Design, synthesis, molecular docking, and in vitro α-glucosidase inhibitory activities of novel 3-amino-2,4-diarylbenzo[4,5]imidazo[1,2-a]pyrimidines against yeast and rat α-glucosidase

In an attempt to find novel, potent α-glucosidase inhibitors, a library of poly-substituted 3-amino-2,4-diarylbenzo[4,5]imidazo[1,2-a]pyrimidines 3a–ag have been synthesized through heating a mixture of 2-aminobenzimidazoles 1 and α-azidochalcone 2 under the mild conditions. This efficient, facile protocol has been resulted into the desirable compounds with a wide substrate scope in good to excellent yields. Afterwards, their inhibitory activities against yeast α-glucosidase enzyme were investigated. Showing IC50 values ranging from 16.4 ± 0.36 µM to 297.0 ± 1.2 µM confirmed their excellent potency to inhibit α-glucosidase which encouraged us to perform further studies on α-glucosidase enzymes obtained from rat as a mammal source. Among various synthesized 3-amino-2,4-diarylbenzo[4,5]imidazo[1,2-a]pyrimidines, compound 3k exhibited the highest potency against both Saccharomyces cerevisiae α-glucosidase (IC50 = 16.4 ± 0.36 μM) and rat small intestine α-glucosidase (IC50 = 45.0 ± 8.2 μM). Moreover, the role of amine moiety on the observed activity was studied through substituting with chlorine and hydrogen resulted into a considerable deterioration on the inhibitory activity. Kinetic study and molecular docking study have confirmed the in-vitro results.

According to results, among derivatives in the first category (compounds 3a-x), it seems the presence of 4-Cl on 2-aryl ring plays a substantial role in anti-α-glucosidase activities. The presence of electron-donating group Additionally, the probable role of amine functional group has been investigated. For this goal, the α-glucosidase potency of compound 3a (IC 50 = 53.8 ± 0.04 μM) was compared with those of compounds 4a and 6a (the IC 50 values were 235.4 ± 0.5 μM and 168.6 ± 1.2 μM, respectively). As it can be observed, the order of activity was NH 2 > H > Cl substituted derivatives. Therefore, the necessary, constructive role of amine moiety on the inhibition of α-glucosidase has been confirmed.
To develop this investigation, the ability of our target compounds to inhibit the rat small intestine α-glucosidase have been evaluated. These inhibitory activities exhibited almost similar trend to that of Saccharomyces cerevisiae α-glucosidase. The most active compound was 3k (IC 50 value of 45.04 μM) which was 3.23 times more potent than acarbose (IC 50 value of 145.74 μM). Moreover, compounds 4a and 6a showed slight inhibitory activities confirmed the significance of amine moiety in targeted compounds 3.

Enzyme kinetic study.
To investigate the inhibition mode of synthesized poly-substituted 3-amino-2,4-diarylbenzo [4,5]imidazo[1,2-a]pyrimidine 3 against α-glucosidase, kinetic study was performed with standard inhibitor, acarbose, and the most potent derivative 3k. To indicate the type of inhibition and K i , Lineweaver-Burk plots and secondary re-plotting of the mentioned plots were presented (Fig. 2). As it was showed in Fig. 2a, while inhibitor concentration increased, the K m value gradually increased, but V m value remained unchanged. Therefore, it can be implied compound 3k was a competitive inhibitor and competes with acarbose for binding to the enzyme active site. Moreover, plot of K m versus different concentration of compound 3k gave an estimate of the inhibition constant, K i of 16 µM (Fig. 2b). Table 1. Substrate scope and α-glucosidase inhibitory activity of compounds 3a-ag. Values are the means of three replicates ± standard deviation (SD). b The activity against Saccharomyces cerevisiae α-glucosidase. c The activity against rat small intestine α-glucosidase.  Docking study. Molecular docking study was performed on the compounds 3a, 3k and 3ad to study the mode of their interaction in the active site of the yeast isomaltase from Saccharomyces cerevisiae (Pdb id:3A4A) with 84% similarity to S. cerevisiae α-glucosidase using Auto Dock Tools (version 1.5.6). These compounds showed similar binding modes of interaction with catalytic residues. The superimposed structure of acarbose as a standard inhibitor and the most potent compound 3k in the active site of isomaltase was shown in Fig. 3. In the most potent compound 3k, benzimidazole and 4-(4Cl-phenyl) ring units created π-π interaction with Phe 303 and Tyr 158, respectively in the active site of the enzyme (Fig. 4). The 2-(4Cl-phenyl) ring formed π-anion interaction with the aromatic side chains of Asp352. Moreover, a π-cation interaction was observed between pyrimidine moiety and Arg 442. Compounds 3a and 3k interacted with similar amino acids in the active site of the enzyme. Benzimidazol, pyrimidine, and 4-phenyl ring of compound 3a interacted with Phe303, Arg442, and Tyr158, respectively (Fig. 5). Compound 3k had additional π-alkyl interaction between 4-(4Cl-phenyl) ring and Arg315, as well as 2-(4Cl-phenyl) ring and Val 216. Higher observed inhibitory activity of compound 3k could be attributed to the formation of stabilizing interactions with specific residues like Arg315 and Val 216, which could be resulted from the presence of chlorine atoms led to the electron-deficiency of phenyl rings. Additionally, chlorine atoms  In compound 3ad, there was a difference in interaction mode of the 2-(4Cl-phenyl) moiety with the active site of enzyme. Insertion of chlorine in 7 and 8 positions on the benzimidazole moiety led to a significant decrease in the inhibitory activity. However, there was not any interaction between 2-(4Cl-phenyl) moiety and Asp352 (Fig. 6).
Further studies on the binding energies of selected compounds exhibited that compound 3k had lower free binding energy (− 9.63 kcal/mol) as compared to compounds 3ad (− 8.89 kcal/mol) and 3a (− 9.14 kcal/mol). As observed from the best docking conformations, showed that all three compounds have a lower free binding energy than acarbose (− 8.20 kcal/mol). Therefore, the results emphasized that the target compounds bind more easily to the target enzyme (α-glucosidase) than the reference drug, acarbose. These findings had good agreement with the obtained results through in vitro experiments.
To assess potential inhibition of human α-glucosidase, compound 3a was docked against the crystal structure for C-terminal domain of human intestinal α-glucosidase (PDB Code: 3TOP) comparing with Acarbose. This study exhibited similar interactions with the yeast isomaltase binding site. The superimposed structure of  Phenyl rings and pyrimidine were involved in several π-anion interactions with Asp 1279, Asp 1420, and Asp 1526. Moreover, compound 3a formed π-π stacking interaction with hydrophobic residue including Tyr1251, Trp1355, and Phe1559.

Conclusion
In conclusion, we introduced a novel, potent series of α-glucosidase inhibitors. Poly-substituted 3-amino-2,4-diarylbenzo [4,5]imidazo[1,2-a]pyrimidines were synthesized through an efficient, short-time, high-yield Michael addition-cyclization between 2-aminobenzimidazoles and α-azidochalcones under the mild conditions. No need to column chromatography led us to obtain a large scope of substrates, all of which exhibited good to excellent inhibitory activity. Among them, compound 3k showed the best inhibitory potency having IC 50 value of 16.4 ± 0.36 μM which was 45.7 times more potent than acarbose as standard inhibitor (IC 50 = 750.0 ± 1.5 μM). The kinetic study for this compound showed there was a competitive mechanism. Moreover, docking studies revealed that 3-amino-2,4-diarylbenzo [4,5]imidazo[1,2-a]pyrimidines could interact with important amino acids in the active site of α-glucosidase.

Experimental
Methods. All chemicals were purchased from Merck (Germany) and were used without further purification. Melting points were measured on an Electrothermal 9100 apparatus and were not corrected. Mass spectra were recorded on an Agilent Technologies (HP) 5973 mass spectrometer operating at an ionization potential of 20 eV. IR spectra were recorded on a Shimadzu IR-460 spectrometer. 1 H and 13 C NMR spectra were measured (DMSO-d 6 solution) with Bruker DRX-300 AVANCE (at 300.1 and 75.1 MHz) spectrometer with TMS as an internal standard. α-Azido chalcones 2 were obtained from the corresponding benzylidene acetophenones in two steps following the literature procedure 15 13 (1,10,20,50, 100, 500 and 1000 μM (5 μL)), and potassium phosphate buffer. Test compounds were dissolved in DMSO (not exceed than 5% of final volume). After 10 min. of pre-incubation at 37 • C, p-nitrophenyl glucopyranoside as substrate (5 μL, 3 mM), was added to the enzyme solution and let to be incubated for one hour at 37 • C. Finally, the change in the absorbance was followed at 405 nm using Cytation 3 hybrid microplate reader (BioTek, USA). DMSO and acarbose were used as the control and standard inhibitor, respectively. IC 50 values of tested compounds were obtained from the nonlinear regression curve using GraphPadprism 6.0 (San Diego, California, USA) (https:// www. graph pad. com/ scien tific-softw are/ prism/). All experimental animal procedures were approved by the Animal Care, use Ethics Committee at Shahid Beheshti University of Medical Sciences (SBMU), and comply with the Animal Research Reporting of In Vivo Experiments (ARRIVE) guidelines. All methods proposed here were performed in accordance with relevant institutional guidelines and regulations.

Molecular docking study.
Since the X-ray crystallographic structure S. cerevisiae α-glucosidase isn't accessible, the 3D structure of S.cerevisiae isomaltase with PDB ID: 3A4A was downloaded from RCSB web site with 84% similarity to S. cerevisiae α-glucosidase 15 .
Docking studies were performed based on previous studies 34,70,71 using Auto Dock Tools (version1.5.6), and the pdb structure of 3A4A and 3TOP were taken from the Brookhaven protein database (http:// www. rcsb. org). The 3D structures of the selected compounds were created by MarvineSketch 5.8.3, 2012, ChemAxon (http:// www. chema xon. com) and converted to pdbqt coordinate using Auto dock Tools. Before preparation of auto dock format of protein, the water molecules and the inhibitors were removed from it. Then, using Auto Dock Tools, polar hydrogen atoms were added, Kollman charges were assigned, and the obtained enzyme structure was used as an input file for the AUTOGRID program. In AUTOGRID for each atom type in the ligand, maps were calculated with 0.375 A spacing between grid points, and the center of the grid box was placed at x = 22.625, y = − 8.069, and z = 24.158 for 3A4A and x = − 51.5, y = 9, and z = − 64.8 for 3TOP. The dimensions of the active site box were set at 50 × 50 × 50 A. Each docked system was carried out by 150 runs of the AUTODOCK search by the Lamarckian genetic algorithm. The best pose of each ligand was selected for analyzing the interactions between α-glucosidase and the inhibitor. The results were visualized using Discovery Studio 4.0 Client (https:// disco ver. 3ds. com/ disco very-studio-visua lizer-downl oad) and LigPlot (https:// www. ebi. ac. uk/ thorn ton-srv/ softw are/ LigPl us/ downl oad. html) (Figs. 3, 4, 5, 6, 7).