Separation and Bioactive Assay of 25R/S-Spirostanol Saponin Diastereomers from Yucca schidigera Roezl (Mojave) Stems

Abstract

In order to find a simple, generic, efficient separation method for 25R/S-spirostanol saponin diastereomers, the liquid chromatographic retention behaviors of C₁₂ carbonylation and C₁₂ unsubstituted 25R/S-spirostanol saponin diastereomers on different stationary phases (C₈, C₁₈, C₃₀ columns) and different mobile phases (MeOH-1% CH₃COOH and CH₃CN-1% CH₃COOH) were investigated. A C₃₀ column was firstly found to offer the highest efficiency for the separation of this kind of diastereomers than C₈ and C₁₈ columns. Meanwhile, the analysis results indicated that both CH₃CN-1% CH₃COOH and MeOH-1% CH₃COOH eluate systems were selective for C₁₂ unsubstituted 25R/S-spirostanol saponin diastereomers, while MeOH-1% CH₃COOH possessed better selectivity for C₁₂ carbonylation ones. Using the abovementioned analysis method, six pairs of 25R/S-spirostanol saponin diastereomers 1a–6a and 1b–6b from Yucca schidigera Roezl (Mojave) were isolated successfully by using HPLC on C₃₀ column for the first time. Among them, three pairs were new ones, named as (25R)-Yucca spirostanoside E₁ (1a), (25S)-Yucca spirostanoside E₁ (1b), (25R)-Yucca spirostanoside E₂ (2a), (25S)-Yucca spirostanoside E₂ (2b), (25R)-Yucca spirostanoside E₃ (3a), (25S)-Yucca spirostanoside E₃ (3b), respectively. Moreover, 3a, 5a, 6a, 3b–6b showed strong inhibitory activities on the growth of SW620 cell lines with the IC₅₀ values of 12.02–69.17 μM.

1. Introduction

Yucca schidigera Roezl (Mojave), belonging to the Yucca genus (Agavaceae family), is mainly distributed in the southwestern United States and the northern desert of Mexico. Steroidal saponins and phenolic acids are reported to be its main constituents. Because of its excellent biological activity and proven safety, it has used worldwide as a kind of additive in foods, beverages, cosmetics and feeds [1]. It is worth mentioning that extracts of Y. schidigera which are enriched in steroidal saponins [2], have already been developed into a commodity for a wide range of applications.

One of the main steroidal saponin aglycone types in Y. schidigera are the C₂₇ spirostanol type steroidal saponins, which can be divided into spirostanol type (25S) and isospiritol type (25R) according to the configuration at C₂₅. The successfully separation of 25R/S-spirostanol saponin diastereomers has scarecely been reported until now, although the different stereo configurations may cause complete different bioactivity and lead to unreliable results. Therefore, the successful separation of 25R/S-spirostanol saponin diastereomers and the determination of their configurations play a crucial role in further pharmacological or molecular biological research on these compounds.

The objective of this study was to establish a simple, generic, efficient separation and analysis method for the 25R/S-spirostanol saponin diastereomers. During the process, the separation ability of three stationary phase (C₈, C₁₈ and C₃₀ columns) as well as two kinds of mobile phases (MeOH-1% CH₃COOH and CH₃CN-1% CH₃COOH) were evaluated. As a result, the separation of 25R/S-spirostanol saponin diastereomers was accomplished by a C₃₀ column, and six pairs of 25R/S-spirostanol saponin diastereomers 1a–6a and 1b–6b were thus obtained, the structures of which were identified by spectroscopy and chemical methods. What’s more, the potent in vitro inhibitory effects of these compounds on human colon cancer cells SW620 were assessed by the MTT method.

2. Results and Discussion

2.1. Selection of Separation Conditions for 25R/S-Spirostanol Saponin Diastereomers by HPLC

On the basis of optimizing stationary and mobile phases, 25R/S-spirostanol saponin diastereomer mixtures 1–6 were isolated successfully by using the C₃₀ column which possessed significant advantages for separating diastereomers. As results, six pairs of 25R/S-spirostanol saponin diastereomers were obtained, and their structures were elucidated to be new ones, (25R)-Yucca spirostanoside E₁ (1a), (25S)-Yucca spirostanoside E₁ (1b), (25R)-Yucca spirostanoside E₂ (2a), (25S)-Yucca spirostanoside E₂ (2b), (25R)-Yucca spirostanoside E₃ (3a), (25S)-Yucca spirostanoside E₃ (3b), and known ones, (25R)-5β-spirostan-3β-ol 3-O-β-d-glucopyranoside (4a) [3], asparagoside A (4b) [3], 25(R)-schidigera-saponin D5 (5a) [4], 25(S)-schidigera-saponin D5 (5b) [4], 25(R)-schidigera-saponin D1 (6a) [4], 25(S)-schidigera-saponin D1 (6b) [4] (Figure 1), respectively.

The 25R/S-spirostanol saponin diastereomers mentioned above could be divided into two classes according to their aglycone types: C₁₂ carbonylated (compounds 1a–3a, 1b–3b) and C₁₂ unsubstituted (compounds 4a–6a, 4b–6b) 25R/S-spirostanol saponins. In the course of comparing separation ability of three stationary phase (C₈, C₁₈ and C₃₀ columns) as well as two kinds of mobile phases (MeOH-1% CH₃COOH and CH₃CN-1% CH₃COOH) [using an evaporating light scattering detector (ELSD) detector] for the abovementioned 25R/S-spirostanol saponin diastereomers, we found that the liquid chromatography retention behaviors of two kinds of aglycone type 25R/S-spirostanol saponin diastereomers were different, and the specific rules are summed up in the following subsections.

2.1.1. General Rules and Characteristics HPLC Analysis for C₁₂ Carbonylation 25R/S-Spirostanol Saponin Diastereomers 1a–3a, 1b–3b

When using a C₃₀ column as stationary phase, better separation effect could be obtained on C₁₂ carbonylated 25R/S-spirostanol saponin diastereomers (Figure 2A, Figure 3A and Figure 4A). During the process of optimizing the separation conditions on the C₃₀ column, MeOH-1% CH₃COOH was found to possess better selectivity. Moreover, the longer for the retention time (t_R), the better resolution (Figure 2B, Figure 3B and Figure 4B) was obtained. On the other hand, we found that the t_R of 25R-spirostanol saponins 1a–3a was always shorter than that of 25S ones 1b–3b with the MeOH-1% CH₃COOH eluate system, which was not related to the type or number of substituted sugars.

2.1.2. General Rules and Characteristics HPLC Analysis for C₁₂ Unsubstituted 25R/S-Spirostanol Saponin Diastereomers 4a–6a, 4b–6b

The C₃₀ column was also found to be more suitable for the separation of C₁₂ unsubstituted 25R/S-spirostanol saponin diastereomers (Figure 5A, Figure 6A and Figure 7A) than the C₈ and C₁₈ columns. The difference from the C₁₂ carbonylated 25R/S-spirostanol saponin diastereomers was that both the CH₃CN-1% CH₃COOH and MeOH-1% CH₃COOH eluate systems were selective to this type of compounds, while CH₃CN-1% CH₃COOH system could guarantee the resolution in shorter t_R (Figure 5B, Figure 6B and Figure 7B).

In addition, the t_R of monosaccharide substituted 25R-spirostanol saponin 4a was shorter than that of 25S-spirostanol saponin 4b whether in MeOH-1% CH₃COOH or in CH₃CN-1% CH₃COOH system. As the number of glycosyl groups increase (compounds 5a, 5b, 6a, 6b), the t_R of 25R-spirostanol saponin was always shorter than that of 25S-spirostanol saponin (5a vs. 5b, 6a vs. 6b) in MeOH-1% CH₃COOH eluate system, a phenomenon in contrast in the CH₃CN-1% CH₃COOH eluate system case.

2.2. Structure Identification for 25R/S-Spirostanol Saponin Diastereomers

Both (25R)-Yucca spirostanoside E₁ (1a) and (25S)-Yucca spirostanoside E₁ (1b) were isolated as white powders with negative optical rotation [([α]${}_{\mathrm{D}}^{25}$ − 11.2°, MeOH) for 1a, ([α]${}_{\mathrm{D}}^{25}$ − 6.6°, MeOH) for 1b]. Their molecular formulae were deduced to be C₃₃H₅₂O₉ by the positive-ion HRESI-MS analysis (m/z 593.3705 [M + H]⁺ for 1a, and 593.3703 [M + H]⁺ for 1b, both calcd. for C₃₃H₅₃O₉, 593.3684). Treatment with 1 M hydrochloric acid (HCl) liberated d-glucose, which was identified by HPLC analysis using an optical rotation detector [5]. Thirty-three carbon signals were displayed in their ¹³C-NMR (

Table 1. ¹³C-NMR data for 1a–3a and 1b–3b in C₅D₅N.

No.	1a	1b	2a	2b	3a	3b	No.	1a	1b	2a	2b	3a	3b
1	30.6	30.6	30.6	30.6	30.6	30.6	21	14.0	13.8	14.0	13.8	13.9	13.8
2	26.7	26.7	26.6	26.7	26.6	26.6	22	109.3	109.8	109.3	109.8	109.3	109.8
3	73.9	73.9	74.0	74.0	74.9	74.9	23	31.8	26.4	31.8	26.4	31.9	26.4
4	30.2	30.2	30.1	30.1	30.7	30.7	24	29.2	26.1	29.2	26.2	29.3	26.2
5	36.5	36.5	36.5	36.5	36.4	36.4	25	30.5	27.5	30.5	27.5	30.6	27.5
6	26.8	26.8	26.8	26.8	26.8	26.8	26	66.9	65.2	66.9	65.2	67.0	65.2
7	26.4	26.4	26.4	26.4	26.4	26.4	27	17.3	16.3	17.3	16.3	17.3	16.3
8	34.7	34.7	34.7	34.7	34.7	34.7	1′	102.9	102.9	102.3	102.4	102.0	101.8
9	41.9	41.9	41.9	41.9	42.0	42.0	2′	75.3	75.3	74.2	74.3	83.1	83.1
10	35.7	35.7	35.7	35.7	35.8	35.8	3′	78.7	78.7	87.7	87.8	78.2	78.2
11	37.7	37.7	37.7	37.8	37.8	37.8	4′	71.7	71.7	69.5	69.5	71.6	71.6
12	213.0	213.0	213.0	213.0	213.0	213.0	5′	78.4	78.4	78.1	78.2	78.3	78.3
13	55.6	55.6	55.6	55.6	55.7	55.6	6′	62.8	62.8	62.3	62.4	62.7	62.7
14	56.0	56.0	56.0	56.0	56.1	56.1	1″			106.3	106.4	106.0	106.0
15	31.5	31.4	31.5	31.4	31.5	31.5	2″			75.3	75.4	77.1	77.1
16	79.8	79.9	79.8	79.9	79.8	79.9	3″			78.2	78.2	78.0	78.0
17	54.3	54.2	54.3	54.2	54.4	54.2	4″			70.9	70.9	71.9	71.9
18	16.1	16.1	16.1	16.1	16.1	16.1	5″			67.4	67.4	78.6	78.6
19	23.0	23.0	23.1	23.1	23.2	23.2	6″					63.0	63.0
20	42.6	43.1	42.6	43.1	42.7	43.2

, C₅D₅N) spectrum. Besides the carbon signals represented by d-glucose, the other twenty-seven ones, especially the quaternary carbon signal at δ_C 109.3 (1a)/109.8 (1b) indicated that they were spirostane-type steroid saponins. The ¹H-NMR spectrum suggested the presence of four methyls [δ 0.70 (3H, d, J = 6.0 Hz, H₃-27), 0.85, 1.09 (3H each, both s, H₃-19 and 18), 1.37 (3H, d, J = 6.5 Hz, H₃-21) for 1a; δ 1.07 (3H, d, J = 6.5 Hz, H₃-27), 0.84, 1.08 (3H each, both s, H₃-19 and 18), 1.37 (3H, d, J = 7.0 Hz, H₃-21) for 1b], two methines bearing an oxygen function [δ 4.32 (1H, m, H-3), 4.55 (1H, q like, ca. J = 8 Hz, H-16) for 1a; δ 4.32 (1H, m, H-3), 4.52 (1H, q like, ca. J = 7 Hz, H-16) for 1b], one oxygenated methene {[δ 3.50 (1H, dd, J = 10.5, 10.5 Hz), 3.60 (1H, dd, J = 4.0, 10.5 Hz), H₂-26] for 1a; [δ 3.38 (1H, br. d, ca. J = 11 Hz), 4.06 (1H, dd, J = 2.5, 11.0 Hz), H₂-26] for 1b} and one β-d-glucopyranosyl [δ 4.93 (1H, d, J = 7.5 Hz, H-1′) for 1a; δ 4.92 (1H, d, J = 7.5 Hz, H-1′) for 1b] in their aglycones. The existence of carbonyl was clarified by the ¹³C-NMR signal at δ_C 213.0 (C-12) (1a/1b). The ¹H-¹H COSY spectra of 1a and 1b suggested the presence of the three partial structures written in bold lines in Figure 8. The planar structure of their aglycons were determined to be spirostan-3-ol-12-one based on the key HMBC correlations from H₂-11, H-14, 17 to C-12; H₃-18 to C-12–14, 17; H₃-19 to C-1, 5, 9, 10; H₃-21 to C-17, 20, 22; H₂-23, 26 to C-22; H₃-27 to C-24–26. Moreover, the β-d-glucopyranosyl was determined to link at C-3 position of aglycone by the long-range correlation from H-1′ to C-3 observed in the HMBC experiment. Their ¹H- and ¹³C-NMR data for the protons and carbons in A–E ring were identical to those of Yucca spirostanoside C₁ [5], and the configuration of A–E ring was determined. Comparing the proton chemical shifts, we found CH₃-27 of 1a (δ 0.70) displayed signal at the higher field than that of 1b (δ 1.07); what’s more, there was a smaller difference between the two protons of CH₂-26 of 1a (∆δ_a,b = 0.10 ppm) than that of 1b (∆δ_a,b = 0.68 ppm). According to the rules summarized by Boll et al. [6] and Schreiber et al. [7], the absolute configuration of C-25 was elucidated to be R and S for 1a and 1b, respectively. On the other hand, the comparision resultsof their ¹³C-NMR data for F ring (C-22–26) and C-27 [δ 17.3 (C-27), 29.2 (C-24), 30.5 (C-25), 31.8 (C-23), 66.9 (C-26), 109.3 (C-22) for 1a; δ 16.3 (C-27), 26.1 (C-24), 26.4 (C-23), 27.5 (C-25), 65.2 (C-26), 109.8 (C-2) for 1b] with those of (25R)-5β-spirostan [δ 17.1 (C-27), 28.8 (C-24), 30.3 (C-25), 31.4 (C-23), 66.8 (C-26), 109.2 (C-22)] and (25S)-5β-spirostan [δ 16.1 (C-27), 25.8 (C-24), 26.0 (C-25), 27.1 (C-23), 65.2 (C-26), 109.7 (C-22)] [8], clarified the absolute configuration of C-25 furtherly. On the basis of above mentioned evidence, the structure of 1a and 1b was elucidated to be (25R)-5β-spirostan-3β-ol-12-one 3-O-β-d-gluco- pyranoside and (25S)-5β-spirostan-3β-ol-12-one 3-O-β-d-glucopyranoside, respectively.

The Q-TOF-ESI-MS analysis results indicated that (25R)-Yucca spirostanoside E₂ (2a) and (25S)-Yucca spirostanoside E₂ (2b), (25R)-Yucca spirostanoside E₃ (3a) and (25S)-Yucca spirostanoside E₃ (3b) had the same molecular formula, C₃₈H₆₀O₁₃ and C₃₉H₆₂O₁₄, respectively. Acid hydrolysis reaction experiments proved that 2a and 2b contained d-glucose and d-xylose [5], while only d-glucose existed in 3a and 3b. Their ¹H-, ¹³C- (Table 1, C₅D₅N) and 2D- (¹H-¹H COSY, HSQC, HMBC) NMR spectra suggested that the aglycons of compounds 2a and 3a, 2b and 3b were (25R)-5β-spirostan-3β-ol-12-one, (25S)-5β-spirostan-3β-ol-12-one, respectively [6,7,8]. Meanwhile, the long-rang correlations from H-1′ to C-3; H-1″ to C-3′ could be observed in the HMBC spectra of compounds 2a and 2b; and in the HMBC spectra of 3a and 3b, the correlations from H-1′ to C-3; H-1″ to C-2′ could be observed. Consequently, the structures of 2a, 2b, 3a, and 3b were elucidated.

The structures of known compounds 4a–b [3], 5a–b [4] and 6a–b [4] were identified by comparing their ¹H-, ¹³C-NMR data with references.

2.3. Inhibitory Activities on the Growth of SW620 Cell Lines Study of Extract, Fractions, and Compounds Obtained from Y. schidigera

The inhibitory effects of Y. schidigera 70% EtOH extract, Y. schidigera 95% EtOH eluate, Y. schidigera H₂O eluate, as well as 25R/S-spirostanol saponin diastereomers 1a–6a, 1b–6b on the growth of SW620 cell lines were measured by the MTT method. As the results in

Table 2. The inhibitory effects of Y. schidigera extract, fractions, and 25R/S-spirostanol saponin diastereomers on the growth of SW620 cell.

Sample	IC50	Sample	IC50
Positive control	10.00 ± 0.15	3a	29.81 ± 0.21
Y. schidigera 70% EtOH extract	85.20 ± 0.95	3b	55.90 ± 0.78
Y. schidigera 95% EtOH eluate	93.04 ± 1.21	4a	>100
Y. schidigera H2O eluate	>100	4b	60.26 ± 4.53
1a	>100	5a	63.37 ± 0.70
1b	>100	5b	33.91 ± 1.27
2a	>100	6a	69.17 ± 1.24
2b	>100	6b	12.02 ± 1.43

n = 4; Positive control: 5-FU; IC₅₀: μg/mL for Y. schidigera 70% EtOH extract, Y. schidigera 95% EtOH eluate, and Y. schidigera H₂O eluate; μM for positive control and compounds1a–6a and 1b–6b.

show, Y. schidigera 70% EtOH extract and Y. schidigera 95% EtOH eluate displayed IC₅₀ values as 85.20 and 93.04 μg/mL, respectively. Meanwhile, compound 6b exhibited strong activity with IC₅₀ value of 12.02 μM comparable with that of the positive control 5-fluorouracil (5-FU, IC₅₀ 10.00 μM), and 3a, 5a, 6a, 3b–5b showed the IC₅₀ values of 29.81–69.17 μM. Moreover, through the summary of structure-activity relationships of 25R/S-spirostanol saponin diastereomers 1a–6a, 1b–6b, it could be found that the configuration of C₂₅ had significant influence of the inhibitory activities towards SW620 cells. For C₁₂ unsubstituted 25R/S-spirostanol saponin diastereomers, the bioactivity of 25S-spirostanol saponins was stronger than that of 25R-ones (4b vs. 4a; 5b vs. 5a; 6b vs. 6a); however, it was exactly opposite for C₁₂ carbonylation 25R/S-spirostanol saponin diastereomers (3b vs. 3a). What’s more, the numbers of substituted glycosyls also affected their activities. For example, as the substituted glycosyls increased, C₁₂ unsubstituted 25R/S-spirostanol saponin diastereomers showed stronger inhibitory effects on SW620 cells (6b vs. 5b vs. 4b; 6a vs. 5a vs. 3a).

3. Materials and Methods

3.1. General Information

The following instruments were used to measure physical data: IR spectra were determined on a 640-IR FT-IR spectrophotometer (Varian Australia Pty Ltd., Mulgrave, Australia). Optical rotations were run on an Autopol^® IV automatic polarimeter (l = 50 mm, Rudolph Research Analytical, Hackettstown, NJ, USA). NMR spectra were obtained on a Bruker 500 MHz NMR spectrometer (Bruker BioSpin AG Industriestrasse 26 CH-8117, Fällanden, Switzerland) at 500 MHz for ¹H- and 125 MHz for ¹³C-NMR (internal standard: TMS). Positive-ion HRESI-TOF-MS were recorded on an Agilent Technologies 6520 Accurate-Mass Q-Tof LC/MS spectrometer (Agilent Corp., Santa Clara, CA, USA). High performance liquid chromatography (HPLC) analyses were performed on an Agilent 1260 Infinity system (Agilent Technologies Inc.) equipped with ELSD (Alltech 2000 ES, Chengdu, China).

Column chromatographies (CC) were performed on macroporous resin D101 (Haiguang Chemical Co., Ltd., Tianjin, China), Silica gel (74–149 µm, Qingdao Haiyang Chemical Co., Ltd., Qingdao, China), and ODS (40–63 μm, YMC Co., Ltd., Tokyo, Japan). High performance liquid chromatography (HPLC) columns: Cosmosil 5C₁₈-MS-II (4.6 mm i.d. × 250 mm, Nacalai Tesque, Inc., Kyoto, Japan), Cosmosil C₈-MS (4.6 mm i.d. × 250 mm, Nacalai Tesque, Inc.), Cosmosil PBr (4.6 mm i.d. × 250 mm, Nacalai Tesque, Inc.), Wacopak Navi C₃₀-5 (4.6 mm i.d. × 250 mm, Wako Pure Chemical Industries, Ltd., Osaka, Japan) were used to analyze the mixture. Preparative high performance liquid chromatography (PHPLC) columns: Cosmosil 5C₁₈-MS-II (20 mm i.d. × 250 mm, Nacalai Tesque, Inc.), Wacopak Navi C₃₀-5 (7.5 mm i.d. × 250 mm, Wako Pure Chemical Industries, Ltd.), and Cosmosil PBr (20 mm i.d. × 250 mm, Nacalai Tesque, Inc.) were used to separate the constituents.

3.2. Plant Material

The stems of Y. schidigera were collected from National City, in southwest Califonia, USA, and identified by Dr. Li Tianxiang (The Hall of TCM Specimens, Tianjin University of TCM, Tianjin, China). The voucher specimen was deposited at the Academy of Traditional Chinese Medicine of Tianjin University of TCM (No. 20160301).

3.3. Extraction and Isolation

3.3.1. Extraction and Isolation of 25R/S-Spirostanol Saponin Diastereomer Mixtures 1–6

The dried stems of Y. schidigera (5.0 kg) were refluxed with 70% ethanol-water for three times. Evaporation of the solvent under pressure provided a 70% ethanol-water (800.0 g). The residue (700.0 g) was dissolved in H₂O, and subjected to D101 CC (H₂O → 95% EtOH) to afford H₂O (380.4 g) and 95% EtOH (310.1 g) eluates, respectively.

The 95% EtOH eluate (200.0 g) was subjected to silica gel CC [CH2Cl2 → CH2Cl2-MeOH (100:1 → 100:3 → 100:7 → 5:1 → 3:1 → 2:1, v/v) → MeOH] to afford thirteen fractions (Fr. 1–Fr. 13). Fraction 6 (12.0 g) was separated by ODS CC [MeOH-H2O (30:70 → 40:60 → 50:50 → 60:40 → 70:30 → 80:20 → 100:0, v/v)], and fourteen fractions (Fr. 6-1–Fr. 6-14) were obtained. Fraction 6-12 (800.9 mg) was isolated by PHPLC [MeOH-1% CH₃COOH (75:25, v/v), Cosmosil 5C₁₈-MS-II column] to yield mixtures 1 and 2. Fraction 6-13 (1.2 g) was subjected to silica gel CC [CH₂Cl₂-MeOH (100:3 → 100:5 → 100:7) → MeOH, v/v] to produce nine fractions (Fr. 6-13-1–Fr. 6-13-9). Fraction 6-13-3 (446.3 mg) was isolated by PHPLC [MeOH-1% CH₃COOH (90:10, v/v), Cosmosil 5C₁₈-MS-II column] to provide mixture 4. Fraction 7 (10.0 g) was subjected to PHPLC [MeOH-1% CH₃COOH (80:20, v/v), Cosmosil 5C₁₈-MS-II column] to produce thirteen fractions (Fr. 7-1–Fr. 7-13). Fraction 7-5 (712.6 mg) was separated by PHPLC [CH₃CN-1% CH₃COOH (40:60, v/v), Cosmosil 5C₁₈-MS-II column] and PHPLC [MeOH-1% CH₃COOH (70:30, v/v), Cosmosil 5C₁₈-MS-II column] to obtain mixture 3. Fraction 7-12 (984.6 mg) was separated by PHPLC [MeOH-1% CH₃COOH (95:5, v/v), Cosmosil PBr column] to give mixtures 5 and 6.

3.3.2. Extraction and Isolation of 25R/S-Spirostanol Saponin Diastereomers 1a–6a, 1b–6b by Using C₃₀ Column

Mixture 1 (190.3 mg) was purified by PHPLC [MeOH-1% CH₃COOH (85:15, v/v)] to give (25R)-Yucca spirostanoside E₁ (1a, 37.8 mg) and (25S)-Yucca spirostanoside E₁ (1b, 23.0 mg). Mixture 2 (160.7 mg) was separated by using the same method as that for mixture 1 to gain (25R)-Yucca spirostanoside E₂ (2a, 47.5 mg) and (25S)-Yucca spirostanoside E₂ (2b, 32.0 mg). Mixture 3 (15.0 mg) was separated by HPLC [MeOH-1% CH₃COOH (70:30, v/v)] to yield (25R)-Yucca spirostanoside E₃ (3a, 3.8 mg) and (25S)-Yucca spirostanoside E₃ (3b, 7.2 mg). Mixture 4 was (180.0 mg) isolated by PHPLC [CH₃CN-1% CH₃COOH (80:20, v/v)] to provide (25R)-5β-spirostan-3β-ol 3-O-β-d-gluco- pyranoside (4a, 10.0 mg) and asparagoside A (4b, 36.5 mg). Mixture 5 (10.0 mg) was purified by HPLC [CH₃CN-1% CH₃COOH (40:60, v/v)] to give 25(R)-schidigera-saponin D5 (5a, 2.0 mg) and 25(S)-schidigera-saponin D5 (5b, 4.2 mg). Using the same HPLC condition, 25(R)-schidigera-saponin D1 (6a, 5.5 mg) and 25(S)-schidigera-saponin D1 (6b, 10.0 mg) were obtained.

(25R)-Yucca spirostanoside E₁ (1a): White powder; [α]${}_{\mathrm{D}}^{25}$ − 11.2° (c = 0.97, MeOH); IR ν_max (KBr) cm⁻¹: 3395, 2929, 2871, 1705, 1454, 1379, 1345, 1244, 1161, 1074, 1027, 985, 921, 897, 867. ¹H-NMR (C₅D₅N, 500 MHz): δ 0.70 (3H, d, J = 6.0 Hz, H₃-27), 0.85 (3H, s, H₃-19), 0.97, 1.33 (1H each, both m, H₂-7), 1.09 (3H, s, H₃-18), 1.11, 1.76 (1H each, both m, H₂-6), [1.28 (1H, m), 1.73 (1H, m, overlapped), H₂-1], 1.37 (3H, d, J = 6.5 Hz, H₃-21), 1.42, 1.86 (1H each, both m, H₂-2), 1.47 (1H, m, H-14), 1.57 (2H, m, overlapped, H₂-24), 1.58 (1H, m, overlapped, H-25), [1.61 (1H, m, overlapped), 2.14 (1H, ddd, J = 5.5, 8.0, 11.5 Hz), H₂-15], 1.63, 1.71 (1H each, both m, H₂-23), 1.73 (2H, m, overlapped, H₂-4), 1.75 (1H, m, overlapped, H-9), 1.83 (1H, m, H-8), 1.94 (1H, quin, J = 6.5 Hz, H-20), 2.08 (1H, m, H-5), [2.21 (1H, dd, J = 4.5, 14.0 Hz), 2.37 (1H, dd, J = 14.0, 14.0 Hz), H₂-11], 2.82 (1H, dd, J = 6.5, 8.5 Hz, H-17), [3.50 (1H, dd, J = 10.5, 10.5 Hz), 3.60 (1H, dd, J = 4.0, 10.5 Hz), H₂-26], 3.95 (1H, m, H-5′), 4.05 (1H, dd, J = 7.5, 8.5 Hz, H-2′), 4.27 (1H, m, overlapped, H-3′), 4.27 (1H, m, overlapped, H-4′), 4.32 (1H, m, H-3), [4.40 (1H, dd, J = 5.0, 11.5 Hz), 4.54 (1H, m, overlapped), H₂-6′], 4.55 (1H, q like, ca. J = 8 Hz, H-16), 4.93 (1H, d, J = 7.5 Hz, H-1′); ¹³C-NMR (C₅D₅N, 125 MHz) spectroscopy data: see Table 1. HRESI-TOF-MS: Positive-ion mode m/z 593.3705 [M + H]⁺ (calcd. for C₃₃H₅₃O₉, 593.3684).

(25S)-Yucca spirostanoside E₁ (1b): White powder; [α]${}_{\mathrm{D}}^{25}$ − 6.6° (c = 1.06, MeOH); IR ν_max (KBr) cm⁻¹: 3391, 2930, 2874, 1704, 1454, 1378, 1345, 1269, 1170, 1070, 1025, 988, 920, 897, 849. ¹H-NMR (C₅D₅N, 500 MHz): δ 0.84 (3H, s, H₃-19), [0.96 (1H, m), 1.35 (1H, m, overlapped), H₂-7] 1.08 (3H, s, H₃-18), 1.07 (3H, d, J = 6.5 Hz, H₃-27), [1.11 (1H, m), 1.76 (m, overlapped), H₂-6], [1.27 (1H, m), 1.72 (1H, m, overlapped), H₂-1], [1.33 (1H, m, overlapped), 1.91 (1H, m, overlapped), H₂-23], [1.33 (1H, m, overlapped), 2.13 (1H, m, overlapped), H₂-24], 1.37 (3H, d, J = 7.0 Hz, H₃-21), [1.41 (1H, m, overlapped), 1.86 (1H, m), H₂-2], 1.45 (1H, m, H-14), 1.59 (1H, m, overlapped, H-25), [1.59 (1H, m, overlapped), 2.12 (1H, m, overlapped), H₂-15], 1.72 (m, overlapped, H₂-4), 1.75 (1H, m, overlapped, H-9), 1.82 (1H, m, H-8), 1.88 (1H, m, overlapped, H-20), 2.08 (1H, m, H-5), [2.20 (1H, dd, J = 4.5, 14.5 Hz), 2.36 (1H, dd, J = 14.5, 14.5 Hz), H₂-11], 2.79 (1H, dd, J = 6.5, 8.5 Hz, H-17), [3.38 (1H, br. d, ca. J = 11 Hz), 4.06 (1H, dd, J = 2.5, 11.0 Hz), H₂-26], 3.94 (1H, m, H-5′), 4.05 (1H, dd, J = 7.5, 8.0 Hz, H-2′), 4.25 (1H, m, overlapped, H-3′), 4.25 (1H, m, overlapped, H-4′), 4.32 (1H, m, H-3), [4.40 (1H, dd, J = 5.0, 11.5 Hz), 4.54 (1H, dd, J = 2.5, 11.5), H₂-6′], 4.52 (1H, q like, ca. J = 7 Hz, H-16), 4.92 (1H, d, J = 7.5 Hz, H-1′); ¹³C-NMR (C₅D₅N, 125 MHz) spectroscopy data: see Table 1. HRESI-TOF-MS: Positive-ion mode m/z 593.3703 [M + H]⁺ (calcd. for C₃₃H₅₃O₉, 593.3684).

(25R)-Yucca spirostanoside E₂ (2a): White powder; [α]${}_{\mathrm{D}}^{25}$ − 0.29° (c = 0.67, MeOH); IR ν_max (KBr) cm⁻¹: 3427, 2928, 2869, 1707, 1454, 1375, 1240, 1162, 1074, 1040, 984, 922, 897, 866. ¹H-NMR (C₅D₅N, 500 MHz): δ 0.70 (3H, d, J = 6.0 Hz, H₃-27), 0.87 (3H, s, H₃-19), 1.10 (3H, s, H₃-18), 1.16, 1.81 (1H each, both m, H₂-6), [1.28 (1H, m), 1.71 (1H, m, overlapped), H₂-1], 1.37 (3H, d, J = 7.0 Hz, H₃-21), [1.37 (1H, m, overlapped), 1.81 (1H, m), H₂-2)], 1.48 (1H, m, H-14), 1.57 (2H, m, overlapped, H₂-24), 1.58 (1H, m, overlapped, H-25), [1.60 (1H, m, overlapped), 2.15 (1H, ddd, J = 6.5, 7.5, 13.5 Hz), H₂-15], [1.64 (1H, m), 1.71 (1H, m, overlapped), H₂-23], 1.73 (2H, m, H₂-4), 1.76 (1H, m, H-9), 1.84 (1H, m, H-8), 1.94 (1H, quin, J = 7.0 Hz, H-20), 2.08 (1H, m, H-5), [2.21 (1H, dd, J = 5.0, 14.5 Hz), 2.38 (1H, dd, J = 14.5, 14.5 Hz), H₂-11], 2.82 (1H, dd, J = 6.5, 8.5 Hz, H-17), [3.50 (1H, dd, J = 10.5, 10.5 Hz), 3.60 (1H, dd, J = 3.5, 10.5 Hz), H₂-26], [3.69 (1H, dd, J = 11.0, 11.0 Hz), 4.31 (1H, m, overlapped), H₂-5″], 3.90 (1H, m, H-5′), 4.02 (1H, dd, J = 7.5, 8.0 Hz, H-2″), 4.07 (1H, dd, J = 8.0, 8.5 Hz, H-2′), 4.14 (1H, dd, J = 8.0, 9.0 Hz, H-3″), 4.16 (1H, m, H-4″), 4.18 (1H, dd, J = 9.0, 9.5 Hz, H-4′), 4.20 (1H, m, overlapped, H-3), 4.25 (1H, dd, J = 8.5, 9.0 Hz, H-3′), [4.34 (1H, dd, J = 5.0, 12.0 Hz), 4.48 (1H, dd, J = 2.0, 12.0 Hz), H₂-6′], 4.55 (1H, q like, ca. J = 7 Hz, H-16), 4.91 (1H, d, J = 8.0 Hz, H-1′), 5.29 (1H, d, J = 7.5 Hz, H-1″); ¹³C-NMR (C₅D₅N, 125 MHz) spectroscopy data: see Table 1. HRESI-TOF-MS: Positive-ion mode m/z 725.4117 [M + H]⁺ (calcd. for C₃₈H₆₁O₁₃, 725.4107).

(25S)-Yucca spirostanoside E₂ (2b): White powder; [α]${}_{\mathrm{D}}^{25}$ − 0.45° (c = 0.45, MeOH); IR ν_max (KBr) cm⁻¹: 3398, 2930, 2869, 1707, 1454, 1375, 1246, 1162, 1074, 1041, 986, 919, 896. ¹H-NMR (C₅D₅N, 500 MHz): δ 0.87 (3H, s, H₃-19), [0.97 (1H, m), 1.35 (1H, m, overlapped), H₂-7], 1.08 (3H, d, J = 8.0 Hz, H₃-27), 1.09 (3H, s, H₃-18), 1.16, 1.80 (1H each, both m, H₂-6), 1.28, 1.71 (1H each, both m, H₂-1), 1.34, 2.13 (1H each, both m, overlapped, H₂-24), 1.38 (3H, d, J = 7.0 Hz, H₃-21), 1.40, 1.85 (1H each, both m, overlapped, H₂-2), 1.41, 1.83 (1H each, both m, overlapped, H₂-23), 1.46 (1H, m, H-14), 1.59 (1H, m, H-25), 1.61, 2.13 (1H each, both m, H₂-15), 1.74 (2H, m, overlapped, H₂-4), 1.75 (1H, m, overlapped, H-9), 1.83 (1H, m, overlapped, H-8), 1.89 (1H, m, H-20), 2.09 (1H, m, H-5), [2.20 (1H, dd, J = 5.0, 14.5 Hz), 2.37 (1H, dd, J = 14.5, 14.5 Hz), H₂-11], 2.80 (1H, dd, J = 7.0, 8.5 Hz, H-17), [3.38 (1H, br. d, ca. J = 12 Hz), 4.05 (1H, dd, J = 3.5, 11.5 Hz), H₂-26], [3.70 (1H, dd, J = 11.0, 11.0 Hz), 4.31 (1H, m, overlapped), H₂-5″], 3.91 (1H, m, H-5′), 4.04 (1H, dd, J = 7.5, 8.0 Hz, H-2″), 4.08 (1H, dd, J = 7.5, 8.5 Hz, H-2′), 4.15 (1H, dd, J = 8.0, 9.0 Hz, H-3″), 4.17 (1H, m, H-4″), 4.19 (1H, dd, J = 9.0, 9.0 Hz, H-4′), 4.26 (1H, dd, J = 8.5, 9.0 Hz, H-3′), 4.30 (1H, m, overlapped, H-3), [4.34 (1H, dd, J = 5.0, 11.5 Hz), 4.49 (1H, dd, J = 2.0, 11.5 Hz), H₂-6′], 4.52 (1H, m, H-16), 4.92 (1H, d, J = 7.5 Hz, H-1′), 5.31 (1H, d, J = 7.5 Hz, H-1″); ¹³C-NMR (C₅D₅N, 125 MHz) spectroscopy data: see Table 1. HRESI-TOF-MS: Positive-ion mode m/z 725.4117 [M + H]⁺ (calcd. for C₃₈H₆₁O₁₃, 725.4107).

(25R)-Yucca spirostanoside E₃ (3a): White powder; [α]${}_{\mathrm{D}}^{25}$ − 12.0° (c = 0.50, MeOH); IR ν_max (KBr) cm⁻¹: 3426, 2927, 2870, 1707, 1456, 1376, 1238, 1161, 1072, 1043, 981, 920, 898, 865. ¹H-NMR (C₅D₅N, 500 MHz): δ 0.70 (3H, d, J = 6.0 Hz, H₃-27), 0.93, 1.31 (1H each, both m, H₂-7), 1.01 (3H, s, H₃-19), 1.09 (3H, s, H₃-18), [1.21 (1H, m), 1.84 (1H each, both m), H₂-6], [1.28 (1H, m), 1.72 (1H, m, overlapped), H₂-1], 1.37 (3H, d, J = 7.0 Hz, H₃-21), 1.37, 1.82 (1H each, both m, overlapped, H₂-2), 1.44 (1H, m, H-14), 1.58 (2H, m, overlapped, H₂-24), 1.58 (1H, m, overlapped, H-25), [1.64 (1H, m), 1.72 (1H, m, overlapped), H₂-23], [1.64 (1H, m), 2.12 (1H, ddd, J = 6.0, 7.0, 12.5 Hz), H₂-15], 1.73 (2H, m, overlapped, H₂-4), 1.74 (1H, m, H-9), 1.84 (1H, m, overlapped, H-8), 1.94 (1H, quin, J = 7.0 Hz, H-20), [2.20 (1H, dd, J = 4.5, 14.0 Hz), 2.38 (1H, dd, J = 14.0, 14.0 Hz), H₂-11], 2.25 (1H, m, H-5), 2.82 (1H, dd, J = 6.5, 8.5 Hz, H-17), [3.50 (1H, dd, J = 10.5, 10.5 Hz), 3.59 (1H, dd, J = 4.0, 10.5 Hz), H₂-26], 3.86 (1H, m, H-5′), 3.98 (1H, m, H-5″), 4.09 (1H, dd, J = 8.0, 9.0 Hz, H-2″), 4.18 (1H, dd, J = 9.0, 9.0 Hz, H-4′), 4.23 (1H, dd, J = 7.5, 8.5 Hz, H-2′), 4.26 (1H, dd, J = 9.0, 9.0 Hz, H-3″), 4.28 (1H, m, H-3), 4.32 (1H, m, overlapped, H-3′), 4.32 (1H, m, overlapped, H-4″), [4.34 (1H, dd, J = 4.5, 11.5 Hz), 4.51 (1H, m, overlapped), H₂-6′], [4.50 (1H, m, overlapped), 4.57 (1H, dd, J = 2.0, 12.0 Hz), H₂-6″], 4.55 (1H, m, H-16), 4.93 (1H, d, J = 7.5 Hz, H-1′), 5.40 (1H, d, J = 8.0 Hz, H-1″); ¹³C-NMR (C₅D₅N, 125 MHz) spectroscopy data: see Table 1. HRESI-TOF-MS: Positive-ion mode m/z 777.4037 [M + Na]⁺ (calcd. for C₃₉H₆₂O₁₄Na, 777.4032).

(25S)-Yucca spirostanoside E₃ (3b): White powder; [α]${}_{\mathrm{D}}^{25}$ − 8.2° (c = 2.7, MeOH); IR ν_max (KBr) cm⁻¹: 3408, 2929, 2987, 1704, 1451, 1373, 1171, 1075, 1032, 990, 919, 896, 850. ¹H-NMR (C₅D₅N, 500 MHz): δ 0.92 (3H, s, H₃-19), [0.94 (1H, m), 1.31 (1H, m, overlapped), H₂-7], 1.07 (3H, d, J = 7.5 Hz, H₃-27), 1.08 (3H, s, H₃-18), 1.22, 1.84 (1H each, both m, H₂-6), 1.29, 1.72 (1H each, both m, H₂-1), 1.31, 1.86 (1H each, both m, overlapped, H₂-2), 1.37 (3H, d, J = 7.0 Hz, H₃-21), 1.38, 2.13 (1H each, both m, overlapped, H₂-24), 1.43 (1H, m, H-14), 1.43, 1.91 (1H each, both m, H₂-23), [1.59 (1H, m, overlapped), 2.10 (1H, m), H₂-15], 1.60 (1H, m, overlapped, H-25), 1.74 (1H, m, overlapped, H-9), 1.83 (2H, m, overlapped, H₂-4), 1.84 (1H, m, overlapped, H-8), 1.88 (1H, m, overlapped, H-20), 2.26 (1H, m, H-5), [2.19 (1H, dd, J = 5.0, 14.0 Hz), 2.37 (1H, dd, J = 14.0, 14.0 Hz), H₂-11], 2.79 (1H, dd, J = 6.5, 8.5 Hz, H-17), [3.37 (1H, br. d, ca. J = 12 Hz), 4.06 (1H, dd, J = 2.5, 11.5 Hz), H₂-26], 3.87 (1H, m, H-5′), 3.97 (1H, m, H-5″), 4.09 (1H, dd, J = 8.0, 8.0 Hz, H-2″), 4.18 (1H, dd, J = 9.0, 9.0 Hz, H-4′), 4.23 (1H, dd, J = 7.5, 8.5 Hz, H-2′), 4.24 (1H, m, H-3), 4.26 (1H, dd, J = 8.0, 9.0 Hz, H-3″), 4.31 (1H, m, overlapped, H-3′), 4.31 (1H, m, overlapped, H-4″), [4.34 (1H, dd, J = 5.0, 11.5 Hz), 4.50 (1H, m, overlapped), H₂-6′], [4.49 (1H, m, overlapped), 4.56 (1H, dd, J = 3.0, 12.0 Hz), H₂-6″], 4.51 (1H, m, H-16), 4.93 (1H, d, J = 7.5 Hz, H-1′), 5.40 (1H, d, J = 8.0 Hz, H-1″); ¹³C-NMR (C₅D₅N, 125 MHz) spectroscopy data: see Table 1. HRESI-TOF-MS: Positive-ion mode m/z 777.4035 [M + Na]⁺ (calcd. for C₃₉H₆₂O₁₄Na, 777.4035).

3.4. Acid Hydrolysis of 1a–3a, 1b–3b

The solution of 1a–3a, 1b–3b (each 2.0 mg) in 1 M HCl (1.0 mL) was treated by using the same method as described in reference [5]: They were heated under reflux for 3 h. Then each reaction mixture was detected by CH₃CN–H₂O (75:25, v/v; flow rate 1.0 mL/min). As a result, d-glucose was found from the aqueous phase of 1a–3a, 1b–3b, and d-xylose was detected from 2a and 3b by comparison of their t_R and optical rotation with those of the authentic sample, d-glucose (t_R 12.4 min (positive)), d-xylose (t_R 6.1 min (positive)).

3.5. Bioassay

3.5.1. Materials

SW620 cell line was obtained from Cell Resource Center of Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences & Peking Union Medical College (Beijing, China). Fetal Bovine Serum (FBS) was purchased from Biological Industries (Beit-Haemek, Israel). Roswell Park Memorial Institute (RPMI) 1640 medium was from Corning (Shanghai, China). Penicillin, streptomycin, and methyl thiazolyl tetrazolium (MTT) were ordered from Thermo Fisher Scientific (Shanghai, China). Dimethyl sulfoxide (DMSO) and 5-fluorouracil were purchased from Sigma-Aldrich (St. Louis, MO, USA).

3.5.2. MTT Assay

The inhibitory effects of Y. schidigera 70% EtOH extract, Y. schidigera 95% EtOH eluate, Y. schidigera H₂O eluate, as well as 25R/S-spirostanol saponin diastereomers 1a–6a, 1b–6b were tested for their individual inhibitory activities on the growth of SW620 cell lines. Cell viability in the presence or absence of tested samples (with the positive control, 5-fluorouracil) was determined using the MTT method [9].

The SW620 cell lines were maintained in RPMI 1640 medium supplemented with 10% FBS, 100 U/mL penicillin and 100 μg/mL streptomycin and kept in a humidified atmosphere of 95% air and 5% CO₂ at 37 °C. For cell viability determination, exponentially growing cells were harvested and plated in 96-well plates (5 × 10⁴/mL) in RPMI 1640 medium for 24 h. After the cells had been washed with PBS, the medium was changed to serially diluted test samples in RPMI. After 48 h of incubation, the cells were washed twice with PBS, and MTT solution was added and incubated for 4 h at 37 °C. Then, MTT was removed. After 100 μL DMSO was added, the 96-well plate was shaken (90 R/S) for 5 min at room temperature under avoiding light, then the absorbance was determined at 490 nm by microplate reader. The tested compounds were independently performed four times. Values are expressed as mean ± S.D. The IC₅₀ values were statistically determined using the SPSS 11.0 software (International business machines corporation (IBM Co.), Armonk, NY, USA).

4. Conclusions

Analysis of spirostanol saponins is usually performed by ultra-high performance supercritical fluid chromatography (UHPSFC) and ultra-high performance liquid chromatography (UHPLC) [10], which are not suitable for mass preparation. Until now, successful preparation examples for (25R/S)-spiromeric epimers were accomplished by SFC isolation [11,12] usually. Meanwhile, it is very rare still for preparing this type of R/S mixture using HPLC methods [13,14,15,16]. On the other hand, among the rare isolation examples, all of them are only for the isolation of C₁₂ unsubstituted 25R/S-spirostanol saponin diastereomers, but no reference was found to separate C₁₂ carbonylated 25R/S-ones. Though the emergence of SFC technology provides a new idea for the qualitative and quantitative analysis of drugs, there are still some difficulties for the further promotion and application of this technology [17]: (1) the types of stationary phases are rare, compounds with strong polarity show poor separation; (2) stationary phase, mobile phase and compounds often do not have good compatibility; (3) not all samples have good solubility in CO₂; (4) when combined with mass spectrometry due to the delayed effect of the porous structure of the stationary phase, a large flow rate is often required, so desirable environmental protection objectives (less mobile phase usage) cannot be achieved. Therofore, it is very important to establish a simple, universal, fast and mass separation analysis method for 25R/S-spirostanol saponin diastereomers.

In this paper, the liquid chromatographic retention behaviors of C₁₂ carbonylated and C₁₂ unsubstituted 25R/S-spirostanol saponin diastereomers on three kinds of stationary phases (C₈, C₁₈, C₃₀ columns) and with two kinds of mobile phases (MeOH-1% CH₃COOH and CH₃CN-1% CH₃COOH) were investigated. A C₃₀ column was firstly found to be more suitable for the separation of this kind of diastereomers than C₈ and C₁₈ column. Meanwhile, the analysis results indicated that both CH₃CN-1% CH₃COOH and MeOH-1% CH₃COOH eluate systems were selective to C₁₂ unsubstituted 25R/S-spirostanol saponin diastereomers, while MeOH-1% CH₃COOH possessed better selectivity for C₁₂ carbonylated ones. Compared with SFC, this method is more simple and versatile. Since 25R/S-spirostanol saponin diastereomers are widely distributed in natural herbs, this research provides a rapid and reliable method for the isolation of similar compounds from plant materials.

On the basis of the abovementioned analysis method, six pairs of 25R/S-spirostanol saponin diastereomers 1a–6a, 1b–6b were isolated from Y. Schidigera succefully. Among them, 1a–3a, 1b–3b were new compounds and 4a and 4b were isolated from Y. schidigera for the first time. The NMR spectrum of compounds: 1a–3a, 1b–3b are in the Supplementary Material.

On the other hand, the study of inhibitory activity and structure-activity relationships of 25R/S-spirostanol saponin diastereomers from Y. schidigera on the growth of SW620 cells will lay a foundation for the further mechanism studies, and it will supply an example for searching for anti-colon cancer drugs in natural products. What’s more, the summary of the structure-activity relationships affords a basis for structural modification, and semi- or total synthesis of new anti-cancer drugs.