Halothiophene benzimidazoles as P1 surrogates of inhibitors of blood coagulation factor Xa

Article abstract of DOI:10.1016/j.bmcl.2004.04.097

Neutral weak halothiophene benzimidazole inhibitors of the serine protease factor Xa were identified via screening of a compound library. The X-ray crystal structure of representative 3a bound to human fXa confirmed the S1 binding mode. Starting from 3a a

Full text of DOI:10.1016/j.bmcl.2004.04.097

Bioorganic & Medicinal Chemistry Letters 14 (2004) 3763–3769
Halothiophene benzimidazoles as P1 surrogates of
inhibitors of blood coagulation factor Xa
Werner W. K. R. Mederski,* Dieter Dorsch, Soheila Anzali, Johannes Gleitz,
Bertram Cezanne and Christos Tsaklakidis
Merck KGaA, Preclinical Pharmaceutical Research, 64271 Darmstadt, Germany
Received 9 January 2004; revised 23 April 2004; accepted 29 April 2004
Available online 25 May 2004
Abstract—Neutral weak halothiophene benzimidazole inhibitors of the serine protease factor Xa were identified via screening of a
compound library. The X-ray crystal structure of representative 3a bound to human fXa confirmed the S1 binding mode. Starting
from 3a a series of halothiophene benzimidazoles was synthesized and investigated for their factor Xa inhibitory activity. This led to
potent and selective achiral inhibitors against fXa such as compounds 9k and 9w.
Ó 2004 Elsevier Ltd. All rights reserved.
1. Introduction
primary specificity (S1) pocket, which is the main
anchoring point for positively charged substrates. Most
of the known first and second generation factor Xa
inhibitors have a benzamidine moiety as a P1 residue.
These amidine groups in P1 position engage in a salt
bridge with the negatively charged Asp189.2 At the outset
of our research program obtaining active fXa inhibitors,
Factor Xa (fXa), a trypsin-like serine protease, is in-
volved in the process of blood coagulation. At the
convergence point of the extrinsic and intrinsic coagu-
lation pathways fXa, as a component of the prothrom-
binase complex, converts prothrombin to thrombin via
proteolysis. Due to the specific mechanism within these
pathways it is assumed that inhibition of this penulti-
mative enzymatic step will allow the effective control of
thrombogenesis with a minimal effect upon bleeding.1
we have synthesized and evaluated
L-phenylglycine
derivatives such as mono-benzamidine 1 (Fig. 1).3 An
X-ray structure of 1 bound to human fXa is illustrated
in Figure 2 and confirmed the depicted binding mode.
As a member of the trypsin-like serine proteases factor
Xa has a negatively charged Asp189 at the bottom of the
Although these benzamidines are potent inhibitors
of fXa they contain a polar basic group, which is
O
N
O
H
H
N
H
N
O
N
H2N
O
N
H
N
H
NH
O
O
R
N
Cl
1 IC50 fXa = 15.0 nM
2a R = Pr IC50 fXa = 5.6 nM
2b R = H IC50 fXa = 32.0 nM
Figure 1. FXa binding affinities of benzamidine 1, and chlorophenyl ureas 2a and 2b.
Keywords: Anti-coagulants; Factor Xa inhibitors; Halothiophene benzimidazoles; X-ray crystal structures.
* Corresponding author. Tel.: +49-(0)6151727683; fax: +49-(0)6151723129; e-mail: mederski@merck.de
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.04.097

3764
W. W. K. R. Mederski et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3763–3769
ꢀ
Figure 2. X-ray structures of 3a and 1 bound to human factor Xa, refined to 2.3 and 2.0 A, respectively.
associated with poor pharmacokinetic (PK) properties. As
part of our research program to obtain inhibitors with
optimized PK profiles, the chiral norvaline derivative 2a
was discovered (Fig. 1).4 A feature of this nonbasic
compound is the chlorophenyl as a surrogate for benz-
The current article discusses the design, synthesis, and
in vitro structure–activity relationships of potent achiral
and nonbasic factor Xa inhibitors containing a halo-
thiophene benzimidazole as P1 residue and tethered
cyclic anilino amides as P4 ligands starting from X-ray
crystallographic data of compound 3a.
amidine in the S1 binding pocket, a D-amino acid as the
central scaffold and a phenyl morpholinone as the P4
ligand, respectively. However, the achiral parent ana-
logue 2b displayed a 6-fold diminished potency in
comparison with 2a indicating that a central branching
contributes to the activity significantly.
At the outset of this study an X-ray crystallographic
analysis has been performed to confirm our binding
hypothesis. Compound 3a was co-crystallized with
human factor Xa (Fig. 2). The chlorothiophene benz-
imidazole of 3a replaces the benzamidine group of
compound 1 in the S1 pocket. Compound 3a does not
have a charged group that forms a salt bridge with the
carboxyl group of Asp189. However, a conserved water
molecule within the S1 pocket was displaced by the
chlorine and makes a hydrophobic contact with the
To search for novel benzamidine mimics a high-
throughput screen of our in-house library was under-
taken. As a result, neutral compounds 3a and 3b as weak
inhibitors of factor Xa were identified. These pyridazi-
nones originally turned out to be potent phosphodi-
esterase-III inhibitors with a positive inotropic action.5
Both of the compounds are comprised of a 2,3,4,5-
tetrahydropyridazin-3-one attached to a halothiophene
benzimidazole from which the latter is believed to oc-
cupy the S1 pocket.
ꢀ
phenyl ring of Tyr228 (3.7 A) and the side chain of
ꢀ
Ala190 (3.6 A). Another major interaction is a hydrogen
bond between the benzimidazole nitrogen and the car-
ꢀ
bonyl of Gly219 (3.1 A). The pyridazinone moiety has
no relevant interactions and points out of the binding

W. W. K. R. Mederski et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3763–3769
3765
O
Me
S4 residue
N
NH
A
N
N
O
(B)n
N
S
N
H
R
from 1
S
N
H
R
A = linkage, e.g.amide bond
B = CH2, O
3a R = Cl , IC50 fXa = 1100.0 nM
3b R = Br , IC50 fXa = 980.0 nM
n = 0, 1
Figure 3. FXa binding affinities of 2,3,4,5-tetrahydro-pyridazin-3-ones 3a,b, and conception of halothiophene benzimidazole analogues.
site. In contrast to benzamidine 1 there are no binding
interactions in the S4 pocket. The piperidone part in
compound 1 has hydrophobic interactions within this
pocket. It is covered between the phenyl rings of Phe174
with benzotriazole derivative TBTU generated the target
amide 9q.
The preparation of key intermediates 11, 13, and 16 not
described in the literature is outlined in Scheme 2.
Benzimidazole 11 was formed in two steps from 4-nitro-
phenylendiamine 10 and aldehyde 5 with pyrosulfite and
subsequent reduction of the nitro group using stannous
chloride. Benzimidazole 13 was prepared from 3,4-di-
amino-benzonitrile 12 and aldehyde 5 again with pyro-
sulfite whereas the cyano intermediate was reduced to
methylamine 13 with lithium aluminum hydride. Start-
ing from 3,4-dinitro-phenol 14 benzimidazole 16 was
obtained in four steps. Alkylation of phenol 14 with
bromo-acetic acid ethyl ester and hydrogenation of the
corresponding nitro groups afforded diamine 15. Con-
densation of 15 with aldehyde 5 under pyrosulfite con-
ditions and subsequent saponification of the ethyl ester
with sodium hydroxide provided acid 16. Intermediates
11, 13, and 16 were converted with the corresponding
acid or amine to the final amides 9m–p, respectively.
ꢀ
ꢀ
ꢀ
(4.0 A), Trp215 (5.0 A), and Tyr99 (5.0 A).
It is well precedented that a variety of strategies can be
employed to optimize interaction with the S4 pocket.
However, many of the S4 residues in our group and in
competitor groups as well are substituted anilines. For
the design of potent inhibitors based on the chlorothio-
phene benzimidazole scaffold the
L-shaped conforma-
tion together with the anilino piperidone as P4 ligand
observed in 1 or the aniline morpholinone in 2a and 2b
has to be build up. For the ease of synthesis a tethered
amide linkage between the 5(6)-position of the benz-
imidazole and the anilino piperidone was envisaged
(Fig. 3).
2. Chemistry
The starting diaminoaryl compounds bearing an alkyl-
ene carbonyl tether were prepared according to litera-
ture procedures or were commercially available. A
representative synthesis of the P4 modified inhibitors
9a–l and 9q–w is described in Scheme 1, exemplified by
morpholinone 9q. Condensation of diaminophenyl
compound 46 with aldehyde 5 in the presence of pyro-
sulfite provided benzimidazole ester 6.5 Hydrolysis of
this ester with sodium hydroxide afforded benzimidazole
acid 7. Coupling of acid 7 and anilino morpholinone 87
3. Results and discussion
The structure–activity relationships surrounding the
inhibitors incorporating the different linkages and the
modified P4 ligands are outlined in Table 1. Compounds
9a–w were assayed against human fXa, thrombin and
trypsin as previously described.8 Briefly, protease
activity was monitored in vitro using protease-specific
chromogenic substrates. Anti-protease activity of drugs
was calculated from the OD ratio of drug and vehicle
O
O
O
a
S
H2N
H2N
Cl
N
OEt
OEt
H
O
+
N
H
S
6
Cl
5
4
b
O
c
N
N
OH
O
O
S
N
H
N
+
Cl
N
H
7
N
H
S
Cl
H2N
N
O
9q
8
O
Scheme 1. General synthesis of the target compounds 9a–l and 9q–w, illustrated at 9q. Reagents and conditions: (a) NaSO3H, NMP, 110 °C, 2 h;
(b) NaOH, MeOH, rt, 12 h; (c) TBTU, DMF, rt, 12 h.

3766
W. W. K. R. Mederski et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3763–3769
NH2
H2N
H2N
NO2
N
a, b
a, c
d, e
S
N
H
Cl
Cl
11
10
12
N
H2N
H2N
CN
NH2
S
N
H
13
O
CO2H
O2N
O2N
OH
N
CO2Et
H2N
H2N
O
a, f
S
N
H
Cl
15
14
16
Scheme 2. Alternative approaches to some thiophene benzimidazole intermediates. Reagents and conditions: (a) 5, NaSO3H, NMP, 110 °C, 12 h; (b)
SnCl2, EtOH, 90 °C, 24 h; (c) LiAlH4, THF, 0 °C, 2 h; (d) BrCH2CO2Et, Cs2CO3, MeCN, rt, 12 h; (e) Raney-Ni, H2, THF, rt, 4 h; (f) 1 N NaOH,
MeOH, rt, 12 h.
Table 1. Factor Xa binding affinity data of selected 5(6)-substituted (5-halo-thiophen-2-yl)-1H-benzimidazoles
A
R1
N
S
N
H
R
Compound
R
A
R1
IC50 fXa (nM)
1100.0
Me
3a
Cl
––
O
N N
H
HN
HN
9a
9b
Cl
Cl
CH2(C@O)
CH2(C@O)
490.0
530.0
MeO2S
N
O
HN
HN
HN
N
O
9c
9d
Cl
Cl
CH2(C@O)
CH2(C@O)
260.0
230.0
N
O
O
O
N
O
9e
9f
Cl
Cl
CH2(C@O)
CH2(C@O)
180.0
120.0
Me
HN
HN
N
O
C
N
O
O
N
9g
9h
Cl
Cl
CH2(C@O)
CH2(C@O)
170.0
14.0
H2
HN
C
N
H2

W. W. K. R. Mederski et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3763–3769
3767
Table 1 (continued)
Compound
R
A
R1
IC50 fXa (nM)
63.0
9i
Br
CH2(C@O)
HN
N
O
Me
9j
Br
CH2(C@O)
41.0
HN
N
O
O
O
HN
HN
C
N
N
9k
9l
Br
Cl
Cl
Cl
CH2(C@O)
C@O
5.2
1900.0
1200.0
440.0
H2
N
O
H2C
O
N
O
9m
9n
NH(C@O)
CH2NH
O
N
O
HN
N
O
O
9o
9p
Cl
Cl
OCH2(C@O)
OCH2(C@O)
150.0
20.0
Me
Me
HN
HN
N
O
O
O
N
O
9q
9r
Cl
Cl
CH2CH2(C@O)
CH2CH2(C@O)
10.0
8.2
HN
N
O
O
Me
Cl
9s
Br
CH2CH2(C@O)
8.4
HN
HN
N
O
O
O
9t
Cl
Cl
CH2CH2(C@O)
CH2CH2(C@O)
7.0
7.6
N
O
F
F
9u
HN
HN
HN
N
O
C
N
N
9v
Br
Cl
CH2CH2(C@O)
CH2CH2(C@O)
58.0
5.9
H2
9w
N
O
O

3768
W. W. K. R. Mederski et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3763–3769
containing assay. For thrombin and trypsin all com-
pounds displayed IC50>10 lM.
that the L-shape conformation and the P4 ligand is
nearly optimized. The presence of substituents at the P4
phenyl ring increased the activity further. A methyl
group (9r) or a chloro substituent (9t) afforded low
nanomolar inhibitors. In contrast to the acetamide series
the substitution of chlorine with a bromo atom at the
thienyl ring in morpholinone 9s did not change the po-
tency relative to the corresponding 9r. The additional
replacement of the phenyl morpholinone with 1-(4-pyr-
idyl)-piperidinyl provided a decrease in potency by one
order of magnitude. Bromide 9v was found to have a 11-
fold less affinity than the chloride 9k. It is assumed that
this substitution pattern in combination with the elon-
gated propionamide tether caused unfavorable interac-
tions within the S1 and/or S4 pocket. Interestingly,
introduction of a 3-fluoro substituent at the phenyl ring,
which increased the activity in other series,9 showed the
same effect in pyridone 9u and morpholinone 9w.
Fluoride 9w represented the most potent inhibitor
within the propionamide series.
The initial compound 9a prepared in this series con-
tained an acetamide in 5(6)-position of the benzimid-
azole and a methylsulfonyl biphenyl9 as S4 template,
which directly led to a 2-fold increase in potency com-
pared to screening compounds 3a and 3b. To explore P4
modification, the methylsulfonyl phenyl was replaced
with pyrrolidinone (9b), piperidinone (9c), morpho-
linone (9d), and pyridinone (9e), respectively, where the
nitrogen is attached to the benzene ring. As a result, the
six-membered rings 9c–e showed a 2- to 3-fold potency
enhancement over biphenylsulfone 9a, whereas the five-
membered 9b showed no change in potency relative to
the corresponding 9a. Methyl substitution on the phenyl
ring increased factor Xa activity further. The methyl
substituted morpholinone 9f displayed a 2-fold increase
in potency over the unsubstituted morpholinone 9d. It is
assumed that the methyl group forced the morpholinone
ring to twist and adopt a more favorable configuration
within the S4 pocket. Homologation of phenyl mor-
pholinone in 9d generated aminomethyl analogue 9g,
which is marginally more potent. However, replacement
of phenyl morpholinone with 1-(4-pyridyl)-piperidinyl10
resulted in a 16-fold increase in binding affinity (9h)
indicating that this functionality could reside deeper in
the S4 binding site of the enzyme. To investigate the
halogen effect observed with the screening compounds
3a and 3b chlorine was replaced by bromine in some
selected inhibitors. The resulting bromo derivatives 9i,j,
and 9k were 2-fold more active than the corresponding
chloro compounds 9e,f, and 9h, respectively, which led
to our first mono-digit nanomolar inhibitor 9k. It is
obvious that the bromo inhibitors gave a better fit in the
S4 pocket than did the chlorides.
Although this class of compounds displayed good
in vitro properties, physicochemical parameters such as
solubility (<10 lg/mL in pH 7 phosphate buffer) and
oral pharmacokinetic profiles (bioavailability <20% in
rats and dogs) have to be improved.
4. Conclusion
Starting from the X-ray structure of halothiophene
benzimidazole 3a emerging from a screening library, a
series of 5(6)-substituted tethered amides was designed
resulting in the discovery of neutral and achiral factor
Xa inhibitors. Potency depends on the length and nature
of the linker between benzimidazole and P4 residue.
Modifications of these parameters led to the identifica-
tion of low nanomolar inhibitors 9k and 9w. Physico-
chemical and pharmacokinetic optimization of these
prototype compounds may lead to the discovery of
novel anti-coagulation drugs.
Further studies on the P4 residue involved the variation
on the length of the tether between the S1 and S4 scaf-
fold. If the phenyl morpholinone is directly linked as an
amide to the benzimidazole significant decrease of
activity was observed. The activity of the amide 9l is
only in the range of the chloro screening compound 3a.
Reversal of the methylene amide bond in 9m or merely
the amide bond in 9n resulted in decreased fXa inhibi-
tory activity compared to analogue 9d. Both compounds
displayed decreased potency by 5- or 2-fold over 9d,
respectively.
Acknowledgements
We wish to extend our thanks to scientists at Proteros
Biostructures GmbH in Planegg-Martinsried, Germany
for their valuable support concerning the crystallization
work during this project. For their skillful experimental
work, we would like to thank Gabriele Czmok, Chris-
Expansion of the linker by an additional oxygen or
methylene group positively affected the potency. The
incorporation of an oxygen directly bound to the 5(6)-
position of the benzimidazole tended to increase the
activity. Although morpholinone ether 9o had only a
minimal effect on potency enhancement over the corre-
sponding 9d the methyl substituted morpholine ether 9p
increased the potency by 6-fold. The ether linkage in the
central region probably to its rotational flexibility may
allow a more pronounced interaction of the S4 residue
with the lipophilic S4 pocket. However, replacement of
the oxygen linker of inhibitor 9o with a methylene group
afforded analogue 9q, which is more than one order of
magnitude potent compared to the parent 9d indicating
ꢁ ^
tina Heiner, Jerome Jonveaux, Dieter Kux, Gabriele
Mahr, and Hy-May Nguyen.
References and notes
1. (a) Harker, L. A.; Hanson, S. R.; Kelly, A. B. Thromb.
Haemost. 1995, 74, 464; (b) Hara, T.; Yokoyoma, A.;
Tanabe, K.; Ishihara, H.; Iwamoto, M. Thromb. Haemost.
1995, 74, 635.
2. (a) Rai, R.; Sprengeler, P. A.; Elrod, K. C.; Young, W. B.
Curr. Med. Chem. 2001, 2, 101; (b) Maignan, S.; Mikol, V.
Curr. Med. Chem. 2001, 2, 161.

W. W. K. R. Mederski et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3763–3769
3769
3. Dorsch, D.; Mederski, W.; Tsaklakidis, C.; Gleitz, J;
Barnes, C. PCT Int. Appl. WO 0257236; Chem. Abstr.
2002, 137, 125096.
8. Hara, T.; Yokoyama, A.; Ishihara, H.; Yokoyama, Y.;
Nagahara, T.; Iwamoto, M. Thromb. Haemost. 1994, 71,
314.
4. Mederski, W. W. K. R.; Germann, M. Bioorg. Med.
Chem. Lett. 2003, 13, 3715.
9. Pinto, 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. J. Med. Chem. 2001, 44, 566.
10. Masters, J. J.; Franciskovich, J. B.; Tinsley, J. M.;
Campbell, C.; Campbell, J. B.; Craft, T. J.; Froelich, L.
L.; Gifford-Moore, D. S.; Hay, L. A.; Herron, D. K.;
Klimkowski, V. J.; Kurz, K. D.; Metz, J. T.; Ratz, A. M.;
Shuman, R. T.; Smith, G. F.; Smith, T.; Towner, R. D.;
Wiley, M. R.; Wilson, A.; Yee, Y. K. J. Med. Chem. 2000,
43, 2087.
€
5. Jonas, R.; Klockow, M.; Lues, I.; Prucher, H.; Schliep, H.
J.; Wurziger, H. Eur. J. Med. Chem. 1993, 28, 129.
6. Analogously prepared from 3-(4-amino-phenyl)-propionic
acid in four steps according to Mederski, W. W. K. R.;
Osswald, M.; Dorsch, D.; Anzali, S.; Christadler, M.;
Schmitges, C.-J.; Wilm, C. Bioorg. Med. Chem. Lett. 1998,
8, 17.
7. Straub, A.; Lampe, T.; Pernerstorfer, J.; Perzborn, E.;
Pohlmann, J.; Roehrig, S.; Schlemmer, K.-H. PCT Int.
Appl. WO 0300256; Chem. Abstr. 2003, 138, 89797.