General Ser/Thr Kinases Pharmacophore Approach for Selective Kinase Inhibitors Search as Exemplified by Design of Potent and Selective Aurora A Inhibitors

Article abstract of DOI:10.1111/cbdd.12733

A general pharmachophore model for various types of Ser/Thr kinases was developed. Search for the molecules fitting to this pharmacophore among ASINEX proprietary library revealed a number of compounds, which were tested and appeared to possess some activity against several Ser/Thr kinases such as Aurora A, Aurora B and Haspin. The possibility of performing the fine-tuning of the general Ser/Thr pharmacophore to desired types of kinase to get active and selective inhibitors was exemplified by Aurora A kinase. As a result, several hits in 3–5?nm range of activity against Aurora A kinase with rather good selectivity and ADME properties were obtained.

Full text of DOI:10.1111/cbdd.12733

Chem Biol Drug Des 2016; 88: 54–65
Research Letter
General Ser/Thr Kinases Pharmacophore Approach for
Selective Kinase Inhibitors Search as Exemplified by
Design of Potent and Selective Aurora A Inhibitors
devoted to selected kinase of serine–threonine family, but
we surmised that there are a number of common features
of all serine–threonine inhibitors so it would be possible to
Natalya I. Vasilevich*, Elena A. Aksenova,
Denis N. Kazyulkin and Ilya I. Afanasyev
construct
a
general pharmacophore model which
Novie Nauchnie Tekhnologii Ltd. (ASINEX Company
Group), 20 Geroev Panfilovtsev Str., Moscow 125480,
Russia
*Corresponding author: Natalya I. Vasilevich,
nvasilevich@asinex.com
subsequently could be optimized for specific kinds of
serine–threonine kinase.
To achieve our goal, we analysed a number of known inhi-
bitors of various serine–threonine kinases including com-
pounds, which have been found both in our laboratory
previously and in literature references cited above. For this
purpose, we used pharmacophore elucidating functionality
of MOE version 2010.10. The results of this work are on
Figure 1.
A general pharmachophore model for various types of
Ser/Thr kinases was developed. Search for the mole-
cules fitting to this pharmacophore among ASINEX
proprietary library revealed a number of compounds,
which were tested and appeared to possess some
activity against several Ser/Thr kinases such as Aurora
A, Aurora B and Haspin. The possibility of performing
the fine-tuning of the general Ser/Thr pharmacophore
to desired types of kinase to get active and selective
inhibitors was exemplified by Aurora A kinase. As a
result, several hits in 3–5 nM range of activity against
Aurora A kinase with rather good selectivity and ADME
properties were obtained.
As it is seen from the figure, the general pharmacophore
looks like a kind of diamond with two opposite hydropho-
bic centres, one aromatic centre and a couple of H-bond
donor and acceptor projections in one corner. The length
of a rhomb side is about 4–5 A.
Application of the general pharmacophore found to ASINEX
proprietary library allowed to identify a scaffold including
two 5- and 6-member heterocyclic aromatic rings linked
through NH group, containing additional pyrrolidine or
piperidine group attached to 6-member ring in meta-
position to NH group. To confirm their activity against Ser/
Thr kinases, a set of compounds belonging to this scaffold
was tested against Aurora A, Aurora B and Haspin kinases.
Key words: kinase, molecular modeling, pharmacophore
modeling, phosphatase, structure-based drug design
Received 8 October 2015, revised 29 December 2015 and
accepted for publication 7 January 2016
Serine–threonine kinases play an important role in signal
transduction pathways in both eukaryotic and prokaryotic
cells; they act in regulation of cell proliferation,
programmed cell death (apoptosis), cell differentiation and
embryonic development. Therefore, inhibitors of Ser/Thr
kinase activity have a broad range of potential therapeutic
uses – from treating cancer to promoting a desired
immunosuppressive effect. As these kinases were found in
a number of mycobacterial organisms, their inhibitors can
be attractive targets to treat bacterial infection (1).
A number of attempts have been done to construct a
pharmacophore model for various serine–threonine kinase
inhibitors. For example, pharmacophore for STPK inhibi-
tors of tuberculosis (2,3), for mTor kinase inhibitors (4), for
Aurora B inhibitors (5), for Aurora A inhibitors (6,7) and for
B-Raf inhibitors (8) was developed. All these works were
Figure 1: General pharmacophore of serine–threonine kinase
inhibitors (MOE 2010.10).
54
ª 2016 John Wiley & Sons A/S. doi: 10.1111/cbdd.12733

General Kinases Pharmacophore Approach
Chem Biol Drug Des 2016; 88: 54–65
55

Vasilevich et al.
56
Chem Biol Drug Des 2016; 88: 54–65

General Kinases Pharmacophore Approach
Table 2: Optimization of compound 2 for Aurora A kinase
IC50, nM (Mean ꢀ SEM,
4 repeats)
2
96.0 ꢀ 9.1
O
N
S
N
N
N
F
S
N
H
N
Cl
11
85.0 ꢀ 9.2
Cl
N
N
F
N
N
H
O
12
13
14
12.0 ꢀ 1.1
193.6 ꢀ 20.5
246.4 ꢀ 21.2
N
S
NH
O
F
Cl
N
N
N
O
F
Cl
N
N
N
N
S
N
N
H
HO
O
O
O
F
F
F
Cl
Cl
Cl
S
S
S
N
N
H
N
H
O
O
O
15
418.3 ꢀ 40.5
N
N
N
N
N
N
H
N
O
16
154.1 ꢀ 11.2
N
N
H
Chem Biol Drug Des 2016; 88: 54–65
57

Vasilevich et al.
Table 2: continued
IC50, nM (Mean ꢀ SEM,
4 repeats)
17
70.4 ꢀ 6.5
Cl
F
N
S
O
N
S
N
O
N
N
N
H
18
19
20
20.5 ꢀ 1.9
Cl
F
F
N
N
S
N
H
N
O
6.0 ꢀ 0.3
Cl
N
N
N
N
H
265.0 ꢀ 22.4
H
N
N
S
N
O
N
21
292.2 ꢀ 24.3
N
N
N
S
N
N
H
22
166.8 ꢀ 15.6
H
N
N
S
N
N
O
23
24
60.0 ꢀ 5.4
N
N
N
O
N
S
N
N
H
22.3 ꢀ 0.15
Cl
N
S
N
N
H
N
58
Chem Biol Drug Des 2016; 88: 54–65

General Kinases Pharmacophore Approach
Table 2: continued
IC50, nM (Mean ꢀ SEM,
4 repeats)
25
95.2 ꢀ 9.1
O
F
N
S
N
N
H
N
26
96.1 ꢀ 8.9
Cl
N
HN
N
N
N
N
S
S
27
35.1 ꢀ 0.36
103.2 ꢀ 10.1
4.1 ꢀ 0.35
NH
F
O
Cl
N
N
N
28
N
S
NH
F
O
Cl
N
N
N
29
30
31
F
O
Cl
N
N
N
Cl
S
N
H
N
5.3 ꢀ 0.46
O
F
Cl
N
N
N
N
H
N
N
H
17.5 ꢀ 2.1
O
F
H
Cl
N
N
N
N
N
H
Chem Biol Drug Des 2016; 88: 54–65
59

Vasilevich et al.
Table 2: continued
IC50, nM (Mean ꢀ SEM,
4 repeats)
32
3.5 ꢀ 0.29
O
F
H
N
Cl
N
N
N
N
H
N
OH
Cl
N
NH
O
O
N
n
NH
2 c
n
n
e
NH2
n
OH
a
N
b
d
n
N
( )m
65-75%
75-80%
( )m
55-65%
45-60%
( )m
80-90%
( )m
( )m
N
N
N
N
boc
N
boc
2'
boc
boc
boc
4'
5'
3'
1'
R1
HN
R1
HN
R1
HN
N
N
N
n
f
g
N
n
N
n
N
95-98%
85-90%
( )m
( )m
N
( )m
N
O
N
boc
6'
H
7'
8'
F
Cl
S
S
S
S
Cl
*
*
*
*
*
*
R1=
N
N N
H
N
N
N
N N
H
n=0, 1; m=1, 2
Scheme 1: Synthesis of compounds 11, 12, 17, 18, 27, 28, 29, 30.
Conditions: (a) (NH4)2CO3, TBTU, Et3N, CH3CN, RT, 8 h; (b) triethyloxonium tetrafluoroborate, CH2Cl2, RT, 2 h; NH3/MeOH, RT, 16 h; (c) CH3COCH2-
COOEt, NaOtBu, EtOH, RF, 10 h; (d) POCl3, N,N-dimethylaniline, toluene, RF, 4 h; (e) R1NH2, Na2CO3, Pd2dba3, Xantphos, toluene/H2O, MW,
140 °C, 2 h; (f) HCl(aq)/MeOH, RT, 4 h; and (g) 3-chloro-2-fluoro-benzoic acid, TBTU, Et3N, CH3CN, RT, 8 h.
Noteworthy, none of these compounds were identified pre-
viously in scientific literature as Ser/Thr kinases inhibitors.
A kinase and tried to optimize the scaffold that we identified
in the screen to be more selective for this target.
Indeed, some of these compounds revealed activity against
all three Ser/Thr kinases, though not very high (Table 1).
Among the primary set of ten compounds depicted in
Table 1, we found compound 2, which possessed the
best inhibitory activity towards Aurora A kinase – about
95 nM and attempted to optimize this compound for Aur-
ora A kinase.
To confirm the possibility of fine-tuning of the general phar-
macophore for selected type of kinases, we chose Aurora
60
Chem Biol Drug Des 2016; 88: 54–65

General Kinases Pharmacophore Approach
boc
R2
O
boc
N
N
a
( )n
b
c
( )n
( )n
( )n
78%
65-70%
N
Br
N
N
75%
HN
R1
N
boc
boc
11'
12'
O
N
9'
10'
H
N
( )n
( )n
d
e
HN
R1
N
HN
R1
N
98%
75-85%
13'
f
14'a
R3
65-75%
N
( )n
HN
R1
N
14'b
n = 2
S
*
*
R1=
N
H
N
N
F
F
Cl
Cl
*
*
*
*
*
R2=
R3=
Cl
*
*
N
N
Scheme 2: Synthesis of compounds 19–22, 24–26, 31.
Conditions: (a) MePh3P+Iꢁ, NaH, THF; (b) 2,6-dibromo-pyridine, 9-BBN, Pd(Ph3P)4, K2CO3, THF, RF; (c) R1NH2, Na2CO3, Pd2dba3, Xantphos, toluene/H2O, MW, 140 °C,
2 h; (d) HCl(aq)/MeOH, RT, 4 h; (e) R2COOH, TBTU, Et3N, CH3CN, RT, 8 h; and (f) R3CHO, NaBH(OAc)3, CH3CN, RT, 8 h.
With this goal, a library of analogues of compound 2 was
prepared (Table 2).
water mixture with 2.5% mol Pd2dba3 и 5% mol xantphos
and potassium carbonate using microwave reactor (yield
65–75%) .
Compounds 11, 12, 17, 18, 27–30 were synthesized in
accordance with Scheme 1. Amides 20 were prepared from
corresponding Boc-protected amino acids 10 were con-
verted into amidine salts 30 by reaction with triethyloxoniun
tetrafluoroborate followed by treatment with solution of
ammonia in methanol. Cyclization of the amidine obtained
with acetoacetic ester in ethanol under reflux gave 6-methyl-
pyrimidones 40 in 55–60% yield. These pyrimidones 40 were
treated with three equivalents of phosphorus oxychloride
and nine equivalents of dimethylaniline in toluene to give
chlorides 50 in 45–50% yield. Buchwald reaction with aro-
matic amines (2-aminothiazoles or 3-aminopyrazoles) led to
new derivatives 60; this reaction was performed in toluene–
For compound 29 with 5-chloro-2-thiazole moiety interme-
diate 60 (where R1 – unsubstituted 2-aminothiazole) was
treated with N-chlorosuccinimide in dichloroethane at
room temperature (yield 70%).
To synthesize compounds 30 and 32, containing 3-ami-
nopyrazole moiety, corresponding 3-amino-1H-pyrazoles
(or 3-amino-5-methyl-1H-pyrazoles) were protected by
tosyl group via reaction with tosyl chloride and sodium
hydrocarbonate in acetonitrile. This protective group was
easily removed from intermediates 60 by sodium hydrox-
ide in methanol treatment.
Chem Biol Drug Des 2016; 88: 54–65
61

Vasilevich et al.
a
N
N
c
b
N
N
R1
R1
45-50%
boc
Cl
N
Cl
60-63%
n
N
H
N
N
H
N
Cl
17'
O
15'
16'
N
d
N
R1
N
n
N
H
N
N
85%
e
NH
R1
N
R2
19'a
n
N
H
18'
70%
N
R1
n
N
N
H
R3
19'b
n = 1, 2
S
*
*
R1=
N N
N
H
O
F
Cl
R2=
R3=
*
*
N
Scheme 3: Synthesis of compounds 2, 23, 32.
Conditions: (a) R1NH2, Na2CO3, Pd2dba3, Xantphos, toluene/H2O, MW, 140 °C, 2 h; (b) alkene (100), 9-BBN, Pd(Ph3P)4, K2CO3, THF, RT; (c) HCl
(aq)/MeOH, RT, 4 h; (d) R2COOH, TBTU, Et3N, CH3CN, RT, 8 h; (e) R3CHO, NaBH(OAc)3, CH3CN, RT, 8 h.
After Boc-deprotection amines 70 were acylated by 3-
chloro-2-fluorobenzoic acid using TBTU as a coupling
reagent (yields 85% and more).
This compound was involved into Suzuki reaction with
alkenes 100 and 9-BBN followed by Boc-deprotection led
to compound 180 which was further acylated or alkylated
to give the target compounds 2, 23, 32.
Synthesis of compounds 19–22, 24–26, 31 is shown in
Scheme 2. Alkene 100 was obtained from ketone 90 via
reaction with methyltriphenylphosphonium iodide and
sodium hydride in 75% yield. Subsequent treatment with
9-BBN in THF, and 2,6-dibromopyridine in Suzuki reaction
conditions gave bromide 110 in 78% yield. Buchwald reac-
tion led to intermediates 120 (65–75% yields). Compounds
19, 20, 22, 24, 23 and 31 were prepared by acylation
reactions using TBTU as a coupling reagent. Compounds
21 and 26 were obtained by reductive amination reaction
in 65–75% yield.
Compounds 13–16 were prepared in accordance with
Scheme 4. Intermediate 210 was obtained by reaction
between compound 200 and 2,6-dibromopyridine in pres-
ence of NaHMDS in THF at 0 °C in 41% yield. Buchwald
reaction with 2-amino-5-methylthiazole gave compound
220 (yield 64%), which was easily converted into alcohol
240 (77%) or amides 270 and 290 (78–84%). Following Boc-
deprotection and acylation led to compounds 13–16 in
74–83% yields.
On the first step, we kept methyl substituent on thiazole ring
and 2-fluoro-3-chloro-benzoyl substituent on aliphatic nitro-
gen (compounds 11–19) varying both 6-member aromatic ring
and an aliphatic N-containing cycle tethered to the aromatic
one either directly or through methylene linker. We found that
the best compound in this series was compound 19.
Compounds 2, 23, 32 were synthesized as shown in
Scheme 3. Buchwald reaction with 2,4-dichloropyrimi-
dines 150 in conditions described above gave a mixture
of regioisomers that were separated by column chro-
matography in 45–50% yield of target compound 160.
62
Chem Biol Drug Des 2016; 88: 54–65

General Kinases Pharmacophore Approach
O
O
F
boc
boc
c
O
N
d
Cl
S
N
b
S
NH
a
S
N
Br
N
N
N
H
N
98%
N
N
H
N
N
64%
82%
N
H
N
41%
N
O
O
O
O
O
O
e
O
O
boc
23'
22'
21'
77%
13
f
20'
96%
boc
S
N
boc
S
N
N
N
N
H
N
N
H
N
O
24'
OH
h
c
98%
OH
26'
84%
g
78%
boc
S
N
S
NH
boc
S
N
N
N
N
H
N
N
H
N
N
N
N
O
N
H
29'
OH
25'
d
O
NH
27'
c
98%
c
74%
97%
S
NH
O
F
S
NH
N
N
H
N
Cl
S
N
N
N
H
N
O
N
30'
N
N
N
H
O
NH
28'
OH
d
83%
d
81%
O
F
14
O
F
Cl
S
N
Cl
S
N
N
N
H
N
O
N
N
H
N
N
O
NH
15
16
Scheme 4: Synthesis of compounds 13–16.
Conditions: (a) 2,6-dibromo-pyridine, NaHMDS, THF, 0 °C; (b) 2-amino-5-methylthiazole, Na2CO3, Pd2dba3, Xantphos, toluene/H2O,
MW, 140 °C, 2 h; (c) HCl(aq)/MeOH, RT, 4 h; (d) 3-chloro-2-fluoro-benzoic acid, TBTU, Et3N, CH3CN, RT, 8 h; (e) LiAlH4, THF, RT,
3 h; (f) NaOH, MeOH/H2O, RT, 4 h; (g) methyl amine hydrochloride, TBTU, Et3N, CH3CN, RT, 8 h; and (h) dimethyl amine hydrochlo-
ride, TBTU, Et3N, CH3CN, RT, 8 h.
Then, a series of compounds with methylene-4-piperidine
linker and various N-acyl and N-alkyl substitution were
made (compounds 20–26). None of these compounds
was superior to compound 19, so the initial 2-fluoro-3-
chloro-benzoyl group appeared to be a substituent of
choice.
One can see that compound 29, containing thiazole
group, possesses better activity against Aurora A kinase
than parent compound 2 (4.1 nM vs. 100 nM); however, its
selectivity in comparison with other kinases was approxi-
mately equal. Replacement of thiazole group for pyrazole
moiety gave compounds 30 and 32, which appeared to
be equally active but more selective that compound 29
(EPHA2: inhibition 70%, 50% and 29%, LYN: 84%, 40%
and 10%, RON: 98%, 10% and 37% for compounds 29,
30 and 32, correspondingly). Moreover, such a replace-
ment improved a solubility of compounds 30 and 32 to
>100 lg/mL.
On the last step, we varied thiazole fragment and discov-
ered pyrazole-containing derivatives to be equally or even
more active than 4-methyl-thiazole analogues.
Therefore, as a result of our attempts to optimize the
general serine–threonine pharmacophore model for
particular Aurora A kinase, a number of compounds with
excellent potency were obtained. The best three
compounds were tested for selectivity against serine/
threonine kinases available in house and some of ADME
properties. For comparison, one non-receptor tyrosine
kinase, RON, was also included. The results are shown
in Table 3.
In conclusion, a general pharmacophore model applicable
to various serine–threonine kinases was designed. A pos-
sibility of fine-tuning of this pharmacophore to particular
types of serine–threonine kinases such as Aurora A kinase
was confirmed. As
a result, compounds with good
potency, selectivity and some ADME parameters were
obtained.
Chem Biol Drug Des 2016; 88: 54–65
63

Vasilevich et al.
64
Chem Biol Drug Des 2016; 88: 54–65

General Kinases Pharmacophore Approach
4. Tanneeru K., Guruprasad L. (2012) Ligand-based
3-D pharmacophore generation and molecular
Acknowledgments
docking of mTOR kinase inhibitors.
Model;18:1611–1624.
5. Wang H.-Y., Li L.-L., Cao Z.-X., Luo S.D., Wei
Y.Q., Yang S.Y. (2009) A specific pharmacophore
J
Mol
The authors gratefully acknowledge support from the
Ministry of Education and Science of the Russian Federa-
tion for funding (agreement 14.576.21.0019 dated July,
27, 2014).
model of Aurora
B kinase inhibitors and virtual
screening studies based on it. Chem Biol Drug
Des;73:115–126.
References
6. Chavan S.R., Dash R.C., Alam M.S., Hirwani R.R.
(2014) Identification of new novel scaffold for Aurora A
inhibition by pharmacophore modeling and virtual
screening. Mol Divers;18:853–863.
ꢀ
ꢀ
1. Magnet S., Hartkoorn R.C., Szekely R. Pato J., Triccas
´
J.A., Schneider P., Szantai-Kis C., Orfi L., Chambon M.,
Banfi D., Bueno M., Turcatti G., Keri G., Cole S.T
ꢀ
7. Deng X.Q., Wang H.Y., Zhao Y.L., Xiang M.L.,
Jiang P.D., Cao Z.X., Zheng Y.Z., Luo S.D., Yu
L.T., Wei Y.Q., Yang S.Y. (2008) Pharmacophore
modelling and virtual screening for identification of
new Aurora-A kinase inhibitors. Chem Biol Drug
Des;71:533–539.
8. Xie H., Chen L., ,Zhang J., Xie X., Qiu K., Fu J. (2015)
A combined pharmacophore modeling, 3D QSAR and
virtual screening studies on imidazopyridines as B-Raf
inhibitors. Int J Mol Sci;16:12307–12323.
(2010) Leads for antitubercular compounds from kinase
inhibitor library screens. Tuberculosis;90:354e360.
2. Seal A., Yogeeswari P., Sriram D., Consortium O., Wild
D.J. (2013) Enhanced ranking of PknB inhibitors using
data fusion methods. Journal of Cheminformatics;5:2.
3. Danilenko V.N., Osolodkin D.I., Lakatosh S.A., Preo-
brazhenskaya M.N., Shtil A.A. et al. (2011) Bacterial
eukaryotic type serine–threonine protein kinases: from
structural biology to targeted anti-infective drug design.
Curr Top Med Chem;11:1352–1369.
Chem Biol Drug Des 2016; 88: 54–65
65