Inhibition of Lysozyme Amyloid Fibrillation by Silybin Diastereoisomers: The Effects of Stereochemistry.

Silybin is a flavonoid lignin compound consisting of two diastereomers with nearly equal molar ratios. It has been reported that silybin can effectively inhibit the aggregation of amyloid protein, but the difference between the two silybin diastereomers has been rarely studied. In this work, the inhibitory ability of silybin to hen egg-white lysozyme (HEWL) was demonstrated, and the difference of kinetic parameters of two diastereomers was analyzed. Fluorescence quenching titration was utilized to analyze the binding differences to native HEWL between the diastereomers, and transmission electron microscopy (TEM) was utilized to analyze the characteristics of the surface of various samples. The differences between hydrophobicity and the secondary structure among several HEWL samples were measured by the 8-anilino-1-naphthalene sulfonic (ANS) acid fluorescence probe, Raman spectra, and far-UV circular dichroism. Moreover, the differences in the binding energy of these two diastereomers with HEWL were analyzed by molecular docking. Also, we have investigated the effect of silybin diastereomers on HEWL fibril-induced cytotoxicity in SH-SY5Y cells. Results show that silybin has a certain inhibitory effect on the HEWL fibrillogenesis process, while silybin B (SB) has a more significant inhibitory effect than silybin A (SA), especially at high concentrations. This work provides some insights into the screening of amyloid inhibitors from complicated natural products and indicates that SB has the prospect of further development as an amyloid inhibitor.

Introduction

The folding of proteins and peptides plays a critical role in maintaining the normal operation of life. If incorrect folding occurs, the protein may lose its original activity and function or even convert into potentially damaging protein aggregates. The initial protein forms monomer states with different folding degrees, resulting in highly disordered, partially structured, or nativelike oligomers. When amyloid occurs, more stable species with a β-sheet structure are formed, along with the increase in density and size. Finally, these proteins further grow to form well-defined fibers with a cross-β-structure and an orderly structure, which can precipitate in the body.[1] This abnormal accumulation of protein is responsible for the majority of amyloidosis such as Alzheimer’s disease, Parkinson’s syndrome, familial amyloidosis, Huntington disease, and type II diabetes disease.[2−4]

Human lysozyme is a bacteriolytic enzyme widely distributed in a variety of tissues and body fluids, including the gastrointestinal tract, kidneys, liver, lymph nodes, spleen, skin, lachrymal, and salivary glands.[5] Lysozyme amyloidosis (ALys) is one of the rarest types of systemic amyloidosis.[6] It is a hereditary, autosomal dominant disease that is associated with a single point mutation in the lysozyme gene. To date, 10 amyloid point mutations have been reported.[6] ALys is considered a slow-moving disease in early adult life with a median survival rate of approximately 18 years, accompanied by life-threatening clinical manifestations, particularly various hemorrhagic complications.[7] Due to the widespread distribution of lysozyme in many organs throughout the body, ALys may cause multiple organ disorders including digestive damage, spontaneous liver rupture, skin-mucosal disease, heart failure,[7] renal dysfunction,[8] repeated pulmonary infectious episodes, and granulomatosis of the bronchi.[5] Despite considerable efforts, the disorders caused by intractable amyloid aggregates remain one of the greatest threats to human health. Currently, the exploration of natural product inhibitors for protein amyloidosis, including ALys, has gradually become a hot research topic. Dozens of natural product inhibitors have shown remarkable therapeutic potential from in vitro studies and in vivo tests, such as curcumin, epigallocatechin-3-gallate (EGCG), baicalein, and so on.[9−12]

Silybin (Silibinin) is a unique flavonoid lignin extracted from milk thistle.[13] In nature, silybin consists of two almost equal amounts of diastereomers, silybin A (SA) and silybin B (SB), with different configurations at C10 and C11 (Figure ). Protective effects of silybin on the β-amyloid peptide such as amyloid β-protein (Aβ) and insulin amyloid fibrillation have been reported.[14−18] As reported, silybin reduced the cerebral plaque burden and brain microglial activation associated with an improvement of the behavioral abnormalities induced by AD pathology.[15,19] A study by Sciacca et al. showed that SB has a better Aβ amyloid aggregation inhibitory effect than SA and has been fully verified in the experiment based on Caenorhabditis elegans.(18) In addition, recent studies have shown that conjugation of a trehalose moiety to SB increases water solubility without significantly compromising its antiaggregation properties.[20] However, there are few reports on the different inhibitory effects of silybin diastereomers on amyloid transformation and other differences in vivo.[18,21,22] This experiment was conducted to investigate the inhibitory effects of silybin diastereomers on the guanidine hydrochloride (GuHCl)-induced aggregation process of hen egg-white lysozyme (HEWL), contributing to the study on the pharmacodynamic differences of silybin diastereomers.

Figure 1

Structure and atom number of silybin diastereomers.

Results

Aggregation Inhibition Kinetic Studies

We investigated the effect of silybin diastereomers (SA and SB) on the fibrillar aggregation of HEWL under 800 rpm of stirring in 3 M GuHCl solution of pH 6.5 at 50 °C. Thioflavin T (ThT) is a well-known molecular probe for the kinetic investigation of protein aggregation reactions because ThT can produce fluorescence specifically upon binding to amyloid fibrils, and the generated fluorescence intensity is directly proportional to the number of fibrils in the solution system. Figure is the scatter plot of the HEWL fibrillogenesis in the presence or absence of the silybin diastereomers. All traces in this figure showed sigmoidal kinetics and can be fitted satisfactorily with a sigmoidal function (eq ) (R2 = 0.992–0.998), which fully represents the progress of amyloid aggregation. Figure shows that the maximum fluorescence intensity, lag time value, and growth rate value changed with various concentrations of silybin diastereomers.

Figure 2

Scatter plot of HEWL fibrillogenesis in the presence or absence of silybin diastereomers. The kinetics of HEWL fibrillogenesis induced by 3 M GuHCl under neutral pH at 50 °C is shown. It was measured by a 96-well plate reader with ThT. The traces of best fit through the data points were obtained by fitting the data with a sigmoidal function (eq ).

Figure 3

Kinetic parameters of HEWL fibrillogenesis in the presence and absence of silybin diastereomers: maximum fluorescence intensity (Ymax, A), lag time (Tlag, B), and maximum growth rate (Kapp, C). Here, red is HEWL in the presence of SA and dark blue is HEWL in the presence of SB.

After a long-term investigation of the insoluble amyloid fibrillogenesis process, a nucleation-dependent aggregation model has been generally accepted,[23−25] and its mathematical kinetic model has become increasingly mature. The model believes that similar to general crystallization, the amyloid growth process begins with monomers coming from an unfolded or folded peptide, and then, they bond with each other into oligomers and protofibrils. The oligomers and protofibrils grow gradually with the addition of monomers until the large fibrils keep an equilibrium in the solution system.[26,27] A typical protein aggregation kinetic pattern follows the characteristic shape of a sigmoidal curve, fitted by time-resolved fluorescence intensity with three obvious steps: a lag time phase, a fast-growth phase, and a plateau phase,[28−30] where the lag phase and the maximum growth rate are characteristic of this sigmoidal curve.

As can be seen in Figure , the sigmoidal curve fitted with the function (eq ) has a strong dependence on silybin concentration, and the maximum fluorescence intensity gradient decreases with the increase of silybin concentration. The lag time and growth rate analyzed by the kinetic curve also changed with the concentration of silybin, indicating that silybin diastereomers can inhibit the final concentration of fibrils, decrease the formation speed of fibrillogenesis, and delay the beginning of nucleation (Figure B,C). Moreover, compared with SA, the maximum fluorescence intensity of SB decreased more significantly (Figure A) at high concentrations, with a longer lag time (Figure B) and a smaller growth rate (Figure C). At a fourfold concentration of silybin diastereomers (n/n), the maximum fluorescence intensity of SA can be reduced to 67% compared to the HEWL fibril in the absence of silybin diastereomers, the maximum growth rate can be reduced to 57%, and the lag period can be extended to 1.78 times, compared with the absence of silybin. By comparison, the corresponding values induced by SB were 58%, 50%, and 2 times (Figure ). In this experiment, the maximum concentration ratio of silybin was taken at 4:1 because the molar concentration of silybin was 560 μM for the poor water solubility of silybin and the insufficient ratio of dimethyl sulfoxide (DMSO). To achieve the values of IC50 for the silybin diastereoisomers, we tried to adjust the concentration of the silybin/HEWL molar ratio to 8 without increasing the DMSO. However, due to the poor water solubility, the actual concentration was inaccurate because silybin precipitated in the aqueous solution. Therefore, the values of IC50 for the silybin diastereoisomers are not achieved in our condition.

ThT is a kind of cationic benzothiazole dye used to detect the amyloid fibril protein content formed in the sample. When ThT binds to amyloid, ThT shows a red-shift in its excitation spectra and emission enhancement.[31] Considering that the fluorescence intensity of ThT bound to native HEWL was less than 1000 a.u., we did not detect its full wavelength measurement spectra. Compared with HEWL fibrils in the absence of silybin diastereomers, the maximum emission wavelength of HEWL fibrils in the presence of silybin diastereomers had a significant blue shift, but there was no significant difference between the two diastereomers. We hypothesized that the HEWL fibrils incubated with silybin diastereomers had less on-folding per unit concentration (Figure ).

Figure 4

Fluorescence intensity of ThT bound to HEWL fibrils in the absence and presence of fourfold concentrations of silybin diastereomers with the scanning performed at the emission wavelength of 460–510 nm with a step length of 1 nm, where black, red, and dark blue represent HEWL fibril, HEWL with SA, and HEWL with SB, respectively.

In addition to the inhibitory ability of silybin diastereoisomers to inhibit HEWL fibrillogenesis, we also investigated their ability to destroy preformed fibrils. First, we found that the moment when silybin diastereoisomers were added to preformed fibrils, the fluorescence intensity of the sample decreased slightly. The fluorescence intensity of fibrils with SA decreased to 79.78% of the original, and the fluorescence intensity of fibrils with SB decreased to 86.85% of the original. Compared with the SB, the fiber with SA decreased more, but this change did not reverse the process of fibrillogenesis completely. After 3 h of incubation, the fluorescence intensity of samples with SA or SB became 79.93 and 87.41% of HEWL fibrils, respectively. Since HEWL fibrils could partially refold under the conditions of not stirring or below 50 °C, the sample still has a little scope for fibrillogenesis. (after 3 h, the fluorescence intensity of HEWL fibrils became 103.10% of the original). We observed that even in the presence of SA or SB, the fibrils did not have the ability to inhibit the further accumulation of mature fibers, so we judged that the ability of silybin diastereoisomers to treat mature fibers is not outstanding compared to their inhibitory ability.

In general, the inhibition of HEWL amyloid by silybin diastereoisomers covers the entire process. No matter the growth rate, lag phase, or maximum fluorescence intensity, all of these kinetic parameters were improved compared to the parameters of lysozyme fibrillogenesis without ligands. These results illustrated that silybin diastereomers, especially SB, can inhibit the initial nucleation of HEWL amyloid aggregation and reduce oligomerization when it was induced by 3 M guanidine hydrochloride (GuHCl).

Fluorescence Quenching

The fluorescence quenching experiment was conducted to examine the binding of silybin diastereomers to native HEWL, in a concentration range of 0–2.16 × 10–5 mol/L in 3 M GuHCl solution at 280 nm excitation, 343 nm emission, and 300–470 nm scanning (Figure ). By wavelength scanning, we found that SB has larger fluorescence quenching changes with the increase of concentration (Figure B,C). After calculation by the Stern–Volmer equation, the bimolecular quenching rate constant (Kq) of SA from Ksv is 1.0842 × 1013 M–1 S–1, while the Kq value of SB is 1.3215 × 1013 M–1 S–1. Both silybin diastereomers are significantly larger than the maximum collisional quenching constant (1010 M–1 S–1), and the static quenching constant of SB is higher. According to the number of binding sites calculated by the equation (eq ), the n value of SB on native HEWL was 0.9399, slightly higher than that of SA (0.9159). In addition, the free energy change results show that SB has a stronger binding energy. The ΔG0 values of SA and SB were −4.941 and −5.232 kcal/mol, respectively.

Figure 5

(A) Silybin diastereomers induced fluorescence quenching of native HEWL with 3 M GuHCl solution at 25 °C. Inset shows the log(F0 – F/F) versus log[CQ] plots. (B, C) The fluorescence quenching of silybin diastereomers was observed in the wavelength scanning at the emission wavelength of 300–470 nm, indicating the concentration changes of 0, 1, 2, 4, 10, and 12 μL of silybin diastereomers.

Transmission Electron Microscopy (TEM) Characterization

Transmission electron microscopy (TEM) is a commonly tool used for the characterization of amyloid fibrils. TEM images of HEWL amyloid fibrils induced by 3 M GuHCl under neutral conditions in the presence or absence of silybin are shown in Figure . We found that the minimum diameters of fibrils are about 10–20 nm. HEWL fibrils in the absence of silybin diastereomers formed a cluster of flakes, while HEWL fibrils in the presence of silybin diastereomers existed as a short spiral stick. In the presence of SB, fibrils are shorter and thicker than those of SA, and the surface is covered by some insoluble particulate matter. We also found that the twisted fibril in neutral conditions with GuHCl was generally larger compared with the HEWL fibril under the strongly acidic condition at high temperature.[32] We speculated that it may be due to the more quick and thorough fibrillogenesis induced by GuHCl, as well as the larger number of mature fibrils entangled together leading to larger fibrils. Also, a partially neutral environment near the isoelectric point of HEWL may encourage the fibers to attract each other and stick together, which could be another reason for their shape.

Figure 6

TEM morphology of the HEWL fibril. The induced fibrillary morphology was obtained at 3 M GuHCl, neutral condition (pH 6.5, 50 °C), and silybin/HEWL molar ratio 4:1. HEWL fibrils incubated in the absence of silybin (A and D), in the presence of SA (B and E), and the presence of SB (C and F).

8-Anilino-1-naphthalene Sulfonic (ANS) Fluorescent Probe

8-Anilino-1-naphthalene sulfonic (ANS) is a special fluorescent probe that can emit strong fluorescence within a hydrophobic environment. Electrostatic forces and hydrophobic and aromatic groups play a very important role in the process of protein fibrillogenesis,[33] so ANS can be used to detect the hydrophobic environment changes between HEWL fibrillogenesis in the absence or presence of silybin diastereomers. Since the ANS fluorescence intensity of the nonfibrotic HEWL was low, the peak value could not be detected when compared with the HEWL amyloid fibril samples, so we further diluted the fibrotic samples for comparison (Figure B). The maximum fluorescence intensity of native HEWL and single ANS was about 520 nm (Figure B), the maximum fluorescence intensity of the HEWL amyloid fibril in the absence of silybin diastereomers was about 475–480 nm (Figure ), and the peak value of silybin diastereomers was about 480–490 nm (Figure ). The ANS fluorescence curve of native HEWL was almost the same as that of the blank ANS curve in the absence of protein, indicating that the native HEWL was in a natural conformation with no conformation changes. As a result of HEWL amyloid fibrillation and the hydrophobic group being exposed, the ANS fluorescence wavelength showed a blue shift to a short wavelength. The blue-shift distances of HEWL in the presence of silybin diastereomers were slightly greater than those of HEWL in the absence of silybin diastereomers. This may be because silybin diastereomers slightly reduced the HEWL internal hydrophobic groups’ exposure and made the HEWL amyloid fibrillation more difficult. There was no significant difference between silybin diastereomers in ANS fluorescence results, suggesting that there may be no significant difference between the two enantiomers in the protection of hydrophobic amino acid residues of HEWL.

Figure 7

(A) ANS fluorescence intensity of HEWL fibrils in the absence and presence of fourfold concentrations of silybin diastereomers with the scanning performed at the emission wavelength of 420–580 nm. HEWL fibrils in the absence and presence of fourfold concentrations of silybin diastereomers at higher concentrations. (B) Fresh HEWL, buffer, and HEWL fibril in the absence and presence of fourfold concentrations of silybin diastereomers were diluted for 1/24 times of the original solution, where red, dark blue, yellow, blue, and light blue represent HEWL fibril, HEWL fibril in the presence of SA, HEWL fibril in the presence of SB, phosphate buffer with 3 M GuHCl, and native HEWL, respectively.

Raman Spectra

Raman spectra of HEWL were recorded in the region from 400 to 3000 cm–1. Figure shows the spectra of HEWL hatched with or without silybin diastereomers. The violet curve at the top is a native HEWL aqueous solution with DMSO and GuHCl, while the green and yellow curves at the bottom are HEWL fibrils incubated in the GuHCl solution with and without DMSO, respectively. The peaks at 2940 and 700 cm–1 represent the DMSO solvent peak. Compared to other literature,[34−38] we found that the peaks of HEWL fibril and native HEWL measured in our experiment generally have some shifts, and this may be due to the influence of high concentrations of GuHCl in the aqueous solution. The frequency at 980–1070 cm–1 is a GuHCl characteristic peak, and it is several orders of magnitude higher than the peak representing the structure and amino acid residues of HEWL. We deducted it from the figure to eliminate its effect on the observation of the protein peak. The most obvious peak in Figure is the peak of amide I at 1666 cm–1. In the spectra coincubated with SB, the peak blue-shifted by 5 cm–1, while in the spectra coincubated with SA, the peak blue-shifted by 7 cm–1. In the spectra of the HEWL fibril without silybin diastereomers, the peak positions of amide I blue-shifted by 24 cm–1 with or without DMSO. The blue shift of the amide I peak often shows that the protein structure changes from an α-helix into a β-sheet or a random coil. The blue shift implies that the secondary structure of HEWL skeletons in the process of fibrillogenesis turns into a rich β-sheet. The presence of silybin can reduce the change in the secondary structure. In particular, SB’s inhibition of protein structure changes is more obvious.

Figure 8

Raman spectra of HEWL in GuHCl buffer at room temperature (violet), in GuHCl buffer with SB at 50 °C (blue), in GuHCl buffer with SA at 50 °C (red), in GuHCl buffer without silybin diastereomers at 50 °C (green), and GuHCl buffer without silybin diastereomers and DMSO at 50 °C (yellow).

Circular Dichroism Studies

Since the noise generated by DMSO and GuHCl seriously affected the structural peaks of HEWL, we also used far-UV circular dichroism (far-UV CD) spectra to analyze the influence of silybin diastereomers on the changes in the secondary structure of HEWL. The far-UV CD spectra of all of the aggregates, except for the natural HEWL, showed an extended β-sheet conformation as revealed by a broad negative band around 219–223 nm. However, native HEWL showed an extended β-sheet conformation as revealed by a broad negative band at 209 nm (Figure ). By analyzing the data in Figure , we obtained the secondary structural elements of all samples, of which natural HEWL had the least β-sheet content, and HEWL with DMSO, HEWL, HEWL with SA, and HEWL with SB had less β-sheet contents accordingly. The results showed that SA and SB could inhibit the HEWL fibrillogenesis to a certain extent and reduce the β-sheet content in the overall secondary elements, while DMSO could improve the content of folding to a certain extent (Table ).

Figure 9

Far-UV CD spectra of HEWL in 3 M GuHCl buffer. In this figure, the violet solid line represents the spectra of native HEWL, the yellow solid line represents the spectra of HEWL fibrils, the green solid line represents the spectra of HEWL with DMSO, the red solid line represents the spectra of HEWL fibrils in the presence of SA, and the blue solid line represents the spectra of HEWL fibrils in the presence of SB.

Table 1

Secondary Structural Element in HEWL in the Absence and Presence of Silybin Diastereomers Obtained by Analyzing Figure

	helix (%)	antiparallel (%)	parallel (%)	β-turn (%)	rndm. coil (%)
Native HEWL	18.7	20.8	5.7	17.6	33.0
HEWL	12.2	28.4	5.6	19.1	35.2
HEWL + DMSO	12.0	29.1	5.6	18.8	35.2
HEWL + SA	13.2	27.3	5.7	18.9	35.0
HEWL + SB	13.3	26.5	5.6	19.1	35.1

Protective Effects of Silybin Diastereomers against HEWL Fibril-Induced Cytotoxicity

In this part of our study, the effects of native HEWL and HEWL fibrils on the viability of SH-SY5Y neuroblastoma cells and the protective effects of silybin diastereomers on the amyloid fibril-induced neurotoxicity were assessed using the MTT assay and bright-field observation (Figure ). Incubation of SH-SY5Y cells with various concentrations (7–112 μM) of HEWL fibrils caused an obvious decrease in cell survival in a dose-dependent manner, but cells exposed to various concentrations (14–112 μM) of native HEWL evidenced no significant difference in appearance compared with the control group, and the viability of cells exposed to native HEWL at all concentrations showed no significant difference compared with the control group. To enhance the difference among the groups, the protective effect of silybin diastereomers was studied on the cells with 112 μM fibrils because it was the maximum concentration of the toxicity test that reduced the cell survival to 34.28 ± 4.45%.

Figure 10

SH-SY5Y cytotoxicity of HEWL fibrils in the presence or absence of silybin diastereomers by the MTT assay and bright-field observation. (A) Optical images of SH-SY5Y cells being maintained in only Dulbecco’s modified Eagle’s medium (DMEM) culture medium (control), culture medium containing 112 μM native HEWL (native), culture medium containing 112 μM HEWL fibrils (model), culture medium containing 112 μM HEWL fibrils in the presence of 224 μM or 448 μM SA (SA1 and SA2), and culture medium containing 112 μM HEWL fibrils in the presence of 224 μM or 448 μM SB (SB1 and SB2). (B) Dose-dependence cytotoxicity of HEWL fibrils and native HEWL. SH-SY5Y cells were exposed to different concentrations (14–112 μM) of native HEWL and different concentrations (7–112 μM) of HEWL fibrils. **p < 0.01, significantly different from control cells. (C) All groups correspond to the treatments of Figure A. **p < 0.01, significantly different from model cells.

Different from untreated control cells, we found that under bright-field observation, cells exposed to fibrils in the presence of SB did not form a network but rather an approximately elliptical cell community. Compared with the model group, the number of SB-group cells was greater, and the contour of the cells could be observed more clearly rather than being blocked by a large number of aggregation. Although the viability of cells exposed to fibrils in the presence of a high concentration of SB was far less than that of the control due to the higher severity of fibril toxicity, SB increased the viability of cells to 57.80 ± 4.25%, by nearly double relative to the untreated control cells. Furthermore, cells exposed to fibrils in the presence of low concentrations of SA showed a lower survival rate than the model cells, and cells exposed to high concentrations of SA showed no significant difference compared to the model cells. The real reason for the decrease in the cell survival rate caused by the low-dose SA is not clear.

Molecular Docking Studies

To further explore the inhibition mechanism of silybin diastereomers on HEWL amyloidosis fibrillogenesis, we conducted molecular docking experiments with Autodock. The results of docking analysis indicated that SB bonded more tightly to HEWL than SA in relative terms. The SA-HEWL binding energy was −7.1 kcal/mol, and the inhibition constant was 6.28 μM, forming hydrogen bonds with six amino acid residues VAL108, ASP52, GLN57, ASP101, ALA107, and ASN59, with bond lengths of 2.85, 1.85, 1.96, 2.03, 2.55, and 2.40 Å, respectively (Figure A1,A2 and Table ). The SB-HEWL binding energy was −8.69 kcal/mol, and the inhibition constant was 0.425 μM, forming hydrogen bonds with five amino acid residues ASN103, ASN59, ILE98, ASP101, and ASP52, with bond lengths of 2.18, 1.81, 1.78, 1.87, and 2.18 Å, respectively (Figure B1,B2 and Table ).

Figure 11

Docking results of silybin diastereomers against HEWL. The figure showed the docking sites were drawn by Autodock, and SA (A1 and A2) or SB (B1 and B2) docking results at HEWL sites were analyzed by Discovery Studio. (A1, B1) The stick models are silybin diastereomers, in which red is oxygen, HEWL is a secondary structure diagram, in which the amino acid residues that form forces with silybin are represented by line models. The surface between the protein and ligands shows the receptor–donor relationship of a hydrogen bond, in which the donor is represented by amaranth, while the acceptor is green. (A2, B2) The diagrams show the interactions between protein and ligands, and the hydrogen bonds are indicated by Viridis dashed lines.

Table 2

Summary of the Molecular Docking Studies of the Silybin Diastereomers against HEWL

name of the compound	binding energy (kcal/mol)	inhibition constant (kI) (μM)	intermolecular energy (kcal/mol)	reference RMSD
SA	–7.1	6.28	–9.78	36.02
SB	–8.69	0.426	–11.38	34.69

As shown in Figure and Table , we found that aromatic rings also play a key role in the interaction between HEWL and ligands, such as pi–donor hydrogen bonds and pi–alkyl bonds, which are also widely found in other docking conformations not shown in this figure. Besides, the amide–pi stacking effect may provide steric hindrance that affects the spread of internal hydrophobic residues in the core of HEWL and inhibit further deterioration of HEWL. According to Frare et al.,[3] the main active sites of HEWL amyloid aggregation are 57–107 amino acid residues, so our docking results are mainly distributed in that area. When silybin diastereomers have hydrogen-bond and hydrophobic interactions with 57–107 amino acid residues, the silybin attaches to the structure of the protein residues and prevents fibrillogenesis.

By analyzing the structure of silybin diastereomers and the relationship between silybin diastereomers and HEWL, we found that the difference of silybin diastereomers in inhibiting amyloid aggregation was caused by the hydroxyl of C20 and C23. Hydrogen bonds between silybin diastereomers and HEWL are mainly produced in C20 hydroxyl groups and hydroxyl C23. Because of its alcohol hydroxyl group, SB has a smaller size to dock in a deep HEWL cavity and is more easily adapted to the relevant location. While the hydrogen-bond position of SA is on the phenol hydroxyl group, the steric hindrance and stereoscopic effect make the corresponding alcohol hydroxyl group less interactive with HEWL.

Discussion

Lysozyme is non-neuropathic hereditary amyloidosis with a wide range of clinical manifestations.[7] As a representative amyloid, HEWL has been receiving wide attention in the investigation of amyloid fibrillation. There are various incubation methods to promote HEWL fibrillation, such as high temperature at low pH conditions,[3,39,40] guanidine hydrochloride or urea under neutral pH conditions,[41,42] and ultrasound to promote fibrillogenesis.[43,44] The factors that influence the kinetics of amyloid fibrillation are complex, including the presence of seeded molecules, foreign surfaces such as test tubes or microwell wall, ionic strength of the solution like buffer concentration and pH, and the intensity of agitation. The method of high temperature at low pH usually requires a continuous reaction for about 2 weeks with taking points per day, which is not conducive to maintaining the solution environment. Therefore, GuHCl is used to induce HEWL fibrillogenesis to promote the reaction process. Because the promoting effect of incubation at 37 °C did not meet the expectation in the previous experiment, a higher incubation temperature at 50 °C was adopted with a moderate reaction speed in this study, considering that the incubation temperature range of HEWL was relatively wide without GuHCl induction. Compared to the HEWL incubated at 65 °C in strongly acidic, aqueous conditions, HEWL incubated by GuHCl at neutral conditions had a faster fibrillogenesis velocity. The fibril further gathered to form larger fibrils that were not evenly dispersed in the system. In macroscopic samples, we observed that the fibril samples under the neutral condition were milky white, while the fibril samples in the acidic condition showed a translucent flocculent precipitate or a gel solution system. Besides, we also observed that HEWL incubated at a high temperature in acidic conditions will be gradually turbid-titrated by alkali, and its form was consistent with the sample induced by GuHCl at the neutral condition. Therefore, we hypothesized that the seeds formed by primary nucleation have alkaline isoelectric points and are more likely to accumulate free monomers, thus accelerating the formation of HEWL fibers in the rapid growth period.

In addition to the above-mentioned conditions, many other factors can also affect the amyloid aggregation of HEWL, and minor environmental variations can enlarge the differences. Moreover, 96-well plates were employed as the standard parallel test method because the results obeyed the Gaussian distribution;[45] however, the dispersion of data is largely due to the change of the minor environment. Therefore, our work adopts the traditional method of environmental reaction. After mixing the system at a certain time, the fixed volume solution was sucked out and diluted for fluorescence measurement.

Besides analyzing amyloid kinetics by measuring time-resolved ThT fluorescence, some researchers also measured the protofilament length by TEM images and the peak displacement caused by protein structure changes by Raman spectroscopy for kinetic analysis.[46] These methods were all obtained through a sigmoidal curve fitting, and researchers could learn in which stage ligands intervene in amyloid aggregation by analyzing the parameters. In addition, several studies obtained particle size distribution based on time-resolved dynamic light scattering (DLS) or atomic force microscopy (AFM) to analyze the kinetics in the process of amyloid aggregation.[47,48] Knowles et al. applied a mathematical model to parse the amyloid self-assembly process from a microscopic perspective.[49] The relative abstract parameters such as elongation constant and fracture constant can be obtained from the amyloid assembly kinetics.[48] In the results of Raman spectra, the signal around 1000 cm–1 is because the high degree of response of GuHCl disguised the peaks of Phe and Trp, which show the side group of the protein. In addition, the peak around 2900–3000 cm–1, which represents DMSO, covered the signal of the C–H stretching vibration; other characteristic peaks were not obvious due to the high signal of GuHCl. We also tried to raise or lower the protein concentrations during the Raman spectra test, but the quality of the protein peak response was still not high, so we only analyzed the signal of amide I.

As shown in Figure , in addition to interactions like H-bond, some of the weaker binding energy dock results between silybin and HEWL show Pi–alkyl and Pi–sigma interactions, while SA shows fewer interactions with HEWL. In addition, in some of our docking results, the interactions between HEWL and the hydroxyl groups on C20 and C23 were really limited and did not produce hydrogen-bonding interactions with HEWL. These findings indicated that although there were combination differences between the two diastereomers with HEWL, the differences were limited. Combined with the previous discussion, these analyses show that SB has larger interactions with HEWL in contrast with SA, which also confirms the results of our experiment.

The differences between two silybin diastereomers were also reflected by the stereoselective metabolism in several pharmacological and metabolic processes. For instance,[22] SB is metabolized by the bovine microsomes ca 2–3 times quicker than SA in the mixture, and both diastereomers influence the respective metabolism of each other; thus, using the two simultaneously may reduce their actual effects due to the interactions between the diastereomers. Moreover, silybin diastereomers possess stereoselective metabolism mainly at C7 and C20 sites; the AUC0→6h value of the total SB was 20-fold higher than that of SA.[50] The interaction between C20 as well as the groups on the same aromatic ring and the receptor may be the key to the difference in the inhibitory effect of silybin diastereomers on lysozyme fibrillogenesis. Considering that silybin diastereomers possess great differences in bioavailability in vivo, and the SB has a better treatment effect than SA in the process of HEWL and Aβ fibrillogenesis,[18] SB alone may have a better effect in the actual treatment, and its high bioavailability compared with SA can reduce the dose of the drug.

Silybin has been applied to treat various liver disorders such as liver injury, hepatitis, and fibrillogenesis for a long time.[51−53] Besides, recent studies have found that silybin has a certain effect in preventing cancer, has antitumor and antioxidant properties, can stabilize the plasma membrane,[54,55] and has a clear inhibitory effect on lipoxygenase and peroxidase.[56] The main therapeutic effect of silybin on neurodegenerative diseases is speculated to be due to its ability to resist oxidation and scavenge free radicals, which can protect the central nervous system from damage and memory impairment. In the test based on C. elegans, silymarin shows the ability to slow down the progress of neurodegeneration by enhancing resistance to oxidative stress.[19] However, according to our results, the protective effect of silybin on amyloidosis should not be neglected in the treatment of neurological diseases. At present, the experimental study based on SH-SY5Y cells found that silybin can block the neuronal toxicity induced by insulin fibrils and protect the cell viability and biological membrane integrity by disabling its ability to destroy the mitochondrial structure.[16] Apart from insulin amyloid, silybin also exhibited similar inhibitory effects in Aβ-related experiments.[15] In our work, we investigated differences between silybin diastereomers based on SH-SY5Y cells. Unlike previously published reports, for the higher concentration of the fibrils we used, we did not have a control-like cell survival rate. Similar to the results of the toxicity test based on C. elegans, SB had a better inhibitory effect than SA.

In conclusion, the inhibitory effect of silybin diastereomers on HWEL fibrillogenesis was studied by fluorescence detection, cytotoxicity test, and molecular docking. We believe that silybin diastereomers have a certain inhibitory effect on HEWL amyloid fibrillation, and SB presented a stronger inhibition and a larger binding force to HEWL. Silybin is likely to have the dual effects of protecting the biological membrane and blocking protein fibrillogenesis in the body, while the high bioavailability of SB in vivo and its superiority over SA may extend this advantage. This study confirms the important role of stereochemistry in the amyloidosis of protein and provides a reference for the study of natural drugs with a complex composition like silybin.

Materials and Methods

Material

Hen egg-white lysozyme (HEWL, L6876) and guanidine hydrochloride (GuHCl, G3272) were purchased from Sigma-Aldrich. Thioflavin T (ThT) was purchased from Aladdin (T168914). 8-Anilino-1-naphthalene sulfonate (ANS) was purchased from Macklin (A800507, Shanghai, China). Phosphate buffer (PB) was prepared from sodium dihydrogen phosphate dodecahydrate and sodium dihydrogen phosphate, purchased from Nanjing Chemical Reagent Co (China). All other reagents are analytically pure and not particularly purified. Statistical analysis was performed by SPSS and Origin.

Separation of Silybin Diastereomers

Silybin diastereomers were separated in the laboratory after optimization on the method of Gazak et al.[57,58] The structural differences between the two diastereomers were increased by adding an acetyl group to silybin at C23 by Novozym 435, and the two were separated by column chromatography; then, the acetyl group was removed separately and, finally, further purified by recrystallization.

Aggregation Inhibition Kinetic Studies

First, 3 M GuHCl was dissolved in 20 mM phosphate buffer to prepare an incubation buffer solution, and the pH was adjusted to 6.5. HEWL was accurately weighed with a final concentration of 2 mg/mL (140 μM). Due to the poor solubility of silybin diastereomers in water, they were dissolved in DMSO with a 100-fold concentration into different gradients for preparation. After the mixture was passed through a 0.22 μm potential energy surface (PES) filter, it was heated in a water bath at 50 °C[42] under 800 rpm. After mixing well from the closed reaction system, 10 μL was pipetted every 20 min, and 7 μM ThT solution was added and diluted to 200 μL for the fluorometric assay. The fluorescence of the resulting samples was measured at room temperature with the excitation wavelength fixed at 440 nm and the emission wavelength fixed at 480 nm. All data were accurately fitted to a sigmoidal curve by the following equation with Origin:where A2 is the original ThT fluorescence intensity (baseline data); A1 – A2 are the maximum fluorescence intensity; x and x0 are measured times and half time of the maximum fluorescence intensity, respectively; the apparent rate constant is given by 1/d; and the lag time is given by x0 – 2d.

We performed the scanning at the emission wavelength of 460–510 nm with a step length of 1 nm to investigate whether the ThT emission wavelength of each system was different after the addition of silybin diastereomers.

Morphological Analysis

The samples of HEWL in the absence and presence of silybin diastereomers were washed with buffer to remove GuHCl after centrifuging at 10 000 rpm for 10 min. Then, the precipitate was diluted to 2 mg/mL (140 μM) with phosphate buffer and mixed well before adding the sample on the copper grid before testing. All images were taken by Je-2100F (Jeol).

The binding relationship between silybin diastereomers and HEWL was investigated using an increasing gradient method for fluorescence quenching with the incubation condition (3 M GuHCl was dissolved in 20 mM phosphate buffer at pH 6.5). An increasing volume (1–20 μL) of 2 mg/mL silybin diastereomer solution in DMSO was added into a fixed volume of the protein solutions (3 μM to 3 mL). The fluorescence intensity was measured at room temperature (25 °C) with the excitation wavelength fixed at 280 nm and the emission wavelength fixed at 343 nm. Also, emission wavelength scanning at 300–470 nm was detected for several concentrations. The fluorescence quenching data were analyzed with the Stern–Volmer equation (eq ) and static quenching of fluorescence (eq ).where F and F0 are the fluorescence intensities in the absence and presence of different concentrations of silybin diastereomers; KSV is the quenching constant (L/mol) of the Stern–Volmer equation, which is the slope of the Stern–Volmer equation; CQ are the different concentrations of silybin diastereomers; K is the y-intercept of the line; and n is the binding-site number.

Furthermore, we calculated the free energy changes (ΔG0) of silybin diastereomers and HEWLwhere R = 1.987 cal/(mol·K), T = 298.15 K, and K is the quenching constant.

ANS Binding Assays

ANS was used for wavelength range scanning at 380 nm excitation wavelength and 420–580 nm emission wavelength to investigate the hydrophobicity on the surface of HEWL (140 μM) in the absence and presence of silybin diastereomers. After diluting the samples for fibrillogenesis dynamics 24 times, ANS with a final concentration of 20 μM was added to the black 96-well plate and incubated in the greenhouse for 1 h before measurement.[59,60]

Raman Spectra

The Raman samples were all 20 mg/mL (1.4 mM) HEWL containing 1% (v/v) DMSO, and the samples were incubated in an environment of 3 M GuHCl. The silybin diastereomers were configured into 2.8 mg/mL DMSO solution and added at 1% (v/v). Besides, the fibril group without DMSO incubation was used as a separate control to investigate the peak effect of DMSO. The sample was detected at a continuous wavelength of 532 nm with a Raman spectrometer (Renishaw, Invia).[38]

Circular Dichroism Studies

Far-UV circular dichroism (far-UV CD) spectra of HEWL samples were measured between 200 and 250 nm using a BRIGHTTIME Chirascan spectrometer at 25 °C. For all measurements, a protein concentration of 0.5 mg/mL (35 μM) and a cell of 0.1 mm pathlength were used. The baseline was determined using the control buffer (20 mM phosphate buffer with 3 M GuHCl) and was subtracted from the results of all samples.

SH-SY5Y Cytotoxicity Assays

SH-SY5Y cells, which were kindly provided by the Stem Cell Bank, Chinese Academy of Sciences, were cultured in the DMEM-F12 medium (Hyclone) with 2.5 mM l-glutamine, 100 IU/mL penicillin, and 100 mg/mL streptomycin, supplemented with 15% fetal bovine serum (FBS), and kept at 37 °C in a 5% CO2 humidified atmosphere. The samples were 2 mg/mL (140 μM) HEWL in the absence and presence of silybin diastereomers dissolved in 20 mM phosphate buffer with 3 M GuHCl. Cells were seeded in a 96-well plate at a density of 1 × 105 cells/well, and the medium was changed before incubation with various samples. For cytotoxicity experiments, cells were treated with increasing amounts (0–112 μM) of HEWL fibrils in the absence and presence of silybin diastereomers and incubated for 24 h. To evaluate the involvement of the antiamyloidogenic activity of silybin diastereomers against toxicity induced by HEWL amyloid fibrils, protein samples aged without or with various concentrations of silybin diastereomers (224 and 448 μM) under amyloidogenic conditions were added to the cells and left for 24 h. Cells treated with native HEWL (112 μM) were used as the control. The data were expressed as a percentage of cell viability in untreated control cells, and each value represents the mean ± SD (n = 3). The statistical analysis was performed using IBM SPSS Statistics 22.

Molecular Docking

Molecular docking studies were conducted to further study the interaction between silybin diastereomers and HEWL proteins as well as the differences between silybin diastereomers. The crystal structure of HEWL (PDB ID: 2lyz) was from the Protein Data Bank, and silybin diastereomers were drawn by ChemDraw and then optimized by Gaussian. The docking was set as semiflexible and calculated with the genetic algorithm 100 times and then analyzed by Discovery Studio 2019.