Synthesis and biological evaluation of selective and potent cyclin-dependent kinase inhibitors

Article abstract of DOI:10.1016/j.ejmech.2012.08.033

A new series of 2,6,9-trisubstituted purines, structurally related to the cyclin-dependent kinase (CDK) inhibitor Roscovitine, has been synthesized. These compounds mainly differ by the substituent on the C-2 position which encompasses a diol group. These

Full text of DOI:10.1016/j.ejmech.2012.08.033

European Journal of Medicinal Chemistry 56 (2012) 210e216
Contents lists available at SciVerse ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Synthesis and biological evaluation of selective and potent cyclin-dependent
kinase inhibitors
,
,
a,
*
Joannah N’gompaza-Diarra a b, Karima Bettayeb c, Nohad Gresh b, Laurent Meijer c , Nassima Oumata
**
a ManRos Therapeutics, Centre de Perharidy, Hôtel de recherche, 29680 Roscoff, France
b Laboratoire de Chimie Organique 2, UMR CNRS 8601, Université Paris Descartes, Sorbonne Paris Cité, Faculté de pharmacie, Paris, France
c C.N.R.S., ‘Protein Phosphorylation & Human Disease’ Group, Station Biologique, B.P. 74, 29682 Roscoff, France
a r t i c l e i n f o
a b s t r a c t
Article history:
A new series of 2,6,9-trisubstituted purines, structurally related to the cyclin-dependent kinase (CDK)
inhibitor Roscovitine, has been synthesized. These compounds mainly differ by the substituent on the
C-2 position which encompasses a diol group. These compounds were screened for kinase inhibitory
activities and antiproliferative effects. They were shown to be potent inhibitors of cyclin-dependent
Received 9 June 2012
Received in revised form
21 August 2012
Accepted 22 August 2012
Available online 1 September 2012
kinases but also, for some of them of casein kinase
1 (CK1) and dual specificity tyrosine-
phosphorylation-regulated kinase 1A (DYRK1A). The inhibition of kinases was accompanied by an
antiproliferative effect against several tumor cell-lines. The most potent derivatives inhibited SH-SY5Y
(neuroblastoma) tumor cell line with an IC50 < 0.5 mM which means approximately a 30 fold increase
compared to Roscovitine. A valine ester was also prepared from the most potent inhibitor to serve as
a prodrug.
Keywords:
Kinase
DYRK1A
Cancer
Ó 2012 Elsevier Masson SAS. All rights reserved.
Prodrug
Purine
Roscovitine
1. Introduction
scaffold [15]. This led us to the identification of CR8 a potent and
selective kinase inhibitor [16].
Cyclin-dependent kinases (CDKs) play an important role in cell
cycle regulation [1] but are also implicated in the regulation of
apoptosis [2], transcription [3] and neural functions [4]. CDKs are
deregulated in numerous proliferative diseases such as cancer [5],
viral infections [6] and kidney disease [7] but also in non prolifer-
ative disease, such as Alzheimer disease in the case of CDK5 [8].
These implications have stimulated the search for pharmaco-
logical inhibitors [9]. Roscovitine [10] and Flavopiridol [11] (Fig. 1)
were the first two to enter clinical trials in the late nineties but up to
now, no product has reached the market. However the interest for
Roscovitine and Flavopiridol remains high as these two products
have now reached phase II/III clinical tests [12]. On the other hand
many products have been terminated due to several toxicities [13].
Most of the ongoing clinical tests are targeting leukemia [14].
Inspired by the continuous interest for Roscovitine we have
conducted several optimization studies which were mostly devoted
to the selection of the substituent at the position 6 of the purine
We now focus on the optimization of groups on the C-2 position
of the purine scaffold. In particular, with the aim to develop less
hydrophobic compounds which could be more soluble and there-
fore display an improved bioavailability.
2. Results and discussion
2.1. Chemistry
Two different approaches were adopted to prepare those of the
aminodiols, which were not commercially available (Scheme 1). In
the first route reaction of cyclohexenylbromide in THF at 60 ꢀC in
the presence of K2CO3 led to the formation of the tertiary amine.
The alkene 1 was then subjected to asymmetric dihydroxylation
with AD-mix b
Ò as described by Sharpless [17] to obtain the diol 2.
In a third step, benzyl groups were removed using catalytic
hydrogenation conditions. This approach was exemplified in the
synthesis of 3-aminocyclohexanediol which was isolated as its
corresponding hydrochloride 3. The second route starts from the
commercial amino acids which were first converted to their ethyl
esters using thionyl chloride in EtOH (4aeb) and the ester group
was reduced into alcohol using AlLiH4 and afforded compounds
* Corresponding author. Tel.: þ33 (0)2 98 72 94 94.
** Corresponding author. Tel.: þ33 (0)6 08 60 58 34.
E-mail addresses: meijer@sb-roscoff.fr (L. Meijer), oumata@manros-
therapeutics.com (N. Oumata).
0223-5234/$ e see front matter Ó 2012 Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.ejmech.2012.08.033

J. N’gompaza-Diarra et al. / European Journal of Medicinal Chemistry 56 (2012) 210e216
211
The selective esterification of one of the alcohols of compound
8a has been investigated with Bocevaline. A mono ester 9 was
formed under classical acylation conditions (DCC, HOBt). Finally,
the Boc protecting group was eliminated using an ethereal hydro-
chloric acid solution and provided compound 10 (Scheme 3).
2.2. Biological evaluation
The prepared compounds were first evaluated against a subset
of CDKs but also against GSK3-a/b (Glycogen synthase kinase 3),
Flavopiridol
Alvocidib
(R)-Roscovitine
Seliciclib
DYRK1A (dual-specificity tyrosine-(Y)-phosphorylation regulated
kinase 1A) and CK1 (casein kinase 1). Compounds were tested for
their antiproliferative effects. Human neuroblastoma SH-SY5Y cells
were exposed for 48 h to various concentrations of each compound.
Viability was then estimated by the MTS reduction assay.
Fig. 1. Representation of (R)-Roscovitine and Flavopiridol the first two CDK inhibitors
which reached clinical tests.
(5aeb). Both methods proved to be reproducible in various scales,
providing the desired compounds in good yields.
The results of the kinase inhibition assays and cell anti-
proliferative effects are gathered in Table 1. The products can be
considered as rather selective CDK inhibitors. Indeed, the products
The substituted purines were then obtained through a classical
three-step synthetic route outlined in Scheme 2. Starting from
commercially available 2,6-dichloropurine, alkylation of N-9 with
2-bromopropane provided 2,6-dichloro-9-iso-propylpurine 6.
Displacement of the 6-chloride by nucleophilic substitution with
4-(2-pyridyl)benzylamine led to compound 7 according to a previ-
ously described procedure [18]. In the final step, reaction of the
chloropurine with diols led to compounds 8aed in moderate yield:
12e53% (Table 1).
do not inhibit GSK3 a/b which is frequently inhibited by most of
CDK inhibitors such as AT5147, a CDK inhibitor which was intro-
duced in clinical phase 1 and showed major toxic effects [19].
Inhibition of GSK3 is generally considered potentially carcinogenic
in relation with the inhibition of the Wnt pathway [20].
The products exhibited roughly the same selectivity as Rosco-
vitine. In particular, some of the prepared compounds appeared to
be also inhibitors of DYRK1A. The role of DYRK1A in cancer is far to
Fig. 2. Molecular docking of compound 10 with CDK9. The construction of the complex was based on the crystal structure determination of the CR8eCDK9 complex.

212
J. N’gompaza-Diarra et al. / European Journal of Medicinal Chemistry 56 (2012) 210e216
Table 1
_
Route A
Kinase inhibition by compounds 8 and 10 against CDKs (Cyclin-dependent kinases):
CDK1/cyclin B, CDK2/cyclin A, CDK7/cyclin H, CDK9/cyclin T, CDK5/p25, CK1, CLK1,
DYRK1A and against the tumor cell line SH-SY5Y. The IC50 M) are the mean of three
Cl
(
m
experiments. (R)-Roscovitine (Ros) was used as a positive control.
a
b
c
Cpd CDK1 CDK2 CDK5 CDK9 GSK3 CK1 CLK1 DYRK1A SHSY-5Y
Ros 0.33
0.21
0.48
0.61
0.65
0.98
0.87
0.28
0.19
0.27
0.61
0.82
0.18
0.23
0.15
0.37
0.45
1.01
0.18
60
7.1
>10
>10
>10
>10
4
4.3
3
17
8a
8b
8c
8d
10
0.60
0.87
0.45
0.37
1.4
0.18 4.0
0.71
0.89
1.6
0.79
1.1
0.52
0.91
0.48
1.4
0.60
0.39
0.66
e
e
e
1
2
3
0.15 12
1.1
d
e
The complexes were constructed using the previously reported
crystal structure of CR8eCDK9 complex [16]. The binding of ligand
10 is stabilized by (Fig. 1):
5a-b
4a-b
- two hydrogen bonds with the Cys106 backbone. As is the case
with CR8, the unsubstituted N of the pentacyclic ring accepts
a proton from the Cys NH group, while the extracyclic N
donates its proton to the Cys carbonyl group;
- a hydrogen bond between the pyridine nitrogen and the Lys35
side-chain;
- an ionic bond between the protonated ammonium group of the
ligand and the side-chain of Asp 109. There are in addition van
der Waals interactions between:
- the bicyclic ring and Phe 105, Ile 25, and Leu 156, while its
substituting isopropyl group interacts with Phe 103;
- the phenyl group and both Phe 105 and Ile 25.
b
a
Scheme 1. Preparation of amino-diols. Reagents and reaction conditions route A. (a)
Dibenzylamine, K2CO3, THF, 60 ꢀC, 6 h; (b) Ad-mix
, MeSONH2, NaHCO3, H2O, tert-
BuOH, r.t, 72 h; (c) 1-HCl/Et2O, 2-H2/Pd/C. Reagents and conditions route B: (d) SOCl2,
EtOH, 60 ꢀC. (e) LiAlH4, THF, H2O, 0 ꢀC- 20 ꢀC.
b
be well understood however some leukemia have been associated
with overexpression of DYRK1A [21]. In contrast to Roscovitine,
they inhibited CK1, a property that probably efficiently contribute
to their high antiproliferative activity.
Several valyl esters have been previously described to improve
bioavailability of the parent compound. The most known is Vala-
The substituent in position 2 of the kinase scaffold is partly
outside the ATP pocket. Such a result could open the way to the
design of new inhibitors with rather diverse substitutions in position
2. Finally to further evaluate the potential interest of 8a, which is the
most potent compound of the series, several preliminary ADME
parameters were determined. Inhibition against a series of CYP 450
cyclovir (ValtrexÒ), the
L-valyl ester prodrug of acyclovir (ACV),
which is widely used to treat viral infections mainly caused by
herpes simplex virus [22].
In the particular case of ester 10, it was observed that 10 dis-
played antiproliferative activity against cancer cells. This is not
really surprising as the ester may undergo a slow hydrolysis in the
culture media. The ester was also found to inhibit kinases. This
result was not anticipated as, up to now, rather small groups have
been introduced in position 2 of the purine.
enzymes was found higher than 5
mM and plasma binding protein
was also favorable with a moderate binding to serum proteins
(Table 2). Evaluation of the ester 10 requires in vivo experiment for
a better understanding of the dual effect of this compound (Fig. 2).
2.3. Docking analysis
3. Conclusion
A docking analysis was performed with the ester 10 for a better
understanding of the unexpected activity of 10 against kinases.
The attractive inhibitory profile against kinase of derivative 8a
with in particular a strong inhibition of CK1 led us to select 8a and
a
b
c
R1
R2
6
8
7
8b
8a
8c
8d
R1
R2
Scheme 2. Synthesis of trisubstituted purines. Reagents and conditions: (a) 2-bromopropane, K2CO3, DMSO, 10e15 ꢀC, 55%; (b) ArCH2NH3þ. TFAꢁ, NEt3, BuOH, 100 ꢀC, 2 h; (c)
R1R2NH, neat or R1R2NH, NEt3,160e170 ꢀC.

J. N’gompaza-Diarra et al. / European Journal of Medicinal Chemistry 56 (2012) 210e216
213
b
a
10
H3N
9
8a
_
Cl
Scheme 3. Synthesis of valine mono ester from 8a. Reagents and conditions: (a) BocevaleOH, DCC/HOBt, THF/AcOEt, rt, 72 h, (b) HCl/Et2O.
its ester 10 for further evaluations. This choice is also supported the
high potency of this product compared to Roscovitine, approxi-
mately 30 fold at inhibiting cell proliferation. Moreover the
preliminary ADME assays in particular high solubility and moderate
serum protein binding were found very favorable.
15 min and then cooled to 0 ꢀC. The alkene 1, (10 g, 36 mmol) was
added and the reaction mixture was stirred for 3 days at 20 ꢀC. The
reaction was quenched with Na2SO3 and the suspension was
allowed to stir for 1e2 h. The suspension was diluted H2O and
AcOEt The organic layer was extracted, washed with water and
brine three times, dried over anhydrous Na2SO4, filtered and
evaporated in vacuum to afford crude product which was purified
by silica gel column chromatography. Cyclohexane/AcOEt (6/4),
4. Experimental
Rf ¼ 0.33, m ¼ 7 g, yield: 62%. 1H NMR (400 MHz, CDCl3)
d : 1.15 (m,
4.1. General methods
1H, CH2), 1.29 (m, 1H, CH2), 1.8 (m, 2H, CH2), 2.10 (d, J ¼ 1.85 Hz, 1H,
CHeNeBn2), 2.79 (m, 1H, CH2eCHOH), 3.32 (d, J ¼ 13.24 Hz, 2H,
CH2eAr), 3.44 (dd, J ¼ 3.12 Hz, J ¼ 10.64 Hz, 1H, CH2eCHOH), 3.6 (s,
1H, CHOH), 3.75 (d, J ¼ 13.28 Hz, 2H, CH2eAr), 4.07 (s, 1H, CHOH),
Melting points were determined on a Kofler hot-stage (Reichert)
and are uncorrected. NMR spectra were recorded on Bruker Avance
400 MHz. Chemical shifts are given in ppm downfield of tetrame-
thylsilane (TMS) used as an internal standard. Reactions were
7.1e7.3 (m, 10H, HeAr). 13C NMR
d : 19.48 (CH2), 22.03 (CH2), 29.51
(CH2), 53.64 (CH2), 57.45 (CH), 68.28 (CH), 70.85 (CH), 127.26 (CH),
127.46 (C), 128.52 (C), 128.98 (CH), 139.29 (C).
monitored by TLC using SDS silica gel 60F-254, 60 A-15
layer plates. Column chromatography was carried out on SDS
Chromagel 60 ACC, 40e63 m. Compounds 8e10 gave satisfactory
mm thin
m
4.2.3. (1S-2R-3R)-3-Aminocyclohexane-1,2-diol hydrochloride (3)
The dibenzylamine derivative (7 g, 21.8 mmol) was mixed with
Et2O and a solution of HCl in Et2O was added to the medium,
a white solid precipitated. The solid was filtered, dried in vacuo,
then diluted in methanol and submitted to hydrogenation (H2/Pd/
C). The reaction mixture was stirred for 48 h. The solution was
filtered over celite, washed with MeOH and evaporated to give
microanalyses ꢂ0.4 calculated values.
4.2. Preparation and dihydroxylation of dibenzylaminocyclohex-2-
ene
4.2.1. N,N-Dibenzylaminocyclohex-2-ene (1)
3-Bromocyclohexenyl (10 mL, 86 mmol) in 20 mL THF was
added to a mixture of dibenzylamine (40 mL, 210 mmol) and then
K2CO3 (14.26 g, 103 mmol) in 120 mL THF. The mixture was stirred
and heated at reflux. The reaction was followed by TLC. THF was
concentrated in vacuo. The residue was extracted by 20 mL H2O and
AcOEt (3 ꢃ 20 mL). The organic layer was washed with H2O, dried
over anhydrous Na2SO4, filtered and concentrated in vacuo to afford
viscous yellow oil which was purified by silica gel column chro-
matography, Cyclohexane/AcOEt (95:5), Rf ¼ 0.7, m ¼ 21.86 g, yield:
a viscous oil. Yield: 93%. 1H NMR (400 MHz, CDCl3)
d : 1.1e1.8 (m,
6H, CH2CH2CH2), 2.9 (m, 1H, CHOH), 3 (m, 1H, CHeNH2), 3.3 (m, 1H,
CHOH), 5.5 (s, 2H, NH2).
4.3. Preparation of aminodiols from amino-acids
4.3.1. Esterification of amino-acids
Thionyl chloride (7.71 mL, 76.2 mmol) was added droplet in cold
(ꢁ10 ꢀC) ethanol (100 mL), the solution was left at ꢁ10 ꢀC for 15 min.
The acid (38.1 mmol) was added to the medium and the mixture was
refluxed at 60 ꢀC for 4 h. The reaction was followed with TLC and the
plates revealed with ninhydrin. The medium was concentrated and
the white crystals were dissolved in DCM, washed with NH4OH aq.
The aqueous layer was extracted with DCM. The organic layer was
dried over anhydrous Na2SO4, filtered and evaporated.
92%. 1H NMR (400 MHz, CDCl3)
d : 1.4e1.6 (m, 3H, CH2eCH2), 1.75
(m, 1H, CHeCHeN), 1.8e2 (m, 2H, CH2eCH]CH), 3.3 (m, 1H, CH2e
CHeN), 3.5 (d, J ¼ 14.05 Hz, 2H, CH2eAr), 3.7 (d, J ¼ 14.05 Hz, 2H,
CH2eAr), 5.6e5.8 (m, 2H, CH]CH), 7.1e7.4 (m,10H, HeAr). 13C NMR
d
: 21.87 (CH2), 23.19 (CH2), 25.38 (CH2), 53.84 (CH2), 54.51 (CH),
126.61 (CH), 127.88 (C), 128.12 (CH), 128.23 (CH), 128.34 (C), 128.49
(CH), 128.75 (C), 130.06 (CH), 130.87 (CH), 140.89 (CH).
4.2.2. (1S-2R-3R)-3-(Dibenzylamino)cyclohexane-1,2-diol (2)
4.3.1.1. (2S-4R)-Ethyl-4-hydroxypyrrolidine-2-carboxylate
(4a).
AD-mix b
Ò (73.6 g, 52 mmol) was added to a solution of NaHCO3
This ester was obtained as viscous yellow oil, 5.27 g, yield: 87%
(10.16 g, 120 mmol), Methanesulfonamide (3.8 g, 95.11 mmol) in
following the above procedure starting with HeHypeOH. 1H NMR
50 mL H2O: tert-butanol (1:1). The medium was stirred at 20 ꢀC for
(400 MHz, CDCl3)
d : 1.1 (t, 3H, CH2eCH3), 1.8 (m, 1H, CH2eCHe
CO2Et), 2 (m, 1H, CH2eCHeCO2Et), 2.2 (m, 1H, CH2eNH), 2.9 (m,
1H, CH2eNH), 3 (m, 1H, CH2eNH), 3.4 (m, 1H, CHeCOH), 4.6 (q, 2H,
CH2eCH3).
Table 2
Preliminary ADME study. CYP 450 inhibitions. IC50 are expressed in
protein binding (rat and human) in %.
mM. Serum
CYP450 inhibition
Serum protein binding
4.3.1.2. (S)-Ethyl 4-amino-hydroxy-butanoate (4b). Starting with
(S)-4-amino-2-hydroxy-butanoic acid (10 g, 0.083 mol) 4b, was
obtained as a colorless oil, m ¼ 6.98 g, 57%. 1H NMR (400 MHz,
1A2
2C19
2C9
2D6
3A4
Rat
Human
94.6
>5
>5
>5
>5
>5
97.4

214
J. N’gompaza-Diarra et al. / European Journal of Medicinal Chemistry 56 (2012) 210e216
CDCl3)
d
: 1.25 (t, 3H, CH2eCH3), 2e2.3 (m, 2H, CH2eCHOH), 3.3 (m,
cooled to r.t; water and CH2Cl2 were added into the medium. The
organic layer was extracted, dried over anhydrous Na2SO4, filtered
and evaporated to provide brown solid. Purification was achieved
by flash chromatography or trituration with diethyl ether to
provide the desired product.
2H, CH2eNH2), 4.2 (q, 2H, CH2eCH3), 4.5 (t, 1H, CHOH), 5.3 (s, 1H,
NH2). 13C NMR (400 MHz, DMSO)
d
: 14.08 (CH3), 31.28(CH2),
35.73(CH2), 60.30 (CH2), 67.34(CH), 173.21(C).
4.3.2. Reduction of esters
Anhydrous THF (35 mL) was added dropwise to LiAlH4 (5.51 g,
145 mmol) at 0 ꢀC. A solution of ester (33 mmol) in THF was added
dropwise under stirring at 0 ꢀC. The reaction mixturewas warmed to
r.t then heated to 60 ꢀC for 3 h. The mediumwas cooled to ꢁ15 ꢀC and
H2O was added dropwise intothe mixture until it became white. The
reaction mixture was then stirred overnight at r.t. The solid was
filtered and alumina was washed with CH2Cl2 and THF. The solution
was concentrated to provide the diol derivative as viscous oil.
4.6.1. (2R)-3-[[9-iso-propyl-6-[[4-(2-pyridyl)phenyl]methylamino]
purin-2-yl]amino]propane-1,2-diol (8a)
Yield 53%, 1H NMR, (CDCl3)
d
: 1.5 (d, J ¼ 6.72 Hz, 6H, 2CH3), 3.4e
3.6 (m, 4H, CH2NHCHCH2OH), 4.1 (m, 1H, CHOH), 4.6 (s, 1H, NCHN),
4.8 (s, 2H, ArCH2NH), 5.2 (s, 1H, CHeiPr), 6.2 (s, 1H, NH), 7.2 (m, 1H,
Hpyridyl), 7.4 (d, J ¼ 8.28 Hz, 2H, Hphenyl), 7.5 (s, 1H, H-8), 7.75 (m, 2H,
Hpyridyl), 7.95 (d, J ¼ 8.28 Hz, 2H, Hphenyl), 8.5 (d, J ¼ 4.12 Hz, 1H,
Hpyridyl). NMR 13C (CDCl3)
d : 22.64 (CH3), 44.78 (CH), 46.46 (CH),
63.61 (CH2), 72.55 (CH), 114.75 (C), 120.54 (CH), 122.13 (CH), 127.18
(CH), 127.98 (CH), 128.23 (C), 129.04 (CH), 134.64 (CH), 136.80 (CH),
138.55 (C), 139.48 (C), 149.66 (CH), 154.88 (C), 157.12 (C).
4.3.2.1. (3R,5S)-5-(Hydroxymethyl)-pyrrolidin-3-ol (5a). The reduc-
tion of (2S-4R)-Ethyl-4-hydroxypyrrolidine-2-carboxylate (5.27 g,
0.033 mol) afforded 5a as yellow oil (1.3 g, 34%) following the above
procedure. 1H NMR (400 MHz, CDCl3)
d
: 1.6 (m, 1H, CH2e
4.6.2. (1S,2R,3R)-3-[[9-iso-propyl-6-[[4-(2-pyridyl)phenyl]
methylamino]purin-2yl]amino]cyclohexane-1,2-diol (8b)
CHCH2OH), 1.85 (m, 1H, CH2eCHCH2OH), 2 (s, 1H, NH), 2.55 (m,
1H, CH2eNH), 2.85 (m, 1H, CH2eNH), 2.9 (m, 1H, CHeCH2OH), 3.3
(m, 1H, CHeOH), 3.4e3.55 (m, 2H, CH2OH), 4.4 (s, 1H, OH).
Compound 3 was first converted into the free base upon extraction
with CH2Cl2 of a mixture of 3 in a saturated solution of Na2CO3 in H2O.
Yield 13%, mp: 100e101 ꢀC,1H NMR (DMSO)
d
: 1.5 (d, J ¼ 6.722 Hz, 6H,
4.3.2.2. (S)-4-Aminobutane-1,2-diol (5b). The reduction of (S)-
Ethyl-4-amino-hydroxy-butanoate (6.9 g, 0.046 mol) afforded 5b as
viscous yellow oil (3.07 g, 63%) following the above procedure. 1H
NMR (400 MHz, DMSO) d : 1.3 (m, 1H, CH2eCH2eCHOH), 1.5 (m, 1H,
CH2eCH2eCHOH), 2.6 (m, 2H, CH2eNH2), 3.2 (m, 1H, CH2eCHOH),
3.3 (m, 1H, CH2OH), 3.4 (m, 1H, CH2OH).
2CH3), 1.8 (m, 4H, CH2CH2CH2), 2 (m, 2H, CH2CH2CH2), 2.8 (m, 1H,
CHNH), 3.45 (dd, J ¼ 3 Hz, J ¼ 9.32 Hz, 1H, CHOH), 3.7 (m, 1H,
CH2NHAr), 4 (d, J ¼ 2.74 Hz, 1H, CHOH), 4.6 (m, 1H, NHN), 4.75 (m,1H,
CHeiPr), 5.7 (s, 1H, NH), 6 (s, 1H, NH), 7.1 (m, 1H, Hpyridyl), 7.4 (d,
J ¼ 8.36 Hz, 2H, Hphenyl), 7.6 (s,1H, Hphenyl), 7.7 (m, 2H, Hpyridyl, Hphenyl),
8 (d, J ¼ 8.32 Hz, 2H, Hpyridyl, H-8), 8.5(d, J ¼ 5 Hz,1H, Hpyridyl).13C NMR
(CDCl3)
d : 18.57 (CH2), 19.05 (CH2), 21.29 (CH), 22.52 (CH3), 22.64
4.4. 2,6-Dichloro-9-iso-propylpurine (6)
(CH3), 23.55 (CH), 29.12(CH2), 31.48 (CH2), 46.67 (CH), 52.75 (CH),
69.33 (CH), 72.71 (C), 79.4 (CH), 120.53 (CH), 122.13 (CH), 127.22 (CH),
127.98 (CH), 134.90 (C), 136.78 (CH), 138.62 (C), 149.68 (CH).
This compound was prepared as described previously [15]. A
mixture of2,6-dichloropurine(10g, 53mmol)inDMSO(80mL)K2CO3
(39.25 g, 280 mmol) was cooled to 15 ꢀC. 2-Bromopropane (35 mL,
60.05 mmol) was introduced droplet. The reaction was monitored by
TLC until completion. The solution was diluted with 100 mL H2O and
extracted with AcOEt (3 ꢃ 25 mL). The organic layer was washed with
H2O (3 ꢃ 10 mL) and dried over anhydrous Na2SO4, filtered and
evaporated under vacuum to provide a yellow solid which was puri-
fied by chromatography on silica gel column (toluene/AcOEt) (9/1) to
4.6.3. (3R,5S)-5-(Hydroxymethyl)-1-[9-iso-propyl-6-[[4-(2-
pyridyl)phenyl]methylamino]- purin-2-yl]pyrrolidin-3-ol (8c)
Yield 43%, mp: 210e212 ꢀC, 1H NMR (DMSO)
d : 1.1e1.2 (t,
J ¼ 7 Hz, 4H, CH2CHCH2OH), 1.55 (d, J ¼ 6.76 Hz, 6H, 2CH3), 4.15 (m,
1H, CHeCH2OH), 4.4 (m, 1H, CHOH), 4.45 (d, J ¼ 5.08 Hz, 2H, CH2e
Ar), 4.6 (m,1H, CH2OH), 4.7 (m,1H, CH2eOH), 4.8 (d, J ¼ 4.08 Hz,1H,
CHeiPr), 7.35 (m, 1H, Hpyridyl), 7.55 (d, J ¼ 8.24 Hz, 2H, Hphenyl), 7.8e
8 (m, 3H, Hpyridyl, H-8), 8.1 (d, J ¼ 5.6 Hz, 2H, Hphenyl), 8.7 (m, 1H,
yield a white solid (6.8 g, 55%).1H NMR (400 MHz, CDCl3)
d 1.67 (s, 6H,
2(CH3)), 4.93 (s, 1H, CH(CH3)2), 8.17 (s, 1H, Hearyl).
Hpyridyl). 13C NMR (DMSO, CDCl3)
d : 21.93(CH3), 37.21 (CH2), 40.10
(CH), 45.86 (CH), 58.29(CH), 68.00 (CH), 119.99 (CH), 122.34 (CH),
126.26 (CH), 127.82 (CH), 135.38 (C), 136.96 (C), 137.12 (CH), 149.42
(CH), 155.93 (C), 157.54 (C).
4.5. Preparation of 2-chloro-9-iso-propyl-N-[[4-(2-pyridyl)phenyl]
methyl]purin-6-amine (7)
To a solution of 6 (2.31 g, 10 mmol) in n-BuOH was added the
primary amine (12 mmol) and NEt3 (2.20 mL, 16 mmol). After
heating at 100 ꢀC for 2 h, n-BuOH was evaporated in vacuo. After
dilution with H2O (10 mL), the mixture was extracted with AcOEt
(3 ꢃ 20 mL). The combined organic extracts were dried (Na2SO4)
and the solvent was removed in vacuo. The residue was chroma-
tographied on a silica gel column using CH2Cl2/AcOEt, 2:1 to 1:1 as
4.6.4. (2S)-4-[[9-iso-propyl-6-[[4-(2-pyridyl)phenyl]methylamino]
purin-2-yl]amino]butane-1,2-diol (8d)
Yield 12%, mp: 186e187 ꢀC, 1H NMR (DMSO)
d : 1.1e1.3 (m, 2H,
CH2eCH2), 1.5 (d, J ¼ 6.92 Hz, 6H, 2CH3), 1.9 (m, 1H, CH2eNH), 2 (m,
1H, CH2NH), 4.4 (s, 1H, CH2eCHeOH), 4.6 (spt, 2H, CH2eOH), 4.8 (s,
2H, CH2eAr), 4.9 (m, CHeIsop), 7.45 (m, 1H, Hpyridyl), 7.55 (d,
J ¼ 8.2 Hz, 1H, Hphenyl), 7.85 (s, 1H, Hepyr), 7.9 (m, 1H, H-8), 8 (m,
2H, Hphenyl), 8.1 (d, J ¼ 8.28 Hz, 2H, Hphenyl), 8.7 (d, J ¼ 4.72 Hz, 1H,
eluent. Yield 56%, mp: 176e179 ꢀC. 1H NMR (CDCl3)
d :1.58 (d, 6H,
J ¼ 6.8 Hz, CH(CH3)2), 4.79 (hept, 1H, CH(CH3)2), 4.85 (bs, 2H,
NHCH2), 6.59 (bs, 1H, NHCH2), 7.20e7.23 (m, 1H, Hpyridyl), 7.49 (d,
2H, J ¼ 8 Hz, Hphenyl), 7.73e7.71 (m, 2H, Hpyridyl), 7.79 (s, 1H, H-8),
7.98 (d, 2H, Hphenyl), 8.71 (d, 1H, J ¼ 4.8 Hz, Hpyridyl).
Hpyridyl).13C NMR (CDCl3)
d : 21.93 (CH3), 33.53 (CH2), 40.10 (CH),
44.44 (CH2), 45.66, 54.95 (CH2), 69.2, 120.00 (CH), 122.34 (CH),
126.23 (CH), 127.9 (CH), 135.09 (CH), 136.92 (C), 137.14 (CH), 141.97
(C), 149.42 (CH), 155.93 (C), 157.17 (C).
4.6. Nucleophilic substitution at C-2 position
4.7. Preparation of the valyl ester from 8a
A mixture of compound 7 (0.756 g, 2 mmol) and aminodiol
(16 mmol) was heated neat or in presence of NEt3 at 170 ꢀC until
completion of the reaction as indicated by TLC. The mixture was
4.7.1. Esterification of 8a with Boc-L-valine
A solution of BocevaleOH (0.434 g, 2 mmol), HOBt (0.229 g,
1.7 mmol) in 30 mL THFeAcOEt (2:1), the medium was cooled to

J. N’gompaza-Diarra et al. / European Journal of Medicinal Chemistry 56 (2012) 210e216
215
0 ꢀC and DCC (0.364 g, 1.7 mmol) was added. The mixture was then
stirred to r.t for 2.5 h. The solid (dicyclohexylurea) was filtered off
and the solution introduced into a round-necked flask containing
with the diol (0.257 g, 0.59 mmol) and triethylamine (0.2 mL) in
THF 20 mL. The solution was allowed to stir 24 h at r.t. THF was
evaporated, the residue dissolved in ethyl acetate, washed with
20 mL citric acid 1 M, a saturated solution of Na2CO3 and finally
with 20 mL H2O. The organic layer was dried over anhydrous
Na2SO4, filtered and evaporated under vacuum to provide yellow
light crystals which were purified by chromatography on silica gel
column, eluent: AcOEt/Toluene/EtOH (5/2/0.3e0.5) and afforded
using Buffer D (10 mM MgCl2, 1 mM EGTA (ethyleneglycoltetra-
acetic acid), 1 mM DTT (dithiothreitol), 25 mM Tris/HCl, 50 g/mL
heparin) at 30 ꢀC, at a final ATP concentration of 15
M. Blank
m
m
values were subtracted and activities expressed as % of the maximal
activity in the absence of inhibitors. Controls were performed with
appropriate dilutions of DMSO. IC50 values were calculated from
dose-response curves. DYRKs (human) and CLKs (mouse)
(recombinant, expressed in E. coli) were extracted from bacteria
and purified by affinity chromatography on glutathione agarose.
Kinase activity was assayed with either 1 mg/mL of RS peptide
(GRSRSRSRSRSR, Proteogenix, Oberhausbergen, France), in the
white crystals (190 mg, 51%).1H NMR (CDCl3)
d
: 0.8e1 (m, 6H,
presence of 15
a final volume of 30
m
M [
g
-33P] ATP (3,000 Ci/mmol; 10 mCi/mL) in
m
2CH3), 1.4 (s, 9H, 3CH3),1. 5 (s, 6H, 2CH3), 2.1 (m, 1H, NH2eCHeiPr),
3.5e3.8 (m, 2H, CH2eNH), 4.05 (m, 1H, CHOH), 4.2 (m, 2H,
CH2OCOꢁ), 4.65 (m, 1H, CONCHCOꢁ), 4.8 (s, 2H, CH2ephe), 5.05 (s,
1H, CHeiPr), 7 (m, 1H, Hpyridyl), 7.5 (d, 2H, Hphenyl), 7.65 (s, 1H, H-8),
L. After 30 min incubation at 30 ꢀC, the
reaction was stopped by harvesting onto P81 phosphocellulose
papers (Whatman) using a FilterMate harvester (Packard) and
papers were washed in 1% phosphoric acid. Scintillation fluid was
added and the radioactivity measured in a Packard counter.
7.75 (m, 2H, Hpyridyl), 7.95 (d, 2H, Hphenyl), 8.7 (m, 1H, Hpyridyl). 13
C
NMR (CDCl3)
d : 17.65(CH3), 19.07 (CH3), 22.55 (CH3), 22.57 (CH3),
22.75(CH3), 28.33, 31.22, 45.88 (CH2), 46.33 (CH), 46.58 (CH), 58.71
(CH), 66.34 (CH2), 70.78 (CH), 79.85(CH), 120.47 (CH), 122.10 (CH),
125.3 (C), 127.15 (C), 128.01 (C), 128.06 (C), 128.23 (CH), 129.04 (CH),
134.82 (C), 136.75 (CH), 138.55 (CH), 139.49 (CH), 149.68 (CH),
154.88 (C), 155.76 (C), 157.10 (C), 160.21(C).
4.10. Cell culture
SH-SY5Y human neuroblastoma cell line was grown in DMEM
supplemented with 2 mM L-glutamine (Invitrogen, Cergy Pontoise,
France) plus antibiotics and a 10% volume of FCS (Invitrogen). Cell
viability was determined by means of the MTS (3-(4,5-
dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfo-
phenyl)-2H-tetrazolium) method as previously described [28].
4.7.2. [(2R)-2-hydroxy-3-[[9-iso-propyl-6-[[4-(2-pyridyl)phenyl]
methylamino]purin-2yl]amino]propyl] (2S)-valyl ester
hydrochloride (10)
1H NMR (CDCl3)
d : (CH2CHCH2), 4.25 (m, 1H, CHOH), 4.75 (m,
4.11. Cytochrome P450 inhibition (CYP 450)
1H, NCHN), 4.8 (m, 2H, CH2eVal), 4.9 (m, 1H, CHeiPr), 7.45 (m, 1 H,
H-8), 7.6 (d, J ¼ 4.68 Hz, 2H, Hpyridyl), 7.9 (m, 2H, H-8), 8.1 (d,
J ¼ 3.6 Hz, 2H, Hphenyl), 8.25e8.4 (m, 2H, Hpyridyl, Hphenyl), 8.7 (d,
The CYP inhibition was determined by a specialized company
following a classical procedure. Briefly microsomal enzyme inhi-
bition was determined by incubation of microsomes the corre-
J ¼ 4.8 Hz, 1H, Hpyridyl). 13C NMR (DMSO)
d : 17.44(CH3), 18.03(CH3),
21.65(CH3), 29.24 (CH), 57.42 (CH), 60.15 (CH2), 66.87 (CH), 67.63
(CH2), 71.96 (CH2), 120.86 (CH), 122.93 (CH), 126.73(CH), 127.87
(CH), 138.42 (CH), 148.60, 155.15 (C), 168.60 (C).
sponding substrate and the compound under study (5 mM).
4.12. Plasma binding protein
4.8. Docking analysis
The product was added to a chamber containing mouse or
human plasma separated from a buffer chamber by a dialysis
membrane. The quantification of the concentration in buffer was
used to calculate the fraction bound to plasma proteins.
Energy-minimization resorted to the 3 Å resolution X-ray
crystal structure of CDK9 complexed with a sub-micromolar
inhibitor (PDB access code RSCB 057590) [16]. It was done using
the DISCOVER (Accelrys) software with the Cff91 force-field [23]
and the BFGS minimizer with a dielectric constant of 1. As in
our previous work using this software [24], the protein backbone
was frozen, while the side-chains of the residues of the recogni-
tion site and the entirety of the ligand were relaxed. The structural
features resulting from these computations are, as such, prelimi-
nary and only aim to show a structural compatibility of this ligand
with the CDK9 binding site. Detailed energy balances bearing on
a series of ligands related to it, also involving structural water
molecules, are planned in the context of the polarizable molecular
mechanics procedure SIBFA [25]. This procedure has previously
been used to investigate the complexation of several related
ligands to protein targets such as Znemetalloeproteins [26] and
the FAK kinase [27].
References
[1] J.W. Harper, P.D. Adams, Chem. Rev. 101 (2001) 2511e2526.
[2] J. Bloom, F.R. Cross, Nat. Rev. Mol. Cell. Biol. 8 (2007) 149e160.
[3] H. Kennedy, C. Dehay, Nat. Rev. Neurosci. 8 (2007) 438e450.
[4] A.G. Rossi, D.A. Sawatzky, A. Walker, Nat. Med. 12 (2006) 1056e1064.
[5] I. Collins, M.D. Garrett, Curr. Opin. Pharmacol. 5 (2005) 366e373.
[6] L.M. Schang, Biochim. Biophys. Acta 1697 (2004) 197e209.
[7] N.O. Bukanov, L.A. Smith, K.W. Klinger, S.R. Ledbetter, O. Ibraghimov-
Beskrovnaya, Nature 444 (2006) 949e952.
[8] (a) E.A. Monaco, M.L. Vallano, Front. Biosci. 10 (2005) 143e159;
(b) J.R. Sundaram, E.S. Chan, C.P. Poore, T.K. Pareek, W.F. Cheong, G. Shui, N. Tang,
C.-M. Low, M.R. Wenk, S. Kesavapany, J. Neurosci. 32 (2012) 1020e1034.
[9] (a) M. Knockaert, P. Greengard, L. Meijer, Trends Pharmacol. Sci. 23 (2002)
417e425;
(b) H. Galons, N. Oumata, L. Meijer, Expert Opin. Ther. Pat 19 (2010) 377e403.
[10] N.S. Gray, L. Wodicka, A.W.H. Thunnissen, T.C. Norman, S. Kwon, F.H. Espinoza,
D.O. Morgan, G. Barnes, S. LeClerc, L. Meijer, S.-H. Kim, D.J. Lockhart,
P.G. Schultz, Science 281 (1998) 533e538.
4.9. Kinase inhibition assays
[11] F. Kashanchi, K. Kehn-Hall, Curr. Pharm. Des. 15 (2009) 2520e2532.
[12] B.A. Christian, K.A. Blum, L.A. Andritsos, B. Fischer, M.E. Moran, M. Brooker-
McEldowney, M.A. Phelps, J.T. Dalton, A.D. Colevas, J.C. Byrd, M.R. Grever,
T.S. Lin, Curr. Opin. Oncol. 19 (2007) 573e578.
[13] M. Malumbres, P. Pevarello, M. Barbacid, J.R. Bischoff, Pharmacol. Sci. 29
(2008) 16e21.
[14] L.R. Kelland, Cyclin-dependent kinase inhibitors and combination therapy:
experimental and clinical status, in: P.J. Smith, E.W. Yue (Eds.), Inhibitors
of Cyclin-Dependent Kinases as Anti-Tumor Agents, CRC Press, 2007, pp.
371e388.
CDK1/cyclin B (native affinity purified from starfish oocytes),
CDK2/cyclin A, CDK7/cyclin H, CDK9/cyclin T (human recombinant,
expressed in baculovirus-infected insect cells), and CDK5/p25
(human recombinant, expressed in E. coli) were assayed in the
presence of 15
were determined from dose-response curves and are expressed in
M. DYRK1A and CLK Kinase activities were assayed in duplicate
mM ATP as previously described [16]. IC50 values
m

216
J. N’gompaza-Diarra et al. / European Journal of Medicinal Chemistry 56 (2012) 210e216
[15] N. Oumata, K. Bettayeb, Y. Ferandin, L. Demange, A. Lopez-Giral, M.-
L. Goddard, L.V. Myrianthopoulos, E. Mikros, M. Flajolet, P. Greengard,
L. Meijer, H. Galons, J. Med. Chem. 51 (2008) 5229e5242.
[16] K. Bettayeb, N. Oumata, A. Echalier, Y. Ferandin, J. Endicott, H. Galons,
L. Meijer, Oncogene 27 (2008) 5797e5807.
[17] H.C. Kolb, M.S. VanNieuwenhze, K.B. Sharpless, Catalytic asymetric dihy-
droxylation, Chem. Rev. 94 (1994) 2483e2547.
[18] N. Oumata, L. Meijer, Y. Ferandin, H. Galons, Practical synthesis of roscovitine
and CR8, Org. Process. Res. Dev. 13 (2009) 641e644.
[22] (a) J.P. Smith, S. Weller, B. Johnson, J. Nicotera, J.M. Luther, D.W. Haas, Anti-
microb. Agents Chemother. 54 (2010) 1146e1151;
(b) E.Y. Kwak, W.S. Shim, J.E. Chang, S. Chong, D.D. Kim, S.J. Chung, C.K. Shim,
Xenobiotica 42 (2012) 603e613.
[23] Accelrys Inc 9685 Scranton Road, San Diego, CA 92121e93752.
[24] J.E. Ombetta, N. Thelier, C.Z. Dong, S. Plocki, L. Tsagris, F. Rannou, F. Massicot,
A. Djimde, E. El-Hayek, Y. Shi, F. Heymans, N. Gresh, C. Chauvet, PLOS One 6
(2010) e10914.
[25] N. Gresh, G. Andres Cisneros, T.A. Darden, J.-P. Piquemal, J. Chem. Theory
Comput. 3 (2007) 1960e1986.
[26] B. de Courcy, J.-P. Piquemal, C. Garbay, N. Gresh, J. Am. Chem. Soc. 132 (2010)
3312e3320.
[19] D. Mahadevan, R. Plumme, M.S. Squires, D. Rensvold, S. Kurtin, C. Pretzinger,
T. Dragovich, J. Adams, V. Lock, D.M. Smith, D. Von Hoff, H. Calvert, Ann. Oncol.
22 (2011) 2137e2143.
[20] D. Wu, W. Pan, Biochem. Sci. 35 (2009) 161e168.
[21] S. Malinge, M. Bliss-Moreau, G. Kirsammer, L. Diebold, T. Chlon, S. Gurbuxani,
J.D. Crispino, J. Clin. Invest. 12 (2012) 948e962.
[27] N. Gresh, B. de Courcy, J.-P. Piquemal, J. Foret, S. Courtiol-Legourd, L. Salmon,
J. Phys. Chem. B 115 (2011) 8604e8616.
[28] J. Ribas, J. Boix, Exp. Cell Res. 29 (2004) 9e24.