Design, modification, and bio-evaluation of salazinic acid derivatives

2.1 General experimental procedures

NMR spectra were recorded on a Bruker Avance III (500 MHz for ¹H NMR and 125 MHz for ¹³C NMR) using residual solvent signals as internal references: acetone‑d₆ at δ_H 2.05, δ_C 29.84 and chloroform-d at δ_H 7.26, δ_C 77.18. HRESIMS were recorded on Bruker MicrOTOF-QII and Agilent ESIQTOF mass spectrometers. Salazinic acid was isolated from the lichen Parmotrema indicum following the procedure in Do et al. (2021). Solvents were distilled before use. All reagents were purchased from Sigma-Aldrich. All reactions were monitored on silica gel 60 F₂₅₄ TLC plates (Merck). Thin layer chromatography (TLC) was carried out on precoated 60 F₂₅₄ silica gel or 60 RP–18 F_254S silica gel (Merck). Spots were visualized by spraying with 10% H₂SO₄ solution followed by heating. Gravity column chromatography was performed using 60 silica gel (0.040–0.063 mm, Himedia). Saccharomyces cerevisiae α-glucosidase (E.C 3.2.1.20), acarbose, and 4-nitrophenyl β-D-glucopyranoside (pNPG) were obtained from Sigma-Aldrich Co (Saint Louis, USA).

2.2 Isolation of salazinic acid from P. indicum

The material (3.8 kg) was cleaned, air-dried, ground, and macerated with ethyl acetate extract at room temperature. The filtrated solution was concentrated under reduced pressure to afford a crude ethyl acetate extract (719.52 g). The filtrated solution was evaporated to dryness and 152.32 g of the precipitate was collected. The crude extract was re-extracted using n-hexane and n-hexane–ethyl acetate extract (1:1, v/v) as solvents, affording three extracts: n-hexane (H, 112 g), n-hexane:ethyl acetate (HEA, 56 g), and a residual (EAR, 70 g).

The EAR was subjected to silica gel CC and eluted with a solvent system of n-hexane-EtOAc-AcOH (1:1:0.02, v/v/v), yielding fractions EAR1-EAR10. EAR10 (7.8 g) was subjected to further silica gel CC using n-hexane-chloroform-EtOAc-acetone-AcOH (1:2:2:2:0.01, v/v/v/v/v) as eluent. This afforded fractions EAR10.1-EAR10.5. Fraction EA10.1 (870 mg) was loaded into a Sephadex LH-20 CC and eluted with methanol to obtain fractions EA10.1.1-EA10.1.3. EA10.1.3 (670 mg) was washed with acetone, yielding compound 1 (480 mg).

2.3 Synthesis of analogues 1a-1c: General procedure

Sodium bromide (21 mg, 0.204 mmol) was added to a solution of salazinic acid (1, 60 mg, 0.155 mmol) in acetic acid (3 mL) at room temperature and dissolved under stirring. Hydrogen peroxide (0.021 mL, 0.182 mmol, 30%) was added each hour for 3.5 h at room temperature with stirring. The reaction was monitored using TLC. Saturated sodium hydrogen carbonate was used to stop the reaction then the neutralized mixture was partitioned between ethyl acetate-water (1:1). The organic layer was pooled, washed with brine, and dried over anhydrous Na₂SO₄. The residue was absorbed onto silica gel in column chromatography using CHCl₃-EtOAc-acetone-AcOH (40:8:5:2, v/v/v/v) as eluent. This yielded compounds 1a (28%), 1b (25%), and 1c (21%).

5,9-Dibromo-1,4,10-trihydroxy-8-methyl-3,7-dioxo-1,3-dihydro-7H-benzo[6,7][1,4]dioxepino[2,3-e]isobenzofuran-11-carbaldehyde (1a). A white amorphous powder with an isolated yield of 28%. Tables 1 and 2 present the ¹H and ¹³C NMR data for 1a: HRESIMS m/z calcd for C₁₇H₇Br₂O₉ [M−H]^-: 512.8456, found: 512.8438.

9,11-Dibromo-1,4,10-trihydroxy-5-(hydroxymethyl)-8-methyl-7H-benzo[6,7][1,4]dioxepino[2,3-e]isobenzofuran-3,7(1H)-dione (1b). A white amorphous powder with an isolated yield of 25%. See Tables 1 and 2 for ¹H and ¹³C NMR data on 1b: HRESIMS m/z calcd for C₁₇H₉Br₂O₉ [M−H]^-: 514.8613, found: 512.8612.

2,4,7-Tribromo-3,8-dihydroxy-9-(hydroxymethyl)-1-methyl-11-oxo-11H-dibenzo[b,e][1,4]dioxepine-6-carbaldehyde (1c). A white amorphous powder with an isolated yield of 21%. See Tables 1 and 2 for ¹H and ¹³C NMR data on 1c: HRESIMS m/z calcd for C₁₆H₈Br₃O₇ [M−H]^-: 548.7820, found: 548.7836.

2.4 Synthesis of analogues 2a and 2b: General procedure

Salazinic acid (1, 30 mg, 0.077 mmol) and corresponding carboxylic trichloroacetic acid (582 mg, 3.56 mmol) and acetic acid (213.5 mg, 3.56 mmol) were added to 6 mL of DMF dissolved with 3.3 mg of AlCl₃. The mixture was heated at 85 °C for 2 h under stirring. The reaction was periodically monitored using TLC. The mixture was partitioned between EtOAc-H₂O (1:1, v/v) and was further processed as described in Section 2.3. The residue was subjected to silica gel CC and eluted with CHCl₃- EtOAc-acetone-AcOH (40:8:5:2, v/v/v/v) to yield compounds 2a (21%) and 2b (32%).

(11-Formyl-1,4,10-trihydroxy-8-methyl-3,7-dioxo-1,3-dihydro-7H-benzo[6,7][1,4]dioxepino[2,3-e]isobenzofuran-5-yl)methyl 2,2,2-trichloroacetate (2a). A white amorphous powder with an isolated yield of 21%. Tables 1 and 2 present the ¹H and ¹³C NMR data for 2a: HRESIMS m/z calcd for C₁₉H₁₂Cl₃O₁₀ [M−H + H₂O-CO₂]^-: 504.9496, found: 504.9497.

Galbinic acid (2b). Another white amorphous powder with an isolated yield of 32%. The NMR data on 2b were consistent with an earlier report (Elix and Engkaninan, 1975; Devi et al., 2020).

2.5 Synthesis of analogues 3a, 3b, 4a, and 4b: General procedure

Salazinic acid (1, 40 mg, 0.103 mmol) was added to a solution of phenylhydrazine (44.6 mg, 0.412 mmol) in 10 mL of ethanol-acetic acid (7:0.08, v/v). The mixture was stirred for 3 h at 60 °C. The reaction was periodically monitored using TLC. The mixture was worked to yield an organic residue, which was subjected to silica gel CC and eluted with n-hexane-EtOAc-acetone (1:6:2, v/v/v) to yield compounds 3a (41%) and 3b (33%). Salazinic acid was reacted with 4-bromophenyl hydrazine (77.0 mg, 0.412 mmol) or 4-chlorophenyl hydrazine (73.7 mg, 0.412 mmol) to afford compounds 4a (64%) or 4b (47%), respectively.

(E)-5,11-Dihydroxy-6-(hydroxymethyl)-9-methyl-3-phenyl-12-((2-phenylhydrazineylidene)methyl)-8H-benzo[6,7][1,4]dioxepino[2,3-f]phthalazine-4,8(3H)-dione (3a). A white amorphous powder with an isolated yield of 41%. Tables 1 and 2 present the ¹H and ¹³C NMR data for 3a: HRESIMS m/z calcd for C₃₀H₂₁N₄O₇ [M−H]^-: 549.1409, found: 549.1410.

3,8-dihydroxy-9-(hydroxymethyl)-1-methyl-11-oxo-4-((E)-(2-phenylhydrazineylidene)methyl)-6-((Z)-(2-phenylhydrazineylidene)methyl)-11H-dibenzo[b,e][1,4]dioxepine-7-carboxylic acid (3b). A white amorphous powder with an isolated yield of 33%. See Tables 1 and 2 for the ¹H and ¹³C NMR data on 3b: HRESIMS m/z calcd for C₃₀H₂₁N₄O₇ [M−H]^-: 549.1409, found: 549.1409.

(E)-11-((2-(4-Bromophenyl)hydrazineylidene)methyl)-1,4,10-trihydroxy-5-(hydroxymethyl)-8-methyl-7H-benzo[6,7][1,4]dioxepino[2,3-e]isobenzofuran-3,7(1H)-dione (4a). A white amorphous powder with an isolated yield of 64%. See Tables 1 and 2 for the ¹H and ¹³C NMR data on 4a: HRESIMS m/z calcd for C₂₄H₁₆BrN₂O₉ [M−H]^-: 557.0019, found: 557.0057.

(E)-11-((2-(4-chlorophenyl)hydrazineylidene)methyl)-1,4,10-trihydroxy-5-(hydroxymethyl)-8-methyl-7H-benzo[6,7][1,4]dioxepino[2,3-e]isobenzofuran-3,7(1H)-dione (4b). A white amorphous powder with an isolated yield of 64%. See Tables 1 and 2 for the ¹H and ¹³C NMR data on 4b: HRESIMS m/z calcd for C₂₄H₁₆ClN₂O₉ [M−H]^-: 511.0544, found: 511.0564.

2.6 Synthesis of compound 5: General procedure

AlCl₃ (5.5 mg, 0.041 mmol) was added to a mixture of salazinic acid (1, 40 mg, 0.103 mmol) and orsellinic acid (17.3 mg, 0.103 mmol) in DMF solvent (2 mL). This was stirred at 90 °C for 3 h and the reaction was periodically monitored using TLC. The mixture was partitioned between ethyl acetate and water (50 mL each). Organic layers were pooled, washed with brine, and dried over anhydrous Na₂SO₄. The residue was purified by normal phase silica gel column chromatography eluted with CHCl₃-EtOAc-acetone-AcOH (100:40:24:8, v/v/v/v), yielding compound 5 (31%). The NMR data were consistent with a previous report (Devi et al., 2020).

2.7 In silico molecular docking: General procedure

The target compounds 1, 1a-1c, 2a-2b, 3a-3b, 4a-4b, and 5 were tested in silico for inhibition of α-glucosidase. Acarbose was used as control. AutoDockTools-1.5.6rc3 was used to model the docking of one receptor to each ligand. The crystal structures of the target proteins and receptors were assumed to be 5KEZ (PDB: https://doi.org/10.2210/pdb5KEZ/pdb), based on the protein data bank. The crystal structure of enzyme 5KEZ was detected as a hydrolase inhibitor, converting 4-nitrophenyl-α-D-glucopyranoside (p-NPG) to 4-nitrophenol and α-D-glucopyranose. If the enzyme was inactive or inhibited, the hydrolase conversion of 4-nitrophenyl-α-D-glucopyranoside (p-NPG) was not observed or proceeded slowly. The inhibitory effect of the ligand was evaluated based on this reaction. One strategy for controlling Type-2 diabetes is to slow the release of glucose from food and limit its absorption into the bloodstream. 5KEZ:PDB exhibited HPA (Human pancreatic alpha-amylase) inhibition and constitutes a novel enzyme inhibitor. This model can be used to detect glycosidase inhibitors that are highly selective for active carbohydrate enzymes (Jongkees et al., 2017).

The receptor was first tested for removal of small molecules like water, small ligands, and heteroatoms. The crystal structures were saved as format files (receptor.pdb) using the Discovery Studio 2019 Client (DSC) package. In the case of ligands, the minimum energy of conformation was investigated using the Molecular Mechanics & Force Fields (MMFF94) method, with optimal conformation discovery via the Avogadro package. This optimal conformation was saved. The AutoDock 4.2 package was used for docking of the ligand to the receptor. In the target procedure, a polar hydrogen and Kollman charge were added to the atoms. For ligands, this was added to all polar hydrogens, the Gasteiger charges were computed, non-polar hydrogen was merged, and the files were saved in pdbqt format. The values used were a grid point spacing of 0.375 Å, 120 x120 × 120 user-specified grid points, and X = –8,214, Y = 20.667, Z = -19.179 coordinates of the central grid point of maps. A Lamarckian genetic algorithm docked the conformation ligand to the macromolecular target. The maximum negative free energy of binding was selected to correspond to the most stable conformation after 2,500,000 energy evaluations and 200 runs (Thiratmatrakul et al., 2014). The Discovery Studio and Molegro (MMV) packages were used to visualize the results and present them as tables and figures. The docking procedure is described in Scheme S1.

2.8 α-Glucosidase inhibition assay

The α-glucosidase (0.2 U/mL) and substrate (5.0 mM p-nitrophenyl-α-D-glucopyranoside) were dissolved in 100 mM pH 6.9 sodium phosphate buffer. The inhibitor (50 µL) was preincubated with α-glucosidase, then the substrate (40 µL) was added to the reaction mixture. The enzymatic reaction was carried out at 37 °C for 20 min and stopped by the addition of 0.2 M Na₂CO₃ (130 μL). Enzymatic activity was quantified by measuring absorbance at 405 nm. All samples were analyzed in triplicate at five different concentrations around the IC₅₀ values, and the mean values were retained. The inhibition percentage (%) was calculated as follows:

Inhibition (%) = [1 – (A sample/A control)] × 100.

2.9 Inhibitory type assay of 4a and 5 on α-glucosidase

The mechanisms by which compounds 4a and 5 inhibits alpha-glucosidase were determined through Lineweaver-Burk plots (Microsoft Excel 2010, Washington, USA), using methods similar to those reported by Tran and co-workers (Tran et al., 2021). Enzyme inhibition at different concentrations of 4a and 5 was evaluated by the effect on the substrate. In Lineweaver–Burk double reciprocal plots of 1/enzyme velocity (1/V) vs. 1/substrate concentration (1/[S]), the inhibition type was determined using pNPG concentrations of 1 mM, 2 mM, and 4 mM in the presence of test compound concentrations of 0, 4.5, 9.0, 18.0, and 36.0 µM for 4a and 0, 3.16, 6.33, and 12.65 µM for 5. All experiments used three replicates. The mixtures were incubated at 37 °C and the 405 nm optical density was measured at 1 min interval for 30 min using a Clariostar Labtech microplate reader (Ortenberg, Germany). Optimal concentrations of the test compounds were chosen based on IC₅₀ values. The inhibition constants were obtained graphically from secondary plots (Microsoft Excel 2010, Washington, USA).

2.10 HPLC experiments proving the purity of compounds 4a and 5

The compounds 4a and 5 were analyzed through a HPLC-DAD using the modified method (Duong et al., 2017). Samples were injected separately. The mobile phase consisted of (ACN + 0.1% HCOOH) as solvent A and (H₂O + 0.1% HCOOH) as solvent B with a gradient of 5% − 10% A over 5 min, 10 $\to$ 30% A over 15 min, 30 $\to$ 80% A over 10 min, 80 $\to$ 100% A over 5 min, then 100% A for 5 min. The flow rate was 1 mL/min, and 10 µL of each sample was injected. The HPLC chromatograms were showed in Figs. S33-34.

3.1 Structural elucidation of synthetic compounds

In this study, ten salazinic acid derivatives were successfully synthesized. These included eight new compounds, 1a-1c, 2a, 3a, 3b, 4a, and 4b, and two known analogs, 2b and 5 (Fig. 1). Synthesis proceeded via four different routes, shown in Scheme 1. In the first route, salazinic acid (1) underwent bromination with sodium bromide and hydrogen hydroxide to yield 1a-1c. The chemical structures of 1a-1c (Tables 1 and 2) were identified from ¹H NMR, ¹³C NMR, HMBC, and HRESI mass spectra. The ¹H NMR data showed 1a-1c and 1 to be highly similar. 1a and 1 showed two structural differences: the absence of the aromatic methine proton (δ_H 6.88, δ_C 117.9, H-5) and the hydroxymethylene group (δ_H 4.64, δ_C 52.8, H₂-8′), suggesting that both H-5 and 8′-CH₂OH were replaced by bromine in 1a. This was confirmed by the HRESI mass data. In 1b, proton H-5 and the formyl group 3-CHO (δ_H 10.45, δ_C 193.1) were replaced by bromine, suggesting that the reaction occurred at C-3 and C-5 of 1. Scheme 2 proposes a possible mechanism for the formation of 1a-1c. First, bromination takes place at C-5, forming the intermediate 1a-1 (Abe et al., 2002; D’Aleo et al., 2013). This intermediate becomes oxidized, then decarboxylated, then brominated at C-3, forming 1b. A similar procedure occurs at C-8′ of the intermediate 1a-1, yielding 1a (Lei and Wang. 2008; Zhao et al., 2018; Perry et al., 2017; Quibell et al., 2018; Kong et al., 2019) (Scheme 2). 1b and 1c differ structurally in the addition of an extra aldehyde group (δ_C 189.8, δ_H 10.73) and the disappearance of the hemiacetal proton H-9′ from 1c. This suggests that the 5-member lactone of 1 becomes hydrolyzed, enabling decarboxylative bromination at C-1′ (Munoz and Murelli, 2012) and the formation of 1c (Scheme 2). In combination, these reactions give 1a-1c their unique scaffolds.

Close

Fig. 1

Chemical structure of compounds 1a-1c, 2a-2b, 3a-3b, 4a-4b, and 5.

Export to PPT

Close

Scheme 1

General synthetic routes to form 1a-1c, 2a-2b, 3a-3b, 4a-4b, and 5.

Export to PPT

Close

Scheme 2

The proposed mechanism of the bromination and nucleophilic addition to form 1a-1c and 3a-3b.

Export to PPT

Salazinic acid (1) esterification yielded 2a when using trichloroacetic acid and 2b when using acetic acid. NMR data showed 2a and 1 to be similar, with the only structural difference being observed in the B-ring. The H₂-8′ was shifted downfield (δ_H 4.88 in 2a compared with δ_H 4.64 in 1), suggesting that esterification of 1 appeared at C-8′.

Hydrazone formation was applied to compound 1, and yielded 3a-3b and 4a-4b. Their chemical structures were elucidated using ¹H NMR, ¹³C NMR, HMBC, and HRESI mass spectra. The active site 3-CHO of 1 was shown to react with 4-chlorophenylhydrazine or 4-bromophenylhydrazine to form 4a and 4b. Interestingly, compounds 3a and 3b were both obtained when phenylhydrazine was used as reagent under the same conditions. This indicated that phenylhydrazine is more reactive than either 4-chlorophenylhydrazine or 4-bromophenylhydrazine. Proton H-8 was shifted upfield in 3a-3b and 4a-4b (δ_H 9.03–9.33) relative to 1 (δ_H 10.45), indicating that the reaction occurred at 3-CHO. This was supported by the substitution of the imine carbon (δ_C 135.4) in 3a-3b and 4a-4b for the 3-CHO group (δ_C 193.1) in 1 and was backed up by the HMBC correlations of H-8 to C-2, C-3, and C-4. Spectroscopic analysis of 3a indicated that a further reaction occurred at C-9′. This was supported by a downfield shift of H-9′ (δ_H 8.48, δ_C 141.4 in 3a compared with δ_H 6.80, δ_C 95.0 in 1), together with the HMBC correlations of NH-1* (δ_H 7.44) to C-9′ (δ_C 141.4), C-2* (δ_C 128.7), and C-6* (δ_C 128.7), and of H-9′ (δ_H 8.48) to C-1′ (δ_C 109.7) and C-6′ (δ_C 122.2). NMR data showed strong structural similarities between 3b and 3a. The absence of the NH-1* proton in 3b indicated that a 6-member lactam was produced by the nucleophilic attack of an active amino group on carbonyl carbon C-7′ (Saundane and Kalpana. 2015; Rafeeq et al., 2015). Scheme 2 presents a possible mechanism for the formation of 3a and 3b.

A reaction between compound 1 and orsellinic acid yielded compound 5. NMR data showed 5 to be identical to parmosidone F (Devi et al., 2020). This supported the biosynthesis reaction proposed by Devi and co-workers. First, orsellinic acid becomes decarboxylated to orcinol and this intermediate undergoes Friedel-Crafts alkylation of the activated 3′ –CH₂OH in 1 to form compound 5.

3.2 In vitro α-glucosidase inhibition of compounds 1–5

Salazinic acid (1) and its derivatives 1a-1c, 2a-2b, 3a-3b, 4a-4b, and 5 were tested for inhibition of α-glucosidase (Table 3). Most of the synthetic products showed inhibition twice as potent as that of the mother compound (IC₅₀ 34.80 µM). In the Br-substituted products 1a-1c, the presence of bromine at C-3 enhanced inhibition. However, as the IC₅₀ values of these compounds were similar, the structural changes in the B-ring may have little effect on potency. The significant increase in potency of compound 2a may be attributed to the presence of chlorine. Compounds 3a, 3b, 4a, and 4b exhibited excellent potency, suggesting that imine moieties play an important role in α-glucosidase inhibition. Compound 5 was the most potent (IC₅₀ 9.72 µM), which was consistent with the findings of Devi et al., (2020), who isolated natural Friedel-Crafts alkylated depsidones similar to 5. They attributed the increase in potency to the transformation of the CH₂OH group in 1 to CH₂-Ph in 5, enhancing the activity of salazinic acid derivatives. In our study, the replacement of the 3-CHO of 1 by substituents in 1a-1c, 3a-3b, and 4a-4b significantly enhanced α-glucosidase inhibition.

3.3 In silico molecular docking

To elucidate the mechanisms of α-glucosidase inhibition, the binding modes in the active site were investigated. Figs. 2-21 and S1-S5, and Tables 4-5, show the interactions of the α-glucosidase enzyme with all ligands and with 1, 1a-1c, 2a-2b, 3a-3b, 4a-4b, 5, and the acarbose control. The α-glucosidase enzyme corresponded to the crystal structure of enzyme 5KEZ: PDB. Figs. 2-21 and Tables 4-5 showed docking of the most active ligands to 5KEZ: PDB. Compounds 3b, 2a, 3a, 1a, 1c, 1b, and 5 were shown In vitro to exhibit strong inhibition of this α-glucosidase enzyme. Further details are presented in Figs. S1−S16 in the supporting information. As can be seen from Table 4, the docking of conformation ligands with the target enzyme were ranked as follows: 5 > 3b > 4b > 4a > 1c > 1b > 1a > 1 > 2a > 3a > 2b. This suggested that a lower value of free energy of binding was associated with stronger and more stable docking with the receptor, and therefore a lower IC₅₀ value. Fig. 2 shows conformation ligands 1, 1a-c, 2a-b, 3a-b, 4, 4a-b, 5 and acarbose docked to the same pocket enzyme in the crystal structure of α-glucosidase enzyme 5KEZ: PDB. Fig. 3 shows the docking poses of ligands including pose 90 (the most stable conformation ligand of entry 1), pose 7 (1a), pose 56 (1b), pose 88 (1c), pose 150 (2a), pose 196 (2b), pose 104 (3a), pose 65 (3b), pose 164 (4a), pose 15 (4b), pose 13 (5), and pose 164 (acarbose). The RMSD values shown in Table 5 confirmed the validity of the model. The poses of the most stable conformation ligands were ranked as follows: 3b > 1c > 2a > 1b > 1a > 3a.

Close

Fig. 2

The docking poses of ligands, 1a-1c, 2a-2b, 3a-3b, 4a-4b, 5 and Acarbose docked to the same cavity in a pocket enzyme of receptor, 5KEZ. Pose 65 or ligand 3b was reference ligand.

Export to PPT

Close

Fig. 3

Ranked docking poses of ligands 1, 1a-1c, 2a-2b, 3a-3b, 4a-4b, 5 and acarbose aligned and calculated the values of RMSD by PyMOL software.

Export to PPT

Close

Fig. 4

The most stable conformation ligand, 3b (ranked pose 65) was immersed in receptor, one enzyme 5KEZ: PDB.

Export to PPT

Close

Fig. 5

The most stable conformation of ligand 3b formed two hydrogen bonds with active sites of receptor, 5KEZ.

Export to PPT

Close

Fig. 6

The interactions between active sites of receptor (5KEZ) and the active site atoms on ligand (3b) were indicated like hydrogen bonds (light green color), Val der Waals, pi-cation, pi-anion, pi-donor hydrogen bond, pi-pi T-shaped, alkyl and pi-alkyl on a 2D diagram.

Export to PPT

Close

Fig. 7

The ligand map showed secondary interactions between the most stable conformation 3b and receptor, 5KEZ.

Export to PPT

Close

Fig. 8

The most stable conformation, ranked pose 150/200, 2a anchored in receptor (5KEZ: PDB), one crystal structure of α-glucosidase enzyme.

Export to PPT

Close

Fig. 9

Five hydrogen bonds formed from residual amino acids of receptor, 5KEZ to active sites in ligand, 2a.

Export to PPT

Close

Fig. 10

The interactions between the best conformation of 2a, pose 150 and receptor, 5KEZ indicated on a 2D diagram, which included hydrogen bonds, Van der Waals, unfavorable acceptor–acceptor, pi-donor hydrogen bond, alkyl, and pi-alkyl.

Export to PPT

Close

Fig. 11

The hydrogen bonds linked from residual amino acids of target protein of receptor, 5KEZ to active atoms in a docking pose 104, one conformation of the ligand 3a.

Export to PPT

Close

Fig. 12

The interactions between the best conformation, pose 104 of ligand 3a and receptor 5KEZ indicated on a 2D diagram.

Export to PPT

Close

Fig. 13

The hydrogen bond bound from residual amino acids of protein of enzyme, 5KEZ to pose 7/200, ranked pose of ligand 1a.

Export to PPT

Close

Fig. 14

The significant interactions established on 2D diagram between residual amino acids of receptor, 5KEZ and active sites of pose 7, a ranked pose 7/200 of ligand 1a.

Export to PPT

Close

Fig. 15

The hydrogen bonds formed from residual amino acids of receptor 5KEZ to active site binding atoms on the most stable conformation ligand 1c, pose 88.

Export to PPT

Close

Fig. 16

The 2D diagram presented the remarkable interactions between pose 88, ligand 1c and receptor 5KEZ.

Export to PPT

Close

Fig. 17

The hydrogen conducted from the conformation ligand 1b, pose 56 to amino acids of protein of enzyme, 5KEZ.

Export to PPT

Close

Fig. 18

The 2D diagram indicated the essential interactions from residual amino acids of receptor, 5KEZ to active sites on the docking pose 88, ligand 1b.

Export to PPT

Close

Fig. 19

Ranked pose 13, the best stable conformation of ligand 5 docked to crystal structure of alpha-glucosidase, 5KEZ.

Export to PPT

Close

Fig. 20

The active amino acids of enzyme 5KEZ bound to active atoms in pose 13 of ligand 5.

Export to PPT

Close

Fig. 21

The 2D diagram showed the essential interactions between pose 13 of ligand 5 and receptor 5KEZ.

Export to PPT

Ranked pose 65, ligand 3b: Figs. 4-7 and Table 4 show the docking profile of 3b. Pose 65 gave the most stable docking with receptor 5KEZ, with free energy of binding of −9,79 kcal.mol⁻¹ and inhibitor constant of 0.067 µM. It was also strongly hydrophilic (shown by the blue range). Pose 65 therefore showed strong thermodynamic bonding to 5KEZ. As seen in Fig. 5 and Table 4, three hydrogen bonds formed from residual amino acids of the A chain of enzyme 5KEZ: Asp300, His201, and Ala 307. These bound to N and O atoms. As seen in Fig. 6, significant ligand interactions between pose 65 and 5KEZ included hydrogen bonds, pi-cation, alkyl, pi-anion, pi-pi T shaped, pi-donor hydrogen, carbon-hydrogen bonds, and Van der Waals forces. The structure had three parts: a cap group (protein identification), a linker (aromatic or aliphatic chains), and functional groups that interacted with the structure enzyme. The cap group (protein identifications) formed interactions including pi-pi T-shaped ligands from the Ile 235: A chain to the aromatic ring, alkyl or pi alkyl ligands from the Ala307: A chain and Cys5: B chain, pi-cation or pi-anion ligands from the Glu233: A chain and Pro2: B chain to aromatics, from the Tyr3: B chain to the aromatic chain via carbon-hydrogen bonding, and from the Gly308: A chain to the aromatic chain via pi-donor hydrogen bonding. The link from Leu162, Ala198 to the methyl group was via pi-pi T shaped ligands, from the Pro2: B chain to the methyl group via pi-cation or pi-anion interactions, from the Pro2: B chain to the oxygen heterocyclic (pi-cation or pi-anion interactions), from the Ile235: A chain with oxygen heterocyclic via pi-pi-T-shaped ligands, from the Ile235: A chain to the nitrogen heterocyclic (pi-pi-T shaped), and from the Ala307: A chain to the nitrogen heterocyclic via alkyl or pi-alkyl interactions. The functional group at the identification part of the pose confirmed the hydrogen bonds, which formed from amino acids Asp300 and His201: A chain to the hydrogen atoms of phenolic hydroxyl group, hydrogen of an amine group, and hydroxy benzyl of the pose. Ligand 3b or pose 65 is the best candidate among ligands to delivery. The ligand map in Fig. 7, indicated the secondary interactions like hydrogen bonds (brown color), steric (light blue), and overlap (violet circles on atoms of ligand 3b) interactions between pose 65 and residual amino acids (distance of less than 4 Å) of the receptor, 5KEZ. Those green lines presented the steric effects to the strength of interactions of the stable pose 65 to the receptor, 5KEZ. The size of violet circles on atoms controlled the strength of overlap interactions and contributed to steric hindrance.

Ranked pose 150, ligand 2a: the interaction profile of pose 150 and target enzyme showed in Figs. 8-10 and Table 4. As seen in Fig. 8, pose 150 anchored to a pocket enzyme with the values of ΔG^o and Ki of −6.90 Kcal.mol⁻¹ and 8.75 µM, respectively. The values of IC₅₀ in vitro and K_i in silico are equivalent. The hydrophilic ability of this pose is also indicated by eight hydrogen bonds, which are linked from active amino acids of A or B chain as seen in Table 4 and Fig. 9. Among hydrogen bonds, 2a:H–A:ASP300:O was the strongest due to its shortest bond length of 1.68 Å. The significant interactions between this pose and 5KEZ indicated in the 2D diagram in Fig. 10. The cap group of pose recognized via Pro 2: B chain with an aromatic ring- pi cation, an electrostatic interaction, Tyr151: A chain with an aromatic ring via pi-pi- stacked. The linker interaction linked from Tyr151: A chain to the methyl group of pose via a pi-sigma ligand interaction. The functional group of pose formed from His305: A chain, Cys5: B chain, Asp300: A chain, Tyr3: B chain, and Tyr151: A chain to hydro atom of alcohol group, oxygen atom of ether group, hydrogen atom of phenolic hydroxy, oxygen of ketone group, and oxygen atom of lactone ring. It indicated that pose 150, a conformation ligand 2a interacted well with target enzyme.

Ranked pose 104, ligand 3a: the ranked pose 104/200, the most stable conformation ligand docked to target enzyme with values of the free Gibbs energy and inhibition constant of − 6.45 Kcal.mol⁻¹ and 18.80 µM. As shown in Table 4 and Fig. 11, pose 104 formed five hydrogen bonds from active amino acids such as Lys200, His201, Glu240: A chain, and Pro2, and Cys5: B chain to oxygen and hydrogen atoms in pose 104. Among hydrogen bonds, one hydrogen bond, 3a:H − A:GLU240:O is the strongest due to its shortest bond length of 1.94 Å. Those hydrogen bonds determined the hydrophilic ability of pose because they were classified by electrostatic interactions. The important interactions between this pose and the crystal structure of the enzyme exposed in the 2D diagram as seen in Fig. 12. Pose 104 considered strong interaction with target enzyme because three parts of pose 104 identified full ligand interactions. The cap group detected by Ala307, Ala198, Ile 235, Lys200 of A chain via pi-alkyl ligand interactions, Tyr151: A chain via pi-pi stacked, Glu233: A chain via pi-cation, an electrostatic, and Cys5: B chain via pi sulfur interaction. The linker was identified by Tyr151 via pi-alkyl interaction. The functional groups in this pose were recognized by hydrogen bonds which bound from Pro2: B chain, His 201: A chain, and Glu240: A chain to the oxygen atom of carboxylic group, the hydrogen atom of an amine, and the hydrogen atoms of phenolic hydroxy group in pose 104. Other residual interactions were weak Van der Waals interactions, which linked from this pose to wall of the cell via residual amino acid around poses such as His305, Gly306, Asp300, Leu162, Val234, and Ser199 of A chain.

Ranked pose 7, ligand 1a: the interaction profile between ranked pose 7, the most stable conformation ligand was selected and docked to receptor after performing docking as shown in Figs. 13-14 and Table 4. The values of ΔG^o and Ki of binding between the best docking pose 7 and receptor 5KEZ conducted −7.30 Kcal.mol⁻¹ and 4.48 µM, respectively as shown in Table 4. As seen in Fig. 13 and Table 4, pose 7 and receptor, 5KEZ showed four hydrogen bonds from active amino acids such as Ala307, Glu233, Glu240: A chain of the pocket enzyme to active atoms in the pose. They are hydrophilic interactions because the hydrogen bonds were electrostatic interactions. The essential interactions between pose 7 and enzyme were indicated in the 2D diagram in Fig. 15. Pose 7 interacted well with enzyme due to three parts of pose 7, which had strong ligand interactions with enzyme, 5KEZ. The cap group of pose 7 recognized by Ile235: A chain and via pi-alkyl or alkyl, and Pro2: B chain via pi-cation (electrostatic interaction), the aromatic identifications of pose 7. The linker part of pose discovered by Tyr151: A chain, His201: A chain, Leu162: A chain, Pro2: B chain, Tyr3; B chain, Ala198: A chain via alkyl or pi-alkyl, and Asp300: B chain via carbon-hydrogen bond. The functional groups of this pose were identified via hydrogen bonds, which linked from Glu240, Lys200, and Glu233: A chain to hydrogen atom of a phenolic hydroxy group, bromine atom of the phenyl ring, and hydrogen atom of a phenolic hydroxy group, respectively. In other sites, the bromine atoms which linked to phenyl rings indicated more ligand interactions with receptors.

Ranked pose 88, ligand 1c: ranked pose 88/200, the best stable conformation ligand of ligand 1c among 200 conformations, which were docked to the enzyme, 5KEZ had the values of free Gibb’s energy and inhibition constant of −8.03 Kcal.mol⁻¹ and 1.03 µ, respectively. As seen in Fig. 15 and Table 4, active atoms in pose 88 and pocket enzyme performed 5 hydrogen bonds, electrostatic interactions, and hydrophilic interactions. The essential ligand interactions between this pose and the crystal structure of an enzyme were indicated in the 2D diagram in Fig. 16. Generally, this pose interacted well with enzyme because it has full ligand interactions at three parts of this pose. The protein identification of ligand or cap group detected aromatic rings of ligand via category interactions such as alkyl or pi-alkyl from Ile235: A chain to aromatic rings in the structure of this pose, pi-cation, an electrostatic interaction, or hydrophilic interaction (yellow lines) from Pro2: B chain to the aromatic ring. The linker part in this pose has been identified via alkyl or pi-alkyl (pink color lines) from Ile235, Ile237, Ala307, His201, Leu162, Ala198 of A chain, and Tyr3, Pro2 of B chain to bromine atoms and methyl group in the pose. The functional group of this pose was classified by hydrogen bonds linked from Ala307, Tyr151, Glu233, and Glu240 to aldehyde, carbonyl, phenolic hydroxyl group, and alcohol benzyl group in pose (green lines). Another special site, this pose has fully characteristic identifications via cap group: aromatic ring, oxygen heterocyclic of 7 members, linker part: (bromine atoms or methyl), and functional groups. Bromine atoms in this pose made more ligand interactions.

Ranked pose 56 or ligand 1b: Ranked pose 56 docked to the pocket enzyme with the calculated values of ΔG^o and K_i of −7.81 Kcal.mol⁻¹ and 1.9 µM, respectively as shown in Table 4. As seen in Table 4 and Fig. 17, pose 56 formed 5 hydrogen bonds with target enzyme, 5KEZ. They are electrostatic interactions, which proved more hydrophilic ligand. The hydrogen bond, 1b:H–A: His305: O was the strongest among them owing to its shortest bond length of 1.98 Å. As shown in Fig. 18, in the 2D diagram, pose 56 of ligand 1b did not interacted ligand as well as pose 65 (ligand 3b) because it did not identify fully the characteristic conformation ligand for instance without ligand interactions with oxygen heterocyclic of seven members, lactone ring. The cap group of this pose identified the aromatic ring via ligand interaction styles such as pi-cation (electrostatic interaction) via Pro2: B chain, alkyl, or pi-alkyl via Tyr3: B chain, Ile235: A chain. The linker part of this pose detected bromo atoms by pi-cation style via Pro 2B with bromo atom, pi-alkyl, or alkyl interactions via Tyr3: B chain, Ala198: A chain, and His201: A chain. The functional groups have been detected by hydrogen bonds via His201, His 305: A chain to hydrogen atoms of phenyl rings. Pose 56 of ligand 1b was the potential drug delivery due to full identification ligand.

Ranked pose 13, ligand 5: The interaction profile of pose 13, the best stable conformation of ligand 5 was shown in Figs. 19-20. In Fig. 19, the best docking pose ligand 5 or pose 13 docked to the same cavity of enzyme 5KEZ as well as other poses with the values of ΔG^o and Ki of −9.24 Kcal.mol⁻¹ and 0.17 µM, respectively. It made clear that this pose interacted with the receptor, 5KEZ in the thermodynamic site. As shown in Fig. 20 and Table 4, five hydrogen bonds linked from active amino acids of A chain of protein of enzyme 5KEZ to active atoms in this pose. Among them, the hydrogen bond, 5:H - A: Asp402: O was the strongest due to its shortest bond length of 2.18 Å. As seen in Fig. 20, the significant interactions between this pose and 5KEZ indicated in the 2D diagram such as hydrogen bond (hydrophobic interaction) pi-cation (electrostatic interaction), pi-alkyl, carbon-hydrogen bond, Van der Waals, and unfavorable donor–donor. In this case, unfavorable interaction (red color), which formed from Gly9: A chain to hydrogen atom of alcohol group of 1,3-dienyl cyclohexyl ring of pose 13, made the system of ligand interactions weaker. The system energy of pose 13 and 5KEZ increased and the complex, which formed from pose 13 to receptor, 5KEZ was unstable. Therefore, pose 13 or ligand 5 did not interact well with enzyme.

Compared docking model and ligand interactions

Ligand piHA-Dm was a type of structure of peptide, composing five amino acids linked to a head-to-side-chain thioether macrocycle. piHA-Dm exposed selected inhibitor of novel typical peptide and occupied in Human pancreatic alpha-amylase (HPA). The HPA/piHA-Dm complex have deposited under the code PDB:5KEZ (target enzyme in silico docking.) It inhibited enzyme 5KEZ with the value of inhibition constant of 7.0 ± 3.5 nM (Jongkees et al., 2017). Ligand 3b or pose 65 was the best ligand, which inhibited enzyme 5KEZ with the value of inhibition constant of 0.067 μM or 67 nM, which indicated ligand 3b inhibited around ten-fold lower than ligand piHA-Dm. Insightful analysis of ligand interactions between ligand piHA-Dm or ligand 3b (pose 65) and crystal structure of enzyme 5KEZ illustrated that ligand piHA-Dm interacted well with five residual amino acids such as Asp197 and 300, His101 and 201, and Glu233 in enzyme 5KEZ. While as shown in the 2D diagram of Fig. 6, ligand 3b (pose 65) interacted well with 5KEZ via Asp300 of A chain: two hydrogen bonds and His 201 of A chain: one hydrogen bond, Glu233: A chain: pi-charge interaction. Other ligand interactions that make ligand 3b to enzyme 5KEZ different from ligand piHA-Dm to the same enzyme included alkyl or pi-alkyl interactions as Leu162, Ala198, Ile235, Ala307: A chain, and Cys 5: B chain, and pi-charge infractions from Pro 2: B chain to ligand 3b as show in Fig. 6. As indicated in Fig. 7, one ligand map presented hydrogen bonds from His 201 and Asp300: A chain (brown color lines) to active atoms in ligand 3b (pose 65). Many steric interactions (green lines) bound from His201, Asp300, Glu233, Tyr151, Gly306, Ala307, His305, Ile235, Cys50, Ser4, and Pro2 of enzyme to ligand 3b, which indeed proved that ligand 3b and target enzyme, 5KEZ interacted strongly through docking processing. At the interaction site, ligand 3b interacted with the same enzyme more strongly than that of ligand piHA-Dm.

The validation of molecular docking model

Ranked docking poses of ligands and Acarbose docked the same cavity of enzyme pocket, 5KEZ (Fig. 2). Pose 65 or ligand 3b, the best docking ligand was selected as reference ligand. Other ligands and ligand Acarbose aligned to pose 65 and calculated the values of RMSD by PyMOL software as illustrated in Fig. 3. The calculation results of redocking can be found in Fig. 3 and Table 5. As shown in Fig. 3, docking poses bound to the same cavity of enzyme without being superimposed because they are not analogue structures. As seen in Table 5, the values of RMSD of pair poses changed from 2.290 Å (ligand 3a, ligand 3b) to 4.453 Å (ligand 1b, ligand 3b). Additionally, the value of RMSD between 3a and 3b was lowest due to their analogue structures. The values of RMSD were considered more suitable and they proved that the docking parameters, which were valuation in terms of conformation, interactions, and reproducing orientation, were predicted well in molecular docking model of interesting compounds (Hidalgo-Figueroa el al., 2021).

From these in vitro data, compounds 4a and 5 were identified as the most active. To elucidate the inhibition mechanism, we investigated activity levels at the concentrations of 0, 4.5, 9.0, 18.0, and 36.0 µM for 4a and of 0, 3.16, 6.33, and 12.65 µM for 5. Lineweaver–Burk plots of 4a showed linearity at each concentration, intersecting the y-axis in the first quadrant (Fig. 22). The kinetics of enzyme inhibition confirmed that 4a acted as a non-competitive inhibitor. The kinetic analysis showed that an increase in the concentration of 4a had no effect on the K_m (1.20), but decreased the v_max from 0.07 to 0.03 µM.min⁻¹. The inhibition constant (K_i) was 28.44 ± 2.50 µM. For compound 5, Lineweaver–Burk plots gave a group of lines with different slopes and intercepts, intersecting the y-axis in the second quadrant (Fig. 23A). The kinetics of enzyme inhibition indicated that 5 acted as a mixed-type inhibitor. The inhibition constants of 5 binding with the free enzyme (K_i) and with the enzyme-substrate complex (K’i) were determined to be 8.22 ± 0.32 μM and 20.32 ± 0.51 μM, respectively (Fig. 23B and Fig. 23C). The value of K_i was lower than that of K’_i, indicating that the binding affinity of α-glucosidase-5 exceeded that of α-glucosidase-PNPG-5 complex.

Close

Fig. 22

Lineweaver-Burk plot for alpha-glucosidase inhibition by compound 4a (A) and the secondary plot of slope vs the inhibitor concentration (B).

Export to PPT

Close

Fig. 23

Lineweaver-Burk plot (A) for α-glucosidase inhibition by 5 and the secondary plots of slope and Y-intercept vs inhibitors concentration (B and C)

Export to PPT