Metabolomic Application in Identification α-Glucosidase Inhibitor from Vegetables and Spices

Abstract

Diabetes mellitus is a degenerative disease with a high prevalence worldwide. Plant-based foods such as vegetables and spices were often reported to be able to reduce the risk of and even cure diabetes mellitus, especially type 2 (T2DM). Indonesia has many natural resources such as vegetables and spices that can be sources of antidiabetic foods. However, the antidiabetic activity of many vegetables and spices have not been evaluated. Besides, studies to observe the bioactive compounds responsible for the antidiabetic activity of vegetables and spices are still limited. One of the antidiabetic activities commonly studied from plant species is the α-glucosidase inhibition. Therefore, the objective of this study was to evaluate the potency of α-glucosidase inhibition of 11 vegetables and 4 spices commonly consumed in Indonesia and to identify their bioactive compound using a metabolomic approach. This study consisted of 3 stages: 1) Screening α-glucosidase inhibition activity of Indonesian vegetables and spices, (2) characterization of antioxidant compound from selected plant (obtained from the first stage) using UHPLC-HRMS and identification of its α-glucosidase inhibitor using NMR- and HPLC combined with UHPLC-HRMS metabolomics, and (3) verification of α-glucosidase inhibition activity of putative compounds (obtained from the second stage) through in silico analysis. The results of the first study showed that the highest antioxidant and α-glucosidase inhibition activity was found in leaves extract of Syzygium polyanthum followed by Pluchea indica with the values of the IC50 antioxidant activity being 2.46 and 4.34 μg/ml, respectively. The values of their IC50 of α-glucosidase inhibition were 11.76 and 12.17 μg/ml, respectively. Interestingly, the total phenolic content of these two samples was also higher than others with 25.37 and 37.99 μg GAE/mg extract, respectively. The 1H-NMR fingerprinting analysis showed that each sample had different spectral profile, especially at the aromatic region (chemical shift (δ) > 5.7). Further metabolite profiling on the 1H-NMR of S. polyanthum revealed the presence of gallic acid, syringic acid, and myricetin, whereas 5-O-caffeoylquinic acid, 4,5-di-O-caffeoylquinic acid methyl ester, and esculetin were found in the 1H-NMR of P. indica. Previous studies reported that all identified compounds had antidiabetic activity especially through α-glucosidase inhibition. Thus, those compounds might be associated with the α-glucosidase inhibition activity of S. polyanthum and P. Indica, which was investigated further in the next step. Based on the results of the first study, S. polyanthum leaves were chosen for the second stage. Dried leaves of S. polyanthum were fractionated using solvents with different polarity sequentially i.e. hexane, combination of hexane-acetone, acetone, combination of acetone-water and water to obtain 14 fractions. Analysis of antioxidant activity through the ability to reduce radical 2,2-diphenyl-1-picrylhydrazyl (DPPH) indicated that at a tested concentration of 500 μg/ml, Fraction 9/F9 (obtained from acetone:water = 4:1) gave the highest antioxidant activity followed by Fraction 1/F1 (obtained from hexane 100%), Fraction 10/F10 (obtained from acetone:water = 3:2) and Fraction 3 (obtained from hexane:acetone = 4:1) with the values of 95.21, 94.82, 93.37 and 93.19%, respectively. The characterization of the UHPLC-HRMS chromatogram of F9 identified the presence of gallic acid, trans-aconitic acid, and myricetin, while only gallic acid and trans-aconitic acid were identified in F10. The UHPLC-HRMS chromatogram of F1 and F3 showed similar profiles. Unfortunately, the bioactive compounds from these fractions could not be identified because of the limited databases. The S. polyanthum fractions were also evaluated for its α-glucosidase inhibition and chemical profile which were analysed using the 1H-NMR and HPLC. The results exhibited that the highest α-glucosidase inhibition was found on F9 followed by F10 with the values of IC50 being 24.8 and 31.8 μg/ml, respectively. Interestingly, the chemical profiles analysed using 1H-NMR and HPLC provided the same patterns whereas the chromatogram of F9 had more diverse and varied signals than other fractions. Principal component analysis (PCA) also strengthens the previous statement by showing that the chemical profile of F9 was different than other fractions. Identification of α-glucosidase inhibitor from S. polyanthum fractions was performed using metabolomic approach. This technique correlates the α-glucosidase inhibition and chemical profile (analysed by 1H-NMR and HPLC) of each fraction using multivariate analysis. This study used orthogonal projection to latent structure (OPLS) as a model of multivariate analysis, which divided chemical profile (X variable) into two blocks: block related to α-glucosidase inhibition (Y variable) as a predictive component and unrelated block as orthogonal component. The validation of 1H-NMR- and HPLC-OPLS-metabolomic indicated that the value of R2Y and Q2 closeness to 1 (0.953 and 0.731 for 1H-NMR-OPLS, and 0.912 and 0.799 for HPLC-OPLS, respectively) reflected the goodness of the models. The score plot both models displayed similar output where F9 and F10 (as more active fractions) were different than others. Moreover, the S-plot of 1H-NMR-OPLS revealed several chemical shifts at 2.6-2.8, 6.6-7.1, and 3.46-4.26, and retention time at 31, 3.5, 3.0, 16.5, 12.5, 12, and 8 min for HPLC-OPLS which were highly correlated with antidiabetic activity. Other outputs of metabolomic, such as X-Variable plot, expressed that the highest abundance at selected signals (obtained from S-plot) was found in F9. Further analysis using 1H-NMR in tandem with 2D-NMR of F9 especially focused on selected signals (obtained from S-plot) identified two putative antidiabetic compounds, which were myricetin-3-O-rhamnoside (myricitrin) and epigallocatechin-3-gallate (EGCG). Interestingly, the presence of these two compounds in F9 was verified by UHPLC-HRMS analysis. The evaluation of α-glucosidase inhibition activity of all identified compounds in F9 (obtained from 1H-NMR and UHPLC-HRMS) were performed in silico through docking analysis. The interaction between ligand (tested compound) and α-glucosidase (PDB code 3TOP) were described by binding energy (Ei) and constant inhibition (Ki) and those values were compared with acarbose as native and positive control ligand. The result showed that the Ei and Ki of acarbose were the lowest with values -10.13 kcal/mol and 0.0375 μM, respectively. All ligands showed the ability to inhibit α-glucosidase, although only myricitrin and EGCG had Ei and Ki close to acarbose with the values of 8.47 kcal/mol and 0.615 μM for myricitrin and 8.19 kcal/mol and 0.991 μM for EGCG. The analysis of docking output using ligplot software expressed the occurrence of hydrogen bond and hydrophobic interaction between ligand and enzyme. Interestingly, myricitrin and EGCG interacted with the amino acid of enzyme similar to acarbose at the catalytic site of the enzyme. Thus, they can be classified as competitive inhibitors for α-glucosidase.