More than just a GPCR ligand: Structure-based discovery of thioridazine derivatives as Pim-1 kinase inhibitors

Article abstract of DOI:10.1039/c4md00030g

Pim-1 kinase is a serine/threonine kinase which plays an important role in cell proliferation and differentiation. The Pim-1 kinase expression is elevated in leukemia and prostate cancer. Accordingly, we employed a structure-based hierarchical virtual screening approach to identify potential unknown Pim-1 kinase activity for existing drugs. Among the LOPAC library of pharmacologically active compounds, one top-ranked drug molecule thioridazine, a well-known antipsychotic agent which exerted its biological function as a dopamine receptor antagonist, showed low micromolar activity in the Pim-1 enzymatic assay. We determined the co-crystal structure of thioridazine bound with Pim-1 kinase, and defined the key elements of the pharmacophore by analyzing the structure-activity relationship of thioridazine analogues. In addition, we also assessed our pharmacophore by successfully predicting the Pim-1 activity of the selective Akt inhibitor, 10-DEBC. Our discovery of the unknown Pim-1 inhibitory activity of thioridazine and 10-DEBC might provide novel insights into understanding their molecular mechanism of action, and inspire the computation-driven multiple-target drug discovery.

Full text of DOI:10.1039/c4md00030g

MedChemComm
View Article Online
View Journal | View Issue
CONCISE ARTICLE
More than just a GPCR ligand: structure-based
discovery of thioridazine derivatives as Pim-1
kinase inhibitors†‡
Cite this: Med. Chem. Commun., 2014,
5, 507
Wei Li,ab Xiaobo Wan,b Fanqi Zeng,b Yuting Xie,b Yanli Wang,b Wei Zhang,b Li Lib
ba
*
and Niu Huang
Pim-1 kinase is a serine/threonine kinase which plays an important role in cell proliferation and
differentiation. The Pim-1 kinase expression is elevated in leukemia and prostate cancer. Accordingly, we
employed a structure-based hierarchical virtual screening approach to identify potential unknown Pim-1
kinase activity for existing drugs. Among the LOPAC library of pharmacologically active compounds, one
top-ranked drug molecule thioridazine, a well-known antipsychotic agent which exerted its biological
function as a dopamine receptor antagonist, showed low micromolar activity in the Pim-1 enzymatic
assay. We determined the co-crystal structure of thioridazine bound with Pim-1 kinase, and defined the
key elements of the pharmacophore by analyzing the structure–activity relationship of thioridazine
analogues. In addition, we also assessed our pharmacophore by successfully predicting the Pim-1 activity
of the selective Akt inhibitor, 10-DEBC. Our discovery of the unknown Pim-1 inhibitory activity of
thioridazine and 10-DEBC might provide novel insights into understanding their molecular mechanism of
action, and inspire the computation-driven multiple-target drug discovery.
Received 23rd January 2014
Accepted 4th February 2014
DOI: 10.1039/c4md00030g
www.rsc.org/medchemcomm
The target binding prole of a drug molecule is essential for including c-KIT and PDGFR kinases, which led to its expanded
understanding its mechanism of action (MOA), as its off-target clinical applications in advanced or metastatic gastrointestinal
binding propensities may lead to better clinical efficacy in some stromal tumor (GIST).2,3
circumstances, while causing side effects in other cases. Many
Experimentally testing all drugs on all potential biomolecular
marketed drugs existing today were discovered serendipitously, targets is practically impossible; therefore, computationally
and the elucidation of their MOA has been usually retrospective. predicting the novel polypharmacology of drug molecules has
Notably, their therapeutic efficacies were oen attributed to recently attracted great interest in the eld of drug discovery.4–6
their multiple target actions. One example is clozapine, it is one One notable method is the development of a statistics-based
of the most successful atypical antipsychotic drugs, it was chemoinformatics approach, similarity ensemble approach
discovered half a century ago, and its clinical efficacy was (SEA), which has been applied in systematically relating the
reported to be associated with its promiscuous binding to ligand chemical similarity to biological targets.7–9 Complemen-
dozens of
G
protein-coupled receptors (GPCRs).1 Many tary to such ligand-based approaches, protein structure-based
compounds even designated as single-target drugs are in fact methods require the structures of known targets to predict
not as specic as originally expected, and have been discovered protein–ligand binding complexes, and to estimate the binding
to bind to other targets. For example, Gleevec was initially affinities, like the inverse docking protocol INVDOCK.10,11
designed to inhibit the abnormal tyrosine kinase BCR-ABL.
GPCRs and kinases are two most important drug target
Subsequently, it was shown to target several other kinases, families, and many of their ligands are promiscuous within
their own protein families. However, the ligand cross-reactivity
between kinase inhibitors and GPCR ligands has rarely been
reported. Interestingly, a well-known kinase drug, sorafenib,
was identied to bind to 5-hydroxytryptamine receptors
(5-HTRs) by computationally screening the FDA approved drug
molecules against the modelled structure of the 5-HT2A
receptor.12 Therefore, we were wondering whether we could
discover known GPCR ligands with unexpected kinase activities
by virtually screening GPCR ligands against therapeutically
relevant kinases.
aGraduate Program in Peking Union Medical College and Chinese Academy of Medical
Sciences, Beijing 100730, China
bNational Institute of Biological Sciences, Beijing No.
7 Science Park Road,
Zhongguancun Life Science Park, Changping, Beijing 102206, China. E-mail:
huangniu@nibs.ac.cn; Fax: +86-10-80720813; Tel: +86-10-80720645
† The authors have declared no conict of interest.
‡ Electronic supplementary information (ESI) available: Pim-1 and thioridazine,
Pim-1 and 10-DEBC complex crystallization, data collection, and structure
determination. Modelled complex structures of chlorprothixene analogues,
thioridazine, propionylpromazine and amitriptyline. Structural similarity
analysis. See DOI: 10.1039/c4md00030g
This journal is © The Royal Society of Chemistry 2014
Med. Chem. Commun., 2014, 5, 507–511 | 507

View Article Online
MedChemComm
Concise Article
The proviral integration site in Moloney murine leukemia docking hits (ESI, Fig. S1‡). Therefore, we measured the enzy-
virus (Pim) kinases are a family of proto-oncogenic serine/ matic inhibition activities of these two compounds using
threonine kinases regulating several signalling pathways that Homogeneous Time Resolved Fluorescence (HTRF) kinase
are fundamental to cancer development and progression.13 As assay as described previously.17 Thioridazine inhibited Pim-1
the most studied member of Pim kinases, Pim-1 kinase is highly kinase with an IC50 value of 6.89 mM (Fig. 2a) while the chlor-
expressed in a wide range of haematopoietic malignancies and prothixene analogue showed no inhibitory activity. Since the
solid cancers, especially in leukemia and prostate cancer.13,14 commercially available thioridazine is a mixture of racemic
Pim-1 kinase adopts a typical kinase fold consisting of enantiomers, we tested the puried forms of both enantiomers
N-terminal and C-terminal domains linked by a hinge region. obtained from Dr Jørn B. Christensen.26 The racemic mixture
However, the presence of residue P123 in the hinge region (À/+), the dextrorotatory form (+) and the levorotatory form (À)
disrupts the canonical hydrogen bonding pattern with the did not show difference in Pim-1 inhibition activity (Table 1 and
adenine moiety of ATP or bound inhibitors due to the lack of the ESI, Fig. S2‡). We also tested two additional dopamine receptor
hydrogen bond donor.15,16 This unique hinge region architec- antagonists, amitriptyline and propionylpromazine. Both
ture suggests that Pim-1 kinase inhibitors might be structurally compounds adopted similar docking poses to thioridazine (ESI,
distinguishable from many classical kinase inhibitors. There- Fig. S1‡); however, the rank of their MM-GB/SA energy scores
fore, we further asked whether we could discover the unex- were below the top 1%. None of the compounds showed activity.
pected Pim-1 kinase activity of GPCR ligands.
Thioridazine represents a novel class of Pim-1 inhibitors. We
In our continuous endeavor to computationally nd novel submitted thioridazine to the SEA server (http://sea.bkslab.org),
ligand cross-reactivity,12,17 we employed a structure-based hier- but none of the kinases were predicted to be thioridazine's
archical virtual screening strategy18,19 in combination with potential target based on ligand structural similarity. We also
experimental evaluation to explore the unknown Pim-1 kinase performed pair-wise similarity comparison of thioridazine
activity of existing drugs and chemical probes with known
biological activity (Fig. 1). Briey, we docked 1280 LOPAC library
compounds (Sigma-Aldrich Corp.) against the ATP binding site
of Pim-1 kinase (PDB ID:2J2I20) using the fast-compute docking
program DOCK 3.5.54.21 The automated docking pipeline as
described previously22,23 includes the sphere generation, scoring
grid and docking calculations. All generated docking poses were
subjected to the MM-GB/SA renement and rescoring proce-
dure.18,24 The energy minimizations were performed using the
Protein Local Optimization Program (PLOP) with an all-atom
molecular-mechanics (MM) force eld and a generalized born
surface area (GB/SA) implicit solvent model with variable
internal dielectric constant.25 The ligand binding energy was
Fig. 2 (a) Dose–response inhibition of Pim-1 by thioridazine. The IC50
computed by subtracting the energies of the optimized free
ligand in solution and the free protein in solution from the
optimized ligand–protein complex's energy in solution,
accounting for protein–ligand interaction energy, desolvation
energies of ligand and protein, and ligand strain energy. Finally,
we visually examined the top 1% ranked poses, and chose the
compounds of interest for experimental testing.
value was determined from three independent tests. (b) Fitting
dextrorotatory form (+) of thioridazine (gray) into an electron density
map in the ATP binding-site of Pim-1. The sigma-A weighted 2Fo–Fc
electron density map contoured at a level of 1.0s. Coordinates and
detailed methods for the solved crystal structure were deposited to the
PDB with the accession ID 4IAA. Molecular images were generated
with PyMOL.27
Interestingly, two well-known GPCR ligands, thioridazine
and chlorprothixene analogues, were identied in the top 1% of
Table 1 Inhibition of Pim-1 by thioridazine derivatives
Compound name
IC50a (mM)
Thioridazine (À/+)
Thioridazine (À)
Thioridazine (+)
NCI186055
NCI186058
NCI64076
b1
b2
b3
b4
10-DEBC
6.89 Æ 1.51
7.39 Æ 2.59
7.03 Æ 2.52
8.16 Æ 2.47
6.73 Æ 0.45
34.0 Æ 7.2% (100 mM)
9.7 Æ 2.1% (100 mM)
22.6 Æ 2.9% (100 mM)
17.6 Æ 8.7% (100 mM)
8.70 Æ 3.39
1.28 Æ 0.28
a
Data are shown as (mean values Æ SD) from three independent
determinations.
Fig. 1 Flowchart of our structure-based virtual screening strategy.
508 | Med. Chem. Commun., 2014, 5, 507–511
This journal is © The Royal Society of Chemistry 2014

View Article Online
Concise Article
MedChemComm
against the co-crystal inhibitors of Pim-1 using ChemMine Web
Tools.28 None of the calculated Tanimoto coefficients are greater
than 0.35 (ESI, Table S1‡). In addition, the structural similarity
analysis against all reported 669 Pim-1 inhibitors in the
ChEMBL database did not identify any structural analogues of
thioridazine. Therefore, thioridazine could not be predicted as a
Pim-1 inhibitor using these ligand-based approaches.
Next, we determined the X-ray crystal complex structure of
˚
Pim-1 kinase bound with thioridazine at 2.85 A resolution
(Fig. 2b). Note that we used the racemic mixture in the crystal-
lization study due to the inadequate availability of pure enan-
tiomers, and we tted the dextrorotatory form (+) of
thioridazine into the electron density map in the ATP-site. The
detailed experimental procedure is included in the ESI (Table
S2‡). The overall structure of the Pim-1 kinase resembles pub-
lished crystal structures. The crystal binding pose of thiorida-
zine is consistent with our docking prediction with an RMSD
˚
value of 3 A using the dextrorotatory form for comparison. The
electron density of the phenothiazine group in thioridazine is
very clear, it interacts favorably with hydrophobic residues V52,
L174 and I185 in the ATP binding-site, while no hydrogen bond
is formed between thioridazine and the kinase hinge region.
Nevertheless, the weak electron density map features at the
aliphatic chain and the piperidine ring of thioridazine suggest
the conformational diversity and/or the ambiguous congura-
tions of the piperidine ring in the racemic mixture.
Fig. 3 (a) The chemical structures of the three thioridazine derivatives
obtained from the National Cancer Institute. (b) Synthesis of pheno-
thiazine derivatives. General synthetic approaches. Reagents: (1) sulfur,
I2, 170 ꢀC, 2 h, 60%; (2) 1,4-dibromobutane, NaH, 25 ꢀC, DMF, over-
night, 78%; (3) NaN3, DMSO–H2O (5 : 1), 25 ꢀC, 48 h, 55%; H2, Pd/C,
MeOH, 50 ꢀC overnight, 50%; (4) NaI, NH3$H2O, CsCO3, DMF, 80 ꢀC, 2
h, 80%. (c) Chemical structure of 10-DEBC.
To explore the key elements of the binding pharmacophore
of thioridazine, we obtained three thioridazine analogues from
the National Cancer Institute (NCI) compound library.
Compounds NCI186055 and NCI186058, with a chloro or bromo the tricyclic scaffolds in both compounds overlap to a great
group replacing the methylsulfanyl group on the phenothiazine extent while the methylsulfanyl group in thioridazine and the
ring of thioridazine, exhibited similar inhibitory activity to chloro group in 10-DEBC point in opposite directions. At the
˚
thioridazine (Fig. 3a and Table 1). This result was consistent same crystallographic resolution of 2.85 A, the electron density
with the crystallographic position of the methylsulfanyl group, map in the 10-DEBC bound structure is sufficiently resolved to
which did not form direct contacts with binding-site residues by place the aliphatic chain and diethylamino group. The charged
extending into the solvent (Fig. 2b). In contrast, compound tertiary amine group in 10-DEBC forms a more favorable elec-
˚
NCI64076 was nearly inactive (Fig. 3a and Table 1). The major trostatic interaction with residue D128 than thioridazine (3.5 A
˚
difference between NCI64076 and thioridazine was the length of in 10-DEBC vs. 4.5 A in thioridazine), which may indicate the
the aliphatic chain bridging the piperidine ring and the structural basis of strong inhibitory activity of 10-DEBC.
phenothiazine scaffold, which may largely reduce the favorable
Thioridazine is an antipsychotic drug known to exert its
electrostatic interaction between the positively charged biological function as a dopamine receptor antagonist. Inter-
nitrogen atom in the piperidine group and negatively charged estingly, thioridazine has recently been found to selectively
residue D128. To further assess the contribution of the aliphatic target the neoplastic cells, and impair human somatic cancer
amines, we synthesized and tested compound b4 along with stem cells capable of in vivo leukemic disease initiation without
three intermediates (Fig. 3b and Table 1). All three intermediate showing effects on normal blood stem cells.30 The authors
compounds were inactive while only compound b4 was as active proposed that dopamine receptors might serve as biomarkers
as thioridazine, which could support the basic components of involved in diverse malignancies. However, it is not clear
the pharmacophore: a tricyclic ring scaffold occupying the whether thioridazine's Pim-1 kinase inhibition activity might
hydrophobic ATP binding-site, a bridging aliphatic chain with also contribute to its effect of eradiating cancer stem cells.
suitable length and an aliphatic amine group interacting with Several pieces of experimental evidence are available to support
negatively charged residue D128.
our hypothesis. Firstly, Pim-1 kinase has been shown to express
Additionally, we noticed that one widely used selective Akt in haematopoietic stem cells and embryonic stem cells.13 Pim-1
inhibitor, 10-DEBC29 (Fig. 3c), satised our designated Pim-1 mediates the homing and migration of hematopoietic cells
pharmacophore. As expected, 10-DEBC showed strong inhibi- through phosphorylation of chemokine receptor 4 (CXCR4).31
tory activity with an IC50 value of 1.28 mM. We also determined Pim-1 kinase is also involved in mouse embryonic stem cell self-
the X-ray crystal complex structure of 10-DEBC bound with Pim- renewal.32 Secondly, Pim-1 plays an important physiological role
1 (Fig. 4 and Table S2‡). Comparing two co-crystal structures, in c-Myc driven tumorigenesis via effective oncogene
This journal is © The Royal Society of Chemistry 2014
Med. Chem. Commun., 2014, 5, 507–511 | 509

View Article Online
MedChemComm
Concise Article
Pim-1 kinase by a structure-based hierarchical virtual screening
approach. The cocrystal structure, SAR study and pharmaco-
phore evaluation indicated that a tricyclic scaffold with a suit-
able aliphatic side chain bridged tertiary amine might present a
novel class of Pim-1 inhibitors. Moreover, a selective Akt
inhibitor, 10-DEBC, possessing the key pharmacophore
elements was shown to exhibit strong in vitro Pim-1 activity.
Therefore, it is intriguing whether the Pim-1 inhibitory activity
of thioridazine and 10-DEBC contributes to their mode of
action. We hope that our present results will not only provide a
novel scaffold for Pim-1 inhibitor design, but also inspire the
computation-driven multiple-target drug discovery.
Acknowledgements
We thank the National Cancer Institute (NCI) for providing the
compounds. We thank Dr Jørn B. Christensen (University of
Copenhagen, Denmark) for generously providing us the pure
Fig. 4 The detailed view of the Pim-1 kinase domain with 10-DEBC
(PDB code: 4MED) compared with thioridazine (PDB code: 4IAA). The
protein is shown in a green cartoon representation. The compound enantiomers of thioridazine. We also thank Dr Andrew Chris-
10-DEBC is represented as orange sticks and thioridazine as cyan
sticks. Several important binding site residues are highlighted. The
sigma-A weighted 2Fo–Fc electron density map contoured at a level
of 1.0s.
tofferson for proofreading. Computational resource was sup-
ported by the Supercomputing Center of Chinese Academy of
Sciences (SCCAS). Financial support from the Chinese Ministry
of Science and Technology “973” Grant 2011CB812402 (to NH).
collaboration, where the elevated expression level of Pim-1 is
oen observed in the context of increased c-Myc levels.13 Pim-1
References
phosphorylates c-Myc to stabilize c-Myc in vivo and to enhance
the c-Myc transforming activity,33 where c-Myc exerts a critical
function in both self-renewal and differentiation of stem cells
and early progenitor cells.34 Thirdly, elevated levels of Pim-1
kinase have been observed in leukemia cancer cells.35 Never-
theless, the mechanism of Pim-1 inhibition by thioridazine in
cancer stem cell differentiation requires further elucidation.
In addition, thioridazine was reported to induce apoptosis in
ovarian, cervical and endometrial cancer cells by targeting the
Akt/mTOR signalling pathway, where the phosphorylation
levels of Akt and 4E-BP1 were decreased by thioridazine treat-
ment.36,37 Notably, both Akt and Pim-1 kinases share several
common substrates that are involved in apoptosis.38 Pim-1
inhibitor has recently been reported to reduce the phosphory-
lation of Akt.39 4E-BP1, one of the best characterized targets of
the mTOR complex, was identied as a substrate of Pim-1
kinase.40 Therefore, it is likely that thioridazine induces
apoptosis by modulating the Akt/mTOR signalling pathway
through Pim-1 inhibition.
In the present study, we also predicted the Pim-1 activity of a
selective Akt inhibitor, 10-DEBC, based on our derived phar-
macophore from thioridazine derivatives. Our biochemical and
structural analysis validated its Pim-1 inhibitory activity.
10-DEBC was reported to be a micromolar inhibitor of Akt
(5 mM).29 However, it was also reported to induce autophagy
through an Akt-independent mechanism in neurons.41 Our
results suggest that the cellular functions of 10-DEBC might be
related to its Pim-1 inhibitory activity.
1 B. L. Roth, D. J. Sheffler and W. K. Kroeze, Nat. Rev. Drug
Discovery, 2004, 3, 353–359.
2 R. Capdeville, E. Buchdunger, J. Zimmermann and A. Matter,
Nat. Rev. Drug Discovery, 2002, 1, 493–502.
3 M. A. Fabian, W. H. Biggs, 3rd, D. K. Treiber, C. E. Atteridge,
M. D. Azimioara, M. G. Benedetti, T. A. Carter, P. Ciceri,
P. T. Edeen, M. Floyd, J. M. Ford, M. Galvin, J. L. Gerlach,
R. M. Grotzfeld, S. Herrgard, D. E. Insko, M. A. Insko,
A. G. Lai, J. M. Lelias, S. A. Mehta, Z. V. Milanov,
A. M. Velasco, L. M. Wodicka, H. K. Patel, P. P. Zarrinkar
and D. J. Lockhart, Nat. Biotechnol., 2005, 23, 329–336.
4 S. Peng, X. Lin, Z. Guo and N. Huang, Curr. Top. Med. Chem.,
2012, 12, 1363–1375.
5 J. B. Brown and Y. Okuno, Chem. Biol., 2012, 19, 23–28.
6 A. L. Hopkins, Nat. Chem. Biol., 2008, 4, 682–690.
7 M. J. Keiser, B. L. Roth, B. N. Armbruster, P. Ernsberger,
J. J. Irwin and B. K. Shoichet, Nat. Biotechnol., 2007, 25,
197–206.
8 M. J. Keiser, V. Setola, J. J. Irwin, C. Laggner, A. I. Abbas,
S. J. Hufeisen, N. H. Jensen, M. B. Kuijer, R. C. Matos,
T. B. Tran, R. Whaley, R. A. Glennon, J. Hert,
K. L. Thomas, D. D. Edwards, B. K. Shoichet and
B. L. Roth, Nature, 2009, 462, 175–181.
9 E. Lounkine, M. J. Keiser, S. Whitebread, D. Mikhailov,
J. Hamon, J. L. Jenkins, P. Lavan, E. Weber, A. K. Doak,
S. Cote, B. K. Shoichet and L. Urban, Nature, 2012, 486,
361–367.
10 Y. Z. Chen and C. Y. Ung, J. Mol. Graph. Modell., 2001, 20,
199–218.
In summary, a selective dopamine receptor antagonist,
thioridazine, was identied to be a low micromolar inhibitor of 11 Y. Z. Chen and D. G. Zhi, Proteins, 2001, 43, 217–226.
510 | Med. Chem. Commun., 2014, 5, 507–511
This journal is © The Royal Society of Chemistry 2014

View Article Online
Concise Article
MedChemComm
12 X. Lin, X. P. Huang, G. Chen, R. Whaley, S. Peng, Y. Wang, 29 K. N. Thimmaiah, J. B. Easton, G. S. Germain, C. L. Morton,
G. Zhang, S. X. Wang, S. Wang, B. L. Roth and N. Huang,
J. Med. Chem., 2012, 55, 5749–5759.
S. Kamath, J. K. Buolamwini and P. J. Houghton, J. Biol.
Chem., 2005, 280, 31924–31935.
13 M. C. Nawijn, A. Alendar and A. Berns, Nat. Rev. Cancer, 30 E. Sachlos, R. M. Risueno, S. Laronde, Z. Shapovalova,
2011, 11, 23–34.
14 N. Ogawa, H. Yuki and A. Tanaka, Expert Opin. Drug
Discovery, 2012, 7, 1177–1192.
15 A. Kumar, V. Mandiyan, Y. Suzuki, C. Zhang, J. Rice, J. Tsai,
D. R. Artis, P. Ibrahim and R. Bremer, J. Mol. Biol., 2005, 348,
183–193.
J. H. Lee, J. Russell, M. Malig, J. D. McNicol, A. Fiebig-
Comyn, M. Graham, M. Levadoux-Martin, J. B. Lee,
A. O. Giacomelli, J. A. Hassell, D. Fischer-Russell,
M. R. Trus, R. Foley, B. Leber, A. Xenocostas, E. D. Brown,
T. J. Collins and M. Bhatia, Cell, 2012, 149, 1284–
1297.
16 S. Schenone, C. Tintori and M. Botta, Curr. Pharm. Des., 31 R. Grundler, L. Brault, C. Gasser, A. N. Bullock, T. Dechow,
2010, 16, 3964–3978.
17 X. Wan, W. Zhang, L. Li, Y. Xie, W. Li and N. Huang, J. Med.
Chem., 2013, 56, 2619–2629.
S. Woetzel, V. Pogacic, A. Villa, S. Ehret, G. Berridge,
A. Spoo, C. Dierks, A. Biondi, S. Knapp, J. Duyster and
J. Schwaller, J. Exp. Med., 2009, 206, 1957–1970.
18 N. Huang, C. Kalyanaraman, J. J. Irwin and M. P. Jacobson, 32 I. Aksoy, C. Sakabedoyan, P. Y. Bourillot, A. B. Malashicheva,
J. Chem. Inf. Model., 2006, 46, 243–253.
19 R. Cao, M. Liu, M. Yin, Q. Liu, Y. Wang and N. Huang,
J. Chem. Inf. Model., 2012, 52, 2730–2740.
J. Mancip, K. Knoblauch, M. Afanassieff and P. Savatier, Stem
Cells, 2007, 25, 2996–3004.
33 Y. Zhang, Z. Wang, X. Li and N. S. Magnuson, Oncogene,
2008, 27, 4809–4819.
20 O. Fedorov, B. Marsden, V. Pogacic, P. Rellos, S. Muller,
A. N. Bullock, J. Schwaller, M. Sundstrom and S. Knapp, 34 M. J. Murphy, A. Wilson and A. Trumpp, Trends Cell Biol.,
Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 20523–20528. 2005, 15, 128–137.
21 D. M. Lorber and B. K. Shoichet, Curr. Top. Med. Chem., 2005, 35 R. Amson, F. Sigaux, S. Przedborski, G. Flandrin, D. Givol
5, 739–749.
and A. Telerman, Proc. Natl. Acad. Sci. U. S. A., 1989, 86,
8857–8861.
36 S. Kang, S. M. Dong, B. R. Kim, M. S. Park, B. Trink,
H. J. Byun and S. B. Rho, Apoptosis, 2012, 17, 989–997.
22 N. Huang, B. K. Shoichet and J. J. Irwin, J. Med. Chem., 2006,
49, 6789–6801.
23 J. J. Irwin, B. K. Shoichet, M. M. Mysinger, N. Huang,
F. Colizzi, P. Wassam and Y. Cao, J. Med. Chem., 2009, 52, 37 S. B. Rho, B. R. Kim and S. Kang, Gynecol. Oncol., 2011, 120,
5712–5720. 121–127.
24 N. Huang, C. Kalyanaraman, K. Bernacki and M. P. Jacobson, 38 M. A. Sussman, Expert Rev. Cardiovasc. Ther., 2009, 7, 929–
Phys. Chem. Chem. Phys., 2006, 8, 5166–5177. 938.
25 K. Zhu, M. R. Shirts and R. A. Friesner, J. Chem. Theory 39 A. T. Fathi, O. Arowojolu, I. Swinnen, T. Sato, T. Rajkhowa,
Comput., 2007, 3, 2108–2119.
D. Small, F. Marmsater, J. E. Robinson, S. D. Gross,
M. Martinson, S. Allen, N. C. Kallan and M. Levis, Leuk.
Res., 2012, 36, 224–231.
26 J. B. Christensen, O. Hendricks, S. Chaki, S. Mukherjee,
A. Das, T. K. Pal, S. G. Dastidar and J. E. Kristiansen, PLoS
One, 2013, 8, e57493.
27 W. L. DeLano, The PyMOL Molecular Graphics System, DeLano
Scientic, San Carlos, CA, USA, 2002, computer program.
28 T. W. Backman, Y. Cao and T. Girke, Nucleic Acids Res., 2011,
39, W486–W491.
40 W. W. Chen, D. C. Chan, C. Donald, M. B. Lilly and
A. S. Kra, Mol. Cancer Res., 2005, 3, 443–451.
41 A. S. Tsvetkov, J. Miller, M. Arrasate, J. S. Wong, M. A. Pleiss
and S. Finkbeiner, Proc. Natl. Acad. Sci. U. S. A., 2010, 107,
16982–16987.
This journal is © The Royal Society of Chemistry 2014
Med. Chem. Commun., 2014, 5, 507–511 | 511