Biochimica et Biophysica Acta 1844 (2014) 670–680

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

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

Curcumin and kaempferol prevent lysozyme fibril formation bymodulating aggregation kinetic parameters

Mohanish S. Borana a, Pushpa Mishra b, Raghuvir R.S. Pissurlenkar c,Ramakrishna V. Hosur a,b,⁎, Basir Ahmad a,⁎⁎a Department of Chemistry, UM-DAE Centre for Excellence in Basic Sciences, University of Mumbai, Vidhyanagari Campus, Mumbai 400098, Indiab Department of Chemical Sciences, Tata Institute of Fundamental Research, Mumbai, Indiac Molecular Simulations Group, Department of Pharmaceutical Chemistry, Bombay College of Pharmacy, Santacruz (East), Mumbai 400098, India

⁎ Correspondence to: R.V. Hosur, UM-DAE Centre foHealth Centre, University of Mumbai, Vidhyanagari Camp+91 22 2280 4545x2488.⁎⁎ Correspondence to: B. Ahmad, UM-DAECentre for ExcCentre, University ofMumbai, Vidhyanagari Campus, Mum2653 2119.

E-mail addresses: hosur@tifr.res.in (R.V. Hosur), basir.

1570-9639/$ – see front matter © 2014 Elsevier B.V. All rhttp://dx.doi.org/10.1016/j.bbapap.2014.01.009

a b s t r a c t

a r t i c l e i n f o

Article history:Received 16 November 2013Received in revised form 1 January 2014Accepted 14 January 2014Available online 24 January 2014

Keywords:Protein misfoldingAggregation inhibitorThioflavin T assayElectron microscopyAggregation kineticsMolecular dynamics simulation

Interaction of small molecule inhibitors with protein aggregates has been studied extensively, but how these in-hibitors modulate aggregation kinetic parameters is little understood. In this work, we investigated the ability oftwo potential aggregation inhibiting drugs, curcumin and kaempferol, to control the kinetic parameters of aggre-gation reaction. Using thioflavin T fluorescence and static light scattering, the kinetic parameters such as ampli-tude, elongation rate constant and lag time of guanidine hydrochloride-induced aggregation reactions of hen eggwhite lysozymewere studied.We observed a contrasting effect of inhibitors on the kinetic parameters when ag-gregation reactions were measured by these two probes. The interactions of these inhibitors with hen egg whitelysozyme were investigated using fluorescence quench titration method and molecular dynamics simulationscoupled with binding free energy calculations. We conclude that both the inhibitors prolong nucleation of amy-loid aggregation throughbinding to region of the proteinwhich is known to form the core of the proteinfibril, butonce the nucleus is formed the rate of elongation is not affected by the inhibitors. This work would provide in-sight into the mechanism of aggregation inhibition by these potential drug molecules.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

From a physicochemical perspective, the process of amyloid-like ag-gregation is a generic feature of polypeptide chains that needs to be fullyunderstood for a thorough characterization of the nature of proteins[1,2]. Molecules that inhibit aggregation through either specific ornon-specific interaction do so by modulating inter- and/or intra-molec-ular interactions between polypeptide chains. Therefore, investigationof inhibition of aggregation is useful for clarifying the mechanisms un-derlying protein aggregation [3]. Protein aggregation is linked to alarge number of pathological conditions in humans [4] and is a majorproblem in the production of many recombinant therapeutic proteins[5]. Therefore, from a biomedical perspective, inhibition of aggregationis a public health priority.

Human lysozyme mutants have been known to form huge amyloiddeposits in the livers and kidneys of individuals affected by hereditarysystemic amyloidosis. Currently, there are no clinical treatments avail-able to prevent or reverse the formation of such amyloid deposits. It is

r Excellence in Basic Sciences,us, Mumbai 400098, India. Tel.:

ellence in Basic Sciences, Healthbai 400098, India. Tel.: +91 22

ahmad@cbs.ac.in (B. Ahmad).

ights reserved.

also known that destabilized lysozyme mutants aggregate faster thanwild type lysozyme both in vitro and in vivo [6,7]. Lysozymes from var-ious sources are also known to form amyloid-like aggregates in a varietyof destabilizing conditions [8–10]. Moreover, despite a large amount ofwork supporting the importance of the small molecule inhibitors inpreventing amyloidogenesis [11,12], little is known about the precisemechanism by which they inhibit aggregation and the effect thatthese inhibitors have on different kinetic parameters of the aggregationprocess, on interaction with native state and aggregation precursorstate, on unfolding pathway and on the overall aggregation pathway[13].

In this work, we investigated the effect of two small polyphenolicmolecules, curcumin and kaempferol, on different kinetic parameterssuch as amplitude, elongation rate constant and lag time of guanidinehydrochloride-induced aggregation process of hen egg white lysozyme(HEWL). We have also shown the effect of interaction of these mole-cules on the unfolding pathway and on the structure of HEWL.Curcumin, a polyphenol found in the spice turmeric, has been shownto have strong antioxidant property and inhibit aggregation of manyproteins and peptides [13–15]. On the contrary, kaempferol, which isalso a natural flavonol found in many plant species and commonlyused in traditional medicine, is a weak antioxidant and its aggregationinhibition properties are not known [16]. We show how the interactionof these molecules with the HEWL affects the kinetic parameters of ag-gregation reactions and equilibrium unfolding pathway of protein. On

671M.S. Borana et al. / Biochimica et Biophysica Acta 1844 (2014) 670–680

the basis of these findings, we emphasize that a full understanding ofthe kinetic parameters is essential for deducing the correct mechanismof aggregation inhibition by potential aggregation inhibition drugs.

2. Materials and methods

2.1. Materials

Hen egg white lysozyme (HEWL), curcumin, kaempferol, guanidinehydrochloride (GuHCl), 1-anilinonapthalene-8-sulfonic acid (ANS),thioflavin T (ThT), and Congo red (CR) were purchased from Sigma-Aldrich Co. All other reagents were of analytical grade with purityN 99%. The concentrations of HEWLwere determined by UV absorbance(at 280 nm, with extinction coefficients (ɛ1%) of 2.63) [17].

2.2. Aggregation inhibition kinetic studies (ThT and light scattering assays)

The aggregation of HEWL in the absence and presence of variousconcentrations of inhibitors was initiated by continuous stirring of theprotein (140 μM) at 230 rpm (rotation per minute) in 4 M GuHCl at37.0 ± 0.1 °C and pH 6.3. Kinetics of aggregation was monitored usingThT and 90° light scattering assays. 100 μL of aggregation reaction mix-tures at different time intervals was added to 1.9 mL of 25 μM ThT solu-tion and pH 7.4 Na-phosphate buffer for ThT assay and light scatteringmeasurements, respectively. The fluorescence of the resulting sampleswas measured at room temperature with excitation wavelength fixedat 440 nm. The 90° light scattering of the resulting sample was mea-sured at room temperature with excitation wavelength and emissionwavelength fixed at 350 nm. In each case, slit width for excitation andemission was 3 nm each.

All the kinetic traceswere satisfactorily fitted to a sigmoidal functionas described by the following equation.

F ¼ F0 þm0xþF1 þm1x

1þ e− x−x1=2ð Þ=τ½ � ð1Þ

where F0 is the observed ThT fluorescence ratio; x and x1/2 are time andtime to 50% of maximum fluorescence, respectively. Therefore, the ap-parent rate constant (kapp) for the aggregation is given by 1/τ, and thelag time is given by x0 − 2τ.

2.3. Dye binding assays of final aggregates

To induce aggregation, HEWL solutions (140 μM) in the absence andpresence of inhibitors were subjected to stirring at 230 rpm for 3 h in4 M GuHCl at 37.0 ± 1.0 °C and pH 6.3. ThT, ANS and CR binding assayswere performed as described previously [18]. Briefly, 100 μL of finalaggregate samples in the absence and presence of inhibitors was addedto 1.9mLof respective solutions prepared in 25mMNa-phosphate bufferat pH 7.4. The ThT fluorescence and ANS fluorescence of the resultingsamples were measured at room temperature with excitation wave-length fixed at 440 nm and 380 nm, respectively. The light scattering ofthe resulting samples was measured with excitation wavelength andemission wavelength fixed at 350 nm. The CR assay was performed byincubating various aggregate samples with CR for 2–3 min and opticalabsorption was monitored between 400 and 700 nm. Samples withoutCR and without protein were used as control to obtain differencespectrum.

2.4. Amyloid intrinsic fluorescence

GuHCl-induced HEWL aggregates prepared as described abovewerealso characterized by amyloid intrinsic fluorescence. 100 μL of final ag-gregate samples in the absence and presence of inhibitors was addedto 1.9 mL of 25 mM Na-phosphate buffer (pH 7.4) and fluorescence

spectra between 365 nm and 500 nm were monitored by excitation ofthe sample at 357 nm with both excitation and emission slit widthskept at 10 nm [19].

2.5. Circular dichroism studies

Far-UV circular dichroism (far-UV CD) spectra of various aggregatesin the absence and presence of inhibitors was measured between 200and 250 nm using Jasco 810 Spectropolarimeter. Far-UV CD spectra ofnative HEWL and 4 M GuHCl unfolded protein without and withdifferent concentrations of inhibitors were also measured. For allmeasurements, a protein concentration of 2.8 μM, slit width of 1 nmand a cell of 1 mm path length were used. The data were presentedas mean residue ellipticity [θ] in deg cm2dmol−1, which is definedas [θ] = CD / (10 × n × l × Cp), where CD is in milli-degree, n is thenumber of amino acid residues (129), l is the path length of thecell in cm, and Cp is the molar concentration of the protein in mono-meric form. The amount of secondary structures (%α-helix) of nativeHEWL in the absence and presence of inhibitors was determined asdescribed in Supplementary information [20].

2.6. Protein intrinsic fluorescence

All aggregate samples (140 μM) were diluted 100 fold in 25 mMsodium phosphate buffer at pH 7.4 immediately before fluorescencemeasurements. Native, 4 M and 6 M GuHCl-incubated HEWL samplesin the absence and presence of different concentrations of inhibitorswere also prepared at the protein concentration of 1.4 μM. Intrinsic fluo-rescence spectra of all these samples were acquired between 300 and400 nm upon exciting the protein at 280 nm.

2.7. Scanning electron microscopy (SEM)

SEM images of various aggregates with and without inhibitors wereobtained as described previously [21]. Aggregate sample was diluted to35 μMusing distilled water. A drop of the resulting solution was depos-ited on aluminum foil and was subsequently air dried. Samples of pro-tein aggregate without inhibitors and those of 2:1 and 5:1 inhibitor toprotein ratios were analyzed at low vacuum using Quanta 200 SEM atdifferent magnifications.

2.8. GuHCl-induced unfolding studies

Solutions for unfolding experiment in the absence and presence ofinhibitors were prepared as described previously [22]. The unfolding/refolding of the proteinwas followed bymeasurements of intrinsic fluo-rescence of the protein. The protein sample was excited at 280 nm andemissionwasmonitored between 300 and 450 nm. All data points weresubtracted with their respective blank. The unfolding profiles were an-alyzed according to two and three state models as described by the fol-lowing equations [23].

F ¼ FN þ FUe− ΔG0−m GuHCl½ �ð Þ=RT

1þ e− ΔG0−m GuHCl½ �ð Þ=RT ð2Þ

F ¼ FN þ FI exp − GI−m1 Gu½ �ð Þ=RTf g þ FU exp − G1−m1 Gu½ �ð Þ=RTf g exp − G2−m2 Gu½ �ð Þ=RTf g1þ exp − GI−m1 Gu½ �ð Þ=RTf g þ exp − GI−m1 Gu½ �ð Þ=RTf g exp − G2−m2 Gu½ �ð Þ=RTf g

ð3Þ

where, F is the observed fluorescence intensity at 340 nm. FN, FI and FUare the fluorescence intensity of native, intermediate and unfolded pro-tein states, respectively. R and T are the gas constant and temperature inKelvin (298 K).ΔG0 is the change in free energy of unfolding of HEWL inthe presence of inhibitors. ΔG1 and ΔG2 are the changes in free energiescorresponding to N ↔ I and I ↔ U transitions, respectively.

672 M.S. Borana et al. / Biochimica et Biophysica Acta 1844 (2014) 670–680

2.9. Protein-inhibitor binding studies

A tock solution of curcumin was freshly prepared in methanol andthe concentration was determined by measuring the absorbance at425 nm, using the extinction coefficient ε425nm = 54, 954 cm−1 M−1.Kaempferol stock solution was prepared by dissolving 5 mg ofkaempferol in 5 mL ethanol. For fluorescence measurements taken atroom temperature, to a fixed volume of 3 μM protein solution (3 mL)previously incubated with 0 and 4 M GuHCl, increasing volumes(1–20 μL) of inhibitor solutions were added. Trp fluorescence between310 and 450 nm was measured by exciting the protein at 295 nm.Both emission and excitation slits were kept at 3 nm. The effect ofhighest concentration of methanol (or ethanol) used in the sampleson the intrinsic fluorescence spectrum on the HEWL was investigated.It was found that the spectra of HEWL in the presence and absence ofmethanol (or ethanol) were almost identical which indicated no alter-ation of HEWL conformation in the presence of highest concentrationof methanol (or ethanol) used in this study.

In order to understand the inhibitor quenching mechanism, datawere analyzed using Stern–Volmer equation [24].

F0F

¼ 1þ KSV � I½ �ð Þ ð4Þ

where, KSV is the Stern–Volmer constant and I is the concentration ofinhibitors, curcumin or kaempferol. The association constant Ka andthe number of binding sites were calculated using non-linear regressionof the data using equation [24].

F0−FF

¼ 12

1þ 1K � n� Pt½ � þ

Lt½ �n� Pt½ �

� �−1

2

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1

K � n� Pt½ �� �2

− 4� Lt½ �n� Pt½ �

sð5Þ

where, P and I are total concentrations of protein and inhibitors,respectively.

2.10. Computational methods

2.10.1. Protein preparationThe crystal structure of the hen egg white lysozyme (HEWL),

retrieved from the Protein Data Bank (PDB ID: 2WAR), which wasbound to NAG multimer was prepared using Protein PreparationWizard (Schrödinger Suite 2013-1). The heavy atom valencies werecompleted by addition of hydrogen atoms. No significant interactionswere observed between the crystal waters and the ligand; hence thewaters were removed from the structure. The atom types and thebond types were rectified as per the OPLS2005 force field. The proton-ation state of the ionizable amino acids in the protein was adjusted topH 7.0. The crystal structure was relaxed until the heavy atoms con-verged to a RMSD of 0.01 Å.

2.10.2. Docking studiesThe binding conformations for curcumin and kaempferol were iden-

tified using docking. Initially a grid was generated centered on the NAGmultimerwith the inner grid cube dimensions of 10Å and the outer gridof 20 Å along x–y–z direction for maximum conformational space ex-ploration of the ligands during docking in the binding pocket. The vanderWaals potentials were softened for the protein binding site by scal-ing the van derWaals radii to 1.0 Å over the non-polar areas of the pro-tein that lie within the grid extents while the partial atomic chargeswere set to 0.25. The receptor atoms beyond the extent of the gridwere unscaled. The probability of hydrogen bonds between the residuesin the protein active site and the ligand was maximized by setting freerotations for the side-chain hydroxyl groups of the amino acids serine,threonine, and tyrosine. After complete parameterization, the ligands

curcumin and kaempferol were docked and the docking poses wereranked by the GlideScoreXP scoring function.

2.10.3. Molecular dynamics simulationsCurcumin and kaempferol in complex with HEWL were simulated

for duration of 5 ns using parallel code of pmemd in AMBER 12 molec-ular modeling suite [25] on a 24 CPU Intel linux_x86_64 computingcluster. The coordinate and parameter files were built using tleap pro-gram. The atom types and parameters were assigned using the FF99SBforce-field. The protein–ligand complexes were initially minimizedand further equilibrated for 1 ns. The system was restrained for heavyatom (protein and inhibitor) motions during minimization and heatingphase to allow solvent homogenization around the complex. The solvat-ed protein–ligand complexes were heated linearly from 0 K to 300 Kover 100 ps in the canonical NVT ensemble using a Langevin thermostat,with collision frequency of 2.0 ps−1 Å−2. The production runs weremade in the NPT ensemble at 27 °C (temperature was controlled withLangevin thermostatwith a 2.0 ps−1 collision frequency). The trajectorywas sampled every 10 ps for a duration of 5 ns. The time step for calcu-lation used in all stageswas set to 2 fs,with hydrogen atoms constrainedby the SHAKE algorithm [26]. The long-range electrostatics was includ-ed on every step using the Particle Mesh Ewald algorithm with a 4thorder B-spline interpolation [27,28]. The trajectories generated duringthe simulations were analyzed for any fluctuations in the rmsd, rmsf,temperature, density and potential energies using the ptraj tool ofAmberTools13 [25].

2.10.4. Binding energy calculations using MM–GBSA methodsThe binding energies (ΔG) were calculated for curcumin and

kaempferol in complex with HEWL over the 500 snapshots sampledevery 10 ps in the molecular dynamics simulation using the MM–

GBSA methods (Molecular Mechanic–Generalized Born Surface Area)in AmberTools13. The binding energies were further decomposedresidue-wise to determine the amino acids that are involved in theinteraction with curcumin and kaempferol.

3. Results

3.1. The effect of polyphenols on GuHCl induced aggregation of lysozyme

Hen egg white lysozyme (HEWL) is known to exist in partiallyfolded conformation in guanidine hydrochloride (GuHCl) solutionsand this is dependent on both GuHCl concentration and temperature.The partially folded state of HEWL accumulated at 3 M and 4 M GuHClconcentrations in the temperature ranges of ~35–60 °C and ~20–45 °C, respectively [29]. Moreover, it was observed that maximumthioflavin T (ThT) binding and rate of amyloid-like aggregation occurredin 3MGuHCl at 50 °C [29]. Since the temperature causes both structuralalterations and aggregation acceleration, we investigated the effect ofGuHCl concentrations on the aggregation of HEWL at physiological tem-perature of 37 °C. Aggregation of HEWL was induced by incubating140 μM of the protein in different concentrations of GuHCl at 37 °C,pH 6.3 and with constant stirring at 230 rpm. When the samples weremonitored by ThT binding assay and turbidity measurements after 3 hof incubation, a huge rise in both ThT fluorescence and turbiditywas ob-served inGuHCl concentration range of 3.5–5Mwith amaximumat 4MGuHCl (Supplementary information, SI Figure S1A and S1B). Both ThTfluorescence and turbidity of the aggregates at ~4 M GuHCl are hugelylarger than the values at 3.5M or 5.0MGuHCl. However no aggregationwas seen below 3.5 M or above 5.5 M GuHCl concentrations. Moreover,no aggregation is seen in the absence of stirring (SI Figure S1A and S1B).These results indicate that HEWL aggregates maximally at 4 M GuHClconcentration at physiological temperature of 37 °C.

We also examined the effects of curcumin and kaempferol, whichhave been known to have strong and weak antioxidant properties, re-spectively on the aggregation process of HEWL [16]. We found that

673M.S. Borana et al. / Biochimica et Biophysica Acta 1844 (2014) 670–680

both inhibitors markedly reduced the ThT fluorescence without signifi-cantly affecting the turbidity of the samples (SI Figure S1A and S1B).

3.2. Morphology of GuHCl induced final aggregate

Given that non-fibrillar aggregates may show high ThT fluorescencesignal, we confirmed the nature of final aggregates by investigating themorphology of the final aggregates in the absence and presence ofcurcumin and kaempferol using scanning electron microscopy (SEM).The SEMhas beenused to show the size and shapes of protein aggregatesformany proteins [21,30,31]. In the absence of inhibitors, predominantlyfibrillar aggregates in the form of mesh are observed (Fig. 1A).

As can be seen from the figure we observed some large fibrils of dif-ferent diameters on themesh of HEWL fibrils. Fig. 1B shows a large fibrilwhich is ~500 nm in diameter and more than 30 μm in length. The for-mation of such large fibrils has been observed in other proteins such aswheat gluten and gliadin upon prolonged incubation of samples for amonth under physiological conditions [21]. The formation of theselarge fibrils in 4 M GuHCl condition also indicates the strength of inter-actions stabilizing these structures. It is known that large fibrils formthrough lateral association of smaller fibrils. It is agreed that proteinsforming large fibrils contain charge-less domains of hydrophobic char-acter, which are the main source of the short-range attraction [21,32].Since GuHCl disrupts mainly hydrophobic interactions at higher con-centrations (4–8 M) [33], we infer that loss of interactions disruptedby GuHCl is compensated by interactions stabilizing lateral associationof HEWL fibrils into large fibrils.

Fig. 1. Scanning electron microscope image of amyloid fibrils with and without inhibitors.Figure A depicts the SEM image of hen egg white lysozyme fibrils in the absence of anyinhibitors. Figure B shows a magnified HEWL fibril. Amyloid fibrils at curcumin/proteinmolar ratios of 2:1 and 5:1 are shown in Figures C and E, respectively. Figures D and Fshow HEWL aggregates in the presence of kaempferol at kaempferol/protein molarratios of 2:1 and 5:1, respectively.

Such large fibrils were not observed in the presence of curcumin/lysozyme molar ratio of 2:1 (Fig. 1C). However, in the presence ofkaempferol/lysozyme molar ratio of 2:1 we observed some fibrillarmesh, which appears smaller than the fibrils present in the absenceof inhibitors (Fig. 1D). At higher inhibitors/protein molar ratio(5:1) no fibrillar structure is observed (Fig. 1E and F). These resultssuggest that both the inhibitors are able to prevent the formationof fibrillar aggregates but curcumin seems to prevent fibrillar aggre-gation more effectively. Curcumin is known to completely inhibit ag-gregation of many proteins mainly occurring in mild denaturingconditions [13,14,34]. In order to understand the mechanism of ag-gregation and interactions involved in stabilization of the aggregatesunder such strong denaturing conditions, we investigated the effectsof these polyphenols on aggregation kinetic parameters.

3.3. Effects of polyphenols on lysozyme aggregation kinetic parameters

Fig. 2 shows kinetic traces of HEWL aggregation in the absence andpresence of various inhibitors/protein molar ratios. The aggregation isinitiated by incubating 140 μMHEWL in 4 M GuHCl at 37 °C and the fi-brillation process was monitored by ThT fluorescence (Fig. 2A and C)and aggregation by 90° light scattering at 350 nm (Fig. 2B and D).Each trace displays lag phase kinetics andfits to a sigmoidal function de-scribed by Eq. (1) satisfactorily (R2= 0.9913 to 0.9987), indicating thatHEWL aggregation is consistent with nucleation-dependent polymeri-zation model [35]. The significant increase in the ThT fluorescence(~55 times as compared to blank ThT fluorescence) in the absence ofinhibitors indicates that the aggregate is mainly fibrillar in nature. Inthe presence of both the inhibitors, ThT fluorescence of the aggregatedecreases significantly but it is still markedly higher (~5 times com-pared to blank ThT) than the ThT fluorescence observed for native pro-tein (Fig. 2A and C). However, no effect of inhibitors on the totalaggregates was observed (Fig. 2B and D) suggesting that both the inhib-itors were modulating the aggregation pathway from fibrillar to amor-phous aggregation, which is also evident from the SEM image (Fig. 1A).

In order to understand the mechanism of GuHCl-induced aggrega-tion process, we compared the kinetic parameters such as amplitude(A), time constant (τ inverse of apparent rate constant, 1/kapp), andlag time. Fig. 3 shows the amplitude, time constant and lag time ex-tracted from 4MGuHCl-induced aggregation kinetics of HEWL in theabsence and presence of different concentrations of inhibitors. Expo-nential decay-type dependence between the amplitude of the aggre-gation reactions and the inhibitor concentrations is observed, whenthe fibrillar aggregation kinetics were followed by ThT fluorescence(Fig. 3A and B, black triangles). On the other hand, the amplitudesof the aggregation kinetics monitored by light scattering (Fig. 3Aand B, cyan circles) exhibited just a small increase. Since ThT fluores-cence monitors fibrillar aggregation and light scattering monitorsboth fibrillar and amorphous aggregation [36], the decrease in theamplitude when the aggregation processes were monitored by ThTindicates that both inhibitors decrease the amount of fibrils formed.

Fig. 3C and D shows the aggregation reaction time constants (τ) ver-sus inhibitor concentration plots. It can be seen that aggregation rates asmonitored by ThT fluorescence are almost unaffected in the presence ofcurcumin at all the concentrations investigated (Fig. 3C), and forkaempferol the same was observed only up to kaempferol/HEWLmolar ratio of 2:1. At the highest kaempferol/HEWLmolar ratio studiedhere (5:1), we observed a significant decrease in the aggregation rate asmonitored by ThTfluorescence (Fig. 3D). The aggregation rate decreasesup to curcumin/HEWLmolar ratio of 2:1 and then becomes almost con-stant at higher molar ratio, when aggregation reaction kinetics wasmonitored by light scattering at 350nm(Fig. 3C). By contrast, the aggre-gation rate decreases linearly above kaempferol/HEWL molar ratio of0.5:1 (Fig. 3D). From these results it appears that the rate of fibrillar ag-gregation is independent of curcumin concentrations but it is sloweddown at higher concentrations of kaempferol.

Fig. 2.Aggregation time courses of hen eggwhite lysozyme. HEWL aggregation kineticsmonitoredwith ThTfluorescence (A and C) and light scattering at 350 nm(B andD) in the absenceand presence of different concentrations of inhibitors. Figures A and B present aggregation kineticswithout andwith inhibitor curcumin, respectively. Figures C andD present aggregationkinetics without and with inhibitor kaempferol, respectively. In all cases data is reported as ThT fluorescence ratio or LS ratio as a function of time. The lines of best fit through the datapoints were obtained by fitting the data with a sigmoidal function (Eq. (5)).

674 M.S. Borana et al. / Biochimica et Biophysica Acta 1844 (2014) 670–680

Fig. 3E and F shows that the lag time of HEWL aggregation asfollowed by ThT assay increases with increasing concentration ofboth the inhibitors. On the contrary, they do not show any signifi-cant change when the process is monitored by 90° LS measurements(Fig. 3E and F, cyan symbols). These results can be interpreted sayingthat both the inhibitors prolong formation of nuclei for fibrillar ag-gregation. No change in the lag time as monitored by LS shows thepresence of partially unfolded species sufficient enough for initiat-ing amorphous aggregation and hence their formation seems to beunaffected by these inhibitors.

The nature of the dependence of aggregation rates and amplitudeson inhibitor concentrations as described above indicates that bothcurcumin and kaempferol are not only inhibiting fibril formation butare also slowing down amorphous aggregation reaction.

3.4. Tinctorial and structural properties of the final aggregates

Thioflavin T and Congo red binding specifically monitor amyloid fi-brillation, whereas 8-anilino-1-naphthalenesulfonate (ANS) fluores-cence monitors both amyloid fibrillation and amorphous aggregation[13,18,36]. The results presented here show that the final HEWL aggre-gate in the absence of the polyphenols (obtained as described inMaterials and methods) binds to specific dyes such as ThT, ANS, andCR which causes the spectral changes associated with protein fibrils(SI Figure S2). In contrast, in the presence of both the polyphenols dyebinding characteristics of the aggregates decrease significantly to an

extent that is observed for unstructured aggregates or amyloid-like olig-omers (Figure S2). From these results it seems that both polyphenolscompletely prevent fibril formation at higher polyphenols/proteinmolar ratio (≥5).

Recently, Chan et al. have discovered fluorescence like emission byamyloid like fibrils in the visible range [19]. This intrinsic fluorescenceof amyloid fibrils is different from the intrinsic fluorescence of proteinsas it does not depend on the presence of aromatic amino acid residueswithin the polypeptide chain. Rather, it seems to originate from elec-tronic levels that become available for transition when the polypeptidechain converts into a cross-β sheet structure. Fig. 4A shows the amyloidintrinsic fluorescence spectra of HEWL aggregates in the absence andpresence of polyphenols, curcumin and kaempferol. In the absence ofthe polyphenols, the spectrum of HEWL aggregate shows a wavelengthof maximum emission (λmax) at 442 nmwhen the aggregate was excit-ed at 355 nm, a typical feature of amyloid intrinsic fluorescence [19].On the other hand, in the presence of either of the polyphenols al-most negligible amyloid intrinsic fluorescence is detected. Thisdata again shows that both polyphenols inhibit the fibrillar aggrega-tion of HEWL.

The secondary structure of the GuHCl-induced HEWL aggregate inthe absence and presence of both the polyphenols was monitoredusing far-UV circular dichroism (CD) in the wavelength range between200 nm and 240 nm (Fig. 4B and C).

The far-UV CD spectrum of the aggregate in the absence of polyphe-nols shows a single negative band at 217 nm, which indicates that the

Fig. 3. Effect of inhibitors, curcumin (Cur) and kaempferol (Kaem), on various kinetic parameters of hen eggwhite lysozyme aggregation. The effects of inhibitors on amplitude (A and B),time constants (C andD) and lag time (E and F) are shown as a function of inhibitor to protein ratio. All valueswere extracted byfitting experimental data in Fig. 1 as described inMaterialsand methods section.

675M.S. Borana et al. / Biochimica et Biophysica Acta 1844 (2014) 670–680

aggregate contains extended β-sheet conformation. In the presence ofcurcumin the band at 217 nm co-exists with another negative CDband at around 222 nm. This indicates that HEWL aggregate formed inthe presence of curcumin is not a pure extended β-sheet structure.Since a strong negative band at 222 nm is a characteristic feature ofalpha-helix, it seems that the aggregate formed in the presence ofcurcumin also contains some helical structure. In contrast, we found asignificant decrease in the β-sheet conformation of the aggregate inthe presence of kaempferol.

3.5. Effect of inhibitors on the unfolding pathway and stability of HEWL

In order to understand the mechanism of fibril formation and its in-hibition in detail, we also studied the effect of these inhibitors on theequilibrium unfolding process of HEWL. We investigated the unfoldingprocess of the protein by monitoring GuHCl-induced denaturation atequilibrium at pH 6.3 using protein intrinsic fluorescence. Fig. 5A

shows the GuHCl-induced unfolding transitions of HEWL in the absenceand presence of inhibitors/lysozyme molar ratio of 2:1. We found thatthe denaturation curves of all the samples are sigmoidal and consist ofsharp single transition.

This indicates that observed transitions occur between two states.The unfolding curves as monitored by intrinsic fluorescence were ana-lyzed by previously described phase diagrammethod that is able to de-tect invisible intermediates present in the unfolding curves [37]. In theabsence of inhibitors, the phase diagram consists of two linear parts(Fig. 5B). The first linear part between 0 and 4 M GuHCl correspondsto transition from native (N) state to a partially folded (I) state, andthe second linear part between 4 and 6M GuHCl corresponds to transi-tion from I to unfolded (U) state. However, the phase diagrams in thepresence of inhibitor are best fitted (regression coefficients R2 in thepresence of curcumin and kaempferol are 0.9901 and 0.9876, respec-tively) with only one linear part. These results suggest accumulationof an intermediate state around 4 M GuHCl concentrations at 37 °C in

Fig. 4. Structural properties of hen egg white lysozyme (L) aggregate. Effect of inhibitors, curcumin (C) and kaempferol (K), on the amyloid intrinsic fluorescence (A) and on the far-UVcircular dichroism spectra of HEWL aggregates in the absence and presence of curcumin (B) and kaempferol (C). For amyloid intrinsic fluorescence the aggregate was excited at 357 nm.

676 M.S. Borana et al. / Biochimica et Biophysica Acta 1844 (2014) 670–680

the absence of any inhibitors which is in agreement with the previousreport [29]. Contrarily, the I state accumulated around 4 M GuHCl isnot detectable in the presence of the inhibitors suggesting that observedtransitions occur between two conformational states, N and U.

We next determined the conformational stabilities of HEWL in theabsence and presence of inhibitors by fitting data in Fig. 5A usingthree states (Eq. (3)) and two states (Eq. (2)), respectively. The valuesof the free energy change (ΔGN–I) and m values of unfolding of HEWLwithout inhibitors determined by extrapolating to 0 M GuHCl for N–Itransition are found to be 41.25 ± 5.34 kJ mol−1 and 13.14 ±1.40 kJ mol−2, respectively. The second transition (I–U) is difficult toquantitate correctly as it produces a very small signal change. The valuesof thermodynamic parameters namely ΔG and m of the N–U transitionin the presence of curcumin are found to be 43.55 ± 2.64 kJ mol−1 and11.09±0.68 kJmol−1M−1, respectively.ΔG andmvalues for unfoldingof HEWL in the presence of kaempferol are observed to be 45.97 ±3.05 kJ mol−1 and 11.39 ± 0.78 kJ mol−1 M−1 respectively. It can benoticed that theΔGs of unfolding in the presence of inhibitors are great-er than that of HEWL in the absence of inhibitors.

The increase in theΔG of unfolding indicates that both the inhibitorsincrease the stability of the protein. In contrast, the m values ofunfolding of HEWL with inhibitors were observed to be lower thanthe m values without inhibitors. These analyses indicate that GuHClunfolded state in the presence of inhibitors is slightly more compactcompared to one without inhibitors [38].

3.6. Inhibitors bind strongly to HEWL and moderately affect its structure

The binding of inhibitors, curcumin and kaempferol, to HEWLwasstudied by fluorescence quench titration method. This method hasbeen extensively used to elucidate the ligand binding behavior ofproteins [13,39]. Intrinsic fluorescence of HEWL shows maximumemission at 343 nm after excitation at 280 nm and the addition ofeither of the inhibitors causes a concentration dependent quenchingof fluorescence (Fig. 6A). Moreover, the bimolecular quenching rateconstants as calculated by analyzing the data by Stern–Volmer equa-tion were found to be in the range of 1013 M−1 s−1 for both theinhibitors (inset, Fig. 6A). This value is significantly larger than themax-imum collisional quenching constant (1010 M−1 s−1) (inset, Fig. 6A)[40].

These analyses suggest that stable complexes formbetween the pro-tein and the inhibitors. Fitting the binding data by non-linear regressionindicates that both curcumin and kaempferol bind strongly to HEWLwith association constants of around 104 M−1. These results are similarto the previous reports of binding of curcumin to HEWL [15].

Effects of curcumin and kaempferol binding on far-UV CD spectra ofnative (N) and 4 M GuHCl unfolded (I) states of HEWL is shown in SIFigure S3. As discussed in the SI, greater increase in secondary structurecontents of the I state compared to that in the N state was observed inthe presence of both the inhibitors (Figure S4). From these results, it ap-pears that evanescence of I–U transition from the unfolding profile

Fig. 5. Hen egg white lysozyme unfolding pathway. Fraction unfolded of hen egg whitelysozyme in the absence and presence of curcumin (cur) and kaempferol (kaem) dena-tured as function of GuHCl at pH 6.3 and 37 °C (A). The lines through the data pointswere obtained by fitting the data to a three-state model (Eq. (4), gray line) and two-state model (Eq. (3), red line for curcumin and dotted line for kaempferol). Figure Bshows the phase diagrams of HEWL unfolding in the absence and presence of inhibitors.The phase diagram was constructed as described in Materials and methods section.

Fig. 6. Inhibitors induced quenching of tryptophan fluorescence of native and 4 M GuHCldenatured HEWL at 25 °C. Fluorescence quenching of native and 4 M denatured HEWLdue to inhibitors is shown inFigureA. Inset in FigureA shows thequenching rate constantsof protein–inhibitors interaction. Nonlinear fitting of inhibitors binding isotherms accord-ing to Eq. (5) as described in Materials and methods (B). Inset in Figure B shows associa-tion constant (Ka) and number of binding site (n) for inhibitors and the native protein (N)and 4 M GuHCl-induced state (I) interactions, where C and K represent curcumin andkaempferol, respectively.

677M.S. Borana et al. / Biochimica et Biophysica Acta 1844 (2014) 670–680

HEWL is due to stabilization of N–I transition; thus, indicating accumu-lation of a new intermediate state (I′) which is characterized by in-creased secondary structure contents.

3.7. Molecular dynamics simulations and binding free energy calculations

3.7.1. DockingThe docking study was performed in the 10 Å active site defined

around the co-crystal NAG multimer. Subsequently curcumin andkaempferol were docked and scoredwithGlideScore XP. It was observedthat kaempferol had higher docking score compared to curcumin (SI,Table S1). The docking poses for the curcumin and kaempferol (2Dand 3D) are depicted in Fig. 7. Curcumin and kaempferol both dockedwell in the hydrophobic groove of the hen egg white lysozyme.Curcumin formed hydrogen bond interaction with Asp52, Trp62 andTrp63 while kaempferol formed H-bonds with Asp52, Leu56, Asn59,Trp62 and Trp63. The phenyl rings in curcumin and kaempferol formedπ–π with the indole ring of Trp62.

3.7.2. Molecular dynamics simulationsThe docked complexes of curcumin and kaempferol in complexwith

hen egg white lysozyme were simulated for a period of 5 ns using thepmemd code in AMBER12 molecular modeling suite. The solvatedHEWL: curcumin andHEWL: kaempferol complexeswere initially heat-ed to 27 °C linearly for 100 ps, and subsequently the density was opti-mized for 100 ps. After which the complexes were equilibrated for1 ns at 27 °C. The RMSD and RMSF were calculated for the simulationtrajectories to evaluate the stability of the complexes during the

equilibrium and the production phases and to monitor atomic fluctua-tions that take place at individual residue level. The RMSD calculatedover the trajectory was 0.6–1.2 Å indicative of complex stability duringequilibrium and production phases of dynamics simulation (Figures S5and S6). Maximum atomic fluctuations were observed for the aminoacids in the binding groove for residue numbers 60–80 and 90–120.The temperature was stable at 27 °C while the density remained at1.0 g/mL for the complexes as seen from the simulation trajectory anal-yses depicted in Figures S5 and S6 during the equilibrium and produc-tion phases, respectively. So also the potential energy of the systemremained stable as depicted in Figures S5 and S6. Overall the total anal-yses indicated the stability of the complexes during the 5 ns moleculardynamics simulation and provided significant confidence to use the tra-jectory frames for calculation of binding free energy.

3.7.3. Binding energy calculationsThe MM–GBSA binding energies were computed on the 500 struc-

tural snapshots captured during the 5 ns molecular dynamics simula-tion using the AmberTools13. The MM–GBSA values indicate that thekaempferol binds more tightly to HEWL than curcumin in relativeterms which is also supported by the docking energies (Table S1).

To get an insight into the amino acids that interact with curcuminand kaempferol the energies were decomposed at residue level(Figure S7 and Table S2). It is seen that curcumin has favored van derWaals interactions with Trp62 and Arg73 while kaempferol has withTrp62, Trp63 and Asn103. As for the electrostatic interactions, curcuminshows favored interactions with Asp48, Trp63 and Arg70 whilekaempferol has with Asp52 and Asp101. Some of these interactionswith Asp52, Trp62 and Trp63 are very well picked up during docking.

C D

BA

Fig. 7. Docking studies of interaction of inhibitors with hen egg white lysozyme. 2D and 3D ligand interaction diagram for curcumin (A and C) and kaempferol (B and D) with HEWL.

678 M.S. Borana et al. / Biochimica et Biophysica Acta 1844 (2014) 670–680

Electrostatic interaction (hydrogen bond) with Asp101 is the key inter-action that anchors kaempferol to HEWL and gives tighter bindingover curcumin which is harmoniously demonstrated by dockingand MM–GBSA calculations.

4. Discussion

The conversion of native protein into insoluble fibrillar structure,commonly known as amyloid aggregates has been observed for mostof the globular and intrinsically unfolded proteins [1,2,41]. Most inves-tigations suggest that the formation of a partially unfolded intermediatestate is a prerequisite for amyloid aggregation [42]. Moreover, mostamyloid aggregation follows nucleation–polymerization reaction,which is described by the sigmoidal growth kinetics. Three importantkinetic parameters namely lag time, growth rate and amplitude can beextracted from sigmoidal kinetics. Lag time is time before the start offibril growth and it corresponds to the formation of the nuclei frompartially folded or partially unfolded intermediate state. The growthrate describes the speed of fibril formation during elongation phase inthe sigmoidal curve. The amplitude represents the total amount of ag-gregates formed during the process. Thus, studying the effect of smallmolecule inhibitors on these kinetic parameters yields unique informa-tion about the mechanism of aggregation and aggregation inhibition.

In this work, we studied the effect of binding of two polyphenols,curcumin and kaempferol, to HEWL on these three kinetic parametersof HEWL aggregation and unfolding process of the protein. The workof Brian A. Vernaglia et al. has shown that HEWL formed amyloid-likefibrils within 2 h of stirring in 3 M GuHCl, pH 6.3 and at 50 °C [29].This time period is immensely suitable for studying the effect of inhibi-tor on the kinetics of protein aggregation. We observe that maximumaggregation occurs on stirring the protein at 230 rpm in 4 M GuHCl,pH 6.3 and at physiological temperature of 37 °C (SI Figure S1). The

aggregation reactions in the absence and presence of the inhibitorswere monitored by ThT fluorescence and 90° light scattering at350 nm. It is known that ThT fluorescence measures the amyloid-likeaggregation process and light scattering measures total aggregation[36].With increasing concentrations of both the inhibitors, we observed(i) increase in lag times, (ii) decrease in amplitude and (iii) no effect onthe elongation rates when the kinetics were followed using ThTfluorescence. In contrast we found (i) a slight decrease in lag time,(ii) no effect on the amplitude and (iii) a decrease in the rate of elonga-tion, when the aggregation processes were monitored with light scat-tering. Taken together, it appears that both the inhibitors prolong theformation of nuclei of amyloid aggregation. But once the nucleus isformed the rate of elongation remains unaffected. Furthermore, a signif-icant decrease in the amplitude when aggregation was followed by ThTfluorescence and no effect in the amplitude when the aggregation wasmonitored by light scattering suggest that the inhibitors significantlydecrease the amount of amyloid-like aggregates and increase theamount of non amyloid aggregates.

This finding is also supported by electron microscope images of theaggregates in the absence and presence of the inhibitors. Our data indi-cate that both the inhibitors are not only decreasing the total amyloid-like aggregates but are also slowing down the non-amyloid aggregation.

The interaction of curcumin and kaempferol by fluorescencequenching method and molecular dynamics simulations coupled withbinding free energy calculations shows that kaempferol has comparablebinding to that of curcumin. Both curcumin and kaempferol bindstrongly to monomeric HEWL with association constant Ka ~ 104 M−1.Moreover, docking and MM–GBSA calculations show that bindinggroove in lysozyme mainly involves amino acid residues from 48to 101. This region of the sequence corresponds to the highlyaggregation-prone region of the sequence identified in hen lysozyme[43–45]. These observations are similar to previous reports of

Fig. 8. Schematic of proposed mechanism. Schematic of the action of inhibitors, curcumin and kaempferol on GuHCl induced hen egg white lysozyme unfolding and aggregation.

679M.S. Borana et al. / Biochimica et Biophysica Acta 1844 (2014) 670–680

interaction of HEWL with polyphenols such as curcumin [15] and (−)-epigallocatechin gallate [46], as both bind to HEWL with Ka ~ 104 M−1.Docking studies show that epigallocatechin gallate, a polyphenol fromgreen tea interacts specifically with residues Trp 62 and Trp 63, whichalso fall in the same region [46]. From unfolding studies in the absenceand presence of inhibitors, it seems that a molecule of the inhibitorsbinds to HEWL and modulates the unfolding pathway in such a waythat the state stabilized at 4 M GuHCl in the presence of inhibitors isless amyloidogenic. Therefore, as depicted in Fig. 8, we conclude thatHEWL without inhibitors accumulates a highly amyloidogenic inter-mediate state in 4 M GuHCl. The inhibitors, curcumin and kaempferol,bind strongly to HEWL in the region that is known to form core of thefibrils and thus modulate the different kinetic parameters of HEWLaggregation-reaction.

5. Conclusion

We have demonstrated a correlation between small molecule inhib-itor binding to aggregation precursor state and aggregation kinetics,indicating that inhibitor binding affects aggregation kinetic parameterssuch as amplitude, elongation time constant and lag time in a concen-tration dependent manner. The results also show how these effects ofinhibitors on kinetic parameters control the nucleation process of amy-loid aggregation. Since these parameters control the aggregation pro-cesses, understanding the control of such interactions and kineticparameters is important for the development of therapeutic and diag-nostic strategies in situations where protein aggregation is associatedwith disease and also for production of medically important proteinsand peptides.

AbbreviationsHEWL Hen egg white lysozymeGuHCl Guanidine hydrochlorideANS 1-Anilinonapthalene-8-sulfonic acidThT Thioflavin TCR Congo redMM–GBSA Molecular mechanic–generalized born surface areaSEM Scanning electron microscopy

Acknowledgement

The authors gratefully acknowledge the UM-DAE-Centre for Excel-lence in Basic Sciences and Tata Institute of Fundamental Research forinstrumentation facilities. RRSP acknowledges DST-SERB (SR/SO/HS-0117/2012) for the computational facility provided at the BombayCollege of Pharmacy. MSB is thankful to the Department of Scienceand Technology, Government of India for Innovation in Science Pursuitfor Inspired Research (INSPIRE) fellowship.

Appendix A. Supplementary data

Computational methods, ThT fluorescence, light scattering, ANSfluorescence and circular dichroism data to highlight the effect of inhib-itors, curcumin and kaempferol, on guanidine hydrochloride inducedaggregation of hen egg white lysozyme are provided as associated con-tent (Figures S1 to S7). Supplementary data to this article can be foundonline at http://dx.doi.org/10.1016/j.bbapap.2014.01.009.

References

[1] C.M. Dobson, Protein misfolding, evolution and disease, Trends Biochem. Sci. 24 (9)(1999) 329–332.

[2] F. Chiti, P.Webster, N. Taddei, A. Clark,M. Stefani, G. Ramponi, C.M. Dobson, Designingconditions for in-vitro formation of amyloid protofilaments and fibrils, Proc. Natl.Acad. Sci. U. S. A. 96 (7) (1999) 3590–3594.

[3] M. Necula, R. Kayed, S. Milton, C.G. Glabe, Small molecule inhibitors of aggregationindicate that amyloid beta oligomerization and fibrillization pathways are indepen-dent and distinct, J. Biol. Chem. 282 (14) (2007) 10311–10324.

[4] F. Chiti, C.M. Dobson, Protein misfolding, functional amyloid, and human disease,Annu. Rev. Biochem. 75 (2006) 333–366.

[5] S. Ventura, A. Villaverde, Protein quality in bacterial inclusion bodies, TrendsBiotechnol. 24 (4) (2006) 179–185.

[6] M.B. Pepys, P.N. Hawkins, D.R. Booth, D.M. Vigushin, G.A. Tennent, A.K. Soutar, N.Totty, O. Nguyen, C.C. Blake, C.J. Terry, T.G. Feest, A.M. Zalin, J.J. Hsuan, Human lyso-zyme gene mutations cause hereditary systemic amyloidosis, Nature 362 (6420)(1993) 553–557.

[7] L.A. Morozova-Roche, J. Zurdo, A. Spencer, W. Noppe, V. Receveur, D.B. Archer, M.Joniau, C.M. Dobson, Amyloid fibril formation and seeding by wild-type humanlysozyme and its disease-related mutational variants, J. Struct. Biol. 130 (2–3)(2000) 339–351.

[8] S. Raccosta, V. Martorana, M. Manno, Thermodynamic versus conformational meta-stability in fibril-forming lysozyme solutions, J. Phys. Chem. B 116 (40) (2012)12078–12087.

[9] L.N. Arnaudov, R. de Vries, Thermally induced fibrillar aggregation of hen egg whitelysozyme, Biophys. J. 88 (1) (2005) 515–526.

[10] Y. Zou, Y. Li, W. Hao, X. Hu, G.Ma, Parallel β-sheet fibril and antiparallel β-sheet olig-omer: new insights into amyloid formation of hen egg white lysozyme under heatand acidic condition from FTIR spectroscopy, J. Phys. Chem. B 117 (15) (2013)4003–4013.

[11] R. Liu, R. Su, M. Liang, R. Huang, M. Wang, W. Qi, Z. He, Physicochemical strategiesfor inhibition of amyloid fibril formation: an overview of recent advances, Curr.Med. Chem. 19 (24) (2012) 4157–4174.

[12] D.E. Ehrnhoefer, J. Bieschke, A. Boeddrich,M. Herbst, L. Masino, R. Lurz, S. Engemann,A. Pastore, E.E. Wanker, EGCG redirects amyloidogenic polypeptides into unstruc-tured, off-pathway oligomers, Nat. Struct. Mol. Biol. 15 (2008) 558–566.

[13] B. Ahmad, L.J. Lapidus, Curcumin prevents aggregation in α-synuclein by increasingreconfiguration rate, J. Biol. Chem. 287 (12) (2012) 9193–9199.

[14] F. Yang, G.P. Lim, A.N. Begum, O.J. Ubeda, M.R. Simmons, S.S. Ambegaokar, P.P. Chen,R. Kayed, C.G. Glabe, S.A. Frautschy, G.M. Cole, Curcumin inhibits formation of amy-loid beta oligomers and fibrils, binds plaques, and reduces amyloid in-vivo, J. Biol.Chem. 280 (7) (2005) 5892–5901.

[15] S.S. Wang, K.N. Liu, W.H. Lee, Effect of curcumin on the amyloid fibrillogenesis ofhen egg-white lysozyme, Biophys. Chem. 144 (1–2) (2009) 78–87.

[16] K.L. Khanduja, A. Bhardwaj, Stable free radical scavenging and antiperoxidativeproperties of resveratrol compared in vitro with some other bioflavonoids, IndianJ. Biochem. Biophys. 40 (6) (2003) 416–422.

[17] M.E. Goldberg, R. Rudolph, R. Jaenicke, A kinetic study of the competition betweenrenaturation and aggregation during the refolding of denatured-reduced eggwhite lysozyme, Biochemistry 30 (11) (1991) 2790–2797.

[18] B. Ahmad, J. Winkelmann, B. Tiribilli, F. Chiti, Searching for conditions to form stableprotein oligomers with amyloid-like characteristics: the unexplored basic pH,Biochim. Biophys. Acta 1804 (1) (2010) 223–234.

680 M.S. Borana et al. / Biochimica et Biophysica Acta 1844 (2014) 670–680

[19] F.T. Chan, G.S. Kaminski Schierle, J.R. Kumita, C.W. Bertoncini, C.M. Dobson, C.F.Kaminski, Protein amyloids develop an intrinsic fluorescence signature duringaggregation, Analyst 138 (7) (2013) 2156–2162.

[20] J.A.Morrow,M.L. Segall, S. Lund-Katz,M.C. Phillips,M. Knapp, B. Rupp, K.H.Weisgraber,Differences in stability among the human apolipoprotein E isoforms determined by theamino-terminal domain, Biochemistry 39 (38) (2000) 11657–11666.

[21] D.M. Ridgley, J.R. Barone, Evolution of the amyloid fiber over multiple length scales,ACS Nano 7 (2) (2013) 1006–1015.

[22] B. Ahmad, M.Z. Ahmed, S.K. Haq, R.H. Khan, Guanidine hydrochloride denaturationof human serum albumin originates by local unfolding of some stable loops indomain III, Biochim. Biophys. Acta 1750 (1) (2005) 93–102.

[23] N.A. Morjana, B.J. McKeone, H.F. Gilbert, Guanidine hydrochloride stabilization of apartially unfolded intermediate during the reversible denaturation of protein disul-fide isomerase, Proc. Natl. Acad. Sci. U. S. A. 90 (6) (1993) 2107–2111.

[24] A. Varlan, M. Hillebrand, Bovine and human serum albumin interactions with3-carboxyphenoxathiin studied by fluorescence and circular dichroism spectroscopy,Molecules 15 (6) (2010) 3905–3919.

[25] D.A. Case, T.A. Darden, T.E. Cheatham I, C.L. Simmerling, J. Wang, R.E. Duke, R. Luo, R.C.Walker,W. Zhang, K.M. Merz, B. Roberts, S. Hayik, A. Roitberg, G. Seabra, J. Swails, A.W.Goetz, I. Kolossváry, K.F.Wong, F. Paesani, J. Vanicek, R.M.Wolf, J. Liu, X.Wu, S.R. Brozell,T. Steinbrecher, H. Gohlke, Q. Cai, X. Ye, J. Wang, M.-J. Hsieh, G. Cui, D.R. Roe, D.H.Mathews, M.G. Seetin, R. Salomon-Ferrer, C. Sagui, V. Babin, T. Luchko, S. Gusarov, A.Kovalenko, P.A. Kollman, AMBER 12, University of California, San Francisco, 2012.

[26] J.P. Ryckaert, G. Ciccotti, H.J.C. Berendsen, Numerical integration of the Cartesian equa-tions of motion of a system with constraints: molecular dynamics of n-alkanes,J. Comput. Phys. 23 (1977) 327.

[27] T. Darden, D. York, L. Pedersen, Particlemesh Ewald: anN·log (N) method for Ewaldsums in large systems, J. Chem. Phys. 98 (1993) 10089–10092.

[28] U. Essmann, L. Perera, M.L. Berkowitz, T. Darden, H. Lee, L.G. Pedersen, A smoothparticle mesh Ewald method, J. Chem. Phys. 103 (19) (1995) 8577–8593.

[29] B.A. Vernaglia, J. Huang, E.D. Clark, Guanidine hydrochloride can induce amyloidfibril formation from hen egg-white lysozyme, Biomacromolecules 5 (4) (2004)1362–1370.

[30] N. Rubin, E. Perugia, M. Goldschmidt, M. Fridkin, L. Addadi, Chirality of amyloidsuprastructures, J. Am. Chem. Soc. 130 (14) (2008) 4602–4603.

[31] A.K. Das, M.G. Drew, D. Haldar, A. Banerjee, The role of the disulfide bond in amyloid likefibrillogenesis in amodel peptide system, Org. Biomol. Chem. 3 (19) (2005) 3502–3507.

[32] J. Adamcik, J.M. Jung, J. Flakowski, P. De Los Rios, G. Dietler, R. Mezzenga, Under-standing amyloid aggregation by statistical analysis of atomic force microscopyimages, Nat. Nanotechnol. 5 (6) (2010) 423–428.

[33] P.E. Mason, C.E. Dempsey, G.W. Neilson, S.R. Kline, J.W. Brady, Preferential interac-tions of guanidinum ions with aromatic groups over aliphatic groups, J. Am. Chem.Soc. 131 (46) (2009) 16689–16696.

[34] N. Pandey, J. Strider, W.C. Nolan, S.X. Yan, J.E. Galvin, Curcumin inhibits aggregationof α-synuclein, Acta Neuropathol. 115 (2008) 479–489.

[35] W.F. Xue, S.W. Homans, S.E. Radford, Systematic analysis of nucleation-dependentpolymerization reveals new insights into the mechanism of amyloid self-assembly,Proc. Natl. Acad. Sci. U. S. A. 105 (26) (2008) 8926–8931.

[36] Y. Yoshimura, Y. Lin, H. Yagi, Y.H. Lee, H. Kitayama, K. Sakurai, M. So, H. Ogi, H. Naiki,Y. Goto, Distinguishing crystal-like amyloid fibrils and glass-like amorphous aggre-gates from their kinetics of formation, Proc. Natl. Acad. Sci. U. S. A. 109 (36) (2012)14446–14451.

[37] I.M. Kuznetsova, K.K. Turoverov, V.N. Uversky, Use of the phase diagram method toanalyze the protein unfolding–refolding reactions: fishing out the “invisible” inter-mediates, J. Proteome Res. 3 (3) (2004) 485–494.

[38] B. Ahmad, I. Vigliotta, F. Tatini, S. Campioni, B. Mannini, J. Winkelmann, B. Tiribilli, F.Chiti, The induction of α-helical structure in partially unfolded HypF-N does notaffect its aggregation propensity, Protein Eng. Des. Sel. 24 (7) (2011) 553–563.

[39] H.C. Hung, C.L. Liu, J.T. Hsu, J.T. Horng,M.Y. Fang, S.Y.Wu, S.H. Ueng, M.Y.Wang, C.W.Yaw, M.H. Hou, Development of an anti-influenza drug screening assay targetingnucleoproteins with tryptophan fluorescence quenching, Anal. Chem. 84 (15)(2012) 6391–6399.

[40] W.R. Ware, Oxygen quenching of fluorescence in solution: an experimental study ofthe diffusion process, J. Phys. Chem. 66 (1962) 455–458.

[41] M. Vendruscolo, T.P. Knowles, C.M. Dobson, Protein solubility and protein homeo-stasis: a generic view of protein misfolding disorders, Cold Spring Harb. Perspect.Biol. 3 (12) (2011).

[42] J.C. Rochet, P.T. Lansbury Jr., Amyloid fibrillogenesis: themes and variations, Curr.Opin. Struct. Biol. 10 (1) (2000) 60–68.

[43] E. Frare, P. Polverino De Laureto, J. Zurdo, C.M. Dobson, A. Fontana, A highlyamyloidogenic region of hen lysozyme, J. Mol. Biol. 340 (2004) 1153–1165.

[44] Y. Tokunaga, Y. Sakakibara, Y. Kamada, Ki. Watanabe, Y. Sugimoto, Analysis of coreregion from egg white lysozyme forming amyloid fibrils, Int. J. Biol. Sci. 9 (2)(2013) 219–227.

[45] E. Frare, M.F. Mossuto, P. Polverino de Laureto, M. Dumoulin, C.M. Dobson, A.Fontana, Identification of the core structure of lysozyme amyloid fibrils by proteol-ysis, J. Mol. Biol. 361 (3) (2006) 551–561.

[46] K.S. Ghosh, B.K. Sahoo, S. Dasgupta, Spectrophotometric studies on the interactionbetween (−)-epigallocatechin gallate and lysozyme, Chem. Phys. Lett. 452 (1–3)(2008) 193–197.