Conformational Ensemble Dynamics of Intrinsically Disordered Full-Length α- and β-Synuclein Monomers

Abstract

Abnormal aggregation of α-synuclein (αS) into amyloid fibrils is a hallmark of neurodegenerative diseases such as Parkinson’s disease. In contrast, its homolog β-synuclein (βS), co-localized at presynaptic terminals, resists amyloid formation and can even inhibit αS fibrillization. However, how sequence variations affect their structural dynamics remains poorly understood. To address this, we conducted 100 independent 1000-ns atomistic discrete molecular dynamics simulations for both αS and βS monomers. Our results revealed that both proteins predominantly adopted intrinsically disordered conformations, punctuated by transient helices and β-sheets. Both αS and βS exhibited a conserved helical tendency in the first half of the N-terminal domain, while the latter half showed dynamic β-sheet characteristics, with αS displaying greater abundance. Notably, the non-amyloid component (NAC) region in αS—critical for its aggregation—frequently adopted dynamic β-sheet structures, whereas the homologous region in βS displayed a greater tendency toward dynamic helices. Despite being largely disordered, the C-terminal regions transiently interacted with β-sheet–prone segments, potentially acting as dynamic caps that limit β-sheet growth in both proteins. Free energy landscape analysis indicated a clear enthalpy–entropy trade-off: structured conformations were stabilized by lower potential energy but penalized by reduced entropy, whereas disordered states, despite higher potential energy, were entropically favored. Importantly, potential energy reduction in αS was primarily associated with β-sheet formation, while in βS it was mainly driven by helix formation. These findings offer mechanistic insight into the distinct conformational landscapes of αS and βS and establish a thermodynamic framework for understanding how sequence differences modulate their structural properties and functional roles.

Graphical Abstract

1. Introduction.

The aberrant aggregation of α-synuclein (αS), a presynaptic protein classified as an intrinsically disordered protein (IDP), represents a key pathological hallmark of synucleinopathies, a group of neurodegenerative disorders that includes Parkinson’s disease (PD), dementia with Lewy bodies (DLB), and multiple system atrophy (MSA)^{1, 2}. In these conditions, misfolded αS assembles into highly ordered cross-β-sheet amyloid fibrils, which accumulate within neurons to form cytoplasmic inclusions such as Lewy bodies (LBs) and Lewy neurites (LNs)—the classical neuropathological features of PD^{3, 4}. In contrast, β-synuclein (βS), a closely related homolog sharing ~78% sequence identity with αS and co-localized at presynaptic terminals, is not found in LBs and does not spontaneously form fibrils under physiological conditions in vitro^{5, 6}. Although expressed at similar levels, βS has not been directly implicated in PD pathogenesis; rather, it has been shown to inhibit αS aggregation both in vitro and in vivo^{6, 7}. Altered expression patterns—elevated αS and reduced βS—have been associated with disease onset, while counter-regulating their levels suppresses αS aggregation in mammalian cell models, suggesting a protective role for βS as a potential endogenous inhibitor^{7, 8}. Nevertheless, despite their sequence homology, the structural basis underlying the markedly different aggregation behaviors of αS and βS remains poorly understood.

Both αS and βS are IDPs lacking stable tertiary structures, yet they can adopt partially dynamic helical and β-sheet conformations⁹. The 140-amino-acid sequence of αS is typically divided into three domains⁵: (i) an N-terminal region (residues 1–60) containing conserved KTKEGV repeats that form helices upon membrane binding¹⁰; (ii) a central hydrophobic domain (residues 61–95), known as the non-amyloid-β component (NAC) region, which is critical for aggregation due to its propensity to form β-sheet structures^{11, 12}; and (iii) a C-terminal domain (residues 96–140), enriched in acidic residues, which remains unstructured and contributes to solubility^{13, 14}. βS, comprising 134 amino acids, shares high sequence homology with αS—especially in the N-terminal domain—and is similarly divided into three regions: the N-terminal (1–60), a shorter NAC domain (61–84), and the C-terminal region (85–134)^{5, 15}. A key structural difference is the absence in βS of an 11-residue segment (⁷⁴VTAVAQKTVEG⁸⁴) within the NAC domain that contains a highly amyloidogenic motif in αS, significantly reducing βS’s aggregation propensity^{5, 16}. The NACore segment (⁶⁸GAVVTGVTAVA⁷⁸) of αS spontaneously forms fibrillar aggregates, while adjacent sequences such as ⁷⁹QKTVEGAGSIA⁸⁹ do not aggregate independently, indicating that fibrillization is primarily driven by the NACore¹⁷. Despite this, familial PD-linked point mutations (e.g., A30P, E46K, H50Q, G51D, and A53T) occur within the N-terminal domain and have been shown to alter αS fibrillization, promoting early disease onset^{1, 18, 19}. Furthermore, naturally occurring post-translational modifications outside the NACore region, such as phosphorylation and C-terminal truncation, are clinically associated with PD pathology^{20, 21}. These findings suggested that inter-domain interactions within the αS monomer play a pivotal role in modulating its aggregation behavior²². Therefore, elucidating the atomic-level structural differences between αS and βS is essential for understanding the molecular mechanisms underlying αS misfolding and aggregation, and for guiding rational drug design strategies aimed at early therapeutic intervention in PD.

The aggregation of αS is widely recognized to follow an irreversible, nucleation-dependent polymerization mechanism characterized by sigmoidal kinetics, including a lag phase, rapid elongation, and a final plateau^{21, 23}. Although mature fibrils were initially viewed as the main pathogenic species, accumulating evidence shows that intermediate oligomeric forms formed early in aggregation are more neurotoxic and may critically initiate neurodegeneration^{2, 24, 25}. Moreover, mature αS fibrils are biologically active, promoting disease progression by facilitating intercellular transmission of pathogenic seeds and driving the spatiotemporal spread of pathology throughout the brain²⁶. Both monomeric and oligomeric αS and βS are intrinsically disordered in solution, with high-resolution structures yet to be resolved. Experimental models reveal αS fibrils adopt an in-register parallel β-sheet structure spanning residues 36–95, while the N-terminal (residues 1–36) and C-terminal domains remain unstructured. Notably, truncation of either the N-terminal (e.g., residues 9–30) or C-terminal (e.g., residues 104–140) regions enhances αS fibrillization, suggesting these regions modulate aggregation^{14, 27–30}. Despite these findings, the conformational differences between amyloidogenic αS and non-amyloidogenic βS—especially in their monomeric forms—remain poorly understood, limiting insight into how sequence variations influence αS aggregation and toxicity, and how βS confers protection.

In this study, we systematically investigated the conformational ensembles and free energy landscapes of αS and βS monomers using 100 independent 1000-ns atomistic discrete molecular dynamics (DMD) simulations³¹. Both proteins predominantly occupied disordered conformations, interspersed with transient secondary structural elements such as helices and β-sheets. αS monomers exhibited a greater propensity for β-sheet formation, whereas βS favored helical structures. A conserved helical tendency was observed in the N-terminal residues 1–35 of both proteins; however, the latter half of the N-terminus adopted β-sheet–rich conformations more prominently in αS than in βS. Notably, the NAC region of αS transiently formed β-sheet structures, while the corresponding region in βS predominantly exhibited helical character, reflecting a sequence-dependent divergence in structural preferences. Although the C-terminal domains of both proteins remained largely disordered, they transiently interacted with β-sheet–prone regions—such as the NAC domain and residues 36–54—thereby capping these amyloidogenic segments and potentially suppressing further β-sheet elongation and amyloid formation. Free energy landscape analysis revealed that structured conformations were energetically favorable but entropically disfavored, whereas disordered states exhibited higher potential energy yet lower free energy, indicating their entropic advantage. In αS, β-sheet formation served as the primary driver of potential energy reduction, whereas in βS, helix formation played a more dominant role in lowering potential energy. Collectively, these findings provide mechanistic insights into the intrinsic structural dynamics governing the distinct aggregation propensities of αS and βS, and establish a thermodynamic framework for understanding how sequence variations modulate amyloidogenic behavior. Together, these findings reveal a sequence-encoded thermodynamic divergence between αS and βS, wherein the absence of the aggregation-prone NACore motif and the enhanced helical propensity of βS underlie its reduced amyloidogenicity and greater resistance to pathological aggregation.

2. Materials and Methods

2.1. Molecular Systems.

The amino acid sequence and initial structure of αS used in our simulations were obtained from the Protein Data Bank (PDB ID: 1XQ8³²). This structure, determined by solution NMR spectroscopy in a micelle-bound state (Figure 1a&b), features two helical regions spanning αS_3–37 and αS_45–92, followed by an extended and disordered C-terminal domain³². In contrast, fibrillar αS structures resolved experimentally exhibit a well-defined cross-β-sheet core primarily spanning residues 36–96, while the remaining regions are too conformationally dynamic to be resolved (Figure 1c) ^{14, 27–30}. The monomeric structure of βS remains unresolved experimentally, and the AlphaFold-predicted model (UniProt ID: Q16143) exhibits low confidence, particularly in the NAC regions. To avoid introducing structural bias, we initialized β-synuclein in a fully extended conformation for simulation.

Figure 1. Comparison of the primary sequences of αS and βS, and structural features of αS.

(a) Sequence alignment of αS and βS. Hydrophobic, positively charged, negatively charged, and polar residues are colored black, blue, red, and green, respectively. Asterisks (*) indicate identical residues, while colons (:) and periods (.) denote conserved and semi-conserved substitutions, respectively. The N-terminal domain is highlighted with light blue and cyan backgrounds, covering residues 1–35 and 36–60, respectively. The NAC domain is shaded in pink, with the NACore region specifically marked in red. The C-terminal domain is highlighted in violet. (b) The initial structure of αS used in our simulations was obtained from the Protein Data Bank (PDB ID: 1XQ8). (c) Three experimentally characterized αS fibril structures (PDB IDs: 2N0A, 6CU7, and 6OSL) all share a common in-register parallel β-sheet architecture spanning residues 36–95, while the remaining regions are presumed to adopt dynamically disordered conformations.

A sequence alignment between αS and βS (Figure 1a) shows high conservation in the N-terminal domain (~90%), but substantially lower identity in the NAC region (~45%) and the C-terminal domain (~42%). Notably, βS lacks most of the amyloidogenic NACore segment present in αS. However, other residues implicated in the formation of the cross-β-sheet fibril core—especially those in the latter half of the N-terminal region and the whole NAC domain^{11, 13, 33}—are largely conserved. To investigate how sequence variations influence structural dynamics, we performed folding simulations for both αS and βS monomers. For each protein, 100 independent DMD simulations were carried out, each spanning 1000 ns and initialized with distinct random velocity distributions. This multiple long-timescale simulation strategy ensures adequate conformational sampling and reduces potential biases introduced by initial conditions.

2.2. DMD Simulations.

Simulations were conducted at 300 K using an all-atom DMD approach^{31, 34} integrated with the Medusa force field^{35, 36}. Although DMD and traditional molecular dynamics (MD) share the same foundational physics, DMD differentiates itself by approximating interatomic potentials with discrete step functions, producing a collision-driven system where atomic velocities change abruptly as energy thresholds are crossed³⁷. When the discretization becomes finer, the all-atom DMD method approaches the continuous potential framework of classical MD. The Medusa force field accurately represents both bonded interactions—including covalent bonds, bond angles, and dihedral angles—and nonbonded interactions such as van der Waals forces, implicit solvation, hydrogen bonding, and electrostatics, similar to established MD force fields^{36, 38}. Solvation contributions were computed using the Lazaridis–Karplus effective energy function (EEF1)³⁹, while hydrogen bonds were modeled explicitly through a reaction-like mechanism³⁸. Electrostatic forces between charged groups were estimated using the Debye–Hückel model, employing a Debye screening length of roughly 10 Å to simulate physiological ionic conditions. The simulation units were defined as 1 Dalton for mass, approximately 50 femtoseconds for time, 1 angstrom for length, and 1 kilocalorie per mole for energy. The DMD software utilized is available to the academic community via Molecules In Action, LLC (www.moleculesinaction.com).

The combination of DMD with the Medusa force field and the EEF1 implicit solvent model has proven highly effective in accurately modeling native protein folding, capturing intricate secondary and tertiary structural details³⁸. This approach generates conformational ensembles that show strong agreement with experimental single-molecule FRET measurements^{40, 41}. Additionally, it has been successfully applied to investigate amyloid formation in calcitonin^42–44 and amylin family proteins^45–47, reliably replicating their experimentally observed amyloidogenic propensities and demonstrating close consistency with NMR structural data⁴⁸. Benchmarking against conventional MD simulations employing force fields such as GROMOS96, OPLS-AA, AMBER99SB-ILDN, and CHARMM36m has confirmed the robustness and enhanced sampling efficiency of the DMD methodology^{47, 49}. Given its high sampling efficiency and proven success in studying protein folding and aggregation—including full-length αS, its amyloidogenic segments, and other amyloid proteins^{12, 13, 20, 50–52}—by our group and others, we employed DMD simulations to investigate the folding dynamics of αS and βS monomers.

2.3. Analysis Methods.

Secondary structure elements were assigned using the Dictionary of Secondary Structure of Proteins (DSSP) algorithm⁵³. Hydrogen bonds were defined by an N···O distance cutoff of 3.5 Å and an N–H···O angle exceeding 150°⁵⁴. Residue contacts were considered formed when non-sequential heavy atoms in side chains or main chains were within 0.65 nm⁵⁵. To capture the conformational heterogeneity of αS and βS monomers, clustering was performed using the Daura method⁵⁶ with a Ca root-mean-square deviation (RMSD) cutoff of 0.8 nm, grouping conformations with pairwise RMSD below this threshold into clusters representing distinct structural states. The conformational free energy landscape was mapped by constructing a two-dimensional potential of mean force (PMF) via F(x,y)=−k_BTlnP(x,y)+C, where F(x,y) is the relative free energy at coordinates x and y, P(x,y) the probability density of observing conformations at those coordinates, k_B the Boltzmann constant, T the temperature, and C an arbitrary constant setting the global minimum to zero energy. Free energies are therefore relative to the most populated state⁵⁷. Furthermore, the local Cα atom density, ρ(r), analogous to a radial distribution function but without normalization to bulk density, was computed around the peptide’s geometric center using ρ(r)=N_r,r+dr/(4πr² dr), where N_r,r+dr counts atoms within a spherical shell of thickness dr at distance r, based on all residue Cα positions²⁰.

3. Results and Discussion

3.1. Equilibrium and convergence assessments for the monomeric simulation of both αS and βS.

The simulation performance was evaluated by monitoring the time evolution of key structural metrics, including the radius of gyration (Rg), the total number of main-chain hydrogen bonds and heavy atom contacts, and the average probabilities of secondary structure elements (i.e., unstructured, helix, and β-sheet) for both αS and βS monomers (Figures S1 and S2). For each system, three trajectories were randomly selected from a total of 100 independent DMD simulations. The absence of long-term trends and the presence of dynamic fluctuations in hydrogen bonding, heavy atom contacts, and secondary structure content suggested that the simulations were not kinetically trapped and that sufficient conformational sampling was achieved. Steady-state behavior was observed during the final 500 ns of the trajectories, where all structural parameters fluctuated around their respective equilibrium values (Figures S1a–c and S2a–c). To further assess convergence, we compared the probability distributions of the above parameters between two time intervals—500–750 ns and 750–1000 ns—using data from all 100 independent simulations. As shown in Figures S1d and S2d, the distributions were highly consistent across the two intervals for both αS and βS monomers, confirming that the simulations reached convergence and the results are statistically robust.

3.2. Both αS and βS monomers exhibited intrinsic structural dynamics, with βS favoring helical structures and αS displaying a slightly greater β-sheet propensity.

The folding dynamics of αS and βS monomers were analyzed by tracking the time evolution of residue-level secondary structure (Figure 2a&b). Both αS and βS displayed highly dynamic structural behavior, lacking stable conformations and frequently transitioning among structural states. Notably, their structural preferences varied across different sequence regions. Specifically, the N-terminal segment of αS (residues 1–35) exhibited a strong tendency to transiently adopt helical conformations, whereas residues 36–60 (the second half of the N-terminal domain) and the NAC region showed a pronounced preference for forming β-sheet structures (Figure 2a). In contrast, the C-terminal domain of αS largely remained unstructured, characterized by random coils and bends, with only fleeting occurrences of helical or β-sheet elements. The βS monomers (Figure 2b) were dominated by dynamic helical structures across both the N-terminal domain (residues 1–60) and the NAC region (residues 61–84). Although short-lived β-sheet structures occasionally appeared, particularly within residues 36–60. Similar to αS, the C-terminal domain of βS exhibited predominantly unstructured conformations, with only transient and minor populations of helical or β-sheet structures.

Figure 2. Folding dynamics analysis of αS and βS.

The folding dynamics of monomeric αS (a) and βS (b) are monitored through the time evolution of the secondary structure at the residue level. Two representative trajectories are randomly selected from a pool of 100 independent DMD trajectories. The snapshots are shown at 200 ns intervals. For clarity, the N-terminal domain is shaded in light blue and cyan, the NAC domain in pink with the NACore highlighted in red, and the C-terminal domain in violet. The N-terminal Cα atom is marked by a spherical bead for orientation. Probability distributions of unstructured, β-sheet, and helical fractions for each αS and βS monomer obtained from equilibrated simulations (c). The horizontal axis denotes the fraction of a specific secondary structural type within an individual conformation. Ensemble-averaged secondary structure contents of unstructured, β-sheet, and helix elements for αS and βS monomers (d). The analysis uses the last 500 ns of equilibrated data from 100 independent 1000 ns DMD simulations.

The probability distributions of unstructured, β-sheet, and helical content in the equilibrium states of αS and βS monomers revealed that both peptides predominantly adopted unstructured conformations. However, αS showed a higher propensity for β-sheet formation, whereas βS exhibited a greater tendency toward helical structures (Figure 2c). The ensemble-averaged secondary structure content further quantified these differences: αS consisted of 48.4% unstructured, 23.8% β-sheet, and 18.4% helix, while βS showed 50.4% unstructured, 10.5% β-sheet, and 30.0% helix content (Figure 2d).

3.3. Dynamic β-sheets in αS extended across residues 36–95, aligning with its fibrillar cross-β core, while in βS they were only restricted to 36–54.

The domain-specific structural dynamics of αS and βS monomers were further characterized by evaluating the average secondary structure propensity of each residue (Figure 3 and Figures S3–S4). In the N-terminal domains, both αS and βS exhibited mixed secondary structures, with notable helix (αS: 27.5%, βS: 34.0%) and β-sheet (αS: 24.8%, βS: 13.6%) content (Figure 3a). In contrast, the NAC region of αS predominantly adopted β-sheet structures (37.8%) with minimal helical content (6.1%), while the NAC region of βS showed the opposite trend—favoring helices (33.9%) over β-sheets (10.1%). Although unstructured conformations were distributed across all domains, they were especially abundant in the C-terminal region (αS: 63.1%, βS: 62.3%), and remained prevalent in the N-terminal (αS: 38.8%, βS: 43.5%) and NAC (αS: 46.6%, βS: 42.5%) domains as well (Figure S3a).

Figure 3. Secondary structure and conformational analysis of αS and βS monomers.

(a) The average helix and β-sheet content in the N-terminal, NAC, and C-terminal domains of αS and βS monomers. (b–d) Residue-level propensities for helix and β-sheet conformations within the N-terminal (b), NAC (c), and C-terminal (d) domains. (e–f) The central structures and corresponding probabilities of the nine most populated conformational clusters for αS (e) and βS (f) monomers. Due to the predominantly unstructured nature of the C-terminus, conformational clustering is limited to the N-terminal and NAC domains. All analyses use the final 500 ns of equilibrated trajectories from 100 independent 1000 ns DMD simulations.

Residue-level analysis of secondary structure propensities within the N-terminal domain revealed that both αS_1–35 and βS_1–35 exhibited similarly high helical tendencies (Figure 3b). However, notable differences were observed in the 36–60 region: αS_36–60 consistently favored β-sheet formation, while βS_36–54 also displayed β-sheet propensity, though to a lesser extent, and βS_55–60 preferentially adopted helical structures (Figure 3b). Despite the high sequence homology (~90%) across the N-terminal domain, the divergent structural preferences within residues 36–60 suggest that subtle sequence variations near the NAC-domain boundary influence inter-domain interactions and contribute to distinct conformational behaviors.

Within the NAC domain, αS predominantly adopted β-sheet structures, with the strongest propensity localized to the NACore segment (average β-sheet propensity reaching ~54.5%), while exhibiting minimal helical character (Figure 3c). In contrast, βS showed limited β-sheet propensity across the NAC region. Notably, residues 56–66 in βS, spanning the NAC head and adjacent N-terminal tail, demonstrated strong helical tendencies, with propensities ranging 60%–96% (Figure 3c). In the C-terminal domain, αS primarily assumed unstructured conformations, interspersed with sporadic regions of weak helix or β-sheet propensity (Figures 3d and S3). βS, by comparison, exhibited a pronounced helical region within residues 95–104 (~62%–97%), while the remainder of the C-terminal region remained largely disordered, with only limited secondary structure features (Figures 3d and S3).

The dynamic β-sheets observed in our simulations were primarily located in the second half of the N-terminal domain (αS_36–60) and the NAC region of the αS monomer. Similar β-sheet regions were also reported in previous classical simulations of αS monomers¹⁹ and in our recent DMD simulations of αS dimerization¹³. Furthermore, standard MD simulations of polymorphic αS dimers, derived from experimentally determined fibril structures, revealed comparable β-sheet segments⁵⁸. Consistently, biophysical and biochemical studies identified β-sheet formation in αS fibrils and unfolded states around residues αS38–43, αS48–65, αS70–79, and αS86–9714, 27–30, 59, 60. Overall, the monomer structures sampled in our simulations exhibited secondary structure patterns resembling those found in fibrillar αS, suggesting that similar interactions which stabilized mature fibrils also contributed to the stabilization of early oligomeric intermediates.

Structural cluster analysis was performed for both αS and βS monomers using a Cα root-mean-square deviation (RMSD) cutoff of 0.80 nm, focusing on the N-terminal and NAC domains. The highly dynamic nature of the C-terminal residues—where the average unstructured propensity significantly exceeded the peptide-wide average (Figure S4)—led to the exclusion of this region to ensure meaningful structural classification. The top nine most populated conformations for both αS and βS exhibited considerable structural diversity and low population ratios, consistent with their intrinsically disordered and dynamic nature (Figure 3e&f). In agreement with the secondary structure analysis, the dominant αS conformations featured β-sheet structures primarily in the second half of the N-terminal domain and the NAC region, flanked by helical motifs in the N-terminal head and an unstructured C-terminal tail (Figure 3e). By contrast, βS conformations exhibited only short β-sheets restricted to the second half of the N-terminal domain, with minimal involvement of the NAC region (Figure 3f). Instead, the first half of the N-terminus, the NAC region, and part of the C-terminal region of βS displayed a higher helical propensity, reflecting a distinct structural preference compared to αS.

3.4. αS monomers maintained low free energy by compensating entropy loss with decreased potential energy from β-sheet formation, whereas βS monomers achieved a similar balance through helix formation.

The ensemble conformational free energy landscapes of both αS and βS monomers were constructed based on the total potential energy plotted against the ratios of secondary structure content (unstructured, β-sheet, and helix), using the last 500 ns of data from 100 independent 1000-ns DMD trajectories (Figure 4). The αS monomers exhibited a narrow yet extended free energy basin associated with unstructured content ranging from 30% to 60% (Figure 4a). Within this basin, αS conformations with a high proportion of unstructured content (>47%) displayed elevated potential energy but still maintained low free energy (snapshots 2 and 3, Figure 4a), indicating an entropically favored ensemble. As the fraction of unstructured conformations decreased (47%–30%), corresponding to the emergence of more structured states, the overall free energy remained low due to a compensatory decrease in potential energy, effectively offsetting the entropic penalty (snapshots 4 and 5, Figure 4a). The correlation between conformational free energy and secondary structure content suggested that the reduction in potential energy in these more structured states primarily originated from β-sheet formation. A clear trend was observed in which increased β-sheet content was associated with lower potential energy, whereas changes in helical content did not exhibit a consistent relationship with potential energy. States characterized by either a very high unstructured content (>60%) or an excessively high β-sheet content (>45%) showed elevated free energy levels (snapshots 1 and 6, Figure 4a). In the former, high free energy resulted from excessive potential energy, while in the latter, it was attributed to insufficient entropy.

Figure 4. Conformational Free Energy Landscape Analysis of αS and βS Monomers.

(a–b) The conformational free energy landscapes of αS (a) and βS (b) monomers are presented as functions of potential energy versus the average fraction of unstructured (left panels), β-sheet (middle panels), and helix (right panels) conformations. Six representative conformational states are labeled 1–6. Structural snapshots corresponding to these states are shown below, selected based on the associated conformational parameters. Only the final 500 ns of simulation data from 100 independent 1000 ns DMD trajectories were used for this analysis. For visual clarity, the N-terminal Cα atom is highlighted with a bead.

Similar to αS, the ensemble free energy landscape of the βS monomer also revealed a dominant population of unstructured conformations (unstructured content >45%) that exhibited high potential energy but remained within a low free energy basin (snapshots 2 and 3, Figure 4b), primarily due to entropic stabilization. As the proportion of unstructured content decreased below 45%, indicating the emergence of more ordered secondary structures, the conformational states demonstrated a reduction in potential energy while maintaining low free energy levels (snapshots 4 and 5, Figure 4b). In contrast to αS, where the potential energy decrease was mainly associated with β-sheet formation, βS showed a distinct pattern. Both increasing helical and β-sheet content in βS correlated with a reduction in potential energy, but this effect was predominantly driven by helix formation, given the inherently low β-sheet propensity in βS. Conformational states with excessively high unstructured content (>60%) or elevated helical content (>42%) exhibited high free energy values (snapshots 1 and 6, Figure 4b), attributable to either high potential energy (in the case of disordered states) or a significant loss of entropy (in the case of highly helical conformations).

Overall, the conformational free energy landscapes of both αS and βS monomers reflected a dynamic equilibrium governed by the interplay between structural potential energy and conformational entropy (Figure 4). Highly ordered states benefited from low potential energy but incurred an entropic penalty, whereas unstructured conformations were entropically favored despite higher potential energy costs. The key difference between the two proteins lay in the structural contributions to potential energy reduction: β-sheet formation played a dominant role in αS, while helix formation contributed more substantially to energy minimization in βS.

3.5. The capping of the NACore by the remaining NAC region and the second half of the N-terminus promoted β-sheet-rich structures in the αS monomer, whereas the βS monomer mostly adopted helix due to the absence of the NACore.

To elucidate the key residue-level interactions supporting the dynamic structural conformations of αS and βS monomers, we analyzed the residue–residue contact frequency maps (Figure 5). In αS, consistently high intra-chain contact frequencies along the diagonal within residues 1–35 (mostly exceeding 40%) indicated stable local interactions, characteristic of helical structures (snapshot 1, Figure 5a). This observation was consistent with the residue-level secondary structure propensities (Figure 3b) and prior experimental evidence⁶¹. In contrast, residues 36–60 of αS exhibited a distinct β-hairpin-like contact pattern, with prominent off-diagonal contacts between residues 37–43 and 48–54, suggesting transient β-sheet formation in this region (snapshot 2, Figure 5a). Notably, a similar β-hairpin spanning residues 37–54 was previously identified by NMR and reported to be stabilized by the β-wrap protein AS69³³.

Figure 5. Residue-Level Contact Frequency Mapping of αS and βS Monomers.

(a–b) Heatmaps of residue-residue contact frequency for αS (a) and βS (b) monomers are generated by analyzing interactions involving both backbone and side-chain atoms, based on the final 500 ns of 100 independent DMD trajectories after equilibrium is reached. Six representative structural motifs (labeled 1–6), corresponding predominantly to α-helical or β-sheet arrangements, are identified as regions with high-frequency intramolecular contacts and are highlighted on the maps. These motifs are selected as exemplars of similar contact-rich regions. For spatial orientation, the N-terminal Cα atom is marked with a bead. Side chains of residues involved in these motifs are color-coded by physicochemical properties: hydrophobic (white), polar (green), positively charged (blue), and negatively charged (red).

The NACore segment played a pivotal role in promoting β-sheet formation within αS monomers by engaging in intramolecular interactions either with the C-terminal portion of the N-terminal domain (snapshot 3, Figure 5a) or with adjacent regions of the NAC domain itself (snapshot 4, Figure 5a). These interactions were predominantly driven by hydrophobic contacts, which acted to cap the elongation edges of the NACore β-sheet, thereby facilitating the extension of β-structure beyond the core segment. Consistent with this, the C-terminal residues of αS were largely disordered, but occasionally adopted transient structural motifs (Figures 2 and 3). Intermittent short helical segments were observed (snapshot 5, Figure 5a), aligning with prior NMR Cα chemical shift data⁶². Additionally, weak β-sheet-like contact patterns involving the C-terminal region were detected (snapshot 6, Figure 5a), typically forming capping interactions with β-sheet-prone regions in the N-terminal or NAC domains. However, these contacts exhibited low β-sheet propensities (<10%), indicating their transient and structurally unstable nature.

The β-sheet structures observed in residues 36–60 and surrounding NAC regions, mediated by NACore interactions, were consistent with earlier computational simulations and experimental evidence, including intramolecular paramagnetic relaxation enhancement (PRE) measurements⁶³. For instance, β-sheet formations within the NAC region were also identified in αS monomer ensembles derived from short-distance cross-linking–guided DMD simulations⁵⁰. Our recent computational study further demonstrated that among the seven imperfect 11-residue repeats (XKTKEGVXXXX motif) within αS, only the C-terminal portion of the third repeat (residues 36–41) and the seventh repeat (NACore, residues 68–78) could independently assemble into β-sheet aggregates¹⁷. The remaining repeats predominantly remained monomeric and structurally disordered. Collectively, these findings reinforce the central role of NACore as a nucleating element that facilitates and stabilizes β-sheet formation through edge-capping interactions, not only in αS monomers but potentially in oligomeric and fibrillar species as well. Furthermore, familial PD-associated mutations located in the N-terminal tail region (e.g., E46K, H50Q, G51D, A53T, A53E) may alter αS aggregation pathways by disrupting critical interactions between the N-terminal β-sheet-prone segment and NACore^{1, 18, 19, 64}. However, the precise molecular mechanisms underlying these pathogenic effects remain to be fully characterized.

In contrast to αS monomers, βS monomers predominantly exhibited residue–residue contact patterns indicative of helical structures across the entire sequence. High intra-chain contact frequencies consistent with helical arrangements were observed in several regions, including the N-terminal residues 1–35 (snapshot 1, Figure 5b), residues 52–66 and 78–88 involving the tail of the N-terminus and the head of the NAC region (snapshot 3, Figure 5b), as well as residues 95–104 and 124–134 in the C-terminal domain (snapshots 4 and 5, Figure 5b). In addition to these helical motifs, the second half of the N-terminal domain also displayed a short β-hairpin contact pattern formed by residues 37–52 (snapshot 2, Figure 5b). This motif closely resembled the β-hairpin observed in αS within the homologous region (residues 37–52), reflecting a conserved structural tendency governed by sequence similarity. This observation was further supported by earlier reports that the αS_35–56 segment independently formed a similar β-hairpin structure³³.

However, due to the absence of the NACore region in βS, the interaction between residues 37–52 and the NAC domain was significantly weaker than that observed in αS monomers. Additionally, occasional capping interactions between the C-terminal residues and the β-sheet growth edge of the β-hairpin were observed (snapshot 6, Figure 5b), although their contact frequencies and stability remained low. Taken together, the residue pairwise contact analysis indicated that the increased helical content and reduced β-sheet formation in βS monomers likely resulted from the lack of a β-sheet-facilitating NACore segment. This structural difference highlighted the sequence-dependent divergence in secondary structure preferences between αS and βS.

3.6. By wrapping around β-sheet–prone regions, the highly negatively charged C-terminal domain may suppress the driving forces for β-sheet elongation, thereby acting as a structural shield that limits aggregation in both αS and βS.

Mounting evidence has indicated that physiological C-terminal truncation enhances the liquid–liquid phase separation (LLPS), amyloid aggregation, and cytotoxicity of αS^{20, 21}. To further investigate the role of the C-terminal domain, its interactions with the N-terminal and NAC regions of both αS and βS were analyzed using residue-pairwise interaction frequencies based on mainchain atoms (Figure 6). The interaction propensity was plotted in the range of 0–10%, as the contact frequencies were too weak to be visualized within the broader 0–50% scale used in Figure 5.

Figure 6. Residue-level contact mapping between the C-terminus and the N-terminus/NAC domain in αS and βS monomers.

(a–b) Heatmaps showing pairwise contact frequencies based on main-chain atom interactions between C-terminal residues and those within the N-terminal and NAC domains of αS (a) and βS (b) monomers. The top panels display the average number of contacts (#NC) formed by each residue in the N-terminal and NAC regions with the C-terminus. Highlighted on the maps are representative β-sheet motifs where the C-terminus caps β-sheet–prone regions situated in the latter half of the N-terminus and the NAC core. Corresponding structural snapshots, selected based on similar contact patterns, are also presented. The N-terminal Cα atom is marked with a bead for spatial orientation. Residues involved in these interactions are color-coded according to their physicochemical properties: hydrophobic (white), polar (green), positively charged (blue), and negatively charged (red).

In αS monomer, the N-terminal residues 36–54 and the NACore region exhibited the strongest β-sheet propensities (Figure 6a) and also showed the highest frequency of interactions with C-terminal residues, particularly αS_109–120 and αS_131–137 (Figure S5a). The pairwise contact map revealed that these highly negatively charged C-terminal segments could directly cap the edge of β-sheet structures formed by residues 36–54 and the NACore region (snapshots in Figure 6a). Notably, both regions are known to play critical roles in driving αS aggregation, whereas the C-terminal domain itself lacks intrinsic aggregation propensity. Solvent-accessible surface area (SASA) analysis further indicated that the β-sheet–prone regions were among the most buried segments of the protein, while the C-terminal residues remained more solvent-exposed (Figure 7a). This trend was corroborated by radial distribution function (RDF) analyses of Cα atoms, which confirmed that aggregation-prone segments tended to localize toward the protein core, while amyloid-resistant regions were positioned near the surface (Figure 7c). Collectively, these results suggest that the dynamic wrapping of the highly charged, aggregation-resistant C-terminal domain around amyloid-prone regions serves as a transient structural shield, limiting spontaneous self-assembly of αS under physiological conditions. This mechanistic insight may explain why C-terminal truncation enhanced αS LLPS and aggregation propensity.

Figure 7. Solvent-accessible surface area and radial distribution function analysis of residues in αS and βS monomers.

(a–b) Per-residue average solvent-accessible surface area (SASA) for αS (a) and βS (b) monomers. (c–d) Radial distribution functions of local Cα atom density with respect to the geometric center of the peptide, calculated separately for the first and second halves of the N-terminal domain, the NAC region, and the C-terminal region in αS (c) and βS (d) monomers.

The βS monomer contained only one β-sheet–rich region spanning residues 36–54, which exhibited the strongest residue-pairwise contacts with the C-terminal domain, particularly around βS_102–114 (Figures 6b and S5b). Contact frequency analysis indicated that this β-sheet–prone segment was capped by the highly charged βS_102–114 region, leading to a reduction in its SASA and likely impeding further β-sheet elongation. Although βS shared six of the seven imperfect 11-residue repeats present in αS, it lacked the seventh and most amyloidogenic NACore repeat. Previous studies reported that only the third repeat in βS showed moderate aggregation potential, whereas the remaining repeats were largely aggregation-incompetent¹⁷. In contrast, αS contained both the NACore and the third repeat, contributing to its strong amyloidogenic propensity. In βS, the sole moderately aggregation-prone region was further shielded by the C-terminal domain (Figure 7b&d), limiting its accessibility for fibril growth. As a result, the overall amyloid aggregation tendency of βS was significantly lower than that of αS—consistent with experimental observations showing that βS was amyloid-resistant, whereas αS functions as an amyloid-prion.

Since the N-terminal residues 36–60 of αS exhibited a strong preference for interacting with the NACore region via β-sheet formation (Figure 5a), and βS shared up to 90% sequence homology with αS in the N-terminal segment, the corresponding region of βS was likely capable of interacting with the NACore of αS. Consequently, the co-existence of αS and βS was expected to lead to the formation of hetero-aggregates through co-aggregation or cross-seeding. Previous studies showed that amyloidogenic core residues—whether in αS^{17, 65}, Aβ^{25, 66}, or hIAPP^{46, 66}—could recruit isolated peptides and extend β-sheet edges with limited sequence specificity^{2, 67}. Given that βS was known to be amyloid-resistant, interactions between αS and βS likely resulted in hetero-aggregates, potentially modulating or inhibiting αS amyloid aggregation. Indeed, similar hetero-aggregates formed by amyloidogenic Aβ with amyloid-resistant peptides such as SEVI⁶⁸, αB-crystallin⁶⁹, or β-endorphin⁵⁵ were reported to inhibit Aβ aggregation, as the amyloidogenic core residues were effectively shielded by the amyloid-resistant species. Therefore, cross-interaction between αS and βS was expected to inhibit αS fibrillization, consistent with experimental reports showing that βS partitioned into αS condensates, promoted liquid–liquid phase separation, and suppressed the liquid–solid transition associated with fibril formation⁵.

3. Conclusion.

In this study, we systematically investigated the ensemble conformational dynamics of αS and βS monomers through 100 independent 1000-ns atomistic DMD simulations. Both proteins predominantly adopted intrinsically disordered conformations, punctuated by transient formation of helices and β-sheets. Overall, αS exhibited a stronger propensity toward β-sheet formation, whereas βS favored helical structures. A shared feature in both proteins was a conserved helical tendency within the N-terminal residues 1–35, followed by a β-sheet–prone character in the latter half of the N-terminal domain. Notably, the NAC region of αS, known to be critical for its pathogenicity, frequently adopted dynamic β-sheet structures, while the corresponding region in βS predominantly assumed transient helical conformations. Despite being highly dynamic and largely disordered, the C-terminal domains of both αS and βS transiently interacted with β-sheet–prone regions—such as the NAC and residues 36–54—serving as flexible structural caps that may limit β-sheet elongation and mitigate aggregation potential. Thermodynamic mapping of the conformational free energy landscapes revealed an enthalpy–entropy compensation mechanism underlying the structural preferences of both proteins. Ordered conformations were stabilized by lower potential energy but incurred significant entropic penalties, whereas disordered states, though higher in potential energy, were favored due to elevated conformational entropy. Crucially, the reduction in potential energy in αS was mainly driven by β-sheet formation, while in βS it was predominantly associated with helix formation. These observations underscore an intrinsic sequence-encoded bias in secondary structure stabilization that not only impacts aggregation behavior but also likely contributes to functional divergence. Collectively, our findings offer mechanistic insights into the sequence-dependent structural dynamics of synucleins and establish a thermodynamic framework for understanding how subtle sequence variations can drive distinct biophysical properties in closely related intrinsically disordered proteins. This work enhances our understanding of synuclein biology and may inform the rational design of modulators targeting conformational equilibria in neurodegenerative diseases.

Supplementary Material

Supporting Information

The Supporting Information is available free of charge at https://pubs.acs.org/doi/….

The equilibrium and convergence analysis of αS monomer simulations (Figure S1); the equilibrium and convergence analysis of βS monomer simulations (Figure S2); the secondary structure analysis of unstructured conformations in αS and βS monomers (Figure S3); Residue-wise unstructured conformation propensity in αS and βS monomers (Figure S4); identification of hotspot binding regions in the C-terminus interacting with residues from the N-terminus and NAC domain in αS and βS Monomers (Figure S5). (PDF)

Data and Software Availability

DMD simulation engine is available at Molecules In Action, LLC. (www.moleculesinaction.com). Initial conformations, input parameter and topology files for DMD simulation, and representative DMD output trajectories for each system are available (https://zenodo.org/records/15833472).