Development of serine protease inhibitors displaying a multicentered short (<2.3 ?) hydrogen bond binding mode: Inhibitors of urokinase-type plasminogen activator and factor Xa

Article abstract of DOI:10.1021/jm0100638

Novel scaffolds that bind to serine proteases through a unique network of short hydrogen bonds to the catalytic Ser195 have been developed. The resulting potent serine protease inhibitors were designed from lead molecule 2-(2-hydroxyphenyl)1H-benzoimidazo

Full text of DOI:10.1021/jm0100638

J . Med. Chem. 2001, 44, 2753-2771
2753
Develop m en t of Ser in e P r otea se In h ibitor s Disp la yin g a Mu lticen ter ed Sh or t
(<2.3 Å) Hyd r ogen Bon d Bin d in g Mod e: In h ibitor s of Ur ok in a se-Typ e
P la sm in ogen Activa tor a n d F a ctor Xa
Erik Verner,† Bradley A. Katz,‡ J effrey R. Spencer,† Darin Allen,† J ason Hataye,† Witold Hruzewicz,†
Hon C. Hui,† Aleksandr Kolesnikov,† Yong Li,† Christine Luong,‡ Arnold Martelli,† Kesavan Radika,§ Roopa Rai,†
Miles She,† William Shrader,† Paul A. Sprengeler,‡ Sean Trapp,† J ing Wang,§ Wendy B. Young,† and
Richard L. Mackman*,†
Departments of Medicinal Chemistry, Structural Biology, and Biochemistry and Enzymology, Axys Pharmaceuticals Inc.,
180 Kimball Way, South San Francisco, California 94080
Received February 8, 2001
Novel scaffolds that bind to serine proteases through a unique network of short hydrogen bonds
to the catalytic Ser195 have been developed. The resulting potent serine protease inhibitors
were designed from lead molecule 2-(2-hydroxyphenyl)1H-benzoimidazole-5-carboxamidine, 6b,
which is known to display several modes of binding. For instance, 6b can recruit zinc and bind
in a manner similar to that reported by bis(5-amidino-2-benzimidazolyl)methane (BABIM)
(Nature 1998, 391, 608-612).1 Alternatively, 6b can bind in the absence of zinc through a
multicentered network of short (<2.3 Å) hydrogen bonds. The lead structure was optimized in
the zinc-independent binding mode toward a panel of six human serine proteases to yield
optimized inhibitors such as 2-(3-bromo-2-hydroxy-5-methylphenyl)-1H-indole-5-carboxamidine,
22a , and 2-(2-hydroxybiphenyl-3-yl)-1H-indole-5-carboxamidine, 22f. Structure-activity rela-
tionships determined that, apart from the amidine function, an indole or benzimidazole and
an ortho substituted phenol group were also essential components for optimal potency. The
affinities (Ki) of 22a and 22f, for example, bearing these groups ranged from 8 to 600 nM toward
a panel of six human serine proteases. High-resolution crystal structures revealed that the
binding mode of these molecules in several of the enzymes was identical to that of 6b and
involved short (<2.3 Å) hydrogen bonds among the inhibitor hydroxyl oxygen, Ser195, and a
water molecule trapped in the oxyanion hole. In summation, novel and potent trypsin-like
serine protease inhibitors possessing a unique mode of binding have been discovered.
In tr od u ction
among the closely related enzymes, and also possess
adequate pharmacodynamic properties for the target of
interest.
Trypsin-like serine proteases are a family of pro-
teolytic enzymes that have been implicated in many
important biological processes, including blood coagula-
tion, digestion, inflammation, metastasis, and cellular
invasion.2 Regulation of their biological action through
small molecule inhibitors is therefore a potentially
attractive way of treating many disease states. Perhaps
the most actively researched member of the family with
respect to the design of small molecule inhibitors is
thrombin.3 Other enzymes in the family, such as factor
Xa, factor VIIa, tryptase, and, more recently, urokinase-
type plasminogen activator (u-PA), are also receiving
considerable attention.3-7 Despite extensive efforts span-
ning several decades, there are few small molecule
inhibitors of trypsin-like serine proteases widely avail-
able as drugs. One reason for the dearth of serine
protease-targeted drugs is the criteria required for a
therapeutically useful small molecule inhibitor of these
enzymes. Inhibitors need to be highly potent, selective
An early approach toward the design of potent inhibi-
tors for serine proteases has been to develop molecules
containing electrophiles, e.g., activated ketones, esters,
aldehydes, boronic acids or esters, that achieve potency
through the formation (or likely formation) of covalent
adducts with the Ser195 or His57 (chymotrypsin num-
bering system) residues of the catalytic triad.8 Figure 1
outlines some examples, 1-4, of inhibitors of u-PA and
other serine proteases that have been noted in the
literature.6,9-11 The covalent interaction tends to pro-
duce highly potent inhibitors which, in some circum-
stances, leads to very low rates of dissociation and
essentially irreversible inhibition.6 Selectivity in this
type of inhibitor becomes critical and is typically gener-
ated through the elaboration of the chemical groups that
occupy the subsites (S4-S3′) around the catalytic
machinery. The mere presence of the requisite electro-
philic moieties in such inhibitors is viewed as detrimen-
tal to the identification of a drug since the chances of
obtaining selectivity, oral availability, and minimal
toxicity are severely compromised, prompting the search
for noncovalent reversible inhibitors.
* Corresponding author: Richard Mackman, Axys Pharmaceuticals
Inc., 180 Kimball Way, South San Francisco, CA 94080. Tel:
(650) 829-1191. Fax: (650) 829-1019. E-mail: richard_mackman@
axyspharm.com.
† Department of Medicinal Chemistry.
There are relatively few reports describing inhibitors
that specifically and rationally exploit noncovalent
‡ Department of Structural Biology.
§ Department of Biochemistry and Enzymology.
10.1021/jm0100638 CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/12/2001

2754 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17
Verner et al.
F igu r e 1. Inhibitors of u-PA and other proteases that contain electrophilic centers.
ments the zinc-mediated inhibitors. The molecule, 2-[2-
hydroxyphenyl]-1H-benzimidazole-5-carboxamidine, 6b
(Table 1), was found using high-resolution X-ray crys-
tallography to display two distinct modes of binding in
trypsin.14 One mode is mediated through recruitment
of a zinc ion similar to that described for BABIM, while
the second mode involves the formation of a unique
network of hydrogen bonds. The hydrogen bonds that
are formed among Ser195, the inhibitor hydroxyl oxy-
gen, and a water molecule trapped in the oxyanion hole
are unusually short (2.0-2.3 Å) compared to ordinary
hydrogen bonds (>2.6 Å). Exploration of structure-
activity relationships (SAR) has resulted in optimization
of the lead 6b into several inhibitors, for example, 22a
and 22f, that display affinities (Ki) in the 8-600 nM
range toward a panel of six common human serine
proteases (Table 2). The binding of the optimized
inhibitors was demonstrated to be through an identical
multicentered network of short and normal range
hydrogen bonds. To our knowledge, the unique multi-
centered short hydrogen-bonding network displayed by
these inhibitors is unprecedented in the literature.
Thus, a complementary class of potent serine protease
inhibitors based on the scaffold 6b, displaying a unique
binding mode, has now evolved. The inhibitor 22f is
currently being exploited for the development of potent
and specific inhibitors of u-PA whereas the alternative
inhibitor, 22a , has been optimized into a series of
selective factor Xa inhibitors.15 This report focuses on
the synthesis and potency optimization of the lead
scaffold 6b that led to the identification of inhibitors
22a and 22f along with the structural characterization
of the short hydrogen bond-mediated binding mode.
F igu r e 2. Schematic of BABIM binding to the catalytic
Ser195 and His57 residues through recruitment of a zinc ion.
interactions with the common catalytic residues of
serine proteases. Recently, a new and unusual high
affinity binding mode for the serine protease inhibitor
bis(5-amidino-2-benzimidazolyl)methane (BABIM) was
identified whereby a zinc ion can be recruited to mediate
binding between the inhibitor and the enzyme (Figure
2).1 The zinc ion is tetrahedrally coordinated between
two chelating nitrogen atoms of the inhibitor and the
active site residues, Ser195 and His57. Molecules de-
signed from the BABIM scaffold have displayed up to 5
orders of magnitude greater affinity toward enzymes in
the serine protease family when assayed in the presence
of zinc ions compared to zinc-free conditions (EDTA
added).12 As a consequence, highly potent and selective
zinc-dependent inhibitors for important targets such as
factor Xa and tryptase have been developed.12 A poten-
tial obstacle to this approach of enzyme inhibition for
future drug molecules is the high dependence on zinc
for tight binding. Insufficient free zinc ions in the local
environment of the target or a slow rate of complex
formation may not be compatible for some drug targets.
Sulfate ions have also been shown to mediate binding
between a small molecule inhibitor and the catalytic
machinery of trypsin.1,13 The affinity enhancements in
the sulfate-mediated binding mode are not as dramatic
as those observed for the zinc-mediated binding mode.
We have sought to engineer inhibitors analogous to
the zinc-mediated inhibitors that can bind to the
catalytic residues through direct, rather than zinc-
mediated, interactions. These inhibitors would poten-
tially eliminate some of the kinetic and physiological
concerns surrounding the zinc-mediated inhibitors and
thus lead to a suite of compounds that nicely comple-
Ch em istr y
A typical procedure for the formation of 5-amidino-
benzimidazoles involves the condensation of 3,4-diamino-
benzamidine or 3,4-diaminobenzonitrile with either an
imidate16,17 or an acid,17,18 followed by conversion of the
nitrile to an amidine using a modified Pinner reaction.16
Two more recent reports19 have utilized the condensa-
tion of 3,4-diaminobenzamidine with a benzaldehyde
followed by oxidation of the intermediate dihydrobenz-
imidazole to yield 5-amidinobenzimidazoles. A combina-
tion of some of these reported methods (general methods

Serine Protease Inhibitors with a Unique Binding Mode
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17 2755
Sch em e 1a
a
Reagents and conditions: General method A: EtOH, reflux, 10-50% yield. General method B: N,N-dimethylacetamide, Na2S2O5,
100 °C, 22% yield. General method C: 1,4-benzoquinone, EtOH, reflux, 89% yield. General method D: PPA, 180 °C, 10-30% yield.
Ta ble 1. Effect of Aryl Substitution on Enzyme Inhibition
R
Ki (µM)
thrombin
no.
2′
3′
4′
5′
6′
formula
procedure
u-PA
t-PA
381
75
56
46
41
113
63
14
10
3.4
69
69
1.0
48
100
25
31
2.4
0.28
factor Xa
plasmin
trypsin
6a
6b
6c
6d
6e
6f
6g
6h
6i
6j
6k
6l
6m
6n
6o
6p
6q
6r
H
H
H
H
H
H
H
H
NO2
F
Br
Me
OMe
Ph
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
C
C
14H12N4
14H12N4O
C14H11N5O3
14H11FN4O
14H11BrN4O
15H14N4O
15H14N4O2
C14H11N5O3
14H11FN4O
14H11BrN4O
15H14N4O
15H14N4O2
20H16N4O
15H14N4O
C18H21N5O
14H12N4O2
18H14N4O
A
A
A
D
D
A
A
D
C
B
D
A
D
D
A
D
A
A
A
55
5.5
3.3
2.8
3.7
7.5
3.6
0.90
0.55
0.25
8.0
3.6
0.40
3.6
13
1.9
3.9
0.28
0.50
125
14
7.8
6.8
5.7
16
8.9
1.3
0.89
0.30
10
340
55
34
33
37
65
41
1.1
7.0
2.3
350
46
41
21
29
60
25
15
5.5
1.7
60
19
5.0
35
80
7.5
33
3.2
4.1
55
23
30
10
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
H
H
H
H
H
H
H
H
H
H
H
H
H
H
OH
H
H
H
NO2
F
Br
Me
OMe
H
H
H
H
H
H
H
H
H
H
Me
Cl
C
C
C
C
14
9.0
6.5
8.5
3.5
1.4
15
30
6.5
18
16
6.0
18
3.3
3.3
C
C
C
C
C
C
48
16
80
15
38
90
17
80
2.1
9.4
25
5.0
9.9
0.57
0.94
Me
NEt2
H
H
H
C
C
C
C
-naphthyl-
Br
Ph
H
H
15H13BrN4O
20H15ClN4O
2.7
12
6s
A-D in Scheme 1), incorporating some minor modifica-
tions, were used to synthesize the benzimidazole ana-
logues 6a -s. General methods A, B, and C involve the
condensation of a substituted benzaldehyde with 3,4-
diaminobenzamidine dihydrochloride 5. In the presence
of 1,4-benzoquinone, the desired benzimidazole was
formed rapidly and in good yield, but removal of trace
amounts of 1,4-benzoquinone and related products was
problematic, often requiring laborious chromatography
(general procedure C). Sodium metabisulfite was found
to facilitate the oxidation step and was easier to remove
compared to 1,4-benzoquinone (general procedure B).
The absence of oxidizing agents in the reaction also gave
the desired product, presumably through slow atmo-
spheric oxidation (general procedure A). Most of the
analogues shown in Table 1 were generated by proce-
dure A, since it eliminated the need to remove an
oxidizing agent in workup. In the examples where it was
more convenient to use a commercially available benzoic
acid, the procedure of Fairley et al. was used.17 Thus,
3,4-diaminobenzamidine was condensed with the cor-
responding benzoic acid in polyphosphoric acid (PPA)
at high temperature (general method D). The 3,4-
diaminobenzamidine reagent 5 which was used in
general methods A-D was synthesized using literature
procedures.17 The products were purified by reverse
phase HPLC, and the yields were typically in the 10-
50% range. Most of the aryl aldehydes and acids were
commercially available except for those obtained ac-
cording to Scheme 2. 2-Bromophenol 7 was treated with
paraformaldehyde in the presence of base and MgCl2
to form benzaldehyde 8.20 The same procedure was also
used to synthesize aldehyde 10 from biphenyl-2-ol 9.
Chlorination of 10 using the reagent N-chlorosuccin-
imide (NCS) gave benzaldehyde 11. Compound 13 was
prepared from commercially available benzaldehyde 12
using bromine in acetic acid.
The synthesis of compounds containing a 5-amidino-
indole group instead of a 5-amidinobenzimidazole was
also based on existing literature procedures (Schemes
3 and 4).16,21 Scheme 3 depicts the sequence of steps
used to make 5-amidinoindole 19, a compound previ-
ously reported by Iwanowicz and co-workers.21 Nitrile
16 was prepared21 and then converted to the 5-ami-
dinoindole product 19, through formation of the hy-
droxyamidine 17 and acetoxyamidine 18. A shorter but

2756 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17
Verner et al.
Sch em e 2a
Sch em e 5a
a
Reagents and conditions: (a) MeOH, NBS; (b) C6H5B(OH)2,
Pd(PPh3)4, base, H2O, 75 °C.
Sch em e 6a
a
Reagents and conditions: (a) MgCl2, Et3N, (CH2O)n, reflux;
(b) NCS, AcOH, reflux; (c) bromine, AcOH.
Sch em e 3a
a
Reagents and conditions: (a) fuming HNO3, AcOH, 0 °C; (b)
NBS, AcOH; (c) C6H5B(OH)2, Pd(PPh3)4, 10% Na2CO3, THF,
80 °C.
a
Reagents and conditions: (a) Iwanowicz;21 (b) CuCN, 220 °C,
quinoline; (c) 50% NH2OH (aq), EtOH, reflux; (d) Ac2O, AcOH,
THF; (e) H2, Pd/C 10%.
of 4-hydrazinobenzonitrile.23 Hydrazine 21 was then
condensed with the desired acetophenones to give
hydrazones. High temperature treatment of the hydra-
zones in PPA then resulted in cyclization to the indoles.
Purification of the products from the PPA reagent
required reverse phase preparative HPLC chromato-
graphy. A drawback of this method was the need to
remove phosphorylated side products, leading to lower
than expected yields, typically in the 10-30% range.
The acetophenones required for analogues 22a -k were
either commercially available or conveniently synthe-
sized using the methods shown in Schemes 5-8. Aryl
bromide 24 was prepared in 85% yield by treatment of
23 with N-bromosuccinimide (NBS) (Scheme 5). The two
aryl bromides 24 and commercially available 25 were
then converted in high yield to biaryls 26 and 27,
respectively, using Suzuki conditions.24 Nitration of 28
with fuming nitric acid yielded 1-(2-hydroxy-5-nitro-
phenyl)ethanone 30 (Scheme 6). Subsequent reaction of
both 28 and 30 with NBS resulted in the formation of
aryl bromides 29 and 31, respectively. The para isomer
of 29 was the main product in the reaction of 28 with
NBS, but sufficient amounts of the desired ortho sub-
Sch em e 4a
a
Reagents and conditions: (a) NaNO2, 6 N HCl, 0 °C then
SnCl2‚2H2O; (b) ArCOCH3, EtOH, Et3N, reflux; (c) PPA, 140-160
°C, 5-71% yield.
lower yielding approach based on the Fischer Indoliza-
tion was used for all the other examples 22a -k in Table
2 (Scheme 4).16,22 Hydrazine 21 was prepared from
aniline 20 according to the hydrazine preparation
conditions described by Castro et al. for the synthesis

Serine Protease Inhibitors with a Unique Binding Mode
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17 2757
Sch em e 7a
Weinreb amide 34 was formed from 33 and was then
treated with methyllithium at 0 °C to generate aceto-
phenone 35 in 47% yield over the two steps. The
acetophenone 35 was then readily transformed to 36,
using NCS, and also to 37, using NBS. Acetophenone
39 was prepared according to Scheme 8. Benzaldehyde
10 was treated with methyllithium in the presence of
cesium carbonate to form 2-methoxybiphenyl-3-carb-
aldehyde 38 in 56% yield. The aldehyde 38 was then
converted to 39 in 70% yield by treatment with vana-
dium(III) chloride and MeMgBr followed by oxidation
of the intermediate alcohol using tetrapropylammonium
perruthenate (TPAP) and 4-methylmorpholine-N-oxide.
To prepare alternative S1 binding heterocycles such
as the methyl substituted benzimidazole 49, the syn-
thetic route in Scheme 9 was employed. Nitrile 43 was
prepared from acid 40 in 69% overall yield through the
formation of the acid chloride 41 and amide 42. Dis-
placement of the methoxy group by methylamine was
readily performed in DMSO at 75 °C to give arylamine
44. Reduction of 44 over Pd/C to diamine 45 was then
followed by condensation with aldehyde 10 according
to general scheme C (see above) to give the benzimid-
azole 46 (77% yield from 43). The final conversion of
benzimidazole 46 to 5-amidinobenzimidazole 49 pro-
ceeded in 33% yield, through the same sequence of steps
used for the synthesis of 19 from 16.
a
Reagents and conditions: (a) N,O-dimethylhydroxylamine
hydrochloride, PyBOP, DIEA, DMF, 0 °C; (b) MeLi, 0 °C, Et2O;
(c) NCS, DMF; (d) NBS, AcOH, 85 °C.
Sch em e 8a
The synthesis of the benzofuran analogue, 58, is
outlined in Scheme 10 and is based on a palladium-
mediated coupling of an alkyne with an aryl halide.
Treatment of 4-hydroxybenzonitrile 50 with N-iodo-
succinimide (NIS) gave aryl halide 51. The hydroxyl
group on benzaldehyde 10 was protected with 2-meth-
oxyethoxymethyl (MEM) chloride. It was important to
protect the hydroxyl group prior to formation of alkyne
53 in order to prevent intramolecular benzofuran for-
mation. The protected aldehyde 52 was converted to
a
Reagents and conditions: (a) CH3I, Cs2CO3, DMF; (b) CH3MgBr,
VCl3, CH2Cl2, -78 °C; (c) TPAP, 4-methylmorpholine-N-oxide,
CH2Cl2.
stituted isomer could be isolated. Similar conditions to
those used for the formation of biaryls 26 and 27 were
then used to generate biaryl 32. The sequence of steps
used to synthesize acetophenones 35-37 from 3-
phenylsalicylic acid 33 is shown in Scheme 7. The
Sch em e 9a
a
Reagents and conditions: (a) (COCl)2, DMF, THF, 0 °C; (b) NH3, THF, 0 °C; (c) TFA, Et3N, THF; (d) 40% methylamine (aq.), DMSO,
75 °C; (e) H2, Pd/C 10%, MeOH, EtOAc; (f) 10, general procedure C; (g) 50% NH2OH (aq), DMSO, 70 °C; (h) Ac2O, AcOH; (i) H2, Pd/C 10%,
AcOH.

2758 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17
Verner et al.
Sch em e 10a
(Table 1) as a low micromolar inhibitor toward a panel
of six human serine proteases. The scaffold has several
notable attributes, including a low molecular weight
(<300 amu), nonpeptidic character, and a convenient
synthesis from diaminobenzamidine and a salicylic acid
(or salicylaldehyde), making structure-activity rela-
tionships relatively straightforward to explore. The
affinity measured at pH ) 7.4 for the six enzymes, in
the absence of free zinc ions, ranged from 5.5 to 75 µM
(Table 1). The high-resolution X-ray structure of the
zinc-independent binding mode of 6b in trypsin and in
u-PA (Figure 3) shows that binding of the inhibitor is
mediated by a series of hydrogen bonds, three of which
are very short (<2.3 Å) (Table 3). This unique multi-
centered short hydrogen bond-mediated binding mode
is shown schematically in Figure 4. Briefly, the short
hydrogen bonds involve a triangle of oxygens comprising
the inhibitor oxygen (O2′), a water co-bound in the
oxyanion hole (H2Ooxy) and the catalytic serine (OγSer195).
The distances in the u-PA complex of 6b are O2′-
OγSer195 ) 2.14 Å, H2Ooxy-OγSer195 ) 2.30 Å, and
H2Ooxy-O2′ ) 2.29 Å, respectively, and those in trypsin
are O2′-OγSer195 ) 2.13 Å, H2Ooxy-OγSer195 ) 2.19 Å,
and H2Ooxy-O2′ ) 2.28 Å, respectively, all well below
that of ordinary hydrogen bonds. The crystallographic
refinements were performed with the energy terms
between the groups involved in the short hydrogen
bonds removed. Conventional crystallographic refine-
ment yielded hydrogen bond lengths of between 2.3 and
2.5 Å, but in this instance, the fit of the inhibitor with
the well-defined density was noticeably suboptimal in
several of the structures. Further characterization
regarding the accuracy and the nature of these unusu-
ally short hydrogen bonds from crystallographic and
mechanistic studies has recently been reported by Katz
et al.14 Upon the basis of these findings it has been
proposed that the binding molecule is actually the
phenolate ion of the inhibitor over a large range of pH.
In thrombin, additional short hydrogen bonds are made
among the inhibitor benzimidazole nitrogen (N1), the
water trapped in the oxyanion hole, and OγSer195 (Table
3). In all three enzymes, u-PA, thrombin, and trypsin,
there are also more typical length hydrogen bonds
involving a distally bound water molecule (H2Odistal),
Nꢀ2His57, and the backbone nitrogens of Gly193 and
Ser195, respectively (Figure 4). Therefore, a complex
multicentered network of hydrogen bonds is formed,
involving at least three unusually short hydrogen bonds.
The X-ray structure of 6b binding in the active site of
u-PA (Figure 3b, i) also reveals a profound effect on the
interactions at the bottom of S1 due to the deeper S1
pocket of u-PA compared to trypsin. The amidine group
of the inhibitor does not make direct hydrogen bonded
salt bridges to Asp189, but rather, a water molecule
mediates the interaction between these two groups. To
date, all the high-resolution X-ray structures of amidine-
or guanidine-bearing inhibitors complexed in the active
site of u-PA reveal direct interactions between the
inhibitor and Asp189.6,28 The short and normal range
hydrogen bond interactions of the inhibitor 6b around
the Ser195 residue in u-PA are presumably strong
enough to compete and perturb the interactions that
would normally occur between the amidine group and
Asp189. Supportive evidence is also provided by the
a
Reagents and conditions: (a) MEM chloride, DIEA, DMF; (b)
(1-diazo-2-oxopropyl)phosphoric acid dimethyl ester, K2CO3, MeOH;
(c) NIS, AcOH; (d) DMF, Pd(PPh3)2Cl2, Et3N, CuI; (e) 50% NH2OH
(aq), EtOH, reflux; (f) Ac2O, THF; (g) H2, Pd/C 10%, MeOH; (h) 6
N HCl, MeOH.
alkyne 53 using the (1-diazo-2-oxopropyl)phosphoric
acid dimethyl ester reagent.25 Coupling of alkyne 53
with aryl halide 51 under similar conditions to those
described by Candiani26 gave the benzofuran, 54, in 23%
yield. The benzofuran 54 was then converted to 57 using
the sequence of steps described for the preparation of
19 from 16. Final removal of the MEM protecting group
with hydrochloric acid gave the desired product, 58, in
43% yield from 54.
Biologica l Activity
All the compounds were tested against a panel of six
human trypsin-like serine proteases: u-PA, tissue-type
plasminogen activator (t-PA), factor Xa, thrombin, plas-
min, and trypsin. Ethylenediaminetetraacetic acid
(EDTA) was used to remove zinc ions thereby eliminat-
ing any zinc-mediated binding effects in the assay that
can occur for some of these scaffolds. The assay was
buffered at pH ) 7.4 in order to mimic physiological pH.
Apparent inhibition constants were converted to Ki
values as described in the Experimental Section.27
Resu lts a n d Discu ssion
The screening of libraries of compounds containing
amidine-bearing heterocycles yielded compound 6b

Serine Protease Inhibitors with a Unique Binding Mode
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17 2759
F igu r e 3. (a) Structure and corresponding (2|Fo| - |Fc|), Rc map contoured at 1.0, 2.4, and 3.8 σ for the active site and oxyanion
hole of the u-PA-6b complex, pH 6.5, 1.75 Å. Short hydrogen bonds are cyan, enzyme-inhibitor and enzyme-water-inhibitor
hydrogen bonds white, and intra-enzyme, intra-inhibitor, water-water, and water-enzyme hydrogen bonds yellow. Note that
the structure and density for the Cys42 and Cys58 side chains have been removed for clarity. (b) Comparison of density maps of
6b complexed in (i) u-PA and (ii) trypsin. The short hydrogen bonds are shown as cyan dashed lines, the intra-enzyme or intra-
inhibitor hydrogen bonds as yellow dashed lines, and enzyme-water-inhibitor hydrogen bonds as white dashed lines. In part
b(i), note the water molecule mediating interactions between the Asp189 residue and the amidine. Another water molecule is
also present in the vicinity of Ser190 in both structures and hydrogen bonds to the amidine group.
short hydrogen bonds. This molecule was found to
establish direct interactions with the Asp189 residue
(data not shown). Comparison of the assay data between
6a and 6b (Table 1) shows that, despite the unfavorable
interaction in S1, the introduction of the hydrogen
bonding network affords a 10-fold potency enhancement
toward u-PA. The structure of trypsin complexed with
6b (Figure 3b, ii) reveals more typical interactions at
the bottom of S1 since the size of the inhibitor is more
compatible with the shorter S1 pocket of trypsin com-
pared to u-PA. Surprisingly though, the Ki value of 6b
compared to 6a for trypsin indicates only a 2-fold
enhancement. These affinity gains for u-PA and espe-
cially trypsin are surprisingly small given that there
are three new direct hydrogen bonds formed between
F igu r e 4. Schematic of 6b, and analogues, binding through
short hydrogen bonds to the serine protease Ser195 and His57
residues in trypsin and u-PA. Short hydrogen bonds are shown
as red lines. Intra-enzyme and intra-inhibitor hydrogen bonds
are shown as blue lines. The hydrogens involved in short
hydrogen bonds are denoted as Hs. The inhibitor hydroxyl
group is deprotonated and the enzyme His57 protonated. The
hydrogen atoms in bold (H) are also involved in ordinary
hydrogen bonds (dashed lines).
the inhibitor and the protein and several via H2Ooxy. It
is therefore likely that the scaffold is suboptimal due
to factors such as, unfavorable van der Waals interac-
tions, desolvation effects, or conformational entropy
changes of the inhibitor upon binding.
Initially, substitutions were made around the phenol
ring of scaffold 6b in an attempt to optimize the potency.
Given the intricate network of hydrogen bonds between
the hydroxyl (phenolate) group of the inhibitor and the
enzyme, it was expected that these substitutions affect-
X-ray structure of 6a , a compound which lacks the
essential hydroxyl group for forming the network of

2760 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17
Verner et al.
ing the electronic characteristics of this key group might
change the inhibition profile for the enzymes. Sterically
small substitutions ortho to the hydroxyl group 6h -l
(Table 1) were indeed found to have profound effects
on potency. In general, the electron withdrawing sub-
stitutions, 6h -j relative to hydrogen, 6b, tended to give
optimal potencies over the electron donating methyl and
methoxy substitutions, which resulted in only minor
(<3-fold for any enzyme) changes. The bromine substi-
tution, 6j, caused the greatest increase in affinity toward
the entire panel of enzymes except thrombin. The
enhancements ranged from 16-fold for trypsin to 47-fold
for factor Xa. Affinity toward u-PA was increased 22-
fold by the bromine substitution. In the case of throm-
bin, the nitro group had a slightly greater effect than
bromine and gave an additional 2-fold enhancement.
Further exploration of larger groups ortho to the hy-
droxyl group resulted in 6m . Substantial potency en-
hancements compared to 6b were noted, but compared
to the bromine substitution 6j, there appears to be no
added benefit. It was noticeable that the effects of ortho
substitution did not correlate in a predictable manner
with the relative inductive potentials of the chemical
groups. One reason for this is that the group ortho to
the hydroxyl is directed toward the S1′ and therefore
introduces additional protein interactions. For example,
the aryl ring in 6m makes contact with the disulfide
bridge between Cys42 and Cys58 at the bottom of S1′,
in addition to interactions with residues His57 and
Val41. There is also the likelihood that the close
proximity of the ortho substituent to the hydrogen bond
network also introduces other effects on potency in
addition to direct inductive effects.
Substitution at other sites on the aryl ring has also
been explored, but only small changes have been
observed. A few examples, analogues 6n -q, are pre-
sented in Table 1. Meta substitution or the presence of
hydroxyl groups at both C-2′ and C-6′, compounds 6n -
p , showed no effects on potency. Larger aromatic rings
such as naphthyl, 6q, tended to reduce affinity. p-
Methyl substitution in addition to o-bromo substitution,
compound 6r , gave essentially the same results as
compound 6j with the o-bromo substitution alone.
Similarly, p-chloro substitution in addition to o-phenyl
substitution, compound 6s, was very similar to com-
pound 6m with only o-phenyl substitution. These latter
results further confirm the lack of effects on affinity at
pH ) 7.4 by para substitutions.
In summary, the SAR performed on the phenolic ring
of the molecule yielded compounds 6j and 6m as two
analogues that display enhanced potency toward the
panel of serine proteases. It was apparent that substitu-
tion ortho to the hydroxyl group had the more dramatic
impact on potency whereas substitution around the
other sites had only marginal effects on affinity. Atten-
tion was now turned to the S1 binding heterocycle part
of the molecule.
The 5-amidinobenzimidazole series of analogues 6a -s
occupy the S1 site of the protease as confirmed by X-ray
analysis (Figure 3). Using a similar panel of trypsin-
like serine protease enzymes, it was reported by Geratz
and co-workers that the intrinsic potency of 5-amidino-
indole compared to 5-amidinobenzimidazole was be-
tween 2- and 350-fold greater.18 In the case of scaffold
6b, changing the benzimidazole heterocycle to an indole
(compound 19), thereby fixing the tautomeric nature of
the nitrogen compared to the benzimidazole analogues,
resulted in at most an 8-fold potency enhancement
(Table 2). The lowest enhancement was for u-PA, where
activity was only 2-fold greater. However, the same
benzimidazole to indole change on the optimized benz-
imidazole scaffolds 6j and 6m , analogues 22b and 22f,
respectively, resulted in larger potency enhancements.
In the case of 22b, the enhancements were modest and
ranged from 5- (trypsin) to 19-fold (thrombin) ultimately
resulting in Ki values below 380 nM toward the whole
panel of enzymes. The enhancements increased to 28-
50-fold in the case of analogue 22f. Analogue 22f has a
Ki toward the whole serine panel of below 320 nM and
is the most potent analogue toward u-PA and plasmin.
For u-PA, the Ki ) 8 nM is a 50-fold increase over the
corresponding benzimidazole analogue, 6m , and a 700-
fold increase over the original lead 6b. Therefore, the
indole ring coupled with the presence of an ortho
substituted phenol resulted in substantial affinity en-
hancements across the panel of enzymes. It is unlikely
that the larger affinity enhancements in this series of
indoles are the result of the proposed greater intrinsic
affinity of the 5-amidinoindole compared to 5-amidino-
benzimidazole since the comparison of 19 and 6b reveals
only small differences in affinity. Furthermore, the
earlier studies18 reported that there were no intrinsic
differences in affinity toward plasmin and trypsin of
these two heterocycles which is not consistent with the
large enhancements observed for indole 22f compared
to benzimidazole 6m in this study. An important factor
to consider in light of these results is the interaction of
To separate these potential effects from the inductive
effects, the same group of substituents, except phenyl,
was introduced onto 6b para to the hydroxyl group,
examples 6c-g. This position on the scaffold is directed
away from the protein and the hydrogen bond network
and toward solvent accessible space. The para substitu-
tions 6c-g had remarkably little effect on affinity
toward all the enzymes. The largest change was for the
p-methoxy analogue, 6g, which was only marginally
(3.5-fold) more potent toward trypsin than lead 6b. The
pKa of the phenol group was measured in several of
these molecules and found to be 4.25 (nitro, 6c), 7.88
(fluoro, 6d ), 8.23 (hydrogen, 6b), and 8.25 (methoxy, 6g).
The lack of a clear correlation between affinity and the
relative inductive-electronic effects of the substituents
is harder to rationalize in this series of analogues.
Clearly other factors in addition to inductive effects
must contribute an important role in binding. Recent
work has shown that the pH of the assay media can
significantly affect the solubility of these analogues in
the same media.14 The benzimidazole group is pro-
tonated at low pH and the phenol is deprotonated at
high pH, leading to greater solubilities at low and high
pH, respectively. Compared to 6b, the p-nitro substi-
tuted analogue, 6c, has a much lower solubility at pH
8.6 (1770-fold lower), whereas the p-methoxy analogue,
6g, is very similar to 6b. These solubility differences
lead to varying free energies of solvation, and conse-
quently varied affects on binding affinity depending on
the inhibitor.14

Serine Protease Inhibitors with a Unique Binding Mode
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17 2761
Ta ble 2. Effect of Modifications on the S1 Binding Element on Enzyme Inhibition
R
Ki (µM)
no.
X
Y
2′
3′
4′
5′
formula
u-PA
t-PA
factor Xa
14
2.7
0.052
0.043
0.089
0.078
0.21
0.078
0.21
0.34
0.014
43
thrombin
plasmin
trypsin
6b
19
N
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
N
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
C
C
C
C
C
14H12N4O
15H13N3O
16H14BrN3O
15H12BrN3O
15H11Br2N3O
C15H11BrN4O3
22H19N3O
21H17N3O
21H16BrN3O
21H16ClN3O
C21H16N4O3
21H17N3
22H19N3O
21H18N4O
21H16N2O2
5.5
2.4
0.055
0.035
0.10
75
9.4
55
13
46
12
23
8.0
0.32
0.29
1.0
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
NMe
CH
22a
22b
22c
22d
22e
22f
22g
22h
22i
22j
22k
49
Br
Br
Br
Br
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Me
H
Br
NO2
Me
H
Br
Cl
NO2
H
0.46
0.38
0.69
1.3
0.11
0.035
0.094
0.16
0.019
138
19
150
2.8
0.18
0.12
0.32
0.24
1.1
0.32
1.4
0.29
0.11
32
0.60
0.25
0.90
2.1
0.25
0.10
0.35
0.29
0.19
26
0.28
4.3
C
C
C
C
0.038
0.008
0.043
0.055
0.025
13
3.2
11
0.55
0.13
0.13
0.70
0.13
0.075
17
1.2
6.5
1.6
C
C
C
C
OMe
OH
OH
H
H
H
19
30
6.8
22
70
7.0
26
12
5.5
58
O
Ta ble 3. Enzyme-Inhibitor and Enzyme-H2Ooxy-Inhibitor Hydrogen Bond Lengths at the Active Sites of Trypsin, Thrombin, and
u-PA Complexes of 6b and Analoguesa
complex
pH
O2′-OγSer195
H2Ooxy-OγSer195
H2Ooxy-O2′
u-PA
H2Ooxy-N1
OγSer195-N1
Nꢀ2His57-O2′
6bb
6.50
6.50
6.50
2.14 (14)
2.33
2.00
2.30 (12)
2.22
2.39
2.29 (08)
2.33
2.21
2.77(08)
2.95
2.87
2.63 (04)
2.72
2.51
2.68 (04)
2.70
2.69
22fc
22a d
Trypsin
6b
22f
22a
3.50-9.45
7.07
2.13 (07)
2.01
2.29
2.19 (11)
2.27
2.08
2.28 (06)
2.31
2.39
2.57 (11)
2.85
2.83
2.58 (07)
2.63
3.02
2.67 (05)
2.75
2.71
8.00
Thrombin
6b
22a
5.34-8.28
2.26 (19)
2.08
2.26 (19)
2.27
2.35 (15)
2.17
2.37 (13)
2.39
2.46 (05)
2.42
2.70 (05)
2.89
7.80
a
O2′ is the inhibitor hydroxyl oxygen. N1 is the nitrogen of the benzimidazole proximal to the Ser195 residue. H2Ooxy is the water in
the oxyanion hole. Nꢀ2His57 is the proximal nitrogen of the catalytic His57 residue. OγSer195 is the oxygen of the Ser195 hydroxyl group.
b
For further direction, see Figure 3. Standard deviations are in parentheses. Resolution ) 1.75 Å; R ) 20.5%. c Resolution ) 1.73 Å; R
d
) 20.2%. Resolution ) 1.65 Å; R ) 24.4%.
the S1 binding heterocycle with the intricate network
of hydrogen bonds. On the basis of the binding mode
determined for 6b, the nitrogen atom N1 donates a
hydrogen bond to OγSer195 and the water molecule in the
oxyanion hole thereby affording some stabilization. The
indole nitrogen has fixed hydrogen bond donating
character, but in contrast, the benzimidazole has two
possible tautomers, only one of which is the preferred
binding conformer (N1 protonated). Gas phase energy
calculations on 6b have indicated that this molecule is
more stable in the unfavorable tautomer conformation
relative to the mode of binding.14 It seems reasonable
to postulate that this tautomeric ambiguity leads to the
weaker affinities of the benzimidazoles compared to
indoles in this series of analogues, e.g., 6m compared
to 22f, and 6r compared to 22a . Consequently, on the
basis of this theory it would be expected that N3
substituted benzimidazoles, e.g., 49, that have a fixed,
unfavorable tautomeric state relative to the preferred
binding mode, would be poor inhibitors. To this end,
compound 49 (Table 2) can be directly compared to the
unsubstituted benzimidazole 6m and indole 22f. The
potency of 49 is consistently lower than the benzimid-
azole analogue 6m across the whole panel of enzymes
except trypsin, and 6m is itself consistently less active
than indole 22f for all enzymes. Assuming that the
additional methyl group was not significantly changing
other binding factors, this result supports the proposed
theory. Furthermore, the benzofuran analogue of 22f,
analogue 58, which cannot donate a hydrogen bond, in
general, displayed a similar affinity profile to that of
6m . Although these results need further investigation,
they strongly suggest that a hydrogen bond donor is
necessary at this position of the scaffold for optimal
affinity.
The scaffolds 22b (o-bromo) and 22f (o-phenyl) were
substituted at the para position to create two series of
analogues, 22a -d and 22e-i, respectively (Table 2).
Comparing and contrasting the effects on affinity of
para substitution on the indole scaffolds to those
observed on the benzimidazole scaffolds reveals a subtle
difference. In the benzimidazole series, 6b-g, para
substitution resulted in potency fluctuations of no more
than 3.5-fold between the most and least potent ana-
logues. However, para substitutions on the o-phenyl
scaffold, analogues 22e-i, resulted in more significant
potency fluctuations. For example, comparing the factor
Xa affinity for 22h (p-chloro) with 22i (p-nitro) indicates
a 24-fold change in potency (Table 2). In the o-bromo
indole analogues 22a -d , the fluctuations were greatest

2762 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17
Verner et al.
toward trypsin where the potency changed by 15-fold
between 22b and 22d . Despite this difference in sensi-
tivity toward para substitution, there is still no clear
correlation with the relative inductive effects of the
substitutions at the para position in either the o-bromo
or o-phenyl series at pH ) 7.4, presumably for reasons
similar to those mentioned earlier. A striking example
of this is the fact that p-nitro analogue in the o-phenyl
series, compound 22i, has optimal potency toward t-PA
and trypsin within that series, but the same p-nitro
substitution on the o-bromo series, 22d , leads to the
lowest comparative potency toward t-PA, trypsin, plas-
min, and u-PA. The analogue 22i is the most potent
inhibitor shown for t-PA, factor Xa, thrombin, and
trypsin displaying a Ki below190 nM across the whole
panel of enzymes.
are all shorter than ordinary hydrogen bond lengths and
therefore similar to that of the u-PA-6b complex (Table
3). Also clearly visible in the density map are two water
molecules in the S1 subsite of u-PA, one mediating
binding from the amidine to Asp189, the other partici-
pating in hydrogen bonding with the amidine group.
Both of these water molecules can be found in the
corresponding complex of 6b in u-PA. The aryl group,
as noted above, is directed toward S1′ and makes van
der Waals contact with the disulfide bridge between
Cys42 and Cys58, His57 and Val41. Figure 5b shows
the overlaid structures of 22f and 6b in u-PA at pH )
6.5, and Figure 5c the overlay of the two inhibitors in
trypsin at pH ) 7.1 and 6.4, respectively. Essentially
the molecules bind in identical modes in both enzymes,
and the key short hydrogen bonding distances compare
very nicely (Table 3). The most significant difference to
note is the small shift of residue His57 in the complexes
of 22f. The histidine is most likely protonated at the
pH of crystallization (pH ) 6.5) and presumably moves
to interact favorably with the aryl ring of 22f. The
affinity of 22f was measured at pH ) 6.5 under con-
ditions identical to those of the crystallization process,
and only minor (<2-fold) differences were observed
compared to the affinities measured at pH 7.4 (data not
shown). Thus, the dramatic potency enhancements of
22f in u-PA (700-fold) and trypsin (177-fold) compared
to 6b are not the result of gross changes in the binding
mode.
To establish the necessity of the hydroxyl group
toward binding in some of the potent inhibitors such as
22f, compound 22j (the analogue of 22f without the
hydroxyl group) was prepared. Analogue 22j was sig-
nificantly less potent than 22f toward all the enzymes
and was also less potent than 19, the basic indole
scaffold with no ortho substitution. The masking of the
hydroxyl group as a methyl ether, compound 22k , also
afforded a reduction in potency across the whole panel.
Removal of the basic amidine group also leads to
significantly poorer inhibition across all the enzymes
in the panel (data not shown). These results, taken
together, indicate that the interactions between the
basic group of the inhibitor and Asp189 in the S1
subsite, and the interactions of the hydroxyl group with
the catalytic residues, are both essential components of
binding for these inhibitors.
A similar structural analysis was performed for 22a ,
whose affinity enhancements over 6b ranged from 72-
fold (trypsin) to 269-fold (factor Xa). The density map
of 22a complexed in u-PA (Figure 6a) is very comparable
to that of 6b complexed in u-PA (Figure 3b, i) and
therefore, by default, 22f. The complex displays the
expected pattern of short hydrogen bonds with the
inhibitor being well ordered and co-binding with the two
water molecules in S1. The distance between O2′ and
OγSer195 is very short, only 2.00 Å, well below a typical
hydrogen bond distance of 2.6 Å. Figure 6b shows the
overlay of 6b and 22a in thrombin and, similar to 22f,
reveals an identical binding mode for the two inhibitors.
The indole nitrogen engages in short hydrogen bonds
with OγSer195 and the H2Ooxy as observed for the inhibi-
tor 6b in thrombin. Table 3 reflects these observations
but also indicates that the largest differences occur
between the complexes of 22a and 6b in trypsin,
whereby the OγSer195-N1 (benzimidazole nitrogen) dis-
tance is 0.4 Å longer. An overlay of these inhibitors in
trypsin is shown in Figure 6c. The inhibitor 22a has
moved away from the catalytic residues slightly and
sinks deeper into the S1 pocket, thus lengthening the
hydrogen bond from the nitrogen of the inhibitor to the
Ser195 residue. The hydrogen bonds involving the
triangle of oxygens remain short, however, since the
water molecule in the oxyanion hole also moves slightly.
Despite these changes, the overall structures are still
very similar. In addition to the structures presented
here, many other high-resolution structures of inhibitors
from this class of scaffold, complexed with different
serine proteases, have displayed similar modes of bind-
ing involving short hydrogen bonds.14 On the basis of
this evidence, the tighter binding observed in the
optimized inhibitors such as 22f and 22a cannot be
The optimized indole scaffolds 22a and 22f do not
show affinity enhancements toward any of the enzymes
in the presence of free zinc ions (data not shown). This
result was expected, since the indole scaffold is a weaker
zinc chelator compared to the benzimidazole scaffold.
Zinc-mediated binding modes are possible for the benz-
imidazoles 6c-s as discovered earlier for 6b.14 However,
at pH ) 7.4, the optimal benzimidazole based inhibitors,
6r and 6m , displayed only small (10-20-fold) affinity
shifts in the presence of free zinc ions, much lower than
the 5 orders of magnitude found for the BABIM-type1,12
compounds (see Supporting Information). Therefore it
is unlikely that the zinc-mediated binding modes of
these inhibitors will contribute toward inhibitor potency
under physiological conditions.
Given the successful optimization of the lead 6b into
several broadly potent zinc-independent inhibitors (22a-
b, 22f, 22h -i), it was necessary to confirm that the
mode of binding was similar to that of 6b and that the
affinity enhancements were not the result of an entirely
different binding mode. From the o-bromo series, 22a
was chosen, and from the o-phenyl series, 22f, since both
had good generic potency profiles but were also highly
potent for two targets of particular interest to this
group, u-PA and factor Xa respectively. Figure 5a shows
a stereoview density map of inhibitor 22f bound in the
active site of u-PA at 1.7 Å resolution. The inhibitor is
well-ordered, and a water molecule, trapped in the
oxyanion hole, sits in close proximity to the inhibitor
hydroxyl oxygen and the Ser195 hydroxyl oxygen. The
hydrogen-bonding distances between the oxygen atoms

Serine Protease Inhibitors with a Unique Binding Mode
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17 2763
F igu r e 5. (a) Stereoview of 22f complexed in the active site of u-PA. The color coding of the hydrogen bonds (dashed lines) near
the catalytic residues is the same as for Figure 3 except the enzyme-water-inhibitor hydrogen bonds are in magenta. Hydrogen
bonds involving H2O2S1 are shown in cyan and those involving H2O1S1, Ser190, and the inhibitor are in green and yellow. (b)
Overlay of the u-PA structures of 6b and 22f at pH ) 6.5. (c) Overlay of the trypsin structures of 6b and 22f at pH ) 6.4 and pH
) 7.1, respectively. Both overlays show the very close (almost identical) binding modes of the two inhibitors in each enzyme. In
part b, a water molecule mediates the interaction between Asp189 and the amidine in both structures and is absent in part c.
The aryl group of 22f is directed toward S1′.
rationalized through gross changes in the binding mode.
These inhibitors therefore represent optimized inhibi-
tors for this binding mode involving short and normal
range hydrogen bonds.
contributor to the large potency enhancements in these
optimized inhibitors. Rather, it is more likely that the
affinity enhancements of inhibitors 22f and 22a , for
example, are due to the combination of both the short
and normal range hydrogen bonds and other factors,
such as the aforementioned inductive effects, free ener-
gies of solvation, and van der Waals interactions.
A key question regarding the potential of these
generic inhibitors for drug development that has not
been addressed relates to the possibility of engineering
selectivity onto the scaffolds. Due to the two key binding
elements, the basic group and the multicentered net-
work of hydrogen bonds interacting with Ser195, it is
unlikely that enzymes other than trypsin-like serine
proteases would be targeted by such molecules. The
critical enzymes that then remain as antitargets would
only be those members of the trypsin-like serine pro-
teases that are closely related to the target of interest,
for instance, t-PA if one were targeting u-PA. In the data
shown for the most potent analogues, there are some
hints of selectivity for one enzyme over another even
though the inhibitors are largely unelaborated. It is
worth noting that these inhibitors are relatively rigid
and possess very few rotatable bonds. The combination
of these properties with the strong 2-fold ‘anchorage’ of
the inhibitors results in a strong dependence of affinity
on substituents placed around the scaffold. Appending
The cumulative SAR data and structural information
presented suggests that four factors contribute to the
overall potency of these new inhibitors. First, the
phenolic group that is involved in an elaborate network
of short and normal range hydrogen bonds, second a
basic group to interact either directly or indirectly with
the Asp189 residue, third ortho substitution to the
phenol, and finally an S1 heterocycle bearing the
appropriate NH functionality. The first three of these
four key functional groups are all essential for optimal
binding, and removal of any one will cause a dramatic
reduction in potency. For this reason, it is extremely
difficult to separate the individual effects on binding
affinity of these four factors from the overall binding
affinity in order to estimate the contribution from the
hydrogen bond network alone. However, by comparing
two inhibitors that differ only in the length of the
hydrogen bonds formed at the active site in thrombin,
the contribution of the short hydrogen bonds over the
normal range hydrogen bonds was estimated to be in
the range of up to 1.7 kcal/mol.14 This relatively small
increase means that the formation of the unusually
short hydrogen bonds is unlikely to be the major

2764 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17
Verner et al.
molecules of the serine proteases identified subtle
differences within the S1 pocket such as the identity of
residue 190 that can be exploited for the generation of
selectivity.28a The enzymes, despite being highly ho-
mologous in this region, do tend to have their own
unique characteristics or ‘fingerprint’. Efforts to apply
this knowledge onto 22f have led to the development of
highly selective inhibitors of u-PA through subtle modi-
fications of the S1 binding group (manuscript in prepa-
ration). Furthermore, substitution at the 3-position of
the indole ring was found to enhance affinity toward
factor Xa but reduce affinity toward u-PA, t-PA, plas-
min, and trypsin thereby generating selective factor Xa
inhibitors.15
Con clu sion
The screening lead 6b has the unique property of
being able to bind to serine proteases in two distinct
and unique binding modes, one zinc-dependent the other
zinc-independent. The zinc-independent binding mode
comprises a multicentered network of short and normal
range hydrogen bonds. Potency optimization through
subtle chemical modifications has yielded several in-
hibitors (22a , 22b, 22f, 22i) that display Ki values below
600 nM toward a panel of six human serine proteases
including important targets such as u-PA and factor Xa.
The binding mode for the optimized inhibitors has also
been demonstrated to involve an almost identical net-
work of short and normal range hydrogen bonds involv-
ing the hydroxyl group of the inhibitor, the catalytic
serine, and a water molecule trapped in the oxyanion
hole. These interactions form the basis of a fully
reversible and unique ‘anchor’ point to the protein. The
basic amidine group interacting with Asp189 at the
bottom of the S1 pocket, either directly or indirectly
through a water molecule, provides a second ‘anchor’
point. These key ‘anchor’ points between the inhibitor
and protein coupled with the commonality of the key
protein residues that are involved affords inhibitors or
scaffolds that are inherently potent toward a whole
family of enzymes. Removal of groups on the inhibitor
that contribute to these noncovalent interactions around
the catalytic residues and van der Waals contacts in S1′
results in sharply decreasing potency. The two inhibi-
tors 22a and 22f, both developed through relatively
minor modifications from the original lead 6b, have
displayed enhancements in potency of up to 2000-fold
over 6b. These novel inhibitors are now serving as leads
in drug discovery efforts toward the development of
selective and orally available u-PA, factor Xa, and factor
VIIa inhibitors. The availability of high resolution
crystal structures of these inhibitors complexed in u-PA
and other targets has provided unique opportunities for
engineering selectivity into the molecules. This, together
with the fully reversible binding mode, the relative
synthetic accessibility, and nonpeptidic nature of the
molecules provides a nice platform for rapidly develop-
ing therapeutically useful molecules for a variety of
protease targets.
F igu r e 6. (a) 22a complexed in the active site of u-PA at pH
) 6.5. The color coding of the hydrogen bonds (dashed lines)
near the catalytic residues are the same as for Figure 3. (b)
Overlay of the 22a and 6b complexes in thrombin at pH )
7.8. The complexes are almost identical with the bromine of
22a directed toward the S1′ pocket and the para methyl group
toward solvent. (c) Overlay of the 22a and 6b complexes in
trypsin at pH ) 8.0 and pH ) 6.4, respectively. The complexes
in this case show the most dramatic differences. The inhibitor
22a relative to 6b sits deeper into the S1 pocket extending
the length of the N1-OγSer195 and N1-H2Ooxy hydrogen bonds
and is tilted away from the catalytic residues slightly. The
short hydrogen bonds are maintained between the O2′-
OγSer195, O2′-H2Ooxy, and OγSer195-H2Ooxy
.
Exp er im en ta l Section
groups onto the S1 binding heterocycle or para to the
phenol group has been found to decrease affinity toward
some enzymes while enhancing affinity toward others.
For example, earlier work using small S1 binding
Reactions were carried out under an inert atmosphere at
room temperature unless indicated otherwise. Solvents and
chemicals were reagent grade or better and were obtained from
commercial sources. All aldehyde, acetophenone, and benzoic

Serine Protease Inhibitors with a Unique Binding Mode
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17 2765
acid reagents were obtained from commercial sources unless
otherwise indicated. Silica gel chromatography was performed
using Merck Kieselgel 60 (230-400 Mesh). Preparative HPLC
purification was carried out on a Gilson HPLC fitted with a
VYDAC C18 (5 × 25 cm) column eluting at 50 mL/min with a
gradient of 2-90% solvent B in solvent A over 60 min (solvent
A was 20 mmol aqueous HCl and solvent B was CH3CN, or
solvent A was 0.1% TFA in water and solvent B 0.1% TFA in
CH3CN). Analytical HPLC analysis was performed using a
HP1100 eluting with a gradient of 2-90% solvent B in solvent
A (solvent A was 0.05% TFA in water and solvent B was
CH3CN). NMR spectroscopy was determined on a GE 300 MHz
NMR fitted with a 300/44 5 mm 31P-15N/1H probe or a J EOL
ECLIPSE 270 MHz NMR. Chemical shifts are reported in
units of δ (ppm) and peak multiplicities are expressed as
follows: s, singlet; d, doublet; t, triplet; br, broad peak; q,
quartet; m, multiplet; dd, doublet of doublet; ddd, doublet of
doublet of doublets. Mass spectra were determined on a
Perkin-Elmer Sciex API-150EX mass spectrometer equipped
with a turbo ion spray source (MS-Sciex) or a Finnigan
TSQ7000 equipped with ElectroSpray ionization source (MS-
ESI). The analytical microanalysis were carried out by Rob-
ertson Microlit Laboratories Inc., Madison, NJ , and were
within 0.4% units of the theoretical values unless otherwise
indicated.
1H), 7.83 (dd, J ) 9.6, 1.9 Hz, 1H), 7.45 (m, 1H), 7.04 (m, 1H).
13C NMR (DMSO-d6) δ 165.9, 152.5, 151.5 (d, J ) 267 Hz),
146.2 (d, J ) 16.6 Hz), 137.8, 134.9, 124.0, 123.8, 123.4, 119.4-
119.7 (2 × C, m), 119.5 (d, J ) 5.2 Hz) 115.9, 114.7, 113.5 (d,
J ) 4.0 Hz). MS-Sciex m/z 271.2 [MH+]. Anal. (C14H11FN4O‚
2.0HCl) C, H, N.
Gen er a l P r oced u r e D: Exa m p le, 2-(2-Hyd r oxy-3-m eth -
ylp h en yl)-1H-ben zoim id a zole-5-ca r boxa m id in e (6k ). 3,4-
Diaminobenzamidine dihydrochloride 5 (0.35 g, 1.57 mmol)
was added to polyphosphoric acid (PPA) (8 mL) followed by
3-methylsalicylic acid (0.24 g, 1.57 mmol). The mixture was
stirred and heated to 180 °C. Analytical HPLC was used to
determine completion of the reaction which was typically
between 2 and 3 h. The mixture was treated with ice water
and neutralized to pH ) 7.0 with NH4OH solution. The
precipitate that formed was collected by filtration and sus-
pended in water. The insoluble material was collected by
filtration and dried. The crude solid was subjected to prepara-
tive HPLC to give 6k (90 mg, 19%). Yields were typically in
the 10-30% range. 1H NMR (DMSO-d6) δ 9.43 (br s, 2H), 9.18
(br s, 2H), 8.20 (s, 1H), 8.07 (d, J ) 7.7 Hz, 1H), 7.83 (d, J )
8.4 Hz, 1H), 7.74 (d, J ) 8.8 Hz, 1H), 7.31 (d, J ) 7.3 Hz, 1H),
6.93 (t, J ) 7.3 Hz, 1H), 2.24 (s, 3H). 13C NMR (DMSO-d6) δ
165.9, 156.5, 154.5, 133.6, 125.9, 124.8, 122.9, 122.2, 118.9,
115.7, 114.3, 111.1, 15.8. MS-Sciex m/z 266.8 [MH+]. Anal.
(C15H14N4O‚1.5HCl‚0.25H2O) C, H, N.
2-P h en yl-1H-ben zoim idazole-5-car boxam idin e (6a). Syn-
thesized using general procedure A. 1H NMR (DMSO-d6) δ 9.59
(br s, 2H), 9.30 (br s, 2H), 8.47 (d, J ) 6.3 Hz, 1H), 8.45 (d, J
) 8.5 Hz, 1H), 8.28 (d, J ) 1.1 Hz, 1H), 7.95 (d, J ) 9.4 Hz,
1H), 7.86 (dd, J ) 9.6, 1.7 Hz, 1H), 7.70-7.66 (m, 3H). 13C
NMR (DMSO-d6) δ 165.7, 152.5, 138.2, 135.0, 132.6, 129.4,
128.0, 125.6, 124.1, 123.6, 115.5, 114.6. MS-Sciex m/z 237.0
[MH+]. Anal. (C14H12N4‚2.0HCl) C, H, N.
2-(2-H yd r oxyp h e n yl)-1H -b e n zoim id a zole -5-ca r b ox-
a m id in e (6b). Synthesized using general procedure A. 1H
NMR (DMSO-d6) δ 9.49 (br s, 2H), 9.21 (br s, 2H), 8.29 (d, J
) 7.3 Hz, 1H), 8.23 (s, 1H), 7.93 (d, J ) 8.4 Hz, 1H), 7.79 (d,
J ) 8.8 Hz, 1H), 7.49 (t, J ) 7.7 Hz, 1H), 7.21 (d, J ) 8.1 Hz,
1H), 7.06 (t, J ) 7.4 Hz, 1H). 13C NMR (DMSO-d6) δ 165.8,
157.9, 151.4, 136.9, 134.3, 133.8, 128.8, 124.1, 123.7, 119.7,
117.3, 115.3, 114.5, 110.2. MS-Sciex m/z 253.0 [MH+]. Anal.
(C14H12N4O‚2.0HCl) C, H, N.
2-(2-Hydr oxy-5-n itr oph en yl)-1H-ben zoim ida zole-5-ca r -
boxa m id in e (6c). Synthesized using general procedure A. 1H
NMR (DMSO-d6) δ 9.51 (br s, 2H), 9.25 (br s, 2H), 9.19 (d, J
) 2.6 Hz, 1H), 8.32-8.27 (m, 2H), 7.92 (d, J ) 8.4 Hz, 1H),
7.79 (d, J ) 8.4 Hz, 1H), 7.39 (d, J ) 9.2 Hz, 1H). 13C NMR
(DMSO-d6) δ 165.9, 163.2, 151.0, 139.8, 138.8, 135.9, 128.1,
124.7, 123.6, 123.1, 118.2, 116.2, 115.0, 112.3. MS-Sciex m/z
297.8 [MH+]. Anal. (C14H11N5O3‚2.0HCl) C, H, N.
3,4-Dia m in oben za m id in e Dih yd r och lor id e (5). This
compound was prepared by the procedure of Fairley et al.17
Compounds 6a -s were made using one of the general
procedures A-D shown below.
Gen er a l P r oced u r e A: Exa m p le, 2-(2-Hyd r oxy-5-m eth -
ylp h en yl)-1H-ben zoim id a zole-5-ca r boxa m id in e (6f). A
suspension of 3,4-diaminobenzamidine dihydrochloride 5 (150
mg, 0.67 mmol) in absolute EtOH (10 mL) at 90 °C was treated
over 2 min with a solution of 2-hydroxy-5-methylbenzaldehyde
12 (91 mg, 0.67 mmol) in absolute EtOH (2 mL). The mixture
was heated for 2-3 h, and the solution was then allowed to
cool. Et2O was added, and the precipitate that formed was
collected by filtration. The precipitate was subjected to reverse
phase HPLC to give 6f (yields typically ranged from 10 to 50%
1
for the analogues presented). H NMR (DMSO-d6) δ 9.56 (br
s, 2H), 9.29 (br s, 2H), 8.27 (s, 1H), 8.18 (s, 1H), 7.96 (d, J )
8.4 Hz, 1H), 7.85 (dd, J ) 8.6, 1.5 Hz, 1H), 7.34 (dd, J ) 8.4,
1.8 Hz, 1H), 7.16 (d, J ) 8.5 Hz, 1H), 2.32 (s, 3H). 13C NMR
(DMSO-d6) δ 165.8, 155.8, 151.6, 137.2, 137.1, 135.0, 134.0,
128.5, 124.0, 123.6, 117.2, 115.3, 114.6, 109.9, 20.0. MS-Sciex
m/z 266.8 [MH+]. Anal. (C15H14N4O‚1.5HCl‚0.75H2O) C, H, N.
Gen er a l P r oced u r e B: Exa m p le, 2-(2-Hyd r oxy-3-br o-
m op h en yl)-1H-ben zoim id a zole-5-ca r boxa m id in e (6j). 3,4-
Diaminobenzamidine dihydrochloride 5 (0.23 g, 1.24 mmol)
was added to N,N-dimethylacetamide (6 mL) followed by
Na2S2O5 (0.31 g, 1.62 mmol) and finally 3-bromo-2-hydroxy-
benzaldehyde 8 (0.25 g, 1.24 mmol). The resultant mixture was
heated at 100 °C for 3 h and then allowed to cool. Et2O (200
mL) was added, and the solution was stirred overnight. The
precipitate that formed was collected by filtration, washed with
cold Et2O, and dried to give a crude solid (0.62 g). Half of the
solid was subjected to preparative HPLC to give 6j (50 mg,
11%, 22% based on crude solid weight). 1H NMR (DMSO-d6) δ
9.44 (br s, 2H), 9.16 (br s, 2H), 8.31-8.25 (m, 2H), 7.88 (d, J
2-(2-H yd r oxy-5-flu or op h en yl)-1H -b en zoim id a zole-5-
ca r boxa m id in e (6d ). Synthesized using general procedure
D. 1H NMR (DMSO-d6) δ 9.56 (br s, 2H), 9.29 (br s, 2H), 8.29-
8.25 (m, 3H), 7.94 (dd, J ) 8.4, 3.3 Hz, 1H), 7.84-7.80 (m,
1H), 7.42-7.35 (m, 1H), 7.30-7.24 (m, 1H). 13C NMR (DMSO-
d6) δ 153.4, 142.7 (d, J ) 234 Hz), 142.0, 138.2, 124.8, 121.7,
111.8, 111.4, 108.7 (d, J ) 23 Hz), 106.4 (d, J ) 7.3 Hz), 103.2,
102.4, 101.9 (d, J ) 26 Hz), 98.5 (d, J ) 8.5 Hz). MS-Sciex m/z
271.0 [MH+]. Anal. (C14H11FN4O‚1.5HCl‚0.5H2O) C, H, N, F.
2-(2-H yd r oxy-5-b r om op h en yl)-1H -b en zoim id a zole-5-
ca r boxa m id in e (6e). Synthesized using general procedure
) 8.6 Hz, 1H), 7.79-7.74 (m, 2H), 7.02 (t, J ) 8.4 Hz, 1H). 13
C
NMR (DMSO-d6) δ 165.8, 154.9, 153.7, 135.6, 126.4, 123.0,
122.4, 120.4, 113.2, 110.7. MS-Sciex m/z 331.0 [79Br, MH+],
333.0 [81Br, MH+]. Anal. (C14H11BrN4O‚HCl‚0.5H2O) C, H, N.
1
D. H NMR (DMSO-d6) δ 9.54 (br s, 2H), 9.28 (br s, 2H), 8.54
(d, J ) 2.5 Hz, 1H), 8.28 (s, 1H), 7.95 (d, J ) 8.7 Hz, 1H), 7.83
(dd, J ) 8.7, 1.5 Hz, 1H), 7.64 (dd, J ) 8.9, 2.5 Hz, 1H), 7.22
(d, J ) 8.9 Hz, 1H). 13C NMR (DMSO-d6) δ 165.8, 157.1, 150.8,
137.8, 136.0, 134.7, 130.5, 123.9, 123.5, 119.6, 115.7, 114.8,
113.0, 110.5. MS-Sciex m/z 331.0 [79Br, MH+], 333.0 [81Br,
MH+]. Anal. (C14H11BrN4O‚2.0HCl) C, H, N, Br.
Gen er a l P r oced u r e C: Exa m p le, 2-(2-Hyd r oxy-3-flu o-
r op h en yl)-1H-ben zoim id a zole-5-ca r boxa m id in e (6i). 3,4-
Diaminobenzamidine dihydrochloride 5 (1.43 g, 7.7 mmol) was
added to absolute EtOH (50 mL) followed by 3-fluoro-2-
hydroxybenzaldehyde (1.07 g, 7.7 mmol) and 1,4-benzoquinone
(0.83 g, 7.7 mmol). The resultant mixture was heated at reflux
for 4 h and then allowed to cool. The solvent was removed
under reduced pressure, and the resultant crude solid was
subjected to preparative HPLC to give 6i (2.34 g, 89%). 1H
NMR (DMSO-d6) δ 9.60 (br s, 2H), 9.34 (br s, 2H), 8.28 (d, J
) 1.7 Hz, 1H), 8.16 (d, J ) 8.8 Hz, 1H), 7.92 (d, J ) 9.6 Hz,
2-(2-Hyd r oxy-5-m eth oxyp h en yl)-1H-ben zoim id a zole-5-
ca r boxa m id in e (6g). Synthesized using general procedure
1
A. H NMR (DMSO-d6) δ 9.44 (br s, 2H), 9.13 (br s, 2H), 8.21
(s, 1H), 7.92-7.89 (m, 2H), 7.76 (d, J ) 8.4 Hz, 1H), 7.09 (s,
2H), 3.80 (s, 3H). 13C NMR (DMSO-d6) δ 165.8, 152.2, 151.5,

2766 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17
Verner et al.
137.1, 134.0, 124.1, 123.7, 122.0, 118.4, 115.4, 114.6, 111.5,
109.9, 56.1. MS-Sciex m/z 283.2 [MH+]. Anal. (C15H14N4O2‚
1.5HCl‚0.5H2O) C, H, N.
2-(2-H yd r oxy-5-ch lor ob ip h en yl-3-yl)-1H -b en zoim id -
a zole-5-ca r b oxa m id in e (6s). Synthesized using general
procedure A. H NMR (DMSO-d6) δ 9.47 (br s, 2H), 9.22 (br s,
2H), 8.42 (d, J ) 2.2 Hz, 1H), 8.25 (s, 1H), 7.87 (d, J ) 8.6 Hz,
1H), 7.78 (d, J ) 8.7 Hz, 1H), 7.64 (d, J ) 7.2 Hz, 2H), 7.50-
7.36 (m, 4H). 13C NMR (DMSO-d6) δ 165.9, 154.5, 153.5, 136.2,
132.3, 131.6, 129.3, 128.2, 127.7, 125.4, 123.1, 122.4, 113.6.
MS-Sciex m/z 363.0 [35Cl, MH+]. Anal. (C20H15ClN4O‚1.5HCl)
C, H, N.
1
2-(2-Hyd r oxy-3-n itr op h en yl)-1H-ben zoim id a zole-5-ca r -
boxa m id in e (6h ). Synthesized using general procedure D. 1H
NMR (DMSO-d6) δ 9.46 (br s, 2H), 9.19 (br s, 2H), 8.63 (dd, J
) 7.7, 1.3 Hz, 1H), 8.27 (s, 1H), 8.08 (dd, J ) 8.2, 1.3 Hz, 1H),
7.90 (d, J ) 8.6 Hz, 1H), 7.79 (dd, J ) 8.6, 1.7 Hz, 1H), 7.21 (t,
J ) 7.9 Hz, 1H). 13C NMR (DMSO-d6) δ 166.3, 153.5, 152.9,
139.7, 139.3, 137.2, 132.5, 128.3, 123.9, 123.2, 119.2, 116.9,
115.3. MS-Sciex m/z 298.0 [MH+]. Anal. (C14H11N5O3‚1.5HCl)
C, H, N.
2-(2-Hyd r oxy-3-m eth oxyp h en yl)-1H-ben zoim id a zole-5-
ca r boxa m id in e (6l). Synthesized using general procedure A.
1H NMR (DMSO-d6) δ 9.51 (br s, 2H), 9.27 (br s, 2H), 8.22 (s,
1H), 7.90-7.79 (m, 3H), 7.18 (d, J ) 8.2 Hz, 1H), 6.97 (t, J )
8.2 Hz, 1H), 3.83 (s, 3H). 13C NMR (DMSO-d6) δ 166.4, 152.7,
149.1, 148.5, 138.1, 135.2, 124.4, 123.9, 120.1, 116.1, 115.1,
111.4, 56.6. MS-Sciex m/z 283.2 [MH+]. Anal. (C15H14N4O2‚
1.5HCl‚0.5H2O) C, H, N.
2-(2-Hyd r oxybip h en yl-3-yl)-1H-ben zoim id a zole-5-ca r -
boxa m id in e (6m ). Synthesized using general procedure D.
1H NMR (DMSO-d6) δ 9.39 (br s, 2H), 9.08 (br s, 2H), 8.28-
8.13 (m, 2H), 7.85 (d, J ) 8.1 Hz, 1H), 7.74 (d, J ) 8.3 Hz,
1H), 7.65-7.56 (m, 2H), 7.52-7.31 (m, 4H), 7.13 (t, J ) 7.6
Hz, 1H). 13C NMR (DMSO-d6) δ 166.0, 155.6, 154.7, 137.5,
133.2, 129.6, 129.2, 128.0, 127.1, 126.3, 122.8, 122.0, 119.3,
112.2. MS-ESI m/z 328.9 [MH+]. Anal. (C20H16N4O‚HCl‚
0.5H2O) C, H, N.
2-(2-H yd r oxy-4-m et h ylp h en yl)-1H -b en zoim id a zole-5-
ca r boxa m id in e (6n ). Synthesized using general procedure
D. 1H NMR (DMSO-d6) δ 9.56 (br s, 2H), 9.29 (br s, 2H), 8.25-
8.22 (m, 2H), 7.95 (d, J ) 8.7 Hz, 1H), 7.84 (d, J ) 8.7 Hz,
1H), 7.09 (s, 1H), 6.93 (d, J ) 8.2 Hz, 1H), 2.35 (s, 3H). 13C
NMR (DMSO-d6) δ 166.3, 158.5, 152.0, 145.8, 137.3, 134.0,
129.2, 124.7, 124.3, 121.5, 118.0, 115.7, 115.0, 108.0, 21.9. MS-
Sciex m/z 266.8 [MH+]. Anal. (C15H14N4O‚2.0HCl) C, H, N.
3-Br om o-2-h ydr oxyben zaldeh yde (8). A solution of 2-bro-
mophenol 7 (6.34 g, 36.6 mmol) in CH3CN (170 mL) was
treated with MgCl2 (5.23 g, 55 mmol), paraformaldehyde (7.4
g, 0.25 mol), and finally Et3N (12.8 mL, 92 mmol). The mixture
was refluxed for 3 h and then allowed to cool. The solution
was diluted with 1 N HCl (200 mL) and then extracted with
Et2O (2 × 200 mL). The organic extracts were washed with
brine, dried over MgSO4, filtered, and concentrated under
reduced pressure to give 8 (6.9 g, 94%) as a yellow oil. This oil
was used in subsequent reactions without further purification.
1H NMR (CDCl3) δ 11.63 (s, 1H), 9.87 (s, 1H), 7.79 (d, J ) 8.1
Hz, 1H), 7.56 (d, J ) 7.5 Hz, 1H), 6.96 (t, J ) 7.8 Hz, 1H).
2-Hyd r oxybip h en yl-3-ca r ba ld eh yd e (10). Biphenyl-2-ol
9 (17 g, 0.1 mol) was dissolved in THF (200 mL) and then
treated with Et3N (27 mL, 0. 2 mol) followed by MgCl2 (14.3
g, 0.2 mol). The mixture was stirred for 20 min, and then
paraformaldehyde (15 g, 0.5 mol) was added. The resultant
mixture was refluxed for 3 h and then diluted with EtOAc (200
mL). The solution was washed with 1 N HCl followed by brine,
dried over Na2SO4, filtered, and concentrated under reduced
pressure. The crude product, 10, (∼100%) was sufficiently pure
(97% by analytical HPLC) to be used in subsequent reactions.
MS-Sciex m/z 199.0 [MH+].
5-Ch lor o-2-h yd r oxybip h en yl-3-ca r ba ld eh yd e (11). The
crude 2-hydroxybiphenyl-3-carbaldehyde 10 (2 g, 10 mmol) was
dissolved in AcOH (30 mL) and then treated with N-chloro-
succinimide (NCS) (1.5 g, 11 mmol). The mixture was then
refluxed for 3 h, allowed to cool, and concentrated under
reduced pressure. The residue was dissolved in Et2O (100 mL)
and the insoluble material removed by filtration. The filtrate
was washed with 1 N HCl, followed by brine, dried over
Na2SO4, filtered, and concentrated under reduced pressure to
2-(2-H yd r oxy-4-d iet h yla m in op h en yl)-1H -b en zoim id -
a zole-5-ca r boxa m id in e (6o). Synthesized using general
1
procedure A. H NMR (DMSO-d6) δ 9.59 (br s, 2H), 9.32 (br s,
1
2H), 8.20 (d, J ) 8.4 Hz, 1H), 8.14 (s, 1H), 7.87 (d, J ) 8.4 Hz,
1H), 7.79 (d, J ) 8.8 Hz, 1H), 6.59 (s, 1H), 6.49 (d, J ) 9.1 Hz,
1H), 3.39 (q, J ) 6.6 Hz, 4H), 1.12 (t, J ) 6.6 Hz, 6H). 13C
NMR (DMSO-d6) δ 165.8, 159.9, 152.8, 149.9, 134.8, 131.1,
130.9, 124.6, 124.1, 113.6, 113.5, 105.4, 97.3, 95.4, 44.4, 12.4.
MS-Sciex m/z 324.0 [MH+]. Anal. (C18H21N5O‚2.0HCl‚0.5H2O)
C, H, N.
give 11 (2.3 g, 98%). H NMR (CDCl3) δ 11.42 (s, 1H), 9.89 (s,
1H), 7.57-7.52 (m, 4H), 7.48-7.38 (m, 3H). MS-Sciex m/z 232.6
[35Cl, MH+].
3-Br om o-2-h yd r oxy-5-m eth ylben za ld eh yd e (13). A so-
lution of 2-hydroxy-5-methylbenzaldehyde 12 (0.5 g, 3.7 mmol)
in AcOH (5 mL) was treated dropwise with bromine (0.21 mL,
4.0 mmol). The resultant mixture was stirred for 1.5 h, and
then water was added. The precipitate that formed was
collected by filtration and washed with water. The solid was
dried to give 13 (0.55 g, 70%). 1H NMR (CDCl3) δ 11.40 (s,
1H), 9.81 (s, 1H), 7.61 (s, 1H), 7.33 (s, 1H), 2.34 (s, 3H). 13C
NMR (CDCl3) δ 195.9, 155.9, 140.6, 132.8, 130.4, 120.9, 110.7,
20.0. MS-Sciex m/z 214.6 [79Br, MH+].
2-(2,6-Dih yd r oxyp h en yl)-1H-ben zoim id a zole-5-ca r box-
a m id in e (6p ). Synthesized using general procedure D. 1H
NMR (DMSO-d6) δ 9.54 (br s, 2H), 9.32 (br s, 2H), 8.31 (d, J
) 1.3 Hz, 1H), 8.01 (d, J ) 8.7 Hz, 1H), 7.82 (dd, J ) 8.5, 1.6
Hz, 1H), 7.30 (t, J ) 8.3 Hz, 1H), 6.72 (d, J ) 8.1 Hz, 2H). 13C
NMR (DMSO-d6) δ 166.1, 158.7, 150.0, 136.1, 134.2, 133.1,
123.8, 123.6, 115.6, 114.7, 107.0, 98.9. MS-Sciex m/z 269.0
[MH+]. Anal. (C14H12N4O2‚2.0HCl) C, H, N.
2-(5-Br om o-1H-in d ol-2-yl)p h en ol (15). Prepared accord-
ing to the procedure of Iwanowicz and co-workers21 from 14
and 28.
2-(1-H yd r oxyn a p h t h a len -2-yl)-1H -b en zoim id a zole-5-
ca r boxa m id in e (6q). Synthesized using general procedure
2-(2-Hydr oxyph en yl)-1H-in dole-5-car bon itr ile (16). The
5-bromo-2-phenyl-1H-indole 15 (1.4 g, 4.86 mmol) was dis-
solved in quinoline (10 mL) and treated with CuCN (1.31 g,
14.6 mmol). The mixture was heated at 220 °C under nitrogen
for 90 min and then cooled. The mixture was treated with
EtOAc (100 mL) and 1 N HCl (100 mL). The organic layer was
separated, washed with brine, dried over MgSO4, filtered, and
concentrated under reduced pressure to give a brown solid.
The solid was subjected to chromatography over silica gel
eluting with 30% EtOAc in hexanes to give 16 (0.56 g, 49%).
1H NMR (DMSO-d6) δ 11.70 (s, 1H), 10.30 (s, 1H), 8.04 (s, 1H),
7.75 (d, J ) 7.7 Hz, 1H), 7.60 (d, J ) 8.4 Hz, 1H), 7.39 (dd, J
) 8.4, 1.8 Hz, 1H), 7.19 (m, 1H), 7.09 (s, 1H), 7.01 (d, J ) 8.1
Hz, 1H), 6.92 (t, J ) 7.7 Hz, 1H).
1
A. H NMR (DMSO-d6) δ 9.47 (br s, 2H), 9.22 (br s, 2H), 8.38
(dd, J ) 7.4, 1.5 Hz, 1H), 8.29 (d, J ) 8.6 Hz, 1H), 8.25 (d, J
) 0.9 Hz, 1H), 7.95-7.87 (m, 2H), 7.78 (dd, J ) 8.5, 1.7 Hz,
1H), 7.67-7.55 (m, 3H). 13C NMR (DMSO-d6) δ 166.0, 156.0,
154.9, 135.2, 128.5, 127.8, 126.0, 124.6, 123.0, 122.9, 122.2,
119.0, 115.6, 114.5, 105.0. MS-Sciex m/z 303.2 [MH+]. Anal.
(C18H14N4O‚2.0HCl) C, H, N.
2-(2-Hyd r oxy-3-br om o-5-m eth ylph en yl)-1H-ben zoim id -
a zole-5-ca r boxa m id in e (6r ). Synthesized using general
1
procedure A. H NMR (DMSO-d6) δ 9.41 (br s, 2H), 9.13 (br s,
2H), 8.21 (s, 1H), 8.09 (s, 1H), 7.84 (d, J ) 8.4 Hz, 1H), 7.74
(d, J ) 8.4 Hz, 1H), 7.57 (s, 1H), 2.31 (s, 3H). 13C NMR (DMSO-
d6) δ 165.9, 153.6, 152.7, 139.5, 136.9, 136.0, 129.6, 126.7,
123.1, 122.4, 116.1, 114.6, 112.7, 110.5, 19.8. MS-Sciex m/z
345.0 [79Br, MH+], 346.8 [81Br, MH+]. Anal. (C15H13BrN4O‚
1.5HCl) C, H, N.
2-(2-Hydr oxyph en yl)-1H-in dole-5-car boxam idin e (19).21
The 2-(2-hydroxyphenyl)-1H-indole-5-carbonitrile 16 (0.56 g,
2.4 mmol) was dissolved in absolute EtOH (30 mL), and excess

Serine Protease Inhibitors with a Unique Binding Mode
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17 2767
50% aqueous hydroxylamine (1.0 mL) was added. The mixture
was stirred for 18 h at room temperature and then refluxed
for 3 h. The mixture was cooled and diluted with EtOAc (150
mL). The organic phase was washed with saturated NaHCO3
solution, dried over MgSO4, filtered, and concentrated under
reduced pressure. The crude solid, 17, was dissolved in AcOH
(15 mL) and treated with acetic anhydride (0.25 mL, 2.64
mmol) followed by THF (20 mL). The mixture was stirred for
30 min to give acetate 18 (observed by LCMS), and then 10%
Pd/C (50 mg) catalyst was added. The mixture was stirred
under hydrogen for 1.5 h. The mixture was then filtered
through Celite and the filtrate concentrated under reduced
pressure. The crude residue was subjected to preparative
HPLC to give 19 (0.24 g, 34%). 1H NMR (DMSO-d6) δ 11.73
(s, 1H), 10.39 (s, 1H), 9.16 (s, 2H), 8.84 (s, 2H), 8.10 (s, 1H),
7.79 (d, J ) 7.7 Hz, 1H), 7.63 (d, J ) 8.4 Hz, 1H), 7.51 (d, J )
8.8 Hz, 1H), 7.19 (t, J ) 8.0 Hz, 1H), 7.14 (s, 1H), 7.05 (d, J )
8.0 Hz, 1H), 6.92 (t, J ) 8.0 Hz, 1H). 13C NMR (DMSO-d6) δ
166.6, 154.7, 139.2, 138.1, 129.0, 127.9, 127.8, 120.9, 120.4,
119.4, 118.1, 117.9, 116.7, 111.8, 101.7. MS-Sciex m/z 252.0
[MH+]. Anal. (C15H13N3O‚HCl‚0.25H2O) C, H, N, Cl.
4-Hyd r a zin oben za m id in e (21). 4-Aminobenzamidine di-
hydrochloride 20 (10.4 g, 50 mmol) was suspended in 6 N HCl
(100 mL) and cooled to 0 °C using an ice bath. A solution of
NaNO2 (3.45 g, 50 mmol) in water (25 mL) was added dropwise
with vigorous stirring of the reaction mixture. After the
addition was completed, the mixture was stirred for 20 min.
Tin(II) chloride dihydrate (22.5 g, 100 mmol) was added in
portions and the final mixture stirred for 30 min at 0 °C. The
mixture was concentrated under reduced pressure and tritu-
rated with absolute EtOH. The decanted solution was concen-
trated again and triturated with absolute EtOH once more.
The residual solid after decantation of the absolute EtOH was
collected by filtration and dried. MS-ESI m/z 252.0 [MH+].
NMR indicated a 60:40 mixture of starting material and
product 4-hydrazinobenzamidine dihydrochloride. This was
used without further purification although a 2-fold excess of
this material had to be used in subsequent reactions.
d6) δ 11.86 (s, 1H), 9.65 (br s, 1H), 9.11 (br s, 2H), 8.96 (br s,
2H), 8.11 (s, 1H), 7.71 (dd, J ) 7.7, 1.5 Hz, 1H), 7.52 (m, 3H),
7.09 (s, 1H), 6.91 (t, J ) 7.9 Hz, 1H). 13C NMR (DMSO-d6) δ
166.4, 150.7, 139.2, 136.9, 132.4, 127.8, 127.7, 122.3, 121.8,
121.3, 120.8, 118.4, 113.3, 111.8, 103.1. MS-Sciex m/z 329.4
[79Br, M+], 331.2 [81Br, M+]. Anal. (C15H12BrN3O‚C2HF3O2)
Calcd: C, 45.97; H, 2.95; N, 9.46; Br, 17.99. Found: C, 46.08;
H, 2.97; N, 9.94; Br, 17.88.
2-(3,5-Dibr om o-2-h yd r oxyp h en yl)-1H-in d ole-5-ca r box-
a m id in e (22c). Synthesized using a commercially available
1-(3,5-dibromo-2-hydroxyphenyl)ethanone and 21. 1H NMR
(DMSO-d6) δ 12.00 (s, 1H), 10.03 (br s, 1H), 9.19 (br s, 2H),
8.84 (br s, 2H), 8.17 (s, 1H), 7.97 (d, J ) 2.4 Hz, 1H), 7.78 (d,
J ) 2.3 Hz, 1H), 7.63-7.56 (m, 2H), 7.26 (s, 1H). 13C NMR
(DMSO-d6) δ 167.0, 150.9, 139.9, 135.8, 134.1, 130.2, 128.1,
124.5, 122.2, 121.8, 119.1, 115.0, 112.7, 112.4, 104.9. MS-Sciex
m/z 407.6 [2 × 79Br, MH+]. Anal. (C15H11Br2N3O‚HCl‚0.5H2O)
C, H, N, Cl.
2-(3-Br om o-2-h yd r oxy-5-n itr op h en yl)-1H-in d ole-5-ca r -
boxa m id in e (22d ). Synthesized using 31 and 21. 1H NMR
(DMSO-d6) δ 12.32 (s, 1H), 9.20 (br s, 2H), 8.85 (br s, 2H),
8.64 (d, J ) 2.7 Hz, 1H), 8.40 (d, J ) 2.7 Hz, 1H), 8.19 (s, 1H),
7.65 (d, J ) 8.4 Hz, 1H), 7.59 (d, J ) 8.4 Hz, 1H), 7.31 (s, 1H).
13C NMR (DMSO-d6) δ 166.4, 158.2, 139.6, 139.3, 135.1, 127.6,
127.3, 122.7, 121.7, 121.4, 121.3, 118.7, 112.8, 111.9, 104.4.
MS-Sciex m/z 374.6 [79Br, MH+], 376.6 [81Br, MH+]. Anal.
(C15H11BrN4O3‚HCl‚0.5H2O) C, H, N.
2-(2-H yd r oxyb ip h en yl-3-yl)-1H -in d ole-5-ca r b oxa m i-
d in e (22f). Synthesized using 35 and 21. 1H NMR (DMSO-
d6) δ 11.96 (s, 1H), 9.24 (br s, 2H), 9.02 (br s, 1H), 8.97 (s, 1H),
8.18 (s, 1H), 7.76 (dd, J ) 7.7, 1.7 Hz, 1H), 7.65 (d, J ) 8.7
Hz, 1H), 7.59 (dd, J ) 3.0, 1.7 Hz, 1H), 7.55 (m, 2H), 7.48 (t,
J ) 7.6 Hz, 2H), 7.41-7.35 (m, 1H), 7.25 (dd, J ) 7.5, 1.7 Hz,
1H), 7.14 (s, 1H), 7.11 (t, J ) 7.7 Hz, 1H). 13C NMR (DMSO-
d6) δ 166.5, 150.8, 139.2, 138.3, 138.0, 131.9, 130.5, 129.3,
128.4, 127.8, 127.1, 121.7, 121.0, 120.5, 118.1 (× 2), 111.7,
102.5. MS-Sciex m/z 328.2 [MH+]. Anal. (C21H17N3O‚HCl) C,
H, N, Cl.
Gen er a l P r oced u r e for th e Syn th esis of In d oles 22a -
k: Exam ple, 2-Hydr oxy-5-m eth ylbiph en yl-3-yl)-1H-in dole-
5-ca r b oxa m id in e (22e). 1-(2-Hydroxy-5-methylbiphenyl-3-
yl)ethanone 26 (0.32 g, 1.4 mmol) was added to absolute EtOH
(10 mL), and 4-hydrazinobenzamidine dihydrochloride 21 (0.66
g, 3.0 mmol) was then added. Triethylamine (1 mL, 7.2 mmol,
5.1 equiv) was added and the resultant solution refluxed for 2
h. The solution was allowed to cool to room temperature and
then diluted with Et2O (100 mL). The precipitate that formed
was collected by filtration and dried to yield the crude
hydrazone as the hydrochloride salt (1.66 g). This material
typically contains some triethylamine hydrochloride contami-
nants. The crude hydrazone (1 g) was added to PPA (5 mL)
and heated at 140-160 °C for 4.5 h. The reaction was
monitored by HPLC for completion. Upon completion the
mixture was allowed to cool and then treated with 1 N HCl.
The precipitate that formed was collected by filtration and
subjected to preparative HPLC to give 22e (150 mg, 28%) as
2-(5-Br om o-2-h yd r oxybip h en yl-3-yl)-1H-in d ole-5-ca r -
boxa m id in e (22g). Synthesized using 37 and 21. 1H NMR
(DMSO-d6) δ 11.98 (s, 1H), 9.18 (br s, 2H), 9.14 (br s, 2H),
8.18 (s, 1H), 7.94 (d, J ) 2.5 Hz, 1H), 7.66-7.37 (m, 8H), 7.25
(d, J ) 1.0 Hz, 1H). 13C NMR (DMSO-d6) δ 167.0, 159.6, 159.1,
151.0, 139.8, 137.5, 136.8, 134.6, 132.7, 129.9, 129.0, 128.3,
124.5, 121.9, 121.5, 119.1, 113.0, 112.4, 104.2. MS-Sciex m/z
405.8 [79Br, MH+]. Anal. (C21H16BrN3O‚C2HF3O2) C, H, N, Br.
2-(5-Ch lor o-2-h yd r oxybip h en yl-3-yl)-1H-in d ole-5-ca r -
boxa m id in e (22h ). Synthesized using 36 and 21. 1H NMR
(DMSO-d6) δ 11.96 (s, 1H), 9.19 (br s, 1H), 9.17 (br s, 1H),
8.92 (br s, 2H), 8.12 (s, 1H), 7.77 (d, J ) 2.5 Hz, 1H), 7.56-
7.34 (m, 7H), 7.20 (d, J ) 2.5 Hz, 1H), 7.18 (s, 1H). 13C NMR
(DMSO-d6) δ 166.3, 149.8, 139.2, 137.0, 136.4, 133.7, 129.3,
128.8, 128.5, 127.7, 126.5, 124.7, 123.5, 121.4, 120.9, 118.3 (×
2), 111.8, 103.6. MS-Sciex m/z 362.0 [35Cl, MH+]. Anal. (C21H16
ClN3O‚HCl‚0.5H2O) C, H, N.
-
1
a brown solid. H NMR (DMSO-d6) δ 11.95 (s, 1H), 9.26 (br s,
2-(5-Nit r o-2-h yd r oxyb ip h en yl-3-yl)-1H -in d ole-5-ca r -
b oxa m id in e (22i). Synthesized using 32 and 21. 1H NMR
(DMSO-d6) δ 12.25 (br s, 1H), 9.19 (br s, 2H), 8.80 (br s, 2H),
8.64 (d, J ) 3.0 Hz, 1H), 8.19 (s, 1H), 8.03 (d, J ) 2.7 Hz, 1H),
7.68-7.44 (m, 7H), 7.31 (s, 1H). 13C NMR (DMSO-d6) δ 166.4,
157.9, 140.4, 139.3, 136.4, 135.6, 131.7, 129.3, 128.7, 128.1,
127.7, 125.2, 122.8, 121.8, 121.6, 121.2, 118.6, 111.9, 104.2.
MS-Sciex m/z 373.0 [MH+]. Anal. (C21H16N4O3‚HCl‚H2O) C, H,
N.
2H), 9.09 (br s, 1H), 8.72 (br s, 1H), 8.18 (s, 1H), 7.66-7.54
(m, 5H), 7.46 (t, J ) 7.5 Hz, 2H), 7.39-7.33 (m, 1H), 7.13 (s,
1H), 7.06 (d, J ) 1.5 Hz, 1H), 2.34 (s, 3H). 13C NMR (DMSO-
d6) δ 166.5, 148.5, 139.2, 138.5, 138.2, 131.9, 131.1, 129.7,
129.3, 128.3, 128.0, 127.8, 127.0, 121.6, 121.0, 120.4, 118.1,
111.7, 102.4, 20.3. MS-Sciex m/z 341.8 [MH+]. Anal. (C22H19N3O‚
2.0HCl) C, H, N.
2-(3-Br om o-2-h yd r oxy-5-m et h ylp h en yl)-1H -in d ole-5-
ca r boxa m id in e (22a ). Synthesized using 24 and 21. 1H NMR
(DMSO-d6) δ 11.97 (s, 1H), 9.43 (br s, 1H), 9.25 (br s, 2H),
9.06 (br s, 2H), 8.19 (s, 1H), 7.65-7.56 (m, 3H), 7.39 (s, 1H),
7.15 (s, 1H), 2.30 (s, 3H). 13C NMR (DMSO-d6) δ 166.5, 148.4,
139.2, 137.0, 132.5, 131.0, 128.1, 127.7, 122.0, 121.2, 120.8,
118.3, 113.3, 111.7, 103.1, 19.8. MS-Sciex m/z 344.0 [79Br,
MH+]. Anal. (C16H14BrN3O‚HCl) C, H, N.
2-Bip h en yl-3-yl-1H-in d ole-5-ca r boxa m id in e (22j). Syn-
thesized using 27 and 21. 1H NMR (DMSO-d6) δ 9.30 (br s,
2H), 9.11 (br s, 2H), 8.29 (s, 1H), 8.19 (s, 1H), 7.97 (d, J ) 7.7
Hz, 1H), 7.83 (d, J ) 7.4 Hz, 2H), 7.68-7.57 (m, 4H), 7.51 (t,
J ) 7.6 Hz, 2H), 7.41 (t, J ) 7.4 Hz, 1H), 7.24 (d, J ) 1.7 Hz,
1H). 13C NMR (DMSO-d6) δ 167.1, 141.5, 140.8, 140.6, 140.3,
132.5, 130.2, 129.5, 128.7, 128.3, 127.5, 126.9, 125.2, 124.1,
121.9, 121.6, 119.2, 112.3, 100.6. MS-Sciex m/z 311.6 [MH+].
Anal. (C21H17N3‚HCl‚0.5H2O) C, H, N.
2-(3-Br om o-2-h yd r oxyp h en yl)-1H-in d ole-5-ca r boxa m i-
1
d in e (22b). Synthesized using 29 and 21. H NMR (DMSO-

2768 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17
2-(2-Met h oxyb ip h en yl-3-yl)-1H -in d ole-5-ca r b oxa m i-
Verner et al.
δ 8.70 (d, J ) 2.7 Hz, 1H), 8.34 (ddd, J ) 9.1, 2.7, 0.5 Hz, 1H),
7.08 (d, J ) 9.1 Hz, 1H), 2.73 (s, 3H). MS-Sciex m/z 179.6
[M-H+].
1
d in e (22k ). Synthesized using 39 and 21. H NMR (DMSO-
d6) δ 12.08 (s, 1H), 9.26 (br s, 2H), 9.09 (br s, 1H), 8.20 (s, 1H),
7.90 (dd, J ) 6.7, 2.5 Hz, 1H), 7.69-7.59 (m, 4H), 7.49 (t, J )
7.0 Hz, 2H), 7.44-7.34 (m, 3H), 7.17 (s, 1H), 3.27 (s, 3H). 13C
NMR (DMSO-d6) δ 166.5, 154.5, 139.5, 137.7, 136.8, 135.8,
130.9, 128.9, 128.4, 128.1, 127.8, 127.5, 125.2, 124.8, 121.3,
120.9, 118.4, 111.9, 102.8, 59.9. MS-Sciex m/z 341.4 [M+]. Anal.
(C22H19N3O‚HCl‚0.25H2O) C, H, N.
1-(3-Br om o-2-h yd r oxy-5-m eth ylp h en yl)eth a n on e (24).
1-(2-Hydroxy-5-methylphenyl)ethanone 23 (10.0 g, 66.6 mmol)
was dissolved in MeOH (200 mL) and treated with N-
bromosuccinimide (NBS) (11.85 g, 66.6 mmol). The mixture
was stirred for 3 h and then water (100 mL) was added. The
resultant pale yellow solid that formed was collected by
filtration and washed with water. The solid was dried to give
24 (13 g, 85%). 1H NMR (DMSO-d6) δ 7.57 (d, J ) 2.2 Hz, 1H),
7.48 (dd, J ) 2.0, 0.7 Hz, 1H), 2.63 (s, 3H), 2.30 (s, 3H). MS-
Sciex m/z 229.0 [79Br, MH+], 231.0 [81Br MH+].
1-(3-Br om o-2-h ydr oxy-5-n itr oph en yl)eth an on e (31). The
acetophenone 30 (1.8 g, 10 mmol) was dissolved in CH3CN (50
mL) and treated with NBS (1.78 g, 10 mmol). The resultant
solution was stirred overnight and then diluted with EtOAc
(250 mL). The organic solution was washed with brine (2 ×
200 mL), water (2 × 200 mL), and brine (2 × 200 mL) again.
The organic solution was then dried over Na2SO4, filtered, and
concentrated under reduced pressure. The residue was dis-
solved in hot MeOH and then treated with water. The solid
precipitate that formed was collected by filtration and dried
to give 31 (1.46 g, 56%). 1H NMR (CDCl3) δ 8.74 (d, J ) 2.7
Hz, 1H), 8.68 (d, J ) 2.7 Hz, 1H), 2.79 (s, 3H). MS-Sciex m/z
257.8 [79Br, M-H+], 259.4 [81Br, M-H+].
1-(2-Hyd r oxy-5-n itr obip h en yl-3-yl)eth a n on e (32). The
nitroarene 31 (464 mg, 1.8 mmol) was mixed with phenyl-
boronic acid (272 mg, 2.2 mmol) in THF (15 mL). The mixture
was then treated with tetrakis(triphenylphosphine)palladium(0)
(103 mg, 0.089 mmol) and 10% Na2CO3 solution (2 mL). The
mixture was heated at 80 °C for 2 h and then a further portion
of tetrakis(triphenylphosphine)palladium(0) (103 mg, 0.089
mmol) was added. After the mixture was heated for another 1
h, 10% Pd/C catalyst (150 mg) was added and heating
continued for 1.5 h. The mixture was allowed to cool, diluted
with EtOAc (60 mL), and then filtered through Celite. The
eluant was washed with brine, 1 N HCl, and finally brine
again. The organic solution was then dried over Na2SO4,
filtered, and concentrated under reduced pressure. The crude
product was subjected to chromatography over silica gel
eluting with 50% EtOAc in hexanes to give 32 (322 mg, 70%).
1H NMR (CDCl3) δ 8.70 (d, J ) 2.7 Hz, 1H), 8.43 (dd, J ) 2.7,
0.7 Hz, 1H), 7.60-7.55 (m, 2H), 7.50-7.41 (m, 3H), 2.79 (s,
3H). 13C NMR (CDCl3) δ 204.4, 164.7, 139.2, 134.6, 132.5,
131.1, 129.2, 128.5, 128.4, 125.8, 118.6, 26.9.
1-(2-Hydr oxy-5-m eth ylbiph en yl-3-yl)eth an on e (26). 1-(3-
Bromo-2-hydroxy-5-methylphenyl)ethanone 24 (0.46 g, 2 mmol)
and phenylboronic acid (0.24 g, 2 mmol) were dissolved in
ethylene glycol dimethyl ether (DME) (10 mL). NaHCO3 (0.50
g, 6 mmol) in water (10 mL) was added followed by tetrakis-
(triphenylphosphine)palladium(0) (0.12 g, 0.1 mmol). The
reaction mixture was refluxed for 10 min and then heated at
75 °C overnight. Most of the DME was then removed under
reduced pressure. The resultant mixture was diluted with Et2O
(100 mL), washed with water and brine, and then dried over
MgSO4. The organic solution was filtered and concentrated
under reduced pressure. The crude solid was subjected to
chromatography over silica gel eluting with 10% EtOAc in
hexanes to give 26 (0.32 g, 71% yield) as a yellow solid. 1H
NMR (DMSO-d6) δ 7.58-7.52 (m, 3H), 7.45-7.31 (m, 4H), 2.66
(s, 3H), 2.35 (s, 3H). MS-Sciex m/z 224.8 [M-H+].
1-Bip h en yl-3-yl-eth a n on e (27). 1-(3-Bromophenyl)etha-
none 25 (2.0 g, 10 mmol) was mixed with phenylboronic acid
(1.2 g, 10 mmol) in THF (100 mL). Tetrakis(triphenylphos-
phine)palladium(0) (1.1 g, 1 mmol) was added followed by
K2CO3 (4.1 g, 30 mmol) and the resultant mixture refluxed
overnight. The mixture was allowed to cool and then filtered
through Celite. The eluant was diluted with EtOAc (500 mL)
and washed with saturated NaHCO3 solution (2 × 300 mL)
followed by brine. The organic solution was dried over Na2SO4,
filtered, and concentrated under reduced pressure. The residue
was subjected to chromatography over silica gel eluting with
10% EtOAc in hexanes to give 27 (1.2 g, 61%). 1H NMR (CDCl3)
δ 8.18 (t, J ) 2.0 Hz, 1H), 7.98-7.91 (m, 2H), 7.76-7.72 (m,
2H), 7.62 (t, J ) 7.9 Hz, 1H), 7.54-7.39 (m, 3H), 2.66 (s, 3H).
1-(3-Br om o-2-h yd r oxyp h en yl)eth a n on e (29). 1-(2-Hy-
droxyphenyl)ethanone 28 (1.36 g, 10 mmol) was dissolved in
AcOH (20 mL) and treated with NBS (1.78 g, 10 mmol). The
mixture was stirred overnight and then concentrated under
reduced pressure. The residual solid was dissolved in Et2O (100
mL) and washed with brine. The organic solution was dried
over Na2SO4, filtered, and concentrated under reduced pres-
sure. The residue was subjected to chromatography over silica
gel eluting with 5% EtOAc in hexanes to give 29 (198 mg, 9%).
1H NMR (CDCl3) δ 7.74 (dd, J ) 7.9, 2.2 Hz, 1H), 7.72 (dd, J
) 7.9, 1.5 Hz, 1H), 6.81 (t, J ) 7.9 Hz, 1H), 2.65 (s, 3H).
1-(2-Hydr oxy-5-n itr oph en yl)eth an on e (30). 1-(2-Hydroxy-
phenyl)ethanone 28 (13.6 g, 0.1 mol) was dissolved in AcOH
(100 mL) and cooled to 0 °C using an ice bath. Fuming HNO3
(6.3 mL, 0.15 mol) was added over 30 min, and the solution
was stirred overnight at room temperature. The solution was
concentrated under reduced pressure and the residual oil
treated with water to yield a solid. The solid was collected and
redissolved in Et2O (200 mL). The organic solution was washed
with water (5 × 100 mL) and then brine. The organic layer
was dried over Na2SO4, filtered, and concentrated under
reduced pressure to yield the crude product. The crude product
was subjected to chromatography over silica gel eluting with
10% EtOAc in hexanes to give 30 (6.5 g, 36%). 1H NMR (CDCl3)
2-Hyd r oxybip h en yl-3-ca r boxylic Acid N-Meth oxy-N-
m eth yla m id e (34). 2-Hydroxy-3-phenylbenzoic acid 33 (5.35
g, 25 mmol) was dissolved in dry DMF (50 mL) and cooled to
0 °C. PyBOP (13.65 g, 26.3 mmol) was added followed by N,O-
dimethylhydroxylamine hydrochloride (2.56 g, 26.8 mmol) and
finally N,N-diisopropylethylamine (DIEA) (11 mL, 63 mmol).
The mixture was stirred for 3 h and allowed to warm to room
temperature. The solvent was removed under reduced pressure
and the residue dissolved in 20% 2-propanol in CH2Cl2. The
solution was washed with saturated NaHCO3, dried over
MgSO4, filtered, and concentrated under reduced pressure to
give a golden oil. The oil was subjected to chromatography over
silica gel eluting with 20% EtOAc in hexanes to give 34 (4.5
g, 71%) as a white solid. 1H NMR (CDCl3) δ 11.42 (s, 1H), 7.94
(dd, J ) 8.1, 1.7 Hz, 1H), 7.61-7.58 (m, 1H), 7.48-7.36 (m,
6H), 6.93 (t, J ) 2.9 Hz, 1H), 3.70 (s, 3H), 3.45 (s, 3H).
1-(2-Hyd r oxybip h en yl-3-yl)eth a n on e (35). 2-Hydroxy-
biphenyl-3-carboxylic acid N-methoxy-N-methylamide 34 (1.96
g, 7.6 mmol) was dissolved in dry Et2O (40 mL) and cooled to
0 °C. A 1.4 M solution of MeLi in Et2O (16.3 mL, 22.8 mmol)
was added dropwise, and the heterogeneous reaction mixture
was then stirred for 1 h. The reaction mixture was quenched
with dilute 1 N HCl. The organic layer was separated, washed
with brine, dried over MgSO4, filtered, and concentrated under
reduced pressure. The crude product was subjected to chro-
matography over silica gel eluting with 5% EtOAc in hexanes
1
to give 35 (1.07 g, 66%) as a white solid. H NMR (CDCl3) δ
7.75 (dd, J ) 8.0, 1.8 Hz, 1H), 7.59-7.53 (m, 3H), 7.46-7.34
(m, 3H), 6.97 (t, J ) 7.7 Hz, 1H), 2.68 (s, 3H).
1-(5-Ch lor o-2-h ydr oxybiph en yl-3-yl)eth an on e (36). 1-(2-
Hydroxybiphenyl-3-yl)ethanone 35 (1.03 g, 4.9 mmol) was
dissolved in DMF (20 mL) and then treated with NCS (1.3 g,
9.7 mmol). The mixture was stirred overnight and then
concentrated under reduced pressure. The residue was dis-
solved in Et2O (100 mL) and washed with water, Na2S2O3
solution, and finally brine. The organic solution was then dried
over Na2SO4, filtered, and concentrated under reduced pres-

Serine Protease Inhibitors with a Unique Binding Mode
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17 2769
sure to give the crude 36 (1.16 g, 96%) which could be used
without further purification.
over Na2SO4, filtered, and concentrated under reduced pres-
1
sure to give 43 (3.26 g, ∼100% from 40) as a yellow solid. H
NMR (DMSO-d6) δ 8.04 (d, J ) 8.1 Hz, 1H), 7.94 (s, 1H), 7.60
(d, J ) 8.1 Hz, 1H), 3.96 (s, 3H).
1-(5-Br om o-2-h ydr oxybiph en yl-3-yl)eth an on e (37). 1-(2-
Hydroxybiphenyl-3-yl)ethanone 35 (0.85 g, 4 mmol) was dis-
solved in AcOH (5 mL) and then treated with NBS (0.71 g, 4
mmol). The mixture was heated at 85 °C for 2 h. The mixture
was concentrated under reduced pressure and the resultant
solid dissolved in Et2O (50 mL). The organic solution was
washed with water followed by brine. The organic layer was
dried over Na2SO4, filtered, and concentrated under reduced
3-Met h yla m in o-4-n it r ob en zon it r ile (44). The crude
3-methoxy-4-nitrobenzonitrile 43 (1.0 g, 5.6 mmol) was stirred
in DMSO (7 mL) and heated to 75 °C. A 40% solution of
methylamine in water (1 mL) was added and the mixture
heated at 75 °C for 2 h. The solution was allowed to cool, and
then EtOAc (200 mL) was added. The solution was washed
with 10% NaHCO3 (3 × 100 mL) followed by brine. The organic
solution was dried over Na2SO4, filtered, and concentrated
under reduced pressure to give 44 (0.91 g, 92%) as an orange
1
pressure to give 37 (1.16 g, ∼100%). H NMR (CDCl3) δ 7.84
(d, J ) 2.5 Hz, 1H), 7.64 (d, J ) 2.5 Hz, 1H), 7.55-7.52 (m,
2H), 7.46-7.36 (m, 3H), 2.66 (s, 3H).
1
solid. H NMR (DMSO-d6) δ 8.24 (br q, J ) 5.1 Hz, 1H), 8.17
2-Meth oxybip h en yl-3-ca r ba ld eh yd e (38). 2-Hydroxybi-
phenyl-3-carbaldehyde 10 (2.0 g, 10 mmol) in DMF (30 mL)
was treated with cesium carbonate (3.9 g, 12 mmol) and MeI
(0.69 mL, 11 mmol). The mixture was stirred overnight and
then diluted with EtOAc (200 mL). The solution was washed
with Na2S2O5 solution and then brine. The organic solution
was dried over Na2SO4, filtered, and concentrated under
reduced pressure. The residual solid was subjected to chro-
matography over silica gel eluting with 5% EtOAc in hexanes
(d, J ) 8.8 Hz, 1H), 7.51 (s, 1H), 7.00 (d, J ) 8.8 Hz, 1H), 2.96
(d, J ) 5.1 Hz, 3H).
4-Am in o-3-m et h yla m in ob en zon it r ile (45). 3-Methyl-
amino-4-nitrobenzonitrile 44 (0.85 g, 4.8 mmol) was dissolved
in MeOH (50 mL) and EtOAc (50 mL). 10% Pd/C (200 mg)
catalyst was added and the mixture stirred under hydrogen
at atmospheric pressure for 1.5 h. The mixture was filtered
through Celite and the eluant concentrated under reduced
pressure to give 45 (0.69 g, ∼98%) as a tan solid. 1H NMR
(DMSO-d6) δ 6.84 (dd, J ) 7.7, 1.4 Hz, 1H), 6.56 (m, 2H), 5.45
(br s, 2H), 4.98 (br s, 1H), 2.71 (s, 3H).
1
to give 38 (1.2 g, 56%). H NMR (CDCl3) δ 10.47 (s, 1H), 7.84
(dd, J ) 7.7, 1.8 Hz, 1H), 7.61-7.55 (m, 3H), 7.48-7.38 (m,
3H), 7.29 (dd, J ) 7.6, 1.0 Hz, 1H), 3.50 (s, 3H).
2-(2-H yd r oxyb ip h en yl-3-yl)-3-m et h yl-3H -b en zoim id -
a zole-5-ca r b on it r ile (46). The 4-amino-3-methylamino-
benzonitrile 45 (0.64 g, 4.35 mmol) and 2-hydroxybiphenyl-3-
carbaldehyde 10 (1.03 g, 5.22 mmol) were reacted according
to general procedure C. The crude product was subjected to
chromatography over silica gel eluting with 25% EtOAc in
hexanes to give 46 (1.22 g, 86%). H NMR (DMSO-d6) δ 8.39
(s, 1H), 7.86 (m, 2H), 7.69 (dd, J ) 8.0, 1.1 Hz, 1H), 7.60 (m,
2H), 7.56-7.35 (m, 4H), 7.14 (m, 1H), 4.03 (s, 3H).
1-(2-Met h oxyb ip h en yl-3-yl)et h a n on e.
(39).
Vana-
dium(III) chloride (0.89 g, 5.65 mmol) was added to CH2Cl2
(10 mL) and cooled to -78 °C. A 3 M solution of MeMgBr in
Et2O (1.9 mL, 5.65 mmol) was added in two portions over 10
min, and the resultant mixture was then stirred for 20 min.
This solution was then added to a solution of the aldehyde 38
(1.2 g, 5.65 mmol) in CH2Cl2 (10 mL) at -78 °C. After stirring
the mixture for 2 h at -78 °C, the solution was allowed to
warm to room temperature. Toluene (20 mL) was added and
the solution refluxed overnight. The solution was allowed to
cool and then diluted with EtOAc (200 mL) and washed with
saturated NaHCO3 (2 × 100 mL), water (2 × 100 mL), and
finally brine. The organic solution was dried over Na2SO4,
filtered, and concentrated under reduced pressure. The re-
sidual solid was subjected to chromatography over silica gel
eluting with 15% EtOAc in hexanes to give the intermediate
alcohol. The alcohol was then dissolved in dry CH2Cl2 and
treated with 4-methylmorpholine-N-oxide (0.79 g, 6.75 mmol)
followed by tetrapropylammonium perruthenate (TPAP). The
mixture was stirred for 60 min and then filtered through silica
gel eluting with CH2Cl2. The eluant was concentrated under
1
2-(2-H yd r oxyp h en yl)-3-m et h yl-3H -b en zoim id a zole-5-
ca r boxa m id in e (49). The 2-(2-hydroxybiphenyl-3-yl)-3-meth-
yl-3H-benzoimidazole-5-carbonitrile 46 (1.22 g, 3.8 mmol) was
dissolved in DMSO (15 mL) and heated to 70 °C. A 50%
aqueous solution of hydroxylamine (2 mL) was added and the
mixture heated at 70 °C for 1 h. The mixture was cooled,
diluted with water (200 mL), and the resultant precipitate that
formed was collected by filtration to give the crude hydroxy-
1
amidine product 47 (1.07 g, 80%). H NMR (DMSO-d6) δ 8.50
(s, 1H), 7.98 (d, J ) 8.5 Hz, 1H), 7.82 (d, J ) 8.5 Hz, 1H), 7.75
(d, J ) 8.1 Hz, 1H), 7.60-7.57 (m, 3H), 7.47-7.35 (m, 3H),
7.20 (t, J ) 7.7 Hz, 1H), 4.00 (s, 3H). A portion of 47 (0.3 g,
0.84 mmol) was dissolved in AcOH (30 mL) and then treated
with acetic anhydride (80 µL, 0.88 mmol). After the solution
was stirred for 50 min, 10% Pd/C (50 mg) catalyst was added.
The mixture was stirred under hydrogen at atmospheric
pressure for 2 h and then filtered through Celite. The eluant
was concentrated under reduced pressure and the crude
residue subjected to preparative HPLC to give 49 (0.12 g, 42%
1
reduced pressure to give 39 (0.9 g, 70%). H NMR (CDCl3) δ
7.60 (dd, J ) 7.7, 1.7 Hz, 1H), 7.57-7.53 (m, 2H), 7.48-7.36
(m, 4H), 7.21 (t, J ) 7.7 Hz, 1H), 3.40 (s, 3H), 2.67 (s, 3H).
3-Meth oxy-4-n itr oben zon itr ile (43). The 3-methoxy-4-
nitrobenzoic acid 40 (5.2 g, 26.4 mmol) was dissolved in THF
(70 mL) and cooled to 0 °C with an ice bath. Oxalyl chloride
(2.53 mL, 29.0 mmol) was added followed by several drops of
DMF. The ice bath was removed and the solution stirred for 1
h. The mixture was concentrated under reduced pressure to
give 41 as a yellow solid. Crude solid 41 was dissolved in THF
(70 mL) and cooled to 0 °C with an ice bath. Ammonia gas
was bubbled through the solution for 10 min, which resulted
in the formation of a pale yellow slurry. The ice bath was
removed and the addition of ammonia gas continued for a
further 5 min. The mixture was then sealed, stirred overnight,
and then diluted with EtOAc (600 mL). The organic mixture
was washed with 1 N HCl, brine, dried over MgSO4, filtered,
and concentrated under reduced pressure to give 42 (3.65 g,
1
from 47) as a tan solid. H NMR (DMSO-d6) δ 9.70 (br s, 2H),
9.41 (br s, 2H), 8.69 (s, 1H), 8.01 (m, 2H), 7.76 (d, J ) 7.7 Hz,
1H), 7.60 (m, 3H), 7.49-7.38 (m, 3H), 7.23 (m, 1H), 4.04 (s,
3H). 13C NMR (DMSO-d6) δ 165.1, 153.4, 152.4, 137.3, 136.6,
135.0, 133.4, 131.5, 130.6, 129.4, 128.3, 127.4, 124.9, 123.9,
120.6, 115.7, 114.2, 112.7, 33.0. MS-Sciex m/z ) 342.8 [MH+].
Anal. (C21H18N4O‚2.0HCl) C, H, N, Cl.
4-Hyd r oxy-3-iod oben zon itr ile (51). The 4-hydroxybenzo-
nitrile 50 (2.4 g, 20 mmol) was dissolved in AcOH (40 mL)
and treated with N-iodosuccinimide (NIS) (4.5 g, 20 mmol).
The mixture was stirred overnight and the precipitate removed
by filtration. The filtrate was concentrated under reduced
pressure and the resultant oil dissolved in Et2O. The insoluble
material was removed by filtration and the filtrate washed
with saturated NaHCO3. The organic phase was dried over
MgSO4, filtered, and concentrated under reduced pressure. The
crude residue was subjected to chromatography over silica gel
eluting with 10% MeOH in CH2Cl2 to give 51 (2.0 g, 40%) (90%
pure by analytical HPLC). 1H NMR (CDCl3) δ 7.96 (d, J ) 2.0
Hz, 1H), 7.54 (dd, J ) 8.7, 2.0 Hz, 1H), 7.04 (d, J ) 8.7 Hz,
1H), 5.92 (s, 1H).
1
71%) as a yellow solid. H NMR (DMSO-d6) δ 8.23 (br s, 1H),
7.94 (d, J ) 8.4 Hz, 1H), 7.72 (d, J ) 1.5 Hz, 1H), 7.71 (br s,
1H), 7.56 (dd, J ) 8.4, 1.5 Hz, 1H), 3.97 (s, 3H). The crude
solid 42 (3.6 g, 18.4 mmol) was stirred in THF (50 mL), and
Et3N (3.3 mL, 23.9 mmol) was added. The yellow slurry was
then treated with TFA (2.85 mL, 20.2 mmol) by slow addition.
After 1 h the solution was concentrated under reduced pres-
sure and diluted with EtOAc (300 mL). The solution was
washed with 1 N HCl (200 mL), saturated NaHCO3 solution
(2 × 200 mL), and then brine. The organic solution was dried

2770 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17
Verner et al.
2-(2-Me t h oxye t h oxym e t h oxy)b ip h e n yl-3-ca r b a ld e -
h yd e (52). The 2-hydroxy-3-phenylbenzaldehyde 10 (1.7 g, 8.6
mmol) was dissolved in DMF (20 mL) and then treated with
DIEA (1.8 mL, 10.1 mmol) followed by 2-methoxyethoxymethyl
chloride (1.2 mL, 10.1 mmol). After 2 h the mixture was diluted
with EtOAc (350 mL) and water (200 mL). The solution was
acidified to pH ) 5 by addition of 1 N HCl. The organic phase
was separated, washed with water, saturated NaHCO3, and
finally brine. The organic phase was dried over Na2SO4,
filtered, and concentrated under reduced pressure to give 52
(2.0 g, 81%). This material was suitable for further use without
purification (90% pure by analytical HPLC).
3-Eth yn yl-2-(2-m eth oxyeth oxym eth oxy)bip h en yl (53).
The crude 2-(2-methoxyethoxymethoxy)biphenyl-3-carbalde-
hyde 52 (2 g, 6.7 mmol) was dissolved in dry MeOH (25 mL)
and treated with K2CO3 (2.9 g, 21 mmol). The (1-diazo-2-
oxopropyl)phosphoric acid dimethyl ester25 (2.7 g, 14.1 mmol)
in MeOH (8 mL) was added dropwise over 5 min. The mixture
was stirred overnight and then diluted with EtOAc. The
mixture was washed with saturated NaHCO3 solution. The
organic phase was then washed with brine, dried over Na2SO4,
filtered, and concentrated under reduced pressure. The residue
was subjected to chromatography over silica gel eluting with
15% EtOAc in hexanes to give 53 (1.6 g, 66% from 10). 1H NMR
(CDCl3) δ 7.52-7.31 (m, 7H), 7.13 (t, J ) 7.4 Hz, 1H), 4.99 (s,
2H), 3.37-3.34 (m, 2H), 3.26 (s, 1H), 3.24 (s, 3H), 3.20-3.17
(m, 2H).
2-[2-(2-Met h oxyet h oxym et h oxy)b ip h en yl-3-yl]b en zo-
fu r a n -5-ca r b on it r ile (54). The 3-ethynyl-2-(2-methoxy-
ethoxymethoxy)biphenyl 53 (1.6 g, 5.7 mmol) was mixed with
4-hydroxy-3-iodobenzonitrile 51 (1.4 g, 5.7 mmol) in dry DMF
(50 mL). The mixture was then treated with dichlorobis-
(triphenylphosphine)palladium(II) (0.4 g, 0.57 mmol) followed
by Et3N (4 mL, 28.7 mmol). The mixture was stirred for 30
min, and then CuI (0.22 g, 1.1 mmol) was added. The resultant
mixture was stirred for 3 h and then diluted with EtOAc (150
mL). The organic solution was washed with brine, dried over
Na2SO4, filtered, and concentrated under reduced pressure.
The residue was subjected to chromatography over silica gel
eluting with 20% EtOAc in hexanes to give 54 (0.53 g, 23%).
1H NMR (CDCl3) δ 7.97 (dd, J ) 8.3, 2.2 Hz, 1H), 7.94 (dd, J
) 1.7, 0.8 Hz, 1H), 7.59-7.57 (m, 2H), 7.56 (dd, J ) 3.3, 1.9
Hz, 1H), 7.52 (d, J ) 0.8 Hz, 1H), 7.47-7.36 (m, 5H), 7.31 (t,
J ) 8.5 Hz, 1H), 4.72 (s, 2H), 3.29-3.25 (m, 2H), 3.21 (s, 3H),
3.20-3.15 (m, 2H).
thrombin (tosyl-Gly-Pro-Lys-p-nitroanilide), and for factor Xa
(CH3OCO-D-CHA-Gly-Arg-pNA-AcOH, “Pefachrome Xa”) were
from Centerchem, Inc. The substrate for HMW u-PA (Glt-Gly-
Arg-AMC) was from Peptide Institute, and the substrate for
t-PA (CH3-SO2-D-HHT-Gly-Arg-pNA, “Spectrozyme tPA”) was
from American Diagnostica.
En zym e Assa ys. u-PA was assayed spectrofluorometrically,
and all other enzymes were assayed spectrophotometrically.
All enzymes were assayed at ambient temperature in 50 mM
Tris, 150 mM NaCl, 0.05% (v/v) Tween-20, 10% (v/v) DMSO,
0.002% antifoam, pH ) 7.4 with EDTA (1 mM). The thrombin
and factor Xa assays also contained 5.0 mM CaCl2. The final
concentrations of u-PA, t-PA, trypsin, thrombin, plasmin and
factor Xa in the respective assays were 4 (or 5), 11 (or 15), 15,
10, 6 and 4 nM. The substrate concentrations of 85 µM (u-
PA), 300 µM (t-PA), 25 or 45 µM (trypsin), 300 µM (thrombin),
130 µM (plasmin), and 550 µM (factor Xa) were chosen based
on the determined Km values of 85, 500, 45, 300, 100, and 600
µM, respectively. For u-PA activity, the rate of change in
relative fluorescence units (RFU) emitted at 460 nm (excitation
at 355 nm) was measured immediately after addition of
enzyme. For all other enzymatic activities, the rate of change
in absorbance at 405 nm was measured immediately after
addition of enzyme. Apparent inhibition constants, Ki′ values,
were calculated from the velocity data generated at the various
inhibitor concentrations using the software package BatchKi
(developed and provided by Dr. Petr Kuzmic, Biokin Ltd.,
Madison, WI) using methodology similar to that described for
tight-binding inhibitors.27 BatchKi applies nonlinear regression
analysis to return nominal likelihood intervals at the 65%
confidence, using the logic described by Bates and Watts.29 The
Ki′ values are accepted where the nominal likelihood intervals
at 65% confidence vary within 0.5 to 1.5-fold from the Ki′
values. These values were converted to Ki values by the
formula, Ki ) Ki′/(1 + S/Km).
Cr ysta lliza tion a n d P r ep a r a tion of P r otea se-In h ibi-
tor Com p lexes. Procedures and conditions for crystallization
have been reported by Katz et al.14
Ack n ow led gm en t. The authors are grateful to Dr.
T. J enkins, R. Day, Dr. D. McGee, and Dr. V. Tai for
the synthesis of screening leads and late stage com-
pounds. The authors also thank Dr. Mike Venuti and
Dr. J ochen Knolle for helpful discussions and insights.
2-(2-H yd r oxyb ip h en yl-3-yl)b en zofu r a n -5-ca r b oxa m i-
d in e (58). The 2-[2-(2-methoxyethoxymethoxy)biphenyl-3-yl]-
benzofuran-5-carbonitrile 54 (0.53 g, 1.3 mmol) was dissolved
in absolute EtOH (75 mL), and excess 50% aqueous hydroxyl-
amine (2 mL) was added. The solution was refluxed for 2.5 h
and then concentrated under reduced pressure to give 55. The
residue 55 was dissolved in THF (75 mL) and treated with
acetic anhydride (0.25 mL, 2.7 mmol). The mixture was stirred
for 30 min to form 56, and then MeOH (25 mL) was added
followed by 10% Pd/C (50 mg) catalyst. The mixture was
stirred under hydrogen for 12 h and then filtered through
Celite. The filtrate was concentrated under reduced pressure
to give 57. The crude residue 57 was dissolved in MeOH (75
mL) and treated with 6 N HCl (5 mL). The resultant mixture
was stirred for 2 h and then concentrated under reduced
pressure. The crude residue was subjected to preparative
HPLC chromatography to give 58 (186 mg, 43%) as an
amorphous white solid. 1H NMR (DMSO-d6) δ 9.54 (br s, 2H),
9.36 (br m, 3H), 8.29 (s, 1H), 7.92-7.80 (m, 3H), 7.59-7.36
(m, 6H), 7.29 (dd, J ) 7.4, 1.5 Hz, 1H), 7.13 (t, J ) 8.6 Hz,
1H). 13C NMR (DMSO-d6) δ 166.6, 156.5, 155.4, 151.9, 138.4,
132.3, 132.2, 130.1, 129.9, 129.1, 127.9, 126.7, 125.0, 123.5,
122.6, 121.5, 119.7, 112.0, 106.4. MS-Sciex m/z 328.6 [MH+].
Anal. (C21H16N2O2‚HCl) C, H, N, O.
Su p p or tin g In for m a tion Ava ila ble: Affinity measure-
ments of selected compounds toward the panel of enzymes in
the presence of zinc ions. This material is available free of
charge via the Internet at http://pubs.acs.org.
Refer en ces
(1) 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
Inhibitors. Nature 1998, 391, 608-612.
(2) (a) Powers, J . C.; Harper, J . W. Inhibitors of serine proteinases.
In Proteinase Inhibitors; Barrett, A. J ., Salvesen, G., Eds.;
Elsevier: Amsterdam, 1986. (b) Andreasen, P. A.; Kjøller, L.;
Christensen, L.; Duffy, M. J . The Urokinase-type Plasminogen
Activator System in Cancer Metastasis: A Review. Int. J . Cancer
1997, 72, 1-22. (c) J ackson, C. M.; Nemerson, Y. Blood Coagula-
tion. Annu. Rev. Biochem. 1980, 49, 765-811. (d) Leung, D.;
Abbenante, G.; Fairlie, D. P. Protease Inhibitors: Current Status
and Future Prospects. J . Med. Chem. 2000, 43, 305-341. (e)
Henkin, J . Proteases and Metastasis. Annu. Rep. Med. Chem.
1993, 28, 151-160.
(3) (a) Sanderson, P. E. J .; Naylor-Olsen, A. M. Thrombin Inhibitor
Design. Curr. Med. Chem. 1998, 5, 289-304. (b) Vacca, J . P.
Thrombosis and Coagulation. Annu. Rep. Med. Chem. 1998, 33,
81-90.
(4) Ripka, W. C.; Vlasuk, G. P. Antithrombotics/Serine Proteases.
Annu. Rep. Med. Chem. 1997, 32, 71-89.
En zym ology. Enzymes were purchased as indicated: high
molecular weight (HMW) u-PA (American Diagnostica), t-PA
(Sigma), trypsin (Worthington), thrombin (Calbiochem), plas-
min (Enzyme Research Labs), and factor Xa (Haematological
Technologies, Inc.). The substrates for trypsin, plasmin, and
(5) Fevig, J . M.; Wexler, R. R. Anticoagulants: Thrombin and Factor
Xa Inhibitors. Annu. Rep. Med. Chem. 1999, 34, 81-100.
(6) Magill, C.; Katz, B. A.; Mackman, R. L. Emerging Therapeutic
Targets in Oncology: Urokinase-type Plasminogen Activator
System. Emerging Ther. Targets 1999, 3, 109-133.

Serine Protease Inhibitors with a Unique Binding Mode
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 17 2771
(7) Rice, K. D.; Tanaka, R. D.; Katz, B. A.; Numerof, R. P.; Moore,
W. R. Inhibitors of Tryptase for the Treatment of Mast Cell-
Mediated Diseases. Curr. Pharm. Des. 1998, 4, 381-396.
(8) Mehdi, S. Synthetic and Naturally Occurring Protease Inhibitors
Containing an Electrophilic Carbonyl Group. Bioorg. Chem.
1993, 21, 249-259.
(9) Wakselman, M.; Xie, J .; Mazaleyrat, J .-P. New Mechanism-
Based Inactivators of Trypsin-like Proteinases. Selective In-
activation of Urokinase by Functionalized Cyclopeptides Incor-
porating a Sulfoniomethyl-Substituted m-Aminobenzoic Acid
Residue. J . Med. Chem. 1993, 36, 1539-1547.
(10) Tamura, S. Y.; Weinhouse, M. I.; Roberts, C. A.; Goldman, E.
A.; Masukawa, K.; Anderson, S. M.; Cohen, C. R.; Bradbury, A.
E.; Bernardino, V. T.; Dixon, S. A.; Ma, M. G.; Nolan, T. G.;
Brunck, T. K. Synthesis and Biological Activity of Peptidyl
Aldehyde Urokinase Inhibitors. Bioorg. Med. Chem. Lett. 2000,
10, 983-987.
(11) Deadman, J . J .; Elgendy, S.; Goodwin, C. A.; Green, D.; Baban,
J . A.; Patel, G.; Skordalakes, E.; Chino, N.; Claeson, G.; Kakkar,
V. V.; Scully, M. F. Characterization of a Class of Peptide
Boronates with Neutral P1 Side Chains as Highly Selective
Inhibitors of Thrombin. J . Med. Chem. 1995, 38, 1511-1522.
(12) (a) J anc, J . W.; Clark, J . M.; Warne, R. L.; Elrod, K. C.; Katz, B.
A.; Moore, W. R. A Novel Approach to Serine Protease Inhibi-
tion: Kinetic Characterization of Inhibitors Whose Potencies and
Selectivities Are Dramatically Enhanced by Zinc(II). Biochem-
istry, 2000, 39, 4792-4800. (b) Axys Pharmaceuticals Inc. Bi-
Amidino Group Substituted Heterocycle Derivatives and their
use as Anticoagulants. Intl. Pat. Appl. 9926932, J une 3, 1999.
(c) Axys Pharmaceuticals Inc. Substituted Amidinoaryl Deriva-
tives and their use as Anticoagulants. Intl. Pat. Appl. 9926933,
J une 3, 1999. (d) Axys Pharmaceuticals Inc. Substituted Ami-
dinoaryl Derivatives and their use as Anticoagulants. Intl. Pat.
Appl. 9926941, J une 3, 1999.
(13) Presnell, S. R.; Patil, G. S.; Mura, C.; J ude, K. M.; Conley, J .
M.; Bertrand, J . A.; Kam, C.-M.; Powers, J . C.; Williams, L. D.
Oxyanion-Mediated Inhibition of Serine Proteases. Biochemistry
1998, 37, 17068-17081.
(14) Katz, B. A.; Elrod, K.; Luong, C.; Rice, M.; Mackman. R. L.;
Sprengeler, P. A.; Spencer, J .; Hataye, J .; J anc, J .; Link, J .;
Litvak, J .; Rai, R.; Rice, K.; Sideris, S.; Verner, E.; Young, W. A
Infections. J . Med. Chem. 1996, 39, 1452-1462. (b) Singh, M.
P.; Sasmal, S.; Lu, W.; Chatterjee, M. N. Synthetic Utility of
Catalytic Fe(III)/Fe(II) Redox Cycling Towards Fused Hetero-
cycles: A Facile Access to Substituted Benzimidazole, Bis-
benzimidazole and Imidazopyridine Derivatives. Synthesis 2000,
10, 1380-1390.
(20) Hofsløkken, N. U.; Skattebøl, L. Convenient Method for the
ortho-Formylation of Phenols. Acta Chem. Scand. 1999, 53, 258-
262.
(21) Iwanowicz, E. J .; Lau, W. F.; Lin, J .; Roberts, D. G. M.; Seiler,
S. M. Derivatives of 5-Amidine Indole as Inhibitors of Thrombin
Catalytic Activity. Bioorg. Med. Chem. Lett. 1996, 6, 1339-1344.
(22) Robinson, B. The Fischer Indole Synthesis; J ohn Wiley and
Sons: New York, 1982.
(23) Castro, J . L.; Baker, R.; Guiblin, A. R.; Hobbs, S. C.; J enkins,
M. R.; Russell, M. G. N.; Beer, M. S.; Stanton, J . A.; Scholey,
K.; Hargreaves, R. J .; Graham, M. I.; Matassa, V. G. Synthesis
and Biological Activity of 3-[2-(Dimethylamino)ethyl]-5-[(1,1-
dioxo-5-methyl-1,2,5-thiadiazolidin-2-yl)-methyl]-1H-indole and
Analogues: Agonists for the 5-HT1D Receptor. J . Med. Chem.
1994, 37, 3023-3032.
(24) Miyaura, N.; Yanagi, T.; Suzuki, A. The Palladium-catalyzed
Cross-coupling Reaction of Phenylboronic Acid with Haloarenes
in the Presence of Bases. Synth. Commun. 1981, 11, 513-519.
(25) Mu¨ller, S.; Liepold, B.; Roth, G. J .; Bestmann, H. J . An Improved
One-pot Procedure for the Synthesis of Alkynes from Aldehydes.
Synlett 1996, 6, 521-522.
(26) Candiani, I.; DeBernardinis, S.; Cabri, W.; Marchi, M.; Bedeschi,
A.; Penco, S. A Facile One-pot Synthesis of Polyfunctionalized
2-Unsubstituted Benzo[b]furans. Synlett 1993, 4, 269-270.
(27) Kuzmic, P.; Sideris, S.; Cregar, L. M.; Elrod, K. C.; Rice, K. D.;
J anc, J . W. High-Throughput Screening of Enzyme Inhibitors:
Automatic Determination of Tight-Binding Inhibition Constants.
Anal. Biochem. 2000, 281, 62-67.
(28) (a) Katz, B. A.; Mackman, R.; Luong, C.; Radika, K.; Martelli,
A.; Sprengeler, P. A.; Wang, J .; Chan, H.; Wong, L. Structural
Basis for Selectivity of a Small Molecule S1-binding Submicro-
molar Inhibitor of Urokinase-type Plasminogen Activator. Chem.
Biol. 2000, 7, 299-312. (b) Nienaber, V.; Wang, J .; Davidson,
D.; Henkin, J . Re-engineering of Human Urokinase Provides a
System for Structure-based Drug Design at High Resolution and
Reveals a Novel Structural Subsite. J . Biol. Chem. 2000, 275,
7239-7248. (c) Nienaber, V. L.; Davidson, D.; Edalji, R.; Gi-
randa, V. L.; Klinghofer, V.; Henkin, J .; Magdalinos, P.; Mantei,
R.; Merrick, S.; Severin, J . M.; Smith, R. A.; Stewart, K.; Walter,
K.; Wang, J .; Wendt, M.; Weitzberg, M.; Zhao, X.; Rockway, T.
Structure-directed Discovery of Potent Non-peptidic Inhibitors
of Human Urokinase that Access a Novel Binding Subsite.
Structure 2000, 8, 553-563. (d) Sperl, S.; J acob, U.; Arroyo de
Prada, N.; Stu¨rzebecher, J .; Wilhelm, O. G.; Bode, W.; Magdolen,
V.; Huber, R.; Moroder, L. (4-Aminomethyl)phenylguanidine
Derivatives as Nonpeptidic Highly Selective Inhibitors of Human
Urokinase. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 5113-5118.
(e) Zeslawska, E.; Schweinitz, A.; Karcher, A.; Sondermann, P.;
Sperl, S.; Stu¨rzebecher, J .; J acob, U. Crystals of the Urokinase
Type Plasminogen Activator Variant âc-uPA in Complex with
Small Molecule Inhibitors Opens the Way towards Structure-
based Drug Design. J . Mol. Biol. 2000, 301, 465-475.
(29) Bates, D. M.; Watts, D. W. Nonlinear Regression and its
Applications. J ohn Wiley and Sons: New York, 1988.
Novel Serine Protease Inhibition Motif Involving
a Multi-
centered Short Hydrogen Bonding Network at the Active Site.
J . Mol. Biol. 2001, 307, 1451-1486.
(15) Rai, R.; Kolesnikov, A.; Li, Y.; Young, W.; Leahy, E.; Sprengeler,
P.; Verner, E.; Shrader, W.; Burgess-Henry, J .; Sangalang, J .;
Allen, D.; Chen, X.; Katz, B.; Luong, C.; Elrod, K.; Cregar, L.
Development of Potent and Selective Factor Xa Inhibitors.
Bioorg. Med. Chem. Lett. In press.
(16) Tidwell, R. R.; Geratz, J . D.; Dann, O.; Volz, G.; Zeh, D.; Loewe,
H. Diarylamidine Derivatives with One or Both of the Aryl
Moieties Consisting of an Indole or Indole-like Ring. Inhibitors
of Arginine-Specific Esteroproteases. J . Med. Chem. 1978, 21,
613-623.
(17) Fairley, T. A.; Tidwell, R. R.; Donkor, I.; Naiman, N. A.;
Ohemeng, K. A.; Lombardy, R. J .; Bentley, J . A.; Cory, M.
Structure, DNA Minor Groove Binding, and Base Pair Specificity
of Alkyl- and Aryl-Linked Bis(amidinobenzimidazoles) and Bis-
(amidinoindoles). J . Med Chem. 1993, 36, 1746-1753.
(18) Geratz, J . D.; Stevens, F. M.; Polakoski, K. L.; Parrish, R. F.;
Tidwell, R. R. Amidino-Substituted Aromatic Heterocycles as
Probes of the Specificity Pocket of Trypsin-Like Proteases. Arch.
Biochem. Biophys. 1979, 197, 551-559.
(19) (a) Lombardy, R. L.; Tanious, F. A.; Ramachandran, K.; Tidwell,
R. R.; Wilson, W. D. Synthesis and DNA Interactions of Benz-
imidazole Dications Which Have Activity against Opportunistic
J M0100638