formation of toxic aβ(1–40) fibrils on gm1 ganglioside-containing membranes mimicking lipid...

doi:10.1016/j.jmb.2008.07.072 J. Mol. Biol. (2008) 382, 1066–1074

Available online at www.sciencedirect.com

Formation of Toxic Aβ(1–40) Fibrils on GM1Ganglioside-Containing Membranes Mimicking LipidRafts: Polymorphisms in Aβ(1–40) FibrilsTakuma Okada†, Keisuke Ikeda, Masaki Wakabayashi†,Mariko Ogawa and Katsumi Matsuzaki⁎

Graduate School ofPharmaceutical Sciences, KyotoUniversity, Sakyo-ku, Kyoto606-8501, Japan

Received 29 March 2008;received in revised form24 July 2008;accepted 24 July 2008Available online30 July 2008

*Corresponding author. E-mail [email protected].† T.O. and M.W. are Research Fel

Society for the Promotion of ScienceAbbreviations used: Aβ, amyloid

Alzheimer's disease; AM, acetoxy1-anilinonaphthalene-8-sulfonate; Ethomodimer-1; Mem-fibrils, Aβ(1–40GM1 ganglioside-containing membrAβ(1–40) fibrils formed with GM1 mgrowth factor; PBS, phosphate-buffewith Tween-20; PC, egg yolk-L-α-phSM, sphingomyelin; Sol-fibrils, Aβ(1solution; TEM, transmission electronthioflavin T.

0022-2836/$ - see front matter © 2008 E

The abnormal aggregation and deposition of amyloid β protein (Aβ) onneuronal cells are critical to the onset of Alzheimer's disease. The entity(oligomers or fibrils) of toxicAβ species responsible for the pathogenesis of thedisease has been controversial. We have reported that the Aβ aggregates onganglioside-rich domains of neuronal PC12 cells as well as in raft-like modelmembranes.Here,we identified toxicAβ(1–40) aggregates formedwithGM1-ganglioside-containing membranes. Aβ(1–40) was incubated with raft-likeliposomes composed of GM1/cholesterol/sphingomyelin at 1:2:2 and 37 °C.After a lag period, toxic amyloid fibrils with a width of 12 nm were formedand subsequently laterally assembled with slight changes in their secondarystructure as confirmed by viability assay, thioflavin-T fluorescence, circulardichroism, and transmission electron microscopy. In striking contrast, Aβfibrils formed without membranes were thinner (6.7 nm) and much less toxicbecause of weaker binding to cell membranes and a smaller surface hydro-phobicity. This study suggests that toxic Aβ(1–40) species formed onmembranes are not soluble oligomers but amyloid fibrils and that Aβ(1–40)fibrils exhibit polymorphisms.

© 2008 Elsevier Ltd. All rights reserved.

Keywords: Alzheimer's disease; amyloid β protein; raft-like membranes;toxicity; polymorphism
Edited by J. Bowie
Introduction

The deposition of aggregated (fibrillar) amyloid βprotein (Aβ) as senile plaques in the brain is one of

ess:

lows of the Japan.β protein; AD,

methyl ester; ANS,hiD-1, ethidium) fibrils formed withanes; Mic-fibrils,icelles; NGF, nerve

red saline; PBS-t, PBSosphatidylcholine;–40) fibrils formed inmicroscopy; ThT,

lsevier Ltd. All rights reserve

the pathological features of Alzheimer's disease(AD), the most common type of dementia.1–3 It isessential to identify the entity of neurotoxic Aβspecies responsible for etiology to understand themechanism of the onset of AD and to devisetherapeutic approaches for the disease. Early studiesclearly demonstrated that formation of Aβ fibrils wasconsidered critical.4 However, it is not necessarilythe case that there is a strong correlation betweenthe severity of dementia and the quantity ofhistological Aβ fibrils in the AD brains,2 althoughAβ fibrils were reported to trigger functionaldisorder in neuronal cells and cell death.5–13 Thediscrepancy has been explained by assuming thatsoluble Aβ oligomers, alternative to amyloid fibrils,play a pivotal role in the onset of AD.7,14–23 AD brainsshowed immunoreactivity against an antioligomericAβ antibody.19 However, it was warned that artifac-tual Aβ oligomers could be misconceived as criticalspecies.24On the other hand, another possibility is that

toxic and nontoxic Aβ fibrils coexist in AD brains.

d.

Fig. 1. Formation of Aβ(1–40) aggregates in thepresence of raft-like liposomes and their cytotoxicity.Aβ(1–40) (50 μM) in PBS was incubated in the presence(•) or absence (▴) of liposomes composed of GM1/cholesterol/SM at 1:2:2 ([GM1]=50 μM) and 37 °C with-out agitation. The aggregation of Aβ was monitored withthe ThT assay (black symbols) (mean±SE; n=3). The cellviability was determined as a function of the preincuba-tion time based on the fluorescence intensity of calcein/AM. Preincubated at 37 °C for various periods were Aβalone ( ), liposomes alone ( ), and a mixture of Aβ andliposomes ( ). The viability of cells treated with the Aβ–liposome mixture was also estimated with EthiD-1 ( )(mean±SE; n=6; *pb0.001 against vehicle treatment).Each sample was mixed with the same volume ofmedium containing NGF before its application to cells[final Aβ(1–40) concentration=25 μM]. The assays wererun in triplicate independently.

Fig. 2. Effects of lipid compositions on the fibrilformation by Aβ. Aβ(1–40) (50 μM) in PBS was incubatedin the presence of liposomes composed of GM1/SM/cholesterol (1:2:2), GM1/PC (2:8), GM1/PC (4:6), and SM/cholesterol (1:1) ([lipid]=250 μM) at 37 °C for 19 h. The fibrilformation was evaluated by the ThT assay.

1067Formation of Toxic Aβ(1–40) Fibrils on GM1-Containing Membranes

Aβ(1–40) fibrils show polymorphisms.25,26 Theextent of mechanical stress during the preparationof Aβ(1–40) fibrils affects their structure and cyto-toxicity.27 The toxicity of Aβ fibrils is also influencedby their surface structures.28 It should be stressedthat the two hypotheses, oligomers versus fibrils, arenot mutually exclusive.The various Aβ oligomers and fibrils discussed

above were formed in aqueous solution. Accumulat-ing evidence suggests that membranes, especiallyganglioside-containing ones, play an important rolein the aggregation of Aβ.29,30 Yanagisawa et al. dis-covered monosialoganglioside GM1-bound Aβ inhuman brains exhibiting early pathological changesassociated with AD.31 The GM1-bound Aβ wasshown to serve as a seed for the formation of amy-loid aggregates in model membranes32,33 and in neu-ronal PC12 cells.34–36 The involvement of protein–membrane interactions has also been reported for thepathogenesis of other amyloid diseases, such as typeII diabetes.37–39

In this study, we characterized toxic Aβ(1–40)species formed on GM1-containing raft-like mem-branes. Circular dichroism (CD), thioflavin-T (ThT)assay, centrifugation, and transmission electronmicroscopy (TEM) revealed that the toxic specieswere typical amyloid fibrils. Furthermore, we foundthat Aβ(1–40) fibrils showed polymorphisms. Aβfibrils formed in solution (Sol-fibrils) exhibited differ-ent physicochemical properties and were much lesstoxic. The importance of various toxic Aβ aggregatesformed with GM1 is discussed.

Results

Cytotoxicity of Aβ(1–40) incubated withGM1-containing liposomes

The formation of Aβ(1–40) aggregates was moni-tored with ThT fluorescence, an amyloid-specificfluorescent dye, in the presence (Fig. 1, black circles)or absence (Fig. 1, black triangles) of liposomes com-posed of GM1/cholesterol/sphingomyelin (SM) at1:2:2 and 37 °C in phosphate-buffered saline (PBS)without agitation. A low GM1-to-Aβ ratio of 1 wasselected to reduce the aggregation rate and therebyclearly observe the onset of aggregation, at which theamount of oligomers is maximal at least during thefibrillization in solution.40 In the absence of lipo-somes, Aβ(1–40) did not aggregate even after 13 h. Incontrast, in the presence of liposomes, the aggregationof Aβ(1–40) started after a lag time of 6 h and reacheda plateau at 9 h. The existence of GM1 in a raft-likeenvironment was essential for the acceleration of Aβaggregation: Neither nonraft GM1/egg yolk-L-α-phosphatidylcholine (PC) mixtures (2:8 and 4:6) norraft-like SM/cholesterol (1:1) membranes withoutGM1 triggered the formation of Aβ amyloids (Fig. 2).PC12 cells were differentiated into neuron-like cells

with nerve growth factor (NGF) and treated withvarious Aβ preparations for 24 h for a viability assayusing calcein acetoxymethyl ester (calcein/AM).GM1-containing liposomes alone had no cytotoxicity(Fig. 1, blue squares). In the absence of liposomes,Aβ(1–40) did not show any cytotoxicity (Fig. 1, bluetriangles). Cytotoxicity of Aβ(1–40) with liposomeswas dependent on the preincubation period (Fig. 1,blue circles). The 7-h incubated mixtures at the onsetof aggregation were almost nontoxic, whereas themixtures that had been incubated for 7.5 h (halfway ofThT fluorescence increase) or 13 h (plateau) exhibitedsignificant cytotoxicity. Cytotoxicity of the mixturesevaluatedwith ethidiumhomodimer-1 (EthiD-1) (Fig.1, red circles) was similar, indicating that the percent-age of cell viability value was not dependent on theassay method.

Fig. 3. Structural change of Aβ(1–40) aggregates. CDspectra of the samples of Fig. 1 at 0 h (trace 1, broken line),7.5 h (trace 2, full line), as well as 13 h (trace 3, dotted line)and Sol-fibrils (trace 4, dashed line).

1068 Formation of Toxic Aβ(1–40) Fibrils on GM1-Containing Membranes

Secondary structures

CDmeasurements were carried out to estimate thesecondary structures of Aβ(1–40) after variousincubation periods (Fig. 3). The spectrum for non-incubated protein was characteristic of a random-coilconformation (Fig. 3, broken line 1). The proteinsincubated for 7.5 h (Fig. 3, full line 2) or 13 h (Fig. 3,dotted line 3) assumed aβ-sheet-rich conformation, asjudged from a minimum around 220 nm, in accor-dance with the increase in ThT fluorescence (Fig. 1,closed circles). Sol-fibrils assumed a β-sheet-rich

structure (Fig. 3, dashed line 4) different from that ofAβ(1–40) aggregates formed in the presence ofliposomes.

Isolation of cytotoxic Aβ(1–40) species

The cytotoxic Aβ aggregates in the 7.5-h or 13-hincubation mixture should be insoluble large aggre-gates, membrane-bound species, or soluble oligo-mers. The incubation mixture was fractionated viatwo steps (Fig. 4, above). First, the sample was gentlycentrifuged and washed five times (10,000g, 10 min)to isolate insoluble aggregates. Next, the supernatantfrom the first step was again fractionated by ultra-centrifugation (200,000g, 3 h) (the second step) toobtain membrane-bound Aβ (pellet). Each fractionwas then resuspended with PBS in the originalvolume. With the pellet from the first step (sample2), larger Aβ aggregates showed a ThT fluorescenceintensity and an Aβ concentration almost equal tothose of the originalmixture (sample 1). The pellet Aβconcentrations of the 7.5-h incubatedmixture and the13-h incubated mixture were 49.7±0.6 and 47.7±0.7 μM (mean±SD; n=2), respectively. In contrast, themembrane (sample 3) and the soluble (sample 4)fractions were ThT negative. GM1-containing lipo-somes were recovered from the membrane fraction(N98% as determined from the GM1 concentration),suggesting that liposomes had not contaminated theinsoluble aggregate fraction.

Fig. 4. Isolation of toxic Aβ spe-cies. The scheme shows fractionationprocedures. The cytotoxicity of thesamples in Fig. 1 at (a) 7.5 h and(b) 13 h was determined with cal-cein/AM.Samples: 1,wholemixture;2, insoluble aggregate fraction—resuspended pellet after gentle cen-trifugation (10,000g, 10 min, fivetimes); 3, membrane fraction—pelletrecovered from ultracentrifugation(200,000g, 4 °C, 3 h); and 4, solublefraction—supernatant after ultracen-trifugation (mean±SE; n=6; *pb0.001against vehicle treatment).

Fig. 6. Cytotoxicity of Aβ aggregates. Aβ cytotoxicitywas estimated with the fluorescence intensity of calcein(mean±SE; n=6; *pb0.001 against vehicle treatment), (1)A2 (7.5 h, insoluble aggregate fraction) and (2) B2 (13 h,insoluble aggregate fraction) of Fig. 4, (3) Sol-fibrils, and(4) the mixture of raft-like liposomes [GM1/cholesterol/SM (1:2:2), [GM1]=25 μM] and sample 3. The finalconcentration of Aβ(1–40) was 25 μM.


Each resuspended fraction was mixed with thesame volume of the NGF-containing serum-freemedium and applied to NGF-differentiated PC12cells. The viability assay was carried out after a 24-htreatment. The insoluble aggregate fraction (sample 2)showed cytotoxicity identical with that of the originalmixture (sample 1), whereas neither the membrane(sample 3) nor the soluble fraction (sample 4) from thesecond step was toxic (Fig. 4a and b). The cytotoxicspecies generated in the presence of GM1-containingliposomeswas successfully recovered in the insolubleaggregate fraction. Hence, this fraction was used inthe subsequent assays.

Characterization of cytotoxic Aβ(1–40) species

TEM images indicated that the cytotoxic speciesisolated from the 7.5-h incubated specimen weretypical nonbranched fibrils (width=12.0±0.7 nm)(Fig. 5a), whereas fibrils in the 13-h incubatedmixtureassembled laterally (Fig. 5b). On the other hand,Sol-fibrils were thinner fibrils or protofilaments(width=6.7±1.3 nm) (Fig. 5c), and the protofila-ments associated laterally and twisted into rope-likefibrils (width=14.5±0.9 nm).Aβ(1–40) fibrils formedwith liposomes (Mem-fibrils)

were cytotoxic (Fig. 6, columns 1 and 2), whereasSol-fibrils weremuch less toxic (Fig. 6, column 3). Thecoexistence of raft-like liposomes did not augmentthe cytotoxicity of Sol-fibrils (Fig. 6, column 4).Next, the binding of the fibrils to cellswas examined

with the use of Congo red assay (Fig. 7). Mem-fibrilsexhibited significantly stronger binding to cell mem-branes (Fig. 7a, b, d1, and d2) compared with Sol-fibrils (Fig. 7c and d3).The 1-anilinonaphthalene-8-sulfonate (ANS) assay

was carried out to estimate the hydrophobicity of Aβaggregates (Fig. 8). ANS is only weakly fluorescent inaqueous phase; however, the intensity increases witha blue shift in hydrophobic environments. Therefore,the dye is frequently used to monitor the presence ofhydrophobic surfaces in proteins.41 No increase influorescence was observed when Aβ(1–40) monomerwas added. In contrast, Aβ fibrils showed a sig-nificant increase in fluorescence accompanied by ablue shift in the rank order Mic-fibrils (Aβ fibrilsprepared in the presence of GM1 micelles)N7.5-hMem-fibrilsN13-h Mem-fibrilsNSol-fibrils.

Discussion

In this study, the formation of cytotoxic Aβ(1–40)species was reproducibly achieved in the presence of

GM1-containing liposomes mimicking lipid rafts inneuronal cell membranes. The toxic species werefound to be typical nonbranched amyloid fibrilswith β-sheet-rich conformations (Figs. 3–5). If oligo-mers coexisting with fibrils, although not observedin TEM, had been cytotoxic, sample 4 in Fig. 4 wouldhave also exhibited cytotoxicity. Aβ–liposome mix-tures were subjected to dot blot assay using theoligomer-specific antibody A11 to confirm theabsence of oligomers.19 Oligomers were not detect-ed during the aggregation process (Fig. 9).The cytotoxic fibrils were formed halfway during

the aggregation process, as judged from ThT fluores-cence (Fig. 1). However, the centrifugation assayrevealed that almost all proteinswere recovered in thefibril fraction at this time point, indicating that thesubsequent increase in ThT fluorescence was due tosome conformational change of Aβ within fibrils, asreflected by a slight change in the CD spectra (Fig. 3).The lateral association of fibrils (Fig. 5b) may also berelated to this structural change. Indeed, the fluores-cence intensity of ThT increases by morphologicalchanges of Aβ fibrils.40 The rearrangement of β-strandswithin β-sheets has also been reported duringthe annealing of a prion protein.42 There was a corre-lation among cytotoxicity, cell binding of fibrils, andthe hydrophobicity of fibrils (Figs. 6–8). Loss of thehydrophobic surface may reduce the affinity of fibrilsfor neuronal cells and thereby their neurotoxicity.We previously reported that Mic-fibrils are also

toxic.43 Contrary to fibrillization mediated by GM1-containing liposomes, in which toxic Aβ fibrils are

Fig. 5. TEM images of cytotoxicAβ(1–40) species. (a) A2 (7.5 h, inso-luble aggregate fraction) and (b) B2(13 h, insoluble aggregate fraction)of Fig. 4 and (c) Sol-fibrils. The barrepresents 50 nm.

Fig. 7. Binding of Aβ aggregatesto NGF-differentiated PC12 cells.(a–c) Confocal laser scanningmicro-graphs of Aβ aggregates bound toNGF-differentiated PC12 cells stainedwith Congo red superimposed ondifferential interference contrastimages. (d) Fluorescence intensityof Congo red per cell (mean±SE;n≈100; *pb0.001). Cells were incu-bated for 30 min with (1) A2 (7.5 h,insoluble aggregate fraction) and (2)B2 (13 h, insoluble aggregate frac-tion) of Fig. 4 and (3) Sol-fibrils. Thefinal concentration of Aβ(1–40) was25 μM.


formed halfway of the ThT fluorescence increase (Fig.1), the toxicity of Mic-fibrils is correlated with theincrease in the fluorescence intensity of ThT, reflectingthat the GM1-mediated aggregation processes aredifferent. Indeed, the fluorescence intensity of ANSwas higher for Mic-fibrils than for Mem-fibrils,indicating a higher surface hydrophobicity of theformer (Fig. 8). Thus, the neuronal membranecomponent GM1 triggers the formation of varioustypes of cytotoxic Aβ fibrils.Polymorphisms of amyloid fibrils are not limited to

Aβ. Amyloid fibrils by mammalian prion protein,44

α-synuclein,45 insulin,46 and human amylin47 alsoexhibit polymorphisms. Distinct structures of HET-sprion amyloids correlate with prion infectivity.48

Not only toxic fibrils but also toxic soluble oligo-mers are formed in GM1-containing membranesdepending on the conditions. Yamamoto et al. isolatedtoxic soluble Aβ oligomers formed in membranescontaining limited amounts of GM1.49 The morphol-

Fig. 8. ANS assay of Aβ aggregates. Fluorescencespectra of ANS (5.0 μM) in PBS were measured in thepresence of Aβ (2.5 μM) with an excitation wavelength of350 nm. Samples: 1, Mic-fibrils; 2, A2 of Fig. 4 (7.5 h,insoluble aggregate fraction); 3, B2 of Fig. 4 (13 h, insolubleaggregate fraction); 4, Sol-fibrils; 5, monomeric Aβ(1–40);and 6, ANS only.

ogy of Aβ aggregates may considerably depend onmembrane composition, especially GM1 content.Yuan and Johnston reported that the shape of GM1clusters in GM1/dipalmitoylphosphatidylcholine/cholesterol membranes changes from round tofilamentous as the proportion of GM1 increases.50

Such differences may lead to the formation of Aβaggregates with different morphologies.This study elucidated that Mem-fibrils have

toxicity and morphology different from those ofSol-fibrils, indicating that the role of membranes inAβ fibrillization is not merely the acceleration of Aβaggregation but also the generation of toxic fibrils.Furthermore, the importance of polymorphisms ofAβ fibrils was suggested in the pathogenesis of AD.

Fig. 9. Immunoreactivity of Aβ incubated with raft-likeliposomes. Amixture of Aβ(1–40) and raft-like liposomes inPBS incubated at 37 °C for various periods was subjected todot blot assay and probed with the anti-Aβ(1–17) antibody6E10 or the oligomer-specific antibody A11 (upper panel).The aggregation ofAβwas alsomonitored by the ThTassay(lower panel). Monomers and fibrils were reacted with6E10, whereas oligomers were absent.


In addition to oligomers, fibrils also appear to con-tribute to neurodegeneration.

Experimental Procedures

Materials

Dulbecco's modified Eagle's medium, horse serum,bovine serum, penicillin, streptomycin, and EthiD-1 werepurchased from Invitrogen (Carlsbad, CA). Aβ(1–40) wasobtained from Peptide Institute (Mino, Japan). Calcein/AMand GM1 from bovine brain were purchased from DojindoLaboratories (Kumamoto, Japan) andWako (Tokyo, Japan),respectively. Cholesterol, SM from bovine brain, and PCwere obtained from Sigma (St. Louis, MO). The rabbit anti-oligomer antibody A11, the mouse anti-Aβ(1–17) antibody6E10, and horseradish peroxidase-conjugated antirabbit orantimouse immunoglobulin G were purchased fromBioSource (Camarillo, CA), Millipore (Billerica, MA), andSanta Cruz Biotechnology (Santa Cruz, CA), respectively.All other chemicals were obtained from Nacalai Tesque(Kyoto, Japan).

Protein solution

Aβwasdissolved in 0.02%ammonia on ice, and any largeaggregate that could act as a seed for aggregation wasremovedbyultracentrifugation in 500-μLpolyallomer tubesat 540,000g and 4 °C for 3 h. The protein concentration of thesupernatant was determined in triplicate with the use ofMicro BCA Protein Assay (Pierce, Rockford, IL). Thesupernatant was collected and stored at −80 °C prior touse. Just before the experiment, the stock solution wasmixed with an equal volume of doubly concentrated PBS(16.0 g/L of NaCl, 0.40 g/L of KCl, 2.30 g/L of Na2HPO4,and 0.40 g/L of KH2PO4, pH 7.4). Aβ(1–40) assumes amonomeric state under these conditions, as previouslydescribed.43

Liposomes

Sonicated liposomes were prepared as reported pre-viously.51 GM1, cholesterol, PC, and SMwere dissolved in achloroform/methanol (1:1, v/v) mixture, chloroform,chloroform, and ethanol, respectively. The concentrationsof GM1, cholesterol, and SM (PC) were determined at leastin triplicate with the use of the resorcinol–hydrochloric acidmethod,52 the cholesterol oxidase method [Free CholesterolE-Test kit by Wako (Osaka, Japan)],53 and the phosphorusassay,54 respectively. Lipids were mixed at the desiredmolar ratio, and the solventwas removed by evaporation ina rotary evaporator. The residual lipid film, after dryingunder vacuum overnight, was hydrated with PBS (8.00 g/LofNaCl, 0.20 g/L of KCl, 1.15 g/L ofNa2HPO4, and 0.2 g/Lof KH2PO4, pH 7.4) and vortex mixed to produce multi-lamellar vesicles, whichwere subsequently sonicated undera nitrogen atmosphere for 9 min (3 min×3 times) using aprobe-type sonicator, Tomy UD-201 (Tokyo, Japan). Metaldebris from the titanium tip of the probe was removed bycentrifugation. The concentration of vesicles was deter-mined based on the concentration of GM1.51

ThT assay

The ThTassaywas carried out as follows:55,56 The sample(final Aβ concentration=0.5 μM) was added to 5 μM ThT

solution in 50 mM glycine buffer (pH 8.5). Fluorescence at490 nm was measured at an excitation wavelength of446 nm on a Shimadzu RF-5300 spectrofluorometer with acuvette holder thermostatted at 25 °C. A calcein solution(1 μM) in 10 mM Tris/150 mM NaCl/1 mM ethylenedia-minetetraacetic acid, pH 7.4, buffer was used as a standard.The statistical analysiswas performedusingANOVA (n=3).

Cell culture

Rat pheochromocytoma PC12 cells were cultured inDulbecco's modified Eagle's medium containing 5% horseserum, 10% bovine serum, 100 U/mL of penicillin, and0.1 mg/mL of streptomycin at 37 °C with 5% CO2. Afterbeing plated at a density of 5000 cells perwell onto a BioCoatpoly-D-lysine-coated 96-well black/clear plate (Becton Dick-inson, England) for a viability assay and at a density of100,000 cells/dish onto a poly-D-lysine-coated 35-mm glass-bottomed dish for confocal laser scan microscopy, cells wereincubated for 24 h at 37 °C with 5% CO2 and differentiatedwith 50 ng/mL of NGF in serum-free medium for 5 days.

Cell viability

Amixture of Aβ(1–40) (50 μM) and GM1-containing raft-like liposomes that had been preincubated for variousperiods was mixed with the same volume of NGF-contain-ing serum-free medium. One hundred microliters of thesample solution [final Aβ(1–40) concentration=25 μM]wasapplied to eachwell and incubated for 24 h at 37 °Cwith 5%CO2. One hundred microliters of a mixture of calcein/AM(final concentration=1 μM) and EthiD-1 (final concentra-tion=1 μM) in PBS was gently added and incubated for 1 hat 37 °C. Calcein/AM is a membrane-permeable dye that iscleaved by intracellular esterases in living cells to producethe fluorophore calcein. EthiD-1 can only penetrate theplasma membranes of dead cells. Fluorescence wasmeasured on a Wallac 1420 ARVO Multilabel Counter(Perkin Elmer). The linearity between the number of livingor dead cells and the fluorescence intensity of calcein(excitation=495 nm; emission=535 nm) or EthiD-1(excitation=530 nm; emission=620 nm) was confirmed. Inthe live cell assay, the fluorescence intensity of cells treatedwith the vehicle only and the background fluorescenceintensity of each well were defined as a positive control(viability=100%) and as a negative control (viability=0%),respectively. In thedead cell assay, the fluorescence intensityof cells treated with the vehicle only and the fluorescenceintensity of cells treatedwith 70%methanol for 30minweredefined as a positive control (viability=100%) and as anegative control (viability=0%), respectively, according tothe manufacturer's instructions. The statistical analysis wasperformed using ANOVA (n=6), and the significance wasassessed against a positive control.

CD measurements

CD spectra were measured at 37 °C on a Jasco J-820apparatus with the use of a 1.0-mm pathlength quartzcuvette to minimize the absorbance due to buffer compo-nents. Eight scans were averaged for each sample. Theaveraged blank spectra (liposomes) were subtracted.

Preparation of amyloid fibrils in solution and withGM1 micelles

Sol-fibrils were prepared as described previously.43 Forthe preparation of seeds, Aβ(1–40) (200 μM) was


incubated for 3 days at 37 °C in PBS, pH 7.4, and theformation of fibrils was confirmed by the ThT assay. TheAβ(1–40) fibril suspension was diluted with ultrapurewater to 7 μM. The solution was sonicated on ice with 20intermittent pulses (pulse=0.6 s; interval=0.4 s; outputlevel=2) using the sonicator equipped with a microtip TP-030. Seed fibrils (1 μM) were mixed with monomericAβ(1–40) (50 μM) in PBS (pH 7.4) and incubated for 12 h at37 °C without any agitation. Aβ fibrils were also preparedin the presence of GM1 micelles (Mic-fibrils). Aβ(1–40)(50 μM)was incubated with GM1micelles (50 μM) for 24 hin PBS at 37 °C as described elsewhere.37 The formation ofamyloid fibrils was confirmed by the ThT assay and TEM.

Transmission electron microscopy

TEM experiments were carried out by the UltrastructureResearch Institute of Hanaichi Co. Ltd. (Okazaki, Japan).Samples were spread on carbon-coated grids, negativelystained with phosphotungstic acid, and examined under aJOEL 1200EX electron microscope with an accelerationvoltage of 100 kV.

Determination of Aβ(1–40) concentration inresuspended pellets

The concentration of Aβ aggregates in resuspendedpellets was determined as follows: The pellets were dis-solved in 90% (v/v)DMSOwith 5% (v/v) aqueous trifluoro-acetic acid to dissociate the aggregates into monomers. Thepeptide concentration was determined from the peak areaon analytical HPLC, which was performed using aCOSMOSIL 5C18-AR-II PackedColumn (4.6×250mm,Naca-lai Tesque) elutedwith a linear gradient of CH3CN (20%–80%CH3CN, 40 min) in 0.1% aqueous trifluoroacetic acid. Theprotein was detected by measuring absorbance at 220 nm.

Congo red assay

NGF-differentiated PC12 cells on a glass-bottomed dishwere incubated with each Aβ sample (final concentra-tion=25 μM) for 30 min, washed twice, stained with20 μMCongo red for 30 min, and washed twice again. Thecells were visualized using 20× C-Apochromat objective ofa Zeiss LSM 510 confocal laser scanning microscope(excitation=543 nm; LP=560 nm). For quantitative esti-mation, cells were scanned in 10 slices along the z-axis andthe fluorescence intensity of Congo red (expressed in 256tones) multiplied by the pixel number was summed andnormalized to the cell number examined (∼100 cells) afterthe background fluorescence had been subtracted.

ANS assay

For estimation of the hydrophobicity of amyloid fibrils,an ANS assay was carried out.41 The amyloid sample (finalAβ concentration=2.5μM)was applied to 5μMANS inPBSin a quartz cuvette and mixed by gentle stirring for 1 min.Fluorescence emission spectra were recorded at an excita-tionwavelength of 350 nmon the spectrofluorometerwith acuvette holder thermostatted at 25 °C.

Dot blot assay

Dot blot assaywas performed as described previously.19

Briefly, 1 μL of the mixture of Aβ (50 μM) and raft-like

liposomes [GM1/cholesterol/SM (1:2:2), [GM1]=50 μM]in PBS at 37 °C at each incubation period was spotted ontonitrocellulose membranes. The membranes were blockedwith 10% nonfat milk in PBS-t (PBS containing 0.01%Tween-20) overnight at 4 °C and washed three times withPBS-t for 5 min. The membranes were reacted with therabbit antioligomer antibody A11 or the mouse anti-Aβ(1–17) antibody 6E10 (diluted 1:500 in 5% nonfat milk PBS-t)for 1 h at room temperature. After washing, the blots wereincubated with horseradish peroxidase-conjugated anti-rabbit or antimouse immunoglobulin G (diluted 1:1000 in5% nonfat milk PBS-t) for 1 h at room temperature. Theblots were developed with Chemi-Lumi One (NacalaiTesque), and chemical luminescence was detected by aLAS-3000mini lumino-image analyzer (FujiFilm, Tokyo,Japan).

Acknowledgements

This study was supported in part by the NaitoFoundation, the Ono Medical Research Foundation,the Takeda Science Foundation, and Grant-in-Aidfor Research Fellowships of the Japan Society for thePromotion of Science for Young Scientists (17-2295).

References

1. Hardy, J. & Selkoe, D. J. (2002). The amyloid hypothesisof Alzheimer's disease: progress and problems on theroad to therapeutics. Science, 297, 353–356.

2. Mattson, M. P. (2004). Pathways towards and awayfrom Alzheimer's disease. Nature, 430, 631–639.

3. Walsh, D. M. & Selkoe, D. J. (2004). Deciphering themolecular basis of memory failure in Alzheimer'sdisease. Neuron, 44, 181–193.

4. Selkoe, D. J. (1991). The molecular pathology ofAlzheimer's disease. Neuron, 6, 487–498.

5. Dahlgren, K. N., Manelli, A. M., Stine, W. B., Jr., Baker,L. K., Krafft, G. A. & LaDu, M. J. (2002). Oligomericand fibrillar species of amyloid-β peptides differen-tially affect neuronal viability. J. Biol. Chem. 277,32046–32053.

6. Davis-Salinas, J. & Van Nostrand, W. E. (1995).Amyloid beta-protein aggregation nullifies its patho-logic properties in cultured cerebrovascular smoothmuscle cells. J. Biol. Chem. 270, 20887–20890.

7. Deshpande, A., Mina, E., Glabe, C. & Busciglio, J.(2006). Different conformations of amyloid β induceneurotoxicity by distinct mechanisms in humancortical neurons. J. Neurosci. 26, 6011–6018.

8. Grace, E. A., Rabiner, C. A. & Busciglio, J. (2002).Characterization of neuronal dystrophy induced byfibrillar amyloid β: implications for Alzheimer'sdisease. Neuroscience, 114, 265–273.

9. Jana,A.&Pahan,K. (2004). Fibrillar amyloid-βpeptideskill human primary neurons via NADPH oxidase-mediated activation of neutral sphingomyelinase.Implications for Alzheimer's disease. J. Biol. Chem. 279,51451–51459.

10. Lorenzo, A. & Yankner, B. A. (1996). Amyloid fibriltoxicity in Alzheimer's disease and diabetes. Ann. N. Y.Acad. Sci. 777, 89–95.

11. Pillot, T., Drouet, B., Queille, S., Labeur, C., Vande-kerchkhove, J., Rosseneu, M. et al. (1999). The


nonfibrillar amyloid β-peptide induces apoptotic neu-ronal cell death: involvement of its C-terminal fusogenicdomain. J. Neurochem. 73, 1626–1634.

12. Tamagno, E., Bardini, P., Guglielmotto, M., Danni, O.& Tabaton, M. (2006). The various aggregation statesof β-amyloid 1–42 mediate different effects on oxida-tive stress, neurodegeneration, and BACE-1 expres-sion. Free Radical Biol. Med. 41, 202–212.

13. Tsai, J., Grutzendler, J., Duff, K. & Gan, W. B. (2004).Fibrillar amyloid deposition leads to local synapticabnormalities and breakage of neuronal branches.Nat. Neurosci. 7, 1181–1183.

14. Chromy, B. A., Nowak, R. J., Lambert, M. P., Viola,K. L., Chang, L., Velasco, P. T. et al. (2003). Self-assembly of Aβ(1–42) into globular neurotoxins.Biochemistry, 42, 12749–12760.

15. Gong, Y., Chang, L., Viola, K. L., Lacor, P. N., Lambert,M. P., Finch, C. E. et al. (2003). Alzheimer's disease-affected brain: presence of oligomeric Aβ ligands(ADDLs) suggests a molecular basis for reversiblememory loss. Proc. Natl. Acad. Sci. USA, 100,10417–10422.

16. Lacor, P. N., Buniel, M. C., Chang, L., Fernandez, S. J.,Gong, Y., Viola, K. L. et al. (2004). Synaptic targeting byAlzheimer's-related amyloid β oligomers. J. Neurosci.24, 10191–10200.

17. Lambert, M. P., Barlow, A. K., Chromy, B. A.,Edwards, C., Freed, R., Liosatos, M. et al. (1998).Diffusible, nonfibrillar ligands derived from Aβ1–42are potent central nervous system neurotoxins. Proc.Natl. Acad. Sci. USA, 95, 6448–6453.

18. Demuro,A.,Mina, E.,Kayed,R.,Milton, S.C., Parker, I.& Glabe, C. G. (2005). Calcium dysregulation andmembrane disruption as a ubiquitous neurotoxicmechanism of soluble amyloid oligomers. J. Biol.Chem. 280, 17294–17300.

19. Kayed, R., Head, E., Thompson, J. L., McIntire, T. M.,Milton, S. C., Cotman, C. W. & Glabe, C. G. (2003).Common structure of soluble amyloid oligomersimplies common mechanism of pathogenesis. Science,300, 486–489.

20. Hoshi, M., Sato, M., Matsumoto, S., Noguchi, A.,Yasutake, K., Yoshida, N. & Sato, K. (2003). Sphericalaggregates of β-amyloid (amylospheroid) show highneurotoxicity and activate tau protein kinase I/glycogen synthase kinase-3β. Proc. Natl. Acad. Sci.USA, 100, 6370–6375.

21. Barghorn, S., Nimmrich, V., Striebinger, A., Krantz, C.,Keller, P., Janson, B. et al. (2005). Globular amyloidβ-peptide oligomer—a homogenous and stable neuro-pathological protein in Alzheimer's disease. J. Neuro-chem. 95, 834–847.

22. Lesne, S., Koh, M. T., Kotilinek, L., Kayed, R., Glabe,C. G., Yang, A. et al. (2006). A specific amyloid-βprotein assembly in the brain impairs memory.Nature,440, 352–357.

23. Walsh, D. M., Klyubin, I., Fadeeva, J. V., Cullen, W. K.,Anwyl, R., Wolfe, M. S. et al. (2002). Naturally secretedoligomers of amyloid β protein potently inhibithippocampal long-term potentiation in vivo. Nature,416, 535–539.

24. Hepler, R. W., Grimm, K. M., Nahas, D. D., Breese, R.,Dodson, E. C., Acton, P. et al. (2006). Solution statecharacterization of amyloid β-derived diffusibleligands. Biochemistry, 45, 15157–15167.

25. Goldsbury, C., Frey, P., Olivieri, V., Aebi, U. & Muller,S. A. (2005). Multiple assembly pathways underlieamyloid-β fibril polymorphisms. J. Mol. Biol. 352,282–298.

26. Goldsbury, C. S., Wirtz, S., Muller, S. A., Sunderji, S.,Wicki, P., Aebi, U. & Frey, P. (2000). Studies on the invitro assembly of Aβ 1–40: implications for the searchfor Aβ fibril formation inhibitors. J. Struct. Biol. 130,217–231.

27. Petkova, A. T., Leapman, R. D., Guo, Z., Yau, W. M.,Mattson, M. P. & Tycko, R. (2005). Self-propagating,molecular-level polymorphism in Alzheimer's β-amyloid fibrils. Science, 307, 262–265.

28. Yoshiike, Y., Akagi, T. & Takashima, A. (2007). Surfacestructure of amyloid-β fibrils contributes to cytotoxi-city. Biochemistry, 46, 9805–9812.

29. Matsuzaki, K. (2007). Physicochemical interactions ofamyloid β-peptide with lipid bilayers. Biochim. Bio-phys. Acta, 1768, 1935–1942.

30. Yanagisawa, K. (2007). Role of gangliosides inAlzheimer's disease. Biochim. Biophys. Acta, 1768,1943–1951.

31. Yanagisawa, K., Odaka, A., Suzuki, N. & Ihara, Y.(1995). GM1 ganglioside-bound amyloid β-protein(Aβ): a possible form of preamyloid in Alzheimer'sdisease. Nat. Med. 1, 1062–1066.

32. Hayashi, H., Kimura,N., Yamaguchi,H.,Hasegawa,K.,Yokoseki, T., Shibata, M. et al. (2004). A seed forAlzheimer amyloid in the brain. J. Neurosci. 24,4894–4902.

33. Kakio, A., Nishimoto, S., Yanagisawa, K., Kozutsumi, Y.& Matsuzaki, K. (2002). Interactions of amyloid β-protein with various gangliosides in raft-like mem-branes: importance of GM1 ganglioside-bound form asan endogenous seed for Alzheimer amyloid. Biochem-istry, 41, 7385–7390.

34. Wakabayashi, M. & Matsuzaki, K. (2007). Formationof amyloids by Aβ-(1–42) on NGF-differentiated PC12cells: roles of gangliosides and cholesterol. J. Mol. Biol.371, 924–933.

35. Wakabayashi, M., Okada, T., Kozutsumi, Y. &Matsuzaki, K. (2005). GM1 ganglioside-mediated accu-mulation of amyloid β-protein on cell membranes.Biochem. Biophys. Res. Commun. 328, 1019–1023.

36. Yamamoto, N., Fukata, Y., Fukata,M. &Yanagisawa, K.(2007). GM1-ganglioside-induced Aβ assembly onsynaptic membranes of cultured neurons. Biochim.Biophys. Acta, 1768, 1128–1137.

37. Brender, J. R., Durr, U. H., Heyl, D., Budarapu, M. B. &Ramamoorthy, A. (2007). Membrane fragmentationby an amyloidogenic fragment of human islet amyloidpolypeptide detected by solid-state NMR spectro-scopy of membrane nanotubes. Biochim. Biophys. Acta,1768, 2026–2029.

38. Brender, J. R., Lee, E. L., Cavitt, M. A., Gafni, A., Steel,D. G. & Ramamoorthy, A. (2008). Amyloid fiberformation and membrane disruption are separateprocesses localized in two distinct regions of IAPP,the type-2-diabetes-related peptide. J. Am. Chem. Soc.130, 6424–6429.

39. Mascioni, A., Porcelli, F., Ilangovan,U., Ramamoorthy,A. & Veglia, G. (2003). Conformational preferences ofthe amylin nucleation site in SDS micelles: an NMRstudy. Biopolymers, 69, 29–41.

40. Benseny-Cases, N., Cocera, M. & Cladera, J. (2007).Conversion of non-fibrillar beta-sheet oligomers intoamyloid fibrils in Alzheimer's disease amyloid peptideaggregation. Biochem. Biophys. Res. Commun. 361,916–921.

41. Turner, D. C. & Brand, L. (1968). Quantitative estima-tion of protein binding site polarity. Fluorescence ofN-arylaminonaphthalenesulfonates. Biochemistry, 7,3381–3390.


42. Petty, S.A.&Decatur, S.M. (2005). Intersheet rearrange-ment of polypeptides during nucleation of β-sheetaggregates.Proc. Natl. Acad. Sci. USA, 102, 14272–14277.

43. Okada, T., Wakabayashi, M., Ikeda, K. &Matsuzaki, K.(2007). Formation of toxic fibrils of Alzheimer'samyloid β-protein-(1–40) by monosialogangliosideGM1, a neuronal membrane component. J. Mol. Biol.371, 481–489.

44. Anderson, M., Bocharova, O. V., Makarava, N., Breydo, L.,Salnikov, V. V. & Baskakov, I. V. (2006). Polymorphismand ultrastructural organization of prion proteinamyloid fibrils: an insight from high resolution atomicforce microscopy. J. Mol. Biol. 358, 580–596.

45. Heise, H., Hoyer, W., Becker, S., Andronesi, O. C.,Riedel, D. & Baldus, M. (2005). Molecular-levelsecondary structure, polymorphism, and dynamicsof full-length α-synuclein fibrils studied by solid-stateNMR. Proc. Natl. Acad. Sci. USA, 102, 15871–15876.

46. Jimenez, J. L., Nettleton, E. J., Bouchard, M., Robinson,C. V., Dobson, C. M. & Saibil, H. R. (2002). Theprotofilament structure of insulin amyloid fibrils. Proc.Natl. Acad. Sci. USA, 99, 9196–9201.

47. Goldsbury, C. S., Cooper, G. J., Goldie, K. N., Muller,S. A., Saafi, E. L., Gruijters, W. T. et al. (1997). Poly-morphic fibrillar assembly of human amylin. J. Struct.Biol. 119, 17–27.

48. Ritter, C., Maddelein, M. L., Siemer, A. B., Luhrs, T.,Ernst, M., Meier, B. H. et al. (2005). Correlation ofstructural elements and infectivity of the HET-s prion.Nature, 435, 844–848.

49. Yamamoto,N.,Matsubara, E.,Maeda, S.,Minagawa,H.,Takashima, A., Maruyama, W. et al. (2007). A ganglio-side-induced toxic soluble Aβ assembly. Its enhancedformation from Aβ bearing the Arctic mutation. J. Biol.Chem. 282, 2646–2655.

50. Yuan, C. & Johnston, L. J. (2000). Distribution ofganglioside GM1 in L-alpha-dipalmitoylphosphati-dylcholine/cholesterol monolayers: a model for lipidrafts. Biophys. J. 79, 2768–2781.

51. Matsuzaki, K., Noguch, T., Wakabayashi, M., Ikeda,K., Okada, T., Ohashi, Y. et al. (2007). Inhibitors ofamyloid β-protein aggregation mediated by GM1-containing raft-like membranes. Biochim. Biophys.Acta, 1768, 122–130.

52. Svennerholm, L. (1957). Quantitative estimation ofsialic acids: II. A colorimetric resorcinol–hydrochloricacid method. Biochim. Biophys. Acta, 24, 604–611.

53. Richmond, W. (1973). Preparation and properties of acholesterol oxidase from Nocardia sp. and its applica-tion to the enzymatic assay of total cholesterol inserum. Clin. Chem. 19, 1350–1356.

54. Bartlett, G. R. (1959). Phosphorus assay in columnchromatography. J. Biol. Chem. 234, 466–468.

55. LeVine, H., III (1993). Thioflavine T interaction withsynthetic Alzheimer's disease β-amyloid peptides:detection of amyloid aggregation in solution. ProteinSci. 2, 404–410.

56. Naiki, H. & Gejyo, F. (1999). Kinetic analysis ofamyloid fibril formation. Methods Enzymol. 309,305–318.

formation of toxic aβ(1–40) fibrils on gm1 ganglioside-containing membranes mimicking lipid...

Documents