Lipid packing defects are necessary and sufficient for membrane binding of α-synuclein

Johnson, David H.; Kou, Orianna H.; White, John M.; Ramirez, Stephanie Y.; Margaritakis, Antonis; Chung, Peter J.; Jaeger, Vance W.; Zeno, Wade F.

doi:10.1038/s42003-025-08622-7

Download PDF

Article
Open access
Published: 07 August 2025

Lipid packing defects are necessary and sufficient for membrane binding of α-synuclein

Communications Biology volume 8, Article number: 1179 (2025) Cite this article

2463 Accesses
1 Citations
12 Altmetric
Metrics details

Subjects

Abstract

α-Synuclein (αSyn), an intrinsically disordered protein implicated in Parkinson’s disease, is thought to initiate aggregation by binding to cellular membranes. Previous studies suggest that anionic lipids are necessary for this binding. However, these studies largely focused on unmodified αSyn, while physiological αSyn is N-terminally acetylated (NTA). Our work challenges the long-standing paradigm that anionic lipids are necessary for αSyn binding by demonstrating that NTA diminishes αSyn’s reliance on anionic membranes, revealing that membrane packing defects (i.e., interfacial hydrophobicity) alone can drive membrane binding. Using fluorescence microscopy and circular dichroism spectroscopy, we monitored the binding of NTA-αSyn to membrane vesicles with different lipid compositions. Phosphatidylcholine and phosphatidylserine concentrations were varied to control surface charge, while phospholipid tail unsaturation and methylation were varied to modulate lipid packing. We also formulated cholesterol-containing membranes that mimicked the lipid composition of synaptic vesicles. In these membranes, all-atom molecular dynamics simulations were used to visualize and quantify membrane packing defects. Our results demonstrate that membrane packing defects are necessary for NTA-αSyn binding and that defect-rich membranes are sufficient for NTA-αSyn binding regardless of membrane charge. These findings provide a molecular mechanism by which lipid structural properties, such as poly-unsaturation, can regulate αSyn binding to physiological membranes.

The role of alpha-synuclein in synucleinopathy: Impact on lipid regulation at mitochondria–ER membranes

Article Open access 30 April 2025

The docking of synaptic vesicles on the presynaptic membrane induced by α-synuclein is modulated by lipid composition

Article Open access 10 February 2021

Abiotic lipid metabolism enables membrane plasticity in artificial cells

Article 22 May 2025

Introduction

α-Synuclein (αSyn) is a membrane-binding intrinsically disordered protein found primarily in the brain, where it localizes to synaptic terminals of neurons¹. Its notoriety originates from its implication and pathogenic function in neurodegenerative diseases, most notably in Parkinson’s Disease (PD), through its inclusion in Lewy bodies^2,3,4. However, in recent years, the intimate relationship between α-synuclein’s physiological function and pathology has become more apparent⁵. This revelation has led to a need to better understand the fundamental interactions that dictate αSyn’s function, specifically in its interaction with lipid bilayers⁶. For example, αSyn is known to assist in the trafficking of synaptic vesicles^7,8,9, but recent research has shown that membrane lipid composition, specifically, affects the role of αSyn in vesicle clustering and docking during trafficking^10,11,12. Additionally, it is known that monomeric αSyn interactions with lipid membranes influence or even induce¹³ fibrilization and aggregation of αSyn^14,15,16. These insights bring attention to the overall importance of how αSyn-membrane interactions dictate αSyn function and pathology.

Over the last twenty years, many studies have described the binding behavior of αSyn to a variety of vesicle compositions to better understand its purpose and function^{10,17,18,19,20,21,22,23,24,25,26,27,28,29}. Upon membrane association, this 140 amino-acid protein transitions from a completely structureless state in solution to a partially folded structure on the membrane surface, where the N-terminal domain consisting of approximately 100 residues adopts an α-helical conformation^21,30. One primary observation from these studies is the critical importance of anionic phospholipids for αSyn binding^{19,20,21,22,24,28}. Previous work has also demonstrated that αSyn binding is influenced by lipid packing defects, which can be defined as transient surface gaps between lipid headgroups where the underlying hydrophobic hydrocarbon tails are exposed to solvent³¹. It is important to note that a lipid packing defect, as defined here, is a term to quantify exposed hydrophobic area of lipid bilayers due to non-ideal packing, highlighting the importance of hydrophobic interactions for aSyn binding. These packing defects have been shown to enhance membrane binding of wild type αSyn (WT-αSyn)^26,27, especially when phospholipids containing methyl-branched tails¹⁷ or single tails^12,32 were used. While there is a substantial body of work collectively describing the effects of familial PD point mutations on αSyn-membrane interactions^20,33,34, fundamental studies on the effects of post-translational modifications (PTMs) have only been more recently considered as many αSyn PTMs are found to be implicated in aggregation, formation of Lewy bodies, and neurodegeneration^35,36,37. As a primary example, physiological αSyn exists with an N-terminal acetylation^{38,39,40,41,42}. Importantly, this PTM has been shown to modify αSyn aggregation kinetics^43,44,45 and affinity for lipid membranes^46,47,48,49 without modifying its structural properties when membrane-bound⁵⁰. This PTM has been overlooked in many previous binding studies, but it is important to consider when understanding αSyn’s function and pathology due to its physiological relevance and abundance^38,39. To date, existing literature comparing the binding behavior between WT-αSyn and N-terminally acetylated αSyn (NTA-αSyn) shows minimal difference in binding behavior to highly charged membranes containing 50 mol% of anionic phosphatidylserine lipids⁴⁷. This difference was more pronounced in moderately charged membranes containing 15–25% phosphatidylserine, where NTA-αSyn exhibited greater membrane binding than WT-αSyn^46,49. However, the effect of NTA on the binding of αSyn in the context of membrane defects remains to be studied, leaving a gap in understanding of how lipid packing defects impact the binding of NTA-αSyn in the context of membrane charge. By considering membrane charge, defect content, and N-terminal acetylation together, we hope to bring new insight into αSyn membrane interactions.

Here, we sought to investigate NTA-αSyn binding as a function of membrane charge and packing defect density by using a variety of lipid compositions in small unilamellar vesicles (SUVs). We first observed experimentally that NTA increased the binding of αSyn to both neutral and anionic membranes when compared to WT-αSyn but saw a disproportionate increase and preference for binding to defect-rich neutral membranes. This observation warranted a more rigorous study into the importance of packing defects for αSyn membrane binding. To quantify packing defect size and quantity, we established model systems using all-atom molecular dynamics simulations of lipid bilayers. The combination of experimental and simulation data revealed that as lipid packing defects increased across membrane compositions, NTA-αSyn binding also increased. Most notably, when anionic charge was added to membranes, NTA-αSyn binding did not increase unless the density of packing defects also increased. Therefore, our results suggest that charge alone is not enough to regulate NTA-αSyn binding and that membrane packing defects are necessary and sufficient for αSyn binding, differing from previous analyses that suggest anionic lipids are the primary promotor of membrane binding^19,28. Lastly, we show that membrane packing defects can be regulated in SUVs mimicking synaptic vesicle compositions by increasing the degree of unsaturation in phospholipid tails. Together, these observations reform the existing narrative about how αSyn binding is regulated for biological membranes and indicate that compositional considerations beyond phospholipid charge are necessary to fully encapsulate NTA-αSyn binding behavior.

Results

NTA-αSyn can exhibit preferential binding to neutral membranes over anionic membranes

To compare membrane compositional preferences of WT-αSyn and NTA-αSyn (Fig. 1A), we utilized a previously developed, fluorescence microscopy-based tethered vesicle assay to monitor αSyn binding to SUVs⁵¹ (Fig. 1B). SUV diameters were determined through calibrations where dynamic light scattering (DLS) distributions and fluorescence intensity distributions were overlaid, while αSyn binding was quantified by performing single-molecule imaging of fluorescently labeled proteins. Detailed explanations for these experimental procedures are provided in the Materials and Methods section. For these fluorescence microscopy experiments, αSyn was fluorescently labeled with ATTO-488 on residue 136, while SUV bilayers were doped with 0.5 mol% ATTO-647N labeled lipid. We used fluorescence microscopy and circular dichroism (CD) spectroscopy to verify that the inclusion of the fluorescent probe negligibly impacted αSyn’s ability to bind membranes (Supplementary Fig. 1). First, we compared the binding of αSyn to SUVs composed of 3:1 (mol:mol) dioleoylphophatidylcholine:dioleoylphosphatidylserine (DOPC:DOPS) and SUVs composed of diphytanoylphosphatidylcholine (DPhPC). The DOPC/DOPC SUVs represent anionic membrane surfaces with a 25% negative charge, while the DPhPC SUVs represent zwitterionic membrane surfaces with zero net charge. These two compositions were chosen because prior studies showed that tail methylation¹⁷ and anionic headgroups¹⁹ increased WT-αSyn binding. However, these properties have never been examined head-to-head, particularly for NTA-αSyn; therefore, we sought to observe αSyn binding preference to both membrane compositions.

**Fig. 1: N-terminal acetylation of α-synuclein enhances overall membrane binding and sensitivity for lipid packing defects.**

In Fig. 1C, αSyn binding was observed via fluorescence microscopy as the fluorescence signals from the protein channel were spatially colocalized with the fluorescence signals from the lipid channel. During our processing and analysis procedures for fluorescence images, we isolated SUV populations with an average diameter of 100 nm to control for membrane curvature effects (Fig. 1D), as αSyn is known to sense membrane curvature⁵². In Fig. 1C WT-αSyn exhibited a higher fluorescence intensity on DOPC/DOPS membranes in comparison to DPhPC membranes, which corresponds to a higher degree of binding on DOPC/DOPS vesicles, as quantified in Fig. 1D. Shown in Fig. 1C, D, N-terminal acetylation increased the overall membrane affinity for both DOPC/DOPS and DPhPC membranes, which is consistent with previous observations^46,47,49. However, unlike WT-αSyn, NTA-αSyn displayed a binding preference for the neutral DPhPC SUVs. NTA resulted in a ~ 2-fold increase for αSyn binding on DOPC/DOPS SUVs, while DPhPC SUVs elicited a ~ 5-fold increase. These relative trends persisted on SUVs that were incubated with lower concentrations of WT-αSyn and NTA-αSyn (Supplementary Fig. 2). While there is a growing body of recent literature showing the importance of neutral membrane lipids for αSyn-membrane interactions^11,12,32, to the best of our knowledge, our results indicate the first reported instance of αSyn exhibiting a preference for neutral membranes over highly charged anionic membranes (Fig. 1E). Importantly, these findings were obtained under conditions that controlled for membrane curvature (Fig. 1D). Therefore, it is clear that the tail chemistry and, ultimately, lipid packing play a significant role in the membrane binding of NTA-αSyn. This fascinating result warranted further analysis to understand the mechanism and potential physiological implications for NTA-αSyn binding preferences to defect-rich membranes over charged membranes.

NTA-αSyn membrane binding increases with higher packing defect content

Due to the various competing factors that control lipid packing, such as membrane curvature and membrane phase behavior, all lipids in Fig. 2 were analyzed at constant curvature (Supplementary Figs. 3–4) in the liquid crystalline phase at room temperature (i.e., 21 °C). Figure 2A provides structures for four phospholipids: dilauroylphosphatidylcholine (DLPC), DOPC, DOPS, and DPhPC. Collectively, these phospholipids represent the three types of tail chemistries analyzed in this work – dilauroyl (DL), dioleoyl (DO), and diphytanoyl (DPh), as well as two classes of headgroup – zwitterionic phosphatidylcholine (PC) and anionic phosphatidylserine (PS). The DL lipids were expected to exhibit the closest packing due to the complete saturation of their carbon chains, while the DO and DPh tails were expected to pack increasingly more loosely due to mono-unsaturation and methylation, respectively. PC and PS headgroups were chosen due to their physiological relevance throughout intracellular membranes⁵³. To quantify the relative quantity of packing defects in each membrane composition, we performed all-atom molecular dynamics simulations of lipid bilayers (Supplementary Tables 1–2). Figure 2B shows selected simulation snapshots for pure DLPC and DPhPC membranes. Each snapshot represents a single frame from a simulation trajectory of 300 ns. In these snapshots, distinct differences in defect content for each composition can be noted, as the DPhPC membrane exhibited greater exposure of the tail groups, which are highlighted in yellow. Together, the combination of DL, DO, DPh tails and PC, PS headgroups yields 6 different phospholipids, 5 of which were examined in Fig. 2C. Using the open-source software package PackMem⁵⁴, we quantified the average size (Fig. 2C, left), average number (Fig. 2C, center), and average membrane surface coverage (Fig. 2C, right) of packing defects (Supplementary Fig. 5). For PC headgroups, the average size, average number, and surface coverage of defects were higher for DO tails than for DL tails. The inclusion of PS in 3:1 DLPC:DLPS membranes did not change the average size of packing defects compared to pure DLPC membranes, nor did it increase the frequency and surface coverage of packing defects. However, the inclusion of PS in 3:1 DOPC:DOPS membranes resulted in significant increases to defect size and quantity when compared to pure DOPC membranes. Interestingly, pure DPhPC membranes possessed substantially larger and more numerous defects than the other 4 systems examined.

**Fig. 2: All-atom molecular dynamics simulations of lipid membranes allow for quantification of membrane packing defects.**

We next examined the binding preference of NTA-αSyn to SUVs as a function of lipid composition (Fig. 3). Using the tethered vesicle assay to generate binding curves, we observed an increase in NTA-αSyn binding as the lipid composition was varied from pure DLPC to pure DOPC, then finally pure DPhPC (Fig. 3A, Supplementary Figs. 3–4). This increased binding was consistent with increased density of packing defects (Fig. 2C). When 25% PS was included in DLPC/DLPS and DOPC/DOPS membranes, the resulting binding trends of the five lipid compositions examined mirrored the trends for defect coverage in Fig. 2C, suggesting that the level of binding for NTA-αSyn may be proportional to the surface coverage of packing defects. Additionally, the results in Figs. 2C and 3B reveal that including 25% PS did not significantly increase the presence of defects nor the amount of NTA-αSyn binding for DL membranes. Conversely, 25% PS substantially increased the surface coverage of packing defects in DO membranes and the levels of NTA-αSyn binding. Nonetheless, neutral DPhPC membranes elicited the highest degree of NTA-αSyn binding for the 5 compositions examined so far.

**Fig. 3: Packing defects are necessary and sufficient for NTA-αSyn binding regardless of membrane charge.**

These trends were mirrored in binding curves that were obtained using CD spectroscopy (Fig. 3C). Here, we directly monitored NTA-αSyn insertion into membrane defects by tracking spectral changes of the non-fluorescently labeled protein as it transitioned from a completely disordered structure to a partially folded a-helical structure upon membrane binding^48,55. To limit curvature-influenced effects, SUVs were prepared such that the vesicle size distributions were comparable among all vesicle compositions (Supplementary Fig. 6). After CD spectra were obtained for lipid-free αSyn and lipid-saturated αSyn, intermediate spectra were fit using Eq. 1 to determine ω, the fraction of αSyn bound to vesicles for a given lipid concentration (Fig. 3D, Supplementary Fig. 6, and Supplementary Table 3). Here, the lipid concentration is reported as a molar excess of lipid (i.e., moles of lipid per mole of protein). In Fig. 3E ω is plotted as a function of the lipid concentration for 5 different compositions. The mirrored binding curves in Fig. 3B, E verify that our fluorescence microscopy-based tethered vesicle assay was appropriate for measuring membrane insertion of αSyn. The binding curves in Fig. 3E were fit to a Langmuir-Hill binding curve (Eq. 2) to determine an apparent dissociation constant (K_d) for each composition (Supplementary Table 4). K_d could not be determined for DLPC and DLPC/DLPS systems due to the exceptionally low binding observed. Of the DOPC, DOPC/DOPS, and DPhPC systems analyzed in Fig. 3F K_d was lowest for DPhPC membranes and highest for DOPC membranes, further indicating that binding affinity was highest for DPhPC membranes.

Negative surface charge synergistically enhances NTA-αSyn binding to membranes

We next investigated whether negative charge affects NTA-αSyn binding to DPh membranes, which are inherently defect-rich. To carry out these experiments, we utilized diphytanoylphosphatidylserine (DPhPS) mixed with DPhPC at a molar ratio of 3:1 DPhPC:DPhPS (Fig. 4). The tethered vesicle assay (Fig. 4A and Supplementary Fig. 4) and CD spectroscopy (Figs. 4B and Supplementary Fig. 7) both revealed that NTA-αSyn exhibited preferential binding to membranes containing 3:1 DPhPC:DPhPS in comparison to membranes containing just DPhPC. Of note, the inclusion of PS increased the surface coverage of packing defects in DPh membranes by increasing the average size of individual membrane defects (Fig. 4C). When comparing the results of PC experiments and PC/PS experiments as a function of lipid tail chemistry, we found that the binding of NTA-αSyn exhibited a strong linear response to surface coverage by defects with both fits intersecting the x-axis at a value of approximately 2.5% (Fig. 4D). This strong linear response is indicative of binding in the sub-saturation regime, where the amount of binding is directly proportional to the number of binding sites available.

**Fig. 4: Packing defects and anionic charge synergistically enhance NTA-αSyn binding to membranes.**

NTA-αSyn binding can be modulated by the degree of lipid unsaturation

We next sought to determine whether the density of membrane packing defects could be modulated by lipid tails that are more relevant than DPh in physiological contexts. Due to αSyn’s localization in the brain, the compositional intricacies of both neuronal membranes and synaptic vesicles must be considered. PUFAs, for example, are enriched in both neurons⁵⁶ and synaptic vesicles^53,57. Since PUFA lipid tails occupy more space than saturated lipid tails for a given phospholipid, we hypothesized that PUFA tails could be used as a means of changing the density of packing defects and subsequently, the amount of NTA-αSyn binding. To test the impact of lipid unsaturation on NTA-αSyn binding, we synthesized and monitored binding to vesicles containing 40/25/25/10 mol% cholesterol/PC/PE/PS to mimic the naturally occurring lipid composition in synaptic vesicles⁵³ (Fig. 5). Here, PE refers to lipids containing phosphatidylethanolamine headgroups. Lipid tails for PE and PS lipids were composed of oleic acid (i.e., DOPE and DOPS). Lipid tails within the PC component were any of the following: oleic acid (i.e., DOPC, also known as 18:1 PC), linoleic acid (18:2 PC), arachidonic acid (20:4 PC), or docosahexaenoic acid (22:6 PC) (Fig. 5A). Our simulation results revealed that the surface coverage by membrane defects indeed increased as the degree of lipid unsaturation increased (Fig. 5B). Interestingly, there was a concomitant increase in NTA-αSyn binding with an increase in the degree of unsaturation as revealed using fluorescence microscopy (Fig. 5C). These results were also observed when SUVs were incubated with lower concentrations of NTA-αSyn (Supplementary Fig. 8). Similar to the results in Fig. 4D, this binding exhibited a direct proportionality to surface coverage by membrane defects but notably intersects the x-axis at a value near 0% (Fig. 5D).

**Fig. 5: Lipid polyunsaturation modulates membrane packing defects in synaptic vesicle mimics, which in turn regulates NTA-αSyn binding.**

Discussion

αSyn’s relevance to PD and other neurodegenerative disorders has been well-studied, and the membrane binding of αSyn has been previously shown to be an essential part of its pathology through the nucleation of aggregation and fibrillization^9,13,14. Here, we provide insights into αSyn’s binding to reconstituted membrane surfaces by modulating the lipid compositions, revealing that membrane packing defects are the primary driver of αSyn binding. Our results demonstrate that NTA-αSyn, the physiologically prevalent form of αSyn, has greater binding to defect-rich, neutrally charged surfaces composed of DPhPC over anionic membranes with fewer packing defects, such as those containing DOPC:DOPS at a 3:1 molar ratio. NTA may selectively enhance αSyn’s binding to DPhPC membranes by increasing its inherent α-helical content in the unbound state^49,58, thus reducing the energy barrier necessary for membrane insertion⁵⁰. NTA also increases the hydrophobicity of the N-terminus of αSyn, replacing the charged amine with a neutral acetyl group, incentivizing additional hydrophobic interactions⁵⁹. While this reduction in positive charge diminishes electrostatic interactions between the unmodified, N-terminal methionine and anionic lipid headgroups, recent MD simulations have shown that this loss in electrostatic contribution to binding energy is offset by increased hydrophobic contributions facilitated by the acetylated methionine⁵⁸. These simulations also demonstrate that acetylation promotes extension of αSyn’s helical structure along the length of the N-terminus, enhancing membrane binding. This state of increased binding activity would, therefore, be highly amenable to defect-rich membranes such as DPhPC, as these membranes contain higher exposure of the underlying lipid tails. It is relevant here to note that we do not evaluate the specific alpha-helical conformation (i.e. broken helix vs extended helix) of αSyn’s N terminus upon membrane binding in this work. This precise structural conformation of the helical N terminus in the membrane-bound state remains context-dependent and lacks strong consensus, as it depends on membrane size and curvature^60,61. In Fig. 1E, αSyn is depicted as an extended helix, which is the likely conformation adopted upon binding to 100 nm vesicles⁶².

The observed preference for membrane defects by NTA-αSyn was preserved across both binding assays used. Fluorescence microscopy and circular dichroism spectroscopy combined with all-atom molecular dynamics simulations demonstrated that the presence of defects dictates the ability of αSyn to insert into membranes, and that negatively charged membranes cannot overcome a lack of membrane defects necessary for αSyn binding. This result is particularly interesting in the context of recent MD simulations that have probed the relative contribution of electrostatic and hydrophobic forces for αSyn membrane binding^{58,63,64,65,66}. This literature jointly describes a multi-state binding event in which electrostatic and hydrophobic interactions serve distinct roles in αSyn binding. Electrostatics dominate when the negative headgroups interact with the lysine-rich N-terminus of αSyn, bringing the protein close to the membrane surface. The hydrophobic interactions then dominate when the N-terminus embeds into the membrane upon binding. These hydrophobic interactions become even more relevant in the context of NTA-αSyn where MD simulations show additional helical propensity increases membrane binding affinity⁵⁸. Throughout this literature, electrostatic interactions were shown to assist in αSyn binding through membrane recognition, but the defects and hydrophobic interactions mediate the physical insertion of the protein beyond surface adsorption⁶⁶. This necessity of defects for binding and insertion was exemplified with our results utilizing DL membranes, where the addition of 25% charge from PS did not increase the density of packing defects nor the amount of bound αSyn. Even in the presence of substantial anionic charge, which assists with αSyn localization to the membrane, insufficient defects resulted in negligible binding to DL membranes. In contrast, we note a substantial increase in defect coverage and binding of αSyn to DO membranes when 25% PS was included. This increase in defect content could be due to repulsion between charged lipid headgroups that open additional defects in the DO system or due to a mismatch in headgroup size between PC and PS⁵⁹. While increased αSyn binding was strongly correlated with an increase in defect content and neutral DPhPC membranes elicited stronger binding than anionic DOPC/DOPS membranes, it must be noted that anionic charge and defect content exhibited synergy in recruitment of αSyn. Interestingly, our CD Spectroscopy results suggest that the addition of PS enhanced cooperativity of αSyn binding in both DO and DPh membranes (Supplementary Table 4).

Notably, αSyn exhibited strong linear binding responses on pure PC membranes, PC/PS membranes, and cholesterol-containing synaptic vesicle mimic membranes. Interestingly, PC and PC/PS membranes yield similar x-intercept values of ~2.5%, indicating that there is a minimum coverage of defects necessary for αSyn to bind to these membranes. This explains why, when PS lipids were added to the DL system, there was a lack of binding response from NTA-αSyn as DL defect coverages were near this minimum value of ~2.5% (Fig. 2C). When comparing the neutral PC systems to the anionic PC/PS systems in Fig. 4D, the anionic membrane systems had a steeper slope than the neutral membrane systems. This steeper slope indicates that for a given increase in defect content, anionic membranes elicit a greater increase in NTA-αSyn binding. This result correlates well with existing literature stating the importance of anionic charge for αSyn binding and further emphasizes synergistic behavior between packing defects (i.e., hydrophobic interactions) and membrane charge (i.e., electrostatic interactions) for αSyn binding. In contrast to the defect minimums seen in the PC membranes (0% PS) and PC/PS membranes (25%), the linear trend for synaptic vesicle mimicking membranes in Fig. 5D exhibits a much lower slope with an x-intercept of approximately 0%. This result was somewhat surprising considering that these membranes contained 10% PS, which would be expected to yield an intermediate slope to the 0% PS and 25% PS systems. This lower slope could be attributed to either the inclusion of PE lipids or cholesterol. It is known that adding DOPE to DOPC/DOPS mixtures increases αSyn binding in SUVs²⁷, so it is likely that the addition of cholesterol is reducing NTA-αSyn’s response to increasing packing defects as adding cholesterol tightens membranes, reducing fluidity, and reducing overall αSyn membrane affinity¹⁸. Other studies have shown cholesterol reduces αSyn’s affinity to membranes by either reducing the interaction of the second half of the N-terminal α-helix with membranes⁶⁷ or attenuating the electrostatic interaction between the membrane and αSyn⁶⁶. In the context of our results, some combination of these mechanisms could be reducing αSyn binding at higher defect content, but it is likely that membrane fluidity reduction and tightened packing is a prominent mechanism with 40 mol% cholesterol membranes. Interestingly, our extrapolation in Fig. 5D suggests that cholesterol-containing membranes promote stronger NTA-αSyn binding in the low defect regime than the pure PC and PC/PS membranes used in Fig. 4D. NTA-αSyn binding to cholesterol-rich compositions with low defects could be due to cholesterol inclusion loosening the membrane in densely packed compositions allowing for αSyn insertion⁶⁴ or the direct interaction between the hydroxyl headgroup and αSyn¹⁸. Additionally, recent research has shown that αSyn can serve as a cholesteryl-ester (CE) specific sensor in lipid droplets and bind specifically to a pure CE lipid droplet⁶⁸. This type of preferential cholesterol interaction may contribute to membrane recruitment of NTA-αSyn in the low defect regime. Taken together, it seems that cholesterol serves as a complementary regulator of αSyn binding. While membrane defects are necessary for αSyn helix insertion, cholesterol could strengthen αSyn binding in defect-poor compositions through a cholesterol specific interaction and attenuate αSyn binding in defect-rich compositions.

The inclusion of PUFAs into synaptic vesicle mimics provides insight into how cells might regulate αSyn binding in physiological contexts. We observed a strong correlation between αSyn binding to membranes and the degree of unsaturation in lipid tails. Incidentally, the enrichment of PUFAs in the brain leads to the formation of oligomeric αSyn^51,52,69,70. More specifically, the 22:6 PUFA, which is the PUFA that elicited the highest degree of αSyn binding in our synaptic vesicle mimic experiments, is known to be further enriched in PD patients with a distinct link in enhancing αSyn neuropathology and oligomerization^{25,71,72,73,74}. We, therefore, hypothesize that this enrichment of PUFAs in the brain may lead to the formation of oligomeric αSyn by promoting higher seeding densities for αSyn on the membrane surface. This proposed mechanism indicates how lipid polyunsaturation could regulate membrane packing defects, thus impacting αSyn interactions with synaptic vesicles and, by extension, aSyn pathology.

Previous studies have neglected the significance of αSyn interactions with zwitterionic membranes, citing evidence of reduced binding compared to anionic membranes. However, this current study, along with a growing body of work^11,12,32, shows that αSyn interactions with zwitterionic membranes, when sufficient defects are present, can be comparable if not greater than a charged membrane. Following our results, future studies should reframe analyses that focus solely on anionic membranes to include interactions with zwitterionic membranes. For example, αSyn’s direct interactions with the endoplasmic reticulum (ER), a noted zwitterionic membrane⁷⁵, have been shown to impact vesicle trafficking and increase ER stress leading to neuroinflammation and neurodegeneration⁷⁶. Although these types of ER interactions have been observed, the underlying cause remains unexplained in the literature. Our study fills this knowledge gap by elucidating mechanisms by which physiological αSyn interacts with electrostatically neutral membranes. Additionally, our work informs molecular mechanisms for therapeutic intervention. Previous work has shown that αSyn fibrillization and toxicity can be reduced in dopaminergic neurons by incubating them with zwitterionic lipid nanoparticles (LNPs)⁷⁷. While the authors considered their findings controversial in the context of αSyn-membrane literature, they hypothesized that αSyn interacts with zwitterionic LNPs to reduce the cytosolic pool of free αSyn available for fibrillation and aggregation. Our current study corroborates these findings by experimentally demonstrating that αSyn can preferentially bind to zwitterionic membranes when the tail chemistry of the phospholipids is favorable. Future work in therapeutic applications could leverage our findings to design and formulate LNPs with enhanced specificity for αSyn interactions.

Conclusion

In conclusion, we show that membrane packing defects are the primary drivers of NTA-αSyn binding in both neutral and anionic phospholipid membranes. Interestingly, NTA-αSyn did not exhibit appreciable binding to anionic membranes until a critical threshold of packing defects was reached. While anionic charge did synergistically enhance NTA-αSyn binding above this minimum defect threshold, these enhancements were coupled to a further increase in membrane defect content likely resulting from electrostatic repulsion between phospholipid head groups. Furthermore, we saw that αSyn binding can be modulated by the degree of unsaturation in lipid tails, which directly controls the density of membrane packing defects. We also noted that cholesterol differentially regulates the membrane binding of NTA-αSyn, enhancing binding in low-defect regimes while inhibiting it in the high-defect regimes. Our results outline molecular mechanisms by which lipid compositions can regulate membrane-nucleated oligomerization and pathology of αSyn. These findings necessitate a reevaluation of the current framework surrounding αSyn-membrane interactions as previous studies primarily focused on anionic membranes. By challenging the prevailing notion that electrostatic interactions with lipid headgroups primarily drive αSyn-membrane interactions, this work underscores the necessity of considering lipid tail chemistry when investigating αSyn-membrane interactions. This mechanism of binding through membrane packing defects likely extends beyond αSyn, pointing to a broader paradigm in protein-membrane interactions. By emphasizing lipid tail chemistry and membrane defect density, this study may uncover how various proteins detect and adapt to specific membrane environments, deepening our understanding of membrane-driven processes.

Materials and methods

Materials

1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1,2-dioleoyl-sn-glycero-3-phospho-L-serine (sodium salt) (DOPS), 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE), 1,2-diphytanoyl-sn-glycero-3-phosphocholine (DPhPC), 1,2-diphytanoyl-sn-glycero-3-phospho-L-serine (sodium salt) (DPhPS), cholesterol (plant), 1,2-dilauroyl-sn-glycero-3-phosphocholine (DLPC), 1,2-dilauroyl-sn-glycero-3-phospho-L-serine (sodium salt) (DLPS), 1,2-dilinoleoyl-sn-glycero-3-phosphocholine (18:2 PC), 1,2-diarachidonoyl-sn-glycero-3-phosphocholine (20:4 PC), and 1,2-didocosahexaenoyl-sn-glycero-3-phosphocholine (22:6 PC) were purchased from Avanti Polar Lipids. Dipalmitoyl-decaethylene glycol-biotin (DP-EG15-biotin) was generously provided by D. Sasaki from Sandia National Laboratories, Livermore, CA⁷⁸. ATTO 647 N 1,2-Dipalmitoyl-sn-glycero-3-phosphoethanolamine (DPPE) and ATTO 488-maleimide were purchased from ATTO-TEC. Biotin-PEG-Silane (5000 MW) and mPEG-Silane (5000 MW) were purchased from Laysan Bio, Inc. Isopropyl-beta-D-thiogalactoside (IPTG) was purchased from Gold Biotechnology. Tris(2-carboxyethyl) phosphine hydrochloride (TCEP), phenylmethanesulfonyl fluoride (PMSF), sodium phosphate monobasic, sodium phosphate dibasic, and β-mercaptoethanol (BME) were purchased from Sigma-Aldrich. 4-(2-hydroxyethyl)-1-piparazineethanesulphonic acid (HEPES) and tris(hydroxymethyl)aminomethane (Tris) were purchased from Fisher Scientific.

α-synuclein expression and purification

The wild-type α-synuclein expression plasmid (pET21a backbone) was provided by Peter Chung (University of Southern California, Los Angeles, CA). The mutated (Y136C) α-synuclein expression plasmid (pRK172 backbone) was provided by Jennifer C. Lee (National Institutes of Health, Bethesda, MD). Each respective plasmid was transformed into BL21 (DE3) cells that were pre-transformed with the pNatB plasmid coding for N-α-acetyltransferase to acetylate the α-synuclein⁷⁹. After incubation in 2xTY media at 37 ° C to OD₆₀₀ ~ 0.6, cells were induced with 1 mM IPTG and incubated for an additional 4 h. Cells were pelleted and resuspended in 25 mM Tris, 20 mM NaCl, 1 mM PMSF, and 1x Roche protease inhibitor cocktail (Sigma-Aldrich) (pH 8.0). 10 mM BME was added for the Y136C mutant to prevent disulfide bonding. The cells were sonicated on ice at 65% amplitude for 15 min, and then 30 μL/10 mL of cell solution of DNAse I (Thermo Fisher Scientific) was added to the solution and incubated for 30 min at 37 °C. The pH of the solution was reduced to pH 3.5, and the cell debris was pelleted for 25 min at 20,000 × g. The pH of the supernatant was set to pH 7.0, and a 50% ammonium sulfate (Sigma Aldrich) precipitation was performed with the protein precipitating out and pelleting after another 25 min, 20,000 × g centrifugation. The pellet was resuspended in 20 mM Tris (pH 8.0), run over a 5 mL HiTrap Q FF column (Cytiva), and eluted with 20 mM Tris, 1 M NaCl (pH 8.0) with 10 mM BME in both buffers for the Y136C mutant. The eluted protein was concentrated using a 3 K MWCO centrifugal filter (Sigma-Aldrich) and run through SEC on a Superdex 200 10/300 column (Cytiva) into 25 mM HEPES, 150 mM NaCl (pH 7.4) with 0.5 mM TCEP for the Y136C mutant. The appropriate fractions were concentrated, and the final concentration was measured using UV-Vis spectroscopy at 280 nm wavelength. The extinction coefficients used were 5960 M⁻¹ cm⁻¹ and 4470 M⁻¹ cm⁻¹ for WT-αSyn and Y136C-αSyn, respectively.

α-synuclein fluorescent labeling

ATTO 488-maleimide was stored as 100 mM aliquots in DMSO at −80 °C. Labeling was performed by combining ATTO-488 maleimide with Y136C-αSyn at a 6x molar excess of dye. The reaction was allowed to proceed overnight at 4 °C in the dark. The total amount of DMSO added to Y136C-αSyn never exceeded 2% v/v. Excess dye was removed by flow through a Superdex 75 10/300 GL (Cytiva) size exclusion chromatography column. The protein:dye labeling ratio was confirmed to be 1:1 by UV-Vis spectroscopy. An extinction coefficient of 90,000 M⁻¹ cm⁻¹ was used for ATTO 488 at 488 nm, with a correction factor of 0.09 for the 280 nm/488 nm absorption ratio.

SUV preparation

Lipid aliquots were thawed from −80 °C and combined in the following molar percentages for use in the tethered vesicle assays:

97.5% DPhPC, 2% DP-EG15-biotin, 0.5% ATTO 647 N
72.5% DPhPC, 25% DPhPS, 2% DP-EG15-biotin, 0.5% ATTO 647 N
97.5% DOPC, 2% DP-EG15-biotin, 0.5% ATTO 647 N
72.5% DOPC, 25% DOPS, 2% DP-EG15-biotin, 0.5% ATTO 647 N
97.5% DLPC, 2% DP-EG15-biotin, 0.5% ATTO 647 N
72.5% DLPC, 25% DLPS, 2% DP-EG15-biotin, 0.5% ATTO 647 N
23.75% DOPC, 23.75% DOPE, 10% DOPS, 40% Cholesterol, 2% DP-EG15-biotin, 0.5% ATTO 647 N
23.75% 18:2 PC, 23.75% DOPE, 10% DOPS, 40% Cholesterol, 2% DP-EG15-biotin, 0.5% ATTO 647 N
23.75% 20:4 PC, 23.75% DOPE, 10% DOPS, 40% Cholesterol, 2% DP-EG15-biotin, 0.5% ATTO 647 N
23.75% 22:6 PC, 23.75% DOPE, 10% DOPS, 40% Cholesterol, 2% DP-EG15-biotin, 0.5% ATTO 647 N

The following molar percentages were used for vesicle preparation in CD spectroscopy experiments:

100% DOPC
75% DOPC, 25% DOPS
100% DPhPC
75% DPhPC, 25% DPhPS
100% DLPC
75% DLPC, 25% DLPS

Solvents were evaporated with a nitrogen stream, followed by vacuum storage for at least 2 h. Dried lipid films were rehydrated using 20 mM sodium phosphate and 150 mM NaCl (pH 7.0) buffer. After 5 min of hydration, the lipid suspensions were subjected to 3 freeze/thaw cycles by submerging the solution in liquid nitrogen for 2 min and thawing at 60 °C for 3 min. The SUVs were then extruded through a 50 nm polycarbonate membrane (Whatman). The average diameter of vesicles used for CD spec across all compositions was 67 ± 5 nm, determined by DLS (Supplementary Fig. 6). When handling PUFA-containing mixtures, special precautions were taken to minimize oxidation. All PUFA aliquots were handled in a N2 glove box, and all buffers were purged with N2 before rehydrating the lipid films. Additionally, freeze-thaw cycles were omitted to limit exposure to any residual oxygen, and lipid extrusion was carried out within a N2 glove box. Samples were then imaged within 30 min of air exposure. All other vesicle preparations were imaged within 48 h of preparation and stored in a 4 °C refrigerator to minimize degradation of quality. Unilamellar structures were verified with a volume labeling assay (Supplementary Fig. 9), and their surface charge was confirmed with a PLL adsorption assay (Supplementary Fig. 10).

Slide passivation and tethered vesicle assay

Glass coverslips were cleaned with a modified RCA protocol and passivated with PEG-silane, as described previously, with slight modifications⁸⁰. Briefly, glass slides were soaked in a concentrated KOH/peroxide bath, followed by a concentrated HCl/peroxide bath each at 80 °C for 10 min. Once cleaned, the slides were passivated by surface treatment with a 7.5 mg/mL PEG-silane solution in isopropanol. The solution was comprised of 5% biotin-PEG-silane (5000 MW) and 95% of mPEG-silane (5000 MW). All materials for creating the PEG-silane solution were handled in an N₂ environment. Immediately before passivating the slide, 1% v/v of glacial acetic acid was added to catalyze the reaction between silanes and hydroxyl groups on the cover slip. 55 µL of the solution was added to the top surface of the cleaned coverslips, spread out, then placed into an oven at ~70 °C for 30–60 min. The slides were then rinsed under DI water to remove excess PEG-silane and dried under N₂. These slides were stored dry under air for a maximum of 1 week.

Imaging wells were created by placing 0.8 mm thick silicone sheets (Grace Bio-Labs) with 5 mm holes on the passivated slide. Each well was hydrated with 6 μg of NeutrAvidin (ThermoFisher) in 30 μL of 20 mM sodium phosphate, 150 mM NaCl (pH 7.0) and incubated for 10 min. After incubation, each well was rinsed to remove excess NeutrAvidin. Following this step, each well was rinsed with a solution containing SUVs to reach a 1 μM lipid concentration. After 10 min of incubation, excess SUVs were rinsed from the wells with the same buffer, and the α-synuclein tagged with ATTO 488 was added at the appropriate concentration.

Confocal microscopy

Imaging was performed on a laser scanning confocal microscope (Leica Stellaris 5). Two excitation lasers were used: 488 nm and 638 nm for labeled αSyn and SUVs, respectively. The wavelength detection bandwidths used were 493–558 nm and 643–749 nm for labeled αSyn and SUVs, respectively. A Leica HC PL APO 63x, 1.4 NA oil-immersion objective was used to acquire images, and the zoom factor was set such that square pixel sizes of 70 nm were obtained. All images were acquired using a scan speed of 400 Hz.

Image processing for the tethered vesicle assay

Image processing was performed similarly to previous work⁸¹. Briefly, images of diffraction-limited, fluorescent lipid and protein puncta were acquired using confocal microscopy. Each image was a three-frame repeat of a single field of view. Using publicly available particle detection software (cmeAnalysis), the 2D fluorescence distribution of all puncta in each image was fit to 2D-gaussian functions⁸². Using this fit, the fluorescent amplitudes were determined for the puncta in both the lipid and protein channels. Only puncta in the lipid channel that persisted through all three consecutive image frames in a single field of view were considered in this analysis, ensuring that all vesicles detected were above the fluorescent noise threshold. As an additional filtering step, any puncta whose amplitude was not significantly above the level of background fluorescence was excluded from the analysis. Our analysis also excluded vesicles with diameters above 250 nm to ensure that all analyzed puncta were truly diffraction-limited. Protein binding for each vesicle was determined by analyzing the colocalization between both fluorescent channels. The search radius for colocalized protein fluorescence amplitudes was three times the standard deviation of the Gaussian fit of the point spread function in the lipid channel.

Calibration of vesicle diameters and number of bound proteins

SUV diameters and the number of proteins bound to each vesicle were determined as previously described^51,81. Briefly, tethered SUVs were imaged to obtain a distribution of the fluorescence intensities for the vesicle population. A conversion factor was then determined by overlaying the square root of the fluorescence distribution and the DLS diameter distribution (Fig. 1B). All tethered vesicles included in the binding analysis (Figs. 1, 3–5, Supplementary Figs. 2–4, 8–9) were subsets of their respective vesicle populations and selected in the diameter range 80–120 nm, resulting in average diameters of 100 nm across all compositions studied. This step was necessary to control for curvature-sensing effects in the interpretation of our results.

To determine the number of proteins bound to each vesicle, we performed single-molecule imaging of fluorescent αSyn that was adsorbed to glass coverslips. The fluorescence intensities of the punctate structures were used to generate a distribution (Fig. 1B). The peak value of this fluorescence intensity distribution was then used to represent the fluorescence intensity of a single ATTO 488 labeled αSyn. Quantitative adjustments were made during the data processing to account for changes in image frame accumulations and linear gain settings on the detector.

Volume labeling assay

To ensure the imaged structures were unilamellar vesicles, a volume labeling assay was used (Supplementary Fig. 9). SUVs containing lipid-anchored ATTO-647N were formed in the presence of 100 μM ATTO-488 dye, encapsulating it within the vesicle lumen. The vesicles were then tethered onto the glass surface as described in Slide Passivation and Tethered Vesicle Assay. All excess ATTO-488 dye in solution was rinsed away from outside the vesicles in the final buffer rinsing step. All vesicle compositions were then imaged under the same two-channel confocal settings with ATTO-488 representing the vesicle lumen (i.e., volume label) and ATTO-647N representing the vesicle bilayer (i.e., surface label). Relevant discussions of these results can be found in the Supplementary Discussion.

Poly-L-Lysine (PLL) adsorption assay

Because DLPS has non-negligible water solubility with an estimated CMC of ~1 μM⁻¹, we performed a tethered vesicle assay with poly-L-lysine (PLL) to verify DLPS was not leaching from the DLPC/DLPS vesicles and influencing the vesicle surface charge (Supplementary Fig. 10). PLL was labeled with ATTO 488 by incubating PLL with ATTO 488 NHS-Ester in a 1:1 molar ratio in 20 mM HEPES, 150 mM NaCl (pH 7.4) at room temperature for 30 min. Excess dye was removed by running the solution through a 7k MWCO Zeba Spin Column. The vesicles were prepared and imaged as described in the SUV Preparation and Slide Passivation and Tethered Vesicle Assay. 500 nM of PLL-ATTO488 was incubated with DOPC, 3:1 DOPC/DOPS, DLPC, and 3:1 DLPC/DLPS, and the fluorescence intensity of the PLL-ATTO488 adsorbed to 100 nm average vesicles was extracted. Relevant discussions of these results can be found in the Supplementary Discussion.

Circular dichroism spectroscopy

Circular dichroism measurements were collected on a JASCO J-1100 instrument with a 1 mm path length quartz cuvette (JASCO). Scans were taken from 195 to 300 nm at a 0.5 nm data pitch with 100 nm/min scan speed. The data integration time was 2 s, and each scan was accumulated 2 times. The buffer used for each experiment was 20 mM sodium phosphate, 150 mM NaCl (pH 7.0). Samples of vesicles and protein were allowed to equilibrate for at least 15 min before spectra were acquired. Each sample contained 5 μM αSyn and was blanked by a buffer solution containing SUVs at the appropriate lipid concentration.

Equation 1 was used to determine ω, the fraction of bound αSyn. Here, C_U is the average spectrum for unbound, disordered αSyn, C_S is the average saturated spectrum for membrane-bound αSyn, and C_I is the intermediate spectrum for αSyn that is partially bound to membranes (i.e., a mixture of bound and unbound states). Sonicated SUVs composed of 3:1 DPhPC/DPhPS were used to obtain saturated spectra, C_S.

$${C}_{I}=\left(1-\omega \right){C}_{U}+(\omega ){C}_{S}$$

(1)

The apparent dissociation constant, K_d, was found using a Langmuir-Hill adsorption isotherm to fit ω as a function of the lipid:protein molar ratio, [L:P] (Eq. 2). Here, n represents the Hill coefficient.

$$\omega =\frac{{[L:P]}^{n}}{{({K}_{d})}^{n}+\,{[L:P]}^{n}}$$

(2)

All-atom molecular dynamics simulation models

Atomistic lipid membrane models were built using the CHARMM-GUI^83,84 webserver’s Membrane Builder⁸⁵ functionality. System geometry, composition, and simulation parameters were inspired by Garten and coworkers who performed similar analysis on other lipid membrane simulations¹⁷. Supplementary Table 1 lists the composition of each model system. The CHARMM36 force field was used to model lipid and ion interactions^86,87. CHARMM TIP3P was used to model water interactions. Cholesterol and all lipids other than DPhPS were pre-defined within the CHARMM-GUI database of lipids and were used without modification. A force field for DPhPS was developed within the CHARMM-GUI ligand modeler⁸⁸ using the CHARMM General Force Field⁸⁹ (CGenFF v 4.6). Because DPhPS is not native to CHARMM-GUI’s Membrane Builder, the initial configuration of DPhPS-containing systems was alternatively generated using Packmol⁹⁰. In both Packmol and Membrane Builder, 50 water molecules per lipid molecule were added along with Na⁺ and Cl^- ions to bring the salt concentration to 150 mM. The number of water molecules was chosen to match the protocol of Garten and coworkers¹⁷ who previously analyzed structures of lipid membranes and their defects. Standard treatment of electrostatics in MD simulations requires the system to be net neutral in charge. Therefore, additional Na⁺ or Cl^- ions were used to neutralize the box in systems where charged lipids were used. Figure 2 illustrates two of the model systems. While the vesicles used in this work had average diameters of ~100 nm and the simulations were performed on flat bilayers, previous work has verified that defect size increases by only ~10% on 100 nm vesicles when compared to flat bilayers, regardless of lipid composition⁵⁴. Therefore, it can be assumed that the flat bilayers simulated in this work serve as accurate representations of their vesicular counterparts with 100 nm diameters.

MD simulation parameters

All MD simulations presented here were conducted in GROMACS 2021.6⁹¹. A standard Lennard-Jones 12-6 model with Lorentz-Berthelot mixing rules was used to for non-bonded interactions. Lennard-Jones interaction potentials were cut off at a distance of 1.2 nm with a force-switch modifier applied from 1.0 to 1.2 nm. Electrostatic interactions were calculated explicitly in normal space at short distances. Long-range electrostatic interactions were calculated using the particle mesh Ewald method. By default, GROMACS automatically adjusts the short-range electrostatic cutoff distance to optimize performance while maintaining a low error. The Verlet leap-frog integrator was used to integrate the equations of motion. An integration timestep of up to 2 fs was facilitated by constraining bonds between hydrogen atoms and heavy atoms using the LINCS algorithm. Thermostats and barostats were used to control temperature and pressure. During all steps, a velocity rescaling (v-rescale) thermostat was used. During some steps of equilibration, a cell-rescaling (c-rescale) barostat was used. A Parinello Rahman barostat was used for production simulations. Temperature was maintained at its set point of 303.15 K with a time constant of 1 ps. Regardless of which barostat was used, pressure was maintained at 1 bar with semi-isotropic coupling and a time constant of 5 ps and a compressibility of 4.5 × 10⁻⁵bar⁻¹. Net motion of the center of mass of the system was removed every 200 fs.

Minimization and equilibration

In addition to generating starting configurations, CHARMM-GUI provides a set of suggested parameters for energy minimization and six steps of equilibration. The suggested parameters were used as-is. During minimization and equilibration simulations, restraining potentials were applied to some lipid atoms to prevent integration from becoming unstable and to prevent the bilayer from deforming from non-optimal initial configurations. The strengths of the restraining potentials and the atoms to which the potentials were applied are listed in Supplementary Table 2. Energy minimization used a steepest descent algorithm for up to 5000 steps or until machine precision was reached. The first two steps of equilibration were conducted in the Canonical ensemble (NVT) for 125 ps each with an integration timestep of 1 fs. Between these steps, the restraining potentials were relaxed. The third step of equilibration was in the isothermal-isobaric ensemble (NPT) and was conducted for 125 ps with an integration timestep of 1 fs. Restraining potentials were further relaxed during this step. The C-rescale barsostat was turned on during the third step and remained turned on for the rest of the equilibration. In both the fourth and fifth steps, the restraining potentials are further relaxed and 250 to 500 ps of simulation were conducted in the NPT ensemble with an integration timestep of 2 fs. In the sixth step, restraints were removed and 500 ps of simulation are conducted in the NPT ensemble with an integration timestep of 2 fs. After equilibration, the box was inspected to determine whether the distance between periodic images of the lipid bilayer was at least 4.0 nm in the z direction. This distance ensured there were no artifacts from system size. Subsequent NPT simulations were used for data production.

MD data production and defect analysis

MD production simulations were conducted for 400 ns. Snapshots containing atomic coordinates of lipids were collected every 10 ps. The first 100 ns of data were excluded from analysis. Exclusion of the first 100 ns of data was meant to ensure that box volume and dimensions had converged to equilibrium values and to decouple the system coordinates from the initial structures produced by CHARMM-GUI or Packmol. Membrane defects were analyzed using the PackMem package⁵⁴. As recommended by the PackMem authors, three thousand evenly spaced simulation snapshots were extracted from the last 300 ns of the simulation trajectories (1 frame per 100 ps). PackMem was used to identify shallow, deep, and all defects within snapshots.

The PackMem software analyzes a top-down, two-dimensional representation of the lipid membrane, detecting areas where aliphatic atoms are exposed. Specifically, PackMem generates a 1 Å by 1 Å grid. Each grid cell is examined to determine whether the atom closest to the membrane surface is an aliphatic atom. If an aliphatic atom is detected before any other atoms, the cell is considered to have a packing defect. The depth of a defect is determined by the vertical distance between the aliphatic atom and the central glycerol atom on the given lipid chain⁵⁴. This vertical distance is compared to a threshold parameter, d. Any defect with a vertical distance less than or equal to d is considered a shallow defect. Any defect with a distance larger than d is considered a deep defect. In our analysis d was set to 1 Å. This distance has been shown to agree well with experimental observations⁵⁴. Because cholesterol differs in structure from a phospholipid, a central glycerol atom is not present. Cholesterol’s hydroxyl oxygen was used as a proxy to mimic the same analysis. The hydroxyl oxygen was chosen because it sits at nearly the same depth in the membrane as the surrounding lipid glycerol groups.

Defects were quantified with three different descriptors, namely percent defect coverage, defect size, and raw counts per frame. To examine the convergence of the defect descriptors, error bars are calculated by examining values calculated from the first third, middle third, and final third of the 3000 analyzed simulation frames. The sample standard deviation of these values is used to approximate the error as prescribed by the PackMem developers. Low error suggests that the simulation has reached an equilibrium state and that the membrane is no longer structurally evolving. The average defect coverage of the lipid membrane surface is calculated. This is done using the expected value for raw number of defects per frame and the expected size of defects (Eq. 3).

$$\% {A}_{{{\mathrm{covered}}},{{\mathrm{avg}}}} = 100 \% * \frac{E\left({A}_{{defect}}\right)E\left(C\right)}{{A}_{{lipid}}} \\ = 100 \% * \frac{{\sum}_{A\in {Areas}}{p}_{A}A{\sum}_{C\in {Counts}}{p}_{C}C}{{A}_{{lipid}}}$$

(3)

P_A denotes the probability any given defect will be of size A, and P_C denotes the probability a given frame will have C counts. E(A_defect) and E(C) represent the expected value for the size and number of defects per frame, respectively. These are used in conjunction with the computed average membrane surface area, A_lipid, to calculate the expected percentage of the membrane surface consisting of defects. The membrane surface area fluctuates by less than 1% after reaching equilibrium, thus this contribution to the error is not considered. Percentage defect coverage allows a straightforward comparison between membranes in terms of area accessible to the ɑ-synuclein. As recommended by the PackMem developers, defects with an area less than 15 Å² are excluded from analysis. Analysis of these small defects tends to be noisy, and convergence of values is difficult to attain. Because our analysis relies on raw defect counts rather than the fit distribution, all defects at or above the 15 Å² threshold are included. We present the “all” defects descriptor from PackMem which includes defects no matter how deep they are in the membrane. Experimentally observed increases in binding correlated with increases in the population and size of all defects in simulations. Simulated defect depth was not an explanatory factor for changes in experimentally measured binding.

Statistics and reproducibility

All details on statistics can be found in the figure captions. All sample sizes are indicated in the figure captions. For comparisons of two data sets, two-tailed unpaired student’s t-tests were performed. For sets with unequal variances, Welch’s t-test was performed. Significance analysis was performed in Excel and MATLAB. Significance levels and p values are provided in the figure captions. For each experimental condition within the tethered vesicle assay, a minimum of ten 35 µm by 35 µm square fields of view were acquired for image analysis.

Data availability

All processed data supporting the findings of this manuscript are included within the manuscript and supplementary information. Raw data for all figures is included in the Supplementary Data file. Fluorescence micrographs or simulation snapshots are available from the corresponding author upon reasonable request.

Code availability

CME Analysis, a free, publicly available MATLAB-associated package from the Danuser lab (University of Texas Southwestern Medical Center), was used to process all fluorescence micrographs from the tethered vesicle assay. MATLAB 2024a was used for all data processing and standard MATLAB statistical packages were used.

References

Maroteaux, L., Campanelli, J. T. & Scheller, R. H. Synuclein: a neuron-specific protein localized to the nucleus and presynaptic nerve terminal. J. Neurosci. 8, 2804 (1988).
Article CAS PubMed PubMed Central Google Scholar
Spillantini, M. G. et al. α-Synuclein in Lewy bodies. Nature 388, 839–840 (1997).
Article CAS PubMed Google Scholar
Spillantini, M., Crowther, R. A., Jakes, R., Hasegawa, M. & Goedert, M. α-synuclein in filamentous inclusions of Lewy bodies from Parkinson’s disease and dementia with Lewy bodies. Proc. Natl. Acad. Sci.USA 95, 6469–6473 (1998).
Article CAS PubMed PubMed Central Google Scholar
Forloni, G. Alpha synuclein: neurodegeneration and inflammation. Int. J. Mol. Sci. 24, 5914 (2023).
Article CAS PubMed PubMed Central Google Scholar
Sharma, M. & Burré, J. α-Synuclein in synaptic function and dysfunction. Trends Neurosci. 46, 153–166 (2023).
Article CAS PubMed Google Scholar
Sarchione, A., Marchand, A., Taymans, J.-M. & Chartier-Harlin, M.-C. Alpha-Synuclein and Lipids: the elephant in the room? Cells 10, 2452 (2021).
Gao, V., Briano, J. A., Komer, L. E. & Burré, J. Functional and pathological effects of α-synuclein on synaptic SNARE complexes. J. Mol. Biol. 435, 167714 (2023).
Article CAS PubMed Google Scholar
Burré, J., Sharma, M. & Südhof, T. α-Synuclein assembles into higher-order multimers upon membrane binding to promote SNARE complex formation. Proc. Natl. Acad. Sci. 111, https://doi.org/10.1073/pnas.1416598111 (2014).
Scott, D. & Roy, S. alpha-Synuclein inhibits intersynaptic vesicle mobility and maintains recycling-pool homeostasis. J. Neurosci. 32, 10129–10135 (2012).
Article CAS PubMed PubMed Central Google Scholar
Man, W. K. et al. The docking of synaptic vesicles on the presynaptic membrane induced by α-synuclein is modulated by lipid composition. Nat. Commun. 12, 927 (2021).
Article CAS PubMed PubMed Central Google Scholar
Lai, Y. et al. Neutral lysophosphatidylcholine mediates α-synuclein-induced synaptic vesicle clustering. Proc. Natl. Acad. Sci. USA 120, e2310174120 (2023).
Article CAS PubMed PubMed Central Google Scholar
Wang, C. et al. N-acetylation of α-synuclein enhances synaptic vesicle clustering mediated by α-synuclein and lysophosphatidylcholine. eLife 13, RP97228 (2024).
Article PubMed PubMed Central Google Scholar
Galvagnion, C. et al. Lipid vesicles trigger α-synuclein aggregation by stimulating primary nucleation. Nat. Chem. Biol. 11, 229–234 (2015).
Article CAS PubMed PubMed Central Google Scholar
Kurochka, A., Yushchenko, D., Bouř, P. & Shvadchak, V. Influence of Lipid membranes on α-synuclein aggregation. ACS Chem. Neurosci. 12, 825–830 (2021).
Article CAS PubMed Google Scholar
Zhu, M. & Fink, A. L. Lipid binding inhibits alpha-synuclein fibril formation. J. Biol. Chem. 278, 16873–16877 (2003).
Article CAS PubMed Google Scholar
Galvagnion, C. The role of lipids interacting with α-synuclein in the pathogenesis of Parkinson’s disease. J. Parkinson’s Dis. 7, 433–450 (2017).
Article CAS Google Scholar
Garten, M. et al. Methyl-branched lipids promote the membrane adsorption of alpha-synuclein by enhancing shallow lipid-packing defects. Phys. Chem. Chem. Phys. 17, 15589–15597 (2015).
Article CAS PubMed Google Scholar
Mahapatra, A., Mandal, N. & Chattopadhyay, K. Cholesterol in synaptic vesicle membranes regulates the vesicle-binding, function, and aggregation of α-synuclein. J. Phys. Chem. B 125, https://doi.org/10.1021/acs.jpcb.1c03533 (2021).
Middleton, E. R. & Rhoades, E. Effects of curvature and composition on α-synuclein binding to lipid vesicles. Biophys. J. 99, 2279–2288 (2010).
Article CAS PubMed PubMed Central Google Scholar
Stöckl, M., Fischer, P., Wanker, E. & Herrmann, A. α-synuclein selectively binds to anionic phospholipids embedded in liquid-disordered domains. J. Mol. Biol. 375, 1394–1404 (2008).
Article PubMed Google Scholar
Davidson, W. S., Jonas, A., Clayton, D. F. & George, J. M. Stabilization of α-synuclein secondary structure upon binding to synthetic membranes. J. Biol. Chem. 273, 9443–9449 (1998).
Article CAS PubMed Google Scholar
Jo, E., McLaurin, J., Yip, C. M., St George-Hyslop, P. & Fraser, P. E. alpha-Synuclein membrane interactions and lipid specificity. J. Biol. Chem. 275, 34328–34334 (2000).
Article CAS PubMed Google Scholar
Rhoades, E., Ramlall, T. F., Webb, W. W. & Eliezer, D. Quantification of α-synuclein binding to lipid vesicles using fluorescence correlation spectroscopy. Biophys. J. 90, 4692–4700 (2006).
Article CAS PubMed PubMed Central Google Scholar
Hellstrand, E. et al. Adsorption of α-synuclein to supported lipid bilayers: positioning and role of electrostatics. ACS Chem. Neurosci. 4, 1339–1351 (2013).
Article CAS PubMed PubMed Central Google Scholar
Manna, M. & Murarka, R. K. Polyunsaturated fatty acid modulates membrane-bound monomeric α-synuclein by modulating membrane microenvironment through preferential interactions. ACS Chem. Neurosci. 12, 675–688 (2021).
Article CAS PubMed Google Scholar
Nuscher, B., et al. α-synuclein has a high affinity for packing defects in a bilayer membrane: a thermodynamics study. J. Biol. Chem. 279, 21966–21975 (2004).
Article CAS PubMed Google Scholar
Ouberai, M. M. et al. α-synuclein senses lipid packing defects and induces lateral expansion of lipids leading to membrane remodeling. J. Biol. Chem. 288, 20883–20895 (2013).
Article CAS PubMed PubMed Central Google Scholar
Andersson, A., Linse, S., Sparr, E., Fornasier, M. & Jonsson, P. The density of anionic lipids modulates the adsorption of alpha-Synuclein onto lipid membranes. Biophys. Chem. 305, 107143 (2024).
Article CAS PubMed Google Scholar
Pirc, K. & Ulrih, N. α-Synuclein interactions with phospholipid model membranes: key roles for electrostatic interactions and lipid-bilayer structure. Biochim. et. Biophys. Acta (BBA) - Biomembranes 1848, 2002–2012 (2015).
Article CAS Google Scholar
Chandra, S., Chen, X., Rizo, J., Jahn, R. & Südhof, T. C. A broken α-helix in folded α-synuclein. J. Biol. Chem. 278, 15313–15318 (2003).
Article CAS PubMed Google Scholar
Bigay, J. & Antonny, B. Curvature, lipid packing, and electrostatics of membrane organelles: defining cellular territories in determining specificity. Dev. Cell 23, 886–895 (2012).
Article CAS PubMed Google Scholar
Zhao, C. et al. Lysophosphatidylcholine binds α-synuclein and prevents its pathological aggregation. Natl Sci. Rev. 11, https://doi.org/10.1093/nsr/nwae182 (2024).
Maltseva, S., Kerr, D., Turke, M., Adams, E. J. & Lee, K. Y. C. Parkinson’s disease-associated mutations in α-synuclein alters its lipid-bound state. Biophys. J. 123, 1610–1619 (2024).
Article CAS PubMed PubMed Central Google Scholar
Ysselstein, D. et al. Effects of impaired membrane interactions on α-synuclein aggregation and neurotoxicity. Neurobiol. Dis. 79, 150–163 (2015).
Article CAS PubMed PubMed Central Google Scholar
Manzanza, N. O., Sedlackova, L. & Kalaria, R. N. Alpha-synuclein post-translational modifications: implications for pathogenesis of lewy body disorders. Front. Aging Neurosci. 13, 690293 (2021).
Article CAS PubMed PubMed Central Google Scholar
Zhang, J., Li, X. & Li, J.-D. The roles of post-translational modifications on α-synuclein in the pathogenesis of Parkinson’s diseases. Front. Neurosci. 13, https://doi.org/10.3389/fnins.2019.00381 (2019).
Ohgita, T., Namba, N., Kono, H., Shimanouchi, T. & Saito, H. Mechanisms of enhanced aggregation and fibril formation of Parkinson’s disease-related variants of α-synuclein. Sci. Rep. 12, 6770 (2022).
Article CAS PubMed PubMed Central Google Scholar
Arnesen, T. et al. Proteomics analyses reveal the evolutionary conservation and divergence of N-terminal acetyltransferases from yeast and humans. Proc. Natl. Acad. Sci. USA 106, 8157–8162 (2009).
Article CAS PubMed PubMed Central Google Scholar
Burré, J. et al. Properties of native brain α-synuclein. Nature 498, E4–E6 (2013).
Article PubMed PubMed Central Google Scholar
Fauvet, B. et al. α-synuclein in central nervous system and from erythrocytes, mammalian cells, and escherichia coli exists predominantly as disordered monomer. J. Biol. Chem. 287, 15345–15364 (2012).
Article CAS PubMed PubMed Central Google Scholar
Theillet, F.-X. et al. Structural disorder of monomeric α-synuclein persists in mammalian cells. Nature 530, 45–50 (2016).
Article CAS PubMed Google Scholar
Bartels, T., Choi, J. & Selkoe, D. α-Synuclein occurs physiologically as a helically folded tetramer that resists aggregation. Nature 477, 107–110 (2011).
Article CAS PubMed PubMed Central Google Scholar
Bell, R., et al. Effects of N-terminal acetylation on the aggregation of disease-related α-synuclein Variants. J. Mol. Biol. 435, 167825 (2023).
Article CAS PubMed Google Scholar
Bell, R. et al. N-terminal acetylation of α-Synuclein slows down its aggregation process and alters the morphology of the resulting aggregates. Biochemistry 61, 1743–1756 (2022).
Article CAS PubMed Google Scholar
Yang, X., Wang, B., Hoop, C., Williams, J. & Baum, J. NMR unveils an N-terminal interaction interface on acetylated-α-synuclein monomers for recruitment to fibrils. Proc. Natl. Acad. Sci.USA 118, https://doi.org/10.1073/pnas.2017452118 (2021).
Dikiy, I. & Eliezer, D. N-terminal Acetylation Stabilizes N-terminal Helicity in Lipid- and Micelle-bound α-synuclein and increases its affinity for physiological membranes. J. Biol. Chem. 289, 3652–3665 (2014).
Article CAS PubMed Google Scholar
Iyer, A. et al. The iImpact of N-terminal acetylation of α-synuclein on phospholipid membrane binding and fibril structure. J. Biol. Chem. 291, 21110–21122 (2016).
Article CAS PubMed PubMed Central Google Scholar
O’Leary, E., Jiang, Z., Strub, M.-P. & Lee, J. Effects of phosphatidylcholine membrane fluidity on the conformation and aggregation of N-terminally acetylated α-synuclein. J. Biol. Chem. 293, 11195–11205 (2018).
Article PubMed PubMed Central Google Scholar
Maltsev, A. S., Ying, J. & Bax, A. Impact of N-Terminal acetylation of α-synuclein on its random coil and lipid binding properties. Biochemistry 51, 5004–5013 (2012).
Article CAS PubMed Google Scholar
Runfola, M., De Simone, A., Vendruscolo, M., Dobson, C. M. & Fusco, G. The N-terminal acetylation of α-synuclein changes the affinity for lipid membranes but not the structural properties of the bound state. Sci. Rep. 10, 204 (2020).
Article CAS PubMed PubMed Central Google Scholar
Hatzakis, N. S. et al. How curved membranes recruit amphipathic helices and protein anchoring motifs. Nat. Chem. Biol. 5, 835–841 (2009).
Article CAS PubMed Google Scholar
Johnson, D., Kou, O., Bouzos, N. & Zeno, W. Protein–membrane interactions: sensing and generating curvature. Trends Biochem. Sci. 49, 401–416 (2024).
Article CAS PubMed PubMed Central Google Scholar
Binotti, B., Jahn, R. & Pérez-Lara, Á An overview of the synaptic vesicle lipid composition. Arch. Biochem. Biophys. 709, 108966 (2021).
Article CAS PubMed Google Scholar
Gautier, R. et al. PackMem: a versatile tool to compute and visualize interfacial packing defects in lipid bilayers. Biophys. J. 115, 436–444 (2018).
Article CAS PubMed PubMed Central Google Scholar
Greenfield, N. J. Using circular dichroism spectra to estimate protein secondary structure. Nat. Protoc. 1, 2876–2890 (2006).
Article CAS PubMed PubMed Central Google Scholar
Bazinet, R. P. & Layé, S. Polyunsaturated fatty acids and their metabolites in brain function and disease. Nat. Rev. Neurosci. 15, 771–785 (2014).
Article CAS PubMed Google Scholar
Takamori, S. et al. Molecular anatomy of a trafficking organelle. Cell 127, 831–846 (2006).
Article CAS PubMed Google Scholar
Tang, Z. et al. Folding of N-terminally acetylated α-synuclein upon interaction with lipid membranes. Biophys. J. 123, 3698–3720 (2024).
Article CAS PubMed Google Scholar
Ree, R., Varland, S. & Arnesen, T. Spotlight on protein N-terminal acetylation. Exp. Mol. Med. 50, 1–13 (2018).
Article CAS PubMed PubMed Central Google Scholar
Lokappa, S. B. & Ulmer, T.S. alpha-Synuclein Populates Both Elongated and Broken Helix States on Small Unilamellar Vesicles. J. Biol. Chem. 286, 21450–21457 (2011).
Article CAS PubMed PubMed Central Google Scholar
Jao, C. C., Hegde, B. G., Chen, J., Haworth, I. S. & Langen, R. Structure of membrane-bound alpha-synuclein from site-directed spin labeling and computational refinement. Proc. Natl. Acad. Sci. USA 105, 19666–19671 (2008).
Article CAS PubMed PubMed Central Google Scholar
Trexler, A. J. & Rhoades, E. α-synuclein binds large unilamellar vesicles as an extended helix. Biochemistry 48, 2304–2306 (2009).
Article CAS PubMed Google Scholar
Kang, C. & Sun, R. Molecular dynamics study of the interaction between the N-terminal of α-synuclein and a lipid bilayer mimicking synaptic vesicles. J. Phys. Chem. B 125, 1036–1048 (2021).
Article CAS PubMed Google Scholar
Liu, J. et al. Membrane packing defects in synaptic vesicles recruit complexin and synuclein. Phys. Chem. Chem. Phys. 23, 2117–2125 (2021).
Article CAS PubMed Google Scholar
Qi, Z., Peng, R., Zhang, J. & Li, Z. Membrane Binding of α-Synuclein Influenced by the Shortening of Phospholipid Acyl Chains. ResearchGate https://doi.org/10.13140/RG.2.2.27706.71367/1 (2025).
Qi, Z., Wan, M., Zhang, J. & Li, Z. Influence of cholesterol on the membrane binding and conformation of α-synuclein. J. Phys. Chem. B 127, https://doi.org/10.1021/acs.jpcb.2c08077 (2023).
Man, W. K. et al. A role of cholesterol in modulating the binding of alpha-synuclein to synaptic-like vesicles. Front. Neurosci. 14, 18 (2020).
Article PubMed PubMed Central Google Scholar
Jacob, R. S. et al. α-Synuclein acts as a cholesteryl-ester sensor on lipid droplets regulating organelle size and abundance. bioRxiv, 2024.2006.2019.599670, https://doi.org/10.1101/2024.06.19.599670 (2024).
Vanni, S. et al. Amphipathic lipid packing sensor Motifs: probing bilayer defects with hydrophobic residues. Biophys. J. 104, 575–584 (2013).
Article CAS PubMed PubMed Central Google Scholar
Kulakowski, G. et al. Lipid packing defects and membrane charge control RAB GTPase recruitment. Traffic 19, 536–545 (2018).
Article CAS PubMed PubMed Central Google Scholar
Yakunin, E. et al. Synuclein neuropathology is controlled by nuclear hormone receptors and enhanced by docosahexaenoic acid in a mouse model for Parkinson’s disease. Brain Pathol. 22, 280–294 (2012).
Article CAS PubMed Google Scholar
Sharon, R., Bar-Joseph, I., Mirick, G., Serhan, C. & Selkoe, D. Altered fatty acid composition of dopaminergic neurons expressing α-synuclein and human brains with α-synucleinopathies. J. Biol. Chem. 278, 49874–49881 (2003).
Article CAS PubMed Google Scholar
Broersen, K., van den Brink, D., Fraser, G., Goedert, M. & Davletov, B. α-synuclein adopts an α-helical conformation in the presence of polyunsaturated fatty acids to hinder micelle formation. Biochemistry 45, 15610–15616 (2006).
Article CAS PubMed Google Scholar
De Franceschi, G. et al. Structural and morphological characterization of aggregated species of α-synuclein induced by docosahexaenoic acid. J. Biol. Chem. 286, 22262–22274 (2011).
Article PubMed PubMed Central Google Scholar
Eisenberg, S. et al. Mapping the electrostatic profiles of cellular membranes. Mol. Biol. Cell 32, 301–310 (2021).
Article CAS PubMed PubMed Central Google Scholar
Colla, E. Linking the endoplasmic reticulum to Parkinson’s disease and Alpha-Synucleinopathy. Front. Neurosci. 13, https://doi.org/10.3389/fnins.2019.00560 (2019).
Aliakbari, F. et al. The potential of zwitterionic nanoliposomes against neurotoxic alpha-synuclein aggregates in Parkinson’s Disease. Nanoscale 10, 9174–9185 (2018).
Article CAS PubMed Google Scholar
Momin, N. et al. Designing lipids for selective partitioning into liquid ordered membrane domains. Soft Matter 11, 3241–3250 (2015).
Article CAS PubMed Google Scholar
Johnson, M., Coulton, A. T., Geeves, M. A. & Mulvihill, D. P. Targeted Amino-terminal acetylation of recombinant proteins in E. coli. PLoS ONE 5, e15801 (2010).
Article CAS PubMed PubMed Central Google Scholar
Hung, J. J. et al. Self-diffusion of a highly concentrated monoclonal antibody by fluorescence correlation spectroscopy: insight into protein–protein interactions and self-association. Soft Matter 15, 6660–6676 (2019).
Article CAS PubMed Google Scholar
Zeno, W. F. et al. Synergy between intrinsically disordered domains and structured proteins amplifies membrane curvature sensing. Nat. Commun. 9, 4152 (2018).
Article PubMed PubMed Central Google Scholar
Aguet, F., Costin, Mettlen, M., Sandra & Danuser, G. Advances in analysis of low signal-to-noise images link dynamin and AP2 to the functions of an endocytic checkpoint. Dev. Cell 26, 279–291 (2013).
Article CAS PubMed PubMed Central Google Scholar
Jo, S., Kim, T., Iyer, V. & Im, W. CHARMM-GUI: a web-based graphical user interface for CHARMM. J. Comput. Chem. 29, 1859–1865 (2008).
Article CAS PubMed Google Scholar
Lee, J. et al. CHARMM-GUI input generator for NAMD, GROMACS, AMBER, OpenMM, and CHARMM/OpenMM simulations using the CHARMM36 additive force field. J. Chem. Theory Comput. 12, 405–413 (2015).
Article PubMed PubMed Central Google Scholar
Wu, E. et al. CHARMM-GUI Membrane Builder toward realistic biological membrane simulations. J. Comput. Chem. 35, 1997–2004 (2014).
Article CAS PubMed PubMed Central Google Scholar
Klauda, J. et al. Update of the CHARMM all-atom additive force field for lipids: validation on six lipid types. J. Phys. Chem. B 114, 7830–7843 (2010).
Article CAS PubMed PubMed Central Google Scholar
Yu, Y. et al. CHARMM36 lipid force field with explicit treatment of long-range dispersion: parametrization and validation for phosphatidylethanolamine, phosphatidylglycerol, and ether lipids. J. Chem. Theory Comput. 17, 1581–1595 (2021).
Article CAS PubMed PubMed Central Google Scholar
Kim, S. et al. CHARMM-GUI ligand reader and modeler for CHARMM force field generation of small molecules. J. Comput. Chem. 38, 1879–1886 (2017).
Article CAS PubMed PubMed Central Google Scholar
Vanommeslaeghe, K. et al. CHARMM general force field: A force field for drug-like molecules compatible with the CHARMM all-atom additive biological force fields. J. Comput. Chem. 31, 671–690 (2009).
Article Google Scholar
Martínez, L., Andrade, R., Birgin, E. G. & Martínez, J. M. P ACKMOL : A package for building initial configurations for molecular dynamics simulations. J. Comput. Chem. 30, 2157–2164 (2009).
Article PubMed Google Scholar
Abraham, M. J. et al. GROMACS: high performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX 1, 19–25 (2015).
Article Google Scholar

Download references

Acknowledgements

Thank you to Jennifer C. Lee for generously providing the pRK172 human αSyn(Y136C) plasmid. Thank you to Brian H. Kim for assisting with the data organization and management. This research was supported by the National Institutes of Health through R35GM147333 to D.H.J., O.H.K., and W.F.Z; R35GM150716 to A.M. and P.J.C.; and R21EY034674 to V.W.J.

Author information

These authors contributed equally: David H. Johnson, Orianna H. Kou.

Authors and Affiliations

Mork Family Department of Chemical Engineering and Materials Science, University of Southern California, Los Angeles, CA, USA
David H. Johnson & Wade F. Zeno
Department of Physics and Astronomy, University of Southern California, Los Angeles, CA, USA
Orianna H. Kou, Antonis Margaritakis & Peter J. Chung
Department of Chemical Engineering, University of Louisville, Louisville, KY, USA
John M. White & Vance W. Jaeger
Department of Biological Sciences, University of Southern California, Los Angeles, CA, USA
Stephanie Y. Ramirez
Department of Chemistry, University of Southern California, Los Angeles, CA, USA
Peter J. Chung
Alfred E. Mann Department of Biomedical Engineering, University of Southern California, Los Angeles, CA, USA
Peter J. Chung

Authors

David H. Johnson
View author publications
Search author on:PubMed Google Scholar
Orianna H. Kou
View author publications
Search author on:PubMed Google Scholar
John M. White
View author publications
Search author on:PubMed Google Scholar
Stephanie Y. Ramirez
View author publications
Search author on:PubMed Google Scholar
Antonis Margaritakis
View author publications
Search author on:PubMed Google Scholar
Peter J. Chung
View author publications
Search author on:PubMed Google Scholar
Vance W. Jaeger
View author publications
Search author on:PubMed Google Scholar
Wade F. Zeno
View author publications
Search author on:PubMed Google Scholar

Contributions

D.H.J., O.H.K., P.J.C., and W.F.Z. designed experiments. D.H.J., O.H.K., S.Y.R., and A.M. performed experiments. J.M.W. and V.WJ. designed and performed the simulations. All authors consulted together on the interpretation of results and manuscript preparation.

Corresponding author

Correspondence to Wade F. Zeno.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Communications Biology thanks the anonymous reviewers for their contribution to the peer review of this work. Primary Handling Editors: Lei Zheng and Tobias Goris. [A peer review file is available.]

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Transparent Peer Review file

Supplementary Information

Description of Additional Supplementary Files

Supplementary Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Johnson, D.H., Kou, O.H., White, J.M. et al. Lipid packing defects are necessary and sufficient for membrane binding of α-synuclein. Commun Biol 8, 1179 (2025). https://doi.org/10.1038/s42003-025-08622-7

Download citation

Received: 23 January 2025
Accepted: 28 July 2025
Published: 07 August 2025
DOI: https://doi.org/10.1038/s42003-025-08622-7