Identification and Structural Analysis of Spirostanol Saponin from Yucca schidigera by Integrating Silica Gel Column Chromatography and Liquid Chromatography/Mass Spectrometry Analysis

Yucca schidigera Roezl (Mojave), a kind of ornamental plant belonging to the Yucca genus (Agavaceae), whose extract exhibits important roles in food, beverage, cosmetic and feed additives owing to its rich spirostanol saponins. To provide a comprehensive chemical profiling of the spirostanol saponins in it, this study was performed by using a multi-phase liquid chromatography method combining a reversed phase chromatography T3 column with a normal phase chromatography silica column for the separation and an ESI-Q-Exactive-Orbitrap MS in positive ion mode as the detector. By comparing the retention time and ion fragments with standards, thirty-one spirostanol saponins were identified. In addition, according to the summary of the chromatographic retention behaviors and the MS/MS cleavage patterns and biosynthetic pathway, another seventy-nine spirostanol saponins were speculatively identified, forty ones of which were potentially new ones. Moreover, ten novel spirostanol saponins (three pairs of (25R/S)-spirostanol saponin isomer mixtures) were targeted for isolation to verify the speculation. Then, the comprehensive chemical profiling of spirostanol saponins from Y. schidigera was reported here firstly.


Introduction
Yucca schidigera Roezl (Mojave), belonging to the genus Yucca, Agavaceae family, is mainly distributed in the southwestern United States and the northern deserts of Mexico [1]. It is a commonly used commercial raw material for steroid saponins, which mainly contains spirostanol, isospirostanol and furostanol saponins. Modern pharmacological studies have shown that the extract of Y. schidigera possessed    * The compounds unambiguously identified with the reference standards comparison; # The potential new compounds; The underline annotation indicated that the structures of compounds were proved by targeted isolation.

Rationale for Structural Characterization
According to the references [1,6,10,12,18], the spirostanol saponins from YS were composed by ten kinds of spirostanol aglycons (YS-1-YS-10, Figure 2) linked with nine kinds of glycosyl groups (A 1 -A 9 , Figure 2, Table S8) which was made up by glucopyranosyl (Glc), galactopyranosyl (Gal) and xylopyranosyl (Xyl) at C-3 of the aglycons. Among them, YS-7, YS-9 and YS-10 were only mentioned in references and haven t been obtained by our lab. The occurrence probability of these three kinds of glycosyl groups was Glc > Gal/Xyl.

Rationale for Structural Characterization
According to the references [1,6,10,12,18], the spirostanol saponins from YS were composed by ten kinds of spirostanol aglycons (YS-1-YS-10, Figure 2) linked with nine kinds of glycosyl groups (A1-A9, Figure 2, Table S8) which was made up by glucopyranosyl (Glc), galactopyranosyl (Gal) and xylopyranosyl (Xyl) at C-3 of the aglycons. Among them, YS-7, YS-9 and YS-10 were only mentioned in references and haven′t been obtained by our lab. The occurrence probability of these three kinds of glycosyl groups was Glc > Gal/Xyl. At the same time, the glycosyls directly linked with aglycons (DG) were limited to Glc and Gal, and whose 2′, 3′ and/or 6′-postions were easily substituted by other kinds of glycosyls, and the characteristic types are listed in Table 1.

Study on MS/MS Cleavage Pattern of Spirostanol Saponins from YSESs
According to what we have summarized, the aglycons of the spirostanol saponins from YSESs were usually substituted by -OH, thus dehydration reactions often happened, and a series of neutral 18 Da losses were produced. Moreover, the E-ring cleavage of ∆ 25(27) Table S9) could be used to quickly identify the different types of spirostanol saponins from YSESs. The two pairs of aglycons YS-2 and YS-4 which exhibited the same molecular and fragment ion composition, could be distinguished by observing the relative intensity of the m/z 283.2420 ion. In detail, because of the 12-OH substitution of YS-2, in its MS/MS spectrum, the intensity of m/z 283.2420 ion was stronger than that of m/z 289.2162, while it was opposite of what was observed in the spectrum of 12-H2 substituted YS-4. This phenomenon has been observed in the spectrum of YS-9 as well. Although a sample possessing the YS-7 aglycon was lacking, we designed a targeted separation, in order to verify the above rule. At the same time, the glycosyls directly linked with aglycons (DG) were limited to Glc and Gal, and whose 2 , 3 and/or 6 -postions were easily substituted by other kinds of glycosyls, and the characteristic types are listed in Table 1.

Study on MS/MS Cleavage Pattern of Spirostanol Saponins from YSESs
According to what we have summarized, the aglycons of the spirostanol saponins from YSESs were usually substituted by -OH, thus dehydration reactions often happened, and a series of neutral 18 Da losses were produced. Moreover, the E-ring cleavage of ∆ 25(27) (Figure 3, Table S9) could be used to quickly identify the different types of spirostanol saponins from YSESs. The two pairs of aglycons YS-2 and YS-4 which exhibited the same molecular and fragment ion composition, could be distinguished by observing the relative intensity of the m/z 283.2420 ion. In detail, because of the 12-OH substitution of YS-2, in its MS/MS spectrum, the intensity of m/z 283.2420 ion was stronger than that of m/z 289.2162, while it was opposite of what was observed in the spectrum of 12-H 2 substituted YS-4. This phenomenon has been observed in the spectrum of YS-9 as well. Although a sample possessing the YS-7 aglycon was lacking, we designed a targeted separation, in order to verify the above rule. Molecules 2020, 25

Study on MS/MS Cleavage Pattern of Spirostanol Saponins from YSESs
By comparing the chromatographic elution order of reference standards (peaks 16, 17, 31, 32, 35,  40, 41, 51, 52, 60, 66, 83-86, 100, 104), it was found that the retention times of this class of compounds were mainly affected by the substitution type of the C-12 and C-25 of the aglycon, the stereochemistry of 25-CH 3 , as well as the type of substituted glycosyl (Table S10).

Structural Elucidation of Spirostanol Saponins from YSESs
Using the rules summarized above, another seventy-nine spirostanol saponins have been tentatively identified. In this section, the tentative identification of ten novel 25(R/S)-spirostanol saponins would be described in detail. What s more, the speculation would be verified by their subsequent targeted isolation.
In  Figure S8). The relative molecular mass of its aglycon was 2 Da more than that of YS-5. The ions of m/z 333.2435, 315.2324, 285.1842 formed by the cleavage of E ring were similar to those of YS-5 as well, which indicated that the A-E ring was the same as in YS-5. As a result, the aglycon of peak 10 was tentatively assumed to be 25(R/S)-5β-spirostan-2β,3β-ol-12-one spirostanol. It was speculated that its glycosyl substituent consisted of two molecules of hexose and one pentose molecule because of its neutral loss of 294 Da and 162 Da. According to the biosynthetic pathway of Y. schidigera, the structure of peak 10 could be tentatively speculated to be either 25(R/S)-5β-spirostan-2β,3β-ol-12-one-3-O-xylopyranosyl(1→3)-[glucopyranosyl(1→2)]-glucopyranoside or 25(R/S)-5β-spirostan-2β,3β-ol-12-one-3-O-xylo-pyranosyl (1→3)-[glucopyranosyl(1→2)]-galactopyranoside. In order to clarify the correctness of above speculation, a targeted separation was conducted, and peak 10 was finally unambiguously identified as a novel compound, named as 25(R)-5β-spirostan-2β,3β-ol-12-one  (Table 1, Figures S9, S10, S12, S13, S15 and S16). According to what we have summarized in the Section "Study on MS/MS Cleavage Pattern of Spirostanol Saponins from YSESs", as the intensities of all their m/z 283.2420 peaks were stronger than those of the m/z 289.2162 ones, their aglycon was proposed to be the 12-OH isomer of YS-9, and in particular, it might be YS-7. According to their neutral ion losses  (Table 1, Figures S11, S14 and S17), which indicated that their aglycon was YS-8. )] suggested they were substituted by two hexose molecules and one pentose molecule, two hexose molecules and one pentose molecule and two hexose molecules, respectively. By referring to references as well as comparing them to reference standards, peaks 24 and 32 were the isomers of peaks 27 and 38, which have been identified to be 25(R)-5β-spirostan-3β-ol-12-  (Table 1) were used to identify another seventy-nine spirostanol saponins, among which, thirty of them were potentially new ones. 10, 14, 15, 27, 34, 36, 38, 40, 41, 51 For the purpose of verifying the above speculations, targeted separation of peaks 10, 14, 15, 27, 34,  36, 38, 40, 41, 51 was carried out. In the light of the MPLC-MS analysis results, peak 10 was found to be enriched in the "Preparation of the YSESs Test Solutions" fraction 9; peaks 14, 15 and 38 were found to be enriched in the "Preparation of the YSESs Test Solutions" fraction 8; peak 27 was found to be enriched in the "Preparation of the YSESs Test Solutions" fraction 7; peaks 34, 36, 40, 41 and 51 were found to be enriched in the "Preparation of the YSESs Test Solutions" fraction 6. Then silica gel, ODS CC and preparative HPLC (pHPLC) were used to isolate these fractions and the spirostanol saponins    (Tables S2, S4 and S6, respectively) spectra of 14/15, 34/36 and 40/41 indicated that the three pairs of spirostanol saponins possessed the same aglycon, which was very similar to that of 10, 25(R/S)-5β-spirostan-3β,12β-diol. The difference was that the 2-OH was lacking in the structures of 14/15, 34/36 and 40/41, meanwhile the 12-C=O of 10 has been changed into a 12-OH group. The long-range correlations from H-14, 17 and H 3 -18 to C-12 observed in their HMBC spectra as well as the correlations between H 2 -2 and H-3 displayed in their 1 H-1 H COSY spectra verified the correctness of the above speculation. By comparing the C-22-26 and 27 carbon signals of 10, 14/15, 34/36 and 40/41 they were determined to be 25R and 25S isomer mixtures. As their 1 H-and 13 C-NMR data were consistent with those of 25(R and S) schidigera-saponin F 1 [6], their aglycon was identified as 25(R/S)-5β-spirostan-3β,12β-diol. Acid hydrolysis yielded d-galactose, d-glucose, and d-xylose; d-glucose and d-xylose; and d-glucose, respectively [1]. According to their 1 H-and 13 C-NMR spectra (Tables S2, S4 and (Tables S5, S7 and S9, C 5 D 5 N) data with each other, that was identical to the speculation in Section 2.2.3. The 1 H-, 13 C-NMR and 2D-NMR ( 1 H 1 H COSY, HSQC, HMBC) spectra of the aglycon indicated that it was similar to that of 14/15, the difference being that the 12-OH group was lacking while a 12-C=O appeared in the structures of these three compounds. This speculation was proven by the long-range correlations from H-14, H-17 and H 3 -18 to C-12 observed in their HMBC spectra (Figure 4). In addition, the pentose signals of 27 and 38 were changed as well. According to the literature, the pentose in the structures of 27 and 38 was elucidated to be β-d-apiofuranosyl [19]. The HSQC, and HSQC-TOCSY combined with COSY spectra were used for the assignment of glycosyl units. Their connections were identified by the long-range correlations from H-1 to C-3, H-1" to C-2 , H-1 to C-3 observed in their HMBC spectra. For 51, there were thirty-nine carbon signals in the 13 C-NMR spectrum. Aside from the twenty-seven carbons belonging to the aglycon, twelve carbons were left to assign. Acid hydrolysis only yielded d-glucose The appearance of a β-d-apiofuranosyl moiety provided new possibilities for speculation about the structure of other compounds.

Targeted Separation of Peaks
These results not only confirmed the speculation based on chromatographic retention behaviors and the MS/MS cleavage patterns, but also made supplemented the assignment of the spirostanol saponin glycosyl substituents of Y. schidigera.
In general, based on the phytochemistry researches reported before, this study accomplished a comprehensive chemical profiling of the spirostanol saponins in Y. schidigera, especially, the targeted isolation experiment made the analysis results more convincing. Moreover, the rules summarized in this paper also provided more accurate references to identify this kind of compounds.

Preparation of Standard Solutions
Standard test solutions of above-mentioned references were prepared in MeOH at a final concentration of approximately 100 ng/mL. All stock solutions were stored at 4 • C in darkness and brought to room temperature before use.

Preparation of YS Test Solutions
An aliquot of 1 kg dried powder of Y. schidigera stems was extracted under reflux in 8, 6, 6 L 70% ethanol (v/v) for 3, 2, 2 h, respectively. The extract was combined and 50 mL was filtered with 0.22 µm microporous membrane to obtain YS test solutions. The rest of them was reserved as YS stock solution. YS test solutions were all stored at 4 • C in darkness and brought to room temperature before use.

Liquid Chromatography Setup
Separation of spirostanol saponins was performed on a Thermo UltiMate 3000 UHPLC instrument equipped with a quaternary pump, an autosampler. After the optimization of stationary phase (BEH C 18 , HSS C 18 , C 18  On the other hand, the chromatographic retention behaviors and the MS/MS cleavage patterns of spirostanol saponins were summarized. In summary, according to the retention time (t R ) and the exact mass-to-charge ratio (m/z), thirty-one compounds were unambiguously identified by comparing them to references. Meanwhile, the MS/MS fragmentation pattern and chromatographic elution order rules have been generalized by using the standard compounds as references, seventy-nine compounds were tentatively identified and forty of them were potential new ones. Among them, ten were targeted for separation to prove the correctness of our speculations. During the verification process, the appearance of a β-d-apiofuranosyl moiety was found for the first time, which provided new possibilities for speculation about the structure of other compounds. As a result, an accurate and comprehensive chemical composition profiling of the aerial part of the Y. schidigera was realized, which lays a foundation for the quality evaluation of the plant.
Supplementary Materials: Supplementary data (Materials and Methods section, BPCs for conditional optimization experiment, 1D, 2D NMR, and HRMS/MS spectra of compounds (10, 14, 15, 27, 34, 36, 38, 40, 41 and 51). The summary of characteristic types of substituted glycosyl groups, characteristic fragment ions of seven aglycone moieties and the chromatographic elution order of spirostanol saponins from YSSs aglycons.) associated with this article can be found in the online version.

Conflicts of Interest:
The authors declare no conflict of interest.