Comparison of the cytotoxic activities of naturally occurring hydroxyanthraquinones and hydroxynaphthoquinones

Article abstract of DOI:10.1016/j.ejmech.2007.08.009

Seven hydroxyanthraquinone derivatives, 1-7, were isolated from the root of Rheum palmatum (Polygonaceae). Two propionated anthraquinone derivatives, 8 and 9, were synthesized. Four hydroxynaphthoquinone derivatives, 13, 14, 16 and 21, were isolated from the root of Lithospermum erythrorhizon Sieb. et Zucc. (Boraginaceae) and also three naphthoquinone derivatives, 19, 22 and 23, were isolated from the root of Macrotomia euchroma (Royle) Pauls. (Boraginaceae). The cytotoxicity of the anthraquinone and naphthoquinone derivatives on P-gp-underexpressing HCT 116 cells and P-gp-overexpressing Hep G2 cells was examined by MTT assay. Among the anthraquinone derivatives, compounds 3-5 which had OH, CH₂OH and COOH substituent groups on the anthraquinone skeletons, respectively, showed potent growth inhibitory activities against both types of cancer cells (IC₅₀ values: 5.7 ± 0.9 to 13.0 ± 0.7 μM in the case of HCT 116 cells and 5.2 ± 0.7 to 12.3 ± 0.9 μM in the case of Hep G2 cells). All hydroxynaphthoquinone derivatives isolated in this study exhibited extremely potent growth inhibitory activities against both types of cancer cells (IC₅₀ values: 0.3 ± 0.09 to 0.46 ± 1.0 μM in the case of HCT 116 cells and 0.22 ± 0.03 to 0.59 ± 0.06 μM in the case of Hep G2 cells) as well as shikonin 10 (IC₅₀ values: 0.32 ± 0.02 μM in the case of HCT 116 cells and 0.24 ± 0.03 μM in the case of Hep G2 cells).

Full text of DOI:10.1016/j.ejmech.2007.08.009

Available online at www.sciencedirect.com
European Journal of Medicinal Chemistry 43 (2008) 1206e1215
http://www.elsevier.com/locate/ejmech
Original article
Comparison of the cytotoxic activities of naturally occurring
hydroxyanthraquinones and hydroxynaphthoquinones
Xing-Ri Cui a, Maiko Tsukada a, Nao Suzuki a, Takeshi Shimamura a, Li Gao b,
Jyunichi Koyanagi c, Fusao Komada d, Setsuo Saito a,
*
a Faculty of Pharmaceutical Sciences, Josai University, Keyakidai 1-1, Sakado, Saitama 350-0295, Japan
b College of Pharmacy, Yanbian University, Yanji 133000, China
c Faculty of Pharmaceutical Sciences, Josai International University, Gumyo 1, Togane, Chiba 281-8555, Japan
d Faculty of Pharmaceutical Sciences, Himeji Dokkyo University, Kami Ohono 7-2-1, Himeji, Hyogo 670-8524, Japan
Received 10 May 2007; received in revised form 24 July 2007; accepted 9 August 2007
Available online 14 September 2007
Abstract
Seven hydroxyanthraquinone derivatives, 1e7, were isolated from the root of Rheum palmatum (Polygonaceae). Two propionated anthra-
quinone derivatives, 8 and 9, were synthesized. Four hydroxynaphthoquinone derivatives, 13, 14, 16 and 21, were isolated from the root of
Lithospermum erythrorhizon Sieb. et Zucc. (Boraginaceae) and also three naphthoquinone derivatives, 19, 22 and 23, were isolated from
the root of Macrotomia euchroma (Royle) Pauls. (Boraginaceae). The cytotoxicity of the anthraquinone and naphthoquinone derivatives on
P-gp-underexpressing HCT 116 cells and P-gp-overexpressing Hep G2 cells was examined by MTT assay. Among the anthraquinone deriv-
atives, compounds 3e5 which had OH, CH2OH and COOH substituent groups on the anthraquinone skeletons, respectively, showed potent
growth inhibitory activities against both types of cancer cells (IC50 values: 5.7 ꢀ 0.9 to 13.0 ꢀ 0.7 mM in the case of HCT 116 cells and
5.2 ꢀ 0.7 to 12.3 ꢀ 0.9 mM in the case of Hep G2 cells). All hydroxynaphthoquinone derivatives isolated in this study exhibited extremely
potent growth inhibitory activities against both types of cancer cells (IC50 values: 0.3 ꢀ 0.09 to 0.46 ꢀ 1.0 mM in the case of HCT 116 cells
and 0.22 ꢀ 0.03 to 0.59 ꢀ 0.06 mM in the case of Hep G2 cells) as well as shikonin 10 (IC50 values: 0.32 ꢀ 0.02 mM in the case of HCT 116
cells and 0.24 ꢀ 0.03 mM in the case of Hep G2 cells).
Ó 2007 Elsevier Masson SAS. All rights reserved.
Keywords: Hydroxyanthraquinone; Hydroxynaphthoquinone; Cytotoxic activity; HCT 116 cells; Hep G2 cells; P-Glycoprotein
1. Introduction
with a comparison of the cytotoxic activities between hydrox-
yanthraquinone derivatives isolated from the root of Rhubarb
(Rheum palmatum (Polygonaceae)) and hydroxynaphthoqui-
none derivatives isolated from the roots of Lithospermum eryth-
rorhizon Sieb. et Zucc. and Macrotomia euchroma (Royle)
Pauls. (Boraginaceae) using human colorectal (HCT 116) [7]
and human hepatoma (Hep G2) [8] cancer cell lines.
Several papers have been published on the cytotoxic activi-
ties of hydroxyanthraquinone derivatives using cancer cell lines
such as L1210 [1], HL-60 [1], LNCap [2], PC3 [3] and A431 [3]
and those of hydroxynaphthoquinone derivatives using cancer
cell lines such as Sarcoma 180 [4], HL-60 [5] and A431 [6].
However, there are no published studies involving a comparison
of the cytotoxic activities between hydroxyanthraquinone and
hydroxynaphthoquinone derivatives. The present study deals
2. Chemical results and discussion
Hydroxyanthraquinone derivatives 1e7 (Fig. 1) were iso-
lated according to the extraction and isolation process (shown
in Fig. 2) which was a modification of the method reported by
Zhou et al. [3]. Dried Rhubarb (1.0 kg) was extracted with
* Corresponding author. Tel.: þ81 49 271 7667; fax: þ81 49 271 7984.
E-mail address: saitoset@josai.ac.jp (S. Saito).
0223-5234/$ - see front matter Ó 2007 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2007.08.009

X.-R. Cui et al. / European Journal of Medicinal Chemistry 43 (2008) 1206e1215
1207
O
OH
OH
1 R1 = H, R2 = CH3
2
3
4
5
R1 = OCH3, R2 = CH3
R1 = OH, R2 = CH3
R1 = H, R2 = CH2OH
R1 = H, R2 = COOH
R1
R2
O
HOH2C
HOH2C
HO
HO
O
O
HO
HO
O
OH
O
O
O
9
OH
OH
12
11
13
14
OH
12
11
13
14
9
2
1
4
8
7
6
2
1
4
8
7
15
COOH
10
15
CH2OH
6
10
3
3
5
5
COOH
3'
O
6
10'
9'
'
15
8'
HOH2C
HO
O
O
O
7
OH
HO
OH
Fig. 1. Structures of compounds 1e7.
MeOH (2.0 l ꢁ 2). The extracts were evaporated to about half
their volume and suspended in H2O. The suspension was
extracted with AcOEt to give AcOEt and H2O phases. A res-
idue was obtained after evaporation of the AcOEt layer and
following the addition of acetone gave a precipitate (ppt,
8.9 g) and a supernatant (sup). The ppt was subjected to col-
umn chromatography to give rhein (5, 2.1 g). The sup was
evaporated to give a residue (9.7 g) which was subjected to
column chromatography to provide three separate fractions,
Fr.1eFr.3. Repeated recrystallization of Fr.1 from acetone
gave chrysophanol (1, 2.4 g). Chromatography of the Fr.2
and Fr.3 gave emodin (3, 1.1 g) and aloe-emodin (4, 0.9 g), re-
spectively. The mother liquor of 1 contained chrysophanol 1
and an another component. Separation of the two components
from the mother liquor by preparative high performance liquid
chromatography (HPLC) was unsuccessful because of the low
solubility of the mixture in the solvent used for the preparative
HPLC. So a separation of the acyl derivatives obtained from
the mixture was tried. Although acetyl derivatives were not
isolated because they were insufficiently soluble for the
preparative HPLC, chromatography (column: ODS-4251-D,
10 ꢁ 250 mm; solvent system, H2OeCH3COCH3eAcOH
(225:271:4); flow rate: 1.0 ml/min; column temperature:
40 ꢂC) of the propionylated mixture gave 1,8-di-O-propionyl
physcion (8, 180 mg) and 1,8-di-O-propionyl chrysophanol
(9, 70 mg) (Fig. 3). After alkaline hydrolysis of 8, physcion
2 (90 mg, 83.3%) was obtained.
the basis of a comparison with their spectral data in the liter-
ature [3]. The H2O layer obtained after extraction with AcOEt
(Fig. 2) was extracted with n-BuOH to afford an n-BuOH layer
and an H2O layer. A residue obtained after evaporation of the
former organic layer was subjected to column chromatography
1
to obtain compound 6 (0.4 g). In the H NMR spectrum of 6,
signals due to a b-D-glucopyranosyl moiety were observed and
signals due to the aglycon were superimposable on the corre-
sponding signals due to the aloe-emodin moiety (see Section
5). In the heteronuclear multiple bond connection (HMBC)
spectrum of 6, a correlation between the anomeric proton sig-
nal at d 5.79 and the carbon signal at d 159.4 due to C-8 was
observed (data not shown). Therefore, compound 6 is sug-
gested to be 8-O-b-D-glucopyranosyl-aloe-emodin [9].
Trimethylamine was added to this aqueous layer to render it
alkaline (pH 9.0) and then it was extracted with n-BuOH. Af-
ter acidification with the addition of 1.0 M HCl and washing
with H2O, n-BuOH solution was evaporated to give a residue
(120 g). Successive column chromatography and preparative
HPLC (column: N(CH3)2-4251-N, 10 ꢁ 250 mm; solvent sys-
tem: CH3COCH3eH2OeAcOH (271:225:4); flow rate:
1.0 ml/min; column temperature: 40 ꢂC) of a portion (7.0 g)
of the residue gave sennoside A (7, 0.05 g), which was identi-
fied by comparison of its spectral data in the literature [10,11].
Hydroxynaphthoquinone derivatives were isolated from
both the roots of L. erythrorhizon Sieb. et Zucc. and M.
euchroma (Royle) Pauls. (Boraginaceae) called Ko-Shikon and
Nan-Shikon, respectively, in Japan. It has been reported that
hydroxynaphthoquinone derivatives obtained from Ko-Shikon
Structures of compounds 1e5 were identified as chrysopha-
nol, physcion, emodin, aloe-emodin and rhein, respectively, on

1208
X.-R. Cui et al. / European Journal of Medicinal Chemistry 43 (2008) 1206e1215
Dried hubarb (1.0 kg)
extracted with hot MeOH (2.0 L x 2)
evaporated to half volume
suspended in H2O (1.0 L)
extracted with AcOEt (1.0 L x 2)
H2O layer
extracted with n-BuOH
AcOEt layer
evaporated
residue
added with acetone
H2O layer
n-BuOH layer
evaporated
added with
1.0 M NaOH
sup
column
ppt (8.9 g)
chromatog.4)
extracted
with n-BuOH
evaporated
residue (9.7 g)
column chromatog.1)
6 (0.4 g)
H2O layer
n-BuOH layer
added
5 (2.1 g)
with 1.0 M HCl
Fr.1
Fr.2
Fr.3
extracted
with n-BuOH
recrystalli-
zation from
acetone
column
evaporated
chromatog.2)
residue (120 g)
3 (1.1 g)
4 (0.9 g)
mother
liquor
1 (2.4 g)
a portion (7.0 g)
of the residue
evaporated
residue (260 mg)
column
chromatog.5)
propionation
depropionation
prep. HPLC3)
prep. HPLC6)
7 (0.05 g)
2 (0.09 g)
Fig. 2. The isolation process of hydroxyanthraquinone derivatives in Rhubarb. (1) Solvent system: a gradient of 0e5% MeOH in CHCl3. (2) Solvent system: a gra-
dient of 0e5% MeOH in CHCl3. (3) Column: ODS-4251-D, 10 ꢁ 250 mm; solvent system: H2OCH3COCH3eAcOH (225:271:4); flow rate: 1.0 mL/min; column
temperature: 40 ꢂC. (4) Solvent system: CHCl3eMeOHeH2O (65:35:10, upper layer). (5) Solvent system: AcOEtepropanoleH2OeAcOH (40:10:30:0.5, upper
layer). (6) Column: N(CH3)2-4251-N, 10 ꢁ 250 mm; solvent system: CH3COCH3eH2OeAcOH (271:225:4); flow rate: 1.0 mL/min; column temperature: 40 ꢂC.
are derivatives of shikonin 10 (absolute configuration at C-11;
S ) and those obtained from Nan-Shikon are derivatives of
alkannin 11 (absolute configuration at C-11; R), an enantiomer
of shikonin 10 [12e15].
isolation process briefly described in Fig. 5. The extraction
of dried Ko-Shikon (1.0 kg) with toluene (4.0 l ꢁ 2), followed
by filtration and evaporation provided a residue (8.7 g) which
was subjected to column chromatography to give a red colored
fraction (1.9 g) and a blue colored fraction (3.8 g). The prepar-
ative HPLC (column: C18 CAPCELPAC, 10 ꢁ 250 mm; flow
rate: 1.0 ml/min; solvent system: 10% H2Oeacetone; column
temperature: 40 ꢂC) of the red colored fraction gave com-
pounds 13 (270 mg), 14 (410 mg), 16 (270 mg) and 21
(250 mg). On the basis of a comparison with their spectral
data in the literature [17], compounds 13, 14 and 16 were iden-
tified as isobutylshikonin, 30-hydroxyisovalerylshikonin and
acetylshikonin, respectively. Compound 21 showed a parent
peak at m/z 316 [M]þ in the EI mass spectrum. In the
Morimoto et al. [16] reported the isolation of the hydroxy-
naphthoquinone derivatives, b,b-dimethylacrylshikonin 12,
isobutylshikonin 13 and b-hydroxyisovalerylshikonin 14
from Ko-Shikon, and that of teracrylshikonin 15 from Nan-
Shikon (Fig. 4). In addition, Kyogoku et al. [17] reported
the isolation of other hydroxynaphthoquinones; shikonin 10,
acetylshikonin 16, isovalerylshikonin 17 and deoxyshikonin
18 from Ko-Shikon, and acetylalkannin 19 and isovalerylal-
kannin 20 from Nan-Shikon. In this study, isolation of the
hydroxynaphthoquinone derivatives was performed using the

X.-R. Cui et al. / European Journal of Medicinal Chemistry 43 (2008) 1206e1215
1209
OH
O
OH
OH
OH
O
+
H3CO
CH3
CH3
O
1
O
2
Mixture (260 mg) contained in mother liquid of 1
(CH3CH2CO)2O/pyridine
preparaitve HPLC*
COCH2CH3
H3CH2COC
H3CH2COC
COCH2CH3
O
O
O
O
O
O
+
H3CO
CH3
CH3
O
O
9 (70 mg)
8 (180 mg)
1.0 M KOH in MeOH/H2O (1:1)
2 (90 mg, 83.3%)
Fig. 3. The propionation of a mixture of 1 and 2, and depropionation of 8. *Column: ODS-4251-D, 10 ꢁ 250 mm; solvent system: H2OeCH3COCH3eAcOH
(225:271:4); flow rate: 1.0 ml/min; column temperature: 40 ꢂC.
1H NMR spectrum of 21, the proton signals on the shikonin
skeleton were superimposable on those of acetylshikonin 16.
Although the methyl proton signal of an ethyl group substituted
at O-11 was observed at d 1.23 as a triplet (J ¼ 7.0 Hz), each
ethylene proton of the ethyl group was observed at d 3.48
and 3.46 as a double quartet (J ¼ 13.7 and 7.0 Hz), which sug-
gests that the rotation of the bond between O-11 and C-10 is hin-
dered. However, these spectral data suggest that compound 21
is ethylshikonin, which is a newly isolated naturally occurring
hydroxynaphthoquinone. Dried Nan-Shikon (1.0 kg) was also
extracted with toluene (4.0 l ꢁ 2) and then evaporated to dry-
ness to give a residue (25.1 g) which was subjected to column
chromatography to provide a red colored fraction (2.5 g) and
a blue colored fraction (18.9 g). The preparative HPLC of the
red colored fraction under the same conditions as those in the
case of Ko-Shikon gave compounds 19 (660 mg), 22
(340 mg) and 23 (400 mg). Compound 19 had the same parent
peak at m/z 330 [M]þ as that of acetylshikonin 16. The 1H NMR
spectrum of 19 had the same spectrum as that of 16, which sug-
gests that 19 is acetylalkannin. Compound 22 exhibited a parent
peak at m/z 358 [M]þ in the EI mass spectrum. In the 1H NMR
spectrum of 22, all proton signals were coincided with those of
isobutylshikonin 13 (see Section 5), which suggests that 22 is
isobutylalkannin. Compound 22 is the first hydroxynaphthoqui-
none isolated from Nan-Shikon, although it has been isolated
from Alkanna tinctoria (Boraginaceae) [18]. Although com-
pound 23 showed a single spot on thin-layer chromatography
(TLC) and a single peak on HPLC (column: C18 CAPCEL-
PAC, 10 ꢁ 250 mm; flow rate: 1.0 ml/min; solvent system:
10% H2Oeacetone; column temperature: 40 ꢂC), the 1H
NMR spectrum showed that it contained some impurities
(7e10% from the ratio of the signal peak heights). However,
the structure of compound 23 was proved by the mass spectrum
and 1H NMR, 13C NMR and HMBC spectra. The EI mass spec-
trum of 23 had a parent peak at m/z 372 [M]þ. In the 1H NMR
spectrum of 23, all proton signals on the alkannin skeleton were
superimposable on those of 22 and furthermore signals of 20-H
(d 2.15, d, J ¼ 1.2 Hz), 30-H (d 2.18, m) and 40-CH3 (d 0.98, dd,
J ¼ 6.7 and 2.1 Hz) due to the 20,30-epoxybutanoyl group were
observed. The substituted position of the group was confirmed

1210
X.-R. Cui et al. / European Journal of Medicinal Chemistry 43 (2008) 1206e1215
OH O
OH
O
1
8
9
16
14
7
6
2
12
11
4
10
3
15
5
13
OH
O
OR
OH
O
OR
R= H shikonin
R = H alkannin
10
12
13
11
15
CH3
CH3
CH3
CH3
R
=
COCH2
COCH
COHC
C
R
R
=
C
C
CH3
CH3
CH3
=
R
R
COCH3
19
20
=
=
CH3
CH3
COCH2
COCH3
14
16
R
R
R
=
=
C
CH3
CH3
COCH2HC
OH
CH3
22 R =
23 R =
COHC
CO
CH3
CH3
CH3
COCH2HC
CH2CH3
OH
17
21
=
=
H
H
C
C
CH3
R
O
OH
O
O
O
OH
O
OH
24
18
Fig. 4. Structures of compounds 10e24.
by the HMBC spectrum as shown in Fig. 6; a correlation
between the H-11 at d 6.05 and carbonyl carbon (C-10) at
d 171.8 due to the carbonyl group of the 20,30-epoxybutanoyl
group was observed.
Shikonin 1 and alkannin 2 were not detected in the toluene
extracts by TLC and analytical HPLC in the present study.
of cell growth and these are listed in Table 1. In the case of
HCT 116 cells, compounds 2e5 showed potent cytotoxicity
(IC50 values: 5.7 ꢀ 0.9 to 19.0 ꢀ 1.2 mM) among non-
glycosidic derivatives, except for compound 1 (IC50 value:
47.4 ꢀ 18.1 mM) which had only weak activity, which suggests
that the presence of substituents such as OCH3, OH, CH2OH
and COOH on the hydroxynaphthoquinone skeleton is required
for enhancement of the inhibition of HCT 116 cells. 1,8-O-Di-
propionated compounds 8 and 9 also exhibited potent cytotoxic
effects (IC50 values: 9.2 ꢀ 0.7 and 14.8 ꢀ 0.5 mM, respec-
tively). On the other hand, both glycosides 6 and 7 showed
weak activities (IC50 values: 75.0 ꢀ 11.5 and 68.2 ꢀ 4.6 mM, re-
spectively). In the case of Hep G2 cells, while compounds 3e5
exhibited almost the same activities (IC50 values: 5.2 ꢀ 0.7,
10.0 ꢀ 0.8 and 12.3 ꢀ 0.9 mM, respectively) as those in the
HCT 116 cells, activity of 1 was reduced (IC50 values: from
47.4 ꢀ 18.1 to 64.4 ꢀ 17.0 mM), as was those of 2 (IC50 values:
from 19.0 ꢀ 1.2 to 77.3 ꢀ 4.5 mM), 6 (IC50 values: from
75.0 ꢀ 11.5 to >100 mM), 8 (IC50 values: from 9.2 þ 0.7 to
>100 mM) and 9 (IC50 values: from 14.8 þ 0.5 to >100 mM).
3. Pharmacological results and discussion
In this study, human colorectal (HCT 116) [7] and human
hepatoma (Hep G2) [8] cancer cell lines were used. It is
known that the former cell lines scarcely express MDR 1
(P-glycoprotein, P-gp) [7] and the latter cell lines overexpress
it [8]. P-gp acts as an efflux pump to remove antitumor agents,
Ca2þ antagonists, cyclosporine, digoxin and other compounds,
from cells [19].
First, the cytotoxic effects of the isolated hydroxyanthraqui-
none derivatives (1e7) and synthetic propionylnaphthoquinone
derivatives (8 and 9) were tested by MTT assay [20], and the
IC50 values were calculated based on the percentage inhibition

X.-R. Cui et al. / European Journal of Medicinal Chemistry 43 (2008) 1206e1215
Drided Nan-Shikon2) (1.0 kg)
1211
Drided Ko-Shikon1) (1.0 kg)
extracted with toluene (4.0 L x 2)
evaporated the extracts
residure (8.7 g)
residure (25.1 g)
SiO2 gel column chromatpgraphy
(a gradient of 0-30% acetone in toluene)
red colored fraction blue colored fraction3)
red colored fraction blue colored fraction3)
(1.9 g) (3.8 g)
(2.5 g)
(18.9 g)
preparative HPLC4)
preparative HPLC
compounds 13 (270 mg),
14 (410 mg), 16 (270 mg)
and 21 (250 mg)
compounds 19 (660 mg),
22 (340 mg) and 23 (400 mg)
Fig. 5. The isolation process of hydroxynaphthoquinone derivatives in Ko-Shikon and Nan-Shikon. (1) Ko-Shikon: Lithospermum erythrorhizon Sieb. et zucc. (2)
Nan-Shikon: Macrotomia euchroma (Royle) Pauls. (3) These blue colored fractions gave obscure spots on TLC and separation of compounds was impossible even
by HPLC. (4) Preparative HPLC: column: C18-CAPCELPAC, 10 ꢁ 250 mm; flow rate: 1.0 mL/min; column temperature, 40 ꢂC.
This indicates that compounds 3e5 together with 7 may be less
sensitive substrates for P-gp transport compared with com-
pounds 1, 2, 6, 8 and 9.
K 24 had weaker activities (IC50 value: 1.19 ꢀ 0.17 in the
case of HCT 116 cells and 2.34 ꢀ 0.38 mM in the case of Hep
G2 cells) than 10 and all hydroxynaphthoquinone derivatives
isolated in this study against HCT 116 and Hep G2 cells, al-
though it still had a potent effect on both cells, which suggests
that the presence of phenolic hydroxyl groups at the 5 and 8
positions of the naphthoquinone skeleton enhances the activity
against both HCT 116 and Hep G2 cells.
The cytotoxic effects of the isolated hydroxynaphthoqui-
none derivatives (13, 14, 16, 19 and 21e23) as well as those
of shikonin (10) and vitamin K (24) were then compared using
the same cancer cells (HCT 116 and Hep G2 cells) (Table 2).
Shikonin 10 showed markedly potent growth inhibitory activity
(IC50 values: 0.23 ꢀ 0.02 and 0.24 ꢀ 0.03 mM, respectively)
against HCT 116 and Hep G2 cells. All the isolated hydroxy-
naphthoquinone derivatives also exhibited potent inhibition of
both cancer cells (IC50 values: 0.30 ꢀ 0.09 to 0.45 ꢀ 0.10 mM
in the case of HCT 116 cells; 0.22 ꢀ 0.03 to 0.59 ꢀ 0.06 mM
in the case of Hep G2 cells), which suggests that the presence
of a substituent at the O-11 position of the naphthoquinone
skeleton is not essential for their activity. However, vitamin
4. Conclusions
Hydroxyanthraquinone derivatives 1, 3, 4, 5 and 7 were
isolated from R. palmatum. Physcion 2 was purified by
Table 1
Cytotoxic effects of anthraquinone derivatives 1e9 on HCT 116 and Hep G2
cells
Compound
IC50 (mM)a
OH
O
HCT 116
Hep G2
1
2
3
4
5
6
7
8
9
47.4 ꢀ 18.1
19.0 ꢀ 1.2
5.7 ꢀ 0.9
64.4 ꢀ 17.0
77.3 ꢀ 4.5*
5.2 ꢀ 0.7
10.0 ꢀ 0.8**
12.3 ꢀ 0.9
>100b
5
H
H
H
4
6
3
2
15
CH3
10
9
8.7 ꢀ 0.8
H
13.0 ꢀ 0.7
75.0 ꢀ 11.5
68.2 ꢀ 4.6
9.2 ꢀ 0.7
7
12
14
8
1
16
51.6 ꢀ 8.0
H
>100b*
CH3
11
13
H
>100b*
H
H
14.8 ꢀ 0.5
4'
CH3
O
*P < 0.01 and **P < 0.05, the significant difference between the activity on
HCT 116 cells and that on Hep G2 cells. Each experiment was performed
in duplicate wells, and drug treatments were performed separately three times.
O
OH
3'
2'
1'
H
a
O
IC50 values (mean ꢀ S.D.) are the concentrations at which 50% of the cells
O
are inhibited from growing. S.D., standard deviation.
b
IC50 values more than 100 mM were indicated as >100.
Fig. 6. 1He13C Long-Range Correlations observed for compound 23.

1212
X.-R. Cui et al. / European Journal of Medicinal Chemistry 43 (2008) 1206e1215
Table 2
Cytotoxic effects of anthraquinone derivatives 13, 14, 16, 19 and 21e23 on
HCT 116 and Hep G2 cells
derivatives are more effective growth inhibitors of both HCT
116 and Hep G2 cells. The activities of these compounds
(IC50 values: 0.30 ꢀ 0.09 to 0.46 ꢀ 0.10 mM in the case of
HCT 116 cells and 0.22 ꢀ 0.03 to 0.42 ꢀ 0.04 mM in the case
Hep G2 cells) were almost identical to those of shikonin 10
(IC50 value: 0.23 ꢀ 0.02 mM in the case of HCT 116 cells and
0.24 ꢀ 0.03 mM in the case Hep G2 cells) which has no substit-
uent at the O-11 position. This indicates that the presence of sub-
stituent at the O-11 position is not essential for the inhibition of
the growth of both HCT 116 and Hep G2 cells. However, the
presence of phenolic OH groups on the naphthoquinone skele-
ton may be essential to enhance the activity because the activity
of vitamin K 24 which has no phenolic OH group was weaker
than that of hydroxynaphthoquinone derivatives. Compounds
13 and 16 are derivatives of shikonin (10, S-configuration at
C-11) and compounds 19 and 22 are derivatives of alkannin
(11, R-configuration at C-11). The activities of isobutylshikonin
13 and isobutylalkanin 22 (isomer of 13) and those of acetylshi-
konin 16 and acetylalkannin 19 (isomer of 16) were, however,
almost identical against both HCT 116 and Hep G2 cells, which
suggests that the difference in the stereochemistry at C-11 has no
influence on the activity.
In this study, we found that naturally occurring hydroxynaph-
thoquinone derivatives are more cytotoxic than naturally occur-
ring hydroxyanthraquinone derivatives. Furthermore, all
hydroxynaphthoquinone derivatives 10, 13, 14, 16, 19 and
21e23 were markedly cytotoxic activities to P-gp overex-
pressed cancer cells. Generally, plane structural molecule and
cationic molecule indicated the substrate for P-gp. Both hydrox-
yanthraquinone and hydroxynaphthoquinone derivatives in this
study were plane structural molecules. The cytotoxic activities
against HCT 116 of almost all hydroxyanthraquinone deriva-
tives (1, 2, 4, 6, 8, and 9) were greater than that against Hep
G2 (P-gp overexpressing cell line). This phenomenon can be ex-
plained by P-gp mediated transport [21,22]. However, Ca antag-
onists, itraconazole, spironolactone, quinidine, propanolol,
lidocaine, cyclosporine A, digoxin, erythromycin, doxorubicin,
and vinblastine were well-documented substrates for P-gP. Mo-
lecular weights, physicochemical properties, lipophilicities and
pharmacological effects of these drugs were completely differ-
ent. Hence properties of the substrate for P-gp have not ever be-
come clear. The cytotoxic activities against both cell lines of all
hydroxynaphthoquinone derivatives investigated in this study
were approximately same levels. Therefore, this demonstrated
that these hydroxynaphthoquinone derivatives may not be trans-
ported by P-gp. That these hydroxynaphthoquinone derivatives
were markedly cytotoxic to P-gp overexpressing cancer cells
may give a new approach to overcome the obstacle posed by
multidrug resistance in cancer.
Compound
IC50 (mM)a
HCT 116
Hep G2
10
0.23 ꢀ 0.02
0.34 ꢀ 0.05
0.37 ꢀ 0.06
0.45 ꢀ 0.10
0.45 ꢀ 0.12
0.34 ꢀ 0.09
0.30 ꢀ 0.09
0.39 ꢀ 0.08
1.19 ꢀ 0.17
0.24 ꢀ 0.03
0.22 ꢀ 0.03*
0.59 ꢀ 0.06
0.38 ꢀ 0.07
0.42 ꢀ 0.04
0.39 ꢀ 0.07
0.42 ꢀ 0.03
0.34 ꢀ 0.04
2.34 ꢀ 0.38
13b
14b
16b
21b
19c
22c
23c
24
*P < 0.01, the significant difference between the activity on HCT 116 cells
and that on Hep G2 cells. Each experiment was performed in duplicate wells,
and drug treatments were performed separately three times.
a
IC50 values (mean ꢀ SD) are the concentrations at which 50% of the cells
are inhibited from growing. S.D., standard deviation.
b
Derivatives isolated from Ko-Shikon.
Derivatives isolated from Nan-Shikon.
c
propionation of the mixture followed by depropionation. 8-
O-b-D-Glucopyranosyl-aloe-emodin 6 was purified by acetyla-
tion of the mixture followed by deacetylation. Furthermore, the
new hydroxynaphthoquinone derivative 21 together with
known derivatives 13, 14 and 16 were isolated from Ko-Shikon
(L. erythrorhizon Sieb. et Zucc.) and a new hydroxynaphthoqui-
none derivative 23 together with known derivatives 19 and 22
were isolated from Nan-Shikon (M. euchroma (Royle) Pauls.).
The cytotoxic activities of the hydroxyanthraquinone and hy-
droxynaphthoquinone derivatives were evaluated using HCT
116 and Hep G2 cancer cell lines. Glycosidic derivatives 6
and 7 had weak activity against both cells. Among hydroxyan-
thraquinone derivatives1e5, while compound 1 had weak activ-
ity, compounds 2e5, which had OCH3, OH, CH2OH and COOH
substituted groups on the hydroxyanthraquinone skeleton,
exhibited potent growth inhibitory activities (IC50 values:
5.7 ꢀ 0.9 to 19.0 ꢀ 1.2 mM) against HCT 116 cells. The activity
of
2
was reduced (IC50 values: from 19.0 ꢀ 1.2 to
77.3 þ 4.5 mM) against Hep G2 cells, whereas the activities of
3e5 on Hep G2 cells were almost similar to that on HCT
116 cells. These results suggest that the presence of polar sub-
stituents, OH, CH2OH and COOH, on the hydroxynaphthoqui-
none skeleton may prevent the removal of compounds 3e5 by
the P-gp in the Hep G2 cells. A similar reduction in activity
was observed in the case of compounds 8 and 9 which
contained weak polar and non-polar OCH3 and CH3 groups,
respectively. Accordingly, the activities of 8-O-propionated de-
rivatives 8 and 9 had potent activities (IC50 values: 9.2 ꢀ 0.7 and
14.8 ꢀ 0.5 mM, respectively) against HCT 116 cells, much more
so than 2 and 1 (IC50 value: 19.0 ꢀ 1.2 and 47.4 ꢀ 18.1 mM), re-
spectively. However, these activities were reduced in the case of
Hep G2 cells (IC50 value: >100 mM).
5. Experimental
All hydroxynaphthoquinone derivatives 13, 14, 16, 19 and
21e23 isolated in this study showed markedly potent activities
against both HCT 116 and Hep G2 cells compared with the
activities of the hydroxyanthraquinone derivatives (compare Ta-
bles 1 and 2). This suggests that hydroxynaphthoquinone
5.1. General procedures
Rhubarb, Ko-Shikon and Nan-Shikon, were purchased from
Tochimoto Amamido Co. Ltd., Japan. Sennoside A and shiko-
nin were purchased from Wako Pure Chemical Industries,

X.-R. Cui et al. / European Journal of Medicinal Chemistry 43 (2008) 1206e1215
1213
Ltd., Japan. Other chemicals and solvents were of reagent
grade and obtained from commercial suppliers. Melting points
(m.p.) were determined using a Yanagimoto micromelting
point apparatus and are uncorrected. Kieselgel 60 F254 (E.
Merck) was used for thin-layer chromatography (TLC). Spots
were detected by spraying plates with 1:9 Ce(SO4)2e10%
H2SO4 reagent followed by heating the plate at 250 ꢂC for
4e5 min. Column chromatography was carried out using
Silica gel 60 (E. Merck) and then the eluates were monitored
using TLC. An SSC-6300 HPLC instrument (Senshu Co. Ltd.)
was employed for analytical HPLC using columns ODS-4251-
D and N(CH3)2-4251-N (Senshu Co. Ltd.) and C18 CAPCEL-
PAC (Shiseido Co. Ltd.) attached to an SSC autoinjector 6310
J ¼ 8.2 Hz, H-5), 7.63 (1H, dd, J ¼ 8.2, 0.9 Hz), 7.59 (1H, d,
J ¼ 1.5 Hz, H-4), 7.31 (1H, dd, J ¼ 8.2, 0.9 Hz, H-7), 7.21
(1H, d, J ¼ 1.5 Hz, H-2), 5.06 (1H, s, CH2OH ), 4.60 (2H, s,
CH2OH); 13C NMR (DMSO-d6) d 191.5 (C-9), 181.2
(C-10), 161.6 (C-8), 161.3 (C-1), 153.7 (C-3), 137.3 (C-14),
133.1 (C-11), 132.9 (C-6), 124.3 (C-5), 120.6 (C-12), 119.3
(C-13), 117.1 (C-4), 115.6 (C-7), 114.2 (C-2), 62.1 (C-15).
1
5.1.1.5. Rhein (5). EIMS: m/z 284 [M]þ; m.p. <300 ꢂC; H
NMR (pyridine-d5) d 11.87 (2H, s, 1- and 8-OH), 8.90 (1H,
t, J ¼ 1.5 Hz, H-4), 8.35 (1H, d, J ¼ 1.5 Hz, H-2), 7.94 (1H,
dd, J ¼ 7.3, 1.2 Hz, H-5), 7.67 (1H, dd, J ¼ 8.5, 7.3 Hz,
H-6), 7.42 (1H, dd, J ¼ 8.5, 1.2 Hz, H-7); 13C NMR (pyri-
dine-d5) d 191.5 (C-9), 181.2 (C-10), 161.6 (C-8), 161.3
(C-1), 153.7 (C-3), 137.3 (C-14), 133.1 (C-11), 132.9 (C-6),
124.3 (C-5), 120.6 (C-12), 119.3 (C-13), 117.1 (C-4), 115.6
(C-7), 114.2 (C-2), 62.1 (C-15).
1
and an SSC fraction collector 6320 for preparative HPLC. H
and 13C NMR experiments were carried out at 500 and
125 MHz, respectively, as well as 1He1H and 1He13C
COSY and HMBC spectra were obtained with a JEOL JNM-
A500 FTNMR spectrometer. Tetramethylsilane was used as
an internal standard. Chemical shifts are given in parts per mil-
5.1.1.6. 8-O-b-D-glucopyranosyl-aloe-emodin (6). FABMS:
m/z 433 [M þ 1]þ; m.p. 232e234 ꢂC (after recrystallization
1
lion. Multiplicities of the H NMR signals are indicated as s
1
(singlet), d (doublet), dd (doublet of doublets), ddd (doublet
of doublet of doublets), t (triplet), q (quartet) and m (multi-
plet). Electron impact mass spectra (EIMS) and fast-atom-
bombardment mass spectra (FABMS) were recorded on
a JEOL JMS-DX 300 mass spectrometer.
from methanol); H NMR (pyridine-d5) d 8.08 (1H, s, H-4),
8.03 (1H, dd, J ¼ 7.6, 0.9 Hz, H-5), 7.99 (1H, dd, J ¼ 8.2,
0.9 Hz, H-7), 7.61 (1H, dd, J ¼ 8.2, 7.6 Hz, H-6), 4.97 (2H, s,
3-CH2OH), 5.79 (1H, d, J ¼ 7.6 Hz, H-10), 4.61 (1H, dd,
J ¼ 12.2, 2.4 Hz, H-60a), 4.55 (1H, dd, J ¼ 8.9, 7.6 Hz, H-20),
4.41 (1H, t, J ¼ 8.9 Hz, H-30), 4.41 (1H, dd, J ¼ 12.2, 5.2 Hz,
H-60b), 4.36 (1H, t, dd, J ¼ 9.2, 8.9 Hz, H-4), 4.23 (1H, ddd,
J ¼ 9.2, 5.2, 2.4 Hz, H-50); 13C NMR (pyridine-d5) d 188.6 (C-
9), 182.6 (C-10), 163.2 (C-3), 159.4 (C-8), 153.2 (C-1), 135.9
(C-6), 133.2 (C-13), 123.9 (C-12), 122.9 (C-7), 121.7 (C-11),
121.5 (C-2), 121.3 (C-5), 116.8 (C-4), 116.2 (C-14), 63.3 (C-
15), 102.8(C-10), 79.3(C-50), 78.3(C-30), 74.9(C-20), 62.4(C-60).
5.1.1. Chemistry
5.1.1.1. Isolation of hydroxyanthraquinone. Hydroxyanthra-
quinone derivatives 1 and 3e7 were isolated according to the
procedure described in this text and the scheme shown in Fig. 2.
5.1.1.2. Chrysophanol (1). EIMS: m/z 254 [M]þ; m.p. 194e
196 ꢂC (after recrystallization from acetone); 1H NMR
(CDCl3) d 12.15 and 12.04 (each 1H, s, 1- and 8-OH), 7.85
(1H, dd, J ¼ 7.6, 1.2 Hz, H-5), 7.69 (1H, dd, J ¼ 8.6, 7.6 Hz,
H-6), 7.68 (1H, br s, H-4), 7.31 (1H, dd, J ¼ 8.6, 1.2 Hz,
H-7), 7.13 (1H, br s, H-2), 2.49 (3H, s, CH3); 13C NMR
(CDCl3) d 192.5 (C-9), 190.7 (C-10), 162.7 (C-8), 162.4
(C-1), 149.3 (C-3), 149.3 (C-3), 133.6 (C-11), 124.3 (C-14),
121.2 (C-6), 119.9 (C-4), 115.8 (C-5), 113.7 (C-12), 113.6
(C-7), 108.2 (C-13), 106.7 (C-2), 22.1 (C-15).
5.1.1.7. Sennoside A (7). FABMS: m/z 885 [M þ Na]þ; 13C
NMR (pyridine-d5) d 186.5 (C-9 and 90), 168.8 (C-15 and
150), 159.1 (C-1 and 10), 157.7 (C-8 and 80), 142.3 (C-13
and 130), 138.2 (C-12 and 120), 135.6 (C-3 and 30), 135.2
(C-6 and 60), 123.9 (C-4 and 40), 122.4 (C-11 and 110),
121.5 (C-14 and 140), 120.5 (C-5 and 50), 118.2 (C-2 and
20), 116.6 (C-7 and 70), 54.0 (C-10 and 100), 103.7 (C-100 an
1000), 77.5 (C-500 and 5000), 75.2 (C-300 and 3000), 73.5 (C-200 and
2000), 69.7 (C-400 and 4000), 60.7 (C-600 and 6000). The 13C NMR
signals agreed with those reported by Nakajima et al. [11].
The retention time (40 min) of this compound agreed with
that of authentic sennoside A on analytical HPLC (column:
N(CH3)2-4251-N, 10 ꢁ 250 mm; solvent system: THFeH2Oe
AcOH (160:60:7); flow rate: 1.0 ml/min; column temperature:
40 ꢂC).
5.1.1.3. Emodin (3). EIMS: m/z 270 [M]þ; m.p. 255e256 ꢂC
1
(after recrystallization from acetone); H NMR (pyridine-d5)
d 12.48 (2H, br s, 1- and 8-OH), 7.72 (1H, d, J ¼ 2.4 Hz,
H-4), 7.70 (1H, dd, J ¼ 0.9, 0.6 Hz, H-5), 7.14 (1H, dd,
J ¼ 0.9, 0.6 Hz, H-7), 7.02 (1H, d, J ¼ 2.4 Hz, H-2), 5.23
(1H, s, 3-OH), 2.24 (3H, s, CH3); 13C NMR (CDCl3)
d 182.2 (C-9), 179.1 (C-10), 167.7 (C-1), 165.2 (C-1), 165.9
(C-3), 162.6 (C-8), 148.3 (C-6), 136.1 (C-14), 136.0 (C-11),
133.7 (C-5), 124.5 (C-12), 121.1 (C-7), 114.1 (C-13), 110.2
(C-2), 108.8 (C-4), 21.7 (C-15).
5.1.1.8. 1,8-O-Dipropionylphyscion (8) and 1,8-O-dipropionyl-
chrysophanol (9). A solution of a mixture (260 mg) of 1 and 2
obtained from the mother liquor of 1 in pyridine (5 ml) was
mixed with propionyl anhydride (5 ml) and then stirred for
24 h at room temperature. The reaction mixture was poured
into ice-water (20 ml) and extracted with AcOEt (20 ml ꢁ 2).
The extracts were successively washed with 5% HCl, 10%
NaHCO3 and H2O, dried over absolute MgSO4 and filtered.
5.1.1.4. Aloe-emodin (4). EIMS: m/z 270 [M]þ; m.p. 220e
221 ꢂC (after recrystallization from acetone); 1H NMR
(DMSO-d6) d 11.85 (2H, s, 1- and 8-OH), 7.75 (1H, t,

1214
X.-R. Cui et al. / European Journal of Medicinal Chemistry 43 (2008) 1206e1215
The filtrate was evaporated to give a residue which was
subjected to preparative HPLC (column: ODS-4251-D,
10 ꢁ 250 mm; solvent system: H2OeCH3COCH3eAcOH
(225:271:4); flow rate: 1.0 ml/min; column temperature:
40 ꢂC) to obtain 1,8-O-dipropionylphyscion (8, 180 mg) and
1,8-O-dipropionylchrysophanol (9, 70 mg). Compound 8:
EIMS: m/z 396 [M]þ; 1H NMR (CDCl3) d 7.99 (1H, dd,
J ¼ 1.8, 0.6 Hz, H-5), 7.66 (1H, d, J ¼ 2.7 Hz, H-4), 7.18
(1H, dd, J ¼ 1.8, 0.6 Hz, H-7), 6.86 (1H, d, J ¼ 2.7 Hz, H-2),
3.95 (3H, s, OCH3), 2.74 (2H, q, J ¼ 7.6 Hz, COCH2CH3),
2.73 (2H, q, J ¼ 7.6 Hz, COCH2CH3), 2.48 (3H, s, 3-CH3),
1.34 (6H, t, J ¼ 7.6 Hz, COCH2CH3 ꢁ 2); 13C NMR (CDCl3)
d 182.3 (C-10), 179.5 (C-9), 172.8 and 172.6 (COCH2CH3 ꢁ
2), 163.9 (C-6), 152.4 (C-8), 150.3 (C-1), 145.6 (C-3), 136.0
(C-11), 134.2 (C-14), 130.8 (C-4), 125.8 (C-13), 123.3 (C-2),
119.4 (C-12), 116.4 (C-5), 116.4 (C-7), 56.1 (6-OCH3), 27.7
(COCH2CH3 ꢁ 2), 21.6 (C-15), 8.7 (COCH2CH3 ꢁ 2). Anal.
Calcd for C22H20O7: C, 66.66; H, 5.09. Found: C, 66.53; H,
12.56 (each 1H, s, 5 and 8-OH), 7.18 (2H, s, H-6 and 7),
6.98 (1H, d, J ¼ 0.9 Hz, H-3), 6.02 (1H, ddd, J ¼ 7.3, 4.6,
0.9 Hz, H-11), 5.13 (1H, br t, J ¼ 7.3 Hz, H-13), 2.48 (1H,
m, H-20), 1.69 (3H, s, 15-CH3), 1.59 (3H, s, 16-CH3), 1.22
(3H, d, J ¼ 7.0 Hz, 30-CH3), 1.21 (3H, d, J ¼ 7.0 Hz, 40-
CH3). The 13C NMR signals are listed in Table 3.
5.1.1.12. 30-Hydroxyisovalerylshikonin (14). EIMS: m/z 388
1
(Mþ); H NMR (CDCl3) d 12.60 and 12.40 (each 1H, s, 5-
and 8-OH), 7.20 (2H, s, H-6 and 7), 7.03 (1H, d, J ¼ 1.0 Hz,
H-3), 6.09 (1H, ddd, J ¼ 7.6, 4.6, 1.0 Hz, H-11), 5.12 (1H,
br t, J ¼ 7.6 Hz, H-13), 2.63 (1H, m, H-120a), 2.49 (1H, dt,
J ¼ 15.3, 7.6 Hz, H-2b0), 1.69 (3H, d, J ¼ 0.9 Hz, 15-CH3),
1.59 (3H, d, J ¼ 0.9 Hz, 16-CH3), 1.31 (3H, s, 40-CH3), 1.30
(3H, s, 50-CH3). The 13C NMR signals are listed in Table 3.
5.1.1.13. Acetylshikonin (16) and acetylalkannin (19). EIMS:
1
m/z 330 [M]þ; H NMR (CDCl3) d 12.60 and 12.40 (each
1
4.95. Compound 9: EIMS: m/z 366 [M]þ; H NMR (CDCl3)
1H, s, 5- and 8-OH), 7.18 (2H, s, H-6 and 7), 6.99 (1H, s,
H-3), 6.02 (1H, ddd, J ¼ 7.4, 4.6, 0.9 Hz, H-11), 5.12 (1H,
br t, J ¼ 7.6 Hz, H-13), 2.62 (1H, br dt, J ¼ 15.0, 7.3 Hz,
H-120a), 2.48 (1H, dt, J ¼ 15.0, 7.3 Hz, H-2b0), 2.14 (3H, s,
COCH3), 1.69 (3H, s, 15-CH3), 1.57 (3H, s, 16-CH3). The
13C NMR signals are listed in Table 3.
d 8.19 (1H, dd, J ¼ 7.6, 1.2 Hz, H-5), 8.00 (1H, dd, J ¼ 1.8,
0.6 Hz, H-4), 7.73 (1H, t, J ¼ 7.9 Hz, H-6), 7.20 (1H, dd,
J ¼ 1.8, 0.6 Hz, H-2), 2.74 (2H, q, J ¼ 7.6 Hz, COCH2CH3),
2.73 (2H, q, J ¼ 7.6 Hz, COCH2CH3), 2.49 (3H, s, 3-CH3),
1.35 (3H, t, J ¼ 7.6 Hz, COCH2CH3), 1.34 (3H, t, J ¼ 7.6 Hz,
COCH2CH3); 13C NMR (CDCl3) d 182.3 and 180.5 (C-9 an
C-10), 172.8 (COCH2CH3), 172.8 (COCH2CH3), 150.3 (C-8),
150.2 (C-1), 146.2 (C-3), 134.6 (C-6), 134.3 (C-11), 134.2
(C-14), 130.7 (C-12), 130.2 (C-4), 125.8 (C-5), 125.8 (C-7),
125.3 (C-13), 123.4 (C-2), 27.8 (COCH2CH3 ꢁ 2), 21.7
(C-15), 8.8 (COCH2CH3 ꢁ 2). Anal. Calcd for C21H18O6: C,
68.85; H, 4.95. Found: C, 66.52; H, 5.06.
1
5.1.1.14. Ethylshikonin (21). EIMS: m/z 316 [M]þ; H NMR
(CDCl3) d 12.60 and 12.51 (each 1H, s, 5- and 8-OH), 7.26
(2H, s, H-6 and 7), 7.20 (1H, s, H-2), 5.20 (1H, dt, J ¼ 7.6,
1.2 Hz, H-13), 4.66 (1H, ddd, J ¼ 7.0, 4.2, 0.9 Hz, H-11),
4.48 (1H, dq, J ¼ 13.7, 7.0 Hz, H-10a), 4.46 (1H, dq,
J ¼ 13.7, 7.0 Hz, H-10b), 2.52 (1H, m, H-120a), 2.30 (1H, dt,
J ¼ 15.0, 7.0 Hz, H-12b0), 1.70 (3H, s, 15-CH3), 1.60 (3H, s,
5.1.1.9. Physcion (2). A solution of compound 8 (150 mg) in
1.0 M KOH (5 ml, MeOHeH2O ¼ 1:1) was stirred for 7 h at
room temperature. The reaction mixture was poured into H2O
(20 ml) and then extracted with AcOEt (20 ml ꢁ 2). The
organic extracts were successively washed with 5% HCl,
10% NaHCO3 and H2O, then dried over absolute MgSO4 and
filtered. The filtrate was evaporated to give a residue which
was subjected to column chromatography to obtain physcion
Table 3
13C NMR spectral data of hydroxynaphthoquinones 13, 14, 16, 19, 22, 21 and
23 in CDCl3
13 and 22
14
16 and 19
21
23
C-1
C-2
176.8
135.9
148.5
178.3
167.3
132.7a
132.6a
166.7
111.5b
111.8b
69.0
175.4
136.4
147.5
177.0
168.7
133.3a
133.1a
168.1
111.6b
111.8b
69.8
176.7
136.7
148.2
178.2
167.5
132.9a
132.7a
167.0
111.6b
111.8b
69.5
176.2
134.7
151.0
179.9
166.2
132.4a
132.3a
165.7
111.6b
112.0b
65.4
176.7
131.6
148.6
178.0
167.0
136.0
136.0
167.5
112.0
112.0
68.7
1
(2, 90 mg, 83.3%). EIMS: m/z 284 [M]þ; H NMR (CDCl3)
C-3
C-4
d 12.31 and 12.11 (each 1H, s, 1- and 8-OH), 7.62 (1H, d,
J ¼ 1.2 Hz, H-5), 7.36 (1H, d, J ¼ 2.4 Hz, H-4), 7.07 (1H, dd,
J ¼ 1.5, 0.9 Hz, H-7), 6.68 (1H, d, J ¼ 2.4 Hz, H-2), 3.94
(3H, s, OCH3), 2.45 (3H, s, CH3); 13C NMR (CDCl3) d 190.8
(C-9), 182.0 (C-10), 166.5 (C-3), 165.2 (C-1), 162.5 (C-8),
145.4 (C-6), 135.2 (C-14), 133.2 (C-11), 124.5 (C-7), 121.3
(C-5), 113.7 (C-12), 110.2 (C-13), 108.2 (C-4), 106.8 (C-2),
56.1 (OCH3), 22.2 (C-15).
C-5
C-6
C-7
C-8
C-9
C-10
C-11
C-12
C-13
C-14
C-15
C-16
C-10
C-20
C-30
C-40
C-50
32.9
32.9
32.8
34.4
33.0
117.8
131.3
25.7
117.7
131.3
25.8
117.7
131.5
25.8
119.1
132.1
26.5
117.9
136.0
25.7
5.1.1.10. Isolation of hydroxynaphthoquinone derivatives. Hy-
droxynaphthoquinone derivatives were isolated according to
the procedure described in this text and the scheme shown
in Fig. 5.
18.9
17.0
171.7
46.5
17.9
169.8
21.0
17.9
65.3
17.9
175.7
34.0
17.9
171.8
25.1
43.2
15.3
e
e
69.2
e
e
18.8
e
29.2c
29.1c
22.4
e
e
e
5.1.1.11. Isobutylshikonin (13) and isobutylalkannin
1
(22). EIMS: m/z 358 [M]þ; H NMR (CDCl3) d 12.58 and
a,b,c: These values may be interchangeable in each column.

X.-R. Cui et al. / European Journal of Medicinal Chemistry 43 (2008) 1206e1215
1215
16-CH3), 1.19 (3H, t, J ¼ 7.0 Hz, 20-CH3). The 13C NMR
experiments involving a control (DMSO only) and the drug
treatments were performed separately 3e5 times. Data repre-
sent mean ꢀ S.D. values.
signals are listed in Table 3.
1
5.1.1.15. 20,30-Epoxyalkannin (23). EIMS: m/z 396 [M]þ; H
NMR (CDCl3) d 12.60 and 12.58 (each 1H, s, 5- and 8-OH),
7.20 (2H, s, H-6 and 7), 6.99 (1H, d, J ¼ 0.9 Hz, H-2), 6.05
(1H, ddd, J ¼ 7.6, 4.6, 0.9 Hz, H-11), 5.14 (1H, m, H-13),
2.62 (1H, m, H-120a), 2.47 (1H, m H-2b0), 2.18 (1H, m,
H-30), 1.69 (3H, s, 15-CH3), 1.59 (3H, s, 16-CH3), 0.98 (3H,
dd, J ¼ 6.7, 2.1 Hz, 4’-CH3). The 13C NMR signals are listed
in Table 3.
References
[1] M. Koyama, K. Takahashi, T.-C. Chou, Z. Darzynkiewicz, J. Kapuscinski,
T.R. Kelly, K.A. Watanabe, J. Med. Chem. 32 (1989) 1594e1599.
[2] T.-L. Cha, L. Qiu, C.-T. Chen, Y. Wen, M.-C. Hung, Cancer Res. 65
(2005) 2287e2295.
[3] X. Zhou, B. Song, L. Jin, D. Hu, C. Diao, G. Xu, Z. Zou, S. Yang, Bio-
org. Med. Chem. Lett. 16 (2006) 563e568.
[4] U. Sankawa, Y. Ebizuka, T. Miyazaki, Y. Isomura, H. Ohtsuka,
S. Shibata, M. Inomata, F. Fukuoka, Chem. Pharm. Bull. 25 (1977)
2392e2395.
5.1.2. Cytotoxic activities
5.1.2.1. Cell lines and culture. The human colorectal carci-
noma cell line (HCT 116, ATCC No. CCL-247) and human
hepatoma cell line (Hep G2 No. RCB0459) were purchased
from Dainippon Pharmaceutical Co. Ltd. (Osaka, Japan) and
RIKEN Cell Bank (Tsukuba, Japan), respectively. Dulbecco’s
modified Eagle’s medium (DMEM), McCoy’s 5A medium, fe-
tal bovine serum (FBS) and penicillinestreptomycin mixture
(100 U/ml penicillin and 100 mg/ml streptomycin) were pur-
chased from Wako Pure Chemical Industries, Ltd. (Osaka,
Japan), Sigma (MO, USA), Biosource International (CA,
USA) and Bio Whittaker (ND, USA), respectively. The HCT
116 cells were maintained in McCoy’s 5A medium and Hep
G2 cells were cultured in DMEM. Each medium was supple-
mented with 10% FBS and a penicillinestreptomycin mixture
at 37 ꢂC in a humidified atmosphere containing 5% CO2 in air.
[5] S. Hashimoto, M. Xu, Y. Masuda, T. Aiuchi, S. Nakajo, J. Cao,
M. Miyakoshi, Y. Ida, K. Nakaya, J. Biochem. 125 (1999) 17e23.
[6] F. Singh, D. Gao, M.G. Lebwohl, H. Wei, Cancer Lett. 200 (2003)
115e121.
[7] M. Shionoya, T. Jimbo, M. Kitagawa, T. Soga, A. Tohgo, Cancer Sci. 94
(2003) 459e466.
[8] G. Lee, M. Piqette-Miller, Can. J. Physiol. Pharmacol. 79 (2001) 876e884.
[9] J. Kinjo, T. Ikeda, K. Watanabe, T. Nohara, Phytochemistry 37 (1994)
1685e1687.
[10] H. Tanaka, R. Murata, A. Yoshida, S. Hayashi, Chem. Pharm. Bull. 30
(1982) 1550e1556.
[11] K. Nakajima, K. Yamauchi, S. Kuwano, J. Pharm. Pharmacol. 37 (1985)
703e706.
[12] M. Tsukada, T. Tanaka, M. Tabata. Abstracts of Papers, 25th Annual
Meeting of the Japanese Society of Pharmacognosy, Fukuoka, Nov.
1978, p. 28.
[13] U. Sankawa, H. Otsuka, Y. Kataoka, Y. Iitaka, A. Hoshi, K. Kuretani,
Chem. Pharm. Bull. 29 (1981) 116e122.
[14] V.P. Papageorgiou, A.N. Assimopoulou, E.A. Couladouros, D. Hepworth,
D.K.C. Nicolaou, Angew. Chem., Int. Ed. 38 (1999) 270e300.
[15] A. Spyros, A.N. Assimopoulou, V.P. Papageorgion, Biomed. Chromatogr.
19 (2005) 498e505.
5.1.2.2. Cytotoxicity test. Aliquots (200 ml) of 5 ꢁ 103 cells/ml
of HCT 116 and Hep G2 cells were seeded in 96-well flat-
bottomed plates (MicrotestÔ Tissue Culture Plate, 96 Well,
Flat Bottom with Low Evaporation Ltd., Falcon, NJ, USA)
and incubated in a medium containing 10% FBS and a penicil-
linestreptomycin mixture at 37 ꢂC in a humidified atmosphere
of 5% CO2 for 24 h. The test drugs were dissolved in dimethyl
sulfoxide (DMSO). The incubation medium was replaced with
each test medium giving a final concentration of 0.1e
500 mmol/l of test compounds and no drug in 2 ml DMSO
over 2 days.
[16] I. Morimoto, T. Kishi, S. Ikegami, Y. Hirata, Tetrahedron Lett. (1965)
4737e4739.
[17] K. Kyogoku, H. Terayama, Y. Tachi, T. Suzuki, M. Komatsu, Shoyaku-
gaku Zasshi 27 (1973) 24e30.
[18] V.P. Papageorgion, A.S. Mellidis, A.N. Assimopoulou, A. Tsarbopoulos,
J. Mass Spectrom. 33 (1998) 89e91.
[19] K. Takara, T. Sakaeda, Y. Tanigawa, K. Nishiguchi, N. Ohmoto,
M. Horinouchi, Eur. J. Pharm. Sci. 16 (2002) 159e165.
[20] T. Mosmann, J. Immunol. Methods 139 (1983) 55e63.
[21] J.W. Smit, B. Weert, A.H. Shinkel, D.K. Meijer, J. Pharmacol. Exp. Ther.
286 (1998) 321e327.
The ability of the drug to inhibit cellular growth was deter-
mined using the MTT assay [20]. The cytotoxic activities of
the test drugs were determined as previously described [23].
Each experiment was performed in duplicate wells, and all
[22] A. Seeling, X.L. Blatter, F. Wohnsland, Int. J. Clin. Pharmacol. Ther. 38
(2000) 111e121.
[23] G.-Z. Jin, H.-J. Quan, J. Koyanagi, K. Takeuchi, Y. Miura, F. Komada,
S. Saito, Cancer Lett. 218 (2005) 15e20.