6-Cyclohexylmethoxy-5-(cyano-NNO-azoxy)pyrimidine-4-amine: A new scaffold endowed with potent CDK2 inhibitory activity

Article abstract of DOI:10.1016/j.ejmech.2013.07.031

Substitution of the cyano-NNO-azoxy moiety (NC-N(O)N-) for the nitroso group in NU6027, a potent and selective CDK2 inhibitor, affords a compound with slightly improved potency and comparable selectivity profile. A molecular modelling study indicates for this new scaffold a binding mode similar to the one adopted by other purine and pyrimidine analogues, and suggests a relevant role for a conserved water molecule in stabilizing the bioactive pose of this and other pyrimidine ligands. The introduction of aminosulfonylphenyl substituents on the 2-amino group of the pyrimidine increased the CDK2 inhibitory potency by two orders of magnitude, while maintaining the same degree of selectivity.

Full text of DOI:10.1016/j.ejmech.2013.07.031

European Journal of Medicinal Chemistry 68 (2013) 333e338
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Short communication
6-Cyclohexylmethoxy-5-(cyano-NNO-azoxy)pyrimidine-4-amine:
A new scaffold endowed with potent CDK2 inhibitory activity
,
Donatella Boschi a, Paolo Tosco a, Naveen Chandra a, Shilpi Chaurasia a, Roberta Fruttero a
*
,
Roger Griffin b, Lan-Zhen Wang b, Alberto Gasco a
a Dipartimento di Scienza e Tecnologia del Farmaco, Università degli Studi di Torino, via Pietro Giuria 9, 10125 Torino, Italy
b Newcastle Cancer Centre, Northern Institute for Cancer Research and School of Chemistry, Newcastle University, Newcastle upon Tyne, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Substitution of the cyano-NNO-azoxy moiety (NCeN](O)Ne) for the nitroso group in NU6027, a potent
and selective CDK2 inhibitor, affords a compound with slightly improved potency and comparable
selectivity profile. A molecular modelling study indicates for this new scaffold a binding mode similar to
the one adopted by other purine and pyrimidine analogues, and suggests a relevant role for a conserved
water molecule in stabilizing the bioactive pose of this and other pyrimidine ligands. The introduction of
aminosulfonylphenyl substituents on the 2-amino group of the pyrimidine increased the CDK2 inhibitory
potency by two orders of magnitude, while maintaining the same degree of selectivity.
Ó 2013 Elsevier Masson SAS. All rights reserved.
Received 23 March 2013
Received in revised form
6 June 2013
Accepted 17 July 2013
Available online 11 August 2013
Keywords:
CDK inhibitor
CDK2 selective inhibitor
Cyano-NNO-azoxy
5-Substituted pyrimidine-4-amine
Water-mediated hydrogen bond
1. Introduction
nitrosopyrimidine-2,4-diamine, NU6027; Fig. 1) has been devel-
oped as a potent and selective CDK2 inhibitor. In this structure, the
The division of eukaryotic cells occurs in four phases (G1, S, G2,
M) and cell-cycle progression is regulated by several members of a
family of cyclin-dependent serine/threonine kinases (CDKs). To
date, eleven members of this family have been identified, respec-
tively named CDK1 (also known as cdc2), and CDK2eCDK11 [1,2].
The activity of these kinases is dependent on their cyclin partners.
The principal CDK/cyclin pairs involved in regulating cell division
are CDK1/cyclinB, CDK2/cyclinA, CDK2/cyclinE, and CDK4,6/cyclins
of the D family. The inhibition of CDKs by small molecules con-
tinues to attract considerable attention as a strategy to exploit in
developing anticancer drugs [3e5]. A variety of these inhibitors
have been described, some of which are currently under clinical
evaluation [4]; an oral selective inhibitor for CDK 4 and 6, palbo-
ciclib [6], has recently received Breakthrough Therapy designation
by the FDA for potential treatment of patients with breast cancer.
presence of an intramolecular hydrogen bond between the adja-
cent 5-nitroso and 4-amino groups forces the molecule to assume
a pseudo-purine geometry, which is reminiscent of the purine
scaffold 2 (6-(cyclohexylmethoxy)-9H-purine-2-amine, NU2058;
Fig. 1), another selective CDK2 inhibitor [7]. X-ray analysis of the
CDK2/NU2058 and CDK2/NU6027 crystal complexes shows that the
two products establish the same key interactions with the back-
bone of amino acid residues within the ATP-binding site of CDK2,
namely a triplet of H-bonds (2-NH2 to Leu83, N3 to Leu83, N9eH to
Glu81, and 2-NH2 to Leu83, N1 to Leu83, 4-NH2 to Glu81, respec-
tively) [7]. Structureeactivity relationships (SAR) have been studied
in detail for these two leads [8e10]. In particular, it has been
observed that a p-aminosulfonylphenyl substituent on the 2-NH2
group (e.g. 3, NU6102; Fig.1) confers high inhibitory potency, due to
additional hydrophobic and hydrogen bonding interactions within
the binding pocket. Extensive structural modification of scaffold 1
has confirmed the crucial role of the nitroso group at the 5-position
of the pyrimidine; only nitro, formyl, and acetyl moieties are
tolerated as possible alternatives at this position [10]. In this paper,
we show how a potent and quite selective CDK2 inhibitor can be
obtained by replacing the NO group in 1 with a cyano-NNO-azoxy
moiety (NCeN(O)]Ne, compound 5; Fig. 1). This function was
The chemotype exemplified by
1 (6-(cyclohexylmethoxy)-5-
* Corresponding author. Tel.: þ39 (0)116707850; fax: þ39 (0)116707286.
E-mail address: roberta.fruttero@unito.it (R. Fruttero).
0223-5234/$ e see front matter Ó 2013 Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.ejmech.2013.07.031

334
D. Boschi et al. / European Journal of Medicinal Chemistry 68 (2013) 333e338
Fig. 1. Structures of compounds 1e5.
discovered while studying an antibiotic isolated from three
different sources, Calvatia lilacina, Calvatia craniformis and Lyco-
perdon pyriforme; the antibiotic was assigned the 4-[(Z)-cyano-
NNO-azoxy]benzoic acid structure 4 [11e13]. The synthesis, and the
structural and biological characterization of some lead compounds
based on this new scaffold are reported below. Structureeactivity
relationships are also discussed, and a binding mode hypothesis
based on a molecular modelling study is proposed.
Fig. 2. Synthetic pathway leading to final products 5, 6 and 9. Reagents and condi-
tions: a) NH2CN, IBA, dry CH3CN; b) HCl gas, THF/H2O, 0 ꢀC; c) Ac2O, 50 ꢀC; d) TsNH2,
IBA, dry CH2Cl2/CH3CN 2/1, 40 ꢀC; e) ZnCl2, EtOH, 65 ꢀC.
starting from 15. For the purpose of the present study, the 1H NMR
resonances of the 4-amino groups in compounds 1, 5, 6, 9, 14 and
16 deserve some comments. The signal of the 4-NH2 group in
derivative 1, recorded in DMSO-d6 at room temperature, is split
into two neat signals, which fall at 10.09 and 8.01 ppm, respec-
tively. The unusually low-field signal at 10.09 ppm can be assigned
to the NeH proton chelated by the NO group [16]: the two protons
experience quite different local magnetic fields, as a consequence
of the strong anisotropic effect exerted by the nitroso function.
2. Chemistry
The preparation of the final products 5, 6, and 9 is outlined in
Fig. 2. The scaffold 5 was easily obtained using a method described
elsewhere [14]. This procedure involves treating the 5-nitroso de-
rivative 1 with (diacetoxyiodo)benzene (IBA) and cyanamide
(NH2CN) in dry CH3CN. The action of HCl on 5, dissolved in a THF/
H2O mixture, afforded the related amide 6. The best yields in the
preparation of the final (4-methylphenyl)sulfonyl derivative 9 were
obtained by protecting the two amino groups in 1 as acetyl de-
rivatives. When the reaction was carried out starting directly from
1, compound 7 was obtained. It is known that derivatives of the
[1,2,5]oxadiazolo-[3,4-d]pyrimidine scaffold, such as compound 7,
can be prepared in anhydrous DMF by action of a slight excess of
IBA on the related 5-nitrosopyrimidine-4-amines, in the presence
of LiH [15]. Conversely, a mixture of derivatives 8 and 7a was ob-
tained by treating the diacetylated intermediate 1a with (4-
methylphenyl)sulfonamide (TsNH2) and IBA dissolved in CH2Cl2/
CH3CN.
Deprotection of the amino groups in 8, carried out with ZnCl2
in ethanol, yielded the desired final product 9. The preparation of
final derivatives 14 and 16, bearing substituted (aminosulfonyl)
phenyl moieties at the N2 position, is outlined in Fig. 3. A
mixture of n-butylsulfonylpyrimidine 10 and 4-amino-N-meth-
ylbenzenesulfonamide 11 dissolved in 2,2,2-trifluoroethanol (TFE)
was treated with trifluoroacetic acid (TFA) to afford 12. Nitro-
sation of the latter yielded the related 5-nitroso-substituted
compound 13, which in turn, upon treatment under the same
conditions used to prepare 5 from 1, yielded the expected product
14. Compound 16 was obtained following the same procedure,
Fig. 3. Synthesis of compounds 14 and 16. Reagents and conditions: a) 11, TFA/TFE,
reflux; b) NaNO2, AcOH/H2O, 100 ꢀC; c) NH2CN, IBA, dry CH3CN.

D. Boschi et al. / European Journal of Medicinal Chemistry 68 (2013) 333e338
335
Table 1
solvent was carried out for all compounds, after which the most
stable geometries were subjected to quantum-mechanical (QM)
refinement. Subsequently, flexible docking was performed. Firstly,
the three-dimensional structure of CDK2 complexed with NU6027
(PDB ID 1E1X) [7] was chosen as the model system. While we were
able to reproduce the co-crystallized pose of NU6027 within a root-
Inhibitory activity of compounds 1, 5, 6, 9, 14, 16 against different CDK isoforms.
CDK isoform
IC50 (m
M)a
1
5
6
9
14
16
CDK2/A
CDK1/B
CDK4/D
CDK7/H
CDK9/T
CDK5/P25
2.2 ꢁ 0.6b
2.9 ꢁ 0.1b
24%e
0.94
ed
53.3
ed
48%c
ed
0.0039
0.14
1.05
1.5
1.66
0.775
0.0025
0.071
0.99
1.55
0.875
0.56
ꢀ
ed
ed
ed
mean-square (RMS) tolerance of 0.76 A, we failed to dock the newly
26%e
ed
ed
ed
synthesized analogue 5. When 5 was manually superposed on
NU6027, a steric clash between the 5-cyanoazoxy group and Lys33
was observed, which prevented its docking. We analysed all crystal
structures of CDK2 available in the Protein Data Bank, finding that
the side chain of Lys33 is quite flexible and can assume different
conformations: in most of the structures this residue does not
protrude into the active site as it does in 1E1X. Therefore, the 1E1X
structure was replaced by 2C6O [20], because of its good resolution
and of the resemblance of the co-crystallised inhibitor NU6102 (3,
Fig. 1) to our compounds. Again, we were able to reproduce the
crystallographic pose of NU6102 inside the binding site within an
36%e
ed
ed
ed
ed
ed
ed
ed
a
CDK-inhibition was determined in duplicate at a minimum of five different
inhibitor concentrations.
b
Ref. [17].
Percent inhibition at 100
Not tested.
c
m
M.
d
e
Percent inhibition at 10
m
M.
Conversely, in models 5, 6, 9, 14 and 16 the 4-NH2 signals appear
as a broadened singlet.
ꢀ
RMS deviation of 0.30 A. Additionally, both NU6027 and 5 could be
docked in an orientation very similar to that occurring with
NU6027 in 1E1X.
3. Results and discussion
The pyrimidine derivatives were evaluated for CDK inhibitory
activity using published procedures [17]. The results are shown in
Table 1, together with the inhibitory potencies of compound 1,
taken as reference. Analysis of the data shows that the new scaffold
5 displays an inhibitory potency on the CDK2 isoform that is about
twice that of reference 1. The substitution of the cyano group on the
azoxy function with different electron-withdrawing moieties, such
as the amide and phenylsulfonyl groups (products 6 and 9,
respectively), induces a marked decrease in the inhibitory potency
in both cases; this is more pronounced in 9, probably as a conse-
quence of the excessive steric bulk of the phenylsulfonyl moiety.
Compounds 14 and 16 were designed based on the knowledge that
the introduction of RNHSO2C6H5e groups on the 2-NH2 group of 1
affords very potent and selective CDK2 inhibitors, as discussed
above. Indeed, the same trend is observed when this structural
manipulation is carried out on scaffold 5: products 14 and 16 are,
respectively, about 240- and 400-fold more potent than the lead.
Moreover, they are highly selective for CDK2, being some thirty
times more potent compared with CDK1 and over two orders of
magnitude with respect to the other isoforms. The same com-
pounds were evaluated for growth inhibitory activity on three
different cell lines, namely A2780 human ovarian cancer, MCF-7
breast cancer, and SKUT-1B human uterine sarcoma [17]. Both py-
rimidines 14 and 16 exhibited growth inhibitory activity against all
three cell lines in the low micromolar concentration range
(Table 2). As expected, these potencies are modest in comparison
with the corresponding activity as inhibitors of CDK2, for reasons
that have yet to be fully established. It is possible that this
discrepancy arises as a consequence of poor cellular permeability of
the pyrimidines [10], or may reflect the partial redundancy of CDK2
as a driver of tumour cell proliferation, as has been reported pre-
viously [18,19].
An attempt to rationalize the activity profile, and to hypothesize
a binding mode for the newly-designed analogues, was made
through molecular modelling. A conformational search in implicit
Table 2
Growth inhibitory potency of compounds 14, 16 against selected tumour cell lines.
Compd
GI50
(
m
M)a
Fig. 4. Binding modes of compounds 5 (a) and 14 (b) (thick sticks) in the active site of
CDK2 (PDB structure ID: 2C6O); the pose of the co-crystallized inhibitor 3 (thin sticks)
is reported for comparison. The binding site is put into evidence through its Connolly
A2780
MCF7
Skut-1B
14
16
1.5 ꢁ 0.6
2.2 ꢁ 0.8
1.7 ꢁ 0.5
3.5 ꢁ 0.6
1.5 ꢁ 0.06
2.4 ꢁ 0.06
surface, while secondary structure elements are represented as flat ribbons (
segments) and tubes (loops). Hydrogen bonds are depicted with dashed lines; their
respective energy magnitudes, as computed by MOE, are reported in kcal molꢂ1
b-sheet
a
Concentration inhibiting cell growth by 50% of control.
.

336
D. Boschi et al. / European Journal of Medicinal Chemistry 68 (2013) 333e338
Analysing the docking results of 5 against CDK2, the N-oxide
NMR signals: s (singlet), d (doublet), t (triplet), q (quartet), br
(broad), vbr (very broad). Flash column chromatography was per-
formed on silica gel (Merck Kieselgel 60, 230e400 mesh ASTM)
using the indicated eluents. Thin layer chromatography (TLC) was
carried out on 5 ꢃ 20 cm plates with 0.25 mm layer thickness.
Anhydrous MgSO4 was used as drying agent for the organic phases.
Analyses (C, H, N) of the new compounds were performed by
REDOX (Monza, Italy); the results were within ꢁ0.4% of the theo-
retical values. Compounds 1 [7], 10 [9], 11 [21], 15 [9] were syn-
thesized following methods described in the literature.
moiety was found to interact with the side chain carboxyl group of
Asp145 via a crystal water molecule, while the cyano group nitro-
gen establishes a charge-enhanced hydrogen bond with the side
chain of Lys33. NU6027 also binds in a similar way, forming a water-
mediated hydrogen bond between the nitroso moiety and Asp145
(Fig. 4). As mentioned above, the 5-nitroso group is essential for
conferring activity to pyrimidine ligands, enabling, through an
intramolecular hydrogen bond, the 6-amino group to establish
optimal interactions with Glu81 [7]. To challenge our hypothesis
regarding the involvement of a water molecule in binding, we
analysed the available crystallographic structures of CDK2 in the
Protein Data Bank, and found the water molecule to be highly
conserved. When we docked both 5 and NU6027 in the absence of
this crystal water, we could not find any pose reminiscent of
the crystallographic pose of NU6027. To further corroborate
this hypothesis, we built a model of 4-cyclohexylmethoxy-6-
aminopyrimidine, namely the NU6027 analogue lacking the 5-
nitroso substituent; again, docking was unsuccessful even after
constraining the amino group in the same conformation as in
NU6027. These findings suggest that the conserved water molecule
may significantly contribute to the binding of pyrimidine ligands
via an interaction with an appropriate hydrogen bond acceptor at
the 5-position. Moreover, the favourable binding affinity of 5 may
benefit from the additional, strong hydrogen bonding interaction
between the cyano group and the positively charged side chain of
Lys33. In the case of compounds 6 and 9, a small number of pro-
ductive conformers with low docking scores were found; in
particular, the phenylsulfonyl group in 9 is barely compatible with
the steric constraints imposed by the size of the cavity. The
sulfonamide-substituted derivatives 14 and 16 dock in the active
site of CDK2 in an orientation largely superimposable upon NU6102
and 5, with better docking scores. In addition to the interactions
described above, the sulfonamido substituent has an additional
hydrogen bonding interaction with the side chain of Lys89, hence
the improved affinity against CDK2.
5.1.1. 5-(Cyano-NNO-azoxy)-6-(cyclohexylmethoxy)pyrimidine-
2,4-diamine (5)
To the stirred suspension of 1 (0.502 g, 2 mmol) and cyanamide
(0.252 g, 6 mmol) in CH3CN (25 mL), IBA was added (0.772 g,
2.4 mmol) in portions at r.t. The colour of the reaction mixture
changed gradually from purple to yellow. The precipitated yellow
solid was filtered off, washed twice with CH3CN and recrystallized
to obtain 2 as a yellow crystalline solid (0.344 g; yield 59%). M.p.
203e204 ꢀC (CH3CN); 1H NMR (300 MHz, DMSO-d6)
d 1.14 (m, 5H,
C6H11), 1.77 (m, 6H, C6H11), 4.13 (d, J ¼ 6 Hz, 2H, CH2O), 7.44 (br s,
1H, NH2, D2O exchangeable), 7.50 (br s, 1H, NH2, D2O exchange-
able), 8.13 (br s, 2H, NH2, D2O exchangeable); 13C NMR (75 MHz,
DMSO-d6)
d 25.1, 25.9, 28.9, 36.4, 72.0, 106.2, 112.3, 159.2, 160.8,
164.2; MS (EI, 70 eV) m/z 291 (Mþ), 251 (Mꢂ0, 100%), 155, 138. Anal.
C12H17N7O2 (C, H, N).
5.1.2. 5-(Aminocarbonyl-NNO-azoxy)-6-(cyclohexylmethoxy)
pyrimidine-2,4-diamine (6)
Hydrogen chloride was bubbled during a few minutes into a
THF/H2O (6/1, 30 mL) solution of 5 (0.291 g, 1 mmol) kept at 0 ꢀC.
EtOAc (10 mL) and a saturated solution of NaHCO3 (20 mL) were
then added, and the cooled reaction mixture was neutralized with
6 M NaOH under stirring. The aqueous phase was separated and
extracted twice with EtOAc. The combined organic phases were
dried with brine, then with MgSO4, and evaporated to give a green
oil that was purified by flash chromatography (CH2Cl2/MeOH 95/5).
The pure yellow solid thus obtained (0.145 g; yield 47%) was
recrystallized. M.p. 149e150 ꢀC (MeOH/H2O); 1H NMR (300 MHz,
4. Conclusion
In this paper we describe how replacing the nitroso group in the
NU6027 lead compound with a cyano-NNO-azoxy function resulted
in derivatives having potent inhibitory activity against CDKs, with
pronounced selectivity for the CDK2 isoform. Substitution of the
cyano group with either carbamoyl or phenylsulfonyl moieties
failed to improve affinity, while p-aminosulfonylphenyl substitu-
tion on the 2-NH2 function proved beneficial, in accordance with
established SARs for NU6027. A molecular modelling study showed
that compounds 5, 14 and 16 may adopt a binding mode similar to
that determined experimentally for other purine and pyrimidine
analogues. This investigation also suggested a significant role for a
conserved water molecule in stabilizing the bioactive pose of py-
rimidine ligands, through interaction with a suitable hydrogen
bond acceptor substituent at the 5-position.
DMSO-d6) d 1.15 (m, 2H, C6H11), 1.22 (m, 3H, C6H11), 1.70 (m, 6H,
C6H11), 4.03 (d, J ¼ 6 Hz, 2H, CH2O), 6.60 (br s, 2H, NH2, D2O
exchangeable), 6.79 (br s, 2H, NH2, D2O exchangeable), 7.53 (br s,
1H, NH2, D2O exchangeable), 7.63 (br s, 1H, NH2, D2O exchange-
able); 13C NMR (75 MHz, DMSO-d6)
d 25.2, 25.8, 28.8, 36.6, 70.8,
106.3,157.7,158.7,160.7,162.3; MS (CI, 70 eV) m/z 310 (Mþþ1). Anal.
C12H19N7O3 (C, H, N).
5.1.3. 7-(Cyclohexylmethoxy)-[1,2,5]oxadiazolo[3,4-d]pyrimidine-
5-amine (7)
To the stirred solution of 1 (0.251 g, 1 mmol) and TsNH2 (0.513 g,
3 mmol) in CH3CN (12 mL), IBA was added (0.386 g, 1.2 mmol) in
portions at r.t. The colour of the reaction mixture gradually changed
from purple to grey to green. The precipitate was filtered off,
washed twice with CH3CN and dried to obtain 7 (0.174 g; yield 70%).
5. Experimental protocols
M.p.198e199 ꢀC (dec.); 1H NMR (300 MHz, DMSO-d6)
d 1.20 (m, 5H,
C6H11), 1.71 (m, 6H, C6H11), 4.38 (d, J ¼ 5.7 Hz, 2H, CH2O), 7.90 (br s,
5.1. Chemistry
1H, NH2, D2O exchangeable), 7.82 (br s, 1H, NH2, D2O exchange-
able); 13C NMR (75 MHz, DMSO-d6)
d 162.7, 160.4, 135.0, 72.8, 36.2,
Melting points (m.p.) were measured with a capillary apparatus
(Büchi 540). M.p. with decomposition were determined after
placing the sample in a bathꢂa1t a temperature 10 ꢀC below the m.p.;
a heating rate of 3 ꢀC min was applied. All compounds were
routinely checked by FT-IR (PerkineElmer SPECTRUM BXII), 1H and
13C NMR (Bruker Avance 300) and mass spectrometry (Finnigan-
Mat TSQ-700). The following abbreviations were used for labelling
28.9, 25.8, 24.7; MS (EI, 70 eV) m/z 249 (Mþ), 154 (100%), 97.
5.1.4. N,N0-[6-(Cyclohexylmethoxy)-5-nitrosopyrimidine-2,4-diyl]
diacetamide (1a)
A solution of 1 (0.251 g, 1 mmol) in Ac2O (3 mL) was stirred at
50 ꢀC for 20 h under dry atmosphere. The solution was then slowly
poured into a stirred iceewater mixture; the separated green oil

D. Boschi et al. / European Journal of Medicinal Chemistry 68 (2013) 333e338
337
was triturated to transform it into a green solid, which was then
collected on a filter, washed twice with water and dried in vacuo on
P2O5 for 3 days (0.299 g; yield 89%). M.p. 171e172 ꢀC; 1H NMR
the mixture was extracted with EtOAc. The combined organic layers
were washed with H2O, followed by brine, dried over MgSO4, and
the solvent was removed in vacuo. The crude oil was purified by
flash chromatography (CH2Cl2/MeOH 9.7/0.3) to obtain an addi-
tional crop of 12 as the free base (0.700 g; overall yield 59%). M.p.
(300 MHz, DMSO-d6) d 1.23 (m, 5H, C6H11), 1.80 (m, 6H, C6H11), 2.34
(s, 3H, CH3), 2.57 (s, 3H, CH3), 4.47 (d, J ¼ 6.6 Hz, 2H, CH2O),11.20 (br
s, 1H, NH, D2O exchangeable), 11.53 (br s, 1H, NH, D2O exchange-
210e212 ꢀC (MeOH); 1H NMR (300 MHz, DMSO-d6)
d 0.99e1.26 (m,
able); 13C NMR (75 MHz, DMSO-d6)
d
25.5, 26.0, 26.3, 27.5, 29.4,
5H, C6H11), 1.69e1.77 (m, 6H, C6H11), 2.39 (s, 3H, CH3), 4.03 (d,
J ¼ 6 Hz, 2H, CH2O), 5.35 (s 1H, H5), 6.90 (vbr s, 2H, NH2, D2O
exchangeable), 7.23 (br s, 2H, NH2, D2O exchangeable), 7.63 (d,
J ¼ 8.7 Hz, 2H, ArH), 7.93 (d, J ¼ 9.0 Hz, 2H, ArH), 9.67 (s,1H, NH, D2O
37.1, 73.5, 139.8, 144.7, 159.7, 169.6, 170.5, 173.1; MS (EI, 70 eV) m/z
335 (Mþ), 197 (100%), 155, 84.
5.1.5. N,N0-{6-(Cyclohexylmethoxy)-5-{[(4-methylphenyl)sulfonyl]-
NNO-azoxy}pyrimidine-2,4-diyl}diacetamide (8) and N-[7-
(cyclohexylmethoxy)[1,2,5[oxadiazolo][3,4-d]pyrimidin-5-yl]
acetamide (7a)
exchangeable); 13C NMR (75 MHz, DMSO-d6)
d 169.4, 163.9 br, 156.5
br, 143.8 br, 130.9 br, 127.5, 118.5 br, 78.4, 71.0, 36.7, 29.1, 28.6, 25.9,
25.1; MS (EI, 70 eV) m/z 391 (Mþ), 361, 295 (100%). Anal.
C18H25N5O3S,CF3COOH (C, H, N).
To the stirred solution of 1a (0.335 g, 1 mmol) and TsNH2
(0.342 g, 2 mmol) in CH2Cl2/CH3CN 2/1 (17 mL), IBA (0.644 g,
2 mmol) was added portion-wise at 40 ꢀC. The colour of the reac-
tion mixture gradually changed from purple to orange. After 20 h,
when all starting materials had been consumed (TLC), water
(10 mL) was added to the reaction mixture, then the aqueous phase
was separated and extracted twice with CH2Cl2. The combined
organic phases were washed with 5% NaHCO3, brine and dried over
MgSO4. The solvent was removed in vacuo, and the resulting crude
oil was purified by flash chromatography (petroleum ether/acetone
7/3) to give, eluting first, 7a (0.079 g; white solid, yield 27%), and,
eluting second, the desired product 8 (0.131 g; yellow solid, yield
26%). Compound 7a: m.p. 290e291 ꢀC; 1H NMR (300 MHz, CDCl3)
5.1.8. 4-(6-Amino-5-nitroso-4-cyclohexylmethoxypyrimidin-2-
ylamino)-amino-N-methylbenzenesulfonamide (13)
A suspension of 12 (1.173 g, 3 mmol) in glacial acetic acid
(20 mL) was solubilized at 95 ꢀC, then a solution of NaNO2 (0.276 g,
4 mmol) in water (3 mL) was added drop-wise while the colour of
the reaction mixture slowly changed to green. The reaction mixture
was stirred at 95 ꢀC for 1 h, then treated with EtOAc (100 mL). The
organic phase was washed with H2O, 5% NaHCO3, H2O, brine, and
dried over MgSO4. The solvent was evaporated and the resulting
crude oil was purified by flash chromatography (CH2Cl2/iPrOH 9.6/
0.4) to afford 13 as a green crystalline solid (0.580 g; yield 46%).
M.p. 157e158 ꢀC (MeOH); 1H NMR (300 MHz, DMSO-d6)
d 1.05e
d
1.25 (m, 5H, C6H11), 1.86 (m, 6H, C6H11), 2.72 (s, 3H, CH3), 4.46 (d,
1.30 (m, 5H, C6H11), 1.70e1.83 (m, 6H, C5H11), 2.40 (d, J ¼ 5.1 Hz, 3H,
NHCH3), 4.43 (d, J ¼ 6.3 Hz, 2H, CH2O), 7.37 (q, J ¼ 5.1 Hz, 1H,
NHCH3, D2O exchangeable), 7.71 (d, J ¼ 8.4 Hz, 2H, ArH), 8.11 (br s,
2H, ArH), 8.61 (vbr s,1H, 6-NH2, D2O exchangeable),10.19 (vbr s,1H,
J ¼ 6.6 Hz, 2H, CH2O), 8.04 (br s, 1H, NH, D2O exchangeable); 13C
NMR (75 MHz, CDCl3) d 25.4, 26.0, 25.9, 26.2, 29.4, 36.8, 75.2, 135.2,
157.3, 161.7, 162.5, 171.8; MS (CI, 70 eV) m/z 292 (Mþþ1). Compound
8: m.p. 170e173 ꢀC (dec.); 1H NMR (300 MHz, CDCl3)
d
1.20 (m, 5H,
6-NH2, D2O exchangeable), 10.57 (s, 1H, NH, D2O exchangeable); 13
C
C6H11), 1.72 (m, 6H, C6H11), 2.13 (s, 3H, COCH3), 2.46 (s, 3H, COCH3),
2.61 (s, 3H, PhCH3), 4.15 (d, J ¼ 6 Hz, 2H, CH2O), 7.36 (d, J ¼ 8.1 Hz,
2H, ArH), 8.01 (d, J ¼ 8.4 Hz, 2H, ArH), 9.40 (br s, 1H, NH, D2O
exchangeable), 10.22 (br s, 1H, NH, D2O exchangeable); 13C NMR
NMR (75 MHz, DMSO-d6) d 26.0, 26.8, 29.5, 30.0, 37.6, 73.1 br, 121.3
br,128.3,134.0,141.0,143.4,150.1 br,160.4 br,171 vbr; MS (EI, 70 eV)
m/z 420 (Mþ), 324 (100%). Anal. C18H24N6O4S,0.5H2O (C, H, N).
(75 MHz, CDCl3)
d
21.8, 23.9, 25.6, 25.9, 26.2, 29.3, 36.9, 74.6, 116.6,
5.1.9. 4-[6-Amino-5-(cyano-NNO-azoxy)-4-
129.4, 130.2, 133.1, 145.6, 153.1, 155.4, 163.9, 168.9, 173.1; MS (CI,
cyclohexylmethoxypyrimidin-2-ylamino]-amino-N-
methylbenzenesulfonamide (14)
70 eV) m/z 505 (Mþþ1), 333 (100%).
To a stirred suspension of 13 (0.270 g, 0.643 mmol) and cyana-
mide (0.084 g, 2.0 mmol, 3.0 equiv) in CH3CN (3.0 mL), IBA (0.350 g,
1.28 mmol, 2.0 equiv) was added portion-wise at r.t. After 2 h the
reaction mixture was extracted with CH2Cl2 (100 mL), washed with
H2O (2 ꢃ 30 mL) and brine, dried (MgSO4) and evaporated to
dryness. The resulting crude product was purified by flash chro-
matography (CH2Cl2/MeOH 9.5/0.5), then recrystallized from iPrOH
to afford 6 as a yellow crystalline solid (0.100 g; yield 34%). M.p.
5.1.6. 6-Cyclohexylmethoxy-5-{[(4-methylphenyl)sulfonyl]-NNO-
azoxy}pyrimidine-2,4-diamine (9)
Compound 7 (0.252 g, 0.5 mmol) was dissolved in EtOH (35 mL)
and anhydrous ZnCl2 (0.239 g, 5 mmol) was added to the solution
as a single portion. The mixture was stirred at 65 ꢀC until all starting
materials had been consumed (48 h, TLC). The yellow solid formed
during the reaction was collected, washed with water, dried and
recrystallized (0.123 g; yield 59%). M.p. 227e228 ꢀC (CH3CN); 1H
140e143 ꢀC (dec.); 1H NMR (300 MHz, DMSO-d6)
d 1.09e1.24 (m,
NMR (300 MHz, DMSO-d6)
d
0.98 (m, 2H, C6H11), 1.22 (m, 3H,
5H, C6H11), 1.60e1.81 (m, 6H, C6H11), 2.40 (d, J ¼ 4.8 Hz, 3H, NHCH3),
4.24 (d, J ¼ 5.4 Hz, 2H, CH2O), 7.35 (br q, 1H, NHCH3, D2O
exchangeable), 7.68 (d, J ¼ 8.4 Hz, 2H, ArH), 8.06 (br s, 2H, ArH), 8.40
(vbr s, 2H, 6-NH2, D2O exchangeable), 10.32 (s, 1H, NH, D2O
C6H11), 1.67 (m, 6H, C6H11), 2.41 (s, 3H, PhCH3), 4.04 (d, J ¼ 6 Hz, 2H,
CH2O), 7.24 (br s, 2H, NH2, D2O exchangeable), 7.44 (d, J ¼ 7.8 Hz,
2H, ArH), 7.64 (br s, 2H, NH2, D2O exchangeable), 7.83 (d, J ¼ 8.1 Hz,
2H, ArH); 13C NMR (75 MHz, DMSO-d6)
d
21.0, 25.1, 25.8, 28.8, 36.5,
exchangeable); 13C NMR (75 MHz, DMSO-d6)
d 25.1, 25.8, 28.5, 28.8,
71.9, 105.8, 128.6, 129.4, 134.9, 144.5, 159.1, 160.7, 164.0; MS (CI,
36.4, 72.7, 107.2, 111.8, 119.7, 127.3, 132.4, 142.7, 156.9, 158.7, 163.9;
MS (EI, 70 eV) m/z 460 (Mþ), 420, 324 (100%). Anal.
C19H24N8O4S$0.25H2O$0.25iPrOH (C, H, N).
70 eV) m/z 421 (Mþþ1). Anal. C18H24N6O4S (C, H, N).
5.1.7. 4-(6-Amino-4-cyclohexylmethoxypyrimidin-2-ylamino)-
amino-N-methylbenzenesulfonamide trifluoroacetate salt (12)
To a solution of 10 (2.640 g, 8 mmol) and 11 (1.500 g, 8 mmol) in
TFE (35 mL), TFA (3.10 mL, 40 mmol) was added. The resulting
mixture was stirred for 10 min at r.t., then refluxed for 24 h at 80 ꢀC.
The precipitated solid was collected on a filter and washed with
EtOAc to afford 12 as a white crystalline solid (1.500 g). Solvents
were removed in vacuo from the filtrate, water (30 mL) was added,
the pH was adjusted to neutral with saturated aqueous NaHCO3 and
5.1.10. 4-[6-Amino-5-(cyano-NNO-azoxy)-4-
cyclohexylmethoxypyrimidin-2-ylamino]-amino-N-(2-hydroxyethyl)-
benzenesulfonamide (16)
To a stirred suspension of 15 (0.150 g, 0.333 mmol) and cyana-
mide (0.042 g, 0.99 mmol, 3.0 equiv) in CH3CN (3.0 mL), IBA (0.021 g,
6.7 mmol, 2.0 equiv) was added portion-wise at r.t. After 2 h the
reaction mixturewas heated to 35 ꢀC for 30 min, then extracted with
CH2Cl2 (100 mL), washed with H2O (2 ꢃ 30 mL) and brine, dried

338
D. Boschi et al. / European Journal of Medicinal Chemistry 68 (2013) 333e338
[2] M. Malumbres, M. Barbacid, Mammalian cyclin-dependent kinases, Trends
Biochem. Sci. 30 (2005) 630e641.
[3] M. Malumbres, M. Barbacid, Cell cycle, CDKs and cancer: a changing paradigm,
Nat. Rev. Cancer 9 (2009) 153e166.
[4] S. Lapenna, A. Giordano, Cell cycle kinases as therapeutic targets for cancer,
Nat. Rev. Drug Discovery 8 (2009) 547e566.
[5] J. Cicenas, M. Valius, The CDK inhibitors in cancer research and therapy,
J. Cancer Res. Clin. Oncol. 137 (2011) 1409e1418.
(MgSO4) and evaporated to dryness. The resulting crude product
was purified by flash chromatography (CH2Cl2/MeOH 9.5/0.5), then
recrystallized from methanol to yield 16 as a yellow crystalline solid
(0.140 g; yield 86%). M.p. 209e211 ꢀC (MeOH); 1H NMR (300 MHz,
DMSO-d6) d 1.03e1.26 (m, 5H, C6H11), 1.65e1.78 (m, 6H, C6H11), 2.78
(q, J ¼ 12.0 Hz, 2H, NHCH2), 3.36 (m, 2H, HOCH2), 4.24 (d, J ¼ 6 Hz, 2H,
CH2O), 4.68 (t, J ¼ 5.4 Hz,1H, OH, D2O exchangeable), 7.49 (t,1H, NH,
D2O exchangeable), 7.69 (d, J ¼ 8.7 Hz, 2H, ArH), 8.04 (br d, J ¼ 7.5 Hz,
2H, ArH), 8.40 (vbr s, 2H, NH2, D2O exchangeable), 10.31 (s, 1H, NH,
[6] M. Guha, Blockbuster dreams for Pfizer’s CDK inhibitor, Nat. Biotechnol. 31
(2013) 187.
[7] C.E. Arris, F.T. Boyle, A.H. Calvert, N.J. Curtin, J.A. Endicott, E.F. Garman,
A.E. Gibson, B.T. Golding, S. Grant, R.J. Griffin, P. Jewsbury, L.N. Johnson,
A.M. Lawrie, D.R. Newell, M.E.M. Noble, E.A. Sausville, R. Schultz, W. Yu,
Identification of novel purine and pyrimidine cyclin-dependent kinase in-
hibitors with distinct molecular interactions and tumor cell growth inhibition
profiles, J. Med. Chem. 43 (2000) 2797e2804.
[8] I.R. Hardcastle, C.E. Arris, J. Bentley, F.T. Boyle, Y.H. Chen, N.J. Curtin,
J.A. Endicott, A.E. Gibson, B.T. Golding, R.J. Griffin, P. Jewsbury, J. Menyerol,
V. Mesguiche, D.R. Newell, M.E.M. Noble, D.J. Pratt, L.Z. Wang, H.J. Whitfield,
N-2-Substituted O-6-cyclohexylmethylguanine derivatives: potent inhibitors
of cyclin-dependent kinases 1 and 2, J. Med. Chem. 47 (2004) 3710e3722.
[9] F. Marchetti, K.L. Sayle, J. Bentley, W. Clegg, N.J. Curtin, J.A. Endicott,
B.T. Golding, R.J. Griffin, K. Haggerty, R.W. Harrington, V. Mesguiche,
D.R. Newell, M.E.M. Noble, R.J. Parsons, D.J. Pratt, L.Z. Wang, I.R. Hardcastle,
Structure-based design of 2-arylamino-4-cyclohexylmethoxy-5-nitroso-6-
aminopyrimidine inhibitors of cyclin-dependent kinase 2, Org. Biomol.
Chem. 5 (2007) 1577e1585.
[10] F. Marchetti, C. Cano, N.J. Curtin, B.T. Golding, R.J. Griffin, K. Haggerty,
D.R. Newell, R.J. Parsons, S.L. Payne, L.Z. Wang, I.R. Hardcastle, Synthesis and
biological evaluation of 5-substituted O4-alkylpyrimidines as CDK2 inhibitors,
Org. Biomol. Chem. 8 (2010) 2397e2407.
[11] A. Gasco, A. Serafino, V. Mortarini, E. Menziani, M.A. Bianco, J. Ceruti Scurti, An
antibacterial and antifungal compound from Calvatia lilacina, Tetrahedron
Lett. 15 (1974) 3431e3432.
[12] H. Umezawa, T. Takeuchi, H. Linuma, M. Ito, M. Ishizuka, Y. Kurakata,
Y. Umeda, Y. Nakanishi, T. Nakamura, A. Obayashi, O. Tanabe, A new antibiotic,
calvatic acid, J. Antibiot. 28 (1975) 87e90.
[13] T. Okuda, N. Nakayama, A. Fujiwara, Calvatic acid production by Lycoperda-
ceae. I. Calvatic acid related compounds produced by Lycoperdon pyriforme,
Trans. Mycol. Soc. Jpn. 23 (1982) 225e234.
[14] R. Fruttero, G. Mulatero, R. Calvino, A. Gasco, A directed synthesis of alkyl, aryl,
and heteroaryl-ONN-azoxycyanides, J. Chem. Soc., Chem. Commun. 5 (1984)
323e324.
[15] M. Sako, S. Oda, K. Hirota, G.P. Beardsley, Convenient and versatile synthetic
methods for furazano[3,4-d]pyrimidines, Synthesis 11 (1997) 1255e1257.
[16] D.L. Lipilin, A.M. Churakov, S.L. Ioffe, Y.A. Strelenko, V.A. Tartakovsky, Nucle-
ophilic aromatic substitution of hydrogen in the reaction of tert-alkylamines
with nitrosobenzenes e synthesis and NMR study of N-(tert-alkyl)-ortho-
nitrosoanilines, Eur. J. Org. Chem. 1 (1999) 29e35.
D2O exchangeable); 13C NMR (75 MHz, DMSO-d6)
d 25.6, 26.4, 29.4,
37.0, 45.5, 60.3, 73.3, 107.5, 112.0, 120.2, 127.7, 134.0, 142.9, 157.2,
158.9, 164.2; MS (EI, 70 eV) m/z 490 (Mþ), 450, 354 (100%). Anal.
C20H26N8O5S$0.5H2O (C, H, N).
5.2. Molecular modelling
The molecular models of the newly synthesized analogues were
generated using standard bond lengths and angles with MOE [22].
Energy minimization was carried out with the MMFF94s force field
(MMFF94 charges, Generalized Born implicit solvent model) using a
truncated NewtoneRaphson geometry optimization method, until
the gradient was below 0.001 kcal molꢂ1. The conformational space
of the models was explored by means of a stochastic conforma-
tional search, as implemented in MOE. Further optimization of the
global minimum geometry was achieved by ab initio quantum-
mechanical (QM) calculations at the UHF/6-31G(d) level of theory
using GAMESS-US [23]. Electrostatic charges were fitted on the QM
electrostatic potential through the RESP method [24] as imple-
mented in AMBER 10 [25]. Crystal structures of human CDK2
complexes with different ligands were retrieved from the Protein
Data Bank. Sections with missing residues (namely, residues 36e43
in 1E1X and residues 38e43, 297, 298 in 2C6O) were not modelled,
since they were far from the active site; instead, the C and N-ter-
minals at the ends of the missing section were capped with N-
methylamide and acetyl groups, respectively. Hydrogen atoms
were added to the proteineligand complex in standard positions,
and then minimized through the SANDER module of the AMBER 10
suite, while keeping heavy atoms restrained to their original posi-
[17] M. Pennati, A.J. Campbell, M. Curto, M. Binda, Y. Cheng, L.-Z. Wang, N.J. Curtin,
B.T. Golding, R.J. Griffin, I.R. Hardcastle, A. Henderson, N. Zaffaroni,
D.R. Newell, Potentiation of paclitaxel-induced apoptosis by the novel cyclin-
dependent kinase inhibitor NU6140: a possible role for survivin down-regu-
lation, Mol. Cancer Ther. 4 (2005) 1328e1337.
ꢀꢂ2
tions by a harmonic potential (103 kcal molꢂ1
A ). All molecular
mechanics calculations on the proteineligand complexes were
carried out using FF99SB and GAFF parameters for protein and
ligand, respectively. After removing the co-crystallized ligand,
docking simulations were performed using AutoDock 4.2 [26]. A
[18] O. Tetsu, F. McCormick, Proliferation of cancer cells despite CDK2 inhibition,
Cancer Cell 3 (2003) 233e245.
[19] S. Ortega, I. Presto, J. Odajima, A. Martín, P. Dubus, R. Sotillo, J.L. Barbero,
M. Malumbres, M. Barbacid, Cyclin-dependent kinase 2 is essential for meiotic
but not for mitotic cell division in mice, Nat. Genet. 35 (2003) 25e31.
[20] C.M. Richardson, D.S. Williamson, M.J. Parratt, J. Borgognoni, A.D. Cansfield,
P. Dokurno, G.L. Francis, R. Howes, J.D. Moore, J.B. Murray, A. Robertson,
A.E. Surgenor, C.J. Torrance, Triazolo[1,5-a]pyrimidines as novel CDK2 in-
hibitors: protein structure-guided design and SAR, Bioorg. Med. Chem. Lett. 16
(2006) 1353e1357.
[21] I.K. Khanna, R.M. Weier, Y. Yu, P.W. Collins, J.M. Miyashiro, C.M. Koboldt,
A.W. Veenhuizen, J.L. Currie, K. Seibert, P.C. Isakson, 1,2-Diarylpyrroles as potent
andselectiveinhibitorsofcyclooxygenase-2,J.Med.Chem. 40(1997)1619e1633.
[22] MOE Version 2010.11, Chemical Computing Group Inc., Montreal, Quebec,
Canada, 2011.
ꢀ
grid with 0.375 A step size encompassing the whole active site was
generated with AutoGrid, then 100 flexible docking runs were
carried out with AutoDock, using the Lamarckian genetic algorithm
with default parameters.
Acknowledgements
The Chemical Computing Group is acknowledged for financial
support to computational work.
[23] M.W. Schmidt, K.K. Baldridge, J.A. Boatz, S.T. Elbert, M.S. Gordon, J.H. Jensen,
S. Koseki, N. Matsunaga, K.A. Nguyen, S.J. Su, T.L. Windus, M. Dupuis,
J.A. Montgomery, General atomic and molecular electronic structure system,
J. Comput. Chem. 14 (1993) 1347e1363.
[24] J. Wang, P. Cieplak, P.A. Kollman, How well does a restrained electrostatic
potential (RESP) model perform in calculating conformational energies of
organic and biological molecules? J. Comput. Chem. 21 (2000) 1049e1074.
[25] D.A. Case, T.A. Darden, T.E. Cheatham III, C.L. Simmerling, J. Wang, R.E. Duke,
R. Luo, M. Crowley, R.C. Walker, W. Zhang, K.M. Merz, B. Wang, S. Hayik,
A. Roitberg, G. Seabra, I. Kolossvary, K.F. Wong, F. Paesani, J. Vanicek, X. Wu,
S.R. Brozell, T. Steinbrecher, H. Gohlke, L. Yang, C. Tan, J. Mongan, V. Hornak,
G. Cui, D.H. Mathews, M.G. Seetin, C. Sagui, V. Babin, P.A. Kollman, AMBER 10,
University of California, San Francisco, USA, 2008.
Appendix A. Supplementary data
Supplementary data associated with this article can be found in
the online version, at http://dx.doi.org/10.1016/j.ejmech.2013.07.
031.
References
[1] V. Mesguiche, R.J. Parsons, C.E. Arris, J. Bentley, F.T. Boyle, N.J. Curtin,
T.G. Davies, J.A. Endicott, A.E. Gibson, B.T. Golding, R.J. Griffin, P. Jewsbury,
L.N. Johnson, D.R. Newell, M.E.M. Noble, L.Z. Wang, I.R. Hardcastle, 4-Alkoxy-
2,6-diaminopyrimidine derivatives: inhibitors of cyclin dependent kinases 1
and 2, Bioorg. Med. Chem. Lett. 13 (2003) 217e222.
[26] G.M. Morris, R. Huey, W. Lindstrom, M.F. Sanner, R.K. Belew, D.S. Goodsell,
A.J. Olson, Autodock4 and AutoDockTools4: automated docking with selective
receptor flexibility, J. Comput. Chem. 16 (2009) 2785e2791.