Effect of a procaspase-activating compound on the catalytic activity of mature caspase-3

Article abstract of DOI:10.1246/bcsj.20150139

Caspase-3, an apoptosis-executing enzyme with a dimeric structure, and its zymogen (procaspase-3) are target proteins in cancer chemotherapy using synthetic compounds. Previous work has shown that PAC-1 (1st procaspaseactivating compound) affects the activation of procaspase-3 in cancerous cells through the Zn²⁺-chelation mechanism. However, we found that the compound also enhances the mature caspase-3 activity in the absence of Zn²⁺. The enhancement of the mature caspase-3 activity by PAC-1 followed a Michaelis-Menten-type saturation dependency with respect to the concentration of the compound. PAC-1 induced the change in the UV-vis spectrum of the mature caspase-3. From these findings, we propose a direct structural interaction of PAC-1 with the protein core. According to a molecular dynamics calculation for the complex of PAC-1 and mature caspase-3, PAC-1 is likely to bind to the interface between the dimeric structure in mature caspase-3. The interactions between PAC-1 and amino acid residues that affect the catalytic activity of mature caspase-3 are suggested. Throughout this work, diverse functions of the apoptosis-inducing compound are disclosed.

Full text of DOI:10.1246/bcsj.20150139

BCSJ Award Article
Effect of a Procaspase-Activating Compound on
the Catalytic Activity of Mature Caspase-3
Takashi Matsuo,* Keita Yamada, Masaya Ishida, Yoshiyuki Miura,
Masaru Yamanaka, and Shun Hirota
Graduate School of Materials Science, Nara Institute of Science and Technology (NAIST), Ikoma, Nara 630-0192
E-mail: tmatsuo@ms.naist.jp
Received: April 18, 2015; Accepted: May 31, 2015; Web Released: June 8, 2015
Caspase-3, an apoptosis-executing enzyme with a dimeric structure, and its zymogen (procaspase-3) are target
proteins in cancer chemotherapy using synthetic compounds. Previous work has shown that PAC-1 (1st procaspase-
activating compound) affects the activation of procaspase-3 in cancerous cells through the Zn2 -chelation mechanism.
However, we found that the compound also enhances the mature caspase-3 activity in the absence of Zn2 . The
enhancement of the mature caspase-3 activity by PAC-1 followed a Michaelis­Menten-type saturation dependency with
respect to the concentration of the compound. PAC-1 induced the change in the UV­vis spectrum of the mature caspase-3.
From these findings, we propose a direct structural interaction of PAC-1 with the protein core. According to a molecular
dynamics calculation for the complex of PAC-1 and mature caspase-3, PAC-1 is likely to bind to the interface between the
dimeric structure in mature caspase-3. The interactions between PAC-1 and amino acid residues that affect the catalytic
activity of mature caspase-3 are suggested. Throughout this work, diverse functions of the apoptosis-inducing compound
are disclosed.
+
+
to have a similar overall structure to that of mature caspase-3.
The idea is based on the fact that procaspase-3 has very weak
Introduction
Caspase-3 is a cysteine protease that cleaves scores of
cellular substrates with the recognition of “DEVD” amino acid
catalytic activity with the lower kcat but similar K values,
m
compared to that of the mature enzyme.
7
,8
1
,2
sequence at the downstream on an apoptosis cascade. This
enzyme is produced as a mature form through the activation of
the corresponding zymogen (procaspase-3) with the help of
Spontaneous maturation (i.e. auto-activation) of procaspase-
3 without the help of effector enzymes is negligible under
9
physiological conditions because of extremely slow velocity.
3
,4
other effector enzymes (caspase-9, granzyme B and so on).
In contrast, the activation triggered by the effector enzymes is
significantly inhibited in several cancerous cells in spite of
higher elevated expression level of procaspase-3 than that in
normal cells, which is a reason for unregulated proliferation
The X-ray crystallographic structure of the mature caspase-3
displays a “homo-heterodimeric dimer”:5 A small domain
(ca. 12 kDa) and large domain (ca. 17 kDa) construct one
9
protomer unit by noncovalent interactions, and the two
identical protomer units interact at the β-sheet-rich interface
of cancer cells. Accordingly, the recovery of the caspase-3
maturation by synthetic compounds has been enlightened as a
new trend of cancer chemotherapy.
(Figures 1a and 1b). The protomer interface has a cavity-
shaped structure, where several amino acid residues contrib-
ute to the hydrophobic interface interactions. In contrast, the
zymogen procaspase-3 additionally has a prodomain at the N-
terminal of a large domain and an inter-domain linker (IL) that
covalently connects the large and small domains. The activa-
tion of procaspase-3 occurs on the cleavage at D9/D28/D175
by the action of effector enzymes (Figure 1c). The X-ray
crystallographic structure of the C163A procaspase-3 mutant
For synthetic compounds affecting the (pro)caspase-3 activa-
tion process, two types of action mechanism are known: (i)
Binding to the core of procaspase-3 to induce structural pertur-
bation in the protein; (ii) Chelation removal of a metal ion that
binds to the procaspase-3 cleavage sites under the situation
of unregulated up-take of metal ions in cancerous cells. As
an example of the former case, a series of synthetic mole-
cules (denoted as ID 1541s) reported by Wells and co-workers
bind to the wild-type procaspase-3 to shift the conformational
equilibrium from the resting state to a mature enzyme-like
(mutated at Cys163 at the active site) recently reported indi-
cates the similarity in the whole structure between procaspase-3
6
10
and mature caspase-3. Although the crystal structure of non-
conformation. Subsequent studies of the compounds have
11­13
mutated procaspase-3 is not available, procaspase-3 is believed
shown the formation of nanofibrils in solutions.
Namely,
Bull. Chem. Soc. Jpn. 2015, 88, 1221–1229 | doi:10.1246/bcsj.20150139
© 2015 The Chemical Society of Japan | 1221

a)
mutations at the cleavage sites and protomer interface: The
1
7,18
change in the polarity of the protomer interface
or the
1
9
insertion of several amino acid residues into IL induce the
mature enzyme-like whole structure in spite of inertness for
cleavages at D9/D28/D175.
As another strategy to artificially regulate the procaspase-3
activation process, Hergenrother and co-workers proposed that
PAC-1 (1st procaspase-activating compound (1); Figure 1d)
recovers the procaspase-3 activation in cancerous cells, where
PAC-1 removes zinc ions coordinated by aspartates at the
cleavage place in procaspase-3 to produce a metal ion-free pro-
b)
2
0­24
tein.
This mechanism is based on the fact that the cleavage
9
site in IL of procaspase-3 is an Asp-rich place. Experimen-
tally, the idea was verified by the occurrence of PAC-1-induced
apoptosis of cancer cells and by the high Zn -chelation
2+
7
¹1 22
activity of 1 (K µ 50 nM, i.e. K µ 2.0 © 10 M ). The clear
d
relationship between IC50 values and expression amounts of
procaspase-3 in various colon cancer tissues supported the
2
0
action of 1 on procaspase-3. However, 1 also induces apopto-
sis by endoplasmic reticulum stress in vivo without the metal-
c)
2
4
2+
chelating mechanism. The previous investigation of the Zn
-
chelation activity and in vitro procaspase-3 activation by 1 was
conducted at neutral conditions (pH 7.2­7.4) although it is
2
5,26
known that lysosome in cancerous cells is weakly acidic.
Consequently, other mechanisms of 1 (e.g. binding mechanism
as seen in ID 1541s) can occur in parallel with metal ion-
chelation mechanism. Furthermore, one noticeable issue in
in vitro studies of zymogen activation is that zymogens pre-
pared by genetic expression are often obtained concomitantly
with a small amount of the corresponding mature enzyme
produced by self-activation during purification: The high cata-
lytic activity of the mature enzyme is unexpectedly observed
even in a small amount of the mature enzyme. In addition,
synthetic molecules that are believed to influence the zymogen
by a structural perturbation mechanism may also affect the
activity of the corresponding mature enzyme if both a zymogen
protein and its mature form have similar whole structure as
seen in (pro)caspase-3. In fact, a recent study demonstrated that
the catalytic activity of mature caspase-3 is also enhanced
by the structural perturbation as well as that of uncleavable
d)
2
7
procaspase-3s. Namely, we should differentiate effects of a
bioactive synthetic molecule on zymogens, their mature
enzymes or both of them by individually investigating effects
of the synthetic molecule on these proteins. Accordingly, this
paper focuses on the possibility of effects of 1 on mature
caspase-3. As a result, we found that 1 is also able to increase
the catalytic activity of mature caspase-3 under weakly acidic
conditions, although previous studies of 1 have focused on the
affection to zymogen activation. From the kinetic data and
molecular dynamics calculations, we propose that the inter-
action of a synthetic molecule to the interface of protomer units
of mature caspase-3 is a possible action mechanism of the small
organic molecule to affect the function of mature caspase-3.
Figure 1. Structures of mature human caspase-3 and PAC-1
1) and activation mechanism for production of the mature
(
caspase-3. a) Structure of human caspase-3 (as a hemi-
thioacetal complex with AcLDESD-CHO at Cys163;
PDB code: 3EDQ). b) Magnified figure of the protomer
interface. c) Production of the mature caspase-3 from
procaspase-3. d) Structure of PAC-1. In fig. b), symbols
“
(A)” and “(B)” stand for each protomer unit.
the increase in the local concentration of procaspase-3 bound
onto the nanofibrils leads to the acceleration of auto-activation.
Several papers have reported the allosteric regulation of various
caspases using synthetic compounds.1
4,15
For proteins other
Experimental
than caspases, the utility of designed synthetic molecules that
interact with the subunit interface of a target protein has been
reported as a strategy for modulating the protein function.
Furthermore, the utility of structural affections has also been
demonstrated using uncleavable procaspase-3 variants with
Materials and Instruments. The plasmid pET-23b coding
full-length human procaspase-3 with His-Tag at the C-terminal
was extracted from E. coli HB101 purchased from ATCC.
Mature human caspase-3 was obtained by the expression in
1
6
1222 | Bull. Chem. Soc. Jpn. 2015, 88, 1221–1229 | doi:10.1246/bcsj.20150139
© 2015 The Chemical Society of Japan

E. coli BL21(DE3) transformed by the above-mentioned plas-
mid and the subsequent auto-activation of procaspase-3 zymo-
gen during the purification.28 The cultivation and the purifica-
tion procedures were carried out using a modified method of
a previously reported method (described below).29 PAC-1 (1),
its building block compounds (compounds 2 and 3) and an
analog of 1 without a phenolic compound (compound 4) were
synthesized by the reported method.20 An analog that lacks a
benzyl group (compound 5) was synthesized by the method
described in Supporting Information. Protein standard sam-
ples for the calibration in size exclusion chromatographic
analysis were purchased from Sigma (for bovine serum
albumin (BSA) and subtilisin Carlsberg (STL)) or obtained
by the method descried in a previous paper (for Hydro-
The release of p-nitroaniline (405 nm) was monitored. For
anaerobic measurements, sample solutions were prepared in a
glove-box filled with nitrogen. For preparation of enzyme solu-
tions with defined concentrations of 2-ME, a frozen enzyme
solution described above was appropriately diluted with buffer
to adjust concentrations of 2-ME of resultant solutions. Other-
wise, enzyme solution was dialyzed against buffer without
2-ME under a N atmosphere and the dilution factor was used
2
for determination of the concentration of 2-ME.
Global Docking and Molecular Dynamics Calculation
for Complex of PAC-1 and Mature Caspase-3. The condi-
tions for docking and MD calculations were set up using the
YASARA structure molecular modeling software package
3
2
(Ver.14.4.15). Global docking calculation (without restriction
3
0
33
genobacter thermophilus cytochrome c552 (HT cyt c) ). STL
was treated with phenylsulfonyl fluoride (PMSF) before use to
suppress autolysis. Other chemicals and columns were obtained
from conventional vendors and used as received unless noted.
UV­vis spectral and kinetic measurements were carried out
using a Shimadzu UV-2550 double beam spectrophotometer
with a thermostated cell holder. Protein purification was con-
ducted using a BioRad Biologic Duoflow in a chromatocham-
ber (4 °C). CD (circular dichroism) spectra were collected using
a JASCO J-725 circular dichroism spectrophotopolarimeter.
Protein Expression and Purification. E. coli BL21(DE3)
harboring the plasmid pET-23b for full-length human
procaspase-3 was cultivated in an LB medium (5 mL) contain-
of binding sites) was run on VINA algorithm with point
3
4
charges assigned by the YASARA2 force field at pH 6.0. As
the initial structure of a receptor protein, the X-ray crystal
structure of the mature caspase-3 with AcLDESD-CHO, an
irreversible inhibitor of caspase-3, (PDB code: 3EDQ) was
employed. To obtain the initial structure of ligand PAC-1 for
docking to the receptor protein, the optimized structure was
calculated by MM2 on Spartan 08 program (Wavefunction Inc.)
and saved as a file with mol2 format. Global docking of the
ligand to the receptor protein was performed with a cubic cell
boundary defined at 5 ¡ from the protein surface. In the dock-
ing simulation, the protein receptor was set as a rigid receptor,
whereas structural flexibility was incorporated into the ligand.
After 100-run repeating calculations, the binding structures
likely to occur were selected based on the calculated binding
energies.
¹
1
ing ampicillin (100 ¯g mL ) overnight at 37 °C. The culture
solution (3 mL) was put into another LB medium (500 mL),
and the solution was aerobically shaken at 37 °C until the
absorbance at 600 nm reached 0.6­0.8 (ca. 3 h). After cooling
to 25 °C, IPTG (isopropyl thiogalactopyranoside) was added
MD simulations were conducted using the YASARA2 force
field, the aqueous solution model with 0.9% NaClaq ion
concentration, point charges assigned at pH 6.0 and additional
(0.3 mM in final) for the induction of protein expression at
+
¹
2
5 °C for 10 h. The culture solution obtained was centrifuged
Na or Cl for charge neutralization in a cubic cell boundary
defined at 5 ¡ from the protein surface. The simulations were
started from the global docking structures obtained above and
conducted for the solution model at 310.15 K. The calculations
were continued over 10­14 ns until the change in Cα-RMSDs
(root-mean-square deviations) reached an equilibrium. The
snapshot figures every 25 ps were stored.
at 4 °C (8000g, 10 min), and the resulting pellet was suspended
in 20 mL of 50 mM Tris buffer (pH 8.0) with 100 mM NaCl
and frozen at ¹80 °C. The suspension was subjected to freeze-
and-thaw cycles (©3). After 2-mercaptoethanol (2-ME) was
added (0.1% (v/v)), the cells were lysed by sonication and
centrifuged at 4 °C (17000g, 45 min). The supernatant was
absorbed onto a cobalt-chelating affinity column (1 mL) with
the elution of 50 mM Tris buffer (pH 8.0) that contains 100 mM
NaCl and 2-ME (0.1% (v/v)) (denoted as “Buffer A”). After
washing with 50 mM Tris buffer (pH 8.0) that contains 500
mM NaCl, 20 mM imidazole and 2-ME (0.1% (v/v)) (denoted
as “Buffer B”), the protein was eluted with Buffer A that con-
tains 200 mM imidazole. Fractions containing pro- and mature
caspase-3 were further purified by gel filtration using a HiLoad
Superdex 75 (GE healthcare) with the Buffer A elution. After
the protein-containing fractions were concentrated and stood
at 4 °C overnight in the presence of 2-ME (0.1% (v/v)) to
promote auto-activation, the solution was frozen by liquid
nitrogen and kept at ¹80 °C. The completion of auto-activation
was checked by the disappearance of a band around 33 kDa on
SDS-PAGE. The enzyme concentration was determined using
Results and Discussion
Mature Caspase-3 Maintains a Homo-Heterodimeric
Dimer in Solutions. The SDS-PAGE profile after the protein
expression with truncated protein induction time (45 min)
showed that the zymogen procaspase-3 was obtained concom-
itant with the mature form (Figure S1 in Supporting Informa-
tion). After the concentrated enzyme solution was stood at 4 °C
overnight, the band derived from the zymogen disappeared
because of slow self-digestion to form the mature form. In
order to avoid complications from the presence of the zymo-
gen, the induction period was prolonged (ca. 3 h) to promote
the self-digestion. All kinetic measurements presented below
were conducted after the full-conversion into the mature form
was confirmed by SDS-PAGE (Figure S2).
¹
1
¹1 31
the extinction coefficient at 280 nm of 26900 M cm .
Kinetic Measurements. The caspase-3 activity was deter-
mined by monitoring the hydrolysis of AcDEVD p-nitroanilide
in MES buffer solutions (pH 6.0) containing NaCl (100 mM).
To confirm whether the obtained mature caspase-3 really
adopts a homo-heterodimeric dimer structure in solution, we
evaluated the apparent molecular size of the isolated enzyme
by gel filtration chromatography (Figure S3) and blue-native
Bull. Chem. Soc. Jpn. 2015, 88, 1221–1229 | doi:10.1246/bcsj.20150139
© 2015 The Chemical Society of Japan | 1223

PAGE analysis (Figure S4). Figure S3a shows the elution of
the mature caspase-3 at the position between bovine serum
albumin (BSA, 66 kDa) and subtilisin Carlesberg (STL, 28
kDa) in gel chromatography. According to the calibration curve
shown in Figure S3b, the molecular weight of the protein in
solution was evaluated to be ca. 54 kDa. The Blue-native PAGE
analysis (Figure S4) also supported the molecular weight deter-
mined by the gel filtration chromatography. The molecular
weight elucidated by the chromatographic analyses is larger
than that expected for a heterodimeric monomer composed of
one set of small and large domains (12 and 17 kDa, respec-
tively). Considering that molecular weights of proteins deter-
mined by gel filtration chromatography contain a 10% error in
a)
Time/s
3
5
b)
maximum, we conclude that the isolated mature caspase-3
surely takes a homo-heterodimeric dimer form in solution state,
not dissociated to two heterodimeric monomers. This finding
highlights the utility of strategies for regulating the caspase-3
activity based on modulations of the dimeric structures.
PAC-1 Affects the Activity of Mature Caspase-3 in the
2
+
Presence of Zn as well as that of Zymogen Procaspase-3.
At first, we investigated the effect of PAC-1 (1) on the catalytic
2+
activity of mature caspase-3 in the presence of Zn . The
caspase-3 assays under weakly acidic conditions will give us
an insight on the action mechanism of 1 in cancerous cells
because it is known that lysosome in cancerous cells is weakly
2
5,26
acidic.
Apoptosis-inducing compounds are expected to
/µM
activate apoptosis-related enzymes under these or similar
conditions. Accordingly, we studied the caspase-3 activity at
pH 6.0.
Figure 2. Dependency of mature caspase-3 activity on
concentrations of Zn2 in the absence and presence of
PAC-1 (1). a) Time-course of product release in the mature
caspase-3-mediated AcDEVD p-nitroanilide at various
+
Figure 2 demonstrates that Zn2+ significantly decreases the
2+
caspase-3 activity at high concentrations of Zn as reported
2+
concentrations of Zn . Dotted and solid lines the data
2
,22,36
in previous work.
The decrease in the catalytic activity
taken in the absence and presence of 1, respectively. b)
Relative mature caspase-3 activity (normalized by the
indicates that a zinc ion blocks the active site. However, the
extent of the decrease in the caspase-3 activity is lower in the
presence of 1 than that without 1 (Figure 2b). A possible reason
of the effect is that 1 still has the ability of Zn2 -binding at
pH 6.0 to contribute to the escape of caspase-3 from the inhi-
2+
activity without the addition of Zn ) in the absence and
presence of 1. The activities were evaluated based on the
increases in absorbance (405 nm) at 60 s in each reaction
solution. Data at each concentration of Zn2 are based on
averages of the values obtained by four-run experiments
and error bars indicate standard deviations. Measurement
conditions: 50 mM MES (pH 6.0), 100 mM NaCl, DMSO
+
+
2
+
2+
bition by Zn . A PAC-1-Zn complex involves the coordi-
nation of the deprotonated phenol in 1 to Zn2
+ 37
. The pKa value
of the phenol is ca. 9.5 (see the pH titration in Figure S5): 1
predominantly adopts the protonated form at pH 6.0. Conse-
2% (v/v), [caspase-3] = 20 nM, [substrate] = 90 ¯M,
2
+
[1] = 0 or 28 ¯M, [2-ME] = 0.037 mM, 37 °C.
quently, the Zn -binding ability is expected to be extremely
low under these conditions. Another noticeable fact is that
2+
the caspase-3 activity is enhanced in the absence of Zn
see curves (A) and (D) in Figure 2a). These facts suggest that
glove-box to avoid the chemical deactivation of Cys163 at
the active site. The reaction condition of [2-ME] µ 15 mM is
(
2
9
other mechanisms in parallel with the metal-ion chelation are
possible in the enhancement of the mature caspase-3 activ-
ity by 1.
similar to that popularly employed in caspase assays. Table 1
summarizes the kinetic parameters for the capsase-3-mediated
hydrolysis of AcDEVD p-nitroanilide without the addition of
2+
PAC-1 Enhances Both Substrate Binding and Catalysis
in Mature Caspase-3-Catalyzed Hydrolysis in the Absence
Zn to reaction media. Figure S6 shows Michaelis­Menten
2+
plots for the kinetic analysis in Zn -free reaction media.
PAC-1 (1) enhances the catalytic activity of mature caspase-
3 at low concentrations of 2-ME (see Entries 1/2 and 3/4). The
overall catalytic activity (k /K ) increases by 5.8-fold and
2+
of Zn
.
To investigate the enhancement of caspase-3 activ-
ity by PAC-1 (1) in detail, we conducted Michaelis­Menten
analyses in the absence and presence of 1. In order to evaluate
effects of the additive used in standard caspase-3 assays on the
PAC-1 ability, we also conducted kinetic measurements in
infinitely diluted 2-ME solutions (i.e. ca. 0 mM after dialysis)
and in solutions that contains a huge excess of 2-ME (15 mM,
cat
m
8.6-fold at 0.037 and ca. 0 mM of 2-ME, respectively. Under
these conditions, PAC-1 brings about positive effects on both
K and k values with respect to the enhancement of catalytic
m
cat
activity. The finding of the activity enhancement even without
2+
0
.12% (v/v)). Sample preparation and kinetic measurements at
Zn in reaction media surely indicates that mechanisms of 1
other than metal chelation should be considered.
[2-ME] µ 0 and 0.037 mM were anaerobically conducted in a
1224 | Bull. Chem. Soc. Jpn. 2015, 88, 1221–1229 | doi:10.1246/bcsj.20150139
© 2015 The Chemical Society of Japan

Table 1. Kinetic parameters of caspase-3-mediated hydrolysis of AcDEVD p-nitroanilide in the absence of Zn2+ a)
2
/
-ME
mM
PAC-1
(1)
Km
/μM
kcat
/s
(kcat/Km)
¹1 ¹1 b)
Entry
b)
¹1 b)
/μM
s
c)
c)
f)
f)
d)
1
2
3
4
5
6
0.037
0.037
ca. 0
ca. 0
(¹)
9.3 « 1.0
5.5 « 0.5
10 « 1
3.2 « 0.2
5.0 « 0.3
4.8 « 0.3
0.70 « 0.03
2.5 « 0.1
0.014 « 0.1
0.039 « 0.01
5.8 « 0.1
0.076 « 0.011
0.44 « 0.05
0.0014 « 0.0003
0.012 « 0.001
1.2 « 0.1
e)
(+)
d)
(¹)
g)
(+)
h)
d)
15
15
(¹)
h)
e)
(+)
5.8 « 0.2
1.2 « 0.1
a) 50 mM MES (pH 6.0); 100 mM NaCl; DMSO 2% (v/v); [substrate] = 1.6­50 ¯M; [caspase-3] = 20 nM (for
measurements in the presence of 2-ME) or 200 nM (for measurements in the absence of 2-ME); 37 °C under a N2
atmosphere ([2-ME] = 0 and 0.037 mM) or under air ([2-ME] = 15 mM). b) Standard errors were determined
from the curve fitting in Michaelis­Menten analyses. c) Corresponding to 0.0003% (v/v). d) [1] = 0 ¯M (not
added). e) [1] = 5 ¯M (i.e. [1]/[caspase-3] = 2500). f) Infinitely diluted 2-ME by dialysis: Calculated
concentration of 2-ME based on dilution factor is <140 pM. g) [1] = 50 ¯M (i.e. [1]/[caspase-3] = 2500).
h) Corresponding to 0.12% (v/v).
The catalytic activity of caspase-3 intrinsically decreases
with lowering the concentration of 2-ME regardless of the
absence and presence of 1 (compare Entries 1, 3, and 5). Since
the sample preparations and kinetic measurements were per-
formed anaerobically, the reduction in the caspase-3 activity is
not associated with chemical deactivation of Cys163 at the
active site. Another noticeable fact is that the presence of 2-ME
with high concentrations abolishes the ability of 1 in the
enhancement of the catalytic activity of caspase-3 (Entries 5
and 6). The effects of 2-ME on the caspase-3 activity in the
absence and presence of PAC-1 will be discussed below.
PAC-1 May Structurally Affect Mature Caspase-3. We
next investigated the concentration effect of PAC-1 (1) on the
catalytic activity at the defined concentration of 2-ME. Figure 3
shows a saturated kinetic behavior of the caspase-3 activity
with respect to concentrations of 1. The curve-fitting using a
/
µM
Michaelis­Menten-type equation shown in Figure 3 yielded the
PAC-1 concentration that shows 50% of the maximum activity
Figure 3. Dependency of initial rate of caspase-3-catalyzed
AcDEVD p-nitroanilide on concentration of PAC-1 (1).
Initial rate values at each concentration of 1 are shown
as averages of the data obtained by three-run experiments
and error bars indicate standard deviations. Measurement
conditions: 50 mM MES (pH 6.0), 100 mM NaCl, DMSO
(
[
KPAC-1) of KPAC-1 = 51 « 15 nM (at [caspase-3] = 20 nM;
2-ME] = 0.037 mM). The saturation dependency of caspase-
3
activity shown in Figure 3 suggests that 1 directly interacts
with mature caspase-3.
Furthermore, we also studied the UV­vis spectrum of a
mixed solution of the protein and 1 to compare with the
simulated spectrum in which no specific interaction between
the two components was assumed (Figure 4). Compound 1
has three absorption bands in the range from 250 to 400
nm (282, 291, and 321 nm; spectrum (a)). The absorbance of
the experimentally obtained spectrum (spectrum (d)) is larger
than the simulated spectrum (a summed line of the spectra of
each component; spectrum (c)) in this region. The difference
between the experimental and simulated spectra (spectrum (e))
stem from some interactions between 1 and the protein.
Furthermore, we observed the appearance of induced CD
signals in the wavelength range of absorption bands derived
from 1 (Figure S7) for the 1/caspase-3 mixture, indicating that
2
%
(v/v), [caspase-3] = 20 nM, [substrate] = 40 ¯M,
[2-ME] = 0.037 mM, 37 °C.
Benzyl and Phenolic Groups in PAC-1 are Important for
the Ability of PAC-1 in the Enhancement of Mature
Caspase-3 Activity. To investigate the structural significance
in PAC-1 (1) for the enhancement of mature caspase-3 activity,
we studied the effects of several analogs of 1 (compounds 2
through 5; Figure 5a) on the mature caspase-3 activity. These
PAC-1 analogs are known to neither activate procaspase-3
2
0,40
nor induce cancerous cell apoptosis.
The time-courses of
caspase-3-mediated hydrolyses in the presence of these com-
pounds are shown in Figure 5b. Compounds that lack a pheno-
lic moiety (2, 3, and 4) have no ability in affecting the mature
caspase-3 activity (curves (B), (C), and (D) in Figure 5b).
Although compound 5 slightly increased the catalytic activity
of caspase-3 (curve (E)), the effect of 5 on the caspase-3
activity is much lower than that of 1 (curve (F)) at the common
1
is located in different environments from the solution bulk as
38,39
the result of interactions with caspase-3.
In other words, 1
structurally affects mature caspase-3. The interaction of 1 with
caspase-3 perturbs the protein structure, leading to the enhance-
ment of the caspase-3 activity.
Bull. Chem. Soc. Jpn. 2015, 88, 1221–1229 | doi:10.1246/bcsj.20150139
© 2015 The Chemical Society of Japan | 1225

a)
/nm
b)
Figure 4. UV­vis spectra of caspase-3, PAC-1 (1) and a
mixed solution of 1/caspase-3 in 50 mM phosphate buffer
(
0
(
pH 6.0) that contains 100 mM NaCl, DMSO 2% (v/v) and
.037 mM 2-ME at 37 °C. a) Spectrum of mature caspase-3
10 ¯M). b) Spectrum of 1 (40 ¯M). c) Summed spectrum
of spectra (a) and (b). d) Observed spectrum for the
mixture of 1 (40 ¯M) and mature caspase-3 (10 ¯M).
e) Difference spectrum of the calculated and observed
spectra (spectrum (d)­spectrum (c)).
measurement conditions. A series of PAC-1 analogs investi-
gated, which are inactive for procaspase-3 maturation, have
negligible effects on the enhancement of mature caspase-3
activity as well. Furthermore, the finding indicates that both the
benzyl and phenolic groups in 1 are essential for the enhance-
ment of mature caspase-3 activity.
Time/s
Additives Used in Caspase-3 Assays Affect the Ability of
PAC-1 in the Enhancement of Mature Caspase-3 Activity.
As described above, the caspase-3 activity decreased at low
concentrations of 2-ME. Furthermore, a huge excess of 2-ME
abolishes the ability of PAC-1 (1) in the enhancement of the
mature caspase-3 activity. A possible reason of these facts is
that 2-ME not only works as a reducing stabilizer for sup-
pressing the oxidation of cysteine thiols but also a structural
effector toward mature caspase-3. A saturation dependency
of the caspase-3 activity on the concentration of 2-ME
Figure 5. Structures of PAC-1 analogs and time-courses
of product release in the mature caspase-3-mediated
AcDEVD p-nitroanilide in the absence and presence of
analogs. a) Structures of building blocks and analogs of
PAC-1 (1) investigated in this work. b) Time-courses of
product release. Measurement conditions: 50 mM MES
(pH 6.0), 100 mM NaCl, DMSO 2% (v/v), [caspase-3] =
2
0 nM, [substrate] = 90 ¯M, [analog] = 0 or 28 ¯M,
[2-ME] = 0.037 mM, 37 °C.
(
1
Figure S8) supports this idea. A huge excess of 2-ME (ca.
000-fold higher concentrations) may interfere the interaction
propyl)dimethylammonio] propanesulfonate), a typical protein
solubilization reagent.
of 1 with mature caspase-3. Additives for caspase-3 assays
are, normally, used with much higher concentrations, compared
with that of caspase-3. In this context, we investigated whether
other additives may diminish the caspase-3 activation ability
of 1. Figure 6 shows the catalytic activities of mature caspase-3
in the absence and presence of CHAPS (3-[(3-chloroamido-
The addition of CHAPS to a reaction solution without 1
increases the catalytic activity of caspase-3 (compare lines
(a) and (b) in Figure 6). The addition of 1 to a solution in
the presence of CHAPS (0.1% (w/v)) slightly enhances the
caspase-3 activity (lines (b) and (c) in Figure 6). However, the
caspase-3 activity in the presence of both CHAPS and 1 is
1226 | Bull. Chem. Soc. Jpn. 2015, 88, 1221–1229 | doi:10.1246/bcsj.20150139
© 2015 The Chemical Society of Japan

the largest binding energies (Structures A, B, and C, see the
binding structures in Figure S10) were subjected to further MD
calculations, where protein fluctuations and solvent molecules
4
4
at 310.15 K are considered.
Figure 7 shows the MD simulation structures at the sta-
ble state (denoted as Structures AMD, BMD, and CMD: These
structures are calculated from the docking structure A, B, and
4
4
C, respectively). The time-courses of Cα-carbon root-mean-
square deviations (RMSDs) are shown in Figure S11. The
snapshot figures at defined simulation periods are demonstrated
in Figures S12 through S14.
In the simulations started from Structures A and B, the
calculations reached stable states at ca. 10 ns. Structures AMD
and BMD (Figures 7a and 7b, respectively) display a hydro-
gen bonding network between PAC-1, Glu124 and Arg164.
In particular, Structure AMD shows several hydrogen bonds
between 1 and the amino acid residues in both protomer units.
The non-phenolic benzene ring of 1, in contrast, tends to
dissociate from the protomer interface in these structures.
In the MD simulation started from Structure C (Figure 7c),
the stable state structure, Structure CMD, was obtained by 12.5-
ns simulation. In this state, two hydrogen bonds between 1 and
caspase-3 are shown at 1­Glu124 and 1­Asp135 with distances
smaller than 3.0 ¡ for each hydrogen bond. Although the non-
phenolic benzene ring stays inside the protomer interface
cavity, the side chain of Lys137 locates far from the benzene
ring. In place of Lys137, the pyrrolidine ring of Pro201 was
found to potentially interact with the benzene ring with a
distance of 3.0­3.5 ¡.
Time/s
Figure 6. Effect of CHAPS on the caspase-3-catalyzed
hydrolysis of AcDEVD p-nitroanilide in 50 mM MES
buffer (pH 6.0). a) Catalytic activity in the absence of
PAC-1 (1) nor CHAPS. b) Catalytic activity in the presence
of 0.1% (v/v) CHAPS (no 1). c) Catalytic activity in the
presence of 0.1% (v/v) CHAPS and 50 ¯M of 1. d)
Catalytic activity in the presence of 50 ¯M of 1 (no
CHAPS). Other conditions: 100 mM NaCl, [substrate] =
5
0 ¯M, [caspase-3] = 10 nM, DMSO 2% (v/v), 37 °C
under a N2 atmosphere. The y-axis was converted from the
absorbance at 405 nm to the product amount using ¾405 =
¹
1
¹1
9440 M cm .
much lower than that with only 1 (lines (c) and (d) in Figure 6).
In other words, the ability of 1 in caspase-3 activation is
significantly decreased by the presence of CHAPS. Apparently,
CHAPS and 1 competitively affect the caspase-3 activity.
According to previous papers, CHAPS, poly(ethylene glycol)
and hydrophobic peptides can modulate the catalytic activities
The computational study demonstrated above indicates that
1 is likely to interact with the protomer interface of mature
caspase-3. It is impossible, at present, to determine which
structure is most probable among the structures shown in
Figures 7a, 7b, and 7c. However, all three structures suggest
the participation of Glu124 and Arg164 in the fixation of 1.
These amino acid residues form a salt-bridge to maintain the
structure of the protomer interface. In fact, the mutation of
4
1,42
of several caspases.
It has also been reported that the bind-
ing of peptides to hydrophobic protomer interface in dimeric
caspases significantly affects the catalytic activity of the
enzymes. Consequently, a possible mechanism for the effect
of CHAPS on the caspase-3 activity is that the hydrophobic
steroid moiety in CHAPS binds to the hydrophobic core of
mature caspase-3 to interfere with the binding of 1.4
5
Glu124 to Ala decreases the catalytic activity of caspase-3.
Arg164 is located in a flexible loop shared with Cys163, a key
amino acid residue for the capase-3 activity. Consequently, the
interaction of 1 with Glu124/Arg164 will affect the positioning
of Cys163 at the active site. This result is similar to previously
reported docking simulations for the complexes of synthetic
molecules with mutant procaspase-3s, where the role of Arg164
3
The catalytic activity assays for cysteine proteases such as
caspase-3 are conducted in the presence of reducing reagents
and detergents. These additives are used with much higher
concentrations than that of targeted proteins. A series of experi-
mental results demonstrated above suggest that such additives
in reaction media may bring about difficulty in correctly evalu-
ating the ability of synthetic molecules targeted in research.
PAC-1 is Possible to Bind to the Protomer Interface of
Mature Caspase-3. On behalf of the fact that the saturation
dependency of the catalytic activity on the concentrations of
PAC-1 (1) and the observation of the differential UV­vis
spectrum, possible interaction modes between mature caspase-3
and 1 were searched by molecular dynamics (MD) calculation.
As the result of 100-run repeating calculations without pre-
restriction of interaction places in a VINA global docking
5
in the rearrangement of the active site structure and an allo-
4
5
steric effect on the binding of synthetic molecules to the
protomer interface were proposed. Considering the similarity
between facts in this and previous reports, we propose that
1 can bind to the protomer interface of mature caspase-3 to
bring about structural perturbation at the active site of mature
caspase-3. Although the X-ray crystallographic structure of
procaspase-3 is not available at present, 1 may influence the
structure of procaspase-3 in a similar manner to mature
caspase-3.
Conclusion
3
3
calculation, all calculated structures displayed that 1 may
bind to the interface between protomer units (Figure S9).
Among the 100 binding models, three possible structures with
Although PAC-1 has been regarded as a synthetic compound
that recovers the activation of procaspase-3 through a metal ion
chelation mechanism, the compound also affects the activity of
Bull. Chem. Soc. Jpn. 2015, 88, 1221–1229 | doi:10.1246/bcsj.20150139
© 2015 The Chemical Society of Japan | 1227

PAC-1 are essential for the enhancement of mature capase-3
activity. Typical additive reagents used in caspase-3 assays
abolish the ability of PAC-1 in the caspase-3 activation.
The MD simulation for the complex of mature caspase-3 and
PAC-1 suggests that PAC-1 is likely to bind to the protomer
interface of mature caspase-3. Glu124 and Arg164 at the
protomer interface may interact with PAC-1 through hydrogen
bonds, resulting in the induction of structural perturbation
at the active site. The structural perturbation to protomer
interfaces in multi-subunit proteins by synthetic molecules is
an effective strategy. The proposed binding structure of an
apoptosis-inducing reagent such as PAC-1 with caspase-3 will
provide a useful insight on the molecular design of next-
generation bioactive compounds.
This work was supported by MEXT, Japan (Grant-in-Aid for
Scientific Research (C) No. 24550189). The authors acknowl-
edge Mr. Leigh McDowell for kind advice during manuscript
preparation.
Supporting Information
This supporting information contains synthesis of com-
pound 5, SDS-PAGE profiles (after IPTG induction of protein
expression and after self-activation), the evaluation of apparent
molecular weights of proteins, blue native PAGE analysis, pH
titration for PAC-1, Michalis­Menten plots for the caspase-3-
mediated hydrolysis of AcDEVD p-nitroanilide, dependency
of caspase-3 activity on concentration of 2-ME, induced CD
spectra, mature caspase-3 and MD simulation details (docking
structures, time-courses of Cα RMSD, snapshot figures dur-
ing calculations). This material is available electronically on
J-STAGE.
References
1
P. R. Mittl, S. Di Marco, J. F. Krebs, X. Bai, D. S.
Karanewsky, J. P. Priestle, K. J. Tomaselli, M. G. Grütter, J. Biol.
Chem. 1997, 272, 6539.
2
5719.
3
H. R. Stennicke, G. S. Salvesen, J. Biol. Chem. 1997, 272,
2
K. M. Boatright, G. S. Salvesen, Curr. Opin. Cell Biol.
Figure 7. Molecular dynamics (MD) simulation structures
Structure AMD, BMD, and CMD) obtained by calculations
started from docking structures, Structures A, B, and C
presented in Figure S8). a) Structure AMD (calculated from
2
003, 15, 725.
W. C. Earnshaw, L. M. Martins, S. H. Kaufmann, Annu.
Rev. Biochem. 1999, 68, 383.
J. Walters, J. L. Schipper, P. Swartz, C. Mattos, A. C. Clark,
Biosci. Rep. 2012, 32, 401.
(
4
(
5
Structure A). b) Structure BMD (calculated from Struc-
ture B). c) Structure CMD (calculated from Structure C).
Structures AMD and BMD were obtained by 10-ns simu-
lation, whereas Structure CMD were obtained by 12.5-ns
simulation. Calculations were performed using YASARA 2
force field, assigned point charges of components at pH 6.0,
the aqueous solution model with 0.9% NaClaq concen-
6
N. D. Thomsen, J. T. Koerber, J. A. Wells, Proc. Natl.
Acad. Sci. U.S.A. 2013, 110, 8477.
7
K. Bose, C. Pop, B. Feeney, A. C. Clark, Biochemistry
003, 42, 12298.
P. Karki, J. Lee, S. Y. Shin, B. Cho, I.-S. Park, Arch.
Biochem. Biophys. 2005, 442, 125.
S. Roy, C. I. Bayly, Y. Gareau, V. M. Houtzager, S.
2
8
+
¹
tration and additional Na and Cl for neutralization of
charges at 310.15 K. Hydrogen atoms, water molecules and
salt atoms are omitted in figures for simplicity.
9
Kargman, S. L. C. Keen, K. Rowland, I. M. Seiden, N. A.
Thornberry, D. W. Nicholson, Proc. Natl. Acad. Sci. U.S.A. 2001,
9
8, 6132.
D. W. Wolan, J. A. Zorn, D. C. Gray, J. A. Wells, Science
the mature caspase-3 as well as its zymogen. The enhancement
of caspase-3 activity by PAC-1 with and without the addition
of Zn indicates other functions of PAC-1 for the caspase-3
activation. A possible mechanism is the direct binding of
PAC-1 to the protein core. The benzyl and phenolic groups in
1
0
2
009, 326, 853.
2
+
11 J. A. Zorn, H. Wille, D. W. Wolan, J. A. Wells, J. Am.
Chem. Soc. 2011, 133, 19630.
12 J. A. Zorn, D. W. Wolan, N. J. Agard, J. A. Wells, J. Biol.
1228 | Bull. Chem. Soc. Jpn. 2015, 88, 1221–1229 | doi:10.1246/bcsj.20150139
© 2015 The Chemical Society of Japan

Chem. 2012, 287, 33781.
O. Julien, M. Kampmann, M. C. Bassik, J. A. Zorn, V. J.
30 Y. Hayashi, S. Nagao, H. Osuka, H. Komori, Y. Higuchi, S.
Hirota, Biochemistry 2012, 51, 8608.
1
3
Venditto, K. Shimbo, N. J. Agard, K. Shimada, A. L. Rheingold,
B. R. Stockwell, J. S. Weissman, J. A. Wells, Nat. Chem. Biol.
31 C. N. Pace, F. Vajdos, L. Fee, G. Grimsley, T. Gray, Protein
Sci. 1995, 4, 2411.
2
014, 10, 969.
J. A. Hardy, J. Lam, J. T. Nguyen, T. O’Brien, J. A. Wells,
Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 12461.
32 E. Krieger, G. Koraimann, G. Vriend, Protein: Struct.,
Funct., Bioinf. 2002, 47, 393.
33 O. Trott, A. J. Olson, J. Comput. Chem. 2010, 31, 455.
34 E. Krieger, K. Joo, J. Lee, J. Lee, S. Raman, J. Thompson,
M. Tyka, D. Baker, K. Karplus, Protein: Struct., Funct., Bioinf.
2009, 77 (Suppl 9), 114.
1
4
15
16
17
A. Shen, Mol. BioSyst. 2010, 6, 1431.
J. Ohkanda, Chem. Rec. 2013, 13, 561.
C. Pop, B. Feeney, A. Tripathy, A. C. Clark, Biochemistry
2
003, 42, 12311.
J. Walters, C. Pop, F. L. Scott, M. Drag, P. Swartz, C.
Mattos, G. S. Salvesen, A. C. Clark, Biochem. J. 2009, 424, 335.
S. H. MacKenzie, J. L. Schipper, E. J. England, M. E.
Thomas, III, K. Blackburn, P. Swartz, A. C. Clark, Biochemistry
013, 52, 6219.
K. S. Putt, G. W. Chen, J. M. Pearson, J. S. Sandhorst,
35 P. Andrews, Biochem. J. 1964, 91, 222.
1
8
36 D. K. Perry, M. J. Smyth, H. R. Stennicke, G. S. Salvesen,
P. Duriez, G. G. Poirier, Y. A. Hannun, J. Biol. Chem. 1997, 272,
18530.
37 L. K. Charkoudian, D. M. Pham, K. J. Franz, J. Am. Chem.
Soc. 2006, 128, 12424.
38 A. Chakrabartty, T. Kortemme, S. Padmanabhan, R. L.
Baldwin, Biochemistry 1993, 32, 5560.
39 F. Zsila, FEBS Lett. 2003, 539, 85.
40 According to previous studies of PAC-1, de-allyl analogs of
PAC-1 show similar activities of the procaspase-3 activation to that
of PAC-1 (see ref 20). Conseqently, compound 5 is also inactive
for the procaspase-3 activation.
41 M. Garcia-Calvo, E. P. Peterson, D. M. Rasper, J. P.
Vaillancourt, R. Zamboni, D. W. Nicholson, N. A. Thornberry,
Cell Death Differ. 1999, 6, 362.
1
9
2
2
0
M. S. Hoagland, J.-T. Kwon, S.-K. Hwang, H. Jin, M. I.
Churchwell, M.-H. Cho, D. R. Doerge, W. G. Helferich, P. J.
Hergenrother, Nat. Chem. Biol. 2006, 2, 543.
Q. P. Peterson, D. C. Hsu, D. R. Goode, C. J. Novotny,
R. K. Totten, P. J. Hergenrother, J. Med. Chem. 2009, 52, 5721.
Q. P. Peterson, D. R. Goode, D. C. West, K. N. Ramsey,
J. J. Y. Lee, P. J. Hergenrother, J. Mol. Biol. 2009, 388, 144.
D. C. West, Y. Qin, Q. P. Peterson, D. L. Thomas, R.
2
1
2
2
2
3
Palchaudhuri, K. C. Morrison, P. W. Lucas, A. E. Palmer, T. M.
Fan, P. J. Hergenrother, Mol. Pharm. 2012, 9, 1425.
42 K. L. Huber, S. Ghosh, J. A. Hardy, Pept. Sci. 2012, 98,
451.
2
4
R. C. Botham, T. M. Fan, I. Im, L. B. Borst, L. Dirikolu,
P. J. Hergenrother, J. Am. Chem. Soc. 2014, 136, 1312.
K. M. Boyle, J. P. Irwin, B. R. Humes, S. W. Runge, J. Cell.
Biochem. 1997, 67, 231.
Y. Urano, D. Asanuma, Y. Hama, Y. Koyama, T. Barrett, M.
43 Although hydrophobic interactions between PAC-1 and
CHAPS in salt-containing solutions are also a possible mechanism
for the reduction of the PAC-1 ability in the enhancement of the
caspase-3 activity, dynamic light scattering measurements for
PAC-1 solutions indicated no formation of nanoparticles and
precipitation in the presence of CHAPS.
2
5
2
6
Kamiya, T. Nagano, T. Watanabe, A. Hasegawa, P. L. Choyke, H.
Kobayashi, Nat. Med. 2009, 15, 104.
2
7
C. Cade, P. Swartz, S. H. MacKenzie, A. C. Clark,
Biochemistry 2014, 53, 7582.
H. J. Kang, Y. Lee, Y.-J. Jeong, K. Park, M. Jang, S. G.
Park, K.-H. Bae, M. Kim, S.-J. Chung, BMC Biotechnol. 2008, 8,
44 Because of the symmetric whole structure of mature
caspase-3, a vertically reserved stucture for the binding PAC-1 is
also possible in each binding structure. Considering that the two
complex structures are intrinsically identical, one of the two
orientations is shown in each structure figure.
2
8
9
2.
2
9
H. R. Stennicke, G. S. Salvesen, Methods (Amsterdam,
45 J. L. Schipper, S. H. MacKenzie, A. Sharma, A. C. Clark,
Biophys. Chem. 2011, 159, 100.
Neth.) 1999, 17, 313.
Bull. Chem. Soc. Jpn. 2015, 88, 1221–1229 | doi:10.1246/bcsj.20150139
© 2015 The Chemical Society of Japan | 1229