Structure-based virtual screening of Src kinase inhibitors

Article abstract of DOI:10.1016/j.bmc.2009.02.054

Src is an important target in multiple processes associated with tumor growth and development, including proliferation, neovascularization, and metastasis. In this study, hit identification was performed by virtual screening of commercial and in-house compound libraries. Docking studies for the hits were performed, and scoring functions were used to evaluate the docking results and to rank ligand-binding affinities. Subsequently, hit optimization for potent and selective candidate Src inhibitors was performed through focused library design and docking analyses. Consequently, we report that a novel compound '43' with an IC₅₀ value of 89 nM, representing (S)-N-(4-(5-chlorobenzo[d][1,3]dioxol-4-ylamino)-7-(2-methoxyethoxy)quin azolin-6-yl)pyrrolidine-2-carboxamide, is highly selective for Src in comparison to EGFR (IC₅₀ ratio > 80-fold) and VEGFR-2 (IC₅₀ ratio > 110-fold). Compound 43 exerted anti-proliferative effects on Src-expressing PC3 human prostate cancer and A431 human epidermoid carcinoma cells, with calculated IC₅₀ values of 1.52 and 0.78 μM, respectively. Moreover, compound 43 (0.1 μM) suppressed the phosphorylation of extracellular signal-regulated kinases and p90 ribosomal S6 kinase, downstream molecules of Src, in a time-dependent manner, in both PC3 and A431 cell lines. The docking structure of compound 43 with Src disclosed that the chlorobenzodioxole moiety and pyrrolidine ring of C-6 quinazoline appeared to fit tightly into the hydrophobic pocket of Src. Additionally, the pyrrolidine NH forms a hydrogen bond with the carboxyl group of Asp348. These results confirm the successful application of virtual screening studies in the lead discovery process, and suggest that our novel compound 43 can be an effective Src inhibitor candidate for further lead optimization.

Full text of DOI:10.1016/j.bmc.2009.02.054

Bioorganic & Medicinal Chemistry 17 (2009) 3152–3161
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Structure-based virtual screening of Src kinase inhibitors
Kyungik Lee a,b, Jongwoo Kim b, Ki-Woong Jeong c, Ki Won Lee c, Yeonjoo Lee b, Ji Yeon Song b,
Maeng Sup Kim b, Gwan Sun Lee b, Yangmee Kim c,
*
a Department of Chemistry, Konkuk University, Seoul 143-701, Republic of Korea
b Hanmi Research Center, Hanmi Pharmaceutical Co. Ltd, 377-1 Yeongcheon-ri, Dongtan-myeon, Hwaseong, Gyeonggi-do 445-813, Republic of Korea
c Department of Bioscience and Biotechnology, Konkuk University, Seoul 143-701, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Src is an important target in multiple processes associated with tumor growth and development, includ-
ing proliferation, neovascularization, and metastasis. In this study, hit identification was performed by
virtual screening of commercial and in-house compound libraries. Docking studies for the hits were per-
formed, and scoring functions were used to evaluate the docking results and to rank ligand-binding affin-
ities. Subsequently, hit optimization for potent and selective candidate Src inhibitors was performed
through focused library design and docking analyses. Consequently, we report that a novel compound
‘43’ with an IC50 value of 89 nM, representing (S)-N-(4-(5-chlorobenzo[d][1,3]dioxol-4-ylamino)-7-(2-
methoxyethoxy)quinazolin-6-yl)pyrrolidine-2-carboxamide, is highly selective for Src in comparison to
EGFR (IC50 ratio > 80-fold) and VEGFR-2 (IC50 ratio > 110-fold). Compound 43 exerted anti-proliferative
effects on Src-expressing PC3 human prostate cancer and A431 human epidermoid carcinoma cells, with
Received 26 November 2008
Revised 26 February 2009
Accepted 27 February 2009
Available online 5 March 2009
Keywords:
Src
Kinase
Inhibitor
Virtual screening
Docking
calculated IC50 values of 1.52 and 0.78 lM, respectively. Moreover, compound 43 (0.1 lM) suppressed the
LIGANDFIT
phosphorylation of extracellular signal-regulated kinases and p90 ribosomal S6 kinase, downstream mol-
ecules of Src, in a time-dependent manner, in both PC3 and A431 cell lines. The docking structure of com-
pound 43 with Src disclosed that the chlorobenzodioxole moiety and pyrrolidine ring of C-6 quinazoline
appeared to fit tightly into the hydrophobic pocket of Src. Additionally, the pyrrolidine NH forms a hydro-
gen bond with the carboxyl group of Asp348. These results confirm the successful application of virtual
screening studies in the lead discovery process, and suggest that our novel compound 43 can be an effec-
tive Src inhibitor candidate for further lead optimization.
Drug discovery
Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction
novel Src kinase inhibitors. To date, purine, pyrrolopyrimidine, pyr-
idopyrimidine, naphthyridone, quinazoline, and quinoline-based
The viral Src gene encoded by Rous sarcoma virus (RSV) was the
first defined oncogene encoding the first recognized tyrosine ki-
nase.1 Tyrosine Kinase (TK)-induced phosphorylation of proteins
is a fundamental mechanism for the control of cell growth and dif-
ferentiation. Src kinase is a member of a structurally homologous
group of nonreceptor TKs present in the cytoplasm, known as the
‘Src family of kinases’.2 Src participates in several signaling path-
ways regulating proliferation, differentiation, and migration,3 and
small-molecule inhibitors of this protein are under investigation
as potential agents for the treatment of a variety of diseases.4,5
It is proposed that inhibitors of Src phosphorylation may im-
pede uncontrolled tumor cell growth, and thus function as novel
therapeutic agents for cancer. Specifically, Src activity is elevated
in breast, pancreatic, ovarian, esophageal, lung, gastric, colon, and
head-and-neck cancers.6–10 A number of planar heteroaromatic
small-molecule inhibitor templates have been exploited to identify
inhibitors have been reported.11 Among these, Dasatinib
(BMS354825; Bristol Meyers Squibb Oncology),12 AZD-0530
(AstraZeneca)13, and Bosutinib (SKI-606; Wyeth Research),14 have
reached Phase 1 or 2 clinical trials (Fig. 1). These small-molecule
inhibitors have similar structures derived from ‘flat’ heterocyclic
cores, and specifically inactivate the ATP-binding site of Src family
kinases. A dual Src/Abl inhibitor, AZD-0530, is based on the anilino-
quinazoline scaffold. This compound displays potency against Src
with an IC50 value of 10 nM. Dasatinib (BMS-354825), a novel oral
and potent multi-targeting inhibitor of Bcr-Abl and Src family ki-
nases (SFK), is a promising cancer therapeutic agent with an IC50
of 0.6 nM against Src. Wyeth has developed a dual Bcr-Abl/Src
inhibitor, Bosutinib (SKI-606), based on a quinoline scaffold, that
is structurally related to the AstraZeneca quinazoline template.15
This compound displays an IC50 of 1.2 nM against Src.
Ligand docking is a method in which the conformational selec-
tion of compounds has a high impact on the feasibility of the pro-
posed binding poses. Virtual screening has been established as a
valuable in silico technique alongside traditional high-throughput
* Corresponding author. Tel.: +82 2 450 3421; fax: +82 2 447 5987.
E-mail address: ymkim@konkuk.ac.kr (Y. Kim).
0968-0896/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2009.02.054

K. Lee et al. / Bioorg. Med. Chem. 17 (2009) 3152–3161
3153
Figure 1. Structures of template inhibitors interacting with the ATP-binding site of Src kinase.
screening for novel active compounds, with increasing success in
the pharmaceutical industry.16,17 Structure-based virtual screening
methods are emerging as reliable and complementary approaches
to high-throughput screening in the lead discovery process, from
hit identification to lead optimization.18 These processes are per-
formed in two stages: (i) accurate prediction of the pose, and (ii)
estimation of tightness of binding (scoring).19 Once a putative pro-
tein–ligand interaction is proposed, a scoring function is used to
estimate relative ligand affinity for the target.20 These screening
hits represent a promising source of candidate molecules for eval-
uation of binding modes by docking and further biological testing.
In this study, we present the computational virtual screening of
molecules to identify new potential hits for Src kinase inhibitors,
using a docking program. The docking program, LIGANDFIT, was em-
ployed to dock hits, and several scoring and consensus scoring
functions were applied to evaluate docking results and rank li-
gand-binding affinities. One novel potential lead was identified
on the basis of docking scores and examination of how key interac-
tions are retained for kinase binding.21 Our results show that this
compound is a potent inhibitor of Src, and induces anti-prolifera-
tive effects in Src-expressing PC3 human prostate cancer and
A431 human epidermoid carcinoma cell lines.
tween the crystal structure (PDB 2H8H) and the most reasonable
binding modes of AZD-0530 docked with LIGANDFIT was calculated.
The results of control docking showed that LIGANDFIT determined
the optimal orientation of the docked inhibitor, AZD-0530 to be
close to that of the original orientation found in the crystal
(Fig. 2). The rms deviation between the experimental docked con-
formation and the calculated docked conformation for AZD-0530 in
Src was 0.34 Å.
Three inhibitors (1–3), Dasatinib (BMS354825; Bristol Meyers
Squibb Oncology, Princeton, NY), AZD-0530 (AstraZeneca, Maccles-
field, United Kingdom) and Bosutinib (SKI-606; Wyeth Research,
Pearl River, NY), were docked onto the optimized structure of
Src, and the resulting complexes further minimized to overcome
protein rigidity. We evaluated various scoring functions (LigScore1,
LigScore2, PLP1, PLP2, and PMF) in an attempt to accurately predict
the binding affinities between ligand molecules and their protein
receptors. LigScore2 was more successful in retrieving an accurate
pose as the top scorer in the LIGANDFIT docking ensemble than were
the other functions, and LigScore2 was consequently employed to
obtain various scores.
Figure 1 depicts the chemical structures of the three known
inhibitors, and interactions between inhibitors and the ATP-bind-
ing site of the Src kinase catalytic domain. Inhibitor molecules oc-
cupy the ATP-binding site. The quinazoline, quinoline, and pyridine
ring templates span the region occupied by the ribose in ATP. The
2. Results
Src is one of the few thoroughly characterized and well-vali-
dated targets in anti-cancer therapy. The main objective of the
present study is to develop a model for virtual screening of Src
kinase inhibitors. To achieve this goal within a reasonable time-
frame, we needed a rapid and robust docking tool. Our initial stud-
ies were performed with the software, LIGANDFIT, widely regarded as
one of the best docking programs. The model was further applied
for screening an in-house database to identify new leads for Src ki-
nase inhibitors.
2.1. Binding models of compounds 1–3
Recently, several X-ray structures of Src kinase have been deter-
mined as complexes with a number of small-molecule inhibitors,
including AP23464 and AP23451, purvalanol, CPG77675 and STI-
571 (imatinib).22–24 The X-ray structure of human Src (entry
2H8H) was selected as the starting reference for molecular docking
analyses.13 For a validation of the accuracy of the docking program
LIGANDFIT and approach used in this study, the RMS deviation be-
Figure 2. Crystal structure (pink) of AZD-0530 bound to Src compared to the
predicted binding mode (blue).

3154
K. Lee et al. / Bioorg. Med. Chem. 17 (2009) 3152–3161
highly conserved hydrogen bond is created by N1 of the ring tem-
plates, which binds to the backbone amide of Met341. The cyano
group of Bosutinib is hydrogen-bonded to the side-chain OH group
of Thr338. Simulations disclose that a network of hydrogen bonds
with Met341 and Thr338 restricts the inhibitors within the hinge
region of Src, with their molecular scaffolds accommodated within
the adenine pocket. Moreover, hydrophobic interactions were
identified between anilino moieties and the hydrophobic region
formed by Met314, Val323, Ile336, Leu393, Ala403, and Phe405.
In Dasatinib, an additional hydrogen bond was identified between
the amide carbonyl group and side-chain NH2 of Lys295. 2-Chloro-
6-methylbenzamide occupies a deep hydrophobic pocket.12 These
structural models were useful in providing concrete starting points
for further ligand design.
ular physical property filters and similarity are very attractive
because of reduced cost, time, reagent, and compound
consumption.
The docking model developed to date provides an accurate idea
of the necessary and sufficient molecular attributes required in a
new lead. Our objective was to screen a large subset of the corpo-
rate collection of compounds. We performed flexible lignad dock-
ing with LIGANDFIT on a relatively large library containing 2300
drug-like molecules extracted from optimized commercial and
in-house databases. Molecular docking rate was about 65%. Among
these, we only considered hit compounds having H-bond with
Met341 (major interactions of the Src binding site) and we selected
the top-scoring compounds by LigScore2.
Docking calculations of the selected compounds with LIGANDFIT
led to the identification of four potent inhibitors with activities
in the nanomolar range, producing 12 hits with high LigScore2
(Table 1). Following the above procedure, a dozen compounds
were finally selected from chemical archives for actual testing in
2.2. Identification of hits by virtual screening
Above all the compounds for virtual screening were filtered by
applying molecular property filters to eliminate compounds with
undesired molecular properties. Src kinase active compounds (can-
didates and compounds in research states) that have MW 400–500,
H-bond donors 62, H-bond acceptors 7–10, 2D polar surface area
65–100, and AlogP 4–5. At first, all compounds were prefiltered
according to the criteria. And then to search for similar structures
with active compounds we pursued 2D-searching, which is based
on chemical similarities. These filters helped in selecting 2300 mol-
ecules from a total of 1,561,000 molecules. The concept of molec-
a Src phosphorylation inhibition assay at a single 10 lM concentra-
tion. This prescreen assay aims to eliminate a large proportion of
the inactive agents, and preserve ‘active’ agents. Four of the se-
lected compounds inhibited Src activity at a concentration of
10 lM. Among the four active compounds (shown in Fig. 3), the
most potent compound was selected for further investigation.
The hit selection process is based not only on potency and novelty,
but also on all other parameters that subsequently influence the
successful outcome of the lead optimization phase. The hit was fur-
ther investigated in terms of docking scores, consensus scores,
binding conformation, and the degree of fit into the Src kinase-
binding site. In particular, how interactions are retained with
important amino acid residues (Met341) in the enzyme binding
site was considered.
Compound 4, a quinazolinone derivative synthesized at Hanmi
Pharm Co. Ltd, displayed a good docking score (Ligscore2 = 6.00)
and most active potency (IC50 = 300 nM), supporting its activity
as a strong inhibitor of Src kinase. Thus, the virtual screening ap-
proach yielded a novel and potent hit class of Src kinase inhibitors
from a limited selection of compounds.
Table 1
Docking scores of the reference compound and the final 12 hits obtained from the
docking study
Compounds
LigScore2
Src tyrosine kinase
(10 lM) positivity
AZD-0530
1
2
3
4
5
6
7
8
6.50
6.29
6.15
6.01
6.00
5.77
5.78
5.72
5.64
5.82
5.31
5.44
5.86
+
+
+
+
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
2.3. Design and synthesis of analogs
9
Compound 4, (S)-N-(4-(3-chloro-4-fluorophenylamino)-7-(2-me-
thoxyethoxy)qui-nazolin-6-yl)pyrrolidine-2-carboxamide, selected
from virtual screening analyses, was employed as a hit for designing
10
11
12
Figure 3. Structures of the final four hits selected from virtual screening and enzyme-based binding affinity assays.

K. Lee et al. / Bioorg. Med. Chem. 17 (2009) 3152–3161
3155
new Src inhibitors. This is a novel anilinoquinazoline compound
substituted atthe C-6 position of thequinazoline ring. Fromthe struc-
turalmodel, itishypothesizedthat thequinazoline ringinteractswith
theadeninebindingsiteofkinase,whereastheanilineringisburiedin
anadjacenthydrophobicpocket. Ahighlyconservedhydrogenbondis
created by N1 of quinazoline, which binds to the backbone amide of
Met341. Docking of quinazoline-based compounds to the Src model
suggests that pyrrolidine rings linked to the C-6 position of quinazo-
line fit the hydrophobic pocket. Pyrrolidine-NH has one additional
interaction with the side-chain carboxyl group of Asp348.
We were able to improve the selectivity of our lead compound
by taking advantage of sequence differences among structurally re-
lated kinases within the extended hydrophobic pocket. Modifica-
tion of substituents on the aniline phenyl ring led to dramatic
changes in the inhibitory activity against Src kinase.25,26 To gain
an insight into the structural basis of the inhibitory activities of
our aniline series, we performed molecular modeling and docking
studies using the available 3D structure of Src kinase.13
We uses sets of pre-defined molecular building blocks that are
connected by a virtual synthesis scheme.19 On the basis of pub-
lished experimental data, we have built a small focused library
(Table 2). The preferred conformation was analyzed by varying
the aniline or bicyclic aniline substituents at C-4 of quinazoline.
Manipulation of the aniline substitution patterns has allowed
the discovery of a potent and selective Src inhibitor. The struc-
ture-guided approaches to fragment optimization rapidly generate
compounds with increased potency, from initial low activity for
fragment hits to nanomolar activities that are usually associated
with high-quality lead molecules. Virtual compounds were gener-
ated from compound 4 with 33 aromatic ring groups at C-4. Next,
docking studies were performed on the compounds. In view of the
structural similarity and binding affinity to Src, compound 43 was
Table 2
Analog design for hit optimization
No.
R
LigScore2
2
3
4
5
6
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
Cl
Cl
Cl
5.81
5.43
5.97
6.08
6.01
5.80
5.50
6.09
5.77
5.86
5.81
5.66
5.90
5.22
5.83
6.09
5.69
5.48
5.49
5.66
5.93
5.71
6.04
5.95
6.05
5.94
5.99
6.01
5.94
5.75
6.14
5.81
5.53
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
F
Br
I
Cl
Cl
I
Br
OMe
Cl
OMe
OMe
OMe
OMe
OMe
Br
Cl
I
I
I
I
OMe
Cl
OMe
OEt
II
II
III
III
III
IV
IV
V
V
VI
VI
VII
VIII
Cl
Cl
Cl
Cl
H
Me
Cl

3156
K. Lee et al. / Bioorg. Med. Chem. 17 (2009) 3152–3161
Scheme 1. Synthesis of (S)-pyrrolidine-2-carboxylic acid [4-(5-chloro-benzo[1,3]dioxol-4-ylamino)-7-(2-methoxy-ethoxy)-quinazolin-6-yl]-amide.
Figure 4. (a) Ribbon diagram illustrating the crystal structure of human Src,13 with compound 43 bound to the active site. (b) Two-dimensional representation of the
interacting mode of compound 43 with Src. The quinazoline ring occupies the adenine binding site. The H-bond interaction between quinazoline N1 and backbone NH of
Met341 is presented as a dotted line. (c) Graphical representation of the binding mode of compound 43 to the ATP-binding site of Src. For clarity, only a few residues are
displayed, and hydrogen bond interactions are represented by yellow dashed lines.

K. Lee et al. / Bioorg. Med. Chem. 17 (2009) 3152–3161
3157
Table 3
established as a new potential lead for Src-selective inhibition on
Selectivity profile of compound 43
the basis of LigScore2 docking scores. The benzodioxole heterocy-
cle is buried in the hydrophobic pocket containing Met314,
Val323, Ile336, Leu393, Ala403, and Phe405. This potential lead is
expected to possess strong selectivity for Src. The compound was
synthesized through the route outlined in Scheme 1, and details
of the synthetic procedures and structural characterizations are
described in Section 4.
Compound
Enzyme (
lM)
Src
EGFR
VEGFR-2
43
Dasatinib
AZD-0530
0.089
0.014
0.003
7.13
2.87
2.60
No inhibition
2.36
21.00
2.4. Binding models
Src signaling is a key pathway during normal and dysregulated
bone functioning, and bone metastases are responsible for sub-
stantial morbidity in advanced prostate and melanoma cancers.
Thus, Src inhibition represents a potentially useful therapeutic
strategy for patients in various stages of prostate cancer. PC3 cells
are derived from a human prostate cancer bone metastasis. Onco-
genic signaling of EGFR in A431 cells may contribute to the ele-
vated activation of Src family kinases. PC3 and A431 cells were
treated with various concentrations of compound 43 to assess its
anti-proliferative effects. We observed dose-dependent inhibitory
effects on the growth of PC3 and A431 cells (Fig. 5a). Moreover,
inhibition of EGF-stimulated total cellular tyrosine phosphoryla-
tion in A431 cells was evident at a submicromolar level of com-
pound 43 (0.78 lM). However, the compound exerted no
inhibitory effects on SK-Br3, a cell line in which the HER2 protein
is highly expressed, or MCF-7, a human breast adenocarcinoma cell
line (<10 lM, data not shown). As extracellular signal-regulated ki-
nase (ERK), which acts downstream of Src kinase, is involved in cell
proliferation, we investigated whether compound 43 affects the
The lead compound 43 was docked onto the Src kinase-binding
site, and regions of favorable protein–ligand interactions were
visually analyzed. The results indicate that the compound docks
well and interacts with important residues of the Src kinase-bind-
ing site (Fig. 4). Whereas hydrogen bonding is essential for this
mode of inhibition, the number of hydrogen bonds between the
protein and ligand is not correlated with potency.13
We additionally used a binding model and conformation analy-
sis to determine the importance of interactions between inhibitors
and enzymes. The docking structure of the active Src kinase do-
main in complex with compound 43 shows that the quinazoline
ring occupies the adenine binding site, similar to results derived
from crystallographic analyses (Fig. 4a). The Src kinase domain
adopts an elongated bilobed structure typical of all kinases.
Binding Mode of quinazolines in Src kinase has been exten-
sively studied.13,25 A highly conserved hydrogen bond is created
by N1 of the quinazoline, which binds to the backbone amide of
Met341. In addition, the compound binds in the predicted orienta-
tion and satisfies the expected hydrophobic and aromatic interac-
tions (Fig. 4b and c). As predicted based on information about
quinazoline inhibitors, a bulky C-4 aniline is necessary to target
Src kinase.7 The aniline NH is not involved in direct hydrogen-
bonding interactions with the protein. The chlorobenzodioxole
moiety fits tightly into the hydrophobic pocket formed by
Met314, Val323, Ile336, Leu393, Ala403, and Phe405, as shown
in Figure 3b. In this study, we further investigate a novel series
of anilinoquinazolines substituted at the C-6 position of the qui-
nazoline ring. The C-6 position allows access to the ATP ribose
binding site, and may thus provide the enzyme with additional
binding affinity. The pyrrolidine ring of the C-6 quinazoline ap-
pears to fit tightly in a hydrophobic pocket. Hydrophobic interac-
tions with the protein are predominant, with one additional
hydrogen bond between pyrrolidine NH and the side-chain car-
boxyl group of Asp348. Buried hydrophobic surface areas between
the protein and ligand in the pyrrolidine ring region may account
for the increase in binding affinity. Docking studies of potential
inhibitors onto our 3D model of Src revealed that heterocycles
linked to the C-6 position of the quinazoline fit the shape of the ri-
bose pocket. Moreover, the C-7 side chain of the molecule is direc-
ted towards the solvent-exposed region. These findings prove that
the solubilizing groups of flexible side chains enhance in vitro po-
tency, and optimize the physicochemical properties of this com-
pound series.27,28,6
2.5. Biological activities
To further establish compound 43 enzyme inhibitory proper-
ties, the compound was evaluated for selectivity against in-house
kinases, as shown in Table 3. The inhibitory activity of compound
43 on Src was more potent than those of EGFR (IC50 = 7.1
lM)
Figure 5. (a) Viability of PC3 and A431 cells treated with compound 43. Cells were
treated with the specified concentrations of compound 43 for 72 h, and viability
determined with the SRB assay. Effects of compound 43 on ERK-p90RSK phos-
phorylation in (b) PC3 and (c) A431 cells. Cells were treated with 0.1 lM compound
43 for the indicated times, lysed, and the levels of phosphorylated and total ERK and
and VEGFR-2 (no inhibition at 10 M). The compound was highly
l
selective for Src, compared to EGFR (IC50 ratio > 80-fold) and VEG-
FR-2 (IC50 ratio > 110-fold). These findings provide valuable infor-
mation for the design of selective and powerful ATP-competitive
inhibitors of Src kinases.
p90RSK proteins determined by Western blot analysis using specific antibodies.

3158
K. Lee et al. / Bioorg. Med. Chem. 17 (2009) 3152–3161
Table 4
of Src family kinases. ERK and p90RSK, downstream signals of Src,
contribute to cell proliferation, motility, and invasion of tumor
cells. In our experiments, compound 43 strongly suppressed cell
proliferation and inhibited phosphorylation of ERK and p90RSK
in a time-dependent manner in PC3 and A431 cells. These results
indicate that inhibition of Src kinase activity by compound 43 leads
to the downregulation of ERK and p90RSK phosphorylation, and
eventual suppression of PC3 and A431 cell proliferation.
Thus, we conclude that large database docking, in conjunction
with appropriate scoring and filtering processes, is useful in medic-
inal chemistry. This approach has reached an advanced stage,
where it may effectively contribute to the lead discovery process.
Further optimization and SAR studies to improve cellular and
in vivo potency are ongoing.
Pharmacokinetic properties of compound 43 in Sprague–Dawley rats
AUC (
lg h/mL)
Cmax
(
lM)
Tmax (h)
t1/2 (h)
F (%)
IV(5 mg/kg)
PO (10 mg/kg)
2.40
1.24
2.01
0.12
0.08
4.0
6.1
5.4
—
20.4
phosphorylation of ERK and p90RSK, a downstream signal of ERK,
in PC3 and A431 cells. Notably, compound 43 (0.1 M) attenuated
l
ERK and p90RSK phosphorylation in a time-dependent manner.
After 6 h, complete inhibition of ERK and p90RSK phosphorylation
was observed in PC3 and A431 cells (Fig. 5b and c). The results col-
lectively indicate that compound 43 strongly inhibits Src kinase
activity, leading to suppression of Src downstream signals, ERK
and p90RSK, and tumor cell proliferation.
4. Experimental
2.6. Rat pharmacokinetic studies
4.1. Generation of ligand and enzyme structures
Based on the initial screening data, a pharmacokinetic (PK) study
on compound 43 was performed in rats (Table 4). Upon administra-
tion of single doses of 10 mg/kg (iv and po), the compound displayed
peak plasma exposure of 0.4 mM (Cmax) with a Tmax of 1 h. Oral bio-
availability of 43 in rats was determined as 20%. Although this novel
compound is not yet considered to be a potential clinical candidate,
the molecular scaffold combined with specific biological properties
allows it to be classified as a promising hit for the further develop-
ment of effective tyrosine kinase inhibitors.
Virtual screening by docking requires an accurate three-dimen-
sional (3D) target structure. The crystal structure of Src kinase
domain bound to its inhibitor, AZD-0530, N-(5-chloro-1,3-ben-
zodioxol-4-yl)-7-[2-(4-methylpipe-razin-1-yl)ethoxy]-5-(tetrahy-
dro-2H-pyran-4-yloxy)qui-nazolin-4-amine, was obtained from
the Protein Data Bank (PDB entry 2H8H).13 Initially, hydrogen atoms
were added to the protein, assuming that all the residues were in the
neutral form. All water molecules in the active site were included.
The protein was subjected to minimization using the steepest des-
3. Conclusions
cent (gradient <0.1), and conjugate gradient algorithms (gradient
29
<0.01) with the CHARMM force field in INSIGHTII
.
The Src active site
A structure-based virtual screening approach has facilitated the
identification of compounds that effectively inhibit Src kinase
activity. A computational docking study was structured to opti-
mize the efficiency of each step in terms of time and costs. A virtual
screening application is presented, leading to the successful identi-
fication of novel nanomolar Src kinase inhibitors. The approach
employed structure-based screening of an in-house kinase data-
base. The software package, LIGANDFIT, was used for the identification
and visualization of protein–ligand interaction sites and database
searching.
was defined using the inhibitor, AZD-0530 N-(5-chloro-1,3-ben-
zodioxol-4-yl)-7-[2-(4-methylpiperazin-1-yl)ethoxy]-5-(tetrahy-
dro-2H-pyran-4-yloxy)quinazo-lin-4-amine, and included all
residues within a 10 Å radius from the center of the ligand. The pro-
tein active site conformations with the ligands were retrieved for
further analysis.
4.2. Compound collection
We analyzed the physicochemical profiles of 1,561,000 com-
pounds for virtual screening. Various commercially compounds
were compiled. Compound collections have been enriched and en-
larged through commercial available databases, namely, AMRI
(175,000), Asinex (358,000), Enamine (400,000), InterBioScreen
(322,000), Life Chemicals (296,000)30 and in-house (10,000) syn-
thesis. Although the docking protocol has been developed inter-
nally in several case studies, and is able to rank known inhibitors
for similar targets at the top of the hit list, in this case, we applied
all the post- processing filtering and reranking tasks recommended
to increase the hit rate in a docking study.31 To design compounds
with good oral bioavailability amenable for chronic oral adminis-
tration, we focused on the molecular properties that affect absorp-
tion, such as 2D polar surface area (PSA), AlogP, molecular weight
(MW), H-bond donor-acceptor (HBDA) properties, and number of
rotatable bonds.32 The physical properties and chemical similari-
ties were calculated with DS Accord for Excel 6 Add-in tools from
Accelrys Inc., San Diego, CA, USA.
LIGANDFIT docking analysis of BMS354825, AZD-0530, and SKI-
606 confirmed that these ligands have a reasonable fit into the
binding site of Src kinase. We identified a C-6 substituted quinaz-
oline compound as a novel strong inhibitor of protein Src kinase by
virtual screening. We selected the hit compound 4 (300 nM) as the
template for the design of new Src kinase inhibitors for further syn-
thetic work.
Substituents at position 4 of the heterocyclic core were synthe-
sized for optimization of anti-proliferative activity, and evaluated
as Src kinase inhibitors. We discovered highly potent and selective
inhibitors of Src through docking studies, synthesis, and enzyme-
based assays. Specifically, compound 43 inhibited Src activity with
enzyme-based assay. The docking structure of a low nanomolar
IC50 values (IC50 = 89 nM) in an complex of compound 43 and Src
disclosed a highly conserved hydrogen bond created by N1 of qui-
nazoline, which interacts with the backbone amide of Met341. In
addition, the pyrrolidine ring of the C-6 quinazoline appears to
fit tightly within a hydrophobic pocket. One additional hydrogen
bond between compound 43 and the side-chain carboxyl oxygen
of Asp348 of the protein is proposed. The buried hydrophobic sur-
face areas of both protein and ligand in the region of the pyrroli-
dine ring may account for the increased binding affinity.
4.3. LIGANDFIT docking and scoring
Virtual screening is one of the important strategies for hit iden-
tification and lead optimization. A schematic diagram representing
the virtual screening strategy is shown in Figure 6. LIGANDFIT from
Discovery Studio 1.7 version (Accelrys, San Diego, USA) was ap-
plied for all runs. Reference protein coordinates for docking were
Src signaling is a major pathway involved in tumor proliferation
and progression. Activation of Src signaling is responsible for con-
siderable morbidity in advanced prostate cancer. Oncogenic signal-
ing of EGFR in A431 cells may contribute to the elevated activation

K. Lee et al. / Bioorg. Med. Chem. 17 (2009) 3152–3161
3159
formed in 96-well polystyrene round-bottomed plates. Briefly,
test compounds were added to a plate in the presence of ATP
and specific substrates in kinase solution, and incubated for 1 h
at room temperature before the addition of protease solution.
After the application of stabilizer solution, fluorescence polariza-
tion values were measured at 485/530 nm using a fluorescence
reader. All compounds were dissolved in DMSO and diluted in
buffer, yielding a final DMSO concentration of 1% in the assay.
4.6. In vitro growth inhibition assay
PC3 cells were obtained from the Korean Cell Line Bank (Seoul,
Republic of Korea). A431 and SK-BR3 cells were obtained from the
American Type Culture Collection (Rockville, MD). PC3 and SK-BR3
cells were maintained in RPMI 1640 supplemented with 10% fetal
bovine serum (FBS, Gibco BRL), 100 units/mL penicillin, and
100 lg/mL streptomycin sulfate. A431 and MCF-7 cells were main-
tained in DMEM containing 10% FBS. For assessment of cell-based
potency, cells were plated on a 96-well plate (Corning costar) at
densities ranging from 10,000 to 15,000 cells/well. The microtiter
plates were incubated at 37 °C under 5% CO2. After 24 h, the drug
was added at the indicated concentrations. After 72 h, reactions
were terminated by the addition of cold 10% trichloro-acetic acid.
Plates were washed with flowing deionized water and air-dried.
Sulforhodamine B solution (0.4% w/v in 1% acetic acid) was added
to each well, and unbound dye washed out. The bound stain was
subsequently solubilized with 10 mM Trizma base, and the absor-
bance read on an automated plate reader. Growth inhibition of 50%
(GI50) was calculated, based on the following equation: [ (Ti ꢀ Tz)/
(C ꢀ Tz)] ꢁ 100 = 50 (Ti: test growth, Tz: time zero, C: control
growth).
Figure 6. Schematic representation of the virtual screening strategy.
4.7. Western blot analysis
obtained from the X-ray structures of Src kinase complexed with 4-
amino-quinazoline. The docking site search was performed in the
shape-based mode based on the crystal structure of the compound,
4-amino-quinazoline. All calculations were performed with the CFF
1.01 force field. Conformations were generated with Monte Carlo
simulations (10,000 trials), and Flexible fit was selected.
The grid resolution was set to 0.5 Å. The default rigid body min-
imization parameters were used to dock the four orientations of
each conformation into the site, followed by 500-step final minimi-
zation of each docked ligand. Scoring was performed for each of the
10 saved ligand conformations using a set of functions, as imple-
mented in Discovery Studio 1.7, including LigScore1, LigScore2,
PLP1, PLP2, JAIN, PMF, and LUDI.33–35
PC3 and A431 cells were cultured in 6 cm dishes for 48 h, fol-
lowed by treatment with compound 43 (0.1 lM) for different
time-periods. Cell lysates were scraped and treated with lysis
buffer (10 mM Tris, pH 7.5; 150 mM NaCl; 5 mM EDTA; 1% Tri-
ton X-100; 1 mM DTT; 0.1 mM PMSF; 10% glycerol; and a prote-
ase inhibitor cocktail tablet) on ice, followed by centrifugation at
14,000 rpm for 10 min. Protein concentrations in the superna-
tant were determined using a dye-binding protein assay kit
(Bio-Rad Laboratories Hercules, CA), as described in the manu-
facturer’s manual. Lysate proteins were subjected to 10% sodium
dodecyl sulfate (SDS)–polyacrylamide gel electrophoresis
(PAGE), and electrophoretically transferred to PVDF membrane
(Millipore Corporation, Bedford, MA, USA). After blotting, the
membrane was blocked in 5% fat-free dry milk for 1 h, and incu-
bated with the specific primary antibody for 2 h at room tem-
perature. Protein bands were detected using the enhanced
chemiluminescence (ECL) detection kit (Amersham, Piscataway,
NJ) after hybridization with the HRP-conjugated secondary
antibody.
4.4. Virtual library construction and docking study
Our initial hits obtained with virtual screening belonged to
the anilinoquinazoline family (Fig. 3, compound 4). Virtual com-
pounds for hit-to-lead were generated from compound 4 with 33
aromatic ring groups at C-4 to ensure activity and selectivity for
Src kinase. Aromatic rings were selected with various substituent
patterns (Table 2).25 The molecular structure was minimized
with the CHARMM force field program (Accelrys, Inc. Sandiego,
USA). All docking simulations for the methods are presented
above. LigScore2 was applied for the selection of scoring func-
tions in further docking studies of Src kinase inhibitors into
the protein active site.
4.8. Rat pharmacokinetics
Adult male Sprague-Dawley rats were obtained from Orient
Co. (Gyeonggi-do, Korea), and had free access to pellet chow
(Picolab rodent diet 5053) and tap water. Throughout the exper-
iments, animals were housed in laminar flow cages (three per
cage) maintained at 22 2 °C, 50 20% relative humidity, under
a 12 h light/12 h dark cycle. Animals were kept in these facilities
for at least one week prior to the experiment. For the oral study,
rats were starved overnight prior to dosing and fed approximately
2 h after dosing. Rats were not subjected to fasting for the intra-
venous study. The vehicle for intravenous (IV) and oral dosing
4.5. In vitro Src kinase inhibition test
An enzyme selectivity screen was performed with the Src ki-
nase assay kit (Promega, Madison, USA). Reactions were per-

3160
K. Lee et al. / Bioorg. Med. Chem. 17 (2009) 3152–3161
was 23% ERi vehicle (PEG 200 with 0.02 5% cHCl/Tween 80 (20:3,
v/v). Blood samples were collected at various times after IV and
oral dosing. Approximately 0.3 mL of blood was collected from
the tail vein in tubes containing heparin, and plasma obtained
by centrifugation. Plasma samples were stored at ꢀ80 °C until
use. Samples were analyzed for compound 43 by LCMS. The time
elapsed to peak (Tmax) and maximum concentration (Cmax) were
4.9.5. (5-Chloro-benzo[1,3]dioxol-4-yl)-[7-(2-methoxy-ethoxy)-
6-nitro-quinazolin-4-yl]-amine
Potassium trimethylsilanoate (3.40 g, 26.5 mmol) was added to a
mixture of (5-chloro-benzo[1,3]dioxol-4-yl)-(7-fluoro-6-nitro-qui-
nazolin-4-yl)-amine (2.4 g, 6.62 mmol) and 2-methoxy-ethanol
(1.57 mL, 19.9 mmol) in dimethylsulfoxide (50 mL). After stirring
for 4 h, the reaction was quenched with water, and the mixture ex-
tracted with ethyl acetate (30 mL ꢁ 3). The organic layer was
washed with water (30 mL ꢁ 4) and brine (30 mL), and dried over
MgSO4. The filtrate was concentrated in vacuo. The residue was puri-
fied by flash chromatography to yield 5-chloro-benzo[1,3]dioxol-4-
yl)-[7-(2-methoxy-ethoxy)-6-nitro-quinazolin-4-yl]-amine (1,.44 g,
52%). 1H NMR (DMSO-d6) d: 10.18 (s, 1H), 9.30 (s, 1H), 8.52 (s, 1H),
7.52 (s, 1H), 7.08 (d, J = 7.5 Hz, 1H), 7.00 (d, J = 7.5 Hz, 1H), 6.25 (s,
2H), 4.43 (m, 2H), 3.75 (m, 2H), 3.31 (s, 3H).
obtained directly from the observed value. The half-life (t1/2
)
was calculated using the elimination rate constant (Ke), which,
in turn, was estimated by regression analysis from the slope of
the line. The half-life value was obtained from 0.693/Ke. The area
under the curve was computed by the linear trapezoidal rule
from 0 to 5 h (AUC5). The value of AUC was calculated as
1
AUC5 + [plasma concentration at 5 h after dosing/elimination rate
constant].
4.9. Chemistry
4.9.6. N4-(5-Chloro-benzo[1,3]dioxol-4-yl)-7-(2-methoxy-etho-
xy)-quinazoline-4,6-diamine
4.9.1. 7-Fluoro-3H-quinazolin-4-one
Concentrated HCl (0.115 mL) was added to a mixture of iron
(961 mg, 17.2 mmol) in 50% ethanol (aq) (20 mL). The mixture
was heated to reflux for 1 h. 5-Chloro-benzo[1,3]dioxol-4-yl)-[7-
(2-methoxy-ethoxy)-6-nitro-quinazolin-4-yl]-amine (1.44 g, 3.43
mmol) was added to the reaction mixture and stirred continuously
for 1 h at the same temperature. The reaction mixture was hot-fil-
tered through Celite, and the filtrate cooled to room temperature.
The mixture was made basic with saturated NaHCO3 (aq), and ex-
tracted with ethyl acetate (20 mL ꢁ 3). The organic layer was
washed with water and brine, and dried over MgSO4. Next, the
mixture was concentrated in vacuo to yield N4-(5-chloro-
benzo[1,3]dioxol-4-yl)-7-(2-methoxy-ethoxy)-quinazoline-4,6-dia-
mine (1.25 g, 95%). 1H NMR (DMSO-d6) d: 9.12 (s, 1H), 8.12 (s, 1H),
7.35 (s, 1H), 7.13 (s, 1H), 7.08 (d, J = 7.5 Hz, 1H), 6.80 (d, J = 7.5 Hz,
1H), 6.05 (s, 2H), 5.23 (s, 2H), 4.43 (m, 2H), 3.75 (m, 2H), 3.31
(s, 3H).
A catalytic amount (1 ml) of N,N-dimethylamide was added to a
solution of 2-amino-4-fluorobenzoic acid (50 g, 322 mmol) and
formamide (77 mL, 1934 mmol), and the resulting solution stirred.
The solution was heated to 180 °C and stirred for 14 h. The solution
was cooled to room temperature, and 300 mL of distilled water
added. The resulting mixture was stirred for about 30 min, and fil-
tered to obtain 7-fluoro-3H-quinazolin-4-one (41.3 g, yield: 78%).
1H NMR (DMSO-d6) d: 7.47–7.35 (m, 2H), 8.20–8.13 (m, 2H),
11.85 (br s, 1H).
4.9.2. 7-Fluoro-6-nitro-3H-quinazolin-4-one
7-Fluoro-3H-quinazolin-4-one (25 g, 152 mmol) was added to a
mixture of concentrated sulfuric acid (50 mL) and fuming nitric
acid (51 mL) at 0 °C. The resulting solution was stirred at room
temperature for 1 h, heated to 110 °C and stirred for 2 h. The solu-
tion was cooled to room temperature, and 300 mL of ice water
added. The resulting mixture was stirred for about 30 min, and fil-
tered to obtain 7-fluoro-6-nitro-3H-quinazolin-4-one (25 g, yield:
79%). 1H NMR (CDCl3) d: 7.79 (d, 1H), 8.32 (s, 1H), 8.72 (d, 1H),
12.83 (br s, 1H).
4.9.7. (S)-2-[4-(5-Chloro-benzo[1,3]dioxol-4-ylamino)-7-(2-me-
thoxy-ethoxy)-quinazolin-6-ylcarbamoyl]-pyrrolidine-1-carbo-
xylic acid tert-butyl ester
N4-(5-chloro-benzo[1,3]dioxol-4-yl)-7-(2-methoxy-ethoxy)-qui-
nazoline-4,6-diamine (200 mg, 0.342 mmol) was added to the reac-
4.9.3. 4-Chloro-7-fluoro-6-nitro-quinazoline
tion mixture of Boc-L-proline (219 mg, 0.512 mmol) and 1-(3-
7-Fluoro-6-nitro-3H-quinazolin-4-one (20 g, 96 mmol), thionyl
chloride (170 mL), phosphorous oxychloride (30 mL) and N,N-
dimethylformamide (1 mL) were added to a reaction bottle and
stirred. The resulting solution was heated to 100 °C until the solid
was dissolved, followed by stirring for 2 h. The reaction was cooled
to room temperature, and the solvent removed under reduced
pressure. The resulting residue was re-distilled after adding
300 mL of toluene, and the procedure repeated three times to ob-
tain 4-chloro-7-fluoro-6-nitro-quinazoline (21 g, yield: 99%). 1H
NMR (CDCl3) d: 7.73 (d, 1H), 8.30 (s, 1H), 8.72 (d, 1H).
diethylaminopropyl)-3-ethylcarbodiimide HCl (295 mg, 0.684
mmol) in pyridine (3 mL). The mixture was stirred for 2 h, and the
reaction quenched with water. Next, the mixture was extracted with
ethyl acetate, and the organic layer was washed with brine, dried over
MgSO4, and evaporated in vacuo. The residue was purified by flash
chromatography to yield 2-[4-(5-chloro-benzo[1,3]dioxol-4-ylami-
no)-7-(2-methoxy-ethoxy)-quinazolin-6-ylcarbamoyl]-pyrrolidine-
1-carboxylic acid tert-butyl ester (150 mg, 75%). 1H NMR (CDCl3) d:
9.08 (s, 1H), 8.60 (s, 1H), 7.26 (m, 1H), 7.03 (s, 1H), 6.95 (d,
J = 7.5 Hz, 1H), 6.65 (d, J = 7.5 Hz, 1H), 6.00 (s, 2H), 4.33 (m, 3H),
3.85 (m, 2H), 3.50 (m, 5H), 2.30–1.80 (m, 4H), 1.70–1.40 (m, 9H).
4.9.4. (5-Chloro-benzo[1,3]dioxol-4-yl)-(7-fluoro-6-nitro-
quinazolin-4-yl)-amine
4.9.8. (S)-Pyrrolidine-2-carboxylic acid [4-(5-chloro-benzo[1,
3]dioxol-4-ylamino)-7-(2-methoxy-ethoxy)-quinazolin-6-yl]-
amide
4-Chloro-7-fluoro-6-nitro-quinazoline (1.67 g, 7.42 mmol) and
5-chloro-benzo-[1,3]dioxloe-4-ylamine23 (1.27 g, 7.42 mmol) were
added to 2-propanol (15 mL). The reaction mixture was heated to
reflux for 2 h after the addition of 6 N HCl (1.5 mL), cooled to room
temperature, and made basic with saturated NaHCO3 (aq). Next,
the mixture was extracted with ethyl acetate (20 mL ꢁ 3), and
washed with brine. The organic layer was dried over MgSO4. The
filtrate was concentrated in vacuo to yield 5-chloro-benzo[1,3]diox-
ol-4-yl)-(7-fluoro-6-nitro-quinazolin-4-yl)-amine (2.4 g, 89%). 1H
NMR (DMSO-d6) d: 10.60 (s, 1H), 10.11(d, J = 8.1 Hz, 1H), 8.61 (s,
1H), 8.32 (s, 1H), 7.12 (d, J = 6 Hz, 1H), 7.00 (d, J = 6 Hz, 1H), 6.10
(s, 2H).
2-[4-(5-Chloro-benzo[1,3]dioxol-4-ylamino)-7-(2-methoxy-
ethoxy)-quinazolin-6-ylcarbamoyl]-pyrrolidine-1-carboxylic acid
tert-butyl ester (150 mg, 0.256 mmol) was diluted with dichloro-
methane (3 mL). Trifluoroacetic acid (3 mL) was added drop-wise
into the reaction mixture, followed by stirring for 1 h. The solvent
was removed through evaporation in vacuo. The residue was di-
luted with ethyl acetate (10 mL) and made basic with saturated
NaHCO3 (aq). Layers were separated, the organic layer washed
with brine and dried over MgSO4. The mixture was concentrated
in vacuo to yield pyrrolidine-2-carboxylic acid [4-(5-chloro-

K. Lee et al. / Bioorg. Med. Chem. 17 (2009) 3152–3161
3161
14. Boschelli, D. H.; Ye, F.; Wang, Y. D.; Dutia, M.; Johnson, S. L.; Wu, B.; Miller, K.;
Powell, D. W.; Yaczko, D.; Young, M.; Tischler, M.; Arndt, K.; Discafani, C.;
Etienne, C.; Gibbons, J.; Grod, J.; Lucas, J.; Weber, J. M.; Boschelli, F. J. Med.
Chem. 2001, 44, 3965.
15. Paul William Manley, P. W.; Cowan-Jacob, S. W.; Mestan, J. Biochim. Biophys.
Acta 2005, 1754, 3.
benzo[1,3]dioxol-4-ylamino)-7-(2-methoxy-ethoxy)-quinazolin-6-
yl]-amide (60 mg, 48%). 1H NMR (CD3OD) d: 8.95 (s, 1H), 8.31 (s,
1H), 7.26 (s, 1H), 7.03 (d, J = 7.5 Hz, 1H), 6.85 (d, J = 7.5 Hz, 1H),
6.01 (s, 2H), 4.39 (m, 2H), 4.10 (m, 1H), 3.91 (m, 2H), 3.46 (s,
3H), 3,29 (m, 2H), 2.30 (m, 1H), 2.00–1.87 (m, 3H).
16. Shoichet, B. K. Nature 2004, 432, 862.
17. Oprea, T. I.; Matter, H. Curr. Opin. Chem. Biol. 2004, 8, 349.
18. Bajorath, J. Nat. Rev. Drug Disc. 2002, 1, 882.
19. Schneider, G.; Bohm, H. J. DDT 2002, 7, 64.
Acknowledgment
20. Ajay, A. J.; Murcko, M. A. J. Med. Chem. 1995, 38, 4953.
21. Yu, H.; Wang, Z.; Zhang, L.; Zhang, J.; Huang, Q. Chem. Biol. Drug Des. 2007, 69,
204.
This work was supported by a Research Program for New Drug
Target Discovery grant (M10601000153-07N0100-15310).
22. Dalgarno, D.; Stehle, T.; Narula, S.; Schelling, P.; van Schravendijk, M. R.;
Adams, S.; Andrade, L.; Keats, J.; Ram, M.; Jin, L.; Grossman, T.; MacNeil, I.;
Metcalf, C., III; Shakespeare, W.; Wang, Y.; Keenan, T.; Sundaramoorthi, R.;
Bohacek, R.; Weigele, M.; Sawyer, T. Chem. Biol. Drug Des. 2006, 67, 46.
23. Breitenlechner, C. B.; Kairies, N. A.; Honold, K.; Scheiblich, S.; Koll, H.; Greiter,
E.; Koch, S.; Schafer, W.; Huber, R.; Engh, R. A. J. Mol. Biol. 2005, 353, 222.
24. Cowan-Jacob, S. W.; Fendrich, G.; Manley, P. W.; Jahnke, W.; Fabbro, D.;
Liebetanz, J.; Meyer, T. Structure 2005, 13, 861.
25. Ple, P. A.; Green, T. P.; Hennequin, L. F.; Curwen, J.; Fennell, M.; Allen, J.;
Lambert-van der Brempt, C.; Costello, G. J. Med. Chem. 2004, 47, 871.
26. Boschelli, D. H.; Wu, B.; Barrios Sosa, A. C.; Durutlic, H.; Chen, J. J.; Wang, Y.;
Golas, J. M.; Lucas, J.; Boschelli, F. J. Med. Chem. 2005, 48, 3891.
27. Hardcastle, I. R.; Arris, C. E.; Bentley, J.; Boyle, F. T.; Chen, Y. H.; Curtin, N. J.;
Endicott, J. A.; Gibson, A. E.; Golding, B. T.; Griffin, R. J.; Jewsbury, P.; Menyerol,
J.; Mesguiche, V.; Newell, D. R.; Noble, M. E. M.; Pratt, D. J.; Wang, L. Z.;
Whitfield, H. J. J. Med. Chem. 2004, 47, 3710.
References and notes
1. Hunter, T.; Sefton, B. M. Proc. Natl. Acad. Sci. U.S.A. 1980, 77, 1311.
2. Schwartzberg, P. L. Oncogene 1998, 17, 1463.
3. Thomas, S. M.; Brugge, J. S. Annu. Rev. Cell Dev. Biol. 1997, 13, 513.
4. Boggon, T. J.; Eck, M. J. Oncogene 2004, 23, 7918.
5. Parsons, S. J.; Parsons, J. T. Oncogene 2004, 23, 7906.
6. Summy, J. M.; Gallick, G. E. Cancer Metastasis Rev. 2003, 22, 337.
7. Zhang, Q.; Thomas, S. M.; Xi, S.; Smithgall, T. E.; Siegfried, J. M.; Kamens, J.;
Gooding, W. E.; Grandis, J. R. Cancer Res. 2004, 64, 6166.
8. Golas, J. M.; Lucas, J.; Etienne, C.; Golas, J.; Discafani, C.; Sridhara, L.; Boghaert,
E.; Arndt, K.; Ye, F.; Boschelli, D. H.; Li, F.; Titsch, C. Cancer Res. 2005, 65, 5358.
9. Brunton, V. G.; Avizienyte, E.; Fincham, V. J.; Serrels, B.; Metcalf, C. A., III;
Sawyer, T. K.; Frame, M. C. Cancer Res. 2005, 65, 1335.
10. Lesslie, D. P.; Summy, J. M.; Parikh, N. U.; Trevino, J. G.; Sawyer, T. K.; Metcalf, C. A.;
Shakespeare, W. C.; Hicklin, D. J.; Ellis, L. M.; Gallick, G. E. Br. J. Cancer 2006, 94, 1710.
11. Sawyer, T. K.; Bohacek, R. S.; Metcalf, C. A., III; Shakespeare, W. C.; Wang, Y.;
Sundaramoorthi, R.; Keenan, T.; Narula, S.; Weigele, M.; Dalgarno, D. C.
Biotechniques 2003, 2–10, 12.
12. Lombardo, L. J.; Lee, F. Y.; Chen, P.; Norris, D.; Barrish, J. C.; Behnia, K.;
Castaneda, S.; Cornelius, L. A.; Das, J.; Doweyko, A. M.; Fairchild, C.; Hunt, J. T.;
Inigo, I.; Johnston, K.; Kamath, A.; Kan, D.; Klei, H.; Marathe, P.; Pang, S.;
Peterson, R.; Pitt, S.; Schieven, G. L.; Schmidt, R. J.; Tokarski, J.; Wen, M. L.;
Wityak, J.; Borzilleri, R. M. J. Med. Chem. 2004, 47, 6658.
28. Hennequin, L. F.; Stokes, E. S.; Thomas, A. P.; Johnstone, C.; Ple´, P. A.; Ogilvie, D. J.;
Dukes, M.; Wedge, S. R.; Kendrew, J.; Curwen, J. O. J. Med. Chem. 2002, 45, 1300.
29. INSIGHT II, 97 Molecular Modelling Program Package; Accelrys, San Diego, CA, 1997.
30. AMRI;
www.amriglobal.com,
Asinex;
www.asinex.com,
Enamine;
www.enamine.net, InterBioScreen; www.ibscreen.com, Life Chemicals;
www.lifechemicals.com.
31. Boehm, H.-J.; Stahl, M.. In Reviews in Computational Chemistry; Wiley-VCH:
New York, 2002; Vol. 18, pp 41–87.
32. Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. Adv. Drug Delivery Rev.
1997, 23, 3.
13. Hennequin, L. F.; Allen, J.; Breed, J.; Curwen, J.; Fennell, M.; Green, T. P.;
Lambert-van der Brempt, C.; Morgentin, R.; Norman, R. A.; Olivier, A.;
Otterbein, L.; Ple, P. A.; Warin, N.; Costello, G. J. Med. Chem. 2006, 49, 6465.
33. Krovat, E. M.; Langer, T. J. Chem. Inf. Comput. Sci. 2004, 44, 1123.
34. Bissantz, C.; Folkers, G.; Rognan, D. J. Med. Chem. 2000, 43, 4759.
35. Wang, R.; Lu, Y.; Wang, S. J. Med. Chem. 2003, 46, 2287.