Design and synthesis of BPR1K653 derivatives targeting the back pocket of Aurora kinases for selective isoform inhibition

Article abstract of DOI:10.1016/j.ejmech.2018.03.064

Twenty five novel chemical analogs of the previously reported Aurora kinase inhibitor BPR1K653 (1-(4-(2-((5-chloro-6-phenylfuro[2,3-d]pyrimidin-4-yl)amino)ethyl)phenyl)-3-(2-((dimethylamino)methyl)phenyl)urea) have been designed, synthesized, and evaluated by Aurora-A and Aurora-B enzymatic kinase activity assays. Similar to BPR1K653, analogs 3b-3h bear alkyl or tertiary amino group at the ortho position of the phenylurea, and showed equal or better inhibition activity for Aurora-B over Aurora-A. Conversely, preferential Aurora-A inhibition activity was observed when the same functional group was moved to the meta position of the phenylurea. Compounds 3m and 3n, both of which harbor a tertiary amino group at the meta position of the phenylurea, showed 10–16 fold inhibition selectivity for Aurora-A over Aurora-B. The in vitro kinase inhibition results were verified by Western blot analysis, and indicated that compounds 3m and 3n were more than 75-fold superior in inhibiting T-loop autophosphorylation of Aurora-A (Thr288), compared to Aurora-B (Thr232) in HCT116 colon carcinoma cells. The computational docking analysis suggested that the tertiary amine at the meta position of the phenylurea formed a more stable interaction with residues in the back pocket of Aurora-A than in Aurora-B, a possible explanation for the observed discrepancy in the selectivity. These results support an alternative small molecule design strategy targeting the back pocket of Aurora kinases for selective isoform inhibition.

Full text of DOI:10.1016/j.ejmech.2018.03.064

Accepted Manuscript
Design and synthesis of BPR1K653 derivatives targeting the back pocket of Aurora
kinases for selective isoform inhibition
Yi-Yu Ke, Chun-Ping Chang, Wen-Hsing Lin, Chia-Hua Tsai, I-Chen Chiu, Wan-Ping
Wang, Pei-Chen Wang, Pei-Yi Chen, Wen-Hsin Lin, Chun-Feng Chang, Po-Chu Kuo,
Jen-Shin Song, Chuan Shih, Hsing-Pang Hsieh, Ya-Hui Chi
PII:
S0223-5234(18)30307-6
10.1016/j.ejmech.2018.03.064
EJMECH 10329
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 18 August 2017
Revised Date: 14 March 2018
Accepted Date: 21 March 2018
Please cite this article as: Y.-Y. Ke, C.-P. Chang, W.-H. Lin, C.-H. Tsai, I.-C. Chiu, W.-P. Wang, P.-
C. Wang, P.-Y. Chen, W.-H. Lin, C.-F. Chang, P.-C. Kuo, J.-S. Song, C. Shih, H.-P. Hsieh, Y.-H. Chi,
Design and synthesis of BPR1K653 derivatives targeting the back pocket of Aurora kinases for selective
isoform inhibition, European Journal of Medicinal Chemistry (2018), doi: 10.1016/j.ejmech.2018.03.064.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Design and synthesis of BPR1K653 derivatives targeting the back pocket of Aurora
kinases for selective isoform inhibition
a, 1
a, 1
a
a
a
Yi-Yu Ke , Chun-Ping Chang , Wen-Hsing Lin , Chia-Hua Tsai , I-Chen Chiu ,
a
a
a
a
a
Wan-Ping Wang , Pei-Chen Wang , Pei-Yi Chen , Wen-Hsin Lin , Chun-Feng Chang ,
a
a,b**
a,c*
Po-Chu Kuo, Jen-Shin Song, Chuan Shih , Hsing-Pang Hsieh , Ya-Hui Chi
Institute of Biotechnology and Pharmaceutical Research, National Health Research
a
Institutes, Zhunan 35053, Taiwan, R. O. C.
b
Department of Chemistry, National Tsing Hua University, Hsinchu 30013, Taiwan,
ROC.
c
Graduate Institute of Basic Medical Science, China Medical University, Taichung
4
0402, Taiwan, R. O. C.
*
*
* Corresponding author.
Corresponding author.
E-mail addresses: ychi@nhri.org.tw (Y.H. Chi), hphsieh@nhri.org.tw (H.P. Hsieh)
1
These authors are equal contributors.
Abbreviations. SAC, spindle assembly checkpoint; CPC, chromosomal passenger
complex; pHH3Ser10, phosphorylation on serine 10 of histone H3; MI, mitotic index.
1

ACCEPTED MANUSCRIPT
Highlights
•
•
•
•
25 new furanopyrimidine analogs were designed and synthesized.
The designed compounds were evaluated for Aurora-A and Aurora-B inhibition activity.
Derivatization on the phenylurea modulates isoform-selective kinase inhibition.
Molecular docking studies were performed.
Graphical abstract
2

ACCEPTED MANUSCRIPT
Abstract
Twenty five novel chemical analogs of the previously reported Aurora kinase inhibitor
BPR1K653 (1-(4-(2-((5-chloro-6-phenylfuro[2,3-d]pyrimidin-4-yl)amino)ethyl)phenyl)-
3
-(2-((dimethylamino)methyl)phenyl)urea) have been designed, synthesized, and
evaluated by Aurora-A and Aurora-B enzymatic kinase activity assays. Similar to
BPR1K653, analogs 3b-3h bear alkyl or tertiary amino group at the ortho position of the
phenylurea, and showed equal or better inhibition activity for Aurora-B over Aurora-A.
Conversely, preferential Aurora-A inhibition activity was observed when the same
functional group was moved to the meta position of the phenylurea. Compounds 3m and
3
n, both of which harbor a tertiary amino group at the meta position of the phenylurea,
showed 10-16 fold inhibition selectivity for Aurora-A over Aurora-B. The in vitro kinase
inhibition results were verified by Western blot analysis, and indicated that compounds 3m
and 3n were more than 75-fold superior in inhibiting T-loop autophosphorylation of
Aurora-A (Thr288), compared to Aurora-B (Thr232) in HCT116 colon carcinoma cells.
The computational docking analysis suggested that the tertiary amine at the meta position
of the phenylurea formed a more stable interaction with residues in the back pocket of
Aurora-A than in Aurora-B, a possible explanation for the observed discrepancy in the
selectivity. These results support an alternative small molecule design strategy targeting
3

ACCEPTED MANUSCRIPT
the back pocket of Aurora kinases for selective isoform inhibition.
Key words: Aurora kinase inhibitor, anticancer agents, docking study
4

ACCEPTED MANUSCRIPT
Introduction
Aurora kinases are serine/threonine kinases which regulate mitotic progression,
centrosome maturation, and spindle assembly [1, 2]. Overexpression of Aurora kinases has
been associated with increased tumor progression, and thus they are appealing targets for
the development of anti-cancer therapies [3-7].
Despite similarities in protein sequence, the sub-cellular localization, expression
patterns, timing, and activity of Aurora-A are markedly different to Auroras -B and -C [1].
Aurora-A primarily localizes to centrosomes and mitotic spindles where it regulates
centrosome duplication, centrosome maturation, and mitotic spindle formation.
Disruption of Aurora-A kinase activity arrests cells in the G2/M stage of the cell cycle,
thereby activating the spindle assembly checkpoint (SAC) and thus growth inhibition [8,
9
]. On the other hand, Aurora-B is part of the chromosomal passenger complex (CPC), a
hetero-tetrameric complex composed of Aurora-B, inner centromere protein (INCENP),
Borealin and Survivin [10, 11]. In prometaphase, the CPC localizes at the centromeric
region of the chromosome, monitoring proper attachment of microtubules emanating from
opposite poles of the cell, activating SAC, as well as promoting chromosome congression
to the metaphase plate [12-14]. Inhibiting the kinase activity of Aurora-B or depleting
INCENP results in dephosphorylation of histone H3 on serine 10 and override of SAC
5

ACCEPTED MANUSCRIPT
[15, 16], leading to premature decondensation of the chromatin, an increased degree of
polyploidy, cell senescence, and apoptosis [17, 18].
Aurora -A or -B selective and pan-aurora inhibitors have demonstrated different
preclinical and clinical therapeutic efficacies [2, 19-23]. For example, clinical trials for a
pan-Aurora inhibitor VX-680 (developed by Vertex) were halted at phase II for toxicity
reasons (one case of heart failure) [6, 24]. VX-689, a newer version of VX-680 which
showed higher Aurora-A selectivity, has entered a phase I clinical trial [25]. MLN8237,
developed by Millennium, exhibited better Aurora-A selectivity over MLN8054. A clinical
trial of MLN8054 was terminated due to severe neutropenia, while MLN8237 has entered
phase III clinical trials [25, 26]. These clinical observations suggest that development of
Aurora-A selective compounds may be beneficial for the treatment of solid tumors due to
the reduced toxicity to cells in the myeloid lineage. On the other hand, higher response
rates against hematologic malignancies have been observed in clinical trials for drugs
with better Aurora-B inhibition efficacy [27].
We previously discovered a pan-Aurora kinase inhibitor named BPR1K653 (i.e.
1
-(4-(2-((5-chloro-6-phenylfuro[2,3-d]pyrimidin-4-yl)amino)ethyl)phenyl)-3-(2-((dimeth
ylamino)methyl)phenyl)urea, Fig. 1) which showed potency against various tumors
independent of the expression of MDR1 (multidrug resistance protein 1) [28]. By synthesis
6

ACCEPTED MANUSCRIPT
of a series of furanopyrimidine analogs followed by activity validation using enzymatic
and cellular assays, and molecular docking, these small molecules were determined to
interact differently with residues in the back pocket of Aurora-A and Aurora-B, accounting
for their selective inhibition activities.
Fig. 1. Molecular structure of BPR1K653, a pan-Aurora kinase inhibitor.
2
. Results and Discussion
2
.1. Chemistry
The preparation of compounds 3a-3y bearing modifications on the urea side chain was
achieved by modifying strategies in previous reports (Scheme1) [29, 30]. S Ar reaction of
N
dichlorofuranopyrimidine 1 [30] with 4-(2-aminoethyl)aniline in ethanol gave the desired
intermediate compound 2, which was converted to the corresponding urea compound
3
a-3y by reaction with various isocyanates or carbamates, respectively.
7

ACCEPTED MANUSCRIPT
Scheme 1. Synthesis of compounds 3a-3y. Reaction conditions and reagents: (a)
4
-(2-aminoethyl)aniline, EtOH, reflux, 1 h, 86%; (b) (i) various isocyanates, CH Cl , rt, 4
2 2
h, 34-66%; or (ii) various carbamates, Et N, 1,4-dioxane, 120 ℃, 8 h, 20-64%.
3
2
2
.2. Biological evaluation
.2.1. Aurora inhibition activity of the compounds
Enzymatic inhibition activities of compounds 3a-3y to Aurora-A and Aurora-B were
evaluated by ATP consumption assays. Table 1 summarizes the structure-activity
relationship (SAR) results for different substituents at the ortho, meta or para position of
the phenylurea. The results indicated that removing all substitutions on the phenylurea (i.e.
compound 3a) resulted in near equal inhibition activity to Aurora-A and Aurora-B. On the
other hand, replacing the N,N-dimethyl tertiary amino group of BPR1K653 (Aurora-A
IC50 = 124.0 nM; Aurora-B IC50 = 45.0 nM) with an isobutyl group (compound 3e)
8

ACCEPTED MANUSCRIPT
dramatically reduced inhibition activity of Aurora-A and Aurora-B (Aurora-A IC50
=
3
000.0 nM; Aurora-B IC = 700.3 nM). This effect was less significant when the number
50
of carbon atoms in R1 is less than four (i.e. compounds 3b-3d). Compound 3h, which
contains a 4-hydroxypiperidinyl group at the ortho position of the phenylurea showed
increased Aurora-B inhibition activity compared to 3f and 3g, implying that improved
hydrophilicity may be beneficial for the interaction between the functional groups and
residues in the back pocket of Aurora-B.
Relocation of the isobutyl group of 3e from the ortho position to the meta position of
phenylurea gave 3l, which was a much more potent inhibitor of Aurora-A than Aurora-B
(
Aurora-A IC = 247.0 nM; Aurora-B IC >10000 nM). Our results for 3i-3l suggested
50 50
that Aurora-B inhibition potency drastically decreased with increased number of carbons at
the meta position of the phenylurea. The Aurora-A inhibition potency was increased 12
fold in 3m (Aurora-A IC = 20.0 nM), which bears a N,N-dimethyl tertiary amino group
5
0
at the meta position of phenylurea (Table 1), compared to 3l (Aurora-A IC = 247.0 nM),
50
which contains an isobutyl group. Similar Aurora-A inhibition potency was observed in 3n
Aurora-A IC = 9.0 nM) and 3o (Aurora-A IC = 21.5 nM). However, further extension
(
5
0
50
of the carbon chain on the tertiary amino group (3p-3q) reduced the Aurora-A inhibition
potency compared to 3o.
9

ACCEPTED MANUSCRIPT
We also synthesized compounds bearing the same alkyl and tertiary amino groups at
the para position of the phenylurea, and measured their inhibition activity to Aurora-A and
Aurora-B. As indicated in Table 1, addition of tertiary amino groups at the para position of
phenylurea (i.e. compounds 3v-3y) did not improve inhibition selectivity to Aurora-A
compared to the meta derivatives (i.e. compounds 3m-3q), or the inhibition selectivity to
Aurora-B compared to the ortho derivatives (i.e. compounds 3f-3h and BPR1K653).
1
0

ACCEPTED MANUSCRIPT
Table 1. Structure and enzymatic evaluation of the prepared compounds.
a
IC values are calculated from duplicated ATP consumption assays of 9 titration points.
5
0
b
c
IC values are calculated from duplicated ATP consumption assays of 7 titration points.
5
0
Less than 50% inhibition at 1000 nM.
1
1

ACCEPTED MANUSCRIPT
2
.2.2 Aurora-A and Aurora-B inhibitory activity of the synthesized compounds in cells
Autophosphorylation of Aurora kinases has been found to be indispensable for their
activation [31, 32]. To test the inhibition selectivity of the synthesized compounds in cells,
we profiled phosphorylated Aurora-A (Thr288), Aurora-B (Thr232) and Aurora-C (Thr198)
using Western blotting. Compounds 3m and 3n were found to be more potent inhibitors of
autophosphorylation of Aurora-A than Aurora-B, while compounds 3h and BPR1K653
favored Aurora-B inhibition (Fig. 2 and Table 2). These data are consistent with our
observations made in the enzymatic kinase activity assays (Table 1). We noted that the
inhibition potency of compounds 3m and 3n to Aurora-B (Thr232) phosphorylation is
much less than compound 3h in HCT116 cells (Table 2) though the cell-free enzymatic
inhibition activity is only 2-fold less (Table 1). This discrepancy may be due to
preference of compounds 3m and 3n to interact with Aurora-A than Aurora-B in cells
which preserve authentic complex interactions.
Inhibiting the kinase activity of either Aurora-A or Aurora-B can halt cell
proliferation, though the mechanisms of action are different [1]. In addition, inhibiting the
kinase activity of Aurora-B has been shown to be more efficient than Aurora-A for cell
growth arrest and apoptosis [33-37]. That is probably why compounds 3a, 3h, 3m and 3n
1
2

ACCEPTED MANUSCRIPT
all inhibited proliferation of HCT116 human colon carcinoma cells at the sub-micro molar
range (Table 2), although they possess various inhibition activities against Aurora
isoforms. BPR1K653 showed the best anti-proliferation activity, perhaps because it is
also the most potent Aurora-B inhibitor amongst the analyzed compounds (Tables 1-2).
To determine if the compounds are cytotoxic to non-cancerous cells, we tested
proliferation inhibition of the compounds to Detroit 551 (normal human skin fibroblasts).
The inhibition concentrations of compounds 3a, 3h, 3m and 3n to the growth of Detroit
5
51 are more than 3 µM, suggesting that these compounds are less toxic to normal human
skin fibroblasts compared to HCT116 colorectal cancer cells (Table 2).
1
3

ACCEPTED MANUSCRIPT
Fig. 2. Representative cellular inhibition profile of the compounds to Aurora kinases.
Immunoblot of phosphorylated Aurora-A (Thr288), Aurora-B (Thr232) and Aurora-C
(Thr198). HCT116 cells were arrested in the M phase with 24 h treatment of 200 nM
nocodazole, followed by 1 h treatment with the indicated compounds at various
concentrations. ACTIN immunoblot was included as a loading control.
Table 2. Biological evaluation of the prepared compounds.
pAURKA pAURKB Cellular HCT116 Detroit 551
a
a
Compd.
R1
H
R2
R3
IC50
IC50
IC Ratio viability
viability
5
0
b
b
(
nM)
(nM)
23.7
A/B
3.1/1
2/1
IC50 (nM) IC50 (µM)
BPR1K653
H
H
H
73.4
389.6
1513.0
59.9
70
3.1
7.3
3.5
5.0
9.7
H
H
3
a
191.2
132.6
153
310
578
342
3
h
H
H
H
11.4/1
3
m
H
H
>10,000 1/(>166.9)
>10,000 1/(>75.6)
3
n
132.2
a
IC values are calculated from Western blot profiles (Fig. 2) of 9 titration points.
IC50 values are calculated from a duplicated, 10-point titration.
5
0
b
1
4

ACCEPTED MANUSCRIPT
2
.2.3 Effect of the synthesized compounds on mitotic progression
Phosphorylation on the serine 10 of histone H3 (pHH3Ser10) is widely used as an
immunomarker specific to cells undergoing mitosis [38]. To determine the effect of the
newly synthesized compounds on mitotic progression, we measured pHH3Ser10-positive
(
i.e. mitotic index, MI) HeLa cells released from thymidine block at various time points
Fig. 3A). We used HeLa instead of HCT116 in mitotic progression analysis due to the high
(
synchronization efficiency of this cell line by thymidine block [39]. Compared to mock and
DMSO-treated controls, treatment with compound 3m significantly increased MI at 10-12
hours post thymidine block (Fig. 3B-C), suggesting compound 3m can delay mitotic
progression – an effect which should be attributed to Aurora-A kinase inhibition. In
contrast, BPR1K653 and compound 3h drastically reduced MI (Fig. 3B-C), suggesting
that these two compounds are more potent Aurora-B kinase inhibitors. Compound 3a,
which does not possess isoform-inhibition selectivity, slightly increased, MI at 10-12 hours
post thymidine block, but to a lesser degree compared to compound 3m. These data
indicate that the cellular phenotype of the compounds on mitotic progression correlated
with their inhibition selectivity to Aurora-A or Aurora-B.
1
5

ACCEPTED MANUSCRIPT
Fig. 3. (A) Schematic illustration of the protocol for the functional study of the
compounds on mitotic progression. (B) Immunofluorescence staining of pHH3Ser10 (red)
on compound-treated cells fixed at 12 hours after release from thymidine block. Nuclei
were counterstained with Hoechst 33342 (blue). Representative images were shown.
Percentage of pHH3Ser10-positive cells were quantified in (C) as shown as mitotic index
(MI). About 2000-3000 cells were quantified using MetaMorph software in each
experimental condition.
1
6

ACCEPTED MANUSCRIPT
2
2
.3. Molecular docking study
.3.1. Structure alignment of Aurora-A and Aurora-B
A typical ATP binding pocket of a kinase can be divided into six sub-pockets: the
adenine pocket, ribose pocket, phosphate pocket, hydrophobic pocket-I, hydrophobic
pocket-II and back pocket [40, 41]. The protein sequences in the ATP-binding site of
Aurora-A and Aurora-B are highly conserved. Only three amino acids are not aligned:
Leu215, Thr217, and Arg220 in Aurora-A; and Arg159, Glu161, and Lys164 in Aurora-B
(
Fig.
S1,
supplementary
data)
[42].
The
interaction
between
imidazo[4,5-b]pyridine-derived small molecules with the Thr217 residue of Aurora-A has
been reported to play a critical role in governing the isoform selectivity for Aurora-A
inhibition [43]. How the other sub-pockets contribute to the inhibition selectivity of
Aurora isoforms is not clear.
Our previous structural analysis suggested that the phenylurea derivative on the
4
-position of furanopyrimidine extended to the back pocket of Aurora-A [29, 44]. To
understand the structural differences in the back pocket of Aurora isoforms, we compared
X-ray profiles of human Aurora-A (PDB ID: 2W1C, in complex with an Aurora inhibitor)
and Aurora-B (PDB ID: 4AF3, in complex with INCENP and VX-680) (Fig. 4) [10, 45].
The protein sequences of Aurora-A and Aurora-B in the region of back pocket are very
1
7

ACCEPTED MANUSCRIPT
well aligned (Fig. 4A). However, a surface modeling analysis determined that the spatial
3
vacancy in the back pocket is about 70% greater in Aurora-B than in Aurora-A (179.59 Å
3
in Aurora-A; 310.33 Å in Aurora-B, Fig. 4B). In addition, residues Gln177, Glu181,
Gln185 in the αC-helix of Aurora-A and the corresponding Gln121, Glu125, Gln129 in
Aurora-B showed more significant deviation than residues in the β3-β4-β5 three-stranded
antiparallel beta sheet (Fig. 4C).
Fig. 4. (A) Alignment of partial amino acid sequence of Aurora-A and Aurora-B.
Residues that contribute to the back pocket are highlighted in yellow background. (B)
Structure alignment of Aurora-A (PDB ID: 2W1C, khaki) and Aurora-B (PDB ID: 4AF3,
cyan). Surface modeling of ATP-binding site and back pocket are presented in the same
color. (C) Conformation variation in the back pocket of Aurora-A (khaki) and Aurora-B
1
8

ACCEPTED MANUSCRIPT
(blue).
2
.3.2. Molecular docking for selective inhibitors
The biological evaluation results indicated that compound 3a (which does not contain
substitution on the phenylurea) showed near equal inhibition to Aurora-A and Aurora-B,
and the addition of substituents on the ortho or meta position of the phenylurea resulted in
increased selectivity for Aurora-B or Aurora-A, respectively (Table 1). The molecular
docking analysis of compound 3a in complex with Aurora-A and Aurora-B suggested that
the ortho position of the phenyl group of 3a presented more space in complex with
Aurora-B than in Aurora-A, while the meta-position of the phenyl group was more distant
from Aurora-A than Aurora-B (Fig. 5). Therefore, extending moieties at the ortho or meta
positions of the phenylurea in 3a may affect the inhibition selectivity for Aurora-A or
Aurora-B.
The molecular docking structure of Aurora-A in complex with compound 3n
suggested that the ethyl amino group not only has a strong CH-π hydrophobic interaction
with Phe144 of Aurora-A, but also forms an intramolecular CH-π interaction with its
phenyl group (Fig. 6A). This intramolecular hydrophobic interaction enables better fitting
of compound 3n in the back pocket of Aurora-A. On the other hand, functional groups
such as the piperidinyl and 4-hydroxypiperidinyl groups in compounds 3p and 3q may be
1
9

ACCEPTED MANUSCRIPT
too large to be accommodated in the back pocket of Aurora-A, thereby decreasing the
potency with which they inhibit Aurora-A (Table 1). In contrast, compound 3h (which
bear a 4-hydroxypiperidinyl derivative on the ortho position of the phenylurea), presented
about 3-fold better enzymatic activity to Aurora-B than Aurora-A (Table, 1). A molecular
docking analysis of compound 3h with Aurora-A and Aurora-B suggested that the -OH at
the 4-position of the piperidine in 3h may form hydrogen bonding with residue Gln121 in
Aurora-B (Fig. 6B), while causing a steric obstacle with Gln177 of Aurora-A. Conversely,
compounds 3f and 3g (which fail to form hydrogen bonding with Gln121 of Aurora-B) did
not show an inhibition preference to either Aurora-A or Aurora-B (Table 1).
2
0

ACCEPTED MANUSCRIPT
Fig. 5. (A) The docking analysis for compound 3a with Aurora-A. The back pocket of
Aurora-A is displayed in transparent surface. (B) The docking analysis for compound 3a
with Aurora-B. The back pocket of Aurora-B is displayed in transparent surface.
Approximate distance of the molecules at the indicated positions are shown.
2
1

ACCEPTED MANUSCRIPT
Fig. 6. (A) The docking analysis for compound 3n (yellow) with Aurora-A. Compound 3n
forms hydrogen bonding with Lys162, Glu181, and Ala213. Compound 3n has a strong
hydrophobic effect with Phe144. (B) The docking analysis for compound 3h (yellow) with
Aurora-B. Compound 3h forms hydrogen bonding with Lys106, Gln121, Glu125, and
Ala157.
2
2

ACCEPTED MANUSCRIPT
3
. Conclusions
The molecular and cellular functions of Aurora kinases -A, -B and -C are distinct in space
and time. Therefore, small molecules that selectively inhibit Aurora isoforms are
expected to show different mechanisms of action, and thus potentially target different
subsets of patients in clinical applications. In this study, we discovered BPR1K653
(1-(4-(2-((5-chloro-6-phenylfuro[2,3-d]pyrimidin-4-yl)amino)ethyl)phenyl)-3-phenyl)ure
a) derivatives 3m and 3n bearing substituents at the meta position of the phenylurea, and
found that they displayed a high degree of inhibition selectivity for Aurora-A over
Aurora-B in both enzymatic and cellular assays, while the superior Aurora-B inhibition
selectivity was maintained in an ortho derivative 3h. The molecular docking analysis of
the compounds with Aurora-A and Aurora-B suggested that the spatial variation in the
back pocket can be used as a design principle for isoform-specific inhibitors. This
selectivity design strategy is different from previous compounds which target the
non-conserved amino acids in the ATP-binding site of Aurora-A and Aurora-B. The
pharmacokinetic and pharmacodynamic properties of the lead compounds with the best
Aurora-A selectivity are currently being evaluated for in vivo animal multi-drug resistant
tumor xenograft models.
2
3

ACCEPTED MANUSCRIPT
2
4

ACCEPTED MANUSCRIPT
4
. Experimental
4
.1. Chemistry
Commercially available reagents were used as supplied, unless otherwise stated.
Reactions requiring anhydrous conditions were performed in flame-dried glassware, and
cooled under an argon or nitrogen atmosphere. All reactions were carried out under
positive pressure of argon or nitrogen, and monitored by analytical thin layer
chromatography using glass-backed plates (5 × 10 cm) pre-coated with silica gel 60 F ,
2
54
as supplied by Merck; zones were detected visually under UV irradiation (254 nm) or by
spraying with phosphomolybdic acid reagent (Sigma-Aldrich, USA) followed by heating
to 80 ℃. Flash column chromatography was used routinely for purification and
separation of product mixtures using silica gel 60 of 230−400 mesh size as supplied by
1
13
Merck. H and C NMR spectra were recorded on Varian Mercury-300 or Varian
Mercury-400 (Fig. S3 supplementary data). Chloroform-d or dimethyl sulfoxide-d was
6
used as the solvent and TMS (δ 0.00 ppm) as an internal standard. Chemical shift values
are reported in ppm relative to the TMS in delta (δ) units. Multiplicities are recorded as s
(
singlet), brs (broad singlet), d (doublet), t (triplet), q (quartet), quint (quintet), sep
septet), dd (doublet of doublets), dt (doublet of triplets), and m (multiplet). Coupling
(
constants (J) are expressed in hertz. Electrospray mass spectra (ESMS) were recorded as
2
5

ACCEPTED MANUSCRIPT
m/z values using an Agilent 1100 MSD mass spectrometer. High-resolution mass spectra
were recorded under ESI-TOF mass spectroscopy conditions. All test compounds
displayed more than 95% purity as determined by a Hitachi 2000 series HPLC system
using a C-18 column (Agilent Eclipse XDB-C18 5 µm, 4.6 mm × 150 mm, USA). Mobile
phase A: acetonitrile; mobile phase B: 2 mM ammonium acetate aqueous solution
containing 0.1% formic acid. The gradient system started from A:B (10% : 90%) to A:B
(90% : 10%) with a flow rate of 0.5 mL/min and the injection volume was 20 ꢀL. The
system was operated at 25 °C. Peaks were detected at 254 nm. IUPAC nomenclature of
compounds was determined with ACD/Name Pro software.
4
.2. N-(4-aminophenethyl)-5-chloro-6-phenylfuro[2,3-d]pyrimidin-4-amine (2)
Compound (1) [30] (1 g, 3.77 mmol) and 4-(2-aminoethyl)aniline (512 mg, 3.77 mmol)
in EtOH (20 mL) were heated to reflux for 1 h. The reaction mixture was poured into
water, the precipitate was filtered and washed with water several times to afford 2 (1.18 g,
1
8
7
3
4
6%); H NMR (400 MHz, CDCl ) δ 8.29 (s, 1H), 8.00 (d, J = 7.2 Hz, 2H), 7.51 (t, J =
3
.2 Hz, 2H), 7.44 (t, J = 7.2 Hz, 1H), 7.05 (d, J = 8.4 Hz, 2H), 6.71 (d, J = 8.4 Hz, 2H),
+
.79 (t, J = 7.0 Hz, 2H), 2.88 (t, J = 7.0 Hz, 2H); LCMS (ESI) m/z: 365.2 [M+H] .
.3. General procedure for the synthesis of compounds 3a, 3c, 3d, 3i, 3r-3t
The general procedure is illustrated below with compound 3a as a specific example.
2
6

ACCEPTED MANUSCRIPT
1
-(4-(2-((5-chloro-6-phenylfuro[2,3-d]pyrimidin-4-yl)amino)ethyl)phenyl)-3-Phenyl)urea
(
3a)
A mixture of compound 2 (100 mg, 0.27 mmol) and phenyl isocyanate (0.059 ml, 0.54
mmol) in CH Cl (2 mL) was stirred under an argon atmosphere at room temperature.
2
2
After 4 h, the reaction mixture was concentrated in vacuo. The residue was purified by
silica gel column chromatography with MeOH/DCM (2:98) to give the title compound 3a
1
(
86 mg, 66%); H NMR (400 MHz, DMSO-d6) δ 8.63 (dd, J = 9.6, 3.6 Hz, 2H), 8.37 (s,
1
H), 7.98 (d, J = 8.3 Hz, 2H), 7.57 (t, J = 8.3 Hz, 2H), 7.48 (t, J = 8.3 Hz, 1H), 7.44 (d, J
8.4 Hz, 2H), 7.39 (d, J = 8.0 Hz, 2H), 7.29-7.25 (m, 3H), 7.19 (d, J = 8.4 Hz, 2H), 6.97
=
1
3
(
t, J = 8.0 Hz, 1H), 3.73 (q, J = 7.6 Hz, 2H), 2.87 (t, J = 7.6 Hz, 2H); C NMR (100
MHz, DMSO-d6) δ 163.46, 156.45, 154.91, 152.54, 144.08, 139.75, 137.89, 132.56,
1
4
29.35, 129.12, 139.00, 128.76, 127.61, 125.66, 121.74, 118.39, 118.14, 104.47, 100.56,
+
2.23, 34.20; LCMS (ESI) m/z: 484.2 [M+H] ; HRMS (ESI) calcd for C H ClN O
27
22
5
2
+
[
M+H] m/z 484.1540; found: 484.1546; HPLC purity = 99.7%, tR = 42.99 min.
.4. General procedure for the synthesis of compounds 3b, 3e-3h, 3j-3q, 3u-3y
The general procedure is illustrated below with compound 3b as a specific example.
-(4-(2-((5-chloro-6-phenylfuro[2,3-d]pyrimidin-4-yl)amino)ethyl)phenyl)-3-(o-tolyl)
urea (3b)
4
1
2
7

ACCEPTED MANUSCRIPT
To a solution of compound 2 (100 mg, 0.27 mmol) in 1,4-dioxane (2.0 mL) at room
temperature were added phenyl o-tolylcarbamate (246 mg, 1.08 mmol) and triethylamine
o
(
0.38 ml, 2.7 mmol). The mixture was stirred under argon atmosphere for 10 h at 120 C.
After the reaction was completed, the reaction mixture was poured into water and
extracted with ethyl acetate. The organic layer was washed with brine, dried over
anhydrous Na SO , and concentrated in vacuo. The residue was purified by silica gel
2
4
column chromatography with MeOH/DCM (2:98) to give the title compound 3b (47 mg,
1
3
2
1
6
5%); H NMR (400 MHz, DMSO-d6) δ 8.97 (s, 1H), 8.38 (s, 1H), 7.99 (d, J = 7.2 Hz,
H), 7.88 (s, 1H), 7.84 (d, J = 7.7 Hz, 1H), 7.57 (t, J = 7.2 Hz, 2H), 7.49 (t, J = 7.2 Hz,
H), 7.40 (d, J = 8.4 Hz, 2H), 7.28 (brs, 1H), 7.20-7.18 (m , 3H), 7.13 (t, J = 7.7 Hz, 1H),
.93 (t, J = 7.7 Hz, 1H), 3.75 (q, J = 7.3 Hz, 2H), 2.89 (t, J = 7.3 Hz, 2H), 2.30 (s, 3H);
C NMR (100 MHz, DMSO-d6) δ 163.44, 156.43, 154.90, 152.66, 144.05, 138.08,
37.47, 132.43, 130.17, 129.33, 129.10, 129.06, 127.62, 127.36, 126.14, 125.63, 122.56,
1
3
1
1
+
20.91, 118.21, 104.47, 100.56, 42.25, 34.20, 17.90; LCMS (ESI) m/z: 498.2 [M+H] ;
+
HRMS (ESI) calcd for C H ClN O [M+H] m/z 498.1697; found: 498.1711; HPLC
28
24
5
2
purity = 96.8%, tR = 43.71 min.
.4.1. 1-(4-(2-((5-chloro-6-phenylfuro[2,3-d]pyrimidin-4-yl)amino)ethyl)phenyl)-3-(2-
ethylphenyl)urea (3c)
4
2
8

ACCEPTED MANUSCRIPT
Following the similar reaction and workup procedures for 3a, the residue was
purified by silica gel column chromatography with MeOH/DCM (2:98) to give 3c. Yield:
1
4
2
1
6
7%; H NMR (400 MHz, DMSO-d6) δ 8.98 (s, 1H), 8.38 (s, 1H), 7.99 (d, J = 7.2 Hz,
H), 7.87 (s, 1H), 7.80 (d, J = 7.6 Hz, 1H), 7.57 (t, J = 7.2 Hz, 2H), 7.49 (t, J = 7.2 Hz,
H), 7.41 (d, J = 8.4 Hz, 2H), 7.28 (brs, 1H), 7.20-7.17 (m, 3H), 7.14 (t, J = 7.6 Hz, 1H),
.99 (t, J = 7.6 Hz, 1H), 3.75 (q, J = 7.4 Hz, 2H), 2.89 (t, J = 7.4 Hz, 2H), 2.60 (q, J = 7.6
1
3
Hz, 2H), 1.17 (t, J = 7.6 Hz, 3H); C NMR (100 MHz, DMSO-d6) δ 163.45, 156.43,
1
1
1
54.91, 152.87, 144.05, 138.11, 136.59, 133.61, 132.40, 129.34, 129.12, 129.06, 128.39,
27.62, 126.04, 125.64, 123.07, 121.94, 118.18, 104.47, 100.56, 42.26, 34.20, 23.80,
+
+
4.29; LCMS (ESI) m/z: 512.3 [M+H] ; HRMS (ESI) calcd for C H ClN O [M+Na]
2
9
26
5
2
m/z 534.1673; found: 534.1682; HPLC purity = 99.5%, tR = 45.37 min.
.4.2. 1-(4-(2-((5-chloro-6-phenylfuro[2,3-d]pyrimidin-4-yl)amino)ethyl)phenyl)-3-(2-
4
isopropylphenyl)urea (3d)
Following the similar reaction and workup procedures for 3a, the residue was
purified by silica gel column chromatography with MeOH/DCM (2:98) to give 3d. Yield:
1
3
2
1
4%; H NMR (400 MHz, DMSO-d6) δ 8.90 (s, 1H), 8.38 (s, 1H), 7.99 (d, J = 8.0 Hz,
H), 7.90 (s, 1H), 7.66 (d, J = 7.6 Hz, 1H), 7.57 (t, J = 8.0 Hz, 2H), 7.49 (t, J = 8.0 Hz,
H), 7.40 (d, J = 8.0 Hz, 2H), 7.28-7.26 (m, 2H), 7.19 (d, J = 8.0 Hz, 2H), 7.14 (t, J = 7.6
2
9