Download - Role of monomer arrangement in the amyloid self-assembly

1Q1

2Q2

3

4Q3Q45678

9

1011121314

15Q516171819202122

4142

43

44

45

46

47

48

49

Biochimica et Biophysica Acta xxx (2014) xxx–xxx

BBAPAP-39488; No. of pages: 11; 4C: 3, 4, 5, 7, 8

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbapap

Role of monomer arrangement in the amyloid self-assembly

OF

Alexander Portillo a, Mohtadin Hashemi a, Yuliang Zhang a, Leonid Breydo b,Vladimir N. Uversky b,c,d, Yuri L. Lyubchenko a,⁎a Department of Pharmaceutical Sciences, University of Nebraska Medical Center, 986025 Nebraska Medical Center, Omaha, NE 68198-6025, USAb Department of Molecular Medicine, USF Health Byrd Alzheimer's Research Institute, Morsani College of Medicine, University of South Florida, 12901 Bruce B. Downs Blvd. MDC07, Tampa,FL 33647, USAc Department of Biological Science, Faculty of Science, King Abdulaziz University, PO Box 80203, Jeddah 21589, Saudi Arabiad Institute for Biological Instrumentation, Russian Academy of Sciences, Pushchino, Moscow Region 142290, Russia

UN

O

Abbreviations: AFM, Atomic Force Microscopy; AP-GNNQQNY; APS, aminopropyl silatrane; CD, circular dichrtroscopy; DMSO, dimethylsulfoxide; HFIP, 1,1,1,3,3,3-herateconstant;MD,moleculardynamics;NHS-PEG-MAL,N-ene glycol-maleimide; TCEP, tris(2-carboxyethyl)phosptrimethylamine-N-oxide;WLC,wormlike chain approximstate to the transition state⁎ Corresponding author at: Department of Pharma

Pharmacy, COP 1012, University of Nebraska Medical CeCenter, Omaha, NE 68198-6025, USA. Tel.: +1 402 559 1(lab); fax: +1 402 559 9543.

E-mail address: [email protected] (Y.L. Lyubche

http://dx.doi.org/10.1016/j.bbapap.2014.12.0091570-9639/© 2014 Elsevier B.V. All rights reserved.

Please cite this article as: A. Portillo, et al., Rdx.doi.org/10.1016/j.bbapap.2014.12.009

Ra b s t r a c t
a r t i c l e i n f o
23

24

25

26

27

28

29

30

31

32

33

34

35

Article history:Received 19 June 2014Received in revised form 24 November 2014Accepted 10 December 2014Available online xxxx

Keywords:Protein aggregationAmyloidsOligomer self-assemblyAFMNanoimagingForce spectroscopyNeurodegenerative diseases

36

37

38

39

40

ECTED P

Assembly of amyloid proteins into aggregates requires the ordering of themonomers in oligomers and especiallyin such highly organized structures asfibrils. This ordering is accompanied by structural transitions leading to theformation of ordered β-structural motifs in proteins and peptides lacking secondary structures. To characterizethe effect of the monomer arrangements on the aggregation process at various stages, we performed compara-tive studies of the yeast prion protein Sup35 heptapeptide (GNNQQNY) alongwith its dimeric formCGNNQQNY-(d-Pro)-G-GNNQQNY. The (d-Pro)-G linker in this construct is capable of adopting a β-turn, facilitating theassembly of the dimer into the dimeric antiparallel hairpin structure (AP-hairpin). We applied Atomic ForceMicroscopy (AFM) techniques to follow peptide–peptide interactions at the single molecule level, to visualizethe morphology of aggregates formed by both constructs, thioflavin T (ThT) fluorescence to follow the aggrega-tion kinetics, and circular dichroism (CD) spectroscopy to characterize the secondary structure of the constructs.The ThT fluorescence data showed that the AP-hairpin aggregation kinetics is insensitive to the external environ-ment such as ionic strength and pH contrary to themonomers the kinetics of which depends dramatically on theionic strength and pH. The AFM topographic imaging revealed that AP-hairpins primarily assemble into globularaggregates, whereas linearfibrils are primary assemblies of themonomers suggesting that both constructs followdifferent aggregation pathways during the self-assembly. These morphological differences are in line with theAFM force spectroscopy experiments and CD spectroscopy measurements, suggesting that the AP-hairpin isstructurally rigid regardless of changes of environmental factors.

© 2014 Elsevier B.V. All rights reserved.
R
50

51

52

53

54

55
CO
R

1. Introduction

Yeast Prion Protein, Sup35p, has been used as a model for the priondisease phenomenon, particularly for understanding of the structuralaspects of such diseases [1,2]. Sup35p misfolding and aggregationmimic those associatedwithmammalian prion diseases [1,3]. Therefore,

56

57

58

59

60

61

62

63

64

65

66

67

68

dimer, CGNNQQNY-(d-pro)G-oism; DFS, dynamic force spec-xafluoro-propan-2-ol; koff, off-hydroxysuccinimide-polyethyl-hine; ThT, thioflavin T; TMAO,ation; xβ, distance of the ground

ceutical Sciences, College ofnter, 986025 Nebraska Medical971 (office), +1 402 559 1973

nko).

ole of monomer arrangemen

the understanding of the transient states of the misfolded Sup35p willpave the way for development of efficient diagnostics and remediesfor the disease [4,5]. The N-terminal domain of the protein plays an im-portant role in the aggregation of the entire Sup35 prion. More specifi-cally, a seven amino acid sequence that spans from residues 7 to 13,GNNQQNY, has a significant involvement in aggregation of the wholeprotein [1,2,6–12]. In fact, the addition of this peptide seeded aggrega-tion of the whole protein, dramatically accelerating the aggregationprocess.

The solid state NMR structural studies emerged to suggest stronglythe origin of polymorphic variation in amyloid fibrils [11]. Differencesin the packing of these β-sheets, originating from differences in theside-chain packing, register or topology of β-sheets, may explainmorphological variants of fibrillar structure. Solid-state NMR studiesclearly showed the coexistence of three distinct conformations of theGNNQQNY peptide within a single fibril. Different packing arrange-ments of peptides within a fibril were proposed to be responsible forthe observed differences in NMR chemical shifts. These coexistingconformations differ by the degree of the local secondary structure

t in the amyloid self-assembly, Biochim. Biophys. Acta (2014), http://

http://dx.doi.org/10.1016/j.bbapap.2014.12.009

mailto:[email protected]

Journal logo


Imprint logo

http://www.sciencedirect.com/science/journal/15709639

www.elsevier.com/locate/bbapap



T

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

146146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

2 A. Portillo et al. / Biochimica et Biophysica Acta xxx (2014) xxx–xxx

UNCO

RREC

(α-helical versus β-sheet) [11]. This structural variability is in line withthe crystallographic studies where eight classes of so-called stericzippers have been identified [13]. Such variation involves mainlyorientation of peptides and β-sheets with respect to each other.

The structure of the peptides and their arrangements withinassemblies can depend on the aggregate size. In crystals, the Sup35heptapeptide forms a steric zipper arrangement between two β-sheetswith theparallel orientation ofmonomerswithin the sheets [3]. However,the monomer arrangement can be different if the aggregates are smalland interact with the environment. This assumption is supported by arecent publication [14], in which molecular dynamic (MD) simulation ofthe Sup35 octapeptide was performed. The authors showed that thesingle β-sheet structure taken from the crystallographic structuredepends on a number of factors including the aggregate size. The experi-ment revealed that an aggregate can disintegrate into smaller-sized olig-omers or the edge peptides can dissociate sequentially. The authors alsoassumed that a heterogeneous mixture of oligomers of different sizesexist prior to the formation of the critical nucleus. The pH and ionicstrength of the surrounding solution can also play a role which was con-firmed by our recent AFM imaging and force spectroscopy studies [15].

These data suggest that assembly of monomeric unit in the aggre-gates and their secondary structure can define the aggregatemorpholo-gy. To test this hypothesis, we performed comparative studies ofheptapeptide Sup35 peptide (monomer) and its covalent dimericform in which two monomers are covalently attached to each other inthe tail-to-head orientation via the (d-Pro)G dipeptide. The latteraccording to NMR and circular dichroism (CD) spectroscopy studies[16] has the β-turn structure forcing the entire dimer to adopt antipar-allel geometry (AP-hairpin). We show here that although bothconstructs are capable of self-assembly into amyloid aggregates, thepre-arrangement of the monomers into dimers changes dramaticallythe AP-hairpin conformational properties limiting the aggregation pro-pensities of the AP-hairpin compared to the monomer.

2. Materials and methods

2.1. Materials

The peptides NH3+-Cys-Gly-Asn-Asn-Gln-Gln-Asn-Tyr-COOH−

(CGNNQQNY, Monomer) and the hairpin peptide NH3+-Cys-Gly-Asn-Asn-Gln-Gln-Asn-Tyr-DPro-Gly-Gly-Asn-Asn-Gln-Gln-Asn-Tyr-COOH−

(CGNNQQNY(d-Pro)G-GNNQQNY, AP dimer) were synthesized byPeptide 2.0, Inc. (Chantilly, VA). Synthesized peptides were purified byVYDAC-C18 reverse-phase HPLC, and their molecular weight was con-firmed by MALDI-TOF mass spectrometry.

The aminopropyl silatrane (APS) was used at a concentration of167 μM for 30 min for mica surface functionalization as described inref. [17]. The 1.67 mM stock solution of NHS-PEG-MAL (N-hydroxysuccinimide-polyethylene glycol-maleimide, MW = 3400 g/mol), purchased from Laysan Bio, Inc. (Arab, AL), was prepared inDMSO (Sigma-Aldrich Inc., St. Louis, MO) and stored at −20 °C. The10 mM Tris(2-carboxyethyl)phosphine (TCEP) hydrochloride (HamptonResearch Inc.) was prepared in water, and was added to peptide solutionand incubated for 10 min prior to deposition on the PEGylated substrates.TCEP was useful for reducing any disulfide bonds that may have formedbetween two peptides to make the thiol group available to the substrate.

2.2. Methods

2.2.1. Aggregation studies using ThT fluorescenceThe extent of peptide aggregation was followed by characteristic

changes in thioflavin T (ThT) fluorescence. First, the lyophilized peptidepowder was dissolved in either water or buffer in an initial volume of100 μL. The initial concentration wasmeasured using a NanoDrop spec-trophotometer (Thermo Scientific, Wilmington, DE), by detecting theabsorbance of the tyrosine residue at 274 nm, and using an extinction

Please cite this article as: A. Portillo, et al., Role of monomer arrangemendx.doi.org/10.1016/j.bbapap.2014.12.009

ED P

RO

OF

coefficient of 1405 L·mol−1·cm−1 for tyrosine. Then, the peptide solu-tionwas diluted down to 500 μM, using either water or a buffer as a sol-vent [1,2]. For the ThTfluorescence assay, 2 μL aliquots from the sampleswere withdrawn periodically and added to 590 μL of 5 μM ThT solution.Fluorescence intensity was measured on a Cary Eclipse spectrofluorim-eter (Varian Inc. Palo Alto, CA) at 485 nmwhile exciting at 450 nm. Eachreported value is an average of 15 values of fluorescence intensity aftersubtracting out the fluorescence contribution from free ThT.

The initial fluorescence readings were taken immediately afterthe solution was prepared. Stirring the solution at intervals wasdone in order to accelerate the aggregation process. Magnetic stirbars were placed into the peptide solutions, and the samples werestirred on a magnetic stirrer (Hanna Instruments, Smithfield, RI)for 20 min once every 2 h. The ThT data were fitted with the follow-ing sigmoidal equation:

y ¼ T0 þT f−T0

1þ 10 log10 T1=2ð Þ−Tð Þ�Smax; ð1Þ

with T0 being a time of 0 h, Tf being the final time, T1/2 the half-timeof aggregation, T the time point being measured, and Smax being themaximum fluorescence attained in the data set.

The working buffer solutions were prepared at four different pHs:pH 2.0 (HCl/KCl), pH 3.7 (sodium acetate/acetic acid), pH 5.6 (sodiumacetate/acetic acid), and pH 7 (HEPES). A fifth bufferwas used occasion-ally at pH 9.8 (sodium carbonate/sodium bicarbonate). The ionicstrength was adjusted by adding NaCl.

2.2.2. AFM topographic imagingAt the plateau levels of ThT fluorescence, 2 μL of the 500 μMaggrega-

tion mixture was deposited on the mica and allowed to sit for 2 min,followed by the addition of 8 μL of distilled water, which was thenallowed to sit for 2 min. The samples were then dried by 2 min of spincoating at 2000 rpm using a Model WS-400BZ-6NPP/LITE spin coater(Laurell Technologies Corporation, North Wales, PA).

Images were acquired in air using a MultiMode SPM NanoScope VMultimode 8 system (Bruker Nano, Santa Barbara, CA) operating inpeak force mode. Silicon nitride (Si3N4) AFM probe tips (MSNL —

Bruker, Santa Barbara, CA, USA) with nominal spring constants of0.6 N/m were used for peak force imaging. Image analysis was doneusing Femtoscan Online (Advanced Technologies Center, Moscow,Russia). The aggregates inwere analyzedmanually by counting elongat-ed fibrils and globular oligomers after subtracting anything in the back-ground that was less than 1 nm in height.

2.2.3. AFM force spectroscopyFreshly cleaved mica (Asheville-Schoonmaker Mica Co., Newport

News, VA, USA) surfaces were treated with amino-propyl-silatrane(APS) for 30 min according to previously reported protocol [17] follow-ed by rinsing with water and drying with argon gas flow. The APSmodified mica surfaces were treated with 167 μM NHS-PEG-MAL inDMSO for 3 h followed by rinsing with DMSO to remove non-boundNHS-PEG-MAL, rinsing with water, and drying with argon gas flow.Maleimide-functionalizedmicawas incubated for 1 h in 190 nMof pep-tide solution (HEPES buffer, pH 7.0). Prior to immobilization of peptide,the peptide solution was treated with 0.25 mM TCEP hydrochloride for10 min in pH 7.0 buffer to reduce any disulfide bonds. After washingwith HEPES (pH 7.0, 10 mM HEPES, 50 mM NaCl) buffer, unreactedmaleimide was quenched with 10 mM β-mercaptoethanol for 10 minfollowed by rinsing with HEPES pH 7.0 buffer.

Silicon nitride (Si3N4) AFM probe tips (MLCT — Bruker, SantaBarbara, CA) were washed in ethanol by immersion for 30 min andthen activated by UV treatment for 30 min. The activated probe tipswere treated with APS for 30 min. The APS modified probe tips werethen treated with 167 μM NHS-PEG-MAL for 3 h followed by rinsingwith DMSO, and thorough rinsing with double distilled water.




T

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

230230

231

232

233

234

235

236

237

239239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

Fig. 1. Schematic of the structures of sequences for both peptides:monomers andAP-hair-pins. Arrows indicate the N-to-C direction of the sequences with no relation to their sec-ondary structure.

3A. Portillo et al. / Biochimica et Biophysica Acta xxx (2014) xxx–xxx

UNCO

RREC

Maleimide-functionalized AFMprobes were incubated for 1 h in 190 nMof peptide solution (HEPES buffer, pH 7.0). After washing with HEPES(pH 7.0, 10 mM HEPES, 50 mM NaCl) buffer, unreacted maleimides onthe probe were quenched with 10 mM β-mercaptoethanol for 10 minfollowed by rinsing with HEPES pH 7.0 buffer.

Force–distance measurements were performed in pH 5.6 (ionicstrength = 11 mM) buffer at room temperature with the Force Robot300 (JPK Instruments, Berlin, Germany). The ramp size was 200 nmwith various loading rates. An application force was kept at a lowvalue (100 pN). Silicon nitride cantileverswith nominal values of springconstants in the range of 0.04–0.07 N/mwere used. The approach veloc-ity was kept at 500 nm/s while the retraction velocity was varied be-tween 100 nm/s and 2000 nm/s, and the corresponding values ofapparent loading rates were between 102 and 105,000 pN/s. The mostprobable rupture force was obtained from the probability function fit[18] of the force distribution compiled in a statistical histogram. Dataanalysis in dynamic force spectroscopy was performed as described inrefs. [19–22].

2.2.4. Dynamic force spectroscopyThe force curves were analyzed using wormlike chain (WLC) ap-

proximation describing behavior of polymer linkers under an appliedexternal force and the Igor Pro 6.04 software package provided by themanufacturer [15,23,24]. The contour lengths and persistence lengthsof the unbinding of peptides were determined from the WLC fit of ex-perimental force–distance curves. The unbinding events where interac-tions between peptides are ruptured upon tip retraction appear at acertain distance in the experimental force–distance curves defined bythe length of the flexible PEG linker. The analysis of the unbindingevents included fitting the part of the force–distance curve with WLCmodel. The fit provided contour length and persistence length of thelinkers.

Each force curve was fitted using the WLC model. The approach ve-locity was kept at 500 nm/swhile the retraction velocity was varied be-tween 80 nm/s and6000nm/s. At each retraction velocity value, 1000 to6000 force curves are obtained. The cantilever–linker–molecule systempossesses an apparent spring constant (kc) due to molecular coupling.The apparent loading rate (r) defined by the slope of the force curve im-mediately before the position of the rupture point on the force curvewas determined using the following equation [7,23]:

1=r ¼ 1=kcv 1þ kcLc=4ffiffiffiffiffiffiffiffiffiffiffiffiffiffiFP=F

3q� �

; ð2Þ

where Fp = kBT/Lp, kc is the cantilever spring constant (pN/nm), v is tipvelocity (nm/s), Lc and Lp are contour length (nm) and persistencelength (nm), respectively, which are parameters ofWLC fitting, F is rup-ture force (pN), and r is loading rate (pN/s).

All of the forces at each apparent rate were put together. The appar-ent loading rates were between 100 and 100,000 pN/s, which corre-spond to pulling velocities of 80 to 6000 nm/s. The range of loadingrates was divided into seven parts, where the force was obtainedusing the probability function through the following equation [25]:

p Fð Þ ¼ koff 0ð Þ exp fxβ=kBT� �

� 1=rð Þ exp −koff 0ð ÞZ F

0exp fxβ=kBT

� �1=rð Þdf Þ;

�ð3Þ

where koff(0) is the off-rate constant of the complex at zero force, xβ isthe distance of the transition state to bound state, kB is Boltzmannconstant, T is the absolute temperature, and r is the loading rate, thatis dF/dt. The probability density of rupture force, p(F) was calculated ac-cording to Eq. (2) in the measured force histogram.


ED P

RO

OF

2.2.5. Circular dichroismCircular dichroism (CD) was used to characterize the secondary

structure of oligomers and compare their structure with monomersand non-nuclei oligomers. In these experiments, 1 mg of peptide wasdissolved in 20 μL of 0.05 M NaOH, and the solution was allowed tostand for 1 min at room temperature. 180 μL of water was then addedso that the final peptide concentration was 5 mg/mL in 5 mM NaOH.The peptide is unstable in the NaOH solution, and was used within anhour.Whenmeasurementsweremade at pH7.0, 20 μL of 100mMphos-phate buffer was added in addition to 160 μL of water, which made theprotein more stable in solution. The CDmeasurements were performedon a Jasco J-815 (Jasco Inc., Easton, MD) and each measurement wasdone at least in triplicate between the wavelengths of 190 to 260 nm.The path length was 0.2 mm, scan speed 20 nm/min, bandwidth 1 nmand response time 16 s. The final peptide concentration for most mea-surements was 0.5 mg/mL unless indicated otherwise. The buffer con-centration was 10 mM. The spectra of buffer or other additives weresubtracted when necessary.

3. Results

In order to evaluate the role of the orientation of the monomers inpeptide self-assembly, a dimeric design in which two Sup35 monomerswere connected in antiparallel orientation via the (d-Pro)G spacer wassynthesized and analyzed, as shown in Fig. 1. This construct was madein order to potentially have the completely antiparallel orientation ofthe monomers. The (d-Pro)G spacer was selected because accordingto refs. [16,26] it adopts a β-turn geometry, with the results of NMRand circular dichroism (CD) spectroscopies confirming theβ-turn struc-ture of this region [19]. Therefore, the oligopeptide with the sequenceCGNNQQNY (d-Pro)G YNNQQNNG is, in principle, capable of forminga β-hairpin structure with an antiparallel orientation of the monomerstermed AP-hairpin (where AP stays for antiparallel).

3.1. Aggregation kinetics

First, the aggregation kinetics were studied with the use of ThT fluo-rescence. The results of the aggregation analysis of the monomer andthe AP-hairpin at pH 5.6, and an ionic strength of 10 mM are shown inFig. 2. This analysis revealed that the aggregation kinetics at pH 5.6was slower for AP-hairpin (green line) compared with the monomer.The aggregation half-time values were 8.9 ± 2.6 h and 15.3 ± 1.0 hfor the monomer and the AP-hairpin, respectively. Non-normalizeddata are shown in Fig. S2A.

Since the pH used in this original set of experiments was close to thepeptide's isoelectric points, the next step was to examine the effects ofpH and ionic strength on the aggregation kinetics. The effect of pHwas tested on theAP-hairpin byperforming the studies at pH2.0 in a so-lution containing also potassium chloride and hydrochloric acid, withthe ionic strength remaining constant at 10 mM. Previously, it wasshown that pH 5.6 and pH 2.0 correspond to the conditions at whichthe monomers had the shortest and longest aggregation half-times,8.9 ± 2.6 h, and 85.3 ± 11.1 h respectively [15]. The experiments forthe AP-hairpin were repeated five times at each pH and the averaged




T293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347Q6

348

Fig. 2. Normalized time-dependent thioflavin (ThT) fluorescence plots. The aggregationkinetics of the monomer (black square), pH 5.6 and AP-hairpin (green). Aggregationhalf-times are indicated by the dashed lines that are color-coded tomatch their respectivepeptide, as they intercept with the x-axis.

t1:1Table 1t1:2The effect of pH and ionic strength on linear monomer and AP antiparallel hairpin dimert1:3half-times (hours) in ThT fluorescence experiments.

t1:4Conditions Half-time for the monomer (h) Half-time for AP dimer (h)

t1:5pH 5.6, 10 mM 8.9 ± 2.6 15.3 ± 1.0t1:6pH 2.0, 10 mM 85.3 ± 11.1 18.1 ± 0.5t1:7pH 2.0, 150 mM 40.8 ± 1.7 18.2 ± 0.6


RREC

results as kinetic plots are shown in Fig. S1. This analysis showed theaggregation half-time of the AP-hairpin at pH 5.6 of 15.3 ± 1.0 h(red circles and fit curve), and the half-time of 18.1 ± 0.5 h forpH 2.0 (green triangles and fit curve), suggesting that the pH hadonly a subtle effect on the aggregation half-time of the AP-hairpin.

The effect of ionic strength on the AP-hairpin aggregation kineticswas evaluated. The ionic strength was tested at pH 2.0 because in theprevious experimentswith themonomer [15], it showed that the aggre-gation half-time was greatly reduced when the ionic strength wasincreased from 10 mM to 150 mM. The experiments for the two ionicstrengths were performed in parallel and the corresponding ThTfluorescence results are shown in Fig. 3 and the non-normalized pointsin Fig. S3. According to these analyses, there is no effect of the ionicstrength on the aggregation kinetics of the AP-hairpin. The differencesin the plateaus values for themonomer andAP-dimers in Fig. S2 suggestthat there are differences in the peptide secondary structures in the ag-gregates formed by these two species. The half-times for the AP-hairpinat high and low ionic strengths were essentially identical (18.2 ± 0.6and 18.1 ± 0.5 h, respectively). Table 1 shows that this is in strong con-trast for the monomer aggregation data, where the half-time of the ag-gregation process of the monomeric peptide is 85.3 ± 11.1 h at lowionic strength and 40.8 ± 1.7 h at high ionic strength.

UNCO

349

350

351

352

Fig. 3. ThT fluorescence curves obtained for AP-hairpin at pH 2 and different ionic strengths.Green triangles are AP dimers in pH 2.0 10 mM, blue triangles are AP-hairpins in pH 2.0150 mM, and magenta diamonds are monomers in pH 2.0 10 mM.


ED P

RO

OF

3.2. Morphology of resulting aggregates assessed by AFM

To get insight into the morphology of aggregates formed by mono-meric and AP-hairpin species, AFM imaging was used to analyze the ag-gregate samples collected at the plateau values of the ThT kinetic curves.

The set of images in Fig. 4A shows the morphologies of aggregatesassembled from the monomers that were incubated in water. Likewise,Fig. 4B shows the morphologies of aggregates assembled from theAP-hairpin in water. The pH of water solutions was approximatelypH 2. Two different morphologies, globular aggregates and fibrils areusually identified in the AFM images of amyloid aggregates, and thesewere present in these samples as well. The results show that in water,therewere nomajor differences inmorphology between the aggregatesformed by monomers and AP-hairpins as fibrillar aggregates tended toform most readily under water for both peptides. This finding is in linewith the ThT aggregation experiments shown in Fig. S3B. Note thatthese are not normalized data demonstrating that both types of fibrilshave similar affinity for ThT binding and pointing to a similarity ofstructures of fibrils.

The morphologies of aggregates formed at pH 2.0 (10 mM NaCl)were also analyzed by AFM. The images of aggregates formed frommonomers are shown in Fig. 5A. These are primarily fibrils. However,the aggregates formed at pH 2.0 10 mM by the AP-hairpin were mainlyglobular species, as shown in Fig. 5B. This plot shows that there are onlya few fibrils within the field. Thus AP peptide with the hairpin designdoes not form fibrils at acidic pH, although it aggregates more thanfive times faster than the monomer samples (Fig. 3). Changing theconditions to pH 5.6 dramatically drops the yield of fibrillar aggregatesfor both samples. The set of images in Fig. S4A shows the morphologiesof aggregates assembled by the monomers at pH 5.6, and the set ofimages in Fig. S4B shows the morphologies of aggregates assembledfrom the AP-hairpin at pH 5.6. Globular aggregates were by far thelargest group of aggregates seen on both images. These small, globularoligomers can be clearly seen in both of the AFM images in Fig. S4 asthe brighter spherical features. There were also some slightly elongatedfeatures that are seen arranged in a vertical line at themiddle of Fig. S4A,as well as some slightly elongated features just below the center ofFig. S4B. These features can form in a relatively short time whencompared to the formation of longer fibrillar assemblies.

Fig. 4. AFM images of aggregates formed in water. Aggregates are formed by the mono-mers (A) and AP-hairpin (B). The scale bars represent 1 μm.




353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

Fig. 5.AFM images of themonomer andAPdimers.Monomers (A) andAPdimers (B)wereimaged in air after incubation in pH 2.0 10 mM NaCl. The scale bars represent 1 μm.


3.3. Force spectroscopy analysis

In order to characterize the interaction between the peptides andwhat role the potential dimerization has with respect to the interpeptide

UNCO

RRECT

Fig. 6. Force spectroscopy of the AP-hairpins. Representative force–distance curve (A)measured aloading rate. The overlay of force distance curves (B) is shownhere. (C) Thehistogramsof the rupt14,477 pN/s, 28,739 pN/s, and 50,300 pN/s. The most probable rupture force (F) for each loadimaxima of 32 pN, 32 pN, 37 pN, 56 pN, 65 pN, 95 pN, and 134 pN from top histogram to botto


OO

F

interactions, AFM force spectroscopywas applied. Since analogous exper-iments for the monomer species were performed earlier [15] at pH 5.6 in11mMacetate buffer,we applied the same approach to theAP species. In-dividual AP molecules were covalently tethered to the AFM tip and themica surface via the flexible PEG linker and their interaction was probedupon the approach–retraction cycles of the AFM tip to the surface. The ex-periments were performed at pH 5.6 and a representative force curve isshown in Fig. 6A. The initial smooth part of the curve corresponds to theextension of the tethers that is extrapolated by the worm-like chain(WLC)model as shownby a solid curve. This entropic process is accompa-nied by the complex dissociation with a rupture force value (32 pN)being the major variable of the force probing experiment. The overlay of428 force curves is shown in Fig. 6B and the distribution of the ruptureforces over 428 rupture events is shown as an inset in Fig. 6A.

Next, we performed pulling experiments at a range of pulling ratesbetween 100 and 2000 nm/s to characterize the complex dissociationprocess using dynamic force spectroscopy (DFS) methodology as de-scribed before [19–22]. The rupture forces increase with the pullingrates. This is demonstrated by Fig. 6C inwhich the force distribution his-tograms obtained at the different loading rates are presented. These dis-tributions were approximated with probability function fit [20–22] and

ED P

R

t pH 5.6with apparent loading rate (ALR) 1941 pN/s, with the inset of forces obtained at thisure force distributionmeasured at pH5.6 at ALR467 pN/s, 1941 pN/s, 4449 pN/s, 7786 pN/s,ng rate was calculated by probability function fitting, resulting in forces at the histogramsm histogram.




T

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

t2:1

t2:2

t2:3

t2:4

t2:5

t2:6

t2:7

t2:8


UNCO

RREC

the values corresponding to themaximawere used to generate the DFSspectra in which the rupture force values are plotted against logarithmof apparent pulling rate values. The major peaks on the force distribu-tions were used in this plot. The second peak on the last distributionwith the force value twice of that for the first peak corresponds to thedouble rupture events appearing in these experiments at the highpulling rates. The data points on DFS plots are fit by two linear lines sug-gesting that the AP hairpin dissociation pathway has two barriers [27].The DFS plots enable us to extract the barrier parameters. From theslope the positions of barriers (xβ) is obtained and the intercepts obtain-ed after the extrapolation to the zero pulling rate produce the off-rateconstant (koff). These values corresponding to both transitions areshown in Table 2.

Fig. 7 summarizes the dynamic force spectroscopy results for theAP-dimer. Fig. 7A shows the plot of Fr obtained for different pullingvelocities versus the apparent loading rates on a logarithmic scale. TheΔG values and maxima locations in Table 2 obtained from the fittingthe plot in Fig. 7A with the Bell–Evans model (see Material andmethods section)were used to reconstruct the energy landscape profilefor the dissociation of the AP dimeric complex shown in Fig. 7B. The first(inner barrier) and second (outer barrier) transient states for thisprocess are located at 0.06 nm and 0.37 nm from the ground state ofthe complex. The barrier heights (ΔG values) for inner barrier(ΔG = 26 ± 0.54 kBT) and outer barrier (ΔG = 28.4 ± 0.61 kBT) wereobtained from the off-rate constant (koff) values (Table 2).

3.4. Secondary structure analysis of the monomer and dimer assessed byCD measurements

We applied circular dichroism (CD) spectroscopy [5,12] to charac-terize the secondary structure of the peptides under a variety of differ-ent conditions, such as various pH values and the presence of differentcosolvents and osmolytes. First, the data shown in Fig. 8 (concentration0.15 mg/mL) demonstrate that structurally both peptides are similarand a deepminimum at ~190 nm indicates a primarily unfolded confor-mation of both peptides. The effect of concentrationwas studied. To thisend, the CD spectrawere recorded at pH 7 at various peptide concentra-tions. This analysis revealed no significant concentration effect on themonomer CD spectra (Fig. 8A). With the increase in the dimer concen-tration the peak intensity decreased with a shift from 195 to 199 nm(Fig. 8B), however, the peak shape remains largely unchanged as seenfrom the comparison of normalized CD spectra in Fig. 8C. These resultsare consistent with formation of oligomeric aggregates by AP dimer athigher concentrations (0.5 mg/mL and above). Formation of oligomersby amyloidogenic proteins and peptides is a common process, andthey vary in their secondary and tertiary structures and stability. It ap-pears that oligomers detected in Fig. 8B are likely to be quite unstablesince all solutions for CD measurements were prepared from the samestock solution (seeMaterials andMethods section) andwhatever oligo-mers were present there must have dissociated upon dilution to lowerconcentration (e.g. 0.15 mg/mL). Investigation of these intermediatesin more detail is beyond the scope of the current manuscript.

Next, the effect of pH on the peptide structures was studied at thepeptide concentration of 0.5 mg/mL. The data in Fig. 9A show no struc-tural changes for the monomer in the pH range of 5–8, although

485

486

487

488

489

490

491

492

493

Table 2Summary of parameters obtained from dynamic force spectroscopy of the AP-hairpin atpH 5.6.

System koff (1) (s−1)a xβ(1) (Å)a koff (2) (s−1)a xβ(2) (Å)a

pH 5.6 AP-hairpin 1.3 ± 3.7 5.8 ± 3.5 59.2 ± 15.1 0.98 ± 0.19

xβ, the position of energy barrier. Indices 1 and 2 in the subscript of the parameters refer tothe outer and inner barriers of the energy landscape, respectively.

a koff, dissociation rate constant for the peptide–peptide pair.


ED P

RO

OF

structural changes do occur at acidic pH down to pH 2.0. The dimerdemonstrates stronger resistance to the pH changes, as seen in Fig. 9B.Even at acidic pH, its CD spectra show very subtle changes. Bothpeptides generally maintain a random coil structure from pH 2.0through pH 8.0.

Additionally, the effect of trimethylamine-N-oxide (TMAO) on thesecondary structure of these peptides was assessed. TMAO is a naturallyoccurring osmolyte that stabilizes proteins against denaturation. It wasshown that osmolytes increase protein stability and that the two-statefolding behavior of proteins is maintained in the presence of these com-pounds [28]. Fig. S5 shows the results obtained for both peptides in abroad range of TMAO concentrations. The increase in the TMAO concen-tration noticeably changed the shape of CD spectra of the monomer(Fig. S5A) indicating that this osmolyte induces structural changes ofthe peptide. Comparatively, no such spectral changes were detectedfor the dimer (Fig. S5B) suggesting that the dimer retains the initial un-folded conformation regardless of the TMAO presence.

In addition to TMAO, the effect of 1,1,1,3,3,3-hexafluoro-propan-2-ol(HFIP) was tested. HFIP is one of the strongest helix-inducing and stabi-lizing co-solvents [29] that is widely used for structural studies ofamyloidogenic proteins. The results shown in Fig. S6A demonstrate achange of the conformation of the monomeric peptide upon theincrease in the HFIP concentration as evidenced by a dramatic changein the shape of far-UV CD spectra. This change corresponds to thetransition from a mainly disordered structure at 0% and 10% HFIP to aprimarily α-helical structure at 50% HFIP. On the other hand, Fig. S6 Bshows that although there are some changes in the CD spectra of theAP-hairpin, these changes are not as dramatic as those detected forthe monomeric peptide.

The CD measurements were performed at a relatively high concen-tration of the peptides raising the concern ofwhether a spontaneous ag-gregation of the peptides occurred during the sample preparation forthe CD measurements. To test this assumption, we performed the CDmeasurements of the aggregated AP-hairpin sample. The peptide wasincubated at pH 5.6 for 24 h corresponding to the plateau level inFig. 2. The sample was imaged with AFM (Fig. S7) and these dataconfirm the globular morphology of the aggregates. The morphologyof the sample did not change over additional two-day incubation atroom temperature or multiple freezing–thawing cycles of the sample towarrant the sample stability during the sample preparation for the CDstudy. The CD spectra of AP aggregates are shown in Fig. 10 (solid line).The CD spectrum of the AP-hairpin sample is shown as a dashed linefor comparison. They show that the two spectra are distinctly different.The AP aggregate spectrum has a sharp minimum at 193 nm and a shal-lowmaximumat 217nmand it is different from that of AP dimer andhadsome resemblance to a typical spectrum of polyproline II helix [30]. Thus,the changes in the CD spectra of the dimer at higher peptide concentra-tions are not due to formation of stable aggregates.

Altogether, the CD analysis leads to the conclusion that theAP-dimeris characterized by very low structural plasticity compared to that of themonomer. While AP-hairpin appears to form transient oligomers athigher protein concentrations, their instability makes it unlikely thatthey account for apparent stability of the conformation of this peptidein a variety of conditions. This increased stability of AP-hairpin maywell interfere with its aggregation to amyloid fibrils. A similar effect ofthe restricted conformational mobility was observed in [31] where adouble cysteine mutant of Aβ(21C/A30C) was unable to form fibrillaraggregates.

4. Discussion

The comparative studies of the linear monomeric and hairpindesigns of Sup35 heptapeptide revealed a number of novel propertiesof these peptides and their interactions that could be related to differ-ences in their aggregation kinetics. We discuss these properties in thesections below.




OF

494

495

496

497

498

499

500

501

502

503

504

505

506

Fig. 7.Dynamic force spectroscopy plot and the energy profile. Dynamic force spectroscopy of AP dimer (A) and the energy profile obtained from the parameters shown in Table 2 (B). TheΔG values for outer barrier and inner barrier are 25.4 kBT and 29.2 kBT, respectively.


4.1. The Sup35 peptide aggregation kinetics probed by ThT fluorescence

Although peptideswith both designs follow typical sigmoidal aggre-gation kinetics, they quantitatively differ from each other. TheAP-hairpin design in which two monomers are connected with ashort-(d-Pro)-G linker forcing the antiparallel arrangement of themonomers change the aggregation propensity of the Sup35 peptide.

UNCO

RRECT

Fig. 8. Concentration dependence of CD spectra of the peptides at pH 7. The concentration depindicate lack of significant structural changes with increased concentration.


PRO

Althoughmonomeric Sup35 peptide appears to aggregate slightly fasterthan the AP-hairpin at pH 5.6 (Fig. 2), the situation is opposite at acidicpH (pH 2.0). According to Fig. 3, at these conditions, the AP-dimers ag-gregate five times faster than the monomeric form does, and the corre-sponding aggregation profile does not depend on the ionic strength ofthe solution. The similarity in the aggregation kinetics at pH 5.6 forboth Sup35 designs is in line with AFM topographic studies (Fig. S4).

ED

endence of monomer (A) and dimer (B) is shown. (C) Normalized spectra of AP dimer to




TED P

RO

OF

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549Q7

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

Fig. 9. pH dependence for monomer and AP-hairpin. Circular dichroism spectra of the monomer pHs in column (A) and AP-hairpin in column (B).


UNCO

RREC

The comparison of all results suggests that kinetics of the aggregationand their morphology depends on ionic strength and pH. The lack of fi-brils formation by both peptides at pH values close to the peptides pI(pH 5.6) indicates to the contribution of electrostatic interactions tothe interpeptide interaction that we identified in our recent publication[15]. Note in this regard studies with Aβ(21C/A30C) peptide in whichthe internal hairpinwas stabilized by disulfide bonds formed by Cys res-idues incorporated into positions 21 and 30 [31]. Such peptides formoligomeric samples, but are not capable of forming amyloid fibrils.

4.2. The structure and morphology of aggregates formed by theSup35 samples

The antiparallel orientation of Sup35 monomers in dimers issupported by Replica exchangeMD simulation described in [32]. The re-sults show that the monomer (CGNNQQNY) adopts the random coilstructure. The d-Pro-G adopts a U-turn geometry keeping both mono-meric units in a close proximity. Similarly, the AP-hairpin also demon-strates rather similar pattern with no clear stable secondary structure.The AP structures with antiparallel β-strands are very rare which is inline with the CD measurements.

In the molecular dynamic (MD) simulation we modeled the assem-bly ofmonomers into dimerswith theβ-sheet conformation and the re-sults are shown in Fig. S9. Replica exchange MD landscape reveals twodeepest minima corresponding to two structures with the antiparalleltopology differing in the length of antiparallel β-strands. Model 2 corre-sponds to the in-register orientation of the monomers with all aminoacids involved in the hydrogen bonding, whereas in Model 1, only fiveresidues out of seven are involved in the antiparallel β-sheet structureformation, termed out-of-register conformation. Energetically bothconformations are very close, but our recent computational analysis ofrupture events with the use of Monte Carlo pulling approach showedthatModel 1 corresponds to themajority of inter-monomer interactions


observed in experiment [32]. In the AP-hairpin, d-Pro-G turn imposessteric limitations for the formation of the out-of-register conformation,which explains the lack of adoption of antiparallel β-strands geometryby the AP-hairpin construct.

The antiparallel arrangements of Sup35 monomers are at odds withcrystallographic data [3]. That studies reveal parallel in-register ar-rangement of monomers in β-sheet conformation, where it was pairedup with another sheet, and the strands in one sheet are antiparallel tothe other sheet [33]. However, we need to take into consideration anumber of factors. First, the steric zipper structure consists of dry inter-faces that have nowater [3,13]. The peptides are fully hydrated in aggre-gation experiments andMD simulations involved necessary amounts ofwater molecules. Second, the crystallographic date corresponds to thesituation in which all monomers are tightly packed, so interactionswith all neighbors contribute the energy minimum of the entire multi-monomer system. So isolated dimers should have different environ-ment that should change the peptide structure and arrangement. Thisconsideration is supported by a recent computational analysis of thehexamer of Sup35 peptide [14]. The initial crystallographic structureof the hexamer was unstable and dissociated over time. Therefore, wehypothesize that Sup35monomers are arranged in antiparallel orienta-tion at early stages of the aggregation process, but the formation of largeaggregates can lead to conformational transition and rearrangements ofmonomers. The water displacement can facilitate such structuraltransition.

The antiparallel arrangement of monomers within dimers has aprofound effect onmorphologies of aggregates. Although inwater fibril-lar aggregates are essentially the only morphology for both species(Fig. 4), the change of conditions reveals the difference between theaggregation propensities of the AP-hairpins and monomers. The mostdramatic difference was observed at acidic pH (Fig. 5). If monomerscontinue forming fibrils with rather straight morphologies, theAP-hairpins assemble into globular aggregates with fibrils being found




T571

572

573

574

575

576

577

578

579

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

600

601

602

603

604

605

606

607

608

609

610

611

612

613

614

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

636

637

638

639

640

641

642

643

644

645

646

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

679

680

Fig. 10. CD spectrum of globular aggregates of AP-hairpin (solid line). Dashed line showsthe spectrum of the monomeric AP-hairpin sample. A circular dichroism spectrum ofglobular AP aggregates suggests a polyproline II-like structure.


UNCO

RREC

very rarely. Similar aggregates are formed by the monomer at pH 5.6alongwith fibrils that are ratherminor species compared to globular ag-gregates. Fibrils appear in the AP samples, but still globular aggregatesremain very predominant species. Some insight into the secondarystructure of these aggregates can be obtained from their CD spectra(Fig. 10) that resemble those of polyproline II helix. While these spectraare not sufficient to tell us about the secondary structure of theseoligomers, it is clear that they are structurally distinct from both non-aggregated AP-hairpin and amyloid fibrils. These data suggest thatantiparallel arrangement of monomers within the dimers facilitatestheir assembly into globular aggregates. It is likely that unfolding ofthe hairpin structure is needed for the fibrils assembly, and purewater provides such conditions.

The structural stability of AP-dimers is supported by far-UV CDmeasurements (Fig. 9), according to which the AP secondary structurechanges very little under the various conditions, suggesting that theAP-hairpin is in a locked configuration. Structural plasticity was also acritical factor in distinguishing different modes of Aβ aggregation. Itwas demonstrated in ref. [34] that phosphorylation of serine 26 impairsfibrillation while stabilizing the nonfibrillar aggregates because thephosphate group at that position diminishes the propensity of a peptideto form a β-hairpin that is necessary for fibril formation. The conforma-tional rearrangement that is needed for fibril formation is not present,much like what was observed with our hairpin dimer [34]. Althoughthe predominant formation of oligomeric aggregates was reported inref. [35], these aggregates did not induce changes in the ThT fluores-cence suggesting that they do not form β-sheet structure. TheAP-hairpin oligomers assembled at pH 2 do induce increase in the ThTfluorescence intensity. Furthermore, the level of ThT fluorescence forAP-hairpin is higher compared to the levels detected for themonomericdesign (Fig. S2) suggesting that the secondary structure existing inoligomeric peptides is favorable for ThT binding and enhancedfluorescence.

Previous studies confirmed that the initial dimer formation favorsthe antiparallel orientation 60% of the time, with the remainder of thearrangements being random in nature [36]. However, beyond thetrimer there is another shift of interactions, and the majority of theinteractions tend to shift from antiparallel to about 70% of the orienta-tions having a majority of a mix between antiparallel and parallelorientations within the β-sheet [36]. The simulations show that theantiparallel arrangement is essential to the initial stages of aggregateformation, stabilizing the initial aggregated forms [37]. The X-ray datashow mature fibrils having a parallel arrangement betweenGNNQQNY sequences. However, when the antiparallel configuration islockedwith a short loop (as in the case of our study), it can then impedethe creation of fibrillar aggregates.


ED P

RO

OF

Another possibility is that the interaction between the β-strandswithin the AP-hairpin peptide may be too strong because of thepresence of the short loop at the center of the peptide. The short loopcould also be very stiff, which can limit possibility to adopt differentstructures, with the notable exception of water. According to the CDspectroscopy data (Fig. 9), there was some variations in the secondarystructure of the linear monomer at low pH, which corresponded tochanges in the half-times in the aggregation experiments. On theother hand, there was no variation in the secondary structure acrossall pHs tested on the antiparallel hairpin dimer, and this correspondedto little change in half-times in the aggregation experiments across allpHs tested. The CD data appear to correlate with the ThT data, as thelinear monomer had changes in kinetics and subtle changes in thesecondary structure with a change in pH. The AP-hairpin had subtlechanges with respect to the aggregation kinetics, and then when testedfor changes in the secondary structure in solutions with different pHvalues, there was no discernable change in the secondary structure.TheAP-hairpinhad a structure thatwasmore stable thanwith the linearmonomer. However, the structure appears to be too stable and rigid thatmost changes in the external environment are possibly preventing anycrucial late stage conformational transition that may be required for ag-gregation. This was hypothesized with the β-lactoglobulin protein [38]where a structural interruption to the hydrophobic core of the proteinwas necessary before the final step leading to aggregation. There werethree phases of aggregation,with the earlier phases consisting ofweaklyassociated aggregates [38].

The tendency of the hairpins to have a different morphology fromthe linear monomer is explained by a different tertiary structure of thehairpinwhich can influence how the fibril self-assembles, and ultimate-ly the morphology of the resulting aggregates. This hypothesis agreeswith the model that has been proposed based on the solid-state NMRstudies of yeast prion proteins, RNQ1 and URE3 [10,39,40]. The se-quence studiedwas scrambled to demonstrate that the secondary struc-ture itself, and not the sequence, is responsible for the templating of themonomer being added to the end of the forming fibril. Along with an-other study involving RNQ1 [41] it was demonstrated that even addinga prion domain of a different protein can result in templating, as wasdemonstrated when adding the prion domain of Sup35 to the end ofRNQ1 or URE3.

4.3. The hairpin assembly of Sup35 and the effect of electrostatics on theaggregation kinetics

X-ray powder diffraction of microcrystal data of GNNQQNY showedin-register and parallel arrangement of the monomers within thecrystals [3]. This parallel in-register arrangement forms a β-sheet,where it was paired up with another sheet, and the strands in onesheet are antiparallel to the other sheet [33]. These experiments wereperformed at acidic pH conditions, with positively charged N terminusand neutral C terminus. In the absence of electrostatic interactionfrom N–C terminal residues and in the presence of the interactionsfrom the aromatic ring of tyrosine residues [42], two monomers atsuch conditions prefer the parallel arrangement of β-strands. Incontrast, at neutral pH, bothN- andC-terminal residues contain charges,the peptides prefer the antiparallel structure due to electrostaticinteraction [43]. On one hand, in comparisonwith the parallel structure,there are more hydrogen bonds within two adjacent monomers. Theantiparallel structure is much more stable than parallel structure. Onthe other hand, the charged C-terminus will disturb the interaction oftyrosine residues at C-terminus and further prevent formation of paral-lel configuration. In the AP-hairpin construct, the C-terminal end of onemonomer is connected with the N-terminus of the second monomer.Therefore, there is no electrostatic contribution to their antiparallel ar-rangementwithin the dimer. As a result, the effect of the ionic strengthsand pH on the aggregation of the AP-construct should be small com-pared to the monomer and this is in line with our experimental data.




T

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

697

698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

721

722

723

724

725

726

727

728

729

730

731

732

733

734

735

736

737

738

739

740

741

742

743Q8

744

745

746

747

748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804Q9805806807808809810811812813814815816817


UNCO

RREC

Overall, the experiments with linear and antiparallel β-hairpindesigns of Sup35 peptide led to a number of novel findings related tothe peculiarities of peptide interactions and the process of its aggrega-tion. One of the unexpected findings was a conformational stiffness ofthe hairpin design of the peptide. Although the AP-hairpin was able toform amyloid aggregates detectable by ThTfluorescence andAFM imag-ing, the range of aggregate morphologies was limited. Specifically,AP-hairpin fibrils were a very rare species compared to the aggregatedform of the linear peptide for which the fibrils were the major species.Thisfinding suggests amodel, according towhich oligomerswith differ-ent conformations are formed at early stages and not all of them lead tothefibrils. Thismodel suggests that there is a set ofmisfolded conforma-tions for the peptide and freezing the conformational mobility of themonomer limits the range of aggregated morphologies. An alternativemodel inwhich the conformational switch can occur at the level of olig-omers should be considered as well. This conformational switch withinoligomers can also be dependent on the monomer conformationalmobility, although intermolecular interactions should play the majorrole in this model. The models can be tested by making other designssuch as hairpins with different loop sizes, mutlifolded peptides, andthese ideas lay a foundation for the future studies with Sup35 peptide.The approaches described in this paper can be extended to longerprion-derived peptides including the full-length protein and suchstudies are our long-term goals.

5. Conclusions

The aggregation kinetics of the AP-hairpin has low variability,regardless of pH or ionic strength. The secondary structure forAP-hairpin remained stable when compared to the linear monomerunder the same changes in the environment, as was tested with CDspectroscopy. The structure of AP-hairpin was not altered even in thepresence of osmolyte HFIP, a known stabilizer of α-helices. The mono-mer was also more noticeably affected than AP-hairpin by changes inconditions such as lowpH, high ionic strength, or osmolytes at high con-centrations. The AFMdata showsmore globular aggregates and a lack offibrils for AP-hairpin than for the linear monomer under comparableconditions. The AP-hairpin had a lower tendency to form fibrils thanthe linear monomer at pH 2.0 at low ionic strength, where the mostfibrils formed by the linear monomer. The AP dimer also was able toform fibrils when incubated in pure water, which was the only condi-tion in which the AP-hairpin could form fibrils. The implication of imag-ing data and the spectroscopic data, is that the antiparallel dimer has anoticeably different behaviors under different conditions. This meansthat the AP-hairpin structure did use an alternate pathway formingaggregates mainly with a globular morphology.

Author contributions

Yuri L. Lyubchenko and Alexander Portillo designed the experi-ments. Alexander Portillo performed the ThT fluorescence, AFM imag-ing, and AFM force spectroscopy experiments, as well as the dataanalysis resulting from these experiments. Mohtadin Hashemi madethe AP-oligomer samples for CD measurements, performed their AFMcharacterization and analyzed DFS data. Yuliang Zhang performed theMD and MC simulations. Leonid Breydo and Vladimir Uverskyperformed the CD experiments and data analyses and contributed tothe CD related discussion of the paper. Yuri L. Lyubchenko, AlexanderPortillo and Vladimir Uversky wrote the paper.

Acknowledgments

The authors thankA. Krasnoslobodtsev, Z. Lv, L. Shlyakhtenko aswellas other Lyubchenko lab members for insightful discussions. Theauthors would also like to thank the Uversky lab members as well fortheir insightful discussions.


The work was supported from grants by EPS-1004094 (NSF) and5R01GM096039-04 (NIH) to YLL, aswell as the UNMC graduate studentresearch fellowship to AP.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbapap.2014.12.009.

ED P

RO

OF

References

[1] M. Balbirnie, R. Grothe, D.S. Eisenberg, An amyloid-forming peptide from the yeastprion Sup35 reveals a dehydrated beta-sheet structure for amyloid, Proc. Natl.Acad. Sci. U. S. A. 98 (2001) 2375–2380.

[2] Eisenberg, An amyloid-forming peptide from the yeast prion Sup35 reveals adehydrated b-sheet structure for amyloid, PNAS 98 (2001) 2375–2380.

[3] R. Nelson, M.R. Sawaya, M. Balbirnie, A.O. Madsen, C. Riekel, R. Grothe, D. Eisenberg,Structure of the cross-beta spine of amyloid-like fibrils, Nature 435 (2005) 773–778.

[4] R. Wetzel, S. Shivaprasad, A.D.Williams, Plasticity of amyloid fibrils, Biochemistry 46(2007) 1–10.

[5] L.O. Tjernberg, A. Tjernberg, N. Bark, Y. Shi, B.P. Ruzsicska, Z. Bu, J. Thyberg, D.J.Callaway, Assembling amyloid fibrils from designed structures containing a signifi-cant amyloid beta-peptide fragment, Biochem. J. 366 (2002) 343–351.

[6] J.R. Glover, A.S. Kowal, E.C. Schirmer, M.M. Patino, J.J. Liu, S. Lindquist, Self-seeded fi-bers formed by Sup35, the protein determinant of [PSI+], a heritable prion-like fac-tor of S. cerevisiae, Cell 89 (1997) 811–819.

[7] S. Narayanan, S. Walter, B. Reif, Yeast prion–protein, sup35, fibril formation pro-ceeds by addition and substraction of oligomers, Chembiochem 7 (2006) 757–765.

[8] S. Hess, S.L. Lindquist, T. Scheibel, Alternative assembly pathways of theamyloidogenic yeast prion determinant Sup35-NM, EMBO Rep. 8 (2007)1196–1201.

[9] J. Krzewska, M. Tanaka, S.G. Burston, R. Melki, Biochemical and functional analysis ofthe assembly of full-length Sup35p and its prion-forming domain, J. Biol. Chem. 282(2007) 1679–1686.

[10] F. Shewmaker, E.D. Ross, R. Tycko, R.B. Wickner, Amyloids of shuffled prion domainsthat form prions have a parallel in-register beta-sheet structure, Biochemistry 47(2008) 4000–4007.

[11] P.C. van der Wel, J.R. Lewandowski, R.G. Griffin, Solid-state NMR study of amyloidnanocrystals and fibrils formed by the peptide GNNQQNY from yeast prion proteinSup35p, J. Am. Chem. Soc. 129 (2007) 5117–5130.

[12] J. Zheng, B. Ma, C.J. Tsai, R. Nussinov, Structural stability and dynamics of anamyloid-forming peptide GNNQQNY from the yeast prion sup-35, Biophys. J. 91(2006) 824–833.

[13] M.R. Sawaya, S. Sambashivan, R. Nelson, M.I. Ivanova, S.A. Sievers, M.I. Apostol, M.J.Thompson, M. Balbirnie, J.J. Wiltzius, H.T. McFarlane, A.O. Madsen, C. Riekel, D.Eisenberg, Atomic structures of amyloid cross-beta spines reveal varied steric zip-pers, Nature 447 (2007) 453–457.

[14] A. Srivastava, P.V. Balaji, Size, orientation and organization of oligomers that nucle-ate amyloid fibrils: clues from MD simulations of pre-formed aggregates, Biochim.Biophys. Acta 1824 (2012) 963–973.

[15] A.M. Portillo, A.V. Krasnoslobodtsev, Y.L. Lyubchenko, Effect of electrostatics on ag-gregation of prion protein Sup35 peptide, J. Phys. Condens. Matter 24 (2012)164205.

[16] H.E. Stanger, S.H. Gellmam, Rules for antiparallel β-sheet design: D-Pro-Gly is supe-rior to L-Asn-Gly for β-hairpin nucleation, J. Am. Chem. Soc. 120 (1998) 4236–4237.

[17] L.S. Shlyakhtenko, A.A. Gall, A. Filonov, Z. Cerovac, A. Lushnikov, Y.L. Lyubchenko,Silatrane-based surface chemistry for immobilization of DNA, protein-DNA com-plexes and other biological materials, Ultramicroscopy 97 (2003) 279–287.

[18] K. Abid, C. Soto, The intriguing prion disorders, Cell. Mol. Life Sci. 63 (2006)2342–2351.

[19] R. Atarashi, V.L. Sim, N. Nishida, B. Caughey, S. Katamine, Prion strain-dependent dif-ferences in conversion of mutant prion proteins in cell culture, J. Virol. 80 (2006)7854–7862.

[20] B.H. Kim, N.Y. Palermo, S. Lovas, T. Zaikova, J.F. Keana, Y.L. Lyubchenko, Single-molecule atomic force microscopy force spectroscopy study of Abeta-40 interac-tions, Biochemistry 50 (2011) 5154–5162.

[21] J. Yu, Y.L. Lyubchenko, Early stages for Parkinson's development: α-synucleinmisfolding and aggregation, J. Neuroimmune Pharmacol. (2008).

[22] J. Yu, S. Malkova, Y.L. Lyubchenko, Alpha-synucleinmisfolding: single molecule AFMforce spectroscopy study, J. Mol. Biol. 384 (2008) 992–1001.

[23] C. Ray, J.R. Brown, B.B. Akhremitchev, Correction of systematic errors in single-molecule force spectroscopy with polymeric tethers by atomic force microscopy, J.Phys. Chem. B 111 (2007) 1963–1974.

[24] S. Jain, J.B. Udgaonkar, Defining the pathway of worm-like amyloid fibril formationby the mouse prion protein by delineation of the productive and unproductive olig-omerization reactions, Biochemistry 50 (2011) 1153–1161.

[25] E. Evans, K. Ritchie, Dynamic strength of molecular adhesion bonds, Biophys. J. 72(1997) 1541–1555.

[26] I.L. Karle, H.N. Gopi, P. Balaram, Peptide hybrids containing alpha- and beta-aminoacids: structure of a decapeptide beta-hairpin with two facing beta-phenylalanineresidues, Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 3716–3719.






818819820821822823824825826827828829Q10830Q11831832833834835836837838839840

841842843844845846847848849850851852853854855856857858859860861862863

864


[27] L. Nielsen, R. Khurana, A. Coats, S. Frokjaer, J. Brange, S. Vyas, V.N. Uversky, A.L. Fink,Effect of environmental factors on the kinetics of insulin fibril formation: elucidationof the molecular mechanism, Biochemistry 40 (2001) 6036–6046.

[28] C.C. Mello, D. Barrick, Measuring the stability of partly folded proteins using TMAO,Protein Sci. 12 (2003) 1522–1529.

[29] N. Hirota, K. Mizuno, Y. Goto, Cooperative alpha-helix formation of beta-lactoglobulinand melittin induced by hexafluoroisopropanol, Protein Sci. 6 (1997) 416–421.

[30] A.L. Rucker, T.P. Creamer, Polyproline II helical structure in protein unfolded states:lysine peptides revisited, Protein Sci. 11 (2002) 980–985.

[31] T. Hard, Protein engineering to stabilize soluble amyloid beta-protein aggregates forstructural and functional studies, FEBS J. 278 (2011) 3884–3892.

[32] Y.L. Zhang, The structure of misfolded amyloidogenic dimers. computational analy-sis of force spectroscopy data, Biophys. J. (2014) (in press).

[33] T.L. Benzinger, D.M. Gregory, T.S. Burkoth, H. Miller-Auer, D.G. Lynn, R.E. Botto, S.C.Meredith, Propagating structure of Alzheimer's beta-amyloid(10–35) is parallelbeta-sheet with residues in exact register, Proc. Natl. Acad. Sci. U. S. A. 95 (1998)13407–13412.

[34] N. Rezaei-Ghaleh, M. Amininasab, K. Giller, S. Kumar, A. Stundl, A. Schneider, S.Becker, J. Walter, M. Zweckstetter, Turn plasticity distinguishes different modes ofamyloid-beta aggregation, J. Am. Chem. Soc. 136 (2014) 4913–4919.

[35] C. Soto, E.M. Castano, The conformation of Alzheimer's beta peptide determines therate of amyloid formation and its resistance to proteolysis, Biochem. J. 314 (Pt 2)(1996) 701–707.

UNCO

RRECT


F

[36] J. Nasica-Labouze, M. Meli, P. Derreumaux, G. Colombo, N. Mousseau, A multiscaleapproach to characterize the early aggregation steps of the amyloid-forming pep-tide GNNQQNY from the yeast prion sup-35, PLoS Comput. Biol. 7 (2011) e1002051.

[37] L. Vitagliano, A. Ruggiero, C. Pedone, R. Berisio, Conformational states and associa-tion mechanism of Yersinia pestis Caf1 subunits, Biochem. Biophys. Res. Commun.372 (2008) 804–810.

[38] J.T. Giurleo, X. He, D.S. Talaga, Beta-lactoglobulin assembles into amyloid through se-quential aggregated intermediates, J. Mol. Biol. 381 (2008) 1332–1348.

[39] F. Shewmaker, R.B. Wickner, R. Tycko, Amyloid of the prion domain of Sup35p hasan in-register parallel beta-sheet structure, Proc. Natl. Acad. Sci. U. S. A. 103(2006) 19754–19759.

[40] R.B. Wickner, F. Shewmaker, D. Kryndushkin, H.K. Edskes, Protein inheritance(prions) based on parallel in-register beta-sheet amyloid structures, Bioessays 30(2008) 955–964.

[41] Y.A. Vitrenko, E.O. Gracheva, J.E. Richmond, S.W. Liebman, Visualization of aggrega-tion of the Rnq1 prion domain and cross-seeding interactions with Sup35NM, J. Biol.Chem. 282 (2007) 1779–1787.

[42] J. Gsponer, U. Haberthur, A. Caflisch, The role of side-chain interactions in the earlysteps of aggregation: molecular dynamics simulations of an amyloid-forming pep-tide from the yeast prion Sup35, Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 5154–5159.

[43] L. Vitagliano, L. Esposito, C. Pedone, A. De Simone, Stability of single sheet GNNQQNYaggregates analyzed by replica exchange molecular dynamics: antiparallel versusparallel association, Biochem. Biophys. Res. Commun. 377 (2008) 1036–1041.
O
ED P

RO




Download - Role of monomer arrangement in the amyloid self-assembly

Top Related