Isolation, structural elucidation, and bioactivity of cholestane derivatives from Ypsilandra thibetica

Abstract

Phytochemical investigation on the whole plant of Ypsilandra thibetica led to the isolation of five new cholestane derivatives. The structures of 1–5 were elucidated by comprehensive spectroscopic analyses of their data and chemical methods. The absolute configurations of compounds 1–3 were ambiguously assigned by single crystal X-ray diffraction analysis. Compound 1 is the first cholestane glycoside with two oxygen bridges between C-16/C-23 and C-18/C-23. Compound 2 is the first 23-spirocholestane derivative with a six-membered ring between C-24/C-23 of the aglycone moiety and C-1′/C-2′ of the D-arabinose ligand, while compounds 3–5 are new 23-spirocholestane derivatives with a D-fucose ligand. The cytotoxic, anti-inflammatory, and anti-angiogenic activities of these isolated compounds were evaluated. Compound 5 showed anti-angiogenic activity in zebrafish embryos.

---

Introduction

Steroidal saponins are rich in approximately 90 plant families, particularly in monocotyledons, such as Liliaceae, Melanthiaceae, Convallariaceae, Dioscoreaceae, Agavaceae, Smilacaceae, and Amaryllidaceae.¹ Various pharmacological activities of these saponins have been investigated, including their antitumor,^2–5 anti-inflammatory,⁶ antifungal,^7,8 and platelet aggregation inducing activities.⁹Ypsilandra thibetica Franch (Melanthiaceae), a perennial herb used as a famous Yi nationality medicine,^10,11 has been used to treat lymphatic tuberculosis, hemostasis, and disorders of hemostasis and thrombus.^12,13 Our previous investigations on this species uncovered a series of structurally diverse steroidal saponins including spirostanol-, furostanol-, and cholestanol-glycosides, more importantly, some of which were revealed to show antibacterial, anticancer, and anti-HIV activities.^6,14–21 In order to obtain more structurally interesting steroidal glycosides, we further investigated the phytochemicals of this species. As a result, five rare cholestane derivatives were isolated and structurally characterized. Herein, we report the isolation, structural elucidation, and bioactivities of these isolated compounds. In the present study, we aimed to evaluate the newly isolated saponins from Ypsilandra thibetica for their potential cytotoxicities in esophageal cancer cells EC109, anti-inflammatory activities in RAW 264.7 macrophages, as well as anti-angiogenic activities of these compounds in zebrafish embryos.

Results and discussion

Compound 1 was isolated as a prismatic crystal. Its molecular formula was determined to be C₃₃H₅₂O₉ based on its quasi-molecular ion peak at m/z 615.3514 [M + Na] ⁺ (calcd 615.3504) in HR-ESI-MS and ¹³C NMR data (Table 2). The IR spectrum showed characteristic absorption bands at 3424 cm⁻¹ and 1709 cm⁻¹, indicating the presence of hydroxyl and carbonyl, respectively. From its HSQC and DEPT spectra, the ¹³C and ¹H NMR spectra of 1 exhibited 33 carbon signals, including five methyls at δ_C 13.2 (C-19), 24.1 (C-21), 22.7 (C-26), 19.2 (C-27), and 17.3 (C-6′), and the corresponding proton signals at δ_H 0.71 (s, CH₃-19), 1.20 (d, J = 7.2 Hz, CH₃-21), 1.24 (d, J = 6.8 Hz, CH₃-26), 1.19 (d, J = 6.8 Hz, CH₃-27), and 1.51 (br d, J = 6.4 Hz, CH₃-6′); nine methylenes including one oxygenated signal at δ_C 66.0 (C-18) and proton signals at δ_H 3.92 (br d, J = 13.2 Hz, CH₂-18a) and 3.32 (br d, J = 13.2 Hz, CH₂-18b); fifteen methines with one acetal signal at δ_C 106.6 (C-1′) and seven oxymethine signals at δ_C 70.0 (C-3), δ_C 74.3 (C-16), δ_C 89.6 (C-24), δ_C 73.1 (C-2′), δ_C 75.6 (C-3′), δ_C 72.7 (C-4′), and δ_C 71.5 (C-5′), corresponding to the proton signals at δ_H 3.82 (m, H-3), δ_H 4.65 (dd, J = 7.7, 2.9 Hz, H-16), δ_H 3.66 (d, J = 3.7 Hz,H-24), δ_H 4.86 (d, J = 7.7 Hz, H-1 H-δ_H 4.31, (t, J = 8.6 Hz, H-2′), δ_H 4.04 (dd, J = 9.4, 3.4 Hz, H-3′), δ_H 3.96 (br d, J = 2.6 Hz, H-4′), and δ_H 3.72 (br q, J = 6.1 Hz, H-5′); one carbonyl signal at δ_C 209.8 (C-6); three quaternary carbons, among which is a ketal signal at δ_C 104.4 (C-23). These spectroscopic data suggested that 1 should be a cholestanol glycoside.¹⁵ The carbonyl group accounted for one degree of unsaturation, and the remaining seven degrees of unsaturation implied compound 1 could be heptacyclic (rings A–G). Furthermore, ¹H–¹H COSY correlations of H₂-1/H₂-2/H-3/H₂-4/H-5, H₂-7/H-8/H-9/H₂-11/H₂-12 and H-8/H-14/H₂-15/H-16/H-17/H-20/H₂-22, as well as H-20/CH₃-21, H-24/H-25/CH₃-26, H-25/CH₃-27, and H-1′/H-2′/H-3′/H-4′/H-5′/CH₃-6′ revealed four spin system (a–d) as shown in Fig. 2. Moreover, the connectivity of the substructures a–c to C-6 (δ_C 209.8)/C-10 (δ_C 40.8)/C-13 (δ_C 47.5)/C-23 (δ_C 104.4) and two tertiary methyls with six rings were verified based on the HMBC correlations (Fig. 2): from H₂-18 (δ_H 3.92 and 3.32) to C-12 (δ_C 34.5)/C-13 (δ_C 47.5)/C-14 (δ_C 53.4)/C-17 (δ_C 52.9)/C-23 (δ_C 104.4); from CH₃-19 (δ_H 0.71) to C-1 (δ_C 36.9)/C-5 (δ_C 57.0)/C-9 (δ_C 54.4)/C-10 (δ_C 40.8); from H₂-4 (δ_H 2.31 and 1.88)/H-5 (δ_H 2.26)/H₂-7 (δ_H 2.43 and 2.07)/H-8 (δ_H 2.11) to C-6 (δ_C 209.8); from H-16 (δ_H 4.65)/H-20 (δ_H 2.05)/H-24 (δ_H 3.66)/H-25 (δ_H 2.22) to C-23 (δ_C 104.4). The above evidence revealed that E- and F-rings shared the oxygen bridge of C-16/C-23, and F- ring possessed an additional rare oxygen bridge of C-18/C-23 as shown in Fig. 1. In addition, a set of six signals corresponding to the D-fucopyranosyl moiety²² were observed at δ_C 106.6 (C-1′), δ_C 73.1 (C-2′), δ_C 75.6 (C-3′), δ_C 72.7 (C-4′), δ_C 71.5 (C-5′), and 17.3 (C-6′), while the HSQC spectrum showed the corresponding proton signals at δ_H 4.86 (d, J = 7.7 Hz, H-1′), δ_H 4.31 (t, J = 8.6 Hz, H-2′), δ_H 4.04 (dd, J = 9.4, 3.4 Hz, H-3′), δ_H 3.96 (br d, J = 2.6 Hz, H-4′), δ_H 3.72, (br q, J = 6.1 Hz, H-5′), and δ_H 1.51 (br d, J = 6.4 Hz, CH₃-6′). The HMBC correlations of H-5′ (δ_H 3.72)/C-1′ (δ_C 106.6) and H-1′ (δ_H 4.86)/C-24 (δ_C 89.6) were assigned for the connectivity of the substructure d and the glycosylated site. As a result, the G-ring has one oxygen bridge between C-1′/C-5′ and C-1′ is connected to C-24.

Fig. 1 Chemical structures of 1–5 and X-ray single crystal structures of 1–3.

Fig. 2 The ¹H–¹H COSY and key HMBC correlations of 1–5.

The relative configuration of 1 was established by the ROESY experiment (Fig. 3). The ROESY correlations of H-1β/CH₃-19/H-8/H-15β/H-20 revealed that these protons were β-oriented. The following protons were α-oriented based on the observed ROESY correlations of H-1α/H-3/H-5/H-9, H- 15α/H-14/H-16/H-17, and H-16/CH₃-21. The ROESY correlations of H-1′ with H-3′, H-4′, and H-5′ indicated that they were co-facial and were arbitrarily assigned to be β-oriented. Fortunately, we obtained the crystal of 1. The structure and the absolute configuration of 1 were further confirmed by single crystal X-ray diffraction with Cu K_α radiation described in Fig. 1. In conclusion, the structure of 1 was elucidated as (23R,24S)-5α-cholest-16β,23:18,23-diepoxy-3β,24-diol-6-one-24-O-β-D-fucopyranoside.

Fig. 3 Key ROESY correlations of 1, 2, and 4.

Compound 2 was obtained as a colorless needle crystal. The positive HR-ESI-MS showed an [M + H] ⁺ ion peak at m/z 563.3584 (calcd 563.3578), corresponding to the molecular formula of C₃₂H₅₀O₈. The ¹H and ¹³C NMR data (Tables 1 and 2), with the assistance of DEPT and HSQC spectra, exhibited the presence of 32 carbon resonances, including five cholestane-type steroid methyl signals at δ_C 14.9 (C-18), 13.3 (C-19), 20.9 (C-21), 22.8 (C-26), and 18.2 (C-27), and the corresponding proton signals at δ_H 0.73 (s, CH₃-18), 0.77 (s, CH₃-19), 0.78 (d, J = 7.1 Hz, CH₃-21), 1.15 (d, J = 6.7 Hz, CH₃-26), and 1.10 (d, J = 6.7 Hz, CH₃-27). Detailed comparison of its ¹H, ¹³C NMR data (Tables 1 and 2) and ¹H–¹H COSY spectra (Fig. 2) with those of 1 revealed that 2 was similar to 1, with the main differences being the disappearance of oxygenated methines replaced by the oxygenated methylene C-5′ (δ_C 69.6), disappearance of oxygenated methylene replaced by methyl C-18 (δ_C 14.9), upfield shift of C-23 (δC 104.4 → 100.3 ppm), C-24 (δC 89.6 → 86.0 ppm), C-2′ (δC 73.1 → 74.0 ppm), and C-3′ (δC 75.6 → 71.8 ppm). Furthermore, compared with 1, the observed HMBC correlations (Fig. 2) from H-16 (δ_H 4.46)/H-20 (δ_H 1.58)/H-22 (δ_H 2.25 and δ_H 1.10)/H-24 (δ_H 3.58)/H-25 (δ_H 2.06)/H-2′ (δ_H 4.19) to C-23 (δ_C 100.3) indicated that compound 2 forms an oxygen bridge of C-2′/C-23 and lacks the oxygen bridge of C-18/C-23. As for the sugar moiety, fragment d showed an acetal proton signal at δ_H 4.52 (br d, J = 7.7 Hz) and an acetal carbon signal at δ_C 102.2. The HMBC correlations from H₂-5′ (δ_H 4.39 and 3.82) to C-1′ (δ_C 102.0), H-1′ (δ_H 4.52) to C-24 (δ_C 86.0), and H-2′ (δ_H 4.19) to C-23 (δ_C 100.3) revealed that the connectivity of the F-ring by two oxygen bridges of C-2′/C-23 and C-1′/C-24 and G-ring had one oxygen bridge of C-1′/C-5′. The relative configuration of the aglycone moiety of 2 was identical to that of 1 by the ROESY experiments. The ROESY correlations of H-24/H-1′/H-3′/H-5hβ/H-4′ disclosed that they were β-oriented, on the contrary, H-2′/H-5′α was α-oriented. Its structure and absolute configuration were equivalently confirmed by X-ray single crystal diffraction with Cu K_α radiation as shown in Fig. 1. Thus, the structure of 2 was elucidated as (23R,24S)-5α-cholest-16β,23-epoxy-23,24-(β-D-arabinopyranose-2,1-di-O-yl)-3β-ol-6-one.

Table 1 ¹H NMR (500 MHz) spectroscopic data of compounds 1–5 (δ in ppm, J in Hz, C₅D₅N)

No.	1	2	3	4	5
s, singlet; d, doublet; t, triplet; br, broad; m, multiple; o, overlapped.
1a	1.58, o	1.64, o	1.62, o	1.63, o	1.52, o
1b	1.14, m	1.16, o	1.17, o	1.15, o	1.01, o
2a	2.04, m	1.65, o	2.00, o	2.03, o	2.05, o
2b	1.64, m	2.05, o	1.61, o	1.64, o	1.59, o
3	3.82, m	3.82, o	3.84, o	3.81, o	3.98, m
4a	2.31, o	2.31, m	2.30, o	2.30, br d (13.2)	2.40 br d (12.9)
4b	1.88, m	1.91, o	1.88, o	1.91, m	1.74, o
5	2.26, o	2.25, o	2.27, o	2.25 br d (12.4)	2.05, o
7a	2.43, dd (12.3,3.7)	2.38, dd (12.9,4.7)	2.41, m	2.36, dd (12.9,2.4)	2.34, dd (13.0,4.5)
7b	2.07, o	1.98, m	2.00, o	2.01, o	1.99, o
8	2.11, m	1.81, m	2.00, o	1.80, o	1.75, o
9	1.20, o	1.14, o	1.24, o	1.15, o	1.11, o
11a	1.54, o	1.51, dd (13.4,3.5)	1.54, m	1.51, o	1.45, m
11b	1.03, o	1.26, o	1.24, o	1.25, o	1.20, o
12a	1.67, o	1.75, dd (12.4,2.9)	2.80, dt (12.5,3.4)	1.75, o	1.75, o
12b	1.03, o	1.07, o	0.96, o	1.06, m	1.02, o
14	1.19, o	0.90, m	1.08, o	0.90, m	0.84, o
15a	2.02, o	1.83, o	1.91, o	1.78, o	1.77, o
15b	1.84, m	1.19, o	1.22, o	1.15, o	1.14, o
16	4.65, td (7.7, 2.9)	4.46, m	4.61, dd (7.2,2.4)	4.55, m	4.51, q (7.8)
17	1.29, br d (7.1)	0.81, o	0.96, o	0.81, d (10.3)	0.83, o
18a	3.92, br d (13.2)	0.73, s	3.91, s	0.74, s	0.73, s
18b	3.32, br d (13.2)		3.91, s
19	0.71, s	0.77, s	0.71, s	0.79, s	0.67, s
20	2.05, o	1.58, m	2.34, o	1.97, m	1.99, o
21	1.20, d (7.2)	0.78, d (7.1)	1.09, d (6.2)	0.86, d (6.4)	0.89 d (6.4)
22a	2.24, o	2.25, o	2.33, o	5.23, d (11.6)	5.25 d (11.7)
22b	1.91, m	1.10, o	1.12, d (6.8)
24	3.66, d (3.7)	3.58, d (5.4)	3.65, br d (4.3)	3.54, d (8.7)	3.56, d (8.8)
25	2.22, m	2.06, o	2.24, m	2.62, m	2.63, m
26	1.24, d (6.8)	1.15, d (6.7)	1.12, d (6.8)	1.25, d (6.7)	1.26, d (8.3)
27	1.19, d (6.8)	1.10, d (6.7)	1.06, d (6.8)	1.19, d (6.7)	1.20, d (8.3)
	24-Fuc	24-Ara	24-Fuc	24-Fuc	24-Fuc
1′	4.86, d (7.7)	4.52, br d (7.7)	4.60, dd (7.2,2.4)	4.51, br d (7.9)	4.56, m
2′	4.31, t (8.6)	4.19, td (9.7,2.0)	4.17, dd (9.9,7.6)	4.30, t (8.7)	4.26, o
3′	4.04, dd (9.4, 3.4)	4.13, m	4.11, br d (10.6)	4.07, d (9.3)	4.26, o
4′	3.96, br d (2.6)	4.26, br d (4.4)	4.06, br s	3.96, br s	5.54, br s
5′a	3.72, br q (6.1)	4.39, dd (12.5,1.9)	3.86, o	3.81, o	3.92, dd (13.3,6.3)
5′b		3.82, br d (12.4)
6′	1.51, br d (6.4)		1.58, br d (6.3)	1.48, br d (6.4)	1.25, br d (4.5)
22-OAc				2.09, s	2.20, s
4′-OAc					2.28, s
3-Glc
1′′					5.06, br d (7.7)
2′′					4.07, t (8.2)
3′′					4.29, o
4′′					4.26, o
5′′					4.02, o
6′′a					4.62, dd (13.4,3.6)
6′′b					4.41, dd (11.8,5.6)

---

Table 2 ¹³C NMR (125 MHz) spectroscopic data of compounds 1–5 (δ in ppm, C₅D₅N)

No.	1	2	3	4	5
1	36.9, CH2	37.0, CH2	37.0, CH2	36.9, CH2	36.6, CH2
2	31.8, CH2	31.8, CH2	31.8, CH2	31.9, CH2	29.5, CH2
3	70.0, CH	70.0, CH	70.0, CH	70.0, CH	76.6, CH
4	31.3, CH2	31.3, CH2	31.3, CH2	31.3, CH2	27.0, CH2
5	57.0, CH	57.0, CH	57.0, CH	57.0, CH	56.4, CH
6	209.8, C	210.1, C	210.1, C	210.0, C	209.5, C
7	46.7, CH2	46.9, CH2	47.2, CH2	46.9, CH2	46.8, CH2
8	37.3, CH	37.0, CH	37.2, CH	36.9, CH	36.8, CH
9	54.4, CH	53.8, CH	54.0, CH	53.7, CH	53.6, CH
10	40.8, C	41.0, C	41.1, C	41.0, C	40.9, C
11	21.8, CH2	21.6, CH2	21.7, CH2	21.6, CH2	21.5, CH2
12	34.5, CH2	40.2, CH2	34.6, CH2	40.2, CH2	40.1, CH2
13	47.5, C	42.2, C	46.9, C	42.7, C	42.7, C
14	53.4, CH	52.6, CH	52.5, CH	52.7, CH	52.6, CH
15	38.3, CH2	32.7, CH2	32.8, CH2	32.8, CH2	32.7, CH2
16	74.3, CH	71.4, CH	71.8, CH	70.1, CH	70.2, CH
17	52.9, CH	58.8, CH	59.5, CH	55.3, CH	55.3, CH
18	66.0, CH2	14.9, CH3	59.9, CH2	14.8, CH3	14.8, CH3
19	13.2, CH3	13.3, CH3	13.3, CH3	13.3, CH3	13.1, CH3
20	21.9, CH	26.1, CH	27.1, CH	33.1, CH	33.0, CH
21	24.1, CH3	20.9, CH3	21.9, CH3	16.3, CH3	16.2, CH3
22	31.1, CH2	37.3, CH2	37.7, CH2	79.6, CH	79.7, CH
23	104.4, C	100.3, C	100.6, C	99.9, C	100.2, C
24	89.6, CH	86.0, CH	86.0, CH	87.4, CH	87.3, CH
25	28.5, CH	28.2, CH	28.1, CH	28.8, CH	28.7, CH
26	22.7, CH3	22.8, CH3	23.1, CH3	22.8, CH3	22.7, CH3
27	19.2, CH3	18.2, CH3	17.9, CH3	20.2, CH3	20.2, CH3
	Fuc	Ara	Fuc	Fuc	Fuc
1′	106.6, CH	102.0, CH	101.5, CH	101.6, CH	101.0, CH
2′	73.1, CH	74.0, CH	73.7, CH	73.5, CH	73.2, CH
3′	75.6, CH	71.8, CH	72.3, CH	72.6, CH	70.5, CH
4′	72.7, CH	70.3, CH	73.0, CH	72.8, CH	74.4, CH
5′	71.5, CH	69.6, CH2	73.5, CH	73.5, CH	71.6, CH
6′	17.3, CH3		17.2, CH3	17.1, CH3	16.7, CH3
22-OAc				170.0, C	170.2, C
				21.2, CH3	21.3, CH3
4′-OAc					171.1, C
					20.9, CH3
3-Glc
1′′					102.2, CH
2′′					75.4, CH
3′′					78.7, CH
4′′					71.9, CH
5′′					78.7, CH
6′′					63.0, CH2

---

Compound 3 was obtained as a colorless needle crystal. By the HR-ESI- MS at m/z 615.3512 [M + Na]⁺ (calcd 615.3504) and ¹³C NMR data (Table 2) analysis, we deduced 3 possessed the same molecular formula as 1. Comparison the NMR (Tables 1 and 2) and ¹H–¹H COSY spectra (Fig. 2) of 3 with those of 1 suggested that 3 was similar to 1 except for the obviously different chemical shift signals for C-18 (δ_C 66.0 → 59.9), C-23 (δ_C 104.4 → 100.6), C-24 (δ_C 89.6 → 86.0), C-2′(δ_C 73.1 → 73.7), and C-3′ (δ_C 75.6 and C-3 Furthermore, the HMBC correlations (Fig. 2) from H-16 (δ_H 4.61)/H-20 (δ_H 2.34)/H-24 (δ_H 3.65)/H-25 (δ_H 2.24)/H-2′ (δ_H 4.17) to C-23 (δ_C 100.6) and H-1′ (δ_H 4.60) to C-24 (δ_C 86.0) revealed that compound 3 has two oxygen bridges of C-1′/C-24 and C-2′/C-23, however, it lacks the oxygen bridge of C-18/C-23. The relative configuration of 3 was confirmed by ROESY correlations and was identical to 2. The absolute stereochemistry of 3 was unambiguous substantiated by X-ray single crystal diffraction with Cu K_α radiation as shown in Fig. 1. Based on the above evidence, the structure of 3 was elucidated as (23R,24S)-5α-cholest-16β,23-epoxy-23,24-(β-D-fucopyranose-2,1-di-O-yl)-3β,18-diol-6-one, named 18-hydroxylypsilanogenin.

The molecular formula of compound 4 was deduced as C₃₅H₅₄O₁₀ according to the ion peak at m/z 657.3616 [M + Na]⁺ (calcd 657.3609) and ¹³C NMR data (Table 2). The IR absorptions at 3440, 1712, and 1742 cm⁻¹ implied the existence of hydroxy, carbonyl, and esteryl groups, respectively. Compounds 4 and 3 had similar structures by contrastive analyses of the ¹H and ¹³C NMR data (Tables 1 and 2), ¹H–¹H COSY and HMBC spectra (Fig. 2). However, detailed analyses of the NMR signals of 4 lead to finding the absences of a methylene replaced by an oxymethine δ_C 79.6 (C-22), an oxymethylene replaced by a methyl δ_C 14.8 (C-18), and the appearance of an acetyl group (δ_C 170.0 and 21.2). The acetoxyl group connected to C-22 proved the correlations of H-14/H-16/CH₃-21/H-22 (Fig. 3). The rest by HMBC correlation from H-22 (δ_H 5.23)/δ_H 2.09 (CH₃) to δ_C 170.0 (Fig. 2) and the acetoxyl group was β-oriented by the ROESY stereochemistry of 4 and was found to be identical to 3 by the ROESY experiment. In a nutshell, the structure of compound 4 was elucidated as (23S,24S)- 5α-cholest-16β,23-epoxy-22β-acetoxy-23,24-(β-D-fucopyranose-2,1-di-O-yl)-3β-ol-6-one, and named 22β-acetoxylypsilanogenin.

Compound 5, a white amorphous solid, has a molecular formula of C₄₃H₆₆O₁₆ hinged on its ion peak at m/z 861.4251 [M + Na] ⁺ (calcd. 861.4243) in the HR-ESI-MS spectrum and ¹³C NMR data (Table 2). Based on the analyses of its ¹H and ¹³C NMR spectra as well as ¹H–¹H COSY and HSQC experiments (Fig. 2), compound 5 has the same aglycone moiety with 4, but the differences are an additional acetyl group (δ_C 171.1 and 20.9) and D-glucopyranosyl moiety at δ_C 102.2 (C-1′′), 75.4 (C-2′′), 78.7 (C-3′′), 71.9 (C-4′′), 78.7 (C-5′′), and 63.0 (C-6′′), which was confirmed by acid hydrolysis experiment. The β-anomeric configuration for the D-glucopyranosyl moiety was confirmed by the coupling constant ³J_1,2 (7.7 Hz) of anomeric proton.²³ The HMBC correlation between H-1′′ (δ_H 5.06) and δ_C 76.6 (C-3) revealed that the D-glucopyranosyl moiety was connected to C-3. The additional acetyl group was attached at C-4′ deduced by the HMBC correlations from δ_H 5.54 (H-4′)/δ_H 2.28 (CH₃) to δ_C 171.1. The stereochemistry of 5 was identical to compound 4 by the ROESY experiment (Fig. 3). Consequently, the structure of 5 was identified as (23S,24S)-5α-cholest-16β,23-epoxy-22β,4′-diacetoxy-23,24-(β-D-fucopyranose-2,1-di-O-yl)-3β-ol-6-one-3-O-β-D-glucopyranoside, and named 22β,4′,-diacetoxylypsilanogenin 3-O-β-D-glucopyranoside.

The possible biosynthetic pathways of compounds 1–5 are proposed in Scheme 1. From the biosynthetic perspective, compound 1–5 had the same synthetic precursor, acetyl-CoA. As deduced in Scheme 1, acetyl-CoA may undergo muti-step reaction to give the intermediate isofucosterol. Then isofucosterol may undergo stepwise oxidization, elimination, and hydrolysis reaction to yield compound VIII. Furthermore, X with fucose or arabinose was obtained through glycosylation and oxidization reaction from VIII. Ketalization reaction of compound X could yield compound 2 and XI (ypsilanogenin) with a different sugar respectively. Subsequently, XI may afford compound 3 by further oxidation. The OH-18 of 3 undergoes ketal exchange reaction to obtain compound 1, which possesses a new ketal carbon and a new hydroxy. Furthermore, ypsilanogenin (XI) undergoes stepwise oxidation, acetylation, and glycosylation to generate compounds 4 and 5.

Scheme 1 The possible biosynthetic pathways of 1–5.

Many cholestanol glycosides have shown cytotoxic activities²⁴ and antibacterial activities.²⁵ Therefore, all these isolated compounds were evaluated in vitro for their cytotoxic activity against esophageal cancer cells EC109 and anti-inflammatory activity in RAW 264.7 macrophages. Furthermore, the inhibitory effect of these compounds on the formation of sub-intestinal vessels (SIV) in zebrafish embryos were also evaluated. Unfortunately, none of these tested compounds showed cytotoxic effects (IC₅₀ > 50 μM) on EC109 cells as compared with the positive drug cisplatin. On the other hand, compounds 1 and 2 exhibited weak inhibitory activities (17.6% and 18.5%, respectively) at a concentration of 50 μM against LPS-induced nitric oxide (NO) production in RAW 264.7 macrophages. The formation of SIV in zebrafish embryos was shown to be inhibited by compound 5 (10 μM) as the length of SIV was significantly decreased after treatment as shown in Fig. 4.

Fig. 4 The length of SIV in zebrafish embryos after being treated with isolates.

Experimental

General experimental procedures

HR-ESI-MS data were recorded on an Agilent 1290 UPLC/6540 Q-TOF mass spectrometer. The NMR spectra were recorded on a Bruker AVANCE III 500. IR spectra were recorded on a Bruker FT-IR Tensor-27 infrared spectrophotometer with KBr disks. X-ray diffraction was conducted using a Bruker APEX DUO diffractometer with graphite-monochromated Cu Kα radiation. Silica gel (200–300 mesh, Qingdao Marine Chemical Co., Ltd, China), RP-18 (50 μm, Merck, Germany), and Sephadex LH-20 (GE Healthcare, USA) were used for column chromatography (CC). An Agilent 1260 system (Agilent Corp., America) with a Zorbax SB-C18 column (5 μm, 9.4 × 250 mm) was used for HPLC separation. TLC was carried out on silica gel HSGF₂₅₄ plates (Qingdao Marine Chemical Co., China) or RP-18 F₂₅₄ (Merck, Darmstadt, Germany).

Plant materials

The plant materials of Y. thibetica, identified by Prof. Xin-Qi Chen (Institute of Botany, Chinese Academy of Sciences, Beijing) were collected in August 2010 from Zhaotong City, Yunnan Province, China A voucher specimen (No. HY0002) was deposited at the State Key Laboratory of Phytochemistry and Plant Resources in West China, Kunming Institute of Botany, Chinese Academy of Sciences.

Extraction and isolation

110 kg of dried whole plants of Y. thibetica were ground into powder and fully extracted three times with 70% EtOH under reflux for 3 h, 2 h, and 2 h. Then, the combined extract was concentrated under reducing pressure to yield 30 kg of crude extract. The extract was dispersed in water and extracted using water-saturated n-butanol three times. The n-BuOH soluble fractions were combined and concentrated. The n-BuOH extract (13 kg) was subjected to column chromatography over silica gel with a mixture (CHCl₃–MeOH–H₂O, 20 : 1 : 0.05 → 7 : 3 : 0.5, v/v/v) to give eleven fractions (Fr. 1–Fr. 11). Fr. 1 (2 kg) was subjected to a polyamide column chromatography, eluting with a gradient system of EtOH–H₂O (20% → 60%, v/v) to afford four main fractions (Fr. 1.1–Fr. 1.4). Fr. 1.3 (800 g) was separated using silica gel column and further fractionated by a RP-C18 silica gel column (MeOH–H₂O, 30% → 100%, v/v) to afford Fr. 1.3.1–Fr. 1.3.11. Fr. 1.3.3 was applied to a Sephadex LH-20 column (MeOH), and further purified by semi-preparative HPLC (35–55% MeCN in H₂O within 30 min, 4 mL min⁻¹) to afford 1 (273 mg, 21.3 min) and 2 (38.6 mg, 26.5 min). Fr. 1.3.5 was repeatedly purified by semi-preparative HPLC (30–65% MeCN in H₂O within 40 min, 4 mL min⁻¹) to afford 3 (88 mg, 20.5 min) and 4 (179 mg, 22.1 min) and 5 (89 mg, 27.2 min).

Compound 1

Colorless prismatic crystal. [α]²⁵_D 12.86 (c 0.07, MeOH). IR (KBr) ν_max 3424, 2934, 2872, 1709, 1633, 1424, 1330, 1273, 1134, 1068 cm⁻¹; ¹H NMR (500 MHz, C₅D₅N) and ¹³C NMR (125 MHz, C₅D₅N) data see Tables 1 and 2; HR-ESI-MS m/z 615.3514 [M + Na] ⁺ (calcd for C₃₃H₅₂O₉Na, 615.3504).

Compound 2

Colorless needle crystal. [α]²⁵_D −22.75 (c 0.08, MeOH). IR (KBr) ν_max 3438, 2931, 2870, 1707, 1634, 1452, 1339, 1285, 1148 cm ⁻¹; ¹H NMR (500 MHz, C₅D₅N) and ¹³C NMR (125 MHz, C₅D₅N) data see Tables 1 and 2; HR-ESI-MS 563.3584 [M + H]⁺ (calcd for C₃₂H₅₁O₈, 563.3578).

Compound 3

Colorless needle crystal. [α]²⁵_D −19.71 (c 0.07, MeOH). IR (KBr) ν_max 3441, 2934, 2870, 1708, 1633, 1453, 1332, 1284, 1141 cm⁻¹; ¹H NMR (500 MHz, C₅D₅N) and ¹³C NMR (125 MHz, C₅D₅N) data see Tables 1 and 2; HR-ESI-MS at m/z 615.3512 [M + Na]⁺ (calcd for C₃₃H₅₂O₉Na, 615.3504).

Compound 4

White amorphous solid. [α]²⁵_D −5.72 (c 0.07, MeOH). IR (KBr) ν_max 3440, 2966, 2937, 2874, 2854, 1742, 1712, 1633, 1442, 1382, 1236, 1066 cm⁻¹; ¹H NMR (500 MHz, C₅D₅N) and ¹³C NMR (125 MHz, C₅D₅N) data see Tables 1 and 2; HR-ESI-MS at m/z 657.3616 [M + Na] ⁺ (calcd for C₃₅H₅₄O₁₀Na, 657.3609).

Compound 5

White amorphous solid. [α]²⁵_D −61.00 (c 0.10, MeOH). IR (KBr) ν_max 3431, 2967, 2939, 2873, 1738, 1716, 1633, 1427, 1373, 1249, 1173 cm⁻¹; ¹H NMR (500 MHz, C₅D₅N) and ¹³C NMR (125 MHz, C₅D₅N) data see Tables 1 and 2; HR-ESI-MS at m/z 861.4251 [M + Na]⁺ (calcd for C₄₃H₆₆O₁₆Na, 861.4243).

Crystallographic data for 1

C₃₃H₅₂O₉·CH₄O, M = 624.78, a = 12.2949(3) Å, b = 9.7000(2) Å, c = 14.6302(3) Å, α = 90°, β = 110.8410(10)°, γ = 90°, V = 1630.65(6) ų, T = 100.(2) K, space group P1211, Z = 2, μ(Cu Kα) = 0.752 mm⁻¹, 30 722 reflections measured, 6401 independent reflections (R_int = 0.0363). The final R₁ values were 0.0279 (I > 2σ(I)). The final wR(F²) values were 0.0702 (I > 2σ(I)). The final R₁ values were 0.0280 (all data). The final wR(F²) values were 0.0703 (all data). The goodness of fit on F² was 1.046. Flack parameter = −0.02(3). The crystallographic data for the structure of 1 were deposited in the Cambridge Crystallographic Data Centre as supplementary publication CCDC 2201542.†

Crystallographic data for 2

C₃₂H₅₀O₈·H₂O, M = 580.73, a = 6.70730(10) Å, b = 11.6828(3) Å, c = 38.5757(8) Å, α = 90°, β = 90°, γ = 90°, V = 3022.79(11) ų, T = 100. (2) K, space group P212121, Z = 4, μ (Cu Kα) = 0.747 mm⁻¹, 27 775 reflections measured, 5969 independent reflections (R_int = 0.0969). The final R₁ values were 0.0372 (I > 2σ(I)). The final wR(F²) values were 0.0845 (I > 2σ(I)). The final R₁ values were 0.0479 (all data). The final wR(F²) values were 0.0898 (all data). The goodness of fit on F² was 1.041. Flack parameter = 0.14(9). The crystallographic data for the structure of 2 were deposited in the Cambridge Crystallographic Data Centre as supplementary publication CCDC 2201543.†

Crystallographic data for 3

C₃₃H₅₂O₉·CH₄O·H₂O, M = 642.80, a = 12.0479(3) Å, b = 7.6358(2) Å, c = 17.9880(5) Å, α = 90°, β = 94.9550(10)°, γ = 90°, V = 1648.63(8) ų, T = 100.(2) K, space group P1211, Z = 2, μ(Cu Kα) = 0.781 mm⁻¹, 19 289 reflections measured, 6224 independent reflections (R_int = 0.0359). The final R₁ values were 0.0285 (I > 2σ(I)). The final wR(F²) values were 0.0726 (I > 2σ(I)). The final R₁ values were 0.0289 (all data). The final wR(F²) values were 0.0731 (all data). The goodness of fit on F² was 1.056. Flack parameter = 0.03(4). Crystallographic data for the structure of 3 have been deposited in the Cambridge Crystallographic Data Centre as supplementary publication CCDC 2201544.†

Acid hydrolysis of compound 5 and determination of the absolute configuration of the sugar by HPLC

Compounds 5 (1.0 mg) was added to 6 M CF₃COOH (1,4-dioxane-H₂O 1 : 1, 1.0 mL) and heated for 2 h at 99 °C. The reaction mixture, diluted with H₂O (1 mL), was extracted with EtOAc (3 × 2 mL). Next, the dried residue of the aqueous layer was dissolved in pyridine (1.0 mL) mixed with L-cysteine methyl ester hydrochloride (1.0 mg) (Aldrich, Japan) and heated for 1 h at 60 °C. Then, O-tolyl isothiocyanate (5.0 μL) (Tokyo Chemical Industry Co., Ltd, Japan) was added to the mixture and followed by heating for 1 h at 60 °C. Each reaction mixture was directly analyzed by analytical HPLC on a Poroshell 120 SB-C18 column (4.6 × 100 mm, 2.7 μm, Agilent) using an elution of MeCN-H₂O (20 : 75 → 40 : 60, v/v) at a flow rate of 0.6 mL min⁻¹. Under UV detection at 254 nm, the retention time for the standard monosaccharides were: D-glucose (13.85 min) was used for comparison with retention times from reaction mixtures for the saponin. The absolute configuration of the sugar for 5 was identified as D-glucose (13.95 min).

Cytotoxicity in human esophageal cancer cells

The evaluation of cytotoxicity was performed in human esophageal cancer cells EC109 by the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay as previously described.²⁶ Cisplatin was used as the positive control. The experiments were conducted for three independent replicates, and IC₅₀ > 50 μM was considered to be inactive.

Anti-inflammatory activities in mouse macrophages

Anti-inflammatory activities of compounds 1–5 were tested in vitro by measuring the LPS-induced NO production in RAW 264.7 mouse macrophages, and cell viability was determined using the MTT method.²⁷ Cells were seeded in 96-well culture plates and treated with compounds at 50 μM with LPS (1 μg mL⁻¹) for 24 h. The cell culture supernatant was collected and then mixed with an equal volume of Griess reagent I and Griess reagent II, and the absorbance of the samples was read at 540 nm with a microplate reader. The results were described as mean ± SD of two independent experiments.

Anti-angiogenic activities in zebrafish embryos

Transgenic zebrafish Tg (fli1: EGFP) y1, which express enhanced Green Fluorescent Proteins (eGFP) in endothelial cells, were purchased from Zebrafish International Resource Center, Univerisity of Orgeon, USA for the experiments.^28,29 The use of zebrafish for the experiments was approved by the Animal Experimentation Ethics Committee of The Chinese University of Hong Kong (Ref No. 21/247/MIS). At 1-hour post-fertilization, embryos were collected and transferred into a 24-well microplate, with 15 embryos per well. Under the treatment of compounds 1–5 (5 or, 10 μM) for 72 h, photos of the sub-intestinal vessels (SIV) of embryos were taken under a fluorescent microscope (Olympus IX71) by a digital camera (Diagnostic Instruments, Inc., USA) at 100× magnification. The length of the SIV was measured using ImageJ software.²⁸ Experiments were performed twice. Data were expressed as mean ± SEM. Differences among treated groups were determined by one-way ANOVA with Tukey’s post-hoc test.

Conclusions

In summary, five rare cholestane derivatives were isolated from the whole plant of Y. thibetica. Compound 1 is the first cholestane glycoside with two oxygen bridges of C-16/C-23 and C-18/C-23. Compound 2 is the first 23-spirocholestane derivative with a ketal carbon at the C-23 link to C-16 and C-2′ of the D-arabinose ligand, and an oxygen bridge between C-24 and C-1′ of the D-arabinose ligand, whereas compounds 3–5 are rare 23-spirocholestane derivatives with a D-fucose ligand. Cytotoxic assays showed that all isolated saponins did not show any cytotoxic effects on EC109 cells. On the other hand, anti-inflammatory assays showed that compounds 1 and 2 exhibited weak inhibitory activities (17.6% and 18.5%, respectively) at a concentration of 50 μM against LPS-induced NO production in RAW 264.7 macrophages. Furthermore, the formation of SIV in zebrafish embryos was blocked by compound 5 at a concentration of 10 μM, suggesting its potential anti-angiogenic activity.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (U1802287, 32000280, and 32170408), the Ten Thousand Talents Plan of Yunnan Province for Industrial Technology Leading Talents (YNWR-CYJS-2019-011), Yunnan Revitalization Talent Support Program “Top Team” Project, and the Project of Yunnan Characteristic Plant Screening and R&D Service CXO Platform (2022YKZY001). This study was also partly supported by a grant from the State Key Laboratory of Research on Bioactivities and Clinical Applications of Medicinal Plants from Hong Kong SAR, China.

References

1. S. G. Sparg, M. E. Light and J. V. Staden, J. Ethnopharmacol., 2004, 94, 219–243.
2. X. J. Qin, Y. A. Si, Y. Chen, H. Liu, W. Ni, H. Yan, T. Shu, Y. H. Ji and H. Y. Liu, Bioorg. Med. Chem. Lett., 2017, 27, 2267–2273.
3. Z. H. Liu, H. Yan, Y. A. Si, W. Ni, Y. Chen, C. X. Chen, L. He, Z. Q. Zhang and H. Y. Liu, Fitoterapia, 2015, 105, 210–214.
4. Y. S. Chen, Y. He, C. Chen, Y. Zeng, D. Xue, F. Y. Wen, L. Wang, H. Zhang and J. R. Du, Steroids, 2014, 83, 39–44.
5. Y. S. Wen, W. Ni, X. J. Qin, H. Yan, C. X. Chen, Y. Hua, Y. C. Cheng, L. He and H. Y. Liu, Phytochem. Lett., 2015, 12, 31–34.
6. Y. A. Si, H. Yan, W. Ni, Z. H. Liu, T. X. Lu, C. X. Chen and H. Y. Liu, Nat. Prod. Bioprospect., 2014, 4, 315–318.
7. D. W. Deng, D. R. Lauren, J. M. Cooney, D. J. Jensen, K. V. Wurms, J. E. Upritchard, R. D. Cannon, M. Z. Wang and M. Z. Li, Planta Med., 2008, 74, 1397–1402.
8. X. J. Qin, W. Ni, C. X. Chen and H. Y. Liu, J. Ethnopharmacol., 2018, 224, 134–139.
9. Z. H. Liu, H. Yan, Y. A. Si, W. Ni, Y. Chen, L. He and H. Y. Liu, Steroids, 2016, 106, 86–92.
10. The Editorial Board of Flora of China, Flora of China, Science Press, Beijing, 1980, p 17.
11. D. Z. Li, The families and genera of Chinese vascular plants, Science Press, 2020, vol. 1, pp. 354–355.
12. State Administration of Chinese Medicine, Chinese Material Medica (VIII), Shanghai Science and Technology Press, 1999, p. 149.
13. Jiangsu New Medical College, Dictionary of Traditional Chinese Crude Drugs, Shanghai Science and Technology Press, 1977, p. 1841.
14. B. B. Xie, H. Y. Liu, W. Ni and C. X. Chen, Steroids, 2009, 74, 950–955.
15. B. B. Xie, C. X. Chen, Y. H. Guo, Y. Y. Li, Y. J. Liu, W. Ni, L. M. Yang, N. B. Gong, Y. T. Zheng, R. R. Wang, Y. Lu and H. Y. Liu, Planta Med., 2013, 79, 1063–1067.
16. Y. Lu, B. B. Xie, C. X. Chen, W. Ni, H. Yan and H. Y. Liu, Helv. Chim. Acta, 2011, 94, 92–97.
17. Y. Lu, C. X. Chen, W. Ni, H. Yan and H. Y. Liu, Steroids, 2010, 75, 982–987.
18. W. T. Gao, L. L. Yu, J. Xie, L. G. Xiao, S. J. Zhang, W. Y. Ma, H. Yan and H. Y. Liu, Nat. Prod. Bioprospect., 2022, 12, 17.
19. T. X. Lu, T. Q. Shu, J. Xu, W. Ni, Y. H. Ji, Q. R. Chen, A. Khan, Q. Zhao and H. Y. Liu, Steroids, 2017, 123, 55–60.
20. Y. Chen, Y. A. Si, W. Ni, H. Yan, X. J. Qin, C. X. Chen and H. Y. Liu, Nat. Prod. Bioprospect., 2016, 6, 173–182.
21. X. D. Zhang, C. X. Chen, Y. J. Yang, W. Ni and H. Y. Liu, Helv. Chim. Acta, 2012, 95, 1087–1093.
22. T. Asano, T. Murayama, Y. Hiral and J. Shoji, Chen. Pharm. Bull., 1993, 41, 566–570.
23. P. K. Agrawal, Phytochemistry, 1992, 31, 3307–3330.
24. T. Iguchi, M. Kuroda, R. Naito, T. Watanabe, Y. Matsuo, A. Yokosuka and Y. Mimaki, Molecules, 2017, 22, 1243.
25. Shamsuzzaman, M. S. Khan, M. Alam, Z. Tabassum, A. Ahmad and A. U. Khan, Eur. J. Med. Chem., 2010, 45, 1094–1097.
26. T. Mosmann, J. Immunol. Methods, 1983, 65, 55–63.
27. D. Li, H. Liu, W. Ni, W. L. Xiao, L. He, Z. Y. Guo, X. J. Qin and H. Y. Liu, Phytochemistry, 2019, 168, 1–10.
28. G. G. L. Yue, B. C. L. Chan, H. F. Kwok, Y. L. Wong, H. W. Leung, C. J. Ji, K. P. Fung, P. C. Leung, N. H. Tan and C. B. S. Lau, Eur. J. Med. Chem., 2013, 59, 243–252.
29. P. Li, G. G. L. Yue, H. F. Kwok, C. L. Long, C. B. S. Lau and E. J. Kennelly, J. Agric. Food Chem., 2017, 65, 8348–8355.

---

Footnote

† Electronic supplementary information (ESI) available: IR, HRESIMS, and NMR (¹H NMR, ¹³C NMR, DEPT, ¹H−¹H COSY, HSQC, HMBC, and ROESY) spectra of compounds 1–5. CCDC 2201542–2201544. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d2nj04576a

---