Evaluation of the Antioxidant, Antidiabetic, and Antiplasmodial Activities of Xanthones Isolated from Garcinia forbesii and Their In Silico Studies

This study aimed to isolate xanthones from Garcinia forbesii and evaluated their activity in vitro and in silico. The isolated compounds were evaluated for their antioxidant activity by DPPH, ABTS and FRAP methods. The antidiabetic activity was performed against α-glucosidase and α-amylase enzymes. The antiplasmodial activity was evaluated using Plasmodium falciparum strain 3D7 sensitive to chloroquine. Molecular docking analysis on the human lysosomal acid-alpha-glucosidase enzyme (5NN8) and P. falciparum lactate dehydrogenase enzyme (1CET) and prediction of ADMET for the active compound, were also studied. For the first time, lichexanthone (1), subelliptenone H (2), 12b-hydroxy-des-D-garcigerrin A (3), garciniaxanthone B (4) and garcigerin A (5) were isolated from the CH2Cl2 extract of the stem bark of G. forbesii. Four xanthones (Compounds 2–5) showed strong antioxidant activity. In vitro α-glucosidase test showed that Compounds 2 and 5 were more active than the others, while Compound 4 was the strongest against α-amylase enzymes. In vitro antiplasmodial evaluation revealed that Compounds 2 and 3 showed inhibitory activity on P. falciparum. Molecular docking studies confirmed in vitro activity. ADMET predictions suggested that Compounds 1–5 were potential candidates for oral drugs. The isolated 2–5 can be used as promising phytotherapy in antidiabetic and antiplasmodial treatment.


ABTS Radical Scavenging Assay
The antioxidant activity was also determined by free radical ABTS (2,2 -azinobis (3-ethyl benzothiazoline-6-sulfonic acid) according to the reported procedure [12]. In this test, 7 mM ABTS solution was prepared and mixed with a buffer solution of potassium peroxydisulfate, incubated for 16 h, and protected from light. A total of 10 µL of the sample and 1 mL of the ABTS was incubated for 4 min and measured at a wavelength of 734 nm. This experiment was performed triplo using quercetin and gallic acid as controls. The ABTS inhibitory activity was calculated by the equation [(Ab-As)/Ab] × 100%.

Rat Intestinal α-glucosidase Inhibitory Activity
The inhibitory activity of rat intestinal α-glucosidase was determined according to previously reported method [14] with slight modifications. This procedure was performed by classifying four mixtures in different groups. The mixture (1) labelled "Blank of enzyme reaction" contained the mixture and DMSO (10 µL), sodium phosphate buffer (50 µL), glucose kit (80 µL), enzyme (20 µL) and substrate (20 µL maltose, sucrose 20 µL). A 10 µL of sample and samples with various concentrations of compounds was added to the well Plates (3) and (4). A 50 µL of sodium phosphate buffer (pH 6.9) was then added to well Plates (1) and (3), and a 30 µL of the buffer was added to well Plates (2) and (4), respectively. The substrate (maltose, 20 µL; sucrose 20 µL) was then added to the Plates (2) and (4). A glucose kit (80 µL each) and the enzyme (20 µL) were added to all plates and were incubated for 10 min for maltose and 40 min for sucrose. Finally, the absorbance was then measured at λ = 520 nm using a microplate reader (BioTek ELx800TM, BioTek Instruments, Inc., Winooski, VT, USA).

α-Amylase Inhibitory Activity
The modified α-amylase enzyme inhibition test was conducted for antidiabetic activity evaluation [15]. A total of 10 mg sample was dissolved in 1 mL DMSO and 0.1 M phosphate buffer and 5 mg α-amylase solution (Porcine pancreatic α-amylase) in 1 mL phosphate buffer pH 6.9 was added. A 100 mg of potato starch (substrate) was heated in 5 mL of 0.1 M phosphate buffer for 5 min and cooled to room temperature. Then, 20 µL of the sample and 50 µL of substrate were mixed into 30 µL of phosphate buffer. The mixture was pre-incubated for 5 min followed by the addition of 20 µL of α-amylase enzyme and then incubation at 37 • C for 15 min. The color development was performed by adding 50 µL of 1M HCl and 50 µL of 10% iodine solution. Finally, the absorbance was measured at a wavelength of 650 nm.

In Vitro Antiplasmodial Assay
The antiplasmodial activity was carried out by the described procedure [16,17] using P. falciparum strain 3D7 sensitive to chloroquine (the LDH method). The samples were made with concentrations of 50, 10, 5, 1, 0.5, 0.1, 0.05, and 0.01 µg/mL. The parasites used in this test were synchronized (Ring stage) with ± 0.3% parasitemia (2% hematocrit). A total of 1 µL of the test solution with various concentrations was taken in to each well (well 96) and then added 99 µL with parasites (each concentration repeated three times). The well plates were put in the chamber with mix gas atmosphere (O 2 5%, CO 2 5% and N 2 90%) and incubated for 72 h at 37 • C. After that, the plate was harvested and stored at −30 • C. After 24 h, the pLDH assay was performed by reading the absorbance at 650 nm using a SpectraMax Paradigm ® Multi-Mode microplate reader.

Statistical Analysis
The study was conducted three times to determine the mean value (mean SD). A linear regression equation was generated for determination of the percentage of α-glucosidase and α-amylase inhibition concentration of each compound. The difference measured to be statistically significant when the p-value < 0.05.

Antioxidant Activity
The antioxidant activity depends on the existence of hydroxyl groups and substi ents on the aromatic ring. The antioxidant activity of phenolic compounds is well know to be related to the presence of ortho-and para-substituted hydroxyl groups on the a matic ring and carbonyl groups as well [25]. The antioxidant activity of isolated xanthon 1-5 was assayed by DPPH, ABTS and FRAP methods. DPPH is known for its stability a free radicals scavenging activity [26]. A radical species containing nitrogen atoms of AB stabilizes the free radicals through proton donors [27]. The ABTS assay can be applied lipophilic and hydrophilic compounds [28]. Furthermore, the iron reductive properties compounds which is an important part of antioxidant activity can be evaluated by usi FRAP method [29].
The free radical scavenging power based on DPPH test with quercetin and gallic a as standards indicated that the inhibitory concentration 50% (IC50) of compounds 2-5 w ranging from 14.1 to 20.4 μM at a concentration of 159.7 μg/mL (Table 3). Compound was inactive with inhibition below 50%. Previously reported garcinoxanthones SV, gar none E, and 1,3,6,7-tetrahydroxyxanthone from G. mangostana L. showed signific DPPH scavenging capacity with IC50 values of 68.55, 63.05, and 28.45 μM, respective compared to ascorbic acid (IC50 = 48.03 μM). The hydroxyl position of groups gave sign icant impact on the antioxidant activity of the latter compound [30]. The ABTS antioxida activity of compounds 2-5 was significant with IC50 values ranging from 0.05-7.9 μM (T ble 3). Compound 1 showed no significant activity with low inhibition at 99 μg/mL (<50% In particular, compound 3 was found to be the most potent with IC50 value 4-fold low than quercetin. Other compounds with xanthone skeleton such as α-mangostin and mangostin isolated from G. mangostana L. showed high activity in ABTS radical scaven ing [31]. In addition, the antioxidant evaluation of compounds 1-5 using FRAP test, shown in Table 3, demonstrated that the reduction power of compound 3 was found to the highest one with value of 203.9 ± 1.19 μM Fe 2+ /g. This activity was followed by co pounds 4, 5, and 2 with values of 192.7, 187.8, and 166.3 μM Fe 2+ /g, respectively. The ir reducing power of compound 3 was approximately 6-fold greater than ascorbic acid standard (30.6 μM Fe 2+ /g). These results implied that compounds 2-5 are potent antio

Antioxidant Activity
The antioxidant activity depends on the existence of hydroxyl groups and substituents on the aromatic ring. The antioxidant activity of phenolic compounds is well known to be related to the presence of orthoand para-substituted hydroxyl groups on the aromatic ring and carbonyl groups as well [25]. The antioxidant activity of isolated xanthones 1-5 was assayed by DPPH, ABTS and FRAP methods. DPPH is known for its stability and free radicals scavenging activity [26]. A radical species containing nitrogen atoms of ABTS stabilizes the free radicals through proton donors [27]. The ABTS assay can be applied to lipophilic and hydrophilic compounds [28]. Furthermore, the iron reductive properties of compounds which is an important part of antioxidant activity can be evaluated by using FRAP method [29].
The free radical scavenging power based on DPPH test with quercetin and gallic acid as standards indicated that the inhibitory concentration 50% (IC 50 ) of compounds 2-5 was ranging from 14.1 to 20.4 µM at a concentration of 159.7 µg/mL (Table 3). Compound 1 was inactive with inhibition below 50%. Previously reported garcinoxanthones SV, garcinone E, and 1,3,6,7-tetrahydroxyxanthone from G. mangostana L. showed significant DPPH scavenging capacity with IC 50 values of 68.55, 63.05, and 28.45 µM, respectively, compared to ascorbic acid (IC 50 = 48.03 µM). The hydroxyl position of groups gave significant impact on the antioxidant activity of the latter compound [30]. The ABTS antioxidant activity of compounds 2-5 was significant with IC 50 values ranging from 0.05-7.9 µM (Table 3). Compound 1 showed no significant activity with low inhibition at 99 µg/mL (<50%). In particular, compound 3 was found to be the most potent with IC 50 value 4-fold lower than quercetin. Other compounds with xanthone skeleton such as α-mangostin and γ-mangostin isolated from G. mangostana L. showed high activity in ABTS radical scavenging [31]. In addition, the antioxidant evaluation of compounds 1-5 using FRAP test, as shown in Table 3, demonstrated that the reduction power of compound 3 was found to be the highest one with value of 203.9 ± 1.19 µM Fe 2+ /g. This activity was followed by compounds 4, 5, and 2 with values of 192.7, 187.8, and 166.3 µM Fe 2+ /g, respectively. The iron reducing power of compound 3 was approximately 6-fold greater than ascorbic acid as standard (30.6 µM Fe 2+ /g). These results implied that compounds 2-5 are potent antioxidant compounds. The reducing power of other xanthones with carbonyl group and halogen substituents, i.e., (R)-6-chloro-2-{[(1-hydroxypropan-2-yl)(methyl)amino]-methyl}-9H-xanthen-9-one hy-drochloride and (R,S)-2-chloro-7-{[(1-hydroxybutan-2-yl)amino]methyl}-9H-xanthen-9-one synthesis results are reported better than vitamin C [32]. The antioxidant evaluation of isolated compounds 1-5 argued that the presence of dimethylallyl moiety and additional hydroxy groups on the structures of 2-5 was assumed to play important roles in enhancing the antioxidant capacity. Notably, the lack of pyran ring in 3 gave a potent ABTS scavenging activity with IC 50 value lower than the standards quercetin and gallic acid.

In Vitro Antidiabetic Assay
The antidiabetic potential of compounds 1-5 was evaluated on α-glucosidase and α-amylase enzymes. The α-glucosidase inhibitory activity of the five tested compounds was performed using sucrose and maltose as the substrates, and acarbose as the positive control [33]. The α-glucosidase is known to break the 1,4-α bond of carbohydrate into its monomers such as glucose and fructose [34]. The results of α-glucosidase assay, as shown in Table 4, showed that compound 5 was the most potent compound with IC 50 value of 37.4 µM followed by compounds 2, 4, and 3 using sucrose as substrate. In contrast, only compounds 2 and 5 inhibited the enzyme activity with IC 50 values lower than 100 µM using maltose as substrate. Compound 1 was inactive in both substrates. A new xanthone namely subelliptenone F had a significant effect on α-glucosidase with an IC 50 value of 4.1 µM compared to acarbose (IC 50 = 900.0 µM) [9]. α-Amylase is one of the key enzymes responsible for degrading starch to glucose in the human body [35,36]. The inhibitory activity of the tested compounds on α-amylase enzyme indicated that all compounds, except compound 1, were active with IC 50 values of 10.8-41.3 µM. The highest activity was shown by compound 4 with IC 50 value 10.8 µM. It is noted that compounds 2-5 are active as α-glucosidase and α-amylase inhibitors. This activity was supposed to be influenced by the dimethylallyl group and pyran rings in the xanthone structure. Compound 1 without dimethylallyl group and pyran rings was not active. The antidiabetic trend of compounds 1-5 was similar with the antioxidant inhibition pattern.
The docking results indicated that the total energy and MDS of compound 2 were 108.388 kcal/mol and −78.81 kcal/mol, and 144.003 kcal/mol and −74.54 kcal/mol for Compound 5 ( Table 6). The binding mode of these compounds, as shown in Table 6 and Figure 2, revealed that compounds 2 and 5 inhibit the acid-alpha-glucosidase through hydrogen bonding and steric interaction. These compounds are bound to the active site of human lysosomal acid-alpha-glucosidase enzyme (5NN8) through hydrogen-bonding interaction of carbonyl, chelate hydroxyl, and free hydroxyl groups with Arg600, Asp518, and Asp404 residues. The steric interaction of compound 2 was formed by the oxygen atom of a carbonyl group with Met519 residue, xanthone backbone and the pyran ring with Asp282(2) and Phe649 residues, and the dimethylallyl group with Trp376 residue. The similar interaction was also formed by compound 5 with Asp282 (2), Asp518, Asp616, Trp613 (2), Trp516, Trp376, Phe649, Met519, and His674 residues. Table 6. The total energy, MDS, and interacting residues of 2 and 5 to human lysosomal acid-alpha-glucosidase enzyme (5NN8).

Ligands
Total of carbonyl, chelate hydroxyl, and free hydroxyl groups with Arg600, Asp518, and Asp404 residues. The steric interaction of compound 2 was formed by the oxygen atom of a carbonyl group with Met519 residue, xanthone backbone and the pyran ring with Asp282(2) and Phe649 residues, and the dimethylallyl group with Trp376 residue. The similar interaction was also formed by compound 5 with Asp282 (2), Asp518, Asp616, Trp613 (2), Trp516, Trp376, Phe649, Met519, and His674 residues.   The molecular docking of Xanthones 2 and 3 was also studied on P. falciparum lactate dehydrogenase enzyme (PDB ID: 1CET). The total energy and MDS of compound 2, as shown in Table 7, was found to be 108.388 kcal/mol and −103.69 kcal/mol. Compound 3 showed values of 79.747 kcal/mol and −86.38 kcal/mol for the total energy and MDS. The interaction analysis indicated that compound 2 bound to 1CET via hydrogen bonding of oxygen atom of hydroxyl chelate group (C1) with Tyr85 residue and oxygen atom of carbonyl with Phe52 residue. The steric interaction was established by xanthone backbone and the dimethylallyl group of compound 2 with Tyr85, Glu122(2), Asp53(2), Gly27, Ala98, and Phe100 residues. The hydrogen bonding was also formed by carbonyl group, chelate hydroxyl, and free hydroxyl (C5) of compound 3 with Tyr85(2) and Phe52 residues. The Glu122, Tyr85, Asp53(2), and Ile119 residues of 1CET bound to carbonyl groups, free hydroxyl (C5), and dimethylallyl group of compound 3 by steric interactions (Figure 3).

Plasmodium Falciparum Lactate Dehydrogenase Enzyme (PDB ID: 1CET)
The molecular docking of Xanthones 2 and 3 was also studied on P. falciparum lactate dehydrogenase enzyme (PDB ID: 1CET). The total energy and MDS of compound 2, as shown in Table 7, was found to be 108.388 kcal/mol and −103.69 kcal/mol. Compound 3 showed values of 79.747 kcal/mol and −86.38 kcal/mol for the total energy and MDS.

ADMET Profiles
The absorption, distribution, metabolism, excretion, and toxicity properties of isolated compounds 1-5, as derived from ProTox online tool, revealed that all compounds had molecular weight less than 500 which is important for penetrability [40]. The Caco-2 permeability for prediction of orally administered drug absorption of compounds 1, 3, and 4 was > 0.90 and, 0.90 for compounds 2 and 5 (Table 8). This implied that compounds 1, 3, and 4 had high permeability in contrast to compounds 2 and 5 [41]. Compounds 1-5 showed high intestinal absorption (80-98%) and would be absorbed in the small intestine [42]. The transdermal efficacy as illustrated by skin permeability of compounds 1-5 was ranging from −2.735 to −2.851 cm/hour (<−2.5) which mean that they will penetrate the skin properly. It is known that molecules will show difficulty in the skin penetration if the logKp value were higher than −2.5 cm/hour [43]. The circulation in blood plasma (VDss) came out to be acceptable for all compounds 1-5 with value higher than −0.15 [44]. An important parameter for reducing side effects and toxicity represented by penetration via blood-brain barrier (BBB) showed that compounds 2 and 4 will sufficiently be able to penetrate, but compounds 2, 3, and 5 were very low to reach the brain. The CNS permeability signified for the permeability of the blood surface to the central nerve defined that compounds 1-4 with values of −2.89 to −1.679 better than acarbose and chloroquine, were able to permeate a central nervous system [45]. Table 8 informed that isolated xanthones, except 4 were noninhibitors for CYP2D6 and for CYP3A4 and will not interfere the CYP450 biotransformation in general, while compound 4 will be metabolized [46]. In terms of excretion, the total clearance of all compounds shared positive values meaning that they will be discharged quickly except for 5. The adverse interactions of compounds 1-5 with OCT2 inhibitors as denoted by OCT2 substrate parameter revealed that compounds 1-5 showed no potential contraindication. The toxicity level of compounds 1-5 predicted by using pkCSM-pharmacokinetics for hepatotoxicity and acute oral toxicity in rats (LD 50 ) represented the value was in the range 1.889 to 2.057 lower than chloroquine and acarbose with LD 50 value 2.888 and 2.495, respectively. Compound 4 had the highest LD 50, of 2.057 and were the most toxic than others [47]. Hepatotoxicity descriptor declared that compounds 4 and 5 cause hepatotoxicity while compounds 1-3 were not hepatotoxic.