Identification, SAR studies, and X-ray Co-crystallographic analysis of a novel furanopyrimidine aurora kinase a inhibitor

Article abstract of DOI:10.1002/cmdc.200900339

Herein we reveal a simple method for the identification of novel Aurora kinase A inhibitors through substructure searching of an in-house compound library to select compounds for testing. A hydrazone fragment conferring Aurora kinase activity and heterocy

Full text of DOI:10.1002/cmdc.200900339

DOI: 10.1002/cmdc.200900339
Identification, SAR Studies, and X-ray Co-crystallographic
Analysis of a Novel Furanopyrimidine Aurora Kinase A
Inhibitor
Mohane Selvaraj Coumar,[a] Ming-Tsung Tsai,[a, b] Chang-Ying Chu,[a] Biing-Jiun Uang,[b] Wen-
Hsing Lin,[a] Chun-Yu Chang,[a] Teng-Yuan Chang,[a] Jiun-Shyang Leou,[a] Chi-Huang Teng,[a]
Jian-Sung Wu,[a] Ming-Yu Fang,[a] Chun-Hwa Chen,[a] John T.-A. Hsu,[a, c] Su-Ying Wu,[a]
Yu-Sheng Chao,[a] and Hsing-Pang Hsieh*[a]
Herein we reveal a simple method for the identification of
novel Aurora kinase A inhibitors through substructure search-
ing of an in-house compound library to select compounds for
testing. A hydrazone fragment conferring Aurora kinase activity
and heterocyclic rings most frequently reported in kinase in-
hibitors were used as substructure queries to filter the in-
house compound library collection prior to testing. Five new
series of Aurora kinase inhibitors were identified through this
strategy, with IC50 values ranging from ~300 nm to ~15 mm, by
testing only 133 compounds from a database of ~125000
compounds. Structure–activity relationship studies and X-ray
co-crystallographic analysis of the most potent compound, a
furanopyrimidine derivative with an IC50 value of 309 nm
toward Aurora kinase A, were carried out. The knowledge
gained through these studies could help in the future design
of potent Aurora kinase inhibitors.
Introduction
The identification of a lead molecule for any given molecular
target is an important step in drug discovery programs. Many
researchers traditionally use high-throughput screening (HTS)
of large compound libraries to identify a lead compound for
drug development. However, HTS of large libraries is a time-
and resource-consuming process, so for the work presented
herein, a faster and more economical knowledge-based ap-
proach was used to selectively screen an in-house HTS library
consisting of 125000 compounds.
used in the treatment of non-small-cell lung cancer (NSCLC).[4,5]
Because of the successful launch of eight small-molecule inhib-
itors that target protein kinases for various cancer therapies
(Figure 1), and given that many inhibitors are currently at vari-
ous stages of clinical trials, the development of protein kinase
inhibitors has taken center stage.
All protein kinases for which structural information is so far
available adopt the same fold, with N- and C-terminal lobes
connected through a short strand of amino acids, referred to
as the hinge region. ATP binds to the cleft formed between
the two lobes and forms hydrogen bonds with residues of the
hinge region.[6] Most of the kinase inhibitors that have been
developed, or are under development, are ATP-competitive in-
Protein kinases are second only to G-protein-coupled recep-
tors as the most sought-after drug targets, and account for
20–30% of all drug discovery efforts.[1] Currently more than
500 human protein kinases have been identified. Protein kinas-
es catalyze the transfer of a phosphate group from adenosine
triphosphate (ATP) to a tyrosine or serine/threonine residue of
the given substrate protein.[2,3] The phosphorylated substrate
protein is thus activated for the particular role it plays among
various cell-signaling processes; thus, protein kinases act as
switches for many cell-signaling processes that regulate cell
proliferation and differentiation. Small molecules designed to
act as ATP-competitive inhibitors could disrupt protein kinase
activity, thereby controlling the cell-signaling process, which is
aberrant in many disease states.[3] In particular, aberrant pro-
tein kinase activity is linked to the development, progression,
and prognosis of various cancers; this has prompted the devel-
opment of small-molecule inhibitors that target such aberrant
protein kinase activity. For example, Bcr-Abl activity is blocked
by Imatinib, which is used for the treatment of chronic mye-
logenous leukemia (CML) and gastrointestinal stromal tumors
(GISTs); EGFR kinase activity is inhibited by Gefitinib, which is
[a] Dr. M. S. Coumar,+ M.-T. Tsai,+ Dr. C.-Y. Chu, Dr. W.-H. Lin, C.-Y. Chang,
T.-Y. Chang, J.-S. Leou, Dr. C.-H. Teng, J.-S. Wu, M.-Y. Fang, C.-H. Chen,
Dr. J. T.-A. Hsu, Dr. S.-Y. Wu, Dr. Y.-S. Chao, Dr. H.-P. Hsieh
Division of Biotechnology and Pharmaceutical Research
National Health Research Institutes
35 Keyan Road, Zhunan, Miaoli County 350, Taiwan (Republic of China)
Fax: (+886)37-586-456
E-mail: hphsieh@nhri.org.tw
[b] M.-T. Tsai,+ Dr. B.-J. Uang
Department of Chemistry, National Tsing Hua University
101, Sect. 2, Guangfu Road, Hsinchu 300, Taiwan (Republic of China)
[c] Dr. J. T.-A. Hsu
Department of Biological Science and Technology
National Chiao Tung University
1001 University Road, Hsinchu 300, Taiwan (Republic of China)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.200900339.
ChemMedChem 2010, 5, 255 – 267
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
255

H.-P. Hsieh et al.
MED
macophore. All the inhibitors de-
veloped so far have a hinge
binding motif and establish at
least one hydrogen bond be-
tween the inhibitor and protein
kinase. Most of this hydrogen
bond acceptor and donor inter-
action is provided by the hetero-
cyclic portion of the inhibitors
(Figure 1).[7–10] In addition to the
ATP binding pocket, protein kin-
ases have other hydrophobic
and solvent-exposed hydrophilic
regions that are not occupied by
ATP. The size, shape, and amino
acid composition of these re-
gions differ markedly between
protein kinases, and interaction
between these regions and the
inhibitor determines the selectiv-
ity of the particular inhibitor for
the kinase in question.[7]
Figure 1. Eight kinase inhibitors marketed for the treatment of cancer: Imatinib (Novartis, targets Bcr-Abl and cKIT,
approved for CML and GISTs); Gefitinib (AstraZeneca, targets EGFR, approved for NSCLC); Erlotinib (Roche, targets
EGFR, approved for NSCLC); Lapatinib (GSK, targets EFGR and HER2, approved for breast cancer); Dasatinib (BMS,
targets Bcr-Abl and Src, approved for CML); Sunitinib (Pfizer, targets Raf, VEGFR, KIT, and FLT3, approved for renal
cell carcinoma (RCC) and GIST); Sorafenib (Bayer, targets Raf, VEGFR-2, VEGFR-3, PDGFR, cKIT, FLT3, and RET, ap-
proved for hepatocellular carcinoma (HCC)); and Nilotinib (Novartis, targets Bcr-Abl, approved for CML). Black dots
indicate the heterocyclic ring that interacts with the kinase hinge region through one or more hydrogen bonds,
as shown by X-ray co-crystallographic studies.
Aurora kinase isoforms A, B,
and C belong to the serine/
threonine subclass of protein
kinases and are key regulators of
mitosis required for genomic sta-
bility. The levels of both Aurora
hibitors in that they bind to the ATP binding pocket of the
kinase, mimicking the ATP–kinase interaction (Figure 2). As evi-
denced from X-ray co-crystal structure data, these kinase inhib-
itors have specific hydrogen bond acceptor and donor compo-
nents that interact with the hinge region formed between the
two terminal lobes of the kinase; they mimic the hydrogen
bonds formed between the kinase and the adenine ring of
ATP. This interaction between the kinase hinge region and the
inhibitors is essential and is part of the kinase inhibitor phar-
kinases A and B are upregulated in various cancers and have
been shown to promote oncogenic transformation in cell lines
in vitro.[11] The dysregulation of Aurora kinase activity has been
linked to genetic instability as well as defects in centrosome
function, spindle assembly, chromosome alignment, and cyto-
kinesis, all of which can lead to tumorigenesis.[11] Preclinical
proof-of-concept studies with VX-680,[12] the first Aurora kinase
inhibitor to enter clinical trials, showed tumor regression in
various animal models of cancer, validating Aurora kinase as a
drug target for cancer treatment. Currently more than ten ATP-
competitive inhibitors, including PHA-739358, SNS-314,
AZD1152, and AT-9283, are at different stages of clinical devel-
opment for the treatment of various cancers (Figure 3).[13–16]
Despite these positive advances, two of the Aurora kinase in-
hibitors, VX-680 and MLN8054, were recently dropped from
clinical trials. VX-680 was found to cause QT prolongation in
phase II clinical trials; this underscores the remaining need and
opportunities for the development of further lead compounds
as Aurora kinase inhibitors.
X-ray co-crystal structures of VX-680,[17] PHA-739358,[18] SNS-
314,[19] and AT-9283[20] in complex with Aurora kinase A reveal
the typical protein kinase–inhibitor hydrogen bonding pattern
between the inhibitor heterocycle and the kinase hinge region
(Figure 3). Based on our knowledge of the nature of the hinge
interaction, the type of heterocyclic systems most commonly
found in protein kinase inhibitors in general, and Aurora kinase
inhibitors in particular, we have started a program to selective-
ly screen, by rationally selecting compounds from our in-house
compound library, for Aurora kinase A inhibitory activity.
Figure 2. For clear visualization of the protein–ligand interaction, the full
structures of ATP and Gefitinib are abbreviated with R1 and R2 groups.
a) Protein kinase–ATP binding through hydrogen bond acceptor and donor
elements between the adenine ring of ATP and the kinase hinge region
amino acids; ATP binds in the cleft formed between the two lobes of the
kinase. b) Typical protein kinase–inhibitor (in this case, EGFR–Gefitinib) bind-
ing through hydrogen bond acceptor and donor elements; Gefitinib binds
in the cleft formed between the two lobes of the kinase, thereby blocking
ATP access to the catalytic site.
256
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 255 – 267

Furanopyrimidine Aurora Kinase A Inhibitors
Figure 3. Aurora kinase inhibitors at various stages of clinical trials. MK-0457/VX-680 (phase II discontinued, pan-selective: Aurora A,B,C); PHA-739358 (pha-
se II, pan-selective); SNS-314 (phase I, pan-selective); AZD1152 (phase II, Aurora B selective); MLN8054 (phase I discontinued, Aurora A selective); MLN8237
(phase I, Aurora A selective); CYC116 (phase I, pan-selective); PF-3814735 (phase I, Aurora A,B selective); GSK1070916 (phase I, Aurora B,C selective); AT-9283
(phase I/II, Aurora A,B); ENMD 2076 (phase I, Aurora A); R-763 (phase I); VX-689/MK-5108 (phase I). Black dots indicate the heterocyclic ring that interacts with
the kinase hinge region through one or more hydrogen bonds, as shown by X-ray co-crystallographic studies.
Herein we present our substructure search strategy to selec-
tively screen a compound library to identify several Aurora
kinase inhibitors, including a novel furanopyrimidine inhibitor,
compound 8. Structure–activity relationship (SAR) exploration
of the hit and X-ray co-crystallographic analysis of the hit in
complex with Aurora kinase A are also detailed.
Aurora kinases. To test this possibility, we searched our in-
house HTS library of ~125000 compounds for a hydrazone
fragment by using the substructure query build-in available in
ISIS/Base. A total of 1250 compounds containing a hydrazone
fragment were retrieved from the HTS library database. Visual
examination of all of them revealed that this could be further
divided into more than 15 subgroups (~500 compounds)
based on the heterocyclic fragment (R1) attached to the hydra-
zone functionality (Scheme 1). Compounds without heterocy-
clic components were omitted, as the kinase hinge binding
motif is essential for efficacy.
Results and Discussion
Substructure searching
In our effort to develop novel Aurora kinase inhibitors, we re-
cently reported the identification of pyrazole hydrazone 1 as
an Aurora kinase inhibitor through virtual screening, and fur-
ther reported our lead optimization efforts.[21] X-ray crystallo-
graphic studies of 1 and related analogues revealed that the
pyrazole ring N and NH groups, along with the hydrazone
linker NH group, form the essential hydrogen bond interaction
with the hinge region amino acids of Aurora kinase. We also
identified compound 2, which has a different heterocycle (tet-
razole) along with the hydrazone linker, that possesses Aurora
kinase inhibitory activity. As it is well known from the vast
amount of published data available for kinase inhibitors that
certain structural motifs are shared among many of the inhibi-
tors developed, we envisaged that structurally diverse ana-
logues incorporating a hydrazone functionality would bind
One or two compounds from each subgroup were initially
assayed for Aurora kinase inhibitory activity. The selected com-
pounds were representative of that subgroup, with the R2
group attached to the hydrazone/benzylidine portion contain-
ing either an indole, a dimethoxy- or monomethoxyphenyl
group if available, as these groups in the benzylidine part of
pyrazole series showed the best activity levels.[21] Compounds
were initially screened for Aurora kinase inhibition at two con-
centrations: 50 and 10 mm. Once compounds with inhibitory
activity were identified, all other compounds with the same
heterocyclic (R1) core structures were also tested for activity. A
total of 96 compounds were assessed, out of which four differ-
ent subgroup compounds possessed >30% Aurora kinase in-
hibition at 10 mm (table 1s, Supporting Information), and 3–6
are the most active compounds in each subgroup. Scheme 1
ChemMedChem 2010, 5, 255 – 267
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
257

H.-P. Hsieh et al.
MED
and develop novel kinase inhibi-
tors.[22–25] A recent publication re-
ports the analysis of kinase in-
hibitor structures (~20000 com-
pounds), and the authors identi-
fied
ments”, such as the bis-
arylanilines, as the most
“kinase-privileged
frag-
common structural feature of
protein kinase inhibitors.[9] They
also identified several heterocy-
clic ring systems such as pyri-
dines, pyrimidines, pyrazoles,
quinazolines, and indoles to be
present with greater frequency
in kinase inhibitor scaffolds than
in non-kinase-inhibiting small
molecules that all contain one or
more heteroatoms such as N, S,
or O.[9]
Accordingly, we filtered our in-
house compound library using
the following criteria: heterocy-
clic ring, 5- or 6-membered,
mono- or polycyclic containing
at least one N, O, or S atom
(Scheme 2, Substructure query-
1). With this filter we generated
Scheme 1. Knowledge-based rational screening strategy applied to the HTS compound library, leading to the
identification of new Aurora kinase A inhibitors 3–6 containing a hydrazone fragment. See the text for details of
the screening strategy.
gives a schematic overview of the knowledge-based screening
process and the most potent Aurora kinase inhibitors 3–6 with
a hydrazone fragment identified through this strategy.
a kinase-biased library of ~3000 compounds. Because this is
still a large collection of compounds for in vitro screening, we
set out to further narrow the field for the screening test. For
this, we used the knowledge gained from the newly identified
Aurora kinase inhibitors 1–6. We started with the general struc-
ture of hydrazone-fragment-containing compounds 1–6, in
Thienopyrimidine compound
3 (Aurora kinase A IC50 =
1.35 mm) showed the highest potency, which is at least an
order of magnitude more potent than the initial leads 1 and 2,
upon which the substructure search was executed.
The next most active compound is tetrahydroinda-
zole core compound 4, which is at least threefold
more active than 1. Compounds 5 and 6—with the
benzimidazole and pyrazole cores, respectively—pos-
sess ranges of activity similar to that of 1. The most
active compounds in each of the four series identi-
fied contain a 3-substitued indole ring in the hydra-
zone part as the R2 group, which was also found to
be the most preferred substituent in the pyrazole
series we investigated previously.[21]
With the successful identification of several new
Aurora kinase inhibitor lead compounds through a
simple substructure search protocol based on two
virtual screening leads 1 and 2, we were interested
to know if this strategy could be used to identify a
diverse set of compounds without a hydrazone frag-
ment. Several research groups have realized a com-
monality in protein kinase inhibitors and have found
general structural features most common in them;
they have used this information to develop various
Scheme 2. Knowledge-based rational screening strategy applied to the HTS compound
library, leading to the identification of novel furanopyrimidine Aurora kinase A inhibitors
expert systems (kinase-targeted libraries, virtual
screening, etc.) that could be employed to identify 7 and 8. See the text for details of the screening strategy.
258
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 255 – 267

Furanopyrimidine Aurora Kinase A Inhibitors
which two rings, A and B, are linked through the hydrazone
group; next we assumed that replacement of the hydrazone
linker with a linker of two to three methylene units would pro-
vide a similar type of molecular arrangement necessary for
binding to Aurora kinase. With this assumption, we searched
the kinase-biased library of ~3000 compounds using substruc-
ture query-2, in which ring B could be either 5- or 6-membered
aromatic or heteroaromatic. Based on these criteria, 62 com-
pounds were retrieved from the kinase-biased library. Visual in-
spection of all candidates allowed their division into 10 major
groups depending on the heterocyclic ring system (A) present.
A total of 21 compounds (figure 1s, Supporting Information)
were selected from this set and were submitted to the Aurora
kinase assay. This led to the identification of compound 7 as
an Aurora kinase A inhibitor, which has a furanopyrimidine
core and an IC50 value of 3.8 mm. With 7 thus identified as an
Aurora kinase A inhibitor, compounds with a furanopyrimidine
core were retrieved (16 compounds) from the kinase-biased li-
brary through substructure query-3 and tested for Aurora
kinase inhibition. Most of the tested compounds exhibited
varying levels of kinase inhibition (table 2s, Supporting Infor-
mation); compound 8 was the most potent, with an IC50 value
of 309 nm (Scheme 2).
ous reports,[26] suggest that the furanopyrimidine hit 8 is a se-
lective Aurora kinase A inhibitor over EGFR kinase, Chk1, PKA,
and CDK1, with additional inhibition at FLT3. This prompted us
to further explore the potential of the furanopyrimidine series
for drug development.
SAR study of furanopyrimidine 8
Structural optimization of the hit 8 was initiated to understand
the structure–activity relationship (SAR) in this novel series.
Compounds 8–24 were initially synthesized, and their ability to
inhibit Aurora kinase A at a concentration of 10 mm was deter-
mined; the results are listed in Table 1 and are expressed as
percentage inhibition. In addition, for those compounds show-
ing >90% inhibition, IC50 values were also determined. Initial
attempts were made to determine the importance of the two
phenyl rings on the furan ring. Removal of either phenyl ring
led to loss of activity for compounds 9 and 10. In particular,
the phenyl group at the 2-position of the furan ring is more es-
sential for maintaining activity than the 3-position phenyl
group, as compound 10 showed complete loss of activity (9:
76.6% inhibition at 50 mm; 10: 35.2% inhibition at 50 mm).
Moreover, replacement of these phenyl groups by a bromine
atom results in maintenance of activity, as shown by com-
pound 11, which retains the phenyl group at the 2-position of
the furan ring, whereas compound 12 is inactive. This clearly
demonstrates the importance of the phenyl ring at the 2-posi-
tion of the furan ring for interacting with Aurora kinase.
Having identified that the 3-position phenyl ring can be re-
Four of the five Aurora kinase inhibitors identified in this
study, compounds 3–6, have never been reported as kinase in-
hibitors. However, the most active compound 8 identified in
this study has been reported earlier as a weak inhibitor of
Chk1 kinase (IC50 ~26 mm).[26] It is well known that many kinase
inhibitors are active toward more than one kinase due to the
close similarity between ATP
binding pockets, as discussed
Table 1. Effect of replacing the phenyl groups of 8 on Aurora kinase A inhibition.
above. We were therefore inter-
ested to know whether the new
compounds identified in this
study are selective for Aurora
kinase inhibition. For this pur-
pose, we screened 3–6 and 8 at
a concentration of 10 mm for in-
hibition of two tyrosine kinases:
epidermal growth factor recep-
tor (EGFR) kinase and FMS-like
tyrosine kinase 3 (FLT3), and the
results are shown in table 3s of
the Supporting Information.
These compounds were found
to inhibit EGFR kinase by <15%;
however, they inhibited FLT3,
with 4 and 6 showing ~70% in-
hibition at 10 mm. FLT3 is inhibit-
ed by most Aurora kinase inhibi-
tors reported, due to the high
degree of similarity in their bind-
ing pockets.[16] Inhibition of FLT3
might be beneficial, as FLT3 is
also a target for acute myeloge-
nous leukemia (AML).[27] Our
present results, along with previ-
Compd
R1
R2
Inhibition [%][a,b]
IC50 [nm][b]
8
9
Ph
H
Ph
Br
Ph
CH3
C2H5
CꢀCH
Ph
Ph
H
Ph
Br
Ph
Ph
Ph
Ph
Ph
Ph
Ph
98.3
17.3
0.0
309
–
–
751
–
–
709
–
–
10
11
12
13
14
15
16
17
18
19
20
21
90.8
36.0
71.0
95.8
89.6
68.5
85.7
96.1
95.7
93.3
95.6
4-methoxyphenyl
4-nitrophenyl
4-hydroxyphenyl
4-acetamidophenyl
4-methoxyphenyl
4-hydroxyphenyl
–
245
272
639
159
4-methoxyphenyl
4-hydroxyphenyl
22
Ph
Ph
83.7
–
23
24
94.7
99.7
732
619
[a] Determined at a compound concentration of 10 mm. [b] Values are expressed as the mean of at least two in-
dependent determinations and are within ꢁ15%.
ChemMedChem 2010, 5, 255 – 267
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
259

H.-P. Hsieh et al.
MED
placed by a bromo group with only a 2.5-fold decrease in ac-
tivity, we further explored other substitutions at this position.
Introduction of alkyl groups such as methyl (compound 13)
and ethyl (compound 14), or an alkynyl group such as acety-
lene (compound 15), however, did not improve activity relative
to 8.
Aurora kinase isoforms B and C, as well as the structurally relat-
ed FLT3 kinase; the results are listed in tables 3s and 4s of the
Supporting Information. Compound 27 inhibited Aurora kina-
se A more strongly than either the B or C isoforms, and also
showed FLT3 kinase inhibition. As it has been established that
both Aurora kinase isoform-selective as well as isoform-nonse-
lective inhibitors possess anticancer activity in preclinical
animal models and are also at various stages in clinical trials,[16]
the new Aurora kinase inhibitors disclosed herein warrant fur-
ther study.
As only the phenyl group substitution at the 3-position of
the furan ring showed maximum activity, and replacement
with other groups had thus far led to loss of activity, explora-
tion of the effect of substitution on this phenyl ring was initiat-
ed. Thus, the introduction of a 4-methoxy (compound 16) or
4-nitro (compound 17) group led to a slight decrease in activi-
ty, whereas the introduction of a 4-hydroxy (compound 18) or
4-acetamido (compound 19) group resulted in retention of
Aurora kinase inhibition levels. We next introduced 4-methoxy
(compound 20) and 4-hydroxy (compound 21) groups in both
phenyl rings and found that compound 21, with two hydroxy
groups, has a twofold higher potency than unsubstituted com-
pound 8. However, replacement of either one or both phenyl
groups with a heterocyclic ring such as pyridine (compound
22) or furan (compounds 23 and 24) did not improve activity
relative to 8.
We next focused our efforts on the effect of replacing the
furan ring with thiophene (compound 25) or pyrrole (com-
pound 27) on Aurora kinase inhibition (Table 2). Replacing the
oxygen atom with sulfur decreased the activity of 25, while re-
placement with an NH group enhanced activity threefold for
27 relative to 8. However, the benzyl-protected pyrrole deriva-
tive 26 lost activity completely, demonstrating the crucial role
of the heteroatom in the bicyclic ring core structure in main-
taining Aurora kinase inhibition.
Synthesis
Compounds 8, 20, and 24, with symmetric substitution on the
furan ring, were synthesized from the appropriate benzoins
through modification of published procedures.[26,28,29] In brief,
cyclization of the appropriate benzoin compounds 28a–c with
malononitrile under basic conditions afforded the required
symmetrically substituted furans 29a–c, which were cyclized
and chlorinated to establish the bicyclic furo[2,3-d]pyrimidine
ring intermediates 31a–c. Nucleophilic reaction with 2-amino-
ethanol gave the desired compounds 8, 20, and 24. The 4-hy-
droxy derivative 21 was prepared from the 4-methoxy deriva-
tive 20 by demethylation using boron tribromide (Scheme 3).
The most potent Aurora kinase A inhibitor identified in this
study, compound 27, was evaluated for the ability to inhibit
Table 2. Effect of replacing the furan ring of 8 on Aurora kinase inhibi-
tion.
Scheme 3. Reagents and conditions: a) malononitrile, DMF, Et2NH, RT, 16 h,
40–67%; b) HCO2H, Ac2O, 08C!reflux, 16 h; c) POCl3, neat, 708C, 3 h, 41–
60% for two steps; d) 2-aminoethanol, nBuOH, MWI, 1758C, 1 h, 63–94%;
e) BBr3, CH2Cl2, 08C!RT, 2 h, 54%.
Compd
R
Inhibition [%][a,b]
98.3
IC50 [nm][b]
8
309
For the preparation of compounds 9 and 10 with a bromo
substituent at either the 2- or 3-position of the furan ring, a
modified reported procedure was used.[30,31] 2-Bromoacetophe-
none (32) or 2-hydroxyacetophenone (33) were condensed
with malononitrile, followed by cyclization to afford the furan
intermediates 34a,b. A reaction sequence similar to that for
the synthesis of compound 8 was then followed to construct
the bicyclic furanopyrimidine system to give the desired com-
pounds 9 and 10 (Scheme 4).
25
26
60.1
2.2
–
–
For the preparation of compounds 11–19, 22, and 23, with
various substitutions at the 3-position of the furan ring, the in-
termediates 36a,b were brominated using N-bromosuccini-
mide to give 37a,b, which could be converted into 11 and 12
by reaction with 2-aminoethanol. Attempts to carry out Suzuki
coupling of 37a with appropriately substituted boronic acids
27
100
104
[a] Determined at a compound concentration of 10 mm. [b] Values are ex-
pressed as the mean of at least two independent determinations and are
within ꢁ15%.
260
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 255 – 267

Furanopyrimidine Aurora Kinase A Inhibitors
For the preparation of thienopyrimidine compound 25, a
modified Gewald reaction was used.[33] Briefly, deoxybenzoin
39 was condensed with malononitrile and then cyclized in the
presence of sulfur to construct the thiophene compound 40,
which could be further elaborated to give 25, as in the case of
8 (Scheme 6).
Scheme 4. Reagents and conditions: a) malononitrile, DMF, Et2NH, RT, 16 h,
38–71%; b) HCO2H, Ac2O, 08C!reflux, 8 h; c) POCl3, neat, 708C, 3 h, 37–
75% for two steps; d) 2-aminoethanol, nBuOH, MWI, 1758C, 1 h, 64–80%.
failed; however, Suzuki coupling was carried out with amino al-
cohol 11 to give compounds 16–19, 22, and 23. An acetylene
group was introduced at the 3-position of the furan ring
through Sonagashira coupling of trimethylsilylacetylene with
11, followed by TMS deprotection to give acetylene 15. Reduc-
tion of the triple bond of 15 gave the ethyl derivative 14. For
the preparation of the methyl derivative 13, TBDMS-protected
38 was cross-coupled with methyl zinc chloride using Negishi
coupling[32] (Scheme 5).
Scheme 6. Reagents and conditions: a) CH2(CN)2, S8, Et3N, EtOH, reflux, 16 h,
64%; b) 1. HCO2H, Ac2O, 08C!reflux, 16 h, 2. POCl3, neat, 708C, 3 h, 20% for
two steps; c) 2-aminoethanol, nBuOH, reflux, 16 h, 23%.
For the preparation of pyrrolopyrimidine compounds 26 and
27, benzoin (28a) was condensed with benzylamine, followed
by cyclization with malononitrile, resulting in the construction
of pyrrole compound 42.[34] Cyclization with formic acid fol-
lowed by chlorination gave the bicyclic pyrrolopyrimidine com-
pound 44, which, upon reaction with 2-aminoethanol, afforded
26. As attempts to debenzylate this compound using alumi-
num trichloride failed, 44 was debenzylated to give 45, which
was treated with 2-aminoethanol to give the desired pyrrolo-
pyrimidine 27 (Scheme 7).
Scheme 7. Reagents and conditions: a) 1. benzylamine, HCl, PhCH3, reflux,
24 h, 72%, 2. CH2(CN)2, PhCH3, reflux, 24 h, 22%; b) HCO2H, reflux, 6 h, 62%;
c) POCl3, reflux, 3 h, 46%; d) 2-aminoethanol, nBuOH, reflux, 16 h, 25–48%;
e) AlCl3, PhCH3, reflux, 2 h, 39%.
Scheme 5. Reagents and conditions: a) NBS, DMF, RT, 3–18 h, 68–85%;
b) ArB(OH)2, Pd(OAc)2, PPh3, Na2CO3, dioxane, H2O, 1008C, 3 h, 42–80%; c) 2-
aminoethanol, nBuOH, MWI, 1758C, 1 h, 28–85%; d) TBDMSCl, imidazole,
DMF, RT, 1 h, 98%; e) 1. Pd(PPh3)4, CH3ZnCl, THF, reflux, 24 h, 52%, 2. TBAF,
CH2Cl2, RT, 16 h, 95%; f) 1. (CH3)3SiCꢀCH, Pd(PPh3)2Cl2, CuI, DIPEA, DMF, 668C,
16 h, 57%, 2. K2CO3, MeOH, RT, 16 h, 85%; g) PtO2, H2 (1 atm), EtOAc/MeOH,
RT, 1 h, 70%.
X-ray co-crystal analysis of the compound 8–Aurora kinase A
complex
Concurrent with the above synthetic explorations, we attempt-
ed to solve the structure of the co-crystal complex between
compound 8 and the Aurora kinase A protein. Co-crystalliza-
ChemMedChem 2010, 5, 255 – 267
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
261

H.-P. Hsieh et al.
MED
tion was carried out as detailed earlier by our research
group,[21] by using an Aurora kinase A construct and the com-
plex structure solved by X-ray crystallography with a resolution
of 2.35 ꢁ. The complex structure (Figure 4) reveals that the hit
were replaced with other hydrophobic groups, the inhibitory
activity was either impaired or retained, with the exception of
compound 21, which showed improved activity (Table 1). As
replacement of the furan ring with pyrrole (in 27) resulted in a
threefold improvement in activity relative to 8 (Table 2), the X-
ray co-crystal complex with compound 8 was used to gather
structural biology insight into the role played by the NH
group. The distance between the oxygen atom in the furan
ring of 8 and the oxygen atom in the main chain of the kinase
residue Ala213 is 3.15 ꢁ. It is possible that upon replacing the
oxygen with an NH group, it forms an additional hydrogen
bond with the hinge residue, and this may contribute to the
improved activity of 27 over that of 8. This can also explain
the inactivity of compound 26, as the bulky N-benzyl group
could effect an unfavorable steric clash between ligand and
protein. Notably, our attempts to co-crystallize compound 27
with Aurora kinase A in order to gain direct evidence for the
interaction between the NH group and the protein failed.
Conclusions
We used preexisting knowledge to select compounds for
screening for inhibitory activity toward Aurora kinase A. This
enabled us to identify the potent Aurora kinase A inhibitor
compound 8, with an IC50 value of 309 nm, by sampling only
133 compounds from a library of ~125000 compounds. Sub-
structure searching has been used by medicinal chemists to
identify analogues of known active core structures, but has not
been documented as a stand-alone tool in the identification of
a potent compound with activity in the nanomolar range, par-
ticularly in the kinase domain. Application of a similar substruc-
ture search strategy for inhibitors of other kinases is therefore
possible. For example, EGFR kinase inhibitors (e.g., Gefitinib, Er-
lotinib, and Lapatinib; Figure 1) are known to have an aniline
group attached to a hinge-binding heterocyclic core;[7,8,27] this
information can be exploited through an appropriate substruc-
ture search. The use of specific structural information gained
in-house, such as the hydrazone fragment information for
Aurora kinase inhibition in our case, could help in the identifi-
cation of novel inhibitors. Computer-aided virtual screening,
the basis of which is substructure/similarity searching, has
been shown to effectively decrease the number of compounds
sampled to identify potent lead compounds, relative to high-
throughput screening of a large library of compounds. Howev-
er, the use of virtual screening and related technologies de-
pends on the availability of resources such as specialized hard-
ware and software, and also on having personnel with special-
ized training in computational chemistry. The approach pre-
sented herein could be used by organic and medicinal chem-
ists who are involved in optimizing lead compounds in drug
discovery projects using simple, everyday chemistry tools such
as ISIS/Draw to identify potential hits more effectively, either
separately or in conjunction with other hit identification tech-
niques such as virtual screening.
Figure 4. X-ray co-crystal complex of compound 8 with Aurora kinase A
(PDB ID: 3K5U). Red lines represent hydrogen bonding interactions between
the inhibitor and protein, with distances indicated.
8 binds in the region formed between the C- and N-terminal
lobes of the protein and occupies the ATP binding pocket of
Aurora kinase A. The furanopyrimidine heterocyclic core forms
two essential hydrogen bonds with the hinge region amino
acid residues (Ala213 and Glu211) from N6 nitrogen (3.26 ꢁ)
and C7 hydrogen (3.29 ꢁ). In addition to these two hydrogen
bonds, another hydrogen bond is observed between the hy-
droxy group and the side chain of Lys162 (3.40 ꢁ). In addition
to the three hydrogen bonds anchoring the molecule to the
protein, extensive hydrophobic interactions with the surround-
ing amino acid residues was observed. The furanopyrimidine
core has hydrophobic interactions with Leu139, Val147,
Ala160, and Leu263. The alkyl chain bearing the hydroxy
group undergoes hydrophobic interactions with Leu210. In
particular, the two phenyl groups form strong hydrophobic
contacts with residues Leu139, Gly140, Val147, Gly216, and
Thr217. The X-ray co-crystal structure of 8 in complex with
Chk1 has already been reported. The main difference between
the Aurora and Chk1 structures is that the side chain hydroxy
group forms an intramolecular hydrogen bond with N8 of the
pyrimidine ring in Chk1 and lacks interaction with the Lys resi-
due.[26]
As apparent from the above description, the two phenyl
groups of compound 8 show extensive hydrophobic contacts.
These contacts explain why compounds 9 and 10, which lack
phenyl groups, are inactive, as such hydrophobic interactions
are absent in these cases. Moreover, when the phenyl groups
In summary, the novel Aurora kinase A inhibitors 3–8 identi-
fied in this study could be used as starting points for the de-
velopment of novel Aurora kinase inhibitors as anticancer
262
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 255 – 267

Furanopyrimidine Aurora Kinase A Inhibitors
with EtOAc; the organic layer was concentrated, and the crude
compound was purified by silica gel column chromatography
using a mixture of n-hexane/EtOAc (95:5) to give white solid 31a
(2.0 g, 63%). H NMR (300 MHz, CDCl3): d=8.77 (s, 1H), 7.61–7.58
(m, 2H), 7.52–7.46 (m, 5H), 7.35–7.32 (m, 3H); LC–MS (ESI) m/z:
307.0 [M+H]+.
agents. To aid in this process, the furanopyrimidine compound
8 was co-crystallized with Aurora kinase A, and the key ligand–
protein interactions were mapped. This structural biology in-
sight, along with the SAR information revealed herein, could
be used in the further design of potent Aurora kinase inhibi-
tors.
1
2-(5,6-Diphenylfuro[2,3-d]pyrimidin-4-ylamino)ethanol
8.
4-
Chloro-5,6-diphenylfuro[2,3-d]pyrimidine (31a, 100 mg, 0.33 mmol)
and 2-aminoethanol (33 mg, 0.54 mmol) in n-butanol (1 mL) were
sealed in a microwave tube and subjected to microwave irradiation
(MWI) at 1758C for 1 h. The reaction mixture was concentrated,
and the residue was purified by silica gel column chromatography
using a mixture of CH2Cl2/MeOH (40:1) to give 8 (80 mg, 74%).
1H NMR (300 MHz, CDCl3): d=8.41 (s, 1H), 7.58–7.48 (m, 7H), 7.30–
7.28 (m, 3H), 5.18 (br, 1H), 3.74 (t, J=4.5 Hz, 2H), 3.62–3.58 (m,
2H); 13C NMR (75 MHz, CDCl3): d=164.8 (C), 157.5 (C), 153.0 (C),
147.8 (C), 132.1 (C), 130.1 (CH), 129.9 (CH), 129.4 (CH), 129.0 (CH),
128.8 (CH), 126.6 (CH), 115.1 (C), 103.7 (C), 77.4 (CH), 63.1 (CH2),
44.9 (CH2); HRMS (EI): m/z [M]+ calcd for C20H17N3O2: 331.1321,
found 331.1320.
Experimental Section
General methods
All commercial chemicals and solvents were reagent grade and
were used without further treatment unless otherwise noted. All
reactions were carried out under an atmosphere of dry N2. Reac-
tions were monitored by TLC using Merck 60 F254 silica gel glass
backed plates (5ꢂ10 cm); zones were detected visually under UV
light (l 254 nm) or by spraying with phosphomolybdic acid re-
agent (Aldrich) followed by heating at 808C. Flash column chroma-
tography was done using silica gel (Merck Kieselgel 60, No. 9385,
1
230–400 mesh ASTM). H NMR spectra were obtained with a Varian
Mercury-300 or Varian Mercury-400 spectrometer operating at 300
and 400 MHz, respectively. Chemical shifts (d) are reported in parts
per million (ppm) relative to the solvent peak or (CH3)4Si. HRMS
data were measured with a Finnigan (MAT-95XL) electron impact
(EI) mass spectrometer. LC–MS data were measured on an Agilent
MSD-1100 electrospray ionization (ESI)-MS–MS system.
2-Amino-4,5-bis-(4-methoxyphenyl)furan-3-carbonitrile
29b.
Compound 29b was prepared in 40% yield from bis-4-methoxy-
1
benzoin (28b) similar to 29a. H NMR (300 MHz, CDCl3): d=7.37–
6.91 (m, 8H), 6.80 (br, 2H), 3.85 (s, 3H), 3.80 (s, 3H); LC–MS (ESI)
m/z: 321.1 [M+H]+.
4-Chloro-5,6-bis-(4-methoxyphenyl)furo[2,3-d]pyrimidine
31b.
Compound 31b was prepared in 41% yield over two steps from
29b, similar to 31a. 1H NMR (300 MHz, CDCl3): d=8.71 (s, 1H),
7.57–7.52 (m, 2H), 7.38–7.33 (m, 2H), 7.04–7.00 (m, 2H), 6.87–6.22
(m, 2H), 3.90 (s, 3H), 3.81 (s, 3H); HRMS (EI): m/z [M]+ calcd for
C20H15ClN2O3: 366.0771, found: 366.0758.
Substructure searching
The in-house HTS compound collection was purchased from
ChemDiversity, and the structure information was stored in Accord
Enterprise Client database 6.0.0. Using the in-build substructure
search option, compounds were retrieved as SDF (structure data
file) format, then imported into MDL ISIS/Base 2.5 database and
further analyzed before selecting compounds for testing.
2-[5,6-Bis-(4-methoxyphenyl)furo[2,3-d]pyrimidin-4-ylamino]e-
thanol 20. Compound 20 was prepared in 94% yield from 31b,
similar to 8. 1H NMR (300 MHz, CDCl3): d=8.35 (s, 1H), 7.49–7.44
(m, 2H), 7.41–7.37 (m, 2H), 7.08–7.03 (m, 2H), 6.83–6.79 (m, 2H),
5.14 (t, J=5.4 Hz, 1H), 4.16 (br, 1H), 3.90 (s, 3H), 3.80 (s, 3H), 3.74
(t, J=5.1 Hz, 2H), 3.57 (dt, J=5.4, 5.1 Hz, 2H); 13C NMR (75 MHz,
CDCl3): d=164.4 (C), 159.8 (C), 159.7 (CH), 157.7 (C), 153.0 (C),
147.3 (CH), 131.0 (CH), 127.7 (CH), 124.0 (CH), 122.0 (C), 115.0 (C),
113.9 (CH), 112.7 (C), 103.4 (C), 62.8 (CH2), 55.3 (CH3), 55.2 (CH3),
44.3 (CH2); HRMS (EI): m/z [M]+ calcd for C22H21N3O4: 391.1532,
found: 391.1536.
Chemistry
2-Amino-4,5-diphenylfuran-3-carbonitrile 29a. Diethylamine
(13.8 g) was added dropwise over a period of 30 min to a mixture
of benzoin (28a, 10.0 g, 47.17 mmol) and malononitrile (3.8 g,
57.58 mmol) in DMF (30 mL) at 08C (the reaction temperature
should not exceed 408C). After the resulting mixture was stirred at
room temperature for 16 h, H2O (100 mL) was added. The resulting
precipitate was filtered, washed with a sufficient amount of H2O,
then with n-hexane, and dried. The solid was recrystallized from
EtOH to provide a yellowish-brown solid product of 29a (6.0 g,
49%). 1H NMR (300 MHz, CDCl3): d=7.47–7.34 (m, 8H), 7.28–7.18
(m, 2H), 4.94 (br, 2H); LC–MS (ESI) m/z: 261.1 [M+H]+.
5’-Amino-[2,2’;3’,2’’]terfuran-4’-carbonitrile 29c. Compound 29c
was prepared in 67% yield from 28c, similar to 29a. 1H NMR
(300 MHz, CD3OD): d=7.58 (dd, J=1.8, 0.6 Hz, 1H), 7.54 (dd, J=
1.8, 0.6 Hz, 1H), 6.88 (dd, J=3.6, 0.6 Hz, 1H), 6.70 (dd, J=3.6,
0.6 Hz, 1H), 6.54 (dd, J=3.6, 1.8 Hz, 1H), 6.52 (dd, J=3.6, 1.8 Hz,
1H); LC–MS (ESI) m/z: 241.0 [M+H]+.
4-Chloro-5,6-diphenylfuro[2,3-d]pyrimidine (31a). A mixture of
29a (2.0 g, 7.69 mmol) and formic acid (24 mL) was cooled to 08C,
and acetic anhydride (24 mL) was added dropwise. The resulting
mixture was stirred for 1 h. The reaction mixture was then warmed
to 1008C and stirred for 16 h. The reaction mixture was cooled,
and H2O was added (40 mL). The precipitate was filtered and
washed thoroughly with H2O and n-hexane to give 30a (2.1 g,
4-Chloro-5,6-difuran-2-ylfuro[2,3-d]pyrimidine (31c). Compound
31c was prepared in 41% yield over two steps from 29c, similar to
31a. 1H NMR (300 MHz, CDCl3): d=8.76 (s, 1H), 7.65 (dd, J=1.8,
0.9 Hz, 1H), 7.57 (dd, J=1.8, 0.6 Hz, 1H), 6.94 (dd, J=3.6, 0.6 Hz,
1H), 6.76 (dd, J=3.6, 0.9 Hz, 1H), 6.61 (dd, J=3.6, 0.9 Hz, 1H), 6.55
(dd, J=3.6, 0.9 Hz, 1H); LC–MS (ESI) m/z: 287.1 [M+H]+.
1
95%). H NMR (300 MHz, CDCl3): d=7.94 (s, 1H), 7.56–7.52 (m, 4H),
2-(5,6-Difuran-2-ylfuro[2,3-d]pyrimidin-4-ylamino)ethanol
24.
7.46–7.43 (m, 3H), 7.32–7.28 (m, 3H), 7.22 (s, 1H); LC–MS (ESI) m/z:
289.1 [M+H]+. A mixture of 30a (3.0 g, 10.41 mmol) and POCl3
(30 mL) was heated at 708C for 3 h. The reaction mixture was
cooled in an ice bath and neutralized by the careful addition of a
saturated solution of NaHCO3. The resulting mixture was extracted
Compound 24 was prepared in 63% yield from 31c, similar to 8.
1H NMR (300 MHz, CD3OD): d=8.29 (s, 1H), 7.76 (dd, J=1.8, 0.6 Hz,
1H), 7.69 (dd, J=1.8, 0.6 Hz, 1H), 6.98 (dd, J=3.6, 0.6 Hz, 1H), 6.94
(dd, J=3.6, 0.6 Hz, 1H), 6.67 (dd, J=3.6, 1.8 Hz, 1H), 6.64 (dd, J=
3.6, 1.8 Hz, 1H), 3.80–3.69 (m, 4H); 13C NMR (150 MHz, CD3OD): d=
ChemMedChem 2010, 5, 255 – 267
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
263

H.-P. Hsieh et al.
MED
165.2 (C), 157.8 (C), 154.1 (CH), 144.8 (C), 144.7 (C), 143.5 (CH),
142.3 (CH), 139.6 (C), 112.4 (CH), 104.9 (C), 104.4 (C), 63.0 (CH2),
44.5 (CH2); HRMS (EI): m/z [M]+ calcd for C16H13N3O4: 311.0906,
found: 311.0924.
20 mL DMF. After the resulting mixture was stirred at room tem-
perature for 3 h, H2O (100 mL) was added. The resulting precipitate
was filtered and washed with H2O. The crude compound was puri-
fied by silica gel column chromatography using a mixture of
EtOAc/n-hexane (1:10), to provide 37a (1.14 g, 85%). 1H NMR
(300 MHz, CDCl3): d=8.75 (s, 1H), 8.20–8.16 (m, 2H), 7.55–7.51 (m,
3H); LC–MS (ESI) m/z: 308.9 [M+H]+, 310.9 [M+2+H]+.
2-[5,6-Bis-(4-hydroxyphenyl)furo[2,3-d]pyrimidin-4-ylamino]etha-
nol (21). BBr3 (0.36 mL, 1m solution in CH2Cl2) was added, while
stirring, to a solution of 20 (20 mg, 0.05 mmol) in CH2Cl2 (1 mL)
cooled to 08C. The reaction mixture was allowed to reach room
temperature while it was stirred for 2 h. H2O (5 mL) was added to
the reaction mixture, and the precipitated material was purified by
silica gel column chromatography using a mixture of CH2Cl2/MeOH
2-(5-Bromo-6-phenylfuro[2,3-d]pyrimidin-4-ylamino)ethanol 11.
Compound 11 was prepared in 85% yield from 37a, similar to 8.
1H NMR (300 MHz, CDCl3): d=8.36 (s, 1H), 8.08–8.05 (m, 2H), 7.52–
7.40 (m, 3H), 6.53 (br, 1H), 3.94–3.89 (m, 2H), 3.54 (br, 2H);
13C NMR (75 MHz, CDCl3): d=164.0 (C), 157.7 (C), 154.3 (CH), 147.0
(C), 129.5 (CH), 128.7 (CH), 128.3 (C), 126.6 (CH), 102.7 (C), 88.8 (C),
62.8 (CH2), 44.3 (CH2); HRMS (EI): m/z [M]+ calcd for C14H12BrN3O2:
333.0113, found: 333.0107.
1
(95:5) to give 21 (10 mg, 54%). H NMR (300 MHz, CD3OD): d=8.22
(s, 1H), 7.36–7.27 (m, 4H), 6.99–6.94 (m, 2H), 6.72–6.67 (m, 2H),
3.63–3.52 (m, 4H); 13C NMR (75 MHz, CD3OD): d=165.1 (C), 159.4
(C), 159.3 (C), 153.8 (CH), 148.9 (C), 132.2 (CH), 129.0 (CH), 123.7 (C),
122.1 (C), 117.6 (CH), 116.4 (CH), 114.2 (C), 104.7 (C), 61.3 (CH2), 44.0
(CH2); HRMS (EI): m/z [M]+ calcd for C20H17N3O4: 363.1219, found:
363.1230.
6-Bromo-4-chloro-5-phenylfuro[2,3-d]pyrimidine 37b. Com-
pound 37b was prepared in 68% yield from 36b, similar to 37a.
1H NMR (300 MHz, CDCl3): d=8.34 (s, 1H), 7.53–7.47 (m, 5H); HRMS
(EI): m/z [M]+ calcd for C12H6BrClN2O: 307.9352, found: 307.9368.
2-Amino-5-phenylfuran-3-carbonitrile 34a. Compound 34a was
prepared in 38% yield from 2-bromoacetophenone (32), similar to
2-(6-Bromo-5-phenylfuro[2,3-d]pyrimidin-4-ylamino)ethanol 12.
Compound 12 was prepared in 28% yield from 37b, similar to 8.
1H NMR (300 MHz, CD3OD): d=8.28 (s, 1H), 7.58–7.51 (m, 5H),
3.65–3.58 (m, 4H); 13C NMR (150 MHz, CD3OD): d=167.1 (C), 158.1
(C), 154.7 (CH), 131.1 (C), 130.5 (CHꢂ2), 130.2 (CH), 124.0 (C), 121.0
(C), 102.9 (C), 61.1 (CH2), 44.1 (CH2); HRMS (EI): m/z [M]+ calcd for
C17H14BrN5O: 333.0113, found: 333.0116.
1
29a. H NMR (300 MHz, CDCl3): d=7.50–7.47 (m, 2H), 7.39–7.33 (m,
2H), 7.27–7.25 (m, 1H), 6.54 (s, 1H), 4.86 (br, 2H); LC–MS (ESI) m/z:
185.0 [M+H]+.
4-Chloro-6-phenylfuro[2,3-d]pyrimidine 36a. Compound 36a was
prepared in 75% yield over two steps from 34a, similar to 31a.
1
Compound 35a: 80% yield from 34a. H NMR (300 MHz, CD3OD):
d=9.00 (br, 1H), 8.36 (br, 1H), 7.67 (d, J=7.5 Hz, 2H), 7.44–7.38
(m, 2H), 7.32–7.25 (m, 1H), 7.10 (s, 1H); LC–MS (ESI) m/z: 235.0
(5-Bromo-6-phenylfuro[2,3-d]pyrimidin-4-yl)-[2-(tert-butyldime-
thylsilanyloxy)ethyl]amine 38. Imidazole (0.51 g, 7.50 mmol) and
tert-butyldimethylsilyl chloride (TBDMSCl, 0.54 g, 3.60 mmol) were
added to a solution of 11 (1.0 g, 3.00 mmol) in DMF (20 mL) at
room temperature and stirred for 1 h. H2O was added to the reac-
tion mixture, which was extracted with CH2Cl2; the organic layer
was separated, washed with brine, and concentrated under
vacuum. The residue obtained was purified by silica gel column
chromatography using a mixture of CH2Cl2/MeOH (30:1) to give 38
1
[M+Na]+. Compound 36a: 94% yield from 35a. H NMR (300 MHz,
CDCl3): d=8.75 (s, 1H), 7.93–7.89 (m, 2H), 7.55–7.26 (m, 3H), 7.09
(s, 1H); LC–MS (ESI) m/z: 231.0 [M+H]+.
2-(6-Phenylfuro[2,3-d]pyrimidin-4-ylamino)ethanol 9. Compound
9 was prepared in 80% yield from 36a, similar to 8. 1H NMR
(300 MHz, CD3OD): d=8.21 (s, 1H), 7.82–7.78 (m, 2H), 7.47–7.36
(m, 3H), 7.19 (s, 1H), 3.81–3.77 (m, 2H), 3.72–3.68 (m, 2H); 13C NMR
(75 MHz, CD3OD): d=166.6 (C), 159.1 (C), 154.3 (CH), 153.3 (C),
130.7 (C), 130.0 (CHꢂ2), 125.5 (CH), 104.6 (C), 99.0 (CH), 61.7 (CH2),
44.6 (CH2); HRMS (EI): m/z [M]+ calcd for C14H13N3O2: 255.1008,
found: 255.0993.
1
(1.31 g, 98%). H NMR (400 MHz, CDCl3): d=8.38 (s, 1H), 8.10–8.07
(m, 2H), 7.51–7.42 (m, 3H), 6.60 (t, J=5.2 Hz, 1H), 3.89 (t, J=
5.2 Hz, 2H), 3.78 (q, J=5.2 Hz, 2H), 0.92 (s, 9H), 0.09 (s, 6H); HRMS
(EI): m/z [M]+ calcd for C20H26BrN3OSi: 447.0978, found: 447.0996.
2-(5-Methyl-6-phenylfuro[2,3-d]pyrimidin-4-ylamino)ethanol 13.
Pd(PPh3)4 (35 mg, 0.03 mmol) and CH3ZnCl (0.9 mL, 2m solution in
THF) were added to a solution of 38 (140 mg, 0.30 mmol) in dry
THF (5 mL) and held at reflux for 24 h. H2O was added to the reac-
tion mixture, which was extracted with EtOAc; the organic layer
was separated, washed with brine, and concentrated under
vacuum. The residue obtained was purified by silica gel column
chromatography using a mixture of n-hexane/EtOAc (4:1) to give
[2-(tert-butyldimethylsilanyloxy)ethyl]-(5-methyl-6-phenylfuro[2,3-
2-Amino-4-phenylfuran-3-carbonitrile 34b. Compound 34b was
prepared in 71% yield from 2-hydroxyacetophenone (33), similar
to 29a. H NMR (300 MHz, CDCl3): d=7.57–7.53 (m, 2H), 7.43–7.30
(m, 3H), 6.98 (s, 1H), 4.92 (br, 2H); LC–MS (ESI) m/z: 207.0
[M+Na]+.
1
4-Chloro-5-phenylfuro[2,3-d]pyrimidine 36b. Compound 36b
was prepared in 37% yield over two steps from 34b, similar to
1
31a. H NMR (300 MHz, CDCl3): d=8.81 (s, 1H), 7.77 (s, 1H), 7.56–
7.47 (m, 5H); HRMS (EI): m/z [M]+ calcd for C12H7ClN2O: 230.0247,
1
d]pyrimidin-4-yl)amine (60 mg, 52%). H NMR (300 MHz, CDCl3): d=
found: 230.0249.
8.37 (s, 1H), 7.73–7.69 (m, 2H), 7.49–7.34 (m, 3H), 5.75 (t, J=5.1 Hz,
1H), 3.87 (t, J=5.1 Hz, 2H), 3.75 (q, J=5.1 Hz, 2H), 2.57 (s, 3H),
0.93 (s, 9H), 0.10 (s, 6H); HRMS (EI): m/z [M]+ calcd for
C21H29N3O2Si: 383.2029, found: 383.2022. The above compound
(30 mg, 0.08 mmol) was dissolved in CH2Cl2 (2 mL), to which was
added tetra-n-butylammonium fluoride (TBAF, 0.16 mL, 1m solution
in CH2Cl2); this was stirred at room temperature for 16 h. H2O was
added to the reaction mixture, which was extracted with CH2Cl2;
the organic layer was separated, washed with brine, and concen-
trated under vacuum. The residue obtained was purified by silica
gel column chromatography using a mixture of acetone/n-hexane
2-(5-Phenylfuro[2,3-d]pyrimidin-4-ylamino)ethanol 10. Com-
pound 10 was prepared in 64% yield from 36b, similar to 8.
1H NMR (300 MHz, CD3OD): d=8.29 (s, 1H), 7.70 (s, 1H), 7.55–7.43
(m, 5H), 3.70–3.61 (m, 4H); 13C NMR (75 MHz, CD3OD): d=167.4
(C), 159.4 (C), 154.6 (CH), 139.4 (CH), 132.2 (C), 130.5 (CH), 129.6
(CHꢂ2), 122.6 (C), 101.3 (C), 61.2 (CH2), 44.1 (CH2); HRMS (EI): m/z
[M]+ calcd for C14H13N3O2: 255.1008, found: 255.1008.
5-Bromo-4-chloro-6-phenylfuro[2,3-d]pyrimidine 37a. N-Bromo-
succinimide (1.16 g, 6.50 mmol) was added portionwise to 4-
chloro-6-phenylfuro[2,3-d]pyrimidine (36a, 1.0 g, 4.33 mmol) in
1
(3:1) to give 13 (20 mg, 95%). H NMR (300 MHz, CD3OD): d=8.20
264
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 255 – 267

Furanopyrimidine Aurora Kinase A Inhibitors
(s, 1H), 7.70–7.67 (m, 2H), 7.50–7.36 (m, 3H), 3.81–3.70 (m, 4H),
2.56 (s, 3H); 13C NMR (75 MHz, CD3OD): d=165.9 (C), 159.8 (C),
154.1 (CH), 148.6 (C), 131.2 (C), 129.8 (CH), 129.5 (CH), 128.1 (CH),
111.1 (C), 105.3 (C), 61.7 (CH2), 44.4 (CH2), 11.2 (CH3); HRMS (EI): m/z
[M]+ calcd for C15H15N3O2: 269.1164, found: 269.1170.
8.1 Hz, 2H), 5.20 (t, J=5.4 Hz, 1H), 3.89 (s, 3H), 3.74 (t, J=5.4, Hz,
2H), 3.57 (q, J=5.4 Hz, 2H); 13C NMR (75 MHz, CDCl3): d=164.5 (C),
159.9 (C), 157.9 (C), 153.5 (CH), 147.0 (C), 120.8 (CH), 129.4 (C),
128.4 (CHꢂ2), 126.2 (CH), 123.7 (C), 115.1 (CH), 114.5 (C), 103.5 (C),
62.6 (CH2), 55.3 (CH3), 44.2 (CH2); HRMS (EI): m/z [M]+ calcd for
C21H19N3O3: 361.1426, found: 361.1412.
2-(5-Ethynyl-6-phenylfuro[2,3-d]pyrimidin-4-ylamino)ethanol 15.
Pd(PPh3)2Cl2 (53 mg, 0.08 mmol), CuI (14 mg, 0.08 mmol), N,N-diiso-
propylethylamine (DIPEA, 5 mL), and (CH3)3SiCꢀCH (0.7 mL,
5.1 mmol) were added to a solution of 11 (500 mg, 1.50 mmol) in
anhydrous DMF (5 mL) and heated at 658C for 16 h. H2O was
added to the reaction mixture, which was extracted with CH2Cl2;
the organic layer was separated, washed with brine, and concen-
trated under vacuum. The residue obtained was purified by silica
gel column chromatography using a mixture of n-hexane/EtOAc
(1:1) to give 2-(6-phenyl-5-trimethylsilanylethynylfuro[2,3-d]pyri-
2-[5-(4-Nitrophenyl)-6-phenylfuro[2,3-d]pyrimidin-4-ylamino]e-
thanol 17. Compound 17 was prepared in 80% yield from p-nitro-
benzene boronic acid and 11, similar to 16. 1H NMR (300 MHz,
CDCl3): d=8.44 (s, 1H), 8.42–8.37 (m, 2H), 7.73–7.68 (m, 2H), 7.47–
7.43 (m, 2H), 7.35–7.28 (m, 3H), 5.01 (t, J=4.8 Hz, 1H), 3.77 (t, J=
4.8 Hz, 2H), 3.64 (q, J=4.8 Hz, 2H), 3.18 (br, 1H); 13C NMR (75 MHz,
CDCl3): d=165.1 (C), 157.6 (C), 154.0 (CH), 148.0 (C), 147.9 (C),
139.4 (C), 130.9 (CH), 129.3 (CH), 128.8 (CH), 128.4 (CH), 126.8 (CH),
124.8 (CH), 112.9 (C), 102.5 (C), 62.1 (CH2), 43.9 (CH2); HRMS (EI):
m/z [M]+ calcd for C20H16N4O4: 376.1172, found: 376.1191.
1
midin-4-ylamino)ethanol (301 mg, 57%). H NMR (300 MHz, CDCl3):
d=8.35 (s, 1H), 8.21–8.18 (m, 2H), 7.50–7.38 (m, 3H), 6.53 (t, J=
5.4 Hz, 1H), 3.92 (t, J=4.5 Hz, 2H), 3.78 (q, J=5.4 Hz, 2H), 3.55 (br,
1H), 0.34 (s, 9H); HRMS (EI): m/z [M]+ calcd for C19H21N3O2Si:
351.1403, found: 351.1393. The above compound (37 mg,
0.11 mmol) was dissolved in MeOH (1 mL), to which K2CO3 (44 mg,
0.32 mmol) was added, and the mixture was stirred at room tem-
perature for 16 h. H2O was added to the reaction mixture, which
was extracted with CH2Cl2; the organic layer was separated,
washed with brine, and concentrated under vacuum. The residue
obtained was purified by silica gel column chromatography using
a mixture of n-hexane/EtOAc (1:2) to give 15 (25 mg, 85%).
1H NMR (400 MHz, CDCl3): d=8.37 (s, 1H), 8.21–8.18 (m, 2H), 7.51–
7.40 (m, 3H), 6.43 (t, J=4.4 Hz, 1H), 3.92 (t, J=4.8 Hz, 2H), 3.80 (td,
J=4.8, 4.4 Hz, 2H), 3.72 (s, 1H), 3.64 (br, 1H); 13C NMR (75 MHz,
CDCl3): d=164.2 (C), 158.0 (C), 154.5 (CH), 154.4 (C), 129.8 (C),
128.8 (CH), 128.7 (C), 125.7 (CH), 103.0 (C), 94.0 (C), 86.2 (C), 76.1
(C), 62.7 (CH2), 44.3 (CH2); HRMS (EI): m/z [M]+ calcd for C16H13N3O2:
279.1008, found: 279.1010.
2-[5-(4-Hydroxyphenyl)-6-phenylfuro[2,3-d]pyrimidin-4-ylami-
no]ethanol 18. Compound 18 was prepared in 94% yield from 16,
similar to 21. H NMR (300 MHz, CD3OD): d=8.26 (s, 1H), 7.52–7.48
(m, 2H), 7.32–7.26 (m, 5H), 6.99–6.95 (m, 2H), 3.62–3.54 (m, 5H);
13C NMR (75 MHz, CD3OD): d=165.5 (C), 159.6 (C), 159.2 (C), 154.5
(CH), 148.2 (C), 132.1 (CH), 130.8 (C), 129.6 (CHꢂ2), 127.3 (CH),
123.3 (C), 117.7 (CH), 116.8 (C), 104.6 (C), 61.2 (CH2), 44.0 (CH2);
HRMS (EI): m/z [M]+ calcd for C20H17N3O3: 347.1270, found:
347.1266.
1
2-[5-(4-Acetamidophenyl)-6-phenylfuro[2,3-d]pyrimidin-4-ylami-
no]ethanol 19. Compound 19 was prepared in 60% yield from p-
acetamidobenzene boronic acid and 11, similar to 16. 1H NMR
(400 MHz, [D6]DMSO): d=10.20 (br, 1H), 8.33 (s, 1H), 7.80–7.77 (m,
2H), 7.45–7.41 (m, 4H), 7.38–7.29 (m, 3H), 5.42 (t, J=5.2 Hz, 1H),
4.70 (t, J=5.2 Hz, 1H), 3.47–3.42 (m, 4H), 2.10 (s, 3H); 13C NMR
(75 MHz, [D6]DMSO): d=168.7 (C), 164.5 (C), 157.4 (C), 154.0 (CH),
145.7 (C), 139.9 (C), 130.1 (CH), 129.2 (C), 128.9 (CH), 128.7 (CH),
126.0 (CH), 125.4 (C), 119.7 (CH), 115.2 (C), 102.4 (C), 59.2 (CH2), 42.8
(CH2), 24.2 (CH3); HRMS (EI): m/z [M]+ calcd for C22H20N4O3:
388.1535, found: 388.1531.
2-(5-Ethyl-6-phenylfuro[2,3-d]pyrimidin-4-ylamino)ethanol
14.
PtO2 (2 mg) was added to a solution of 15 (17 mg, 0.06 mmol) in a
mixture of EtOAc/MeOH (1:1, 2 mL) and hydrogenated at atmos-
pheric pressure (using a balloon of H2 gas) for 1 h at room temper-
ature. The reaction mixture was filtered over Celite, the solvents
were removed under vacuum, and the residue obtained was puri-
fied by silica gel column chromatography using a mixture of n-
hexane/EtOAc (1:2) to give 14 (12 mg, 70%). 1H NMR (300 MHz,
CDCl3): d=8.34 (s, 1H), 7.69–7.65 (m, 2H), 7.49–7.35 (m, 3H), 5.66
(t, J=4.5 Hz, 1H), 4.09 (br, 1H), 3.92 (t, J=4.5 Hz, 2H), 3.81 (q, J=
4.5 Hz, 2H), 2.89 (q, J=4.5 Hz, 2H), 1.42 (t, J=4.5 Hz, 3H); 13C NMR
(75 MHz, CDCl3): d=165.3 (C), 158.0 (C), 153.4 (CH), 147.3 (C), 129.8
(C), 128.8 (CH), 128.5 (CH), 126.9 (CH), 115.5 (C), 103.7 (C), 62.6
(CH2), 44.3 (CH2), 18.6 (CH2), 15.4 (CH3); HRMS (EI): m/z [M]+ calcd
for C16H17N3O2: 283.1321, found: 283.1329.
2-(6-Phenyl-5-pyridin-3-ylfuro[2,3-d]pyrimidin-4-ylamino)ethanol
22. Compound 22 was prepared in 53% yield from 3-pyridyl bor-
1
onic acid and 11, similar to 16. H NMR (300 MHz, CDCl3): d=8.74–
8.70 (m, 2H), 8.38 (s, 1H), 7.87–7.83 (m, 1H), 7.50–7.42 (m, 3H),
7.30–7.26 (m, 3H), 5.07 (t, J=5.1 Hz, 1H), 4.07 (br, 1H), 3.75 (t, J=
5.1 Hz, 2H), 3.62 (q, J=4.5 Hz, 2H); 13C NMR (75 MHz, CDCl3): d=
165.0 (C), 157.7 (C), 153.9 (CH), 150.3 (C), 149.9 (CH), 148.1 (C),
137.5 (CH), 129.0 (CH), 128.7 (CH), 128.7 (C), 126.6 (CH), 124.2 (CH),
111.0 (C), 102.9 (C), 61.9 (CH2), 43.8 (CH2); HRMS (EI): m/z [M]+ calcd
for C19H16N2O2: 332.1273, found: 332.1266.
2-(5-Furan-2-yl-6-phenylfuro[2,3-d]pyrimidin-4-ylamino)ethanol
23. Compound 23 was prepared in 42% yield from 2-furanyl bor-
2-[5-(4-Methoxyphenyl)-6-phenylfuro[2,3-d]pyrimidin-4-ylami-
no]ethanol 16. p-Methoxybenzene boronic acid (105 mg,
0.46 mmol), Pd(OAc)2 (10 mg, 0.04 mmol), PPh3 (44 mg, 0.16 mmol),
and Na2CO3 (0.46 mL, 2m solution) were added to a solution of 11
(140 mg, 0.42 mmol) in dioxane/H2O (1:1, 4 mL) mixture under N2
and heated at 1008C for 3 h. After completion, the reaction mix-
ture was cooled to room temperature, H2O (10 mL) was added,
and the reaction was extracted with CH2Cl2 (3ꢂ20 mL). Combined
organic phases were dried over Na2SO4, concentrated under
vacuum, and the residue obtained was purified by silica gel
column chromatography using CH2Cl2/MeOH (30:1) to give 16
1
onic acid and 11, similar to 16. H NMR (300 MHz, CDCl3): d=8.38
(s, 1H), 7.70–7.67 (m, 2H), 7.62 (dd, J=1.5, 0.6 Hz, 1H), 7.41–7.36
(m, 3H), 6.56 (dd, J=3.3, 1.5 Hz, 1H), 6.54 (dd, J=3.3, 0.6 Hz, 1H),
6.43 (t, J=4.8 Hz, 1H), 3.87 (t, J=4.8 Hz, 2H), 3.74 (q, J=4.8 Hz,
2H); 13C NMR (75 MHz, CDCl3): d=165.2 (C), 157.8 (C), 153.9 (CH),
145.4 (C), 142.8 (CH), 129.4 (CH), 129.1 (CH), 128.6 (CH), 127.3 (CH),
112.1 (CH), 110.5 (CH), 105.2 (C), 102.1 (C), 62.9 (CH2), 44.4 (CH2);
HRMS (EI): m/z [M]+ calcd for C18H15N3O3: 321.1113, found:
321.1117.
2-Amino-4,5-diphenylthiophene-3-carbonitrile 40. A mixture of
deoxybenzoin (39, 558 mg, 2.90 mmol), malononitrile (198 mg,
3.00 mmol), powdered sulfur (96 mg, 3.00 mmol), and triethylamine
1
(81 mg, 53%). H NMR (300 MHz, CDCl3): d=8.33 (s, 1H), 7.52–7.49
(m, 2H), 7.38 (d, J=8.1 Hz, 2H), 7.26–7.24 (m, 3H), 7.04 (d, J=
ChemMedChem 2010, 5, 255 – 267
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
265

H.-P. Hsieh et al.
MED
(300 mg, 3.00 mmol) in EtOH were held at reflux for 16 h. The sol-
vents were removed under vacuum, H2O was added, the mixture
was acidified with concentrated HCl, and the solid residue ob-
tained was filtered. The crude product was purified by silica gel
column chromatography using n-hexane/EtOAc (3:1) to give 40
(500 mg, 64%). 1H NMR (300 MHz, CDCl3): d=7.44–7.35 (m, 10H);
LC–MS (ESI) m/z: 277.1 [M+H]+.
CDCl3): d=8.58 (s, 1H), 7.47–7.24 (m, 11H); LC–MS (ESI) m/z: 306.3
[M+H]+.
2-(5,6-Diphenyl-7H-pyrrolo[2,3-d]pyrimidin-4-ylamino)ethanol
27. 2-Aminoethanol (1 mL, 16.60 mmol) was added to a solution of
45 (60 mg, 0.20 mmol) in n-butanol (10 mL) and held at reflux for
16 h. The reaction mixture was evaporated, and the residue ob-
tained was purified by silica gel column chromatography using n-
hexane/EtOAc (3:1) to give 27 (31 mg, 48%) 1H NMR (300 MHz,
CD3OD): d=8.16 (s, 1H), 7.43–7.23 (m, 10H), 3.61–3.48 (m, 2H),
3.34–3.30 (m, 2H); LC–MS (ESI) m/z: 331.1 [M+H]+.
4-Chloro-5,6-diphenylthieno[2,3-d]pyrimidine 41. Compound 41
was prepared in 20% yield over two steps from 40, similar to 31a.
1H NMR (300 MHz, CDCl3): d=8.88 (s, 1H), 7.38–7.28 (m, 10H); LC–
MS (ESI) m/z: 323.2 [M+H]+.
2-(5,6-Diphenylthieno[2,3-d]pyrimidin-4-ylamino)ethanol
25.
Enzyme inhibition assay
Compound 25 was prepared in 23% yield from 41, similar to 8.
1H NMR (300 MHz, CDCl3): d=8.45 (s, 1H), 7.45–7.17 (m, 10H), 3.75
(bs, 1H), 3.66–3.63 (m, 2H), 3.51–3.46 (m, 2H); LC–MS (ESI) m/z:
348.1 [M+H]+.
Aurora kinase A inhibition assays[21,35] and EGFR kinase inhibition
assays[36] for the compounds were carried out as described by us
previously.
2-Amino-1-benzyl-4,5-diphenyl-1H-pyrrole-3-carbonitrile 42.
A
mixture of benzoin (1.18 g, 5.60 mmol), benzylamine (2.12 g,
19.80 mmol), and HCl (1 mL) in toluene was held at reflux for 24 h,
after which time the solvents were evaporated under vacuum. The
residue obtained was purified by silica gel column chromatography
using n-hexane/EtOAc (3:1) to give 2-benzylamino-1,2-diphenyl-
X-ray co-crystal structure determination
Compound 8 was co-crystallized with Aurora kinase A construct,
and the complex structure was solved in a manner similar to that
described by us before.[21]
1
ethanone (1.2 g, 72%). H NMR (300 MHz, CDCl3): d=8.00–7.43 (m,
15H), 6.31–6.29 (d, 1H), 4.24–4.09 (m, 2H); LC–MS (ESI) m/z: 302.1
[M+H]+. A mixture of the above compound (1.20 g, 4.0 mmol) and
malononitrile (316 mg, 4.8 mmol) in toluene was held at reflux for
24 h, and the solvents were then evaporated under vacuum. The
residue obtained was purified by silica gel column chromatography
using n-hexane/EtOAc (3:1) to give 42 (300 mg, 22%). 1H NMR
(300 MHz, CDCl3): d=7.41–7.08 (m, 15H), 4.93 (m, 2H); LC–MS (ESI)
m/z: 350.1 [M+H]+.
Acknowledgements
We thank the staff of beamline BL13B1 at the National Synchro-
tron Radiation Research Centre (NSRRC), Taiwan and SP12B2 at
SPring-8, Japan for technical assistance. We thank Mark Swofford
for help with the English editing. The authors acknowledge finan-
cial support from the National Science Council, Taiwan (Grant
Nos. NSC-95-2113M-400-001-MY3 for S.Y.W. and NSC-98-2119M-
400-001-MY3 for H.P.H.).
7-Benzyl-5,6-diphenyl-7H-pyrrolo[2,3-d]pyrimidin-4-ol 43. A solu-
tion of 42 (300 mg, 0.90 mmol) in formic acid (20 mL) was held at
reflux for 6 h. After cooling the reaction mixture, ice H2O (200 mL)
was added, and the precipitate was filtered, washed well with H2O,
1
Keywords: aurora kinase inhibitors
·
hit identification
·
·
and dried to give 43 (200 mg, 62%). H NMR (300 MHz, [D6]DMSO):
d=7.99 (s, 1H), 7.35–6.81 (m, 15H), 5.31 (s, 2H); LC–MS (ESI) m/z:
378.1 [M+H]+.
structural biology
substructure searches
·
structure–activity relationships
7-Benzyl-4-chloro-5,6-diphenyl-7H-pyrrolo[2,3-d]pyrimidine 44.
A solution of 43 (644 mg, 1.60 mmol) in POCl3 (15 mL) was held at
reflux for 3 h. After cooling the reaction mixture, ice H2O (100 mL)
was added, and the precipitate was collected by filtration and puri-
fied by silica gel column chromatography using n-hexane/EtOAc
(3:1) to give 44 (310 mg, 46%). 1H NMR (CDCl3): d=8.74 (s, 1H),
7.29–6.91 (m, 15H), 5.49 (m, 2H); LC–MS (ESI) m/z: 396.2 [M+H]+.
[1] P. Cohen, Nat. Rev. Drug Discovery 2002, 1, 309–315.
[2] G. Manning, D. B. Whyte, R. Martinez, T. Hunter, S. Sudarsanam, Science
2002, 298, 1912–1934.
[3] D. Fabbro, S. Ruetz, E. Buchdunger, S. W. Cowan-Jacob, G. Fendrich, J.
Liebetanz, J. Mestan, T. O’Reilly, P. Traxler, B. Chaudhuri, H. Fretz, J. Zim-
mermann, T. Meyer, G. Caravatti, P. Furet, P. W. Manley, Pharmacol. Ther.
2002, 93, 79–98.
[4] R. Capdeville, E. Buchdunger, J. Zimmermann, A. Matter, Nat. Rev. Drug
Discovery 2002, 1, 493–502.
[5] M. Ranson, W. Mansoor, G. Jayson, Expert Rev. Anticancer Ther. 2002, 2,
161–168.
[6] S. Cheek, K. Ginalski, H. Zhang, N. V. Grishin, BMC Struct. Biol. 2005, 5, 6.
[7] J. J. Liao, J. Med. Chem. 2007, 50, 409–424.
[8] M. E. Noble, J. A. Endicott, L. N. Johnson, Science 2004, 303, 1800–1805.
[9] A. M. Aronov, B. McClain, C. S. Moody, M. A. Murcko, J. Med. Chem.
2008, 51, 1214–1222.
2-(7-Benzyl-5,6-diphenyl-7H-pyrrolo[2,3-d]pyrimidin-4-ylami-
no)ethanol 26. 2-Aminoethanol (202 mg, 3.30 mmol) was added to
a solution of 44 (115 mg, 0.30 mmol) in n-butanol (15 mL), and
held at reflux for 16 h. The reaction mixture was evaporated, and
the residue obtained was purified by silica gel column chromatog-
raphy using n-hexane/EtOAc (3:1) to give 26 (30 mg, 25%).
1H NMR (300 MHz, CD3OD): d=8.24 (s, 1H), 7.31–7.14 (m, 15H),
5.38 (s, 2H), 3.65–3.54 (m, 2H), 3.32–3.30 (m, 2H); LC–MS (ESI) m/z:
421.1 [M+H]+.
[10] I. Akritopoulou-Zanze, P. J. Hajduk, Drug Discovery Today 2009, 14, 291–
297.
[11] J. Fu, M. Bian, Q. Jiang, C. Zhang, Mol. Cancer Res. 2007, 5, 1–10.
[12] E. A. Harrington, D. Bebbington, J. Moore, R. K. Rasmussen, A. O. Ajose-
Adeogun, T. Nakayama, J. A. Graham, C. Demur, T. Hercend, A. Diu-Her-
cend, M. Su, J. M. Golec, K. M. Miller, Nat. Med. 2004, 10, 262–267.
[13] O. Gautschi, J. Heighway, P. C. Mack, P. R. Purnell, P. N. Lara, Jr., D. R.
Gandara, Clin. Cancer Res. 2008, 14, 1639–1648.
4-Chloro-5,6-diphenyl-7H-pyrrolo[2,3-d]pyrimidine 45. A solution
of 44 (202 mg, 0.50 mmol) and AlCl3 (133 mg, 1.00 mmol) in tolu-
ene (10 mL) was held at reflux for 2 h. After cooling the reaction
mixture, ice H2O (50 mL) was added, and the separated solid was
1
collected by filtration to give 45 (60 mg, 39%). H NMR (300 MHz,
266
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 255 – 267

Furanopyrimidine Aurora Kinase A Inhibitors
[14] C. H. Cheung, M. S. Coumar, H. P. Hsieh, J. Y. Chang, Expert Opin. Invest.
Drugs 2009, 18, 379–398.
[26] N. Foloppe, L. M. Fisher, R. Howes, P. Kierstan, A. Potter, A. G. Robertson,
A. E. Surgenor, J. Med. Chem. 2005, 48, 4332–4345.
[15] M. S. Coumar, C. H. Cheung, J. Y. Chang, H. P. Hsieh, Expert Opin. Ther.
Pat. 2009, 19, 321–356.
[27] C. Ustun, D. L. DeRemer, A. P. Jillella, K. N. Bhalla, Expert Opin. Invest.
Drugs 2009, 18, 1445–1456.
[16] J. R. Pollard, M. Mortimore, J. Med. Chem. 2009, 52, 2629–2651.
[17] G. M. Cheetham, P. A. Charlton, J. M. Golec, J. R. Pollard, Cancer Lett.
2007, 251, 323–329.
[28] K. Gewald, Chem. Ber. 1966, 99, 1002.
[29] M. M. Ali, M. A. Zahran, Y. A. Ammar, Y. A. Mohamed, A. T. Seleim, Ind. J.
Hetero. Chem. 1995, 4, 191–194.
[18] D. Fancelli, J. Moll, M. Varasi, R. Bravo, R. Artico, D. Berta, S. Bindi, A. Ca-
meron, I. Candiani, P. Cappella, P. Carpinelli, W. Croci, B. Forte, M. L. Gior-
gini, J. Klapwijk, A. Marsiglio, E. Pesenti, M. Rocchetti, F. Roletto, D. Se-
verino, C. Soncini, P. Storici, R. Tonani, P. Zugnoni, P. Vianello, J. Med.
Chem. 2006, 49, 7247–7251.
[19] J. D. Oslob, M. J. Romanowski, D. A. Allen, S. Baskaran, M. Bui, R. A.
Elling, W. M. Flanagan, A. D. Fung, E. J. Hanan, S. Harris, S. A. Heumann,
U. Hoch, J. W. Jacobs, J. Lam, C. E. Lawrence, R. S. McDowell, M. A. Nan-
nini, W. Shen, J. A. Silverman, M. M. Sopko, B. T. Tangonan, J. Teague,
J. C. Yoburn, C. H. Yu, M. Zhong, K. M. Zimmerman, T. O’Brien, W. Lew,
Bioorg. Med. Chem. Lett. 2008, 18, 4880–4884.
[20] S. Howard, V. Berdini, J. A. Boulstridge, M. G. Carr, D. M. Cross, J. Curry,
L. A. Devine, T. R. Early, L. Fazal, A. L. Gill, M. Heathcote, S. Maman, J. E.
Matthews, R. L. McMenamin, E. F. Navarro, M. A. O’Brien, M. O’Reilly,
D. C. Rees, M. Reule, D. Tisi, G. Williams, M. Vinkovic, P. G. Wyatt, J. Med.
Chem. 2009, 52, 379–388.
[21] M. S. Coumar, J. S. Leou, P. Shukla, J. S. Wu, A. K. Dixit, W. H. Lin, C. Y.
Chang, T. W. Lien, U. K. Tan, C. H. Chen, J. T. Hsu, Y. S. Chao, S. Y. Wu,
H. P. Hsieh, J. Med. Chem. 2009, 52, 1050–1062.
[30] Y. Miyazaki, S. Matsunaga, J. Tang, Y. Maeda, M. Nakano, R. J. Philippe,
M. Shibahara, W. Liu, H. Sato, L. Wang, R. T. Nolte, Bioorg. Med. Chem.
Lett. 2005, 15, 2203–2207.
[31] E. F. DiMauro, J. Newcomb, J. J. Nunes, J. E. Bemis, C. Boucher, J. L. Bu-
chanan, W. H. Buckner, A. Cheng, T. Faust, F. Hsieh, X. Huang, J. H. Lee,
T. L. Marshall, M. W. Martin, D. C. McGowan, S. Schneider, S. M. Turci,
R. D. White, X. Zhu, Bioorg. Med. Chem. Lett. 2007, 17, 2305–2309.
[32] A. O. King, N. Okukado, E. Negishi, J. Chem. Soc. Chem. Commun. 1977,
683–684.
[33] K. Gewald, E. Schinke, H. Bçttcher, Chem. Ber. 1966, 99, 94–100.
[34] P. M. Traxler, P. Furet, H. Mett, E. Buchdunger, T. Meyer, N. Lydon, J. Med.
Chem. 1996, 39, 2285–2292.
[35] M. S. Coumar, J. S. Wu, J. S. Leou, U. K. Tan, C. Y. Chang, T. Y. Chang,
W. H. Lin, J. T. Hsu, Y. S. Chao, S. Y. Wu, H. P. Hsieh, Bioorg. Med. Chem.
Lett. 2008, 18, 1623–1627.
[36] W. H. Lin, J. S. Song, T. Y. Chang, C. Y. Chang, Y. N. Fu, C. L. Yeh, S. H. Wu,
Y. W. Huang, M. Y. Fang, T. W. Lien, H. P. Hsieh, Y. S. Chao, S. F. Huang,
S. F. Tsai, L. M. Wang, J. T. Hsu, Y. R. Chen, Anal. Biochem. 2008, 377, 89–
94.
[22] A. M. Aronov, G. W. Bemis, Proteins Struct. Funct. Bioinf. 2004, 57, 36–50.
[23] A. M. Aronov, M. A. Murcko, J. Med. Chem. 2004, 47, 5616–5619.
[24] J. F. Lowrie, R. K. Delisle, D. W. Hobbs, D. J. Diller, Comb. Chem. High
Throughput Screening 2004, 7, 495–510.
Received: August 17, 2009
Revised: November 24, 2009
[25] R. Gozalbes, L. Simon, N. Froloff, E. Sartori, C. Monteils, R. Baudelle, J.
Med. Chem. 2008, 51, 3124–3132.
Published online on December 28, 2009
ChemMedChem 2010, 5, 255 – 267
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
267