Antioxidant and α-Glucosidase Inhibitory Activities Guided Isolation and Identification of Components from Mango Seed Kernel

In the present study, petroleum ether, dichloromethane, ethyl acetate, and n-butanol fractions of mango seed kernel exhibited different degrees of antioxidant and α-glucosidase inhibitory activity. Thus, quantitative and qualitative analysis of the petroleum ether fraction was conducted by GC-MS. Among identified components, four unsaturated fatty acids had never been reported in natural products before, together with 19 known components. In addition, 17 compounds were isolated and elucidated from other active fractions. Compounds 2, 9, 15, and 17 were isolated for the first time from Mangifera genus. Compounds 1 and 2 exhibited prominent DPPH radical scavenging and α-glucosidase inhibitory effects. In order to further explore their mechanism of α-glucosidase inhibition, their enzyme kinetics and in silico modeling experiments were performed. The results indicated that 1 inhibited α-glucosidase in a noncompetitive manner, whereas 2 acted in a competitive manner. In molecular docking, the stability of binding was enhanced by π-π T-shaped, π-alkyl, π-π stacked, hydrogen bond, and electrostatic interactions. Thus, compounds 1 and 2 were determined to be new potent antioxidant and α-glucosidase inhibitors for preventing food oxidation and enhancing hypoglycemic activity.


Introduction
Type 2 diabetes is a chronic and complex metabolic disease characterized mainly by hyperglycemia, insulin resistance, and insufficient insulin secretion [1]. When carbohydrates are swallowed, dietary polysaccharides are absorbed by the hydrolysis effects of enteral α-glucosidase to obtain free monosaccharides [2]. Free monosaccharides assimilated in the blood can lead to hyperglycemia caused by type 2 diabetes [3]. One of the diabetes treatments is to inhibit the enzyme activity of α-glucosidase to minimize the formation of blood glucose. α-Glucosidase inhibitors extend the overall carbohydrate digestion time and decrease the absorption of glucose, which slow postprandial glucose elevation [4]. Several drugs, α-glucosidase inhibitors, such as voglibose, acarbose, and miglitol, are now available to treat the patients who suffer from postprandial hyperglycemia. However, a long-term use of these inhibitors has been associated with severe side effects, such as diarrhea and vomiting [5]. Hyperglycemiainduced oxidative stress is implicated in the onset and progression of diabetes and, if left untreated, can lead to severe complications [6]. Free radicals and reactive oxygen species are products of normal cellular metabolism and are extremely reactive and potentially damaging transient chemical species [7]. Numerous studies showed that over generation of free radicals and reactive oxygen species could be harmful to human health and trigger many diseases, such as cancer, arteriosclerosis, inflammatory disorders, and aging processes [7]. Therefore, the development of highefficiency and low-toxicity antioxidant and α-glucosidase inhibitors from natural resources has become one research hotspot.
Mango (Mangifera indica L.) is one of the most popular fruit worldwide, which belongs to the Anacardiaceae and grown in many places around the world, particularly in tropical and subtropical countries. This fruit is available as a dietary supplement due to their high nutritional value, such as providing dietary fiber, fats, proteins, carbohydrates, and phenolic compounds, which are vital to human growth, development, and health [8]. Mango seed kernel, a traditional Chinese medicine, is widely available and used in China, and its bioactivities had been recorded. Chaianun and Praphan [9] indicated that the 95% ethanol extracts from two different kinds of mango seed kernel, namely, Kaew and Choke-Anan, both had good α-glucosidase inhibitory activity with IC 50 values of 163:19 ± 2:33 and 113:51 ± 5:85 μg/mL, respectively. Emmanuel, Ganiyu, Afolabi, Aline, and Margareth [10] had demonstrated that the methanol extract of mango seed kernel could significantly inhibit the activity of α-glucosidase. Although it was used for centuries as a traditional folk prescription in China, there are few studies on chemical components in mango seed kernel and bioactivities of its monomeric compounds. Meanwhile, no investigations on the composition and activity of its petroleum ether fraction have been previously reported.
Consequently, a bioassay-guided fractionation method was used by us to research the bioactive components in mango seed kernel. Its crude extract was successively partitioned with petroleum ether, dichloromethane, ethyl acetate, and n-butanol, respectively. The ethyl acetate and n-butanol fractions showed significant DPPH radical scavenging activity. Petroleum ether, dichloromethane, and ethyl acetate fractions revealed moderate α-glucosidase inhibitory activity. These finding prompted us to perform a detailed chemical evaluation, leading to the identification of 23 components from the petroleum ether fraction by LC-MS and the isolation and purification of 17 compounds from other active fractions of mango seed kernel. Antioxidant and α-glucosidase inhibitory activities of the identified components from the petroleum ether fraction had been reported in many references. Bioactivities of 13 isolated compounds from other active fractions were tested, and compounds 1 and 2 were found to be active. The enzyme kinetic parameters of active compounds were also determined. The results of the molecular docking were used to visualize how they bind to α-glucosidase, providing insights into α-glucosidase inhibition mechanism of the active compounds.

Antioxidant and α-Glucosidase Inhibitory Activities
Evaluation of the Extract and Fractions from Mango Seed Kernel. The seeds of Mangifera indica L. (1 kg) were removed from the shell, which homogenized with a homogenizer. Dried powders (600 g) of mango seed kernel were exhaustively extracted with 60% ethanol for 1 h (20 L × 2 times) at room temperature and were concentrated under reduced pressure. The concentrated solution was successively partitioned with petroleum ether, dichloromethane, ethyl acetate, and n-butanol.
2.3.1. DPPH Assay. The DPPH radical scavenging activity assay was determined according to the reported method with minor modifications [11]. Briefly, extract and four fractions were dissolved in methyl alcohol at concentrations ranging from 62.5 to 1000 μg/mL, which were prepared to mix with a methyl alcohol solution of DPPH (0.2 mM, 100 μL) on a 96-well plate. After that, shaken gently and stood at room temperature for 30 min, the optical density (OD) 517 nm in each well were determined by a microplate reader. Triplicates of each sample were run, and the mean values were calculated. Ascorbic acid was chosen as positive control. The percentage inhibition (%) for each sample was calculated by the following formula: Oxidative Medicine and Cellular Longevity DPPH solution was replaced by a methanol solution, which was used as control, and samples were replaced by the methanol solution as the black. All the results were expressed as means ± standard deviation (SD). IC 50 values denoted the concentration of the sample required to scavenge 50% of the DPPH radical.
2.3.2. α-Glucosidase Inhibitory Assay. The α-glucosidase inhibitory assay was performed on a 96-well plate by using a microplate reader according to the published method [12] with minor modifications. Extract and four fractions were required to dissolve in 5% DMSO at the concentrations ranging from 62.5 to 1000 μg/mL. In brief, 20 μL of α-glucosidase (1.3 U/mL) and 20 μL sample were added to 20 μL of 0.1 M potassium phosphate buffer (pH 6.8). After 5 min incubation at 37°C, 20 μL of 4-pNPG (p-nitrophenyl-α-D-glucopyranoside) (2.5 mM) was added and incubated for 15 min; then, 80 μL of Na 2 CO 3 (0.2 M) was added to stop the reaction. The optical density (OD) values were determined using a microplate reader at 405 nm. Triplicates of each sample were run, and the mean values were calculated. Acarbose was used as positive control. The percentage inhibition (%) was calculated as the following formula: A a : absorbance of the sample group with enzyme: A b : absorbance of the sample control group without enzyme: A c : absorbance of the control group without samples: A d : absorbance of the blank control group without samples and enzyme: The IC 50 was calculated by using SPSS software.

GC-MS Analysis of the Bioactive Fraction (Petroleum
Ether Fraction) from Mango Seed Kernel. The GC-MS analysis of constitutes in the bioactive fraction (Petroleum ether fraction) of mango seed kernel was carried out on a Shimadzu (TQ-8040) series GC-MS system (n) equipped with an AOC-20i autosampler [13]. The operational process of GC-MS analysis was as follows: rising from 40°C to 130°C at a rate of 4°C/min and holding for 2 min, rising from 130°C to 150°C at a rate of 2°C/min and holding for 3 min, rising from 150°C to 180°C at a rate of 2°C/min and holding for 3 min, and finally enhanced to 210°C and held isothermally for 5 min. Chemical constituents were then pointed out based on the mass spectra and retention times with already known compounds in the NIST14 and NIST14s (National Institute of Standards and Technologies, Mass Spectra Libraries, Gaithersburg, MD, USA) [13].  The n-butanol fraction (25.3 g) was subjected to a reversed-phase ODS, and elution with a gradient of MeOH (5 ⟶ 50%, v/v) in water to obtain four subfractions (Fr.1-Fr.4). Compounds 9 (6.2 mg) and 17 (5.2 mg) were obtained from Fr.1 by preparative TLC (EtOAc-MeOH-H 2 O, 5 : 1 : 0.9). 1, 2, 3, 4, 6-Penta-O-galloyl-β-D-glucoside (1): maple powder; 1 H and 13 C NMR data matched literature values [14].
1-Glycerol gallate (9): pale yellow powder; 1 H and 13 C NMR data were in agreement with the literature values [21].
Ethyl gallate (10): colorless powder; 1 H and 13 C NMR data were similar to those recorded by Chen et al. [22].
Caffeic acid (12): white needle crystal; 1 H and 13 C NMR data were in agreement with those dealt with in the literature [24].

Determination of Antioxidant and α-Glucosidase
Inhibitory Activities of the Isolated Compounds. The antioxidant and α-glucosidase inhibitory activities of each isolated compound at concentrations ranging from 15.625 to 500 μM (DPPH) and from 125 to 2000 μM (α-glucosidase inhibitory), respectively, were determined using the methods described in Section 2.3.

Inhibitory Kinetic Analysis.
The kinetic mechanisms of active compounds 1 and 2 towards α-glucosidase were determined by the graphical views of Dixon and Lineweaver-Burk plots [30]. The concentration ranges for the substrates were 0.8-5.0 mM. The concentration of the enzyme was kept constant at 1.3 U/mL. Lineweaver-Burk plots were established to evaluate the type of inhibition. Dixon plots were used to calculate inhibitory constant (K i ) values of the tested compounds. Each kinetic analysis was implemented in triplicate from which mean K i values.

Molecular Docking Calculation.
To investigate the interaction between active compounds and α-glucosidase, molecular simulations were generated by using AutoDock Vina (version 1.1.2). The crystal structure of isomaltase from Saccharomyces cerevisiae (PDB ID: 3A4A) was diffusely used in the molecular docking analysis [31]. The 3D structure of compounds 1-2 and acarbose was built by ChemBioDraw Ultra 14.0 and energy minimized by ChemBio3D Ultra 14.0 software. The ligands were modified for docking by incorporating nonpolar hydrogen and setting revolvable bonds. As for receptor, all the water molecules were removed and necessary hydrogen atoms were added to the protein. Compounds 1 and 2 and acarbose were considered fully flexible. The grid for docking was formed using X-, Y-, and Z-axes. The docking runs were performed using the Lamarckian genetic algorithm (LGA). In addition, acquiescent arguments were used unless otherwise indicated. The docked formation with the lowest energy value between α-glucosidase and active compounds was obtained and visually evaluated using PyMoL 1.7.6 software (http://www.pymol.org/).

Statistical
Analysis. The statistical analysis was conducted by the Statistical Program for Social Sciences (SPSS; Chicago, IL) version 13.0 for Windows. All the data were expressed as mean values ± standard deviation (n = 3). Statistical significance between groups was assessed by using oneway analysis of variance followed by the LSD test in the condition of variance homogeneity and Dunnett's T3 test in the condition of variance heterogeneity. Differences were considered significant at p < 0:05.

10
Oxidative Medicine and Cellular Longevity played a crucial role in α-glucosidase inhibition. Compound 6 (IC 50 = 313:03 ± 3:71 μM) showed low α-glucosidase inhibitory activity. Compounds 5, 9, and 10 were reported no inhibitory activity. These results indicated that the substitution of methoxyl, ethyoxyl, and glyceryl at the hydroxyl group on the carboxylic acid of gallic acid could reduce the inhibition of α-glucosidase. Meanwhile, compound 6 exhibited higher α-glucosidase inhibitory potential than 7 (IC 50 = 425:12 ± 31:50 μM). From this study, it indicated that increasing the number of hydroxyl group on the aromatic ring could enhance the inhibitory activity of α-glucosidase.
3.6. α-Glucosidase Inhibition Kinetics of Compounds 1 and 2. Compounds 1 and 2 both had strong inhibitory activities on α-glycosidase, especially 2. Therefore, it was necessary to investigate the inhibition types and calculate the inhibition kinetics constants (K i ). Results of enzyme kinetics study were performed by using Dixon plots and Lineweaver-Burk plots.
The results were displayed in Table 3 and Figure 2. In Figure 2, K m values of α-glucosidase in the existence of 1 did not change but V max values of α-glucosidase with the compounds progressively reduced, pointing that 1 was in a noncompetitive inhibitor manner with the K i value of 732:95 ± 0:22 nM. In addition, 2 restrained α-glucosidase in competitive manner as the V max maintained constant and K m improved. Meanwhile, K i value of this compound was calculated as 98:37 ± 0:31 nM. Competitive inhibitors were known to compete with substrates for the active sites of enzymes, whereas noncompetitive inhibitors primarily bind to allosteric sites of enzyme-substrate complexes [40].
3.7. Molecular Docking Studies. In order to understand how isolated compounds 1 and 2 conjugate with α-glucosidase, molecular docking was implemented. Because of the existence of the galloyl moiety, compound 2 reached deeper the active pocket than acarbose (Figure 3). In the interaction of compound 2 with α-glucosidase, 2 was encompassed by the residues VAL-232, PHE-303, PHE-178, VAL-216, GLN279, LYS-156, LEU-313, ASP-307, PHE-321, VAL-308, and ALA-329. In addition, some active hydroxyl groups were involved in the formation of hydrogen bonds, such as the galloyl at the 3-OH of sugar moiety at the C-1 and C-6 with side 11 Oxidative Medicine and Cellular Longevity chains of PRO-312 and SER-304, respectively, the galloyl at the 3-O on sugar moiety at the C-3 position and the galloyl at the 4-O on sugar moiety at the C-6 position with side chains of SER-240 and GLY-309, and the carbonyl group on the galloyl on sugar moiety at the C-2 and oxygen atom on the sugar moiety with HIS-280 and THR-310, respectively. Meanwhile, π-π T-shaped, π-alkyl, and electrostatic interactions facilitated the tight bonding between the enzyme and compound 2. By contrast, acarbose could only form seven hydrogen bonds.
To investigate the allosteric site of the α-glucosidase bond with noncompetitive inhibitors (1), we implemented blind docking by establishing a grid involved in all enzymes. Compound 1 was surrounded by a number of catalytic amino acid residues (Figure 3), including PRO-467, PHE-469, GLU-408, MET-70, and ASP-68. Eight hydrogen bonds were present between the galloyl at the 1, 2, 3-OH and 1-O on sugar moiety at the C-1 position, the galloyl at the 1-OH on sugar moiety at the C-2 position, the galloyl at the 1-O on sugar moiety at the C-3 position, the galloyl at the 3, 4-OH on sugar moiety  67, PRO-66, ARG-413, LYS-406, LYS-466, THR-83, TRP-81, and SER-65, respectively. The aromatic rings of the galloyl on sugar moiety at the C-4 and C-2 positions of compound 1 formed π-π T-shaped and π-π stacked interactions with TRP-36 and TYR-470, respectively. In recent years, molecular docking simulations have become an important tool for understanding the interaction mode and the structure-activity relationships of ligands with receptors. The in silico research results of compounds 1 and 2 were in agreement with the IC 50 data and the enzyme kinetic study in vitro. The integration of enzyme activity, kinetics, and molecular docking studies provided principle insights into the molecular basis underlying ligand binding affinity and α-glucosidase inhibition.

Conclusions
In summary, antioxidant and α-glucosidase inhibitory activities of the extract, fractions, and isolated compounds from mango seed kernel were first documented. The petroleum ether, dichloromethane, ethyl acetate, and n-butanol fractions exhibited different degrees of antioxidant and α-glucosidase inhibitory activity. In order to further clarify the chemical basis, a total of 23 ingredients were identified by efficient and fast GC-MS analysis from the petroleum ether fraction. As far as we know, it was the first report on the detailed compounds of the petroleum ether fraction of mango seed kernel. Among all the identified components, (Z)-3-(Heptadec-10-en-1-yl) phenol, eicosane, 9-octadecenoic acid methyl ester, and 9-hexadecenoic acid methyl ester had never been reported before. In addition, 17 compounds were isolated and identified from other active fractions of mango seed kernel. 1, 2, 3, 4, 6-penta-O-galloyl-α-D-glucoside (2), 1-glycerol gallate (9), gallocatechin (15), and hexyl-β-D-glucoside (17) were isolated from Mangifera for the first time. Among isolated components, 1, 2, 3, 4, 6penta-O-galloyl-β-D-glucoside (1) and 1, 2, 3, 4, 6-penta-O-galloyl-α-D-glucoside (2) showed stronger DPPH radical scavenging and α-glucosidase inhibitory capacities than others. The structure-activity relationship has been discussed. According to the kinetic research, 2 inhibited α-glucosidase in a competitive manner, whereas 1 acted in a noncompetitive manner. Molecular docking studies showed that compound 1 accommodated well in the allosteric site of the α-glucosidase interacting with a number of crucial amino residues. Compound 2 reached deeper the active site pocket of α-glucosidase than acarbose. The stability of binding was enhanced by π-π T-shaped, π-alkyl, hydrogen bond, and electrostatic interactions. These results might be useful to investigate the pharmacological activity in this fruit. Meanwhile, our study provided a chemical basis for the further development and utilization of mango seed kernel in the food and medicine fields.

Data Availability
The underlying data supporting our findings can be found and generated during the study.