The design of potent hydrazones and disulfides as cathepsin S inhibitors

Article abstract of DOI:10.1016/S0968-0896(02)00468-6

The design and synthesis of dipeptidyl disulfides and dipeptidyl benzoylhydrazones as selective inhibitors of the cysteine protease Cathepsin S are described. These inhibitors were expected to form a slowly reversible covalent adduct of the active site cysteine of Cathepsin S. Formation of the initial adduct was confirmed by mass spectral analysis. The nature and mechanism of these adducts was explored. Kinetic analysis of the benzoyl hydrazones indicate that these inhibitors are acting as irreversible inhibitors of Cathepsin S. Additionally, the benzoylhydrazones were shown to be potent inhibitors of Cathepsin S processing of Class II associated invariant peptide both in vitro and in vivo.

Full text of DOI:10.1016/S0968-0896(02)00468-6

Bioorganic & Medicinal Chemistry 11 (2003) 733–740
The Design of Potent Hydrazones and Disulfides as Cathepsin S
Inhibitors
Charles L. Cywin,a,* Raymond A. Firestone,a Daniel W. McNeil,a Christine A. Grygon,b
Kathryn M. Crane,b Della M. White,b Peter R. Kinkade,a Jerry L. Hopkins,b
Walter Davidson,b Mark E. Labadia,b Jessi Wildeson,b Maurice M. Morelock,b
Jeffrey D. Peterson,c Ernest L. Raymond,c Maryanne L. Brownb and Denice M. Speroa
aDepartment of Medicinal Chemistry, Research and Development Center, Boehringer Ingelheim Pharmaceuticals, Inc.,
900 Ridgebury Road, Ridgefield, CT 06877-0368, USA
bDepartment of Biology, Research and Development Center, Boehringer Ingelheim Pharmaceuticals, Inc., 900 Ridgebury Road,
Ridgefield, CT 06877-0368, USA
cDepartment of Pharmacology, Research and Development Center, Boehringer Ingelheim Pharmaceuticals, Inc.,
900 Ridgebury Road, Ridgefield, CT 06877-0368, USA
Received 24 July 2002; revised 6 September 2002; accepted 27 September 2002
Abstract—The design and synthesis of dipeptidyl disulfides and dipeptidyl benzoylhydrazones as selective inhibitors of the cysteine
protease Cathepsin S are described. These inhibitors were expected to form a slowly reversible covalent adduct of the active site
cysteine of Cathepsin S. Formation of the initial adduct was confirmed by mass spectral analysis. The nature and mechanism of
these adducts was explored. Kinetic analysis of the benzoyl hydrazones indicate that these inhibitors are acting as irreversible
inhibitors of Cathepsin S. Additionally, the benzoylhydrazones were shown to be potent inhibitors of Cathepsin S processing of
Class II associated invariant peptide both in vitro and in vivo.
# 2002 Elsevier Science Ltd. All rights reserved.
Introduction
were irreversible and thus concern for the potential
haptenization made the vinyl sulfones less desirable
from a drug development point of view.
Cathepsin S, a cysteine protease, is responsible for the
final proteolytic step of MHC Class II invariant chain
processing, cleaving the precursor fragment (p10) of
invariant chain to CLIP(Class II Invariant chain
peptide).1ꢀ4 The MHC Class II-CLIPcomplex can
associate with another protein, HLA-DM, which facil-
itates the release of CLIPand loading of antigenic pep-
tide. The MHC Class II-peptide complex is transported
to the cell surface where it engages CD4+ T cells and
elicits an immune response. Inhibition of Cathepsin S
should attenuate MHC Class II’s ability to load and
present antigenic peptide and thus have a beneficial
effect in autoimmune diseases. As part of our ongoing
efforts to target autoimmune diseases we set out to
design novel inhibitors of Cathepsin S. Palmer and
Bromme recently reported vinyl sulfones were potent
inhibitors of Cathepsin S.5 However, these inhibitors
*Corresponding author. Tel.:+1-203-791-6401; fax:+1-203-791-5868;
e-mail: ccywin@rdg.boehringer-ingelheim.com
0968-0896/03/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(02)00468-6

734
C. L. Cywin et al. / Bioorg. Med. Chem. 11 (2003) 733–740
In an attempt to prevent the irreversible binding seen
with the vinyl sulfone inhibitors, we investigated repla-
cing the vinyl sulfone moiety with a reactive thiophile
capable of forming an adduct, after addition of the
cysteine thiol group, that could slowly revert to active
enzyme under physiological conditions and thus avoid
Chemistry
The disulfide inhibitors were synthesized as outlined in
Scheme 1. Tosylate 3, available from known N-Boc
homophenylalaninol by the reaction of tosyl chloride in
pyridine, was reacted with thiolacetic acid to introduce
the first sulfur.9ꢀ11 Deacetylation of the intermediate
acetate and disulfide formation with 2,20-dithiodipyr-
idine 4 gives the labile disulfide 5.12 The desired dis-
ulfides, 1 and 7, were then produced by displacement
with the appropriate thiol followed by removal of the
amino protecting group and condensation with known 6
under standard peptide coupling conditions.
potential haptenization. We took
a two prong
approach. First we explored disulfides which can
undergo nucleophilic attack by cysteine. There was pre-
cedent for inactivating a cysteine protease (papain) with
thiolating agents.6 Target compound 1 could form a
disulfide with the Cathepsin S active site cysteine. Inac-
tivation would be expected to be slowly reversible by
cleavage with endogenous sulfides such as glutathione
(Fig. 1). Compound 1 requires nucleophilic attack by
the active site cysteine at the a-sulfur atom. This posi-
tion adjoins the usual site of attack of a protease on a
peptide substrate, but there was precedent for this in
other proteases.7 The second approach involved incor-
poration of a benzoyl hydrazone which could undergo
attack by the cysteine. This would form a thioaminal
adduct which theoretically would be reversible or could
undergo hydrolysis to the aldehyde (Fig. 2).8 The pro-
posed peptidyl hydrazone inhibitors, such as 2, had a
potential advantage over the disulfide inhibitors in that
the right hand section of the molecule could be capable
of binding to the prime side of the active site and unlike
the disulfides would be retained after covalent addition
of the active site cysteine.
Preparation of the hydrazones is outlined in Scheme 2.
Condensation of known Weinreb amide 8 with 6 gave
the new Weinreb amide 9.5 This was converted to alde-
hyde 10 with lithium aluminum hydride in ether which
was then condensed with the appropriate benzoic
hydrazide to give the hydrazone inhibitors. Addition-
ally, hydrazone 2 was reduced with sodium cyanoboro-
hydride to give hydrazide 13 (Scheme 3) which was used
in mechanistic studies vide infra.5c
Biology
The potential inhibitors were screened against human
recombinant Cathepsin S as the primary assay. Elastase
Figure 1.
Figure 2.
Scheme 1.

C. L. Cywin et al. / Bioorg. Med. Chem. 11 (2003) 733–740
735
Scheme 2.
Scheme 3.
(porcine pancreatic elastase), as well as the more closely
related cysteine protease human Cathepsin B were used
as selectivity screens.13 The primary and selectivity
assays were fluorescence based enzymatic assays which
provided reproducible IC50’s within 2-fold. Neither 1
nor 7 showed elastase activity and only disulfide 1
showed measurable Cathepsin S activity using standard
assay conditions (Table 1). Compound 1 however did
show selectivity for Cathepsin S over Cathepsin B. Since
the standard assay conditions employ a reducing agent
to maximize the free active site cysteine it was possible
that the reducing agent could react with the inhibitor. In
fact if DTT is removed from the assay, both compound
1 and 7 showed significantly increased potencies. These
in vitro results with DTT would likely be mirrored in
cellular and in vivo systems by endogenous, ubiquitous
thiol nucleophiles such as glutathione.
Cathepsin B. By contrast ent-2 (the enantiomer of 2),
prepared as outlined in Scheme 2 starting from
enantiomeric starting materials, showed at least a 1000-
fold loss in activity. Further evaluation of hydrazone 2
showed it to be selective in a panel of protease assays
performed by an independent lab with IC50’s>7 mM
observed for Calpain, Angiotensin Converting Enzyme,
Cathepsin G, Collagenase IV, neutral endopeptidase
and neutrophil elastase.14
Interestingly, the aldehyde precursor 10 was also active
at similar potencies to 2. This led to speculation that the
activity of 2 resulted from hydrolysis to the aldehyde
under assay conditions. To test this possibility we
developed HPLC conditions to distinguish 2 and 10 in
standard assay buffer as shown in Figure 3 graph A.
Compound 2 was then assayed for Cathepsin S activity
under standard experimental conditions including IC50
determination. An aliquot of the assay solution was
then analyzed and shown to contain no aldehyde
within the limits of detection of the HPLC assay
(ꢁ1%, graph B).
We next turned our attention to the benzoyl hydrazone
series. Compound 2 showed good potency against
Cathepsin S in vitro as well as good selectivity over
Table 1. In vitro enzyme inhibition
Compounds were also tested in a cellular system. The
assay measured inhibition of Cathepsin S via the accu-
mulation of substrate invariant chain fragment (p10)
and a decrease in product (CLIP) as observed in a
human B cell line (HOM2) after treatment with com-
pound.3 This pulse chase experiment uses 35S methio-
nine, immuno-precipitation, and SDS-PAGE to
determine the increase of the physiologic substrate (p10)
quantified by a StormR phosphor-imager. In this assay
the Minimal Inhibitory Concentration (MIC) of the
drug required to detect an increase in p10 is measured.
Disulfide 1 showed a minimum inhibitory concen-
tration of 1 mM, similar to the IC50 determined with
Compd
Cathepsin S
IC50 (nM)
Cathepsin B
IC50 (nM)
Elastase
1
7
2
4500 (98)a
NAb (15,000)a
87,000
NTc
700
NAb
NAb
NAb
NAb
NAb
NAb
NAb
NAb
11
13,000
7
35
565
10
ent-2
11
12
13
10
NA @ 500 mM
180
720
30,000
640
aIC50’s in assay without the reducing agent DTT.
bNA=not active @ 5 mg/mL.
cNT=not tested.

736
C. L. Cywin et al. / Bioorg. Med. Chem. 11 (2003) 733–740
formation which was expected to be reversible. Cathe-
psin S mediated hydrolysis of the resultant covalent
adduct to give the aldehyde was discounted based on
our HPLC analysis of 2 after in vitro assay. An irrever-
sibility assay was devised in which the compounds were
incubated with Cathepsin S at a 10-fold excess at pH
4.5. The enzyme mixtures were then subjected to reverse
phase HPLC. Cathepsin S and/or Cathepsin S/com-
pound complexes were recovered and analyzed using
electrospray MS. Irreversible inhibitors such as E-64
result in a new HPLC peak with a molecular weight
increase corresponding to addition of the inhibitor.16
Hydrazone 2, 11 and 12 as well as the disulfide inhibi-
tors 1 and 7 showed new HPLC peaks with corre-
sponding increases in molecular weight that were
consistent with the proposed mechanisms in Figures 1
and 2. The hydrazones showed a new HPLC peak with
a molecular weight increase corresponding to the addi-
tion of the entire molecule. In separate experiments,
incubating protein and 2 for 30 min then adding a 10-
fold molar excess of E-64 showed no significant dis-
placement. Although this confirmed that the hydrazone,
and not the aldehyde, was the active species it also
indicated the compound is acting as essentially an irre-
versible inhibitor. The carbonyl of the benzoyl hydra-
zone is an alternative site for attack by the active site
cysteine as shown in Figure 6 which could result in a
more hydrolytically stable covalent adduct. Alter-
natively it might be expected that the loss of water
would result in the acyl enzyme complex instead. How-
ever, both scenarios would be inconsistent with the mass
spectral data that showed water was not lost. To test the
alternative site of attack we assayed compound 13. It
also showed a new HPLC peak with an increase in mass
corresponding to the molecular weight of 13. However,
when a 10-fold excess of E-64 was added there was a
significant displacement, 82% after 36 h, noted by
HPLC with a corresponding change in molecular weight
Figure 3. Samples were injected on a 2.1ꢂ75 mm, 3.5 mm, XDB-C18
column equilibrated in 90% (0.02% phosphoric acid)/10% acetoni-
trile. A solvent gradient was run to 10% (0.02% phosphoric acid)/
90% acetonitrile. Signals were monitored at 254, 280 and 214 nm.
Shown are 254 nm signals. Graph A shows the separation of 2 and 10
under these conditions. Graph B is a sample taken from actual in vitro
Cathepsin S assay after determining IC50. Graphs A and B show only
the region on interest, from 5–6.8 min.
DTT present. The hydrazones 2, 11 and 12 show nearly
complete reduction of CLIPand an accumulation of
p10 at 10 nM as shown in Figure 4. By contrast ent-2
shows a significant CLIPband and no significant accu-
mulation of p10 at 100 nM.
Compound 2 was also evaluated in vivo to determine if
an increase in p10 could be observed in several mouse
strains including C57Bl/6 and Balb/c (I-Ab and d allele
mice).3,15 Compound 2 showed a significant accumula-
tion of p10 at 100 mg/kg while ent-2 showed no activity
at the same dose (Fig. 5).
Although the hydrazone series showed excellent activity
in vitro and in vivo the issue of reversibility of these
inhibitors was still unresolved. As postulated in Figure 2
addition of the active site cysteine should result in acetal
Figure 5. Balb/c male mice were given a single oral dose of 100 mg/kg
compound. Splenocytes were collected following dosing, lysed and
immunoprecipitated. Western blot analysis for the detection of mouse
invariant chain was conducted.
Figure 4. Immunoprecipitation of Class II MHC fragments p10 and
CLIPfrom HOM-2 cells following a 30 min 35S methionine pulse and
4 h chase in the presence of inhibitors at the concentrations shown.

C. L. Cywin et al. / Bioorg. Med. Chem. 11 (2003) 733–740
737
Figure 6.
Table 2. Inhibition kineticsa
of the adduct indicating this compound was reversible.
The decreased activity of 13 relative to 2 shows a pref-
erence for addition at the normal P1 site, but clearly the
MS data indicates that attack at the benzoyl group is
possible.17 Additionally, follow-up stop-flow kinetics
showed that compound 2 did not have a measurable off
rate over a 1.5 h time course (data not shown) confirm-
ing these inhibitors are acting in an irreversible fash-
ion.18 As such, we determined inactivation constants for
the more promising hydrazones and the data are repor-
ted in Table 2. The fact that these inhibitors remained
bound even in denatured protein meant there was still a
possibility that, upon dosing of these inhibitors in vivo,
haptenization could result. Thus, this series was de-
emphasized and the pursuit of new rapidly reversible
inhibitors continued and will be reported elsewhere.19
Compd
Cathepsin S
K inactivation (sꢀ1
Cathepsin S
Ki (mM)
)
2
11
12
0.035
0.055
0.079
0.050
0.033
0.11
aFrom Cat
S enzymatic assay: Substrate Val-Arg-Coumarin:
Km=6.2eꢀ6 M, kcat=2.1eꢀ2.
(GIBCO) supplemented with 10% FBS, penicillin-
streptomycin, and glutamine for 30 min in presence of
Cathepsin S inhibitor prior to pulsing for 30 min with
0.25 mCi 35S methionine (Amersham). Cells were cen-
trifuged and complete RPMI 1640 media plus inhibitor
was added for a 4 h chase.
In vivo p10 Western. Male Balb/c mice were dosed
orally with compound at 100 mg/kg in 30% Cremo-
phor. The mice were sacrificed after 3.5 h and spleno-
cytes were isolated and lysed. Invariant chain bound to
MHC Class II was immunoprecipitated with anti-MHC
Class II antibody. Proteins were separated by electro-
phoresis, transferred to 0.2 mm nitrocellulose and pro-
bed with invariant chain antibody followed by
peroxidase labeled goat anti-rat Ig (Biosource, ari4404).
Blots were developed with ECL Detection Reagents
(Amersham, RPN 2108).
Conclusions
Two new series of Cathepsin S inhibitors have been
described with the goal of identifying novel, selective
and reversible enzyme inhibitors. The benzoyl hydra-
zone series was shown to be a potent and selective inhi-
bitor in enzymatic assays. The hydrazones also
displayed potent cellular and in vivo activity. The
mechanistic aspects of these inhibitors were investigated
and both series displayed behavior consistent with irre-
versible inhibition. Although the possibility that these
inhibitors are reversible but with extremely long off-
rates has not been completely discounted, the fact that
an adduct forms that is not readily reversible or degra-
ded increases the potential risk of haptenization and
therefore further synthetic efforts on these series were
discontinued.
Reversibility assessment. The compounds were prepared
as 5 mg/mL stock solutions in DMSO. Cathepsin S-
hexaHis in 100 mM sodium acetate, pH 4.5 with 2 mM
DTT and 1mM EDTA was diluted to 1.6 mg/mL with
the same buffer containing 10 mM cysteine. Aliquots of
Cathepsin S-hexaHis and each of the compound stocks
were combined in a ratio of 12 to 1. This results in a
final molar ratio for Cathepsin S-hexaHis to compound
of approximately 1:10. The samples were incubated for
30 min at 37 ꢃC. The entire sample was injected into a
Hewlett Packard HP-1090HPLC operated at RT. A
Delta-Pak HPI C4, 300 A, 2.0ꢂ150 mm column (Waters
Associates, Milford, MA) was operated at a flow rate of
0.2 mL/min. A 25 min linear gradient from 10 mM
ammonium acetate, pH 4.5 to 10 mM ammonium ace-
tate in acetonitrile: water (60:40) was employed. Detec-
tion was by UV at 210 nm. Fractions corresponding in
retention time to Cathepsin S-hexaHis (15.2–16.6 min)
were collected and stored at ꢀ80ꢃC until analyzed. Elec-
trospray ionization mass spectrometry was performed on
Experimental
Biology
Assay conditions for the enzymatic assays can be found
in ref 19. The cellular assays were performed as outlined
below.
Immunoprecipitation of Class II from HOM-2 Cells.
Class II MHC was immunoprecipitated as previously
reported3a with the following changes; HOM-2 cells
were preincubated with methionine-free RPMI 1640

738
C. L. Cywin et al. / Bioorg. Med. Chem. 11 (2003) 733–740
a Micromass (Manchester, UK) AutoSpec OATOF
mass spectrometer. The recovered fractions were intro-
duced using a syringe pump operated at 0.85 mL/h. The
magnet was scanned from m/z=3000–1000 at 8 s/dec-
ade and 1000 resolution. Cone voltage was 23 V. On
observation of adducts the voltage was increased to 50
V to determine whether the adduction could be removed
by this addition of additional collision energy. The
multiply charged spectra (+8 to +11) were deconvo-
luted by application of the Micromass Maxent program
from mass 24000 to 26000.
concentrated and purified by chromatography (EtOAc/
Hexane=1/9) to yield 1.338 g (31%) of tosylate 3 which
was used directly. H NMR (DMSO-d6) 7.76 (d, J=8.0
Hz, 2H), 7.47 (d, J=8.0 Hz, 2H), 7.30–7.20 (m, 2H),
7.14–7.10 (m, 3H), 3.90 (d, J=5.0 Hz, 2H), 3.59–3.51
(m, 1H), 2.61–2.50 (m, 2H), 2.41 (s, 3H), 1.63–1.55 (m,
2H), 1.36 (s, 9H); ESMS 420 (MH+).
1
(b) To 10 mL of DMF cooled in an ice bath was added
0.365 g (5.7 mmol) of 87.5% KOH. After the slow
addition of 0.60 mL (8.4 mmol) of thiolacetic acid, the
mixture was stirred for 15 min. The bath was removed
and the mixture was stirred until all the KOH dissolved
giving an orange solution. The ice bath was returned
and a solution of 1.012 g (2.41 mmol) of 3 in 10 mL
DMF was added dropwise. The reaction was allowed to
warm to rt overnight. The mixture was then diluted to
300 mL with CH2Cl2 and 75 mL of 5% aq citric acid
added. The resulting mixture was filtered and the
organic layer of the filtrate was washed with 5% citric
acid (2ꢂ75 mL), satd aq NaHCO3 (3ꢂ75 mL) and satd
aq NaCl (75 mL). The organics were dried over
Na2SO4, filtered and concentrated to afford 0.560 g
(72%) of the thioacetate which was used directly in the
next reaction. 1H NMR (CDCl3) 7.30–7.26 (m, 2H), 7.20–
7.16 (m, 3H), 4.51 (br d, J=7.0 Hz, 1H), 3.80 (br s, 1H),
3.13–3.01 (m, 2H), 2.75–2.61 (m, 2H), 2.35 (s, 3H), 1.87–
1.74 (m, 2H), 1.45 (s, 9H); CIMS(NH3) 268 (M-(t-Bu)).
Substrate kinetics. The assay buffer employed in all
experiments consisted of 100 mM NaOAc, pH 4.5; 2.5
mM EDTA, 1 mM TCEP, 10% glycerol, 2% DMSO
and 0.025% CHAPS. The substrate used in these studies
was 7-amino-4-methyl coumarin, CBZ-Val-Arg, hydro-
chloride salt. Substrate Km and kcat values were deter-
mined using a stopped-flow method. Increase in
fluorescence, reflecting product formation, were recor-
ded as a function of time using a KinTek stopped-flow
instrument model # SF-2001 (KinTek Corp., Austin,
TX). Experiments involved enzyme concentrations of
80, 250 and 750 nM, and substrate concentrations of
2.5, 5, and 10 mM.
Inhibitor kinetics. Inactivation of Cathepsin S occurs in
a 2-step process: formation of an equilibrium species
(EI), followed by an inactivation step in which the inhi-
bitor irreversibly binds to the enzyme (EI*). Inactiva-
tion reactions were initiated by adding Cathepsin S into
a solution containing substrate and inhibitor. Fluores-
cence was measured as a function of time using a spec-
trofluorimeter (instrumentation: SLM Aminco 8100). A
series of progress curves were obtained for each com-
pound (15–260 nM, depending inhibitor potency) at
constant enzyme and substrate concentrations (60 nM
and 10 mM, respectively). Data were analyzed in a
manner similar to the method described by Palmer and
Bromme.5
(c) A solution of 0.560 g of the thioacetate (1.73 mmol)
of 6 in 40 mL of MeOH was degassed by repeatedly
subjecting it to vacuum and nitrogen purge. Addition of
0.480 g (2.18mol) of 2,20-dithiodipyridine 4 was fol-
lowed by addition of a solution of 0.110 g (2.62 mmol)
.
of LiOH H2O in 2 mL water. The reaction was left stir-
ring at rt overnight. MeOH was removed by rotary
evaporation and the residue was extracted with EtOAc.
After evaporation of the solvent, the residue was chro-
matographed eluting with EtOAc/Hexane (1:3) to give
compound 5 (0.195 g, 29%). 1H NMR (CDCl3) 8.53 (d,
J=4.5 Hz, 1H), 7.64–7.55 (m, 2H), 7.29–7.25 (m, 2H),
7.20–7.16 (m, 3H), 7.12–7.09 (m, 1H), 5.55 (br s, 1H),
3.87 (br s, 1H), 3.14–3.04 (m, 2H), 2.74–2.60 (m, 2H),
1.99–1.88 (m, 2H), 1.44 (s, 9H); CIMS(CH4) 391
(MH+).
Chemistry
General chemical methods. Chromatography was per-
formed on silica gel. Melting points are reported in
Celsius (ꢃC) and are uncorrected. 1H NMR spectra were
determined at 400 MHz and were referenced to solvent,
reported in d, with J values reported in Hz. Mass spec-
tra were performed by the department of Analytical
Sciences of Boehringer Ingelheim Pharmaceuticals. Ele-
mental analyses were performed by Quantitative
Technologies, Inc., Whitehouse, N.J.
(d) A solution of 0.094 g (0.241 mmol) of 5 and 26 mL
(0.25 mmol) of thiophenol in 1 mL of CH2Cl2 was stir-
red for 3 days. The product was diluted with 50 mL of
EtOAc and was washed with 3ꢂ15 mL of satd aq
NaHCO3 and 1ꢂ15 mL of satd aq NaCl. The organic
solution was dried over Na2SO4, filtered and con-
centrated to a residue that was filtered through SiO2
eluting 10% ethyl acetate in hexane. This yielded 0.047
Preparation of morpholine-4-carboxylic acid [3-methyl-
1S-(3-phenyl-1S-phenyldisulfanylmethyl-propylcarbamoyl)-
butyl]-amide (1). (a) To a solution of N-Bochomophe-
nylalaninol9 (10.4 mmol) in 50 mL of CH2Cl2 was
added 2.636 g (13.8 mmol) of TsCl and 1.20 mL (14.8
mmol) of pyridine. After stirring at rt overnight, the
reaction mixture was diluted to 300 mL with EtOAc and
washed with 5% aq citric acid (3ꢂ50 mL), satd aq
NaHCO3 (3ꢂ50 mL) and satd aq NaCl (50 mL). The
combined organic layers were dried over Na2SO4, filtered,
1
g of the phenyl disulfide. H NMR (CDCl3) 7.56 (d,
J=7.5 Hz, 2H), 7.37–7.17 (m, 6H), 7.17 (d, J=7.0 Hz,
2H), 4.59 (br s, 1H), 3.91 (br s, 1H), 2.98 (br s, 2H),
2.74–2.58 (m, 2H), 2.00–1.91 (m, 1H), 1.81 (br s, 1H),
1.48 (s, 9H); CIMS(NH3) 390 (MH+). N-deprotection
was accomplished by dissolving the phenyldisulfide in 5
mL of 20% trifluoroacetic acid in CH2Cl2 and stirring
for 45 min. The reaction mixture was concentrated in
vacuo and used as its TFA salt directly in the next

C. L. Cywin et al. / Bioorg. Med. Chem. 11 (2003) 733–740
739
1
reaction. H NMR (CDCl3) 7.46 (d, J=6.5 Hz, 2H),
temperature, the reaction mixture was poured into 100
mL of 5% aq citric acid. This mixture was extracted
with 2ꢂ75 mL of EtOAc and the extract was washed
with 5% aq citric acid (2ꢂ25 mL), satd aq NaHCO3
(3ꢂ25 mL) and satd aq NaCl (25 mL). The organic
solution was dried over Na2SO4, filtered, concentrated
to an oil, and purified by chromatography (5%MeOH/
CH2Cl2) to afford 0.636 g (84%) of product 9, mp 121–
7.37–7.22 (m, 6H), 7.12 (d, J=7.0 Hz, 2H), 3.54 (br s,
1H), 3.07 (br d, J=12.5 Hz, 1H), 2.88 (dd, J=14.6 Hz,
9.0, 1H), 2.78–2.62 (m, 2H), 2.17–2.06 (m, 2H);
CIMS(NH3) 290 (MH+).
(e) A solution of 0.033 g (0.14 mmol) of N-(morpholine-
4-carbonyl)-l-leucine (6) in 1 mL THF was maintained
at ꢀ15 to ꢀ30 ꢃC. Addition of 0.060 mL (0.55 mmol) of
N-methylmorpholine (NMM) followed by 0.020 mL
(0.15 mmol) of isobutyl chloroformate resulted in a
precipitate.5 After 15 min, a solution of the amine from
previous reaction in 1.0 mL of THF was added. After
0.5 h, the reaction mixture was poured into 10 mL of
5% aq citric acid. This mixture was extracted with 50
mL of EtOAc and the extract was washed with 5% aq
citric acid (2ꢂ10 mL), satd aq NaHCO3 (3ꢂ10 mL) and
satd aq NaCl (10 mL). The organic solution was dried
over sodium sulfate, filtered and concentrated to a resi-
due. The residue, after purification by HPLC (Micro-
sorb C18RP, 21.4 mmꢂ30 cm, 10 mL/min, isocratic
85%AcCN/H2O, 254 nm) afforded 0.030 g (24% over 3
1
123; H NMR (270 MHz, CDCl3) 7.30–7.24 (m, 2H),
7.20–7.15 (m, 3H), 6.74 (d, J=8.4 Hz, 1H), 4.95–4.92
(m, 2H), 4.46–4.39 (m, 1H), 3.69–3.62 (m, 4H), 3.62 (s,
3H), 3.39–3.35 (m, 4H), 3.16 (s, 3H), 2.71–2.61 (m, 2H),
2.10–1.88 (m, 2H), 1.71–1.47 (m, 3H), 0.94 (d, J=5.8
Hz, 6H); CIMS (NH3) 449 (MH+). Anal. (C23H36N4O5)
C, H, N.
(b) To a suspension of LiAlH4 in Et2O (10 mL) at
ꢀ30 ꢃC was added a solution of 9 in Et2O/THF (2:1, 6
mL). After warming to ꢀ20 ꢃC over 30 min, the mixture
was quenched with an aq KHSO4 solution and filtered.
The filtrate was diluted with EtOAc (50 mL), washed
with 1 N aq HCl (3ꢂ10 mL), satd aq NaHCO3 (3ꢂ10
mL) and satd aq NaCl (1ꢂ10 mL) then dried over
Na2SO4, filtered, and concentrated in vacuo to afford 10
1
steps) of product 1, mp 94–95; H NMR (CDCl3) 7.52
(d, J=8.5 Hz, 2H), 7.32 (t, J=7.5 Hz, 2H), 7.28–7.22
(m, 2H), 7.18 (t, J=7.5 Hz, 1H), 7.12 (d, J=6.5 Hz,
2H), 6.33 (d, J=8.5 Hz, 1H), 4.83 (d, J=8.0 Hz, 1H),
4.32–4.26 (m, 1H), 4.21–4.12 (m, 1H), 3.69–3.60 (m,
4H), 3.36–3.33 (m, 4H), 2.98–2.90 (m, 2H), 2.65–2.50
(m, 2H), 2.01–1.91 (m, 1H), 1.87–1.78 (m, 1H), 1.74–
1.65 (m, 2H), 1.54–1.46 (m, 1H), 0.96 (d, J=6.5 Hz,
3H), 0.95 (d, J=6.0 Hz, 3H); CIMS(NH3) 516 (MH+).
Anal. (C27H37N3O3S2) C, H, N.
1
as a foam (0.118g, 91%). Mp 58–60; H NMR (CDCl3)
9.51 (s, 1H), 7.31–7.26 (m, 2H), 7.23–7.17 (m, 3H), 7.01
(br s, 1H), 4.97 (s, 1H), 4.49–4.39 (m, 2H), 3.71–3.63 (m,
4H), 3.43–3.37 (m, 4H), 2.69 (t, J=7.8 Hz, 2H), 2.29–
2.20 (m, 1H), 2.00–1.90 (m, 1H), 1.74–1.66 (m, 2H),
1.61–1.55 (m, 1H), 0.97 (t, J=6.4 Hz, 6H); ESMS 390
(MH+). Anal. (C21H31N3O4) C, H, N.
(c) A solution of 0.381 g (0.98 mmole) of 10 and 0.132 g
(0.97 mmole) of benzoic hydrazide was refluxed for 4 h.
The product was concentrated to a foam and chroma-
tographed eluting with a gradient of 0–4% methanol in
CH2Cl2 afforded 0.282 g (57%) of 2. Recrystallization
from 20% IPA in hexane gave 0.157 g (32%) of purified
product, mp 175–176; 1H NMR (DMSO-d6, 80 ꢃC)
11.28 (s, 1H), 7.84 (app d, J=7.5 Hz, 3H), 7.73 (br s,
1H), 7.57–7.53 (m, 1H), 7.47 (t, J=7.5 Hz, 2H), 7.28 (t,
J=7.5 Hz, 2H), 7.22–7.16 (m, 3H), 6.27 (d, J=8.0 Hz,
1H), 4.45–4.39 (m, 1H), 4.27–4.22 (m, 1H), 3.60–3.51
(m, 4H), 3.38–3.28 (m, 4H), 2.73–2.59 (m, 2H), 2.08–
1.92 (m, 2H), 1.74–1.62 (m, 1H), 1.58–1.54 (m, 2H), 0.93
(d, J=6.5 Hz, 3H), 0.90 (d, J=6.5 Hz, 3H); CIMS
(NH3) 508 (MH+). Anal. (C28H37N5O4) C, H, N.
Preparation of morpholine-4-carboxylic acid [1S-(1S-
isopropyldisulfanylmethyl-3-phenyl-propylcarbamoyl)-3-
methyl-butyl]-amide (7). The synthesis of 7 was directly
analogous to 1 starting from 5 (0.101 g, 0.268 mmol)
and isopropyl mercaptan to yield, after final purification
by HPLC (Microsorb C18RP, 21.4 mmꢂ30 cm, 10 mL/
min, isocratic 85% AcCN/H2O, 254 nm, retention time
9.5 min), 0.019g (16% over 3 steps) of 7, mp 121–123;
1H NMR (CDCl3) 7.29–7.25 (m, 2H), 7.20–7.16 (m,
3H), 6.35 (d, J=8.5 Hz, 1H), 4.89 (d, J=8.0 Hz, 1H),
4.34–4.29 (m, 1H), 4.20–4.13 (m, 1H), 3.69–3.60 (m,
4H), 3.38–3.29 (m, 4H), 2.98 (septet, J=7.0 Hz, 1H),
2.90 (d, J=5.5 Hz, 2H), 2.70–2.57 (m, 2H), 2.01–1.92
(m, 1H), 1.89–1.79 (m, 1H), 1.75–1.65 (m, 2H), 1.55–
1.47 (m, 1H), 1.28 (d, J=7.0 Hz, 6H), 0.96 (d, J=6.5
Hz, 3H), 0.95 (d, J=6.0 Hz, 3H); CIMS(NH3) 482
(MH+). Anal. (C24H29N3O3S2) C, H, N.
Synthesis of morpholine-4-carboxylic acid (1S-{1S-[(2-
hydroxy-benzoyl)-hydrazonomethyl]-3-phenyl-propylcarb-
amoyl}-3-methyl-butyl)-amide (11). Synthesis same as
for 2 starting from 10 (201 mg, 0.52 mmol) and
o-hydroxybenzoic hydrazide (79 mg, 0.52 mmol) yield
Synthesis of morpholine-4-carboxylic acid {1S-[1S-
(benzoyl-hydrazonomethyl)-3-phenyl-propylcarbamoyl]-3-
methyl-butyl}-amide (2). (a) 4-Methyl-2-[(morpholine-4-
carbonyl)-amino]-pentanoic acid. A solution of 0.425 g
(1.704 mmol) of N-(morpholine-4-carbonyl)-l-leucine
(6) in 15 mL THF was maintained at ꢀ15 to ꢀ30 ꢃC.
Addition of 0.75 mL (6.82 mmol) of N-methylmorpho-
line (NMM) followed by 0.25 mL (1.93 mmol) of iso-
1
12%; mp 107–111; H NMR (DMSO-d6) 11.86 (s, 1H),
11.56 (s, 1H), 8.12 (d, J=8.1 Hz, 1H), 7.79 (d, J=6.5
Hz, 1H), 7.66 (d, J=4.7 Hz, 1H), 7.39 (t, J=7.4 Hz,
1H), 7.27–7.13 (m, 5H), 6.92–6.87 (m, 2H), 6.50 (d,
J=8.2 Hz, 1H), 4.38–4.34 (m, 1H), 4.20–4.14 (m, 1H),
3.54–3.49 (m, 4H), 3.30–3.24 (m, 4H), 2.69–2.53 (m,
2H), 1.99–1.84 (m, 2H), 1.67–1.51 (m, 2H), 1.47–1.40
(m, 1H), 0.88 (d, J=6.5 Hz, 3H), 0.84 (d, J=6.5 Hz,
butyl chloroformate resulted in formation of
a
precipitate. After 15 min, a solution of the 2S-amino-N-
methoxy-N-methyl-4-phenyl-butyramide5 in 5 mL of
THF was added. After 0.5 h of stirring at the reduced
+
ꢄ
3H); ESMS 524 (MH ). Anal. (C23H36N4O5 0.5 equiv
H2O) C, H, N.

740
C. L. Cywin et al. / Bioorg. Med. Chem. 11 (2003) 733–740
Synthesis of morpholine-4-carboxylic acid (1S-{1S-[(4-
hydroxy-benzoyl)-hydrazonomethyl]-3-phenyl-propylcarb-
amoyl}-3-methyl-butyl)-amide (12). Synthesis same as
for 2 starting from 10 (197 mg, 0.51 mmol) and
p-hydroxybenzoic hydrazide (78 mg, 0.51 mmol) yield
Ploegh, H. L. J. Exp. Med. 1997, 4, 186. (c) Riese, R. J.;
Mitchell, R. N.; Villadangos, J. A.; Shi, G.; Palmer, J. T.;
Karp, E. R.; De Sanctis, G. T.; Ploegh, H. L.; Chapman, H. A.
J. Clin. Invest. 1998, 101, 2351.
4. For a recent review see: Marquis, R. W.; In Doherty, A. M.,
Ed.; Annual Reports in Medicinal Chemistry Vol. 35, Ed.;
Academic: San Diego, 2000; pp 309.
1
44%; mp 152–155; H NMR (DMSO-d6) 11.31 (s, 1H),
10.05 (s, 1H), 8.13 (d, J=8.2 Hz, 1H), 7.73 (d, J=8.7
Hz, 2H), 7.64 (br s, 1H), 7.29–7.16 (m, 5H), 6.81 (d,
J=8.7 Hz, 2H), 6.51 (d, J=8.3 Hz, 1H), 4.39–4.32 (m,
1H), 4.22–4.16 (m, 1H), 3.58–3.49 (m, 4H), 3.37–3.23
(m, 4H), 2.70–2.52 (m, 2H), 2.01–1.85 (m, 2H), 1.69–
1.53 (m, 2H), 1.48–1.41 (m, 1H), 0.90 (d, J=6.6 Hz,
3H), 0.87 (d, J=6.5 Hz, 3H); CIMS (CH4) 524 (MH+).
5. (a) Palmer, J. T.; Rasnick, D.; Klaus, J. L.; Bromme, D. J.
Med. Chem. 1995, 38, 3193. (b) Bromme, D.; Klaus, J. L.;
Okamoto, K.; Rasnick, D.; Palmer, J. T. Biochem. J. 1996,
315, 85. (c) For a review of cysteine protease inhibitors
including the related semicarbazide inhibitors see;Marquis,
R. W. In Inhibition of Cysteine Protease; Doherty, A. M.,
Trainor, G. L., Eds.; Annual Reports in Medicinal Chemistry,
Vol. 35; Academic Press: New York, 2000; pp 309.
6. (a) Johnson, F. A.; Lewis, S. D.; Shafer, J. A. Biochemistry
1981, 20, 44. (b) Smith, D. J.; Maggio, E. T.; Kenyon, G. L.
Biochemistry 1975, 14, 767.
7. Lynas, J.; Walker, B. Biochem. J. 1997, 325, 609.
8. A reversible cruzain inhibitor, 3-Hydroxynaphthalene-2-
carboxylic acid (4-dimethylaminonaphthalen-1-ylmethylene)-
hydrazide, has been reported see Cohen, F. E.; McKerrow, J.
H.; Ring, C. S.; Rosenthal, P. J.; Kenyon, G. L.; Li, Z. Inhi-
bitors of Metazoan Parasite Proteases U.S. Patent 5,610,192,
1997.
9. Sasaki, N. A.; Hashimoto, C.; Potier, P. Tetrahedron Lett.
1987, 28, 6069.
10. Palmer, J. T.; Rasnick, D.; Klaus, J. L. U.S. Patent,
5,776,718,1998
11. Chapman, J. H.; Owen, L. N. J. Chem. Soc 1950, 579.
12. Fournie-Zaluski, M.-C.; Coric, C.; Turcand, S.; Lucas, E.;
Noble, F.; Maldonado, R.; Roques, B. P. J. Med. Chem. 1992,
35, 2473.
13. (a) Cywin, C. L.; Emmanuel, M. J.; Frye, L. L.; Spero, D.
M.; Thomson, D. S.; Ward, Y. U. S. Patent 6,395,897 (b) The
substrate for the cathepsin S and B assays was Z-L-valyl-L-
valyl-L-arginine-(4-methyl-7-coumarinyl)amide while casein-
BODIPY(fluorescein) was used for elastase assays.
14. Performed by MDS Panlabs Pharmacology Services,
11804 North Creek Parkway South, Bothell, WA 98011-8890.
The results from this testing are as follows; human Calpain
(IC50=7.6 mM), rabbit Angiotensin Converting Enzyme (5%
inhibition @ 10 mM), human Cathepsin G (6% inhibition @
30 mM), human Collagenase IV (ꢀ8% inhibition @ 100 mM),
human neutral endopeptidase (0% inhibition @ 10 mM) and
human neutrophil elastase (2% inhibition @ 30 mM).
15. Complete details of this method will be reported else-
where.
16. E-64 is N-(trans-epoxysuccinyl)-L-leucine 4-guanidinobu-
tylamide from Sigma (Catalog# E 3132).
17. It should be noted that compounds lacking the additional
covalent binding domain of these inhibitors such as
MuLeuN(CH2)3Ph were inactive.
18. Morelock, M. M.; Pargellis, C. A.; Graham, E. T.;
Lamarre, D.; Jung, G. J. Med. Chem. 1995, 38, 1751.
19. Ward, Y. D.; Thomson, D. S.; Frye, L. L.; Cywin, C. L.;
Morwick, T. M.; Emmanuel, M. J.; Zindell, R.; McNeil, D.;
Bekkali, Y.; Giradot, M.; Hrapchak, M.; DeTuri, M.; Crane,
K.; White, D.; Pav, S.; Wang, Y.; Hao, M.-H.; Grygon, C. A.;
Labadia, M. E.; Freeman, D.; Davidson, W.; Hopkins J.;
Brown, M. L.; Spero, D. M. J. Med. Chem. in press.
ꢄ
Anal. (C23H36N4O5 0.5 equiv H2O) C, H, N.
Synthesis of morpholine-4-carboxylic acid {1-[1S-(N-
benzoyl-hydrazinomethyl)-3-phenyl-propylcarbamoyl]-3-
methyl-butyl}-amide (13). Compound 2 (26 mg, 0.051
mmol) was dissolved in EtOH (1 mL) and treated with
NaCNBH3 (7 mg, 0.11 mmol) and the mixture was
stirred at rt for 2 h. The mixture was diluted with
EtOAc and washed with satd aq NaHCO3 (3ꢂ), the
organic layer was dried over Na2SO4 and concentrated
and the residue was purified by HPLC (Microsorb
C18RP, 21.4 mmꢂ30 cm, 10 mL/min, isocratic
70%AcCN/H2O for 10 min then gradient to 95%
AcCN/H2O over 5 min, 254 nm, rentention time 11.29
min), HPLC purity >99%, mp 72–74; 1H NMR
(CDCl3) 8.79 (s, 1H), 7.82 (d, J=7.0 Hz, 2H), 7.52–7.49
(m, 1H), 7.45–7.42 (m, 2H), 7.26–7.23 (m, 2H), 7.19–
7.13 (m, 3H), 6.41(d, J=8.0 Hz, 1H), 4.78 (d, J=7.5
Hz, 1H), 4.37–4.30 (m, 1H), 4.11–4.06 (m, 1H), 3.69–
3.61 (m, 4H), 3.39–3.30 (m, 4H), 3.13 (dd, J=12.5 Hz,
4.0, 1H), 2.72–2.60 (m, 3H), 1.84–1.69 (m, 5H), 0.97 (d,
J=6.5 Hz, 3H), 0.95 (d, J=6.5 Hz, 3H); CIMS (NH3)
510 (MH+); HRMS (MH+, C28H40N5O4) calcd
510.3080 found 510.3078.
Acknowledgements
Lana Smith for the synthesis of the ent-2, and David
Thomson for useful discussions during the preparation
of this manuscript.
References and Notes
1. (a) Villadangos, J. A.; Ploegh, H. L. Immunity 2000, 12,
233. (b) Amigorena, S.; Webster, P.; Drake, J.; Newcomb, J.;
Cresswell, P.; Mellman, I. J. Exp. Med. 1995, 181, 1729.
2. (a) Barrett, A. J.; Rawlings, N. D. Perspectives in Drug
Discovery and Design 1996, 6, 1. (b) Chapman, H. A.; Reise,
R. J.; Shi, G.-P. Ann. Rev. Physiol. 1997, 59, 63.
3. (a) Riese, R. J.; Wolf, P. R.; Bromme, D.; Natkin, L. R.;
Villadangos, J. A.; Ploegh, H. L.; Chapman, H. A. 1996;4:357
(b) Villadangos, J. A.; Riese, R. J.; Peters, C.; Chapman, H. A.;