Antioxidant constituents from rhubarb: Structural requirements of stilbenes for the activity and structures of two new anthraquinone glucosides

Article abstract of DOI:10.1016/S0968-0896(00)00215-7

The methanolic extracts from five kinds of rhubarb were found to show scavenging activity for DPPH radical and ·O₂^-. Two new anthraquinone glucosides were isolated from the rhizome of Rheum undulatum L. together with two anthraquinone glucosides, a naphthalene glucoside, and 10 stilbenes. In the screening test for radical scavenging activity of rhubarb constituents, stilbenes and a naphthalene glucoside showed activity, but anthraquinones and sennosides did not. In addition, most stilbenes inhibited lipid peroxidation of erythrocyte membrane by tert-butyl hydroperoxide. Detailed examination of the scavenging effect on various related compounds suggested the following structural requirements; 1) phenolic hydroxyl groups are essential to show the activity; 2) galloyl moiety enhances the activity; 3) glucoside moiety reduces the activity; 4) dihydrostilbene derivatives maintain the scavenging activity for the DPPH radical, but they show weak activity for ·O₂^-. In addition, several stilbenes with both the 3-hydroxyl and 4′-methoxyl groups inhibited xanthine oxidase.

Full text of DOI:10.1016/S0968-0896(00)00215-7

Bioorganic & Medicinal Chemistry 9 (2001) 41±50
Antioxidant Constituents from Rhubarb:
Structural Requirements of Stilbenes for the Activity and
Structures of Two New Anthraquinone Glucosides
Hisashi Matsuda, Toshio Morikawa, Iwao Toguchida, Ji-Young Park,
Shoichi Harima and Masayuki Yoshikawa*
Kyoto Pharmaceutical University, Misasagi, Yamashina-ku, Kyoto 607±8412, Japan
Received 12 June 2000; accepted 29 July 2000
ꢀ
AbstractÐThe methanolic extracts from ®ve kinds of rhubarb were found to show scavenging activity for DPPH radical and O2 .
Two new anthraquinone glucosides were isolated from the rhizome of Rheum undulatum L. together with two anthraquinone glu-
cosides, a naphthalene glucoside, and 10 stilbenes. In the screening test for radical scavenging activity of rhubarb constituents,
stilbenes and a naphthalene glucoside showed activity, but anthraquinones and sennosides did not. In addition, most stilbenes
inhibited lipid peroxidation of erythrocyte membrane by tert-butyl hydroperoxide. Detailed examination of the scavenging e€ect on
various related compounds suggested the following structural requirements; 1) phenolic hydroxyl groups are essential to show the
activity; 2) galloyl moiety enhances the activity; 3) glucoside moiety reduces the activity; 4) dihydrostilbene derivatives maintain the
ꢀ
scavenging activity for the DPPH radical, but they show weak activity for O2 . In addition, several stilbenes with both the 3-
hydroxyl and 40-methoxyl groups inhibited xanthine oxidase. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
Active oxygen species and free radicals react with bio-
molecular constituents (e.g., lipids, protein, and DNA)
to cause certain clinical diseases, such as cerebral ische-
mia, atherosclerosis, in¯ammation, diabetes, and
cancer,4 8 which are regarded as damage associated
with Oketsu syndrome in Chinese traditional medicine.
We examined the scavenging e€ects of the methanolic
extract from the dried rhizome of R. palmatum, R. tan-
guticum, R. ocinale, R. coreanum, and R. undulatum on
1,1-diphenyl-2-picrylhydrazyl (DPPH) radical and
superoxide anion radical (ꢀ O2 ) generated by the xan-
thine±xanthine oxidase system and/or on lipid perox-
idation by tert-butyl hydroperoxide (t-BuOOH) in the
erythrocyte membrane ghost system. As a result, these
extracts showed DPPH radical and ꢀ O2 scavenging
activities and/or inhibition of lipid peroxidation. From
the rhizome of R. undulatum, of which neither chemical
nor pharmacological studies have been adequately
reported, we isolated 10 known stilbene constituents
from the active fraction together with a known naph-
thalene glucoside and two known and two new anthra-
quinone glucosides. This study dealt with the isolation
and characterization of active constituents from the
rhizome of R. undulatum, and structural requirements of
active constituents for antioxidant activity.
Rhubarbs, the rhizomes of Rheum palmatum L., R. tan-
guticum Maxim., R. ocinale Baill., R. coreanum Nakai,
and R. undulatum L., are used in remedies for blood
stagnation syndrome (called `Oketsu syndrome' in
Japanese traditional medicine) as well as a purgative
agent in Japanese, Korean, and Chinese traditional
medicines. Among them, the rhizome of R. undulatum, a
Korean rhubarb, is considered to have less purgative
e€ect but more potent e€ect on Oketsu syndrome than
other kinds of rhubarbs.1 Previously, the nitric oxide
(NO) production inhibitory activity in lipopolysaccharide-
activated mouse macrophages, anti-platelet aggregation in
rabbit platelets, and anti-allergic and anti-in¯ammatory
e€ects in mice and rats were reported as its anti-Oketsu
e€ect.1 3 Many studies about the pharmacological
properties and bioactive constituents of the former four
rhubarbs, which are listed in Japanese Pharmacopoeia
XIII, have been reported, but those of the latter rhu-
barb have not been studied suciently.
*Corresponding author. Tel.: +81-75-595-4633; fax: +81-75-595-4768;
e-mail: shoyaku@mb.kyoto-phu.ac.jp
0968-0896/01/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00215-7

42
H. Matsuda et al. / Bioorg. Med. Chem. 9 (2001) 41±50
Results and Discussion
fraction was subjected to silica gel column chromato-
graphy [CHCl3±MeOH (10:1, v/v)!(4:1)! CHCl3±
MeOH±H2O (10:3:1, lower layer)!MeOH] to give ®ve
fractions (Fr. 1±Fr. 5). Each fraction was subjected to
ODS column chromatography (MeOH±H2O) and
®nally HPLC [YMC-pack R&D ODS-5-A, 250Â20 mm
i.d., MeOH±H2O or CH3CN±H2O] to give two new
anthraquinone glucosides, chrysophanol 8-O-b-d-(60-
galloyl)-glucopyranoside (1, 0.092%) and aloe-emodin
1-O-b-d-glucopyranoside (2, 0.0065%) together with
two anthraquinone glucosides: chrysophanol 1-O-b-d-
Isolation of chemical constituents from the rhizome of
R. undulatum
The dried rhizome of R. undulatum (5.8 kg, cultivated in
Korea) was extracted with methanol at room temperature.
The methanolic extract (33.8% from the natural medi-
cines) was subjected to Diaion HP-20 column chroma-
tography (H2O!MeOH!acetone) to give the H2O-
eluted fraction (13.3%), MeOH-eluted fraction (18.7%),
and acetone-eluted fraction (1.8%). The MeOH-eluted
Chart 1.

H. Matsuda et al. / Bioorg. Med. Chem. 9 (2001) 41±50
43
glucopyranoside (5,9 0.25%) and chrysophanol 8-O-b-d-
glucopyranoside (6,9 0.16%), a naphthalene glucoside:
torachrysone 8-O-b-d-glucopyranoside (7,10 0.12%), and
10 stilbenes: rhaponticin (8,11 3.5%), piceatannol 30-O-b-
d-glucopyranoside (9,11 2.0%), desoxyrhaponticin (10,11
0.048%), isorhapontin (11,12 0.36%), rhapontigenin (13,11
0.58%), piceatannol (14,11 0.073%), desoxyrhap-onti-
genin (15,13 0.015%), resveratrol (16,14 0.048%), rha-
ponticin 200-O-gallate (17,11 0.12%), and rhaponticin 600-
O-gallate (18,11 0.087%).
ion peak (M+Na)+ and by high-resolution MS mea-
surement. In the UV spectrum of 1, absorption maxima
were observed at 220 (log e 4.58), 260 (4.32), 284 (sh,
4.17), and 409 (3.79) nm, suggestive of an anthraqui-
none structure. The IR spectrum of 1 showed absorption
bands ascribable to hydroxyl, free and chelated carbonyl,
and aromatic functions at 3410, 1709, 1641, 1637, 1466
and 1075 cm 1. The 1H NMR (DMSO-d6) and 13C
NMR (Table 1) spectra of 1 showed signals assignable
to a tertiary methyl [d 2.40 (s, 3-CH3)], an anomeric [d
5.24 (d, J=7.7 Hz, Glc-10-H)], a galloyl group [d 6.99
(2H, s, galloyl-200, 600-H)], aromatic protons [d 7.16 (1H,
br s, 4-H), 7.46 (1H, br s, 2-H), 7.67 (1H, d, J=8.6 Hz, 5-
H), 7.72 (1H, dd, J=7.6, 8.6 Hz, 6-H), 7.82 (1H, d,
J=7.6 Hz, 7-H)], and a chelated hydroxyl proton [d
12.82 (br s, 1-OH)]. The proton and carbon signals in
the NMR data of 1 were superimposable on those of
chrysophanol 8-O-b-d-glucopyranoside (6), except for
the signal due to a galloyl group. Comparison of the 1H
NMR and 13C NMR (Table 1) spectra for 1 with those
for 6 revealed an acylation shift around the 60-position
[d 4.27, 4.50 (1H each, both m, Glc-60-H2)]. Alkaline
treatment of 1 with 1.0% sodium methoxide (NaOMe)±
methanol furnished 6. Furthermore, in the hetero-
nuclear multiple bond connectivity (HMBC) experiment
of 1, long-range correlations were observed between the
following proton and carbon pairs: 10-H and 8-C; 60-H
and galloyl carbonyl carbon [dc 165.6 (700-C)], as shown
in Figure 1. On the basis of this evidence, 1 was deter-
mined to be 60-galloyl chrysophanol 8-O-b-d-glucopyr-
anoside.
Structures of chrysophanol 8-O-ꢀ-D-(60 -galloyl)-gluco-
pyranoside (1) and aloe-emodin 1-O-ꢀ-D-glucopyran-
oside (2)
Chrysophanol 8-O-b-d-(60-galloyl)-glucopyranoside (1)
was isolated as yellow needles of mp 207±210 ꢁC with
positive optical rotation ([a]2D5 +95.0ꢁ). The positive-ion
fast atom bonbardment (FAB)±MS of 1 showed a quasi-
molecular ion peak at m/z 591 (M+Na)+, while a quasi-
molecular ion peak was observed at m/z 567 (M H) in
the negative-ion FAB±MS. The molecular formula
C28H24O13 of 1 was determined from the quasimolecular
Table 1. 13C NMR Data for chrysophanol 8-O-b-d-(60-galloyl)-glu-
copyranoside (1) and aloe-emodin 1-O-b-d-glucopyranoside (2)a
1
2
1
2
C-1
C-2
C-3
C-4
C-5
C-6
C-7
C-8
161.5
119.3
147.5
123.9
122.2
138.5
120.5
157.8
187.4
181.9
132.0
119.3
114.6
134.7
21.4
158.3
119.2
151.8
118.0
118.3
136.1
124.2
161.3
187.6
182.1
134.5
116.7
119.0
132.4
Glc-10
Glc-20
Glc-30
Glc-40
Glc-50
Glc-60
100.2
73.2
76.3
69.7
74.0
63.3
100.5
73.2
76.5
69.3
77.2
60.4
Aloe-emodin 1-O-b-d-glucopyranoside (2) was isolated
as yellow needles of mp 174±177 ꢁC with negative optical
rotation ([a]D24 56.4ꢁ). The molecular formula C21H20
O10 of 2 was determined from the positive- and nega-
tive-ion FAB±MS [m/z 455 (M+Na)+, 431 (M H) ]
and by high-resolution MS measurement. The UV
spectrum of 2 indicated the presence of the anthraqui-
none moiety from the characteristic absorption maxima
at 222 (log e 4.55), 257 (4.37), 282 (sh, 4.06), and 407
(3.84) nm. The IR spectrum showed absorption bands at
3410, 1645, 1638, 1603, 1458, and 1072 cm 1, suggesting
the presence of hydroxyl, free and chelated carbonyl, and
Galloyl-100
Galloyl-200,600
Galloyl-300,500
Galloyl-400
119.2
108.6
145.5
135.6
165.1
C-9
C-10
C-4a
C-8a
C-9a
C-10a
3-CH3
3-CH2OH
Galloyl-700
62.1
aMeasured in DMSO-d6 at 125 MHz.
Figure 1. HMBC and NOE correlations of 1 and 2.

44
H. Matsuda et al. / Bioorg. Med. Chem. 9 (2001) 41±50
1
aromatic functions. The H NMR (DMSO-d6) and 13C
NMR (Table 1) spectra of 2 indicated the presence of a
methylene bearing a hydroxyl group [d 4.65 (2H, s, 3-
DPPH radical and O2 scavenging activities of MeOH
extracts from rhubarbs and the fractions from
R. undulatum
ꢀ
CH2OH)],
a b-d-glucopyranosyl [d 5.14 (1H, d,
J=7.6 Hz, Glc-10-H)], and an anthraquinone structure
[d 7.36 (1H, dd, J=1.2, 8.2 Hz, 7-H), 7.62 (1H, d,
J=1.2 Hz, 2-H), 7.68 (1H, dd, J=1.2, 7.6 Hz, 5-H), 7.75
(1H, dd, J=7.6, 8.2 Hz, 6-H), 7.91 (1H, d, J=1.2 Hz, 4-
H)]. Acid hydrolysis of 2 with 5% sulfuric acid furn-
ished d-glucose, which was identi®ed by GLC analysis
of the trimethylsilyl (TMS) thiazolidine derivative.15
Enzymatic hydrolysis of 2 with b-glucosidase liberated
aloe-emodin (3).16 The carbon signals of the aglycone
part in the 13C NMR spectrum of 2 were super-
imposable on those of aloe-emodin 8-O-b-d-glucopyr-
anoside9,17 or aloe-emodin !-O-b-d-glucopyranoside,9
except for the signals due to a glycoside linked position.
In the HMBC experiment of 2, long-range correlation
was observed between 10-proton and 1-carbon, as
shown in Figure 1. In addition, in the nuclear Over-
hauser e€ect spectroscopy (NOESY) experiment on 2,
the NOE correlation was observed between the follow-
ing proton pairs: 10-H and 2-H; 3-CH2OH and 2,4-H, as
shown in Figure 1. Consequently, 2 was determined to
be 1-O-b-d-glucopyranosyl-aloe-emodin.
The DPPH radical, which is stable and shows an
absorption at 517 nm, has been used as a convenient
tool for the radical scavenge assay, and this assay is
independent of any enzyme activity.23,24 When this
compound accepts an electron or hydrogen radical to
become a more stable compound, the absorption van-
ishes. The xanthine±xanthine oxidase system was con-
ventionally used for generation of ꢀ O2 , which was
detected by reduction of nitroblue tetrazolium (NBT) in
the present study.25,26 The DPPH radical and ꢀ O2
scavenging activities of the methanolic extracts from
®ve kinds of rhubarb, R. palmatum, R. tanguticum, R.
ocinale, R. coreanum, and R. undulatum, were exam-
ined. As shown in Table 2, all extracts scavenged both
ꢀ
the DPPH radical and O2 . However, the methanolic
extract of R. undulatum did not show the most activity
among them.
Next, we isolated the chemical constituents from the
rhizome of R. undulatum, since neither chemical nor
pharmacological studies have been adequately reported.
The H2O- and methanol-eluted fractions showed potent
scavenging activity for the DPPH radical and ꢀ O2 ,
while the acetone-eluted fraction showed little activity.
A reference compound, a-tocopherol, showed scaven-
ging activity for the DPPH radical, but showed little
Preparation of anthraquinone and stilbene derivatives
To clarify the structure±activity relationships in the
radicals scavenging activity, the following related com-
pounds were prepared. Aloe-emodin (3), chrysophanol
(4),17,18 and isorhapontigenin (12)19 were derived by
enzymatic hydrolysis of 2, 6 and 11 with b-glucosidase
in 0.2 M acetate bu€er (pH 5.0) for 16 h at 37 ꢁC,
respectively. Dihydrostilbenes (8a, 9a, 13a, 14a,20 16a,20
and 19a) were derived by hydrogenation of 8, 9, 13, 14,
16, and trans-stilbene (19) in the presence of palladium±
carbon. The completely methylated derivatives (8b, 9b,
13b,21 and 16b20) were prepared by CH3I methylation of
8, 9, 13 and 16. In addition, partially methylated deri-
vatives (8c,20 9c22) were prepared by methanolysis of 8b
and 9b.
ꢀ
activity for O2 . However, gallic acid and (+)-catechin
showed potent scavenging of both radicals. These poly-
phenols were reported to inhibit the xanthine oxidase
activity,27 but they did not inhibit the enzyme activity in
the present conditions at less than 10 mM.
Scavenging e€ects of chemical constituents from
R. undulatum and their derivatives on the DPPH radical
ꢀ
and O2
An anthraquinone glucoside (1) with a 60-galloyl moiety
showed potent DPPH radical and ꢀ O2 scavenging
Table 2. DPPH radical and ꢀ O2 scavenging activities and xanthine oxidase inhibitory activity of ®ve rhubarbs
DPPH radical
ꢀ O2
Formazan formation
IC50
Xanthine oxidase
IC50
a
SC50
R. palmatum MeOH ext.
R. tanguticum MeOH ext.
R. ocinale MeOH ext.
R. coreanum MeOH ext.
R. undulatum MeOH ext.
H2O-eluted fraction
MeOH-eluted fraction
Acetone-eluted fraction
a-Tocopherol
5.2 mg/mL
2.6 mg/mL
3.3 mg/mL
5.9 mg/mL
7.2 mg/mL
19 mg/mL
3.8 mg/mL
95 mg/mL
11 mM
5.0 mg/mL
4.1 mg/mL
3.8 mg/mL
8.5 mg/mL
6.3 mg/mL
12 mg/mL
3.5 mg/mL
>100 mg/mL (9%)d
>100 mg/mL (3%)d
>100 mg/mL ( 4%)d
>100 mg/mL (11%)d
>100 mg/mL ( 6%)d
±
±
±
±
>100 mg/mL (35%)d
>100 mM (16%)c
1.8 mM
Gallic acid
(+)-Catechin
3.9 mM
6.0 mM
>10 mM (0%)b
>10 mM ( 7%)b
5.3 mM
aConcentration required for 50% reduction of 40 mM DPPH radical.
bValues in parentheses represent the inhibition (%) at 10 mM.
cValue in parentheses represents the inhibition (%) at 100 mM.
dValues in parentheses represent the inhibition (%) at 100 mg/mL.

H. Matsuda et al. / Bioorg. Med. Chem. 9 (2001) 41±50
45
activity, but other anthraquinone glucosides (2±6) as well
as sennosides A and B lacked such an e€ect (Table 3). A
naphthalene glucoside, torachrysone 8-O-b-d-glucopyr-
anoside (7), also showed DPPH radical scavenging
activity (Table 3). Gallic acid also shows potent radical
scavenging activity.28,29 In agreement with these pre-
vious studies, gallic acid showed potent radical scaven-
ging activity in the present study (Table 2), while 5
lacked the activity. These ®ndings indicate that the 60-
galloyl moiety is essential to show the radical scaven-
ging activity. Similarly, two stilbenes with a galloyl
moiety [rhaponticins 200-O-gallate (17) and 600-O-gallate
(18)] showed equipotent activity, and they were also
more potent than rhaponticin (8). Among the stilbene
constituents lacking a galloyl moiety, piceatannol (14),
with four hydroxyl groups, showed the most potent
activity, and desoxyrhaponticin (10) showed the least
activity (Table 4).
hydroxyl functions of stilbenes are essential for the
radical scavenging activity.
In addition, we examined their inhibitory activities of
xanthine oxidase to clarify that their e€ects were due to
ꢀ
the scavenging e€ect of O2 or the inhibition of xan-
thine oxidase. As a result, 8c, 13 and 15 showed the
stronger inhibition of xanthine oxidase among them.
These ®ndings indicated that inhibition of formazan
formation by 8c, 13 and 15 was due to their inhibitions
for xanthine oxidase. With regard to their structural
requirements for enzyme inhibition, both 3-hydroxyl
and 40-methoxyl groups are important for the activity.
Antioxidant e€ects of the MeOH extract from
R. undulatum and its stilbene constituents on lipid
peroxidation by the erythrocyte membrane ghost system
As shown in Table 5, the methanolic extract and the
H2O- and methanol-eluted fractions, but not acetone-
eluted fraction, inhibited lipid peroxidation of ery-
throcyte membrane by t-BuOOH. Most stilbene con-
stituents (8±18), except for 10, showed inhibition of lipid
peroxidation of erythrocyte membranes by t-BuOOH.
Scavenging e€ects of anthraquinone and stilbene
ꢀ
derivatives on DPPH radical and O2
As shown in Table 3, dihydrostilbene derivatives (8a,
9a, 13a, 14a and 16a) maintained the scavenging activity
ꢀ
for the DPPH radical, but their O2 scavenging activ-
ities were reduced. Since trans-stilbene (19) and 19a
lacked the oxygen function and methylated stilbenes
(8b, 9b, 9c, 13b and 16b) lacked the activity, the phenolic
Several stilbenes, including piceatannol (14) and resver-
atrol (16), were recently reported to show antioxidant
activity (e.g., DPPH radical scavenging activity, Cu2+
-
Table 3. Radical scavenging activities of sennosides, anthraquinones and a naphthalene
DPPH radical
ꢀ O2
Formazan
formation
IC50 (mM)
Xanthine
oxidase
IC50 (mM)
10±100
SC50 (mM)a
Sennoside A
Sennoside B
threo
erythro
>40 (5%)b
>40 (6%)b
>100 (0%)c
>100 ( 3%)c
±
±
R1
R2
R3
R4
Chrysophanol 8-O-(60-galloyl)-Glc (1)
Aloe-emodin 1-O-Glc (2)
Aloe-emodin (3)
Chrysophanol (4)
Chrysophanol 1-O-Glc (5)
Chrysophanol 8-O-Glc (6)
Emodin
H
Glc
H
H
Glc
H
H
H
H
CH3
CH2OH
CH2OH
CH3
CH3
CH3
CH3
COOH
CH3
H
H
H
H
H
H
OH
H
Glc(6-gallate)
13
23
>100 ( 6%)c
H
H
H
H
Glc
H
>40 (5%)b >100 (8%)c
>40 (5%)b >100 ( 3%)c
±
±
>40 (19%)b >100 (32%)c >100 (20%)c
>40 (9%)b >100 ( 6%)c
>40 (18%)b >100 (12%)c
>40 (40%)b >100 (17%)c
>40 (2%)b >100 (1%)c
>40 (2%)b >100 (9%)c
±
±
±
±
±
Rhein
Physcion
H
H
OCH3
Torachrysone 8-O-Glc (7)
16
>100 (38%)c >100 (20%)c
aConcentration required for 50% reduction of 40 mM DPPH radical.
bValues in parentheses represent the reduction or inhibition (%) at 40 mM.
cValues in parentheses represent the reduction or inhibition (%) at 100 mM.
dGlc, b-d-glucopyranosyl.

46
H. Matsuda et al. / Bioorg. Med. Chem. 9 (2001) 41±50
Table 4. Radical scavenging activities of stilbene constituents and their derivatives
DPPH radical
ꢀ O2
Formazan
formation
Xanthine
oxidase
a±b
R1
R2
R3
R4
SC50 (mM)a
IC50 (mM)
IC50 (mM)
Rhaponticin (8)
CˆC
O-Glce
O-Glc
O-Glc(CH3)4
OH
OH
OH
OCH3
OCH3
OH
OH
OCH3
OCH3
OCH3
OCH3
29
28
38
>30 (0%)b
8a
8b
8c
C
C
>100 (38%)d >100 (0%)d
CˆC
OCH3 >40 (1%)c
OCH3 >40 (10%)c
>100 (27%)d
24
±
39
CˆC
Piceatannol 30-O-Glc (9)
9a
9b
9c
CˆC
C C
OH
OH
OCH3
OCH3
OH
OH
OCH3
OCH3
O-Glc
O-Glc
O-Glc(CH3)4
OH
OH
OH
15
26
16
58
>30 ( 3%)b
>100 ( 4%)d
±
CˆC
OCH3 >40 (1%)c
>100 (9%)d
CˆC
OCH3 >40 (43%)c >100 (36%)d >30 ( 1%)b
Desoxyrhaponticin (10)
Isorhapontin (11)
12
CˆC
CˆC
CˆC
O-Glc
O-Glc
OH
OH
OH
OH
H
OCH3
OCH3
OCH3 >40 (9%)c
OH
OH
>100 (33%)d
±
>40 (29%)c >100 (49%)d >100 (-6%)d
32
21
>30 (29%)b
Rhapontigenin (13)
13a
13b
CˆC
OH
OH
OCH3
OH
OH
OCH3
OH
OH
OCH3
OCH3
OCH3
24
20
20
34
C
C
>100 (49%)d >30 (23%)b
CˆC
OCH3 >40 (1%)c
>30 (26%)b
>30 (32%)b
Piceatannol (14)
14a
Desoxyrhapontigenin (15)
CˆC
OH
OH
OH
OH
OH
OH
OH
OH
H
OH
OH
11
8.6
6.8
34
11
>30 (34%)b
>100 (17%)d
24
C-C
CˆC
OCH3 >40 (45%)c
Resveratrol (16)
16a
16b
CˆC
OH
OH
OCH3
OH
OH
OCH3
H
H
H
OH
OH
24
>100 (40%)d >30 (9%)b
C
C
>40 (31%)c >100 (11%)d
±
±
CˆC
OCH3 >40 (2%)c
>100 ( 5%)d
Rhaponticin 200-O-gallate (17) CˆC O-Glc(2-gallate)
OH
OH
H
OH
OH
H
OCH3
OCH3
H
4.8
4.8
>40 (1%)c
>40 (0%)c
11
13
>30 ( 8%)b
>30 ( 3%)b
Rhaponticin 600-O-gallate (18) CˆC O-Glc(6-gallate)
Trans-stilbene (19)
19a
CˆC
H
H
>100 (2%)d
>100 ( 4%)d
±
±
C
C
H
H
H
aConcentration required for 50% reduction of 40 mM DPPH radical.
bValues in parentheses represent the reduction or inhibition (%) at 30 mM.
cValues in parentheses represent the reduction or inhibition (%) at 40 mM.
dValues in parentheses represent the reduction or inhibition (%) at 100 mM.
eGlc, b-d-glucopyranosyl.
induced lipid peroxidation of LDL, TPA-induced free
radical formation in HL-60 cells).30,31 Since most stil-
benes from R. undulatum showed antioxidant activities,
the ®ndings in this study present signi®cant evidence in
support of the previous study.30,31
Table 5. Antioxidative activity of the extracts from R. undulatum and
its stilbene constituents for lipid peroxidation by the erythrocyte
membrane ghost system
Sample
IC50
MeOH Ext.
5.3 mg/mL
8.3 mg/mL
3.3 mg/mL
55 mg/mL
Stilbenes from the rhizome of R. undulatum were repor-
ted to inhibit NO production in lipopolysaccharide-
activated macrophages.3 Recently, it was revealed that
peroxynitrite (ONOO ), which is easily produced by the
H2O-eluted fraction
MeOH-eluted fraction
Acetone-eluted fraction
Rhaponticin (8)
9.5 mM
40 mM
ꢀ
reaction of NO with O2 , is a highly reactive oxidant
Piceatannol 30-O-b-d-glucopyranoside (9)
Desoxyrhaponticin (10)
Isorhapontin (11)
and induces various oxidative damage.32 This rhubarb
and its stilbene constituents with the antioxidant and
NO production inhibitory activities may be important
evidence substantiating these traditional e€ects.
>100 mM (9%)a
33 mM
Rhapontigenin (13)
Piceatannol (14)
Desoxyrhapontigenin (15)
Resveratrol (16)
Rhaponticin 200-O-gallate (17)
Rhaponticin 600-O-gallate (18)
18 mM
6.0 mM
49 mM
47 mM
11 mM
12 mM
In conclusion, the methanolic extracts from ®ve kinds of
rhubarb were found to show scavenging activity for the
ꢀ
DPPH radical and O2 . In screening tests for DPPH
ꢀ
a-Tocopherol
(+)-Catechin
6.9 mM
2.4 mM
radical and O2 scavenging activity, most stilbenes and
a naphthalene glucoside, torachrysone 8-O-glucoside,
isolated from R. undulatum showed the activity. However,
aValue in parentheses represents inhibition (%) at 100 mM.

H. Matsuda et al. / Bioorg. Med. Chem. 9 (2001) 41±50
47
anthraquinone glucosides (chrysophanol 1-O-b-d-glu-
copyranoside and chrysophanol 8-O-b-d-glucopyrano-
side) and sennosides A and B did not. In addition, most
stilbenes, except for desoxyrhaponticin (10), showed
lipid peroxidation of erythrocyte membranes by t-
BuOOH. To elucidate the structure±activity relation-
ships in anthraquinones and stilbenes, we examined
various related compounds. The ®ndings of these
experiments suggested the following structural require-
ments: 1) phenolic hydroxyl groups of stilbene are
essential to show the activity; 2) the galloyl moiety
enhanced the activity; 3) the glucoside moiety of stilbene
reduced the activity; 4) dihydrostilbene derivatives
maintained the scavenging activity for DPPH radical,
pressure gave the MeOH extracts [110 g (22.1%) from
R. palmatum, 147 g (29.4%) from R. tanguticum, 162 g
(32.4%) from R. ocinale, 141 g (28.1%) from R. cor-
eanum, and 1960 g (33.8%) from R. undulatum].
The MeOH extract (850 g) from R. undulatum was sub-
jected to Diaion HP-20 column chromatography
(3.0 kg, H2O!MeOH!acetone) to a€ord the H2O-
eluted fraction (335 g, 13.3%), MeOH-eluted fraction
(470 g, 18.7%), and acetone-eluted fraction (45 g, 1.8%).
The MeOH-eluted fraction (150g) was separated by
ordinary-phase silica gel column chromatography [3.0kg,
CHCl3±MeOH (10:1, v/v)!(4:1)!CHCl3±MeOH±H2O
(10:3:1, lower layer)!MeOH] to give ®ve fractions [Fr.
1 (4.4 g), Fr. 2 (16.5 g), Fr. 3 (14.8 g), Fr. 4 (71.5 g), Fr. 5
(42.8 g)]. Fraction 2 (12.0 g) was further subjected to
reversed-phase silica gel column chromatography [360 g,
MeOH±H2O (50:50 v/v)!(80:20)!MeOH] to a€ord
®ve fractions [Fr. 2-1 (5.3 g), Fr. 2-2 (0.6 g), Fr. 2-3
(3.2 g), Fr. 2-4 (1.2 g), Fr. 2-5 (1.7 g)]. Finally, fraction
2-1 (500 mg) was subjected to HPLC [YMC-pack R&D
ODS-5-A, 250Â20 mm id, MeOH±H2O (50:50 v/v)] to
give rhapontigenin (13, 292 mg, 0.58%), piceatannol (14,
40 mg, 0.073%), and resveratrol (16, 26 mg, 0.048%).
Through a similar procedure, fraction 2-2 (300 mg) was
subjected to HPLC [MeOH±H2O (55:45 v/v)] to furnish
desoxyrhapontigenin (15, 41 mg, 0.015%), fraction 2-3
(300 mg) was subjected to HPLC [CH3CN±H2O
(40:60 v/v)] to give chrysophanol 1-O-b-d-glucopyrano-
side (5, 136 mg, 0.25%) and chrysophanol 8-O-b-d-glu-
copyranoside (6, 87 mg, 0.16%). Fraction 3 (12.0 g) was
further separated by reversed-phase silica gel column
chromatography [360 g, MeOH±H2O (60:40 v/v)!
MeOH] to yield ®ve fractions [Fr. 3-1 (3.3 g), Fr. 3-2
(3.6 g), Fr. 3-3 (3.3 g), Fr. 3-4 (1.4 g), Fr. 3-5 (0.4 g)].
Fraction 3-2 (300 mg) and fraction 3-4 (300 mg) were
further subjected to HPLC [MeOH±H2O (55:45 v/v)] to
furnish chrysophanol 8-O-b-d-(60-galloyl)-glucopyrano-
side (1, 129 mg, 0.092%) and torachrysone 8-O-b-d-
glucopyranoside (7, 65 mg, 0.12%), respectively. Frac-
tion 4 (28.0 g) was subjected to reversed-phase silica gel
column chromatography [1.0 kg, MeOH±H2O (30:70 v/
v) !(40:60)!MeOH] to a€ord seven fractions [Fr. 4-1
(1.5 g), Fr. 4-2 (6.3 g), Fr. 4-3 (11.0 g), Fr. 4-4 (2.1 g), Fr.
4-5 (0.7 g), Fr. 4-6 (3.2 g), Fr. 4-7 (3.2 g)]. Fractions 4-2
and 4-3 were identi®ed as piceatannol 30-O-b-d-gluco-
pyranoside (9, 2.0%) and rhaponticin (8, 3.5%), respec-
tively. Fraction 4-4 (1.45 g) was subjected to HPLC
[MeOH±H2O (50:50 v/v) and (40:60)] to give aloe-emodin
1-O-b-d-glucopyranoside (2, 14 mg, 0.0065%), desoxy-
rhaponticin (10, 104 mg, 0.048%), isorhapontin (11, 782
mg, 0.36%), rhaponticin 200-O-gallate (17, 261mg, 0.12%),
and rhaponticin 600-O-gallate (18,189 mg, 0.087%). These
constituents were identi®ed by comparison of their
physical data with those of authentic samples (14, 16) or
with reported values.9 14
ꢀ
but they showed weak activity for O2 . In addition,
several stilbenes with both the 3-hydroxyl and 40-meth-
oxyl groups inhibited xanthine oxidase.
Experimental
The following instruments were used to obtain physical
data: speci®c rotations, Horiba SEPA-300 digital polari-
meter (l=5 cm); UV spectra, Shimadzu UV-1200 spec-
trometer; IR spectra, Shimadzu FTIR-8100 spectrometer;
EI±MS and high-resolution MS, JEOL JMS-GCMATE
mass spectrometer; FAB±MS and high-resolution MS,
JEOL JMS-SX 102A mass spectrometer; 1H NMR spec-
tra, JNM-LA500 (500 MHz) spectrometer and JEOL EX-
270 (270MHz); 13C NMR spectra, JNM-LA500 (125
MHz) and JEOL EX-270 (68 MHz) spectrometers with
tetramethylsilane as an internal standard.
The following experimental conditions were used for
chromatography: ordinary-phase silica gel column
chromatography, Silica gel BW-200 (Fuji Silysia Che-
mical, Ltd., 150±350 mesh); reversed-phase silica gel
column chromatography, Chromatorex ODS DM1020T
(Fuji Silysia Chemical, Ltd., 100±200 mesh); TLC, pre-
coated TLC plates with Silica gel 60F254 (Merck,
0.25 mm) (ordinary phase) and Silica gel RP-18 60F254
(Merck, 0.25 mm) (reversed phase); reversed-phase
HPTLC, pre-coated TLC plates with Silica gel RP-18
60WF254S (Merck, 0.25 mm); detection was achieved by
spraying with 1% Ce(SO4)2±10% aqueous H2SO4 and
heating.
Extraction and isolation
The dried rhizome of R. palmatum L. [500 g, cultivated
in China (Kansu province) and purchased from MAE
CHU Co., Ltd., Nara, Japan], R. tanguticum Maxim.
[500 g, cultivated in China (Chinghai province) and
purchased from MAE CHU Co., Ltd.], R. ocinale
Baill. [500 g, cultivated in China (Szechwan province)
and purchased from MAE CHU Co., Ltd.], R. cor-
eanum Nakai [500 g, cultivated in Korea and purchased
from MAE CHU Co., Ltd.], and R. undulatum (5.8 kg,
cultivated in Korea), which were botanically identi®ed
by comparison of their taxonomical features with
authentic rhubarb samples, were crushed and then
extracted three times with methanol at room tempera-
ture for 24 h. Evaporation of the solvent under reduced
Chrysophanol 8-O-b-d-(60-galloyl)-glucopyranoside (1):
Yellow needles from MeOH, mp 207±210 ꢁC, [a]D25
+95.0ꢁ (c=0.1, MeOH). High-resolution positive-ion
FAB±MS: calcd for C28H24O13Na (M+Na)+: 591.1114.
Found: 591.1168. UV [MeOH, nm (log e)]: 220 (4.58),

48
H. Matsuda et al. / Bioorg. Med. Chem. 9 (2001) 41±50
260 (4.32), 284 (sh, 4.17), 409 (3.79). IR (KBr): 3410,
1
Enzymatic hydrolysis of aloe-emodin 1-O-ꢀ-D-glucopyra-
noside (2), 6 and 11. A solution of 2 (10 mg, 0.023
mmol), 6 (10 mg, 0.024 mmol), or 11 (10 mg, 0.024
mmol) in 0.2 M acetate bu€er (pH 5.0, 2.0 mL) was
treated with b-glucosidase (10 mg, SIGMA G-0395 from
almonds) and the solution was stirred at 37 ꢁC for 16 h.
The reaction mixture was extracted with AcOEt. The
AcOEt extract was washed with brine, then dried over
MgSO4 powder and ®ltered. Removal of the solvent
from the ®ltrate under reduced pressure furnished a
residue, which was puri®ed by silica gel column chro-
matography [200 mg (n-hexane±acetone=1:1)] to give 3
(5.8 mg, 95%), 4 (6.0 mg, 98%), or 12 (5.9 mg, 96%),
respectively. Compounds 3, 4 and 12 were identi®ed
by comparison of the physical data with reported
values.16,18,19
2923, 1709, 1641, 1637, 1466, 1075 cm
.
1H NMR
(500 MHz, DMSO-d6) d: 2.40 (3H, s, 3-CH3), 4.27, 4.50
(1H each, both m, Glc-60-H2), 5.24 (1H, d, J=7.7, Glc-
10-H), 6.99 (2H, s, galloyl-200,600-H), 7.16 (1H, br s, 4-H),
7.46 (1H, br s, 2-H), 7.67 (1H, d, J=8.6 Hz, 5-H), 7.72
(1H, dd, J=7.6, 8.6 Hz, 6-H), 7.82 (1H, d, J=7.6 Hz, 7-
H), 12.82 (1H, br s, 1-OH). 13C NMR (125 MHz,
DMSO-d6) dC: given in Table 1. Positive-ion FAB±MS
m/z: 591 (M+Na)+. Negative-ion FAB±MS m/z: 567
(M H) .
Aloe-emodin 1-O-b-d-glucopyranoside (2): Yellow
needles from MeOH, mp 174±177 ꢁC, [a]D24 56.4ꢁ
(c=0.1, MeOH). High-resolution positive-ion FAB±MS:
calcd for C21H20O10Na (M+Na)+: 455.0954. Found:
455.0952. UV [MeOH, nm (log e)]: 222 (4.55), 257
(4.37), 282 (sh, 4.06), 407 (3.84). IR (KBr): 3410, 2924,
1645, 1638, 1603, 1458, 1072 cm 1. 1H NMR (500 MHz,
DMSO-d6) d: 4.65 (2H, s, 3-CH2OH), 5.14 (1H, d,
J=7.6 Hz, Glc-10-H), 7.36 (1H, dd, J=1.2, 8.2 Hz, 7-H),
7.62 (1H, d, J=1.2 Hz, 2-H), 7.68 (1H, dd, J=1.2,
7.6 Hz, 5-H), 7.75 (1H, dd, J=7.6, 8.2 Hz, 6-H), 7.91
(1H, d, J=1.2 Hz, 4-H), 12.90 (1H, br s, 8-OH). 13C
NMR (125 MHz, DMSO-d6) dC: given in Table 1. Posi-
tive-ion FAB±MS m/z: 455 (M+Na)+. Negative-ion
FAB±MS m/z: 431 (M H) , 269 (M+±C6H11O5) .
Hydrogenation of 8, 9, 13, 14, 16, and trans-stilbene
(19). A solution of 8, 9, 13, 14, 16, or trans-stilbene
(19) (20 mg each) in EtOH (2.0 mL) was treated with
10% palladium carbon (Pd±C, 10 mg) and the whole
mixture was stirred at room temperature under an H2
atmosphere for 2 h. The reaction mixture was ®ltered
and then evaporation of the solvent under reduced
pressure furnished dihydrostilbene derivatives (8a, 9a,
13a, 14a, 16a, or 19a) quantitatively. Compounds 14a,
16a and 19a were identi®ed by comparison of the phy-
sical data with reported values.20
Alkaline hydrolysis of chrysophanol 8-O-ꢀ-D-(60-galloyl)-
glucopyranoside (1). A solution of 1 (10 mg, 0.018
mmol) in 1.0% NaOMe±MeOH (2.0 mL) was stirred at
room temperature for 1 h. The reaction mixture was
neutralized with Dowex HCR-W2 (H+ form) and the
resin was removed by ®ltration. Evaporation of the sol-
vent from the ®ltrate under reduced pressure gave a
residue, which was puri®ed by silica gel column chro-
matography [1.0 g (CHCl3±MeOH±H2O=10:3:1, lower
layer)] to furnish 6 (7.2 mg, 98%), which was identi®ed
by comparison of the physical data with reported
values.9
8a: Light yellow powder, [a]2D5 37.2ꢁ (c=0.1, MeOH).
High-resolution EI±MS: calcd for C21H26O9 (M+):
422.1576. Found: 422.1588. UV [MeOH, nm (log e
(4.30), 282 (3.76), 322 (3.60). IR (KBr): 3480, 1636,
1559, 1509, 1075 cm 1. 1H NMR (270 MHz, CD3OD) d:
2.74 (4H, br s, a- and b-H2), 3.85 (3H, s, -OMe), 4.77
(1H, d, J=7.3 Hz, Glc-100-H), 6.30 (1H, t, J=2.0 Hz, 4-
H), 6.37, 6.38 (1H each, both d, J=2.0 Hz, 2, and 6-H),
6.57 (1H, dd, J=2.3, 8.2 Hz, 60-H), 6.64 (1H, d,
J=2.3 Hz, 20-H), 6.78 (1H, d, J=8.2 Hz, 50-H). EI±MS
m/z: 422 (M+, 2), 260 (64), 137 (100).
Acid hydrolysis of aloe-emodin 1-O-ꢀ-D-glucopyranoside
(2). A solution of aloe-emodin 1-O-b-d-glucopyrano-
side (2, 2 mg) in 5% aq H2SO4±1,4-dioxane (1:1, v/v,
1 mL) was heated under re¯ux for 1 h. After cooling, the
reaction mixture was neutralized with Amberlite IRA-
400 (OH form) and the insoluble portion was removed
by ®ltration. After removal of the solvent from the ®l-
trate under reduced pressure, the residue was separated
on a Sep-Pack C18 cartridge column (H2O, MeOH).
The H2O eluate was concentrated under reduced pres-
sure to give a residue, which was treated with l-cysteine
methyl ester hydrochloride (2 mg) in pyridine (0.02 mL)
and the mixture was left standing at 60 ꢁC for 1 h. The
reaction solution was then treated with N,O-bis(tri-
methylsilyl)tri¯uoroacetamide (BSTFA, 0.01 mL) and
the whole mixture was left standing at 60 ꢁC for 1 h. The
supernatant of the reaction mixture was subjected to
GLC analysis to identify the thiazolidine derivatives of
d-glucose. GLC conditions: column, Supelco SPBTM-1,
0.25 mm idÂ30 m; injection temperature, 230 ꢁC; detec-
tion temperature, 230 ꢁC; column temperature, 230 ꢁC;
He ¯ow rate 15 mL/min. tR: 24.2 min.
9a: Light yellow powder, [a]2D3 33.3ꢁ (c=0.1, MeOH).
High-resolution positive-ion FAB±MS: calcd for C20
H24O9Na (M+Na)+: 431.1319. Found: 431.1310. UV
[MeOH, nm, (log e)]: 213 (4.31), 282 (3.75), 301 (3.55).
1
IR (KBr): 3480, 1636, 1559, 1509, 1081 cm 1. H NMR
(270 MHz, CD3OD) d: 2.75 (4H, m, a- and b-H2), 4.63
(1H, d, J=7.6 Hz, Glc-100-H), 6.09 (3H, br s, 2, 4, and 6-
H), [6.72 (2H, br d), 6.93 (1H, br s), 20, 50, and 60-H].
Positive-ion FAB±MS m/z: 431 (M+Na)+.
13a: Light yellow powder. High-resolution EI±MS:
calcd for C15H16O4 (M+): 260.1048. Found: 260.1041.
UV [MeOH, nm (log e)]: 223 (4.25), 295 (4.05), 305
(4.10). IR (KBr): 3478, 1614, 1559, 1509 cm 1. 1H NMR
(270 MHz, CD3OD) d: 2.69 (4H, m, a- and b-H2), 3.85
(3H, s, -OMe), 6.16 (1H, t, J=2.0 Hz, 4-H), 6.43 (2H, d,
J=2.0 Hz, 2 and 6-H), 6.88 (1H, br d, 50-H), 6.88 (1H,
dd-like, 60-H), 7.00 (1H, d, J=1.7 Hz, 20-H). EI±MS m/z:
260 (M+, 25), 137 (100).
Complete methylation of 8, 9, 13 and 16. A solution of 8
or 9 (150mg each, 0.36mmol) in N,N-dimethylformamide

H. Matsuda et al. / Bioorg. Med. Chem. 9 (2001) 41±50
49
(DMF, 5.0 mL) was treated with methyl iodide (CH3I,
Bioassay Methods
0.5 mL) in the presence of sodium hydride (NaH, 58 mg)
and the mixture was stirred at room temperature for 2 h.
The reaction mixture was poured into ice-water and the
whole was extracted with AcOEt. The AcOEt extract
was successively washed with saturated aqueous
NaHCO3 and brine, then dried over MgSO4 powder
and ®ltered. Removal of the solvent from the ®ltrate
under reduced pressure furnished a residue, which was
puri®ed by silica gel column chromatography [5.0 g
(n-hexane±AcOEt=1:1)] to give 8b (115 mg, 64%) or 9b
(125 mg, 68%). Through a similar procedure, a solution
of 13 (30 mg, 0.12 mmol) or 16 (30 mg, 0.13 mmol) in
DMF (3.0 mL) using CH3I (0.3 mL) and NaH (20 mg)
was stirred at room temperature for 2 h. The reaction
mixture was poured into ice-water and the whole was
extracted with AcOEt. The AcOEt extract was treated in
the usual manner to give a residue, which was puri®ed by
silica gel column chromatography [500 mg, (n-hexane±
AcOEt=5:1)] to furnish 13b (33 mg, 86%) or 16b
(36 mg, quant.). Compounds 13b and 16b were identi-
®ed by comparison of the physical data with reported
values.20,21
DPPH radical scavenging activity
The free radical scavenging activity of the constituents
of rhubarb was assessed using the DPPH radical.23,24
An ethanol solution of DPPH (100 mM, 1.0 mL) was
mixed with di€erent concentrations of each test com-
pound (0±200 mM, 0.5 mL) and 0.1 M acetate bu€er (pH
5.5, 1.0 mL), and the absorbance change at 517 nm was
measured 30 min later. The reaction solution without
DPPH was used as a blank test. Measurements were
performed in duplicate, and the concentration required
for a 50% reduction (50% scavenging concentration,
SC50) of 40 mM DPPH radical solution was determined
graphically.
ꢀ O2 scavenging activity
The improved assay method for superoxide dismutase
described by Imanari et al. was used.25 Brie¯y, a reac-
tion mixture containing 100 mM xanthine, 100 mM
EDTA, 25 mM NBT, 0.005% bovine serum albumin,
and ca. 1.8 mU/mL xanthine oxidase in 33.3 mM
sodium carbonate bu€er (pH 10.2) was incubated with
or without each test sample for 20 min at 25 ꢁC (total
volume: 3.0 mL). After incubation, the solution was
mixed with 0.1 mL of 6 mM CuCl2 to stop the reaction.
The formazan formation was monitored at 560 nm. In
addition, inhibitory e€ects of test compounds on xan-
thine oxidase activity were examined to clarify whether
the inhibition of formazan formation was due to inhi-
bition of xanthine oxidase. The reaction mixture with-
out NBT was incubated in similar conditions described
above and 0.1 mL of 2 M HCl was added to stop the
reaction. Uric acid formation was monitored at 290 nm.
Several stilbenes (8, 8c, 9, 9c, 12, 13, 13a, 13b, 14, 15, 16,
17, 18) at 100 mM showed high optical density in the
blank test, therefore 30 mM or 50 mM was chosen as the
maximum concentration.
8b: Yellow powder, [a]2D5 68.0ꢁ (c=0.3, MeOH). High-
resolution EI±MS: calcd for C27H36O9 (M+): 504.2359.
Found: 504.2373. UV [MeOH, nm, (log e)]: 220 (4.38),
305 (4.35), 323 (4.44). IR (KBr): 1610, 1592, 1516,
1
1065 cm 1. H NMR (270 MHz, CDCl3) d: 3.39, 3.56,
3.66, 3.68, 3.82, 3.89, 3.94 (3H each, both s, -OMe), 4.86
(1H, d, J=7.3 Hz, Glc-100-H), 6.53 (1H, t, J=2.0 Hz, 4-
H), 6.72, 6.82 (1H each, both br s, 2, and 6-H), 6.85
(1H, d, J=8.2 Hz, 50-H), 6.88, 7.03 (1H, each, both d,
J=16.2 Hz, a, and b-H), 7.03 (1H, dd, J=2.0, 8.2 Hz,
60-H), 7.06 (1H, br s, 20-H). EI±MS m/z: 504 (M+, 23),
286 (100).
9b: Yellow powder, [a]2D6 33.2ꢁ (c=0.1, MeOH). High-
resolution EI±MS: calcd for C27H36O9 (M+): 504.2359.
Found: 504.2357. UV [MeOH, nm (log e)]: 213 (3.68),
301 (3.03), 325 (2.98). IR (KBr): 1630, 1582, 1516,
Antioxidant activity for lipid peroxidation by the
erythrocyte membrane ghost system
1
1060 cm 1. H NMR (270 MHz, CDCl3) d: 3.70, 3.78,
3.79, 3.85, 3.86 (3H each, both s, -OMe), 3.82 (6H, s,
-OMe), 4.82 (1H, d, J=7.6 Hz, Glc-100-H), 6.38 (1H, t,
J=2.2 Hz, 4-H), 6.64 (2H, d, J=2.2 Hz, 2- and 6-H),
6.81, 6.83 (1H, each, both d, J=17.3 Hz, a- and b-H),
[7.02 (1H, br s), 7.17 (2H, br s), 20, 50, and 60-H]. EI±MS
m/z: 504 (M+,1), 286 (100).
The method described by Osawa et al. was used.33
Brie¯y, rabbit blood was washed with isotonic bu€er
solution (10 mM phosphate/152 mM NaCl, pH 7.4)
three times, and lysed in hypotonic bu€er solution
(10 mM phosphate bu€er, pH 7.4). Erythrocyte mem-
brane ghosts were pelleted by centrifugation (20,000Âg,
40 min), and the precipitate was diluted to give a sus-
pension (2.5 mg protein/mL) as determined by the
Lowry method. The reaction mixture containing ery-
throcyte membrane ghost suspension (0.85 mL), t-
BuOOH (24 mM, 0.05 mL) and di€erent concentrations
of each test compound in ethanol (0.1 mL) was incu-
bated for 30 min at 37 ꢁC. The solution was cooled in an
ice-cold water bath, and 10 mL of 2,6-di-tert-butyl-4-
hydroxytoluene (BHT, 50 mM), 1.5 mL of thiobarbi-
turic acid (TBA) reagent composed of 0.375% TBA,
15% trichloroacetic acid, and 8.8 mg/mL BHT in
0.25 M HCl were added to the solution. After heating in
boiling water for 15 min, the solution was cooled and
centrifuged (4 ꢁC, 1000Âg, 15 min) and the quantity of
Methanolysis of 8b and 9b
A solution of 8b or 9b (50 mg each, 0.10mmol) in 9%
HCl±dry MeOH (4.0 mL) was heated under re¯ux for 2 h.
The reaction mixture was poured into ice-water and the
whole was extracted with AcOEt. The AcOEt extract was
successively washed with saturated aqueous NaHCO3 and
brine, then dried over MgSO4 powder and ®ltered.
Removal of the solvent from the ®ltrate under reduced
pressure furnished a residue, which was puri®ed by
silica gel column chromatography [500 mg (n-hexane±
AcOEt=3:1)] to give 8c (27 mg, 98%) and 9c (22 mg,
81%), respectively. 8c and 9c were identi®ed by com-
parison of the physical data with reported values.20,22

50
H. Matsuda et al. / Bioorg. Med. Chem. 9 (2001) 41±50
TBA-reacting substance in supernatant was determined
at 535 nm.
17. Ko, S. K.; Whang, W. K.; Kim, I. H. Arch. Pharm. Res.
1995, 18, 282.
18. Kato, K.; Morita, Y. Shoyakugaku Zasshi 1987, 41, 67.
19. Reimann, E. Liebigs Ann. Chem. 1971, 750, 109.
20. Mannila, E.; Talvitie, A.; Kolehmainen, E. Phytochem-
istry 1993, 33, 813.
References
21. Talvitie, A.; Mannila, E.; Kolehmainen, E. Liebigs Ann.
Chem. 1992, 399.
22. Gill, M. T.; Bajaj, R.; Chang, C. J.; Nichols, D. E.;
McLaughlin, J. L. J. Nat. Prod. 1987, 50, 36.
23. Uchiyama, M.; Suzuki, Y.; Fukuzawa, K. Yakugaku Zasshi
1968, 88, 678.
1. Kubo, M.; Ko, S.; Hiraba, K.; Harima, S.; Matsuda, H.;
Kim, I. J. Traditional Medicines 1997, 14, 237.
2. Ko, S. K.; Lee, S. M.; Whang, W. K. Arch. Pharm. Res.
1999, 22, 401.
3. Matsuda, H.; Kageura, T.; Morikawa, T.; Toguchida, I.;
Harima, S.; Yoshikawa, M. Bioorg. Med. Chem. Lett. 2000,
10, 323.
4. Kehrer, J. P. Crit. Rev. Toxicol. 1993, 23, 21.
5. Sotomatsu, A. Saishin-Igaku 1990, 45, 1748.
6. MacCord, J. M. Adv. Free Radic. Biol. Med. 1986, 2, 325.
7. Kensler, T. W.; Ta€e, B. G. Adv. Free Radic. Biol. Med.
1986, 2, 347.
8. Paolisso, G.; Giugliano, D. Diabetologia 1996, 39, 357.
9. Okabe, H.; Matsuo, K.; Nishioka, I. Chem. Pharm. Bull.
1973, 21, 1254.
24. Blois, M. S. Nature 1958, 181, 1199.
25. Imanari, T.; Hirota, M.; Miyazaki, M.; Hayakawa, K.;
Tamura, Z. Igaku No Ayumi 1977, 101, 496.
26. McCord, J. M.; Fridovich, I. J. Biol. Chem. 1969, 244,
6049.
27. Hatano, T.; Yasuhara, T.; Yoshihara, R.; Agata, I.; Noro,
T.; Okuda, T. Chem. Pharm. Bull. 1990, 38, 1224.
28. Nogaki, A.; Satoh, K.; Iwasaka, K.; Takano, H.; Taka-
hama, M.; Ida, Y.; Sakagami, H. Anticancer Res. 1998, 18,
3487.
10. Tuboi, M.; Minami, M.; Nonaka, G.; Nishioka, I. Chem.
Pharm. Bull. 1977, 25, 2708.
11. Kashiwada, Y.; Nonaka, G.; Nishioka, I. Chem. Pharm.
Bull. 1984, 32, 3501.
12. Pan, H.; Lundgren, L. N. Phytochemistry 1995, 39, 1423.
13. Banks, H. J.; Cameron, D. W. Aust. J. Chem. 1971, 24,
2427.
29. Ohsugi, M.; Fan, W.; Hase, K.; Xiong, Q.; Tezuka, Y.;
Komatsu, K.; Namba, T.; Saitoh, T.; Tazawa, K.; Kadata, S.
J. Ethnopharmacol. 1999, 67, 111.
30. Lee, S. K.; Mbwambo, Z. H.; Chung, H.; Luyengi, L.;
Gamez, E. J. C.; Mehta, R. G.; Kinghorn, A. D.; Pezzuto, J.
M. Combinatorial Chemistry & High Throughput Screening
1998, 1, 35.
14. Nakajima, K.; Taguchi, H.; Endo, T.; Yoshioka, I. Chem.
Pharm. Bull. 1978, 26, 3050.
15. Hara, S.; Okabe, H.; Mihashi, K. Chem. Pharm. Bull.
1986, 34, 1843.
31. Teguo, P. W.; Fauconneau, B.; Deeux, G.; Huguet, F.;
Vercauteren, J.; Merillon, J. J. Nat. Prod. 1998, 61, 655.
32. Gross, S. P. A.; Singh, R. J.; Hogg, N.; Kalyanaraman, B.
Free Radic. Res. 1999, 31, 597.
16. Uchibayashi, M.; Matsuoka, T. Chem. Pharm. Bull. 1961,
9, 234.
33. Osawa, T.; Ide, A.; Su, J. D.; Namiki, M. J. Agric. Food
Chem. 1987, 35, 808.