The GST-α-syn fusion construct in pGEX-4T1 vector (kindly provided by Dr. Hyangshuk Rhim, The Catholic University College of Medicine, Seoul, Korea) was inserted into E.coli BL21bacteria for protein expression by heat shock transformation. The transformed bacteria were grown in LB medium supplemented with 0.1 mg/ml ampicillin at 37°C in an orbital shaker until the OD600 was 0.5. GST-α-syn expression was then induced by adding 0.5 mM IPTG (Sigma-Aldrich Chemie GmbH, Germany), and the culture was incubated for 2 h at 37°C. The cells were harvested by 15 min centrifugation at 9000 × g (Beckman Coulter, Avanti J-26 XP). The pellets were resuspended in lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM EDTA, 1% NP-40, 0.1% DTT), and the lysed pellets were shaken for 10 min at room temperature. The lysed pellets were then subjected to 6 freeze-thaw cycles in liquid nitrogen and a 37°C water bath. The lysates were centrifuged further at 27,000 × g for 15 min. The supernatant was kept for purification with affinity chromatography. Glutathione sepharose 4B beads (Amersham, Sweden) were centrifuged at 500 × g at 4°C for 8 min to remove the storage buffer. The beads were then washed with 20 ml cold PBS followed by spinning at 500 × g at 4°C for 10 min. The cell lysate was mixed with the washed beads and incubated for 1 h at room temperature, followed by centrifugation at 500 × g at 4°C for 8 min. The beads were then washed twice with column washing buffer (50 mM Tris-HCl, 150 mM NaCl, 10 mM EDTA, 1% Triton X-100, pH 8.0), twice with 50 mM Tris-HCl, pH 8.0 and once with 1X PBS. The washed beads were resuspended in 5 ml of 1X PBS. The GST tag, which interacts with glutathione beads, was cleaved by thrombin (1 unit/μl human plasma thrombin, Sigma). The thrombin-catalyzed reaction was incubated overnight at room temperature with continuous mixing. The mixture was then incubated for 5 min at 37°C and centrifuged for 8 min at 500 × g at 4°C. Benzamidinesepharose beads (Amersham, Sweden) were used to remove the thrombin from the solution. Pure α-syn was collected by centrifugation at 500 × g for 8 min at 4°C. The α-syn concentration was estimated using the BCA assay (Pierce Biotechnology, Rockford, IL).

Stock solutions of the tested compounds (10 mM) were prepared in DMSO. Solutions of lower concentrations were prepared by diluting the stock solutions to final concentrations of 25–100μM. The amount of DMSO in the final samples was 1%. Samples of 25μM (unless otherwise stated) α-syn in PBS were aged either alone or with phenolic acids at various molar ratios (phenolic acids to protein molar ratios of 4:1, 2:1, and 1:1). The samples were placed in 1.5 ml sterile polypropylene tubes, sealed with parafilm and incubated at 37°C for 5 days with continuous shaking at 800 rpm in a Thermomixer (Eppendorf). Samples were collected at regular time points. Thioflavin-S (Th-S) fluorescence was measured immediately, while the rest of the samples were stored at −80°C until needed for further analysis.

α-Syn fibril formation was monitored by Th-S binding. Th-S is a fluorescent dye that interacts with fibrils in a β-sheet structure. Each sample (10 μl) was mixed with 40 μl of Th-S (25μM) in PBS. Fluorescence was measured in a 384-well, non-treated, black micro-well plate (Nunc, Denmark) using a Victor X3 2030 (Perkin Elmer) microplate reader with excitation and emission wavelengths of 450 and 486 nm, respectively. To allow for background fluorescence, the fluorescence intensity of a blank well containing only PBS solution was subtracted from all readings.

Electron microscopy images were produced from α-syn aged in the presence or absence of GA. The samples (5 μL) were deposited on Formvar-coated 400 mesh copper grids, fixed briefly with 0.5% glutaraldehyde (5 μl), negatively stained with 2% uranyl acetate (Sigma-Aldrich) and examined in a Philips CM-10 TEM electron microscope.

Samples (20 ng) of α-syn incubated alone or with GA were mixed with loading sample buffer (250 mM Tris-HCl, pH 6.8, 30% glycerol, 0.02% bromophenol blue) without SDS or boiling and then separated on 1 mm 15% SDS-PAGE gels. The separated proteins were transferred to nitrocellulose membranes (0.45μm, WhatmanGmbh-Germany) at 90 V for 80 min. The membranes were boiled for 5 min in PBS and then blocked for 1 h with 5% non-fat milk prepared in PBS-Tween-20 (0.05% PBST). The membranes were incubated overnight at 4°C with the primary antibody, namely the mouse monoclonal anti-α-syn (211) that recognizes human α-syn (121–125) (Santa Cruz Biotechnology, USA), at a dilution of 1:1000. The membranes were then washed several times with PBST, followed by incubation with HRP-conjugated goat anti-mouse antibody (Dako Ltd., Ely, UK) at a dilution of 1:70,000 for 60 min at room temperature and gentle agitation. The membranes were then extensively washed for 25 min. The immunoreactive bands were visualized with a SuperSignal West FemtoChemiluminescent Substrate Kit (Pierce, Rockford, USA) according to the manufacturer's instructions.

A384-well ELISA plate (NuncMaxisorb,Nunc, Denmark) was coated with 1 μg/ml of non-biotinylated mouse monoclonal anti-α-syn antibody [mAb, 211—recognizes amino acid residues 121–125 of human α-syn (Santa Cruz Biotechnology, California, USA)]diluted in 200 mM NaHCO₃, pH 9.6 (50 μl/well) and incubated overnight at 4°C. The plate was then washed 4 times with PBST and blocked with 100 μl/well of blocking buffer (5% gelatin from cold water fish skin, 0.05% Tween 20 in 1X PBS pH 7.4) for 2 h at 37°C. After washing 4 times with PBST, 50 μl of the samples were dispensed in each well, and each sample was tested in duplicate. The plates were then incubated at 37°C for another 3 h. After washing 4 times with PBST, 50 μl of biotinylated 211 antibody diluted in blocking buffer to a concentration of 0.4 μg/ml was added and incubated at 37°C for 2 h. The wells were washed 4 times with PBST and incubated with 50 μl/well of Extravidin-Peroxidase (Sigma-Aldrich, GmbH- Germany) diluted 1:7500 in blocking buffer and incubated for 1 h at 37°C. The wells were then washed 4 times with PBST before adding SuperSignal ELISA Femto Maximum Sensitivity Substrate (Pierce Biotechnology, Rockford, USA) (50 μL/well). The chemiluminescence in relative light units was immediately measured using amicroplate reader (Perkin Elmer).

BE (2)-M17 human neuroblastoma cells were routinely cultured in Dulbecco's MEM/Nutrient Mix F-12 (1:1) (Gibco BRL, Rockville, MD) supplemented with 15% fetal bovine serum and 1% penicillin-streptomycin (100 U/ml penicillin, 100 mg/ml streptomycin). The cells were maintained at 37°C in a humidified incubator with 5% CO2/95% air.

To evaluate the cell viability of BE (2)-M17 human neuroblastoma cells treated with α-syn aggregates, the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction assay was performed as described previously (Mosmann, 1983). Briefly, BE (2)-M17 cells in DMEM medium were plated at a density of 15,000 cells/100 μl/well in a 96-well plate. After 24 h, the medium was replaced with 200 μ l of OPTI-MEM (Gibco BRL) serum-free medium containing either α-syn aged in the presence or absence of GA or pre-aged α-syn incubated with different concentrations of GA. Samples were diluted in OPTI-MEM to obtain the desired concentration. Cells were then returned into the incubator and incubated for 48 h. A total of 20 μ l of MTT (6 mg/ml) in PBS was added to each well, and the plate was incubated at 37°C for 4.5 h. The medium-MTT solution was removed, cell lysis buffer (100 μ l/well; 15% (w/v) SDS/50% (v/v) N,N-dimethylformamide, pH 4.7) was added, and the plate was incubated overnight at 37°C. Absorbance values at 590 nm were determined using a plate reader (Perkin Elmer).

Congo red (20μM) was dissolvedin PBS (pH 7.4) and filtered through a 0.45μm filter. Samples of α-syn (5μM), aged alone or with GA at different molar ratios, were mixed with Congo red (final concentration 5μM), and the reaction samples were thoroughly mixed. The UV absorbance spectrum was recorded from 400 to 600 nm in a spectrophotometer (DU-800, Beckman-Coulter) using 10 mm quartz cuvettes (Hellma Analytics-Germany). Monomeric α-syn and Congo red alone were used as negative controls.

Recombinant α-syn dissolved in sterilized PBS (pH 7.4) was aggregated at a concentration of 25μM as indicated above. Briefly, the α-syn samples were placed in 1.5 ml sterile polypropylene tubes sealed with parafilm and incubated at 37°C for 5 days, with continuous shaking at 800 rpm in a Thermomixer (Eppendorf). The resulting aggregated α-syn was incubated either alone or with GA at various molar ratios (GA: α-syn molar ratios 6:1, 4:1, and 2:1). It should be noted that for the purpose of the experiment, the concentration of α-syn was assumed to be the same as that of the fresh α-syn. The samples were incubated at 37°C for 3 days on a thermomixer with continuous shaking at 800 rpm. Samples were collected at regular time points, and Th-S fluorescence was measured immediately.

The aggregation of monomeric α-syn with or without seeding was carried out as described elsewhere (Di Giovanni et al., 2010). Mature α-syn fibrils were fragmented by sonication to obtain short fibrils, which were employed as ‘seeds’. Briefly, 100μM monomeric α-syn was seeded with 2μMseeds and incubated in the presence or absence of GA (10μM and 50μM) at 37°C for 6 h with continuous shaking. Fibrillation was monitored by Th-S binding as described above.

Recombinant α-syn dissolved in sterilized PBS (pH 7.4) was aggregated at a concentration of 100μM. Sample was placed in 1.5 ml sterile polypropylene tubes, sealed with parafilm and incubated at 37°C for 2 days, with continuous shaking at 800 rpm in a Thermomixer (Eppendorf).

Samples of α-syn oligomers, α-syn: GA oligomers and α-syn monomers (15 ng), were separated in 3–1% Bis-Tris Native PAGE gel (Novox, Life technologies) according to the manufacturer's instructions. The separated proteins were transferred to PVDF membrane (0.45μm, Thermo Scientific) at 100 V for 120 min, after soaking the methanol for 30 s, and followed by soaking in 1x methanol free transfer buffer for 10 min. The membrane then blocked for 1 h with 5% non-fat milk prepared in PBS-Tween-20 (0.05% PBST). The membrane were incubated overnight at 4°C with the primary antibody, namely the mouse monoclonal anti-α-syn (211) that recognizes human α-syn (121–125) (Santa Cruz Biotechnology, USA), at a dilution of 1:1000. The membranes were then washed several times with PBST, followed by incubation with HRP-conjugated goat anti-mouse antibody (Dako Ltd., Ely, UK) at a dilution of 1:120,000 for 60 min at room temperature and gentle agitation. The membranes were then extensively washed for 25 min. The immunoreactive bands were visualized with a SuperSignal West FemtoChemiluminescent Substrate Kit (Pierce, Rockford, USA) according to the manufacturer's instructions.

SEC was carried out using an AKTA FPLC system (GE Healthcare-Sweden) and a superdex 200 column at 4°C, in order to separate the oligomers generated from the aggregation of α-syn with GA (GA: α-syn molar ratio of 4:1). Monomeric α-syn at a concentration of 100μM was aggregated in the presence of GA for 5 days as described above. At the end of the aggregation process, the sample was centrifuged for 45 min at 14,000 × g at 4°C generating a supernatant free from insoluble material. Prior to injecting 80% of the generated supernatant, the column was thoroughly equilibrated with SEC running buffer (1 × PBS, pH 7.4) and the flow rate was set to 0.1 ml/min (0.5 ml/fraction). α-Syn elution was monitored at absorbance wavelengths of 215, 254, and 280 nm. To determine the elution time of monomeric α-syn, molecular weight standards (ferritin 440 kDa, aldolase 171 kDa, abmumin 68 kDa and chymotrypsinogen A 25 kDa) and monomeric α-syn were co-injected into the column and eluted at the same conditions mentioned above. The fractions eluting between 7-9 ml CV were combined and labeled as oligomers (sample P1), whereas the fractions eluting in the 13–15 ml CV were combined and labeled as monomers (sample P2). The P1 and P2 fractions were further characterized by western blotting and TEM.

The P1 and P2 samples, representing the oligomeric and monomeric fractions of SEC, respectively, were concentrated using a speed vac (CentriVap, Labconco). Their protein content concentration was estimated by the BCA assay. The UV absorbance spectrum was recorded from 200 to 600 nm in a spectrophotometer (DU-800, Beckman-Coulter) using 10 mm quartz cuvettes (Hellma Analytics-Germany) and employing equal concentrations of both P1 and P2. Fresh monomeric α-syn was used as negative control.

For NMR studies, recombinant 15N-labeled α-syn was expressed and purified as previously described (Eliezer et al., 2001; Bussell and Eliezer, 2003), resuspended in PBS at pH 6.5. Two-dimensional proton-nitrogen correlation spectra were acquired for α-syn at 200μM concentration in the absence of GA and in the presence of increasing GA: α-syn stoichiometries of 0.5:1, 1:1, 2:1, 5:1, and 10:1. Data were collected on a Varian 600 MHz Unity Inova spectrometer equipped with a cold probe.

To investigate the effect of GA on α-syn fibrillation, 25μM of α-syn was incubated with GA at varying molar ratios of GA: α-syn of 4:1, 2:1, and 1:1 for a period of 6 days. GA inhibited the formation of α-syn fibrils in a concentration-dependent manner as indicated by the Th-S assay (Figure 1A). Taking into account the lack of Th-S signal at the 4:1 ratio, GA exhibited an excellent inhibitory effect on α-syn fibrillation, inhibiting it completely during the 6-day incubation (Figure 1A). At a 2:1 ratio, GA also suppressed the formation of fibrils to a great extent, and the Th-S signal was detected only after 4 days of incubation, while after 6 days, GA reduced the Th-S counts to approximately one fifth of the control counts (Figure 1A). Even at a 1:1 ratio, GA hindered the fibrillation of α-syn but to a smaller extent than the other ratios (Figure 1A). The anti-fibrillogenic activity of GA was further assessed by the Congo red binding assay. Congo red (CR) is a dye with high affinity for amyloid fibrils. The absorption maximum of CR alone (5μM) was 490 nm, while when incubated with α-syn amyloid fibrils, the absorption maximum shifted to 508 nm (Figure 1B). This shift in the peak absorption wavelength represents the binding of CR to the β-sheet-rich fibrils. However, this prominent shift was not observed for the fresh α-syn sample or when α-syn was aged in the presence of GA (Figure 1B), indicating that GA inhibited the formation of structures with a β-sheet conformation. In fact, GA blocked the formation of β-sheets in a dose-dependent manner, similar to the Th-S results. It was observed that the lower the GA concentration, the more prominent the peak absorbance wavelengths shift was (485, 493, and 500 nm for 100, 50, and 25μM GA, respectively). Electron microscopy images also confirmed that GA not only affected the extent of fibrillation in a dose-dependent fashion but also the morphology of α-syn fibrils. Indeed, instead of forming dense meshes of long fibrils (longer than 1μm) as observed in aged α-syn alone (Figure 1C), α-syn aged in the presence of 100μM (Figure 1D) and 50μM of GA (Figure 1E) generated thin, sheared fibrils that were approximately 0.1–0.2μm in length. α-Syn aged in the presence of 25μM GA (Figure 1F) formed longer fibrils (approximately 0.5μm long), which were arranged in dense networks.

To evaluate the effect of GA on α-syn oligomer formation, samples of α-syn aged alone or in the presence of GA at different molar ratios were assessed for their oligomeric content by oligomer-specific ELISA and immunoblotting. For the detection of α-syn oligomeric species in the samples, we employed a novel ELISA developed in our laboratory (El-Agnaf et al., 2006), which specifically recognizes the oligomeric species present in the samples. The ELISA results (Figure 2A) indicated that GA could only inhibit the formation of α-syn oligomers when employed at the highest concentration, namely 100μM. Interestingly, at lower concentrations, GA appeared to enhance α-syn oligomerization compared to the control (Figure 2A). The samples of α-syn aged in the presence of 25 and 50μM GA exhibited gradually increasing oligomeric content, which was much higher than the control (Figure 2A).

Similarly, immunoblotting of the same samples (Figure 2B) indicated a decrease in the oligomeric species when α-syn was aged with 100μM GA (Figure 2B), while there was a characteristic increase in the band corresponding to dimeric α-syn (Figure 2B). The same sample was also characterized by the presence of strong monomeric and trimeric (approximately 50 kDa) bands and the absence of the bands corresponding to larger α-syn aggregates with high molecular weight. However, in the presence of 50 and 25μM GA, the bands corresponding to oligomeric species were much stronger, with a concentration-dependent decrease in the bands corresponding to monomeric, dimeric and trimeric species. Additionally, the intensity of the band corresponding to larger α-syn aggregates increased in a dose-dependent manner. These data indicate that GA at low concentrations may have the ability to stabilize α-syn oligomers.

Given that GA was shown to be such an effective inhibitor of α-syn fibrillation, we investigated whether it could also reverse fibrillation. Therefore, 25μM of preformed α-syn fibrils were incubated at 37°C in the presence of GA at molar ratios of GA: α-syn of 6:1, 4:1, and 2:1 for a period of 48 h. By measuring the Th-S fluorescence counts (Figure 3A), we estimated the fibril content at the indicated time points. At time 0, the Th-S counts were approximately 8000 for all samples (Figure 3A). The α-syn fibrils that were incubated alone continued to aggregate further, as indicated by the increase in Th-S counts (Figure 3A). However, the α-syn fibrils that were incubated in the presence of GA disaggregated over time as shown by the decrease in the Th-S counts (Figure 3A). It is noteworthy that after 24 h of incubation, α-syn fibrils incubated without GA gave approximately 18,000 Th-S counts, while the fibrils incubated with all tested concentrations of GA produced less than 2000 Th-S counts (Figure 3A). Thus, GA disaggregated preformed α-syn fibrils in a dose-dependent fashion.

After 48 h of incubation, the samples were also tested for their CR binding (Figure 3B). As expected, α-syn fibrils incubated in the presence of GA produced spectra with a peak absorbance wavelength that was slightly shifted compared to the controls, indicating minimal β-sheet content. In contrast, the fibrils of α-syn that were incubated alone produced a peak absorbance at 550 nm, indicating a high β-sheet content (Figure 3B). These findings were confirmed by EM (Figure 3C). All samples at time 0 were characterized by the presence of dense meshes of fibrils longer than 0.5μm. However, the α-syn fibrils incubated in the presence of GA gradually became thinner and shorter and appeared fragmented, isolated and scarce, unlike the control, which continued to demonstrate networks of long fibrils (Figure 3C). Furthermore, similar results were obtained when the experiment was performed using α-syn fibrils after reaching plateau in Th-S assay (Figure S1).

To assess whether GA disaggregation of α-syn fibrils was accompanied by a decrease in α-syn toxicity, we then evaluated the toxicity of the α-syn species that resulted from the disaggregation experiment on human neuroblastoma M17 cells (Figure 3D). The samples were diluted to final α-syn concentrations of 0.5, 1, and 5μM. Preformed α-syn fibrils incubated for 6 days in the absence of GA decreased cell viability in a dose-dependent manner. However, preformed α-syn fibrils incubated with GA for 6 days generated species that were less toxic compared to the control. This trend was most prominent for 5μM α-syn, which, when incubated in the absence of GA, induced death of 45% of the cells, but in the presence of GA, cell death was less than 25%.

It has been previously shown that the process of amyloid fibril formation follows a nucleation-dependent polymerization (Jarrett and Lansbury, 1992). According to this model, soluble species generated via the nucleation of oligomeric species (nucleation or lag time phase), which in turn polymerize (polymerization or growth phase) to generate fibrils, reaching thus a final plateau known as the equilibrium phase (Harper et al., 1999). Small aggregates or seeds have been shown to accelerate the nucleation phase of amyloid formation both in vitro and in vivo by a process known as seeding (Jarrett and Lansbury, 1993; Harper and Lansbury, 1997; Volpicelli-Daley et al., 2011; Luk et al., 2012). Given that GA inhibited both α-syn fibrillation and disaggregated preformed α-syn fibrils, we sought to identify the effect of this phenolic acid on the seeding of α-syn aggregation. More specifically, mature α-syn fibrils were fragmented by sonication to obtain short fibrils, which were employed as “seeds” (Figure 4B). These short fibrillar “seeds” were then added to monomeric α-syn, which was allowed to aggregate as described above. As expected, the addition of short fibrillar seeds accelerated the fibrillation of α-syn as indicated by the increased Th-S counts. Indeed, with seeding, the extent of α-syn fibrillation after 6 h of incubation was comparable to the fibrillation of the protein incubated for 72 h without seeding (Figures 1, 4B, respectively).

In order to assess the effect of GA on the seeding of α-syn aggregation, 10 and 50μM GA was added to 100μM monomeric α-syn containing seeds at a final concentration of 2μM, and the mixture was incubated with continuous mixing at 37°C for 6 h. GA at 50μM inhibited the seeded fibrillation of α-syn by approximately 90% as indicated by the extremely low Th-S counts. At lower concentrations (10μM), GA also had an inhibitory effect on the seeded fibrillation of α-syn but to a smaller extent. At 10μM, GA inhibited seeded fibrillation by 40–50% (Figure 4A). These findings were confirmed by TEM (Figure 4B).

It has been reported that the oligomeric intermediates are the neurotoxic species in the amyloid fibrillation pathway (Allsop et al., 2001; El-Agnaf et al., 2001, 2006; Vendruscolo et al., 2001; Walsh et al., 2002; Argyriou et al., 2012; Breydo et al., 2012; Colla et al., 2012). As described above, at high concentration (4:1), GA inhibited both early and late aggregate formation, whereas at low concentrations (2:1 and 1:1), it stabilized α-syn oligomers (Figure 2A). To determine the effect of GA on toxicity conferred by α-syn aggregates, a cell-based toxicity assay, the MTT assay, was conducted with human neuroblastoma M17 cells. The samples were diluted to final α-syn concentrations of 5 and 0.5μM. The MTT assay showed that α-syn aged in the absence of GA decreased cell viability in a dose-dependent manner (Figure 5). However, when α-syn was aged in the presence of a high concentration of GA (4:1), there was visible neuroprotection of the cells observed (Figure 5). At low concentrations (2:1 and 1:1), GA exhibited a minor protective effect against α-syn toxicity, possibly due to the increase in oligomers as demonstrated by the ELISA results and the immunoblotting analysis (Figures 2A,B, respectively). Our results clearly demonstrate that GA reduced the toxicity of α-syn at a high concentration (100μM) by significantly inhibiting the formation of toxic α-syn oligomers. The toxicity of the oligomers resulted from α-syn: GA 1:4 was compared with α-syn oligomers prepared in the absence of GA as described in materials and methods section by MTT assay (Figure 5B). Both samples were characterized for the presence of oligomers by SDS PAGE (Figure 5C) and native PAGE (Figure 5D). It was clear that the oligomers resulting from the incubation of α-syn with GA were not toxic comparing with oligomers in absence of GA. Moreover, these oligomers were stable under denaturing conditions (Figure 5C) compared those formed in the absence of GA.

The strong inhibitory effect that GA exerted on fibrillation together with the stimulating effect it had on α-syn oligomerization at lower GA concentrations, prompted us to investigate further the interaction of GA with α-syn oligomers. For this purpose, monomeric α-syn (100μM) was aggregated in the presence of GA (GA: α-syn 4:1). After 5 days of incubation the samples were centrifuged and the supernatant was injected in a superdex 200 SE column. The elution volume for monomeric α-syn was determined by molecular weight standards (Figure S2A), and was eluted in a peak corresponding to column volume of 13–15 mL (Figure S2B), while oligomeric α-syn eluted in a peak corresponding to column volume of approximately 7–9 ml (Figure 6A). The fractions corresponding to the oligomeric and monomeric α-syn peaks were separately pooled together giving rise to P1 and P2 samples, respectively (Figure 6A), which were concentrated using a speed vac. The α-syn species in samples P1 and P2 were characterized by western blotting (Figure 6B). According to the western blotting results, the oligomers generated during the incubation of α-syn with GA are stable under denaturing conditions (Figures 5C, 6B). Electron microscopy of the same samples indicates the presence of different species of oligomers in P1 (Figure 6D), in agreement with the imunoblotting results (Figure 6B). In order to detect the incorporated GA in the P1 and P2 samples, we exploited the property of GA to produce UV absorbance spectra with two peaks, one at 225 nm and one at 260 nm. In the sample containing GA: α-syn at 4:1 molar ratio, we detected GA only in the oligomeric P1 samples (Figure 6C). These findings strongly support the hypothesis that GA binds to the oligomeric species and stabilizes them. To further evaluate whether GA interacts with α-syn monomers, we monitered a titration of GA into a solution of monomeric α-syn using two-dimensional NMR spectroscopy, which provides signals covering the entire amino acid sequence of α-syn (Figure 7). At stoichiometries of up to 10:1 GA: α-syn we observed no significant chemical shift or resonance intensity changes (Figures S3A–E), confirming that GA does not interact significantly with monomeric α-syn. This result is largely in agreement with NMR studies of related compounds Entacapone, Tolcapone and Quercetin (Di Giovanni et al., 2010).

To define the most effective molecular scaffolds against α-syn fibril formation and establish a structure-activity relationship for GA, we exploited the structural diversity of benzoic acid derivatives (i.e., hydroxybenzoic acids). Twelve hydroxybenzoic acids, including GA, salicylic acid and gentisic acid (see Table 1) were assessed for their anti-fibrillogenic properties by the Th-S binding assay (Figure 8). The structure of the tested hydroxybenzoic acids is characterized by the presence of a phenyl ring that has a carboxyl group and hydroxyl moieties (OH) attached to the ring at different positions. The selection of the compounds examined in the present study was based on the number (between 0 and 3) and position of the hydroxyl groups attached to the phenyl ring (see Table 2).

The kinetics of α-syn fibrillation was studied over a period of 5 days in the presence or absence of the phenolic acids. Based on the Th-S assay results (Figure 8 and Table 3), the effect of the twelve phenolic acids on α-syn fibrillation varied greatly. The extent of the inhibition of α-syn fibrillation was ~ 99% for GA, ~ 72% for 2,4,6-trihydroxybenzoic acid, ~ 60% for 3,4-dihydroxybenzoic acid, ~ 30% for five compounds including 2,6-dihydroxybenzoic acid and 4-hydroxybenzoic acid and only 5% for benzoic acid (see Figure 8 and Table 3). Moreover, three compounds, 2-hydroxybenzoic acid (salicylic acid), 3-hydroxybenzoic acid and 3,5-dihydroxybenzoic acid, failed to inhibit α-syn fibrillation, with salicylic acid appearing to enhance the aggregation of α-syn compared to the control (see Figure 8 and Table 3). By comparing the structures of the most effective compounds and the least effective ones, we conclude that better inhibition of fibrillation is observed when there are a large number of hydroxyl groups (-OH) attached to the phenyl ring of the compound. Moreover, the positions at which the hydroxyl groups are attached to the phenyl ring also appear to play a critical role. GA has hydroxyl groups attached at positions 3 and 5 (see Table 2) and is much more effective than 2,4,6-trihydroxybenzoic acid, which also bears three -OH groups but at different positions (Figure 8). Interestingly, 3,4-dihydroxybenzoic acid, which only lacks an -OH group at position 5 compared to the very effective GA, is the third best inhibitor of α-syn fibrillation (Figure 8). Among all the dihydroxybenzoic acids tested, the ones with -OH groups at two consecutive positions, i.e., 3,4-dihydroxybenzoic acid (60% inhibition) and 2,3-dihydroxybenzoic acid (44% inhibition), are more potent inhibitors compared to those that have -OH groups at non-consecutive positions, i.e., 2,5-dihydroxybenzoic acid (30% inhibition) and 3,5-dihydroxybenzoic acid (0% inhibition) (Figure 8). Additionally, all compounds with one -OH group, as well as the benzoic acid without a -OH group, failed to show any inhibitory effect on α-syn aggregation (Figure 8).

To further investigate the significance of the hydroxyl moieties on the activity of the GA and the other active phenolic acids, three additional compounds with either fluorides or methoxy groups rather than hydroxyl groups attached to their phenyl rings (4-methoxybenzoic acid, 3,4,5-trimethoxybenzoic acid and 3,4,5-trifluorobenzoic acid) (see Table 2) were tested for their ability to inhibit α-syn fibril formation. As expected, none of the three tested compounds in any of the three molar ratios tested (compound: α-syn 4:1, 2:1, and 1:1) could inhibit α-syn fibrillation (Figure 8). This finding provides additional support for the importance of the presence of the hydroxyl group in the phenyl ring.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.