2.1 Plant materials

The fresh stems of Glycosmis parviflora (Sims) Little were collected in March 2021 from Daklak, Vietnam. The plant voucher (VGP202105) was authenticated by Dr. Thanh-Hoa Vo (Vietnam National University Ho Chi Minh City, Vietnam), deposited at the School of Medicine, Vietnam National University Ho Chi Minh City, Vietnam, and the Department of Pharmacognosy, College of Pharmacy, Taipei Medical University, Taipei, Taiwan.

2.2 Extraction

The dried stems of G. parviflora (2.5 kg) were extracted thrice with ethanol 95 % at 50 °C. The crude extract (120.0 g) was obtained by evaporating the filtrate under vacuum. The crude extract was suspended in the distilled water and partitioned with ethyl acetate (EtOAc) and n-butanol (BuOH) to obtain an EtOAc residue (42.0 g), a BuOH residue (49.0 g), and an aqueous residue (21.0 g), respectively. The anti-AChE activity of these layers was investigated.

2.3 Cholinesterase inhibitory activity assay

The cholinesterase inhibitory activity analysis was performed as described by Ellman et al. (Ellman et al., 1961) with slight modifications. AChE from electric eel and BuChE from equine serum was provided by Sigma Aldrich (CAS No. 9000-81-1, 9001-08-5, respectively). In brief, assays were performed using the concentration at 0.03 U/mL of AChE (Enzyme was dissolved in Tris buffer solution pH 8.0) in the presence of 1.5 mM 5,5′-Dithiobis (2-nitrobenzoic acid),1.5 mM acetylthiocholine iodide, and inhibitors in 200 µL reaction mixtures, and continuously monitored for 10 mins at 405 nm, 37 °C. 5,5′-Dithiobis (2-nitrobenzoic acid) was used for color development, caused by a reaction between it and thiocholine (a product of the hydrolysis by AChE and BChE). BChE activity was assayed using the same method, but butyrylthiocholine iodide, and Tris buffer solution pH 7.4 were used instead of acetylthiocholine iodide, and Tris buffer solution pH 8.0, respectively. Physostigmine (Santa Cruz, Texas, USA, CAS No. 57-47-6) at the dose of 2.75 µg/mL was used as a positive control. The control group was assigned 100 %, and the compound (or the fraction)-treated groups were calculated relative to the control. Percentages of inhibitory activity were calculated with the following formula: $$\%\mathrm{C}\mathrm{h}\mathrm{E}\mathrm{i}\mathrm{n}\mathrm{h}\mathrm{i}\mathrm{b}\mathrm{i}\mathrm{t}\mathrm{o}\mathrm{r}\mathrm{y}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{i}\mathrm{t}\mathrm{y}=\frac{{\mathrm{A}}_{\mathrm{E}\mathrm{C}}-{\mathrm{A}}_{\mathrm{B}\mathrm{C}}}{{\mathrm{A}}_{\mathrm{S}\mathrm{E}}-{\mathrm{A}}_{\mathrm{S}\mathrm{C}}}\times 100$$ where in:

A_SE = absorbance of sample experiment (with ChE).

A_SC = absorbance of sample control (without ChE).

A_EC = absorbance of experimental control (with ChE).

A_BC = absorbance of blank control (without ChE).

2.4 Preparation of acetylcholinesterase-substracted EtOAc layer extract for UPLC-MS/MS

The previously outlined method (Hsu et al., 2021) for removing acetylcholinesterase binding components was adhered to. The blank solution was made up of 50 µL of EtOAc extract of the stem of G. parviflora (20.000 ppm in dimethyl sulfoxide), 325 µL of phosphate buffer, and 150 µL of tris buffer. In parallel, instead of utilizing 150 µL of Tris buffer, 50 µL of Tris buffer and 100 µL of AChE were added to the test solution. Both solutions underwent stirring and were incubated for 40 min in a 37 °C water bath. Regarding that, 475 µL of acetonitrile was added to each solution to terminate the enzyme reaction. Both test solutions were filtered through 0.22 µm with a final concentration of 1000 ppm. UPLC-MS/MS was implemented to analyze the membrane working solutions.

2.5 Acetylcholinesterase dereplication of UPLC/MS/MS metabolome profiles analysis

The sample was analyzed using a UPLC system (ACQUITY UPLC; Waters, USA) coupled with a linear ion trap-orbitrap mass spectrometer (Q Exactive Plus Hybrid Quadrupole Orbitrap Mass (Thermo Fisher Scientific Inc., USA). Separation was achieved using an AcclaimTM RSLC PA2 column (2.2 μm, 120 Å, 2.1 mm x 150) (Thermo Fisher Scientific Inc., Waltham, MA, USA) at a flow rate of 0.3 mL/min with a gradient program of mobile phase A (dd water) and B (MeOH). The gradient program included a 5–30 % phase B from 0-3 min, 30–60 % phase B from 3 − 7 min, and 60–100 % phase B from 7 − 11 min. The column was washed with 100 % phase B for 2 min and re-equilibrated for 3 mins. The mass spectrometer was operated in both positive and negative ion modes using the Xcalibur software (version 2.0, Thermo Fisher Scientific, Bremen, Germany) with high-resolution full scan cycles (m/z 100–1500) and dd MS2. The sample was injected at a volume of 2 µL.

For finding the disparity in metabolome profiles between the EtOAc extract (Blank) and the AChE protein-subtracted extract (Test), data containing retention time, m/z, and peak intensity from UPLC–ESI–MS analyses were exported as text files. Subsequent processing involved the utilization of Xcalibur software and Microsoft Excel 2010 for manual pre-processing. Comparative analysis was conducted to discern differences in composition profiles between the Blank and Test extracts. The intensity ratio for each peak in the Test extract chromatogram, relative to the Blank extract, was calculated using the provided equation. Peaks exhibiting an intensity change exceeding 10 % were considered significantly altered due to AChE protein subtraction, designating them as potential AChE-interacting compounds (Hsu et al., 2021). Identification of chemical formulas and structures corresponding to these peaks was achieved by interrogation the ChemSpider and Reaxys databases.

2.6 Isolation and identification

The bioactive EtOAC residue (42.0 g) was fractionated into twelve fractions (Fr.1 ∼ 12) using the MPLC system (Isolera ONE, Uppsala, Sweden) with a silica gel column (Biotage SNAP, KP-Sil, 340.0 g), the gradient mixture of n-hexane–ethyl acetate (70:30, 60:40, 50:50, 25:75, 0:100, v/v) and the mixture of acetone–methanol (MeOH) (1:1, v/v). The semi-preparative HPLC method was conducted on six fractions with Luna silica gel column (250 × 10 nm, 5 µm Silica (2) 100 Å). The mixture of n-hexane – EtOAc (7:3, v/v) was eluted to obtain compounds 6 (5.6 mg), compound 8 (6.7 mg), and 9 (4.6 mg) from fraction Fr.2. Four pure compounds were yielded from Fr.4 by using the similar mobile phase, including compounds 3 (13.0 mg), 4 (13.8 mg), 5 (5.9 mg) and 7 (6.1 mg). When using the mixture of n-hexane – EtOAc (6:4, v/v), compound 1 (32.4 mg) was purified from Fr.5, and from Fr.7 and Fr.8 obtained compound 10 (46.0 mg) when using the mobile phase of n-hexane – EtOAc (5:5, v/v); Eluted by 100 % EtOAc for Fr.12, compounds 11 (17.6 mg) and 2 (25.6 mg) were yielded, respectively as show in Table S1. The extraction and isolation procedure had been repeated to obtain the required amount of each compound for chemical and biological experiments.

The structures of isolated compounds were elucidated by ESI-MS, 1D, and 2D NMR data and compared with the literature references. All NMR spectra were recorded on Bruker spectrometers (Rheinstetten, Germany) at 300 MHz, 500 MHz or 600 MHz with methanol‑d₄, dimethyl sulfoxide-d₆, acetone‑d₆, or chloroform-d as solvent and tetrametylsilane as reference. The UV–Vis spectra were recorded with a Hitachi L-7455 Photo Diode Array Detector (Japan). ESI-MS data was obtained on a Q Exactive Plus Hybrid Quadrupole Orbitrap Mass (Thermo Fisher Scientific Inc., USA).

2.7 Kinetic studies of acetylcholinesterase inhibition

Kinetic studies were performed with the same test conditions, using seven concentrations of substrate acetylthiocholine iodide (0.1, 0.125, 0.167, 0.25, 0.5, 1, and 2 mM) and two concentrations of inhibitor at around their respective IC₅₀ values. The initial rates were determined based on the absorbance increase at 405 nm observed between 0.5 and 5 min. The Lineweaver-Burk plots were plotted as Eq: $$\frac{1}{V}=\frac{\mathit{Km}}{\mathit{Vmax}}\frac{1}{[S]}+\frac{1}{\mathit{Vmax}}$$ Where v, [S], and Km values represent the rate of the enzymatic reaction, the concentration of substrate, and the Michaelis-Menten constant, respectively.

2.8 Molecular docking

The CDOCKER algorithm in Discovery Studio 2021 (Accelrys Software, Inc., San Diego, CA, USA) was used to assess the probable molecular binding mode between four potential AChE inhibitors, physostigmine as a positive control, and the acetylcholinesterase enzyme. The Protein Data Bank (https://www.rcsb.org/pdb/) was used to obtain the crystal structure of acetylcholinesterase (PDB ID: 1C2B). The compounds' 2D structures were created in ChemBioDraw Ultra 13.3 and converted to standard 3D formats in DS 2021. The transformed structures were then subjected to energy reduction in the CHARMm force field using the conjugate gradient approach (convergence criterion of 0.001 kcal/mol) (Brooks et al., 1983). Following that, the energy-minimized structures were implemented in the molecular docking experiments.

The apo-crystal structure of the acetylcholinesterase receptor (PDB code: 1C2B) was created by eliminating water molecules and adding hydrogen atoms. The available module represented missing loop sections in the prepared protein. Subsequently, the protein structure was subjected to ionization and protonation computations, followed by a final round of energy minimization optimized for molecular docking. The docking analysis binding site was determined by referring to the PDB site records and then updated using the binding site module. Upon this, the molecular docking analysis focused on the discovered binding site sphere within the 1C2B structure (coordinates: x = 40.1361, y = 84.2962, z = 29.1651; radius = 18.9). Then, using the CDOCKER program included in DS 2021, the energy-minimized structures of the four selected candidate compounds, positive control, were docked into the binding site of 1C2B. Each compound, 10 docking poses were generated and ranked based on CDOCKER energy which was computed by the CHARMm energy (−CDOCKER_ENERGY, interaction energy plus ligand strain) and the binding energy (−CDOCKER_INTERACTION_ENERGY). The pose with the highest CDOCKER energy score was chosen as the best match for each chemical. Furthermore, the ligand-binding free energy was calculated for each complex using the Generalized Born with Molecular Volume (GBMV) approach (Lee et al., 2003).

2.9 ADMET profiles using in silico method

In silico ADMET (Absorption, distribution, metabolism, excretion, and toxicity) profiles of each bioactive compound were forecasted using admetSAR 2.0 (https://lmmd.ecust.edu.cn/admetsar2/) to anticipate human intestinal absorption, Caco-2 permeability, blood–brain barrier permeability, P-gp inhibition, CYP3A4 substrate activity, inhibition of CYP3A4, CYP2C19, CYP2D6, and CYP1A2, hepatotoxicity, respiratory toxicity, reproductive toxicity, nephrotoxicity, plasma protein binding, and acute oral toxicity (Khanal et al., 2019; Dwivedi et al., 2021).

2.10 Statistical analysis

All data were repeated three times (n = 3) and denoted as means ± standard deviation (SD). Data was analyzed using GraphPad Prism edition 9.5.1 (GraphPad Software Inc, California, USA). The inhibitory activities of the crude extract and layers were compared using one-way ANOVA and Dunnett's Test (p-values < 0.05 and < 0.01, respectively).

3.1 Inhibitory against cholinesterase activity of the extracts and layers from G. parviflora

To trail the activity efficacy of the extract of G. parviflora stem in preventing AD, it was fractionated into three layers: the ethyl acetate layer (EtOAc layer), the butanol layer (BuOH layer), and the aqueous layer. The findings demonstrated that the ethyl acetate layer extracted from the stem of G. parviflora displayed the highest AChE inhibition, reaching 58.79 ± 4.31 % at the concentration of 125 µg/mL, surpassing both the aqueous and n-butanol layer (Fig. 1). In comparison, the anti-acetylcholinesterase activity displayed significantly greater potency towards butyrylcholinesterase inhibition when utilizing identical extraction concentrations and layers. The BuChE inhibitory effect demonstrated a value of 26.26 ± 4.00 % for the ethyl acetate layer, representing more half the magnitude of the activity against AChE (Fig. S1). Based on these findings, it is inferred that the compounds present in the EtOAc layer possess the potential to serve as AChE inhibitors.

Close

Fig. 1

Comparative on acetylcholinesterase (AChE) inhibition between crude extract, and fractionated layers in G. parviflora stems. The crude extracts and fraction layers at 125 µg/mL, Physostigmine as a positive control at 2.75 µg/mL. The data were presented as mean ± S.D. (n = 3). EtOAc stands for ethyl acetate, and BuOH stands for n-butanol. * p-value < 0.05, ** p-value < 0.01 in comparison with the ethyl acetate layer.

Export to PPT

3.2 Fast screening potential acetylcholinesterase-inhibiting constituents by enzyme-interacting dereplication of UPLC-MS/MS profiling

The dereplication approach was employed to rapidly screen the compounds that inhibit AChE from the EtOAc layer of the stem of G. parviflora; comparing enzyme-treated and untreated samples (Hsu et al., 2021). This process involved analyzing UPLC-MS/MS chromatographic profiles to efficiently isolate and identify active compounds inhibiting AChE. The study showed the differences in signal intensities change between AChE-treated and untreated extracts at specific peaks a–f; the representative BPI chromatograms, as depicted in Fig. 2, and Table S2, are provided herein. These differences suggested that the components associated with these peaks which were caused by the AChE protein subtraction could be potential AChE inhibitor candidates.

Close

Fig. 2

The base peak chromatograms (BPI) of untreated (A) and acetylcholinesterase −treated samples (B) of the ethyl acetate extraction from G. parviflora stems in positive mode.

Export to PPT

3.3 Isolation and identification of the structure of the potential acetylcholinesterase inhibitors

According to the above, by this dereplication approach, potential targets were quickly narrowed down to eleven compounds from the six significant peaks. Next, merging using MPLC and semi-preparative HPLC isolated eleven compounds from the EtOAc layer of the stem of G. parviflora. As a result, eight pure alkaloids (1–9), a flavonoid (10), and a phytosterol (11) were successively isolated. Compound 9 is demonstrated as an undescribed alkaloid.

Compound 9 generated a colorless oil; [α] ²⁸_D − 27.6 (c 0.1, MeOH). The UV spectrum (MeOH) revealed three strong absorption peaks at the values of λ_max 219, 261, 334 nm and one medium at λ_max 294 nm, respectively (Fig. S9). The chemical formula of C₂₀H₂₃NO₂ was determined by the positive HR ESI-MS [M + H]⁺ at m/z 310.1803 (calcd for a C₂₀H₂₄NO₂ 310.1807) (Fig. S2). On the ¹H NMR spectrum, the signal of a para-tetrasubstituted benzene ring and an ortho-tetrasubstituted benzene ring was determined from two aromatic protons at δ_H [6.99 (d, 8.6, H-7); 7.16 (d, 8.6, H-7)]. Additionally, two methoxy signals were detected at δ_H 3.91 (s, 3H), 3.87 (s, 3H), one methyl group attached to an aromatic ring at δ_H 2.35 (s, 3H), one proton of NH group at δ_H 7.77 (s, 1H) and signals of prenyl group at δ_H 3.93 (m, 2H, H-1′), 5.33 (tqq, 6.5, 1.5, 1.5, H-2′), 1.93 (brs, 3H, H-4′) were also exhibited on the proton NMR analysis (Fig. S3). The structure of compound 9 was found using ¹³C NMR, DEPT, and HSQC data to yield five CH₃, one CH₂, five CH, and nine quaternary C groups (Fig. S4−6). The positions of two methoxy groups and a methyl group were established using HMBC data correlation between δ_H 3.91/ δ_C-2, δ_H 3.87/ δ_C-6 and δ_H 2.35/ δ_C-2,C-3,C-4, respectively (Fig. S7−8). Furthermore, the HMBC spectrum crosspeak of H-1′/C-5,C-6 δ_H-1ʹ/ δ_C-5,C-6 was used to identify the link of the prenyl group to the aromatic ring at the C-5 location. Other important HMBC and COSY signals are shown in Fig. 3, and 1D NMR data is clearly displayed in Table 1. Moreover, the 1D NMR spectroscopic data of compound 9 was quite comparable to that of compound 8 (Ito et al., 2004), except for substituting the –OH group to –OCH₃ at the C-2 position. Based on the data presented above, the structure of the undescribed carbazole alkaloid 9 was found to be 2,6-dimethoxy-3-methyl-5-(3-methylbut-2-en-1-yl)-9H-carbazole, namely glybomine D (Fig. 4).

Close

Fig. 3

Key HMBC and COSY of compound 9.

Export to PPT

Close

Fig. 4

The structures of isolated compounds 1–11 from G. parviflora stems.

Export to PPT

The other compounds were identified as 1,3-dimethoxy-2-hydroxy-10-methyl-9(10H)-acridinone (1), arborine (2), dictamnine (3), skimmianine (4), glycoborinine (5), O-methylglycosolone (6), glybomine C (7), glybomine B (8), 4′-O-Methyl-gallocatechin (10), stigmast-5,22-dien-3-O-β-D-glucopyranoside (11) (Fig. 4) by comparison with spectroscopic (1D and 2D-NMR) and LC-MS data from the literature.

Compound 1: C₁₆H₁₅NO₄; HR-ESI-MS [M + H]⁺ m/z 286.1069; for ¹H NMR (300 MHz, CDCl₃): 8.49 (dd, 8.1, 1.8, H-8), 7.74 (ddd, 8.7, 7.0, 1.7, H-6), 7.52 (d, 8.7, H-5), 7.32 (ddd, 7.9, 7.0, 0.9, H-7), 6.31 (s, H-4), 4.03 (s, 3-OMe), 3.94 (s, 1-OMe), 3.87 (s, N-Me) (Bunalema et al., 2017) (Fig. S10).

Compound 2: C₁₆H₁₄N₂O; HR-ESI-MS [M + H]⁺ m/z 251.1173; for ¹H NMR (300 MHz, CDCl₃): 8.38 (dd, 7.9, 1.7, 0.5, H-4), 7.72 (ddd, 8.7, 7.2, 1.7, H-1), 7.46 (d, 8.1, 7.2, 1.0, H-2, H-3), 7.25–7.38 (m, H-8, 9, 10, 11, 12, 13, 14), 4.25 (s, H-6), 3.61 (s, N-Me) (Murugan et al., 2020) (Fig. S11).

Compound 3: C₁₂H₉NO₂; HR-ESI-MS [M + H]⁺ m/z 200.0706; for ¹H NMR (500 MHz, CDCl₃): 8.28 (dd, 8.5, 1.5, H-5), 8.01 (dd, 8.4, 1.0, H-8), 7.69 (ddd, 8.4, 6.7, 1.5, H-7), 7.63 (d, 2.8, H-2′), 7.45 (ddd, 8.2, 6.7, 1.2, H-6), 7.09 (d, 2.8, H-3′), 4.46 (s, 4-OMe); ¹³C NMR (125 MHz, CDCl₃): 164.0 (C-2), 157.0 (C-4), 145.8 (C-10), 143.7 (C-2′), 129.8 (C-7), 127.9 (C-8), 123.9 (C-6), 122.5 (C-5), 118.9 (C-9), 104.9 (C-3′), 103.6 (C-3), 59.2 (C-4) (de Oliveira et al., 2016) (Fig. S12−13).

Compound 4: C₁₄H₁₃NO₄; HR-ESI-MS [M + H]⁺ m/z 260.0911; for ¹H NMR (300 MHz, CDCl₃): 8.02 (d, 9.4, H-5), 7.58 (d, 2.8, H-2′), 7.23 (d, 9.4, H-6), 7.05 (d, 2.8, H-3′), 4.44 (s, 4-OMe), 4.10 (s, 7-OMe), 4.03 (s, 8-OMe) (Cardoso-Lopes et al., 2010) (Fig. S14).

Compound 5: C₁₈H₁₇NO₂; HR-ESI-MS [M + H]⁺ m/z 280.1332; for ¹H NMR (500 MHz, CDCl₃): 7.83 (s, H-4), 7.21 (d, 9.8, H-1′), 7.10 (d, 8.5, H-8), 6.83 (d, 8.4, H-7), 6.80 (s, H-1), 5.80 (d, 9.8, H-2′), 2.40 (s, 3-Me), 1.48 (s, Me-4′, Me-5′); ¹³C NMR (125 MHz, CDCl₃): 153.0 (C-2), 146.7 (C-6), 140.7 (C-9a), 134.9 (C-8a), 131.3 (C-2′), 124.0 (C-4), 120.4 (C-1′), 119.1 (C-4b), 117.5 (C-3), 116.3 (C-4a), 115.1 (C-5), 114.1 (C-7), 110.2 (C-8), 96.6 (C-1), 75.2 (C-3′), 27.4 (C-4, 5′), 16.4 (Me-3) (Chakravarty et al., 1999) (Fig. S15−16).

Compound 6: C₁₇H₂₁NO₃; HR-ESI-MS [M + H]⁺ m/z 288.1595; for ¹H NMR (500 MHz, CDCl₃): 7.43 (dd, 8.0, 1.4, H-5), 7.16 (t, 8.0, H-6), 7.04 (dd, 8.0, 1.4, H-7), 5.26 (tq, 5.5, 1.5, H-2′), 3.95 (s, N-Me), 33.89 (s, 8-OMe), 3.87 (s, 4-OMe), 3.39 (d, 6.9, H₂-1′), 1.80 (s, 4′ −Me), 1.68 (s, 5′ −Me); ¹³C NMR (125 MHz, CDCl₃): 165.0 (C-2), 160.0 (C-4), 148.8 (C-8), 132.5 (C-3′), 130.6 (C-8a), 122.7 (C-3), 122.4 (C-6), 121.5 (C-2′), 120.2 (C-4a), 116.0 (C-5), 113.5 (C-7), 8-OMe), 56.7 (4-OMe), 35.5 (N-Me), 25.73 (5′-Me), 24.4 (C-1′), 18.0 (4′-Me) (Chakravarty et al., 1999) (Fig. S17−18).

Compound 7: C₁₈H₁₉NO₂; HR-ESI-MS [M + H]⁺ m/z 282.1485; for ¹H NMR (300 MHz, (CD₃)₂CO): 7.81 (s, H-4), 7.07 (d, 8.5, H-8), 6.85 (d, 8.4, H-7), 6.89 (s, H-1), 5.34 (m, H-2′), 3.94 (d, 6.4, H₂-1′), 2.34 (s, 3-Me), 1.94 (s, 4′ −Me), 1.69 (s, 5′-Me) (Ito et al., 2004) (Fig. S19).

Compound 8: C₁₉H₂₁NO₂; HR-ESI-MS [M + H]⁺ m/z 294.1496; for ¹H NMR (600 MHz, CDCl₃: 7.80 (s, H-4), 7.09 (d, 8.6, H-8), 6.97 (d, 8.6, H-7), 6.72 (s, H-1), 5.31 (m, H-2′), 3.91 (t, 5.7, H₂-1′), 3.86 (s, MeO-6), 2.38 (s, Me-3), 1.92 (d, 1.4, Me-3′), 1.69 (d, 1.6, 5′ −Me); ¹³C NMR (150 MHz, CDCl₃): 152.8 (C-2), 151.0 (C-6), 140.6 (C-9a), 135.2 (C-8a), 132.2 (C-3′), 124.5 (C-4), 122.8 (C-4b), 122.5, (C-2′) 117.5 (C-4a), 115.7 (C-3), 107.7 (C-8), 96.23 (C-1), 58.0 (6-OMe), 25.7 (C-1′), 25.6 (C-5′), 18.2 (C-4′), 16.3 (Me-3) (Brutting et al., 2016) (Fig. S20−21).

Compound 10: C₁₆H₁₆O₇; HR-ESI-MS [M + H]⁺ m/z 319.0809; for ¹H NMR (300 MHz, (CD₃OD): 6.41 (s, H-2′, 6′), 5.93(d, 2.3, H-6), 5.88 (d, 2.3, H-8), 4.56 (d, 7.1, H-2), 3.96 (m, H-3), 3.79 (s, 4′-OMe), 2.82 (dd, 16.2, 5.3, H-4b), 2.51 (dd, 16.2, 7.7, H-4a) (Garcia et al., 1993) (Fig. S22).

Compound 11: C₃₅H₅₈O₆; HR-ESI-MS [M + Na]⁺ m/z 597.4724; for ¹H NMR (600 MHz (CD₃)₂SO): 5.31 (br d, 4.8, H-6), 5.16 (dd, 15.2, 8.7, H-22), 5.02 (dd, 15.2, 8.7, H-23), 4.85 (m, OH-2′, 3′, 4′), 4.42 (s, OH-6′), 4.22 (d, 7.8H-1′), 3.64 (m, H-6′), 3.46 (td, 11.2, 5.6, H-3), 3.12 (t, 8.9, H-3′), 3.07 (m, H-4′), 3.01 (m, H-5′), 2.89 (td, 8.4, 3.6, H-2′), 2.37 (m, H-1), 2.12 (t, 12.5, H-1), 1.93 (m, H₂-16), 1.80 (m, H-4), 1.63 (m, H-8), 1.50 (m, H-7), 1.39 (ddd, 11.8, 8.0, 5.0, H-20), 1.23 (m, H-17), 1.15 (m, H-2), 0.99 (d, 6.6, H-19), 0.95 (s, H-9, 24), 0.89 (t, 6.3, H-29), 0.82 (s, H-26, 27), 0.65 (s, H-18).; ¹³C NMR (150 MHz, (CD₃)₂SO): 140.5 (C-5), 138.0 (C-22), 128.9 (C-23), 121.2 (C-6), 100.8 (C-1′), 77.0 (C-3), 76.8 (C-3′), 76.8 (C-5′), 73.5 (C-4′), 70.2 (C-2′), 61.1 (C-6′), 56.3 (C-14), 55.4 (C-17), 49.6 (C-9), 41.9 (C-12), 41.8 (C-13), 38.3 (C-1), 36.9 (C-4), 36.3 (C-10), 35.5 (C-20), 33.4 (C-2), 31.5 (C-24, 25), 31.4 (C-8), 31.4 (C-7), 29.3 (C-16), 25.5 (C-15), 23.9 (C-28), 22.6 (C-11), 19.1 (C-26), 19.0 (C-27), 18.9 (C-19), 18.7 (C-21), 11.8 (C-18), 11.7 (C-29) (Ridhay et al., 2012) (Fig. S23−24).

3.4 Assessment of acetylcholinesterase inhibitory activities of the eleven potential compounds

To further validate the inhibitory activities of potential AChE inhibitors identified from the EtOAc layer of the stem of G. parviflora, the inhibitory abilities of these eleven potential AChE inhibitors were measured through biochemical assays. This study utilized an acetylcholinesterase inhibition assay with physostigmine as the positive control to evaluate the effectiveness of purified compounds from the stem of G. parviflora EtOAc extracts in preventing AD. The results show that among eleven compounds, 6, 1, 4, and 2 showed lower IC₅₀ values ​​of 39.81 ± 1.76 µM, 41.53 ± 8.84 µM, 49.40 ± 7.10 µM, and 59.92 ± 12.71 µM, respectively, indicating that these four compounds exhibited more notable inhibitory activity towards AChE. As shown in Table 2, the remaining compounds seemed to have slight inhibition effects (IC₅₀ > 100 µM). In particular, compound 6 showed the greatest inhibition, indicating its potential as a lead compound for acetylcholinesterase inhibitors.

3.5 Evaluation of acetylcholinesterase kinetics for four highly potential compounds

Acetylcholinesterase inhibitors have been reported to exhibit competitive, noncompetitive, and mixed-type inhibitory (Heo et al., 2020; Zhang et al., 2023). In the present study, Lineweaver–Burk double reciprocal plots revealed that compounds 6, 1, 4, and 2 were mixed-type AChE inhibitors with a mode of action similar to physostigmine (Fig. 5) (Robaire and Kato, 1975). Subsequently, the interaction of the active compounds with the AChE enzyme was determined through molecular docking. The results suggest that compounds 6, 1, 4, and 2 inhibited the enzyme by binding with the free enzyme and the enzyme-substrate complex (Tabrez and Damanhouri, 2019).

Close

Fig. 5

Lineweaver–Burk plots for the inhibitions of AChE by (A) O-methylglycosolone (6), (B) 1,3-dimethoxy-2-hydroxy-10-methyl-9(10H)-acridinone (1), (C) skimmianine (4), (D) arborine (2) and (E) Physostigmine; Substrates were employed at seven diferent concentrations (0.1–2.0 mM). Experiments were carried out at two inhibitor concentrations at around their respective IC₅₀ values. Initial reaction rates are expressed as increases in absorbance per min. Each point is the average value from three independent experiment.

Export to PPT

3.6 Mechanism of four AChE inhibitors in the binding potency of acetylcholinesterase

A molecular docking system was used to further understand the binding affinity and mechanism of the four inhibitors (compounds 6, 1, 4, and 2) inhibitory on AChE. For comparison, physostigmine was employed as a reference standard. The higher negative value of CDOCKER energy points out higher binding affinity and, consequently, higher interaction potency with acetylcholinesterase protein. As a result, the positive control, namely physostigmine and four of the potential compounds from the stem of G. parviflora including O-methylglycosolone (6), 1,3-dimethoxy-2-hydroxy-10-methyl-9(10H)-acridinone (1), skimmianine (4), and arborine (2) were reported to be as –23.749, −20.213, −13.006, −11.815, and −11.387 kcal/mol, respectively (Table 3), had negative values of CDOCKER energy, which were considered as potential AChE inhibitors. O-methylglycosolone displayed the highest docking score among the four potential compounds, with a ΔG (binding free energy) as high as −20.213 Kcal/mol, indicating its highest binding interaction. For the AChE complex, Physostigmine formed hydrogen bonds with the side chains Gln291, Ile294, Phe295, and Ser293 of the complex, and the phenyl ring formed pi-pi T-shaped with Tyr286, pi-alkyl interactions with key residues Leu286, Trp341 and Phe338 in CAS and PAS region of the complex. O-methylglycosolone (6) interaction with the AChE complex formed a hydrogen bond with Tyr341, Tyr72, Ser293, and Arg296, the phenyl ring formed pi-pi T-shaped with Tyr341, pi-alkyl interactions with Leu76, Trp286 and Phe297. Compounds 1 and 4 have similar CDOCKER binding energy that might form H-bonding with the side chains Tyr72 and Ser293, and the phenyl ring formed pi-pi T-shaped with Trp 286 of the complex. Arborine (2) only formed hydrogen bonds with the side chains Tyr341 of the complex, which might cause low binding affinity (Fig. 6, Fig. S25). The results show that hydrogen bonds and hydrophobic forces may play an important role, especially the inhibitor interaction with amino acid residues such as Tyr341, Tyr72, Ser293, and Arg296 in the active cavity, which is crucial for effective and selective inhibition of AChE activity. The docking conformation image of each compound in the selected pose is shown in Fig. 6. In addition, the CDOCKER interaction energies and binding free energies (ΔG) of the four compounds had the same trend as the IC₅₀ values ​​in the in vitro method, supporting those the subsequent molecular docking predictions were reliable.

Close

Fig. 6

The binding modes between sites of acetylcholinesterase (PDB code: 1C2B) and AChE-inhibiting components were calculated by molecular docking. View of docking conformations for the four inhibitors binding to the active sites of active sites of acetylcholinesterase (1C2B), (A): O-methylglycosolone (6); (B): skimmianine (4); (C): 1,3-dimethoxy-2-hydroxy-10-methyl-9(10H)-acridinone (1) and (D): arborine (2). Key residues are depicted as line models against a background, with hydrogen bonds denoted by green dashed lines. Pi-pi interactions are illustrated with purple dashed lines, while pi-alkyl interactions are indicated by pink dashed lines.

Export to PPT

3.7 Prediction of pharmacokinetic properties

The pharmacokinetic characteristics of each bioactive compound are depicted in a heat map, illustrating human intestinal absorption (HIA), Caco-2 permeability (Caco-2p), blood–brain barrier (BBB) permeability, P-glycoprotein (P-gp) inhibition, CYP3A4, CYP2C19, CYP1A2 inhibition/substrate activity, hepatotoxicity, respiratory toxicity, reproductive toxicity, nephrotoxicity, plasma protein binding, and acute oral toxicity. The ADMET profile of each phytoconstituent is visualized as a heat map (Fig. 7). According to the ADMET prediction results, all compounds exhibit a high probability of human intestinal absorption (>90 %), which is equivalent to the result of the physostigmine-positive control. Among the compounds, 1,3-dimethoxy-2-hydroxy-10-methyl-9(10H)-acridinone (1) has the lowest probability of blood–brain barrier penetration (42.5 %), while arborine (2) demonstrates the highest probability (97.5 %), followed by O-methylglycosolone (6) (75 %), and skimmianine (4) (65 %). Regarding pharmacokinetic distribution, all compounds are predicted to have a high probability of plasma protein binding (>70 %). Metabolism prediction suggests that the compounds primarily undergo metabolism via CYP3A4 substrate/inhibition, except for O-methylglycosolone (6), metabolized through CYP2C19 inhibition (probability = 72.9 %). O-methylglycosolone (6) is predicted to have the lowest probability of hepatoxicity (50.34 %) compared to other compounds.

Close

Fig. 7

ADMET profile of bio-actives via heat map presenting 1,3-dimethoxy-2-hydroxy-10-methyl-9(10H)-acridinone (1), arborine (2), skimmianine (4), O-methylglycosolone (6), physostigmine (positive control).

Export to PPT