New dammarane-type triterpenoid saponins from Panax notoginseng saponins

Abstract

Background

Panax notoginseng saponin (PNS) is the extraction from the roots and rhizomes of Panaxnotoginseng (Burk.) F. H. Chen. PNS is the main bioactive component of Xuesaitong, Xueshuantong, and other Chinese patent medicines, which are all bestselling prescriptions in China to treat cardiocerebrovascular diseases. Notoginsenoside R₁ and ginsenoside Rg₁, Rd, Re, and Rb₁ are the principal effective constituents of PNS, but a systematic research on the rare saponin compositions has not been conducted.

Objective

The objective of this study was to conduct a systematic chemical study on PNS and establish the HPLC fingerprint of PNS to provide scientific evidence in quality control. In addition, the cytotoxicity of the new compounds was tested.

Methods

Pure saponins from PNS were isolated by means of many chromatographic methods, and their structures were determined by extensive analyses of NMR and HR-ESI-MS studies. The fingerprint was established by HPLC-UV method. The cytotoxicity of the compounds was tested by 3-(4,5-dimethylthiazol-2-yl)-2,5 -diphenyltetrazolium bromide assay.

Results and Conclusion

Three new triterpenoid saponins (1–3) together with 25 known rare saponins (4–28) were isolated from PNS, except for the five main compounds (notoginsenoside R₁ and ginsenoside Rg₁, Rd, Re, and Rb₁). In addition, the HPLC fingerprint of PNS was established, and the peaks of the isolated compounds were marked. The study of chemical constituents and fingerprint was useful for the quality control of PNS. The study on antitumor activities showed that new Compound 2 exhibited significant inhibitory activity against the tested cell lines.

Highlights

• •A systematic research on the rare saponin compositions of Panax notoginseng saponin was conducted for the first time.
• •Three new and 25 known triterpenoid saponins were isolated.
• •The HPLC fingerprint of Panax notoginseng saponin was established.
• •The antitumor activities of the new compounds were tested.
• •New Compound 2 showed significant activity.

1. Introduction

Panax notoginseng (Burk.) F.H. Chen (P. notoginseng), commonly called “Sanqi” or “Tianqi” in Chinese is a species of the genus Panax, family Araliaceae [1]. P. notoginseng is one of the most widely used Chinese herbal drugs for the treatment of cardiovascular diseases, such as occlusive vasculitis, coronary diseases, atherosclerosis, and cerebral infarction in China, Korea, and Russia for a long time [2]. There are about 200 chemical compositions that have been isolated from P. notoginseng, including saponins, flavonoids, and cyclopeptides [3]. Dammarane triterpenoidal saponins are the major bioactive ingredients of P. notoginseng [4].

Panax notoginseng saponin (PNS) are developed into the traditional Chinese medicine agents with the trademarks of Xuesaitong injection, Xueshuantong injection, and Xuesaitong tablet, which are all bestselling prescriptions used for treatment of cardiovascular and cerebrovascular diseases in China [5], [6]. Notoginsenoside R₁ and ginsenoside Rg₁, Rd, Re, and Rb₁ are regarded as the main active constituents [7], [8]. But, there is a lack of references about the extraction and structure identification of the rare ginsenosides in PNS. As the study on microconstituents is important to control the quality and safety of clinical medication, we investigated the rare content in the PNS; three new and 25 known dammarane-type triterpenoids were isolated and identified (Fig. 1). We also established the HPLC fingerprint and marked the compounds isolated. In addition, we tested the cytotoxic activity of the three new compounds against three human cancer cell lines.

Fig. 1.

The structure of the Compounds 1–28. Glc, β-D-glucopyranosyl; Xyl, β-D-xylopyranosyl; Ara, α-L-arabinofuranosyl; Rha, α-L-rhamnopyranose.

2. Materials and methods

2.1. General experimental procedures

UV spectra: Shimadzu UV-2401A spectrophotometer (Shimadzu Instruments Co., Ltd, Tokyo, Japan). IR spectra: Nicolet FT-IR-360 spectrometer (Thermo Nicolet, Inc., Waltham, MA, USA). NMR spectra: Bruker ARX-400 spectrometers (Bruker Ltd, Karlsruhe, Germany). High resolution electrospray ionization mass spectrum (HR-ESI-MS) were taken using, Agilent G6230 instruments (Agilent Technologies, Santa Clara, CA, USA). Preparative HPLC was performed on an Agilent 1100 apparatus equipped with a Rheodyne injector and with UV detectors using a Thermo C₁₈ column (10 × 250 mm, 5 μm). Column chromatography was performed using silica (200–300 mesh and 100–200 mesh; Qingdao Marine Chemical, Inc., Qingdao, People's Republic of China) and octadecylsilyl (ODS) (50 μm; YMC Co., Kyoto, Japan). The reagents were of HPLC grade or analytical grade (Sinopharm Chemical Reagent Co., Ltd, Beijing, China).

2.2. Plant material

The PNS was provided by Yunnan Baiyao Group Co., Ltd.

2.3. Extraction and isolation

The PNS (2 kg) was separated by silica gel column using a gradient of CH₂Cl₂/CH₃OH (100:1→50:1→25:1→10:1→5:1→2:1→1:1, v/v) to obtain eight fractions (Fr.1–Fr.8). Fr.1 (10 g) was chromatographed subsequently over silica gel chromatography with CHCl₃-MeOH (30:1→20:1→10:1, v/v) to get five major fractions (Fr.1-1–Fr.1-5) based on thin-layer chromatography (TLC) analysis. Fr.1-2 was purified by ODS eluted with MeOH-H₂O (40:60→55:45→70:30, v/v) to provide Compounds 2 (5.1 mg) and 4 (75.5 mg). Fr.1-3 was then separated into three major fractions (Fr.1-3a–Fr.1-3c) by silica gel chromatography with CHCl₃-MeOH (15:1 and 13:1, v/v) as eluent. Fr.1-3a and Fr.1-3b were further separated by preparative HPLC (p-HPLC) eluting with MeCN-H₂O 28:72 to yield Compounds 3 (10.2 mg) and 16 (25.3 mg), while 25 (3.0 mg) and 8 (4.5 mg) were prepared from Fr.1-3c with MeCN-H₂O (29:71, v/v) as a solvent system. Fr.2 (18.8 g) was subjected to chromatography on ODS gel to afford 10 factions (Fr.2-1–Fr.2-10). Fr.2-3 was subjected to chromatography on silica gel to yield Compound 6 (7.7 mg). Compounds 7 (5.3 mg) and 14 (3.6 mg) were isolated from Fr.2-6 by p-HPLC with MeCN-H₂O (30:70, v/v), and Compounds 18 (9.8 mg) and 20 (5.6 mg) were isolated from Fr.2-7 by p-HPLC with MeCN-H₂O (28:72, v/v). Fr.2-8 was chromatographed over silica gel to obtain 10 major fractions with CHCl₃-MeOH (50:1 to 5:1, v/v) as eluent (Fr.2-8-1–Fr.2-8-10). Fr.2-8-5 was separated subsequently by ODS chromatography eluting with MeOH-H₂O (40:60→55:45→70:30, v/v) to afford 19 (2.3 mg). Fr.2-4 was fractionated by ODS eluted with MeOH-H₂O (40:60→60:40→80:20, v/v) to afford five major fractions (Fr.2-4-1–Fr.2-4-5). The analysis of a combined fraction of Fr.2-4-1–Fr.2-4-3 was performed by p-HPLC. Compound 27 (9.5 mg) was isolated from Fr.2-4-1 by HPLC system of MeCN-H₂O (28:72, v/v). Compounds 10 (8.5 mg) and 15 (3.3 mg) were prepared from Fr.2-4-2 with MeCN-H₂O (20:80, v/v) as a solvent system, whereas 9 (11.2 mg) and 12 (3.5 mg) were obtained from Fr.2-4-2 with MeCN-H₂O (37:63, v/v) as a solvent system. Compounds 11 (3.8 mg) and 28 (2.2 mg) were got from Fr.2-4-3 by HPLC system of MeCN-H₂O (43:57 and 35:65, v/v). Fr.5 (23.5 g) was subjected to chromatography on ODS gel to provide ten factions (Fr.5-1–Fr.5-10). Compound 5 (1115.8 mg) was purified by recrystallizing from Fr.5-1. The analysis of other compounds isolated from Fr.5 was performed by p-HPLC: Compound 22 (mg) was purified from Fr.5-2, and Compounds 17 (51.1 mg) and 21 (12.7 mg) were isolated from Fr.5-3 by HPLC system of MeCN-H₂O (30:70, v/v). Compound 13 (13.4 mg) were isolated from Fr.5-5 HPLC system of MeCN-H₂O (34:66, v/v). Compounds 1 (4.3 mg) and 26 (8.5 mg) were prepared from Fr.5-7 with MeCN-H₂O (27:73, v/v) as a solvent system, whereas 23 (36.7 mg) and 24 (8.9 mg) were prepared from Fr.5-8 with MeCN-H₂O (37:63, v/v) as a solvent system.

2.4. Notoginsenoside Ab₁ (1)

3β,6α,12β,22S-tetrahydroxy-dammar-20(21),24-diene-6-O-β-D-glucopyranoside: white amorphous powder; [α]$20\text{D}$ : +10.5, (c = 0.20, MeOH); IR ν_max 3420, 2931, 1634, 1454, 1384, 1074, 1032 cm⁻¹; ¹H and ¹³C NMR: see Table 1; HR-ESI-MS m/z 659.4130 [M+Na]⁺ (calculated for C₃₆H₆₀O₉Na, 659.4135).

Table 1.

¹H [ δ in ppm, multiplicity (J in Hz)] and ¹³C NMR (δ in ppm) spectroscopic data of Compounds 1–3.¹⁾

Position	1		2		3
	δC	δH	δC	δH	δC	δH
1	39.9	1.70 (1H, m) 1.05 (1H, m)	39.7	1.65 (1H, m) 1.02 (1H, m)	40.0	1.71 (1H, m) 1.05 (1H, m)
2	28.4	1.94 (1H, m) 1.87 (1H, m)	28.2	1.91 (1H, m) 1.84 (1H, m)	28.4	1.95 (1H, m) 1.88 (1H, m)
3	79.0	3.55 (1H, m)	78.7	3.52 (1H, m)	79.0	3.55 (1H, m)
4	40.9		40.7		40.9
5	61.8	1.46 (1H, m)	61.7	1.43 (1H, m)	61.9	1.46 (1H, m)
6	80.5	4.46 (1H, m)	80.2	4.45 (1H, m)	80.6	4.45 (1H, m)
7	45.8	2.59 (1H, m) 1.98 (1H, m)	45.6	2.57 (1H, m) 1.94 (1H, m)	45.8	2.55 (1H, m) 1.96 (1H, m)
8	41.7		41.5		41.7
9	51.1	1.59 (1H, m)	50.9	1.59 (1H, m)	51.1	1.58 (1H, m)
10	40.1		40.0		40.2
11	32.7	1.50 (2H, m)	32.8	1.88 (1H, m) 1.52 (1H, m)	31.5	1.98 (1H, m) 1.57 (1H, m)
12	80.2	4.29 (1H, m)	80.0	4.25 (1H, m)	80.2	4.28 (1H, m)
13	59.1	2.13 (1H, m)	55.4	2.29 (1H, m)	52.8	2.16 (1H, m)
14	52.2		51.7		51.6
15	33.4	1.92 (1H, m) 1.26 (1H, m)	32.9	1.52 (1H, m) 1.22 (1H, m)	33.0	1.76 (1H, m) 1.16 (1H, m)
16	34.7	2.16 (1H, m) 1.60 (1H, m)	35.9	2.21 (1H, m) 1.43 (1H, m)	35.1	2.18 (1H, m) 2.05 (1H, m)
17	40.2	3.05 (1H, m)	38.4	2.91 (1H, m)	48.7	2.29 (1H, m)
18	17.8	1.25 (3H, s)	17.6	1.28 (3H, s)	17.8	1.24 (3H, s)
19	18.2	1.04 (3H, s)	18.1	1.03 (3H, s)	18.2	1.05 (3H, s)
20	160.2		156.3		156.5
21	111.2	5.31 (2H, d, J = 12.5)	113.9	5.29 (2H, d, J = 12.5)	108.4	5.15 (1H, s) 4.97 (1H, s)
22	77.0	4.52 (1H, m)	91.3	4.71 (1H, t, J = 7.1)	33.1	2.77 (1H, m) 2.51 (1H, m)
23	36.3	2.73 (1H, m) 2.54 (1H, m)	31.3	2.60 (1H, m) 2.32 (1H, m)	33.2	2.06 (1H, m) 2.46 (1H, m)
24	122.4	5.42 (1H, t, J = 6.9)	121.1	5.34 (1H, t, J = 6.9)	75.5	4.46 (1H, m)
25	132.8		133.4		150.0
26	26.4	1.71 (3H, s)	26.1	1.65 (3H, s)	110.5	5.29 (1H, s) 4.96 (1H, s)
27	18.6	1.65 (3H, s)	18.3	1.59 (3H, s)	18.7	1.91 (3H, s)
28	32.2	2.10 (3H, s)	32.0	2.08 (3H, s)	32.2	2.10 (3H, s)
29	16.8	1.63 (3H, s)	16.7	1.60 (3H, s)	16.8	1.63 (3H, s)
30	17.1	0.88 (3H, s)	16.9	0.82 (3H, s)	17.2	0.82 (3H, s)
6-O-sugar
1	106.5	5.07 (1H, d, J = 8.0)	106.3	5.04 (1H, d, J = 7.8)	106.4	5.05 (1H, d, J = 7.8)
2	73.0	3.99 (1H, m)	73.2	3.93 (1H, m)	73.0	3.91 (1H, m)
3	78.7	3.99 (1H, m)	78.5	3.96 (1H, m)	78.6	3.98 (1H, m)
4	72.3	4.26 (1H, m)	72.1	4.22 (1H, m)	72.3	4.27 (1H, m)
5	75.9	4.13 (1H, m)	75.7	4.10 (1H, m)	75.9	4.13 (1H, m)
6	63.5	4.57 (1H, m) 4.41 (1H, m)	63.4	4.54 (1H, m) 4.36 (1H, m)	63.5	4.54 (1H, m) 4.44 (1H, m)

NMR, nuclear magnetic resonance; s, singlet; d,doublet; t, triplet; m, multiplet

¹⁾Measured in pyridine-d₅, 500 MHz for ¹H, 125 MHz for ¹³C, The assignment was based on DEPT, correalation spectroscopy (COSY), HSQC, and HMBC experiments.

2.5. Notoginsenoside Ab₂ (2)

22S-hydroperoxyl-3β,6α,12β-trihydroxy-dammar-20(21),24-diene-6-O-β-D-glucopyranoside: white amorphous powder; [α]$20\text{D}$ : +11.6, (c = 0.18, MeOH); IR ν_max 3422, 2933, 1637, 1452, 1384, 1075, 1031 cm⁻¹; ¹H and ¹³C NMR: see Table 1; HR-ESI-MS 675.4077 [M + Na]⁺ (calcd. for C₃₆H₆₀O₁₀Na 675.4084).

2.6. Notoginsenoside Ab₃ (3)

3β,6α,12β,24R-tetrahydroxy-dammar-20(21),25-diene-6-O-β-D-glucopyranoside: white amorphous powder; [α]$20\text{D}$ : +3.4, (c = 0.25, MeOH); IR ν_max 3416, 2941, 1636, 1452, 1386, 1163, 1076, 1041 cm⁻¹; ¹H and ¹³C NMR: see Table 1; HR-ESI-MS 659.4132 [M+Na]⁺ (calcd. For. C₃₆H₆₀O₉Na 659.4135).

2.7. Acid hydrolysis and HPLC analysis

The absolute configurations of the sugar moieties in Compounds 1–3 were determined by the method of literature reported [9]. Compounds 1–3 (2.0 mg/sample) were refluxed with 10 mL of 60% aqueous dioxane with 5% HCl for 2 h. The reaction mixture was evaporated under vacuum and then suspended in H₂O and extracted with CHCl₃. After drying in vacuum, the residue of aqueous layer was melted in 0.2 mL of C₅H₅N with 2 mg of L-cysteine methyl ester hydrochloride followed by warming at 60°C for 1 h. After that, 5 mL of o-tolylisothiocyanate is added and warmed up at 60°C for another hour. The reaction mixture was analyzed directly by reversed-phase HPLC on a Thermo C₁₈ column (250 × 4.6 mm, 5 μm), with 20% CH₃CN at a flow rate of 1.0 mL/min at 30°C, and the detection wavelength was 254 nm. The analysis of standard monosaccharide, D-glucose, followed the same procedure. The value of t_R of the standard monosaccharide derivatives was 17.8 min, and the derivatives of 1–3 gave peaks at t_R 17.7–17.9 min, respectively.

2.8. Computational studies

Conformational searches were performed with Gaussian 09W program (Gaussian Inc., USA). The geometry of each conformer in the energy window of the conformational search was optimized with Gaussian 09W in vacuum, at the B3LYP-6-31g (d,p) level. Imaginary vibrational frequency of each conformer was checked, and no such frequency indicates true energy minima. Isotropic magnetic shielding was calculated with the GIAO (gauge-independent atomic orbital) method at the B3LYP/6-31G (d, P) level by using Gaussian 09W [10], [11].

2.9. Fingerprint analysis

Chromatographic conditions: Waters 1525 HPLC system (Waters Corp., Milford, Massachusetts, USA); Chromatographic column: VP ODS C₁₈ (250 mm × 4.6 mm, 5 μm; Agilent Technologies, Santa Clara, CA, USA); volume flow: 1.0 mL/min; column temperature: 30°C; detection wavelength: 203 nm; injection volume: 10 μL. The samples were eluted with the mixture of Solvent A (water) and Solvent B (acetonitrile). The elution rate using Solvent B was 20–45% for 0–60 min.

Preparation of samples: Accurately weighed 25 mg of powder sample was diluted with 10 mL of 70% methanol. Before injection, the samples were filtered through a 0.45-μm membrane filter.

2.10. Cell line

HepG-2 (human hepatic cancer cell line), NCI-H460 (human lung cancer cell), and MCF-7 (human breast cancer cell) were purchased from Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences (Shanghai, China). The HepG-2 and NCI-H460 cancer cells were maintained in Roswell Park Memorial Institute 1640 medium, and MCF-7 cancer cells were maintained in high-glucose Dulbecco's minimum essential medium, supplemented with 10% fetal bovine serum. The cells grew in a 5% CO₂ incubator at 37°C. The cells were routinely digested and passaged every 3 days.

2.11. Cell viability assay

The cells were plated in 96-well plates (1 × 10⁴ cells/well) overnight, then 1-3 at various concentrations of 0.01, 0.1, 1, 10, and 100 μg/mL and the positive control cisplatin at concentrations of 0.5, 1, 2, 4, and 8 μg/mL were treated in the plates for 72 h. Subsequently, 20 μL of 3-(4,5-dimethylthiazol-2-yl)-2,5 -diphenyltetrazolium bromide reagent (5 mg/mL) was added to each well for 4 h, and then 100 μL of triple liquid containing 10 mg SDS (sodium dodecyl sulfate), 1.2 μL of 36–37% concentrated hydrochloric acid, and 50 μL of isobutanol were added. After the coculture for 12 h, the reduction of cell viability was determined at 570 nm using a microplate reader (Bio-Rad, USA). The cell proliferation inhibition rate was calculated according to the following formula: Inhibition rate (%) = (Acontrol − Asample)/Acontrol × 100 [12].

3. Result and discussion

Compound 1 was obtained as white amorphous powder. The molecular formula of 1 was deduced to be C₃₆H₆₀O₉ by positive mass spectrometry (HR-ESI-MS) data at m/z 659.4130 [M+Na]⁺ (calculated for C₃₆H₆₀O₉Na, 659.4135). The ¹³C NMR (Table 1) showed 36 carbon signals. The distortionless enhancement by polarization transfer (DEPT) spectrum exhibited 7 methyls, 9 methylenes, 14 methines, and 6 quaternary carbons signals. Four olefinic carbon signals at δ_C 160.2, 132.8, 122.4, and 111.2 ppm suggested two double bonds in the molecule. The ¹H NMR showed signals of seven methyl groups at δ_H 0.88 (3H, s), 1.04 (3H, s), 1.25 (3H, s), 1.63 (3H, s), 1.65 (3H, s), 1.71 (3H, s), and 2.10 (3H, s); four oxygen substituted protons at δ_H 3.55 (1H, m), 4.29 (1H, m), 4.46 (1H, m), and 4.52 (1H, m); and one anomeric proton at δ_H 5.07 (1H, d, J = 8.0). The ¹H and ¹³C signals were fully assigned according to heteronuclear signal quantum correlation (HSQC) spectra (Table 1). Methylene carbon signal at δ_C 111.2 ppm showed correlation spots with protons at δ_H 5.18 (H-21) and 4.52 (H-21) ppm in HSQC spectrum. These two proton signals showed connections with carbon signals at δ_C 160.2 (C-20), 77.0 (C-22), and 40.2 (C-17) ppm in heteronuclear multiple bond correlation (HMBC) spectrum, and δ_H 4.52 (H-22) showed connection with δ_C 111.2 (C-20), 122.4 (C-24), and 40.2 (C-17) (Fig. 2). Thus, the signals at δ_C 160.2, 111.2, and 77.0 ppm were assigned to be the signals of C-20, C-21, and C-22, respectively. δ_H 5.42 (H-24) showed connections with δ_C 132.8 (C-25), 26.4 (C-26), and 18.6 (C-27). Therefore, it was concluded that the two double bonds were at △20(21) and △24(25). The signals of Compound 1 were quiet similar to those of ginsenoside Rk₃, except for the chemical shift of C-20, C-22, and C-23, which were at δ_C 160.2, 77.0, and 36.3 of Compound 1 but were at δ_C 155.4, 33.7, and 27.0 of ginsenoside Rk₃, respectively [13]. The shift to downfield of C-20 (+4.8 ppm), C-22 (+43.3 ppm), and C-23 (+9.3 ppm) indicated that C-22 of 1 was linked to hydroxyl. In addition, the β configuration was prompted by the large coupling constant observed for the anomeric proton δ_H 5.07 (1H, d, J = 8.0). The absolute configurations of sugar was elucidated as D-glucose through acid hydrolysis and HPLC analysis. Moreover, the linkages between H-1" (δ_H 5.07) and C-6 (δ_C 80.5) were determined by HMBC correlations. The configuration of OH at C-6 was α based on the correlations between H-6 with H-18β, 19β in rotating frame overhauser effect (ROESY) spectrum. In addition, H-17 was deduced as α-forms by correlations between H-17 and Me-30 in the ROESY spectrum. The configuration of C-22 was identified by the comparison of the calculated and experimental chemical shifts of ¹³C. The calculated chemical shifts of C-22 about (22R)-1 and (22S)-1 were δ_C 72.0 and δ_C 77.8, whereas the experimental result was δ_C 77.0. Therefore, the configuration of C-22 was identified as S. On the basis of the aforementioned analyses, Compound 1 could be deduced to be 3β,6α,12β,22S-tetrahydroxy-dammar-20(21),24-diene-6-O-β-D-glucopyranoside and named as notoginsenoside Ab₁.

Fig. 2.

The important HMBC correlations of Compounds 1–3. HMBC, heteronuclear multiplebond correlation.

Compound 2 was isolated as white amorphous powder. The HR-ESI-MS spectrum showed [M+Na]⁺ at m/z 675.4077 (calcd. for C₃₆H₆₀O₁₀Na 675.4084). It was proposed to possess a hydroperoxyl group due to positive response to N,N-dimethyl-p-phenylenediammonium dichloride reagent [14]. The ¹³C NMR (Table 1) showed 36 carbon signals. The DEPT spectrum exhibited 7 methyls, 9 methylenes, 14 methines, and 6 quaternary carbons signals. The ¹H NMR showed signals of seven methyl groups at δ_H 0.82 (3H, s), 1.03 (3H, s), 1.28 (3H, s), 1.59 (3H, s), 1.60 (3H, s), 1.65 (3H, s), and 2.08 (3H, s). The signals of Compound 2 were quiet similar to those of Compound 1, except for the chemical shift of C-20 (−3.9 ppm), C-22 (+14.3 ppm), and C-23 (−5 ppm), indicating the hydroperoxyl substitution of C-22. C-22 (δ_H 4.71, δ_C 91.3) bearing hydroperoxyl group was also supported by chemical shift comparison with those of related hydroperoxylated triterpenes [15], [16], [17]. In addition, the long-range correlations between H-21 and C-17, 22, H-22 and C-17, 21, 23, 24, H-24 and C-26, 27, in the HMBC spectrum further confirmed the structure (Fig. 2). The monosaccharide was determined to be D-glucose by HPLC analysis of chiral derivatives of sugars in the acid hydrolyzate. The β-anomeric configuration for the glucosyl unit was established from the coupling constant for the anomeric proton δ_H 5.04 (d, J = 7.8 Hz). Compared with the calculated chemical shifts of C-22 about (22R)-2 (δ_C 82.8) and (22S)-2 (δ_C 88.2), the configuration of C-22 in Compound 2 was identified as S. Compound 2 could be deduced to be 22S-hydroperoxyl-3β,6α,12β-trihydroxy-dammar-20(21),24-diene-6-O-β-D-glucopyranoside and named as notoginsenoside Ab₂.

Compound 3, was obtained as white amorphous powder. The molecular formula was assigned as C₃₆H₆₀O₉ by the HR-ESI-MS spectrum at m/z 659.4132 [M+Na]⁺ (calcd. For. C₃₆H₆₀O₉Na 659.4135). The ¹³C NMR (Table 1) showed 36 carbon signals, and the ¹H NMR showed six methyl groups signals. The signals were similar to those of ginsenoside-Rh₅ [18], except for the position of the double bond. The corrections between H-21 (δ_H 4.97, 5.15) and C-20, 17, 22 in HMBC spectrum (Fig. 2) demonstrated there is a double bond in C-20, 21. What is more, the chemical shift of C-20, C-21, and C-22 were at δ_C 156.5, 108.4, and 33.1 in Compound 3 but that of ginsenoside-Rh₅ were at δ_C 142.5, 13.2, and 122.2 in ginsenoside-Rh₅, further indicating that the double bond was in C-20, 21 instead of in C-20, 22. There was an additional β-D-glucose in Compound 3 based on the results of acid hydrolysis analysis and the coupling constant for the anomeric proton. The correlation of H-6/H-18β, H-6/H-19β in ROESY spectrum revealed the α-orientation of OH at C-6. The correlation between H-17 and Me-30 in ROESY spectrum indicated the α-forms of H-17. The configuration of C-24 was identified as R by the comparison of the calculated chemical shifts of (24R)-3 (δ_C 72.3) and (24S)-3 (δ_C 67.9) with experimental chemical shift of C-24 (δ_C 75.5). Compound 3 could be deduced to be 3β,6α,12β,24R-tetrahydroxy-dammar-20(21),25-diene-6-O-β-D-glucopyranoside and named as notoginsenoside Ab₃.

The known compounds were identified as notoginsenoside R₂ (4) [19], 20(S)-ginsenoside Rh₁ (5) [20], 20(S)-ginsenoside Rg₂ (6) [21], ginsenoside Rg₃ (7) [22], pseudoginsenoside RT₃ (8) [23], gypenoside-XVII (9) [24], vina-ginsenoside-R₄ (10) [25], notoginsenoside R₃ (11) [26], quinquenoside L14 (12) [27], 20(R)-ginsenoside Rh₁ (13) [28], ginsenoside F₂ (14) [29], ginsenoside U (15) [30], ginsenoside Rg6 (16) [31], ginsenoside Rk₃ (17) [13], notoginsenoside T₅ (18) [32], ginsenoside Rk₁ (19) [33], ginsenoside Rg₄ (20) [34], ginsenoside-Rh₄ (21) [35], ginsenoside-Rh₅ (22) [18], 20(S)-ginsenoside-ST₂ (23) [36], floralquinquenoside A (24) [37], ginsenoside-Rh₁₄ (25) [38], notopanaxoside A (26) [39], notopanaxoside G (27) [40], sanchirhinoside D (28) [41] by comparison of their data with the literature (Fig. 1).

As typical multiple-constituent and multiple-action traditional Chinese medicine, the quality is difficult to control, so a systematic study about the complicated ingredients of PNS is much essential. This is the first study to explore the rare content of PNS comprehensively. In the study, we isolated 3 new and 25 known dammarane-type triterpenes from PNS. In addition, a method of HPLC-UV was successfully applied to the determination of rare ginsenosides in PNS. As the Fig. 3 showed, we also marked the compounds isolated in the fingerprint. The results of this study are significant for quality improvement and evaluation of clinical medicines made of PNS.

Fig. 3.

(A) The HPLC fingerprint of PNS. (B) The magnified HPLC fingerprint of PNS with compounds isolated from PNS marked. (1, notoginsenoside Ab₁; 2, notoginsenoside Ab₃; 3, notoginsenoside Ab₃; 4, notoginsenoside R₂; 5, 20(S)-ginsenoside Rh₁; 6, 20(S)-ginsenoside Rg₂; 7, ginsenoside Rg₃; 8, pseudoginsenoside RT₃; 9, gypenoside-XVII; 10, vina-ginsenoside-R₄; 11, notoginsenoside R₃; 12, quinquenoside L₁₄; 13, 20(R)-ginsenoside Rh₁; 14, ginsenoside F₂; 15, ginsenoside U; 16, ginsenoside Rg₆; 17, ginsenoside Rk₃; 18, notoginsenoside T₅; 19, ginsenoside RK₁20, ginsenoside Rg₄; 21, ginsenoside-Rh₄; 22, ginsenoside-Rh₅; 23, 20(S)-ginsenoside-ST₂; 24, floralquinquenoside A; 25, ginsenoside-Rh₁₄; 26, notopanaxoside A; 27, notopanaxoside G; 28, sanchirhinoside D). All the Compounds 1-28 isolated form PNS and the sample of PNS were analysis by HPLC at the same condition. In addition, 1-28 were marked on the HPLC spectrum of PNS.

The new Compounds (1–3) were evaluated for their cytotoxic activity against HepG2, NCI-H460, and MCF-7 cancer cell lines by using the modified 3-(4,5-dimethylthiazol-2-yl)-2,5 -diphenyltetrazolium bromide method, with cisplatin as the positive control (Table 2). New Compound 2 showed significant activity against HepG2, NCI-H460, and MCF-7 with IC₅₀ 4.49, 8.06, and 7.38 μg/mL, respectively.

Table 2.

IC₅₀ values of the new Compounds (1–3) against HepG2, NCI-H460, and MCF-7 cells.

Comp.	IC50 (μg/mL)1)
	HepG2	NCI-H460	MCF-7
1	>100	>100	>100
2	4.49	8.06	7.38
3	>100	>100	>100
Cisplatin2)	0.97	1.34	2.66

Cell inhibition activity was determined by the MTT assay. MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5 -diphenyltetrazolium bromide

¹⁾IC₅₀ is the concentration of the compound inhibiting 50% of cell proliferation.

²⁾Cisplatin was used as positive control.

Conflicts of interest

All authors declare no conflicts of interest.

Appendix A. Supplementary data

The following is the supplementary data to this article: