Structure-cytotoxic activity relationship for the toad poison bufadienolides

Article abstract of DOI:10.1016/S0968-0896(98)00067-4

The toad poison bufadienolides including natural and derivatized compounds were tested for their cytotoxic effects on primary liver carcinoma cells PLC/PRF/5 and their structure-cytotoxic activity relationships were studied. For this study, a ligand-binding model was developed by using a pharmacophore mapping program, Distance Comparisons (DISCO). The structural features that are common to the 3D structures of active bufadienolides were identified to provide approach to a 3D QSAR method by using Comparative Molecular Field Analysis (CoMFA) study and to correlate the steric and electrostatic fields of the molecules to their activities. A valuable model which enables prediction of their activities was obtained from the CoMFA analysis, which may be employed for the drug designs of new bufadienolide analogues. Copyright (C) 1998 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(98)00067-4

BMC 960p
BIOORGANIC &
MEDICINAL
CHEMISTRY
Bioorganic & Medicinal Chemistry 6 (1998) 1103±1115
Structure±Cytotoxic Activity Relationship for the Toad Poison
Bufadienolides
Yoshiaki Kamano,a,* Ayano Kotake,a Hirofumi Hashima,a Masuo Inoue,a
Hiroshi Morita,b Koichi Takeya,b Hideji Itokawa,b Nobuyo Nandachi,c
Toshiaki Segawa,d Ayako Yukita,d Kyoko Saitou,e Mariko Katsuyamae
and George R. Pettitf
aFaculty of Science, Kanagawa University, 2946 Tsuchiya, Hiratsuka, Kanagawa 259-12, Japan
bSchool of Pharmacy, Tokyo University of Pharmacy & Life Science, 1432-1 Horinouchi, Hachioji, Tokyo 192-03, Japan
cOhme Institute, Tobishi Pharmaceutical Co. Ltd., 1-7-1 Suehiro, Ohme, Tokyo 198, Japan
dTsukuba Institute, Toagousei Chemical Industry Co., 2 Ohkubo, Tsukuba, Ibaragi 300-33, Japan
eSumisho Electronics Co., Ltd., 2-23 Shimomiyabi-cho, Shinjuku-ku, Tokyo 162, Japan
fCancer Research Institute and Department of Chemistry, Arizona State University, Tempe, AZ 85287-1604, USA
Received 13 February 1998; accepted 6 April 1998
AbstractÐThe toad poison bufadienolides including natural and derivatized compounds were tested for their cytotoxic
e€ects on primary liver carcinoma cells PLC/PRF/5 and their structure±cytotoxic activity relationships were studied.
For this study, a ligand-binding model was developed by using a pharmacophore mapping program, Distance
Comparisons (DISCO). The structural features that are common to the 3D structures of active bufadienolides were
identi®ed to provide approach to a 3D QSAR method by using Comparative Molecular Field Analysis (CoMFA)
study and to correlate the steric and electrostatic ®elds of the molecules to their activities. A valuable model which
enables prediction of their activities was obtained from the CoMFA analysis, which may be employed for the drug
designs of new bufadienolide analogues. # 1998 Published by Elsevier Science Ltd. All rights reserved.
Introduction
In the present studies, cytotoxic activities of 80 bufa-
dienolides, including derivatives and isomers of naturally
occurring bufadienolides were studied on the cell line,
primary liver carcinoma PLC/PRF/5. Of them, 16 were
shown to have potent cytotoxicities (IC50<10 3) and
their structure±activity relationships were revealed to
show the necessary structural features for the activities.
The Chinese drug Ch'an Su, called `Senso' in Japanese,
is a product of the skin gland of toads such as Bufo bufo
gargarizans etc., which has been used traditionally as a
cardiotonic, diuretic, anodyne and hemostatic agent. Its
major e€ective components, generally called, the toad
poison or bufadienolides, have a novel steroidal A/B cis
and C/D cis structure with an a-pyrone ring at C17-
position and exhibit a range of biological activities, such
as cardiotic, blood pressure stimulating, respiration, and
antineoplastic activities. Of the bufadienolides, resibufo-
genin is now used as a cardiotonic drug, and bufalin has
recently been reported to have a strong surface anes-
thetic activity1 and cytotoxic e€ect and di€erentiation-
apoptosis activity on murine leukemia HL-60 cells.2
The three dimensional (3D) quantitative structure±
activity relationships (QSAR) of these active com-
pounds were studied by computational methods, in
which, ®rst, calculation was made to ®nd the possible
pharmacophores for those 16 potently cytotoxic com-
pounds by using the Distance Comparisons (DISCO)
program.3 On the basis of the proposed pharmacophore
model, 3D QSAR analysis was conducted by using
Comparative Molecular ®eld (CoMFA) analysis.4
A
suitable model was developed, which might provide
reliable information about the activity and be useful for
*Corresponding author.
0968-0896/98/$19.00 # 1998 Published by Elsevier Science Ltd. All rights reserved
PII: S0968-0896(98)00067-4

1104
Y. Kamano et al./Bioorg. Med. Chem. 6 (1998) 1103±1115
drug designing of active bufadienolides. The cell growth
inhibitory activities of the bufadienolides derived from
the calculations and those of bioassay were compared to
show good agreement of these two sets of data.
extracted with CHCl3. After removal of the solvent,
the residue (16 mg) was chromatographed on a silica gel
column. Elution with n-hexane-acetone (3:1) and
recrystallization from acetone gave 9.5 mg of 4, as a
colorless prism, mp 227±229ꢀC, [a]d 9.9ꢀ (c 0.5, CHCl3),
m/z 486(M)+; lmax(MeOH)/nm 295 (log e 3.20);
1
nmax(KBr)/cm
3380±3110, 3038, 1742, 1729, 1718,
Materials and Methods
General details
1
1675, 1620, 1545, 1265, 1245, 955, 800, 750; H NMR
(CDCl3) d 0.70 (3H, s, 18-CH3), 0.95 (3H, s, 19-CH3), 2.47
(1H, m, 17-H), 2.66 (4H, m, 2ÂCH2 of succinate), 5.13
(1H, bs, 3-H), 6.28 (1H, d, J=9.8, 23-H), 7.24 (1H, d,
J=1.5, 21-H), 7.84 (1H, dd, J=9.8, 2.4, 22-H);
13C NMR (CDCl3) d 16.8 (C-18), 21.3 (C-19), 51.2 (C-
17), 71.0 (C-3), 85.4 (C-14), 115.3 (C-23), 122.7 (C-20),
146.8 (C-22), 148.6 (C-21), 162.5 (C-24), 171.6 (3-OCO-
(CH2)2COOH), 176.9 (3-OCO(CH2)2COOH); Anal. calcd
for C28H38O7: C, 69.11; H, 7.87. Found: C, 69.02; H, 7.95.
For biological test, chromatography and reactions,
commercial solvents of analytical grade were used after
redistillation. All melting points were determined on a
micro hot-stage apparatus (Reichert, Austria) and
uncorrected. Ultra violet spectra were measured with a
SHIMAZU, UV 2500PC spectrometer, optical rota-
tions, with a HORIBA, High Sensitive Polarimeter with
the [a]d values given in 10 1 deg cm2 g 1. EI and FAB
mass spectra were taken with a JEOL JMS-AX505H
mass spectrometer and IR spectra on a Jasco, FT/IR-
300 spectrophotometer. Preparative high-pressure liquid
chromatography (HPLC) was performed with an Inert-
sil PREP-ODS column (20 mm i.d.Â250 mm, GL Sci-
ence Inc.) packed with 10 mm ODS. TLC was conducted
on precoated Kieselgel 60 F254 (Art. 5715; Merck) and
the spots were detected by spraying with 5% sulfuric
acid±ethanol solution and heating (hot plate). 1H and
13C NMR spectra were recorded on a JEOL 400EX
spectrometer at 303 K. The NMR coupling constants (J)
are given in Hz.
Bufalin 3-methacrylate (5). To a solution of bufalin (1)
(10 mg) in pyridine (3mL), methacryloyl chloride (0.2mL)
was added. The mixture was allowed to stand for 3 h at
room temperature and poured into ice-water. The solid
thus obtained was collected and washed with water.
Recrystallization of the crude product from MeOH gave
8.5 mg of 3-methacrylate (5), as a colorless needles: [a]d
5.1ꢀ (c 1.0, CHCl3); m/z 454(M)+; lmax (CH3CN)/nm
300 (log e 3.33); nmax(KBr)/cm 1 3509, 2929, 2359, 1714,
1
1453, 1260, 1024, 798; H NMR (CDCl3) d 0.71 (3H, s,
18-CH3), 0.97 (3H, s, 19-CH3), 1.96 (3H, s, 3-OCO
(CH3)C=CH2), 2.49 (1H, m, 17-H), 5.16 (1H, bs, 3-H),
5.55 and 6.10 (2H, 3-OCO (CH3)C=CH2), 6.27 (1H, d,
J=9.8, 23-H), 7.23 (1H, d, J=1.5, 21-H), 7.83 (1H, dd,
J=9.8, 2.9, 22-H); 13C NMR (CDCl3): d 16.5 (C-18),
21.4 (C-19), 30.5 (3-OCO (CH3)C=CH2), 51.3 (C-17),
70.5 (C-3), 85.4 (C-14), 115.4 (C-23), 122.6 (C-20), 125.0
(3-OCO(CH3)C=CH2), 137.1 (3-OCO(CH3)C=CH2),
146.7 (C-22), 148.6 (C-21), 162.5 (C-24), 166.8 (3-OCO-
(CH3)C=CH2); Anal. calcd for C28H38O5: C, 73.98; H,
8.43. Found: C, 73.77; H, 8.41.
Materials
Ch'an Su, bufadienolides and cardenolides: Ch'an Su
was obtained in Hong Kong folk-medical market. Of
the 80 compounds tested, some were isolated from Ch'an
Su, some were commercially obtained and some were
derivatized from known bufadienolides in our laboratory.
Bufalin 3-succinate (4), bufalin 3-methacrylate (5),
bufalin 3-[N-(tert-butoxycarbonyl)hydrazido] succinate
(8),9 14b, 15b-epoxy-3-oxo-bufa-1, 20, 22-trienolide
(Á1-3-oxo-resibufogenin) (27), 14b, 15b-epoxy-3-oxo-
bufa-1, 4, 20, 22-tetraenolide (Á1,4-3-oxo-resibufogenin)
(28), desacetyl-cinobufagin 3-acetate-16-succinate (48),
3-oxo-14a-artebufogenin (56), digitoxigenin 3-methyl-
suberate (67), methyl 3b-hydroxy-14b, 15b; 16b, 22b-
diepoxy-21-nor-5b, 14b-cholanoate (CMKA) (76), and
methyl(E)-3b, 16b-dihydroxy-14b, 15b-epoxy-21-nor-
chol-20(22)-enoate (CMKB) (77) were newly prepared
from the corresponding starting materials as described
below.
Bufalin 3-[N-(tert-butoxycarbonyl)hydrazido]succinate (8).
To a cold solution of bufalin 3-hemi-succinate (38 mg)
and tert-butylcarbazoate (14 mg) in CH2Cl2 (2.8 mL) in
ice bath was added triethylamine (0.01 mL) and 1-ethyl-
3-[3a-(dimethylamino)propyl] carbodiimide hydro-
chlorine (EDCI) (15 mg) and the mixture was stirred for
5.5 h. Another 15 mg portion of EDCI was added to
accelerate the formation of butoxycarbonyl hydrazide
(8). Stirring was continued for another 2 h in ice-bath
and for a day at room temperature. The mixture was
then diluted with CH2Cl2 (5 mL) and poured into water.
The CH2Cl2 layer was washed successively with water,
2% Na2CO3 and water. Evaporation of the solvent gave
a slightly yellowish oily residue (55 mg). Silica gel col-
umn chromatography of the residue with n-hexane-
acetone (3:1) gave 12 mg of pure butoxycarbonyl
Bufalin 3-succinate (4). A mixture of bufalin (1: 15 mg)
and succinic anhydride (30 mg) in pyridine (5 mL) was
re¯uxed for 2 h and then poured into ice-water. It was

Y. Kamano et al./Bioorg. Med. Chem. 6 (1998) 1103±1115
1105
hydrazide (8) as a colorless prism, mp 255±258 ꢀC, [a]d
1.6, 22-H); 13C NMR(CDCl3) d 16.8 (C-18), 18.8 (C-19),
59.7 (C-17), 73.6 (C-15), 115.4 (C-23), 121.8 (C-20), 124.3
(C-1), 128.0 (C-2), 146.7 (C-22), 149.7 (C-21), 154.4 (C-4),
161.9 (C-24), 167.6 (C-5), 186.0 (C-3); Anal. calcd. for
C24H26O4: C,76.16; H, 6.93. Found: C, 75.89; H, 7.07.
8.3ꢀ (c 1.0, CHCl3); m/z 600(M)+; lmax(MeOH)/nm
1
297 (log e 3.40); nmax(KBr)/cm
3600, 3355, 1740,
1678, 1655, 1552, 1240, 970, 835, 797; 1H NMR (CDCl3)
d 0.70 (3H, s, 18-CH3), 0.95 (3H, s, 19-CH3), 2.17 (1H, m,
17-H), 1.46 (9H, s, COOC(CH3)3), 2.53 and 2.70 (4H,
each m, OCO(CH2)2CO), 5.10 (1H, bs, 3-H), 6.27 (1H, d,
J=9.8, 23-H), 6.59 (1H, bs, 3-OCO(CH2)2CONHNH-
COOC(CH3)3), 7.23 (1H, d, J=1.5, 21-H), 7.76 (1H, bs,
3-OCO(CH2)2CONHNHCOOC(CH3)3), 7.84 (1H, dd,
J=9.8, 2.9, 22-H); 13C NMR(CDCl3) d 16.6 (C-18),
21.3 (C-19), 28.1 (COOC(CH3)3), 53.8 (C-17), 71.1
(C-3), 81.9 (COOC(CH3)3), 85.4 (C-14), 115.3 (C-23),
122.7 (C-20), 146.8 (C-22), 148.6 (C-21), 155.4
(3-OCO(CH2)2CONHNHCOOC(CH3)3), 162.5 (C-24),
171.4 (3-OCO(CH2)2CONHNHCOOC(CH3)3), 172.3
(3-OCO(CH2)2CONHNHCOOC(CH3)3); Anal. calcd
for C33H48N2O8: C, 65.97; H, 8.05; N, 4.66. Found: C,
66.03, H, 7.92, N, 4.63.
Desacetyl-cinobufagin 3-acetate-16-succinate (48). A mix-
ture of 47 (500 mg),1 succinic anhydride (500 mg) and
pyridine (10 mL) was re¯uxed for 4 h. The mixture was
poured into ice-water and extracted with CHCl3. After
the removal of solvent, the residue (553 mg) was chro-
matographed on a silica gel column with n-hexane-acet-
one (3:1). Recrystallization from n-hexane-acetone gave
495 mg of colorless needles of 48: [a]d +10.7ꢀ (c 1.0,
CHCl3); m/z 542(M)+; lmax(MeCN)/nm 295 (log e
1
3.38); nmax(KBr)/cm
3400±3100, 1740, 1720, 1700,
1645, 1545, 1250, 1235, 958, 894, 790; 1H NMR (CDCl3)
d 0.83 (3H, s, 18-CH3), 1.00 (3H, s, 19-CH3), 2.07 (3H,
s, 3-OCOCH3), 2.55 (4H, m, 2ÂCH2 of succinate), 2.81
(1H, d, J=9.5, 17-H), 3.68 (1H, d, J=1.5, 15-H), 5.10
(1H, bs, 3-H), 5.53 (1H, dd, J=9.5, 1.5, 16-H), 6.25
(1H, d, J=10.2, 23-H), 7.19 (1H, d, J=3.0, 21-H), 7.95
(1H. dd, J=10.2, 3.0, 22-H); 13C NMR (CDCl3) d 17.2
(C-18), 23.7 (C-19), 50.4 (C-17), 59.5 (C-15), 70.2 (C-3),
72.3 (C-14), 75.1 (C-16), 114.1 (C-23), 116.2 (C-20),
148.3 (C-22), 151.4 (C-21), 162.3 (C-24), 170.7 (3-
OCOCH3), 171.5 (16-OCO(CH2)2COOH), 175.5 (16-
OCO(CH2)2COOH); Anal. calcd for C30H38O9: C,66.40;
H, 7.06. Found: C, 66.44; H, 7.02.
14b,15ꢀ-Epoxy-3-oxo-bufa-1,20,22-trienolide(Á1-3-oxo-
resibufogenin) (27) and 14ꢀ, 15ꢀ-epoxy-3-oxo-bufa-
1,4,20,22-tetraenolide(Á1,4-3-oxo-resibufogenin) (28). A
mixture of 3-oxo-resibufogenin (26)12 (1.25 g, 3.27 mmol)
and 2,3-dichloro-5,6-dicyano-1,4-benzoquinone(DDQ)
(816 mg, 3.6 mmol) in dioxane (30 mL) was re¯uxed for
4 h. The reaction mixture was mixed with silica gel
(10 g), which was dried and placed on a silica gel column
(700Â30 mm). The column was eluted with n-hexane-
acetone (4:1) to give four major fractions. 27 was
isolated from the second fraction, which was further
puri®ed by preparative TLC. Compound 27: colorless
needles, mp 221±223 ꢀC, [a]d +70ꢀ (c 0.83, CHCl3); m/z
3-Oxo-14ꢁ-artebufogenin (56). To a solution of 14a-
artebufogenin (55)15 (50 mg) in acetic acid (3 mL), a
solution of chromic acid (10 mg) in acetic acid (1 mL)
was added with stirring. The oxidation was conducted
at 15±18 ꢀC for 4 h. After decomposition of the excess
reagent with MeOH (0.25 mL, with stirring), the mixture
was poured into ice-water and extracted with CHCl3.
The extract was washed with water and concentrated to
dryness. Recrystallization of the crude product from
MeOH gave 37 mg of 3-oxo-14a-artebufogenin (56) as a
colorless needles: mp 210±213 ꢀC, [a]d +47.6ꢀ (c 1.0,
380 (M)+; lmax(MeOH)/nm 297 (log e 2.42), 223 (log e
1
2.99); nmax(KBr)/cm
2963, 2936, 2872, 1742, 1723,
1696, 1680, 1538, 1124, 952, 825, 830, 785; 1H
NMR(CDCl3) d 0.75 (3H, s, 18-CH3), 1.16 (3H, s, 19-
CH3), 2.45 (1H, m, 17-H), 3.46 (1H, s, 15-H), 5.94(1H,
d, J=10.2, 2-H), 6.18 (1H, dd, J=9.7, 0.9, 23-H), 6.78
(1H, d, J=10.2, 1-H), 7.17 (1H, dd, J=2.7, 0.9, 21-H),
7.70 (1H, dd , J=9.7, 2.8, 22-H); 13C NMR (CDCl3) d
16.8 (C-18), 20.7 (C-19), 59.8 (C-17), 73.8 (C-15), 115.4
(C-23), 122.0 (C-20), 127.7 (C-1, C-2), 146.8 (C-22),
149.7 (C-21), 160.1 (C-24), 200.1 (C-3); Anal. calcd for
C24H28O4: C,75.76; H, 7.42. Found: C, 75.98; H, 7.35.
CHCl3); m/z 382 (M)+ for C24H30O4; lmax(MeCN)/nm
1
298 (log e 3.70); nmax(KBr)/cm
1639, 1541, 1450, 1386, 1322, 1249, 1227, 1204, 1123,
2933, 2855, 1742,
1
990, 950, 833, 789; H NMR (CDCl3) d 0.67 (3H, s, 18-
CH3), 1.05 (3H, s, 19-CH3), 1.82 (1H, m, 14-H), 2.79
(1H, m, 17-H), 6.34 (1H, d, J=1.0, 23-H), 7.27 (1H, dd,
J=12.7, 2.9, 22-H), 7.33(1H, m, 21-H); 13C NMR
(CDCl3) d 13.9 (C-18), 20.3 (C-19), 64.8 (C-14), 115.8
(C-20), 116.1 (C-23), 144.1 (C-22), 149.1 (C-21), 161.4
(C-24), 212.4 (C-3), 212.5 (C-15). Anal. calcd for
C24H30O4: C,75.36; H, 7.91. Found: C, 75.57; H, 7.95.
The fourth fraction was dried and the residue was
recrystallized from acetone-hexane to give 170 mg
(14%) of tetraenolide (28) as a colorless needles, [a]d
+20.0ꢀ (c 1.0, CHCl3); m/z 378 (M)+; lmax(MeOH)/
nm 298 (log e 3.73), 236 (log e 4.19); nmax(KBr)/cmx1
2940, 2871, 1746, 1722, 1662, 1642, 1538, 1124, 950, 832,
785; 1H NMR(CDCl3) d 0.86 (3H, s, 18-CH3), 1.26 (3H,
s, 19-CH3), 2.17 (1H, m, 17-H), 3.46 (1H, s, 15-H), 6.35
to 6.0 (3H, m, 23-H, 2-H, 4-H), 7.05 (1H, d, J=10.0, 1-H),
7.25 (1H, dd, J=1.6, 0.9, 21-H), 7.75 (1H, dd, J=9.9,
Digitoxigenin 3-methylsuberate (67). To a solution of
digitoxigenin 3-suberate (66) (100 mg) in Et2O (10 mL),
a solution of CH2N2 (ca. 6 mg) in Et2O (15 mL) was

1106
Y. Kamano et al./Bioorg. Med. Chem. 6 (1998) 1103±1115
1
added and the mixture was allowed to stand at room
temperature for 10 min. The solvent was evaporated
carefully in vacuo and the residue was puri®ed by silica gel
column chromatography with n-hexane-acetone mixture
to give 85 mg of 67 as colorless prisms: [a]d +42.2ꢀ (c 1.0,
CHCl3); m/z 544 (M)+ for C32H48O7; lmax(MeCN)/nm
n
max(KBr)/cm
3560, 3500±3400, 3050, 2720, 1730,
1
1300, 1226, 1095, 1030, 978, 768; H NMR (CDCl3) d
0.98 (3H, s, 18-CH3), 1.05 (3H, s, 19-CH3), 2.64 (1H, dd,
J=10.8, 8.4, 17-H), 2.84 (2H, dd, J=16.1, 7.3, 22-CH2),
3.04 (1H, ddd, J=9.9, 9.5, 3.3, 20-H), 3.41 (1H, bs, 15-
H), 3.69 (3H, s, 24-OCH3), 4.13 (1H, br m, 3-H), 4.45
(1H, ddd, J=9.2, 7.3, 5.5, 21-H), 4.72 (1H, dd, J=8.4 ,
1.1, 16-H), 9.58(1H, d, J=3.3, 25-H). 13C NMR
(CDCl3) d 15.7 (C-19), 20.5 (C-11), 20.5 (C-7), 23.8 (C-
18), 25.6 (C-6), 27.8 (C-2), 29.5 (C-1), 33.2 (C-8), 33.3
(C-4), 35.5 (C-10), 35.9 (C-5), 39.2 (C-9), 39.3 (C-12),
39.9 (C-22), 42.0 (C-13), 51.8 (C-24-OCH3), 54.1 (C-17),
60.0 (C-15), 60.1 (C-20), 66.8 (C-3), 76.7 (C-14), 78.3 (C-
21), 82.9 (C-16),171.2 (C-23), 200.7(C-25). Anal. calcd
for C25H36O6: C, 69.42; H,8.39. Found: C, 69.48; H,
8.41.
1
216 (log e 4.22), 272 (log e 3.01)nm; nmax(KBr)/cm
3550, 3098, 2941, 2877, 1782, 1749, 1718, 1619, 1446,
1380, 1255, 1213, 1170, 1154, 1065, 1024, 956, 925. 887;
1H NMR (CDCl3) d 0.88 (3H, s, 18-CH3), 0.96 (3H, s,
19-CH3), 2.78 (1H, m, 17-H), 3.67 (3H, s, 3-OCO(CH2)6-
COOCH3), 4.81 and 4.99 (2H, dd, J=18.1, 1.46, 21-H),
5.10 (1H, m, 3-H), 5.88 (1H, s, 22-H), 13C NMR
(CDCl3) d 15.8 (C-18), 23.7 (C-19), 28.8 (3-OCO(CH2)6-
COOCH3), 50.9 (C-17), 70.1 (C-3), 73.4 (C-21), 85.5 (C-14),
117.7 (C-22), 173.2 (C-23), 174.1 (3-OCO(CH2)6COOCH3),
174.5 (C-20 and 3-OCO(CH2)6COOCH3). Anal. calcd
for C32H48O7: C,70.56; H, 8.88. Found: C, 70.48; H, 8.95.
Bioassay
Methyl 3ꢀ-hydroxy-14ꢀ,15ꢀ; 16ꢀ,22ꢀ-diepoxy-21-nor-
5ꢀ,14ꢀ-cholanoate (CMKA) (76) and methyl(E)-3ꢀ,16ꢀ-
dihydroxy-14ꢀ, 15ꢀ-epoxy-21-nor-chol-20(22)-enoate
(CMK B) (77). Cinobufagin (36) (1.0 g) was dissolved
in 0.1 N KOH in MeOH (50 mL) and the mixture was
allowed to stand for 6 h at room temperature. The
reaction mixture was acidi®ed to pH 3±5 by addition of
10% p-toluenesulfonic acid in MeOH±water (1:1). By
evaporation, the volume was reduced to about one
third. The mixture was poured into ice-water and
extracted with CHCl3. The CHCl3 layer was washed
with aqueous Na2CO3 and water and evaporated to
dryness to give a residue (1.02 g) which was chromato-
graphed on a silica gel column. Elution with n-hexane-
acetone (7:1, 6:1, and 4:1) gave 313 mg of crude CMKA
(76), and 142 mg of CMKB (77). Recrystallization of the
crude CMKA from acetone gave 296 mg of needles of
76 and recrystallization of the crude CMKB (77) from
n-hexane-acetone gave 281 mg of needles of 77. CMKA
(76): colorless needles, [a]d +25.0ꢀ (c 1.0, CHCl3); m/z
Primary liver carcinoma PLC/PRF/5 cells were main-
tained in tissue culture ¯asks and grown in 96-well
microtiter plates for the assay. Appropriate dilutions of
the test compounds (102 10 6 mg/mL) were added to
the culture. After incubation at 37 ꢀC, 5%CO2, for 72 h,
the survival rate of the cells in the cultures were eval-
uated by the MTT method. The e€ect was shown as
IC50, which is the concentration of test compound (mg/
mL) to give 50% inhibition of the growth of PLC cells.
Computational methods
Sybyl molecular modeling software (version 6.3) was
used for all molecular modeling techniques and CoMFA
studies.4 The compounds were built from fragments in
the SYBYL database. Each structure was fully geo-
metry-optimized by using the standard Tripos molecular
mechanics force ®eld with a distance-dependent dielec-
tric function and a 0.01 kcal/mol energy-gradient con-
vergence criterion. In addition, each stable conformer of
side chains including d-lactone ring at C-17 was
obtained by using the systematic search routine in
SYBYL. Partial atomic charges required for calculation
of the electrostatic interaction energies were calculated
by using the Gasteiger±Marsili method.30
1
404 (M)+, 386 (M-H2O)+; nmax(KBr)/cm 3500, 3050,
1745, 1190, 1145, 1095, 1075, 1037, 1003, 937, 915, 757;
1H NMR (CDCl3) d 0.98 (3H, s, 18-CH3), 1.06 (3H, s,
19-CH3), 1.80 and 1.67 (2H, m, H-20), 2.29 (1H, ddd,
J=11.5, 8.3, 3.2, 17-H), 2.57 and 2.74 (2H, d J=6.0,
7.0, 22-CH2), 3.40 (1H, bs, 15-H), 3.69 (3H, s, 24-
OCH3), 4.13 (1H, br m, 3-H), 4.27 (1H, m, 21-H), 4.55
(1H, dd, J=8.0, 1.5, 16-H), 13C NMR (CDCl3) d 15.4
(C-19), 20.5 (C-7), 20.6 (C-11), 23.8 (C-18), 25.7 (C-6),
27.8 (C-2), 29.5 (C-1), 33.3 (C-4), 33.4 (C-8), 35.5 (C-
20), 35.5 (C-10), 36.5(C-5), 39.2 (C-9), 39.5 (C-12), 40.3
(C-22), 41.3 (C-13), 51.6 (C-24-OCH3), 52.7 (C-17), 60.3
(C-15), 66.8 (C-3), 77.0 (C-14), 78.1 (C-21), 82.1 (C-
16),172.0 (C-23); Anal. Calcd for C24H36O5: C, 71.25;
H, 8.97. Found: C, 71.26; H, 9.02. CMKB (77): color-
less needles, [a]d 5.7ꢀ (c 1.0, CHCl3); m/z 432 (M)+,
414(M-H2O)+, 404(M-CO)+, 386 (M-CO-H2O)+;
Calculations of ring centroids, least square ®tting, and
excluded volume analyses were also carried out on 16
active bufadienolides (IC50<10 3) by using DISCO
program3 in SYBYL. The pharmacophore mapping
strategy was the calculation of the location of the ligand
and site points, followed by execution of DISCO to ®nd
pharmacophore maps. Other compounds were aligned
via this pharmacophore model.
For each of the alignment sets, the steric and Coulombic
potential energy ®elds were separately calculated at each

Y. Kamano et al./Bioorg. Med. Chem. 6 (1998) 1103±1115
1107
lattice intersection on a regularly spaced grid of 2.0‹
Structure±activity relationships (SAR) consideration
units in all x, y, and z directions. The steric term repre-
sents the van der Waals interactions, whereas the Cou-
lombic term represents the electrostatic interactions for
which a distance-dependent dielectric expression e=Rij
was adopted. The grid pattern was generated auto-
matically by the SYBYL/CoMFA routine. An sp3 car-
bon atom with a +1.0 charge was selected as the probe
for calculation of the steric and electrostatic ®eld.
Values of the steric and electrostatic energies were trun-
cated at 30 kcal/mol.
Bufadienolides, cardenolides and their derivatives were
divided into A±E groups, group A: 14-hydroxy com-
pounds (1±20, 60±62); group B: 14, 15-epoxy com-
pounds (21±51); group C: cardenolides and their
derivatives (63±68); group D: other bufadienolides (52±
59), and group E: derivatives with cleaved a-pyrone ring
(69±80). The structure and activity relationship revealed
for each group may be summarized as follows.
Group A (1±20, 60±62)
To obtain a 3D-QSAR, partial least squares (PLS)31
method was used. The PLS method was used success-
fully in many QSAR studies for rationalization of those
structural features a€ecting the biological activity. The
PLS algorithm was initially used with the cross-valida-
tion option to obtain the optimal number of compo-
nents needed for the subsequent analysis of the data. In
the leave-one-out cross-validation, each compound was
systematically excluded from the set and its activity
predicted by a model derived from the rest of the com-
pounds. The optimal number of components was then
chosen as that which yielded either the smallest rms
error or the largest cross-validated r2 value. A ®nal PLS
analysis was then performed by using the reported
optimum number of components, with no cross-valida-
tion. This generated a ®tted correlation of the entire
training set with conventional r2 values. The steric and
electrostatic ®elds were scaled according to the CoMFA
standard deviations in order to give the same potential
weights on the resulting QSAR. The 3D-QSAR calibra-
tion model so derived was then employed to give theo-
retical inhibitory e€ect values of 10 bufadienolide
derivatives which were compared with their bioassay
data.
To bufalin (1) having a basic structure, introduction of
19-aldehyde (19) or 11a-hydroxy (15) group indicated a
tendency of the enhancive activity in 10
4
order,
whereas that of 16b-acetoxy (13), 16b-hydroxy (14) or
5b-hydroxy (18) group indicated a tendency to reduce it.
In addition to introduction of 15a-hydroxy (60 and 61)
group, epimerization of 3b-hydroxy or 14b-hydroxy
group to corresponding a-isomers (2 and 10) remarka-
blely reduced activity. On the other hand, introduction
of a double bond to C-4 position (11) did not reduce the
activity of bufalin (1). The activity of bufalin (1) was
increased by acetylation of 3b-hydroxy group (3), but
other 3-esteri®ed derivatives (4±8) and 3-keto derivative
(9) had lower activities than bufalin (1). It is interesting
that the activity of 3-[N-(tert-butoxycarbonyl)hydrazide]-
succinate (8) and 3-suberate (6) were stronger than that
of 3-succinate (4).
Group B (21±51)
In this group, only cinobufagin (36) showed the strong
4
activity of 10
order. Also, cinobufotalin (50) and
3
resibufagin (21) kept their activity to 10 order. Thus,
introduction of 19-aldehyde (21) group to resibufogenin
(23) having a basic structure increased the activity, but
that of 19-CH2OH (22) group decreased it. For C-16
position, introduction of acetoxy group increased
remarkabely the activity, as in the case of resibufogenin
(23) and cinobufagin (16-acetoxy resibufogenin: 36), or
marinobufagin (35) and cinobufotalin (16-acetoxy mar-
inobufagin: 50), but that of hydroxy (46, 47, 49 and 51)
or ester group (48) reduced it. Introduction of 5b-
hydroxy (35 and 50) group reduced the activity.
Also, introduction of an ester (38±41, 43, and 47) or
oxo-group (26, 32, 42, and 44) to C-3 position and that
of double bond to C-1 and C-4 (27 or 28) decreased the
activity. In generally, the activity of cinobufagin deriva-
tives kept higher value than that of resibufogenin
derivatives by reason of the e€ect of 16b-acetoxy group.
Combination of various substituents at C-3 and C-16
may a€ect the potency of cytotoxicities on PLC cells.
The sudden drop in activity appeared on epimerization
of 3b-hydroxy (30 and 36) group and 14b, 15b-epoxy
Results and Discussion
Cytotoxicity
The cytotoxic activities (IC50 mg/mL) of the 80 natural
and derivatized bufadienolides and cardenolides on
PLC cells are shown in Table 1. Of the natural bufadi-
enolides, 20 (1, 6, 9, 13, 14, 15, 18, 19, 21, 23, 35, 36, 40,
46, 50, 51, 53, 55, 63, and 64) were active on PLC cells
(IC50 <8.2 mg/mL). Hellebrigenin (19) having an alde-
hyde group at C-19, bufalin 3-acetate (3), gama-
bufotalin (15), and bufalin (1) were signi®cantly active
(IC50 1.6±2.8Â10 4 mg/mL). 14b-Hydroxy derivatives
were generally more active than 14b, 15b-epoxy com-
pounds and open a-pyrone ring compounds. Similar
structure±activity relations have been reported on the
surface anaesthic activities1 and rhinovirus inhibitory
activities9 of bufadienolides.

1108
Y. Kamano et al./Bioorg. Med. Chem. 6 (1998) 1103±1115
(25, 26, and 38) group into corresponding 3a-hydroxy
(31 and 37) and 14a, 15a-epoxy (30, 33 and 49) group.
the activity. It is interesting that the activity of 3-
methylsuberate (67) was higher than that of 3-suberate
(66).
Group C (63±68)
Group D (52±59)
The activities of cardenolides were generally weaker
than those of corresponding bufadienolides having the
same steroidal skeletons. Introduction of 5b-hydroxy
(63) group or an ester function at C-3 (65±67) reduced
Miscellaneous bufadienolides and cardenolides, of
which C-14 position did not carry any oxygen group,
were collected. From that the compounds in this group
Table 1. The cytotoxicity of bufadienolides and related compounds against PLC/PRF/5 cell
No. Compds
IC50(mg/mL)
No. Compds
IC50(mg/mL)
4
1
1
2
3
4
5
6
7
8
Bufalin5
2.8Â10
39 Cinobufagin-3-succinate19
40 Cinobufagin-3-suberate8
41 Cinobufagin-3-cinnamate19
42 3-oxo-Cinobufagin20
43 Cinobufagin-3,5-dinitrobenzoate19
44 3,16-Diketo-cinobufagin19
45 16-oxo-Cinobufagin-3-acetate17,19
46 Desacetyl-cinobufagin5
47 Desacetyl-cinobufagin-3-acetate19
48 Desacetyl-cinofufagin-3-acetate-
16-succinate
8.5Â10
2
2
3a-Hydroxy-bufalin6
7.6Â10
1.0Â10
4
Bufalin-3-acetate7
2.0Â10
6.4
5.6Â10
2
2
Bufalin-3-succinate
Bufalin-3-methacrylate
Bufalin-3-suberate8
Bufalin-3-methylsuberate8
Bufalin-3-[N-(tert-butoxycarbonyl)
hydrazido]succinate9
3-oxo-Bufalin10
6.6Â10
2
1
5.9Â10
7.8Â10
4
1
7.5Â10
8.5Â10
3
9.0Â10
2.4
5.2Â10
4
1
6.2Â10
2
9.8Â10
3
9
9.3Â10
50
10 14a-Hydroxy-bufalin-3b,16b-
diacetate7
11 Scillarenin10
12 3-oxo-Sillarenin10,11
13 Bufotalin5
14 Desacetyl-bufotalin5
15 Gamabufotalin5
16 Gamabufotalin-3-acetate12
17 3-oxo-Gamabufotalin-11-acetate12
18 Telocinobufagin5
19 Hellebrigenin5
20 Acetyl-arenobufagin13a,13b
21 Resibufagin14,15
7.5
1
49 Desacetyl-14a,15a-epxoy-
cinobufagin-3-acetate17
50 Cinobufotalin5
51 Desacetyl-cinobufotalin5
52 b-Chlorohydrin21
8.8Â10
4
3.1Â10
4
3
1.9Â10
1.0Â10
4
1
3.4Â10
6.7Â10
4
2
7.8Â10
7.0Â10
4
1
2.3Â10
53 14b-Artebufogenin5,16
54 14b-Artebufogenin-3-acetate15
55 14a-Artebufogenin5,15
56 3-oxo-14a-Artebufogenin
57 Á14-Bufalin18
8.2Â10
2
1
6.0Â10
5.5Â10
1
7.9Â10
8.2
7.0
4
3.5Â10
4
1
1.6Â10
8.3Â10
1
8.7Â10
58 Á14-3-oxo-Bufalin18
59 Á14-Bufotalin-3-acetate23
60 15a-Hydroxy-bufalin22
61 15a-Hydroxy-bufalin-3-acetate23
62 15-oxo-Bufalin-3-acetate24
63 Periplogenin25
4.9
5.0
35
3
8.9Â10
1
22 Resibufaginol14,15
23 Resibufogenin5
6.1Â10
2
1
7.7Â10
7.1Â10
2
24 3a-Hydroxy-resibufogenin6
25 Resibufogenin-3-acetate16
26 3-oxo-Resibufogenin6
27 Á1-3-oxo-Resibufogenin
28 Á1,4-3-oxo-Resibufogenin
29 16a-Hydroxy-resibufogenin-3-
acetate17
30 14a,15a-Epoxy-resibufogenin18
31 3a-Hydroxy-14a,125a-epoxy-
resibufogenin6
8.3
7.7Â10
1
2
3.2Â10
6.1Â10
4
5.8
2.6
6.7
6.7
64 Digitoxigenin5
8.7Â10
1
65 Digitoxigenin-3-acetate26
66 Digitoxigenin-3-suberate19
67 Digitoxigenin-3-methylsuberate
68 Á14-Digitoxigenin26
5.5Â10
5.2
5.3Â10
1
1
3.6Â10
20
69 Isobufalin27
10
4.7
70 Isobufalin Me-ester27
71 Isobufalin Me-ester-3-acetate27
72 Isobufalin Et-ester27
73 Isobufalin Et-ester-3-acetate27
74 Isobufotalin Me-ester-3-acetate27
75 Isogamabufotalin Me-ester-3,11-
diacetate27
9.1
6.6
7.5
6.4
49
32 3-oxo-14a,15a-Epoxy-
resibufogenin6
8.7
5.9
4.7
33 14a,15a-Epoxy-resibufogenin-3-
acetate18
9.7
34 14a,15a-Epoxy-resibufogenin-3a-
acetate6
35 Marinobufagin5,14
36 Cinobufagin5
37 3a-Hydroxy-cinobufagin6
38 Cinobufagin-3-acetate19
76 CMKA
77 CMKB
78 RRA-128
79 14b-arte-MK29
80 14b-arte-Et-K-Ac29
9.5
26
1
5.2Â10
4
7.4Â10
>50
8.4
22
2
9.8Â10
2
4.0Â10

Y. Kamano et al./Bioorg. Med. Chem. 6 (1998) 1103±1115
1109
Figure 1. Structures of bufadienolides (1±62), cardenolides (63±68) and a-pyrone opening compounds (69±80).

1110
Y. Kamano et al./Bioorg. Med. Chem. 6 (1998) 1103±1115
Figure 1.Ðcontd.

Y. Kamano et al./Bioorg. Med. Chem. 6 (1998) 1103±1115
1111
showed lower activity than that of other compounds
having oxygen group at C-14, it was suggested that
oxygen at C-14 position had more important e€ect than
the steroidal structure on the activity. On the other
those of the corresponding parent compounds. The
activity was increased when 3b-hydroxy substituents
were converted to 3b-acetate in the 14b-hydroxy deriva-
tives, but decreased in the 14b, 15b-epoxy derivatives.
It is distinctly possible that a combination of various
substituents at C3, C19, and D ring may occur potency
of inhibition in cell growth. The important structural
features which in¯uence the cytotoxic activities of natu-
rally occurring bufadienolide, cardenolide, and their
derivatives were summarized in Figure 2. Thus, 17-a-
pyrone ring, 3b, 14b-hydroxyls, and the ring junction
(C/D cis) in bufadienolides were apparently most crucial
requirement for the potent cytotoxic activity.
2
hand, the activity of 10 order of b-chlorohydrin (52)
having 14b-Cl group was slightly higher than that of
resibufogenin (23) having 14b, 15b-epoxy group. Also,
the IC50 value of 14b-artebufogenin (53) (14b-H) was
stronger than that of corresponding 14a-H isomer (55).
This demonstrated that the compounds having cis C/D
ring junction had higher cytotoxicities than those with
trans junction.
Group E (69±80)
DISCO analysis
Those compounds having each opened a-pyrone ring
moiety (69±80) had more or less the same e€ect on the
cells, suggesting that a-pyrone ring was apparently not a
necessary requirement for the cytotoxicities on PLC
cells of bufadienolides.
For preparation of pharmacophore model, 2D struc-
tures was converted into the 3D structures to translate
the 2D structure±activity information into the 3D
requirements for the activity. In the present study, 16
bufadienolides, showing potent cytotoxic activities
(IC50<10 3), were chosen as active compounds. For
each of the preferred conformations the points to be
considered in the superposition step were chosen and
their locations were determined by calculation. These
points included not only the atoms in the molecule but
also the central points of aromatic ring centroids, and
the points locating between the hydrogen-bond donors
and acceptors in hydrogen bondings. The points of
hydrogen bonds involved in this calculation were carbonyl
In an over view, the biological evaluation of 14-hydroxy
and 14, 15-epoxy as well as 14-artebufogenins revealed
that the con®guration is crucial for the cytotoxicity:
those having b type con®guration is superior to those
having a con®guration. Deoxygenation at C-14 reduced
the activity. The 14b-hydroxy derivatives having 16-
acetoxy group reduced activity, whereas the 14b, 15b-
epoxy derivatives increased it. The activity of 3b-ester
and 3-ketone derivatives were reduced, comparing with
Figure 2. E€ects of various partial structures and groups on cytotoxic activities of naturally occurring bufadienolides, cardenolides
and their related analogues. The structure is a bufalin.

1112
Y. Kamano et al./Bioorg. Med. Chem. 6 (1998) 1103±1115
Figure 3. Superimposed structures of 16 bufadienolides analysed by DISCO.
oxygens normally interacted with hydrogen-bond donors
in the directions of an angle of 120ꢀ, having the distance
of 2.9±3.0 A between the carbonyl oxygen and the
hydrogen. On the basis of these chemical similarities
such as hydrophobic, hydrogen-bond donating or accept-
ing groups, the preferred conformers were examined for
their superimposability by the program DISCO.3
features was well superimposed (Figure 3). The mean of
the distances among these features was 5.94 A.
CoMFA analysis
The above SAR studies indicated that a combination of
various substituents may increase the cell growth inhi-
bition. In applying these results in QSAR studies, one of
the important goals is the quantitative correlation
In the present survey, for the least squares ®tting, three
points, i.e. the hydrophobic ring centroid of steroidal
backbone, and the points locating between the donor
and acceptor atoms of hydrogen bonds, involving the
two electron lone pairs of the oxygen atoms, of bufalin
3-acetate (3) as a template molecule were chosen.
DISCO is usually instructed to iterate the tolerance at
which two inter-point distances are considered the same.
The maximum tolerance of these features, i.e. hydro-
phobic Â5, acceptor atom Â2, and donor site Â2 shown
in Figure 3, of the 16 bufadienolides was 1.5 A indicat-
ing that the three dimensioned structures including these
Table 2. Summary of CoMFA-PLS results (PLC cells)
No. of compounds
Opt. no. of components
Prob atom
Cross-validated r2
Std error of estimate
r2
70
5
c (sp3,+1)
0.326
0.729
0.803
52.297
F values
Contributions
Steric
Electrostatic
Figure 4. Plot of actual vs predicted inhibitory activities on
cells of primary liver carcinoma PLC/PRF/5 (r2=0.803). The
0.468
0.532
activities are expressed by log 1/IC50
.

Y. Kamano et al./Bioorg. Med. Chem. 6 (1998) 1103±1115
1113
between 3D molecular structure and biological function,
which is subsequently predictable of this property for
novel compounds.
is evident that the CoMFA-derived QSAR manifested a
relatively good cross-validated r2 and indicating thereby
a considerable predictive and correlative capacity for
growth inhibition of the PLC cells, regardless of its large
number of samples (n=70). The relative contributions
of the steric and electrostatic ®elds to this model were
almost equal. Evidence for the predictive performance
of the CoMFA-derived model is provided in Figure 4
which shows plots of actual versus predicted growth
inhibitions [log (1/IC50)].
Analysis was performed for 70 kinds of natural and
semisynthetic analogues, which excluded 10 analogues
synthesized in this study. First, `leave-one-out' cross-
validation was used so that each compound would be
systematically excluded and its activity predicted by the
analysis during formulation of the regression equation.
For this set, acceptable q2 (0.326) was obtained. Second
PLS analysis was performed by using the optimum
number of components with no cross-validation. The
results of a non-cross-validated PLS analysis using the
70 compounds are listed in Table 2, and the actual,
predicted activity and residuals are shown in Table 3. It
The major steric and electrostatic features of the QSAR
are represented in Figure 5 in the form of three-dimen-
sional contour maps, displayed as transparent surfaces.
The surfaces at the top in Figure 5 indicate the areas in
space around the template molecule (19) where increases
Table 3. Actual, predicted activities and residuals from CoMFA analysis
log 1/IC50 (PLC)
Compds
Actual
Predicted
Residual
Compds
Actual
Predicted
Residual
1
2
3.55
1.12
3.70
3.12
2.05
2.03
0.88
3.51
0.72
3.47
3.11
3.64
1.22
0.10
3.46
3.80
0.06
2.05
0.22
1.11
0.92
0.50
0.76
0.83
1.30
0.67
0.94
0.77
0.67
0.28
3.13
1.01
1.40
0.07
2.00
2.74
1.25
1.67
3.68
2.54
2.17
0.72
2.88
1.18
3.47
3.15
2.65
1.23
0.01
2.96
3.62
0.80
1.12
0.33
0.27
0.38
0.27
0.13
1.05
0.18
1.24
1.09
1.06
0.95
0.67
1.99
0.84
1.37
0.88
2.37
0.81
0.13
2.03
0.56
0.49
0.14
0.16
0.63
0.46
0.00
0.04
0.99
0.01
0.11
0.50
0.17
0.74
0.93
0.11
0.84
0.54
0.76
0.63
0.22
1.48
0.56
0.15
0.29
0.28
0.39
1.15
0.17
0.03
0.81
0.37
41
42
43
44
45
46
47
49
50
51
52
53
54
55
57
58
59
60
61
62
63
64
65
66
68
69
70
71
72
73
74
75
78
79
80
0.81
1.25
0.11
0.07
0.38
0.28
1.01
0.06
3.00
0.17
1.15
0.09
0.26
0.91
0.08
0.69
0.70
1.54
0.15
1.11
1.22
3.06
0.26
0.72
0.44
1.00
0.96
0.82
0.88
0.81
1.69
0.99
1.70
0.92
1.34
0.78
1.15
0.07
0.35
0.16
0.13
0.63
0.16
2.34
1.28
1.61
0.13
0.07
0.23
0.28
1.17
0.61
1.81
0.71
1.02
1.67
1.98
0.48
0.91
0.39
1.22
0.72
1.16
1.44
1.27
0.80
0.75
1.84
0.66
1.27
0.03
0.10
0.04
0.43
0.54
0.42
0.38
0.11
0.66
1.11
0.46
0.04
0.33
0.68
0.20
0.48
0.09
3.35
0.56
0.09
0.45
1.08
0.22
0.20
0.06
0.22
0.24
0.34
0.56
0.47
0.89
0.24
0.14
0.27
0.08
3
6
7
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
29
30
31
32
33
34
35
36
37
38
39
40

1114
Y. Kamano et al./Bioorg. Med. Chem. 6 (1998) 1103±1115
(green region) and/or decreases (yellow region) in steric
bulk could enhance growth inhibition of PLC cells.
Contour maps of the electrostatic ®eld contributions are
also provided in Figure 5.
ring and the side chain at C-3 lowers the cytotoxicity.
On the other hand, more bulky side chain at C-14 and
-15 of b con®guration enhances the activity, suggesting
that b type con®guration favors the activity more than
that of a type for the C-14 and/or -15 substituents of the
compounds.
The steric ®eld map indicated that presence of steric
bulk of a type con®guration at C-14 and -15 has a
measurable negative e€ect on the activity, as mentioned
in the SAR consideration as discussed above. Less bulk
favorable regions are indicated around the a-pyrone
In the electrostatic CoMFA map, the region where a
more positive electrostatic interaction might be expected
to enhance the growth inhibition activity, was encompassed
Figure 5. Stereoviews of CoMFA contour plots from the PLS analysis. Above: steric, below: electrostatic. Regions where increased
steric bulk is associated with enhanced activity are indicated in green, whereas regions where increased steric bulk is associated with
diminished activity are indicated in yellow. On the other hand, regions where increased positive charge is favorable for activity are
indicated in blue, whereas regions where increased negative charge is favorable are indicated in red.

Y. Kamano et al./Bioorg. Med. Chem. 6 (1998) 1103±1115
1115
around the a-face around the steroidal backbone skele-
ton. The regions of negative electrostatic interactions
were scattered around the ester side chains at C-3 and
-5, and the a-pyrone ring.
6. Kamano, Y.; Pettit, G. R.; Inoue, M.; Tozawa, M.; Komei-
chi, Y. J. Chem. Research(S), 1977, 78±79, and J. Chem.
Research(M), 1977, 0843±0859.
7. Pettit, G. R.; Kamano, Y.; Inoue, M.; Komeichi, Y.; Nas-
simbeni, L. R.; Niven, M. L. J. Org. Chem., 1982, 47, 1503±1505.
8. Kamano, Y.; Yamamoto, H.; Tanaka, Y.; Komatsu, M.
Tetrahedron Lett., 1986, 54, 5673±5676.
The CoMFA model derived from the above bufadieno-
lides was used to estimate the IC50 values of the test
bufadienolides, we synthesized in this study. The observed
and corresponding CoMFA-predicted IC50 values for
the test-set bufadienolides are as follows (observed:
predicted): bufalin 3-succinate (4): (1.18:1.41), bufalin 3-
methacrylate (5) (1.20:0.94), bufalin 3-[N-(tert-butoxy-
carbonyl) hydrazido] succinate (8) (1.12:0.94), Á1-3-
oxo-resibufogenin (27) (1.12:0.94), Á1,4-3-oxo-resibufo-
genin (28) (0.35:0.84), desacetyl-cinobufagin 3-acetate-
16-succinate (48) (1.12:0.94), 3-oxo-14a-artebufogenin
(56) (1.12:0.94), digitoxigenin 3-methylsuberate (67)
(1.12:0.94), CMKA (76) (1.12:0.94), and CMKB (77)
(1.12:0.94). These comparison shows that this model can
generally predict the activity of the active derivatives,
though it may need to test for inactive derivatives.
9. Kamano, Y.; Satoh, N.; Nakayoshi, H.; Pettit, G. R.; Smith,
C. R. Chem. Parm. Bull., 1988, 36, 326±332.
10. Kamano, Y.; Pettit, G. R. J. Am. Chem, Soc., 1972, 94,
8592.
11. Kamano, Y.; Pettit, G. R. J. Org. Chem., 1974, 39, 2629±
2631.
12. Pettit, G. R.; Kamano, Y. J. Chem., Soc., Perkin Trans. I,
1973, 725±727.
13. a) Hofer, P.; Linde, H.; Meyer, K. Tetrahedron Lett., 1959,
8. b) Hofer, P.; Linde, H.; Meyer, K. Helv. Chim. Acta., 1960,
43, 1950.
14. Kamano, Y.; Yamamoto, H.; Hatayama, K.; Tanaka, Y.;
Shinihara, M.; Komatsu, M. Tetrahedron Lett., 1968, 54, 5669±
5672.
15. Kamano, Y.; Hatayama, K.; Shinohara, M; Komatsu, M.
Chem. Pharm. Bull., 1971, 19(12), 2478±2484.
In summary, we established 3D-QSAR models by using
the CoMFA methodology for a set of 70 bufadienolides
that are known to inhibit the growth of primary liver
carcinoma PLC/PRF/5 cells in vitro. To identify a
common pharmacophore and to elucidate a common
binding model, DISCO was used to systematically
compare low-energy conformations of the analogues.
The results provided a correlation between the inhibi-
tory activity of these bufadienolides and the steric and
electrostatic ®elds around them. The QSAR models
reveal regions in three dimensional space around these
analogues. The models obtained in this study are rea-
sonably predictive, as indicated by the cross-validated r2
values, and also provided predictions which agreed well
with experimental values. This model should be useful
for the new drug designing of this series which possess
higher anity and selectivity.
16. Kamano, Y.; Kumon, S.; Arai, T.; Komatsu, M. Chem.
Pharm. Bull., 1973, 21(9), 1960±1964.
17. Kamano, Y.; Pettit, G. R.; Inoue, M.; Tozawa, M.; Smith,
C. R.; Weislder, D. J. Chem., Soc., Perkin Trans, 1988, 2037±
2041.
18. Kamano, Y. Chem. Pharm. Bull., 1969, 17(8), 1711±1719.
19. Pettit, G. R.; Kamano, Y. J. Org. Chem., 1972, 37, 4040±
4044.
20. Kamano, Y.; Dralsar. P.; Pettit, G. R.; Tozawa, M. Col-
lection Czechoslovak Chem. Commum., 1987, 52, 1325±1330.
21. Kamano, Y.; Pettit, G. R. Chem. Pharm. Bull., 1973, 21(4),
895±898.
22. Horiger, N.; Zivanov, D.; Linde, H.H.A.; Meyer, K. Helv.
Chim. Acta, 1970, 53, 1993.
23. Kamano, Y.; Pettit, G. R.; Brown P.; Inoue, M. Tetra-
hedron, 1975, 31, 2359±2361.
24. Kamano, Y.; Pettit, G. R.; Tozawa, M.; Komeichi, Y.;
Inoue, M. J. Org. Chem., 1975, 40, 2136±2138.
25. Kamano, Y.; Pettit, G. R. J. Chem. Soc., Perkin Trans,
1975, 1976±1978.
References
26. Kamano, Y.; Tozawa, M.; Pettit, G. R. J. Org. Chem.,
1975, 40, 793±795.
1. Yoshida, S.; Kamano, Y.; Sakai, T. Chem. Pharm. Bull.,
1976, 24, 1714±1717.
27. Kamano, Y.; Komatsu, M. Chem. Pharm. Bull., 1969,
17(8), 1698±1705.
2. Jing, Y.; Ohizumi, H.; Kawazoe, N. et al., Jpn. J. Cancer
Res., 1994, 85/645.
28. Kamano, Y.; Yamamoto, H.; Komatsu, M. Chem. Pharm.
Bull., 1969, 17(6), 1246±1250.
3. Martin, Y. C.; Bures, M. G.; Danaher, E. A.; DeLazzer, J.;
Lico, I.; Pavlik, P. A., J. Computer-Aided Mol. Design, 1993, 7,
83.
29. Kamano, Y.; Tanaka, Y.; Komatsu, M. Chem. Pharm.
Bull., 1969, 17(8), 1706±1710.
4. Cramer, R. D.; Patterson, D. E.; Bunce, J. D. J. Am. Chem.
Soc., 1988, 110, 5959±5967.
30. Gasteiger, J.; Marsili, M. Tetrahedron, 1980, 36, 3219±
3228.
5. Kamano, Y.; Kagaku No Rhoiki, 1970, 24, 339±354 and
421±432.
31. Stahle, L.; Wold, S. Progr. Med. Chem., 1988, 25, 292±334.