Synthesis of new 3-alkoxy-7-amino-4-chloro-isocoumarin derivatives as new β-amyloid peptide production inhibitors and their activities on various classes of protease

Article abstract of DOI:10.1016/S0968-0896(03)00235-9

A series of new 7-substituted-4-chloro-3-alkoxy isocoumarin derivatives were synthesized and evaluated as inhibitors of representative classes of proteases: serine protease (α-chymotrypsin, trypsin), cysteine protease (Caspase-3), and aspartyl protease (H

Full text of DOI:10.1016/S0968-0896(03)00235-9

Bioorganic & Medicinal Chemistry 11 (2003) 3141–3152
Synthesis of New 3-Alkoxy-7-amino-4-chloro-isocoumarin
Derivatives as New ꢀ-Amyloid Peptide Production Inhibitors and
Their Activities on Various Classes of Protease
Frederic Bihel,a Gilles Quelever,a Hugues Lelouard,e Agnes Petit,b
Cristine Alves da Costa,b Olivier Pourquie,c Frederic Checler,b
Annie Thellend,d Philippe Pierree and Jean-Louis Krausa,*
a
´
´
INSERM U-382, Developmental Biology Institute of Marseille (CNRS-INSERM-Univ. Mediterranee- AP Marseille),
´
´
Laboratoire de Chimie Biomoleculaire, Faculte des Sciences de Luminy, case 907, 13288 Marseille Cedex 09, France
´
b
Institut de Pharmacologie Moleculaire et Cellulaire du CNRS, UMR6097, 660 route des Lucioles, 06560 Valbonne, France
c
´
´
LGPD Unite Mixte de Recherche 6545, Faculte des Sciences de Luminy, case 907, 13288 Marseille Cedex 09, France
dLaboratoire de Chimie Organique Biologique, UMR 7613, 4, place Jussieu, 75252 Paris Cedex 5, France
e
´ ´ ´
Centre d’Immunologie de Marseille-Luminy, CNRS-INSERM- Univ. Mediterranee, Faculte des Sciences de Luminy, Case 906,
13288 Marseille Cedex 09, France
Received 3 January 2003; accepted 4 April 2003
Abstract—A series of new 7-substituted-4-chloro-3-alkoxy isocoumarin derivatives were synthesized and evaluated as inhibitors of
representative classes of proteases: serine protease (a-chymotrypsin, trypsin), cysteine protease (Caspase-3), and aspartyl protease
(HIV-protease), 20S proteasome and also as inhibitors of amyloid peptide g-secretase-mediated production. Protease inhibition
selectivity is directly related to the structure of the substituent at the 7-position of the isocoumarin nucleus. 7-Nitro-isocoumarin
derivatives (4c, 4d, 4f) are potent a-chymotrypsin inhibitors but slightly active or inactive on HIV-protease, as well as on cysteine
protease. In contrast, only derivatives bearing a free amino (5d, 5f) or a substituted amino group( 6f) at the 7-position of the iso-
coumarin nucleus, were found weakly active or inactive on a-chymotrypsin, trypsin, Caspase-3 and HIV-protease, but prevent
g-secretase-mediated production of Ab 40/42 amyloid peptides, which is known to be involved in Alzheimer’s disease. Moreover,
the most active compounds on b-amyloid peptide production [JLK6 (5d), JLK2 (5f) and JLK7 (6f)] show only weak or moderate
inhibitory activity on the 20S proteasome. The obtained results suggest that the described new isocoumarin analogues could be of
interest, since compounds like JLK6 (5d), JLK2 (5f) and JLK7 (6f) can be considered as possible hits for the development of new
agents directed towards Alzheimer’s disease.
# 2003 Elsevier Science Ltd. All rights reserved.
Introduction
the non-peptidic scaffolds coumarin and isocoumarin
which occur naturally in many food sources, including
citrus fruits, herbs, vegetables or isolated from bacteria or
marine species have demonstrated promising protease
inhibitory properties.3ꢀ5 Schematically, protease inactiva-
tion by these coumarin or isocoumarin nuclei seems to
result from a nucleophilic attack on coumarin or iso-
coumarin lactone carbonyl groupby the active serine, or
cysteine residues constitutive of the enzyme active sites.6
In the case of aspartyl protease, inactivation could be
due to the binding of the coumarin lactone carbonyl
group to the protease aspartate residues. Depending on
the nature of the substituents borne by the coumarinic
and isocoumarinic rings, the resulting derivatives may
Structure-based design of peptidomimetic peptidase inhi-
bitors has proven to be highly successful for developing
new drugs as illustrated by the development of HIV pro-
tease inhibitors.1 Recent advances have led to protease
inhibitors of reduced peptidic nature, which generally
exhibit improved pharmacokinetic properties (oral bio-
availability, excretion) and constitute often the main
obstacles to drug development.2 Among these inhibitors,
*Corresponding author. Tel.: +33-491-82-91-41; fax: +33-491-82-94-
16; e-mail: kraus@luminy.univ-mrs.fr
0968-0896/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0968-0896(03)00235-9

3142
F. Bihel et al. / Bioorg. Med. Chem. 11 (2003) 3141–3152
act as general or specific protease inhibitors. Considerable
number of papers dealing with discrimination between
different classes of proteases (a-chymotrypsin, human
leukocyte elastase, . . .) by coumarin or by isocoumarin
derivatives have been published.6ꢀ11
Chemistry
All the compounds described are based on a 4-chloro-
isocoumarin scaffold. These compounds were synthe-
sized as described in Scheme 2. 4-Nitrohomophthalic
acid 2 was obtained by regioselective nitration of com-
mercially available homophthalic acid in 88% yield. The
monoesterification of 2 was achieved by a kinetic
approach involving the use of a primary alcohol both
reagent and solvent in the presence of a catalytic
amount of concentrated sulphuric acid, leading to
2-alkoxy-carbonylmethylbenzoic acid 3a–f. The mixture
including compound 2 and the selected alcohol was
heated until complete disappearance of the diacid. The
reaction was then quenched before the formation of the
corresponding diester (TLC analysis control) by addi-
tion of a large amount of bicarbonate solution. The
esterification rate was dependent on the nature of the
substituent at the 4-position of the aromatic ring, as the
4-nitro groupdecreased the esterification rate. Isolation
and purification of intermediates 3a–f required a careful
work up described in the experimental part, the yields
ranging from 70 to 98%.
We have recently reported new promising isocoumarin
derivatives JLK6 (5d), JLK2 (5f) and JLK7 (6f) as
potential inhibitors of amyloid peptide production,
which is the major component of the senile plaques
associated with the progression of Alzheimer’s neuro-
degenerative disease.12 Indeed, isocoumarin compounds
JLK6 (5d), JLK2 (5f) and JLK7 (6f) appear to inhibit
the Amyloid Protein Precursor (APP) cleavage medi-
ated by the g-secretase, enzyme considered as a member
of aspartyl protease family.13 Moreover these com-
pounds represent the first non-peptidic inhibitors able to
prevent g-secretase cleavage of b-APP without affecting
processing of mÁEnotch or endoproteolysis of pre-
senilins as shown in Scheme 1.14 Furthermore, Esler et
al. have shown that while the isocoumarin analogues
lower Ab production in cell culture, these compounds
do not inhibit isolated g-secretase activity in a solubi-
lized membrane preparation.15 All these results seem to
indicate that the isocoumarin analogues are not
g-secretase inhibitors but we do not know if the com-
pounds may be upstream of g-secretase as enzyme inhi-
bitors or by another yet undefined mechanism.
The cyclization of derivatives 3a–f was achieved by
using PCl5 refluxed in toluene,16 leading to 3-alkoxy-4-
chloro-isocoumarin derivatives 4a–f in yields ranging
from 30 to 70%. Quantitative reduction of 7-nitro-iso-
coumarin 4d–f into the corresponding amino 5a–f was
performed using a catalytic amount of palladium on
activated carbon under atmospheric pressure of hydro-
gen in THF. Compound 6f resulted from the condensa-
tion of 1-adamantylfluoroformate on the corresponding
7-amino-isocoumarin derivative 5f in the presence of
Et3N in THF. Due to the low nucleophilicity of the
aniline function, this condensation required 3 days at
room temperature (increased temperature led to pro-
duct decomposition) in 50% yield. Alternatively, the use
of 1-adamantylchloroformate or the use of 1-adamantol
in the presence of phosgene gave the expected com-
pound in lower yields. It should be also underlined the
high chemical reactivity of the lactone function in com-
pounds which incorporated at the 7-position an unsub-
stituted amino group. Indeed when a solution of 5f in
DMSO, not completely anhydrous, is left standing at
room temperature for few days, we observed the
hydrolysis of lactone group, as shown by the loss of UV
It was therefore of interest to investigate the inhibitory
properties of new isocoumarin compounds including
JLK6 (5d), JLK2 (5f) and JLK7 (6f) on representative
classes of proteases: aspartyl proteases (HIV-protease);
serine proteases (a-chymotrypsin and trypsin); cysteine
protease (Caspase-3). The selectivity of these new iso-
coumarin derivatives in their inhibitory activity was
further tested on the proteasome in vivo. The 20S pro-
teasome is an intracellular barrel shape proteinase com-
plex involved in cytosolic proteolysis and displaying
three main catalytic activities defined as chymotrypsin-
like, trypsin-like and peptidyl–glutamyl–peptide hydro-
lyzing enzyme (PGPH). The detection of an inhibitory
activity of the isocoumarin compounds on central cel-
lular proteases such as the proteasome, would impair
significantly the clinical development of these drugs as
anti-neurodegenerative agents.
Scheme 2. (i) Fuming nitric acid, 2 h, rt; (ii) R1OH, [H2SO4], heat; (iii)
PCl5, toluene, reflux, 15 h; (iv) H2/Pd/C, THF, rt; (v) adamantyl
fluoroformate, TEA, THF, rt, 3 days.
Scheme 1. APP and Notch secretase proteolysis, according to Wolfe
et al.14

F. Bihel et al. / Bioorg. Med. Chem. 11 (2003) 3141–3152
3143
absorption at 360 nm, which is a characteristic absorp-
tion for the isocoumarin nucleus. The proposed
mechanism for the opening of the isocoumarin nucleus
is the formation of an a-chloro-ester which led to an
unstable quinonimide methide compound. Due to the
highly electrophilicity of the carbon in a-position of the
ester group, the quinonimide methide compound gave,
in presence of water, the a-ketoester 8 characterized by
LC-MS (Scheme 3).17 In contrast, the stability of ana-
logue 6f in which the 7-amino groupis substituted by
the adamantyl groupthrough a carbamate function, is
increased because the electron doublet of the nitrogen
atom is less engaged in the isocoumarin nucleus con-
jugation (Scheme 4).
pharmacological agents to influence presenilin endo-
proteolysis and Notch cleavage. In addition, we have
evaluated the activity and the selectivity of the newly
synthesized isocoumarin derivatives on representative
proteases (serine, aspartyl or cysteine). Isocoumarin
derivatives were tested with the serine proteases,
a-chymotrypsin and trypsin, the HIV aspartyl protease
and a cysteine protease, Caspase-3. In several experi-
ments, enzymes were exposed to the novel isocoumarins
using the preincubation method, in which the enzyme
and the inhibitors are pre-incubated prior determination
of the remaining enzyme activity at intervals of time.
Inactivation potencies versus proteases were obtained
and expressed by the determination of second order rate
constants kobs=½IŠ Mꢀ1 minꢀ1, as depicted in Table 1.
Since coumarin nucleus and analogues are inhibitors of
various protease families, coumarin derivative 7 (Fig. 1)
was also tested on the different enzyme models for
comparison with the isocoumarin nucleus.18
Results and Discussion
At first, activity of the isocoumarin derivatives was
evaluated for their ability to affect ß-amyloid peptide
production in HEK293 cells overexpressing wild-type
bAPP. We have also assessed the potential of such
Aꢀ-Peptide production measurements
Total Ab secreted from HEK293 cells overexpressing
wild-type bAPP (962 fmol mLꢀ1 in 35-mm wells) was
markedly reduced by JLK6 (5d), JLK2 (5f) and JLK7
(6f) (Fig. 2a), whereas 4f (data not shown) were appar-
ently ineffective. A 100 mM concentration of inhibitors
achieved 70–80% inhibition of total Ab recovered (Fig.
2b). Interestingly, JLK6 (5d) and JLK2 (5f), but not 4f,
Scheme 3.
Figure 1. Structure of the coumarin analogue.
Table 1. Inhibitory activities of isocoumarin and coumarin compounds versus various proteases
Compd
R1
R2
Chymotrypsina
Trypsinb
HIV-proteasec
% inhibition
10 mM 50 mM 100 mM 50 mM 100 mM
Caspase-3d
[I] (mM) kobs=½IŠ (M-1 sꢀ1
)
[I] (mM) kobs=½IŠ (M-1 sꢀ1
)
4e
4f
4c
4d
5e
5f
5d
4a
4b
6f
7
NO2
NO2
NO2
NO2
NH2
NH2
NH2
H
OCH2CH2Br
OCH2CH2OCH3
OCH2CH2CN
OCH3
OCH2CH2Br
OCH2CH2OCH3
OCH3
0.5
1
1
1
10
10
100
100
100
260,000
1,400,000
307,000
247,000
33,500
37,800
2100
Inactive
Inactive
Inactive
Inactive
Inactive
Inactive
Inactive
2650
3280
Inactive
50% (10 mM)
—
55
—
—
—
—
—
—
—
35
50
—
79
—
—
21
—
32
—
58
85
—
47
100
51
57
77
57
70
20
84
—
17
—
—
—
32
77
48
51
24
Inactive
23
26
—
—
68
89
25
20
OCH2CH2I
OCH2CH2CN
16,300
9200
830
100
100
H
Adamantyl-CO–NH OCH2CH2OCH3 100
Coumarin analogue
100
—
50% (10 mM)
Inactive
—, not determinated; rate constants were obtained as described in Materials and methods.
aConditions were as followed: 0.1 M Hepes, 0.5 M NaCl, pH 6 at 25 ꢁC.
b0.1 M phosphate, pH 8 at 25 ꢁC.
c0.05 M NaOAc, 0.2 M NaCl, 0.005 M DTT, pH 4.9, 10% (v/v) glycerol at 37 ꢁC.
d0.02 M Hepes, pH 7.4, 0.1% CHAPS, 0.005 M DTT, 0.002 M EDTA at 37 ꢁC.

3144
F. Bihel et al. / Bioorg. Med. Chem. 11 (2003) 3141–3152
Scheme 4.
Figure 2. Effect of JLK inhibitors on bAPP maturation in HEK293
cells expressing wild-type bAPP. (a) Stably transfected HEK293 cells
overexpressing wild-type bAPP (see Methods) were incubated for 7 h
with 1% dimethylsulfoxide (control) or the indicated JLK inhibitors at
100 or 10 mM. Total Ab was immunoprecipitated from conditioned
media using FCA18 antibody and analysed on a 16.5% Tris–tricine
gel. (b) Histograms show quantitative densitometric analyses of total
Ab levels secreted by HEK293 cells expressing wild-type bAPP; data
are expressed as percentages of control secretion in the absence of JLK
inhibitors. As far as the diagrams result directly from SDS-Page,
standard-deviations (SD) do not appear on the diagrams.
active isocoumarin inhibitors do not affect a-secretase.
Although the possibility that isocoumarin inhibitors can
affect secretory processes cannot be definitely ruled out,
these results indicate that isocoumarin agents may
directly target g-secretase cleavage of bAPP.
As previously reported, HEK293 cells revealed the
expected endogenous N-terminal (NPS1; Fig. 4a) and
C-terminal (CPS1; Fig. 4b) fragments of presenilin 1
potentiated the recovery of two products with relative
molecular masses of ꢂ12,000–14,000 (Mr 12–14 K; Fig.
3a). Previous immunological characterization of such
fragments indicated that the upper-Mr band was label-
led with FCA18,19 a specific antibody against the Asp1
residue of Ab, whereas the lower-Mr fragment was
not.20 The presence of these two products, which
probably correspond to the b- and a-secretase-derived
C-terminal stubs, respectively, was increased by active
isocoumarin inhibitors (Fig. 3a). It should be noted
that most intracellular Ab-related species were not
detected by FCA18, and therefore probably correspond
to N-terminally truncated fragments of Ab (data not
shown).
Inhibitors did not affect the expression of wild-type
bAPP (Fig. 2a). Moreover, APPa, the a-secretase-
derived bAPP catabolite, was not affected by any of the
isocoumarin inhibitors (Fig. 3b). This indicates that
Figure 4. Effect of JLK inhibitors and DFK167 on endoproteolysis of
endogenous or overexpressed presenilins. a, b, HEK293 cells were
incubated for 7 h without inhibitor (control) or with the indicated
inhibitor (JLK at 100 mM; MW167 at 200 mM). Endoproteolysis of
endogenous presenilin 1 was monitored by western blotting for N-
terminal (NPS1) (a) and C-terminal (CPS1) (b) fragments using Ab444
and Ab111 antibodies, respectively (see Methods). No endogenous
holoprotein was detected. (c) Stably transfected HEK293 cells over-
expressing wild-type presenilin 2 were incubated with the indicated
JLK inhibitors as in (a) and (b). Overexpressed presenilin 2 holopro-
tein (Mr ꢂ56 K) and CPS2 (Mr ꢂ22 K) were detected by western
blotting with Ab333 antibody.
Figure 3. Effect of JLK inhibitors on b-, a-secretase products and
APPa. a- and b-secretase-derived C-terminal fragments (a) were
revealed using BR188 antibody. Secreted APPa was immunoprecipi-
tated with 207 antibody and was detected by western blotting with
10D5C antibody (b) (see Methods).

F. Bihel et al. / Bioorg. Med. Chem. 11 (2003) 3141–3152
3145
without holoprotein-like immunoreactivity (data not
ꢁ-Chymotrypsin (ꢁ-CT)
shown). Figure 4a and b clearly shows that isocoumarin
inhibitor did not affect proteolytic cleavage of endo-
genous presenilin 1. MW167, an inhibitor of Ab pro-
duction did not influence endoproteolysis of
endogenous presenilin 1 (Fig. 4a and b).21 We also
examined the putative effect of inhibitors JLK6 (5d) and
JLK2 (5f) on the hydrolysis of presenilin 2 in stably
transfected cells. As expected, a strong immunor-
eactivity corresponding to CPS2 was observed, together
with PS2 holoprotein-like immunoreactivity (Fig. 4c).
Again, isocoumarin agents were unable to affect pre-
senilin 2 proteolysis (Fig. 4c).
It is noteworthy that the presence of a nitro group as sub-
stituent at the 7-position of the isocoumarin nucleus
greatly enhances the inactivation rate constant magnitude
comparatively to other groups like NH2, N-adamantyl-
oxycarbonyl or H, whatever the nature of the substituent
at the 3-position of the isocoumarin ring. For example,
compound 4f the most potent inhibitor of a-CT inhibitor
ðkobs=½IŠ ¼ 1;400;000Þ is about 40 times more potent than
its homologue JLK2 (5f) ðkobs=½IŠ ¼ 37;800Þ and more
than three orders of magnitude than its analogue JLK7
(6f). It can be also observed that inside the series of nitro
analogues, inactivation rates also depend on both nature
and length of the side arm located at the 3-position of
the isocoumarin nucleus; for example, analogue 4e
bearing a 3-(2-bromoethoxy) substituent is 5 times less
efficient than its corresponding analogue 4f bearing a 3-
(2-methoxyethoxy) substituent and as effective as ana-
logue 4d bearing a 3-methoxy substituent. These results
suggest that the substitution at the 7-position of the
isocoumarin nucleus appears to be determinative for
these compounds to be processed by a-chymotrypsin.
Compounds bearing electron-withdrawing groups at the
7-position such as NO2 are highly preferentially pro-
cessed compared to the corresponding analogues bear-
ing electron-donating substituents at the 7-position like
NH2. All these structural observations are in agreement
with the known primary specificity of a-CT related to
the presence of an hydrophobic pocket,8 in which the 7-
nitro and 3-(2-methoxyethoxy) substituents of the com-
pound 4f give the best fit with the S1 site of a-chymo-
trypsin. Interestingly, it should be noted that compound
JLK7 (6f) bearing a N-adamantyloxycarbonyl amino
substituent at the 7-position of the nucleus, which has
been reported as inhibitor of the production of b-amyl-
oid peptides, displays only a very weak inhibitory
activity on a-CT ðkobs=½IŠ ¼ 830Þ.
As stated above, some lines of evidence indicate that
presenilins may contribute either directly or indirectly
to proteolysis of Notch, as well as of bAPP. Particu-
larly, invalidation of presenilin genes leads to full
abolishment of the generation of NICD.22,23 We have
established stably transfected cells lines overexpressing
Myc-tagged mÁEnotch-1. As previously described,24
these cells produce three bands that exhibit Myc
immunoreactivity (Fig. 5a). The intermediate is derived
from initiation at methionine 1727 and was unaffected
by any of our cell treatments. The low-Mr protein
migrated as overexpressed control NICD (Fig. 5a).
Pretreatment of these cells with MW167 effectively pre-
vented NICD production (by ꢂ80%, Fig. 5b), and its
mÁENotch precursor was thereby strongly protected
(Fig. 5a). However, none of the isocoumarin inhibitors
that affected Ab secretion had an effect on NICD gen-
eration (Fig. 5a and b).
Trypsin
The most potent a-CT inhibitors (4c–f) failed to efficiently
inhibit trypsin. Only compounds (4a, 4b) which are not
substituted at the 7-position display a weak or moderate
inhibitory potency on trypsin. Second-order kinetic rate
ðkobs=½IŠÞ for these two compounds were respectively 2600
and 3200. These values are 3 and 8 times weaker than the
values observed in the case of a-CT for these analogues. It
can be also underlined that the three Ab secretion inhibi-
tors JLK6 (5d), JLK2 (5f) and JLK 7 (6f) were devoided of
inhibitory efficiency on this enzyme.
Comparatively to the isocoumarin nucleus, coumarin
analogue 7 tested on both serine protease enzyme mod-
els (a-CT and trypsin), failed to demonstrate significant
inhibitory activity and did not displayed any selectivity
between the two serine proteases.
Figure 5. Effect of JLK inhibitors and MW167 on NICD generation
by HEK293 cells overexpressing Myc-tagged mÁENotch. (a) Cells
overexpressing Myc-tagged mÁENotch were incubated without inhi-
bitor (control) or with JLK inhibitors or MW167 as described in Fig-
ure 4. Intact mÁENotch and NICD were detected by western blotting
with 9E10 antibody. The rightmost lane corresponds to HEK293 cells
transiently transfected with 2 mg of NICD cDNA and analysed 72 h
after transfection for their intracellular NICD content by western
blotting with 9E10. (b) Histograms showing quantitative densitometric
analyses of NICD content; data are expressed as percentages of con-
trol values. As far as the diagrams result directly from SDS–PAGE,
standard-deviations (SD) do not appear on the diagrams.
HIV-aspartate protease
As already reported, 4-hydroxycoumarin rings have
been used in the design of non-peptidic HIV protease
inhibitors.25 Indeed it has been shown that salient

3146
F. Bihel et al. / Bioorg. Med. Chem. 11 (2003) 3141–3152
hydrogen bonding interactions occurring on the one
hand between the lactone oxygen atoms of the coumarin
with the NH groups of the flap region isoleucine resi-
dues 50=500 of the HIV protease and on the other hand
between the 4-hydroxy groupof the coumarin ring and
the catalytic aspartate residues 25=250 of the enzyme to
which the 4-hydroxy groupis speudosymmetrically
bonded. Some of these leads display anti-HIV protease
activity at submicromolar concentrations. As illustrated
in Table 1, the newly synthesized isocoumarin deriva-
tives display weak inhibitory activities on recombinant
HIV protease, IC50 ranging from 10 to 100 mM. Prob-
ably, the presence of the lipophilic chlorine at the 4-
position of the isocoumarin ring is not favourable for
the binding to the catalytic aspartate residues 25=250
comparatively to the 4-hydroxyl substituent of the cou-
marin nucleus. The most active compounds of this new
series was derivative JLK7 (6f) (IC50 ꢂ10 mM). Similar
inhibitory activity was also obtained with the coumarin
derivative 7 (IC50=10 mM). Besides no significant dif-
ference in inhibitory activity was found whatever the
nature of the substituents incorporated at the 3- or 7-
positions on the isocoumarin ring. These low inhibitory
anti-HIV protease potencies clearly indicate that the fit
of these isocoumarin rings onto the surface of the active
site of the HIV-protease is not optimised.
Individual compounds were incubated with cells for 12 h
prior visualization. Interestingly, the high anti-chymo-
trypsin activity of compounds 4c and 4e did not corre-
late with proteasome inhibition, while the related
product, 4d, displayed a strong proteasome inhibition
equivalent to the one induced by epoxomicin. Two of
the inhibitors, 5e and JLK2 (5f), were slightly cytotoxic
for HeLa cells at the concentration of 10 mM and were
therefore tested at 1 mM. At this low concentration,
JLK2 (5f) had a weak inhibitory effect on the ubiquitin–
proteasome pathway (not shown). The other two iso-
coumarin derivatives, JLK6 (5d) and JLK7 (6f), able to
block Ab secretion, were also found to have some pro-
teasome inhibitory effect. The inhibition was stronger
(proteasome re-localization) for JLK6 (5d) than for
JLK7 (6f), although the magnitude of these effects was
much weaker than the one induced by the specific pro-
teasome inhibitor epoxomicin or even by compound 4d.
General anti-protease isocoumarin derivatives activity
We have described a new generation of non-peptidic
inhibitors that have distinct effects on Ab secretion and
Notch cleavage. Furthermore, the fact that an identical
protective effect was observed on the C-terminal bAPP
fragments C99 and C83, which are generated by b- and
a-secretase, respectively, indicates that these inhibitors
[JLK6 (5d), JLK2 (5f) and JLK7 (6f)] probably target g-
secretase-mediated cleavage. However, these inhibitors
do not affect production of NICD in identical cells,
although we confirmed that this cleavage can be blocked
by MW167.32 The ability of isocoumarin inhibitors to
discriminate between Ab production and Notch clea-
vage was recently confirmed by Esler et al., which have
proved that isocoumarin analogues JLK 2 (5f) and
JLK6 (5d) do not block Ab production by directly
inhibiting the g-secretase enzyme but their observations
strongly suggest that these inhibitors reduce Ab released
from cells by interacting with an unknown target
upstream of g-secretase.15
Caspase-3
The activity of isocoumarin compounds was also eval-
uated on the enzyme Caspase-3, which belongs to the
cysteine protease family.26 As shown in Table 1, all
compounds inhibited only moderately Caspase-3. Their
inhibitory potencies were around one order magnitude
less active than the standard Caspase-3 peptidic inhi-
bitor (Acetyl-Asp-Glu-Val-Asp-CHO; IC50=1 mM).
Compound JLK7 (6f) is a very weak Caspase-3 inhi-
bitor, only 20% inhibition observed at 100 mM and
JLK2 (5f) was devoided of inhibitory efficiency on this
enzyme. This result is interesting because it has been
first reported that Caspases were involved in the pro-
teolytic cascade events leading to b-amyloid secretion,27
but recently it has been demonstrated that b-amyloid
peptide resulting from the cleavage pathway of its pre-
cursor APP is independent of its cleavage by Caspases,28
our results seem in agreement with this last report.
Coumarin derivative 7 was found inactive on Caspase-3.
However, other processes such as destabilization of the
reported protein complex associated with the g-secretase
cannot be ruled out.33 Since we have shown the unex-
pected properties of JLK6 (5d), JLK2 (5f) and JLK7
(6f) compounds on the inhibition of the b-amyloid pep-
tide secretion, and their possible use as new tools in the
understanding of Alzheimer’s disease, it was of interest
to explore the selectivity of this series of isocoumarin
derivatives on representative families of proteolytic
enzymes. First of all, it can be observed differential
behavior of isocoumarin derivatives towards the repre-
sentative proteolytic enzymes. Formation of the acyl-
enzyme intermediate which results from the attack of
the active serine residue of a-CT or trypsin on the lac-
tone carbonyl groupisocoumarin ring, and conse-
quently the nature of substituents at 3- and 7-position
induced the inhibition selectivity between the two serine
proteases. Second-order kinetic rate constants ðkobs=½IŠÞ
determined for a-CT and trypsin inactivation clearly
support these findings. The observed differences in inhi-
bition potencies between a-CT and trypsin give rise by
these derivatives can be explained by the structure of
Proteasome
Inhibition of the proteasome by inhibitors such as
epoxomicin or lactacystin greatly enhances the cytosolic
accumulation of poly-ubiquitinated proteins which are
the natural substrates of the proteasome.29 We therefore
evaluated the effect of different isocoumarin inhibitors
on the total protein ubiquitination levels in HeLa cells
by confocal microscopy using the FK2 monoclonal
antibody. FK2 is specific for mono- or poly-ubiquiti-
nated proteins and, importantly, does not react with
free ubiquitin (Ub).30 Interference of the isocoumarin
derivatives with proteasome activity lead to an increase
in FK2 staining and important re-localization together
with the proteasome in the perinuclear area (Fig. 6).31

F. Bihel et al. / Bioorg. Med. Chem. 11 (2003) 3141–3152
3147
Figure 6. Effect of different isocoumarin inhibitors on proteasomal activity. HeLa cells were treated for 12 h without (control) or with the different
inhibitors and as a positive control with epoxomicin, a selective inhibitor of proteasomal activity. Cells were fixed, permeabilized and stained for
proteasome with a rabbit polyclonal antibody (green) and for poly-ubiquitinated proteins (Ub-protein) using the FK2 monoclonal antibody (red).
Effects of the different inhibitors were visualized on a confocal microscope by an increase in FK2 staining and an important re-localization together
with the proteasome in the perinuclear area (arrows).
their hydrophobic S1 site (nomenclature of Schechter
and Berger).34 In trypsin, the base of S1 contains the
negatively charged carboxylate side chain of Asp-189
which is not the case for the S1 a-CT hydrophobic
pocket. This structural S1 difference accounts for the
specificity of these two enzymes. In contrast, in the case
of the other classes of protease (HIV-aspartyl protease
or cysteine protease Caspase-3), no significant inhibi-
tory selectivity was obtained with this series of com-
pounds. Interestingly, high chymotrypsin inhibition

3148
F. Bihel et al. / Bioorg. Med. Chem. 11 (2003) 3141–3152
potency was not correlated with proteasome inhibition
in vivo, suggesting a different enzymatic selectivity on
this proteolytic complex. The weak proteasome inhibi-
tion observed with derivatives JLK6 (5d), JLK2 (5f) and
JLK7 (6f) cannot be precisely quantified in our rela-
tively crude assay, however this effect could have a
profound significance for ß-amyloid peptide secretion.
A correlation of the inhibition of the chymotrypsin-like
activity of the proteasome with suppression of ß-amy-
loid peptide secretion has been described,35 however the
mechanism by which the proteasome exert this effect
remain unclear.
mixtures as eluent. Preparative flash column chromato-
graphies were carried out on silica gel (230–240 mesh,
G60 Merck). Analytical purity was assessed by LC-MS
using a Waters 2790 system. The stationary phase was a
Waters Symmetry C-18 column (2.1 ꢃ 50 mm). The
mobile phase employed 0.1% formic acid with acetoni-
trile as the organic modifier and a flow rate of 0.3 mL/
min. Low resolution mass spectra were recorded on a
Micromass ZMD mass spectrometer. Mass spectra were
acquired in either the positive or negative ion mode
under electrospray ionization (ESI).
4-Nitrohomophthalic acid (2). To a solution of fuming
nitric acid (20 mL) at 0 ꢁC was added homophthalic acid
(5 g, 31 mmol). The mixture was stirred at room tem-
perature for 6 h. The solution was poured onto ice and
filtered. The resulting solid was dried under reduced
pressure to give 4-nitrohomophthalic acid 2 (3.88 g,
88% yield). Rf (methanol/methylene chloride [1:1]) 0.60;
Nevertheless, it should be pointed out that in the case of
the inactivation of proteases involved in the secretion of
amyloid peptide, the presence of an amino group at the
7-position of the isocoumarin ring fulfilled the criteria
for suicide-type inactivation at least for compounds 5d
and 5f. Indeed after the nucleophilic attack of the car-
bonyl groupleading to the acyl-enzyme, the presence of
the electron doublet of the amino groupat the 7-posi-
tion allows a new possible enzyme alkylation reaction,
consequently after the departure of the chlorine at the
position 4 of the isocoumarin ring (Scheme 3). Never-
theless, this mechanism cannot be validated for different
reasons: isocoumarin target is not identified, purified
and structurally resolved, and also in compound 6f
delocalization of the amino lone pair electron is exclu-
ded.
mp217 ꢁC; H NMR (250 MHz, DMSO) d 8.43 (d, 1H,
1
J=2.4 Hz), 8.18 (dd, 1H, J=2.4, 8.4 Hz), 7.48 (d, 1H,
J=8.4 Hz), 3.93 (s, 2H); FAB-MS (GT/FAB+)
(M+H)+ 226.
2-(2-Iodo-ethoxycarbonylmethyl)-benzoic
acid
(3a).
Homophthalic acid (1 g, 5.55 mmol) was dissolved in 2-
iodoethanol (2.5 mL) and sulphuric acid (95%, 100 mL,
1.84 mmol) was added to the solution. The mixture was
stirred at 70 ꢁC for 7 min. EtOAc (10 mL) was added to
the mixture and the solution was extracted with a
NaHCO3 solution (5%, 3 ꢃ 10 mL). The aqueous layer
was acidified by a citric solution upto pH 3. This solu-
tion was extracted with EtOAc (3 ꢃ 10 mL), dried over
MgSO4 and the solvent was removed in vacuum to give
The present results bring to light that isocoumarin
compounds such as JLK6 (5d), JLK2 (5f) and JLK7
(6f), which have demonstrated potent anti-secretive
activity of ß-amyloid peptide through the inactivation
of protease, show moderate or weak inhibitory activities
on the whole different tested protease models, including
the 20S proteasome. This specificity could be of interest
in the case of some of these isocoumarin derivatives
which are considered as possible hits for the develop-
ment of new anti-neurodegenerative drugs.
2-(2-iodo-ethoxycarbonylmethyl)-benzoic
acid
(1.32 g, 71% yield). Rf (methanol/methylene chloride
3a
[1:9]) 0.70; mp87 ꢁC; H NMR (250 MHz, CDCl3) d
1
8.09 (d, 1H, J=6.7 Hz), 7.49 (t, 1H, J=6.7 Hz), 7.35 (t,
1H, J=6.6 Hz), 7.23 (d, 1H, J=7.3 Hz), 4.30 (t, 2H,
J=6.9 Hz) 4.00 (s, 2H), 3.23 (t, 2H, J=6.9 Hz); 13C
NMR (60 MHz, CDCl3) d 172.91, 171.28, 137.20,
134.69, 133.70, 131.86, 130.11, 128.21, 70.76, 41.63,
ꢀ0.28; FAB-MS (NOBA/FAB+) (M+H)+ 335. The
following compounds 3b–f were prepared as described
previously.
Experimental
Synthesis
All common chemicals and solvents were reagent grade
or better. The purity of each compound was checked by
1H and 13C NMR spectroscopy, mass spectroscopy, IR
spectroscopy, thin-layer chromatography, melting
point, and elemental analysis and results are consistent
with the proposed structures. The 1H NMR spectra
were recorded on either a Bruker AC-250 MHz instru-
ment, and 13C NMR spectra, on a Bruker AC-60 MHz
instrument. Fast atom bombardment mass spectra
(FAB) were recorded on a Jeol DX-100 spectrometer
using a caesium ion source and glycerol/thioglycerol
(GT) (1:1) or m-nitrobenzylalcohol (NOBA) as matrix.
Elemental analysis were performed by Service Central
d’Analyses du CNRS (Venaison, France) and were
within 0.4% of the theoretical values. Thin-layer chro-
matographic analysis (TLC) were conducted on silica
gel plates 0.2 mm thick (60F254 Merck) with various
2-(2-Cyano-ethoxycarbonylmethyl)-benzoic acid (3b).
(70% yield) Rf (methanol/methylene chloride [1:9])
ꢁ
0.58; mp94 C; 1H NMR (250 MHz, CDCl3) d 8.10 (dd,
1H, J=1.3, 7.7 Hz), 7.50 (td, 1H, J=1.4, 7.5 Hz), 7.37
(td, 1H, J=1.1, 7.5 Hz), 7.23 (d, 1H, J=7.5 Hz), 4.25 (t,
2H, J=6.3 Hz) 4.02 (s, 2H), 2.65 (t, 2H, J=6.3 Hz); 13C
NMR (60 MHz, CDCl3) d 172.81, 171.88, 136.96,
134.34, 133.39, 132.86, 129.11, 128.71, 117.73, 59.76,
41.46, 18.69; FAB-MS (GT/FAB-) (MꢀH) 232.
2-(2-Cyano-ethoxycarbonylmethyl)-5-nitro-benzoic acid
(3c). (98% yield)ꢁ Rf (methanol/methylene chloride
1
[1:9]) 0.65; mp142 C; H NMR (250 MHz, CD3OD) d
8.99 (d, 1H, J=2.5 Hz), 8.52 (dd, 1H, J=2.5, 8.4 Hz),
7.77 (d, 1H, J=8.4 Hz), 4.38 (s, 2H), 3.91 (t, 2H,
J=6.1 Hz), 2.78 (t, 2H, J=6.1 Hz); 13C NMR (60 MHz,
CD3OD) d 173.90, 168.41, 149.57, 145.90, 135.11,

F. Bihel et al. / Bioorg. Med. Chem. 11 (2003) 3141–3152
3149
133.26, 128.17, 127.85, 117.79, 59.26, 41.12, 18.19;
FAB-MS (NOBA/FAB+) (M+H)+ 277.
elemental analysis calcd (%) for C12H8ClNO3: C 57.73,
H 3.23, N 5.61; found C 57.75, H 3.22, N 5.64.
2-Methoxycarbonylmethyl-5-nitro-benzoic acid (3d).
(76% yield) Rf (ethyl acetate/acetic acid [14.5:0.5]) 0.57;
3-(4-Chloro-7-nitro-1-oxo-1H-isochromen-3-yloxy)-pro-
pionitrile (4c). (36% yield) Rf (ethyl acetate/cyclohex-
ane [4:12]) 0.92; mp150 ꢁC; 1H NMR (250 MHz,
CDCl3) d 8.99 (d, 1H, J=2.3 Hz), 8.52 (dd, 1H, J=2.3,
8.9 Hz), 7.84 (d, 1H, J=8.9 Hz), 4.63 (t, 2H, J=6.1 Hz),
2.91 (t, 2H, J=6.1 Hz); 13C NMR (60 MHz, CDCl3) d
157.43, 155.81, 149.93, 143.37, 129.61, 126.93, 123.72,
117.29, 116.33, 93.35, 65.21, 19.17; FAB-MS (NOBA/
FAB+) (M+H)+295; elemental analysis calcd (%) for
C12H7ClN2O5: C 48.92, H 2.39, N 9.51; found C 48.89,
H 2.37, N 9.58.
ꢁ
mp158 C; 1H NMR (250 MHz, CD3OD) d 8.71 (d, 1H,
J=2.4 Hz), 8.24 (dd, 1H, J=2.4, 8.4 Hz), 7.48 (d, 1H,
J=8.4 Hz), 4.07 (s, 2H), 3.57 (s, 3H); 13C NMR
(60 MHz, CD3OD) d 173.80, 168.91, 149.27, 145.50,
135.91, 133.86, 128.27, 127.65, 53.43, 41.82; FAB-MS
(GT/FAB+) (M+H)+ 240.
2-(2-Bromo-ethoxycarbonylmethyl)-5-nitro-benzoic acid
(3e). (73% yield) Rf (methanol/methylene chloride
[2:8]) 0.69; mp90 ꢁC; H NMR (250 MHz, CDCl3) d
1
8.78 (d, 1H, J=2.6 Hz), 8.21 (dd, 1H, J=2.6, 8.4 Hz),
7.39 (d, 1H, J=8.4 Hz), 4.11 (s, 2H), 4.31 (t, 2H,
J=6.0 Hz), 3.39 (t, 2H, J=6.0 Hz); 13C NMR (60 MHz,
CD3OD) d 173.20, 168.11, 149.77, 145.20, 135.71,
133.16, 127.97, 127.15, 70.31, 41.22, 28.99; FAB-MS
(GT/FAB+) (M+H+) 332.
4-Chloro-3-methoxy-7-nitro-isocoumarin (4d). (70%
yield) Rf (ethyl acetate/hexane [3:12]) 0.57; mp115 ꢁC;
1H NMR (250 MHz, CDCl3) d 9.20 (d, 1H, J=2.3 Hz),
8.69 (dd, 1H, J=2.3, 8.9 Hz), 7.99 (d, 1H, J=8.9 Hz),
4.33 (s, 3H); 13C NMR (60 MHz, CD3OD) d 157.42,
156.19, 145.12, 143.16, 129.89, 126.41, 123.53, 116.59,
89.65, 57.41; FAB-MS (NOBA/FAB+) (M+H)+ 256;
elemental analysis calcd (%) for C10H6ClNO5: C 46.99,
H 2.37, N 5.48; found C 46.96, H 2.38, N 5.45.
2-(2-Methoxy-ethoxycarbonylmethyl)-5-nitro-benzoic acid
(3f). (87% yield) Rf (methanol/methylene chloride
[2:14]) 0.70; mp55 ꢁC; H NMR (250 MHz, CD3OD) d
1
8.99 (d, 1H, J=2.5 Hz), 8.52 (dd, 1H, J=2.5, 8.4 Hz),
7.77 (d, 1H, J=8.4 Hz), 4.38 (s, 2H), 4.38 (t, 2H,
J=4.6 Hz ), 3.76 (t, 2H, J=4.6 Hz), 3.52 (s, 3H); 13C
NMR (60 MHz, CD3OD) d 174.10, 168.31, 149.67,
145.10, 135.81, 133.19, 127.37, 127.23, 70.10, 69.80,
59.87, 41.42; FAB-MS (NOBA/FAB+) (M+H)+ 284.
3-(2-Bromo-ethoxy)-4-chloro-7-nitro-isocoumarin (4e).
(44% yield) Rf (ethyl acetate/cyclohexane [5:12]) 0.31;
mp120 ꢁC; H NMR (250 MHz, CDCl3) d 8.97 (d, 1H,
1
J=2.3 Hz), 8.48 (dd, 1H, J=2.4, 8.9 Hz), 7.80 (d, 1H,
J=8.9 Hz), 4.69 (t, 2H, J=6.1 Hz), 3.63 (t, 2H,
J=6.0 Hz); 13C NMR (60 MHz, CDCl3) d 157.57,
145.94, 143.20, 130.29, 127.12, 124.28, 118.07, 117.60,
91.26, 70.01, 28.09; FAB-MS (NOBA/FAB+)
(M+H)+ 349; elemental analysis calcd (%) for
C11H7BrClNO5: C 37.91, H 2.02, N 4.02; found C
37.94, H 2.04, N 3.99.
4-Chloro-3-(2-iodo-ethoxy)-isocoumarin (4a). 3a (1.23 g,
3.68 mmol) and phosphorus pentachloride (1.7 g,
8.3 mmol) were added to toluene (20 mL). The mixture
was refluxed for 15 h. The solvent was removed under
reduced pressure and ice was poured onto the residue.
The solution was refluxed for 1 h and the mixture was
extracted with EtOAc (20 mL) and the organic layer
was washed with brine (3 ꢃ 10 mL), dried over
MgSO4, and the solvent was removed in vacuum. The
residue was recrystallized from methanol to give
4-chloro-3-(2-iodo-ethoxy)-isocoumarin 4a (380 mg,
30% yield). Rf (ethyl acetate/cyclohexane [6:12]) 0.68;
4-Chloro-3-(2-methoxy-ethoxy)-7-nitro-isocoumarin (4f).
(69% yield) Rf (ethyl acetate/methylene chloride/hex-
ane [7:3:12]) 0.37; mp127 ꢁC; 1H NMR (250 MHz,
CDCl3) d 8.95 (d, 1H, J=2.3 Hz), 8.45 (dd, 1H, J=2.4,
8.9 Hz), 7.78 (d, 1H, J=8.9 Hz), 4.54 (t, 2H, J=4.4 Hz),
3.71 (t, 2H, J=4.4 Hz), 3.37 (s, 3H); 13C NMR
(60 MHz, CDCl3) d 157.51, 155.82, 149.98, 143.14,
129.80, 126.43, 123.76, 117.03, 90.45, 70.51, 70.09,
59.31; FAB-MS (NOBA/FAB+) (M+H)+ 300; ele-
mental analysis calcd (%) for C12H10ClNO6: C 48.10, H
3.36, N 4.67; found C 48.20, H 3.35, N 4.66.
mp67 ꢁC; H NMR (250 MHz, CDCl3) d 8.32 (d, 1H,
1
J=7.9 Hz), 7.86 (m, 2H), 7.54 (td, 1H, J=1.6, 8.1 Hz),
4.72 (t, 2H, J=7.0 Hz), 3.56 (t, 2H, J=7.0 Hz); 13C
NMR (60 MHz, CDCl3) d 159.32, 152.17, 137.52,
135.86, 130.25, 126.76, 122.64, 117.62, 92.17, 70.33,
ꢀ0.34; FAB-MS (NOBA/FAB+) (M+H)+ 347; ele-
mental analysis calcd (%) for C11H8ClIO3: C 37.69, H
2.30; found C 37.68, H 2.30. The following com-
pounds 4b–f were prepared according to the same
procedure.
7-Amino-4-chloro-3-methoxy-isocoumarin (5d). 4d (2 g,
7 mmol) was dissolved in THF (30 mL) and palladium
on activated carbon (10% Pd, 50 mg) was added to the
solution, and the reaction mixture was stirred under
atmospheric pressure of hydrogen at room temperature
for 6 h, filtered and concentrated in vacuo to give a
residue which was recrystallized from ethanol, to give 7-
3-(4-Chloro-1-oxo-1H-isochromen-3-yloxy)-propionitrile
(4b). (67% yield) Rf (ethyl acetate/cyclohexane [7:16])
amino-4-chloro-3-methoxy-isocoumarin
5d
(1.6 g,
ꢁ
0.36 ; mp115 C; 1H NMR (250 MHz, CDCl3) d 8.20 (d,
7 mmol, quantitative yield) as a yellow powder. Rf (ethyl
acetate/hexane [7:12]) 0.56; mp115 ꢁC; 1H NMR
(250 MHz, CDCl3) d 9.10 (d, 1H, J=2.3 Hz), 8.69 (dd,
1H, J=2.3, 8.9 Hz), 7.99 (d, 1H, J=8.9 Hz), 4.33 (s,
3H), 3.57 (s, 2H); 13C NMR (60 MHz, CD3OD) d
157.43, 156.12, 145.17, 141.31, 129.85, 126.48, 123.56,
1H, J=7.5 Hz), 7.75 (m, 2H), 7.44 (dd, 1H, J=1.4,
8.1 Hz), 4.57 (t, 2H, J=6.3 Hz), 2.90 (t, 2H, J=6.3 Hz);
13C NMR (60 MHz, CDCl3) d 159.93, 152.46, 138.01,
136.73, 131.03, 127.82, 123.59, 118.44, 116.93, 93.55,
65.29, 19.57; FAB-MS (NOBA/FAB+) (M+H)+ 250;

3150
F. Bihel et al. / Bioorg. Med. Chem. 11 (2003) 3141–3152
116.58, 89.66, 57.41; FAB-MS (NOBA/FAB+)
(M+H)+ 226; elemental analysis calcd (%) for
C10H8ClNO3: C 53.23, H 3.57, N 6.21; found C 53.26,
H 3.54, N 6.24. The following compounds 5e–f were
prepared according to the same procedure.
type bAPP were incubated for 7 h in the presence of
various inhibitor concentrations. Media were collected,
diluted in 1/10 RIPA 10ꢃ buffer and incubated over-
night with a 200-fold dilution of FCA18 antibody for
total Ab. After incubation for a further 3 h with protein
A-sepharose (Zymed) and subsequent centrifugation, pel-
leted proteins were submitted to 16.5% Tris–tricine gels
and western blotted for 45min. Nitrocellulose membranes
were heated in boiling PBS for 5 min and capped with 5%
skimmed milk in PBS containing 0.05% Tween 20 for 1 h.
Membranes were then incubated overnight with WO2
antibody and blots were visualized by enhanced chemilu-
minescence (ECL; Amersham).
7-Amino-4-chloro-3-(2-bromo-ethoxy)-isocoumarin (5e).
(Quantitative yield) Rf (ethyl acetate/hexane [4:6]) 0.35;
mp110 ꢁC; H NMR (250 MHz, CDCl3) d 8.18 (d, 1H,
1
J=8.3 Hz), 7.89 (d, 1H, J=2.7 Hz), 7.25 (dd, 1H,
J=2.8, 8.3 Hz), 4.68 (t, 2H, J=4.3 Hz), 3.67 (t, 2H,
J=4.4 Hz); 13C NMR (60 MHz, CDCl3) d 157.42,
155.91, 149.89, 141.28, 129.97, 126.13, 123.64, 117.60,
91.26, 70.01, 28.09; FAB-MS (NOBA/FAB+)
(M+H)+ 318; elemental analysis calcd (%) for
C11H9BrClNO3: C 41.47, H 2.85, N 4.40; found C
41.50, H 2.82, N 4.46.
Immunoprecipitation of sAPPꢁ. Conditioned media
from HEK293 cells overexpressing wild-type bAPP751
were incubated with a 700-fold dilution of 207 antibody
(against the N-termini of bAPP and sAPPa) in the pre-
sence of pansorbin (Calbiochem) as described.20 After
centrifugation, pellets were resuspended with loading
buffer, subjected to 8% Tris–glycine electrophoresis and
western blotted for 2 h. Nitrocellulose membranes were
incubated overnight with a 500-fold dilution of 10D5C
antibody and blots were visualized as described above.
7-Amino-4-chloro-3-(2-methoxy-ethoxy)-isocoumarin (5f).
(Quantitative yield) Rf (ethyl acetate/hexane [9:12]) 0.39;
mp127 ꢁC; H NMR (250MHz, CDCl3) d 7.48 (d, 1H,
1
J=8.6 Hz), 7.38 (d, 1H, J=2.4 Hz), 7.05 (dd, 1H, J=2.4,
8.6 Hz), 4.39 (t, 2H, J=4.5 Hz), 3.68 (t, 2H, J=4.5 Hz),
3.38 (s, 3H); 13C NMR (60MHz, CDCl3) d 157.51, 155.83,
149.92, 141.37, 129.94, 126.33, 123.77, 117.04, 90.48, 70.53,
70.04, 59.31; FAB-MS (NOBA/FAB+) (M+H)+ 270;
elemental analysis calcd (%) for C12H12ClNO4: C 53.44, H
4.49, N 5.19; found C 53.45, H 4.48, N 5.18.
Detection of ꢀAPP and C-terminal fragments. Cells
were lysed in a 50 mM Tris buffer containing 150 mM
NaCl, 5 mM EDTA, 0.5% Triton X-100 and 0.5%
deoxycholate, and 30 mg of proteins were resuspended in
the loading buffer, electrophoresed on 8% SDS-poly-
acrylamide gel and western blotted for 2 h. Nitrocellulose
sheets were incubated in skimmed milk and exposed
overnight to a 2000-fold dilution of WO2 (bAPP); blots
were visualized as described above. C83 and C99 frag-
ments from the same lysates were analyzed by 12% SDS-
polyacrylamide gels, western blotted and then incubated
with BR188; blots were visualized using with horseradish
peroxidase (HRP)-conjugated goat-anti rabbit antibody
as described.20
[4-Chloro-3-(2-methoxy-ethoxy)-1-oxo-1H-isochromen-
7-yl]-carbamic acid adamantan-1-yl ester (6f). 5f (50 mg,
0.22 mmol) was dissolved in THF (1.25 mL) and 1-ada-
mantyl fluoroformate (225 mg, 2.1 mmol) and triethyla-
mine (140 mL, 1 mmol) were added to this solution. The
reaction mixture was stirred at room temperature for 3
days. The solvent was removed under reduced pressure
and the residue was dissolved in EtOAc (10 mL). The
organic layer was washed with a NaHCO3 solution
(5%, 2 ꢃ 10 mL), brine (2 ꢃ 10 mL), and a citric acid
solution (5%, 2 ꢃ 10 mL), dried over MgSO4 and the
solvent was removed in vacuo to give a residue which
was purified by flash chromatography on silica gel,
using EtOAc–cyclohexane [6:12], to give [4-chloro-3-(2-
methoxy-ethoxy)-1-oxo-1H-isochromen-7-yl]-carbamic
acid adamantan-1-yl ester 6f (50 mg, 50% yield) as a
yellow powder. Rf (ethyl acetate/hexane [6:12]) 0.59; mp
140 ꢁC; 1H NMR (250 MHz, CDCl3) d 7.93 (d, 1H,
J=2.2 Hz), 7.73 (dd, 1H, J=8.7 Hz), 7.47 (d, 1H,
J=8.7 Hz), 6.59 (s, 1H), 4.31 (t, 2H, J=4.5 Hz), 3.58 (t,
2H, J=4.5 Hz), 3.27 (s, 3H), 2.00 (bs, 9H), 1.51 (bs,
6H); 13C NMR (60 MHz, CDCl3) d 157.52, 155.87,
154.85, 149.91, 139.39, 129.94, 126.36, 123.78, 117.00,
90.48, 70.51, 70.00, 65.86, 59.34, 40.82, 37.82, 22.61,
21.65; FAB-MS (NOBA/FAB+) (M+H)+ 449; ele-
mental analysis calcd (%) for C23H26ClNO6: C 61.67, H
5.85, N 3.13; found C 61.69, H 5.86, N 3.12.
Transfection of mDEnotch-1 and analysis of clones.
HEK293 cells were cotransfected by DAC30 (Euro-
gentec, Brussels, Belgium) with 2 mg of Myc-tagged
mÁEnotch-1-pCS2 complementary DNA and 0.3 mg of
pZeoSV2 cDNA. Zeocin-resistant clones were washed
with PBS and lysed, and ꢂ30 mg of proteins were elec-
trophoresed on large 8% SDS–polyacrylamide gels and
western blotted for 2 h. Stable transfectants were identi-
fied after exposure of membranes to a 2000-fold dilution
of 9E10 (anti-Myc) antibody, anti-mouse immunoglobu-
lin G coupled to HRP ð1=5000Þ, and ECL.
Transient transfection of NICD. Mock HEK293 cells
were transiently transfected with 2 mg of cDNA encod-
ing Myc-tagged NICD. After 2 days, NICD expression
in cell lysates was detected as described above. This cel-
lular extract was used as standard control to estimate
NICD production in mÁEnotch-expressing cells by
SDS–PAGE.
7-Benzyloxy-3-(6-benzyloxy-benzo[1,3]dioxol-5-yl)-chro-
men-2-one (7). Synthesis of coumarin derivative 7 has
already been reported.18
Detection of presenilins and their endoproteolytic frag-
ments. HEK293 cells expressing endogenous presenilins
or overexpressing wild-type presenilin 2 were lysed and
Inhibitor treatment of cells and Aꢀ immunoprecipitation.
Stably transfected HEK293 cells overexpressing wild-

F. Bihel et al. / Bioorg. Med. Chem. 11 (2003) 3141–3152
3151
50 mg of proteins were subjected to 12% SDS–PAGE
and western blotted for 1.5 h. Nitrocellulose sheets were
incubated in skimmed milk and exposed overnight to
0.2–1.0 mg mL-1 Ab444 (for NPS1), Ab111 (for CPS1)
or Ab333 (for CPS2).
Ala-Ala-Pro-Phe-pnitroanilide in Hepes buffer (0.1 M
Hepes, 0.5 M NaCl, pH 6) and N-a-Benzoyl-l-Arginine-
pnitroanilide in 0.1 M phosphate buffer (pH 8), respec-
tively. kobs=½IŠ values were determined in the presence of
substrate as described by Tian and Tsou.37 For exam-
ple, inactivation of chymotrypsin (0.147 mM) by 4f
(0,5 mM) in the presence of 52 mM N-succinyl-Ala-Ala-
Pro-Phe-pnitroanilide was measured by addition of a
70 mL of the enzyme solution to a mixture of substrate
and inhibitor containing 2.5% DMSO. The increase in
absorbance was monitored (410 nm) until no further
release of p-nitroanilide was observed. The kobs=½IŠ
values were calculated from plots of log ([P]ꢀ[P]t) ver-
sus time, where [P] and [P]t are the concentrations of p-
nitroanilide after total inactivation and at time t,
respectively. The reaction rates were based on two
determinations, and the standard errors are ꢄ10% or
less.
Antibodies. The 207 antibody recognizes the N-termini
of bAPP and sAPPa. WO2 interacts with a sequence of
five to eight amino acids in Ab. BR188 was raised
against the 12 C-terminal residues of bAPP, and BR188
recognizes whole bAPP as well as its a- and b-secretase-
derived fragments. 10D5C is directed against the C ter-
minus of APPa. FCA18 specifically interacts with free
Asp1 residues of all Ab species, but does not interact
with N-terminally truncated Ab-related fragments.21
FCA3340 and FCA3542 specifically label Ab40 and
Ab42, respectively.19
Detection of poly-ubiquitinated proteins and proteasome
Immunofluorescence confocal microscopy with Leica
TCS 4D was performed as described.36 mAb FK2 and
anti-proteasome were obtained from Affinity (Mam-
head, UK). All secondary antibodies were from Immu-
notech (Marseille, France). For pharmacological
treatments, HeLa cells grown in DMEM with 5% FCS
were incubated respectively with 1 mM epoxomicin, or
10 mM of the tested compounds.
Acknowledgements
Ministere de la Recherche et de la Technologie (ACI
Molecules et Cibles Therapeutiques, France 2001) are
greatly acknowledged for financial support, and we are
appreciative of the scientific contribution offered by
Trophos SA (Parc Scientifique de Luminy, Marseille,
France). We thank R. Kopan for mÁNotch-1 and
NICD constructs, and the following for antibodies: T.
Hartmann and K. Beyreuther for WO2, M. Goedert for
BR188, L. Pradier for 9E10, M. Savage for 207, D.
Schenk for 10D5C, and T. Tabira and W. Araki for
Ab111, Ab333 and Ab444.
Enzyme inhibition kinetics—materials and methods
Chymotrypsin, was purchased from ICN Biomedicals,
Ohio. Trypsin and Caspase-3 were purchased from
Sigma Chemical Co., St Louis, MO. HIV-protease was
purchased from Bachem, USA.
References and Notes
Enzyme inhibition. For the HIV-protease inhibition
assay, the enzyme was preincubated for 10 min at 37 ꢁC
with inhibitor by adding an aliquot (10 mL) of the inhi-
bitor in DMSO to the enzyme (20 mL of a 0.294 mM
solution) in buffer (370 mL, 0.05 M NaOAc, pH 4.9,
0.2 M NaCl, 0.005 M DTT, 10% (v/v) glycerol). Per-
centage of inhibition was spectrophotometrically mea-
sured (l=300 nm) on a Uvikon 930 spectrophotometer
by adding an aliquot of the substrate solution (20 mL of
a 573 mM solution of H-His-Lys-Ala-Arg-Val-Leu-
pNO2Phe-Glu-Ala-Nle-Ser-NH2) to the enzyme-inhi-
bitor solution into a cuvette.
1. Leung, D.; Abbenante, G.; Fairlie, D. P. J. Med. Chem.
2000, 43, 305.
2. Turner, S. R.; Strohbach, J. W.; Tommasi, R. A.; Aristoff,
P. A.; Johnson, P. D.; Skulnick, H. I.; Dolak, L. A.; Seest,
E. P.; Tomich, P. K.; Bohanan, M. J.; Horng, M. M.; Lynn,
J. C.; Chong, K. T.; Hinshaw, R. R.; Watenpaugh, K. D.;
Janakiraman, M. N.; Thaisrivongs, S. J. Med. Chem. 1998, 41,
3467.
3. Murray R., Mendez J., Brown S., In The Natural Coumar-
ins: Occurrence, Chemistry and Biochemistry; R. Murray, J.
Mendez, S. Brown, eds.; John Wiley & Sons: New York, 1982;
p97.
4. Kang, S. Y.; Lee, K. Y.; Sung, S. H.; Park, M. J.; Kim,
Y. C. J. Nat. Prod. 2001, 64, 683.
For the Caspase-3 inhibition assay, the enzyme was
preincubated for 5 min at 37 ꢁC with inhibitor by adding
an aliquot (10 mL) of the inhibitor in DMSO to the
enzyme (10 mL of a 2 mg/mL solution) in buffer (970 mL,
0.02 M Hepes, pH 7.4, 0.1% CHAPS, 0.005 M DTT,
0.002 M EDTA). Percentages of inhibition were spec-
trophotometrically measured (l=405 nm) by adding an
aliquot of the substrate solution (10 mL of a 20 mM
solution of Acetyl-Asp-Glu-Val-Asp-p-nitroanilide) to
the enzyme-inhibitor solution into a cuvette.
5. Thrash, T. P.; Welton, T. D.; Behar, V. Tetrahedron Lett.
2000, 41, 29.
6. Kam, C.-M.; Kerrigan, J. E.; Plaskon, R. R.; Duffy, E. J.;
Lollar, P.; Suddath, F. L.; Powers, J. C. J. Med. Chem. 1994,
37, 1298.
7. Doucet, C.; Pochet, L.; Thierry, N.; Pirotte, B.; Delarge, J.;
Reboud-Ravaux, M. J. Med. Chem. 1999, 42, 4161.
8. Pochet, L.; Doucet, C.; Schynts, M.; Thierry, N.; Boggetto,
N.; Pirotte, B.; Jiang, K. Y.; Masereel, B.; de Tullio, P.;
Delarge, J.; Reboud-Ravaux, M. J. Med. Chem. 1996, 39,
2579.
9. Harper, J. W.; Powers, J. C. J. Am. Chem. Soc. 1984, 106,
7618.
10. Harper, J. W.; Hemmi, K.; Powers, J. C. J. Am. Chem.
Soc. 1983, 105, 6518.
Determination of inactivation rates in the presence of
substrate. Progress curve method. For the chymotrypsin
and trypsin inhibition assays, substrates are N-succinyl-

3152
F. Bihel et al. / Bioorg. Med. Chem. 11 (2003) 3141–3152
11. Hernandez, M. A.; Powers, J. C.; Glinski, J.; Oleksyszyn,
J.; Vijayalakshmi, J.; Meyer, E. F. J. Med. Chem. 1992, 35,
1121.
12. Petit, A.; Bihel, F.; Alves da Costa, C.; Pourquie, O.;
Checler, F.; Kraus, J.-L. Nat. Cell Biol. 2001, 3, 507.
13. Wolfe, M. S. J. Med. Chem. 2001, 44, 2039.
14. Kimberly, W. T.; Esler, W. P.; Ye, W.; Ostaszewski, B. L.;
Gao, J.; Diehl, T.; Selkoe, D. J.; Wolfe, M. S. Biochemistry
2003, 42, 137.
15. Esler, W. P.; Das, C.; Campbell, W. A.; Kimberly, W. T.;
Kornilova, A. Y.; Diehl, T. S.; Ye, W.; Ostaszewski, B. L.;
Xia, W.; Selkoe, D. J.; Wolfe, M. S. Nat. Cell Biol. 2002, 4,
E110.
16. Tirodkar, R. B.; Usgaonkar, R. N. Indian J. Chem. 1969,
7, 1114.
Lynn, J. C.; Mulichak, A. M.; Spinelli, P. A.; Hinshaw,
R. R.; Pagano, P. J.; Moon, J. B.; Ruwart, M. J.;
Wilkinson, K. F.; Rush, B. D.; Zipp, G. L.; Dalga, R. J.;
Schwende, F. J.; Howard, G. M.; Padbury, G. E.; Toth,
L. N.; Zhao, Z.; Koeplinger, K. A.; Kakuk, T. J.; Cole,
S. L.; Zaya, R. M.; Piper, R. C.; Jeffrey, P. J. Med.
Chem. 1994, 37, 3200.
26. Sakahira, H.; Enari, M.; Nagata, S. Nature 1998, 391, 96.
27. Gervais, F. G.; Xu, D.; Robertson, G. S.; Vaillancourt,
J. P.; Zhu, Y.; Huang, J. Q.; LeBlanc, A.; Smith, D.; Rigby,
M.; Shearman, M. S.; Clarke, E. E.; Zheng, H.; Van Der
Ploeg, L. H. T.; Ruffolo, S. C.; Thornberry, N. A.; Xanthou-
dakis, S.; Zamboni, R. J.; Roy, S.; Nicholson, D. W. Cell
1999, 97, 395.
28. Soriano, S.; Lu, D. C.; Chandra, S.; Pietrzik, C. U.; Koo,
E. H. J. Biol. Chem. 2001, 276, 29045.
29. Hirsch, C.; Ploegh, H. L. Trends Cell Biol. 2000, 10, 268.
30. Fujimuro, M.; Sawada, H.; Yokosawa, H. FEBS Lett.
1994, 349, 173.
17. Lopez-Alvarado, P.; Avendano, C.; Menendez, J. C.
Synthesis-Stuttgart 1998, 186.
18. Pinard, E.; Gaudry, M.; Henot, F.; Thellend, A. Tetra-
hedron Lett. 1998, 39, 2739.
19. Barelli, H.; Lebeau, A.; Vizzavona, J.; Delaere, P.; Che-
vallier, N.; Drouot, C.; Marambaud, P.; Ancolio, K.; Bux-
baum, J. D.; Khorkova, O.; Heroux, J.; Sahasrabudhe, S.;
Martinez, J.; Warter, J. M.; Mohr, M.; Checler, F. Mol. Med.
1997, 3, 695.
20. Ancolio, K.; Dumanchin, C.; Barelli, H.; Warter, J. M.;
Brice, A.; Campion, D.; Frebourg, T.; Checler, F. Proc. Natl.
Acad. Sci. U.S.A. 1999, 96, 4119.
31. Wojcik, C.; Schroeter, D.; Wilk, S.; Lamprecht, J.; Pawe-
letz, N. Eur. J. Cell Biol. 1996, 71, 311.
32. De Strooper, B.; Annaert, W.; Cupers, P.; Saftig, P.;
Craessaerts, K.; Mumm, J. S.; Schroeter, E. H.; Schrijvers, V.;
Wolfe, M. S.; Ray, W. J.; Goate, A.; Kopan, R. Nature 1999,
398, 518.
33. Checler, F. Mol. Neurobiol. 1999, 19, 255.
34. Schechter, J.; Berger, A. Biochem. Biophys. Res. Commun.
1967, 27, 157.
35. Christie, G.; Markwell, R. E.; Gray, C. W.; Smith, L.;
Godfrey, F.; Mansfield, F.; Wadsworth, H.; King, R.;
McLaughlin, M.; Cooper, D. G.; Ward, R. V.; Howlett, D. R.;
Hartmann, T.; Lichtenthaler, S. F.; Beyreuther, K.; Under-
wood, J.; Gribble, S. K.; Cappai, R.; Masters, C. L.;
Tamaoka, A.; Gardner, R. L.; Rivett, A. J.; Karran, E. H.;
Allsop, D. J. Neurochem. 1999, 73, 195.
21. Wolfe, M. S.; Citron, M.; Diehl, T. S.; Xia, W.; Donkor,
I. O.; Selkoe, D. J. J. Med. Chem. 1998, 41, 6.
22. Herreman, A.; Serneels, L.; Annaert, W.; Collen, D.;
Schoonjans, L.; De Strooper, B. Nat. Cell Biol. 2000, 2, 461.
23. Zhang, Z.; Nadeau, P.; Song, W.; Donoviel, D.; Yuan,
M.; Bernstein, A.; Yankner, B. A. Nat. Cell Biol. 2000, 2, 463.
24. Kopan, R.; Schroeter, E. H.; Weintraub, H.; Nye, J. S.
Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 1683.
25. Thaisrivongs, S.; Tomich, P. K.; Watenpaugh, K. D.;
Chong, K.-T.; Howe, W. J.; Yang, C.-P.; Strohbach,
J. W.; Turner, S. R.; McGrath, J. P.; Bohanon, M. J.;
36. Lelouard, H.; Gatti, E.; Cappello, F.; Gresser, O.;
Camosseto, V.; Pierre, P. Nature 2002, 417, 177.
37. Tian, W. X.; Tsou, C. L. Biochemistry 1982, 21, 1028.