Advertisement

The Flavonoid Baicalein Inhibits Fibrillation of α-Synuclein and Disaggregates Existing Fibrils*

Open AccessPublished:April 19, 2004DOI:https://doi.org/10.1074/jbc.M403129200
      The aggregation of α-synuclein has been implicated as a critical step in the development of Parkinson's disease. Parkinson's disease is a progressive neurodegenerative disorder caused by the loss of dopaminergic neurons from the substantia nigra; currently, no cure exists. Baicalein is a flavonoid with antioxidant properties; upon oxidation, it forms several products including quinones. We show here that low micromolar concentrations of baicalein, and especially its oxidized forms, inhibit the formation of α-synuclein fibrils. In addition, existing fibrils of α-synuclein are disaggregated by baicalein. The product of the inhibition reaction is predominantly a soluble oligomer of α-synuclein, in which the protein molecules have been covalently modified by baicalein quinone to form a Schiff base with a lysine side chain in α-synuclein. The binding of baicalein was abolished by conversion of the Tyr residues into Phe, demonstrating that Tyr is involved in the interaction of α-synuclein with baicalein. In disaggregation baicalein causes fragmentation throughout the length of the fibril. These observations suggest that baicalein and similar compounds may have potential as therapeutic leads in combating Parkinson's disease and that diets rich in flavonoids may be effective in preventing the disorder.
      Parkinson's disease (PD)
      The abbreviations used are: PD, Parkinson's disease; AFM, atomic force microscopy; ThT, thioflavin T; TRITC, tetramethylrhodamine isothiocyanate; EM, electron microscopy; HPLC, high pressure liquid chromatography; SEC, size exclusion chromatography; CV, cyclic voltammogram.
      1The abbreviations used are: PD, Parkinson's disease; AFM, atomic force microscopy; ThT, thioflavin T; TRITC, tetramethylrhodamine isothiocyanate; EM, electron microscopy; HPLC, high pressure liquid chromatography; SEC, size exclusion chromatography; CV, cyclic voltammogram.
      is a neurodegenerative disorder caused by the progressive loss of dopaminergic neurons from the substantia nigra region of the brain and is characterized by cytosolic inclusions known as Lewy bodies (
      • Forno L.S.
      ,
      • Braak H.
      • Braak E.
      ,
      • da Costa C.A.
      ,
      • Goedert M.
      • Jakes R.
      • Crowther R.A.
      • Spillantini M.G.
      ). α-Synuclein has been shown to be a major fibrillar component of Lewy bodies (
      • Spillantini M.G.
      • Schmidt M.L.
      • Lee V.M.
      • Trojanowski J.Q.
      • Jakes R.
      • Goedert M.
      ), and a variety of evidence implicates the aggregation of α-synuclein as a key step in the etiology of PD (
      • Trojanowski J.Q.
      • Lee V.M.
      ). In vitro experiments have demonstrated that α-synuclein undergoes aggregation via partially folded intermediates, which form a critical nucleus, resulting in protofibrils (7 nm wide) and mature fibrils (10 nm wide), or via a pathway of soluble oligomers, resulting in amorphous aggregates (
      • Khurana R.
      • Ionescu-Zanetti C.
      • pope M.
      • Li J.
      • Nielson L.
      • Ramirez-Alvarado M.
      • Regan L.
      • Fink A.L.
      • Carter S.A.
      ). Recently, it has been suggested that the precursors of fibrils, oligomers of α-synuclein, might be more toxic than fibrils (
      • Lashuel H.A.
      • Petre B.M.
      • Wall J.
      • Simon M.
      • Nowak R.J.
      • Walz T.
      • Lansbury Jr., P.T.
      ,
      • Volles M.J.
      • Lansbury Jr., P.T.
      ).
      Currently, no preventative therapy is available for Parkinson's disease; consequently, a compound that slows and/or prevents fibrillation and aggregation of α-synuclein could lead to a therapeutic strategy for PD. Current treatments ameliorate the motor symptoms by supplementation of the deficient neurotransmitter, dopamine. A precursor of dopamine, l-dopa, has been used in the treatment of PD since the 1960s. However, severe side effects are frequently observed within a few years of l-dopa therapy.
      Oxidative stress is thought to be an important factor PD due to the destructive effect of free radicals (
      • Butterfield D.
      • Kanski J.
      ,
      • Giasson B.I.
      • Ischiropoulos H.
      • Lee V.M.
      • Trojanowski J.Q.
      ,
      • Zhang Y.
      • Dawson V.L.
      • Dawson T.M.
      ), and enhanced fibrillation of α-synuclein due to oxidized stress has been reported (
      • Ostrerova-Golts N.
      • Petrucelli L.
      • Hardy J.
      • Lee J.M.
      • Farer M.
      • Wolozin B.
      ). Both l-dopa and dopamine are known to produce free radicals during normal metabolism (
      • Asanuma M.
      • Miyazaki I.
      • Ogawa N.
      ,
      • Graham D.G.
      ). Antioxidants, acting as free radical scavengers, have been suggested to prevent or reduce the rate of progression of this disease (
      • Prasad K.N.
      • Cole W.C.
      • Kumar B.
      ,
      • Abbott R.A.
      • Cox M.
      • Markus H.
      • Tomkins A.
      ,
      • Nie G.
      • Cao Y.
      • Zhao B.
      ,
      • Spencer J.P.
      • Jenner A.
      • Butler J.
      • Aruoma O.I.
      • Dexter D.T.
      • Jenner P.
      • Halliwell B.
      ,
      • Lange K.W.
      • Rausch W.D.
      • Gsell W.
      • Naumann M.
      • Oestreicher E.
      • Riederer P.
      ). Accordingly, a combination of l-dopa and antioxidants has been recommended in the treatment of PD.
      Flavonoids, a group of polyphenolic compounds, are important components in the human diet and medicinal plants. The intake of flavonoids is in the range of 50–800 mg/day, depending on the types of vegetables, fruits, and beverages consumed. Flavonoids have broad pharmacological activities, due to the inhibition of certain enzymes and/or antioxidant activity. Dietary flavonoids have been demonstrated as potential neuroprotective agents (
      • Joseph J.A.
      • Shukitt-Hale B.
      • Denisova N.A.
      • Bielinski D.
      • Martin A.
      • McEwen J.J.
      • Bickford P.C.
      ,
      • Schroeter H.
      • Williams R.J.
      • Matin R.
      • Iversen L.
      • Rice-Evans C.A.
      ,
      • Datla K.P.
      • Christidou M.
      • Widmer W.W.
      • Rooprai H.K.
      • Dexter D.T.
      ). For example, the consumption of flavonoid-rich blueberries or strawberries can reverse cognitive and motor behavior deficits in rats (
      • Joseph J.A.
      • Shukitt-Hale B.
      • Denisova N.A.
      • Bielinski D.
      • Martin A.
      • McEwen J.J.
      • Bickford P.C.
      ), and intake of antioxidant flavonoids is associated with a lower incidence of dementia (
      • Commenges D.
      • Scotet V.
      • Renaud S.
      • Jacqmin-Gadda H.
      • BarbergerGateau P.
      • Dartigues J.F.
      ).
      Baicalein, a typical flavonoid compound (Fig. 1), is the main component of a traditional Chinese herbal medicine Scutellaria baicalensis and has multiple biological activities including antiallergic, anticarcinogenic, and anti-HIV properties (
      • Li B.Q.
      • Fu T.
      • Gong W.H.
      • Dunlop N.
      • Kung H.
      • Yan Y.
      • Kang J.
      • Wang J.M.
      ,
      • Wu J.A.
      • Attele A.S.
      • Zhang L.
      • Yuan C.S.
      ,
      • Ikezoe T.
      • Chen S.S.
      • Heber D.
      • Taguchi H.
      • Koeffler H.P.
      ,
      • Gao Z.
      • Huang K.
      • Xu H.
      ,
      • Shieh D.E.
      • Liu L.T.
      • Lin C.C.
      ). Recent studies report that baicalein protects rat cortical neurons from amyloid β-induced toxicity by its inhibition of lipoxygenase (
      • Lebeau A.
      • Esclaire F.
      • Rostene W.
      • Pelaprat D.
      ). Natural medicines containing these compounds have been reported to have beneficial effects in treating memory loss and dementia (
      • Perry E.K.
      • Pickering A.T.
      • Wang W.W.
      • Houghton P.J.
      • Perry N.S.
      ,
      • Watanabe C.M.
      • Wolffram S.
      • Ader P.
      • Rimbach G.
      • Packer L.
      • Maguire J.J.
      • Schultz P.G.
      • Gohil K.
      ).
      Figure thumbnail gr1
      Fig. 1Chemical structures of baicalein, baicalin, baicalein quinone, and the Schiff base resulting from attack of a protein lysine on the quinone.
      In this study, we show that baicalein inhibits the fibril formation of α-synuclein. In particular, the quinone oxidation form of baicalein forms a Schiff base with a lysine in α-synuclein and prevents the progress of fibrillation by stabilizing an oligomeric form. Thioflavin T assays, light scattering, circular dichroism, electron microscopy, and in situ atomic force microscopy (AFM) were used to verify the inhibiting effects. Further, we show that baicalein disaggregates existing fibrils of α-synuclein, again leading to formation of a soluble oligomer.

      EXPERIMENTAL PROCEDURES

      Purification of Human α-Synuclein—Human wild-type α-synuclein was expressed in the Escherichia coli BL21 (DE3) cell line and purified as described previously (
      • Yamin G.
      • Uversky V.N.
      • Fink A.L.
      ).
      Materials—Baicalein, baicalin, and thioflavin T (ThT) were obtained from Sigma. TRITC was purchased from Molecular Probes, Inc. (Eugene, OR).
      Aggregation/Fibrillation Studies and ThT Assays—The flavonoids were dissolved in Me2SO to make a stock solution at a concentration of 10 mm. Lyophilized α-synuclein was dissolved in 0.001 m NaOH and adjusted to pH 7.4 prior to centrifugation at 95,000 rpm with a Beckman Airfuge ultracentrifuge to remove any aggregated material. The protein solution (1 mg/ml; 70 μm) was mixed with flavonoids to give a final concentration of 5, 20, 50, or 100 μm with a final Me2SO concentration of 1%. A control α-synuclein solution containing 1% Me2SO was also prepared. All of the solutions consisted of 20 mm Tris-HCl buffer solution at pH 7.4, 100 mm NaCl, and were stirred at 37 °C with a mini-Teflon stir bar. Aliquots of 5 μl were removed from the incubated solution and added to 1 ml of 10 μm ThT solutions in 50 mm Tris-HCl buffer (pH 8.0) as a function of time to monitor the fibrillation kinetics. ThT fluorescence was recorded at 482 nm with excitation at 450 nm and slits of 2.5 nm for both excitation and emission using a FluoroMax-3 spectrofluorometer (Jobin Yvon Horiba). The lag times for all of the incubated solutions were estimated from curve-fitting the kinetics using the equation described previously (
      • Nielsen L.
      • Khurana R.
      • Coats A.
      • Frokjaer S.
      • Brange J.
      • Vyas S.
      • Uversky V.N.
      • Fink A.L.
      ). The solution was examined for fibril formation by electron microscopy (EM) and AFM. Anaerobic conditions were obtained by flushing the incubation solution with nitrogen gas for 10 min in a sealed vial.
      Light Scattering Measurements—Light scattering of the incubated solutions (in a 220-μl cuvette) was monitored over time using the spectrofluorometer with both excitation and emission at 300 nm.
      Circular Dichroism Measurements—Far-UV CD spectra were collected on an Aviv 60DS spectrophotometer using a 0.01-cm path length cell. CD spectra were recorded with a step size of 1 nm, a bandwidth of 1.5 nm, and an averaging time of 5 s. An α-synuclein concentration of 1 mg/ml was used for all spectra, and an average of five scans after subtracting the appropriate buffer was obtained.
      Size Exclusion HPLC Measurements and Collection of Fractions—An aliquot of each sample was removed from the α-synuclein incubation at various times, and the insoluble material was removed by centrifugation for 20 min at 14,000 rpm. Sample volumes of 18 μl of supernatant were eluted from a TSK-GEL G2000SWXL size exclusion column (7.8-mm inner diameter × 30 cm) in 50 mm phosphate buffer, pH 7.0, and 100 mm Na2SO4 using a Waters 2695 separations module with a Waters 996 photodiode array detector. The HPLC system was controlled, and data were collected and analyzed by Millennium software. The column was eluted at a flow rate of 0.5 ml/min, and the absorbance of the mobile phase was monitored over the wavelength range from 220 to 450 nm with a bandwidth of 1.2 nm. The retention times were calibrated with the following protein molecular mass standards: ribonuclease A (13.6 kDa), chymotrypsinogen A (25 kDa), ovalbumin (43 kDa), albumin (bovine serum) (67 kDa), aldolase (158 kDa), and blue dextran 2000 (∼2000 kDa). The elution peaks were collected manually and concentrated ∼2-fold with an SVC 100H SpeedVac Concentrator (Savant Instruments Inc.). The concentrated protein solutions were then analyzed by AFM.
      UV-visible Spectrophotometric Titration Experiments—The binding ratio and equilibrium dissociation constants Kd for ligand-α-synuclein complexes were measured by a spectrophotometric method. For the measurement of binding ratio, samples of α-synuclein/inhibitor in the range from 0.05 to 20 were prepared, and the change in absorbance was plotted versus protein/inhibitor ratio. For estimation of dissociation constants, 18 samples with different α-synuclein concentrations ranging from 0 to 8 times Kd were prepared in 20 mm Tris-HCl buffer, pH 7.4, and 100 mm NaCl. Absorption spectra over the range from 220 to 500 nm were measured using a UV-2401 PC UV-visible recording spectrophotometer (Shimadzu, Japan) immediately following the addition of 25 μm baicalein to the samples. The binding constant, Kd, was calculated with the Benesi-Hilderbrand equation (
      • Benesi A.H.
      • Hilderbrand J H.
      ),
       CαsynCint/ΔA=Kd/Δε+(1/Δε)Cαsyn 


      where Cα-syn and Cint are concentrations of α-synuclein and inhibitors, respectively. By linearly plotting Cα-synCintA versus Cα-syn, Kd can be estimated from the slope and the intercept.
      Cyclic Voltammetry Measurements—Three electrode systems were used in these experiments. The working electrode was a glass-carbon electrode (3-mm diameter) with a saturated calomel electrode reference and Pt counter electrode. Voltammetry experiments were performed on a CH Instruments model 440 electrochemical work station in a 5-ml electrochemical cell containing 10 mm phosphate buffer (pH 7.4) in 100 mm NaCl as supporting electrolyte. A cyclic voltammogram (CV) was employed at 25 mV/s with the potential window set from –0.4 to 1.0 V. All of the electrolyte solutions were purged with N2 gas prior to the voltammetric measurements.
      Electron Microscopy Measurements—Transmission electron microscopy was used to estimate the size and structural morphology of α-synuclein. Aliquots of 5-μl sample were applied to carbon-coated pioloform copper grids and incubated for 10 min. Salts were washed out with distilled water, and samples were dried, negatively stained with 1% (w/v) uranyl acetate, and visualized on a JEOL JEM-100B transmission electron microscopy operated at 80 kV. Typical magnifications ranged from × 75,000 to 300,000. The grids were thoroughly examined to obtain an overall evaluation of the samples.
      AFM Measurements—In situ AFM images were obtained using a flow cell sample chamber with a mica bottom and collected with a PicoScan SPM microscope (Molecular Imaging, Phoenix, AZ) equipped with the MAC mode in which the magnetically coated probe oscillates near its resonant frequency under an alternating magnetic field. Probes with a 2.8 newton/m spring constant and a 75-kHz resonance frequency (Molecular Imaging) were used for the MAC mode imaging. The imaging was carried out at a scan rate of 0.5–1 line/s with 512 data points per line, at a driver current of 10 ± 5 A. The amplitude change of the probe was sufficiently low that the imaging was essentially nondestructive to the sample. Heights ranging from 1.0 to 100 nm were estimated by section analysis, and lateral sizes were calibrated with various standard gold colloid particles, using the same type of probe and similar set point for imaging. At least four regions of the mica surface were examined to verify that similar structures existed through the sample. No filter treatment was used to modify the images. For non-in situ AFM imaging, aliquots of 5 μl of sample were placed on a freshly cleaved mica substrate. After incubation for 10 min, the substrate was rinsed with water twice to remove salt and loosely bound protein and blown dry with N2. Some AFM images were collected with an Autoprobe CP Multiple AFM (Park Scientific) in tapping mode using silicon cantilevers with a spring constant of 50 newtons/m and a resonance frequency of 290–350 kHz.
      Mass Spectrometry—Mass spectra were obtained using a MicroMass Quattro II electrospray instrument. Samples for mass spectrometry analysis were prepared by diluting 2 μl of protein solution in 200 μl of 50% acetonitrile, 50% HCl, pH 2.0, mixture and introduced via a Harvard Apparatus (Holliston, MA) syringe pump at a flow rate of 6 μl/min. The source temperature was set to 50 °C, and the capillary voltage was 3.0 kV.
      Labeling α-Synuclein with TRITC—TRITC is an amine-reactive fluorescent label, forming thiourea upon reaction with the amine side chain of lysine residues. The absorption maximum is 544 nm, and the emission maximum 571 nm. A molar excess of TRITC (dissolved in N,N-dimethylformamide) was added slowly to the protein solution while stirring. The dye was allowed to react with α-synuclein for l h in 0.1 m sodium bicarbonate buffer, pH 9.0, at room temperature. The reaction was stopped by adding 0.1 ml of freshly prepared 1.5 m hydroxylamine, pH 8.5, to the reaction solution. The labeled protein was separated from free dye by dialysis against 10 mm phosphate buffer (pH 8.0) for 48 h. Labeling was confirmed by UV-visible spectra. The protein concentration was estimated by Lowry assay. The fibrillation of dye-labeled protein was monitored by the ThT assay.

      RESULTS

      Baicalein Binds Tightly to α-Synuclein—The chemical structures of baicalein and baicalin are shown in Fig. 1. The binding of baicalein to α-synuclein results in changes in the absorbance spectrum of baicalein, in particular a substantial red shift in λmax, from 320 to 360 nm. Thus, the affinity of α-synuclein for baicalein was measured by spectral titration (Fig. 2A), from which a Kd of 500 nm was determined. The protein-ligand binding ratio was measured at two wavelengths; only one binding site for α-synuclein was observed. Since α-synuclein is intrinsically unstructured (i.e. natively unfolded), it is likely that baicalein binding is associated with some structural change in the protein. This was confirmed by circular dichroism experiments (see below).
      Figure thumbnail gr2
      Fig. 2UV-visible changes in baicalein in response to binding to α-synuclein (A) or autoxidation (B).A, 25 μm baicalein was titrated with increasing concentrations of α-synuclein (1.5, 3.0, 4.5, 7.5, 15, 22.5, 30, and 45 μg/ml). Absorbance (Abs) change is indicated by the arrow direction when α-synuclein was added to the solution. B, 50 μm baicalein was incubated under standard conditions used to generate α-synuclein fibrils (pH 7.4, 100 mm NaCl, 37 °C, with agitation). Spectra were collected at 0, 0.5, 3, 11, 23, and 45 h intervals; as oxidation proceeds, the absorbance decreases. The solid line represents the spectrum at 0 h. The inset shows the decrease in absorbance at 330 nm as a function of increasing time.
      α-Synuclein Fibril Formation Is Inhibited by Baicalein and Baicalin—The kinetics of fibril formation were monitored by the increase in ThT fluorescence associated with dye binding to fibrils (
      • Naiki H.
      • Higuchi K.
      • Hosokawa M.
      • Takeda T.
      ); upon binding to α-synuclein fibrils, ThT fluorescence is typically enhanced by 2 orders of magnitude. When monitored by ThT fluorescence in the absence of baicalein, the incubation of α-synuclein (70 μm) at pH 7.5 and 37 °C shows a gradual rise with a sigmoidal shape, with a lag time of ∼18 h (Fig. 3), consistent with a nucleated polymerization mechanism. In the presence of baicalein (5–100 μm), α-synuclein fibrillation was inhibited, as indicated by the negligible or reduced ThT fluorescence over the time course (Fig. 3A). With 50 or 100 μm baicalein, no fibrillation was observed for >70 h, compared with the 18-h lag time in its absence.
      Figure thumbnail gr3
      Fig. 3Baicalein and baicalin inhibit fibrillation of α-synuclein.A, inhibition by baicalein; B, inhibition by baicalin; C, inhibition by preincubated (16 h) baicalein. α-Synuclein (70 μm) was incubated in 20 mm Tris-HCl buffer, pH 7.4, 100 mm NaCl with stirring at 37 °C in the presence of 0 (open circles), 5 μm (inverted triangles), 20 μm (squares), 50 μm (triangles), and 100 μm (diamonds). Fibril formation was monitored by the increases in ThT fluorescence.
      Baicalin has the same backbone structure as baicalein but with an additional sugar ring (Fig. 1). As with baicalein, fibril formation of α-synuclein was markedly inhibited when incubated with baicalin (Fig. 3B).
      The aggregation of α-synuclein incubated with baicalein and baicalin was also monitored by static light scattering, in order to confirm the inhibition of fibrillation observed with the ThT assay. The kinetics of α-synuclein aggregation monitored by light scattering are shown in Fig. 4 and are consistent with the data from the ThT assays. In the presence of baicalein or baicalin, the dose-dependent decrease in scattering indicates that aggregation is inhibited by baicalein and baicalin. However, even at higher concentrations of baicalein or baicalin, the light scattering signal does not disappear, suggesting the formation of nonfibrillar aggregates, probably large soluble oligomers.
      Figure thumbnail gr4
      Fig. 4Inhibition of α-synuclein aggregation by baicalein (A) and baicalin (B). The kinetics of aggregation were monitored by static light scattering. The experimental conditions were the same as those in . Inhibitor concentrations were 0 (open circles), 5 μm (plus signs), 20 μm (inverted triangles), 50 μm (triangles), and 100 μm (squares).
      Additional confirmation of the inhibition was obtained from centrifugation and SDS gels, which showed essentially all of the protein in the supernatant at the end of the incubation in the presence of 100 μm baicalein, compared with essentially all of the protein in the pellet in the control, lacking baicalein. The SDS gels of both the supernatant and dissolved pellet showed only monomeric α-synuclein, indicating that under the experimental conditions, no detectable intermolecular cross-linking occurred.
      Oxidation of Baicalein Increases Its Inhibitory Potency— Since flavonoids are susceptible to oxidation, we examined the stability of baicalein under the conditions used to form α-synuclein fibrils. Baicalein oxidation was monitored by change in absorbance (Fig. 2B), which demonstrated that it underwent oxidization under the incubation conditions used (pH 7.5, 37 °C). The half-life for loss of baicalein was 15.2 h. Baicalin showed changes in its absorbance spectrum upon incubation at 37 °C similar to those seen with baicalein, indicating a similar rate of oxidation, t½ = 14.3 h. In order to determine if the active inhibitory species was baicalein or its oxidation products, we preincubated baicalein for 12–24 h prior to the addition to α-synuclein incubation solutions. Aged baicalein required lower concentrations to effect similar inhibition of fibril formation (Fig. 3C). This observation suggests that the oxidized form of baicalein may play an important role in the inhibition of fibrillation. When the incubation of α-synuclein was monitored by light scattering, similar results were obtained with aged baicalein (preincubated for 12 or 24 h) as with those of freshly prepared baicalein (data not shown).
      To address the question of the role of flavonoid oxidation in inhibiting α-synuclein fibrillation, the incubations were carried out under anaerobic conditions. The data (Fig. 5) indicate that in the absence of oxygen there is much less inhibition observed in the presence of baicalein, demonstrating that oxidized products are the major inhibitors of fibrillation, although the parent compound does have some inhibitory effects.
      Figure thumbnail gr5
      Fig. 5Baicalein is less effective as an inhibitor of fibrillation under anaerobic conditions. Fibrillation of α-synuclein was monitored under the same conditions as in , except under anaerobic conditions. The presence of 100 μm baicalein (inverted triangles) led to less inhibition than observed under comparable aerobic conditions () for the same concentration of baicalein. The control, without baicalein, is shown by the open circles.
      Both freshly prepared and aged baicalein alone at a concentration of 100 μm showed no effect on ThT fluorescence itself (i.e. baicalein does not interfere with the ThT assay for fibrils).
      Baicalein Affects α-Synuclein Nucleation but Not Fibril Elongation—The inhibition of α-synuclein fibril formation by baicalein may involve two broadly different mechanisms, either inhibition of nucleus formation or inhibition of the growth/extension of fibrils. To resolve this issue, we initially added baicalein at various times during the incubation of α-synuclein. As shown in Fig. 6, we found that baicalein stopped the progress of fibrillation regardless of the time it was added, namely at 0, 6, 12, 21, 25, and 45 h. Thus, baicalein rapidly prevented further fibril growth whenever it was added to the solution. Further, the subsequent decrease in ThT signal suggests that the addition of baicalein led to disaggregation of the fibrils that were present (see below). However, these results do not allow us to determine which stage of fibrillation is inhibited by baicalein.
      Figure thumbnail gr6
      Fig. 6Baicalein can reverse α-synuclein fibrillation. The addition of baicalein (50 μm) to incubating α-synuclein (56 μm) not only stopped further fibrillation (monitored by ThT fluorescence) but also caused a decrease in the ThT signal, due to the disaggregation of existing fibrils (see “Results”). baicalein was added at 0 (inverted triangle), 6 (square), 12 (diamond), 21 (filled triangle), 25 (hexagon), and 45 (filled circle) h after incubation was initiated. Conditions were pH 7.4, 100 mm NaCl, and 37 °C.
      To address this question more directly, we used seeding experiments. The addition of seed fibrils to α-synuclein solutions eliminates the lag observed when starting with soluble monomeric α-synuclein. Thus, if baicalein affects the rate of fibril growth, it would be apparent in such experiments. When α-synuclein was seeded in the presence of baicalein, there was no difference in rate of fibril growth in comparison with the control sample lacking baicalein. This demonstrates that there is no effect of baicalein (either fresh or aged) on the rate of fibril elongation (Fig. 7), which means the inhibitory effect must be on or prior to nucleation. This was confirmed in subsequent experiments that show the build-up of an oligomer (see below). The data in Fig. 7 show that after the initial growth of the seeds, the ThT signal declines; we attribute this to disaggregation of the fibrils, as discussed below.
      Figure thumbnail gr7
      Fig. 7The addition of seed fibrils abolishes the lag in α-synuclein fibrillation, and the presence of baicalein has no effect on this. α-Synuclein (70 μm) was incubated at pH 7.4, 100 mm NaCl, 37 °C and seeds (sonicated preformed α-synuclein fibrils) were added at zero time in the presence of 0 (filled circles) or 100 μm aged (inverted open triangles), 100 μm fresh (filled squares), 200 μm aged (inverted filled triangles), or 200 μm fresh (open circles) baicalein. The aged baicalein had been preincubated for 16 h. Conditions were pH 7.4, 100 mm NaCl, and 37 °C, and the reaction was monitored with ThT fluorescence.
      Baicalein Stabilizes a Partially Folded Conformation of α-Synuclein—Circular dichroism spectroscopy was used to investigate the effect of baicalein on the time-dependent changes in the secondary structure of α-synuclein. Both baicalein-treated (50 μm) and control solutions containing 70 μm α-synuclein were prepared in 20 mm Tris-HCl buffer, pH 7.4, and 100 mm NaCl. Freshly prepared α-synuclein gave a CD spectrum characteristic of an unfolded protein conformation (Fig. 8). The spectrum for α-synuclein incubated for 21 h without baicalein showed a gradual change in molar ellipticity, especially at 218 nm, indicating the presence of partially folded structures with significant β-structure (Fig. 8A). In contrast, for the incubation solution containing baicalein, the CD spectrum showed an initial rapid change to the spectrum of a partially folded intermediate and then no further changes, indicating that baicalein stabilized a partially folded conformation of α-synuclein (Fig. 8B).
      Figure thumbnail gr8
      Fig. 8Baicalein induces a conformational change in α-synuclein. Far-UV circular dichroism spectra of α-synuclein were measured in the absence (A) or presence of baicalein (B) of α-synuclein (70 μm). A, spectra were collected after incubation for 0 h (solid line), 6 h (dotted line), and 21 h (dashed line). B, spectra were collected after incubation in the presence of baicalein (50 μm): α-synuclein alone (solid line) and incubation with baicalein for 0 h (dotted line), 6 h (dashed line), and 21 h (dashed and dotted line).
      Baicalein Stabilizes an Oligomer of α-Synuclein—The soluble aggregated forms of α-synuclein in the presence and absence of baicalein were further studied by size exclusion HPLC. Monomeric α-synuclein eluted at 18 min in this experiment. α-Synuclein solution was incubated at 37 °C, pH 7.4, for 3 days; at appropriate time periods, an aliquot of the solution was centrifuged and loaded on the SEC HPLC column. The elution profile of α-synuclein incubated for 6 h showed a major peak corresponding to the monomer and a much smaller peak eluting at 14 min (Fig. 9A). With increasing time, the monomer peak decreased in size and the oligomer peak area increased, suggesting that at later stages of the fibrillation, most of the soluble protein is present as large soluble aggregates (oligomers) (however, the majority of the protein was insoluble fibrils).
      Figure thumbnail gr9
      Fig. 9Baicalein stabilizes high molecular weight oligomeric species. Increasing amounts of the oligomers were formed in the presence of baicalein with increasing time and in a dose-dependent manner. A, size exclusion HPLC of α-synuclein (70 μm) after incubation for 0, 6, 25, and 53 h in the presence of 20 μm baicalein. The peaks at 18 min correspond to monomer, and peaks at 14 min correspond to oligomers. The spectra have been offset for clarity. B, time-dependent increase in the amount of soluble oligomer, measured with SEC HPLC, as a function of baicalein concentration in the incubation. Filled symbols, fresh baicalein; open symbols, baicalein preincubated for 16 h. Concentrations were 100 μm (squares), 50 μm (triangles), and 20 μm (circles). C, UV spectra of oxidized baicalein (dashed line) and baicalein-bound soluble oligomer (solid line) from SEC HPLC. Abs, absorbance.
      In the presence of 20 μm baicalein, an oligomer peak eluting at 14 min was observed immediately after the addition of baicalein to the protein solution, and its area increased over the time course of the reaction (Fig. 9A); ultimately, essentially all of the protein was present as the oligomer. The large intensity of the oligomer peak could be due to an increased amount of oligomer or to increased absorbance at 275 nm (the wavelength monitored) due to the presence of baicalein or a derivative. Spectral analysis of the peak eluting at 14 min indicated that baicalein or its breakdown products was present in the oligomer (Fig. 9B). Since the presence of baicalein in complex with α-synuclein will increase the absorbance at 275 nm, it will contribute to the increased peak intensity observed with monitoring at 275 nm.
      The effect of baicalein on oligomer formation was examined as a function of baicalein concentration, as shown in Fig. 9C. Both freshly prepared and preincubated baicalein led to formation of the soluble oligomeric aggregates. To verify that the high molecular weight peaks are produced by protein aggregates, and not oligomers of the flavonoids, we did two control experiments. First, we loaded preincubated baicalein onto the HPLC column, and no peaks were observed at 14 min. Second, we mixed the incubation sample with 0.1 mm Coomassie Blue and monitored the elution profile at 520 nm, which is the adsorption wavelength for Coomassie Blue-protein complex. Baicalein does not contribute to the optical adsorption at this wavelength. The peak with retention time at 14 min in Fig. 9A is consistent with the peak at the same retention time in the presence of Coomassie Blue, confirming that it corresponds to high molecular weight protein.
      Baicalein Forms a Schiff Base with α-Synuclein—The absorbance spectrum of the oligomer formed by incubation of α-synuclein with baicalein (Fig. 9B) indicated that baicalein or its core structure was present, presumably covalently bound to the protein.
      To confirm this, we ran electrospray ionization mass spectrometry on the oligomer. The results showed a peak (corresponding to ∼25% of the protein) in addition to that of unmodified α-synuclein, with a molecular mass increased by 269 Da (Fig. 10). There are two likely mechanisms to account for the covalent addition of baicalein to α-synuclein. Both involve formation of baicalein quinone via spontaneous oxidation by oxygen. The quinone can react with protein amines either via Michael addition or through formation of a Schiff base (imine) (Fig. 1). The observed molecular mass of the α-synuclein-baicalein adduct is consistent with Schiff base formation and not Michael addition. No other significant peaks were observed in the mass spectrometry, indicating that only a single modification by baicalein occurred to α-synuclein.
      Figure thumbnail gr10
      Fig. 10Electrospray ionization mass spectrum of α-synuclein after incubation with baicalein. Unmodified α-synuclein appears at a molecular mass of 14,460; the peak at 14,729 Da is attributed to the baicalein quinone adduct (see “Results”).
      AFM and EM Measurements Show That α-Synuclein Oligomers Are Stabilized by Baicalein—Confirmation that baicalein inhibited α-synuclein fibrillation was also obtained from EM and AFM images. In the absence of baicalein, typical fibrillar structures were detected (Fig. 11, a and d). α-Synuclein was incubated with fresh and aged baicalein, and aliquots of samples were removed for EM measurements over time. Individual oligomers of 15–20-nm diameter were initially formed at early times, and then numerous oligomer units associated with each other, resulting in amorphous protein aggregates after incubation for 36 h (Fig. 11b). Very little fibrillar material was observed in comparison with the incubation without baicalein; no difference between the effects of aged and fresh baicalein was observed.
      Figure thumbnail gr11
      Fig. 11Electron microscopy (a–c) and AFM (d–h) images of baicalein-inhibited and disaggregated α-synuclein.a and d, fibrils of α-synuclein grown in the absence of baicalein. b, nonfibrillar deposits after 36 h of α-synuclein (70 μm) incubation in the presence of baicalein (20 μm) for 36 h. c, loss of α-synuclein fibrils (compare with a) after treating fibrils with baicalein (100 μm) for 6 h. e, oligomers formed after incubation of α-synuclein (70 μm) with 20 μm baicalein for 18 h; individual baicalein-stabilized spherical and annular oligomers after separation by SEC HPLC are shown in f and g. f corresponds to the late fraction, and g corresponds to the early fraction of the 14-min peak. h, the disaggregating effect of baicalein (100 μm) after a 6-h incubation with α-synuclein fibrils (compare with d). The scale bar for a–d and f is 200 nm. e and g are 500 nm2, and h is 6 μm2.
      Tapping mode atomic force microscopy permits imaging at higher resolution than EM. In one type of experiment, α-synuclein was incubated with 20 μm baicalein for 18 and 36 h; SEC HPLC analysis of these samples showed over 90% of the protein in the form of oligomers by 18 h, at which time there was no detectable monomer. The AFM images of samples of the incubation mixture show globular oligomers with uniform size in the range of 3.8–5.6 nm in height and 14–19 nm in width (Fig. 11e), consistent with the sizes measured by EM. After an additional 18 h of treatment, these oligomers associated each other. Some of them formed larger globular aggregates, which became aligned, but did not form fibrils. In another type of experiment α-synuclein was treated with same amount of baicalein for 18 h, and the oligomers were isolated by HPLC. The early fraction and late fraction of the oligomeric peak were collected separately after elution from the TSK-GEL G2000SWXL column. The fractions was concentrated and then measured with Mac-mode AFM. A torus-like structure was observed, 3.5–5.4 nm in height, 20–23 nm in outer diameter, and 11–14 nm in inner diameter in the late eluting fraction (Fig. 11f). These annular oligomers represent a small fraction of the total oligomeric fraction. Globular oligomers with sizes in the range from 32 to 45 nm were observed in the early eluting fraction (Fig. 11g). Thus, at least three different types of soluble oligomers are observed during the inhibition of α-synuclein aggregation by baicalein.
      α-Synuclein was incubated in the presence of baicalein for 1 month, and no fibrils were formed, which demonstrates the stability of the oligomers and shows that they do not convert to fibrils. Clearly, therefore, the baicalein-stabilized oligomers are not on the direct pathway to fibrillation. The concentration of baicalein required to inhibit fibrillation is much lower than the concentration of α-synuclein monomer. This is consistent with baicalein stabilizing an oligomer of α-synuclein in a nonstoichiometric manner (i.e. one baicalein molecule for several α-synuclein molecules).
      Cyclic Voltammograms of α-Synuclein Show That Baicalein Binds near Tyrosine—α-Synuclein was not found to be electrochemically active on a glassy carbon electrode with protein concentrations up to 1 mg/ml at a potential range of –0.2–1.0 V (versus saturated calomel electrode). The CV of baicalein at a glassy carbon electrode is shown in Fig. 12A. The CV indicates that the oxidation of baicalein is irreversible. The oxidation of baicalein proceeds via two electron transfer steps, since its CV shows two anodic (–0.07 and 0.45 V) and one cathodic wave (–0.05 V). The oxidation peaks are believed to be due to the oxidation of the phenolic hydroxyl groups and the pyranoid ring. The first oxidation peak at –0.07 V in Fig. 12A corresponds to the oxidation of two phenol groups to quinone. The anodic peak potential, Epa, was –0.07 V.
      Figure thumbnail gr12
      Fig. 12Electrocatalytic oxidation of baicalein is mediated by α-synuclein and tyrosine. Cyclic voltammograms of 500 μm baicalein in the absence or presence of 6, 12, 18, 24, and 30 μg/mg α-synuclein (A) or 5, 10, 15, 20, and 25 μm tyrosine (B). Supporting electrolyte was 10 mm phosphate buffer (pH 7.4) and 100 mm NaCl. Scan rate was 25 mV/s.
      α-Synuclein was titrated into a baicalein solution, and CVs were collected. The addition of α-synuclein to baicalein and subsequent scanning over the same potential range produced an increase in the peak current for oxidation without introducing a peak potential, but not for the reduction current (Fig. 12). This suggests the absence of a direct redox reaction between baicalein and α-synuclein and rather that the concentration of baicalein increased at the electrode surface due to binding to α-synuclein. No change in peak potential and current was observed as the solution was degassed for 10 min, indicating that the electrocatalysis was not mediated by oxygen in solution. No change in peak potential and current of the baicalein solution was induced by the addition of EDTA up to 100 μm, proof that the electrocatalysis was not due to contaminating metal ions, such as iron and copper.
      The interaction of an electroactive small molecule with a nonelectroactive biomacromolecule may result in an increase of the voltammetric peak. Proteins may contain a variety of electroactive moieties including disulfides, thiols, and metals ions. In α-synuclein, tyrosine and methionine are the most likely residues to mediate the oxidation of baicalein. Tyrosine itself has a very poor electrochemical response, but the signal can be greatly enhanced by a mediator, usually an active oxidant. To address the question of which amino acid residue is responsible for the electrocatalytic oxidation of α-synuclein in the presence of baicalein, tyrosine and methionine were added to baicalein solution. CVs were collected under the same condition as with α-synuclein. Direct electrooxidation of tyrosine alone gave a poorly behaved irreversible voltammogram with an anodic peak located at 0.75 V. However, when tyrosine was added to baicalein solution, the anodic peak current of baicalein was dramatically enhanced but not the cationic peak (Fig. 12B), and no peak shift was observed, the same results as observed with α-synuclein. These results suggest that both α-synuclein and tyrosine interact with baicalein. No change in potential and current was observed by the addition of methionine to baicalein.
      We propose that baicalein was first oxidized at a lower oxidation potential than tyrosine, generating a phenolic hydroxyl free radical in situ and further oxidizing tyrosine to the corresponding phenoxyl radical via electron transfer. The phenolic hydroxyl free radical was reduced to the parent baicalein, forming a redox cycle. The phenoxyl radical of tyrosine can be further oxidized to dopaquinone and/or final polymerized products. We suggest that α-synuclein or tyrosine was adsorbed on the electrode and consequently increased the surface catalyst concentration, resulting in the distinct catalytic electrochemical response at much lower concentrations. Therefore, we conclude that baicalein mediates the electrocatalysis of tyrosine residues in α-synuclein and that baicalein binding to α-synuclein involves one or more Tyr residues.
      This was confirmed in experiments using a mutant of α-synuclein in which the four tyrosine residues were replaced with phenylalanine. Titration of α-synuclein (1.5–60 μg/ml) into 25 μm baicalein solution in the presence of 100 mm NaCl, pH 7.4, at room temperature does not induce any change in the spectrum of baicalein, in contrast to the similar titration with wild type (Fig. 2A), suggesting that the interaction is abolished in the absence of tyrosine. When this variant was incubated under standard conditions, it formed fibrils with a similar rate to that of wild-type α-synuclein. However, in the presence of 50–100 μm baicalein, no inhibition of fibrillation was observed.
      Baicalein Disaggregates Existing Fibrils of α-Synuclein— The data in Fig. 5 suggest that the addition of baicalein to solutions containing fibrils of α-synuclein might lead to their loss, based on the decrease in the ThT signal. To test this directly, we formed α-synuclein fibrils under standard conditions and then resuspended them after washing in the presence of 20, 50, or 100 μm freshly prepared and preincubated baicalein. As shown in Fig. 13, A and B, a time- and dose-dependent decrease in ThT signal were observed, suggesting loss of fibrils, compared with the control in which the absence of the baicalein led to no significant loss of ThT signal. This was confirmed with comparable light scattering experiments that indicated that baicalein induces dose-dependent disaggregation (Fig. 13C). However, the light scattering signal does not completely disappear, suggesting the conversion of fibrils into nonfibrillar aggregates, probably large soluble oligomers. AFM and EM imaging confirmed the loss of fibrils in the presence of baicalein (100 μm) over a period of time (Fig. 11, c and h). In situ AFM (i.e. AFM with the sample in aqueous solution) was used to monitor the baicalein-induced disaggregation of individual fibrils (Fig. 14). Interestingly, the fibrils did not shrink from the ends but rather broke up into fragments that disaggregated, indicating that the effect of baicalein involves internal disaggregation sites of the fibrils as well as the ends. In order to ascertain the end state of the α-synuclein molecules after baicalein-induced disaggregation, we ran the disaggregated samples on SEC HPLC (Fig. 15). These show that the monomer is the major initial product, followed by slower time-dependent increases in the amount of soluble oligomers. In addition, SDS-PAGE of the pellet and supernatant after incubating α-synuclein fibrils in the presence of 10–100 μm baicalein shows the time-dependent loss of insoluble material and increased soluble monomer (data not shown), further confirming that baicalein causes the disaggregation of α-synuclein fibrils.
      Figure thumbnail gr13
      Fig. 13Baicalein disaggregates existing fibrils of α-synuclein.A, control (filled circles) and 20 μm (circles), 50 μm (triangles), and 100 μm (diamonds) baicalein (fresh (open symbols) and preincubated (16 h) (filled symbols)). B, control (filled circles) and 50 μm (circles) and 100 μm (filled triangles) baicalein. C, control (filled circles) and 50 μm (circles) and 100 μm (filled triangles) baicalein. α-Synuclein fibril concentration was 70 μm. Note the different time scales in B and C.
      Figure thumbnail gr14
      Fig. 14In situ AFM images of α-synuclein fibrils disaggregating in the presence of baicalein.Top panels, images of α-synuclein fibrils in the same region of the sample cell taken at 30-min intervals in the presence of 200 μm baicalein. Bottom, expanded view of two adjacent fibrils disaggregating, starting at 1 h. Conditions were pH 7.5 (20 mm Tris) and room temperature in a MAC mode AFM flow cell.
      Figure thumbnail gr15
      Fig. 15SEC HPLC traces showing formation of monomer and a soluble oligomer from baicalein-disaggregated α-synuclein fibrils. α-Synuclein fibrils (70 μm) were incubated at 37 °C with 20 μm baicalein at pH 7.5. Samples of supernatant at 0, 0.5, 2, and 12 h were injected onto the SEC column, and the eluant was monitored at 275 nm. Increasing concentration of oligomer is seen with increasing time of incubation. The elution time scale is different from that in the previous SEC HPLC figures, because a different flow rate was used. The identities of the peaks eluting after the monomer are unknown. Abs, absorbance.
      Covalent Modification of α-Synuclein Leads to Inhibition of Fibrillation—Since our results indicate that covalent modification, via Schiff base formation, leads to inhibition of α-synuclein fibrillation, we covalently modified Lys residues with a hydrophobic label. The success of the TRITC-labeling of α-synuclein was confirmed by the absorbance spectrum, which demonstrated the presence of the rhodamine derivative with peaks at 515 and 550 nm; the relative intensity of the former increased with increasing number of labels. We estimate an average of one label per α-synuclein. This dye was chosen for labeling, since it does not interfere with the ThT assay for fibril formation. When TRITC-labeled α-synuclein was incubated under standard conditions, no increase in ThT signal was observed, suggesting that labeling of Lys prevented fibrillation. Interestingly, and in contrast to the results with baicalein, no evidence of oligomers was observed by SEC HPLC when samples from the incubation of TRITC-labeled α-synuclein were analyzed. It is possible that such oligomers were present but dissociated faster than the time for the chromatography.

      DISCUSSION

      This investigation demonstrates that baicalein and baicalin are capable of preventing α-synuclein fibril formation in vitro and, more significantly, that they can disaggregate existing fibrils. Several critical mechanistic questions arise, such as the following. What is the active inhibitory species? What is the nature of the inhibited α-synuclein? What is the mechanism of disaggregation? Our data suggest that an oxidized product (a quinone) of baicalein or baicalin is the most active species in inhibiting fibrillation and that the major end products of both inhibition and disaggregation are oligomeric and may involve covalent modification of a lysine side chain.
      The spectral titration of baicalein with α-synuclein indicates that α-synuclein binds baicalein with a submicromolar dissociation constant, and the circular dichroism data indicate that the resulting complex is associated with a small increase in secondary structure of α-synuclein. The electrochemical results demonstrate that a tyrosine residue (or residues) is involved in the observed electrocatalysis of baicalein by α-synuclein, suggesting that baicalein is bound in proximity to one or more of the four tyrosines of α-synuclein (three are located in close proximity at the C terminus). The essentiality of the tyrosine residues in α-synuclein for baicalein binding and inhibition of fibrillation was confirmed in experiments using a Tyr-free mutant. It is also possible that baicalein binds to the central hydrophobic NAC region (residues 61–92); perhaps the structure induced by baicalein involves both this hydrophobic region and some of the tyrosines forming some local structure driven by hydrophobic interactions.
      Autoxidation of Baicalein Is Required for Inhibitory Activity— Flavonoids such as baicalein undergo autoxidation in the presence of oxygen. The spontaneous oxidation of baicalein was monitored by absorbance changes. From the monoexponential decrease in absorbance at 352 nm, baicalein and baicalin had half-lives of 15.2 and 14.3 h, respectively, under the conditions of α-synuclein incubation. Thus, a significant amount of the flavonoid would be oxidized during the nucleation and early fibrillation stages of aggregation. The most likely oxidation product is the corresponding quinone (Scheme 1), since other species such as the semiquinone are much shorter lived. Such quinones are expected to be susceptible to nucleophilic attack via Michael addition or imine (Schiff-base) formation by the lysine side chains of α-synuclein, leading to covalent modification. The spectrum of α-synuclein in the oligomer stabilized by baicalein shows evidence for such covalent modification, in that the absorbance spectrum of the oligomer from α-synuclein plus baicalein shows the presence of the baicalein ring structure (Fig. 9). The fact that samples of baicalein that had been preincubated for 12–24 h were more effective inhibitors and that much less inhibition was observed by baicalein under anaerobic conditions, in which the oxidation is prevented, indicates that an oxidation product, presumably the quinone, is the major species responsible for inhibition of fibrillation.
      Although the oxidation of baicalein could generate reactive oxygen species, these will be absent in the preincubated (aged) samples and thus can be eliminated as potential inhibitors, since any reactive oxygen species generated during the oxidative decomposition of the catechols would have reacted prior to the incubation with α-synuclein (especially short lived reactive species such as hydroxyl radical and superoxide anion).
      Further evidence that the antioxidation potential of baicalein is not directly involved in the inhibition of fibrillation comes from the observation that ascorbic acid, glutathione, or tocopherols, even when used at a 10-fold higher concentration, are not as effective as baicalein in inhibiting fibrillation.
      The Inhibited Form of α-Synuclein Is Oligomeric—Several lines of evidence indicate that the presence of baicalein during fibril formation or disaggregation leads to an oligomeric end product, including the light scattering results, which are consistent with formation of a soluble aggregate. Similarly, the results of the seeding experiments reveal that the effect of baicalein is prior to fibril formation and are consistent with formation of an off-pathway oligomer. The SEC HPLC data show that in the presence of baicalein or its oxidation products, α-synuclein rapidly generates a slowly dissociating oligomer. It is the presence of this stable oligomer that is most likely responsible for the lack of fibrillation. The oligomers were also observed and characterized by EM and AFM, the latter suggesting that multiple forms of baicalein-stabilized oligomers may be present. Interestingly, significantly less than a stoichiometric amount of baicalein is required to inhibit fibrillation, suggesting that one covalently modified α-synuclein molecule is capable of forming an oligomer with several nonmodified molecules. Precedent for such “catalysis” of inhibition exists in the action of nitrated α-synuclein (
      • Yamin G.
      • Uversky V.N.
      • Fink A.L.
      ). By preferentially stabilizing an oligomer, baicalein could shift the flux of α-synuclein from the fibrillation pathway to the formation of oligomers, thus effectively leading to inhibition of fibrillation (Scheme 2). The stability of the oligomers is illustrated by the absence of fibrils for a period of at least 1 month of incubation of α-synuclein in the presence of baicalein.
      Recently, it has been suggested that protofibrils on the aggregation pathway in neurodegenerative diseases might be more toxic than fibrils (
      • Lashuel H.A.
      • Petre B.M.
      • Wall J.
      • Simon M.
      • Nowak R.J.
      • Walz T.
      • Lansbury Jr., P.T.
      ,
      • Anguiano M.
      • Nowak R.J.
      • Lansbury Jr., P.T.
      ,
      • El Agnaf O.M.
      • Nagala S.
      • Patel B.P.
      • Austen B.M.
      ,
      • Kayed R.
      • Head E.
      • Thompson J.L.
      • McIntire T.M.
      • Milton S.C.
      • Cotman C.W.
      • Glabe C.G.
      ,
      • Volles M.J.
      • Lee S.J.
      • Rochet J.C.
      • Shtilerman M.D.
      • Ding T.T.
      • Kessler J.C.
      • Lansbury Jr., P.T.
      ,
      • Chromy B.A.
      • Nowak R.J.
      • Lambert M.P.
      • Viola K.L.
      • Chang L.
      • Velasco P.T.
      • Jones B.W.
      • Fernandez S.J.
      • Lacor P.N.
      • Horowitz P.
      • Finch C.E.
      • Krafft G.A.
      • Klein W.L.
      ). Thus, understanding the relative toxicity of fibrils, protofibrils, soluble or non-soluble oligomers, and baicalein-stabilized oligomers will be critical for the development of therapeutic strategies. If the baicalein-stabilized oligomers are not toxic to human neurons, then baicalein and related compounds may have potential in treatment of PD. Since baicalein and related compounds are present in many plant foods, if the baicalein-stabilized oligomer is toxic, the consumption of flavonoid-rich foods should increase the risk of PD. However, the fact that there are many observations that flavonoids act as a neuroprotective agents and ameliorate dementia suggests that this is unlikely to be the case (
      • Joseph J.A.
      • Shukitt-Hale B.
      • Denisova N.A.
      • Bielinski D.
      • Martin A.
      • McEwen J.J.
      • Bickford P.C.
      ,
      • Schroeter H.
      • Williams R.J.
      • Matin R.
      • Iversen L.
      • Rice-Evans C.A.
      ,
      • Datla K.P.
      • Christidou M.
      • Widmer W.W.
      • Rooprai H.K.
      • Dexter D.T.
      ,
      • Commenges D.
      • Scotet V.
      • Renaud S.
      • Jacqmin-Gadda H.
      • BarbergerGateau P.
      • Dartigues J.F.
      ,
      • Lebeau A.
      • Esclaire F.
      • Rostene W.
      • Pelaprat D.
      ,
      • Perry E.K.
      • Pickering A.T.
      • Wang W.W.
      • Houghton P.J.
      • Perry N.S.
      ,
      • Watanabe C.M.
      • Wolffram S.
      • Ader P.
      • Rimbach G.
      • Packer L.
      • Maguire J.J.
      • Schultz P.G.
      • Gohil K.
      ). In fact, there are reports that elevated intake of antioxidant flavonoids is associated with a lower incidence of dementia and neurodegenerative disease (
      • Commenges D.
      • Scotet V.
      • Renaud S.
      • Jacqmin-Gadda H.
      • BarbergerGateau P.
      • Dartigues J.F.
      ,
      • Perry E.K.
      • Pickering A.T.
      • Wang W.W.
      • Houghton P.J.
      • Perry N.S.
      ,
      • Watanabe C.M.
      • Wolffram S.
      • Ader P.
      • Rimbach G.
      • Packer L.
      • Maguire J.J.
      • Schultz P.G.
      • Gohil K.
      ).
      The inhibition of α-synuclein fibril formation by these flavonoids could arise through two different general mechanisms, either inhibition of nucleus formation or inhibition of the growth/extension of fibrils. The addition of seed fibrils to α-synuclein solutions eliminates the lag observed when starting with soluble monomeric α-synuclein. This demonstrates that there is no effect of baicalein (either fresh or aged) on the rate of fibril elongation, which means that the inhibitory effect must be on or prior to nucleation. This is consistent with kinetic partitioning of the aggregation reaction to the off-pathway soluble oligomers (Scheme 2).
      Baicalein Quinone Is Covalently Incorporated into α-Synuclein—The absorbance spectrum of the stabilized soluble oligomer indicates the incorporation of baicalein or an oxidation product into the oligomer and the mass spectrometry results are consistent with formation of a Schiff base between the quinone of baicalein and a Lys residue of α-synuclein. The mass spectrometry data also indicate that only a single residue is modified, although the presence of low concentrations of additional incorporation cannot be excluded. It is also possible that intermolecular cross-linking could occur through the interaction of a single baicalein quinone molecule with the Lys side chains of two different α-synuclein molecules, although we observed no evidence for this at the relatively low concentrations of baicalein used. The results of the experiments with the TRITC-labeled α-synuclein indicate that covalently adding a relatively nonpolar group to a Lys side chain of α-synuclein prevents its fibrillation, thus demonstrating that it is the presence of the added moiety that prevents inhibition and nothing specific to the general molecular structure of baicalein (other than the quinone formation). However, the fact that fresh baicalein and baicalein under anaerobic conditions also have some inhibitory effect indicates that in addition to the inhibition due to covalent modification there are probably additional modes of inhibition by unoxidized baicalein.
      Baicalein Disaggregates α-Synuclein Fibrils—When α-synuclein fibrils were resuspended in the presence of baicalein, a time- and dose-dependent decrease in ThT signal was observed, suggesting loss of fibrils. This was confirmed with comparable light scattering experiments that indicated that baicalein induces dose-dependent disaggregation. However, the light scattering signal does not disappear, suggesting the conversion of fibrils into nonfibrillar aggregates, probably soluble oligomers. Unlike inhibition of fibrillation, which was sensitive to the oxidation state of baicalein, the disaggregation occurred significantly, even under anaerobic conditions, suggesting that the initial stages of disaggregation are different from those of inhibition. However, the end product, a mixture of monomer and soluble oligomers, is similar, although the initial product in disaggregation is the monomer, which might then oligomerize in the same manner that monomeric α-synuclein reacts with baicalein to form oligomers.
      AFM images clearly show the disaggregation of α-synuclein fibrils induced by baicalein; interestingly, these images show that the disruption of the fibrils occurs not only from the ends of the fibril, as might be expected, but also from internal regions (i.e. baicalein causes both exo- and endodisaggregation of the fibrils). This probably reflects a specific intercalation of the relatively planar baicalein, with its many hydrogen bond donor/acceptors, into the β-sheet structure of the fibril. Our observations suggest that this does not require oxidation of baicalein, unlike inhibition, further supporting the notion of different mechanisms for inhibition of fibrillation and disaggregation of fibrils.

      References

        • Forno L.S.
        J. Neuropathol. Exp. Neurol. 1996; 55: 259-272
        • Braak H.
        • Braak E.
        J. Neurol. 2000; 247: II3-II10
        • da Costa C.A.
        Curr. Top. Med. Chem. 2003; 3: 17-24
        • Goedert M.
        • Jakes R.
        • Crowther R.A.
        • Spillantini M.G.
        Parkinson's Disease: Methods and Protocols. Humana Press Inc., Totowa, NJ2001
        • Spillantini M.G.
        • Schmidt M.L.
        • Lee V.M.
        • Trojanowski J.Q.
        • Jakes R.
        • Goedert M.
        Nature. 1997; 388: 839-840
        • Trojanowski J.Q.
        • Lee V.M.
        Ann. N. Y. Acad. Sci. 2003; 991: 107-110
        • Khurana R.
        • Ionescu-Zanetti C.
        • pope M.
        • Li J.
        • Nielson L.
        • Ramirez-Alvarado M.
        • Regan L.
        • Fink A.L.
        • Carter S.A.
        Biophys. J. 2003; 85: 1135-1144
        • Lashuel H.A.
        • Petre B.M.
        • Wall J.
        • Simon M.
        • Nowak R.J.
        • Walz T.
        • Lansbury Jr., P.T.
        J. Mol. Biol. 2002; 322: 1089-1102
        • Volles M.J.
        • Lansbury Jr., P.T.
        Biochemistry. 2003; 42: 7871-7878
        • Butterfield D.
        • Kanski J.
        Mech. Ageing Dev. 2001; 122: 945-962
        • Giasson B.I.
        • Ischiropoulos H.
        • Lee V.M.
        • Trojanowski J.Q.
        Free Radic. Biol. Med. 2002; 32: 1264-1275
        • Zhang Y.
        • Dawson V.L.
        • Dawson T.M.
        Neurobiol. Dis. 2000; 7: 240-250
        • Ostrerova-Golts N.
        • Petrucelli L.
        • Hardy J.
        • Lee J.M.
        • Farer M.
        • Wolozin B.
        J. Neurosci. 2000; 20: 6048-6054
        • Asanuma M.
        • Miyazaki I.
        • Ogawa N.
        Neurotox. Res. 2003; 5: 165-176
        • Graham D.G.
        Mol. Pharmacol. 1978; 14: 633-643
        • Prasad K.N.
        • Cole W.C.
        • Kumar B.
        J. Am. Coll. Nutr. 1999; 18: 413-423
        • Abbott R.A.
        • Cox M.
        • Markus H.
        • Tomkins A.
        Eur. J. Clin. Nutr. 1992; 46: 879-884
        • Nie G.
        • Cao Y.
        • Zhao B.
        Redox. Rep. 2002; 7: 171-177
        • Spencer J.P.
        • Jenner A.
        • Butler J.
        • Aruoma O.I.
        • Dexter D.T.
        • Jenner P.
        • Halliwell B.
        Free Radic. Res. 1996; 24: 95-105
        • Lange K.W.
        • Rausch W.D.
        • Gsell W.
        • Naumann M.
        • Oestreicher E.
        • Riederer P.
        J. Neural. Transm. Suppl. 1994; 43: 183-201
        • Joseph J.A.
        • Shukitt-Hale B.
        • Denisova N.A.
        • Bielinski D.
        • Martin A.
        • McEwen J.J.
        • Bickford P.C.
        J. Neurosci. 1999; 19: 8114-8121
        • Schroeter H.
        • Williams R.J.
        • Matin R.
        • Iversen L.
        • Rice-Evans C.A.
        Free Radic. Biol. Med. 2000; 29: 1222-1233
        • Datla K.P.
        • Christidou M.
        • Widmer W.W.
        • Rooprai H.K.
        • Dexter D.T.
        Neuroreport. 2001; 12: 3871-3875
        • Commenges D.
        • Scotet V.
        • Renaud S.
        • Jacqmin-Gadda H.
        • BarbergerGateau P.
        • Dartigues J.F.
        Eur. J Epidemiol. 2000; 16: 357-363
        • Li B.Q.
        • Fu T.
        • Gong W.H.
        • Dunlop N.
        • Kung H.
        • Yan Y.
        • Kang J.
        • Wang J.M.
        Immunopharmacology. 2000; 49: 295-306
        • Wu J.A.
        • Attele A.S.
        • Zhang L.
        • Yuan C.S.
        Am. J Chin Med. 2001; 29: 69-81
        • Ikezoe T.
        • Chen S.S.
        • Heber D.
        • Taguchi H.
        • Koeffler H.P.
        Prostate. 2001; 49: 285-292
        • Gao Z.
        • Huang K.
        • Xu H.
        Pharmacol. Res. 2001; 43: 173-178
        • Shieh D.E.
        • Liu L.T.
        • Lin C.C.
        Anticancer Res. 2000; 20: 2861-2865
        • Lebeau A.
        • Esclaire F.
        • Rostene W.
        • Pelaprat D.
        Neuroreport. 2001; 12: 2199-2202
        • Perry E.K.
        • Pickering A.T.
        • Wang W.W.
        • Houghton P.J.
        • Perry N.S.
        J. Pharm. Pharmacol. 1999; 51: 527-534
        • Watanabe C.M.
        • Wolffram S.
        • Ader P.
        • Rimbach G.
        • Packer L.
        • Maguire J.J.
        • Schultz P.G.
        • Gohil K.
        Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 6577-6580
        • Yamin G.
        • Uversky V.N.
        • Fink A.L.
        FEBS Lett. 2003; 542: 147-152
        • Nielsen L.
        • Khurana R.
        • Coats A.
        • Frokjaer S.
        • Brange J.
        • Vyas S.
        • Uversky V.N.
        • Fink A.L.
        Biochemistry. 2001; 40: 6036-6046
        • Benesi A.H.
        • Hilderbrand J H.
        J. Am. Chem. Soc. 1949; 71: 2703-2707
        • Naiki H.
        • Higuchi K.
        • Hosokawa M.
        • Takeda T.
        Anal. Biochem. 1989; 177: 244-249
        • Anguiano M.
        • Nowak R.J.
        • Lansbury Jr., P.T.
        Biochemistry. 2002; 41: 11338-11343
        • El Agnaf O.M.
        • Nagala S.
        • Patel B.P.
        • Austen B.M.
        J. Mol. Biol. 2001; 310: 157-168
        • Kayed R.
        • Head E.
        • Thompson J.L.
        • McIntire T.M.
        • Milton S.C.
        • Cotman C.W.
        • Glabe C.G.
        Science. 2003; 300: 486-489
        • Volles M.J.
        • Lee S.J.
        • Rochet J.C.
        • Shtilerman M.D.
        • Ding T.T.
        • Kessler J.C.
        • Lansbury Jr., P.T.
        Biochemistry. 2001; 40: 7812-7819
        • Chromy B.A.
        • Nowak R.J.
        • Lambert M.P.
        • Viola K.L.
        • Chang L.
        • Velasco P.T.
        • Jones B.W.
        • Fernandez S.J.
        • Lacor P.N.
        • Horowitz P.
        • Finch C.E.
        • Krafft G.A.
        • Klein W.L.
        Biochemistry. 2003; 42: 12749-12760