GC-MS Analysis and Inhibitory Evaluation of Terminalia catappa Leaf Extracts on Major Enzymes Linked to Diabetes

Terminalia catappa leaves are used in managing both diabetes mellitus and its complications in Southwest Nigeria. However, its inhibitory activity on enzymes implicated in diabetes is not very clear. This study investigated the in vitro inhibitory properties and mode of inhibition of T. catappa leaf extracts on enzymes associated with diabetes. The study also identified some bioactive compounds as well as their molecular interaction in the binding pocket of these enzymes. Standard enzyme inhibition and kinetics assays were performed to determine the inhibitory effects of aqueous extract (TCA) and ethanol extract (TCE) of T. catappa leaves on α-glucosidase and α-amylase activities. The phytoconstituents of TCA and TCE were determined using GC-MS. Molecular docking of the phytocompounds was performed using Autodock Vina. TCA and TCE were the most potent inhibitors of α-glucosidase (IC50 = 3.28 ± 0.47 mg/mL) and α-amylase (IC50 = 0.24 ± 0.08 mg/mL), respectively. Both extracts displayed a mixed mode of inhibition on α-amylase activity, while mixed and noncompetitive modes of inhibition were demonstrated by TCA and TCE, respectively, on α-glucosidase activity. The GC-MS analytic chromatogram revealed the presence of 24 and 22 compounds in TCE and TCA, respectively, which were identified mainly as phenolic compounds, terpenes/terpenoids, fatty acids, and other phytochemicals. The selected compounds exhibited favourable interactions with the enzymes compared with acarbose. Overall, the inhibitory effect of T. catappa on α-amylase and α-glucosidase may be ascribed to the synergistic action of its rich phenolic and terpene composition giving credence to the hypoglycaemic nature of T. catappa leaves.


Introduction
Diabetes mellitus (DM) is an endocrine, chronic, noncommunicable disease plaguing the world populace with a rapid increase. A reported 425 million individuals were globally affected by DM, while 629 million people have been projected to be affected by 2045 [1]. DM is characterized by hyperglycaemia as a consequence of impaired insulin secretion (as experienced in type 1 diabetes) or insulin resistance (as experienced in type 2 diabetes) resulting in diabetic complications such as diabetic retinopathy, neuropathy, and nephropathy [2]. Type 2 diabetes (T2D) is the most prevalent type of DM affecting over 90% of people diagnosed with this disease [3]. Lifestyle modification through exercise and diet as well as oral medications such as metformin, pioglitazone, and acarbose to decrease hepatic glucose output and insulin sensitivity improvement and reduce starch digestibility, respectively, are management methods currently employed in T2D [4].
Terminalia catappa Linn, commonly known as Indian almond, belongs to the Combretaceae family and grows in the tropics of Asia, Africa, and Australia [5]. In urban regions where these trees are found, the leaves form a menace and are the major constituents of generated lignocellulosic waste. In Southwest Nigeria, it is commonly called "igi furutu" or "igifuruntu," and various plant parts are used to treat diabetic complications by the locals [6]. Several studies have reported different activities of T. catappa extracts such as hepatoprotective effects, anticancer property, antimutagenic activity, and antiaging property [7]. Divya and Anand [8] have also reported on the inhibitory property of T. catappa methanolic leaf extract on diabetic-linked enzymes. Despite this antidiabetic claim by the locals, the elaborate antidiabetic mechanism is far from clear. is study assessed the inhibitory properties of T. catappa leaf extracts on α-glucosidase and α-amylase, the mode of enzyme inhibition, as well as identified phytocompounds present and proposed the molecular mechanism of binding in the active sites of the enzymes.

Materials and Methods
2.1. Materials. α-Glucosidase, α-amylase enzymes, and their substrates were acquired from Solarbio Life Sciences, Beijing, China. Other chemicals were products of Sigma-Aldrich, St. Louis, USA.

Plant Collection, Identification, and Extraction.
Mature T. catappa leaves were sourced between October and December 2016, from Covenant University compound. ey were identified by Dr. J. O. Popoola of Biological Sciences Department and voucher specimen deposited at Biological Sciences Department herbarium, Covenant University, Ota, Ogun State, with herbarium number TC/CUBio/H809. Aqueous T. catappa (TCA) and ethanol T. catappa (TCE) leaf extracts were prepared as reported by Iheagwam et al. [9]. e leaves were cut, air-dried, pulverised, and macerated in distilled water and ethanol (80%), respectively, at 1 : 10 (w/ v) ratio for 72 hrs. e obtained filtrates were concentrated using a rotary evaporator.

Antidiabetic Assessment
2.3.1. α-Glucosidase Inhibitory Activity. α-Glucosidase inhibitory activity of the extracts was evaluated according to the method described by Ibrahim and Islam [10] with slight modification. Various extract concentration and acarbose (1-5 mg/mL, 250 μL) were incubated at 37°C for 15 min with α-glucosidase solution (1 U/mL, 500 μL). ρ-Nitrophenylα-D-glucopyranoside (pNPG) solution (5 mM, 250 μL) was thereafter added, and the resulting mixture was incubated for 20 min at 37°C. e reaction was terminated by adding Na 2 CO 3 (0.2 M, 100 μL), and absorbance was measured at 405 nm. Phosphate buffer (100 mM) was used as control in place of inhibitors. Inhibitory activity was calculated using the following equation: where A s = absorbance in the presence of sample and A c = absorbance of control. All solutions were prepared in 0.1 M phosphate buffer (pH 6.8). e method of Sabiu and Ashafa [11] was adopted for α-glucosidase inhibitory kinetics. Extract (5 mg/mL, 250 μL) was preincubated with α-glucosidase solution (1 U/mL, 500 μL) for 10 min at 25°C. Varying pNPG concentrations (0.15-5 mg/mL, 250 μL) were added and incubated for 10 min at 25°C to both sets of reaction mixtures to start the reaction. ereafter, Na 2 CO 3 (0.2 M, 500 μL) was added to stop the reaction. For the control kinetic reaction, 100 mM phosphate buffer (pH 6.8, 250 μL) was used in place of the extract. Reaction rates (v) were calculated, and double reciprocal plots of α-glucosidase inhibition kinetics were determined.

α-Amylase Inhibitory
Activity. α-Amylase inhibitory activity of the extracts was evaluated by adopting the method described by Ibrahim and Islam [10] with slight modification. Various extract concentrations (1-5 mg/mL, 250 μL) and acarbose were incubated at 37°C for 20 min with amylase solution (2 U/mL, 500 μL). Starch solution (1%, 250 μL) was later added to the reaction mixture and incubated at 37°C for 1 h. Dinitrosalicylic acid (DNS) colour reagent (1 mL) was added to stop the reaction. e resulting mixture was boiled for 10 min, and absorbance was measured at 540 nm. Phosphate buffer (100 mM) was used as control in place of inhibitors. e α-amylase inhibitory activity was calculated using the following formula: where A s � absorbance in the presence of sample and A c � absorbance of control. All solutions were prepared in 100 mM phosphate buffer (pH 6.8). e method of Sabiu and Ashafa [11] was adopted for α-amylase inhibitory kinetics. In brief, extract (250 μL, 5 mg/ mL) was incubated with α-amylase (2 U/mL, 500 μL) for 10 min, before the addition of various substrate concentrations (0.3-10 mg/mL, 250 μL). e reaction proceeded as highlighted for α-glucosidase. α-Amylase inhibition kinetics was determined from the Lineweaver-Burk double reciprocal plot.
e GC-MS analysis of T. catappa extracts was carried out using GCMS-QP2010SE SHIMADZU, Japan, fused with the Optima 5 ms capillary column (30 × 0.25 mm) of 0.25 μm film thickness following the described method of Ajiboye et al. [12] with slight modifications. e gas chromatography conditions were as follows: pure helium carrier gas (flow rate: 1.56 mL/min; linear velocity: 37 cm/s), initial column oven temperature (60°C) programmed to increase to 160°C at the rate of 10°C/min and then finally to 250°C with a hold time of 2 min/increment, and an injection volume of 0.5 μL in the splitless mode with a split ratio of 1 : 1 and injector temperature set at 200°C. Mass spectrophotometer conditions were as follows: ion source temperature (230°C), interface temperature (250°C), solvent delay at 4.5 min, and acquisition in a scan range of 50-700 amu. Electron ionization mode and multiplier voltage were set at 70 eV and 1859 V, respectively. Retention time, fragmentation pattern, and mass spectral data of the unknown components in the extracts were compared with those in Wiley and National Institute of Standards and Technology (NIST) libraries for compound identification.

Ligand and Protein
Modelling. e structures of the GC-MS identified compounds with ≥5% abundance were prepared as reported by Iheagwam et al. [13]. e 3D structure of α-glucosidase and α-amylase was modelled using the crystal structures with PDB codes 5kzw and 1b2y, respectively, obtained from RCSB protein data bank as templates in SWISS-MODEL [14].

Virtual Screening, Drug-Likeness, and Molecular
Docking. Virtual screening of selected identified ligands, analysis of drug-likeness using the rule of five (RO5), and molecular docking were carried out according to the methodology of Iheagwam et al. [13]. However, grid dimensions of the binding pockets were 60 × 40 × 32 and 40 × 34 × 40 points separated by 1Å for α-glucosidase and α-amylase, respectively. Inhibition constant (K i ) of docked ligands were calculated by using the following formula: 2.6. Statistical Analysis. Data were analysed using SPSS version 25 (IBM Corp., New York, USA) and subjected to one-way analysis of variance (ANOVA) using the Duncan multiple range post hoc test. Values were reported as mean ± standard deviation (SD) of three (3) replicates and considered significantly different at p < 0.05.

Results
For the α-glucosidase inhibitory activity of TCA and TCE as shown in Figure 1, a significantly (p < 0.05) lower inhibition by the extracts was observed at all concentrations relative to control. TCA exhibited a significantly (p < 0.05) higher inhibition of α-glucosidase activity compared to TCE. Nonetheless, at lower concentrations (1-3 mg/mL), there was no difference between the inhibitory activities of TCA and TCE. ese were further supported by a lower IC 50 value (2.23 ± 0.21 mg/mL) for acarbose when compared with TCA (3.28 ± 0.47 mg/mL) and TCE (3.78 ± 0.26 mg/mL (Table 1). e kinetic study on the inhibition mode using the double reciprocal plot revealed TCE exhibited a noncompetitive mode of inhibition with a common K m value of 0.19 mM and V max value of 0.13 mM/min, while TCA exhibited a mixed mode of inhibition with a K m value of 0.77 mM and V max value of 0.1 mM/min ( Figure 2). e percentage inhibition of α-amylase activity by T. catappa leaf extracts is presented in Figure 3. ough a concentration-dependent effect was observed, TCE inhibitory activity was significantly (p < 0.05) higher than TCA and acarbose at all concentrations. TCA elicited inhibitory effects that competed favourably with the standard drug (acarbose). ese results were supported with an IC 50 of 0.24 ± 0.08, 0.75 ± 0.14, and 0.85 ± 0.18 mg/mL recorded for TCE, TCA, and acarbose, respectively (Table 1) Tables 2 and 3, respectively, based on their retention time, abundance, and compound classification. GC-MS analysis revealed the presence of 24 compounds in TCE and 22 compounds in TCA. Seven compounds were found in both extracts; however, phytol and n-hexadecanoic acid were higher in TCE, while 4H-pyran-4-one, 2,3-dihydro-3,5-dihydroxy-6-methyl-, benzofuran, 2,3-dihydro-, 2methoxy-4-vinylphenol, and 9,12-octadecadienoic acid (Z,Z)-were higher in TCE. It was also observed that there was no much difference in the abundance of vitamin E in both extracts.
When the hit compounds were screened for their druglikeness, they all obeyed Lipinski's RO5. However, phytol Evidence-Based Complementary and Alternative Medicine     Evidence-Based Complementary and Alternative Medicine and vitamin E, on the one hand, violated only the octanolwater partition coefficient due to higher values than the RO5 threshold as presented in Table 6. Acarbose, on the other hand, violated 3 variants. e binding affinity of the selected compounds as shown in Table 7 using Autodock Vina ranged from − 6.0 to 8.0 kcal/ mol and − 5.1 to 5.9 kcal/mol for α-amylase and α-glucosidase, respectively. ese values though lower were comparable with acarbose where − 8.3 was recorded for α-amylase and − 7.4 for α-glucosidase. Concomitantly, 1.39 to 40.51 μM was the α-amylase inhibition constant (K i ) recorded for the compounds compared to 0.84 μM for acarbose, while 47.95 to 184.70 μM was the α-glucosidase K i recorded for the compounds compared to 3.83 μM for acarbose.
As depicted in Figure 7, the ligands bound to both the active and allosteric sites of the enzymes. It further justified the in vitro results as the majority of the ligands favoured active site binding compared to the allosteric site. Hydrogen, van der Waals, and π bonds were the   (Figures 8 and 9).

Discussion
α-Glucosidase and α-amylase are major enzymes that metabolise carbohydrate in the digestive tract thereby affecting carbohydrate metabolism. Drugs which illicit their pharmacological action by inhibiting these enzymes are used as therapeutic control in managing diabetes through the control of postprandial hyperglycaemia [15,16]. Research on inhibitors of these enzymes especially from medicinal plants has been intensified due to their claim of being inexpensive and less toxic compared to synthetically derived medications such as acarbose and miglitol with similar mechanisms of action [17]. Promising inhibitory activity of T. catappa leaf extracts was exhibited on α-glucosidase and α-amylase as previously reported in a dose-dependent manner [8]. Nonetheless, this potential was more portrayed in α-amylase activity as T. catappa leaf extracts exhibited a better inhibitory potential than acarbose. is was corroborated by various studies that have previously reported a higher inhibitory potential of medicinal plant extracts than acarbose [16,18,19]. It was also noteworthy that our extracts had better α-glucosidase and α-amylase inhibitory activities than those reported for Nicotiana tabacum and Calotropis procera leaf extracts [20,21]. e reported α-glucosidase and α-amylase inhibitory activities of Sutherlandia montana and Aerva lanata (ethanol) leaf extracts were higher than our extracts except for A. lanata aqueous leaf extract α-amylase inhibitory activity which was reported to be lower than ours [4,22]. Contrary to the reports of Xu et al. [23] and Wan et al. [24], the inhibitory activity of T. catappa leaf extracts was higher on α-amylase than on α-glucosidase at the varied   Evidence-Based Complementary and Alternative Medicine concentrations and may be attributed to the different mechanism of action on these enzymes. is was further buttressed by the kinetic studies, where the TC extracts exhibited a mixed mode of inhibition on α-amylase, while mixed and uncompetitive inhibition mechanisms were observed for TCA and TCE, respectively, on α-glucosidase. e mixed mechanisms exhibited by TCA and TCE may suggest the bioactives present in the extracts may bind in the active site of these enzymes thereby reducing the affinity of the substrate [25,26]. Binding of these phytochemicals in the allosteric site is also a possible mechanism of action which may lead to a conformational change of these enzymes leading to a reduction in substrate affinity for the active site concomitantly hampering enzyme catalysis [25,26]. e results suggest these extracts may have more affinity for the enzyme (E) than the enzyme-substrate complex (ES). e noncompetitive inhibition by TCE would suggest the phytochemicals    [27]. e observed inhibitory action observed for TC extracts may be attributed to the synergistic action of identified phytochemicals from the gas chromatogram. Fatty acids, phenolic compounds, and terpenes/terpenoids were the majority classes of identified phytochemicals in both extracts. Phenolic compounds and terpenoids have also been reported to elicit antioxidant properties and alleviate oxidative stress accumulation, in the process preventing the progression  Figure 7: Binding of ligands in the active and allosteric pockets of (a) α-glucosidase and (b) α-amylase. e ligands ethyl-α-D-glucopyranoside, vitamin E, n-hexadecanoic acid, phytol, and acarbose were colour coded as black, blue, purple, green, and red, respectively. of diabetic complications [28]. Compounds such as phytol [29,30], various terpenes and terpenoids [11], hexadecanoic acid, ethyl ester, and 9,12-octadecadienoic acid (Z,Z)- [31] have been reported to exhibit various antidiabetic activities. Furthermore, reports have it that hydrolysis of phenolic compounds leads to the  Evidence-Based Complementary and Alternative Medicine generation of shorter phenolic groups which accumulate, reduce oxidative stress, and inhibit amylase activity as well as other digestive enzymes reducing starch digestion [28,32,33]. is could also explain the better amylase inhibitory property of the extracts when compared with the glucosidase inhibitory activity. Pharmaceutical industries use structure-based drug design to solve challenges affecting integrated and classical drug design [34].
In lead compound development, compliance of test compound physicochemical properties (molecular mass, number of hydrogen bond donors and acceptors and so on) to Lipinski rule of 5 (RO5) is imperative to avoid failure during clinical trials [35,36]. Compounds that pass RO5 (usually with none or one default) are predicted to have optimal pharmacokinetic properties, consequently subjecting them further to molecular docking [13]. Since all compounds passed RO5, they may exhibit good pharmacokinetic properties. Molecular docking further gave us a better understanding of the binding interaction between some identified phytochemicals and the key carbohydrate hydrolysing enzymes. e relatively lower binding affinity and inhibitory constant of the individual bioactives than acarbose could be due to the lesser number of hydrogen bonds present between the amino acids and the hydrogen donor/acceptor atoms in the ligands. is finding was contrary to what Pérez-Nájera et al. [37] reported on Smilax aristolochiifolia root extract and its compounds where the number of hydrogen bonds did not affect binding affinity. Vitamin E had the lowest free energy and K i in amylase and glucosidase binding pockets which was comparable to acarbose. Consequently, it exhibited a more stable affinity with only a small concentration required to inhibit these enzymes [38]. Molecular docking further affirmed the in vitro inhibitory mechanisms as more identified compounds bound to the active site than the allosteric site signifying a preference for the (E) to elicit their potential pharmacological action [39]. e common interaction between Trp, Tyr, Ile, Ala, and Asp in the binding pockets of the enzymes and ligands (acarbose, vitamin E, and phytol) suggests nonpolar bonds (van der Waals force) are the major interactions occurring between the extracts and enzymes. Trp and Asp have previously been identified as common amino acids stabilising the interactions between glucosidase and various ligands, while Tyr was reported for amylase [39][40][41].

Conclusion
is is the first time, to the best of our knowledge, the inhibitory mechanism of T. catappa leaf extracts on glucosidase and amylase is being reported, making it an effective agent in managing postprandial hyperglycaemia. ese extracts preferably bind to the active site of these enzymes where their various identified compounds synergistically illicit their inhibitory action. From the different GC-MS identified compounds, vitamin E was the most potent ligand that qualified as a potential drug candidate after docking studies. ese plants can be leveraged upon as a natural source of not only vitamin E but other antidiabetic compounds for drug formulation. On the other hand, isolation and characterisation of these identified phytocompounds in addition to in vivo studies are still required to confirm these findings.

Data Availability
e data used to support the findings of this study are included in the article.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.