Identification of potential cellular targets of aloisine A by affinity chromatography

Article abstract of DOI:10.1016/j.bmc.2009.06.024

Affinity chromatography was used to identify potential cellular targets of aloisine A (7-n-butyl-6-(4′-hydroxyphenyl)-5H-pyrrolo[2,3b]pyrazine), a potent inhibitor of cyclin-dependent kinases. This technique is based on the immobilization of the drug on a

Full text of DOI:10.1016/j.bmc.2009.06.024

Bioorganic & Medicinal Chemistry 17 (2009) 5572–5582
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Identification of potential cellular targets of aloisine A by
affinity chromatography
Caroline Corbel a, Rose Haddoub b, Damien Guiffant a, Olivier Lozach a, David Gueyrard b, Jérôme Lemoine c,
,
*
b,
Morgane Ratin a, Laurent Meijer a, Stéphane Bach a, , Peter Goekjian
*
*
a CNRS USR-3151, Protein Phosphorylation and Human Disease, Station Biologique, B.P.74, F-29682 Roscoff Cedex, Bretagne, France
b Université de Lyon, CNRS UMR 5246 ICBMS, Laboratoire de Chimie Organique 2—Glycochimie, Université Claude Bernard Lyon 1, CPE-Lyon, Bâtiment 308,
43 Boulevard du 11 Novembre 1918, Villeurbanne F-69622, France
c Université de Lyon, CNRS UMR 5180, Université Claude Bernard Lyon 1, Bâtiment 308, 43 Boulevard du 11 Novembre 1918, Villeurbanne F-69622, France
a r t i c l e i n f o
a b s t r a c t
Affinity chromatography was used to identify potential cellular targets of aloisine A (7-n-butyl-6-(40-
hydroxyphenyl)-5H-pyrrolo[2,3b]pyrazine), a potent inhibitor of cyclin-dependent kinases. This tech-
nique is based on the immobilization of the drug on a solid matrix, followed by identification of specif-
ically bound proteins. To this end, both aloisine A and the protein-kinase inactive control N-methyl
aloisine, bearing extended linker chains have been synthesized. We present the preparation of such ana-
logues having the triethylene glycol chain at different positions of the molecule, as well as their immo-
bilization on an agarose-based matrix. Affinity chromatography of various biological extracts on the
aloisine matrices allowed the identification of both protein kinases and non-kinase proteins as potential
cellular targets of aloisine.
Article history:
Received 6 February 2009
Revised 5 June 2009
Accepted 14 June 2009
Available online 18 June 2009
Keywords:
Aloisine
Immobilized protein kinase inhibitor
Mass spectrometry
Click chemistry
Ó 2009 Elsevier Ltd. All rights reserved.
Pyrrolo[2,3-b]pyrazine
Cyclin-dependent kinase
Pyridoxal kinase
Affinity chromatography
GSK-3
Fuse binding protein 1
Glycogen-debranching enzyme
ERK1
1. Introduction
rodegenerative tauopathies.5 Furthermore, cyclin-dependent ki-
nases are conserved across species during evolution, providing
Cyclin-dependent kinases (CDKs) are low-molecular-weight
serine-threonine kinases that require association to a regulatory
subunit, a cyclin, in order to be activated. CDKs play important
roles in the cell division cycle, apoptosis, differentiation, neuronal
functions, and transcription.1–3 Abnormalities and dysregulations
of CDK activities have been associated with many diseases includ-
ing cancers, viral infections, ischemia, and neurodegenerative dis-
orders such as Alzheimer’s and Parkinson’s disease.4 Recent data
have notably implicated CDK5 in aberrant tau phosphorylation
and the association of the latter with microtubules. Hyperphosph-
orylation, aggregation, and formation of the filamentous form of
tau protein (paired helical filaments or PHF-Tau) are all observed
in Alzheimer’s disease but also in other diseases referred to as neu-
new insights and new pharmacological opportunities for the treat-
ment of infectious diseases. These observations have encouraged
an intensive search for potent and selective pharmacological inhib-
itors of CDKs. A large number of small molecular weight inhibitors
have been characterized to date, all of which appear to act by direct
competition with ATP.6–8
The selectivity of kinase inhibitors is usually addressed by test-
ing the inhibitor against a panel of purified kinases.9 But even for
the largest panels, these only reflect a fraction of the 518 kinases
that constitute the human kinome. In addition, the chemical com-
pound might interact with some of the other ꢀ1500 ATP utilizing
enzymes and numerous other nucleotide-binding proteins present
in the human proteome. As a consequence many of the potential
targets are not evaluated using this method which, in addition, re-
quires time- and budget- consuming expression, purification and
assay set up for each individual kinase. Finally, in vitro assay
against recombinant enzymes may not adequately reflect the
distribution or the phosphorylation states found in vivo. This
* Corresponding authors. Fax: +33 472448319 (J.L.); fax: +33 298292526 (S.B.);
fax: +33 478898914 (P.G.).
E-mail addresses: jerome.lemoine@univ-lyon1.fr (J. Lemoine), bach@sb-roscoff.fr
(S. Bach), goekjian@univ-lyon1.fr (P. Goekjian).
0968-0896/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2009.06.024

C. Corbel et al. / Bioorg. Med. Chem. 17 (2009) 5572–5582
5573
enzymatic targets of aloisines. The N-methyl analogues (substitu-
tion of the hydrogen atom in position 5 by a methyl group) were
also prepared as negative controls as this substitution results in a
dramatic decrease in the protein kinase inhibitory activity.18
7
N
N
2
3
6
4
'
OH
N
R
'
3
2.2. Synthesis
Scheme 1. Structure of aloisine A (R = H) and N-methyl-aloisine A (R = Me).
2.2.1. Synthesis of aloisine A analogues with a linker at the 40-
OH position
method, although useful to give a first idea about the selectivity of
such ATP-competitive inhibitors, is thus clearly not comprehen-
sive. Using this approach, the exact mechanism of action and the
full spectrum of their targets remain largely unknown. Determina-
tion of the in vivo selectivity and identification of their intracellular
targets is a prerequisite in order to understand their cellular effects
and to improve their efficiency on a rational basis. Different ap-
proaches have therefore been developed to study the spectrum of
in vivo intracellular targets.10,11 In particular the purification of
intracellular targets by affinity chromatography using immobilized
drugs12 has been applied successfully notably to the study of pur-
valanol,13 paullones,14 hymenialdisine,15 Iressa16 and roscovitine.17
In this paper, we report the results of our investigation on the
targets of aloisine A. Aloisine A (7-n-butyl-6-(40-hydroxyphenyl)-
The synthesis of the pyrrolo[2,3-b]pyrazine core is typically per-
formed by condensation between an alkylpyrazine and a benzoni-
trile derivative.20 Introduction of the linker in the 40-OH position
can be performed either before or after the condensation step. As
the condensation is the limiting step and proceeds in moderate
yield, we decided to introduce the linker at an early stage (Scheme
2). Therefore, 4-hydroxybenzonitrile was reacted with the mono-
tosylated triethylene glycol 1 in the presence of potassium carbon-
ate in DMF to obtain 2 in 75% yield. The aloisine core was then
prepared by condensation of 2 with n-pentylpyrazine according
to the literature with a slight modification. Indeed, initial attempts
of condensation provided the expected product in low yield. Fur-
ther investigations revealed that an excess of LDA was required
for the formation of the pyrrolo[2,3-b]pyrazine ring 3 in acceptable
yield, presumably via to the formation of a dianion intermediate.21
The protected alcohol was then converted into an amine suitable
for coupling with the agarose-based matrix. To this end, compound
3 was treated with a catalytic amount of p-toluenesulfonic acid (p-
TSA) in methanol to afford the corresponding alcohol. Iodination by
treatment with triphenylphosphine/imidazole/iodine in DMF fol-
lowed by displacement with sodium azide gave the corresponding
alkyl azide which, in turn, yielded the primary amine 5 upon
hydrogenation. Methylation of 3 with methyl iodide and sodium
hydride gave the N-methyl analog 4, which after an identical se-
quence of reactions, gave the N-methyl control compound 6 in sim-
ilar yields.
5H-pyrrolo[2,3-b] pyrazine) is
a
potent inhibitor of CDKs,18
whereas its N-methyl derivative is completely inactive (Scheme
1). In vitro assays have shown that aloisine A is highly selective
for CDK1/cyclin B, CDK2/cyclin A-E, CDK5/p25 and GSK-3, the
two latter enzymes being implicated in Alzheimer’s disease.
Moreover this drug displays antiproliferative activity thus war-
ranting further investigation of its selectivity for a better under-
standing of its cellular and physiological action and optimization
of its pharmacological properties. To this aim, we have designed
aloisine A analogues conjugated with a linker suitable for immobi-
lization onto a matrix and used them for the purification of the
cellular targets of aloisine A. Finally, we have investigated the
activity of aloisine A against targets from simpler organisms, such
as budding yeast cdk’s and the Plasmodium falciparum homologue
of GSK-3.
2.2.2. Synthesis of aloisine A analogues with a linker at the C-7
position
In order to introduce the linker at the C-7 position, the strategy
was again to synthesize a pyrazine core conjugated with the TEG,
followed by condensation with the appropriate benzonitrile deriv-
ative. Synthesis of the pyrazinyl derivative was achieved by alkyl-
ation between 2-methyl pyrazine and the chain 722 to afford the
desired pyrazinyl derivative 9 in 46% yield (Scheme 2). However,
attempts at condensation with 4-methoxybenzonitrile gave yields
between 5% and 15%. We reasoned that these disappointing results
may be due to a complexation of the lithium by the TEG chain (e.g.,
structure A), thus hampering its reaction with the benzonitrile
derivative. We therefore decided to synthesize a longer chain with
the hope to prevent this complexation phenomenon. Briefly, com-
pound 114 was substituted with 1-pentenol in the presence of so-
dium hydride in DMF to give the corresponding ether in 56%
yield. Ozonolysis, reduction with sodium borohydride, tosylation,
and iodination gave the extended chain iodide 8, which was subse-
quently converted to the pyrazine 10. We were delighted to ob-
serve that in this case, the expected pyrrolo[2,3-b]pyrazine core
11 was formed in 45% yield. The amino-functionalized aloisine
derivative 12 was obtained by following the same route as de-
scribed above. The N-methyl analogue was not prepared in this
case.
2. Results
2.1. Design
A triethylene glycol (TEG) chain was selected as a linker for
immobilization due to its hydrophilic properties which allow a
good presentation of the inhibitor with minimal unfavorable fold-
ing on the matrix and non-specific adsorption of proteins. Further-
more, the linker was introduced to avoid possible steric hindrance
between target(s) and the matrix.12 From a chemical point of view,
either the NH of the pyrrolo[2,3-b]pyrazine ring or the hydroxyl
group of the phenol moiety can be functionalized directly with
the linker. The crystal structure of aloisine A complexed to CDK2/
5 shows that the NH group is involved in a key hydrogen-bond
with the protein, whereas the 40-OH group points out of the ATP
binding cleft and is not involved in any specific interactions.18,19
The butyl chain at the C-7 position is pointing toward a small
hydrophobic pocket, but may be accessible in other target proteins.
Crystal structure and QSAR studies have also shown that modifica-
tions on the C-2 and C-3 of the pyrazine core lead to a decrease of
the inhibition activity of the aloisine A.
We thus decided to introduce a linker on the positions 40, 30, and
7 for our investigations. Although one can anticipate a slight per-
turbation in the biological activity of aloisine A by such modifica-
tions, the combined affinity profiles of the differently linked
compounds should give a fairly complete view of the potential
2.2.3. Synthesis of aloisine A analogues with a linker at the C-30
position
Our initial strategy was to introduce the linker by a Sonogashira
coupling of an alkynyl-substituted triethyleneglycol chain with a
bromobenzonitrile derivative, followed by condensation with

5574
C. Corbel et al. / Bioorg. Med. Chem. 17 (2009) 5572–5582
O
TsO
THP
3
1
N
N
N
iv-vii
O
O
ii
i
O
O
NH2
THP
NC
OR
N
3
2
N
N
R
R
3 : R = H
4 : R = Me
R = H
R = (CH2CH2O)3THP
iii
2 :
5 : R = H
6 : R = Me
O
O
N
N
N
O
n
OTHP
Li
+
I
O
RO
m
N
viii
7: n = 1, m = 1
8: n = 3, m = 2
A
O
3
O
O
O
n
NH2
THP
O
3
m
N
N
N
N
x-xiii
OTHP
ix
OMe
OMe
N
H
N
H
9: n = 2, m = 1
10: n = 4, m = 2
N
N
11
12
N
N
N
N
N
xx-xxi
NC
OTHP
X
OTHP
R
xvii
OH
xv
OTHP
N
N
R
N
N
H
R
N
N
N
13: X = Br
14: X = CCSiMe3
15: R = SiMe3
16: R = H
N
xiv
xvi
N
N
O
H2N
O
X
O
O
20: R = H
21: R = Me
17: R = H, X = OTs
18: R = H, X = N3
19: R = Me, X = N3
xviii
xix
Scheme 2. Synthesis of aloisine A and N-methyl aloisine A analogs bearing sidechains for conjugation. (i) K2CO3, 1, DMF, 60 °C (75%); (ii) LDA, n-pentylpyrazine, THF, ꢁ40 °C
(55%); (iii) CH3I, NaH, DMF (67%); (iv) p-TSA, MeOH, rt (83% R = H, 86% R = CH3); (v) I2, PPh3, imidazole, DMF, 60 °C (40% R = H, 39% R = CH3); (vi) NaN3, DMF, 70 °C (70% R = H,
69% R = CH3); (vii) H2, 10%Pd/C, MeOH, rt (62% R = H, 66% R = CH3); (viii) 2-methylpyrazine, LDA, THF, ꢁ45 °C (46%); (ix) LDA, THF, ꢁ45 °C, then 4-methoxybenzonitrile (45%);
(x) TsOH, MeOH, rt (83%); (xi) I2, TPP, imidazole, DMF, 70 °C (39%); (xii) NaN3, DMF, 70 °C (73%); (xiii) H2, 10%Pd/C, MeOH, rt (52%); (xiv) PPh3, PdCl2(PPh3)2,
(trimethylsilyl)acetylene, TEA, CuI, THF (77%); (xv) 14, LDA, n-pentylpyrazine, THF, ꢁ40 °C; (xvi) TBAF, THF (84% for two steps); (xvii) N3CH2(CH2OCH2)2CH2OTs, DIPEA,
MeCN, CuI, rt (86%); (xviii) NaN3, DMF, 70 °C (77%); (xix) CH3I, Cs2CO3, DMF (85%); (xx) p-TSA, MeOH (84% R = H, 83% R = CH3); (xxi) H2, 10%Pd/C, MeOH, rt (64% R = H, 73%
R = CH3).
n-pentylpyrazine. However, all attempts at the Sonogashira reac-
tion with a triethylene glycol-bearing acetylene failed as only the
starting materials and/or a diyne product were recovered, presum-
ably due to the incompatibility of the reaction with the TEG chain.
Bearing in mind the challenges associated with the synthesis of
pyrrolo[2,3-b]pyrazine core in the presence of a TEG chain, we
turned our attention toward a more efficient conjugation method
to introduce the linker. Amongst the common methods reported,
the copper(II)-catalyzed 1,3-dipolar cycloaddition between an al-
kyne and an azide (so-called ‘click’ chemistry) was investigated.
To this aim, the bromobenzonitrile 13 was reacted under Sono-
gashira conditions with (trimethylsilyl)acetylene to afford the
3-alkynylbenzonitrile 14 in 77% yield. Condensation of 14 with
pentylpyrazine using the optimized conditions gave a mixture of
15 and the desilylated product 16. Further treatment of this mix-
ture with TBAF allowed obtention of pure 16 in 84% over two steps.
The cycloaddition step with the tosylate of an azido-substituted
triethylene glycol23 was then performed using CuI and DIPEA in
acetonitrile at room temperature and afforded the desired product
17 in 86% yield. Finally, displacement of the tosyl with sodium
azide, cleavage of the 40-OTHP with p-TSA, and reduction of the
azide gave the amino-terminated linked aloisine 20. Using the
same strategy, the N-methyl analogue 21 was obtained by methyl-
ation of 18 prior to reduction.
Incorporating the chain at a late stage by click chemistry over-
comes the problems associated with the combination of a basic
and hygroscopic pyrrolo[2,3-b]pyrazine core and
a chelating,
hydrophilic TEG, responsible for the often poor yields obtained pre-
viously. In view of the well-documented functional group compat-
ibility of both the Sonogashira reaction and the ‘click’ chemistry, as

C. Corbel et al. / Bioorg. Med. Chem. 17 (2009) 5572–5582
5575
Table 1
well as the high yields obtained, this latter approach provides a
more satisfactory solution for the incorporation of TEG chains onto
inhibitor structures for the purposes of affinity chromatography.24
Before using these PEG extension-linked analogues for affinity
chromatography analysis, their inhibitory efficacy on CDK1/Cy-
Kinase inhibition by aloisine A and analogues
Compound tested
CDK1/cyclinB
Protein kinase IC50 (lM)
CDK5/p25
GSK-3
Aloisine A
N-Methyl-aloisine A
5
6
20
21
0.15
140
0.23
100
6.7
0.20
160
0.66
50
>10
>10
0.65
>200
0.34
45
5.8
>10
clinB, CDK5/p25 and GSK-3a/b was tested and compared to
aloisine A. The results obtained are presented in Table 1. Introduc-
tion of the linker at the 40-OH of aloisine A (analogue 5) did not
greatly modify the CDK1, CDK5 and GSK-3 inhibition properties.
In the case of an addition of the linker at the C-30 position of aloi-
>10
MeAl(6)
Al(5)
N
N
A
N
O
O
O
O
N
O
O
N
H
N
Al(12)
MeAl(21)
Al(20)
N
N
N
OH
OH
O
N
N
N
H
O
O
N
N
N
N
N
N
N
OH
O
O
N
H
O
O
N
6
5
5
21 20
MW
(kDa)
B
MeAl Al Ald MeAl Al
191 -
Crude
extracts
6
5
5
21 20
MeAl Al
97 -
64 -
51 -
MeAl Al Ald
α
β2
β
GSK-3
Erk1
Erk2
39 -
28 -
Pyridoxal Kinase
(PDXK_PIG/046560)
CDK5
PDXK
G2
M
lysate
C
G2
M
MeAl Al MeAl Al
PSTAIRE
Crude
extracts
6
5
Al
D
MeAl
+
+
+
WT
WT
WT
PDXK
CDK1
Figure 1. Affinity chromatography purification of the targets of aloisine A. (A) Matrices used in this study. Methyl-aloisine control beads are indicated by MeAl followed by
the number small chemical compound immobilized (e.g., MeAl(6)). Aloisine A matrices are indicated by Al followed by the compound number (e.g., Al(5)). Different
compounds correspond to different positions of the linker chain. Extracts from porcine brain (B), and starfish G2 and M phase oocytes (C) HEK293 cells overexpressing (+) or
not (WT) the human PDXK (D) were incubated with immobilized-drug matrices. After extensive washing, the affinity matrix-bound proteins were resolved by SDS–PAGE. The
identification of the proteins detected after silver staining was realized by western blotting with specific antibodies or mass spectrometry in the case of PDXK. The Anti-
PSTAIRE antibody cross-reacts with CDK1, CDK2 and CDK3. Concerning the Ald lane, the porcine brain lysate was depleted on methyl-aloisine beads (MeAl(6)) before the
purification on aloisine beads (Al(5)).

5576
C. Corbel et al. / Bioorg. Med. Chem. 17 (2009) 5572–5582
sine A (analogue 20), the inhibition properties are significantly de-
creased. This modification is probably the consequence of steric
hindrance inside the ATP binding pocket of the selected kinases.
As already known,18 the substitution of the hydrogen atom in posi-
tion 5 by a methyl group results in a dramatic decrease in the
inhibitory activity of the analogues 6 and 21. The aloisine A ana-
logues 5 and 6 were thus chosen as the best tools for the initial
selectivity analysis by affinity chromatography. However, the
remaining analogues were used for identifying cellular targets that
bind in a different manner than the CDK’s.
subsequently removed in a series of washing steps, and specifically
bound proteins were eluted using SDS–PAGE sample buffer to dis-
rupt the interaction between the target proteins and the immobi-
lized small-molecule ligand. Finally, proteins were identified by
immunodetection (Fig. 1B) or by mass spectrometry (Fig. 2).
Methyl-aloisine is unable to form the canonical hydrogen bond
pair with the hinge region in the ATP-binding site of protein ki-
nases,25 and can thus be used in biological studies as a control
for those effects not related to protein kinases. It is therefore of
interest to focus on those bands that are absent from methyl-
aloisine control beads MeAl(6) but present on the aloisine A matrix
Al(5). Thus it is possible to reduce the complexity of the profile ob-
tained by depletion of the porcine brain lysates on control beads
(lane Ald). Under these experimental conditions, the purified pro-
teins are related to those biological activities of aloisine A that
would not be observed with the methyl-aloisine A as a negative
control. As expected, confirming the selectivity analysis of Mettey
2.2.4. Preparation of affinity chromatography matrices
In order to identify the cellular targets of aloisine A, the chem-
ical analogues substituted with a linker at the different positions
(compounds 5 and 6, 12, 20 and 21 in Scheme 2) were coupled
to sepharose beads to provide the matrices Al(5), MeAl(6),
Al(12), Al(20), and MeAl(21) (Fig. 1A). Coupling of the compounds
to the matrix were realized according to the recommendations of
the manufacturer (Sigma–Aldrich) and is described in the experi-
mental procedures section. Sepharose beads conjugated with eth-
anolamine were prepared as controls.
et al.18 using a panel of 26 different protein kinases, CDK5, GSK-3
a,
GSK-3b and b2 (a splicing isoform of GSK-3b26) and also ERK1 and
ERK2, were purified on aloisine A beads. Using starfish oocyte ly-
sates at G2 and M phases of the cell division cycle, we showed also
that CDK1 can bind to the aloisine A matrix Al(5) (Fig. 1C). Only a
very small amount of CDK1 was detected on the methyl-aloisine
matrix MeAl(6).
2.3. Identification of potential cellular targets of aloisine A
2.3.1. Purification of potential aloisine A cellular targets by
affinity chromatography
The immobilized small-molecule ligand matrices were first
incubated with porcine brain extracts. Unbound proteins were
Three other matrices, Al(12) and Al(20), and MeAl(21) sephar-
ose beads, were used in this study to evaluate the influence of
the linker position on the chromatographic process. The matrix
bearing the aloisine analog Al(12) with the linker on the C-7
MW
(kDa)
MeAl(6)Al(5)d
MeAl(6) beads
a. Clathrin heavy chain 1
(CLH1_Human) and Glycogen
Debranching Enzyme
(GDE_CANFA)
Al(5)d matrix
k. Clathrin heavy chain 1
(CLH1_Human)
191 -
97 -
k
l
a
b
c
d
e
64 -
51 -
b. Exportin 1 (XP01_HUMAN),
Exportin 7 (XP07_PONPY) and
Oxoglutarate dehydrogenase
(ODO1_Human)
l. Glycogen phosphorylase
(PYGM_BOVIN)
f
m
n
g
o
h
p
i
39 -
28 -
j
c. Exportin 2 (XP02_HUMAN)
m. Tubulin and GSK-3α
(GSK3a_HUMAN)
d. Glycogen phosphorylase
(PYGM_BOVIN)
n. GSK-3β (GSK-3b_HUMAN)
e. Fuse binding protein 1
(FUBP1_MOUSE)
o. MAPK3 (ERK1)
(MK03_MOUSE)
f. Tubulin beta chain (TBB_PIG)
g. Glutamine synthase (GLNA_PIG)
h. Glutamine synthase
p. MAPK1 (ERK2) (MK01_RAT)
(GLNA_PIG) and beta Actine
(ACTB_CERPY)
i. Glial Fibrillary acidic protein
(GFQP_BOVIN)
j. Pyridoxal kinase (PDXK_PIG)
Figure 2. Aloisine A can bind different protein kinases and pyridoxal kinase but also various other classes of cellular proteins. Porcine brain lysates were subjected to affinity
chromatography purification on MeAl(6) beads or on Al(5) beads. In the latter case, the lysate was first depleted on MeAl(6) methyl-aloisine beads as described in the legend
of Figure 1. Gel was stained and the major bands were excised and analyzed by mass spectrometry. The results of this analysis are reported in the table. The proteins were
identified using MALDI-TOF peptide mass fingerprint mapping, thus the accession codes indicate the closest protein sequences in the databases (the genome of the pig
(Sus scrofa) is currently only partially sequenced). Mowse scores, sequence coverages, calculated masses, and additional data using heavier loading are provided in the
Supplementary data.

C. Corbel et al. / Bioorg. Med. Chem. 17 (2009) 5572–5582
5577
position did not show any specific binding and gave results identi-
ATP (mM) on Al(5) beads
A
cal to the ethanolamine control (not shown). As shown in Figure 1,
the matrix Al(20) with a C-40 linker cannot bind CDK5 but can bind
GSK-3. This is most likely due to steric hindrance. Indeed, the study
of the binding mode of CDKs/GSK-3 inhibitors by co-crystallization
experiments have indicated that the major difference is located in
the entrance of the ATP-binding pocket and thus that GSK-3 can
accommodate bulkier compounds compared to CDK5.27,28 The re-
sult obtained for the inhibition of GSK-3 by 20 confirms this data
(see Table 1). Indeed compound 20 is still able to inhibit the
Eth
0
0.1
1
2.5
5
10
0
α
β2
β
GSK-3
Erk1
Erk2
PDXK
CDK5
GSK-3 kinase activity (IC50 = 5.8
lM) but less efficiently than com-
MeAl(6) beads
Al(5) beads
B
pound 5 (IC50 = 0.34 M). This indicates also that modification of
l
ATP (mM)
Pyridoxal (mM)
Pyridoxal (mM)
the structure at this position by addition of bulky group could be
used to increase the selectivity of aloisines for GSK-3. In addition,
a band that corresponds to a ꢀ37 kDa protein is detected on all
the matrices evaluated. This band was excised from the gel for
identification by mass spectrometry and corresponds to the por-
cine pyridoxal kinase, PDXK. We further confirmed this interaction
with human PDXK stably over-expressed in human HEK293 cells
(Fig. 1D). In this case, the high level of PDXK expression results
in reduced binding of CDK1, suggesting a competition for binding
to aloisine. PDXK may thus act as a sink for aloisine and cells over-
expressing PDXK may thus lose some sensitivity to aloisine. As ex-
pected, the methyl-aloisine gel MeAl(21) did not bind GSK-3, but
does bind PDXK.
0
7.5
15
0
7.5
15
10
α
β2
GSK-3
β
Erk1
Erk2
PDXK
CDK5
Figure 3. Analysis of the mode of interaction between cellular components and
affinity matrices by competition assays. Porcine brain extracts were supplemented
with increasing concentrations of ATP (A and B) or pyridoxal (B), prior to loading on
the indicated matrices (Al(5) or MeAl(6)).
2.3.3. Mode of binding of aloisine and methyl-aloisine to their
putative cellular targets
2.3.2. Identification of potential cellular targets by MALDI-TOF
mass spectrometry
In order to analyse the mode of binding of aloisine and methyl-
aloisine, we have analyzed the effect of increasing the concentra-
tions of ATP and pyridoxal in the brain extract, prior to and during
incubation with aloisine beads, on the binding of several protein
targets (Fig. 3). First, ATP is able to compete for the binding of
GSK-3 isoforms, ERK1/2, PDXK and CDK5 to Al(5) beads (Fig. 3A).
These results are consistent with kinetic studies, as well as the res-
olution of a CDK2-aloisine co-crystal structure, that demonstrate
that aloisine acts by competitive inhibition of ATP binding to the
catalytic subunit of protein kinase.18 In the case of PDXK the mode
of interaction is less predictable. ATP is able to compete for the
binding of PDXK to both Al(5) beads and MeAl(6) beads (Fig. 3A
and B). Pyridoxal can only compete for the binding of PDXK to
the MeAl(6) matrix (Fig. 3B) beads suggesting that methyl-aloisine
A and aloisine A have two different binding sites on pyridoxal
kinase.
Eleven bands obtained from the MeAl(6) and 12 bands obtained
from methylaloisine-depleted Al(5) matrices were excised and
analyzed by MALDI-TOF peptide mass fingerprint mapping. The
bands that yielded conclusive identifications are reported in Figure
2. All the mass spectrometry data obtained are reported in the S1
(for MeAl(6) matrix) and S2 and S3 (for Al(5) matrix) tables of
the Supplementary data section.
Methyl-aloisine A interacts with several classes of non-kinase
proteins. The methyl-aloisine A matrix MeAl(6) binds notably to:
(i) glycogen debranching enzyme that helps facilitate the break-
down of glycogen;29 (ii) oxoglutarate dehydrogenase, a key control
point in the citric acid cycle that uses NADP+ and coenzyme A as
co-factors;30 (iii) glycogen phosphorylase, catalyzes phosphoro-
lytic cleavage of the
a(1?4) glycosidic linkages of glycogen and
uses pyridoxal phosphate for acid/base catalysis by the enzyme,
which furthermore has a well characterized indole-binding site;31
iv) fuse binding protein 1 that has been shown to function as an
ATP-dependent DNA helicase;32 (v) glutamine synthase, an en-
zyme that plays an essential role in the metabolism of nitrogen
by catalyzing the condensation of glutamate and ammonia to form
glutamine in the presence of ATP.33
Similar analysis of protein bands on the parent aloisine matrix
Al(5) after depletion on the MeAl(6) matrix (Fig. 2) confirm the
immunoblot results in Figure 1: both GSK-3 and Erk isoforms
were identified by peptide mass fingerprint mapping. Removing
more abundant and less selective targets by depletion on the
MeAl(6) matrix will permit to better identify less abundant tar-
gets such as protein kinases, and to focus on those targets that
rely on the canonical ATP-competitive H-bonding pattern. How-
ever, no additional targets bound on the depleted Al(5) inhibitor
matrix were identified, with the exception of tubulin, glycogen
phosphorylase and clathrin heavy chain 1, all of which were also
identified on the MeAl(6) matrix.34 Heavier loading of the matrix
in order to study some of the minor bands led only to the iden-
tification of weakly binding, relatively abundant metabolic pro-
teins (Table S3 of the Supplementary data), but no additional
protein kinases.
2.4. Aloisine A inhibition of non-mammalian targets
2.4.1. Aloisine A inhibits two budding yeast CDKs
There are six CDKs in the budding yeast Saccharomyces cerevisiae.
Among these CDKs are Cdc28 and Pho85 that bind two distinct
classes of cyclins to perform different functions namely regulation
of the cell cycle transitions (Cdc28) or regulation of the cellular re-
sponse to nutrient levels and environmental conditions (Pho85).35
Pho85 is the yeast functional homologue of the mammalian
CDK536 and is the most closely related kinase in the yeast kinome
to Cdc28 (for review see37). Cdc28 is the yeast homologue of CDK1
and regulates each cell cycle transition in yeast.38 As shown in Fig-
ure 4A, both Pho85 and Cdc28 can be purified on aloisine matrix
from yeast cell lysates. As expected, methyl-aloisine cannot bind
to these CDKs. To confirm these targets, we have analyzed the inhi-
bition of Pho85 and Cdc28. An in vivo assay using the transcription
factor Pho4 tagged with GFP was used to analyse the effect on
Pho85 (the principle is described in the legend of Fig. 4). A
250
lM aloisine treatment induces a nuclear localization of the
GFP-tagged transcription factor Pho4 showing that Pho85 is

5578
C. Corbel et al. / Bioorg. Med. Chem. 17 (2009) 5572–5582
Eth MeAl Al EthMeAl Al
Eth MeAl Al
A
B
Pho85
Cdc28
α-PSTAIRE
α-Cdc28
α-Pho85
Inhibition
P
Diffuse
localisation
Nuclear
localisation
Pho85
Pho85
Pho4
Pho4
Methanol 5%
Aloisine 250µM
C
125
100
75
50
25
0
0.01
0.1
1
10
100
Me Aloisine
Aloisine A
Drug concentration (µM)
IC50(Aloisine A)=2.5 µM
IC50 (N-methyl-Aloisine A)>33µM
Figure 4. Aloisine A can bind and inhibit Cdc28 and Pho85 yeast CDKs. (A) Extracts from budding yeast Saccharomyces cerevisiae were incubated with immobilized-drug
matrices MeAl(6) and Al(5). A control matrix, sepharose beads quenched with ethanolamine, was used (Eth). After extensive washings, the affinity matrix-bound proteins
were resolved by SDS–PAGE. The identification of the proteins was realized by western blotting with specific antibodies. Note that both Pho85 and Cdc28 primary sequences
have the PSTAIRE motif. (B) The transcription factor Pho4 is one of the Pho85 substrates. Its phosphorylation by Pho85 triggers its export from the nucleus by the receptor
Msn5. Therefore, when Pho85 is inhibited, the absence of phosphate leads to a Pho4 nuclear localization. Using a GFP tag, it provides a sensitive fluorescence-based assay to
follow inhibition of Pho85 (assay described in39). The treatment of yeast cells with aloisine A (right panel) leads to a significant nuclear localization of Pho4-GFP. Note that the
treatment with methanol (solvent of the drug in this assay) does not modify the localization (left panel). (C) Cdc28/Clb2 activity was assayed in the presence of increasing
concentrations of methyl-aloisine and aloisine A. Kinase activities are expressed as a % of the maximal activity that is measured in the absence of inhibitor.
inactive in presence of the drug, whereas 5% methanol (solvent of
the drug used in this study) does not affect the localization. Con-
cerning Cdc28, the active kinase was purified by immunoprecipita-
tion using anti-HA antibody and was used for the determination of
the IC50 of both aloisine A and methyl-aloisine A (Fig. 4C). The IC50
the numerous and complex factors involved (permeability, efflux,
subcellular localization, etc.).39
2.4.2. Aloisine A inhibits the Plasmodium falciparum homolo-
gue of GSK-3
for aloisine A is 2.5
l
M, and, as expected, methyl-aloisine cannot
Increasing resistance worldwide of P. falciparum to common
anti-malaria agents calls for the urgent identification of new drugs.
The perspective is now to identify new molecular targets which
could be screened for selective and potent inhibitors of a physio-
logical process essential to the parasite. Among the estimated
6000 genes of P. falciparum, protein kinases constitute attractive
candidates as potential targets (see review in ref.40). Using
P. falciparum extracts obtained during the red blood cell stage of
the infection, the PfGSK-3 was bound onto aloisine A matrix
(Fig. 5A). PfGSK-3 shares a high degree of similarity with human
GSK-3.41 This direct interaction with aloisine matrix was
inhibit Cdc28/Clb2. Cdc28 is essential for yeast and inhibition of
its activity leads to cell cycle arrest and thus to growth defect that
can be easily observed on agar plate using halo assay.39 We have
tested the ability of aloisine A to inhibit the yeast growth. Even
using a drug-sensitized yeast strain, no growth defect was ob-
served following treatment of cells with aloisine A, suggesting that
aloisine A may not enter yeast cell very well. This can explain the
relative high concentration needed (250 lM) to observe a pheno-
type in the Pho85 functional assay. It remains difficult to precisely
evaluate the concentrations of small molecules within cells due to

C. Corbel et al. / Bioorg. Med. Chem. 17 (2009) 5572–5582
5579
6
5
A
MW
(kDa)
Eth MeAl Al
191 -
97 -
6
5
Eth MeAl Al
64 -
51 -
Pf GSK-3
39 -
28 -
B
C
6
5
Eth MeAl Al
Rec. Pf GSK-3
125
100
75
50
25
0
Pf
0.1
1
10
100
Me-Aloisine A
Drug concentration (µM)
Aloisine A
IC50(Aloisine A)=5.5 µM
IC50 (N-methyl-Aloisine A)>100µM
Figure 5. Aloisine A can bind and inhibit GSK-3 from Plasmodium falciparum. (A) Plasmodium falciparum lysates obtained during the red blood cell stage of the infection were
subjected to affinity chromatography purification on ethanolamine (Eth), methyl-aloisine (MeAl(6)) or aloisine (Al(5)) beads. The identification of the band corresponding to
PfGSK-3 on the silver-stained gel was realized by Western blotting with specific antibody. Note that all the major protein bands detected on aloisine A matrix by silver
staining of the polyacrylamide gel are also present on methyl-aloisine beads. (B) A recombinant PfGSK-3 can bind to the aloisine matrix. C, PfGSK-3 recombinant enzyme can
be inhibited by aloisine with an IC50 of 5.5
maximal activity.
lM. Note that methyl-aloisine A cannot inhibit recombinant PfGSK-3 enzymatic activity. Kinase activities are expressed in% of
confirmed using recombinant PfGSK-3 produced in bacteria (
Fig. 5B). PfGSK-3 recombinant enzyme can be inhibited by aloisine
aloisine can also bind to other kinds of cellular proteins and nota-
bly to PDXK.
with an IC50 of 5.5
itor of PfGSK-3 although it is 8.5-fold less potent than on mamma-
lian GSK-3 (Table 1).
l
M. Aloisine is thus considered as a new inhib-
These results show that PDXK could be considered as a frequent
off-target of small chemical inhibitors of CDKs. Indeed, a recent
study has shown the interaction of (R)-roscovitine with PDXK.17,42
This drug is frequently studied and used as a CDK inhibitor. Like
aloisine A, it inhibits protein kinases by direct competition with
ATP for binding to the catalytic cleft (reviewed in Ref. 43). Concern-
ing the interaction with PDXK, co-crystal structures showed that
roscovitine occupies the ATP-pyridoxal binding groove. Interest-
ingly, roscovitine is not directly binding at the ATP site, but at
the pyridoxal site.42 The authors showed that both ATP and pyri-
doxal can compete for the binding of PDXK to roscovitine affinity
matrix.17 Thus our data indicate that methyl-aloisine seems to
share the same mode of binding. In the case of aloisine A, we can
suggest that the small chemical compound may use the ATP
3. Discussion
Aloisine A (7-n-butyl-6-(40-hydroxyphenyl)-5H-pyrrolo[2,3-b]
pyrazine) is a potent inhibitor of different protein kinases. Prior
to this study, its specificity was known essentially from its
in vitro effect on fewer than 30 protein kinases. Such an assay only
reflects a part of the 518+ kinases that constitute the human ki-
nome, but indicates that aloisine A appears to be remarkably spe-
cific to CDK1/2/5 and GSK-3.18 The current study shows that
despite the relative selectivity of aloisine for CDKs and GSK-3,

5580
C. Corbel et al. / Bioorg. Med. Chem. 17 (2009) 5572–5582
binding site. A co-crystal structure of PDXK-aloisine A would
clearly shed light on the position of atoms involved in the interac-
tion. These data would be helpful to obtain derived structures of
aloisine that cannot bind to PDXK but still to protein kinases to
avoid potential side effects during preclinical/clinical studies. In-
deed, PDXK catalyses the phosphorylation of the three forms of
vitamin B6 (pyridoxine, pyridoxal and pyridoxamine) in the pres-
ence of ATP and Zn2+. This phosphorylation constitutes an essential
step in the synthesis of pyridoxal 50-phosphate, the intracellular
active form of vitamine B6, a key factor for at least 140 enzymes,
such as aminotransferases and decarboxylases. But, in the case of
the roscovitine, despite the interaction of this drug with PDXK,
its effects on the catalytic activity of PDXK are rather limited.
4. Conclusion
Potential applications of CDK/GSK-3 inhibitors are currently
evaluated against cancers, neurodegenerative disorders such as
Alzheimer’s disease, proliferation of protozoan parasites, and viral
infections (HIV, cytomegalovirus, and herpes virus). Using affinity
chromatography on immobilized drug, we show here the charac-
terization of the putative cellular targets of aloisine A. This com-
pound was initially selected for its ability to be a potent inhibitor
of CDK1/2/5 and GSK-3. This study indicates that despite the rela-
tive selectivity of aloisine for CDKs and GSK-3, aloisine can also
bind to other kind of cellular proteins and notably PDXK. Analysis
of various biological models indicates that aloisine selectivity pro-
file is well conserved from the yeast S. cerevisiae to human. Each of
these interactions must be carefully investigated in order to im-
prove the biological activity of the drug and also to evaluate and
avoid the possible side effects preliminary to further assays on ani-
mal models of human diseases involving CDK5 and/or GSK-3. How-
ever, our results indicate that methyl-aloisine can be used
effectively as a control for many of these non-kinase activities.
Experimental data indicate that only 100 lM of ATP is sufficient
to compete roscovitine out, therefore resulting in a weak effect
on PDXK activity.42
This study of putative cellular targets of aloisine and methyl-
aloisine using an affinity chromatography method has further
shown that these compounds can also interact with non kinase
proteins: for instance glycogen phosphorylase, glycogen-debran-
ching enzyme, oxoglutarate dehydrogenase, glutamine synthase,
and Fuse binding protein 1. We feel that these findings illustrate
the advantages of this method for identifying cellular targets of a
given family of compounds. Indeed, these results were unexpected
before this study and can only be obtained using such a compre-
hensive approach. However, these results must still be analyzed
carefully in order to evaluate the contribution of the binding of
aloisine A to these proteins to the pharmacological properties of
the drug, such as the inhibition of cell proliferation as shown by
Mettey et al.18 The use of methyl-aloisine as a negative control in
such studies, however, can help restrict the origin of the observed
effect to the protein kinases that were identified after depletion on
the MeAl(6) matrix (Ald), namely the CDK, ERK and GSK isoforms.
Using various models, from unicellular eukaryote to mamma-
lian cells and tissues, we have also shown the conservation of tar-
gets of aloisine A through evolution, as illustrated by the inhibition
of two budding yeast CDKs (Pho85 and Cdc28), and also the GSK-3
counterpart of P. falciparum.
The simultaneous evaluation of all classes of proteins expressed
in one tissue (from dehydrogenases to protein kinases) is one of the
strengths of this method. But as such, it also raises the question of
the relative expression of the putative targets. Indeed, this method
reflects the distribution of the inhibitor within target proteins in
the cell, meaning that highly abundant detected targets can mask
less abundant, though perhaps relevant candidates. In our case,
we chose to focus on those targets that do not bind the methyl-
aloisine negative control by depleting the cell lysates on the
methyl-aloisine matrix. For the most part these were limited to
the expected CDK, GSK-3 and ERK isoforms. It is particularly inter-
esting to note that the off-target activities of the aloisines are not
due to the typical ATP-competitive binding mode, as these targets
all bind equally well to methyl-aloisine.
5. Experimental procedures
5.1. Synthesis
Compounds 2–21 were synthesised by following the pathway
described in Scheme 2. Detailed synthetic procedures and spectro-
scopic data are available in the Supplementary data.
5.2. Affinity chromatography on immobilized drugs
Homogenization buffer: 60 mM b-glycerophosphate, 15 mM p-
nitrophenylphosphate, 25 mM Mops (pH 7.2), 15 mM EGTA,
15 mM MgCl2, 1 mM DTT, 1 mM sodium vanadate, 1 mM NaF,
1 mM phenylphosphate, 10
10 g soybean trypsin inhibitor/ml and 100
Bead buffer: 50 mM Tris pH 7.4, 5 mM NaF, 250 mM NaCl, 5 mM
EDTA, 5 mM EGTA, 0.1% NP-40, 10 g/ml of leupeptin, aprotinin
and soybean trypsin inhibitor and 100 M benzamidine.
l
g leupeptin/ml, 10
lg aprotinin/ml,
l
lM benzamidine.
l
l
5.2.1. Immobilization of drugs on sepharose matrix
The drug attached to the linker (e.g., 5 for active aloisine; 6 for
negative control N-methyl analogue) are coupled in 0.1 M NaHCO3,
0.2 M NaCl, pH 8.3 on CNBr-Activated Sepharose 4B (Sigma–Al-
drich, #C9142) under agitation at room temperature for 8 h. The
final calculated loading was 3 lmol/ml of resin. After the coupling
process, the remaining active groups are blocked in pH 8.0 buffer
containing 1 M ethanolamine, under agitation for 3 h at 20 °C.
The beads are then washed three times with bead buffer and stored
in bead buffer (supplemented with 0.05% NaN3 to avoid microbial
growth) at 4 °C until use.
It is also worth pointing out here that while these matrices
could be used to provide a comprehensive determination of all cel-
lular targets (preferably using human tissues), it is not the objec-
tive of the current study. Larger amounts of proteins and matrix
associated with separation of the putative targets using two-
dimensional gel electrophoresis have been used to increase the
sensitivity of the analysis in the study of chemical inhibitors.44,45
Instead, we aim here to provide only those new targets that were
identified conclusively, and an insight into the distribution of the
compound among its principal cellular targets. Furthermore, in
conjunction with other such studies,13–17,46 we hope to identify
those recurring targets, such as PDXK and NADH-dependent
enzymes, that should be included systematically in the panel of
enzymes used for the study of protein kinase inhibitor selectivity.
5.2.2. Preparation of extracts
Pork brains were obtained from a local slaughterhouse and di-
rectly homogenized and processed for affinity chromatography or
stored at ꢁ80 °C prior to use. Tissues were weighed, homogenized
and sonicated in homogenization buffer (2 ml per g of material).
Homogenates were centrifuged for 10 min at 14,000g at 4 °C. The
supernatant was recovered, assayed for protein content (Bio-Rad
protein assay) and immediately loaded batch wise on the affinity
matrix. Yeast S. cerevisiae extracts were prepared using a ‘One Shot’
cell disrupter (Constant Systems LTD, Daventry, England).
5.2.3. Affinity chromatography of aloisine-interacting proteins
Just before use, 10
ll of packed aloisine beads were washed
with 1 ml of bead buffer and resuspended in 600
ll of this buffer.

C. Corbel et al. / Bioorg. Med. Chem. 17 (2009) 5572–5582
5581
The cell or tissue extract supernatant (3 mg total protein) or puri-
fied protein was then added; the tubes were rotated at 4 °C for
30 min. After a brief spin at 10,000g and removal of the superna-
tant, the beads were washed 4 times with bead buffer before
addition of 60 ll 2X Laemmli sample buffer. Following heat dena-
turation for 3 min, the bound proteins were analyzed by SDS–PAGE
5.4.2. Preparation of extracts
Yeast S. cerevisiae were grown to OD600nm ꢃ 2 (exponential
phase). yeast pellet is then obtained by centrifugation
A
(3000 rpm, 15 min) and diluted in breaking buffer (25 mM Tris
pH 7.4, 100 mM NaCl, 0.2% Triton X100, 15 mM p-nitrophenyl-
phosphate, 0.1 mM Na-Vanadate, 1 mM PMSF and protease inhib-
itor cocktail (CompleteTM, Roche)). Cells were broken using a ‘One
Shot’ cell disrupter (Constant Systems LTD, Daventry, England)
with a 2.5 kbar pressure. Extracts were clarified by centrifugation
(13,000 rpm, 20 min) before protein quantification using a Brad-
ford assay.
and western blotting or silver staining as described below.
5.2.4. Electrophoresis, western blotting, and silver staining
Following heat denaturation for 3 min, the proteins bound to
the aloisine matrices were separated by 10% SDS–PAGE (1 mm
thick gels) followed by immunoblotting analysis or silver staining
using a GE Healthcare SDS–PAGE silver staining kit. For immuno-
5.4.3. Immunoprecipitation
blotting, proteins were transferred to 0.45
l
m nitrocellulose filters
Protein A/G PLUS-agarose (Santa Cruz, SC-2003) and monoclo-
nal anti-HA (Covance, MMS-101R) were used for immunoprecipi-
tation of the complex Cdc28/Clb2-HA before kinase assay.
(Schleicher and Schuell). These were blocked with 5% low fat milk
in Tris-Buffered Saline—Tween-20, incubated for 1 h with antibod-
ies and analyzed by Enhanced Chemiluminescence (ECL, GE
Healthcare). The following antibodies were used: anti-GSK-3
(1:1000, Stressgen), anti-CDK5 (1:500; Santa Cruz), anti-ERK1/2
(1:4000; Sigma), anti-p35/p25 (1:500; Santa Cruz), anti-PSTAIRE
(1:3000, Sigma), anti-PDXK (1:500;17), anti-Pho85 and anti-Cdc28
(respectively given by E. O’Shea and C. Mann).
5.4.4. Kinase assays
Purified Cdc28/Clb2-HA complex was incubated for 30 min at
30 °C in a 30
(1 mg/ml), 6
l of [ -33P]ATP (15
tions of drugs in buffer (60 mmol/l b-glycerophosphate,
l
l reaction mix containing 5
l
l of Histone H1
c
l
lmol/l) and varying concentra-
C
15 mmol/l p-nitrophenylphosphate, 25 mmol/l MOPS (pH, 7.2),
5 mmol/l EGTA, 15 mmol/l MgCl2, 1 mmol/l DTT, 1 mmol/l sodium
vanadate, 1 mmol/l phenylphosphate). After 30 min incubation at
5.2.5. Protein identification by mass spectrometry
Silver stained protein bands were excised from the correspond-
ing 1D gel. In-gel digestion was carried out using porcine trypsin
from Promega (Madison, WI, USA): Gel fragments were washed
30 °C, 20 ll of supernatant were spotted onto pieces of Whatman
P81 phosphocellulose paper. The filters were washed three times
(for at least 5 min each time) in a 1% phosphoric acid solution.
The wet filters were counted in the presence of 1 ml ACS (GE
Healthcare) scintillation fluid.
with 2 ꢂ 200
with CH3CN, and dehydrated under vacuum. Then, 10
solution (12.5 ng/ l in 25 mM ammonium hydrogen carbonate)
l
l of 25 mM ammonium hydrogen carbonate, dried
l
l of trypsin
l
was added and the digestion was carried out overnight at 37 °C.
After separation of the supernatant, gel fragments were washed
5.4.5. Pho4-GFP assay
YRP1 carrying EB0347 plasmid were grown over-night under
selection (w/o uracil) then diluted at OD600nm = 0.5. The culture is
treated with 250 lM aloisine A. Samples are observed at 3 h and
analyzed with fluorescence microscopy: GFP-tagged proteins were
visualized using a Chroma GFP4 filter (excitation 455–495 nm) on
an Olympus BX61 microscope and photographed with a Diagnostic
Instruments Spot RT cooled CCD camera.
twice with 50 ll of H2O/ACN (v/v) containing 0.5% formic acid.
The fractions were concentrated under vacuum and stocked at
ꢁ20 °C. Prior to analysis on an Ultraflex MALDI-TOF-TOF instru-
ment (Bruker Daltonics, Bremen), the peptides were diluted in
20
ll of H2O/ACN (v/v) containing 0.5% formic acid. One microliter
of the peptide solution was mixed directly onto the target with 1
ll
of DHB matrix (10 mg/mL of DHB in H2O/CH3OH (v/v)). All peptide
mass fingerprint spectra were internally calibrated with the tryptic
auto-digestion ions. Identification of proteins was achieved using
the Mascot algorithm. Only those bands yielding a correlation fac-
tor above 60 and a genomic predicted mass that is consistent with
the mass observed on the gel are shown in Figure 2. Mowse scores,
sequence coverages, calculated masses, and additional data using a
heavier loading are provided in the Supplementary data.
Acknowledgments
We thank Dr. E. O’Shea for Pho4-GFP expressing vector and
Pho85 antibody, Dr. K. Kuchler for yeast highly permeable strain
YRP1 and Dr. M.N. Simon for yeast strains expressing Clb2-HA pro-
tein. Special thanks to Dr. C. Mann for Cdc28 antibody and also for
helpful suggestions. C.C. and D.G. are doctoral research fellows
supported by the ‘Région Bretagne—CNRS Département des Sci-
ences du Vivant’ and ‘Ligue contre le Cancer, Comité des Côtes d’Ar-
mor—Association pour la Recherche contre le Cancer’, respectively.
S.B. is supported by ‘Association pour la Recherche contre le Can-
cer’ (Contract ARC No. 3889). This research was also supported
by grants from the EEC (FP6-LSHB-CT-2004-503467, PRO-KINASE
Research (R.H., L.M., P.G.) and FP6-LSH-2003-2.3.0_2, ANTIMAL
(L.M.)) and the Canceropole Grand-Ouest (L.M.).
5.3. Protein kinase assays
The description of protein kinase preparations and assays
(CDK1/cyclin B, CDK5/p25, GSK-3 and PfGSK-3) is provided in the
Supplementary data section.
5.4. Biochemical and genetic experiments using the yeast S.
cerevisiae
Supplementary data
5.4.1. Strains and plasmids
MNY629 is a gift from Marie-Noëlle Simon (IGC, Marseille,
France) and allows the expression of CLB2 fused to HA-tag. YRP1
is a gift from Karl Kuchler (Vienna University, Austria) and corre-
Supplementary data (The synthesis and characterization data of
compounds 2–21) associated with this article can be found, in the
online version, at doi:10.1016/j.bmc.2009.06.024.
sponds to
a drug-sensitized yeast strain (snq2
Dpdr5Derg6D).
EB0347 is a gift from Erin O’Shea (Howard Hughes Medical Insti-
tute, Harvard University, Cambridge, MA, USA). It contains PHO4
promoter and the entire PHO4 ORF fused in-frame to N terminus
of GFP on pRS316 plasmid.
References and notes
1. Dhavan, R.; Tsai, L. H. Nat. Rev. Mol. Cell Biol. 2001, 2, 749.
2. Malumbres, M.; Ortega, S.; Barbacid, M. Biol. Chem. 2000, 381, 827.

5582
C. Corbel et al. / Bioorg. Med. Chem. 17 (2009) 5572–5582
3. Morgan, D. O. Annu. Rev. Cell Dev. Biol. 1997, 13, 261.
27. Fischer, P. M. Chem. Biol. 2003, 10, 1144.
4. Knockaert, M.; Greengard, P.; Meijer, L. Trends Pharmacol. Sci. 2002, 23, 417.
5. Mazanetz, M. P.; Fischer, P. M. Nat. Rev. Drug Discovery 2007, 6, 464.
6. Cohen, P. Nat. Rev. Drug Discovery 2002, 1, 309.
7. Fischer, P. M.; Lane, D. P. Curr. Med. Chem. 2000, 7, 1213.
8. Gray, N.; Detivaud, L.; Doerig, C.; Meijer, L. Curr. Med. Chem. 1999, 6, 859.
9. Davies, S. P.; Reddy, H.; Caivano, M.; Cohen, P. Biochem. J. 2000, 351, 95.
10. Bach, S.; Blondel, M.; Meijer, L.. In Monographs on Enzyme Inhibitors; CRC Press:
Boca Raton, FL, 2006; Vol. 2, p 103.
28. Meijer, L.; Skaltsounis, A. L.; Magiatis, P.; Polychronopoulos, P.; Knockaert, M.;
Leost, M.; Ryan, X. P.; Vonica, C. A.; Brivanlou, A.; Dajani, R.; Crovace, C.;
Tarricone, C.; Musacchio, A.; Roe, S. M.; Pearl, L.; Greengard, P. Chem. Biol. 2003,
10, 1255.
29. Watanabe, Y.; Makino, Y.; Omichi, K. J. Biochem. 2008, 143, 435.
30. Araujo, W. L.; Nunes-Nesi, A.; Trenkamp, S.; Bunik, V. I.; Fernie, A. R. Plant
Physiol. 2008, 148, 1782.
31. Lukacs, C. M.; Oikonomakos, N. G.; Crowther, R. L.; Hong, L.-N.; Kammlott, R. U.;
Levin, W.; Li, S.; Liu, C. M.; McGady, D. L.; Pietranico, S.; Reik, L. Proteins:
Structure, Function and Bioinformatics 2006, 63, 1123.
11. Terstappen, G. C.; Schlupen, C.; Raggiaschi, R.; Gaviraghi, G. Nat. Rev. Drug
Discovery 2007, 6, 891.
12. Guiffant, D.; Tribouillard, D.; Gug, F.; Galons, H.; Meijer, L.; Blondel, M.; Bach, S.
Biotechnol. J. 2007, 2, 68.
32. Chung, H. J.; Liu, J.; Dundr, M.; Nie, Z.; Sanford, S.; Levens, D. Mol. Cell. Biol.
2006, 26, 6584.
13. Knockaert, M.; Gray, N.; Damiens, E.; Chang, Y. T.; Grellier, P.; Grant, K.;
Fergusson, D.; Mottram, J.; Soete, M.; Dubremetz, J. F.; Le Roch, K.; Doerig, C.;
Schultz, P.; Meijer, L. Chem. Biol. 2000, 7, 411.
14. Knockaert, M.; Wieking, K.; Schmitt, S.; Leost, M.; Grant, K. M.; Mottram, J. C.;
Kunick, C.; Meijer, L. J. Biol. Chem. 2002, 277, 25493.
33. Berlicki, L. Mini.-Rev. Med. Chem. 2008, 8, 869.
34. Reprocessing of the data with a recent version of the database led to the
possible identification of MEK-1 (Table S2 of the Supplementary data).
However, Mettey et al. have shown previously that this kinase is not a target
of aloisine (Ref. 18).
15. Wan, Y.; Hur, W.; Cho, C. Y.; Liu, Y.; Adrian, F. J.; Lozach, O.; Bach, S.; Mayer, T.;
Fabbro, D.; Meijer, L.; Gray, N. S. Chem. Biol. 2004, 11, 247.
35. Wu, D.; Dou, X.; Hashmi, S. B.; Osmani, S. A. J. Biol. Chem. 2004, 279,
37693.
16. Brehmer, D.; Greff, Z.; Godl, K.; Blencke, S.; Kurtenbach, A.; Weber, M.; Muller,
S.; Klebl, B.; Cotten, M.; Keri, G.; Wissing, J.; Daub, H. Cancer Res. 2005, 65, 379.
17. Bach, S.; Knockaert, M.; Reinhardt, J.; Lozach, O.; Schmitt, S.; Baratte, B.; Koken,
M.; Coburn, S. P.; Tang, L.; Jiang, T.; Liang, D. C.; Galons, H.; Dierick, J. F.; Pinna,
L. A.; Meggio, F.; Totzke, F.; Schachtele, C.; Lerman, A. S.; Carnero, A.; Wan, Y.;
Gray, N.; Meijer, L. J. Biol. Chem. 2005, 280, 31208.
36. Huang, D.; Patrick, G.; Moffat, J.; Tsai, L. H.; Andrews, B. Proc. Natl. Acad. Sci.
U.S.A. 1999, 96, 14445.
37. Huang, D.; Friesen, H.; Andrews, B. Mol. Microbiol. 2007, 66, 303.
38. Ceccarelli, E.; Mann, C. J. Biol. Chem. 2001, 276, 41725.
39. Kung, C.; Kenski, D. M.; Krukenberg, K.; Madhani, H. D.; Shokat, K. M. Chem.
Biol. 2006, 13, 399.
18. Mettey, Y.; Gompel, M.; Thomas, V.; Garnier, M.; Leost, M.; Ceballos-Picot, I.;
Noble, M.; Endicott, J.; Vierfond, J. M.; Meijer, L. J. Med. Chem. 2003, 46, 222.
19. Mapelli, M.; Massimiliano, L.; Crovace, C.; Seeliger, M. A.; Tsai, L. H.; Meijer, L.;
Musacchio, A. J. Med. Chem. 2005, 48, 671.
20. Vierfond, J. M.; Mettey, Y.; Mascrier-Demagny, L.; Miocque, M. Tetrahedron Lett.
1981, 22, 1219.
21. Davis, M.; Wakefield, B.; Wardell, J. Tetrahedron 1992, 48, 939.
22. Fuchter, M.; Beall, L.; Baum, S.; Montalban, A.; Sakellariou, E.; Mani, N.; Miller,
T.; Vesper, B.; White, A.; Williams, D.; Barrett, A.; Hoffman, B. Tetrahedron
2005, 61, 6115.
23. Eduardo, F. M.; Correa, J.; Irene, R. M.; Ricardo, R. Macromolecules 2006, 39,
2113.
40. Doerig, C.; Meijer, L.; Mottram, J. C. Trends Parasitol. 2002, 18, 366.
41. Droucheau, E.; Primot, A.; Thomas, V.; Mattei, D.; Knockaert, M.; Richardson,
C.; Sallicandro, P.; Alano, P.; Jafarshad, A.; Baratte, B.; Kunick, C.; Parzy, D.;
Pearl, L.; Doerig, C.; Meijer, L. Biochim. Biophys. Acta 2004, 1697, 181.
42. Tang, L.; Li, M. H.; Cao, P.; Wang, F.; Chang, W. R.; Bach, S.; Reinhardt, J.;
Ferandin, Y.; Galons, H.; Wan, Y.; Gray, N.; Meijer, L.; Jiang, T.; Liang, D. C. J. Biol.
Chem. 2005, 280, 31220.
43. Meijer, L.; Raymond, E. Acc. Chem. Res. 2003, 36, 417.
44. Godl, K.; Wissing, J.; Kurtenbach, A.; Habenberger, P.; Blencke, S.; Gutbrod, H.;
Salassidis, K.; Stein-Gerlach, M.; Missio, A.; Cotten, M.; Daub, H. Proc. Natl. Acad.
Sci. U.S.A. 2003, 100, 15434.
45. Brehmer, D.; Godl, K.; Zech, B.; Wissing, J.; Daub, H. Mol. Cell. Proteomics 2004,
3, 490.
46. Trapp, J.; Jochum, A.; Meier, R.; Saunders, L.; Marshall, B.; Kunick, C.; Verdin, E.;
Goekjian, P.; Sippl, W.; Jung, M. J. Med. Chem. 2006, 49, 7307.
24. Haddoub, R.; Gueyrard, D.; Goekjian, P. Tetrahedron Lett. 2009, 50, 741.
25. Traxler, P.; Furet, P. Pharmacol. Therapeut. 1999, 82, 195.
26. Mukai, F.; Ishiguro, K.; Sano, Y.; Fujita, S. C. J. Neurochem. 2002, 81, 1073.