Novel β-amino acid derivatives as inhibitors of cathepsin A

Article abstract of DOI:10.1021/jm300663n

Cathepsin A (CatA) is a serine carboxypeptidase distributed between lysosomes, cell membrane, and extracellular space. Several peptide hormones including bradykinin and angiotensin I have been described as substrates. Therefore, the inhibition of CatA has

Full text of DOI:10.1021/jm300663n

Article
pubs.acs.org/jmc
Novel β‑Amino Acid Derivatives as Inhibitors of Cathepsin A
Sven Ruf,*,† Christian Buning,† Herman Schreuder,† Georg Horstick,†,§ Wolfgang Linz,† Thomas Olpp,†
Josef Pernerstorfer,† Katrin Hiss,† Katja Kroll,† Aimo Kannt,† Markus Kohlmann,† Dominik Linz,‡
Thomas Hubschle,† Hartmut Rutten,† Klaus Wirth,† Thorsten Schmidt,† and Thorsten Sadowski†
̈
̈
†Sanofi-Aventis Deutschland GmbH, Industriepark Hochst, 65926 Frankfurt, Germany
̈
‡Universitatsklinikum des Saarlandes, Homburg/Saar, Germany
̈
S
* Supporting Information
ABSTRACT: Cathepsin A (CatA) is a serine carboxypepti-
dase distributed between lysosomes, cell membrane, and
extracellular space. Several peptide hormones including
bradykinin and angiotensin I have been described as substrates.
Therefore, the inhibition of CatA has the potential for
beneficial effects in cardiovascular diseases. Pharmacological
inhibition of CatA by the natural product ebelactone B
increased renal bradykinin levels and prevented the develop-
ment of salt-induced hypertension. However, so far no small
molecule inhibitors of CatA with oral bioavailability have been
described to allow further pharmacological profiling. In our work we identified novel β-amino acid derivatives as inhibitors of
CatA after a HTS analysis based on a project adapted fragment approach. The new inhibitors showed beneficial ADME and
pharmacokinetic profiles, and their binding modes were established by X-ray crystallography. Further investigations led to the
identification of a hitherto unknown pathophysiological role of CatA in cardiac hypertrophy. One of our inhibitors is currently
undergoing phase I clinical trials.
tension.6 Similar effects were observed by antisense oligonucleo-
tides suppressing the expression of CatA.7 In addition to peptide
hormones, the lysosomal receptor protein Lamp2a has also been
identified as a substrate of this protease, linking CatA activity to
autophagic processes.8
Since locally increased bradykinin levels have been reported to
play a beneficial role in cardiac diseases,9 we developed interest
in the inhibition of CatA as a potential therapeutic intervention
in this area. Contrary to established targets in the cardiac renin−
angiotensin system such as ACE and NEP,10 CatA is only weakly
expressed in the vasculature and has an acidic pH optimum of
activity. Both properties possibly exclude this enzyme from
systemic peptide hormone turnover and suggest a role for CatA
in the regulation of the local kallikrein−kinin and renin−
angiotensin system.
INTRODUCTION
■
CatA is a lysosomal serine protease, which is also found in the
extracellular space. The enzyme is a member of the α/β
hydrolase fold family and shares structural homology with yeast
carboxypeptidase Y.1 Further research has established CatA as a
multicatalytic enzyme with additional deamidase and esterase
activities.2 Distinct from the catalytic functions of CatA is its
protective function in the lysosome: CatA stabilizes the high
molecular weight enzyme complex made up of β-galactosidase
and neuraminidase, and deficiencies of this complex lead to the
severe human disease galactosialidosis.3 A recent study in mice
carrying a catalytically inactive CatA mutant demonstrated that
CatA’s protective function is distinct from its carboxypeptidase
activity.4
Several peptide hormones have been described as substrates
for CatA in vitro such as bradykinin, angiotensin I, and
oxytocin.5 Pharmacological inhibition of CatA by the natural
product ebelactone B (Figure 1) increased renal bradykinin
levels and prevented the development of salt-induced hyper-
In this article we describe the discovery of novel CatA
inhibitors, which allowed us to identify a hitherto unknown
pathophysiological role of this protease in cardiac hypertrophy
and atrial arrhythmogenesis.
CHEMISTRY
■
Our CatA inhibitors are easily recognized as products of amide
coupling reactions starting from β-amino acids 4 and the
pyrazole heterocycles 3 (Scheme 1).
Received: May 11, 2012
Published: August 3, 2012
Figure 1. Structure of ebelactone B.
© 2012 American Chemical Society
7636
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649

Journal of Medicinal Chemistry
Article
a
Scheme 1. General Synthetic Access to Cathepsin A Inhibitors 2 and 8
a
Reagents and conditions: (a) H2SO4, rt, 4 h; (b) HCl, MeOH; (c) THF, NEt3; (d) R1X, (X = Br, Cl), Cs2CO3, THF; (e) NaOH, MeOH; (f) β-AS,
TOTU, NEM, DMF, rt; (g) R4= Cl: SO2Cl2, HOAc, R4 = F: CH3CN, Selectfluor, (h) LiOH, dioxane, water.
Synthetic pathways to β-amino acids have been reviewed
elsewhere. A significant number of them are now commercially
available.11
The pyrazole heterocycle 3 was accessible by addition of the
corresponding hydrazine derivative 6 to oxalacetic acid 5 in
diluted H2SO4. The condensation products 7 were easily isolated
by filtration from the reaction mixture. An amide coupling of the
intermediate 7 with β-amino acids using TOTU12 as activating
reagent delivered the inhibitors 8.
ylate 9. Further functionalization of the hydroxy group was then
accomplished by alkylation of the methyl esters 11 with alkyl
halogenides in acetone with Cs2CO3 as base. Under these
reaction conditions a selective O-alkylation of the pyrazolone
precursors 11 was observed.
Introduction of a halogen in the 4-position of the pyrazolone
core was accomplished with sulforyl chloride or Selectofluor
applied on the 5-methoxy-1-(2-fluorophenyl)pyrazol-3-carbox-
ylic acid methyl ester.
In an alternative approach methyl esters 11 of the
pyrazolonecarboxylic acids were synthesized by the addition of
the corresponding hydrazines 10 to dimethyl acetylenedicarbox-
Introduction of a hydroxyl group in the 5-alkoxy-substituent
resulted from the reduction of 13 with NaBH4 in methanol
(Scheme 2). 13 was synthesized according to Scheme 1 with
7637
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649

Journal of Medicinal Chemistry
Article
a
Scheme 2. Synthesis of Cathepsin A Inhibitors 13−16
a
13−16a: R1 = CH3, R2 = H. 13−16b: R1 = R2 = Cl. Reagents and conditions: (a) NaBH4, MeOH; (b) chiral chromatography.
Figure 2. Identification of preferred chemical motifs (virtual fragments) from HTS hits (blue and purple circles) and subsequent combinations by H2L
chemistry.
bromopinacolone used as alkyl bromide. The mixture of
diastereomers 14 obtained after reduction was separated by
chiral chromatography, yielding pure stereoisomers 15 and 16.
The configuration of the chiral alcohol carbon atom was
arbitrarily assigned R to the isomer with the shortest retention
time on the chiral column.
we identified certain chemical subunits present in several hit
clusters. Chemical building blocks were selected with high
similarity to the virtual fragments for suitable combinations by
parallel chemistry.
The chosen procedure and project progression is shown in
Figure 2. The hit compounds 20 and 21 revealed that aliphatic
chains with a carboxylic acid terminus and the pyrazole scaffold
were preferred chemical motifs for CatA inhibitors. A first series
of parallel syntheses led to the identification of the rescaffolded
hit 8q, which had lost activity by a factor of 10 compared to the
starting point 20, but because of the reduced molecular weight,
the ligand efficiency was in the same range. From a chemical
perspective 8q was a very attractive starting point for us. We
initiated X-ray crystallographic studies in order to obtain
information of the binding mode of 8q and started parallel
synthesis activities focusing on variations of the phenyl residues
in the β-amino acid and pyrazole parts of the molecule.
By starting from 8q and applying the strategy described above,
we were able to increase the activity by a factor of 100 and to
identify (S)-β-amino acids as the preferred enantiomeric form
for the construction of CatA inhibitors (see derivatives 8s and 8t
in Table 1).
1
The H NMR spectra of 15/16a and of 15/16b were very
similar. However, these compounds could be clearly distin-
guished by their different retention times on an LCMS system
with an elongated gradient method on an achiral column.
IDENTIFICATION OF THE LEAD STRUCTURE 8A
■
With no small molecule inhibitors of CatA besides ebelactone B
described in the literature we initiated a HTS of our compound
collection, which was successful in identifying several hit series
with high activity (<0.5 μM) on CatA confirming the
druggability of CatA with small molecules. Unfortunately the
most active hits series from the HTS turned out to be of only
limited use for the following reasons: (A) toxicological issues of
hits known from other projects; (B) observed proteolytic
cleavage during crystallization; (C) high MW combined with
high lipophilicity and low ligand efficiency (not leadlike).13
Therefore, on the basis of our HTS results, we initiated
explorative hit to lead chemistry in order to identify a suitable
lead structure. Working in this early phase without reliable SAR
patterns, we were looking for guiding principles to direct our
chemical activities. By detailed analysis of the HTS hit structures,
As can be seen from Table 1, the highest inhibitory activity in
this series was observed with 8b featuring a biphenyl substituent
in the β-amino acid residue.
The introduction of an oxygen linker between the two phenyl
groups (8j) led to somewhat reduced inhibitory activity. By a
7638
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649

Journal of Medicinal Chemistry
Article
Table 1. Biological Data of Compounds 2, 8, 13,15, 16
compd
R1
R2
R3
R4
configuration of β-amino acid
IC50 (μM)
2a
-OCH3
2-F
2-F
2-F
2-F
2-F
H
2-Me
2-Me
2-Me
2-Me
2-Me
2-Me
2-Me
2-Me
H
S
0.026
0.015
0.015
0.03
2b
2c
-OC2H5
H
S
-OCH2CH(CH2)2
H
S
2d
2e
2f
-OCH3
F
S
-OCH3
Cl
-OH
-OMe
H
S
0.110
0.110
0.100
0.038
0.025
0.057
0.059
0.076
0.100
0.087
0.120
0.102
0.200
0.226
0.252
0.274
0.357
0.375
0.803
5.38
H
S
2g
8a
H
H
S
OH
2-F
2-F
2-Cl
3-F
2-F
H
S
8b
8c
OH
4-(4-F-phenyl)
2-Me
H
racemate
OH
H
S
8d
8e
8f
OH
2-Me
H
S
OH
2-methyl-5-fluoro
2,3-dichloro
2-methyl
2,3-dichloro
2-methyl
4-phenoxy
H
H
racemate
OH
H
racemate
8g
8h
8i
OH
4-F
H
S
OH
2,4-difluoro
H
racemate
OH
3-Cl
H
S
8j
OH
H
H
racemate
8k
8l
OH
2-F
H
racemate
OH
2-F
3-Me
H
S
8m
8n
8o
8p
8q
8r
OH
2-F
4-Me
H
S
OH
2-F
2-OMe
H
S
OH
2-F
2-CF3
H
S
OH
2,5-dimethyl
4-phenyl
4-F
H
racemate
OH
2,5-dimethyl
H
racemate
OH
4-Cl
H
2,5-dichloro
2-Me
H
racemate
>30
8s
OH
H
S
R
S
S
S
S
S
S
0.070
7.8
8t
OH
H
2-Me
H
13a
13b
15a
15b
16a
16b
(R)-OCH2-C(O)tBu
(R)-OCH2-C(O)tBu
(R)-OCH2-CH(OH)tBu
(R)-OCH2-CH(OH)tBu
(S)-OCH2-CH(OH)tBu
(S)-OCH2-CH(OH)tBu
2-F
2-F
2-F
2-F
2-F
2-F
2-Me
H
0.003
0.06
2,4-dichloro
2-Me
H
H
0.005
0.050
0.01
2,4-dichloro
2-Me
H
H
2,4-dichloro
H
0.07
switch to simple phenyl residues on the β-amino acid scaffold,
the ortho-tolyl derivative 8a showed a dramatic improvement in
inhibitory activity compared to the unsubstituted one 8k.
Putting the methyl group in meta or para position resulted in a
drastic reduction in activity as seen with examples 8l,m. The
same observation was made after replacement of the ortho
methyl group by sterically more demanding or polar groups, e.g.,
in 8n or 8o. Also an introduction of additional substituents did
not lead to an affinity improvement as demonstrated by 8e,g,h.
Looking at the phenyl residue bound to the pyrazole heterocyclic
core, we determined that the ortho-fluoro derivative 8a was the
most promising. It was slightly more active than the ortho-chloro
and meta-fluoro analogues 8c and 8d and also more active than
the unsubstituted derivative 8s. In good agreement with this
observation is the further decreased biological activity of the m-
Cl derivative 8i or the p-F derivative 8g. As observed in the
phenyl residue on the β-amino acid, the introduction of a second
substituent on this N-phenyl ring of the pyrazol was not
beneficial in terms of activity. That was especially the case for
two substituents in the 2- and 5- position of the phenyl ring, as
can be seen in 8p and 8q. On the basis of these data, we chose 8a
for a full lead profile, and our obtained data are presented in
Table 2. Typical for this series of compounds was the low
metabolic lability in human microsomes and the absence of
CYP3A4 inhibition. We observed no undesired activities on the
human hERG channel and no hints for genotoxicity in the
micronucleus and for mutagenicity in the AMES II test system.
In addition, this compound had low log D and PSA values and a
high solubility. The downside of this compound, however, was
the limited cell permeability as measured in the CACO-214 assay.
Regarding selectivity, 8a showed submicromolar activity on
neutral endopeptidase (NEP).
7639
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649

Journal of Medicinal Chemistry
Article
Table 2. ADME and Safety Data of Lead 8a
Figure 3. (a) X-ray structure of lead 8a showing important hydrogen bonding interactions with the CatA enzyme. The top panel shows the solvent
accessible surface of CatA, while the lower panel shows the atomic model. (b) Two views of the Fo − Fc omit map of 15a in CatA as well as the atomic
model. The contour level is 3σ. Superimposed are the two conformations of 15a we fitted (see text).
7640
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649

Journal of Medicinal Chemistry
Article
CRYSTAL STRUCTURES OF 8A AND 15A BOUND TO
Table 4. Interactions between 15a and Cathepsin A
■
CATA15
atom 1
atom 2
distance (Å)
type
The crystal structure of the CatA precursor (known as human
protective protein precursor) was solved by Rudenko et al. in
1995.16,17 This precursor consists of a single chain of 542 amino
acids. However, the precursor is inactive because the active site is
blocked by the excision peptide. To obtain the active form of the
enzyme, the 54 kDa CatA precursor was proteolytically
transformed into the active mature form by cleavage of a 2
kDa excision peptide with trypsin-Sepharose followed by gel
filtration. The activated protein was crystallized, and the initial
structure was solved using the structure of the precursor as a
search model. Details of the active crystal structure will be
published elsewhere. The mature enzyme consists of two
polypeptides, which are linked together by disulfide bonds to
form a monomer. The catalytic triad of cathepsin A is located in
the 32 kDa chain and consists of the amino acids Ser150, His429,
and Asp372.
a
F7A
CA Pro301
O Pro58
3.4
van der Waals
van der Waals
van der Waals
van der Waals
van der Waals
van der Waals
H bond
a
F7B
2.7
ND2 Asn339
SG Cys60
2.7
N9A/B
C11A
C12A
O13A
O16A
3.4/3.3
3.1
O Gly57
O Gly57
3.5
ND2 Asn339
OD1 Asn339
ND2 Asn339
NH2 Arg334
OH Tyr247
OE2 Glu149
ND2 Asn55
N Gly56
3.4
3.4
H bond
3.2
H bond
2.8
H bond
O22A/B
C26A/B
O27A/B
3.3/3.4
3.4/3.5
2.8/2.8
2.8/2.8
2.8/2.9
3.2/3.1
3.4
H bond
van der Waals
H bond
H bond
b
O28A/B
OE2 Glu149
NE2 His429
CE Met430
CE Met430
OD1 Asp64
H bond
b
ionic
The crystal structure of 8a in complex with the CatA protein
was solved at a resolution of 2.17 Å (Figure 3a, compound−
protein interactions listed in Table 3) and is the first published
C30A
van der Waals
van der Waals
van der Waals
C31A/B
C32A
3.2/3.5
3.3
a
The inhibitor has been fitted in two conformations: A and B.
Depending on the protonation state of the residues involved.
b
Table 3. Interactions between 8a and Cathepsin A
atom 1
F28
atom 2
distance (Å)
type
However, if we look at Figure 3b, lower panel, we see in
conformation A the hydroxyl making decent hydrogen bonds
with the side chains of Asn339 and Arg344 whereas in
conformation B the lipopilic part of the 3,3-dimethylbutane-2-
ol substituent makes van der Waals interactions with the side
chain of Phe336. The increased potency of 2c (IC50 = 15 nM)
with a lipophilic cyclopropylmethoxy residue in the 5-position of
the pyrazole gives an additional confirmation for the beneficial
influence of van der Waals interactions in this region of the
enzyme on ligand activity. In the case of 15a we therefore assume
that both interactions identified above will contribute positively
to the binding affinity. These observed binding interactions have
the additional advantage that the loss of entropy typically
accompanied upon ligand binding is reduced because the
fixation of the flexible 3,3-dimethylbutane-2-ol-side chain of 15a
to a single conformation is avoided. The overall result from these
combination of ligand binding descriptors is the observed
significantly stronger binding.
N Pro301
3.4
3.3
3.4
3.3
2.7
3.2
2.8
2.8
van der Waals
van der Waals
van der Waals
H bond
CA Pro301
SG Cys60
N2
O15
O26
OH Tyr247
OE2 Glu149
NE2 His429
ND2 Asn55
N Gly56
a
H bond
a
ionic
O27
H bond
H bond
a
Depending on the protonation state of the residues involved.
structure of the active CatA enzyme with a small molecule ligand.
The carboxylate group of 8a is accommodated in the carboxylate
recognition site making hydrogen bonds with the OE2 of
Glu149, the NE2 of His429, the NH2 of Asn55, and the main-
chain N of Gly56. The ortho-methylbenzene substituent in the
β-amino acid part of the inhibitor is located in the S1′ pocket.
This S1′ pocket is part of a deep lipophilic tunnel within the
enzyme. Finally, the carbonyl oxygen of the central amide group
makes a hydrogen bond with the OH of Tyr247 and the hydroxy
group in the 5-position of the pyrazole ring points toward the
solvent and does not interact with the protein.
The crystal structure of 15a in complex with the CatA protein
was solved at a resolution of 2.04 Å (Figure 3b, compound−
protein interactions listed in Table 4).
The binding mode of the common core of 15a and 8a is
analogous. The 3,3-dimethylbutane-2-ol substituent has weak
electron density and is very likely disordered.
We modeled two orientations: one with the OH group
interacting with the side chains of Asn339 and Arg344 and one
where this group faces the solvent. In the latter case the tertiary
butyl group has the potential to undergo van der Waals
interactions with the side chain of Phe336. Also the ortho-
fluorobenzene group is present in two orientations: one
orientation where the fluoro group interacts with the main-
chain O of Pro59 and the OD1 of Asp339 and one orientation
where this fluoro group faces the solvent.
LEAD OPTIMIZATION
■
The lead compound 8a already has a remarkably clean overall in
vitro ADME and safety profile (see Table 2). However an
obvious parameter for improvement is the low cell permeability
of 8a, which results in a low oral bioavailability of below <5%. We
also aimed for a further improvement of the target affinity.
Our strategies to overcome the low cell permeability of our
lead 8a targeted the low log D of −0.16. From a chemical point
of view the free carboxylate group and the pyrazolone core are
ideal starting points for the introduction of lipophilicity by
suitable structural modifications.
A classical approach to overcome low permeability in the case
of aliphatic carboxylic acids is to use prodrugs.18 The conversion
of the acid into an ester drastically reduces the polarity and in
many cases allows the compound to cross cellular membranes. In
the case of 8a the use of an ester prodrug would result in the
formation of a CatA inhibitor with low cell permeability in the
plasma. However, the CatA expression in the systemic
circulation is rather low and the low cell permeability of the
From these crystal structures, it is not immediately clear why
15a is much more potent than 8a (IC50 of 5 nM vs 38 nM).
7641
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649

Journal of Medicinal Chemistry
Article
Figure 4. Dependency of cell permeability (Caco) and CYP3A4 induction on 5-substituents on the pyrazole heterocycle. Green values are favorable,
and red values are disfavorable.
metabolized prodrug could prevent our inhibitor from entering
targeted tissues with a significant CatA expression. For these
reasons we saw a serious risk that the use of ester prodrugs would
limit our pharmacological evaluation of this novel target and
discarded this approach.
the GLIDE docking software to design a suitable lipophilic
linker20 (see Figure 5b).
Consequently, we synthesized and tested alkyl chains having
either hydroxyl group or alternatively a carbonyl functional
group to form hydrogen bonds with the proposed amino acids.
By doing so, we identified the two compound pairs 13a,b and
15a,b. Both pairs of compounds have cell permeabilities of >20
× 10−7 cm/s, and derivatives with R3 = Me (see Table 1) have a
significantly improved affinity compared to the initial lead
compound 8a. As important has been the observation that both
15a,b show no significant CYP3A4 induction.
As predicted by our rational design approach, 13a and 15a
show the highest in vitro biological activity on CatA measured in
our series. 13b and 15b with a less preferred phenyl substitution
pattern in the β-amino acid part of the molecule are comparable
to 2a regarding the biological activity.
Interestingly, the configuration on the chiral alcohol carbon
atom has only a minor influence on the in vitro CatA activity,
with the stereoisomer 16 having lost approximately a factor of 2
compared to 15.
In addressing the 4-position of the pyrazolone core, we
investigated halogen and oxygen substituents. While fluorine was
well tolerated in terms of activity (2d), compounds with chlorine
(2e), hydroxyl (2f), and methoxy substituent (2g) lost activity
by a factor of 3 compared to 2a (Table 1).
Our novel CatA inhibitors 2a, 15a, 15b show high selectivity
toward other targets. From an array of 10 proteases tested we
only found micromolar activity on NEP for 2a, which was no
longer present in 15a or 15b. As shown in Figure 6 the NEP
activity in our lead series is limited to a small substructural class
featuring a free hydroxyl group on the pyrazole core in
combination with chlorinated β-phenylalanine residues. In
general the introduction of alkoxy residues into the 5-position
of the pyrazole resulted in IC50(NEP) > 1 μM.
Compounds 2a, 15a, and 15b represent promising combina-
tions of high cell permeability and acceptable CYP3A4 induction
profile. We also found no hints for CYP inhibition or hERG
channel interactions, and the compounds were clean in AMES
and MNT testing. A broad study of 2a and 15a on the
commercial CEREP profile21 of a diverse set of 100 ion channels,
We therefore focused our investigations on the 4- and 5-
position of the pyrazolone for the introduction of structural
modifications. First, we introduced lipophilic extensions at the
oxygen in the 5-position. Figure 5 indicates the general trends we
observed regarding target affinity and permeability. Both
permeability and target affinity improve when substituting the
pyrazolone with either ethyl or methylcyclopropyl, resulting in
perfectly permeable and highly active compounds. Compounds
2b,c have an IC50 of 15 nM combined with a high Caco
classification as an indicator for good cell permeability. However,
the improvement in biological activity and cell permeability was
correlated with induction of CYP3A4, a strong indicator of
potential later issues with drug−drug interactions as demon-
strated with derivatives (2b,c) (see Figure 4).
Since the lead compound does not show any CYP3A4
induction, we concluded that the introduced lipophilicity in the
5-position of the pyrazolone had to be counterbalanced with
some polarity in this part of the molecule.
Given the X-ray structural information of the lead compound
8a, we extensively used structure-based design technologies to
identify the most promising polar substituent in the 5-position of
the pyrazolone. Initially, we analyzed the CatA binding pocket in
detail using the knowledge-based approach Superstar.19 Super-
star proposes hot spots for the placement of chemical
functionalities in the active site based on a thorough statistical
analysis of published X-ray data. We used the alcohol −OH and
the methyl probe to identify promising regions for additional
polarity combined with a lipophilic linker (see Figure 5a). The
regions as detected by SuperStar correspond to a moderately
increased probability of binding of either lipophilic or alcoholic
groups as indicated by the propensity level. From this analysis we
concluded that the extension of the pyrazolone oxygen by a
lipophilic linker is favored including a terminal hydroxyl group
addressing hydrogen bonding interactions to ASN339 and/or
ARG344. Extensive docking studies have been performed using
7642
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649

Journal of Medicinal Chemistry
Article
enzymes, and receptors revealed no hits above 70% inhibition at
10 μM.
For a further differentiation between the compounds 2a, 15a,
and 15b we initiated pharmacokinetic studies in rat, and the
results are presented in Table 5. Cmax and AUC values of 2a in
Table 5. Pharmacokinetic Data after po Administration of 10
mg/kg 2a, 15a,b in 25% Glycofurol/Cremophor (75/25)
Solution to Male Sprague Dawley Rats
2a
15a
15b
plasma
heart
plasma heart plasma heart
C
max (ng/mL)
22200
3720
1360
5300
426
2670
9300
872
AUC(0−inf)
130000
16000
1600
2600
(ng·h/mL)
the plasma are considerably higher compared to those of 15a and
15b. In addition the overall tissue distribution profile is similar in
all three compounds with the highest tissue exposure in the heart
observed with 2a.
BIOLOGICAL ACTIVITY: IN
VIVO/PHARMACOLOGICAL RESULTS
■
To test whether CatA inhibition would increase urinary
bradykinin, we collected urine from rats over a 6 h period
before and after treatment with compounds 2a, 15a, 15b at 30
mg/kg po. Bradykinin was then determined by a specific RIA.
Figure 7a shows that although basal urinary bradykinin levels
varied between the three treatment groups, all compounds
increased the amount of bradykinin by 2- to nearly 3-fold. This
effect seemed not to be mediated by compound-induced
alterations of glomerular filtration rate, as creatinine excretion
was not changed by compound treatment (data not shown). In
our opinion different excretion rates of parent compound into
the distal tubule of the kidney might explain the observed
differences in diuretic activity.
Bradykinin is known for its diuretic effects, which are
mediated by increased renal blood flow and inhibition of
sodium reabsorption in the collecting duct.22 To determine
whether the increased amount of urinary bradykinin observed
under CatA inhibition translated into enhanced diuresis,
anesthetized rats received bolus injection of compounds 2a,
15a, 15b at 1 mg/kg iv and urine was collected through cannulas
inserted into the ureters in five succeeding time periods of 20
min. Albumin in physiological saline was infused throughout the
experimental period to ensure urine flow. In vehicle-treated
Figure 5. (a) Superstar propensity plot of cathepsin A binding pocket
and compound 8a as determined by X-ray analysis (red, OH probe,
propensity level 1.8; green, methyl probe, propensity level 1.6). Atom
colors for compound 8a are as follows: carbon, gray,; oxygen, red;
nitrogen, blue; fluorine, green. Atom colors for CatA binding site are as
follows: carbon, green; oxygen, red; nitrogen, blue; sulfur, yellow. The
black arrow is the extension vector that is directed to the preferred area
for a hydroxyl group marked by the red circle originating from the
oxygen in the 5-position of the pyrazolone. Additionally, positions of
relevant amino acids are indicated. (b) Cathepsin A binding pocket
including compound 8a and docking poses from the structure-based
design cycle. Atom colors are as follows: carbon, gray; oxygen, red;
nitrogen, blue; fluorine, green. In cyan the solvent accessible surface of
the binding site is shown.
Figure 6. IC50 on CatA and NEP of several compounds from our lead series.
7643
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649

Journal of Medicinal Chemistry
Article
demonstrated in cardiac tissue24 and could be involved in the
local degradation of bradykinin. We speculated that CatA
inhibition could therefore be protective in cardiac hypertrophy.
Compound 2a was chosen because of its favorable ADME
profile that makes it especially suitable for chronic in vivo testing.
The heart failure animal model used in this study makes use of a
continuous angiotensin II (Ang II) infusion in male hyper-
cholesteremic ApoE ko mouse. This animal model combines the
two well-known cardiovascular risk factors hypercholesterolemia
and hypertension to experimentally establish cardiac hyper-
trophy and heart failure in a short period of time. Compound 2a
was administered at 100 mg kg−1 day−1 po and compared to the
aldosterone antagonist spironolactone (100 mg kg−1 day−1
b.i.d.). The results are summarized in Table 6.
To judge cardiac hypertrophy parameters, biometric weight
data were collected 4 weeks after compound treatment at the
study end. While there was no difference in the overall body
weights between the tested groups, clear and significant
differences in the left ventricular wet weight and the left
ventricular wet weight to body weight (LV wwt/BW) ratios were
found. Treatment with compound 2a significantly reduced the
Ang II induced increase of both LV wwt (p = 0.001) and LV
wwt/BW ratio (p = 0.0038) almost to control levels.
In line with the improvement of left ventricular hypertrophy
parameters the functional hemodynamic pressure−volume (p−
v) loop characterization revealed additional beneficial effects of
CatA inhibition in this heart failure mouse model. As first
functional parameters we measured mean arterial pressure
(MAP) and heart rate (HR) within the ascending aortic arch.
This was done just before the tip catheter was finally inserted
into the left ventricle for p−v loop analysis. While there existed a
clear and significant elevation of MAP in the Ang II infused
placebo group compared to the NaCl control group (p =
0.0299), compound 2a did not induce any significant changes in
MAP and HR under Ang II stimulated conditions. In contrast,
compound 2a led to a significant improvement of left ventricular
end diastolic pressure (p = 0.0012) and the left ventricular
relaxation time τ (Weiss) (τw, p = 0.0014) when compared to the
Ang II placebo group. No significant changes, however, could be
determined in the ejection fraction between all groups. And the
Ang II induced elevation in left ventricular end-systolic pressure
(LVESP) was not corrected by compound 2a treatment. Overall,
inhibition of CatA was not inferior to treatment with the
marketed aldosterone antagonist spironolactone.
Additionally, CatA expression has been shown to be increased
in human cardiac atrial tissue.21 To investigate the effect of CatA
inhibition by compound 2a on the development of an atrial
substrate for atrial fibrillation (AF), we used male Wistar rats
with ventricular ischemia/reperfusion (I/R), a rat model for
atrial structural remodeling. Three months after surgery the
median duration of atrial fibrillation induced by burst pacing, a
functional parameter describing the progression of an
arrhythmogenic substrate in the atrium, was decreased in 2a
treated I/R rats in comparison to placebo treated I/R rats (0.69
3.08 s vs 25.00 18.09 s, p < 0.01) (Figure 8). This indicates
that 2a displays potent atrial antiarrhythmic effects in a rat model
for atrial structural remodeling which leads to less stable AF
episodes and attenuates the progression of AF.
Figure 7. CatA inhibition leads to bradykinin-dependent diuresis in rats.
(a) Compounds 2a, 15a, and 15b increased urinary bradykinin levels
compared to untreated rats. Data are expressed as the mean SEM (n =
9−10). Statistical analysis was done by paired t test. (b) Anesthetized
rats received iv bolus injections of compounds 2a, 15a, and 15b. Urine
was collected for the given intervals from cannulated ureters and
quantified. Data are expressed as the mean SEM (n = 5−7). Statistical
analysis was done by two-way ANOVA with Bonferroni post-tests
relative to vehicle. (c) Compound 2a was given iv at 10 mg/kg in the
presence of HOE140 given at 0.5 mg/kg iv. Urine samples were
collected as described in (b). Statistical analysis was done by two-way
ANOVA with Bonferroni post-tests relative to compound x: (∗)
significance was set to p < 0.05.
animals the urine volume increased only slightly over time. In
contrast, all compounds significantly increased urine flow over
time (2a, 14.3 μL/min; 15a, 27.7 μL/min; 15b, 36.5 μL/min; vs
vehicle, 11.6 μL/min, p < 0.001). The magnitude of these effects
correlated well with the compounds' ability to increase urinary
bradykinin levels (Figure 7b). Finally, the diuretic activity
induced by 10 mg/kg 2a could be significantly attenuated by the
simultaneous application of the bradykinin receptor antagonist
HOE140 (15.3 μL/min vs 18.3 μL/min, p < 0.001),
demonstrating the kinin dependency of this effect (Figure 7c).
Besides its renal effects, bradykinin has been shown to play an
important role in reducing the development of cardiac
hypertrophy.23 Moreover, expression of CatA has been
CONCLUSION
■
The identification of novel small molecule inhibitors of
cathepsin A is reported for the first time. By employing a
fragment approach on HTS results, we rapidly identified 8a as a
7644
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649

Journal of Medicinal Chemistry
Article
a
Table 6. Effects of Compound 2a and Spironolactone in a Mouse Model of Ang II Induced Cardiac Hypertrophy
NaCl + placebo (N = 10)
Ang II + placebo (N = 13)
Ang II + 2a (N = 14)
Ang II + spironolactone (N = 9)
BW (g)
31.95 0.69
30.48 0.52
182.03 8.45
5.99 0.28
29.63 0.62
28.84 0.91
b
b
b
LV wwt (mg)
LV wwt/BW (mg/g)
MAP (mmHg)
heart rate (bpm)
EF (%)
140.06 9.71
142.69 4.13
145.96 9.27
b
b
4.38 0.29
4.84 0.18
5.07 0.30
80.77 2.54
406.41 15.83
56.49 2.13
92.92 3.04
99.03 6.82
403.23 7.98
51.14 2.49
139.36 8.79
15.49 1.95
8.73 0.41
95.88 5.03
407.11 10.37
57.89 1.80
129.95 7.64
103.99 6.66
410.78 11.70
58.59 1.40
137.39 6.77
14.93 1.30
7.52 0.32
LVESP (mmHg)
LVEDP (mmHg)
τw (ms)
b
b
7.31 0.78
8.32 1.10
b
b
7.19 0.33
7.05 0.25
a
Data are shown as mean values with standard error of the mean (SEM). The homogeneity of variances was tested using the Levene test. In the case
of homogenous variances, one way ANOVA followed by a Dunnett’s post hoc test versus the Ang II placebo group was performed. In the case of
heterogeneous variances, one way ANOVA on rank-transformed values followed by Dunnett’s post hoc test versus the ANG II placebo group was
b
performed. p < 0.05 was defined as significant. All analyses were performed using SAS (version 8.2) for SUN 4 via interface software EverStat,
version 5.0. Abbreviations: BW, body weight; MAP, mean arterial pressure; LV, left ventricle; wwt, wet weight; EF, ejection fraction; LVESP, left
ventricle end systolic pressure; LVEDP, left ventricle end diastolic pressure.
standard. Purity and characterization of compounds were established by
a combination of LCMS and NMR analytical techniques. The purity of
the final compound is >95% unless otherwise noted. The employed
HPLC and LCMS methods (columns, eluents, gradients, and flow
rates) are described below and are specified for each described
compound. The HPLC−MS instrumentation was used with one of the
two following setups: (A) Waters Aquity UPLC system equipped with a
Waters Aquity autosampler, UPLC pump, and a diode array and the
HPLC eluent coupled with a 1:3 split to a Waters Aquity SQD single
quadrupole mass sepctrometer with electrospray ionization, with
spectra scanned from 100 to 1300 amu in positive and negative
modes (LC3, LC5, LC6, LC7); (B) Agilent series 1100 system
equipped with a degasser G1322A, bin pump G1312A, ALS G1313A,
col comp G1316A, DAD G1315A, and MSD G1946A, with the HPLC
eluent coupled to a mass spectrometer with electrospray ionization and
with spectra scanned from 110 to 1000 amu in positive and negative
modes (LC1, LC2, LC4).
Figure 8. CatA inhibition reduces duration of atrial fibrillation (AF)
LC1. Column: YMC-Pack Jsphere H80, 33 mm × 2.1 mm, 4 μm.
induced by burst-pacing in rats with ischemia/reperfusion (I/R) (n = 14
Flow: 1.3 mL/min. Room temperature. Eluent A: water + 0.05 % TFA.
per group). Individual AF duration is shown in black diamonds. Median
Eluent B: ACN + 0.05 % TFA. Gradient: from 95% A + 5% B to 5% A +
AF duration is shown in white squares (A). Representative atrial
95% B within 2.5 min. MS ionization method: ES+.
electrograms during electrophysiological measurements in 2a-treated I/
LC2. Column: YMC-Pack Jsphere H80, 33 mm × 2.1 mm, 4 μm.
R rats (B) and in placebo treated I/R rats (C). Statistical analysis was
Flow: 1.0 mL/min. Room temperature. Eluent A: water + 0.05% TFA.
done by ANOVA. Significance was set to p < 0.05.
Eluent B: ACN + 0.05% TFA. Gradient: 98% A + 2% B for 1.0 min, then
to 5% A + 95% B within 4.0 min, then 5% A + 95% B for 1.25 min. MS
lead for our CatA inhibitor program. Rational design using X-ray
crystallographical data guided the following lead optimization
strategy, which was successful in overcoming the cell
permeability issue associated with 8a and delivered CatA
inhibitors with high nanomolar in vitro activity and oral
bioavailability. Pharmacological studies undertaken with 2a,
15a, 15b confirmed the reported influence of CatA inhibition on
renal bradykinin levels and resulted in the identification of
beneficial effects in models of cardiac hypertrophy and of atrial
fibrillation after administration of 2a. On the basis of more
detailed pharmacological investigations, which will be reported
elsewhere, we were able to select one of our novel CatA
inhibitors as a drug candidate, which is currently in phase I
clinical trials.
ionization method: ES+.
LC3. Column: Waters XBridge C18, 50 mm × 4.6 mm, 2.5 μm. Flow:
1.7 mL/min. 40 °C. Eluent A: water + 0.05% TFA. Eluent B: ACN +
0.05% TFA. Gradient: 95% A + 5% B for 0.3 min, then to 5% A + 95% B
within 3.2 min, then 5% A + 95% B for 0.5 min. MS ionization method:
ES+.
LC4. Column: Merck Chromolith FastGrad RP-18e, 50 mm × 2 mm.
Flow: 2.0 mL/min. Room temperature. Eluent A: water + 0.05% TFA.
Eluent B: ACN + 0.05% TFA. Gradient: 98% A + 2% B for 0.2 min, then
to 2% A + 98% B within 2.2 min, then 2% A + 98% B for 0.8 min, then to
98% A + 2% B within 0.1 min, then 98% A + 2% B for 0.7 min. MS
ionization method: ES+.
LC5. Column: Waters XBridge C18, 50 mm × 4.6, 2.5 μm. Flow: 1.3
mL/min. Room temperature. Eluent A: water + 0.1% FA. Eluent B:
ACN + 0.08% FA. Gradient: from 97% A + 3% B to 2% A + 98% B
within 18.0 min, then 2% A + 98% B for 1.0 min, then to 97% A + 3% B
within 0.5 min, then 97% A + 3% B for 0.5 min. MS ionization method:
ES+.
LC6. Column: Waters XBridge C18, 50 mm × 4.6 mm, 2.5 μm. Flow:
1.7 mL/min. 50 °C. Eluent A: water + 0.05% TFA. Eluent B: ACN +
0.05% TFA. Gradient: 95% A + 5% B for 0.2 min, then to 5% A + 95% B
within 2.2 min, then 5% A + 95% B for 1.1 min, then to 95% A + 5% B
within 0.1 min, then 95% A + 5% B for 0.9 min. MS ionization method:
ES+.
EXPERIMENTAL SECTION
■
All commercially available reagents and solvents were used without
further purification. Reversed phase high pressure chromatography was
conducted on an Abimed Gilson instrument using a LiChrospher 100
RP-18e (5 μm) column from Merck. Nuclear magnetic resonance (1H
NMR) spectra were recorded in DMSO-d6 on a Bruker Avance DRX
500, an Avance II 400, or Avance III 600 at room temperature. Chemical
shifts are reported in as δ values from an internal tetramethylsilane
7645
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649

Journal of Medicinal Chemistry
Article
LC7. Column: Waters XBridge C18, 50 mm × 4.6 mm, 2.5 μm. Flow:
1.7 mL/min. 40 °C. Eluent A: water + 0.05% TFA. Eluent B: ACN +
0.05% TFA. Gradient: from 95% A + 5% B to 5% A + 95% B within 3.3
min, then 5% A + 95% B for 0.55 min, then to 95% A + 5% B within 0.15
min. MS ionization method: ES+.
(S)-3-{[1-(2-Fluorophenyl)-5-hydroxy-1H-pyrazole-3-
carbonyl]amino}-3-o-tolylpropionic Acid (8a). General Proce-
dure A for the Synthesis of Compounds 8. Step 1. Sulfuric acid
(400 mL) was added slowly to water (400 mL). After the mixture was
cooled to 5 °C, 2-fluorophenylhydrazine hydrochloride (123 g, 757
mmol) was added resulting in a brown suspension. Then a solution of
oxalacetic acid (100 g, 757 mmol) in water (400 mL) was added slowly
during a period of 25 min. After 2 h the conversion was complete and
the solid was filtered. After a wash with water, the solid was dried. The
product 1-(2-fluorophenyl)-5-hydroxy-1H-pyrazole-3-carboxylic acid
7a was obtained as light brown solid (151 g, 90%) and used directly
in the next step.
Step 2. An amount of 0.45 mmol of 1-(2-fluorophenyl)-5-hydroxy-
1H-pyrazole-3-carboxylic acid was dissolved in 5 mL of DMF, and 0.54
mmol of TOTU and 1.125 mmol of NEM were added. The mixture was
stirred for 5 min at room temperature. Then 0.495 mmol of (S)-3-
amino-3-o-tolylpropionic acid was added and the mixture stirred
overnight at room temperature. The solvent was evaporated in vacuo
and the residue subjected to preparative HPLC to give 8a in a yield of
28%. 1H NMR (500 MHz, DMSO) δ 8.60 (d, 1H), 7.55 (m, 2H), 7.45
(m, 2H), 7.35 (m, 1H), 7.15 (m, 3H), 5.80 (s, 1H), 5.60 (q, 1H), 2.90
(dd, 2H), 2.70 (dd, 2H), 2,40 (s, 3H). LCMS (LC3): m/z 383,13 (M +
H).
General Procedure B for the Synthesis of Compounds 2 and
13−16 (Steps 1−4). (S)-3-{[1-(2-Fluorophenyl)-5-((R)-2-hy-
droxy-3,3-dimethylbutoxy)-1H-pyrazole-3-carbonyl]amino}-3-
o-tolylpropionic Acid (15a). Step 1. Dimethyl acetylenedicarboxylate
(87.4 g, 757 mmol) was added to a solution of 2-fluorophenylhydrazine
hydrochloride (100 g, 615 mmol) in methanol (1 L) at 0 °C. Then
triethylamine (125 g, 1.23 mol) was added slowly during 60 min. The
solution was stirred for 16 h at room temperature. The solvent was then
removed under reduced pressure and the residue dissolved in EA (500
mL). After washing with aqueous hydrochloric acid (500 mL), the
solvent was removed under reduced pressure and methyl 1-(2-
fluorophenyl)-5-hydroxy-1H-pyrazole-3-carboxylate 11a was obtained
as an off white solid in quantitative yield.
Step 2. An amount of 200 mg (0.85 mmol) of methyl 1-(2-
fluorophenyl)-5-hydroxy-1H-pyrazole-3-carboxylate 11a was dissolved
in 5 mL of DMF. Then 552 mg (1.7 mmol) of cesium carbonate and
152 mg (0.85 mmol) of 1-bromo-3,3-dimethylbutan-2-one were added,
and the mixture was heated to 50 °C for 6 h. Then the mixture was
filtered and the solvent removed in vacuo. The obtained crude title
compound 5-(3,3-dimethyl-2-oxo-butoxy)-1-(2-fluorophenyl)-1H-pyr-
azole-3-carboxylic acid methyl ester was subjected to hydrolysis without
further purification.
Step 3. An amount of 100 mg (0.3 mmol) of 5-(3,3-dimethyl-2-
oxobutoxy)-1-(2-fluorophenyl)-1H-pyrazole-3-carboxylic acid methyl
ester was dissolved in 5 mL of MeOH. Then 0.7 mmol of lithium
hydroxide and 2 mL of water were added, and the mixture was stirred
overnight at room temperature. The solvent was removed in vacuo and
the residue subjected to aqueous workup with a 10% solution of citric
acid and DCM. The organic phase was dried and the solvent removed in
vacuo to give 5-(3,3-dimethyl-2-oxobutoxy)-1-(2-fluorophenyl)-1H-
pyrazole-3-carboxylic acid, which was used without further purification
in the next step.
1H), 7.65 (m, 1H), 7.60 (m, 1H), 7.50 (m, 2H), 7.40 (m, 1H), 7.10 (m,
3H), 6.10 (s, 1H), 5.60 (q, 1H), 5.25 (s, 2H), 2.90 (dd, 1H), 2.70 (dd,
1H), 2.40 (s, 3H), 1.10 (s, 9H) LCMS (LC4): m/z 482,18 (M + H).
Step 5. An amount of 50 mg (0.1 mmol) of (S)-3-{[5-(3,3-dimethyl-
2-oxobutoxy)-1-(2-fluorophenyl)-1H-pyrazole-3-carbonyl]amino}-3-o-
tolylpropionic acid 13a was dissolved in 5 mL of methanol, and 0.05
mmol of sodium borohydride was added. The mixture was stirred at
room temperature overnight. The solvent was removed in vacuo and the
residue subjected to preparative HPLC to give 52% of (S)-3-{[1-(2-
fluorophenyl)-5-(2-hydroxy-3,3-dimethylbutoxy)-1H-pyrazole-3-
carbonyl]amino}-3-o-tolylpropionic acid 14a. An amount of 100 mg
14a was separated into the two diastereomeric title compounds by
HPLC on a chiral column (Chiralpak AD-H/55, 250 mm × 4.6 mm) at
30 °C (eluent, heptane/ethanol/MOH (15:1:1) + 0.1% TFA; flow, 1.0
mL/min). Then 45 mg of the first diastereomer (retention time, 16.35
min) eluted from the column and 40 mg of the second diastereomer
(retention time, 21.15 min) eluted from the column were obtained, one
of them being the diastereomer with R configuration in the alcohol
moiety and the other being the diastereomer with S configuration in the
alcohol moiety. The stereochemistry in the alcohol moiety was not
determined; it was arbitrarily assigned R configuration in the first
diastereomer eluted from the column 15a and S configuration in the
second diastereomer eluted from the column 16a.
(S)-3-{[1-(2-Fluorophenyl)-5-((R)-2-hydroxy-3,3-dimethylbu-
toxy)-1H-pyrazole-3-carbonyl]amino}-3-o-tolylpropionic Acid
1
(15a). H NMR (500 MHz, DMSO) δ 8.65 (d, 1H), 7.60 (m, 2H),
7.60 (m, 1H), 7.45 (m, 2H), 7.35 (m, 1H), 7.10 (m, 3H), 6.20 (s, 1H),
5.60 (q, 1H), 4.90 (d, 1H), 4.25 (dd, 1H), 3.95 (dd, 1H), 2.90 (dd, 1H),
2.70 (dd, 1H), 2.45 (s, 3H), 0.85 (s, 9H). LCMS (LC5): m/z 484.34 (M
+ H), tR = 10.16 min.
(S)-3-{[1-(2-Fluorophenyl)-5-((S)-2-hydroxy-3,3-dimethylbu-
toxy)-1H-pyrazole-3-carbonyl]amino}-3-o-tolylpropionic Acid
1
(16a). H NMR (500 MHz, DMSO) δ 8.65 (d, 1H), 7.6 (m, 2H),
7.45 (m, 2H), 7.35 (m, 1H), 7.10 (m, 3H), 6.20 (s, 1H), 5.60 (q 1H),
4.90 (bs, 1H), 4.20 (dd, 1H), 3.95 (dd, 1H), 2.90 (dd, 1H), 2.70 (dd,
1H), 2.45 (s, 3H), 0.85 (s, 9H). LCMS (LC5): m/z 484.34 (M + H), tR
= 10.33 min.
(S)-3-(2,4-Dichlorophenyl)-3-{[5-(3,3-dimethyl-2-oxobu-
toxy)-1-(2-fluorophenyl)-1H-pyrazole-3-carbonyl]amino}-
propionic Acid (13b). Following general procedure B and using (S)-
1
3-amino-3-(2,4-dichlorophenyl)propionic acid in step 4 gave 13b. H
NMR (500 MHz, DMSO) δ 8.90 (d, 1H), 7.7 (m, 1H), 7,55 (m, 3H),
7.5 (m, 1H), 7.4 (m, 2H), 6.20 (s, 1H), 5.70 (q 1H), 5.30 (s, 2H), 2.90
(dd, 1H), 2.65 (dd, 1H), 1.10 (s, 9H). LCMS (LC4): m/z 536.20 (M +
H).
(S)-3-(2,4-Dichlorophenyl)-3-{[1-(2-fluorophenyl)-5-((R)-2-
hydroxy-3,3-dimethylbutoxy)-1H-pyrazole-3-carbonyl]amino}-
propionic Acid (15b). Following general procedure B and using (S)-
1
3-amino-3-(2,4-dichlorophenyl)propionic acid in step 4 gave 15b. H
NMR (500 MHz, DMSO) δ 8.90 (d, 1H), 7.60 (m, 4H), 7.45 (m, 2H),
7.40 (m, 1H), 6.25 (s, 1H), 5.70 (q, 1H), 4.90 (d, 1H), 4.25 (dd, 1H),
3.95 (dd, 1H), 2.90 (dd, 1H), 2.65 (dd, 1H), 0.85 (s, 9H). LCMS
(LC5): m/z 538.24 (M + H), tR = 11.13 min.
(S)-3-(2,4-Dichlorophenyl)-3-{[1-(2-fluorophenyl)-5-((S)-2-hy-
droxy-3,3-dimethylbutoxy)-1H-pyrazole-3-carbonyl]amino}-
propionic Acid (16b). Following general procedure B and using (S)-
1
3-amino-3-(2,4-dichlorophenyl)propionic acid in step 4 gave 16b. H
NMR (500 MHz, DMSO) δ 8.85 (d, 1H), 7.60 (bm, 4H), 7.45 (m, 2H),
7.40 (m, 1H), 6.25 (s, 1H), 5.70 (q, 1H), 4.90 (bs, 1H), 4.25 (dd, 1H),
3.95 (dd, 1H), 2.90 (dd, 1H), 2.65 (dd, 1H), 0.85 (s, 9H). LCMS
(LC5): m/z 538,24 (M + H), tR = 11.36 min.
Step 4. An amount of 324 mg (0.675 mmol) of 5-(3,3-dimethyl-2-
oxobutoxy)-1-(2-fluorophenyl)-1H-pyrazole-3-carboxylic acid was dis-
solved in 5 mL of DMF, and 0.675 mmol of TOTU and 1.35 mmol of
EDIA were added. The mixture was stirred for 5 min at room
temperature. Then 0.675 mmol of (S)-3-amino-3-(2-methylphenyl)-
propionic acid was added, and the mixture was stirred overnight at room
temperature. After evaporation of the solvent, the residue was subjected
to preparative HPLC to give (S)-3-{[5-(3,3-dimethyl-2-oxobutoxy)-1-
(2-fluorophenyl)-1H-pyrazole-3-carbonyl]amino}-3-o-tolylpropionic
acid (13a) in a yield of 67%. 1H NMR (400 MHz, DMSO) δ 8.70 (d,
(S)-3-{[1-(2-Fluorophenyl)-5-methoxy-1H-pyrazole-3-
carbonyl]amino}-3-(2-methylphenyl)propionic Acid (2a). Fol-
lowing steps 1−4 from general procedure B using MeI in step 2 gave 2a.
1H NMR (500 MHz, DMSO) δ 8.70 (d, 1H), 7.60 (m, 2H), 7,45 (m,
2H), 7.35 (m, 1H), 7.15 (m, 3H), 6.20 (s, 1H), 5.60 (q 1H), 3.90 (s,
3H), 2.90 (dd, 1H), 2.70 (dd, 1H), 2.45 (s, 3H). LCMS (LC3): m/z
398,18 (M + H).
Expression and Purification of Recombinant Human Proca-
thepsin A. ORF coding for human procathepsin A (proCatA, Ala29-
Tyr480, BC000597) was cloned from full-length cDNA in frame with
7646
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649

Journal of Medicinal Chemistry
Article
insect prepromelitin signal sequence and Myc-tag followed by a 10xHis
at the C-terminal end of the coding protein. The DNA was ligated into
the multiple cloning site of the baculovirus transfer vector25 pVL1393
vector (AB vector). After co-transfection of plasmids with baculovirus-
DNA (FlashBac GOLD, Oxford Expression Technologies),26 the virus
was amplified in two steps in Spodoptera frugiperda27 cell line SF9 in
SF900II medium (Invitrogen) supplemented with 5% FCS. Recombi-
nant virus was harvested 5 days post-transfection. The virus titers were
determined by plaque assay method28 and reached about 1 × 108 pfu/
mL.
20−30 min of incubation at room temperature fluorenscence intensity
was measured in a Tecan Safire2 plate reader (various emission/
excitation wavelengths based on use substarte). The % inhibition and
IC50 values were calculated based on internal high and low controls.
Crystallography. To obtain the active form of the enzyme, the 54
kDa CatA precursor was proteolytically transformed into the active
mature form by cleavage of a 2 kDa excision peptide with trypsin-
Sepharose followed by gel filtration.
CatA was crystallized using the hanging drop method. Then 1 μL of
protein solution, containing 6.5 mg/mL cathepsin A, 25 mM Tris-HCl
(pH 8.0), and 300 mM NaCl, was mixed with 1 μL of reservoir solution,
containing 100 mM sodium acetate (pH 4.5), 18−20% PEG400, and
100 mM CdCl2, and set to equilibrate at 4 °C. Rod-shaped crystals
appeared in about 1 week. Crystals were soaked with inhibitors by
transferring them to a drop of 9 μL reservoir solution with 1 μL of a 100
mM solution of 8a or 15a in DMSO. For cryoprotection, 20% glycerol
was added to the soaking solution and the crystal was picked with a
small nylon loop and flash frozen in liquid nitrogen. Data were
measured at the ESRF at beamlines ID29 and ID14-4 and processed
with XDS/XSCALE31 as implemented in APRV.32 The structures were
solved using a previously solved cathepsin A structure as a starting
model. The initial model was obtained by molecular replacement with
Phaser33 using the CatA precursor structure14 (PDB code 1IVY) as a
starting model. Model building was done with Coot,34 and refinement
was done with Buster.35
ProCatA was expressed for 72 h dpi at MOI 3 in 1.1 × 106 cells/mL
High Five cells growing in ExCell405 medium at a 4 L scale using four
vented 3000 mL flasks (Corning) at 110 rpm at 27 °C.
̈
Purification steps were carried out at 4 °C using a AKTA explorer
system (GE) for chromatography. His-tagged protein was bound
directly to metal chelate ligand29 from the clarified insect cell
supernatant using a 5 mL HisTrap column HP (GE) with a flow rate
of 3 mL/min. After a wash with 10 CV, 20 mM NaPO4, pH 8.0, 500 mM
NaCl within a linear gradient, bound protein was eluted from 0% to
100% buffer, 20 mM NaPO4, pH 8.0, 500 mM NaCl, and 1000 mM
imidazole. Fractions were collected and analyzed on SDS−PAGE,30 and
fractions containing proCatA were pooled. Protein was polished by size
exclusion chromatography (SEC) on Superdex 26/60 200 HR (GE)
using 300 mM NaCl, 50 mM Tris-HCl, pH 7.5, as buffer.
Cathepsin A Activity Assays. Recombinant human proCatA was
proteolytically activated with recombinant human cathepsin L (R&D
Systems, no. 952-CY) according to the manufacturer's instructions.
Briefly, proCatA was incubated at 10 μg/mL with cathepsin L (CatL) at
1 μg/mL in 25 mM 2-(N-morpholino)ethanesulfonic acid (MES), 5
mM dithiothreitol (DTT), pH 6.0, for 15 min at 37 °C. CatL activity
was then stopped by the addition of 10 μM N-(trans-epoxysuccinyl)-L-
leucine-4-guanidinobutylamide (E64, Sigma). The activated CatA was
diluted in 25 mM MES, 5 mM DTT, pH 5.5, to a final concentration of
833 ng/mL and mixed with compounds or vehicle in a multiple assay
plate. After incubation for 15 min at room temperature a bradykinin
peptide carrying a N-terminal BodipyFL label (Jerini Peptide
Technology) was added to the assay mixture to a final concentration
of 2 μM. After incubation for 15 min at room temperature the reaction
was stopped by the addition of 130 mM 2-(4-(2-hydroxyethyl)-1-
piperazinyl)ethansulfonic acid, pH 7.4, containing 0.013% Triton X-
100, 0.13% coating reagent (Caliper Life Sciences), 6.5% DMSO, and
20 μM ebelactone B (Sigma, no. E0886). Uncleaved substrate and
product were then separated by a microfluidic capillary electrophoresis
on a LabChip 3000 drug discovery system (12-Sipper-Chip, Caliper Life
Sciences) and quantified by determination of the respective peak areas.
Substrate turnover was calculated by dividing product peak area by the
sum of substrate and product peak areas.The enzyme activity and the
inhibitory effect of any test compound could thus be quantified. The
efficacy of test compounds is expressed as the concentration that
induces 50% inhibition of enzyme activity (IC50).
Biochemical Selectivity Assays. CatA inhibitor selectivity was
determined against several proteases. Recombinant human cathepsin L
(CathL, R&D Systems, no. 952-CY), cathepsin S (CathS, R&D
Systems, no. 1183-CY), ACE (R&D Systems, no. 929-ZN), ACE2
(R&D Systems, no. 933-ZN), DPPlV (R&D Systems, no. 1180-SE),
NEP (R&D Systems, no. 1182-ZNC), Kallikrein (R&D Systems, no.
2337-SE), human cathepsin D from liver (CathD, Enzo Life Sciences,
no. BML-SE199), human chymotrypsin from pancreatic tissue (US
Biological, no. C5070-08), factor Xa from human plasma (Xa, US
Biological, no. F0018-15), and recombinant HCV NS3/4a (AnaSpec,
no. 61017-10) were tested accoding to manufacturers' instructions
(adapted to 384-well format). Briefly, enzymes (3 μL) and inhibitor (3
μL), diluted in appropriate reaction buffer, were transferred into black
384-well low volume plates (medium binding) and were incubated for
15 min at room temperature followed by the addition of substrate
solution (diluted in reaction buffer: CathL (ES008, R&D), CathS
(ES002, R&D), CathD (M-1895, Bachem), ACE (ES005, R&D), ACE2
(ES007, R&D), DPPIV (L-1225, Bachem), NEP (ES005, R&D),
Kalikrein (P9273, Sigma), chymotrypsin (J-1355, Bachem), Xa (ES008,
R&D), NS3/4a (61017-10, AnaSpec)) to a final volume of 9 μL. After
Molecular Modeling and Visualization. Docking was performed
with GLIDE 3.5 in standard precision mode without further
minimization, using Glidescore for ranking. All ligands were prepared
using LigPrep including the Epik approach.36 All protomers and
tautomers existing at pH 7.4 were included in the docking studies. The
docking results were filtered manually. All graphics were generated
using Sybyl 8.037
In Vivo Experiments. All experiments using animals were
performed in accordance with the German law for the protection of
animals and adhered to the international animal welfare legislation and
rules.
Pharmacokinetic Studies. Pharmacokinetic parameters of 2a, 15a,
and 15b were determined in male Sprague−Dawley rats (Harlan
Winkelmann, Germany). After single oral administration of 10 mg/kg
in a solution consisting of glycofurol 75 (19%), Cremophor RH40
(6%), and water (75%), plasma and tissue samples were taken at eight
time points over 24 h. Three animals per time point were used in a
nonserial sampling design. Heart tissue was homogenized with water (2
g of water for 1 g tissue) using a ball mill (Retsch). The respective
compound concentrations in plasma and tissue homogenates were
determined using an exploratory LCMS/MS method using a
structurally related compound as internal standard. The area under
the plasma concentration−time curve was calculated using the software
WinNonlin (Pharsight Corporation).
Pharmacological Studies. For the diuresis experiments male
Wistar rats were anesthetized with urethane (1.25 mg/kg im) and the
carotid artery and the jugular vein were cannulated for blood pressure
control and for infusion with albumine in physiologic saline (0.28 mg
kg−1 h−1) and bolus injection of compound. Animals were placed on a
heating water pad. Infusion with albumin started directly after
cannulation. After median lapratomy both ureters were cannulated
with tubes and urine was sampled. Rats received iv bolus injections of
compounds in 0.15% Na2CO3 20 min after start of albumin infusion.
Urine was collected in five succeeding time periods of 20 min. The
volume of each urine sample was measured.
Cardiac hypertrophy was induced in male B6.129P2-ApoE/J mice
(ApoE ko mice) by continuous 28 days of infusion of Ang II (1 μg kg−1
min−1 sc) via osmotic mini pumps (model 2004, ALZET Osmotic
Pumps, Durect Corporation, U.S.). Before subcutaneous implantation
minipumps were filled with either sterile saline (NaCl placebo group) or
Ang II diluted in sterile saline (all other groups). The minipumps were
subcutaneously implanted under short isoflurane anesthesia and were
thereafter active for the next 4 weeks. Oral treatment started on the day
of surgery approximately 2 h after recovery from surgery. 2a was
administered at 100 mg kg−1 day−1 in HEC 0.6% + Tween 0.5%
7647
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649

Journal of Medicinal Chemistry
Article
(vehicle) by oral gavage, while within the Ang II placebo group and the
NaCl placebo control group animals received vehicle only. As a
reference compound spironolactone at a dose of 100 mg/kg po twice
daily was chosen.
ABBREVIATONS USED
■
CatA, cathepsin A; ANG II, angiotensin II; EA, ethyl acetate;
EDIA, ethyldiisoproylamine; EF, ejection fraction; MAP, mean
arterial pressure; NEP, neprylisine; HPLC, high pressure liquid
chromatography; HR, heart rate; LV, left ventricle; LVEDP, left
ventricular end diastolic pressure; LVESP, left ventricular end
systolic pressure; RIA, radioactive immunoassay; NEM, N-
ethylmorpholine; RP, reverse phase; TOTU, O-
((ethoxycarbonyl)cyanomethyleneamino)-N,N,N′,N′-tetrame-
thyluronium tetrafluoroborate; wwt, wet weight
After 28 days, the final functional examinations by aortic arch tip
catheter analysis and left ventricular pressure−volume (p−v) loop
measurements were performed using a 1.4 French catheter (SPR-839,
Millar Instruments, U.S.). Briefly, the animals were anesthetized with
thiopental at 100 mg/kg ip. As a next step the right carotid artery was
prepared for catheter insertion and the p−v loop catheter was initially
placed within the ascending aortic arch for MAP and HR registration
and finally placed within the left ventricular chamber for steady-state
characterization of p−v loop parameters. Afterward the abdomen was
opened below the diaphragm and the vena cava was exposed for
ligation. Using complete occlusion of the inferior vena cava the preload-
recruitable left ventricular performance was registered. All parameters
were digitally recorded using the Notocord hem evolution software, and
the analysis of the different functional parameters was performed offline
at a later time point.
Atrial arrhythmogenic substrate was induced by ventricular ischemia
reperfusion (I/R). In 28 male Wistar rats (3−4 months old) a left-sided
thoracotomy was performed and myocardial ischemia was induced by
temporary occlusion of the left coronary artery close to its origin from
the aorta for 30 min. Ischemia induction was followed by reperfusion.
CatA inhibitor 2a (30 mg kg−1 day−1) was administered via oral lavage
once a day in 14 I/R rats. Fourteen placebo treated I/R rats received
vehicle only (HEC 0.6% + Tween 0.5%). After 3 months of treatment,
all rats were anesthetized (intraperitoneal injection of pentobarbital)
and electrophysiological measurements were performed in open chest
experiments. Atrial pacing threshold was determined and unipolar
pacing was performed from the surface of the left atrial appendage using
a pacing electrode (tip electrode) next to a custom-made multiple
action potential catheter (Franz-like electrode) with a pulse of 1 ms and
at twice the diastolic threshold. The longest episode of atrial fibrillation
induced by five repetitive 1 s bursts of stimuli (very rapid pacing, cycle
length of 10 ms) was determined. When atrial electrograms showed a
rapid atrial rate, cycle length of <70 ms, and duration of >5 beats, atrial
fibrillation was diagnosed.
REFERENCES
■
(1) Hiraiwa, M. Cathepsin A/protective protein: an unusual lysosomal
multifunctional protein. Cell. Mol. Life Sci. 1999, 55, 894−907.
(2) (a) Itoh, K.; Takiyama, N.; Kase, R.; Kondoh, K.; Sanao, A.;
Oshima, A.; et al. Purification and characterization of human lysosomal
protective protein expressed in stably transformed Chinese hamster
ovary cells. J. Biol. Chem. 1993, 268, 1180−1186. (b) Jackman, H. J.;
Tan, F.; Tamei, H.; Beurling-Harbury, C.; Li, X.-Y.; Skidgel, R. A.; et al.
A deamidase in human platelets that deamidates tachykinins: probable
identity with the lysosomal protective protein. J. Biol. Chem. 1990, 265,
11265−11272. (c) Jackman, H. J.; Morris, P. W.; Deddish, P. A.;
Skidgel, R. A.; Erdos, E. G. Inactivation of endothelin I by deamidase
̈
(lysosomal protective protein). J. Biol. Chem. 1992, 267, 2872−2875.
(d) Itoh, K.; Kase, R.; Simmoto, M.; Satake, A.; Sakaruba, H.; Suzuki, Y.
Protective protein as an endogenous endothelin degradation enzyme in
human tissues. J. Biol. Chem. 1995, 270, 515−518.
(3) (a) d’Azzo, A.; Hoogeveen, A. T.; Reuser, J. J.; Robinson, H.;
Galjaard, H. Molecular defect in combined β-galactosidase and
neuraminidase deficiency in man. Proc. Natl. Acad. Sci. U.S.A. 1982,
79, 4535−4539. (b) Zhou, X. Y.; Morreau, H.; Rottier, R.; Davis, D.;
Bonten, E.; Gillemans, N.; et al. Mouse model for the lysosomal
disorder galactosialidosis and correction of the phenotype with
overexpressing erythroid precursor cells. Genes Dev. 1995, 9, 2623−
2634.
(4) Seyrantepe, V.; Hinek, A.; Junzheng, P.; Fedjaev, M.; Sheila, E.;
Yoshito, K.; Canuel, M.; Itoh, K.; Morales, C. R.; Lavoie, J.; Tremblay,
J.; Pshezhetsky, A. V. Enzymatic activity of lysosomal carboxypeptidase
(cathepsin) A is required for proper elastic fiber formation an
inactivation of endothelin-1. Circulation 2008, 117, 1973−1981.
(5) Miller, J. J.; Changaris, D. G.; Levy, R. S. Conversion of angiotensin
I to angiotensin II by cathepsin A isozymes of porcine kidney. Biochem.
Biophys. Res. Commun. 1988, 154, 1122−1129.
(6) Ito, H.; Majima, M.; Nakajima, S.-i.; Hayashi, I.; Katori, M.; Izumi,
T. Effect of prolonged administration of a urinary kininase inhibitor
ebelactone B on the development of deoxycorticosterone acetate-salt
hypertension in rats. Br. J. Pharmacol. 1999, 126, 613−620.
(7) Hayashi, I.; Majima, M.; Fujita, T.; Okumura, T.; Kumagai, Y.;
Tomita, N.; et al. In vivo transfer of antisense oligonucleotide against
urinary kininase blunts deoxycorticosterone acetate-salt hypertension in
rats. Br. J. Pharmacol. 2000, 131, 820−826.
Urinary Bradykinin. Bradykinin in urine was determined from male
Wistar rats collected over a period of 6 h before and after treatment of
rats with compounds 2a and 15a,b at 30 mg kg−1 day−1. A commercially
available RIA (Bachem) was used according to the manufacturer's
instructions.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental and spectoscopic data for nonkey compounds and
X-ray crystallographic data. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
(8) Cuervo, A. M.; Mann, L.; Bonten, E. J.; d’Azzo, A.; Dice, J. F.
Cathepsin A regulates chaperone-mediated autophagy through cleavage
of the lysosomal receptor. EMBO J. 2003, 22, 47−59.
*Phone: +49 69 305 12982. E-mail: sven.ruf@sanofi.com.
Present Address
§Evangelisches Krankenhaus, Oberhausen, Germany.
(9) Westermann., D.; Schultheiss, H. P.; Tschope, C. New perspective
̈
on the tissue kallikrein-kinin system in myocardial infarction: role of
angiogenesis and cardiac regeneration. Int. Immunopharmacol 2008, 8,
148−154.
Notes
The authors declare no competing financial interest.
(10) Kam, J.-D. Dual ACE/NEP inhibitorsmore than playing the
ACE card. J. Hum. Hypertens. 2006, 20, 478−481.
ACKNOWLEDGMENTS
■
(11) Weiner, B.; Szymanski, W.; Jansen, D. B.; Minnard, A. J.;
Ferninga, B. L. Recent advances in the catalytic asymmetric synthesis of
β-amino acids. Chem. Soc. Rev. 2010, 39, 1656−1691.
(12) Montalbetti, C. A. G. N.; Falque, V. Amide bond formation and
peptide coupling. Tetrahedron 2005, 61, 10827−10852.
The authors thank Marion Zerlin and Hartmut Mors for the
CatA HTS campaign. The authors also thank Karolina Daton,
Daniel Hein, Alexander Liesum, and Anke Hullman for excellent
technical assistance.
7648
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649

Journal of Medicinal Chemistry
Article
(13) Oprea, T. I.; Davis, M.; Teague, S. J.; Leeson, P. D. Is there a
difference between leads and drugs? A historical perspective. J. Chem.
Inf. Comput. Sci. 2001, 41, 1308−1315.
(14) Artursson, P. Epithelial transport of drug in cell culture: a model
for studying the passive diffusion of drugs over intestinal absorbtive
(Caco-2) cells. J. Pharm. Sci. 1990, 79, 476−482.
(15) Protein Data Bank deposition codes for 8a and 15a: 4az0 and
4az3.
(16) Rudenko, G.; Bonten, E.; d’Azzo, A.; Hol, W. G. J. Three-
dimensional structure of the human protective protein: structure of the
precursor form suggests a complex activation mechanism. Structure
1995, 3, 1249−1259.
(17) Rudenko, G.; Bonten, E.; d’Azzo, A.; Hol, W. G. J. Structure
determination of the human protective protein: twofold averaging
reveals the three-dimensional structure of a domain, which was entirely
absent in the initial model. Acta Crystallogr. 1996, D52, 923−936.
(18) Stella, V. J.; Himmelstein, J. Prodrugs and site-specific drug
delivery. J. Med. Chem. 1980, 23, 1275−1282.
(19) Verdonk, M. L.; Cole, J. C.; Taylor, R. SuperStar: a knowledge-
based approach for identifying interaction sites in proteins. J. Mol. Biol.
1999, 289 (4), 1093−108.
(20) Friesner, R. A.; Banks, J. L.; Murphy, R. B.; Halgren, T. A.; Klicic,
J. J.; Mainz, D. T.; Repasky, M. P.; Knoll, E. H.; Shelley, M.; Perry, J. K.;
Shaw, D. E.; Francis, P.; Shenkin, P. S. Glide: a new approach for rapid,
accurate docking and scoring. 1. Method and assessment of docking
accuracy. J. Med. Chem. 2004, 47, 1739.
(21) See www.cerep.fr for more details.
(22) Katori, M.; Majima, M. The renal kallikrein−kinin system: its role
as a safety valve for excess sodium intake, and its attenuation as a
possible etiologic factor in salt-sensitive hypertension. Crit. Rev. Clin.
Lab. Sci. 2003, 40, 43−115.
(23) Linz, W.; Schoelkens, B. A. A specific B2-bradykinin receptor
antagonist HOE 140 abolishes the antihypertrophic effect of ramipril.
Br. J. Pharmacol. 1992, 105, 771−772.
(24) Jackman, H. L.; Massad, M. G.; Sekosan, M.; Tan, F.;
Brovkovych, V.; Marcic, B. M.; Erdos, E. G. Angiotensin 1-9 and 1-7
̈
release in human heart. Role of cathepsin A. Hypertension 2002, 39,
976−981.
(25) Kitts, P. A.; Possee, R. D. A method for producing recombinant
baculovirus expression vectors at high frequency. BioTechniques 1993,
14, 810−817.
(26) Patent applications EP1144666, WO0112829, and AU6460800.
(27) Vaughn, J. L.; Goodwin, R. H.; Tompkins, G. J.; McCawley, P.
The establishment of two cell lines from the insect Spodoptera frugiperda
(Lepidoptera; Noctuidae). In Vitro 1977, 13, 213−7.
(28) Brown, M; Faulkner, P. A plaque assay for nuclear polyhedrosis
viruses using a solid overlay. J. Gen. Virol. 1977, 36, 361−364.
(29) Hochuli, E.; Bannwarth, W.; Dobeli, H.; Gentz, R.; Stuber, D.
̈
̈
Genetic approach to facilitate purification of recombinant proteins with
a novel metal chelate adsorbent. Nat. Biotechnol. 1988, 6, 1321−1325.
(30) Laemmli, U. K. Cleavage of structural proteins during assembly of
the head of bacteriophage T4. Nature 1970, 227, 680−685.
(31) Kabsch, W. XDS. Acta Crystallog. 2010, D66, 125−132.
(32) Kroemer, M.; Dreyer, M. K.; Wendt, K. U. APRVa program for
automated data processing, refinement and visualization. Acta
Crystallogr. 2004, D60, 1679−1682.
(33) McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P. D.; Winn, M.
D.; Storoni, L. D.; Read, R. J. Phaser crystallographic software. J. Appl.
Crystallog. 2007, 40, 658−674.
(34) Emsley, P.; Cowtan, K. Coot: model-building tools for molecular
graphics. Acta Crystallog D 2004, 60, 2126−2132.
(35) Bricogne, G.; Blanc, E.; Brandl, M.; Flensburg, C.; Keller, P.;
Paciorek, W.; Roversi, P; Sharff, A.; Smart, O. S.; Vonrhein, C.;
Womack, T. O. BUSTER; Global Phasing Ltd.: Cambridge, U.K., 2011.
(36) Shelley, J. C.; Choletti, A.; Frye, L. L.; Greenwood, J. R.; Timlin,
M. R.; Uchimaya, M. Epik: a software program for pKa prediction and
protonation state generation for drug-like molecules. J. Comput.-Aided
Mol. Des. 2007, 21, 681−691.
(37) Sybyl 8.0; Tripos (1699 South Hanley Road, St. Louis, MO, U.S.).
7649
dx.doi.org/10.1021/jm300663n | J. Med. Chem. 2012, 55, 7636−7649