- Email: [email protected]
Steroidal saponins from Yucca gloriosa L. rhizomes: LC–MS profiling, isolation and quantitative determination
Phytochemistry 72 (2011) 126–135
Contents lists available at ScienceDirect
Phytochemistry journal homepage: www.elsevier.com/locate/phytochem
Steroidal saponins from Yucca gloriosa L. rhizomes: LC–MS profiling, isolation and quantitative determination Alexandre Skhirtladze a, Angela Perrone b, Paola Montoro b, Mariam Benidze a, Ether Kemertelidze a, Cosimo Pizza b, Sonia Piacente b,⇑ a b
Iovel Kutateladze Institute of Pharmacochemistry, P. Sarajishvili 36, 0159 Tbilisi, Georgia Dipartimento di Scienze Farmaceutiche, Università degli Studi di Salerno, Via Ponte Don Melillo, I-84084 Fisciano, Italy
a r t i c l e
i n f o
Article history: Received 30 July 2010 Received in revised form 18 October 2010 Accepted 19 October 2010 Available online 19 November 2010 Keywords: Yucca gloriosa Steroidal glycosides LC–MS profiling Quantitative determination
a b s t r a c t The occurrence of steroidal saponins in the rhizomes of Yucca gloriosa has been detected by LC–MS. On the basis of the LC–MS analysis, five steroidal glycosides, including three spirostane, one furostane and one cholestane glycosides, along with seven known compounds have been isolated and characterized by ESI-MS and by the extensive use of 1D- and 2D-NMR experiments. Quantitative analysis of the steroidal glycosides in Y. gloriosa rhizomes was performed by an LC–MS method validated according to European Medicines Agency (EMEA) guidelines. The dried BuOH extract obtained from rhizomes contains more than 25% w/w of glycosides, thus Y. gloriosa rhizomes can be considered a rich source of steroidal glycosides. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction Yucca gloriosa L. is largely cultivated in Eastern Georgia, where industrial plantations of this species occupy a total area of about 150 ha (Benidze et al., 1991). Y. gloriosa produces a strong underground system, whose main part is represented by rhizomes. In the rhizomes, biosynthesis of 5b steroids occurs, whereas in leaves and flowers only 5a steroids (Skhirtladze et al., 2006; Kemertelidze and Benidze, 2001) have been detected. The roots and bark of Y. gloriosa are free from saponins, but they are rich in phenolic constituents related to resveratrol, with high antioxidant and antiproliferative activities (Bassarello et al., 2007a,b; Nigro et al., 2007; Montoro et al., 2008). Due to the high content of steroidal saponins (5–6%), the rhizomes of Y. gloriosa are used as raw material for the semisynthesis of steroidal hormones. Continuing our studies on Yucca spp. (Montoro et al., 2008, 2010; Skhirtladze et al., 2006), here we report on the phytochemical investigation of the steroidal saponins of the rhizomes of Y. gloriosa. A preliminary analysis of the BuOH enriched saponin fraction obtained by Y. gloriosa rhizomes was performed by LC–ESI-MS in order to detect the presence of saponins in this part of the plant. Successively, five new steroidal glycosides, including three spirostane, one furostane and one cholestane glycosides, along with seven known compounds were isolated. Finally, an LC–MS/MS ⇑ Corresponding author. Tel.: +39 089 969763; fax: +39 089 969602. E-mail address: [email protected] (S. Piacente). 0031-9422/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.phytochem.2010.10.013
method was developed for the quantitative analysis of steroidal glycosides in the extract of Y. gloriosa rhizomes.
2. Results and discussion From the analytical perspective, it is clear that steroidal saponins are not detectable by HPLC–UV analysis for the lack of a strong UV chromophore and that HPLC analysis of all compounds requires gradient elution. Mass spectrometry represents an effective detection method, and in addition improvement in selectivity and specificity can be raised by using tandem mass spectrometry. LC–MS is selective and sensitive enough to carry out the analysis of saponins (Yan and Guo, 2005; Tor et al., 2005; Montoro et al., 2010). On the basis of these observations, a preliminary analysis of the extract obtained by Y. gloriosa rhizomes was performed by LC–ESI-MS in order to detect the presence of saponins in this part of the plant. The ESI-MS spectra of saponins show adducts together with pseudomolecular ions; in the ESI-MS spectra in positive-ion mode, mainly [M+H]+, [M+NH4]+, [M+Na]+, and [M+K]+ ions can be observed (Miao et al., 2002). Column and mobile phase selection is generally determined by the combination of the compounds to be analyzed and the matrix where they occur. In our work the use of an Atlantis C18 column and a gradient elution allowed us to obtain a good separation of the steroidal glycosides occurring in the BuOH enriched saponin fraction of Y. gloriosa rhizomes, without any other compound interfering with the analyte detection.
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
The profile of the Total Ion Current (TIC) showed 12 main peaks. For these peaks MS spectra were obtained by extracting them from the total profile. Successively, on the basis of the m/z values, Reconstructed Ion Chromatograms (RIC) were extracted. Positive-ion electrospray LC–MS analysis obtained in the acquisition range of 190–1200 amu, TIC profile and RICs are shown in Fig. 1. Twelve main compounds were detected: at m/z 917, compounds 1 and 3; at m/z 915, compound 2; at m/z 1199, compound 4; at m/z 903, compounds 5 and 7; at m/z 757, compounds 6 and 9; at m/z 755, compound 8; at m/z 741, compounds 10–12. On the basis of the results of the on-line screening and identification by LC–MS, compounds 1–12 were isolated and their structures were unambiguously elucidated from NMR spectroscopic data (Figs. 2 and 3). The BuOH extract was fractionated by silica gel column chromatography to yield compounds 10 and 11. The fractions obtained were chromatographed by reversed-phase HPLC to yield compounds 1–9 and 12 (see Section 4). The absolute configurations of the sugar units were assigned after acid hydrolysis of the crude saponin mixture and identification with authentic samples by TLC, followed by preparative separation of each sugar. The D configuration of galactose and glucose and the L configuration of rhamnose were established by comparison of their optical rotation values with those reported in the literature (Belitz et al., 2009; Wang et al., 2008). The HRMALDITOF mass spectrum of 2 (m/z 937.4412 [M+Na]+, calcd. for C45H70O19Na, 937.4409) supported a molecular formula of C45H70O19. The ESI-MS mass spectrum showed the major ion peak
127
at m/z 937.4 which was assigned to [M+Na]+. The MS/MS of this ion showed a peak at m/z 775.4 [M+Na162]+, corresponding to the loss of an hexose unit. In the MS3 spectrum a peak at m/z 613.4 [M+Na162162]+, corresponding to the loss of a second hexose unit, was observed. Finally, in the MS4 spectrum a peak at m/z 451.4 [M+Na162162162]+, indicating the loss of a third hexose unit, was observed. The 1H NMR spectrum of compound 2 showed signals for two tertiary methyl groups at d 1.11 (3H, s) and 1.12 (3H, s), a secondary methyl group at d 1.05 (3H, d, J = 6.6 Hz), exomethylene protons at d 4.81 and 4.72 (each 1H, br s), two methine proton signals at d 4.43 (1H, m) and 4.12 (1H, m), two methylene proton signals at d 4.31 and 3.87 (each 1H, d, J = 12.1 Hz), along with three anomeric protons at d 4.96 (1H, d, J = 7.5 Hz), 4.68 (1H, d, J = 7.5 Hz) and 4.51 (1H, d, J = 7.5 Hz). The 13C NMR spectrum displayed, for the aglycon moiety, signals ascribable to a keto group at d 215.6, a ketal function at d 110.1, two secondary alcoholic functions at d 75.4 and 80.7, and one primary alcoholic function at d 65.3, suggesting as aglycon a spirostanol skeleton characterized by the occurrence of a keto group (Table 1). On the basis of the HSQC and HMBC correlations, the aglycon moiety of compound 2 was identified as 5b-spirost-25(27)-en-3b-ol-12-one or schidigera-genin B, previously isolated from Y. schidigera (Miyakoshi et al., 2000). The chemical shifts of all the individual protons of the three sugar units were ascertained from a combination of 1D-TOCSY and DQF-COSY spectral analysis, and the 13C chemical shifts of their relative attached carbons were assigned unambiguously from the HSQC spectrum (Table 2). These data showed the presence of three bglucopyranosyl units (d 4.96, 4.68 and 4.51). Glycosidation shifts
Fig. 1. Positive-ion mode LC–MS analysis of the BuOH extract of Yucca gloriosa rhizomes. Total ion current and reconstructed ions chromatogram for compounds 1–12.
128
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
Fig. 2. Compounds 2, 4, 6, 7, 8 from Yucca gloriosa L. rhizomes.
were observed for C-3glcI (d 87.6), C-2glcI (d 78.9) and C-3 (d 75.4). An unambiguous determination of the sequence and linkage sites was obtained from the HMBC spectrum, which showed key correlation peaks between the proton signal at d 4.51 (H-1glcI) and the carbon resonance at d 75.4 (C-3), d 4.96 (H-1glcII) and d 78.9 (C-2glcI), and the proton signal at d 4.68 (H-1glcII) and the carbon resonance at d 87.6 (C-3glcI). On the basis of this evidence, the structure of the new compound 2 was established as 5b-spirost-25(27)-en-3b-ol12-one 3-O-{b-D-glucopyranosyl-(1 ? 2)-O-[b-D-glucopyranosyl(1 ? 3)]-b-D-glucopyranoside}. The molecular formula of compound 8 was established as C39H62O14 by HRMALDITOFMS analysis (m/z 777.4041 [M+Na]+,
calcd. for C39H62O14Na, 777.4037). The ESI-MS mass spectrum showed the major ion peak at m/z 777.4 which was assigned to [M+Na]+. In the MS/MS spectrum a peak at m/z 615.3 [M+Na162]+, corresponding to the loss of an hexose unit, was observed. The MS3 spectrum displayed a peak at m/z 453.4 [M+Na162162]+, due to the loss of a second hexose unit. The 1 H and 13C NMR data of the aglycon moiety of 8 compared to those of 2 showed that the two compounds differed only by the replacement of the exomethylene group with a secondary methyl group at C-27 (dH 0.79, dC 17.0). The 1H and 13C NMR data of the aglycon of 8 were in agreement with those reported for (25R)-5b-spirostan-3bol-12-one, named gloriogenin (Nakano et al., 1989, 1991a). For the
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
129
Fig. 3. Compounds 1, 3, 5, 9, 10, 11, 12 from Yucca gloriosa L. rhizomes.
sugar portion, the 1H NMR spectrum showed signals for two anomeric protons at d 4.68 (1H, d, J = 7.5 Hz) and 4.41 (1H, d, J = 7.5 Hz). Complete assignments of the 1H and 13C NMR signals of the sugar portion were accomplished by HSQC, HMBC, DQFCOSY and 1D-TOCSY experiments which led to the identification of one b-galactopyranosyl (d 4.41) unit and one b-glucopyranosyl (d 4.68) unit. Once again, direct evidence of the sugar sequence and the linkage sites was derived from HSQC and HMBC experi-
ments. The glycosidation shifts on C-3 (d 75.8) and C-2gal (d 78.6) indicated the linkage sites. Key correlation peaks were observed in the HMBC spectrum of 8 between the proton signal at d 4.41 (H-1gal) and the carbon resonance at d 75.8 (C-3) and the proton signal at d 4.68 (H-1glc) and the carbon resonance at d 78.6 (C-2gal). Therefore, compound 8 was established as the new (25R)-5b-spirostan-3b-ol-12-one 3-O-b-D-glucopyranosyl-(1 ? 2)O-b-D-galactopyranoside.
130
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
Table 1 13 C and 1H NMR data (J in Hz) of the aglycon moieties of compounds 2, 4, 6–8 (CD3OD, 600 MHz). 2
4
dC
dH
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
30.4 27.3 75.4 30.9 36.7 27.0 26.7 35.7 43.2 36.0 38.4 215.6 56.7 57.6 32.1 80.7 54.8 16.3 23.8 43.2 13.5 110.1
1.92, 1.57 1.71, 1.60 4.12 m 1.89, 1.55 1.93 m 2.00, 1.47 1.25, 1.17 1.80 m 1.65 m
23 24 25 26
33.9 29.0 144.4 65.3
27
108.6
m m m m m
2.25, 2.04 m
1.45 m 2.00, 1.32 m 4.43 m 2.32 m 1.12 s 1.11 s 1.98 m 1.05 d (6.6)
1.78, 1.75 m 2.57, 2.20 m 4.31 d (12.1) 3.87 d (12.1) 4.81, 4.72 br s
6
7
dC
dH
dC
dH
37.8 30.9 79.6 38.8 143.9 127.4 73.3 40.8 49.9 37.7 21.8 40.5 44.0 55.1 39.0 83.3 61.5 13.4 19.1 28.4 19.6 50.5
1.69, 0.79 m 1.88, 1.58 m 3.50 m 2.42, 2.26 m
31.3 27.4 75.9 31.4 36.8 27.8 27.0 37.3 40.3 36.3 37.4 80.6 46.9 56.0 32.7 82.0 62.9 16.3 24.5 43.3 13.5 110.6
1.88, 1.59 1.70, 1.63 4.11 m 1.89, 1.54 1.93 m 2.00, 1.48 1.26, 1.19 1.61 m 1.55 m
1.65, 1.44 m 3.29 m
2.41 (2H) dd (7.1, 2.3) 2.13 m 0.95 d (6.6)
32.3 29.3 31.5 67.6
0.95 d (6.6)
17.3
216.2 53.5 25.1 22.9 22.9
5.24 3.69 1.49 0.94
br d (2.1) dd (8.2, 2.1) m m
1.50 (2H) m 2.06, 1.22 m 1.06 m 2.48, 1.88 m 4.16 m 1.26 dd (11.6, 7.9) 0.95 s 1.04 s 2.55 m 0.99 d (6.6) 2.77 br d (15.4, 2.0) 2.30 dd (15.4, 10.4)
The HRMALDITOF mass spectrum of 6 (m/z 779.4199 [M+Na]+, calcd. for C39H64O14Na, 779.4194) supported a molecular formula of C39H64O14. The ESI-MS mass spectrum showed the major ion peak at m/z 779.4 which was assigned to [M+Na]+. In the MS/MS spectrum a peak at m/z 617.4 [M+Na162]+, corresponding to the loss of an hexose unit, was observed. The MS3 spectrum displayed a peak at m/z 455.3 [M+Na162162]+, due to the loss of a second hexose unit. A comparison of the NMR data of compound 6 with those of compound 8 showed that in the aglycon of 6 the keto group at C-12 was replaced by the 12b-hydroxy group, indicating as aglycon the 12b-hydroxysmilagenin (Nakano et al., 1991b). For the sugar portion signals of two anomeric protons at d 4.68 (1H, d, J = 7.5 Hz) and 4.47 (1H, d, J = 7.5 Hz) were observed in the 1H NMR spectrum. A combination of HSQC, 1D-TOCSY and DQF-COSY spectra led to the identification of two b-glucopyranosyl (d 4.68 and 4.47) units. Key correlation peaks between the proton signal at d 4.47 (H-1glcI) and the carbon resonance at d 75.9 (C-3) and the proton signal at d 4.68 (H-1glcII) and the carbon resonance at d 80.6 (C-2glcI) were observed in the HMBC spectrum. Thus, the structure of the new compound 6 was identified as (25R)-5b-spirostan-3b,12b-diol 3-O-b-D-glucopyranosyl-(1 ? 2)-O-b-D-glucopyranoside. The molecular formula of compound 7 was established as C45H74O18 by HRMALDITOFMS analysis (m/z 925.4778 [M+Na]+, calcd. for C45H74O18Na, 925.4773). The ESI-MS mass spectrum showed the major ion peak at m/z 925.5 which was assigned to [M+Na]+. In the MS/MS spectrum a peak at m/z 763.5 [M+Na162]+, corresponding to the loss of an hexose unit, was observed. The MS3 spectrum displayed a peak at m/z 601.5 [M+Na162162]+, due to the loss of a second hexose unit. The 1 H NMR spectrum of compound 7 showed signals for two tertiary methyl groups at d 0.72, 1.02 (3H, s) and 1.63 (3H, s), a secondary methyl group at d 0.98 (3H, d, J = 6.6 Hz), two methine proton signals at d 4.74 (1H, m) and 4.12 (1H, m), and two methylene proton signals at d 3.35 and 3.82 ascribable to a primary alcoholic func-
8
dC
dH
31.0 26.8 76.0 31.0 37.3 27.3 27.4 36.3 41.2 36.2 21.9 40.9 44.8 56.0 34.9 85.5 65.7 14.8 24.2 105.4 11.5 153.0
1.89, 1.59 1.69, 1.63 4.12 m 1.89, 1.53 1.91 m 1.98, 1.49 1.23, 1.18 1.64 m 1.47 m
1.49, 1.38 m 1.84, 1.32 m
1.64 (2H) m 1.48 (2H) m 1.61 m 3.45, 3.30 m
23.9 31.7 34.1 75.8
2.15 (2H) m 1.64, 1.30 m 1.81 m 3.80, 3.35 m
32.3 29.3 31.1 67.5
1.60 (2H) m 1.48 (2H) m 1.63 m 3.45, 3.30 m
0.82 d (5.8)
17.1
0.98 d (6.6)
17.0
0.79 d (5.8)
m m m m m
1.20 m 2.02, 1.38 m 4.40 m 1.85 m 0.75 s 1.03 s 1.92 m 1.02 d (6.6)
m m m m m
1.16 m 2.19, 1.42 m 4.74 m 2.52 d (10.1) 0.72 s 1.02 s 1.63 s
dC
dH
30.7 27.5 75.8 30.8 36.6 27.2 26.8 35.3 43.1 36.7 37.9 215.8 56.5 57.1 32.2 80.6 54.8 16.0 23.5 43.0 12.9 109.3
1.90, 1.55 1.70, 1.60 4.11 m 1.91, 1.58 1.91 m 1.98, 1.45 1.27, 1.17 1.82 m 1.67 m
m m m m m
2.20, 2.03 m
1.42 m 2.02, 1.35 m 4.34 m 2.29 m 1.07 s 1.08 s 1.91 m 1.02 d (6.6)
tion, along with three anomeric protons at d 4.70 (1H, d, J = 7.5 Hz), 4.48 (1H, d, J = 7.5 Hz) and 4.26 (1H, d, J = 7.5 Hz). The 13 C NMR spectrum displayed for the aglycon signals ascribable to two sp2 carbons at d 105.4 and 153.0, two secondary alcoholic functions at d 76.0 and 85.5, and one primary alcoholic function at d 75.8, suggesting as aglycon a furostanol skeleton (Table 1). On the basis of the HSQC and HMBC correlations, the aglycon moiety of compound 7 was identified as (25R)-5b-furost-20(22)-en3b,26-diol named pseudosmilagenin. The C-25 configuration was deduced to be R based on the difference of chemical shifts (Dab = da db) of the geminal protons at H2-26 (Dab = 0.45 ppm). It has been described that Dab is usually >0.57 ppm in 25S compounds and <0.48 in 25R compounds (Agrawal, 2004). The NMR data of compound 7 in comparison with those reported for macrostemonoside F (Peng et al., 1993) showed that the two compounds differed only by the replacement of the galactose unit with of the glucose unit as inner sugar at C-3. Thus, compound 7 was established as the new (25R)-26-O-b-D-glucopyranosyl-5b-furost20(22)-en-3b,26-diol 3-O-b-D-glucopyranosyl-(1 ? 2)-O-b-D-glucopyranoside. For the first time a pseudosmilagenin glycoside has been isolated from Yucca species. Compound 4 showed in the positive ESI-MS a major ion peak at m/z 1221.6 [M+Na]+ and a significant fragment in MS/MS analysis at m/z 1075.6 [M+Na146]+, ascribable to the loss of a deoxyhexose unit. The MS3 fragmentation showed peak at m/z 1057.5 [M+Na14618]+ due to the loss of a water molecule. Finally, in the MS4 a peak at m/z 911.5 [M+Na14618146]+, indicating the loss of a second deoxyhexose unit, was observed. Its molecular formula was established unequivocally as C59H90O25 by HRMALDITOF mass spectrum (m/z 1221.5672 [M+Na]+, calcd. for C59H90O25Na, 1221.5669). The 1H NMR spectrum of compound 4 displayed signals for two tertiary methyl groups at d 0.95 (3H, s) and 1.04 (3H, s), three secondary methyl groups at d 0.99 (3H, d, J = 6.6 Hz) and 0.95 (6H, d, J = 6.4 Hz), three methine proton signals at d 4.16 (1H, m), 3.50 (1H, m) and 3.69 (1H, dd, J = 8.2, 2.1 Hz), an
131
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135 Table 2 13 C and 1H NMR data (J in Hz) of the sugar portions of compounds 2, 4, 6–8 (CD3OD, 600 MHz). 2 1 2 3 4 5
100.6 78.9 87.6 69.9 78.2
6
62.7
1 2 3 4 5
103.4 75.8 77.9 72.3 77.7
6
63.4
1 2 3 4 5
104.3 75.0 77.9 71.4 77.8
6
62.8
b-D-GlcI 4.51 d (7.5) 3.72 dd (9.0, 7.5) 3.76 dd (9.0, 9.0) 3.38 dd (9.0, 9.0) 3.33 ddd (9.0, 4.5, 2.0) 3.87 dd (12.0, 2.0) 3.68 dd (12.0, 4.5) b-D-GlcII 4.96 d (7.5) 3.13 dd (9.0, 7.5) 3.39 dd (9.0, 9.0) 3.16 dd (9.0, 9.0) 3.32 ddd (9.0, 4.5, 2.0) 3.85 dd (12.0, 2.0) 3.65 dd (12.0, 4.5) b-D-GlcIII 4.68 d (7.5) 3.31 dd (9.0, 7.5) 3.38 dd (9.0, 9.0) 3.31 dd (9.0, 9.0) 3.38 ddd (9.0, 4.5, 2.0) 3.87 dd (12.0, 2.0) 3.67 dd (12.0, 4.5)
4 101.9 75.1 76.5 80.4 74.1 64.6
102.7 72.1 72.1 73.7 70.5
b-D-GlcI 4.46 d (7.5) 3.27 dd (9.0, 7.5) 3.55 dd (9.0, 9.0) 3.55 dd (9.0, 9.0) 3.75 ddd (9.0, 4.5, 2.0) 4.56 dd (12.0, 2.0) 4.56 dd (12.0, 4.5) a-L-RhaI 4.86 d (1.2) 3.92 dd (1.2, 3.2) 3.69 dd (3.2, 9.3) 3.44 t (9.3) 4.02 m
17.7
1.31 d (6.0)
125.5 114.0 151.4 151.1 117.6
Vanillic acid – 7.65 d (2.0) – – 7.24 d (8.1)
124.1
7.68 dd (2.0, 8.1)
7 OMe
167.5 56.2
1 2 3 4 5 6
100.4 71.8 72.1 73.5 71.0 17.7
1 2 3 4 5
106.5 75.2 78.0 71.3 77.6
6
62.7
– 3.92 s a-L-RhaII 5.52 d (1.2) 4.12 dd (1.2, 3.2) 3.92 dd (3.2, 9.3) 3.52 t (9.3) 3.72 m 1.25 d (6.0) b-D-GlcII 4.22 d (7.5) 3.21 dd (9.0, 7.5) 3.37 dd (9.0, 9.0) 3.28 dd (9.0, 9.0) 3.25 ddd (9.0, 4.5, 2.0) 3.87 dd (12.0, 2.0) 3.67 dd (12.0, 4.5)
6 100.7 80.6 77.8 71.5 78.2 62.4
104.2 76.1 77.5 71.7 78.0 62.4
olefinic proton at d 5.24 (1H, br d). The 1H and 13C NMR data of the aglycon portion of compound 4 were in agreement with those reported for 3b,7b,16b-trihydroxycholest-5-en-23-one (Mimaki et al., 1992). Additionally, three aromatic protons at d 7.68 (1H, dd, J = 2.0, 8.1 Hz), d 7.65 (1H, d, J = 2.0 Hz) and d 7.24 (1H, d, J = 8.1 Hz) and a methoxy group at d 3.92 (3H, s) along with four anomeric protons at d 5.52 (1H, d, J = 1.2 Hz), 4.86 (1H, d, J = 1.2 Hz), 4.46 (1H, d, J = 7.5 Hz) and 4.22 (1H, d, J = 7.5 Hz) were observed in the 1H NMR spectrum. The HSQC, HMBC and DQFCOSY experiments allowed us to determine the presence of a 3,4substituted benzoyl moiety, identified as vanilloyl unit, two b-glucopyranosyl units (d 4.46 and 4.22) and two a-rhamnopyranosyl units (d 5.52 and 4.86). The HMBC spectrum showed key correlation peaks between the proton signal at d 4.46 (H-1glcI) and the carbon resonance at d 79.6 (C-3), d 4.86 (H-1rhaI) and d 80.4 (H-4glcI), and the proton signal at d 4.22 (H-1glcII) and the carbon resonance at d 83.3 (C-16). The HMBC correlation between the proton signals at d 4.56 (H2-6glcI) and the carbon resonance at d 167.5 (C-7van) indicated that the b-glucose unit at C-3 was esterified by the vanilloyl unit at C-6. Finally, the correlation between the signal proton at d 5.52 (H-1rhaII) and the carbon resonance at d 151.1 (C-4van) was observed in the HMBC spectrum. On the basis of these data, the structure of compound 4 was identified as the new 16-O-b-Dglucopyranosyl-colest-5-en-3b,7b,16b-triol-23-one 3-O-{6-O-[3-
7
b-D-GlcI 4.47 d (7.5) 3.58 dd (9.0, 7.5) 3.30 dd (9.0, 9.0) 3.29 dd (9.0, 9.0) 3.57 ddd (9.0, 4.5, 2.0) 3.87 dd (12.0, 2.0) 3.68 dd (12.0, 4.5) b-D-GlcII 4.68 d (7.5) 3.22 dd (9.0, 7.5) 3.38 dd (9.0, 9.0) 3.24 dd (9.0, 9.0) 3.28 ddd (9.0, 4.5, 2.0) 3.85 dd (12.0, 2.0) 3.68 dd (12.0, 4.5)
100.5 80.4 78.0 71.8 78.2 62.9
103.4 75.5 77.9 71.4 77.7 62.9
104.3 75.4 77.7 71.5 77.7 62.9
b-D-GlcI 4.48 d (7.5) 3.59 dd (9.0, 7.5) 3.35 dd (9.0, 9.0) 3.25 dd (9.0, 9.0) 3.59 ddd (9.0, 4.5, 2.0) 3.88 dd (12.0, 2.0) 3.69 dd (12.0, 4.5) b-D-GlcII 4.70 d (7.5) 3.24 dd (9.0, 7.5) 3.35 dd (9.0, 9.0) 3.34 dd (9.0, 9.0) 3.28 ddd (9.0, 4.5, 2.0) 3.88 dd (12.0, 2.0) 3.69 dd (12.0, 4.5) b-D-GlcIII 4.26 d (7.5) 3.22 dd (9.0, 7.5) 3.38 dd (9.0, 9.0) 3.31 dd (9.0, 9.0) 3.29 ddd (9.0, 4.5, 2.0) 3.88 dd (12.0, 2.0) 3.69 dd (12.0, 4.5)
8 101.4 78.6 74.9 70.2 76.3 62.2
104.2 76.1 77.5 71.7 78.0 62.4
b-D-Gal 4.41 d (7.5) 3.92 dd (8.5, 7.9) 3.71 dd (8.5, 2.9) 3.83 dd (2.9, 1.2) 3.54 m 3.77 dd (12.0, 2.0) 3.70 dd (12.0, 4.5) b-D-Glc 4.68 d (7.5) 3.22 dd (9.0, 7.5) 3.39 dd (9.0, 9.0) 3.26 dd (9.0, 9.0) 3.30 ddd (9.0, 4.5, 2.0) 3.88 dd (12.0, 2.0) 3.70 dd (12.0, 4.5)
methoxy-4-O-(a-L-rhamnopyranosyl)-benzoyl]-a-L-rhamnopyranosyl-(1 ? 4)-b-D-glucopyranoside}. Additionally, five known spirostanol derivatives namely (25R)5b-spirostan-3b-ol-12-one-3-O-{b-D-glucopyranosyl-(1 ? 2)-O-[bD-glucopyranosyl-(1 ? 3)]-b-D-glucopyranoside} (YS-VII) (1) (Nakano et al., 1991a), (25R)-5b-spirostan-3b-ol-12-one-3-O-b-Dglucopyranosyl-(1 ? 2)-[b-D-glucopyranosyl-(1 ? 3)]-b-D-galactopyranoside (YS-VIII) (3) (Nakano et al., 1991a), (25R)-5b-spirostan3b-ol-3-O-b-D-glucopyranosyl-(1 ? 2)-b-D-galactopyranoside (YSII) (10) (Nakano et al., 1989), (25R)-5b-spirostan-3b-ol-3-O-b-D-glucopyranosyl-(1 ? 2)-b-D-glucopyranoside (YS-I) (11) (Nakano et al., 1989) and (25S)-5b-spirostan-3b-ol-3-O-b-D-glucopyranosyl(1 ? 2)-b-D-galactopyranoside (timosaponin A III) (12), along with the furostanol derivative namely 26-O-b-D-glucopyranosyl-furost20(22)-en-3b-ol-3-O-[b-D-galactopyranosyl-(1 ? 2)-O-b-D-glucopyranoside] (macrostemonoside F) (5) (Peng et al., 1993) and the cholestane derivative namely 3,16-di-O-b-D-glucopyranosylcholest-5-en-23-one-3b,7b,16b-triol (camassioside) (Mimaki et al., 1992) (9) (Saito et al., 1994) were isolated. A better accuracy in liquid chromatography tandem mass spectrometry for quantitative analyses is recognized for tandem in space mass spectrometry. In particular Multiple Reaction Monitoring (MRM), the tandem mass spectrometric technique in which a specific transition from a
132
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
precursor ion to a product ion is monitored for each analyte, guarantees a very high selectivity and sensitivity. Thus for quantitative purposes, an accurate method on a mass spectrometer equipped with an electrospray source and a triple quadrupole analyzer has been developed for the analysis of steroidal glycosides in Y. gloriosa rhizomes. To optimize tandem mass spectrometry conditions, fragmentation patterns were studied by analyzing a standard solution of 0.1 lg/ml for each investigated compound by electrospray-triple quadrupole-mass spectrometry (ES-QqQ-MS). Although for compounds 1–12, the loss of the sugar chain was the predominant fragmentation of the pseudomolecular ion [M+H]+, the prevalent ion in MS1 was the sodiated pseudomolecular ion [M+Na]+ that was difficult to fragment. For these reasons the quantitative analysis was performed by working in pseudo Multiple Reaction Monitoring (pseudo-MRM), that is a tandem mass spectrometry detection method that uses both the quadrupoles, as working in MRM, but setting precursor and product ion to the same values (Haug et al., 2009). With this MS monitoring technique, the precursor ion is measured after application of a voltage leading to fragmentation of interfering compounds with the exception of the analyte (Andreoli et al., 1999). This approach results in an improved signal to noise compared with Single Ion Monitoring (SIM) and is particularly advantageous for saponins, as these compounds (with a large number of heteroatoms) often appear in ESI-MS as sodium or potassium adducts, which are more resistant to fragmentation than their protonated molecular ions. Product ions and precursor ions were set to the same values corresponding to the sodiated molecular ion adducts: m/z 937.4 for compound 2, m/z 939.4 for compound 3, m/z 1221.6 for compound 4, m/z 925.5 for compounds 5 and 7, m/z 779.4 for compound 9, m/z 777.4 for compound 8, m/z 763.2 for compound 12. For compound 6 the pseudo-transition of the sodiated ion at m/z 779.4 produced a noisier trace than that of the pseudo-transition of the protonated ion at m/z 757.4 (Fig. 4), thus this last transition was chosen. Compounds 2–9 and 12 were selected for quantitative analysis since they had a purity >99%. Compounds 1, 10 and 11 were less pure and were excluded from the calibration plot assessment. Fig. 4 shows LC/MS/MS profiles for the nine compounds under investigation. Calibration curves were performed for all the compounds by using as external standards at five different concentration levels (1, 2.5, 5, 10, 25 and 50 lg/ml). The calibration curves obtained by plotting the external standard areas versus the known concentration of each compound were linear in the range 1–25 lg/ml for all the compounds. Five aliquots of the BuOH enriched saponins fraction of Y. gloriosa rhizomes were analyzed in order to quantify their glycosidic content. Table 3 reports quantitative analysis results. The HPLC–MS/MS assay was validated according to the European Medicines Agency guidelines relating to the validation of analytical methods (EMEA). The method based on the pseudo-MRM transitions was specific with no peak interference from other compounds at the retention times of the marker compounds (2–9 and 12) in the chromatograms. The intra-day accuracy and precision were calculated by analyzing three samples of compound 2 at three different concentration levels, namely, 1, 5 and 10 lg/ml, on the same day. Inter-day estimates were performed over three consecutive days. The standard deviation was <5%. The calibration graphs, obtained by plotting the area obtained from external standard against the known concentration of external standard (for each compound) was linear in the range of 1–50 lg/ml. The limit of quantification (LOQ), defined as the lowest concentration of compound quantifiable with acceptable accuracy and precision, was determined by injection of a series of diluted standard solutions until a signal-to-noise ratio of 10 was attained. The LOQ values calculated for the six compounds under investigation were less than 20 ng/ml. Table 3 reports validation data of the method developed for quantitative analysis of compounds 2–9 and 12.
3. Conclusion The dried extract obtained from rhizomes contain more than 25% w/w of glycosides, thus Y. gloriosa rhizomes can be considered a rich source of steroidal glycosides which characteristically show a cis junction between rings A and B, belonging to the 5b series of cholestane, furostane and spirostane. The quantitative method here described is straightforward and convenient because it requires a very fast sample preparation procedure, and it can be applied to crude extract obtained from dried plant material. In particular, pseudo-MRM monitoring in tandem MS gave very good results in terms of specificity and sensitivity for saponins, and similar MS approaches could be applied to other plant extracts characterized by the occurrence of this class of compounds, affording robust and accurate quantitative results. 4. Experimental 4.1. General Optical rotations were measured on a JASCO DIP 1000 polarimeter. IR measurements were obtained on a Bruker IFS-48 spectrometer. NMR experiments were performed on a Bruker DRX-600 spectrometer (Bruker BioSpinGmBH, Rheinstetten, Germany) equipped with a Bruker 5 mm TCI CryoProbe at 300 K. All 2DNMR spectra were acquired in CD3OD (99.95%, Sigma–Aldrich) and standard pulse sequences and phase cycling were used for DQF-COSY, HSQC, HMBC and ROESY spectra. The NMR data were processed using UXNMR software. Exact masses were measured by a Voyager DE mass spectrometer. Samples were analyzed by matrix-assisted laser desorption ionization time-of-flight (MALDITOF) mass spectrometry. A mixture of analyte solution and a-cyano-4-hydroxycinnamic acid (Sigma) was applied to the metallic sample plate and dried. Mass calibration was performed with the ions from ACTH (fragment 18–39) at 2465.1989 Da and angiotensin III at 931.5154 Da as internal standard. Column chromatography was performed over silica gel (100/ 160 lm, Merck). TLC was performed on silica gel plates (Merck precoated silica gel 60 F254) and developed in the solvent system, CHCl3:MeOH:H2O (26:14:3). All solvents for extraction and chromatographic separation were of analytical grade and purchased from Carlo Erba (Rodano, Italy). HPLC separations were carried out on a Waters 590 system equipped with a Waters R401 refractive index detector, a Waters XTerra Prep MSC18 column (300 7.8 mm i.d.) and a Rheodyne injector. HPLC grade methanol, acetonitrile and trifluoroacetic acid were purchased from J.T. Baker (Baker Mallinckrodt, Phillipsburg, NJ, USA). HPLC grade water (18 mX) was prepared using a Millipore Milli-Q purification system (Millipore Corp., Bedford, MA). Standards of compounds 1–12 were isolated in our laboratory. 4.2. Plant material The rhizomes of Y. gloriosa were collected in February 2008 in Shiraki in the experimental field of the Institute of Pharmacochemistry, Georgia. Samples of Y. gloriosa were identified by Dr. Jemal Aneli, Department of Pharmacobotany, Institute of Pharmacochemistry, Tbilisi, Georgia. A voucher specimen (No. 259) has been deposited at this department. 4.3. Extraction and isolation Two hundred grams barkless powdered rhizomes of Y. gloriosa were extracted with MeOH 80% (1 l) once at room temperature
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
133
Fig. 4. Positive-ion mode LC–ESI-MS/MS pseudo-MRM analyses of the BuOH extract of Yucca gloriosa rhizomes. (A) Total profile of all the Q1–Q2 specific transitions. (B1–B8) Profiles of the specific transition from parent ions to product ions in pseudo-MRM mode.
and twice at 60 °C. The collected extracts have been dried under vacuum and the concentrate was partitioned between hexane (4 g), water (32 g) and n-butanol (11 g). Part of n-butanolic extract (2 g) containing steroidal compounds was subjected to a silica gel column chromatography (100
2.5 cm, 100/160 lm, Merck) eluting with gradient system chloroform–methanol-(9:1 ? 0:10) yielding 225 fractions (8 ml each) which were combined in fractions A (25 mg), B (36 mg), C (120 mg) and D (37 mg) including two major spirostanol glycosides, compounds 10 (40 mg) and 11 (62 mg). Fractions A–D were
134
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
Table 3 Quantitative results for compounds 2–9 and 12, and their concentration in Yucca gloriosa rhizomes extracts. Compound
r2
Calibration curve equation
LOQ (ng ml1)
LOD (ng ml1)
Dried extract mg/g (D.S.)
2 3 4 5 6 7 8 9 12
0.998 0.999 0.997 0.994 0.999 0.999 0.998 0.999 0.999
y = 443x + 778 y = 1890x + 53.3 y = 16.1x + 101 y = 2170x 435 y = 1930x 588 y = 1270x + 483 y = 1420x 125 y = 1310x 959 y = 1270x + 483
7.83 5.38 13.71 9.81 11.35 8.52 9.99 11.11 7.41
1.85 0.75 1.17 1.56 1.38 0.98 0.91 2.01 0.86
24.96 6.93 8.82 45.12 28.36 14.43 31.44 9.55 67.20
chromatographed by RP-HPLC on a C18 column (Waters XTerra Prep MSC18, 7.8 300 mm), using different percentage of MeOH in isocratic conditions. From fraction A (57% MeOH as eluent, flow rate 2 ml/min) compounds 2 (3.5 mg, tR = 30.4 min) and 1 (5.2 mg, tR = 36.8 min) were obtained; from fraction B (65% MeOH as eluent, flow rate 2 ml/min) 6 (3.8 mg, tR = 16.7 min), 8 (4.2 mg, tR = 23.8 min), 12 (7.4 mg, tR = 38.6 min), 3 (1.2 mg, tR = 24.2 min) were isolated; fraction C (60% MeOH, as eluent, flow rate 2 ml/ min) yielded 5 (5.2 mg, tR = 38 min) and 7 (2.2 mg, tR = 41.2 min); fraction D (65% MeOH, as eluent, flow rate 2 ml/min) yielded 9 (1.5 mg, tR = 25.9 min), 4 (1.2 mg, tR = 28.8 min). In order to obtain the enriched saponin fraction for qualitative and quantitative analysis, dried rhizomes extract was prepared with the following procedure: 200 g air dried barkless powdered rhizomes of Y. gloriosa were extracted with MeOH 80% once at room temperature and twice at 60 °C. The collected extracts have been dried under vacuum. The residue (47 g) was dissolved in 150 ml water and was defatted with hexane and then with n-butanol to yield 11 g of a steroidal glycoside enriched fraction (5.5% from dry rhizomes; the water content in fresh rhizomes is 75%). One milligram of the BuOH extract was dissolved in 1 ml of MeOH before analysis of 20 ll in chromatographic systems. 4.4. ESI-MS and ESI-MS/MS analyses Full scan ESI-MS and collision induced dissociation (CID) ESIMS/MS analyses of standard steroidal glycosides were performed on an Applied Biosystems (Foster City, CA, USA) API2000 spectrometer equipped with a triple quadrupole analyzer. The analytical parameters were optimized by infusing a standard solution of compound 3 (1 lg/ml in methanol) into the source at a flow rate of 5 ll/min. The optimized parameters were declustering potential 100 eV, focusing potential 120 eV, entrance potential 8 eV, collision energy 30 eV and collision cell exit potential (CXP) 15 eV. Data were acquired in the positive-ion MS and MS/MS modes. 4.5. Qualitative HPLC–ESI-MS and quantitative HPLC–ESI-MS/MS analyses Qualitative on-line HPLC–ESI-MS analyses of the saponin enriched fraction was obtained as reported above and was performed on a Thermo Finnigan (Thermo Fisher) Spectra System HPLC coupled to a Thermo Fisher LCQ Deca IT spectrometer. Analyses were carried out using a Waters (Milford, MA, USA) Atlantis C18 column (150 2.0 mm i.d.; 5 lm particle size) eluted with mixtures of water containing 0.05% trifluoroacetic acid, TFA (solvent A) and acetonitrile containing 0.05% TFA (solvent B) at a flow rate of 0.2 ml/min. Gradient elution started with 100% A and changed to 80:20 (A:B) in 35 min, then from 80:20 (A:B) to 100% B in 5 min. The flow from the chromatograph was injected directly into the ESI source, maintained at a temperature of 280 °C, and MS were measured under the optimized parameters indicated for the ESI-MS with nitrogen
(2.26) (0.28) (0.31) (2.11) (2.52) (0.16) (2.81) (1.02) (2.03)
supplied at a flow rate of 80 (arbitrary units). MS data were acquired using the software provided by the manufacturer, and reconstructed ion chromatograms (RICs) were elaborated in order to identify steroid glycosides from their protonated molecular ions. Quantitative on-line HPLC–ESI-MS/MS analyses of the extracts were performed using an Agilent 1100 HPLC system interfaced to an Applied Biosystems (Foster City, CA, USA) API2000 instrument. The chromatographic conditions were as described above for HPLC–UV analyses. The API2000 ES source was tuned by infusing a standard solution of compound 3 (1 lg/ml in methanol) into the source at a flow rate of 10 ll/min. The optimized parameters were declustering potential 100 eV, focusing potential 120 eV, entrance potential 8 eV, collision energy 30 eV and collision cell exit potential (CXP) 15 eV The spectrometer was used in the MS/MS mode by using multiple reaction monitoring, in the same way as the pseudo-MRM experiment. 4.6. Calibration and quantification In order to prepare the specific calibration plot for each compound under investigation, a sample (10 mg) of each standard (2–9 and 12) was weighted accurately into a 10 ml volumetric flask, dissolved in methanol and the volume made up to the mark with methanol. The resulting stock solution was diluted with methanol in order to obtain reference solutions containing 2.5, 5, 10, 25 and 50 lg/ml of external standard. Peak areas of the external standard were plotted against the corresponding standard concentration using weighted linear regression to generate standard curves. All quantitative data were elaborated with the aid of Analyst software (Applied Biosystems). 4.7. 5b-Spirost-25(27)-en-3b-ol-12-one 3-O-{b-D-glucopyranosyl(1 ? 2)-O-[b-D-glucopyranosyl-(1 ? 3)]-b-D-glucopyranoside} (2) Amorphous white solid; C45H70O19; ½a22 D – 20.7° (c 0.1 MeOH); IR 1 mKBr : 3435 (>OH), 2949 (>CH), 1281 and 1030 (C–O–C); for 1H max cm
and 13C NMR data (CD3OD, 600 MHz) of the aglycon moiety and sugar portion, see Tables 1 and 2, respectively; ESI-MS m/z 937.4 [M+Na]+, MS/MS m/z 775.4 [M+Na162]+, MS3 m/z 613.4 [M+Na 162162]+, MS4 m/z 451.4 [M+Na162162162]+; HRMALDITOFMS [M+Na]+ m/z 937.4412 (calcd. for C45H70O19Na, 937.4409). 4.8. 16-O-b-D-Glucopyranosyl-colest-5-en-3b,7b,16b-triol-23-one 3O-{6-O-[3-methoxy-4-O-(a-L-rhamnopyranosyl)-benzoyl]-a-Lrhamnopyranosyl-(1 ? 4)-b-D-glucopyranoside} (4)
Amorphous white solid; C59H90O25; ½a22 D – 57.1° (c 0.1 MeOH); UV (MeOH) kmax (log e) 295 (4.63), 262 (4.60), 218 (4.50) nm; IR 1 mKBr : 3424 (>OH), 2931 (>CH), 1284 and 1056 (C–O–C); for max cm 1 H and 13C NMR data (CD3OD, 600 MHz) of the aglycon moiety and sugar portion, see Tables 1 and 2, respectively; ESI-MS m/z 1221.6 [M+Na]+, MS/MS m/z 1075.6 [M+Na146]+, MS3 m/z
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
135
1057.5 [M+Na14618]+, MS4 m/z 911.5 [M+Na14618146]+, m/z 581.4 [M+Na14618146330]+; HRMALDITOFMS [M+Na]+ m/z 1221.5672 (calcd. for C59H90O25Na, 1221.5669).
6-162). Any idea in this publication is possessed by the author and may not represent the opinion of Georgian National Science Foundation itself.
4.9. (25R)-5b-Spirostan-3b,12b-diol 3-O-b-D-glucopyranosyl-(1 ? 2)O-b-D-glucopyranoside (6)
References
22 D
Amorphous white solid; C39H64O14; ½a – 38.7° (c 0.1 MeOH); 1 IR mKBr : 3443 (>OH), 2921 (>CH), 1274 and 1068 (C–O–C); for max cm 1 H and 13C NMR data (CD3OD, 600 MHz) of the aglycon moiety and sugar portion, see Tables 1 and 2, respectively; ESI-MS m/z 779.4 [M+Na]+, MS/MS m/z 617.4 [M+Na162]+, MS3 m/z 455.3 [M+Na162162]+; HRMALDITOFMS [M+Na]+ m/z 779.4199 (calcd. for C39H64O14Na, 779.4194). 4.10. (25R)-26-O-b-D-Glucopyranosyl-5b-furost-20(22)-en-3b,26-diol 3-O-b-D glucopyranosyl-(1 ? 2)-O-b-D-glucopyranoside (7) Amorphous white solid; C45H74O18; ½a22 D – 45.0° (c 0.1 MeOH); 1 IR mKBr : 3424 (>OH), 2925 (>CH), 1290 and 1044 (C–O–C); for max cm 1 H and 13C NMR data (CD3OD, 600 MHz) of the aglycon moiety and sugar portion, see Tables 1 and 2, respectively; ESI-MS m/z 925.5 [M+Na]+, MS/MS m/z 763.5 [M+Na162]+, MS3 m/z 601.5 [M+Na162162]+; HRMALDITOFMS [M+Na]+ m/z 925.4778 (calcd. for C45H74O18Na, 925.4773). 4.11. (25R)-5b-Spirostan-3b-ol-12-one 3-O-b-D-glucopyranosyl(1 ? 2)-O-b-D-galactopyranoside (8) Amorphous white solid; C39H62O14; ½a22 D – 40.8° (c 0.1 MeOH); 1 IR mKBr : 3450 (>OH), 2921 (>CH), 1274 and 1057 (C–O–C); for max cm 1 H and 13C NMR data (CD3OD, 600 MHz) of the aglycon moiety and sugar portion, see Tables 1 and 2, respectively; ESI-MS m/z 777.4 [M+Na]+, MS/MS m/z 615.3 [M+Na162]+, MS3 m/z 453.4 [M+Na162162]+; HRMALDITOFMS [M+Na]+ m/z 777.4041 (calcd. for C39H62O14Na, 777.4037). 4.12. Acid hydrolysis The crude saponin mixture (100 mg) was refluxed with 15 ml of 2 N HCl for 4 h. The sapogenins were extracted with EtOAc (3 15 ml), and the organic layer was neutralized by washing with H2O, and evaporated to dryness. The acid aqueous layer was neutralized with 1 N NaOH and freeze-dried. Three sugars were identified with authentic samples by TLC in MeCOEt–isoPrOH– Me2CO–H2O (20:10:7:6) as rhamnose, glucose and galactose. After a preparative TLC of sugar mixture (50 mg) in this solvent, the optical rotation of each purified sugar was measured. The D configuration of galactose and glucose and the L configuration of rhamnose were established by comparison of their optical rotation values with those reported in the literature: D-glucose ½a23 D +52.5, L-rham23 nose ½a23 D 4.4 (Wang et al., 2008), D-galactose ½aD +80.2 (Belitz et al., 2009). The optical rotations were determined after dissolving the sugars in H2O and allowing them to stand for 24 h: D-glucose 23 23 ½a23 D +54.5 (c 0.1), D-galactose ½aD +82.1 (c 0.1), L-rhamnose ½aD 4.8 (c 0.1). Acknowledgements The designated project has been fulfilled by financial support of Georgian National Science Foundation (Grant #GNSF/PRE07/
Agrawal, P.K., 2004. NMR spectral investigations: Part 51. Dependence of 1H NMR chemical shifts of geminal protons of glycosyloxy methylene (H2–26) on the orientation of the 27-methyl group of furostane-type steroidal saponins. Magn. Reson. Chem. 42, 990–993. Andreoli, R., Manini, P., Bergamaschi, E., Mutti, A., Franchini, I., Niessen, W.M.A., 1999. Determination of naphthalene metabolites in human urine by liquid chromatography–mass spectrometry with electrospray ionization. J. Chromatogr. A 847, 9–17. Bassarello, C., Bifulco, G., Montoro, P., Skhirtladze, A., Kemertelidze, E., Pizza, C., Piacente, S., 2007a. Gloriosaols A and B, two novel phenolics from Yucca gloriosa: structural characterization and configurational assignment by a combined NMR-quantum mechanical strategy. Tetrahedron 63, 148–154. Bassarello, C., Bifulco, G., Montoro, P., Skhirtladze, A., Benidze, M., Pizza, C., Piacente, S., 2007b. Yucca gloriosa: a rich phenolic derivatives with strong antioxidant activities. J. Agric. Food Chem. 55, 6636–6642. Belitz, H.-D., Grosch, W., Schieberle, P., 2009. Food Chemistry, fourth ed. SpringerVerlag, Heidelberg, Berlin. Benidze, M.M., Pkheidze, T.A., Kemertelidze, E.P., 1991. Spirostanol glycosides of Yucca gloriosa. Khim. Prir. Soedin. 2, 295–296. EMEA Quality Guidelines: Validation of Analytical Procedures: Text and Methodology (ICH Q2). Available from:
Contents lists available at ScienceDirect
Phytochemistry journal homepage: www.elsevier.com/locate/phytochem
Steroidal saponins from Yucca gloriosa L. rhizomes: LC–MS profiling, isolation and quantitative determination Alexandre Skhirtladze a, Angela Perrone b, Paola Montoro b, Mariam Benidze a, Ether Kemertelidze a, Cosimo Pizza b, Sonia Piacente b,⇑ a b
Iovel Kutateladze Institute of Pharmacochemistry, P. Sarajishvili 36, 0159 Tbilisi, Georgia Dipartimento di Scienze Farmaceutiche, Università degli Studi di Salerno, Via Ponte Don Melillo, I-84084 Fisciano, Italy
a r t i c l e
i n f o
Article history: Received 30 July 2010 Received in revised form 18 October 2010 Accepted 19 October 2010 Available online 19 November 2010 Keywords: Yucca gloriosa Steroidal glycosides LC–MS profiling Quantitative determination
a b s t r a c t The occurrence of steroidal saponins in the rhizomes of Yucca gloriosa has been detected by LC–MS. On the basis of the LC–MS analysis, five steroidal glycosides, including three spirostane, one furostane and one cholestane glycosides, along with seven known compounds have been isolated and characterized by ESI-MS and by the extensive use of 1D- and 2D-NMR experiments. Quantitative analysis of the steroidal glycosides in Y. gloriosa rhizomes was performed by an LC–MS method validated according to European Medicines Agency (EMEA) guidelines. The dried BuOH extract obtained from rhizomes contains more than 25% w/w of glycosides, thus Y. gloriosa rhizomes can be considered a rich source of steroidal glycosides. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction Yucca gloriosa L. is largely cultivated in Eastern Georgia, where industrial plantations of this species occupy a total area of about 150 ha (Benidze et al., 1991). Y. gloriosa produces a strong underground system, whose main part is represented by rhizomes. In the rhizomes, biosynthesis of 5b steroids occurs, whereas in leaves and flowers only 5a steroids (Skhirtladze et al., 2006; Kemertelidze and Benidze, 2001) have been detected. The roots and bark of Y. gloriosa are free from saponins, but they are rich in phenolic constituents related to resveratrol, with high antioxidant and antiproliferative activities (Bassarello et al., 2007a,b; Nigro et al., 2007; Montoro et al., 2008). Due to the high content of steroidal saponins (5–6%), the rhizomes of Y. gloriosa are used as raw material for the semisynthesis of steroidal hormones. Continuing our studies on Yucca spp. (Montoro et al., 2008, 2010; Skhirtladze et al., 2006), here we report on the phytochemical investigation of the steroidal saponins of the rhizomes of Y. gloriosa. A preliminary analysis of the BuOH enriched saponin fraction obtained by Y. gloriosa rhizomes was performed by LC–ESI-MS in order to detect the presence of saponins in this part of the plant. Successively, five new steroidal glycosides, including three spirostane, one furostane and one cholestane glycosides, along with seven known compounds were isolated. Finally, an LC–MS/MS ⇑ Corresponding author. Tel.: +39 089 969763; fax: +39 089 969602. E-mail address: [email protected] (S. Piacente). 0031-9422/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.phytochem.2010.10.013
method was developed for the quantitative analysis of steroidal glycosides in the extract of Y. gloriosa rhizomes.
2. Results and discussion From the analytical perspective, it is clear that steroidal saponins are not detectable by HPLC–UV analysis for the lack of a strong UV chromophore and that HPLC analysis of all compounds requires gradient elution. Mass spectrometry represents an effective detection method, and in addition improvement in selectivity and specificity can be raised by using tandem mass spectrometry. LC–MS is selective and sensitive enough to carry out the analysis of saponins (Yan and Guo, 2005; Tor et al., 2005; Montoro et al., 2010). On the basis of these observations, a preliminary analysis of the extract obtained by Y. gloriosa rhizomes was performed by LC–ESI-MS in order to detect the presence of saponins in this part of the plant. The ESI-MS spectra of saponins show adducts together with pseudomolecular ions; in the ESI-MS spectra in positive-ion mode, mainly [M+H]+, [M+NH4]+, [M+Na]+, and [M+K]+ ions can be observed (Miao et al., 2002). Column and mobile phase selection is generally determined by the combination of the compounds to be analyzed and the matrix where they occur. In our work the use of an Atlantis C18 column and a gradient elution allowed us to obtain a good separation of the steroidal glycosides occurring in the BuOH enriched saponin fraction of Y. gloriosa rhizomes, without any other compound interfering with the analyte detection.
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
The profile of the Total Ion Current (TIC) showed 12 main peaks. For these peaks MS spectra were obtained by extracting them from the total profile. Successively, on the basis of the m/z values, Reconstructed Ion Chromatograms (RIC) were extracted. Positive-ion electrospray LC–MS analysis obtained in the acquisition range of 190–1200 amu, TIC profile and RICs are shown in Fig. 1. Twelve main compounds were detected: at m/z 917, compounds 1 and 3; at m/z 915, compound 2; at m/z 1199, compound 4; at m/z 903, compounds 5 and 7; at m/z 757, compounds 6 and 9; at m/z 755, compound 8; at m/z 741, compounds 10–12. On the basis of the results of the on-line screening and identification by LC–MS, compounds 1–12 were isolated and their structures were unambiguously elucidated from NMR spectroscopic data (Figs. 2 and 3). The BuOH extract was fractionated by silica gel column chromatography to yield compounds 10 and 11. The fractions obtained were chromatographed by reversed-phase HPLC to yield compounds 1–9 and 12 (see Section 4). The absolute configurations of the sugar units were assigned after acid hydrolysis of the crude saponin mixture and identification with authentic samples by TLC, followed by preparative separation of each sugar. The D configuration of galactose and glucose and the L configuration of rhamnose were established by comparison of their optical rotation values with those reported in the literature (Belitz et al., 2009; Wang et al., 2008). The HRMALDITOF mass spectrum of 2 (m/z 937.4412 [M+Na]+, calcd. for C45H70O19Na, 937.4409) supported a molecular formula of C45H70O19. The ESI-MS mass spectrum showed the major ion peak
127
at m/z 937.4 which was assigned to [M+Na]+. The MS/MS of this ion showed a peak at m/z 775.4 [M+Na162]+, corresponding to the loss of an hexose unit. In the MS3 spectrum a peak at m/z 613.4 [M+Na162162]+, corresponding to the loss of a second hexose unit, was observed. Finally, in the MS4 spectrum a peak at m/z 451.4 [M+Na162162162]+, indicating the loss of a third hexose unit, was observed. The 1H NMR spectrum of compound 2 showed signals for two tertiary methyl groups at d 1.11 (3H, s) and 1.12 (3H, s), a secondary methyl group at d 1.05 (3H, d, J = 6.6 Hz), exomethylene protons at d 4.81 and 4.72 (each 1H, br s), two methine proton signals at d 4.43 (1H, m) and 4.12 (1H, m), two methylene proton signals at d 4.31 and 3.87 (each 1H, d, J = 12.1 Hz), along with three anomeric protons at d 4.96 (1H, d, J = 7.5 Hz), 4.68 (1H, d, J = 7.5 Hz) and 4.51 (1H, d, J = 7.5 Hz). The 13C NMR spectrum displayed, for the aglycon moiety, signals ascribable to a keto group at d 215.6, a ketal function at d 110.1, two secondary alcoholic functions at d 75.4 and 80.7, and one primary alcoholic function at d 65.3, suggesting as aglycon a spirostanol skeleton characterized by the occurrence of a keto group (Table 1). On the basis of the HSQC and HMBC correlations, the aglycon moiety of compound 2 was identified as 5b-spirost-25(27)-en-3b-ol-12-one or schidigera-genin B, previously isolated from Y. schidigera (Miyakoshi et al., 2000). The chemical shifts of all the individual protons of the three sugar units were ascertained from a combination of 1D-TOCSY and DQF-COSY spectral analysis, and the 13C chemical shifts of their relative attached carbons were assigned unambiguously from the HSQC spectrum (Table 2). These data showed the presence of three bglucopyranosyl units (d 4.96, 4.68 and 4.51). Glycosidation shifts
Fig. 1. Positive-ion mode LC–MS analysis of the BuOH extract of Yucca gloriosa rhizomes. Total ion current and reconstructed ions chromatogram for compounds 1–12.
128
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
Fig. 2. Compounds 2, 4, 6, 7, 8 from Yucca gloriosa L. rhizomes.
were observed for C-3glcI (d 87.6), C-2glcI (d 78.9) and C-3 (d 75.4). An unambiguous determination of the sequence and linkage sites was obtained from the HMBC spectrum, which showed key correlation peaks between the proton signal at d 4.51 (H-1glcI) and the carbon resonance at d 75.4 (C-3), d 4.96 (H-1glcII) and d 78.9 (C-2glcI), and the proton signal at d 4.68 (H-1glcII) and the carbon resonance at d 87.6 (C-3glcI). On the basis of this evidence, the structure of the new compound 2 was established as 5b-spirost-25(27)-en-3b-ol12-one 3-O-{b-D-glucopyranosyl-(1 ? 2)-O-[b-D-glucopyranosyl(1 ? 3)]-b-D-glucopyranoside}. The molecular formula of compound 8 was established as C39H62O14 by HRMALDITOFMS analysis (m/z 777.4041 [M+Na]+,
calcd. for C39H62O14Na, 777.4037). The ESI-MS mass spectrum showed the major ion peak at m/z 777.4 which was assigned to [M+Na]+. In the MS/MS spectrum a peak at m/z 615.3 [M+Na162]+, corresponding to the loss of an hexose unit, was observed. The MS3 spectrum displayed a peak at m/z 453.4 [M+Na162162]+, due to the loss of a second hexose unit. The 1 H and 13C NMR data of the aglycon moiety of 8 compared to those of 2 showed that the two compounds differed only by the replacement of the exomethylene group with a secondary methyl group at C-27 (dH 0.79, dC 17.0). The 1H and 13C NMR data of the aglycon of 8 were in agreement with those reported for (25R)-5b-spirostan-3bol-12-one, named gloriogenin (Nakano et al., 1989, 1991a). For the
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
129
Fig. 3. Compounds 1, 3, 5, 9, 10, 11, 12 from Yucca gloriosa L. rhizomes.
sugar portion, the 1H NMR spectrum showed signals for two anomeric protons at d 4.68 (1H, d, J = 7.5 Hz) and 4.41 (1H, d, J = 7.5 Hz). Complete assignments of the 1H and 13C NMR signals of the sugar portion were accomplished by HSQC, HMBC, DQFCOSY and 1D-TOCSY experiments which led to the identification of one b-galactopyranosyl (d 4.41) unit and one b-glucopyranosyl (d 4.68) unit. Once again, direct evidence of the sugar sequence and the linkage sites was derived from HSQC and HMBC experi-
ments. The glycosidation shifts on C-3 (d 75.8) and C-2gal (d 78.6) indicated the linkage sites. Key correlation peaks were observed in the HMBC spectrum of 8 between the proton signal at d 4.41 (H-1gal) and the carbon resonance at d 75.8 (C-3) and the proton signal at d 4.68 (H-1glc) and the carbon resonance at d 78.6 (C-2gal). Therefore, compound 8 was established as the new (25R)-5b-spirostan-3b-ol-12-one 3-O-b-D-glucopyranosyl-(1 ? 2)O-b-D-galactopyranoside.
130
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
Table 1 13 C and 1H NMR data (J in Hz) of the aglycon moieties of compounds 2, 4, 6–8 (CD3OD, 600 MHz). 2
4
dC
dH
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
30.4 27.3 75.4 30.9 36.7 27.0 26.7 35.7 43.2 36.0 38.4 215.6 56.7 57.6 32.1 80.7 54.8 16.3 23.8 43.2 13.5 110.1
1.92, 1.57 1.71, 1.60 4.12 m 1.89, 1.55 1.93 m 2.00, 1.47 1.25, 1.17 1.80 m 1.65 m
23 24 25 26
33.9 29.0 144.4 65.3
27
108.6
m m m m m
2.25, 2.04 m
1.45 m 2.00, 1.32 m 4.43 m 2.32 m 1.12 s 1.11 s 1.98 m 1.05 d (6.6)
1.78, 1.75 m 2.57, 2.20 m 4.31 d (12.1) 3.87 d (12.1) 4.81, 4.72 br s
6
7
dC
dH
dC
dH
37.8 30.9 79.6 38.8 143.9 127.4 73.3 40.8 49.9 37.7 21.8 40.5 44.0 55.1 39.0 83.3 61.5 13.4 19.1 28.4 19.6 50.5
1.69, 0.79 m 1.88, 1.58 m 3.50 m 2.42, 2.26 m
31.3 27.4 75.9 31.4 36.8 27.8 27.0 37.3 40.3 36.3 37.4 80.6 46.9 56.0 32.7 82.0 62.9 16.3 24.5 43.3 13.5 110.6
1.88, 1.59 1.70, 1.63 4.11 m 1.89, 1.54 1.93 m 2.00, 1.48 1.26, 1.19 1.61 m 1.55 m
1.65, 1.44 m 3.29 m
2.41 (2H) dd (7.1, 2.3) 2.13 m 0.95 d (6.6)
32.3 29.3 31.5 67.6
0.95 d (6.6)
17.3
216.2 53.5 25.1 22.9 22.9
5.24 3.69 1.49 0.94
br d (2.1) dd (8.2, 2.1) m m
1.50 (2H) m 2.06, 1.22 m 1.06 m 2.48, 1.88 m 4.16 m 1.26 dd (11.6, 7.9) 0.95 s 1.04 s 2.55 m 0.99 d (6.6) 2.77 br d (15.4, 2.0) 2.30 dd (15.4, 10.4)
The HRMALDITOF mass spectrum of 6 (m/z 779.4199 [M+Na]+, calcd. for C39H64O14Na, 779.4194) supported a molecular formula of C39H64O14. The ESI-MS mass spectrum showed the major ion peak at m/z 779.4 which was assigned to [M+Na]+. In the MS/MS spectrum a peak at m/z 617.4 [M+Na162]+, corresponding to the loss of an hexose unit, was observed. The MS3 spectrum displayed a peak at m/z 455.3 [M+Na162162]+, due to the loss of a second hexose unit. A comparison of the NMR data of compound 6 with those of compound 8 showed that in the aglycon of 6 the keto group at C-12 was replaced by the 12b-hydroxy group, indicating as aglycon the 12b-hydroxysmilagenin (Nakano et al., 1991b). For the sugar portion signals of two anomeric protons at d 4.68 (1H, d, J = 7.5 Hz) and 4.47 (1H, d, J = 7.5 Hz) were observed in the 1H NMR spectrum. A combination of HSQC, 1D-TOCSY and DQF-COSY spectra led to the identification of two b-glucopyranosyl (d 4.68 and 4.47) units. Key correlation peaks between the proton signal at d 4.47 (H-1glcI) and the carbon resonance at d 75.9 (C-3) and the proton signal at d 4.68 (H-1glcII) and the carbon resonance at d 80.6 (C-2glcI) were observed in the HMBC spectrum. Thus, the structure of the new compound 6 was identified as (25R)-5b-spirostan-3b,12b-diol 3-O-b-D-glucopyranosyl-(1 ? 2)-O-b-D-glucopyranoside. The molecular formula of compound 7 was established as C45H74O18 by HRMALDITOFMS analysis (m/z 925.4778 [M+Na]+, calcd. for C45H74O18Na, 925.4773). The ESI-MS mass spectrum showed the major ion peak at m/z 925.5 which was assigned to [M+Na]+. In the MS/MS spectrum a peak at m/z 763.5 [M+Na162]+, corresponding to the loss of an hexose unit, was observed. The MS3 spectrum displayed a peak at m/z 601.5 [M+Na162162]+, due to the loss of a second hexose unit. The 1 H NMR spectrum of compound 7 showed signals for two tertiary methyl groups at d 0.72, 1.02 (3H, s) and 1.63 (3H, s), a secondary methyl group at d 0.98 (3H, d, J = 6.6 Hz), two methine proton signals at d 4.74 (1H, m) and 4.12 (1H, m), and two methylene proton signals at d 3.35 and 3.82 ascribable to a primary alcoholic func-
8
dC
dH
31.0 26.8 76.0 31.0 37.3 27.3 27.4 36.3 41.2 36.2 21.9 40.9 44.8 56.0 34.9 85.5 65.7 14.8 24.2 105.4 11.5 153.0
1.89, 1.59 1.69, 1.63 4.12 m 1.89, 1.53 1.91 m 1.98, 1.49 1.23, 1.18 1.64 m 1.47 m
1.49, 1.38 m 1.84, 1.32 m
1.64 (2H) m 1.48 (2H) m 1.61 m 3.45, 3.30 m
23.9 31.7 34.1 75.8
2.15 (2H) m 1.64, 1.30 m 1.81 m 3.80, 3.35 m
32.3 29.3 31.1 67.5
1.60 (2H) m 1.48 (2H) m 1.63 m 3.45, 3.30 m
0.82 d (5.8)
17.1
0.98 d (6.6)
17.0
0.79 d (5.8)
m m m m m
1.20 m 2.02, 1.38 m 4.40 m 1.85 m 0.75 s 1.03 s 1.92 m 1.02 d (6.6)
m m m m m
1.16 m 2.19, 1.42 m 4.74 m 2.52 d (10.1) 0.72 s 1.02 s 1.63 s
dC
dH
30.7 27.5 75.8 30.8 36.6 27.2 26.8 35.3 43.1 36.7 37.9 215.8 56.5 57.1 32.2 80.6 54.8 16.0 23.5 43.0 12.9 109.3
1.90, 1.55 1.70, 1.60 4.11 m 1.91, 1.58 1.91 m 1.98, 1.45 1.27, 1.17 1.82 m 1.67 m
m m m m m
2.20, 2.03 m
1.42 m 2.02, 1.35 m 4.34 m 2.29 m 1.07 s 1.08 s 1.91 m 1.02 d (6.6)
tion, along with three anomeric protons at d 4.70 (1H, d, J = 7.5 Hz), 4.48 (1H, d, J = 7.5 Hz) and 4.26 (1H, d, J = 7.5 Hz). The 13 C NMR spectrum displayed for the aglycon signals ascribable to two sp2 carbons at d 105.4 and 153.0, two secondary alcoholic functions at d 76.0 and 85.5, and one primary alcoholic function at d 75.8, suggesting as aglycon a furostanol skeleton (Table 1). On the basis of the HSQC and HMBC correlations, the aglycon moiety of compound 7 was identified as (25R)-5b-furost-20(22)-en3b,26-diol named pseudosmilagenin. The C-25 configuration was deduced to be R based on the difference of chemical shifts (Dab = da db) of the geminal protons at H2-26 (Dab = 0.45 ppm). It has been described that Dab is usually >0.57 ppm in 25S compounds and <0.48 in 25R compounds (Agrawal, 2004). The NMR data of compound 7 in comparison with those reported for macrostemonoside F (Peng et al., 1993) showed that the two compounds differed only by the replacement of the galactose unit with of the glucose unit as inner sugar at C-3. Thus, compound 7 was established as the new (25R)-26-O-b-D-glucopyranosyl-5b-furost20(22)-en-3b,26-diol 3-O-b-D-glucopyranosyl-(1 ? 2)-O-b-D-glucopyranoside. For the first time a pseudosmilagenin glycoside has been isolated from Yucca species. Compound 4 showed in the positive ESI-MS a major ion peak at m/z 1221.6 [M+Na]+ and a significant fragment in MS/MS analysis at m/z 1075.6 [M+Na146]+, ascribable to the loss of a deoxyhexose unit. The MS3 fragmentation showed peak at m/z 1057.5 [M+Na14618]+ due to the loss of a water molecule. Finally, in the MS4 a peak at m/z 911.5 [M+Na14618146]+, indicating the loss of a second deoxyhexose unit, was observed. Its molecular formula was established unequivocally as C59H90O25 by HRMALDITOF mass spectrum (m/z 1221.5672 [M+Na]+, calcd. for C59H90O25Na, 1221.5669). The 1H NMR spectrum of compound 4 displayed signals for two tertiary methyl groups at d 0.95 (3H, s) and 1.04 (3H, s), three secondary methyl groups at d 0.99 (3H, d, J = 6.6 Hz) and 0.95 (6H, d, J = 6.4 Hz), three methine proton signals at d 4.16 (1H, m), 3.50 (1H, m) and 3.69 (1H, dd, J = 8.2, 2.1 Hz), an
131
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135 Table 2 13 C and 1H NMR data (J in Hz) of the sugar portions of compounds 2, 4, 6–8 (CD3OD, 600 MHz). 2 1 2 3 4 5
100.6 78.9 87.6 69.9 78.2
6
62.7
1 2 3 4 5
103.4 75.8 77.9 72.3 77.7
6
63.4
1 2 3 4 5
104.3 75.0 77.9 71.4 77.8
6
62.8
b-D-GlcI 4.51 d (7.5) 3.72 dd (9.0, 7.5) 3.76 dd (9.0, 9.0) 3.38 dd (9.0, 9.0) 3.33 ddd (9.0, 4.5, 2.0) 3.87 dd (12.0, 2.0) 3.68 dd (12.0, 4.5) b-D-GlcII 4.96 d (7.5) 3.13 dd (9.0, 7.5) 3.39 dd (9.0, 9.0) 3.16 dd (9.0, 9.0) 3.32 ddd (9.0, 4.5, 2.0) 3.85 dd (12.0, 2.0) 3.65 dd (12.0, 4.5) b-D-GlcIII 4.68 d (7.5) 3.31 dd (9.0, 7.5) 3.38 dd (9.0, 9.0) 3.31 dd (9.0, 9.0) 3.38 ddd (9.0, 4.5, 2.0) 3.87 dd (12.0, 2.0) 3.67 dd (12.0, 4.5)
4 101.9 75.1 76.5 80.4 74.1 64.6
102.7 72.1 72.1 73.7 70.5
b-D-GlcI 4.46 d (7.5) 3.27 dd (9.0, 7.5) 3.55 dd (9.0, 9.0) 3.55 dd (9.0, 9.0) 3.75 ddd (9.0, 4.5, 2.0) 4.56 dd (12.0, 2.0) 4.56 dd (12.0, 4.5) a-L-RhaI 4.86 d (1.2) 3.92 dd (1.2, 3.2) 3.69 dd (3.2, 9.3) 3.44 t (9.3) 4.02 m
17.7
1.31 d (6.0)
125.5 114.0 151.4 151.1 117.6
Vanillic acid – 7.65 d (2.0) – – 7.24 d (8.1)
124.1
7.68 dd (2.0, 8.1)
7 OMe
167.5 56.2
1 2 3 4 5 6
100.4 71.8 72.1 73.5 71.0 17.7
1 2 3 4 5
106.5 75.2 78.0 71.3 77.6
6
62.7
– 3.92 s a-L-RhaII 5.52 d (1.2) 4.12 dd (1.2, 3.2) 3.92 dd (3.2, 9.3) 3.52 t (9.3) 3.72 m 1.25 d (6.0) b-D-GlcII 4.22 d (7.5) 3.21 dd (9.0, 7.5) 3.37 dd (9.0, 9.0) 3.28 dd (9.0, 9.0) 3.25 ddd (9.0, 4.5, 2.0) 3.87 dd (12.0, 2.0) 3.67 dd (12.0, 4.5)
6 100.7 80.6 77.8 71.5 78.2 62.4
104.2 76.1 77.5 71.7 78.0 62.4
olefinic proton at d 5.24 (1H, br d). The 1H and 13C NMR data of the aglycon portion of compound 4 were in agreement with those reported for 3b,7b,16b-trihydroxycholest-5-en-23-one (Mimaki et al., 1992). Additionally, three aromatic protons at d 7.68 (1H, dd, J = 2.0, 8.1 Hz), d 7.65 (1H, d, J = 2.0 Hz) and d 7.24 (1H, d, J = 8.1 Hz) and a methoxy group at d 3.92 (3H, s) along with four anomeric protons at d 5.52 (1H, d, J = 1.2 Hz), 4.86 (1H, d, J = 1.2 Hz), 4.46 (1H, d, J = 7.5 Hz) and 4.22 (1H, d, J = 7.5 Hz) were observed in the 1H NMR spectrum. The HSQC, HMBC and DQFCOSY experiments allowed us to determine the presence of a 3,4substituted benzoyl moiety, identified as vanilloyl unit, two b-glucopyranosyl units (d 4.46 and 4.22) and two a-rhamnopyranosyl units (d 5.52 and 4.86). The HMBC spectrum showed key correlation peaks between the proton signal at d 4.46 (H-1glcI) and the carbon resonance at d 79.6 (C-3), d 4.86 (H-1rhaI) and d 80.4 (H-4glcI), and the proton signal at d 4.22 (H-1glcII) and the carbon resonance at d 83.3 (C-16). The HMBC correlation between the proton signals at d 4.56 (H2-6glcI) and the carbon resonance at d 167.5 (C-7van) indicated that the b-glucose unit at C-3 was esterified by the vanilloyl unit at C-6. Finally, the correlation between the signal proton at d 5.52 (H-1rhaII) and the carbon resonance at d 151.1 (C-4van) was observed in the HMBC spectrum. On the basis of these data, the structure of compound 4 was identified as the new 16-O-b-Dglucopyranosyl-colest-5-en-3b,7b,16b-triol-23-one 3-O-{6-O-[3-
7
b-D-GlcI 4.47 d (7.5) 3.58 dd (9.0, 7.5) 3.30 dd (9.0, 9.0) 3.29 dd (9.0, 9.0) 3.57 ddd (9.0, 4.5, 2.0) 3.87 dd (12.0, 2.0) 3.68 dd (12.0, 4.5) b-D-GlcII 4.68 d (7.5) 3.22 dd (9.0, 7.5) 3.38 dd (9.0, 9.0) 3.24 dd (9.0, 9.0) 3.28 ddd (9.0, 4.5, 2.0) 3.85 dd (12.0, 2.0) 3.68 dd (12.0, 4.5)
100.5 80.4 78.0 71.8 78.2 62.9
103.4 75.5 77.9 71.4 77.7 62.9
104.3 75.4 77.7 71.5 77.7 62.9
b-D-GlcI 4.48 d (7.5) 3.59 dd (9.0, 7.5) 3.35 dd (9.0, 9.0) 3.25 dd (9.0, 9.0) 3.59 ddd (9.0, 4.5, 2.0) 3.88 dd (12.0, 2.0) 3.69 dd (12.0, 4.5) b-D-GlcII 4.70 d (7.5) 3.24 dd (9.0, 7.5) 3.35 dd (9.0, 9.0) 3.34 dd (9.0, 9.0) 3.28 ddd (9.0, 4.5, 2.0) 3.88 dd (12.0, 2.0) 3.69 dd (12.0, 4.5) b-D-GlcIII 4.26 d (7.5) 3.22 dd (9.0, 7.5) 3.38 dd (9.0, 9.0) 3.31 dd (9.0, 9.0) 3.29 ddd (9.0, 4.5, 2.0) 3.88 dd (12.0, 2.0) 3.69 dd (12.0, 4.5)
8 101.4 78.6 74.9 70.2 76.3 62.2
104.2 76.1 77.5 71.7 78.0 62.4
b-D-Gal 4.41 d (7.5) 3.92 dd (8.5, 7.9) 3.71 dd (8.5, 2.9) 3.83 dd (2.9, 1.2) 3.54 m 3.77 dd (12.0, 2.0) 3.70 dd (12.0, 4.5) b-D-Glc 4.68 d (7.5) 3.22 dd (9.0, 7.5) 3.39 dd (9.0, 9.0) 3.26 dd (9.0, 9.0) 3.30 ddd (9.0, 4.5, 2.0) 3.88 dd (12.0, 2.0) 3.70 dd (12.0, 4.5)
methoxy-4-O-(a-L-rhamnopyranosyl)-benzoyl]-a-L-rhamnopyranosyl-(1 ? 4)-b-D-glucopyranoside}. Additionally, five known spirostanol derivatives namely (25R)5b-spirostan-3b-ol-12-one-3-O-{b-D-glucopyranosyl-(1 ? 2)-O-[bD-glucopyranosyl-(1 ? 3)]-b-D-glucopyranoside} (YS-VII) (1) (Nakano et al., 1991a), (25R)-5b-spirostan-3b-ol-12-one-3-O-b-Dglucopyranosyl-(1 ? 2)-[b-D-glucopyranosyl-(1 ? 3)]-b-D-galactopyranoside (YS-VIII) (3) (Nakano et al., 1991a), (25R)-5b-spirostan3b-ol-3-O-b-D-glucopyranosyl-(1 ? 2)-b-D-galactopyranoside (YSII) (10) (Nakano et al., 1989), (25R)-5b-spirostan-3b-ol-3-O-b-D-glucopyranosyl-(1 ? 2)-b-D-glucopyranoside (YS-I) (11) (Nakano et al., 1989) and (25S)-5b-spirostan-3b-ol-3-O-b-D-glucopyranosyl(1 ? 2)-b-D-galactopyranoside (timosaponin A III) (12), along with the furostanol derivative namely 26-O-b-D-glucopyranosyl-furost20(22)-en-3b-ol-3-O-[b-D-galactopyranosyl-(1 ? 2)-O-b-D-glucopyranoside] (macrostemonoside F) (5) (Peng et al., 1993) and the cholestane derivative namely 3,16-di-O-b-D-glucopyranosylcholest-5-en-23-one-3b,7b,16b-triol (camassioside) (Mimaki et al., 1992) (9) (Saito et al., 1994) were isolated. A better accuracy in liquid chromatography tandem mass spectrometry for quantitative analyses is recognized for tandem in space mass spectrometry. In particular Multiple Reaction Monitoring (MRM), the tandem mass spectrometric technique in which a specific transition from a
132
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
precursor ion to a product ion is monitored for each analyte, guarantees a very high selectivity and sensitivity. Thus for quantitative purposes, an accurate method on a mass spectrometer equipped with an electrospray source and a triple quadrupole analyzer has been developed for the analysis of steroidal glycosides in Y. gloriosa rhizomes. To optimize tandem mass spectrometry conditions, fragmentation patterns were studied by analyzing a standard solution of 0.1 lg/ml for each investigated compound by electrospray-triple quadrupole-mass spectrometry (ES-QqQ-MS). Although for compounds 1–12, the loss of the sugar chain was the predominant fragmentation of the pseudomolecular ion [M+H]+, the prevalent ion in MS1 was the sodiated pseudomolecular ion [M+Na]+ that was difficult to fragment. For these reasons the quantitative analysis was performed by working in pseudo Multiple Reaction Monitoring (pseudo-MRM), that is a tandem mass spectrometry detection method that uses both the quadrupoles, as working in MRM, but setting precursor and product ion to the same values (Haug et al., 2009). With this MS monitoring technique, the precursor ion is measured after application of a voltage leading to fragmentation of interfering compounds with the exception of the analyte (Andreoli et al., 1999). This approach results in an improved signal to noise compared with Single Ion Monitoring (SIM) and is particularly advantageous for saponins, as these compounds (with a large number of heteroatoms) often appear in ESI-MS as sodium or potassium adducts, which are more resistant to fragmentation than their protonated molecular ions. Product ions and precursor ions were set to the same values corresponding to the sodiated molecular ion adducts: m/z 937.4 for compound 2, m/z 939.4 for compound 3, m/z 1221.6 for compound 4, m/z 925.5 for compounds 5 and 7, m/z 779.4 for compound 9, m/z 777.4 for compound 8, m/z 763.2 for compound 12. For compound 6 the pseudo-transition of the sodiated ion at m/z 779.4 produced a noisier trace than that of the pseudo-transition of the protonated ion at m/z 757.4 (Fig. 4), thus this last transition was chosen. Compounds 2–9 and 12 were selected for quantitative analysis since they had a purity >99%. Compounds 1, 10 and 11 were less pure and were excluded from the calibration plot assessment. Fig. 4 shows LC/MS/MS profiles for the nine compounds under investigation. Calibration curves were performed for all the compounds by using as external standards at five different concentration levels (1, 2.5, 5, 10, 25 and 50 lg/ml). The calibration curves obtained by plotting the external standard areas versus the known concentration of each compound were linear in the range 1–25 lg/ml for all the compounds. Five aliquots of the BuOH enriched saponins fraction of Y. gloriosa rhizomes were analyzed in order to quantify their glycosidic content. Table 3 reports quantitative analysis results. The HPLC–MS/MS assay was validated according to the European Medicines Agency guidelines relating to the validation of analytical methods (EMEA). The method based on the pseudo-MRM transitions was specific with no peak interference from other compounds at the retention times of the marker compounds (2–9 and 12) in the chromatograms. The intra-day accuracy and precision were calculated by analyzing three samples of compound 2 at three different concentration levels, namely, 1, 5 and 10 lg/ml, on the same day. Inter-day estimates were performed over three consecutive days. The standard deviation was <5%. The calibration graphs, obtained by plotting the area obtained from external standard against the known concentration of external standard (for each compound) was linear in the range of 1–50 lg/ml. The limit of quantification (LOQ), defined as the lowest concentration of compound quantifiable with acceptable accuracy and precision, was determined by injection of a series of diluted standard solutions until a signal-to-noise ratio of 10 was attained. The LOQ values calculated for the six compounds under investigation were less than 20 ng/ml. Table 3 reports validation data of the method developed for quantitative analysis of compounds 2–9 and 12.
3. Conclusion The dried extract obtained from rhizomes contain more than 25% w/w of glycosides, thus Y. gloriosa rhizomes can be considered a rich source of steroidal glycosides which characteristically show a cis junction between rings A and B, belonging to the 5b series of cholestane, furostane and spirostane. The quantitative method here described is straightforward and convenient because it requires a very fast sample preparation procedure, and it can be applied to crude extract obtained from dried plant material. In particular, pseudo-MRM monitoring in tandem MS gave very good results in terms of specificity and sensitivity for saponins, and similar MS approaches could be applied to other plant extracts characterized by the occurrence of this class of compounds, affording robust and accurate quantitative results. 4. Experimental 4.1. General Optical rotations were measured on a JASCO DIP 1000 polarimeter. IR measurements were obtained on a Bruker IFS-48 spectrometer. NMR experiments were performed on a Bruker DRX-600 spectrometer (Bruker BioSpinGmBH, Rheinstetten, Germany) equipped with a Bruker 5 mm TCI CryoProbe at 300 K. All 2DNMR spectra were acquired in CD3OD (99.95%, Sigma–Aldrich) and standard pulse sequences and phase cycling were used for DQF-COSY, HSQC, HMBC and ROESY spectra. The NMR data were processed using UXNMR software. Exact masses were measured by a Voyager DE mass spectrometer. Samples were analyzed by matrix-assisted laser desorption ionization time-of-flight (MALDITOF) mass spectrometry. A mixture of analyte solution and a-cyano-4-hydroxycinnamic acid (Sigma) was applied to the metallic sample plate and dried. Mass calibration was performed with the ions from ACTH (fragment 18–39) at 2465.1989 Da and angiotensin III at 931.5154 Da as internal standard. Column chromatography was performed over silica gel (100/ 160 lm, Merck). TLC was performed on silica gel plates (Merck precoated silica gel 60 F254) and developed in the solvent system, CHCl3:MeOH:H2O (26:14:3). All solvents for extraction and chromatographic separation were of analytical grade and purchased from Carlo Erba (Rodano, Italy). HPLC separations were carried out on a Waters 590 system equipped with a Waters R401 refractive index detector, a Waters XTerra Prep MSC18 column (300 7.8 mm i.d.) and a Rheodyne injector. HPLC grade methanol, acetonitrile and trifluoroacetic acid were purchased from J.T. Baker (Baker Mallinckrodt, Phillipsburg, NJ, USA). HPLC grade water (18 mX) was prepared using a Millipore Milli-Q purification system (Millipore Corp., Bedford, MA). Standards of compounds 1–12 were isolated in our laboratory. 4.2. Plant material The rhizomes of Y. gloriosa were collected in February 2008 in Shiraki in the experimental field of the Institute of Pharmacochemistry, Georgia. Samples of Y. gloriosa were identified by Dr. Jemal Aneli, Department of Pharmacobotany, Institute of Pharmacochemistry, Tbilisi, Georgia. A voucher specimen (No. 259) has been deposited at this department. 4.3. Extraction and isolation Two hundred grams barkless powdered rhizomes of Y. gloriosa were extracted with MeOH 80% (1 l) once at room temperature
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
133
Fig. 4. Positive-ion mode LC–ESI-MS/MS pseudo-MRM analyses of the BuOH extract of Yucca gloriosa rhizomes. (A) Total profile of all the Q1–Q2 specific transitions. (B1–B8) Profiles of the specific transition from parent ions to product ions in pseudo-MRM mode.
and twice at 60 °C. The collected extracts have been dried under vacuum and the concentrate was partitioned between hexane (4 g), water (32 g) and n-butanol (11 g). Part of n-butanolic extract (2 g) containing steroidal compounds was subjected to a silica gel column chromatography (100
2.5 cm, 100/160 lm, Merck) eluting with gradient system chloroform–methanol-(9:1 ? 0:10) yielding 225 fractions (8 ml each) which were combined in fractions A (25 mg), B (36 mg), C (120 mg) and D (37 mg) including two major spirostanol glycosides, compounds 10 (40 mg) and 11 (62 mg). Fractions A–D were
134
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
Table 3 Quantitative results for compounds 2–9 and 12, and their concentration in Yucca gloriosa rhizomes extracts. Compound
r2
Calibration curve equation
LOQ (ng ml1)
LOD (ng ml1)
Dried extract mg/g (D.S.)
2 3 4 5 6 7 8 9 12
0.998 0.999 0.997 0.994 0.999 0.999 0.998 0.999 0.999
y = 443x + 778 y = 1890x + 53.3 y = 16.1x + 101 y = 2170x 435 y = 1930x 588 y = 1270x + 483 y = 1420x 125 y = 1310x 959 y = 1270x + 483
7.83 5.38 13.71 9.81 11.35 8.52 9.99 11.11 7.41
1.85 0.75 1.17 1.56 1.38 0.98 0.91 2.01 0.86
24.96 6.93 8.82 45.12 28.36 14.43 31.44 9.55 67.20
chromatographed by RP-HPLC on a C18 column (Waters XTerra Prep MSC18, 7.8 300 mm), using different percentage of MeOH in isocratic conditions. From fraction A (57% MeOH as eluent, flow rate 2 ml/min) compounds 2 (3.5 mg, tR = 30.4 min) and 1 (5.2 mg, tR = 36.8 min) were obtained; from fraction B (65% MeOH as eluent, flow rate 2 ml/min) 6 (3.8 mg, tR = 16.7 min), 8 (4.2 mg, tR = 23.8 min), 12 (7.4 mg, tR = 38.6 min), 3 (1.2 mg, tR = 24.2 min) were isolated; fraction C (60% MeOH, as eluent, flow rate 2 ml/ min) yielded 5 (5.2 mg, tR = 38 min) and 7 (2.2 mg, tR = 41.2 min); fraction D (65% MeOH, as eluent, flow rate 2 ml/min) yielded 9 (1.5 mg, tR = 25.9 min), 4 (1.2 mg, tR = 28.8 min). In order to obtain the enriched saponin fraction for qualitative and quantitative analysis, dried rhizomes extract was prepared with the following procedure: 200 g air dried barkless powdered rhizomes of Y. gloriosa were extracted with MeOH 80% once at room temperature and twice at 60 °C. The collected extracts have been dried under vacuum. The residue (47 g) was dissolved in 150 ml water and was defatted with hexane and then with n-butanol to yield 11 g of a steroidal glycoside enriched fraction (5.5% from dry rhizomes; the water content in fresh rhizomes is 75%). One milligram of the BuOH extract was dissolved in 1 ml of MeOH before analysis of 20 ll in chromatographic systems. 4.4. ESI-MS and ESI-MS/MS analyses Full scan ESI-MS and collision induced dissociation (CID) ESIMS/MS analyses of standard steroidal glycosides were performed on an Applied Biosystems (Foster City, CA, USA) API2000 spectrometer equipped with a triple quadrupole analyzer. The analytical parameters were optimized by infusing a standard solution of compound 3 (1 lg/ml in methanol) into the source at a flow rate of 5 ll/min. The optimized parameters were declustering potential 100 eV, focusing potential 120 eV, entrance potential 8 eV, collision energy 30 eV and collision cell exit potential (CXP) 15 eV. Data were acquired in the positive-ion MS and MS/MS modes. 4.5. Qualitative HPLC–ESI-MS and quantitative HPLC–ESI-MS/MS analyses Qualitative on-line HPLC–ESI-MS analyses of the saponin enriched fraction was obtained as reported above and was performed on a Thermo Finnigan (Thermo Fisher) Spectra System HPLC coupled to a Thermo Fisher LCQ Deca IT spectrometer. Analyses were carried out using a Waters (Milford, MA, USA) Atlantis C18 column (150 2.0 mm i.d.; 5 lm particle size) eluted with mixtures of water containing 0.05% trifluoroacetic acid, TFA (solvent A) and acetonitrile containing 0.05% TFA (solvent B) at a flow rate of 0.2 ml/min. Gradient elution started with 100% A and changed to 80:20 (A:B) in 35 min, then from 80:20 (A:B) to 100% B in 5 min. The flow from the chromatograph was injected directly into the ESI source, maintained at a temperature of 280 °C, and MS were measured under the optimized parameters indicated for the ESI-MS with nitrogen
(2.26) (0.28) (0.31) (2.11) (2.52) (0.16) (2.81) (1.02) (2.03)
supplied at a flow rate of 80 (arbitrary units). MS data were acquired using the software provided by the manufacturer, and reconstructed ion chromatograms (RICs) were elaborated in order to identify steroid glycosides from their protonated molecular ions. Quantitative on-line HPLC–ESI-MS/MS analyses of the extracts were performed using an Agilent 1100 HPLC system interfaced to an Applied Biosystems (Foster City, CA, USA) API2000 instrument. The chromatographic conditions were as described above for HPLC–UV analyses. The API2000 ES source was tuned by infusing a standard solution of compound 3 (1 lg/ml in methanol) into the source at a flow rate of 10 ll/min. The optimized parameters were declustering potential 100 eV, focusing potential 120 eV, entrance potential 8 eV, collision energy 30 eV and collision cell exit potential (CXP) 15 eV The spectrometer was used in the MS/MS mode by using multiple reaction monitoring, in the same way as the pseudo-MRM experiment. 4.6. Calibration and quantification In order to prepare the specific calibration plot for each compound under investigation, a sample (10 mg) of each standard (2–9 and 12) was weighted accurately into a 10 ml volumetric flask, dissolved in methanol and the volume made up to the mark with methanol. The resulting stock solution was diluted with methanol in order to obtain reference solutions containing 2.5, 5, 10, 25 and 50 lg/ml of external standard. Peak areas of the external standard were plotted against the corresponding standard concentration using weighted linear regression to generate standard curves. All quantitative data were elaborated with the aid of Analyst software (Applied Biosystems). 4.7. 5b-Spirost-25(27)-en-3b-ol-12-one 3-O-{b-D-glucopyranosyl(1 ? 2)-O-[b-D-glucopyranosyl-(1 ? 3)]-b-D-glucopyranoside} (2) Amorphous white solid; C45H70O19; ½a22 D – 20.7° (c 0.1 MeOH); IR 1 mKBr : 3435 (>OH), 2949 (>CH), 1281 and 1030 (C–O–C); for 1H max cm
and 13C NMR data (CD3OD, 600 MHz) of the aglycon moiety and sugar portion, see Tables 1 and 2, respectively; ESI-MS m/z 937.4 [M+Na]+, MS/MS m/z 775.4 [M+Na162]+, MS3 m/z 613.4 [M+Na 162162]+, MS4 m/z 451.4 [M+Na162162162]+; HRMALDITOFMS [M+Na]+ m/z 937.4412 (calcd. for C45H70O19Na, 937.4409). 4.8. 16-O-b-D-Glucopyranosyl-colest-5-en-3b,7b,16b-triol-23-one 3O-{6-O-[3-methoxy-4-O-(a-L-rhamnopyranosyl)-benzoyl]-a-Lrhamnopyranosyl-(1 ? 4)-b-D-glucopyranoside} (4)
Amorphous white solid; C59H90O25; ½a22 D – 57.1° (c 0.1 MeOH); UV (MeOH) kmax (log e) 295 (4.63), 262 (4.60), 218 (4.50) nm; IR 1 mKBr : 3424 (>OH), 2931 (>CH), 1284 and 1056 (C–O–C); for max cm 1 H and 13C NMR data (CD3OD, 600 MHz) of the aglycon moiety and sugar portion, see Tables 1 and 2, respectively; ESI-MS m/z 1221.6 [M+Na]+, MS/MS m/z 1075.6 [M+Na146]+, MS3 m/z
A. Skhirtladze et al. / Phytochemistry 72 (2011) 126–135
135
1057.5 [M+Na14618]+, MS4 m/z 911.5 [M+Na14618146]+, m/z 581.4 [M+Na14618146330]+; HRMALDITOFMS [M+Na]+ m/z 1221.5672 (calcd. for C59H90O25Na, 1221.5669).
6-162). Any idea in this publication is possessed by the author and may not represent the opinion of Georgian National Science Foundation itself.
4.9. (25R)-5b-Spirostan-3b,12b-diol 3-O-b-D-glucopyranosyl-(1 ? 2)O-b-D-glucopyranoside (6)
References
22 D
Amorphous white solid; C39H64O14; ½a – 38.7° (c 0.1 MeOH); 1 IR mKBr : 3443 (>OH), 2921 (>CH), 1274 and 1068 (C–O–C); for max cm 1 H and 13C NMR data (CD3OD, 600 MHz) of the aglycon moiety and sugar portion, see Tables 1 and 2, respectively; ESI-MS m/z 779.4 [M+Na]+, MS/MS m/z 617.4 [M+Na162]+, MS3 m/z 455.3 [M+Na162162]+; HRMALDITOFMS [M+Na]+ m/z 779.4199 (calcd. for C39H64O14Na, 779.4194). 4.10. (25R)-26-O-b-D-Glucopyranosyl-5b-furost-20(22)-en-3b,26-diol 3-O-b-D glucopyranosyl-(1 ? 2)-O-b-D-glucopyranoside (7) Amorphous white solid; C45H74O18; ½a22 D – 45.0° (c 0.1 MeOH); 1 IR mKBr : 3424 (>OH), 2925 (>CH), 1290 and 1044 (C–O–C); for max cm 1 H and 13C NMR data (CD3OD, 600 MHz) of the aglycon moiety and sugar portion, see Tables 1 and 2, respectively; ESI-MS m/z 925.5 [M+Na]+, MS/MS m/z 763.5 [M+Na162]+, MS3 m/z 601.5 [M+Na162162]+; HRMALDITOFMS [M+Na]+ m/z 925.4778 (calcd. for C45H74O18Na, 925.4773). 4.11. (25R)-5b-Spirostan-3b-ol-12-one 3-O-b-D-glucopyranosyl(1 ? 2)-O-b-D-galactopyranoside (8) Amorphous white solid; C39H62O14; ½a22 D – 40.8° (c 0.1 MeOH); 1 IR mKBr : 3450 (>OH), 2921 (>CH), 1274 and 1057 (C–O–C); for max cm 1 H and 13C NMR data (CD3OD, 600 MHz) of the aglycon moiety and sugar portion, see Tables 1 and 2, respectively; ESI-MS m/z 777.4 [M+Na]+, MS/MS m/z 615.3 [M+Na162]+, MS3 m/z 453.4 [M+Na162162]+; HRMALDITOFMS [M+Na]+ m/z 777.4041 (calcd. for C39H62O14Na, 777.4037). 4.12. Acid hydrolysis The crude saponin mixture (100 mg) was refluxed with 15 ml of 2 N HCl for 4 h. The sapogenins were extracted with EtOAc (3 15 ml), and the organic layer was neutralized by washing with H2O, and evaporated to dryness. The acid aqueous layer was neutralized with 1 N NaOH and freeze-dried. Three sugars were identified with authentic samples by TLC in MeCOEt–isoPrOH– Me2CO–H2O (20:10:7:6) as rhamnose, glucose and galactose. After a preparative TLC of sugar mixture (50 mg) in this solvent, the optical rotation of each purified sugar was measured. The D configuration of galactose and glucose and the L configuration of rhamnose were established by comparison of their optical rotation values with those reported in the literature: D-glucose ½a23 D +52.5, L-rham23 nose ½a23 D 4.4 (Wang et al., 2008), D-galactose ½aD +80.2 (Belitz et al., 2009). The optical rotations were determined after dissolving the sugars in H2O and allowing them to stand for 24 h: D-glucose 23 23 ½a23 D +54.5 (c 0.1), D-galactose ½aD +82.1 (c 0.1), L-rhamnose ½aD 4.8 (c 0.1). Acknowledgements The designated project has been fulfilled by financial support of Georgian National Science Foundation (Grant #GNSF/PRE07/
Agrawal, P.K., 2004. NMR spectral investigations: Part 51. Dependence of 1H NMR chemical shifts of geminal protons of glycosyloxy methylene (H2–26) on the orientation of the 27-methyl group of furostane-type steroidal saponins. Magn. Reson. Chem. 42, 990–993. Andreoli, R., Manini, P., Bergamaschi, E., Mutti, A., Franchini, I., Niessen, W.M.A., 1999. Determination of naphthalene metabolites in human urine by liquid chromatography–mass spectrometry with electrospray ionization. J. Chromatogr. A 847, 9–17. Bassarello, C., Bifulco, G., Montoro, P., Skhirtladze, A., Kemertelidze, E., Pizza, C., Piacente, S., 2007a. Gloriosaols A and B, two novel phenolics from Yucca gloriosa: structural characterization and configurational assignment by a combined NMR-quantum mechanical strategy. Tetrahedron 63, 148–154. Bassarello, C., Bifulco, G., Montoro, P., Skhirtladze, A., Benidze, M., Pizza, C., Piacente, S., 2007b. Yucca gloriosa: a rich phenolic derivatives with strong antioxidant activities. J. Agric. Food Chem. 55, 6636–6642. Belitz, H.-D., Grosch, W., Schieberle, P., 2009. Food Chemistry, fourth ed. SpringerVerlag, Heidelberg, Berlin. Benidze, M.M., Pkheidze, T.A., Kemertelidze, E.P., 1991. Spirostanol glycosides of Yucca gloriosa. Khim. Prir. Soedin. 2, 295–296. EMEA Quality Guidelines: Validation of Analytical Procedures: Text and Methodology (ICH Q2). Available from: