Hepatoprotective Triterpene Saponins from the Roots of Glycyrrhiza inflata

Two novel oleanane-type triterpene saponins, licorice-saponin P2 (1) and licorice-saponin Q2 (3), together with nine known compounds 2, 4–11, have been isolated from the water extract of the roots of Glycyrrhiza inflata. The structures of these compounds were elucidated on the basis of spectroscopic analysis, including 2D-NMR experiments (1H–1H COSY, HSQC, HMBC and ROESY). In in vitro assays, compounds 2–4, 6 and 11 showed significant hepatoprotective activities by lowering the ALT and AST levels in primary rat hepatocytes injured by D-galactosamine (D-GalN). In addition, compounds 2–4, 6, 7 and 11 were found to inhibit the activity of PLA2 with IC50 values of 6.9 μM, 3.6 μM, 16.9 μM, 27.1 μM, 32.2 μM and 9.3 μM, respectively, which might be involved in the regulation of the hepatoprotective activities observed.


Introduction
The genus Glycyrrhiza consists of about 30 species with a nearly global distribution, of which 18 species are found in China. Among them, three species named Glycyrrhiza uralensis, Glycyrrhiza glabra and Glycyrrhiza inflata, have been used as traditional Chinese medicine for the treatment of hepatitis, spasmodic cough, gastric ulcer, and so on. Phytochemical studies have showed that triterpenoid saponins and flavonoids were the two of major kinds of active substances of Glycyrrhiza, which have a variety of pharmacological activities, including hepatoprotective [1,2], antiviral [3], anti-inflammatory [4] and antioxidative [5] effects. Recently, we reported the chemical constituents of G. uralensis and G. glabra, as well as their cytotoxic or neuraminidase bioactivities [6,7]. As part of our ongoing research on the genus Glycyrrhiza, an extensive phytochemical investigation on the roots of G. inflata has now led to the isolation of two new oleanane-type saponins 1, 3 and nine known saponins 2, 4-11. All compounds were screened for their protective activities against D-galactosamine (D-GlaN) induced toxicity in vitro. In addition, the inhibitory activities on phospholipase A2 (PLA2) were presented. Herein, we report the isolation and structural elucidation of these saponins, along with the investigation of their protective activities.

Results and Discussion
The total saponin fraction of G. inflata was prepared by co-application of polyamide and macroporous resin column chromatography [7]. The resulting extract was subjected to ODS column chromatography and preparative HPLC to afford two new oleanane-type saponins 1, 3 together with nine known ones 2, 4-11. Their structures were shown in Figure 1.

Structural Determination
Compound 1 was obtained as a white amorphous powder and showed a protonated peaks in the low-resolution positive HR-ESI-MS spectrum at m/z 861. 3929  , five methines (including one oxygenated methine and one unsaturated methine), and nine quaternary carbons (including one carbonyl quaternary carbon, one unsaturated quaternary carbon and one carboxyl carbon). Therefore, compound 1 was considered to be an oleanane-type triterpene glucuronide bearing a 12(13)-double bond and a keto group at C-11. In the HMBC spectrum, correlations of δH 5.03 (H-1′) to δC 91.3 (C-3) and δH 5.38 (H-1′′) to δC 85.5 (C-2′) could be observed. In addition, the correlations in the HMBC spectrum from H-1′ at δH 5.03, H-23 at δH 1.27 and H-24 at δH 1.08 to C-3 at δC 91.3 helped in assigning one oxygenated methine at C-3. Detailed analysis of the above 1D-NMR data and 2D-NMR correlations indicated that 1 is an oleanane-type saponin derivative and is structurally related to the known compound licorice-saponin G2 (4). The comparison of the NMR data of 1 with those of 4 suggested that the hydroxyl group at C-24 in 4 was transposed to C-29 in 1. The HMBC correlations from δH 3.98, 4.06 (H-29) to δC 39.1 (C-19) and δC 180.2 (C-30) and the 1 H-1 H COSY correlations between the proton signal at δH 2.49 (H-18) and δH 1.98, 2.24 (H-19) confirmed that hydroxyl group was connected to C-29 in compound 1 ( Figure 2).

Hepatoprotective Activity
All the separated compounds were assessed for their hepatoprotective activities against the increase of AST and LDH levels in primary rat hepatocytes injured by D-GalN. The maximum nontoxic concentrations of tested compounds on primary rat hepatocytes were in the range of 120-240 μM. A set of cells in culture medium treated with D-GalN was used as the model group, and in comparison to the model group, macedonoside A (2), licorice-saponin Q2 (3), licorice-saponin G2 (4), 22β-acetoxy-glycyrrhizin (6) and glycyrrhizin (11) notably lowered AST (10.3-16.5 U·L −1 ) and LDH (200.7-242.8 U·L −1 ) in the range of concentration 30-120 μM. (Table 2). Comparing the activities of these saponins, compound 5 and 7 was shown to have significantly weaker hepatoprotective activities than the compound 2 and 11 owing to presence of a lactone ring at position 22(30). Compound 11 showed stronger activity than 1. That might be because an additional CH2OH group is preferable to improve the steric hindrance, thus resulting in a decrease in the bonding capacity with active targets. Interestingly, compound 3 displayed higher activity than compound 4. The reason might be that compound 3 with a 18α-H group was found to be favorable for the anti-liver injury activity. On the basis of the above analysis, it seemed that a carboxyl residue at position 29 or 30 was possibly the necessary group for hepatoprotective activity.

Enzyme Inhibition Activity
As a regulator associated with the stability of the liver cell membrane, phospholipase A2 (PLA2) is a promising target for hepatoprotective drug development [16]. To examine whether the compounds inhibit activities on PLA2, the enzyme inhibitory potency of all isolated compounds was conducted and the results were summarized in Table 3. Among these, two saponins (compounds 2 and 3) and glycyrrhizin (11) exhibited efficient inhibitory activity with IC50 value of 6.9 μM, 3.6 μM and 9.3 μM, respectively. Compounds 4, 6 and 7 showed moderate inhibitory activities with IC50 values of 16.9 μM, 27.1 μM and 32.2 μM, respectively.
What was noteworthy, is that analysis of the two assays of 1-11 showed that there was good relationship between PLA2 inhibitory activities and hepatoprotective effects, leading to the hypothesis that inhibition of PLA2 was one of the possible mechanisms of the hepatoprotective effect of licorice saponins. Table 3. Inhibitory activities of isolated saponins on PLA2.

Material
The roots of Glycyrrhiza inflata were collected in Weli County, Xinjiang Uygur Autonomous Region, China, October 2013. A voucher sample (No. 20131015) was preserved in Nanjing University of Chinese Medicine, and identified by Prof. Qi-Nan Wu.

Extraction and Isolation
The roots of G. inflata (dry weight, 25 kg) were exhaustively extracted two times with boiling water (

Acid Hydrolysis
The configuration of the sugars of compounds 1 and 3 was determined by acid hydrolysis and GC experiments based of the literature procedure [6,9]. The specific steps were as follows: a solution of compounds 1-3 (1.0 mg each) in 1 N HCl (1 mL) was stirred at 90 °C for 2 h. After cooling, the solution was evaporated under a stream of N2. Anhydrous pyridine solutions (0.1 mL) of each residue and L-cysteine methyl ester hydrochloride (0.06 N) were mixed and warmed at 60 °C for 1 h. The trimethylsilylation reagent trimethylsilylimidazole (0.15 mL) was added, followed by warming at 60 °C for another 30 min. After drying the solution, the residue was partitioned between H2O and CH2Cl2 (1 mL, 1:1 v/v). The CH2Cl2 layer was analyzed by GC/MS. The peaks of authentic sample of D-glucuronic acid after treatment in the same way were detected at 14.23 min. The final result was to compare the retention times of monosaccharide derivatives with standard sample. The absolute configuration of sugar was confirmed to be D-glucuronic acid (D-glucuronic acid for compound 1 with retention time 14.21 min; D-glucuronic acid for compound 3 with retention time 14.22 min).

Cell Assay
Isolated rat hepatocytes were prepared from male Wistar rats by a collagenase perfusion technique as described previously [17]. The D-GalN concentration used for cell culture treatment was previously determined according to a modification of the method of Morikawa et al. [18]. The cultured cells in logarithmic growth phase were made into a single-cell suspension and seeded in 96-well plates (1 × 10 4 cells/well) in the DMEM/F 12 with 2% FBS complete medium for 24 h at 37 °C. Then, the hepatocytes were exposed to 2 mM D-GalN for 2 h to induce hepatotoxocity. The medium with silibin meglumine (as positive drug, purity 95.6%, Hunan Xieli Pharmaceutical Co., Ltd., Zhuzhou, China) and different concentrations of test compounds was mixed in cell medium (final test compounds concentration were 30 μM, 60 μM and 120 μM, respectively), and incubated for 24 h. The obtained reacted supernatant was directly used to detect ALT and AST levels. The control group was a set of cells maintained in culture medium, while the model group was a set of cells maintained in culture medium and treated only with D-GalN. All data are expressed as the mean ± SD of at least three independent experiments as indicated. The test for the paired samples was used to determine statistical difference between parameters. These differences were considered significant for p < 0.05 or 0.01.

Assay for Inhibition against PLA2
The PLA2 inhibitory assays of compounds 1-11 and the positive drug diethylenetriaminepentaacetic acid (Purchased from Aladdin, Los Angeles, CA, USA, purity > 98.0%) were carried out according to the literature [19]. First of all, each tube was added with 1 mL fresh substrate buffer solution (pH = 8.2).
After that, 50 μL tested compounds at various concentrations were placed at reaction tube and blank tube, respectively. As for control tube, 50 μL deionized water was instead. Then each tube incubated at 40 °C for 10 minutes. The reaction tube and blank tube were followed by the treatment with PLA2 enzyme (5 μL) at the concentration of 5 μg/mL. Before put them into the incubator at the temperature of 40 °C to react 30 minutes, the content of the tube should be fully blending. The optical density value of each tube was then read in an ELISA plate reader using a wavelength of 495 nm. The IC50 values were