Probing the subpockets of factor Xa reveals two binding modes for inhibitors based on a 2-carboxyindole scaffold: A study combining structure-activity relationship and X-ray crystallography

Article abstract of DOI:10.1021/jm0490540

Structure-activity relationships within a series of highly potent 2-carboxyindole-based factor Xa inhibitors incorporating a neutral P1 ligand are described with particular emphasis on the structural requirements for addressing subpockets of the factor Xa enzyme. Interactions with the subpockets were probed by systematic substitution of the 2-carboxyindole scaffold, in combination with privileged P1 and P4 substituents. Combining the most favorable substituents at the indole nucleus led to the discovery of a remarkably potent factor Xa inhibitor displaying a K_i value of 0.07 nM. X-ray crystallography of inhibitors bound to factor Xa revealed substituent-dependent switching of the inhibitor binding mode and provided a rationale for the SAR obtained. These results underscore the key role played by the P1 ligand not only in determining the binding affinity of the inhibitor by direct interaction but also in modifying the binding mode of the whole scaffold, resulting in a nonlinear SAR.

Full text of DOI:10.1021/jm0490540

J. Med. Chem. 2005, 48, 4511-4525
4511
Articles
Probing the Subpockets of Factor Xa Reveals Two Binding Modes for Inhibitors
Based on a 2-Carboxyindole Scaffold: A Study Combining Structure-Activity
Relationship and X-ray Crystallography
Marc Nazare´,* David W. Will, Hans Matter, Herman Schreuder, Kurt Ritter, Matthias Urmann,
Melanie Essrich, Armin Bauer, Michael Wagner, Jo¨rg Czech, Martin Lorenz, Volker Laux, and Volkmar Wehner
Aventis Pharma Deutschland GmbH, a company of the Sanofi-Aventis group, Industriepark Ho¨chst, Building G878,
D-65926 Frankfurt am Main, Germany
Received November 22, 2004
Structure-activity relationships within a series of highly potent 2-carboxyindole-based factor
Xa inhibitors incorporating a neutral P1 ligand are described with particular emphasis on the
structural requirements for addressing subpockets of the factor Xa enzyme. Interactions with
the subpockets were probed by systematic substitution of the 2-carboxyindole scaffold, in
combination with privileged P1 and P4 substituents. Combining the most favorable substituents
at the indole nucleus led to the discovery of a remarkably potent factor Xa inhibitor displaying
a Ki value of 0.07 nM. X-ray crystallography of inhibitors bound to factor Xa revealed
substituent-dependent switching of the inhibitor binding mode and provided a rationale for
the SAR obtained. These results underscore the key role played by the P1 ligand not only in
determining the binding affinity of the inhibitor by direct interaction but also in modifying
the binding mode of the whole scaffold, resulting in a nonlinear SAR.
1. Introduction
for the development of orally bioavailable compounds
without resorting to prodrug6 approaches. These com-
pounds either contain nonbasic amidine isosteres, for
example, razaxaban,7 or exploit the so-called chloro
binding mode in the S1 pocket.8
Thromboembolic diseases, including deep vein throm-
bosis, myocardial infarction, and pulmonary embolism,
are major causes of mortality in Europe and North
America.1 Unfortunately the standard anticoagulants
currently used for the treatment and prevention of these
diseases have several disadvantages. These include slow
onset of action, parenteral mode of administration, and
severe, potentially life-threatening side effects such as
excessive bleeding requiring stringent monitoring of
drug levels.2 The serine protease factor Xa (fXa) has
emerged as an attractive target for the therapy of
thrombosis-related diseases because it is a key enzyme
in the activation cascade of the blood coagulation
system, linking the extrinsic and intrinsic activation
pathways.3 It is anticipated that inhibition of fXa should
prevent thrombus formation without compromising
normal hemostasis and platelet function by maintaining
a basal thrombin level.4 An orally available fXa inhibitor
should therefore be a superior antithrombotic agent and
overcome the shortcomings associated with the current
treatment. A major obstacle to the development of an
orally available drug from the many potent and selective
inhibitors available is often the requirement for a highly
basic benzamidine or guanidine serving as an arginine
mimetic in the S1 pocket of fXa. In general, these
moieties are linked to the extremely poor bioavailability
observed for the inhibitors bearing this functionality.5
Fortunately, recently a range of potent nonbasic fXa
inhibitors has emerged, opening up exciting possibilities
In previous communications,9 we described 2-carboxy-
indole fXa inhibitors that carried the P1 ligand at the
indole N1 and the P4 ligand as 2-carboxamides and that
were otherwise undecorated. We described the optimi-
zation of various neutral P1 ligands using the chloro
binding mode and piperidine-derived P4 ligands at-
tached to the carboxyindole scaffold leading to a set of
interchangeable P1 and P4 ligands and thus highly
active inhibitors. The key primary interactions neces-
sary for potent inhibition of fXa are those in the S1 and
S4 pockets. The choice of ligands that provide optimal
interactions within these pockets combined with a
suitable scaffold ensuring the correct trajectories for the
ligands has been a major focus in fXa inhibitor re-
search.10 In most cases the scaffold is accommodated at
the S3/S4 â-sheet region and acts as an amino acid
mimetic with differing degrees of conformational con-
straint (Figure 1).
The indole scaffold is highly rigid compared to many
other scaffolds. We believed that the lack of flexibility
and adaptability of the indole scaffold could be a
superior prerequisite for addressing secondary interac-
tions. Although so far we had not exploited the pos-
sibilities of addressing subsites of the fXa enzyme by
substituting the cyclic perimeter of the indole nucleus,
we were intrigued by the potential for increasing
binding affinity through interactions in this region. In
addition, only a few reports have systematically ex-
* To whom correspondence should be addressed. Phone:
+49(0)69 30515150. Fax: +49(0)69 331399. E-mail: Marc.Nazare@
sanofi-aventis.com.
10.1021/jm0490540 CCC: $30.25 © 2005 American Chemical Society
Published on Web 06/11/2005

4512 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14
Nazare´ et al.
Figure 1. Highly potent fXa inhibitors 1,9 2,11 312 with amino acid mimetic scaffolds.
Scheme 1. Synthetic Routes to the P1 and P4 Ligandsa
used throughout this investigation. The P4 ligand 5 was
prepared from commercially available N-Boc protected
4-aminopiperidine 4 in two steps by reductive amination
with acetone followed by acid-catalyzed removal of the
Boc group. The biaryl bromide 9 was synthesized by
condensation of 1-(5-chlorothiophen-2-yl)ethanone 6
with oxalic acid diethyl ester followed by regioselective
formation of the isoxazole ring by treatment with
hydroxylamine. Reduction of the ester in 8 and subse-
quent bromination yielded bromide 9.
The benzo-substituted indole nuclei were either com-
mercially available or readily accessible through stan-
dard indole synthesis procedures.16 Scheme 2 gives a
representative illustration of key transformations for
the synthesis of the substituted indole scaffolds, to-
gether with the two principal reaction sequences to
attach the P1 and P4 ligands to give the fully decorated
inhibitors. The 2-carboxy-6-methylindole scaffold 11 was
synthesized by reaction of 2-chloro-5-methylaniline 10
with 2-oxopropionic acid under palladium catalysis.17
Inhibitor 13 was then assembled by amide coupling with
5 using TOTU, to give amide 12, followed by N-
alkylation of the indole with bromide 9 in the presence
of sodium hydride. The 3-methylindole 17 was con-
structed employing a similar procedure, exemplifying
the generality of this strategy. The use of 2-oxobutanoic
acid allowed the introduction of the 3-methyl substitu-
ent directly during the annulation reaction to give
compound 15. The other pathway for constructing the
fully decorated inhibitors is illustrated by the synthesis
of compound 21. 7-Methyl-2-carboxyindole 18 was con-
verted into its methyl ester under standard conditions
followed by selective iodination of the 3-position and
palladium-catalyzed exchange with cyanide to yield 19.
This precursor was first alkylated with bromide 9 using
sodium hydride, and after saponification to the acid 20,
amide coupling with 5 led to the desired compound 21.
All other compounds described were synthesized analo-
gously.
a Reagents and conditions: (a) acetone, NaBH3CN, AcOH,
methanol, room temp, 16 h, 80%; (b) HCl/MeOH, room temp, 16
h, 100%; (c) KOt-Bu, then oxalic acid diethyl ester, toluene, room
temp, 3 h, 62%; (d) NH2OH‚HCl, ethanol, reflux, 6 h, 89%; (e)
NaBH4, ethanol, room temp, 12 h, 67%; (f) NBS, PPh3(polystyrene-
bound), CH2Cl2, room temp, 1 h, 77%.
plored and elucidated by X-ray crystallography the
opportunities of secondary interactions provided by
means of scaffold substitutions.13,14 We intended, there-
fore, to explore the influence of various substituents and
substitution patterns at the indole scaffold on the
inhibition of fXa and to characterize the interactions
using X-ray crystallography and molecular modeling.
The inhibitors synthesized contained an optimal com-
bination of P1 and P4 ligands, allowing us to probe the
interplay of secondary interactions of the modified
indole nucleus with the fXa protein, both in terms of
inhibitory activity and through X-ray crystallography.
Here, we report our findings on the effect of several
substitution patterns resulting in a dramatic enhance-
ment in binding affinity by specifically addressing
subsites of the fXa enzyme, leading to a 0.07 nM15
inhibitor 21, and provide a stringent analysis of key
interactions based on X-ray crystal structures of inhibi-
tor/fXa complexes.
The 3-chloro- and bromo-substituted inhibitors 23 and
24 were synthesized by selective halogenation of inhibi-
tor 1 using N-chlorosuccinimide or N-bromosuccinimide,
respectively. The 3-fluoroindole inhibitor 22 was syn-
thesized analogously, using N-fluoropyridinium triflate.
Using these synthetic routes, we were able to sys-
tematically synthesize a wide range of fXa inhibitors
and obtain dense SAR data that revealed several key
features of favorable and unfavorable interactions.
2. Chemistry
We first set out to systematically probe the influence
on fXa activity of a diverse set of simple substituents
on the indole nucleus. The biarylic P1 ligand and the
isopropylpiperidine P4 ligand were kept constant in this
series of compounds and served as strong anchor points
for the putative inhibitors. Scheme 1 illustrates the
synthesis of the key P4 and P1 building blocks 5 and 9

Two Binding Modes for Factor Xa Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14 4513
Scheme 2. Synthetic Routes to Substituted Indole Scaffolds and Inhibitorsa
a Reagents and conditions: (a) 2-oxopropionic acid, [Pd(t-Bu3P)2], K3PO4, AcOH, MgSO4, DMA 140 °C, 44%; (b) 5, TOTU,
N-ethylmorpholine, CH2Cl2, room temp, 16 h, 59%; (c) 9, NaH, DMF, 70 °C, 16 h, 15% for 13 and 10% for 17; (d) 2-oxobutanoic acid,
Pd(OAc)2, DABCO, MgSO4, DMF 105 °C, 55%; (e) 5, BOP-Cl, NEt3, CH2Cl2, room temp, 16 h, 99%; (f) HCl/MeOH, room temp, 16 h, 100%;
(g) (i) I2, DMF, room temp, 16 h, 94%; (ii) CuCN, Et4NCN, dppf, Pd2(dba)3, DMF/THF, 120 °C, 5 h, 15%; (h) 9, NaH, DMF, room temp,
16 h, then LiOH, THF/H2O, 60 °C, 2 h, 39% two steps; (i) 5, TOTU, N-ethylmorpholine, CH2Cl2, room temp, 16 h, 59%; (j) (i) Hal ) F,
N-fluoropyridinium triflate, CH2Cl2, room temp, 4 days, 50%; (ii) Hal ) Cl, NCS, CH2Cl2, room temp, 16 h, 36%; (iii) Hal ) Br, NBS,
CH2Cl2, room temp, 16 h, 40%.
3. Results and Discussion
in this series. Compounds 13, 29, 36, 37, with chlorine
replaced by a methyl group with different steric and
electrostatic properties, showed a different affinity
pattern. The 4-methylindole 36 was equipotent to refer-
ence compound 1, whereas compound 29 was 3 times
less potent. Surprisingly the 6-methyl compound 13
with a Ki of 18 nM was the least active methyl-
substituted derivative in this study. Most remarkably,
the 7-methyl derivative 37, with a Ki of 0.25 nM, was
more than an order of magnitude more potent than 1.
Combining the substituents in the 4- and 7-position of
compounds 36 and 37 to give dimethyl derivative 38 did
not improve on the remarkable affinity of 37. The most
potent compound in this series was the 7-nitro deriva-
tive 39 (Ki ) 0.1 nM), which, because of the presence of
the undesirable nitro group, was not considered as a
suitable druglike structure for further investigation. A
comparison of the affinity of compounds 35 and 37
suggests that the chloro substituent (Ki ) 0.7 nM) fits
almost as well into the putative enzyme subpocket as
the methyl substituent (Ki ) 0.25 nM). The structural
basis for the observed SAR was elucidated using X-ray
crystallography of selected inhibitors complexed with
fXa and will be discussed below.
3.1. Structure-Activity Relationship at the In-
dole Scaffold. Table 1 shows the effect on fXa inhibi-
tory activity of substitution at the benzo ring of the
indole. The P1 and P4 ligands of the compounds in Table
1 were kept constant to allow the influence of the indole
substitution to be observed in isolation. We first chose
to investigate substitution at the 5-position of the indole
because many of the starting materials were com-
mercially available or easily accessible in high yield by
standard indole synthesis. A fluoro substituent in this
position (25) did not affect fXa binding affinity in
comparison to the unsubstituted reference compound 1
(Ki ) 3 nM). Increasing the size of the halogen (com-
pounds 26-27, Ki ) 5 and 9 nM) reduced fXa binding
affinity, and the introduction of a large benzyloxy
substituent was extremely deleterious to affinity (28,
Ki ) 835 nM). The 5-methyl derivative 29 (Ki ) 9 nM)
was 3-fold less potent than 1. Compounds 30-32 show
the effect of increasing the polarity of the substituent
at this site of the indole ring, with small polar moieties
being most favorable at this position (32, NO2 Ki ) 1
nM) and the methylsulfone derivative 31 clearly being
too large (20 nM). Introduction of a single chloro
substituent at positions 4, 5, and 6 of the indole
(compounds 26, 33, and 34) resulted in a 2-fold loss in
binding affinity compared to the unsubstituted reference
compound 1. The position of the chloro substituent had
almost no influence on the potency of the compounds
3.2. Structure-Activity Relationship at the In-
dole 3-Position. The remarkable increase in fXa bind-
ing affinity resulting from introduction of an optimal
substituent in the 7-position of the indole encouraged
us to investigate substitution at the 3-position. The

4514 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14
Table 1. Substitution on the Benzo Ring of the Indole Scaffold
Nazare´ et al.
Table 2. Substitution at the 3-Position of the Indole Scaffold
a Ki values were measured as described in ref 15.
1. Remarkably, the introduction of a nitrile moiety in
the 3-position resulted in a 10-fold increase in fXa
binding affinity, giving compound 41 with a Ki value of
0.3 nM. Combining the favorable interactions of the
7-methyl group used in compound 37 and the 3-nitrile
group yielded the most active inhibitor 21 with a Ki
value for fXa inhibition of 0.07 nM. The 3-nitrile
substitution was also capable of compensating for less
favorable substitutions such as methylsulfone 42, which
is approximately 3 times more potent than its parent
compound 31.
3.3. Exploring Alternative P1 Substituents. With
this encouraging structure-activity relationship data in
hand, we then intended to determine how general these
interactions would be in the series and in particular if
this substitution pattern would also display a superior
fXa binding affinity in combination with other P1
ligands. Table 3 shows the results for a set of six
2-carboxyindole fXa inhibitors, containing three differ-
ent P1 ligands and that are, in each case, unsubstituted
at the indole core or contain the 3-cyano,7-methyl
substitution pattern that was found to be optimal in
combination with the biaryl P1 ligand.
Surprisingly the attempt to transfer the optimal
indole substitution pattern led to compounds that were
significantly less active than the initial unsubstituted
ones. For example, when combined with the m-meth-
oxybenzyl P1 ligand in 44, a 4-fold drop in activity was
a Ki values were measured as described in ref 15.
results are shown in Table 2. Introduction of lipophilic
steric bulk in the form of halogen or methyl substituents
(compounds 22-24, 17) did not influence potency sig-
nificantly compared to the unsubstituted reference 1
(e.g., 23, 3-Cl, Ki ) 2 nM). Compound 40, with a larger
planar phenyl substituent was 18 times less potent than

Two Binding Modes for Factor Xa Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14 4515
Table 3. Effect of 3-Cyano,7-Methyl Substitution of the Indole
Scaffold in Combination with Different P1 Ligands
Table 4. Data Collection and Refinement Statistics for Five
Factor Xa X-ray Structures
inhibitor
1
21
43
45
47
detector
ID14-1a mar345 mar345 mar300 mar300
P212121 P212121 P212121 P212121 P212121
space group
cell dimensions
a, Å
b, Å
c, Å
56.1
72.0
79.2
55.40
70.65
76.27
55.94
72.05
78.55
56.38
71.90
78.63
56.3
72.0
78.3
obsd reflns
unique reflns
resolution, Å
Rsym, %
completeness
protein atoms
inhibitor atoms
calcium ions
water molecules
R-factor, %
R-free, %
72 973 18 168 73 686 33 430 45 424
16 670 5984
15 828 6939
12 329
2.4
6.7
95.2
2249
32
2.2
7.9
3.0
7.2
2.2
2.95
10.3
97.4
2240
32
5.1
98.8
2249
33
94.0
2240
36
94.8
2240
30
1
1
73
1
1
1
221
20.7
25.3
290
21.8
27.2
0.009
1.30
24.5
0.72
126
18.4
26.1
0.009
1.58
24.3
0.67
291
18.1
28.0
0.008
1.46
24.7
1.90
20.2
29.8
0.008
1.29
24.0
0.78
rmsd bond lengths, Å 0.009
rmsd bond angles, deg 1.51
rmsd dihedrals, deg
rmsd impropers, deg
25.2
0.75
a Beam line ID14-1 at the European Synchrotron Radiation
Facility (ESRF).
(Ki ) 89 nM), and 45 (Ki ) 3 nM). In addition, one
previous structure of an indole-2-carboxamide with an
m-benzamidine in S1 was also used for analysis (PDB
ID code 1lqd, 4718). All crystals diffracted reasonably
well with resolutions between 2.2 and 3.0 Å (see
Experimental Section, Table 4). In general, the inhibitor
binding modes are well-defined. Thus, the 2.2 Å struc-
tures are accurately defined, while the 3.0 Å structure
is sufficiently accurate to indicate the binding mode.
While many first-generation fXa inhibitors rely on the
interaction of a basic amidine with Asp189 at the bottom
of the S1 pocket, these structures indicate favorable
nonbasic interactions with residues in this subsite.
Collectively these X-ray structures reveal that the
chlorothiophene, chlorophenyl, or m-methoxyphenyl
substituents are accommodated in S1. However, differ-
ent interaction patterns for chloro or methoxy substit-
uents were identified.
We will start our discussion with the X-ray structure
of the unsubstituted reference compound 1 (Ki ) 3 nM
2.2 Å) to rationalize the structure-activity relationship
in this series. Its overall binding mode in factor Xa is
shown in Figure 2. While Figure 2A summarizes key
interactions of 1 with important amino acids, the
detailed binding mode is shown in parts B-D of Figure
2. The protein binding site in Figure 2B,D is displayed
as a solvent-accessible surface with cavity depth mapped
onto it. Orange parts indicate the very deep and buried
S1 pocket, while green and cyan parts refer to more
shallow and solvent-exposed subsites, such as the S4
and EBP (“ester binding pocket”). Here, the protease
S2 pocket is present as a small, solvent-exposed subsite
inaccessible for inhibitors because of the aromatic Tyr99
separating S4 from S2.
a Ki values were measured as described in ref 15.
observed compared to the unsubstituted compound 43
(Ki ) 406 vs 89 nM). Even more remarkable was the
6-fold loss in activity in conjunction with the 4-chlo-
rophenylacetamide as the P1 ligand (46 Ki ) 26 nM vs
45 Ki ) 3 nM). This unexpectedly divergent SAR
behavior strongly suggested that the inhibitors in Table
3 with different P1 ligands have different binding modes
to the FXa enzyme and emphasized the dominance of
the P1 ligand interaction with fXa.
3.4. Binding Modes of Indoles with Biaryl P1
Substituents. These conflicting SAR observations
and the exceptional binding affinity of the 3-cyano,7-
methyl derivative 21 with a chlorobiaryl P1 substituent
prompted us to investigate the binding modes of selected
derivatives by X-ray crystal structure analysis of their
complexes with fXa.
Although the electron density of compound 1 is well-
defined, the indole benzo ring and the S4 isopropyl
moiety reveal a weaker density, which points to some
disorder or conformational mobility in those parts filling
the solvent-exposed regions of S4 and EBP.
We obtained four X-ray crystal structures of factor Xa
with compounds 1 (Ki ) 3 nM), 21 (Ki ) 0.07 nM), 43

4516 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14
Nazare´ et al.
Figure 2. A 2.2 Å X-ray crystal structure of indole-2-carboxamide 1 (Ki ) 3 nM) in complex with human fXa. (A) Schematic view
of essential intermolecular interactions. (B) Binding mode with fXa shown as solvent-accessible surface color-coded by cavity
depth (from orange to blue). Essential pockets are indicated as S1, S2, S4, and EBP (ester binding pocket). Structurally conserved
water is displayed with cyan spheres. Essential amino acids are indicated. (C) Comparison of X-ray derived binding modes for 1
(Ki ) 3 nM, white carbons) and 21 (Ki ) 0.07 nM, orange carbons) with the P1 biaryl substituent in complex with factor Xa. (D)
Same as (C) but with fXa solvent-accessible surface indicated. (E) Detailed analysis of the binding site from the fXa/21 complex
(white carbons) for favorable interactions with a GRID methyl probe (orange contours). Only key amino acids surrounding the
EBP are shown. The favorable contour is highlighted using an orange circle. Key lipophilic contacts to Arg143 and Gln192 side
chains are indicated and shown as dotted lines. In addition, the best docking mode for 35 (Ki ) 0.7 nM, orange carbons) is indicated.
(F) Comparison of the fXa/21 (ligand with white carbons, protein with purple carbons) and best docking mode for 39 (ligand and
protein with cyan carbons). Polar interactions of the 7-NO2 group with fXa side chains are indicated.
The chlorothiophene substituent of 1 and related
molecules is deeply buried inside the S1 pocket with the
chloro atom pointing toward the center of the Tyr228
aromatic ring. This atom replaces a structurally con-

Two Binding Modes for Factor Xa Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14 4517
served water molecule present in benzamidine-contain-
ing factor Xa structures. It is engaged in a lipophilic
contact to the aromatic ring of Tyr228 with distances
of 3.43-4.52 Å to the aromatic carbons and 3.76 Å to
the ring centroid, in agreement with interaction data
from force field methods such as GRID, which consis-
tently predict a region of favorable interaction to Cl at
the bottom of this pocket (see Figure 2F). The carbon-
chlorine bond is directed toward the plane of the Tyr228
both the Gln192 side chain and the indole is weak,
which indicates some flexibility in this region. Inspec-
tion of Figure 2B also reveals an unoccupied subpocket
between indole C7 and the aliphatic part of the Gln192
side chain.
In Figure 2C,D the binding mode of 1 (white carbons)
is compared with the X-ray structure of the most active
inhibitor 21 (Ki ) 0.07 nM, orange carbons). The latter
structure with a resolution of 3.0 Å provided additional
insights into favorable interactions of both the 7-methyl
and 3-nitrile groups. The hydrophobic 7-methyl group
now fills the space between the indole C7 atom and the
aliphatic portion of the Gln192 side chain. This flexible
side chain is reoriented compared to its position in the
fXa/1 complex (distance Gln192-Cδ 1.92 Å) and is in
lipophilic contact with the methyl substituent (distance
of 7-CH3 group to Gln192-Câ is 3.81 Å). This favorable
contribution of a methyl carbon to binding affinity is in
agreement with the analysis of favorable binding site
interactions using GRID (cf. Figure 2E). Lipophilic
contacts between the alkyl part of the Gln192 side chain
and this methyl group or a chlorine atom contribute
favorably to binding affinity ((35) 0.7 nM vs (37) 0.25
nM), as can be seen from the superimposition of the
experimental binding mode of 21 in comparison to the
best docking mode of 37 (orange carbons) in Figure 2E.
From docking studies it can be concluded that the
chlorine substituent in position 7 of the indole system
also occupies this small subpocket and is able to interact
via lipophilic contacts to both amino acid side chains.
Moreover, the addition of a hydrogen bond acceptor to
the indole-C7 position additionally favors binding af-
finity because of the surrounding hydrogen bond donors
from the flexible Arg143 and Gln192-NꢀH2 side chains
(e.g., distance C7-Gln192-Nꢀ is 2.75 Å in the fXa/1
complex). This is in good agreement with the remark-
ably high affinity of compound 39 (Ki ) 0.10 nM) having
an NO2 substituent in this position. The best docking
mode of 39 is shown in Figure 2F (cyan carbons) in
comparison to the experimental fXa/21 complex. Be-
cause all docking studies were carried out using flexible
key residues, the side chains of Gln192 and Arg143
(cyan carbons in Figure 2F) move in order to undergo
the proposed polar interaction with the NO2 group 39
in this subpocket.
Interestingly, dimethyl substitution at positions 7 and
4 of the indole system results in reduced affinity
compared to the 7-monomethyl substitution (38 Ki ) 1
nM vs 37 Ki ) 0.25 nM). Because the 7-methyl substitu-
tion causes a minor shift of the indole scaffold within
the EBP toward Arg222, the reduced affinity can be
attributed to the closer contact between any bulky C4
substituent and the Arg222 guanidinine (distance C4-
Arg222-Nꢀ of 4.34 Å in the fXa/21 complex). The
inhibitor with the same 4-methyl group but without
7-substitution does not undergo this slight shift in the
EBP and thus shows the same binding affinity as its
7-unsubstituted derivative (36 Ki ) 3 nM vs 1 Ki ) 3
nM). Substitutions at the indole 5-position also reduce
affinity because of less favorable steric contacts to amino
acids at the edge of the EBP, namely, the Arg222 side
chain and the backbone carbonyl group of Glu147
(distances C5-Arg222-Nꢀ of 4.44 Å and C5-Glu147-
CdO of 3.80 Å in the fXa/1 complex).
ring (dihedral angle C-Cl‚‚‚Tyr228(centroid)-Tyr228(Cz)
:
118°). Additional details are obvious from inspection of
Figure 2C, where 1 (white carbons) and 21 (orange
carbons, Ki ) 0.07 nM) are superimposed. This binding
mode of chlorine in S1 is similar to that previously
described for 3-oxybenzamides with a dichlorophenyl
substituent in S1.13,14 This “chloro-binding mode” has
been described by different groups for thrombin19 and
fXa,8,14 indicating that electrostatic interactions between
a benzamidine and Asp189 in S1 are not mandatory for
high affinity. The thiophene sulfur atom is engaged in
close contact with Val213-Cγ (3.40 Å) and Trp215-CR
(3.59 Å) lining the S1 pocket, while its carbon atoms
are oriented toward the Asp189 side chain at the bottom
of S1 (3.31-3.47 Å to Asp189-Oγ1 and -Oγ2). The
isoxazole group is situated close to the active site
Ser195-Oγ, allowing reasonable hydrogen bonds to be
formed between the Oγ and both the isoxazole nitrogen
(3.23 Å) and oxygen (3.04 Å) atoms (cf. Figure 2A and
Figure 4E). Profiling the ligand 1 using the GRID force
field and a hydrogen bond donating probe like amide
NH or phenolic OH suggests, however, that the isox-
azole nitrogen is a better hydrogen bonding acceptor.
The isoxazole carbon is oriented toward Gly219-CdO
(4.15 Å). Replacing the isoxazole ring by an amide, as
in 45, results in a short hydrogen-bonding interaction
between the ligand amide NH group and Gly219 (2.91
Å), while interaction with Ser195 is no longer possible
(CdO‚‚‚Oγ: 4.70 Å). This may imply the involvement
of a structurally conserved water, although this is not
detected in the X-ray structure.
No other direct hydrogen bonds were observed be-
tween factor Xa and 1. The indole-2-carboxamide car-
bonyl oxygen is 3.65 Å from the Gly216-N, which is
probably too long for a significant interaction. Again, a
structurally conserved water molecule bridging both
partners could be postulated here. The twist between
the planes of the indole and amide groups is 25°. The
isopropylpiperidine substituent is bound in the S4
pocket, surrounded by aromatic residues Phe174, Tyr99,
and Trp215 at the bottom of the S4 pocket. Hydrophobic
interactions are dominant in this pocket, but basic
nitrogen atoms are also accommodated on top of the
Trp215 ring system, as present in this and related
structures. In contrast to the X-ray structures of 3-oxy-
benzamides with a 4-pyridylpiperidine substituent,14 the
basic pyridyl nitrogen is missing here. Hence, no polar
interaction with the network of hydrogen-bonding ac-
ceptors at the edge of this pocket or structurally
conserved water is possible.
The indole benzo ring is situated on top of the
Cys191-Cys220 disulfide bridge forming the bottom of
the so-called “ester binding pocket” (EBP).13 This pocket
is lined by the flexible side chains of the fXa residues
Gln192, Arg143, and Glu147. The electron density of

4518 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14
Nazare´ et al.
Figure 3. Comparison of X-ray binding modes for different indole-2-carboxamides in complex with human factor Xa. The following
compounds are shown: (A) 21 (0.07 nM), (B) 45 (3 nM), (C) 43 (89 nM), (D) 4718 (25 nM). The experimental factor Xa binding site
is shown as a solvent-accessible surface color-coded by cavity depth (from orange to blue). Structurally conserved water is displayed
as cyan spheres.
The 3-cyano substituent is oriented on top of the
Glu217-CR atom (3.59 Å) close to the position of a
structurally conserved water in the fXa/1 complex (2.03
Å), which has been replaced in the fXa/21 complex.
Profiling ligand 21 using a GRID C.3 probe additionally
suggests that this cyano substituent is involved in
favorable interactions with aliphatic carbons, such as
Glu217-CR. This cyano group with a larger dipole
moment has a positive effect on binding affinity, shown
by comparing the unsubstituted derivative 1 (Ki ) 3 nM)
with 41 (Ki ) 0.30 nM). Its nitrogen atom with a partial
negative charge is close to the CR carbon of Glu217 (3.59
Å) and its attached HR (2.63 Å), which both are more
positively charged. Furthermore, the guanidine of the
neighboring Arg222 creates a more positive electrostatic
field in this region. Analysis using GRID and an sp
nitrogen with lone pair indeed indicates favorable
interactions in this area close to the cyano group.
Obviously this electrostatic complementarity contributes
positively to the fXa binding affinity because shorter
and/or less charged substituents such as 3-methyl,
chloro, and bromo are less potent fXa inhibitors with
Ki values of 3 nM (17), 2 nM (23), and 2 nM (24) in
comparison to 41 (Ki ) 0.30 nM). Furthermore, the
longer cyano substituent is directly situated on the
surface of the binding site and thus also increases
affinity by two favorable contacts. The final effect
contributing to fXa binding affinity originates from an
enhanced twist of the indole-2-carboxamide dihedral
angle from 25° to 46° caused by the sterically more
demanding indole 3-substitution. Although this leads
to a strained ligand geometry, its overall contribution
for binding affinity is positive because an additional
hydrogen bond between the amide CdO and Gly216-
NH is formed (3.41 Å).
3.5. Binding Modes of Indoles with Other P1
Substituents. To understand the nonlinear structure-
activity relationship upon transfer of the indole 3-cy-
ano,7-methyl substitution pattern from 21 to different
S1-directed substituents (Table 3), the X-ray crystal
structure of fXa complexed with 45, carrying a 4-chlo-
rophenylacetamide substituent directed toward S1, was
solved (resolution 2.95 Å, Ki ) 3 nM). The reasonably
strong electron density of inhibitor 45 is clear, although
the position of the terminal S4 isopropyl is less well-
defined and a mixture of conformations may be present,
which may explain why a different rotamer was fitted
at this position compared to the fXa/1 complex.
In Figure 3 the binding modes for indole-2-carbox-
amide based inhibitors 21, 45, 43, and 4718 in fXa,

Two Binding Modes for Factor Xa Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14 4519
Figure 4. Superpositions of experimentally derived inhibitor binding modes from Figure 3 (orange carbons) with compound 1
(white carbons) from the crystal structure of its complex with human factor Xa. All superpositions are shown without protein
binding sites (A-D). For comparison, each individual ligand is additionally displayed with key amino acids highlighted and with
a focus on the factor Xa S1 pocket plus important interactions highlighted (E-H): (A, E) 21 (0.07 nM); (B, F) 45 (3 nM); (C, G)
43 (89 nM); (D, H) 4718 (25 nM).
indicated by the binding site solvent-accessible surfaces
and cavity depth for color coding, are compared. The
superposition of these inhibitor binding modes (orange
carbon atoms) onto the fXa/1 complex is shown in Figure

4520 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14
Nazare´ et al.
4A-D (white carbons). This superposition was done
using rms fitting of the entire protein domain, while
only essential amino acids in S1 are displayed in Figure
4E-H.
molecules and one with the backbone nitrogens of
Ala190 and Ala221, respectively. The observation that
the methyl group points toward Asp189 (Figure 4G)
could possibly be explained by the involvement of both
Asp189 carboxylate oxygens in two hydrogen bonds
each, one with a bound water molecule and one with
the backbone nitrogens of Ala190 and Ala221. Thus, no
unsatisfied hydrogen-bonding partners are left in S1.
Because of a slight movement (∼0.5 Å) of the struc-
turally conserved water in the S1 pocket, together with
a small rotation of the benzene ring, both benzamidine
and methoxyphenyl moieties are able to form a good
hydrogen bond with the bound water molecule, as
indicated in Figure 4G,H, which may explain the
observed binding affinity of these compounds.
The chlorine atom of the 4-chlorophenylacetamide
moiety in 45 is situated 0.55 Å deeper in S1 compared
to the fXa/1 complex; its distance to the Tyr228 ring
centroid is reduced to 3.44 Å. The acetamide nitrogen
is at a distance of 2.84 Å from the backbone Gly219-
CdO, suggesting a positive contribution from this
hydrogen bond. It causes the acetamide carbonyl oxygen
to be 4.59 Å away from the Oγ atom of Ser195, which
was interacting with the isoxazole ring in the biaryl-
P1 substituent in 1. Surprisingly, the orientation of the
indole scaffold in 45 with this P1 substituent is perpen-
dicular to the binding mode of indole-2-carboxamides
with biaryl substituents. The binding mode of the
isopropyl piperidine in S4 is, however, similar to that
in the fXa/1 and fXa/21 complexes.
While Figures 2 and 4A,E demonstrate the similar
binding modes for the biaryl compounds 1 and 21, other
P1 substituents led to a significant change of the indole
orientation. Detailed inspection of Figures 3B and 4B,F
for the fXa/45 complex reveals the indole C7 atom in
its perpendicular orientation to be in close contact with
the Cys191-Cys220 disulfide bridge (C7-Cys220-Sγ of
3.92 Å; C7-Cys191-Sγ ïφ 4.22 Å) at the bottom of the
EBP. The neighboring C6 indole atom is involved in
close contacts to Gln192-Câ (3.56 Å) and to the carbonyl
oxygen of Glu147 (3.74 Å), while the indole C5 carbon
is close to the flexible side chain of Arg143 (e.g., to the
guanidine carbon, 3.61 Å). The entire indole benzo ring
plane is thus packed against the aliphatic part of the
shifted Gln192 side chain. This binding mode now
orients the indole C3 and C4 positions toward the
solvent. It allows the 2-carboxyamide oxygen to be
engaged in a hydrogen bond with Gly216-NH (3.37 Å),
with a dihedral angle of -37.1° between the indole and
amide planes. These changes in binding mode and
significant interactions collectively contribute to the
observed binding affinity of 3 nM for 45. Interest-
ingly, in all structures the isopropylpiperidine moiety
is consistently situated in the S4 pocket between
Tyr99, Phe174, and Trp215 without conformational
changes.
The X-ray structure of 43 from the m-methoxybenzyl
P1 series (Ki ) 89 nM) was solved to a resolution of 2.2
Å. The electron density of the inhibitor is clear, its
binding mode similar to the fXa/45 complex with a
perpendicular orientation of the indole scaffold (cf.
Figures 3C and 4C,G). The m-methoxyphenyl group is
located in S1 with its aromatic carbons at positions
similar to positions of the corresponding benzamidines
(e.g., 47,18 Figures 3D and 4D,H). The oxygen atom is
involved in a hydrogen bond to a structural water
molecule (3.00 Å), which in turn is hydrogen-bonded to
both the backbone nitrogen and carboxyl oxygen of
Ile227 with distances of 3.41 and 3.08 Å, respectively
(cf. Figure 4C,G). The methoxy oxygen atom is also close
to the carbonyl oxygen of Ala190 (3.44 Å). As can be
seen from Figure 4G, the carbon of the methoxy sub-
stituent points toward Asp189 at the bottom of S1 (2.93
and 3.17 Å). Both carboxylate oxygens of Asp189 are
also involved in hydrogen bonds: one with bound water
These different indole orientations of 43 and 45 in
complex with fXa compared to 1 suggest that any
substituent attached to the aromatic C7 carbon affects
this binding mode and results in a reduced affinity in
comparison with unsubstituted analogues. This is in
perfect agreement with the SAR of the m-methoxyben-
zyl P1 series with a 4-fold reduction of affinity for the
corresponding 3-cyano,7-methyl derivative 44 (Ki ) 406)
versus the unsubstituted derivative 43 (Ki ) 89 nM). It
also agrees with the experimental observations for the
4-chlorophenylacetamide P1 series with the correspond-
ing 3-cyano,7-methyl derivative 46 (Ki ) 26 nM) com-
pared to the unsubstituted compound 45 (Ki ) 3 nM).
Again, the perpendicular orientation of the indole scaf-
fold explains the reduced binding affinity by introducing
substituents into the indole C3 and C7 positions. Hence,
the divergent SAR of the compounds shown in Table 3
is explained by differing binding modes of the central
scaffold, depending on the nature of the P1 substituent.
Finally, the binding mode of 4718 with a 3-amidi-
nobenzyl substituent attached at the indole N1 indicates
that this substituent is located in the S1 pocket and
interacts with Asp189 at the bottom of this pocket (cf.
Figures 3D and 4D,H). This is the typical binding mode
observed for other benzamidine-based factor Xa inhibi-
tors. The indole scaffold itself is solvent-exposed and
stacked upon the flexible side chain of Gln192. This
orientation is similar to that observed for compounds
43 and 45 with the indole C7 in van der Waals contact
to the Cys191-Cys220 disulfide bond at the bottom of
the EBP. For compound 47, the terminal dichlorobenzyl
substituent is oriented almost parallel to the aromatic
side chain of Trp215 at the bottom of S4 and perpen-
dicular to the rings of Tyr99 and Phe174. The two Cl
atoms attached to the benzyl ring in this pocket have
strong electron density. Cl-1 points deep into the S4
pocket, while Cl-2 points toward the solvent and in the
direction of Glu217. Hence, all alternative P1 motifs
except the biaryl substituent result in this perpendicular
indole orientation, while this rigid moiety provides an
ideal topology to orient the indole scaffold in a position
optimally suited to access previously unoccupied pockets
of the factor Xa binding site by additional substituents.
This finding offers many possibilities for further design
of potent fXa inhibitors with neutral P1 substituents.
4. Conclusion
In this study we have explored the SAR of a series of
2-carboxyindole factor Xa inhibitors by systematic sub-

Two Binding Modes for Factor Xa Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14 4521
of piperidin-4-yl-carbamic acid tert-butyl ester (4) (5.0 g, 25
mmol) in 15 mL of methanol (15 mL) were added acetone (8
mL), Na(CN)BH3 (3.14 g, 50 mmol), and acetic acid (0.3 mL).
After the mixture was stirred for 16 h at room temperature,
the solvent was removed under reduced pressure and the
residue was partitioned between water (30 mL) and ethyl
acetate (30 mL). The organic layer was washed with saturated
Na2CO3 solution and water and then dried over Na2SO4. The
solvent was removed under reduced pressure to give 4.8 g
(80%) of product as a white solid. MS (ES) m/z: 243.4 (MH+).
stitution at all positions of the indole scaffold. Address-
ing subpockets in the enzyme, such as the EBP with
suitable substituents, which allowed additional lipo-
philic and hydrogen-bonding interactions, resulted in a
significant improvement in binding affinity. Further-
more, combination of optimal substituents resulted
in an even more dramatic increase in potency. The
3-cyano,7-methyl substitution pattern was found to be
an optimal arrangement in combination with the chlo-
rothienylisoxazole P1 moiety and the 4-isopropylpiperi-
dine P4 portion (21, Ki ) 0.07 nM). Surprisingly this
substitution pattern was deleterious when combined
with a 4-chloroanilinacetamide or an m-methoxybenzyl
P1 ligand. This divergent SAR behavior is explained by
X-ray crystallographic analysis of fXa/inhibitor com-
plexes. It became apparent that, in contrast to analo-
gous P1-benzamidine-containing inhibitors, the chlo-
rothienylisoxazole P1 ligand induced an unexpected 90°
rotation of the indole scaffold, making the subpockets
identified accessible with this alternative binding mode.
The inhibitors with 4-chloroanilinacetamide or m-meth-
oxybenzyl P1 ligands displayed the upright positioning
of the indole core described previously for analogous
P1-benzamidine-containing inhibitors, and thus, the
3-cyano,7-methyl substitution pattern resulted in un-
favorable interactions. Another surprising feature of this
inhibitor series is that the 4-isopropylpiperidine P4
ligand was capable of mediating the required interac-
tions in the S4 pocket for both binding modes. These
results underscore the key role played by the P1 ligand
not only in determining the binding affinity by direct
interaction of its neutral chlorine atom to the aromatic
ring of Tyr228 but also in modifying the binding mode
of the entire scaffold, resulting in a nonlinear SAR.
Exploiting this behavior may potentially allow the
generation of other highly potent fXa inhibitors with
novel P1-ligand-scaffold combinations.
(ii) 1-Isopropylpiperidin-4-ylamine (5). To a solution of
(1-isopropylpiperidin-4-yl)carbamic acid tert-butyl ester (4.8 g,
20 mmol) in methanol (15 mL) was added methanolic hydro-
chloric acid (8 M, 20 mL), and the mixture was stirred for 16
h. Removal of the solvent under reduced pressure, followed
by removal of residual volatiles by twice coevaporating with
1
toluene, gave 4.4 g (100%) of product 5. H NMR (DMSO-d6):
δ 8.52 (s, br., 2H), 3.36 (d, J ) 6.5 Hz, 3H), 3.03 (d, J ) 11 Hz,
2H), 2.51 (s, 1 H), 2.15-2.01 (m, 4H), 1.26 (d, J ) 6.5 Hz, 6H).
MS (ES) m/z: 143.2 (MH+).
Ethyl 4-(5-Chlorothiophen-2-yl)-2,4-dioxobutanoate (7).
Potassium tert-butoxide (17.5 g, 156 mmol) was added to a
solution of 2-acetyl-5-chlorothiophene (6) (25 g, 156 mmol) in
anhydrous toluene (400 mL) at 0 °C under nitrogen in four
portions. The mixture was stirred at 0 °C for 10 min. Then
diethyl oxalate (25 mL, 187 mmol) was added via syringe, and
the resulting mixture was stirred at room temperature for 3
h. The precipitated product was collected by filtration, washed
with toluene, and then dissolved in ethyl acetate (500 mL).
The organic layer was washed with diluted aqueous hydro-
chloric acid (100 mL) and then dried over Na2SO4. Removal of
the solvents afforded the desired product 7 as a yellow solid
(25 g, 62%), which was used in the following reaction without
1
further purification. H NMR (CD3OD): δ 7.87 (m, 1H), 7.15
(m, 1H), 6.95 (s, 1H), 4.35 (q, 2H, J ) 7.1 Hz), 1.36 (t, 3H, J )
7.1 Hz). MS (ES) m/z: 261.0 (MH+).
Ethyl 5-(5-Chlorothiophen-2-yl)isoxazole-3-carboxy-
late (8). A mixture of ethyl 4-(5-chlorothiophen-2-yl)-2,4-
dioxobutanoate (7) (25 g, 96 mmol) and hydroxylamine hydro-
chloride (25 g, 360 mmol) in anhydrous ethanol (300 mL) was
refluxed for 5 h. The solvents were removed under reduced
pressure, H2O (150 mL) was added to the residue, followed by
ammonium hydroxide (28-30%) to give pH 7. The aqueous
mixture was then extracted with ethyl acetate (3 × 100 mL),
and the combined organic portions were dried over MgSO4.
After the solvents were removed, the crystalline product 8 (22
g, 89%) was used in the following reaction without further
purification. 1H NMR (CDCl3): δ 7.31 (d, 1H, J ) 4.0 Hz), 6.95
(d, 1H, J ) 4.0 Hz), 6.71 (s, 1H), 4.44 (q, 2H, J ) 7.1 Hz), 1.41
(t, 3H, J ) 7.1 Hz). MS (ES) m/z: 258.0 (MH+).
(5-(5-Chlorothiophen-2-yl)isoxazol-3-yl)methanol. So-
dium borohydride (16.1 g, 84.4 mmol) was added cautiously
to the solution of ethyl 5-(5-chlorothiophen-2-yl)isoxazole-3-
carboxylate (8) (22 g, 85 mmol) in anhydrous ethanol (250 mL)
at 0 °C. The mixture was stirred at room temperature for 12
h. Ethanol was removed under reduced pressure, the residue
was taken up in H2O (100 mL), and dilute aqueous hydrochlo-
ric acid was added to give pH 3-4. The precipitated product
was collected by filtration, washed with water, and dried in
vacuo to give a pale-yellow solid (16 g, 67%). 1H NMR
(CDCl3): δ 7.27 (d, 1H, J ) 3.9 Hz), 6.94 (d, 1H, J ) 3.9 Hz),
6.40 (s, 1H), 4.78 (s, 2H). MS (ES) m/z: 216.3 (MH+).
3-(Bromomethyl)-5-(5-chlorothiophen-2-yl)isoxazole (9).
N-Bromosuccinimide (14.5 g, 82 mmol) was added to a mixture
of (5-(5-chlorothiophen-2-yl)isoxazol-3-yl)methanol (3.50 g, 16.2
mmol) and polystyrene bound triphenyl phosphine (27 g, 81
mmol) in anhydrous methylene chloride (250 mL) at 0 °C. The
mixture was stirred at room temperature for 1 h. The reaction
mixture was filtered, and the filtrate was concentrated in
vacuo. The residue was purified by flash column chromatog-
raphy (10% of ethyl acetate/heptane) to yield 16.1 g (77%) of
the product 9 as a white solid. 1H NMR (CDCl3): δ 4.42 (s,
2H), 6.42 (s, 1H), 6.95 (d, 1H, J ) 4.0 Hz), 7.28 (d, 1H, J ) 4.0
5. Experimental Section
5.1. Chemistry. Solvents and other reagents were used as
received without further purification. Column chromatography
was carried out on Merck silica gel 60 (230-400 mesh).
Reversed-phase high-pressure chromatography was conducted
on a Abimed Gilson instrument using a LiChrospher 100 RP-
18e (5 µm) column from Merck. Thin-layer chromatography
was carried out on TLC aluminum sheets with silica gel
60F254 from Merck. LC-MS analyses were performed on an
Agilent series HP1100 system using a YMC J sphere ODS H80
20 mm × 2.1 mm (4 µm) column and a Merck Purosphere 55
mm × 2 mm (3 µm) column or on a Waters 2790 HPLC, using
a Waters Symmetry C18, 50 mm × 2.1 mm (3.5 µm) column.
Varying ratios of acetonitrile and 0.05% trifluoroacetic acid
or 0.1% formic acid in water were used as solvent systems.
NMR spectra were recorded in CD3OD, CDCl3, or DMSO-d6
either on a Bruker DRX 400 or Varian Unity Plus 300.
Chemical shifts are reported as δ values from an internal
tetramethylsilane standard. Mass spectral data were obtained
on either a VG Bio-Q triple quadropole mass spectrometer
using electrospray ionization or a VG ZAB 2-SEQ mass
spectrometer using FAB ionization. Accurate mass measure-
ments were conducted with a Bruker Apex III FTICR mass
spectrometer. Purity and characterization of compounds were
established by a combination of LC-MS, high-resolution mass
spectrometry (HRMS), and NMR analytical techniques. All
compounds were found to be >95% pure by LC-MS and HRMS
analysis.
1-Isopropylpiperidin-4-ylamine (5). (i) (1-Isopropylpi-
peridin-4-yl)carbamic Acid tert-Butyl Ester. To a solution

4522 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14
Nazare´ et al.
Hz), 6.95 (d, 1H, J ) 4.0 Hz), 6.42 (s, 1H), 4.42 (s, 2H). MS
(ES) m/z: 279.1 (MH+).
sealed with a serum cap, and argon was bubbled through the
solution for 10 min. Pd(OAc)2 (93 mg, 0.2 mmol) was added to
the solution, and argon was bubbled through the mixture for
an additional 10 min. The reaction mixture was heated to the
105 °C for 16 h. After cooling to room temperature, the reaction
mixture was filtered, dichloromethane (150 mL) was added,
and the mixture was extracted twice with 2 M NaOH (2 ×
150 mL). The combined aqueous layers were acidified to pH 2
with 2 M hydrochloric acid and extracted with dichlo-
romethane (3 × 150 mL). The combined organic layers were
dried over MgSO4 and filtered. After removal of the solvents
under reduced pressure, the crude product 15 (770 mg, 55%)
was used without further purification for the next reaction
Preparation of 1-[5-(5-Chlorothiophen-2-yl)isoxazol-
3-ylmethyl]-6-methyl-1H-indole-2-carboxylic Acid (1-Iso-
propylpiperidin-4-yl)amide (13) (Scheme 2). (i) 6-Methyl-
1H-indole-2-carboxylic Acid (11). 2-Chloro-5-methylaniline
(10) (2 g, 14.1 mmol), 2-oxopropionic acid (3.7 g, 42.4 mmol),
acetic acid (1.2 mL, 21.2 mmol), and MgSO4 (0.85 g, 7.1 mmol)
were suspended in dimethylacetamide (100 mL). The tube was
sealed with a serum cap, and argon was bubbled through the
solution for 10 min. K3PO4 (3.9 g, 18.4 mmol) and Pd(t-Bu3P)2
(715 mg, 1.4 mmol) were added to the solution, and argon was
bubbled through the mixture for an additional 10 min. The
reaction mixture was heated to 140 °C in a preheated oil bath
for 24 h. After cooling to room temperature, the reaction
mixture was filtered, dichloromethane (150 mL) was added,
and the mixture was extracted twice with 2 M NaOH (2 ×
150 mL). The combined aqueous layers were acidified to pH 2
with 2 M hydrochloric acid and extracted with dichlo-
romethane (3 × 150 mL). The combined organic layers were
dried over MgSO4 and filtered. After removal of the solvents
under reduced pressure, the crude product 11 (1.1 g, 44%) was
used without further purification for the next reaction step.
1H NMR (DMSO-d6): δ 12.75 (s, 1H), 8.53 (d, J ) 7.6 Hz, 1H),
7.50 (d, 1H, J ) 7.9 Hz, 1H), 7.21 (s, 1H), 7.02 (s, 1H), 6.90 (d,
J ) 7.7 Hz, 1H), 2.50 (s, 3H). MS (ES) m/z: 173.9 (M - H).
1
step. H NMR (CDCl3): δ 7.62 (d, J ) 8.2 Hz, 1H), 7.38 (d, J
) 7.9 Hz, 1H), 7.18 (dd, J ) 8.0 Hz, J ) 7.0 Hz, 1H), 7.05 (dd,
J ) 8.0 Hz, J ) 7.0 Hz, 1 H), 2.48 (s, 3H). MS (ES) m/z: 176.1
(MH+).
(ii) N-(1-Isopropylpiperidin-4-yl)-3-methyl-1H-indole-
2-carboxamide (16). To a solution of 3-methyl-1H-indole-2-
carboxylic acid (15) (300 mg 1.72 mmol) in dichloromethane
(10 mL) and NEt3 (1 mL, 6.8 mmol) was added bis(oxo-3-
oxazolidinyl)phosphoryl chloride (BOP-Cl) (435 mg, 1.72 mmol)
at room temperature, and the mixture was stirred for 5 min.
After addition of 5 (487 mg 2.2 mmol), the mixture was stirred
for 16 h. The reaction mixture was diluted with dichlo-
romethane (100 mL) and washed with brine (50 mL). The
organic layer was dried over MgSO4 and filtered. After removal
of the solvents under reduced pressure, the crude product 16
(506 mg, 99%) was used without further purification for the
(ii) 6-Methyl-1H-indole-2-carboxylic Acid (1-Isopropyl-
piperidin-4-yl)amide (12). To
a solution of 6-methyl-
1H-indole-2-carboxylic acid (11) (250 mg, 1.4 mmol) and
N-ethylmorpholine (726 µL, 5.7 mmol) in dichloromethane (8
mL) was added O-((cyano(ethoxycarbonyl)methylene)amino)-
N,N,N′,N′-tetramethyluronium tetrafluoroborate (TOTU) (468
mg, 1.4 mmol), and the mixture was stirred for 30 min at room
temperature. 5 (304 mg, 1.4 mmol) was added, and the reaction
was stirred for 16 h. After removal of the solvent under
reduced pressure, the residue was purified by preparative
reversed-phase HPLC, eluting with a gradient of acetonitrile
in water (+0.01% trifluoroacetic acid). The product 12 was
obtained in 250 mg (59%) yield as its trifluoroacetate salt. 1H
NMR (DMSO-d6): δ 11.46 (s, 1H), 8.53 (d, J ) 7.6 Hz, 1H),
7.48 (d, 1H, J ) 7.9 Hz, 1H), 7.20 (s, 1H), 7.16 (s, 1H), 6.87 (d,
J ) 7.7 Hz, 1H), 6.57 (s, 1H), 4.09 (m, 1H), 3.43 (m, 2H), 3.14
(m, 2H), 2.50 (s, 3H), 2.09 (m, 2H), 2.02 (m, 1H), 1.80 (m, 2H),
1.25 (d, J ) 6.5 Hz, 6H). MS (ES) m/z: 300.2 (MH+).
1
next reaction step. H NMR (CDCl3): δ 7.60 (d, J ) 8.0 Hz,
1H), 7.36 (d, J ) 7.8 Hz, 1H), 7.18 (dd, J ) 8.0 Hz, J ) 7.0 Hz,
1H), 7.02 (dd, J ) 8.1 Hz, J ) 7.3 Hz, 1 H), 4.09 (m, 1H), 2.93
(m, 2H), 2.84 (m, 2H), 2.51 (s, 3H), 2.41 (m, 2H), 2.12 (m, 1H),
1.70 (m, 2H), 1.25 (d, J ) 6.5 Hz, 6H). MS (ES) m/z: 300.4
(MH+).
(iii) 1-[5-(5-Chlorothiophen-2-yl)isoxazol-3-ylmethyl]-
3-methyl-1H-indole-2-carboxylic Acid (1-Isopropylpiper-
idin-4-yl)amide (17). N-(1-Isopropylpiperidin-4-yl)-3-methyl-
1H-indole-2-carboxamide (16) (100 mg, 0.33 mmol) was dis-
solved in DMF (2 mL), and sodium hydride (60% in oil, 15 mg,
0.4 mmol) was added. The solution was stirred at room
temperature for 20 min. Then 9 (107 mg, 0.4 mmol) was added,
and the reaction mixture was heated to 70 °C for 5 h. Then,
after the mixture was cooled to room temperature, a saturated
NaHCO3 solution (10 mL) was added. The reaction mixture
was then extracted with ethyl acetate (3 × 50 mL). The
combined organic layers were dried over MgSO4 and filtered.
After removal of the solvents under reduced pressure, the
product was purified by preparative reversed-phase HPLC,
eluting with a gradient of acetonitrile in water (+0.01%
trifluoroacetic acid). Following lyophilization, the product 17
was obtained in 16 mg (10%) yield. 1H NMR (DMSO-d6): δ
8.56 (d, J ) 7.7 Hz, 1H), 7.66-7.60 (m, 3H), 7.57-7.52 (m,
2H), 7.30 (ddd, J ) 8.0 Hz, J ) 7.0 Hz, J ) 1.0 Hz, 1H), 7.27
(d, J ) 4.1 Hz, 1H), 7.13 (dd, J ) 8.0 Hz, J ) 7.0 Hz, 1 H),
6.59 (s, 1H), 5.64 (s, 2H), 4.11 (m, 1H), 3.43 (m, 2H), 3.13 (m,
2H), 2.38 (s, 3H), 2.07 (m, 2H), 2.02 (m, 1H), 1.82 (m, 2H),
1.26 (d, J ) 6.6 Hz, 6H). MS (ES) m/z: 497.1 (MH+). HRMS
calcd for C26H30ClN4O2S 497.1778; found 497.1776.
Preparation of 1-[5-(5-Chlorothiophen-2-yl)isoxazol-
3-ylmethyl]-3-cyano-7-methyl-1H-indole-2- carboxylic Acid
(1-Isopropylpiperidin-4-yl)amide (21) (Scheme 2). (i)
7-Methyl-1H-indole-2-carboxylic Acid Methyl Ester.
7-Methyl-1H-indole-2-carboxylic acid (18) (1.0 g, 57 mmol) was
dissolved in methanolic hydrochloric acid (8 M, 20 mL), and
the mixture was stirred at room temperature for 16 h. After
removal of the solvent under reduced pressure, residual
volatiles were removed by codistillation twice with toluene (2
× 10 mL). The remaining slightly yellow solid (1.1 g, 100%)
was subjected to the subsequent reaction without further
purification. MS (ES) m/z: 190.4 (MH+).
(iii) 1-[5-(5-Chlorothiophen-2-yl)isoxazol-3-ylmethyl]-
6-methyl-1H-indole-2-carboxylic Acid (1-Isopropylpi-
peridin-4-yl)amide (13). 6-Methyl-1H-indole-2-carboxylic
acid (1-isopropylpiperidin-4-yl)amide (12) (222 mg; 0.8 mmol)
was dissolved in DMF (2 mL), and sodium hydride (60% in
oil, 31 mg, 0.8 mmol) was added. The solution was stirred at
room temperature for 20 min, then 9 (214 mg, 0.8 mmol) was
added and the reaction mixture was heated to 80 °C for 16 h.
After the mixture was cooled to room temperature, saturated
NaHCO3 solution (10 mL) was added, and the reaction mixture
was extracted with ethyl acetate (3 × 50 mL). The combined
organic layers were dried over MgSO4 and filtered. After
removal of the solvents under reduced pressure, the product
was purified by preparative reversed-phase HPLC, eluting
with a gradient of acetonitrile in water (+0.01% trifluoroacetic
acid). Following lyophilization, the product 13 was obtained
in 30 mg (15%) yield as its trifluoroacetate salt. 1H NMR
(DMSO-d6): δ 8.58 (d, J ) 7.6 Hz, 1H), 7.56 (m, 3H), 7.39 (s,
1H), 7.27 (d, 4.2 Hz, 1H), 7.00 (s, 1H), 6.98 (d, J ) 7.7 Hz,
1H), 5.86 (s, 2H), 4.05 (m, 1H), 3.43 (m, 2H), 3.14 (m, 2H),
2.50 (s, 3H), 2.09 (m, 2H), 2.02 (m, 1H), 1.80 (m, 2H), 1.25 (d,
J ) 6.5 Hz, 6H). MS (ES) m/z: 497.2 (MH+). HRMS calcd for
C26H30ClN4O2S 497.1773; found 497.1769.
1-[5-(5-Chlorothiophen-2-yl)isoxazol-3-ylmethyl]-3-
methyl-1H-indole-2-carboxylic Acid (1-Isopropylpiperi-
din-4-yl)amide (17) (Scheme 2). (i) 3-Methyl-1H-indole-
2-carboxylic Acid (15). 2-Iodophenylamine (14) (0.9 g, 4
mmol), 2-oxobutanoic acid (1.2 g, 12 mmol), 1,4-diazabicyclo-
[2.2.2]octane (1.3 g, 12 mmol), and MgSO4 (0.85 g, 7.1 mmol)
were suspended in DMF (12 mL). The round-bottom flask was
(ii) 3-Iodo-7-methyl-1H-indole-2-carboxylic Acid Meth-
yl Ester. 7-Methyl-1H-indole-2-carboxylic acid methyl ester
(1.08 g, 5.7 mmol) was dissolved in dichloromethane (15 mL),

Two Binding Modes for Factor Xa Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14 4523
and N-iodosuccinimide (1.41 g, 6.3 mmol) was added. The
solution was stirred for 16 h at room temperature, then washed
with aqueous sodium thiosulfate solution. The organic phase
was dried (Na2SO4) and filtered through a silica gel cartridge,
washing with dichloromethane. The solvent was removed
under reduced pressure to give the product in 1.69 g (94%)
codistillation with toluene (2 × 10 mL). The remaining slightly
yellow solid (2.3 g, 100%) was subjected to the subsequent
reaction without further purification. MS (ES) m/z: 176.6
(MH+).
(ii) 1-[5-(5-Chlorothiophen-2-yl)isoxazol-3-ylmethyl]-
1H-indole-2-carboxylic Acid Methyl Ester. To a solution
of 1H-indole-2-carboxylic acid methyl ester (244.2 mg, 1.4
mmol) in DMF (2 mL) was added sodium hydride (60% in oil,
52.2 mg, 1.3 mmol) at room temperature. After the mixture
was stirred for 30 min, 9 (500 mg, 1.8 mmol) was added and
the mixture was heated for 1 h at 80 °C. After cooling of the
reaction mixture to room temperature and addition of water
(5 mL), the mixture was filtered through a pad of Celite,
washing with ethyl acetate. After concentration under reduced
pressure the product (288 mg, 55%) was directly subjected to
the subsequent saponification reaction without further puri-
fication. MS (ES) m/z: 373.6 (MH+).
(iii) 1-[5-(5-Chlorothiophen-2-yl)isoxazol-3-ylmethyl]-
1H-indole-2-carboxylic Acid. To a solution of 1-[5-(5-chlo-
rothiophen-2-yl)isoxazol-3-ylmethyl]-1H-indole-2-carboxylic acid
methyl ester (288 mg, 0.77 mmol) in THF (10 mL) were added
water (3 mL) and lithium hydroxide monohydrate (57.0 mg,
1.4 mmol). After the mixture was stirred for 2 h at 60 °C, the
reaction mixture was cooled to room temperature. The mixture
was acidified to pH 3 with half-concentrated hydrochloric acid.
The resulting precipitate was collected by filtration and
washed with water (3 mL). The product (253 mg, 92%) was
obtained as a white solid, which was dried under reduced
pressure. MS (ES) m/z: 359.4 (MH+).
1
yield. H NMR (DMSO-d6): δ 12.09 (s, 1H), 7.21 (d, J ) 8.0
Hz, 1H), 7.12 (d, 1H, J ) 7.3 Hz, 1H), 7.21 (m, 1H), 3.92 (s,
3H), 2.54 (s, 3H). MS (ES) m/z: 316 (MH+).
(iii) 3-Cyano-7-methyl-1H-indole-2-carboxylic Acid
Methyl Ester (19). 3-Iodo-7-methyl-1H-indole-2-carboxylic
acid methyl ester (1.69 g, 5.4 mmol), CuCN (0.96 g, 10.8 mmol),
and tetraethylammonium cyanide (0.42 g, 2.7 mmol) were
dissolved in DMF (15 mL) and THF (15 mL), and the mixture
was purged with argon for 15 min. 1,1′-Bis(diphenylphos-
phino)ferrocene (0.45 g, 0.8 mmol) and Pd2(dba)3 (0.246 g, 0.3
mmol) were added. The reaction mixture was heated to 120
°C for 5 h, then the solvent was removed under reduced
pressure. The residue was dissolved in ethyl acetate and
washed with a saturated aqueous NaHCO3 solution. The
organic phase was dried (Na2SO4), filtered, and evaporated.
The product was purified first by silica gel chromatography,
eluting with a gradient of ethyl acetate in heptane, followed
by preparative reversed-phase HPLC, eluting with a gradient
of acetonitrile in water (+0.01% trifluoroacetic acid). The
product 19 was obtained in 0.15 g (13%) yield. 1H NMR
(DMSO-d6): δ 12.97 (s, 1H), 7.82 (d, J ) 8.3 Hz, 1H), 7.46 (m,
1H), 7.25 (d, J ) 8.4 Hz, 1H), 3.92 (s, 3H), 2.54 (s, 3H). MS
(ES) m/z: 215 (MH+).
(iv) 1-((5-(5-Chlorothiophen-2-yl)isoxazol-3-yl)methyl)-
3-cyano-7-methyl-1H-indole-2-carboxylic Acid (20). 3-Cy-
ano-7-methyl-1H-indole-2-carboxylic acid methyl ester (19)
(150 mg, 0.7 mmol) was dissolved in DMF (2 mL), and sodium
hydride (60% in oil, 31 mg, 0.8 mmol) was added. The solution
was stirred at room temperature for 20 min, then cooled to
-70 °C. 9 (214 mg, 0.8 mmol) was added, and the reaction
mixture was allowed to warm to room temperature. The
reaction mixture was stirred for 3 h. Then 2 M aqueous NaOH
solution (0.5 mL) was added, and the reaction mixture was
stirred for 16 h. The product was purified by preparative
reversed-phase HPLC, eluting with a gradient of acetonitrile
in water (+0.01% trifluoroacetic acid). Following lyophilization
(iv) 1-[5-(5-Chlorothiophen-2-yl)isoxazol-3-ylmethyl]-
1H-indole-2-carboxylic Acid (1-Isopropylpiperidin-4-yl)-
amide (1). To a solution of 1-[5-(5-chlorothiophen-2-yl)isoxazol-
3-ylmethyl]-1H-indole-2-carboxylic acid (117 mg, 0.33 mmol)
in dichloromethane (1 mL) were added NEt3 (0.17 mL, 1.2
mmol) and bis(oxo-3-oxazolidinyl)phosphoryl chloride (BOP-
Cl) (79 mg, 033 mmol) at room temperature, and the mixture
was stirred for 30 min. After addition of 5 (81 mg, 0.37 mmol),
the mixture was stirred for 16 h. After removal of the solvent
under reduced pressure, the residue was purified by prepara-
tive HPLC (C18 reverse phase column, eluting with a gradient
of acetonitrile in water (+0.01% trifluoroacetic acid). The
fractions containing the product were evaporated and lyo-
philized to yield the product 1 (93 mg, 60%) as a white solid.
The product was obtained as its trifluoroacetate salt. 1H NMR
(DMSO-d6): δ 8.65 (d, J ) 7.5 Hz, 1H), 7.69 (d, J ) 8.0 Hz,
1H), 7.60 (d, J ) 8.0 Hz, 1H), 7.54 (d, J ) 4.0 Hz, 1H), 7.30
(ddd, J ) 8.0 Hz, J ) 7.0 Hz, J ) 1.0 Hz, 1H), 7.26 (d, J ) 4.0
Hz, 1H), 7.23 (s, 1H), 7.15 (ddd, J ) 8.0 Hz, J ) 7.0 Hz, J )
1.0 Hz, 1H), 6.58 (s, 1H), 5.91 (s, 2H), 4.06 (m, 1H), 3.43 (m,
2H), 3.10 (m, 2H), 2.07 (m, 2H), 2.02 (m, 1H), 1.80 (m, 2H),
1.26 (d, J ) 6.5 Hz, 6H). MS (ES) m/z: 483.3 (MH+). HRMS
calcd for C25H28ClN4O2S 483.1621; found 483.1616.
1-[5-(5-Chlorothiophen-2-yl)isoxazol-3-ylmethyl]-3-flu-
oro-1H-indole-2-carboxylic Acid (1-Isopropylpiperidin-
4-yl)amide (22). To a solution of 1 (40 mg, 0.08 mmol) in
dichloromethane (1 mL) was added N-fluoropyridinium triflate
(22 mg, 0.09 mmol), and the mixture was stirred at room
temperature for 4 days. The reaction mixture was directly
purified by preparative reversed-phase HPLC, eluting with a
gradient of acetonitrile in water (+0.01% trifluoroacetic acid)
After lyophilization the product 22 (22 mg, 50%) was obtained
as white solid. 1H NMR (DMSO-d6): δ 8.59 (d, J ) 7.8 Hz,
1H), 7.75 (dd, J ) 8.1 Hz, J ) 7.0 Hz, 1H), 7.54 (d, J ) 4.0 Hz,
1H), 7.41 (ddd, J ) 8.0 Hz, J ) 7.0 Hz, J ) 1.0 Hz, 1H), 7.28
(m, 3H), 6.60 (s, 1H), 5.76 (s, 2H), 4.08 (m, 1H), 3.43 (m, 2H),
3.10 (m, 2H), 2.07(m, 2H), 2.02 (m, 1H), 1.82 (m, 2H), 1.26 (d,
J ) 6.5 Hz, 6H). MS (ES) m/z: 501.2 (MH+). HRMS calcd for
C25H27ClFN4O2S 501.1527; found 501.1522.
1
the product 20 was obtained in 113 mg (39%) yield. H NMR
(DMSO-d6): δ 14.3 (s, 1H), 7.56 (m, 2H), 7.28 (m, 3H), 6.79 (s,
1H), 6.21 (s, 2H), 2.66 (s, 3H). MS (ES) m/z: 398 (MH+).
(v) 1-[5-(5-Chlorothiophen-2-yl)isoxazol-3-ylmethyl]-3-
cyano-7-methyl-1H-indole-2-carboxylic Acid (1-Isopro-
pylpiperidin-4-yl)amide (21). To a solution of 1-((5-(5-
chlorothiophen-2-yl)isoxazol-3-yl)methyl)-3-cyano-7-methyl-
1H-indole-2-carboxylic acid (20) (30 mg, 75 µmol) and N-ethyl-
morpholine (38 µL) in dichloromethane (8 mL) was added
O-((cyano(ethoxycarbonyl)methylene)amino)-N,N,N′,N′-tet-
ramethyluronium tetrafluoroborate (TOTU) (25 mg, 75 µmol),
and the mixture was stirred for 30 min at room temperature.
Then 5 (32 mg, 0.15 mmol) was added, and the reaction
mixture was stirred for 16 h. After removal of the solvent
under reduced pressure, the residue was purified by prepara-
tive reversed-phase HPLC, eluting with a gradient of aceto-
nitrile in water (+0.01% trifluoroacetic acid). The product 21
was obtained in 22 mg (46%) yield as its trifluoroacetate salt.
1H NMR (DMSO-d6): δ 9.37 (d, J ) 7.7 Hz, 1H), 7.56 (m, 2H),
7.29 (d, J ) 4.0 Hz, 1H), 7.26 (m, 2 H), 7.20 (d, J ) 8.4 Hz,
1H), 6.67 (s, 1H), 5.94 (s, 2H), 4.11 (m, 1H), 3.42 (m, 2H), 3.12
(m, 2H), 2.70 (s, 3H), 2.07 (m, 2H), 2.02 (m, 1H), 1.84 (m, 2H),
1.26 (d, J ) 6.5 Hz, 6H). MS (ES) m/z: 522.2 (MH+). HRMS
calcd for C27H29ClN5O2S 522.1730; found 522.1725.
Representative Procedure: Preparation of 1-[5-(5-
Chlorothiophen-2-yl)isoxazol-3-ylmethyl]-1H-indole-2-
carboxylic Acid (1-Isopropylpiperidin-4-yl)amide (1). (i)
1H-Indole-2-carboxylic Acid Methyl Ester. 1H-Indole-2-
carboxylic acid (2.0 g, 124 mmol) was dissolved in methanolic
hydrochloric acid (8 M, 30 mL), and the mixture was stirred
at room temperature for 16 h. After removal of the solvent
under reduced pressure, residual volatiles were removed by
3-Chloro-1-[5-(5-chlorothiophen-2-yl)isoxazol-3-yl-
methyl]-1H-indole-2-carboxylic Acid (1-iSopropylpiperi-
din-4-yl)amide (23). To a solution of 1 (40 mg, 0.08 mmol) in
dichloromethane (1 mL) was added N-chlorosuccinimide (17
mg, 0.13 mmol), and the mixture was stirred at room temper-
ature for 16 h. The reaction mixture was directly purified by

4524 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14
Nazare´ et al.
preparative reversed-phase HPLC, eluting with a gradient of
acetonitrile in water (+0.01% trifluoroacetic acid). After lyo-
philization the product 23 (15 mg, 36%) was obtained as a
internal database and were used for flexible docking studies.
After analysis of protein-ligand interactions using the pro-
gram GRID,23 molecules were manually oriented and mini-
mized within the binding site or automatically docked into the
active site using the program QXP24 employing a modified
version of the AMBER force field.25 Selected protein side chains
were treated as flexible for minimization and Monte Carlo
searching after comparative analysis of fXa X-ray structures.
Selected conformers are subsequently optimized using the
MMFF94s force field26 implemented in Sybyl.27 Protein/ligand
complexes were minimized using quasi-Newton-Raphson
(BFGS) or conjugate gradient (CG) procedures with all protein
atoms as rigid. The program MOLCAD28 was used to visualize
properties such as lipophilicity29,30 and cavity depth on solvent-
accessible protein surfaces.31
1
white solid. H NMR (DMSO-d6): δ 8.81 (d, J ) 7.8 Hz, 1H),
7.75 (d, J ) 8.5 Hz, 1H), 7.60 (d, J ) 7.5 Hz, 1H), 7.57 (d, J )
7.7 Hz, 1 H), 7.55 (d, J ) 4.0 Hz, 1H), 7.41 (ddd, J ) 8.0 Hz,
J ) 7.0 Hz, J ) 1.0 Hz, 1H), 7.28 (d, J ) 4.1 Hz, 1H), 6.62 (s,
1H), 5.70 (s, 2H), 4.06 (m, 1H), 3.43 (m, 2H), 3.10 (m, 2H),
2.07 (m, 2H), 2.02 (m, 1H), 1.82 (m, 2H), 1.26 (d, J ) 6.6 Hz,
6H). MS (ES) m/z: 517.2 (MH+). HRMS calcd for C25H27-
Cl2N4O2S 517.1231; found 517.1226.
3-Bromo-1-[5-(5-chlorothiophen-2-yl)isoxazol-3-yl-
methyl]-1H-indole-2-carboxylic Acid (1-Isopropylpiperi-
din-4-yl)amide (24). To a solution of 1 (40 mg, 0.08 mmol) in
dichloromethane (1 mL) was added N-bromosuccinimide (176
mg, 0.13 mmol), and the mixture was stirred at room temper-
ature for 16 h. The reaction mixture was directly purified by
preparative reversed-phase HPLC, eluting with a gradient of
acetonitrile in water (+0.01% trifluoroacetic acid). After lyo-
philization the product 24 (18 mg, 40%) was obtained as a
Acknowledgment. We thank A. Sihorsch, M. Ka¨m-
merer-Dienst, L. Bayer, N. Laub, D. Timme, F. Gerst-
mann, A. Liesum, V. Brachvogel, P. Loenze, and S.
Engel for their excellent technical assistance.
1
white solid. H NMR (DMSO-d6): δ 8.82 (d, J ) 7.8 Hz, 1H),
7.72 (d, J ) 8.5 Hz, 1H), 7.57-7.52 (m, 3H), 7.41 (ddd, J ) 8.0
Hz, J ) 7.0 Hz, J ) 1.0 Hz, 1H), 7.27 (d, J ) 4.1 Hz, 1H), 6.52
(s, 1H), 5.99 (s, 2H), 4.06 (m, 1H), 3.43 (m, 2H), 3.10 (m, 2H),
2.07 (m, 2H), 2.02 (m, 1H), 1.82 (m, 2H), 1.26 (d, J ) 6.6 Hz,
6H). MS (ES) m/z: 561.1 (MH+). HRMS calcd for C25H27-
BrClN4O2S 561.0726; found 561.0721.
5.2. X-ray Structure Analysis. Crystallization. Purified
human fXa was purchased from Enzyme Research Lab (South
Bend, IN). The Gla-domain was removed, and the Gla-less
factor Xa was crystallized in hanging drops as described
earlier.18
Data Collection and Processing. Crystals were soaked
in 5 µL of reservoir solution containing ∼20 mM saturated
inhibitor (depending on the solubility) for 24-72 h. Data were
collected at cryotemperatures. The crystals were picked up
with a fiber loop, soaked for a few seconds in a solution
containing 20% glycerol and 5-20 mM inhibitor in reservoir
solution, and flash-frozen in a stream of gaseous nitrogen at
100 K.
The X-ray intensity data were collected on a 130 mm Mar
CCD detector mounted on an Elliot GX-13 “big wheel” rotating
anode generator (Nonius, The Netherlands), operating at 40
kV and 55 mA, or on an Mar300 imaging plate (X-ray research,
Germany) mounted on an FR591 rotating anode (Nonius, The
Netherlands), operating at 40 kV and 80 mA. Data processing
and scaling were carried out using the XDS package.20 Data
collection and refinement statistics for selected crystal struc-
tures are presented in Table 4.
Supporting Information Available: Experimental data
for the compounds 25-46 and purity data of all compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org. The coordinates of the fXa complexes were
deposited in the Brookhaven Protein Data Bank (accession
codes 2BOH, 2BQ6, 2BQ7, 2BQW, 1LQD).
References
(1) Vlasuk, G. P. Direct Factor Xa inhibitors. In New Therapeutic
Agents in Thrombosis and Thrombolysis; Sasahara, A. A.,
Loscalzo, J., Eds.; Marcel Dekker: New York, 1997; pp 261-
283.
(2) (a) Hirsh, J. Heparin. N. Engl. J. Med. 1991, 324, 1565-1574.
(b) Freedman, M. D. Warfarin and other “anti”-vitamin K
anticoagulants. Pharmacodynamics and clinical use. Am J. Ther.
1996, 3, 771-783. (c) Raj, G.; Kumar, R.; McKinney, W. P. Long-
term oral anticoagulant therapy: update on indications, thera-
peutic ranges, and monitoring. Am. J. Med. Sci. 1994, 307, 128-
32.
(3) (a) Coleman, R. W., Marder, V. J., Salzman, E. W., Eds.
Hemostasis and Thrombosis: Basic Principles and Clinical
Practice; J. B. Lippincott: Philadelphia, PA, 1994. (b) Davie, E.
W.; Fujikawa, K.; Kisiel, W. The coagulation cascade: initiation,
maintenance, and regulation. Biochemistry 1991, 30, 10363-
10370. (c) Leadley, R. J. Coagulation factor Xa inhibition:
biological background and rationale. Curr. Top. Med. Chem.
2001, 1, 151-159. (d) Stassen, J. M.; Arnout, J.; Deckmyn, H.
The hemostatic system. Curr. Med. Chem. 2004, 11, 2245-2260.
(4) Kaiser, B. Thrombin and factor Xa inhibitors. Drugs Future
1998, 23, 423-436.
(5) (a) Pinto, D. J. P.; Orwat, M. J.; Wang, S.; Fevig, J. M.; Quan,
M. L.; Amparo, E.; Cacciola, J.; Rossi, K. A.; Alexander, R. S.;
Smallwood, A. M.; Luettgen, J. M.; Liang, L.; Aungst, B. J.;
Wright, M. R.; Knabb, R. M.; Wong, P. C.; Wexler, R. R.; Lam,
P. Y. S. Discovery of 1-[3-(Aminomethyl)phenyl]-N-[3-fluoro-2′-
(methylsulfonyl)-[1,1′-biphenyl]-4-yl]-3-(trifluoromethyl)-1H-pyra-
zole-5-carboxamide (DPC423), a highly potent, selective, and
orally bioavailable inhibitor of blood coagulation factor Xa. J.
Med. Chem. 2001, 44, 566-578. (b) Pauls, H. W.; Ewing, W. R.
The design of competitive, small-molecule inhibitors of coagula-
tion Factor Xa. Curr. Top. Med. Chem. 2001, 1, 83-100.
(6) (a) Song, Y.; Clizbe, L.; Bhakta, C.; Teng, W.; Wong, P.; Huang,
B.; Tran, K.; Sinha, U.; Park, G.; Reed, A.; Scarborough, R. M.;
Zhu, B.-Y. Design and synthesis of factor Xa inhibitors and their
prodrugs. Bioorg. Med. Chem. Lett. 2003, 13, 297-300. (b)
Gustafsson, D.; Elg, M. The pharmacodynamics and pharmaco-
kinetics of the oral direct thrombin inhibitor Ximelagatran and
its active metabolite Melagatran: a mini-review. Thromb. Res.
2003, 109 (Suppl. 1), S9-S15. (c) Koshio, H.; Hirayama, F.;
Ishihara, T.; Kaizawa, H.; Shigenaga, T.; Taniuchi, Y.; Sato, K.;
Moritani, Y.; Iwatsuki, Y.; Uemura, T.; Kaku, S.; Kawasaki, T.;
Matsumoto, Y.; Sakamoto, S.; Tsukamoto, S.-I. Orally active
factor Xa inhibitor: synthesis and biological activity of masked
amidines as prodrugs of novel 1,4-diazepane derivatives. Bioorg.
Med. Chem. 2004, 12, 5415-5426.
Structure Solution and Crystallographic Refinement.
The structures were solved by molecular replacement. The
search models were made from the coordinates of refined
structures of fXa complexes solved previously with the same
crystal packing as the current complexes. The bound inhibitors
were omitted from the search models. Energy-restrained least-
squares refinement was carried out using X-PLOR.21 This
refinement was started with rigid body refinement to adjust
for small differences in cell dimensions, followed by energy
minimization and individual temperature factor refinement.
At this stage the 2Fo - Fc and Fo - Fc maps were inspected
and the inhibitors were fitted. Solvent molecules were included
if they were on sites of difference electron density with values
above 3.5σ and if they were within 3.5 Å of the protein
molecule or a water molecule. After two to three additional
rounds of manual inspection, rebuilding, and refinement, final
models were obtained with R-factors between 18.1% and 21.8%
and free R-factors between 25.3% and 29.8% plus good
geometry. The EGF-1 domain is not visible in the electron
density maps probably because of disorder, and the rather high
free R-factors might relate to this disordered EGF-1 domain.
The statistics of the crystallographic refinement are listed in
Table 4.
(7) (a) Quan, M. L.; Lam, P. Y. S.; Han, Q.; Pinto, D. J. P.; He, M.
Y.; Li, R.; Ellis, C. D.; Clark, C. G.; Teleha, C. A.; Sun, J.-H.;
Alexander, R. S.; Bai, S.; Luettgen, J. M.; Knabb, R. M.; Wong,
P. C.; Wexler, R. R. Discovery of 1-(3′-aminobenzisoxazol-5′-yl)-
3-trifluoromethyl-N-[2-fluoro-4-[(2′-dimethylaminomethyl)imida-
zol-1-yl]phenyl]-1H-pyrazole-5-carboxyamide hydrochloride (raza-
3. Computational procedures. Docking Studies. Factor
Xa crystal structures were taken from RSCB22 and from our

Two Binding Modes for Factor Xa Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 14 4525
xaban), a highly potent, selective, and orally bioavailable factor
Xa inhibitor. J. Med. Chem. 2005, 48, 1729-1744. (b) Lam, P.
Y. S.; Clark, C. G.; Li, R.; Pinto, D. J. P.; Orwat, M. J.; Galemmo,
R. A.; Fevig, J. M.; Teleha, C. A.; Alexander, R. S.; Smallwood,
A. M.; Rossi, K. A.; Wright, M. R.; Bai, S. A.; He, K.; Luettgen,
J. M.; Wong, P. C.; Knabb, R. M.; Wexler, R. R. Structure-based
design of novel guanidine/benzamidine mimics: potent and
orally bioavailable factor Xa inhibitors as novel anticoagulants.
J. Med. Chem. 2003, 46, 4405-4418.
(17) (a) Chen, C.-Y.; Lieberman, D. R.; Larsen, R. D.; Verhoeven, T.
R.; Reider, P. J. Syntheses of indoles via a palladium-catalyzed
annulation between iodoanilines and ketones. J. Org. Chem.
1997, 62, 2676-2677. (b) Nazare´, M.; Schneider, C.; Linden-
schmidt, A.; Will, D. W. A flexible, palladium-catalyzed indole
and azaindole synthesis by direct annulation of chloroanilines
and chloroaminopyridines with ketones. Angew. Chem, Int. Ed.
2004, 43, 4526-4528.
(18) Matter, H.; Defossa, E.; Heinelt, U.; Blohm, P.-M.; Schneider,
D.; Mueller, A.; Herok, S.; Schreuder, H.; Liesum, A.; Brachvogel,
V.; Loenze, P.; Walser, A.; Al-Obeidi, F.; Wildgoose, P. Design
and Quantitative structure-activity relationship of 3-amidi-
nobenzyl-1H-indole-2-carboxamides as potent, nonchiral, and
selective inhibitors of blood coagulation factor Xa. J. Med. Chem.
2002, 45, 2749-2769.
(19) 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.
(20) Kabsch, W. Evaluation of single-crystal X-ray diffraction data
from a position sensitive detector. J. Appl. Crystallogr. 1988,
21, 916-924.
(21) Bru¨nger, A. T. X-PLOR: A System for X-ray Crystallography
and NMR, version 3.1; Yale University Press: New Haven, CT,
1992.
(22) PDB files from RSCB Protein Database: http://www.rcsb.org/
pdb. (a) Bernstein, F. C.; Koetzle, T. F.; Williams, G. J. B.; Meyer,
E. F.; Brice, M. D.; Rodgers, J. R.; Kennard, O.; Shimanouchi,
T.; Tasumi, M. The protein data bank: a computer-based
archival file for macromolecular structures. J. Mol. Biol. 1977,
112, 535-542. (b) Berman, J. W.; Zukang, F.; Gilliland, G.; Bhat,
T. N.; Weissig, H.; Shindyalov, I. N.; Bourne, P. E. The protein
data bank. Nucleic Acids Res. 2000, 28, 235-242.
(23) Goodford, P. J. A computational procedure for determining
energetically favorable binding sites on biologically important
macromolecules. J. Med. Chem. 1985, 28, 849-857.
(24) McMartin, C.; Bohacek, R. S. QXP: powerful, rapid computer
algorithms for structure-based drug design. J. Comput.-Aided
Mol. Des. 1997, 11, 333-344.
(25) Weiner, S. J.; Kollman, P. A.; Case, D. A.; Singh, U. C.; Ghio,
C.; Algona, G.; Profeta, S.; Weiner, P. A new force field for
molecular mechanical simulation of nucleic acids and proteins.
J. Am. Chem. Soc. 1984, 106, 765-784.
(26) (a) Halgren, T. Maximally diagonal force constants in dependent
angle-bending coordinates. II. Implications for the design of
empirical force fields. J. Am. Chem. Soc. 1990, 112, 4710-4723.
(b) Halgren, T. MMFF VI. MMFF94s option for energy minimi-
zation studies. J. Comput. Chem. 1999, 20, 720-729.
(27) SYBYL Molecular Modelling Package, version 6.9; Tripos: St.
Louis, MO, 2003.
(28) (a) Heiden, W.; Goetze, T.; Brickmann, J. Fast generation of
molecular surfaces from 3D data fields with enhanced “marching
cube” algorithm. J. Comput. Chem. 1993, 14, 246-250. (b)
Brickmann, J.; Bertling, H.; Bussian, B. M.; Goetze, T.; Kno-
blauch, M.; Waldherr-Teschner, M. MOLCADsinteractive mo-
lecular computer graphics on high-performance computers.
Tagungsber.sVortragstag., Ges. Dtsch. Chem., Fachgruppe Chem.-
Inf. 1987, 93-111.
(29) Heiden, W.; Moeckel, G.; Brickmann, J. A new approach to the
display of local lipophilicity/hydrophilicity mapped on molecular
surfaces. J. Comput.-Aided Mol. Des. 1993, 7, 503-514.
(30) (a) Ghose, A.; Crippen, G. Atomic physicochemical parameters
for three-dimensional structure-directed quantitative structure-
activity relationships. 1. Partition coefficients as a measure of
hydrophobicity. J. Comput. Chem. 1986, 7, 565-577. (b) Viswa-
nadhan, V. N.; Ghose, A. K.; Revankar, G. R.; Robins, R. K.
Atomic physicochemical parameters for three-dimensional struc-
ture-directed quantitative structure-activity relationships. 4.
Additional parameters for hydrophobic and dispersive interac-
tions and their application for an automated superposition of
certain naturally occurring nucleoside antibiotics. J. Chem. Inf.
Comput. Sci. 1989, 29, 163-172.
(8) (a) Adler, M.; Kochanny, M. J., Ye, B.; Rumnenik, G.; Light, D.
R.; Biancalana, S.; Whitlow, M. Crystal structures of two potent
nonamidine inhibitors bound to factor Xa. Biochemistry 2002,
41, 15514-15523. (b) Maignan, S.; Guilloteau, J.-P.; Choi-
Sledeski, Y. M.; Becker, M. R.; Ewing, W. R.; Pauls, H. W.;
Spada, A. P.; Mikol, V. Molecular structures of human factor
Xa complexed with ketopiperazine inhibitors: Preference for a
neutral group in the S1 pocket. J. Med. Chem. 2003, 46, 685-
690.
(9) (a) Nazare´, M.; Essrich, M.; Will, D. W.; Matter, H.; Ritter, K.;
Urmann, M.; Bauer, A.; Schreuder, H.; Czech, J.; Lorenz, M.;
Laux, V.; Wehner, V. Novel factor Xa inhibitors based on a
2-carboxyindole scaffold: SAR of P4 substituents in combination
with a neutral P1 ligand. Bioorg. Med. Chem. Lett. 2004, 14,
4197-4201. (b) Nazare´, M.; Essrich, M.; Will, D. W.; Matter,
H.; Ritter, K.; Urmann, M.; Bauer, A.; Schreuder, H.; Dudda,
A.; Czech, J.; Lorenz, M.; Laux, V.; Wehner, V. Factor Xa
inhibitors based on a 2-carboxyindole scaffold: SAR of neutral
P1 substituents. Bioorg. Med. Chem. Lett. 2004, 14, 4191-4195.
(10) (a) Fevig, J. M.; Wexler, R. R. Anticoagulants: thrombin and
factor Xa inhibitors. Annu. Rep. Med. Chem. 1999, 34, 81-100.
(b) Zhu, B.-Y.; Scarborough, R. M. Recent advances in inhibitors
of factor Xa in the prothrombinase complex. Curr. Opin. Car-
diovasc., Pulm. Renal Invest. Drugs 1999, 1, 63-87. (c) Ewing,
W. R.; Pauls, H. W.; Spada, A. P. Progress in the design of
inhibitors of coagulation factor Xa. Drugs Future 1999, 24, 771-
787. (d) Zhu, B.-Y.; Scarborough, R. M. Factor Xa inhibitors:
recent advances in anticoagulant agents. Annu. Rep. Med. Chem.
2000, 35, 83-102. (e) Sanderson, P. E. J. Anticoagulants:
Inhibitors of thrombin and factor Xa. Annu. Rep. Med. Chem.
2001, 36, 79-88. (f) Walenga, J. M.; Jeske, W. P.; Hoppensteadt,
D.; Fareed, J. Factor Xa inhibitors: Today and beyond. Curr.
Opin. Invest. Drugs 2003, 4, 272-281. (g) Ries, U. J.; Priepke,
H. W. M. Factor Xa inhibitorssa review of the recent patent
literature. IDrugs 2000, 3, 1509-1524.
(11) Choi-Sledeski, Y. M.; Kearney, R.; Poli, G.; Pauls, H.; Gardner,
C.; Gong, Y.; Becker, M.; Davis, R.; Spada, A.; Liang, G.; Chu,
V.; Brown, K.; Collussi, D.; Leadley, R., Jr.; Rebello, S.; Moxey,
P.; Morgan, S.; Bentley, R.; Kasiewski, C.; Maignan, S.; Guillo-
teau, J.-P.; Mikol, V. Discovery of an orally efficacious inhibitor
of coagulation factor Xa which incorporates a neutral P1 ligand.
J. Med. Chem. 2003, 46, 681-684.
(12) Jia, Z. J.; Wu, Y.; Huang, W.; Zhang, P.; Clizbe, L. A.; Goldman,
E. A.; Sinha, U.; Arfsten, A. E.; Edwards, S. T.; Alphonso, M.;
Hutchaleelaha, A.; Scarborough, R. M.; Zhu, B.-Y. 1-(2-Naph-
thyl)-1H-pyrazole-5-carboxylamides as potent factor Xa inhibi-
tors. Part 2: A survey of P4 motifs. Bioorg. Med. Chem. Lett.
2004, 14, 1221-1227.
(13) Nazare´, M.; Matter, H.; Klingler, O.; Al-Obeidi, F.; Schreuder,
H.; Zoller, G.; Czech, J.; Lorenz, M.; Dudda, A.; Peyman, A.;
Nestler, H. P.; Urmann, M.; Bauer, A.; Laux, V.; Wehner, V.;
Will, D. W. Novel factor Xa inhibitors based on a benzoic acid
scaffold and incorporating a neutral P1 ligand. Bioorg. Med.
Chem. Lett. 2004, 14, 2801-2805.
(14) Matter, H.; Will, D. W.; Nazare´, M.; Schreuder, H.; Laux, V.;
Wehner, V. Structural requirements for factor Xa inhibition by
3-oxybenzamides with neutral P1 substituents: Combining
X-ray crystallography, 3D-QSAR and tailored scoring functions.
J. Med. Chem. 2005, 48, 3290-3312.
(15) In vitro assays for determination of fXa activity were performed
as described in the following: Ostrem, J. A.; Al-Obeidi, F. A.;
Safar, P.; Safarova, A.; Stringer, S. K.; Patek, M.; Cross, M. T.;
Spoonamore, J.; LoCascio, J. C.; Kasireddy, P.; Thorpe, D.;
Sepetov, N.; Lebl, M.; Wildgoose, P.; Strop, P. Discovery of a
novel, potent, and specific family of factor Xa inhibitors via
combinatorial chemistry. Biochemistry 1998, 37, 1053-1059. The
Ki values were averaged from triplicate determination, and the
standard deviation of the positive control was <15% of the mean.
(16) (a) Joule, J. A. Product class 13: indole and its derivatives. Sci.
Synth. 2001, 10, 361-652. (b) Gribble, G. W. Recent develop-
ments in indole ring synthesissmethodology and applications.
J. Chem. Soc., Perkin Trans. 1 2000, 1045-1075.
(31) (a) Connolly, M. L. Solvent-accessible surfaces of proteins and
nucleic acids. Science 1983, 221, 709-713. (b) Connolly, M. L.
Analytical molecular surface calculation. J. Appl. Crystallogr.
1983, 16, 548-558.
JM0490540