Inhibition of amyloid-β(16–22) aggregation by polyphenols using replica permutation with solute tempering molecular dynamics simulation

Abstract

Aggregates of amyloid-β (Aβ) peptides are thought to cause Alzheimer’s disease. Polyphenolic compounds are known to inhibit Aβ aggregation. We applied replica permutation with solute tempering (RPST) to the system of Aβ fragments, Aβ(16–22), and polyphenols to elucidate the mechanism of inhibition of Aβ aggregation. The RPST molecular dynamics simulations were performed for two polyphenols, myricetin (MYC) and rosmarinic acid (ROA). Two Aβ fragments were distant, and the number of residues forming the intermolecular β-sheet was reduced in the presence of MYC and ROA compared with that in the absence of polyphenols. MYC was found to interact with glutamic acid and phenylalanine of Aβ fragments. These interactions induce helix structure formation of Aβ fragments, making it difficult to form β-sheet. ROA interacted with glutamic acid and lysine, which reduced the hydrophilic interaction between Aβ fragments. These results indicate that these polyphenols inhibit the aggregation of Aβ fragments with different mechanisms.

Significance

Alzheimer’s disease is thought to be caused by the aggregates of amyloid-β peptides that accumulate in the brain. The replica permutation with solute tempering molecular dynamics method efficiently visualizes the aggregation and disaggregation of molecules using a computer. This paper uses this method to simulate myricetin and rosmarinic acid, which are the aggregation inhibitors of amyloid-β peptides. We explore the difference in the mechanism by which these molecules inhibit the amyloid-β aggregation.

Introduction

Alzheimer’s disease (AD) is a human brain disorder that interferes with memory, thinking, and behavior. Senile plaques have been observed in the cerebral cortex of patients with AD. Amyloid fibrils, which are needle-like aggregates of amyloid-β (Aβ) peptides, are contained in the senile plaques [1]. Aβ peptides have a propensity to aggregate into harmful oligomers and amyloid fibrils. Inhibiting Aβ aggregations is crucial for the treatment of AD. Many experimental and computational studies have been reported to elucidate the structures of these oligomers and amyloid fibrils and to clarify the aggregation and disaggregation processes of Aβ peptides [2–22]. Several inhibitors for the aggregation of Aβ peptides have been studied, such as polyphenols [23–25], epigallocatechin-3-gallate (EGCG) [26–30], and indanone-carbamate-based molecules [31,32]. In the present paper, we focus on polyphenols. Ono et al. [24] found that polyphenolic compounds inhibit the oligomerization of Aβ. Their results show that the efficacy of phenolic compounds as inhibitors is as follows: myricetin (MYC)>rosmarinic acid (ROA)>nordihydroguaiaretic acid=ferulic acid≥curcumin. The presence of rings in their chemical structures allows for hydrophobic and hydrophilic interactions with amino acid residues [33,34]. The binding residues of the polyphenol compounds were investigated using nuclear magnetic resonance (NMR) spectroscopy and molecular dynamics (MD) simulations [24]. They found that MYC promotes the chemical shift changes for the sequential residues Lys16–Val24 and Ile31–Gly33 in the Aβ peptide. Several computational analyses have been performed on the effects of polyphenols on the Aβ peptide and Aβ(16–22) peptide, which is a fragment of Aβ. Ghorbani et al. modeled a helix structure of the Aβ peptide and placed polyphenols near it by a molecular docking technique. They showed that the helix structure is maintained by using MD simulations [35]. Replica-permutation MD simulations have been performed for polyphenols and one Aβ(16–22) peptide, but the effect of polyphenols on aggregation between two or more Aβ(16–22) peptides has not been investigated [25]. Despite these studies, the detailed mechanism of aggregation inhibition has not been elucidated so far.

It is difficult to simulate the interaction between Aβ (16–22) peptides and polyphenols by the conventional MD method. Replica-exchange method (REM) [36,37], one of the generalized ensemble algorithms [38–45], is useful for such simulations. As an advanced method of REM, the replica-permutation method (RPM) [46] has been developed. This method uses the Suwa–Todo algorithm [47] to permute the temperature between three or more replicas. Replica exchange with solute tempering (REST) [48] and its variants [49–54] have been developed to reduce the number of replicas required in REM. In REST, the target system is divided into a solute and a solvent region. Only the solute temperature, the temperature of the solute region, is virtually exchanged between the two replicas, leaving the temperature of the solvent region unchanged by using the modified potential energy. Recently, the replica permutation with solute tempering (RPST) [55,56] has been developed by combining RPM and REST, allowing only the solute temperature to be permuted between three or more replicas based on the Suwa–Todo algorithm. As a result, further improvements in the efficiency of temperature transitions and structural sampling have been reported. RPST provides statistically reliable data on the conformations of biomolecules.

In this study, we explore the mechanism of the inhibition of Aβ(16–22) peptide aggregation by polyphenols using MD simulations. RPST [55,56] is applied to an aqueous solution system containing Aβ(16–22) peptides and MYC or ROA. Aβ(16–22) peptide contains the central hydrophobic core (Lys-Leu-Val-Phe-Phe) in the full-length Aβ. This core is essential for the aggregation of Aβ [57]. We consider the dimerization of two Aβ(16–22) peptides, which is the early stage of their aggregation. MYC and ROA are polyphenols that are effective in inhibiting the aggregation of Aβ, as mentioned above. The effects of aggregation inhibition by polyphenols are investigated from the conformations obtained by the simulations. We see the effects of the polyphenols on the distance between Aβ(16–22) peptides and the secondary structures of Aβ(16–22) peptides. The detailed interactions are focused on by analyzing contact maps between two Aβ(16–22) peptides and those between Aβ(16–22) peptides and polyphenols. These results show which interactions are important to inhibit the aggregation of the Aβ(16–22) peptides.

Computational Details

We applied RPST with a randomly assigned list [55,56] to systems consisting of two Aβ(16–22) peptide molecules and four polyphenol molecules in an aqueous solution, including five sodium ions and five chloride ions. For the details of RPST, please refer to Refs. [55,56]. The N-termini and the C-termini of the Aβ(16–22) peptide molecules were blocked by the acetyl group and N-methyl group, respectively. A fully extended conformation was used for each Aβ(16–22) peptide as the initial structure. MYC or ROA was used for the polyphenol molecules. These molecular structures are shown in Fig. 1. As a comparison, we also calculated the system without polyphenols. The systems were placed in a cubic unit cell with a side length of 37.3 Å with the periodic boundary conditions.

Figure 1 

Initial conformations of (A) myricetin (MYC) and (B) rosmarinic acid (ROA).

We performed the MD simulations using the Generalized-Ensemble Molecular Biophysics (GEMB) program developed by the authors. The MD simulations were performed for 510.0 ns per replica including the first 10.0 ns as the equilibration period. The number of replicas was four. The solute temperatures T_m of the replicas were distributed in the range of 300.0 to 500.0 K. T₁–T₄ were set at 300.0, 355.7, 421.7, and 500.0 K. The temperature of the system was set to T₀=300 K. We here use the same notation for T_m and T₀ as in references [55,56]. The AMBER parm14SB force field [58] was used for the Aβ(16–22) peptide molecules, the sodium ions, and the chloride ions. The generalized AMBER force fields were applied to the polyphenol molecules. The TIP3P rigid-body model [59] was used for the water molecules. We employed the symplectic [60] quaternion scheme for the rigid-body water molecules [61,62]. The force field validation for the combination of AMBER parm14SB force field and TIP3P water has been demonstrated for Aβ42 dimer in Ref. [63]. It was found that although the combination of AMBER parm14SB force field and TIP3P water overestimates the helix structure for Aβ42, the β-sheet structure is also reproduced. The temperature was controlled by the Nosé–Hoover thermostat [64–66]. Coulomb interactions were calculated using the particle mesh Ewald (PME) method [67,68]. The cutoff distance for the LJ potential energy was set to 12.0 Å. The multiple-time-step method [69] was employed as follows. The time step was set to 0.5 fs for the bonded interactions between peptide atoms, 2.0 fs for the nonbonded interactions between peptide atoms and those between peptide atoms and solvent atoms (water atoms and ions), and 4.0 fs for interactions between solvent atoms. The time step was taken as long as 4.0 fs because the symplectic rigid-body algorithm was used for the water molecules [62]. Trials of replica-permutations were carried out every 1.0 ps. The LJ and real part of the Coulomb energy terms between the Aβ(16–22) peptide atoms and polyphenol atoms were employed as the solute region. Five different initial momenta were prepared with different random numbers for statistical analysis. Only replicas at the solute temperatures T_m of 300 K were used for the data analysis. The mean values and errors of several physical quantities were calculated using the bootstrap method [70] for the five simulations. The number of bootstrap cycles was set to 1×10⁶. The define secondary structure of proteins (DSSP) criteria [71] were utilized to determine the secondary structures of the Aβ(16–22) peptides. As for the Aβ(16–22) peptide system without polyphenols, the trajectory data in reference [55] was used.

Results and Discussion

First, we focused on the distance between Aβ(16–22) peptides to investigate the effect of polyphenols on Aβ aggregation. The probability distribution for the minimum inter-peptide distance P(d_αα) is shown in Fig. 2(A). The minimum inter-peptide distance d_αα is defined as the shortest intermolecular distance between C_α atoms in Aβ(16–22) peptides. The peak is seen at 4.3–4.4 Å in this figure, which means Aβ(16–22) peptides tend to interact with each other. The peak height was reduced in the polyphenol-containing systems, which means a decrease in the fraction of Aβ(16–22) peptides in contact. This decrease was observed in the systems that include either MYC or ROA, with a larger decrease in the system with MYC. Figure 2(B) shows the free-energy landscape for d_αα determined by the following equation:

Figure 2 

(A) Probability distribution of d_αα and (B) free-energy landscape as a function of d_αα in each system.

  

$$F\left(d_{\alpha \alpha }\right)=-\frac{1}{{\beta }_0}\mathrm{l}\mathrm{n}\frac{P\left(d_{\alpha \alpha }\right)}{4\pi {d_{\alpha \alpha }}^2}\text{.}$$(1)

The free-energy landscapes have the global-minimum states at 4.2–4.4 Å. The free-energy difference between the dimer and monomer states is small in the order of the following systems: Aβ(16–22) peptides with MYC<Aβ(16–22) peptides with ROA<Aβ(16–22) peptides without polyphenols. These findings correspond to the result in the experimental study that MYC contributes to the inhibition of Aβ aggregation more than ROA [24].

We also calculated the number of dimerization events of Aβ(16–22) peptides per replica with the same criteria in Ref. Ref. [55], as listed in Table 1. The number of dimerization events is reduced by the existence of MYC or ROA. It means that MYC and ROA inhibit the dimerization of Aβ(16–22) peptides.

Table 1 

The number of dimerization events per replica during the 500 ns simulations

Aβ(16–22)	25.0±2.0
Aβ(16–22) with MYC	14.3±1.4
Aβ(16–22) with ROA	4.4±1.1

Next, we analyzed the effect of polyphenols on the secondary structures of Aβ(16–22) peptides. Figure 3 shows the number of residues forming the secondary structure at the corresponding d_αα. From now on, each data was calculated in 1 Å increments. The result for the intermolecular antiparallel β-sheet is shown in Fig. 3(A). It can be seen that Aβ(16–22) peptides aggregate by forming the intermolecular antiparallel β-sheet as Aβ(16–22) peptides approach each other in each system. In other simulation works for Aβ(16–22) peptides, the intermolecular antiparallel β-strand structures were also mainly observed in the oligomer states [15,72–75]. When MYC or ROA is included in the system of Aβ(16–22) peptides, the number of residues forming the β-sheet is reduced. This result means a decrease in the stable Aβ(16–22) aggregates by the presence of polyphenols. We focused on the number of residues forming the helix structure in Fig. 3(B). When d_αα≥6 Å, the helix structure is formed more. The helix structure was more likely to form when d_αα≥5 Å in the system with MYC compared to that without polyphenols. The increase in helix structures makes it more difficult for the peptides to form intermolecular hydrogen bonds. It allows us to understand why MYC is effective in pulling the distance between Aβ(16–22) peptides apart, as shown in Fig. 2.

Figure 3 

Ensemble averages of the numbers of residues that have the (A) intermolecular antiparallel β-sheet and (B) helix structures at the corresponding d_αα are shown for each system. The data for the Aβ(16–22) peptide system without polyphenols in panel (a) was replotted with permission from Ref. [55]. Copyright 2022 AIP Publishing.

We have so far discussed the probability distribution and conformational changes using a single reaction coordinate, d_αα. Here we use an additional parameter to describe the binding process better. We employ the end-to-end distance as the second reaction coordinate, which was calculated as the distance between two C_α atoms of Lys16 and Glu22. The two-dimensional probability distributions are shown in Fig. 4. There is a higher probability distribution with d_αα≥6 Å in the presence of MYC or ROA than in the absence of these polyphenols. It means that the two Aβ(16–22) peptides are less likely to bind with each other in the presence of these polyphenols. In the absence of polyphenols, the structure with d_αα=4 Å (i.e. d_αα=3.5–4.5 Å) has two peaks at the end-to-end distances of 11–13 Å and 18–19 Å. The peak height at 18–19 Å was considerably reduced in the MYC and ROA systems. A conformation with the end-to-end distance of 18–19 Å is an elongated structure, so it is easy to form an intermolecular β-sheet structure, as the snapshot in Fig. 3(A). It means it is difficult for two Aβ(16–22) molecules to form an intermolecular β-sheet structure even when they are close in the MYC and ROA systems.

Figure 4 

Two-dimensional probability distribution as a function of the minimum inter-peptide distance d_αα and end-to-end distance in the systems consisting of (A) Aβ(16–22) peptides, (B) Aβ(16–22) peptides with MYC, and (C) Aβ(16–22) peptides with ROA.

To focus on the detailed interactions between Aβ(16–22) peptides in the dimer state, the intermolecular side-chain contact maps at d_αα=4 Å (i.e. d_αα=3.5–4.5 Å) are shown in Fig. 5. This is because this range of d_αα includes the peaks of the distributions in Fig. 2(A) and the two Aβ(16–22) peptides can be regarded as a dimer. When the shortest distance between heavy atoms of the side chains of a residue in one peptide and those in the other peptide was less than 4.0 Å, we regarded it as an intermolecular side-chain contact. These contact maps show which residue pairs form the intermolecular side-chain contacts when two Aβ(16–22) peptides exist close to each other. Figure 5(A) shows the contact map for the system without polyphenols. The hydrophilic side-chain contacts between Lys16 and Glu22 and the hydrophobic side-chain contacts, particularly between Val18 and Phe20 and between Leu17 and Phe19, are formed. These side-chain contacts are also seen in other works [15,76]. In a simulation study, the antiparallel β-sheet of Aβ(16–22) peptides was observed due to the electrostatic attraction between the side chains of Lys16 and Glu22 [15]. An experimental study indicated that the β-strand structure extends across the hydrophobic segment from Leu17 through Ala21 of Aβ(16–22) peptides [76]. As shown even in these previous studies, these contacts facilitate the formation of the antiparallel β-bridges between Aβ(16–22) peptides. In Fig. 5(B), some hydrophobic side-chain contacts decreased in the system with MYC compared to that without polyphenols in Fig. 5(A). The hydrophilic side-chain contacts were not reduced. In contrast, the hydrophilic interactions decreased, and hydrophobic interactions between Leu17 and Phe20 increased for the system with ROA in Fig. 5(C). At d_αα=4 Å, two Aβ(16–22) peptides almost always form contacts between any of the side chains. Therefore, Fig. 5 shows which residues are likely to form intermolecular side-chain contacts when two Aβ(16–22) peptides exist nearby. In the absence of polyphenols, contacts are not inhibited and are formed both between hydrophobic residues and between hydrophilic residues. Most of the hydrophobic contacts are inhibited in the system with MYC. In the presence of ROA, contacts between hydrophilic residues are inhibited, but contacts are formed only between hydrophobic residues, especially between Leu17 and Phe20. A representative snapshot where the contact between Leu17 and Phe20 was formed is shown in Fig. 6. The conformation does not form hydrogen bonds between main chains. As seen in this structure, the β-sheet formation is not promoted in response to the increase of the hydrophobic side-chain interaction in the system with ROA because ROA penetrates between these hydrophobic residues and inhibits Aβ(16–22) peptides from aggregating. It results in the aggregates being less likely to form amyloid fibrils. Except in this case, side-chain contacts between Aβ(16–22) peptides are reduced for the systems with polyphenols, although the interactions between Aβ(16–22) peptides are different depending on whether MYC or ROA is included in the system.

Figure 5 

Intermolecular side-chain contact maps at d_αα=4 Å in the systems consisting of (A) Aβ(16–22) peptides, (B) Aβ(16–22) peptides with MYC, and (C) Aβ(16–22) peptides with ROA. The abscissa and ordinate are the residues of the Aβ(16–22) peptide. Because the two Aβ(16–22) peptides are identical, the probability distributions were calculated, taking this symmetry into account.

Figure 6 

A representative conformation of the Aβ(16–22) peptides and ROA at d_αα=4 Å when a side-chain contact was formed between Leu17 and Phe20.

We then focused on the interactions between Aβ(16–22) peptides and polyphenols. The intermolecular contact maps between side chains of Aβ(16–22) peptides and polyphenols at d_αα=4 Å (i.e. d_αα=3.5–4.5 Å) are shown in Fig. 7. When the shortest distance between heavy atoms of the side chains of a residue and those of a polyphenol molecule was less than 4.0 Å, an intermolecular contact was regarded as formed. The label for each heavy atom of polyphenols is shown in Fig. 1. As seen in Fig. 7(A), O18, O19, and O20 of MYC interact with Glu22 of Aβ(16–22) with the highest probabilities in the system containing MYC. C1–C9 of MYC also interact with Phe19 and Phe20 with high probability. This result is consistent with our previous MD simulations on Aβ(16–22) monomer: a system containing one Aβ(16–22) molecule and four MYC molecules showed that the Aβ(16–22) residue most likely to bind MYC was Glu22, and Phe19 and Phe20 also bound MYC relatively well [25]. Our present result means that MYC interacted hydrophobically with phenylalanine at the benzene ring and hydrophilically with glutamic acid at the hydroxyl group. The result can explain that the decrease in hydrophobic interactions between Aβ(16–22) peptides is due to the interaction between MYC and the phenylalanine of Aβ(16–22) peptides. On the other hand, Fig. 7(B) shows the contact map regarding the system with ROA. The hydroxyl group (O3, O5, O22, and O24) of ROA tended to interact with Glu22, and the carboxylate group (O12 and O13) interacted with Lys16. These results are also consistent with our previous work on Aβ(16–22) monomer [25]. In the present study, we found that ROA interacted with both hydrophilic residues of Aβ(16–22) peptides, glutamic acid, and lysine, thereby reducing the hydrophilic interaction between Aβ(16–22) peptides. Note that ROA is less likely to approach Phe20 compared to MYC. This is because the benzene rings in ROA are kept away from Phe20 due to the stronger interaction between the hydroxyl groups on the benzene rings and Glu22.

Figure 7 

Intermolecular contact maps between polyphenol atoms and atoms of the side chains of Aβ(16–22) peptides at d_αα=4 Å in the systems consisting of (A) Aβ(16–22) peptides with MYC and (B) Aβ(16–22) peptides with ROA. The abscissa represents the polyphenol atoms, and the ordinate represents the residues of Aβ(16–22) peptide.

Some representative snapshots are shown in Fig. 8, summarizing the results in this study. Figure 8(A) shows the structure of Aβ(16–22) peptides forming the β-strand to aggregate. In the snapshot of Fig. 8(B), MYC accesses Aβ(16–22) peptide by the hydrophilic interactions between the hydroxyl group of MYC and Glu22 and the hydrophobic interaction between the benzene ring of MYC and Phe19. We see that these hydrophilic and hydrophobic interactions facilitate the formation of helix structures. Therefore, in the presence of MYC, Aβ(16–22) peptides are less likely to aggregate as a sheet structure, as shown in Fig. 8(A). Figure 8(C) shows the structure of ROA interacting with Aβ(16–22) peptide, where the hydroxyl group and carboxylate group of ROA interact hydrophilically with Glu22 and Lys16, respectively. These interactions inhibit the hydrophilic interactions between Aβ(16–22) peptides, which results in the inhibition of aggregation between Aβ(16–22) peptides.

Figure 8 

Snapshots of representative conformations at d_αα=4 Å of (A) the Aβ(16–22) peptides, (B) the Aβ(16–22) peptide and MYC, and (C) the Aβ(16–22) peptide and ROA.

Conclusion

Polyphenols are known to inhibit Aβ peptide aggregation. Elucidation of the mechanism contributes to developing therapeutic agents for Alzheimer’s disease caused by Aβ peptide aggregation. In this study, RPST was applied to systems of Aβ(16–22) peptides and polyphenols, myricetin (MYC) or rosmarinic acid (ROA). We analyzed the structures and focused on the interactions between Aβ(16–22) peptides and between Aβ(16–22) peptides and polyphenols. The aggregated structure of Aβ(16–22) peptides and the binding structures of Aβ(16–22) peptides and polyphenols were clarified at the atomic level.

The distributions of the distance between Aβ(16–22) peptides d_αα indicated that polyphenols tend to keep Aβ(16–22) peptides away from each other. Analysis on the secondary structures when d_αα≤5 Å shows that both polyphenols reduce the number of residues forming the intermolecular antiparallel β-sheet. In particular, it was suggested that MYC can induce the helix structures of Aβ(16–22) peptides. We investigated the side-chain contact maps between Aβ(16–22) peptides and found that MYC reduces hydrophobic interactions between Aβ fragments and ROA reduces hydrophilic interactions between Aβ fragments. As shown in the contact maps between polyphenols and Aβ(16–22) peptides, MYC accessed to Aβ(16–22) peptides by the hydrophilic interactions between hydroxyl groups of MYC and glutamic acid and the hydrophobic interactions between benzene rings of MYC and phenylalanine. On the other hand, ROA interacted with Aβ(16–22) peptides by the hydrophilic contacts between the hydroxyl group and glutamic acid and those between the carboxylate group and lysine. Although the inhibition mechanisms of MYC and ROA are thus different, both polyphenols inhibit Aβ(16–22) aggregation. However, it was also clarified that MYC inhibits aggregation more. The reason can be interpreted as follows: Aβ(16–22) peptides form antiparallel β-sheets due to the electrostatic attraction between Lys16 and Glu22 and the hydrophobic interaction between Lys17–Ala21 residues. MYC attaches to the Lys17–Ala21 hydrophobic residues, too, in the central part of the Aβ(16–22) peptide. On the other hand, ROA interacts mainly with the N- and C-terminal residues, Lys16 and Glu22 only. Thus, MYC inhibits Aβ(16–22) aggregation more. Based on this finding, it is expected that more potent aggregation inhibitors of Aβ and other fibrillizing proteins can be explored, which could lead to advances in treating diseases associated with protein misfolding and aggregation.

Conflict of Interest

D. F., S. G. I., and H. O. declare that they have no conflict of interest.

Author Contributions

S. G. I. and H. O. designed the research. D. F. performed the simulations, analyzed the results, and wrote the paper. All authors discussed the results and revised the paper.

Data Availability

The evidence data generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Acknowledgements

The computations were performed using supercomputers at Research Center for Computational Science, Okazaki Research Facilities, National Institutes of Natural Sciences, Japan (21-IMS-C172, 22-IMS-C186, and 23-IMS-C198) and the Institute of Statistical Mathematics (ISM), Research Organization of Information and Systems, Japan. This study was carried out under the ISM General Cooperative Research 1 (2021-ISMCRP-1009, 2022-ISMCRP-1002, and 2023-ISMCRP-1003).

References

• [1]   Koffie,  R. M.,  Meyer-Luehmann,  M.,  Hashimoto,  T.,  Adams,  K. W.,  Mielke,  M. L.,  Garcia-Alloza,  M., et al. Oligomeric amyloid beta associates with postsynaptic densities and correlates with excitatory synapse loss near senile plaques. Proc. Natl. Acad. Sci. U.S.A. 106, 4012–4017 (2009). https://doi.org/10.1073/pnas.0811698106
• [2]   Lührs,  T.,  Ritter,  C.,  Adrian,  M.,  Riek-Loher,  D.,  Bohrmann,  B.,  Döbeli,  H., et al. 3D structure of Alzheimer’s amyloid-&#x3b2;(1&#x2013;42) fibrils. Proc. Natl. Acad. Sci. U.S.A. 102, 17342–17347 (2005). https://doi.org/10.1073/pnas.0506723102
• [3]   Petkova,  A. T.,  Yau,  W.-M.,  Tycko,  R. Experimental constraints on quaternary structure in Alzheimer’s β-amyloid fibrils. Biochemistry 45, 498–512 (2006). https://doi.org/10.1021/bi051952q
• [4]   Bernstein,  S. L.,  Dupuis,  N. F.,  Lazo,  N. D.,  Wyttenbach,  T.,  Condron,  M. M.,  Bitan,  G., et al. Amyloid-β protein oligomerization and the importance of tetramers and dodecamers in the aetiology of Alzheimer’s disease. Nat. Chem. 1, 326–331 (2009). https://doi.org/10.1038/nchem.247
• [5]   Viet,  M. H.,  Li,  M. S. Amyloid peptide Aβ40 inhibits aggregation of Aβ42: Evidence from molecular dynamics simulations. J. Chem. Phys. 136, 245105 (2012). https://doi.org/10.1063/1.4730410
• [6]   Itoh,  S. G.,  Okumura,  H. Coulomb replica-exchange method: handling electrostatic attractive and repulsive forces for biomolecules. J. Comput. Chem. 34, 622–639 (2013). https://doi.org/10.1002/jcc.23167
• [7]   Okumura,  H.,  Itoh,  S. G. Amyloid fibril disruption by ultrasonic cavitation: Nonequilibrium molecular dynamics simulations. J. Am. Chem. Soc. 136, 10549–10552 (2014). https://doi.org/10.1021/ja502749f
• [8]   Itoh,  S. G.,  Okumura,  H. Dimerization process of amyloid-β(29–42) studied by the Hamiltonian replica-permutation molecular dynamics simulations. J. Phys. Chem. B 118, 11428–11436 (2014). https://doi.org/10.1021/jp505984e
• [9]   Tycko,  R. Amyloid polymorphism: Structural basis and neurobiological relevance. Neuron 86, 632–645 (2015). https://doi.org/10.1016/j.neuron.2015.03.017
• [10]   Okumura,  H.,  Itoh,  S. G. Structural and fluctuational difference between two ends of Aβ amyloid fibril: MD simulations predict only one end has open conformations. Sci. Rep. 6, 38422 (2016). https://doi.org/10.1038/srep38422
• [11]   Itoh,  S. G.,  Okumura,  H. Oligomer formation of Amyloid-β(29–42) from its monomers using the hamiltonian replica-permutation molecular dynamics simulation. J. Phys. Chem. B 120, 6555–6561 (2016). https://doi.org/10.1021/acs.jpcb.6b03828
• [12]   Dutta,  M.,  Mattaparthi,  V. S. K. In silico investigation on the inhibition of Aβ(42) aggregation by Aβ (40) peptide by potential of mean force study. J. Biomol. Struct. Dyn. 36, 741–752 (2018). https://doi.org/10.1080/07391102.2017.1296783
• [13]   Bhattacharya,  S.,  Xu,  L.,  Thompson,  D. Molecular simulations reveal terminal group mediated stabilization of helical conformers in both Amyloid-β42 and α-Synuclein. ACS Chem. Neurosci. 10, 2830–2842 (2019). https://doi.org/10.1021/acschemneuro.9b00053
• [14]   Itoh,  S. G.,  Yagi-Utsumi,  M.,  Kato,  K.,  Okumura,  H. Effects of a hydrophilic/hydrophobic interface on Amyloid-β peptides studied by molecular dynamics simulations and NMR experiments. J. Phys. Chem. B 123, 160–169 (2019). https://doi.org/10.1021/acs.jpcb.8b11609
• [15]   Okumura,  H.,  Itoh,  S. G. Molecular dynamics simulations of amyloid-β (16–22) peptide aggregation at air-water interfaces. J. Chem. Phys. 152, 095101 (2020). https://doi.org/10.1063/1.5131848
• [16]   Tachi,  Y.,  Okamoto,  Y.,  Okumura,  H. Conformational change of amyloid-β40 in association with binding to GM1-glycan cluster. Sci. Rep. 9, 6853 (2019). https://doi.org/10.1038/s41598-019-43117-6
• [17]   Okumura,  H.,  Itoh,  S. G.,  Nakamura,  K.,  Kawasaki,  T. Role of water molecules and helix structure stabilization in the laser-induced disruption of amyloid fibrils observed by nonequilibrium molecular dynamics simulations. J. Phys. Chem. B 125, 4964–4976 (2021). https://doi.org/10.1021/acs.jpcb.0c11491
• [18]   Nguyen,  H. L.,  Linh,  H. Q.,  Krupa,  P.,  La Penna,  G.,  Li,  M. S. Amyloid β dodecamer disrupts the neuronal membrane more strongly than the mature fibril: Understanding the role of oligomers in neurotoxicity. J. Phys. Chem. B 126, 3659–3672 (2022). https://doi.org/10.1021/acs.jpcb.2c01769
• [19]   Zhang,  D. Y.,  Wang,  J.,  Fleeman,  R. M.,  Kuhn,  M. K.,  Swulius,  M. T.,  Proctor,  E. A., et al. Monosialotetrahexosylganglioside promotes early Aβ42 oligomer formation and maintenance. ACS Chem. Neurosci. 13, 1979–1991 (2022). https://doi.org/10.1021/acschemneuro.2c00221
• [20]   Itoh,  S. G.,  Yagi-Utsumi,  M.,  Kato,  K.,  Okumura,  H. Key residue for aggregation of Amyloid-β peptides. ACS Chem. Neurosci. 13, 3139–3151 (2022). https://doi.org/10.1021/acschemneuro.2c00358
• [21]   Yagi-Utsumi,  M.,  Itoh,  S. G.,  Okumura,  H.,  Yanagisawa,  K.,  Kato,  K.,  Nishimura,  K. The double-layered structure of Amyloid-β assemblage on GM1-containing membranes catalytically promotes fibrillization. ACS Chem. Neurosci. 14, 2648–2657 (2023). https://doi.org/10.1021/acschemneuro.3c00192
• [22]   Okumura,  H.,  Itoh,  S. G.,  Zen,  H.,  Nakamura,  K. Dissociation process of polyalanine aggregates by free electron laser irradiation. PLoS One 18, e0291093 (2023). https://doi.org/10.1371/journal.pone.0291093
• [23]   Ono,  K.,  Yoshiike,  Y.,  Takashima,  A.,  Hasegawa,  K.,  Naiki,  H.,  Yamada,  M. Potent anti-amyloidogenic and fibril-destabilizing effects of polyphenols in vitro: Implications for the prevention and therapeutics of Alzheimer’s disease. J. Neurochem. 87, 172–181 (2003). https://doi.org/10.1046/j.1471-4159.2003.01976.x
• [24]   Ono,  K.,  Li,  L.,  Takamura,  Y.,  Yoshiike,  Y.,  Zhu,  L.,  Han,  F., et al. Phenolic compounds prevent amyloid β-protein oligomerization and synaptic dysfunction by site-specific binding. J. Biol. Chem. 287, 14631–14643 (2012). https://doi.org/10.1074/jbc.M111.325456
• [25]   Ngoc,  L. L. N.,  Itoh,  S. G.,  Sompornpisut,  P.,  Okumura,  H. Replica-permutation molecular dynamics simulations of an amyloid-β (16–22) peptide and polyphenols. Chem. Phys. Lett. 758, 137913 (2020). https://doi.org/10.1016/j.cplett.2020.137913
• [26]   Bieschke,  J.,  Russ,  J.,  Friedrich,  R. P.,  Ehrnhoefer,  D. E.,  Wobst,  H.,  Neugebauer,  K., et al. EGCG remodels mature α-synuclein and amyloid-β fibrils and reduces cellular toxicity. Proc. Natl. Acad. Sci. U.S.A. 107, 7710–7715 (2010). https://doi.org/10.1073/pnas.0910723107
• [27]   Ehrnhoefer,  D. E.,  Bieschke,  J.,  Boeddrich,  A.,  Herbst,  M.,  Masino,  L.,  Lurz,  R., et al. EGCG redirects amyloidogenic polypeptides into unstructured, off-pathway oligomers. Nat. Struct. Mol. Biol. 15, 558–566 (2008). https://doi.org/10.1038/nsmb.1437
• [28]   Zhang,  T.,  Zhang,  J.,  Derreumaux,  P.,  Mu,  Y. Molecular mechanism of the inhibition of EGCG on the Alzheimer Abeta (1–42) dimer. J. Phys. Chem. B 117, 3993–4002 (2013). https://doi.org/10.1021/jp312573y
• [29]   Ngo,  S. T.,  Truong,  D. T.,  Tam,  N. M.,  Nguyen,  M. T. EGCG inhibits the oligomerization of amyloid beta (16–22) hexamer: Theoretical studies. J. Mol. Graph. Model. 76, 1–10 (2017). https://doi.org/10.1016/j.jmgm.2017.06.018
• [30]   Li,  F.,  Zhan,  C.,  Dong,  X.,  Wei,  G. Molecular mechanisms of resveratrol and EGCG in the inhibition of Aβ42 aggregation and disruption of Aβ42 protofibril: similarities and differences. Phys. Chem. Chem. Phys. 23, 18843–18854 (2021). https://doi.org/10.1039/D1CP01913A
• [31]   Shahrivar-Gargari,  M.,  Hamzeh-Mivehroud,  M.,  Hemmati,  S.,  Shahbazi Mojarrad,  J.,  Notash,  B.,  Tüylü Küçükkılınç,  T., et al. Design, synthesis, and biological evaluation of novel indanone-based hybrids as multifunctional cholinesterase inhibitors for Alzheimer’s disease. J. Mol. Struct. 1229, 129787 (2021). https://doi.org/10.1016/j.molstruc.2020.129787
• [32]   Paul,  R.,  Bera,  S.,  Devi,  M.,  Paul,  S. Inhibition of Aβ16–22 peptide aggregation by small molecules and their permeation through POPC lipid bilayer: Insight from molecular dynamics simulation study. J. Chem. Inf. Model. 62, 5193–5207 (2022). https://doi.org/10.1021/acs.jcim.1c01366
• [33]   Morel,  I.,  Abaléa,  V.,  Sergent,  O.,  Cillard,  P.,  Cillard,  J. Involvement of phenoxyl radical intermediates in lipid antioxidant action of myricetin in iron-treated rat hepatocyte culture. Biochem. Pharmacol. 55, 1399–1404 (1998). https://doi.org/10.1016/S0006-2952(97)00648-5
• [34]   Bennett,  C. J.,  Caldwell,  S. T.,  McPhail,  D. B.,  Morrice,  P. C.,  Duthie,  G. G.,  Hartley,  R. C. Potential therapeutic antioxidants that combine the radical scavenging ability of myricetin and the lipophilic chain of vitamin E to effectively inhibit microsomal lipid peroxidation. Bioorg. Med. Chem. 12, 2079–2098 (2004). https://doi.org/10.1016/j.bmc.2004.02.031
• [35]   Ghorbani,  M.,  Soleymani,  H.,  Allahverdi,  A.,  Shojaeilangari,  S.,  Naderi-Manesh,  H. Effects of natural compounds on conformational properties and hairpin formation of amyloid-β (42) monomer: Docking and molecular dynamics simulation study. J. Biomol. Struct. Dyn. 38, 3371–3383 (2020). https://doi.org/10.1080/07391102.2019.1664934
• [36]   Hukushima,  K.,  Nemoto,  K. Exchange Monte Carlo method and application to spin glass simulations. J. Phys. Soc. Jpn. 65, 1604–1608 (1996). https://doi.org/10.1143/Jpsj.65.1604
• [37]   Sugita,  Y.,  Okamoto,  Y. Replica-exchange molecular dynamics method for protein folding. Chem. Phys. Lett. 314, 141–151 (1999). https://doi.org/10.1016/S0009-2614(99)01123-9
• [38]   Berg,  B. A.,  Neuhaus,  T. Multicanonical algorithms for 1st order phase-transitions. Phys. Lett. B 267, 249–253 (1991). https://doi.org/10.1016/0370-2693(91)91256-u
• [39]   Berg,  B. A.,  Neuhaus,  T. Multicanonical ensemble - a new approach to simulate 1st-order phase-transitions. Phys. Rev. Lett. 68, 9–12 (1992). https://doi.org/10.1103/PhysRevLett.68.9
• [40]   Hansmann,  U. H. E.,  Okamoto,  Y.,  Eisenmenger,  F. Molecular dynamics, langevin and hybrid Monte Carlo simulations in a multicanonical ensemble. Chem. Phys. Lett. 259, 321–330 (1996). https://doi.org/10.1016/0009-2614(96)00761-0
• [41]   Lyubartsev,  A. P.,  Martsinovski,  A. A.,  Shevkunov,  S. V.,  Vorontsovvelyaminov,  P. N. New approach to Monte-Carlo calculation of the free-energy - method of expanded ensembles. J. Chem. Phys. 96, 1776–1783 (1992). https://doi.org/10.1063/1.462133
• [42]   Marinari,  E.,  Parisi,  G. Simulated tempering - a new Monte-Carlo scheme. Europhys. Lett. 19, 451–458 (1992). https://doi.org/10.1209/0295-5075/19/6/002
• [43]   Mitsutake,  A.,  Sugita,  Y.,  Okamoto,  Y. Generalized-ensemble algorithms for molecular simulations of biopolymers. Biopolymers 60, 96–123 (2001). https://doi.org/10.1002/1097-0282(2001)60:2<96::AID-BIP1007>3.0.Co;2-f
• [44]   Nakajima,  N.,  Nakamura,  H.,  Kidera,  A. Multicanonical ensemble generated by molecular dynamics simulation for enhanced conformational sampling of peptides. J. Phys. Chem. B 101, 817–824 (1997). https://doi.org/10.1021/jp962142e
• [45]   Yamauchi,  M.,  Mori,  Y.,  Okumura,  H. Molecular simulations by generalized-ensemble algorithms in isothermal-isobaric ensemble. Biophys. Rev. 11, 457–469 (2019). https://doi.org/10.1007/s12551-019-00537-y
• [46]   Itoh,  S. G.,  Okumura,  H. Replica-permutation method with the suwa-todo algorithm beyond the replica-exchange method. J. Chem. Theory Comput. 9, 570–581 (2013). https://doi.org/10.1021/ct3007919
• [47]   Suwa,  H.,  Todo,  S. Markov chain Monte Carlo method without detailed balance. Phys. Rev. Lett. 105, 120603 (2010). https://doi.org/10.1103/PhysRevLett.105.120603
• [48]   Liu,  P.,  Kim,  B.,  Friesner,  R. A.,  Berne,  B. J. Replica exchange with solute tempering: a method for sampling biological systems in explicit water. Proc. Natl. Acad. Sci. U.S.A. 102, 13749–13754 (2005). https://doi.org/10.1073/pnas.0506346102
• [49]   Best,  R. B.,  Zhu,  X.,  Shim,  J.,  Lopes,  P. E.,  Mittal,  J.,  Feig,  M., et al. Optimization of the additive CHARMM all-atom protein force field targeting improved sampling of the backbone phi, psi and side-chain chi(1) and chi(2) dihedral angles. J. Chem. Theory Comput. 8, 3257–3273 (2012). https://doi.org/10.1021/ct300400x
• [50]   Camilloni,  C.,  Provasi,  D.,  Tiana,  G.,  Broglia,  R. A. Exploring the protein G helix free-energy surface by solute tempering metadynamics. Proteins 71, 1647–1654 (2008). https://doi.org/10.1002/prot.21852
• [51]   Moors,  S. L. C.,  Michielssens,  S.,  Ceulemans,  A. Improved replica exchange method for native-state protein sampling. J. Chem. Theory Comput. 7, 231–237 (2011). https://doi.org/10.1021/ct100493v
• [52]   Terakawa,  T.,  Kameda,  T.,  Takada,  S. On easy implementation of a variant of the replica exchange with solute tempering in GROMACS. J. Comput. Chem. 32, 1228–1234 (2011). https://doi.org/10.1002/jcc.21703
• [53]   Wang,  L.,  Friesner,  R. A.,  Berne,  B. J. Replica exchange with solute scaling: a more efficient version of replica exchange with solute tempering (REST2). J. Phys. Chem. B 115, 9431–9438 (2011). https://doi.org/10.1021/jp204407d
• [54]   Kamiya,  M.,  Sugita,  Y. Flexible selection of the solute region in replica exchange with solute tempering: Application to protein-folding simulations. J. Chem. Phys. 149, 072304 (2018). https://doi.org/10.1063/1.5016222
• [55]   Fukuhara,  D.,  Itoh,  S. G.,  Okumura,  H. Replica permutation with solute tempering for molecular dynamics simulation and its application to the dimerization of amyloid-β fragments. J. Chem. Phys. 156, 084109 (2022). https://doi.org/10.1063/5.0081686
• [56]   Fukuhara,  D.,  Yamauchi,  M.,  Itoh,  S. G.,  Okumura,  H. Ingenuity in performing replica permutation: How to order the state labels for improving sampling efficiency. J. Comput. Chem. 44, 534–545 (2023). https://doi.org/10.1002/jcc.27020
• [57]   Tjernberg,  L. O.,  Naeslund,  J.,  Lindqvist,  F.,  Johansson,  J.,  Karlstroem,  A. R.,  Thy-berg,  J., et al. Arrest of amyloid fibril formation by a pentapeptide ligand. J. Biol. Chem. 271, 8545–8548 (1996). https://doi.org/10.1074/jbc.271.15.8545
• [58]   Maier,  J. A.,  Martinez,  C.,  Kasavajhala,  K.,  Wickstrom,  L.,  Hauser,  K. E.,  Simmerling,  C. ff14SB: Improving the accuracy of protein side chain and backbone parameters from ff99SB. J. Chem. Theory Comput. 11, 3696–3713 (2015). https://doi.org/10.1021/acs.jctc.5b00255
• [59]   Jorgensen,  W. L.,  Chandrasekhar,  J.,  Madura,  J. D.,  Impey,  R. W.,  Klein,  M. L. Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 79, 926–935 (1983). https://doi.org/10.1063/1.445869
• [60]   Yoshida,  H. Construction of higher order symplectic integrators. Phys. Lett. A 150, 262–268 (1990). https://doi.org/10.1016/0375-9601(90)90092-3
• [61]   Miller,  T.,  Eleftheriou,  M.,  Pattnaik,  P.,  Ndirango,  A.,  Newns,  D.,  Martyna,  G. Symplectic quaternion scheme for biophysical molecular dynamics. J. Chem. Phys. 116, 8649–8659 (2002). https://doi.org/10.1063/1.1473654
• [62]   Okumura,  H.,  Itoh,  S. G.,  Okamoto,  Y. Explicit symplectic integrators of molecular dynamics algorithms for rigid-body molecules in the canonical, isobaric-isothermal, and related ensembles. J. Chem. Phys. 126, 084103 (2007). https://doi.org/10.1063/1.2434972
• [63]   Man,  V. H.,  Nguyen,  P. H.,  Derreumaux,  P. High-resolution structures of the Amyloid-β 1–42 dimers from the comparison of four atomistic force fields. J. Phys. Chem. B 121, 5977–5987 (2017). https://doi.org/10.1021/acs.jpcb.7b04689
• [64]   Hoover,  W. G. Canonical dynamics: Equilibrium phase-space distributions. Phys. Rev. A 31, 1695–1697 (1985). https://doi.org/10.1103/physreva.31.1695
• [65]   Nosé,  S. A molecular dynamics method for simulations in the canonical ensemble. Mol. Phys. 52, 255–268 (1984). https://doi.org/10.1080/00268978400101201
• [66]   Nosé,  S. A unified formulation of the constant temperature molecular dynamics methods. J. Chem. Phys. 81, 511–519 (1984). https://doi.org/10.1063/1.447334
• [67]   Darden,  T.,  York,  D.,  Pedersen,  L. Particle mesh Ewald: An N⋅log (N) method for Ewald sums in large systems. J. Chem. Phys. 98, 10089–10092 (1993). https://doi.org/10.1063/1.464397
• [68]   Essmann,  U.,  Perera,  L.,  Berkowitz,  M. L.,  Darden,  T.,  Lee,  H.,  Pedersen,  L. G. A smooth particle mesh Ewald method. J. Chem. Phys. 103, 8577–8593 (1995). https://doi.org/10.1063/1.470117
• [69]   Tuckerman,  M.,  Berne,  B. J.,  Martyna,  G. J. Reversible multiple time scale molecular dynamics. J. Chem. Phys. 97, 1990–2001 (1992). https://doi.org/10.1063/1.463137
• [70]   Efron,  B. Bootstrap methods: Another look at the jackknife. Ann. Stat. 7, 1–26 (1979).
• [71]   Kabsch,  W.,  Sander,  C. Dictionary of protein secondary structure: Pattern recognition of hydrogen-bonded and geometrical features. Biopolymers 22, 2577–2637 (1983). https://doi.org/10.1002/bip.360221211
• [72]   Nguyen,  P. H.,  Li,  M. S.,  Derreumaux,  P. Effects of all-atom force fields on amyloid oligomerization: replica exchange molecular dynamics simulations of the Abeta (16–22) dimer and trimer. Phys. Chem. Chem. Phys. 13, 9778–9788 (2011). https://doi.org/10.1039/c1cp20323a
• [73]   Gnanakaran,  S.,  Nussinov,  R.,  Garcia,  A. E. Atomic-level description of amyloid beta-dimer formation. J. Am. Chem. Soc. 128, 2158–2159 (2006). https://doi.org/10.1021/ja0548337
• [74]   Klimov,  D. K.,  Thirumalai,  D. Dissecting the assembly of Abeta16-22 amyloid peptides into antiparallel beta sheets. Structure 11, 295–307 (2003). https://doi.org/10.1016/s0969-2126(03)00031-5
• [75]   Nguyen,  P. H.,  Li,  M. S.,  Stock,  G.,  Straub,  J. E.,  Thirumalai,  D. Monomer adds to preformed structured oligomers of Abeta-peptides by a two-stage dock-lock mechanism. Proc. Natl. Acad. Sci. U.S.A. 104, 111–116 (2007). https://doi.org/10.1073/pnas.0607440104
• [76]   Balbach,  J. J.,  Ishii,  Y.,  Antzutkin,  O. N.,  Leapman,  R. D.,  Rizzo,  N. W.,  Dyda,  F., et al. Amyloid fibril formation by A beta 16–22, a seven-residue fragment of the Alzheimer’s beta-amyloid peptide, and structural characterization by solid state NMR. Biochemistry 39, 13748–13759 (2000). https://doi.org/10.1021/bi0011330