Fast-forwarding hit to lead: Aurora and epidermal growth factor receptor kinase inhibitor lead identification

Article abstract of DOI:10.1021/jm1000198

A focused library of furanopyrimidine (350 compounds) was rapidly synthesized in parallel reactors and in situ screened for Aurora and epidermal growth factor receptor (EGFR) kinase activity, leading to the identification of some interesting hits. On the basis of structural biology observations, the hit 1a was modified to better fit the back pocket, producing the potent Aurora inhibitor 3 with submicromolar antiproliferative activity in HCT-116 colon cancer cell line. On the basis of docking studies with EGFR hit 1s, introduction of acrylamide Michael acceptor group led to 8, which inhibited both the wild and mutant EGFR kinase and also showed antiproliferative activity in HCC827 lung cancer cell line. Furthermore, the X-ray cocrystal study of 3 and 8 in complex with Aurora and EGFR, respectively, confirmed their hypothesized binding modes. Library construction, in situ screening, and structure-based drug design (SBDD) strategy described here could be applied for the lead identification of other kinases.

Full text of DOI:10.1021/jm1000198

4980 J. Med. Chem. 2010, 53, 4980–4988
DOI: 10.1021/jm1000198
Fast-Forwarding Hit to Lead: Aurora and Epidermal Growth Factor Receptor Kinase Inhibitor
Lead Identification†
Mohane Selvaraj Coumar,‡,# Chang-Ying Chu,‡,# Cheng-Wei Lin,‡,§ Hui-Yi Shiao,‡ Yun-Lung Ho,‡
Randheer Reddy,‡ Wen-Hsing Lin,‡ Chun-Hwa Chen,‡ Yi-Hui Peng,‡ Jiun-Shyang Leou,‡ Tzu-Wen Lien,‡ Chin-Ting Huang,‡
Ming-Yu Fang,‡ Szu-Huei Wu,‡ Jian-Sung Wu,‡ Santhosh Kumar Chittimalla,‡ Jen-Shin Song,‡
John T.-A. Hsu,‡, Su-Ying Wu,‡ Chun-Chen Liao,§,^ Yu-Sheng Chao,‡ and Hsing-Pang Hsieh*,‡
‡Division of Biotechnology and Pharmaceutical Research, National Health Research Institutes, 35 Keyan Road, Zhunan, Miaoli County 350,
Taiwan, ROC, §Department of Chemistry, National Tsing Hua University, 101 Guangfu Road Sec. 2, Hsinchu 300, Taiwan, ROC, Department of
Biological Science and Technology, National Chiao Tung University, 1001 University Road, Hsinchu 300, Taiwan, ROC, and ^Department of
Chemistry, Chung Yuan Christian University, 200 Zhongbei Road, Zhongli 320, Taiwan, ROC. # These authors contributed equally to this work.
Received January 7, 2010
A focused libraryof furanopyrimidine (350 compounds) wasrapidly synthesizedin parallel reactorsand in
situ screened for Aurora and epidermal growth factor receptor (EGFR) kinase activity, leading to the
identification of some interesting hits. On the basis of structural biology observations, the hit 1a was
modified to better fit the back pocket, producing the potent Aurora inhibitor 3 with submicromolar
antiproliferative activity in HCT-116 colon cancer cell line. On the basis of docking studies with EGFR hit
1s, introduction of acrylamide Michael acceptor group led to 8, which inhibited both the wild and mutant
EGFR kinase and alsoshowedantiproliferativeactivityin HCC827lung cancercellline. Furthermore, the
X-ray cocrystal study of 3 and 8 in complex with Aurora and EGFR, respectively, confirmed their
hypothesized binding modes. Library construction, in situ screening, and structure-based drug design
(SBDD) strategy described here could be applied for the lead identification of other kinases.
Introduction
screening of new compounds, targeting the inhibition of
Aurora and epidermal growth factor receptor (EGFRa)
kinase as an example. Considering the modular nature of
the synthesis andscreening method, similar protocolscouldbe
applied to other protein kinase drug discovery programs.
Protein kinases are enzymes that catalyze the transfer of
phosphate groups from adenosine triphosphate (ATP) to
sereine, threonine, or tyrosine residues of predetermined
target proteins. Phosphorylation of proteins is an important
step in cellular signal transduction, regulating cell differentia-
tion, cell growth, and cell migration and influencing apoptotic
mechanisms. Deregulation of protein kinase is implicated in a
number of diseases, including cancer, diabetes, and inflam-
mation. Thus, targeted inhibition of the deregulated protein
kinase has become an attractive therapeutic strategy in cancer
therapy.3 Aurora kinases A, B, and C, members of sereine/
threonine kinase, are key mitotic regulators involved in
maintaining the genomic integrity of daughter cells. Because
overexpression of Aurora A and Aurora B is frequently
associated with tumorigenesis, these molecules have been
targeted for cancer therapy.4,5 EGFR, a tyrosine kinase, is
overexpressed and also mutated in non-small-cell lung cancer
(NSCLC), which leads to aberrant EGFR signaling. Since
lung cancer is the leading cause of death due to cancer and
80% of them are NSCLC, targeting EGFR has become a
major treatment strategy for NSCLC.6,7
Drug discovery is a complex, risky, costly, and time-
consuming process, with estimates for the discovery and
development of a drug exceeding 10 years and approximately
U.S. $1 billion.1 Researchers continually seeks methods and
technologies to discover drugs more quickly and efficiently. In
this context, application of focused library synthesis to speed
up both lead identification and optimization is of proven
value. However, several caveats for the approaches described
so far exist, such as the difficulty of both standardization
procedures for solid-phase synthesis and the purification step
for solution-phase synthesis. Alternatively solution-phase
synthesis followed by in situ screening without the need for
isolation of the product has successfully identified potent
enzyme inhibitors.2 These recent methods rely on the identi-
fication and application of reactions suitable for carrying out
in water-miscible solvent, as after completion of the reaction,
the reaction mixture is diluted in water and screened for
activity. We here describe our strategies to speed up kinase
drug lead identification applied to the design, synthesis, and
†The atomic coordinates and crystal structure have been deposited in
the Protein Data Bank as entry 3M11.
*To whom correspondence should be addressed. Phone: þ886-37-
246-166, extension 35708. Fax: þ886-37-586-456. E-mail: hphsieh@
nhri.org.tw.
aAbbreviations: ATP, adenosine triphosphate; DM, double mutant;
EGFR, epidermal growth factor receptor; LCMS, liquid chromato-
graphy coupled with mass spectrometry; NSCLC, non-small-cell lung
cancer; PDB ID, Protein Data Bank identification number; SBDD,
structure-based drug design; WT, wild type.
Results and Discussion
Through substructure searching for kinase-privileged frag-
ments in our in-house compound library, we found that
r
pubs.acs.org/jmc
Published on Web 06/15/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13 4981
Table 1. Optimization of Screening Protocol for Library Synthesis
a Values are expressed as the mean of at least two independent determinations and are within (15%.
Scheme 1. Construction of 350 Furanopyrimidine Library
Compounds Based on Aurora Hit 1aa
nucleophilic reaction of various amines with 2 to give the
target compound 1 should proceed smoothly and to comple-
tion; (b) Aurora kinase A enzyme assay should be specific to
the target compounds 1 and should not be influenced by the
presence of either the starting material 2 or the amines used.
To test the feasibility of such a library synthesis and in situ
screening for activity without purification, we first ran a trial
reaction using representative primary amines, secondary
amines, aniline derivative, and heterocyclic amines to assess
the completion of the nucleophilic reaction. For this purpose,
5 mg of the chloro compound 2, 4 equiv of the nucleophile,
and 1 mL of n-butanol were combined in a 5 mL screw-capped
vial and heated at 80 ꢀC for 16 h in a parallel reactor.
Completion of reaction was assessed by the amount of
remaining chloro compound 2 and the formation of desired
product using LCMS. It was found that the reactions with
both primary amines (1a,b) and the secondary amine (1c) were
completed after 16 h of reaction. However, in the case of
aniline derivative (1d), the reaction was only partially com-
pleted and the starting material 2 was observed in LCMS even
after prolonged heating for 24 h. In contrast to this, amino
acid (1e) and heterocyclic amines (1f) did not react to give the
desired products (Table 1s, Supporting Information). These
results showed that only primary and secondary amines were
suitable for use as nucleophiles for library synthesis and that
for heterocyclic and aniline derivatives alternative synthetic
strategies would be essential.
To optimize the screening protocol of the library com-
pounds and to verify potential interference of the excess
amine, five compounds (1a-c,g,h) were synthesized in the
parallel reactor as mentioned above. After 16 h of reaction,
n-butanol was removed from the reaction mixture under va-
cuum and the residue obtained was diluted with DMSO to
make 10 mM stock samples. The product concentration was
considered the same as that of compound 2 because all of the
starting material was consumed in the reaction as assessed by
LCMS. The five compounds were tested at 10 μM for Aurora
A inhibition and showed similar levels of activity as those of
the pure compounds, suggesting that assay results were not
distorted by the presence of excess amine (Table 1). Moreover
a Amines used are shown in Chart 1s of Supporting Information.
compound 1a (Table 1), which has a furanopyrimidine scaf-
fold, possesses nanomolar (IC50 = 309 nM) Aurora kinase A
inhibition.8 To identify compounds with improved potency,
we began modification of the hit 1a. Because preliminary
investigationrevealed thatboththe phenylrings were essential
for activity,8 modification of the side chain was carried out. In
general, compound 1 was synthesized by nucleophilic sub-
stitution reaction of 2 with various amines in n-butanol
(Scheme 1). The necessary chloro compound, 2, was synthe-
sized by a modified procedure reported by Floppe et al.8,9
Since individual synthesis of the target compound followed by
purification would be a time-consuming task, we envisaged
synthesizing compounds in a parallel reactor and testing them
for Aurora kinase inhibition without purification in a combi-
natorial fashion. Once compounds with a certain level of
activity were identified in this primary screen, they would be
resynthesized individually, purified, and retested.
Microtiter-plate-based reaction followed by in situ bio-
chemical estimation of enzyme inhibition without the neces-
sity of isolating the products has been successfully utilized to
identify enzyme inhibitors.2 The key aspect for the successful
use of this methodology is the selection of a high-yielding
reaction, such as acid amine coupling or 1,3-dipolar addition
reaction.2,10,11 Similarly, we envisaged synthesis and screening
protocol, using nucleophilic substitution reaction of amines
with 4-chlorofuranopyrimidine 2. The reaction was carried
out in standard glass vials at millimole scale. Two key factors
were considered essential for such a work hypothesis: (a) the

4982 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13
Coumar et al.
Figure 1. Percentage inhibition of Aurora A (at 10 μM, blue color)
and EGFR (at 20 μM, red color) kinase by furanopyrimidine library
compounds 1.
Chart 1. Aurora Kinase A Inhibitors 1i-r and EGFR Kinase
Inhibitor 1s Hits Identified from Library Synthesis with IC50
1 μM
<
Figure 2. X-ray cocrystal structure of Aurora A protein in complex
with inhibitor 1a (PDB ID 3K5U). Inhibitor 1a binds to Aurora A
using two hydrogen bonds to the hinge region and another hydro-
gen bond to Lys162. The region commonly referred to as the back
pocket (circled yellow) is unoccupied.
Scheme 2. Modification of Library Hit 1j to Give Analogues
3-5a
a (a) Phenyl isocyanate, CH2Cl2, room temp, 8 h for 3; triethylamine,
benzoyl chloride, CH2Cl2, room temp, 8 h for 4; triethylamine, sulfonyl
chloride, CH2Cl2, room temp, 8 h for 5.
the starting material 2 was found to have 0% inhibition at
10 μM and ∼20% inhibition at 50 μM, further supporting the
library design and screening protocol.
Having found the right condition for library synthesis and
screening protocols, we synthesized around 350 compounds
using a variety of primary and secondary amines (Chart 1s,
Supporting Information) in the parallel reactor with a maxi-
mum capacity of 70 reactions in a single run. To confirm the
integrity of the reactions between different runs, compounds
1a and 1b were included as internal control in different
batches. Screening the compounds as mentioned at 10 μM
showed that 46% of the library compounds possessed greater
than 50% inhibition at 10 μM and around 20 compounds
showed over 90% Aurora inhibition (Figure 1). All 20 com-
pounds were resynthesized, purified, and tested for Aurora
inhibition; among them 10 compounds 1i-r (Chart 1) showed
submicromolar enzyme inhibition.
Compound 1a has been cocrystallized in complex with
Aurora A and the structure solved by X-ray (Figure 2; PDB
ID 3K5U),8 revealing that the hit binds to the ATP binding
pocket with the furanopyrimidine core forming two essential
hydrogen bonds with the hinge region. In addition, another
hydrogen bond has been observed between the OH group
present in the side chain and the Lys172 residue.8 The two
phenyl groups form extensive hydrophobic interaction with
the surrounding amino acid residues. Subsequently, the X-ray
cocrystal structure of 1i (data not shown) in complex with
Aurora A was also solved, which revealed that the hinge-
binding interaction is similar to that of 1a, while the carbonyl
group of the ester function interacts with the Lys162. Struc-
tures 1a and 1i both occupy the ATP binding site, commonly
referred to inkinase literature as the front cleft/pocket,12 while
the portion commonly referred to as the back cleft/pocket12
is unoccupied, similar to the situation with ATP binding.
The back pocket, which is also referred to as an allosteric site13
or a deep pocket,14 has been shown to be occupied by several
kinase inhibitors, and exploiting this interaction has been
used to enhance the potency and selectivity of the inhi-
bitors to specific kinases.12-14 On the basis of the similarity
in the binding of 1a and 1i in Aurora A, we assumed that
other potent compounds identified from library synthesis
could have a similar binding mode in Aurora A and that all
of them have an unoccupied back pocket. Moreover, the side
chains of 1a and 1i are directed toward the entrance of the
back pocket.
On the basis of this structural biology information and the
fact that the hits 1i,j possess diverse side chains amenable for
further structural modification, we contemplated introduc-
tion of functional groups, such as amide, urea, and sulfon-
amide, common functional groups found among kinase
inhibitors that could exploit back-pocket binding.13,14 With
this aim, the aniline functional group of 1j was converted into
urea, amide, and sulfonamide functionality by acylation with
phenyl isocyanate, benzoyl chloride, or sulfonyl chloride to
give compounds 3-5 (Scheme 2). Similarly, the ester function

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13 4983
Scheme 3. Modification of Library Hit 1i to Give Analogue 6a
a (a) LiOH, MeOH/H2O (4:1), reflux, 2 h; (b) HOBt, EDC, DMF,
aniline, room temp, 16 h.
Table 2. Identification of Potent Aurora Kinase A Inhibitor 3 with
Antiproliferative Activity in HCT-116
Figure 3. X-ray cocrystal structure of Aurora A protein in complex
with inhibitor 3 (PDB ID 3M11). Inhibitor 3 binds to Aurora A
using two hydrogen bonds to the hinge region and another hydro-
gen bond to Lys162. The region commonly referred to as back
pocket (circled yellow) is occupied by the terminal phenylurea
group.
The urea carbonyl group forms a hydrogen bond with Lys162
of Aurora A, and the terminal phenyl ring of 3 forms hydro-
phobic contacts with the surrounding residues in the back
pocket of Aurora A.
Aurora based inhibition,a IC50 (nM)
Concurrent with the above Aurora inhibitor development,
the librarywas screened for EGFR kinase inhibition at20μM.
Most of the compounds were poor inhibitors of EGFR
kinase, particularly those that are potent Aurora kinase
inhibitors (Figure 1). Three structurally related compounds
showed >80% inhibition at 20 μM. Resynthesis identified
compound 1s (Chart 1) as a potent EGFR kinase inhibitor
with IC50 ≈ 200 nM, without Aurora kinase inhibition.
Structurally 1s is an S-isomer and, when compared to the
Aurora hit 1a, has an additional phenyl group in the side
chain. The novel EGFR hit 1s was evaluated for its ability to
inhibit the double mutant (DM, L858R/T790M) EGFR
kinase activity, which is resistant to gefitinib (clinically ap-
proved treatment for NSCLC), and found to be inactive
(IC50 >10 μM; Table 3). Identification of compounds that
can overcome gefitinib resistant mutant EGFR kinase is
therapeutically relevant, as a large number of patients who
are initially responsive to gefitinib become resistant over the
course of treatment due to emergence of T790M gatekeeper
mutation.7 To aid in this process, we docked 1s into the active
site of the EGFR protein (Figure 4; PDB ID 1M17) cocrystal
complex with erlotinib (another clinically approved treatment
for NSCLC). The docking results suggest that the furanopyr-
imidine ring oxygen forms the required hinge binding inter-
action with the Met769 hinge residue, while the hydroxyl
group is within hydrogen bonding distance to Lys721 residue.
The three phenyl rings form extensive hydrophobic interac-
tion with the surrounding residues. In addition to this infor-
mation, modeling results suggest that the phenyl group at the
3-position of the furan ring is suitably positioned to anchor a
Michael acceptor group, which could be helpful in improving
the activity through covalent modification of the Cys773
residue. It is known that interaction between a suitably
positioned Michael acceptor group and the Cys773 residue
of the EGFR could improve the EGFR inhibition and also
impart activity toward Gefitinib resistant T790M mutant
kinase through irreversible covalent bond formation.7,16,17
compd
Aurora kinase A
HCT-116
1i
1j
3
796
640
>10,000
>10,000
400
43
304
4
>10000
>10000
>10000
120
5
5831
>10000
20
6
tozasertib15
a Values are expressed as the mean of at least two independent
determinations and are within (15%. Tozasertib (VX-680):15 Aurora
kinase inhibitor tested in phase II clinical trials.
of 1i was hydrolyzed to the carboxylic acid 10, which was
coupled with aniline to give the amide 6 (Scheme 3).
Compound 3 with the urea functional group showed potent
Aurora kinase A inhibition with IC50 ≈ 50 nM. Most im-
portantly, 3 inhibited the growth of the HCT-116 coloncancer
cell line in vitro with IC50 ≈ 400 nM (Table 2). Moreover,
HCT-116 cells treated with 3 showed the hallmarks of Aurora
kinase B inhibition, as evidenced by the DNA content analysis
using flow cytometry, which showed a higher percentage of
cells with 4N and 8N DNA content (polyploidy) due to
failure in cytokinesis. Western blot analysis showed lower
levels of posphorylated Aurora A (at Thr288) and histone
(H3 at Ser10) proteins, which are Aurora A and B down-
stream substrates, respectively (Figures 1s and 2s, Supporting
Information). These results clearly show that 3 inhibits both
Aurora A and B inside the cell. However, the other analogues,
4-6, showed poor Aurora A inhibition and did not inhibit
HCT-116 cell growth.
Next, the X-ray cocrystal structure of 3-Aurora A complex
was solved in order to verify if 3 indeed binds to Aurora kinase
A as hypothesized. The complex structure shows that 3 binds to
Aurora A in a similar way as the initial hit 1a except that the
back pocket of Aurora A is occupied by the terminal phenyl-
urea moiety of 3 (Figure 3; PDB ID 3M11), as hypothesized.

4984 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13
Coumar et al.
Table 3. Identification of Potent Double Mutant EGFR Kinase
Inhibitor 8
Scheme 4. Synthesis of Furanopyrimidine Analogues 7 and 8a
EGFR based inhibition,a IC50 nM
EGFR kinase
compd
WT
DM
HCC827
1s
7
223
82
>10000
>10000
28b
518
394
8b
12b
8b
7b
a (a) (S)-(þ)-Phenylglycinol, n-butanol, 80 ꢀC, 16 h; (b) ArB(OH)2,
Pd(dppf)2Cl2, Na2CO3, dioxane/H2O, 95 ꢀC, 16 h; (c) H2, Pd/C, EtOH,
atm, 16 h; (i) acryloyl chloride, pyridine, Et2O, room temp, 16 h.
8
gefitinib18
20b
11b
2815
15b
9 (BIBW-2992)19
a Values are expressed as the mean of at least two independent
determinations and are within (15%. b Values are expressed as the
mean of at least two independent determinations and are within (25%.
WT: wild type EGFR kinase. DM: double mutant (L858R/T790M)
EGFR kinase. Gefitinib: reversible EGFR inhibitor used in clinics for
the treatment of non-small-cell lung cancer (NSCLC). Compound 9:
irreversible EGFR inhibitor in phase III clinical testing for NSCLC.
EGFR kinase inhibition with development potential for
gefitinib-resistant EGFR overexpressing tumors.
We further confirmed that the designed-in Michael accep-
tor group of 8 could covalently interact with the Cys773
residue by solving the X-ray cocrystal complex of 8 with
EGFR kinase (Figure 3s, Supporting Information). A clear
covalent linkage between the Michael acceptor group of 8 and
Cys773 residue sulfur atom of EGFR protein was observed,
which would result in irreversible inhibition of EGFR kinase,
explaining the potent activity of 8, compared to the hit 1s.
Furthermore, the molecule was also anchored by a hydrogen
bond interaction between the furan ring oxygen and Met769
of hinge region. (However, it is noted that some parts of the
ligand (phenyl group) were not clear in the density map,
indicating that they might be flexible in the EGFR pocket,
precluding deposition of the coordinates to PDB.)
Conclusion
In summary, through rapid construction of a furanopyr-
imidine kinase-targeted library and screening it without iso-
lation of the compounds, several synthetically tractable and
structurally distinct leads for Aurora and EGFR kinase
inhibition were quickly identified. Potent hits identified from
the library were further modified based on structure biology/
docking studies in the kinase. On the basis of the structural
insight that the region commonly referred to as back pocket in
Aurora kinase A is unoccupied by Aurora hit 1a, functional
groups that could interact in the back pocket were introduced.
This drug-design strategy led to the identification of potent
Aurora kinase A inhibitor 3, which has antiproliferative
activity in the HCT-116 colon cancer cell line. Docking study
of the EGFR hit 1s from the library revealed possible en-
hancement of EGFR kinase activity through the introduction
of Michael acceptor functionality. On the basis of this finding,
introduction of acrylamide Michael acceptor group led to the
synthesis of 8, which showed potent activity in both wild and
mutant EGFR kinase, with potent antiproliferative activity in
HCC827 lung cancer cell line. X-ray cocrystal complex of 3
with Aurora kinase A showed thatthe back pocket is occupied
by the inhibitor 3, while X-ray cocrystal complex of 8 with
EGFR kinase showed the presence of a covalent bond be-
tween the Michael acceptor region of the inhibitor and Csy733
residue of EGFR, as predicted. Thus, library construction,
in situ screening, and structure-based drug design provided a
Figure 4. Docking studies of 1s in active site of EGFR kinase (PDB
ID 1M17) cocrystal complex with erlotinib. The 3-phenyl group on
the furan ring is close to the Cys773 residue.
On the basis of insights from molecular modeling of 1s, we
introduced an acrylamide Michael acceptor group into the
phenyl ring (on the 3-position of the furan ring) of hit 1s to
give compounds 7 and 8 (Scheme 4). It was found that
introduction of acrylamide group at the para position of 1s
improved the WT EGFR kinase inhibition around 3-fold for
7; however, it was inactive toward DM EGFR kinase (IC50
>
10μM; Table3). Moving the Michael acceptor groupfrom the
para to the meta position improved the WT EGFR kinase
activity dramatically: over 30-fold for 8, compared to 1s.
Moreover, the presence of the Michael acceptor group ren-
dered8 activetowardthe DM EGFR kinasewithanIC50 of28
nM, similar to 9, which is a second-generation EGFR kinase
inhibitor under development for gefitinib-resistant NSCLC.
Compound 8 also inhibited the cell growth of HCC827, an
NSCLC cell line in the range of single-digit nanomoles. These
results suggest that 8 is a novel lead compound with potent

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13 4985
tool to identify compounds with the same scaffold but dis-
playing distinct biological target/activities. Both 3 and 8 are
potential leads that are development candidates in our drug
discovery program. In addition, the library compounds could
be subjected to screening to identify suitable hits to act as
starting points for a lead-optimization program for other
therapeutically relevant kinases. Moreover, such a library-
construction and screening approach could be applied to
different core structures known to possess kinase hinge region
binding, thereby allowing effective exploration of kinome
chemical space.
(5,6-Diphenylfuro[2,3-d]pyrimidin-4-yl)-[2-(1H-imidazol-4-yl)-
ethyl]amine 1h. 1H NMR (300 MHz, CDCl3 þ CD3OD): δ 8.30
(s, 1H), 7.81 (s, 1H), 7.46-7.39 (m, 5 H), 7.36-7.33 (m, 2H),
7.21-7.19 (m, 3H), 6.61 (s, 1H), 3.65-3.60 (t, J = 6.6 Hz, 2H),
2.77-2.73 (t, J = 6.6 Hz, 2H). LCMS (ESI) m/z 382.4 (M þ H)þ.
3-(5,6-Diphenylfuro[2,3-d]pyrimidin-4-ylamino)propionic Acid
Ethyl Ester 1i. 1H NMR (300 MHz, CDCl3): δ 8.41 (s, 1H),
7.56-7.44 (m, 8H), 7.28-7.24 (m, 2H), 5.16-5.13 (t, J = 5.4 Hz,
1H), 4.11-4.04 (q, J = 7.2 Hz, 2H), 3.76-3.70 (q, J = 6.0 Hz,
2H), 2.57-2.53 (t, J = 6.0 Hz, 2H), 1.23-1.18 (t, J = 7.2 Hz, 3H).
HRMS (EI): calcd for C23H21N3O3 387.1583, found 387.1580.
[2-(4-Aminophenyl)ethyl]-(5,6-diphenylfuro[2,3-d]pyrimidin-4-yl)-
amine 1j. 1H NMR (300 MHz, CDCl3): δ 8.43 (s, 1H), 7.50-7.23
(m, 10H), 6.79 (d, J = 8.4 Hz, 2H), 6.59 (d, J = 8.4 Hz, 2H), 4.68 (t,
J= 5.2 Hz, 1H), 3.65 (q, J= 6.4 Hz, 2H), 3.62 (bs, 2H), 2.67 (t, J=
6.4 Hz, 2H). 13C NMR (100 MHz, CDCl3): δ 164.9 (C), 157.5 (C),
154.2 (CH), 146.5 (C), 144.8 (C), 132.3 (C), 129.6 (CH), 129.5 (CH),
129.4 (CH), 128.7 (CH), 128.4 (CH), 128.3 (CH), 126.3 (CH), 115.4
(CH), 115.0 (C), 103.1 (C), 42.1 (CH2), 34.2 (CH2). HRMS (EI)
calcd. for C26H22N4O 406.1794, found 406.1789.
S-2-(5,6-Diphenylfuro[2,3-d]pyrimidin-4-ylamino)propan-1-ol
1k. 1H NMR (300 MHz, CDCl3): δ 8.40 (s, 1H), 7.61-7.49 (m, 8
H), 7.32-7.27 (m, 2H), 4.86 (bs, 1H), 3.72-3.68 (m, 1H), 3.44-
3.38 (m, 1H), 1.11 (d, J = 6.3 Hz, 3H). HRMS (EI) calcd for
C21H19N3O2 345.1477, found 345.1474.
2-[2-(5,6-Diphenylfuro[2,3-d]pyrimidin-4-ylamino)ethoxy]ethanol
1l. 1H NMR (300 MHz, CDCl3):δ8.41 (s, 1H), 7.59-7.50 (m, 8H),
7.27-7.24 (m, 2H), 5.06 (bs, 1H), 3.69-3.59 (m, 4H), 3.54 (t, J =
5.1 Hz, 2H), 3.44 (t, J = 4.5 Hz, 2H), 1.77 (bt, J = 6.0 Hz, 1H).
HRMS (EI) calcd for C22H21N3O3 375.1583, found 375.1575.
(5,6-Diphenylfuro[2,3-d]pyrimidin-4-yl)-(2-pyridin-2-ylethyl)-
amine 1m. 1H NMR (300 MHz, CDCl3): δ 8.42 (s, 1H), 8.32 (d,
J = 3.9 Hz, 1H), 7.57-7.52 (m, 1H), 7.48-7.32 (m, 8H),
7.25-7.22 (m, 2H), 7.12-7.08 (m, 1H), 7.02 (d, J = 7.5 Hz,
1H), 5.29 (bt, J = 6.6 Hz, 1H), 3.90 (q, J = 6.0 Hz, 2H), 2.98 (t,
J = 6.3 Hz, 2H). LCMS (ESI) m/z 393.1 (M þ H)þ. HPLC
purity, 94.7%.
Experimental Section
General Methods. All commercial chemicals and solvents are
reagent grade and were used without further treatment unless
otherwise noted. All reactions were carried out under an atmo-
sphere of dry nitrogen. Reactions were monitored by TLC using
Merck 60 F254 silica gel glass backed plates (5 cm ꢀ 10 cm); zones
were detected visually under ultraviolet irradiation (254 nm) or
by spraying with phosphomolybdic acid reagent (Aldrich) fol-
lowed by heating at 80 ꢀC. Flash column chromatography was
done using silica gel (Merck Kieselgel 60, no. 9385, 230-400
mesh ASTM). 1H and 13C NMR spectra were obtained with a
Varian Mercury-300 or Varian Mercury-400 spectrometer.
Chemical shifts were recorded in parts per million (ppm, δ)
and were reported relative to the solvent peak or TMS. High-
resolution mass spectra (HRMS) were measured with a Finni-
gan (MAT-95XL) electron impact (EI) mass spectrometer or
using Finnigan/Thermo Quest MAT 95XL FAB mass spectro-
meter. LCMS data were measured on an Agilent MSD-1100
ESI-MS/MS system. Parallel synthesis was carried out in par-
allel block reactors (KEM Scientific Inc.), and the solvents were
evaporated in a Savant SpeedVac Plus system. The purity of the
final compounds was determined using a Hitachi 2000 series
HPLC system using a C-18 column (Agilent ZORBAX Eclipse
XDB-C18 5 μm, 4.6 mm ꢀ 150 mm) and was found to be >95%
unless otherwise stated.
Parallel Synthesis Protocol. The chloro compound 2 (5 mg)
was taken into a 5 mL screw-capped vial, and 4 equiv of the
nucleophile/amine and n-butanol (1 mL) were added. The
mixture was heated at 80 ꢀC for 16 h in a parallel block reactor.
After reaction completion, n-butanol was removed from the
reaction mixture under vacuum, and the residue obtained was
diluted with DMSO to make 10 mM stock samples and tested at
10 μM for Aurora A inhibition and at 20 μM for EGFR
inhibition.
Synthesis of Pure Furanopyrimidine Analogues 1b,c,g-s. Pure
compounds 1b,c,g-s were prepared from 4-chloro-5,6-diphenyl-
furo[2,3-d]pyrimidine (2) and appropriate amines in 70-90%
yield in a manner similar to that of 1a reported by us.8
(2,5-Dimethoxy-2,5-dihydrofuran-2-ylmethyl)-(5,6-diphenylfuro-
1
[2,3-d]pyrimidin-4-yl)amine 1n. H NMR (400 MHz, CDCl3): δ
8.40 (s, 1H), 7.56-7.48 (m, 7H), 7.28-7.23 (m, 3H), 7.00 (dd, J =
6.0, 1.2 Hz, 1H), 5.81 (dd, J = 6.0, 0.8 Hz, 1H), 5.67 (dd, J = 1.2,
0.8 Hz, 1H), 5.07 (bt, J = 5.6 Hz, 1H), 3.85 (dd, J = 13.6, 6.4 Hz,
1H), 3.78 (dd, J = 13.6, 5.2 Hz, 1H), 3.22 (s, 3H), 3.04 (s, 3H).
HRMS (EI) calcd for C25H23N3O4 429.1689, found 429.1686.
HPLC purity, 94.9%.
4-(3,5-Dimethylpiperazin-1-yl)-5,6-diphenylfuro[2,3-d]pyrimidine
1
1o. H NMR (400 MHz, CDCl3): δ 8.48 (s, 1H), 7.49-7.40 (m,
8H), 7.28-7.24 (m, 2H), 3.73-3.69 (m, 2H), 2.54-2.47 (m, 2H),
2.22 (dd, J = 12.4, 10.8 Hz, 2H), 0.81 (d, J = 6.4 Hz, 6H). 13
C
NMR (100 MHz, CDCl3): δ 166.9 (C), 160.6 (C), 152.5 (CH),
147.6 (C), 133.5 (C), 130.3 (CH), 129.6 (C), 129.2 (CH), 128.5
(CH), 128.3 (CH), 128.2 (CH), 127.1 (CH), 115.7 (C), 105.6 (C),
55.5 (CH2), 50.1 (CH), 19.2 (CH3). HRMS (EI) calcd for
C24H24N4O 384.1950, found 384.1943.
4-[2-(5,6-Diphenylfuro[2,3-d]pyrimidin-4-ylamino)ethyl]phenol
1p. 1H NMR (300 MHz, CDCl3): δ 8.43 (s, 1H), 7.49-7.36 (m,
6H), 7.32-7.23 (m, 4H), 6.89 (d, J = 9.0 Hz, 2H), 6.74 (d, J =
9.0 Hz, 2H), 4.65 (bt, J = 5.1 Hz, 1H), 3.68 (q, J = 6.0 Hz, 2H),
2.71 (t, J = 6.3 Hz, 2H). HRMS (EI) calcd for C26H21N3O2
407.1634, found 407.1629.
(5,6-Diphenylfuro[2,3-d]pyrimidin-4-yl)-(3-imidazol-1-ylpropyl)-
amine 1b. 1H NMR (300 MHz, CDCl3 þ CD3OD): δ 8.31 (s, 1H),
7.68 (s, 1H), 7.54-7.51 (m, 3 H), 7.46-7.43 (m, 4H), 7.24-7.20
(m, 3H), 7.03 (s, 1H), 6.90 (s, 1H), 3.95-3.90 (t, J = 6.6 Hz, 2H),
3.39-3.32 (m, 2H), 1.97-1.92 (m, 2H). HRMS (EI): calcd for
C24H21N5O 395.1746, found 395.1746.
1-(5,6-Diphenylfuro[2,3-d]pyrimidin-4-yl)piperidin-4-ol 1c. 1H
NMR (300 MHz, CDCl3): δ 8.52 (s, 1H), 7.51-7.41 (m, 8H),
7.40-7.27 (m, 2H), 3.76-3.71 (m, 1H), 3.67-3.59 (m, 2H),
3.03-2.97 (m, 2H), 1.63-1.59 (m, 2H), 1.30-1.18 (m, 2H).
HRMS (EI): calcd for C23H21N3O2 371.1634, found 371.1628.
(5,6-Diphenylfuro[2,3-d]pyrimidin-4-yl)-(2-morpholin-4-ylethyl)-
amine 1g. 1H NMR (300 MHz, CDCl3): δ 8.42 (s, 1H), 7.58-7.47
(m, 7 H), 7.28-7.24 (m, 3H), 5.36 (s, 1H), 3.57-3.56 (m, 6H),
2.50-2.34 (m, 6H). 13C NMR (75 MHz, CDCl3): δ 165.1 (C),
157.6 (C), 154.3 (C), 147.0 (C), 132.6 (C), 130.1 (CH), 129.8 (CH),
129.7 (C), 129.1 (CH), 128.7 (CH), 128.6 (CH), 126.6 (CH), 115.2
(C), 77.4 (CH), 66.5 (CH2), 56.7 (CH2), 53.2 (CH2), 37.1 (CH2).
LCMS (ESI) m/z 401.5 (M þ H)þ.
1-(5,6-Diphenylfuro[2,3-d]pyrimidin-4-ylamino)propan-2-ol 1q.
1H NMR (300 MHz, CDCl3): δ 8.34 (s, 1H), 7.54-7.46 (m, 8 H),
7.25-7.22 (m, 2H), 5.06 (bt, J = 5.1 Hz, 1H), 3.92-3.86 (m, 1H),
3.58-3.51 (m, 1H), 3.32-3.23 (m, 1H), 1.11 (d, J = 6.6 Hz, 3H).
LCMS (ESI) m/z 346.1 (M þ H)þ.
(4-Aminobenzyl)-(5,6-diphenylfuro[2,3-d]pyrimidin-4-yl)amine 1r.
1H NMR (300 MHz, CDCl3): δ 8.44 (s, 1H), 7.53-7.44 (m, 8H),
7.26-7.25 (m, 2H), 6.93 (d, J = 7.2 Hz, 2H), 6.61 (d, J = 7.2 Hz,
2H), 4.88 (bs, 1H), 4.51 (d, J = 3.6 Hz, 2H), 3.26 (bs, 2H). HRMS
(EI) calcd for C25H20N4O 392.1637, found 392.1634.

4986 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13
Coumar et al.
S-2-(5,6-Diphenylfuro[2,3-d]pyrimidin-4-ylamino)-2-phenyl-
ethanol 1s. 1H NMR (300 MHz, CDCl3): δ 8.40 (s, 1H), 7.59-
7.28 (m, 13 H), 7.04-7.00 (m, 2H), 5.36 (bd, J = 6.0 Hz, 1H),
5.30-5.24 (m, 1H), 3.86-3.74 (m, 2H). LCMS (ESI) m/z
408.2 (M þ H)þ.
added, and the mixture was neutralized with dilute HCl and
extracted with ethyl acetate (3 ꢀ 20 mL). The organic layers were
combined, washed with water and then brine, and dried over
MgSO4. Removal of solvents under vacuum provided crude pro-
duct 10 (83 mg, 90%). 1H NMR (300 MHz, CDCl3 þ CD3OD): δ
8.43 (s, 1H), 7.55-7.41 (m, 8H), 7.28-7.24 (m, 2H), 3.77 (t, J = 6.0
Hz, 2H), 2.6 (t, J = 6.0 Hz, 2H). LCMS (ESI) m/z 360.1 (M þ H)þ.
3-(5,6-Diphenylfuro[2,3-d]pyrimidin-4-ylamino)-N-phenylpro-
pionamide 6. N-Hydroxybenzotriazole (34 mg, 0.22 mmol) and
1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride
(42 mg, 0.22 mmol) were added to a solution of 10 (80 mg,
0.22 mmol) in anhydrous DMF (4.0 mL), and the mixture was
stirred under nitrogen atmosphere for 1 h at room temperature.
Then aniline (30 mg, 0.32 mmol) was added, with continued
stirring for 16 h. After completion of the reaction, water was
added and extracted with ethyl acetate (3 ꢀ 30 mL). The organic
layers were combined, washed with water (2 ꢀ 20 mL) and then
brine, and dried over MgSO4. Removal of solvents under
vacuum provided crude product, which was purified on silica
gel flash column chromatography using EtOAc/hexane = 3:7 to
1-{4-[2-(5,6-Diphenylfuro[2,3-d]pyrimidin-4-ylamino)ethyl]-
phenyl}-3-phenylurea 3. To a solution of compound 1j (50 mg,
0.123 mmol) in CH2Cl2 (5 mL) was added phenyl isocyanate
(18 mg, 0.147 mmol). The mixture was stirred for 8 h at room
temperature. Then water (10 mL) was added and extracted
with CH2Cl2 (3 ꢀ 10 mL). The combined organic layers were
dried over MgSO4 and concentrated under reduced pressure.
The residue was purified by gradient flash column chroma-
tography on silica gel (acetone/CH2Cl2 = 1:25 to MeOH/
CH2Cl2 = 1:50) to furnish 3 (59 mg, 92%). 1H NMR (CDCl3,
300 MHz): δ 8.43 (s, 1H), 7.48-7.22 (m, 16H), 7.12 (tt, J = 6.8,
2.0 Hz, 1H), 6.95 (d, J = 8.4 Hz, 2H), 6.72 (bs, 2H), 4.66 (bt,
J = 5.2 Hz, 1H), 3.69 (q, J = 6.4 Hz, 2H), 2.75 (t, J = 6.4 Hz,
2H). 13C NMR (100 MHz, CDCl3): δ 164.5 (C), 157.4 (C),
154.0 (C), 153.8 (CH), 146.6 (C), 138.4 (C), 136.9 (C), 133.3
(C), 131.8 (C), 129.6 (CH), 129.4 (CH), 129.1 (C), 129.0 (CH),
128.9 (CH), 128.4 (CH), 126.2 (CH), 123.3 (CH), 120.4 (CH),
120.2 (CH), 114.1 (C), 103.1 (C), 41.9 (CH2), 34.3 (CH2).
1
give 6 (70 mg, 72%). H NMR (300 MHz, CDCl3): δ 8.55 (s,
1H), 8.36 (s, 1H), 7.65-7.62 (m, 2H), 7.42-7.29 (m, 8H),
7.22-7.08 (m, 5H), 5.19 (bs, 1H), 3.80 (q, J = 6.0 Hz, 2H),
2.68 (t, J = 6.0 Hz, 2H). LCMS (ESI) m/z: 435.2 (M þ H)þ.
HPLC purity, 92.6%.
2-(5-Bromo-6-phenylfuro[2,3-d]pyrimidin-4-ylamino)-2-phenyl-
ethanol 12. Compound 12 was synthesized in 69% yield from 118
and (S)-(þ)-phenylglycinol using similar reaction conditions as
those for compound 1. 1H NMR (300 MHz, CDCl3): δ 8.30 (s,
1H), 8.07-8.02 (m, 2H), 7.51-7.30 (m, 8H), 6.89 (d, J = 6.9 Hz,
1H), 5.50-5.45 (m, 1H), 4.05 (d, J = 4.8 Hz, 2H), 3.39 (brs, 1H).
LCMS (ESI) m/z 411.1 (M þ H)þ.
HRMS (FAB) calcd for
C33H27N5O2 525.2165, found
526.2249 (M þ H)þ.
N-{4-[2-(5,6-Diphenylfuro[2,3-d]pyrimidin-4-ylamino)ethyl]-
phenyl}benzamide 4. To a solution of compound 1j (50 mg,
0.123 mmol) and triethylamine (25 mg, 0.246 mmol) in CH2Cl2
(5 mL) was added benzoyl chloride (26 mg, 0.184 mmol), and
the mixture was stirred for 8 h at room temperature. Then
water (10 mL) was added and extracted with CH2Cl2 (3 ꢀ
10 mL). The combined organic layers were dried over MgSO4
and concentrated under reduced pressure. The residue was
purified by gradient flash column chromatography on silica gel
(EtOAc/hexane = 1:10 to MeOH/CH2Cl2 = 1:50) to furnish 4
(53 mg, 85%). 1H NMR (400 MHz, CDCl3): δ 8.44 (s, 1H), 7.90
(d, J = 8.0 Hz, 2H), 7.79 (bs, 1H), 7.23-7.60 (m, 15H), 7.02 (d,
J = 8.0 Hz, 2H), 4.67 (bt, J = 5.2 Hz, 1H), 3.72 (q, J = 6.4 Hz,
2H), 2.80 (t, J = 6.4 Hz, 2H). 13C NMR (100 MHz, CDCl3): δ
165.8 (C), 164.7 (C), 157.3 (C), 154.0 (CH), 146.5 (C), 136.4 (C),
134.8 (C), 134.7 (C), 132.0 (C), 131.8 (CH), 129.6 (CH), 129.5
(CH), 129.3 (C), 129.1 (CH), 128.8 (CH), 128.7 (CH), 128.4
(CH), 128.3 (CH), 127.0 (CH), 126.2 (CH), 120.5 (CH), 114.8
(C), 103.1 (C), 41.8 (CH2), 34.4 (CH2). HRMS (EI) calcd. for
2-[5-(4-Nitrophenyl)-6-phenylfuro[2,3-d]pyrimidin-4-ylamino]-
2-phenylethanol 13a. A mixture of compound 12 (410 mg,
1.0 mmol) and p-nitrobenzeneboronic acid (195 mg, 1.2 mmol)
was dissolved in dioxane (2.0 mL) under a nitrogen atmosphere.
Pd(dppf)2Cl2 (4.2 mg, 0.05 mmol) and Na2CO3 solution (2 M,
1.0 mL) were added to this mixture and heated at 95 ꢀC for 16 h
under nitrogen. After completion, the reaction mixture was
cooled to room temperature, water was added, and the mix-
ture was extracted with ethyl acetate (3 ꢀ 20 mL). Combined
organics were dried over NaSO4, concentrated under vacuum,
and the residue obtained was purified over silica gel flash col-
umn chromatography using hexane/ethyl acetate = 1:1 to give
13a (346 mg, 77%). 1H NMR (300 MHz, CDCl3): δ 8.41
(s, 1H), 8.29 (d, J = 8.4 Hz, 2H), 7.67 (d, J = 8.4 Hz, 2H),
7.50-7.47 (m, 2H), 7.34-7.26 (m, 6H), 7.09-7.05 (m, 2H),
5.23 (dd, J = 10.2, 5.4 Hz, 1H), 5.16 (d, J = 5.4 Hz, 1H),
3.88-3.84 (m, 2H), 3.31 (t, J = 5.4 Hz, 1H). LCMS (ESI) m/z
453.1 (M þ H)þ.
C
33H26N4O2 510.2056, found 510.2055.
N-{4-[2-(5,6-Diphenylfuro[2,3-d]pyrimidin-4-ylamino)ethyl]-
phenyl}benzenesulfonamide 5. To a solution of compound 1j
(50 mg, 0.123 mmol) and triethylamine (25 mg, 0.246 mmol) in
CH2Cl2 (5 mL) was added benzenesulfonyl chloride (32 mg,
0.184 mmol) and the mixture stirred for 8 h at room tempera-
ture. Then water (10 mL) was added and extracted with
CH2Cl2 (3 ꢀ 10 mL). The combined organic layers were dried
over MgSO4 and concentrated under reduced pressure. The
residue was purified by gradient flash column chromatography
2-[5-(3-Nitrophenyl)-6-phenylfuro[2,3-d]pyrimidin-4-ylamino]-
2-phenylethanol 13b. Compound 13b was synthesized in 75%
yield in a manner similar to that for 13a, using m-nitrobenzene-
boronic acid. 1H NMR (300 MHz, CDCl3): δ 8.41 (s, 2H), 8.30
(ddd, J = 8.1, 1.2, 1.2 Hz, 1H), 7.82 (dd, J = 7.8, 1.2 Hz, 1H),
7.65 (dd, J = 8.1, 7.8 Hz, 1H), 7.50-7.47 (m, 2H), 7.34-7.26 (m,
6H), 7.11-7.07 (m, 2H), 5.27-5.20 (m, 2H), 3.88-3.84 (m, 2H),
3.31 (t, J = 5.4 Hz, 1H). LCMS (ESI) m/z 453.1 (M þ H)þ.
N-{4-[4-(2-Hydroxy-1-phenylethylamino)-6-phenylfuro[2,3-d]-
pyrimidin-5-yl]phenyl}acrylamide 7. A mixture of 13a (136 mg,
0.3 mmol) and 5% Pd/C (10 mg) in EtOH (3 mL) was hydro-
genated at atmospheric pressure for 16 h. The reaction mixture
was filtered over Celite, and the solvents removed under vacuum
to give 2-[5-(4-aminophenyl)-6-phenylfuro[2,3-d]pyrimidin-4-
ylamino]-2-phenylethanol. Acryloyl chloride (0.03 mL, 0.33
mmol) was added to a mixture of the above compound and
pyridine (0.05 mL, 0.6 mmol) in a solution of ether (3 mL). The
reaction mixture was then stirred at room temperature for 16 h.
After completion of the reaction, water was added and extracted
with ethyl acetate (3 ꢀ 20 mL). The combined organics were
on silica gel (acetone/CH2Cl2 = 1:25 to MeOH/CH2Cl2
=
1:10) to furnish 5 (60 mg, 90%). 1H NMR (400 MHz, CDCl3): δ
8.42 (s, 1H), 7.77 (d, J = 8.4 Hz, 2H), 7.46-7.53 (m, 3H),
7.34-7.43 (m, 5H), 7.24-7.27 (m, 5H), 6.96 (d, J = 8.4 Hz,
2H), 6.88 (d, J = 8.4 Hz, 2H), 6.70 (bs, 1H), 4.59 (bt, J = 5.2
Hz, 1H), 3.66 (q, J = 6.4 Hz, 2H), 2.71 (t, J = 6.4 Hz, 2H). 13C
NMR (100 MHz, CDCl3): δ 164.6 (C),157.3 (C), 154.0 (CH),
146.5 (C), 139.2 (C), 135.5 (C), 135.0 (C), 132.9 (C), 132.0 (C),
129.6 (CH), 129.4 (CH), 129.3 (CH), 129.3 (C), 129.0 (CH),
128.8 (CH), 128.4 (CH), 127.1 (CH), 126.2 (CH), 121.7 (CH),
114.8 (CH), 103.1 (C), 41.6 (CH2), 34.3 (CH2). HRMS (EI)
calcd for C32H26N4O3S 546.1726, found 546.1729.
3-(5,6-Diphenylfuro[2,3-d]pyrimidin-4-ylamino)propionic Acid 10.
LiOH (18 mg, 0.78 mmol) was added to a solution of 1i (100 mg,
0.26 mmol) in a mixture of 10 mL of methanol/water (4:1) and
refluxed for 2 h. Solvents were removed under vacuum. Water was

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13 4987
dried over MgSO4, concentrated under vacuum and the residue
was purified over silica gel flash column chromatography using
hexane/ethyl acetate = 2:3 to give7 (29 mg, 20%). 1H NMR(300
MHz, CDCl3) δ 8.33 (s, 1H), 8.00 (brs, 1H), 7.80-7.71 (m, 2H),
7.57 (d, J = 8.4 Hz, 2H), 7.47-7.44 (m, 3H), 7.30-7.19 (m, 5H),
7.07 (d, J = 8.4 Hz, 2H), 6.50 (d, J = 16.8 Hz, 1H), 6.32 (dd, J =
16.8, 10.2 Hz, 1H), 5.83 (d, J = 10.2 Hz, 1H), 5.45 (d, J = 6.3 Hz,
1H), 5.20-5.18 (m, 1H), 3.86-3.73 (m, 2H). LCMS (ESI) m/z
477.1 (M þ H)þ. HPLC purity, 90.4%.
with 20 μL of a MTS/PMS mixture (MTS/PMS ratio of 20:1) in
each well of the 96-well plate for 2 h at 37 ꢀC in a humidified
incubator with 5% CO2 to allow viable cells to convert the tetra-
zolium salt (MTS) into formazan. The amount/concentration of
formazan, which indicates the number of live cells, was determined
by measuring its absorbance at 490 nm using a PerkinElmer
Victor2 plate reader (PerkinElmer, Shelton, CT).
X-ray Cocrystal Structure Determination of 3 in Complex with
Aurora Kinase A. Cocrystal complex structure determination
was carried out as reported in our previous publication using
compound 3 as the inhibitor.22
N-{3-[4-(2-Hydroxy-1-phenylethylamino)-6-phenylfuro[2,3-d]-
pyrimidin-5-yl]phenyl}acrylamide 8. Compound 8 was synthe-
sized in 25% yield from 13b in a manner similar to that for 7. 1H
NMR (300 MHz, CDCl3): δ 8.58 (brs, 1H), 8.27 (s, 1H),
7.79-7.64 (m, 2H), 7.49-7.46 (m, 2H), 7.40 (t, J = 7.5 Hz,
1H), 7.27-7.20 (m, 8H), 7.05-7.02 (d, J = 7.5 Hz, 2H), 6.46 (d,
J = 16.8 Hz, 1H), 6.28 (dd, J = 16.8, 10.2 Hz, 1H), 5.73 (d, J =
10.2 Hz, 1H), 5.64 (d, J = 6.0 Hz, 1H), 5.30 (brs, 1H), 3.86 (dd,
J = 11.4, 2.7 Hz, 1H), 3.69 (dd, J = 11.4, 6.9 Hz, 1H). HRMS
(EI) calcd for C29H24N4O3 476.1848, found 476.1843.
X-ray Cocrystal Structure Determination of 8 in Complex with
EGFR Kinase. (a) Expression and Purification of EGFR. EGFR
catalytic domain (residues 696-1022) was cloned into the pBac-
PAK8-MT-EGFP-GST vector and expressed in insect cells Hi5.
The protein was purified by GST affinity column. The bound
protein was washed with 1-fold PBS buffer solution and eluted with
20 mM GSH in 50 mM Tris-Cl (pH 8.0) buffer solution. The eluted
fractions were then exchanged to PreScission digestion buffer
(50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1 mM DTT,
pH 8.0) by desalting column, followed by treatment with PreScis-
sion protease (GE Healthcare) at 4 ꢀC for 16 h to remove the GST
tag, and then loaded to GST affinity column to recover follow-
through. The follow-through was then buffer-exchanged and con-
centrated to 3-6 mg/mL in 20 mM Tris-Cl (pH 8.0), 5 mM DTT,
100 mM sodium chloride, and 0.5 mM EDTA buffer solution.
(b) Crystallization and structure determination. The hanging
drop method was used to obtain crystals of the wild type EGFR
catalytic domain. The crystals were grown at 18 ꢀC for 1 week.
Crystals of EGFR kinase complexed with compound were
obtained by soaking apo-EGFR kinase crystals in reservoir
solution containing 1.0 M ammonium citrate tribasic, pH 7.0,
0.1 M Bis-Tris propane, pH 7.0 (Hampton research), and 0.6
mM compound 8. Before being flash-frozen in liquid nitrogen,
the crystal was immersed briefly in a cryoprotectant containing
22.2% glycerol in soaking drop. Diffraction data were collected
on beamlines BL13B1 (Taiwan). The data were processed by
DENZO23 and reduced with SCALEPACK. The structure was
solved by molecular replacement in MOLREP24 using the
published EGFR structure (PDB ID 1M17) as the search model.
The refinement calculations were performed by REFMAC5,25
and model building was carried out with the program O11.26
Aurora Kinase A Assay. Aurora kinase A enzyme inhibition
assay was performed as reported by us earlier.20
EGFR WT/DM Kinase Assay. This assay was performed by
modification of the method reported by us earlier.21 GST-
EGFR-KDWT containing the EGFR kinase catalytic domain
(residues 696-1022) and GST-EGFR-KDL858R/T790M contain-
ing the EGFR kinase catalytic domain (residues from 696 to
1022 and with L858R/T790M) were expressed in Sf9 insect cells
transfected with baculovirus containing pBac-PAK8-GST-
EGFR-KD plasmid, respectively. GST-EGFR-KDWT protein
expression and purification and kinase assay were done in a
manner as reported earlier.21 The EGFRL858R/T790M Kinase-Glo
assays were carried out in 96-well plates at 37 ꢀC for 60 min in a
final volume of 50 μL, including the following components:
200 ng of GST-EGFR-KDL858R/T790M proteins, 25 mM
HEPES, pH 7.4, 4 mM MnCl2, 2 mM DTT, 10 mM MgCl2,
0.1 mg/mL bovine serum albumin, 10 μM poly(Glu,Tyr) 4:1, 0.5
mM Na3VO4, and 1 μM ATP. Following incubation, 50 μL
Kinase-Glo Plus reagent (Promega) was added, and the mixture
was incubated at 25 ꢀC for 20 min. A 70 μL aliquot of each
reaction mixture was transferred to a black microtiter plate, and
the luminescence was measured using a Wallac Vector 1420
multilabel counter (PerkinElmer).
HCT-116 Antiproliferative Assay. HCT-116 cell viability was
examined by MTS assay (Promega, Madison, WI). Two-thousand
HCT-116 cells in 100 μL of McCoy’s 5a medium were seeded
in each well of a 96-well plate. After 96 h of incubation with a test
compound, the cells were incubated with 20 μL of a MTS/PMS
mixture (MTS/PMS ratio of 20:1) for 2 h at 37 ꢀC in a humidified
incubator with 5% CO2 to allow viable cells to convert the tetra-
zolium salt (MTS) into formazan. The amount/concentration of
formazan, which indicates the number of live cells, was determined
by measuring the absorbance at 490 nm using a PerkinElmer
Victor2 plate reader (PerkinElmer, Shelton, CT).
HCT-116 Flow Cytometry Analysis. Flow cytometry analysis
of HCT-116 cells treated with compound 3 was performed as
follows. HCT116 cells were seeded at low confluency in six-well
plates and treated with compounds or DMSO for 48 h. Cells
were harvested and stained with 0.1% propidium iodide (Sigma
Aldrich) hypotonic buffer at 4 ꢀC for 15 min. Samples were
subjected to flow cytometry (FACS Calibus, BD) to measure
DNA content of the cells (FL2-A). Cell cycle profiling was
analyzed by CellQuest Pro.
Acknowledgment. We thank the staff of beamline BL13B1
at the National Synchrotron Radiation Research Centre
(NSRRC), Taiwan, and SP12B2 at SPring-8, Japan, for
technical assistance, as well as Chung-Yu Chang and Teng-
Yuan Chang of the National Health Research Institutes for
their assistance with Aurora and EGFR assay development.
We also thank Mark Swofford for helping with the English
editing. The authors acknowledge financial support by the
National Science Council (Grant No. NSC-98-2119-M-400-
001-MY3 to H-P.H.) and Department of Health (Grant No.
DOH98-TD-G-111-019 to Y.-S.C. and Grant No. DOH98-
TD-G-111-020 to H.-P.-H.), Taiwan, ROC.
Supporting Information Available: List of amines used for
library synthesis, flow cytometry and Western blotting analysis
for compound 3, X-ray cocrystal structure of EGFR protein in
complex with inhibitor 8, X-ray refinement statistics for com-
pounds 3 and 8, and HPLC purity determination conditions.
This material is available free of charge via the Internet at http://
pubs.acs.org.
HCT-116 Western Blot Analysis. Western blot analysis of
HCT-116 cells treated with compound 3 was performed as
reported by us earlier.22
HCC-827 Antiproliferative Assay. HCC827 cell viability was
examined by MTS assay (Promega, Madison, WI). One-thousand
HCC827 cells in 100 μL of RPMI1640 with 10% FBS medium
were seeded in each well of a 96-well plate. After 96 h of incu-
bation with the test compound, the cells were further incubated
References
(1) Watkins, K. J. Fighting the clock. Chem. Eng. News 2002, 80, 27–33.
(2) Brik, A.; Wu, C. Y.; Wong, C. H. Microtiter plate based chemistry
and in situ screening: a useful approach for rapid inhibitor dis-
covery. Org. Biomol. Chem. 2006, 4, 1446–1457.

4988 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 13
Coumar et al.
(3) Noble, M. E.; Endicott, J. A.; Johnson, L. N. Protein kinase
inhibitors: insights into drug design from structure. Science 2004,
303, 1800–1805.
(4) Fu, J.; Bian, M.; Jiang, Q.; Zhang, C. Roles of Aurora kinases in
mitosis and tumorigenesis. Mol. Cancer Res. 2007, 5, 1–10.
(5) Gautschi, O.; Heighway, J.; Mack, P. C.; Purnell, P. R.; Lara, P. N.,
Jr.; Gandara, D. R. Aurora kinases as anticancer drug targets. Clin.
Cancer Res. 2008, 14, 1639–1648.
(6) Tiseo, M.; Loprevite, M.; Ardizzoni, A. Epidermal growth factor
receptor inhibitors: a new prospective in the treatment of lung
cancer. Curr. Med. Chem.: Anti-Cancer Agents 2004, 4, 139–148.
(7) Sharma, S. V.; Bell, D. W.; Settleman, J.; Haber, D. A. Epidermal
growth factor receptor mutations in lung cancer. Nat. Rev. Cancer
2007, 7, 169–181.
(8) Coumar, M. S.; Tsai, M. T.; Chu, C. Y.; Uang, B. J.; Teng, C. H.;
Chang, C. Y.; Chang, T. Y.; Lin, W. H.; Hsu, J. T. A.; Wu, J. S.;
Wu, S. Y.; Chao, Y. S.; Hsieh, H. P. Identification, SAR studies
and X-ray cocrystal analysis of a novel furano-pyrimidine Aurora
kinase A inhibitor. ChemMedChem 2010, 5, 255–267.
(9) Foloppe, N.; Fisher, L. M.; Howes, R.; Kierstan, P.; Potter, A.;
Robertson, A. G.; Surgenor, A. E. Structure-based design of novel
Chk1 inhibitors: insights into hydrogen bonding and protein-
ligand affinity. J. Med. Chem. 2005, 48, 4332–4345.
(10) Wu, C. Y.; Chang, C. F.; Chen, J. S.; Wong, C. H.; Lin, C. H. Rapid
diversity-oriented synthesis in microtiter plates for in situ screen-
ing: discovery of potent and selective alpha-fucosidase inhibitors.
Angew. Chem., Int. Ed. 2003, 42, 4661–4664.
(11) Lee, L. V.; Mitchell, M. L.; Huang, S. J.; Fokin, V. V.; Sharpless,
K. B.; Wong, C. H. A potent and highly selective inhibitor of
human alpha-1,3-fucosyltransferase via click chemistry. J. Am.
Chem. Soc. 2003, 125, 9588–9589.
(12) Liao, J. J. Molecular recognition of protein kinase binding pockets
for design of potent and selective kinase inhibitors. J. Med. Chem.
2007, 50, 409–424.
(16) Wissner, A.; Mansour, T. S. The development of HKI-272 and
related compounds for the treatment of cancer. Arch. Pharm.
(Weinheim, Ger.) 2008, 341, 465–477.
(17) Fry, D. W.; Bridges, A. J.; Denny, W. A.; Doherty, A.; Greis,
K. D.; Hicks, J. L.; Hook, K. E.; Keller, P. R.; Leopold, W. R.;
Loo, J. A.; McNamara, D. J.; Nelson, J. M.; Sherwood, V.; Smaill,
J. B.; Trumpp-Kallmeyer, S.; Dobrusin, E. M. Specific, irreversible
inactivation of the epidermal growth factor receptor and erbB2, by
a new class of tyrosine kinase inhibitor. Proc. Natl. Acad. Sci. U.S.A.
1998, 95, 12022–12027.
(18) Ciardiello, F.; Tortora, G. EGFR antagonists in cancer treatment.
N. Engl. J. Med. 2008, 358, 1160–1174.
(19) Li, D.; Ambrogio, L.; Shimamura, T.; Kubo, S.; Takahashi, M.;
Chirieac, L. R.;Padera, R. F.;Shapiro, G. I.;Baum, A.;Himmelsbach,
F.; Rettig, W. J.; Meyerson, M.; Solca, F.; Greulich, H.; Wong,
K. K. BIBW2992, an irreversible EGFR/HER2 inhibitor highly
effective in preclinical lung cancer models. Oncogene 2008, 27, 4702–
4711.
(20) Coumar, M. S.; Wu, J. S.; Leou, J. S.; Tan, U. K.; Chang, C. Y.;
Chang, T. Y.; Lin, W. H.; Hsu, J. T.; Chao, Y. S.; Wu, S. Y.; Hsieh,
H. P. Aurora kinase A inhibitors: identification, SAR exploration
and molecular modeling of 6,7-dihydro-4H-pyrazolo-[1,5-a]-
pyrrolo[3,4-d]pyrimidine-5,8-dione scaffold. Bioorg. Med. Chem.
Lett. 2008, 18, 1623–1627.
(21) Lin, W. H.; Song, J. S.; Chang, T. Y.; Chang, C. Y.; Fu, Y. N.; Yeh,
C. L.; Wu, S. H.; Huang, Y. W.; Fang, M. Y.; Lien, T. W.; Hsieh,
H. P.; Chao, Y. S.; Huang, S. F.; Tsai, S. F.; Wang, L. M.; Hsu,
J. T.; Chen, Y. R. A cell-based high-throughput screen for epider-
mal growth factor receptor pathway inhibitors. Anal. Biochem.
2008, 377, 89–94.
(22) Coumar, M. S.; Leou, J. S.; Shukla, P.; Wu, J. S.; Dixit, A. K.; Lin,
W. H.; Chang, C. Y.; Lien, T. W.; Tan, U. K.; Chen, C. H.; Hsu,
J. T.; Chao, Y. S.; Wu, S. Y.; Hsieh, H. P. Structure-based drug
design of novel Aurora kinase A inhibitors: structural basis for
potency and specificity. J. Med. Chem. 2009, 52, 1050–1062.
(23) Otwinowski, Z.; Minor, W. Processing of x-ray diffraction data
collected in oscillation mode. Methods Enzymol. 1997, 276, 307–326.
(24) Vagin, A; Teplyakov, A. MOLREP: an automated program for
molecular replacement. J. Appl. Crystallogr. 1997, 30, 1022–1025.
(25) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Refinement of
macromolecular structures by the maximum-likelihood method.
Acta Crystallogr. 1997, D53, 240–255.
(13) Liu, Y.; Gray, N. S. Rational design of inhibitors that bind to
inactive kinase conformations. Nat. Chem. Biol. 2006, 2, 358–364.
(14) Backes, A. C.; Zech, B.; Felber, B.; Klebl, B.; Muller, G. Small-
molecule inhibitors binding to protein kinase. Part II: the novel
pharmacophore approach of type II and type III inhibition. Expert
Opin. Drug Discovery 2008, 3, 1427–1449.
(15) Harrington, E. A.; Bebbington, D.; Moore, J.; Rasmussen, R. K.;
Ajose-Adeogun, A. O.; Nakayama, T.; Graham, J. A.; Demur, C.;
Hercend, T.; Diu-Hercend, A.; Su, M.; Golec, J. M.; Miller, K. M.
VX-680, a potent and selective small-molecule inhibitor of the
Aurora kinases, suppresses tumor growth in vivo. Nat. Med. 2004,
10, 262–267.
(26) Jones, T. A.; Zou, J.-Y.; Cowan, S. W.; Kjeldgaard Improved
methods for building protein models in electron density maps and
the location of errors in these models. Acta Crystallogr. 1991, A47,
110–119.