A Dual Non-ATP Analogue Inhibitor of Aurora Kinases A and B, Derived from Resorcinol with a Mixed Mode of Inhibition

Article abstract of DOI:10.1111/cbdd.12728

Aurora kinases are the most commonly targeted mitotic kinases in the intervention of cancer progression. Here, we report a resorcinol derivative, 5-methyl-4-(2-thiazolylazo) resorcinol (PTK66), a dual inhibitor of Aurora A and Aurora B kinases. PTK66 is a surface binding non-ATP analogue inhibitor that shows a mixed pattern of inhibition against both of Aurora A and B kinases. The in vitro IC₅₀ is approximately 47 and 40 μm for Aurora A and Aurora B kinases, respectively. In cellular systems, PTK66 exhibits a substantially low cytotoxicity at micromolar concentrations but it can induce aneuploidy under similar dosages as a consequence of Aurora kinase inhibition. This result was corroborated by a drop in the histone H3 (S10) phosphorylation level detected via Western blot analysis using three different cell types. Altogether, our findings indicate that the ligand containing resorcinol backbone is one of the novel scaffolds targeting the Aurora family of kinases, which could be a target for antineoplastic drug development.

Full text of DOI:10.1111/cbdd.12728

Chem Biol Drug Des 2016; 87: 958–967
Research Article
A Dual Non-ATP Analogue Inhibitor of Aurora Kinases
A and B, Derived from Resorcinol with a Mixed Mode
of Inhibition
1
1
Dhanasekaran Karthigeyan , Sudhevan Surabhi ,
function leads to spindle pole defects, chromosomal
mis-segregation, aneuploidy and apoptosis. In many can-
cers, Aurora kinases are reported to be overexpressed
and actively involved in the events of tumour onset and
progression (1). Recent investigations demonstrated the
efficacy of dual or pan Aurora kinase inhibitors in the
development of anticancer therapeutics (2). Many such
small molecules inhibitors under various stages of clinical
trials have shown little success (3). Thus, we were moti-
vated to search for novel scaffolds targeting the Aurora
kinases.
1
2
Pushpak Mizar , Siddhanta Soumik ,
Amrita Banerjee , Sarmistha Halder Sinha ,
Dipak Dasgupta , Chandrabhas Narayana and
3
1
3
2
1,
Tapas K. Kundu *
1
Transcription and Disease Laboratory, Molecular Biology
and Genetics Unit, Jawaharlal Nehru Centre for Advanced
Scientific Research, Jakkur, Bangalore 560064, India
Light Scattering Laboratory, Chemistry and Physics of
Materials Unit, Jawaharlal Nehru Centre for Advanced
Scientific Research, Jakkur, Bangalore 560064, India
Biophysics & Strucural Genomics Division, Saha Institute
2
3
Based on the previously reported small molecules target-
ing the Aurora family kinases and research performed by
our group on the requirements of the chemical scaffold
that could inhibit the catalytic activity of Aurora kinases,
the following criteria need to be fulfilled. First, the small
molecule should possess a heterocyclic scaffold. Second,
the ligand should be able to make H-bonds with the hinge
region residues that are conserved among the Aurora-inhi-
bitor complex cocrystal structures and are reported to be
significant for ligand binding. In addition, the scaffold
should make hydrophobic interactions with the surround-
ing residues by the core scaffold itself or by the incorpora-
tion of hydrophobic linkages (4). The 5-Methyl-4-(2-
thiazolylazo) resorcinol (PTK66) is one of the resorcinol
scaffold containing ligands that contains a heterocyclic ring
capable of hydrogen bonding with the hinge region resi-
dues, while the thiazolylazo side chain can possibly inter-
act with the hydrophobic residues around the hinge region
of Aurora kinases.
of Nuclear Physics, Block-AF, SectorÀ1, Bidar Nagar,
Kolkata 700 064, West Bengal, India
*Corresponding author: Tapas K. Kundu,
tapas@jncasr.ac.in
Aurora kinases are the most commonly targeted mito-
tic kinases in the intervention of cancer progression.
Here, we report a resorcinol derivative, 5-methyl-4-
(
2-thiazolylazo) resorcinol (PTK66), a dual inhibitor of
Aurora A and Aurora B kinases. PTK66 is a surface
binding non-ATP analogue inhibitor that shows
a
mixed pattern of inhibition against both of Aurora A
and B kinases. The in vitro IC50 is approximately 47
and 40 lM for Aurora A and Aurora B kinases, respec-
tively. In cellular systems, PTK66 exhibits a substan-
tially low cytotoxicity at micromolar concentrations but
it can induce aneuploidy under similar dosages as a
consequence of Aurora kinase inhibition. This result
was corroborated by a drop in the histone H3 (S10)
phosphorylation level detected via Western blot analy-
sis using three different cell types. Altogether, our find-
ings indicate that the ligand containing resorcinol
backbone is one of the novel scaffolds targeting the
Aurora family of kinases, which could be a target for
antineoplastic drug development.
In vitro screening of PTK66 against the Aurora family of
kinases revealed a dose-dependent inhibition of kinase
activity of Aurora A kinase and Aurora B kinase with an
IC50 of ~ 40 lM. The kinetic analysis of this inhibition
showed a mixed pattern of inhibition against both of these
kinases. The hinge binding of PTK66 was successfully ver-
ified via surface-enhanced Raman spectroscopy (SERS) of
the enzyme–ligand complex. Furthermore, the binding was
corroborated using isothermal titration calorimetry (ITC).
The thermodynamic parameters obtained from the ITC
measurements show a binding stoichiometry of N = 2 and
N = 1 for Aurora A and B kinases, respectively. In silico
analysis of the ligand–enzyme complex, using the
Key words: Aurora kinase, non-ATP analogue inhibitor,
PTK66, surface-enhanced Raman spectroscopy
Received 21 July 2015, revised 13 November 2015 and
accepted for publication 30 December 2015
Aurora kinases play an important role in chromosomal
segregation during mitosis. The disruption of kinase
9
58
ª 2016 John Wiley & Sons A/S. doi: 10.1111/cbdd.12728

PTK66 Inhibits Aurora Kinase A and B
Autodock tool also revealed two potential binding sites for
Aurora A kinase and one favourable binding site for Aurora
B kinase as expected. Additionally, PTK66 was found to
induce aneuploidy in a dose and time-dependent manner
presumably, due to the dual inhibition of Aurora kinases
inside the cell.
USA) as a Raman excitation source. For SERS study, a
609
infinity-corrected objective (Nikon Plan Apo, Japan, NA
0.9) and ~ 5 mW laser power were used. The Ag sol was
prepared by modified Lee–Meisel method as described
elsewhere (5). SERS samples were prepared by keeping
the nanoparticle to analyte ratio at 5:1 using 40 nM of
protein.
Materials and Methods
PTK-66 synthesis
Docking studies
2
-amino thiazole (2 g) was dissolved in 50 mL of 6.0 mol/L
For the docking studies, crystal structure of Aurora A
kinase and Aurora B kinase was obtained from Protein
Data Bank with codes 1MQ4 (8) and 4AF3 (9), respec-
tively. Preparation of the protein structures prior to docking
is mentioned elsewhere (6). The structure of the inhibitor
PTK66 was optimised using the GAUSSIAN 03 program (10).
Docking analysis was performed with AUTODOCK 4.2 (11)
using Lamarckian genetic algorithm (LGA) and a grid spac-
hydrochloric acid solution at 0–5 °C. A solution of 0.53 g
of sodium nitrite in 20 mL water was added dropwise. The
mixture was stirred and kept at 0–5 °C for 45 min. In the
meantime, 0.96 g of 5-methyl resorcinol was dissolved in
1
5
0 mol/L sodium carbonate solution and cooled to 0–
°C. The solution was then added to the diazotised com-
pound under vigorous stirring. The reaction mixture was
allowed to stand overnight in a refrigerator at 0–5 °C. The
resultant precipitate was filtered and rinsed with cold
water. The reddish precipitate was further purified by
recrystallisation with ethanol and active charcoal. Please
refer to Scheme S1. Red Solid. The compound was char-
ꢀ
ing of 0.375 A. Docking was performed with flexible ligand
and 100 GA runs, a population size of 300, 25 9 105
evaluations and a maximum of 27 000 generations. Clus-
ter analysis was performed using MGL AUTODOCK tools ver
ꢀ
1.5.4 (12) using a cut-off of 2A. The figures were gener-
1
13
1
acterized by H and C NMR spectra (Figure S1).
DMSO-d6, 400 MHz) d: 2.49 (s, 3H), 6.23 (d, 1H, J= 2.4
Hz), 6.49 (d, 1H, J= 2.4 Hz), 7.73 (d, 1H, J= 3.2 Hz),
H
ated using the PyMOL molecular graphics system (13).
(
7. 93 (d, 1H, J = 3.2 Hz), 13.30 (s, 2H). 87% yield.
Isothermal Titration Calorimetry (ITC)
Calorimetric titrations were performed in a Microcal iTC
200 microcalorimeter. Freshly purified 5 lM protein (Aurora
A/B kinases) was titrated against 100 lM ligand (felodipine/
PTK66) in each case at 25 °C. Proteins and the ligands
were diluted in the final dialysate (20 mM Tris–HCl, pH 8.0,
0.2 mM EDTA and 100 mM KCl) against which the proteins
were extensively dialysed to avoid any artefact due to the
minor differences in buffer composition. DMSO content
was kept exactly same in the cell (not exceeding 0.2% v/v)
and syringe to rule out any interference due to DMSO con-
tent mismatch. All samples were thoroughly degassed
prior to titration and were prepared with the same batch of
dialysate to avoid any artefacts from the minor differences
in buffer composition. Dilutions of 100 lM felodipine/
PTK66 in dialysate containing equal percentage of DMSO
served as controls. The stirring speed was 300 rpm. The
resulting thermograms (background enthalpy subtracted)
were analysed using ‘one set of binding sites’ model of
Levenberg–Marquardt nonlinear least squares curve fitting
algorithm, inbuilt in the MICROCAL LLC software (Northamp-
ton, MA, USA).
1
3
C (DMSO-d6, 400 MHz) d: 175.39, 166.03, 157.32,
1
44.43, 143.04, 130.42, 120.73, 112.0, 101.02, 17.60.
Kinase assay
Bacterially expressed recombinant Xenopus histone H3
(1 lg) (5) was incubated along with either human Aurora A
kinase (40 ng) or Aurora B kinase (40 ng) (6) purified from
recombinant baculovirus-infected Sf9 cells in a 30 lL
reaction mixture containing 50 mM Tris–HCl, 100 mM NaCl,
0
.1 mM EGTA, 10 mM MgCl2, 0.2% 2-mercaptoethanol
and [c-32P] ATP (Specific Activity 3.5 Ci/mMol). The reac-
tion mixture along with varying concentrations of PTK66
was incubated at 30 °C for 5 min followed by addition of
3
2
[
c- P] ATP and further incubation for 10 min. The reaction
was stopped using 5XSDS-dye and resolved using 12%
PAGE followed by autoradiography. The intensity of the
signal was quantified using Alphaimager. For kinetic analy-
sis, increasing concentrations of [c-32P] ATP (5–30 lM)
were used in the presence of either DMSO or 10, 20, 40
and 80 lM of the inhibitor in the reaction mixture and the
kinetics was studied using Lineweaver–Burk plot.
Cell culture, treatment, FACS analysis and
Western blotting analysis
5
Surface-enhanced Raman spectroscopy
For the treatment, 1 9 10 cells of the indicated cell lines
Surface-enhanced Raman spectroscopy spectra were
recorded using a custom-made Raman system (7) with
were seeded in 30 mm dishes and maintained at 37 °C in
a 5% CO2 incubator using DMEM with 10% FBS. After
24-h post seeding, either DMSO (solvent control) or the
indicated concentration of the compound mentioned in the
180ꢀ backscattering geometry and 632.8 nm He-Ne laser
(model 30995, Research Electro Optics, Inc., Boulder,CO,
Chem Biol Drug Des 2016; 87: 958–967
959

Karthigeyan et al.
study was added and further incubated for the indicated
time and collected. For cell cycle/DNA content analysis,
cells were harvested and washed with 1XPBS and fixed in
control. The catalytic inhibition was analysed using a
Lineweaver–Burk plot shown in Figure 1d (panels I and II for
Aurora A kinase and Aurora B kinase, respectively). The
kinetic plot indicates that PTK66 is a mixed-type inhibitor of
both Aurora A and B kinases. Altogether, these results
demonstrate the dual inhibition of Aurora A and B kinases
mediated by the resorcinol scaffold containing ligand, PTK66.
7
0% ethanol. Subsequently, stained with (6 lM/mL) pro-
pidium iodide (Sigma, St Louis, MO, USA) with simultane-
ous RNAse treatment using (10 lg/mL) RNAse A (Sigma)
and analysed in BD FACS Calibur using CELL QUEST soft-
ware (BD Bioscience, San Jose, CA, USA). For Western
blotting, analysis cells were lysed in Laemmli buffer (60 mM
Tris–HCl pH 6.8, 2% SDS, 10% glycerol, 5% b-mercap-
toethanol, 0.01% bromophenol blue) and probed using the
indicated antibodies.
SERS of Aurora A kinase and Aurora B kinase and
their binding with PTK66
To verify the direct binding of ligand over the enzyme of inter-
est, SERS was employed to assess the PTK66-Aurora A
kinase (Figure 2A) and PTK66-Aurora B kinase (Figure 2B)
complex individually. The SERS spectrum of Aurora A kinase
and Aurora B kinase is dominated by bands from the aro-
matic amino acid side chains such as Trp, Tyr, Phe and His
(5). Other prominent bands include the amide bands which
are important markers for protein secondary structures and
bands corresponding to the aliphatic side chains. The amide
modes are indicative of the secondary structure of the pro-
teins, specifically the amide I mode which consists of CO
stretching vibrations with some contribution from N–H bend-
ing, out of phase CN stretching vibration and CCN deforma-
MTS assay
3
1
9 10 HEK293T cells were seeded in 96-well plates and
incubated in DMEM containing 5% FBS with 5% CO2 at
7 °C for 24 h and then treated with the indicated con-
3
centration of PTK66 or an equal volume of DMSO as sol-
vent control and maintained further for 24 h, after which
TM
2
0 lL of Cell Titre 96 Aqueous one reagent (Promega,
Madison, WI, USA) was added to each well and pro-
cessed as per manufacturer’s instructions. Considering
the untreated wells as 100% viability, a bar graph was
plotted using absorbance measure across the sample.
À1
tion. The 1620/1619 cm band in Aurora A kinase/Aurora B
kinase can be assigned to the amide I band (5). The modes
À1
at approximately 1033, 1189 and 1586 cm
can be
Apoptosis assay
assigned to C-H bending, a combination of in-plane C-H
bending with ring stretching and phenyl stretching modes,
HEK293T cells were treated with either DMSO or 10 lM
PTK66 or 5 lM garcinol for 24 h and then fixed in metha-
nol. Briefly, After fixation, the cells were stained using the
live/dead cell double staining reagents, namely Annexin
V-Cy3 and 6-carboxyfluorescein diacetate, provided in the
kit (Sigma, APOAC-1KT) by following the manufacturer’s
protocol.
respectively, for both Aurora A and B kinases. The region
À1
from 1300 to 1500 cm can be assigned for the d(CH
2
)
mode for phenylalanine. The SERS spectra also contain the
À
stretching vibration of C-COO which appears as a doublet
À1
at 910 and 931 cm . The asymmetric stretching vibration of
À
À1
COO at 1586 cm overlaps with the histidine, tryptophan
and phenylalanine bands. The distinct doublets arising at
À1
1
443/1471 cm
correspond to the CH2 deformation
Results
modes. The C-N stretching modes are observed at approxi-
À1
À1
mately 1030 cm . The mode at 1349 cm
assigned to a complex mode of C -H bending and C-C
a
can be
PTK66 is a dual inhibitor of Aurora kinases
a
In the search for novel scaffolds that can putatively target the
Aurora family kinases, we investigated the inhibition status of
stretching internal co-ordinates. We previously reported (6)
that the event of small molecule binding brings about subtle
structural rearrangements and alter the way a protein origi-
nally binds to the surface of the nanoparticles in their free
form (6). The changes in SERS spectra are a direct indication
of the binding of the small molecule to the protein. In the case
of PTK66 binding to Aurora A kinase, there are considerable
changes in the SERS spectra of the kinases, which validates
the binding of the ligand (see Figure 2, changes indicated by
arrows).
5
-Methyl-4-(2-thiazolylazo) resorcinol (Figure 1A) against Aur-
ora A kinase (Figure 1B, panel I) as well as Aurora B kinase
Figure 1C, panel I) using radiolabel-based in vitro kinase
(
assays. Interestingly, PTK66 could inhibit both Aurora A and
B kinases equally as shown in panel I of Figure 1B and 1C,
respectively. The same was quantified and plotted as the rel-
ative activity against the concentration of PTK66 ranging from
5
to 100 lM (Figure 1B and 1C, panel II). Using the relative
activity measured for three independent experiments, the
in vitro Log IC50 values were calculated using Graphpad.
PTK66 shows an IC50 of approximately 47 lM against Aurora
A kinase and 40 lM against Aurora B kinase. The kinetic
characterisation of the enzyme inhibition was performed
using a fixed concentration of enzyme and histone H3 with
an increasing concentration of [c-32P] ATP in the presence of
varying concentrations of PTK66 using DMSO as a solvent
On gross visual inspection, the changes in the modes in
Aurora B-PTK66 are more pronounced than that of Aurora
A-PTK66 complex. This is also supported by the ITC stud-
ies (discussed later) which show greater change in
enthalpy involving the binding of PTK66 to Aurora B kinase
than Aurora A kinase. This result points to the greater
structural alterations in case of PTK66 binding to Aurora B
9
60
Chem Biol Drug Des 2016; 87: 958–967

PTK66 Inhibits Aurora Kinase A and B
A
B
C
D
Figure 1: PTK66 inhibits Aurora kinases A and B. (A) Chemical structure of PTK66. (B) In vitro kinase assay with Aurora A kinase against
increasing concentration of PTK66 (lM) is represented in panel (I) and the quantification is shown in panel (II). (C) In vitro kinase assay with
Aurora B kinase against increasing concentration of PTK66 (lM) is represented in panel (I) and its quantification in panel (II). AU represents
arbitrary units. The error bars represent the standard deviation. (D) Lineweaver–Burk plot for the kinetic analysis of PTK66 using the
indicated concentration against Aurora A (I) and B (II) kinases, respectively.
Chem Biol Drug Des 2016; 87: 958–967
961

Karthigeyan et al.
Figure 2: Surface binding mode of PTK66.
A) Surface-enhanced Raman spectroscopy
(
spectra of Aurora A kinase (black) and its
complex with PTK66 (red) and (B) Aurora B
kinase (black) and its complex with PTK66
(red). The spectral accumulation time was
30 second.
kinase than Aurora A kinase including hydrogen bond
rearrangements and change in hydration energy which are
observed in ligand-binding events (14,15). Interestingly, our
SERS spectra also pick up these binding events thereby
exhibiting large-scale changes in the vibrational modes. It
has been shown earlier that both Aurora A kinase and
Aurora B kinase bind to the b-sheet domain of the protein
kinase (Figure S2d). By contrast, PTK66 associates with
both Aurora A kinase (Figure S2a) and Aurora B kinase
(Figure S2b) exhibiting 2:1 binding stoichiometry (ligand:
protein) with Aurora A kinase and 1:1 with Aurora B
kinase. All the associations are characterised by micromo-
lar dissociation constants. The interactions of Aurora A
kinase and Aurora B kinase with PTK66 are enthalpy dri-
ven with unfavourable entropy contributions, whereas the
association of Aurora A kinase with felodipine is entropy
driven at 25 °C with both favourable enthalpy and entropy
contributions. Enthalpy–entropy compensation results in
comparable free energy changes making DG negative and
favouring the processes.
(5). And the docking studies (discussed later) show that
the molecule binds very near to the hinge region which is
close to the surface and on the periphery of the b-sheet
domain. Therefore, the changes in the SERS spectra can
be attributed to two main reasons: (i) the direct interfer-
ence of the modes of PTK66 in the spectra. Many of the
modes of PTK66 molecule (Figure S2) can be seen in the
SERS spectra of PTK66 complexed with Aurora A kinase
and Aurora B kinase, similar to the observation in case of
a competitive inhibitor, reversine from our previous work
The docking studies based on the low binding energies
and high number of conformations in the clustering analy-
sis and the ITC experiments, point to two probable binding
sites on Aurora A kinase: one site (site A) near the activa-
tion loop region the other (site B) in the proximity of the
hinge region (Figure 3, panel A). In site A, the PKT66
molecule is hydrogen bonded to residues Arg285 and
Met300 and is also in hydrophobic contact with Cys290,
Gly291, Thr292, Leu293 and Leu296. These residues lie in
close proximity to Thr288 which is autophosphorylated
upon activation. Trp277, another crucial residue, folds into
the active site pocket thereby forming a hydrophobic
pocket. The side chain of this residue also forms hydrogen
bonds with the bound adenosine molecule (6). Thus, bind-
ing of the PTK66 molecule in this region may significantly
inhibit Aurora A kinase by preventing it from adopting an
active conformation. Site B is also significant as it is very
close to the active site of Aurora A kinase. On this site,
PTK66 is hydrogen bonded to Tyr212 and is in hydropho-
bic contact with Leu139, Lys141, Val147, Tyr212, Ala213,
Gly216 and Leu263. Ala213 is a critical residue for binding
with the 6-amino and 1-imido groups of adenosine to the
hinge region of the active site (see Figure S4) through
direct hydrogen bonds with the main chain amides of the
residue. Hence, the binding of PTK66 might bring about a
conformational change in the ATP binding pocket, thus
resulting in inhibition similar to that of felodipine (6). In the
case of Aurora B kinase, the ITC data suggest the pres-
ence of a single binding site. The docking results also
show the presence of a favourable binding site with the
PTK66 molecule hydrogen bonded with Arg81 and Ala157
(6); and (ii) the structural alterations which occur when the
PTK66 binds to the protein. As SERS is highly orientation
dependent, the change in orientation of the amino acid
residues due to the structural perturbation brings about
changes in the spectra.
Moreover, the amide I band, which is sensitive to the pro-
À1
tein structure, at approximately 1620 cm in both Aurora
A and B kinases, undergoes change in the intensity on
complexation with PTK66. This indicates a structural alter-
ation or conformational change near the binding site which
is consistent with the kinetic studies showing mixed type
of inhibition.
Binding sites of PTK66 to Aurora A kinase and
Aurora B kinase
Given that PTK66 inhibits Aurora kinases in a mixed fash-
ion through the surface binding mode in case of both Aur-
ora A and B kinases, we confirmed the physical binding
using ITC. Freshly purified enzymes were titrated against
PTK66. Similarly, felodipine which is known to specifically
bind Aurora A kinase was used as a control. The binding
stoichiometry and the thermodynamic parameters of the
interaction of Aurora A/B kinases with felodipine and
PTK66 were obtained using ITC. The thermograms and
the relevant parameters are shown in Figure S3. Felodipine
binds to Aurora A kinase (Figure S2c) but not Aurora B
9
62
Chem Biol Drug Des 2016; 87: 958–967

PTK66 Inhibits Aurora Kinase A and B
Figure 3: PTK66 docks favourably with
Aurora kinases in silico. (A) Aurora A kinase
docked with PTK66 in two binding sites,
namely site A and site B. The right panels
show the individual close-up snapshots. (B)
Aurora B kinase docked with PTK66 along
with the close-up snapshot in the right.
PTK66 is represented in sticks with carbon,
oxygen, nitrogen and hydrogen atoms
coloured as green, red, blue and white,
respectively. The residues in hydrophobic
contact and hydrogen bonded are coloured
in orange and shown as sticks (right). The
hydrogen bond is shown as dotted lines. (C)
The binding energy, inhibition constant, total
internal energy, electrostatic energy and
hydrogen bond forming residues of the
docked conformation of PTK66 to Aurora A
and B kinases are shown in tabulated form.
(
Figure 3B). This binding pocket also lies in close proximity
plex formation than that of Aurora A–PTK66 complexation.
Although Aurora A–PTK66 complex has more number of
hydrogen bond formations due to an additional binding
site, this does not make its enthalpy change greater than
that of Aurora B–PTK66 complex. This is because of the fact
that the total enthalpy change is not merely an additive
entity of the new hydrogen bond formations but a com-
bined effect of the hydrogen bond rearrangements and
change in hydration energy due to the displacement of
water molecules from the binding pockets. The docking is
performed with a rigid protein model which does not take
into consideration these structural perturbation and
dynamics of bound water molecules.
to the ATP binding active site of Aurora B kinase (refer to
Figure S3), thereby suggesting the possible mechanism of
inhibition where PTK66 upon binding to this site may block
the binding of ATP as well due to a structural change in
the proximity of those binding sites. Hence, a mixed fash-
ion inhibition could be observed with this class of small
molecules. In case of Aurora B kinase, this inhibition
seems to be greater than that in Aurora A kinase due to
the binding of PTK66 in spite of having less number of
binding sites. This is supposedly due to the greater struc-
tural changes on ligand binding which is reflected by
greater change in enthalpy during Aurora B–PTK66 com-
Chem Biol Drug Des 2016; 87: 958–967
963

Karthigeyan et al.
PTK66 inhibits Aurora kinases in cells
the PTK66-treated HEK293T cells, which shows a dose-
dependent (0, 12.5, 25, 50 lm) increase in the aneuploid
population represented by the M5 population in (Figure 4A)
panel I and the >4N population in the panel II. Additionally,
the time-dependent (0, 12, 24, 36 and 48 h) drop in cellu-
lar proliferation (Figure 4B, panel I) along with an increase
in the >4N population was detected in the same cell type
upon treatment with 20 lm PTK66 (Figure 4B, panel II).
Further, the histone phosphorylation status upon treatment
with increasing concentrations of PTK66 was verified via
Western blot analysis using antiphosphor-H3(S10) anti-
body normalised against Histone H3 with an anti-H3 anti-
body using Laemmli lysates prepared 24 h post-treatment
with either HEK293T, H1299 or U2OS cell lines (Figure 5,
panels A, B and C, respectively). This increase in polyploid
population along with a strong drop in the histone H3-S10
phosphorylation is highly suggestive of Aurora B kinase
inhibition predominantly as compared to Aurora A kinase.
The majority of the aneuploidy culminated from mitotic
defects such as chromosome mis-segregation and spindle
pole abnormalities which were observed via indirect
Upon establishing Aurora kinase inhibition biochemically,
PTK66 was tested in cells using the HEK293T cell line.
The viability was assessed for a wide range of micromolar
concentrations of ranging from 5 to 100 lm of PTK66
using MTS assay. These results indicated least effect over
the viability achieving approximately 30% inhibition at
1
00 lm concentration (Figure S3a). This result was further
corroborated for the onset of apoptosis using fluorophore
labelled early markers of apoptosis in DMSO versus
PTK66 (10 lM)-treated HEK293T cells (Figure S5b, panels
I and II, respectively) after 24 h of treatment using Annexin
V-Cy3 (red) while the viable cells were labelled with 6-car-
boxyfluorescein diacetate (6-CFDA) (green). Garcinol
(5 lm) was used as a positive control (Figure S5b, panel
III). PTK66-treated cells showed only 6-CFDA staining
there by ruling out the possibility of toxicity. Furthermore,
dual Aurora kinase inhibition in cells was followed by two
significant hallmarks, namely cell cycle defects and histone
H3 (S10) phosphorylation (16). The DNA content was anal-
ysed via propidium iodide staining and FACS analysis of
Figure 4: PTK66 induces aneuploidy in
mammalian cell lines. (A) HEK293T cells
treated with either DMSO or the indicated
concentration of PTK66 for 24 h and
subjected to (panel I) FACS analysis using
propidium iodide staining. The number of
cells (%) displaying different ploidy levels
among the analysed population in each cell
line was (panel II) tabulated. (B) HEK293T
cells treated with 20 lM PTK66 for the
indicated time and (Panel I) DIC images
(209 magnification). (Panel II) DNA content
analysed using the BD FACS calibre. The
number of cells (%) under each cell cycle
stage was plotted as a bar graph
representing the same (panel III) with the
indicated DNA content.
9
64
Chem Biol Drug Des 2016; 87: 958–967

PTK66 Inhibits Aurora Kinase A and B
immunofluorescence staining using an a-tubulin antibody
Figure 5D, panel I). The quantification revealed a signifi-
same extent. These results suggest that the resorcinol
scaffold could be an important molecular entity to design
Aurora kinase inhibitors for therapeutic purposes.
(
cant increase in mitotic defects in all three cell types upon
PTK66 treatment (Figure 5D, panel II). Further to assess
the specificity of PTK66, the residual activity against a
panel of 20 kinases related to Aurora family kinases and
those which are known to affect histone phosphorylation
were compared with the solvent (DMSO) control (Proqi-
nase, GmbH). As tabulated in Table S1, all the three mem-
bers of Aurora were inhibited by approximately 50% at a
Conclusion
Together, our experimental results and observations led
to the discovery of a novel resorcinol scaffold that can
target the hinge region of Aurora kinases A and B. Fur-
thermore, upon using ITC and molecular docking, this
scaffold was shown to have at least two favourable bind-
ing sites with Aurora A kinase and one site with Aurora B
(IC50) concentration of 40 lM. However, there were a few
other kinases, related to Aurora kinases, such as NEK2
and JNK1 which were also found to be inhibited to the
Figure 5: PTK66 inhibits H3S10 phosphorylation. Western blot analysis of (A) HEK293T, (B) H1299 and (C) U2OS cells treated with
DMSO/PTK66 for 24 h using the corresponding cell lysates using phospho-H3(S10) antibody and H3 antibody (loading control). Panel I
shows the autoradiogram, and panel II represents the corresponding quantification. The error bars represent the standard deviation for
triplicate experiments. (D) Indirect immunofluorescence assay using a-tubulin antibody and DAPI counter staining of H1299 cells (panel I)
to visualise the mitotic defects. Panel II shows the quantification of mitotic defects in the indicated cell lines after 40 lM PTK66 treatment
for 24 h.
Chem Biol Drug Des 2016; 87: 958–967
965

Karthigeyan et al.
kinase. Further, by means of SERS technique, we could
establish the surface binding of this scaffold with the Aur-
ora family kinases in general. Moreover, this binding
could effectively inhibit the kinase activity of both
enzymes in vitro. Similarly, the inhibition was also con-
vincingly established in cell lines upon measuring the
ploidy status and the histone H3 S10 phosphorylation
level after treatment with PTK66. Thus, the novel resorci-
nol scaffold containing PTK66 is proven to target the
Aurora family kinases. Future research may be directed
to design more potent and specific inhibitors of Aurora
kinases. Most of the reported inhibitors possess adenine-
like scaffolds with two 5 or 6 member rings that make
contact with the hinge region through residues such as
Leu210, Glu211, Tyr212, Ala213, Leu139, Val147 and
Leu263 (17). Similarly, the resorcinol scaffold has been
shown for the first time to contact some of these key
residues from the outer surface rather than the ATP
pocket. This binding may be one of the reasons for the
low micromolar inhibition of Aurora A kinase by PTK66,
which can be tailored to increase the potency by suitable
modification of this scaffold in future. Importantly, the
same scaffold can be targeted to a similar surface pocket
over Aurora B kinase, which can effectively hydrogen
bond with Ala157, a key residue that is often targeted by
Aurora B kinase inhibitor molecules (18). This finding
could be considered as noteworthy information to the
future design of potent dual inhibitors of Aurora kinases
using this resorcinol scaffold.
2. Kelly K.R., Ecsedy J., Mahalingam D., Nawrocki S.T.,
Padmanabhan S., Giles F.J., Carew J.S. (2011) Target-
ing aurora kinases in cancer treatment. Curr Drug Tar-
gets;12:2067–2078.
3. Cheung C.H., Sarvagalla S., Lee J.Y., Huang Y.C.,
Coumar M.S. (2011-2013) Aurora kinase inhibitor
patents and agents in clinical testing: an update.
Expert Opin Ther Pat;24:1021–1038.
4. Pollard J.R., Mortimore M. (2009) Discovery and devel-
opment of aurora kinase inhibitors as anticancer
agents. J Med Chem;52:2629–2651.
5. Soumik S., Dhanasekaran K., Partha P.K., Tapas K.K.,
Chandrabhas N. (2013) Surface enhanced Raman
spectroscopy of Aurora kinases: direct, ultrasensitive
detection of autophosphorylation. RSC Adv;3:4221–
4230.
6. Karthigeyan D., Siddhanta S., Kishore A.H., Perumal
ꢀ
S.S., Agren H., Sudevan S., Bhat A.V., Balasubra-
manyam K., Subbegowda R.K., Kundu T.K., Narayana
C. (2014) SERS and MD simulation studies of a kinase
inhibitor demonstrate the emergence of a potential
drug discovery tool. Proc Natl Acad Sci.
U S
A;111:10416–10421.
7. Pavan Kumar G.V., Narayana C. (2007) Adapting a flu-
orescence microscope to perform surface enhanced
Raman spectroscopy. Curr Sci;96:778–781.
8. Nowakowski J., Cronin C.N., McRee D.E., Knuth
M.W., Nelson C.G., Pavletich N.P., Rogers J., Sang
B.C., Scheibe D.N., Swanson R.V., Thompson D.A.
(2002) Structures of the cancer-related Aurora-A, FAK,
and EphA2 protein kinases from nanovolume crystal-
lography. Structure;10:1659.
Acknowledgments
9
. Elkins J.M., Santaguida S., Musacchio A., Knapp S.
(2012) Crystal structure of human Aurora B in complex
with INCENP and VX-680. J Med Chem;55:7841–
7848.
T.K.K. acknowledges JNCASR (Govt of India) and the
financial support of Department of Biotechnology, Grant
BT/01/CEIB/10/III/01, and J. C. Bose National Fellowship.
D.K. acknowledges Council of Scientific and Industrial
Research Senior Research Fellowship.
10. Frisch M.J., Trucks G.W., Schlegel H.B., Scuseria
G.E., Robb M.A., Cheeseman J.R., Montgomery J.A.
Jr. et al. (2004) Gaussian 03, Revision C.02. Walling-
ford, CT.
Author Contributions
11. Morris G.M., Huey R., Lindstrom W.W, Sanner M.F.,
Belew R.K., Goodsell D.S., Olson A.J. (2004)
AutoDock4 and AutoDockTools4: automated docking
with selective receptor flexibility. J Comp Chem;30:
2785.
The manuscript was written through contributions of all
authors. All authors have given approval to the final version
of the manuscript.
1
2. Schrodinger L.L.C. (2010) The PyMOL Molecular
Graphics System, Version 1.3r1.
Conflict of Interest
13. Cheetham G.M.T., Knegtel R.M.A., Coll J.T., Renwick
S.B., Swenson L., Weber P., Lippke J.A., Austen D.A.
(2002) Crystal structure of Aurora-2, an oncogenic ser-
ine/threonine kinase. J Biol Chem;277:42419–42422.
The authors declare no competing financial interest.
1
4. Connelly P.R., Aldape R.A., Bruzzese F.J., Chambers
S.P., Fitzgibbon M.J., Fleming M.A., Itoh S., Livingston
D.J., Navia M.A., Thomson J.A. (1994) Enthalpy of
hydrogen bond formation in a protein-ligand binding
reaction. Proc Natl Acad Sci U S A;91:1964.
References
1
. Karthigeyan D., Prasad S.B., Shandilya J., Agrawal S.,
Kundu T.K. (2011) Biology of Aurora A kinase: implica-
tions in cancer manifestation and therapy. Med Res
Rev;31:757–793.
15. Poornima C.S., Dean P.M. (1995) Hydration in drug
design. 1. Multiple hydrogen-bonding features of water
9
66
Chem Biol Drug Des 2016; 87: 958–967

PTK66 Inhibits Aurora Kinase A and B
Figure S1. NMR spectra of PTK66.
molecules in mediating protein-ligand interactions.
J Comput Aided Mol Des;9:500.
1
6. Crosio C., Fimia G.M., Loury R., Kimura M., Okano Y.,
Zhou H., Sen S., Allis C.D., Sassone-Corsi P. (2002)
Mitotic phosphorylation of histone H3: spatio-temporal
regulation by mammalian aurora kinases. Mol Cell
Biol;22:874.
Figure S2. SERS spectrum of PTK66 only.
Figure S3. PTK66 shows thermodynamically favorable
binding with Aurora kinases.
1
1
7. Yan A., Wang L., Xu S., Xu J. (2011) Aurora-A kinase
inhibitor scaffolds and binding modes. Drug Discov
Today;16:260.
Figure S4. PTK66 complex with Aurora kinases in the
ATP binding pocket.
8. Sidra B., Saba F., Mohammad A.K., Hira I., Sajid R.
Figure S5. Cellular effects of PTK66 in HEK 293 cells.
(2013) In silico screening for identification of novel Aur-
ora kinase inhibitors by molecular docking, dynamics
simulations and ligand-based hypothesis approaches.
Enzyme Eng;2:1.
Table S1. Profiling of PTK66 at 40 lM concentration
against 30 protein kinases.
Scheme S1. Shown above is the scheme followed to syn-
thesis PTK66.
Supporting Information
Additional Supporting Information may be found online in
the supporting information tab for this article.
Chem Biol Drug Des 2016; 87: 958–967
967