In vitro and in silico studies of bis (indol-3-yl) methane derivatives as potential α-glucosidase and α-amylase inhibitors

Abstract In this paper, bis (indol-3-yl) methanes (BIMs) were synthesised and evaluated for their inhibitory activity against α-glucosidase and α-amylase. All synthesised compounds showed potential α-glucosidase and α-amylase inhibitory activities. Compounds 5 g (IC50: 7.54 ± 1.10 μM), 5e (IC50: 9.00 ± 0.97 μM), and 5 h (IC50: 9.57 ± 0.62 μM) presented strongest inhibitory activities against α-glucosidase, that were ∼ 30 times stronger than acarbose. Compounds 5 g (IC50: 32.18 ± 1.66 µM), 5 h (IC50: 31.47 ± 1.42 µM), and 5 s (IC50: 30.91 ± 0.86 µM) showed strongest inhibitory activities towards α-amylase, ∼ 2.5 times stronger than acarbose. The mechanisms and docking simulation of the compounds were also studied. Compounds 5 g and 5 h exhibited bifunctional inhibitory activity against these two enzymes. Furthermore, compounds showed no toxicity against 3T3-L1 cells and HepG2 cells. Highlights A series of bis (indol-3-yl) methanes (BIMs) were synthesised and evaluated inhibitory activities against α-glucosidase and α-amylase. Compound 5g exhibited promising activity (IC50 = 7.54 ± 1.10 μM) against α-glucosidase. Compound 5s exhibited promising activity (IC50 = 30.91 ± 0.86 μM) against α-amylase. In silico studies were performed to confirm the binding interactions of synthetic compounds with the enzyme active site.


Introduction
Diabetes mellitus (DM) is one of common metabolic disease characterised by hyperglycaemia. 1 The main clinical treatment strategies for DM is to control the blood glucose level using drugs. 2 The catalytic hydrolysis of carbohydrates by enzymes such as a-glucosidase and a-amylase is the most important reason for the increase of glucose in blood. 3 a-Glucosidase(EC 3.2.1. 20), existing in the surface of small intestine, is an important hydrolase enzyme. [4][5] It catalyses the hydrolysis of carbohydrates into absorbable glucose monomers by spliting the bond between glucosidic oxygen and glucosyl residues of carbohydrates. 6-7 a-Amylase (E.C.3.2.1.1), one hydrolase enzyme, is mainly secreted by the pancreas and salivary glands. [8][9] a-Amylase catalytic hydrolyse the starch to produces maltose and glucose by breaking a-1,4-glucosidic bonds. [10][11] The inhibition the activity of a-glucosidase or a-amylase delay the hydrolase of polysaccharides, consequently, the postprandial blood glucose level can be reduced, which is believed as an effective approach for the treatment strategy of DM. To date, a large amount of a-glucosidase and a-amylase inhibitors are obtained from natural products and chemical synthesis. [12][13][14][15] However only few have further application, such as acarbose, voglibose, and miglitol. However, these clinical drugs are often associated with side-effects. Therefore, it is still worth further investigation for the development of more effective inhibitors towards a-glucosidase and a-amylase.
Bis (indol-3-yl) methanes (BIMs), as the key skeletons, present in a variety of bioactive natural products isolated from marine organisms, land plants and microorganisms. [16][17][18] And such a fact has also stimulated the synthesis of different BIMs leading to the reveal of a wide range of bio-pharmacological activities, including anti-fungal, anti-inflammatory, anti-oxidant, anti-cancer, and antibacterial activities. [19][20][21][22] Especially, it is notable that BIMs are reported to process activities of lowering blood lipids and preventing obesity 23 as well as inhibiting a-glucosidase activity, showing great potential in the treatment of DM. In order to find potential a-glucosidase inhibitors, some BIMs were synthesised and presented effective inhibitory activity. [24][25] The cyano group, a carbon-nitrogen triple bond, has been widely used in the structural modification and transformation of small drug molecules. It can change the physical and chemical properties of small molecules, enhance the interaction between drug molecules and target proteins to improve drug efficacy, improve the metabolic stability of compounds in the body, and so on. 26 The cyano group also is used as an important substitution group in some reported a-glucosidase inhibitors. [27][28] Based on the principle of combination of active structural moieties in drug design and our synthetic methodology for BIMs derivatives, 29 we synthesised BIMs derivatives to evaluate them for their a-glucosidase and a-amylase inhibitory activity.

Chemistry
Based on the effective synthetic method of BIMs reported by our group, 5a~c, 6a~c, and 7a~c were synthesised from N,N-dimethylaminomalononitrile (1) with substituted indole (2), substituted pyrrole (3) or N-phenylaminoacetic acid ethyl ether (4) under the catalysis of Al(OTf) 3 . The synthetic route was shown in Scheme 1. All synthetic compounds are known compounds, which were reported in references. 29 2.2. Evaluation of activity against a-glucosidase

Structure activity relationships (SAR) analysis
The SAR of compounds was discussed by analysing the substitution pattern on indole moiety. Compound 5a (IC 50 ¼ 55.53 ± 0.40 mM) with no substitution on indole ring showed $5 fold stronger compared to acarbose (IC 50 ¼ 261.45 ± 2.17 mM). Compound 5r (IC 50 ¼ 100.96 ± 1.69 mM) with a methyl group at 1position of indole presented decreased activity. However, introducing bromine group on the indole ring of 5r, compound 5 s (IC 50 ¼ 30.48 ± 1.27 mM), caused obviously increase in inhibition activity. The introduction of two phenyl groups at the 2-position of indole ring (compound 5 h, IC 50 ¼ 9.57 ± 0.62 mM) resulted in markedly increase of activity ( Figure 1). The above results indicated possible hydrogen bond interaction between compound 5a and a-glucosidase, halogen bond between 5 s and glucosidase, as well strong p-p stacking effect between 5 h and a-glucosidase.
Compared to compound 5a, compounds 5c~5f with chlorine substituent on the benzene ring of indole had more potent activities. This indicated that chlorine atom, a typical halogen atom, might be well interact with a-glucosidase through halogen bond like that of compound 5 s ( Figure 2). Among chlorine substituted derivatives, 5e (IC 50 ¼ 9.00 ± 0.97 mM) with the 6-chlorine group was found to be most active and 29 times better activity than acarbose. It's positional isomer 5d (IC 50 ¼ 10.22 ± 0.63 mM) with 5chlorine group, presented similar activity. However, 5c (IC 50 ¼ 45.10 ± 1.42 mM) and 5f (IC 50 ¼ 20.65 ± 0.42 mM) with 4-chlorine group and 7-chlorine group respectively showed lower activities.
Analysing effect of different type of substituent at same substituent position (Figure 4), halogen atom substituents lead to strengthen of a-glucosidase inhibitory activities, while electron donating and withdrawing groups resulted in decrease of the activity, as compared to compound 5a. Compound 5 g (IC 50 ¼ 7.54 ± 1.10 mM) with 5-bromine groups presented 34 times stronger activity than acarbose, and 5 b (IC 50 ¼ 40.86 ± 0.27 mM) with 5-Scheme 1 Synthesis of BIMs. Reagents and conditions: Al(OTf) 3 fluorine groups displayed 6 times stronger activity than acarbose. However, compound 5k (IC 50 ¼ 55.99 ± 0.78 mM) with 5-methyl group and 5p (IC 50 ¼ 50.37 ± 1.06 mM) with 5-methoxy group exhibited lower activities than that of 5a. In addition, the introduction of electron withdrawing group, methyl formate, 5q (IC 50 ¼80.12 ± 0.79 mM) resulted in a decrease on the a-glucosidase inhibitory activity.
Compared to series of 5a$s, 6a$c and 7a$c only presented moderate a-glucosidase inhibitory activities with IC 50 values from 100.38 ± 0.53 mM to 242.78 ± 5.14 mM, indicating that indole ring might be more beneficial than pyrrole ring and benzene ring on a-glucosidase inhibitory activities.

Inhibition mechanism study of a-glucosidase inhibitor
To gain further insight into the interaction between BIMs and a-glucosidase, the inhibition mechanism of compounds 5e, 5 g, and 5 h with highly potent inhibitory were investigated. The relationship of enzyme concentration with enzyme activity in presence of compounds 5e, 5 g, and 5 h was firstly investigated. As shown in Figure 5, the plots of enzyme concentration vs remaining enzyme activity at different inhibitor concentrations give a group of straight lines that all passed through the origin point. Those results indicated that the a-glucosidase inhibition of compounds 5e, 5 g, and 5 h was reversible.
The inhibitory kinetics of compounds 5e, 5 g, and 5 h on a-glucosidase were studied using Lineweaver-Burk plots. For compounds 5e, 5 g, and 5 h, the plots of 1/v vs 1/[S] give a lot of straight lines that intersected at the same point in the third quadrant respectively ( Figure 6), revealing that compounds 5e, 5 g, and 5 h were all mixed-type inhibitors. These results indicated that compounds 5e, 5 g, and 5 h could bind with free enzyme, as well as, enzyme-substrate complex to reduce the catalytic activity of a-glucosidase.
Besides, we also determined the equilibrium constants of binding of inhibitors to free enzymes (K I ) and enzyme-substrate complexes (K IS ) through plots of slope (K m /V m ) and vertical intercept (1/V m ) vs inhibitor concentration. The results were presented in Table 2. The K I values of compounds 5e, 5 g, and 5 h were higher than their K IS values, suggesting that the affinity of compounds 5e, 5 g, and 5 h with free enzyme was lower than that with enzyme-substrate complex.

Molecular docking study
Sybyl molecular docking program was used to simulate the binding modes of a-glucosidase with the topmost active compounds, i.e. 5e, 5 g, and 5 h to further understand the inhibition mechanism. As shown in Figure 7(a-d), compound 5e (orange compound), 5 h (yellow compound), and 5 g (cyans compound), were well nested into the active site of a-glucosidase and presented similar coordination with the active site of enzyme. The docking of compound 5e was presented in Figure 7(e). One indole ring nitrogen formed a hydrogen bond (2.0 Å) with Asp307, another indole ring nitrogen formed a hydrogen bond (1.9 Å) with Glu277, and cyano nitrogen formed a hydrogen bond (2.6 Å) with Arg315. Also, chlorine on indole ring formed a halogen bond with Arg442 (3.3 Å). For compound 5 g (Figure 7(f)), The two indole ring nitrogen made two hydrogen bonds with Gln353 (1.9 Å) and Glu411 (2.0 Å), respectively, and the two chlorines on indole ring made two halogen bond with Arg315 (2.0 Å) and Gln279 (3.7 Å), respectively. Moreover, one indole ring made p-p interactions with Phe303 (3.9 Å). Compounds 5 h (Figure 7(g)) also formed a hydrogen bond with Asp307 (2.0 Å), and formed p-p interactions with Phe303 (4.3 Å) and Tyr158 (4.5 Å), respectively. The hydrogen bond, p-p interactions, or halogen bond existed between a-glucosidase with compounds 5e, 5 g, or 5 h played very important roles in the binding of compounds and proteins.     . SAR analysis of compounds 5a, 5 b, 5d, 5 g, 5q, 5p, and 5k.

Sar analysis
The SAR of all compounds against a-amylase were also analysed with compound 5a (IC 50 : 493.59 ± 10.34 mM) as template compound. The introduction of halogen (F, Cl, Br) on indole ring effectively increased the inhibitory activity, Br group (compound 5 g, IC 50 : 32.18 ± 1.66 mM) is most beneficial for the inhibitory activity, and Cl group at 6-position (compound 5e, IC 50 : 40.03 ± 2.14 mM) of indole presented better inhibitory activity compared to other position. The introduction of methyl group also changed the inhibitory activity, and compound 5 l with methyl group at 6-position (IC 50 : 36.35 ± 1.37 mM) presented the best inhibitory activity. The introduction of methoxy, Phenyl, and ester group also increased the inhibitory activity, especially, compound 5 h with Phenyl group (IC 50 : 31.47 ± 1.42 mM) extended good inhibitory activity. Furthermore, compounds 6a$c and 7a$c presented effectively lower a-amylase inhibitory activities than compounds 5a$s, indicating that indole ring was more helpful for a-amylase inhibitory activities.

Inhibition mechanism study of a-amylase inhibitor
The highest potent inhibitory compounds 5 g, 5 h, and 5 s were evaluated their inhibition mechanism. As shown in Figure 8, the remaining enzyme activity at different inhibitor concentrations was measured and the obtained straight lines passed through the origin point, suggesting that compounds 5 g, 5 h, and 5 s reversibly inhibited the a-amylase. Lineweaver-Burk plots were used to analyse the inhibitory kinetics of compounds 5 g, 5 h, and 5 s. The plots of 1/v vs 1/[S] intersected a point at the second quadrant, indicating that 5 g, 5 h, and 5 s were mixed inhibitors ( Figure 9). Besides, the K I and K IS values of compounds 5 g, 5 h, and 5 s were measured, showing K I values were higher than K IS values (Table 3).

Docking simulation for a-amylase
As mentioned above, molecular docking between compounds 5 g, 5 h, and 5 s with a-amylase were used to explain their mechanism. Docking results revealed that compounds 5 g (yellow compound), 5 h (cyang compound), and 5 s (wheat compound) were well located within the active site of a-amylase (Figure 10(a-d)). Binding models of compounds 5 g, 5 h, and 5 s with a-amylase were depicted in Figure 10 were helpful to stabilise the formation of the ligand-protein complex. Based on above assays results of compounds targeting a-glucosidase and a-amylase, compounds 5e, 5 g, and 5 h showed highest a-glucosidase inhibitory and compounds 5 g, 5 h, and 5 s presented highest a-amylase inhibitory. Thence, compounds 5 g and 5 h displaying the potential bifunctional could be considered as the lead compounds for bifunctional drugs.

In vitro cytotoxicity assay
Moreover, the preliminary in vitro safety of compounds 5e, 5 g, 5 h, and 5 s were tested on Mouse Preadipocytes cells (3T3-L1) and Human liver cancer cells (HepG2) using MTT method and the results were shown in Figure 11. Compounds 5e, 5 g, 5 h, and 5 s displayed no effect on the viability of 3T3-L1 cells and HepG2 cells at concentration range of 2 0 $ 80 lM, suggesting these promising inhibitors had non-toxic towards live cells.

a-Glucosidase inhibition activities and mechanism assay
The a-glucosidase inhibitory activity of compounds 5a~c, 6a~c, and 7a~c were performed according to our previous report. 30,31 The 10 lL of a-glucosidase enzyme (final concentration 0.1 U/mL) and 10 lL of test compounds (dissolved in DMSO) were added into 130 lL of phosphate buffer (0.1 M, pH 6.8), followed 10 min incubation at 37 C. Then 50 lL of PNPG (final concentration 0.25 mM) was added, and mixture was continuely incubated for 20 min at 37 C. The absorbance at 405 nm was measured using Multimodel Reader. The percentage of enzyme inhibition was calculated: % Inhibition ¼ [(A 1 -A 0 )/A 0 ] Â 100%, where A 1 was the absorbance with the test compound, and A 0 was the absorbance without the test compound. The IC 50 value of compound was obtained from the plot of inhibition percentage vs test compound at different concentrations. Acarbose was used as the positive control. The experiment was performed in duplicate. The inhibition mechanism of compounds 5 g, 5e, and 5 h was analysed using similar above method. The enzyme inhibitory kinetics was detected using plots of enzyme concentration vs remaining enzyme activity at different inhibitor concentrations, and the substrate inhibitory kinetics was obtained from Lineweaver-Burk plot of remaining enzyme activity vs substrate concentration in the presence of different inhibitor concentrations. 32

a-Amylase inhibition activities and mechanism assay
The a-amylase inhibitory activity of all tested compounds was performed according to previous report. 33,34 A 10 lL of a-amylase enzyme solution (final concentration 0.25 U/mL), 10 lL of test compounds (dissolved in DMSO), and 80 lL of phosphate buffer (20 mM, pH 6.9) were mixed and incubated for 10 min at 37 C. Next, 100 lL starch solution (final concentration 0.5%) was added into the mixture followed by an incubation of 10 min. After 100 lL DNS (containing 1 M Potassium sodium tartrate and 48 mM 3,5-Dinitrosalicylic acid) was added, the mixture was kept incubated in boiling water for 15 min. Finally, the absorbance was measured at 540 nm after dilution of solution by adding 900 lL distilled water. The percentage of enzyme inhibition was calculated: % Inhibition ¼ [(A 1 -A 0 )/A 0 ] Â 100%, where A 1 was the absorbance with the test compound, and A 0 was the absorbance without the test compound. The IC 50 value of compound was obtained from the plot of inhibition percentage vs test compound at different concentrations. Acarbose was used as the positive control. The experiment was performed in duplicate.
The inhibition mechanism of compounds 5 g, 5 h, and 5S was also analysed. The enzyme inhibitory kinetics was detected using plots of enzyme concentration vs remaining enzyme activity at different inhibitor concentrations, and the substrate inhibitory kinetics was obtained from Lineweaver-Burk plot of remaining enzyme activity vs substrate concentration in the presence of different inhibitor concentrations.

Molecular docking
Molecular docking was conducted to explore the interaction of inhibitors with a-glucosidase and a-amylase using Sybyl (Version 2.1.1, Tripos, US) according to our previous report. 33 The crystal structure of a-glucosidase (PDB: 3AJ7) 33,35 and a-amylase (PDB: 3BAJ) 36,37 were obtained from the Protein Data Bank. Compounds were prepared with addition of hydrogen atoms, addition of charge with Gasteiger-H€ uckle mode, and energy minimisation. Then the protein was prepared by procedure of removing H 2 O, fixing side chain amides, and adding hydrogens. The active site of a-glucosidase was simulated out using automatic mode. The active site of a-amylase was simulated out using Ligen mode. Then, the docking simulations between compounds and a-glucosidase or a-amylase were carried out with the default format of Pymol program.

Cell cytotoxicity assay
The 3T3-L1 cells or HepG2 cells were cultured in DMEM medium containing 10% foetal bovine serum, 100 U/mL penicillin and 0.1 mg/mL streptomycin in a humidified incubator with a 5% CO 2 atmosphere at 37 C. Cells in the logarithmic growth phase were used for this assay. After 5 Â 10 3 3T3-L1 cells were seeded in 96 well plates for 24 h, compound with different concentration was added into each well for 24 h. MTT reagent (100 lL, 0.5 mg/mL) was added to each well for 4 h incubation. After the supernatant was discarded, 100 lL of DMSO was added. The absorbance was measured at 490 nm. Each sample was performed in 3 parallel experiments.