Combinatorial library of serine and cysteine protease inhibitors that interact with both the S and S' binding sites

Article abstract of DOI:10.1021/jm990272g

A combinatorial library of 400 inhibitors has been synthesized and screened against several serine and cysteine proteases including plasmin, cathepsin B, and papain. The inhibitors are based upon a cyclohexanone nucleus and are designed to probe binding i

Full text of DOI:10.1021/jm990272g

J . Med. Chem. 1999, 42, 4001-4009
4001
Com bin a tor ia l Libr a r y of Ser in e a n d Cystein e P r otea se In h ibitor s Th a t In ter a ct
w ith Both th e S a n d S′ Bin d in g Sites
Paul Abato, J effrey L. Conroy, and Christopher T. Seto*
Department of Chemistry, Brown University, 324 Brook Street, Box H, Providence, Rhode Island 02912
Received J une 1, 1999
A combinatorial library of 400 inhibitors has been synthesized and screened against several
serine and cysteine proteases including plasmin, cathepsin B, and papain. The inhibitors are
based upon a cyclohexanone nucleus and are designed to probe binding interactions in the S2
and S2′ binding sites. This methodology has led to the discovery of inhibitor 15A, which
incorporates Trp at both the P2 and P2′ positions and has an inhibition constant against plasmin
of 5 µM. Data from screening of the library shows that plasmin has a strong specificity for Trp
at the S2 subsite and prefers to bind hydrophobic and aromatic amino acids such as Ile, Phe,
Trp, and Tyr at the S2′ subsite. In contrast, the S2′ subsites of cathepsin B and papain do not
show a strong preference for any particular amino acid.
In tr od u ction
Combinatorial chemistry has emerged as a powerful
method for generating lead compounds for drug discov-
ery and for optimizing the biological activity of those
leads.1 This technique has been used to develop new
F igu r e 1. General structure of the compounds that are
ligands for a variety of biological targets including
proteases, kinases, various receptors, and antibodies,
among others. Proteases are particularly interesting
targets because they are involved with a wide variety
of important diseases that include AIDS, cancer, and
malaria. Many of the libraries that have been generated
for screening against proteases incorporate a chemical
functionality that mimics the tetrahedral intermediate
that occurs during enzyme-catalyzed peptide hydrolysis.
For example, phosphonic acids have been screened
against the metalloprotease thermolysin,2 and statine,3
(hydroxyethyl)amine,4 and diamino diol5 isosteres have
been used to synthesize libraries against the aspartic
protease HIV protease. In addition, a peptide aldehyde
library has been targeted against the cysteine protease
interleukin-1â converting enzyme.6
We have recently designed a new class of inhibitors
for serine and cysteine proteases that are based upon a
4-heterocyclohexanone pharmacophore.7,8 These inhibi-
tors react with the enzyme active site nucleophile to
generate a reversibly formed hemiketal or hemithioketal
adduct that also mimics the tetrahedral intermediate.9
One attractive feature of the 4-heterocyclohexanone-
based inhibitors is that they can be derivatized in two
directions, allowing them to make contacts with both
the S and S′ subsites.10-12 Thus, the 4-heterocyclohex-
anone pharmacophore, with its bidirectional nature, can
easily be incorporated into a combinatorial synthesis of
inhibitors. In this paper we describe the synthesis and
screening of a 400-membered library of inhibitors that
are based upon a cyclohexanone nucleus (Figure 1). The
Xaa position in compound 1 is designed to fit into the
S2 specificity pocket, the Yaa position will fit into the
S2′ site, and the carbonyl moiety of the cyclohexanone
ring is designed to react with the active site nucleophile
to give a reversibly formed covalent adduct.
present in the 400-member combinatorial library of inhibitors.
Xaa and Yaa are each one of 20 different amino acids.
This work has three objectives. First, we demonstrate
that the cyclohexanone nucleus can be a useful platform
for developing protease inhibitors that interact with
both the S and S′ binding sites using combinatorial
chemistry. Second, the library is screened against
several medicinally relevant serine and cysteine pro-
teases in order to discover potential leads. Third, we
explore the S2′ specificity of these proteases. For many
of these enzymes the specificity of this site has not been
well defined.
Resu lts a n d Discu ssion
Design a n d Syn th esis of th e Libr a r y. Before we
began constructing the library, we needed to devise an
efficient synthesis of a building block such as compound
4 (Scheme 1). This molecule incorporates the cyclohex-
anone nucleus, is amenable to solid-phase peptide
synthesis, and carries the ketone functionality in a
suitably protected form. We have reported previously
that compound 2 can be converted to 4 by oxidative
cleavage of the double bond and replacement of the Boc
protecting group with Fmoc.10 However, we have found
that on larger scale, reaction of the amino acid with
Fmoc chloride or Fmoc N-hydroxysuccinimide ester
under a variety of conditions gave relatively low and
inconsistent yields of 4. This problem can be circum-
vented by switching the protecting groups first and then
oxidizing the alkene to the acid as shown in Scheme 1.
Compound 4 is a mixture of two diastereomers, each of
which has the substituents on the cyclohexanone ring
in the 2,6-cis configuration.10
We have chosen the “split synthesis” strategy, first
described by Furka,13 for constructing the library and
10.1021/jm990272g CCC: $18.00 © 1999 American Chemical Society
Published on Web 09/03/1999

4002 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 19
Abato et al.
Sch em e 1a
Sch em e 2a
a
Reagents: (a) TFA, CH2Cl2; (b) FmocCl, DIEA; (c) NaIO4,
KMnO4, NaHCO3. One of two enantiomers is shown.
a
Reagents: (a) piperidine, DMF; (b) 4, HBTU, DIEA; (c) Cbz-
Xaa-OH, HBTU, DIEA; (d) TFA; (e) TFA, H2O. One of two
diastereomers is shown.
core was attached to all of the pools, followed by a second
amino acid, each pool receiving a different amino acid.
Finally, the inhibitors were cleaved from the beads, and
the protecting groups were removed. This resulted in
20 pools of inhibitors, each containing 20 different
compounds. The inhibitors in an individual pool all have
the same amino acid on their N-terminus, but a mixture
of the 20 different amino acids on the C-terminus. The
pools were assayed for inhibitory activity, and one or
several of the pools which showed the highest activity
were for chosen for deconvolution.
The second round of synthesis again began with 20
pools of resin, each with a different amino acid attached.
These were coupled to the inhibitor core; then every
compound was coupled to the N-terminal amino acid
that corresponded to the pool from the first round of
synthesis that was chosen for deconvolution. After
cleavage from the solid support and removal of the
protecting groups, these pools were assayed to deter-
mine which amino acid was optimal for the C-terminal
side of the inhibitor. Although the iterative deconvolu-
tion strategy does not always result in identification of
the best inhibitor in the library,1e it is a straightforward
and reliable method for determining the structure of
molecules that have high activity compared to the other
members of the library.
The details of the solid-phase synthesis are shown in
Scheme 2. Resin was purchased preloaded with the
Fmoc-protected amino acids. The N-terminus was depro-
tected with piperidine and then coupled to inhibitor core
4 using HBTU to give 6. Piperidine deprotection fol-
lowed by coupling to an N-Cbz amino acid gave com-
pound 7. The side chains of the amino acids were then
deprotected using TFA, which also released the inhibi-
tors from the solid support. Finally, the ketal protecting
group was removed by adding H2O to the cleavage
cocktail from the previous reaction and allowing the
solution to stand at room temperature overnight. Using
this same chemistry, we have reported previously the
solid-phase synthesis of an N-acetylated inhibitor that
contained Orn and Pro at the Xaa and Yaa positions,
respectively.10 This inhibitor was fully characterized,
and the overall yield of the isolated and HPLC-purified
compound was 50%.
F igu r e 2. Schematic diagram of the deconvolution strategy
used in the synthesis and screening of the combinatorial
library of protease inhibitors. A-D correspond to four different
amino acids. In the actual library 20 amino acids were used,
but the figure has been limited to 4 amino acids for the sake
of clarity. Inh represents the cyclohexanone pharmacophore,
and TFA is F3CCO2H.
the iterative deconvolution strategy, developed by
Houghten, for assaying its biological activity.14 These
techniques, as applied to the cyclohexanone-based in-
hibitor library, are outlined in Figure 2. Synthesis of
the library began with 20 batches of peptide synthesis
resin, each with a different amino acid attached. The
resin was mixed and then split into 20 pools that
contained a mixture of all 20 amino acids. The inhibitor
Eighteen of the 20 common amino acids were incor-
porated into the library. Cysteine and methionine were

Library of Serine and Cysteine Protease Inhibitors
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 19 4003
omitted to avoid problems associated with sulfur oxida-
tion, and they were replaced with hydroxyproline (Hyp)
and ornithine (Orn).
Assa y of th e Libr a r y. The initial library of inhibi-
tors was assayed against five proteases: cathepsin B,
plasmin, urokinase, kallikrein, and papain. The first
four of these enzymes have all been implicated in the
progression of cancer.15 These enzymes promote the
processes of angiogenesis and metastasis, either directly
by degrading components of the basement membrane
which surround blood vessels or indirectly by activating
other proteases which in turn attack the basement
membrane.16 Compounds which inhibit these proteases
may have potential as anticancer chemotherapeutic
agents.17 Papain was chosen as a control protease since
it is well-established that this enzyme prefers aromatic
amino acids such as phenylalanine at the P2 position.18
Thus we expected that in the screening of the initial
library against papain, the pools which incorporated
aromatic amino acids at the Xaa position would show
good activity against this protease.
Figure 3 shows the results of the solution-phase
assays of the library against plasmin, cathepsin B, and
papain. The assays were performed using p-nitroanilide
substrates and were monitored by UV spectroscopy. A
single concentration of the inhibitors was used, and the
runs were performed in duplicate or triplicate. The
concentration of each individual inhibitor in these
assays was 50 µM, giving a total inhibitor concentration
for all 20 inhibitors in each pool of 1 mM. The inhibitors
in several of the pools were not soluble at this concen-
tration, so the assays were performed in more dilute
solution as noted in the captions to Figures 3, 5, and 6.
For papain the three best pools of inhibitors incorpo-
rate Phe, Trp, and Tyr at the Xaa position as expected,
based upon the known specificity of the S2 subsite of
this protease.18 We infer from these results that for
papain, the inhibitors are binding in the active site in
the anticipated manner with Xaa in the S2 subsite.
F igu r e 3. Assay of 20 pools of 20 compounds each against
plasmin, cathepsin B, and papain. Each pool contains com-
pounds in which the Xaa position is defined by the amino acids
on the x-axes of the graphs, and Yaa is a mixture of all 20 amino
acids. The error in these measurements is approximately (5%.
For cathepsin B the pools in which Xaa ) Phe and Tyr were
assayed at 1/2 concentration and for Xaa ) Trp at 1/4
concentration, compared to the other pools in the library due
to low solubility of the inhibitors in the assay solution.
Cathepsin B displays relatively small differences in
activity between the various pools, with most showing
between 10% and 30% inhibition. Ile and Leu are the
only pools that show significantly better activity. These
data are consistent with information from previous
inhibition studies, which suggest that cathepsin B
prefers hydrophobic amino acids such as Leu, Ile, and
Phe at the P2 position.16
hydrogen-bonding interaction that occurs with the N-H
functionality of Trp.
Unlike papain and cathepsin B, plasmin shows very
high selectivity for one particular pool within the
library. This pool contains Trp at the Xaa position. Thus
plasmin prefers the extended aromatic side chain of Trp
at the P2 position to the exclusion of all other amino
acids, including those with smaller aromatic groups
such as Phe and Tyr and those with simple hydrophobic
side chains such as Ile and Leu. The X-ray crystal
structure of the active site portion of plasmin has not
been reported. However it will be interesting to see, once
the structure has been solved, if this selectivity is caused
by specific aromatic stacking interactions between the
inhibitor and aromatic side chains in the S2 subsite.
Alternately, S2 could comprise a simple hydrophobic
pocket that has good shape complementarity to ex-
tended aromatic rings, or there may be a specific
We have also assayed the inhibitor library against
urokinase and kallikrein. With these two proteases only
low levels of inhibition were observed, and the variations
between the pools were similar to the approximate (5%
error in the assays (data not shown). There are several
possible explanations for this lack of inhibition. First,
the structure or conformation of the inhibitors in the
library may not be sterically compatible with the active
sites of kallikrein and urokinase. Second, the inhibitors
may bind to the enzyme through weak noncovalent
interactions but not react with the active site nucleo-
phile to give the reversibly formed covalent adduct. In
the 4-heterocyclohexanone series of inhibitors, such
compounds have inhibition constants in the millimolar
range against papain. Thus inhibitors of this type are

4004 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 19
Abato et al.
F igu r e 4. Assay of the Trp deconvolution against plasmin.
Each bar represents the assay of one compound which has Trp
at the Xaa position and an amino acid at the Yaa position that
is indicated by the x-axis of the graph.
not likely to give significant activity under our assay
conditions.7,9
On the basis of the results of the initial assays, we
chose several pools for deconvolution. For plasmin, only
the Trp pool was worth exploring further. For cathepsin
B, we chose to deconvolute the Ile pool and the three
pools that had aromatic amino acids at the Xaa position,
Phe, Trp, and Tyr. The later three pools had activities
comparable to that of several other pools in the library,
but they were assayed at lower concentrations and thus
should contain better inhibitors on average when com-
pared to the other pools. For papain, the Phe and Trp
pools showed the best activity and were thus chosen for
deconvolution.
F igu r e 5. Assay of the Ile, Tyr, and Phe deconvolutions
against cathepsin B. Several of the assays were performed at
1/2 or 1/4 concentration as noted on the x-axes.
The results from the assays of the deconvolutions are
shown in Figures 4-6. Plasmin (Figure 4) clearly
prefers hydrophobic amino acids at the Yaa position of
these inhibitors, with Ile, Phe, Trp, and Tyr showing
significantly higher activities than the other compounds
in the deconvolution. Of the three enzymes, plasmin
again shows the largest variation in activity among the
20 pools of the library.
Cathepsin B displays limited variations in activity
among the four different deconvolutions, as shown in
Figure 5. In general, we observe that this protease has
a small preference for amino acids such as Arg, Gln,
His, Ile, Leu, Phe, and Trp at the Yaa position of the
inhibitors. In addition, the inhibitors with Gly, Hyp, and
Ser at this position all have low activity.
S2′ site of cathepsin B, the specificity of the site only
becomes visible by comparing the results of multiple
deconvolutions and is not easily apparent from a single
deconvolution.
For papain, hydrophobic amino acids including Ile,
Leu, Phe, Trp, and Tyr are preferred at the Yaa position
of the inhibitors. In addition, when Xaa ) Trp, residues
such as Glu, Pro, and Val at the Yaa position have good
activity. For this protease, comparison of the deconvo-
lutions shown in Figure 6 indicates that the specificity
patterns at the Yaa positions for the two sets of inhibitors
have some similarity. This similarity suggests that
binding interactions at the P2′ position are not signifi-
cantly perturbed by differences that may occur when
Phe or Trp is present at P2.
One of the assumptions that we make in using the
iterative deconvolution strategy is that binding interac-
tions at one part of the inhibitor do not significantly
perturb binding at another part. If this assumption were
strictly true, then binding of the inhibitors to the S2
and S2′ subsites should be completely decoupled from
one other, with the result that all four of the deconvo-
lutions in Figure 5 would have identical binding profiles
and differ only in magnitude. For cathepsin B, this does
not appear to be the case. However, on the basis of the
combined data from the four deconvolutions, we con-
clude that this protease has a limited preference for
hydrophobic amino acids at the S2′ subsite. When a
binding site has low specificity as we observed for the
R esyn t h esis a n d E va lu a t ion of In h ibit or s 15-
18. After evaluating all of the data from the assays of
the combinatorial library, we have chosen to resynthe-
size four of the inhibitors (15-18, Scheme 3) using
solution-phase methods and to examine the biological
activity of these compounds in greater detail. The four
inhibitors were selected based upon a combination of
their activity against the individual proteases and the
presence of overlapping activity among the three en-
zymes. The solution-phase synthesis of inhibitors 15-
18, shown in Scheme 3, follows closely the solid-phase
synthesis that was used to construct the library. The
only significant differences in the solution-phase chem-
istry were that Boc protecting groups were used for

Library of Serine and Cysteine Protease Inhibitors
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 19 4005
Ta ble 1. Inhibition of Proteases by Inhibitors 15-18
Ki (µM)
compda
Xaa
Yaa
plasmin
papain
cathepsin B
15A
15B
16A
16B
17b
Trp
Trp
5 ( 0.6
10 ( 2
100 ( 20
150 ( 25
420 ( 40
380b ( 60
Trp
Phe
250b ( 80
Phe Trp
Phe Phe 3000 ( 1000 870 ( 90
490 ( 70
18b
1100 ( 200
a
b
A and B represent two different diastereomers. Assayed as
a mixture of two diastereomers.
Compounds 15-18 all have moderate activity against
papain with inhibition constants that range from 250
to 870 µM. These values are also consistent with the
assays of the library against papain (Figure 6).
Compounds 17 and 18, both of which have Phe at the
Xaa position, are expected to be weak inhibitors of
F igu r e 6. Assay of the Phe and Trp deconvolutions against
papain. Several of the assays were performed at 1/2 or 1/4
concentration as noted on the x-axes.
plasmin as indicated by the low activity of the Xaa
)
Phe pool against this protease (Figure 3). Analysis of
the inhibitors that have been resynthesized by solution-
phase methods gives inhibition constants of 420 and
3000 µM, respectively, for these two compounds. In
contrast inhibitors 15 and 16, which have Trp at the
Xaa position, should have good activity against plasmin
as indicated by the data in Figure 4. We have used
HPLC to separate the two diastereomers of each of these
compounds, although we have not determined their
absolute configurations. Compounds 16A,B have Ki
values of 100 and 150 µM, while compounds 15A,B have
values of 5 and 10 µM, respectively. These results
demonstrate that we have been able to generate an
inhibitor with good activity against plasmin, such as
15A, by synthesizing a combinatorial library of inhibi-
tors around a cyclohexanone nucleus that take advan-
tage of binding interactions in both the S2 and S2′
subsites.
Sch em e 3a
We have recently reported the synthesis and evalu-
ation of a “rationally designed” inhibitor of plasmin that
is based upon a tetrahydrothiopyranone ring system.19
This inhibitor incorporates an aminoalkyl side chain at
the P1 position that is designed to interact with an
aspartic acid residue at the base of the S1 subsite.
Plasmin is specific for substrates that have a positively
charged amino acid at the P1 position, and interactions
with the S1 subsite are critical for recognition and
binding of both substrates and inhibitors.20 In the
tetrahydrothiopyranone-based inhibitors, we have found
that if the aminoalkyl side chain at the P1 position is
removed, there is a 200-300-fold decrease in potency.19
In light of these observations, it is interesting to note
that compound 15A has significant affinity for plasmin
with an inhibition constant of 5 µM, although it does
not incorporate a positively charged P1-like side chain.
Using the combinatorial library described in this
paper, we have discovered 15A as a potential lead
compound for developing high-affinity inhibitors of
plasmin. Two modifications to the structure of 15A are
likely to increase its potency. First, we can incorporate
an aminoalkyl side chain that is positioned to bind in
the S1 subsite. The amide nitrogen at the 2-position of
the cyclohexanone ring should be an appropriate place
to attach such functionality.19 Second, we can substitute
the methylene group at the 4-position of the cyclohex-
anone ring with an electronegative functionality such
a
Reagents: (a) HBTU, DIEA; (b) TFA; (c) Cbz-Xaa-OH where
Xaa ) Trp and Phe, HBTU, DIEA; (d) LiOH; (e) TFA, H2O. One of
two diastereomers is shown.
amines and the C-termini of the inhibitors were pro-
tected as methyl esters. The esters were removed using
basic hydrolysis near the end of the synthesis.
Compounds 15-18 were assayed against plasmin,
cathepsin B, and papain, and the resulting inhibition
data are shown in Table 1. Compound 18 has low
activity against cathepsin B, with a Ki value of 1.1 mM.
This result is expected based upon the low activity of
the compound during the library screening (Figure 5).

4006 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 19
Abato et al.
at water aspirator pressure. Wang resins that were prederiva-
tized with Fmoc amino acids were purchased from Calbiochem-
Novabiochem Corp. Amino acids and their side chain protect-
ing groups were used as follows: Ala, Arg(PMC), Asn(Trt),
Asp(t-Bu), Gln(Trt), Glu(t-Bu), Gly, His(Trt), Hyp(t-Bu), Ile,
Leu, Lys(Boc), Orn(Boc), Phe, Pro, Ser(t-Bu), Thr(t-Bu), Trp-
(Boc), Tyr(t-Bu), and Val.
Syn th esis of th e Libr a r y. Dry Wang resin that was
prederivatized with the 20 Fmoc amino acids (0.2 mmol each)
was combined in a flask and shaken with 125 mL of methylene
chloride. The solvent was removed; the beads were dried under
vacuum and then split into 20 even batches. The beads were
placed into 20 1- × 10-cm Econo-Columns (Biorad) which
served as synthesis vessels. The columns were fitted with
Teflon stopcocks and connected to a 24-port vacuum manifold
(Burdick and J ackson) that was used to drain solvents and
reagents from the columns. The resin in each column was
swelled in DMF; then the Fmoc protecting groups were
removed by treatment with 2 mL of a 1:1 solution of DMF and
piperidine for 10 min. After washing with 5 × 5 mL of DMF,
a positive Kaiser test indicated the presence of free amines
on the resin.
Cou p lin g of Com p ou n d 4. To couple compound 4 to the
resin, a stock solution was prepared that contained compound
4 (5.6 g, 12 mmol, 3 equiv), HBTU (4.54 g, 12 mmol, 3 equiv),
diisopropylethylamine (DIEA; 2.76 mL, 16 mmol, 4 equiv), and
DMF solvent in a total volume of 40 mL. Aliquots (2 mL) of
this stock solution were added to each of the synthesis vessels
and the reactions were gently agitated for 1.5 h. A negative
Kaiser test indicated that the reaction had gone to completion.
The resin was washed with 5 × 5 mL of DMF; then any
unreacted amino groups were capped by treating the resin for
10 min with 2 mL of a freshly prepared stock solution that
contained acetic anhydride (1.5 mL, 20 mmol, 5 equiv), DIEA
(2.76 mL, 16 mmol, 4 equiv), and DMF in a total volume of 40
mL. After capping, the resin was then washed with 5 × 5 mL
of DMF, and the N-terminal Fmoc group was deprotected as
described above.
Cou p lin g of th e Secon d Am in o Acid . To couple the
second amino acid to the resin, the 20 synthesis vessels were
treated with 3 equiv of an N-Cbz amino acid, each vessel
receiving a different amino acid. The reactions also contained
3 equiv of HBTU and 4 equiv of DIEA in 2 mL of DMF and
were agitated for 1.5 h. After the reactions were complete as
judged by a negative Kaiser test, the resin was washed with
5 × 5 mL of DMF, 5 × 10 mL of methylene chloride, and 5 ×
10 mL of MeOH, and dried under vacuum overnight.
Clea va ge fr om th e Resin a n d Rem ova l of P r otectin g
Gr ou p s. The inhibitors were cleaved from the resin and the
protecting groups on the amino acid side chains were removed
by treating each batch of resin with 3 mL of cleavage cocktail
for 2 h. The cleavage cocktail contained 95% TFA, 2.5% water,
and 2.5% triisopropylsilane. The solutions were filtered to
remove the resin, and the beads were washed with 3 × 1 mL
of fresh TFA. The ketal protecting groups on the cyclohexanone
rings were removed by adding 3 mL of water to the TFA
solutions from the previous reactions and agitating the solu-
tions for 30 h at room temperature. After the initial addition
of water to the TFA solutions, the reactions became cloudy.
However, after several hours at room temperature, the reac-
tions became homogeneous. After the reactions were complete,
the TFA and water were removed and the residual material
was dried at 0.02 mmHg for 48 h. The 20 batches of inhibitors
were each dissolved in 1 mL of DMSO, filtered, and stored in
a freezer at -48 °C. If a total yield of 50% is assumed for the
synthesis, then each inhibitor stock solution contained 20
different compounds, with a concentration of 5 mM for each
individual compound in the mixtures.
as S, O, or SO2. This electronegative functionality will
increase the electrophilicity of the ketone by inducing
an unfavorable through-space electrostatic repulsion
between the electronegative group and the dipole of the
ketone.7 We have demonstrated previously that there
is a good correlation between ketone electrophilicity and
inhibition constants in 4-heterocyclohexanone-based
inhibitors.7
Con clu sion s
In this paper we have described the construction and
screening of a combinatorial library of protease inhibi-
tors that extend into both the S and S′ specificity sites.
Using this method we have found an inhibitor that has
significant affinity for the serine protease plasmin. In
addition, this work has shown that combinatorial chem-
istry is an efficient method for probing the specificity
of the S2′ subsite. Currently there is little information
available concerning the binding specificity of this
position for many proteases.
Our data have shown that for plasmin, the S2′ subsite
prefers to bind hydrophobic and especially aromatic
amino acids. Furthermore, binding of such hydrophobic
residues in this site significantly increases the affinity
of the inhibitor for the enzyme. On the other hand, the
S2′ subsites of cathepsin B and papain do not appear
to have strong preferences for any particular amino acid,
and binding in this position leads to only incremental
increases in affinity. Concerning the S2 subsites, the
data that we have presented are consistent with the
known specificity of cathepsin B and papain, which
prefer hydrophobic amino acids at this position.18 For
plasmin, it has been believed that Phe binds well in the
S2 subsite. However our results show that Trp, with
its larger aromatic surface and potential for hydrogen
bonding, provides up to an 80-fold increase in affinity
when compared to Phe. Clearly, combinatorial chemis-
try has provided useful information concerning the
binding interactions at the S2 and S2′ subsites of several
proteases. A similar approach should be equally ame-
nable to exploring the specificities of the other leaving
group subsites and for incorporating nonpeptidic func-
tionality into the inhibitors.
Exp er im en ta l Section
Gen er a l Meth od s. NMR spectra were recorded on Bruker
Avance-300 or Avance-400 instruments. Spectra were cali-
brated using TMS (δ ) 0.00 ppm) for 1H NMR. Mass spectra
were recorded on a Kratos MS 80 under fast-atom bombard-
ment (FAB) conditions or were performed using electrospray
ionization at the Harvard University Chemistry Department
Mass Spectrometry Facility. HPLC analyses were performed
on a Rainin HPLC system with Rainin Microsorb silica or C18
columns and UV detection. Semipreparative HPLC was per-
formed on the same system using semipreparative columns
(21.4 × 250 mm). UV spectra were recorded on a Perkin-Elmer
8452A diode array UV-vis spectrophotometer.
Reactions were conducted under an atmosphere of dry
nitrogen in oven-dried glassware. Anhydrous procedures were
conducted using standard syringe and cannula transfer tech-
niques. THF and toluene were distilled from sodium and
benzophenone. Piperidine was distilled from KOH. Other
solvents were of reagent grade and were stored over 4 Å
molecular sieves. All other reagents were used as received.
Organic solutions were dried over MgSO4 unless otherwise
noted. Solvent removal was performed by rotary evaporation
Decon volu tion Syn th eses. The solid-phase syntheses of
the various deconvolutions were performed using a similar
procedure as described above. The only changes that were
made were that the mix and split step at the beginning was
omitted and the syntheses were performed on a 0.04-mmol
scale per inhibitor.

Library of Serine and Cysteine Protease Inhibitors
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 19 4007
P a p a in Assa ys. Papain (EC 3.4.22.2; purchased from
Sigma) was assayed using L-BAPNA as the substrate, and the
reaction progress was monitored by UV spectroscopy.11a Initial
rates were determined by monitoring the formation of p-
nitroaniline at 412 nm from 30 to 120 s after mixing. Assay
solutions contained 1.5 mM substrate, 5 mM EDTA, 5 mM
cysteine, 50 mM sodium phosphate at pH 7.5, and 15% DMSO.
Assays of the initial library, which constituted 20 pools of
inhibitors with 20 compounds per pool, also contained a total
inhibitor concentration of 1 mM (50 µM of each individual
inhibitor) and were performed in triplicate. Assays of the
deconvolutions contained an inhibitor concentration of 200 µM.
Assays of compounds 15-18 contained 1.7 mM substrate and
inhibitor concentrations that ranged from 2 to 500 µM. Ki
values were calculated using a Dixon analysis. Data analysis
was performed with the commercial graphing package Grafit
(Erithacus Software Ltd.). The Km value under these conditions
was measure to be 6.8 mM. In the absence of DMSO, the Km
value has been reported as 3.2 mM in 100 mM phosphate
buffer at pH 7.5.21
Ca th ep sin B Assa ys. Cathepsin B (EC 3.4.22.1; purchased
from Sigma) was assayed using L-BAPNA as the substrate.22
Assay solutions contained 1.5 mM substrate, 1.5 mM EDTA,
3 mM DTT, 50 mM sodium phosphate buffer at pH 7.4, and
15% DMSO. Assays of compounds 15-18 contained 0.7 mM
substrate. All other conditions were the same as indicated for
the assays of papain. The Km value under these conditions was
measure to be 3.3 mM. The Km value in the absence of DMSO
and in sodium acetate buffer at pH 5.1 is reported to be 1.3
mM.22
2H), 7.62 (m, 2H), 7.78 (d, J ) 7.5 Hz, 2H); 13C NMR (100
MHz, CDCl3) δ 18.8, 22.0, 25.0, 28.4, 32.1, 47.3, 59.0, 66.7,
98.8, 119.9, 125.1, 127.0 127.6, 141.3, 144.0, 156.5, 179.0;
HRMS-FAB (M + Na+) calcd for C27H31NNaO6 488.2049, found
488.2041.
Tr yp top h a n Meth yl Ester 9. To carboxylic acid 8 (250 mg,
0.73 mmol) were added H-Trp-OMe (250 mg, 0.87 mmol),
HBTU (398 mg, 0.88 mmol), and 2 mL of DMF, followed by
DIEA (0.40 mL, 1.75 mmol). The reaction was stirred for 3 h
at room temperature, then diluted with 50 mL of EtOAc. The
organic layer was washed with 35 mL of 1 N KHSO4, 35 mL
of saturated NaHCO3, and 35 mL of brine and dried over
MgSO4. The resulting solution was concentrated by rotary
evaporation, and the residual material was purified by flash
chromatography (EtOAc) to yield methyl ester 9 (400 mg, 0.68
mmol, 74%): 1H NMR (300 MHz, MeOH-d4) δ 1.30-1.75 (m,
19H), 1.76-2.00 (m, 1H), 2.05-2.10 (m, 2H), 2.15-2.40 (m,
2H), 3.12-3.21 (m, 1H), 3.62-3.70 (m, 3H), 3.72-3.95 (m, 4H),
4.72-4.79 (m, 1H), 6.89-7.21 (m, 3H), 7.35 (d, J ) 8.0 Hz,
1H), 7.53-7.56 (d, J ) 7.5 Hz, 1H); 13C NMR (75 MHz, MeOH-
d4) δ 18.9, 25.2, 27.4, 29.7, 33.8, 51.7, 53.7, 58.9, 79.1, 99.2,
109.8, 111.4, 118.1, 118.8, 121.4, 123.5, 127.6, 137.1, 157.2,
173.2, 175.2, 209.2; HRMS-ESI (M + Na+) calcd for C29H41N3-
NaO7 566.2840, found 566.2858.
P h en yla la n in e Meth yl Ester 10. To carboxylic acid 8 (500
mg, 1.46 mmol) were added H-Phe-OMe (312 mg, 1.75 mmol),
HBTU (796 mg, 1.75 mmol), and 4 mL of DMF, followed by
DIEA (0.8 mL, 3.5 mmol). The procedure was identical to that
used for the preparation of compound 9. The crude material
was purified by flash chromatography (2:3 hexanes/EtOAc) to
yield compound 10 (1.10 g, 2.14 mmol, 74%): 1H NMR (300
MHz, DMSO-d6) δ 1.20-2.83 (m, 19H), 1.77-2.00 (m, 2H),
2.01-2.34 (m, 2H), 2.93-3.02 (m, 1H), 3.22-3.14 (m, 1H),
3.55-3.71 (m, 3H), 3.74-3.86 (m, 4H), 4.66-4.74 (m, 1H),
7.22-7.33 (m, 5H); 13C NMR (75 MHz, MeOH-d4) δ 19.0, 25.3,
27.8, 28.7, 33.8, 37.4, 51.8, 54.2, 59.0, 17.2, 99.2, 126.9, 128.5,
129.2, 137.2, 157.1, 172.7, 175.3; HRMS-ESI (M + Na+) calcd
for C41H48N4NaO8 527.2732, found 527.2751.
P la sm in Assa ys. Plasmin (EC 3.4.21.7; purchased from
Sigma) was assayed using D-Val-Leu-Lys-p-nitroanilide as
substrate.23 Assay solutions contained 250 µM substrate, 50
mM sodium phosphate buffer at pH 7.4, and 10% DMSO.
Assays of compounds 15-18 contained 180 µM substrate. All
other conditions were the same as indicated for the assays of
papain. The Km value under these conditions was measure to
be 180 µM. The Km value in the absence of DMSO and in Tris-
HCl buffer at pH 8.3 is reported to be 270 µM.24
Tr p -X-Tr p Ester 11. To compound 9 (370 mg, 0.68 mmol)
was added a solution of CH2Cl2 (11.5 mL), TFA (4.5 mL), and
triisopropylsilane (0.38 mL). The reaction was stirred at room
temperature for 1 h; then the solvents were removed under
vacuum. The crude amine was dissolved in 2 mL of DMF, and
DIEA (∼0.5 mL) was added to neutralize the residual TFA
and to raise the pH to 8 (as measured with moist pH paper).
Cbz-Trp-OH (276 mg, 0.816 mmol), HBTU (309 mg, 0.816
mmol), and DIEA (0.28 mL, 1.6 mmol) were added to the flask
and the reaction was allowed to stir for 2 h at room temper-
ature. The workup was identical to the procedure used in the
preparation of compound 9. The crude material was purified
by flash chromatography (1:5 hexanes/EtOAc) to yield 11 (0.40
g, 0.52 mmol, 76%): 1H NMR (300 MHz, MeOH-d4) δ 0.94 (t,
J ) 6.9 Hz, 1H), 1.12-1.65 (m, 10H), 1.75-1.90 (m, 2H), 2.10-
2.14 (m, 1H), 2.20-2.29 (m, 1H), 3.11-3.20 (m, 3H), 3.28-
3.30 (m, 1H), 3.50-3.70 (m, 6H), 4.50-4.58 (m, 1H), 4.75-
4.85 (m, 1H), 4.98-5.08 (m, 2H), 7.01-7.35 (m, 13H), 7.51-
7.61 (m, 2H); 13C NMR (75 MHz, MeOH-d4) δ 13.5, 18.3, 18.7,
19.9, 24.9, 25.2, 27.4, 28.3, 29.1, 33.7, 51.9, 53.7, 56.4, 58.9,
60.5, 66.6, 99.8, 109.8, 111.0, 118.1, 118.4, 118.8, 121.4, 121.5,
127.7, 128.5, 137.2, 157.1, 172.9, 173.2, 173.5, 175.2; HRMS-
FAB (M + H+) calcd for C43H50N5O8 764.3659, found 764.3647.
Tr p -X-P h e Ester 12. Compound 10 (387 mg, 0.54 mmol)
was deprotected and coupled to Cbz-Trp-OH using a procedure
that was similar to that used for the preparation of compound
11. The crude material was purified by flash chromatography
(1:5 hexanes/EtOAc) to give compound 12 (255 mg, 0.391
mmol, 73%): 1H NMR (300 MHz, MeOH-d4) δ 0.82-0.91 (t, J
) 6.3 Hz, 1H), 1.21-1.60 (m, 10H), 1.90-1.96 (m, 1H), 1.97-
2.03 (m, 1H), 2.05-2.30 (m, 2H), 2.90-2.99 (m, 1H), 3.07-
3.30 (m, 3H), 3.60-3.85 (m, 6H), 4.53 (t, J ) 7.3 Hz, 1H), 4.66-
4.72 (m, 1H), 5.01-5.12 (m, 2H), 7.00-7.35 (m, 14H), 7.63 (d,
J ) 7.7 Hz, 1H); 13C NMR (75 MHz, MeOH-d4) δ 13.5, 18.8,
19.8, 25.3, 28.3, 33.7, 37.3, 51.8, 54.2, 59.0, 60.5, 66.6, 98.8,
110.0, 111.3, 118.4, 118.5, 121.4, 123.7, 126.9, 127.7, 128.5,
F m oc Alk en e 3. To a solution of 210 (9.9 g, 30.5 mmol) in
CH2Cl2 (75 mL) was added TFA (25 mL) at 0 °C. The solution
was warmed to room temperature, stirred for 1 h, and
concentrated to remove the TFA. The resulting oil was
redissolved in 20 mL of CH2Cl2 and washed with saturated
Na2CO3 and brine (200 mL). The solution was dried over
Na2CO3 and filtered, and then DIEA (6 mL) and FmocCl (7.89
g, 30.5 mmol) were added. This solution was stirred for 3 h,
then washed with 1 N HCl, saturated Na2CO3, and brine, and
dried. The solution was then reduced to approximately 20 mL,
diluted with hexanes (200 mL), and placed in a refrigerator
overnight to allow crystallization. White crystals of 3 (9.22 g,
67%) were isolated by vacuum filtration and dried in vacuo:
1H NMR (400 MHz, CDCl3) δ 1.26-1.30 (m, 1H), 1.42-1.62
(br m, 7H), 1.62-1.90 (br m, 3H), 1.93-2.02 (m, 1H), 2.14-
2.23 (m, 1H), 3.85-3.90 (m, 4H), 4.25-4.44 (m, 3H), 4.97-
5.07 (m, 2H), 5.15 (br s, 1H), 5.78-5.88 (m, 1H), 7.31 (t, J )
7.4 Hz, 2H), 7.39 (t, J ) 7.4 Hz, 2H), 7.62 (t, J ) 6.8 Hz, 2H),
7.76 (d, J ) 7.5 Hz, 2H); 13C NMR (100 MHz, CDCl3) δ 14.1,
18.8, 22.6, 25.1, 28.8, 31.6, 31.8, 47.3, 58.8, 59.0, 66.5, 99.1,
114.8, 119.9, 125.1, 127.0, 127.6, 138.7, 141.3, 156.4; HRMS-
FAB (M + Na+) calcd for C28H33NNaO4 470.2307, found
470.2322.
F m oc Ca r boxylic Acid 4. To compound 3 (0.46 g, 1.1
mmol) were added 25 mL of 7:3 acetone/H2O, NaIO4 (1.1 g,
5.2 mmol), KMnO4 (32 mg, 0.20 mmol), and NaHCO3 (0.10 g,
1.0 mmol) and the reaction was stirred at room temperature
for 7 h. The reaction was diluted with EtOAc (150 mL) and
washed with 100 mL of 1 N HCl and 100 mL of brine, and the
organic layer was dried over MgSO4. The crude material was
purified by flash chromatography (3:2 EtOAc/hexanes) to yield
compound 4 (308 mg, 0.66 mmol, 63%): 1H NMR (300 MHz,
CDCl3) δ 1.43-1.71 (m, 10H), 2.07-2.12 (m, 3H), 2.35-2.53
(m, 2H), 3.91 (br s, 4H), 4.14-4.30 (m, 2H), 4.37-4.48 (m, 2H),
5.17 (br s, 1H), 7.31 (t, J ) 7.4 Hz, 2H), 7.41 (t, J ) 7.4 Hz,

4008 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 19
Abato et al.
129.3, 137.1, 137.4, 172.0, 172.7, 175.3; HRMS-ESI (M + Na+)
calcd for C41H48N4NaO8 747.4932, found 747.4967.
194.4, 209.0; HRMS-ESI (M + Na+) calcd for C39H41N5NaO7
714.2900, found 714.2883 for the mixture of diastereomers
15A,B.
P h e-X-Tr p Ester 13. Compound 9 (450 mg, 0.83 mmol) was
deprotected and coupled to Cbz-Phe-OH (297 mg, 1.0 mmol)
using a procedure that was similar to that used for the
preparation of compound 11. The crude material was purified
by flash chromatography (1:4 hexanes/EtOAc) to yield 13 (280
mg, 0.38 mmol, 46%): 1H NMR (300 MHz, DMSO-d6) δ 1.15-
1.55 (m, 10H), 1.63-1.66 (m, 1H), 1.70-1.90 (m, 1H), 1.96-
2.00 (m, 2H), 2.14-2.22 (m, 1H), 2.70-2.78 (m, 1H), 2.90-
3.19 (m, 3H), 3.40 (s, 3H), 3.56-3.60 (m, 4H), 3.74 (m, 3H),
4.32-4.40 (m, 1H), 4.47-4.55 (m, 1H), 4.93-4.97 (m, 2H),
6.97-7.02 (t, J ) 7.3 Hz, 1H), 7.05-7.10 (t, J ) 7.9 Hz, 1H),
7.15-7.35 (m, 10H), 7.46-7.52 (m, 2H), 8.23-8.28 (m, 1H),
10.9 (s, 1H); 13C NMR (75 MHz, MeOH-d4) δ 15.8, 20.3, 22.4,
26.4, 28.7, 35.0, 39.3, 53.4, 54.8, 57.7, 60.0, 61.4, 66.8, 100.0,
111.3, 113.1, 119.7, 120.1, 122.6, 125.2, 127.8, 128.7, 129.1,
129.4, 129.6, 129.7, 129.9, 130.0, 130.9, 137.8, 138.7, 139.9,
157.5, 172.9, 174.3; HRMS-ESI (M + Na+) calcd for C41H48N4-
NaO8 747.3367, found 747.3386.
In h ibitor 16. Compound 12 (291 mg, 0.40 mmol) was
deprotected using the procedure outlined for the preparation
of inhibitor 15. The diastereomers were separated using
preparative reverse-phase HPLC (40% MeCN, 60% H2O, 0.1%
TFA) to give 16A (25 mg, 0.04 mmol, 10%) and 16B (25 mg,
0.04 mmol, 10%). Similar to inhibitor 15, inhibitor 16 was not
very soluble in the solvent used during the purification so that
a majority of the material was left in crude form and was not
purified. 16A: 1H NMR (300 MHz, DMSO-d6) δ 1.15-1.20 (m,
1H), 1.25-1.35 (m, 1H), 1.36-1.60 (m, 1H), 1.65-1.90 (m, 3H),
1.97-2.13 (m, 3H), 2.13-2.25 (m, 1H), 2.35-2.40 (m, 1H),
2.80-2.96 (m, 2H), 3.05-3.21 (m, 2H), 3.34 (s, 3H), 4.32-4.50
(m, 3H), 4.95 (s, 2H), 6.95-7.47 (m, 12H), 7.64-7.67 (d, J )
7.4 Hz, 1H), 8.05 (d, J ) 7.1 Hz, 1H), 8.14 (d, J ) 8.0 Hz, 1H),
10.82 (s, 1H); 13C NMR (75 MHz, DMSO-d6) δ 24.1, 25.5, 28.9,
33.3, 34.9, 35.6, 49.0, 54.2, 56.4, 58.2, 66.0, 111.0, 112.1, 119.0,
119.4, 121.7, 124.7, 127.3, 127.8, 128.2, 128.5, 129.0, 129.2,
130.0, 136.8, 138.0, 138.7, 156.6, 172.2, 172.8, 174.1, 209.0.
16B: 1H NMR (300 MHz, DMSO-d6) δ 1.10-1.21 (m, 1H),
1.25-1.34 (m, 2H), 1.65-1.80 (m, 3H), 1.91-2.11 (m, 4H),
2.25-2.35 (m, 1H), 2.80-2.97 (m, 2H), 3.05-3.12 (m, 2H), 3.30
(s, 3H), 4.29-4.42 (m, 3H), 4.89-4.99 (m, 2H), 6.95-7.42 (m,
12H), 7.67 (d, J ) 7.5 Hz, 1H), 8.05 (d, J ) 7.4 Hz, 1H), 8.16
(d, J ) 8.2 Hz, 1H), 10.82 (s, 1H), 12.61 (s, 1H); 13C NMR (75
MHz, DMSO-d6) δ 24.2, 25.5, 28.9, 33.2, 34.6, 25.3, 37.5, 48.9,
54.1, 56.4, 58.1, 66.0, 111.1, 112.1, 119.0, 119.4, 121.7, 124.7,
127.2, 128.3, 128.5, 129.0, 129.9, 136.9, 137.8, 138.7, 156.7,
172.4, 172.9, 174.1, 209.1; HRMS-ESI (M + Na+) calcd for
P h e-X-P h e Ester 14. Compound 10 (302 mg, 0.60 mmol)
was deprotected and coupled to Cbz-Phe-OH (244 mg, 0.82
mmol) using a procedure that was similar to that used for the
preparation of compound 11. The crude material was purified
by flash chromatography (1:4 hexanes/EtOAc) to yield 14 (267
mg, 0.39 mmol, 65%): 1H NMR (300 MHz, MeOH-d4) δ 0.86-
0.95 (m, 1H), 1.20-1.55 (m, 11H), 1.82-1.90 (m, 1H), 2.03-
2.30 (m, 2H), 2.82-3.02 (m, 2H), 3.14-3.22 (m, 2H), 3.68-
3.85 (m, 6H), 4.44-4.49 (q, J ) 5.1 Hz, 1H), 4.67-4.70 (m,
1H), 5.04 (s, 2H), 7.21-7.32 (m, 13H), 7.51 (d, J ) 8.3 Hz,
1H), 8.30 (d, J ) 7.6 Hz, 1H); 13C NMR (75 MHz, MeOH-d4) δ
13.5, 18.8, 24.6, 25.3, 33.7, 38.2, 51.7, 54.2, 56.9, 59.0, 59.1,
60.6, 66.6, 99.0, 126.7, 126.8, 126.9, 127.7, 128.0, 128.4, 128.5,
129.2, 129.3, 129.4, 137.2, 137.6, 157.1, 172.8, 175.2, 175.3;
HRMS-ESI (M + Na+) calcd for C39H47N3NaO8 708.3259, found
708.3251.
C
37H40N4NaO7 675.2792, found 675.2784 for the mixture of
diastereomers 16A,B.
In h ibitor 17. Compound 13 (240 mg, 0.33 mmol) was
deprotected using the procedure outlined for the preparation
of inhibitor 15. The purification was accomplished using
preparative reverse-phase HPLC (40% MeCN, 60% H2O, 0.1%
TFA) to yield 17 (99 mg, 0.15 mmol, 45%): 1H NMR (300 MHz,
DMSO-d6) δ 1.10-1.20 (m, 1H), 1.25-1.35 (m, 2H), 1.67-1.72
(m, 2H), 1.80-1.87 (m, 2H), 2.00-2.15 (m, 4H), 2.35-2.38 (m,
2H), 2.72-2.79 (m, 2H), 2.94-3.01 (m, 2H), 3.14 (d, J ) 4.6
Hz, 1H), 3.18-3.20 (m, 1H), 4.35-4.49 (m, 2H), 4.95 (s, 1H),
6.95-7.35 (m, 11H), 7.50-7.55 (m, 2H), 8.10 (q, J ) 3.4 Hz,
1H), 8.82 (s, 1H); 13C NMR (75 MHz, DMSO-d6) δ 23.1, 24.1,
25.5, 27.9, 33.3, 34.7, 35.3, 35.6, 49.1, 53.7, 57.0, 58.2, 66.0,
110.9, 112.2, 119.0, 121.8, 124.3, 127.1, 128.1, 128.2, 128.8,
129.1, 130.1, 136.9, 137.9, 138.8, 139.1, 156.6, 156.7, 171.8,
172.2, 172.9, 174.4, 207.3, 209.1; HRMS-ESI (M + Na+) calcd
for C37H40N4NaO7 675.2792, found 675.2797.
In h ibitor 18. Compound 14 (271 mg, 0.40 mmol) was
deprotected using the procedure outlined for the preparation
of inhibitor 15. The purification was accomplished using
preparative reverse-phase HPLC (40% MeCN, 60% H2O, 0.1%
TFA) to yield 18 (100 mg, 0.15 mmol, 39%): 1H NMR (300
MHz, DMSO-d6) δ 1.65-1.80 (m, 3H), 2.00-2.10 (m, 3H),
2.35-2.42 (m, 3H), 2.95-3.10 (m, 2H), 4.30-4.44 (m, 2H), 4.95
(s, 2H), 7.20-7.30 (m, 13H), 7.56 (d, J ) 8.3 Hz, 1H), 8.03-
8.16 (m, 1H), 12.40 (s, 1H); 13C NMR (75 MHz, DMSO-d6) δ
24.2, 25.5, 33.3, 33.4, 33.7, 34.9, 35.7, 38.6, 49.0, 54.1, 57.0,
58.1, 66.0, 71.0, 127.1, 127.2, 128.2, 128.4, 129.0, 130.1, 137.9,
138.7, 138.9, 139.1, 156.7, 164.2, 171.8, 172.1, 172.7, 209.0;
HRMS-ESI (M + Na+) calcd for C35H39N3NaO8 636.2684, found
636.2654.
In h ibitor 15. Compound 11 (0.39 g, 0.511 mmol) was
dissolved in 20 mL of MeOH. To this solution was added a
solution of LiOH (80 mg, 3.3 mmol) dissolved in 5 mL of water,
and the reaction was stirred for 12 h at room temperature.
The basic solution was neutralized with 1 N HCl to pH 7, and
the solvents were removed under vacuum at 25 °C. The crude
carboxylic acid was dissolved in 8 mL of TFA and 3 mL of
water and the solution was stirred for an additional 12 h at
room temperature. The solvents were again removed under
vacuum at 25 °C and the crude material was purified by flash
chromatography (7% MeOH in CH2Cl2). Diastereomers 15A,B
were separated by preparative HPLC on a silica column (96.5%
CH2Cl2, 3.4% MeOH, 0.1% TFA). The compound had a low
solubility in this solvent system, so the crude material was
suspended in solvent and filtered to remove the product that
did not dissolve, and the portion that remained in solution was
purified by HPLC. This procedure gave 15A (6 mg, 0.009
mmol, 2%) and 15B (2 mg, 0.003 mmol, 1%). The large majority
of the material was left in crude form and not purified. 15A:
1H NMR (300 MHz, DMSO-d6) δ 1.25-1.75 (m, 2H), 1.81-
2.05 (m, 3H), 2.13-2.45 (m, 4H), 3.06-3.23 (m, 2H), 3.34-
3.42 (m, 2H), 4.42-4.65 (m, 3H), 5.03-5.13 (m, 2H), 7.10-
7.56 (m, 12H), 7.68 (d, J ) 7.6 Hz, 1H), 7.81 (d, J ) 7.8 Hz,
1H), 8.17-8.31 (m, 1H), 10.97 (s, 1H); 13C NMR (75 MHz,
DMSO-d4) δ 24.1, 25.4, 27.9, 28.8, 33.2, 34.7, 49.0, 53.5, 56.3,
58.2, 65.4, 111.1, 112.2, 119.0, 121.7, 124.3, 124.7, 128.3, 129.2,
136.9, 137.9, 156.6, 172.4, 172.8, 209.1. 15B: 1H NMR (300
MHz, DMSO-d6) δ 0.92-1.22 (m, 1H), 1.20-1.29 (m, 3H),
1.52-1.60 (m, 2H), 1.63-1.70 (m, 1H) 1.91-2.07 (m, 4H),
2.25-2.34 (m, 1H) 2.74-3.08 (m, 4H), 4.22-4.39 (m, 6H), 4.84
(s, 2H), 6.83-7.31 (m, 11H), 7.42 (d, J ) 7.6 Hz, 1H), 7.52 (d,
J ) 7.7 Hz, 1H), 7.92 (d, J ) 7.4 Hz, 1H), 7.98 (d, J ) 8.0 Hz,
1H), 10.70 (s, 2H); 13C NMR (75 MHz, DMSO-d4) δ 24.1, 25.5,
27.9, 28.9, 33.3, 34.8, 35.5, 46.9, 48.2, 49.1, 53.7, 56.4, 58.2,
66.1, 110.9, 112.2, 119.0, 119.4, 121.6, 124.4, 124.7, 128.0,
128.2, 128.5, 129.1, 136.9, 137.9, 156.6, 172.3, 172.9, 174.5,
Ack n ow led gm en t. This research was supported by
the NIH (Grant 1 R01 GM57327-01), the Petroleum
Research Fund administered by the American Chemical
Society (Grant 30544-G4), and the U.S. Army Medical
Research and Materiel Command - Breast Cancer
Research Initiative (Grant DAMD17-96-1-6161, Career
Development Award to C.T.S.). J .L.C. and P.A. were
supported by GAANN Fellowships from the Department
of Education. J .L.C. was also supported by a University

Library of Serine and Cysteine Protease Inhibitors
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 19 4009
Human Rhinovirus 3C Protease Inhibitors. 3. Structure-Activ-
ity Studies of Ketomethylene-Containing Peptidomimetics. J .
Med. Chem. 1999, 42, 1203. (b) Roush, W. R.; Gwaltney, S. L.,
II; Cheng, J .; Scheidt, K. A.; McKerrow, J . H.; Hansell, E. Vinyl
Sulfonate Esters and Vinyl Sulfonamides: Potent, Irreversible
Inhibitors of Cysteine Proteases. J . Am. Chem. Soc. 1998, 120,
10994. (c) Liu, S.; Hanzlik, R. P. Structure-Activity Relation-
ships for Inhibition of Papain by Peptide Michael Acceptors. J .
Med. Chem. 1992, 35, 1067. (d) Harbeson, S. L.; Abelleira, S.
M.; Akiyama, A.; Barrett, R., III; Carroll, R. M.; Straub, J . A.;
Tkacz, J . N.; Wu, C.; Musso, G. F. Stereospecific Synthesis of
Peptidyl R-Keto Amides as Inhibitors of Calpain. J . Med. Chem.
1994, 37, 2918. (e) Bromme, D.; Klaus, J . L.; Okamoto, K.;
Rasnick, D.; Palmer, J . T. Peptidyl Vinyl Sulfones: A New Class
of Potent and Selective Cysteine Protease Inhibitors. S2P2
Specificity of Human Cathepsin O2 in Comparison With Cathe-
psins S and L. Biochem. J . 1996, 315, 85. (f) Palmer, J . T.;
Rasnick, D.; Klaus, J . L.; Bromme, D. Vinyl Sulfones as
Mechanism-Based Cysteine Protease Inhibitors. J . Med. Chem.
1995, 38, 3193. (g) Chatterjee, S.; Gu, Z.-Q.; Dunn, D.; Tao, M.;
J osef, K.; Tripathy, R.; Bihovsky, R.; Senadhi, S. E.; O’Kane, T.
M.; McKenna, B. A.; Mallya, S.; Ator, M. A.; Bozyczko-Coyne,
D.; Siman, R.; Mallamo, J . P. D-Amino Acid Containing, High-
Affinity Inhibitors of Recombinant Human Calpain I. J . Med.
Chem. 1998, 41, 2663. (h) Angelastro, M. R.; Mehdi, S.; Burkhart,
J . P.; Peet, N. P.; Bey, P. R-Diketone and R-Keto Ester Deriva-
tives of N-Protected Amino Acids and Peptides as Novel
Inhibitors of Cysteine and Serine Proteinases. J . Med. Chem.
1990, 33, 11.
Fellowship from Brown University. We thank Dr.
Andrew Tyler of the Harvard University Chemistry
Department Mass Spectrometry Facility for providing
several of the mass spectra. This facility is supported
by grants from the NSF (CHE-9020045) and the NIH
(SIO-RR06716).
Su p p or tin g In for m a tion Ava ila ble: HPLC characteriza-
tion for compounds 15-18. This material is available free of
charge via the Internet at http://pubs.cas.org.
Refer en ces
(1) For several recent reviews, see: (a) Lowe, G. Combinatorial
Chemistry. Chem. Soc. Rev. 1995, 24, 309. (b) Ellman, J . A.
Design, Synthesis, and Evaluation of Small-Molecule Libraries.
Acc. Chem. Res. 1996, 29, 132. (c) Gordon, E. M.; Gallop, M. A.;
Patel, D. V. Strategy and Tactics in Combinatorial Organic
Synthesis. Applications to Drug Discovery. Acc. Chem. Res. 1996,
29, 144. (d) Balkenhohl, F.; von dem Bussche-Hunnefeld, C.;
Lansky, A.; Zechel, C. Combinatorial Synthesis of Small Organic
Molecules. Angew. Chem., Int. Ed. Engl. 1996, 35, 2288. (e)
Thompson, L. A.; Ellman, J . A. Synthesis and Applications of
Small Molecule Libraries. Chem. Rev. 1996, 96, 555.
(2) Campbell, D. A.; Bermak, J . C.; Burkoth, T. S.; Patel, D. V. A
Transition State Analogue Inhibitor Combinatorial Library. J .
Am. Chem. Soc. 1995, 117, 5381.
(3) Owens, R. A.; Gesellchen, P. D.; Houchins, B. J .; Dimarchi, R.
D. The Rapid Identification of HIV Protease Inhibitors Through
the Synthesis and Screening of Defined Peptide Mixtures.
Biochem. Biophys. Res. Commun. 1991, 181, 402.
(4) Kick, E. K.; Ellman. J . A. Expedient Method for the Solid-Phase
Synthesis of Aspartic Acid Protease Inhibitors Directed toward
the Generation of Libraries. J . Med. Chem. 1995, 38, 1427.
(5) Wang, J . T.; Li, S.; Wideburg, N.; Krafft, G. A.; Kempf, D. J .
Synthetic Chemical Diversity: Solid-Phase Synthesis of Librar-
ies of C2 Symmetric Inhibitors of HIV Protease Containing
Diamino Diol and Diamino Alcohol Cores. J . Med. Chem. 1995,
38, 2995.
(6) Rano, T. A.; Timkey, T.; Peterson, E. P.; Rotonda, J .; Nicholson,
D. W.; Beckeer, J . W.; Chapman, K. T.; Thornberry, N. A. A
Combinatorial Approach Defines Specificities of Members of the
Caspase Family and Granzyme B. Functional Relationships
Established for Key Mediators of Apoptosis. Chem. Biol. 1997,
4, 149.
(13) (a) Sebestyen, F.; Dibo, G.; Kovacs, A.; Furka, A. Chemical
Synthesis of Peptide Libraries. Bioorg. Med. Chem. Lett. 1993,
3, 413. (b) Furka, A.; Sebestyen, F.; Asgedom, M.; Dibo, G.
General Method for Rapid Synthesis of Multicomponent Peptide
Mixtures. Int. J . Pept. Protein Res. 1991, 37, 487.
(14) Dooley, C. T.; Chung, N. N.; Wilkes, B. C.; Schiller, P. W.;
Bidlack, J . M.; Pasternak, G. W.; Houghten, R. A. An All
D-Amino Acid Opioid Peptide with Central Analgesic Activity
from a Combinatorial Library. Science 1994, 266, 2019.
(15) (a) Yamaguchi, N.; Chung, S.; Shiroeda, O.; Koyama, K.;
Imanishi, J . Characterization of a Cathepsin L-Like Enzyme
Secreted from Human Pancreatic Cancer Cell Line HPC-YP.
Cancer Res. 1990, 50, 658. (b) Liotta, L. A.; Steeg, P. S.; Stetler-
Stevenson, J . G. Cancer Metastasis and Angiogenesis: An
Imbalance of Positive and Negative Regulation. Cell 1991, 64,
327. (c) Baricos, W. H.; Zhou, Y.; Mason, R. W.; Barrett, A. J .
Human Kidney Cathepsins B and L. Characterization and
Potential Role in Degradation of Glomerular Basement Mem-
brane. Biochem. J . 1988, 252, 301.
(16) Pepper, J . S.; Montesano, R.; Mandriots, S. J .; Orci, L.; Vassalli,
J . Angiogenesis: A Paradigm for Balanced Extracellular Pro-
teolysis During Cell Migration and Morphogenesis. Enzyme
Protein 1996, 49, 138.
(17) (a) Zou, Z.; Anisowicz, A.; Hendrix, M. J . C.; Thor, A.; Neveu,
M.; Sheng, S.; Rafidi, K.; Seftor, E.; Sager, R. Maspin, a Serpin
with Tumor-Suppressing Activity in Human Mammary Epithe-
lial Cells. Science 1994, 263, 526. (b) Kimura, T.; Fuchimoto,
S.; Iwagaki, H.; Hizuta, A.; Orita, K. Inhibitory Effect of
Nafamostat Mesilate in Metastasis into the Livers of Mice and
on Invasion of the Extracellular Matrix by Cancer Cells. J . Int.
Med. Res. 1992, 20, 343.
(18) Otto, H.-H.; Schirmeister, T. Cysteine Proteases and Their
Inhibitors. Chem. Rev. 1997, 97, 133.
(19) Sanders, C. T.; Seto, C. T. 4-Heterocyclohexanone-Based Inhibi-
tors of the Serine Protease Plasmin. J . Med. Chem. 1999, 42,
2969-2976.
(20) Teno, N.; Wanaka, K.; Okada, Y.; Tsuda, Y.; Okamoto, U.;
Hijikata-Okunomiya, A.; Naito, T.; Okamoto, S. Development
of Active Center-Directed Inhibitors Against Plasmin. Chem.
Pharm. Bull. 1991, 39, 2340.
(21) Mole, J . E.; Horton, H. R. Kinetics of Papain-Catalyzed Hydroly-
sis of R-N-Benzoyl-L-arginine-p-nitroanilide. Biochemistry 1973,
12, 816.
(7) Conroy, J . L.; Sanders, T. C.; Seto, C. T. Using the Electrostatic
Field Effect to Design a New Class of Inhibitors for Cysteine
Proteases. J . Am. Chem. Soc. 1997, 119, 4285.
(8) For reviews of serine and cysteine protease inhibitors, see: (a)
Mehdi, S. Synthetic and Naturally Occurring Protease Inhibitors
Containing an Electrophilic Carbonyl Group. Bioorg. Chem.
1993, 21, 349. (b) Wells, G. J .; Bihovsky, R. Calpain Inhibitors
as Potential Treatment for Stroke and Other Neurodegenerative
Diseases: Recent Trends and Developments. Exp. Opin. Ther.
Pat. 1998, 8, 1707.
(9) Conroy, J . L.; Seto, C. T. Demonstration by 13C NMR Studies
that Tetrahydropyranone-Based Inhibitors Bind to Cysteine
Proteases by Reversible Formation of a Hemithioketal Adduct.
J . Org. Chem. 1998, 63, 2367.
(10) Conroy, J . L.; Abato, P.; Ghosh, M.; Austermuhle, M. I.; Kiefer,
M. R.; Seto, C. T. Synthesis of Cyclohexanone-Based Cathepsin
B Inhibitors that Interact with Both the S and S′ Binding Sites.
Tetrahedron Lett. 1998, 39, 8253.
(11) For other reversible inhibitors of serine and cysteine proteases
that can make contacts with both the S and S′ subsites, see: (a)
Hu, L.-Y.; Abeles, R. H. Inhibition of Cathepsin B and Papain
by Peptidyl R-Keto Esters, R-Keto Amides, R-Diketones, and
R-Keto Acids. Arch. Biochem. Biophys. 1990, 281, 271. (b)
Imperiali, B.; Abeles, R. H. Extended Binding Inhibitors of
Chymotrypsin that Interact with Leaving Group Subsites S1′-
S3′. Biochemistry 1987, 26, 4474. (c) Yamashita, D. S.; Smith,
W. W.; Zhao, B.; J anson, C. A.; Tomaszek, T. A.; Bossard, M. J .;
Levy, M. A.; Oh, H.-J .; Carr, T. J .; Thompson, S. K.; Ijames, C.
F.; Carr, S. A.; McQueney, M.; D′Alessio, K. J .; Amegadzie, B.
Y.; Hanning, C. R.; Abdel-Meguid, S.; DesJ arlais, R. L.; Gleason,
J . G.; Veber, D. F. Structure and Design of Potent and Selective
Cathepsin K Inhibitors. J . Am. Chem. Soc. 1997, 119, 11351.
(12) For several other examples of serine and cysteine protease
inhibitors, see: (a) Dragovich, P. S.; Prins, T. J .; Zhou, R.;
Fuhrman, S. A.; Patick, A. K.; Matthews, D. A.; Ford, C. E.;
Meador, J . W., III; Ferre, R. A.; Worland, S. T. Structure-Based
Design, Synthesis, and Biological Evaluation of Irreversible
(22) Bajkowski, A. S.; Frankfater, A. Specific Spectrophotometric
Assays for Cathepsin B1. Anal. Biochem. 1975, 68, 119.
(23) Hitomi, Y.; Ikari, N.; Fujii, S. Inhibitory Effect of a New
Synthetic Protease Inhibitor (FUT-175) on the Coagulation
System. Haemostasis 1985, 15, 164.
(24) Ohno, K.; Kosaki, G.; Kambayashi, J .; Imaoka, S.; Hirata, F.
FOY: [Ethyl p-(6-guanidinohexanoyloxy)benzoate] Methane-
sulfonate as
a Serine Proteinase Inhibitor. I. Inhibition of
Thrombin and Factor Xa In Vitro. Thrombosis Res. 1980, 19, 579.
J M990272G