Thienylhalomethylketones: Irreversible glycogen synthase kinase 3 inhibitors as useful pharmacological tools

Article abstract of DOI:10.1016/j.bmc.2009.08.042

Thienylhalomethylketones, whose chemical, biological, and pharmaceutical data are here reported, are the first irreversible inhibitors of GSK-3β described to date. Their inhibitory activity is likely related to the cysteine residue present in the ATP-binding site, which is proposed as a relevant residue for modulation of GSK-3 activity. The good cell permeability of the compounds allows them to be used in different cell models. Overall, the results presented here support the potential use of halomethylketones as pharmacological tools for the study of GSK-3β functions and suggest a new mechanism for GSK-3β inhibition that may be considered for further drug design.

Full text of DOI:10.1016/j.bmc.2009.08.042

Bioorganic & Medicinal Chemistry 17 (2009) 6914–6925
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Thienylhalomethylketones: Irreversible glycogen synthase kinase 3 inhibitors
as useful pharmacological tools
Daniel I. Perez , Santiago Conde , Concepción Pérez , Carmen Gil , Diana Simon b,c, Francisco Wandosell b,c
a
a
a
a
,
c
d
e
a,
*
Francisco J. Moreno , José L. Gelpí , Francisco J. Luque , Ana Martínez
Instituto de Química Médica-CSIC, Juan de la Cierva 3, 28006 Madrid, Spain
Centro de Investigación Biomédica en Red sobre Enfermedades Neurodegenerativas (CIBERNED), Spain
Centro de Biología Molecular ‘Severo Ochoa’, CSIC-UAM, Universidad Autonoma de Madrid, 28049 Madrid, Spain
Departamento de Bioquímica y Biología Molecular, Facultad de Biología, Universidad de Barcelona, Avda. Diagonal 645, 08028 Barcelona, Spain
Departamento de Fisicoquímica e Instituto de Biomedicina (IBUB), Facultad de Farmacia, Universidad de Barcelona, Avda. Diagonal 643, 08028 Barcelona, Spain
a
b
c
d
e
a r t i c l e i n f o
a b s t r a c t
Article history:
Thienylhalomethylketones, whose chemical, biological, and pharmaceutical data are here reported, are
the first irreversible inhibitors of GSK-3b described to date. Their inhibitory activity is likely related to
the cysteine residue present in the ATP-binding site, which is proposed as a relevant residue for modu-
lation of GSK-3 activity. The good cell permeability of the compounds allows them to be used in different
cell models. Overall, the results presented here support the potential use of halomethylketones as phar-
macological tools for the study of GSK-3b functions and suggest a new mechanism for GSK-3b inhibition
that may be considered for further drug design.
Received 1 April 2009
Revised 21 July 2009
Accepted 13 August 2009
Available online 22 August 2009
Keywords:
GSK-3b inhibitors
Alzheimer’s disease
Halomethylketones
Tau phosphorylation
Ó 2009 Elsevier Ltd. All rights reserved.
1
. Introduction
Human cells contain 491 different protein kinases, which are
is a serine/threonine kinase involved in many cellular signaling
pathways. Although it has been known for more than 30 years
ago for its participation in glycogen synthesis, it has recently
emerged as one of the most fascinating targets for medicinal chem-
responsible for phosphorylating certain amino acids. These en-
zymes are critically involved in a wide variety of cellular processes,
including metabolism, transcription, and cell movement. They are
linked to a number of diseases, making them important drug tar-
8
,9
ists. Today it is well recognized that GSK-3b over-activity is re-
lated to several unmet diseases such as type II diabetes, chronic
inflammatory diseases and AD. Consequently, its specific inhibitors
could be powerful drugs for the effective treatments of such severe
1
gets. Nearly all kinase inhibitors developed so far target the bind-
0
2
10,11
ing site for the nucleotide adenosine 5 -triphosphate (ATP), which
has a similar structure in all protein kinases and makes kinase
inhibitors unspecific. However, the search for inhibitors of protein
kinases has recently culminated in the registration of several com-
pounds such as imatinib (gleevec), or erlotinib (tarceva), for cer-
tain cancer therapies. The challenge now is to develop kinase
inhibitors for other therapeutic indications where an urgent need
for effective treatment is required.5 Neurodegenerative diseases
are one of such indications.
Alzheimer’s disease (AD) is the most common cause of demen-
tia in adults. The current therapy for AD has only moderate efficacy
in the palliative treatment of the disease, but an effective therapeu-
tic approach is not available yet.6 Recent studies have suggested
the central role of glycogen synthase kinase 3b (GSK-3b) in the
pathogenesis of both sporadic and familial forms of AD.7 GSK-3b
pathologies.
In AD, the over-activity and/or over-expression of GSK-3b ac-
counts for memory impairment, tau hyperphosphorylation, in-
creased b-amyloid production and local plaque-associated
microglial-mediated inflammatory responses which are character-
3
4
7
istic hallmarks of the disease. As a causal mediator of AD, inhibi-
tors of GSK-3b would provide a novel avenue for therapeutic
12
intervention in this devastating disorder. Currently, only one
compound with this mechanism of action is in clinical trials for
AD. The thiadiazolidindione (TDZD) candidate called NP-12 has
completed phase I clinical trials with single and multiple doses
studies and phase II studies for AD has now started in Europe. A
pivotal study on progressive supranuclear palsy, where abnormal
hyperphosphorylation of tau in the brain plays a vital role in its
13
molecular pathogenesis, is also planned.
Our research group has been involved for many years in dis-
ease-modifying therapeutic approaches for the treatment of
*
Corresponding author. Tel.: +34 91 5854624; fax: +34 91 5644853.
E-mail address: amartinez@iqm.csic.es (A. Martínez).
1
4,15
AD.
One of our most active programs is focused on the search,
0
968-0896/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2009.08.042

D. I. Perez et al. / Bioorg. Med. Chem. 17 (2009) 6914–6925
6915
design, synthesis, and evaluation of GSK-3 inhibitors, which led to
the first ATP non-competitive GSK-3b inhibitors known to date.
Based on this encouraging finding, a second set of 13 com-
pounds was selected from our library and those bearing a thio-
phene or phenyl ring with a carbonyl moiety directly attached
were further tested. The highest activities were found for the group
of chloromethylhalothienyl ketones (chloroacetylhalothiophenes
19–21; Fig. 2), while non-halogenated alkyl ketones, esters, and
aldehydes did not inhibit GSK-3b at the maximum concentration
1
6
In this manuscript chemical, biological, and pharmaceutical data
about thienylhalomethylketones are reported. The results support
the use of these compounds as useful pharmacological tools with
good cell permeability to gain insight into the role played by
GSK-3b in physiological and pathological processes.
used (100 lM). These data suggested that the chloroacetylthioph-
ene moiety could be a structural feature required for activity, and it
was selected as a chemical scaffold for lead optimization.
2
2
. Results and discussion
.1. Lead discovery
2
.2. Lead optimization and SAR
Our first hit in this series of compounds was discovered search-
ing for GSK-3 inhibitors in our compound library. Preliminary re-
sults together with structure–activity relationships studies of this
new family of compounds were published in a previous work.
Herein we report the synthesis of new compounds (37–44) and
additional experiments carried out for the HMK family.
To explore the structure–activity relationships (SAR) on this
new class of GSK-3b inhibitors, a variety of 2- or 3-substituted
halomethylthienylketones (Scheme 1) were synthesized.
1
7
We first examined the effect of some substituents, mainly alkyl
or halide, on the thiophene heterocycle. These compounds,
although previously known, were here synthesized by chloroacet-
ylation of the corresponding thiophene following the Friedel–
Crafts method (compounds 22–36; Scheme 1) to be evaluated as
GSK-3 inhibitors. In some cases, when 2.1 equivalents of chloroace-
tyl chloride and aluminum trichloride were used, mixtures of
mono- and di-chloroacylated compounds were obtained (com-
pounds 22–28). Isolation was then performed by chromatographic
techniques and the chemical structure unequivocally elucidated by
their spectroscopic properties. GSK-3b inhibitory activities of all
the 15-substituted chloromethyl thienyl ketones prepared are pre-
sented in Table 1.
In the previous work, the initial search was focused on eight
unrelated chemical structures (compounds 1–8, Fig. 1). Commer-
cially available GSK-3b was incubated with ATP and the small pep-
tide GS-1 (based on residues of glycogen synthase phosphorylated
1
8
by GSK-3b) as substrate, in the presence and absence of the com-
pound, and the inhibitory activity was measured following a previ-
ously described method.18 GSK-3b activity was expressed as
picomoles of phosphate incorporated per 20 min of incubation or
as the percentage of maximal activity. All the compounds tested
were inactive or poorly active other than 2-chloroacetyl-4,5-
dichlorothiophene 8, which inhibited GSK-3b at a very low
concentration.
Besides the chloroacetyl group, a second substituent is required
to enhance activity, as noted in the comparison of the activities
measured for compound 22 versus the activity of 21 and 8. More-
over, this second substituent must be electron-withdrawing, such
as halide atoms or chloroacetyl groups, in order to maintain or in-
crease the inhibition of GSK-3b. Thus, when an electron-donating
group such as the methyl moiety is present (compounds 24, 26,
Ph
N
O
Ph
N
N
N
H
Ph
Ph
NH
N
S
Cl
2
7, 36), the inhibition of GSK-3b is very low or completely lost. This
Et
fact might suggest that the electron distribution on the molecular
surface should be important for the inhibitory activity, as also
1
2
1
9
noted for TDZDs derivatives.
IC50>100 µM
IC50>100 µM
To further examine the influence of the chloroacetyl moiety on
GSK-3b inhibition, derivatives 37, 38, and 43, in which the distance
between the carbonyl group and the halide atom is enlarged, were
synthesized. For the synthesis of these compounds 4-chlorobutyryl
chloride was used as acylating agent. In vitro evaluation showed a
complete lack of activity (Table 2).
Ph
N
Cl
NH .HCl
N
2
N
Ph
S
Me
Ph
3
4
Furthermore, the nature of the halide atom joined to the acetyl
moiety was also explored. Different substituted bromomethyl thi-
enyl ketones 39–42 were synthesized employing bromoacetyl bro-
mide, all of them being inhibitors of GSK-3b. While the change of
chlorine by bromine does not alter the IC50 (compounds 39–42),
substitution by an ester moiety (44) led to a drastic decrease in
inhibitory activity (Table 2), which points out that the halide atom
is crucial for activity. Compound 44 was synthesized when chloro-
acetylthiophene 35 reacted in the presence of sodium acetate and
sodium anhydride in acetic acid (Scheme 1). The substitution of the
halide atom in the chloroacetyl chain confirms the halomethylke-
tone (HMK) moiety as the structural feature required for activity.
Considering the HMK chemical feature key for GSK-3b inhibi-
IC50>100 µM
IC50>100 µM
Br
Br
Br
N
S
Cl
N
Ph
OH
O
5
6
IC50=75 µM
IC50>100 µM
Cl
Cl
N
Cl
S
Cl
tion and the Hammett inductive substituent constant (
can explain the lack of activity for methyl-substituted thiophenes
= ꢀ0.17) and the activity of chloro-, bromo- and chloroacetyl-
substituted thiophenes ( = 0.23, 0.23, and 0.50, respectively) as
I
r ), we
NO
O
7
8a
IC50=2.5 µM
(r
I
IC50>100 µM
r
I
a consequence of their electronic influence on the reactivity of
HMK moiety directly attached to the aromatic ring.
Figure 1. First compounds used for the GSK-3b inhibition screening with their
experimental IC50 values. aRef. 17.

6
916
D. I. Perez et al. / Bioorg. Med. Chem. 17 (2009) 6914–6925
O
O
O
Br
Br
MeO
MeO
Me
Me
OEt
H
S
Cl
O
O
9
10
11
12
13
IC50>100 µM
IC50>100 µM IC50>100 µM
IC50>100 µM
IC50>100µM
OH
Me
Me
OH
H
O
OEt
O
OEt
OEt
S
AcS
S
S
S
O
O
1
4
15
16
17
IC50>100 µM
IC50=25 µM
IC50>100 µM
IC50>100 µM
O
O
O
Cl
Cl
Cl
Cl
Cl
Br
Br
Br
Cl
Cl
Br
S
Br
S
S
S
O
1
8a
IC50=25 µM
19a
IC50=2 µM
20a
IC50=1 µM
21a
IC50=3 µM
Figure 2. Focused library used in discovery of GSK-3b inhibitors with their experimental IC50 values. aRef. 17.
Table 1
a)
a
Inhibition of GSK-3 by chloromethyl thienyl ketones
R1
R2
R1
R2
Compd
R
R1
R2
IC50 (lM)
i)
X
R
R1
R2
S
1
R
n
S
2
O
R, R , R = H,ClAc,
Me, Cl, Br
22-34, 37-42,
R
S
Cl
O
O
X
n
b)
R
22
23
H
H
Me
Me
H
H
ClAc
H
ClAc
H
Me
ClAc
H
H
Br
Br
H
H
50
i)
H
1.5
24
H
H
Me
H
Me
H
>100
5.0
75
>100
5.0
10
R
R
R
S
S
25
26
27
28
R=Cl n=1 X=Cl 35
R=Me n=1 X=Cl 36
R=Cl n=3 X=Cl 43
R = Me, Cl
H
H
29
Cl
Br
H
Br
Ac
H
Me
ii)
30
31
H
10
O
O
Br
H
H
0.5
1.0
50
3
3
3
2
3
4
O
Ac
H
8.0
R
S
R
O
R2
Cl
4
4
Scheme 1. Synthesis of compounds 22–44. Reagents and conditions: (i)
XCO(CH X, AlCl , CS , (ii) (AcO) O, NaOAc, AcOH,
R1
S
R
2
)
n
3
2
2
D.
3
3
5
6
Cl
Me
Cl
Me
H
H
5.0
>100
a
Ref. 17.
2
2
.3. Mechanism of GSK-3 inhibition and selectivity
.3.1. Kinetic studies on GSK-3b
To investigate the inhibitory mechanism of HMKs on GSK-3b
maximal rate (Vmax) decreases in the presence of the inhibitor, the
Michaelis–Menten constant (K ) remains unaltered. These results
would suggest that HMKs act as non-competitive inhibitors of
ATP binding, because an increase in the ATP concentration (from
and further confirm our previous results, several kinetic experi-
ments were performed using two different HMKs (compounds 19
and 23). Lineweaver–Burk plots of enzyme kinetics are shown in
Figure 3.
m
6.5 to 100 lM) does not interfere with enzymatic inhibition.
Kinetic experiments varying both ATP levels (6.5, 10, 15 25, 50,
and 100 lM) and HMKs (1 and 2 lM) were performed. Double-re-
ciprocal plotting of the data are depicted in Figure 3a. The intercept
We have also studied the substrate dependence of the kinase
activity in the presence of these inhibitors using the peptide GS-
1 as substrate. Kinetic experiments were performed varying the
of the plot in the vertical axis (1/V) rises when the HMK concentra-
concentrations of both GS-1 (6.5, 10, 15, 25, 50, and 100
inhibitors 19 and 23 (from 1 to 2 M), while the ATP concentration
was kept constant (15 M). Double-reciprocal plotting of the data
lM) and
tion increases (from 1 to 3
lM), whereas the intercept in the hor-
l
izontal axis (1/S) does not change, meaning that, while the enzyme
l

D. I. Perez et al. / Bioorg. Med. Chem. 17 (2009) 6914–6925
6917
Table 2
2.3.2. Kinase profiling
Inhibition of GSK-3 by substituted thiophenes
Selectivity of kinase inhibition is critical for pathway analysis
in vivo and in cellular systems. To examine the selectivity of these
compounds for GSK-3b, many of them (compounds 18, 19, 23, 25,
28–32, 34, 35, 39, and 40) were tested as inhibitors of serine/thre-
onine cAMP-dependent protein kinase (PKA). All of them remained
Compd
R
R1
R2
n
X
IC50 (lM)
R1
R2
R
S
n
X
inactive at the highest concentration (100 lM) used. Furthermore,
the same set of compounds was evaluated against a panel of eight
different serine/threonine and tyrosine kinases (CaM-K II, MAP-K,
EGFR-K, IR-K MeK1-K, Abl-K, PKp56, and Src). These kinases are in-
volved in very different signaling pathways implicated not only in
neurodegeneration (Cam-KII and MAP-K) but also in other physio-
logical processes and diseases as insulin signal pathway (IR-K) or
cancer (Abl-K, EGFR-K, MeK1-K, and Src). The inhibitory assays
O
3
3
3
4
4
4
7
8
9
0
1
2
Cl
Br
H
Br
Br
Br
H
H
Br
Br
H
H
H
Br
H
H
3
3
1
1
1
1
Cl
Cl
Br
Br
Br
Br
>100
>100
1.0
1.0
1.0
H
Br
1.0
were performed using a 10 lM concentration for the different
O
inhibitors. The results reported in Table 3 show that, in general,
HMKs do not exhibit a significant inhibitory effect on the whole
set of kinases, except in the proto-oncogenic Src, which is elevated
in some human tumors.
R2
R1
X
n
S
R
4
4
3
4
Cl
Cl
Cl
Cl
H
H
3
1
Cl
OCOMe
>100
>100
2.3.3. Neurotransmitter receptors binding
In order to determine the potential off-target effects of these
GSK-3b inhibitors, due to the chemical structure of our leads or
to the enzyme targeted, we evaluated the in vitro binding to differ-
ent neurotransmitter receptors. We selected
2
a-adrenergic (a ),
(
Fig. 3b), in which each point is the mean of three different exper-
dopaminergic (hD2 and hD3), glutamatergic (AMPA and NMDA),
muscarinic (hM), nicotinic (bungarotoxin sensitive and insensi-
tive), and serotoninergic (5-HT) receptors. Human recombinant
iments, suggest that HMKs act also as non-competitive inhibitors
of GS-1 binding.
Figure 3. Double-reciprocal plots of kinetic data from assays of GSK-3b activity at different concentrations of HMKs (19 and 23). (a) ATP concentrations in the reaction
mixture varied from 6.5 to 100 M. HMKs concentrations are depicted in the plot, and the concentration of GS-1 was kept constant at 15 M. V is picomoles of phosphate/
0 min. (b) GS-1 concentrations in the reaction mixture varied from 6.5 to 100 M. HMKs concentrations are depicted in the plot, and the concentration of ATP was kept
l
l
2
l
constant at 15 lM. V is picomoles of phosphate/20 min.

6
918
D. I. Perez et al. / Bioorg. Med. Chem. 17 (2009) 6914–6925
Table 3
Inhibitory activity (%inhibition)a exhibited by selected HMKs compounds for several kinases
Compd (10
lM)
Abl-K
CAM-K II
EGFR-K
IR-K
MAP-K
MEK 1 K
PK p56
Src-K
18
19
23
25
28
29
30
31
32
34
35
39
40
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
21
—
19
—
16
—
—
—
10
—
—
—
16
25
—
—
—
—
—
—
—
17
—
—
—
12
12
11
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
16
—
—
—
—
12
—
—
—
31
33
49
17
—
—
—
20
—
—
24
26
94
54
61
—
19
n.d
69
22
12
66
76
—
—
12
41
—
—
—
a
The symbol – indicates an inhibition of less than 10%. n.d. value not determined because of interference with the detection method due to compound autofluorescence.
ones were selected when available. The results (see Table 4) show
that HMKs do not bind to any of these biological targets, thus
reducing the potential secondary effects in vivo and pointing to a
selective mechanism of action.
This binding mode might not a priori justify the non-ATP com-
petitive nature of HMKs. However, the presence of the thiol group
of Cys199, lying close to the HMK in some of the structures docked
into the ATP-binding site suggests an alternative interpretation to
the mechanism of action of HMKs. Although cysteine residues
within active sites of kinases are relatively rare, several cases are
2
1,22
2
.3.4. Enzyme mapping
well known,
and the glutathionylation of cysteines present in
To gain deeper insight into the inhibitory mechanism of HMKs
to GSK-3b, a mapping study was performed to identify potential
the activation loop is a general regulation mechanism for protein
2
3
kinases. Moreover, recent studies revealed that a unique active
site cysteine (Cys199) is present in GSK-3b, which is replaced by
other residues in other kinases, such as alanine in cycline depen-
cavities suitable for binding. To this end, the docking module of
2
0
classical molecular interaction potential (CMIP) program was
used, taking advantage of the availability of X-ray crystallographic
data for GSK-3b. This computational approach has been previously
validated for several chemically diverse GSK-3 inhibitors whose
binding mode had been previously determined by solving the X-
ray structure of the complexes with GSK-3b.1 The CMIP analysis
was performed for the active HMK 32 and the inactive and chem-
ically related derivative 16. While there seems to be a rather
unspecific interaction with HMK 16, which could reflect its lack
of GSK-3b inhibition, a binding site is found for 32 near the ATP-
binding site (see Supplementary data). To further check this find-
ing, the interaction of 32 at the ATP-binding site was reexamined
by means of alternative docking studies performed with rDock.
In one of the two main clusters of poses, 32 is inserted in a highly
hydrophobic pocket (formed by residues Ile62, Val70, Ala83,
Met101, Val110, Leu132, Tyr134, Val 135, and Leu188) at the rear
of the ATP-binding site, and the carbon atom bearing the chlorine
is positioned above the sulfur atom of Cys199 (distance of 3.7 Å;
see Fig. 4).
2
4
dent kinase 2 (CDK-2; see Fig. 5).
Considering this unique structural motif on GSK-3 and the abil-
2
5
26
ity of peptidyl-HMKs to inhibit serine and cysteine proteases,
we propose a covalent interaction between Cys199 and HMKs in
their specific binding to GSK-3. This should consist in the forma-
tion of a thioether carbon–sulfur (C–S) bond and subsequent elim-
ination of the halide atom, which might be favored by the
stabilization of the leaving chloride anion by the protonated amino
group of Lys85 (Fig. 4). The covalent modification here proposed
may explain the observed inhibitory kinetics and kinase selectivity.
Among the kinases assayed, only Src was inhibited by HMKs. The
chemical feature critical for Src activity is a specific cysteine cluster
9
2
7
in the catalytic site. Therefore, a covalent modification by HMKs
could inhibit this tyrosine kinase.
Modification of this unique cysteine on the GSK-3 catalytic site
has been postulated previously, although not demonstrated, in
2
8
other thiol-reactive compounds. We propose here that Cys199
can be a key residue for modulation of GSK-3b activity and offers
Table 4
Inhibitory activity (%inhibition)a exhibited by selected HMKs compounds for several neurotransmitter receptors
Compd (10
lM)
a
2
H D2
H D3
AMPA
NMDA
M
N (aBGT insensitive)
N (
aBGT sensitive)
5 HT
18
19
23
25
28
29
30
31
32
34
35
39
40
—
—
10
—
—
—
—
—
—
—
—
—
—
13
31
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
12
—
—
—
—
—
—
—
25
—
—
17
—
—
—
—
17
—
—
16
13
—
—
10
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
11
—
—
—
—
—
—
—
—
—
11
—
—
—
—
17
—
22
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
12
—
—
a
The symbol – indicates an inhibition of less than 10%.

D. I. Perez et al. / Bioorg. Med. Chem. 17 (2009) 6914–6925
6919
covalently the specific Cys199. Further studies are in progress,
including directed mutagenesis to confirm these first results.
2
.3.5. Chemical reactivity of HMKs versus thiol groups
Halomethylketones have been classified as reactive chemical
groups. To study to some extent the potential reactivity of these
compounds, we analyzed the reaction between the HMK 29 and
mercaptoethanol. Reactions were carried out at room temperature
and the progress was followed by HPLC–MS. In a HMK 29/mercap-
toethanol (1:1) reaction mixture, no S-alkylated product was found
after 30 min. Only 7% of the sulfur derivative was detected when
the mixture was stirring for 24 h. When the concentration of
mercaptoethanol was increased 10-fold, 16% of the S-alkylated
product was formed after 18 h. However, in equimolar concentra-
tions of HMK 29 and mercaptoethanol, when a tertiary amine was
present in the reaction media, 75% of the S-alkylated product was
observed after the first 30 min. These results confirm that the reac-
tivity of our HMKs versus thiol groups is mediated by a basic envi-
ronment and explain the crucial role of Lys85 in the binding of our
HMKs to Cys199 in the ATP site of GSK-3b.
Figure 4. Representation of the putative binding mode of HMK 32 in the ATP-
binding site of GSK-3b. The inhibitor fits a hydrophobic cavity (surface in gray)
delineated by hydrophobic residues (shown green). The carbon atom bearing the
chloride is above the sulfur atom of Cys199, and the chlorine atom is close to the
amino group of Lys85 (shown in blue).
2.3.6. Derivatization of sulfhydryl groups on GSK-3
To check the relevance of Cys199 in the inhibition of GSK-3 we
have alkylated the free thiol groups in GSK-3b by the treatment
with iodoacetamide. We have observed a complete lack of activity
after this treatment, showing the relevance of this chemical motif
in mediating the inhibition mechanism of the enzyme. Two differ-
ent inhibitors, lithium and the HMK 32 were used as control
a new way to selectively target this important enzyme. According
2
9
to the previous SAR and the GSK-3 3D surface potential maps, it
is reasonable to hypothesize that HMKs inhibitors are first recog-
nized near the enzymatic catalytic site and subsequently modify
(Fig. 6).
Figure 5. The unique Cys residue near the ATP-biding site in GSK-3 (a and b) is shown by their alignment with different CDKs.

6
920
D. I. Perez et al. / Bioorg. Med. Chem. 17 (2009) 6914–6925
2
.4.1. Tau phosphorylation
Since GSK-3b is shown to phosphorylate tau both in transfected
3
0
cells and in vivo, the ability of HMKs to interfere with the GSK-
b-mediated tau phosphorylation in cells was explored using pri-
3
mary granule cerebellar neurons (GCNs).
The effect of HMK addition on tau phosphorylation was deter-
mined by Western blotting. Detection was carried out using the
PHF-1 phosphospecific antibody and Tau-1 (see Section 4). Lithium
chloride was added as positive control to the medium at 10 mM.
The bands were quantified by densitometric analysis. Data showed
that the addition of different HMKs (compounds 18, 19, 23, 28–30,
and 32) reduced (PHF-1) or increased (Tau-1) the tau-immunore-
activity (see Fig. 8). The effect of HMKs on tau phosphorylation
was similar to the trend observed when lithium chloride was used
as a GSK-3b inhibitor (IC50 1.5 mM).
Figure 6. The sulfhydryl groups of GSK-3b were derivatized by reaction with
iodoacetamide. The kinase activity from GSK-3 derivatized by iodoacetamide (GSK-
3
-SH TREATED), or without any modification (GSK-3-CONTROL) was assayed with
myelin binding protein. The autoradiography shows that HMK 32 reduced the
kinase activity, in the range of 1–10 mM, and that GSK-3 treated with iodoacet-
amide (GSK-3-SH TREATED) lost its kinase activity.
2
.3.7. Reversibility studies
To confirm experimentally the irreversible enzymatic inhibition
2
.4.2. Increased neuritogenesis
hypothesized for HMKs, we have studied the recovery of GSK-3b
activity after HMK treatment. To this end, an immobilized recom-
binant histidine-tagged GSK-3b on a nickel agarose gel was used,
and the activity of the solid-supported enzyme was measured by
the scintillation method. In the presence of HMKs 19 and 32, a
reduction of the phosphorylation activity was observed (after
GSK-3b is involved in cellular differentiation processes and pro-
3
1
motes neurite outgrowth in different cell lines. In some neuro-
blastoma cell lines, such as N2A, a neuronal-like phenotype may
be generated by either incubation in a serum-free medium, or by
lithium chloride treatment in the presence of serum (DEMEN-
1
0%FCS). To evaluate whether some HMKs may have a similar lith-
2
0 min the activity was only 20% of the maximal phosphorylation
ium-like effect, we added these compounds to N2A in the presence
of serum. As indicated in Figure 9, treatment of N2A cell cultures
with compounds 19 and 23 for 16 h showed an increase in neurite
outgrowth. These results suggest that HMKs may have similar
intracellular targets as lithium chloride and indirectly show good
cell permeability.
activity). Then, several careful washes were performed on the
immobilized enzyme to eliminate the inhibitor and the enzymatic
activity was subsequently measured. The GSK-3b phosphorylating
ability was not recovered, suggesting that the inhibitor remains
irreversibly bound to the enzyme.
Moreover, a second experiment was performed using recombi-
nant GSK-3b in solution. In that case, we pre-incubated for 30 min
the enzyme and the HMK 32. Then, the reaction mixture was di-
luted 20 times and the final kinase activity determined. Control
was used following the same strategy but without inhibitor. We
have found a complete lack of activity after the HMK pretreatment,
showing a potential covalent or very strong binding to the enzyme
2
.5. ADME properties
Early ADME considerations in preclinical development would
help to avoid costly late stage preclinical and clinical failures,
and a number of in silico, in vitro and in vivo techniques are avail-
able to screen compounds for ADME characteristics.
purposes, the oral absorption and blood brain barrier permeability
properties of HMKs were explored using the previously reported
CODES/neural network methodology, which was able to predict
3
2,33
For our
(
Fig. 7).
These findings suggest that HMKs are the first irreversible GSK-
b inhibitors described up to now, leading to a new interpretation
3
3
4
of the plots depicted in Figure 3a. When GSK-3b is incubated in the
presence of an HMK derivative, covalent modification of Cys199
abolishes its phosphorylating activity, presumably by altering the
correct positioning of ATP in the ATP-binding site, thus explaining
why an increase in ATP concentration does not vary the enzymatic
inhibition.
those properties for structurally diverse drugs. Using our in silico
ADME model based on CODEs descriptors, we have determined
that compounds 29 and 35 are able to cross the blood brain barrier
and be orally absorbed (see Table 5). Therefore, HMKs have poten-
tial value to be used in animal studies aimed at determining the
functional implications of GSK-3 in certain neurological disorders.
At this point it is worth noting that peptidyl chloromethylketones,
which have limited biological utility due to their potential toxicity,
are useful tools to study processes mediated by cysteine protease
2
.4. Cellular effects of GSK-3 inhibition
3
5
To evaluate the cell permeability of HMK inhibitors and their
in animal models.
potential use as pharmacological tools and/or leads for drug design,
we further explored their effect on tau phosphorylation and
neuritogenesis.
3
. Conclusions
HMKs represent a new family of GSK-3b inhibitors that are the
first irreversible inhibitors reported to date for this therapeutically
valuable kinase. We postulate that their inhibitory activity is med-
iated by covalent modification of Cys199 residue, which might rep-
resent a new strategy to be considered for the future design of new
specific GSK-3 inhibitors. We have also shown their ability to pro-
mote neurite outgrowth and to decrease tau phosphorylation. In
addition, they exhibit the ability to activate the promoter IV of
brain-derived growth factor on neurons. Furthermore, these
compounds are predicted to cross the blood–brain barrier, increas-
ing their value as tools to be used in animal models for neurode-
generative disorders. Finally, it is worth mentioning that they are
now available from several commercial sources, that is, EMD-Cal-
Figure 7. GSK-3b was incubated with HMK 32 (10 lM), afterward was diluted 20
3
6
times and then the kinase activity assayed. The kinase activity of GSK-3 pre-
inhibited (GSK-3-HMK PRE-TREATED); or without any previous inhibition (GSK-3-
CONTROL) were assayed using MBP, as substrate. The autoradiography shows the
kinase activity, in both circumstances, after 10 or 30 min of reaction (C10, C30).
Subsequent addition of HMK 32 at 10
lM (3), or lithium chloride at 1 mM (4) or
HMK 32 (0.5 M) (5) were also shown.
l

D. I. Perez et al. / Bioorg. Med. Chem. 17 (2009) 6914–6925
6921
Figure 8. A representative Western blot of soluble extracts from Granule cerebellar neurons (GCNs), obtained after 16 h in the presence of lithium chloride (Li) or different
HMKs (compounds 18, 19, 23, 28, 29, 30, 32). C represents cell extracts from control GCNs. Identical samples were incubated with anti-tau-1, or anti PHF-1; or with anti-b-
actin as a loading control. In each experiment, PHF-1 was normalized with respect to the amount of actin present in each cell extract. The data are expressed as the mean of
three independent experiments and the data from control cells were considered 100 relative units (r.u.).
Figure 9. Phase-contrast photography from N2A cells treated for 16 h with different HMKs. (a) Control + DMSO treatment, (b) treatment with compound 23 at 0.2 lM, (c)
treatment with compound 19 at 2 lM. The arrows illustrated some cytoplasmic extension detected only after compound 19 or 23 additions (b, c).
sition time 0.99 s, pulse width 9
shifts are reported in values (ppm) relative to internal Me
l
s (57°), data size 32 K. Chemical
Si and J
Table 5
Predicted percentage of oral absorption and calculated log [BB] for compounds 29 and
5 using neural network/CODES model
4
3
values are reported in hertz. Elemental analyses were performed
by the analytical department at CENQUIOR (CSIC), and the results
obtained were within ±0.4% of the theoretical values.
Mass spectra were obtained by electronic impact in a Hewlett–
Packard 5973 spectrophotometer.
Compound Oral absorption (%)
Log [BB]
2
3
9
5
100
85
0.166
0.085
HPLC analyses for HMKs were performed on a Waters 6000
instrument, with UV detector (214–254 nm), using different col-
umns; Delta Pack C18, 5
l
m, 300 Å, (150 ꢁ 3.9 mm),
m, 125 Å, (300 ꢁ 3.9 mm) and Symmetry C18, 5
O [(0.05% H PO + 0.04%
3
N)] 50/50 was used as mobile phase. For the chemical reactivity
l
Bondapack
biochem and Merck, as GSK-3b inhibitor VI and VII. These two
C18, 10
l
lm,
compounds have been successfully used to study different aspects
1
Et
00 Å, (150 ꢁ 3.9 mm). Acetonitrile/H
2
3
4
3
6,37,38
of GSK-3 functions.
Overall, HMKs represent useful pharma-
cological tools for use in proof of concept studies for neurological
disorders where GSK-3 might be involved.
of HMKs versus thiol groups, experiments were performed in a
Waters 2695 diode array, with positive electrospray using the sun-
fire column C18, 3.5
mic acid) and MiliQ water (0.1% formic acid) as mobile phase in a
l
m (50 ꢁ 4.6 mm) and acetonitrile (0.08% for-
4
4
. Experimental
1
0–100 gradient.
The synthesis of chloroacetyl thiophenes 22–36 were previ-
.1. Chemistry
1
7
ously described and their analytical and spectroscopic data are
reported in Supplementary data.
Melting points were determined with a Reichert-Jung Therm-
ovar apparatus and are uncorrected. Flash column chromatography
was carried out at medium pressure using silica gel (E. Merck,
Grade 60, particle size 0.040–0.063 mm, 230–240 mesh ASTM).
Compounds were detected with UV light (254 nm). H NMR spec-
tra were obtained on a Gemini-200 spectrometer working at
4.1.1. General procedure for the synthesis of 4-chlorobutyryl-
thiophenes (37, 38, 43) and bromoacetylthiophenes (39–42)
The corresponding thiophene was added to a stirred mixture of
4-chlorobutyryl chloride (compounds 37, 38, 43) or bromoacetyl
bromide (BrAcO-Br) (compounds 39–42) and anhydrous aluminum
1
2
00 MHz. Typical spectral parameters: spectral width 10 ppm,
13
pulse width 9
l
s (57°), data size 32 K. C NMR experiments was
3 2
chloride (AlCl ) in dry carbon disulfide (CS ). Then the mixture was
carried out on the Varian Gemini-200 spectrometer operating at
stirred at room temperature overnight, refluxed for 3 h and cooled
in an ice bath. Under vigorous stirring, HCl 0.1 N was cautiously
5
0 MHz. The acquisition parameters: spectral width 16 kHz, acqui-

6
922
D. I. Perez et al. / Bioorg. Med. Chem. 17 (2009) 6914–6925
added to this cooled mixture, the present solid disappeared and the
reaction mixture separated in two layers. The aqueous layer was
4.1.1.5. 2-Bromo-1-(5-bromothiophen-2-yl)ethanone (41).
Bromoacetyl bromide (1.91 g, 9.50 mmol), anhydrous AlCl
3
extracted with CH
and washed again with water (25 mL), HCl 0.1 N (2 ꢁ 15 mL), and
a saturated solution of NaHCO (15 mL), dried over magnesium
sulfate and evaporated. The compounds were purified by column
chromatography.
2
Cl
2
(50 mL). The organic layers were combined
(1.26 g, 9.50 mmol), 2-bromothiophene (2.0 g, 8.26 mmol) and
dry CS (50 mL). Purified by column chromatography (hex-
ane:ethyl acetate 8:1), 0.64 g (42%), colorless solid, mp 90–91 °C.
2
3
1
H NMR (200 MHz, CDCl
J = 4.02 Hz, 1H, Ar), 4.30 (s, 2H, CH
d 183.6 (CO), 142.2 (C–CO), 133.8, 131.6 (CH), 124.6 (C–Br), 29.8
3
): d 7.57 (d, J = 4.02 Hz, 1H, Ar), 7.17 (d,
1
3
2
); C NMR (75 MHz, CDCl
3
):
+
4
.1.1.1. 4-Chloro-1-(5-chlorothiophen-2-yl)butan-1-one (37).
-Chlorobutyryl chloride (3.24 g, 0.023 mol), anhydrous AlCl
3.06 g, 0.023 mol), 2-chlorothiophene (2.37 g, 0.020 mol) and dry
CS (50 mL). Purified by column chromatography (hexane:ethyl
acetate 4:1), 2.06 g (46%), colorless solid, mp 46–47 °C. H NMR
200 MHz, CDCl ): d 7.53 (d, J = 4.0 Hz, 1H, Ar), 6.96 (d, J = 4.0 Hz,
H, Ar), 3.65 (m, 2H, CH ), 3.05 (m, 2H, CH ), 2.20 (m, 2H, CH );
): d 191.2 (CO), 142.5 (C–CO), 139.8 (C–
Cl), 131.8 (CH), 128.0 (CH), 44.1 (CH ), 35.0 (CH ), 26.3 (CH ); m/
z (EI): 226, 224, 222 (M , 1, 3, 5%), 147, 145 (M–CH CH CH Cl,
8, 100%); HPLC: Column Bondapack C18, 10 m, 125 Å,
300 ꢁ 3.9 mm), purity 99%, rt = 4.16 min, acetonitrile/H O (0.05%
PO + 0.04% Et N) 50/50. Anal. Calcd for C Cl OS: C, 43.06;
H, 3.61; S, 14.37. Found: C, 42.98; H, 3.49; S, 14.41.
(CH
99, 100%); HPLC: Column
(300 ꢁ 3.9 mm), purity 98%, rt = 4.78 min, acetonitrile/H
PO + 0.04% Et N) 50/50. Anal. Calcd for C Br OS: C, 25.38;
H, 1.42; S, 11.29. Found: C, 25.29; H, 1.49; S, 11.40.
2
); m/z (EI): 286, 284, 282 (M , 14, 26, 13%), 191, 189 (M–CH
Bondapack C18, 10 m, 125 Å,
O (0.05%
2
Br,
4
3
l
l
(
2
2
H
3
4
3
H
6 4
2
1
(
3
1
2
2
2
4.1.1.6. 2-Bromo-1-(3,5-dibromothiophen-2-yl)ethanone (42).
1
3
C NMR (50 MHz, CDCl
3
Bromoacetyl bromide (0.32 g, 1.59 mmol), anhydrous AlCl
3
2
2
2
(0.21 g, 1.59 mmol), 2,4-dibromo thiophene (0.35 g, 1.45 mmol)
and dry CS (50 mL). Purified by column chromatography (hex-
ane:ethyl acetate 7:1), 0.39 g (72%), colorless solid, mp 75–76 °C.
+
2
2
2
2
3
l
l
1
(
2
H NMR (200 MHz, CDCl
C NMR (75 MHz, CDCl
3
3
): d 7.22 (s,1H, Ar), 4.53 (s, 2H, CH
): d 182.8 (CO), 141.4 (C–CO), 136.2
(CH), 123.8, 114.8 (C–Br), 32.9 (CH ); m/z (EI): 366, 364, 362, 360
M , 7, 20, 7, 20%), 271, 269, 267 (M–CH Br, 61, 100, 61%); HPLC:
Column Bondapack C18, 10
m, 125 Å, (300 ꢁ 3.9 mm), purity
99%, rt = 4.23 min, acetonitrile/H O (0.05% H PO + 0.04% Et N)
50/50. Anal. Calcd for C Br OS: C, 19.86; H, 0.83; S, 8.84. Found:
C, 19.81; H, 0.77; S, 8.78.
2
);
1
3
H
3
4
3
8
H
8
2
2
+
(
2
4
.1.1.2. 1-(5-Bromothiophen-2-yl)-4-chlorobutan-1-one (38).
-Chlorobutyryl chloride (3.24 g, 0.023 mol), anhydrous AlCl
3.06 g, 0.023 mol), 2-bromothiophene (3.26 g, 0.020 mol) and
dry CS (50 mL). Purified by column chromatography (hex-
ane:ethyl acetate 4:1), 2.00 g (38%), colorless solid, mp 43–44 °C.
l
l
4
3
2
3
4
3
(
6
H
3
3
2
1
H NMR (200 MHz, CDCl
3
): d 7.40 (d, J = 4.0 Hz, 1H, Ar), 7.05 (d,
), 3.0 (m, 2H, CH ), 2.15 (m,
): d 191.5 (CO), 145.8 (C–CO),
32.2 (CH), 131.4 (CH), 122.9 (C–Br), 44.5 (CH ), 35.3 (CH ), 26.9
); m/z (EI): 270, 268, 266 (M , 4, 13, 10%), 191, 189 (M–
CH CH Cl, 100, 100%); HPLC: Column Bondapack C18,
m, 125 Å, (300 ꢁ 3.9 mm), purity 99%, rt = 5.85 min, acetoni-
trile/H (0.05% PO + 0.04% Et N) 50/50. Anal. Calcd for
BrClOS: C, 35.91; H, 3.01; S, 11.98. Found: C, 35.78; H, 2.89;
S, 11.90.
4.1.1.7. 4-Chloro-1-(2,5-dichlorothiophen-3-yl)butan-1-one (43).
4-Chlorobutyryl chloride (3.24 g, 0.023 mol), anhydrous AlCl
(3.06 g, 0.023 mol), 2,5-dichlorothiophene (3.06 g, 0.020 mol) and
dry CS (50 mL). Purified by column chromatography (hexane:ethyl
acetate 4:1), 2.10 g (41%), oil. H NMR (200 MHz, CDCl
1H, Ar), 3.70 (m, 2H, CH ), 3.14 (m, 2H, CH ), 2.20 (m, 2H, CH
): d 192.0 (CO), 136.6 (C–CO), 132.3 (C–
Cl), 130.0 (C–Cl), 126.9 (CH), 44.4 (CH ), 38.9 (CH ), 26.4 (CH ); m/
z (EI): 262, 260, 258, 256 (M , 1, 4, 10, 9%), 183, 181, 179 (M–
CH CH CH Cl, 18, 79, 100%); HPLC: Column Bondapack C18,
m, 125 Å, (300 ꢁ 3.9 mm), purity 98%, rt = 5.46 min, acetoni-
trile/H (0.05% PO + 0.04% Et N) 50/50. Anal. Calcd for
Cl OS: C, 37.31; H, 2.74; S, 12.45. Found: C, 37.40; H, 2.64; S,
12.42.
J = 4.0 Hz, 1H, Ar), 3.60 (m, 2H, CH
2
2
3
1
3
2
1
2 3
H, CH ); C NMR (50 MHz, CDCl
2
2
2
+
1
(
CH
2
3
): d 7.20 (s,
);
CH
2
2
2
l
2
2
2
1
3
1
0
l
C NMR (50 MHz, CDCl
3
2
O
H
3
4
3
2
2
2
+
8 8
C H
2
2
2
l
10 l
4
.1.1.3. 2-Bromo-1-(3,4-dibromothiophen-2-yl)ethanone (39).
2
O
H
3
4
3
Bromoacetyl bromide (2.40 g, 0.012 mol), anhydrous AlCl
3
C H
8 7
3
(
1.57 g, 0.012 mol), 3,4-dibromothiophene (2.5 g, 0.010 mol) and
dry CS (50 mL). Purified by column chromatography (hex-
ane:ethyl acetate 8:1), 3.60 g (96%), colorless solid, mp 82–83 °C.
2
4.1.2. 2-(2,5-Dichlorothiophen-3-yl)-2-oxoethyl acetate (44)
To a solution of 2-chloro-1-(2,5-dichlorothiophen-3-yl) etha-
none (35) (0.2 g, 1.31 mmol) in acetic acid (50 mL) was added so-
dium acetate (0.32 g, 3.94 mmol) and acetic anhydride (0.27 g,
1
H NMR (200 MHz, CDCl
C NMR (75 MHz, CDCl
3
): d 7.71 (s, 1H, Ar), 4.62 (s, 2H, CH
): d 181.5 (CO), 140.2 (C–CO), 131.3
CH), 119.2, 118.5 (C–Br), 32.8 (CH ); m/z (EI): 366, 364, 362, 360
M , 4, 10, 4, 10%), 271, 269, 267 (M–CH Br, 52, 100, 52%); HPLC:
Bondapack C18, 10
m, 125 Å, (300 ꢁ 3.9 mm), purity
9%, rt = 4.00 min, acetonitrile/H O (0.05% H PO + 0.04% Et N)
0/50. Anal. Calcd for C Br OS: C, 19.86; H, 0.83; S, 8.84. Found:
2
);
1
3
3
(
(
2
+
2
3.94 mmol). The mixture was stirred and refluxed for 5 h. CH
(100 mL) and H O (100 mL) was added to the reaction. The organic
layer was washed with H
O (2 ꢁ 100 mL), a saturated solution of
NaHCO (100 mL) and a saturated solution of NaCl (100 mL). The
2 2
Cl
Column
9
5
l
l
2
2
3
4
3
2
6
H
3
3
3
C, 19.79; H, 0.70; S, 8.69.
organic layer was dried over magnesium sulfate and the solvent
was evaporated. Compound 44 was purified by column chromatog-
1
4
.1.1.4. 2-Bromo-1-(4,5-dibromothiophen-2-yl)ethanone (40).
raphy (hexane:ethyl acetate 7:1), 0.30 g (91%), oil. H NMR
Bromoacetyl bromide (0.95 g, 4.74 mmol), anhydrous AlCl
3
(200 MHz, CDCl
3
): d 7.28 (s, 1H, Ar), 5.22 (s, 2H, CH
); C NMR (50 MHz, CDCl ): d 184.5 (CO), 169.2 (CO), 133.6
(C–CO), 132.0 (C–Cl), 126.7(C–Cl), 125.5 (CH), 66.6 (CH ), 19.4
); m/z (EI): 256, 254, 252, (M , 5, 27, 54%), 185, 183 (M–
, 100, 100%); HPLC: Column Symmetry C18, 5 m, 100 Å,
(150 ꢁ 3.9 mm), purity 99%, rt = 8.96 min, acetonitrile/H O (0.05%
PO + 0.04% Et N) 50/50. Anal. Calcd for C Cl S: C, 37.96;
2
), 2.25 (s, 3H,
1
3
(
0.63 g, 4.74 mmol), 2,3-dibromothiophene (1.0 g, 4.13 mmol)
and dry CS (50 mL). Purified by column chromatography (hex-
ane:ethyl acetate 9:1), 1.39 g (93%), colorless solid, mp: 67–68 °C.
CH
3
3
2
2
+
(CH
3
1
H NMR (200 MHz, CDCl
C NMR (75 MHz, CDCl
3
): d 7.59 (s, 1H, Ar), 4.27 (s, 2H, CH
): d 182.9 (CO), 141.0 (C–CO), 135.5
CH), 123.4, 115.6 (C–Br), 29.4 (CH ); m/z (EI): 366, 364, 362, 360
M , 4, 10, 4, 10%), 271, 269, 267 (M–CH Br, 52, 100, 52%); HPLC:
Bondapack C18, 10
m, 125 Å, (300 ꢁ 3.9 mm), purity
8%, rt = 4.20 min, acetonitrile/H O (0.05% H PO + 0.04% Et N)
0/50. Anal. Calcd for C Br OS: C, 19.86; H, 0.83; S, 8.84. Found:
2
);
3 5
C H O
2
l
2
1
3
3
(
(
2
H
3
4
3
H
8 6
2 3
O
+
2
H, 2.39. Found: C, 37.89; H, 2.42.
Column
l
l
9
5
2
3
4
3
4.1.3. Chemical reactivity of HMKs
Hundred microliters of an acetonitrile solution of HMK 29
(50 mM) and 100 L of an acetonitrile solution of 2-mercap-
6
H
3
3
C, 19.69; H, 0.74; S, 8.64.
l

D. I. Perez et al. / Bioorg. Med. Chem. 17 (2009) 6914–6925
6923
toethanol (50 mM) were added to a mixture of methanol/water (1/
, 0.8 mL) and stirred for 30 min or 18 h. The reaction progress was
followed by HPLC–MS injecting aliquots after 30 min or 18 h. HMK
9 (MW = 195) showed a retention time of 4.80 min, while the S-
alkylated derivative (MW = 236) presented a retention time of
.06 min. Additional experiment was carried out when the 2-
mercaptoethanol concentration was increased 10 times. In this
case 100 L of an acetonitrile solution of 2-mercaptoethanol
500 mM) was added. The same procedure described above was
followed when triethylamine was used. In this case, 100 L of an
This concentration was previously determined to produce the
maximal inhibition. After that, the activity of GSK-3b, treated with
HMK 32 or without any inhibitor (control), diluted 20 times in ki-
nase buffer, and maintained for 30 min to equilibrate the different
components, was determined. In parallel, the control reaction
(GSK-3 without inhibitor) was incubated at a final concentration
of HMK 32 similar to those obtained after dilution (0.5 lM); or
low lithium chloride concentration. The final kinase activity was
assayed over Myelin basic protein (MBP), as described below.
1
2
4
l
(
l
acetonitrile solution of triethylamine (50 mM) was added to a
4.2.3. Derivatization of sulfhydryl groups of GSK-3
methanol/water (1/1, 0.7 mL) solution.
The sulfhydryl groups were derivatized by reaction with 15 mM
iodoacetamide in a buffer containing 50 mM tris, pH 7.4, 1 mM
EGTA. GSK-3b enzyme (Sigma) was incubated in this buffer, for
4
4
0
.2. Biological studies
2
0 or 30 min and then the kinase activity was assayed. GSK-3b en-
zyme (Sigma) derivatized with iodoacetamide or without any
modification was assayed for its kinase activity.
GSK-3b activity was determined in 50 mM tris, pH 7.5, 10 mM
MgCl2, 1 mM EGTA buffer (kinase buffer) at 37 °C, in the presence
.2.1. Inhibition of GSK-3
GSK-3b enzyme (Sigma) was incubated with 15 lM of ATP,
3
2
.2 lCi of [c- P]ATP, GS-1 substrate, and different concentration
of test compound. GSK-3b activity was assayed in 50 mM tris, pH
.5, 10 mM MgCl , 1 mM EGTA and 1 mM EDTA buffer at 37 °C,
in the presence of 15 mM GS-1 (substrate), 15 M of ATP, 0.2 Ci
L. After 20 min of incubation
L aliquots of the supernatant were spotted onto 2x2
of 4
l
g of myelin basic protein (MBP) as substrate, 15
l
M of ATP,
7
2
32
0
.2
l
Ci of [
c
-
P] ATP in a final volume of 25
l
l. As an internal con-
l
l
trol, the inhibitory capacity of HMK 32 or lithium chloride was
used as reference.
3
2
of [
c- P] ATP in a final volume of 12 l
at 37 °C, 4
l
pieces of Whatman P81 phosphocellulose paper, and the filter
was washed four times (at least 10 min each time) in 1% phospho-
ric acid. The dried filters were transferred into scintillation vials,
and the radioactivity was measured in a liquid scintillation coun-
ter. Blank values were subtracted, and the GSK-3b activity was ex-
pressed in percentage of maximal activity. The IC50 is defined as
the concentration of each compound that reduces enzyme activity
by 50% with respect to that without inhibitor present.
4
.2.4. Neurotransmitter receptors binding assays
The neurotransmitter binding assays are described in each
paragraph:
–
–
–
–
–
–
–
–
–
a
-2 (Non-selective) adrenergic receptor:39 From rat cerebral
3
cortex, yohimbine was used as reference compound and [ H]
RX 821002 as a ligand (0.5 nM).
4
0
D2 (h) dopamine receptor: Human recombinant (CHO cells),
3
(
+)-buctaclamol was used as reference compound and [ H] spip-
4
.2.2. Reversibility inhibition studies of GSK-3
erone as a ligand (0.3 nM).
D3 (h) dopamine receptor: Human recombinant (CHO cells),
4
1
Immobilized recombinant histidine-tagged GSK-3b on a nickel
3
agarose gel (Sigma) was incubated with 15
[c-
l
M of ATP, 0.2
P]ATP, GS-1 substrate, and different concentration of test
compounds. GSK-3b activity was assayed in 50 mM tris, pH 7.5,
0 mM MgCl , 1 mM EGTA and 1 mM EDTA buffer at 37 °C, in a fi-
nal volume of 20 L. After 20 min of incubation at 37 °C, the reac-
lCi of
(
+)-buctaclamol was used as reference compound and [ H] spip-
erone as a ligand (0.3 nM).
AMPA glutaminergic receptor: From rat cerebral cortex,
3
2
4
2
L
-glu-
3
1
2
tamate was used as reference compound and [ H] AMPA as a
ligand (8 nM).
NMDA glutaminergic receptor: From rat cerebral cortex, CGS
l
4
3
tion mixture was centrifuged at 12,000 rpm. for 5 min at 4 °C. The
supernatant was collected to evaluate the radioactivity incorpo-
rated in the GS-1 peptide. Thus, a 4 lL aliquot of the supernatant
were spotted onto 2 ꢁ 2 pieces of Whatman P81 phosphocellulose
paper, and the filter was washed four times (at least 10 min each
time) in 1% phosphoric acid. The dried filters were transferred into
scintillation vials, and the radioactivity was measured in a liquid
scintillation counter. Blank values were subtracted, and the GSK-
3
1
9755 was used as reference compound and [ H] CGP 39653
as a ligand (5 nM).
M (non-selective) muscarinic receptor: From rat cerebral cor-
4
4
3
tex, atropine was used as reference compound and [ H] QNB as a
ligand (0.05 nM).
N (neuronal) (
cerebral cortex, nicotine was used as reference compound and
a
-BGTX-insensitive) nicotinic receptor:45 From rat
3
3b activity was expressed in percentage of maximal activity.
[
H] cytisine as a ligand (1.5 nM).
N (neuronal) (
-BGTX-sensitive) nicotinic receptor:46 From rat
cerebral cortex,
On the other hand, the pellet with the immobilized GSK-3 on
the nickel agarose gel was washed three times (at least 10 min
each time) with abundant volume of 50 mM tris, pH 7.5, 10 mM
a
a
a
-bungarotoxin was used as reference com-
-bungarotoxin as a ligand (1 nM).
1
25
pound and [ I]
4
7
MgCl
After the last washing, the pellet enzymatic activity was deter-
mined in 50 mM tris, pH 7.5, 10 mM MgCl , 1 mM EGTA and
mM EDTA buffer containing 15 M of ATP, 0.2 lCi of [c- P]ATP,
2
buffer, to eliminate the inhibitor unbounded to the enzyme.
5-HT (non-selective) serotoninergic receptor: From rat cere-
bral cortex, serotonine was used as reference compound and
3
2
[
H] serotonine as a ligand (2 nM).
3
2
1
l
and 15 mM GS-1 substrate. After 15 min incubation the radioactiv-
ity incorporated in the GS-1 peptide was determined following the
method above described. The GSK-3b activity in the presence of the
compound is expressed as a percentage of the GSK-3b activity of
the control without the compound but with the same work-up.
A second study was performed to determine the irreversibility
of HMK inhibition. Thus, GSK-3b enzyme (Sigma) was incubated
in a buffer containing 20 mM Tris, pH 7.4, for 30 min, before deter-
4
.2.5. Inhibition of protein kinases
The experimental procedures for the inhibition of different pro-
tein kinases are described in each paragraph:
–
Abl kinase:48 Mouse recombinant (E. coli), staurosporine was
used as reference compound and poly GT (0.4 lg.mL ) as a sub-
strate. Fluorescence polarization was used as the method of
detection for the reaction product (phosphopolyGT).
ꢀ1
mine its activity, with HMK 32, at a final concentration of 10 lM.

6
924
D. I. Perez et al. / Bioorg. Med. Chem. 17 (2009) 6914–6925
–
–
–
–
–
–
CAM kinase II:49 From rat brain, staurosporine was used as ref-
teins (10–50 lg) were separated by SDS–PAGE before being
33
erence compound and [
c
-P] ATP + autocamtide-2 (5
l
M) as a
transferred to a nitrocellulose filter. The filters were blocked with
5% non-fat milk in PBS, 0.1% Tween-20 (PBS-T) and then incubated
with primary antibodies overnight at 4 °C. Subsequently, the filters
were rinsed three times in PBST buffer before being exposed to the
corresponding peroxidase-conjugated secondary antibody (diluted
1:5000, Promega) for 1 h at room temperature. Immunoreactivity
was visualized using an enhanced chemiluminescence detection
system (Perkin–Elmer Life Sci.). Each experiment was normalized
with respect to the amount of actin present in each cell extract.
The data are expressed as the mean of three independent experi-
ments and the data from control cells were considered 100 relative
unit (r.u.).
substrate. Liquid scintillation was used as method of detection
3
3
for the reaction product ([
c
-P] autocamtide-2).
5
0
EGFR kinase: From A-431 cells, PD 153035 was used as refer-
33
ꢀ1
ence compound and [
c -P] ATP + poly GAT (0.48 mg mL ) as a
substrate. Liquid scintillation was used as method of detection
3
3
for the reaction product ([
c
-P] poly GAT).
5
1
IR-K (h): Human recombinant, staurosporine was used as ref-
33
ꢀ1
erence compound and [
c -P] ATP + poly GT (0.03 mg.mL ) as a
substrate. Liquid scintillation was used as method of detection
3
3
for the reaction product ([
c -P] poly GT).
MAP kinase (ERK 42):52 Rat recombinant (E. coli), staurosporine
was used as reference compound and
0.5 mg mL ) as a substrate. Liquid scintillation was used as
method of detection for the reaction product ([
3
3
[
c
-P] ATP + MBP
ꢀ
1
(
4.3. Molecular modeling. CMIP calculations
33
c -P] MBP).
5
3
MEK1 kinase: Rabbit recombinant (E.coli), staurosporine was
Preferential binding sites of HMKs on GSK-3 protein surface
were obtained using the docking module of the program CMIP.
The protein (PDB entry 1I09) was mapped onto a 3D grid of the
appropriate size with a 0.5 Å regular spacing (ca. 4,000,000 grid
positions). Binding was assayed by exhaustive search of 8000 ori-
entations (using non-redundant Euler angles) for every grid posi-
tion. When necessary, flexible molecules were assayed as a
family of standard rotamers. Docked positions were scored using
interaction energies between the ligand and the protein, which
were evaluated by adding electrostatic and van der Waals contri-
butions. The solvent-screened electrostatic interaction was deter-
mined by using the Mehler-Solmajer sigmoidal distance-
used as reference compound and ATP + unactivated MAP kinase
ꢀ
1
(
0.01 mg mL ) as a substrate. Liquid scintillation was used as
33
method of detection for the reaction product ([
Protein kinase p56
c -P] MBP).
From bovine thymus, staurosporine
lck 54
:
33
was used as reference compound and [
c
-P] ATP + poliGT
ꢀ
1
(
0.5 mg mL ) as a substrate. Liquid scintillation was used as
33
method of detection for the reaction product ([
c -P] polyGT).
4
.2.6. Neuronal cell cultures and treatments
Neuro2A mouse neuroblastoma cells were routinely grown in
5
7
Dulbecco’s modified Eagle’s medium (DMEM) with 10% heat-inac-
tivated fetal calf serum containing glutamine (2 mM) and antibiot-
ics (Pen./Strept.).
dependent dielectric model, and the van der Waals component
was evaluated by using a Lennard–Jones expression. Parameters
for protein atoms were obtained from the amber98 force field.58
For the ligand molecules, RESP atomic charges determined at the
HF/6-31G(d) level were used in conjunction with van der Waals
parameters taken for related atoms in the amber98 force field. Fi-
nally, refinement of the binding mode was carried out by means of
docking computations with the program rDock, which is an exten-
To obtain a neuronal-like phenotype, cells were maintained for
2
2
4–48 h in serum-free neurobasal medium supplemented with B-
7 (NB-B27). Alternatively N2A have specific properties and neuro-
nal-like phenotype may be induced by addition of lithium chloride
0 mM, in the growing media (DEMEN-10%FCS). The effect was
1
5
9
analyzed 24 h after addition. Similarly, some HMKs were assayed
for the potential effect of producing neurite extension as lithium
chloride did, and the compounds were added as indicated for 16 h.
Granule cerebellar neurons (GCNs) were obtained as described
previously. Briefly, cerebellar granule neurons were isolated from
postnatal day 6 (P6) mice as described. The cells were plated onto
sion of the program RIBODOCK using an empirical scoring function
6
0
calibrated based on protein–ligand complexes. To this end, 100
docking computations were run, and the output docking modes
were analyzed by visual inspection in conjunction with the dock-
ing scores.
5
5
poly-L-lysine coated dishes and grown for 48 h in serum-free neu-
Acknowledgement
robasal medium supplemented with B-27 and 12.5 mM KCl.
Lithium chloride (Sigma) was used in this study as the GSK-3 ki-
We acknowledge partial financial support from NeuroPharma
5
6
nase inhibitor. As lithium also inhibits inositol monophospha-
tase, all of our experiments with lithium were performed in the
presence of an excess of myo-inositol to avoid the effects provoked
by inositol depletion.
(
currently named NOSCIRA).
Supplementary data
The inhibition of GSK-3 was inferred from tau phosphorylation
and the effects were compared with parallel cultures that were ex-
posed to selected HMKs, at different concentration (ranging from
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2009.08.042.
0
.5 to 25 lM).
References and notes
4
.2.7. Antibodies and Western blot analysis
1. Cohen, P. Nat. Rev. Drug Disc. 2002, 1, 309.
2. Sawa, M. Mini-Rev. Med. Chem. 2008, 8, 1291.
3. Hochhaus, A. J. Natl. Compr. Canc. Netw. 2008, 6(Suppl 2), S1.
The antibodies used in this study were: anti-b-actin mAb (Sig-
ma); PHF-1, which is an antiphospho Ser3
supplied by Dr. P. Davies, Albert Einstein Coll, Bronx, NY, USA);
and Tau monoclonal antibody Tau-1, from Chemicon, recognizes
Residues Ser 195,198,199,202 non-phosphorylated.
96/404
Tau mAb (kindly
4
5
.
.
Danovi, S. A.; Wong, H. H.; Lemoine, N. R. Br. Med. Bull. 2008, 87, 97.
Noble, M. E.; Endicott, J. A.; Johnson, L. N. Science 2004, 303, 1800.
6. Schmidt, R.; Neff, F.; Lampl, C.; Benke, T.; Anditsch, M.; Bancher, C.; Dal-Bianco,
P.; Reisecker, F.; Marksteiner, J.; Rainer, M.; Kapeller, P.; Dodel, R.
Neuropsychiatr. 2008, 22, 153.
7. Hooper, C.; Killick, R.; Lovestone, S. J. Neurochem. 2008, 104, 1433.
8. In Glycogen Synthase Kinase 3 and its Inhibitors; Martinez, A., Castro, A., Medina,
Cell extracts were prepared from cells washed with 1ꢁ PBS and
then resuspended in a buffer containing: 20 mM HEPES, pH 7.4;
M., Eds.; John Wiley & Sons: New Jersey, 2006.
1
5
00 mM NaCl; 100 mM NaF; 1 mM sodium ortho-vanadate;
mM EDTA; Okadaic Acid 1 M; 1% Triton X-100; and a protease
9.
Huang, H. C.; Klein, P. S. Curr. Drug Targets 2006, 7, 1389.
0. Cohen, P.; Goedert, M. Nat. Rev. Drug Disc. 2004, 3, 479.
11. Martinez, A. Med. Res. Rev. 2008, 28, 773.
2. Martinez, A.; Perez, D. I. J. Alzheimers Dis. 2008, 15, 181.
l
1
inhibitor cocktail (Complete, Roche). The soluble fraction was ob-
tained by centrifugation at 14,000g for 10 min at 4 °C, and the pro-
1

D. I. Perez et al. / Bioorg. Med. Chem. 17 (2009) 6914–6925
6925
1
1
3. A. Martinez, M. Alonso, A. Castro, I. Dorronsoro, in: A.E. Martinez (Ed.),
Medicinal Chemistry of Alzheimer’s disease, Transworld Research Network:
Kerala, 2008; p 225.
36. Yasuda, S.; Liang, M.-H.; Marinova, Z.; Yahyavi, A.; Chuang, D.-M. Mol.
Psychiatry 2009, 14, 51.
37. Lochhead, P. A.; Kinstrie, R.; Sibbet, G.; Rawjee, T.; Morrice, N.; Cleghon, V. Mol.
Cell 2006, 24, 627.
4. Castro, A.; Conde, S.; Rodriguez-Franco, M. I.; Martinez, A. Mini-Rev. Med. Chem.
2
002, 2, 37.
38. Li, Z.; Tan, F.; Thiele, C. J. Mol. Cancer Ther. 2007, 6, 3113.
39. Uhlen, S.; Wikberg, J. E. Pharmacol. Toxicol. 1991, 69, 341.
40. Grandy, D. K.; Marchionni, M. A.; Makam, H.; Stofko, R. E.; Alfano, M.;
Frothingham, L.; Fischer, J. B.; Burke-Howie, K. J.; Bunzow, J. R.; Server, A. C.,
et al Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 9762.
41. MacKenzie, R. G.; VanLeeuwen, D.; Pugsley, T. A.; Shih, Y. H.; Demattos, S.;
Tang, L.; Todd, R. D.; O’Malley, K. L. Eur. J. Pharmacol. 1994, 266, 79.
42. Murphy, D. E.; Snowhill, E. W.; Williams, M. Neurochem. Res. 1987, 12, 775.
43. Sills, M. A.; Fagg, G.; Pozza, M.; Angst, C.; Brundish, D. E.; Hurt, S. D.; Wilusz, E.
J.; Williams, M. Eur. J. Pharmacol. 1991, 192, 19.
1
1
5. Castro, A.; Martinez, A. Mini-Rev. Med. Chem. 2001, 1, 267.
6. Martinez, A.; Alonso, M.; Castro, A.; Perez, C.; Moreno, F. J. J. Med. Chem. 2002,
4
5, 1292.
7. Conde, S.; Perez, D. I.; Martinez, A.; Perez, C.; Moreno, F. J. J. Med. Chem. 2003,
6, 4631.
1
4
1
1
8. Woodgett, J. R. Anal. Biochem. 1989, 180, 237.
9. Martinez, A.; Alonso, M.; Castro, A.; Dorronsoro, I.; Gelpi, J. L.; Luque, F. J.;
Perez, C.; Moreno, F. J. J. Med. Chem. 2005, 48, 7103.
2
0. Rosenfeld, R.; Vajda, S.; DeLisi, C. Annu. Rev. Biophys. Biomol. Struct. 1995, 24,
6
77.
44. Richards, M. H. Br. J. Pharmacol. 1990, 99, 753.
2
2
1. Ahn, N. G.; Resing, K. A. Science 2005, 308, 1266.
2. Gattis, J. L.; Ruben, E.; Fenley, M. O.; Ellington, W. R.; Chapman, M. S.
Biochemistry 2004, 43, 8680.
45. Pabreza, L. A.; Dhawan, S.; Kellar, K. J. Mol. Pharmacol. 1991, 39, 9.
46. Sharples, C. G.; Kaiser, S.; Soliakov, L.; Marks, M. J.; Collins, A. C.; Washburn, M.;
Wright, E.; Spencer, J. A.; Gallagher, T.; Whiteaker, P.; Wonnacott, S. J. Neurosci.
2000, 20, 2783.
2
3. Anselmo, A. N.; Cobb, M. H. Biochem. J. 2004, 381, e1.
2
4. Vulpetti, A.; Crivori, P.; Cameron, A.; Bertrand, J.; Brasca, M. G.; D’Alessio, R.;
Pevarello, P. J. Chem. Inf. Model. 2005, 45, 1282.
47. Peroutka, S. J.; Snyder, S. H. Mol. Pharmacol. 1979, 16, 687.
48. Parker, G. J.; Law, T. L.; Lenoch, F. J.; Bolger, R. E. J. Biomol. Screen. 2000, 5, 77.
49. Lengyel, I.; Nairn, A.; McCluskey, A.; Toth, G.; Penke, B.; Rostas, J. J. Neurochem.
2001, 76, 1066.
50. Carpenter, G.; King, L., Jr.; Cohen, S. J. Biol. Chem. 1979, 254, 4884.
51. Al-Hasani, H.; Passlack, W.; Klein, H. W. FEBS Lett. 1994, 349, 17.
52. Robbins, D. J.; Zhen, E.; Owaki, H.; Vanderbilt, C. A.; Ebert, D.; Geppert, T. D.;
Cobb, M. H. J. Biol. Chem. 1993, 268, 5097.
25. Rauber, P.; Angliker, H.; Walker, B.; Shaw, E. Biochem. J. 1986, 239, 633.
26. Drenth, J.; Kalk, K. H.; Swen, H. M. Biochemistry 1976, 15, 3731.
27. Rahman, M. A.; Senga, T.; Oo, M. L.; Hasegawa, H.; Biswas, M. H.; Mon, N. N.;
Huang, P.; Ito, S.; Yamamoto, T.; Hamaguchi, M. Oncol. Rep. 2008, 19, 975.
8. Mazanetz, M. P.; Fischer, P. M. Nat. Rev. Drug Disc. 2007, 6, 464.
9. Dajani, R.; Fraser, E.; Roe, S. M.; Young, N.; Good, V.; Dale, T. C.; Pearl, L. H. Cell
2
2
2
001, 105, 721.
0. Kockeritz, L.; Doble, B.; Patel, S.; Woodgett, J. R. Curr. Drug Targets 2006, 7,
377.
1. Orme, M. H.; Giannini, A. L.; Vivanco, M. D.; Kypta, R. M. Mol. Cell Neurosci.
003, 24, 673.
53. Williams, D. H.; Wilkinson, S. E.; Purton, T.; Lamont, A.; Flotow, H.; Murray, E. J.
Biochemistry 1998, 37, 9579.
54. Cheng, H. C.; Nishio, H.; Hatase, O.; Ralph, S.; Wang, J. H. J. Biol. Chem. 1992, 267,
9248.
55. Sayas, C. L.; Avila, J.; Wandosell, F. J. Neurosci. 2002, 22, 6863.
56. Klein, P. S.; Melton, D. A. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 8455.
57. Mehler, E. L.; Solmajer, T. Protein Eng. 1991, 4, 903.
58. Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K. M.; Ferguson, D. M.;
Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.; Kollman, P. A. J. Am. Chem. Soc. 1995,
117, 5179.
3
3
1
2
3
2. Di, L.; Kerns, E. H. Curr. Opin. Chem. Biol. 2003, 7, 402.
3. Veber, D. F.; Johnson, S. R.; Cheng, H.-Y.; Smith, B. R.; Ward, K. W.; Kopple, K. D.
J. Med. Chem. 2002, 45, 2615.
3
3
4. Dorronsoro, I.; Chana, A.; Abasolo, I.; Castro, A.; Gil, C.; Stud, M.; Martinez, A.
QSAR Comb. Sci. 2004, 23, 89.
3
5. Powers, J. C.; Asgian, J. L.; Ekici, O. D.; James, K. E. Chem. Rev. 2002, 102,
59. Morley, S. D.; Afshar, M. J. Comput. Aided Mol. Des. 2004, 18, 189.
60. Barril, X.; Hubbard, R. E.; Morley, S. D. Mini-Rev. Med. Chem. 2004, 4, 779.
4
639.