Kinetic and structural characterization of caspase-3 and caspase-8 inhibition by a novel class of irreversible inhibitors

Article abstract of DOI:10.1016/j.bbapap.2010.05.007

Because of their central role in programmed cell death, the caspases are attractive targets for developing new therapeutics against cancer and autoimmunity, myocardial infarction and ischemic damage, and neurodegenerative diseases. We chose to target casp

Full text of DOI:10.1016/j.bbapap.2010.05.007

Biochimica et Biophysica Acta 1804 (2010) 1817–1831
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbapap
Kinetic and structural characterization of caspase-3 and caspase-8 inhibition by a
novel class of irreversible inhibitors
Zhigang Wang a, , William Watt b,1, Nathan A. Brooks b,2, Melissa S. Harris c, Jan Urban d, Douglas Boatman e,
⁎
Michael McMillan f, Michael Kahn f,g,h, Robert L. Heinrikson i, Barry C. Finzel b,3, Arthur J. Wittwer j,
,5
James Blinn j, Satwik Kamtekar j,4, Alfredo G. Tomasselli j,
⁎
a
Oligonucleotide Therapeutics Unit, Pfizer, Inc., 620 Memorial Drive, Cambridge, MA 02139, USA
Global Research and Development, Pfizer, Inc., 301 Henrietta Street, Kalamazoo, MI 49007, USA
Global Research and Development, Pfizer, Inc., 216 Eastern Pointe Rd, Groton, CT 06840, USA
8841 Helen James Avenue, San Diego, CA 92126, USA
b
c
d
e
Arena Pharmaceuticals, 6166 Nancy Ridge Dr., San Diego, CA 92121, USA
f
Eli and Edythe Broad Center for Regenerative Medicine and Stem Cell Research, University of Southern California, 1501 San Pablo St., Los Angeles, CA 90033, USA
Department of Biochemistry and Molecular Biology, Keck School of Medicine, University of Southern California, 1501 San Pablo St., Los Angeles, CA 90033, USA
Department of Molecular Pharmacology and Toxicology, School of Pharmacy, University of Southern California, 1501 San Pablo St., Los Angeles, CA 90033, USA
g
h
i
Proteos, Inc., 4717 Campus Drive, Kalamazoo, MI 49008, USA
j
Global Research and Development, Pfizer, Inc., 700 Chesterfield Parkway West, Chesterfield, MO 63017, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Because of their central role in programmed cell death, the caspases are attractive targets for developing new
therapeutics against cancer and autoimmunity, myocardial infarction and ischemic damage, and
neurodegenerative diseases. We chose to target caspase-3, an executioner caspase, and caspase-8, an
initiator caspase, based on the vast amount of information linking their functions to diseases. Through a
structure-based drug design approach, a number of novel β-strand peptidomimetic compounds were
synthesized. Kinetic studies of caspase-3 and caspase-8 inhibition were carried out with these urazole ring-
containing irreversible peptidomimetics and a known irreversible caspase inhibitor, Z-VAD-fmk. Using a
stopped-flow fluorescence assay, we were able to determine individual kinetic parameters of caspase-3 and
caspase-8 inhibition by these inhibitors. Z-VAD-fmk and the peptidomimetic inhibitors inhibit caspase-3 and
caspase-8 via a three-step kinetic mechanism. Inhibition of both caspase-3 and caspase-8 by Z-VAD-fmk and
of caspase-3 by the peptidomimetic inhibitors proceeds via two rapid equilibrium steps followed by a
relatively fast inactivation step. However, caspase-8 inhibition by the peptidomimetics goes through a rapid
equilibrium step, a slow-binding reversible step, and an extremely slow inactivation step. The crystal
structures of inhibitor complexes of caspases-3 and -8 validate the design of the inhibitors by illustrating in
detail how they mimic peptide substrates. One of the caspase-8 structures also shows binding at a secondary,
allosteric site, providing a possible route to the development of noncovalent small molecule modulators of
caspase activity.
Received 29 January 2010
Received in revised form 1 May 2010
Accepted 17 May 2010
Available online 24 May 2010
Keywords:
Caspase-3
Caspase-8
Peptidomimetic inhibitors
Irreversible inhibition
Urazole
Crystal structure
© 2010 Published by Elsevier B.V.
1. Introduction
Apoptosis, or programmed cell death [1], is an essential biological
process occurring in cells as a component of animal development,
tissue homeostasis, and immune responses [2–4]. Although the
apoptotic process is a normal and essential physiological event in
healthy organisms, in pathological states, it can be abrogated as in
cancer [5] and autoimmunity [6], or exacerbated as in stroke [7],
neurodegeneration [8], retinal cell death [9], myocardial and liver
ischemia [10,11], and inflammatory diseases such as sepsis [12],
osteoarthritis [13], rheumatoid arthritis [14], and asthma [15].
Promoting apoptosis in cancer [16] or attenuating it in stroke,
neurodegeneration, and inflammation [17–19] may be viable
Abbreviations: AMC, 7-amino-4-methyl-coumarin; DMSO, dimethyl sulfoxide; DTT,
dithiothreitol; fmk, flouromethyl ketone; MMX, Molecumetics Company, Ltd; RMSD,
root mean squared deviation
⁎ Corresponding authors.
E-mail addresses: zhigang.wang@pfizer.com (Z. Wang), atomasse@gmail.com
(A.G. Tomasselli).
1
Present address: 7262 Atlantic Ave., South Haven, MI 49090, USA.
Present address: Lilly Research Laboratories, Eli Lilly and Co., Indianapolis, Indiana
2
46285, USA.
3
Present address: Medicinal Chemistry Department, College of Pharmacy, University
of Minnesota, 308 Harvard St., S.E., Minneapolis, MN 55455, USA.
4
Present address: Pacific Biosciences, 1505 Adams Drive, Menlo Park, CA 94025, USA.
Present address: 1540 Garden Valley, Wildwood, MO 63038, USA.
5
1570-9639/$ – see front matter © 2010 Published by Elsevier B.V.
doi:10.1016/j.bbapap.2010.05.007

1818
Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
approaches to treat these devastating diseases, most of which
continue to defy therapeutic intervention.
injury, myocardial infarction, and stroke [17–19]. Pralnacasan (VX-
740), an inhibitor of caspase-1, reached clinical Phase II for
rheumatoid arthritis and osteoarthritis but was subsequently discon-
tinued due to liver problems with its prolonged use. VX-765 has
completed Phase IIa studies for psoriasis, but results have not been
reported [17,18]. IDN-6556, a peptidomimetic irreversible inhibitor of
caspase-1, -3, -6,-7,-8, and -9, was very effective in rodent model of
liver diseases and has reached Phase II in trials for patients with
chronic hepatitis C and trials for patients undergoing liver transplan-
tation [17,18].
Development of nonpeptide inhibitors is being pursued by various
pharmaceutical companies to circumvent poor membrane penetra-
tion, short half-life in vivo, and other undesirable pharmacokinetic
properties characteristic of peptide-based inhibitors [17–19]. For
most proteases, substrates and peptidomimetic inhibitors adopt a β-
strand conformation, and such is the case for the caspases [32,38]. The
recognition element for a caspase inhibitor, therefore, might mimic a
β-strand and include an aspartic acid or mimic thereof at the P1
position. In a search for nonpeptide chemical scaffolds that mimic
peptide secondary structures, Kim and Kahn [45] used solid-phase
organic synthesis to construct bicyclic templates reflecting an
extended β-strand conformation. This approach was previously
validated with the design, synthesis, and evaluation of serine
proteinase thrombin inhibitors with KI values in the subnanomolar
range [46,47]. We have investigated an alternate β-strand peptido-
mimetic scaffold of Kim and Kahn [45] incorporating a urazolopyr-
idazine template for structure-based design of caspase inhibitors.
Both types of inhibitors block the enzyme by forming a thiohemiketal
complex, but the irreversible inhibitors have a leaving group, 2,5-
dichlorobenzoate, and inactivate the enzyme permanently through its
expulsion (Scheme 1).
Here we report kinetic studies and structural characterization of
caspase-3 and caspase-8 inhibition by these novel β-strand
irreversible peptidomimetic inhibitors (hereafter referred to as
Compound-1, -2, -3, -4, -5, -6, -7, -8, and -9, or “MMX” inhibitors),
and by Z-VAD-fmk, a widely used irreversible caspase inhibitor. The
executioner caspase-3 and initiator caspase-8 were chosen because
they have similar active sites and common processing patterns of
activation, yet operate at distinct points in the apoptotic cascade.
Using a stopped-flow fluorescence assay, we were able to determine
the individual kinetic parameters of caspase-3 and caspase-
8 inhibition by Z-VAD-fmk and a number of irreversible peptido-
mimetic inhibitors. Structures of three complexes of caspase-3 and
caspase-8 with these novel peptidomimetic inhibitors have been
solved to high resolution. The kinetic and structural results are
discussed in light of the subtle difference between the two caspases
and their inhibition mechanisms.
Caspases, a family of related cysteinyl proteinases, mediate critical
steps of the apoptotic process [20–23]. These enzymes are exquisitely
specific molecular scissors that employ the cysteinyl thiol group as a
nucleophile to catalyze the cleavage of peptide bonds having an
aspartic acid at the P1 position. Fourteen caspases have been reported
to date and human orthologs have been identified for twelve of these
[24]. Caspases have been classified with respect to their functions into
three categories: (a) initiator caspases, (b) executioner caspases, and
(c) inflammatory or cytokine processing caspases. All caspases are
synthesized in cells as inactive zymogens that contain a prosegment, a
large N-terminal, and a smaller C-terminal segment that are processed
during enzyme activation to produce the A and B chains, respectively,
of the caspase heterodimer. The prosegment, large in the initiator
caspases and short in the executioner caspases, appears to direct the
course of activation and determine the location of the proenzyme in
the cell. The initiator procaspase-8 participates in the extrinsic
pathway that is set in motion by ligand-induced activation of the
death receptor at the plasma membrane and proceeds with
procaspase-8 recruitment. These events foster procaspase-8 dimer-
ization and activation in a mechanism independent of intersubunit
cleavage [22,25,26]. Another apoptotic pathway involving an initiator
caspase is the intrinsic pathway that is triggered by an insult to the
cell, e.g., to the mitochondria, leading to the release of cytochrome c
with the subsequent formation of a complex consisting of Apaf-1,
procaspase-9, and cytochrome c. This results in procaspase-9
activation and is also independent of intersubunit cleavage [27]. In
both pathways, the activated initiator caspases transduce the
apoptotic signal by proteolytically processing and activating the
downstream executioner procaspases, caspase-3, -6, and -7. The
executioner procaspases are constitutive dimers, and cleavage of
the intersubunit linkers by initiator caspases allows the formation of
catalytically competent enzyme heterodimers [22,26]. These, in turn,
lead to the degradation of cell cycle proteins, DNA repair enzymes,
and cytoskeleton proteins in the apoptotic process of programmed
cell death. The inflammatory procaspases undergo an activation
process similar to that of initiator procaspases to acquire the catalytic
activity for cytokine processing [25].
An enormous body of information has been generated for caspases,
and the X-ray structures of caspase-1 [28,29], caspase-2 [30], caspase-
3 [31–34], caspase-6 [35], caspase-7 [36,37], caspase-8 [38–40], and
caspase-9 [41] have been determined.
Intense efforts during the past 15 years to find drugs that inhibit
caspases have produced several clinical candidates, but no caspase
inhibitor has yet reached the market [17–19]. Because these enzymes
have shallow and widely dispersed active sites, it has been necessary
to design chemically reactive compounds to achieve potent inhibition
of enzyme activity. Generally speaking, caspase inhibitors are
composed of three parts: an active site recognition element, the P1
aspartic acid to confer caspase specificity, and a reactive electrophilic
warhead that is capable of forming a covalent linkage (reversible or
irreversible) with the nucleophilic active site thiol of the enzyme.
Electrophilic warheads include aldehydes, chloromethyl ketones [42],
epoxides [43], and vinyl sulphones [44], all of which react readily with
thiols. To deliver these warheads selectively to the thiol in the target
of interest, it is necessary to build in a recognition element that will
concentrate the inhibitor at the active site of the target. Many
inhibitor design strategies have been used to devise recognition
elements mostly by changing the P1–P3 amino acids in a tetrapeptide
of the natural caspase substrates with a myriad of different chemical
moieties. These efforts have resulted in the generation of inhibitors
with various degrees of peptidic features. Addition of warheads to
such peptide-based inhibitors of caspases has produced molecules
that were efficacious when tested in animal models representative of
various human diseases: they include liver diseases, traumatic brain
2. Materials and methods
Active and highly pure recombinant, human caspase-3 and
caspase-8 were obtained as previously described [48,49], although
in the present work, we have used the wild type sequence of caspase-
3 containing Asp190 and Trp206 instead of to Glu190 and Arg206
previously used [48]. Upon purification and activation, caspase-3
subunits A and B were processed to yield the sequences S29GIS…
IETD175 and S176GVD…YFYH277-(H)7, respectively (nomenclature of
Rank et al. [48]). On the other hand, the caspase-8 sequence is the one
expressed earlier [49]; upon refolding and purification, mature
caspase-8 yielded the sequences S18PRE…VETD181 and L192SSP…
FPSD286, for subunits A and B, respectively (nomenclature of
Koeplinger et al. [49]). Purified caspase-3 and purified and refolded
caspase-8 were about 90% and 85% active, respectively by active site
titration. Caspase substrates, Ac-DEVD-AMC and Ac-IETD-AMC, were
purchased from Peptide Institute, Inc. Z-VAD-fmk was purchased
from Bachem Bioscience, Inc. Substrates and inhibitors are dissolved

Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
1819
Scheme 1. β-strand peptidomimetic scaffold incorporating a urazolopyridazine template for structure-based design of caspase reversible and irreversible inhibitors.
in DMSO to make concentrated stock solutions. DTT-containing
buffers were made fresh every day by dissolving DTT in buffer stocks
without readjusting the pH. The buffers were then degassed and used
to dilute either the enzyme or substrate and inhibitor. All other
reagents were of the highest purity available from commercial
sources.
concentrated under vacuum. The residue was dissolved in ethyl
acetate and washed with saturated aqueous NaHCO3 and brine, then
dried over anhydrous magnesium sulfate, filtered, and concentrated
under vacuum to a brown oil. The oil was purified by flash
chromatography to give 21 g (96% yield) of the product as an off-
white solid. The solid was dissolved in 400 ml of 1:1 TFA/methylene
chloride and the solution was stirred for 3.5 h at ambient temperature
then concentrated under vacuum to 19.1 g of an off-white solid. This
product was used in subsequent steps without further purification.
2-Chlorotrityl chloride resin (3.7 g, 1.33 mmol/g) was shaken with
a solution of 5 ml of thionyl chloride in 95 ml of methylene chloride
for 30 min. The resin was filtered then washed with methylene
chloride (5×50 ml). The resin was taken up in methylene chloride
(40 ml), then the acid was added followed by diisopropylethylamine
and the mixture was shaken for 2.5 h. Methanol (10 ml) and
diisopropylethylamine were added, and shaking was continued for
30 min. The resin was filtered and washed with dimethylformamide
(5×30 ml), methylene chloride (5×30 ml), and diethyl ether
(3×30 ml) then dried in a vacuum desiccator. Fmoc analysis indicated
0.33 mmol/g attached amino acid. The resin was taken up in a
mixture of tetrahydrofuran, methylene chloride, and methanol (5:4:1,
40 ml), and sodium borohydride was added with vigorous gas
evolution. The mixture was shaken for 45 min then filtered, and the
resin was washed with DMF (5×20 ml), methanol (3×20 ml), and
methylene chloride (3×20 ml). The resin was shaken with 25%
piperidine in DMF (20 ml) for 5 min. then filtered and treated for
another 5 min with 25% piperidine in DMF. The resin was filtered and
washed with DMF (5×20 ml), methylene chloride (5×20 ml), and
diethyl ether (3×20 ml) then dried in a vacuum desiccator.
2.1. Synthesis of inhibitors
Scheme 2 provides a representation of inhibitor synthesis
following the procedures described in earlier works [50,51]. Isobutyl-
chloroformate (5.0 ml, 39 mmol) was added to a solution of N-Fmoc-
(O-t-Bu)-aspartic acid (14.5 g, 35.2 mmol) stirred in THF (350 ml) in a
flask cooled to −25 °C and followed by N-methylmorpholine (4.3 ml
39 mmol). The resulting mixture was stirred for 10 min, and then a
solution of diazomethane in diethyl ether (140 ml, ca. 1 mmol/ml)
was added. The solution was kept stirring in the cold bath for 1 h, then
allowed to warm to room temperature and stirred an additional 1.5 h.
The solution was diluted with ethyl acetate and washed with
saturated aqueous NaHCO3 solution, water, and saturated aqueous
NaCl solution. The organic solution was dried over anhydrous sodium
sulfate, filtered, and concentrated under vacuum to a pale yellow
foam. The foam was dissolved in THF (250 ml) and the flask was
immersed in a dry ice/acetone bath. Hydrogen bromide (4.6 ml, 48%
aqueous) was added slowly and the solution was stirred for 30 min. in
the dry ice/acetone bath, then for 1.5 h in an ice/water bath. The
solution was concentrated on a rotary evaporator to ca. 100 ml then
diluted with ethyl acetate and washed with saturated aqueous
NaHCO3 solution, water and saturated aqueous NaCl solution. The
organic layer was dried over anhydrous magnesium sulfate, filtered,
and concentrated under vacuum to give 20.6 g of the product as a pale
viscous oil. The crude product was dissolved in DMF (375 ml), and
2,6-dichlorobenzoic acid was added followed by solid NaHCO3 (4.23 g,
50.4 mmol) and potassium iodide (7.66 g, 46.1 mmol). The mixture
was stirred under an anhydrous nitrogen atmosphere for 16 h then
The resin was divided into portions (50 mg, ca. 0.22 mmol), and
the appropriate dienoic acid (0.6 mmol) was added. The mixture was
then treated with a solution of 0.6 M PyBOP, 0.6 M HOBt, and 1.2 M
diisopropylethylamine in DMF (1 ml). The mixture was shaken until
Kaiser analysis of a resin sample indicated complete conversion of the
amine (1 h) when the resins were rinsed with DMF (3×1 ml) and

1820
Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
Scheme 2. Representation of inhibitor synthesis.
methylene chloride (3×1 ml). Each resin was treated with 1 ml of
a freshly prepared solution resulting from combining the appro-
priate urazole (0.5 mmol) with bis(trifluoroacetoxy)iodobenzene
(0.5 mmol) in DMF (10 ml). These mixtures were shaken for 1 h,
then the resins were washed with DMF (4×1 ml) and methylene
chloride (3×1 ml). The resins were then treated with 1 ml of a 0.7 M
solution of Dess–Martin periodinane in methylene chloride and DMSO
(1:1) for 3 h. The resins were filtered and washed with DMF (4×1 ml)
and methylene chloride (3×1 ml). Each resin was treated for 30 min
with 1 ml of 95% TFA/methylene chloride (3:1) and then washed
twice with 95% TFA/methylene chloride. The combined filtrates were
concentrated under vacuum then the residue was dissolved in acetic
acid and lyophilized to give the product as a powder. All inhibitors
employed in the present study were N95% pure as judged by HPLC/MS.
containing 0.001% Tween-20, 2% DMSO, and 50 mM DTT at pH 7.5
(for caspase-3) and pH 7.2 (for caspase-8) at 25 °C. Ac-DEVD-AMC
and Ac-IETD-AMC were employed as substrates of caspase-3 and
caspase-8, respectively. Upon cleavage, both substrates released the
fluorescent AMC, which was then measured by its fluorescence
emission at 450 nm. Fluorescence measurements employed an
excitation wavelength that was set at 350 nm with a monochromator,
and the emission wavelength was set by using a 450 10 nm band-
pass filter mounted on the fluorescence detector. The enzyme,
substrate, and inhibitor concentrations were chosen so that all the
reactions were under steady-state conditions. For the KM measure-
ments of both caspase-3 and caspase-8, the final reaction mixtures
contained 10 nM enzyme. Caspase-3 inhibition experiments were
performed at 10 nM enzyme and 50 μM substrate whereas caspase-
8 inhibition experiments were run at 10 nM enzyme and 25 μM
substrate. In a typical reaction, equal volumes (about 50 μl each) of
the 2× substrate, inhibitor, and DMSO mixture and 2× enzyme in the
reaction buffer were mixed rapidly on the stopped-flow apparatus.
The reaction progress curve, i.e., the fluorescence emission intensity-
2.2. Stopped-flow kinetic experiments
All kinetic experiments were carried out on a Hi-Tech SF-61DX2
stopped-flow system in 100 mM NaPO4/10 mM Tris–HCl buffer

Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
1821
versus-time curve, was then monitored up to 4 min. A minimum of
three curves were averaged for each substrate/inhibitor concentra-
tion, with at least six concentrations examined for each inhibitor.
Scheme 4. A two-step slow binding reversible inhibition mechanism.
2.3. Kinetic data analysis
The initial velocities (V) of the caspase-3 and caspase-8 catalyzed
reactions were calculated from the slopes of the linear initial progress
curves. KM values of both enzymes toward their corresponding
substrates were then determined by nonlinear least squares regres-
sion fittings of the initial velocity (V)-versus-substrate concentration
([S]) curves according to the Michaelis–Menten equation (Eq. (1))
using the program Prism (GraphPad).
Crystals of caspase-3 in the presence of Compound-1 were grown by
vapor diffusion at 4 °C from 5-μl drops containing equal volumes of
16% ethanol, 0.1 M Tris buffer at pH 7.8 and the protein complex
solution.
2.5. Crystallization of caspase-8/inhibitor complexes
About 2.0 μl of 100-mM stocks of MMX inhibitors (Compound-4 or
Compound-9) dissolved in DMSO were added to 100-μl aliquots of
8.4 mg/ml caspase-8 (0.30 mM) in 20 mM Tris, 100 mM DTT, pH 8.0.
The protein/inhibitor solutions were incubated at 4 °C for 60 minutes.
The solutions were then centrifuged at 3000 rpm for 3 minutes before
setting up the sitting drops. Drops of 2.5 μl were mixed with an equal
volume of reservoir solution [1.0–1.1 M citrate, 0.1 M HEPES (or
PIPES), pH 6.5] and incubated at 4°C. The average size of the crystals
was 0.2×0.2×0.3 mm.
Vmax½Sꢀ
V = K
:
ð1Þ
+ ½Sꢀ
M
Progress curves consistent with first-order inhibition kinetics
(two-step irreversible inhibition mechanism; Scheme 3) were fit
according to Eq. (2) by nonlinear least squares regressions to
determine kobs values at each inhibitor concentrations:
t
ð1−e−k
Þ
;
obs
Ft = F0 + V0
ð2Þ
kobs
2.6. Crystallographic data collection and refinement
where Ft is the fluorescence intensity at time t, F0 is the background
fluorescence intensity, V0 is a constant related to the steady-state rate
(in fluorescence units) of the uninhibited reaction, and kobs is the
observed rate constant of enzyme inactivation. The inhibition or
dissociation constant of the enzyme/inhibitor complex (KI) and the
first-order rate constant of inactivation at infinite inhibitor concen-
tration (kinact) values were then obtained by fitting the kobs-versus-[I]
curves according to Eq. (3),
The caspase-3 crystals were flash frozen directly in liquid nitrogen.
The caspase-8 crystals were slowly introduced to cryoprotectant
before freezing (final solution: 1.0 M citrate, 50 mM HEPES (or PIPES),
pH 6.5, and 25 mM DTT). The data sets were collected at the Advanced
Photon Source (APS; Argonne National Laboratory, Argonne, IL)
IMCA-CAT beamlines and scaled with the HKL2000 software package
[52]. The data collection and refinement statistics are given in Table 2.
The structures of the complexes were solved by molecular replace-
ment using the program AMoRe in the CCP4 program suite [53,54].
The caspase-3 complexes were solved with a search model based on
the published caspase-3 (1CP3 in the Brookhaven Data Bank)
structure [32]. Some of the caspase-8 complexes were solved with a
search model based on the caspase-8/Ac-IETD-H structure [38].
Refinements were carried out by alternating rounds of refinement
using SHELXTL97 [55], CNX, 2002 (Accelrys, San Diego, CA) or
REFMAC [53,56], with manual rebuilding using CHAIN, LORE, or
COOT [57–59]. Parameter files generated using the software package
Afitt (Openeye Scientific Software, Santa Fe, NM) were manually
edited prior to use. All structures were checked for integrity using
PROCHECK [60].
kinact½Iꢀ
½Iꢀ + KI 1 +
ꢀ
ꢁ
kobs
=
:
ð3Þ
½Sꢀ
KM
Progress curves consistent with first-order, followed by a steady-
state phase kinetics (slow-binding reversible inhibition mechanism;
Scheme 4) were fit according to Eq. (4) by nonlinear least squares
regressions to determine kobs values:
t
ð1−e−k
Þ
+ Vst;
obs
Ft = F0 + ðV0−VsÞ
ð4Þ
kobs
where Ft, F0, and V0 have the same meanings as in Eq. (2) and Vs is the
steady-state rate of the inhibited reactions. KI and k2 and k−2 values
were then obtained by fitting the kobs-versus-[I] curves according to
Eq. (5),
3. Results
3.1. Steady-state kinetic parameters (KM)
k2½Iꢀ
½Iꢀ + KI 1 +
ꢀ
ꢁ
kobs
=
+ k−2
:
ð5Þ
Steady-state kinetic analyses of caspase-3- and caspase-8-
catalyzed reactions in the absence of inhibitors were first carried
out to determine KM values for both enzymes toward their
respective substrates. Under the experimental conditions, both
enzymes followed Michaelis–Menten kinetics (data not shown). KM
values of 33.7 3.9 μM and 4.52 0.47 μM were obtained for the
caspase-3 and caspase-8 reactions, respectively.
½Sꢀ
KM
2.4. Crystallization of caspase-3/Compound-1 complex
About 2 μl of a 100-mM DMSO stock solution of Compound-1 was
added to 100 μl of caspase-3 (6.5 mg/mL in 20 mM Tris, 10 mM DTT,
pH 8.0) and incubated at 4 °C for 60 minutes. The solution was
centrifuged at 3000 rpm for 3 minutes before setting up sitting drops.
3.2. Stopped-flow kinetics of caspase inhibition by Z-VAD-fmk
Caspase-3-catalyzed Ac-DEVD-AMC hydrolysis reactions were
measured in the presence of 0, 1, 2, 5, 10, 20, 40, 80, 150, and 300 μM
Z-VAD-fmk (Fig. 1) and a series of progress curves (fluorescence
intensity vs. time) were obtained (Fig. 2A). It was noted that the
progress curve at 0 μM Z-VAD-fmk was slightly curved. Simulation of
caspase-3 reaction progress using the determined kinetic parameters
Scheme 3. A two-step irreversible inhibition mechanism.

1822
Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
Fig. 1. Chemical structures of selected peptidomimetic caspase inhibitors.

Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
1823
3.3. Caspase inhibition by the MMX irreversible inhibitors
The chemical structures of selected irreversible inhibitors and their
inhibition kinetic parameters against caspase-3 and caspase-8 are
summarized in Fig. 1 and Table 1, respectively. When the caspase-3-
catalyzed Ac-DEVD-AMC hydrolysis reactions were measured in the
presence of 0, 0.2, 0.5, 1, 2, 5, 10, and 20 μM Compound-6, progress
curves similar to those in the presence of Z-VAD-fmk were obtained
(Fig. 4A). The first-order progress curves fit Eq. (2) well (Fig. 4A), and
kobs values at different Compound-6 concentrations were obtained
(Fig. 4B). KI of 13.4 3.7 μM and kinact of 0.094 0.018 s−1 were
calculated by fitting the kobs-versus-Compound-6 concentration curve
according to Eq. (3) (Fig. 4B).
Similar experiments were carried out on caspase-8 inhibition by
Compound-6, and the progress curves at different Compound-6
concentrations are shown in Fig. 5A. The linear progress curve at 0 μM
Compound-6 concentration again confirms the stability of caspase-8 at
low concentration. However, none of the progress curves obtained in
the presence of the inhibitor is first-order. They all exhibit an initial
“burst” phase followed by a steady-state increase of the signal. This
behavior is consistent with a slow-binding reversible inhibition model
(Scheme 4). In fact, all the progress curves can be fit nicely according to
Eq. (4) (Fig. 5A) to generate an array of kobs values (Fig. 5B). The kobs
-
versus-Compound-6 concentration curve exhibited saturation behavior
at high inhibitor concentrations and KI, k2, and k−2 values were
determined unambiguously using Eq. (5) (Fig. 5B). The kinetic
parameters of caspase-8 inhibition by Compound-6 are as follows:
KI=9.27 1.03 μM, k2=0.18 0.01 s−1, and k−2=0.043 0.007 s−1
.
Inhibition of caspases-3 and -8 by more than three dozen other
irreversible MMX inhibitors was also examined using the stopped-
flow technique. It was found that all the compounds behave like
Compound-6, i.e., they act like two-step irreversible inhibitors for
caspase-3 and slow-binding reversible inhibitors for caspase-8.
Fig. 2. Caspase-3 inhibition by Z-VAD-fmk. A. Progress curves of Ac-DEVD-AMC
hydrolysis catalyzed by caspase-3 in the presence of various concentrations of Z-VAD-
fmk as indicated. The data were obtained under the conditions described in the
Materials and methods section (50 μM substrate and 10 nM enzyme) and analyzed
according to Eq. (2). B. Values of kobs obtained from panel A plotted versus Z-VAD-fmk
concentrations. The plot was fit according to Eq. (3) to generate KI and kinact values.
3.4. Caspase/inhibitor co-crystal structures
We have determined the structures of three caspase bound
peptidomimetic inhibitors. Compound-1 bound to caspase-3 crystal-
lized in space group P21212 and yielded data to a resolution of 2.0 Å.
Compound-4 and Compound-9 bound to caspase-8 crystallized in a
trigonal crystal form (P3121) previously reported by Watt et al. [38]
and yielded data sets to 1.8 Å resolution. The structures have been
refined to crystallographic R values of 17.0% to 18.6% (Rfree 21.0–
22.6%). Compound structures are illustrated in Fig. 6 with R1, R2, and
R3 representing substitution groups at the 1, 4, and 5 positions,
respectively. Summaries of the diffraction data and crystallographic
refinement statistics are presented in Table 2.
These structures agree well with previously determined caspase-3
and caspase-8 structures. Calculated over all pairs of Cα atoms in the
structures, the RMSD between the caspase-3 structure described here
and that of an Ac-DVAD caspase-3 complex [32] is 0.3 Å. The similarly
calculated RMSDs between the Compound-4 and Ac-IETD [38], or
Compound-9 and Ac-IETD bound caspase-8 structures are 0.3 and
0.2 Å, respectively. The inhibitors described here thus do not induce
global, novel protein conformational changes.
The interpretation of the electron density maps and analysis of
bound ligand geometry was aided by the high resolution of the data
(Fig. 7 and Table 2). Given the long timescale of crystallographic
experiments, it is unsurprising that the reactions proceeded through
to the irreversible loss of the dichlorobenzoate moiety. There is thus
no evidence of the 2,5-dichlorobenzoate leaving group in any of the
electron density maps. For the caspase-3 complex structure, and the
structure of Compound-9 bound to caspase-8, all other inhibitor
atoms are visible in the electron density. In the Compound-4/caspase-
8 structure, although the urazolopyridazine core is well ordered, the
R3 group (a 4-chlorophenyl acetic acid amide methylene moiety) is
with Program DYNAFIT [61] ruled out measurable substrate depletion
during the 240-second data acquisition period, suggesting that
caspase-3 undergoes autoinactivation at a low concentration. The
progress curves at different Z-VAD-fmk concentrations appeared to be
first order, consistent with an irreversible inhibition mechanism
(Scheme 3). Eq. (2) was successfully fit to these curves by nonlinear
least squares regression (Fig. 2A), and kobs values at different inhibitor
concentrations were obtained (Fig. 2B). The kobs-versus-inhibitor
curve approaches saturation at high inhibitor concentrations. This
allows the KI and kinact values of caspase-3 inhibition by Z-VAD-fmk to
be determined unambiguously by fitting the data according to Eq. (3)
(Fig. 2B). The KI and kinact of the inhibition of caspase-3 by Z-VAD-fmk
are 18.4 1.4 μM and 0.078 0.002 s−1, respectively (Table 1).
Inhibition studies of caspase-8 by Z-VAD-fmk were also carried out
and the progress curves obtained at different inhibitor concentrations
are shown in Fig. 3A. The linear progress curve at 0 μM inhibitor
demonstrated that caspase-8 is stable at low concentration. Similar to
the caspase-3 study, the progress curves in the presence of Z-VAD-
fmk are first order. However, much less inhibitor was required to
inhibit caspase-8 to the same extent as caspase-3. This suggests that
Z-VAD-fmk is a more potent inhibitor for caspase-8. Fitting Eq. (2) to
the progress curves (Fig. 3A), a kobs-versus-inhibition concentration
curve was obtained (Fig. 3B). This curve shows saturable behavior at
high Z-VAD-fmk concentrations and KI of 0.45 0.05 μM and kinact of
0.23 0.01 s−1 were determined by fitting the data according to
Eq. (3) (Fig. 3B).

1824
Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
Table 1
Inhibition constants for caspases-3 and -8 by selected inhibitors.
Caspase-3
Inhibitors
KI (μM)
k2 (s−1
)
k−2 (s−1
)
kinact (s−1
)
kinact/KI (M−1 s−1
)
Z-VAD-fmk
18.4 1.4
13.4 3.7
9.13 1.01
0.66 0.02
5.05 0.32
5.24 0.39
2.03 0.25
6.81 1.20
7.54 0.68
17.2 1.6
N/Aa
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0.078 0.002
0.094 0.018
0.082 0.004
0.058 0.001
0.082 0.002
0.074 0.002
0.067 0.003
0.086 0.007
0.026 0.001
0.023 0.001
4,200
7,000
9,000
88,000
16,000
14,000
33,000
13,000
3,400
Compound-1
Compound-2
Compound-3
Compound-4
Compound-5
Compound-6
Compound-7
Compound-8
Compound-9
1,300
Caspase-8
Inhibitors
KI (μM)
k2 (s−1
)
k−2 (s−1
)
kinact (s−1
)
KI[k−2 /(k−2 +k2)] (μM)
Z-VAD-fmk
0.45 0.03
9.27 1.03
0.33 0.13
3.54 0.88
1.32 0.12
1.52 0.54
3.63 0.15
3.10 0.22
3.57 0.33
6.44 0.29
N/A
N/A
0.004 0.001
b0.009
0.011 0.002
0.007 0.002
b0.009
0.008 0.001
0.012 0.002
b0.003
0.23 0.01
E/Sc
E/S
E/S
E/S
E/S
E/S
E/S
E/S
N/Ab
Compound-1
Compound-2
Compound-3
Compound-4
Compound-5
Compound-6
Compound-7
Compound-8
Compound-9
0.18 0.01
0.16 0.01
0.45 0.08
0.18 0.01
0.19 0.02
0.14 0.01
0.56 0.02
0.21 0.01
0.54 0.02
0.202
b0.018
0.084
b0.049
0.069
0.196
0.065
b0.050
0.047
0.004 0.001
E/S
a
Not available.
b
kinact/KI =510,000 M−1 s−1
.
c
Extremely slow (b0.005 s−1).
not visible and presumably disordered. In all three structures, the
electron density is very clear for the catalytic cysteine sulfur atom and
for the proximal ketone moiety. There is continuous electron density
connecting the ketone with the sulfur atom, confirming the existence
of a covalent bond. We have placed the methylene carbon atom
linking the ketone to the sulfur in a position consistent with geometric
constraints, although the electron density is perhaps somewhat
weaker than might be expected. It is unlikely that the methylene
group would be more mobile than its neighboring atoms, and so the
weaker electron density is most parsimoniously explained by a
limited degree of chemical heterogeneity in this region of the
structure. Radiation decay due to X-ray exposure during data
collection has previously been shown to be responsible for the
modification of a thiomethylketone in the active site of a caspase-3
structure determined at 1.06 Å resolution [62]. A similar phenomenon
may be occurring in the structures described here. The identities of
the alternate species could not be unambiguously determined either
in the atomic resolution structure of capsase-3 or in those described
here.
inhibitors bind to caspase-3 and caspase-8 in similar ways. These
interactions closely mimic those of caspases with peptidic sub-
strates (Fig. 8).
Beyond these core regions, the different MMX inhibitors mimic
different elements of caspase peptide substrates. For example, the
butyrate moiety in Compound-4 superimposes well on the P3
glutamic acid of the Ac-IETD caspase-8 structure (Fig. 8). Indeed,
both of these moieties hydrogen bond with their neighboring waters
in an identical fashion. In contrast, Compound-9 lacks a glutamic acid
mimic, but has an amide nitrogen in the R1 position where it forms a
hydrogen bond with Arg413. Compound-9 also forms a hydrogen bond
between an oxygen atom from the urazolopyridazine moiety and the
amide nitrogen of Arg413. It thus matches the antiparallel backbone
hydrogen bonding interactions observed between the P3 glutamic
acid and Arg413 in the Ac-IETD caspase-8 structure. A similar pattern
of interactions is apparent when the structures of Compound-1 and
Ac-DVAD in caspase-3 are compared. Of course, in addition to
mimicking peptide substrates, the MMX inhibitors also form novel
interactions with the caspases, such as those of the benzimidazole
moiety of Compound-9 with caspase-8.
3.5. Caspase/inhibitor interactions
3.6. Compound-4 also binding to caspase-8 in a pocket far from the
active site
The inhibitors share a number of common interactions with
caspase-3 and -8 in addition to their covalent linkage to the catalytic
cysteine. In all three adducts the “warhead” ketone oxygen
hydrogen bonds with the backbone amide NH of both the catalytic
cysteine and conserved glycine (caspase-3 Gly122, caspase-8 Gly318).
Similarly, the P1 carboxylate moiety forms hydrogen bonds with the
side chains of three conserved residues (caspase-3 Arg64, caspase-
In addition to covalent binding at the active site, Compound-4 also
binds noncovalently close to the interface between two p18–p11
heterodimers (Fig. 9). One ring of the urazolopyridazine and the R3
group that emerges from this ring (a 4-chlorophenyl acetic acid amide
ethylene moiety) are clearly visible in the electron density. The
urazolopyridazine binds to a relatively flat and featureless surface of
the protein. The oxygen of the amide forms a hydrogen bond with the
side chain of Glu396 while the nitrogen forms a hydrogen bond with
the side chain of Thr337. The chlorophenyl moiety occupies a well-
defined, hydrophobic pocket, stacking on Phe399 and forming an edge
8 Arg260; caspase-3 Arg207, caspase-8 Arg413; caspase-3 Gln161
,
caspase-8 Gln358). The adjacent amide nitrogen hydrogen bonds
with a main chain carbonyl (caspase-3 Ser205, caspase-8 Ser411) as
well as forming a close contact with the catalytic cysteine sulfur
atom (3.0 Å). One of the carbonyls of the urazolopyridazine forms a
hydrogen bond with a backbone main chain amide (caspase-3
Arg207, caspase-8 Arg413). As can be surmised from this enumeration
of interactions, the structures show that the core elements of the
to face interaction with Tyr334
.
When compared with other caspase-8 structures, it is apparent
that the presence of Compound-4 appears to induce a conformational

Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
1825
Fig. 4. Caspase-3 inhibition by Compound-6. A. Progress curves of Ac-DEVD-AMC
hydrolysis catalyzed by caspase-3 in the presence of various concentrations of
Compound-6 as indicated. The data were obtained under the conditions described in
the Materials and methods section (50 μM substrate and 10 nM enzyme) and analyzed
according to Eq. (2). B. Values of kobs obtained from panel A plotted versus Compound-6
concentrations. The plot was fit according to Eq. (3) to generate KI and kinact values.
Fig. 3. Caspase-8 inhibition by Z-VAD-fmk. A. Progress curves of Ac-IETD-AMC
hydrolysis catalyzed by caspase-8 in the presence of various concentrations of Z-
VAD-fmk as indicated. The data were obtained under the conditions described in the
Materials and methods section (25 μM substrate and 10 nM enzyme) and analyzed
according to Eq. (2). B. Values of kobs obtained from panel A plotted versus Z-VAD-fmk
concentrations. The plot was fit according to Eq. (3) to generate KI and kinact values.
of potent irreversible peptidomimetic inhibitors of caspase-3 and
caspase-8. Kinetic analysis of irreversible enzyme inhibition is
complicated by the fact that it is usually a multi-step process; kinetic
parameters of each step are required to describe the inhibition kinetic
mechanism. Previous kinetic studies of irreversible caspase-3 and
caspase-8 inhibitors failed to determine the individual kinetic
parameters; only kinact/KI values were obtained. As we observed
from this work, the inhibition processes of caspase-3 and caspase-8 by
irreversible inhibitors were usually fast at high inhibitor concentra-
tions so that after a very short time, virtually no substrate was turning
over. The consequence is that the product release progress curve
exhibited an extremely short first-order portion before becoming
saturated (irreversible) or entering the linear steady-state stage
(reversible). This makes the determination of the kobs values at high
inhibitor concentrations extremely difficult. Without kobs values at
high inhibitor concentrations, it is impossible to determine each
individual inhibition kinetic parameter (KI and kinact in the two-step
irreversible and KI, k2, and k−2 in the slow-binding reversible
mechanism described in Schemes 3 and 4, respectively) by nonlinear
least squares regression fittings using Eq. (3) or (5). In this work, we
employed a stopped-flow apparatus to acquire the progress curves so
that the kobs values at high inhibitor concentrations were determined
accurately. Our stopped-flow system generated reliable data for AMC
release from both the caspase-3- and caspase-8-catalyzed reactions in
the presence of Z-VAD-fmk and the irreversible MMX inhibitors. The
mathematical models fit the corresponding progress curves well and a
change in this pocket. In the structure of caspase-8 bound to Ac-IETD-
aldehyde [38], the pocket is occupied by an oxidized molecule of DTT,
and Tyr334 is in an alternate rotomeric state. This tyrosine is also in
this alternate rotomeric state in the structure of caspase-8 bound to
Compound-9, where the pocket is occupied by a reduced DTT
molecule. This alternate tyrosine rotomer is incompatible with the
binding of Compound-4 as the tyrosine would clash with the
urazolopyridazine moiety.
4. Discussion
This study presents a detailed kinetic analysis of the inhibition of
caspase-3 and caspase-8 by a series of irreversible inhibitors that form
thioether bonds with the active site thiol. Moreover, the structures of
three of the resulting covalently modified caspase derivatives have
been solved by X-ray crystallography, providing a glimpse into the
catalytic sites of these important enzymes.
4.1. Stopped-flow kinetic studies of caspase inhibition by irreversible
inhibitors
To achieve potent inhibition of cysteinyl proteases, it has been
necessary to make inhibitors containing an electrophilic group to
enable the formation of a covalent bond with the active site thiol.
Here, this principle has been successfully applied to the development

1826
Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
Table 2
Summary of crystallographic statistics.
Protein
Caspase-8
Caspase-8
Caspase-3
Compound
Compound-9
Compound-4
Compound-1
Space group
Unit cell
dimensions
No. of molecules
in A.S.
P3121
P3121
P21212
62.6, 62.6, 129.39,
90.0, 90.0, 120.0
1
64.0, 64.0, 130.9,
90.0, 90.0, 120.0
1
96.69, 68.2, 44.8,
90.0, 90.0, 90.0
1
Resolution
I/sigma
Completeness
rmsd bond
length,
20.0–1.8 (1.85–1.80) 20.0–1.8 (1.85–1.80) 20–2.0 (2.05–2.0)
13.8 (2.9)
95.7 (81.7)
0.007
13.8 (2.9)
99.6 (96.9)
0.009
19.7 (4.3)
89.2 (60.2)
0.009
Rmsd angles
Rcryst
Rfree
PDB ID code
Ramachandran
Core
Allowed
Generous
Disallowed
1.20
1.24
1.35
17.0 (21.0)
20.2 (21.8)
3KJN
18.5 (24.1)
20.8 (29.3)
3KJQ
18.3 (23.8)
20.6 (23.7)
3KJF
89.5
10.5
0
89.5
10.5
0
91.7
8.3
0
0
0
0
The values in parentheses are for the high-resolution shell.
tion by Compound-6 and other irreversible peptidomimetic inhibi-
tors; Table 1).
The unambiguous determination of these parameters makes it
possible for us to compare the potencies of irreversible caspase
inhibitors in ways that have not previously been possible. For
example, Table 1 includes the calculation of kinact/KI for caspase-3
inhibitors and KI[k−2 /(k−2 +k2)] for caspase-8 inhibition. These
values can be viewed as the overall second-order rate constant for the
irreversible inhibition of caspase-3 and the overall Ki for the reversible
phase of caspase-8 inhibition, respectively. It is of interest that for
Compounds 1–9 kinact/KI for caspase-3 inhibition correlates very well
with 1/KI (R2 =0.9865). There is no correlation between kinact/KI and
kinact (R2 =0.0007). This suggests that the overall potency of these
compounds for caspase-3 is driven primarily by the rapidly reversible
affinity component, KI, instead of by differences in kinact. Examination
of Table 1 confirms that while kinact/KI varies 66-fold for these
compounds, kinact varies only 4-fold, but KI varies 23-fold. A similar
analysis of caspase-8 inhibition by these compounds is not possible,
Fig. 5. Caspase-8 inhibition by Compound-6. A. Progress curves of Ac-IETD-AMC
hydrolysis catalyzed by caspase-8 in the presence of various concentrations of
Compound-6 as indicated. The data were obtained under the conditions described in
the Materials and methods section (25 μM substrate and 10 nM enzyme) and analyzed
according to Eq. (4). B. Values of kobs obtained from panel A plotted versus Compound-6
concentrations. The plot was fit according to Eq. (5) to generate KI, k2, and k−2 values.
saturable kobs-versus-inhibition concentration ([I]) curve was
obtained for each enzyme-inhibitor pair. Fitting kobs- versus-[I] curves
according to the relevant equations, we were able to obtain all the
individual kinetic parameters involved in two-step irreversible
inhibition (caspase-3 and -8 inhibition by Z-VAD-fmk and caspase-3
inhibition by Compound-6 and other irreversible peptidomimetic
inhibitors) and slow-binding reversible inhibition (caspase-8 inhibi-
Fig. 7. Electron density map for Compound-9 bound to caspase-8. The gray mesh is 3
standard deviations above the mean contour in a weighted Fo–Fc electron map phased
using a model with inhibitor omitted and calculated from 20.0 to 1.8 Å resolution. The
catalytic cysteine is shown using spheres. Protein main chain is indicated using a green
ribbon trace. Figs. 6–8 were made using PyMol Molecular Graphics System (DeLano
Scientific, San Carlos, CA).
Fig. 6. Representation of the substitution groups, R1, R2, and R3, in MMX compounds,
especially the carbocyclic fusion of R2 and R2 utilized in inhibitors such as Compound-1
and Compound-6.

Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
1827
Fig. 8. Comparison of MMX and peptidic inhibitors. The peptide mimicry of the MMX inhibitors is illustrated by superposing the caspase Cα atoms in peptidic inhibitor- and MMX-
containing structures. The ribbon trace and protein surface of the peptidic inhibitor complexes are shown in gray, and the inhibitors themselves are shown in white and colored ball-
and-stick representation. The catalytic cysteines are shown as spheres, and the MMX inhibitors are shown in colored ball-and-stick representation. The chemical diagrams indicate
hydrogen bonding interactions (dashed lines). A disordered 4-chlorophenyl acetic acid amide methylene moiety in Compound-4 is shown in the diagram in gray. Hydrogen bonds
with water have been omitted for clarity. The P3 carboxylate of Ac-IETD forms two water interactions with Asn261 and Asp259 in caspase-8; the identical interactions are formed by
the butyrate moiety of Compound-4.
since kinact/KI is not known. However, if the structure–activity
relationships for caspase-3 and caspase-8 inhibition were similar,
one might expect either 1/KI or 1/KI([k−2 /(k−2 +k2)] for caspase-
8 inhibition to correlate with kinact/KI for caspase-3. These correlations
are very poor, however (R2 =0.0259 and 0.0600, respectively),
suggesting that Compounds 1–9 may have different structure–activity

1828
Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
Fig. 9. Inhibitor binding at a secondary pocket in caspase-8. An overall view of the biological dimer of heterodimers of caspase-8 with Compound-4 is shown on the left. A two-fold
symmetry axis projecting out of the plane of the page is indicated by a blue oval. Beneath the surface, one heterodimer is shown with a cyan cartoon, and the other with a green
cartoon. Ligands bound at the active site and secondary binding sites are shown in sphere representation. A zoomed-in view of one of the secondary binding sites is shown on the
right. The ligand is shown in stick representation, with carbon atoms colored yellow. Residues that form the binding pocket are shown as sticks and transparent spheres with carbon
atoms colored green. The heterodimer subunits are colored in the same way as in the overall view on the left. Hydrogen bonds are indicated by dashed red lines.
relationships for caspase-3 and 8 inhibition and that selective
inhibition of these enzymes by this class of inhibitors may be possible.
Indeed, Table 1 suggests that Compound-3 is the most potent inhibitor
of caspase-3, but one of the three weakest inhibitors of caspase-8, while
the most potent inhibitor of caspase-8 (Compound-2) is among the
four weakest inhibitors of caspase-3.
pathways shown in Scheme 5. Both mechanisms consist of the rapid
formation of a noncovalent enzyme-inhibitor (Michaelis) complex
(E·I) and the possibility of the formation of a thiohemiketal
intermediate (E-I). They differ in the route by which the cysteine
residue of the enzyme active site reacts to form a thioether product
(EI*) with the concomitant release of the leaving group (fluoride from
Z-VAD-fmk and dichlorobenzoate from the peptidomimetic inhibi-
tors). In one case (path B in Scheme 5), the Michaelis complex leads to
direct attack of the methylene carbon by the active site cysteine
residue to form the thioether product (EI*). In the other case (path A
in Scheme 5), the thiohemiketal is the immediate precursor to
this final product, either directly as shown or via a thiiranium ion
4.2. Kinetic mechanism of caspase inhibition by irreversible inhibitors
Z-VAD-fmk is a widely used irreversible caspase inhibitor. Like
Compound-6 and the other irreversible peptidomimetic inhibitors,
the irreversible modification could take place via one of the two
Scheme 5. A three-step irreversible inhibition mechanism.

Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
1829
intermediate. We cannot rule out either mechanism with just the
kinetic data, and to our knowledge, there is no literature precedent to
favor one over the other. Given that either of these pathways may
be possible, the question arises: why do we observe only two-step
inhibitory kinetic mechanisms? More specifically, why does Z-VAD-
fmk inhibition of both caspases follow the two-step irreversible
kinetic mechanism described by Scheme 3, while the MMX com-
pounds follow the same kinetics in caspase-3 inhibition, but fit a two-
step slow-binding reversible mechanism in caspase-8 inhibition
(Scheme 4)? The answer may be found, in part, by comparing the
kinetic schemes. Schemes 3 and 4 are actually subsets of the more
general mechanism in Scheme 5. For example, if the rate constants k2
and k−2 of Scheme 5 are large enough, as we believe is the case with
Z-VAD-fmk inhibition and inhibition of caspase-3 by the peptidomi-
metic compounds, then the formation of the thiohemiketal becomes
very fast and its formation and that of the Michaelis complex
approximate a single rapid equilibrium step. In this case, Scheme 5
would be indistinguishable from Scheme 3, no matter how the final
thioether is formed. On the other hand, if the rate constants k2 and
k−2 are small and kinact is extremely small in Scheme 5, as we believe
is the case with MMX compound inhibition of caspase-8, then
thiohemiketal formation could not be combined in one rapid
equilibrium step and the final thioether formation would not be
observable under our experimental conditions. Scheme 5 would
simplify to Scheme 4. The crystallographic results, which show the
absence of the leaving group, demonstrate that in this latter case, the
unobserved rate constant, kinact, must nonetheless be present. Thus,
the data are all consistent with Scheme 5, the apparent differences
being due solely to the magnitudes of the individual rate constants.
with the phenyl moiety of Compound-1 deeper in the pocket, and the
azaindole of Compound-9 extending further along the peptide-
binding groove. The positions of the oxygen atoms of the amides
linking the aromatic groups to the urazole rings in the two
compounds, however, differ by 3.8 Å. The orientation of the linker
in Compound-1, in particular, suggests the feasibility of appending a
group from the C4 position that might also access the S4 pocket.
The shallow depression that is the S2 subsite in both caspase-3 and
caspase-8 imparts selectivity for substrates with small hydrophobic P2
groups. This subsite is partially occupied by the C4, C5 and C6 atoms of
the urazolopyridazine core. The saturated carbocyclic fusion of R2 to
R3 utilized in some inhibitors, e.g., Compound-1 and Compound-6
(Fig. 6), presents a good opportunity to extend the inhibitors past this
pocket and provides opportunities for further elaboration. The
cyclohexyl moiety of Compound-9 extends even further from the
bicyclic core, and approaches the S1′ subsite. Trying to directly engage
the S1′ subsite using structure based drug design with the series of
inhibitors described here presents an interesting challenge because it
is reasonable to assume that this subsite overlaps the binding site of
the dichlorobenzoic acid leaving group that is displaced during the
course of the reaction.
4.4. Inhibitor binding at a secondary pocket in caspase-8
Hardy et al. [64] have proposed the existence of allosteric sites in
caspases. These sites were initially identified while screening caspase-
3 against a collection of thiol containing compounds. Subsequent
analysis, and the determination of cocrystal structures with caspase-7
revealed a binding site close to the heterodimer–heterodimer
interface, 14 Å from the active site. Mutagenesis experiments, and
the synthesis of thiol-free analogs, both demonstrated that the
formation of a cysteine:ligand covalent bond in the allosteric pocket
is required for inhibition by this class of compounds [64]. Further
mutational, kinetic, and structural studies in caspase-1 suggested that
it contains an allosteric circuit of residues connecting two active sites
[65]. Recent studies in caspase-7 have highlighted the role of this
allosteric switch in the transition from its inactive, zymogen state to
its proteolytically active form [66]. The shapes and precise locations of
concave pockets at heterodimer–heterodimer interfaces are not
conserved across different caspases [66]. Indeed, DTT, which is
bound in one such pocket in caspase-8 [38] (and described above)
does not appear to bind to caspase-3.
Our observation that Compound-4 has a secondary binding site on
caspase-8 demonstrates that it is possible to target a drug-like, non-
covalent small molecule to the proposed allosteric pockets of caspases
(Fig. 9). The volume of the concave pocket occupied by Compound-4
is small: modeling shows that it is of an appropriate size, shape, and
hydrophobicity to accommodate a single phenylalanyl or tyrosyl
residue. The affinity of Compound-4 for this site would therefore be
expected to be comparatively weak. Exploiting the proximal two-fold
symmetry axis (Fig. 9) may provide an approach to enhancing the
potency of compounds containing chlorophenyl acetic acid amide or
similar moieties. Joining two such moieties with an appropriate linker
could provide an advantage through avidity. Because the residues
which form the binding pocket are not conserved across caspases, it is
likely that such inhibitors would be caspase-specific.
4.3. Enzyme active site–inhibitor interactions
As detailed in the Results section, the urazolopyridazine inhibitors
described here mimic the peptide substrates of caspases. For some of
the moieties in the inhibitors, the nature of this mimicry is apparent
from inspection of their chemical structures. Thus the conserved
ketone warhead, P1 aspartic acid, and amide moiety emerging from
the C7 ring atom of the inhibitors has an identical counterpart in
peptidic caspase substrates. Fig. 8 shows that the corresponding
atoms are superimposable. Perhaps more subtly, the correspondence
between inhibitors and peptidic substrates also extends to the bicyclic
cores. For example, as designed, the interaction of C9 carbonyl oxygen
with a conserved arginine parallels a peptidic P3 backbone-carbonyl
interaction (Fig. 8). Substitutions off the N1 position of the urazole
ring allow the inhibitors to access the S3, S4, and S5 pockets, and afford
opportunities both to mimic peptide substrates and design novel
interactions. On the basis of positional scanning using a combinatorial
substrate library, Thornberry et al. [63] identified a glutamic acid as
the optimal group for the P3 position of caspase inhibitors. This
preference can be explained structurally by pointing to the hydrogen
bond that exists between the P3 carboxylate in the Ac–IETD peptide
complex with caspase-8 [38], and the exposed face of the conserved
guanidinium of Arg413 (Fig. 8). The butyrate side chains that emerge
from the N1 position of the urazole ring in Compounds-4 and -9 are
nearly ideal mimics of this glutamic acid. In contrast, although the
branched valerate of Compound-1 was designed to afford a similar
interaction, the carboxylic moiety instead binds to Ser209 of caspase-3.
Superposition of the two inhibitors suggests that replacing the
valerate with a butyrate moiety may restore S3 pocket binding.
However, there appears to be little correlation in potency accompa-
nying the presence or absence of this particular S3–arginine
interaction in the inhibitors described here.
In conclusion, this study examines a novel urazolopyridazine
template for caspase inhibition that mimics the β-strand structure of
substrates for these enzymes. These compounds are potent and fairly
selective inhibitors of caspase-3 and caspase-8, and we provide herein
a detailed kinetic analysis of enzyme inhibition and high-resolution X-
ray crystallographic structures of three enzyme/inhibitor complexes.
Interest in the caspases as therapeutic targets in a number of diseases
where apoptosis plays a defining role is fueled by the obvious link
between caspase action and cell death. Biologically, the picture is
A comparison of the Compound-1/caspase-3 and Compound-9/
caspase-8 structures demonstrates that substitutions at the N1
urazole are capable of following different trajectories to access the
S4 pocket. Both inhibitors place aromatic groups within this pocket,

1830
Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
[22] S.J. Riedl, Y. Shi, Molecular mechanism of cell regulation during apoptosis, Nat.
Rev. Mol. Cell Biol. 5 (2004) 897–907.
[23] S. Kumar, Caspase function in programmed cell death, Cell Death Differ. 14 (2007)
32–43.
[24] S.K. Rupinder, A.K. Gurpreet, S. Manjeet, Cell suicide and caspases, Vascul.
Pharmacol. 46 (2007) 383–393.
[25] D.W. Chang, D. Ditsworth, H. Liu, S.M. Srinivasula, E.S. Alnemri, X. Yang,
Oligomerization is a general mechanism for the activation of apoptosis initiator
and inflammatory procaspases, J. Biol. Chem. 278 (2003) 16466–16469.
[26] P. Fuentes-Prior, G.S. Salvesen, The protein structure that shapes caspases activity,
specificity, activation and inhibition, Biochem. J. 384 (2004) 201–232.
[27] S.J. Riedl, G.S. Salvesen, The apoptosome: signaling platform of cell death, Nat. Rev.
Mol. Cell Biol. 8 (2007) 405–413.
[28] N.P.C. Walker, R.V. Talanian, K.D. Brady, L.C. Dang, N.J. Bump, C.R. Ferenz, S.
Franklin, T. Ghayur, M.C. Hackett, L.D. Hammill, L. Herzog, M. Hugunin, W. Houy, J.A.
Mankovich, L. McGuiness, E. Orlewicz, M. Paskind, C.A. Pratt, P. Reis, A. Summani, M.
Terranova, J.P. Welch, L. Xiong, A. Möller, D.E. Tracey, R. Kamen, W.W. Wong,
Crystal structure of the cysteine protease interleukin-1β-converting enzyme:
a (p20/p10)2 homodimer, Cell 78 (1994) 343–352.
complicated by the existence of over a dozen potential caspase targets
that are similar in mechanism and substrate recognition.
There is still much to be learned about the precise roles and
interrelationships of individual caspases in cellular dysfunction, and
selective inhibition of any one caspase may prove to be a daunting
task. The more we can learn about the detailed interactions between
caspases and their substrates or inhibitors, the closer we come to
being able to design specific drugs for specific diseases. The class of
caspase inhibitors described in the present work expands our view of
the catalytic sites in both an initiator and an executioner caspase.
Hopefully, these findings will provide insights as to the design of
drugs for multiple therapeutic indications that intervene at specific
points in the apoptotic cascade.
[29] K.P. Wilson, J. Black, J.A. Thomson, E.E. Kim, J.P. Griffith, M.A. Navia, M.A. Murcko,
S.P. Chambers, R.A. Aldape, S.A. Raybuck, S.J. Livingston, Structure and mechanism
of interleukin-1β converting enzyme, Nature 370 (1994) 270–275.
[30] A. Schweizer, C. Briand, M.G. Gruetter, Crystal structure of caspase-2, apical
initiator of the intrinsic apoptotic pathway, J. Biol. Chem. 278 (2003)
42441–42447.
[31] J. Rotonda, D.W. Nicholson, K.M. Fazil, M. Gallant, Y. Gareau, M. Labelle, E.P.
Peterson, D.M. Rasper, R. Ruel, J.P. Vaillancourt, N.A. Thornberry, J.W. Becker, The
three dimensional structure of apopain/CPP32, a key mediator of apoptosis, Nat.
Struct. Biol. 3 (1996) 619–625.
[32] P.R. Mittl, S. Di Marco, J.F. Krebs, X. Bai, D.S. Karanewsky, J.P. Priestle, M.G. Grütter,
Structure of recombinant human CPP32 in complex with the tetrapeptide acetyl-
Asp-Val-Ala-Asp fluoromethyl ketone, J. Biol. Chem. 272 (1997) 6539–6547.
[33] D. Lee, S.A. Long, J.L. Adamsa, G. Chanc, K.S. Vaidyac, T.A. Francisc, K. Kiklyd, J.D.
Winklere, C.-M. Sunge, C. Debouckf, S. Richardsonf, M.A. Levyg, W.E. DeWolf Jr., P.
M. Kellerg, T. Tomaszekg, M.S. Headh, M.D. Ryanh, R.C. Haltiwangerh, P.-H. Liangi,
C.A. Jansoni, P.J. McDevitti, K. Johansoni, N.O. Conchaj, W. Chanj, S.S. Abdel-
Meguidj, A.M. Badgerk, M.W. Larkk, D.P. Nadeauk, L.J. Suvak, M. Gowenk, M.E.
Nuttallkl, Potent and selective nonpeptide inhibitors of caspases 3 and 7 inhibit
apoptosis and maintain cell functionality, J. Biol. Chem. 275 (2000) 16007–16014.
[34] R. Ganesan, P.R.E. Mittl, S. Jelakovic, M.G. Gruetter, Extended substrate recognition
in caspase-3 revealed by high resolution X-ray structure analysis, J. Mol. Biol. 359
(2006) 1378–1388.
Acknowledgments
We thank Thomas L. Emmons for helpful discussions, and also the
staff of the IMCA-CAT beamlines. Use of the IMCA-CAT beamline 17-ID
at the Advanced Photon Source was supported by the companies of
the Industrial Macromolecular Crystallography Association through a
contract with the Illinois Institute of Technology. Use of the Advanced
Photon Source was supported by the US Department of Energy, Office
of Science, Office of Basic Energy Sciences, under contract no. W-31-
109-Eng-38.
References
[1] S. Elmore, Apoptosis: a review of programmed cell death, Toxicol. Pathol. 35
(2007) 495–516.
[2] P. Meier, A. Finch, G. Evan, Apoptosis in development, Nature 407 (1997)
796–801.
[3] N.N. Danial, S.J. Korsmeyer, Cell death: critical control points, Cell 116 (2004)
205–219.
[4] R.M. Siegel, Caspases at the crossroads of immune-cell life and death, Nat. Rev.
Immunol. 6 (2006) 308–317.
[35] R. Baumgartner, G. Meder, C. Briand, A. Decock, A. D'Arcy, U. Hassiepen, R. Morse,
M. Renatus, The crystal structure of caspase-6, a selective effector of axonal
degeneration, Biochem. J. 423 (2009) 429–439.
[5] J.F. Kerr, C.M. Winterford, B.V. Harmon, Apoptosis: its significance in cancer and
cancer therapy, Cancer 73 (1994) 2013–2026.
[6] A. Worth, A.J. Thrasher, H.B. Gaspar, Autoimmune lymphoproliferative syndrome:
molecular basis of disease and clinical phenotype, Br. J. Hematol. 133 (2006)
124–140.
[7] A. Benchoua, C. Guégan, C. Couriaud, H. Hosseini, N. Sampaïo, D. Morin, B.
Onténiente, Specific caspase pathways are activated in the two stages of cerebral
infarction, J. Neurosci. 21 (2001) 7127–7134.
[8] J. Yuan, B.A. Yanker, Apoptosis in the nervous system, Nature 407 (2000)
802–809.
[36] J. Chai, Q. Wu, E. Shiozaki, S.M. Srinivasula, E.S. Alnemri, Y. Shi, Crystal structure of
procaspase-7 zymogen: mechanism of activation and substrate binding, Cell 107
(2001) 399–407.
[37] S.J. Riedl, P. Fuentes-Prior, M. Renatus, N. Kairies, S. Krapp, R. Huber, G.S. Salvesen,
W. Bode, Structural basis for the activation of human procaspase-7, Proc. Natl
Acad. Sci. USA 98 (2001) 14790–14795.
[38] W. Watt, K.A. Koeplinger, A.M. Mildner, R.L. Heinrikson, A.G. Tomasselli, K.D.
Watenpaugh, The atomic resolution structure of caspase-8, a key activator of
apoptosis, Structure 7 (1999) 1135–1143.
[39] H. Blanchard, L. Kodandapani, P.R.E. Mittl, S. Di Marco, J.K. Krebs, J.C. Wu, K.J.
Tomaselli, M.G. Grütter, The three dimensional structure of caspase-8: an initiator
enzyme in apoptosis, Structure 7 (1999) 1125–1133.
[40] H. Blanchard, M. Donepudi, M. Tschopp, L. Kodandapani, J.C. Wu, M.G. Grutter,
Caspase -8 specificity probed at subsite S4: crystal structure of the caspase-8–Z-
DEVD-cho complex, J. Mol. Biol. 302 (2000) 9–16.
[41] M. Renatus, H.R. Stennicke, F.L. Scott, R.C. Liddington, G.S. Salvesen, Dimer
formation drives the activation of the cell death protease caspase-9, Proc. Natl
Acad. Sci. USA 98 (2001) 14250–14255.
[42] A.S. Ripka, D.H. Rich, Peptidomimetic design, Curr. Opin. Chem. Biol. 2 (1998) 441–452.
[43] A. Albeck, S. Kliper, Mechanism of cysteine protease inactivation by peptidyl
epoxides, Biochem. J. 322 (1997) 879–884.
[44] D. Bromme, J.L. Klaus, K. Okamoto, D. Rasnick, J.T. Palmer, Peptidyl vinyl
sulphones: a new class of potent and selective cysteine protease inhibitors-
S2P2 specificity of human cathepsin O2 in comparison with cathepsins S and L,
Biochem. J. 315 (1996) 85–89.
[9] G. Tezel, M. Wax, Inhibition of caspase activity in retinal cell apoptosis induced by
various stimuli in vitro, Invest. Ophthalmol. Vis. Sci. 40 (1999) 2660–2667.
[10] J. Narula, P. Panday, E. Arbstini, N. Haider, N. Narula, F.D. Kolodgie, B. Dal Bello, M.J.
Semigran, A. Bielsa-Masdeu, G.W. Dec, S. Israels, M. Ballester, R. Virmani, S.
Saxena, S. Kharbanda, Apoptosis in heart failure: release of cytochrome c from
mitochondria and activation of caspase-3 in human cardiomyopathy, Proc. Natl
Acad. Sci. USA 96 (1999) 8144–8149.
[11] T.L. Vanden, Y. Qin, K. Wojcik, C.Q. Li, Z.H. Shao, T. Anderson, L.B. Becker, K.J.
Hamann, Reperfusion and not simulated ischemia initiates apoptosis injury in
chick cardiomyocytes, Am. J. Physiol. Heart Circ. Physiol. 284 (2003) H141–H150.
[12] R.S. Hotchkiss, D.W. Nicholson, Apoptosis and caspases regulate death and
inflammation in sepsis, Nat. Rev. Immunol. 6 (2006) 813–822.
[13] H.A. Kim, F.J. Blanco, Cell death and apoptosis in ostearthritic cartilage, Curr. Drug
Targets 8 (2007) 333–345.
[14] C.A. Dinarello, Therapeutic strategies to reduce IL-1 activity in treating local and
systemic inflammation, Curr. Opin. Pharmacol. 4 (2004) 378–385.
[15] S. Nakae, Y. Komiyama, H. Yokoyama, A. Nambu, M. Umeda, M. Iwase, I. Homma,
K. Sudo, R. Horai, M. Asano, Y. Iwakura, IL-1 is required for allergen-specific Th2
cell activation and the development of airway hypersensitivity response, Int.
Immunol. 15 (2003) 483–490.
[16] K. Takeda, J. Stagg, H. Yagita, K. Okumura, M.J. Smyth, Targeting death-inducing
receptors in cancer therapy, Oncogene 26 (2007) 3745–3757.
[17] S. Cornelius, K. Kersse, N. Festjens, M. Lamkanfi, P. Vandenabeele, Inflammatory
caspases: targets for novel therapies, Curr. Pharm. Des. 13 (2007) 367–385.
[18] S.D. Linton, Caspase inhibitors: a pharmaceutical industry perspective, Curr. Top.
Med. Chem. 5 (2005) 1697–1717.
[45] H.-O. Kim, M. Kahn,
A merger of rational drug design and combinatorial
chemistry: development and application of peptide secondary structure
mimetics, Comb. Chem. High Throughput Screen. 3 (2000) 167–183.
[46] P.D. Boatman, C.O. Ogbu, M. Eguchi, H.-O. Kim, H. Nakanishi, B. Cao, J.P. Shea,
M. Kahn, Secondary structure peptide mimetics: design, synthesis, and
evaluation of β-strand mimetic thrombin inhibitors, J. Med. Chem. 42 (1999)
1367–1375.
[47] R. St. Charles, J.H. Matthews, E. Zhang, A. Tulinsky, Bound structures of novel P3–P′
1
β-strand mimetic inhibitors of thrombin, J. Med. Chem. 42 (1999) 1376–1383.
[48] K.B. Rank, A.N. Mildner, W.J. Leone, K.A. Koeplinger, K.C. Chou, A.G. Tomasselli, R.L.
Heinrikson, S.K. Sharma, [W206R]-Procaspase-3; an inactivatable substrate for
caspase-8, Protein Expr. Purif. 22 (2001) 258–266.
[49] K.A. Koeplinger, A.M. Mildner, J.W. Leone, J.S. Wheeler, R.L. Heinrikson, A.G.
Tomasselli, Caspase-8: an efficient method for autoactivation of recombinant
procaspase-8 by matrix absorption and characterization of the active enzyme,
Protein Expr. Purif. 18 (2000) 378–387.
[19] T. O'Brien, D. Lee, Prospects for caspase inhibitors, Mini-Rev. Med. Chem. 4 (2004)
153–165.
[20] N.A. Thornberry, Y. Lazebnik, Caspases: enemies within, Science 281 (1998)
1312–1316.
[21] M.O. Hengartner, The biochemistry of apoptosis, Nature 407 (2000) 770–776.

Z. Wang et al. / Biochimica et Biophysica Acta 1804 (2010) 1817–1831
1831
[50] T. Little, J. Meara, F. Ruan, M. Nguyen, M. Qabar, Efficient synthesis of novel 4-
substituted urazoles, Synthetic Communications 32 (2002) 1741–1749.
[51] P.D. Boatman, J. Urban, M. Nguyen, M. Qabar, M. Kahn, High-throughput synthesis
and optimization of thrombin inhibitors via urazole α-addition and Michael
addition, Bioorg. & Med. Chem. Lett. 13 (2003) 1445–1449.
[59] P. Emsley, K. Cowtan, Coot: model-building tools for molecular graphics, Acta
Cryst. D60 (2004) 2126–2132.
[60] R.A. Laskowski, M.W. MacArthur, D.S. Moss, J.M. Thornton, PROCHECK: a program
to check the stereochemical quality of protein structures, J. Appl. Crystallogr. 26
(1993) 283–291.
[52] Z. Otwinowski, W. Minor, Processing of X-ray diffraction data collected in
oscillation mode, Meth. Enzymol. 276 (1997) 307–326.
[61] P. Kuzmic, Program DYNAFIT for the analysis of enzyme kinetic data: application
to HIV proteinase, Anal. Biochem. 237 (1996) 260–273.
[53] Collaborative Computational Project No. 4, The CCP4 suite: programs for protein
crystallography, Acta Crystallogr., Sect D: Biol. Crystallogr 50 (1994) 760–763.
[54] J. Nazava, AMoRe: An automated package for molecular replacement, Acta
Crystallogr. A50 (1994) 157–163.
[55] Bruker Analytical X-ray Systems, SHELXTL Software Reference Manual, Bruker
Analytical X-ray Systems, Madison, WI, USA, 1997.
[56] G.N. Murshudov, A.A. Vagin, E.J. Dodson, Refinement of macromolecular
structures by maximum-likelihood method, Acta Crystallogr. D Biol. Crystallogr.
53 (1997) 240–255.
[62] R. Ganesan, P.R.E. Mittl, S. Jelkovic, M.G. Grutter, Extended substrate recognition in
caspase-3 revealed by high resolution X-ray analysis, J. Mol. Biol. 359 (2006)
1378–1388.
[63] N.A. Thornberry, T.A. Rano, E.P. Peterson, D.M. Rasper, T. Timkey, M. Garcia-Calvo,
V.M. Houtzager, P.A. Nordstrom, S. Roy, J.P. Vaillancourt, K.T. Chapman, D.W.
Nicholson, A combinatorial approach defines specificities of members of the
caspase family and granzyme B, J. Biol. Chem 272 (1997) 17907–17911.
[64] J.A. Hardy, J. Lam, T. O'Brien, J.A. Wells, Discovery of an allosteric site in the
caspases, Proc. Natl Acad. Sci. USA 101 (2004) 12461–12466.
[57] J.S. Sack, F.A. Quiocho, CHAIN Crystallographic Modeling Program, Baylor College
of Medicine, Houston, TX, USA, 1992.
[65] D. Datta, J.M. Scheer, M.J. Romanowski, J.A. Wells, An allosteric circuit in caspase-
1, J. Mol. Biol. 381 (2008) 1157–1167.
[58] B.C. Finzel, Mastering the LORE of protein structure, Acta Crystallogr. D Biol.
Crystallogr. 51 (1995) 450–457.
[66] J.A. Hardy, J.A. Wells, Dissecting an allosteric switch in caspase-7 using chemical
and mutational probes, J. Biol. Chem. 284 (2009) 26063–26069.