Soluble Epoxide Hydrolase Inhibitory Activity of Components Isolated from Apios americana Medik

A new compound 1, 5-methoxy-2,5,7,4′-tetrahydroxy-coumaronochromone, along with seven known compounds (2–8), were isolated from Apios americana using open column chromatography. Their structures were established based on an analysis of 1D and 2D NMR, and MS spectra. Among these, two compounds 1 and 2 showed inhibitory activity on soluble epoxide hydrolase (sEH) at a concentration below 50 μM. The respective competitive (1) and mixed (2) inhibitors were revealed to have Ki values of 21.0 ± 0.8 and 14.5 ± 1.5 μM, based on the Dixon plot. The potential inhibitor (2) was visually presented in a predicted binding pose in the receptor by molecular docking. Additionally, molecular dynamics were performed for a detailed understanding of their complex by Gromacs 4.6.5 package.

The objective of our study is to isolate and identify isoflavonoids from A. americana, to evaluate the inhibitory activity of new-type compounds on sEH in vitro, and to provide the information of the interaction between receptor and ligand.

Enzyme Kinetics
To determine the nature of the binding mechanism between the inhibitors (1 and 2) and the enzyme, two inhibitors were diluted as concentrations of 0, 25, and 50 μM in methanol, and the respective samples were then evaluated for their inhibitory activity at various substrate concentrations ranging from 3.1 to 25 μM. The enzyme kinetics of compounds 1 and 2 were determined by generating double-reciprocal plots (Lineweaver-Burk plot and Dixon plot). As shown in Figure 3C,D, Lineweaver-Burk plots of 1

The Inhibitory Activity on sEH
7.6 ± 2.5 nM a All compounds were tested in a set of triplicated experiment; b Positive control; c Not Tested.

Enzyme Kinetics
To determine the nature of the binding mechanism between the inhibitors (1 and 2) and the enzyme, two inhibitors were diluted as concentrations of 0, 25, and 50 µM in methanol, and the respective samples were then evaluated for their inhibitory activity at various substrate concentrations ranging from 3.1 to 25 µM. The enzyme kinetics of compounds 1 and 2 were determined by generating double-reciprocal plots (Lineweaver-Burk plot and Dixon plot). As shown in Figure 3C,D, Lineweaver-Burk plots of 1 yielded a family of straight lines with different slopes through a common point on the Y-axis; therefore, 1 was designated as a competitive inhibitor with preferential binding toward activity site of receptor. On the other hand, a family of straight lines of 2 cross over one point in the second quadrant, and hence 2 was confirmed to interact with the receptor as a mixed inhibitor. Additionally, the inhibition constant (K i ) values of isoflavonoids 1 and 2 were calculated by using Dixon plots ( Figure 3E,F). The K i values for the inhibition of sEH were 21.0 ± 0.8 and 14.5 ± 1.5 µM, respectively.

Molecular Docking
The study for visualizing the complex between receptor (code ID: 3ANS) and ligand (2) was subjected to molecular simulation using the Autodock 4.2 program. According to the result of enzyme kinetics, a grid was set up to contain the full receptor for scholarly molecular docking. The result was represented graphically by LIGPLOT v.4.5.4 and Chimera 1.10rc. As indicated in Figure 4A and Table 3, ligand (2) was docked to form a stable pose, with −8.02 kcal/mol of Autodock score, into the right pocket of the receptor, and was surrounded around twelve amino acids (Tyr466, Val498, Asp335, Trp525, His524, Arg410, Ser407, Val416, Ser415, Leu417, Leu408, and Met419). Among these, Asp335, Ser407, Val416, Tyr466, and Trp525 were hydrogen bonds at distances of 3.21, 2.72, 3.11, 3.34, and 3.05 Å from the inhibitor (2) ( Figure 4B). Molecular docking helped keep track of the allosteric site where compound 2 may be combined. Our result showed that the complex calculated with the lowest Autodock score was the best pose, and the potential inhibitor (2) was found to make five hydrogen bonds with Asp335, Ser407, Val416, Tyr466, and Trp525 in the right pocket next to the active site. This study showed similar results with previous reports. The right pocket proved to be the position of sEH preferring the interaction with compounds, such as flavonoids, stilbene, and anthraquinone derivatives again [8,16]. yielded a family of straight lines with different slopes through a common point on the Y-axis; therefore, 1 was designated as a competitive inhibitor with preferential binding toward activity site of receptor. On the other hand, a family of straight lines of 2 cross over one point in the second quadrant, and hence 2 was confirmed to interact with the receptor as a mixed inhibitor. Additionally, the inhibition constant (Ki) values of isoflavonoids 1 and 2 were calculated by using Dixon plots ( Figure 3E,F). The Ki values for the inhibition of sEH were 21.0 ± 0.8 and 14.5 ± 1.5 μM, respectively.

Molecular Docking
The study for visualizing the complex between receptor (code ID: 3ANS) and ligand (2) was subjected to molecular simulation using the Autodock 4.2 program. According to the result of enzyme kinetics, a grid was set up to contain the full receptor for scholarly molecular docking. The result was represented graphically by LIGPLOT v.4.5.4 and Chimera 1.10rc. As indicated in Figure 4A and Table 3, ligand (2) was docked to form a stable pose, with −8.02 kcal/mol of Autodock score, into the right pocket of the receptor, and was surrounded around twelve amino acids (Tyr466, Val498, Asp335, Trp525, His524, Arg410, Ser407, Val416, Ser415, Leu417, Leu408, and Met419). Among these, Asp335, Ser407, Val416, Tyr466, and Trp525 were hydrogen bonds at distances of 3.21, 2.72, 3.11, 3.34, and 3.05 Å from the inhibitor (2) ( Figure 4B). Molecular docking helped keep track of the allosteric site where compound 2 may be combined. Our result showed that the complex calculated with the lowest Autodock score was the best pose, and the potential inhibitor (2) was found to make five hydrogen bonds with Asp335, Ser407, Val416, Tyr466, and Trp525 in the right pocket next to the active site. This study showed similar results with previous reports. The right pocket proved to be the position of sEH preferring the interaction with compounds, such as flavonoids, stilbene, and anthraquinone derivatives again [8,16].

Molecular Simulation
Molecular dynamics simulation was performed to investigate the corresponding interaction of the receptor with the ligand, using the Gromacs 4.6.5 package. The key results are presented in Figure 5. The complexes were superposed at snap shot in 1 ns intervals during a 10 ns simulation cycle. Figure 5A shows the corresponding flexibility according to the loop movement at allosteric site of the receptor.

Molecular Simulation
Molecular dynamics simulation was performed to investigate the corresponding interaction of the receptor with the ligand, using the Gromacs 4.6.5 package. The key results are presented in Figure 5. The complexes were superposed at snap shot in 1 ns intervals during a 10 ns simulation cycle. Figure 5A shows the corresponding flexibility according to the loop movement at allosteric site of the receptor. This simulation was made and the final value of stable potential energy was about −6.2 × 10 5 kJ/mol, with RMSD under 0.25 nm, and RMSF of 0.32 nm ( Figure 5B-D). These results proved that molecular dynamics of the complex was perfectly achieved. Additionally, this study observed the hydrogen bonds to have important roles at the interaction of the receptor and the ligand. During the simulation, they were mainly built of 1-3 hydrogen bonds, and sometimes formed 0, 4, and 5 hydrogen bonds ( Figure 5E). Finally, the distances of the key amino acids in molecular docking with ligand (2) were calculated during the 10 ns trajectory. As shown in Figure 5F,H, Asp335, Ser407, and Tyr466 were far away, at a distance of over 4.5 Å from the ligand (2) in the course of the energy minimization. Also, during molecular simulation, they were separated from the ligand at over 10 Å distance. Conversely, Val416 and Trp525 maintained a distance of 3.5 Å up to 1.8 ns into molecular dynamics. Interestingly, after 2 ns, Val416 was maintained within 3.5 Å distance of the ketone of the ligand ( Figure 5G). Molecular dynamics revealed that the potential inhibitor (2) maintained a constant 3.5 Å distance fromVal416 of the receptor during the 10 ns simulation, and with Trp525 at about 2 ns. Resvertol, desoxyrhapontigenin, and rhapontigenin of stilbene are composed of hydrogen bonds with the Val416 residue in the right pocket [8]. Our findings were analyzed with compound 2 dependently moved according to the mobility of the loop containing Val416. These results suggest that Val416 acts as the key amino acid while 2 is interacting with the allosteric site of the receptor. Finally, compound 2 isolated from A. americana was observed to block the catalytic reaction of sEH, which could be targeted for the treatment of cardiovascular disease. Two hydroxyl and ketone groups of compound 2 may be closely related to Val416 of the receptor.

Molecular Simulation
Molecular dynamics simulation was performed to investigate the corresponding interaction of the receptor with the ligand, using the Gromacs 4.6.5 package. The key results are presented in Figure  5. The complexes were superposed at snap shot in 1 ns intervals during a 10 ns simulation cycle. Figure 5A shows the corresponding flexibility according to the loop movement at allosteric site of the receptor. This simulation was made and the final value of stable potential energy was about −6.2 × 10 5 kJ/mol, with RMSD under 0.25 nm, and RMSF of 0.32 nm ( Figure 5B-D). These results proved that molecular dynamics of the complex was perfectly achieved. Additionally, this study observed the hydrogen bonds to have important roles at the interaction of the receptor and the ligand. During the simulation, they were mainly built of 1-3 hydrogen bonds, and sometimes formed 0, 4, and 5 hydrogen bonds ( Figure 5E). Finally, the distances of the key amino acids in molecular docking with ligand (2) were calculated during the 10 ns trajectory. As shown in Figure 5F,H, Asp335, Ser407, and Tyr466 were far away, at a distance of over 4.5 Å from the ligand (2) in the course of the energy minimization. Also, during molecular simulation, they were separated from the ligand at over 10 Å distance. Conversely, Val416 and Trp525 maintained a distance of 3.5 Å up to 1.8 ns into molecular dynamics. Interestingly, after 2 ns, Val416 was maintained within 3.5 Å distance of the ketone of the ligand ( Figure 5G). Molecular dynamics revealed that the potential inhibitor (2) maintained a constant 3.5 Å distance fromVal416 of the receptor during the 10 ns simulation, and with Trp525 at about 2 ns. Resvertol, desoxyrhapontigenin, and rhapontigenin of stilbene are composed of hydrogen bonds with the Val416 residue in the right pocket [8]. Our findings were analyzed with compound 2 dependently moved according to the mobility of the loop containing Val416. These results suggest that Val416 acts as the key amino acid while 2 is interacting with the allosteric site of the receptor. Finally, compound 2 isolated from A. americana was observed to block the catalytic reaction of sEH, which could be targeted for the treatment of cardiovascular disease. Two hydroxyl and ketone groups of compound 2 may be closely related to Val416 of the receptor.

Plant Material
In October 2016, the roots of A. americana (voucher specimen RBRC 001) were cultivated and collected locally from Jeoungeup, Jeollabuk-do, Korea. The species was identified by Dr. S.Y. Kang of Radiation Breeding Research Center, Korea Atomic Energy Research Institute.

Plant Material
In October 2016, the roots of A. americana (voucher specimen RBRC 001) were cultivated and collected locally from Jeoungeup, Jeollabuk-do, Korea. The species was identified by Dr. S.Y. Kang of Radiation Breeding Research Center, Korea Atomic Energy Research Institute.

sEH Assay
The sEH assay was performed as described previously [8]. Briefly, 130 µL of sEH in 25.0 mM Bis-Tris-HCl buffer (pH 7.0) and 20.0 µL of the compounds (1-0.06 mM concentration) diluted in MeOH, were added in 96-well plate, to which 50.0 µL of 20.0 µM PHOME was added in the mixture. After initiating the enzyme reaction at 37 • C, the products by hydrolysis of the substrate were monitored at excitation and emission of 330 and 465 nm for one hour.
where C 40 and S 40 were the fluorescence of the control and inhibitor, respectively, after 40 min, S 0 and C 0 is the fluorescence of inhibitor and control, respectively, at 0 min. AUDA was used as the positive control.

Molecular Docking
The molecular docking simulation was performed as previously described [17]. The 3D structure of the ligand was built and minimized by MM2 using Chem 3D Pro. (Ver. 14.0). The flexible bonds of the ligand were assigned with AutoDockTools. The 3D structure of sEH (PDB ID: 3ANS) was obtained from RCSB (protein data bank) after substrates were removed from the original enzyme. All hydrogen atoms and gasteiger charges were assigned using AutoDockTools. Briefly, compound 2 was subjected to the simulation of the grid containing full enzyme (X: 126, Y: 126, and Z: 126) at 0.375 Å. This docking study was simulated using the Lamarckian Genetic Algorithm with Runs 50 and the long maximum number.

Molecular Dynamics
Molecular dynamics simulation was performed as previously reported, with few modifications [18]. The topology files of the ligand and receptor were built as GROMOS 53A6 force-field by prodrg and Gromacs 4.6.5, respectively. The gro complex of them was placed in the center of a cubic box (7.8 7.8 7.8) solvated with water molecules containing six Cl − (1.0 Å distance). Energy minimization was performed to end when the minimization reached with the maximum force under 10.0 kJ/mol. Following this, the Particle Mesh Ewald (PME) method was used for the treatment of long-range electrostatic interactions, and the linear constraint solver (LINCS) algorithm was used for covalent bond constraints. Each NVT and NPT was performed for 0.1 ns to equilibrate the system with constant volume, temperature (300 K), and pressure (1 bar). The MD ran for 10 ns (10,000 ps).

Statistical Analysis
All assays in the presence of inhibitors were performed in triplicate. The results are presented as the means ± standard error of the mean. Sigma Plot (SPP Inc., Chicago, IL, USA) was used for analysis of results.

Conclusions
Using column chromatography, a new compound, 5-methoxy-2,5,7,4 -tetrahydroxycoumaronochromone (1); along with genestein (2); 3 -methoxy-4 ,5,7-trihydroxyisoflavone (3); gerontoisoflavone A (4); 5-methoxygenistein-7-O-glucoside (5); 2 -hydroxygenistein-7-O-glucoside (6); 2 -hydroxygenistein-7-O-gentibioside (7); and 4-hydroxybenzoic acid (8); were purified from A. americana. This study was undertaken to evaluate their inhibitory activity on sEH. Of these, compounds 1 and 2 had dose-dependent IC 50 values of less than 50 µM, and acted as competitive and mixed inhibitors, respectively. Based on the Dixon plot, the K i value of compound 2 was calculated at about 10 µM. Flavonoids, stilbene, and anthraquinone derivatives from medicinal plants have been revealed to be potential inhibitors of sEH [8,16]. Through this study, components from A. americana, which has been used as food source, were proved to be new inhibitors of the sEH enzyme which is targeted for treatment of cardiovascular disease. As indicated in Figure 3A, these aglycon derivatives were observed to have more inhibitory activity on sEH than that of glycosides. Furthermore, our study suggests for the first time the predicted pose of isoflavonoid binding to sEH by using in silico skills. The mixed-type inhibitor (2) was suitably fitted in the right pocket next to the sEH active site. This interaction was constantly held via the interaction of the inhibitor with residue Val416 of the enzyme. These results suggest that it is necessary to consider the hydrophobic interaction with Val416 when developing synthetic compounds based on the pharmacophore of the isoflavonoid type. Additionally, this right pocket as an allosteric site needs to be considered for the development of non-competitive or mixed-type inhibitors of sEH.
Finally, for development of a new type inhibitor of sEH, as an alternative to urea, this study suggests that the structure of compound 2 is eligible to be the backbone of the lead compound.