4.3. Isolation Procedure

After drying, the sample materials (5 g) were pulverised and extracted with methanol (50 mL) under two cycles of refluxing. The extract was then filtered, and the filtrate was evaporated to yield a syrup (1.16 g). A portion of the extract (500 mg) was suspended in water (50 mL) and then partitioned with n-hexane to remove the lipids. The aqueous layer was then further partitioned with ethyl acetate to extract low-polarity compounds, and the resulting aqueous layer was partitioned with n-butanol to extract glycoside compounds such as saponins. Finally, the remaining aqueous layer contained polar compounds, such as sugars and tannins. Subsequently, the n-butanol layer was evaporated, and the residue was subjected to column chromatography on silica gel (methanol-chloroform). The fractions (eluted with 20–50% methanol/chloroform) were combined and re-chromatographed using HPLC (column: Imtakt UK C-18, 20 mm × 250 mm, solvent: 80% methanol/water, flow rate: 8 mL/min; Portland, OR, USA) to obtain compounds 1 (27 mg) and 2 (6 mg).

The chemical structures of 1 and 2 were determined by spectroscopic techniques, including 2D-NMR (COSY, HMQC, and HMBC) and LC/MS. Using high-resolution MS, 1 and 2 were formulated as C₆₁H₉₆O₂₇. The ¹³C-NMR spectra of compounds 1 and 2 were similar. Four moles of hexose signals and triterpene moiety signals were observed; moreover, the existence of tigloyl and acetyl groups was deduced. Therefore, compounds 1 and 2 were supposed to be analogs of jegosaponin. However, there were large differences in signals around C-21, 22, and 28 between 1 and 2; 1 and 2 were deduced to be isomers of tigloyl and acetyl groups. The positions of the tigloyl and acetyl groups were determined using the HMBC spectrum. Finally, 1 and 2 were identified as jegosaponin A and B by comparison with the literature data [17]. Details of these compounds are described in Supplementary Materials.