Structures, chemotaxonomic significance, cytotoxic and Na+,K+-ATPase inhibitory activities of new cardenolides from Asclepias curassavica

Article abstract of DOI:10.1039/c4ob01545b

Five new cardenolide lactates (1-5) and one new dioxane double linked cardenolide glycoside (17) along with 15 known compounds (6-16 and 18-21) were isolated from the ornamental milkweed Asclepias curassavica. Their structures were elucidated by extensive spectroscopic methods (IR, UV, MS, 1D- and 2D-NMR). The molecular structures and absolute configurations of 1-3 and 17 were further confirmed by single-crystal X-ray diffraction analysis. Simultaneous isolation of dioxane double linked cardenolide glycosides (17-21) and cardenolide lactates (1-5) provided unique chemotaxonomic markers for this genus. Compounds 1-21 were evaluated for the inhibitory activities against DU145 prostate cancer cells. The dioxane double linked cardenolide glycosides showed the most potent cytotoxic effect followed by normal cardenolides and cardenolide lactates, while the C21 steroids were non-cytotoxic. Enzymatic assay established a correlation between the cytotoxic effects in DU145 cancer cells and the K_i for the inhibition of Na⁺,K⁺-ATPase. Molecular docking analysis revealed relatively strong H-bond interactions between the bottom of the binding cavity and compounds 18 or 20, and explained why the dioxane double linked cardenolide glycosides possessed higher inhibitory potency on Na⁺,K⁺-ATPase than the cardenolide lactate. This journal is

Full text of DOI:10.1039/c4ob01545b

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
Structures, chemotaxonomic significance,
cytotoxic and Na+,K+-ATPase inhibitory activities
of new cardenolides from Asclepias curassavica†
Cite this: Org. Biomol. Chem., 2014,
12, 8919
Rong-Rong Zhang,a Hai-Yan Tian,a Ya-Fang Tan,a Tse-Yu Chung,b Xiao-Hui Sun,a
Xue Xia,a Wen-Cai Ye,a David A. Middleton,c Natalya Fedosova,d Mikael Esmann,*d
Jason T. C. Tzen*b and Ren-Wang Jiang*a
Five new cardenolide lactates (1–5) and one new dioxane double linked cardenolide glycoside (17) along
with 15 known compounds (6–16 and 18–21) were isolated from the ornamental milkweed Asclepias cur-
assavica. Their structures were elucidated by extensive spectroscopic methods (IR, UV, MS, 1D- and
2D-NMR). The molecular structures and absolute configurations of 1–3 and 17 were further confirmed by
single-crystal X-ray diffraction analysis. Simultaneous isolation of dioxane double linked cardenolide
glycosides (17–21) and cardenolide lactates (1–5) provided unique chemotaxonomic markers for this
genus. Compounds 1–21 were evaluated for the inhibitory activities against DU145 prostate cancer cells.
The dioxane double linked cardenolide glycosides showed the most potent cytotoxic effect followed by
normal cardenolides and cardenolide lactates, while the C21 steroids were non-cytotoxic. Enzymatic
assay established a correlation between the cytotoxic effects in DU145 cancer cells and the Ki for the inhi-
bition of Na+,K+-ATPase. Molecular docking analysis revealed relatively strong H-bond interactions
between the bottom of the binding cavity and compounds 18 or 20, and explained why the dioxane
double linked cardenolide glycosides possessed higher inhibitory potency on Na+,K+-ATPase than the
cardenolide lactate.
Received 22nd July 2014,
Accepted 17th September 2014
DOI: 10.1039/c4ob01545b
www.rsc.org/obc
Asclepias curassavica L. (commonly called Mexican butterfly
weed or tropical milkweed) is a species of flowering plant in
Introduction
Cardenolides represent a class of steroids with a five-mem- the milkweed family Asclepiadaceae. This species is native to
bered lactone ring at C-17. This class of compounds bind to the American tropics. Currently, it is a common ornamental
the extracellular surface of the housekeeping membrane garden plant and cultivated as a source of food for butterflies
protein Na+,K+-ATPase.1 In recent years, the anticancer effects worldwide.3 This species is a rich source of cardenolides. Early
of this class of compounds received wide attention, and was phytochemical studies on this herb collected from North
found to be related to the modifications in the Na+,K+-ATPase America was carried out 50 years ago, which resulted in the
signalosome and the subsequent multiple downstream signal- isolation of calotropin as a cytotoxic principle against the
ing pathways.2
nasopharynx cancer cells.4 Subsequent studies on this herb
collected from Taiwan and Yunnan provinces of China led to
the identification of a series of cardenolides and their glyco-
sides with strong cytotoxicities against human A549 (lung car-
cinoma cell), MCF-7 and MDA-MB-231 (both breast carcinoma
cells), and HepG2 (hepatoma cell).5–7 In our continuous search
for structurally unique and biologically interesting cardiotonic
steroids from the plant kingdom, this plant cultivated in
Guangdong province was chosen for phytochemical investi-
gation. Five new cardenolide lactates (1–5) and one dioxane
double linked cardenolide glycoside (17) along with 15 known
compounds, i.e. calotropagenin (6),8 5α-card-20(22)-enolide,
11β-hydroxycorotoxigenin (7),9 frugoside (8),10 ascleposide
(9),11 calactinic acid (10),12 digitoxigenin (11),13 sarcostin
(12),14 12-O-benzoyldeacylmetaplexigenin (13),15 curassavoside
aInstitute of Traditional Chinese Medicine and Natural Products, College of
Pharmacy, Jinan University, Guangzhou 510632, P. R. China.
E-mail: trwjiang@jnu.edu.cn; Fax: +86-20-85221559; Tel: +86-20-85221016
bGraduate Institute of Biotechnology, National Chung-Hsing University, Taichung
40227, Taiwan, China. E-mail: tctzen@dragon.nchu.edu.tw
cDepartment of Chemistry, University of Lancaster, Lancaster LA1 4YB, UK.
E-mail: d.a.middleton@liverpool.ac.uk
dDepartment of Biomedicine, Aarhus University, DK-8000 Aarhus, Denmark.
E-mail: me@biophys.au.dk
†Electronic supplementary information (ESI) available: HRESIMS, IR, UV, 1D
and 2D NMR spectra of 1–5 and 17, the dose–response curves showing inhi-
bition of Na+,K+-ATPase by compounds 2, 3, 10 and 18 . In addition, crystallo-
graphic data of 1–3 and 17 in standard CIF. CCDC 1014923–1014926. For ESI
and crystallographic data in CIF or other electronic format see DOI: 10.1039/
c4ob01545b
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8919–8929 | 8919

View Article Online
Paper
Organic & Biomolecular Chemistry
Scheme 1 Structural formulae of compounds 1–21.
A (14),15 12-O-benzoylsarcostin (15),15 curassavoside B (16),15 3430, 1733 and 1716 cm−1, indicating the presence of hydroxyl
calotropin (18),16 16α-acetoxycalotropin (19),17 calactin (20)18 and carbonyl functionalities. 1H and 13C NMR spectra of 1
and asclepin (21)19 (Scheme 1), were isolated and structurally (Table 1) showed the typical signals for a butenolactone ring
characterized by spectroscopic analysis in combination with [δH 5.89 (1H, s, H-22), 4.99 (1H, dd, J = 18.4, 1.5 Hz, H-21β),
single-crystal X-ray diffraction. Herein we report the structural and 4.94 (1H, dd, J = 18.4, 1.5 Hz, H-21α), and δC 177.3, 178.5,
elucidation, chemotaxonomic significance and biological 117.8 and 75.3], two methyl singlets [δH 0.85 (3H, s, H-19) and
activities of these compounds.
0.88 (3H, s, H-18), δC 12.5 and 16.4], one methyl doublet
[δH 1.35 (3H, d, J = 6.9 Hz, H-3′), δC 20.6], resonances for two
oxygenated carbons [δH 4.18 (1H, q, J = 6.9 Hz, H-2′) and 4.73
(1H, m, H-3), and δC 67.9 and 75.5], and an additional ester car-
bonyl at δC 176.0. The 13CNMR chemical shifts of all the hydro-
genated carbons could be assigned unambiguously from the
HSQC spectrum. Comparison of the NMR data for 1 with des-
glucouzarin20 indicated that the 1H and 13CNMR signals for the
protons and carbons in the A, B, C, D and E rings had similar
Results and discussion
Compound 1 was isolated as colorless crystals. HRESIMS ana-
lysis of 1 showed a quasi-molecular ion peak at m/z 469.2563
[M + Na]+, corresponding to a molecular formula C26H38O6.
The IR spectrum of 1 exhibited prominent absorption bands at
8920 | Org. Biomol. Chem., 2014, 12, 8919–8929
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8919–8929 | 8921

View Article Online
Paper
Organic & Biomolecular Chemistry
chemical shifts. The major differences in the 13C NMR spectral
data between 1 and desglucouzarin20 were the presence of a
lactate unit (one carbonyl, one oxygenated methine and one
methyl) in 1 instead of a sugar unit in desglucouzarin.21
The full NMR data assignments of 1 were determined by
1
the analyses of H–1H COSY, HSQC and HMBC spectroscopic
data (Fig. 1). The 1H–1H COSY spectrum showed three spin
coupling systems (in bold face in Fig. 1): (i) H2-1 → H2-2 → H-3
→ H2-4 → H-5 → H2-6 → H2-7 → H-8 → H-9 → H2-11 → H2-12,
(ii) H2-15 → H2-16 → H-17, and (iii) H3-3′ → H-2′, which were
consistent with the cardenolide lactate skeleton. The HSQC
Fig. 2 X-ray structure of 1 showing the atom labeling scheme. The
spectrum revealed that the proton at δH 5.89 (H-22) was attached C and O atoms are drawn as 30% thermal ellipsoids.
to the carbon at δC 117.8 (C-22), and the HMBC spectrum
showed that H-22 was correlated to C-20, C-21, and C-23,
assignment of the absolute configuration (Fig. 2). Accordingly,
the structure of compound was established as
suggesting that the furan-2(5H)-one unit could be formed by
ring closure involving an oxygen atom bridged to C-21 and
1
C-23. Specially, HMBC correlations from H3-3′ (δH 1.35) to C-2′ (3S,5S,8R,9S,10S,13R,14S,17R,2′S)-3-O-urarigeninlactate.
Compound 2 was obtained as colorless crystals. HRESIMS
analysis of 1 showed a quasi-molecular ion peak at m/z
(δC 67.9) and C-1′ (δC 176.0), confirmed the presence of a lactate
unit in compound 1, and the correlation between H-3 [δH 4.73
(1H, m)] and C-1′ established location of the lactate unit at C-3. 447.2745 [M + H]+, corresponding to a molecular formula
C26H38O6, which was the same as compound 1. Its IR spectrum
Taken together, compound 1 was determined as a new cardeno-
lide with a lactate moiety at C-3 as shown in Fig. 1.
exhibited absorption bands for hydroxyl (3412 cm−1), carbonyls
The relative configuration of 1 was determined by analysis (1733 cm−1 and 1717 cm−1) functionalities. Proton signals at δH
4.66 (1H, q, J = 6.8 Hz, H-2′) and 1.65 (d, J = 6.8 Hz, H-3′) and
carbon signals at δC 175.5, 67.7 and 21.3 were assignable to a
of NOESY data, which showed correlations H3-19 ↔ H-8, H-8
↔ H3-18, indicating that H3-19, H-8 and H3-18 were the same
β-oriented. Similarly, the NOESY correlations H-3 ↔ H-5, H-5 lactate unit. Both the 1H and 13C-NMR of 2 were similar to those
of 1; however, the retention times were different (26.8 min for 1
and 27.8 min for 2, Fig. 3), indicating that 2 was an isomer of 1.
Crystals of 2 suitable for X-ray diffraction analysis were
obtained from methanol solution. The absolute configuration
of compound 2 could be unambiguously established by single
↔ H-9, but absence of correlation H-3 ↔ H3-18 suggested that
H-3, H-5, H-9 were mutually oriented on the other side.
In spite of the above NOESY correlation data, we were unable
to assign the configuration of C-2′ in the lactate unit. Fortunately,
crystals suitable for X-ray diffraction analysis were obtained from
methanol solution of 1. Final refinement of the CuKα diffraction crystal X-ray analysis based on the low Flack parameter −0.04
(17) (Fig. 4). Accordingly, the structure of compound 2 was
established as (3S,5S,8R,9S,10S,13R,14S,17R,20Z,2′R)-3-O-urari-
geninlactate, which is an isomer of 1 at the chiral center of the
lactate unit.
data resulted in a small Flack parameter 0.1(2), allowing the
The molecular formula of compound 3 was determined to
be C26H39O7 from its HRESIMS (m/z 463.2677, [M + H]+). The
IR spectrum indicated hydroxyl and carbonyl functionalities at
1
3469, 1733 and 1714 cm−1, respectively. The H- and 13C-NMR
(Table 1) spectra were similar to those of 2, indicating that 3
was also a cardenolide bearing a lactate unit with proton
signals at δH 4.42 (1H, m, H-2′) and 1.37 (d, J = 6.9 Hz, H-3′),
and carbon signals at δC 176.3, 67.9 and 20.6; however, there
was only one methyl group [δH 0.86 (3H, s, H-18); δC 16.4] in 3,
and the other one was replaced by an oxygenated methylene
[δH 4.27 (H, d, J = 12.3 Hz, H-19α) and δH 4.42 (H, d, J = 12.3
Hz, H-19β); δC 63.5]. The location of the oxygenated methylene
at C-19 was revealed by HMBC correlations from Hα-19 to C-1
(δC 32.1), C-5 (δC 45.7) and C-9 (δC 51.2) (see ESI†). Further-
more, HMBC correlations from C-1″ (δC 176.3) to both H-19α
and H-19β indicated the location of the lactate unit at C-19.
Finally the absolute configuration of 3 was confirmed by
single crystal X-ray analysis (Fig. 5) with a low Flack parameter
0.17(17). Accordingly, compound
(3S,5S,8R,9S,10R,13R,14S,17R,20Z,2″S)-19-O-coroglaucigeninlactate.
3 was determined as
Fig. 1 Key 1H–1H COSY, HMBC and NOESY correlations of 1.
8922 | Org. Biomol. Chem., 2014, 12, 8919–8929
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 3 Comparison of compounds 1 and 2 by HPLC. HPLC conditions: Agilent 1200 system, reverse phase C-18 column (5 μm, 20 × 250 mm; Cos-
mosil, Japan), mobile phase: CH3CN–H2O (7 : 3), and detection at 220 nm.
(d, J = 6.9 Hz, H-3″), δC 176.3, 67.9 and 20.5 for another unit.
HMBC correlations between H-19 [δH 4.34 (1H, m, H-19α), δH
4.43 (1H, m, H-19β)] and C-1″ (δC 176.3) suggested one lactate
unit at C-19. Similarly, HMBC correlation between H-3 [δH 4.79
(1H, m)] and C-1′ (δC 175.9) (see ESI†) suggested another lactate
unit at C-3. In order to confirm the configuration of the lactate
units at C-3 and C-19, compound 4 was hydrolyzed in a solution
of NaOH in MeOH. After removing the steroid moiety, the
Fig. 4 X-ray structure of 2 showing the atom labeling scheme. The
C and O atoms are drawn as 30% thermal ellipsoids.
specific optical rotation value of the aqueous solutions was
−4.5°, which was consistent with the specific optical rotation
value of R-lactic acid (−3.8°), suggesting both lactate units were
the same. Accordingly, the structure of 4 was identified as
(3S,5S,8R,9S,10S,13R,14S,17R,20Z,2′R,2″R)-coroglaucigenin-3,19-
O-dilactate.
The molecular formula of
5 was established to be
C32H48O11 by analysis of its HR-ESIMS which exhibited a
quasi-molecular ion at m/z 608.2531 [M + H]+. Comparison of
the NMR data of 5 (Table 1) with those of 3 showed that their
signals for the protons and carbons in the A, B, C, D and E
rings were similar, and compound 5 also has a lactate unit
[δH 4.42 (1H, m, H-2″), 1.36 (d, J = 6.9 Hz, H-3″), δC 176.4, 67.8,
20.5] at C-19, which was confirmed by the HMBC correlation
between H-19 and C-1″. In contrast, the signals for H-3
(δH 3.71) and C-3 (δC 78.7) were shifted to downfield as com-
pared with those of 3, suggesting that there might be a sugar
unit at C-3. The characteristic signals δC 99.8, 72.4, 72.8, 74.3,
70.5, 18.2 suggested the presence of a rhamnose unit, which
was also observed in frugoside (8),10 and the coupling constant
of the anomeric proton (δH 4.71, d, J = 8.0 Hz, H-1′′′) suggested
the β-configuration of the glycosidic bond.
The HMBC correlations between H-1′′′ and C-3 further con-
firmed the location of the sugar unit at C-3 (see ESI†). In order
to confirm the configuration of the lactate unit at C-19, similar
to 4, compound 5 was hydrolyzed in a solution of NaOH in
MeOH, and the specific optical rotation value of the aqueous
solutions was −3.9°, which was consistent with that of R-lactic
acid (−3.8°). Accordingly, the structure of 5 was determined as
(3R,5S,8R,9S,10R,13R,14S,17R,20Z,2″R)-19-O-frugosidelactate.
The molecular formula of compound 17 was deduced to be
C28H40O9 by the analysis of HR-ESIMS (m/z 521.2581 [M + H]+).
Fig. 5 X-ray structure of 3 showing the atom labeling scheme. The C
and O atoms are drawn as 30% thermal ellipsoids.
Compound 4, colorless powder, had a molecular formula
C29H42O9 as determined by HRESIMS (m/z 557.2448 [M + H]+).
Absorption bands at 3446, 1733 and 1711 cm−1 in the IR spec-
trum demonstrated the presence of hydroxyl and carbonyl
groups, respectively. The 13C NMR and DEPT spectrum con-
tained 29 signals assigned to three methyls, eleven methyl-
enes, eight methines, and seven quaternary carbons including
four carbonyl carbons (δC 175.9, 176.3, 177.2 and 178.3). Com-
parison of the NMR data of 4 with 2 and 3 revealed that 4 was
a cardenolide bearing two lactate units with signals at δH 4.19
(1H, q, J = 6.9 Hz, H-2′) and 1.37 (d, J = 6.9 Hz, H-3′), δC 175.5,
67.9 and 20.6 for one unit, and δH 4.25 (1H, m, H-2″) and 1.37
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8919–8929 | 8923

View Article Online
Paper
Organic & Biomolecular Chemistry
Absorption bands at 3420 and 1725 cm−1 in the IR spectrum
of 17 demonstrated the presence of hydroxyl and carbonyl
groups, respectively. The 1H NMR spectrum of the aglycone
portion of 17 showed characteristic signals of a butenolactone
ring at δH 5.89 (1H, d, J = 1.5 Hz), 4.90 (1H, dd, J = 17.6,
1.5 Hz), and 5.03 (1H, dd, J = 17.6, 1.5 Hz), as well as a methyl
signal at δ 1.22 (3H, s), indicating a cardenolide skeleton with
a loss of methyl at C-19. The 13C NMR spectrum of 17 showed
28 carbon signals, of which 22 could be assigned to the agly-
cone moiety and six assigned to a sugar unit. Analysis of the
13C and DEPT NMR spectra confirmed clearly the occurrence
of two oxygenated quaternary carbon atoms (δC 75.3 and 86.0)
along with signals typical of a butenolactone ring: a carbonyl
group (δ 177.2), an olefinic quaternary carbon (δC 178.4), an
Fig. 7 X-ray structure of 17 showing the atom labeling scheme. The
C and O atoms are drawn as 30% thermal ellipsoids.
olefinic methine (δC 117.8), and an oxygenate methylene
1
(δC 74.0) in the aglycone moiety. The H and 13C NMR chemi- H-1′ and H-5′ were β-oriented. However, the configurations of
cal shifts of the aglycone moiety of 17 were similar to those of the tertiary hydroxyl group at C-10, C-14 and C-2′ could not be
the known compound calotropin (18)16 except that the alde- established by NOESY spectrum. Fortunately, single crystals
hyde group at C-10 of 18 was replaced with a hydroxyl group in were obtained from methanol solution. X-ray analysis with
17, which was further confirmed by HMBC correlations CuKα diffraction data resulted in a small Flack parameter 0.0
between C-10 (δC 75.3) and H-2 [δH 3.92 (1H, m)], H-5[δH 1.69 (2), which confirmed the configurations deduced from NMR
(1H, m)] and H-8 [δH 1.50 (1H, m)] (Fig. 6). Additionally, the 1H analysis and revealed the configurations for the tertiary
NMR spectrum of 17 displayed a signal corresponding to an hydroxyl groups (Fig. 7). Accordingly, the structure of 17 was
anomeric proton at δH 4.45 (1H, s, H-1′),16 which suggested identified as (2R,3R,5S,8R,9S,10R,13R,14S,17R,20Z,1′S,2′S,3′S,5′R)-
that 17 is a doubly linked glycoside. The relative configuration 19β-hydroxycalotropin.
of C-3′ was established by analysis of 1H NMR coupling con-
stants (δH 3.61, dd, J = 12.1, 4.7 Hz, H-3′), which indicated an the physical and spectroscopic data with the literature as shown
axial position and thus an S-configuration at C-3′.
in the Introduction section. It is noteworthy that 1–5 are five new
Other known compounds were identified by comparison of
In the NOESY spectrum, correlations Hα-1 ↔ H-5, H-5 ↔ cardenolide lactates. Among them, compounds 1 and 2 have a
H-9 and H-3 ↔ H-5 suggested these Hα-1, H-3, H-5 and H-9 lactate unit at C-3, the lactate unit of 3 is linked at C-19, com-
was all α-oriented (Fig. 6). NOESY correlation H-5′ ↔ H-1′ and pound 4 has two lactate units at both C-3 and C-19, while com-
absence of correlations between H-1′ and H-3′ indicated that pound 5 is a cardenolide glycoside with a lactate unit at C-19.
Lactic acid is widely present in the human body, animals, plants
and microorganisms.22 It is commonly used as an additive in
food, drug salt, e.g. milrinone lactate in pharmaceutical indus-
tries,23 and serves as a raw material for the production of biode-
gradable polymers.24 However, this is the first report that natural
bioconjugates of cardenolide and lactic acid exist in plants.
Cardenolides are endogenously produced in about
60 genera of 12 families of the angiosperms, for examples
Asclepiadaceae and Apocynaceae families.25 Especially, carde-
nolides from the Asclepiadaceae family often form a dioxane
ring between the cardenolide aglycone and the sugar unit (e.g.
compounds 17–21 in this study), particularly in genera of
Asclepias and Calotropis, two close genera in the mega-tree cla-
dogram of Apocynaceae.25 This structure feature is rarely
found in the cardenolides from other families. Simultaneous
isolation of dioxane double linked cardenolide glycosides
(17–21) and cardenolide lactates (1–5) provided unique chemo-
taxonomic markers for this genus.
MTT colorimetric assay was performed to test the inhibitory
activities of 1–21 against DU145 prostate cancer cells. As
shown in Table 2, the dioxane double linked cardenolide
glycosides 17–21 showed the most potent cytotoxic effect
against DU145 cells (IC50 values in the range 0.03–0.29 μM),
the normal cardenolides and related glycosides 6–11 also
Fig. 6 Key 1H–1H COSY, HMBC and NOESY correlations of 17.
8924 | Org. Biomol. Chem., 2014, 12, 8919–8929
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Table 2 Inhibitory activities against DU145 cancer cells of compounds
1–21
Compound
IC50a (μM)
Compound
IC50a (μM)
1
2
3
4
5
6
7
8
14.38 2.45
1.66 0.13
5.31 0.41
16.96 0.89
16.80 1.43
0.48 0.08
0.92 0.09
0.34 0.06
0.33 0.05
0.39 0.04
0.33 0.05
12
13
14
15
16
17
18
19
>25.00
>25.00
>25.00
>25.00
>25.00
0.29 0.03
0.03 0.01
0.28 0.02
0.04 0.01
0.21 0.01
0.12 0.02
9
10
11
20
21
Taxolb
a The experiments were performed three times, each in triplicate.
b Positive control.
demonstrated strong cytotoxic effects with IC50 values in the
range 0.33–0.92 μM, and the cardenolide lactates 1–5 exhibited
moderate cytotoxic effects with IC50 values in the range
1.66–16.96 μM, while the C21 steroids 12–16 were inactive with
IC50 values larger than 25 μM.
The cytotoxic effects of cardenolides were reported to be
related to the modifications in the Na+,K+-ATPase signalo-
some.2 Thus, representative compounds 2 and 3 from the car-
denolide lactate group, 8 and 10 from the normal cardenolide
group and 18–21 from the dioxane double linked cardenolide
glycoside group were tested for their inhibitory effects on Na+,
K+-ATPase. As shown in Table 3, the inhibitions of Na+,K+-
ATPase were generally consistent with the cytotoxic effects, and
thus establishes a correlation between the cytotoxic effect in
human cancer cells and the Ki for the inhibition of Na+,K+-
ATPase. Particularly, compounds 8 and 18–21 showed very
potent inhibitory activities with Ki values less than 0.05 μM.
The dose-response curves showing inhibition of Na+,K+-ATPase
by representative compounds were shown in Fig. S46 (ESI†).
The interactions between normal cardenolides, e.g. ouabain,
and Na+,K+-ATPase have been studied by solid state NMR,26 X-
ray crystallography27 and molecular docking analysis,28 while
interactions between the cardenolide lactates, dioxane double
linked cardenolide glycosides and Na+,K+-ATPase have not been
reported. In this study, we explored the binding modes of com-
pounds 2, 18 and 20 with the molecular docking approach.
Similar to the complex structure of ouabain with Na+,K+-
ATPase,29 the head five-membered lactone rings of compounds
2, 18 and 20 penetrated deeply into the binding cavity (Fig. 8).
Fig. 8 Detailed molecular interactions between the binding pocket of
Na+,K+-ATPase and compounds 2, 18 and 20. The binding mode of
ouabain with Na+,K+-ATPase was shown in upper panel to depict the
localization of the ouabain site. (Left panels) The amino acids of Na+,K+-
ATPase close to the ligand compounds (ball-and-stick structure) are
shown in stick structure. Mg2+ close to the ligand is shown in CPK
(green ball). (Right panels) Interactions between Na+,K+-ATPase and
ligand compounds are depicted in simple drawings. Amino acid residues
of Na+,K+-ATPase involved in formation of hydrogen bonds are shown in
blue squares. Distances of hydrogen bonds (green lines) between
ligands and Na+,K+-ATPase (from donor hydrogen to receptor) are indi-
cated. The distance between Mg2+ and ligands (ouabain) is rather long
(about 6.4 Å), and no direct interaction was observed between Mg2+ and
ligands (ouabain).
Table 3 Na+,K+-ATPase inhibitory activities of the representative
compounds
Compounds
Ki (µM)
Compounds
Ki (µM)
2
3
8
0.163
10.750
<0.05
0.960
<0.05
18
19
20
21
<0.05
<0.05
<0.05
<0.05
10
Ouabaina
a Positive control.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8919–8929 | 8925

View Article Online
Paper
Organic & Biomolecular Chemistry
H-bonds were formed between the binding cavity of Na+,K+- taxonomic markers for this genus. The dioxane double linked
ATPase and compounds 2, 18 and 20. In the docking simu- cardenolide glycosides showed the most potent cytotoxic effect
lation of compound 2, two H-bonds were found; one was against DU145 cancer cells followed by normal cardenolides
formed between the hydroxyl group at C-14 and G319 (distance and cardenolide lactates, while the C21 steroids were non-cyto-
2.04 Å, all distances were shown from donor hydrogen to toxic. Enzymatic assay established a correlation between the
receptor), and the other was formed between the lactate cytotoxic effect and the inhibitory effect on Na+,K+-ATPase.
moiety at C-3 and E117 (distance 1.98 Å). In the docking simu- Molecular docking analysis revealed strong H-bond inter-
lation of compound 18, three H-bonds were found; one was actions in the binding cavity of dioxane double linked cardeno-
formed between the hydroxyl group at C-14 and T797 (distance lide glycosides.
1.69 Å), one was formed between the aldehyde group at C-10
and Q111 (distance 1.89 Å), and the other one was formed
between the hydroxyl group at C-2′ and Q111 (distance 2.15 Å).
Experimental section
General
In accord with the comparable Na+,K+-ATPase inhibitory
potency of compounds 18 and 20 (Table 3), similar intermole-
cular interaction was observed when compound 18 was
replaced with compound 20 in the docking simulation though
more H-bonds were observed. Six H-bonds were found: three
temperature. UV spectra were measured with a Jasco V-550 UV/
were formed between the hydroxyl group at C-14 and T797/
D121 (distances 1.75, 1.88, and 2.13 Å, respectively), one was
Jasco FTIR-480 Plus spectrometer. NMR spectra were obtained
formed between the aldehyde group at C-10 and Q111 (dis-
tance 1.87 Å), and the other two were formed between E116
and the hydroxyl groups at C-2′ and C-3′ (distances 1.87 and
1.93 Å). However, the three H-bonds formed with the hydroxyl
group at C-14 of compound 20 were relatively weak in compari-
son with the strong H-bond (short distance of 1.69 Å) formed
with the hydroxyl group at C-14 of compound 18. To compare
Melting points were measured on an X-5 melting point appar-
atus without correction. Optical rotations were determined
using MeOH solutions on a Jasco P-1020 polarimeter at room
VIS spectrophotometer, while IR spectra were collected from a
on a Bruker AV-400 spectrometer. ESIMS and HRESIMS spectra
were obtained on a Finnigan LCQ Advantage Max ion trap
mass spectrometer and an Agilent 6210 ESI/TOF mass spectro-
meter, respectively. Silica gel for column chromatography
(200–300 mesh) was produced by Qingdao Marine Chemical
Industrials, and Sephadex LH-20 was purchased from Pharma-
cia Biotech (Pharmacia, Kalamazoo, MI, USA). Precoated silica
gel GF254 plates (Qingdao Marine Chemical Plant, Qingdao,
relative binding affinities of compounds 2, 18 and 20 in the
binding cavity of Na+,K+-ATPase, LigScore 1 and 2 were calcu-
P. R. China) were used for TLC analysis. Preparative HPLC was
lated. The results showed that LigScore 1 were 5.22, 5.75 and
performed on a Varian Prostar system equipped with a pre-
parative Cosmosil C18 column (5 μm, 20 × 250 mm) column.
6.13 while LigScore 2 were 6.13, 7.38 and 6.61 for compounds
2, 18 and 20, respectively. On the basis of the LigScore calcu-
lations (average 5.675, 6.515 and 6.370 for 2, 18 and 20), the
binding affinity of compound 2 was apparently lower than those
of compounds 18 or 20, and compounds 18 and 20 were likely
to possess similar binding affinity. The theoretical calculation
might partly explain why compound 18 and compound 20 pos-
sessed comparable inhibitory potency on Na+,K+-ATPase.
Taken together, the relatively strong H-bond interaction and
high binding affinity in the binding cavity seemed to explain
why compounds 18 and 20 possessed higher inhibitory
potency on Na+,K+-ATPase than compound 2. Detailed inter-
Plant material
The whole plants of Asclepias curassavica were collected in
Zhongshan city of Guangdong province, P. R. China in August
2012, and authenticated by Prof. Guang-Xiong Zhou (Jinan
University). A specimen (no. 2012081001) was deposited in the
Institute of Traditional Chinese Medicine and Natural Pro-
ducts, College of Pharmacy, Jinan University, P. R. China.
Extraction and isolation
molecular H-bond geometries in the docking simulation of The dried and powdered whole plants of Asclepias curassavica
compounds 2, 18, and 20 were shown in Table S1 (ESI†).
(20 kg) were extracted with 70% (V/V) ethanol at 90 °C. The
solution was concentrated under reduced pressure to afford a
crude extract (3 kg), which was subsequently partitioned
between EtOAc and H2O. The EtOAc extract (200 g) was sub-
jected to silica gel column chromatography eluted with gradi-
Conclusions
In summary, this paper describes the isolation and structure ent mixtures of CHCl3–MeOH (100 : 0→1 : 4) to yield 18
elucidation of five new cardenolide lactates and one dioxane fractions (Fr. 1 to Fr. 18).
double linked cardenolide glycoside along with 15 known com-
Fr. 9 (18 g) was subjected to silica gel column chromato-
pounds from the ornamental milkweed Asclepias curassavica. graphy, eluted with a gradient CHCl3–CH3OH (from 30 : 1 to
The structures and absolute stereochemistry were elucidated 1 : 1) system to give seven subfractions (Fr. 9A to Fr. 9G). Fr. 9E
by NMR spectroscopic data, and were confirmed by ORD spec- was subjected to preparative HPLC eluted by CH3CN–H2O
tral data as well as X-ray crystallographic analysis. Simul- (7 : 3) to afford compounds 1 (34 mg), 2 (18 mg) and 16
taneous isolation of dioxane double linked cardenolide (39 mg). Fr. 12 (18 g) was subjected to silica gel column chrom-
glycosides and cardenolide lactates provided unique chemo- atography eluted with gradient mixtures of CH3Cl–MeOH
8926 | Org. Biomol. Chem., 2014, 12, 8919–8929
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
(20 : 0→1 : 4) to yield ten subfractions (Fr. 12A to Fr. 12J). Sub- The hydrolysis of compounds 4 and 5
fraction Fr. 12E (2.8 g) was separated by Sephadex LH-20 and
Compounds 4 (3.012 mg) and 5 (3.734 mg) were submitted to
hydrolysis by adding to a solution of NaOH in MeOH (0.5 N,
4 mL). The mixture was stirred for 10 hours at room tempera-
ture (no 4 or 5 was detected by TLC analysis). After the reac-
tion, the mixture was acidified with dilute HCl (4%) to pH 5–6
and saturated with NaCl, followed by extraction with CHCl3.
Then the aqueous layer was concentrated in vacuo and the
residue was dissolved in 0.5 mL water. The optical rotation of
an aliquot of the aqueous solution (200 μL) containing the
lactic acid (c = 0.195 for 4 and 0.111 for 5) was determined
using a Jasco P-1020 spectrometer (l = 0.1 dm).
preparative HPLC eluted by CH3CN–H2O (70 : 30) to yield com-
pounds 6 (24 mg), 9 (23 mg) and 12(18 mg). Subfraction Fr.
12H was purified by preparative HPLC eluted by CH3CN–H2O
(70 : 30) to afford 3 (59 mg) 13 (24 mg) and 15 (23 mg). Subfrac-
tion Fr.12G (320 mg) was re-separated by Sephadex LH-20
column (CH3OH) and preparative HPLC [CH3CN–H2O (70 : 30)]
to yield compounds 18 (27 mg) and 19 (21 mg). Subfraction
Fr.12H was purified by preparative HPLC [CH3CN–H2O
(70 : 30)] to give 4 (26 mg), 20 (27 mg) and 21 (20 mg). Fr. 15
(36 g) was subjected to ODS column chromatography
eluted with gradient mixtures of MeOH–H2O (1 : 4→4 : 1) to
yield 13 subfractions (Fr. 15A to Fr. 15M). Subfraction Fr.15F
was subjected to preparative HPLC eluted by CH3CN–H2O
X-ray analysis
(70 : 30) to afford compounds 7 (34 mg), 8 (18 mg) and 10 X-ray diffraction data were collected on an Agilent Gemini S
(13 mg). Subfraction Fr.15H was purified by preparative HPLC ultra spharrie CCD diffractometer using graphite monochro-
eluted by CH3CN–H2O (70 : 30) to give 5 (21 mg), 11 (24 mg) mated radiation (λ = 1.54178 Å) under room temperature. The
and 14 (23 mg). Fr. 18 (28 g) was subjected to silica gel column crystal structures were elucidated by direct methods using
chromatography, eluted with a CHCl3–CH3OH gradient system SHELXS-97 and refined by full-matrix least-squares method on
(30 : 1–1 : 1) to give 10 fractions (Fr. 15A to Fr. 15J). Subfraction F2 using SHELXS-97. In the structure refinement, non-hydro-
Fr.15E was subjected to preparative HPLC [CH3CN–H2O (7 : 3)] gen atoms were refined anisotropically. Hydrogen atoms
to give compound 17 (30 mg).
bonded to carbons were placed at their geometrically ideal
(3R,5S,8R,9S,10S,13R,14S,17R,20Z,2′S)-3-O-Urarigenin lactate positions. Hydrogen atoms bonded to oxygen were located by
(1). Colorless crystals from MeOH; [α]2D5 −49 (0.1, CHCl3); employing the difference Fourier method and were included in
UV (MeOH) λmax (log ε) 217(4.05) nm; IR (KBr) νmax 3430, the calculation of structure factors with isotropic temperature
2939, 1733, 1716 cm−1 1H and 13C NMR data, see Table 1; factors.
;
HRESIMS m/z 469.2563 [M + Na]+ (calcd for C26H37O6Na,
469.2561).
(3R,5S,8R,9S,10S,13R,14S,17R,20Z,2′S)-3-O-Urarigenin lactate
(1). Colorless columnar crystals from MeOH, C26H38O6, ortho-
(3R,5S,8R,9S,10S,13R,14S,17R,20Z,2′R)-3-O-Urarigenin lactate rhombic, P212121, a = 11.4718(5), b = 12.9514(5), c = 15.8446(8)
(2). Colorless crystals; [α]2D5 −67 (0.1, MeOH); UV (MeOH) λmax Å, β = 90.0, V = 2354.13(18) Å3, Z = 4, dx = 1.260 Mg m−3
,
(log ε) 216 (4.24) nm; IR (KBr) νmax 3412, 2939, 1733, μ(CuKα) = 0.711 mm−1, F(000) = 968. 3588 unique reflections
1
1717 cm−1; H and 13C NMR data, see Table 1; HRESIMS m/z were collected to θmax = 62.76°, in which 3276 reflections were
447.2745 [M + H]+ (calcd for C26H38O6, 447.2741).
observed [F2 > 4σ(F2)]. The final R = 0.0353, Rw = 0.0873, S =
(3S,5S,8R,9S,10R,13R,14S,17R,20Z,2″S)-19-O-Coroglaucigenin 1.062 and CCDC 1014923.
lactate (3). Colorless crystals; [α]2D5 −67 (0.1, MeOH); UV
(3R,5S,8R,9S,10S,13R,14S,17R,20Z,2′R)-3-O-Urarigenin lactate
(MeOH) λmax (log ε) 218 (4.16) nm; IR (KBr) νmax 3469, 2930, (2). Colorless columnar from MeOH, C26H38O6, orthorhombic,
1733, 1714 cm−1; 1H and 13C NMR data, see Table 1; HRESIMS P212121, a = 10.4301(3), b = 11.8072(3), c = 18.7440(5) Å, β =
m/z 463.2677 [M + Na]+ (calcd for C26H39O7, 463.2690).
(3S,5S,8R,9S,10R,13R,14S,17R,20Z,2′R,2″S)-3,19-O-Coroglau-
90.0, V = 2308.33(11) Å3, Z = 4, dx = 1.325 Mg m−3, μ(CuKα) =
0.755 mm−1, F(000) = 996. 3400 unique reflections were col-
cigenin dilactate (4). Colorless powder; [α]2D5 −46 (0.1, MeOH); lected to θmax = 62.74°, in which 3514 reflections were observed
UV (MeOH) λmax (log ε) 218 (4.08) nm; IR (KBr) νmax 3446, [F2 > 4σ(F2)]. The final R = 0.0315, Rw = 0.0810, S = 1.065 and
2932, 1733, 1711 cm−1 1H and 13C NMR data, see Table 1; CCDC 1014924.
;
HRESIMS m/z 557.2448 [M + Na]+ (calcd for C29H41O9Na,
(3S,5S,8R,9S,10R,13R,14S,17R,20Z,2″S)-19-O-Coroglaucigenin
lactate (3). Colorless columnar from MeOH, C26H38O7, mono-
534.2451).
(3R,5S,8R,9S,10R,13R,14S,17R,20Z,2″R)-19-O-Frugoside lactate clinic, P21, a = 6.3498(3), b = 19.5019(8), c = 9.5651(5) Å, β =
(5). Colorless powder; [α]2D5 −54 (0.1, MeOH); UV (MeOH) λmax 102.297(4), V = 1157.30(9) Å3, Z = 2, dx = 1.327 Mg m−3
,
1
(log ε) 217 (4.17); IR (KBr) νmax 3446, 2934, 1733 cm−1; H and μ(CuKα) = 0.776 mm−1, F(000) = 500. 2600 unique reflections
13C NMR data, see Table 1; HRESIMS m/z 608.2531 [M + H]+ were collected to θmax = 62.66°, in which 2721 reflections were
(calcd for C32H48O11).
observed [F2 > 4σ(F2)]. The final R = 0.0310, Rw = 0.0783, S =
(2R,3R,5S,8R,9S,10R,13R,14S,17R,20Z,1′S,2′S,3′S,5′R)-19β-Oxhy- 1.057 and CCDC 1014925.
drylcalotropin (17). Colorless crystals from MeOH solution;
(2R,3R,5S,8R,9S,10R,13R,14S,17R,20Z,1′S,2′S,3′S,5′R)-19β-Hydro-
[α]2D5 −49 (0.1, MeOH); UV (MeOH) λmax (log ε) 205(2.59) nm; IR xycalotropin
(17). Colorless
crystals
from
MeOH,
(KBr) νmax 3420, 1725 cm−1; 1H and 13C NMR data, see Table 1; C28H40O9·4H2O, monoclinic, P21, a = 8.9982(4), b = 12.3600(7),
HRESIMS m/z 521.2581 [M
521.2589).
+
H]+ (calcd for C28H40O9, c = 13.3864(8) Å, β = 98.572(5), V = 1472.17(14) Å3, Z = 2, dx =
1.319 Mg m−3, μ(CuKα) = 0.882 mm−1, F(000) = 624. 3354
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8919–8929 | 8927

View Article Online
Paper
Organic & Biomolecular Chemistry
unique reflections were collected to θmax = 62.65°, in which Discover Studio 2.1 package (http://accelrys.com/products/dis-
2798 reflections were observed [F2 > 4σ(F2)]. The final R = covery-studio/). The 2D structures of steroid-like compounds
0.0459, wR2 = 0.1181, S = 1.107 and CCDC 1014926.
used in this study were constructed by using the ChemDraw
program, and their corresponding 3D structures were con-
verted by the Chem3D program (http://www.cambridgesoft.
Cytotoxicity assay
The MTT assay was done as described previously30 with taxol com/). The binding site for the steroid-like compounds in the
served as the positive control. Briefly, the prostate cancer cells Na+,K+-ATPase subunit was defined as ouabain binding site
were plated into 96-well plates at a density of 3 × 103 cells per among the extracellular loops linking transmembrane seg-
well for androgen independent cells DU145. Cells were main- ments in the ouabain-Na+,K+-ATPase complex structure. In the
tained in RPMI-1640 medium supplemented with 10% fetal docking simulation of compounds, the ligand-binding domain
bovine serum (Gibco, Rockville, MD, USA), 100 units mL−1 was defined as the region of the sphere with a 12.5 Å radius
penicillin and streptomycin at 37 °C in a 5% CO2 humidified from the center of the binding pocket. The LibDock method-
atmosphere. Following incubation for 48 h, 20 μL of the MTT ology effectively executed the docking of combinatorial
solution [5 mg mL−1 in phosphate buffered saline (PBS)] was libraries of compounds in a high-throughput manner while
added to each well, and the cells were further incubated for keeping the protein structure fixed. In all complex structures
4 h. Then the medium was removed and replaced by 150 μL of generated by LibDock module, the binding orientation and
DMSO in each well to dissolve the formazan crystals. The rela- conformation of compounds with the similar ouabain binding
tive cell viability was determined by measuring the optical den- mode were selected. The energy of the ligand and residues in
sities at 570 nm on microplate reader (SPECTRAmax 250, the binding site of the selected docking complex structures
Molecular Devices, Minnesota, USA), and was expressed as a were further energy-minimized by smart minimize algorithm
percentage relative to the control. The experiments were per- with CHARMm force field in the Discover Studio 2.1
formed three times, each in triplicate.
package.34 The distances of intermolecular hydrogen bonds
(from proton to acceptor) were set as less than 2.5 Å. The
binding affinities of compound 2, 18 and 20 in the binding
Na+,K+-ATPase inhibition assay
Na+,K+-ATPase from pig kidney microsomal membranes was pocket of Na+,K+-ATPase were scored by LigScore 1 and Lig-
prepared by treatment with SDS and purified by differential Score 2.35
centrifugation.31 The specific activity of the enzyme prepa-
ration was approximately 30 µmol ATP hydrolysed per mg
protein per min at 37 °C.31,32 The Na+,K+-ATPase inhibitory
activities of 2, 3, 8, 10, and 18–21 were determined essentially
Acknowledgements
as previously reported.26 In brief, Na+,K+-ATPase is incubated This work was supported by the 111 Project (no. B13038) from
at 37 °C for 2 hours in the presence of 3 mM MgCl2, 3 mM Na- the Ministry of Education of P. R. China.
phosphate and 40 mM Tris (pH 7.0) with increasing concen-
trations of inhibitor. The Na+,K+-ATPase activity is sub-
sequently determined by 40-fold dilution into a standard assay
medium containing 130 mM NaCl, 20 mM KCl, 4 mM MgCl2
Notes and references
and 3 mM ATP (in triplicate). For each inhibitor concentration
the inactivation was determined in two or three independent
experiments and the error bars in Fig. S46† indicate the stan-
dard deviation. Data in Fig. 46S† are given as percent of the
Na+,K+-ATPase activity in the absence of inhibitor (see legend
to Fig. S46† for details).
The inhibitors were solubilised in DMSO as concentrated
solutions, typically as 10 mM solutions. The dilution into the
Na+,K+-ATPase incubation medium was at least 40-fold, giving
a maximal DMSO concentration of 2.5% in the incubation
medium. Control experiments showed that incubation for
2 hours at 37 °C as above with 2.5% DMSO (absence of inhibi-
tor) led to less than 10% inhibition of Na+,K+-ATPase activity.
1 L. Yatime, M. Laursen, J. P. Morth, M. Esmann, P. Nissen
and N. U. Fedosova, J. Struct. Biol., 2011, 174, 296.
2 R. A. Newman, P. Yang, A. D. Pawlus and K. I. Block, Mol.
Interv., 2008, 8, 36.
3 S. B. Malcolm, Chemoecology, 1990, 1, 12.
4 S. M. Kupchan, J. P. Knox, J. E. Kelsey and
J. A. Saenzrenauld, Science, 1964, 146, 1685.
5 F. Abe, Y. Mori and T. Yamauchi, Chem. Pharm. Bull., 1991,
39, 2709.
6 M. C. Roy, F. R. Chang, H. S. Huang, M. Y. N. Chiang and
Y. C. Wu, J. Nat. Prod., 2005, 68, 1494.
7 J. Z. Li, C. Qing, C. X. Chen, X. J. Hao and H. Y. Liu, Med.
Chem. Lett., 2009, 19, 1956.
8 H. T. A. Cheung and C. J. Nelson, J. Chem. Soc., Perkin
Trans. 1, 1989, 1563.
Molecular modeling and docking
The crystal structure of pig kidney Na+,K+-ATPase-ouabain
complex with Mg2+ (PDB code 4HYT) were downloaded from
9 Y. L. Liu and P. M. Loshkarev, Meditsinskaya Promyshlennost
SSSR, 1963, 17, 18.
Protein Data Bank.29 Ouabain in this complex structure was 10 F. Kiuchi, Y. Fukao, T. Maruyama, T. Obata, M. Tanaka,
removed first, and the modified Na+,K+-ATPase after hydrogen
saturation was applied with CHARMm force field33 using the
T. Sasaki, M. Mikage, M. E. Haque and Y. Tsuda, Chem.
Pharm. Bull., 1998, 46, 528.
8928 | Org. Biomol. Chem., 2014, 12, 8919–8929
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
11 T. Reichstein, A. M. Kuritzkes, Ch. Tamm and H. Jaeger, 24 G. A. Fleming, K. T. Murray, C. Yu, J. G. Byrne,
Helv. Chim. Acta, 1963, 46, 8.
12 D. H. G. Crout, R. F. Curtis, C. H. Hassall and T. L. Jones,
Tetrahedron Lett., 1963, 63.
J. P. Greelish, M. R. Petracek, S. J. Hoff, S. K. Ball,
N. J. Brown and M. Pretorius, J. Am. Heart Assoc., 2008, 118,
1619.
13 R. M. Padua, A. B. Oliveira, J. D. S. Filho, G. J. Vieira, 25 A. A. Agrawal, G. Petschenka, R. A. Bingham, M. G. Weber
J. A. Takahashi and F. C. Braga, J. Brazil. Chem. Soc., 2005,
16, 614.
14 X. Y. Li, H. X. Sun, Y. P. Ye, F. Y. Chen, J. Tu and Y. J. Pan,
Chem. Biodiversity, 2005, 2, 1701.
15 J. Z. Li, H. Y. Liu, Y. J. Lin;, X. J. Hao, W. Ni and
Ch. X. Chen, Steroids, 2008, 73, 594.
16 H. T. A. Cheung, C. J. Nelson and T. R. Watson, J. Chem.
Sci., 1988, 7, 1851.
17 J. Z. Li, C. Qing, C. X. Chen, X. J. Hao and H. Y. Liu, Bioorg.
Med. Chem. Lett., 2009, 19, 1956.
18 H. T. A. Cheung and T. R. Watson, J. Chem. Soc., 1980, 2162.
19 H. T. A. Cheung, F. C. K. Chiu, T. R. Watson and R. J. Wells,
J. Chem. Soc., Perkin Trans., 1983, 2827.
and S. Rasmann, New Phytol., 2012, 194, 28.
26 D. A. Middleton, S. Rankin, M. Esmann and A. Watts, Proc.
Natl. Acad. Sci. U. S. A., 2000, 97, 13602.
27 H. Ogawa, T. Shinoda, F. Cornelius and C. Toyoshima, Proc.
Natl. Acad. Sci. U. S. A., 2009, 106, 13742.
28 F. Cornelius, R. Kanai and C. Toyoshima, J. Biol. Chem.,
2013, 288, 6602.
29 M. Laursen, L. Yatime, P. Nissen and N. U. Fedosova, Proc.
Natl. Acad. Sci. U. S. A., 2013, 110, 10958.
30 T. Mosmann, J. Immunol. Methods, 1983, 65, 55.
31 I. Klodos, M. Esmann and R. L. Post, Kidney Int., 2002, 62,
2097.
32 M. Esmann, Methods Enzymol., 1988, 156, 105.
20 S. D. Jolad, R. B. Bates, J. R. Cole, J. J. Hoffmann, T. J. Siahaan 33 B. R. Brooks, R. E. Bruccoleri, B. D. Olafson, D. J. States,
and B. N. Timmermann, Phytochemistry, 1986, 25, 2581.
21 K. Koike, C. Bevelle, S. K. Talapatra, G. A. Cordell and
N. R. Farnsworth, Chem. Pharm. Bull., 1980, 28, 401.
S. Swaminathan and M. Karplus, J. Comput. Chem., 1983, 4,
187.
34 S. L. Dixon and K. M. Merz Jr., J. Med. Chem., 2001, 44,
22 B. C. Chao and J. L. Xu, Food Fermentation Ind., 1993, 3, 56.
3795.
23 I. M. Mujtaba, E. A. Edreder and M. Emtir, Appl. Energy, 35 S. L. Mayo, B. D. Olafson and W. A. Goddard, J. Phys.
2012, 89, 74.
Chem., 1990, 94, 8897–8909.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 8919–8929 | 8929