Synthesis, characterization, and structure-activity relationships of amidine-substituted (bis)benzylidene-cycloketone olefin isomers as potent and selective factor Xa inhibitors

Article abstract of DOI:10.1021/jm990456v

Factor Xa (FXa) is a trypsin-like serine protease that plays a key role in blood coagulation linking the intrinsic and extrinsic pathways to the final common pathway of the coagulation cascade. During our initial studies, we observed facile photochemical conversion of the known FXa/tPA inhibitor, BABCH [(E,E)-2,7-bis(4-amidinobenzylidene)cycloheptan-1-one, 1a], to the corresponding (Z,Z) olefin isomer, 1c (FXa K(i) = 0.66 nM), which was over 25 000 times more potent than the corresponding (E,E) isomer (1a, FXa K(i) = 17 000 nM). In order to determine the scope of this observation, we expanded on our initial investigation through the preparation of the olefin isomers in a homologous series of cycloalkanone rings, 4-substituted cyclohexanone analogues, and modified amidine derivatives. In most cases the order of potency of the olefin isomers was (Z,Z) > (E,Z) > (E,E) with the cycloheptanone analogue (1c) showing the most potent factor Xa inhibitory activity. In addition, we found that selectivity versus thrombin (FIIa) can be dramatically improved by the addition of a carboxylic acid group to the cycloalkanone ring as seen with 8c (FXa K(i) = 6.9 nM, FIIa K(i) > 50 000 nM). Compounds with one or both of the amidine groups substituted with N-alkyl substituents or replaced with amide groups led to a significant loss of activity. In this report we have demonstrated the importance of the two amidine groups, the cycloheptanone ring, and the (Z,Z) olefin configuration for maximum inhibition of FXa within the BABCH template. The results from this study provided the foundation for the discovery of potent, selective, and orally active FXa inhibitors.

Full text of DOI:10.1021/jm990456v

J . Med. Chem. 1999, 42, 5415-5425
5415
Syn th esis, Ch a r a cter iza tion , a n d Str u ctu r e-Activity Rela tion sh ip s of
Am id in e-Su bstitu ted (Bis)ben zylid en e-Cyclok eton e Olefin Isom er s a s P oten t
a n d Selective F a ctor Xa In h ibitor s1,2
William J . Guilford,* Kenneth J . Shaw,* J erry L. Dallas, Sunil Koovakkat, Wheeseong Lee, Amy Liang,
David R. Light, Margaret A. McCarrick, Marc Whitlow, Bin Ye, and Michael M. Morrissey
Discovery Research, Berlex Biosciences, 15049 San Pablo Avenue, P.O. Box 4099, Richmond, California 94804-0099
Received September 7, 1999
Factor Xa (FXa) is a trypsin-like serine protease that plays a key role in blood coagulation
linking the intrinsic and extrinsic pathways to the final common pathway of the coagulation
cascade. During our initial studies, we observed facile photochemical conversion of the known
FXa/tPA inhibitor, BABCH [(E,E)-2,7-bis(4-amidinobenzylidene)cycloheptan-1-one, 1a ], to the
corresponding (Z,Z) olefin isomer, 1c (FXa Ki ) 0.66 nM), which was over 25 000 times more
potent than the corresponding (E,E) isomer (1a , FXa Ki ) 17 000 nM). In order to determine
the scope of this observation, we expanded on our initial investigation through the preparation
of the olefin isomers in a homologous series of cycloalkanone rings, 4-substituted cyclohexanone
analogues, and modified amidine derivatives. In most cases the order of potency of the olefin
isomers was (Z,Z) > (E,Z) > (E,E) with the cycloheptanone analogue (1c) showing the most
potent factor Xa inhibitory activity. In addition, we found that selectivity versus thrombin
(FIIa) can be dramatically improved by the addition of a carboxylic acid group to the
cycloalkanone ring as seen with 8c (FXa Ki ) 6.9 nM, FIIa Ki > 50 000 nM). Compounds with
one or both of the amidine groups substituted with N-alkyl substituents or replaced with amide
groups led to a significant loss of activity. In this report we have demonstrated the importance
of the two amidine groups, the cycloheptanone ring, and the (Z,Z) olefin configuration for
maximum inhibition of FXa within the BABCH template. The results from this study provided
the foundation for the discovery of potent, selective, and orally active FXa inhibitors.
In tr od u ction
FXa in purified systems of 40 nM9 and is efficacious in
several species.10 We recently reported the synthesis of
N-[2-[5-[amino(imino)methyl]-2-hydroxyphenoxyl-3,5-di-
flu-oro-6-[3-(4,5-dihydro-1-methyl-1H-imidazol-2-yl)phe-
noxy]pyridin-4-yl]-N-methylglycine, ZK807834 (CI 1031),
a novel inhibitor of FXa, which inhibits purified human
FXa with a Ki of 0.11 nM.7b Preliminary studies of the
selectivity, pharmacokinetics, oral availability, and ef-
ficacy have been reported.11 In a direct comparison in
rabbits, ZK-807834 was shown to be efficacious at a 25-
fold lower dose than DX-9065a, consistent with the
observed differences of the in vitro potencies.12
Factor Xa (FXa), a trypsin-like serine protease, links
the intrinsic and extrinsic pathways to the final common
pathway in the blood coagulation cascade. The primary
role of FXa is the proteolytic activation of prothrombin,
after combining with factor Va and calcium on a
phospholipid membrane to form the prothrombinase
complex.3 Thrombin then promotes blood clot formation
by catalyzing the formation of polymerizable fibrin from
fibrinogen in addition to activating platelets. Ex vivo
data demonstrate that prothrombinase, not thrombin,
is the major determinant of the procoagulant activity
of human whole-blood clots.4
Progression of thrombosis plays an important role in
the pathophysiology of unstable angina and myocardial
infarction and may be a factor in the failure of coronary
fibrinolysis. Progression of deep vein thrombosis can
contribute to pulmonary embolism. Recent data indicate
FXa may be more important than thrombin as a
mediator of thrombus progression.4,5 This hypothesis is
consistent with studies showing persistent inhibition of
procoagulant activity and thrombus progression by
inhibitors of FXa.6 These studies and others stimulated
interest in novel approaches to nonpeptide-based inhibi-
tors of human FXa as potential candidates for preclini-
cal testing.7
In our previous communication we reported that the
known FXa/tPA inhibitor, BABCH [(E,E)-2,7-bis(4-
amidinobenzylidene)cycloheptan-1-one, 1a ], could be
photochemically converted to a mixture of the (E,Z) and
(Z,Z) double bond isomers (Scheme 1). When these
isomers were tested in the absence of light we found
that the (Z,Z) isomer, 1c (FXa Ki ) 0.66 nM), was over
25 000 times more potent then the corresponding (E,E)
isomer (1a , FXa Ki ) 17 µM). The high potency and
selectivity of 1c was very encouraging, and the confor-
mational rigidity of this system provided an excellent
template leading to the design of a variety of heterocyclic-
based FXa inhibitor templates.13 In order to understand
this large disparity between the FXa inhibitory activi-
ties of the geometrical isomers, we investigated other
structurally related analogues. In this report we de-
scribe the synthesis, photoisomerization, and in vitro
activities of a series homologous cycloalkanone rings,
The most clinically8 advanced nonpeptide FXa inhibi-
tor is (2S)-2-[4-[((3S)-1-acetimidoyl-3-pyrrolidinyl)oxo]-
phenyl]-3-(7-amidino-2-naphthyl)propanoic acid hydro-
chloride pentahydrate, DX-9065a, which has a Ki for
10.1021/jm990456v CCC: $18.00 © 1999 American Chemical Society
Published on Web 12/10/1999

5416 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 26
Guilford et al.
Sch em e 1
monobenzylidene and the corresponding (E,Z) isomer,
and therefore they were purified by preparative reverse-
1
phase HPLC. The H NMR spectra are consistent with
a symmetrical configuration, and the downfield chemical
shift of the vinylic protons (δ ) 7.08-7.69 ppm, Table
4) is consistent with the 1H NMR spectra of other
reported R,â-unsaturated cyclic ketones having (E,E)
double bond configurations.15 Unambiguous assignment
of the (E,E) double bond geometry was confirmed by
NOESY NMR studies of 1a ,2 3a , and 4a . In these
studies, NOE cross peaks were observed between the
cycloheptanone ring protons and the protons meta to
the amidine substituent, with no observed NOE between
the cycloheptanone ring protons and the vinylic protons.
Solutions of the (E,E) isomers (1a -6a ) were irradi-
ated with a 450 W high-pressure mercury lamp to
produce an equilibrium mixture of the corresponding
(E,Z) 1b-6b and (Z,Z) isomers 1c, 3c-6c with a small
amount of the starting (E,E) isomer. The isomerization
could be effected on a preparative scale using either a
450 W high-pressure mercury lamp or a 150 W standard
flood lamp. The reaction mixtures were irradiated until
a constant product ratio was obtained as determined
by HPLC, typically 2-4 h with the 450 W high-pressure
mercury lamp. The product ratio could be determined
4-substituted cyclohexanone derivatives, and substi-
tuted amidine analogues (Tables 1-3).
Ch em istr y
The symmetrical (E,E)-bis-(amidinobenzylidene)cyclic
ketone analogues in Table 1 (1a -6a ) were prepared by
acid-catalyzed condensation of 4-amidinobenzaldehyde
(7) with the appropriate cycloalkanone (Scheme 2).14
The poor solubility of the (E,E) isomers facilitated the
purification of the lower ring analogues (1a -3a ) which
precipitated from the crude reaction mixture and were
recrystallized to afford analytically pure samples as
salts. The larger ring analogues (4a -6a ) could not be
effectively precipitated from a mixture that included the
1
by integration of the vinyl resonances in the H NMR
spectrum of the crude photolysis mixtures. For example,
photolysis of 1a afforded a product ratio of <5%, 55%,
and 40% for 1a , 1b, and 1c, respectively. The isomers
were separated by preparative HPLC, and purified
fractions were handled with care to exclude light and
prevent reequilibration during isolation. Irradiation of
2a failed to afford any of the corresponding (Z,Z) isomer,
Ta ble 1. In Vitro Inhibitory Activities of Bisarylamidine Cyclic Ketone Isomers Tested in the Absence of Light
Ki (nM)a
Ki (nM)
Ki (nM)
n
no.
FXa
FIIa
Trp
no.
FXa
FIIa
Trp
no.
FXa
FIIa
Trp
0
1
2
3
4
5
2a
3a
1a
4a
5a
6a
4600
5600
17000
1800
390
2200
3000
6200
2100
3700
730
590
170
280
310
1500
3000
2b
3b
1b
4b
5b
6b
230
420
200
130
320
510
19200
120
130
200
350
730
260
29
30
340
2100
2c
3c
1c
4c
5c
6c
NDb
18
0.66
4.6
2.8
110
NDb
810
530
300
290
740
NDb
440
33
100
190
2200
960
150
a
Ki values for these competitive inhibitors are averaged from multiple determinations (n > 2), and the standard deviations are <30%
b
of the mean. Compound not available for assay.

Synthesis of Selective Factor Xa Inhibitors
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 26 5417
Ta ble 2. In Vitro Inhibitory Activities of Bisarylamidine Cyclic Ketone Isomers Tested in the Absence of Light
Ki (nM)a
Ki (nM)a
Ki (nM)a
X
no.
FXa
FIIa
Trp
no.
FXa
FIIa
Trp
no.
FXa
FIIa
Trp
CH-CO2H
CH-CO2Me
CH-Ph
8a
9a
10a
11a
24400
>5000
3300
21500
4400
630
880
72
56
8b
9b
10b
11b
NDb
>500
395
NDb
600
325
630
NDb
224
160
61
8c
9c
10c
11c
6.9
4.8
5.2
2.5
>50000
1100
1700
380
430
100
100
20
N-CO2Et
>5000
3300
125
225
a
Ki values for these competitive inhibitors are averaged from multiple determinations (n g 2), and the standard deviations are <30%
b
of the mean. Compound not available for assay.
Ta ble 3. In Vitro Inhibitory Activities of Mono- and Bisarylamidine Cycloheptanone Analogues Tested in the Absence of Light
(E,E) isomer Ki (nM)a
(E,Z) and (Z,E) isomers Ki (nM)
FXa FIIa Trp
500 360 155
(Z,Z) isomer Ki (nM)
no.
FXa
FIIa
Trp
no.
no.
FXa
FIIa
Trp
30
90
690
13a
14a
15a
17a
18a
19a
29000
350
12850
3700
>500
ND
840
500
>500
ND
13b
14b
15b
17b
18b
19b
13c
14c
15c
17c
18c
19c
12
18
50
1600
1400
110
52
1850
>5000
>5000
140
120
15
500
NA
>500
350
NDb
NAc
>5000
2900
>5000
1400
>5000
>5000
44
>5000
1400
>5000
2100
>5000
410
>5000
1000
>5000
200
1600
5.6
a
Ki values for these competitive inhibitors are averaged from multiple determinations (n g 2), and the standard deviations are <30%
of the mean. Compound not available for assay. c Compound not assayed on this protease.
b
Sch em e 2
2c, which may be attributable to a competing dimer-
ization reaction that has been reported for 2,5-difur-
furylidenecyclopentanone.15a
cycloheptanone ring which unambiguously support the
assignment of the (Z,Z) double bond configuration.
The 4-substituted cyclohexanone and piperidinone
analogues in Table 2 were prepared in a similar fashion
as outlined above. Condensation of the requisite cyclic
ketones with 4-amidinobenzaldehyde afforded the ap-
propriate (E,E) isomer. Subsequent photochemical
isomerization, followed by preparative HPLC afforded
the corresponding (E,Z) and (Z,Z) isomers. One excep-
tion was compound 8c which could not be obtained pure
after photoisomerization and chromatography. This
compound was prepared by saponification of the purified
ester analogue 9c, followed by HPLC purification.
Assignment of double bond geometry for this series was
carried out by 1H NMR and is consistent with the
unsubstituted cycloalkanone series (Table 4).
The three isomeric adducts could be differentiated and
unambiguously identified by their characteristic NMR
1
spectra. The H NMR spectra of the (E,Z) isomers (1b-
6b) are significantly more complex, showing separate
resonances for many of the hydrogen atoms, illustrating
the unsymmetrical nature of these isomers. The 1H
NMR spectra of the (Z,Z) isomers (1c, 3c-6c) contain
the same elements of symmetry as the (E,E) isomers,
but the upfield chemical shifts of the vinylic protons
(δ ) 6.82-6.96, Table 4) are consistent with related
compounds containing the (Z,Z) double bond configur-
ation.15b Studies with 1c, 3c, and 4c showed NOE cross
peaks between the vinylic and allylic protons on the

5418 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 26
Guilford et al.
Ta ble 4. Vinyl Proton 1H NMR Chemical Shift Values of
Bisarylamidine Cyclic Ketone Isomers
with a single equivalent of 4-cyanobenzaldehyde af-
forded (4-cyanobenzylidene)cycloheptanone (12). Con-
version of nitrile to the substituted N-methyl and N,N-
dimethylamidine via Pinner reaction was followed by
condensation with 4-amidinobenzaldehyde to afford the
corresponding (E,E) isomers (13a and 14a , respectively).
Condensation of 12 with 7 in the presence of phosphoric
acid at 100 °C also effected conversion of the nitrile to
afford the amide 15a . Surprisingly, the major compo-
nent isolated after condensation of cycloheptanone with
an excess of 4-cyanobenzaldehyde in phosphoric acid at
100 °C was the bis-nitrile (16) and not the corresponding
amide. The bis-amide (17a ) was prepared by hydrolysis
of 16 with sulfuric acid at 0-5 °C. The bis-dimethyl-
amidine analogue 18a was obtained by condensation of
cycloheptanone with 4-dimethylamidinobenzaldehyde
(20). In a similar manner, the bis-3-amidino analogue,
19a , was prepared by condensation of cycloheptanone
with 3-amidinobenzaldehyde.16 Photolysis of the (E,E)
isomers afforded an equilibrium mixture of isomers that
were separated by preparative HPLC as described
above. Due to the nonredundancy of the two aryl groups,
photolysis of the unsymmetrical analogues (13a -15a )
no.
δa
no.
δa
no.
δa
1a
2a
3a
4a
5a
6a
8a
9a
10a
11a
13a
14a
15a
17a
18a
19a
7.35
7.56
7.69
7.08
7.10
7.20
7.85
7.85
7.75
7.75
7.39
7.38
7.34
NDc
7.33
7.33
1b
2b
3b
4b
5b
6b
8b
9b
l0b
11b
13b
14b
15b
17b
18b
19b
6.84, 7.63
1c
2c
3c
4c
5c
6c
8c
9c
10c
l1e
13c
14c
15c
17c
18c
19c
6.96
7.54, 7.7(m)b
6.90, 7.42
NDc
6.82
6.83
6.86
6.96
6.85
6.85
6.9
7.13
6.70, 7.65(m)b
6.86, 7.65(m)b
7.16, 7.65(m)b
NDc
6.95, 7.48
6.95, 7.45
7.15, 7.55
6.82, 7.7(m)b
6.8, 7.6(m)
6.74, 6.8, 7.6(m)b
6.73, 7.56
6.98
6.95, 7.00
6.89, 6.96
6.84
6.93
6.95
6.78, 7.6
6.8, 7.7(m)
a
Chemical shift of vinyl protons. b Chemical shift of vinyl proton
could not be separated unambiguously from aromatic multiplet
centered at listed resonance. c Compound not analyzed by 1H
NMR.
The symmetrical and unsymmetrical amide and alkyl-
amidine (E,E) analogues were prepared as illustrated
in Schemes 3 and 4. Condensation of cycloheptanone
Sch em e 3a
a
Reagents: (i) 4-OHC-C6H4-CN, H3PO4; (ii) HCl, EtOH; (iii) MeNHR, EtOH, heat; (iv) H3PO4, 7.
Sch em e 4a
a
Reagents: (i) 4-OHC-C6H4-CN, H3PO4; (ii) H2SO4; (iii) HCl, 4-OHC-C6H4-C(NH)NMe2 (20); (iv) H3PO4, 3-OHC-C6H4-C(NH)NH2.

Synthesis of Selective Factor Xa Inhibitors
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 26 5419
Sch em e 5
Ta ble 5. In Vitro Inhibitory Activities of Bisarylamidine Cyclic
Ketone Isomers Tested in the Absence of Light against tPA
and uPA
produces two unique (E,Z) isomers in an approximate
1:1 ratio (Scheme 5). Due to the difficulties involved with
the separation of these (E,Z) isomers, these analogues
were tested as a mixture.
(E,E) isomer
(E,Z) isomer
(Z,Z) isomer
no. tPAa
uPAb
no. tPAa uPAb no. tPAa
uPAb
Resu lts a n d Discu ssion
1a 39300 >30000
1b
720 10000
1c
16
61
20000
>30000
19a 66000
20000 19b 7300 15000 19c
All of the compounds illustrated in Tables 1-3 were
tested for their ability to inhibit human factor Xa in a
purified enzyme system. In addition, all compounds
were assayed against human thrombin and bovine
trypsin. Assays were run in duplicate, and Ki values
were determined as described previously.17 All assays
were run in the absence of light, and extra care was
taken to minimize exposure of solutions to light before
and during the bioassays to minimize isomerization of
the inhibitors. The ability to avoid isomerization was
verified by HPLC analysis of selected test compounds
after completion of the assays.
The data in Table 1 illustrate that the order of potency
in FXa for the double bond isomers of this series of cyclic
ketone analogues is (Z,Z) > (E,Z) > (E,E). In addition,
the FXa inhibitory activity of the (Z,Z) isomers varied
significantly with the cycloalkanone ring size as seen
by the 27-fold drop in activity for the cyclohexanone
analogue (3c, FXa Ki ) 18 nM) and the 170-fold lowered
activity of the cyclodecanone analogue (6c, FXa Ki )
110 nM) when compared with 1c (FXa Ki ) 0.66 nM).
The corresponding cyclooctanone and cyclononanone
analogues (4c and 5c, respectively) were within 7-fold
of the activity of 1c. A similar trend for potency of this
series was seen with trypsin; however, the trypsin Ki
values are 20-70 fold higher than the corresponding
values for FXa. The measured Ki values were all within
a factor of three for the (Z,Z) analogues against throm-
bin. These results suggest that the (Z,Z) inhibitors may
adopt a similar binding mode in trypsin and FXa, but
may bind in a different mode in thrombin.
The higher affinity of the (Z,Z) isomers for FXa as
compared to trypsin suggests some favorable interac-
tions in the FXa active site which are not conserved in
the trypsin active site. Identification of the specific
residues which are responsible for the higher potencies
of the (Z,Z) compounds in FXa is complicated by a
number of amino acid differences in the binding sites
of these two proteases. The binding mode of 1c in FXa
has not been experimentally determined, and it is not
known whether 1c binds in the same mode in trypsin
as in FXa. In our previous report, we proposed two
possible binding modes to account for the potent FXa
inhibitory activity of 1c.2 To gain further understanding
of the binding mode for this series of (Z,Z) inhibitors,
we attempted to resolve the crystal structure of 1c in
a
Ki values (nM) for these competitive inhibitors are averaged
from multiple determinations (n g 2), and the standard deviations
are <30% of the mean.15 IC50 values (nM).
b
trypsin. Since photoisomerization may occur during
crystal data collection, good resolution was needed to
determine which isomer was bound in the active site of
trypsin. Although both benzamidine rings were visible
in the structure, occupying the S1 and S4 binding sites
of the extended binding mode, the electron density for
the central ring was absent in the solved structure
resulting in an ambiguous determination of the double
bond geometry of the inhibitor. Recently, the crystal
structure of a commercial source of BABCH in trypsin
was reported by another group at 1.7 Å resolution.18
These investigators identified the trypsin-bound inhibi-
tor as the (E,Z) isomer (1b) which binds in an extended
mode in trypsin, consistent with their initial structural
data.18
The crystal structure of BABCH complexed with
tissue-type plasminogen activator (tPA), another related
serine protease, has been reported recently.19 The
observed binding mode in this structure is similar to
that in the crystal structure of 1b with trypsin, except
for the geometry of the cycloheptanone ring. Although
the bound inhibitor was originally thought to be the
(E,E)-isomer (1a ), the authors later proposed that it may
in fact be 1c, based on the (Z,Z) preference of FXa and
its similarity to tPA.18 These workers suggest that in
tPA, Tyr 99 prevents the (E,Z)-isomer from binding well
but allows the (Z,Z) isomer to bind. Tyr 99 is also found
in FXa, but is Leu in trypsin and thrombin. The data
in Table 5 show that tPA has a 45-fold preference for
binding 1c as compared to 1b, and a 2500-fold prefer-
ence for 1c over 1a . This is the same trend seen for FXa,
although the magnitude of the (Z,Z) preference is lower
for tPA than for FXa. Our data support the suggested
similarity between FXa and tPA in their relative affini-
ties for (Z,Z), (E,Z), and (E,E) isomers of BABCH, and
the corresponding expectation that the isomer bound in
the tPA crystal complex is 1c.
Table 5 shows that all of the geometrical isomers of
1 and 19 inhibit urokinase-type plasminogen activator
(uPA) very weakly. In our earlier communication we
proposed two potential binding modes for the BABCH
olefin isomers in the active site of FXa and, based on

5420 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 26
Guilford et al.
the 3D structure of uPA, neither of these binding modes
would be favorable for these inhibitors, explaining the
poor uPA inhibitory activity. Unlike the other serine
proteases discussed previously, uPA has its S4 site
effectively blocked by residues Leu 97B, Ala 98, and His
99,20 so that an extended binding mode spanning the
SI and S4 sites is not possible for uPA. In addition, the
Glu 217 in FXa is an arginine residue in uPA, excluding
the U-shaped binding mode.
The data in Table 1 clearly demonstrate that the (E,E)
double bond isomers are significantly weaker inhibitors
of factor Xa compared to the corresponding (E,Z) or (Z,Z)
isomers. The difference is most pronounced in the five-,
six-, and seven-membered rings (1a -3a ) where the Ki
values for these inhibitors are all within a factor of 40
of benzamidine (FXa Ki ) 158 000 nM). Benzamidine
has been shown to bind to serine protease inhibitors
through a combination of a salt bridge interaction and
hydrophobic interactions in the S1 pocket.21 The high
Ki values for 1a -3a suggest that the interactions of
these inhibitors with FXa are based primarily on one
of the arylamidine substituents in the S1 pocket, with
no other productive contacts of the inhibitor with the
protease. This is consistent with the binding mode
proposed in our recent communication,2 with one of the
arylamidine substituents bound in the S1 pocket, and
the second arylamidine substituent being solvent ex-
posed. The increased activity of the higher ring ana-
logues, 4a -6a , can be attributed to the increased
flexibility of the cycloalkanone ring, allowing the inhibi-
tors to potentially adopt conformation binding modes
similar to the (E,Z) isomers. The lower ring analogues
(1a -4a ) are more potent inhibitors of trypsin when
compared to FXa or thrombin, illustrating the difference
between the binding pockets of these serine proteases.
Interestingly, the selectivity difference between FXa and
trypsin is reversed with the more flexible cyclononanone,
and cyclodecanone analogues (5a , 6a ) demonstrate an
approximate 4-fold increased inhibitory activity for FXa
compared to trypsin.
F igu r e 1. Model of binding mode of 8c in FXa, from molecular
dynamics simulation. The Cerius2 program from MSI was used
to create this figure.
carboxylate on inhibitor binding to these two proteases.
The final structure of the FXa/8c complex is shown in
Figure 1. In FXa, the binding modes of 3c and 8c were
essentially the same, and they strongly resemble the
extended binding mode previously found with MD
simulations of 1c.2 Gln 192, which is Glu in thrombin,
makes extensive hydrophobic contacts with 8c and 3c
and forms a weak hydrogen bond with the carboxylate
of 8c. In contrast to the FXa results, the binding of 8c
in thrombin was significantly different from 3c. During
the simulation, the amidine-Asp 189 salt bridge in the
S1 pocket began to break, indicating the instability of
the bound structure. The inhibitor binding modes of 3c
and 8c in FXa and 3c in thrombin showed no such
instability during the 500 ps simulations. Likewise, a
200 ps simulation of 10c, which has a 4-phenyl sub-
stituent, in thrombin demonstrated no instability in its
binding mode. These simulations support the data in
Table 2 and suggest that the low affinity of 8c for
thrombin may be a result of an electrostatic repulsion
between the carboxy group of the inhibitor and that of
Glu 192, which is Gln in FXa and trypsin. The increased
thrombin selectivity of Daiichi’s compound, DX-9065a,
was also significantly improved with the addition of a
carboxylic acid group, and similar conclusions were
reached from their modeling studies.23 This observation
provided us with insights that were advantageously
used for the design of novel low molecular inhibitors
with excellent thrombin selectivity.7b
The highly basic nature of the bisarylamidine inhibi-
tors in Table 1 and 2 would present formidable obstacles
toward the development of orally active agents. Re-
placement of one or both of the basic amidine substit-
uents with a less polar substituent would increase the
log P of the molecule and present a molecule that would
have a better chance for absorption after oral adminis-
tration. For example, the calculated logD (pH ) 7.0)
values for bisamidine 1c and monoamidine 15c are
-2.93 and 0.45, respectively.24 Therefore, we investi-
gated a series of analogues where one or both of the
amidine substituents are modified to better understand
the structural requirements of the amidine substituents
The inhibitory activities for a series of 4-substituted
cyclic analogues (8-11) are illustrated in Table 2.
Consistent with the trend discussed above, the rank
potency of FXa inhibitory activity for the geometrical
isomers in this series is (Z,Z) > (E,Z) > (E,E). Each of
the diverse substituents at the C-4 position of the (Z,Z)
isomer, 8c-11c, afforded a modest increase in the FXa
inhibitory potency when compared with 3c, suggesting
that substitution at this position may favorably influ-
ence binding to FXa. In addition, the carboxylate
analogue 8c showed a dramatic increase in its thrombin
selectivity (thrombin Ki > 50 000 nM). This selectivity
toward thrombin appears to be based on a negative
electrostatic interaction and not a steric interaction
since neither the corresponding ester (9c) nor the benzyl
or carbamate analogues (10c and 11c, respectively)
showed any significant difference in their thrombin
activity.
Molecular dynamics simulations were performed to
investigate the effects of a 4-carboxy substituent on (Z,Z)
cyclohexanone binding to FXa and thrombin.22 Com-
plexes of FXa and thrombin with 3c and its 4-carboxy-
late analogue 8c were minimized and subjected to 500
ps of molecular dynamics to compare the effect of the

Synthesis of Selective Factor Xa Inhibitors
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 26 5421
22.8 mmol) in 20 mL of 85% phosphoric acid was stirred as
neat cycloheptanone (1.27g, 11.4 mmol) was added. The
reaction was heated at 100 °C for 3 h, allowed to cool, and
diluted with methanol (50 mL) and ether (500 mL). The
resulting solid was isolated by decanting. Purification was
accomplished by treating a methanolic solution with charcoal,
filtering through Celite, and precipitating with ether to yield
1a (0.6 g). An aqueous acetonitrile solution of 1a (0.53 g, 1.2
mmol) was irradiated with the 450 W lamp for 2 h. Purification
by preparative HPLC using a 20-30% gradient of acetonitrile
(0.1% TFA) in water (0.1% TFA) yielded two new compounds,
1b (0.2 g) and 1c (0.25 g), which were isolated as white solids
after lyophilization. 1a : NMR (DMSO, TMS) δ 1.92 (s, 4H),
2.67 (s, 4H), 7.35 (s, 2H), 7.72 (d, 4H), 7.92 (d, 4H), 9.50 (br s,
8H). Anal. (C23H24N4O‚2HCl‚H2O) C, H, N. 1b: NMR (DMSO,
TMS) δ 1.90 (br s, 4H), 2.51 (s, 2H), 2.79 (br s, 2H), 6.84 (s,
1H), 7.35 (d, 2H) 7.63 (s, 1H), 7.72 (m, 4H), 7.91 (d, 2H), 9.09
(s, 2H), 9.19 (s, 2H), 9.29 (s, 2H), 9.40 (s, 2H). Anal.
(C23H24N4O‚2C2HF3O2‚0.5H2O) C, H, N, F. 1c: NMR (DMSO,
TMS) δ 1.90 (m, 4H), 2.58 (m, 4H), 6.96 (s, 2H), 7.44 (d, 4H),
7.71 (d, 4H), 9.17 (br m, 8H). Anal. (C23H24N4O‚2C2HF3O2) C,
H, N, F.
Dou ble Bon d Isom er s of 2,7-Bis(4-am idin oben zyliden e)-
cyclop en ta n -1-on e (2a , 2b). A suspension of 7 (4.2 g, 22.8
mmol) in 20 mL of 85% phosphoric acid was stirred as neat
cyclopentanone (0.94, 11.1 mmol) was added. The reaction was
heated at 100 °C for 3 h, allowed to cool, diluted with methanol,
and cooled in an ice bath for 30 min. The resulting solid was
isolated by filtration and dried to give 2a . An aqueous
acetonitrile solution of 2a (0.1 g, 0.24 mmol) was irradiated
with a 150 W lamp for 24 h. Purification by preparative HPLC
using 17% acetonitrile (0.1% TFA) in water (0.1% TFA) gave
44 mg of a white solid, 2b, after lyophilization. 2a : NMR
(DMSO, TMS) δ 2.52 (m, 2H), 3.24 (m, 2H), 7.56 (s, 2H), 8.0
(m, 8H), 9.50 (br s, 8H). Anal. (C21H20N4O‚2HCl‚H2O) C, H,
N, Cl. 2b: NMR (DMSO, TMS) δ 1.38 (m, 1H), 2.30 (m, 1H),
2.45 (m, 1H), 2.88 (m, 1H), 4.69 (s, 1H), 7.54 (s, 1H), 7.7 (m,
9H), 9.3 (m, 8H). Anal. (C21H20N4O‚1.9C2HF3O2‚2.5H2O) C, H,
N, F.
Dou ble Bon d Isom er s of 2,7-Bis(4-am idin oben zyliden e)-
cycloh exa n -1-on e (3a , 3b, 3c). A suspension of 7 (4.2 g, 22.8
mmol) in 20 mL of 85% phosphoric acid was stirred as
cyclohexanone (1.11 g, 11.3 mmol) was added at once. The
reaction was heated at 100 °C for 3 h. The reaction was allowed
to cool, diluted with 50 mL of methanol, and stirred for 30
min. The resulting solid was isolated by filtration and titrated
with methanol and ether to yield 0.54 g of 3a .
A aqueous acetonitrile solution of 3a was irradiated with a
450 W lamp for 3 h. Purification by preparative HPLC using
a 17-27% gradient of acetonitrile (0.1% TFA) in water (0.1%
TFA) yielded 0.3 g of 3b and 80 mg of 3c as solids after
lyophilization. 3a : NMR (DMSO, TMS) δ 1.77 (m, 2H), 2.94
(s, 4H), 7.69 (s, 2H), 7.8 (m, 8H), 9.36 (s, 4H), 9.54 (s, 4H).
Anal. (C22H22N4O‚2HCl‚0.25H2O) C, H, N, Cl. 3b: NMR
(DMSO, TMS) δ 1.04 (m, 2H), 2.80 (m, 2H), 2.92 (m, 2H), 6.90
(s, IH), 7.42 (s, 1H) 7.60 (d, 2H), 7.75 (m, 4H), 7.88 (d, 2H),
9.13 (d, 4H), 9.38 (d, 4H). Anal. (C22H22N4O‚1.9C2HF3O2‚H2O)
C, H, N, F. 3c: NMR (DMSO, TMS) δ 2.30 (m, 2H), 2.82 (m,
4H), 6.82 (s, 2H), 7.6 (m, 8H), 9.4 (m, 8H). Anal. (C22H22N4O‚2C2-
HF3O2‚0.6H2O) C, H, N, F.
with the hope that this data could be applied to other
related amidine-based templates. The data from Table
3 again demonstrates the trend of potency for the
different geometrical isomers ((Z,Z) > (E,Z) > (E,E)).
In the (Z,Z) series, bis-replacement of the amidine
substituent afforded inhibitors that were more than 3
orders of magnitude less active than 1c, as seen with
the bis-amide derivative 17c (FXa Ki >5000 nM) and
the bis-N,N-dimethylamidine analogue 18c (FXa Ki )
1600 nM). This illustrates the importance of the salt
bridge interaction of the amidine substituent in the S1
pocket for this series of inhibitors. Substitution of only
one of the amidine groups with N-methyl or N,N-
dimethyl (13c and 14c, respectively), resulted in ap-
proximately 20-fold loss of activity against FXa when
compared with 1c, with minimal effect on the thrombin
or trypsin inhibition. Further substitution of an amide
group, 15c, resulted in a 75-fold loss in activity against
FXa and trypsin but showed a 10-fold increased potency
toward thrombin.
Con clu sion s
SAR studies of these cycloalkanone series have shown
that the in vitro FXa potency and selectivity are strongly
dependent on the double bond geometry with the (Z,Z)
isomers significantly more potent than the (E,Z) iso-
mers. In general, the corresponding (E,E) isomers are
weakly active inhibitors of FXa. Comparison of a series
of (Z,Z) cycloalkanone analogues demonstrated that the
FXa inhibitory activity varies significantly with the ring
size with the cycloheptanone analogue (1c) as the most
potent and selective compound from this series. Throm-
bin selectivity was achieved through addition of a
carboxylate group to the 4-position of the cyclohexanone
ring to afford compound 8c (FXa Ki ) 6.9 nM, thrombin
Ki > 50 000 nM). In a series of (Z,Z)-cycloheptanone
analogues, it was found that modification of both
amidine substituents afforded inhibitors that were more
than 3 orders of magnitude less active than 1c, while
modification of only one amidine substituent led to an
18-75-fold loss of potency. The (Z,Z) geometric isomers
represent a new class of potent and selective inhibitors
of FXa, and on the basis of good in vitro potency and
selectivity, this conformationally rigid series has been
used as a template for the design of a variety of
chemically stable, potent, and selective inhibitors of FXa
will be reported in future publications.
Exp er im en ta l Section
Gen er a l Meth od s. Proton magnetic resonance spectra (1H
NMR) were recorded in CDCl3 or DMSO, as noted, on a Varian
300-MHz spectrometer (internal standard TMS, δ ) 0).
Elemental analyses were performed by Robertson Microlit
Laboratories, Inc., Madison, NJ . HPLC analyses and prepara-
tive purifications were carried out using reverse-phase columns
(Rainin, dynamax) using mixtures of acetonitrile and water,
both of which contained 0.1% trifluoroacetic acid. Typically,
the elution order for the three isomers was (E,E), (E,Z), and
(Z,Z) isomer with the notable exception of 8a , 8b, and 8c where
the order is (E,E), (Z,Z), and (E,Z). Photochemical irradiation
was carried out with a 150 W flood lamp or a medium-
pressure, quartz, mercury vapor, 450 W lamp purchased from
Ace Glass. Typically the irradiations were monitored by HPLC
and continued until an equilibrium mixture of isomers was
obtained.
Dou ble Bon d Isom er s of 2,7-Bis(4-am idin oben zyliden e)-
cycloocta n -1-on e (4a , 4b, a n d 4c). A suspension of 7 (1.0 g,
5.4 mmol) in 2.4 N hydrochloric acid was stirred as neat
cyclooctanone (0.34 g, 2.6 mmol) was added. The reaction was
heated at 100 °C for 8 h and allowed to cool over 10 h. The
slurry was concentrated, dissolved in methanol, treated with
charcoal, and filtered through Celite. Purification by reverse-
phase HPLC using 30% acetonitrile (0.1% TFA) in water (0.1%
TFA) yielded three compounds, 4a , 4b, and the monoalkylation
product. A solution of the (E,Z) isomer (4b, 0.15 g, 0.4 mmol)
in 20 mL of a 1:1 mixture of acetonitrile and water was
irradiated with
a 450 W lamp for 8 h. Purification by
Dou ble Bon d Isom er s of 2,7-Bis(4-am idin oben zyliden e)-
cycloh ep ta n -1-on e (1a , 1b, a n d 1c). A suspension of 7 (4.2g,
preparative HPLC using 30% acetonitrile (0.1% TFA) in water
(0.1% TFA) gave 4c, 40 mg, which was isolated as a white solid

5422 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 26
Guilford et al.
after lyophilization. 4a : NMR (DMSO, TMS) δ 1.62 (m, 2H),
1.75 (m, 4H), 2.72 (m, 4H), 7.08 (s, 2H), 7.65 (d, 4H), 7.87 (d,
4H), 9.13 (s, 4H), 9.36 (s, 4H). Anal. (C24H26N4O‚2.3C2HF3O2‚
0.9H2O) C, H, N, F. 4b: NMR (DMSO, TMS) δ 1.57 (m, 2H),
1.70 m, 4H), 2.56 (m, 4H), 6.70 (s, 1H), 7.65 (m, 9H). Anal.
(C24H26N4O‚1.9C2HF3O2‚H2O) C, H, N, F. 4c: NMR (DMSO,
TMS) δ 1.60 (m, 2H), 1.71 (m, 4H), 2.53 (m, 4H), 6.83 (s, 2H),
7.65 (m, 8H). Anal. (C24H26N4O‚2.5C2HF3O2‚1H2O) C, H, N,
F.
4-oxocyclohexancarboxylate (0.34 g, 2 mmol) and 7 (0.74 g, 4
mmol) in 5 mL of 4 N hydrochloric acid was heated at reflux
for 50 min. Desired product precipitated upon cooling to yield
8a , 0.4g, as a yellow solid. 8a : NMR (DMSO, TMS) δ 2.85 (m,
1H), 3.1 (m, 4H), 7.7 (s, 2H), 7.85 (m, 8H), 9.18 (s, 4H), 9.32
(s, 4H). Anal. (C23H22N4O3‚2HCl‚1.65H2O) C, H, N, Cl.
A solution of 9c (40 mg, 0.056 mmol) in 1 mL of aqueous
THF was stirred at 0 °C as 0.8 mL of an aqueous lithium
hydroxide solution (12.8 mg, 0.3 mmol) was added. After 30
min at 0 °C and 30 min at ambient temperature, the reaction
was acidified with trifluoroacetic acid. Purification by prepara-
tive HPLC using a 5-35% gradient of acetonitrile (0.1% TFA)
in water (0.1% TFA) gave 0.03 g of the corresponding acid.
8c: NMR (DMSO, TMS) δ 3.0 (m, 5H), 6.85 (s, 2H), 7.75 (m,
8H), 9.2 (m, 8H). Anal. (C23H22N4O3‚2.35C2HF3O2‚1H2O) C, H,
N, F.
Dou ble Bon d Isom er s of 2,7-Bis(4-am idin oben zyliden e)-
cyclon on a n -1-on e (5a , 5b, a n d 5c). A suspension of 7 (2.0
g, 10.8 mmol) in 2.4 N hydrochloric acid was stirred as neat
cyclononanone (0.6 mL, 5.4 mmol) was added. The reaction
was heated at 100 °C for 6 d and allowed to cool over 10 h.
The slurry was concentrated, dissolved in methanol, treated
with charcoal, and filtered through Celite. Product was
precipitated with the addition of 1 N hydrochloric acid and
purified by reverse-phase HPLC using 30% acetonitrile (0.1%
TFA) in water (0.1 % TFA) to give 0.42 g of 5a as a white solid
after lyophilization. A solution of 5a (0.42g, 0.6 mmol) in 15
mL of a 1:3 mixture of acetonitrile and water was irradiated
with a 450 W lamp for 24 h. Purification by preparative HPLC
using 30% acetonitrile (0.1% TFA) in water (0.1% TFA) yielded
0.1 g of 5b and 80 mg of 5c as white solids after lyophilization.
5a : NMR (DMSO, TMS) δ 1.38 (m, 4H), 1.63 (m, 4H), 2.78
(m, 4H), 7.10 (s, 2H), 7.70 (m, 4H), 7.8 (m, 4H), 9.05 (br s,
4H), 9.32 (m, 4H). Anal. (C25H28N4O‚2.1C2HF3O2‚H2O) C, H,
N, F. 5b: NMR (DMSO, TMS) δ 1.8 (m, 8H), 2.45 (m, 2H),
2.63 (m, 2H), 6.86 (s, 1H), 7.65 (m, 9H). Anal. (C25H28N4O‚2.3C2-
HF3O2‚1.3H2O) C, H, N, F. 5c: NMR (DMSO, TMS) δ 1.56
(m, 4H), 1.68 (m, 4H), 2.43 (m, 4H), 6.86 (s, 2H), 7.60 (m, 8H).
Anal. (C25H28N4O‚2.7C2HF3O2‚2H2O) C, H, N, F.
Dou ble Bon d Isom er s of 2,7-Bis(4-am idin oben zyliden e)-
cyclod eca n -1-on e (6a , 6b, a n d 6c). A suspension of 7 (2.0
g, 10.8 mmol) in 2.4 N hydrochloric acid was stirred as neat
cyclodecanone (0.8 mL, 5.4 mmol) was added. The reaction was
heated at 100 °C for 3 d and allowed to cool over 10 h. The
slurry was concentrated, dissolved in methanol, treated with
charcoal, filtered through Celite, and precipitated with ether.
Purification by reverse-phase HPLC using a gradient of 20-
75% acetonitrile (0.1% TFA) in water (0.1% TFA) gave the
(E,E) isomer (6a ), 0.5 g, after concentration and lyophilization.
A solution of 6a (0.5 g, 0.7 mmol) in 15 mL of a 2:1 mixture of
acetonitrile and water was irradiated with a 450 W lamp for
10 h. Purification by preparative HPLC using 30% acetonitrile
(0.1% TFA) in water (0.1% TFA) gave 0.36 g of 6b and 20 mg
of 6c as a white solid after lyophilization. 6a : NMR (DMSO,
TMS) δ 1.36 (m, 6H), 1.69 (m, 4H), 2.82 (m, 4H), 7.20 (s, 2H),
7.7 (m, 4H), 7.9 (m, 4H), 9.30 (s, 4H), 9.39 (s, 4H). Anal.
(C26H30N4O‚2.1C2HF3O2‚H2O) C, H, N, F. 6b: NMR (DMSO,
TMS) δ 1.5 (m, 10H), 2.82 (m, 2H), 2.63 (m, 2H), 7.16 (s, 1H),
7.65 (m, 9H), 9.5 (m, 8H). Anal. (C26H30N4O‚2.1C2HF3O2‚
0.5H2O) C, H, N, F. 6c: NMR (DMSO, TMS) δ 1.36 (m, 4H),
1.55 (m, 6H), 2.35 (m, 4H), 6.96 (s, 2H), 7.65 (m, 8H) ), 9.5 (m,
8H). Anal. (C26H30N4O‚3.2C2HF3O2‚H2O) C, H, N, F.
Dou ble Bon d Isom er s of 2,7-Bis(4-am idin oben zyliden e)-
4-m eth oxyca r boxy-cycloh exa n -1-on e (9a , 9b, a n d 9c). A
solution of 8a (0.52g, 1 mmol) in 50 mL of methanol was stirred
at ambient temperature for 3 d. Solvent was removed, and
residue was recrystallized from ethyl acetate to give 0.47 g of
9a as a white solid. A solution of 9a (0.52g, 1 mmol) in 300
mL of 50% aqueous acetonitrile containing 0.1% trifluoroacetic
acid was stirred and irradiated with a 450 W lamp for 4 h.
Purification by preparative HPLC using a 5-35% gradient of
acetonitrile (0.1% TFA) in water (0.1 % TFA) gave 0.12 g of
9b and 0.1 g of 9c after lyophilization. 9a : NMR (DMSO, TMS)
δ 2.95 (m, 1H), 3.1 (m, 4H), 3.58 (s, 3H), 7.7 (s, 2H), 7.85 (m,
8H), 9.3 (s, 4H), 9.56 (s, 4H). Anal. (C24H24N4O3‚2HCl‚2.4H2O)
C, H, N, Cl. 9b: NMR (DMSO, TMS) δ 2.95 (m, 1H), 3.1 (m,
4H), 3.62 (s, 3H), 6.95 (s, 1H), 7.48 (s, 1H), 7.6 (d, 2H), 7.72
(m, 4H), 7.88 (d, 2H), 9.1 (d, 4H), 9.36 (d, 4H). Anal.
(C24H24N4O3‚2C2HF3O2‚0.8H2O) C, H, N, F. 9c: NMR (DMSO,
TMS) δ 3.0 (m, 4H), 3.2 (m, 1H), 3.58 (s, 3H), 6.85 (s, 2H),
7.75 (m, 8H), 9.3 (m, 8H). Anal. (C24H24N4O3‚2.4C2HF3O2‚
1.5H2O) C, H, N, F.
Dou ble Bon d Isom er s of 2,7-Bis(4-am idin oben zyliden e)-
4-p h en yl-cycloh exa n -1-on e (10a , 10b, a n d 10c). A suspen-
sion of 7 (4.2 g, 22.8 mmol) and 4-phenylcyclohexanone (2 g,
11.5 mmol) in 45 mL of 2 N hydrochloric acid was stirred and
heated at 100 °C for 2 h. The reaction was allowed to cool,
diluted with 50 mL of methanol, and concentrated. The
resulting solid was recrystallized from ethanol to give 10a . A
mixture of 10a (0.5 g, 0.8 mmol) and 222 mL of a 50:50:1
mixture of acetonitrile, water, and trifluoroacetic acid was
irradiated with 450 W lamp for a total of 5 h. Purification by
preparative HPLC using a 20-45-95% gradient of acetonitrile
(0.1% TFA) in water (0.1% TFA) gave 40 mg of 10b and 150
mg of 10c as solids after lyophilization. 10a : NMR (DMSO,
TMS) δ 3.05 (m, 3H), 3.24 (m, 2H), 7.25 (m, 5H), 7.75 (m, 6H),
7.95 (m, 4H), 9.36 (s, 4H), 9.54 (s, 4H). Anal. (C28H26N4O-
EtOAc‚2.05HCl‚1.5H2O) C, H, N, Cl. 10b: NMR (DMSO, TMS)
δ 3.1 (m, 5H), 6.95 (s, 1H), 7.22 (m, 1H) 7.35 (m, 4H), 7.45 (s,
1H), 7.6 (d, 2H), 7.7 (d, 2H), 7.84 (m, 4H), 9.18 (d, 4H), 9.38
(d, 4H). Anal. (C28H26N4O‚2C2HF3O2‚1.5H2O) C, H, N, F. 10c:
NMR (DMSO, TMS) δ 3.1 (m, 5H), 6.9 (s, 2H), 7.22 (m, 1H)
7.4 (m, 8H), 7.7 (m, 4H), 9.24 (d, 4H), 9.42 (d, 4H). Anal.
(C28H26N4O‚2C2HF3O2‚0.3H2O) C, H, N, F.
4-Am id in oben za ld eh yd e (7). A solution of 4-cyanobenz-
aldehyde (50 g, 0.37 mol) and methanol (44 mL, 1.09 mol) in
200 mL of a 1:3 mixture of ether and p-dioxane was cooled in
an ice bath as hydrogen chloride gas was bubbled into the
solution. The reaction was allowed to warm to ambient
temperature over 16 h before solvent was removed by concen-
tration. The residue was dissolved in 350 mL of methanol,
saturated with ammonia gas, and heated at reflux for a total
of 1.5 h with additional ammonia added in portions. The
reaction was cooled to ambient temperature, the volume was
reduced, and diluted with 1 L of acetone. The solid ammonium
chloride was removed by filtration, and solvent was removed
under reduced pressure. The residue was dissolved in 0.2 N
aqueous hydrochloric acid solution, extracted with ethyl
acetate (4×), concentrated, and lyopholyzed to yield 25 g of a
white solid. 7: NMR (DMSO, TMS) δ 8.1 (m, 4H), 9.5 (s, 2H),
9.70 (s, 2H), 10.13 (s, 1H).
Dou ble Bon d Isom er s of 3,5-Bis(4-am idin oben zyliden e)-
N-eth oxyca r bon y-p ip er id -4-on e (1a , 11b, a n d 11c). A
suspension of 7 (3.26 g, 17.7 mmol) and N-ethoxycarbonyl-
piperidone (1 g, 5.8 mmol) in 45 mL of 2 N hydrochloric acid
was stirred and heated at reflux for 4 h. The reaction was
allowed to cool and was concentrated. Purification by precipi-
tation with ethyl acetate from a methanolic solution gave 2.1
g of 11a . A mixture of 11a (0.5 g, 0.9 mmol) in 402 mL of a
100:100:1 mixture of acetonitrile, water, and trifluoroacetic
acid was irradiated with a 450 W lamp for a total of 4 h.
Purification by preparative HPLC using a 20-45-95% gradi-
ent of acetonitrile (0.1% TFA) in water (0.1% TFA) gave 64
mg of 11b lyophilization and 50 mg of 11c as white solids after
lyophilization. 11a : NMR (DMSO, TMS) δ 0.9 (t, 3H), 3.84
(q, 2H), 4.75 (m, 4H), 7.65 (d, 4H), 7.75 (s, 2H), 7.9 (s, 4H),
9.06 (s, 4H), 9.39 (s, 4H). Anal. (C24H25N5O3‚0.2EtOAc‚2HCl‚
Dou ble Bon d Isom er s of 2,7-Bis(4-am idin oben zyliden e)-
4-ca r boxycycloh exa n -1-on e (8a a n d 8c). A slurry of ethyl

Synthesis of Selective Factor Xa Inhibitors
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 26 5423
0.5H2O) C, H, N, Cl. 11b: NMR (DMSO, TMS) δ 1.1 (m, 3H),
4.08 (m, 2H), 4.56 (s, 2H), 4.74 (s, 2H), 7.15 (s, 1H), 7.55 (s,
1H) 7.75 (m, 6H), 7.95 (d, 2H), 9.38 (m, 8H). Anal. (C24H25N5-
O3‚1.97C2HF3O2‚1.8H2O) C, H, N, F. 11c: NMR (DMSO, TMS)
δ 1.22 (t, 3H), 4.12 (q, 2H), 4.52 (m, 4H), 7.13 (s, 2H), 7.62 (d,
4H), 7.75 (d, 4H), 9.2 (s, 4H), 9.35 (s, 4H). Anal. (C24H25N5O3‚2C2-
HF3O2‚1.2H2O) C, H, N, F.
2-(4-Cya n oben zylid en e)cycloh ep ta n on e (12). A mixture
of 4-cyanobenzaldehyde (10 g, 76 mmol) and cycloheptanone
(29 mL, 76 mmol) in 100 mL of 85% phosphoric acid is heated
for 1.5 h at 100 °C. The mixture is cooled to room temperature
and filtered. The filtrate is poured into water (800 mL), and
the resulting solid is isolated by filtration, washed with water,
and dried. Purification by flash chromatography using a
mixture of ether:hexane (1:2) as the eluant afforded 4.8 g of
12. 12: NMR (CDCl3, TMS) δ 1.84 (m, 6H), 2.62 (m, 2H), 2.73
(m, 2H), 7.43 (m, 3H), 7.71 (d, 2H).
Dou ble Bon d Isom er s of 2-(4-Am id in oben zylid en e)-7-
(4-(N-m eth ylam idin oben zyliden e)-cycloh eptan -1-on e (13a,
13b, a n d 13c). A solution of 12 (2.00 g, 9 mmol) in a mixture
of anhydrous ethanol (17 mL) and anhydrous dioxane (34 mL)
was cooled to 0 °C and saturated with anhydrous hydrogen
chloride gas. The reaction mixture is allowed to stand for 60
h at 0-5 °C and then slowly added to 800 mL of a 3:1 mixture
of ether:hexane. The resulting solid is isolated by filtration
and dried in vacuo to afford the ethyl imidate as a solid. An
ether solution of ethyl imidate (1.5 g) was washed with 10%
aqueous sodium bicarbonate, dried, and concentrated. The
residue is added to a solution of methylamine hydrochloride
(0.42 g, 6.2 mmol) in 40 mL of methanol and heated for 3 h at
95 °C in a sealed vessel. The reaction mixture is cooled and
the solvent removed. Purification of the residue by flash
chromatography eluting with a 14:2:1:1 mixture of ethyl
acetate:methanol:water:aqueous ammonium hydroxide af-
forded 2-(N-methylamidinobenzylidene)-cycloheptan-1-one as
a solid.
A suspension of 7 in 15 mL of 85% phosphoric acid is stirred
as 2-(N-methylamidinobenzylidene)-cycloheptan-1-one (1.24 g,
4.3 mmol) is added. The mixture is heated for 3.5 h at 100 °C
and then cooled to room temperature. The reaction mixture is
then diluted with 20 mL of methanol and ether. The resulting
solid was isolated by decanting the liquid and purified by ether
precipitation from a methanolic hydrochloric acid solution to
afford 13a . A solution of 13a (0.27 g, 0.6 mmol) in methanol
(55 mL) was irradiated with a 450 W lamp for 3 h. The solution
is concentrated and purified by preparative reverse-phase
HPLC, eluting with a 1:4 mixture of acetonitrile (0.1% TFA)
in water (0.1% TFA) to give 0.1 g of 13b and 97 mg of 13c
after lyophilization. 13a : NMR (DMSO, TMS) δ 1.96 (m, 4H),
2.67 (m, 4H), 3.02 (m, 3H) 7.39 (s, 2H), 7.8 (m, 8H), 9.02 (s,
1H), 9.27 (s, 2H), 9.39 (s, 2H), 9.55 (s, 1H), 9.85 (m, 1H). Anal.
(C24H26N4O‚2C2HF3O2‚1.25H2O) C, H, N, F. 13b: NMR (DMSO,
TMS) δ 1.92 (m, 4H), 2.55 (m, 2H), 2.71 (m, 2H), 3.02 (m, 3H),
6.82 (s, 1H), 7.6 (m, 9H), 9.45 (m, 7H). Anal. (C24H26N4O‚2C2-
HF3O2‚2H2O) C, H, N. 13c: NMR (DMSO, TMS) δ 1.95 (m,
4H), 2.58 (m, 4H), 3.02 (m, 3H), 6.98 (s, 2H), 7.45 (m, 4H),
7.65 (m, 2H), 7.73 (m, 2H), (8.92 s, 1H), 9.15 (s, 2H), 9.34 (s,
2H), 9.42 (s, 1H), 9.78 (m, 1H). Anal. (C24H26N4O‚2.2C2HF3O2‚
1.5H2O) C, H, N, F.
A mixture of 7 (1.0 g, 5.4 mmol) in 15 mL of 85% phosphoric
acid was stirred as 2-(N,N-dimethylamidinobenzylidene)-cy-
cloheptan-1-one (1.1 g, 5.4 mmol) was added. The mixture is
heated for 3.5 h at 100 °C, then cooled to room temperature.
The reaction mixture is then diluted with 20 mL of methanol
and slowly added to a stirring ether solution. After 16 h, the
solvent is decanted. Purification by preparative reverse-phase
HPLC, eluting with a gradient of 15-40% of acetonitrile (0.1%
TFA) in water (0.1% TFA), gave 160 mg of the (E,E) isomer
(14a ) and 60 mg of 14b as solids. A solution of 14b (0.2 g, 0.6
mmol) in methanol (20 mL) was irradiated with a 450 W lamp
for 3 h. The solution is concentrated, and the residue purified
by preparative reverse-phase HPLC, eluting with 20% of
acetonitrile (0.1% TFA) in water (0.1% TFA). The (Z,Z) isomer
(14c), 20 mg, was isolated as an amber solid after lyophiliza-
tion. 14a : NMR (DMSO, TMS) δ 1.90 (m, 4H), 2.5 (m, 2H),
2.65 (m, 2H), 3.02 (s, 3H), 3.25 (s, 3H) 7.38 (s, 2H), 7.75 (m,
6H), 7.95 (d, 2H), 9.02 (s, 1H), 9.3 (m, 5H). Anal. (C25H28N4-
O‚3C2HF3O2‚0.3H2O) C, H, N, F. 14b: NMR (DMSO, TMS) δ
1.90 (m, 4H), 2.5 (m, 2H), 2.65 (m, 2H), 2.95 (s, 1.5H), 3.02 (s,
1.5H), 3.22 (s, 1.5H), 3.27 (s, 1.5H), 6.8 (s, 1H), 7.6 (m, 11H),
8.95 (m, 1H), 9.56 (m, 5H). Anal. (C25H28N4O‚2.2C2HF3O2‚
1.0H2O) C, H, N, F. 14c: NMR (DMSO, TMS) δ 1.97 (m, 4H),
2.61 (m, 4H), 2.98 (s, 3H), 3.22 (s, 3H), 6.95 (s, 1H) 7.00 (s,
1H), 7.65 (m, 6H), 7.76 (d, 2H), 8.93 (s, 1H), 9.3 (m, 5H). Anal.
(C25H28N4O‚2.5C2HF3O2‚2.5H2O) C, H, N.
Dou ble Bon d Isom er s of 2-(4-Am id in oben zylid en e)-7-
(4-car boxam ideben zyliden e)-cycloh eptan -1-on e (15a, 15b,
a n d 15c). A suspension of 7 (2.18 g, 11.8 mmol) in 16 mL of
85% phosphoric acid is stirred as 12 (2.66 g, 12 mmol) is added.
The mixture is heated for 4 h at 100 °C and then cooled to 50
°C. The reaction mixture is slowly poured into water (75 mL).
The resulting solid is isolated by filtration, washed sequen-
tially with methanol and ether, and dried in vacuo to afford
1.9 g of 15a as the phosphoric acid salt. The hydrochloride
salt is obtained by treating a methanolic suspension (30 mL)
with anhydrous hydrogen chloride gas and precipitating the
salt with ether (300 mL). The resulting precipitate is isolated
by filtration and dried in vacuo to afford 15a . A solution of
15a (0.50 g, 1.2 mmol) in methanol (5 mL) was irradiated with
a 450 W lamp for 3 h. The solution is concentrated and purified
by preparative reverse-phase HPLC, eluting with 23% of
acetonitrile (0.1% TFA) in water (0.1% TFA), to give 97 mg of
15b and 44 mg of 15c after lyophilization. 15a : NMR (DMSO,
TMS) δ 1.94 (m, 4H), 2.69 (m, 4H), 7.34 (s, 2H), 7.46 (s, 1H),
7.60 (d, 2H), 7.74 (d, 2H), 7.95 (m, 4H), 8.10 (s, 1H), 9.28 (s,
2H), 9.49 (s, 2H). Anal. (C23H23N3O2‚HCl‚0.5H2O) C, H, N, Cl.
15b: NMR (DMSO, TMS) δ 1.95 (m, 4H), 2.42 (m, 2H), 2.70
(m, 2H), 6.74 (s, 0.5H), 6.80 (s, 0.5H), 7.67 (m, 11 H), 9.42 (m,
4H). Anal. (C23H23N3O2‚1.2C2HF3O2‚H2O) C, H, N, F. 15c:
NMR (DMSO, TMS) δ 1.94 (m, 4H), 2.60 (m, 4H), 6.89 (s, 1H),
6.96 (s, 1H), 7.38 (d, 2H), 7.40 (s, 1H), 7.45 (d, 2H), 7.76 (m,
4H), 7.97 (s, 1H), 9.03 (s, 2H), 10.29 (s, 2H). Anal. (C23H23N3-
O2‚1.2C2HF3O2‚H2O) C, H, N.
Dou ble Bon d Isom er s of 2,7-Bis(4-a m in oca r boxyben -
zylid en e)cycloh ep ta n -1-on e (17b a n d 17c). A mixture of
4-cyanobenzaldehyde (20.0 g, 153 mmol) and cycloheptanone
(29 mL, 77 mmol) in 200 mL of 85% phosphoric acid is heated
for 1.5 h at 100 °C. The mixture is cooled to room temperature.
The resulting solid was isolated by filtration and washed with
water and methanol to afford 2,7-bis-(4-cyanobenzylidene)-
cycloheptanone (16). A solution of 16 (2.00 g, 6 mmol) in 15
mL of concentrated sulfuric acid is stirred for 3 h at room
temperature. The reaction mixture is poured into 200 mL of
water. The resulting solid was isolated by filtration, washed
sequentially with water and methanol, and dried to afford the
bis-benzamide, 17a , as a solid.
A suspension of 17a (0.60 g, 1.6 mmol) in methanol (40 mL)
was irradiated with a 450 W lamp for 3 h. Purification by
preparative reverse-phase HPLC, eluting with a gradient of
20-45% of acetonitrile (0.1% TFA) in water (0.1% TFA),
yielded 61 mg of 17b and 59 mg of 17c after drying. 17b: NMR
(DMSO, TMS) δ 1.82 (m, 4H), 2.42 (m, 2H), 2.70 (m, 2H), 6.73
(s, 1H), 7.26 (d, 2H), 7.37 (br s, 1H), 7.48 (br s, 1H) 7.56 (s,
Dou ble Bon d Isom er s of 2-(4-Am id in oben zylid en e)-7-
(4-(N,N-Dim et h yla m id in ob en zylid en e)-cycloh ep t a n -1-
on e (14a , 14b, a n d 14c). A solution of 12 (1.2 g, 5.3 mmol) in
100 mL of anhydrous ethanol was cooled to 0 °C and saturated
with anhydrous hydrogen chloride gas. The reaction mixture
is allowed to warm to room temperature over 16 h and
concentrated under reduced pressure. The residue was dis-
solved in 150 mL of absolute ethanol, treated with dimethyl-
amine, and heated at reflux for 1 h. The reaction mixture is
cooled and concentrated. The residue was dissolved in water,
acidified with 1 N HCl, washed with ether, and neutralized
with 1 N NaOH. Ether extraction gave 1.1 g of 2-(N,N-
dimethylamidinobenzylidene)-cycloheptan-1-one. NMR (DMSO,
TMS) δ 1.74 (m, 6H), 2.69 (m, 2H), 3.3 (m, 2H), 7.3 (s, 1H),
7.4 (m, 4H).

5424 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 26
Guilford et al.
1H), 7.61 (d, 2H), 7.80 (d, 2H), 7.96 (m, 3H), 8.08 (br s, 1H).
Anal. (C23H22N2O3) C, H, N. 17c: NMR (DMSO, TMS) δ 1.82
(m, 4H), 2.60 (m, 4H), 6.84 (s, 2H), 7.38 (m, 6H), 7.78 (d, 4H),
7.97 (br s, 2H). Anal. (C23H22N2O2‚0.3H2O) C, H, N.
Dou b le Bon d Isom er s of 2,7-Bis[4-(N,N-d im et h yl)-
a m id in oben zylid en elcycloh ep ta n -1-on e (18a , 18b, a n d
18c). A slurry of cycloheptanone (0.56 mL, 4.8 mmol) and 20
(2.04g, 9.6 mmol) in 15 mL of 3 N hydrochloric acid was heated
at 100 °C for 3 h. The mixture was concentrated and purified
by preparative reverse-phase HPLC to yield 190 mg of the
(E,E) isomer, 18a , and 30 mg of the (E,Z) isomer, 18b. A
solution of 18b (50 mg, 0.1 mmol) in methanol (15 mL) was
at room temperature (∼23 °C) in a THERMOmax 96-well
microplate reader (Molecular Devices, Sunnyvale, CA). All
substrates were Chromogenix substrates from Kabi Pharmacia
Hepar, Inc. (Franklin, OH). The final concentration of enzyme
and substrate used are as follows: factor Xa (1 nM) and S-2222
(164 µM), thrombin (16 nM) and S-2302 (300 µM), trypsin (16
nM) and S-2266 (127 µM), tPA (10 nM) and S-2288 (371 µM),
and uPA (2 nM) and S-2444 (32 µM). These bisamidine-type
inhibitors were found to have competitive mechanism against
the arginine containing substrates used (data not shown);
therefore, the Km for each substrate was determined in these
assay conditions, and the substrate concentration equal to Km
was used to assay varying concentrations of the inhibitors to
determine the Ki. In the case of the very potent compound 1c,
the Ki for factor Xa was determined by both decreasing the
enzyme concentration to 0.1 nM and by fitting data obtained
at 0.3 and 1 nM enzyme to a modification of the Morrison
equation.17
irradiated with
a 450 W lamp for 3 h. Purification by
preparative reverse-phase HPLC, eluting with a gradient of
20-45% of acetonitrile (0.1% TFA) in water (0.1% TFA),
yielded 10 mg of 18c after drying. 18a : NMR (DMSO, TMS)
δ 1.92 (m, 4H), 2.7 (m, 4H), 3.0 (s, 6H), 3.2 (s, 6H), 7.33 (s,
2H), 7.68 (m, 8H), 9.1 (s, 2H), 9.4 (s, 2H). Anal. (C27H32N4O‚
3.2C2HF3O2‚H2O) C, H, N, F. 18b: NMR (DMSO, TMS) δ 1.92
(s, 4H), 2.5 (s, 2H), 2.76 (s, 2H), 2.98 (s, 3H), 3.2 (m, 3H), 3.22
(s, 3H), 3.26 (s, 3H), 6.78 (s, 1H), 7.46 (m, 4H), 7.6 (s, 1H),
7.68 (m, 4H), 8.92 (s, 2H), 9.3 (s, 2H). Anal. (C27H32N4O‚3C2-
HF3O2) C, H, N. 18c: NMR (DMSO, TMS) δ 1.82 (m, 4H), 2.56
(m, 4H), 3.0 (s, 6H), 3.2 (s, 6H), 6.93 (s, 2H), 7.54 (m, 8H),
8.92 (s, 2H), 9.3 (s, 2H). Anal. (C27H32N4O‚2.75C2HF3O2) C,
H, N, F.
Ack n ow led gm en t. We thank Drs. Les Browne,
William Dole, J ohn Morser, and Mark Sullivan for their
support and encouragement of this effort and Dr. Gary
Phillips for helpful discussions in preparation of this
manuscript.
Refer en ces
Dou ble Bon d Isom er s of 2,7-Bis(3-am idin oben zyliden e)-
cyclon oh ep ta n -1-on e (19a , 19b, a n d 19c). A suspension of
3-amidinobenzaldehyde (2.32 g, 12.5 mmol) in 12 mL of 85%
phosphoric acid is stirred as cycloheptanone (0.70 g, 6.2 mmol)
was added. The mixture is heated for 3.5 h at 100 °C, then
cooled to room temperature. The dark red reaction mixture is
diluted with methanol (20 mL) and ether (75 mL). The
resulting solid was isolated by filtration, washed with methanol/
ethyl acetate (1:1), and dried in vacuo to afford 1.2 g of 19a .
The hydrochloride salt was obtained by treatment of a metha-
nolic solution with anhydrous hydrogen chloride gas and
isolated by filtration. A solution of 19a (0.4 g, 0.9 mmol) in
500 mL of methanol was irradiated with a 150 W lamp for 14
d. Purification by preparative HPLC using a 15-22% gradient
of acetonitrile (0.1% TFA) in water (0.1% TFA) gave 27 mg of
19b and 57 mg of 19c after lyophilization. 19a : NMR (DMSO,
TMS) δ 1.95 (m, 4H), 2.71 (m, 4H), 7.33 (s, 2H), 7.8 (m, 6H),
7.93 (s, 2H), 9.38 (s, 4H), 9.53 (s, 4H). Anal. (C23H24N4O‚2HCl)
C, H, N, Cl. 19b: NMR (DMSO, TMS) δ 1.92 (m, 4H), 2.50
(m, 2H), 2.76 (m, 2H) 6.80 (s, 1H), 7.70 (m, 9H), 9.19 (s, 4H),
9.38 (d, 4H). Anal. (C23H24N4O‚2C2HF3O2‚1.4H2O) C, H, N.
19c: NMR (DMSO, TMS) δ 1.92 (m, 4H), 2.62 (m, 4H), 6.95
(s, 2H), 7.66 (m, 8H), 9.23 (s, 4H), 9.38 (s, 4H). Anal.
(C23H24N4O‚2C2HF3O2‚H2O) C, H, N.
4-(N,N-Dim eth yla m id in o)ben za ld eh yd e (20). A solution
of 4-(methoxyimidate)benzaldehyde (10 g, 50 mmol) in 18 mL
of trimethylorthoformate and 100 mL of methanol was stirred
for 20 h. The reaction was filtered into ether (500 mL). The
resulting precipitate was isolated by filtration to give 5.6 g of
the corresponding acetal hydrochloride salt after drying. An
ethanolic solution of the acetal hydrochloride salt (5 g) was
treated with dimethylamine and heated at 90 °C for 2.5 h.
Solvent was removed under reduced pressure. The residue was
dissolved in 100 mL of 1 N hydrochloric acid and was heated
at 60 °C for 3 h. The reaction was concentrated and lyophi-
lyzed to yield 2.8 g of 20. 20: NMR (DMSO, TMS) δ 2.95 (s,
3H), 3.25 (s, 3H), 7.8 (d, 2H), 8.10 (d, 2H), 9.4 (s, lH), 9.58 (s,
1H), 10.12 (s, 1H).
(1) Some of this data has been previously reported: Guilford, W.
J .; Shaw, K. J .; Dallas, J .; Koovakkat, S.; Liang, A.; Light, D.;
Morrissey, M. M. Design, Synthesis, and Activity of Novel Factor
Xa Inhibitors I. (Z,Z)-Substituted Bis(amidino-benzylidene)-
cycloketone Analogues. Presented at the 215th American Chemi-
cal Society National Meeting, March 29-April 2, 1998, Dallas,
TX; abstract #121.
(2) A communication introducing this material has been pub-
lished: Shaw, K. J .; Guilford, W. J .; Dallas, J . L.; Koovakaat, S.
K.; McCarrick, M. A.; Liang, A.; Light, D. R.; Morrissey, M. M.
(Z,Z)-2,7-Bis-(4-amidinobenzylidene)-cycloheptan-1-one: iden-
tification of a highly active inhibitor of blood coagulation factor
Xa. J . Med. Chem. 1998, 41, 3551-3556.
(3) Mann, K. G.; Nesheim, M. E.; Church, W. R.; Haley, P.;
Krishnaswamy, S. Surface-dependent reactions of the vitamin
K-dependent enzyme complexes. Blood 1990, 76, 1-16.
(4) Eisenberg, P. R.; Siegel, J . E.; Abendschein, D. R.; Miletich J .
P. Importance of factor Xa in determining the procoagulant
activity of whole-blood clots. J . Clin. Invest. 1993, 91, 1877-
1883.
(5) Prager, N. A.; Abendschein, D. R.; McKenzie, C. R.; Eisenberg,
P. R. Role of thrombin compared with factor Xa in the pro-
coagulant activity of whole blood clots. Circulation 1995, 92,
962-967.
(6) (a) McKenzie, C. R.; Abendschein, D. R.; Eisenberg, P. R.
Sustained inhibition of whole-blood clot procoagulant activity
by inhibition of thrombus-associated factor Xa. Arterioscler.
Thromb. Vasc. Biol. 1996, 16, 1285-1291. (b) Mellott, M. J .;
Holahan, M. A.; Lynch, J . J .; Vlasuk, G. P.; Dunwiddie, C. T.
Acceleration of recombinant tissue-type plasminogen activator-
induced reperfusion and prevention of reocclusion by recombi-
nant antistasin, a selective factor Xa inhibitor, in a canine model
of femoral arterial thrombosis. Circ. Res. 1992, 70, 1152-1160.
(c) Nicolini, F. A.; Lee, P.; Malycky, J . L.; Lefkowitz, J .; Kottke-
Marchant, K.; Plow, E. F.; Topol, E. J . Selective inhibition of
factor Xa during thrombolytic therapy markedly improves
coronary artery patency in a canine model of coronary throm-
bosis. Blood Coag. Fibrinol. 1996, 7, 39-48.
(7) (a) Nagahara, T.; Yokoyama, Y.; Inamura, K.; Katakura, S.;
Komoriya, S.; Yamaguchi, H.; Hara, T.; Iwamoto, M. Dibasic
(amidinoaryl)propanoic acid derivatives as novel blood coagula-
tion factor Xa inhibitors. J . Med. Chem. 1994, 37, 1200-1207.
(b) Phillips, G. B.; Buckman, B. O.; Davey, D. D.; Eagen, K. A.;
Guilford, W. J .; Hinchman, J .; Ho, E.; Koovakaat, S.; Liang, A.;
Light, D. R.; Mohan, R.; Ng, H. P.; Post, J . M.; Shaw, K. J .;
Smith, D.; Subramanyam, B.; Sullivan, M. E.; Trinh, L.; Ver-
gona, R.; Walters, J .; White, K.; Whitlow, M.; Wu, S.; Xu, W.;
Morrissey, M. M. Discovery of N-[2-[5-[Amino(imino)methyl]-2-
hydroxyphenoxy]-3,5-difluoro-6-[3-(4,5-dihydro-1-methyl-1H-imi-
dazol-2-yl)phenoxy]pyridin-4-yl]-N-methylglycine (ZK-807834):
A Potent, Selective and Orally Active Inhibitor of the Blood
Coagulation Enzyme Factor Xa. J . Med. Chem. 1998, 41, 3557-
3562. (c) Sato, K.; Kawasaki, T.; Hisamichi, N.; Taniuchi, Y.;
Hirayama, F.; Koshio, H.; Matsumoto, Y. Antithrombotic effects
of YM-60828, a newly synthesized factor Xa inhibitor, in rat
thrombosis models and its effects on bleeding time. British J .
Biologica l Mea su r em en ts. Ser in e P r otea se In h ibition
Assa ys. Care was taken during the assays to minimize
exposure of cyclic ketone analogue solutions to light. Human
factor Xa and human thrombin were from Enzyme Research
Laboratories, Inc. (South Bend, IN), bovine trypsin was from
Roche Molecular Biochemicals (Indianapolis, IN), and tissue
plasminogen activator (tPA) and urokinase (uPA) were both
from Sigma (St. Louis, MO). All five serine proteases were
assayed in 150 mM NaCl, 2.5 mM CaCl2, 50 mM Tris-HCl,
pH 7.5, and 0.1% PEG 6000, in a final assay volume of 200 µL

Synthesis of Selective Factor Xa Inhibitors
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 26 5425
Pharm. 1998, 123, 92-96. (d) Klein, S. I.; Czekaj, M.; Gardner,
C. J .; Guertin, K. R.; Cheney, D. L.; Spada, A. P.; Bolton, S. A.;
Brown, K.; Colussi, D.; Heran, C. L.; Morgan, S. R.; Leadley, R.
J .; Dunwiddie, C. T.; Perrone, M. H.; Chu, V. Identification and
initial structural-activity relationships of a novel class of non-
peptide inhibitors of blood coagulation factor Xa. J . Med. Chem.
1998, 41, 437-450. (e) Maduskuie, T. P.; McNamara, K. J .; Ru,
Y.; Knabb, R. M.; Stouten, R. W. Rational design and synthesis
of novel, potent bis-phenylamidine carboxylate factor Xa inhibi-
tors. J . Med. Chem. 1998, 41, 53-62. (f) Katz, B. A.; Clark, J .
M.; Finer-Moore, J . S.; J enkins, T. E.; J ohnson, C. R.; Ross, M.
J .; Luong, C.; Moore, W. R.; Stroud, R. M. Design of potent
selective zinc-mediated serine protease inhibiors. Nature 1998,
391, 608-612.
Conformationally Restrained Cinnamoyl Compounds as Inhibi-
tors of HIV-1 Integrase: Synthesis, Biological Evalation, and
Molecular Modeling. J . Med. Chem. 1998, 41, 3948-3960. (d)
Dimmock, J . R.; Kandepu, N. M.; Nazarali, A. J .; Kowalchuk,
T. P.; Motaganahalli, N.; Quail, J . W.; Mykytiuk, P. A.; Audette,
G. F.; Prasad, L.; Perjesi, P.; Allen, T. M.; Santos, C. L.;
Szydlowski, J .; De Clercq, E.; Balzarini, J . Conformational and
Quantitative Structure-Activity Relationship Study of Cytotoxic
2-Arylidenebenzocycloalkanones J . Med. Chem. 1999, 42, 1358-
1366.
(16) Prepared in a manner similar to 7.
(17) Ki values for tight-binding inhibitors (inhibitors with Ki values
less than 3 nM) were analyzed with the Morrison equation to
correct for the proportion of inhibitor bound to the enzyme
relative to the free inhibitor concentration. Morrison, J . F.
Kinetics of the reversible inhibition of enzyme-catalyzed reac-
tions by tight-binding inhibitors. Biochem. Biophys. Acta 1969,
185, 269-286.
(18) Renatus, M.; Bode, W.; Huber, R.; Sturzebecher, J .; Stubbe, M.
T. Structural and Functional Analyses of Benzamidine-Based
Inhibitors in Complex with Trypsin: Implications for the Inhibi-
tion of Factor Xa, tPA, and Urokinase J . Med. Chem. 1998, 41,
5445-5456.
(19) Renatus, M.; Bode, W.; Huber, R.; Sturzebecher, J .; Prasa, D.;
Fischer, S.; Kohnert, U.; Stubbs, M. T. Structural Mapping of
the Active Site Specificity Determinants of Human Tissue-type
Plasminogen Activator. Implications for the Design of Low
Molecular Weight Substrates and Inhibitors. J . Biol. Chem.
1997, 272, 21713-21719.
(20) Spraggon, G.; Phillips, C.; Nowak, U. K.; Ponting, C. P.;
Saunders, D.; Dobson, C. M.; Stuart, D. I.; J ones, E. Y. The
Crystal Structure of the Catalytic Domain of Human Urokinase-
type Plasminogen Activator. Structure 1995, 3 (7), 681-691.
(21) (a) Bode, W.; Schwager, P. The refined crystal structure of bovine
beta-trypsin at 1.8 Å resolution. II. Crystallographic refinement,
calcium binding site, benzamidine binding site and active site
at pH 7.0. J . Mol. Biol. 1975, 98, 693-717. (b) Marquart, M.;
Walter, J .; Deisenhofer, J .; Bode, W.; Huber, R. The geometry
of the active site and of the peptide groups in trypsin, trypsi-
nogen and its complexes with inhibitors. Acta. Crystallogr. Sect.
B. 1983, 39, 480-490. (c) Banner, D. W.; Hadvary, P. Crystal-
lographic analysis at 3.0 Å - Resolution of the binding to human
thrombin of four active site-directed inhibitors. J . Biol. Chem.
1991, 266, 20085-20091.
(8) Ikeda, Y.; Fujita, M.; Murayama, N.; Yamada, H.; Inoue, T.;
Sasano, H.; Kunitada, S. Human pharmacokinetics, pharmaco-
dynamics, and tolerability of activated factor X inhibitor DX-
9065a. Circulation 1998, 98, 1454.
(9) Hara, T.; Yokoyama, A.; Ishihara, H.; Yokoyama, Y.; Nagahara,
T.; Iwamoto, M. DX-9065a, a new synthetic, potent anticoagulant
and selective inhibitor for factor Xa. Thromb. Haemost. 1994,
71, 314-319.
(10) (a) Hara, T.; Yokoyama, A.; Tanabe, K.; Ishihara, H.; Iwamoto,
M. DX-9065a, an orally active, specific inhibitor of factor Xa,
inhibits thrombosis without affecting bleeding time in rats.
Thromb. Haemost. 1995, 74, 635-639. (b) Yokoyama, T.; Kelly,
A. B.; Marzee, U. M.; Hanson, S. R.; Kunitada, S.; Harker, L. A.
Antithrombotic effects of orally active synthetic antagonist of
activated factor X in nonhuman primates. Circulation 1995, 92,
485-491.
(11) (a) Light, D.; Liang, A.; Rumennik, G.; Clarke, J .; McLean, K.;
Post, J .; Trinh, L.; Phillips, G.; Davey, D. D.; Guilford, W. J .;
Morser, J .; Morrissey, M. M.; Sullivan, M. E. In vitro inhibition
of factor Xa and prothrombinase by BX-807834, a potent,
selective factor Xa inhibitor. FASEB J . 1998, 12, A719 #4170.
(b) Subramanyam, B.; Ho, E.; Walters, J .; Cheesman, S.;
Guilford, W. J .; Vergona, R.; White, K.; Phillips, G.; Davey, D.
D.; Morrissey, M. M.; Light, D.; Morser, J .; Sullivan, M. E. The
pharmacokinetics of BX-807834, a potent, selective factor Xa
inhibitor in animals. FASEB J . 1998, 12, A719 #4169. (c)
Sullivan, M. E.; Morser, J .; Light, D.; Guilford, W. J .; Vergona,
R.; Herrmann, M.; White, K.; Phillips, G.; Davey, D. D.; Mor-
rissey, M. M.; Fredrich, M.; Baldus, B. The efficacy and safety
of BX-8107834, a potent, selective factor Xa inhibitor, in animal
models. FASEB J . 1998, 12, A-719 #4171.
(22) The AMBER program package was used for all molecular
dynamics simulations (Case, D. A.; Pearlman, D. A; Caldwell,
J . W.; Cheatham, T. E.; Ross, W. S.; Simmerling, C. L.; Darden,
T. A.; Merz, K. M.; Stanton, R. V.; Cheng, A. L.; Vincent, J . J .;
Crowley, M.; Ferguson, D. M.; Radmer, R. J .; Seibel, G. L.; Singh,
U. C.; Weiner, P. K.; Kollman, P. A. AMBER 5; University of
California: San Francisco, 1997) with the parm94 force field
(Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K.
M.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J . W.;
Kollman, P. A. A second generation force field for the simulation
of proteins, nucleic acids, and organic molecules. J . Am. Chem.
Soc. 1995, 117, 5179-5197.) Additional parameters were devel-
oped for the inhibitors (see ref 2). All simulations consisted of
molecular dynamics of the active site region, inhibitor, and
explicit solvent, with the remainder of the protein fixed. A time
step of 0.002 ps was used with all bond lengths constrained to
their equilibrium values. Simulations were either 200 ps or 500
ps in length, as noted in the text. A cutoff of 9.0 A was used for
nonbonded interactions. The initial coordinates for thrombin and
FXa were from the Brookhaven protein databank entries 1AHT
and 1HCG, respectively.
(12) Baum, P.; Light, D.; Verhallen, P.; Sullivan, M.; Eisenberg, P.;
Abendschein, D. Antithrombotic effects of a novel antagonist of
factor Xa. Circulation 1998, 98, 1800.
(13) (a) Phillips, G.; Davey, D. D.; Eagen, K.; Koovakkat, S. K.; Liang,
A.; Ng, H. P.; Pinkerton, M.; Trinh, L.; Whitlow, M.; Beatty, A.
M.; Morrissey, M. M. Design, Synthesis, and Activity of 2,6-
Diphenoxypyridine-Derived Factor Xa Inhibitors J . Med. Chem.
1999, 42, 1749-1756. (b) Buckman, B. O.; Mohan, R.; Koo-
vakkat, S. K.; Liang, A.; Trinh, L.; Morrissey, M. M. Design,
Synthesis, and Biological Activity of Novel Purine and Bicyclic
Pyrimidine Factor Xa Inhibitors. Bioorg. Med. Chem. Lett. 1998,
2235-2240.
(14) (a) Stuerzebecher, J .; Markwardt, F.; Walsmann, P. Synthetic
Inhibitors of Serine Proteinases XXIII. Inhibition of factor Xa
by diamidines. Thrombosis Res. 1980, 17, 545-548. (b) Stuer-
zebecher, J .; Struerzebecher, U.; Vieweg, H.; Wagner, H.; Haupt-
mann, J .; Markwardt, F. Synthetic Inhibitors of bovine factor
Xa and thrombin. Comparison of their anticoagulant efficiency.
Thrombosis Res. 1989, 54, 245-252.
(15) (a) George, H.; Roth, H. J . Photoisomerisierung und cyclo-1,2-
addition R,â-ungesaettigter cyclanonen. Tetrahedron Lett. 1971,
43, 4057-4060. (b) Aizenshtat, Z.; Hausmann, M.; Pickholtz, Y.;
Tai, D.; Blum, J . Chlorocarbonylbis(triphenylphosphine)iridium-
catalyzed isomerization, isoaromatization, and disproportion-
ation of some cycloalkanones having exocyclic double bonds. J .
Org. Chem. 1977, 42, 2386-2394. (c) Artico, M.; Di Santo, R.;
Costi, R.; Novellino, E.; Greco, G.; Massa, S.; Tramontano, E.;
Marongiu, M. E.; De Montis, A.; La Colla, P. Geometrically and
(23) Katakura, S.; Nagahara, T.; Hara, T.; Iwamoto, M. A Novel
Factor Xa Inhibitor: Structure-activity Relationships and Se-
lectivity Between Factor Xa and Thrombin. Biochem. Biophys.
Res. Commun. 1993, 197 (2), 965-972.
(24) LogD calculations were performed using LogD Predictor Version
4.0 - demo from Advanced Chemistry Development Inc.
J M990456V