Ypsiyunnosides A–E, Five New Cholestanol Glycosides from Ypsilandrayunnanensis

Abstract Phytochemical investigation on the whole plants of Ypsilandra yunnanensis was carried out for the first time and led to the isolation of five new cholestanol glycosides, ypsiyunnosides A–E (1–5), and one known analogue. Their structures were determined mainly by detailed spectroscopic analysis, including extensive 1D and 2D NMR, MS and UV, as well as chemical methods. Among them, compound 1 possessed a rare 6/6/6/5/5 fused-rings cholestanol sketelon, which was identified as (23R,25R)-3β,16α,26-triol-16,23-cyclocholest-5,17(20)-dien-22-one. Their induced platelet aggregation activities and cytotoxicities were evaluated. Graphical Abstract Five new cholestanol glycosides, ypsiyunnosides A–E (1–5), were isolated from the whole plants of Ypsilandra yunnanensis. Compound 1 possessed a rare 6/6/6/5/5 fused-rings cholestanol sketelon. Their structures were elucidated by a combination of 1D and 2D NMR, MS and chemical analysis. Electronic supplementary material The online version of this article (doi:10.1007/s13659-016-0098-2) contains supplementary material, which is available to authorized users.

antifungal, hemostatic, and anti-HIV activities [2][3][4][5][6][7][8][9]. However, the phytochemicals and the biological activity of Y. yunnanensis have not been reported so far. The HPLC analysis revealed that the secondary metabolites of Y. thibetica and Y. yunnanensis were very different, especially in the water-soluble part of the crude extract. As part of our continuing search for structurally diverse and bioactive steroidal glycosides from the Ypsilandra plants, the chemical constituents of Y. yunnanensis were investigated. The result led to the isolation of five new cholestane glycosides (1)(2)(3)(4)(5) and one known analogue, parispseudoside C (6) [10] as shown in Fig. 1 from the 70 % EtOH extract of the whole title plant. Their structures were determined by the analysis of spectroscopic data (HR-ESI-MS, 1D and 2D NMR) and chemical methods. This paper presents herein the isolation, structural elucidation, and the induced platelet aggregation activities and cytotoxicities of these compounds.

Results and Discussion
Ypsiyunnoside A (1) was obtained as a white amorphous powder, a ½ 23 D -139.6 (c 0.05, MeOH); UV (MeOH) k max (log e) 237 (3.03) and 197 (2.76) nm. Its molecular formula was determined as C 57 H 90 O 26 on the basis of a positive-ion at m/z 1213.5616 [M ? Na] ? (calcd. 1213.5618) in its HR-ESI-MS and 13 C NMR data (Table 2), corresponding to 13°of unsaturation. The IR spectrum of 1 exhibited absorption bands for hydroxyl (3440 cm -1 ) and a,b-unsaturated ketone (1701 and 1658 cm -1 ) functionalities, which was confirmed by the UV absorption at k max (MeOH) 237 nm. The 1 H NMR spectrum of the aglycone moiety of 1 (Table 1) displayed signals for four characteristic steroidal methyls at d H 1.08 (s, Me-19), 1.46 (s, Me-18), 1.83 (s, Me-21), and 1.12 (d, J = 6.5 Hz, Me-27) and an olefinic proton at d H 5.35 (br s). The 13 C NMR and DEPT spectra (Table 2) showed a total of 57 carbon signals, which were classified as 7 methyls, 11 methylenes, 32 methines, and 7 quaternary carbons. Among them, 27 carbon signals were assigned to the aglycone including four methyls at d C 8.6 (q, Me-21), 15 (Table 2), respectively. Acid hydrolysis of 1 gave D-glucose and L-rhamnose, which were determined by GC chromatographic analysis of their  Ypsiyunnosides A-E, Five New Cholestanol Glycosides from Ypsilandra yunnanensis 175 L-cysteine methyl ester-TMS derivates. The above spectroscopic information hinted that compound 1 was a cholestanol pentaglycoside. Comparison of the 1 H and 13 C NMR spectra of the aglycone of 1 with those of parispseudoside C (6) [10] led to find the absence of one carbonyl group, one methylene and the appearance of an oxygen-bearing quaternary carbon (d C 83.0) and one methine (d H 2.44; d C 57.6) in 1. In addition, one carbonyl group, two double bonds and five monosaccharides in compound 1 accounted for 8 degrees of unsaturation, and the remaining five degrees of unsaturation required that the aglycone of 1 be pentacyclic. Therefore, it was supposed that compound 1 was a 23,16aldol condensation product of parispseudoside C (6), which was verified by the 1 H-1 H COSY and HMBC spectra.  [11].
The b-configurations for the two glucopyranosyls were deduced by large J 1H-2H values ( 3 J 1,2 = 7.4-7.7 Hz) of their anomeric protons, while the anomeric configuration of the three rhamnopyranosyls were assigned as a-configuration from the 13 C NMR data of C-3 00 , C-5 00 , C-3 000 , C-5 000 , C-3 0000 , and C-5 0000 with those of the corresponding carbons of methyl aand b-rhamnopyranoside [12,13]. The sequence of the sugars and the linkage sites to the aglycone were in good agreement with those of 6, which was supported by their almost identical NMR data and the HMBC correlations from H-1 0 (d H 4.97) of 3-Glu to C-3 (d C 78.1) Table 1 continued    Table 2). Inspection of the 1 H and 13 C NMR spectroscopic data of 3 (Tables 1 and 2) with those of 6 revealed their considerable structural similarity. The major differences were observed for the replacement of a methene in 6 by an oxygenated methine (d H 4.04; d C 73.2). The oxymethine was placed at C-7 on the basis of the 1 H-1 H COSY correlation of d H 5.69 (H-6) with d H 4.04 (H-7). The relative configuration of OH-7 was determined to be b-oriented by the chemical shift of C-7 (d C 73.2), while the signals for C-7 would be at d C 64.7 of the 7a-isomer [7]. Other parts were identical to those of 6 based on 2D NMR experiments. Therefore, the structure of ypsiyunnoside C  1051.5090) and 13 C NMR data ( Table 2). The 1 H and 13 C NMR spectroscopic data of 4 (Tables 1 and 2) were similar to those of 6, differing only in the disappearance of glucopyranosyl signals and the upfield shift of C-26 (d C 75.1 ppm ? 67.9 ppm). This indicated that no sugar moiety was attached to C-26. By the detailed analysis of 1D and 2D-NMR data ( Tables 1 and 2 Table 2). The 1 H and 13 C NMR spectroscopic data of 5 showed its aglycone resembling that of compounds 4 and 6. However, the 1 H NMR spectrum (Table 1)  Considering the traditional use of Ypsilandra plants as hemostatic medicine for Yi minority [14] and the cytotoxic activity of steroidal glycosides previously obtained from Y. thibetica [3,6], compounds 1-6 were evaluated for their induced platelet aggregation activities and cytotoxicities against HEK293 and HepG2 human cancer cell lines. Unfortunately, the results showed that none of them had any obviously induced platelet aggregation activity at the concentration of 300 lg/mL and cytotoxic activity against the two human cancer cell lines (HepG2 and HEK293, IC 50 [ 20 lM).

Acid Hydrolysis and GC Analysis
Compounds 1-5 (1-2 mg, each) were refluxed with 2 M HCl (1:1 v/v, 2 mL) on water bath for 2 h. After cooling, the reaction mixture was neutralized with 1 M NaOH and filtered. The filtrate was extracted with CHCl 3 (3 9 2 mL). The aqueous layer was evaporated to dryness. The dried residue was dissolved in 0.5 mL anhydrous pyridine and treated with L-cysteine methyl ester hydrochloride (1.0 mg) stirred at 60°C for 1 h. Trimethylsilylimidazole (0.5 mL) was added to the reaction mixtures, and these were kept at 60°C for 30 min. The supernatants (2 lL) were analyzed by GC, respectively, under the following conditions: H 2 flame ionization detector. Column: 30QC2/AC-5 quartz capillary column (30 m 9 0.32 mm). Column temperature: 180-280°C with the rate of 3°C/min, and the carrier gas was N 2 (1 mL/min); injector temperature: 250°C; split ratio: 1/50. The configurations of D-glucose and L-rhamnose for compounds 1-5 were determined by comparison of the retention times of the corresponding derivatives with those of standard D-glucose and L-rhamnose giving a single peak at 19.01 and 15.43 min, respectively. These assignments of absolute configurations are based on the assumption that the corresponding enantiomeric sugar derivatives of D-cysteinyl methyl ester would in fact be separable from the L-cysteinyl derivatives under our GC conditions.

Platelet Aggregation Assays
Turbidometric measurements of platelet aggregation of the samples were performed in a Chronolog Model 700 Aggregometer (Chronolog Corporation, Havertown, PA, USA) according to Born's method [15,16]. The blood from the rabbits by cardiac puncture, was anticoagulated with 3.8 % sodium citrate (9:1, v/v). Platelet-rich plasma (PRP) was prepared shortly after blood collection by spinning the sample at 180 g for 10 min at 22°C. The PRP was carefully removed and the remaining blood centrifuged at 2400 g for 10 min to obtain platelet-poor plasma (PPP). The centrifuge temperature was maintained at 22°C. Platelet counts were adjusted by the addition of PPP to the PRP to achieve a count of 500 9 10 9 L -1 . Platelet aggregation studies were completed within 3 h of preparation of PRP. Immediately after preparation of PRP, 250 lL was incubated in each of the test tubes at 37°C for 5 min and then 2.5 lL of inducers (or compounds) was added. The change of optical density as a result of platelet aggregation was recorded. The extent of aggregation was estimated by the percentage of maximum increase in light transmission, with the buffer representing 100 % transmittance. Arachidonic acid (AA) was used as a positive control.

Cytotoxic Assay
Compounds 1-6 were evaluated for their cytotoxicities against two human cancer cell lines (HepG2 and HEK293) using the MTT method described in the literature elsewhere [17]. (-)-OddC (Troxacitabine) was used as a positive control with IC 50 values of 0.17 ± 0.02 lM and 0.30 ± 0.03 lM to the two cell lines, respectively. The experiments were conducted in three independent replicates, and IC 50 [ 20 lM was considered to be inactive.