Synthesis, enzymatic evaluation, and docking studies of fluorogenic caspase 8 tetrapeptide substrates

Article abstract of DOI:10.1002/cmdc.200900356

The synthesis, enzymatic evaluation, and molecular modeling studies of new fluorogenic tetrapeptide-based substrates selective for caspase 8, having the general structure Ac-IETD-AXX, are described. Various fluorescent reporter groups (AXX), i.e., 3- and 4-substituted coumarins and quinolin-2(1H)-ones were synthesized by von Pechmann condensation. They were subsequently coupled with the caspase-8-selective tetrapeptide Ac-IETD-OH under newly developed synthetic conditions to give the desired substrates in good yields and in high enantiomeric purity. Based on K_M and V_max values, the new compounds proved to be excellent substrates for recombinant human caspase 8. In contrast, the K_M values for the same compounds as substrates for human caspase 3 were approximately 10-20-fold higher. Molecular modeling studies based on the X-ray crystal structures of both human caspases 3 and 8 revealed that there is sufficient room within both active sites to accommodate substrates with moderately bulky substituents in the 3- and 4-positions of the fluorogenic coumarins and quinolin-2(1H)-ones. Automated docking of the substrates into the active sites of both human caspases 3 and 8 with the program Auto-Dock 3 gave structures similar to the published crystallographic structures for the same tetrapeptide bound to caspase 8 in the form of an irreversible inhibitor. The calculated binding energies for the new substrates to either caspase 3 or 8 showed little difference between the substrates, consistent with the K _M data. In addition, the calculated binding energies (ΔG) to caspase 8 were considerably more negative than those to caspase 3, also consistent with the K_M data. A possible molecular interaction that might explain the selectivity of the IETD tetrapeptide motif for caspase 8 over caspase 3 is discussed.

Full text of DOI:10.1002/cmdc.200900356

DOI: 10.1002/cmdc.200900356
Synthesis, Enzymatic Evaluation, and Docking Studies of
Fluorogenic Caspase 8 Tetrapeptide Substrates
Przemysław Reszka,[a] Riad Schulz,[a] Karen Methling,[a] Michael Lalk,[b] and
Patrick J. Bednarski*[a]
The synthesis, enzymatic evaluation, and molecular modeling
studies of new fluorogenic tetrapeptide-based substrates selec-
tive for caspase 8, having the general structure Ac-IETD–AXX,
are described. Various fluorescent reporter groups (AXX), i.e.,
3- and 4-substituted coumarins and quinolin-2(1H)-ones were
synthesized by von Pechmann condensation. They were subse-
quently coupled with the caspase-8-selective tetrapeptide Ac-
IETD-OH under newly developed synthetic conditions to give
the desired substrates in good yields and in high enantiomeric
purity. Based on KM and Vmax values, the new compounds
proved to be excellent substrates for recombinant human cas-
pase 8. In contrast, the KM values for the same compounds as
substrates for human caspase 3 were approximately 10–20-fold
higher. Molecular modeling studies based on the X-ray crystal
structures of both human caspases 3 and 8 revealed that there
is sufficient room within both active sites to accommodate
substrates with moderately bulky substituents in the 3- and 4-
positions of the fluorogenic coumarins and quinolin-2(1H)-
ones. Automated docking of the substrates into the active
sites of both human caspases 3 and 8 with the program Auto-
Dock 3 gave structures similar to the published crystallograph-
ic structures for the same tetrapeptide bound to caspase 8 in
the form of an irreversible inhibitor. The calculated binding en-
ergies for the new substrates to either caspase 3 or 8 showed
little difference between the substrates, consistent with the KM
data. In addition, the calculated binding energies (DG) to cas-
pase 8 were considerably more negative than those to cas-
pase 3, also consistent with the KM data. A possible molecular
interaction that might explain the selectivity of the IETD tetra-
peptide motif for caspase 8 over caspase 3 is discussed.
Introduction
Caspases (cysteine-dependent aspartate-specific proteases)
belong to the group of cysteine proteases. About a dozen
human members of these enzymes are known, which play es-
sential roles in apoptotic cell death[1] and inflammation.[2] Fur-
thermore, it is well recognized that many diseases, such as
cancer and autoimmune disorders, are associated with de-
creased caspase activity, resulting in enhanced cell survival.
Conversely, extensive or inappropriate caspase activity can
lead to accelerated cell death (e.g., neurodegenerative diseases
and ischemic tissue injury).[3]
through an aspartate residue to the reporter amine, such as 7-
amino-4-methylcoumarin (AMC), 7-amino-4-trifluoromethylcou-
marin (AFC), para-nitroaniline (pNA), or aminoluciferin. Proteo-
lytic cleavage of the substrate by a caspase leads to the libera-
tion of the reporter amine and an increase in fluorescence
(AMC and AFC), absorbance (pNA), or luminescence (aminolu-
ciferin). Microtiter plate assays based on fluorescence, absorb-
ance, and luminescence detection have become routine meth-
ods for measuring the activities of various caspases in apoptot-
ic cells.[8]
Caspases are synthesized as inactive pro-enzymes that are
converted during apoptosis into active proteases by autoacti-
vation or by amplification cascade through a highly regulated
process.[4] In such a caspase cascade, initiator caspases (such as
caspase 8 or 9) activate so-called effector caspases (such as
caspase 3 or 6), which are responsible for the cleavage of vari-
ous cellular proteins, leading to apoptotic death.[5]
Until now, little attention has been directed at the influence
of the structure of the reporter groups on the kinetics parame-
ters of a particular substrate. However, it has been observed
that the substitution of the same tetrapeptide with different
leaving groups can lead to substantial differences in KM values
[a] Dr. P. Reszka, R. Schulz, Dr. K. Methling, Prof. Dr. P. J. Bednarski
Department of Pharmaceutical and Medicinal Chemistry
Institute of Pharmacy, University of Greifswald
F.-L.-Jahn Strasse 17, 17487 Greifswald (Germany)
Fax: (+49)3834-864874
Caspases make use of a polarized cysteine side chain at the
catalytic site to cleave the peptide bond immediately following
an aspartate residue in the substrate. These proteases selec-
tively recognize a tetrapeptide sequence within the target
peptides.[6] Based on these observations, the specific tetrapep-
tide substrates and inhibitors for each caspase have been de-
signed and synthesized.[6,7] Synthetic tetrapeptide substrates
and inhibitors have found wide use as probes for caspase ac-
tivity in cells undergoing apoptosis. The commercially available
substrates contain a specific tetrapeptide motif attached
E-mail: bednarsk@pharmazie.uni-greifswald.de
[b] Dr. M. Lalk
Department of Pharmaceutical Biology
Institute of Pharmacy, University of Greifswald
F.-L.-Jahn Strasse 15, 17487 Greifswald (Germany)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.200900356: CD spectra of final products
6a–f and 7a–e.
ChemMedChem 2010, 5, 103 – 117
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
103

P. J. Bednarski et al.
MED
when tested against the same caspase. For example, the KM
value for Ac-IETD–AMC was found to be 10-fold lower than
that for the corresponding Ac-IETD–pNA on caspase 8 (6 mm
and 60 mm respectively).[9]
tuted coumarins 2a–2e. Deprotection of compounds 2a–2e
under acidic conditions gave the final products 3a–3e in mod-
erate overall yields [AMC (compound 3a) was synthesized as
well].
In our efforts to develop a new and highly sensitive analyti-
cal method that will allow the simultaneous detection of multi-
ple caspase activities (multiplex assay), we report herein the
design and synthesis of a series of new caspase 8 substrates
Ac-IETD–AXX, for which AXX is a fluorogenic coumarin or qui-
nolin-2(1H)-one moiety. The enzymatic constant KM and Vmax
values were determined for the new substrates with human re-
combinant caspases 3 and 8, and compared with the enzymat-
ic parameters obtained with the caspase 3 and 9 substrates
Ac-DMQD–AMC and Ac-LEHD–AMC, respectively. Molecular
modeling studies were carried out to help rationalize the selec-
tivity of the various substrates for caspase 8 over caspase 3.
Synthesis of the quinolin-2(1H)-ones (carbostyrils) proceeded
analogously to that of the coumarins. However, protection of
the amino group was not required, and the desired products
were obtained directly by condensation of 1,3-phenylenedia-
mine with the appropriate b-keto esters (Scheme 1). Com-
pounds 4a–4d were obtained by simple heating of the sub-
strates without solvent and catalyst.[11–14] This procedure, how-
ever, did not furnish the desired products if ethyl 4,4,4-trifluor-
oacetoacetate or ethyl 4-chloroacetoacetate were used (esters
a and b, respectively, in Scheme 1). The reaction of either
esters a or b with 1,3-phenylenediamine led to a heterogene-
ous mixture and resulted in hazardous bumping of the reactor.
A change in reaction conditions gave the desired product, but
only in the case of ester a. The trifluoromethyl derivative 4e
was obtained by heating 1,3-phenylenediamine with ester a
and zinc chloride in dimethyl sulfoxide at 1508C.[15] Further ef-
forts to synthesize the chloromethyl derivative from ester b
were not performed.
Results
Chemistry
The synthetic routes to prepare the fluorescent markers 3a–3e
and 4a–4e are depicted in Scheme 1. Synthesis of the coumar-
in derivatives required modification of the structure at the 4-
and 3-positions of the ring. The coumarin derivatives were pre-
pared by a modification of the von Pechmann procedure.[10,11]
3-Carbethoxyaminophenol 1 was subjected to reaction with a
series of b-keto esters in sulfuric acid to afford 4- and 3-substi-
Compounds 3d and 4d are new structures, whereas the
others have already been described.[12,13,16–18] Nevertheless,
some of these compounds are not very well characterized; for
example, little is known about the fluorescence properties of
phenyl and propyl derivatives such as 3b, 3c, and 4c. Thus,
the influence of substituents on the fluorescence properties
(maxima and relative intensities) of the synthesized com-
pounds was determined, and the results are summarized in
Table 1. The absorption spectra revealed that most derivatives,
Table 1. Fluorescence properties of the newly synthesized compounds.
[a]
Compd
Maxima [nm]
Frel
lex
lem
AMC[b]
3b
3c
3d
354
356
365
354
356
385
348
348
356
344
361
434
434
491
452
447
486
409
408
462
424
450
1.00
0.99
0.05
0.60
0.02
0.50
0.92
0.91
0.21
0.70
0.48
3e
AFC[c]
4a[d]
4b[e]
4c[f]
4d
4 f[g]
[a] Fluorescence intensities (at the optimal lem) are relative to the intensi-
ty of AMC on a equimolar basis. [b] Lit. lex/lem =345/445 nm.[16] [c] Lit. lex/
lem =365/495 nm,[19] 365/490 nm.[11] [d] Lit. lex/lem =344/418 nm,[11] 348/
381 nm.[20] [e] Lit. lex/lem =348/382 nm.[20] [f] Lit. lex/lem =402/410 nm.[13]
[g] Lit. lex/lem =360/470 nm,[12] 364/429 nm.[20]
with the exception of AFC, absorb in a similar region: l 345–
365 nm. However, coumarins absorb at longer wavelengths
than the corresponding quinolin-2(1H)-ones. The emission
spectra are much more characteristic, ranging from l 408 to
491 nm. The reasons for this are the different Stokes shifts,
Scheme 1. Synthesis of fluorescent markers. Reagents and conditions:
a) 70% H2SO4, room temperature, 4 h; b) concd H2SO4/CH3COOH (1:1), reflux,
2 h; c) solvent-free conditions, reflux, 24–64 h; d) DMSO, ZnCl2, 1508C, 24 h.
104
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 103 – 117

Fluorogenic Caspase 8 Tetrapeptide Substrates
which, once again, are much stronger for coumarins than for
quinolin-2(1H)-ones.
[4,5-b]pyridin-1-yl]methylene}-N-methylmethanaminium hexa-
fluorophosphate N-oxide (HATU),[22] a standard coupling re-
agent in peptide synthesis,[23–25] in the presence of the catalytic
base 2,4,6-trimethylpyridine (TMP)[26] in DMF. However, the re-
action conditions could be optimized by using a new, more
hindered base, 2,4,6-tri-tert-butylpyridine (TBP), instead of TMP
in the less polar solvent mixture of dichloromethane/DMF (in-
stead of DMF) and in the presence of an excess of the less pre-
cious fluorescent amine (instead of excess tetrapeptide).[22] The
new caspase 8 substrates 6a–6 f and 7a–7e were obtained in
satisfactory yields and with negligible quantities of undesired
d-Asp-containing epimer (<0.5% in most cases by HPLC). All
substrates gave very characteristic negative signals in the CD
spectra (see Supporting Information), confirming the expected
all-l configuration, whereas the d-Asp-containing substrates
give positive signals.[22]
The influence of the substituents on the fluorescence spec-
tra may be described with an intramolecular charge-transfer
transition (push–pull effect).[19–21] Owing to this phenomenon,
electronegative substituents at positions 3 and 4, such as CF3,
CN, F, and Ph, cause a red shift in fluorescence, whereas elec-
tropositive substituents such as CH3, OCH3, and C3H7 cause a
blue shift in fluorescence. The results obtained in our study are
in accordance with these observations.
New 3-substituted derivatives 3d and 4d proved to be
good fluorophores and had Stokes shifts about 15 nm higher
than the corresponding unsubstituted AMC and 4a. However,
they retained only about 60–70% of the fluorescence intensity
of the parent 3-unsubstituted compounds. None of the synthe-
sized compounds exceeded AMC in terms of fluorescence in-
tensity. Nevertheless, propyl derivative 3b proved to have fluo-
rescence properties nearly identical to those of AMC and was
found to be the most promising reporter group for a caspase
substrate for our new HPLC assay.
Enzymology
To determine the quality and selectivity of the newly synthe-
sized caspase substrates, the kinetics parameters KM, Vmax, and
the relative Vmax/KM values of human recombinant caspases 3
and 8 were determined (Table 2). The results are compared
with the standard caspase substrates (AMC conjugates such as
Ac-IETD–AMC for caspase 8), which were tested simultaneously
under the same experimental conditions (rel.; thus the Vmax/KM
value can be treated as a relative measure of the kcat/KM value).
From the data in Table 2 it is clear that for the IETD-based
substrates 6a–7e there are no great differences between the
KM values for caspase 8; the KM values for IETD derivatives
differ only by about twofold, and are in the range of 11–25 mm.
The KM value of 7c, with a bulky phenyl group at the 4-posi-
tion, was found to be approximately twice that of the methyl
derivate 7d, while the Vmax values are nearly the same
(Table 2). The differences in the rel. Vmax/KM values are slightly
more pronounced for IETD derivatives (about threefold). Thus,
the nature of the leaving group amine is not critical for good
substrate activity. The standard 6a (Ac-IETD–AMC) was found
to be the best substrate for caspase 8 followed by 6b.
The next step was to couple the tert-butyl-protected tetra-
peptide Ac-IETD-OH 5 [the detailed solid-phase peptide syn-
thesis (SPPS) of the tetrapeptide is published elsewhere[22]] by
its free carboxylic acid with the amino group of the fluorescent
markers (Scheme 2). We recently observed an extensive epime-
rization at the C-terminal aspartate residue under the standard
reaction conditions with N-{[(dimethylamino)-1H-1,2,3-triazolo-
As expected, the caspase-3-selective substrate Ac-DMQD–
AMC[7] was resistant to catalysis by caspase 8 at concentrations
as high as 1600 mm. However, the caspase 9 substrate Ac-
LEHD–AMC[27] was as good a substrate for caspase 8 as Ac-
IETD–AMC.
The KM values of the IETD substrates were found to be gen-
erally 10-fold higher for caspase 3 than those for caspase 8, in-
dicating modest selectivity for caspase 8. The only substrate
not turned over by caspase 3 was the reported caspase 9 sub-
strate Ac-LEHD–AMC (no turnover found at 1600 mm). The re-
maining compounds showed very similar KM and rel. Vmax/KM
values, including the “caspase 3 substrate” Ac-DMQD–AMC.
By comparing the three reference substrates (the first three
AMC conjugates in Table 2), it is clear that there is substantial
overlap between their activities. Ac-LEHD–AMC, a supposed
caspase 9 substrate, is also a good substrate for caspase 8; in
fact it is as good as the caspase 8 substrate Ac-IETD–AMC. Like-
wise, Ac-IETD–AMC, a supposed caspase 8 substrate, is cleaved
Scheme 2. Preparation of caspase 8 substrates. Reagents and conditions:
a) HATU/TBP in DMF/CH2Cl2, 5 days; b) 95% TFA/H2O, 1 h.
ChemMedChem 2010, 5, 103 – 117
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
105

P. J. Bednarski et al.
MED
Table 2. Kinetics parameters for cleavage of caspase substrates.[a]
Substrate
Caspase 3
Caspase 8
KM [mm]
rel. Vmax
rel. Vmax/KM
KM [mm]
rel. Vmax
rel. Vmax/KM
Ac-LEHD–AMC
Ac-DMQD–AMC
Ac-IETD–AMC (6a)
>1600
–
1.00 (24.92ꢀ16.13)[b]
1.07ꢀ0.71
1.02ꢀ0.74
0.74ꢀ0.46
1.04ꢀ0.60
ND
–
18.01ꢀ5.01
>1600
1.28ꢀ0.13
1.08
–
308.0ꢀ109.4
238.8ꢀ62.92
252.5ꢀ85.43
333.2ꢀ126.4
164.2ꢀ36.9
ND[c]
286.1ꢀ149.1
199.6ꢀ71.3
203.4ꢀ67.7
311.7ꢀ178.3
223.1ꢀ74.6
1.00
1.01
1.03
0.66
1.08
ND
0.82
0.95
0.99
0.75
0.58
–
16.48ꢀ4.14[b]
17.92ꢀ4.63
15.47ꢀ1.05
14.85ꢀ1.71
59.84ꢀ11.48[d]
13.78ꢀ2.41
11.89ꢀ2.29
24.40ꢀ5.33
12.41ꢀ3.33
20.09ꢀ4.06
1.00 (30.53ꢀ12.24)[b]
0.93ꢀ0.14
0.53ꢀ0.08
0.70ꢀ0.07
ND
1.00
0.87
0.57
0.78
ND
0.69
0.86
0.29
0.51
0.34
6b
6d
6e
6 f
7a
7b
7c
7d
7e
0.88ꢀ0.62
0.68ꢀ0.39
0.72ꢀ0.50
0.81ꢀ0.57
0.39ꢀ0.11
0.55ꢀ0.04
0.61ꢀ0.07
0.42ꢀ0.08
0.38ꢀ0.04
0.41ꢀ0.04
[a] KM and Vmax values were calculated from the S/V against S plot (Hanes–Wolff plot) and are averages of 3–4 independent determinations; the kinetics of
all substrates with the exception of 6 f were determined by measuring the fluorescence of the leaving amine. [b] Absolute value for Vmax in pmolminꢁ1
[c] Not determined. [d] Determined by measuring absorption of the 7-aminocoumarin at l=405 nm.
.
with nearly equal efficiency by caspase 3 as Ac-DMQD–AMC, a
typical caspase 3 substrate. Seemingly only Ac-DMQD–AMC de-
serves to be termed a caspase-3-specific substrate, as it was
found to be a good substrate for caspase 3 but not for cas-
pase 8. These results raise questions regarding the selectivity
of the commercially available substrates, which are usually re-
ferred to by the vendor as “specific”. There are numerous re-
ports in the biological literature in which the conclusions are
based largely on results obtained with so-called “specific” cas-
pase substrates.[27–32]
both caspases 3 and 8. With the exception of the four peptide
bonds, all single bonds of the substrates were conformational-
ly flexible during docking. Due to the large number of rotata-
ble bonds present in each substrate, which leads to a great
deal of freedom in the docked structure and multiple binding
modes for the substrate, there was no clearly preferred confor-
mation for any substrate. To solve this problem we used the
known X-ray crystal structure of the Ac-IETD inhibitor in the
caspase 8[33] active site as a template. With this template we
defined criteria for acceptable docking solutions, that is, “pro-
ductive conformations” (see Experimental Section for criteria).
With the exception of Ac-DMQD–AMC and 7c docked to cas-
pase 8, all docked substrate structures fulfilled these criteria.
An acceptable docked structure for Ac-DMQD–AMC was not
found, which was not surprising, as this tetrapeptide motif is
not a good substrate for caspase 8. A detailed discussion for
substrate 7c follows below.
The differences in the KM values for the tetrapeptides as sub-
strates for caspases 3 and 8 may well be explained by different
binding energies to the active site of the respective enzyme.
Thus, we undertook molecular modeling studies in an effort to
rationalize the differences or lack of differences in the KM
values of the substrates toward caspases 3 and 8.
Figure 1 shows a representative result of the least-squares fit
of the Ac-IETD motif from the X-ray structure overlapped with
the same atoms of the substrate Ac-IETD–AMC (3) in the active
site of caspase 8. Table 3 lists all root-mean-squared (RMS)
values for the modeled Ac-IETD substrates for caspase 8 com-
pared with the X-ray crystal conformation of the Ac-IETD inhib-
itor bound to the same active site. With one exception, 6b,
these values ranged between 2 and 3 ꢁ, indicating a good
degree of overlap of the heavy atoms of the tetrapeptide.
By clustering the various docking solutions from the genetic
algorithm in AutoDock it was found that in up to 50% of all
cases the docked structure was flipped in the binding pocket
by 1808 relative to the X-ray crystal structure. In these binding
geometries, however, the substrates cannot be hydrolyzed;
they are “nonproductive conformations”. The conformation
with the lowest DG value from all 80 runs per substrate was
considered further. Additionally, the lowest-energy conforma-
tions that gave structures suitable for peptide hydrolysis were
selected for further study. The binding energies of the lowest-
energy structure and the lowest-energy productive structures
are listed in Table 3. In three out of nine cases for caspase 3,
Molecular modeling: docking
The X-ray crystal structure of caspase 8 (PDB code: 1QTN), ob-
tained at 1.20 ꢁ resolution,[33] is covalently modified with the
inhibitor Ac-IETD-CHO. The inhibitor resides in a cleft that bi-
sects the active site. The thiohemiacetal bond between the
active site Cys360 of caspase 8 and the IETD-based inhibitor
was broken and replaced with the appropriate reporter group
(AXX) to model the caspase substrate (Ac-IETD–AXX). This ap-
proach was applied because the structures of caspase 8 sub-
strates Ac-IETD–AXX obtained with conventional minimization
methods do not reflect the extended conformation of the tet-
rapeptide structure of the inhibitor found within the active site
of caspase 8.[33]
The structure of caspase 8 with the inhibitor removed was
minimized with the help of the program INSIGHT II. The Ac-
DEVD-based inhibitor in the X-ray crystal structure of caspase 3
(PDB code: 1RHK), obtained at 2.50 ꢁ resolution,[34] was also re-
moved, and the structure was minimized as with caspase 8 in
preparation for docking simulations. AutoDock 3 was used to
automatically dock the various substrates in the active sites of
106
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 103 – 117

Fluorogenic Caspase 8 Tetrapeptide Substrates
was impossible to calculate the Ks values for our substrates
binding to either of the two caspases.[36] Instead, we used the
experimentally determined KM values as an estimate of sub-
strate binding affinity. Thus, the binding energies calculated by
AutoDock cannot be directly converted into KM values. Howev-
er, others have also used KM and Vmax values as relative meas-
ures of substrate binding affinity in docking studies.[37–39] Like-
wise, we found that the calculated binding energies of the
most stable productive structures bound to either caspase 3 or
8 were consistent with the KM values for the respective sub-
strates. Importantly, the binding energies of the Ac-IETD sub-
strates were significantly more negative when docked to cas-
pase 8 than to caspase 3, consistent with the experimental KM
values (see Table 3).
With the exception of the substrate 7c, all tetrapeptide units
bound to the active site of both caspases in a similar extended
conformation that was deemed suitable for catalysis (Figure 2).
In general, the leaving group amine has only a modest influ-
Figure 1. Two views of the best least-squares fit of the Ac-IETD motif bound
to the active site of human caspase 8, showing the X-ray crystal structure
(gray) of the inhibitor (Ac-IETD-CHO) superimposed with the best-ranked
docking structure (black) of a substrate (Ac-IETD–AMC). Only the heavy
atoms and amide hydrogen atoms were used for the comparison and are
shown in the figure; the RMSD was 2.077 ꢁ.
the most stable productive structure was also the most stable
overall. For caspase 8, all of the most stable productive struc-
tures were higher in energy than the overall most stable
docked structures.
The binding energies calculated by AutoDock 3 are estimat-
ed by an empirical free energy function that was calibrated by
using a set of 30 structurally known protein–ligand complexes
with experimentally determined binding constants.[35] The re-
sidual standard error for predicting binding energies was re-
ported to be 2.177 kcalmolꢁ1. In our case, the exact enzyme
concentration used in the enzyme assay was not known and
also varied between batches of recombinant caspase, so that it
Figure 2. Low-energy conformations of substrates 6a, 6b, 6d, 6e, 7a, 7b,
7d, and 7e deemed to be in a productive orientation for catalysis by cas-
pase 8.
Table 3. Calculated changes in free energy of binding (DG) compared with KM values of various tetrapeptide substrates for caspases 3 and 8.
Compd
Caspase 3
Caspase 8
KM [mm]
DG [kcalmolꢁ1
]
KM [mm]
DG [kcalmolꢁ1
]
Productive Conformation[a]
Best of All
Productive Conformation[a]
Best of All
RMSDrel [ꢁ][b]
Ac-LEHD–AMC
Ac-DMQD–AMC
Ac-IETD–AMC (6a)
>1600
308.0
238.8
252.5
333.2
164.2
ND
286.1
199.6
203.4
311.7
223.1
+1.95
ꢁ5.12
ꢁ3.80
ꢁ3.44
ꢁ2.29
ꢁ3.39
ND
ꢁ2.51
ꢁ3.93
ꢁ3.17
ꢁ2.89
ꢁ2.91
ꢁ7.96
ꢁ7.67
ꢁ4.48
ꢁ6.74
ꢁ4.51
ꢁ5.30
ND
ꢁ4.19
ꢁ3.93
ꢁ3.17
ꢁ3.61
ꢁ2.91
17.92
>1600
16.48
17.92
15.47
14.85
59.84
13.78
11.89
24.40
12.41
20.09
ꢁ5.90[c]
ꢁ0.99[d]
ꢁ6.04
ꢁ5.46
ꢁ6.64
ꢁ6.18
ꢁ6.53
ꢁ7.40
ꢁ5.02
ꢁ3.27[d]
ꢁ5.83
ꢁ5.03
ꢁ7.64
ꢁ7.95
ꢁ7.04
ꢁ6.04
ꢁ7.63
ꢁ7.14
ꢁ8.12
ꢁ8.15
ꢁ7.87
ꢁ6.11
ꢁ7.53
ꢁ6.89
ND
ND
2.077
4.026
2.811
2.864
1.962
2.730
2.939
2.727
3.060
2.939
6b
6d
6e
6 f
7a
7b
7c
7d
7e
[a] A conformation is considered productive if the substrate–enzyme orientation can lead to catalysis (see Experimental Section for details. [b] RMSD of
heavy atoms of the productive conformation of the Ac-IETD tetrapeptide relative to the conformation of the same atoms of the irreversibly bound IETD in-
hibitor in the X-ray crystal structure of caspase 8 (PDB code: 1QTN). [c] Alternative binding mode needed for catalysis. [d] Not all conditions for a produc-
tive structure were reached (see text).
ChemMedChem 2010, 5, 103 – 117
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
107

P. J. Bednarski et al.
MED
ence on the binding energies, which is reflected again in the
experimental KM values (Table 3). However, there was no signifi-
cant correlation between the binding energy of the productive
conformation and the KM values.
A truly productive structure for 7c in caspase 8 could not be
identified because the phenyl ring prohibits the amide bond
from adopting the optimal orientation for catalysis (compare
Figure 3A and 3B). A reason could be the rigid binding site
Figure 4. Schematic representation of the hydrogen bonds between cas-
pase 8 and the Ac-IETD substrate 6a (in bold). The interaction with Arg258
is believed to be critical for the selectivity of IETD substrates for caspase 8.
The black arrow indicates the monodentate hydrogen bond that may be re-
sponsible for this selectivity.
predict relative trends in the binding affinities of other tetra-
peptide motifs with the same AMC leaving group. Thus, we in-
vestigated the binding properties of the Ac-LEHD[27] and Ac-
DMQD motifs that have been reported to be selective for cas-
pases 3 and 9, respectively. Figure 5 shows the best productive
structures for Ac-LEHD–AMC, Ac-DMQD–AMC, and Ac-IETD–
AMC bound to caspases 3 and 8, and Table 3 lists the calculat-
ed binding energies for these structures. In keeping with the
kinetics data, Ac-LEHD–AMC gave a very poor binding energy
(+1.95 kcalmolꢁ1) with caspase 3 (Figure 5A), for which it is
not a substrate (KM >1600 mm). On the other hand, this same
Figure 3. Phenyl-containing Ac-IETD–quinazolone substrate 7c bound to
A) caspase 3 and B) caspase 8. Substrate 7c shows a productive conforma-
tion bound to caspase 3, whereas the conformation in the active site of cas-
pase 8 is not optimal for catalysis because the peptide bond is opposite the
correct orientation.
modeled by AutoDock, which does not allow for conformation-
al flexibility of the amino acid side chains of the enzyme in the
active site. This limitation has been discussed by the authors of
AutoDock 3.[35] Hence, a useful evaluation of the docking re-
sults for substrate 7c is hardly possible.
substrate shows
a
comparably good binding energy
(ꢁ5.90 kcalmolꢁ1) toward caspase 8, again in keeping with it
being a good substrate for this enzyme (KM =18.01 mm). Curi-
ously, the productive structure of Ac-LEHD–AMC that gave the
best binding energy does not bind to the active site cleft of
caspase 8 in the usual extended conformation (e.g., as with
caspase 3 in Figure 5A), but rather kinks back on itself by pro-
jecting the imidazole ring of the histidine residue into the cleft
normally occupied by the last two amino acid residues (Fig-
ure 5D). The stability of this productive conformation was as-
sessed by molecular dynamics (MD) simulations, which showed
very little movement of the substrate within the active site of
caspase 8 over 20 ps (1 fs time-step) at 310 K, supporting the
idea that this conformation is stable and could lead to cataly-
sis.
With the exception of 7c, AutoDock successfully predicts
that IETD substrates (6a, 6b, 6e, 6 f, 7a, 7b, 7d, and 7e) are
selective for caspase 8 over caspase 3 (Table 3). The 10-fold
higher KM values for caspase 3 relative to caspase 8 are reflect-
ed in the lower binding energy for the productive conforma-
tion, ranging between 2 and 5 kcalmolꢁ1 for 6b and 7a, re-
spectively. Figure 4 shows the important hydrogen bonding in-
teractions between caspase 8 and substrate 6a, which is repre-
sentative of the other IETD substrates. As discussed in detail
below, the selectivity for caspase 8 over caspase 3 may be the
result of the additional interaction between the glutamic acid
of the substrate and Arg258 (Figure 4). The guanidinium
group of Arg258 is conspicuous in the structure of caspase 8
as the large protrusion that sticks out in the top-middle of the
active site cleft (Figure 3B). A corresponding arginine is not
present in caspase 3 (Figure 3A). With caspase 3, nearly the
same interactions between IETD substrates and enzyme take
place, with the exception of the missing interaction with the
third arginine (compare Figure 5C and 5F).
Figure 5B and 5E show the structures that were found to be
optimal for catalysis by caspases 3 and 8, respectively. Ac-
DMQD–AMC is a poor substrate for caspase 8 (KM >1600 mm)
and also a poor binder (DG=ꢁ0.99 kcalmolꢁ1). It is also a poor
substrate for human caspase 9, with a KM value of >1600 mm
(data not shown). In the case of caspase 3, for which Ac-
DMQD–AMC is a known substrate (we determined a KM value
of 308 mm; see Table 2), a good substrate binding affinity was
calculated (DG=ꢁ5.12 kcalmolꢁ1) by AutoDock. In addition,
After finding that AutoDock can explain the relative selectivi-
ty of IETD substrates, we wanted to know if the program could
108
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 103 – 117

Fluorogenic Caspase 8 Tetrapeptide Substrates
Figure 5. Best productive structures of Ac-LEHD–AMC (A,D), Ac-DMQD–AMC (B,E) and Ac-IETD–AMC (C,F) bound to caspase 3 (A,B,C) and caspase 8 (D,E,F).
The large structure protruding down from the middle of caspase 8 into the active site cleft is the guanidinium group of Arg258.
the tetrapeptide of our docked structure of Ac-DMQD–AMC
(Figure 5B) made the same contacts with the active site as did
the tetrapeptide sequence reported in the X-ray crystal struc-
ture of human caspase 9 bound with the irreversible inhibitor
Ac-DMQD-CHO (data not shown).[40]
carbonyl carbon of the amide bond. Strong salt bridges be-
tween guanidinium and carboxylate groups, as in the case of
methylguanidinium formate, have distances of ~2.9 ꢁ. In the
case of the caspases, a strong salt bridge forms between the
carboxylate of the aspartate residue and the guanidinium ni-
trogen of Arg260 (distance c in Figure 4), because the aspar-
tate side chain can bury itself into a cationic hole in the
enzyme (Figure 5F); as a result, most of the time relatively
little movement occurs, and all substrates achieve distances
<4 ꢁ at some time (Figure 6C). In contrast, the substrate glu-
tamate is directed out of the binding cleft and interacts less
strongly with the active site (Figure 5F). Figure 6A shows that
more movement occurs between Arg413 and the glutamate
carboxyl group of the substrates (distance a in Figure 4). All
substrates achieve a distance of 4 ꢁ or less part of the time,
whereas all substrates except 6e achieve distances just under
3 ꢁ at some time. In the case of the monodentate interactions
between Arg258 and the glutamate groups of the various
IETD substrates (distance b in Figure 4), all substrates at some
time achieve distances just under 3 ꢁ (Figure 6B), evidence
that supports an interaction between these two groups. Thus,
we conclude that the guanidinium group of Arg258 in cas-
pase 8 is capable of interacting with the glutamate group of
the IETD substrates, and we propose that this interaction is a
key reason for the selectivity of IETD substrates for caspase 8
over caspase 3.
Ac-IETD–AMC displayed an extended conformation in both
caspase 3 (Figure 5C) and caspase 8 (Figure 5F). However, in
the case of caspase 8, an additional point of attachment for
the glutamate residue is offered by the guanidinium group of
Arg258. Together with Arg413, which makes bidentate contact
with the carboxylate group of the substrate glutamate, the
guanidinium group of Arg258 can form a monodentate hydro-
gen bond with the same glutamate (see Figure 4). Because an
arginine residue is missing at the analogous location in cas-
pase 3 (Figure 5C), this additional monodentate hydrogen
bond may be a key factor that imparts the IETD motif with se-
lectivity for caspase 8.
Molecular modeling: molecular dynamics
To help understand the interaction between the guanidinium
group of Arg258 and the carboxylate group of the substrate
glutamate for the selectivity of IETD substrates for caspase 8,
we performed MD simulations with the various substrate–
enzyme complexes. To assess the importance of the interaction
between these groups, the distances were measured between
the guanidinium nitrogen atoms of Arg413 and Arg258, and
the carboxylate group of the substrate glutamate residue (dis-
tances a and b in Figure 4) during the 20 ps simulation.
Figure 6 shows that the distances between the guanidinium
and carboxylate groups of various salt bridges as well as the
distance between the catalytic cysteine sulfur atom and the
Finally, the distance between the catalytic sulfur atom and
the carbonyl carbon (distance d in Figure 4) varies between 4
and 6 ꢁ over most of the 20 ps for the various substrates. This
distance is not yet small enough for a reaction to occur, but in-
dicates good proximity of the catalytic unit to the cleavable
group.
ChemMedChem 2010, 5, 103 – 117
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
109

P. J. Bednarski et al.
MED
Figure 6. Distances [ꢁ] between the guanidinium nitrogen atom of A) Arg413 or B) Arg258 and a carboxylate oxygen atom of the glutamate residue in the
IETD substrate (distances a and b, respectively, in Figure 4) determined during an MD simulation with seven tetrapeptide substrates for 20 ps at 310 K. For
comparison, panel C) shows the distances between the guanidinium nitrogen atom of Arg260 and the carboxylate oxygen atom of the aspartate substrate
residues (distance c in Figure 4), and panel D) shows the distances between the cysteine sulfur atom and the carbonyl carbon of the amide bond (distance d
in Figure 4).
Discussion
the ester protecting groups under acidic conditions. This flexi-
ble synthetic strategy allows a variety of fluorogenic amines to
be coupled to the more precious tetrapeptide at the end of
the solid-phase peptide synthesis. This is in contrast to the tra-
ditional synthesis of fluorogenic peptides, which involves initial
coupling of the 7-aminocoumarin to the first amino acid (e.g.,
aspartic acid) followed by the solution-phase peptide synthe-
sis.[17] An alternative method for coupling pNA to a peptide at
the end of a solid-phase synthesis was described a few years
ago.[42]
Artificial tetrapeptide substrates that are more or less selective
for various caspases are routinely used in the evaluation of cel-
lular activation of caspase-dependent pathways.[8] A substrate
with either a chromogenic (typically pNA), fluorogenic (typi-
cally AMC or AFC), or chemiluminogenic leaving amine (typi-
cally aminoluciferin) is used. Fluorescent leaving amines have
an advantage of sensitivity over pNA and do not require fur-
ther enzymatic activation as is the case with aminoluciferin,
which requires activation by luciferase.[41] With the aim of de-
veloping new analytical methods for quantifying muliplex cas-
pase activity, we synthesized tetrapeptide substrates selective
for caspase 8 bearing new fluorogenic 7-aminocoumarins and
7-aminoquinolin-2(1H)-ones and evaluated the importance of
the leaving amine on the kinetics of hydrolysis as well as
enzyme selectivity.
The new tetrapeptides proved to be excellent caspase 8 sub-
strates, with hydrolysis kinetics similar to that of commercially
available Ac-IETD–AMC. The KM values were compared for two
different human recombinant caspases (3 and 8), revealing
modest selectivity toward caspase 8 for all IETD substrates. Se-
lectivity is known to be dependent on the IETD tetrapeptide.[7]
This work shows that selectivity for caspase 8 over caspase 3 is
not greatly influenced by the leaving amine. Nor does the
nature of the leaving amine greatly affect the kinetics parame-
ters of the IETD substrates with either caspase, indicating suffi-
cient room in their active sites to accommodate a variety of bi-
cyclic amines. The quinolin-2(1H)-one-based tetrapeptides had
somewhat lower KM values than the corresponding coumarin
derivatives as caspase 8 substrates, but the coumarin deriva-
tives had somewhat higher Vmax values.
The different fluorescent markers (AXX) based on coumarin
and quinolin-2(1H)-ones were synthesized by standard meth-
ods, and their fluorescent properties were characterized (lex,
lem, relative fluorescence in comparison with AMC). None
showed superiority in their photophysical properties to either
AMC or AFC. We prepared the protected tetrapeptide sub-
strate after the coupling of tert-butyl-protected Ac-IETD-OH via
its free carboxylic group with fluorescent markers (via their
free amino group) under optimized reaction conditions.[22] The
substrates were obtained in high optical purity after removing
Due to their overlapping substrate specificities, earlier re-
searchers grouped caspases 8 and 9 together with caspase 6,
110
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 103 – 117

Fluorogenic Caspase 8 Tetrapeptide Substrates
while caspase 3 was included in a separate group together
with caspases 2 and 7.[6] Consistent with this classification, we
found that the caspase 9 substrate Ac-LEHD–AMC[27] is also a
good substrate for caspase 8, but shows no turnover with cas-
pase 3. Similarly, the caspase 3 motif DMQD[40] is not recog-
nized by caspase 8. However, we found that the caspase 8
motif IETD is also recognized by caspase 3, albeit at 10-fold
higher concentrations than caspase 8, which is not consistent
with the classification by Thornberry et al.[6]
For cases in which small changes were made to the struc-
tures of the caspase substrates (i.e., substrates for which only
the leaving amines are varied), AutoDock also predicted that
the amine leaving group will have only a minor influence on
the KM value of the substrate. The substrate for which this was
not the case is 7c, where phenyl substituents in the leaving
amine result in productive structures with poor theoretical
binding affinities. However, the KM values of 7c are similar to
those of substrates with smaller substituents and good binding
affinities, most likely because the protein must adapt to the
more bulky phenyl side chain. Induced-fit adaptations have
not been considered in our docking attempts. We do not yet
know why this discrepancy occurred, but suspect that because
in our docking run with AutoDock, no flexibility of the protein
was allowed.
Molecular modeling studies were performed by starting with
the X-ray crystal structures of human caspases 3 and 8 in an at-
tempt to rationalize the similarities and differences in the KM
values of tetrapeptide-based substrates. The software program
AutoDock[35,43] was used to perform the automated docking of
tetrapeptide substrates into the active sites of the two proteas-
es. Docking methods such as AutoDock have found important
use in the virtual screening[44,45] and design[46] of enzyme inhib-
itors and substrates.
We were surprised to find the caspase 9 substrate Ac-LEHD–
AMC[27] to be a good substrate for caspase 8, with a KM value
similar to those of the substrates bearing the IETD motif for
caspase 8. A possible explanation comes from the modeling
studies with Ac-LEHD–AMC binding to caspase 8. We observed
an alternative binding mode for Ac-LEHD–AMC in the active
site of caspase 8, whereby the histidine imidazole side chain
takes the place of the last two dipeptides in the active site
cleft; the DG value of ꢁ5.90 kcalmolꢁ1, on the same order of
magnitude as those of the other examined substrates, indi-
cates a similar binding affinity. MD simulations show this alter-
native binding mode to be stable, suggesting that it may be a
productive conformation.
AutoDock has been used by others to analyze enzyme–sub-
strate interactions in order to rationalize kinetics data. One of
the first uses of AutoDock for predicting substrate binding was
the automated docking of monosaccharide substrates into the
active site of glucoamylase.[47] More recently, Reilly and co-
workers used AutoDock to study the specificity of various gly-
coside hydrolases for disaccharide[48] and cello-oligosacchar-
ide[49] substrates. Chica et al. used a combined kinetics/molecu-
lar modeling approach to generate a model for the binding of
small acyl donor dipeptide substrates to tissue transglutami-
nase.[38] Kanade and colleagues also used AutoDock to study
the structural features of diphenol substances to distinguish
between substrates and inhibitors of polyphenol oxidase.[39]
Substrate binding to the cytochrome P450 (CYP) 2D6 isoform
has also been studied by automated docking of known sub-
strates to the active site by using AutoDock followed by mo-
lecular dynamics.[50] The predicted binding energy was correlat-
ed with the observed binding energy calculated from Ks or KM
values. The specificity of CYP 3A4 in the oxidative metabolism
of indapamide was recently rationalized with the help of Auto-
Dock.[51] Thus, AutoDock has become useful in the study of
enzyme interactions with substrates in addition to the more
traditional uses of inhibitor search and design.
Conclusions
The new tetrapeptide substrates of general structure Ac-IETD–
AXX are excellent substrates for caspase 8, similar to the com-
mercially available Ac-IETD–AMC. However, selectivity for cas-
pase 8 over caspase 3 is only modest. Computer modeling of
the molecular interactions between the new substrates and
caspase 8 reveal that there is sufficient room within the cas-
pase 8 active site to accommodate substrates with substituents
at the 3- and 4-positions of the coumarin and quinolin-2(1H)-
one rings. AutoDock has helped to rationalize KM values of tet-
rapeptide substrates for two caspases, 3 and 8, by predicting
productive structures and binding affinities. We propose that
the reason for the selectivity of IETD substrates for caspase 8
over caspase 3 is the additional arginine residue (Arg258) in
the active site of caspase 8, which can hydrogen bond to the
carboxyl group of the substrate glutamate. AutoDock may be
useful for designing more potent and selective caspase sub-
strates in the future. The use of these new IETD substrates in
an assay for simultaneous determination of multiple caspase
activities (multiplex assay) will be published in due course.
The design of enzyme substrates is more challenging than
the design of inhibitors, however, not only because good bind-
ing but also the ability to undergo efficient catalysis must be
considered. Furthermore, our substrates are linear peptides
with many degrees of conformational freedom, adding greatly
to the number of possible docked structures. Nevertheless, our
results show good agreement between the KM values and
binding energies, calculated with AutoDock, if a conformation-
ally productive structure is considered instead of simply choos-
ing the structure with the overall lowest binding energy. Auto-
Dock predicted correctly the selectivity of the LEHD, DMQD,
and IETD tetrapeptide sequences for the two caspases. The se-
lectivity of IETD substrates for caspase 8 over caspase 3 ap-
pears to be the result of an additional interaction between the
carboxylate group of the substrate glutamate and the guanidi-
nium group of Arg258 of caspase 8.
Experimental Section
Chemistry
Reagents were obtained from commercial suppliers and used with-
out further purification. Ac-LEHD–AMC and Ac-DMQD–AMC were
ChemMedChem 2010, 5, 103 – 117
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
111

P. J. Bednarski et al.
MED
obtained from Bachem (Weil am Rhein, Germany). Ac-IETD-based
substrates were synthesized as described below. Recombinant
human caspases 3 and 8 were purchased from Alexis Biochemicals
7.1 Hz, 3H, CH3), 0.97 (t, J=7.4 Hz, 3H, 13-coum); Anal. (C15H17NO4)
C, H, N.
(4-Phenyl-2-oxo-2H-1-benzopyran-7-yl)carbamic acid ethyl ester
(2c): From ethyl benzoylacetate (2.30 g, 12 mmol). Yield: 1.22 g
(36%); mp: 167–168.58C; Rf =0.76 (EA), 0.76 (EA/toluene 1:1);
1H NMR (200 MHz, [D6]DMSO): d=10.18 (bs, 1H, NH), 7.65–7.36 (m,
8H, arom), 6.25 (s, 1H, 3-coum), 4.18 (m, 2H, CH2), 1.28 (m, 3H,
CH3); Anal. (C18H15NO4) C, H, N.
(Lçrrach, Germany) and had a reported activity of 3–4ꢂ105 Umgꢁ1
.
The proteases were stored in 5–10 mL aliquots at ꢁ808C according
to the supplier’s data and used only once after thawing.
Melting points were determined with a MEL-TEMP melting point
apparatus (Mel-Temp Laboratories Inc., USA), and are uncorrected.
Elemental analyses were performed on a PerkinElmer 2400 CHN el-
emental analyzer. Mass spectra were collected on a microTOF 112
instrument (ESI). UV/Vis spectra were recorded on a Spekol 1200
instrument in CH3OH. Fluorescence spectra were recorded on a
PerkinElmer LS 50B luminescence spectrometer. Fluorescent mark-
ers were analyzed at a concentration of 10ꢁ6 m in CH3OH. CD spec-
tra were recorded on JASCO-J 810 circular dichroism chiroptical
spectrometer. The compounds were analyzed at a concentration of
~0.1 gLꢁ1 in H2O (6a, 6b, 6e, 6 f) or CH3OH (6c, 6d, 7a–7e).
NMR spectra were recorded with a Bruker DPX 200 (1H NMR:
200 MHz; 13C NMR: 50 MHz); measurements were carried out in
[D6]DMSO. The chemical shifts (d) were related to tetramethylsilane
(TMS) as internal standard (eventually to DMSO signal), and are re-
ported in parts per million (ppm). Coupling constants (J) are re-
ported in hertz (Hz). The following abbreviations are used: s, sin-
glet; bs, broad singlet; d, doublet; dd, doublet of doublets; t, trip-
let; q, quartet; m, multiplet; FJ, coupling constant with 19F atom.
(3-Fluoro-4-methyl-2-oxo-2H-1-benzopyran-7-yl)carbamic
acid
ethyl ester (2d): From ethyl 2-fluoroacetoacetate (1.78 g,
12 mmol). Yield: 0.82 g (28%); mp: 247–2508C; Rf =0.73 (EA/tolu-
1
ene 1:1); H NMR (200 MHz, [D6]DMSO): d=10.12 (bs, 1H, NH), 7.69
(d, J=8.8 Hz, 1H, 6-coum), 7.55 (d, J=2.0 Hz, 1H, 9-coum), 7.44
(dd, J=8.8 and 2.0 Hz, 1H, 7-coum), 4.18 (q, J=7.1 Hz, 2H, CH2),
F
2.34 (d, 3H, J=2.8 Hz, 11-coum), 1.27 (t, J=7.1 Hz, 3H, CH3); Anal.
(C13H12FNO4) C, H, N.
[4-(Chloromethyl)-2-oxo-2H-1-benzopyran-7-yl]carbamic
acid
ethyl ester (2e): From ethyl 4-chloroacetoacetate (1.98 g,
12 mmol). Yield: 2.16 g (70%); mp: 246–2478C dec. (Lit. 2478C
1
dec.[52]); H NMR (200 MHz, [D6]DMSO): d=10.19 (bs, 1H, NH), 7.76
(d, J=8.8 Hz, 1H, 6-coum), 7.60 (d, J=2.0 Hz, 1H, 9-coum), 7.42 (m,
J=8.8 and 2.0 Hz, 1H, 7-coum), 6.52 (s, 1H, 3-coum), 4.98 (s, 2H,
CH2Cl), 4.18 (q, J=7.1 Hz, 2H, CH2), 1.27 (t, J=7.1 Hz, 3H, CH3);
Anal. (C13H12ClNO4) C, H, N.
General procedure for the synthesis of coumarin derivatives
3a–3e: Protected coumarin (3 mmol) was heated for 2 h at
~1258C in a mixture of 5 mL concentrated H2SO4 and 5 mL glacial
acetic acid. After cooling to room temperature, the solution was
poured into 50 mL ice water, and the resulting suspension was
made slightly basic with 1n NaOH. The precipitate was filtered and
crystallized from CH3OH to give yellowish crystals.
RP-HPLC was carried out with a Merck–Hitachi system consisting of
a D-7000 interface, L-7100 pump, L-7360 thermostat column
(308C), L-7612 solvent degasser, L-4500 diode array detector, and a
7125 Rheodyne sample injector fitted with a 20 mL injection loop
(250 mL for preparative use). Separations were done with a CC 250/
4 Nucleosil 120–5 C18 column or a CC 250/4 Hypersil 120–5 C18
column (Macherey–Nagel, Dꢃren, Germany). Mobile phases: A,
CH3OH; B, 20 mm KH2PO4/H2O (pH 3.3); C, CH3CN+0.1% TFA; D,
H2O+0.1% TFA. HPLC conditions with the Hypersil column (flow
rate: 1.0 mLminꢁ1): System 1, 45% A/B; HPLC conditions with the
Nucleosil column (flow rate: 0.7 mLminꢁ1): System 2, 30!53% C/D
over 20 min; System 3, 20!50% C/D over 22 min; System 4, 20!
47% C/D over 20 min; System 5, 20!54% C/D over 25 min.
7-Amino-4-methyl-2H-1-benzopyran-2-one (3a): UV/Vis, Fl. re-
corded for commercial AMC (Aldrich). From 2a (0.74 g, 3 mmol).
Yield 0.30 g (58%); mp: 221–2258C (Lit. 220–2248C[10]); Rf =0.46
(EA/toluene 1:1); 1H NMR (200 MHz, [D6]DMSO): d=7.42 (d, J=
8.6 Hz, 1H, 6-coum), 6.60 (dd, J=8.6, 2.0 Hz, 1H, 7-coum), 6.45 (d,
J=2.0 Hz, 1H, 9-coum), 5.92 (s, 1H, 3-coum), 5.81 (bs, 2H, NH2),
2.30 (s, 3H, 11-coum); UV/Vis (nm, CH3OH): lmax =208, 230, 353 (e=
16180 mꢁ1 cmꢁ1); HPLC, System 1: tR =4.73 min (identical to com-
mercial AMC), purity: 99.2% at l=230 nm.
(3-Hydroxyphenyl)carbamic acid ethyl ester (1): This compound
was prepared as described previously.[11]
General procedure for the synthesis of protected coumarins 2a–
2e: Compound 1 (2.0 g, 11 mmol) and b-keto ester (12 mmol)
were first suspended in 60 mL 70% H2SO4. The reaction was then
stirred while further portions of 70% H2SO4 (final volume: 100–
120 mL) were added to give a clear yellow to orange solution. The
reaction was maintained for 4 h at room temperature and poured
into 300 mL ice water, giving a precipitate. The solid was filtered
and crystallized from EtOH (2e), EtOH/H2O (2a, 2d) or CH3OH (2b,
2c) to give colorless to yellowish (2c) crystals.
7-Amino-4-propyl-2H-1-benzopyran-2-one (3b): From 2b (0.83 g,
3 mmol). Yield: 0.33 g (55%); mp: 206–208.58C; Rf =0.58 (EA/tolu-
ene 1:1); H NMR (200 MHz, [D6]DMSO): d=7.45 (d, J=8.8 Hz, 1H,
6-coum), 6.57 (m, J=8.8, 2.0 Hz, 1H, 7-coum), 6.41 (m, J=2.0 Hz,
1H, 9-coum), 6.10 (bs, 2H, NH2), 5.86 (s, 1H, 3-coum), 2.64 (t, J=
7.5 Hz, 2H, 11-coum), 1.59 (m, 2H, 12-coum), 0.96 (t, J=7.4 Hz, 3H,
13-coum); UV/Vis (nm, CH3OH): lmax =208, 232, 353 (e=
16300 mꢁ1 cmꢁ1); HPLC, System 1: tR =12.85 min, purity: 98.2% at
l=230 nm; Anal. (C12H13NO2) C, H, N.
1
(4-Methyl-2-oxo-2H-1-benzopyran-7-yl)carbamic acid ethyl ester
(2a): From ethyl acetoacetate (EA; 1.55 g, 12 mmol). Yield: 2.14 g
(79%); mp: 189–1938C (Lit. 186–1888C[11]); Rf =0.77 (EA), 0.62 (EA/
toluene 1:1).
7-Amino-4-phenyl-2H-1-benzopyran-2-one (3c): From 2c (0.93 g,
3 mmol). Yield: 0.37 g (53%); mp: 222–2238C; Rf =0.60 (EA), 0.63
(EA/toluene 1:1); 1H NMR (200 MHz, [D6]DMSO): d=7.62–7.54 (m,
5H, arom), 7.19 (d, J=8.2 Hz, 1H, 6-coum), 6.66 (d, J=2.4 Hz, 1H,
7-coum), 6.61 (s, 1H, 9-coum), 6.35 (bs, 2H, NH2), 6.01 (s, 1H, 3-
coum); UV/Vis (nm, CH3OH): lmax =209, 253, 364 (e=
15470 mꢁ1 cmꢁ1); HPLC, System 1: tR =25.60 min, purity: 99.3% at
l=230 nm; Anal. (C15H11NO2) C, H, N.
(4-Propyl-2-oxo-2H-1-benzopyran-7-yl)carbamic acid ethyl ester
(2b): From ethyl butyrylacetate (1.90 g, 12 mmol). Yield: 1.15 g
(38%); mp: 154.5–1568C; Rf =0.77 (EA/toluene 1:1); 1H NMR
(200 MHz, [D6]DMSO): d=10.13 (bs, 1H, NH), 7.74 (d, 1H, J=8.8 Hz,
6-coum), 7.56 (d, 1H, J=2.0 Hz, 9-coum), 7.40 (dd, 1H, J=8.8 and
2.0 Hz, 7-coum), 6.18 (s, 1H, 3-coum), 4.17 (q, J=7.1 Hz, 2H, CH2),
2.73 (t, J=7.5 Hz, 2H, 11-coum), 1.64 (m, 2H, 12-coum), 1.27 (t, J=
3-Fluoro-7-amino-4-methyl-2H-1-benzopyran-2-one (3d): From
2d (0.80 g, 3 mmol). Yield: 0.20 g (35%); mp: 248–2518C; Rf =0.60
112
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 103 – 117

Fluorogenic Caspase 8 Tetrapeptide Substrates
1
(EA), 0.57 (EA/toluene 1:1); H NMR (200 MHz, [D6]DMSO): d=7.42
tion mixture was magnetically stirred and gradually heated to 140–
1508C. The reaction was monitored with TLC, and after 24 h the
crude product was purified by column chromatography with
EtOAc as eluent. Caution: the reaction of 1,3-phenylenediamine with
ethyl 4,4,4-trifluoroacetoacetate (or ethyl 4-chloroacetoacetate) with-
out solvent can lead to a heterogeneous mixture and result in haz-
ardous bumping of the reactor. Yield: 0.675 g (32%); mp: 273–
2768C (Lit. 268–2708C[11]); Rf =0.35 (EA); 1H NMR (200 MHz,
[D6]DMSO): d=11.82 (bs, 1H, NH), 7.35 (m, 1H, 6-quin), 6.57 (dd,
J=8.8, 2.2 Hz, 1H; 7-quin), 6.46 (d, J=2.2 Hz, 1H, 9-quin), 6.43 (s,
1H, 3-quin), 6.16 (bs, 2H, NH2); UV/Vis (nm, CH3OH): lmax =224, 364
(e=14760 mꢁ1 cmꢁ1); HPLC, System 1: tR =10.83 min; HPLC purity:
98.9% at l=230 nm; Anal. (C10H7F3N2O) C, H, N.
(d, J=8.6 Hz, 1H, 6-coum), 6.64 (dd, J=8.6, 2.2 Hz, 1H, 7-coum),
6.47 (d, J=2.2 Hz, 1H, 9-coum), 6.05 (bs, 2H, NH2), 2.29 (d, FJ=
2.8 Hz, 3H, CH3); UV/Vis (nm, CH3OH): lmax =208, 232, 351 (e=
13090 mꢁ1 cmꢁ1); HPLC, System 1: tR =5.60 min, purity: 97.5% at
l=230 nm; Anal. (C10H8FNO2) C, H, N.
7-Amino-4-(chloromethyl)-2H-1-benzopyran-2-one (3e): From 2e
(0.85 g, 3 mmol). Yield: 0.30 g (47%); mp: 183–1858C dec. (Lit.
1878C dec.[36]); H NMR (200 MHz, [D6]DMSO): d=7.48 (d, J=8.6 Hz,
1
1H, 6-coum), 6.62–6.55 (m, 1H, 7-coum), 6.45–6.43 (m, 1H, 9-
coum), 6.23 (bs, 2H, NH2), 6.18 (s, 1H, 3-coum), 4.87 (s, 2H, CH2Cl);
UV/Vis (nm, CH3OH): lmax =208, 237, 365 (e=14210 mꢁ1 cmꢁ1);
HPLC, System 1: tR =5.84 min, purity: 98.7% at l=230 nm; Anal.
(C10H8ClNO2) C, H, N.
Synthesis of Ac-Ile-Glu(tBu)-Thr(tBu)-Asp(tBu)-OH (5): This com-
pound was prepared as described previously.[22]
General procedure for the synthesis of quinolin-2(1H)-one deriv-
atives 4a–4d: A mixture of 1,3-phenylenediamine (5.4 g, 50 mmol)
and b-keto ester (50 mmol) was gradually heated to 1508C (4a,
18 h), 1808C (4b, 18 h), 1908C (4d, 18 h), or 2008C (4c, 64 h);
(after the first hour of reaction the product usually begins to pre-
cipitate). The reaction mixture was then allowed to cool to room
temperature, and the crude solid was filtered and washed with
CH3OH. The product was crystallized from CH3OH (4a, 4b) or DMF/
H2O (4c, 4d).
Optimized coupling procedure of 5 with aromatic amines: After
dissolving 5 (0.1 mmol), HATU (0.2 mmol), and TBP (0.2 mmol) in
1 mL DMF, the resulting solution was stirred for 5 min at 208C. The
aromatic amine was added (0.15 mmol per ~200 mL DMF), and the
reaction mixture was immediately diluted with 600 mL CH2Cl2. After
five days the solution was transferred to 50 mL EtOAc and washed
three times with 1n HCl, 1n NaHCO3, and saturated NaCl in H2O.
After drying with Na2SO4, the organic layer was evaporated in va-
cuo to afford a white–yellow solid. The precipitate was washed
with Et2O to remove any remaining aromatic amine and dried. As
found by NMR and RP-HPLC studies, the product was sufficiently
pure and could be directly deprotected without further purifica-
tion. The solid was dissolved in 3 mL ice-cold 95% TFA/H2O and re-
acted for 60 min at room temperature. The resulting mixture was
poured into 30 mL ice-cold Et2O with vigorous stirring. The precipi-
tate was collected by filtration and washed with Et2O to provide a
white solid. The product was purified by analytical RP-HPLC and
lyophilized. Overall purity of the products, including the content of
the d-Asp epimer, was determined by RP-HPLC.
7-Amino-4-methyl-2(1H)-quinolinone (4a): From ethyl acetoace-
tate (6.5 g, 50 mmol). Yield: 3.48 g (40%); mp: 276–2788C dec. (Lit.
2798C[13]); 1H NMR (200 MHz, [D6]DMSO): d=10.72 (bs, 1H, NH),
7.20 (d, J=8.6 Hz, 1H, 6-quin), 6.41 (dd, J=8.6, 2.0 Hz, 1H; 7-quin),
6.27 (d, J=2.0 Hz, 1H, 9-quin), 5.87 (s, 1H, 3-quin), 5.64 (bs, 2H,
NH2), 2.28 (s, 3H, 11-quin); UV/Vis (nm, CH3OH): lmax =223, 347 (e=
15490 mꢁ1 cmꢁ1); HPLC, System 1: tR =3.71 min, purity: 99.9% at
l=230 nm; Anal. (C10H10N2O) C, H, N.
7-Amino-4-propyl-2(1H)-quinolinone (4b): From ethyl butyrylace-
tate (7.91 g, 50 mmol). Yield: 2.22 g (22%); mp: 278–2798C dec.;
1H NMR (200 MHz, [D6]DMSO): d=11.16 (bs, 1H, NH), 7.39 (d, J=
8.8 Hz, 1H, 6-quin), 6.46 (dd, J=8.8, 2.0 Hz, 1H; 7-quin), 6.38 (d, J=
2.0 Hz, 1H, 9-quin), 5.92 (s, 1H, 3-quin), 5.74 (bs, 2H, NH2), 2.63 (t,
J=7.5 Hz, 2H, 11-quin), 1.59 (m; 2H; 12-quin), 0.95 (t, J=7.3 Hz,
3H, 13-quin); UV/Vis (nm, CH3OH): lmax =223, 348 (e=
15600 mꢁ1 cmꢁ1); HPLC, System 1: tR =8.95 min, purity: 99.9% at
l=230 nm; Anal. (C12H14N2O) C, H, N.
Ac-Ile-Glu-Thr-Asp–AMC (Ac-IETD–AMC; 6a): The analytical data
of this compound are described in detail elsewhere.[22]
Ac-Ile-Glu-Thr-Asp–3b (6b): From 5 and 3b. Yield 17 mg (31%);
1H NMR (200 MHz, [D6]DMSO): d=10.16 (s, NH-coum), 8.39 (d, J=
7.4 Hz, Asp-NH), 8.17 (d, J=7.6 Hz, Glu-NH), 7.96 (d, J=8.4 Hz, Ile-
NH), 7.77 (d, J=8.8 Hz, 6-coum), 7.75 (s, 9-coum), 7.70 (d, J=
7.8 Hz, Thr-NH), 7.54 (dd, J=8.8, 1.6 Hz, 7-coum), 6.23 (s, 3-coum),
4.72 (m, Asp-Ca-H), 4.40–3.88 (m, Glu-Ca-H, Ile-Ca-H, Thr-Ca-H, Thr-
Cb-H), 2.88–2.55 (m, Asp-Cb-H2; 11-coum), 2.24 (m, Glu- Cg-H2), 1.91
(m, Glu-Cb-H1), 1.85 (s, Ac-Ile), 1.80–1.58 (m, Ile-Cb-H, Glu-Cb-H2, 12-
coum), 1.41 (m, Ile-Cg-H1), 1.11–0.94 (m, Ile-Cg-H2, Thr-CH3, 13-
coum), 0.83–0.75 (m, Ile-Cb-CH3, Ile-Cg-CH3); 13C NMR (50 MHz,
[D6]DMSO): d=174.0, 171.5, 171.4, 171.2, 169.8, 169.7, 169.4, 160.1,
156.4, 153.8, 141.8, 125.6, 115.5, 114.5, 111.5, 106.1, 66.7 (CbH-Thr),
58.0 (CaH-Thr), 56.9 (CaH-Ile), 52.0 (CH-Glu), 50.5 (CH-Asp), 36.3
(CH2-Asp), 35.8 (CbH-Ile), 32.6 (11-coum), 30.1 (CgH2-Glu), 26.9 (CbH2-
Glu), 24.4 (CH2-Ile), 22.4 (CH3-acetyl), 21.2 (12-coum), 19.1 (CH3-Thr),
15.3 (Cb-CH3-Ile), 13.6 (13-coum), 10.9 (Cg-CH3-Ile); HPLC (DAD,
lmax =230, 325 nm) System 2: tR =12.35 min, <0.1% of d-Asp
isomer in purified product, HPLC purity: 99.7% at l=325 nm; MS
(ESI) m/z calcd for C33H45N5O12: 703.8, found: m/z 702.3 [MꢁH]ꢁ,
816.3 [M+CF3COOꢁ]ꢁ.
7-Amino-4-phenyl-2(1H)-quinolinone (4c): From ethyl benzoyla-
cetate (9.60 g, 50 mmol). Yield: 2.24 g (19%); mp: >3008C (Lit.
334–3368C[12]); 1H NMR (200 MHz, [D6]DMSO): d=11.42 (bs, 1H,
NH), 7.57–7.37 (m, 5H, arom), 7.03 (d, J=8.6 Hz, 1H, 6-quin), 6.45
(dd, J=8.6, 2.0 Hz, 1H; 9-quin), 6.39 (d, J=2.0 Hz, 1H, 7-quin), 5.94
(s, 1H, 3-quin), 5.88 (bs, 2H, NH2); UV/Vis (nm, CH3OH): lmax =225,
252, 355 (e=13940 mꢁ1 cmꢁ1); HPLC, System 1: tR =15.07 min,
purity: 99.5% at l=230 nm; Anal. (C15H12N2O) C, H, N.
3-Fluoro-7-amino-4-methyl-2(1H)-quinolinone (4d): From ethyl 2-
fluoroacetoacetate (7.40, 50 mmol). Yield: 3.65 g (38%); mp:
1
>3008C; H NMR (200 MHz, [D6]DMSO): d=11.67 (bs, 1H, NH), 7.37
(d, J=8.6 Hz, 1H, 6-quin), 6.55 (d, J=8.6, 2.2 Hz, 1H; 7-quin), 6.42
(d, J=2.2 Hz, 1H, 9-quin), 5.71 (bs, 2H, NH2), 2.27 (d, FJ=2.8 Hz,
3H, 11-qiun.); UV/Vis (nm, CH3OH): lmax =221, 344 (e=
13710 mꢁ1 cmꢁ1); HPLC, System 1: tR =4.24 min, purity: 99.1% at
l=230 nm; Anal. (C10H9FN2O) C, H, N.
Ac-Ile-Glu-Thr-Asp–3c (6c): From 5 and 3c. Yield 19 mg (33%);
1H NMR (200 MHz, [D6]DMSO): d=10.20 (s, NH-coum), 8.39 (d, J=
7.4 Hz, Asp-NH), 8.16 (d, J=7.6 Hz, Glu-NH), 7.95 (d, J=8.4 Hz, Ile-
NH), 7.89 (d, J=1.6 Hz, 9-coum), 7.72 (d, J=7.6 Hz, Thr-NH), 7.57–
7.48 (m, 6H, 7-coum, arom), 7.41 (d, J=8.8 Hz, 6-coum), 6.31 (s, 3-
Synthesis of 7-amino-4-(trifluoromethyl)-2(1H)-quinolinone (4e):
1,3-Phenylenediamine (1.0 g, 9.25 mmol), ZnCl2 (1.40 g, 10 mmol),
DMSO (5 mL) and ethyl 4,4,4-trifluoroacetoacetate (1.70 g,
9.25 mmol) were added sequentially to a reaction flask. The reac-
ChemMedChem 2010, 5, 103 – 117
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
113

P. J. Bednarski et al.
MED
coum), 4.72 (m, Asp-Ca-H), 4.38–3.82 (m, Glu-Ca-H, Ile-Ca-H, Thr-Ca-
H, Thr-Cb-H), 2.74 (m, Asp-Cb-H2), 2.24 (m, Glu- Cg-H2), 1.91 (m, Glu-
Cb-H1), 1.85 (s, Ac-Ile), 1.71–1.63 (m, Ile-Cb-H, Glu-Cb-H2), 1.40 (m,
Ile-Cg-H1), 1.14 (m, Ile-Cg-H2), 1.05 (d, J=6.2 Hz, Thr-CH3), 0.83–0.75
(m, Ile-Cb-CH3, Ile-Cg-CH3); 13C NMR (50 MHz, [D6]DMSO): d=174.0,
171.6, 171.43, 171.36, 170.0, 169.8, 169.4, 159.9, 154.8, 154.3, 142.1,
134.8, 129.7, 128.9, 128.5, 127.3, 115.7, 113.9, 112.7, 106.3, 66.7
(CbH-Thr), 58.0 (CaH-Thr), 56.9 (CaH-Ile), 52.0 (CH-Glu), 50.5 (CH-
Asp), 36.3 (CH2-Asp), 35.8 (CbH-Ile), 30.1 (CgH2-Glu), 26.9 (CbH2-Glu),
24.4 (CH2-Ile), 22.4 (CH3-acetyl), 19.2 (CH3-Thr), 15.3 (Cb-CH3-Ile),
10.9 (Cg-CH3-Ile); HPLC (DAD, lmax =248, 332 nm) System 2: tR =
15.15 min, 0.18% of d-Asp isomer in purified product, HPLC purity:
97.4% at l=325 nm; MS (ESI) m/z calcd for C36H43N5O12: 737.8,
found: m/z 736.3 [MꢁH]ꢁ, 850.3 [M+CF3COOꢁ]ꢁ.
Ac-Ile-Glu-Thr-Asp–3d (6d): From 5 and 3d. Yield 16 mg (33%);
1H NMR (200 MHz, [D6]DMSO): d=10.15 (s, NH-coum), 8.38 (d, J=
7.6 Hz, Asp-NH), 8.16 (d, J=7.6 Hz, Glu-NH), 7.95 (d, J=8.4 Hz, Ile-
NH), 7.80 (d, J=1.6 Hz, 9-coum), 7.77–7.67 (m, Thr-NH, 6-coum),
7.60 (dd, J=8.6, 1.6 Hz, 7-coum), 4.72 (m, Asp-Ca-H), 4.39–3.88 (m,
Glu-Ca-H, Ile-Ca-H, Thr-Ca-H, Thr-Cb-H), 2.74 (m, Asp-Cb-H2), 2.36 (d,
FJ=2.6 Hz, 11-coum), 2.23 (m, Glu- Cg-H2), 1.91 (m, Glu-Cb-H1), 1.85
(s, Ac-Ile), 1.79–1.61 (m, Ile-Cb-H, Glu-Cb-H2), 1.40 (m, Ile-Cg-H1), 1.11
(m, Ile-Cg-H2), 1.05 (d, J=6.0 Hz, Thr-CH3), 0.83–0.75 (m, Ile-Cb-CH3,
Ile-Cg-CH3); 13C NMR (50 MHz, [D6]DMSO): d=174.0, 171.5, 171.4,
171.3, 169.9, 169.7, 169.5, 154.8, 154.2, 150.2, 144.8, 140.8, 135.8,
131.3, 114.63, 114.57, 105.9, 66.6 (CbH-Thr), 58.1 (CaH-Thr), 57.0
(CaH-Ile), 51.9 (CH-Glu), 50.5 (CH-Asp), 36.2 (CH2-Asp), 35.6 (CbH-Ile),
30.0 (CgH2-Glu), 26.7 (CbH2-Glu), 24.3 (CH2-Ile), 22.3 (CH3-acetyl), 19.1
(CH3-Thr), 15.3 (Cb-CH3-Ile), 10.8 (Cg-CH3-Ile), 9.7 (CH3-coum); HPLC
(DAD, lmax =231, 322 nm) System 3: tR =15.87 min, 0.45% of d-Asp
isomer in purified product, HPLC purity: 99.3% at l=325 nm; MS
(ESI) m/z calcd for C31H40F1N5O12: 693.7, found: m/z 692.3 [MꢁH]ꢁ,
806.2 [M+CF3COOꢁ]ꢁ.
(50 MHz, [D6]DMSO): d=173.9, 171.5, 171.4, 171.3, 170.2, 169.8,
169.3, 158.5, 154.6, 142.8, 125.3, 119.0, 116.2, 114.5, 108.5, 106.5,
66.7 (CbH-Thr), 58.2 (CaH-Thr), 56.9 (CaH-Ile), 52.0 (CH-Glu), 50.6
(CH-Asp), 36.2 (CH2-Asp), 35.7 (CbH-Ile), 30.0 (CgH2-Glu), 26.9 (CbH2-
Glu), 24.4 (CH2-Ile), 22.4 (CH3-acetyl), 19.2 (CH3-Thr), 15.3 (Cb-CH3-
Ile), 10.8 (Cg-CH3-Ile); HPLC (DAD, lmax =234, 337 nm) System 3: tR =
19.01 min, 0.22% of d-Asp isomer in purified product, HPLC purity:
98.0% at l=325 nm; MS (ESI) m/z calcd for C31H38F3N5O12: 729.7,
found: m/z 728.2 [MꢁH]ꢁ, 842.2 [M+CF3COOꢁ]ꢁ.
Ac-Ile-Glu-Thr-Asp–4a (7a): From 5 and 4a. Yield 34 mg (53%);
1H NMR (200 MHz, [D6]DMSO): d=11.54 (s, 1-quin), 9.99 (s, NH-
quin), 8.32 (d, J=7.6 Hz, Asp-NH), 8.14 (d, J=7.8 Hz, Glu-NH), 7.94
(d, J=8.4 Hz, Ile-NH), 7.86 (d, J=1.6 Hz, 9-quin), 7.68 (d, J=7.8 Hz,
Thr-NH), 7.62 (d, J=8.8 Hz, 6-quin), 7.27 (dd, J=8.8, 1.6 Hz, 7-quin),
6.26 (s, 3-quin), 4.73 (m, Asp-Ca-H), 4.38–4.11 (m, Glu-Ca-H, Thr-Ca-
H, Ile-Ca-H), 3.99 (m, Thr-Cb-H), 2.72 (m, Asp-Cb-H2), 2.37 (s, 11-quin),
2.23 (m, Glu-Cg-H2), 1.92 (m, Glu-Cb-H1), 1.86 (s, Ac-Ile), 1.81–1.65
(m, Ile-Cb-H, Glu-Cb-H2), 1.42 (m, Ile-Cg-H1), 1.11 (m, Ile-Cg-H2), 1.05
(d, J=6.2 Hz, Thr-CH3), 0.83–0.75 (m, Ile-Cb-CH3, Ile-Cg-CH3);
13C NMR (50 MHz, [D6]DMSO): d=173.9, 171.5, 171.4, 171.2, 169.6,
169.4, 169.3, 161.9, 147.5, 140.2, 139.3, 125.1, 119.1, 115.7, 113.7,
104.8, 66.6 (CbH-Thr), 58.0 (CaH-Thr), 56.9 (CaH-Ile), 51.9 (CH-Glu),
50.5 (CH-Asp), 36.2 (CH2-Asp), 35.9 (CbH-Ile), 30.1 (CgH2-Glu), 26.9
(CbH2-Glu), 24.4 (CH2-Ile), 22.4 (CH3-acetyl), 19.1 (CH3-Thr), 18.3 (11-
quin), 15.3 (Cb-CH3-Ile), 10.9 (Cg-CH3-Ile); HPLC (DAD, lmax =229,
286, 327, 342 nm) System 4: tR =12.03 min, <0.1% of d-Asp
isomer in purified product, HPLC purity: 94.4% at l=325 nm; MS
(ESI) m/z calcd for C31H42N6O11: 674.7, found: m/z 673.3 [MꢁH]ꢁ,
787.3 [M+CF3COOꢁ]ꢁ.
Ac-Ile-Glu-Thr-Asp–4b (7b): From 5 and 4b. Yield 7 mg (19%);
1H NMR (200 MHz, [D6]DMSO): d=11.55 (s, 1-quin), 9.99 (s, NH-
quin), 8.32 (d, J=7.4 Hz, Asp-NH), 8.15 (d, J=7.4 Hz, Glu-NH), 7.94
(d, J=8.4 Hz, Ile-NH), 7.84 (s, 9-quin), 7.69–7.65 (m, Thr-NH, 6-quin),
7.27 (d, J=8.2 Hz, 7-quin), 6.22 (s, 3-quin), 4.73 (m, Asp-Ca-H), 4.36–
4.10 (m, Glu-Ca-H; Thr-Ca-H; Ile-Ca-H), 3.99 (m, Thr-Cb-H), 2.86–2.56
(m, Asp-Cb-H2; 11-quin), 2.23 (m, Glu-Cg-H2), 1.92 (m, Glu-Cb-H1),
1.80 (s, Ac-Ile), 1.82–1.53 (m, Ile-Cb-H; Glu-Cb-H2, 12-quin), 1.39 (m,
Ile-Cg-H1), 1.09–0.94 (m, Ile-Cg-H2; Thr-CH3; 13-quin), 0.83–0.75 (m,
Ile-Cb-CH3, Ile-Cg-CH3); 13C NMR (50 MHz, [D6]DMSO): d=173.9,
171.5, 171.4, 171.2, 169.6, 169.4, 169.3, 161.9, 151.1, 140.1, 139.7,
124.9, 118.2, 114.9, 113.8, 105.1, 66.6 (CbH-Thr), 58.0 (CaH-Thr), 56.9
(CaH-Ile), 52.0 (CH-Glu), 50.5 (CH-Asp), 36.2 (CH2-Asp), 35.9 (CbH-Ile),
33.2 (11-quin), 30.1 (CgH2-Glu), 26.9 (CbH2-Glu), 24.4 (CH2-Ile), 22.4
(CH3-acetyl), 21.8 (12-quin), 19.1 (CH3-Thr), 15.3 (Cb-CH3-Ile), 13.7
(13-quin), 10.8 (Cg-CH3-Ile); HPLC (DAD, lmax =229, 287, 329,
340 nm) System 5: tR =16.11 min, 0.81% of d-Asp isomer in puri-
fied product, HPLC purity: 99.0% at l=325 nm; MS (ESI) m/z calcd
for C33H46N6O11: 702.8, found: m/z 701.3 [MꢁH]ꢁ, 815.3
[M+CF3COOꢁ]ꢁ.
Ac-Ile-Glu-Thr-Asp–3e (6e): From 5 and 3e. Yield 15 mg (35%);
1H NMR (200 MHz, [D6]DMSO): d=12.30 (bs, COOH), 10.21 (s, NH-
coum), 8.39 (d, J=7.6 Hz, Asp-NH), 8.17 (d, J=7.4 Hz, Glu-NH), 7.96
(d, J=8.4 Hz, Ile-NH), 7.81–7.77 (m, 9-coum, 6-coum), 7.69 (d, J=
7.6 Hz, Thr-NH), 7.55 (d, J=8.8 Hz, 7-coum), 6.56 (s, 3-coum), 4.99
(s, CH2Cl), 4.73 (m, Asp-Ca-H), 4.38–3.91 (m, Glu-Ca-H, Ile-Ca-H, Thr-
Ca-H, Thr-Cb-H), 2.74 (m, Asp-Cb-H2), 2.24 (m, Glu- Cg-H2), 1.93 (m,
Glu-Cb-H1), 1.85 (s, Ac-Ile), 1.81–1.63 (m, Ile-Cb-H; Glu-Cb-H2), 1.41
(m, Ile-Cg-H1), 1.11 (m, Ile-Cg-H2), 1.05 (d, J=6.0 Hz, Thr-CH3), 0.83–
0.75 (m, Ile-Cb-CH3, Ile-Cg-CH3); 13C NMR (50 MHz, [D6]DMSO): d=
174.0, 171.5, 171.4, 171.2, 170.0, 169.7, 169.4, 159.8, 154.0, 150.5,
142.2, 125.8, 115.5, 113.3, 112.6, 106.1, 66.7 (CbH-Thr), 58.0 (CaH-
Thr), 56.9 (CaH-Ile), 52.0 (CH-Glu), 50.5 (CH-Asp), 41.6 (CH2Cl), 36.3
(CH2-Asp), 35.8 (CbH-Ile), 30.1 (CgH2-Glu), 26.9 (CbH2-Glu), 24.4 (CH2-
Ile), 22.4 (CH3-acetyl), 19.1 (CH3-Thr), 15.3 (Cb-CH3-Ile), 10.9 (Cg-CH3-
Ile); HPLC purity: (DAD, lmax =232, 330 nm) System 3: tR =
16.11 min, 3.62% of d-Asp isomer in purified product, purity:
95.1% at l=325 nm; MS (ESI) m/z calcd for C31H40Cl1N5O12: 710.4,
found: m/z 708.2 [Mꢁ2H]2ꢁ, 822.2 [MꢁH+CF3COOꢁ]2ꢁ. The struc-
ture was further identified by DEPT.
Ac-Ile-Glu-Thr-Asp–4c (7c): From 5 and 4c. Yield 9 mg (26%);
1H NMR (200 MHz, [D6]DMSO): d=12.29 (bs, COOH), 11.82 (s, 1-
quin), 10.03 (s, NH-quin), 8.32 (d, J=7.4 Hz, Asp-NH), 8.14 (d, J=
7.6 Hz, Glu-NH), 8.00 (s, 9-quin) 7.93 (d, J=8.4 Hz, Ile-NH), 7.69 (d,
J=7.6 Hz, Thr-NH), 7.54–7.44 (m, 5H, arom), 7.31 (d, J=8.8 Hz, 6-
quin), 7.21 (d, J=8.8 Hz, 7-quin), 6.26 (s, 3-quin), 4.73 (m, Asp-Ca-
H), 4.40–4.10 (m, Glu-Ca-H, Thr-Ca-H, Ile-Ca-H), 3.98 (m, Thr-Cb-H),
2.73 (m, Asp-Cb-H2), 2.24 (m, Glu-Cg-H2), 1.91 (m, Glu-Cb-H1), 1.85 (s,
Ac-Ile), 1.81–1.63 (m, Ile-Cb-H, Glu-Cb-H2), 1.42 (m, Ile-Cg-H1), 1.10
(m, Ile-Cg-H2), 1.05 (d, J=6.0 Hz, Thr-CH3), 0.82–0.74 (m, Ile-Cb-CH3,
Ile-Cg-CH3); 13C NMR (50 MHz, [D6]DMSO): d=173.9, 171.5, 171.4,
171.3, 169.7, 169.6, 169.3, 161.6, 151.1, 140.5, 140.1, 136.8, 128.63,
128.57, 126.6, 119.3, 114.3, 114.1, 105.0, 66.6 (CbH-Thr), 58.1 (CaH-
Ac-Ile-Glu-Thr-Asp–AFC (6 f): From 5 and AFC. Yield 13 mg (19%);
1H NMR (200 MHz, [D6]DMSO): d=12.29 (bs, COOH), 10.29 (s,
coum-NH), 8.39 (d, J=7.4 Hz, Asp-NH), 8.15 (d, J=7.6 Hz, Glu-NH),
7.94 (d, J=8.4 Hz, Ile-NH), 7.89 (s, 9-coum), 7.72–7.59 (m, Thr-NH,
6-coum, 7-coum), 6.92 (s, 3-coum), 4.73 (m, Asp-Ca-H), 4.38–3.70
(m, Glu-Ca-H, Ile-Ca-H, Thr-Ca-H, Thr-Cb-H), 2.75 (m, Asp-Cb-H2), 2.24
(m, Glu- Cg-H2), 1.92 (m, Glu-Cb-H1), 1.85 (s, Ac-Ile), 1.81–1.61 (m,
Ile-Cb-H, Glu-Cb-H2), 1.41 (m, Ile-Cg-H1), 1.11 (m, Ile-Cg-H2), 1.05 (d,
J=6.0 Hz, Thr-CH3), 0.79 (m, Ile-Cb-CH3; Ile-Cg-CH3); 13C NMR
114
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 103 – 117

Fluorogenic Caspase 8 Tetrapeptide Substrates
Thr), 56.9 (CaH-Ile), 52.0 (CH-Glu), 50.5 (CH-Asp), 36.2 (CH2-Asp),
35.9 (CbH-Ile), 30.1 (CgH2-Glu), 26.9 (CbH2-Glu), 24.4 (CH2-Ile), 22.4
(CH3-acetyl), 19.1 (CH3-Thr), 15.3 (Cb-CH3-Ile), 10.8 (Cg-CH3-Ile); HPLC
(DAD, lmax =231, 294, 334, 347 nm) System 5: tR =17.81 min, 0.41%
of d-Asp isomer in purified product, HPLC purity: 99.2% at l=
325 nm; MS (ESI) m/z calcd for C36H44N6O11: 736.8, found: m/z 735.3
[MꢁH]ꢁ, 849.3 [M+CF3COOꢁ]ꢁ.
Abicap-Messplatz) at lex =355 nm, lem =460 nm in black 96-well
microtiter plates with clear well bottoms (BD Biosciences: BD Fal-
conTM MicrotestTM). Photometric detection was performed with a
2010 plate reader (Anthos) at l 405 nm in 96-well microtiter plates
(Sarstedt). Assays were performed at a final volume of 60 mL per
well.
Twofold dilutions (with assay buffer) of the appropriate substrates
were prepared. Each substrate (overall six different concentrations,
20 mL each) was added per well (the final substrate concentrations
were first obtained after the dilution with an enzyme). The concen-
tration range was chosen so that at least two substrate concentra-
tions were above and below the KM value (KM values were provi-
sionally determined in the preliminary tests). The maximal final
substrate concentration applied was 1600 mm. Aliquots of appropri-
ate enzyme were diluted with assay buffer and incubated for
~5 min at room temperature. Enzyme solution (40 mL) was added
with a multichannel pipette (thus all the substrates could be as-
sayed simultaneously) to the wells containing appropriate sub-
strate. The microtiter plate was briefly shaken, and the release of
reporter groups was immediately measured. Measurements were
done in 2 min intervals for 30 min. Initial velocities were usually
calculated from increases obtained within the first 10 min. As a
negative control the measurements at 0 min for appropriate sub-
strate dilutions were made (initial control experiments revealed
that there is no significant change in the fluorescence/absorption
of the substrate solution when assayed without enzyme). The ini-
tial velocity (V, pmolminꢁ1) at each substrate concentration (S) was
calculated from the calibration curves, each determined for the ap-
propriate reporter group under the appropriate conditions. The
same experimental KM and Vmax values were calculated from S/V
against S plot (Hanes–Wolff plot).[36] The rel. Vmax value for appropri-
ate substrate was calculated by dividing Vmax (determined for this
substrate) by Vmax determined for reference substrate (i.e., Ac-
DMQD–AMC for caspase 3 and Ac-IETD–AMC for caspase 8). The re-
ported values are the average of three to four independent deter-
minations.
Ac-Ile-Glu-Thr-Asp–4d (7d): From 5 and 4d. Yield 16 mg (47%);
1H NMR (200 MHz, [D6]DMSO): d=12.11 (d, JF =5.4 Hz, 1-quin),
10.02 (s, quin-NH), 8.33 (d, J=7.4 Hz, Asp-NH), 8.15 (d, J=7.8 Hz,
Glu-NH), 7.98–7.93 (m, Ile-NH, 9-quin), 7.70–7.63 (m, Thr-NH, 6-
quin), 7.35 (dd, J=8.8, 1.8 Hz, 7-quin), 4.73 (m, Asp-Ca-H), 4.38–421
(m, Glu-Ca-H, Ile-Ca-H), 4,15 (m, Thr-Ca-H), 3.98 (m, Thr-Cb-H), 2.73
(m, Asp-Cb-H2), 2.34 (d, FJ=2.6 Hz, 11-quin), 2.25 (m, Glu- Cg-H2),
1.91 (m, Glu-Cb-H1), 1.85 (s, Ac-Ile), 1.81–1.62 (m, Ile-Cb-H, Glu-Cb-
H2), 1.41 (m, Ile-Cg-H1), 1.11 (m, Ile-Cg-H2), 1.05 (d, J=6.2 Hz, Thr-
CH3), 0.83–0.75 (m, Ile-Cb-CH3, Ile-Cg-CH3); 13C NMR (50 MHz,
[D6]DMSO): d=174.0, 171.6, 171.4, 171.2, 169.7, 169.5, 169.4, 155.5,
154.9, 149.8, 144.9, 139.5, 135.8, 127.3, 125.2, 114.7, 105.0, 66.7
(CbH-Thr), 58.1 (CaH-Thr), 56.9 (CaH-Ile), 52.0 (CH-Glu), 50.5 (CH-
Asp), 36.3 (CH2-Asp), 35.9 (CbH-Ile), 30.1 (CgH2-Glu), 26.9 (CbH2-Glu),
24.4 (CH2-Ile), 22.4 (CH3-acetyl), 19.1 (CH3-Thr), 15.3 (Cb-CH3-Ile),
10.9 (Cg-CH3-Ile), 9.7 (11-quin); HPLC (DAD, lmax =225, 282, 324,
336 nm) System 3: tR =12.45 min, <0.1% of d-Asp isomer in puri-
fied product, HPLC purity: 96.9% at l=325 nm; MS (ESI) m/z calcd
for C31H41F1N6O11: 692.7, found: m/z 691.3 [MꢁH]ꢁ, 805.3
[M+CF3COOꢁ]ꢁ.
Ac-Ile-Glu-Thr-Asp–4e (7e): From 5 and 4e. Yield 16 mg (40%);
1H NMR (200 MHz, [D6]DMSO): d=12.26 (s, 1-quin), 10.18 (s, NH-
quin), 8.34 (d, J=7.4 Hz, Asp-NH), 8.15 (d, J=7.6 Hz, Glu-NH), 8.06
(s, 9-quin) 7.94 (d, J=8.4 Hz, Ile-NH), 7.71–7.62 (m, Thr-NH; 6-quin),
7.38 (d, J=8.8 Hz, 7-quin), 6.83 (s, 3-quin), 4.73 (m, Asp-Ca-H), 4.35–
4.11 (m, Glu-Ca-H, Thr-Ca-H, Ile-Ca-H), 3.98 (m, Thr-Cb-H), 2.77 (m,
Asp-Cb-H2), 2.23 (m, Glu-Cg-H2), 1.92 (m, Glu-Cb-H1), 1.85 (s, Ac-Ile),
1.81–1.62 (m, Ile-Cb-H, Glu-Cb-H2), 1.41 (m, Ile-Cg-H1), 1.12 (m, Ile-
Cg-H2), 1.05 (d, J=6.2 Hz, Thr-CH3), 0.83–0.75 (m, Ile-Cb-CH3, Ile-Cg-
CH3); 13C NMR (50 MHz, [D6]DMSO): d=173.9, 171.5, 171.4, 171.2,
169.8, 169.7, 169.3, 160.3, 149.5, 141.4, 140.7, 136.1, 124.7, 119.5,
115.0, 108.9, 105.2, 66.6 (CbH-Thr), 58.1 (CaH-Thr), 56.9 (CaH-Ile),
52.0 (CH-Glu), 50.5 (CH-Asp), 36.2 (CH2-Asp), 35.8 (CbH-Ile), 30.0
(CgH2-Glu), 26.9 (CbH2-Glu), 24.4 (CH2-Ile), 22.4 (CH3-acetyl), 19.2
(CH3-Thr), 15.3 (Cb-CH3-Ile), 10.8 (Cg-CH3-Ile); HPLC (DAD, lmax =228,
340, 355 nm) System 5: tR =16.35 min, 1.21% of d-Asp isomer in
purified product, HPLC purity: 98.6% at l=325 nm; MS (ESI) m/z
calcd for C31H39F3N6O11: 728.7, found: m/z 727.3 [MꢁH]ꢁ, 841.2
[M+CF3COOꢁ]ꢁ.
Molecular modeling
The structures of leaving amines (AXX) and substrates Ac-IETD–
AXX were constructed and minimized with the INSIGHT II software
package (2005) from Accelrys (San Diego, CA, USA) by using the
Discover module and the consistent valence force field (CVFF). Au-
toDock 3[35,42] (autodock.scripps.edu) was used for the docking sim-
ulations and to calculate changes in free energy of binding (DG)
for each docked substrate to either caspase 3 or 8. Root-mean-
square values were calculated with PCModel (v. 8.5) (Serena Soft-
ware, Bloomington, IN, USA). INSIGHT II and AutoDock were run on
either SGI Octane2 or Fuel workstations.
Determination of KM, Vmax, and rel. Vmax/KM
Caspases 3 and 8 were used at ~0.05–0.1 U per assay (per well),
whereas caspase 9 was present at ~1.0 U per assay (per well). It is
important to note that the same catalytic units of different charges
of appropriate caspase were found to cleave the same substrate
with substantially different velocities (if any). Thus the appropriate
AMC conjugate was always added as reference substrate. Assay
buffers were prepared freshly (or frozen at ꢁ208C for a maximum
of one month in 1–1.5 mL aliquots and thawed prior to use) accord-
ing to the supplier’s information. Caspase substrates were stored
in 10–40 mm stock solutions in DMSO at ꢁ208C. The final concen-
tration of DMSO under assay conditions was never >4%. Reporter
groups were stored in 5 mm stock solutions in DMSO at 48C. Fluo-
rescence measurements were done with a Fluoroskan II (ABION,
The X-ray crystal structures of caspases 3 (PDB code: 1RHK[34]) and
8 (PDB code: 1QTN[33]), covalently modified with the respective in-
hibitor, were downloaded from the Brookhaven Protein Data Bank.
Caspase 3 is covalently modified with an Ac-DEVD-based inhibitor,
whereas caspase 8 is covalently bound to an Ac-IETD inhibitor.
With the Biopolymer module of INSIGHT II, all water oxygen atoms
were removed from the X-ray structures, hydrogen atoms were
added to the enzyme and inhibitor, the bond orders were auto-
matically corrected for, and atomic partial charges were calculated
with the CVFF. The thiohemiacetal bond between the active site
cysteine and inhibitor was broken to give the free enzyme and a
tetrapeptide derivative with a free hydroxy group. The enzyme thi-
olate was re-protonated, and the inhibitor was removed. Next, the
ChemMedChem 2010, 5, 103 – 117
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
115

P. J. Bednarski et al.
MED
enzymes were soaked in two shells of water, the inner shell with
movable waters, and the outer shell with fixed waters. Both
enzyme structures were then minimized to an RMSD of 0.001 by
the conjugated gradient method with cross-terms, but no Morse
potential, using the CVFF. Water molecules were then removed,
and the structures were saved in .pdb format for docking with Au-
toDock.
Keywords: AutoDock · caspases · fluorescence · molecular
modeling
[1] S. Kumar, Cell Death Differ. 2007, 14, 32–43.
[2] A. M. Scott, M. Saleh, Cell Death Differ. 2007, 14, 23–31.
[3] R. V. Talanian, K. D. Brady, V. L. Cryns, J. Med. Chem. 2000, 43, 3351–
3371.
To create the caspase substrates (Ac-IETD–AXX), the Builder
module of INSIGHT II was used to convert the hydroxy group of
the IETD inhibitor to an aldehyde and to attach it to the appropri-
ate reporter group via the aromatic amine to form the amide
bond. For each substrate the partial charges were recalculated by
INSIGHT II and minimized before saving in .pdb format for docking
with AutoDock.
[4] D. W. Nicholson, Cell Death Differ. 1999, 6, 1028–1042.
[5] H. Y. Chang, X. Yang, Microbiol. Mol. Biol. Rev. 2000, 64, 821–846.
[6] N. A. Thornberry, T. A. Rano, E. P. Peterson, D. M. Rasper, T. Timkey, M.
Garcia-Calvo, V. M. Houtszager, P. A. Nordstrom, S. Roy, J. P. Vaillancourt,
K. T. Chapman, D. W. Nicholson, J. Biol. Chem. 1997, 272, 17907–17911.
[7] R. V. Talanian, C. Quinlan, S. Trautz, M. C. Hackett, J. A. Mankovich, D.
Banach, T. Ghayur, K. D. Brady, W. W. Wong, J. Biol. Chem. 1997, 272,
9677—9682.
The substrate docking simulations were performed within the in-
hibitor binding pocket as identified by the X-ray structures of cas-
pases 3 and 8. The most important AutoDock parameter settings
were as follows: the grid size was set to include the binding site of
the tetrapeptide inhibitors as well as the catalytic cysteine (AA 285
in caspase 3 and AA 360 in caspase 8). The resolution of the grid
was 70ꢂ70ꢂ80 grid points with a spacing of 0.275 ꢁ. Each sub-
strate was individually docked into this grid with the Lamarckian
algorithm as implemented in AutoDock. For the docking simula-
tion, all sigma bonds of the inhibitor with the exception of the
peptide bonds were freely rotatable. Peptide bonds were fixed in
the trans conformation. The genetic-based algorithm ran 80 simu-
lations per substrate with 3000000 energy evaluations and a maxi-
mum number of generations of 27000. The crossover rate was in-
creased to 0.85, and the number of individuals in each population
to 150. All other parameters were left at the AutoDock default set-
tings.
[8] A. L. Niles, R. A. Moravec, T. L. Riss, “Caspase Activity Assays” in: Apopto-
sis and Cancer (Eds.: G. Mor, A. B. Alvero), Humana Press, Totowa, 2008,
pp. 137–150.
[9] G. K. Smith, D. G. Barrett, K. Blackburn, M. Cory, W. S. Dallas, R. Davis, D.
Hassler, R. McConnell, M. Moyer, K. Weaver, Arch. Biochem. Biophys.
2002, 399, 195–205.
[10] H. von Pechmann, O. Schwarz, Chem. Ber. 1899, 32, 3696–3699.
[11] R. L. Atkins, D. E. Bliss, J. Org. Chem. 1978, 43, 1975–1980.
[12] G. Kokotos, R. Geiger, C. Tzougraki, Biol. Chem. Hoppe–Seyler 1990, 371,
835–840.
[13] L. L. Woods, M. M. Fooladi, J. Chem. Eng. Data 1968, 13, 440–442.
[14] A. Chilin, P. Rodighiero, G. Pastorini, A. Guiotto, J. Org. Chem. 1991, 56,
980–983.
[15] J. Chen, P. R. Selvin, J. Photochem. Photobiol. A 2000, 135, 27–32.
[16] M. Zimmerman, E. Yurewicz, G. Patel, Anal. Biochem. 1976, 70, 258–262.
[17] E. R. Bissell, A. R. Mitchell, R. E. Smith, J. Org. Chem. 1980, 45, 2283–
2287.
[18] P. R. Hammond, A. N. Fletcher, R. A. Henry, R. L. Atkins, Appl. Phys. 1975,
8, 311–314.
[19] R. E. Smith, E. R. Bissell, A. R. Mitchell, K. W. Pearson, Thromb. Res. 1980,
17, 393–402.
[20] H.-K. Lee, H. Cao, T. M. Rana, J. Comb. Chem. 2005, 7, 279–284.
[21] W. M. F. Fabian, K. S. Niederreiter, G. Uray, W. Stadlbauer, J. Mol. Struct.
1999, 477, 209–220.
[22] P. Reszka, K. Methling, M. Lalk, Z. Xiao, K. Weisz, P. J. Bednarski, Tetrahe-
dron: Asymmetry 2008, 19, 49–59.
[23] L. A. Carpino, A. El-Faham, Tetrahedron Lett. 1994, 35, 2279–2282.
[24] A. Speicher, T. Klaus, T. Eicher, J. Prakt. Chem. 1998, 340, 581–583.
[25] F. Albericio, J. M. Bofill, A. El-Faham, S. A. Kates, J. Org. Chem. 1998, 63,
9678–9683.
[26] L. A. Carpino, D. Ionescu, A. El-Faham, J. Org. Chem. 1996, 61, 2460–
2465.
[27] H. Zou, R. Yang, J. Hao, J. Wang, C. Sun, S. W. Fesik, J. C. Wu, K. J. Toma-
selli, R. C. Armstrong, J. Biol. Chem. 2003, 278, 8091–8098.
[28] P. Martinsson, G. Liminga, P. Nygren, R. Larsson, Anticancer Drugs 2001,
12, 699–705.
[29] S. Zhuang, G. Simon, Am. J. Physiol. 2000, 279, 341–351.
[30] A. Komoriya, B. Z. Packard, M. J. Brown, M.-L. Wu, P. A. Henkart, J. Exp.
Med. 2000, 191, 1819–1828.
[31] K. B. Kwon, E. K. Kim, M. J. Han, B. C. Shin, Y. K. Park, K. S. Kim, Y. R. Lee,
J. W. Park, B. H. Park, D. G. Ryu, Biol. Pharm. Bull. 2006, 29, 1082–1086.
[32] H. Feng, Y. Zeng, M. Graner, L. Whitesell, E. Katsanis, J. Biomed. Biotech-
nol. 2004, 2004, 41–51.
For evaluation of the docking results, the substrate orientation had
to fulfill the following four conditions to pass as a potentially pro-
ductive conformation for catalysis: 1) the distance between the C
atom that forms the semi-thiohemiketal with the S atom of
Cys(285/360) is <4.00 ꢁ; 2) the O atom in the peptide bond of the
Asp residue must be orientated toward His(237/317) with a dis-
tance <4.50 ꢁ; 3) the Asp side chain must point into the S1 bind-
ing pocket, and the distance to Arg(179/260), Arg(341/413), and
Gln(283/358) must be <4.75 ꢁ (numbers in brackets refer to
amino acid positions, the first for caspase 3, second for caspase 8);
and 4) the tetrapeptide backbone of the substrate must be orien-
tated in the binding site analogous to the tetrapeptide inhibitor.
MD simulations were performed at 310 K for 20 ps with a 1 fs
time-step by using the CVFF, as implemented in the Discovery
module of INSIGHT II. The simulations were started with the pro-
ductive enzyme–substrate complex obtained from the AutoDock
simulations. Each complex was surrounded by a shell of explicit
water molecules, the ionization of basic and acidic groups was
based on a system at pH 7.4, and all bonds were unconstrained
during the simulation.
[33] W. Watt, K. A. Koeplinger, A. M. Mildner, R. L. Heinrikson, A. G. Tomasselli,
K. D. Watenpaugh, Structure 1999, 7, 1135–1143.
[34] J. W. Becker, J. Rotonda, S. M. Soisson, R. Aspiotis, C. Bayly, S. Francoeur,
M. Gallant, M. Garcia-Calvo, A. Giroux, E. Grimm, Y. Han, D. McKay, D. W.
Nicholson, E. Peterson, J. Renaud, S. Roy, N. Thornberry, R. Zamboni, J.
Med. Chem. 2004, 47, 2466–2474.
Acknowledgements
[35] G. M. Morris, D. S. Goodsell, R. S. Halliday, R. Huey, W. E. Hart, R. K.
Belew, A. J. Olson, J. Comput. Chem. 1998, 19, 1639–1662.
[36] R. A. Copeland, Enzymes: A Practical Introduction to Structure, Mecha-
nism and Data Analysis, 2nd Ed., John Wiley & Sons, New York, 2000,
pp. 109–145.
We thank Dr. S. Jakobs for acquiring the CD spectra, M.-A. Omer-
Adam for lyophilization of the final products, P. Pfeffer and Prof.
G. Klebe of the University of Marburg, (Germany) for introducing
us to AutoDock, and Prof. V. Liebscher for valuable discussions.
116
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 103 – 117

Fluorogenic Caspase 8 Tetrapeptide Substrates
[37] J. Liu, S. S. Ericksen, D. Besspiata, C. W. Fisher, G. D. Szklarz, Drug Metab.
Dispos. 2003, 31, 412–420.
[38] R. A. Chica, P. Gagnon, J. W. Keillor, J. N. Pelletier, Protein Sci. 2004, 13,
979–991.
of Success for Diverse Protein Targets” in: Drug Discovery Research (Ed.:
Z. Huang), John Wiley & Sons, Hoboken, 2007, pp. 24–62.
[46] D. R. J. Hauser, T. Scior, D. M. Domeyer, B. Kammerer, S. A. Laufer, J. Med.
Chem. 2007, 50, 2060–2066.
[47] P. M. Coutinho, M. K. Dowd, P. J. Reilly, Proteins 1997, 27, 235–248.
[48] A. D. Hill, P. J. Reilly, Biopolymers 2008, 89, 1021–1031.
[49] B. Mertz, A. D. Hill, C. Mulakala, P. J. Reilly, Biopolymers 2007, 87, 249–
260.
[39] S. R. Kanade, V. L. Suhas, N. Chandra, L. R. Gowda, FEBS J., 2007, 274,
4177–4187.
[40] B. Fang, P. I. Boross, J. Tozser, I. T. Weber, J. Mol. Biol. 2006, 360, 654–
666.
[50] Y. Ito, H. Kondo, P. S. Goldfarb, D. F. V. Lewis, J. Mol. Graph. Model. 2008,
26, 947–956.
[51] H. Sun, C. Moore, P. M. Dansette, S. Kumar, J. R. Halpert, G. S. Yost, Drug
Metab. Dispos. 2009, 37, 672–684.
[52] B. Curten, P. Kullmann, M. E. Bier, K. Kandler, B. F. Schmidt, Photochem.
Photobiol. 2005, 81, 641–648.
[41] M. A. O’Brien, W. J. Daily, P. E. Hesselberth, R. A. Moravec, M. A. Scurria,
D. H. Klaubert, R. F. Bulleit, K. V. Wood, J. Biomol. Screen. 2005, 10, 137–
148.
[42] Y. Kwon, K. Welsh, A. R. Mitchell, J. A. Camarero, Org. Lett. 2004, 6,
3801–3804.
[43] D. S. Goodsell, D. S. Morris, A. J. Olson, J. Mol. Recognit. 1996, 9, 1–5.
[44] C. L. Li, L. Xu, D. W. Wolan, I. A. Wilson, A. J. Olson, J. Med. Chem. 2004,
47, 6681–6690.
Received: August 26, 2009
[45] S. Ghosh, A. Nie, J. An, Z. Huang, “Recent Advances in Drug Discovery
Research Using Structure-Based Virtual Screening Techniques: Examples
Revised: October 16, 2009
Published online on November 13, 2009
ChemMedChem 2010, 5, 103 – 117
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
117