Inhibition of Butyrylcholinesterase and Human Monoamine Oxidase-B by the Coumarin Glycyrol and Liquiritigenin Isolated from Glycyrrhiza uralensis

Eight compounds were isolated from the roots of Glycyrrhiza uralensis and tested for cholinesterase (ChE) and monoamine oxidase (MAO) inhibitory activities. The coumarin glycyrol (GC) effectively inhibited butyrylcholinesterase (BChE) and acetylcholinesterase (AChE) with IC50 values of 7.22 and 14.77 µM, respectively, and also moderately inhibited MAO-B (29.48 µM). Six of the other seven compounds only weakly inhibited AChE and BChE, whereas liquiritin apioside moderately inhibited AChE (IC50 = 36.68 µM). Liquiritigenin (LG) potently inhibited MAO-B (IC50 = 0.098 µM) and MAO-A (IC50 = 0.27 µM), and liquiritin, a glycoside of LG, weakly inhibited MAO-B (>40 µM). GC was a reversible, noncompetitive inhibitor of BChE with a Ki value of 4.47 µM, and LG was a reversible competitive inhibitor of MAO-B with a Ki value of 0.024 µM. Docking simulations showed that the binding affinity of GC for BChE (−7.8 kcal/mol) was greater than its affinity for AChE (−7.1 kcal/mol), and suggested that GC interacted with BChE at Thr284 and Val288 by hydrogen bonds (distances: 2.42 and 1.92 Å, respectively) beyond the ligand binding site of BChE, but that GC did not form hydrogen bond with AChE. The binding affinity of LG for MAO-B (−8.8 kcal/mol) was greater than its affinity for MAO-A (−7.9 kcal/mol). These findings suggest GC and LG should be considered promising compounds for the treatment of Alzheimer’s disease with multi-targeting activities.


Introduction
Acetylcholinesterase (AChE, EC 3.1.1.7) catalyzes the breakdown of acetylcholine (ACh), a neurotransmitter found in synapses of the cerebral cortex [1]. AChE inhibitors reduce AChE activity and maintain or increase ACh levels, which are typically deficient in Alzheimer's disease (AD), and thus, enhance cholinergic transmission in brain [2,3]. AD is a chronic, devastating manifestation of neuronal dysfunction and is characterized by progressive mental failure, disordered cognitive functions, and speech impairment. Various cholinesterase inhibitors (e.g., tacrine, donepezil, galantamine, and rivastigmine), immunotherapies, antisense oligonucleotides, phyto-pharmaceuticals, and nutraceuticals are being used to treat AD [4].

Analysis of Inhibitory Activities
The AChE, BChE, MAO-A, and MAO-B inhibitory activities of the eight compounds were investigated at a concentration of 10 µM. Most of the eight inhibited AChE, BChE, MAO-A, and MAO-B by less than 50%. However, GC and LG potently inhibited AChE and BChE, and MAO-A and MAO-B, respectively (Table 1). GC inhibited BChE and AChE with IC 50 values of 7.22 and 14.77 µM, respectively, with a selectivity index (SI) of 2.0 for BChE with respect to AChE, and also moderately inhibited MAO-B (29.48 µM). Other compounds showed weak inhibitory activities against AChE or BChE, except liquiritin apioside, which moderately inhibited AChE (IC 50 = 36.68 µM).
LG potently inhibited MAO-B (IC 50 = 0.098 µM) and MAO-A (IC 50 = 0.27 µM). The SI value of LG for MAO-B with respect to MAO-A was 2.8 (Table 1). Liquiritin, a LG glycoside, weakly inhibited MAO-A and MAO-B (>40 µM). Thus, GC and LG were found to be effective inhibitors of BChE and MAO-B, respectively.

Analysis of the Reversibilities of BChE and MAO-B Inhibitions
Reversibilities of BChE inhibition by GC and of MAO-B inhibition by LG were investigated by dialysis and dilution methods. Residual BChE activity after GC inhibition recovered partially from 34.6% (undialyzed activity; AU) to 58.4% (dialyzed activity; AD) by dialysis, whereas inhibition by tacrine (a known reversible inhibitor) significantly recovered from 10.3% to 74.1% (Figure 2A). We also confirmed reversibility using the dilution method by measuring and comparing residual BChE activities of a sample preincubated with GC at a concentration of 50 × IC50 and then diluted to a concentration of 1 × IC50 with a control sample treated at a GC concentration of 1 × IC50. We found that residual activities were similar before and after dilution (51.1% and 40.1%, respectively), and that the activity of the sample at a concentration of 50 × IC50 was 10.3% ( Figure 2A). These results suggested that GC is a reversible inhibitor of BChE, because if it acted as an irreversible inhibitor, activity would have been reduced by dilution. On the other hand, the relative residual activity of MAO-B after LG inhibition recovered from 38.4% (AU) to 87.2% (AD) by dialysis, which was similar to activity recovery observed for the reversible MAO-B inhibitor lazabemide (from 36.1% to 88.0%). On the other hand, values for the irreversible inhibitor pargyline were 17.0% and 8.4%, respectively ( Figure 2B). These results showed that GC and LG reversibly inhibited BChE and MAO-B, respectively.

Analysis of Inhibitory Patterns
Modes of BChE inhibition by GC and of MAO-B inhibition by LG were investigated by analyzing Lineweaver-Burk plots. Plots of BChE inhibition by GC were linear and intersected the x-axis ( Figure  3A). Secondary plots of the slopes of Lineweaver-Burk plots against inhibitor concentration showed the Ki value of GC for BChE inhibition was 4.47 ± 0.29 µM ( Figure 3B).

Analysis of the Reversibilities of BChE and MAO-B Inhibitions
Reversibilities of BChE inhibition by GC and of MAO-B inhibition by LG were investigated by dialysis and dilution methods. Residual BChE activity after GC inhibition recovered partially from 34.6% (undialyzed activity; A U ) to 58.4% (dialyzed activity; A D ) by dialysis, whereas inhibition by tacrine (a known reversible inhibitor) significantly recovered from 10.3% to 74.1% (Figure 2A). We also confirmed reversibility using the dilution method by measuring and comparing residual BChE activities of a sample preincubated with GC at a concentration of 50 × IC 50 and then diluted to a concentration of 1 × IC 50 with a control sample treated at a GC concentration of 1 × IC 50 . We found that residual activities were similar before and after dilution (51.1% and 40.1%, respectively), and that the activity of the sample at a concentration of 50 × IC 50 was 10.3% ( Figure 2A). These results suggested that GC is a reversible inhibitor of BChE, because if it acted as an irreversible inhibitor, activity would have been reduced by dilution. On the other hand, the relative residual activity of MAO-B after LG inhibition recovered from 38.4% (A U ) to 87.2% (A D ) by dialysis, which was similar to activity recovery observed for the reversible MAO-B inhibitor lazabemide (from 36.1% to 88.0%). On the other hand, values for the irreversible inhibitor pargyline were 17.0% and 8.4%, respectively ( Figure 2B). These results showed that GC and LG reversibly inhibited BChE and MAO-B, respectively.

Analysis of Inhibitory Patterns
Modes of BChE inhibition by GC and of MAO-B inhibition by LG were investigated by analyzing Lineweaver-Burk plots. Plots of BChE inhibition by GC were linear and intersected the x-axis ( Figure 3A). Secondary plots of the slopes of Lineweaver-Burk plots against inhibitor concentration showed the K i value of GC for BChE inhibition was 4.47 ± 0.29 µM ( Figure 3B).  These results indicate GC acted as a noncompetitive inhibitor of BChE and bound to a site other than the understood substrate binding site of BChE. On the other hand, plots of MAO-B inhibition by LG were linear and intersected the y-axis ( Figure 3C) and secondary plots showed the Ki value of LG for MAO-B inhibition was 0.023 ± 0.00061 µM ( Figure 3D), indicating LG is a competitive inhibitor of MAO-B.

Molecular Docking Simulation
Docking simulations showed that GC located at the binding site of 3-[(1S)-1-(dimethylamino)ethyl]phenol (SAF) in AChE (PDB: 1GQS) and the binding site of N- The binding affinity (−7.8 kcal/mol) of GC for BChE was greater than its affinity for AChE (−7.1 kcal/mol) as determined by AutoDock Vina (Table 2), and these binding affinity values concurred with the IC50 values (Table 1). Docking simulation results suggested that GC did not form a hydrogen bond with AChE ( Figure 4A), but that GC forms two hydrogen bonds  These results indicate GC acted as a noncompetitive inhibitor of BChE and bound to a site other than the understood substrate binding site of BChE. On the other hand, plots of MAO-B inhibition by LG were linear and intersected the y-axis ( Figure 3C) and secondary plots showed the K i value of LG for MAO-B inhibition was 0.023 ± 0.00061 µM ( Figure 3D), indicating LG is a competitive inhibitor of MAO-B.

Molecular Docking Simulation
Docking simulations showed that GC located at the binding site of 3-[(1S)-1-(dimethylamino)ethyl] phenol (SAF) in AChE (PDB: 1GQS) and the binding site of The binding affinity (−7.8 kcal/mol) of GC for BChE was greater than its affinity for AChE (−7.1 kcal/mol) as determined by AutoDock Vina (Table 2), and these binding affinity values concurred with the IC 50 values (Table 1). Docking simulation results suggested that GC did not form a hydrogen bond with AChE ( Figure 4A), but that GC forms two hydrogen bonds with the Thr284 and Val288 residues of BChE (distances: 2.42 and 1.92 Å, respectively) ( Figure 4B). These results explain the preference of GC for BChE. with the Thr284 and Val288 residues of BChE (distances: 2.42 and 1.92 Å, respectively) ( Figure 4B). These results explain the preference of GC for BChE.
LG and liquiritin located at the binding site of 7-methoxy-1-methyl-9H-beta-carboline complexed with MAO-A (PDB: 2Z5X) and of pioglitazone complexed with MAO-B (PDB: 4A79). The binding affinities of LG and liquiritin with MAO-B were greater than their binding affinities with MAO-A (Table 2), and LG binding affinities were in-line with the IC50 values shown in Table 1    LG and liquiritin located at the binding site of 7-methoxy-1-methyl-9H-beta-carboline complexed with MAO-A (PDB: 2Z5X) and of pioglitazone complexed with MAO-B (PDB: 4A79). The binding affinities of LG and liquiritin with MAO-B were greater than their binding affinities with MAO-A (Table 2), and LG binding affinities were in-line with the IC 50 values shown in Table 1. However, docking simulations did not predict hydrogen bond formation between LG or liquiritin with MAO-A or MAO-B ( Figure 4C-F).

Discussion
In the present study, GC (a coumarin) was isolated from G. uralensis and its BChE inhibitory activity was evaluated. Coumarins are characterized by the presence of 1,2-benzopyrone or benzopyran-2-one groups, which are the most common oxygen-containing heterocyclic compounds found in Nature. The ChE inhibitory activities by coumarins have been previously reviewed for synthetic and natural compounds [38]. Most of the known coumarins have a lower IC 50 value for AChE than for BChE, and selectivity for AChE or BChE is dependent on scaffold substituents, as exemplified by 3-(4-aminophenyl)-coumarin derivatives [39]. Furthermore, potencies of natural coumarins for AChE or BChE are much weaker than those of synthetic analogues. Nevertheless, natural coumarins exhibit significant inhibitory activities against AChE, examples include xanthotoxin from Ferula lutea (IC 50 = 0.76 µM) [40] and a 4-phenylcoumarin mesuagenin B from Mesua elegans (IC 50 = 0.70 µM) [41]. Based on the classification of natural AChE inhibitors, those with IC 50 values ≤ 15 µM are termed high potency inhibitors and those with values ranging from 15 to 50 µM moderate potency inhibitors [42]. According to this classification, GC is a high potency AChE inhibitor (IC 50 = 14.77 µM), though the value is near the threshold. In a previous study, osthenol, a prenylated coumarin obtained from Angelica pubescens, selectively inhibited MAO-A, and exhibited moderate AChE inhibitory activity (IC 50 = 25.3 µM) [43].
Natural coumarins have been reported to have low BChE inhibitory activities; sphondin and pimpinellin from Heracleum platytaenium inhibited BChE by 63.69% and 78.02%, respectively, at a concentration of 25 µg/mL concentration (115.7 and 101.5 µM, respectively) [44], and notably, all these IC 50 values are higher than that of GC (IC50 = 7.22 µM) as determined in the present study.
Dual inhibitions of ChE and MAO-B have been investigated in the context of AD [11,16]. In the present study, GC potently inhibited BChE with an IC 50 value of 7.22 µM, and moderately inhibited AChE and MAO-B, indicating GC should be considered as a multi-function inhibitor of BChE, AChE, and MAO-B.
Pan et al. reported that MAO-B inhibition by LG in rat liver mitochondria was weaker than MAO-A inhibition by a mixed type [32]. However, in our study, LG more potently inhibited human MAO-B (IC 50 = 0.098 µM) than human MAO-A (IC 50 = 0.27 µM) and functioned as a competitive inhibitor. The IC 50 of LG for MAO-B was lower than that of the flavonoid acacetin (IC 50 = 0.17 µM) [53], which is one of the lowest IC 50 values reported for a natural compound to date. Liquiritin was less effective than LG, aglycone of liquiritin, likely observed in acacetin and acacetin 7-O-(6-O-malonylglucoside) [53].
In our docking analysis, GC showed greater binding affinity with BChE than with AChE, and LG and liquiritin were predicted to bind to MAO-B more strongly than to MAO-A, and these results agreed well with determined IC 50 values. In particular, our kinetic study showed that GC noncompetitively inhibited BChE. Docking simulation was performed to identify BChE binding sites. The docked pose for GC indicated that it interacted with BChE beyond the active site and hydrogen bonded with Thr284 and Val288. The active-site of BChE is composed of 4 subdomains, i.e., a peripheral site, a choline binding pocket, a catalytic site, and an acyl binding pocket [54], and the acyl binding pocket contains Trp231, Leu286, and Val288, which permit binding and hydrolysis of ligands and substrates bulkier than those of AChE [54], which is considered to be largely responsible for the different ligand-binding specificities of AChE and BChE [55]. Jannat et al. reported that (2S,3R)-pretosin C is a noncompetitive inhibitor of BChE and that it hydrophobically interacts with Val288, Lue286, and Phe357, and hydrogen bonds with Gly283 and Asn397, and docks at a non-ligand binding site [56]. It was also observed that hydrogen bond formation was the main driving force behind BChE-coumarin complex formation, whereas hydrophobic and halogen interactions underpinned AChE interactions with N1-(coumarin-7-yl) derivatives [57]. Similarly, we found that GC hydrogen bonded with Thr284 and Val288 located outside the ligand binding site. Such results suggest that GC might bind noncompetitively at the acyl binding pocket of BChE.
MAO-A and MAO-B activities were measured continuously at 316 nm for 20 min, and at 250 nm for 30 min, respectively, as described previously [60,61]. The concentrations used were; kynuramine (0.06 mM) for MAO-A and benzylamine (0.3 mM) for MAO-B. AChE activity was assayed continuously for 10 min at 412 nm using 0.2 U/mL of enzyme in the presence of 0.5 mM DTNB and 0.5 mM ATCI in 0.5 mL of reaction mixture, as previously described [49,58], based on the method developed by Ellman et al. [62]. BChE activity was assayed using the same method as AChE, except using BTCI [49]. Substrate concentrations of BTCI for BChE and benzylamine for MAO-B were 2.3-and 2.1-fold of the respective K m values (0.22 and 0.14 mM).

Inhibitory Activities and Enzyme kinetics
Inhibitions of MAO-A, MAO-B, AChE, and BChE were initially observed at an inhibitor concentration of 10 µM. IC 50 values of compounds exhibiting >50% inhibition were determined. Kinetic parameters, inhibition types, and K i values were determined for the most potent inhibitors, i.e., GC for BChE and LG for MAO-B, as previously described [49,58]. The kinetics of BChE and MAO-B inhibitions were investigated at five different substrate concentrations; 0.05, 0.1, 0.25, 0.5 or 1.0 mM for BChE, and 0.03, 0.06, 0.15, 0.3, or 0.6 mM for MAO-B. Inhibition studies were conducted in the absence or presence of each inhibitor at about 0.5×, 1.0×, and 2.0× their IC 50 values [58]. Inhibitory patterns and K i values were determined using Lineweaver-Burk plots and secondary derivative plots.

Analysis of Inhibitor Reversibility
The reversibilities of BChE inhibition by GC and of MAO-B inhibition by LG were investigated by dialysis at concentrations of 2 × IC 50 values, as previously described [63]. After preincubating GC or LG with BChE or MAO-B, respectively, for 30 min, residual activities for undialyzed and 6 h-dialyzed samples were measured; relative values for A U and A D were then calculated and compared with each control without inhibitor. Reversibilities were determined by comparing A U and A D values of inhibitors with those of references. In addition, the dilution method was used to access BChE activity recovery after inhibition by GC (i.e., after preincubating BChE with GC at 50 × IC 50 for 15 min) and diluting to a GC concentration of 1 × IC 50 [60]. Residual activity of the preincubated and then diluted mixture was measured and compared to those of mixtures at 1× or 50 × IC 50 concentrations.
LG potently inhibited MAO-B (IC 50 = 0.098 µM) and MAO-A (IC 50 = 0.27 µM). GC was found to be a noncompetitive inhibitor of BChE and LG to be a competitive inhibitor of MAO-B. The binding affinity of GC for BChE (−7.8 kcal/mol) was higher than its affinity for AChE (−7.1 kcal/mol), and this binding was driven by hydrogen bond formation with Thr284 and Val288 of BChE. These findings regarding the multi-inhibitory effects of GC and LG suggest that they be considered potential candidates for the treatment of Alzheimer's disease.