3,6-Disubstituted coumarins as mechanism-based inhibitors of thrombin and factor Xa

Article abstract of DOI:10.1021/jm050448g

In this work, coumarins were screened on thrombin (THR) and factor Xa (FXa), two of the most promising targets for the development of anticoagulant drugs. This allowed us to highlight compound 30, characterized by a 2,5-dichlorophenyl ester in the 3-position and a chloromethyl moiety in the 6-position, as a very potent THR inhibitor (k_i/K_I = 37 000 M^-1 s^-1). Moreover, this compound exhibits good selectivity over FXa (168-fold) and trypsin (54-fold). The mechanism of inactivation was investigated in this series and significantly differs from that previously observed with α-chymotrypsin. Indeed, the addition of hydrazine on the THR-inhibitor complex promotes a partial induced THR reactivation. This reactivation, confirmed by LC/MS, showed the resurgence of the native THR and a new dihydrazide complex. Docking experiments were then efficiently used to explain the trends observed in the enzymatic assays as well as to corroborate the postulated inhibition mechanism. Finally, the cell permeability of our derivatives was estimated using a computational approach.

Full text of DOI:10.1021/jm050448g

7592
J. Med. Chem. 2005, 48, 7592-7603
3,6-Disubstituted Coumarins as Mechanism-Based Inhibitors of Thrombin and
Factor Xa
Raphae¨l Fre´de´rick,†,‡ Se´verine Robert,†,‡ Caroline Charlier,§ Je´roˆme de Ruyck,§ Johan Wouters,§
Bernard Pirotte,# Bernard Masereel,† and Lionel Pochet*,†
Department of Pharmacy and CBS Laboratory, University of Namur, FUNDP, 61, Rue de Bruxelles, B-5000 Namur, Belgium,
and Natural and Synthetic Drug Research Centre, University of Lie`ge, ULG, 1, Avenue de l’Hoˆpital baˆt B36, Tour 4,
B-4000 Lie`ge, Belgium
Received May 11, 2005
In this work, coumarins were screened on thrombin (THR) and factor Xa (FXa), two of the
most promising targets for the development of anticoagulant drugs. This allowed us to highlight
compound 30, characterized by a 2,5-dichlorophenyl ester in the 3-position and a chloromethyl
moiety in the 6-position, as a very potent THR inhibitor (ki/KI ) 37 000 M-1 s-1). Moreover,
this compound exhibits good selectivity over FXa (168-fold) and trypsin (54-fold). The mechanism
of inactivation was investigated in this series and significantly differs from that previously
observed with R-chymotrypsin. Indeed, the addition of hydrazine on the THR-inhibitor complex
promotes a partial induced THR reactivation. This reactivation, confirmed by LC/MS, showed
the resurgence of the native THR and a new dihydrazide complex. Docking experiments were
then efficiently used to explain the trends observed in the enzymatic assays as well as to
corroborate the postulated inhibition mechanism. Finally, the cell permeability of our derivatives
was estimated using a computational approach.
Introduction
and (ii) 4-OH-coumarins, as warfarin and acenocouma-
rol,12 which inhibit the hepatic synthesis of vitamin-K
dependent proteins including THR and FXa. Owing to
their indirect and nonselective mechanism of action,
both classes of anticoagulants require careful clinical
monitoring. Moreover, they are often associated with
numerous drug interactions and side effects. Therefore,
many efforts have been devoted finding novel classes
of low molecular weight direct THR and FXa inhibi-
tors.13,14 These could be safer and more selective than
current therapies.
To date, two types of low molecular weight active site
directed THR inhibitors have been developed and clas-
sified according to their mechanism of action. The first
class inhibits THR in a noncovalent manner. These
compounds fill the enzyme active site and are stabilized
within the S1-S3 specificity pockets through hydrogen-
bonding, hydrophobic, and electrostatic interactions.
Representative compounds of this class include arga-
troban,15 NAPAP,16 inogatran,17 melagatran,18 and nap-
sagatran.19 The second class consists of inhibitors that
covalently interact with Ser195 of the catalytic triad.
In addition, these compounds usually possess an elec-
trophilic center such as a chloromethyl ketone as
observed in PPACK,20 a C-terminal aldehyde,21 a fluo-
romethyl ketone,22 a R-ketoester,23 an amide,24 a phos-
phonate,25 a phosphonite,25 or a boronic acid function.26
A large variety of small molecular weight active site
directed FXa inhibitors, distributed in two groups, have
also emerged in the literature: (i) covalent inhibitors
derived from the peptide substrate27,28 and (ii) nonco-
valent inhibitors such as the bis-arylamidine DX-
9065a29 or the bis-amidine ZK-807834.30 Whatever the
targeted enzyme, many of these inhibitors have a basic
P1 side chain to facilitate the formation of a salt-bridge
with Asp189 located at the bottom of the S1 pocket.
Thrombin (THR) and factor Xa (FXa) constitute
important trypsin-like serine proteases involved in the
coagulation cascade. THR plays a key role in thrombosis
and hemostasis.1-3 Indeed, as the terminal enzyme of
the coagulation cascade, THR is directly responsible for
the conversion of glycoprotein fibrinogen to fibrin.4-7
Moreover, besides the cleavage of fibrinogen, THR
amplifies its own generation by feedback activation of
factors V, VIII, and IX and stimulates platelet activa-
tion. In contrast, when bound to thrombomodulin, it
causes a feedback inhibition of the coagulation cascade
by activation of protein C. This, in turn, inactivates
factors Va and VIIIa by proteolytic degradation. Because
of its importance in thrombosis, THR has been recog-
nized early as a major target for the development of
inhibitors.
FXa is the point of convergence of the intrinsic and
extrinsic pathways. In association with phospholipids,
factor Va, and calcium, it forms the prothrombinase
complex and catalyzes the conversion of prothrombin
in THR. Therefore, FXa constitutes another key target
for the development of anticoagulants, which might be
even more efficient than THR inhibition in interrupting
the blood coagulation cascade.8-10
Current anticoagulant therapies clinically used in the
treatment of acute and chronic thrombosis involve the
use of (i) heparins and related compounds11 acting
through activation of antithrombin III, a glycoprotein
that nonselectively inhibits THR, FXa, and factor IXa,
* To whom correspondence should be addressed. Phone: +32 (0)81
72 42 92. Fax: +32 (0)81 72 42 99. E-mail: lionel.pochet@fundp.ac.be.
† Department of Pharmacy, University of Namur.
‡ These authors equally contributed to this work.
§ University of Lie`ge.
# CBS Laboratory, University of Namur.
10.1021/jm050448g CCC: $30.25 © 2005 American Chemical Society
Published on Web 11/05/2005

Coumarins as Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7593
Scheme 1. Synthesis of the Esters 6-32, Amide 33, and Thioester Derivatives 34a
a Reagents and conditions: (i) HCOH, HCl, 80 °C, 20 min; (ii) diethyl malonate, piperidine, AcOH, EtOH, reflux, 17 h; (iii) HCl (3 N),
EtOH, reflux, 3 h; (iv) SOCl2, reflux, 3 h; (V) R′OH or R′′NH2, or R′′′SH, C5H5N, dioxane, 25 °C, 1 h 30 min.
Table 1. In Vitro THR and FXa Inhibitory Potency of Aryl
Nevertheless, this basic moiety is usually responsible
Esters
for poor bioavailability. So several laboratories aimed
to develop less basic, low molecular weight, selective
inhibitors of THR and FXa.31-35
Previously, we reported that particular coumarins
were powerful inhibitors of R-chymotrypsin (R-CT) and
human leukocyte elastase (HLE), two serine proteases
residual activity (%)
compd
R
THR
FXa
implicated in numerous biological processes such as
digestion (R-CT) and phagocytosis (HLE).36-39 These
compounds are characterized by an alkyl or an aryl
moiety linked to the heterocycle by an ester, a thioester,
an amide, or a ketone function in the 3-position and an
electrophilic chloromethyl group in the 6-position. They
were found to act as mechanism-based inhibitors.
Interestingly, THR and FXa share the same catalytic
mechanism with R-CT and HLE. Moreover, it was
shown that a proper pharmacomodulation of isocou-
marins, initially developed as HLE inhibitors, could
afford selective THR inhibitors.40,41 Therefore, in this
paper, we have evaluated the efficiency of 3,6-disubsti-
tuted coumarins, characterized by a nonbasic side chain,
as potent THR and FXa inhibitors. For this purpose,
new compounds were synthesized. The postulated in-
hibition mechanism toward THR was investigated using
reactivation and mass spectrometry experiments (LC/
MS), and compared with that of R-CT. Molecular dock-
ing was performed on relevant compounds to corrobo-
rate our hypothesis and to explain the trends observed
in the enzymatic assays. Finally, we have estimated the
cell permeability of some derivatives using a computa-
tional approach. On the basis of the results as a whole,
it was shown that according to the nature of the
substituent in the 3-position, coumarins could act as
very potent THR inhibitors, with a partially induced
reactivation profile and good cell permeation.
6
7
phenyl
60 ( 9
59 ( 18
110 ( 6
89 ( 15
114 ( 15
109 ( 12
113 ( 5
125 ( 7
114 ( 11
134 ( 9
29 ( 3
benzyl
95 ( 5
8
2-pyridyl
3-pyridyl
1-naphthyl
2-naphthyl
6-quinolyl
8-quinolyl
100 ( 5
113 ( 11
115 ( 18
109 ( 14
98 ( 10
107 ( 14
78 ( 8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
o-chlorophenyl
m-chlorophenyl
p-chlorophenyl
o-iodophenyl
10 ( 2
107 ( 12
59 ( 13
27 ( 8
102 ( 6
92 ( 13
73 ( 9
m-iodophenyl
p-iodophenyl
99 ( 10
79 ( 3
35 ( 3
80 ( 8
o-toluyl
103 ( 12
78 ( 16
72 ( 8
m-toluyl
p-toluyl
111 ( 20
24 ( 6
m-fluorophenyl
m-bromophenyl
m-nitrophenyl
m-methoxyphenyl
m-trifluoromethylphenyl
3-chloro-5-methoxyphenyl
2,3-dichlorophenyl
2,5-dichlorophenyl
2,6-dichlorophenyl
3,5-dichlorophenyl
61 ( 16
33 ( 3
8 ( 3
80 ( 2
124 ( 32
108 ( 21
134 ( 13
126 ( 28
125 ( 22
39 ( 6
66 ( 4
66 ( 21
103 ( 5
100 ( 12
1.0 ( 0.1
101 ( 11
102 ( 10
109 ( 21
73 ( 8
gel type reaction gave the 6-hydroxymethyl-2-oxo-2H-
1-benzopyran-3-carboxylic acid ethyl ester 3, which was
further hydrolyzed to obtain 4. Reaction of 4 with
thionyl chloride yielded the acyl chloride 5, which
reacted with the required alcohol, amine, or thiol to give
the esters 6-32 (Table 1), the amide 33 (Table 2), and
the thioester 34 (Table 2).
The synthesis of ketones 40-41 followed the strategy
illustrated in Scheme 2. According to this pathway, we
prepared the â-ketoester 37 by reaction of the acyl
chloride 36 of 3-chlorophenylacetic acid 35 with Mel-
Chemistry
6-Chloromethylcoumarins 6-34 were obtained using
the strategy depicted in Scheme 1.36-38 5-Hydroxym-
ethylsalicylaldehyde 2 was achieved by reaction of
salicylaldehyde 1 with formaldehyde. Then the conden-
sation of 2 with diethyl malonate through a Knoevena-

7594 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24
Fre´de´rick et al.
Table 2. In Vitro THR and FXa Inhibitory Potency of
investigated (Table 2). For this purpose, we synthesized
compounds 40 and 41 characterized by a ketone link
between the coumarin ring and the aromatic side chain.
The inhibitory potency of these molecules was compared
with the corresponding ester 15, amide 33, and thioester
34. In each case, the replacement of the ester function
strongly decreases the inhibitory potency toward THR
and FXa. As observed for R-CT and HLE, the amide link
suppresses the inhibitory potency. We previously showed
that this amide function highly stabilizes an anti
conformation in which both carbonyl groups point in
opposite directions.43 Therefore, the lack of inhibitory
potency of the amide derivative 33 could be associated
with a poor recognition of this conformation in THR and
FXa binding sites. Surprisingly, the thioester 34 is found
to be inactive on THR at 2 µM and moderately active
on FXa at 5 µM, whereas this compound was as potent
as the ester on R-CT.38 The new synthesized ketones
(40, 41) are less active than the corresponding ester 15.
The ketone 40 is inactive at 2 µM on THR and not very
active on FXa at 5 µM, while the methylene ketone 41
is fairly active on both proteases.
Thioester, Amide, and Ketone Derivatives
residual activity (%)
compd
Y
THR
FXa
15
33
34
40
41
COO
CONH
COS
10 ( 2
29 ( 3
89 ( 7
60 ( 7
83 ( 2
77 ( 3
97 ( 2
107 ( 12
100 ( 15
77 ( 5
CO
COCH2
drum’s acid followed by decarboxylation in refluxing
ethanol. Then the ketone derivatives 38-39 were
obtained by condensation of 2 with 37 or commercially
available 3-(3-chlorophenyl)-3-oxopropionic acid ethyl
ester. Derivative 40 was achieved through the reaction
of 38 with refluxing thionyl chloride. For the conversion
of alcohol 39 into the alkyl halide 41, we performed the
one-pot procedure described by Iranpoor.42 This method,
which employs the PPh3/DDQ/(n-hexyl)4NCl system,
allowed us to obtain the chloromethyl 41.
We also considered the influence of the moiety intro-
duced in the 6-position (Table 3). The removal (43) or
the substitution of a methyl (42) or an acetoxymethyl
(44) moiety for the chloromethyl group (15) strongly
decreases the inhibitory potency on both proteases.
These results suggest that the chloromethyl function
plays a key role in the inhibition process, as postulated
in the inhibition mechanism proposed for other serine
proteases.44 From earlier work, it appeared that the
importance of this moiety depends on the targeted
enzyme: the replacement of the chloromethyl leads to
a complete loss of the activity toward R-CT, whereas
with HLE, the activity is preserved.37
In conclusion, similar structure-activity relationships
have been drawn from both enzymes. Although the
compounds are generally more potent on THR than on
FXa, the substitution of the aryl moiety in the 3-position
by a halide in the meta or in the 2,5-position improved
the inhibitory potency on both enzymes. In each case,
the replacement of the ester link between the coumarin
ring and the side chain usually decreases the activity
of the compounds. And finally, the removal or the
substitution of the chloromethyl in the 6-position always
lead to a drastic loss in inactivator potential.
Inactivation Kinetics and Selectivity. For the
most active compounds, we determined the ki/KI ratio,
which is the index of the inhibitory potency for mech-
anism-based inhibitors.45 The kinetic parameters ki and
KI were determined using either the progress curve or
the preincubation method.37 In the progress curve
method, the inhibitor competes with a chromogenic
substrate to bind the enzyme. In the preincubation
method, in contrast, the enzyme and the inhibitor are
incubated together before the determination of the
remaining activity at regular time intervals.
Derivatives 42-44 (Table 3) were synthesized accord-
ing to the strategy presented in Scheme 1 using the
corresponding commercially available salicylaldehyde
and â-ketoester.
Results and Discussion
In earlier studies, we found that the most potent
inhibitors of R-CT and HLE possessed an aryl ester
substituent in the 3-position.36,38 Therefore, we decided
to evaluate the potency of various aryl derivatives (6-
32) against human THR and FXa.
Enzyme Inhibition. To evaluate their inhibitory
potency toward THR and FXa, a screening test was
developed. In this test, the proteolytic activity is mea-
sured 10 min after incubation with the inhibitor at 2
and 5 µM for THR and FXa, respectively. The results,
displayed in Table 1, are expressed as the percentage
of enzyme residual activity.
The phenyl ester 6 is a moderately efficient THR and
FXa inhibitor. The replacement of the phenyl ring by a
benzyl 7, a pyridyl (8, 9), or a bulkier aryl group such
as a naphthyl (10, 11) or a quinolyl group (12, 13) leads
to a complete loss of inhibitory potency on both pro-
teases. The influence of substitution on the phenyl ring
was then investigated. As observed with HLE and R-CT,
the type and the position of the substituent strongly
influence THR and FXa inhibitory potency. Indeed,
whatever their nature, ortho (14, 17, 20) and para
substituents (16, 19, 22) seem to be always unfavorable.
The most active compounds, on both proteases, possess
a halogen group in the meta position (15, 18, 23, 24).
Molecules possessing a nitro 25, a methoxy 26, or a
trifluoromethyl moiety 27 in the meta position only
moderately inhibit THR, whereas a complete loss of FXa
inhibitory potency is observed. Among the dichlorophe-
nyl derivatives (29-32), it should be noted that only the
2,5-dichlorophenyl 30 is strongly efficient on THR. This
latter, which fairly inhibits FXa at 5 µM, was found to
be the most potent THR inhibitor in this series (1.0 (
0.1% of THR residual activity at 2 µM).
For THR, the most potent inhibitor was the 2,5-
dichlorophenyl ester 30 with a ki/KI ratio of 37 000 M-1
s-1 (Table 4). The THR inhibitory potency of the
m-bromo 24 (ki/KI ) 10 540 M-1 s-1) and m-chloro 15
(ki/KI ) 7720 M-1 s-1) derivatives were in the same
The influence of the link between the coumarin ring
and the phenyl side chain in the 3-position was also
range, whereas the m-fluoro 23 (ki/KI ) 2970 M-1 s-1
)
and m-iodo 18 (ki/KI ) 1210 M-1 s-1) derivatives were

Coumarins as Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7595
Scheme 2. Synthesis of the Ketone Derivatives 40 and 41a
a Reagents and conditions: (i) SOCl2, reflux, 1 h; (ii) (a) Meldrum’s acid, pyridine, CH2Cl2, 0 °C to room temp, 2 h; (b) EtOH, reflux, 3
h; (iii) corresponding â-ketoester, piperidine, AcOH, EtOH, reflux, 17 h; (iv) SOCl2, reflux, 2 h; (V) PPh3, DDQ, (n-hexyl)4NCl, CH2Cl2,
room temp, 5 min.
Table 3. In Vitro THR and FXa Inhibitory Potency of
6-Substituted Ester Derivatives
anism-based inhibitor for 30 vs competitive inhibitor for
argatroban), the comparison should be analyzed with
care.
Reactivation. The reversible or irreversible charac-
ter of the THR inhibition by compounds 15 and 30 was
analyzed. Previously, we showed that 6-chloromethyl-
coumarins irreversibly inhibited R-CT.44 Conversely,
reactivation of HLE arose, sped up by hydroxylamine.37
In the case of THR inhibited by 15 or 30 (Figure 1), we
studied the spontaneous reactivation during 2 days.
Only a poor recovery of activity, from 10% to 35% for
15 (Figure 1a) and from 1% to 11% for 30 (Figure 1b),
was observed within 7 h. To promote the reactivation,
an aliquot of hydrazine was added after 85 min. This
addition induced a rapid and partial reactivation from
15% to 55% for 15 and from 1% to 40% for 30 in 30 min
(Figure 1). Surprisingly, this reactivation did not exceed
65% and 40% for 15 and 30, respectively. These results
suggest the presence of several binding modes in the
enzyme-inhibitor complexes.
Mass Spectrometry. The 6-chloromethylcoumarins
were designed as mechanism-based inhibitors.44 Ac-
cording to their postulated inhibition mechanism (Scheme
3), it is proposed that the lactone carbonyl is first
attacked by the nucleophilic Ser195. The subsequent
lactone ring opening results in an increase of the leaving
properties of the chlorine atom and formation of an
electrophilic quinone methide (Scheme 3, pathway a).
This latter compound can then form a covalent bond
with a nucleophilic residue within the active site.
Alternatively, water-mediated deacylation of the acyl
enzyme can restore the enzyme activity (Scheme 3,
pathway b). Since mass spectrometry is a powerful
means to confirm the presence of covalently bound
groups in proteins, we measured the mass of the
thrombin-inhibitor complexes formed between THR
and the two derivatives, 15 and 30.
residual activity (%)
compd
X
THR
FXa
15
42
43
44
CH2Cl
10 ( 2
70 ( 4
107 ( 7
91 ( 15
29 ( 3
61 ( 3
77 ( 7
82 ( 7
CH3
H
CH2OCOCH3
less active. Compound 30 also displayed a good selectiv-
ity profile because its selectivity ratio of THR to FXa
and trypsin (TRY) was 168 and 54, respectively.
For compounds 15, 24, and 30, the weak FXa inhibi-
tion associated with the poor solubility of coumarins did
not allow us to use the progress curve method for the
determination of the kinetic parameters. Therefore, the
preincubation method was used to determine the ki/KI
ratio expressed as kobs/[I]. A good correlation between
these two methods was obtained because the ki/KI and
kobs/[I] ratios on FXa inhibition for 15 were 655 and 620
M-1 s-1, respectively. The herein obtained results
confirmed that these coumarins possessed only a weak
FXa inhibitory potency. The best FXa inhibitors 15 (kobs
/
[I] of 620 M-1 s-1) appeared to be also a good THR
inhibitor (ki/KI ) 7720 M-1 s-1).
For comparison purposes, we determined IC50 values
on THR for the best derivative 30 in this series and for
argatroban, a well-known THR inhibitor.15 IC50 values
were 103 ( 32 and 331 ( 22 nM for 30 and argatroban,
respectively. According to the Cheng-Prusoff equa-
tion,46 we estimated a Ki value for argatroban of 13 nM,
which is in agreement with the published value.15
According to this experiment, 30 seemed to be at least
as potent as argatroban. Nevertheless, since the type
of inhibition of these two derivatives is different (mech-
The enzyme-inhibitor complexes were produced by
portionwise addition of 15 or 30 to THR until an 85%
(15) or 95% (30) inactivation level was reached. An

7596 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24
Fre´de´rick et al.
Table 4. Kinetic Parameters ki and KI for Inactivation of THR, FXa, and Trypsin (TRY) by Ester Derivatives 15, 18, 23, 24, and 30a
selectivity
THRb ki/KI
(M-1 s-1
FXa kobs/[I]
trypsin kobs/[I]
compd
R
)
(M-1 s-1
220
)
(M-1 s-1
685
)
THR/FXa
168
THR/TRY
54
30
2,5-dichlorophenyl
37000
(0.016 min-1, 0.424 µM)
10540
24
15
23
18
m-bromophenyl
m-chlorophenyl
m-fluorophenyl
m-iodophenyl
430
620
nd
190
310
nd
25
12
55
25
(0.010 min-1, 0.983 µM)
7720
(0.014 min-1, 1.865 µM)
2970
(0.006 min-1, 2.002 µM)
1210c
nd
nd
a nd: not determined. Standard errors are less than 20%. b In parentheses: ki (min-1) and KI (µM). c ki and KI were not determined
because of weak inhibitory potency for 18.
Figure 1. THR enzymatic activity (52.2 nM) after incubation: (a) without (0) or with 15 (10 µM) spontaneous evolution (9),
spontaneous evolution followed by hydrazine (0.6 M) addition after 85 min incubation (b); (b) without (0) or with 30 (10 µM)
spontaneous evolution (9), spontaneous evolution followed by hydrazine (0.6 M) addition after 85 min incubation (b). Results are
expressed as the mean ( SE, n ) 3.
Scheme 3
aliquot of each complex was then treated with hydrazine
(0.6 M) during 1 h. The masses of the native THR and
of the different complexes were obtained by LC/MS
(Table 5).
For both inhibitors (15 and 30), the mass shift
observed between the native THR and the THR-
inhibitor complex corresponds to their exact mass (348.0
and 382.0 Da for 15 and 30, respectively) less one
chlorine atom (exact mass ) 313.0 Da and 348.0 for 15
and 30, respectively) (Scheme 4). These results exclude
the formation of a stable acyl enzyme in which Ser195
would have attacked the exocyclic ester (not shown on

Coumarins as Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7597
Table 5. Observed Mass of Native THR and Enzyme-Inhibitor Complexes
THR-15 treated
THR-30 treated
native THR
36036.2
THR-15
with hydrazine
THR-30
with hydrazine
mass (Da)
36349.4
36036.2
36285.9
0.0
36384.0
36036.2
36285.1
0.0
∆ mass between complex and
313.2
347.8
native THRa (Da)
249.7
248.9
a Mass shifts were calculated using the mixture between the complex and the native THR.
Scheme 4
Scheme 3). Treatment of these complexes with hydra-
zine leads to two entities possessing the same masses
whatever the chosen inhibitor, 15 or 30 (Table 5). The
lowest one (36 036.2 Da) corresponds to the native THR,
explaining the partial reactivation of the enzyme. The
second one (36 285.9 and 36 285.1 Da) presents a
positive mass shift of 249.7 and 248.9 Da for 15 and
30, respectively. This shift could be explained by the
formation of a dihydrazide complex that would result
from a nucleophilic attack of hydrazine on the lactone
ring and on the exocyclic ester (Scheme 4).
In comparison with R-CT, the inhibition mechanism
starts the same way. In each case, we observed the
leaving of the chlorine atom, probably resulting from
the opening of the lactone. However, with R-CT, no
reactivation was observed, even after treatment with
hydrazine.44 Moreover, the acyl enzyme complex formed
between R-CT and 42, devoid of any leaving group, is
irreversible. These results suggest the formation of a
strong interaction between R-CT and 42, and it can be
hypothesized that the acyl function, deeply buried in
the binding pocket, cannot be attacked by hydrazine.
On the opposite, with THR, we observed a partial
induced reactivation that suggests the existence of two
binding modes. The first one would be responsible for
the irreversible inactivation of THR probably through
the formation of a stable alkyl enzyme complex. In the
second one, an acyl enzyme could be slowly deacetylated.
the automated GOLD docking program.47 The enzyme-
inhibitor complexes were then refined by molecular
mechanics (DISCOVER348 module from INSIGHTII,49
CVFF force field), allowing the active site’s side chains
to accommodate the ligand.
The study revealed that compounds 15 and 30 inter-
act in a similar manner within the thrombin active site
(Figure 2). Their aryl side chain is deeply buried in the
specificity binding pocket (S pocket), while the chlorom-
ethyl group fits the proximal pocket (P pocket). The
distal pocket (D pocket) is left empty. In this bound
conformation, both carbonyl moieties of the inhibitors
adopt a syn conformation. This is in correlation with
the results from a previous structural study43 showing
that the syn orientation of the exo carbonyl function
versus the lactone in coumarinic derivatives is optimal
for the inhibition of R-CT.
Key interactions stabilize both compounds in the
binding cleft. A strong hydrogen bond involves the
exocyclic carbonyl moiety and the hydroxyl group of
Ser195. The coumarin ring favorably interacts through
π-π interactions (stacking and CH-π interactions) with
the aromatic residues delimiting the P pocket, namely,
Tyr60A and Trp60D. Furthermore, an interesting lipo-
philic contact is observed between Tyr228 located on the
back wall of the S binding pocket and the chlorine atom
in the meta position in 15 and in the 5-position in 30.
This type of interaction has already been described by
the group of Vacca.31 This group showed the potent
driving nature of such an interaction for thrombin
inhibitors with nonbasic P-1 substructures.
Molecular Modeling. In an attempt to support
reactivation and mass spectrometry data and to ratio-
nalize the trends observed in the enzymatic assays, the
molecular interactions between two of the more potent
derivatives, 15 and 30, and the THR active site were
investigated. First, binding modes were explored using
The global interaction energy of the complexes formed
between THR and 15 and 30 is -36.5 and -37.5 kcal
mol-1, respectively. On the basis of these energies, it is

7598 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24
Fre´de´rick et al.
Figure 2. Docking of compounds 15 (left) and 30 (right) into the active site of THR.
Figure 3. Molecular electrostatic potential (MEP) mapped onto the Connolly surface of (a) fluorophenyl, (b) bromophenyl, (c)
chlorophenyl, and (d) 2,5-dichlorophenyl moieties of 23, 24, 15, 30, respectively. The potentials are shown with identical chromatic
scales (red is more negative, greater electron density; blue is more positive, less electron density).
difficult to explain the difference in activity observed
in the biological results. Therefore, to estimate the
molecular parameters that could influence the affinity
of both compounds, the molecular electrostatic potential
(MEP) was calculated for the part differing among the
two inhibitors, i.e., the m-chlorophenyl of 15 and 2,5-
dichlorophenyl of 30, and compared to those of m-fluoro-
and m-bromophenyl of 23 and 24, respectively (Figure
3). Analysis of the MEP profiles correlates the trends
observed in the biological results (reported in Table 4).
Indeed, the compounds can be separated into three
groups. The fluorophenyl, which corresponds to the
lesser active derivative (23), exhibits the most electron-
rich aromatic ring. In contrast, in the 2,5-dichorophenyl,
which corresponds to the most active derivative (30),
the surface above the aromatic ring is much less
attractive (weaker electron density). The chlorophenyl
(15) and bromophenyl (24) present similar MEP profiles
and inhibitory potency, between the two others.
On the basis of these observations, the 2,5-dichlo-
rophenyl of 30 is thought to be much more stabilized in
the S pocket, in the vicinity of negatively charged
Asp189, than the m-chlorophenyl of 15 (the distance
between the carboxylate of Asp189 and the centroid of
the two rings is 4.6-4.7 Å). This could, in part, explain
the difference in activity of 15 vs 30.
Finally, in each complex, the lactone carbonyl is in
proximity to the hydroxyl moiety of the catalytic Ser195.
For each case, a distance of 3.2 Å between the carbon
of the lactone and the oxygen atom of Ser195 is
consistent with the nucleophilic attack of this residue
on the coumarin lactone, leading to the acyl enzyme.
This result corroborates the postulated inhibition mech-
anism presented above, as well as the existence of a
reversible complex shown in the reactivation study
(Figure 1, Scheme 3). Furthermore, in the surrounding
of the chloromethyl moiety of the ligand, two amino acid
residues (Tyr60A and His57) were identified as potential
nucleophilic residues. These should be able to attack the
quinone methide formed after the departure of the
chlorine atom (Scheme 3, pathway a), leading to the
irreversible complex (alkylated enzyme) revealed by the
reactivation study (Figure 1).
Predicted Cell Permeability. Since the final goal
in the development of new antithrombotics is bioavail-
ability, we estimated cell permeation using a computa-
tional approach for compounds 15, 18, 23, 24, 30 and
compared them with argatroban. For each compound,
a set of 94 molecular determinants relevant to the
process of membrane partitioning were computed with
the VOLSURF software50-52 and projected onto a pre-
dictive model of Caco2 permeation (Figure 4a). This
model, which was developed elsewhere,53 correlates the
94 molecular descriptors with the experimentally ob-
tained permeation of nearly 750 chemically diverse
compounds. Derivatives computed to have a value
greater than 0.4 are predicted to cross the cell mem-
brane, whereas those with a value lower than -0.4 will
have a highly reduced probability of penetrating the
Caco2 cell by passive diffusion.
In this model, 15, 18, 23, 24, and 30 were predicted
to behave as very good Caco2 penetrating agents with
values of 0.96, 1.03, 0.72, and 1.03, respectively. Arga-
troban, in contrast, possessed a value of -0.3, which is
in agreement with its known poor bioavailability.
For 30, the influence of individual molecular deter-
minants on the partitioning behavior was analyzed and
compared with argatroban. Figure 4b plots the variable
weight profile for the Caco2 model. The vertical bars
represent the relative contribution of each individual
molecular descriptor to the global Caco2 score. Positive
bars indicate descriptors whose values are directly
proportional to membrane permeability, while descrip-
tors with negative bars show an inverse correlation with
permeation. It appears from this plot that the enhance-
ment of Caco2 permeability is mainly due to an increase
of compound’s hydrophilic integy moment (Iw, which
represents the offset between the barycenter of the

Coumarins as Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7599
Figure 4. (a) Projection of Caco2 PLS scores of 15, 18, 23, 24, 30, and argatroban on Caco2 PLS scores of compounds with
known properties. (b) PLS weights profile for VOLSURF descriptors of the Caco2 model: V, volume; S, molecular surface; G,
molecular globularity; W, hydrophilic regions; Iw, hydrophilic integy moment; Cw, capacity factor; D, hydrophobic regions; BV,
best volumes; ID, hydrophobic integy moment; CP, critical packing parameter; HB, hydrogen-bonding; MW, molecular weight. (c)
Profile of molecular descriptors computed for 30. (d) Hydrophilic molecular interaction field (MIF) of 30 calculated with a water
probe, represented at -3 kcal/mol. Red vectors correspond to the hydrophilic integy moments Iw. (e) Profile of molecular descriptors
computed for argatroban.
hydrophilic regions and the center of mass of the
molecule), hydrophobic regions (D), and log P. On the
other hand, descriptors of hydrophilic regions (W),
hydrophilic capacity factor (Cw), and H-bonding (HB)
are inversely correlated with Caco2 permeability.
Parts c and e of Figure 4 shows the profile of these
molecular descriptors computed for 30 and argatroban,
respectively. The good predicted permeability of 30 can
be notably explained by the combined effects of (i) the
presence of a localized hydrophilic region around the
two carbonyl moieties (W, Figure 4c,d), which provides
an anisotropy of distribution of hydrophilic regions along
the molecular axis of 30 (Iw, Figure 4c,d), (ii) the
hydrophobic nature of the coumarin ring (D, Figure 4d),
and (iii) the absence of H-bond interaction. Conversely,
the profile of molecular descriptors for argatroban
(Figure 4e) shows that the highly hydrophilic nature (W,
Cw, Figure 4d) with H-bonding interaction and low log P
value highly contributes to the poor cell permeability
of this compound.
inhibitors of THR. Their kinetic parameters ki/KI were
determined, and it appeared that the most potent
compounds on THR were characterized by a 2,5-dichlo-
rophenyl or a 3-halogenophenylester on the 3-position
and a chloromethyl function on the 6-position. Moreover,
the best inhibitor in this series, 30, possessed good
selectivity over THR because FXa inactivation (kobs/[I]
) 220 M-1 s-1) and TRY inactivation (kobs/[I]) 685 M-1
s-1) were 168- and 54-fold lower than the THR one (ki/
KI ) 37 000 M-1 s-1). This study did not allow us to
obtain highly potent FXa inhibitors because the best
compound (15) inhibited FXa with a kobs/[I] of 620 M-1
s-1
.
The inhibition mechanism toward THR was investi-
gated. Albeit similar results were obtained using com-
pounds 15 and 30 as inhibitors, they significantly differ
from that described previously with R-CT. Although the
leaving of the chlorine atom was observed with both
enzymes, we showed hydrazine-induced partial reacti-
vations for THR. These results were confirmed by LC/
MS, which showed the resurgence of the native THR
and a new dihydrazide complex. To support these
assumptions, a docking study of 15 and 30 within THR
was performed. Key interactions (H-bonds, stacking,
and lipophilic contacts) stabilizing the compounds in the
active site of THR were highlighted. Moreover, com-
Conclusion
In this work, we evaluated the activity of coumarin-
type derivatives as direct inhibitors of THR and FXa.
A first screening at 2 and 5 µM for THR and FXa,
respectively, highlighted several compounds as strong

7600 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24
Fre´de´rick et al.
MHz, CDCl3): δ 8.70 (s, 1H), 8.00-7.35 (m, 9H), 7.30 (d, 1H),
parison of the molecular electrostatic potential (MEP)
generated around the substituted aryl side chain of
inhibitors helped to rationalize the trends observed in
the enzymatic assays. This study also showed that the
favored conformation and orientation of the inhibitors
within THR allow the formation of either an acyl
enzyme or an alkyl enzyme and thus corroborate the
partially reversible character of the THR inactivation
process in this series.
Finally, we have estimated the cell permeability of
our derivatives using a computational approach. Gener-
ally, our compounds should present very good mem-
brane permeation compared with argatroban, a known
THR inhibitor. Nevertheless, the method used here is
theoretical and parameters other than cell penetration
such as first-pass metabolism should be taken into
account when evaluating the bioavailability of cou-
marins that possess a lactone and a phenyl ester
vulnerable to esterase-catalyzed hydrolysis.
4.60 (s, 2H). Anal. (C21H13ClO4) C, H
3-Quinol-6-yl-6-chloromethyl-2-oxo-2H-1-benzopyran-
3-carboxylate (12). This was prepared according to the
general procedure presented above. Yield: 32%. Mp: 205-
206 °C. IR (KBr): 3056 (C-H arom), 1771 (CdO ester), 1729
(CdO lactone), 1620, 1575, 1501, 1376, 1294, 1246, 1225, 1210.
1H NMR (80 MHz, CDCl3) δ 9.00-8.95 (m, 2H), 8.79 (s, 1H),
8.21-8.18 (m, 2H), 7.80-7.40 (m, 5H), 4.70 (s, 2H). Anal.
(C20H12NClO4) H, N. C: calcd 65.67; found 64.64
3-Quinol-8-yl-6-chloromethyl-2-oxo-2H-1-benzopyran-
3-carboxylate (13). This was prepared according to the
general procedure presented above. Yield: 80%. Mp: 127 °C
(dec). IR (KBr): 3058 (C-H arom), 1770 (CdO ester), 1733 (Cd
1
O lactone), 1622, 1575, 1372, 1246, 1214. H NMR (80 MHz,
CDCl3): δ 9.00 (s, 1H), 8.95-8.90 (m, 1H), 8.25-8.17 (m, 1H),
7.80-7.40 (m, 7H), 4.70 (s, 2H). Anal. (C20H12NClO4) C, H, N.
4-(3-Chlorophenyl)-3-oxobutyric Acid Ethyl Ester (37).
To a flask containing 4.01 g (23.5 mmol) of 3-chlorophenylacetic
acid 35 was added thionyl chloride (20 mL). This solution was
stirred and refluxed under an argon atmosphere during 2 h.
After cooling, the mixture was evaporated and anhydrous
toluene (20 mL) was added to the residue and evaporated. This
operation is repeated three times. The acyl chloride was then
dissolved in dry CH2Cl2 (20 mL) and added via syringe to an
ice-cooled solution containing 3.72 g (25.8 mmol) of Meldrum’s
acid and 47.0 mmol of dry pyridine dissolved in dry CH2Cl2
(20 mL). The mixture was stirred under an argon atmosphere
and allowed to warm at room temperature during 2 h. The
resulting orange solution was then diluted with CH2Cl2 (50
mL) and carefully poured onto 2 M HCl and ice (150 mL). The
phases were separated, and the aqueous phase was extracted
twice with methylene chloride (2 × 50 mL). The combined
organic phases were washed with 1 M HCl (100 mL) and brine
(100 mL), dried over MgSO4, and evaporated. The resulting
orange crystals, which mainly consisted of the acetylated
Meldrum’s acid, were dissolved in EtOH (50 mL) and refluxed
for 3 h. The solvent was removed under reduced pressure and
furnished 3.01 g (13.4 mmol) of the desired product as a clear
Experimental Section
Chemistry. 1H NMR spectra were performed on Bruker AW
80 (80 MHz) or JEOL JNM EX 400 (400 MHz). The spectra
were measured in CDCl3 with TMS as an internal standard
(δ: 0.00 ppm). IR data reported in cm-1 were obtained using a
BIO-RAD FTS 165 spectrophotometer. Melting points were
determined with a Bu¨chi-Tottoli apparatus (B540) in open
capillary tubes and are uncorrected. Analyses (C, H, N) were
performed on a ThermoFinnigan flash EA 112 analyzer and
were within (0.4% of the theoretical values. Analytical thin-
layer chromatography (TLC) experiments were performed on
premade, glass-backed plates SiO2, 60PF254, 250 µm (Merck
5719). Compounds were visualized by UV illumination and by
heating to 150 °C after spraying phosphomolybdic acid in
ethanol. All the reactions were performed in two-necked round-
bottomed flasks equipped with a septum stopper and an argon-
filled balloon. All glassware was warmed prior to use and
degassed at 0.1 mmHg. All transfers of reagents were per-
formed via syringes.
1
oil. Yield: 57%. H NMR (CDCl3): δ 7.30-7.26 (m, 3H), 7.11
(m, 1H), 4.20 (q, J ) 7.2 Hz, 2H), 3.84 (s, 1H), 3.48 (s, 1H),
1.28 (t, J ) 7.2 Hz, 3H)
3-(3-Chlorobenzoyl)-6-hydroxymethyl-2-oxo-2H-1-ben-
zopyran-2-one (38). This compound was prepared according
to the general procedure presented above and used in the next
step in a one-pot procedure without any further purification.
3-(3-Chlorobenzoyl)-6-chloromethyl-2-oxo-2H-1-ben-
zopyran-2-one (40). To a flask containing 0.1 g of 38 was
added 1 mL of thionyl chloride. This solution was stirred and
refluxed under an argon atmosphere during 2 h. After cooling,
the mixture was evaporated and anhydrous toluene (20 mL)
was added to the residue and evaporated to dryness. This
operation is repeated three times. The residue was then
dissolved in hot acetone. The suspension was filtered off onto
a pad of Celite, and the filtrate was evaporated to half the
volume. The precipitate was filtered off, giving 63 mg of the
desired pure product. Yield: 10% in two steps. Mp: 198-200
General Method for the Synthesis of Esters 6-32,
Amide 33, and Thioester 34. To a flask containing 1 g (4.54
mmol) of 6-hydroxymethyl-2-oxo-2H-chromene-3-carboxylic
acid was added thionyl chloride (10 mL). This solution was
stirred and refluxed under an argon atmosphere during 3 h.
After cooling, the mixture was evaporated and anhydrous
toluene (10 mL) was added to the residue and evaporated. This
operation was repeated three times. Then, to the dried acyl
chloride dissolved in dry dioxane (20 mL) was added the
required alcohol, thiol, or amine (1.1 equiv) dissolved in
anhydrous dioxane (5 mL) and dry pyridine (1.1 equiv). After
2 h, the solvents were evaporated and the residue was
dissolved in AcOEt (150 mL) and washed three times with 0.1
N HCl (50 mL). The combined organic phases were dried over
MgSO4 and evaporated to dryness. The products were crystal-
lized in AcOEt/petroleine (40-60 °C).
This method allowed us to obtained derivatives 6-9, 14-
34, 42-44, which displayed structural features in agreement
with literature.36-38
3-Naphth-1-yl-6-chloromethyl-2-oxo-2H-1-benzopyran-
3-carboxylate (10). This was prepared according to the
general procedure presented above. Yield: 25%. Mp: 143-
146 °C. IR (KBr): 3054 (C-H arom), 1746 (CdO ester), 1713
(CdO lactone), 1622, 1574, 1239, 1222. 1H NMR (80 MHz,
CDCl3): δ 8.75 (s, 1H), 8.20-7.80 (m, 10H), 4.60 (s, 2H). Anal.
(C21H13ClO4) C, H.
1
°C. H NMR (CDCl3): δ 8.12 (s, 1H), 7.85 (s, 1H), 7.75-7.66
(m, 3H), 7.60 (d, 1H), 7.44 (m, 2H), 4.67 (s, 2H). Anal. (C17H10-
Cl2O3) C, H.
3-[2-(3-Chlorophenyl)acetyl]-6-hydroxymethyl-2-oxo-
2H-1-benzopyran-2-one (39). This was prepared according
to the general procedure presented above. Yield: 26%. Mp:
1
139-140 °C. H NMR (CDCl3): δ 8.52 (s, 1H), 7.68-7.66 (m,
2H), 7.38 (d, J ) 8.8 Hz, 1H), 7.29-7.23 (m, 3H), 7.18 (d, J )
6.8 Hz, 1H), 4.79 (d, J ) 4.4 Hz, 1H), 4.47 (s, 1H), 1.95 (s,
1H). Anal. (C18H13ClO4) H. C: calcd 65.76; found 65.11.
3-[2-(3-Chlorophenyl)acetyl]-6-chloromethyl-2H-1-ben-
zopyran-2-one (41). To a flask containing 559 mg of triph-
enylphosphine (2.13 mmol) and 483 mg of 2,3-dichloro-5,6-
dicyanobenzoquinone (2.13 mmol) dissolved in anhydrous
methylene chloride (10 mL) was added 831 mg of tetrahexy-
lammonium chloride (2.13 mmol) with stirring at room tem-
perature under an argon atmosphere. Then, an amount of 500
3-Naphth-2-yl-6-chloromethyl-2-oxo-2H-1-benzopyran-
3-carboxylate (11). This was prepared according to the
general procedure presented above. Yield: 30%. Mp: 187-
190 °C. IR (KBr): 3054 (C-H arom), 1777 (CdO ester), 1734
(CdO lactone), 1624, 1578, 1247, 1240, 1223. 1H NMR (80

Coumarins as Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7601
mg of 39 (1.52 mmol) dissolved in anhydrous methylene
chloride (15 mL) was added, and the mixture was vigorously
stirred. The yellow color immediately changed to deep red. The
solution was diluted with CH2Cl2 (10 mL) and washed twice
with water and once with brine. The combined organic phases
were dried over MgSO4 and evaporated to dryness. Finally,
AcOEt (10 mL) was added to the residue and the yellowish
precipitate filtered off and washed twice with AcOEt (10 mL)
to give 100 mg of pure product. Yield: 20%. Mp: 162-166 °C.
1H NMR (CDCl3): δ 8.51 (s, 1H), 7.71-7.67 (m, 2H), 7.40 (d,
J ) 8.4 Hz, 1H), 7.29-7.24 (m, 3H), 7.18 (d, J ) 6.8 Hz, 1H),
4.65 (s, 1H), 4.47 (s, 1H). Anal. (C18H12Cl2O3) C, H.
Enzymatic Studies. Human THR and bovine TRY were
purchased from Roche, and human FXa was purchased from
Kordia. The proteases were assayed spectrophotometrically
using p-nitroanilide substrates: D-phenylalanylpipecolylargi-
nyl-p-nitroanilide (S-2238 from Chromogenix) and tosylgly-
cylprolylarginyl-p-nitroanilide (Chromozym TH from Roche)
for THR, N-methoxycarbonyl-D-norleucylglycyl-L-arginyl-p-
nitroanilide for FXa (Chromozym X from Roche), and N-R-
benzoyl-DL-arginyl-p-nitroanilide (B4875 from Sigma) for TRY.
The reactions were performed in 0.01 M Tris-HCl, 0.01 M
HEPES, 0.1 M NaCl, 0.1% (w/v) PEG 6000 and at pH 7.4 for
THR, 0.05 M Tris-HCl, 0.15 M NaCl, 0.005 M CaCl2, 0.001 M
EDTA, 0.05% (v/v) Tween-20 and at pH 7.4 for FXa, and 0.1
M Tris-HCl, 0.01 M CaCl2 and at pH 8.0 for TRY. Assays
contained 10% (v/v) DMSO and were run at 25 °C in a Lambda
20 (Perkin-Elmer) spectrophotometer equipped with a ther-
mostated cell holder. Argatroban was purchased from Sequoia
Research Products.
experimental data to the equations were performed with
GraphPad prism, version 3.02 (GraphPad Software). At low
inhibitor concentrations, the ratio ki/KI was obtained as kobs
/
[I] (preincubation method). For the progress curve method, the
experimental conditions were as follows: [THR]0 ) 1.16 nM,
[I]0 ) 0.025-10 µM, [Chromozym TH] ) 80 µM, and [FXa]0 )
1.16 nM, [I]0 ) 0.1-10 µM, [Chromozym X] ) 200 µM. For
the preincubation method, the enzymes and inhibitors con-
centrations were as follows: for FXa, [FXa]0 ) 5.8 nM, [I]0 )
1-5 µM; for TRY, [TRY]0 ) 145 nM, [I]0 ) 2.5-10 µM.
Spontaneous and Hydrazine Reactivation. To evaluate
the irreversible character of the inhibition, enzyme and
inhibitor are incubated until a 75% inactivation was reached.
The enzyme-inhibitor complexes were then incubated in the
presence or absence of hydrazine for 48 h. Enzyme activity of
aliquots was monitored and compared to that of the control.
The THR activity of each group was expressed as a percent of
control at t0. For reactivation studies, the experimental condi-
tions were as follows: [THR]0 ) 52.2 nM, [I] ) 10 µM, and
[hydrazine] ) 0.6 M.
Mass Spectrometry. Preparation of Enzyme-Inhibi-
tor Complexes. The enzyme-inhibitor complexes were pre-
pared by portionwise addition of inhibitor until an 85%
inactivation was reached. These complexes were then incu-
bated with hydrazine for 1 h. For each condition, enzyme
aliquots were taken and analyzed by LC/MS. Note that in this
case the THR kinetic buffer does not contain PEG-6000. The
experimental conditions for the preparation of the complexes
were as follows: [THR]0 ) 1.28 µM, [15] ) 40 µM, [30] ) 10
µM, and [hydrazine] ) 0.6 M.
Screening Test. Inactivation measurement at one inhibitor
concentration was performed by incubating the inhibitor and
the enzyme at 25 °C for 10 min. Then, the residual activity of
the enzyme was measured with the appropriate chromogenic
substrate (40 µM of S-2238 for THR and 300 µM of Chromozym
X for FXa) and compared to a control (enzyme solution without
inhibitor). The residual activity percentage was obtained by
dividing the measured residual activity with the control
activity. The concentrations of enzymes and inhibitors for the
screening test were the following: for THR, [THR] ) 2.32 nM,
[I] ) 2 µM; for FXa, [FXa] ) 5.8 nM, [I] ) 5 µM.
IC50 Determination. The 20 µL of test compound in DMSO
(or DMSO alone in the control) and 20 µL of human thrombin
(Roche, 23.2 nM) in 140 µL of THR kinetic buffer were
incubated for 20 min in a 96-well assay plate. An amount of
20 µL of S-2238 (Chromogenix, 2.5 mM) was then added to
the mixture. After 5 min of substrate hydrolysis, the reaction
was stopped by the addition of 20 µL of 10% acetic acid and
the absorbance at 405 nm was measured in a microplate reader
(Bio-Rad). The background absorbance was determined with
20 µL of test compound in DMSO (or DMSO alone in the
control) mixed with 180 µL of THR kinetic buffer and 20 µL of
10% acetic acid. All compounds were assayed in triplicate at
six concentrations. Percent inhibition at each concentration
was calculated from the OD at 405 nm from the experimental
and control samples. IC50 values were calculated from a
sigmoidal dose response nonlinear regression equation using
GraphPad prism, version 3.02 (GraphPad Software). The
reported IC50 values are expressed as the mean ( SEM.
Kinetic Parameters Determination. Suicide substrate
inactivation can be represented by the following minimum
scheme:
LC/MS. The molecular mass of the enzyme-inhibitor
complexes was obtained by injecting 1 µg of these complexes
into an Agilent SB-C18 column, pore size 300 Å, particle size
3.5 µm, 50 mm × 2 mm i.d. at 25 °C. Solvent A consisted of
water with 1% formic acid and solvent B of acetonitrile. The
gradient profile for solvent B was 5-80% in 20 min. The flow
rate was 500 µL min-1. After elution from the column, the
sample was ionized by an ESI source of the mass spectrometer.
ESI/MS experiments were performed on a Agilent 1100 LC/
MSD trap (Agilent Technologies) operating in positive ion
mode. Instrumental parameters were as follows: nebulizer
pressure 40 psi, auxiliary gas flow 6 L min-1, temperature of
auxiliary gas 325 °C, and capillary voltage 4.5 kV.
Molecular Modeling. Molecular modeling studies were
carried out using INSIGHTII software,49 version 2000, running
on a Silicon graphics workstation. The structures of the
inhibitors were obtained by means of the BUILDER module
and optimized using the CFF91 force field. Docking simula-
tions were performed inside the human THR enzymes (PDB
code: 1A4W) with the automated GOLD program.47 For each
inhibitor, the binding mode with the best fit as well as
stereoelectronic matching quality was selected. To take into
account protein flexibility, all the complexes were then refined
with DISCOVER3.48 The minimization process, using the
CVFF force field (dielectric constant 1r), includes two steps:
the steepest descent algorithm, reaching a convergence of 10.00
kcal mol-1 Å-1, followed by the conjugated gradient algorithm
to reach a final convergence of 0.01 kcal mol-1 Å-1
.
Quantum chemical calculations were performed using the
Gaussian 98 program.54 The geometries of the four fragments
(fluoro-, bromo-, chloro-, and 1,4-dichlorophenyl) were fully
optimized, and the molecular electrostatic potential (MEP) was
calculated at the HF/6-31G* level of theory. These were plotted
onto the Connolly surface (probe 1.4 Å) using the INSIGHTII
software.
Cell Permeability. Prediction of Caco2 cell permeation of
15, 18, 23, 24, 30, and argatroban was performed with the
VOLSURF algorithm,50-52 as implemented in SYBYL, version
7.0. The three-dimensional structures of 15, 18, 23, 24, 30 were
obtained using the INSIGHTII software by means of the
BUILDER module and optimized using the CFF91 force field.
The three-dimensional structure of argatroban was extracted
from the active site of thrombin (PDB code: 1DWC). The three-
dimensional molecular interaction field (MIF) was first cal-
E and I are the free forms of enzyme and inhibitor, respec-
tively, E*I is a kinetic chimere of the Michaelis complex and
the acyl enzyme, E-I is the inactivated enzyme, and P is the
product of hydrolysis of I. The kinetic constants ki and KI were
determined using the preincubation method or the progress
curve method.37 Linear and nonlinear regression fits of the

7602 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24
Fre´de´rick et al.
(19) Hilpert, K.; Ackermann, J.; Banner, D. W.; Gast, A.; Gubernator,
K.; Hadvary, P.; Labler, L.; Muller, K.; Schmid, G.; Tschopp,
T. B.; et al. Design and synthesis of potent and highly
selective thrombin inhibitors. J. Med. Chem. 1994, 37, 3889-
3901.
(20) Kettner, C.; Shaw, E. D-Phe-Pro-ArgCH2C1-A selective affinity
label for thrombin. Thromb. Res. 1979, 14, 969-973.
(21) Bajusz, S.; Szell, E.; Bagdy, D.; Barabas, E.; Horvath, G.;
Dioszegi, M.; Fittler, Z.; Szabo, G.; Juhasz, A.; Tomori, E.; et al.
Highly active and selective anticoagulants: D-Phe-Pro-Arg-H,
a free tripeptide aldehyde prone to spontaneous inactivation, and
its stable N-methyl derivative, D-MePhe-Pro-Arg-H. J. Med.
Chem. 1990, 33, 1729-1735.
(22) Neises, B.; Broersma, R. J.; Tarnus, C.; Piriou, F.; Remy, J. M.;
Lintz, C.; Heminger, E. F.; Kutcher, L. W. Synthesis and
comparison of tripeptidylfluoroalkane thrombin inhibitors. Bioorg.
Med. Chem. 1995, 3, 1049-1061.
(23) Iwanowicz, E. J.; Lau, W. F.; Lin, J.; Roberts, D. G.; Seiler, S.
M. Retro-binding tripeptide thrombin active-site inhibitors:
discovery, synthesis, and molecular modeling. J. Med. Chem.
1994, 37, 2122-2124.
(24) Boatman, P. D.; Ogbu, C. O.; Eguchi, M.; Kim, H. O.; Nakanishi,
H.; Cao, B.; Shea, J. P.; Kahn, M. Secondary structure peptide
mimetics: design, synthesis, and evaluation of beta-strand
mimetic thrombin inhibitors. J. Med. Chem. 1999, 42, 1367-
1375.
(25) Li, M.; Lin, Z.; Johnson, M. E. Structure-based design and
synthesis of novel thrombin inhibitors based on phosphinic
peptide mimetics. Bioorg. Med. Chem. Lett. 1999, 9, 1957-
1962.
culated around the target molecule, using the energy (Lennard-
Jones, hydrogen-bonding, and Coulombic electrostatics) of
interaction with a molecular probe placed at each node of a
regular 3D lattice. A grid resolution of 0.5 Å was employed,
and three distinct molecular probes (H2O, DRY, and carbonyl)
were used to evaluate both polar and nonpolar molecular
regions. In a subsequent step, VOLSURF transforms the MIF
into a set of scalar descriptors and projects their values onto
a predictive model of Caco2 cell permeation. This predictive
model was developed by correlating the same set of descriptors
with the experimental permeation of nearly 750 known drugs,
using a cross-validated discriminant partial least-squares
procedure.
Acknowledgment. This work was supported in part
by the “Fonds Spe´cial de Recherche” (F.S.R., FUNDP,
Namur, Belgium), the “Fonds de la Recherche Scienti-
fique Me´dicale Belge” (F.R.S.M., Grant No. 3.4570.99),
the F.R.I.A. (Bruxelles, Belgium) from which J. de
Ruyck is a Research Fellow, and the “Fonds National
de la Recherche Scientifique” (F.N.R.S., Bruxelles,
Belgium) from which C. Charlier is a Research Fellow.
Supporting Information Available: Microanalytical data
for all synthesized compounds not included in the Experimen-
tal Section. This material is available free of charge via the
Internet at http://pubs.acs.org.
(26) Kettner, C.; Mersinger, L.; Knabb, R. The selective inhibition
of thrombin by peptides of boroarginine. J. Biol. Chem. 1990,
265, 18289-18297.
(27) Tamura, S. Y.; Levy, O. E.; Uong, T. H.; Reiner, J. E.; Goldman,
E. A.; Ho, J. Z.; Cohen, C. R.; Bergum, P. W.; Nutt, R. F.; Brunck,
T. K.; Semple, J. E. Guanylpiperidine peptidomimetics: potent
and selective bis-cation inhibitors of factor Xa. Bioorg. Med.
Chem. Lett. 2000, 10, 745-749.
(28) Marlowe, C. K.; Sinha, U.; Gunn, A. C.; Scarborough, R. M.
Design, synthesis and structure-activity relationship of a series
of arginine aldehyde factor Xa inhibitors. Part 1: structures
based on the (D)-Arg-Gly-Arg tripeptide sequence. Bioorg. Med.
Chem. Lett. 2000, 10, 13-16.
(29) Hara, T.; Yokoyama, A.; Ishihara, H.; Yokoyama, Y.; Nagahara,
T.; Iwamoto, M. DX-9065a, a new synthetic, potent anticoagulant
and selective inhibitor for factor Xa. Thromb. Haemostasis 1994,
71, 314-319.
(30) Adler, M.; Davey, D. D.; Phillips, G. B.; Kim, S. H.; Jancarik, J.;
Rumennik, G.; Light, D. R.; Whitlow, M. Preparation, charac-
terization, and the crystal structure of the inhibitor ZK-807834
(CI-1031) complexed with factor Xa. Biochemistry 2000, 39,
12534-12542.
(31) Tucker, T. J.; Brady, S. F.; Lumma, W. C.; Lewis, S. D.; Gardell,
S. J.; Naylor-Olsen, A. M.; Yan, Y.; Sisko, J. T.; Stauffer, K. J.;
Lucas, B. J.; Lynch, J. J.; Cook, J. J.; Stranieri, M. T.; Holahan,
M. A.; Lyle, E. A.; Baskin, E. P.; Chen, I. W.; Dancheck, K. B.;
Krueger, J. A.; Cooper, C. M.; Vacca, J. P. Design and synthesis
of a series of potent and orally bioavailable noncovalent thrombin
inhibitors that utilize nonbasic groups in the P1 position. J. Med.
Chem. 1998, 41, 3210-3219.
(32) Lumma, W. C., Jr.; Witherup, K. M.; Tucker, T. J.; Brady, S. F.;
Sisko, J. T.; Naylor-Olsen, A. M.; Lewis, S. D.; Lucas, B. J.;
Vacca, J. P. Design of novel, potent, noncovalent inhibitors of
thrombin with nonbasic P-1 substructures: rapid structure-
activity studies by solid-phase synthesis. J. Med. Chem. 1998,
41, 1011-1013.
(33) Scozzafava, A.; Briganti, F.; Supuran, C. T. Protease inhibitors.
Part 3. Synthesis of non-basic thrombin inhibitors incorporating
pyridinium-sulfanilylguanidine moieties at the P1 site. Eur. J.
Med. Chem. 1999, 34, 939-952.
(34) Clare, B. W.; Scozzafava, A.; Supuran, C. T. Protease inhibitors,
part 13: Specific, weakly basic thrombin inhibitors incorporating
sulfonyl dicyandiamide moieties in their structure. J. Enzyme
Inhib. 2001, 16, 1-13.
References
(1) Hirsh, J.; Weitz, J. I. New antithrombotic agents. Lancet 1999,
353, 1431-1436.
(2) Hirsh, J.; Weitz, J. I. Thrombosis and anticoagulation. Semin.
Hematol. 1999, 36, 118-132.
(3) Thiagarajan, P.; Wu, K. K. Mechanisms of antithrombotic drugs.
Adv. Pharmacol. 1999, 46, 297-324.
(4) Dahlback, B. Blood coagulation. Lancet 2000, 355, 1627-1632.
(5) Davie, E. W.; Fujikawa, K.; Kisiel, W. The coagulation cascade:
initiation, maintenance, and regulation. Biochemistry 1991, 30,
10363-10370.
(6) Davie, E. W. Biochemical and molecular aspects of the coagula-
tion cascade. Thromb. Haemostasis 1995, 74, 1-6.
(7) Mann, K. G. Biochemistry and physiology of blood coagulation.
Thromb Haemostasis 1999, 82, 165-174.
(8) Porcari, A. R.; Chi, L.; Leadley, R. Recent advances in clinical
trials of the direct and indirect selective Factor Xa inhibitors.
Expert Opin. Invest. Drugs 2000, 9, 1595-1600.
(9) Tan, K. T.; Makin, A.; Lip, G. Y. Factor X inhibitors. Expert Opin.
Invest. Drugs 2003, 12, 799-804.
(10) Gould, W. R.; Leadley, R. J. Recent advances in the discovery
and development of direct coagulation factor xa inhibitors. Curr.
Pharm. Des. 2003, 9, 2337-2347.
(11) Hirsh, J.; Warkentin, T. E.; Shaughnessy, S. G.; Anand, S. S.;
Halperin, J. L.; Raschke, R.; Granger, C.; Ohman, E. M.; Dalen,
J. E. Heparin and low-molecular-weight heparin: mechanisms
of action, pharmacokinetics, dosing, monitoring, efficacy, and
safety. Chest 2001, 119, 64S-94S.
(12) Hirsh, J.; Dalen, J.; Anderson, D. R.; Poller, L.; Bussey, H.;
Ansell, J.; Deykin, D. Oral anticoagulants: mechanism of action,
clinical effectiveness, and optimal therapeutic range. Chest 2001,
119, 8S-21S.
(13) Vacca, J. P. New advances in the discovery of thrombin and
factor Xa inhibitors. Curr. Opin. Chem. Biol. 2000, 4, 394-400.
(14) Sanderson, P. E. Anticoagulants: Inhibitors of thrombin and
factor Xa. Annu. Rep. Med. Chem. 2001, 36, 79-88.
(15) Kikumoto, R.; Tamao, Y.; Tezuka, T.; Tonomura, S.; Hara, H.;
Ninomiya, K.; Hijikata, A.; Okamoto, S. Selective inhibition of
thrombin by (2R,4R)-4-methyl-1-[N2-[(3-methyl-1,2,3,4-tetrahy-
dro-8-quinolinyl) sulfonyl]-l-arginyl)]-2-piperidinecarboxylic acid.
Biochemistry 1984, 23, 85-90.
(16) Sturzebecher, J.; Markwardt, F.; Voigt, B.; Wagner, G.; Wals-
mann, P. Cyclic amides of N-alpha-arylsulfonylaminoacylated
4-amidinophenylalanine. Tight binding inhibitors of thrombin.
Thromb. Res. 1983, 29, 635-642.
(17) Eriksson, U. G.; Renberg, L.; Bredberg, U.; Teger-Nilsson, A.
C.; Regardh, C. G. Animal pharmacokinetics of inogatran, a low-
molecular-weight thrombin inhibitor with potential use as an
antithrombotic drug. Biopharm. Drug Dispos. 1998, 19, 55-64.
(18) Eriksson, B. I.; Carlsson, S.; Halvarsson, M.; Risberg, B.;
Mattsson, C. Antithrombotic effect of two low molecular weight
thrombin inhibitors and a low-molecular weight heparin in a
caval vein thrombosis model in the rat. Thromb. Haemostasis
1997, 78, 1404-1407.
(35) Sanderson, P. E.; Stanton, M. G.; Dorsey, B. D.; Lyle, T. A.;
McDonough, C.; Sanders, W. M.; Savage, K. L.; Naylor-Olsen,
A. M.; Krueger, J. A.; Lewis, S. D.; Lucas, B. J.; Lynch, J. J.;
Yan, Y. Azaindoles: moderately basic P1 groups for enhancing
the selectivity of thrombin inhibitors. Bioorg. Med. Chem. Lett.
2003, 13, 795-798.
(36) Pochet, L.; Doucet, C.; Schynts, M.; Thierry, N.; Boggetto, N.;
Pirotte, B.; Jiang, K. Y.; Masereel, B.; de Tullio, P.; Delarge, J.;
Reboud-Ravaux, M. Esters and amides of 6-(chloromethyl)-2-
oxo-2H-1-benzopyran-3-carboxylic acid as inhibitors of R-chy-
motrypsin: significance of the “aromatic” nature of the novel
ester-type coumarin for strong inhibitory activity. J. Med. Chem.
1996, 39, 2579-2585.

Coumarins as Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 24 7603
(37) Doucet, C.; Pochet, L.; Thierry, N.; Pirotte, B.; Delarge, J.;
(46) Cheng, Y.; Prusoff, W. H. Relationship between the inhibition
constant (K1) and the concentration of inhibitor which causes
50% inhibition (I50) of an enzymatic reaction. Biochem. Phar-
macol. 1973, 22, 3099-3108.
Reboud-Ravaux, M. 6-Substituted 2-oxo-2H-1-benzopyran-3-
carboxylic acid as
a core structure for specific inhibitors
of human leukocyte elastase. J. Med. Chem. 1999, 42, 4161-
4171.
(47) Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R.
Development and validation of a genetic algorithm for flexible
docking. J. Mol. Biol. 1997, 267, 727-748.
(38) Pochet, L.; Doucet, C.; Dive, G.; Wouters, J.; Masereel, B.;
Reboud-Ravaux, M.; Pirotte, B. Coumarinic derivatives
as mechanism-based inhibitors of alpha-chymotrypsin and
human leukocyte elastase. Bioorg. Med. Chem. 2000, 8, 1489-
1501.
(39) Pochet, L.; Frederick, R.; Masereel, B. Coumarin and isocou-
marin as serine protease inhibitors. Curr. Pharm. Des. 2004,
10, 3781-3796.
(40) Kam, C. M.; Fujikawa, K.; Powers, J. C. Mechanism-based
isocoumarin inhibitors for trypsin and blood coagulation serine
proteases: new anticoagulants. Biochemistry 1988, 27, 2547-
2557.
(48) Discover3, version 2.98; Accelrys Inc.: San Diego, CA, 1998.
(49) InsightII, version 2000; Accelrys Inc.: San Diego, CA, 2000.
(50) Cruciani, G.; Pastor, M.; Clementi, S. Handling Information from
3D Grid Maps for QSAR Studies. Molecular Modeling and
Prediction of Bioactivity; Kluwer Academic/Plenum Publishers:
New York, 2000; pp 73-82.
(51) Cruciani, G.; Crivori, P.; Carrupt, P.-A.; Testa, B. Molecular
fields in quantitative structure-permeation relationships: the
VolSurf approach. J. Mol. Struct.: THEOCHEM 2000, 503, 17-
30.
(41) Kam, C. M.; Kerrigan, J. E.; Plaskon, R. R.; Duffy, E. J.; Lollar,
P.; Suddath, F. L.; Powers, J. C. Mechanism-based isocoumarin
inhibitors for blood coagulation serine proteases. Effect of the
7-substituent in 7-amino-4-chloro-3-(isothioureidoalkoxy)isocou-
marins on inhibitory and anticoagulant potency. J. Med. Chem.
1994, 37, 1298-1306.
(42) Iranpoor, N.; Firouzabadi, H.; Aghapour, G.; Vaez Zadeh, A. R.
Triphenylphosphine/2,3-dichloro-5,6-dicyanobenzoquinone as a
new selective and neutral system for the facile conversion of
alcohols, thiols and selenols to alkyl halide in the presence of
halide ions. Tetrahedron 2002, 58, 8689-8693.
(43) Wouters, J.; Huygens, M.; Pochet, L.; Pirotte, B.; Durand, F.;
Masereel, B. Structural approach of the mechanism of inhibition
of a-chymotrypsin by coumarins. Bioorg. Chem. Lett. 2002, 12,
1109-1112.
(44) Pochet, L.; Dieu, M.; Fre´de´rick, R.; Murray, A. M.; Kempen, I.;
Pirotte, B.; Masereel, B. Investigation of the inhibition mecha-
nism of coumarins on chymotrypsin by mass spectrometry.
Tetrahedron 2003, 59, 4557-4561.
(45) Silverman, R. B. Mechanism-based enzyme inactivators. Meth-
ods Enzymol. 1995, 249, 240-283.
(52) Crivori, P.; Cruciani, G.; Carrupt, P. A.; Testa, B. Predicting
blood-brain barrier permeation from three-dimensional molec-
ular structure. J. Med. Chem. 2000, 43, 2204-2216.
(53) URL: http://www.moldiscovery.com.
(54) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery,
J. A., Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam,
J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.;
Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.;
Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G.
A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck,
A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;
Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.;
Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.;
Challacombe, M.; Gill, P. M. W.; Johnson, B. G.; Chen, W.; Wong,
M. W.; Andres, J. L.; Head-Gordon, M.; Replogle, E. S.; Pople,
J. A. Gaussian 98, revision A.11; Gaussian, Inc.: Pittsburgh,
PA, 1998.
JM050448G