N-Terminal Acetylation Affects α-Synuclein Fibril Polymorphism

Abstract

Parkinson’s disease etiology involves amyloid formation by α-synuclein (αSyn). In vivo, αSyn is constitutively acetylated at the α-amino N-terminus. Here, we find N-terminally acetylated αSyn (Ac-αSyn) aggregates more slowly than non-acetylated αSyn (NH₃-αSyn) with significantly reduced sensitivity to thioflavin T (ThT). Fibril differences were characterized by transmission electron microscopy, circular dichroism spectroscopy and limited-proteolysis. Interestingly, the low-ThT Ac-αSyn fibrils seed both acetylated and non-acetylated αSyn and faithfully propagate the low-ThT character through several generations, indicating a stable fibril polymorph. In contrast, the high-ThT NH₃-αSyn seeds lose fidelity over subsequent generations. Despite being outside of the amyloid core, the chemical nature of the N-terminus modulates αSyn aggregation and fibril polymorphism.

Graphical Abstract

Intracellular accumulation of α-synuclein (αSyn) amyloid fibrils is a pathological feature of Parkinson’s disease and other synucleinopathies.¹ Monomeric αSyn is intrinsically disordered and is thought to play a role in synaptic vesicle regulation and transmission through its membrane binding properties.² αSyn is characterized by three regions (Figure 1A): the KXKEGVXXXXX repeats, which form an amphipathic helix upon membrane binding,³ followed by the non-amyloid-β component (NAC) region, which is highly amyloidogenic,⁴ and the C-terminal acidic tail. In the fibrillar state, residues 37–97 fold into a β-strand-rich, Greek-key-like structure (Figure 1B), while the N- and C-terminal residues are localized to the fibril exterior, remaining unstructured.^5–7

Figure 1.

(A) Schematic representation of the αSyn primary sequence, highlighting the acetylated N-terminus, amphipathic repeat region (residues 1–60), NAC region (residues 61–95), and C-terminal tail (residues 96–140). Locations of acidic and basic residues are colored red and blue, respectively. (B) CryoEM structure of Ac-αSyn fibril (PDB code 6OSJ).⁵ Nʹ and Cʹ denote the N- and C-termini of one of the protomers. Residues (K58 and E61) forming an internal salt bridge are indicated.

Unique among the various post-translational modifications that αSyn can undergo,^8–10 acetylation of the α-amino N-terminus is constitutive.^{8, 11} This is a common modification in eukaryotic proteins that is co-translational and irreversible.^{12, 13} Several studies have investigated the effect of N-terminal acetylation on the conformational properties of αSyn. As expected, the acetyl-group stabilizes transient N-terminal helical structure.^14–18 Increased membrane vesicle binding also has been reported.^{14, 17–19} Recently, it was reported that N-terminally acetylated αSyn (Ac-αSyn) binds to proteoglycans on cellular membranes with higher affinity than non-acetylated αSyn (NH₃-αSyn), highlighting a biological impact.²⁰ In the literature, N-terminal acetylation is reported to have little to no effect on the properties of αSyn fibrils.^{15–17, 21–23} Here, we reexamine the effect of N-terminal acetylation on αSyn fibril formation by comparing aggregation kinetics, secondary structural characterization by circular dichroism and Raman spectroscopy, limited-proteolysis, and seeding experiments.

αSyn fibril formation was monitored by thioflavin T (ThT), an extrinsic fluorophore which increases in intensity upon amyloid binding,²⁴ and visualized by transmission electron microscopy (TEM, Figure 2A). Ac-αSyn exhibited a prolonged lag phase compared to NH₃-αSyn, consistent with prior reports of slower aggregation of Ac-αSyn.^{15–17, 23} While both proteins formed fibrils of similar widths (~15 nm), twisted fibrils were more easily identified in NH₃-αSyn, whereas Ac-αSyn fibrils appeared rod-like.

Figure 2.

Effect of N-terminal acetylation on αSyn aggregation. (A) Aggregation kinetics of NH₃-αSyn (blue) and Ac-αSyn (red) measured by ThT fluorescence ([αSyn] = 50 μM, [ThT] = 5 μM in 20 mM NaPi, 100 mM NaCl, pH 7.0, 37 °C). Average (line) and standard deviation (shaded) are shown (n = 4). TEM images of NH₃-αSyn (top) and Ac-αSyn (bottom). Scale bar is 200 nm. (B) Absorbance (solid) and emission (dashed) spectra of ThT-containing fibrils after aggregation.

Strikingly, a near ten-fold reduction in the final ThT fluorescence intensity was observed for Ac-αSyn relative to that of NH₃-αSyn. This trend was consistent with that of the pelleted fibrils (Figure 2B) and seen across multiple protein preparations at varying protein concentrations (Figure S1) and in different buffer conditions (Figure S2). Further, this high- vs. low-ThT behavior is reproduced by post-staining NH₃-αSyn and Ac-αSyn fibrils formed in the absence of ThT (Figure S3). Spectroscopic analysis of the pelleted material taken after 60 h found comparable amounts (~85%) of protein (assessed using absorbance at 280 nm) in the insoluble fraction (Figure 2B), indicating that the reduced ThT fluorescence is not because of decreased Ac-αSyn fibrillar material. While fewer bound ThT molecules (assessed using absorbance at 420 nm) were found for Ac-αSyn (1:11 ThT:Ac-αSyn) compared to NH₃-αSyn (1:3 ThT:NH₃-αSyn), it does not fully account for the order of magnitude loss in ThT signal. On a per molecule basis, ThT bound to Ac-αSyn fibrils has fluorescence intensity that is ~4-fold weaker than that of ThT bound to NH₃-αSyn fibrils. Together, these results suggest that Ac-αSyn forms a different fibril polymorph with distinct ThT binding site(s).

Because ThT is believed to act as a molecular rotor, where its emission is enhanced when the rotation about the C–C bond between the benzathiole and benzylamine rings becomes restricted,²⁵ we postulate that the low-ThT fibril polymorph formed by Ac-αSyn is attributed to a less hindered binding site(s) (i.e. more dynamics). Moreover, static quenching may be more prevalent (e.g. interactions with specific sidechains), contributing to the observed lower intensity.

Secondary structural differences between NH₃-αSyn and Ac-αSyn fibrils were investigated by circular dichroism (CD) and Raman spectroscopies. CD spectra showed that NH₃-αSyn and Ac-αSyn shared the expected single spectral minimum near 220 nm for β-sheet structure (Figure 3). The acetylated protein, however, has a less-negative mean residue ellipticity at 220 nm and more-negative mean residue ellipticity at shorter wavelengths, suggesting a higher proportion of random coil character. While overall similar, Raman spectra (Figure S4) also indicated less β-sheet content in Ac-αSyn fibrils due to a lower intensity of the amide-I band (1664 cm⁻¹). Our interpretation is that either there is a smaller β-sheet core or more disordered segments in the Ac-αSyn fibrils. In either case, it would corroborate that Ac-αSyn adopts a different fibril polymorph.

Figure 3.

Comparison of β-sheet content and limited-PK digestion of αSyn fibrils. CD spectra of NH₃-αSyn (blue) and Ac-αSyn (red) fibrils. Average (line) and standard deviation (shaded) are shown (from 3 independent microplates). Inset: SDS-PAGE of limited protease digestions of αSyn fibrils (40 μM) with decreasing PK (200 to 0.2 ng/mL for 16 h at 37 °C) visualized by Coomassie staining. N and A are abbreviations NH₃-αSyn and Ac-αSyn.

Limited proteolysis experiments were then performed using proteinase-K (PK), a broad spectrum protease widely used to evaluate amyloid structures and their protease resistance.²⁶ The highly ordered, compact structure of amyloid fibrils renders segments of the peptide backbone in the amyloid core inaccessible to proteases; differences in digestion rates or patterns therefore reflect the abundance or localization of disordered sequences that are susceptible to proteolysis. SDS-PAGE analysis clearly showed that Ac-αSyn fibrils are 26 ± 6% more proteolyzed than NH₃-αSyn at all concentrations examined, indicating greater PK-accessibility (Figure 3 inset). Despite its enhanced susceptibility to PK digestion, peptide mapping by liquid chromatography mass spectrometry showed that Ac-αSyn had similar PK-resistant cores (e.g. 18–125, 19–113 and 31–125) to NH₃-αSyn (Tables S1 and S2). This indicates that Ac-αSyn likely has a similar structured amyloid core but with more disordered regions than NH₃-αSyn. Since ThT binding is dependent on ordered β-strands, this could explain both reductions in ThT binding and fluorescence enhancement in Ac-αSyn fibrils.

To establish whether the low-ThT and high-ThT species represent distinct polymorphs dependent on N-terminal acetylation, seeding experiments were conducted. By introducing seeds of either the high- (NH₃-αSyn) or low-ThT (Ac-αSyn) fibrils, it is possible to evaluate whether templating and replication of fibril structure to the other protein occurs (cross-seeding). Second generation fibrils were prepared by seeding either Ac-αSyn (Figure 4A) or NH₃-αSyn (Figure 4B) with 10% fibrillar NH₃-αSyn or Ac-αSyn that were prepared in the absence of ThT. In all cases, the lag phase is abolished, indicating that both acetylated or non-acetylated αSyn can template on to either seeds. Moreover, seeding either monomer with fibrillar Ac-αSyn led to low-ThT intensity identical to that of the seeds. Similarly, seeding of either monomer with fibrillar NH₃-αSyn also reproduces the high ThT-fluorescence character of the seeds. Again, comparable amounts of fibrils were found post-aggregation, supporting that the observed ThT signals represent different fibril polymorphs. These results unequivocally demonstrate that the reduced ThT fluorescence is a structural characteristic of the initial Ac-αSyn fibrils which can be propagated in the absence of N-terminal acetylation.

Figure 4.

Self- and cross-seeding reactions. (A) Ac-αSyn and (B) NH₃-αSyn (n = 4) seeded with either 10% Ac-αSyn (red) or NH₃-αSyn (blue). Unseeded aggregation (black, 1^st) for (A) Ac-αSyn and (B) NH₃-αSyn are also shown for reference. First through fourth generation reactions are shown where 2^nd generation is seeded by 1^st generation, 3^rd generation is seeded by 2^nd generation and so on.

Interestingly, while the low-ThT structure was faithfully propagated by both Ac-αSyn and NH₃-αSyn over several generations (red curves, Figures 4A and 4B), seeding with high-ThT fibrils lacked fidelity, leading to a loss of ThT fluorescence with each successive generation (blue curves, Figures 4A and 4B). This implies that although Ac-αSyn is compatible with the high-ThT polymorph, the low ThT polymorph dominates in the end, and therefore, is a more stable structure for αSyn regardless of the acetylation state of the N-terminus. Another possibility is that multiple polymorphs exist in the first generation of aggregated NH₃-αSyn and that the low-ThT polymorph is a more efficient seed, resulting in a more stable fibril structure. Overall, TEM images of the seeded fibrils showed long, well-dispersed homogeneous fibrils with little discernible twist in all cases, similar to the 1^st generation Ac-αSyn fibrils (Figure S5). We verified that the final overall ThT intensity for each generation is unchanged over weeks; thus, it does not appear that these two forms are rearranging on this timescale.

Acetylation of the N-terminus apparently changes the energy landscape and guides Ac-αSyn to the more stable, low-ThT structure. Once formed, it serves as an efficient seed that can faithfully propagate this polymorph even in the absence of N-terminal acetylation. Curiously, the more stable, low-ThT Ac-αSyn structure contains more disordered structure and is more susceptible to proteolysis than the high-ThT polymorph. The high-ThT polymorph formed by NH₃-αSyn appears to represent a local minimum on the energy landscape of αSyn. This makes sense as NH₃-αSyn aggregates faster, so the fibrils may be kinetically trapped. Ac-αSyn bypasses this structure, possibly due to the increased helical structure in its N-terminal region altering intra- and intermolecular interactions. The impact of N-terminal acetylation on αSyn fibril structure is surprising in light of recent near-atomic resolution cryoEM structures,^5–7 none of which have suggested that the N-terminal region of the protein is directly involved in the structured amyloid core. Interestingly, while the acetylated and non-acetylated cryoEM structures are highly similar, there is one noted difference of a buried intramolecular salt bridge between K58 and E61 for Ac-αSyn (Figure 1B).^{5, 7} In the NH₃-αSyn structure, K58 is solvent-exposed, eliminating the salt bridge.⁶ We hypothesize that because ThT is a negatively charged molecule, an additional surface-accessible lysine would favor electrostatic interaction between ThT and the non-acetylated fibrils, which could explain the higher ThT binding stoichiometry for NH₃-αSyn.

In conclusion, this work offers new insights into the remarkable polymorphic nature of αSyn fibrils, which reinforces the pathological relevance of fibril polymorphism. In the complex milieu of a cell, subtle environmental changes can lead to the formation of different polymorphs associated with different disease phenotypes, which have been documented in the cases of amyloid-β and Tau, where specific fibril polymorphs are associated with Alzheimer’s, chronic traumatic encephalopathy, and frontotemporal dementia.^27–29 The observation that a small acetyl group at a single residue outside the amyloid core can impact αSyn fibril structure is surprising, and highlights the potential importance of post-translational modifications in modulating αSyn fibril polymorphism. Thus, we believe that the effect of N-terminal acetylation should not be overlooked, and moving forward, Ac-αSyn should be used as the standard for αSyn research.