Y. Matsuo et al. / Phytochemistry 96 (2013) 244–256
249
O-b-
O-b-
D
-galactopyranosyl-(1 ? 2)-O-[b-
-glucopyranosyl-(1 ? 4)-b- -galactopyranoside.
-glucosidase gave the corre-
-glucose, whereas acid hydro-
lysis of 7 gave spirostanol sapogenin 2a, -galactose, -glucose, and
D
-glucopyranosyl-(1 ? 3)]-
and did not affect cell growth of the TIG-3 normal cells. Although
previously reported pseudo-furostanol glycosides have not shown
cytotoxic activity (Liu et al., 2003; Yokosuka et al., 2002), pseudo-
furostanol glycoside 9 exhibited significant tumor-selective cyto-
toxicity. Comparing cytotoxic activities of 6 and 13, those of 1
and 2, and those of 7 and 13 suggested that the presence of a
D
D
Enzymatic hydrolysis of 7 with b-
sponding spirostanol saponin 3 and
D
D
D
D
D
-xylose. The HMBC correlation of 7 between the anomeric proton
(H-100000) of one b-
D-glucopyranosyl group at dH 4.83 (d, J = 7.8 Hz)
C-2a
hydroxy group did not affect cytotoxic activity, whereas
position reduced the
and C-26 of the aglycone at dC 75.3 confirmed that a b-
D-glucopyr-
introducing a hydroxy group to the C-14a
anosyl unit was attached at C-26. NOE correlations between the
signals of the H-20 proton at dH 2.36 and the H2-23 protons at dH
cytotoxicity. Preliminary analysis showed that the C-6 hydroxy
group or C-6 O-glucosyl group significantly reduced cytotoxic
activity in the spirostanol glycosides (Lu et al., 2011; Yokosuka
et al., 2009b). In the present study, these results also suggested
that introduction of a hydroxy group to the steroidal aglycone re-
duced cytotoxicity. Compounds 1 and 11, as well as 6 and 14, are
2.09 (2H) indicated a C-22
a
configuration. Thus, 7 was identified
-glucopyranosyl)oxy]-2 ,14 ,22 -trihydroxy-
O-b- -galactopyranosyl-(1 ? 2)-O-[b- -xylo-
-glucopyranosyl-(1 ? 4)-b- -galactopy-
as
(25R)-26-[(b-
-furostan-3b-yl
D
a
a
a
5a
D
D
pyranosyl-(1 ? 3)]-O-b-
D
D
ranoside.
related; the terminal b-
D-glucopyranosyl residue of 1 and 6 was re-
The spectroscopic data for 8 (C56H92O28) showed that it had a
structure similar to 14. However, the 13C NMR signals at dC 152.3
(C) and 103.6 (C) indicated the presence of an olefinic moiety in
8. Furthermore, the Me-21 methyl doublet observed at dH 1.32
(J = 6.8 Hz) in the 1H NMR spectrum of 14 was absent from that
of 8, and was replaced by a methyl singlet at dH 1.62. The H-17 pro-
ton resonance appeared at dH 2.42 as a doublet with a coupling
constant of 10.1 Hz. These spectroscopic properties and the HMBC
correlations of C-20 (dC 103.6) and C-22 (dC 152.3) with H-17 (dH
2.42, d, J = 10.1 Hz), Me-21 (dH 1.62, s), and H2-23 (dH 2.23, m) sug-
placed by a b- -xylopyranosyl group in 11 and 14. No difference in
D
cytotoxicity was observed between 1 and 11, and between 6 and
14. On the other hand, the aglycones, 1a, 2a, 4a, 5a, and 9a, did
not show cytotoxicity against any cells.
Fig. 2 shows the time course of the antiproliferative effects of
the tumor-selective compounds 9 in HL-60 cells and A549 cells
at 3.0, 10, and 20 lg/ml. Interestingly, pseudo-furostanol glycoside
9 appeared to inhibit HL-60 and A549 cell growth in a time-depen-
dent manner. Morphological observation of HL-60 and A549 cells
stained with 40,6-diamidino-2-phenylindole (DAPI) suggested that
9 and 14 induced apoptosis in both HL-60 and A549 cells; the cells
displayed nuclear chromatin condensation and apoptotic bodies
(data not shown). Furthermore, an increase of caspase-3 activity
in HL-60 cells and A549 cells was evident after 16 h of treatment
with 9 and 14 (Fig. 3). Finally, the cell cycle distribution of HL-60
cells and A549 cells treated with 9 and 14 was analyzed by flow
cytometry (Table 5). Compounds 9 and 14 induced a sub-G1 pop-
ulation of 51.9 3.4% and 55.9 2.1% in HL-60 cells and
6.0 0.3% and 20.6 3.3% in A549 cells, respectively. This suggests
that 9 and 14 may induce apoptotic cell death through caspase-3
gested that 8 was the corresponding
coside of 14. This was supported by enzymatic hydrolysis of 8 with
b- -glucosidase giving spirostanol glycoside 11 and -glucose,
whereas enzymatic hydrolysis using naringinase gave spirostanol
sapogenin 1a (Gvazava and Kikoladze, 2006), -galactose,
-glucose, and -xylose. Thus, was assigned as (25R)-26-
[(b- -glucopyranosyl)oxy]-2 -hydroxy-5 -furost-20(22)-en-3b-yl
O-b- -galactopyranosyl-(1 ? 2)-O-[b- -xylopyranosyl-(1 ? 3)]-O-
b- -glucopyranosyl-(1 ? 4)-b- -galactopyranoside.
Compound 9 (C61H100O30) was also expected to be a D20(22)
D
20(22)-pseudo-furostanol gly-
D
D
D
D
D
8
D
a
a
D
D
D
D
-
pseudo-furostanol glycoside, based on the characteristic 1H and
13C NMR signals at dH 2.44 (d, J = 10.1 Hz, H-17) and 1.64 (s, Me-
21), and at dC 152.3 (C) and 103.6 (C), and the HMBC correlations
of C-20 (dC 103.6) and C-22 (dC 152.3) with H-17 (dH 2.44, d,
J = 10.1 Hz), Me-21 (dH 1.64, s), and H2-23 (dH 2.23, m). Enzymatic
activity. Additionally,
9
induced
a
G0/G1 population of
84.7 0.2% in A549 cells. This result implies that 9 also induce
cell-cycle arrest at the G0/G1 phase in A549 cells.
3. Concluding remarks
hydrolysis of 9 with b-
10 and -glucose, whereas acid hydrolysis of 9 gave the spirostanol
sapogenin (25R)-5 -spirostan-3b-ol (9a, tigogenin) (Gvazava and
Kikoladze, 2010), -arabinose, -galactose, -glucose, -rhamnose,
and -xylose. Therefore, 9 was the corresponding
20(22)-pseudo-
furostanol glycoside of 10: (25R)-26-[(b- -glucopyranosyl)oxy]
-5 -furost-20(22)-en-3b-yl O- -arabinopyranosyl-(1 ? 2)-O-[b-
-xylopyranosyl-(1 ? 3)]-O-b- -glucopyranosyl-(1 ? 4)-O-[
-galactopyranoside.
D-glucosidase gave the spirostanol glycoside
D
In conclusion, nine new and five known steroidal glycosides
were isolated from the bulbs of B. elegans and evaluated for their
cytotoxic activities against HL-60 and A549 tumor cells, and TIG-
3 normal cells. Compounds 9 and 14 increased caspase-3 activity
and induced cell accumulation at the sub-G1 phase of cell cycle
in HL-60 cells and A549 cells, respectively. These results indicated
that 9 and 14 had potent apoptotic activity in HL-60 cells and A549
cells. Previous studies of steroidal glycosides found that pseudo-
furostanol glycoside did not show cytotoxic activity (Liu et al.,
2003; Yokosuka et al., 2002). Thus, it is notable that pseudo-furost-
anol glycoside 9 appeared to inhibit tumor cells growth in a time-
dependent manner and induce apoptosis in HL-60 and A549 cells.
Compound 9 also induced cell-cycle arrest at the G0/G1 phase in
A549 cells.
a
L
D
D
L
D
D
D
a
a-L
D
D
a-L-
rhamnopyranosyl-(1 ? 2)]-b-
D
2.2. Cytotoxic activity
It has been reported that steroidal glycosides have potent anti-
cancer activity. Relationships between chemical structure and
cytotoxicity of steroidal glycosides have been studied (Yokosuka
et al., 2009a). The aglycone moiety and the sugar moiety play an
important role in the cytotoxicity.
Isolated compounds 1–14 were evaluated for their cytotoxic
activities against HL-60 and A549 tumor cells, and TIG-3 normal
cells. Etoposide and cisplatin were used as positive controls, and
4. Experimental
4.1. General
had IC50 values of 0.3 0.01 and 4.1 0.23
lM against HL-60 cells,
and 1.2 0.03 and 2.3 0.04 M against A549 cells, respectively.
Compounds 1, 5, 6, 9–11, 13, and 14 showed cytotoxic activity
against both HL-60 and A549 cells with IC50 values ranging from
l
Optical rotations were measured using an automatic digital
polarimeter (P-1030, Jasco, Tokyo, Japan). IR spectra were recorded
on a spectrophotometer (FT-IR 620, Jasco). NMR spectra were
recorded at 500 MHz for 1H NMR using standard Bruker pulse pro-
grams at 300 K (DRX-500, Bruker, Karlsruhe, Germany). Chemical
0.5 0.01 to 8.1 0.28
lM, respectively (Table 4). Compounds 6,
9, 11, 13, and 14 were selectively cytotoxic to the tumor cells