Flavonoids with Vicinal Hydroxyl Groups Inhibit Human Calcitonin Amyloid Formation

Abstract

Human calcitonin (hCT) is a 32-residue peptide hormone that can aggregate into amyloid fibrils and cause cellular toxicity. In this study, we investigated the inhibition effects of a group of polyphenolic molecules on hCT amyloid formation. Our results suggest that the gallate moiety in epigallocatechin-3-gallate (EGCG), a well-recognized amyloid inhibitor, is not critical for its inhibition function in the hCT amyloid formation. Our results demonstrate that flavonoid compounds, such as myricetin, quercetin, and baicalein, that contain vicinal hydroxyl groups on the phenyl ring effectively prevent hCT fibrillization. This structural feature may also be applied to non-flavonoid polyphenolic inhibitors. Moreover, our results indicate a plausible mechanistic role of these vicinal hydroxyl groups which might include the oxidation to form a quinone and the subsequent covalent linkage with amino acid residues such as lysine or histidine in hCT. This may further disrupt the crucial electrostatic and aromatic interactions involved in the process of hCT amyloid fibril formation. The inhibition activity of the polyphenolic compounds against hCT fibril formation may likely be attributed to a combination of factors such as covalent linkage formation, aromatic stacking, and hydrogen bonding interactions.

Introduction

Human calcitonin (hCT) is a 32-residue polypeptide hormone produced by the parafollicular cells of the thyroid gland.^[1] It plays a regulatory role in calcium-phosphorus metabolism and skeletal bone formation.^[1–2] Therefore, hCT injections were once used as a therapeutic treatment for bone-related disorders such as osteoporosis, hypercalcemia, and Paget’s disease.^[3] However, hCT has a tendency to self-associate and form amyloid fibrils which may cause cellular toxicity.^[4–5] The elevated production and subsequent amyloid deposition of hCT has been found to be associated with medullary thyroid carcinoma.^[1] Due to its high aggregation propensity, hCT injections are no longer used as a therapeutic treatment for bone disorders.^{[3, 6]} Instead, a different species of calcitonin, salmon calcitonin (sCT) is utilized because of its considerably lower tendency to form amyloid fibrils.^[6–8] Unfortunately, the use of sCT is accompanied by potential side effects such as vomiting, anorexia, and unwanted immune responses.^[9–12] It has been reported that hCT exhibits a higher potency of preventing osteoclast bone resorption than sCT when hCT fibrillation is avoided, making it a superior treatment for bone-related disorders.^[13] Hence, developing effective inhibition strategies against hCT amyloidogenesis is of considerable importance for treating hCT fibrillization related diseases and the therapeutic applications of hCT as an effective drug for bone-related disorders.

A large number of small molecules have been reported as inhibitors of the amyloid formation of aggregation-prone proteins.^[14–16] One class of these molecular inhibitors is polyphenols which contain multiple phenol units and are known to often carry antioxidant and anticancer activities.^[17] Polyphenols are found to be abundant in a wide variety of fruits, vegetables, and beverages including wine and tea. A variety of polyphenols have been found to inhibit fibril formation of amyloidogenic proteins, such as α-synuclein, amyloid-b (Ab), prion, tau, and islet amyloid polypeptide (IAPP).^[18] So far, however, there have been very few polyphenol-based molecules reported as inhibitors of hCT aggregation.^[19–21] Among these is epigallocatechin-3-gallate (EGCG), a polyphenol that belongs to the flavonoid family and makes up 50–80% of all the polyphenols in green tea.^[22] The abundance of polyphenols in green tea has been suggested to be associated with many health benefits including neuroprotective, antioxidant, anti-inflammatory, and anti-carcinogenic effects.^[23] EGCG has been found to be capable of preventing fibril formation of a variety of amyloidogenic proteins including hCT.^{[19, 24–27]} However, the inhibiting mechanism of EGCG in protein amyloid formation is not fully clear. EGCG base structure contains three phenyl rings (A, B, and D) and a heterocyclic ring (C) (Figure 1A). The B-ring contains a vicinal trihydroxyl structure and the D-ring is composed of a gallic acid moiety with a vicinal trihydroxyl structure. It is known that these trihydroxyl groups of EGCG can auto-oxidize to form quinones under neutral or alkaline pH.^[24] The formed quinone structures may further react with free SH groups in cysteine, or form covalent bonds with lysine, histidine, or arginine residues to produce adducts like Schiff base (Figure S1).^{[24, 27–28]} It has been suggested that the oxidized EGCG is able to covalently bind with amyloidogenic proteins such as lysozyme which results in amyloid inhibition.^[29] However, this is still under debate as it has also been reported that Schiff base formation may occur but may not be necessary for amyloid inhibition activity.^{[24, 27]} Alternatively, it has also been suggested that hydrogen bonding and/or aromatic interactions may be actively involved in the amyloid inhibiting function of EGCG.^{[27, 30]} Regarding hCT, it has been reported that EGCG may inhibit hCT fibrillization by interacting with the aromatic residues Tyr12, Phe16, and Phe22 which are known to be crucial in the formation and stabilization of the fibrillar structure of hCT.^{[19, 31]} Identifying the molecular mechanism of EGCG and other polyphenols in inhibiting hCT amyloid formation will illuminate novel strategies for the development of effective inhibitors of hCT amyloid formation. In this study, we investigated the hCT amyloid inhibition activities of a group of polyphenols, most with flavonoid skeletons, and assessed the structural features of the effective polyphenol inhibitors that may be crucial for the amyloid inhibiting function.

Figure 1.

(A) Chemical structures of EGCG, ECG and EGC. (B) Effect of EGCG on aggregation kinetics of hCT (5 μm) followed by ThT fluorescence in sodium phosphate buffer (5 mm, pH 7.4) at 25°C. (C and D) Tapping mode AFM images of hCT (5 μm) in the absence (C) or the presence (D) of 100 μm EGCG. The samples were incubated for 22–24 h at 25°C before imaging. (E and F) Effect of ECG (E) or EGC (F) on the aggregation kinetics of hCT (5 μm) followed by ThT fluorescence in sodium phosphate buffer (5 mm, pH 7.4) at 25°C. (G) Normalized maximum ThT fluorescence intensity in the aggregation of hCT in the absence or the presence of 100 μm EGCG, ECG, or EGC. Data are reported as means:standard deviation (SD) of triplicate results. (H and I) Tapping mode AFM images of hCT (5 μm) in the presence of 100 μm ECG (H) or 100 μm EGC (I). The samples were incubated in sodium phosphate buffer (5 mm, pH 7.4) at 25°C for 24 h before imaging.

Results and Discussion

Inhibitory effect of EGCG analogues on hCT amyloidogenesis

The aggregation kinetics of hCT (5 μm) with differing concentrations of EGCG in phosphate buffer (5 mm sodium phosphate, pH 7.4) was monitored using the fluorescence of thioflavin T (ThT). ThT selectively binds to the amyloid fibrillar structure, leading to an increase in the fluorescence intensity in the vicinity of 480 nm.^[32–34] As shown in Figure 1B, the aggregation kinetics of hCT shows a fast increase in fluorescence intensity until it reaches maximum intensity at ≈9 h. This result is in accordance with previous reports indicating rapid hCT amyloid formation.^{[6, 35–36]} Notably, after reaching the maximum, the fluorescence intensity slowly declines over time. This is consistent with other reports where the decrease of the signal is likely attributed to the precipitation of hCT aggregates over the incubation period.^[36–37] When 5 μm EGCG is added, there is a ≈50% decrease in the maximum ThT fluorescence intensity compared to that of hCT only (Figure 1B). In the presence of 50 μm EGCG, the maximum fluorescence intensity is reduced by ≈80%. Treatment with 100 μm EGCG results in negligible ThT fluorescence over the period of measurement, indicating a strong inhibition of hCT fibril formation. The morphology of the aggregates of hCT without or with EGCG was further assessed by atomic force microscopy (AFM) imaging. As shown in Figure 1C, hCT aggregated and formed long, thin and stacked fibrillar configurations. When hCT was incubated with 100 μm EGCG, no fibrillar structures were observed (Figure 1D). Instead, only some spherical oligomeric aggregates were seen. Overall, these results are consistent with previous findings that EGCG is an effective inhibitor of hCT amyloid fibril formation.^[19]

A recent 2D nuclear magnetic resonance (NMR) study of Huang et al. showed that EGCG binds largely to the central and C-terminus regions of the hCT peptide.^[19] Their results suggest the formation of favorable π–π stacking interactions between the aromatic rings of EGCG and the side chains of the aromatic residues in hCT.^[19] It has been reported that intermolecular interactions between the aromatic residues, for example, Phe16 and Phe19 (Figure 2) from the adjacent strands in the antiparallel β-sheet structure of hCT amyloids, play an important role in fibril formation of hCT, and disrupting such interactions can significantly affect the formation and stability of the aggregated structures.^{[31, 38]} On the other hand, EGCG has also been indicated to undergo auto-oxidation to form superoxides and quinones.^{[39, 40]} Palhano et al. demonstrated that EGCG is fully oxidized after an incubation period of 6 h in pH 7.4 phosphate buffer at 25 °C.^[24] Oxidized EGCG may covalently bind to the peptide via Schiff base formation and/or reaction with sulfhydryl groups in cysteine residues.^{[24, 41, 42]} hCT contains cysteine residues at positions 1 and 7. However, the peptide used in this study was oxidized to form an intramolecular disulfide bond between Cys1 and Cys7, and there is no free SH group available in hCT to react with the oxidized EGCG structure. Nevertheless, hCT contains Lys18 and His20 residues in the central region of the sequence which may form covalent adducts such as Schiff base with EGCG. These residues have been reported to be critical in the formation and stabilization of hCT amyloid fibrils by forming electrostatic interactions with the Asp15 residue on the adjacent monomeric chain in the β-sheet structure of hCT fibrils (Figure 2).^{[43, 44]} In addition, these residues are in close proximity to the Phe16 and Phe19 aromatic residues. If the covalent linkage were to form on the lysine or histidine residues, it might potentially interfere with both the electrostatic and aromatic interactions that are critical for fibril formation. To identify the covalent linkage of EGCG and hCT peptide, the hCT (5 μm) sample incubated with 100 μm EGCG overnight was examined using matrix assisted laser desorption ionization (MALDI). The results do not show an observable mass signal corresponding to the covalent adduct of hCT and the oxidized EGCG (data not shown). Palhano et al. previously reported that the products between Aβ40 and EGCG likely form high molecular weight cross-linked aggregates that could not be volatilized with MALDI.^[24] Therefore, our current results could not rule out the possibility of covalent bond formation that may be responsible for the amyloid inhibiting function of EGCG. Furthermore, it has been suggested that EGCG can form nonspecific interactions with the backbone and/or side chains of α-synuclein, IAPP, and Aβ via hydrogen bonding.^{[25–26, 30]} Shaham-Niv et al. reported that EGCG can also affect both early and later stages of metabolite fibril self-assembly through inhibitor-metabolite interactions possibly including hydrogen-bonding interaction.^[45] In this regard, nonspecific hydrogen bonding interactions between EGCG and hCT may also contribute to the inhibiting function of EGCG on hCT fibrillization.

Figure 2.

Plausible interstrand aromatic (green line) and electrostatic (red line) interactions between the adjacent monomeric strands in the antiparallel β-sheet structure of hCT amyloids.

In order to investigate the structural features that are crucial for the inhibition function of EGCG, we studied two flavanol analogues of EGCG, epicatechin gallate (ECG) and epigallocatechin (EGC). ECG does not contain a trihydroxyl group on the B-ring, and instead possesses a catechol moiety (3’,4’-OH on the B-ring, Figure 1A). EGC lacks the gallate moiety compared to EGCG (Figure 1A). Both compounds are also found in green tea but in less abundance than EGCG.^[41] ECG has been found to be a potent amyloid inhibitor for Aβ while a less potent inhibitor for IAPP compared to EGCG.^{[25, 46]} In contrast, EGC is much less effective in inhibiting aggregation of IAPP and Aβ, implying that the gallate moiety is critical in the inhibition of these amyloids.^{[25, 46]} Our results show that the addition of ECG or EGC leads to the decrease of the maximum ThT fluorescence intensity in hCT aggregation in a dose-dependent manner (Figures 1E and F), suggesting that both are able to in hibit fibril formation of hCT. Nevertheless, the decrease in ThT fluorescence with ECG or EGC is not as dramatic as that with EGCG (Figure 1G). When hCT is incubated with 100 μm EGCG, the maximum fluorescence intensity falls by ≈83%. When 100 μm ECG or EGC is applied, the fluorescence intensity falls by ≈50% for ECG and ≈65% for EGC, respectively. These results are consistent with the report of Cao and Raleigh that ECG and EGC are less effective inhibitors of IAPP compared to EGCG.^[25] In addition, our results indicate that removal of the hydroxyl group at the 5’ position of the B-ring (ECG) impairs the inhibiting function of the compound more significantly than elimination of the gallate moiety (EGC) (Figure 1G). Nonetheless, the AFM imaging of hCT incubated with either ECG or EGC does not indicate amyloid fibril formation; only spherical oligomeric aggregates were identified (Figures 1H and I). Taken together, these results demonstrate that ECG and EGC are also effective inhibitors of hCT amyloid formation.

To further evaluate the impact of the gallate moiety of EGCG in hCT amyloid inhibition, we tested the effect of gallic acid (GA) and a series of GA derivatives (Figure S2A) on hCT aggregation. Neither 3-hydroxybenzoic acid (3-HBA) or 3,5-hydroxybenzoic acid (3,5-HBA) shows an appreciable influence on the aggregation of hCT (Figure S2B). The maximum ThT fluorescence intensity of hCT drops of ≈25% in the presence of 100 μm 3,4-hydroxybenzoic acid (3,4-HBA) (Figure S2B). The AFM imaging exhibits that amyloid fibrils still formed steadily after incubating hCT with 100 μm 3,4-HBA or GA (Figure S2). Taken together, these results show that these compounds are not efficient inhibitors of hCT amyloid formation. This is consistent with the aforementioned result that removal of the gallate moiety in EGCG does not significantly weaken the amyloid inhibiting activity of the molecule. It is noteworthy that 3,4-HBA and GA both contain vicinal hydroxyl groups on the benzene ring with the potential to be oxidized and covalently linked to lysine or histidine residues in hCT to form adducts like Schiff base. Meanwhile, these compounds also contain a carboxylic acid group on the benzene ring. It is possible that the negatively charged carboxylic acid group under physiological conditions may form transient electrostatic interactions with Lys18 or His20 of hCT and thus hinder the covalent bond formation of the compound with the peptide. Interestingly, methyl gallate, a molecule that has a similar structure to the gallate moiety of EGCG, is also a weak inhibitor of Aβ amyloid formation,^[47] in accordance with our results.

Effect of the vicinal hydroxyl groups in flavonoids on hCT amyloid inhibition

To identify the roles of the aromatic skeletons and hydroxyl groups of flavonoids in hCT amyloid inhibition, we studied a group of flavone molecules that contain a similar base structure to EGCG without or with a hydroxyl group at different positions. Flavones are a class of flavonoids most of which can be readily found in tea, citrus fruits, berries, red wine, apples, and legumes.^[48] The base structure of the flavone (Fla) molecule contains two phenyl rings and a heterocyclic ring similar to that of EGCG (Figure 3A). As shown in Figure 3B and Figure S3, Fla does not show a noticeable effect on the aggregation of hCT. Since there is no hydroxyl group in Fla, aromatic interactions would be speculated to play a major role in the interplay between Fla and the peptide. Nonetheless, the carbonyl group on the heterocyclic ring may be involved in hydrogen bonding and possibly Schiff base formation. However, it should be noted that most Schiff base reactions with aldehydes or ketones require acidic pH conditions.^[49] Our results indicate that the interactions between the scaffold aromatic structure of flavone and hCT are not sufficient for inhibiting the peptide amyloid formation. These results are also consistent with the report of Malisauskas et al. that Fla does not influence the rate of human insulin fibril formation.^[50] Addition of a hydroxyl group at the 5-, 6-, or 7-position of the A-ring of flavone (5-hydroxyflavone (5-HF), 6-hydroxyflavone (6-HF), and 7-hydroxyflavone (7-HF), Figure 3A) does not improve the inhibition activity of the molecule on hCT aggregation (Figures 3B and S3). These compounds have been previously reported to be ineffective at inhibiting the aggregation of amylin or human insulin as well.^[50–51]

Figure 3.

(A) Chemical structures of Fla, 5-HF, 6-HF, 7-HF, and Nar. (B) Comparison of the maximum ThT fluorescence intensity in the aggregation of hCT (5 μm) in the absence or the presence of 100 μm Fla, 5-HF, 6-HF, 7-HF, or Nar. Data are reported as means±SD of triplicate results.

On the other hand, a flavanone molecule naringenin (Nar) that contains two non-vicinal hydroxyl groups on the A-ring and one hydroxyl group on the B-ring (Figure 3A), shows a mild effect on hCT aggregation. A ≈19% drop in ThT fluorescence intensity was observed with Nar compared to hCT alone (Figures 3B and S3). Nar has been found to be able to inhibit Aβ amyloid-induced mitochondrial dysfunction.^[47] However, Velander et al. reported that this compound is not effective at inhibiting amylin amyloid formation.^[51] As Nar contains three non-adjacent hydroxyl groups, the formation of hydrogen bonding interactions of Nar and the backbone and/or sidechains of hCT may putatively account for its mild inhibitory effect on hCT fibrillization.

Next, we studied two flavonoid molecules, luteolin (Lut) and quercetin (Quer) that contain adjacent hydroxyl groups on the same benzene ring (Figure 4A). Lut contains 4 hydroxyl groups, with two vicinal hydroxyls located on the 3’ and 4’ position of the B-ring. Lut is found in different dietary sources including chamomile tea, carrots, broccoli, celery, green pepper, and olive oil.^{[52, 53]} Lut has been recognized to be antiangiogenic and inhibit Aβ and insulin amyloidogenesis.^{[47, 50, 54]} Compared to Lut, Quer contains one additional hydroxyl group at the C-3 position of the C-ring. Quer is found in red onions, apples, berries, grapes, tea, tomatoes, and other dietary sources.^[55] Quer has been reported to possess strong anti-inflammatory effects and inhibit insulin fibril formation and destabilize the mature fibrils.^[54–56] As shown in Figure 4B, Lut shows an efficient inhibitory activity on hCT amyloidogenesis in a dose-dependent manner. In the presence of 5 μm Lut, the maximum ThT fluorescence intensity decreases by ≈40% compared to that of hCT only (Figure 4B). When 50 μm and 100 μm of Lut is added, the maximum ThT fluorescence intensity decreased by ≈65% and ≈75%, respectively. Quer also significantly interferes with the fibril formation of hCT and exhibits a slightly stronger inhibitory effect compared to Lut and EGCG at higher concentrations (Figure 4C and D). The ThT fluorescence trace of hCT aggregation is nearly flat in the presence of 100 μm Quer, implying an almost complete inhibition of the amyloid formation of hCT (Figure 4C). Only a few spherical oligomeric aggregates were observed in AFM imaging in the presence of Lut or Quer (Figure 4E and F), confirming the strong inhibition activity of these molecules on hCT fibril formation. Based on the structural features of these compounds, it appears that the catechol moiety on the B-ring likely makes both Lut and Quer stronger inhibitors of hCT fibril formation compared to Nar. Interestingly, Akaishi et al. previously reported that the 3’,4’-dihydroxyl group in fisetin is critical for its inhibitory effect on Aβ fibril formation.^[57] Our results are in accord with their findings. The emerging crucial impact of the catechol moiety in these effective inhibitors indicates that the covalent linkage may likely be an important factor in the mechanism of inhibition of hCT aggregation as has been indicated by previous studies with other amyloidogenic proteins.^{[29, 58, 59]} Quer contains an additional hydroxyl group on the heterocyclic ring, and this hydroxyl group could be involved in hydrogen bonding interactions with hCT, leading to a slightly stronger inhibition on hCT aggregation than Lut. In addition, Wang et al. suggested that factors such as hydrophobic interactions and aromatic stacking are important for Quer to inhibit bovine insulin fibril formation.^[56] These interactions should also be taken into account as well in the inhibitory effects of these compounds on hCT amyloid formation.

Figure 4.

(A) Chemical structures of Lut and Quer. (B,C) Effect of Lut (B) or Quer (C) on the aggregation kinetics of hCT (5 μm) followed by ThT fluorescence in sodium phosphate buffer (5 mm, pH 7.4) at 25 °C. (D) Normalized maximum ThT fluorescence intensity in the aggregation of hCT in the absence or the presence of the flavonoid molecules (100 μm). Data are reported as means±SD of triplicate results. (E and F) Tapping mode AFM images of hCT (5 μm) in the presence of 100 μm Lut (E) or 100 μm Quer (F). The samples were incubated in sodium phosphate buffer (5 μm, pH 7.4) at 25°C for 24 h before imaging.

Myricetin (Myr) is a flavonol that contains one additional hydroxyl at the C-5’ position of the B-ring in comparison to Quer (Figure 5A). It is found in fruits and vegetables such as onions, scallions, broccoli, apples, and berries.^[48] Myr has been shown to be a strong amyloid inhibitor of Aβ and α-synuclein where it can reversibly bind to oligomers through noncovalent interactions such as hydrogen bonding and aromatic stacking.^{[60, 61]} Here, Myr shows a remarkably potent inhibitory activity on hCT fibril formation in a dose-dependent manner (Figure 5B). There is a ≈60% drop in the maximum ThT fluorescence intensity of hCT aggregation when 5 μm Myr is added. In the presence of 100 μm Myr, the ThT fluorescence is negligible throughout the period of the measurement (Figure 5B), indicating a complete inhibition of hCT amyloid formation. This inhibition activity is stronger than that of EGCG (Figures 5D). AFM imaging shows that only few spherical oligomeric structures were observed when 5 μm hCT was incubated with 100 μm Myr (Figure 5E). These results show that Myr and Quer are the most potent inhibitors of hCT aggregation tested in this study. In light of the structural commonality of these efficient inhibitors, our results suggest that the presence of vicinal hydroxyls, either a catechol moiety or a trihydroxyl group on the phenyl ring, is critical in the inhibiting function of these flavonoid molecules on hCT amyloid formation. In addition to the possible contribution of hydrogen bonding interactions with the backbone and side chains of the peptide, the adjacent hydroxyl groups on the benzyl ring may also favor the formation of a quinone structure to form a covalent linkage with the peptide.

Figure 5.

(A) Chemical structures of Myr and Pyro. (B and C) Effect of Myr (B) or Pyro (C) on the aggregation kinetics of hCT (5 μm) followed by ThT fluorescence in sodium phosphate buffer (5 mm, pH 7.4) at 25°C. (D) Normalized maximum ThT fluorescence intensity in the aggregation of hCT in the absence or the presence of the small molecules (100 μm). Data are reported as means±SD of triplicate results. (E and F) Tapping mode AFM images of hCT (5 μm) in the presence of 100 μm Myr (E) or 100 μm Pyro (F). The samples were incubated in sodium phosphate buffer (5 mm, pH 7.4) at 25°C for 24 h before imaging.

To further investigate the impact of the multiple-hydroxyl phenyl structure on hCT fibrillization, we examined pyrogallol (Pyro) which is composed of a single vicinal trihydroxyl phenyl ring (Figure 5A). Remarkably, Pyro also inhibits hCT fibril formation (Figure 5C), although its inhibitory activity is weaker compared to that of EGCG (Figure 5D). No fibers were observed in the AFM imaging of 5 μm hCT incubated with 100 μm Pyro (Figure 5F), validating the inhibitory effect of this small molecule on hCT amyloid formation. Phan et al. reported that the pyrogallol moiety of gallocatechin gallate is one of the most potent groups that bind to Aβ through a combination of interactions such as aromatic and hydrogen bonding.^[62] It is intriguing that the addition of a negatively charged carboxylic acid group on the trihydroxyl phenyl ring, such as in GA, dramatically hinders the inhibitory activity on hCT amyloid formation (Figure S2).

Effect of baicalein and polyphenols with flexible phenyl rings on hCT fibrillization

The effective flavonoid inhibitors on hCT amyloid formation tested so far all contain a catechol or pyrogallol moiety on the B-ring. Baicalein (Baic) is a flavone compound that has a vicinal trihydroxyl group on the A-ring (Figure 6A). Baic is used in traditional Chinese herbal medicine and exhibits antiallergic, anticarcinogenic, and anti-HIV effects.^[63–67] Baic has also been shown to have inhibitory activities on the fibrillization of Aβ, α-synuclein, insulin, and amylin.^{[47, 50, 51, 61, 68–72]} Our results here demonstrate that Baic is a potent inhibitor of hCT amyloidogenesis, validated by both the ThT kinetics results and AFM imaging (Figures 6B and D). Its inhibition efficiency is comparable to that of EGCG (Figure 6C). In comparison to the results of Fla (Figure 3), these results indicate that the vicinal hydroxyl groups located on the A-ring of flavonoids also significantly enhance the inhibition efficiency of the molecule in hCT fibril formation.

Figure 6.

(A) Chemical structure of Baic. (B) Effect of Baic on the aggregation kinetics of hCT (5 μm) followed by ThT fluorescence in sodium phosphate buffer (5 mm, pH 7.4) at 25°C. (C) Normalized maximum ThT fluorescence intensity in the aggregation of hCT in the absence or the presence of EGCG or Baic (100 μm). Data are reported as means±SD of triplicate results. (D) Tapping mode AFM image of hCT (5 μm) in the presence of 100 μm Baic. The sample was incubated in sodium phosphate buffer (5 mm, pH 7.4) at 25°C for 24 h before imaging.

Besides flavonoids, we also studied a polyphenol rosmarinic acid (RA) that is comprised of two flexible phenol rings (Figure S4A). RA is found in herbs such as basil and rosemary^[73] and has been reported to exhibit inhibitory activity in amyloidogenesis of Aβ and IAPP.^[74] Each phenol group in RA contains a catechol moiety which could be involved in Schiff base formation as suggested in the vicinal hydroxyl containing flavonoids such as EGCG.^[58] Our result shows that RA also blocks hCT amyloid formation, with only a few of spherical aggregates formed (Figure S4B). Resveratrol (Resv) is another polyphenol that has been found to inhibit Aβ and IAPP fibril formation and even disaggregate the preformed fibrils.^[75] Resv is comprised of two polyphenolic residues with hydroxyl groups that could be involved in aromatic interactions and hydrogen bonding, but does not contain vicinal hydroxyl structures (Figure S4C). Interestingly, Resv is not capable of inhibiting hCT fibril formation (Figure S4D). These results indicate that the presence of vicinal hydroxyl groups on the benzyl ring may also be crucial for non-flavonoid polyphenolic compounds to efficiently inhibit hCT amyloid formation. It is worth noting that the amyloids formed when incubating hCT with Resv were thicker and more crystalline-like compared to that of hCT alone. This suggests that Resv can modulate the morphology of the hCT fibrils, although it cannot block the amyloid formation of hCT.

As summarized in Figure 7, the examined polyphenols lacking the vicinal hydroxyl groups are not efficient inhibitors of hCT amyloid formation, whereas the presence of the vicinal hydroxyl groups on the phenyl rings can dramatically enhance the inhibitory activity. Our study suggests that the vicinal hydroxyl groups on the phenyl ring is a critical structural feature of the flavonoids and likely non-flavonoid polyphenolic molecules for their inhibitory activity of hCT fibril formation. It has been indicated that the inhibition ability of polyphenols on protein amyloid formation is influenced by various factors such as aromatic stacking, hydrogen bond formation, and Schiff base formation.^{[51, 69, 71]} Our results imply that the aromatic and hydrogen bonding interactions between hCT and the flavonoids are not sufficient to prevent hCT amyloid formation. Therefore, the molecular mechanism of the vicinal hydroxyl groups of these polyphenol inhibitors may likely involve oxidation and subsequent covalent bond formation with hCT which consequently interferes with the crucial electrostatic and aromatic interactions of amino acid residues in hCT fibril formation.^{[31, 38, 43, 44]} In addition, it is conceivable that the non-vicinal hydroxyl groups and the carbonyl group on the heterocyclic ring of flavonoids may also be involved in hydrogen bonding with the backbone and/or side chains of hCT. All these interactions need to be taken into account with regards to hCT amyloid inhibition. The inhibition activity of polyphenolic compounds on hCT fibril formation may likely be attributed to a combination of factors such as formation of covalent linkage, aromatic stacking, and hydrogen bonding interactions.

Figure 7.

Summary of the effect of the polyphenols examined on hCT amyloidogenesis. The left group is nonefficient inhibitors of hCT amyloid formation, and the right group is efficient inhibitors.

In this study, we investigated the effect of the polyphenolic compounds on the amyloid formation of hCT in aqueous sodium phosphate buffer solution. The aggregation of hCT in vivo, however, occurs under rather complex physiological conditions. For instance, metal ions, such as calcium, zinc, copper, and iron, that steadily exist in the cellular environment have been implicated to play an important role in the formation of amyloid deposits in amyloid diseases.^[76] A number of novel amyloid inhibitors have been reported to target metal chelation and metal-associated amyloid formation.^[76–78] EGCG has been also found to be an iron chelator and has been shown to reduce iron levels and Aβ amyloidogenesis.^[79] Moreover, it has been well recognized that cellular membranes can significantly manipulate protein amyloid formation in vivo. Lorenzen et al. reported that EGCG moderately reduces the binding of α-synuclein oligomers to the surface of phospholipid membranes and thus blocks the toxicity of α-synuclein aggregates.^[80] On the other hand, it has been found that EGCG becomes a much less efficient inhibitor of IAPP amyloid formation at a phospholipid interface than in bulk solution.^[81] A systematic evaluation of the efficacy of the reported polyphenolic inhibitors in the presence of metal ions and/or in a membrane environment will provide mechanistic insight into the inhibiting function of these molecules on hCT amyloid formation in cellular environment. Assessment of the effects of these compounds against the hCT amyloid-induced cellular toxicity will provide more direct information on the potential application of these molecules in therapeutics. However, it should also be noted that these polar flavonoids may possibly exhibit poor pharmacokinetics in therapeutic applications.^[82] The bioavailability of these compounds might be limited due to issues such as poor systematic absorption.^[83–84] These factors should, therefore, be considered in designing novel flavonoid-based molecules for inhibiting hCT amyloidogenesis in vivo.

Conclusions

In summary, we report a group of polyphenolic compounds that inhibit hCT amyloid formation. Our results show that the gallate moiety in EGCG is not critical for the inhibition function on hCT amyloidogenesis. Remarkably, our results suggest that the presence of vicinal hydroxyl groups on the phenyl rings of flavonoids is essential for their capabilities of preventing hCT amyloid fibril formation. This structural feature may also be applied to non-flavonoid polyphenolic inhibitors. In addition to participating in hydrogen bonding interactions, this structural motif may favor the covalent bond formation with hCT to form adducts such as Schiff base. This could disrupt critical electrostatic and aromatic interactions in the process of amyloid formation of hCT. The inhibitory efficiency of polyphenolic molecules is likely dependent on the interplay of hCT and the compounds through synergistic interactions including covalent bond formation, hydrogen bonding, and aromatic interactions. Future studies with high resolution approaches, such as NMR spectroscopy and molecular dynamics simulation, will provide new insight into the interactions of the compounds and hCT at the molecular level. The knowledge obtained may facilitate the design of novel polyphenol molecules as effective inhibitors of hCT amyloidogenesis with appropriate bioavailability for treating hCT aggregation-related diseases. These advances would also allow for the return of hCT use in the clinic.

Experimental Section

Materials

All chemical reagents were purchased from commercial suppliers and used without further purification. Epigallocatechin-3-gallate (EGCG), epigallocatechin (EGC), epicatechin gallate (ECG), pyrogallol (Pyro), gallic acid (GA), 3-hydroxybenzoic acid (3-HBA), 3,4-hydroxybenzoic acid (3,4-HBA), 3,5-hydroxybenzoic acid (3,5-HBA), flavone (Fla), 5-hydroxyflavone (5-HF), 6-hydroxyflavone (6-HF), 7-hydroxyflavone (7-HF), naringenin (Nar), quercetin (Quer), myricetin (Myr), luteolin (Lut), baicalein (Baic), rosmarinic acid (RA), and resveratrol (Resv) were purchased from Sigma–Aldrich. 3-HBA, 3,4-HBA, and 3,5-HBA were prepared by dissolving the compound powder in water. All other compounds were dissolved in dimethyl sulfoxide (DMSO).

Synthesis and preparation of hCT

The hCT peptide was synthesized on a PS3 solid-phase peptide synthesizer (Protein Technologies Inc., Woburn, MA) using the appropriate Fmoc-protected amino acids. The hCT peptide was cleaved from the resin, purified by reversed phase high-performance liquid chromatography (RP-HPLC) utilizing a C18 column, and lyophilized until further use. The molecular weight of the peptide was confirmed with matrix-assisted laser desorption ionization (MALDI) mass spectrometry. Air oxidation was used to form the hCT disulfide bond as previously described.^[36] Briefly, hCT was dissolved under dilute conditions in 6m urea (pH 8.1) to prevent fibrillization.^{[19, 85]} The solution was incubated overnight with air bubbling to form the disulfide bond between Cys1 and Cys7. An Ellman test was utilized to confirm disulfide bond formation and MALDI mass spectrometry was performed to ensure that no intermolecular disulfide bond formation occurred.^[86] The oxidized hCT peptide was further purified by HPLC and subsequently lyophilized to form dry white powder. The peptide was then dissolved in hexafluoroisopropanol (HFIP), sonicated for 5 min, and quiescently incubated on ice for 1 h to ensure monomerization. The HFIP-peptide solution was pipetted into centrifuge tubes each containing ≈50 μL of the solution. The caps were left open and a stream of air was used to evaporate off the HFIP. To ensure the evaporation of the HFIP, the samples were stored in a vacuum desiccator overnight. The resulting clear crystals/film was stored at −80°C until further use. For all of the assays, oxidized and monomerized hCT peptide solutions were prepared by dissolving in 5 mm sodium phosphate buffer (pH 7.4) and the concentration was determined by UV absorbance at 275 nm (ε=1531 cm⁻¹m⁻¹).^[87]

Aggregation kinetics by ThT fluorescence

The aggregation kinetics of hCT, without or with additional small molecules, was prepared with 5 μm oxidized hCT, 5 mm sodium phosphate buffer (pH 7.4) and 20 μm ThT dye. 3-HBA, 3,4-HBA, and 3,5-HBA were dissolved in water and all of the other compounds were dissolved in DMSO to prepare 10 mm stock solutions, respectively. Additional dilutions were applied to make stock solutions with different concentrations. A particular amount of stock solutions was then added to the peptide solution to make a solution with 5 μm hCT, 20 μm ThT and 5 mm sodium phosphate buffer (pH 7.4). For the assays that contained DMSO, all of the solutions contained the same volume of DMSO to ensure consistency. 100 μL of solution was pipetted into each well of a 96-well microplate (Costar black, clear-bottom). The microplate was capped with a cover and sealed with para-film. The plate was then loaded into a Gemini SpectraMax EM fluorescence plate reader (Molecular Devices, Sunnyvale, California). The ThT fluorescence was measured every 10 min after shaking for 5 s with an excitation wavelength of 440 nm and an emission of 480 nm at 25°C for 22–24 h. All fluorescence kinetics data are reported as a means of triplicate results.

Atomic force microscopy

hCT (15–20 μL) without or with the small molecules were taken directly from the aggregation kinetics assay (after incubating for 22–24 h) and were adsorbed onto the surface of freshly cleaved mica (5×5 mm). The samples were covered and left to dry overnight. AFM imaging of the samples was acquired in tapping mode using the Asylum Research MFP-3D-BIO AFM system with MikroMasch NSC15/Al BS cantilevers.