Polyglutamine aggregates impair lipid membrane integrity and enhance lipid membrane rigidity

    Polyglutamine aggregates impair lipid membrane integrity and enhance lipid membrane rigidity Chian Sing Ho, Nawal K. Khadka, Fengyu She, Jianfeng Cai, Jianjun Pan PII: DOI: Reference:

S0005-2736(16)30015-3 doi: 10.1016/j.bbamem.2016.01.016 BBAMEM 82110

To appear in:

BBA - Biomembranes

Received date: Revised date: Accepted date:

30 October 2015 8 January 2016 20 January 2016

Please cite this article as: Chian Sing Ho, Nawal K. Khadka, Fengyu She, Jianfeng Cai, Jianjun Pan, Polyglutamine aggregates impair lipid membrane integrity and enhance lipid membrane rigidity, BBA - Biomembranes (2016), doi: 10.1016/j.bbamem.2016.01.016

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

PT

Polyglutamine Aggregates Impair Lipid Membrane Integrity and Enhance Lipid Membrane Rigidity Chian Sing Ho1, Nawal K. Khadka1, Fengyu She2, Jianfeng Cai2, and Jianjun Pan1* 1

RI

Department of Physics, University of South Florida, Tampa, FL 33620 Department of Chemistry, University of South Florida, Tampa, FL 33620

NU

SC

2

MA

*To whom correspondence should be addressed:

AC CE P

TE

D

Jianjun Pan, Ph.D. E-mail: [email protected]

Keywords: polyglutamine, amyloid, oligomer, lipid bilayer, solution AFM, fluorescence microscopy

1

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Abstract Lipid membranes are suggested as the primary target of amyloid aggregates. We study aggregates formed by a polyglutamine (polyQ) peptide, and their disruptive effect on lipid membranes. Using solution atomic force microscopy (AFM), we observe polyQ oligomers coexisting with short fibrils, which have a twisted morphology that likely corresponds to two intertwined oligomer strings. Fourier transform infrared spectroscopy reveals that the content of β-sheet enriched aggregates increases with incubation time. Using fluorescence microscopy, we find that exposure to polyQ aggregates results in deflated morphology of giant unilamellar vesicles. PolyQ aggregates induced membrane disruption is further substantiated by timedependent calcein leakage from the interior to the exterior of lipid vesicles. Detailed structural and mechanical perturbations of lipid membranes are revealed by solution AFM. We find that membrane disruption by polyQ aggregates proceeds by a two-step process, involving partial and full disruption. In addition to height contrast, the resulting partially and fully disrupted bilayers have distinct rigidity and adhesion force properties compared to the intact bilayer. Specifically, the bilayer rigidity increases as the intact bilayer becomes partially and fully disrupted. Surprisingly, the adhesion force first decreases and then increases during the disruption process. By resolving individual fibrils deposited on bilayer surface, we show that both the length and the number of fibrils can increase with incubation time. Our results highlight that membrane disruption could be the molecular basis of polyQ aggregates induced cytotoxicity.

2

ACCEPTED MANUSCRIPT

SC

RI

PT

Introduction To date, it is known that misfolding of expanded polyglutamine (polyQ) stretches is responsible for at least ten types of neurodegenerative disorders in human [1]. Among them, Huntington’s disease (HD) is the most notable one. HD is caused by an abnormal polyQ stretch located near the N-terminus of the protein huntingtin (htt). It was found that the risk of developing HD, as well as the severity of the disease, is greatly enhanced when the number of glutamine repeats reaches a pathogenic threshold [2]. Numerous biochemical and biophysical studies have indicated that polyQ stretches have the generic amyloidogenetic propensity of forming protease resistant aggregates, a common theme shared by many amyloid proteins and peptides, including amyloid-β in Alzheimer’s disease, α-synuclein in Parkinson’s disease, islet amyloid polypeptide in type II diabetes mellitus, and prion protein in Creutzfeldt-Jakob disease.

MA

NU

One hallmark of HD is the formation of fibrillar inclusion bodies in intranuclear and cytoplasmic spaces. Kinetic studies using model peptides have established that polyQ fibrils are formed following classic nucleated polymerization mechanism [3, 4]. Additional studies suggested that the nucleus size is dependent on polyQ repeat length [5]; flanking sequences associated with htt can further alter polyQ aggregation behavior and kinetics [6, 7]. The existence of polymorphic aggregates complicates structure-dysfunction relationship in polyQ associated disorders, i.e., what is the molecular species responsible for cytotoxicity along the pathogenic pathway [8]? Indeed, both mature fibrils [9-11] and elusive oligomers [12-18] have been documented as the toxic determinant.

AC CE P

TE

D

Studies comparing molecular bases of amyloidogenic diseases indicated that despite sequence diversity, all amyloid oligomers share a common conformation that is ubiquitously recognized by specifically designed antibodies [13]. Electrophysiological measurements revealed that all soluble oligomers increased membrane ion conductivity, while fibrils had no such effect. Based on these observations, a common mechanism of membrane targeting and permeabilization was conceptualized [19]. This view was reinforced by observations that amyloid pathogenesis can be caused by aggregates localized either in the extracellular or cytosolic space. In addition, cytotoxicity was conserved when cytosolic aggregates were administered externally [13, 19]. In line with the membrane permeabilization mechanism, many amyloidogogenic proteins and peptides were found to reorganize membrane structure, such as α-synuclein [20], amyloid-β [21], islet amyloid polypeptide [22], and prion protein fragment [23]. Membrane disruption by polyQ aggregates has also been observed. In particular, both polyQ repeat length [24] and flanking sequence [25] were found to play a role in modulating membrane integrity. Despite the accumulating knowledge, molecular basis of polyQ induced membrane disruption remain poorly understood. For example, it is unclear what molecular species of polyQ causes membrane depolarization. Large amorphous polyQ aggregates that do not correspond well to either oligomeric or fibrillar structures were observed on bilayer surfaces [24]. Are these amorphous aggregates causing lipid bilayer disruption, or are they just loosely associated with bilayer surface due to electrostatic interactions? Similarly, large aggregates of polyQ with betterdefined morphologies have been observed on bilayer surfaces [25]. The authors referred to these aggregates as “oligomers”. One intriguing fact is that these “oligomers” are much larger compared to soluble oligomers observed on mica surface. How are these “oligomers” generated? Are they responsible for lipid bilayer permeabilization? AFM based studies have shown that polyQ aggregates enhance lipid membrane roughness, a structural feature that might be related to lipid membrane depolarization [24, 25]. However, it is unclear what species is causing membrane roughening. Finally, the details of polyQ aggregates induced membrane disruption are unknown. Are membranes permeabilized by amyloid pores [26] or by nonspecific membrane disruption [19]? The molecular mechanism of membrane disruption is further

3

ACCEPTED MANUSCRIPT confounded by observations that membrane surface can alter amyloid aggregation [27, 28]. By suppressing aggregates’ mobility, lipid membranes can either catalyze amyloid aggregation [29] or reverse the aggregated product [30].

D

MA

NU

SC

RI

PT

Here we use a simple lipid bilayer system to study the interplay between polyQ aggregates and the underlying lipid matrix. We first use solution AFM to obtain structures of polyQ aggregates generated by our modified preparation protocol. Our solution AFM measurement shows that globular oligomers and short fibrils coexist on mica surface. The secondary structure of polyQ as a function of incubation time is revealed by attenuated total reflection Fourier transformation infrared (FTIR) spectroscopy. To qualitatively assess membrane perturbation imparted by polyQ aggregates, we use fluorescence microscopy of ~30 μm sized giant unilamellar vesicles (GUVs) and fluorescence spectroscopy of ~120 nm sized unilamellar vesicles (ULVs). Membrane perturbation is confirmed by morphological changes of GUVs and time-dependent calcein leakage from ULVs. Finally, we use solution AFM to investigate polyQ induced structural and mechanical perturbation of mica supported planar lipid bilayers. We find that membrane disruption is a two-step process involving partial and full disruption. The resulting partially and fully disrupted bilayers have district structural, mechanical, and adhesion force properties. Our solution AFM also shows individual fibrils deposited on bilayer surfaces. Interestingly, we find that bilayer associated fibrils can grow in length. Based on our high-resolution data, we speculate that polyQ oligomers are the toxic species causing lipid bilayer disruption. Our experimental data highlight that membrane disruption by polyQ aggregates can be the molecular basis of polyQ induced cytotoxicity.

AC CE P

TE

Materials and Methods 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and rhodamine-DPPE (1,2-dipalmitoyl-snglycero-3-phosphoethanolamine), in powder form, were purchased from Avanti Polar Lipids (Alabaster, AL). All other reagents were purchased from Fisher Scientific (Pittsburgh, PA). PolyQ35 peptide (KK-Q35-KK) was synthesized by solid phase synthesis method. To enhance water solubility, the polyQ stretch was flanked by two lysine residues at the N- and C-termini. The flanking residues likely modify aggregation kinetics [31], although the resulting aggregates seem to be unchanged [32]. Synthesized peptide was purified by HPLC, followed by lyophilization. Peptide molecular weight was confirmed by a mass spectrometer (Bruker AutoFlex MALDI-TOF). PolyQ35 aggregates were prepared by dissolving peptide powder (100 μM) in 10 mM HEPES at pH 7.4. The stock solution was kept at 4.0oC. Aliquots incubated for different days were used for FTIR, fluorescence microscopy, and solution AFM measurements. (All measurements were performed at room temperature). For calcein leakage experiment, peptide aliquots were collected at different time points, snap-frozen in liquid nitrogen, and stored at -80oC. Ice-thawed aliquots were mixed with calcein-encapsulated ULVs for calcein leakage assay. Attenuated total refection FTIR Attenuated total reflection FTIR measurements at room temperature (23oC) were performed using a Vertex 70 Bio ATRII spectrometer (Bruker, Billerica, MA). A fresh aliquot (30 μl) of polyQ35 incubated for different days at 4oC was loaded into the silicon crystal sample cell. For pelleted aggregates, an aliquot incubated for 7 days was centrifuged at 13000 rpm for 10 min. The pellet was washed for 3 times before bringing to the initial volume. Infrared absorption spectra were collected at wavenumbers from 400 to 4000 cm-1 with a resolution of 2 cm-1 (averaged over 1500 scans). The time takes for one dataset collection is ~40 min. Spectra were corrected for background absorption arising from the 10 mM HEPES buffer.

4

ACCEPTED MANUSCRIPT

NU

SC

RI

PT

Fluorescence microscopy DOPC+0.2mol% rhodamine-DPPE were prepared by mixing appropriate amount of lipid stock solutions. After depositing onto ITO coated glass slides, lipid dry films were vacuumed for > 2 h to remove organic solvents. A GUV production chamber filled with 100 mM sucrose and 10 mM HEPES (pH 7.4) was formed by sealing lipid dry films with a Viton O-ring. The assemblage was heated to 60oC using a Digi-Sense temperature controller. GUVs were generated by lipid swelling under an AC field of 10 Hz and 2.0 Volt (duration of 2 h). GUV solution was transferred to 3 mL buffer containing 100 mM glucose and 10 mM HEPES (pH 7.4). After settling for > 1 h, GUV aliquot from the bottom was transferred onto a coverslip and examined by fluorescence microscopy using a Nikon Eclipse-U microscope and an Andor iXon Ultra 897 EM-CCD. To retard liquid evaporation, the cover slip was housed in a homemade hollow chamber sealed by copper plates and thermal insulation foam. To study the effect of polyQ35 aggregates, an equal volume of peptide solution (after adjusting glucose concentration to 100 mM) was mixed with a GUV aliquot (final peptide concentration was 15 μM). Immediately after mixing, time-lapse or single-shot fluorescence micrographs were collected. Control experiments were performed by mixing buffer-only solution with the GUV aliquot.

AC CE P

TE

D

MA

Calcein leakage assay DOPC dry film was hydrated by 1 mL 10 mM HEPES buffer (pH 7.4) containing 20 mM NaCl and 30 mM calcein. Freeze-thaw cycles (6 times total) between -80 and 50oC were carried out to ensure uniform distribution of calcein. The resultant lipid suspension was extruded using an Avanti mini-extruder outfitted with a 100-nm-diameter pore filter to produce ULVs. An ÄKTA pure FPLC (GE Healthcare, Piscataway, NJ) and a gel filtration column (Superdex 200 10/300 GL) were used to remove exterior calcein. The elution buffer contained 108 mM NaCl and 10 mM HEPES pH 7.4 (elution buffer). The equality of the osmotic pressure in the interior and at the exterior of calcein-encapsulated ULVs was confirmed using a Wescore 5500 vapor pressure osmometer (Logan, Utah). Dynamic light scattering measurement using a Dynapro Nanostar (Wyatt Technology, Santa Barbra, CA) indicated that the size exclusion chromatography (SEC) separated vesicles had an average diameter of 120 nm. Snap-frozen polyQ35 aliquots collected at different time points were thawed on ice. In addition to these aliquots, we also investigated the effect of pelleted aggregates. This was done by centrifuging 100 μL peptide solution incubated for 7 days at 13000 rpm for 10 min. Peptide pellet was washed by SEC elution buffer for 3 times before bringing the final volume to 100 μL. ULV sample was prepared by mixing 100 μL peptide solution (100 μM) with 880 μl SEC elution buffer; 20 μl calcein-encapsulated ULVs was added at last (final peptide concentration was 10 μM). The mixture was immediately transferred into a 1 mL quartz cuvette. Time-course fluorescence spectrum (23oC) was obtained using a Jasco FP-8300 fluorometer (Easton, MD). The excitation and emission wavelengths were 494 and 515 nm, respectively, with a bandwidth of 2.5 nm for each wavelength. Fluorescence signal I was collected every 20 s until the intensity reached a plateau. To obtain the maximum intensity Imax when all calcein molecules become unenclosed, an aliquot of Triton X-100 was added to the peptide treated ULV sample. The percentage of calcein leakage is defined as: , where I0 is the ma fluorescence intensity from calcein-encapsulated ULVs without peptide treatment. Solution AFM A liquid-compatible Multimode 8 AFM (Bruker, Santa Barbara, CA) operated at the PeakForce quantitative nanomechanics (QNM) mode was used for solution AFM study (23oC). Fresh peptide aliquot (incubated for different days at 4oC) was diluted to 10 μM (10 mM HEPES at pH 7.4) and loaded into the AFM liquid cell, which contained a freshly cleaved mica substrate. After equilibration (~10 min), peptide aggregates were scanned (in solution) using a special silicon

5

ACCEPTED MANUSCRIPT nitride tip that works with the PeakForce QNM mode (model: Scanasyst-Fluid+). Square images were acquired at a scan rate of 0.6 Hz.

NU

SC

RI

PT

We also study the effect of polyQ35 aggregates on mica supported lipid bilayers. Experimental procedure for lipid bilayer preparation has been described before [33, 34]. Briefly, lipid small unilamellar vesicles (SUVs) in 10 mM HEPES pH 7.4 were generated by ultrasonication using a Sonic Dismembrator. A mica supported lipid bilayer was formed by injecting SUVs into the AFM liquid cell. After bilayer formation (~30 min, examined by preliminary AFM scans), polyQ35 aggregates (20 μM) were injected into the liquid cell using a syringe pump (50 μL/min). In situ AFM scans using PeakForce QNM mode were performed with a peak force of ~300 pN. During each scan, the probe oscillates at 1 kHz while it scans across the bilayer surface. On fly analysis of the oscillating force curves at each position yields a map of mechanical property (i.e., the reduced Young’s modulus E*) and another map of the adhesion force Fadh between the bilayer and the AFM tip (Fig. S1, SI). Both maps have the same spatial resolution as the height image. All maps (e.g., height, reduced Young’s modulus, and adhesion force) with the same resolution were collected in a single scan.

AC CE P

TE

D

MA

Results Oligomer and fibril morphologies of polyQ35 revealed by solution AFM

Figure 1 PolyQ35 aggregates on mica surface obtained by solution AFM. (A) Height image. Scale bar = 50 nm. (B) Height profiles along the six trajectories indicated in (A). (Only the central portion containing aggregates is shown.) Globular oligomers have a height of ~2 nm and a width of ~10 nm; fibrils have a height of ~4 nm and a width of ~12 nm. Twisted morphology is seen for the fibril with a periodicity of ~13 nm along the fibril axis. Amyloidogenic peptides can undergo different aggregation pathways, yielding polymorphic oligomers, prefibrils, and fibrils [35]. Monomeric disaggregation using organic solvents is typically used to study aggregation kinetics of polyQ peptides [36]. Here we are only interested in different forms of polyQ35 aggregates interacting with lipid bilayers; therefore, we prepare polyQ35 aggregates by directly dissolving lyophilized powder in HEPES buffer (pH 7.4) without the intervening disaggregation process. To slow down polyQ aggregation, we keep peptide stock solution at 4oC when not in use. By taking aliquots incubated for different days at 4oC, we are able to obtain different proportions of polyQ35 aggregates (i.e., oligomers and fibrils).

6

ACCEPTED MANUSCRIPT

RI

PT

High-resolution structures of polyQ35 aggregates are obtained by solution AFM (Fig. 1). Height profiles indicate that oligomers have a height of ~2 nm and an apparent width of ~10 nm. We use a spherical model to correct oligomer width influenced by the finite size of the AFM tip (nominal radius of 2 nm): W oligomer=W 2/8rtip, where W is the apparent width. The corrected oligomer width is 6.2 nm, which gives rise to an oligomer volume, 2π(W oligomer/2)2h/3, of 40 nm3. To estimate the number of monomers per oligomer, we need to know monomer volume. Assuming a density range of 1.2 to 1.5 g/ml for polyQ35 [37], the resulting monomer volume is 3.8 to 3.0 nm3. This gives rise to an estimation of ~10 to 13 monomers per oligomer. We note that in addition to peptide density, AFM tip size also contributes greatly to the uncertainty of the estimated number of monomers per oligomer.

AC CE P

TE

D

MA

NU

SC

Fibrils are observed to coexist with oligomers (Fig. 1). Height profiles in two directions indicate that fibrils have a height of ~4 nm, about twice of the oligomer height. Moreover, twisted morphology is revealed by periodic height distribution along the fibril axis (periodicity of ~13 nm). These observations suggest that polyQ35 fibrils might be formed by intertwining two oligomer strings.

Figure 2 Solution AFM height images of polyQ35 aggregates incubated for 0 day (A), 1 day (B), 2 days (C), and 9 days (D) at 4oC. Insets show representative fibrils and oligomers with 3-time magnification. Images are obtained after similar equilibration times in the AFM liquid cell at 23oC. Scale bars = 100 nm. (E–H) Height probabilities for the corresponding height images in (A–D). At 4oC, polyQ35 undergoes slow aggregation, yielding different proportions of monomers, oligomers, and fibrils. Figure 2 shows four aliquots incubated for different days at 4oC. We note that polyQ are known to aggregate at room temperature when equilibrating in the AFM liquid cell; more aggregates (oligomers and fibrils) are observed with longer incubation at room temperature. The images in Fig. 2 are obtained with similar incubation time after injection into the liquid cell. It is clear that the amount of aggregates increases with the incubation days at 4oC. Amorphous aggregates are observed when there are abundant fibrils (Fig. 2C and 2D). Globular oligomers with similar sizes are observed in all images. Height probabilities indicate that the fibril height is independent on incubation time (Fig. 2E–2H). Note that fibril peaks are located at ~3 nm, a value smaller than the fibril height estimated from Fig. 1. This is due to the cylindrical shape of fibrils (only fibril apex has the maximum height of ~4 nm).

7

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Secondary structure of polyQ35 aggregates revealed by FTIR To evaluate the secondary structure of polyQ35 aggregates after incubating for different days, we use attenuated total reflection FTIR in solution. Similar to the AFM experiment, it is expected (and experimentally confirmed) that polyQ35 undergoes aggregation when equilibrating in the FTIR liquid cell. For this reason, we only compare the first scan after transferring a fresh aliquot from the stock solution at 4oC. The absorption spectra for aliquots incubated from 0 to 7 days are shown in Fig. 3. We also pelleted aggregates for an aliquot incubated for 7 days (see Materials and Methods). The absorption spectrum for the pellet is also displayed. Two major bands located at 1655 and 1607 cm-1 corresponding to loop/turn and intermolecular β-sheet structures, respectively, are highlighted by dashed lines [38, 39]. It is clear that the loop/turn band is progressively weakened while the β-sheet band is enhanced (the largest change takes place during the first day of incubation). The observed transition is consistent with our AFM data, which show that the amount of β-sheet enriched aggregates increases with incubation days. Interestingly, the pelleted aggregates also display a small loop/turn band. This can be explained by considering that polyQ aggregates contain a small fraction of loop/turn motif.

Figure 3 Attenuated total reflection FTIR absorption spectra of polyQ35 incubated for different days at 4oC (indicated by the values next to each spectrum). The spectrum of the pelleted aggregates is obtained by centrifuging an aliquot incubated for 7 days. GUV disruption revealed by fluorescence microscopy We use fluorescence microscopy of DOPC GUVs (diameter of ~30 μm) to qualitatively study membrane perturbation endowed by polyQ35 aggregates. For the intact GUV, fluorescent lipid rhodamine-DPPE is uniformly distributed over the vesicle (Fig. 4). Addition of an equal volume of polyQ35 solution (incubated for 5 days at 4oC) results in lipid bilayer disruption. This is clearly illustrated by the five vesicles with deflated shapes. We note that addition of the same volume of HEPES buffer does not induce any vesicle deformation. Moreover, the observed perturbation does not take place immediately following the introduction of polyQ35; the majority of vesicle breakage occurs in ~10 min after mixing.

8

ACCEPTED MANUSCRIPT Figure 4 Fluorescence microscopy of DOPC GUVs before and after mixing with polyQ35 aliquot. The peptide has been incubated for 5 days at 4oC, and the final peptide concentration is 15 μM. Scale bars = 10 μm. The left panel shows an intact GUV. After mixing with polyQ35, GUVs become deflated (five representative vesicles on the right).

AC CE P

TE

D

MA

NU

SC

RI

PT

Vesicle disruption revealed by calcein leakage Our GUV experiment clearly shows that polyQ35 can induce lipid bilayer disruption. To assess the kinetic impact of polyQ35 on membrane integrity, we use fluorescence spectroscopy. Calcein-encapsulated ULVs (diameter of ~120 nm) are prepared by the extrusion method, followed by SEC separation. Fluorescence intensity from intact vesicles is small due to calcein self-quenching. When vesicles become leaky to calcein, the fluorescence intensity increases; the increased amount is correlated with calcein leakage or the degree of membrane disruption. We monitor calcein fluorescence for vesicles mixed with polyQ35 incubated for 0, 1, and 6 days at 4oC, as well the pelleted aggregates from a 7-day aliquot. The final peptide concentration in mixture is 10 μM. The same SEC portion of calcein-encapsulated ULVs is used to eliminate lipid concentration variation. For the 0-day aliquot, the degree of calcein leakage increases rapidly during the first few hours of mixing (Fig. 5). The degree of calcein leakage levels off near ~8 h. Interestingly, only a fraction of calcein molecules are released from vesicle interior even when the maximum leakage is reached (i.e., maximum leakage < 100%). This indicates that even at maximum leakage, the impaired lipid membrane is not completely damaged, i.e., a membrane barrier still exists to prevent enclosed calcein from diffusing away. For aliquots incubated with 1 day and 6 days, the maximum calcein leakage becomes less compared to the 0-day aliquot. Interestingly, calcein leakage is also observed for the pelleted aggregates, albeit with longer equilibration time required to reach leakage saturation.

Figure 5 Calcein leakage of DOPC ULVs induced by polyQ35 incubated for different days at 4oC, as well as the pelleted aggregates from a 7-day aliquot. The excitation and emission wavelengths are 494 and 515 nm, respectively. Calcein-encapsulated ULVs are prepared using SEC separation. Snap-frozen peptide aliquots are mixed with the same portion of ULVs to eliminate lipid concentration variation. The final peptide concentration is 10 μM. Time-course measurements are performed at 23oC. Fluorescence curve for ULV solution without polyQ35 is also shown. Disruption of mica supported lipid bilayers revealed by solution AFM Both GUV and calcein leakage data indicate that polyQ35 confer a time-dependent disruptive effect on lipid bilayers. To obtain detailed structural and mechanical impact of polyQ35 on lipid bilayers, we use solution AFM. Mica supported planar DOPC bilayers are prepared in the AFM liquid cell. To test whether polyQ35 monomers play a role in modulating lipid bilayer structure,

9

ACCEPTED MANUSCRIPT

RI

PT

we first expose the DOPC bilayer to monomeric supernatant, which was obtained by ultracentrifuging 20 μM polyQ35 powder dissolved in 10 mM HEPES pH 7.4 (maximum RCF of 266,000×g for 5 h). After incubating for 2 h in the liquid cell, very little change of the DOPC bilayer was observed (data not shown). This leads us to conclude that polyQ35 monomers do not disrupt bilayers, a result consistent with previous findings [13, 19]. We next expose DOPC bilayers to aliquots of 20 μM polyQ35 incubated for different days at 4oC. Time-dependent bilayer remodeling is observed. This is illustrated in Fig. 6 for the 0-day aliquot. Since monomers are not responsible for bilayer disruption, the observed bilayer restructuring must be due to aggregated forms of polyQ35 (i.e., oligomers and fibrils).

MA

NU

SC

In addition to height images, high-resolution maps of the reduced Young’s modulus E* and the adhesion force Fadh are also shown in Fig. 6 (see Supplementary Material for the derivation of E* and Fadh from PeakForce QNM mode scanning). Using the thin shell theory for lipid bilayers, the reduced Young’s modulus is proportionally associated with bilayer lateral compressibility modulus KA and bilayer bending modulus KC: and 2, where H is the bilayer thickness and ν is the Poisson ratio. It is clear that both KA and KC are proportional to E*; a larger E* reflects a larger rigidity of the bilayer. The adhesion force Fadh from AFM retracting force-distance curves characterizes the strength of interactions between the bilayer surface and the AFM tip. Typically, molecules such as lipids that have a tendency of being pulled out by the AFM tip exhibit larger Fadh than those that are rigid such as amyloid fibrils.

AC CE P

TE

D

The time-dependent height images in Fig. 6 show that after exposing to polyQ35 aggregates, the DOPC bilayer is gradually modified. Regions corresponding to the intact bilayer (e.g., region a) are first converted into partially disrupted organization (e.g., region b), which is then further disrupted to form the fully disrupted bilayer (e.g., region c). During the disruption process, fibrils are observed to deposit on bilayer surfaces, mainly in partially and fully disrupted regions. This indicates that the interactions between fibrils and lipid bilayers may not be strong enough for fibril deposition. Exposure of buried lipid motifs facilitates the sorting of fibrils onto the bilayer surface. Height probabilities of the three height images (Fig. 6J) reveal that the disrupted bilayer has a smaller height than the intact bilayer. This is further illustrated by height profiles crossing the bilayer surface (Fig. 7). For intact bilayer, the height variation is small (rmsd ≈ 0.06 nm). Similar height variation is observed for the unperturbed region a; partial (region b) and full (region c) disruption results in ~0.2 and ~0.6 nm height decrease, respectively. Notice that both the partially and fully disrupted regions exhibit larger roughness compared to the intact bilayer.

10

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC CE P

TE

D

Figure 6 Solution AFM images of mica supported DOPC bilayers exposed to 20 μM polyQ35 aggregates incubated for 0 day at 4oC. (A–C) Height images, (D–F) reduced Young’s modulus E* maps, and (G–I) adhesion force maps. Scale bars = 200 nm. The three rows correspond to three time points (~20 min apart) when the measurements are performed. Three regions are indicated as a, b, and c in (A), (D), and (G) to highlight the intact bilayer, the partially disrupted bilayer, and the fully disrupted bilayer, respectively. (J) Histograms (normalized) of the corresponding height, reduced Young’s modulus, and adhesion force maps at the three time points shown in (A–I). Examination of Fig. 6C indicates that there are numerous small patches on the bilayer surface. AFM scans with larger magnification reveal that these small patches can have circular or irregular shapes depending on the degree of bilayer disruption and the incubation days of polyQ35 at 4oC (Fig. 8). After extracting the three species present in Fig. 8 (i.e., fibrils, circularly or irregularly shaped patches, and the fully disrupted region) using appropriate height thresholds, we find that the patches have an average height larger than that of the fully disrupted region by ~0.6 nm (Fig. 8G and 8H). Compared to the height profiles in Fig. 7, the height difference of ~0.6 nm suggests that these patches correspond to intact bilayer regions, rather than polyQ35 oligomers. This statement is further supported by our analysis of bilayer material and adhesion force properties as discussed below.

11

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

Figure 7 Solution AFM height images of a DOPC bilayer modulated by polyQ35 aggregates. (A) DOPC bilayer, and (B) DOPC bilayer e posed to 20μM polyQ35 aggregates incubated for 0 day at 4oC. Scale bars = 200 nm. (C) and (D) are height profiles (in nm) along the solid green lines in (A) and (B), respectively. Three regions in (D) are indicated. Region a corresponds to intact bilayer; region b corresponds to partially disrupted bilayer; region c corresponds to fully disrupted bilayer. Compared to the intact bilayer, the height of the partially and fully disrupted bilayers is smaller by ~0.2 and ~0.6 nm, respectively.

Figure 8 Solution AFM height images of DOPC bilayers exposed to 20 μM polyQ35 incubated for 0 day (A) and 1 day (B). Intact lipid patches (C and D) and polyQ35 fibrils (E and F) are digitally extracted by setting appropriate height thresholds. (G) and (H) are normalized height probabilities for the three species: a, fibrils; b, intact lipid patches; and c, disrupted lipid regions. The average height difference between the disrupted and the intact lipid regions is ~0.6 nm. Scale bars = 100 nm. In addition to height contrasts between the intact, the partially and the fully disrupted bilayers in Fig. 6, the simultaneously obtained reduced Young’s modulus maps indicate that the fully

12

ACCEPTED MANUSCRIPT

MA

NU

SC

RI

PT

disrupted region has the largest rigidity, followed by the partially disrupted and the intact regions (Fig. 6D). We note that (i) the obtained rigidity is a collective mechanical property incorporating lipids and peptide aggregates in each pixel, and (ii) the absolute values of E* are dependent on scanning parameters and the AFM tip. Nevertheless, the relative rigidities between different regions on the bilayer surface are accurately reflected by the reduced Young’s modulus map. The observed rigidity contrasts between the three regions in Fig. 6D can be understood by considering that regions with larger amount of polyQ35 aggregates, which causes bilayer disruption, have larger rigidity than that of fluid phase lipids. Similar to the rigidity map, the adhesion force map shows that there are distinct contrasts between the three types of bilayer regions (Fig. 6G). While it is expected that the intact region has the largest adhesion force, the smaller adhesion force for the partially disrupted region than that of the fully disrupted region is quite surprising (Fig. 6G). This observation can be reconciled by considering that the fully disrupted region exposes adhesive lipid motifs such as the interfacial hydrocarbon chains, which are still buried in the partially disrupted region. Consequently, the adhesion force experienced by the AFM tip can be larger in the fully disrupted region than in the partially disrupted region – assuming that hydrocarbon chains are more adhesive than polar headgroups. Histograms of the reduced Young’s modulus and the adhesion force maps (Fig. 6J) are consistent with our qualitative assessment that larger bilayer disruption results in larger bilayer rigidity and smaller bilayer adhesion force.

AC CE P

TE

D

Figure 6 shows the overall contrast between the three types of regions depending on the degree of bilayer disruption. Scans with larger magnifications reveal more details of bilayer mechanical and adhesion force properties (Fig. S2, Supplementary Material). Compared to the fully disrupted region, the circularly or irregularly shaped patches exhibit smaller rigidity and larger adhesion force (Fig. S2). These properties are consistent with our earlier assertion that the protruded patches (by ~0.6 nm) correspond to intact bilayer regions, rather than polyQ35 oligomers. In addition to the intact patches, fibrils are also observed in the large-magnification images. In particular, we find that compared to the fully disrupted region, fibrils have a smaller adhesion force (Fig. S2). Moreover, the rigidity of fibrils is comparable to that of the fully disrupted bilayer. This observation supports the notion that the disrupted region is enriched with polyQ35 aggregates. DOPC bilayers exposed to polyQ35 aliquots incubated for 1 day and 9 days are shown in Fig. S3 (Supplementary Material). Similar to the 0-day aliquot, fibrils deposited on bilayer surfaces are observed in all images. Importantly, all aliquots are able to induce bilayer disruption. This is confirmed by comparing height variances between the intact and disrupted bilayers using a twosample F-test (after excluding regions with deposited fibrils). The resulting p values are < 0.01 for all of the disrupted bilayers in Fig. S3. One difference is the intact lipid patches induced by the 0-day and 1-day aliquots, while no patches are observed for the 9-day aliquot. The difference could be due to different equilibration time in the AFM liquid cell. One interesting question is whether lipid bilayer surface can promote or inhibit polyQ aggregate formation. Figure 9 shows two consecutive scans (30 min apart) for a DOPC bilayer exposed to 20 μM polyQ35 incubated for 0 day at 4oC. It is clear that the later scan contains more fibrils. Close examination indicates that some of the fibrils in the earlier scan (Fig. 9A) become longer in the later scan (Fig. 9B). To quantify fibril elongation, we calculate the aspect ratio of each fibril. This is done by digitally tracing fibril contours (Fig. S4, Supplementary Material). The length of each fibril is estimated based on the fibril area and the assumption that each fibril has the same width (e.g., 12 nm). Once the length is determined, the aspect ratio is the ratio of the fibril length over the constant fibril width. Note that a different estimation of the fibril width only shifts the aspect ratio histograms (Fig. 9D and 9E) to the left or right. Consistent with our

13

ACCEPTED MANUSCRIPT

TE

D

MA

NU

SC

RI

PT

preliminary assessment of fibril elongation, there are more fibrils with larger aspect ratios in the later scan. It seems that after deposition onto the bilayer surface, fibrils can continuously grow in length by recruiting oligomers to the vicinity of the bilayer surface.

AC CE P

Figure 9 PolyQ35 fibril elongation on DOPC bilayer surface. Scale bars = 200 nm. Five ellipses (a–e) are shown to highlight where fibril elongation occurs. (C) Height profiles along two trajectories depicted in (A). (D) and (E) are aspect ratio histograms of polyQ35 fibrils in (A) and (B), respectively. (F) and (G) are height probabilities in rectangular regions containing all fibrils (Fig. S4, Supplementary Material). The average height difference between fibrils and the bilayer surface is ~2 nm. Height profiles along two directions of a fibril in Fig. 9A indicates that the bilayer associated fibril has a height of ~2 nm and an apparent width of ~12 nm (Fig. 9C). Height probability distributions within rectangular regions containing all the fibrils support that the fibrils have an average height of ~2 nm larger than that of the bilayer surface (Fig. 9F and 9G). (The rectangular regions used for histogram calculation are shown in Fig. S5, Supplementary Material.) This value is smaller than the fibril height when adsorbed on mica surface (Figs. 1 and 2). The difference can be explained by considering that the bottom of the cylindrical fibrils protrudes slightly into the headgroup region of the bilayer. As a result, the relative height between the bilayer surface and the apex of the fibril is smaller than the actual height of the fibril. Fibril indentation could contribute to bilayer disruption, although this effect is minimal compared to the extensive disruption where fibrils are not present (fully disrupted region in Figs. 6 and 8). Discussion Globular oligomers of polyQ Identifying different forms of polyQ aggregates is relevant to the development of therapeutic strategies targeting specific toxic species in expanded CAG repeat diseases [40-42]. Earlier kinetic studies on synthetic polyQ peptides suggested that the nucleated growth of fibrillar aggregates involves a critical nucleus size of one monomer [4]. Fibrils are spontaneously

14

ACCEPTED MANUSCRIPT

PT

formed by the metastable monomeric nucleus. Other studies challenged this view by showing that soluble aggregates do form during the monomer-fibril conversion [43]. The apparent discrepancy was later reconciled (to some degree) by a systematic study showing that nucleus size is repeat length dependent [5]. Short stretches aggregate via nuclei up to 4 monomers, while longer stretches undergo monomeric nucleation. Other studies indicated that flanking sequences could also modify nucleus size [6, 44, 45].

D

MA

NU

SC

RI

One source of confusion is the difficulty of visually detecting the various fibrillar precursors with polymorphic and possibly transient nature. In this regard, high-resolution AFM is invaluable for revealing different amyloid aggregates present on or off aggregation pathways of polyQ [40]. Indeed, oligomers of polyQ without or with the htt N-terminal have been observed using AFM [7, 46]. Although the reported dimensions vary, AFM studies support that multi-monomeric oligomers of long polyQ stretches do form [47]. This, however, does not exclude the possibility of monomeric nucleation pathway, since the preparation protocol [48, 49] and the mica surface [50] can alter polyQ aggregation behavior, as well as the peptide sequence used in each study [5, 6]. We note that our polyQ35 stock solution was incubated at 4oC and examined at room temperature. It has been reported that htt-polyQ complex contains more loop/turn structures when incubated at 4oC than at 37oC [39]. This observation has little effect on our data since we clearly see the intermolecular β-sheet band in all of the aliquots examined (Fig. 3), while this band is diminished in the reported study at 4oC. One possibility is that all our measurements are performed at room temperature, which could bring the aliquots to “normal temperature” behavior.

AC CE P

TE

Lipid bilayer perturbation Cell membranes have been proposed as the primary target by an array of amyloidogenic proteins and peptides [13, 19, 27]. To explore membrane perturbation induced by polyQ35 aggregates, we use a number of experimental approaches. We find that polyQ35 aggregates can perturb lipid bilayer integrity, yielding deflated GUVs. Our fluorescence spectroscopy data indicate that after exposing to polyQ35 aggregates, lipid bilayers become permeable to calcein. Moreover, the efficacy of calcein leakage is inversely related to the incubation time of polyQ35 at 4oC. The disruptive effect of polyQ35 aggregates on lipid bilayers is further revealed by highresolution solution AFM measurements. Interestingly, we find that lipid bilayer disruption is a two-step process. Lipid motifs such as phosphate group residing closer to the water-bilayer interface are first exposed, resulting in partially disrupted morphology; full disruption is achieved by exposing deeper lipid motifs such as the glycerol-carbonyl group. The observed two-step process highlights that peptide concentration and equilibration time can be important parameters in explaining discrepant membrane toxicities of amyloid aggregates. Similar lipid bilayer disruption/remodeling has been observed for a broad spectrum of amyloid aggregates [51-53]. Studies on interactions between polyQ and lipid bilayers have also been performed [25]. The authors found that polyQ monomers do not appreciably aggregate and disrupt lipid bilayers; addition of a flanking sequence significantly augments bilayer disruption. Since our data clearly show that simple polyQ stretches is fully capable of disrupting lipid bilayers, the discrepancy can be attributed to the incubation time that is required to generate toxic polyQ aggregates. We overcome the problem by skipping monomeric disaggregation process. In addition to topographical information, our AFM measurements also reveal bilayer mechanical and adhesion force properties. We find that bilayer rigidity increases continuously when the bilayer undergoes partial and full disruption by polyQ35 aggregates. Interestingly, the adhesion force between the bilayer and the AFM tip first decreases and then increases during the twostep disruption process. Since proper membrane function requires balanced structural and mechanical properties, our AFM data provide additional perspective on how polyQ aggregates

15

ACCEPTED MANUSCRIPT

NU

SC

RI

PT

might hamper cellular activity. Our observation of the enhanced bilayer rigidity by polyQ adsorption and insertion is consistent with previous AFM force measurements, which showed that larger puncture forces are required for bilayers exposed to amyloid-β prefibrillar oligomers [21, 54]. Using an AFM mode referred to as the scanning probe acceleration microscopy (SPAM), Legleiter and coworkers measured the maximum tapping force (Fmax) and minimal tapping force (Fmin) of lipid bilayers during each tapping event [24, 25, 49, 55]. The authors found that addition of amyloid aggregates decreases Fmax and Fmin. By correlating Fmax to bilayer rigidity and Fmin to bilayer-tip adhesion force, the authors concluded that amyloid aggregates simultaneously decrease bilayer rigidity and adhesion force. While the later conclusion is in accordance with ours – although we obtain a two-step process, which was not seen in their studies – their result of smaller bilayer rigidity contradicts our data. One explanation is that unlike the reduced Young’s modulus, which is obtained by fitting force-distance curves, Fmax from SPAM only measures one force during a tapping cycle. An alternative explanation is that the studies by Legleiter and coworkers used a different lipid bilayer, as well as the addition of a flanking sequence to polyQ stretches.

AC CE P

TE

D

MA

It is clear that polyQ aggregates can modulate structural and mechanical properties of lipid bilayers. The vexing question is what aggregate is responsible for the observed change. Our solution AFM measurements on mica indicate that globular oligomers and short fibrils are the dominant aggregated forms of polyQ35. Exposing lipid bilayers to polyQ35 causes bilayer disruption in an equilibration time dependent manner. The final bilayer is composed of intact lipid patches, fully disrupted region, and polyQ fibrils (Fig. 8). Where do the oligomers go? The answer is hinted by the fully disrupted region without fibrils. Since monomers do not cause bilayer disruption [13, 19], the only candidate for the observed change in the fully disrupted region is oligomers, i.e., polyQ35 oligomers are the toxic aggregates causing bilayer disruption. This conclusion is supported by several lines of evidence: (i) oligomers can increase the rigidity of the fully disrupted region (Fig. 6), and (ii) bilayer thickness (~3-4 nm) is compatible with oligomers but not fibrils (Fig. 1). Although controversies remain pertaining to cytotoxicity of different forms of amyloid aggregates [56], the implication from our data that globular oligomers cause membrane disruption agrees with many studies showing that formation of inert fibrils is a coping response to toxic oligomers [57]. Several compelling mechanisms have been suggested to explain membrane permeabilization induced by amyloidogenic proteins and peptides: (i) nonspecific membrane disruption [19, 58], (ii) transmembrane pore formation [26], and (iii) detergent-like micellization [59]. Although the exact mechanism is most likely protein and aggregate dependent, our experimental data support nonspecific membrane disruption by polyQ aggregates. The absence of detergent-like effect is evident based on the calcein leakage data, which indicates that even when the leakage reaches maximum, there are still significant amount of calcein molecules enclosed by lipid vesicles. This will not be the case if lipids are extracted to form micelle-like structures with polyQ aggregates. The absence of annular pores is confirmed by our high-resolution solution AFM measurements (Fig. S3, Supplementary Material). Aside from occasional defects, the majority of the fully disrupted bilayer is homogeneous. We note that our solution AFM has enough resolution to see transmembrane protein oligomers of ~5 nm diameter in lipid bilayers (unpublished data). The mechanism of generic membrane disruption by polyQ aggregates is consistent with several lines of experimental data [19, 58, 60, 61]. Conclusions In this study, we prepared aggregates of a polyQ stretch containing 35 glutamine residues at 4oC. Depending on the incubation time, globular oligomers and short fibrils are observed using solution AFM on a mica surface. Size and volume estimation reveals that polyQ35 oligomers

16

ACCEPTED MANUSCRIPT

MA

NU

SC

RI

PT

contain ~10-13 monomers, although the estimation is complicated by uncertainties due to oligomer size heterogeneity and the finite size of AFM tips. PolyQ35 fibrils have a twisted morphology and are most likely formed by intertwining two oligomer strings. Using attenuated total refection FTIR, we show that polyQ35 aggregates are enriched in intermolecular β-sheet structure; the content of β-sheet increases with incubation time. We also study the impact of polyQ35 aggregates on vesicular and mica supported planar lipid membranes. Vesicle breakage is observed using fluorescence microscopy of ~30 μm sized GUVs. Similarly, timedependent fluorescence spectroscopy measurement reveals that lipid vesicles become permeable to calcein after introducing polyQ35 aggregates. Interestingly, we observe that even at the maximum leakage, significant amount of calcein molecules are still encapsulated by lipid vesicles. This suggests that polyQ35 aggregates only modulate lipid membrane integrity, but do not cause membrane fragmentation. Detailed structural and mechanical perturbation of lipid membranes is obtained by solution AFM. We find that membrane disruption proceeds by partially decreasing bilayer height by ~0.2 nm, followed by a more aggressive decrease of ~0.4 nm (total of ~0.6 nm). In addition to the height contrast, the resulting partially and fully disrupted bilayers have distinct rigidity and adhesion force properties compared to the intact bilayer. Specifically, the bilayer rigidity increases as the intact bilayer becomes partially and fully disrupted. Surprisingly, the adhesion force first decreases and then increases during the disruption process. Finally, we show that membrane associated fibrils can grow in length, although the role of lipid bilayer surface on polyQ aggregation (e.g., promotion or suppression of fibril growth compared to the aqueous solution) is unclear based on our study [56, 62].

AC CE P

TE

D

Acknowledgments We thank Dr. Martin Muschol and Ms Tatiana Miti for helping osmotic pressure measurements. J. Pan acknowledges the startup fund from the University of the South Florida. J. Cai is supported by NSF 1351265 and NIH 1R01GM112652-01A1. J. Pan is supported by NIH R15GM117531.

17

ACCEPTED MANUSCRIPT References [1] P.O. Bauer, N. Nukina, The pathogenic mechanisms of polyglutamine diseases and current therapeutic strategies, J Neurochem, 110 (2009) 1737-1765.

PT

[2] H.Y. Zoghbi, H.T. Orr, Glutamine repeats and neurodegeneration, Annu Rev Neurosci, 23 (2000) 217-247.

RI

[3] E. Scherzinger, A. Sittler, K. Schweiger, V. Heiser, R. Lurz, R. Hasenbank, G.P. Bates, H. Lehrach, E.E. Wanker, Self-assembly of polyglutamine-containing huntingtin fragments into amyloid-like fibrils: Implications for Huntington's disease pathology, P Natl Acad Sci USA, 96 (1999) 4604-4609.

SC

[4] S.M. Chen, F.A. Ferrone, R. Wetzel, Huntington's disease age-of-onset linked to polyglutamine aggregation nucleation, P Natl Acad Sci USA, 99 (2002) 11884-11889.

NU

[5] K. Kar, M. Jayaraman, B. Sahoo, R. Kodali, R. Wetzel, Critical nucleus size for diseaserelated polyglutamine aggregation is repeat-length dependent, Nat Struct Mol Biol, 18 (2011) 328.

MA

[6] A.K. Thakur, M. Jayaraman, R. Mishra, M. Thakur, V.M. Chellgren, I.J.L. Byeon, D.H. Anjum, R. Kodali, T.P. Creamer, J.F. Conway, A.M. Gronenborn, R. Wetzel, Polyglutamine disruption of the huntingtin exon 1 N terminus triggers a complex aggregation mechanism, Nat Struct Mol Biol, 16 (2009) 380-389.

TE

D

[7] J. Legleiter, E. Mitchell, G.P. Lotz, E. Sapp, C. Ng, M. DiFiglia, L.M. Thompson, P.J. Muchowski, Mutant Huntingtin Fragments Form Oligomers in a Polyglutamine Lengthdependent Manner in Vitro and in Vivo, J Biol Chem, 285 (2010) 14777-14790. [8] T. Takahashi, S. Katada, O. Onodera, Polyglutamine Diseases: Where does Toxicity Come from? What is Toxicity? Where are We Going?, J Mol Cell Biol, 2 (2010) 180-191.

AC CE P

[9] M.F.M. Engel, L. Khemtemourian, C.C. Kleijer, H.J.D. Meeldijk, J. Jacobs, A.J. Verkleij, B. de Kruijff, J.A. Killian, J.W.M. Hoppener, Membrane damage by human islet amyloid polypeptide through fibril growth at the membrane, P Natl Acad Sci USA, 105 (2008) 6033-6038. [10] P.H. Ren, J.E. Lauckner, I. Kachirskaia, J.E. Heuser, R. Melki, R.R. Kopito, Cytoplasmic penetration and persistent infection of mammalian cells by polyglutamine aggregates, Nat Cell Biol, 11 (2009) 219. [11] L. Pieri, K. Madiona, L. Bousset, R. Melki, Fibrillar alpha-Synuclein and Huntingtin Exon 1 Assemblies Are Toxic to the Cells, Biophys J, 102 (2012) 2894-2905. [12] D.M. Walsh, I. Klyubin, J.V. Fadeeva, W.K. Cullen, R. Anwyl, M.S. Wolfe, M.J. Rowan, D.J. Selkoe, Naturally secreted oligomers of amyloid beta protein potently inhibit hippocampal longterm potentiation in vivo, Nature, 416 (2002) 535-539. [13] R. Kayed, E. Head, J.L. Thompson, T.M. McIntire, S.C. Milton, C.W. Cotman, C.G. Glabe, Common structure of soluble amyloid oligomers implies common mechanism of pathogenesis, Science, 300 (2003) 486-489. [14] I. Sanchez, C. Mahlke, J.Y. Yuan, Pivotal role of oligomerization in expanded polyglutamine neurodegenerative disorders, Nature, 421 (2003) 373-379. [15] A. Demuro, E. Mina, R. Kayed, S.C. Milton, I. Parker, C.G. Glabe, Calcium dysregulation and membrane disruption as a ubiquitous neurotoxic mechanism of soluble amyloid oligomers, J Biol Chem, 280 (2005) 17294-17300. [16] C. Haass, D.J. Selkoe, Soluble protein oligomers in neurodegeneration: lessons from the Alzheimer's amyloid beta-peptide, Nat Rev Mol Cell Bio, 8 (2007) 101-112. 18

ACCEPTED MANUSCRIPT [17] T. Takahashi, S. Kikuchi, S. Katada, Y. Nagai, M. Nishizawa, O. Onodera, Soluble polyglutamine oligomers formed prior to inclusion body formation are cytotoxic, Hum Mol Genet, 17 (2008) 345-356.

PT

[18] I. Benilova, E. Karran, B. De Strooper, The toxic A beta oligomer and Alzheimer's disease: an emperor in need of clothes, Nat Neurosci, 15 (2012) 349-357.

RI

[19] R. Kayed, Y. Sokolov, B. Edmonds, T.M. McIntire, S.C. Milton, J.E. Hall, C.G. Glabe, Permeabilization of lipid bilayers is a common conformation-dependent activity of soluble amyloid oligomers in protein misfolding diseases, J Biol Chem, 279 (2004) 46363-46366.

SC

[20] B.D. van Rooijen, M.M.A.E. Claessens, V. Subramaniam, Lipid bilayer disruption by oligomeric alpha-synuclein depends on bilayer charge and accessibility of the hydrophobic core, Bba-Biomembranes, 1788 (2009) 1271-1278.

NU

[21] C. Canale, S. Seghezza, S. Vilasi, R. Carrotta, D. Bulone, A. Diaspro, P.L.S. Biagio, S. Dante, Different effects of Alzheimer's peptide A beta(1-40) oligomers and fibrils on supported lipid membranes, Biophys Chem, 182 (2013) 23-29.

MA

[22] M.F.M. Engel, Membrane permeabilization by Islet Amyloid Polypeptide, Chem Phys Lipids, 160 (2009) 1-10.

D

[23] P. Walsh, G. Vanderlee, J. Yau, J. Campeau, V.L. Sim, C.M. Yip, S. Sharpe, The Mechanism of Membrane Disruption by Cytotoxic Amyloid Oligomers Formed by Prion Protein(106-126) Is Dependent on Bilayer Composition, J Biol Chem, 289 (2014) 10419-10430.

TE

[24] K.A. Burke, K.M. Hensal, C.S. Umbaugh, M. Chaibva, J. Legleiter, Huntingtin disrupts lipid bilayers in a polyQ-length dependent manner, Bba-Biomembranes, 1828 (2013) 1953-1961.

AC CE P

[25] K.A. Burke, K.J. Kauffman, C.S. Umbaugh, S.L. Frey, J. Legleiter, The Interaction of Polyglutamine Peptides with Lipid Membranes Is Regulated by Flanking Sequences Associated with Huntingtin, J Biol Chem, 288 (2013) 14993-15005. [26] H.A. Lashuel, P.T. Lansbury, Are amyloid diseases caused by protein aggregates that mimic bacterial pore-forming toxins?, Q Rev Biophys, 39 (2006) 167-201. [27] C. Cecchi, M. Stefani, The amyloid-cell membrane system. The interplay between the biophysical features of oligomers/fibrils and cell membrane defines amyloid toxicity, Biophys Chem, 182 (2013) 30-43. [28] C. Galvagnion, A.K. Buell, G. Meisl, T.C.T. Michaels, M. Vendruscolo, T.P.J. Knowles, C.M. Dobson, Lipid vesicles trigger alpha-synuclein aggregation by stimulating primary nucleation, Nature chemical biology, 11 (2015) 229. [29] J.D. Knight, A.D. Miranker, Phospholipid catalysis of diabetic amyloid assembly, J Mol Biol, 341 (2004) 1175-1187. [30] I.C. Martins, I. Kuperstein, H. Wilkinson, E. Maes, M. Vanbrabant, W. Jonckheere, P. Van Gelder, D. Hartmann, R. D'Hooge, B. De Strooper, J. Schymkowitz, F. Rousseau, Lipids revert inert A beta amyloid fibrils to neurotoxic protofibrils that affect learning in mice, Embo J, 27 (2008) 224-233. [31] R.H. Walters, R.M. Murphy, Examining Polyglutamine Peptide Length: A Connection between Collapsed Conformations and Increased Aggregation, J Mol Biol, 393 (2009) 978-992. [32] S.M. Chen, V. Berthelier, J.B. Hamilton, B. O'Nuallain, R. Wetzel, Amyloid-like features of polyglutamine aggregates and their assembly kinetics, Biochemistry-Us, 41 (2002) 7391-7399.

19

ACCEPTED MANUSCRIPT [33] C.S. Ho, N.K. Khadka, J. Pan, Sub-ten-nanometer heterogeneity of solid supported lipid membranes determined by solution atomic force microscopy, Bba-Biomembranes, 1858 (2015) 181-188.

PT

[34] N.K. Khadka, C.S. Ho, J.J. Pan, Macroscopic and Nanoscopic Heterogeneous Structures in a Three-Component Lipid Bilayer Mixtures Determined by Atomic Force Microscopy, Langmuir, 31 (2015) 12417-12425.

RI

[35] C.G. Glabe, Structural Classification of Toxic Amyloid Oligomers, J Biol Chem, 283 (2008) 29639-29643.

SC

[36] B. O'Nuallain, A.K. Thakur, A.D. Williams, A.M. Bhattacharyya, S.M. Chen, G. Thiagarajan, R. Wetzel, Kinetics and thermodynamics of amyloid assembly using a high-performance liquid chromatography-based sedimentation assay, Method Enzymol, 413 (2006) 34-74.

NU

[37] H. Fischer, I. Polikarpov, A.F. Craievich, Average protein density is a molecular-weightdependent function, Protein Sci, 13 (2004) 2825-2828. [38] M. Jackson, H.H. Mantsch, The Use and Misuse of Ftir Spectroscopy in the Determination of Protein-Structure, Crit Rev Biochem Mol, 30 (1995) 95-120.

MA

[39] Y. Nekooki-Machida, M. Kurosawab, N. Nukina, K. Ito, T. Oda, M. Tanaka, Distinct conformations of in vitro and in vivo amyloids of huntingtin-exon1 show different cytotoxicity, P Natl Acad Sci USA, 106 (2009) 9679-9684.

D

[40] M.A. Poirier, H.L. Li, J. Macosko, S.W. Cai, M. Amzel, C.A. Ross, Huntingtin spheroids and protofibrils as precursors in polyglutamine fibrilization, J Biol Chem, 277 (2002) 41032-41037.

AC CE P

TE

[41] B. Caughey, P.T. Lansbury, Protofibrils, pores, fibrils, and neurodegeneration: Separating the responsible protein aggregates from the innocent bystanders, Annu Rev Neurosci, 26 (2003) 267-298. [42] C. Zuccato, M. Valenza, E. Cattaneo, Molecular Mechanisms and Potential Therapeutical Targets in Huntington's Disease, Physiol Rev, 90 (2010) 905-981. [43] C.C. Lee, R.H. Walters, R.M. Murphy, Reconsidering the mechanism of polyglutamine peptide aggregation, Biochemistry-Us, 46 (2007) 12810-12820. [44] D. Bulone, L. Masino, D.J. Thomas, P.L.S. Biagio, A. Pastore, The Interplay between PolyQ and Protein Context Delays Aggregation by Forming a Reservoir of Protofibrils, Plos One, 1 (2006) e111. [45] M. Jayaraman, R. Mishra, R. Kodali, A.K. Thakur, L.M.I. Koharudin, A.M. Gronenborn, R. Wetzel, Kinetically Competing Huntingtin Aggregation Pathways Control Amyloid Polymorphism and Properties, Biochemistry-Us, 51 (2012) 2706-2716. [46] K.A. Burke, J. Godbey, J. Legleiter, Assessing mutant huntingtin fragment and polyglutamine aggregation by atomic force microscopy, Methods, 53 (2011) 275-284. [47] Y. Nagai, T. Inui, H.A. Popiel, N. Fujikake, K. Hasegawa, Y. Urade, Y. Goto, H. Naiki, T. Toda, A toxic monomeric conformer of the polyglutamine protein, Nat Struct Mol Biol, 14 (2007) 332-340. [48] K.N. Dahlgren, A.M. Manelli, W.B. Stine, L.K. Baker, G.A. Krafft, M.J. LaDu, Oligomeric and fibrillar species of amyloid-beta peptides differentially affect neuronal viability, J Biol Chem, 277 (2002) 32046-32053.

20

ACCEPTED MANUSCRIPT [49] E.A. Yates, J. Legleiter, Preparation Protocols of A beta(1-40) Promote the Formation of Polymorphic Aggregates and Altered Interactions with Lipid Bilayers, Biochemistry-Us, 53 (2014) 7038-7050.

PT

[50] K.A. Burke, E.A. Yates, J. Legleiter, Biophysical insights into how surfaces, including lipid membranes, modulate protein aggregation related to neurodegeneration, Front Neurol, 4 (2013) 17.

RI

[51] A. Relini, O. Cavalleri, R. Rolandi, A. Gliozzi, The two-fold aspect of the interplay of amyloidogenic proteins with lipid membranes, Chem Phys Lipids, 158 (2009) 1-9.

SC

[52] F. Hane, E. Drolle, R. Gaikwad, E. Faught, Z. Leonenko, Amyloid-beta Aggregation on Model Lipid Membranes: An Atomic Force Microscopy Study, J Alzheimers Dis, 26 (2011) 485494.

NU

[53] M.S. Saponetti, M. Grimaldi, M. Scrima, C. Albonetti, S.L. Nori, A. Cucolo, F. Bobba, A.M. D'Ursi, Aggregation of A beta(25-35) on DOPC and DOPC/DHA Bilayers: An Atomic Force Microscopy Study, Plos One, 9 (2014) e115780.

MA

[54] J. Lee, A.L. Gillman, H. Jang, S. Ramachandran, B.L. Kagan, R. Nussinov, F.T. Arce, Role of the Fast Kinetics of Pyroglutamate-Modified Amyloid-beta Oligomers in Membrane Binding and Membrane Permeability, Biochemistry-Us, 53 (2014) 4704-4714.

D

[55] K.A. Burke, E.A. Yates, J. Legeiter, Amyloid-Forming Proteins Alter the Local Mechanical Properties of Lipid Membranes, Biochemistry-Us, 52 (2013) 808-817.

TE

[56] S.M. Butterfield, H.A. Lashuel, Amyloidogenic Protein Membrane Interactions: Mechanistic Insight from Model Systems, Angew Chem Int Edit, 49 (2010) 5628-5654.

AC CE P

[57] M. Arrasate, S. Mitra, E.S. Schweitzer, M.R. Segal, S. Finkbeiner, Inclusion body formation reduces levels of mutant huntingtin and the risk of neuronal death, Nature, 431 (2004) 805-810. [58] Y. Sokolov, J.A. Kozak, R. Kayed, A. Chanturiya, C. Glabe, J.E. Hall, Soluble amyloid oligomers increase bilayer conductance by altering dielectric structure, J Gen Physiol, 128 (2006) 637-647. [59] E. Sparr, M.F.M. Engel, D.V. Sakharov, M. Sprong, J. Jacobs, B. de Kruijff, J.W.M. Hoppener, J.A. Killian, Islet amyloid polypeptide-induced membrane leakage involves uptake of lipids by forming amyloid fibers, Febs Lett, 577 (2004) 117-120. [60] J.D. Green, L. Kreplak, C. Goldsbury, X.L. Blatter, M. Stolz, G.S. Cooper, A. Seelig, J. KistLer, U. Aebi, Atomic force microscopy reveals defects within mica supported lipid bilayers induced by the amyloidogenic human amylin peptide, J Mol Biol, 342 (2004) 877-887. [61] M.R.R. de Planque, V. Raussens, S.A. Contera, D.T.S. Rijkers, R.M.J. Liskamp, J.M. Ruysschaert, J.F. Ryan, F. Separovic, A. Watts, beta-sheet structured beta-amyloid(1-40) perturbs phosphatidylcholine model membranes, J Mol Biol, 368 (2007) 982-997. [62] J. Fantini, N. Yahi, Molecular insights into amyloid regulation by membrane cholesterol and sphingolipids: common mechanisms in neurodegenerative diseases, Expert Rev Mol Med, 12 (2010) e27.

21

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Graphical abstract

22

ACCEPTED MANUSCRIPT Highlights

TE

D

MA

NU

SC

RI

PT

PolyQ forms oligomers and fibrils PolyQ aggregates permeabilize lipid bilayers PolyQ aggregates disrupts lipid bilayer organization by a two-step process PolyQ aggregates modulate lipid bilayer mechanical and adhesion force properties

AC CE P

● ● ● ●

23