Selective and orally bioavailable phenylglycine tissue factor/factor VIIa inhibitors

Article abstract of DOI:10.1016/j.bmcl.2005.04.079

We describe the structure-based design and synthesis of highly potent, orally bioavailable tissue factor/factor VIIa inhibitors which interfere with the coagulation cascade by selective inhibition of the extrinsic pathway.

Full text of DOI:10.1016/j.bmcl.2005.04.079

Bioorganic & Medicinal Chemistry Letters 15 (2005) 5344–5352
Selective and orally bioavailable phenylglycine tissue
factor/factor VIIa inhibitors
Katrin Groebke Zbinden,* Ulrike Obst-Sander, Kurt Hilpert,
Holger Kuhne, David W. Banner, Hans-Joachim Bo¨hm, Martin Stahl,
¨
Jean Ackermann, Leo Alig, Lutz Weber, Hans Peter Wessel, Markus A. Riederer,
´
Thomas B. Tschopp and Thierry Lave,
F. Hoffmann—La Roche Ltd., CH-4070 Basel, Switzerland
Received 12 January 2005; revised 25 April 2005; accepted 28 April 2005
Available online 4 October 2005
Abstract—We describe the structure-based design and synthesis of highly potent, orally bioavailable tissue factor/factor VIIa inhib-
itors which interfere with the coagulation cascade by selective inhibition of the extrinsic pathway.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Thromboembolic disorders are among the leading caus-
es for morbidity and mortality in the industrialized
world. Thrombosis of either the venous or the arterial
vascular system may cause pulmonary embolism, myo-
cardial infarction or ischemic stroke.1 The development
of new anticoagulant therapies therefore represents an
important medical need. For a long period of time the
coumarins have been the only orally active anticoagu-
lants interfering with the coagulation cascade which
have been available on the market. They act indirectly,
inhibiting the c-carboxylation in the biosynthesis of
coagulation factors prothrombin, VII, IX, and X. This
mechanism leads to a slow on-set of action. Additional-
ly, the coumarins suffer from substantial food and drug
interaction, and high protein binding,2 rendering it very
difficult to maintain a balanced plasma exposure. There-
fore, careful and regular monitoring of the patient is re-
quired.3 During the last decades, substantial efforts have
been devoted to the search for an orally bioavailable
replacement of the coumarins. Out of numerous re-
search programs, the only compound to have reached
the market is Ximelagatran from AstraZeneca, which
is a double prodrug of the direct thrombin inhibitor
Melagatran.4 Ximelagatran has been launched in France
in 2004 under the tradename Exanta, but so far has not
been accepted by the FDA due to concerns about long-
term liver damage and possible risk of heart attacks.
It is known that excessive inhibition of the coagulation
cascade at the final stages (factor Xa, thrombin) can lead
to bleeding complications.5 An alternative target would
be the inhibition of the tissue factor/factor VIIa (TF/
F.VIIa) complex which is the main trigger of thrombotic
events.6 This complex is part of the extrinsic pathway of
the coagulation cascade and causes the activation of fac-
tors IX and X, ultimately resulting in the generation of
thrombin and the thrombin-mediated conversion of
fibrinogen to fibrin. Proof-of-concept experiments in
animal models from our laboratories7 as well as other
research groups8 demonstrate that specific inhibition
of the TF/F.VIIa complex results in an antithrombotic
effect without enhancing bleeding propensity. These re-
sults suggest that F.VIIa is a very promising target for
a novel anticoagulant, and many research programs in
the pharmaceutical industry aim at the discovery of a
low molecular weight TF/F.VIIa inhibitor.9 Here,
we report the design, synthesis, and profiling of the
first highly potent, orally bioavailable TF/F.VIIa
inhibitors from our laboratories.
Results and Discussion. In a previous publication10, com-
pounds 1–3 (Table 1) were described. They were the first
TF/F.VIIa inhibitors discovered in our laboratories
exhibiting an activity in the single digit nanomolar
range. In addition to their good inhibitory activity,
*
Corresponding author. Tel.: +41 61 688 83 70; fax: +41 61 688 97
48; e-mail: katrin.groebke_zbinden@roche.com
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.04.079

K. Groebke Zbinden et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5344–5352
5345
Table 1. Inhibition of TF/F.VIIa and related serine proteases by low nanomolar compounds 1–3
OBn
OBn
O
O
R
H
N
H
N
α
α
NH
NH
O
O
HO
O
HO
O
1 epimer
H2N
H2N
NH
NH
1
2: R = -SO2-Ph
3: R = -CO-CH2-Ph
Compound
Ki (lM)
Thrombin
F.VIIa
2 · PTa (lM)
F.VIIa
Thrombin
F.Xa
Trypsin
1
2
3
0.004
0.002
0.007
0.035
0.54
6.4
8.0
12.4
19.0
0.04
0.05
0.08
8.8
270
913
4
12
44
a Human citrated plasma is spiked with at least six 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.11
A
B
His57
S2 Pocket
Lys192
Ser195
Asp102
Cα
Cα
Gln143
S3 Pocket
S1 Pocket
Trp215
Asp189
Asp189
Figure 1. Crystal structure (1w7x.pdb) of compound 1 bound to the active site of F.VIIa. (A) Ball-and-stick representation. (B) Surface
representation. Crystals of compound 1 complexed with TF/F.VIIa were originally produced as described.10 Crystals of the complex of 1 with short
form recombinant F.VIIa were produced as described12, except that the enzyme was activated before complex formation. Crystals were measured on
˚
˚
the Swiss Norwegian Beam Lines (SNBL) at the ESRF, Grenoble, to resolutions of 2.5 A and 2.85 A, respectively. The inhibitor binding mode could
be determined in both cases, but the data sets were incomplete and difficult to refine satisfactorily. In the complex with short form recombinant
F.VIIa, the inhibitor binding is disturbed by a crystal contact. Nevertheless, this structure without TF has been deposited in the Protein Data Bank as
1w8b.pdb. Recently, the structure determination in a new crystal form and at high resolution has been repeated for this publication. To this end
crystals of short form recombinant human F.VIIa were prepared as described.13 Crystals of the complex with 1 were frozen and data were measured
14
˚
˚
˚
in-house to 1.8 A resolution and processed with XDS. The unit cell is P41212, with a = b = 95.0 A, and c = 115.85A. The structure was refined,
starting from coordinates of other in-house structures, using Refmac515 and ARP/wARP16, to final overall crystallographic R-factors of 16.9%
(working) and 19.5% (free), with values in the outer shell of 25.8% and 27.7%, respectively. The inhibitor density is clear. Coordinates of this new
short form recombinant human F.VIIa complex have been deposited in the Protein Data Bank as 1w7x.pdb. An unexpected observation is that the
peptide bond between Lys192 and Gly193 is flipped by 180ꢁ relative to its usual ÔactiveÕ conformation into the ÔinactiveÕ conformation shown. The
carbonyl oxygen of Lys192 is hydrogen bonded to the –OH of Ser195 and the –NH of Gly193, which normally donates a hydrogen bond as part of
the oxyanion-hole, now donates a hydrogen bond to the side-chain oxygen of Gln143. Factor VIIa is unique among the related serine proteases in
having a glutamine in position 143 which can turn inwards to stabilize the ÔinactiveÕ conformation. This ÔinactiveÕ conformation is seen both in the
presence and absence of tissue factor for this inhibitor, although in the complex with tissue factor the loop containing Gln143 is poorly ordered. We
speculate that the ÔinactiveÕ conformation is the more stable conformation in the factor VIIa/tissue factor complex and that factor X, the canonical
macromolecular substrate, must expend some binding energy to promote the active conformation. Such an ÔinactiveÕ conformation was described in
the literature17 for the epidermolytic toxin A, a functionally unrelated serine protease, including an estimate of the energy required to flip the peptide.
In our hands, some inhibitor series bind to the ÔinactiveÕ conformation and others to the ÔactiveÕ conformation (unpublished results).

5346
K. Groebke Zbinden et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5344–5352
compounds 2 and 3 showed promising selectivity against
thrombin and factor Xa (F.Xa). The drawback of these
selective TF/F.VIIa inhibitors is their insufficient PK
profile. Additionally, their poor physicochemical behav-
ior and their high molecular weight cause a moderate
functional activity in the PT (prothrombin time) assay
which is a measure of the ability of an inhibitor to pre-
vent clotting via the extrinsic pathway of the coagula-
tion cascade.11 In this report, we describe how we were
able to reduce the molecular weight and the complexity
of the phenylglycine amide TF/F.VIIa inhibitors while
maintaining selectivity and improving plasma activity.
61 nM. However, the selectivity against thrombin was
lost. Activity could be improved by a factor of 2 by
replacement of the methoxy group in the 3 position by
an ethoxy group (compound 5).
Much better selectivity against thrombin was observed
for compounds with two meta substituents in both posi-
tions 3 and 5 as exemplified by compounds 6 and 7–20
(Table 3). This observation can be explained by the un-
ique positioning and shape of the small S2 pocket of
F.VIIa. In all X-ray structures of complexes between
F.VIIa and phenylglycine as well as phenylglycine amide
inhibitors (unpublished results), the S2 pocket is occu-
pied by a small substituent in the position meta to Ca
of the phenylglycine moiety (Fig. 2B). In thrombin,
the S2 pocket is lined by an insertion loop made up of
Tyr60A-Pro60B-Pro60C-Trp60D. It is located in a
slightly different region of the active site and looks com-
pletely different from the S2 pocket of F.VIIa (Fig. 2C).
The X-ray crystal structure of compound 1 bound to the
active site of TF/F.VIIa is shown in Figure 1. The ami-
dine forms the expected salt bridge with the carboxylate
of Asp189 at the bottom of the S1 pocket. A hydrogen
bond is formed between the aniline NH and the hydrox-
yl group of the active site serine (Ser195). The methoxy
group meta to Ca of the phenylglycine moiety is located
in the small S2 pocket of F.VIIa. The carboxylate is
located in close proximity of the side chain of Lys192
and might thus be engaged in a favorable electrostatic
interaction, although the end of this side chain is quite
disordered in most X-ray crystal structures of complexes
between F.VIIa and inhibitors. Notably a ÔflippedÕ con-
formation of the peptide bond between Lys192 and
Gly193 is observed (see Fig. 1A).
In the case of F.VIIa, compounds 1–20 most likely
accommodate their respective 3-substituents in the S2
pocket. In the active site of thrombin, the 3,4-dialk-
oxy-substituted phenyl residue of compounds 4 and 5
can be rotated by 180ꢁ around the bond between Ca
and the phenyl glycine phenyl ring, in such a way that
the alkoxy group in position 3 is turned out of the active
site and thus points toward the solvent. This explains the
good inhibitory activity of these two derivatives toward
thrombin. In the case of the 3,5-disubstituted phenylgly-
cine derivatives 6–20, the accommodation of the 3,5-
dialkoxyphenyl residue in the active site of thrombin
by means of such a rotation is no longer possible be-
cause two meta substituents are present, thus explaining
their selectivity against thrombin and F.Xa which lacks
the small S2 pocket completely.
From examination of the X-ray structure it became
obvious that the carboxylate could also be attached
directly to the central benzylic position (Ca of the central
phenylglycine moiety) and still contribute to the binding
affinity by an electrostatic interaction with the quite flex-
ible lysine side chain.
This hypothesis was confirmed by the preparation of the
phenylglycine derivative 4 (Table 2). The replacement
of the phenylglycine amide side chain by a simple
carboxylate still resulted in a binding affinity of
There is not much room for optimization of the meta
substituent in position 3, occupying the small S2 pocket.
Besides the ethoxy group, methoxy, methyl, ethyl, vinyl,
Table 2. Inhibition of TF/F.VIIa and related serine proteases by phenylglycine derivatives 4–6
OBn
OBn
4
4
O
O
O
O
3
3
3
5
α
α
NH
α
NH
HO
HO
HO
NH
O
O
O
H2N
NH
H2N
NH
H2N
NH
4
5
6
Compound
Ki (lM)
Thrombin
F.VIIa
2 · PTa (lM)
F.VIIa
Thrombin
F.Xa
Trypsin
4
5
6
0.061
0.028
0.19
0.074
0.085
>35
0.73
0.79
5.6
0.31
1.0
1.2
3.0
11.9
8.8
16.9
184
9.1

K. Groebke Zbinden et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5344–5352
Table 3. Inhibition of TF/F.VIIa by meta,meta-disubstituted derivatives 7–20
5347
R3
R5
HO
NH
O
H2N
NH
Compound R3
R5
Ki (lM)
F.VIIa Thrombin F.Xa Trypsin F.VIIa
Thrombin 2 · PT (lM) 2 · aPTTa (lM) aPTT/PT
O
7
8
EtO
0.14
>6.6
1.6
2.9
>7.5
13.3
2.9
7.0
4.8
8.7
1.9
49.1
27.2
4.8
4.8
5.9
3.4
19.8
15.6
23.7
14.3
4.1
3.3
4.0
4.2
O
O
O
O
O
O
O
Ethyl
Vinyl
0.060
0.069
0.038
9
2.9
41.4
10
4.7
123.7
O
O
11
EtO
0.11
5.5
12.2
4.0
50.1
5.4
29.4
5.4
OH
S
12
13
EtO
EtO
0.18
0.48
>3.5
9.9
4.4
1.6
3.5
8.6
20.1
20.1
4.8
4.6
26.9
22.3
5.6
4.8
N
N
O
O
O
O
O
14
15
16
17
18
EtO
EtO
EtO
0.33
8.6
9.5
2.3
2.6
7.3
5.4
9.9
6.6
4.5
25.8
211.8
26.2
4.5
4.3
8.6
9.1
2.7
17.2
12.9
51.7
34.0
10.3
3.8
3.0
6.0
3.7
3.8
O
O
0.065
0.43
R
R
O
O
O
11.3
8.8
9.1
O
O
O
0.129
0.12
15.3
1.5
68.1
s
4.6
40.2
R enantiomer at Ca
19
20
EtO
EtO
0.13
0.51
7.3
1.4
2.2
>68
57.1
20.2
35.6
11.8
69.9
57.6
2.0
4.9
10.2
8.0
N
a Human citrated plasma is spiked with at least six concentrations of the inhibitor. Clotting via the intrinsic pathway is initiated by the addition of
Actinꢂ FS (ellagic acid in soy phosphatides). Clotting time is determined by a turbidity measurement. The concentration of inhibitor necessary to
double control clotting time (EC50) is determined by fitting the data to an exponential regression.11
ethinyl, propyl, cyclopropyl, and 2-propenyl are allowed
in this position as demonstrated by comparing a series
of compounds with a tetrahydropyranyloxy substituent
in position 5 (Table 3, entries 7–10, and unpublished re-
sults). Regarding in vitro and plasma potency (2 · PT)
the best compound out of this series is the ethinyl
derivative 10. Additionally it is characterized by a good
selectivity against thrombin and F.Xa. Consequently,
the intrinsic pathway is inhibited to a much lesser extent
than the extrinsic pathway as can be seen by the quite
favorable aPTT/PT ratio. The aPTT (2 · prolongation
of activated partial thromboplastin time) serves as a
measure for the ability of the inhibitor to interfere with
the coagulation cascade via the intrinsic pathway.11
Larger moieties in position 3 lead to a sharp drop in
affinity (unpublished data). According to modeling and

5348
K. Groebke Zbinden et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5344–5352
S2 Pocket
A
B
C
His57
S2 Pocket
Ser195
Asp102
S3 Pocket
Trp215
Asp189
S1 Pocket
S1 Pocket
Figure 2. (A) Compound 18 docked in the active site of F.VIIa (magenta: relevant amino acid residues of enzyme). (B) The same representation with
the surface of the active site of F.VIIa shown in magenta. (C) Superposition of compound 18 in the F.VIIa binding conformation with the active site
of thrombin. The meta ethoxy group cannot be accommodated as in F.VIIa because the S2 pocket is different in thrombin.
X-ray structural analysis of selected derivatives (unpub-
lished data), the meta substituent in position 5 is located
in the shallow, solvent-exposed S3 pocket allowing the
introduction of a large variety of larger substituents.
Compounds 11–13 all featuring six-membered heterocy-
clic rings are characterized by good PT potencies. Com-
pound 13 which has the poorest in vitro affinity of these
compounds is, surprisingly, the most active compound
in plasma. This might be attributed to its additional
positive charge which enhances its hydrophilicity.
activity of S-compounds 14 and 15 might be attributed
to a residual F.Xa inhibitory activity. Separation of
compound 14 into its two epimers led to the most active
derivative found within this series. Compound 18 exhib-
its a plasma activity of 2.7 lM. In all X-ray structures
generated from compounds of the phenylglycine series
(unpublished results), it is always the stereoisomer with
configuration R at Ca of the phenylglycine motive which
is found in the active site of F.VIIa. From analogy it can
be assumed that compound 18 also is most likely the epi-
mer with configuration R at Ca.
Going from six-membered to five-membered rings, the
tetrahydrofuranyloxy substituent with configuration S
is found to be optimal (compound 14), while derivative
16 with an R-configured tetrahydrofuranyloxy substitu-
ent is much less active in plasma. The same tendency
was observed for the two derivatives 15 and 17 bearing
an ethinyl substituent in position 3. The better plasma
Good in vitro affinity was observed for the biaryl deriv-
ative 19. It suffered, however, from poor plasma activity.
Replacing the phenyl ring in position 5 by a pyridyl moi-
ety led to a substantial loss in activity, while at the same
time the plasma activity was restored (compound 20).
This clearly shows that not only the in vitro affinity,
Table 4. Pharmacokinetic parameters (rat) of parent compounds 6 and 14 as well as of the corresponding prodrugs 21–23
O
R5
O
R'
NH
O
H2N
N
R''
Compound
R5
R0
R00
Dose (mg/kg)
T1/2 iv (h)
T1/2 po (h)
Cl (ml/min/kg)
Vdss (l/kg)
F (%)
6
21
22
EtO
EtO
EtO
H
H
Et
H
6.1
10.3
10
2.9
—
—
2.6
—
—
0.3
—
—
2
30
OH
OH
5.0
4.8
3.5
100
s
O
14
23
H
H
3
1.4
0.9
—
5.7
—
0.5
—
—
15
O
O
O
Et
OH
10
4.1

K. Groebke Zbinden et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5344–5352
5349
R4
but to a large extent also the hydrophilicity influences
R3
R5
the functional activity of a TF/VIIa inhibitor in plasma.
The same effect has been observed by other groups
working in the field of thrombin and factor Xa
inhibition.18
O
R4
NH
R3
R5
N
The zwitterionic character of the phenylglycine deriva-
tives renders them very polar. Not surprisingly, the oral
bioavailability of 6 amounted to only 2% in the rat (Ta-
ble 4). When the strongly basic benzamidine group was
masked as amidoxime prodrug (compound 21), the oral
bioavailability increased to 30%, while the amidoxime/
ester double prodrug 22 displayed an oral bioavailabil-
ity of 100%. Despite this excellent bioavailability, a
very high dose would be required to cover the plasma
exposure needed for sufficient PT prolongation because
the activity of the parent compound 6 is relatively low
(2 · PT = 9.1 lM). Therefore, the PK profiles of the
more active tetrahydrofuranyloxy derivative 14 and its
amidoxime/ester double prodrug 23 were determined.
The bioavailability of 23 in rats was found to be
15%, with a good half-life of about 4 h. Conversion effi-
ciency to parent compound after iv dosing of double
prodrug was found to be about 50% for both 22 and
23. Double prodrug 22 has a bioavailability which is
higher than its conversion efficiency observed after
intravenous administration of the double prodrug. This
indicates very good absorption of the double prodrug
and suggests additional conversion of the prodrug dur-
ing intestinal absorption. In contrast, only partial
absorption seems to occur in the case of 23 leading to
a bioavailability about 3-fold lower than conversion
efficiency. Therefore, dose adjustment and formulation
optimization are needed for compound 23 to maintain
plasma concentration for sustained inhibition of
coagulation activity and are currently being further
evaluated.
a
O
H
N
24
b
R4
R4
R4
R3
R5
R3
R5
R3
R5
HO
d,e
c
O
O
NH
NH
NH
O
O
O
H2N
NH
H2N
N
N
OH
4: R3 = OMe, R4 = OBn, R5 = H
5: R3 = OEt, R4 = OBn, R5 = H
6: R3 =R5 = OEt, R4 = H
26
25
Scheme 1. Preparation of phenylglycine derivatives 4–6. Reagents and
conditions: (a) i—4-aminobenzonitrile, MeOH; ii—benzylisonitrile,
BF3-OEt2, 0 ꢁC; (b) 20 equiv H2O, 52–80%; (c) H2N-OH HCl,
triethylamine, EtOH, reflux, 66–82%; (d) H2, Raney nickel, EtOH/
H2O/HOAc, 56–86%; (e) LiOH, THF, 80–98%.
nickel and subsequent ester hydrolysis finally provided
the zwitterionic phenylglycines 4–6.
The synthesis of the 3-ethoxy-substituted compounds 7,
11–14, 16, and 18 started with the monoethylation of 5-
hydroxymethyl-benzene-1,3-diol, followed by a benzyla-
tion of the remaining phenol and oxidation of the
hydroxymethyl group to give benzaldehyde 27, which
was condensed with 4-aminobenzonitrile and toluene-
4-sulfonylisocyanide to furnish phenylglycine ester 28
(Scheme 2). After removal of the benzyl group, phenol
29 was derivatized by a Mitsunobu reaction with various
alcohols. Conversion of the nitrile into the respective
amidine by a Pinner reaction and hydrolysis of the ester
functionality lead to the zwitterionic phenylglycine
derivatives 7, 11, 13, 14, and 16. The (S)-tetrahydrofura-
nyloxy-substituted derivative 16 was separated into its
epimers by chiral HPLC. Compound 12 was prepared
starting by derivatization of phenol 29 with 4-hydroxy-
piperidine-1-carboxylic acid tert-butyl ester. After cleav-
age of the Boc-protecting group and reaction of the
piperidine with methane sulfonylchloride, the nitrile
was converted into the amidine by a Pinner reaction.
Hydrolysis of the ester led to the desired phenylglycine
product 12.
In summary, we have found novel small molecule factor
VIIa inhibitors with good inhibitory activity against fac-
tor VIIa and decent in vitro selectivity against other ser-
ine proteases of the coagulation cascade and against
trypsin. The best derivatives displayed low micromolar
functional activity in inhibiting the coagulation cascade
via the extrinsic pathway (2 · PT prolongation), while
interfering with the intrinsic pathway to a much lesser
extent (2 · aPTT prolongation). The pharmacokinetic
profile of this compound class is characterized by a
promising oral bioavailability and a long half-life. Fur-
ther optimization with the goal to reach sufficient
plasma exposure and antithrombotic activity in
relevant animal models will be the subject of further
publications.
Chemistry. Phenylglycine derivatives 4–6 were synthe-
sized starting by a Lewis acid-catalyzed condensation
between an appropriately substituted benzaldehyde, 4-
aminobenzonitrile, and benzylisonitrile (Scheme 1).
The intermediate iminoether 24 was hydrolyzed in situ
to the corresponding phenylglycine ester 25 by the addi-
tion of an excess of water. By reaction with hydroxyl-
amine hydrochloride compound 25 was converted to
the corresponding amidoxime 26. Reduction with Raney
The 3-ethinyl-substituted derivatives 10, 15, and 17 were
prepared starting with the reaction of 3,5-dihydroxy-
benzaldehyde with trifluoromethanesulfonic anhydride

5350
K. Groebke Zbinden et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5344–5352
HO
OH TfO
OTf
HO
O
OBn
b, c
a
O
H
O
H
O
H
d
O
NH
31
32
HO
O
OBn
OH
a, b, c
d
O
HO
O
H
O
OH
O
R
R
e
N
27
28
f, g
e
O
O
HO
NH
NH
NH
O
O
O
O
O
O
OH
O
O
R
R
O
HO
HO
NH
i
f, g, h
NH
NH
H2N
NH
N
N
O
O
O
10, 15, 17
34
33
Scheme 3. Preparation of 3-ethinyl-substituted derivatives 10, 15, and
17. Reagents and conditions: (a) trifluoromethanesulfonic anhydride,
TEA, CH2Cl2, 0 ꢁC, 88%; (b) trimethylsilylacetylene, Pd(PPh3)2Cl2,
Cu(I)I, TEA, THF, 70%; (c) 1 N LiOH, MeOH, 66%; (d) i—4-
aminobenzonitrile, MeOH; ii—toluene-4-sulfonylisocyanide, BF3-
OEt2, 0 ꢁC; iii—20 equiv H2O, 65%; (e) R0-OH, PPh3, DEAD, THF,
62–85%; 10: R = R0; 15: R0-OH = (R)-3-hydroxy-tetrahydrofurane; 17:
R0-OH = (S)-3-hydroxy-tetrahydrofurane (Derivatives 15 and 17:
Mitsunobu reaction with the respective alcohols R0-OH proceeds with
inversion of stereochemistry.) (f) HCl-gas, CHCl3–MeOH (3:1),
ꢀ10 ꢁC, evaporate, then 2 N NH3 in MeOH, 41–50%; (g) LiOH,
THF, 60–88%.
H2N
NH
H2N
NH
N
18: R = (S)-3-tetra-
7,11, 13, 14, 16
29
f, j
hydrofuranyl
O
O
O
O
O
O
N
NH
S
O
HO
k, g, h
NH
NH
O
O
starting by ethylene glycol protection of 3,5-dibromo-
benzaldehyde, followed by a mono-Stille coupling with
tributyl(vinyl)tin and subsequent conversion of the
remaining bromo substituent into a hydroxy group by
H2N
NH
N
12
30
Scheme 2. Preparation of 3-ethoxy-substituted phenylglycine deriva-
tives 7, 11–14, 16, and 18. Reagents and conditions: (a) ethylbromide,
K2CO3, DMF, 60 ꢁC, 30%; (b) benzylbromide, K2CO3, DMF, 65 ꢁC,
83%; (c) MnO2, 1,2-dichloroethane, 50 ꢁC, 81%; (d) i—4-aminobenzo-
nitrile, MeOH; ii—toluene-4-sulfonylisocyanide, BF3-OEt2, 0 ꢁC; iii—
20 equiv H2O, 63%; (e) H2, Pd/C, THF, EtOH, 70%; (f) R00-OH, PPh3,
DEAD, THF, 63–68%; 7 and 13: R00 = R0 = R; 14 and 16: R0 = R; 11:
R00-OH = cis-4-trityloxy-cyclohexanol; 14: R00-OH = (R)-3-hydroxy-
tetrahydrofurane; 16: R00-OH = (S)-3-hydroxy-tetrahydrofurane; 30:
O
Br
Br
OH
a, b, c
d, e
O
35
36
O
H
O
H
O
O
R00-OH = 4-hydroxy-piperidine-1-carboxylic
acid
tert-butylester
f
(Derivative 11: The trityl group of R00-OH is lost during the Pinner
reaction. Derivatives 11, 14, 16: Mitsunobu reaction with the respective
alcohols R00-OH proceeds with inversion of stereochemistry.) (g) HCl-
gas, CHCl3–MeOH (3:1), ꢀ10 ꢁC, evaporate, then 2 N NH3 in MeOH,
45–92%; (h) LiOH, THF, 62–97%; (i) separation of 16 by chiral HPLC,
31%; (j) 4 N HCl in EtOAc, 94%; (k) CH3-SO2-Cl, DIPEA, THF, 44%.
O
O
O
O
O
O
HO
HO
O
NH
i
g, h
NH
NH
O
O
O
(Scheme 3). Mono-Sonogashira coupling and cleavage
of the remaining triflate lead to intermediate 32 which
was condensed in a Lewis acid-catalyzed three-compo-
nent reaction with 4-aminobenzonitrile and toluene-4-
sulfonylisocyanide to give phenylglycine ester 33. The
phenol was derivatized with various alcohols via a Mits-
unobu reaction. Finally, the nitrile was converted into
the corresponding amidine by a Pinner reaction. The
final products 10, 15, and 17 were generated by hydroly-
sis of the corresponding phenylglycine esters.
H2N
NH
H2N
NH
N
8
9
37
Scheme 4. Preparation of 3-ethyl-substituted derivative 8 and 3-vinyl-
substituted 9. Reagents and conditions: (a) p-TsOH, ethylene glycol,
benzene, reflux, 98%; (b) tributyl(vinyl)tin, tetrakis(triphenylphos-
phin)palladium(0), toluene, 80 ꢁC, 64%; (c) n-butyllithium, trimeth-
ylborate, THF, ꢀ78 ꢁC, then acetic acid, H2O2, 68%; (d) 4-hydroxy-
tetrahydropyrane, PPh3, DEAD, THF, 69%; (e) 2 N HCl, THF, 60%;
(f) i—4-aminobenzonitrile, MeOH; ii—toluene-4-sulfonylisocyanide,
BF3-OEt2, 0 ꢁC; iii—20 equiv H2O, 66%; (g) HCl-gas, CHCl3–MeOH
(3:1), ꢀ10 ꢁC; evaporate, then 2 N NH3 in MeOH, 86%; (h) LiOH,
THF, 74%; (i) H2, Pd/C, EtOH/H2O, 66%.
The vinyl-substituted phenylglycine derivative 8 and
the corresponding ethyl compound 9 were obtained

K. Groebke Zbinden et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5344–5352
5351
lithiation, reaction with trimethylborate, and subse-
quent oxidation with hydrogen peroxide (Scheme 4).
Intermediate 35 was reacted in a Mitsunobu reaction
with 4-hydroxy-tetrahydropyrane and then deprotected.
The resulting aldehyde 36 was converted in a Lewis acid-
catalyzed condensation with 4-aminobenzonitrile and
toluene-4-sulfonylisocyanide to provide phenylglycine
ester 37. Conversion of the nitrile into an amidine group
and hydrolysis of the ester moiety led to the phenylgly-
cine compound 8 which was converted into the ethyl
analogue 9 by hydrogenation of the double bond.
The 5-aryl-substituted compounds 19 and 20 were syn-
thesized starting from phenol 29 (Scheme 2) which after
conversion to triflate 38 was reacted with phenylboronic
acid and 3-(tri-n-butylstanyl)-pyridine (Scheme 5). Con-
version of the nitrile into the amidine via Pinner reaction
and hydrolysis of the phenylglycine ester provided the
desired products.
O
O
O
O
OTf
OH
The amidoxime prodrug 21 was synthesized starting
from the 3,5-diethoxy-substituted phenylglycine ester
40 (prepared as described in Scheme 1) by hydrolysis
of the ester group and subsequent reaction of the nitrile
with hydroxylamine hydrochloride (Scheme 6). The ami-
doxime/ester double prodrugs 22 and 23 were prepared
by reaction of the appropriately substituted phenylgly-
cine esters 40 (prepared as described in Schemes 1 and
2) with hydroxylamine hydrochloride.
a
NH
NH
O
O
N
N
29
38
b
O
Aryl
O
O
Aryl
Acknowledgments
HO
NH
NH
c, d
The authors wish to thank Felix Gruber, Isabelle Guil-
laumat, Heidi Hoffmann-Maiocchi, Roland Keller, Fri-
dolin Mehlin, Rudolf E. Minder, Caroline Prassler,
Noemi Raschetti, Bettina Roser, Ernst Schaffter, Patri-
cia von Arx, and Jitka Weber for their skilful technical
O
O
H2N
NH
H2N
NH
19, 20
39
´
assistance as well as Allan DÕArcy, Andre Alker, Fiona
Gruninger, Daniel Kirchhofer, Dirk Kostrewa, and the
¨
Scheme 5. Preparation of biaryl derivatives 19 and 20. Reagents and
conditions: (a) trifluoromethanesulfonic anhydride, TEA, dichloro-
methane, 0 ꢁC, 99%; (b) 19: phenylboronic acid, K2CO3, tetrakis(tri-
phenylphosphine)palladium(0), toluene, 90 ꢁC, 93%; 20: 3-(tri-n-
butylstannyl)-pyridine, triphenylarsine, Pd2(dba)3-CHCl3, LiCl,
NMP, 80 ꢁC, 60%; (c) HCl-gas, CHCl3–MeOH (3:1), ꢀ10 ꢁC, evap-
orate, then 2 N NH3 in MeOH, 77–87%; (d) LiOH, THF, 90–94%.
staff of the Swiss Norwegian Beam Lines at the ESRF
for their respective contributions to protein production,
crystal growth, and data collection.
References and notes
O
O
1. (a) Murray, C. J.; Lopez, A. D. Lancet 1997, 349, 1269; (b)
Braunwald, E.; Califf, R. M.; Cannon, C. P.; Fox, K. A.;
Fuster, V.; Gibler, W. B.; Harrington, R. A.; King, S. B.;
Kleimann, N. S.; Theroux, P.; Topol, E. J.; Van de Werf,
F.; White, H. D.; Willerson, J. T. Am. J. Med. 2000, 108,
41.
O
O
O
R
R
HO
a, b
NH
NH
O
O
2. Wells, P. S.; Holbrook, A. M.; Crowther, N. R.; Hirsh, J.
Ann. Int. Med. 1994, 121, 676.
H2N
N
3. (a) Hyers, T. M.; Agnelli, G.; Hull, R. D.; Morris, T. A.;
Samama, M.; Tapson, V.; Weg, J. G. Chest 2001, 119,
176S; (b) Landefeld, C. S.; Rosenblatt, M. W.; Goldman,
L. Am. J. Med. 1989, 87, 153; (c) Hylek, E. M.; Go, A. S.;
Chang, Y.; Jensvold, N. G.; Henault, L. E.; Selby, J. V.;
Singer, D. E. New Engl. J. Med. 2003, 349, 1019.
4. (a) Crowther, M. A.; Weitz, J. I. Expert Opin. Invest.
Drugs 2004, 13, 403; (b) McCall, K. L.; MacLaughlin, E.
J. J. Pharm. Technol. 2003, 19, 222.
N
OH
R = Et: 21
b
40
O
O
R
O
NH
O
5. (a) Eriksson, B. I.; Bergqvist, D.; Dahl, O. E.; Lindbratt,
S.; Bylock, A.; Frison, L.; Eriksson, U. G.; Welin, L.;
Gustafsson, D. Lancet 2002, 360, 1441; (b) Turpie, G. G.
Expert Opin. Pharmacother. 2004, 5, 1373.
R = Et: 22
s
R =
: 23
O
H2N
N
6. (a) Nemerson, Y. Sem. Hematol. 1992, 29, 170; (b)
Edgington, T. S.; Dickinson, C. D.; Ruf, W. Thromb.
Haemost. 1997, 78, 401.
7. (a) Himber, J.; Refino, C. J.; Bucklen, L.; Roux, S.
Thromb. Haemost. 2001, 85, 475; (b) Himber, J.; Kirchho-
OH
Scheme 6. Preparation of prodrugs 21–23. Reagents and conditions:
(a) LiOH, THF, 86%; (b) H2N-OH hydrochloride, TEA, EtOH, reflux,
21: 22%, 22: 81%, 23: 65%.

5352
K. Groebke Zbinden et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5344–5352
fer, D.; Riederer, M.; Tschopp, T. B.; Steiner, B.; Roux, S.
P. Thromb. Haemost. 1997, 78, 1142.
E.; Granberg, K.; Nilsson, I. Bioorg. Med. Chem. Lett.
2002, 12, 2907; (j) Young, W. B.; Kolesnikov, A.; Rai, R.;
Sprengeler, P. A.; Leahy, E. M.; Shrader, W. D.; Sanga-
lang, J.; Burgess-Henry, J.; Spencer, J.; Elrod, K.; Cregar,
L. Bioorg. Med. Chem. Lett. 2001, 11, 2253; (k) Kohrt, J.
T.; Filipski, K. J.; Rapundalo, S. T.; Cody, W. L.;
Edmunds, J. J. Tetrahedron Lett. 2000, 41, 6041; (l)
Jakobsen, P.; Horneman, A. M.; Persson, E. Bioorg. Med.
Chem. 2000, 8, 2803; (m) Jakobsen, P.; Ritsmar Pedersen,
B.; Persson, E. Bioorg. Med. Chem. 2000, 8, 2095; (n)
Roussel, P.; Bradley, M.; Kane, P.; Bailey, C.; Arnold, R.;
Cross, A. Tetrahedron 1999, 55, 6219.
8. (a) Suleymanov, O. D.; Szalony, J. A.; Salyers, A. K.;
LaChance, R. M.; Parlow, J. J.; South, M. S.; Wood, R.
S.; Nicholson, N. S. J. Pharmacol. Exp. Ther. 2003, 306,
1115; (b) Szalony, J. A.; Taite, B. B.; Girard, T. J.;
Nicholson, N. S.; LaChance, R. M. J. Thrombosis
Thrombolysis 2002, 14, 113; (c) Zoldhelyi, P.; McNatt,
J.; Shelat, H. S.; Yamamoto, Y.; Chen, Z.-Q.; Willerson, J.
T. Circulation 2000, 101, 289; (d) Golino, P.; Ragni, M.;
Cirillo, P.; DÕAndrea, D.; Scognamiglio, A.; Ravera, A.;
Buono, C.; Ezban, M.; Corcione, N.; Vigorito, F.;
Condorelli, M.; Chiariello, M. Circ. Res. 1998, 82, 39;
(e) Kelley, R. F.; Refino, C. J.; OÕConnell, M. P.; Modi,
N.; Sehl, P.; Lowe, D.; Pater, C.; Bunting, S. Blood 1997,
89, 3219; (f) De Guzman, L.; Refino, C. J.; Steinmetz, H.;
Bullens, S.; Lipari, T.; Smyth, R.; Eaton, D.; Bunting, S.
Thromb. Haemost. 1997(Supp. 292); (g) Harker, L. A.;
Hanson, S. R.; Wilcox, J. N.; Kelly, A. B. Haemostasis
1996, 26, 76; (h) Pawashe, A. B.; Golino, P.; Ambrosio,
G.; Migliaccio, F.; Ragni, M.; Pascucci, I.; Chiariello, M.;
Bach, R.; Garen, A.; Konigsberg, W. K.; Ezekowitz, M.
D. Circ. Res. 1994, 74, 56.
10. Groebke Zbinden, K.; Banner, D.W.; Ackermann, J.;
Kirchhofer, D.; Ji, Y.-H.; Tschopp, T.B.; Wallbaum, S.;
Weber, L. Bioorg. Med. Chem. Lett., in press.
11. White, G. C.; Marder, V. J.; Coleman, R. W.; Hirsh, J.;
Salzman, E. In Haemostasis and Thrombosis Basic Prin-
ciples and Clinical Practice, 2nd ed., 1987; pp 1048–1060.
12. Eigenbrot, C.; Kirchhofer, D.; Dennis, S. M.; Santell, L.;
Lazarus, R. A.; Stamos, J.; Ultsch, M. H. Structure 2001,
9, 627.
13. Sichler, K.; Banner, D. W.; DÕArcy, A.; Hopfner, K. P.;
Huber, R.; Bode, W.; Kresse, G. B.; Kopetzki, E.;
Brandstetter, H. J. Mol. Biol. 2002, 322, 591.
9. Review: (a) Robinson, L. A.; Saiah, E. M. K. Annu. Rep.
Med. Chem. 2002, 37, 85; (b) Parlow, J. J.; Kurumbail, R.
G.; Stegeman, R. A.; Stevens, M. A.; Stallings, W. C.;
South, M. S. J. Med. Chem. 2003, 46, 4696; (c) Parlow, J.
J.; Case, B. L.; Dice, T. A.; Fenton, R. L.; Hayes, M. J.;
Jones, D. E.; Neumann, W. L.; Wood, R. S.; LaChance,
R. M.; Girard, T. J.; Nicholson, N. S.; Clare, M.;
Stegemann, R. A.; Stevens, A. M.; Stallings, W. C.;
Kurumbail, R. G.; South, M. S. J. Med. Chem. 2003, 46,
4050; (d) Parlow, J. J.; Dice, T. A.; LaChance, R. M.;
Girard, T. J.; Stevens, A. M.; Stegemann, R. A.; Stallings,
W. C.; Kurumbail, R. G.; South, M. S. J. Med. Chem.
2003, 46, 4043; (e) Parlow, J. J.; Stevens, A. M.;
Stegemann, R. A.; Stallings, W. C.; Kurumbail, R. G.;
South, M. S. J. Med. Chem. 2003, 46, 4297; (f) Klingler,
O.; Matter, H.; Schudok, M.; Bajaj, S. P.; Czech, J.;
Lorenz, M.; Nestler, H. P.; Schreuder, H.; Wildgoose, P.
Bioorg. Med. Chem. Lett. 2003, 13, 1463; (g) Carroll, A.
R.; Pierens, G. K.; Fechner, G.; de Almeida Leone, P.;
Ngo, A.; Simpson, M.; Hyde, E.; Hooper, J. N. A.;
Bostro¨m, S.-L.; Musil, D.; Quinn, R. J. J. Am. Chem. Soc.
2002, 124, 13340; (h) Hanessian, S.; Margarita, R.; Hall,
A.; Johnstone, S.; Tremblay, M.; Parlanti, L. J. Am.
Chem. Soc. 2002, 124, 13342; (i) Hanessian, S.; Therrien,
14. Kabsch, W. J. Appl. Cryst. 1993, 26, 795.
15. (a) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Acta
Cryst. 1997, D53, 240; (b). Acta Cryst. 1994, D50, 760.
16. Lamzin, V. S.; Wilson, K. S. Acta Cryst. 1993, D49, 129.
17. Cavarelli, J.; Prevost, G.; Bourguet, W.; Moulinier, L.;
Chevrier, B.; Delagoutte, B.; Bilwes, A.; Mourey, L.;
Rifai, S.; Piemont, Y.; Moras, D. Structure 1997, 5,
813.
18. (a) 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. Bioorg. Med. Chem. Lett. 2004, 14, 1221; (b)
Zhang, P.; Bao, L.; Zuckett, J. F.; Goldman, E. A.; Jia, Z.
J.; Arfsten, A.; Edwards, S.; Sinha, U.; Hutchaleelaha, A.;
Park, G.; Lambing, J. L.; Hollenbach, S. J.; Scarborough,
R. M.; Zhu, B.-Y. Bioorg. Med. Chem. Lett. 2004, 14, 983;
(c) Zhang, P.; Bao, L.; Zuckett, J. F.; Jia, Z. J.; Woolfrey,
J.; Arfsten, A.; Edwards, S.; Sinha, U.; Hutchaleelaha, A.;
Lambing, J. L.; Hollenbach, S. J.; Scarborough, R. M.;
Zhu, B.-Y. Bioorg. Med. Chem. Lett. 2004, 14, 989; (d)
Sanderson, P. E. J.; Cutrona, K. J.; Dyer, D. L.; Krueger,
J. A.; Kuo, L. C.; Lewis, S. D.; Lucas, B. J.; Yang, Y.
Bioorg. Med. Chem. Lett. 2003, 13, 161.