2.1 Preparation of proteins and ligands

The 3-D coordinates of target peptides (2BEG: 3D structure of Aβ_1-42 proto-filament, 2MXU: a S-shaped 42-residue β-amyloid dodecamer, and 2NAO: a disease-relevant Aβ_1-42 amyloid fibril) were acquired from the PDB RCSB database (https://www.rcsb.org). AutoDock Tools (ADT) (https://autodocksuite.scripps.edu/adt/) was employed to prepare protein for molecular docking by adding Kollman charges. It should be noted that there no water molecules and heteroatoms in the structure file of target proteins. Finally, Merck Molecular Force Field (MMFF) was used to reduce the energy of protein molecules. The 2D structures of ligands were downloaded from PubChem (Table 1) (https://pubchem.ncbi.nlm.nih.gov). Ligands were optimized for docking process by designating bond ordering and degrees using ADT. The component charges of Gasteiger were established. Universal Force Field (UFF) was employed to minimize the energy levels of the selected ligands using PyRx (https://pyrx.sourceforge.io).

2.2 Molecular docking studies

Molecular docking was achieved utilizing Autodock-Vina enabled PyRx (https://pyrx.sourceforge.io), where the Lamarckian Genetic Algorithm (LGA) was employed as a scoring function (Trott and Olson, 2010). The molecular docking of ligands with target proteins namely 2BEG, 2MXU, and 2NAO was performed inside a grid-box of dimensions (46 × 23 × 27), (68 × 50 × 66), and (35 × 70 × 64) Å, respectively. The most optimal configuration of each ‘protein–ligand complex’ was created and evaluated using Discovery Studio 2020 (BIOVIA) (https://discover.3ds.com/discovery-studio-visualizer-download). The ligands’ dissociation coefficient (Kd) for the target protein was determined from binding energy (ΔG) employing the following formula (Rehman et al., 2014):

2.3 Molecular dynamics simulation

The stability of the 2BEG-Diosmetin, 2MXU-Rutin, and 2NAO-Genistein complexes was evaluated using GROMACS 2022.4 (https://zenodo.org/records/6451564) for molecular dynamics (MD) simulations. Protein structures were generated with the CHARMM36 force field and TIP3P water models, and the CHARMM General Force Field (CGenFF) was used for the compounds. Simulations were carried out in a dodecahedron-shaped box with the protein–ligand complex centered and at least 1.0 nm from the edges, using TIP3P water and Na⁺ or Cl^- ions to neutralize and mimic physiological conditions with 150 mM NaCl. Energy minimization involved up to 50,000 “steepest descent” steps. The system was equilibrated in NVT and NPT ensembles at 300 K and 1.0 bar, respectively, with temperature and pressure controlled by the Berendsen thermostat and Parrinello-Rahman barostat. A 100 ns production run was performed with a 2-fs time step using the Leapfrog integrator and LINCS for constraints, with a van der Waals cutoff of 1.2 nm. Results were analyzed and presented as means with standard deviations, based on triplicate studies.

2.4 Determination of physicochemical and ADMET properties

The physicochemical and ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) properties of the selected compounds were analyzed using the pkCSM web server, (https://biosig.lab.uq.edu.au/pkcsm/). The pkCSM tool provided predictions for key properties such as molecular weight, logP, and aqueous solubility, along with evaluations of critical pharmacokinetic parameters like gastrointestinal absorption, blood–brain barrier permeability, and hepatotoxicity (Pires et al., 2015).

3.1 Molecular docking study

The validity of the docking method utilized in this investigation was established through a validation process involving the molecular docking between 2NAO and a known compound, namely CID_9998128. Our findings demonstrated that CID_9998128 occupied the identical binding site on 2NAO, consistent with earlier reports (Thai et al., 2018). The interaction of CID_9998128 with amino acid residues, including Phe4, His6, Ser8, Gly9, Tyr10, Glu11, Val12, and His13 of chains D, E, and F, was observed. The compound CID_9998128 established three conventional hydrogen bonds and seven hydrophobic interactions (such as Pi-Pi T-shaped, Pi-Pi Stacked, Alkyl, and Pi-Alkyl). Given the reproducibility of these results with the previously published report (Thai et al., 2018), it suggests that the in-silico docking principle employed in this investigation is indeed valid. The analyses of molecular docking of target proteins with natural compounds suggest that all the studied ligands were bound between different filaments of the Aβ_1-42 fiber (Table 1, Figs. 1-3). The docking energy of ligands was in the range of −4.7 to −7.2 kcal/mol, −4.8 to −8.2 kcal/mol, and −5.0 to −8.0 kcal/mol for 2BEG, 2MXU, and 2NAO, respectively. In addition, docking energies of the most promising compounds namely Diosmetin, Rutin, and Genistein were calculated using the Glide SP module (Schrodinger, LLC, NY, USA), as reported earlier (AlAjmi et al., 2018). For the 2BEG target, the energies were −7.9, −7.2, and −7.6 kcal/mol for Diosmetin, Rutin, and Genistein, respectively. For the 2MXU target, the energies were −7.8, −8.9, and −8.1 kcal/mol, respectively. For the 2NAO target, the energies were −7.7, −7.5, and −7.9 kcal/mol, respectively.

Close

Fig. 1

Molecular docking of Diosmetin with 2BEG. (A) 2D docking pose, (B) 3D docking pose, and (C) Molecular interaction between 2BEG and Diosmetin.

Export to PPT

Close

Fig. 2

Molecular docking of Rutin with 2MXU. (A) 2D docking pose, (B) 3D docking pose, and (C) Molecular interaction between 2MXU and Rutin.

Export to PPT

Close

Fig. 3

Molecular docking of Genistein with 2NAO. (A) 2D docking pose, (B) 3D docking pose, and (C) Molecular interaction between 2NAO and Genistein.

Export to PPT

Amongst the studied ligands, Diosmetin, Rutin, and Genistein displayed the lowest docking energy towards 2BEG, 2MXU, and 2NAO, respectively. Diosmetin, a bioactive flavonoid, has potential therapeutic applications including anti-inflammatory, anti-oxidant, anti-microbial, and analgesic effects (Garg et al., 2022). Further, Genistein, an isoflavone from legumes, exhibits anti-oxidant, anti-inflammatory, anti-proliferative, and anti-microbial properties (Khan et al., 2022). Rutin, a flavonol, is known for its anti-oxidant, cytoprotective, anti-carcinogenic, neuroprotective, and cardioprotective properties (Ganeshpurkar and Saluja, 2017). Flavonoids like Genistein (AlFaris et al., 2021) and Rutin (Al Juhaimi et al., 2018) have been reported in Saudi Arabian dates, while Diosmetin derivatives have been extracted from date palms using UPLC/MS/MS techniques (Alshwyeh, 2020).

The 2BEG-Diosmetin pair was held by two standard hydrogen bonds (E:Val39:HN, and E:Val18:O), and ten hydrophobic interactions with E:Leu17:CD2 (one Pi-Sigma interaction), C:Phe19 (one Pi-Pi T-Shaped interaction), D:Phe19 (two Pi-Pi T-Shaped interactions), E:Phe19 (one Pi-Pi T-Shaped interaction), E:Leu17 (one Pi-Alkyl interaction), E:Val40 (one Pi-Alkyl interaction), D:Val17 (one Pi-Alkyl linkage), D:Val40 (one Pi-Alkyl linkage), and E:Val40 (one Pi-Alkyl interaction) (Fig. 1, Table 2). Further, Diosmetin formed several van der Waals’ interactions with B:Phe19, C:Leu17, C:Val18, C:Val40, D:Gly38, E:Gly37, and E:Gly38. The docking energy of the 2BEG:Diosmetin complex is −7.2 kcal/mol, corresponding to a binding affinity of 1.91 × 10⁵ M⁻¹.

The 2MXU-Rutin system was stabilized by one classical hydrogen linkage with C:Gly33:HN, and one carbon-hydrogen bond with C:His14:CE1 (Fig. 2A, 2B). The complex was also stabilized by six hydrophobic linkages with C:Ile32-CD1 (one Pi-Sigma interaction), C:His14 (one Pi-Pi stacked interaction), C:Gly33:C-O; Leu34:N (one Amide-Pi interaction), D:Val12 (one Alkyl interaction), D:His14 (one Pi-Alkyl interaction), and A:Val12 (one Pi-Alkyl interaction) (Fig. 2, Table 2). The system further revealed numerous van der Waals connections , such as A:Leu17, A:Ile32, A:Gly33, A:Leu34, B:Val12, B:His14, B:Leu17, B:Ile32, B:Gly33, B:Leu34, C:Val12, C:Leu34, D:Leu17, D:Ile32, D:Gly33, E:His14, and E:Leu17. The binding free energy of the 2MXU-Rutin combination was −8.2 kcal/mol, consistent with the dissociation constant of 1.03 × 10⁶ M⁻¹ (Table 2). Furthermore, the 2NAO-Genistein complex was stabilized by six conventional hydrogen bonds with A:Tyr10:O, A:His13:HD1, A:His14:HN, A:Gln15:O, B:His14:HN, and B:His14:O; and three hydrophobic interactions with B:Val12:CG2 (Pi-Sigma interaction), A:Lys16, (Pi-Alkyl interactions) and B:Lys16 (Pi-Alkyl interactions) (Fig. 3, Table 2). Genistein also established several van der Waals’ connections such as A:Ser8, A:Gly9, A:Glu11, B:Tyr10, B:His13, and B:Gln15. The docking energy of the 2NAO-Genistein interaction was −8.0 kcal/mol, equivalent to a dissociation constant of 7.37 × 10⁵ M⁻¹ (Table 2).

3.2 Molecular dynamics simulation

3.2.1

3.2.1 Root mean square deviation (RMSD)

The RMSD value of a protein–ligand pair is commonly used to assess the complex's durability and dynamical character. Thus, we calculated the RMSD of targeted peptides combined with their respective ligand over a time duration of 100 ns (Fig. 4). The RMSD of 2BEG alone displayed some fluctuation during the first 0–15 ns and then achieved symmetry from 20 ns onwards. Also, the RMSD of 2BEG in conjunction with Diosmetin stayed constant during the 20–100 ns simulation time, suggesting the unchanging character of the 2BEG-Diosmetin pair. The mean RMSD values of 2BEG alone, and 2BEG-Diosmetin complex during 20–100 ns were 5.48 ± 0.32 Å, and 6.89 ± 0.48 Å respectively (Fig. 4A).

Close

Fig. 4

Molecular dynamics simulation 2BEG, 2MXU, and 2NAO with their respective bioactive compounds Diosmetin, Rutin, and Genistein. (A) Root mean square deviation (RMSD) of 2BEG in the absence and presence of Diosmetin, (B) RMSD of 2MXU in the absence and presence of Rutin, and (C) RMSD of 2NAO in the absence and presence of Genistein.

Export to PPT

Likewise, the RMSD of 2MXU alone displayed some fluctuation during 0–10 ns, afterward attaining stability. The RMSD of 2MXU in the presence of Rutin showed some variation during 0–20 ns and then continued to be steady across 20–100 ns simulation time, hinting at the complex’s stability. The mean RMSD estimates of 2MXU only, and 2MXU-Rutin conjugate during 20–100 ns were 5.97 ± 0.26 Å, and 6.07 ± 0.37 Å respectively (Fig. 4B). Furthermore, the RMSD of 2NAO alone or in the presence of Genistein was consistent during the whole MS study. The mean RMSD estimates of 2NAO only, and 2NAO-Genistein conjugate during 0–100 ns were found to be 2.14 ± 0.13 Å, and 1.88 ± 0.09 Å, respectively (Fig. 4C). During the MD simulation, the steady RMSD values indicated the formation of a strong association between the protein and its ligand. Furthermore, it revealed that there existed no significant alterations in the overall conformation of the peptide-ligand system.

3.2.2

3.2.2 Radius of gyration (Rg) and solvent accessible surface area (SASA)

The Rg represents the peptide-ligand's compact dimensions, whereas the SASA evaluates the protein–ligand complex's interaction with the liquid. Each of them, consequently, offers details on the resilience of the peptide-ligand composite during the MS study. Herein, the Rg of 2BEG, 2MXU, and 2NAO was evaluated in combination with their respective ligands i.e. Diosmetin, Rutin, and Genistein respectively (Fig. 5A). The Rg of 2BEG-Diosmetin, 2MXU-Rutin, and 2NAO-Genistein conjugates varied from 4.14-5.16 Å, to 4.43–4.91 Å, and 4.52–4.64 Å respectively, with a mean value of 4.58 ± 0.34 Å, 4.86 ± 0.17 Å and 4.57 ± 0.06 Å, respectively (Fig. 5A). The SASA of 2BEG, 2MXU, and 2NAO in the company of their respective ligands i.e. Diosmetin, Rutin, and Genistein during 20–100 ns simulation fluctuated in the range of 423.4–503.7 Å², 491.7–508.5 Å², and 444.6–476.4 Å² respectively, with a mean of 470.8 ± 17.3 Å², 478.3 ± 18.1 Å², and 467.4 ± 13.9 Å² correspondingly (Fig. 5B). The robustness of the protein–ligand complex is confirmed by these data, which show that the differences in Rg and SASA of 2BEG, 2MXU, and 2NAO combined with their respective ligands did not diverge considerably.

Close

Fig. 5

(A) Radius of gyration, and (B) Solvent accessible surface area (SASA) of 2BEG, 2MXU, and 2NAO in the absence and presence of Diosmetin, Rutin, and Genistein.

Export to PPT

3.2.3

3.2.3 Total number of contacts between proteins and ligands

The overall number of permanent connections that are generated between a specific protein and its binding partner determines how strong the system is. Here, the complete set of interactions between 2BEG and Diosmetin, 2MXU and Rutin, as well as 2NAO and Genistein over simulation time was computed (Fig. 6). In 2BEG-Diosmetin system, the total number of contacts varied in the 1–15 range, with no less than 7 contacts. Likewise, the total contacts between 2MXU and Rutin varied from 3 to 14, with an average of 8 contacts. Further, the contacts between 2NAO and Genistein were in the range of 0–9, with an average of 4 interacting regions. These findings point to the establishment of a persistent ligand–protein combination.

Close

Fig. 6

Variation in total number of contacts between proteins and ligands. (A) 2BEG-Diosmetin, (B) 2MXU-Rutin, and (C) 2NAO-Genistein complexes.

Export to PPT

The Aβ peptide exits in several polymorphic forms; each characterized by a unique conformation and motif (Miller et al., 2010). Numerous pivotal conformational features of Aβ_1-42 fibrils are now available. For example, Aβ_1-42 fibrils can form a trifold-β pattern consisting of three separate overlapping parallel β-sheets (Xiao et al., 2015). Alternatively, they can also adopt a β-sheet–β-turn–β-sheet motif, featuring two inter-molecular, parallel, and in-register β-strands (Lührs et al., 2005). In our investigation, the model peptides 2MXU and 2NAO are S-shaped dodecamers and trimer respectively, whereas 2BEG is a U-shaped fibril. In addition, it was discovered that the residues including Val12, His13, His14, Lys16, Leu 17, Phe19, Ile32, Gly33, Leu34, and Val40 interacted actively with the ligands (Table 2), which are in concordance with previous studies (Kuang et al., 2015; Xi et al., 2016). In 2BEG protofilament, the residues Leu134, Phe19, Val36, and Ala21 of one β-sheet mediate the hydrophobic intermolecular contacts with the even-numbered residues of a parallel β-strand (Petkova et al., 2002). Residue pair Asp23-Lys28, connects these two strands via an intermolecular salt bridge, contributing to the overall stability of the “U” shape of the protofibril (Jarmuła et al., 2022). Further, a framework of inter-and intra-strand connections stabilizes this U-shaped proto-fibril. Notably, this network includes a π–π interaction involving the aromatic rings of internally present Phe19 and Phe20 located on the exterior of the peptide (Ahmed et al., 2010). In our study, Phe19 is involved in binding the ligand via π-π hydrophobic interaction, while Leu17 and Val40 form π-Alkyl bonding with the Diosmetin (Fig. 1).

The 3-D conformation of 2MXU significantly deviates from a typical β-loop-β motif generally obtained in high-resolution structures of Aβ_1-40 peptides. For instance, in the 2MXU conformer, specific side chains interactions like those concerning Phe19 and Leu34 that are previously reported in solid state NMR spectra of Aβ_1-42 fibrils, are lacking (Villalobos Acosta et al., 2018). Moreover, cross-peaks of several key residues including Val12, His13, and His14 are also absent from the spectra (Lührs et al., 2005). Interestingly, in this study, these residues are shown to be involved in strong binding with the ligand through H-bond (His14) or Pi-Alkyl interaction (Val12) (Fig. 2). Additionally, the 2MXU-fibril structure reveals the presence of an ionic bond between the amino moiety of Lys28 and the C terminus of Ala42 (Xiao et al., 2015); this interaction is absent in Aβ_1-40 fibrils. In the secondary structure, Gly33 is present at the loop region connecting the three β-strands, one of which involves residues from Val12 to Phe20. We observed that Gly33 interacts with several atoms of Rutin via both H-bonding and amide-Pi stacked hydrophobic interactions (Fig. 2).

In the 2NAO structure, residues 1–14 are organized in a β-sheet, while residues 15–42 construct a cross–β-sheet with hidden hydrophobic side chains. For both Aβ_1-42 peptide types A and B, inter-sheet salt bridges involving Lys16, Glu22, and Ala42 stabilize the structure by linking separate cross-β sheets. Additionally, Met35 of one β-sheet interacts with Leu17 or Gln15 from the second β-strand (Villalobos Acosta et al., 2018). Our research identified hydrogen bonding between the ligand and charged residues His16, His13, and Gln15, with Lys16 engaging in Pi-Alkyl interactions with Genistein (Fig. 3).

As a rule, the lower the docking energy, the better a ligand’s binding affinity. The docking scores in Table 2 reveal that the 2NAO and 2MXU peptides have high binding affinity for the ligands, while 2BEG shows the lowest binding energy. Both 2NAO and 2MXU fibrils are stabilized by strong hydrogen bonds and significant hydrophobic interactions (π-Alkyl and π-Sigma). Previous studies noted that high docking scores were due to pi-pi stacking (in His14), hydrogen bonding between the ligand’s amine and Gly33, and hydrophobic interactions with residues Val12, Leu17, Leu34, and Ile32 (Marondedze et al., 2020). In contrast, 2BEG's stability relies mainly on hydrophobic interactions, with minimal hydrogen bonding, which may account for the lower binding affinity of Diosmetin.

3.3 Prediction of Physicochemical Properties, and Drug-Likeness potential

Molecular descriptors provide critical insights into the pharmacokinetic and pharmacodynamic properties of compounds, key for evaluating their drug candidacy. Diosmetin, Rutin, and Genistein have molecular weights of 300.27, 610.52, and 270.24 g/mol respectively (Table 3), highlighting significant variations in size that could impact their pharmacokinetics. Their LogP values range from −1.69 to 2.58, reflecting differences in lipid solubility that might affect their absorption and membrane permeability. The compounds also differ in structural flexibility, with rotatable bonds ranging from 1 to 6, suggesting they can adopt various conformations that could influence receptor binding. Hydrogen bonding capacities vary, with Diosmetin, Rutin, and Genistein having between 3 and 10 hydrogen bond donors and 5 to 16 acceptors, affecting their potential to form bonds with other molecules. The total polar surface areas (TPSA) of these compounds range from 112.52 to 240.90 Å², indicating differences in solubility and interactions with aqueous environments (Table 3). These diverse molecular characteristics underscore the unique biological activities and therapeutic potentials of each compound, highlighting the importance of understanding these properties in drug development.

3.4 Analysis of ADMET properties

The ADMET analysis of Diosmetin, Rutin, and Genistein highlights their potential as therapeutic agents (Table 3). Their solubility values of −3.238, −2.892, and −3.595 mol/L suggest good bioavailability. Caco-2 permeability results show Diosmetin (0.326 × 10⁻⁶ cm/s) and Genistein (0.900 × 10⁻⁶ cm/s) have better potential for oral absorption than Rutin, which has a negative value. This is reflected in their intestinal absorption rates: Diosmetin at 79.9 %, Rutin at 23.4 %, and Genistein at 93.4 %, indicating Rutin’s lower efficiency. All three compounds have a similar skin permeability value of −2.735, suggesting comparable potential for topical use. They are substrates for P-glycoprotein but do not inhibit P-glycoprotein I or II, minimizing concerns about drug interactions.

Regarding distribution, Diosmetin (0.709 L/Kg) and Rutin (1.663 L/Kg) have higher volumes of distribution compared to Genistein (0.094 L/Kg), indicating more extensive body distribution for Rutin (Table 3). The fraction unbound (Fu) values are 0.068 for Diosmetin, 0.187 for Rutin, and 0.087 for Genistein, suggesting that Rutin has a higher proportion of unbound drug in the plasma, which could enhance its activity but also increase the risk of drug interactions. Metabolically, Diosmetin inhibits CYP1A2, CYP2C19, and CYP2C9, while Genistein inhibits CYP1A2 and CYP2C19. These interactions may affect the metabolism of other drugs, necessitating careful consideration in combination therapies. None of the compounds show AMES toxicity or hepatotoxicity. Diosmetin, Rutin, and Genistein have clearance rates of 0.598 ml/min/kg, −0.369 ml/min/kg, and 0.151 ml/min/kg, respectively. Rutin’s negative clearance rate suggests possible elimination issues, and its inhibition of hERG II channels may pose a risk of cardiotoxicity, warranting further investigation (Table 3).

Diosmetin and Genistein exhibit favorable pharmacokinetic profiles for oral use, owing to their higher permeability and absorption rates. Conversely, Rutin's lower absorption rate could limit its effectiveness. Enhancing Rutin's bioavailability and validating these results in vivo are needed. While Diosmetin, Rutin, and Genistein show promising safety profiles with no hepatotoxicity or skin sensitization, further investigation is necessary to assess metabolic interactions and Rutin's potential cardiotoxicity to ensure their clinical safety and efficacy.

3.5 Potential mechanism of action

The therapeutic action of date palm is believed to stem from its diverse botanicals with various physiological effects. Some compounds directly inhibit Aβ fibril formation. Flavonoids like resveratrol can prevent Aβ_1-42 fibril formation and convert the peptide into non-toxic aggregates (Phan et al., 2019). Diosmetin reduces Aβ accumulation by protecting against Advanced Glycation End Products (AGEs)-induced reactive oxygen species (ROS), downregulating Aβ production, and promoting its degradation (Lai et al., 2022). Rutin improves cognitive function in mice by reducing tau deposits and modulating its phosphorylation (Sun et al., 2021). Genistein protects against Aβ toxicity by inhibiting protein kinase B inactivation and tau hyperphosphorylation (Petry et al., 2020). Further, Genistein also acts as a dual inhibitor of Aβ and hIAPP aggregation (Ren et al., 2018) and has shown potential in crossing the blood–brain barrier to reduce Aβ-related neurotoxicity in vivo (Duan et al., 2021). These findings suggest that date palm metabolites have potential as treatments for Alzheimer's disease, but further research is needed to clarify the specific roles and mechanisms of each flavonoid in detail.

3.6 Limitations of the study

Computational approaches offer significant potential to streamline drug development and clinical research, reducing the need for animal models and human trials while predicting toxicities more efficiently. Molecular simulations, key to these methods, can predict binding affinities and interactions with Alzheimer's-related proteins. However, molecular docking has limitations, including limited ligand and receptor conformation sampling and reliance on estimated scoring algorithms, which may not always align with experimental results. Moreover, it is important to note that the docking and simulation studies were performed in isolation, without considering the presence of other biomolecules. These studies do not mimic the cellular milieu as interactions inside the cell are different from those encountered in the lab. Therefore, in vitro as well as in vivo studies are required to validate the findings of the present study.