Isolation of a New Phenolic Ester and Biflavonoids From Taxus mairei Leaves: Dual Mechanisms Targeting α-Glucosidase and High-Glucose-Induced VEGF Expression

Abstract

Objective

Diabetes mellitus and its complications, specifically diabetic retinopathy (DR), require novel therapeutics with enhanced safety profiles. Guided by the traditional use of Taxus mairei, this study aimed to isolate bioactive constituents from its leaf extracts and investigate their dual potential in regulating metabolic enzymes and protecting retinal cells against hyperglycemic stress.

Methods

The bioactive n-butanol fraction of T. mairei leaves was subjected to phytochemical isolation and structural elucidation. The isolated compounds were evaluated for inhibitory activities against α-glucosidase and α-amylase, supported by molecular docking analysis to understand structure–activity relationships. Furthermore, the potential retinal-protective effects were assessed using ARPE-19 retinal pigment epithelial cells under high-glucose conditions, focusing on cell viability and vascular endothelial growth factor A (VEGF-A) secretion.

Results

Nine compounds were identified, including a novel phenolic ester, 3-hydroxy-2-(3-hydroxy-4-methoxyphenyl)propyl 2-(4-hydroxyphenyl)acetate (1), and eight known constituents. The biflavonoids—amentoflavone (6), 7,7″-di-O-methylamentoflavone (7), and sotetsuflavone (8)—exhibited potent α-glucosidase inhibition (IC₅₀ values of 0.51–1.10 μg/mL) superior to acarbose, along with moderate α-amylase suppression. Docking studies indicated that compact folded conformations and extensive hydrogen-bonding networks strongly enhance binding affinity. Conversely, while compound 1 lacked enzyme inhibitory activity, it significantly attenuated high-glucose-induced VEGF-A overproduction and restored cell viability in ARPE-19 cells without cytotoxicity.

Conclusions

This study highlights T. mairei as a source of bioactive compounds with complementary antidiabetic potential. The identified biflavonoids offer promise for postprandial glycemic control, whereas the novel phenolic ester showed a distinct protective activity profile against high-glucose-induced VEGF-A overproduction. These findings suggest that T. mairei constituents have the potential to be developed into multifunctional therapeutics for diabetes and its microvascular complications.

Keywords

amentoflavone α-glucosidase inhibition diabetic retinopathy retinal pigment epithelium

Introduction

Diabetes mellitus is a global metabolic disorder characterized by chronic hyperglycemia resulting from impaired insulin secretion, insulin resistance, or both. It is broadly classified into Type 1 diabetes mellitus (T1DM), involving autoimmune destruction of pancreatic β-cells, and Type 2 diabetes mellitus (T2DM), which is strongly associated with obesity, insulin resistance, and progressive β-cell dysfunction.^1,2 Diabetes leads to serious long-term complications affecting both macrovascular systems (heart, brain, peripheral vasculature) and microvascular tissues (retina, kidney, nerves), which collectively contribute to disability and premature mortality.³ Although pharmacologic agents such as metformin, thiazolidinediones, sulfonylureas, and α-glucosidase inhibitors are widely used, their clinical application is often limited by adverse effects including gastrointestinal discomfort, hypoglycemia, or undesired weight changes. Therefore, the discovery of safer, natural compounds with multifunctional antidiabetic properties remains an important therapeutic goal.

Postprandial hyperglycemia control is central to T2DM management, and digestive enzyme inhibitors targeting α-glucosidase and α-amylase play a crucial role in delaying carbohydrate absorption and reducing glucose excursions. At the same time, early diabetic retinopathy (DR), one of the most prevalent microvascular complications, is increasingly recognized as a disease not only of endothelial dysfunction but also of retinal pigment epithelial (RPE) impairment. RPE cells serve as a major component of the blood–retinal barrier and are highly sensitive to metabolic stress. High-glucose exposure induces oxidative damage, mitochondrial dysfunction, and pathological overexpression of vascular endothelial growth factor A (VEGFA), all of which contribute to early DR progression.^4,5 Therapeutic compounds capable of simultaneously regulating glucose metabolism and protecting retinal cells may therefore offer dual benefits in diabetes management.

Taxus mairei, a species traditionally valued for its production of paclitaxel, contains abundant flavonoids—including flavones, biflavones, flavonols, dihydroflavones, and related subclasses—that exhibit notable anti-inflammatory, antioxidant, antimicrobial, and anticancer activities.^6,7 However, despite extensive phytochemical characterization of Taxus species, their potential roles in glycemic regulation and retinal protection remain insufficiently explored. In particular, the antidiabetic potential of biflavonoids and phenolic derivatives from T. mairei has not yet been comprehensively evaluated.

Preliminary bioactivity screening in our laboratory indicated that the n-butanol fraction of T. mairei leaves exhibited the most pronounced α-glucosidase and α-amylase inhibitory activities among all evaluated solvent fractions (data not shown), thereby guiding our decision to further isolate and characterize its major constituents.

The present study aimed to investigate the antidiabetic potential of compounds isolated from the bioactive n-butanol fraction of Taxus mairei leaves, focusing on both metabolic and retinal-protective mechanisms. To achieve this, we isolated and characterized a novel phenolic ester, 3-hydroxy-2-(3-hydroxy-4-methoxyphenyl)propyl 2-(4-hydroxyphenyl)acetate (1), along with eight known compounds (2–9). We evaluated their inhibitory activities against α-glucosidase and α-amylase, and further assessed the effects of compound 1 on ARPE-19 cell viability and VEGF-A secretion under high-glucose stress. By integrating enzymatic assays, retinal cell models, and molecular docking, this study provides new insights into the multifunctional antidiabetic potential of T. mairei.

Materials and Methods

General Experimental Procedures and Reagents

The human ARPE-19 cell line was purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA; CRL-2302™). Cells were maintained in DMEM/F12 (1:1) medium supplemented with 10% fetal bovine serum (FBS) and 2 mM L-glutamine. Cell detachment was performed using Trypsin-EDTA (all culture reagents were from Gibco, USA), and cells were washed with phosphate-buffered saline (PBS). Cultures were maintained in a CO₂ incubator at 37°C with 5% CO₂. Low-glucose (5 mM) and high-glucose (30 mM) media were prepared using D-glucose (Sigma-Aldrich, USA). Cell viability was assessed using a CCK-8 assay kit. VEGF-A secretion was quantified using a commercial ELISA kit. Total RNA was extracted using the RNeasy Total RNA Isolation Kit (QIAGEN, Germany), followed by reverse transcription with the MultiScribe™ Reverse Transcriptase Kit (Applied Biosystems, USA). qRT-PCR was conducted using SYBR Green PCR Master Mix (Applied Biosystems, USA) or Luna Universal Probe qPCR Master Mix (New England Biolabs, USA).

For enzyme inhibition assays, α-glucosidase (1 U/mL), p-Nitrophenyl-α-D-glucopyranoside (pNPG), Na₂CO₃, and acarbose (positive control) were obtained from Sigma-Aldrich (USA). α-Amylase and its corresponding substrate were used for the α-amylase inhibition assay. Compounds were isolated using silica gel, Sephadex LH-20^®, and RP-18 chromatographic media.

Absorbance for α-glucosidase and α-amylase inhibition assays was measured using a BioTek Synergy 4 Multi-Detection Microplate Reader (BioTek Instruments, USA). Quantitative real-time PCR was performed using an ABI PRISM^® 7900HT Fast Real-Time PCR System (Applied Biosystems, Italy). Compound purification was carried out using preparative high-performance liquid chromatography (HPLC) systems in combination with silica gel, Sephadex LH-20^®, and RP-18 column chromatography setups.

Structural elucidation of isolated compounds was supported by analytical instrumentation. Infrared (IR) spectroscopy was used to identify functional groups, while high-resolution electrospray ionization mass spectrometry (HRESIMS) provided accurate mass measurements (e.g., compound 1, m/z 355.1150 [M+Na]⁺). Nuclear magnetic resonance (NMR), including ¹H NMR, ¹³C NMR, HSQC, and HMBC experiments, was employed for structural confirmation.

Plant Materials

The leaves of T. mairei were purchased from Nantou County, Taiwan in January 2019. The raw plant material was identified by Distinguished Prof. Fang-Rong Chang. A voucher specimen (code no. KMU-GINP-Taxus-001) was stored in the Graduate Institute of Natural Products, College of Pharmacy, Kaohsiung Medical University, Taiwan.

Isolation of Phytochemicals

The general phytochemical isolation and chromatographic purification procedures were adapted from our previously established protocols.⁸ The dried leaves (6.44 kg) were extracted three times with MeOH (3 × 70 L), and the combined extracts were concentrated under reduced pressure to yield a crude MeOH extract (555 g). The extract was suspended in H₂O and partitioned successively with EtOAc and n-BuOH to afford the EtOAc-soluble fraction (185.68 g) and the n-BuOH-soluble fraction (136.09 g), respectively. The n-BuOH fraction was subjected to Diaion^® column chromatography and eluted with a gradient of H₂O to MeOH (100:0 → 0:100) to yield seven fractions (Fraction A to Fraction G). Fraction B (8.74 g) was chromatographed on a silica gel column and eluted with CH₂Cl₂/MeOH (15:1 → 0:1) to obtain eight subfractions (Fraction B-1 to Fraction B-8). Fraction B-2 was further purified by preparative RP-HPLC (CH₃CN/H₂O, 24:76) to yield compounds 1 (2.0 mg) and 2 (1.6 mg). Fraction C (11.78 g) was separated by silica gel column chromatography using CH₂Cl₂/MeOH (10:1 → 0:1) to produce six subfractions (Fraction C-1 to Fraction C-6). Fraction C-4 was further fractionated with CH₂Cl₂/MeOH (8:1 → 0:1) to obtain six subfractions (Fraction C-4-A to Fraction C-4-F). Fraction C-4-B was purified by RP-HPLC (MeOH/H₂O, 30:70) to yield compounds 3 (84.3 mg) and 4 (16.9 mg). Fraction C-5 was subjected to RP-18 column chromatography (MeOH/H₂O, 30:70 → 100:0) to give nine fractions (Fraction C-5-A to Fraction C-5-I). Fraction C-5-B was purified by Sephadex LH-20^® (MeOH) followed by a preparative HPLC step (CH₃CN/H₂O, 30:70) to obtain compound 5 (1.9 mg). Fraction G (11.10 g) was chromatographed on a Sephadex LH-20^® column (MeOH) to afford six fractions (Fraction G-1 to Fraction G-6). Fraction G-5 was further separated on a silica gel column eluted with CH₂Cl₂/MeOH (18:1 → 0:1) to produce five subfractions (Fraction G-5-A to Fraction G-5-F). Fraction G-2 was purified using silica gel chromatography eluted with CH₂Cl₂/MeOH (100:0 → 0:100) to yield compound 6 (58.9 mg). Finally, Fraction G-5-F was purified by preparative RP-HPLC (CH₃CN/H₂O, 50:50) to yield compounds 7 (32.7 mg), 8 (3.3 mg), and 9 (7.4 mg).

α-Glucosidase Activity Inhibition Assay

The α-glucosidase inhibitory activity was evaluated following a modified method described in the literature. Briefly, an α-glucosidase solution (1 U/mL in phosphate buffer, pH 6.8) was mixed with multiple concentrations of test compounds and pre-incubated at room temperature for 5 min. The reaction was initiated by adding 250 µL of p-nitrophenyl-α-D-glucopyranoside (pNPG) solution, followed by incubation at 37°C for 20 min. The enzymatic reaction was terminated by adding 500 µL of 9.4 mM Na₂CO₃. Acarbose was used as the positive control under the same conditions. The release of p-nitrophenol was quantified by measuring the absorbance at 405 nm using a microplate reader (BioTek Synergy 4 Multi-Detection Microplate Reader).⁹

α-Amylase Activity Inhibition Assay

The α-amylase inhibitory activity was determined according to a modified method described by Acar et al.¹⁰ with slight adjustments. Briefly, a human α-amylase enzyme solution (1 U/mL in phosphate buffer, pH 6.9) was mixed with multiple concentrations of test compounds and pre-incubated at 37°C for 10 min. The reaction was initiated by adding soluble starch solution as the substrate, followed by incubation at 37°C for 10 min. For color development, 3,5-dinitrosalicylic acid (DNS) reagent was added to terminate the enzymatic reaction, and the mixture was heated in a boiling water bath for 5 min to form the colored complex. After cooling to room temperature, the absorbance was measured at 540 nm using a BioTek Synergy 4 Multi-Detection Microplate Reader. Acarbose was used as the positive control. The percentage of α-amylase inhibition was calculated relative to the untreated enzyme control, and IC₅₀ values were obtained by nonlinear regression of inhibition curves.

Quantitative RT-PCR (qRT-PCR)

Total RNA was extracted from ARPE-19 cells cultured under different conditions using the RNeasy Total RNA Isolation Kit (QIAGEN, Valencia, CA, USA) following the manufacturer’s protocol.^4,5 RNA purity and concentration were determined spectrophotometrically, and samples with an A260/A280 ratio ≥ 1.8 were used for downstream analysis.

For miRNA and mRNA expression analysis, reverse transcription was performed using the MultiScribe™ Reverse Transcriptase Kit. The resulting cDNA was diluted and subjected to qRT-PCR using SYBR Green PCR Master Mix. All reactions were conducted on an ABI PRISM^® 7900HT Fast Real-Time PCR System (Applied Biosystems, Monza, Italy). Gene expression levels were quantified using the 2^-ΔΔCt method,⁴ with GAPDH serving as the internal control.⁵

All qRT-PCR reactions were conducted on an ABI PRISM^® 7900HT Fast Real-Time PCR System (Applied Biosystems, Monza, Italy). Gene expression levels were quantified using the 2^-ΔΔCt method. The expression of VEGF-A was assessed by qRT-PCR amplification using Luna Universal Probe qPCR Master Mix (New England Biolabs, MA, USA) with the following primers:

VEGF-A

Forward (F): TTCTGAGTTGCCCAGGAGAC

Reverse (R): TGGTTTCAATGGTGTGAGGA

GAPDH (internal control)

Forward (F): 5′-TGTGGGCATCAATGGATTTGG-3′

Reverse (R): 5′-ACACCATGTATTCCGGGTCAAT-3′

qRT-PCR analysis for Figure 1C was performed using three independent biological replicates, with one qPCR measurement obtained from each biological replicate. The qRT-PCR reaction conditions were as follows: an initial denaturation step at 95°C for 1 min, followed by 43 cycles of denaturation at 95°C for 15 s and extension at 60°C for 30 s. GAPDH was used as an internal control for normalization. Gene expression comparisons were analyzed using the 2^-ΔΔCt method. qRT-PCR analysis for Figure 1C was performed using three independent biological replicates, with one qPCR measurement obtained from each biological replicate.

Figure 1.

Effects of compound 1 on ARPE-19 cell viability, VEGF-A secretion, and VEGFA mRNA expression under high-glucose and low-glucose conditions. (A) ARPE-19 cells were cultured in high-glucose (30 mM) or low-glucose (5 mM) medium for 24, 48, and 72 h and treated with compound 1 at concentrations of 1, 5, 10, 20, 40, and 80 μg/mL. Cell viability was assessed using the CCK-8 assay. (B) ARPE-19 cells were cultured in high-glucose or low-glucose medium for 72 h and treated with compound 1 (5 and 20 μg/mL). VEGF-A concentration in the culture supernatant was measured using ELISA. (C) ARPE-19 cells were cultured in high-glucose or low-glucose medium for 72 h and treated with compound 1 (5 and 20 μg/mL). VEGFA mRNA expression was determined by qRT-PCR. Data are presented as mean ± SEM from three independent biological replicates (n = 3). For the CCK-8 assay, each biological replicate included three technical replicates; for ELISA, each biological replicate included two technical replicates; and qRT-PCR was performed without technical replicates (one measurement per biological replicate). Statistical comparisons between two independent groups were performed using the unpaired Student’s t-test. Exact p-values are provided in the Results section

Molecular Docking

The general in silico molecular docking framework, including protein preparation and energy grid generation using Discovery Studio, was guided by validated computational methodologies.¹¹ Molecular docking was performed to elucidate the structural basis underlying the inhibitory differences among compounds 6, 7, and 8. The crystal structure of human α-glucosidase (PDB ID: 5NN8, resolution < 3.0 Å) was retrieved from the Protein Data Bank. Protein preparation was conducted using Discovery Studio 3.0 (BIOVIA, Dassault Systèmes), including the removal of water molecules, ions, and alternate conformers, followed by hydrogen addition and force-field minimization.

Docking simulations were carried out using the LigandFit module, a shape-directed cavity-based docking algorithm.¹² Energy grid calculations were generated using Dreiding, CFF, and PLP1 scoring potentials. For each ligand, the top 100 docking poses were ranked using the Ludi_3 empirical scoring function, which evaluates hydrogen bonding, hydrophobic interactions, and entropic contributions to binding affinity.¹³ Torsion angle analysis was performed to classify poses with dihedral angles < 90° as cis- and > 90° as trans- conformations. The best-scoring conformations were compared across compounds 6, 7, and 8.

Statistical Analysis

All experimental data are presented as the mean ± standard error of the mean (SEM) from three independent biological replicates (n = 3). For the CCK-8 assay, each biological replicate included three technical replicates; for ELISA, each biological replicate included two technical replicates; and for qRT-PCR, one measurement was obtained from each biological replicate without technical replicates. For assays with technical replicates, the mean of the technical replicates was treated as a single biological replicate for statistical analysis. Statistical comparisons between two independent groups were performed using the unpaired Student’s t-test. Exact p-values for the comparisons related to Figure 1 are provided in the Results section. A value of p < 0.05 was considered statistically significant.

Results

Chemical Constituents of the Bioactive n-Butanol Fraction

In this study, the constituents of the bioactive n-butanol fraction obtained from the leaves of T. mairei were investigated, as this fraction exhibited the most pronounced anti-α-glucosidase and anti-α-amylase activities among all solvent partitions. Successive chromatographic separation of the n-butanol extract yielded one new compound, 3-hydroxy-2-(3-hydroxy-4-methoxyphenyl)propyl 2-(4-hydroxyphenyl)acetate (1), together with eight known constituents: methyl-4-hydroxyphenylacetate (2), roseoside (3), rhododendrin (4), trans-4-methoxycinnamic acid (5), amentoflavone (6), 7,7ʺ-di-O-methylamentoflavone (7), sotetsuflavone (8), and β-sitosterol (9) (Figure 2).

Figure 2.

Structures of compounds 1–9 isolated from the T. mairei

Structural Elucidation of Compound 1

Compound 1 was isolated from subfraction Fr. B-2 as a yellow powder. Its molecular formula, C₁₈H₂₀O₆, was determined from the HRESIMS quasi-molecular ion peak at m/z 355.11496 [M+Na]⁺ (calcd. 355.11521), indicating nine degrees of unsaturation. The IR spectrum of 1 displayed absorptions characteristic of hydroxyl (3381 cm^-1), carbonyl (1717 cm^-1), and aromatic functionalities (1614, 1453 cm^-1).

The ¹³C NMR spectrum showed 18 carbon resonances, and the combined DEPT and HSQC analyses classified these into one methyl, three methylenes (two oxygenated), eight methines (seven olefinic), and six quaternary carbons including an ester carbonyl. The ¹H and ¹³C NMR data (Table 1) indicated the presence of a methoxy group (δ_H 3.72, 3H, s; δ_C 55.3), a 1,4-disubstituted benzene ring [δ_H 7.15 (2H, d, J = 8.4 Hz, H-3′/H-5′), 7.28 (2H, d, J = 8.4 Hz, H-2′/H-6′); δ_C 115.8 (CH-3′/CH-5′), 130.5 (CH-2′/CH-6′), 124.8 (C-1′), 157.5 (C-4′)], and a 1,2,4-trisubstituted aromatic ring [δ_H 7.14 (¹H, m, H-2), 7.23 (¹H, d, J = 8.0 Hz, H-5), 7.02 (¹H, dd, J = 8.0, 2.0 Hz, H-6); δ_C 112.4 (CH-2), 115.9 (CH-5), 120.9 (CH-6), 131.6 (C-1), 146.6 (C-3), 148.1 (C-4)], as well as an ester carbonyl carbon at δ_C 171.7 (C-8′).

Table 1.

¹H (400 MHz, Pyridine-d₅) and ¹³C (100 MHz, Pyridine-d₅) NMR Data of 1

Position	δ_H¹, J in Hz³	δ_C (mult.)²	COSY	HMBC
1		131.6, C⁴	-
2	7.14, m	112.4, CH	-	C3, C4, C7
3		146.6, C	-	-
4		148.1, C	-	-
5	7.23, d, J = 8.0 Hz	115.9, CH	H6	C1, C4, C6
6	7.02, dd, J = 8.0, 2.0 Hz	120.9, CH	H5	C2, C3, C5, C7
7	3.46, quint, J = 13.2 Hz	47.4, CH	H8, H9	C1, C2, C6, C8, C9
8	4.79, dd, J = 10.8, 7.2 Hz 4.85, dd, J = 10.8, 7.2 Hz	65.9, CH₂	H7	C1, C7, C9
9	4.20, d, J = 6.4 Hz	63.4, CH₂	H7	C1, C7, C8, C8′
10	3.72, s	55.3, CH₃	-	C4
1′		124.8, C	-	C7′
2′	7.28, d, J = 8.0 Hz	130.5, CH	H3′	C4′, C6′, C7′
3′	7.15, d, J = 8.0 Hz	115.8, CH	H2′	C1′
4′		157.5, C	-	-
5′	7.15, d, J = 8.0 Hz	115.8, CH	H6′	C1′
6′	7.28, d, J = 8.4 Hz	130.5, CH	H5′	C2′, C4′, C7′
7′	3.68, s	40.2, CH	-	C1′, C2′, C6′, C8′
8′		171.7, C	-	-

¹Spectra recorded at 400 MHz in Pyridine-d₅.

²Spectra recorded at 100 MHz in Pyridine-d₅.

³J values (in Hz) in parentheses.

⁴Attached protons were deduced by DEPT experiment.

Figure 3 illustrates the key COSY and HMBC correlations used for establishing the planar structure of 1. The COSY spectrum established the spin system H₂-8/H-7/H₂-9, defining the connectivity within the aliphatic chain linking the two aromatic moieties. Further structural confirmation was obtained from detailed HMBC correlations. The methylene protons H₂-7′ showed long-range correlations to C-1′, C-2′, C-6′ and the ester carbonyl C-8′, while H₂-9 correlated with C-1, C-7, C-8, and C-8′, establishing the attachment of the aliphatic chain to both aromatic systems. The aromatic protons H-2′ and H-6′ correlated with C-4′ and C-7′, and H-3′ and H-5′ showed correlations to C-1′, confirming the substitution pattern of the 1,4-disubstituted benzene ring. Additionally, H-7′ correlated to C-1′, C-2′, and C-6′, supporting the same framework. For the 1,2,4-trisubstituted benzene ring, H-2 correlated with C-3, C-4, and C-6; H-5 with C-1, C-4, and C-6; and H-6 with C-2, C-3, and C-5. A diagnostic HMBC correlation from the methoxy protons H₃-10 to C-4 confirmed the methoxy group substitution at C-4.

Figure 3.

Key COSY and HMBC of 1

Taken together, these MS, IR, and comprehensive NMR analyses unequivocally established the structure of 1 as 3-hydroxy-2-(3-hydroxy-4-methoxyphenyl)propyl 2-(4-hydroxyphenyl)acetate. The measured optical rotation value was close to zero, suggesting that 1 may occur as a racemic mixture; therefore, its absolute configuration could not be determined.

Along with the new compound (1), eight previously reported constituents were also isolated from the bioactive n-butanol fraction. By comparing their spectroscopic data (¹H NMR, ¹³C NMR, and MS) with established literature values, these compounds were unambiguously identified as methyl-4-hydroxyphenylacetate (2), roseoside (3), rhododendrin (4), trans-4-methoxycinnamic acid (5), amentoflavone (6), 7,7ʺ-di-O-methylamentoflavone (7), sotetsuflavone (8), and β-sitosterol (9). These literature comparisons confirmed their structural identities (Figure 2). Detailed NMR assignments and literature comparisons for known compounds 2–9 are presented in the Supporting Information.

Inhibitory Effects of Compounds 1-9 on α-Glucosidase and α-Amylase

Inhibition of α-glucosidase and α-amylase is a well-established therapeutic strategy for delaying carbohydrate digestion and postprandial hyperglycemia in diabetes management. In this study, all isolated compounds (1–9) were evaluated for their inhibitory activities against these two key digestive enzymes, and the results are summarized in Table 2.

Table 2.

Inhibition of α-Glucosidase and α-Amylase Activity by Compounds 1 to 9

Compound	α-glucosidase		α-amylase
Compound	IC₅₀ (μg/mL)^a	Inh % (60 μg/mL)	IC₅₀ (μg/mL)^a	Inh % (60 μg/mL)
1	> 60	7.3 ± 4.28	> 60	24.71 ± 0.88^*
2	> 60	7.12 ± 2.25	> 60	4.18 ± 0.16
3	> 60	16.36 ± 4.09	> 60	2.78 ± 0.35
4	> 60	10.00 ± 1.63	> 60	4.18 ± 2.28
5	> 60	5.18 ± 1.92	> 60	0.58 ± 0.25
6	0.51 ± 0.04	99.90 ± 0.86^***	> 60	40.48 ± 1.45^**
7	0.93 ± 0.02	91.89 ± 8.87^***	> 60	39.48 ± 0.39^**
8	1.1 ± 0.11	89.89 ± 2.26^***	> 60	38.62 ± 2.17^**
9	> 60	-6.06 ± 4.84	> 60	9.30 ± 0.43
Acarbose ^b	118.7 ± 4.08	91.47 ± 0.19^***	7.09 ± 0.27	118.83 ± 1.66^***

Percentage of inhibition (Inh %) at 60 μg/mL concentration. Results are presented as mean ± S.E.M. (n = 3).

* p < 0.05, ** p < 0.01, *** p < 0.001 compared with the control.

^aConcentration necessary for 50 % inhibition (IC₅₀).

^bAcarbose (200 μg/mL) used as a positive control.

For α-glucosidase, compounds 6, 7, and 8 exhibited potent inhibitory activity with IC₅₀ values of 0.51 ± 0.04, 0.93 ± 0.02, and 1.10 ± 0.11 μg/mL, respectively, which were substantially stronger than that of the positive control acarbose (118.7 ± 4.08 μg/mL). At 60 μg/mL, compounds 6, 7, and 8 also showed exceptionally high inhibition percentages (99.90 ± 0.86%, 91.89 ± 8.87%, and 89.89 ± 2.26%, respectively). All other compounds (1–5, 9) displayed negligible inhibition (IC₅₀ > 60 μg/mL), indicating weak or no activity.

For α-amylase, compounds 6, 7, and 8 again demonstrated the highest activity among the tested isolates, with inhibition percentages of 40.48 ± 1.45%, 39.48 ± 0.39%, and 38.62 ± 2.17%, respectively, at 60 μg/mL. Although acarbose exhibited markedly stronger inhibition (118.83 ± 1.66%), compounds 6–8 still showed moderate and comparable suppression of α-amylase activity, whereas the other isolates displayed minimal inhibition (<10%).

Taken together, the biflavonoids amentoflavone (6), 7,7ʺ-di-O-methylamentoflavone (7), and sotetsuflavone (8) emerged as the principal enzyme inhibitors in the n-butanol fraction. Their dual inhibitory activities—strong against α-glucosidase and moderate against α-amylase suggest a potential role in regulating postprandial glucose levels and highlight these compounds as promising natural agents for diabetes management.

Molecular Docking Analysis

The markedly different inhibitory potencies of compounds 6–8 against α-glucosidase suggested that structural variations, particularly in substitution patterns, may play a critical role in modulating their binding behavior. To explore this structure–activity relationship, molecular docking was performed using the human α-glucosidase structure (PDB ID: 5NN8) and the LigandFit docking protocol with the Ludi_3 empirical scoring function.

Docking results indicated that all three biflavonoids, amentoflavone (6), 7,7ʺ-di-O-methylamentoflavone (7), and sotetsuflavone (8), adopt relatively compact folded conformations within the catalytic pocket, enabling efficient accommodation in the confined active site. These conformations arise from rotational flexibility between the two flavone units, rather than representing distinct stereoisomeric forms. No discrete comparison of cis/trans isomers was performed in this study.

Amentoflavone (6), which contains multiple free hydroxyl groups distributed across both flavone units, exhibited an enhanced capacity to form hydrogen-bonding interactions with polar residues within the active site. These hydroxyl functionalities can act as both hydrogen bond donors and acceptors, facilitating the formation of an extended hydrogen-bonding network with key catalytic and binding-site residues (e.g., Asp, Glu, and His), which are known to play essential roles in substrate recognition and catalysis. In contrast, 7,7ʺ-di-O-methylamentoflavone (7) bears methoxy substitutions at the 7 and 7ʺ positions, replacing hydroxyl groups and thereby reducing the number of available hydrogen-bonding sites. While these methoxy groups diminish hydrogen-bonding capacity, they increase the overall hydrophobic character of the molecule, potentially enhancing van der Waals and hydrophobic interactions with nonpolar residues (e.g., Phe, Leu, and Val) within the binding pocket.

Sotetsuflavone (8), although structurally related to amentoflavone, differs in the positional arrangement of hydroxyl and methoxy substituents across the biflavonoid scaffold. These positional variations may lead to subtle changes in molecular orientation and spatial accommodation within the catalytic pocket, influencing how the ligand aligns with key residues. As a result, even when the total number of functional groups is similar, differences in their spatial distribution can significantly affect the interaction pattern, binding stability, and overall affinity.

Consistent with these observations, the Ludi_3 scores were 949 for 6, 824 for 7, and 771 for 8 (Table 3), which aligned with their experimentally determined α-glucosidase inhibitory activities (IC₅₀ values of 0.51 ± 0.04, 0.93 ± 0.02, and 1.10 ± 0.11 μg/mL, respectively). Amentoflavone (6) showed the strongest predicted binding affinity, corresponding to the lowest IC₅₀ value.

Table 3.

The Ludi_3 Scores for Compound 6, 7, and 8

Compounds	6	7	8
Ludi_3 scores	949	824	771

Protective Effects of Compound 1 on ARPE-19 Cell Viability and VEGF-A Expression and Secretion

To evaluate the potential cytoprotective and anti-angiogenic effects of compound 1, ARPE-19 retinal pigment epithelial cells were exposed to either high-glucose (30 mM) or low-glucose (5 mM) culture conditions. As shown in Figure 1A, high-glucose exposure reduced ARPE-19 cell viability over time. Compared with the 30 mM glucose-alone group, compound 1 at 80 μg/mL significantly increased cell viability at 24 h (p = 0.001), 48 h (p = 0.031), and 72 h (p = 0.024), whereas compound 1 at 40 μg/mL significantly increased cell viability at 72 h (p = 0.017).

We next examined whether compound 1 modulates VEGF-A production under glucose stress. As shown in Figure 1B, VEGF-A secretion was significantly increased in the 30 mM glucose-alone group compared with the blank/low-glucose group (p = 0.018), whereas treatment with compound 1 at 20 μg/mL significantly reduced VEGF-A secretion compared with the 30 mM glucose-alone group (p = 0.024). In addition, qRT-PCR analysis showed that VEGFA mRNA expression was significantly elevated in the 30 mM glucose-alone group compared with the blank/low-glucose group (p = 0.006). Treatment with compound 1 significantly reduced VEGFA mRNA expression at 5 μg/mL (p = 0.047) and 20 μg/mL (p = 0.013), compared with the 30 mM glucose-alone group (Figure 1C). Taken together, these results indicate that compound 1 attenuates high-glucose-associated cellular injury and suppresses high-glucose-induced VEGF-A production in ARPE-19 cells.

Discussion

Diabetes management involves both the regulation of postprandial glucose metabolism and the protection of vulnerable retinal tissues from hyperglycemia-induced injury. In this context, our study revealed two mechanistically distinct yet complementary bioactivities among the isolated compounds from T. mairei. While the biflavonoids amentoflavone (6), 7,7ʺ-di-O-methylamentoflavone (7), and sotetsuflavone (8) exhibited strong inhibition of digestive enzymes, compound 1 demonstrated a unique cytoprotective role under glucose-induced stress in retinal pigment epithelial cells.

High-glucose exposure is well established to trigger oxidative stress, mitochondrial dysfunction, and pathological VEGFA upregulation in retinal cells-key events in the early development of diabetic retinopathy.^14-16 Consistent with these mechanisms, compound 1 significantly improved ARPE-19 cell viability and attenuated VEGFA hypersecretion under high-glucose conditions, suggesting potential anti-angiogenic and retinal-protective actions. Similar protective effects have been reported for phenolic esters and related phenylpropanoid derivatives in retinal or endothelial models.^17-19 However, the present study did not directly investigate oxidative stress- or inflammation-related signaling pathways. Therefore, although compound 1 showed protective effects against high-glucose-induced cellular injury and VEGF-A overproduction in ARPE-19 cells, the underlying mechanism remains to be clarified.

In contrast, compounds 6-8 primarily target gastrointestinal glucose digestion. Their potent α-glucosidase inhibitory activities align with prior findings that amentoflavone-type biflavonoids exhibit strong affinity toward glucosidase catalytic sites.^20-22 The IC₅₀ values we observed were markedly lower than that of acarbose, underscoring their therapeutic promise. Their moderate α-amylase inhibition complements this activity, supporting a balanced dual-enzyme inhibition pattern. As previously noted, excessive α-amylase inhibition may provoke gastrointestinal side effects, whereas moderate suppression is desirable for safe attenuation of postprandial glucose spikes.²³ Since the α-amylase used in the present assay was human α-amylase, the moderate inhibition observed for biflavonoids 6–8 reflects activity in a human enzyme system. However, direct comparisons with studies using α-amylases from other species should be interpreted cautiously, as enzyme origin may influence the apparent inhibitory potency.

Molecular docking analyses further supported the structure–activity relationships of these biflavonoids. Compounds 6-8 could be adopted within the active pocket of human α-glucosidase (PDB: 5NN8), and their Ludi_3 scores indicated favorable hydrogen-bonding and π–π interactions. This agrees with prior computational and structural studies demonstrating that biflavonoid rigidity, ortho-substitution, and flavone–flavone orientation strongly influence binding affinity toward glucosidases.^24,25

Importantly, molecular docking in the present study was restricted to α-glucosidase, and no docking analysis was conducted for α-amylase. Human α-amylase was used in the in vitro enzyme inhibition assay; however, the computational component of this work was specifically designed to examine the structural basis of α-glucosidase inhibition among compounds 6–8. Accordingly, the α-amylase-related results should be interpreted solely on the basis of the experimental inhibition assay, rather than docking-based structural prediction.^26,27

Collectively, our findings indicate that T. mairei contains two classes of bioactive compounds with divergent but synergistic antidiabetic potentials. The biflavonoids (6-8) offer potent control over carbohydrate digestion through α-glucosidase inhibition and moderate α-amylase suppression-an approach widely recognized as central to managing postprandial hyperglycemia.^28,29 Meanwhile, compound 1 provides retinal cytoprotection by enhancing cellular resilience and regulating VEGFA expression during glucose stress. The coexistence of metabolic regulation and retinal-protective effects within a single plant source positions T. mairei as a promising natural reservoir for multifunctional antidiabetic and vasculoprotective agents.

Amentoflavone, one of the major biflavonoids identified in the present study, has been previously reported to possess antidiabetic activity. In an earlier study, amentoflavone isolated from Cycas pectinata showed inhibitory activity against both α-glucosidase and α-amylase, with IC₅₀ values of 8.09 ± 0.023 μM and 73.6 ± 0.48 μM, respectively. A review on amentoflavone also summarized both its in vivo antidiabetic activity in diabetic mice and its digestive enzyme inhibitory activity, while a more recent systematic review reported α-glucosidase IC₅₀ values ranging from 3.28 to 24.43 μM across four studies. In the present study, compound 6 (amentoflavone) showed strong α-glucosidase inhibition (IC₅₀ = 0.51 ± 0.04 μg/mL) together with moderate α-amylase inhibition at 60 μg/mL, which is generally consistent with the reported antidiabetic potential of this biflavonoid. However, direct cross-study comparison should be interpreted cautiously because assay conditions, enzyme sources, and reporting units may differ among studies. For 7,7″-di-O-methylamentoflavone (compound 7) and sotetsuflavone (compound 8), directly comparable prior enzyme-based antidiabetic data were not readily identifiable in our focused literature search, so the present findings may help extend the currently limited comparative dataset for these methylated biflavonoids.

Flavonoids are recognized as an important class of natural α-glucosidase inhibitors. A review by Yin et al summarized 411 natural α-glucosidase inhibitors from medicinal plants, including 103 flavonoids, supporting the flavonoid scaffold as a major source of inhibitory activity. In this context, the strong α-glucosidase inhibitory activity of compounds 6–8 in the present study is consistent with the broader literature on flavonoid-type inhibitors. Since compounds 6–8 are biflavonoids, our findings further extend this evidence to biflavonoid constituents from T. mairei. However, direct numerical comparison across studies should be made cautiously because enzyme sources, assay conditions, controls, and reporting units differ.

In a comparative context, several agents have been reported to suppress VEGF/VEGFA-related responses under glucose stress in retinal cells. In the present study, compound 1 also reduced VEGF-A/VEGFA under high-glucose conditions and improved ARPE-19 viability, indicating a similar protective phenotype at the level of VEGF-related output.³⁰ However, because no upstream mechanistic assays were performed in the present study, it remains unclear whether compound 1 acts through pathways similar to those reported for other agents.

However, the present study has several limitations. First, the isolated compounds were evaluated individually without combination experiments, precluding the assessment of potential synergistic or antagonistic interactions on α-glucosidase or α-amylase inhibition. Additionally, differences in solubility and bioavailability were not rigorously controlled beyond nominal assay concentrations, and retinal safety evaluation was performed only for compound 1, with no data available for biflavonoids 6–8 in ARPE-19 cells. Second, although compound 1 was evaluated over a broad concentration range (1–80 μg/mL) in the CCK-8 assay, VEGF-related analyses were limited to two concentrations (5 and 20 μg/mL). Consequently, a full dose–response relationship for VEGF-A regulation could not be established. VEGF-A reduction was not directly correlated with cell viability, and the absence of a positive control limits the interpretation of the magnitude and specificity of the observed effects. Furthermore, the temporal pattern and kinetic parameters of VEGF-A secretion, as well as the effects of compound 1 on other angiogenesis-related factors, were not evaluated, as the present study focused primarily on specific VEGF-A endpoints. Third, no mechanistic assays were conducted to determine whether the effects of compound 1 are mediated by antioxidant, anti-inflammatory, or direct VEGF-A-inhibitory pathways. Key markers of oxidative stress and inflammation (e.g., ROS, MDA, antioxidant enzymes, and inflammatory signaling) were not assessed; thus, the findings are limited to a protective phenotypic effect under high-glucose conditions. In addition, in silico molecular docking or simulations targeting VEGF-A-related signaling proteins were not performed for compound 1, leaving its precise molecular targets to be clarified in future studies. Fourth, the cell-based experiments evaluating the protective effects of compound 1 were conducted using a relatively small number of biological replicates (n = 3). Although this sample size was sufficient to detect statistically significant differences under highly controlled in vitro conditions, it inherently limits the statistical power and the generalizability of our findings to more complex biological systems. Future studies incorporating larger sample sizes and in vivo models are necessary to robustly validate these cytoprotective and anti-angiogenic effects. Finally, the molecular docking analyses provided only a static representation of ligand–enzyme interactions. Although the docking results qualitatively aligned with experimental activity, no statistical correlation analysis was performed due to the limited number of compounds, and structure–activity relationships were not quantitatively analyzed. Moreover, docking does not capture dynamic or thermodynamic behavior; therefore, further studies, such as molecular dynamics simulations or experimental binding assays, are required to validate binding stability and mechanistic relevance. Additionally, docking-based interpretations are inherently constrained by the rigid-receptor approximation, which neglects potential induced-fit conformational adjustments of the enzyme upon ligand binding. Empirical scoring functions like Ludi_3 provide a qualitative ranking of binding poses rather than a definitive thermodynamic binding free energy; thus, these scores should be interpreted as predictive indicators rather than absolute measures of in vitro affinity. Regarding future perspectives, in vivo validation is essential to translate these in vitro and in silico findings into preclinical realities. Specifically, the postprandial glycemic control of the potent biflavonoids (6–8) will be evaluated in animal models of diabetes, such as streptozotocin (STZ)-induced or db/db mice, to verify their oral efficacy and pharmacokinetics. Additionally, the retinal-protective effects of the novel phenolic ester (1) against high-glucose-induced VEGF-A overproduction will be validated using diabetic retinopathy mouse models by assessing retinal vascular permeability and microvascular leakage.

Conclusions

Taxus species are known to contain a chemically diverse array of flavonoids, including flavones, biflavones, flavonols, dihydroflavones, dihydroflavonols, and flavanols, many of which have been associated with important biological functions. Among these constituents, amentoflavone (6) has attracted considerable scientific interest due to its broad-spectrum pharmacological activities, including reported antibacterial, antileishmanial, and antiviral effects. In the present investigation, one new phenolic ester (1) and eight known compounds (2–9) were isolated from the n-butanol fraction of T. mairei leaves, further supporting the chemical richness of this species. Our findings highlight two complementary antidiabetic mechanisms within the T. mairei constituents. The biflavonoids amentoflavone (6), 7,7ʺ-di-O-methylamentoflavone (7), and sotetsuflavone (8) exhibited strong α-glucosidase inhibition and moderate α-amylase suppression, with compound 6 showing the highest activity. Molecular docking analyses demonstrated that the potent biflavonoids adopt compact folded conformations and utilize extensive hydrogen-bonding networks to achieve superior binding interactions with α-glucosidase, providing a structural explanation for the observed potency. In contrast, compound 1 did not demonstrate digestive enzyme inhibition but instead exerted a distinct cytoprotective role in ARPE-19 retinal pigment epithelial cells. It significantly improved cell viability under high-glucose stress and attenuated VEGF-A overproduction, suggesting a protective effect under hyperglycemic conditions, although the precise mechanism was not investigated in the present study. Collectively, these results support T. mairei as a promising natural source of multifunctional antidiabetic agents. The biflavonoids 6–8 may contribute to postprandial glucose regulation, while compound 1 offers retinal protective effects under hyperglycemic stress. Further studies—including mechanistic analyses, in vivo efficacy evaluation, and safety profiling—are warranted to validate the therapeutic potential of these compounds and explore their development as complementary agents for diabetes management and microvascular protection.

Supplemental Material

Supplemental Material - Isolation of a New Phenolic Ester and Biflavonoids From Taxus mairei Leaves: Dual Mechanisms Targeting α-Glucosidase and High-Glucose-Induced VEGF Expression

Supplemental Material for Isolation of a New Phenolic Ester and Biflavonoids From Taxus mairei Leaves: Dual Mechanisms Targeting α-Glucosidase and High-Glucose-Induced VEGF Expression by Mon-Der Cho, Shang-Yu Chou, Yu-Ming Hsu, Chi-Ying Li, Yang-Chang Wu, Yi-Hong Tsai, Fang-Rong Chang in Natural Product Communications

Footnotes

This research was financially supported by grants from the National Science and Technology Council (NSTC), the Drug Development and Value Creation Research Center at Kaohsiung Medical University, and the Kaohsiung Municipal Min-Sheng Hospital Project. The authors gratefully acknowledge Jing Koou Enterprise Co., Ltd. (No. 4, Aly. 23, Ln. 288, Juemin Rd., Sanmin District, Kaohsiung, Taiwan; TEL: +886-7-3854349) for their technical assistance and support in resource coordination that contributed to this work. The authors also appreciate the Center for Research Resources and Development in Kaohsiung Medical University for the assistance in protein identification, flow cytometry and confocal image analysis. We also sincerely thank Ho Ni Cheng and part-time project assistants in Prof. Fang-Rong Chang’s lab for their valuable assistance in the experimental work conducted in this study.

ORCID iDs

Yi-Hong Tsai

Fang-Rong Chang

Ethical Considerations

Ethical approval is not applicable for this article.

Consent to Participate

There are no human subjects in this article and informed consent is not applicable.

Author Contributions

Conceptualization, M.D. Cho, F.R. Chang, and Y.H. Tsai; methodology, Y.M. Hsu; software, S.Y. Chou; validation, Y.H. Tsai, F.R. Chang, and Y.C. Wu; formal analysis, Y.H. Tsai; investigation, M.D. Cho; resources, S.Y. Chou; data curation, M.D. Cho and Y.M. Hsu; writing—original draft preparation, M.D. Cho and C.Y. Li; writing—review and editing, Y.H. Tsai and F.R. Chang; visualization, Y.H. Tsai; supervision, Y.C. Wu and F.R. Chang; project administration, M.D. Cho and S.Y. Chou; funding acquisition, F.R. Chang and Y.C. Wu. All authors have read and agreed to the published version of the manuscript.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was partly supported by funds from the National Science and Technology Council (113-2320-B-127-001, 113-2321-B-037-002, 112-2321-B-037-007, 111-2321-B-037-004, 115-2320-B-127-001), Ministry of Science and Technology (MOST 111-2320-B-037-020-MY3), and the Drug Development and Value Creation Research Center, Kaohsiung Medical University (KMU-TC112A03-3). This work was also supported by the Kaohsiung Municipal Min-Sheng Hospital Project (KMSH-11303), awarded to Mon-Der Cho.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The data presented in this study are available in the article and the Supplementary Materials.*

Statement of Human and Animal Rights

This article does not contain any studies with human or animal subjects.

Institutional Review Board Statement

Ethical review and approval were not required for this study because no human participants, human-derived clinical specimens, identifiable personal data, or animal experiments were involved.

Supplemental Material

Supplemental material for this article is available online.

Appendix

References

American Diabetes Association . Classification and diagnosis of diabetes. Diabetes Care. 2015;38:S8-S16.

American Diabetes Association . Diagnosis and classification of diabetes mellitus. Diabetes Care. 2014;37:S81-S90.

Sakran

Abu-Elheja

Khateeb

, et al. The many faces of diabetes: Is there a need for re-classification? A narrative review. BMC Endocr. Disord. 2022;22:9.

Gong

Xie

Liu

. Differentially expressed microRNAs in the development of early diabetic retinopathy. J. Diabetes Res. 2017;2017:4727942.

Bahrami

Shen

Zhu

Zhang

Chang

Gillies

. Effects of VEGF inhibitors on human retinal pigment epithelium under high glucose and hypoxia. Clin. Exp. Ophthalmol. 2019;47:1074-1081.

Wang

Liu

Zhang

, et al. Transcriptome sequencing reveals regulatory mechanisms of Taxol synthesis in Taxus wallichiana var. mairei. Int. J. Genomics. 2019;2019:1596895.

Wei

Wang

. Flavonoid components, distribution, and biological activities in Taxus: A review. Molecules. 2023;28:1-22.

Huang

Lin

Liu

, et al. Lipid metabolism and its mechanism triggered by supercritical CO₂ extract of adlay (Coix lacryma-jobi var. ma-yuen (Rom. Caill.) Stapf) bran in high fat diet-induced hyperlipidemic hamsters. Front. Pharmacol. 2021;12:785944.

Zheng

Wang

Liu

Sun

Zhao

Chen

. Screening and extraction process optimization for potential α-glucosidase inhibitors from quinoa seeds. Food Med. Homology. 2024;1:9420004.

10.

Xiao

Storms

Tsang

. A quantitative starch-iodine method for measuring α-amylase and glucoamylase activities. Anal. Biochem. 2006;351:146-148.

11.

Premnath

Liu

Perkins

Abbott

Anderson

McInnes

. Fragment based discovery of arginine isosteres through REPLACE: towards non-ATP competitive CDK inhibitors. Bioorg. Med. Chem. 2014;22:616-622.

12.

Venkatachalam

Jiang

Oldfield

Waldman

. LigandFit: A shape-directed docking algorithm for ligand–protein active sites. J. Mol. Graph. Model. 2003;21:289-307.

13.

Bohm

. Development of a simple empirical scoring function to estimate binding constants. J. Comput. Aided Mol. Des. 1994;8:243-256.

14.

Kowluru

Chan

. Oxidative stress and diabetic retinopathy. Exp. Diabetes Res. 2007;2007:43603.

15.

Cunha-Vaz

Bernardes

Lobo

. Blood–retinal barrier in diabetic retinopathy: Vasogenic vs inflammatory mechanisms. Prog. Retin. Eye Res. 2011;31:473-488.

16.

Guo

, et al. High glucose induces abnormal VEGF expression and apoptosis in RPE cells. Mol. Med. Rep. 2017;16:7027-7034.

17.

Lim

Wang

, et al. Phenolic acids protect retinal pigment epithelial cells from oxidative stress. Curr. Eye Res. 2019;44:264-271.

18.

Kim

Roh

, et al. Resveratrol inhibits VEGF expression and angiogenesis in retinal cells. Biochem. Biophys. Res. Commun. 2011;412:454-460.

19.

Ribeiro

Freitas

Silva

AMS

, et al. Antioxidant and antiangiogenic activity of phenolic compounds. Food Chem. 2015;183:235-241.

20.

Liu

Wang

Luo

, et al. Amentoflavone as a natural α-glucosidase inhibitor. Food Funct. 2019;10:1940-1950.

21.

Saito

Yoshizaki

Matsunami

. Biflavonoids as potent α-glucosidase inhibitors. Bioorg. Med. Chem. Lett. 2021;31:127722.

22.

Boonnak

Karalai

Chantrapromma

, et al. Structure–activity relationship of biflavonoids as α-glucosidase inhibitors. Phytochemistry. 2014;108:147-152.

23.

Oki

Matsui

Osajima

. Inhibitory effects of α-glucosidase and α-amylase inhibitors from natural sources. J. Agric. Food Chem. 1999;47:5504-5509.

24.

Chen

Mukwaya

Wong

Zhang

. Molecular docking of flavonoids with α-glucosidase. J. Agric. Food Chem. 2012;60:9190-9199.

25.

Dai

Mumper

. Structure–activity relationships of flavonoids. Molecules. 2010;15:7313-7352.

26.

Hou

Wang

. Assessing protein homology models using docking reliability. J. Comput. Chem. 2011;32:866-877.

27.

MacGregor

Janeček

Svensson

. Relationship of sequence and structure to specificity in α-amylases. Biochim. Biophys. Acta. 2001;1543:283-292.

28.

Ceriello

. Postprandial hyperglycemia and diabetes complications. Diabetes. 2005;54:1-7.

29.

Tundis

Loizzo

Menichini

. Natural products as α-amylase and α-glucosidase inhibitors. Mini Rev. Med. Chem. 2010;10:315-331.

30.

Yin

Zhang

Feng

Zhang

Kang

. α-Glucosidase inhibitors isolated from medicinal plants. Food Sci. Hum. Wellness. 2014;3:136-174.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

1.09 MB