Potent inhibitors of CDK5 derived from roscovitine: Synthesis, biological evaluation and molecular modelling

Article abstract of DOI:10.1016/j.bmcl.2012.10.141

Cyclin dependent kinase 5 (CDK5) is a serine/threonine kinase belonging to the cyclin dependent kinase (CDK) family. CDK5 is involved in numerous neuronal diseases (including Alzheimer's or Parkinson's diseases, stroke, traumatic brain injury), pain signaling and cell migration. In the present Letter, we describe syntheses and biological evaluations of new 2,6,9-trisubstituted purines, structurally related to roscovitine, a promising CDK inhibitor currently in clinical trials (CDK1/Cyclin B, IC₅₀ = 350 nM; CDK5/p25, IC ₅₀ = 200 nM). These new molecules were synthesized using an original Buchwald-Hartwig catalytic procedure; several compounds (3j, 3k, 3l, 3e, 4k, 6b, 6c) displayed potent kinase inhibitory potencies against CDK5 (IC₅₀ values ranging from 17 to 50 nM) and showed significant cell death inducing activities (IC₅₀ values ranging from 2 to 9 μM on SH-SY5Y). The docking of the inhibitors into the ATP binding domain of the CDK5 catalytic site highlighted the discriminatory effect of a hydrogen bond involving the CDK5 Lys-89. In addition, the calculated final energy balances for complexation measured for several inhibitors is consistent with the ranking of the IC ₅₀ values. Lastly, we observed that several compounds exhibit submicromolar activities against DYRK1A (dual specificity, tyrosine phosphorylation regulated kinase 1A), a kinase involved in Down syndrome and Alzheimer's disease (3g, 3h, 4m; IC₅₀ values ranging from 300 to 400 nM).

Full text of DOI:10.1016/j.bmcl.2012.10.141

Bioorganic & Medicinal Chemistry Letters 23 (2013) 125–131
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Potent inhibitors of CDK5 derived from roscovitine: Synthesis, biological
evaluation and molecular modelling
Luc Demange a, , Fatma Nait Abdellah a, Olivier Lozach b, Yoan Ferandin b, Nohad Gresh a,
⇑
c,d
Laurent Meijer b,c, , Hervé Galons
⇑
a Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques (LCBPT), UMR 8601 CNRS, Université Paris Descartes, Sorbonne Paris Cité, UFR Biomédicale des Saints
Pères, 45 rue des Saints Pères, 75270 Paris cedex 06, France
b CNRS, USR3151, ‘Protein Phosphorylation & Human Disease’ group, Station Biologique, 29680 Roscoff, Bretagne, France
c ManRos Therapeutics, Hôtel de Recherche, Centre de Perharidy, 29680 Roscoff, France
d Laboratoire de Pharmacochimie, Université Paris Descartes, 4, avenue de l’observatoire, 75006 Paris, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Cyclin dependent kinase 5 (CDK5) is a serine/threonine kinase belonging to the cyclin dependent kinase
(CDK) family. CDK5 is involved in numerous neuronal diseases (including Alzheimer’s or Parkinson’s dis-
eases, stroke, traumatic brain injury), pain signaling and cell migration. In the present Letter, we describe
syntheses and biological evaluations of new 2,6,9-trisubstituted purines, structurally related to roscovi-
tine, a promising CDK inhibitor currently in clinical trials (CDK1/Cyclin B, IC50 = 350 nM; CDK5/p25,
IC50 = 200 nM). These new molecules were synthesized using an original Buchwald–Hartwig catalytic
procedure; several compounds (3j, 3k, 3l, 3e, 4k, 6b, 6c) displayed potent kinase inhibitory potencies
against CDK5 (IC50 values ranging from 17 to 50 nM) and showed significant cell death inducing activities
Received 26 September 2012
Revised 29 October 2012
Accepted 31 October 2012
Available online 10 November 2012
Keywords:
CDK5
Kinase
Roscovitine
(IC50 values ranging from 2 to 9 lM on SH-SY5Y). The docking of the inhibitors into the ATP binding
domain of the CDK5 catalytic site highlighted the discriminatory effect of a hydrogen bond involving
the CDK5 Lys-89. In addition, the calculated final energy balances for complexation measured for several
inhibitors is consistent with the ranking of the IC50 values. Lastly, we observed that several compounds
exhibit submicromolar activities against DYRK1A (dual specificity, tyrosine phosphorylation regulated
kinase 1A), a kinase involved in Down syndrome and Alzheimer’s disease (3g, 3h, 4m; IC50 values ranging
from 300 to 400 nM).
Buchwald–Hartwig amination
Ó 2012 Elsevier Ltd. All rights reserved.
Cyclin-dependent kinases (CDKs) form a family of 20 ubiquitous
serine/threonine kinases (STKs),1–3 involved in cell cycle,4,5 apop-
tosis6 and gene transcription.7 Their regulation is maintained by
structural modifications,8–11 but their activation is related to their
association with regulatory proteins called cyclins.1 In the past
decade, their deregulation has been observed in multiple patholo-
gies,5 which has promoted the search for potent and selective
inhibitors of their serine/threonine kinase (STK) activity.
Among the CDKs family, an unusual member, the cyclin-
dependent kinase 5 (CDK5), was discovered in 1992.12–14 CDK5 is
activated by proteins identified as p35, its isoform p39, and their
respective proteolytic fragments p25 and p29.14–16 The resulting
complexes participate in cell–cell communication and play a key
function in neuronal survival, migration and in the development
of the cerebral cortex. CDK5 is partially responsible for the hyper-
phosphorylation of the protein tau, and for the production of the
b-amyloïd peptide, the two major hallmarks of the Alzheimer’s
disease.17–19 Its implication in other neuronal diseases like
Parkinson’s disease,20,21 amyotrophic lateral sclerosis,22 Hunting-
ton’s disease,23 stroke,24 has been strongly suggested. In addition,
CDK5 is involved in the regulation of pain signalling25,26 and in
the pancreatic secretion of insulin.27 Finally CDK5 is involved in tu-
mor cell metastatic migration, such as prostate cancer cells.28 Thus,
pharmacological inhibitors of CDK5 have a great potential as new
drugs against several major pathologies.
Purines, the scaffold of which is widely used in medicinal chem-
istry,29 have provided several CDKs inhibitors. Thus, (R)-Roscovi-
tine, 6-benzylaminopurine developed by Cyclacel Pharmaceuticals
(Chart 1), is one of the most promising molecule.30–32 Based on this
scaffold, we studied new potent CDKs inhibitors, such as DRF 53
(Chart 1).33 Taking advantage of these structure–activity relation-
ship results, we report here the synthesis of a new library of hybrid
structures
including
6-aminopyridyl,
6-aminopyrimidinyl,
6-aminopyrazinyl and 6-aminopyridazinyl.
⇑
Corresponding authors. Tel.: +33 (0)1 42 86 40 81 (L.D.); tel.: +33 (0)2 98 72 94
92 (L.M.).
The new trisubstituted purines were synthesized following the
efficient three-step convergent procedure outlined in scheme 1.
Briefly, the commercially available 2,6-dichloropurine was
E-mail addresses: luc.demange@parisdescartes.fr (L. Demange), meijer@manros-
therapeutics.com (L. Meijer).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.10.141

126
L. Demange et al. / Bioorg. Med. Chem. Lett. 23 (2013) 125–131
Aryl
H
Aryl
H
Cl
Cl
N
N
N
N
6
7
9
5
N
N
N
N
N
N
2
N
N
N
a
1
b
N
c
N
8
4
2
N
H
R
Cl
N
Cl
N
Cl
3
NH
1
2 a-h
3 a-m, 4 a-p, 5 a-c, 6 a-f
Scheme 1. Reagents and conditions: (a) 2-bromopropane, K2CO3, DMSO 15–18 °C, 5 days; (b) aminoaryl, Pd(OAc)2, ( ) BINAP, tBuOK, toluene, 100 °C; (c) R2-NH2, NEt3, 110-
160 °C.
Table 1
Optimization of the Buchwald–Hartwig catalytic conditions for regioselective C-6 amination of compound 1
Amine
Product
Pd(OAc)2 (%)
BINAP (%)
tBuOK (equiv)
Time
Yield (%)
2-Aminopyridine
3-Aminopyridine
4-Aminopyridine
2-Aminopyrimidine
4-Aminopyrimidine
5-Aminopyrimidine
Aminopyrazine
2a
2b
2c
2d
2e
2f
4
4
4
8
8
0.7
2
8
8
8
16
16
1.4
4
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
6 h
6 h
6 h
1 day
1 day
3 h
1 day
5 h
7335
7635
7335
22
26
57
2g
2h
30
38
4-Aminopyridazine
2
4
All reactions were carried out under argon atmosphere in refluxing dry toluene.
regioselectively alkylated using 2-bromopropane at 15–18 °C in
DMSO for 5 days to afford the 2,6-dichloro-9-isopropylpurine 1
in 70% yield.33 Then, this compound was submitted to the catalytic
Buchwald–Hartwig amination we explored previously using func-
tionalized aminopyridines, which led to compounds 2a–c with
good yields using 4% of Pd(OAc)2 as catalyst and 8% of BINAP as li-
gand.35 In the present study, we extended the scope of this reaction
to aminoaryls including two nitrogen atoms (Table 1). The amina-
tion of 1 with 2- and 4-aminopyrimidines required double concen-
trations of catalysts and ligand (8% of Pd(OAc)2, 16% of BINAP) to
give products 2d–e with moderate yields (22–26%). Conversely,
the use of the previously observed catalytic conditions (4% of
Pd(OAc)2 and 8% of BINAP) with 5-aminopyrimidine and amino-
pyrazine afforded exclusively compounds 2f0 and 2g0 (Scheme 2).
This double amination is probably related to the nucleophilicity
of the amino group in the aminopyrimidine ring which is lower
than those of the 5-aminopyrimidine or those of the aminopyr-
azine. In our hands, lowering the catalytic amount of Pd(OAc)2
led to the expected purines 2f and 2g (Table 1, Scheme 2) with
good yields and reduced reaction time in the case of 5-aminopy-
rimidine. Also, the use of a reduced amount of catalyst and ligand
with 4-aminopyridazine afforded exclusively the expected com-
pound 2h in 38% yield. Lastly, reaction of aminated products 2a–
h with appropriate amino-alcohols afforded trisubstituted purines
3a–m, 4a–p, 5a–c and 6a–f in 50–60% yield.36
The inhibition of CDK5/p25, GSK-3ab and CK1de serine/threo-
nine kinase (STK) activity was determined in the presence of a
range of concentrations of newly synthesized products using an as-
say with [c-
33P]-ATP as described in the experimental section. In
order to evaluate the selectivity of the compounds, their inhibition
of STK activity was controlled under the same conditions with
other purified kinases (DYRK1A, CDK2/cyclin A). IC50 values were
determined from dose–response curves. Globally, as shown in Ta-
bles 2–4, newly synthesized purines exhibit potent activity against
N
N
N
N
HN
HN
Pd(OAc)2 4% BINAP 8%
Pd(OAc)2 4% BINAP 8%
N
N
N
tBuOK 1.5 eq.
tBuOK 1.5 eq.
N
N
N
Cl
N
N
N
N
N
NH
N
NH
N
N
N
H N
N
N
H2N
2g', 45%
2f', 25%
2
N
N
N
N
Cl
N
N
N
HN
HN
1
N
N
N
N
N
Pd(OAc) 0.7 % BINAP 1.4%
tBuOK 1.5 eq.
Pd(OAc) 2 % BINAP 4%
tBuOK 1.5 eq.
N
Cl
N
Cl
N
2g, 30%
2f, 57%
Scheme 2. C-6 amination of dichloropurine 1 and its side products.

L. Demange et al. / Bioorg. Med. Chem. Lett. 23 (2013) 125–131
127
Table 2
Effect of trisubstituted purines on five protein kinases activity
Compd.
Scaffold
R2
CDK2
CDK5
GSK-3
ab
DYRK1A
CK1
2a
3a
Cl
6.2
1
2.1
2.3
>10
33
3.8
3.3
3
6.2
(R)-1-Hydroxy-but-2-ylamino
3b
3c
3d
(S)-1-Hydroxy-but-2-ylamino
(R)-1-Hydroxy-3-methylbut-2-ylamino
(S)-1-Hydroxy-3-methylbut-2-ylamino
0.53
0.8
1.2
4.1
1.1
1.9
21
>10
>10
2.1
2.0
1.2
12
6.2
11
HN
N
N
N
N
2
N
R
2b
3e
Cl
0.68
0.05
1.1
0.025
>10
10
2.1
1.0
2.1
1.0
(R)-1-Hydroxy-but-2-ylamino
3f
(S)-1-Hydroxy-but-2-ylamino
0.014
0.018
0.12
0.09
0.022
0.3
>10
7.3
>10
>10
1.1
0.4
0.4
1.2
2
N
3g
3h
3i
(R)-1-Hydroxy-3-methylbut-2-ylamino
(S)-1-Hydroxy-3-methylbut-2-ylamino
1,3-Dihydroxyprop-2-ylamino
2.2
2.7
1.8
HN
N
0.22
0.3
N
N
N
2
R
2c
3j
Cl
0.7
0.038
7.4
0.017
18
>10
3.9
4.1
2.1
1
N
(R)-1-Hydroxy-but-2-ylamino
3k
3l
3m
(S)-1-Hydroxy-but-2-ylamino
(R)-1-Hydroxy-3-methylbut-2-ylamino
(S)-1-Hydroxy-3-methylbut-2-ylamino
0.05
0.02
0.13
0.05
0.021
0.2
6
12
7.8
0.8
0.78
0.9
>1
1.8
3.3
HN
N
N
N
N
2
R
Purines were tested at various concentrations on kinases as described in the Experimental Section. IC50 values, calculated from the dose–response curves, are reported in
-, not tested. IC50 value reported as >10 indicates that the compound did not display any inhibitory activity at the highest concentration tested (10 M).
lM.
l
the STK activity of CDK2 (IC50 values ranging from 14 nM to 9
except for 4a and 4c) and CDK 5 (IC50 values ranging from 17 nM to
11 M except for 4a and 4c) and average inhibition on the STK
activity of CK1 and DYRK1A (IC50 values ranging from 0.5 to
20 M). Similar to roscovitine, purvalanol and DRF53, these com-
pounds are inactive on GSK-3 /b STK activity, in contrast to mole-
l
M
Docking of 3j in the CDK5 active site (IC50 = 17 nM) confirms the
basic interactions between the purine scaffold and the kinase ATP
binding site (Fig. 2),33 but the most important observation relates
l
to the onset of an H-bond between Lys 89 Ne and the aromatic
l
nitrogen of pyridine. Such a bond could explain the potent inhibi-
tion of CDK5 by compounds bearing 3- and 4-aminopyridine. We
also observed that the H-bond between Cys 83 and the N6–H is
preserved and the onset of two additional H-bond which involve
on the one hand Cys 83 and purine N7 and on the other hand Gln
130 and the purine amino alcohol hydroxyl group of the R2 substi-
tuent. Docking furthermore confirms the presence of a hydropho-
bic pocket which involves Ala 30, Phe 80 and Val 64.
a
cules belonging to other CDKs inhibitor families, such as
indirubins.37
Derivatives bearing an aminopyridine in the C6 position are en-
dowed with IC50 values for CDK5 STK inhibition ranging from 17
nM to 4.1
cient inhibitors for this kinase than roscovitine (IC50 = 0.2
purvalanol (IC50 = 0.07 M) and DRF53 (IC50 = 0.08 M) (Chart 1),
l
M (Table 2). Compounds 3e, 3g and 3j–l are more effi-
lM),
l
l
Studies on purines substituted at C6 by aminopyrimidines,
aminopyrazines or aminopyridazines revealed that nitrogen effects
are cumulative for STK inhibition (Tables 3 and 4). Thus, com-
pounds 4k–m and 6a–c belonging, respectively to the 5-aminopy-
rimidine serial and the 4-pyridazinyl series, are potent CDK2 and
CDK5 inhibitors including IC50 values in the 18-76 nM range. This
result is consistent with the docking of these inhibitors in the
CDK5 ATP binding site which highlights the persistence of an H-
bond between Lys89 and, respectively N3 and N4 of the aminoarylic
structure of 4k and 6a (Supplementary data Fig. 1A and 1B). In
marked contrast, compounds 4a–d belonging to the 2-aminopy-
rimidine serial are inactive, owing to the unfavorable position of
their two nitrogens; their IC50 values are close or superior to
10,000 nM.
and showed no selectivity between CDK2 (IC50 values ranging from
18 to 50 nM), CDK5 (IC50 values ranging from 17 to 50 nM), and
CDK1 (data not shown, IC50 values ranging from 19 to 120 nM).
This result is consistent with the strong structural homology of
the ATP binding site for those kinases.33
Interestingly, the potency of the CDK STK inhibition depends on
the nitrogen position of the aminopyridine core. Thus, compounds
3a–d (on CDK5, IC50 values ranging from 1100 to 4100 nM) are
globally 10 times less active than compounds 3e–i (on CDK5, IC50
values ranging from 22 to 300 nM) which are globally equipotent
than compounds 3j–l (on CDK5, IC50 values ranging from 17 to
50 nM). Among the latter inhibitors, 3e, 3g, 3j and 3l, are about
tenfold more active than roscovitine.

128
L. Demange et al. / Bioorg. Med. Chem. Lett. 23 (2013) 125–131
Table 3
Effect of trisubstituted purines bearing an N6 aminopyrimidinyl on five protein kinases activity
Compd.
Scaffold
R2
CDK2
CDK5
GSK-3
ab
DYRK1A
CK1
2d
4a
Cl
>10
>10
>10
>10
>10
>10
>10
>10
8
>10
N
(R)-1-Hydroxy-but-2-ylamino
4b
4c
4d
(S)-1-Hydroxy-but-2-ylamino
(R)-1-Hydroxy-3-methylbut-2-ylamino
(S)-1-Hydroxy-3-methylbut-2-ylamino
7.3
>10
9
11
>10
10
>10
>10
>10
2.0
>10
7
>10
>10
>10
HN
N
N
N
2
N
R
N
2e
4e
Cl
13
1.2
>10
1.8
>10
>10
>10
3.2
8
5.8
N
(R)-1-Hydroxy-but-2-ylamino
4f
4g
4h
4i
(S)-1-Hydroxy-but-2-ylamino
2.0
4.8
0.73
3.2
8.5
3.2
>10
>10
22
>10
>10
2.8
1
1.1
5.8
16
8.3
2.8
4.9
7
(R)-1-Hydroxy-3-methylbut-2-ylamino
(S)-1-Hydroxy-3-methylbut-2-ylamino
1,3-Dihydroxyprop-2-ylamino
0.33
0.79
6
HN
N
N
N
4j
1-Hydroxy-2-methylprop-2-ylamino
0.73
9
N
2
N
N
R
2f
2f’
Cl
0.18
0.17
2.8
0.36
>10
>10
5.3
0.48
2.3
0.22
5-Aminopyrimidinyl
4k
4l
4m
4n
4o
4p
(R)-1-Hydroxy-but-2-ylamino
(S)-1-Hydroxy-but-2-ylamino
(R)-1-Hydroxy-3-methylbut-2-ylamino
(S)-1-Hydroxy-3-methylbut-2-ylamino
1-Hydroxy-2-methylprop-2-ylamino
1,3-Dihydroxyprop-2-ylamino
0.018
0.023
0.02
0.11
0.10
0.042
0.076
0.038
0.37
0.18
0.3
>10
>10
5
16
>10
>10
0.28
0.27
0.3
0.49
2.5
0.6
0.93
1.3
2
0.82
2.1
N
HN
N
N
N
N
0.26
1.9
2
R
See legend of Table 2 for details. All values are reported in lM.
Table 4
Effect of trisubstituted purines bearing an N6 aminopyrazinyl or aminopyridazinyl on five protein kinases activity
Compd.
Scaffold
R2
CDK2
CDK5
GSK-3
ab
DYRK1A
CK1
N
2g
5a
Cl
3.1
0.21
12
0.9
>10
>10
9
0.4
7
2.8
(R)-1-Hydroxy-but-2-ylamino
5b
5c
2g0
(R)-1-Hydroxy-3-methylbut-2-ylamino
1-Hydroxy-2-methylprop-2-ylamino
Aminopyrazinyl
0.088
1.0
2.8
0.3
3
13
>10
>10
10
0.3
4
0.5
4
4
1.1
HN
N
N
N
N
2
N
R
N
2h
6a
Cl
1.6
0.072
3
0.06
>10
13
3.4
0.45
1.1
0.5
N
(R)-1-Hydroxy-but-2-ylamino
6b
6c
6d
6e
6f
(S)-1-Hydroxy-but-2-ylamino
0.029
0.028
0.12
0.11
0.046
0.042
0.041
0.23
0.2
7.3
6
4.3
>10
13
0.3
0.28
0.48
1.0
3
1
(R)-1-Hydroxy-3-methylbut-2-ylamino
(S)-1-Hydroxy-3-methylbut-2-ylamino
1,3-Dihydroxyprop-2-ylamino
0.8
1.9
0.7
0.6
HN
N
1-Hydroxy-2-methylprop-2-ylamino
0.1
N
N
N
2
R
See legend of Table 2 for details. All values are reported in lM.

L. Demange et al. / Bioorg. Med. Chem. Lett. 23 (2013) 125–131
129
N
HN
N
HN
N
Cl
HN
N
N
N
N
N
N
N
N
N
N
HN
HN
HN
OH
OH
OH
purvalanol A
DRF 53
(R)-roscovitine
CDK1/cyclin B
CDK2/cyclin A
0.35 µM
0.70 µM
0.004 µM34
0.070 µM34
0.075 µM34
0.22 µM
-
CDK5/p25
CK1
0.20 µM
2.30 µM
0.08 µM
0.014 µM
100% inhibition at 0.1 µM
SH-SY5Y
17.4 µM
-
17.2 µM
Chart 1. Structure of (R)-roscovitine, purvalanol A and DRF 53, and their activities on relevant kinases and in cell survival assays. IC50 values are reported in
l
M. (See
above-mentioned reference for further information.)
For each series, the influence of the C-2 amino alcohol appears
to be limited. Nevertheless, as reported in the case of roscovitine,
the (R) stereoisomer appears to have a higher affinity for CDK5
than the (S) one. This is particularly highlighted upon comparing
3l (CDK5:IC50 = 21 nM) and 3m (CDK5:IC50 = 200 nM), 4m
(CDK5:IC50 = 38 nM) and 4n (CDK5:IC50 = 370 nM) and lastly 3g
(CDK5:IC50 = 22 nM) and 3h (CDK5:IC50 = 300 nM). Finally, the
substitution of an amino alcohol by a halogen is very detrimental
in term of CDK5 STK inhibition (IC50 >2000 nM for 2a, 2c–h). Sur-
prisingly, this effect is less detrimental on CDK2 inhibition (2b
IC50 = 680 nM; 2c IC50 = 700 nM and 2f IC50 = 180 nM). Lastly, com-
pound 2f0 (CDK5:IC50 = 360 nM), which bears on C-2 a 5-amino-
pyrimidinyl group retains and average activity (Table 3).
We calculated dEs and dE, the final energy balances for complex-
ation in presence or in absence of solvation. It is the difference be-
tween, on the one hand, the inhibitor-CDK5 intermolecular
interaction energy plus the continuum solvation energy of the
complex; and on the other hand, the sum of the separately en-
ergy-minimized ligand and protein conformational energies plus
their continuum solvation energies at the conformational energy
minima. The ranking of energy balances of the outcome of energy
minimization is globally consistent with that of the IC50 values
(Table 5). However while the dE ranking of the structurally related
compounds 4a, 2f0, 6a and 3j is the same as that of the IC50 values
observed for CDK5, the affinity of DRF 53 which belongs to a differ-
ent family of inhibitor appears overestimated with respect to this
series. This indicates the possibilities and limitations of the compu-
tational approach used in this study.
and 1
l
M)40 or cyclohexyl-thiophene moiety linked with triazole
M).41 The latter compounds
(IC50 range between 35 nM and 1
l
have complex chemical structures and their preparation involves
multi-step syntheses; consequently, large-scale syntheses are
more difficult, and subsequent pharmacomodulations, in order to
improve activity or selectivity, might be very challenging. By con-
trast, the purine scaffold we used here is chemically accessible, sta-
ble and adjustable fur further optimizations.
Interestingly, DYRK1A, a Serine Threonine kinase involved in
neurodegenerative pathologies, such as the Down syndrome,42 is
targeted by several compounds including a C6 5-aminopyrimidinyl
(Table 3), aminopyrazinyl or 4-aminopyridazinyl (Table 4). With
the significant exception of the analogs of Lamellarin D which in-
clude in their structure a complex chromeno[3,4-b]indole skeleton
(IC50 = 0.07 l
M),43 there are few sub-micromolar DYRK1A inhibi-
tors reported in the recent literature, and none of them has reached
the stage of clinical evaluation.44,45 Thus, products 4k–n, 5a, 5b
and 6a–d are the first purine-like sub-micromolar DYRK1A inhibi-
tors (IC50 values ranging from 280 to 480 nM), in contrast to rosco-
vitine (87% of remaining STK activity for DYRK1A at 1
purvalanol (88% of remaining STK activity for DYRK1A at
0.1
M).32 Thus, these molecules might pave the way for the design
lM) or
l
of novel families of DYRK1A inhibitors.
Lastly, we also studied our compounds as inhibitors of CK1, a
kinase involved in multiple physiological events and responsible
Globally, our highest-affinity compounds exhibits activities
against CDK5 similar or better than those previously reported by
our group (DRF 53, Chart 1)33,38 and by others, including specific
CDK5 inhibitors with scaffolds such as pyrazolopyrimidine ring
(IC50 = 30 nM),39 2,4-diaminothiazoles (IC50 range between 15 nM
Table 5
Comparison between the inhibition of CDK5 (IC50 values) and the energy minimi-
zation (calculated without and with energy of solvation) for relevant compounds
Compd.
IC50
(lM) dE Energy minimization dEs Energy minimization
(without solvation)
(with solvation)
(kcal/mol)
(kcal/mol)
Roscovitine
Purvalanol
DRF 53
4a
2f0
0.2
0.07
0.08
ꢀ41
ꢀ32
ꢀ49.5
ꢀ52.5
ꢀ25.7
ꢀ30.8
ꢀ38.2
ꢀ41.9
ꢀ30.7
ꢀ34.0
ꢀ15.9
ꢀ21
Figure 1. Effect of prepared purines on the survival of SH-SY5Y cells. The
compounds were tested at various concentrations for their effects on SH-SY5Y cell
survival after 48 h incubation estimated using the MTS reduction assay as described
in experimental section. IC50 values, calculated from the dose–response curves, are
10.0
0.36
0.06
0.01
6a
3j
ꢀ26.1
ꢀ28.9
reported in lM.

130
L. Demange et al. / Bioorg. Med. Chem. Lett. 23 (2013) 125–131
Figure 2. Compound 3j docked in the ATP binding pocket of CDK5. The classical hydrogen bonds network between the inhibitor and the kinase, and the interaction between
Lys 89 and the relevant nitrogen of the aminoarylic core (orange arrow) are outlined as black dashed lines. H-Bond lengths are also specified. Hydrophobic interactions are
outlined as red dashed lines.
for amyloid-b formation.32 To date, one of the most potent inhibi-
tors is DRF 53 (IC50 = 80 nM, Chart 1), which encompasses a C6
aminobiaryl including a terminal pyridine ring involved in a H-
bond with the ATP binding pocket. By contrast, compounds synthe-
sized in the present study lacking such an entity, are less potent
against CK1 STK activity, and exhibits IC50 values in the 0.800–
molecules exhibit a significant cell death inducing activity. In con-
clusion we suggest that this new series of molecules constitutes an
incentive for the development of novel therapeutic agents against
several diseases involving CDKs deregulation.
Acknowledgments
10
lM range.
In addition to these STK inhibition studies on purified kinases,
This research was supported by Grants from the ‘Cancéropole
Grand-Ouest’ Grant (L.M.), the ‘Institut National du Cancer’ (INCa
« Cancer Détection d’innovations 2006 ») (L.M.), the ‘Association
France-Alzheimer Finistère’ (L.M.) and the ‘Ligue Nationale contre
le Cancer (Grand-Ouest)’ (L.M.).
the most active purines were studied for their cellular effects on
the survival of human neuroblastoma SH-SY5Y cells (Fig. 1). Thus,
SH-SY5Y cells were exposed for 48 h to various concentrations of
purines and cell viability was estimated by the MTS reduction as-
say as described in the Experimental Section. All selected com-
pounds displayed significant cell death inducing activity, with
Professor Christiane Garbay is gratefully acknowledged for crit-
ical reading of the manuscript.
IC50 values ranging from 1.8 to 8.8 lM. Moreover, these molecules
are globally 5 to 10-fold more potent than roscovitine and DRF 53.
This result suggests a good correlation between CDKs inhibition
and cell death inducing activity.
In conclusion, we have described procedures to introduce a new
aminoaryl core in the 6 position of the purine scaffold by resorting
to an innovative variant of the Buchwald–Hartwig amination pro-
cedure. Several compounds exhibit a strong potency to inhibit
CDK5 with IC50 values in the 20 nM range and, by contrast to
DRF53, are selective for CDK5 compared to CK1. The docking of
these compounds in the CDK5 ATP binding site highlights an essen-
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2012.10.
141.
References and notes
1. Malumbres, M.; Barbacid, M. Trends Pharmacol. Sci. 2005, 30, 630.
2. Malumbres, M.; Harlow, E.; Hunt, T.; Hunter, T.; Lahti, J. M.; Manning, G.;
Morgan, D. O.; Tsai, L.-H.; Wolgemuth, D. J. Nat. Cell Biol. 2009, 11, 1275.
3. Carassou, P.; Meijer, L.; Le Moulec, S.; Aoun, J.; Bengrine-Lefèvre, L. Bull. Cancer
2012, 99, 163.
4. Malumbres, M.; Barbacid, M. Nat. Rev. Cancer 2001, 1, 222.
5. Meijer, L. Bull. Cancer 2006, 93, 41.
6. Borgne, A.; Golsteyn, R. M. Prog. Cell Cycle Res. 2003, 5, 453.
7. Garriga, J.; Grana, X. Gene 2004, 337, 15.
tial hydrogen bond involving the Lys 89 N
e and a nitrogen on the
6-aminoaryl which impacts the structure–activity relationship. For
several inhibitors, we have shown a significant correlation be-
tween the ranking of IC50 values and that of the energy balances
at the outcome of energy-minimization. Moreover, the most potent

L. Demange et al. / Bioorg. Med. Chem. Lett. 23 (2013) 125–131
131
8. Abbas, T.; Dutta, A. Cell Cycle 2006, 5, 1123.
9. Lolli, G.; Johnson, L. N. Cell Cycle 2005, 4, 572.
10. Lee, M. H.; Yang, H. Y. Cell. Mol. Life Sci. 1907, 2001, 58.
11. Coqueret, O. Trends Cell Biol. 2003, 13, 41.
12. Hellmich, M. R.; Pant, H. C.; Wada, E.; Battey, J. F. Proc. Natl. Acad. Sci. 1992, 89,
10867.
13. Lew, J.; Beaudette, K.; Litwin, C. M.; Wang, J. H. J. Biol. Chem. 1992, 267, 13383.
14. Dhariwala, F. A.; Rajadhyaksha, M. S. Cell Mol. Neurobiol. 2008, 28, 351.
15. Brown, N. R.; Noble, M. E.; Endicott, J. A.; Johnson, L. N. Nat. Cell Biol. 1999, 1,
438.
16. Ko, J.; Humbert, S.; Bronson, R. T.; Takahashi, S.; Kulkarni, A. B.; Li, E.; Tsai, L. H.
J. Neurosci. 2001, 21, 6758.
17. Cruz, J.; Tsai, L. H. Trends Mol. Med. 2004, 10, 452.
18. Cruz, J. C.; Kim, D.; Moy, L. Y.; Dobbin, M. M.; Sun, X.; Bronson, R. T.; Tsai, L. H. J.
Neurosci. 2006, 26, 10536.
35. Demange, L.; Oumata, N.; Quinton, J.; Bouaziz, S.; Lozach, O.; Meijer, L.; Galons,
H. Heterocycles 2008, 75, 1735. General procedure for Buchwald Hartwig
amination: a solution of Pd(OAc)2 and BINAP in 10 mL of dry toluene was
warmed at 45 °C for 5 min. 2,6-Dichloro-9-iso-propylpurine was then added
under N2 bubbling. The mixture was kept at 45 °C for 10 min and tert-BuOK
was added. After 10 min, the required aminoaryl was added. The mixture was
heated under N2 until the reaction was completed (3 h to 2 days depending on
the amine) at 100 °C. The proportion for each reactant, and the reaction time
depend on the nature of the aminoaryl and are summarized in Table 1 After
cooling at room temperature, the mixture was filtrated through celite, and
toluene was evaporated. The residue was diluted in CH2Cl2 (75 mL) and
washed (1 ꢁ 10 mL) with water and brine (2 ꢁ 10 mL). The organic layer was
dried and concentrated under vacuum. The residue was purified by
chromatography on silica gel using various amount of AcOEt/cyclohexane/
ethanol as eluent.
19. Sadleir, K. R.; Vassar, R. J. Biol. Chem. 2012, 287, 7224.
20. Wong, A. S. L.; Lee, R. H. K.; Cheung, A. Y.; Yeung, P. K.; Chung, S. K.; Cheung, Z.
H.; Ip, N. Y. Nat. Cell Biol. 2011, 13, 568.
21. Zelda, H.; Ip, C.; Ip, H. Y. Trends Cell Biol. 2012, 22, 169.
22. Nguyen, M. D.; Lariviere, R. C.; Julien, J.-P. Neuron 2001, 30, 135.
23. Anne, S. L.; Saudou, F.; Humbert, S. J. Neurosci. 2007, 27, 7318.
24. Menn, B.; Bach, S.; Blevins, T. L.; Campbell, M.; Meijer, L.; Timsit, S. PLoS ONE
2010, 5.
25. Pareek, T. K.; Kulkarni, A. B. Cell Cycle 2006, 5, 585.
26. Utreras, E.; Futatsugi, A.; Rudrabhatla, P.; Keller, J.; Iadarola, M. J.; Pant, H. C.;
Kulkarni, A. B. J. Biol. Chem. 2009, 284, 2275.
27. Wei, F.-Y.; Nagashima, K.; Ohshima, T.; Saheki, Y.; Lu, Y.-F.; Matsushita, M.;
Yamada, Y.; Mikoshiba, K.; Seino, Y.; Matsui, H.; Tomizawa, K. Nat. Med. 2005,
11, 1104.
36. General procedure for 2 chlorine nucleophilic substitution: A mixture of 2-chloro-
6-arylamino-9-iso-propylpurine (0.3 mmol), NEt3 (0.45 mmol) and amino-
alcohol (1.5 mmol) in DMSO (from 0 to 1 mL, depending on the aminoaryl
stability) was heated at 120 °C until the reaction was completed (3 h to 5 days).
After cooling at room temperature, the mixture was diluted with CH2Cl2
(35 mL) and washed with water (5 mL) and brine (5 mL). The organic layer was
dried and concentrated under vacuum. The residue was purified by
chromatography on silica gel using various amount of AcOEt/ethanol/NEt3 as
eluent.
37. Leclerc, S.; Garnier, M.; Hoessel, R.; Marko, D.; Bibb, J. A.; Snyder, G. L.;
Greengard, P.; Biernat, J.; Wu, Y. Z.; Mandelkow, E.-M.; Eisenbrand, G.; Meijer,
L. J. Biol. Chem. 2001, 276, 251.
38. Demange, L.; Lozach, O.; Ferandin, Y.; Hoang, N.-T.; Meijer, L.; Galons, H. Med.
Chem. Res. 2012. http://dx.doi.org/10.1007/s00044-012-0334-1.
39. Heathcore, D. A.; Patel, H.; Kroll, S. H. B.; Hazel, P.; Periyasamy, M.; Alikian, M.;
Kanneganti, S. K.; Jogalekar, A. S.; Scheiper, B.; Barbazanges, M.; Blum, A.;
Brackow, J.; Siwicka, A.; Pace, R. D. M.; Fuchter, M. J.; Snyder, J. P.; Liotta, D. C.;
Freemont, P. S.; Aboagye, E. O.; Coombes, R. C.; Barrett, A. G. M.; Ali, S. J. Med.
Chem. 2010, 53, 85082.
40. Laha, J. K.; Zhang, X.; Qiao, L.; Liu, M.; Chatterjee, S.; Robinson, S.; Kosic, K. S.;
Cuny, G. D. Bioorg. Med. Chem. Lett. 2011, 21, 2098.
41. Shiradkar, M.; Thomas, J.; Kanase, V.; Dighe, R. Eur. J. Med. Chem. 2011, 46,
2066.
42. Wegiel, J.; Gong, C. X.; Hwang, Y. W. FEBS J. 2011, 278, 239.
43. Neagoie, C.; Vedrenne, E.; Buron, F.; Mérour, J. Y.; Rosca, S.; Bourg, S.; Lozach,
O.; Meijer, L.; Baldeytou, B.; Lansiaux, A.; Routier, S. Eur. J. Med. Chem. 2012, 49,
379.
28. Strock, C. J.; Park, J. I.; Nakakura, E. K.; Bova, G. S.; Isaacs, J. T.; Ball, D. W.;
Nelkin, B. D. Cancer Res. 2006, 66, 7509.
29. Legraverend, M.; Grierson, D. S. Bioorg. Med. Chem. 2006, 14, 3987.
30. Meijer, L.; Raymond, E. Acc. Chem. Res. 2003, 36, 417.
31. Meijer, L.; Bettayeb, K.; Galons, H. Roscovitine (CYC202, Seliciclib) In
Monographs on enzyme inhibitors. CDK inhibitors and their potential as anti-
tumor agents; Smith, P. J., Yue, E., Eds.; CRC Press: Taylor & Francis: Boca Raton,
Fl.;, 2006; Vol. 2, pp 187–226. chapter 9.
32. Bain, J.; Plater, L.; Elliott, M.; Shpiro, N.; Hastie, C. J.; McLauchlan, H.; Klevernic,
I.; Arthur, J. S.; Alessi, D. R.; Cohen, P. Biochem. J. 2007, 408, 297.
33. Oumata, N.; Bettayeb, K.; Ferandin, Y.; Demange, L.; Lopez-Giral, A.; Goddard,
M.-L.; Myrianthopoulos, V.; Mikros, V.; Flajolet, M.; Greengard, P.; Meijer, L.;
Galons, H. J. Med. Chem. 2008, 51, 5229.
34. Popowycz, F.; Fournet, G.; Schneider, C.; Bettayeb, K.; Ferandin, Y.; Lamigeon,
C.; Tirado, O. M.; Mateo-Lozano, S.; Notario, V.; Colas, P.; Bernard, P.; Meijer, L.;
Joseph, B. J. Med. Chem. 2009, 52, 655.
44. Becker, W.; Sippli, W. FEBS Lett. 2011, 278, 246.
45. Wang, D.; Wang, F.; Tan, Y.; Dong, L.; Chen, L.; Zhu, W.; Wang, H. Bioorg. Med.
Chem. Lett. 2012, 22, 168.