Structure-based design and synthesis of pyrazinones containing novel P 1 'side pocket' moieties as inhibitors of TF/VIIa

Article abstract of DOI:10.1016/j.bmcl.2005.04.037

We describe the structure-based design, synthesis, and enzymatic activity of a series of substituted pyrazinones as inhibitors of the TF/VIIa complex. These inhibitors contain substituents meta to the P₁ amidine designed to explore additional i

Full text of DOI:10.1016/j.bmcl.2005.04.037

Bioorganic & Medicinal Chemistry Letters 15 (2005) 3006–3011
Structure-based design and synthesis of pyrazinones
containing novel P1 Ôside pocketÕ moieties
as inhibitors of TF/VIIa
Barbara A. Schweitzer,a,* William L. Neumann,a Hayat K. Rahman,a Carrie L. Kusturin,a
Kirby R. Sample,a Gennadiy I. Poda,b Ravi G. Kurumbail,b Anna M. Stevens,b
Roderick A. Stegeman,b William C. Stallingsb and Michael S. Southa
aDepartment of Medicinal and Combinatorial Chemistry, Pfizer Corp., 800 N. Lindbergh Blvd., St. Louis, MO 63167, USA
bStructural and Computational Chemistry, Pfizer Corporation, 700 Chesterfield Village Pkwy., St. Louis, MO 63198, USA
Received 7 November 2004; revised 21 April 2005; accepted 22 April 2005
Abstract—We describe the structure-based design, synthesis, and enzymatic activity of a series of substituted pyrazinones as inhib-
itors of the TF/VIIa complex. These inhibitors contain substituents meta to the P1 amidine designed to explore additional interac-
tions with the VIIa residues in the so-called ÔS1 side pocketÕ. A crystal structure of the designed inhibitors demonstrates the ability of
the P1 side pocket moiety to engage Lys192 and main chain of Gly216 via hydrogen bond interactions, thus, providing additional
possibility for chemical modification to improve selectivity and/or physical properties of inhibitors.
Ó 2005 Elsevier Ltd. All rights reserved.
Cardiovascular disease remains the most common cause
of mortality in the western world. The use of standard
therapies such as heparin or warfarin is severely limited
due to slow onset of action and lack of selectivity lead-
ing to bleeding side effects that requires close patient
monitoring.1 The search for safer more efficacious anti-
thrombotics has focused on direct inhibitors targeted to
specific inhibition of enzymes within the extrinsic coagu-
lation cascade—the ultimate goal being discovery of a
drug with an improved safety profile and pharmacody-
namics over existing treatments, specifically, better oral
bioavailability and specificity.
responsible for activation of IX to IXa and X to Xa
through proteolytic cleavage. This ultimately results in
the amplified conversion of prothrombin to thrombin
and fibrinogen to fibrin in the latter stages of clot
formation.
Several approaches have been described in the literature
for antagonizing the activity of TF/VIIa complex. These
include active site inhibited factor VIIa (VIIai), tissue
factor mutants, antibodies directed against TF or VIIa,
naturally occurring protein-based inhibitors (for exam-
ple, tissue factor pathway inhibitor or TFPI, hookworm
anticoagulant protein, and NAPc2), peptide-based exo-
site inhibitors, and small molecule active site
inhibitors.5d
Several enzymes within the coagulation cascade have
been targets for antithrombotic therapies including
thrombin,2 Xa,3 and IXa.4 Recently, TF/VIIa has
emerged as an attractive target.5 Factor VIIa (TF/VIIa)
is the initiator of the extrinsic coagulation cascade. Pla-
que rupture exposes tissue factor (TF) to zymogen VII
in circulating plasma, which eventually forms the active
TF/VIIa complex. The assembled TF/VIIa complex is
Evidence from our laboratories and others indicate that
selective inhibition of the TF/VIIa complex may provide
prevention of thrombosis with the advantage of a lower
risk of bleeding side effects versus other coagulation cas-
cade targets.6
Recent papers from our laboratories, both published
and in progress, describe the design and synthesis of
compounds that inhibit the TF/VIIa complex.7 These
articles discuss optimization of the core structure and
Keywords: Pyrazinone; TFVIIa.
*
Corresponding author. Tel.: +1 (314) 274 8970; fax: +1 (314) 274
1621; e-mail: barbara.a.schweitzer@pfizer.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.04.037

B. A. Schweitzer et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3006–3011
3007
dine also forms good hydrogen bonds with main chain
carbonyl of Gly219, a structural water and a weak
hydrogen bond with the Ser190 side chain.
Analysis of structures also revealed a solvent exposed re-
gion outside of the S1 pocket formed by Gln143,
Asp146, Arg147, Lys192, Gly216, Gly219, and Cys220
(ÔS1 side pocketÕ) that could be reached off of the ortho
position of the P1-needle (Fig. 3). Analysis of crystal
structures of related serine proteases revealed that a sim-
ilar binding pocket exists in thrombin, factor Xa, and
trypsin. However, only VIIa has a lysine at position
192. This is replaced by a glutamine in Xa and trypsin,
while thrombin has a glutamic acid at this position.
The S1 side pocket is very shallow in trypsin while
thrombin has a bulky tryptophan residue that forms
one of the walls of the pocket. Interestingly, the side
pocket in Xa appears to be very similar to that in VIIa.
We believed that exploration of this region could pro-
vide additional specific interactions with VIIa leading
to enhanced potency, selectivity, and improved physical
properties.
Figure 1. Pyrazinone core structure.
subsequent optimization at the P1, P2, and P3 sites of the
inhibitors with ultimate achievement of low nanomolar,
pyrazinone-based inhibitors (Fig. 1). In this article, we
report the design, synthesis, and biological activity of
new inhibitors containing a Ôside pocketÕ residue on P1.
These compounds arose from further optimization of
the P1 moiety based on some of our most potent
analogs.
Our approach to the design of these inhibitors origi-
nated from examination of the crystal structure of exist-
ing active compounds described elsewhere.7d The
common feature of all the structures is the anchoring
interaction of benzamidine via bidentate ionic interac-
tion with conserved Asp189 located deeply in the S1
pocket of VIIa (as shown in Figs. 2 and 3). The benzami-
We began with synthesis of compound 4b, a pyridyl
substituted P1. The pyridyl residue can participate in
hydrogen bonds with Lys192 through the nitrogen of
the pyridine ring as well as through its amide carbonyl.
Preparation of 4b began with compound 18 (Scheme 1).
Protection of the basic nitrogens using di-tert-butyldi-
carbonate gave us a mixture of mono and bis-protected
products, 2a and b.9 We carried both of these forward to
the mono and bis protected versions of compound 3.
This was achieved through activation of the carboxylate
2 using HBTU followed by reaction with 4-pyridyl-
amine. Final deprotection using TFA followed by re-
verse phase HPLC purification afforded us compound
4b as the trifluoroacetate salt. Despite the overall low
Figure 2. Schematic representation of the binding mode of the P1
moiety of inhibitor with TF/VIIa complex.
Figure 3. (a) Crystal structure of compound 4b bound in the active site of TF/VIIa complex. The structure has been refined to an Rfree of 26.8% at
2.0 A resolution (Rcrystal: 22.3%). Some of the key side chains of factor VIIa are displayed (C: gold, N: dark blue, O: red, and S: pink). The carbon
atoms of the inhibitor are shown in dark pink color. The hydrogen bonds formed by the inhibitor are shown in dotted (white) lines. Two bound
solvent molecules are shown as cyan spheres. One of them is bound near the active site serine and histidine side chains while the other solvent
molecule (structure water) is bound in the S1 pocket (b) close-up view of the S1 side pocket (multi-colored as shown in Fig. 3a). This figure is rotated
relative to Figure 3a to clarify the position of the S1 side pocket. Also, for clarity, only the benzamidine and the pyridyl amide of the inhibitor are
shown. The black arrow indicates the approximate directionality of the connection between the benzamidine moiety and the rest of the inhibitor. The
hydrogen bonds are shown in black dotted lines.

3008
B. A. Schweitzer et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3006–3011
Scheme 1. Synthesis of compound 4b.
˚
terminal protonated nitrogen of Lys192 (3.6 A). The
yield of 4b, probably due to low reactivity of the amino-
pyridyl nucleophile, we obtained sufficient quantities for
biological testing and X-ray structure determination.
interactions formed by the inhibitor with the lysine side
chain result in a well-defined conformation of the side
chain. The side chain of Lys192 is disordered in most
of the previous crystal structures of TF/VIIa with pyrazi-
none inhibitors (unpublished observation). While sta-
bilization of one specific conformation of the
disordered side chain of Lys192 is entropically unfavor-
able, it could be compensated by enthalpic gains
afforded by specific interactions between the amino acid
side chain and the inhibitor.
The new inhibitor 4b containing the P1 pyridyl side
pocket moiety maintained low nanomolar activity
against TF/VIIa (9.4 nM) as well as excellent selectivity
against Xa (>30 lM) and thrombin (>30 lM).
The crystal structure of compound 4b bound to TF/VIIa
is shown in Figure 3a.10 The benzamidine moiety of the
inhibitor forms an ion-pair with the carboxylate of
Asp189 as observed in the structures of other pyrazi-
none complexes with TF/VIIa.7b,d In addition, the ami-
dine moiety also forms hydrogen bonds to the side chain
of Ser190 and the main chain carbonyl of Gly219. The
inhibitor forms three hydrogen bonds to main chain
atoms of Ser214 and Gly216 of VIIa as observed in
other serine protease structures.11 The meta amine of
the phenyl group of P2 forms hydrogen bonds with the
side chains of Asp60 and Tyr94 and the main chain car-
bonyl of Thr99, and provides enhanced potency and
selectivity against factor Xa and thrombin as described
previously.7d Factor VIIa has a well-defined S2 pocket
that is electronegative due to the presence of Asp60.
The S1 amide linker also forms a hydrogen bond to main
chain carbonyl of His57.
Figure 3b shows a close view of the S1 side pocket of fac-
tor VIIa formed by residues Gln143, Asp146, Arg147,
Cys191, Lys192, Gly216, Gly219, and Cys220. As pre-
dicted, the pyridyl moiety from the meta position of
the P1 phenyl fills up most of the S1 side pocket.
A careful analysis of the data presented in Table 1 shows
that most of the compounds have comparable binding
affinities for VIIa and trypsin. This is consistent with
the crystal structures of these two serine proteases.
The active site in trypsin is relatively open with a shal-
low S1 side pocket and S2 sub site. While trypsin does
not have an aspartic acid in the S2 pocket, comparable
to Asp60 of VIIa, the side chain of Tyr94 and the car-
bonyl oxygen of Ser96 are optimally positioned for
forming good hydrogen bonding interactions with the
aniline moiety of the inhibitor. Similarly, the shallow
S1 side pocket in trypsin could accommodate bulky sub-
stituents from the P1 position of the inhibitor without
any steric clashes. Also, both trypsin and factor VIIa
contain a serine at position 190 in the S1 pocket that is
replaced by an alanine in thrombin and Xa. These struc-
tural similarities in the active sites of VIIa and trypsin
could explain the difficulties encountered in engineering
selectivity over trypsin in our current series of inhibitors.
The amide moiety from the P1 ortho phenyl position
and the pyridyl group extend toward the solvent ex-
posed S1 side pocket. While the pyridyl ring is oriented
almost orthogonal to the P1 phenyl ring of the inhibitor
(112°), the amide of the P1 side chain is involved in three
hydrogen-bonding interactions. One hydrogen bond is
formed between the amide nitrogen and main chain of
˚
Gly216 (3.3 A). The two others are between the amide
˚
carbonyl and the main chain nitrogen (3.4 A) and the

B. A. Schweitzer et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3006–3011
3009
Table 1. TF/VIIa Inhibition by 4
coupling to the pyrazinone core (Schemes 2 and 3). This
allowed rapid assembly of side pocket amides from a
single intermediate. In addition, this would allow us to
attach our P1 analogs to other interesting cores as well.
The commercially available dibromocarboxylic acid 5
was transformed in five high yielding steps to the key
intermediate 10 (Scheme 2). One particularly interesting
step in this pathway involves differentiation of the two
nitriles of compound 7. Interestingly, we obtained reac-
tion of hydroxyl amine exclusively at the less hindered
nitrile to give 8, even though the nitrile ortho to the acid
is fully conjugated and hence more electron poor. A
more in-depth discussion of the selectivity of this reac-
tion will be presented at a later date.
Compd
R
IC50 (lM)
VIIa
Xa IIa
Trypsin
We explored several traditional amide bond forming
reactions to generate 12, including HBTU, DCC/HOBt,
and EDC/HOBt. The cleanest, highest yielding reaction
came from generation of the pentafluorophenol ester 11
followed by addition of the desired amine. In addition,
the pentafluorophenol ester could be purified and
stored, allowing for easy parallel functionalization at
this position.
4a
4b
C(O)NH2
0.0062 >30
13.7 0.020
0.0094 >30 >30
0.028
4c
4d
CF3CH2NHC(O)
0.012
0.020
>30
>30
22.6 0.015
16.7 0.021
The pentafluorophenol ester was reacted with a small di-
verse set of amines with potential for H-bond interac-
tion. The compounds were deprotected to give 13 and
then coupled to the pyrazinone core (Scheme 3). The
coupling reaction typically employed HBTU with DIEA
in DMF at ambient temperature, with preactivation of
the acid first followed by addition of the P1 amine 15.
Catalytic hydrogenation of 16 in methanol with HCl/
dioxane or in acetic acid yielded our final products that
were purified by reverse phase chromatography. Prod-
ucts were isolated as the trifluoroacetate salts.
4e
4f
HO2CCH2NHC(O) 0.021
MeO2CCH2NHC(O) 0.023
>30 >30
>30 26.7 0.031
>30 >30
0.016
4g
4h
C(O)OH
H
0.152
0.0017 >30
0.050
6.1 0.015
Based on the biological and structural evidence from
compound 4b, the next logical step was to prepare a ser-
ies of analogs to further probe this region. To accom-
plish this goal a more parallel friendly synthesis for
the P1 side pocket compounds was developed.
Thus, the carboxylic acid functionalized P1 benzamidine
10 was prepared in a protected form followed by cou-
pling of the P1 to the desired amine, deprotection, and
Table 1 summarizes the biological activity of the P1 side
pocket analogs. Excellent levels of TF/VIIa inhibition
(6–21 nM) were observed for all amide side pocket ana-
Scheme 2. Synthesis of P1 side pocket analogs.

3010
B. A. Schweitzer et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3006–3011
Scheme 3. Coupling of side pocket P1 analogs to pyrazinone core.
logs. In addition, all compounds showed moderate to
excellent selectivity versus Xa and thrombin.
Source, supported by the U.S. Department of Energy,
Basic Energy Sciences, Office of Science, under Contract
No. W-31-109-Eng-38.
For comparison we synthesized a non-amide analog, en-
try 4g, and found that it was 24 times less potent than
our most potent analog. The difference in potency proba-
bly arises from the fact that the amide carbonyl forms a
better hydrogen bond to the enzyme than the carbonyl
of the acid.
References and notes
1. (a) Hirsh, J. Thromb. Res. 2003, 109, S1; (b) Gage, B. F.;
Fihn, S. D.; White, R. H. Am. J. Med. 2000, 109, 481.
2. For reviews see: (a) Sanderson, P. E. J.; Naylor-Olsen, A.
M. Curr. Med. Chem. 1998, 5, 289; (b) Menear, K. Curr.
Med. Chem. 1998, 5, 457; (c) Kaplan, K. L. Exp. Opin.
Pharmacother. 2003, 4, 653–666.
3. (a) Samana, M. M. Throm. Res. 2002, 106, V267; (b)
Quan, M. L.; Wexler, R. R. Curr. Top. Med. Chem. 2001,
1, 137.
4. (a) Benedict, C. R.; Ryan, J.; Wolitzky, B.; Ramos, R.;
Gerlach, M.; Tijburg, P.; Stern, D. J. Clin. Invest. 1991,
88, 1760; (b) Feurstein, G. X.; Patel, A.; Toomey, J. R.;
Bugelski, P.; Nichols, A. J.; Church, W. R.; Valocik, R.;
Koster, P.; Baker, A.; Blackburn, M. N. Arterioscler.
Thromb. Vasc. Biol. 1999, 19, 2554.
5. (a) Robinson, L. A.; Saiah, E. M. K. Annu. Rep. Med.
Chem. 2002, 37, 85; (b) Golino, P. Thromb. Res. 2002, 106,
V257; (c) Moll, S.; Roberts, H. R. Sem. Hematol. 2002, 39,
145; (d) Lazarus, R. A.; Olivero, A. G.; Eigenbrot, C.;
Kirchhofer, D. Curr. Med. Chem. 2004, 11, 2275.
Previous studies indicate the amidine is responsible for
poor oral bioavailability. In our search for more bioavail-
able compounds we have attempted to modulate the basi-
city of the amidine. Replacement of the amidine with less
basic or non-basic groups has led to a drastic decrease of
potency. It is possible that this newly discovered site of
interaction within TF/VIIa could provide a starting point
for making successful changes to the amidine.
In conclusion, we have identified yet another site of sub-
stitution of the P1 benzamidine moiety able to engage
additional interactions in the S1 side pocket of factor
VIIa. These additional molecular interactions can pro-
vide the basis for design of new potent and selective
TF/VIIa inhibitors.
6. (a) Szalony, J. A.; Taite, B. B.; Girard, T. J.; Nicholson,
N. S.; LaChance, R. M. J. Thromb Thrombolysis. 2002, 14,
113; (b) Zoldhelyi, P.; McNatt, J.; Shelat, H. S.; Yamam-
oto, Y.; Chen, Z.-Q.; Willerson, J. T. Circulation 2000,
101, 289; (c) Himber, J.; Kirchhofer, D.; Riedereer, M.;
Tschopp, T. B.; Steiner, B.; Roux, S. P. Thromb. Haemost.
1997, 7, 1142; (d) Harker, L. A.; Hanson, S. R.; Wilcox, J.
N.; Kelly, A. B. Haemostasis 1996, 26, 76.
7. (a) Parlow, J. J.; South, M. S. Tetrahedron 2003, 59, 7695;
(b) South, M. S.; Case, B. L.; Wood, R. S.; Jones, D. E.;
Hayes, M. J.; Girard, T. J.; Lachance, R. M.; Nicholson, N.
S.; Clare, M.; Stevens, A. M.; Stegeman, R. A.; Stallings, W.
C.; Kurumbail, R. G.; Parlow, J. J. Bioorg. Med. Chem.
Lett. 2003, 13, 2319; (c) South, M. S.; Dice, T. A.; Girard, T.
J.; Lachance, R. M.; Stevens, A. M.; Stegeman, R. A.;
Stallings, W. C.; Kurumbail, R. G.; Parlow, J. J. Bioorg.
Med. Chem. Lett. 2003, 13, 2363; (d) Parlow, J. J.; Case, B.
Acknowledgments
We thank Rhonda LaChance for running the biological
assays. We also thank Dr. Huey Shieh for some of the
early crystallographic refinements of the TF/VIIa struc-
ture. Diffraction data for the TF/VIIa complex with the
different inhibitors were collected at beamline 17-ID in
the facilities of the Industrial Macromolecular Crystal-
lography Association Collaborative Access Team
(IMCA-CAT) at the Advanced Photon Source.
IMCA-CAT facilities are supported by the corporate
members of the IMCA and through a contract with
the Illinois Institute of Technology (IIT), executed
through the IITÕs Center for the Advanced Photon

B. A. Schweitzer et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3006–3011
3011
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.; Stegeman, R. A.; Stevens, A.
M.; Stallings, W. C.; Kurumbail, R. G.; South, M. S. J.
Med. Chem. 2003, 46, 4050; (e) Parlow, J. J.; Stegeman, R.
A.; Stevens, A. M.; Stallings, W. C.; Kurumbail, R. G.;
South, M. S. J. Med. Chem. 2003, 46, 4297; (f) Parlow, J. J.;
Stegeman, R. A.; Stevens, A. M.; Stallings, W. C.; Kurum-
bail, R. G.; South, M. S. Bioorg. Med. Chem. Lett. 2003, 13,
3721; (g) Parlow, J. J.; Dice, T. A.; Lachance, R. M.; Girard,
T. J.; Stevens, A. M.; Stegeman, R. A.; Stallings, W. C.;
Kurumbail, R. G.; South, M. S. J. Med. Chem. 2003, 46,
4043; (h) Parlow, J. J.; Kurumbail, R. G.; Stegeman, R. A.;
Stevens, A. M.; Stallings, W. C.; South, M. S. J. Med. Chem.
2003, 46, 4696; (i) Parlow, J. J.; Stevens, A. M.; Stegeman,
R.; Stallings, W. C.; Kurumbail, R. G.; South, M. S. J. Med.
Chem. 2003, 46, 4297.
D. J.; Rueppel, M. L.; Neumann, W. L.; Long, S.; Wei,
H.; Trujillo, J.; Parlow, J. J.; Jones, D. E.; Case, B.; Hayes,
M. J.; Zeng, W.; Abbas, Z.; Fenton, R. L.; Kusturin, C.
L.; Hayat, R. K.; Sample, K. R.; Schweitzer, B. A.; Wood,
R. S.; Szalony, J.; Suleymanov, O. D.; Salyers, A.;
Nicholson, N. S. WO Patent 2003029224, 2003. Chem.
Abstr. 2003, 138, 304305.
9. Four possible products could be present in this mixture:
mono-aniline protected, mono-amidine protected,
bis-amidine
protected,
mono-aniline/mono-amidine.
Ultimately, deprotection of the 3a and b mixture would
give us one compound, 4b. Therefore, we did not pursue
extensive characterization of these products.
10. Crystallographic data for the structures in this paper have
been deposited with the RCSB Protein Data Bank as
PDB ID 1Z6J. The structural details can be viewed at
www.rcsb.org using the ID number above.
8. For synthesis of this and other pyrazinone cores see:
South, M. S.; Webber, R. K.; Huang, H.; Toth, M. V.;
Moormann, A. E.; Snyder, J. S.; Scholten, J. A.; Garland,
11. Banner, D. W.; DÕArcy, A.; Chene, C.; Winkler, F. K.;
Guha, A.; Konigsberg, W. H.; Nemerson, Y.; Kirchhofer,
D. Nature 1996, 380, 41.