for island A. Interestingly, there are no direct contacts between the ordered residues of both protofilaments. Like in dimeric filaments of TMEM106B [31], an additional disconnected density of unknown identity resides on the symmetry axis and mediates filament interactions (Fig. 3). This non-proteinaceous density, which is co-ordinated by neutral polar residues Q24 (Q31 in mutant α-synuclein) from both protofilaments, is probably of a different chemical nature than the large non-proteinaceous densities within each protofilament. The density of island B, which is farthest from the doublet filament axis, is less resolved than in the singlet filament map, but is similar in overall shape.

The in vitro assembly of recombinant human wild-type α-synuclein, the 7-amino acid insertion mutant and a 1:1 mixture of both, yielded filaments that did not reproduce the structures of those from frontal cortex of the JOS case (Fig. 4).

Filaments of wild-type recombinant α-synuclein were of a single type (Fig. 4a). Solved at 3.5 Å resolution, their structures closely resemble known structures of wild-type recombinant α-synuclein filaments (PDB:6H6B; PDB:6CU7) [11, 18] and have minor differences with two others (PDB:6A6B; PDB:6OSJ) [19, 24]. Filaments made of the insertion mutant were also of a single type, but their structure at 3.6 Å resolution was different from that of wild-type α-synuclein filaments (Fig. 4a). This type resembled most closely one of the filaments from in vitro seeded aggregation of wild-type recombinant α-synuclein with seeds of MSA filaments (PDB 7NCG) [20] The same protofilament fold was first observed for filaments assembled from recombinant E46K α-synuclein (PDB:6UFR) [2]. This fold, which comprises residues G36-D98 of wild-type α-synuclein, lacks the N-terminal sequence with the JOS insertion.

Filaments made from a 1:1 mixture of wild-type and mutant human α-synucleins were of at least three Types. We solved the structures of two Types, with the third lacking the twist necessary for helical reconstruction (Fig. 4b, c; Figure S6). The structures were similar in appearance to those of recombinant wild-type α-synuclein and had a left-handed helical twist. The structure of the majority filament (Type 1, 57% of total) was solved to 2.7 Å resolution and revealed two similar, but non-identical, protofilaments labelled A and B. The ordered core of protofilament A spans residues L38-D98, with amino acids L38-G93 adopting almost the same fold as the corresponding segment in MSA protofilament IA (PDB:6XYO) [30] (Figure S5b). The ordered core of protofilament B extends from G36-D98. Both protofilaments differ at the turn K80-A85, with a reversal of inside/outside orientations of the side chains of T81 and V82, and a shift of the segment A85-D98 along the filament axis by one rung. It is not feasible to attribute the distinct protofilament folds to the differences between wild-type and mutant α-synucleins, because the ordered cores are composed of sequences present in both.

We also observed two non-proteinaceous densities in pseudo-symmetrical locations that were surrounded by the side chains of K43 and K45 from one protofilament and K58 and K60 from the other, and vice versa. They probably correspond to phosphate ions from the buffer and are similar in structure to the densities located between the JOS core and island A. JOS segment K43-E83 aligns well with the corresponding segment of protofilament B, resulting in the superposition of both structures, in which residues I88-K97 of JOS structurally align with residues A85-G93 of protofilament B. JOS island A structurally aligns with residues H50-Q62 of protofilament A, with K21 and K23 of the former superimposing on K58 and K60 of the latter (Figure S5c).

The structure of the minority filaments (Type 2, 17% of total) was solved to 2.8 Å resolution. Two identical protofilaments are linked by pseudo-2₁ symmetry. The ordered protofilament core spans residues G41-V95 and adopts a similar fold to that of protofilament A of Type 1 filament, with minor differences in the relative positions of their amino- and carboxy-terminal parts. Two putative phosphate densities were also present at the inter-protofilament interface (Fig. 4c).