Do Catechins (ECG and EGCG) Bind to the Same Site as Thioflavin T (ThT) in Amyloid Fibril? Answer From Saturation Transfer Difference NMR

Amyloidogenesis has been linked to various pathological neurodegenerative diseases, as well as metabolic disorders. ¹ Several studies have been trying to understand the mechanism of amyloidogenesis, finding inhibitors to slow down the aggregation kinetics. ² Structural insights into the essential motifs underlying amyloidogenesis have often led to the screening of several compounds as targeted therapeutics. A number of organic and small inorganic molecules have been identified, tried and tested for their potency in targeting the amyloidogenic species to inhibit disease propagation. ^3-6 Among the several ‘wonder’ molecules studied till date, the catechins such as epicatechin-3-gallate (ECG) and epigallocatechin-3-gallate (EGCG) (Figure 1A), ⁷ from tea extract have been shown to serve as potential therapeutics against several amyloidogenic disorders. ² These studies have shown the potency of the catechins in effectively inhibiting the amyloid fibril formation for several proteins including amylin, or, human Islet Amyloid Polypeptide (hIAPP), Amyloid beta 40/42 (Aβ40/42), and α-synuclein. ⁸ Experimental evidence has often suggested their ability to disaggregate the mature amyloid fibrils to form oligomeric aggregates that hinder the overall aggregation kinetics. ^9,10 Despite these advances, a complete understanding of the exact mechanism of inhibition by the catechins has often been hindered due to competitive binding with extrinsic molecular probes like Thioflavin T (ThT). ^9,10

Figure 1

(a) Chemical structure of the catechins- ECG and EGCG (b) Chemical structure of Thioflavin T.

Notably, ThT (Figure 1B) is among the most widely and effectively used molecular probes for detecting amyloid fibrils in vitro. ^11-13 Because of its remarkable specificity for the amyloid fibrils and the enhanced fluorescence emission upon binding to β-sheet fibrils, ThT is considered as the “gold standard” ^11-13 for amyloid detection. Despite its capability to bind to several fibrillar forms, non-specific binding affinity for any high-molecular-weight species has often been ruled out. ^14,15 Studies have found a correlation between ThT binding affinity and the chemical properties of the target fibrillar species. ¹⁶ In fact, reports show the prominent role played by the aromatic residues in defining substantial hydrophobicity that might underlie the structural epitope. ^11-13 The aromatic side-chains have also been known to partake in forming the characteristic structural motifs, rudimentary to ThT binding. ^11-13 On the other hand, parallel studies have demonstrated the importance of the aliphatic side-chain interactions that are likely to be cooperative in defining these ThT binding sites. ^11-13 However, the various target motifs and complexities of the systems, have restricted our understanding of the exact structural specificities. Nevertheless, these studies provide evidence for possible sequence specificity in ThT binding that is common to most amyloid fibrils, making ThT an active probe. In fact, hydrophobic interactions have been known to dictate amyloidogenesis in several of the known proteins and peptides. ^17-19 Particularly, the amyloid fibrils have been identified to share the cross-β sheet architecture that is responsible for ThT-binding. ²⁰

Interestingly, this sequence-specificity may be corroborated with the loss of efficiency for ThT in probing the kinetics in case of several tested amyloid inhibitors. Studies have often shown the effectiveness of the small molecules in targeting the aggregation-prone hydrophobic sequences. ² Thus molecular interaction with the inhibitor molecules results in compromising the availability of the structural motifs underlying ThT affinity. ²¹ This potential competitive binding interferes with efficient probing of the aggregation kinetics in the presence of the inhibitor molecules, particularly the catechins and other polyphenols. Several inhibitory studies with the catechins against an array of amyloids have shown the incompetence of ThT as a useful probe. ^9,10 Innumerable evidence has been collected, most of which suggest that ThT competes for the same binding site as that of the catechins that are often reflected in false positive data on the inhibitory activity. ²² However, despite the advances, it has been challenging to find a direct proof of any competitive interaction between ThT and the catechins.

In this study we have used Saturation Transfer Difference (STD) NMR to identify the binding epitope of catechin molecules (ECG and EGCG) for α-synuclein amyloid fibrils that are capable of binding to ThT. ¹⁶ ECG and EGCG have been previously studied for their inhibitory potency against α-synuclein amyloidogenesis in vitro. ^2,23–25 However, the exact mechanism of inhibition remains elusive. Thus to gain an in-depth understanding of a shared epitope for a possible competitive binding with ThT, STD NMR technique was employed. This technique can be effectively used to monitor ligand binding (over a range of affinities, K_D ∼ mM–μM) via non-scalar magnetization transfer from a large protein to smaller ligands. ^26–28 Further, theoretical quantification has also been made possible using the complete relaxation and conformational exchange matrix (CORCEMA) algorithm. ^26–30 CORCEMA-ST allows deduction of the theoretical STD values considering the binding kinetics and thermodynamics of the participating protons in the binding site. ^26–30 Recently, differential epitope mapping saturation transfer difference (DEEP-STD) NMR spectroscopy even allows the residue-specific identification of the protein residues directly involved in the binding site. ^31,32 This, further, helps in understanding the in-depth differential binding epitope for the protein-ligand systems. Active probing requires only a small concentration of the protein (µM–nM) in the presence of ~20–1000 times excess of the ligand molecule.

Initially, on-resonance and off-resonance radiofrequencies were analyzed for their effect on the free ThT sample (without the protein fibrils). At an on-resonance frequency of -0.5 ppm, no signal was observed from the ligand alone in the absence of α-synuclein fibril (Figure S1). However, strong STD signal intensities were obtained in the presence of protein fibrils with broad NMR resonances. Thus, all the subsequent STD experiments were performed using -0.5 ppm as on-resonance frequency and 40 ppm as off-resonance frequency, respectively. In due course of the experiment, any close molecular interaction with ThT resulted in the transfer of magnetization via spin-diffusion from the mature α-synuclein fibril to the ThT protons (Figure 2b and S1). Figure 2B shows the 1D STD spectra of ThT bound to α-synuclein fibrils. The saturation transfer signals were observed for the aromatic protons around ~7.7 and ~7.9 ppm (Figure 2B) as well as the aliphatic side-chains around ~2.55 and ~3.07 ppm. The difference spectra, obtained by phase cycling, show the presence of positive peaks corresponding to ThT protons that suggest the close molecular interaction with the fibrillar forms. Similar profiles were obtained for the catechins (ECG and EGCG) indicating their direct molecular interaction with the α-synuclein fibril (data not shown).

Figure 2

1D STD NMR spectra for matured α-synuclein fibrils in the presence of ThT and ECG. (a) Reference 1D NMR spectrum for α-synuclein fibrils in the presence of ThT and ECG at a molar ratio of α-synuclein:ThT:ECG = 1:300:300. (b) STD NMR spectrum (on-resonance: −0.5 ppm and off-resonance: 40 ppm) for ThT added to matured protein fibrils at a molar ratio of 300:1 (ThT:fibril). (c) Competition STD NMR spectrum (on-resonance: −0.5 ppm and off-resonance: 40 ppm) for ECG added to the sample containing α-synuclein fibrils and ThT, resulting in a molecular ratio of α-synuclein:ThT:ECG = 1:300:300, reveal complete displacement of the ThT peaks giving way to only peaks from ECG.

To study the possible competitive binding of the two ligands viz. ThT and ECG (also EGCG), competition STD NMR was performed. ^33,34 The binding of multiple ligands to α-synuclein fibrils was analyzed simultaneously, mainly, since the signals from the different ligands did not overlap. For the purpose, equimolar concentrations of ECG (or EGCG) and ThT was added to the α-synuclein fibrils at a molar ratio of α-synuclein:ThT:ECG = 1:300:300. This resulted in a decrease in the ThT signal that was accompanied by the appearance of the strongly positive signals from ECG at ~6.9 to 7.1 ppm (Figure 2c). A similar profile was obtained for EGCG (Figure S1). This is suggestive of competitive binding of ECG (or EGCG) for the fibrillating forms that displace the ThT molecules that were already residing within their fibril-binding epitope. In addition, the reverse STD NMR experiment was also performed where ThT was added to a solution containing ECG (or EGCG) and α-synuclein fibril. It was interesting to see that in contrast to the previous case the signals from ThT did not displace the signals from ECG in the STD spectra (Figure 2c). This observation implicates a higher affinity of the catechin molecules for the binding epitopes as opposed to ThT, when present in equimolar concentrations.

Thus our experimental observations reinstate an overlap of the structural motifs in mature α-synuclein fibrils that resulted in competitive binding of ECG (and EGCG) and ThT. This is in compliance with the previous reports ^1,2,9 where equimolar concentration of the catechins showed reduction of ThT binding intensity, apparently suggestive of dose-dependent inhibition. This reduction in ThT emission might be attributed to the competitive binding for the structural motifs- crucial for ThT association. Alternatively, since the experiments have been performed using pre-formed matured fibrils, the ECG/EGCG mediated consumption of the fibrillar forms, disaggregating them into smaller aggregates, cannot be completely ruled out as a possibility. ^23–25 It is worth mentioning that the ThT does not interact to the catechin-mediated suspension of non-fibrillar aggregates. ^23–25 In agreement to this reduction of the binding epitope in the sample-space, no saturation transfer signals for ThT was obtained when added to the fibrils already interacting with ECG. Thus, even though STD NMR enabled us to conclusively identify the competitive binding of the two ligands for the α-synuclein fibrillar forms, the data does not help in unraveling the inhibitory mechanism of ECG (or EGCG). Further in-depth analyses using DEEP-STD NMR technique should enable us to gain a clearer understanding of the binding epitope. ^31,32 Future application of the CORCEMA algorithm ^26-30 might enable us to get a quantitative interpretation of the difference in interaction for the two ligands. Nevertheless, STD NMR can serve as a useful molecular probe to screen potential inhibitor molecules against the amyloidogenic counterparts, helping to understand a possible mechanism of action at the atomic resolution.

Experimental

Chemicals

Thioflavin T, ECG and EGCG were purchased from Sigma Aldrich Co. (St. Louis, MO). A stock solution of 5 mM was made using a deuterated buffer for all the STD NMR experiments. Phosphate buffer in MilliQ water was first lyophilized overnight to remove the solvent molecules. The resultant phosphate salts were then dissolved in a similar volume of 100% D₂O as that of the water to keep similar molarity of the buffer solution.

α-Synuclein protein was expressed in E.coli BL21(DE3) strain and purified as described previously. ³⁵ The protein was purified by repeated cycles of ammonium acetate precipitation and resolubilization before lyophilization. The powdered protein in the desired concentration was dissolved in 20 mM Phosphate buffer (pH 6.8, 0.01% sodium azide), and dialysed against the same buffer at 4°C overnight using a 10 kDa MWCO membrane. The resultant sample was filtered off any pre-formed oligomers using pre-washed 100 kDa Molecular Weight Cut-off (MWCO) filters (Centricon YM-100, Millipore). The sample solution was made up to the desired concentration of 300 µM and allowed to aggregate. Matured fibrils were obtained after allowing aggregation for seven days upon incubation at 37°C and 250 rpm shaking condition. A minimal volume of only 1.6 µL was added to 100% D2O to make up the required sample condition.

NMR Spectroscopy

All NMR experiments were performed using the Bruker Avance III 500 MHz NMR spectrometer with a SMART probe. Topspin v3.1 software (Bruker Biospin GmbH, Switzerland) was used for data acquisition and processing. All NMR experiments were performed at 25°C.

Saturation Transfer Difference Experiment

The samples contained 1.6 µM amyloid fibril, and the molar ratios of protein to ligand were maintained as 1:300. For 1D STD NMR spectra, the spectral width was 12 ppm. Selective irradiation of the protein was achieved by a train of Gaussian-shaped pulses with a 1% truncation. Each pulse measured up to 49 ms in duration, separated by 1 ms delays in-between. The protein was irradiated at −0.5 ppm (on-resonance) and 40 ppm (off-resonance). Total saturation time was 2 seconds. Unless stated otherwise, the total numbers of scans used were 1024 for the reference spectra and 2048 for the difference spectra. Chemical shift values of all the ligands were assigned under the same conditions as the STD NMR spectra. The size of the STD effect was determined from signal intensities in the STD spectrum concerning those in the reference spectrum. Therefore, an STD effect of 100% is defined as when the signals in both ranges have the same intensity. For competitive STD experiments, titrations were performed by repeating the STD spectra in the presence of a second ligand at different molar ratios concerning the first ligand under the conditions stated above.

Competitive Binding Assay

Initially, ThT was titrated against α-synuclein fibrils up to a molar concentration of 300:1 (ligand to fibril). Later, the catechin (either EGC or EGCG) was added to this sample at a molar ratio of 1:1 (catechin to ThT) to study the competitive binding. The reverse titration was also performed to confirm the competitive affinity of the catechins for the fibrillary forms. In the reverse titrations, the fibrils were initially titrated with increasing concentrations of the catechin up to a molar ratio of 300:1 (ligand to fibril). Later, ThT was added at a molar ratio of 1:1 (ThT to catechin), to rule out any non-competitive binding of the catechins.