Alpha-synuclein structure and Parkinson's disease - lessons and emerging principles

Abstract

Alpha-synuclein (αS) is the major constituent of Lewy bodies and a pathogenic hallmark of all synucleinopathathies, including Parkinson's disease (PD), dementia with Lewy bodies (DLB), and multiple system atrophy (MSA). All diseases are determined by αS aggregate deposition but can be separated into distinct pathological phenotypes and diagnostic criteria. Here we attempt to reinterpret the literature, particularly in terms of how αS structure may relate to pathology. We do so in the context of a rapidly evolving field, taking into account newly revealed structural information on both native and pathogenic forms of the αS protein, including recent solid state NMR and cryoEM fibril structures. We discuss how these new findings impact on current understanding of αS and PD, and where this information may direct the field.

Keywords: Alpha-synuclein; Amyloid; CryoEM; Oligomers; Parkinson’s disease; Protein-protein interactions.

Fig. 1

Change in Circular Dichroism (CD) signal in the far UV caused by the binding of αS to an excess of DMPS vesicles. This demonstrates a shift from a random coil structure in the absence of lipid vesicles (green), towards an alpha-helical secondary structure in the presence of DMPS lipid vesicles (red). Meade et. al. unpublished data reproducing data from Galvagnion et al. [34]

Fig. 2

Structure of a single layer within a mature αS fibril. Based on the CryoEM structure published by Li et al [26] displaying formation of the ‘Greek Key’ topology with rotational symmetry about the axis of the fibril. The early onset mutations (E46K, H50Q, G51D/E, A53T) are highlighted (pink) in addition to three key electrostatic interactions that are perturbed in early onset PD (K58-E61, E46-K80 and K45/H50-E57)

Fig. 3

CryoEM structures of four distinct types of full length αS fibril. The four structures are known as type 1a ‘rod’ [25, 26], type 1b ‘twister’ [27], type 2a and type 2b polymorphs [53]. Single layer density slices within the rod structure have revealed a Greek Key topology with rotational symmetry about the axis of the fibril. In contrast, single layers within the twister structure reveal a β-arch motif. Both type 1 polymorphs contain two protofilaments composed of stacked β-sheets with rotational symmetry about the fibril axis. In contrast, type 2 polymorphs lack the steric zipper geometry identified in type 1 polymorphs and are instead characterised by a hydrophobic cleft that is stabilised by intermolecular salt bridges and additional interactions between the NAC and the N-terminus. Left Box) Shown is the 3D model of the type 1a (rod) and type 1b (twister) fibril polymorphs, respectively, with their distinctively different helical pitches depicted. Top) Shown are representative regions of density maps of both polymorphs are superimposed with their models showing match of side chains with cryoEM densities. Bottom) How a 5 nm protofilament [24] may represent a shared fibril kernel from which both rod and twister fibrils can develop. In rod fibrils the interface is composed of residues within the preNAC region (blue, residues 47–56), an area in which most of the early onset PD mutations are located (cyan). In the twister fibrils the interface is composed of residues within the NAC core region (red, residues 68–78). This suggests that early onset mutations disfavour the rod like fibrils over the twister structures, offering the possibility for fibril morphogenesis and the potential to shift the aS population towards a more toxic polymorph. The left hand panel has been adapted from Li et al. 2018 [27] (CC-BY 4.0). The right hand panels are adapted from Guerrero-Ferreira et al 2019 [53] (CC-BY-NC-ND 4.0) and show schematic representations of all four currently characterized αS polymorphs with the with the N-terminus in blue, the NAC region in red and the C-terminus in yellow

Fig. 4

The KTKEGV imperfect repeats found within the αS structure. a Model of conformational transition proposed by Singh et al. [47] (CC BY-NC 4.0) of the transition of a 4-peptide bundle into amyloid fibrils, from an alpha-helix into a β-sheet fibril via aggregation induced stabilisation of anti-parallel 3₁₀-helix bundles. This model may be representative of the transitions which occur with aS from an alpha–helix membrane bound monomer to β-sheet fibril. b Structure of the micelle bound human aS, published by Ulmer et al., determined by solution NMR spectroscopy [99], highlighting the antiparallel α-helices of the membrane bound αS monomer, helix 1 spanning from Val [3]-Val [38] and helix 2 spanning from Lys [46]-Thr [93], connected by a well ordered linker. c The linear 140 residues of human aS arranged into KTKEGV imperfect repeats 1–9. Blue = basic; light blue = his; red = acidic; purple = polar uncharged; black = nonpolar. d Shown is a colour coded schematic with repeats 1–7 arranged into two 11/3 helix (3 turns over 11 residues), adapted from the αS helical wheels proposed by Dettmar 2018 [100] and Bendor et al. 2013 [101] representative of the membrane induced amphipathic helix. It has been proposed that lysine rich positions (blue) interact with negatively charged lipid head groups, while hydrophobic regions (black, grey area) interact with membrane lipids. Interestingly the Gly residues are found at the hydrophobic-water boundaries of the core, and are found on the adjacent helix face, which may be important in facilitating alpha to β switching at the water membrane, as previously seen in amyloid beta [102]. The position of single amino acid changes associated with early onset PD mutations might destabilise sidechain-sidechain packing that promotes formation of the helix and thereby accelerate the pathway toward amyloidosis. e Proposed structure of 2 × 3₁₀ helical wheel, formed by constriction of the α–helical domains seen in the micelle structure, clearly shows that the separation of the Lys and Glu residues in the aS amino acid sequence causes then to stack on top of each other stabilising the 3₁₀ intermediate, driving the energetic landscape towards the β-sheet fibril. Most interesting here is that the the first of the ‘ionic locks’ observed in the cryoEM structures is already formed in this structure, between K58-E61. In this proposed structure there does not appear to be a membrane binding domain. Potentially this structural change from α-helix to 3₁₀ intermediate could cause membrane disruption and mediate toxicity of αS