Tài liệu Báo cáo khoa học: Plasticity of S2–S4 specificity pockets of executioner caspase-7 revealed by structural and kinetic analysis pdf

Plasticity of S2–S4 specificity pockets of executioner
caspase-7 revealed by structural and kinetic analysis
Johnson Agniswamy, Bin Fang and Irene T. Weber
Department of Biology, Molecular Basis of Disease, Georgia State University, Atlanta, GA, USA
Caspases, the key effector molecules in apoptosis, are
potential targets for pharmacological modulation of
cell death. Uncontrolled apoptosis due to enhanced
caspase activity occurs in nerve crush injury, stroke
and neurodegenerative diseases such as Alzheimer’s,
Parkinson’s and Huntington’s diseases [1–3]. On the
other hand, inadequate caspase activity is implicated in
cancer, autoimmune diseases and viral infections [4–6],
and a number of potential drugs are being developed
for selective induction of apoptosis in cancer cells [7,8].
The substrate-based peptide inhibitor zVAD-fmk pro-
vides substantial protection against stroke, myocardial
infarction, osteoarthritis, hepatic injury, sepsis, and
amyotrophic lateral sclerosis in animal models [9–11].
Small nonpeptide inhibitors are preferred for their
superior metabolic stability and cell permeability. Two
nonpeptide inhibitors are currently in phase II clinical
trials: IDN6556 for treatment of acute-tissue injury
disease and liver diseases [12], and VX-740 for treat-
ment of rheumatoid arthritis [13]. Knowledge of the
molecular basis for substrate specificity of caspases is
critical for design of therapeutic agents for selective
control of cell death.
Caspases are cysteine proteases that hydrolyze the
peptide bond after an aspartate residue [14–17]. Thir-
teen human caspases have been cloned and character-
ized to varying extents [18,19]. Caspases are classified

into three groups based on their function and
Keywords
allosteric site; apoptosis; cysteine protease;
enzyme
Correspondence
I. T. Weber, Department of Biology, Georgia
State University, PO Box 4010, Atlanta,
GA 30302, USA
Fax: +1 404 413 5301
Tel: +1 404 413 5411
E-mail:
Database
The atomic coordinates and structure
factors have been deposited in the Protein
Data Bank under the accession codes 2QL5
for caspase-7 ⁄ DMQD, 2QL9 for caspase-7 ⁄
DQMD, 2QLF for caspase-7 ⁄ DNLD, 2QL7
for caspase-7 ⁄ IEPD, 2QLB for caspase-7 ⁄
ESMD, and 2QLJ for caspase-7 ⁄ WEHD
(Received 18 April 2007, revised 3 July
2007, accepted 17 July 2007)
doi:10.1111/j.1742-4658.2007.05994.x
Many protein substrates of caspases are cleaved at noncanonical sites in
comparison to the recognition motifs reported for the three caspase sub-
groups. To provide insight into the specificity and aid in the design of
drugs to control cell death, crystal structures of caspase-7 were determined
in complexes with six peptide analogs (Ac-DMQD-Cho, Ac-DQMD-Cho,
Ac-DNLD-Cho, Ac-IEPD-Cho, Ac-ESMD-Cho, Ac-WEHD-Cho) that
span the major recognition motifs of the three subgroups. The crystal
structures show that the S2 pocket of caspase-7 can accommodate diverse

residues. Glu is not required at the P3 position because Ac-DMQD-Cho,
Ac-DQMD-Cho and Ac-DNLD-Cho with varied P3 residues are almost as
potent as the canonical Ac-DEVD-Cho. P4 Asp was present in the better
inhibitors of caspase-7. However, the S4 pocket of executioner caspase-7
has alternate regions for binding of small branched aliphatic or polar resi-
dues similar to those of initiator caspase-8. The observed plasticity of the
caspase subsites agrees very well with the reported cleavage of many pro-
teins at noncanonical sites. The results imply that factors other than the
P4–P1 sequence, such as exosites, contribute to the in vivo substrate speci-
ficity of caspases. The novel peptide binding site identified on the molecular
surface of the current structures is suggested to be an exosite of caspase-7.
These results should be considered in the design of selective small molecule
inhibitors of this pharmacologically important protease.
Abbreviation
PARP, poly(ADP-ribose) polymerase.
4752 FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS
prodomain structures. Caspases-1, 4, 5 and 11 have
roles in cytokine maturation and inflammatory
responses and consequently are called inflammatory
caspases (group I) [20]. The other family members are
involved in apoptosis. Caspases-2, 8, 9, 10 and 12
function upstream within the apoptotic signaling
pathways and are termed initiator caspases (group II).
Caspases-3, 6, 7 and 14 are activated by initiator
caspases and act as the immediate executioners of the
apoptotic process. These caspases are termed execu-
tioner or effector caspases (group III). The caspases
are reported to recognize tetrapeptide motifs in their
substrates. Caspases-1, 4 and 5 prefer the tetrapeptide
WEHD, whereas caspase-2, 3 and 7 have a preference

for DEXD, and caspase-6, 8 and 9 prefer (L⁄V)EXD
[21,22]. However, a peptide library search with sub-
strate-phage identified DLVD with 170% faster hydro-
lysis by caspase-3 than the canonical DEVD peptide,
thereby challenging the specificity of group III execu-
tioner caspases [23]. Similarly, screening of peptide
inhibitors based on amino acid positional fitness scores
predicted DFPD as a potent inhibitor of caspase-7
[24]. Also, the DNLD peptide was shown to have simi-
lar potent inhibitory activity on caspase-3 as the
canonical DEVD [24]. The crystal structure of caspase-
8 in complex with the assumed non-optimal inhibitor
Ac-DEVD-Cho questioned the original specificity clas-
sification [25]. Moreover, the specificity extends beyond
the P4–P1 tetrapeptide. Our previous study has identi-
fied a preference of caspase-3 for hydrophobic P5 resi-
dues, unlike caspase-7 [26], but similar to caspase-2
[27]. Recently, an increasing number of caspase sub-
strates have been found to be cleaved at noncanonical
sites [28]. These studies undermine the current classifi-
cation of caspase specificity into three groups. There-
fore, new structural data are needed to determine the
detailed interactions that define caspase specificity.
The structure of caspase-7 in complex with the
canonical tetrapeptide inhibitor DEVD is known [29].
In further studies of the molecular basis for the sub-
strate specificity of executioner caspases, we have
determined the crystal structures of human recombi-
nant caspase-7 in complexes with six tetrapeptide
analogs: Ac-DMQD-Cho, Ac-DQMD-Cho, Ac-DNLD-

Cho, Ac-IEPD-Cho, Ac-ESMD-Cho and Ac-WEHD-
Cho. The sequences of these peptidyl inhibitors span
the range of recognition motifs reported for the three
groups of caspases. These new structures reveal that
non-optimal peptides for group III and optimal pep-
tides of group I and II can bind and form favorable
interactions within S2, S3 and S4 subsites of group III
caspase-7. Also, a new peptide binding site was identi-
fied for on the molecular surface distal to the active
site. The results demonstrate the plasticity of substrate
recognition by caspase-7, and will be valuable for the
design of inhibitors of this pharmacologically impor-
tant enzyme.
Results
Inhibition of caspase-7 by tetrapeptide aldehydes
Six substrate analog reversible inhibitors, Ac-DMQD-
Cho, Ac-DQMD-Cho, Ac-DNLD-Cho, Ac-IEPD-Cho,
Ac-ESMD-Cho and Ac-WEHD-Cho, which span the
three functional and phylogenetic classes of caspase sub-
strates, were evaluated for inhibition of caspase activity.
The selection of tetrapeptide sequences used in the pres-
ent study was based on the known protein cleavage sites
of caspases. DMQD is the reported executioner caspase
cleavage site of protein kinase C delta, and caspase
cleavage at DQMD in baculovirus p35 transforms it to
a pancaspase inhibitor [30,31]. ESMD forms the N-ter-
minal cleavage site in caspase-3, whereas DNLD was
identified as a potent substrate for executioner caspases
by computational studies [24,32]. IEPD has been identi-
fied as the optimal cleavage sequence of granzyme B

and caspase-8, whereas WEHD forms the optimal sub-
strate sequence of inflammatory caspases [22]. The
kinetic parameters of caspase-7 were measured for the
canonical substrate Ac-DEVD-pNA. The K
m
value for
this substrate is 54.5 ± 2.18 lm whereas the k
cat
⁄ K
m
is
4071.6 mm
)1
Æmin
)1
. The inhibitory potency of the six
aldehydes together with the canonical Ac-DEVD-Cho
can be divided into two groups based on the K
i
values.
The first group consists of stronger inhibitors with rela-
tively low K
i
values in the order of Ac-DEVD-Cho
(0.7 ± 0.03 nm)  Ac-DQMD-Cho (0.94 ± 0.04 nm)
< Ac-DNLD-Cho (1.4 ± 0.06 nm)<< Ac-DMQD-
Cho (8.0 ± 0.3 nm). The second group contains weaker
inhibitors with much higher K
i
values in the order of

Ac-IEPD-Cho (550 ± 22 nm) < Ac-ESMD-Cho (1300
±50nm)<< Ac-WEHD-Cho (4400 ± 175 nm).
Overall structure of the six caspase-7 complexes
Caspase-7 was crystallized in complex with six sub-
strate analog reversible inhibitors, Ac-DMQD-Cho,
Ac-DQMD-Cho, Ac-DNLD-Cho, Ac-IEPD-Cho, Ac-
ESMD-Cho and Ac-WEHD-Cho. All the complexes
crystallized in the trigonal space group of P3
2
21
(Table 1). The structures were refined to the resolu-
tions of 2.14–2.8 A
˚
and R-factors from 18.7–21.2%.
The overall structure of the six independently refined
complexes is essentially identical with a complete
catalytic unit of two p20–p10 heterodimers in the
J. Agniswamy et al. Plasticity of caspase-7 specificity pockets
FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS 4753
asymmetric unit (Fig. 1A). The two heterodimers are
arranged side by side in the opposite orientation to
form a central 12 stranded b-sheet surrounded by 10 a
helices. The refined overall structures are very similar
to the reported structure of caspase-7 with the canoni-
cal inhibitor DEVD [29]. The complete catalytic unit
of two heterodimers can be superimposed with that of
caspase-7 ⁄ DEVD with rmsd of 0.28–0.38 A
˚
for 460
topologically equivalent Ca atoms. The individual

heterodimers were even more similar to those of cas-
pase-7 ⁄ DEVD complex showing rmsd of 0.24–0.33 A
˚
and 0.24–0.35 A
˚
for 230 equivalent Ca atoms. The
conformations of the four prominent surface loops
that form the substrate-binding cleft of caspase-7 agree
well with those of caspase-7 ⁄DEVD complex. There-
fore, the substrate analogs of diverse sequences can be
accommodated in the substrate binding cleft without
major changes in overall conformation.
Conformation of peptide analogs
The peptide analog inhibitors of the six complexes were
clearly visible in the electron density maps. The electron
density for the inhibitor in the caspase-7 ⁄DQMD struc-
ture is shown in Fig. 1B. The peptide inhibitors adopt
an extended conformation in all the complexes. The
inhibitors bind with thiohemiacetal bonds between the
aldehyde group (-CHO) and the mercapto group (-SH)
of Cys186 of caspase-7. The peptide inhibitors make
most of their contacts with the residues from the p10
subunit, except for those with the catalytic dyad
(Cys186 and His144) and the residues anchoring the
aspartate residue in the P1 position (Arg87, Ala145 and
Gln184) which are in the p20 chain. The inhibitors bind
the two active sites in the heterotetramer in a similar
fashion except for the alternate conformations for
P3 Ser and P4 Trp in ESMD and WEHD complexes,
which are described below. The Ca positions of the

peptides from all the six complexes are very similar and
superimpose with rmsd of 0.08–0.41 A
˚
. The peptides
and the interacting caspase-7 residues have very similar
conformations for the three stronger inhibitors,
whereas more structural variation is observed for the
weaker inhibitors compared to the canonical caspase-
7 ⁄ DEVD structure (Fig. 2A,B). The main chain atoms
of the inhibitors in all the six complexes exhibit similar
hydrogen bond interactions, except for P4 N in cas-
pase-7 ⁄ WEHD that cannot interact with the carbonyl
of Gln276 (Fig. 3, supplementary Table S1). Caspases
are unique among proteases in their stringent specificity
Table 1. Crystallographic data collection and refinement statistics.
Caspase-7 ⁄
Ac-DQMD-Cho
Caspase-7 ⁄
Ac-IEPD-Cho
Caspase-7 ⁄
Ac-ESMD-Cho
Caspase-7 ⁄
Ac-DMQD-Cho
Caspase-7 ⁄
Ac-DNLD-Cho
Caspase-7 ⁄
Ac-WEHD-Cho
Protein databank code 2QL9 2QL7 2QLB 2QL5 2QLF 2QLJ
Space group P3
2

21 P3
2
21 P3
2
21 P3
2
21 P3
2
21 P3
2
21
a ¼ b(A
˚
) 87.16 87.94 88.25 87.26 87.46 88.50
c(A
˚
) 187.42 187.55 188.29 187.71 185.83 187.22
b (°) 120 120 120 120 120 120
Resolution range 50–2.14 50–2.4 50–2.25 50–2.34 50–2.8 50–2.6
Total observations 295 498 90 499 126 436 136 298 63 618 86 151
Unique reflections 44 778 31 049 38 990 32 117 18 047 24 974
Completeness 97.6 (80.2)
a
92.3 (59.7) 94.3 (68.3) 90.2 (52.0) 90.3 (91) 92.1 (60.7)
<I ⁄ r(I)> 24.0 (3.2) 12.2 (2.2) 16.0 (2.6) 15.5 (3.0) 13.8 (2.6) 16.7(3.3)
R
sym
(%)
b
7.2 (31.9) 9.8 (39.2) 9.0 (34.3) 9.8 (34.2) 9.2 (43.0) 6.7 (34.8)

Refinement statistics
Resolution range 50–2.14 50–2.4 50–2.25 50–2.34 50–2.8 50–2.6
R
cryst
(%)
c
19.1 19.6 18.7 21.2 19.6 19.6
R
free
(%)
d
22.5 23.7 22.3 23.3 22.9 23.4
Mean B ) factor (A
˚
2
) 45.3 51.0 49.4 62.3 56.3 70.7
Number of atoms
Protein 3825 3823 3821 3828 3828 3853
Water 296 220 215 125 59 52
Citrate ion
rmsd
111101
Bond length (A
˚
) 0.006 0.006 0.006 0.006 0.006 0.007
Angles (°) 1.3 1.3 1.3 1.3 1.3 1.3
a
Values in parentheses are given for the highest resolution shell.
b
R

sym
¼ S
hkl
|I
hkl
) ÆI
hkl
æ| ⁄S
hkl
I
hkl
.
c
R ¼ S|F
obs
) F
cal
| ⁄SF
obs
.
d
R
free
¼ S
test
(|F
obs
| ) |F
cal
|)

2
⁄S
test
|F
obs
|
2
.
Plasticity of caspase-7 specificity pockets J. Agniswamy et al.
4754 FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS
for P1 Asp. The side chain atoms of aspartate at P1
have very similar positions and superpose well.
However, the side chain positions of P2, P3 and P4 res-
idues of the peptides differ significantly in the com-
plexes. These differences will be discussed separately
for each subsite.
Ace
P4 Asp
P3 Gln
P2 Met
P1 Asp
Cys 186
p10
A
B
p20
p10’
p20’
Fig. 1. Structure of caspase-7 and peptide analog inhibitors. (A)
Ribbon diagram of caspase-7 tetrameric assembly. The p20, p10

subunits and their symmetry related equivalents p20¢ and p10¢ are
shown in green, blue, red and cyan, respectively. The catalytic
cysteines in the p20 subunits are shown in green and red ball and
stick. The magenta ball and stick model represents the Ac-DQMD-
CHO inhibitor. (B) 2F
o
-F
c
electron density map of Ac-DQMD-CHO in
the caspase-7 ⁄ DQMD structure contoured at a level of 1r. The cat-
alytic Cys186 of caspase-7 forms a thiohemiacetal bond with the
acetyl group of the inhibitor.
Cys 186A
B
P1
P2
P3
P4
Tyr 230
Arg 87
Pro 235
Arg 233
Trp 232
Phe 282
Gln 276
T
rp 240
Cys 186
P1
Arg 87

P2
P3
Trp 232
Trp 240
Gln 276
Phe 282
Tyr 230
Pro 235
P4
Fig. 2. Superposition of inhibitors bound at the active site.
(A) Superposition of stronger inhibitors and surrounding caspase-7
residues. The inhibitor and active site residues in the caspase-
7 ⁄ DEVD complex (protein databank code: 1F1J) are colored by ele-
ment type whereas those of caspase-7 ⁄ DQMD, caspase-7 ⁄ DMQD
and caspase-7 ⁄ DNLD are colored blue, cyan and green, respec-
tively. The inhibitors are in ball and stick representation and the cas-
pase residues are shown in a stick model. (B) Superposition of
weaker inhibitors and active site residues. The caspase-7 ⁄ IEPD,
caspase-7 ⁄ ESMD, caspase-7 ⁄ WEHD complexes are colored red,
magenta and yellow, respectively. For sake of clarity, residues
His144 and Gln184 in S1 subsite are not shown.
J. Agniswamy et al. Plasticity of caspase-7 specificity pockets
FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS 4755
S2 subsite
By contrast to the highly specific S1 subsite, the S2
pocket of caspases is the only subsite to show substan-
tial alteration upon substrate binding, which indicates
its importance in substrate recognition and regulation
of activity. In the unliganded structure of caspase-3,
the side chain of a tyrosine residue occupies the S2

pocket and must rotate approximately 90° to accom-
modate the P2 residue [33]. Similarly, the S2 pocket
Fig. 3. Schematic diagram of caspase-7 interactions with peptide analog inhibitors. (A) Caspase-7 ⁄ DQMD. (B) Caspase-7 ⁄ DNLD. (C) Cas-
pase-7 ⁄ DMQD. (D) Caspase-7 ⁄ IEPD. (E) Caspase-7 ⁄ ESMD. (F) Caspase-7 ⁄ WEHD. Thicker lines represent the peptide analog inhibitors. The
inhibitors are covalently bound to catalytic Cys186. Dashed lines represent hydrogen bonds and salt bridges, while curved lines indicate van
der Waals interactions.
Plasticity of caspase-7 specificity pockets J. Agniswamy et al.
4756 FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS
was blocked in the crystal structure of caspase-3 in
complex with the inhibitor of apoptosis protein XIAP
[34]. The S2 subsites of caspase-7 and 3 are formed by
identical aromatic residues (Tyr230, Trp232 and
Phe282 in caspase-7). The inflammatory and initiator
caspases have larger S2 subsites due to the substitution
of Tyr230 with Val and tolerate bulkier side chains.
The S2 subsites in caspase-3 and 7 were predicted to
preferentially accommodate small b-branched aliphatic
residues such as Val and Thr29.
To investigate the plasticity of the S2 pocket of
caspase-7, it was probed with P2 Gln, Met, Pro, Leu
and His in the six complexes with different tetra-
peptides (Figs 2 and 3). The complexes of caspase-7
with DNLD, DQMD, and ESMD contain the long
aliphatic Leu and Met at P2. The S2 subsite accommo-
dates Met and Leu with good hydrophobic interactions
by the rotation of the Phe282 and Tyr230 side chains to
expand the pocket (Fig. 2 and supplementary Fig. S1).
The Met side chains exhibit different conformations in
the complexes with DQMD and ESMD as shown in
supplementary Fig. S1. Caspase-7 ⁄ DMQD contains

the polar P2 Gln, which is a mismatch in the hydro-
phobic S2 pocket. However, the CB and CG atoms of
Gln form favorable van der Waals interactions with
residues in the subsite, similar to those of Met P2, with
minor changes in the S2 aromatic side chains (supple-
mentary Fig. S1). The polar atoms of Gln are directed
away from the pocket. The smaller hydrophobic
P2 Pro is present in the caspase-7 ⁄IEPD complex. The
Tyr230 side chain has a similar conformation to that
in the canonical caspase-7 ⁄DEVD structure, but the
side chains of Trp232 and Phe282 adjust to form
favorable van der Waals interactions with P2 Pro
(Fig. 4A). The S2 pockets of caspase-7 and 3 were pre-
dicted not to accommodate aromatic residues. A com-
putational study on amino acid preference at different
subsites of caspase-7 based on positional fitness scores
predicted His as the amino acid with the least score
for binding in the S2 subsite [24]. However, the P2 His
in the caspase-7 ⁄ WEHD complex can clearly be
accommodated in the S2 subsite, although there are
relatively large movements of the three aromatic side
chains forming S2 (Fig. 4B). The CB of histidine is in
a similar position as the CB of valine in the caspase-
7 ⁄ DEVD complex. The v
2
angle of Tyr230 rotates
more than 70° to form an aromatic stacking interac-
tion with the P2 His. This stacking interaction is fur-
ther strengthened by the hydrogen bond between NE2
of P2 His and OH of Tyr230 in one binding site. These

structures demonstrate that the size of the S2 pocket
of caspase-7 can be enlarged or reduced by rotating
Tyr 230 A
D
B
E
C
F
Phe 282
Trp 232
Trp 232
Phe 282
Tyr 240
P2 Pro
P2 His
Arg 233
P3 Ser
Pro 235
Val 86
Gln 276
Trp 232 Trp 232 Trp 240
Trp 240
P4 Glu
Gln 276
P4 Ile
Trp 240
Trp 420
P4 Ile
P4 Ile
P3 Glu

P3 Glu
Trp 232
Trp 412
P2 Pro
P3 Thr
Fig. 4. Key variations in S2, S3 and S4 subsites of caspase-7. (A) Pro in S2 subsite of casepase-7 ⁄ IEPD. The new structure is colored by ele-
ment type and caspase-7 ⁄ DEVD is shown in cyan. (B) His in the S2 subsite of caspase-7 ⁄ WEHD. Dashed lines represent the hydrogen bond
and ion pair interactions. (C) Ser in the S3 subsite of caspase-7 ⁄ ESMD. (D) Glu in the S4 subsite caspase-7 ⁄ ESMD and (E) Ile in the S4 sub-
site of caspase-7 ⁄ IEPD. (F) Comparison of P4 Ile in the S4 subsites of caspase-7 and caspase-8 ⁄ IETD. The caspase-7 residues are colored
by atom type, whereas those of caspase-8 are shown in green.
J. Agniswamy et al. Plasticity of caspase-7 specificity pockets
FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS 4757
the v angles of Tyr230, Trp232 and Phe282. This
adjustment enables caspase-7 to accommodate smaller
or larger aliphatic as well as long polar or aromatic
residues in the S2 subsite.
S3 subsite
Glutamic acid is considered to be the preferred P3
residue for all human caspases [21]. The P3 Glu is
anchored by multiple interactions with the conserved
Arg233 of caspase-7, which also is critical for binding
of P1 Asp. However, several physiological substrates
of caspase have been identified with different amino
acids at the P3 position (supplementary Table S2) [32].
Therefore, the specificity of the S3 subsite in caspase-7
was probed with Met, Gln, Asn, Glu and Ser in the
six complexes (Figs 2 and 3). The main chain amide
and carbonyl oxygen of the P3 residue form strong
hydrogen bonds with the main chain carbonyl and
amide of Arg233 in all structures. These hydrogen

bonds stabilize and fix the P3 residue in position, inde-
pendent of the type of side chain. The side chain of
the canonical glutamic acid in the caspase-7 ⁄IEPD and
caspase-7 ⁄ WEHD complexes forms favorable ionic
interactions with the guanidinium group of Arg233
(supplementary Fig. S2). The long polar P3 residues
Gln and Asn in complexes with DQMD and DNLD
form similar hydrogen bond interactions with Arg233.
However, the hydrophobic side chain of P3 Met in the
complex with DMQD exhibits different conformations
in the two active sites. In one conformation, it is direc-
ted into the subsite whereas, in the second active site,
the Met side chain is directed out of the S3 groove
and has hydrophobic interactions with Pro235. The
CB and CG atoms of Met P3 are positioned similarly
to the equivalent atoms of P3 Glu. The smaller polar
P3 Ser in the caspase-7 ⁄ ESMD complex is positioned
to form a hydrogen bond interaction with Arg233 in
one binding site, in addition to water-mediated interac-
tions with both Arg233 and Val86 (Fig. 4C). In sum-
mary, the P3 residue is anchored by the main chain
hydrogen bonds with Arg233, and key interactions
necessary for the specificity of the S3 subsite are con-
served for both smaller and longer polar P3 residues.
Moreover, hydrophobic P3 residues can form favor-
able van der Waals interactions with Pro235.
S4 subsite
The structural divergence at the S4 subsite provides
major specificity conferring elements to the three
groups of caspases. The S4 subsite of inflammatory

caspases of group I is an extended, shallow hydropho-
bic depression suitable for the binding of a P4 Trp
residue. By contrast, the two bulky tryptophan resi-
dues lining the S4 subsite of group II and III apoptotic
caspases considerably reduce the size of the groove.
Experimental and theoretical studies have suggested
that caspase-3 and 7 have very high specificity for
aspartic acid at P4 [22,24]. Absolute specificity of Asp
over Glu at P4 was shown for caspase-7 using fluores-
cent substrates [35]. Caspase-8, which belongs to
group II, was shown to have high specificity for Leu at
the P4 position [22]. However, structural analysis sub-
sequently showed that caspase-8 tolerates both hydro-
phobic Ile and acidic Asp at P4 [25].
The specificity of the S4 subsite was probed with
Asp, Ile, Glu and Trp in the six caspase-7 structures,
and demonstrated greater flexibility in this subsite
compared to S2 and S3 (Figs 2 and 3). The main chain
amide of P4 residue forms a hydrogen bond with the
carbonyl oxygen of Gln276 in all the structures except
caspase-7 ⁄ WEHD. The side chain of P4 Asp binds cas-
pase-7 through interaction with the main chain amide
of Gln276. The reported interaction between the side
chain of Gln276 and P4 Asp in the caspase-7 ⁄DEVD
complex [29] is absent in all these complexes, even in
those with P4 Asp. A network of three ordered water
molecules deep in the subsite interacts with the side
chain of P4 Asp, suggesting that caspase-7 can accom-
modate residues larger than Asp at P4. The P4 Glu in
the caspase-7⁄ESMD complex extends into the subsite

with the formation of a hydrogen bond with the main
chain amide of Gln276 (Fig. 4D). The P4 Glu also
forms a hydrogen bond with the NE1 of Trp240.
P4 Trp in the caspase-7 ⁄ WEHD complex was accom-
modated in the S4 subsite by rotation of the side
chains of Trp232 and Ser234 in addition to 1.2–1.6 A
˚
shifts of residues 278–281. The P4 Trp exhibits differ-
ent orientations in the two binding sites. However,
there is no hydrogen bond between the P4 amide and
carbonyl oxygen of Gln276 in both orientations.
Therefore, the structural changes confirm that large
aromatic residues are not favored at the S4 subsite of
caspase-7 (supplementary Fig. S3). The Ile P4 in the
caspase-7 ⁄ IEPD complex fits snugly between Trp232
and Trp240 and forms favorable van der Waals inter-
actions with Trp232, Trp240 and CB of Ser275
(Fig. 4E). Thus, Trp232 plays a dual role in caspase-7
by interacting with both P2 and small, branched ali-
phatic P4 residues. Interestingly, a similar hydrophobic
S4 subsite and mode of binding of P4 Ile was observed
with the initiator caspase-8 where the P4 Ile side chain
lies between Trp420 and Tyr412 (Fig. 4F). This struc-
tural analysis implies that the S4 subsite of caspase-7
is well suited for Glu as well as Asp at P4. Furthermore,
Plasticity of caspase-7 specificity pockets J. Agniswamy et al.
4758 FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS
small aliphatic residues are well tolerated in the S4
pocket and form favorable hydrophobic interactions
with Trp232 and Trp240. However, the large aromatic

Trp cannot be accommodated without structural
changes and loss of a hydrogen bond interaction.
Correlation of structural interactions and
inhibition of the peptides
The 6 substrate analog inhibitors can be divided into
strong (Ac-DEVD-Cho  Ac-DQMD-Cho < Ac-DNLD-
Cho << Ac-DMQD-Cho) and weak (Ac-IEPD-Cho <
Ac-ESMD-Cho << Ac-WEHD-Cho) inhibitors (Fig. 2).
The stronger inhibitors all contain Asp at P4. However,
the predicted specificity of caspase-7 for Glu at P3
position is not required because Ac-DMQD-Cho,
Ac-DQMD-Cho and Ac-DNLD-Cho with different P3
residues are almost as potent as the canonical Ac-
DEVD-Cho with P3 Glu. Similarly, the structural and
kinetic data show that S2 can accommodate longer P2
residues than the predicted small, b-branched Val and
Thr. The intermediate K
i
value of Ac-IEPD-Cho con-
firms the structural identification of the small aliphatic
binding region of the S4 subsite of caspase-7. The struc-
ture of caspase-7 ⁄IEPD shows only one hydrogen bond
between the main chain atoms of caspase-7 and P4, and
lacks the second hydrogen bond observed for the P4
Asp side chain in the four better inhibitors. The second
weakest inhibitor is Ac-ESMD-Cho, although the crys-
tal structure shows little change compared to the com-
plexes with more potent inhibitors. Although Glu is
structurally well suited at the P4 position due to an addi-
tional weak hydrogen bond compared to Asp, the Ac-

ESMD-cho is a weaker inhibitor than those with P4
Asp. The polar interaction of Arg233 with the P3 Ser
side chain hydroxyl of Ac-ESMD-cho is lost in one
binding site. Also, the P2 Met in one of the binding sites
of caspase-7 ⁄ ESMD complex is moved out relative to
the position of the P2 Met side chain in the caspase-
7 ⁄ DQMD complex. These changes at the S2 and S3
subsites reduce the overall inhibitory potential of
Ac-ESMD-cho compared to P4 Asp containing inhibi-
tors. The weakest inhibitor is Ac-WEHD-Cho, which
agrees well with the larger structural changes in S4 when
the bulky tryptophan residue is at the P4 position.
Moreover, unlike the inhibitors with P4 Asp, there are
no hydrogen bond interactions of caspase-7 with P4 Trp
in the caspase-7 ⁄ WEHD structure (Fig. 3).
Bound citrate at the allosteric site of caspase-7
A citrate molecule was found at the allosteric inhibi-
tory site at the dimer interface of caspase-7. The
compounds DICA and FICA can bind covalently to
Cys290 resulting in movement of Tyr233 and the cata-
lytic Cys186 away from the active conformation, which
prevented binding of tetrapeptide inhibitors at the
active site [36]. A similar allosteric site and inhibitory
mechanism were observed in inflammatory caspase-1
[37]. In five of our caspase-7 complexes, a citrate mole-
cule, presumably an artifact from the crystallization
solution, was observed at the allosteric site (Fig. 5A).
O
4
and O

6
of the citrate molecule form close O S
interactions with Cys290 from the two p10 subunits.
The citrate oxygens have ionic interactions with the
side chain of Arg187 from both p10 subunits. Tyr233,
the third important residue at the allosteric site, inter-
acts with O
1
of citrate and a water molecule in the two
subunits, respectively. However, the active conforma-
tion was observed for the catalytic Cys186 and the
loops L1 to L4 forming the substrate binding site.
Thus, the citrate ion binds and forms favorable inter-
actions within the allosteric site despite the occupation
of the active site by tetrapeptide inhibitors.
Putative exosite in caspase-7
In all six caspase-7 structures, extended difference den-
sity was observed at a surface pocket between the two
p20–p10 heterodimers (Fig. 5B,C). This surface pocket
is approximately 22 A
˚
distant from the allosteric site
at the dimer interface where citrate is bound, and on
the opposite side of the molecular surface. The differ-
ence density was fit by a five-residue peptide in
extended conformation with the sequence of Gln-Gly-
His-Gly-Glu. The identity of the residues was deduced
from the shape of the electron density and the poten-
tial interactions with caspase-7 residues. However, due
to the resolution limit of the structure and surface

binding, the sequence could not be identified without
ambiguity. The two glycine residues in the sequence
cannot be distinguished in the electron density from
larger amino acids with disordered side chains. The
peptide is presumed to be part of a bacterial protein
trapped from cell lysate during purification of caspase-7.
The pentapeptide buries approximately 270 A
˚
2
of
accessible surface area, mostly from the p10 subunit.
The central His is buried in a deep cavity, which is
equidistant at approximately 30 A
˚
from the two cata-
lytic cysteines (Fig. 5E). This deep cavity is formed by
the residues Glu257, Gln260, Glu298 and Tyr300 from
both p10 subunits (Fig. 5D), whereas Gln59 from both
p20 subunits flanks the surface of the cavity. The His
side chain interacts with the side chains of Glu298 of
one p10 subunit and Gln260 of the other p10 subunit.
It forms water-mediated interactions with Glu298 and
J. Agniswamy et al. Plasticity of caspase-7 specificity pockets
FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS 4759
Tyr300 of the first p10 subunit and with Glu257 from
both p10 subunits. The N-terminal amide of the penta-
peptide interacts with the side chain of Gln59 from the
p20 subunits. The numerous interactions observed for
the bound peptide suggest that this cavity may have a
functional role.

A similar cavity is also present in the structures of
inflammatory and initiator caspases. The size and
charge of the cavity varies among the caspases. In
caspase-3, substitution of Gln260 by His completely
closes the cavity, indicating either the absence of a
ligand binding site or a difference in the preferred
ligand. Exosites, which are binding sites distant from
the active site, often play an essential role in the sub-
strate recognition and processing by proteases [38].
Exosites have been proposed to explain the discrepan-
cies between in vivo protein cleavage sites and peptide
substrates preferred by in vitro studies of caspases
[16,17,39]. A role for an exosite on caspase-7 has
been proposed for the abundant nuclear enzyme
poly(ADP-ribose) polymerase (PARP), which is
cleaved at DEVD-213flG by caspase-3 and 7 resulting
in a form that cannot synthesize ADP-ribose poly-
mers in response to DNA damage [40]. Caspase-7
processes PARP modified with long branched poly
(ADP-ribose) chains much more efficiently than does
caspase-3, suggesting the presence of specific interac-
tions between poly(ADP-ribose) and caspase-7 [41].
The small peptide binding site identified in the cur-
rent structures is a putative exosite of caspase-7.
However, further studies will be needed to identify
the protein substrate for the exosite and possible
effect on caspase-7 activity.
Discussion
An increasing number of caspase substrates have been
shown to be cleaved at noncanonical sites, which chal-

lenges the specificity requirements suggested by in vitro
studies with short synthetic peptides. These six new
structures have demonstrated that the S2, S3 and S4
specificity subsites of executioner caspase-7 are more
flexible than anticipated. The enzyme accommodates
noncanonical tetrapeptides in a similar manner as the
Cys 290
Tyr 233
Tyr 233’
Arg 187’
Cys 290’
AB
DE
C
CIT
O
6
O
5
O
2
O
1
O
3
O
4
O
7
~ 30 Å

~ 30 Å
Ac
tive
site
Active
site
Arg 187
Tyr 300’
Glu 257 Glu 257’
Glu 298
Glu 298’
Gln 260
Gln 260’
Gln 59
Pentapeptide
Tyr 300
Ser 302’
Gln 59’
Surface potential >–<
Fig. 5. Allosteric site and putative exosite of caspase-7. (A) Interaction of citrate ion bound at the allosteric site of caspase-7 ⁄ DMQD.
Despite occupation of the allosteric site, the catalytic residues are in the active conformation. (B) 2F
o
-F
c
electron density map of pentapep-
tide. (C) Peptide bound at the putative exosite on the surface of caspase-7. The caspase-7 surface is colored according to the element type.
The central histidine of the pentapeptide is buried deep in the cavity. (D) The interactions of the pentapeptide in the cavity. The water mole-
cules are represented as red balls. (E) The molecular surface of the caspase-7 around the putative exosite colored according to the electro-
static potential. Blue depicts areas of positive electrostatic potential, red depicts areas of negative electrostatic potential and white
represents areas of neutral potential. The putative exosite is highly electronegative and equidistant from the two active sites.

Plasticity of caspase-7 specificity pockets J. Agniswamy et al.
4760 FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS
canonical DEVD without dramatic conformational
changes. The S2 subsite of caspases was suggested to
contribute to substrate recognition [17,42]. The aro-
matic S2 subsite of caspase-3 and 7 containing bulky
Tyr230, Trp232 and Phe282 was predicted to preferen-
tially accommodate small aliphatic P2 residues such as
Ala, Val or Thr. By contrast, the substitution of
Tyr230 by Val in inflammatory and initiator caspases
was suggested to account for the accommodation of
large P2 side chains. Furthermore, screening of peptide
inhibitors based on amino acid positional fitness score
predicted that bulky polar His was the least favorable
residue at P2 for caspase-7 [24]. However, our crystal
structures show that executioner caspases can accom-
modate both smaller and larger P2 residues with favor-
able interactions in the S2 subsite. The rotation of the
side chains of Tyr230, Trp232 and Phe282 can reduce
or expand the S2 subsite of caspase-7 to accommodate
various sizes of hydrophobic P2 residue. The aromatic
side chain of P2 His forms stacking interactions and a
hydrogen bond with Tyr230, which suggests that a
P2 Tyr is likely to bind in a similar orientation. Several
physiological substrates of executioner caspase have
His or Tyr at P2 (supplementary Table S2) supporting
the structural observation; a more detailed list of cas-
pase substrates is provided elsewhere [28,32]. In addi-
tion, inhibition data show that Ac-DMQD-Cho,
Ac-DQMD-Cho, Ac-DNLQ-Cho are as potent as the

canonical Ac-DEVD-Cho, indicating that the S2
pocket of caspase-7 can harbor longer P2 residues than
the predicted small b-branched aliphatic Val and Thr.
Apart from the P1 residue, Glu P3 was optimal in
all three groups of caspases using the combinatorial
tetrapeptide substrate library search [22]. The long Glu
side chain is considered necessary to form ionic inter-
actions with the conserved Arg233. However, our
results suggest that the main chain interactions
between the P3 residue and Arg233 are more impor-
tant for proper positioning of the P3 residue. All six
inhibitors, irrespective of the P3 side chain, have con-
served main chain interactions and are positioned simi-
larly in the S3 subsite of caspase-7. Other polar
residues (Ser, Asn and Gln) form favorable hydrogen
bond interactions with the guanidinium group of
Arg233. The kinetic studies show that the presence of
Gln, Asn or Met at P3 does not alter the inhibitory
potency of the substrate analogs. In fact, the N-termi-
nal processing sites in procaspase-7 and procaspase-3
have the sequences DSVD and ESMD, respectively,
which implies that P3 Ser is physiologically acceptable
by initiator caspases. In some cells, caspase-3 was
shown to remove the N-terminal peptide of caspase-7
before the activation by granzyme B [43]. Moreover,
both caspase-3 and 7 show autoprocessing, which con-
firms that P3 Ser is recognized by executioner caspases.
The S4 pocket exhibits significant variability in both
substrate specificity and inhibitor selectivity among the
three groups of caspases. Inflammatory caspases rea-

dily accommodate P4 Trp. By contrast, executioner
caspases showed a absolute specificity for Asp at the
P4 position in studies with fluorescent peptide sub-
strates [35]. A combinatorial tetrapeptide study showed
that executioner caspases prefer Asp at P4, whereas
initiator caspases accommodate Leu ⁄ Ile ⁄ Val. This
classification was challenged by the observation that
caspase-8 tolerates small hydrophobic Ile and acidic
Asp residues at the P4 position [25]. Our results con-
firm that P4 Trp can bind but results in unfavorable
structural distortions of the S4 pocket of caspase-7.
Importantly, we show that the S4 pocket of caspase-7
has polar and nonpolar regions, which bind polar resi-
dues or short-branched aliphatic side chains in a man-
ner similar to that of initiator caspase-8. Trp232 of
caspase-7 plays a dual role by interacting with hydro-
phobic residues at both P2 and P4. Interestingly,
several physiological substrates of executioner caspases
have short-branched aliphatic P4 residues (supplemen-
tary Table S2). For example, DCC (deleted in colorec-
tal cancer), a candidate tumor suppressor gene, is
cleaved by caspase-3 at the noncanonical LSVD
sequence with an aliphatic P4 residue, and the cleavage
product is proapoptotic [44]. P4 Glu is also observed
in several physiological substrates, in agreement with
our structure and kinetic data for the tetrapeptide
ESMD.
The plasticity of the specificity subsites of execu-
tioner caspases demonstrated here suggests that factors
other than the P4–P1 sequence contribute to substrate

specificity. Indeed, solvent exposed, partially ordered
regions of proteins with non-optimal sequences might
be processed by active caspases without the need of
high binding affinity. However, the large number of
substrates processed at noncanonical sites implies that
exosites may contribute to caspase recognition of their
substrates. For example, Bid, the pro-apoptotic Bcl2
family member, is cleaved by caspase-8 at an LQTD
motif in a flexible loop, but a second potential site
IGAD in the same loop is not processed. The second
site is certainly accessible because it is targeted by
granzyme B in the mitochondrial pathway of apoptosis
[45]. Thus, it is postulated that important exosite-medi-
ated interactions preferentially guide caspase to the
first site or conversely steer the caspase away from the
second site [17]. Similarly, the more efficient cleavage
of PARP by caspase-7 rather than caspase-3 also sug-
gests the existence of exosites [17]. In addition, PARP
J. Agniswamy et al. Plasticity of caspase-7 specificity pockets
FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS 4761
modified with long branched poly(ADP-ribose) chain
has a much higher affinity for caspase-7 compared to
caspase-3, implying specific interactions between cas-
pase-7 and ADP-ribose moiety. However, no exosite
has been identified in caspases so far. The symmetric
pentapeptide binding pocket identified in the present
study is equidistant from the two active sites and pos-
sibly serves as an exosite for caspase-7. The size and
charge of the central cavity differs between caspase-3
and 7. Furthermore, the two extended N-termini of the

large subunits flank the pocket in executioner caspases
and no role has been established for these short exten-
sions. Thus, it is tempting to speculate that the N-ter-
minal extensions of the procaspases might block the
potential exosite and their removal starts the process
of executioner caspase activation, as well as the expo-
sure of the exosite. Clearly, further studies are required
to identify the molecular basis for substrate recogni-
tion of caspases.
Experimental procedures
Expression and purification of caspase-7
The recombinant human caspase-7 was expressed in Escher-
ichia coli grown at 37 °C in 5 L of induction media
(20 gÆL
)1
tryptone, 10 gÆL
)1
yeast extract, 5 g Æ L
)1
NaCl,
0.4% glucose, 1 mm MgCl
2
, 0.1 mm CaCl
2
, pH 7.4) supple-
mented with 100 mgÆL
)1
of ampicillin. Expression of
recombinant protein was induced by adding isopropyl thio-
b-d-galactoside to an absorbance of 0.6 at 600 nm. Cells

were harvested after 4.5 h and suspended in 200 mL of
25 mm Tris ⁄ HCl, 5 mm imidazole, 25 mm NaCl, 0.1% tri-
ton X-100, 0.1 mgÆmL
)1
lysozyme, pH 7.5. The cell lysate
obtained by centrifugation was loaded on to a nickel affin-
ity column (HisTrap
TM
HP, Amersham, NJ, USA). The
caspase-7 was eluted from the column using a 20 mm to
1 m imidazole gradient. The sample was dialyzed against
50 mm Tris, 100 mm NaCl, 20 mm imidazole, 10 mm
dithiothreitol, pH 7.5 to remove excess imidazole. The sam-
ple was further purified by size exclusion chromatography
on a Superdex-75 column (Amersham) with 50 mm Tris,
100 mm NaCl, 10 mm dithiothreitol, pH 7.5 as buffer. The
purity of the resulting sample was assessed by SDS ⁄ PAGE.
Enzyme kinetic assays
Enzymatic activity of caspase-7 was determined using the
colorimetric caspase-3 ⁄ 7 substrate Ac-DEVD-pNA (Bio-
mol, Plymouth Meeting, PA, USA), where Ac is the acetyl
group and pNA is p-nitroanilide. Caspase-7 was preincu-
bated in assay buffer (50 mm Hepes, 100 mm NaCl, 0.1%
Chaps, 10% glycerol, 1 mm EDTA and 10 mm dithiothrei-
tol, pH 7.5) at room temperature for 5 min prior to the
addition of substrate at different concentrations. p-Nitro-
anilide released by the substrate hydrolysis was measured at
a wavelength of 405 nm using a Polarstar Optima micro-
plate reader (BMG Labtechnologies, Durham, NC, USA).
sigmaplot 9.0 (SPSS Inc., Chicago, IL, USA) was used to

obtain the K
m
and V
max
values by fitting reaction velocities
as described [46]. The catalytic constant k
cat
for Ac-DEVD-
pNA was determined using the equation k
cat
¼ V
max
⁄ [E],
where the enzyme concentration [E] was determined by
active site titration during K
i
determination, as described
below.
The peptide analogs Ac-DMQD-Cho, Ac-DQMD-Cho,
Ac-DNLD-Cho, Ac-IEPD-Cho, Ac-ESMD-Cho and Ac-
WEHD-Cho form a covalent bond between the aldehyde
group (Cho) and the mecapto (-SH) group of Cys186 in
caspase-7. The aldehyde inhibitors of caspases are classified
as reversible inhibitors. For the measurement of inhibition
constant K
i
, caspase-7 was preincubated with the peptide
analogs in assay buffer at room temperature for 15 min.
After the addition of substrate, the reaction velocity was
measured based on substrate hydrolysis. The inhibition

constants of each inhibitor were determined by a
dose-dependent curve described by K
i
¼ (IC
50
) 0.5[E]) ⁄
(1 + [S] ⁄ K
m
), where [E], [S] and IC
50
, respectively, corre-
spond to active enzyme concentration, substrate concentra-
tion and the inhibitor concentration needed for half
maximum enzyme activity [47].
Crystallization, X-ray data collection, structure
determination and analysis
Caspase-7 was incubated at room temperature with each of
the six inhibitors at a 1 : 20 molar ratio. Crystals of the cas-
pase-7 complexes were grown in hanging drops at room
temperature by mixing 1 lL of protein solution (6 mgÆmL
)1
of protein) and reservoir solution (12.6–14.5% poly(ethylene
glycol) 3350, 0.3 m diammonium hydrogen citrate, 10 mm
dithiothreitol). The crystals were frozen with cryoprotectant
of 18% poly(ethylene glycol) 3350, 0.3 m diammonium
hydrogen citrate and 21% glycerol. Diffraction data were
collected at 100 °K on beamline 22-ID (SER-CAT) at the
Advance Photon Source, Argonne National Laboratory
(Argonne, IL, USA). All data were integrated and scaled
with HKL2000 [48].

The crystal structures were solved by molecular replace-
ment with the published structure of caspase-7 (1F1J) as
the initial model using phaser [49]. The inhibitors were
fitted into unambiguous electron density. The models were
subjected to several rounds of refinement in cns [50] and
model building with o [51]. Solvent molecules were
inserted at stereochemically reasonable positions. The final
refined models have good protein geometry with no disal-
lowed / ⁄ w-values on the Ramachandran plots. Hydrogen
bond interactions were identified by distances of 2.6 –
3.5 A
˚
between hydrogen donor and acceptor atoms.
Plasticity of caspase-7 specificity pockets J. Agniswamy et al.
4762 FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS
Surface area calculations were made with contact and
areaimol [52]. The structures were superposed globally
on the canonical caspase-7 ⁄ DEVD complex using 460
topologically equivalent Calpha atoms with lsqman of the
Uppsala software factory. Molecular figures were prepared
with molscript, raster3d [53], and pymol (http://
www.pymol.org).
Acknowledgements
BF was supported in part by the Georgia State Uni-
versity Research Program Enhancement award. ITW is
a Georgia Cancer Coalition Distinguished Cancer
Scholar. This research was supported in part by the
Georgia State University Molecular Basis of Disease
Program, the Georgia Research Alliance, and the
Georgia Cancer Coalition. We thank the staff at the

SER-CAT beamline at the Advanced Photon Source,
Argonne National Laboratory, for assistance during
X-ray data collection. Use of the Advanced Photon
Source was supported by the US Department of
Energy, Office of Science, Office of Basic Energy Sci-
ences, under Contract No. DE-AC02-06CH11357.
References
1 Tacconi S, Perri R, Balestrieri E, Grelli S, Bernardini S,
Annichiarico R, Mastino A, Caltagirone C, Macchi B
(2004) Increased caspase activation in peripheral blood
mononuclear cells of patients with Alzheimer’s disease.
Exp Neurol 190, 254 – 262.
2 Hartmann A, Troadec JD, Hunot S, Kikly K, Faucheux
BA, Mouatt-Prigent A, Ruberg M, Agid Y, Hirsch EC
(2001) Caspase-8 is an effector in apoptotic death of
dopaminergic neurons in Parkinson’s disease, but path-
way inhibition results in neuronal necrosis. J Neurosci
21, 2247–2255.
3 Hermel E, Gafni J, Propp SS, Leavitt BR, Wellington
CL, Young JE, Hackam AS, Logvinova AV, Peel AL,
Chen SF et al. (2004) Specific caspase interactions and
amplification are involved in selective neuronal vulnera-
bility in Huntington’s disease. Cell Death Differ 11,
424–438.
4 Vucic D, Stennicke HR, Pisabarro MT, Salvesen GS &
Dixit VM (2000) ML-IAP, a novel inhibitor of apopto-
sis that is preferentially expressed in human melanomas.
Curr Biol 10, 1359–1366.
5 Volkmann X, Cornberg M, Wedemeyer H, Lehner F,
Manns MP, Schulze-Osthoff K, Bantel H (2006)

Caspase activation is required for antiviral treatment
response in chronic hepatitis C virus infection. Hepato-
logy 43, 1311–1316.
6 Lakhani SA, Masud A, Kuida K, Porter GA Jr, Booth
CJ, Mehal WZ, Inayat I, Flavell RA (2006) Caspases 3
and 7: key mediators of mitochondrial events of apopto-
sis. Science 311, 847–851.
7 Nguyen JT & Wells JA (2003) Direct activation of the
apoptosis machinery as a mechanism to target cancer
cells. Proc Natl Acad Sci USA 100, 7533–7538.
8 Jia LT, Zhang LH, Yu CJ, Zhao J, Xu YM, Gui JH,
Jin M, Ji ZL, Wen WH, Wang CJ, Chen SY, Yang AG
(2003) Specific tumoricidal activity of a secreted
proapoptotic protein consisting of HER2 antibody
and constitutively active caspase-3. Cancer Res 63,
3257–3262.
9 Endres M, Namura S, Shimizu-Sasamata M, Waeber C,
Zhang L, Gomez-Isla T, Hyman BT, Moskowitz MA
(1998) Attenuation of delayed neuronal death after mild
focal ischemia in mice by inhibition of the caspase
family. J Cereb Blood Flow Metab 18, 238–247.
10 Wiessner C, Sauer D, Alaimo D & Allegrini PR, (2000)
Protective effect of a caspase inhibitor in models for
cerebral ischemia in vitro and in vivo. Cell Mol Biol
(Noisy-le-Grand) 46, 53–62.
11 Rabuffetti M, Sciorati C, Tarozzo G, Clementi E, Manf-
redi AA & Beltramo M (2000) Inhibition of caspase-1-
like activity by Ac-Tyr-Val-Ala-Asp-chloromethyl ketone
induces long-lasting neuroprotection in cerebral ischemia
through apoptosis reduction and decrease of proinflam-

matory cytokines. J Neurosci 20, 4398–4404.
12 Hoglen NC, Chen LS, Fisher CD, Hirakawa BP,
Groessl T & Contreras PC (2004) Characterization of
IDN-6556 (3-[2-(2-tert-butyl-phenylaminooxalyl)-
amino]-propionylamino]-4-oxo-5-(2,3,5,6-tetrafluoro-
phenoxy)-pentanoic acid): a liver-targeted caspase
inhibitor. J Pharmacol Exp Ther 309, 634–640.
13 Leung-Toung R, Li W, Tam TF & Karimian K (2002)
Thiol-dependent enzymes and their inhibitors: a review.
Curr Med Chem 9, 979–1002.
14 Denault JB & Salvesen GS (2002) Caspases: keys in the
ignition of cell death. Chem Rev 102, 4489–4500.
15 Nicholson DW, (1999) Caspase structure, proteolytic
substrates, and function during apoptotic cell death.
Cell Death Differ 6, 1028–1042.
16 Thornberry NA & Lazebnik Y (1998) Caspases: enemies
within. Science 281, 1312–1316.
17 Fuentes-Prior P & Salvesen GS, (2004) The protein
structures that shape caspase activity, specificity, activa-
tion and inhibition. Biochem J 384, 201–232.
18 Degterev A, Boyce M & Yuan J (2003) A decade of
caspases. Oncogene 22, 8543–8567.
19 Lamkanfi M, Kalai M & Vandenabeele P (2004)
Caspase-12: an overview. Cell Death Differ 11, 365–368.
20 Martinon F, Holler N, Richard C & Tschopp J (2000)
Activation of a pro-apoptotic amplification loop
through inhibition of NF-kappaB-dependent survival
signals by caspase-mediated inactivation of RIP. FEBS
Lett 468, 134–136.
J. Agniswamy et al. Plasticity of caspase-7 specificity pockets

FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS 4763
21 Lavrik IN, Golks A & Krammer PH (2005) Caspases:
pharmacological manipulation of cell death. J Clin
Invest 115, 2665–2672.
22 Thornberry NA, Rano TA, Peterson EP, Rasper DM,
Timkey T, Garcia-Calvo M, Houtzager VM, Nordstrom
PA, Roy S, Vaillancourt JP et al. (1997) A combinato-
rial approach defines specificities of members of the cas-
pase family and granzyme B. Functional relationships
established for key mediators of apoptosis. J Biol Chem
272, 17907–17911.
23 Lien S, Pastor R, Sutherlin D & Lowman HB (2004) A
substrate-phage approach for investigating caspase spe-
cificity. Protein J 23, 413–425.
24 Yoshimori A, Takasawa R & Tanuma S (2004) A novel
method for evaluation and screening of caspase inhibi-
tory peptides by the amino acid positional fitness score.
BMC Pharmacol 4,7.
25 Blanchard H, Donepudi M, Tschopp M, Kodandapani
L, Wu JC & Grutter MG (2000) Caspase-8 specificity
probed at subsite S(4): crystal structure of the caspase-
8-Z-DEVD-cho complex. J Mol Biol 302, 9–16.
26 Fang B, Boross PI, Tozser J & Weber IT (2006) Struc-
tural and kinetic analysis of caspase-3 reveals role for s5
binding site in substrate recognition. J Mol Biol 360,
654 – 666.
27 Schweizer A, Briand C & Grutter MG, (2003) Crystal
structure of caspase-2, apical initiator of the intrinsic
apoptotic pathway. J Biol Chem 278, 42441– 42447.
28 Fischer U, Janicke RU & Schulze-Osthoff K (2003)

Many cuts to ruin: a comprehensive update of caspase
substrates. Cell Death Differ 10, 76 – 100.
29 Wei Y, Fox T, Chambers SP, Sintchak J, Coll JT, Golec
JM, Swenson L, Wilson KP, Charifson PS (2000) The
structures of caspases-1-3-7 and -8 reveal the basis for
substrate and inhibitor selectivity. Chem Biol 7, 423–432.
30 Bump NJ, Hackett M, Hugunin M, Seshagiri S, Brady
K, Chen P, Ferenz C, Franklin S, Ghayur T, Li P et al.
(1995) Inhibition of ICE family proteases by baculovirus
antiapoptotic protein p35. Science 269, 1885–1888.
31 Emoto Y, Manome Y, Meinhardt G, Kisaki H, Khar-
banda S, Robertson M, Ghayur T, Wong WW, Kamen
R, Weichselbaum R et al. (1995) Proteolytic activation
of protein kinase C delta by an ICE-like protease in
apoptotic cells. EMBO J 14, 6148–6156.
32 Earnshaw WC, Martins LM & Kaufmann SH (1999)
Mammalian caspases: structure, activation, substrates,
and functions during apoptosis. Annu Rev Biochem 68,
383–424.
33 Ni CZ, Li C, Wu JC, Spada AP & Ely KR (2003) Con-
formational restrictions in the active site of unliganded
human caspase-3. J Mol Recognit 16, 121–124.
34 Riedl SJ, Renatus M, Schwarzenbacher R, Zhou Q, Sun
C, Fesik SW, Liddington RC, Salvesen GS (2001) Struc-
tural basis for the inhibition of caspase-3 by XIAP. Cell
104, 791–800.
35 Stennicke HR, Renatus M, Meldal M & Salvesen GS
(2000) Internally quenched fluorescent peptide substrates
disclose the subsite preferences of human caspases 1, 3,
6, 7 and 8. Biochem J 350, 563–568.

36 Hardy JA, Lam J, Nguyen JT, O’Brien T & Wells JA
(2004) Discovery of an allosteric site in the caspases.
Proc Natl Acad Sci USA 101, 12461–12466.
37 Scheer JM, Romanowski MJ & Wells JA (2006) A com-
mon allosteric site and mechanism in caspases. Proc
Natl Acad Sci USA 103, 7595–7600.
38 Lopez-Otin C & Overall CM (2002) Protease degrado-
mics: a new challenge for proteomics.
Nat Rev Mol Cell
Biol 3, 509–519.
39 Stennicke HR & Salvesen GS (2000) Caspases ) con-
trolling intracellular signals by protease zymogen activa-
tion. Biochim Biophys Acta 1477, 299–306.
40 Lazebnik YA, Kaufmann SH, Desnoyers S, Poirier GG
& Earnshaw WC (1994) Cleavage of poly(ADP-ribose)
polymerase by a proteinase with properties like ICE.
Nature 371, 346–347.
41 Germain M, Affar EB, D’Amours D, Dixit VM, Salve-
sen GS & Poirier GG (1999) Cleavage of automodified
poly (ADP-ribose) polymerase during apoptosis. Evi-
dence for involvement of caspase-7. J Biol Chem 274,
28379–28384.
42 Chereau D, Kodandapani L, Tomaselli KJ, Spada AP
& Wu JC (2003) Structural and functional analysis of
caspase active sites. Biochemistry 42, 4151 – 4160.
43 Denault JB & Salvesen GS (2003) Human caspase-7
activity and regulation by its N-terminal peptide. J Biol
Chem 278, 34042–34050.
44 Mehlen P, Rabizadeh S, Snipas SJ, Assa-Munt N,
Salvesen GS & Bredesen DE (1998) The DCC gene

product induces apoptosis by a mechanism requiring
receptor proteolysis. Nature 395, 801–804.
45 Barry M, Heibein JA, Pinkoski MJ, Lee SF,
Moyer RW, Green DR, Bleackley RC (2000)
Granzyme B short-circuits the need for caspase 8
activity during granule-mediated cytotoxic T-lympho-
cyte killing by directly cleaving Bid. Mol Cell Biol 20,
3781–3794.
46 Howard AD, Kostura MJ, Thornberry N, Ding GJ,
Limjuco G, Weidner J, Salley JP, Hogquist KA, Chap-
lin DD, Mumford RA et al. (1991) IL-1-converting
enzyme requires aspartic acid residues for processing of
the IL-1 beta precursor at two distinct sites and does
not cleave 31-kDa IL-1 alpha. J Immunol 147, 2964 –
2969.
47 Maibaum J & Rich DH (1988) Inhibition of porcine
pepsin by two substrate analogues containing statine.
The effect of histidine at the P2 subsite on the inhibition
of aspartic proteinases. J Med Chem 31, 625–629.
48 Otwinowski Z & Minor W (1997) Processing of X-ray
diffraction data collected in oscillation mode. Methods
Enzymol 276, 307–326.
Plasticity of caspase-7 specificity pockets J. Agniswamy et al.
4764 FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS
49 McCoy AJ, Grosse-Kunstleve RW, Storoni LC & Read
RJ (2005) Likelihood-enhanced fast translation func-
tions. Acta Crystallogr D Biol Crystallogr 61, 458–464.
50 Brunger AT, Adams PD, Clore GM, DeLano WL, Gros
P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J, Nilges
M, Pannu NS et al. (1998) Crystallography and NMR

system: a new software suite for macromolecular struc-
ture determination. Acta Crystallogr D Biol Crystallogr
54, 905–921.
51 Jones TA, Zou JY, Cowan SW & Kjeldgaard M (1991)
Improved methods for building protein models in elec-
tron density maps and the location of errors in these
models. Acta Crystallogr A 47, 110 – 119.
52 Collaborative Computational Project, Number 4 (1994)
The CCP4 suite: programs for protein crystallography.
Acta Crystallogr D Biol Crystallogr 50, 760 – 763.
53 Merritt EA & Murphy ME, (1994) Raster3d, Version
2.0. A program for photorealistic molecular graphics.
Acta Crystallogr D Biol Crystallogr 50, 869–873.
Supplementary material
The following supplementary material is available
online:
Fig. S1. S2 subsite of caspase-7.
Fig. S2. S3 subsite of caspase-7.
Fig. S3. S4 subsite of caspase-7.
Table S1. Polar interactions of caspase-7 with peptide
analogs.
Table S2. Examples of physiological caspase substrates
with studied substitutions at P2–P4.
This material is available as part of the online article
from
Please note: Blackwell Publishing is not responsible
for the content or functionality of any supplementary
materials supplied by the authors. Any queries (other
than missing material) should be directed to the corres-
ponding author for the article.

J. Agniswamy et al. Plasticity of caspase-7 specificity pockets
FEBS Journal 274 (2007) 4752–4765 ª 2007 The Authors Journal compilation ª 2007 FEBS 4765