Heterocyclic thrombin inhibitors. Part 1: Design and synthesis of amidino-phenoxy quinoline derivatives

Article abstract of DOI:10.1016/S0960-894X(03)00442-6

Amidino-phenoxy quinoline derivatives represent a new class of potent thrombin inhibitors with good selectivity and remarkably low molecular weight (M_W: 335-391). X-ray analyses of thrombin-bound inhibitors revealed that enzyme inhibition is mainly based on hydrophobic interactions.

Full text of DOI:10.1016/S0960-894X(03)00442-6

Bioorganic & Medicinal Chemistry Letters 13 (2003) 2291–2295
Heterocyclic Thrombin Inhibitors. Part 1: Design and Synthesis of
Amidino-Phenoxy Quinoline Derivatives
Uwe J. Ries,a,* Henning W. M. Priepke,a Norbert H. Hauel,a Eric E. J. Haaksma,a
a
Jean M. Stassen,b Wolfgang Wienenb andHerbert Nar
aDepartment of Chemical Research, Boehringer Ingelheim Pharma KG, Birkendorfer Straße 65, D-88397 Biberach/Riß, Germany
bDepartment of Biological Research, Boehringer Ingelheim Pharma KG, Birkendorfer Straße 65, D-88397 Biberach/Riß, Germany
Received16 January 2003; revised9 April 2003; accepted24 April 2003
Abstract—Amidino-phenoxy quinoline derivatives represent a new class of potent thrombin inhibitors with good selectivity and
remarkably low molecular weight (MW: 335–391). X-ray analyses of thrombin-boundinhibitors revealedthat enzyme inhibition is
mainly basedon hydrophobic interactions.
# 2003 Publishedby Elsevier Ltd.
The search for potent andorally active thrombin inhib-
itors for antithrombotic therapy has been one of the
main challenges in medicinal chemistry for more than a
decade. In the past synthetic efforts have been con-
centratedon tripeptide analogues (transition state ana-
logues) and arginine amide derivatives.1 The early
inhibitor types possessedreactive functional groups
such as aldehydes, activated ketones or boronic acids.
These common structural features may be responsible
for several problems like lack of selectivity, side effects,
rapidclearance or low oral bioavailability. In recent
years new types of non-peptidic thrombin inhibitors
have been described having heterocyclic core structures,2
which are expectedto show advantages concerning the
requiredproperties in vitro andin vivo.
we present crystal structures of two of these new
inhibitors boundto thrombin.
The short andefficient syntheses of amidinophenoxy
quinoline derivatives are outlined in Scheme 2. 2-
Chloro-4-methyl-7-bromo quinoline4 was reactedwith
4-hydroxy-benzonitrile at 160 ꢀC to yieldthe cyanophe-
noxy analogue 6. This central intermediate was used for
Suzuki-type couplings5 yielding substituted 7-phenyl-
quinoline derivatives like 3a. Similarly, 6 was reacted
with (E)-ethyl crotonate under Heck-conditions6 to
affordthe cinnamic acidderivatives (e.g.,
4a). Cyclo-
pentyl-substitutedquinolines like 5a were preparedby
the following sequence: Palladium-catalyzed coupling of
6 with tributyl allyltin gave the corresponding allyl
derivative, which was oxidized with RuCl3/NaIO4. The
resulting phenylacetic acidderivative was convertedinto
the corresponding ethyl ester and dialkylated. The
cyano groups in all intermediates were reacted under
Recently, we described the design and synthesis of novel
benzimidazole derivatives as potent, direct thrombin
inhibitors with favorable pharmacokinetics andoral
activity.3 Basedon the leadstructure 1 we designed a
number of alternative heterocyclic analogues, from
which the key compound, quinoline derivative 2a, was
discovered (Scheme 1). Surprisingly 2a showeda 22-fold
higher affinity for thrombin comparedto the benzimi-
dazole lead 1, although it lacks the important inter-
action of the phenyl sulfonamide moiety with the
S4-subsite of the enzyme. Herein we wish to report our
efforts to optimize the in vitro andin vivo potency, and
*Corresponding author. Tel.: +49-7351-544535; fax: +49-7351-
547169; e-mail: uwe.ries@bc.boehringer-ingelheim.com
Scheme 1. Benzimidazole and quinoline lead structures.
0960-894X/03/$ - see front matter # 2003 Publishedby Elsevier Ltd.
doi:10.1016/S0960-894X(03)00442-6

2292
U. J. Ries et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2291–2295
Scheme 2. (a) 4-Hydroxybenzonitrile, DMF, 160 ꢀC, 1 h, 74%. (b) 2-methyl-phenylboronic acid, Pd(PPh3)4, Na2CO3, toluene, 110 ꢀC, 6 h, 91%. (c)
(E)-ethylcrotonate, Pd(OAc)2, tris-(o-tolyl)phosphine, TEA, xylene, 32 h, 150 ꢀC, 23%. (d) (1) HCl (g), EtOH, 25 ꢀC; (2) (NH4)2CO3, EtOH, 12 h,
25 ꢀC, 60–95%. (e) allyltributyltin, Pd(PPh3)4, toluene, 15 h, 120 ꢀC, 96%. (f) RuCl3, NaIO4, H2O/CH2Cl2/CH3CN, 24 h, 25 ꢀC, 40%. (g) carbonyl-
diimidazole, THF, EtOH, 2 h, 25 ꢀC, 39%. (h) Br(CH2)4Br, NaH, DMSO, 1 h, 40 ꢀC, 39%.
ꢀ
water from the active site. The o-tolyl ring is turned72
Pinner conditions7 yielding the corresponding amidines.
Compounds 2b and 2c were preparedon identical routes
using 4-amino-benzonitrile or 4-mercapto-benzonitrile
in the first step of the reaction sequence.
relative to the quinoline plane andthe o-methyl group
points into the distal pocket. Thereby, a favorable
binding geometry is adopted. The o-methyl moiety is
able to only partially occupy the S4-subsite.
The starting point for the optimization of in vitro
potency was the leadcompound 2a (Scheme 1). Mole-
cular modeling studies of 2a indicated that the amidino
group interacts with the specificity pocket (S1 subsite)
via a salt bridge with Asp 189 and the 4-methyl group
points into the proximal hydrophobic pocket (S2). As
expected, the oxygen atom between the quinoline and
benzamidine moieties could be replaced by nitrogen or
sulfur with little effect on inhibitory potency (2b/2c vs
2a). In the proposed binding mode the inhibitor cannot
reach the distal pocket (S4 subsite) due to its small size.
In contrast to these compounds nearly all thrombin
inhibitors of comparable potency reportedin the past
contain an additional lipophilic group which con-
tributes to the binding energy via an interaction with the
S4 subsite.8 However, a first report describing nanomo-
Surprisingly, the p-tolyl analogue 3b also turnedout to
be a highly potent thrombin inhibitor, although it
shouldnot be able to interact with the S4 subsite.
Modeling studies based on the 3c-co-crystal structure
suggests that the p-tolyl group in 3b shouldnot occupy
the S4 subsite but shouldinteract with a lipophilic
groove in front of the distal pocket consisting of the
aliphatic portions of the side chains of Ile 124 and Glu
217. Obviously the interaction of 3b with this binding
area8b makes an important contribution to the overall
binding energy.
These findings encouraged us to design additional
quinoline derivatives containing small substituents able
to interact only with this lipophilic groove. Introduction
of an ethyl 2-methylcrotonate residue led to an equally
potent inhibitor (4a). In contrast, the corresponding
lar thrombin inhibitors lacking a group to fill the S4
9
pocket has been publishedrecently.
In order to
improve enzyme inhibition we designed substituents for
the quinoline position 7, which shouldplace a lipophilic
residue into the distal pocket. In addition, these sub-
stituents shouldalso contain a polar group interacting
with the water phase in order to allow modifications of
the inhibitor polarity, which has a strong influence on
important parameters like solubility, protein binding
10
andfree inhibitor concentration in plasma.
For the first step two different tolyl substituents were
introduced at the quinoline position 7 resulting in 11-
and4-foldimprovement of enzyme inhibition, respec-
tively, (3a/3b, Table 1). The X-ray structure of the
thrombin/3c complex, the amino analogue of 3a (Fig. 1),
confirmed the anticipated binding mode of the proximal
half of the inhibitor.11 Apart from the salt bridge
between the amidino group and Asp 189 no polar inter-
action between the inhibitor andthe enzyme is observed.
Therefore the high binding affinity is probably mainly
basedon lipophilic interactions andidsplacement of
Figure 1. X-ray structure of the thrombin/3c complex.

U. J. Ries et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2291–2295
Table 1. Thrombin inhibition of quinoline derivative
2293
CompdR
X
Thrombin inhibition
Trypsin inhibitiona
Plasmin inhibitiona
IC50, mM
IC50, mM
IC50, mM
3a
3b
3c
4a
4b
5a
5b
5c
5d
O
O
0.006
2.9
8.3
12.2
14.8
6.4
0.018
0.037
0.016
0.6
4.2
NH
O
3.5
6.7
O
8.1
25
O
0.003
0.006
0.01
0.56
2.4
5.8
O
20
O
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
O
0.092
3.8
5e
O
an.d.=not determined.
acid 4b was only weakly active indicating that the
negatively chargedcarboxylate idsturbs the enzyme-
inhibitor interaction.
Several ethyl phenylacetate derivatives (5a–5d) also
turned out to be potent thrombin inhibitors, depending
on the substitution at the a-carbon atom. The dialky-
latedderivatives 5a and 5b showedenzyme inhibition in
the lower nanomolar range, whereas the mono ethyl
derivative 5c was slightly less active. A comparison with
the unsubstitutedethyl phenylacetate 5d indicates that
the alkyl groups in 5a–5c strongly contribute to the
enzyme inhibition. As in the case of 4a/4b, transforma-
tion of the ester 5d into the corresponding acid 5e ledto
a substantial loss of activity. Most of the quinoline
derivatives listed in Table 1 showedgoodselectivity
versus trypsin andplasmin.
In order to gain insight into the molecular interactions
of 5a with thrombin, we co-crystallizedthis compound
Figure 2. X-ray structure of the thrombin/5a complex.

2294
U. J. Ries et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2291–2295
Table 2. Thrombin inhibition andantithrombotic activity in vitro
CompdR
Thrombin inhibition
aPTT
ED200, mM
LogP
(octanol/water)
IC50, mM
3a
3d
3e
H
0.006
0.010
0.008
7.9
3.3
0.9
>3.5
1.3
0.5
CONHCH2CO2Et
CONHCH2CO2H
with thrombin (Fig. 2). Surprisingly, the cyclopentyl
group of 5a, insteadof fitting into the lipophilic S4-
pocket, interacts with the lipophilic groove in front of
the S4 subsite. As a consequence the ethyl ester points
into the S4 subsite indicating that 5a uses these two
unexpectedinteractions which results in the best
thrombin inhibition within this series of compounds. In
addition, 5a was the only example which also showeda
weak inhibition of factor Xa (IC50=1.5 mM). The
binding mode observed in thrombin also seemed to be
very important for the inhibition of factor Xa and
turned out to be the basis for the design of potent dual
inhibitors of both coagulation enzymes.16
Acknowledgements
We thank Gunter Hebel, Simpert Koßler, Georg Scha-
dler and Angela Schmid for their excellent technical
assistance andDr. Peter Sieger andMonika Wollrath
for the determination of log P values.
References and Notes
1. Wiley, M. R.; Fisher, M. J. Exp. Opin. Ther. Patents 1997,
7, 1265.
2. (a) Sanderson, P. E. J.; Dyer, D. L.; Naylor-Olsen, A. M.;
Vacca, J. P.; Gardell, S. J.; Lewis, S. D.; Lucas, B. J., Jr.; Lyle,
E. A.; Lynch, J. J., Jr.; Mulichak, A. M. Bioorg. Med. Chem.
Lett. 1997, 7, 1497. (b) Sanderson, P. E. J.; Cutrona, K. J.;
Dorsey, B. D.; Dyer, D. L.; McDonough, C. M.; Naylor-
Olsen, A. M.; Chen, I.-W.; Chen, Z.; Cook, J. J.; Gardell, S. J.;
Krueger, J. A.; Lewis, S. D.; Lin, J. H.; Lucas, B. J., Jr.; Lyle,
E. A.; Lynch, J. J., Jr.; Stranieri, M. T.; Vastag, K.; Shafer,
J. A.; Vacca, J. P. Bioorg. Med. Chem. Lett. 1998, 8, 817. (c)
Sanderson, P. E. J.; Lyle, T. A.; Cutrona, K. J.; Dyer, D. L.;
Dorsey, B. D.; McDonough, C. M.; Naylor-Olsen, A. M.;
Chen, I.-W.; Chen, Z.; Cook, J. J.; Cooper, C. M.; Gardell,
S. J.; Hare, T. R.; Krueger, J. A.; Lewis, S. D.; Lin, J. H.;
Lucas, B. J., Jr.; Lyle, E. A.; Lynch, J. J., Jr.; Stranieri, M. T.;
Vastag, K.; Yan, Y.; Shafer, J. A.; Vacca, J. P. J. Med. Chem.
1998, 41, 4466. (d) Tamura, S. Y.; Semple, J. E.; Reiner, J. E.;
Goldman, E. A.; Brunck, T. K.; Lim-Wilby, M. S.; Carpenter,
S. H.; Rote, W. E.; Oldeshute, G. L.; Richard, R. M.; Nutt,
R. F.; Ripka, W. C. Bioorg. Med. Chem. Lett. 1997, 7, 1543.
(e) Wagner, J.; Kallen, J.; Ehrhardt, C.; Evenou, J.-P.;
Wagner, D. J.Med.Chem. 1998, 41, 3664. (f) Iwanowicz, E. J.;
Lau, W. F.; Lin, J.; Roberts, D. G. M.; Seiler, S. M. Bioorg.
Med. Chem. Lett. 1996, 6, 1339.
3. Hauel, N. H.; Nar, H.; Priepke, H. M. W.; Ries, U. J.;
Stassen, J. M.; Wienen, W. J. Med. Chem. 2002, 45, 1757.
4. Davis, S. E.; Rauckman, B. S.; Chan, J. H.; Roth, B. J.
Med. Chem. 1989, 32, 1936.
5. Stanforth, S. P. Tetrahedron 1998, 54, 263.
6. Heck, R. F. Org. React. 1982, 345.
7. Dox, A. W. Org. Synth., Coll. 1941, 1, 5.
8. (a) For a selection of references see: Hilpert, K.; Ack-
ermann, J.; Banner, D. W.; Gast, A.; Gubernator, K.; Had-
vary, P.; Labler, L.; Muller, K.; Schmid, G.; Tschopp, T. B.;
van de Waterbeemd, H. J. Med. Chem. 1994, 37, 3889. (b)
Tucker, T. J.; Lumma, W. C.; Mulichak, A. M.; Chen, Z.;
In order to reduce the high lipophilicity and to increase
the antithrombotic potency in vitro we introduced
additional hydrophilic groups. As indicated by the X-
ray structure of 3c in thrombin (Fig. 1) the C-5 carbon
of the o-tolyl residue points into the water phase and
offers the possibility of introducing polar substituents.
The results of these modifications are shown in Table 2.
As expectedthe introduction of ester or amide moieties
at C-5 hadno influence on enzyme inhibition ( 3d,3e). In
contrast the antithrombotic potency in human plasma
measuredby the activatedpartial thromboplastin time
(aPTT) was strongly affectedandcorrelatedvery well
with lipophilicity (log P octanol/water). The compound
with the highest hydrophilicity (3e) was also the most
potent example within this series. However, similar to
other quinoline derivatives, 3e showedonly short
duration of action after iv-administration to rats.
The quinoline derivatives described in this paper repre-
sent a new class of potent andselective thrombin inhi-
bitors with remarkably low molecular weight (MW: 335–
391). Due to their high lipophilicity these compounds
showedonly moedst antithrombotic activity in vitro
which prevents their therapeutic use. By co-crystal-
lization of two quinoline derivatives with thrombin we
gainednew knowlegde concerning enzyme-inhibitor
interactions, which are the structural basis for the fur-
ther design of improved coagulation inhibitors. The
progress we made with modified heterocyclic dual inhi-
bitors of thrombin andfactor Xa is reportedin the
following paper.16

U. J. Ries et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2291–2295
2295
ꢀ
˚
constants: 71.2, 72.0, 73.9 A, 101.0 ; Data collection statistics:
Naylor-Olson, A. M.; Lewis, S. D.; Lucas, R.; Freidinger,
R. M.; Kuo, L. C. J. Med. Chem. 1997, 40, 830. (c) Sall, D. J.;
Bastian, J. A.; Briggs, S. L.; Buben, J. A.; Chirgadze, M. L.;
Clawson, D. K.; Denney, M. L.; Giera, D. D.; Gifford-Moore,
D. S.; Harper, R. W.; Hauser, K. L.; Klimkowski, V. J.; Kohn,
T. J.; Lin, H.-S.; McCowan, J. R.; Palkowitz, A. D.; Smith,
G. F.; Takeuchi, K.; Thrasher, K. J.; Tinsley, J. M.; Utterback,
B. G.; Yan, S.-C.; Zhang, M. J. Med. Chem. 1997, 40, 3489.
9. Sanderson, P. E. J.; Cutrona, K. J.; Dyer, D. L.; Krueger,
J. A.; Kuo, L. C.; Lewis, S. D.; Lucas, B. J.; Yan, Y. Bioorg.
Med. Chem. Lett. 2003, 13, 161.
10. Tucker, T. J.; Lumma, W. C.; Lewis, S. D.; Gardell, S. J.;
Lucas, B. J.; Sisko, J. T.; Lynch, J. J.; Lyle, E. A.; Baskin,
E. P.; Woltmann, R. F.; Appleby, S. D.; Chen, I.-W.; Dan-
chek, K. B.; Naylor-Olson, A. M.; Yan, Y.; Stauffer, K. J.;
Cook, J. J.; Stranieri, M. T.; Krueger, J. A.; Cooper, C. M.;
Vacca, J. P. J. Med. Chem. 1997, 40, 3693. (b) Tucker, T. J.;
Brady, S. F.; Lumma, W. C.; Lewis, S. D.; Gardell, S. J.;
Naylor-Olson, A. M.; Yan, Y.; Sisko, J. T.; Stauffer, K. J.;
Lucas, B. J.; Lynch, J. J.; Cook, J. J.; Stranieri, M. T.; Hola-
han, M. A.; Lyle, E. A.; Baskin, E. P.; Chen, I.-W.; Danchek,
K. B.; Krueger, J. A.; Cooper, C. M.; Vacca, J. P. J. Med.
Chem. 1998, 41, 3210.
˚
resolution limits 20.0–1.75 A, total observations 73544, unique
reflections 32119, completeness 87.0%,
Refinement statistics: number of atoms in model: protein 2357,
inhibitor 28, water molecules 194, number of reflections (free
R merge 0.042;
set) 31988 (1616), rms deviation ꢀfrom ideal geometry: bond
2
˚
length 0.005A, bondangles 1.24 , temperature factors (A ):
˚
2
2
˚
˚
27.5A , rms bonded Bs 1.6 A , R factor 20.0%, R free 22.4%;
for the thrombin/5a complex: space group C2, cell constants:
ꢀ
˚
70.0, 71.1, 72.6 A, 100.3 Data collection statistics: resolu-
˚
tion limits 20.0–1.80 A, total observations 58452, unique
reflections 28685, completeness 87.6%, R merge 0.040;
Refinement statistics: number of atoms in model: protein
2357, inhibitor 31, water molecules 146, number of reflec-
tions (free set) 28488 (1441), rms deviation from ideal geo-
metry: bondlength 0.005A , bondangles 1.28 ꢀ, temperature
˚
factors (A ): 24.1A , rms bonded Bs 1.6 A , R factor 22.4%,
R free 25.4%).
12. Skrzypczak-Jankun, E.; Carperos, V. E.; Ravichandran,
K. G.; Tulinsky, A.; Westbrook, M.; Maraganore, J. M. J.
Mol. Biol. 1991, 221, 1379.
2
2
2
˚
˚
˚
13. Gewirth, D. The HKL Manual, 3rdEdition; Yale Uni-
versity: New Haven, CT, 1994.
11. Human a-Thrombin was obtainedfrom Enzyme Research
Laboratories (South Bend, IN USA) and crystallized in com-
plex with the exosite binding hirudin fragment residues 55–
65.12 Co-crystals were generatedby soaking crystals with
mother liquor containing 1 mM of inhibitor. Data were col-
lectedon a MAR Research imaging plate (X-Ray Research,
Hamburg, Germany) mountedon a Rigaku RU200 rotating
anode generator and processed and scaled with HKL.13 Model
building and refinement were carried out with MAIN14 and
CNS15; for the thrombin/3c complex: space group C2, cell
14. Burk, D. PhD. Thesis; TU Muenchen: Muenchen, 1992.
15. Brunger, A. T.; Adams, P. D.; Clore, G. M.; DeLano,
W. L.; Gros, P.; Grosse-Kunstleve, R. W.; Jiang, J. S.; Kus-
zewski, J.; Nilges, M.; Pannu, N. S.; Read, R. J.; Rice, L. M.;
Simonson, T.; Warren, G. L. Acta Crystallogr. D Biol. Crys-
tallogr. 1998, 54, 905.
16. Ries, U. J.; Priepke, H. M. W.; Hauel, N. H.; Handschuh,
S.; Mihm, G.; Stassen, J. M.; Wienen, W.; Nar, H. Bioorg.
Med. Chem. Lett. See preceding paper. doi:10.1016/S0960-
894X(03)00443-8.