Design of novel aminopyrrolidine factor Xa inhibitors from a screening hit

Article abstract of DOI:10.1016/j.ejmech.2008.12.025

Starting from a hit identified by focused screening, 3-aminopyrrolidine factor Xa inhibitors were designed. The binding mode as determined by X-ray structural analysis as well as the pharmacokinetic behaviour of selected compounds is discussed.

Full text of DOI:10.1016/j.ejmech.2008.12.025

European Journal of Medicinal Chemistry 44 (2009) 2787–2795
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Design of novel aminopyrrolidine factor Xa inhibitors from a screening hit
*
Katrin Groebke Zbinden , Lilli Anselm, David W. Banner, Jo¨rg Benz, Francesca Blasco,
´
Guillaume Decoret, Jacques Himber, Bernd Kuhn, Narendra Panday,
Fabienne Ricklin, Philippe Risch, Daniel Schlatter, Martin Stahl,
Stefan Thomi, Robert Unger, Wolfgang Haap
F. Hoffmann – La Roche Ltd., CH-4070 Basel, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history:
Starting from a hit identified by focused screening, 3-aminopyrrolidine factor Xa inhibitors were
designed. The binding mode as determined by X-ray structural analysis as well as the pharmacokinetic
behaviour of selected compounds is discussed.
Received 7 November 2008
Received in revised form
28 November 2008
Accepted 17 December 2008
Available online 3 January 2009
Ó 2009 Elsevier Masson SAS. All rights reserved.
Keywords:
Anticoagulant
Oral bioavailability
Factor Xa inhibitor
Biased screen
Computer-aided design
X-ray analysis
Clotting assay
1. Introduction
reached the market was Ximelagatran [3] from AstraZeneca, which
is a double prodrug of the direct thrombin inhibitor Melagatran.
Prevention of coagulation is required in a wide range of patients
presenting with conditions such as venous thromboembolism,
atrial fibrillation and acute coronary syndromes. Anticoagulants
currently in clinical use include the parenterally administered
unfractionated heparin and low molecular weight heparins
(LMWHs) as well as the orally administered coumarins. Heparin
and the LMWHs are indirect inhibitors of thrombin and factor Xa.
Ximelagatran was launched in France in 2004 under the trade name
Exanta, but has been withdrawn from the market due to concerns
about long-term liver damage and possible risk of heart attacks [4].
In the meantime PradaxaÒ [5], (developed by Boehringer Ingel-
heim), like Ximelagatran a double prodrug of a direct thrombin
inhibitor, has been approved by the European Medicines Agency in
spring 2008 and has since been launched in Germany and the UK
[6]. Johnson & Johnson/Bayer have submitted an application for US
new drug approval (NDA) of its oral F.Xa inhibitor XareltoÒ [7], and
just recently the drug was approved in Canada [8]. Bristol-Myers
Squibb’s oral F.Xa inhibitor Apixaban [9] is currently in phase III
clinical studies [10]. Oral anticoagulants which have been studied in
phase III clinical trials have been discussed in recent review articles
[11]. Several oral factor Xa inhibitors are being investigated in phase
II clinical trials. The results of these studies have been reviewed
extensively [12].
The coumarins act by inhibiting g-carboxylation in the biosynthesis
of coagulation factors prothrombin, VII, IX and X. This mechanism
leads to a slow on-set of action. Additionally, the coumarins suffer
from substantial food and drug interaction and high protein
binding [1], rendering it very difficult to maintain a balanced
plasma exposure. Careful and regular monitoring of the patient is
therefore required [2].
During the last decades substantial efforts have been devoted to
the search for an orally bioavailable replacement of the coumarins.
Out of countless research programs, the first compound to have
The serine protease factor Xa has a unique role in the coagula-
tion cascade. It is located at the convergence point of the intrinsic
and extrinsic pathways and therefore represents an interesting
target for the development of a novel anticoagulant. During the
past few years enormous efforts have been devoted to the search
* Corresponding author. Tel.: þ41 61 688 83 70; fax: þ41 61 688 97 48.
E-mail address: katrin.groebke_zbinden@roche.com (K. Groebke Zbinden).
0223-5234/$ – see front matter Ó 2009 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2008.12.025

2788
K.G. Zbinden et al. / European Journal of Medicinal Chemistry 44 (2009) 2787–2795
inhibitors has been published by a research group from Bristol-
Myers Squibb [14].
O
A
S1 pocket
B
Tyr99
O
N
S
2. Results and discussion
HN
O
Like other proteases of the coagulation cascade, factor Xa
possesses an S1 specificity pocket recognizing basic motifs by
means of the Asp189 side chain at its bottom. Inhibitors with
typical basic S1 binding elements such as amidines suffer often
from poor bioavailability. At the outset of this project, ample
precedence in the literature indicated that chlorinated phenyl or
heteroaromatic rings would be good neutral motifs for binding to
the S1 pocket [15]. A focused screen of the Roche compound
collection was performed under the following procedure: (i)
selecting all chlorinated planar cycles without substituents in
either ortho position, (ii) docking these into the coordinates of
a published protein crystal structure (PDB code 1nfy) with the
constraint of holding the C–Cl bond in a position allowing a chlo-
rine–aromatic interaction with Tyr228 and (iii) removing typical
GPCR ligands and highly flexible compounds from the solution set.
This resulted in about 8000 compounds, for which single point
Trp215
Tyr228
Gly218
Cl
S4 pocket
(hydrophobic box)
K (F.Xa) = 35 µM
Phe174
i
Fig. 1. (A) Hit from biased screen. (B) Model of screening hit bound to the active site of
F.Xa.
for an orally active and selective factor Xa inhibitor resulting in
a great number of disclosed structures [13].
Here we report the identification of novel 3-aminopyrrolidine
factor Xa inhibitors starting from a hit originated from a biased
screen. Recently a similar series of aminopyrrolidine factor Xa
Table 1
Binding affinity, functional activity and selectivity of aminothiazoles 1–6.
S4
O
N
S
HN
O
S1
Compound
S1
S4
Ki (
mM)
Thrombin F.VIIa
2 ꢀ PT (
m
M)
2 ꢀ aPTT (
m
M)
log D
Solubility (mg/mL)
F.Xa
Thrombin
2.8
O
N
1
2
0.442
0.162
6.3
53
17.8
8.9
32.0
20.9
2.2
0.6
>580
>620
N
N
N
N
N
8.6
Cl
F
H
N
N
O
3
4
0.112
0.099
>35
>310
45.6
12.2
70.1
33.8
>3.0
<1.0
O
N
5.9
60
2.1
>640
N
N
N
N
S
N
5
6
0.032
0.018
15.3
480
4.4
10.4
73.3
1.0
>620
<1.0
Cl
F
H
N
>35
>1950
43.8
>3.0
N
O

K.G. Zbinden et al. / European Journal of Medicinal Chemistry 44 (2009) 2787–2795
2789
Table 2
Variation of central heterocycle (N-linked derivatives): influence of binding affinity and functional activity of compounds 7–13.
O
O
S
Cl
S4
N
N
H
Compound
Central heterocycle
S4
F.Xa
2 ꢀ PT (
m
M)
2 ꢀ aPTT (
m
M)
log D
Solubility (
m
g/mL)
N
N
7
0.074
0.006
0.175
4.2
5.6
0.1
>601
N
N
F
H
N
N
O
8
9
6.4
2.7
>3
<1
N
N
N
11.4
22.6
ꢁ0.5
>600
N
F
N
N
H
N
N
10
11
12
0.006
0.011
4.7
2.0
2.4
2.4
1.0
O
F
F
F
N N
H
N
N
O
N
13.3
19.5
<1.0
N
H
N
N
O
N
N
0.008
1.272
1.5
1.8
2.3
2.3
197
166
H
N
13
n.d.
n.d.
N
O
measurements at 20
chromogenic enzyme assay. A total of 60 compounds gave signals
below 70% control at this concentration.
m
M concentration were run in a standard
3-fluorophenyl-pyridone [16] moiety proved to be most suitable for
enzyme affinity, the methyl–piperidinyl–piperazine [17] group is
by far superior with respect to activity in the ex vivo clotting assays
PT and aPTT. This can be explained by the substantially lower lip-
ophilicity of compound 2 which is also reflected in a reduced log D
value and an increased aqueous solubility. Enzyme affinity and ex
vivo activity could be further improved by replacing the 4-chlor-
ophenyl S1 moiety with a 2-chlorothiophene substituent [18]
(entries 4–6, Table 1). Again, the methyl–piperidinyl–piperazine
derivative 5 is superior with respect to functional activity in plasma.
As in the chlorophenyl series, the 3-fluoro-phenyl-pyridone S4
moiety results in good enzyme affinity, but is – due to its lip-
ophilicity – poorly active in the clotting assays.
The aminothiazole hit shown in Fig. 1 is a weak inhibitor of
factor Xa. The modeled binding mode of this compound indicated
that it essentially only binds into the S1 pocket and does not extend
into the lipophilic S4 pocket formed by the aromatic side chains of
Tyr99, Phe174 and Trp215 (‘hydrophobic box’). Typical factor Xa
inhibitors gain a significant amount of their affinity by binding into
this pocket. Based on this model, it was hypothesized that inhibi-
tory activity of the screening hit could be improved when replacing
the ester functionality by an amide group with the potential to
extend into the S4 binding pocket. To that end, a great variety of
amide analogues of the screening hit were prepared, using struc-
turally diverse commercial amines as well as amines which are
described in the literature as good S4 pocket binding elements.
Compared to the screening hit, amide derivatives 1–3 (Table 1)
show an >100-fold increase in binding affinity. While the
In an attempt to further improve activity in the clotting assays,
the central thiazole core was replaced by various nitrogen hetero-
cycles (Table 2). Results for the pyrazole derivative 7 and the triazole
derivative 9 show that this core replacement leads to a sub-
stantial loss of binding affinity in the methyl–piperidinyl–piperazine

2790
K.G. Zbinden et al. / European Journal of Medicinal Chemistry 44 (2009) 2787–2795
Table 3
Variation of substituents on central pyrrolidine moiety influence of binding affinity, functional activity, microsomal stability and CYP inhibition profile of compounds 12–26.
N
R2
R1
O
O
O
S
Cl
N
H
N
N
H
F
Compound
R1
R2
F.Xa
(nM)
2 ꢀ PT
M)
2 ꢀ aPTT
Microsomal clearance
human ( l/min/mg
Microsomal clearance
rat ( l/min/mg
(
m
(
m
M)
m
m
microsomal protein)
microsomal protein)
12
14
15
16
17
18
19
20
21
22
23
24
H
OH
OMe
OEt
–O–CH2–CHF2
–O–CH2–CH2–OH
–CH2–OH
Me (rac)
Cyclopropyl (rac)
F
H
H
H
H
H
H
H
H
H
H
H
H
8
10
3
1.5
1.8
1.7
3.6
4.3
2.4
2.9
5.5
12.9
8.2
4.9
7.2
1.8
2.0
0.5
2.0
3.3
1.3
2.6
3.0
9.2
6.2
10.9
8.3
15.2
0.0
19.6
74.7
9.0
79.2
15.8
42.3
68.3
35.0
12.6
35.4
165.8
109.3
65.9
17.6
10
20
22
27
14
43
31
95
94
8.4
21.7
76.3
267.2
16.1
7.0
–CH2–OH
–CH2–OMe
34.5
102.9
N
25
26
H
H
132
93
5.6
5.4
10.9
10.4
61.4
7.5
115.5
19.5
O
N
H
subclass. However, the 3-fluorophenyl-pyridone derivatives 8, 10
and 11 are all characterized by excellent potency. For the triazole
compound 10 an optimal balance between binding affinity and lip-
ophilicity was achieved leading to a favourable plasma activity.
Unfortunately, this compound is characterized by a low bioavail-
ability in the rat (Table 4).
Replacing the aromatic by an saturated central core resulted in
the aminopyrrolidine derivative 12 with R-configuration, display-
ing excellent functional activity in plasma. The corresponding
S-enantiomer 13 has 160-fold lower enzyme affinity. The pyrroli-
dine moiety allows the introduction of additional substituents. This
set of compounds is shown in Table 3. The methoxy derivative 15 is
the most potent compound of this series.
the S4 substituent lies with its plane normal to that of the fluo-
rophenyl moiety and the fluorine atom is found exclusively on the
‘right-hand’ side of the hydrophobic box as depicted. The interac-
tion with the enzyme is mostly governed by van der Waals and
p–p
interactions with few polar contacts, namely two rather long
hydrogen bonds between the carbonyl of Gly216 and the amide NH
adjacent to the S4 moiety of the inhibitor (d w 2.9 Å) and between
the carbonyl of Gly218 and the amide NH adjacent to the S1 moiety
(d w 3.3 Å) as well as a charge dipole interaction between the
protonated pyrrolidine N and the carbonyl of Gly216 (d w 3.5 Å).
The methoxy group packs well over the disulphide bridge 191:220
and against the side chain of Gln192 which is somewhat
disordered.
X-ray structural analysis confirmed its absolute configuration
and shows that the ligand binding mode (Fig. 2) is as expected from
the literature precedence [19]. The peripheral aromatic moieties are
found in their usual positions. The chlorothiophene ring is located
in the S1 pocket, with the chlorine atom displacing a structural
water molecule and being engaged in a hydrophobic interaction
with Tyr228 (not visible in figure). 3-Fluorophenyl-pyridone is
situated in the S4 pocket (‘hydrophobic box’). The aminopyrrolidine
linker is in a strain-free conformation. The amide group adjacent to
Compared to the methoxy derivative 15, the hydroxy compound
14, the ethoxy compound 16 and the methyl compound 20 show
slightly reduced binding affinity which is also reflected in a lower
functional activity. For compounds bearing larger substituents such
as the difluoro-ethoxy moiety, the hydroxyethoxy moiety, the
hydroxymethyl moiety and the cyclopropyl group (compounds 17–
19 and 21) the reduction in binding and functional activity is more
pronounced indicating that these groups do not fit well in the
shallow cavity formed by disulphide bridge 191:220 and the side
Table 4
Physicochemical properties and pharmacokinetic profile of compounds 10, 14 and 18.
Compound Solubility Permeability log D Human
no.
Rat microsomal SDPK rat i.v.
SDPK rat p.o.
(mg/mL) (10ꢁ6 cm/s)
microsomal clearancea clearancea
Dose (mg) Clearance (mL/min//kg) T1/2 (h) Vss (l/kg) Dose (mg) T1/2 (h) Tmax (h) F (%)
(mL/min/kg)
2.4 0.4
n.m. 0.0
1.6 6.8
(mL/min/kg)
10
14
18
1.0
516
>713
0.9
1.5
0.4
6.0
15.8
12.6
0.5
0.5
1.0
25.0
74.0
57.1
0.3
0.25
0.35
0.3
1.1
1.1
1.0
1.0
1.0
0.2
0.9
2.7
1.1
0.25
0.25
10
38
16
a
In vitro CL was extrapolated into in vivo CL by physiological scaling [33].

K.G. Zbinden et al. / European Journal of Medicinal Chemistry 44 (2009) 2787–2795
2791
shown that for both compounds renal clearance contributes for at
least 50% of drug elimination. Active secretion is assumed. Also
other non-P450 elimination pathways like hydrolytic cleavage
might be involved. Detailed investigation of elimination processes
has however not been performed.
Gln192
Cys191:Cys220
3. Chemistry
Tyr99
The synthesis of the aminothiazole derivatives 1–6 (Scheme 1)
started by coupling (2-amino-thiazol-4-yl)-acetic acid ethyl ester
with either 4-chlorobenzoic acid (entries 1–3) or 5-chloro-thio-
phene-2-carboxylic acid (entries 4–6) using BOP as coupling
reagent. Subsequently the esters were hydrolyzed to the corre-
sponding acids 27a and 27b which were then coupled with the
appropriate amines in the presence of optimized coupling reagents
(2-piperazin-1-yl-1-pyrrolidin-1-yl-ethanone and BOP for intro-
duction of S4 moiety A, 1-(1-methyl-piperidin-4-yl)-piperazine and
TBTU for moiety B, 1-(4-amino-3-fluoro-phenyl)-1H-pyridin-2-one
[20] and BOP-Cl for moiety C).
Gly216
Gly218
Trp215
S4 pocket
(hydrophobic box)
Phe174
The preparation of derivatives 7–11 (Scheme 2) started with the
alkylation of the respective aminoheterocycles under optimized or
literature conditions. 3-Aminopyrazole was alkylated with ethyl
bromoacetate in the presence of potassium tert-butylate. 3-Amino-
1,2,4-triazole was alkylated with ethyl iodoacetate using sodium
hydride as base. 5-Amino-1H-tetrazole was reacted with methyl
chloroacetate with KOH in refluxing MeOH [21]. All three alkyl-
ations provided the desired regioisomers in modest yield. Struc-
tures of 28a, 28b and 28c were assigned with the help of NOE
experiments. The aminoheterocycles were coupled with 5-chloro-
thiophene-2-carboxylic acid under optimized conditions (BOP for
28a and 28b, EDCI/DMAP for 28c) to give 29a, 29b and 29c. Esters
29a and 29b were hydrolyzed and coupled with the appropriate
amines in the presence of optimized coupling reagents (1-(1-
methyl-piperidin-4-yl)-piperazine and BOP for the introduction of
S4 moiety B, 1-(4-amino-3-fluoro-phenyl)-1H-pyridin-2-one and
BOP-Cl for moiety C) to give the final products 7–10. The tetrazole
derivative 29c was directly reacted with the amine leading to S4
moiety C in the presence of AlMe3 to give final product 11.
Fig. 2. Crystal structure of compound 15 bound to the active site of F.Xa. Methods are
described in Chapter 4.2.
chain of Gln192. The reduced binding affinity of fluorine compound
22 might be explained by a diminished interaction between the
dipole of the carbonyl of Gly216 and the fluorinated pyrrolidine
which can be attributed to a reduced basicity caused by the elec-
tron-withdrawing substituent. Introduction of a substituent next to
the ring nitrogen of the pyrrolidine core leads to a substantial loss
in activity (compounds 23–26).
Unfortunately, the most active aminopyrrolidine derivative –
compound 15 – is rapidly cleared by human and rat microsomes
(Table 3). The hydroxy derivative 14 and the hydroxyethoxy
derivative 18 are somewhat less potent in the clotting assays but
have improved microsomal stability.
Pharmacokinetic parameters of compounds 14 and 18 are given
in Table 4. Derivative 14 has a moderate oral bioavailability, while
only poor plasma exposure is reached after oral administration of
derivative 18. This can be explained by its reduced membrane
permeability. Low microsomal clearance of compounds 14 and 18
did unfortunately not transfer to low in vivo clearance (Table 4).
Both derivatives have high plasma clearance in the rat. It was
The aminopyrrolidine derivatives 12 and 13 (Scheme 2) were
prepared starting with the alkylation of R- or S-3-Boc-amino-
pyrrolidine with ethyl bromoacetate. Cleavage of the Boc protecting
group, coupling with 5-chloro-thiophene-2-carboxylic acid using
BOP as coupling reagent and AlMe3-mediated reaction with 1-(4-
O
O
O
O
OH
S4
a (for 1 and 4)
c (for 2 and 5)
N
N
N
d (for 3 and 6)
a, b
S
S
S
HN
S1
HN
H2N
O
O
S1
1 S1 = 4-Chlorophenyl, S4 = A
2 S1 = 4-Chlorophenyl, S4 = B
3 S1 = 4-Chlorophenyl, S4 = C
27a S1 = 4-Chlorophenyl
27b S1 = 5-Chlorothiophen-2-yl
S
Moieties
4
4 S1 = 5-Chlorothiophen-2-yl, S4 = A
5 S1 = 5-Chlorothiophen-2-yl, S4 = B
6 S1 = 5-Chlorothiophen-2-yl, S4 = C
N
N
N
N
F
N
O
N
A
B
C
O
N
N
H
Scheme 1. Preparation of 2-aminothiazole derivatives 1–6. Conditions: (a) BOP, DIPEA, DMF, r.t.; (b) NaOH, MeOH/H2O, 0 ꢂC; (c) TBTU, TEA, CH2Cl2, r.t.; (d) BOP-Cl, DIPEA, MeCN,
DMF, r.t.

2792
K.G. Zbinden et al. / European Journal of Medicinal Chemistry 44 (2009) 2787–2795
O
O
O
Y
R
R
O
a (for 28a)
b (for 28b)
Y
c (for 28c)
O
H
N
S4
d (for 29a and 29b)
e (for 29c)
f, d (for 7 and 9)
f, g (for 8 and 10)
N
N
N
N
Y
X
N
Y
X
N
N
X
X
H2N
H2N
HN
HN
O
O
29a X = Y = C, R = Et
29b X = N, Y = C, R = Et
29c X = Y =N, R = Me
28a X = Y = C, R = Et
28b X = N, Y = C, R = Et
28c X = Y =N, R = Me
S
S
Cl
Cl
7
8
9
X = Y = C, S4 = B
X = Y = C, S4 = C
X = N, Y = C, S4 = B
h
O
O
S4
10 X = N, Y = C, S4 = C
S4
H
N
N
i, j, d, h
N
N
Y
X
*
*
HN
HN
O
HN
O
O
O
S
11 X = Y = N, S4 = C
S
Cl
Cl
12 R enantiomer
13 S enantiomer
Scheme 2. Preparation of derivatives 7–13. Conditions: (a) ethyl bromoacetate, KOtBu, DMF, 0 ꢂC; (b) ethyl iodoacetate, NaH, DMF, 0 ꢂC; (c) methyl chloroacetate, KOH, MeOH,
reflux; (d) BOP, DIPEA, THF or DMF, r.t.; (e) EDCI, TEA, DMAP, CH2Cl2, r.t.; (f) NaOH, EtOH/H2O, r.t.; (g) BOP-Cl, DIPEA, MeCN, r.t.; (h) AlMe3, toluene, 90 ꢂC; (i) ethyl bromoacetate,
TEA, THF, r.t. to 50 ꢂC; (j) 4 N HCl in dioxane, r.t.
N
O
O
O
O
O
O
N
H
N
F
N
N
c, d
a, b
HN
OH
HN
OH
N
OH
N+
O
O
N
14
S
S
30
Cl
Cl
N
O
O
O
O
N
H
e (for 32a)
f (for 32b)
N
F
N
c, d
HN
R
16 OEt
17 OCH2CHF2
HN
R
O
O
S
O
g (for 32c)
h (for 32d)
i (for 32e)
S
32a OEt
32b OCH2CHF2
O
Cl
N
F
Cl
Br
+
Br
H2N
N
O
O
O
O
a, b, c, d
N
H
j
N
N
F
O
N
N+
R
HN
R
F
N
O
15 OMe
18 OCH2CH2OH
22 F
O
32c OMe
32d OCH2CH2OTHP
32e F
N
Br
N
H
S
31
Cl
Scheme 3. Preparation of aminopyrrolidine derivatives 14–18 and 22. Conditions: (a) H2, PtO2, MeOH, r.t.; (b) BOP, DIPEA, DMF, r.t.; (c) 4 N HCl in dioxane, r.t.; (d) 31, TEA, DMF, r.t.;
(e) EtI, Ag2O, MeCN/THF 3:1, r.t.; (f) 2,2-difluoroethyl trifluoromethanesulfonate, NaH, DMF, 0 ꢂC; (g) MeI, Ag2O, MeCN/THF 5:1, r.t.; (h) 2-(2-bromoethyl)tetrahydro-2H-pyran, NaH,
DMF, r.t.; (i) DAST, CH2Cl2, ꢁ78 ꢂC; (j) TEA, THF, r.t.

K.G. Zbinden et al. / European Journal of Medicinal Chemistry 44 (2009) 2787–2795
2793
N
N
O
O
O
O
N
H
N
H
O
O
O
O
N
F
N
F
N
N
a
b, c, d, e
f
O
S
O
S
O
O
OH
N
H
N
H
O
O
O
N
O
H
O
OH O
O
Si
Cl
Cl
33
34
19
Scheme 4. Preparation of aminopyrrolidine derivative 19: (a) diphenylphosphorylazide, 2-(trimethylsilyl)ethanol, TEA, toluene, 80 ꢂC; (b) TBAF, acetonitrile, 80 ꢂC; (c) isobutyl
chloroformate, N-methyl morpholine, acetonitrile, r.t.; (d) 4 N HCl in dioxane; (e) 31, K2CO3, MeCN; (f) NaBH4, EtOH, r.t.
amino-3-fluoro-phenyl)-1H-pyridin-2-one leading to S4 moiety C
provided final products 12 and 13.
coupling with 5-chloro-thiophene-2-carboxylic acid with isobutyl
chloroformate as coupling reagent, deprotection of the pyrrolidine
moiety and alkylation with bromide 31 provided the ester inter-
mediate 34 which was reduced with sodium borohydride to
provide the desired final product 19.
The preparation of the hydroxy-substituted aminopyrrolidine
derivative 14 (Scheme 3) involved the hydrogenation of (3S,4S)-3-
azido-4-hydroxy-pyrrolidine-1-carboxylic acid tert-butyl ester [22]
as a first step. The resulting aminoalcohol was then coupled with 5-
chloro-thiophene-2-carboxylic acid using BOP as coupling reagent
to give intermediate 30. Deprotection of the pyrrolidine moiety and
subsequent alkylation with bromide 31 yielded the final
compound. Bromide 31 which was also used in many of the
syntheses described below was obtained by reaction of 1-(4-
amino-3-fluoro-phenyl)-1H-pyridin-2-one [20] with bromoacetyl
bromide in the presence of TEA.
The synthesis of alkylsubstituted aminopyrrolidine derivatives
20 and 21 (Scheme 5) was accomplished starting with a 1,3-
dipolar cycloaddition reaction between the appropriate
a,b-
unsaturated esters and the azomethine ylide generated by TFA
treatment of N-(methoxy-methyl)-N-(trimethylsilylmethyl)ben-
zyl-amine [24]. The resulting pyrrolidines 35a and 35b were
debenzylated by hydrogenation and Boc-protected. After ester
hydrolysis,
a
Curtius rearrangement provided the amino-
The first step in the preparation of the substituted amino-
pyrrolidine derivatives 16 and 17 was the alkylation of the hydroxy
intermediate 30 with ethyl iodide or 2,2-difluoroethyl tri-
fluoromethanesulfonate respectively. The pyrrolidine ring of 32a
and 32b was deprotected and then alkylated with bromide 31 to
lead to the desired final products.
pyrrolidines 36a and 36b. Coupling with 5-chloro-thiophene-
2-carboxylic acid, cleavage of the Boc protecting group and
alkylation of the pyrrolidine with bromide 31 gave the targeted
final products.
The proline-derivedcompounds23–26(Scheme 6) were prepared
starting with a coupling reaction between N-Boc-trans-4-amino-L-
Derivatives 15 and 18 were obtained starting by alkylation of
(3S,4S)-3-azido-4-hydroxy-pyrrolidine-1-carboxylic acid tert-butyl
ester with methyl iodide and Ag2O or 2-(2-bromoethyl)tetrahydro-
2H-pyran and NaH respectively. Intermediates 32c and 32d were
hydrogenated and subsequently coupled with 5-chloro-thiophene-
2-carboxylic acid. Then the pyrrolidine ring was deprotected and
alkylated with bromide 31.
The fluoro compound 22 was obtained by treating (3S,4S)-3-
azido-4-hydroxy-pyrrolidine-1-carboxylic acid tert-butyl ester
with DAST which according to the literature [23] leads to the trans
isomer. The resulting intermediate 32e was converted to the final
product as described for compounds 15 and 18.
proline methyl ester and 5-chloro-thiophene-2-carboxylic acid. The
resulting intermediate 37 was deprotected and alkylated with
bromide 31 to give ester 38 which was subsequently either reduced
with sodium borohydride to give the hydroxymethyl derivative 23 or
hydrolyzed and coupled with cyclopropylmethylamine to give amide
26. Intermediate 37 could alternatively be reduced to the corre-
sponding hydroxymethyl compound using NaBH4 in the presence of
lithium chloride which was then transformed into the mesylate 39
and subsequently reacted with methanolate to give 24 or with pyr-
rolidine to give 25. Intermediate 37 could alternatively be reduced to
the corresponding hydroxymethyl compound using NaBH4 in the
presence of lithium chloride which was then transformed into the
mesylate 39 and subsequently reacted with methanolate to give 40a
or with pyrrolidine to give 40b. The final products 24 and 25 were
obtained by cleavage of the Boc protecting group and alkylation with
bromide 31.
The synthesis of the hydroxymethyl compound 19 (Scheme 4)
started with a Curtius rearrangement of pyrrolidine-1,3,4-tricar-
boxylic acid 1-tert-butyl ester 3-ethyl ester in the presence of 2-
(trimethylsilyl)ethanol. Cleavage of the Teoc group with TBAF,
O
S4
N
H
O
O
O
N
N
Si
N
N
a
b,c,d,e
f,g,h
HN
R
+
O
O
H2N
R
R
R
O
S
Cl
rac
rac
rac
O
O
35a Me
35b cyclopropyl
36a Me
36b cyclopropyl
20 Me
21 cyclopropyl
Scheme 5. Preparation of aminopyridine derivatives 20 and 21. Conditions: (a) TFA, CH2Cl2, 0 ꢂC; (b) H2, Pd/C 10%, MeOH, r.t.; c) Boc2O, DMAP, CH2Cl2, r.t.; (d) 1 N NaOH THF, 0 ꢂC;
(e) DPPA, DIPEA, toluene or dioxane, 90 ꢂC; (f) BOP, DIPEA, DMF, r.t.; (g) 4 N HCl, dioxane, r.t.; (h) 31, TEA, THF, r.t.

2794
K.G. Zbinden et al. / European Journal of Medicinal Chemistry 44 (2009) 2787–2795
O
O
O
O
O
O
S4
S4
N
H
N
H
O
O
O
N
N
N
N
a
b,c
d
O
O
O
HO
NH2
NH
NH
NH
O
O
O
37
38
23
S
S
S
Cl
Cl
Cl
g,h
O
S4
O
O
O
O
O
N
H
S4
e,f
O
S
N
N
N
N
O
O
H
i (for 40a)
j (for 40b)
R
R
O
N
b,c
N
NH
S
NH
NH
H
NH
O
O
O
O
S
S
S
26
Cl
25 1-Pyrrolidinyl
Cl
Cl
24 OMe
40a OMe
39
Cl
40b 1-Pyrrolidinyl
Scheme 6. Preparation of proline-derived aminopyridine derivatives 23–26. Conditions: (a) BOP, DIPEA, DMF, r.t.; (b) 4 N HCl in dioxane, r.t.; (c) 31, TEA, DMF, r.t.; (d) NaBH4, MeOH,
ꢂC; (e) 1 N NaOH, THF, r.t.; (f) cyclopropylmethylamine, BOP, DIPEA, DMF (g) NaBH4, LiCl, EtOH, THF, r.t.; (h) MsCl, DIPEA, CH2Cl2, 0 ꢂC; (i) NaOMe, MeOH/THF, reflux; (j) pyrrolidine,
0
THF, r.t.
4. X-ray structural analysis: experimental
from the crystal and scanned with 0.15 mm pixel size. Exposure
times were 600 s for 0.5ꢂ frames. Data from 360 frames were pro-
cessed to 1.95 Å resolution using XDS [27]. The space group is P21
with unit cell dimensions a ¼ 48.96, b ¼ 77.19, c ¼ 74.76 Å,
4.1. Clotting assays
Prothrombin time PT: 2 ꢀ PT prolongation is a measure for the
ability of an inhibitor to prevent clotting via the extrinsic pathway
of the coagulation cascade. Experimental procedure: human citra-
ted plasma is spiked with at least 6 concentrations of inhibitor.
Clotting is initiated by addition of exogenous tissue factor (Innovin).
Clotting time is determined by a turbidity measurement. The
concentration of inhibitor necessary to double control clotting time
is determined by fitting the data to an exponential regression.
Activated partial thromboplastin time aPTT: 2 ꢀ prolongation of
aPTT serves as a measure for the ability of the inhibitor to interfere
with the coagulation cascade via the intrinsic pathway. Experi-
mental procedure: human citrated plasma is spiked with at least 6
concentrations of inhibitor. Clotting via the intrinsic pathway is
initiated by addition of ActinÒ FS (ellagic acid in soy phosphatides).
Clotting time is determined by a turbidity measurement. The
concentration of inhibitor necessary to double control clotting time
is determined by fitting the data to an exponential regression.
See Ref. [25] for further experimental details.
b
¼ 92.12ꢂ. For 133,552 observations of 36,457 reflections, the
merging R factor on intensities was 11.9% (38.5% in the outermost
shell,1.95–2.01 Å), with completeness 90.3% (95.9%) and I/s 7.8 (3.0).
Data reduction used the CCP4 package [28]. The structurewas solved
by molecular replacement using 2bok.pdb as model. Model building
with Moloc [29,30] and Coot [31] and refinement with Refmac5 [32]
gave final overall crystallographic R factors of 20.3% (working) and
27.5% (free), with values in the outer shell (1.95–2.0 Å) of 24.1% and
34.4%, respectively, for 5088 atoms, including one Caþþ ion, two Naþ
ions, and 576 water molecules. The observed monoclinic form is
similar in packing to the commonly found orthorhombic form (e.g.
2bok.pdb) and may arise due to partial degradation of the calcium
loop. In one unit cell two independent, but very similar protein
molecules are observed. The inhibitor density in both molecules is
clear and unambiguous. Coordinates have been deposited at the
Protein Data Bank with the PDB code 2vvc.
X-ray structures of further complexes between factor Xa and
selected inhibitors from Tables 2 and 3 are deposited in the Protein
Data Bank as well under the following PDB codes: 2vvv (compound
10), 2vvu (compound 12), 2vwn (compound 14), 2vwo (compound
22), 2vwl (compound 23), 2vwm (compound 26).
4.2. X-ray structural analysis
Short form Factor Xa was produced as previously described [26]
as the Arg150Glu mutant. Compound 15 was pre-incubated with the
protein and crystals grown by addition of 0.1 M bis–tris pH 6.5, 25%
PEG3350 in microbatch mode. Crystals appeared after a week and
grew to full size within two weeks. Before data collection, crystals
were transferred to crystallization buffer supplemented with 20%
glycerol and flash-frozen in liquid N2. Diffraction data were
measured on a Bruker FR591 X-ray generator, with 0.2 mm focus,
run at 50 kV/60 mA and equipped with an Osmic focussing mono-
chromator and an Oxford Cryostream cooler run at 100 K. The
Marresearch345 (dtb) image plate detector was placed 150 mm
5. Conclusion
A weakly active aminothiazole factor Xa inhibitor was identified
from a biased screen. Guided by molecular modeling, the binding
affinity of the initial lead was improved. Aminothiazole derivatives
with nanomolar affinity were identified. Replacement of the central
aminothiazole core by various nitrogen-containing aromatic
heterocycles resulted in inhibitors with even better binding affinity
and good functional activity in plasma clotting assays. Further
improvement of functional activity was achieved by the

K.G. Zbinden et al. / European Journal of Medicinal Chemistry 44 (2009) 2787–2795
2795
(d) A.G.G. Turpie, Arterioscler. Thromb. Vasc. Biol. 27 (2007) 1238–1247;
(e) A.C. Spyropoulos, Expert Opin. Investig. Drugs 16 (4) (2007) 431–440;
(f) S. Haas, S. Schellong, Ha¨mostaseologie 27 (2007) 41.
introduction of a pyrrolidine moiety as central core. The presence of
the pyrrolidine ring allowed the modulation of binding affinity,
functional activity and in vitro pharmacokinetic parameters by
introduction of a large variety of ring substituents. Potent inhibitors
with good microsomal stability were identified.
´
[13] A. Casimiro-Garcıa, A.D. Dudley, R.J. Heemstra, K.J. Filipski, F.C. Bigge,
J.J. Edmunds, Expert Opin. Ther. Pat. 16 (2) (2006) 119–145.
[14] Y. Shi, D. Sitkoff, J. Zhang, W. Han, Z. Hu, P.D. Stein, Y. Wang, L.J. Kennedy,
S.P. O’Connor, S. Ahmad, E.C.-K. Liu, S.M. Seiler, P.Y.S. Lam, J.A. Robl, J.E. Macor,
K.S. Atwal, R. Zahler, Bioorg. Med. Chem. Lett. 17 (2007) 5952.
[15] (a) M. Adler, M.J. Kochanny, B. Ye, G. Rummenik, D.R. Light, S. Biancalana,
M. Whitlow, Biochemistry 41 (2002) 15514;
Acknowledgements
(b) S. Maignan, J.-P. Guilloteau, Y.M. Choi-Sledeski, M.R. Becker, W.R. Ewing,
H.W. Pauls, A.P. Spada, V. Mikol, J. Med. Chem. 46 (2003) 685–690;
(c) M. Nazare´, D.W. Will, H. Matter, H. Schreuder, K. Ritter, M. Urmann,
M. Essrich, A. Bauer, M. Wagner, J. Czech, M. Lorenz, V. Laux, V. Wehner, J. Med.
Chem. 48 (2005) 4511–4525;
The authors wish to thank Louis Burcklen, Olivier Kuster, Mar-
tine Stihle, and Martin Weber for skillful technical assistance and
Thomas Tschopp for helpful discussion.
´
(d) H. Matter, D.W. Will, M. Nazare, H. Schreuder, V. Laux, V. Wehner, J. Med.
Chem. 48 (2005) 3290–3312.
References
[16] C.F. Bigge, D.A. Dudley, J.J. Edmunds, C.A. Van Huis, A. Casimiro-Garcı´a, K.J.
Filipski, J.T. Kohrt, WO2003045912.
[17] J.W. Liebeschuetz, A.J. Lyons, C.W. Murray, A.D. Rimmer, S.C. Young, N.P. Camp,
S.D. Jones, P.J. Morgan, S.J. Richards, W.A. Wylie, S.E. Lively, M.J. Harrison, B.
Waszkowycz, J.J. Masters, M.J. Wiley, WO2000076970.
[1] P.S. Wells, A.M. Holbrook, N.R. Crowther, J. Hirsh, Ann. Intern. Med. 121 (1994)
676–683.
[2] (a) T.M. Hyers, G. Agnelli, R.D. Hull, T.A. Morris, M. Samama, V. Tapson,
J.G. Weg, Chest 119 (2001) 176S–193S;
[18] S. Roehrig, A. Straub, J. Pohlmann, T. Lampe, J. Pernerstorfer, K.-H. Schlemmer,
P. Reinemer, E. Perzborn, J. Med. Chem. 48 (2005) 5900–5908.
[19] (a) J.X. Qiao, C.-H. Chang, D.L. Cheney, P.E. Morin, G.Z. Wang, S.R. King,
T.C. Wang, A.R. Rendina, J.M. Luettgen, R.M. Knabb, R.R. Wexler, P.Y.S. Lam,
Bioorg. Med. Chem. Lett. 17 (2007) 4419–4427;
(b) C.S. Landefeld, M.W. Rosenblatt, L. Goldman, Am. J. Med. 87 (1989) 153–159;
(c) E.M. Hylek, A.S. Go, Y. Chang, N.G. Jensvold, L.E. Henault, J.V. Selby,
D.E. Singer, N. Engl. J. Med. 349 (2003) 1019–1026.
[3] (a) B. Eriksson, Drugs 64 (Suppl. 1) (2004) 27–35;
(b) M.A. Crowther, J.I. Weitz, Expert Opin. Investig. Drugs 13 (2004) 403–413;
(c) K.L. McCall, E.J. MacLaughlin, J. Pharm. Technol. 19 (2003) 222–228;
(d) S. Sander, C.M. White, Formulary 39 (8) (2004) 398–404;
(e) D. Gustafsson, R. Bylund, T. Antonsson, I. Nilsson, J.-E. Nystroem, U. Eriksson,
U. Bredberg, A.-C. Teger-Nilsson, Nat. Rev. Drug Discov. 3 (8) (2004) 649–659;
(f) C.J. Boos, A. Hinton, G.Y.H. Lip, Eur. J. Intern. Med. 16 (2005) 267.
[4] P. Mitchell, Chem. Ind. 18 (2004) 5.
[5] (a) M. Ieko, Curr. Opin. Investig. Drugs 8 (9) (2007) 758–768;
(b) L.A. Sorbera, J. Bozzo, J. Castan˜er, Drugs Future 30 (9) (2005) 877–885.
[6] Boehringer ingelheim press release from 2008-03-27: <http://www.boehringer-
ingelheim.com/corporate/news/press_releases/detail.asp?ID ¼ 5334>.
[7] (a) E. Perzborn, D. Kubitza, F. Misselwitz, Ha¨mostaseologie 27 (2007) 282–289;
(b) V. Laux, E. Perzborn, D. Kubitza, F. Misselwitz, Semin. Thromb. Hemost. 33
(2007) 515–523;
(b) J.X. Qiao, T.C. Wang, G.Z. Wang, D.L. Cheney, K. He, A.R. Rendina, B. Xin,
J.M. Luettgen, R.M. Knabb, R.R. Wexler, P.Y.S. Lam, Bioorg. Med. Chem. Lett. 17
(2007) 5041–5048;
´
(c) C.A. Van Huis, C.F. Bigge, A. Casimiro-Garcıa, W.L. Cody, D.A. Dudley,
K.J. Filipski, R.J. Heemstra, J.T. Kohrt, L.S. Narasimhan, R.P. Schaum, E. Zhang,
J.W. Bryant, S. Haarer, N. Janiczek, R.J. Leadley Jr., T. McClanahan,
J.T. Petersen, K.M. Welch, J.J. Edmunds, Chem. Biol. Drug Des. 69 (2007)
444–450.
[20] K. Filipski, J.T. Kohrt, A. Casimiro-Garcia, C.A. Van Huis, D.A. Dudley, W.L. Cody,
C.F. Bigge, S. Desiraju, S. Sun, S.N. Maiti, M.R. Jaber, J.J. Edmunds, Tetrahedron
Lett. 47 (44) (2006) 7677–7680.
[21] F. Einberg, J. Org. Chem. 35 (11) (1970) 3978–3980.
[22] S.E. Schaus, J.F. Larrow, E.N. Jacobsen, J. Org. Chem. 62 (12) (1997) 4197–4199.
[23] D. Bouzard, P. Di Cesare, M. Essiz, J.P. Jacquet, J.R. Kiechel, P. Remuzon,
A. Weber, T. Oki, M. Masuyoshi, et al., J. Med. Chem. 33 (5) (1990) 1344.
[24] A. Padwa, W. Dent, Org. Synth. 67 (1989) 133.
[25] G.C. White, V.J. Marder, R.W. Coleman, J. Hirsh, E. Salzman, Haemostasis and
Thrombosis, Basic Principles and Clinical Practice, second ed. (1987)
1048–1060.
[26] (a) K.-P. Hopfner, H. Brandstetter, A. Karcher, E. Kopetzki, R. Huber, R.A. Engh,
W. Bode, EMBO J. 16 (1997) 6626–6635;
(c) P. Kakar, T. Watson, G.Y.H. Lip, Drugs Today 43 (3) (2007) 129–136;
(d)P. Kakar, T. Watson, G.Y.H. Lip, Curr. Opin. Investig. Drugs 8 (3) (2007) 256–265;
(e) G. Escolar, J. Villalta, F. Casals, J. Bozzo, N. Serradell, J. Bolo´s, Drugs Future 31 (6)
(2006) 484–493.
[8] FDA news from 2008-08-06: <http://www.fdanews.com/newsletter/
article?issueId ¼ 11832&articleId ¼ 109191>; Bayer press release from 2008-
09-16: <http://www.press.bayer.com/baynews/baynews.nsf/id/CEE29C7D543A
B693C12574C5005133F1?Open&;ccm ¼ 001>.
(b) K. Scha¨rer, M. Morgenthaler, R. Paulini, U. Obst-Sander, D.W. Banner,
D. Schlatter, J. Benz, M. Stihle, F. Diederich, Angew. Chem., Int. Ed. Engl. 44
(2005) 4400–4404.
[9] (a) M.R. Lassen, B.L. Davidson, A. Gallus, G. Pineo, J. Ansell, D. Deitchman,
J. Thromb. Haemost. 5 (2007) 2368;
(b) D.J.P. Pinto, M.J. Orwat, S. Koch, K.A. Rossi, R.S. Alexander, A. Smallwood,
P.C. Wong, R. Alan, J.M. Luettgen, R.M. Knabb, K. He, B. Xin, R.R. Wexler,
P.Y.S. Lam, J. Med. Chem. 50 (22) (2007) 5339–5356;
[27] W. Kabsch, J. Appl. Crystallogr. 26 (1993) 795–800.
[28] Colloborative Computational Project Number 4, Acta Crystallogr. D50 (1994)
760–763.
(c) S.M. Bates, J.I. Weitz, Drugs Future 33 (2008) 293–301.
[29] P.R. Gerber, K. Mu¨ller, J. Comput. Aided Mol. Des. 9 (1995) 251–268.
[30] Gerber Molecular Design, <http://www.moloc.ch>.
[31] G.N. Murshudov, A.A. Vagin, E.J. Dodson, Acta Crystallogr. Sect. D D53 (1997)
240–255.
[10] <http://www.clinicaltrials.gov/ct2/results?term ¼ apixaban>.
[11] (a) K.A. Bauer, Curr. Opin. Hematol. 15 (2008) 509;
(b) K.A. Bauer, J. Thromb. Haemost. 5 (2007) 2175–2177;
(c) Q. Ma, Br. J. Clin. Pharmacol. 64 (3) (2007) 263–265.
[32] P. Emsley, K. Cowtan, Acta Crystallogr. Sect. D 60 (2004) 2126–2132.
[33] R.S. Obach, J.G. Baxter, T.E. Liston, B.M. Silber, B.C. Jones, F. Macintyre,
D.J. Rance, P. Wastall, J. Pharmacol. Exp. Ther. 283 (1) (1997) 46.
[12] (a) D. Kubitza, S. Haas, Expert Opin. Investig. Drugs 15 (8) (2006) 843–855;
(b) S.M. Bates, J.I. Weitz, Br. J. Haematol. 134 (2006) 3–19;
(c) J.I. Weitz, L.-A. Linkins, Expert Opin. Investig. Drugs 16 (3) (2007) 271–282;