Cytotoxic steroidal saponins from Agave sisalana

Article abstract of DOI:10.1055/s-0030-1250672

Two new steroidal saponins, 8 and 10, along with 7 known steroidal sapogenins and saponins (1-7) and a furostanol saponin (9) were isolated from Agave sisalana Perrine ex Engelm. The structures of these two new compounds were identified and characterized by 1D and 2D NMR spectroscopy and mass spectrometry. In addition, acid hydrolysis and GC-FID were used to confirm the sugar moieties of 8 and 10. The cytotoxic effects of 1-10 on MCF-7, NCI-H460, and SF-268 cancer cells were evaluated, and among them, compound 10 proved to be the most cytotoxic with IC₅₀ values of 1.2, 3.8, and 1.5 μM, respectively. Georg Thieme Verlag KG Stuttgart.

Full text of DOI:10.1055/s-0030-1250672

Original Papers 929
Cytotoxic Steroidal Saponins from Agave sisalana
Authors
Pi-Yu Chen1, Chin-Hui Chen1, Ching-Chuan Kuo2, Tzong-Huei Lee3, Yueh-Hsiung Kuo4,5, Ching-Kuo Lee1,3
1
2
3
4
5
Affiliations
School of Pharmacy, Taipei Medical University, Taipei, Taiwan
National Institute of Cancer Research, National Health Research Institutes, Tainan, Taiwan
Graduate Institute of Pharmacognosy Science, Taipei Medical University, Taipei, Taiwan
Tsutzki Institute for Traditional Medicine, China Medical University, Taichung, Taiwan
Agricultural Biotechnological Research Center, Academia Sinica, Taipei, Taiwan
Key words
Abstract
sugar moieties of 8 and 10. The cytotoxic effects
of 1–10 on MCF-7, NCI-H460, and SF-268 cancer
cells were evaluated, and among them, com-
pound 10 proved to be the most cytotoxic with
IC50 values of 1.2, 3.8, and 1.5 µM, respectively.
"
l Agavaceae
!
"
l Agave sisalana
Two new steroidal saponins, 8 and 10, along with
7 known steroidal sapogenins and saponins (1–7)
and a furostanol saponin (9) were isolated from
Agave sisalana Perrine ex Engelm. The structures
of these two new compounds were identified
and characterized by 1D and 2D NMR spectros-
copy and mass spectrometry. In addition, acid hy-
drolysis and GC‑FID were used to confirm the
"
l steroidal saponin
"
l MCF‑7 cells
"
l NCI‑H460 cells
"
l SF‑268 cells
Supporting information available online at
http://www.thieme-connect.de/ejournals/toc/
plantamedica
Introduction
also evaluated for their cytotoxicity against MCF-
7, NCI-H460, and SF-268 cancer cells. Our results
were consistent with those reported previously
(unpublished data). In this study, we purified the
polar fraction of the methanolic extract to isolate
two new steroidal saponins, 8 and 10, along with
seven known steroidal sapogenins and saponins,
1–7, and a furostanol saponin, 9. Herein, we re-
port the isolation and identification of these com-
pounds and assess their cytotoxic effects on three
human cancer cell lines.
received
revised
accepted
June 6, 2010
Nov. 15, 2010
Dec. 7, 2010
!
Steroidal saponins are secondary metabolites
commonly found in several botanical species of
the family Agavaceae [1,2]. The basic chemical
structure of steroidal saponins is a glycosylated
C27 steroidal aglycone skeleton. On the basis of
the ring structural features, C27 steroidal agly-
cones are further categorized into spirostanes
and furostanes. Furostanol saponins, character-
ized by a pentacyclic system, are the biosynthetic
precursors of spirostanes, which have a 6–ring
structure [3]. In most steroidal saponins, a hy-
droxyl group located at the C-3 or C-26 has a gly-
cosidic linkage to a sugar moiety. Steroidal sapo-
nins have been reported to possess many useful
properties, such as antifungal, antibacterial, and
cytotoxic activities [1].
The family Agavaceae includes more than 300
species and is common in tropical and subtropical
regions [4]. In southern Taiwan, two species of the
genus, Agave–Agave americana and Agave sisala-
na, have been cultivated since 1918 for the fiber
industry. Previously, it was shown that metha-
nolic extracts of A. sisalana containing homoiso-
flavonoids and flavonoids have immunomodula-
tory effects [5]. Besides, steroidal saponins are re-
ported to be rich in A. sisalana [6], and they exhib-
ited potent cytotoxicities [1,7]. In our preliminary
study, the methanolic extracts of A. sisalana were
Bibliography
DOI http://dx.doi.org/
10.1055/s-0030-1250672
Published online January 17,
2011
Planta Med 2011; 77: 929–933
© Georg Thieme Verlag KG
Stuttgart · New York ·
ISSN 0032‑0943
Correspondence
Prof. Ching-Kuo Lee
School of Pharmacy,
Taipei Medical University
No. 250 Wu-Xin Street
Taipei 110
Taiwan
Phone: + 886227361661
ext. 6150
Materials and Methods
!
General experimental procedures
Optical rotations were measured in methanol
(MeOH) on a JASCO P-1020 digital polarimeter. IR
spectra were measured on a JASCO FT/IR 4100
spectrometer. 1H- and 13C‑NMR spectra were ac-
quired on a Bruker DRX-500 SB spectrometer. LC/
MS/MS spectra were recorded on an Agilent 1100
HPLC system and an ABI API 4000Q Trap. Mass
spectra were recorded on a JEOL JMS-700 mass
spectrometer using the positive ion mode and
high-resolution fast atom bombardment (FAB).
Semipreparative HPLC was performed on a Hitachi
L-7110 HPLC with a refractive index detector
(Thermo Separation Products). A Phenomenex Lu-
na silica column (5 µm, 10 × 250 mm, 3 mL·min−1
Fax: + 886223772265
cklee@tmu.edu.tw
Correspondence
Prof. Yueh-Hsiung Kuo
Tsutzki Institute
for Traditional Medicine
China Medical University
No. 91 Hsueh-Shih Road
Taichung 404
Taiwan
Phone: + 886422053366
ext. 5701
Fax: + 886422071693
yhkuo@ntu.edu.tw
Chen P-Y et al. Cytotoxic Steroidal Saponins… Planta Med 2011; 77: 929–933

930 Original Papers
Fig. 1 Chemical structures of compounds 8
and 10.
flow rate) and a Merck Lichrospher 100 RP-8 reverse-phase col-
umn (10 µm, 10 × 250 mm, 3 mL·min−1 flow rate) were used for
normal-phase and reverse-phase chromatography, respectively.
Silica gel (Merck, Geduran® Si 60 0.063–0.2 mm) and octadecyl-
silanized (ODS)-silica gel (BioSil, ODS‑W 45–55 µm) were used
for column chromatography. GC analysis was performed on an
Agilent GC 7890A gas chromatograph equipped with a flame ion-
ization detector. An Agilent DB-5 GC column was chosen for GC
analysis.
phase HPLC on a semipreparative normal phase column (Luna sili-
ca, 5 µm, 10 × 250 mm; Phenomenex) at a flow rate of 3 mL·min−1
,
with a mixture of n-hexane:ethyl acetate (7:1, v/v) followed by
n-hexane:ethyl acetate:acetone (14:2:1, v/v/v) to obtain com-
pound 1 (5.2 mg, tR = 11.62 min) and compound 2 (3.5 mg, tR =
11.65 min). Compound 3 (8.4 mg, tR = 12.15 min) was obtained
from portion 7 (fractions 89–99) by HPLC with an isocratic mix-
ture of n-hexane-acetone (16:7, v/v).
Silica gel column chromatography of the n-butanol layer (546.0 g)
was eluted stepwise with chloroform-methanol (14:1, 7:1, 7:2,
7:3, 7:4, 7:5, 7:6, 7:7, and 7:14, v/v) and finally with pure
methanol. Fractions of 500 mL each were collected and moni-
tored by TLC using chloroform-methanol-water (14:6:1, v/v/v)
and were observed at 254 nm. Similar fractions were pooled into
10 portions. Portion 1 (fractions 1–9) was separated by normal-
phase HPLC with a mixture of n-hexane-chloroform-ethyl ace-
tate-acetone (18:2:5:1, v/v/v/v) to obtain compound 4 (3.5 mg,
tR = 30.85 min) and compound 5 (1.1 mg, tR = 32.21 min). Portion
4 (fractions 65–76) was separated on a Sephadex LH-20 column
(3 × 50 cm) with chloroform-methanol (5:2, v/v) and collected
in 200 mL of each subfraction. The precipitate in subfraction 6
was washed with methanol to obtain compound 6 (143.5 mg).
Then, the supernatant from subfraction 6 was separated further
by reverse-phase HPLC on a semipreparative reverse-phase col-
umn (Lichrospher® 100 RP-8 column, 10 µm, 10 × 250 mm;
Merck) at a flow rate of 3 mL·min−1, with a mixture of water-
Plant material
Leaves of Agave sisalana Perrine ex Engelm (Agavaceae) were col-
lected in Heng-Chun Town, Ping-Tung County, Taiwan, in July
2004. Dr. Chi-I Chang (Graduate Institute of Biotechnology, Na-
tional Pintung Science and Technology University, Taiwan) iden-
tified the material. A voucher specimen (LCK9306) was deposited
in the Graduate Institute of Pharmacy, Taipei Medical University,
Taipei, Taiwan.
Extraction and isolation
The air-dried leaves (61.1 kg) of A. sisalana were cut into small
pieces and extracted with 100 L methanol at room temperature
for seven days. After removal of the solvent under reduced pres-
sure, the methanolic extract was dissolved in 4 L of water and
then partitioned between ethyl acetate and water (3 × 4 L,
30 min each). Then, the extracted water layer was partitioned
with n-butanol (3 × 4 L, 30 min each). Silica gel column chroma-
tography of the ethyl acetate layer (69.6 g) was eluted stepwise
with n-hexane-ethyl acetate (85:15, 80:20, 75:25, 65:35,
55:45, 40:60, and 20:80, v/v) and finally with pure ethyl ace-
tate. Fractions of 500 mL each were collected and monitored by
TLC using n-hexane-ethyl acetate (1:1, v/v) and observed at
254 nm. Fractions with similar components were pooled into 8
portions. Portion 2 (fractions 54–65) was separated by normal-
methanol (1:4, v/v) to obtain compound
7 (16.6 mg, tR =
19.65 min). Portion 8 (fractions 85–95) was separated by silica
gel column chromatography with a mixture of chloroform-meth-
anol-water (7:3:0.5, v/v/v) to obtain compound 8 (22.7 mg). Por-
tion 9 (fractions 96–105) was subjected to (ODS) silica gel column
chromatography and eluted with water-methanol-acetonitrile
(2:1:1, v/v/v) to give compound 9 (793.2 mg). Finally, the wash
subfraction of portion 9 was subjected again to silica gel column
Chen P-Y et al. Cytotoxic Steroidal Saponins… Planta Med 2011; 77: 929–933

Original Papers 931
chromatography and eluted with chloroform-methanol-water
(10:3:0.5, v/v/v) to obtain compound 10 (22.2 mg).
Table 1 1H and 13C‑NMR chemical shift assignments for the aglycone moiety
of compounds 8 and 10 in C5D5N.
Compound 8: colorless, amorphous powder; [α]2D3 − 50.0 (c 0.10;
8
10
MeOH); IR (film) νmax 3410, 2925, 1700, 1645, 1374, 1042 cm−1
;
1H NMR (C5D5N, 500 MHz) and 13C NMR (C5D5N, 125 MHz) as-
Position
δC
δH
δC
δH
"
1
2
36.7
29.8
77.3
34.7
44.6
28.7
31.5
34.4
55.7
36.4
38.1
0.69, 1.30
1.55, 1.97
3.89
37.2
30.0
77.5
34.9
44.7
29.0
32.5
35.3
54.5
35.9
21.3
0.77, 1.49
1.61, 2.04
3.93
signments (l Tables 1 and 2); ESI‑MS (neg. ion mode) m/z
1325.4 [M – H]−, 1193.6, 1031.5, 885.5, 753.5, 591.4, HR FAB‑MS
(pos. ion mode) m/z 1349.5981 [M + Na]+, (calcd. for C61H98O31Na,
1349.5990).
3
4
1.22, 1.76
0.84
1.33, 1.77
0.87
5
Compound 10: colorless, amorphous powder; [α]2D3 − 63.0 (c 0.10;
6
1.11
1.03, 1.10
0.77, 1.50
1.38
MeOH); IR (film) νmax 3449, 2928, 1631, 1062 cm−1 1H NMR
;
7
1.56, 2.10
1.76
(C5D5N, 500 MHz) and 13C NMR (C5D5N, 125 MHz) assignments
8
−
"
(l Tables 1 and 2); ESI‑MS (neg. ion mode) m/z 1341.4 [M – H] ,
1209.6, 1047.6, 901.5, 739.5, 577.4, HR FAB‑MS (pos. ion mode)
m/z 1365.6293 [M + Na]+ (calcd. for C62H102O31Na, 1365.6303).
9
0.88
0.48, t, J = 10.7
10
11
2.21, brd, J = 13.6
2.36, t, J = 13.6
1.19, 1.38
1.01, 1.64
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
213.2
55.5
56.0
31.9
79.8
54.4
16.2
11.8
42.7
14.0
109.5
31.8
29.3
30.7
67.1
40.2
40.8
56.5
32.2
81.3
62.9
16.7
12.4
42.5
15.0
109.8
26.4
26.3
27.6
65.2
Acid hydrolysis and sugar analysis
Compounds 8 and 10 (2.0 mg each) were hydrolyzed with 4 N
aqueous HCl (0.5 mL) at 60°C for 1 h. The reaction mixture was
partitioned between ethyl acetate and water. The water layer
was neutralized by passing it through an ion exchange solid
phase extraction (SPE) cartridge (Strata X‑AW 33 µm polymeric
Weak Anion; Phenomenex) [8]. After removal of the solvent
under reduced pressure, the evaporation residue of the aqueous
layer was dissolved in anhydrous pyridine (1 mL), and L-cysteine
methyl ester hydrochloride (2.3 mg) was added. The mixture was
stirred at room temperature for 24 h. Then 300 µL of HMDS-TMCS
(hexamethyldisilazane-trimethylchlorosilane, 2:1) was added,
and the mixture was stirred at room temperature for 30 min.
The precipitate was centrifuged off, and 20 µL of supernatant
was added into ethyl acetate (950 µL). After being well mixed
and filtered, the solution was analyzed by GC‑FID [9]. The pro-
gram of oven temperature was 180 to 300°C at 4°C/min; injec-
tion temperature 250°C; the carrier had a heat flow rate of
0.8 mL/min. The silylated standards of D-galactose, D-glucose,
D-xylose, and L-rhamnose were detected at 19.75, 19.32, 15.27,
and 16.56 min, respectively. Identification of D-galactose, D-glu-
cose, D-xylose, and L-rhamnose was carried out for compounds 8
and 10, giving peaks respectively at 19.76, 19.32, 15.27, and
16.56 min for compound 8, and at 19.80, 19.36, 15.32, and
16.60 min for compound 10.
1.34
1.00
0.75, 1.56
4.47
1.38, 2.00
4.50
2.74, t, J = 7.4
1.07, s
1.77
0.80, s
0.63, s
1.88
0.65, s
1.91, t, J = 6.7
1.33, d, J = 6.7
1.13, d, J = 6.7
1.60, 1.70
1.56
1.44, 1.89
1.34, 2.12
1.57
1.56
3.47, 3.59
3.35, d, J = 10.8
4.06
27
17.5
0.68, d, J = 5.4
16.4
1.06, d, J = 7.1
number data as percentage of control versus compound concen-
tration, and the IC50 was the point on the graph where 50%
growth inhibition occurred. Each data point represented the
mean of at least 3 independent experiments run in triplicate
[10]. Actinomycin D (98%, Sigma) was used as the positive con-
trol. The purity of the test samples was greater than 95% as veri-
fied by HPLC and NMR.
Growth inhibition assay
Supporting information
Human MCF-7 breast cancer and NCI-H460 non-small cell lung
cancer cell lines were purchased from the American Type Culture
Collection (ATCC). Human SF-268 glioblastoma cell line was pur-
chased from the National Cancer Institute (NCI). These three cell
lines were maintained in Dulbeccoʼs modified Eagleʼs medium
supplemented with 10% fetal calf serum and nonessential amino
acids (Life Technologies, Inc.) at 37°C in a humidified incubator
with 5% CO2 and utilized by the Developmental Therapeutics
Program of the NCI for anticancer compound prescreen. Cells in
logarithmic growth phase were cultured at a density of 10000
cells·mL−1 in a 24-well plate. The cells were treated with various
concentrations of compounds 1–10 for 72 h. Subsequently, the
cells were fixed and stained with 50% ethanol containing 0.5%
methylene blue for 30 min. The plates were washed 5 times with
water and then air-dried. The resulting residue was dissolved in
1% N-lauroyl-sarcosine, and then the optical density was mea-
sured at 570 nm using a microplate reader. Cell number data
were normalized to the percentage of vehicle-treated control
and then graphed. IC50 values were determined by graphing cell
1H, 13C NMR, and selective TOCSY spectra of compounds 8 and 10
are available as Supporting Information.
Results and Discussion
!
Chromatographic fractionation of a methanolic extract of A. sisa-
lana leaves yielded 2 new compounds, 8 and 10 (l Fig. 1), and
"
8 known compounds, 1–7 and 9. The structures of the com-
pounds were identified by using spectroscopic analysis, 2D NMR,
and acid hydrolysis. Compounds 1–7 and 9 were identified by
comparing their spectroscopic data with published results for
tigogenin (1) [11], neotigogenin (2) [11], hecogenin (3) [12], neo-
hecogenin (4) [13], rockogenin (5) [14], cantalasaponin-1 (6)
[15], hecogenin 3-O-β-D-glucopyranosyl-(1→2)-[β-D-xylo-
pyranosyl-(1→3)]-β-D-glucopyranosyl-(1→4)-β-D-galactopyra-
noside (7) [7], and polianthosides E (9) [16].
Compound 8 was obtained as a colorless amorphous powder. Its
molecular formula was assigned as C61H98O31 on the basis of the
Chen P-Y et al. Cytotoxic Steroidal Saponins… Planta Med 2011; 77: 929–933

932 Original Papers
spectrum also showed 6 anomeric proton signals at δH 4.84, 5.07,
5.11, 5.20, 5.41, and 5.55 indicating that there are 6 monosac-
charide moieties in 8. Acid hydrolysis of compound 8 with 4 N
HCl yielded hecogenin (3), the aglycone of 8, which was identi-
fied by comparing its 1H- and 13C‑NMR spectra with published
results [12]. The water layer of hydrolysis products of compound
8 was silylated and then analyzed by GC‑FID. GC analyses showed
that the sugar composition of 8 includes L-rhamnopyranose, D-
xylopyranose, D-glucopyranose, and D-galactopyranose. In the
HMBC spectrum, the anomeric proton signals of galactose at δH
4.84 correlated with the C-3 (δC 77.3) of 8. The sequence of the
oligosaccharide chain was deduced from HMBC correlations of
δH 5.11 (Glc I H-1)/δC 80.0 (Gal C-4), δH 5.20 (Xyl I H-1)/δC 80.8
(Glc I C-2), δH 5.41 (Rha H-1)/δC 74.5 (Xyl I C-3), δH 5.55 (Glc II H-
1)/δC 86.9 (Glc I C-3) and δH 5.07 (Xyl II H-1)/δC 87.1 (Glc II C-3).
Furthermore, the δH of each proton of the monosaccharides was
fully assigned by a selective TOCSY experiment (shown in Sup-
porting Information). In the TOCSY spectrum, the nearby protons
would be enhanced when exciting the H-1 of the monosugar. As
a result, 8 was determined to be hecogenin 3-O-α-L-rhamno-
pyranosyl-(1 → 3)-β-D-xylopyranosyl-(1 → 2)-[β-D-xylopyrano-
syl-(1 → 3)-β-D-glucopyranosyl-(1 → 3)]-β-D-glucopyranosyl-
(1 → 4)-β-D-galactopyranoside.
Compound 10 was obtained as a colorless amorphous powder. Its
molecular formula was assigned as C62H102O31 on the basis of
13C‑NMR and HR‑FAB‑MS data (m/z 1365.6293 [M + Na]+, calcd.
1365.6303). In the 1H‑NMR spectrum of 10, the 2 methyl singlets
at δH 0.63 and 0.80 as well as the 2 methyl doublets at δH 1.06
(J = 7.1 Hz), 1.13 (J = 6.8 Hz) were in agreement with the charac-
teristic signals of the aglycone moiety of neotigogenin (2) [11].
This result was also reflected in the 13C‑NMR data of 10, in which
the carbonyl group at δC 213.2 of 8 was absent. The type of mono-
saccharide of 10 was the same as that of 8 as shown by 1H‑NMR
and GC‑FID data. However, the oligosaccharide linkages of 10
were quite different from those of 8. In the HMBC spectrum of
10, the long range correlations of δH 4.86 (Gal-1)/δC 77.5 (C-3),
Table 2 1H- and 13C‑NMR chemical shift assignments for the sugar moieties of
compounds 8 and 10 in C5D5N.
8
10
Posi-
tion
Gal
δC
δH
Posi-
tion
Gal
δC
δH
1
102.6
4.84, d,
J = 7.4
1
102.5
4.86, d,
J = 7.5
2
73.2
75.6
80.0
75.1
60.9
4.37, m
4.10, m
4.47, m
3.99, m
4.63, m
4.19, m
5.11, d,
J = 8.0
2
73.2
75.6
80.0
75.4
60.8
4.41, m
4.11, m
4.56, m
3.96, m
4.65, m
4.21, m
5.08, d,
J = 7.3
3
3
4
4
5
5
6a
6b
1
6a
6b
1
GlcI
GlcII
Xyl I
Rha
Xyl II
104.8
GlcI
104.8
2
80.8
86.9
70.4
77.4
62.6
4.29, m
4.05, m
3.70, m
3.89, m
4.38, m
4.04, m
5.55, d,
J = 7.9
2
80.9
88.0
70.6
78.2
63.0
4.29, m
4.12, m
3.72, m
3.81, m
4.41, m
3.96, m
5.17, m
3
3
4
4
5
5
6a
6b
1
6a
6b
1
104.0
GlcII
104.2
2
75.3
87.1
69.3
78.3
62.2
4.10, m
4.09, m
4.01, m
4.05, m
4.48, m
4.29, m
5.20, d,
J = 7.5
2
75.6
76.6
78.0
75.4
61.2
4.11, m
4.06, m
4.29, m
3.98, m
4.21, m
4.03, m
5.74,
3
3
4
4
5
5
6a
6b
1
6a
6b
1
104.7
Rha
102.7
brs
2
75.4
74.5
76.0
64.2
3.92, m
4.00, m
4.10, m
4.16, m
3.44, m
2
3
4
5
6
72.6
72.8
74.0
70.5
18.6
4.61, m
4.50, m
4.31, m
4.91, m
1.67, d,
J = 5.9
3
4
5a
5b
1
99.7
5.41,
GlcIII
1
104.0
5.56, d,
J = 6.5
brs
δ
H 5.08 (Glc I-1)/δC 80.0 (Gal-4), δH 5.17 (Glc II-1)/δC 80.9 (Glc I-
2
3
4
5
6
72.5
72.6
74.0
70.0
18.7
4.47, m
4.48, m
4.27, m
4.79, m
1.62, d,
J = 6.0
2
75.1
87.1
70.8
77.5
62.1
4.05, m
4.06, m
4.08, m
3.91, m
4.46, m
2), δH 5.74 (Rha-1)/δC 78.0 (Glc II-4), δH 5.56 (Glc III-1)/δC 88.0
(Glc I-3), and δH 5.09 (Xyl-1)/δC 87.1 (Glc III-3) were used to de-
duce the sequence of the oligosaccharide chain. In addition, the
3
4
5
δ
H of each proton of the monosaccharide was determined by the
6a
selective TOCSY spectrum (shown in Supporting Information).
Compared with the reference of Ding et al. [17], the sugar moiety
of compound 10 was the same as dongnoside A. The chemical
shifts of C-22 to C-27 of compound 10 were more up field than
the signals of dongnoside A, and that was the characteristic dif-
ference between C-25R and C-25S. Besides, the proton data
matched those of the 25S spirostane-type steroidal saponins (H-
26a-H26b = 0.71 > 0.35) [11,18]. The 13C‑NMR data of the agly-
cone of compound 10 was the same as the compound 2, neotigo-
genin [11]. Consequently, 10 was determined to be neotigogenin
3-O-α-L-rhamnopyranosyl-(1→4)-β-D-glucopyranosyl-(1→2)-
[β-D-xylopyranosyl-(1→3)-β-D-glucopyranosyl-(1→3)]-β-D-
glucopyranosyl-(1→4)-β-D-galactopyranoside.
1
106.2
5.07, d,
J = 7.5
6b
4.29, m
2
75.4
77.6
70.4
67.2
3.92, m
4.05, m
4.05, m
4.17, m
3.52, m
Xyl
1
106.2
75.6
77.8
69.2
67.3
5.09, m
3.96, m
4.07, m
4.06, m
4.19, m
3.54, m
3
2
4
3
5a
5b
4
5a
5b
13C‑NMR and HR‑FAB‑MS data (m/z 1349.5981 [M + Na]+, calcd.
1349.5990). The 1H‑NMR spectrum of 8 showed 2 methyl singlet
signals at δ 0.65 and 1.07 as well as 2 methyl doublet signals at δ
0.68 (J = 5.4 Hz) and 1.33 (J = 6.7 Hz), which is characteristic of a
The growth inhibition assays of purified compounds 1–10 using
MCF-7, NCI-H460, and SF-268 cancer cells showed that 4 com-
pounds (7–10) were significantly cytotoxic compared to the pos-
13
"
steroidal skeleton. In the C‑NMR spectrum of 8 (l Table 1), a
quaternary carbon at δ 109.5 indicated the distinctive hemiacetal
C-22 of the spirostanol skeleton. The heteronuclear multiple bond
coherence (HMBC) correlations of the proton signals at H2-11 (δH
2.21 and 2.36) and H3-18 (δH 1.07) with the carbonyl signal at δC
213.2 indicated that a carbonyl was located at C-12. The 1H‑NMR
"
itive control actinomycin D (l Table 3). In particular 10 was the
most cytotoxic compound; its IC50 values against MCF-7, NCI-
H460, and SF-268 cells were 1.2, 3.8, and 1.5 µM, respectively. In
addition, steroidal saponins with a sugar moiety (compounds 7–
"
10) were more cytotoxic than their aglycones (1–5) (l Table 3).
Chen P-Y et al. Cytotoxic Steroidal Saponins… Planta Med 2011; 77: 929–933

Original Papers 933
References
Table 3 IC50 values (µM SD) of compounds 1–10 and actinomycin D in hu-
man NCI-H460, MCF-7, and SF268 cancer cellsa.
1 Sparg SG, Light ME, Van Staden J. Biological activities and distribution of
plant saponins. J Ethnopharmacol 2004; 94: 219–243
2 Hostettmann K, Marston A. Saponins, Appendix 3. Cambridge: Cam-
bridge University Press; 1995
3 Yang CR, Zhang Y, Jacob MR, Khan SI, Zhang YJ, Li XC. Antifungal activity
of C‑27 steroidal saponins. Antimicrob Agents Chemother 2006; 50:
1710–1714
4 Jin JM, Zhang YJ, Yang CR. Four new steroid constituents from the waste
residue of fibre separation from Agave americana leaves. Chem Pharm
Bull 2004; 52: 654–658
Compound
NCI-H460
> 20
MCF-7
> 20
SF268
> 20
1
2
> 20
> 20
> 20
> 20
> 20
> 20
> 20
> 20
> 20
> 20
> 20
> 20
> 20
> 20
> 20
3
4
5
6
5 Chen PY, Kuo YC, Chen CH, Kuo YH, Lee CK. Isolation and immunomodu-
latory effect of homoisoflavones and flavones from Agave sisalana Per-
rine ex Engelm. Molecules 2009; 14: 1789–1795
6 Zou P, Fu J, Yu HS, Zhang J, Kang LP, Ma BP, Yan XZ. The NMR studies on
two new furostanol saponins from Agave sisalana leaves. Magn Reson
Chem 2006; 44: 1090–1095
7
5.3 1.8
6.5 1.1
> 20
11.9 2.6
9.5 4.8
> 20
4.0 2.2
8
8.2 1.2
7.5 1.4
1.5 0.8
7.5 5.2
9
10
3.8 2.7
2.6 1.6
1.2 0.1
Actinomycin Db
31.1 2.9
7 Mimaki Y, Kanmoto T, Kuroda M, Sashida Y, Nishino A, Satomi Y, Nishino
H. Steroidal saponins from the underground parts of Hosta longipes
and their inhibitory activity on tumor promoter-induced phospholipid
metabolism. Chem Pharm Bull 1995; 43: 1190–1196
8 Yokosuka A, Mimaki Y. Steroidal glycosides from the underground parts
of Trillium erectum and their cytotoxic activity. Phytochemistry 2008;
69: 2724–2730
a Cells were treated with various concentrations of test drugs for 72 hours. Cell growth
was determined with the methylene blue dye assay. The IC50 value resulting from 50%
inhibition of cell growth was calculated. Each value represents the means SD of three
independent experiments. b Actinomycin D was used as the positive control
9 Ito A, Chai HB, Kardono LBS, Setowati FM, Afriastini JJ, Riswan S, Farns-
worth NR, Cordell GA, Pezzuto JM, Swanson SM, Kinghorn AD. Saponins
from the bark of Nephelium maingayi. J Nat Prod 2004; 67: 201–205
10 Finlay GJ, Baguley BC, Wilson WR. A semiautomated microculture
method for investigating growth inhibitory effects of cytotoxic com-
pounds on exponentially growing carcinoma cells. Anal Biochem
1984; 139: 272–277
These results confirm that the sugar moiety plays an important
role in the bioactivity of steroidal saponins.
Acknowledgements
!
11 Tori K, Seo S, Terui Y, Nishikawa J, Yasuda F. Carbon-13 NMR spectra of
5β-steroidal saponins. Reassignment of the F-ring carbon signals of
(25S)-spirostans. Tetrahedron Lett 1981; 22: 2405–2408
12 Wu TS, Shi LS, Kuo SC. Alkaloids and other constituents from Tribulus
terrestris. Phytochemistry 1999; 50: 1411–1415
This investigation was supported by the grant of the National Sci-
ence Council of the Republic of China, Taiwan Department of
Health Clinical Trial Research Center of Excellence (DOH99-TD-
B-111-004), and Health Cancer Research Center of Excellence
(DOH99-TD-B-111-005). We are appreciative of Ms. Shwu-Huey
Wang and Ms. Shou-Ling Huang for the NMR data acquisition in
the Instrumentation Center of Taipei Medical University and Na-
tional Taiwan University, and the YungShin Group for supported
MS instrumentation. We also thank Prof. Shang-Tzen Chang for
supported GC‑FID instrumentation.
13 Jaffer AJ, Crabb TA, Turner CH. Hydroxy group substituent effects on 1
H
NMR chemical shifts in the structural elucidation of spirostanes. Org
Magn Reson 1983; 21: 576–579
14 Nakano K, Hara Y, Murakami K, Takaishi Y, Tomimatsu T. 12-Hydroxy
steroidal glycosides from the caudex of Yucca gloriosa. Phytochemistry
1991; 30: 1993–1995
15 Sati OP, Pant G. Cantalasaponin-1, a novel spirostanol bisdesmoside
from Agave cantala. J Nat Prod 1985; 48: 395–399
16 Jin JM, Zhang YJ, Yang CR. Spirostanol and furostanol glycosides from
the fresh tubers of Polianthes tuberose. J Nat Prod 2004; 67: 5–9
17 Ding Y, Tian RH, Yang CR, Chen YY, Nohara T. Two new steroidal sapo-
nins from dried fermented residues of leaf-juices of Agave sisalana for-
ma Dong No. 1. Chem Pharm Bull 1993; 41: 557–560
18 Agrawal PK. Spectral assignments and reference data. Magn Reson
Chem 2003; 41: 965–968
Chen P-Y et al. Cytotoxic Steroidal Saponins… Planta Med 2011; 77: 929–933