Mechanism-based thrombin inhibitors: Design, synthesis, and molecular docking of a new selective 2-oxo-2H-1-benzopyran derivative

Article abstract of DOI:10.1021/jm061368v

New 2-oxo-2H-1-benzopyran derivatives were prepared to optimize 2a,b, initially developed as mechanism-based α-chymotrypsin (α-CT) inhibitors, into potent and selective thrombin (THR) inhibitors. From this study, 22, characterized by a 2-(N-ethyl-2′-oxoacetamide)-5′- chlorophenyl ester side chain, was shown to be a good THR inhibitor (k _i/K_I = 3455 M^-1·s^-1), displaying an excellent selectivity profile against other serine proteases such as factor Xa, trypsin, and α-CT. Docking analysis of this compound into the different protein structures revealed the molecular basis responsible for its potency and selectivity.

Full text of DOI:10.1021/jm061368v

J. Med. Chem. 2007, 50, 3645-3650
3645
Mechanism-Based Thrombin Inhibitors: Design, Synthesis, and Molecular Docking of a New
Selective 2-Oxo-2H-1-benzopyran Derivative
Raphae¨l Fre´de´rick,*,†,‡ Se´verine Robert,† Caroline Charlier,§ Johan Wouters,§ Bernard Masereel,† and Lionel Pochet†
Department of Pharmacy and CBS Laboratory, Drug Design and DiscoVery Center (D3C), UniVersity of Namur, FUNDP, 61,
Rue de Bruxelles, B-5000 Namur, Belgium
ReceiVed NoVember 27, 2006
New 2-oxo-2H-1-benzopyran derivatives were prepared to optimize 2a,b, initially developed as mechanism-
based R-chymotrypsin (R-CT) inhibitors, into potent and selective thrombin (THR) inhibitors. From this
study, 22, characterized by a 2-(N-ethyl-2′-oxoacetamide)-5′-chlorophenyl ester side chain, was shown to
be a good THR inhibitor (ki/KI ) 3455 M-1‚s-1), displaying an excellent selectivity profile against other
serine proteases such as factor Xa, trypsin, and R-CT. Docking analysis of this compound into the different
protein structures revealed the molecular basis responsible for its potency and selectivity.
Introduction
(R-CT) inhibitors,9 were also potent nonbasic THR inhibitors.10
In particular, 2a, strongly active on R-CT (ki/KI ) 762 700
Thrombin (THR) is a key enzyme involved in the last step
of the coagulation cascade. By converting the glycoprotein
fibrinogen into fibrin monomers, THR is responsible for clot
stabilization and therefore plays a major role in hemostasis and
thrombosis. In addition, this enzyme is one of the most potent
stimulators of platelet secretion and aggregation. THR is thus
clearly recognized as an attractive target for the development
of inhibitors in thrombotic disorders.1,2
Most THR inhibitors designed up to now bind the enzyme
through formation of a salt bridge between Asp189 at the bottom
of the specificity pocket (S1 or S pocket) and a positively
charged group in the P1 position of the ligand.1 Recently,
different groups have shown that THR inhibitors deprived of
such a basic moiety demonstrate improved bioavailability and
pharmacokinetics.3-8 In this context, Merck has reported 1a,
characterized by a 2′,5′-dichlorophenyl moiety, as one of the
first potent nonbasic THR inhibitors.8 The binding mode of this
compound, revealed by X-ray diffraction, implicates a strong
hydrophobic interaction between the chlorine in the 5′-position
and the aromatic ring of Tyr228 deeply inserted within the
S-pocket. Replacement of the 2′-chlorine by a 2-(N-ethyl-2′-
oxoacetamide) moiety (1b) further improved the potency and
the solubility in comparison to 1a.
M-1‚s-1), displays an interesting inhibitory potency against
THR, with inhibitory parameters ki/KI of 7720 M-1‚s-1 9,10
.
Introduction of a chlorine in the 2′-position (2b) enhanced THR
inhibition (ki/KI ) 37 000 M-1‚s-1) about 5-fold without really
modifying the potency against R-CT (ki/KI ) 220 000 M-1‚s-1).
Taking advantage of the structure-activity relationships
(SAR) described by Merck for 1a,b, we aimed to optimize
coumarins 2a,b into more potent and selective THR inhibitors.
In this paper, we thus report the development of new coumarins
characterized by a 2′-substituted 5′-chlorophenyl side chain
linked to the coumarin ring in the 3-position by an ester, methyl
ester, or an amide. To rationalize the SAR and selectivity
observed, a molecular docking study was efficiently realized
using the automated GOLD docking software.
Chemistry
Targeted compounds 21-29 were prepared by reaction
between the acid chloride and the appropriate amine or alcohols
following the previously reported method.9,10 However, this
protocol required the prior synthesis of intermediates 4, 7, and
11 (Schemes 1 and 2). The 2-hydroxymethyl-4-chlorophenol 4
was obtained by reduction of the corresponding acid 3 with
LiAlH4 in THF (Scheme 1). The phenol 7 was prepared from
4 in three steps (Scheme 1): (a) 4 reacted with TBDMSCl in
the presence of DMAP to give the silyl ether 5, this latter being
(b) alkylated with R-bromoacetic acid ethyl ester in dioxane
and (c) deprotected with tetra-N-butylammonium fluoride in
THF to afford the desired intermediate 7.
The hydroxyl compound 11 was obtained from 2-methoxy-
4-chlorophenol 8 (Scheme 2). This latter compound was first
alkylated using R-bromoacetic acid ethyl ester and cesium
carbonate in dioxane and then demethylated with boron tribro-
mide. This reaction, accompanied by hydrolysis of the ester,
led to acid 10, which reacted with ethylamine in the presence
of HOBT, NMM, and EDC in anhydrous DMF to give the
targeted intermediate 11.
In this context, our group has recently reported that cou-
marins, initially developed as mechanism-based R-chymotrypsin
Some 6-phenoxycoumarins have also been prepared. The
intermediate 6-phenoxy-2-oxo-2H-1-benzopyrane-3-carboxylic
acid 15 was synthesized according to Scheme 3. First, the
4-phenoxyphenol 12 reacted according to a Reimer-Tiemann
reaction with chloroform and sodium hydroxide in an ethanolic/
aqueous solution to afford the 2-hydroxy-5-phenoxybenzalde-
* To whom correspondence should be addressed. Phone: +32 (0)81 72
42 99. Fax: +32 (0)81 72 42 99. E-mail: raphael.frederick@gmail.com.
† Department of Pharmacy.
‡ Present address: Auckland Cancer Society Research Centre (ACSRC),
University of Auckland, PO92319, Auckland, New Zealand.
§ CBS Laboratory.
10.1021/jm061368v CCC: $37.00 © 2007 American Chemical Society
Published on Web 06/20/2007

3646 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15
Fre´de´rick et al.
Table 1. Synthesis, Physicochemical Parameters, and in Vitro THR
Scheme 1a
Residual Activity of Ester and Amide of 6-Chloromethyl- and
6-Phenoxycoumarins 21-29
yield
(%)
mp
(°C)
[I]
THR residual
compd
R
X
R′
(µM) activity (%)
2aa
2ba
21b
CH2Cl
CH2Cl
CH2Cl
O
O
O
H
Cl
NH2
2
2
10 ( 2
1 ( 0.1
60 ( 9
6 ( 3
55 254-258 100
100
10
2
22
CH2Cl
O
OCH2CONHEt 15 124-127
3 ( 3
11 ( 2
103 ( 4
101 ( 1
95 ( 6
91 ( 4
110 ( 9
94 ( 6
90 ( 3
23
24
25
26
27
28
29
CH2Cl OCH2 Cl
CH2Cl OCH2 OH
CH2Cl OCH2 OCH2COOEt
CH2Cl NH
CH2Cl NH
85 212-215 100
26 183-184 100
38 134-136 100
39 281-282 100
Cl
OH
Cl
H
a Reagents and conditions: (i) LiAlH4, THF, room temp, 16 h; (ii)
TBDMSCl, NEt3, DMAP, CH2Cl2, 0 °C f room temp, 2.5 h; (iii)
BrCH2COOEt, Cs2CO3, dioxane, room temp, 4 h; (iv) NBu4F, THF, room
temp, 1 h.
72 307c
100
OPh
OPh
O
O
17 152-153 100
31 119-122 100
a From ref 10. b Obtained after deprotection of the amine using the
mixture trifluoroacetic acid/CH2Cl2. c Decomposition.
Scheme 2a
µM inhibitor. For the drugs displaying high potency, lower
concentrations were investigated. The results, displayed in Table
1, are expressed as the percentage of enzyme residual activity.
Replacement of the 2′-chlorine atom of 2b by an amine (21)
greatly reduces THR inactivation potential (60% of THR
residual activity at 100 µM vs 1% at 2 µM for 2b). In contrast,
introduction of a 2-(N-ethyl-2′-oxoacetamide) in this position
leads to a good THR inhibitor (22) (11% of THR residual
activity at 2 µM) as potent as 2a (Table 1).
Compounds characterized by a benzyl ester (23-25) or a
phenylamide (26, 27) in the 3-position of the coumarin ring
are all inactive. This emphasizes the importance of the ester
linker to suitably orient the compounds in the enzyme active
site (Table 1).
a Reagents and conditions: (i) BrCH2COOEt, Cs2CO3, dioxane, room
temp, 4 h; (ii) BBr3, CH2Cl2, room temp, 1 h; (iii) EtNH2.HCl, HOBt, NMM,
EDC, DMF, room temp, 18 h.
From a previous modeling study,10 2a,b were shown to poorly
occupy the D-pocket within the THR active site. On the basis
of this observation, we designed 28 and 29, analogous to 2a
and 2b, respectively, and both are characterized by a phenoxy
in the 6-position in place of a chloromethyl. The aromatic moiety
might form an additional π-π interaction with Trp215 inside
the D-pocket, further stabilizing the compounds within the cleft.
This modification, however, did not afford potent THR inhibi-
tors. Indeed, up to now, all examples designed to replace the
chloromethyl group led to inactive compounds. This result
confirmed the importance of the CH2Cl moiety in this series
for the inhibition mechanism, as discussed later.
For the most active compound (22), a detailed enzymatic
study was performed. The ki/KI ratio, which reflects the index
of inhibitory potency for mechanism-based inhibitors,11 was
determined on THR but also on R-CT, factor Xa (FXa), and
trypsin (TRY) to assess the selectivity of this compound against
other serine proteases. The kinetic parameters ki and KI were
determined using the progress curve or the preincubation
method, depending on the protease.12 In the progress curve
method, the inhibitor competes with a chromogenic substrate
to bind the enzyme. In the preincubation method, however, the
enzyme and the inhibitor are incubated together before the
determination of the remaining activity at regular time intervals.
To assess the efficiency of our inhibitors, the partition ratio,
defined as the ratio of product released to inactivation (kcat/ki),
was also determined. The results are summarized in Table 2.
This study confirmed that 22 is a good THR inhibitor (ki/KI
) 3455 M-1‚s-1, partition ratio of 13.5) even if slightly less
Scheme 3a
a Reagents and conditions: (i) CHCl3, NaOH, H2O, EtOH, reflux, 12 h;
(ii) diethyl malonate, piperidine, HOAc, reflux, 14 h; (iii) NaOH, EtOH,
H2O, reflux, 1 h.
hyde 13. This latter was then condensed with diethyl malonate
under classical Knoevenagel conditions to give 14 and finally
saponified to generate the desired acid 15.
Finally, 4, 7, 11, the 2,5-dichlorobenzyl alcohol 16, 2,5-
dichlorophenol 17, 2,5-dichloroaniline 18, and N-tert-butoxy-
carbonyl-2-hydroxy-4-chloroaniline 19 reacted with the acyl
chloride of 15 or the 6-chloromethyl-2-oxo-2H-1-benzopyrane-
3-carboxylic acid 20 obtained as previously described9 to afford
the targeted coumarins 21-29 (Table 1).
Results and Discussion
The inhibitory potency of 21-29 was first evaluated on
isolated and purified THR. In this screening, the proteolytic
activity of THR is measured 10 min after incubation with 100

Mechanism-Based Thrombin Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15 3647
Table 2. Comparison of Kinetic Parameters ki and KI for Inactivation of THR, FXa, R-CT, and TRY by Derivatives 22 and 2a,b
THR
compd
R
ki/KI (M-1‚s-1
7716 ( 952a
)
kcatki
FXa kobs/[I] (M-1‚s-1
)
TRY kobs/[I] (M-1‚s-1
)
R-CT ki/KI (M-1‚s-1
)
selectivity THR/R-CT
2a
2b
22
H
Cl
2.8
1.8
13.5
620 ( 36a
219 ( 7a
59 ( 2
313 ( 55a
685 ( 66a
118 ( 2
762700b
288000b
755 ( 45
0.01
0.13
5
37007 ( 5150a
3457 ( 835
N-ethyloxoacetamide
a From ref 10. b From refs 9 and 13. Standard errors are less than 15%.
Table 3. Computational Data
potent and efficient than 2a,b (ki/KI ) 7720 M-1‚s-1 and
partition ratio of 2.8 for 2a; ki/KI ) 37 000 M-1‚s-1 and partition
ratio of 1.8 for 2b) (Table 2). Of great interest, 22 was shown
to exhibit a significantly improved selectivity against R-CT with
a ki/KI of 755 M-1‚s-1 (vs 762 700 M-1‚s-1 for 2a and 220 000
M-1‚s-1for 2b). So 22 is about 5 times more potent on THR
than on R-CT (selectivity ratio THR/R-CT of 5) compared to
selectivity ratios of 0.01 and 0.13 for 2a and 2b, respectively.
occur-
rence
total
energy ET
OγSer195‚‚‚COlactone 5′Cl‚‚‚Tyr228
enzyme
THR
I
mode
(%)
(kcal‚mol-1
)
(Å)
(Å)
2b normal
inverse
22 normal
inverse
2b normal
inverse
22 normal
inverse
2b normal
inverse
22 normal
inverse
2b normal
inverse
22 normal
inverse
63
37
5
91
84
14
2
29
3
26
2
23
67
25
16
51
-1815
-1808
-1838
-1819
-1891
-1870
-1878
-1866
-2402
-2405
-2411
-2400
-512
2.56
5.32
2.56
5.23
2.63
4.42
3.37
4.46
5.23
5.12
4.69
5.11
3.38
4.92
3.38
2.99
3.60
3.67
4.20
6.47
4.15
4.52
4.20
4.43
R-CT
FXa
Furthermore, 22 is almost completely inactive on both FXa (kobs
/
[I] ) 60 M-1‚s-1, selectivity ratio THR/FXa of 58) and TRY
(kobs/[I] ) 120 M-1‚s-1, selectivity ratio THR/TRY of 29).
The inhibition profile of 22 was further established on THR
by enzymatic and mass spectrometry experiments, according
to the well-established criteria for mechanism-based enzyme
inactivators.11 First, a typical saturation kinetic was observed
for 22. Second, the irreversibility of the THR inhibition by 22
was confirmed by ultracentrifugation experiments. Indeed, the
inhibition was persistent even after removing the excess of
inhibitor. Third, an enzyme-inhibitor complex with a 1:1
stoichiometry was obtained by mass spectrometry. A mass shift
of 414 Da was observed between the THR-22 complex
(36 450.2 Da) and the native protein (36 036.2 Da) which
corresponds to the exact mass of 22 (449 Da) less one chlorine
atom. These findings confirm the mechanism-based behavior
of 22 as already observed for these coumarin derivatives.10
Details on these studies are given in the Supporting Information.
To our knowledge, 22 constitutes a unique example of selective
mechanism-based and nonbasic THR inhibitor.
Because all the serine proteases share the same catalytic
mechanism, it is generally particularly difficult to achieve
selectivity in mechanism-based inhibitors. To understand these
results, i.e., the lower THR inhibitory potency of 22 compared
to 2b and its greatly improved selectivity profile, the binding
modes of both compounds were investigated in the four serine
proteases THR, R-CT, FXa, and TRY by molecular modeling.
The mechanism of inhibition proposed for these coumarins14
implies the nucleophilic attack of the active Ser195 Oγ atom
on the lactone carbonyl moiety. The resulting lactone ring
opening leads to increased leaving properties of the 6-chlorine
atom and the subsequent formation of an electrophilic quinone
methide. This latter can then form a covalent bond with a
nucleophilic residue, probably His57, within the active site. For
this reaction to occur, the coumarin ring needs to be properly
positioned in the neighborhood of Ser195; i.e., the lactone
coumarin should be close to the hydroxyl group of Ser195
(distance of <3 Å).
Compounds 22 and 2b were docked within the THR, FXa,
R-CT, and TRY active sites using the automated GOLD program
(PDB codes 1A4W, 1LPG, 4GCH, and 1EB2, respectively).15
In each case, 100 conformations were produced and clustered
by similarity. Two main modes were identified among the
solutions: (i) the normal mode, which refers to the binding of
TRY
-510
-512
-511
the 2′-substituted 5′-chlorophenyl side chain within the S-pocket
and the lactone coumarin close to the Ser195 residue and is the
binding mode in agreement with the mechanism of inhibition
described above and (ii) the inverse mode, which corresponds
to the binding of the chloromethylcoumarin within the S-pocket
and the lateral side chain outside the active site. Taking into
account the protein flexibility, the conformations with the best
score (GoldScore) in the normal and inverse modes were further
refined using the MINIMIZE module as implemented in Sybyl,
version 7.2.3 (Tripos force field and Gasteiger-Huckel charges).16
To assess the feasibility of the inhibition mechanism, the
distance between the oxygen Oγ of Ser195 and the lactone
carbonyl of the coumarin ring (OγSer195-COlactone) was evalu-
ated. Furthermore, to appraise the stabilization of the compound
within the S-pocket, we also measured the distance between
the 5′-chlorine on the coumarin side chain and the centroid of
Tyr228 aromatic ring (5′Cl-Tyr228). Such a hydrophobic
contact implying a chlorine atom and Tyr228 has been shown
to strengthen the interaction of 1b with THR (observed 5′Cl-
Tyr228 ) 4.69 Å).8 The obtained results are reported in
Table 3.
When 2b is docked within the THR active site, 63% of the
conformations generated by GOLD adopt the normal mode. The
superposition of the first 50 solutions (Figure 1, left) emphasizes
the highly conserved orientation. Key interactions stabilize the
compound inside the active site (Figure 2a). The 2′,5′-dichlo-
rophenyl is deeply inserted within the S-pocket and, similar to
1b, interacts with Tyr228 (distance 5′Cl-Tyr228 ) 3.60 Å).
The coumarin ring forms a π‚‚‚π stacking contact with Tyr60a
in the P pocket (not shown in Figure 2a for clarity), while the
lactone carbonyl is involved in an H-bond network with the
so-called oxyanion hole (formed by the backbone NH of residues
Ser195 and Gly193). In this conformation, taking into account
the intrinsic flexibility of the protein,17 a distance OγSer195
COlactone of 2.56 Å makes nucleophilic attack of Ser195 onto
the coumarin ring possible and is thus consistent with the
-

3648 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15
Fre´de´rick et al.
Figure 1. Superposition of the first 50 solutions generated by GOLD
in the THR active site for 2b (left) and 22 (right).
Figure 3. Docking of 2b (left) and 22 (right) in the active site of
R-CT (a, b) and FXa (c, d). Pictures were made using Pymol.18
(Figure 2c). The low occurrence (5%) of the normal binding
mode might be partly explained by the high flexibility of the
oxoacetamide side chain of 22. Actually, many papers have
reported the strong influence of ligand flexibility on the accurate
prediction of binding poses, as well as the limitations of using
a rigid protein during docking simulations.19,20 However, this
flexibility might also trigger the binding of 22 in the inverse
orientation (Figure 2d).
In the inverse mode, two similarly occurring conformations
are observed (Figure 2d), one possessing the oxoacetamide side
chain on the left side (43%, in magenta) and the other on the
right side (48%, in yellow) regarding the orientation of the
coumarin ring. Both conformations are equally stabilized (ET
) -1819 and -1816 kcal‚mol-1, respectively), notably through
an H-bond between the hydroxyl group of Ser195 and the O
atom of the exocyclic ester. In each case, the particular
orientation of the coumarin ring excludes the nucleophilic attack
of Ser195 onto the lactone (distance OγSer195-COlactone > 5 Å).
Therefore, the binding of 22 in the inverse mode prevents the
inactivation of THR through the postulated inhibition mecha-
nism. Nevertheless, in both inverse orientations, Ser195 lies in
Figure 2. Normal binding mode of (a) 2b and (b) 22 in the active site
of THR. (c) Superimposition of 22 (green) and 1b (magenta) obtained
from the X-ray crystal structure (PDB code 1TA6). (d) Superimposition
of both conformations obtained in the inverse mode for 22. Pictures
were made using Pymol.18
proposed inhibition mechanism. In contrast, in the inverse mode,
the distance OγSer195-COlactone, increased to 5.32 Å, prevents
the reaction from occurring. This conformation is consequently
in disagreement with the proposed mode of action of this
compound. Furthermore, the difference of 7 kcal‚mol-1 in the
total energy of both complexes confirms the higher stabilization
of 2b in the normal mode compared to the inverse one.
When 22 is docked within the THR active site, only 5% of
the generated conformations adopt the normal mode whereas
91% lie in the inverse one. The superposition of the first 50
conformations (Figure 1, right) does not show such a clear
convergence as observed for 2b. However, when the energy of
both modes is compared, the less occurring normal orientation
appears to be much more favorable than the inverse one (∆ET
) 19 kcal‚mol-1).
In the normal mode, 22 is similarly to 2b stabilized through
an H-bond involving the lactone carbonyl and the oxyanion hole
as well as a lipophilic Caromatic-Cl‚‚‚π interaction (5′Cl-Tyr228
) 3.67 Å) (Figure 2b). The distance OγSer195-COlactone of 2.56
Å is in agreement with the inhibition mechanism. Additionally,
22 is stabilized through an H-bond between the oxoacetamide
carbonyl and the backbone NH of Gly219. A similar interaction
observed in the case of 1b crystallized with THR corroborates
this binding mode.8 Both 2-(N-ethyl-2′-oxoacetamide)-5′chlo-
rophenyl side chains adopt an analogous orientation in the
S-pocket, as observed in the superimposition of 1b and 22
the vicinity of the exocyclic ester (distance OγSer195-COexocy
-
clic ester ) 3.5 and 3.8 Å). The nucleophilic attack of the catalytic
serine onto the exocyclic carbonyl becomes achievable and
might lead to the hydrolysis of the side chain, releasing the
phenol and the coumarin carboxylic acid.
On the basis of these results, it can be assumed that when 22
adopts the normal orientation within the THR active site, it leads
to the complete inactivation of the protein. Alternatively, the
binding of 22 in the inverse mode results in a catalysis leading
to its hydrolysis and the formation of side products. Therefore,
the lower inhibitory activity of 22 versus 2b could be explained
by a higher decomposition rate (resulting from a binding in the
inverse mode) rather than a weaker stabilization when binding
in the normal orientation. The high partition ratio (kcat/ki ) 13.5)
experimentally observed for 22 clearly supports this hypothesis.
In R-CT, 2b adopts again primarily the normal mode (84%
of solutions), in agreement with the proposed mode of action
(distance OγSer195-COlactone ) 2.63 Å). In addition to interac-
tions similar to those observed in THR, an H-bond occurs
between the lactone carbonyl and the backbone NH of Asp194
(Figure 3a). Furthermore, the 2′-chlorine atom is surrounded
by the Met192 side chain (not shown in Figure 3a for clarity;
see Figure S1 in Supporting Information). This lipophilic

Mechanism-Based Thrombin Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15 3649
Experimental Section
environment produces an extra hydrophobic stabilization of the
2′,5′-dichlorophenyl group. Together, these additional interac-
tions might account for the high R-CT inhibitory potency
observed for 2b (ki/KI ) 220 000 M-1‚s-1).
Detailed synthesis of intermediates 4-7, 9-11, 13-15 and the
targeted compounds 21-29 can be found in the Supporting
Information.
In contrast, in the case of 22 docked in R-CT, only 2% of
the solutions generated adopt the normal mode (Figure 3b).
Because of Met192 bordering the S-pocket, the 5′-chlorophenyl,
with its bulky 2-(N-ethyl-2′-oxoacetamide) substituent, is clearly
less deeply inserted within the cleft. Consequently, the coumarin
ring, and in particular the lactone carbonyl, moves away from
Ser195 (distance OγSer195-COlactone ) 3.37 Å). Moreover, in
this conformation, the lipophilic Caromatic-Cl‚‚‚π interaction
(distance 5′Cl-Tyr228 ) 6.47 Å) and the H-bond network with
the oxyanion hole cannot be formed. These observations plainly
explain the severe loss in the R-CT inhibitory potency for 22
vs 2b (∼290-fold) and therefore its greatly improved selectivity
profile (THR/R-CT ) 5 for 22 vs 0.17 for 2b).
For comparison purposes, we also analyzed the binding modes
of 2b and 22 within the FXa and TRY active sites. In the first
protease, only 3% and 2% of the docked conformations obtained
for 2b and 22, respectively, show the normal mode, with the
2′-substituted-5′-chlorophenyl inserted in the S-pocket (parts c
and d of Figure 3, respectively). However, because of the
absence of a P pocket (structural feature particular to FXa), the
coumarin ring is rotated by 180°, positioning the lactone next
(2-Hydroxy-4-chlorophenoxy)-N-ethylacetamide (11). To a
solution of ethyl 2-methoxy-4-chlorophenoxyacetate 9 (1 g, 4.1
mmol) in CH2Cl2 (10 mL) is added a solution of BBr3 prepared
from 1.46 g of BBr3 (5.5 mmol) diluted in dry CH2Cl2 (5.55 mL).
The mixture is vigorously stirred for 1 h at room temperature. After
this time, the solvents are removed and the residue is dissolved in
CH2Cl2 (30 mL), washed with water and brine, dried over MgSO4,
and evaporated to dryness. The crude product is purified by column
chromatography (eluant, 100% AcOEt) to afford 420 mg (1.99
mmol) of pure 2-hydroxy-4-chlorophenoxyacetic acid 10 (yield,
47%). Then, to a solution of 10 (300 mg, 1.48 mmol), hydroxy-
benzotriazole (HOBT) (220 mg, 1.63 mmol), ethylamine chlorhy-
drate (133 mg, 1.63 mmol), and N-methylmorpholine (NMM) (330
mg, 3.26 mmol) in dry DMF (10 mL) is added the ethyl-3-(3-
(dimethylamino)propyl)carbodiimide chlorhydrate (EDC) (340 mg,
1.77 mmol). The mixture is stirred for 18 h under an inert
atmosphere and then diluted by AcOEt (50 mL), washed with water
(3 × 50 mL) and brine, dried over MgSO4, and evaporated. The
crude product is purified by chromatography column (eluant, AcOEt
100%) to give 290 mg (1.26 mmol) of the desired 11 (yield, 86%).
General Method for the Synthesis of Compounds 21-29.10
To a flask containing 1 g (4.54 mmol) of 6-hydroxymethyl-2-oxo-
2H-chromene-3-carboxylic acid is added thionyl chloride (10 mL).
This solution is refluxed for 3 h under an argon atmosphere. After
cooling, the mixture is evaporated and dry toluene (10 mL) is added
to the residue and evaporated. This operation is repeated three times.
Then, to the dried acyl chloride dissolved in dry dioxane (20 mL)
are added the required alcohol or amine (1.1 equiv) and dry pyridine
(1.1 equiv). After 2 h, the solvents are evaporated and the residue
is dissolved in AcOEt (150 mL) and washed three times with 0.1
N HCl (50 mL). The combined organic phases are dried over
MgSO4 and evaporated to dryness. The final compounds are purified
by crystallization (acetonitrile, 21; AcOEt, 24-29), preparative
HPLC (22), or collected filtration (23).
to Gly216, far from the catalytic Ser195 (distance OγSer195
-
COlactone ) 5.23 and 4.69 Å for 2b and 22, respectively).
Therefore, whatever the substituent in the P1 position, the
coumarin binds to FXa in a conformation that prevents Ser195
from reacting with the lactone moiety. This observation helps
to clarify the poor inhibitory potency against this enzyme
observed in all the coumarin series.10
In TRY, 65% of the docked solutions for 2b exhibit the
normal orientation in comparison to only 16% for 22 (Table
3). Nevertheless, in both cases, the lactone moiety is more distant
from Ser195 than within the THR active site (distance OγSer195
-
Acknowledgment. C.C. is greatly indebted to the Belgian
National Foundation for Scientific Research (FNRS) for the
award of a research fellowship.
COlactone ) 3.38 Å) (see Supporting Information for figures).
This might explain the lower inhibitory activity on TRY than
on THR observed for these compounds.
Supporting Information Available: Detailed description of the
synthesis of 21-29 and intermediates 4-7, 9-11, 13-15, molec-
ular modeling methods, biological assays, mass spectra, and
analytical data of all compounds. This material is available free of
charge via the Internet at http://pubs.acs.org.
Conclusion
To design selective mechanism-based inhibitors specifically
targeting THR, we considered the optimization of our lead 2b.
On the basis of this parent compound, a series of new coumarins
were synthesized and their biological activity assayed against
THR. Among these, 22, characterized by a 2-(N-ethyl-2′-
oxoacetamide)-5′chlorophenyl ester moiety in the 3-position of
a 6-chloromethyl-2-oxo-2H-1-benzopyran core, constitutes the
most promising derivative. 22 is slightly less potent on THR
than 2b (THR ki/KI ) 3455 M-1‚s-1 vs THR ki/KI ) 37 000
M-1‚s-1), but of great interest, it exhibits a drastically improved
selectivity profile, particularly regarding R-CT (selectivity ratios
of 58, 29, and 5 for FXa, TRY, and R-CT, respectively). As
shown by molecular docking, the bulky 2-(N-ethyl-2′-oxoac-
etamide) chain prevents suitable orientation of the coumarin
nucleus in the R-CT active site for the required nucleophilic
attack by the active Ser195.
References
(1) Hirsh, J.; O’Donnell, M.; Weitz, J. I. New anticoagulants. Blood 2005,
105 (2), 453-463.
(2) Weitz, J. I. New anticoagulants for treatment of venous thromboem-
bolism. Circulation 2004, 110 (9), I19-I26.
(3) De Simone, G.; Menchise, V.; Omaggio, S.; Pedone, C.; Scozzafava,
A.; Supuran, C. T. Design of weakly basic thrombin inhibitors
incorporating novel P1 binding functions: molecular and X-ray
crystallographic studies. Biochemistry 2003, 42 (30), 9013-9021.
(4) 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 (7), 1011-1013.
(5) 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 (5), 795-
798.
To our knowledge, such selectivity has never been reported
previously in mechanism-based THR inhibitors deprived of a
basic moiety. 22 thus represents a remarkable pharmacological
tool to further explore selective mechanism-based THR inhibi-
tors.
(6) Scozzafava, A.; Briganti, F.; Supuran, C. T. Protease inhibitors. Part
3. Synthesis of non-basic thrombin inhibitors incorporating pyri-
dinium-sulfanilylguanidine moieties at the P1 site. Eur. J. Med. Chem.
1999, 34 (11), 939-952.

3650 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15
Fre´de´rick et al.
(7) Supuran, C. T.; Scozzafava, A.; Briganti, F.; Clare, B. W. Protease
inhibitors: synthesis and QSAR study of novel classes of nonbasic
thrombin inhibitors incorporating sulfonylguanidine and O-methyl-
sulfonylisourea moieties at P1. J. Med. Chem. 2000, 43 (9), 1793-
1806.
(8) 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
(17), 3210-3219.
(9) 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 alpha-chymotrypsin:
significance of the “aromatic” nature of the novel ester-type coumarin
for strong inhibitory activity. J. Med. Chem. 1996, 39 (13), 2579-
2585.
(10) Frederick, R.; Robert, S.; Charlier, C.; de Ruyck, J.; Wouters, J.;
Pirotte, B.; Masereel, B.; Pochet, L. 3,6-Disubstituted coumarins as
mechanism-based inhibitors of thrombin and factor Xa. J. Med. Chem.
2005, 48 (24), 7592-7603.
(11) Silverman, R. B. Mechanism-based enzyme inactivators. Methods
Enzymol. 1995, 249, 240-283.
(12) Doucet, C.; Pochet, L.; Thierry, N.; Pirotte, B.; Delarge, J.; 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 (20), 4161-4171.
(13) Pochet, L.; Doucet, C.; Dive, G.; Wouters, J.; Masereel, B.; Reboud-
Ravaux, M.; Pirotte, B. Coumarinic derivatives as mechanism-based
inhibitors of alpha-chynotrypsin and human leukocyce elastase.
Bioorg. Med. Chem. 2000, 8 (6), 1489-1501.
(14) Pochet, L.; Dieu, M.; Fre´de´rick, R.; Murray, A. M.; Kempen, I.;
Pirotte, B.; Masereel, B. Investigation of the inhibition mechanism
of coumarins on chymotrypsin by mass spectrometry. Tetrahedron
2003, 59, 4557-4561.
(15) 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 (3), 727-748.
(16) Sybyl, version 7.2.3; Tripos Inc. (1699 South Hanley Rd, St. Louis,
MO, 63144).
(17) Teague, S. J. Implications of protein flexibility for drug discovery.
Nat. ReV. Drug DiscoVery 2003, 2 (7), 527-541.
(18) DeLano, W. L. The PyMOL Molecular Graphics System; DeLano
Scientific: San Carlos, CA, 2002.
(19) Perola, E.; Walters, W. P.; Charifson, P. S. A detailed comparison
of current docking and scoring methods on systems of pharmaceutical
relevance. Proteins 2004, 56 (2), 235-249.
(20) Erickson, J. A.; Jalaie, M.; Robertson, D. H.; Lewis, R. A.; Vieth,
M. Lessons in molecular recognition: the effect of ligand and protein
flexibility on molecular docking accuracy. J. Med. Chem. 2004, 47,
45-55.
JM061368V