Effects of tannins from Geum japonicum on the catalytic activity of thrombin and factor Xa of blood coagulation cascade

Article abstract of DOI:10.1021/np9801458

Bioassay-guided fractionation of the MeOH extract of the whole plant of Geum japonicum led to the isolation of seven known tannins. They were identified by spectroscopic methods as penta-O-galloyl-β-glucoside (1); pedunculagin (2), 2,3-(S)-hexahydroxydiphenoyl-D-glucose (3), tellimagrandin II (4), 2,6-di-O-galloyl-D-glucose (5), casuariin (6), and 5- desgalloylstachyurin (7). Compounds 1, 2, 4, 6, and 7 showed potent anticoagulant activity by significantly prolonging the clotting of rabbit plasma. The inhibitory effect of 2 was competitively directed against thrombin. Its IC₅₀ values for inhibition of the enzymatic activity of thrombin on synthetic substrate and fibrinogen were 0.18 and 0.15 μM, respectively. On the other hand, compounds 1, 4, 6, and 7 are mixed noncompetitive inhibitors of thrombin. Their IC₅₀ values for inhibition of fibrinogen hydrolysis were twofold to sevenfold lower than those for the inhibition of synthetic substrate hydrolysis. Factor Xa was competitively inhibited by compounds 1, 2, 4, 6, and 7. The phenolic hydroxyl groups of the active tannins appear to play an important role in their inhibitory effect on the enzymes.

Full text of DOI:10.1021/np9801458

1356
J . Nat. Prod. 1998, 61, 1356-1360
Effects of Ta n n in s fr om Geu m ja pon icu m on th e Ca ta lytic Activity of Th r om bin
a n d F a ctor Xa of Blood Coa gu la tion Ca sca d e
Hui Dong,† Shao-Xing Chen,†,§ R. Manjunatha Kini,*,† and Hong-Xi Xu‡
Bioscience Centre, Faculty of Science, National University of Singapore, 10 Kent Ridge Crescent, Singapore 119260, Singapore,
and Department of Chemistry, Dalhousie University, Halifax, Nova Scotia B3H 3J 5, Canada
Received April 13, 1998
Bioassay-guided fractionation of the MeOH extract of the whole plant of Geum japonicum led to the
isolation of seven known tannins. They were identified by spectroscopic methods as penta-O-galloyl-â-
glucoside (1), pedunculagin (2), 2,3-(S)-hexahydroxydiphenoyl-D-glucose (3), tellimagrandin II (4), 2,6-
di-O-galloyl-D-glucose (5), casuariin (6), and 5-desgalloylstachyurin (7). Compounds 1, 2, 4, 6, and 7
showed potent anticoagulant activity by significantly prolonging the clotting of rabbit plasma. The
inhibitory effect of 2 was competitively directed against thrombin. Its IC50 values for inhibition of the
enzymatic activity of thrombin on synthetic substrate and fibrinogen were 0.18 and 0.15 µM, respectively.
On the other hand, compounds 1, 4, 6, and 7 are mixed noncompetitive inhibitors of thrombin. Their
IC50 values for inhibition of fibrinogen hydrolysis were twofold to sevenfold lower than those for the
inhibition of synthetic substrate hydrolysis. Factor Xa was competitively inhibited by compounds 1, 2,
4, 6, and 7. The phenolic hydroxyl groups of the active tannins appear to play an important role in their
inhibitory effect on the enzymes.
Blood coagulation is a complex process involving initia-
tion, the formation of the fibrin clot and platelet plug, and
localization of the clot to the area of vascular injury.
cascade by catalyzing the conversion of fibrinogen to fibrin,
leading to fibrin thrombus formation. A safe and effective
inhibitor of thrombin could be a useful tool in the treatment
of venous thrombosis, arterial fibrillation, restenosis, and
arterial thrombosis and in the prevention of myocardial
infarction. Because of this, the modulation of thrombin by
direct, small molecule inhibitors is a widely sought goal in
the development of anticoagulant agents.1 The regulation
of thrombosis by using thrombin inhibitors has been
studied extensively. All these inhibitors reported so far
are either peptides or proteins from both natural and
synthetics sources.2,3
The whole plant of Geum japonicum Thunb. (Rosaceae)
has been used as a diuretic and an astringent in China
and J apan.4 Several tannins and anti-HIV triterpenes
have been isolated from this plant.4-6 In the course of
screening for anticoagulant agents from plants, we found
that the methanol extract of the whole plant of G. japoni-
cum exhibited potent anticoagulant activity. In this paper,
we report the isolation of seven tannins from G. japonicum
and evaluate their anticoagulant effects, particularly on
thrombin and factor Xa activities.
Resu lts a n d Discu ssion
Seven known tannins, penta-O-galloyl-â-D-glucose (1),7
pedunculagin (2),8 2,3-(S)-hexahydroxydiphenoyl-D-glucose
(3),9 tellimagrandin II (4),8 2,6-di-O-galloyl-D-glucose (5),10
casuariin (6),8,11 and 5-desgalloylstachyurin (7)12 were
isolated from the methanol extract of the whole plant.
Their structures were determined by spectroscopic methods
and by comparison with data reported in the literature.
The anticoagulant activity of these seven tannins was
determined by measuring their ability to prolong clotting
of rabbit plasma. Compound 1 significantly prolonged the
clotting times of rabbit plasma over 300 s in the prothrom-
bin time (PT), the thrombin time (TT), and the activated
partial thromboplastin time (APTT) assays with concentra-
tions of 0.45, 0.50, and 0.47 mM, respectively (Figure 1).
Compounds 2, 4, 6, and 7 also showed significant antico-
agulant activity with clotting times greater than 300 s in
Thrombin plays a central role in the blood coagulation
* To whom correspondence should be addressed. Tel.: 0065-8745235.
Fax: 0065-7770052. E-mail: bsckinim@leonis.nus.edu.sg.
† National University of Singapore.
‡ Dalhousie University.
§ Present address: Department of Scientific Service, Institute of Science
and Forensic Medicine, 11 Outram Road, Singapore 169078, Singapore.
10.1021/np9801458 CCC: $15.00
© 1998 American Chemical Society and American Society of Pharmacognosy
Published on Web 10/03/1998

Coagulation Effects of Tannins in Geum japonicum
J ournal of Natural Products, 1998, Vol. 61, No. 11 1357
positive control, heparin, showed significant anticoagulant
effects in APTT and TT assays rather than in the PT assay,
with a lower concentration (1 (g/mL) (Table 1). These
results are in accordance with those reported previously.13
As a first step in characterizing the anticoagulant effects
of these tannins, their effect on proteolytic activity of
thrombin on synthetic peptide substrate was studied. The
IC50 values of 1-7 are presented in Table 2. Compounds
1, 2, 4, 6, and 7 exhibited a dose-dependent inhibition of
thrombin-catalyzed hydrolysis of the chromogenic sub-
strate, H-D-Phe-Pip-Arg-p-nitroanilide (S-2238). Among
them, compound 4 showed the most potent activity, with
an IC50 of 0.070 µM. Compounds 6 and 7 are anomers with
a different configuration at the C-1 position of the sugar
moiety. The inhibitory activity of compound 7 on thrombin
is about twofold more potent than that of compound 6
(Student’s t-test, p < 0.05). Under the same experimental
conditions, compounds 3 and 5 exhibited no measurable
inhibition of thrombin-catalyzed reaction at a higher
concentration of 200 µM.
F igu r e 1. Effect of 1 on prothrombin time (PT, O), thrombin time
(TT, 9), and activated partial thromboplastin time (APTT, 0) clotting
assays. Each point is reported as mean (seconds) ( SD (n ) 3-5); * p
< 0.05, ** p < 0.01, significantly different from control, Student’s t-test.
*** clotting time > 300 s.
The kinetic studies of compounds 1, 2, 4, 6, and 7 were
conducted to determine the type of inhibition and the
inhibition constants (Ki). Compound 2 competitively in-
hibited S-2238 hydrolysis by bovine thrombin, with a Ki
value of 0.042 µM. On the other hand, compounds 1, 4, 6,
and 7 showed mixed noncompetitive inhibition of S-2238
hydrolysis by bovine thrombin, with Ki values of 3.9, 0.42,
0.98, and 0.83 µM, respectively. These results indicate that
compound 2 binds to the active site of thrombin, whereas
the PT, TT, and APTT assays at a concentration of 1 mM.
On the other hand, compounds 3 and 5 had no effect on
plasma clotting at concentrations up to 2 mM (Table 1).
These results indicated that compounds 1, 2, 4, 6, and 7
inhibited either both intrinsic and extrinsic pathways or
the common pathway in the blood coagulation cascade. The
Ta ble 1. Effects of 1-7, and Heparin on Prothrombin Time (PT), Thrombin Time (TT), and Activated Partial Thromboplastin Time
(APTT) Assays In Vitroa
compounds (mM)
PT (s)
TT (s)
APTT (s)
42.6
>300
50.6 ( 0.9
>300
60.2 ( 12.0
43.7 ( 2.3
35.3 ( 6.5
39.9 ( 4.7
>300
control
b1 (0.5)
(0.25)
2 (1.0)
(0.75)
(0.50)
(0.25)
3 (2.0)
4 (1.0)
(0.50)
5 (2.0)
6 (1.0)
(0.5)
7 (1.0)
(0.5)
10.2 ( 0.1
>300c
13.9 ( 0.8**
>300
15.2 ( 1.4**
13.5 ( 0.2**
10.2 ( 1.2
10.0 ( 0.2
>300
11.9 ( 0.1**
11.1 ( 0.2
>300
13.6 ( 0.3**
>300
13.9 ( 0.9**
33.6 ( 0.5**
23.7 ( 0.2**
10.1 ( 0.3
11.2 (0.3
>300
16.5
>300
33.5 ( 2.4**
22.7 ( 1.6**
12.6 ( 1.5
9.43 ( 0.1
>300
18.7 ( 0.6**
10.7 ( 0.1
>300
45.7 ( 4.1
40.8 ( 3.2
>300
12.3 ( 0.3*
>300
72.3 ( 8.0*
>300
16.2 ( 0.9**
>300
31.6 ( 2.1
>300
dheparin, 500 µg/mL
250 µg/mL
1 µg/mL
>300
>300
60.4 ( 6.7**
63.2 ( 6.3*
a
PT, TT, and APTT were determined in the presence of the indicated concentrations of 1-7. Data are reported as clotting times
b
(seconds) of plasma (mean ( SD, n ) 3-5); * p < 0.05, ** p < 0.01 significantly different from control, Student’s t-test. Other data
shown in Figure 1. c No clotting was observed (clotting time >300 s). Heparin sodium salt from porcine intestinal, 187 USP units/mg.
d
Ta ble 2. Inhibitory Activities of Compounds 1-7 on Thrombin-Catalyzed Hydrolysis of S-2238 and Fibrinogena
IC50 (µM)
type of
compounds
for S-2238
for fibrinogen
inhibition
Ki (µM)
1
2
3
4
5
6
7
0.11 ( 0.02b
0.18 ( 0.01
>200
0.050 ( 0.006*
0.15 ( 0.03
>200
MNCc
CCd
3.9 ( 0.6
0.042 ( 0.004
N. D.e
N. D.e
MNCc
N. D.e
MNCc
MNCc
0.070 ( 0.01
>200
0.014 ( 0.001**
>200
0.42 ( 0.06
N. D.e
0.28 ( 0.02
0.19 ( 0.02
0.059 ( 0.002*
0.026 ( 0.002**
0.98 ( 0.2
0.83 ( 0.2
a
IC50 values were the concentration inhibiting 50% of the initial activities of enzymes; Ki values were determined as reported under
b
Experimental Section. The data presented are those obtained from the analysis with best fitting; shown as mean ( SD (n ) 3); * p <
0.05, ** p < 0.01 significantly different from their IC50 values for S-2238 substrate, Student’s t-test. c MNC: mixed noncompetitive. CC:
d
classical competitive. e Data not detected.

1358 J ournal of Natural Products, 1998, Vol. 61, No. 11
Dong et al.
F igu r e 2. Effect of 1 (O), 2 (b), 4 (4), 6 (0), and 7 (9) on the inhibitory
activity of thrombin-induced fibrin-clotting. Points shown are means
of triplicate ( standard derivation.
compounds 1, 4, 6, and 7 bind to two different sites, at both
the active site of thrombin and a secondary site.
It is known that thrombin in the blood coagulation
cascade catalyzes the conversion of fibrinogen to fibrin
clots. Therefore, the effect of these tannins on the pro-
teolytic activity of thrombin on its macromolecular sub-
strate, fibrinogen, was examined. Compounds 1, 2, 4, 6,
and 7 exhibited a dose-dependent inhibition of thrombin-
catalyzed conversion of fibrinogen to fibrin clots (Table 2,
Figure 2). Compounds 1, 4, 6, and 7, which exhibited
mixed noncompetitive inhibition, showed a decrease (two-
fold to sevenfold) in their IC50 values of inhibition of fibrin
clot formation (Table 2) compared to those of inhibition of
S-2238 hydrolysis. This indicates that the second binding
site is most probably the anion binding exosite (ABE)14 of
thrombin. On the other hand, the IC50 values of compound
2 for hydrolysis of synthetic and macromolecular substrates
were not significantly different (Student’s t-test, p > 0.05).
This is expected, as compound 2 is a competitive inhibitor
of thrombin and only binds to the active site of thrombin.
Among these active compounds, 4 showed the most potent
inhibitory activity on the thrombin-induced fibrin clotting,
with an IC50 of 0.014 µM.
F igu r e 3. Schematic representation of interaction of tannins with
thrombin (A) and factor Xa (B). Compound 2 binds only to the active
site of thrombin, whereas compounds 1, 4, 6, and 7 bind to both the
active site and ABE (A). On the other hand, all the compounds bind
only to the active site of factor Xa (B).
Ta ble 3. Inhibitory Activities of Compounds 1, 2, 4, 6, and 7
on Factor Xa-Catalyzed Hydrolysis of S-2222a
type of
compounds
IC50 (µM)
inhibition
Ki (µM)
1
2
4
6
7
0.17 ( 0.08
0.56 ( 0.05
0.21 ( 0.02
0.72 ( 0.3
0.28 ( 0.04
CCb
CCb
CCb
CCb
CCb
0.57 ( 0.08
0.99 ( 0.04
0.44 ( 0.1
0.45 ( 0.05
0.69 ( 0.04
a
Ki values were determined as reported under Experimental
Section. The data presented are shown as mean ( SD (n ) 3).
b
CC: classical competitive.
To test the specificity of these tannins, their effects on
proteolytic activity of factor Xa were determined. Com-
pounds 1, 2, 4, 6, and 7 inhibited factor Xa-catalyzed
hydrolysis of synthetic substrate S-2222 (N-benzoyl-L-
isoleucyl-L-arginine-p-nitroaniline hydrochloride and its
methyl ester). Their IC50 values were 0.17, 0.56, 0.21, 0.72,
and 0.28 µM, respectively. These values are 1.5 to 3 times
less than those for the inhibition of thrombin, indicating
only a small preference for thrombin over factor Xa. The
kinetics of inhibition of factor Xa for these tannins was also
examined. Compounds 1, 2, 4, 6, and 7 are competitive
inhibitors of factor Xa, indicating that they bind only to
the active site of the enzyme. Compounds 1, 4, 6, and 7,
which apparently bound to ABE and the active site of
thrombin (see above), could bind only to the active site of
factor Xa, inasmuch as they are similar in nature. Com-
pound 2, however, binds to the active site of both enzymes.
Thus, these results can be explained by the fact that factor
Xa, unlike thrombin, does not have ABE (Figure 3). The
Ki value of compound 2 for inhibition of factor Xa was 0.99
µM, and about 23-fold more than that of thrombin. In
contrast, compounds 1, 6, and 7 showed that the Ki values
for inhibition of thrombin are about 1.2-6.8 times that of
factor Xa (Tables 2, 3). Compound 4 competitively inhib-
ited factor Xa, with similar Ki values for inhibition of
thrombin and factor Xa.
The formation of tannin-enzyme complexes are usually
mediated by hydrogen bonding between phenolic hydroxyl
groups and polar groups in the active site of enzyme.15 To
determine the role of the hydroxyl groups of phenolic
moieties, compound 1a , with no phenolic hydroxyl groups,
as well as the methylated derivatives, 1b and 2a , were
synthesized. These three compounds showed no inhibition
of thrombin at a concentration up to 100 µM. These results
indicate that the phenolic hydroxyl groups of the tannins
play a key role in their inhibitory effect on thrombin.
Tannins are well known for their physiological and
pharmacological actions.16 Here, the ability of tannins from
G. japonicum to inhibit two key serine proteinases of the
blood coagulation cascade, thrombin and factor Xa, is
demonstrated. The seven tannins studied here can be
classified into four types, namely gallotannins (1 and 5),
ellagitannins (2 and 3), C-glycoside ellagitannins (6 and
7), and 4, which bears both the galloyl and the hexahy-

Coagulation Effects of Tannins in Geum japonicum
J ournal of Natural Products, 1998, Vol. 61, No. 11 1359
of column chromatography on Sephadex LH-20 and cellulose
of fraction II, eluted with 50% MeOH, using CHCl3-MeOH
(1:1), MeOH, 3% HOAc as eluents, followed by purification on
Bio-Gel P-4 (extra fine, <45 µM) column with 0.5% HOAc,
results in the isolation of compounds 2 (60.8 mg), 3 (36.2 mg),
5 (9.6 mg), 6 (8.1 mg), and 7 (10.2 mg). Fraction III, eluted
with MeOH, was chromatographed over cellulose with MeOH-
HOAc-H2O (5:5:90) and then over Toyopearl HW-40s with
30% aqueous MeOH to afford 1 (120 mg) and 4 (9 mg).
Of these compounds, the assignments of the 1H or 13C NMR
spectral data of the sugar moieties of 68,20 and 712 were
different from the data reported in the literature. The 1H and
13C NMR data of the sugar moiety were reassigned with the
help of DEPT, 1H-1H COSY, HMQC, and HMBC spectra data.
The 13C NMR data of the sugar moieties of 6: (CD3COCD3,
175 MHz,) δ 66.1 (C-1), 81.7 (C-2), 73.8 (C-3), 77.0 (C-4), 69.6
(C-5), 68.7 (C-6); and 7: (CD3COCD3, 175 MHz) δ 68.8 (C-1),
77.5 (C-2), 71.6 (C-3), 78.0 (C-4), 69.2 (C-5), 68.9 (C-6). 1H
NMR data of 7: (CD3COCD3, 500 MHz) δ 5.64 (1H, d, J ) 4.8
Hz, H-1), 5.45 (1H, t, J ) 2.5 Hz, H-3), 5.01 (1H, dd, J ) 2.8,
8.6 Hz, H-4), 4.72 (1H, dd, J ) 2.3, 4.8 Hz, H-2), 4.64 (1H, dd,
J ) 3.2, 12.3 Hz, H-6), 4.12 (1H, dd, J ) 2.5, 8.6 Hz, H-5),
3.83 (1H, d, J ) 11.8 Hz, H-6).
Syn th esis of P en ta-O-ben zoyl-â-D-glu copyr an oside (1a).
A solution of 1 g (5.6 mmol) of D(+)-glucopyranoside, 3.6 g (29.5
mmol) of benzoic acid, 0.33 g (2.7 mmol) of DMAP (4-
dimethylaminopyridine), 0.41 g (2.7 mmol) of DMAP‚HCl, and
3.7 g (17.9 mmol) of DCC (1,3-dicyclohexylcarbodiimide) in 100
mL of CH2Cl2 was purged under N2 and then refluxed for 24
h. The reaction mixture was cooled to room temperature and
filtered. Solvent was removed to give a white foam. The crude
product was purified by chromatography on Si gel (CH2Cl2) to
afford 1.5 g of penta-O-benzyol-â-D-glucopyranoside (1a ). ES-
IMS: 723.5 [M + Na]+; 1H NMR (CDCl3, 500 MHz) δ 6.34 (1H,
d, J ) 8.4 Hz, Glu H-1), 6.09 (1H, t, J ) 9.4 Hz, Glu H-3), 5.90
(1H, t, J ) 9.8 Hz, Glu H-4), 5.73 (1H, dd, J ) 3.7, 10.2 Hz,
Glu H-2), 4.43-4.68 (3H, m, Glu H-5, 6), 7.31-8.07 (25H, m,
aromatic protons).
droxydiphenoyl (HHDP) groups. Among the active com-
pounds, only the 2,3- and 4,6-coupled ellagitannin, pedun-
culagin (2), showed competitive inhibition on thrombin. The
hydroxyl groups of the sugar moiety are highly acylated
by either galloyl or HHDP groups in these active com-
pounds. In contrast, compound 3, having only two hydroxyl
groups acylated by two galloyl groups at C-2,C-6, and
compound 5, with one HHDP group at C-2, C-3, are
inactive. Similarly, compounds 1a , 1b, and 2a , bearing no
phenyl hydroxyl groups, but with methylated hydroxyl
groups, also showed no measurable inhibition on thrombin
at a higher concentration of 100 µM. It is indicated that
phenolic hydroxyl groups of the active tannins play an
important role in their inhibitory activity on thrombin.
From a pharmaceutical point of view, compound 2 can be
considered as a good direct thrombin inhibitor for control-
ling thrombosis because of its potent anticoagulant activity,
relative selectivity, and lower toxicity, with an LD50 of >100
mg/kg p.o. in mice and rats.17 The total synthesis of
compound 2 has recently been reported.18 Although the
bioactivities of tannins have been studied extensively,19 this
is the first report of the effects of tannins from a plant
source on the enzymes of blood coagulation cascade.
Further work on the selective effects of other serine
proteinases and on the effects of the blood coagulation
system in vivo should be evaluated individually.
Exp er im en ta l Section
Gen er a l Exp er im en ta l P r oced u r es. ESIMS were col-
lected on a PE SCIEX API-300 mass spectrometer. All spectra
(1H NMR, 13C NMR, COSY, HMQC, and HMBC) were recorded
on a Bruker AMX 500 spectrometer (500 MHz for 1H and 125
MHz for 13C), and the chemical shifts are reported in parts
per million using TMS as an internal standard. TLC was done
on HPTLC-Fertigplaten Cellulose F (Merck) plates in the
solvent systems: (a) H2O and (b) 7% aqueous HOAc.
Ma ter ia l a n d Rea gen ts. The plant of Geum japonicum
was collected from Guizhou Province of the People’s Republic
of China in August 1994. It was identified by Dr. Dao-Feng
Chen, Department of Pharmacognosy, Shanghai Medical
University, People’s Republic of China. A voucher specimen
is deposited in the Department of Pharmacognosy, Shanghai
University of Traditional Chinese Medicine, People’s Republic
of China.
Heparin sodium salt (from porcine intestinal mucosa, 187
USP units/mg), thromboplastin with calcium reagent (for PT
test), TT reagent, APTT reagent, thrombin from bovine plasma
(3 NIH units/vial), and fibrinogen, fraction I, type IV, from
bovine plasma were purchased from Sigma Chemical Co. (St.
Louis, MO). Human Factor Xa was purchased from ICN
Biomedicals Inc. (Aurora, OH). Substrates S-2238 (H-D-Phe-
Pip-Arg-p-nitroanilide) and S-2222 (N-benzoyl-L-isoleucyl-L-
arginine-p-nitroaniline hydrochloride and its methyl ester)
were obtained from Chromogenix Co. (Mo¨lndal, Sweden).
(Sweden).
Extr a ction a n d Isola tion . Dried whole plant (2.6 kg) was
chopped into small pieces and percolated three times with
MeOH (20 L) at room temperature, and the extract was
evaporated in vacuo to yield the MeOH dried extract (350 g).
This latter was resuspended in distilled H2O (1 L) and
successively partitioned with hexane (2 L (5), EtOAc (2 L ×
5) and n-BuOH (1 L × 5). The n-BuOH-soluble fraction was
filtered, and the filtrate evaporated under reduced pressure
(50 °C) to give a brown residue (73 g). The n-BuOH extract
was subjected to Sephadex LH-20 column chromatography and
eluted with H2O-MeOH starting with H2O (increasing amount
of MeOH), giving nine fractions, fractions 1-9. Fraction 9,
eluted with 80% MeOH (3.3 g), showed potent anticoagulant
activity. It was further chromatographed over Sephadex LH-
20, eluted with the above solvent system, to give three
fractions, I (0.08 g), II (1.7 g), and III (1.2 g). A combination
Meth yla tion of 1. An EtOH solution (0.5 mL) of 1 (5 mg)
was treated with ethereal CH2N2 (1.5 mL) for 1 h at room
temperature. The solvent was evaporated off, and the residue
was subjected to preparative TLC (CH2Cl2-CH3COCH3, 7:3)
to give a pentadecamethylate (1b) (2.1 mg, Rf ) 0.6). ESIMS:
1173.5 [M + Na]+; 1H NMR (CD3COCD3, 500 MHz) δ 7.36 (2H,
s, galloyl), 7.32 (2H, s, galloyl), 7.26 (2H, s, galloyl), 7.25 (2H,
s, galloyl), 7.21 (2H, s, galloyl), 6.41 (1H, d, J ) 8.1 Hz, Glu
H-1), 6.21 (1H, t, J ) 9.7 Hz, Glu H-3), 5.90 (1H, t, J ) 9.7
Hz, Glu H-4), 5.78 (1H, dd, J ) 8.1, 9.4 Hz, Glu H-2), 4.87
(1H, dd, J ) 3.1, 12.1 Hz, Glu H-6), 4.74 (1H, m, Glu H-5),
4.44 (1H, dd, J ) 5.3, 12.1 Hz, Glu H-6), 3.91 (9H, 3 × OCH3),
3.89 (6H, 2 × OCH3), 3.88 (3H, OCH3), 3.83 (6H, 2 × OCH3),
3.82 (9H, 2 × OCH3), 3.80 (3H, OCH3), 3.79 (3H, OCH3), 3.76
(3H, OCH3), 3.73 (3H, OCH3), and 3.72 (3H, OCH3).
Meth yla tion of 2. An EtOH solution (0.5 mL) of 1 (6 mg)
was treated with ethereal CH2N2 (1.5 mL) for 2 h at room
temperature. The solvent was evaporated off, and the residue
was subjected to preparative TLC (CH2Cl2-CH3COCH3 7:3)
to give an isomer mixture of trideca-O-methylpedunculagin
(2a ) (1.6 mg). ESIMS: 989.5 [M + Na]+; 1H NMR (CD3COCD3,
500 MHz) δ 6.96 (1H, s, HHDP), 6.93 (1H, s, HHDP), 6.89 (1H,
s, HHDP), 6.85 (2H, s, HHDP), 6.81 (1H, s, HHDP), 6.62 (1H,
s, HHDP), 6.60 (1H, s, HHDP), 5.53 (1H, t, J ) 10.5 Hz, Glu
H-3), 5.28 (1H, t, J ) 9.8 Hz, Glu H-3), 5.24-5.08 (5H, m, Glu
H-1R, 2R, 2â, 6 × 2), 5.02 (1H, t, J ) 10.1 Hz, Glu H-4), 4.85
(1H, t, J ) 8.7 Hz, Glu H-4), 4.88 (1H, d, J ) 6.4 Hz, Glu H-1â),
4.65 (1H, m, Glu H-5), 4.29 (1H, m, Glu H-5), 4.13 (2H, m,
Glu H-6), 3.61-3.93 (78H, m, 26 × OCH3)
Biologica l Assa ys. Blood -clottin g a ssa ys:21 The effects
of these compounds on coagulation of rabbit plasma in vitro
were measured at 37 °C on a fibrometer (BBL FibroSystem,
Becton Dickinson Co., Cockeysville, MD) using the PT assay;
150 µL of 0.05 M Tris buffer containing 0.1 M NaCl (pH 7.5),
50 µL of sample solution (compound dissovled in 0.05 M Tris
buffer), and 100 µL rabbit plasma were incubated for 2 min
at 37 °C. Clotting was initiated by addition of 100 µL of

1360 J ournal of Natural Products, 1998, Vol. 61, No. 11
Dong et al.
thromboplastin with calcium reagent (Sigma). Similarly, the
TTs were measured by addition of 100 µL bovine thrombin
(Sigma, 3-4 NIH units/mL) to a mixture of rabbit plasma (100
µL), 0.05 M Tris buffer (50 µL), and compounds dissolved in
Tris buffer (50 µL), which was preincubated for 2 min. APTT
was measured by incubating rabbit plasma (100 µL), 0.05 M
Tris buffer (50 µL), and compounds dissolved in Tris buffer
(50 µL) with APTT reagent (100 µL) for 2 min. Clotting was
initiated by adding 25 mM CaCl2 (100 µL). Clotting times are
based on the average of three separate determinations.
Deter m in a tion of th e in h ibitor y a ctivity of ta n n in s for
th r om bin :13 The activity of thrombin from bovine plasma was
measured using S-2238 as substrate on a Ceres 900C Micro-
plate Autoreader (Bio-Tek Instruments, Inc., Winooski, VT).
The assay mixture of 200 µL comprises 10 µL thrombin (0.19
NIH units/mL) in 0.05 M Tris buffer containing 0.1 M NaCl
(pH 7.5) and 10 µL of various concentrations of compounds,
which was preincubated with thrombin for 2 min at 37 °C.
The reaction was started by the addition of 20 µL S-2238 (0.66
mM). Activity against the substrate was determined by
measurement of the release of p-nitroaniline from S-2238 as
indicated by the increase in optical density at 405 nm over 5
min. The results were expressed as percentage of inhibition
of thrombin activity. The kinetics of thrombin from bovine
plasma were studied, with a range of substrate concentrations
from 0.6 to 1.7 mM.22 Assays were performed by using a 0.05
M Tris buffer containing 0.1 M NaCl (pH 7.5) and a fixed
concentration of thrombin from bovine plasma (0.19 NIH units/
mL) at 37 °C. The total volume of reaction mixture was
maintained at 200 µL. Reactions were initiated by addition
of substrate to the cuvette containing thrombin and compound
preincubated for less than 2 min. Data from initial rate
experiments were used to construct Lineweaver-Burk plots,23
while the relationship of (substrate concentration)-1 versus
(initial velocity)-1 was analyzed by linear regression. The type
of inhibition and Ki values were determined from the plots.24
Compounds 1a , 1b, and 2a were tested in 0.1% DMSO in Tris
buffer containing 0.1 M NaCl (pH 7.5), which was used as
control.
with S-2222 as the substrate. The assay mixtures of 200 µL
contained 0.20 mM S-2222 in 0.05 M Tris buffer containing
0.1 M NaCl (pH 8.3) and various concentrations of tannins.
The reaction was started by addition of factor Xa (50 ng/mL).
Activity against the substrate was determined by measure-
ment of the release of 4-nitroanilide as indicated by the
increase in optical density at 405 nm over 5 min at 37 °C. For
determination of the Ki, increasing concentrations of S-2222
were added (0.04 mM-0.20 mM). The IC50 value was the
concentration inhibiting 50% of the initial activity of enzymes.
The data were collected as mean ( SD of three separate
experiments.
Ack n ow led gm en t. This work was supported by a grant
from the National University of Singapore (RP950389).
Refer en ces a n d Notes
(1) Kimball, S. D. Blood Coagulation Fibrinolysis 1995, 6, 511-519.
(2) Beretz, A.; Cazenave, T. P. Planta Med. 1991, 57 (suppl. 1), 68-72.
(3) Maraganore, J . M.; Bourdon, P.; J ablonski, J .; Ramachandran, K. L.;
Fenton, J . W. Biochemistry 1990, 29, 7095-7101.
(4) Yoshida, T.; Okuka, T. J . Chem. Soc., Perkin Trans. 1 1985, 315-
321.
(5) Yoshida, T.; Maruyama, Y.; Usman Memon, M.; Shingu, T.; Okuda,
T. Phytochemistry 1985, 5, 1041-1046.
(6) Xu, H. X.; Zeng, F. Q.; Wang, M.; Sim, K. Y. J . Nat. Prod. 1996, 59,
1-3.
(7) Yoshida, T.; Chen, L.; Shingu, T.; Okuda, T. Chem. Pharm. Bull. 1988,
8, 2940-2949.
(8) Okuda, T.; Yoshida, T.; Ashida, M.; Yazaki, K. J . Chem. Soc., Perkin
Trans 1 1983, 1765-1772.
(9) Tanaka, T.; Nonaka, G. I.; Nishioka, I. J . Chem. Res. (M) 1985, 2001-
2009.
(10) Lee, M. W.; Tanaka. T.; Nonaka, G. I.; Nishioka, I. Phytochemistry
1992, 8, 2835-2839.
(11) Nonaka, G.; Skai, T., Tanaka, T., Mihashi, K.; Nishioka, I. Chem.
Pharm. Bull. 1990, 8, 2151-2156.
(12) Lee, S. H.; Tanaka, T.; Nonaka, G.; Nishioka, I. Phytochemistry 1990,
11, 3625.
(13) Akoum, A.; J osefonvicz, J .; Sigot, M. Thromb. Res. 1990, 60, 9-18.
(14) Fenton, J . W., II; Ofosu, F. A.; Moon, D. G.; Maraganore, J . M. Blood
Coagulation Fibrinolysis 1991, 2, 69-75.
(15) Spencer, C. M.; Cai, Y.; Martin, R.; Gaffney, S. H.; Goulding, P. N.;
Magnolato, D.; Lilley, T. H.; Haslam, E. Phytochemistry 1988, 27,
2397-2409.
Th r om bin -in d u ced fibr in -clot a ssa ys:25 All solutions
used were prepared with 0.05 M Tris buffer containing 0.1 M
NaCl (pH 7.5), except for fibrinogen, which was reconstituted
in ultrapure H2O. For the assay, 190 µL of different concen-
trations of a compound were preincubated for 3 min with 10
µL of thrombin (0.05 NIH units/mL) at 37 °C, and then 50 µL
fibrinogen (2 mg/mL) were added to initiate enzymatic reac-
tion. The turbidity of clots was measured for 5 min at 405
nm, and velocities were calculated as milli-optical density units
per minute (mOD/min). Each assay was performed in tripli-
cate. The percentage of inhibition of thrombin-induced fibrin
clotting was calculated according to the formula: percentage
inhibition (%) ) [1 - (velocity of samples/velocity of control)]
× 100.
(16) Haslam, E. J . Nat. Prod. 1996, 59, 205-215.
(17) Ogata, T.; Kato, A. Chem. Abstr. 1993, 118, 45735q; J apan Patent
04, 264, 026A2, 18 Sep 1992.
(18) Feldman, K. S., Smith, R. S. J . Org. Chem. 1996, 61, 2606-2612.
(19) Okuda, T.; Yoshida, T.; Hatono, T. Planta Med. 1989, 55, 117-122.
(20) Hatano, T.; Yoshida, T.; Shingu, T.; Okuda, T. Chem. Pharm. Bull.
1988, 10, 3849-3856.
(21) Kettner, C.; Mersinger, L.; Knabb, R. J . Biol. Chem. 1990, 30, 18289-
18297.
(22) Furie, B.; Bing, D. H.; Feldman, R. J .; Robison, D. J .; Burnier, J . P.;
Furie, B. C. J . Biol. Chem. 1982, 257, 3875-3882.
(23) Lineweaver, H.; Burk, D. J . Am. Chem. Soc. 1934, 56, 658.
(24) Todhunter, J . A. Methods Enzymol. 1979, 63(A), 383-411.
(25) Rupin, A.; Mennecier, P.; Nanteuil, de G.; Laubie, M.; Verbeuren, T.
J . Thromb. Res. 1995, 3, 217-225.
(26) Herbert, J . M.; Bernat, A.; Dol, F.; Herault, J . P.; Crepon, B.; Lormeau,
J . C. J . Pharm. Exp. Therap. 1996, 276, 1030-1038.
Deter m in a tion of th e in h ibitor y a ctivity of ta n n in s for
fa ctor Xa :26 The activity of human factor Xa was measured
NP9801458