Identification of a new series of potent diphenol HSP90 inhibitors by fragment merging and structure-based optimization

Article abstract of DOI:10.1016/j.bmcl.2014.03.100

Heat shock protein 90 (HSP90) is a molecular chaperone to fold and maintain the proper conformation of many signaling proteins, especially some oncogenic proteins and mutated unstable proteins. Inhibition of HSP90 was recognized as an effective approach to simultaneously suppress several aberrant signaling pathways, and therefore it was considered as a novel target for cancer therapy. Here, by integrating several techniques including the fragment-based drug discovery method, fragment merging, computer aided inhibitor optimization, and structure-based drug design, we were able to identify a series of HSP90 inhibitors. Among them, inhibitors 13, 32, 36 and 40 can inhibit HSP90 with IC₅₀ about 20-40 nM, which is at least 200-fold more potent than initial fragments in the protein binding assay. These new HSP90 inhibitors not only explore interactions with an under-studied subpocket, also offer new chemotypes for the development of novel HSP90 inhibitors as anticancer drugs.

Full text of DOI:10.1016/j.bmcl.2014.03.100

Bioorganic & Medicinal Chemistry Letters 24 (2014) 2525–2529
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Identification of a new series of potent diphenol HSP90 inhibitors
by fragment merging and structure-based optimization
Jing Ren a, , Jian Li b,d, , Yueqin Wang c, , Wuyan Chen d, Aijun Shen c, Hongchun Liu c, Danqi Chen a,
Danyan Cao a, Yanlian Li a, Naixia Zhang a, Yechun Xu d, Meiyu Geng c, , Jianhua He b, , Bing Xiong a,
,
⇑
⇑
⇑
Jingkang Shen a,
⇑
a Department of Medicinal Chemistry, National Key Laboratory of New Drug, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, 555 Zuchongzhi Road,
Shanghai 201203, China
b Shanghai Institute of Applied Physics, Chinese Academy of Sciences, 239 Zhangheng Road, Shanghai 201204, China
c Division of Antitumor Pharmacology, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, 555 Zuchongzhi Road, Shanghai 201203, China
d CAS Key Laboratory of Receptor Research, Drug Discovery and Design Center, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, 555 Zuchongzhi Road,
Shanghai 201203, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Heat shock protein 90 (HSP90) is a molecular chaperone to fold and maintain the proper conformation of
many signaling proteins, especially some oncogenic proteins and mutated unstable proteins. Inhibition of
HSP90 was recognized as an effective approach to simultaneously suppress several aberrant signaling
pathways, and therefore it was considered as a novel target for cancer therapy. Here, by integrating sev-
eral techniques including the fragment-based drug discovery method, fragment merging, computer aided
inhibitor optimization, and structure-based drug design, we were able to identify a series of HSP90 inhib-
itors. Among them, inhibitors 13, 32, 36 and 40 can inhibit HSP90 with IC50 about 20–40 nM, which is at
least 200-fold more potent than initial fragments in the protein binding assay. These new HSP90 inhib-
itors not only explore interactions with an under-studied subpocket, also offer new chemotypes for the
development of novel HSP90 inhibitors as anticancer drugs.
Received 8 January 2014
Revised 26 February 2014
Accepted 31 March 2014
Available online 8 April 2014
Keywords:
Heat shock protein
HSP90
Fragment-based drug discovery
Anticancer
Ó 2014 Elsevier Ltd. All rights reserved.
Heat shock protein 90 (HSP90) belongs to a large family of heat
shock proteins with molecular weight around 90 kDa, and func-
tions as a molecular chaperone to fold and maintain proper confor-
mations of many ‘client’ proteins.1–3 HSP90 is upregulated by heat
and other stressors to protect cells against these damaging effects,
oncogenic proteins, it has been considered as a promising target
for cancer therapy.7,8 The most attractive strength of HSP90 inhib-
itors is that they can not only simultaneously affect several aber-
rant signaling pathways essential for cancer development, may
also reduce the possibility of acquiring resistance in tumor cells
induced by some targeting drugs treatment.9
and it consists of two isoforms: HSP90a is an inducible form over-
expressed in many cancer cells, while HSP90b is the constitutive
form.4,5 As one of the most abundant member of heat shock pro-
teins, HSP90 has been extensively studied and reported to have
at least 280 client proteins.6 Among them, 48 are involved in cell
growth and various signaling cascades. Some of its clients are noto-
rious oncogenes, including several validated cancer drug targets
such as HER-2, Bcr-Abl, VEGFR, EGFR. Besides, the more sensitive
clients are usually those mutated unstable proteins involved in
aberrant signaling transduction in tumor cells. Since inhibition of
HSP90 can lead to the degradation of a large collection of
Currently, many HSP90 inhibitors have been discovered and
more than 10 drugs are now in clinical trials. Figure 1 lists several
diversified HSP90 inhibitors,10,11 and they are all bound to the
ATP site of the HSP90 N-terminal ATPase domain to impede the
chaperone function of HSP90. As the prototype of several syn-
thetic inhibitors, radicicol (2) contains a core fragment, dihy-
droxyphenyl group which forms extensive hydrogen bonding
interactions with the ATP binding site of N-terminus of HSP90.
Beside of these inhibitors inspired by natural products, frag-
ment-based drug discovery (FBDD) is also a popular strategy in
the HSP90 inhibitor development, which was adopted to discover
AT13387(4) in Astex pharmaceutical12 and SNX-5422(5) in Sere-
nex.13 Here we report an application of FBDD approach to identify
new series of dihydroxyphenyl-based HSP90 inhibitors, which ini-
tially designed by merging two bioactive fragments identified
⇑
Corresponding authors. Tel./fax: +86 21 50806600 5412 (B. X).
E-mail addresses: mygeng@simm.ac.cn (M. Geng), hejh@sinap.ac.cn (J. He),
bxiong@simm.ac.cn (B. Xiong).
These authors contributed equally to this work.
http://dx.doi.org/10.1016/j.bmcl.2014.03.100
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

2526
J. Ren et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2525–2529
Figure 1. HSP90 inhibitors.
from crystallography-based fragment screening, and followed
with the optimization guided by structure-based drug design
methods.
structure it rotates around the phenol ring and enables the
[12,4]triazolo[4,3-a]pyridine group situating at a small subpocket
that usually accommodates the isopropyl group in compounds
AUY922 and AT-13387. Besides, it was found the triazolopyridine
To apply the FBDD approach, we firstly analyzed crystal struc-
tures of complexes of HSP90 and its inhibitors deposited in PDB
database. Through this, it was found that all inhibitors have an
essential chemical moiety forming two hydrogen bonds with the
residue ASP93 of HSP90. Based on this critical interaction pattern,
we screened the ChemDiv commercial available database with this
pharmacophore, and scrutinized its novelty and synthesis accessi-
bility of fragment hits. After that, we purchased and synthesized
about twenty promising fragments as our small HSP90 focused
library. By utilizing the co-crystallization and screening with X-
ray crystallography method,14 we were able to identify several
fragments bound in the ATP site of HSP90 in solved crystal struc-
tures, and two of them were illustrated in Figure 2. To our surprise,
fragment 6 binds to HSP90 in an unexpected way. Before obtained
the co-crystal structure of 6 with HSP90, we speculated that it
would bind to the position similar to the isoxazole ring of
AUY922; while as illustrated in Figure 2A, in the complex crystal
ring forms a
p–p stacking interaction with residue PHE-138 and
one of the nitrogen atoms in triazole ring form two hydrogen
bonds with residue ASN-51 through a structural water molecule.
By overlapping two complex structures of fragments 6 and 7
(Fig. 2D), we can identify the phenol rings match very well. Conse-
quently, we designed compounds 8 and 9 by merging these two
fragments, and expected it to have improved binding affinity to
HSP90. As listed in Table 1, from the fluorescence polarization
(FP) assay, compound 8 and its analog 9 surprisingly showed lower
activities against HSP90 than the fragments.
To gain an understanding why the binding affinity is consider-
ably low, we crystallized and successfully solved the co-structure
of HSP90 with compound 9 (Fig. 2E). As shown, the compound 9
binds to the N-terminal of HSP90 as expected, with the triazolo-
pyridine ring situated at the same position of fragment 6. Based
on this crystal structure, we use the computational chemistry
N
N
Br
N
HO
HO
6
7
Cl
O
OH
N
N
A
D
B
E
C
F
Figure 2. Co-crystal structures of HSP90 with compounds 6, 7 and 9. The HSP90 was shown in cartoon style and the ligands 6, 7 and 9 were shown in stick. (A) The co-crystal
structure of HSP90 with 6 (PDB ID: 4L91). (B) The co-crystal structure of HS90 with 7 (PDB ID: 4L94). (C) The chemical structures of fragments 6 and 7. (D) Superimposition of
co-crystal structures of 6 and 7. (E) Co-crystal structure of HSP90 with 9 (PDB ID: 4L8Z). (F) The 2D diagram of interactions between HSP90 and compound 9.

J. Ren et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2525–2529
2527
Table 1
calculations, it was shown that the binding of triazolopyridine at
the ATP site of HSP90 is enthalpy favorable and the electrostatic
interaction between triazolopyridine group and HSP90 is less
important than the van der Waals interactions. As pointed by
numerous analyses that the enthalpy and entropy compensation
is a common phenomenon in protein–ligand interactions16 that
says favorable enthalpy interactions such as hydrogen bonding,
The fluorescence polarization assay of the identified active fragments and compounds
from initial optimization
R1
HO
R2
p–p stacking can countpoise by the reduction in entropy contribu-
tion due to restriction of the flexibility of the receptor. Therefore,
the actually improvement in binding free energy of compound 9
may be low than the enthalpy contribution since the triazolopyri-
dine group clearly hinders the mobility of its surrounding residues
of HSP90. Nevertheless, we thought the binding interaction of
triazolopyridine group is not the destructive factor accounting for
the low activity of compound 9.
OH
O
6-15
R2
a
Compd
R1
FP IC50 (lM)
N
N
N
Br
Br
8
8.8 2.1%@5
18.7 5.1%@5
0.63 0.08
l
Mb
Mb
N
N
N
N
N
N
N
Then we performed the solvation free energy calculations on
compound 9. By utilizing the density functional M06-2X at 6-
31+G(d) level combined with the SMD solvation model17 imple-
mented in Gaussian program, we found that compound 9 has a
low solvation free energy about ꢀ26.66 kcal/mol, which indicated
that when binding to the HSP90, the compound needs to overcome
a large desolvation effect to transfer from aqueous solution to the
protein binding site. The solvation free energy of the virtual com-
pound without the triazolopyridine group is ꢀ19.72 kcal/mol. Tak-
ing together, it means that adding the triazolopyridine group needs
to supply about 7 kcal/mol desolvation free energy. Therefore, in
the next round of optimization, we focused on the issue of the
hydrophilicity of the compounds, and tried to reduce the ligand
solvation free energy to improve the binding affinity to HSP90.
Based on the previous SAR of HSP90 inhibitors,18 the tetrahy-
droisoquinoline in compound 9 is not the optimal for binding.
Therefore, together with above mentioned rationale, several het-
ero-aromatic rings were used to substitute the [12,4]triazolo[4,3-
a]pyridine group, and produced compounds 10–15. Comparison
of compounds 10 to 8 or 11 to 9, the only difference occurs at
the triazolopyridine ring substituted with the indole group; the
binding activities from FP assay are dramatically improved. This
is consistent with the calculation that the solvation free energy
of compound 11 is ꢀ13.98 kcal/mol, much less than the value of
compound 9 (ꢀ26.66 kcal/mol). Similar trend can be deduced from
the inhibition activities of 13–15. Compounds 14 and 15 are about
3 and 7-fold less active than ligand 13 respectively, which rein-
forces the hydrophobic property of the fused bicyclic ring is critical
for the binding. From compounds 10–13 in Table 1, it was found
that compound 13 with isoindoline group at R2 is the most potent
inhibitor, consistent with the data reported in the previous struc-
ture–activity relationship about R2 group.12 Consequently, we
selected compound 13 to carry out the further optimization by
attaching various functional groups to the indole moiety.
The synthesis route of compounds in Table 2 is outlined in
Scheme 1 (see the Supporting material for details). Generally, the
synthesized compound 16 was hydrolyzed with lithium hydroxide
to give compound 17. Compound 18 was prepared by coupling the
key intermediate 17 with isoindoline using standard EDC/HOBt
coupling conditions. After coupling of 18 with different substituted
indoles,19 followed by de-protection of the benzyl protecting
groups, the final product 28–36 were obtained.
Inspecting the co-crystal structure, it was thought that the sub-
pocket accommodating the indole ring cannot permit a large sub-
stituent on the indole, especially at 6-, 7-position of indole ring.
Based on the in silico modeling, we prepared 9 compounds with
various substituents at three positions of indole. As listed in
Table 2, modification of the R1-position of indole with linear polar
groups (compounds 28–30) decreases the binding affinity consid-
erably, at least 30-fold less active than the parent compound 13,
especially the compound with acidic group (30). Due to the limited
N
N
9
l
N
10
11
12
13
N
0.13 0.01
N
N
N
S
0.33 0.02
0.038 0.006
N
N
N
14
15
0.13 0.01
0.25 0.03
N
N
N
N
a
All the inhibition ratios or IC50 values were obtained from triple measurements.
The compounds Geldanamycin and NVP-AUY922 were used as control, with IC50
values 74.0 7 nM and 8.0 0.2 nM, respectively.
b
The inhibition ratio measured with FP assay method at ligand concentration
5 lM.
methods to quantify two important components of the protein–
ligand binding free energy, namely the binding interaction energy
and solvation free energy, of ligand 9 with or without the triazolo-
pyridine group. The Glide docking program15 in Schrödinger soft-
ware package was used to calculate the protein–ligand binding
energy. Since we solved the crystal structure of ligand 9 bound
to the N-terminal of HSP90, the SCORE-ONLY mode was used to
obtain the binding interaction energy. From the calculation, it
was found that the non-covalent interaction energy between com-
pound 9 and the N-terminal of HSP90 is ꢀ68.78 kcal/mol, and the
large contribution stems from van der Waals interactions
(ꢀ56.83 kcal/mol). While for the virtual compound by deleting
the triazolopyridine group (see chemical structure 9v in Support-
ing materials) from ligand 9, its interaction energy is
ꢀ48.77 kcal/mol. The large difference of non-covalent interactions
between two compounds is rooted from the difference of their van
der Waals interactions (ligand 9 is ꢀ20.25 kcal/mol lower), while
the coulomb electrostatic interactions are almost the same (ligand
9: ꢀ11.95 kcal/mol; the virtual compound without the triazolo-
pyridine group: ꢀ12.18 kcal/mol). From these docking energy

2528
J. Ren et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2525–2529
Table 2
changed it to phenyl-amino group. With this effort, compounds
40–42 were prepared via palladium catalyzed Buchwald–Hartwig
amination20,21 from the intermediate 18, and followed by
deprotection.
The fluorescence polarization binding assay of compounds 28–36l
R2
R3
R1
4
5
3
6
As shown in Table 3, the compound (40) bearing aniline moiety
is the most potent inhibitor against HSP90 in current study, with
the IC50 about 21 nM in the FP assay. Compound 41 with methyl
group substituted at R1 position shows moderate inhibition
(IC50 = 71 nM), while the compound (42) with di-chlorine substi-
tuted phenylamino group decreases the potency about 6-fold
(IC50 = 190 nM). Although we only synthesized few compounds,
to the best of our knowledge, this modification is interesting as it
provided new ideas for HSP90 inhibitor development, which may
also enable other researchers to design substituted phenylamino
groups or novel heterocycles at this position.
We investigated the druglikeness of these series by selecting 8
potent compounds to check their metabolic stability properties—
the clearance ratio in human and mouse liver microsomes as well
as the inhibitions against five cytochrome P450 isoenzymes. In
Table 4, all tested molecules demonstrated low clearance ratio in
human or rat liver microsomes. Although all compounds showed
2
7
1N
HO
N
OH
O
R1
R2
H
R3
FP IC50 (lM)
a
Compd
28
OMe
1.5 0.33
NH2
O
29
H
H
OMe
56.9% 6.3%@1
l
Mb
Mb
N
H
O
30
H
14.1% 7.7%@1
l
OH
31
32
H
H
Br
H
H
0.07 0.01
0.039 0.0005
NO2
O
33
34
H
H
H
H
0.13 0.01
0.62 0.09
O
H
N
no more than 50% inhibition at concentration 10 lM against the
abundant CYP isoenzymes 3A4 and 2D6, they did process unfavor-
able inhibitions towards other three isoenzymes 2C9, 1A2, and
2C19. We also checked the time-dependent inhibition (TDI) of
these compounds, and found that only compound 32 has TDI inhi-
bition to 2C19. Taking together, we selected compounds 13, 31, 36
and 42 for further cellular activity test, as these compounds dem-
onstrate slightly better stability profiles than others.
O
H
35
36
H
H
Br
Me
0.078 0.005
0.038 0.003
H
a
All the inhibition ratios or IC50 values were obtained from triple measurements.
The inhibition ratio measured with FP assay method at ligand concentration
b
1
l
M.
From Table 5, it was found that the selected four compounds
volume of the subpocket accommodating the indole ring, as indi-
cated in Table 2, the inhibitory activity is very sensitive to the sub-
stitution at R2- or R3-position of indole. The bromine atom at these
positions of the indole ring is tolerable, both with IC50 value about
70 nM (31 and 35). The 4-nitro indole can maintain the equipoten-
cy against HSP90 in FP assay, but the activities of compounds 33
and 34 with a slightly large group at R2-position of indole is lower
than compound 13, decrease about 4- and 20-fold, respectively. 36
with methyl group at R3-position of indole is also tolerable with
similar IC50 value (38 nM).
have similar inhibitory activities, with GI50 around 10 lM against
four cancer cell lines. By comparing to the potent cellular prolifer-
ation activities of the HSP90 inhibitors in clinical trials,11 we spec-
ulate these compounds may have poor cell permeability issues as
there is a large gap between cellular activity and protein binding
activity. Further membrane permeability assay will verify this
hypothesis, and may shed the light on the direction of
optimization.
In summary, by employing the fragment-based drug discovery
approach, we were able to identify two binding fragments.
Scrutinizing the co-crystal structures of these two fragments with
N-terminal ATPase domain of HSP90, it was found that two frag-
ments situated at different subpocket in ATP binding site, and their
common pharmacophore-phenol ring fitted very well. Apparently
merging two fragments to form compound 9 were proved not to
From the co-crystal structure of HSP90 bound with compound
9, the
p–p stacking involving the residue PHE138 is important
for van der Waals interactions. Due to the ligand stacking part is
largely at the six-membered ring of indole, therefore, the ring-
opening approach was taken to further modify the indole ring by
R1
R2
Br
Br
Br
BnO
BnO
BnO
3a
3b
3g
NH
BnO
O
OH
N
N
OBn O
16
OBn O
17
OBn O
18
OBn O
37-39
3c
3d or 3e
R2
R2
R3
R3
R1
R1
R1
R2
NH
N
3d,3e or 3f
N
HO
BnO
HO
N
N
N
OH
O
OBn O
19-27
OH
O
28-36
40-42
Scheme 1. Synthesis of compounds 28–36 and 40–42. Reagents and conditions: (3a) LiOH, MeOH,H2O, reflux, 99%; (3b) isoindoline, EDCI, HOBt, DIPEA, DCM, rt, 70%; (3c)
indoles, K3PO4, CuI, (+/ꢀ)-trans-1,2-Diaminocyclohexane or (1R,2R)-N,N0-Dimethyl-1,2-cyclohexane-diamine, toluene, reflux, 40–85%; (3d) H2, Pd/C, MeOH, rt, 35–85%; (3e)
BCl3, DCM, 0 °C, 19–42%; (3f) BBr3, DCM, ꢀ78 °C, 13%; (3g) aniline, Pd2(dba)3, t-Bu3P or X-phos, NaOBu-t, dioxane, 120 °C, 36–75%.

J. Ren et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2525–2529
2529
Table 3
200-fold more active than the initial fragments. These new series
of HSP90 inhibitors are interesting as the ligands extend to a less
explored subpocket, which may provide several potential chemo-
types for HSP90 inhibitor development.
The fluorescence polarization binding assay of compounds 40–42
R1
R2
NH
HO
Acknowledgments
N
The authors wish to thank Dr. Haiyan Liu for her help in per-
forming the in vitro PK study. We are grateful for financial support
from ‘Interdisciplinary Cooperation Team’ Program for Science and
Technology Innovation of the Chinese Academy of Sciences; the
‘100 Talents Project’ of CAS to Y.X. The National Natural Science
Foundation of China (Grant Nos. 81072580, 81273368 and
21272246); the National Science & Technology Major Project ‘Key
New Drug Creation and Manufacturing Program’ of China (Grant
No. 2014ZX09507-002).
OH
O
a
Compd
R1
R2
FP IC50 (lM)
40
41
42
H
Me
Cl
H
H
Cl
0.021 0.005
0.071 0.003
0.19 0.04
a
All the inhibition ratios or IC50 values were obtained from triple measurements.
Table 4
The in vitro metabolic stability study of the selected compounds
Supplementary data
Compd HLMa
RLMb
Direct inhibition
Time-
(10
l
M) Stability
Stability
mean (inhibition ratio) dependent
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2014.03.
100.
Clint(ul/min/ Clint(ul/min/
mg protein) mg protein)
inhibition
3A4 2D6 2C9 1A2 2C19
13
31
32
35
36
40
41
42
13
0
25
49
6
37%/ 14% 73% 52% 60% No inhibition
17%
33%/ 19% 54% 34% 70% No inhibition
22%
35%/ 50% 74% 47% 89% 2C19
8%
15%/ 32% 78% 48% 71% No inhibition
13%
26%/ 7% 68% 28% 63% No inhibition
4%
19%/ 22% 68% 72% 52% No inhibition
29%
40%/ 7% 75% 61% 58% No inhibition
38%
39%/ 5% 73% 40% 23% No inhibition
33%
References and notes
1. Feder, M. E.; Hofmann, G. E. Annu. Rev. Physiol. 1999, 61, 243.
2. Picard, D. Cell. Mol. Life Sci. 2002, 59, 1640.
3. Garrido, C.; Gurbuxani, S.; Ravagnan, L.; Kroemer, G. Biochem. Biophys. Res.
Commun. 2001, 286, 433.
22
18
0
16
42
12
0
4. Pearl, L. H.; Prodromou, C. Annu. Rev. Biochem. 2006, 75, 271.
5. Moulick, K.; Ahn, J. H.; Zong, H.; Rodina, A.; Cerchietti, L.; Gomes DaGama, E.
M.; Caldas-Lopes, E.; Beebe, K.; Perna, F.; Hatzi, K.; Vu, L. P.; Zhao, X.; Zatorska,
D.; Taldone, T.; Smith-Jones, P.; Alpaugh, M.; Gross, S. S.; Pillarsetty, N.; Ku, T.;
Lewis, J. S.; Larson, S. M.; Levine, R.; Erdjument-Bromage, H.; Guzman, M. L.;
Nimer, S. D.; Melnick, A.; Neckers, L.; Chiosis, G. Nat. Chem. Biol. 2011, 7, 818.
6. da Silva, V. C.; Ramos, C. H. J. Proteomics 2012, 75, 2790.
7. Travers, J.; Sharp, S.; Workman, P. Drug Discov. Today 2012, 17, 242.
8. Hwang, M.; Moretti, L.; Lu, B. Curr. Med. Chem. 2009, 16, 3081.
9. Xu, W.; Neckers, L. Clin. Cancer Res. 2007, 13, 1625.
2
0
0
16
a
10. Hong, D. S.; Banerji, U.; Tavana, B.; George, G. C.; Aaron, J.; Kurzrock, R. Cancer
Treat. Rev. 2013, 39, 375.
11. Biamonte, M. A.; Van de Water, R.; Arndt, J. W.; Scannevin, R. H.; Perret, D.; Lee,
W. C. J. Med. Chem. 2010, 53, 3.
HLM is human liver microsome.
RLM is rat liver microsome.
b
12. Woodhead, A. J.; Angove, H.; Carr, M. G.; Chessari, G.; Congreve, M.; Coyle, J. E.;
Cosme, J.; Graham, B.; Day, P. J.; Downham, R.; Fazal, L.; Feltell, R.; Figueroa, E.;
Frederickson, M.; Lewis, J.; McMenamin, R.; Murray, C. W.; O’Brien, M. A.;
Parra, L.; Patel, S.; Phillips, T.; Rees, D. C.; Rich, S.; Smith, D. M.; Trewartha, G.;
Vinkovic, M.; Williams, B.; Woolford, A. J. J. Med. Chem. 2010, 53, 5956.
13. Huang, K. H.; Veal, J. M.; Fadden, R. P.; Rice, J. W.; Eaves, J.; Strachan, J. P.;
Barabasz, A. F.; Foley, B. E.; Barta, T. E.; Ma, W.; Silinski, M. A.; Hu, M.; Partridge,
J. M.; Scott, A.; DuBois, L. G.; Freed, T.; Steed, P. M.; Ommen, A. J.; Smith, E. D.;
Hughes, P. F.; Woodward, A. R.; Hanson, G. J.; McCall, W. S.; Markworth, C. J.;
Hinkley, L.; Jenks, M.; Geng, L.; Lewis, M.; Otto, J.; Pronk, B.; Verleysen, K.; Hall,
S. E. J. Med. Chem. 2009, 52, 4288.
Table 5
The cellular antiproliferative activity of compounds 13, 31, 36 and 42a
Compd
GI50
NCI-H3122
(l
M)
A549
HCT116
BT-474
13
31
36
42
10.08 0.17
12.26 2.15
8.6 1.69
7.53 0.82
9.67 0.01
7.53 1.19
10.12 0.03
14.31 1.57
21.01 2.02
11.1 1.69
24.17 1.69
7.25 0.37
8.52 0.52
7.07 3.48
6.31 1.22
10.63 0.3
14. Zhao, L.; Cao, D.; Chen, T.; Wang, Y.; Miao, Z.; Xu, Y.; Chen, W.; Wang, X.; Li, Y.;
Du, Z.; Xiong, B.; Li, J.; Xu, C.; Zhang, N.; He, J.; Shen, J. J. Med. Chem. 2013, 56,
3833.
a
Values represent mean standard deviation are calculated from experiments
performed in triplicate.
15. Halgren, T. A.; Murphy, R. B.; Friesner, R. A.; Beard, H. S.; Frye, L. L.; Pollard, W.
T.; Banks, J. L. J. Med. Chem. 2004, 47, 1750.
be successful as the compound 9 is surprisingly less active than its
parent fragment. To gain the knowledge about the reason of this
phenomenon, computational modeling was adopted to calculate
two important components in binding free energy: the binding
interaction energy and ligand solvation free energy. Through the
calculations, it was found that the destructive factor may be rooted
from ligand solvation free energy. Therefore, in the next rounds of
optimization, we focused on reducing the hydrophilic tendency of
ligands, and successfully obtained several compounds were at least
16. Chodera, J. D.; Mobley, D. L. Annu. Rev. Biophys. 2013, 42, 121.
17. Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B 2009, 113, 4538.
18. Murray, C. W.; Carr, M. G.; Callaghan, O.; Chessari, G.; Congreve, M.; Cowan, S.;
Coyle, J. E.; Downham, R.; Figueroa, E.; Frederickson, M.; Graham, B.;
McMenamin, R.; O’Brien, M. A.; Patel, S.; Phillips, T. R.; Williams, G.;
Woodhead, A. J.; Woolford, A. J. J. Med. Chem. 2010, 53, 5942.
19. Antilla, J. C.; Klapars, A.; Buchwald, S. L. J. Am. Chem. Soc. 2002, 124, 11684.
20. Hartwig, J. F.; Kawatsura, M.; Hauck, S. I.; Shaughnessy, K. H.; Alcazar-Roman,
L. M. J. Org. Chem. 1999, 64, 5575.
21. Hanthorn, J. J.; Valgimigli, L.; Pratt, D. A. J. Org. Chem. 2012, 77, 6908.