Nephelium lappaceum L., belonging to the genus Nephelium and thus the Sapindaceae family, is widely distributed in Southeast Asia’s tropical regions, including Vietnam [1]. The fruit of this species, commonly known as “Rambutan”, is recommended for severe dysentery, as an astringent, an antifebrile, and a warm carminative in dyspepsia [2]. When rambutan fruit is processed, the fruits are deseeded first, and the seeds become a waste by-product [3, 4]. Interestingly, rambutan seeds have been reported to have narcotic effects, and roasted rambutan seeds are consumed in the Philippines [5]. Previous studies on the seeds of N. lappaceum reported the presence of alkaloids, phenolic compounds, glycosides, carbohydrates, and saponins that are responsible for these antioxidant and antibacterial activities [1, 6, 7]. As part of our ongoing search for new glycosides from natural sources, we report the isolation of a new oleanane-type triterpene glycoside from the seeds of N. lappaceum, and the spectral evidence leading to elucidating the structure was carried out.

Compound 1 was isolated as an amorphous powder, exhibiting a molecular ion (positive-ion HR-ESI-MS) at m/z 1315.6298 [M + Na]+ corresponding to the molecular formula C62H100O28. The 1H NMR spectrum of 1 displayed signals of seven tertiary methyl proton signals at δ 1.26 (s, H3-23), 1.13 (s, H3-24), 0.86 (s, H3-25), 0.97 (s, H3-26), 1.30 (s, H3-27), 0.95 (s, H3-29), 1.01 (s, H3-30), an olefinic proton at δ 5.46 (br.t, J = 3.5 Hz, H-12), corresponding to 13C NMR data of seven tertiary methyl carbon signals at δ 28.1 (C-23), 17.2 (C-24), 15.6 (C-25), 17.5 (C-26), 26.1 (C-27), 33.4 (C-29) and 23.6 (C-30), two olefinic carbon signals at δ 122.2 (C-12) and 144.6 (C-13), which were typical signals of the oleanolic acid skeleton [8]. The downfield chemical shift at δ 88.6 (C-3) and the upfield chemical shift at δ 178.1 (C-28) in the 13C NMR spectrum of 1 (Table 1) indicated that 1 was a 3,28-bidesmosidic glycoside [9].

TABLE 1. 1H (600 MHz) and 13C (150 MHz) NMR Data of 1 (Py-d5, δ, ppm, J/Hz)

The 1H NMR spectrum of 1 showed six anomeric protons at δ 4.84 (d, J = 7.6 Hz), 4.86 (d, J = 6.0 Hz), 4.88 (d, J = 7.0 Hz), 5.01 (d, J = 7.6 Hz), 6.02 (d, J = 8.2 Hz), and 6.13 (br.s) which showed HSQC correlations with six anomeric carbon signals at δ 103.7, 104.9, 103.8, 105.0, 96.0, and 101.6, respectively, indicating the presence of six sugar units. These sugars were determined to be two L-arabinoses (Ara I, Ara II), two D-xyloses (Xyl I, Xyl II), one D-glucose (Glc), and one L-rhamnose (Rha) by acid hydrolysis and TLC comparison with authentic samples. The β-anomeric configurations for the Xyl and Glc units were deduced from their JH-1, 2 coupling constants ranging from 7.6 to 8.0 Hz, and the Ara units were determined to be α anomeric configurations based on the JH-1, 2 values (6.0–7.0 Hz) observed in the 4C1 forms [10]. The α-configuration of the Rha unit was determined from the broad singlet observed for the anomeric proton and the JC-1, H-1 value of 168 Hz [8]. Assignment for all 1H and 13C NMR signals and determination of the structure were achieved by 2D NMR analyses, mainly in HMBC and ROESY. In the HMBC spectrum, a correlation between δH 4.86 (d, J = 6.0 Hz, Ara I H-1) and δC 88.6 (C-3 aglycone) indicated that Ara I was linked to the C-3 of the aglycone. The linkage of the Rha unit at the C-2 of Ara I was determined by the correlation between δH 6.13 (br.s, Rha H-1) and δC 75.5 (Ara I C-2). Similarly, the linkage of Ara II at C-3 of the Rha was indicated by the correlation between δH 4.88 (d, J = 7.0 Hz, Ara II H-1) and δC 83.1 (Rha C-3). This conclusion was further confirmed by the observation of three cross-peaks in the ROESY spectrum between δH 4.86 (d, J = 6.0 Hz, Ara I H-1) and 3.28 (H-3 aglycone), δH 6.13 (br.s, Rha H-1) and 4.53 (Ara I H-2), δH 4.88 (d, J = 7.0 Hz, Ara II H-1) and δH 4.67 (dd, J = 9.6, 2.1 Hz, Rha H-3). Thus, the glycosidic sequence linked to the C-3 of the aglycone was established as α-L-arabinopyranosyl-(1→3)-α-L-rhamnopyranosyl-(1→2)-α-L-arabinopyranosyl. The three remaining Xyl I, Xyl II, and Glc were identified as β-D-xylopyranosyl-(1→3)-β-D-xylopyranosyl-(1→2)-β-D-glucopyranosyl sequence linked to C-28 of the aglycone. This assumption was verified according to the chemical shift at δC 96.0 (Glc C-1), a typical value suggesting an ester linkage with C-28 of the aglycone. The linkage was confirmed by the observation of three cross-peaks in the HMBC spectrum between δH 6.02 (d, J = 8.0 Hz, Glc H-1) and δC 178.1 (C-28 aglycone), δH 5.01 (d, J = 7.6 Hz, Xyl I H-1) and δC 75.8 (Glc C-2), δH 4.84 (d, J = 7.6 Hz) and δC 75.7 (Xyl I C-3), and two cross-peaks in the ROESY spectrum between δH 5.01 (d, J = 7.6 Hz, Xyl I H-1) and δC 4.28 (Glc H-2), δH 4.84 (d, J = 7.6 Hz) and δC 4.07 (Xyl I H-3) (Fig. 1). The structure of 1 was thus characterised as 3-O-α-L-arabinopyranosyl-(1→3)-α-L-rhamnopyranosyl-(1→2)-α-Larabinopyranosyloleanolic acid 28-O-β-D-xylopyranosyl-(1→3)-β-D-xylopyranosyl-(1→2)-β-D-glucopyranosyl ester.

Fig. 1.
figure 1

Key HMBC and ROESY correlations of compound 1.

Experimental

General Experimental Procedures. Optical rotation was measured with an AA-10R automatic polarimeter (Optical Activity LTD, Ramsey, UK). NMR spectra were measured on a Varian VNMRS 600 MHz spectrometer (Agilent Technologies, Santa Clara, California, USA). HR-ESI-MS spectrum was recorded on Bruker micrOTOF II mass spectrometer (Bruker, Mannheim, Germany). Separation and purification were performed by vacuum liquid chromatography (VLC) on RP-18 silica gel (75–200 μm, Silicycle, Quebec, Canada) and NP-60 silica gel 60 (60–200 μm, Merck, Germany), column chromatography (CC) on silica gel 60 (15–40 μm, Merck, Germany) and Sephadex LH-20 (GE Healthcare Bio-Sciences AB, Merck, Germany). Chemical shifts are given on the δ-scale with pyridine-d5 as the internal standard.

Plant Material. The seeds of N. lappaceum were collected in a fruit boutique in Thai Nguyen City, Vietnam, in 2021 (21°34′01′′ N, 105°48′35′′ E) and identified by one of the authors (Dr. Hung Duc Nguyen). A voucher specimen (No. NELASE1121) was deposited in our lab.

Extraction and Isolation. Dried seeds (156.8 g) of N. lappaceum were successively extracted three times with 500 mL of EtOH–H2O (75%–35%) each time. The aqueous-ethanolic extract was evaporated under reduced pressure to give a thick syrup, then dissolved in a minimum volume of H2O and freeze-dried to yield 8.9 g of crude extract. This was dissolved in a minimum volume of H2O and submitted to a VLC over silica gel RP-18 eluting with each 500 mL of solvents, including EtOH–H2O (0:1, 1:1, 0:1), yielding 3 fractions (NLS.1, NLS.2 and NLS.3). Fraction NLS.2 (617.1 mg) was dissolved in a minimum volume of CHCl3–MeOH–H2O, 75:25:3 and subjected to a VLC over silica gel NP-60 eluted with CHCl3–MeOH– H2O, 75:25:3, 70:30:5, 60:32:7 (300 mL each) to collect three subfractions (NLS.2.1–NLS.2.3). Subfraction NLS.2.2 (89.2 mg) was subjected again on a CC over silica gel NP-60 eluted with CHCl3–MeOH–H2O, 70:30:5 affording five subfractions (NLS.2.2.1–NLS.2.2.5). Subfraction NLS.2.2.2 (6.3 mg) rich in saponin was applied to a CC over Sephadex LH-20 with EtOH 96% to remove pigments to afford 1 (2.9 mg) as a pure compound.

3-O-α-L-Arabinopyranosyl-(13)-α-L-rhamnopyranosyl-(12)-α-L-arabinopyranosyloleanolic acid 28-O-β-D-xylopyranosyl-(13)-β-D-xylopyranosyl-(12)-β-D-glucopyranosyl ester (1), white amorphous powder, [α]25D –22° (c 0.75, MeOH). 1H (600 MHz, Py-d5) and 13C (150 MHz, Py-d5) NMR spectral data are shown in Table 1. HR-ESI-MS m/z 1315.6298 [M + Na]+ (calcd for C62H100NaO28, 1315.6293).

Acid Hydrolysis and GC Analysis. The protocol of identification of sugar moieties and those absolute configurations of isolated compound was detailed as previously referenced [11,12,13].