Synthesis, α-glucosidase inhibition and molecular docking studies of natural product 2-(2-phenyethyl)chromone analogues

2.1 Chemistry

The synthetic pathway of the title compounds 3–34 was shown in Scheme 1. The starting materials 1 was treated with POCl₃ in the presence of dimethylformamide solution to obtain intermediates 2 via Vilsmeier-Haack reaction (Xiang et al., 2012). Then, intermediates 2 condensed with various anilines in dichloromethane in the presence of Na(AcO)₃BH. After the completion of the reaction, the mixture was washed with sodium bicarbonate and purified by chromatography on silica gel to afford the corresponding (2-phenyethyl)chromones derivatives 3–34 (McGonagle et al., 2013). The chemical structures of all the target products 3–34 were characterized by spectroscopic techniques such as ¹H NMR, ¹³C NMR, and HRMS (Supplementary material, Fig. S1-S96). For instance, the ¹H NMR spectrum of 3 showed a doublet peak at δ_H 4.09 ppm with a coupling constant of 5.6 Hz representing the methylene. The proton of the –NH appeared at δ_H 6.04 ppm with a coupling constant of 5.6 Hz as a triplet peak. Two doublet peaks at δ_H 6.64 ppm and δ_H 7.24 ppm with a coupling constant of 8.8 Hz were attributed to the benzene ring protons of C2′,6′-H and C3′,5′-H, respectively. The protons of the C2-H of the chromone ring were observed as a singlet at δ_H 8.15 ppm. Two doublet signals at δ_H 7.97 ppm (J = 8.8 Hz) and δ_H 6.87 ppm (J = 2.0 Hz) and a doublet–doublet signal at δ_H 6.98 ppm (J = 8.8 Hz, 2.0 Hz) were attributed to C5-H, C8-H, and C6-H of chromone ring, respectively. The ¹³C NMR spectrum of 3 showed the methylene carbon appeared at δ_C 38.12 ppm. The characteristic peak of carbonyl carbon appeared at δ_C 175.51 ppm and the signal at δ_C 162.48 ppm was attributed to C7. The correlation signal of benzene ring carbons were observed to assign the C-2′,6′ and C-3′,5′ signal to δ_C 131.24 and 114.23 ppm, respectively. The remaining aromatic and benzopyran ring carbons resonate at δ_C 102.08–157.66 ppm. Furthermore, the mass spectrum of compound 3 showed a molecular ion peak at m/z 343.9934 as [M + H]^- which also supported the proposed chemical structure of the product.

Close

Scheme 1

Reagents and conditions: (a) POCl₃, dimethyl formamide, 50℃, 2 h; (b) different anilines (R₂ArNH₂), Na(AcO)₃BH, CH₂Cl₂, rt, 0.5 h.

Export to PPT

2.2 α-Glucosidase inhibitory activity and structure–activity relationships (SARs)

All the synthesized compounds 3–34 were screened for their α-glucosidase inhibitory activity by in vitro enzyme assay (Table 1). Some of the synthesized derivatives displayed inhibitory activity against α-glucosidase with IC₅₀ values ranging from 11.72 ± 0.08 to 85.58 ± 2.30 μM. Acarbose as positive control (IC₅₀ = 832.22 ± 2.00 μM), the value of IC₅₀ is similar to prevenient literature report (Khan et al., 2014; Rahim et al., 2015; Taha et al., 2015). Among them, compound 4 showed the most active inhibitory activity against α-glucosidase with an IC₅₀ of 11.72 ± 0.08 μM. From the SAR point of view (Fig. 2), compounds 3–18 showed higher inhibition compared with 19–34, which demonstrated the importance of OH at the C-7 position for the inhibitory activity. The introduction of electron-withdrawing groups such as –NO_2, -Br, -Cl, -CF_3, and -F on the phenyl ring enhances the inhibitory activity against α-glucosidase. Comparing 20 (IC₅₀＞100 μM), 25 (IC₅₀＞100 μM), and 26 (IC₅₀ = 30.63 ± 0.19 μM), shows that substitution at para-position displayed the more potent inhibitory effect than other positions. Moreover, the introduction of the electron-donating group (–CH₃) on the phenyl ring could enhance the inhibitory activity against α-glucosidase, while the oxygen-containing electron donor group (–OCH₃ and –OCH₂CH₃) resulted in the reduction of the activity. In conclusion, these results show that there are differences in the biological activities of this type of compound due to the pattern of substitution between aniline and chromone.

Table 1 α-Glucosidase inhibitory activity of compounds 3–34. .

Compound	R1	R2	IC50 (μM)a,b
3	OH	4-Br	31.88 ± 0.85
4	OH	4-Cl	11.72 ± 0.08
5	OH	2-OCH2CH3	＞100
6	OH	4-OCH2CH3	＞100
7	OH	4-CF3	25.90 ± 1.05
8	OH	4-OCH3	85.58 ± 2.30
9	OH	H	85.37 ± 2.21
10	OH	4-OC6H5	75.33 ± 0.19
11	OH	4-CH3	22.62 ± 0.58
12	OH	3-CH3	60.47 ± 2.27
13	OH	2-CH3	65.06 ± 1.24
14	OH	3-OCH3	48.78 ± 1.71
15	OH	3-CF3	31.50 ± 3.64
16	OH	3, 4, 6-(OCH3)3	＞100
17	OH	4-F	23.90 ± 1.00
18	OH	2, 4-(CH3)2	72.81 ± 2.54
19	H	4-CH3	＞100
20	H	2-CF3	＞100
21	H	2,4-(CH3)2	＞100
22	H	4-F	＞100
23	H	2-F	＞100
24	H	H	＞100
25	H	3-CF3	＞100
26	H	4-CF3	30.63 ± 0.19
27	H	4-OCH2CH3	＞100
28	H	2, 5-Cl2	59.93 ± 4.23
29	H	2-OCH2CH3	＞100
30	H	4-NO2	44.45 ± 1.38
31	H	2-OCH3	＞100
32	H	4-OH	＞100
33	H	3-OH, 4-OCH3	30.63 ± 0.19
34	H	1-methyl-1H-indole	＞100
Acarbose	–	–	832.22 ± 2.00

a Acarbose is standard for α-glucosidase inhibition activity.

b IC₅₀ values represent as mean ± SD of three determinations.

Close

Fig. 2

SARs analysis of this class of compounds.

Export to PPT

2.3 Enzyme kinetic study

To further understand the inhibition mechanism, the most active compound 4 on α-glucosidase for the hydrolyzation of 4-nitrophenyl-β-d-galactopyranoside (pNPG) was studied. As shown in Fig. 3A, the lines of Lineweaver-Burk plot had a fixed intercept on the Y-intercept and X-slopes which increase with the concentration of the inhibitor. The values of V_max remained the same and the values of K_m increased with increasing concentration of the inhibitor, which suggested that compound 4 was a competitive inhibitor of α-glucosidase (Fig. 3B). Besides, the K_i value was determined as 1.88 × 10^-6 M by using Dixon plots analysis of enzymatic reactions (El Ashry et al., 2022). The inhibitory mechanism is similar to previous article (Sun et al., 2015). The result revealed that compound 4 could only bind with the free enzyme.

Close

Fig. 3

(A) Lineweaver-Burk plot for compound 4 (the concentrations of 4 were 0, 4, 6, 8, and 10 μM, and pNPG were 0.15, 0.2, 0.3, and 0.4 mM). (B) Secondary replot of slopes (Lineweaver-Burk plot) vs various concentrations of the compound.

Export to PPT

2.4 Circular dichroism spectra

Circular dichroism (CD) is an excellent tool for researching the interactions between proteins and small molecules by analyzing the changes in the protein secondary structure (such as α-helix, β-sheet, β-turn, and random coil) and microenvironment (Zheng et al., 2020). The CD spectra of α-glucosidase with and without compound 4 were shown in Fig. 4. The CD spectrum of the α-glucosidase sample showed two negative peaks at 209 nm and 220 nm, which were the characteristic peaks of α-helix. Furthermore, compound 4 changed the intensity of two peaks, especially at peak of 209 nm by the change of concentration. Results of CD spectra indicated that the binding of 4 and α-glucosidase impelled a modification in the secondary structure of the enzyme. Besides, the calculation result shows that the α-helix decreased from 41.2 to 39.0 %. The parameter of β-sheet (%) showed slight changes increased from 15.5 % to 15.8 %, and β-turn (%) showed irregular changes (Table 2). These results suggested that the structure of α-glucosidase became looser and unstable (Wang et al., 2020). In addition, in molecular docking, it is shown that compound 4 linked α-glucosidase enzyme to form strong hydrophobic binding, which emphasized that α-helix is the appropriate site for 4 binding in α-glucosidase (Jiang et al., 2014). In previous articles, malvidin-3-O-galactoside and diacylated anthocyanins also show a similar mechanism of action (Xue et al., 2022; Yang et al., 2021).

Close

Fig. 4

The circular dichroism spectra of 4. The concentration of α-glucosidase was 4 μM and its mixtures with different concentrations of 4 (0, 2, 4, 12 μM).

Export to PPT

2.5 Fluorescence quenching

The fluorescence spectrum can be used to research the interaction between biological macromolecules protein and small molecules (Nan et al., 2022). In order to study the interaction between the 4 and α-glucosidase, the fluorescence quenching experiment of α-glucosidase with 4 at various concentrations was conducted. According to Fig. 5A, the fluorescence intensity of α-glucosidase decreased with the increased concentrations of compound 4, indicating it was able to change the polarity and conformation of α-glucosidase. Regarding the addition of compound 4, the blue shift toward a lower wavelength is attributed to descending microenvironment polarity and increasing hydrophobicity of the tryptophan residue (Nan et al., 2022). To further understand the quenching mechanism of compound 4 on α-glucosidase, the quenching rate constant (K_sv) was calculated using the Stern-Volmer equation (Han et al., 2022).

Close

Fig. 5

(A) Fluorescence emission spectra of protein α-glucosidase at different concentrations of 4, T = 298 K, α-glucosidase = 3.5 U/mL. (B) The Stern-Volmer plots of 4 and α-glucosidase. (C) The double logarithm regression plots of 4 and α-glucosidase.

Export to PPT

F₀ stands for the fluorescence intensity without a quencher. F is the fluorescence intensity with the different concentrations of quencher. K_q denotes the quenching rate constant of the biomolecule. τ₀(10^-8 s) is the lifetime of the fluorophore in the absence of the quencher. [Q] is the concentration of quencher. Accordingly, a good linear relationship is observed in Fig. 5B which demonstrates the interactions that existed in a single quenching mechanism, static or dynamic. Moreover, according to Table 3, the value of K_q (92.3 × 10¹¹ Lmol⁻¹) is much>2.0 × 10¹⁰ Lmol⁻¹_. This result demonstrated that the ligand quenching type of compound 4 was a static quenching mechanism. Using the formula (II), the value of binding sites (n) and binding constant (K_a) with α-glucosidase were obtained (Fig. 5C).

log(F-F)/F = nlog[Q] + log K_a (II).

The binding constant (Ka) and the numerical value of binding sites (n) of 4 were 2.500 × 10⁴ L/mol and 1.33, respectively. The value of n was approximately equal to 1, which suggested just one binding site on the enzyme for 4 (Han et al., 2022).

2.6 Molecular docking

To further explain the interaction between compound 4 and α-glucosidase. The theoretical binding modes of compound 4 with α-glucosidase are shown in Fig. 6. Because the 3D structure of a-glucosidase is still unclear, the homology model was constructed using the known crystal structures of the Saccharomyces cerevisiae isomaltase (PDB ID: 3AJ7), according to our previous article (Wang et al., 2017). The binding energy of 4 with the enzyme was evaluated to be −7.7 kcal•mol⁻¹. Compound 4 adopted a Y-shaped conformation in the pocket of the α-glucosidase. It was located at the hydrophobic site, surrounded by the residues Phe-157, Phe-177, Leu-218, Pro-240, and Phe-300, forming a strong hydrophobic binding. Detailed analysis showed that the 4-chlorophenyl group of 4 formed hydrogen bond acceptor with the residue Lys-155. In addition, the chromone scaffold of compound 4 and the residues Glu-276 and Asp-349 formed hydrogen bond donor, and formed π-π stacking interaction with the residue Phe-157 (El Ashry et al., 2022). In conclusion, the above molecular docking explained the interactions between 4 and α-glucosidase, which afforded more information for further development of α-glucosidase inhibitors.

Close

Fig. 6

Compound 4 was docked to the binding pocket of the α-glucosidase. (A) The chemical structure of compound 4. (B) The overall structure of α-glucosidase with 4. (C) Binding pose of 4 at the binding site. (D) Binding pose of 4 in the surface of the binding pocket.

Export to PPT

2.7 Effects of compound 4 on postprandial blood glucose in mice

The reduction of postprandial blood glucose by 4 in vivo was measured using the sucrose loading test, and the level of glucose was tested at 0, 15, 30, 60, 90, and 120 min, respectively. The results were presented in Fig. 7, acarbose served as the positive control. In the experiment, the blood glucose levels of the mice in all groups rose to a peak value at 15 min and returned to the pretreatment level at 120 min. It was displayed that the level of blood glucose of the blank group was no remarkable changes throughout the whole process. Furthermore, the groups of compound 4 and acarbose were significantly lower than the control group, which could be concluded that with oral administration at a 20 mg/kg dose, acarbose and 4 were able to reduce glucose uptake in normal mice fed with saccharose.

Close

Fig. 7

(A) Effect of 4 on blood glucose in Kunming mice. The animal assays were conducted with the administration of sucrose at a dose of 2.5 g/kg. The mice of each assay were orally administrated along with water (control group), 4 (20 mg/mL), and acarbose (20 mg/mL), respectively. Data are presented as the mean ± SEM (n = 7, **p < 0.01); (B) The inhibition of postprandial sucrose response by 4 in normal Kunming mic. The incremental AUC_0-120 min in normal mice after sucrose administration. Asterisks indicate a significant difference (*p < 0.05,) compared with the control group.

Export to PPT

2.8 In vitro cytotoxicity

In vitro cytotoxicity testing is one of the important methods to measure the toxicity of the drug. The toxicity of compound 4 (0, 25, 75, 125, and 675 μM) toward the growth of the human embryonic kidney (HEK 293) cell line was evaluated by the CCK-8 method (Wen et al., 2016). The result was shown that 4 exhibited low cytotoxicity against normal human cell lines with an IC₅₀ value of 556.12 ± 6.42 μM, which was much higher than that of glucosidase inhibition (11.72 ± 0.08 μM). It was demonstrated that 4 exhibited an inhibitory effect on α-glucosidase at non-cytotoxic concentrations. Therefore, it could be concluded that 4 had a good safety drug for the application as an anti-diabetic agent.

2.9 ADME predictions

Optimization of absorption, distribution, metabolism, and excretion (ADME) profiles of biotherapeutics is critical in modern drug discovery. To measure the ADME properties of compounds 3–34, the pharmacokinetic parameters of all synthesized compounds were calculated by using the Molinspiration online property calculation toolkit (https://www.molinspiration.com/) and the results are presented in Table 4. According to the result, all the synthesized compounds 3–34 showed an excellent absorption range from 78.62 % to 94.43 %. The ADME properties of all the compounds have shown satisfactory results. The Lipinski rule of five states that an orally active drug has no more than one violation of the following criteria: molecular weight (MV) < 500, values of the octanol–water partition coefficient (log P) < 5, the number of hydrogen bond donor (NHD) < 5 and number of hydrogen bond acceptor (NHA) < 10 (Caron et al., 2018). From all these parameters, it can be predicted that among synthesized compounds, all compounds are likely to be orally active as they obeyed Lipinski’s rule of five.

4.1 Chemistry

All needful starting materials and reagents were supplied from commercial suppliers. TLC was performed on 0.20 mm Silica Gel 60 F254 plates (Qingdao Ocean Chemical Factory, Shandong, China). Nuclear magnetic resonance spectra (NMR) were recorded on a JEOL spectrometer (400 MHz) with TMS as an external reference and reported in parts per million. The Supplemental Materials contain sample ¹H NMR, ¹³C NMR, and HRMS spectrum for target products 3–34.

4.2 α-Glucosidase inhibitory activities

The α-glucosidase inhibitory activities of these compounds were screened according to the previous methods with minor modifications, acarbose was used as a positive control (Ji et al., 2021; Wang et al., 2017a, 2017b; Xu et al., 2020). The α-glucosidase (750 U) was purchased by Sigma Aldrich (St. Louis, MO, USA). Different concentrations of the tested compounds or acarbose (10 μL) and 150 μL of α-glucosidase solutions (0.2 U/mL) were added to a 96-well plate and the mixture was incubated at 37.5 °C for 15 min. Then, 40 μL of pNPG (1.25 mM) was added to the mixture. After incubation at 37 °C for 30 min, the absorbance of released p-nitrophenyl was measured at 405 nm using a microplate reader (Varioskan LUX, Thermo, USA); The inhibition percentage was obtained by the formula: Inhibition (%) = (1-ΔA sample/ΔA control) / *100 %. Where ΔA sample and ΔA control represent the absorbance with and without the test compounds, respectively. The IC₅₀ value was calculated by the linear with inhibition percentage and the concentrations of test compounds.

4.3 Enzyme kinetic study

The enzyme kinetics assays for α-glucosidase were performed using the method the above-mentioned. The substrate pNPG (0.15, 0.2, 0.3, and 0.4 mM) and compound 4 (0, 4, 6, 8, and 10 μM) were used in varying concentrations. The Michaelis constant (K_m) and maximum velocity of the reaction (V_max) were calculated by using the Lineweaver-Burk plot.

4.4 Circular dichroism spectra

The circular dichroism (CD) spectrum was measured from 190 to 260 nm by using Chirascan Plus (Applied Photophysics, Leatherhead, UK) at 25 °C. The α-glucosidase solution (4 μM) and its mixtures with different concentrations of 4 (0, 2, 4, and 12 μM) were measured. The CD spectra were recorded in the far UV region between 190 and 260 nm with a bandwidth of 1 nm on a Chirascan qCD. All collected CD spectra were subtracted from the PBS buffer signal. The CDNN 2.1 software was used to investigate the α-helix content change of α-glucosidase (Zhang et al., 2018).

4.5 Fluorescence quenching

Fluorescence spectra of samples were recorded on Cary Eclipse (Agilent Technologies) at 25 °C. Briefly, 1.5 uL of 4 stock solution (10 mM) was added seven times to the α-glucosidase solution (3.5 U/mL, 3 mL). The concentrations of 4 varied from 0 to 40 μM. In the following step, the spectra were recorded with the excitation wavelength of 280 nm and the emission wavelength ranging from 300 to 450 nm. Noteworthy, the fluorescence spectrum of PBS (pH = 6.8) was tested using the same method.

4.6 Molecular docking

Autodock Vina 1.1.2 was used for molecular docking to study the binding mode between compound 4 and α-glucosidase (Trott & Olson, 2010). The 3D structure of compound 4 was drawn by ChemBioDraw Ultra 14.0 and ChemBio3D Ultra 14.0. The AutoDockTools 1.5.6 package is used for generating the docking input files (David S. Goodsell, 1996; Morris et al., 2009; Sanner, 1999). The search grid of α-glucosidase was identified as center_x: −19.676, center_y: −7.243, and center_z: −21.469 with dimensions size_x: 15, size_y: 15, and size_z: 15. The value of exhaustiveness was set to 16. For Vina docking, the default parameters were used if it was not mentioned. The best-scoring pose as judged by the Vina docking score was chosen and visually analyzed using PyMOL 1.7.6 software (https://www.pymol.org);

4.7 Effects of 4 on postprandial blood glucose

To further study the potency of compound 4 to the reduction of postprandial blood glucose levels in vivo, the saccharose loading test in normal Kunming mice was performed by using the modifying previous reported method (Huang et al., 2020a, 2020b; Ji et al., 2021; Yu et al., 2020). The Kunming (25–30 g and eight-week-old) mice were purchased from Guizhou Medical University (Guiyang, China). All mice for experiments were fasted for 12 h but had free access to water, before treatment. In this experiment, all mice were randomly divided into four groups with 7 animals in each group. Gave different drugs or blank solutions to the four groups: the control group (0.5 % CMC-Na solution), acarbose group (20 mg/kg, dissolved in 0.5 % CMC-Na solution), compound 4 treated group (20 mg/kg, dissolved in 0.5 % CMC-Na solution), and blank group (0.5 % CMC-Na solution). Forty minutes after intragastric administration, the all animals were orally administrated with saccharose (2.5 g/kg, dissolved in 0.5 % CMC-Na solution), Their postprandial blood glucose was tested from the tail tip of mice by using Roche Accu-Chek Instant at 0 h, 0.25 h, 0.5 h, 1 h, 1.5 h, and 2 h. Kunming mice (25–30 g) were purchased from Laboratory Animal Center of Guizhou Medical University with the license number SCXK (Qian) 2018–0001. All animal procedures were performed in accordance with the Guidelines for Care and Use of Laboratory Animals and the experiments were approved by the animal ethics committee of Guizhou Medical University.

4.8 In vitro cytotoxicity

The cytotoxicity in vitro was evaluated by the CCK-8 method. First, HEK 293 cells were seeded in 96-well plates at a density of 1 × 10⁵ cells per well and then incubated for 24 h. The cells were incubated with 4 at various concentrations (0, 25, 75, 125, and 675 µM) for 48 h (37 °C and 5 % CO₂), respectively. 10 µL of CCK-8 reagent was added to the cell culture medium and then incubated at 37 °C for 1 h. Finally, absorbance was recorded at 450 nm using a multifunctional microplate reader (Varioskan LUX, Thermo Fisher Scientific, Vantaa, Finland). The relative cell viability (%) was calculated as the formula: (A_sample - A_blank)/(A_control - A_blank)*100.

4.9 ADME predictions

Calculations of ADME properties were done by using the molinspiration online property toolkit (https://www.molinspiration.com/). Pharmaceutically important properties of new drugs including mi log P, H-bond donors, H-bond acceptors, number of rotatable bonds, molecular volume, and molecular weight were analyzed in accordance with Lipinski’s rule of five. The rule of five indicates that poor absorption or permeation is expected when MW > 500, NHD > 5, NHA > 10 or log P > 5 (Choy & Prausnitz, 2011). The percentage of absorption (% ABS) was calculated using the expression: % ABS = 109–0.345 × TPSA.