Development of 3-phenyl-N-(2-(3-phenylureido)ethyl)-thiophene-2-sulfonamide compounds as inhibitors of antiapoptotic Bcl-2 family proteins

Article abstract of DOI:10.1002/cmdc.201400058

Antiapoptotic Bcl-2 family proteins, such as Bcl-x_L, Bcl-2, and Mcl-1, are often overexpressed in tumor cells, which contributes to tumor cell resistance to chemotherapies and radiotherapies. Inhibitors of these proteins thus have potential applications in cancer treatment. We discovered, through structure-based virtual screening, a lead compound with micromolar binding affinity to Mcl-1 (inhibition constant (K_i)=3 μM). It contains a phenyltetrazole and a hydrazinecarbothioamide moiety, and it represents a structural scaffold not observed among known Bcl-2 inhibitors. This work presents the structural optimization of this lead compound. By following the scaffold-hopping strategy, we have designed and synthesized a total of 82 compounds in three sets. All of the compounds were evaluated in a fluorescence-polarization binding assay to measure their binding affinities to Bcl-x_L, Bcl-2, and Mcl-1. Some of the compounds with a 3-phenylthiophene-2-sulfonamide core moiety showed sub-micromolar binding affinities to Mcl-1 (K_i=0.3-0.4 μM) or Bcl-2 (K_i≈1 μM). They also showed obvious cytotoxicity on tumor cells (IC ₅₀<10 μM). Two-dimensional heteronuclear single quantum coherence NMR spectra of three selected compounds, that is, YCW-E5, YCW-E10, and YCW-E11, indicated that they bind to the BH3-binding groove on Bcl-x _L in a similar mode to ABT-737. Several apoptotic assays conducted on HL-60 cells demonstrated that these compounds are able to induce cell apoptosis through the mitochondrial pathway. We propose that the compounds with the 3-phenylthiophene-2-sulfonamide core moiety are worth further optimization as effective apoptosis inducers with an interesting selectivity towards Mcl-1 and Bcl-2. Hopping to inhibition: Several sets of derivatives of a lead compound have been designed and synthesized by following the scaffold-hopping strategy, and active compounds containing a 3-phenylthiophene-2-sulfonamide core moiety have been obtained. The most potent compounds have sub-micromolar binding affinity to the antiapoptotic protein Mcl-1 and are effective apoptosis inducers in living cells.

Full text of DOI:10.1002/cmdc.201400058

CHEMMEDCHEM
FULL PAPERS
DOI: 10.1002/cmdc.201400058
Development of 3-Phenyl-N-(2-(3-phenylureido)ethyl)-
thiophene-2-sulfonamide Compounds as Inhibitors of
Antiapoptotic Bcl-2 Family Proteins
Chengwen Yang,[a] Sha Chen,[b] Mi Zhou,[a] Yan Li,[a] Yangfeng Li,[a] Zhengxi Zhang,[a]
Zhen Liu,[a] Qian Ba,[c] Jingquan Li,[c] Hui Wang,[c] Xiaomei Yan,*[b] Dawei Ma,*[a] and
Renxiao Wang*[a]
Antiapoptotic Bcl-2 family proteins, such as Bcl-xL, Bcl-2, and
Mcl-1, are often overexpressed in tumor cells, which contrib-
utes to tumor cell resistance to chemotherapies and radio-
therapies. Inhibitors of these proteins thus have potential ap-
plications in cancer treatment. We discovered, through struc-
ture-based virtual screening, a lead compound with micromo-
lar binding affinity to Mcl-1 (inhibition constant (Ki)=3 mm). It
contains a phenyltetrazole and a hydrazinecarbothioamide
moiety, and it represents a structural scaffold not observed
among known Bcl-2 inhibitors. This work presents the structur-
al optimization of this lead compound. By following the scaf-
fold-hopping strategy, we have designed and synthesized
a total of 82 compounds in three sets. All of the compounds
were evaluated in a fluorescence-polarization binding assay to
measure their binding affinities to Bcl-xL, Bcl-2, and Mcl-1.
Some of the compounds with a 3-phenylthiophene-2-sulfon-
amide core moiety showed sub-micromolar binding affinities
to Mcl-1 (Ki =0.3–0.4 mm) or Bcl-2 (Ki ꢀ1 mm). They also showed
obvious cytotoxicity on tumor cells (IC50 <10 mm). Two-dimen-
sional heteronuclear single quantum coherence NMR spectra
of three selected compounds, that is, YCW-E5, YCW-E10, and
YCW-E11, indicated that they bind to the BH3-binding groove
on Bcl-xL in a similar mode to ABT-737. Several apoptotic
assays conducted on HL-60 cells demonstrated that these com-
pounds are able to induce cell apoptosis through the mito-
chondrial pathway. We propose that the compounds with the
3-phenylthiophene-2-sulfonamide core moiety are worth fur-
ther optimization as effective apoptosis inducers with an inter-
esting selectivity towards Mcl-1 and Bcl-2.
Introduction
Apoptosis, or programmed cell death, is an essential physio-
logical process required for the development and maintenance
of tissue homeostasis. Deregulation of this process is known to
be associated with various types of cancer.[1–5] The B-cell lym-
phoma 2 (Bcl-2) family of proteins is a major group of apopto-
sis regulators, which includes both antiapoptotic members
(such as Bcl-2, Bcl-xL, Mcl-1, and Bcl-w), proapoptotic members
(such as Bak and Bax), and BH3-only proteins.[6] Overexpression
of antiapoptotic Bcl-2 family proteins contributes to resistance
to chemotherapies and radiotherapies in many tumor cells.
Hence, the Bcl-2 family of proteins has become an attractive
molecular target for the development of new cancer therapies
since the beginning of this century.[7,8]
The antiapoptotic members in the Bcl-2 family conduct their
biological functions through protein–protein interactions with
the proapoptotic members.[6] Designing small-molecule inhibi-
tors of protein–protein interactions has been considered as
reaching for the high-hanging fruits in drug discovery.[9–11] De-
spite the challenge, a number of small-molecule inhibitors of
Bcl-2 family proteins have been publicly reported in the past
ten years or so,[12–34] which makes the Bcl-2 family proteins per-
haps the most popular targets in this category. Some of the
best-known Bcl-2 family protein inhibitors are summarized in
Figure 1. A few of them have successfully entered phase II clini-
cal trials, such as Navitoclax (ABT-263) developed by Abbott
Laboratories.[12–14] But no small-molecule compound has en-
tered phase III trials so far. Despite all of the progress, Bcl-2 in-
hibitors are still very limited as compared with other well-es-
tablished drugs. In particular, novel chemical scaffolds are de-
[a] Dr. C. Yang, M. Zhou, Dr. Y. Li, Y. Li, Z. Zhang, Dr. Z. Liu, Prof. D. Ma,
Prof. R. Wang
State Key Laboratory of Bioorganic & Natural Products Chemistry
Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences
345 Lingling Rd, Shanghai 200032 (P. R. China)
E-mail: madw@main.sioc.ac.cn
wangrx@mail.sioc.ac.cn
[b] S. Chen, Prof. X. Yan
MOE Key Laboratory of Spectrochemical Analysis & Instrumentation
Key Laboratory for Chemical Biology of Fujian Province
Department of Chemical Biology, College of Chemistry & Chemical Engi-
neering, Xiamen University
422 Siming South Rd, Xiamen, Fujian 361005 (P. R. China)
E-mail: xmyan@xmu.edu.cn
[c] Dr. Q. Ba, Dr. J. Li, Prof. H. Wang
Institute of Nutritional Sciences
Shanghai Institutes of Biological Sciences, Chinese Academy of Sciences
319 Yueyang Rd, Shanghai 200031 (P. R. China)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201400058.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1436

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Figure 1. Some known small-molecule inhibitors of antiapoptotic Bcl-2 family proteins and their binding data reported in the literature.[12–34]
sired to increase the chance of discovering successful drug
candidates.
obtain a potent, orally available, selective inhibitor of Bcl-2,
that is, ABT-199.[30] This compound inhibits the growth of Bcl-
2-dependent tumors in vivo and spares human platelets. But,
so far, many Bcl-2 family inhibitors are still more potent to-
wards Bcl-2 or Bcl-xL than Mcl-1. For example, ABT-737 and
ABT-263 were found to be highly potent inhibitors of Bcl-2,
Bcl-xL, and Bcl-w (inhibition constant (Ki)<1 nm) but only mod-
estly affected Mcl-1 (Ki =0.55 mm).[12–14] They thus lack efficacy
on some types of cancer cells with Mcl-1 overexpression.
Indeed, down-regulation of Mcl-1 makes cancer cells sensitive
to ABT-737 or other cytotoxic agents.[37–39] Thus, inhibitors of
In addition, the selectivity of Bcl-2 inhibitors has drawn
public attention recently.[35] Although all antiapoptotic Bcl-2
family proteins have similar 3D structures, the sequence identi-
ties among them are only modest, typically around 30–70%,
which implies that they may target different types of tissues in
apoptosis and may also differ in responses to different stress
stimuli.[36] In fact, more selective inhibitors of certain Bcl-2
family proteins have been revealed recently. For example,
Souers et al. reported the re-engineering of Navitoclax to
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1437

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Mcl-1 and other less-noted antiapoptotic Bcl-2 family proteins
are expected to supplement current Bcl-2/Bcl-xL inhibitors to
provide more effective or more selective cancer therapies.
In order to discover inhibitors of Bcl-2 family proteins with
new chemical scaffolds, we conducted virtual screening of
small-molecule compounds by targeting Mcl-1 in our previous
study. Mcl-1 was chosen as the target because it has been dis-
covered in recent years that the biological functions of Mcl-
1 are notably different from those of other antiapoptotic Bcl-2
family proteins, such as Bcl-2 or Bcl-xL. In fact, the sequence
obtained some compounds with a 3-phenylthiophene-2-sulfo-
namide core moiety that exhibited both promising binding af-
finities to the target proteins and cytotoxicity on tumor cells.
The best compounds have sub-micromolar binding affinities to
Mcl-1 and Bcl-2 and are about 10-fold more potent than the
lead compound. We also demonstrated in several apoptosis
assays that these compounds effectively induced cell apoptosis
through the mitochondrial pathway.
similarity between Mcl-1 and Bcl-2 is only 30%, which indicates Results and Discussion
its independent characteristics.[35] Over 56000 compounds
1. Hopping from the lead compound to the thiophene-2-sul-
from the Maybridge chemical catalogue (released in August
2008) were docked into the BH3-binding groove on Mcl-1 by
using GOLD software (version 5.1). After visual examination of
the top-ranked candidates suggested by molecular docking,
we manually selected 200 compounds. Samples of these com-
pounds were ordered and then tested in a fluorescence-polari-
zation (FP)-based binding assay. Among them, one compound,
BCL-VS-156 (Scheme 1), showed clear dose-dependent binding
to Mcl-1 with a Ki value of 3.0 mm. Although this level of bind-
ing affinity is only modest, this compound consists of a phenyl-
tetrazole and a hydrazinecarbothioamide moiety, which are
not observed among known Bcl-2 inhibitors. By taking this
compound as the lead compound, we synthesized a total of
82 derivative compounds in three sets (Scheme 1). Finally, we
fonamide compounds
In order to confirm the binding affinity of the lead compound
(BCL-VS-156) to the Bcl-2 family proteins, the first set of com-
pounds synthesized by us were direct derivatives of the lead
compound, which all contain a phenyltetrazole moiety as well
as a hydrazinecarbothioamide moiety (Table 1). The synthetic
methods for preparing this set of compounds are outlined in
Scheme 2. A pair of regioisomers, 3a and 3b, were actually ob-
tained after compound 2 reacted with methyl bromoacetate.
These two regioisomers were carefully separated and applied
in the subsequent reactions, and two sets of compounds were
thus obtained, that is, YCW-A1–A13 and YCW-B1–B12. For both
sets of compounds, the phenyl rings connected to the thio-
urea moiety were modified with
small-size
electron-donating
(such as methyl and methoxy) or
electron-withdrawing (such as
trifluoromethyl and nitro) sub-
stituent groups. In addition, the
thiourea moiety was replaced by
a urea moiety. The overall yields
of the A and B compounds were
generally higher than 80% (with
only two exceptions). However,
some steps in the synthetic
route were not completely prac-
tical. For example, azides were
needed to introduce the tetra-
zole ring and some reactions re-
quired a long period of heating
at reflux. The purity of most of
the A compounds ranged from
90–99%. But the purifications of
some of the B compounds were
relatively troublesome, and the
purity of six B compounds was
only around 75%.
Compounds YCW-A1–A13 and
YCW-B1–B12 were all tested in
an FP-based binding assay,
which was applied successfully
in our previous studies.[40–42] The
results of our binding assay indi-
cated that these compounds
Scheme 1. An overview of how the thiophene-2-sulfonamide compounds were derived from the lead compound
in our study.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1438

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
pounds were selected based on the MTT assay re-
sults, and they were tested in more sophisticated
assays to explore their apoptosis-inducing mecha-
nism. Our results showed that none of the A or B
compounds exhibited obvious cytotoxicity on any
cell line under test. Thus, it seems that these two
sets of compounds are not very cell permeable,
and they are not likely to target Bcl-2 family pro-
teins effectively in living cells.
Table 1. Chemical structures and binding data for the A and B compounds.
Compd
X
R1
R2
Yield
[%]
Ki [mm][a]
Bcl-2
Due to the difficulties in organic synthesis as
well as the poor cell permeability of the A and B
compounds, as our next attempt, we synthesized
another set of compounds by replacing the phe-
nyltetrazole moiety with a phenyltriazole moiety
(Table 2). These compounds (YCW-C1–C18) were
synthesized as outlined in Scheme 3. In this case,
the popular “click-chemistry”[43] method was adopt-
ed to introduce the triazole ring. The other change
in the structural scaffold was that the amide group
is attached to the triazole ring directly, and it is
separated from the thiourea moiety by one or two
methylene units. With these changes, this set of
compounds was generally more convenient to syn-
thesize than the A and B compounds. The overall
yields of most of the compounds in this set (only
one exception) were over 80%, and their purities
were all above 90%.
Bcl-xL
Mcl-1
YCW-A1
YCW-A2
YCW-A3
YCW-A4
YCW-A5
YCW-A6
YCW-A7
YCW-A8
YCW-A9
YCW-A10
YCW-A11
YCW-A12
YCW-A13
YCW-B1
YCW-B2
YCW-B3
YCW-B4
YCW-B5
YCW-B6
YCW-B7
YCW-B8
YCW-B9
YCW-B10
YCW-B11
YCW-B12
S
S
S
S
S
S
S
S
O
O
O
O
O
S
S
S
S
S
3-Me
3-Cl
3-MeO
3-NO2
4-Cl
4-MeO
4-NO2
3-CF3
3-Me
4-Cl
4-MeO
4-NO2
3-Cl
3-Me
3-Cl
3-MeO
3-NO2
4-Cl
4-MeO
4-NO2
4-Cl
4-MeO
4-NO2
3-Me
3-Cl
H
H
H
H
H
H
H
5-CF3
5-Me
H
H
H
5-Cl
H
H
H
H
H
H
H
H
H
H
5-Me
5-Cl
86
88
85
89
87
86
89
61
85
59
85
86
89
88
89
87
91
89
87
91
89
84
91
88
90
weak
9.3ꢁ1.9
weak
3.1ꢁ0.6
weak
3.6ꢁ0.8
5.2ꢁ0.2
5.8ꢁ2.0
4.1ꢁ0.2
2.8ꢁ0.2
10.2ꢁ7.0
1.9ꢁ0.4
0.80ꢁ0.06
N.A.
N.A.
N.A.
N.A.
N.A.
5.6ꢁ1.1
4.3ꢁ1.5
20ꢁ6.0
2.5ꢁ0.20
3.0ꢁ0.4
7.2ꢁ2.7
2.4ꢁ0.1
N.A.
N.A.
N.A.
N.A.
N.A.
5.0ꢁ1.3
6.7ꢁ2.4
10.6ꢁ0.4
2.7ꢁ0.1
4.3ꢁ0.4
9.7ꢁ3.7
2.2ꢁ0.2
6.5ꢁ4.1
N.A.
N.A.
N.A.
N.A.
N.A.
6.1ꢁ0.92
5.0ꢁ1.4
5.2ꢁ2.2
2.8ꢁ1.5
4.5ꢁ0.8
4.6ꢁ1.3
2.0ꢁ0.1
N.A.
N.A.
N.A.
N.A.
N.A.
weak
3.5ꢁ0.8
13ꢁ1.2
N.A.
N.A.
N.A.
N.A.
N.A.
9.5ꢁ5.1
9.4ꢁ7.1
5.2ꢁ2.4
7.5ꢁ1.1
weak
10ꢁ7.5
3.0ꢁ1.3
N.A.
N.A.
N.A.
N.A.
N.A.
S
S
Compounds YCW-C1–C18 were also tested in
our FP-based assay to measure their binding to the
three antiapoptotic Bcl-2 family proteins. Unex-
pectedly, all of them exhibited none or very weak
binding to the three target proteins. Our molecular
modeling results suggested that the low-energy
conformation of the C compounds is more rigid
than those of the A or B compounds. As indicated
by the predicted binding mode of the lead com-
pound (Figure 2a), a certain level of conformation
flexibility in the backbone is necessary for main-
O
O
O
O
O
[a] The mean values and standard deviations of the Ki values were derived based on the
outcomes of three parallel measurements. N.A.: No obvious sign of binding was ob-
served; weak: dose-dependent FP signals were observed, but the change in the FP sig-
nals (up to 100 mm concentration) was not significant enough to determine the Ki value
accurately.
generally have Ki values ranging between 1–20 mm for Bcl-2
and between 2–11 mm for Mcl-1. As for Bcl-xL, some of these
compounds have micromolar binding affinities, whereas the
binding affinities of others were too weak to determine accu-
rately. These results confirmed that the A and B compounds,
which share the same structural scaffold as the lead com-
pound, are able to bind to the Bcl-2 family proteins. But no
compound was found to have really improved binding affini-
ties relative to that of the lead compound. Another notable
observation was that, when the thiourea moiety was replaced
by a urea moiety, the resulting compounds lost their binding
affinities completely.
taining binding to Mcl-1. The rigidity is perhaps the reason
why the C compounds cannot bind to the desired targets. This
set of compounds were also tested on tumor cell lines in MTT
assays. Most of them did not exhibit obvious cytotoxicity on
tumor cell lines. Thus, this set of compounds were not further
pursued in our study.
We then continued applying the scaffold-hopping strategy
to develop new derivative compounds. This time, we chose
the thiophene-2-sulfonamide moiety to replace the five-mem-
bered tetrazole or triazole moiety in the A/B/C compounds.
Both the thiophene and sulfonamide groups are common in
many drug molecules. They were chosen also because of their
synthetic feasibility. As a result, a new set of compounds, that
is, YCW-D1–D18 (Table 3), were synthesized as outlined in
Scheme 4. In this case, the thiophene-2-sulfonamide moiety (in
16) was prepared by sulfonylation of the available starting ma-
terial 14. Compound 16 was then connected to a phenylboron-
ic acid derivative, 13, 21a, or 21b, through the Suzuki cross-
coupling reaction. The product after deprotection, 18, 23a, or
All of the A and B compounds were also screened by 3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) vi-
ability assays on six tumor cell lines. It needs to be emphasized
that the outcomes of an MTT assay should not be related di-
rectly to a certain molecular target. In our study, this assay
served as a basic measure for characterizing the cell permeabil-
ity and cytotoxicity of our compounds. Some of the com-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1439

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Scheme 2. Synthesis of compounds YCW-A1–A13 and YCW-B1–B12. Reagents and conditions: a) NaN3, Et3N·HCl, DMF, reflux, 2 h, 45%; b) methyl bromoacetate,
NaH, THF, reflux, 5 h, 56% for 3a, 34% for 3b; c) N2H4·H2O, THF, reflux, 12 h, 4a: 92%, 4b: 91%; d) aryl isothiocyanate or aryl isocyanate, THF, reflux.
23b, was connected to an aryl isothiocyanate to obtain the
final compounds. With this synthetic route, the whole chemical
Table 2. Chemical structures for the C compounds.
structure of this set of compounds was dissected into three
building blocks. These building blocks were prepared in a paral-
lel manner, and combination of them yielded the final com-
pounds with the desired structural diversity. The overall yields
Table 3. Chemical structures for the D compounds.
Compd
R1
R2
Yield [%]
Compd
R1
R2
R3
Yield [%]
YCW-C1
YCW-C2
YCW-C3
YCW-C4
YCW-C5
YCW-C6
YCW-C7
YCW-C8
YCW-C9
YCW-C10
YCW-C11
YCW-C12
YCW-C13
YCW-C14
YCW-C15
YCW-C16
YCW-C17
YCW-C18
H
3-Me
3-Cl
3-MeO
3-NO2
4-Cl
4-MeO
4-NO2
3-CF3
H
3-Me
3-Cl
3-MeO
3-NO2
4-Cl
4-MeO
4-NO2
3-CF3
H
H
H
H
H
H
H
H
5-CF3
H
H
H
H
90
93
88
93
89
93
86
92
92
92
88
86
84
91
76
83
87
87
YCW-D1
YCW-D2
YCW-D3
YCW-D4
YCW-D5
YCW-D6
YCW-D7
YCW-D8
YCW-D9
YCW-D10
YCW-D11
YCW-D12
YCW-D13
YCW-D14
YCW-D15
YCW-D16
YCW-D17
YCW-D18
3-Me
3-Cl
3-MeO
3-NO2
4-Cl
H
H
H
H
H
H
H
5-CF3
H
H
H
H
5-CF3
H
H
H
H
5-CF3
2,4-dichlorophenyl
2,4-dichlorophenyl
2,4-dichlorophenyl
2,4-dichlorophenyl
2,4-dichlorophenyl
2,4-dichlorophenyl
2,4-dichlorophenyl
2,4-dichlorophenyl
1-naphthyl
1-naphthyl
1-naphthyl
1-naphthyl
1-naphthyl
99
98
99
95
99
96
99
97
90
95
91
93
90
99
97
90
91
89
4-MeO
4-NO2
3-CF3
3-MeO
3-NO2
4-MeO
4-NO2
3-CF3
3-MeO
3-NO2
4-MeO
4-NO2
3-CF3
H
H
H
H
2-naphthyl
2-naphthyl
2-naphthyl
2-naphthyl
5-CF3
2-naphthyl
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1440

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
The binding mode of the lead compound to Mcl-1 predicted
by molecular modeling suggested that the 2,4-dichlorophenyl
group on the structure went into a cavity formed mainly by
residues Met231, Met250, Val253, and Phe270 (Figure 2a). In
order to explore the role of this moiety in binding, another
similar set of thiophene-2-sulfonamide compounds, that is,
YCW-E1–E21 (Table 4), were synthesized, in which a 3,4-dihy-
droxyphenyl or 3,4-dimethoxyphenyl moiety was introduced to
replace the 2,4-dichlorophenyl moiety. In this case, the R1 and
R2 groups are still the small-size electron-donating and elec-
tron-withdrawing substituent groups as in the other sets of
compounds described above. The synthetic route for this set
of compounds is outlined in Scheme 5 and is similar to that for
the D compounds. The overall yields of the compounds with
the 3,4-dihydroxyphenyl moiety turned out to be case depen-
dent and ranged from 28–90%, whereas the overall yields of
the compounds with the 3,4-dihydroxyphenyl moiety were
generally over 90%. The purities of most of the compounds in
this set are above 90% (with four exceptions).
Encouragingly, our binding assay results indicated that some
E compounds have promising binding affinities to Mcl-1 and
Bcl-2. The most potent compounds have inhibition constants
of 0.3–0.4 mm for Mcl-1 (Table 4), values that are about 10-fold
more potent than the lead compound. Some E compounds
also exhibited apparent cytotoxicity (IC50 <10 mm) on several
tumor cell lines (Table 5). After a few failed attempts, we finally
obtained some compounds, especially those E compounds
with the 3,4-dihydroxyphenyl moiety, that have better potency
for the target proteins and tumor cells than the lead com-
pound. The binding of the E compounds to the target proteins
and characterization of their apoptosis-inducing mechanism
will be described and discussed in the following sections.
2. Binding of the thiophene-2-sulfonamide compounds to
the Bcl-2 family proteins
Compounds YCW-E1–E21 were tested in our FP-based assay to
measure their binding affinities to the three target proteins,
that is, Bcl-xL, Bcl-2, and Mcl-1. Most of them have micromolar
or sub-micromolar binding affinities to Bcl-2 and Mcl-
1 (Table 4). The most potent compounds for Mcl-1 are YCW-E4,
-E5, -E10, and -E11, which have inhibition constants between
0.3–0.5 mm. These four compounds are also the most potent
ones on Bcl-2 with Ki values of around 1 mm. Another notable
observation was that compounds YCW-E1–E11 have no or
rather weak binding affinities to Bcl-xL. The selectivity of our E
compounds towards Mcl-1 and Bcl-2 is interesting among
other known Bcl-2 inhibitors.
Figure 2. Binding modes of a) the lead compound BCL-VS-156 and b) one of
our most potent compounds (YCW-E5) with Mcl-1, as predicted by molecular
modeling. The ligand molecule is rendered as a ball-and-stick model. The
molecular surface of Mcl-1 is colored on a lipophilicity scale, in which polar
regions are blue and green and lipophilic regions are brown. The locations
of the binding pocket residues are labeled: hydrophobic residues have
white labels and polar residues have cyan labels. These pictures were pre-
pared by using the SYBYL software. Other representations of these two
binding modes are given in the Supporting Information (part II).
of most of the compounds in this set (only one exception) are
over 90%, and their purities are all above 90%.
Compounds YCW-D1–D18 were then tested in our FP-based
assay to measure their binding affinities with the three antia-
poptotic Bcl-2 family proteins. Again, none of them exhibited
obvious activity in our binding assay. But quite a few of the D
compounds exhibited obvious cytotoxicity (IC50 <10 mm) on
tumor cell lines (Table 5), which indicated improved cell perme-
ability of this structural scaffold relative to the A/B/C com-
pounds described above. This observation encouraged us to
explore other modifications on this structural scaffold.
Two things regarding the structure–activity relationship of
the E compounds need to be discussed in particular. Firstly, rel-
ative to the D compounds, the major difference in the YCW-
E1–E11 compounds is the introduction of the 3,4-dihydroxy-
phenyl moiety. Given the fact that the D compounds do not
have binding affinity to the three target proteins, this moiety
apparently plays a critical role here. In fact, compounds YCW-
E12–E21 were synthesized intentionally to replace the two hy-
droxy groups on this moiety with methoxy groups. The result-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1441

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
ing compounds indeed lost their
binding affinities to Mcl-1 and
Bcl-2 completely, which proved
our speculation. Secondly, some
of the active E compounds have
a urea moiety instead of the thi-
ourea moiety inherited from the
lead compound. The binding af-
finities of these urea-containing
compounds (YCW-E6–E11) are
comparable to those of the thio-
urea-containing
compounds
(YCW-E1–E5). This trend is differ-
ent from what was observed
among the A/B compounds,
among which only the thiourea-
containing compounds have
binding affinities to the target
proteins. In addition, some A/B
compounds have obvious bind-
ing affinities (Ki =5–10 mm) to
Bcl-xL. In contrast, all of the E
compounds have no or rather
weak binding affinities to Bcl-xL.
Thus, the structure–activity rela-
tionship of the E compounds is
notably different from that of
the A/B compounds, despite the
fact that these two sets of com-
pounds share a certain level of
structural similarity.
Scheme 3. Synthesis of compounds YCW-C1–C18. Reagents and conditions: a) 1. NaNO2, H2SO4, 08C; 2. NaN3,
08C!RT, 94%; b) CuI, CH3CN, ethyl propiolate, RT, 76%; c) LiOH, MeOH, reflux, 99%; d) 4-methylmorpholine, EDCI,
DMF, glycinamide hydrochloride, RT, 79%; e) CH3CN, MeOH, PIFA, RT, 35%; f) aryl isothiocyanate, THF, RT; g) 1,2-
ethanediamine, MeOH, reflux, 73%. EDCI: 3-(3-dimethylaminopropyl)-1-ethylcarbodiimide; PIFA: phenyliodine (III)
bis(trifloroacetate).
Scheme 4. Synthesis of compounds YCW-D1–D18. Reagents and conditions: a) 1. NaNO2, H2SO4, 08C; 2. KI, 08C!RT; b) 1. nBuLi, THF, ꢂ788C; 2. B(OMe)3;
3. HCl, RT; c) Et3N, CH2Cl2, 08C!RT, 85%; d) Pd(OAc)2, Sphos, K3PO4, THF, reflux, 17: 50%, 22a: 44%, 22b: 61% ; e) CF3COOH, CH2Cl2, RT, 18: 95%, 23a: 95%,
23b: 94%; f) aryl isothiocyanate, THF, RT. Sphos: dicyclohexylphosphino-2’,6’-dimethoxybiphenyl; Boc: tert-butoxycarbonyl.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1442

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
The obtained 15N-HSQC spectrum of Bcl-xL with
and without YCW-E5 is shown in Figure 3a as an ex-
ample. The 15N-HSQC spectra of Bcl-xL with YCW-E10,
YCW-E11, and ABT-737 are given in the Supporting
Information (part I). Many residues on Bcl-xL exhibited
apparent chemical shifts upon addition of YCW-E5
(Figure 3b). An overall similarity can be observed be-
tween the chemical shift patterns with YCW-E5 and
ABT-737. To illustrate this point, the nine residues
showing the largest chemical shifts (in W values)
upon the addition of YCW-E5 are labeled on the
complex structure formed by Bcl-xL and ABT-737 in
Figure 3c. One can see that these residues indeed
surround the BH3-binding groove on Bcl-xL, which
suggests that YCW-E5 binds with Bcl-xL with the
same pattern as ABT-737. The results obtained on
Bcl-xL, of course, need to be transferred to Mcl-1 or
Bcl-2 with care. However, because Mcl-1 and Bcl-2 do
not really have a second major binding site on the
molecular surface, it is reasonable to assume that
YCW-E5, and the other active E compounds, also
bind to Mcl-1 and Bcl-2 in the BH3-domain binding
groove.
Table 4. Chemical structures and binding data for the E compounds.
Compd
X
R1
R2
R3
Yield
[%]
Ki [mm][a]
Bcl-2
Bcl-xL
Mcl-1
YCW-E1
YCW-E2
YCW-E3
YCW-E4
YCW-E5
YCW-E6
YCW-E7
YCW-E8
YCW-E9
YCW-E10
YCW-E11
YCW-E12
YCW-E13
YCW-E14
YCW-E15
YCW-E16
YCW-E17
YCW-E18
YCW-E19
YCW-E20
YCW-E21
S
S
S
S
3-MeO
3-NO2
4-MeO
4-NO2
3-CF3
3-MeO
3-NO2
4-MeO
4-NO2
3-CF3
3-Cl
H
H
H
H
OH
OH
OH
OH
39
57
45
69
85
66
82
90
35
28
69
99
97
79
97
95
97
92
94
91
87
N.A.
N.A.
weak
weak
weak
N.A.
N.A.
N.A.
N.A.
weak
weak
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
3.9ꢁ0.8
1.7ꢁ1.0
4.2ꢁ0.9
1.2ꢁ0.2
1.3ꢁ0.2
3.2ꢁ0.3
1.2ꢁ0.2
3.1ꢁ0.7
1.2ꢁ0.3
1.2ꢁ0.1
0.83ꢁ0.05
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
17ꢁ8.3
1.5ꢁ0.2
2.0ꢁ0.5
0.50ꢁ0.07
0.39ꢁ0.04
12.0ꢁ4.4
2.2ꢁ0.9
weak
0.98ꢁ0.03
0.33ꢁ0.03
0.33ꢁ0.13
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
S
5-CF3 OH
O
O
O
O
O
O
S
S
S
S
S
O
O
O
O
O
H
H
H
H
OH
OH
OH
OH
5-CF3 OH
5-Cl
H
H
H
OH
H
OMe
OMe
OMe
OMe
3-NO2
4-MeO
4-NO2
3-CF3
3-NO2
4-MeO
4-NO2
3-Cl
H
5-CF3 OMe
H
H
H
5-Cl
5-CF3 OMe
OMe
OMe
OMe
OMe
We then employed molecular modeling to derive
the binding model of YCW-E5 to Mcl-1. This job was
conducted through molecular docking, followed by
molecular dynamics simulation to allow local confor-
mation flexibility on the protein molecule. To make
a comparison, the binding mode of the lead com-
pound (BCL-VS-156) to Mcl-1 was also derived with
the same methods. The final binding modes of the
lead compound and YCW-E5 are shown in Figure 2a
and b, respectively. Two subsites can be observed in
3-CF3
N.A.
N.A.
[a] The mean values and standard deviations of the Ki values were derived based on
the outcomes of three parallel measurements. N.A.: No obvious sign of binding was
observed; weak: dose-dependent FP signals were observed, but the change in the FP
signals (up to 100 mm concentration) was not significant enough to determine the Ki
value accurately.
In order to obtain some experimental proof for the binding
mode of our active compounds, two-dimensional heteronu-
clear single quantum coherence (HSQC) spectroscopy was em-
ployed in our study. Three of the most promising compounds,
that is, YCW-E5, -E10, and -E11, were selected to test their in-
teractions with 15N-labeled Bcl-xL in this experiment. ABT-737,
a nanomolar binder to Bcl-xL/Bcl-2, was used as the positive
control here. Ideally, Mcl-1 or Bcl-2 should have been used in
this experiment because our compounds exhibited tighter
binding to Mcl-1 or Bcl-2 than Bcl-xL. However, samples of 15N-
labeled Mcl-1 or Bcl-2 were not available to us, so 15N-labeled
Bcl-xL was applied instead. Note that we concluded that YCW-
E5, -E10, and -E11 have “weak” binding affinities to Bcl-xL,
based on the outcomes of our FP-based binding assay
(Table 4). This means that a dose-dependent response in the
FP signal was observed for the given compound, but the gross
change in the FP signal was not significant enough, even at
the highest concentration (100 mm) tested in our assay. Conse-
quently, the inhibition constant of this compound could not
be derived accurately. However, HSQC spectroscopy is a more
sensitive technique for detecting weak bindings in the high-
micromolar range. Thus, we expected that HSQC measure-
ments for our compounds on Bcl-xL could provide useful infor-
mation.
the BH3-binding groove on Mcl-1, which will be referred to as
the
“L-site” and the “R-site” below according to their locations on
the left or right in Figure 2. In a very recent work published by
Tanaka et al.,[34] these two sites are also referred to as the
“western region” and the “eastern region”, respectively. The
L-site is essentially hydrophobic and is formed mainly by
Phe228, Met231, Met250, Val253, and Phe270. The R-site is
more amphiphilic and is formed by hydrophobic residues, such
as Val216, Val265, and Phe319, as well as the polar residues
Asn260 and Arg263. In terms of shape, the L-site is deeper and
more constrained, whereas the R-site is more open to the sol-
vent. These two subsites are connected by a shallow channel,
with His224 and Thr266 located on either side of it. Our pre-
dicted binding modes suggest that the 2-phenoxyphenyl
moiety of both BCL-VS-156 and YCW-E5 is buried at the L-site,
whereas the phenylurea (or phenylthiourea) moiety occupies
the R-site. The remaining parts of the chemical structure, in-
cluding the five-membered ring and the substituent group on
it, go through the channel between the two subsites.
We hope to reach a qualitative explanation of the basic
structure–activity relationship observed among our com-
pounds with these predicted binding modes. The most impor-
tant aspect in the binding modes of our compounds is how
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1443

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Scheme 5. Synthesis of compounds YCW-E1–E21. Reagents and conditions: a) 10% Pd/C, 85% N2H4·H2O, EtOH, RT, 85%; b) 1. NaNO2, H2SO4, 08C; 2. KI, 08C!
RT, 87%; c) BBr3, CH2Cl2, RT; d) TBSCl, imidazole, DMAP, CH2Cl2, RT, 93% for 2 steps; e) 1. nBuLi, THF, ꢂ788C; 2. B(OMe)3; 3. HCl, RT, 28: 93%, 32: 55%; f) 16,
Pd(OAc)2, Sphos, K3PO4, THF, reflux, 30: 70%, 33: 70%; g) CF3COOH, CH2Cl2, RT, 31: 99%, 34: 90%; h) aryl isothiocyanate, CH2Cl2, RT; i) 40% HF, MeCN, RT. TBS:
tert-butyldimethylsilyl; DMAP: 4-dimethylaminopyridine.
they fit into the L-site. Without exception, all known potent in-
hibitors of Bcl-2 family proteins are able to occupy the L-site.
That is why we kept the 2-phenoxyphenyl moiety unchanged
among the several sets of compounds synthesized by us. In
particular, although the L-site is essentially hydrophobic, our
model suggests that one of the two hydroxy groups on the
terminal phenyl ring of YCW-E5 forms a hydrogen bond with
the backbone amide group of Met250 (Figure 2b). This hydro-
gen bond becomes possible only after the L-site of Mcl-1 has
undergone considerable conformational change. This seems to
be a new feature of some small-molecule inhibitors of Bcl-2
family proteins that has been reported recently. For example,
Lessene et al. reported that WEHI-539, a selective Bcl-XL inhibi-
tor with nanomolar potency, forms hydrogen bonds with the
backbone amide groups of Ser106 and Leu108 at the L-site on
Bcl-xL (Protein Data Bank (PDB) entry 3ZLR).[33] Our results
showed that blocking these two hydroxy groups, as in com-
pounds YCW-E12–E21, led to a total loss of binding affinity to
either Mcl-1 or Bcl-2, which indicates the importance of this
possible hydrogen bond. This observation also suggests that
the terminal 3,4-dihydroxyphenyl moiety on our E compounds
fits into the L-site so tightly that even small extra methyl
groups are not allowed.
The phenylurea moiety, which is at the other end of the
chemical structure in our compounds, is predicted to occupy
the R-site on Mcl-1. The binding data for the E compounds
show that the introduction of a few different types of substitu-
ent groups onto the phenyl ring produces up to 50-fold differ-
ence in binding affinity to Mcl-1 and 5-fold difference in bind-
ing affinity to Bcl-2 (Table 4). The R-site is relatively wide and
open, so the phenylurea moiety does not fit into it in a precise
binding pose. In fact, it was observed to “swing” inside the
R-site during the molecular dynamics simulation. It is not
straightforward to interpret the effect on the binding affinity
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1444

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
enhances the binding to Mcl-1. These two hydrogen bonds
were fairly stable during the molecular dynamics simulations
of this complex model (see the Supporting Information,
part II). Note that His224 is conserved among Bcl-2 family pro-
teins, but Thr266 is unique on Mcl-1. This may explain the
moderate selectivity of the E compounds towards Mcl-1.
Table 5. Cytotoxicity of the D and E compounds on six cell lines as mea-
sured in MTT assays.
Compd[a]
IC50 [mm][b]
BEAS-2B
A549
H1299
HL-7702
7404
Hep3B
YCW-D9
YCW-D10
YCW-D11
YCW-D12
YCW-D13
YCW-D14
YCW-D15
YCW-D16
YCW-D17
YCW-D18
YCW-E1
>100
17
>100
63
4
>100
0.3
>100
11
>100
8
>100
50
>100
>100
>100
>100
15
>100
19
>100
36
22
>100
86
>100
76
>100
>100
>100
>100
33
>100
26
>100
64
>100
16
>100
20
>100
21
>100
33
10
>100
21
>100
15
12
>100
16
3. Characterization of the proapoptotic effects of thiophene-
2-sulfonamide compounds
5
2
>100
12
>100
23
>100
8
>100
11
Three compounds, that is, YCW-E5, YCW-E10, and YCW-E11,
were studied in four different types of assays to characterize
their apoptosis-inducing effects on HL-60 cells. The HL-60 cell
line was chosen because it is known to overexpress Bcl-2
family proteins and is often used for testing Bcl-2 inhibitors.[7,8]
These three compounds were chosen because they were the
most potent ones in the binding assay and they also exhibited
obvious cytotoxicity on tumor cells. ABT-737 was used as the
positive reference in all four assays.
12
2
1
7
2
6
12
9
>100
55
>100
>100
4
>100
>100
>100
>100
8
>100
61
>100
25
YCW-E2
YCW-E3
YCW-E4
YCW-E5
>100
>100
14
4
7
6
YCW-E6
YCW-E7
YCW-E8
YCW-E9
3
2
8
3
0.3
0.9
4
45
>100
>100
>100
>100
13
>100
>100
>100
>100
10
>100
>100
>100
>100
7
10
>100
30
17
>100
>100
>100
10
21
>100
18
>100
20
10
26
>100
13
>100
>100
>100
13
>100
>100
>100
>100
>100
13
In the first assay, flow cytometry was employed to detect
whether the compounds under test could induce apoptosis in
HL-60 cells. In this assay, viable cells were negative for both
propidium iodide (PI) and annexin V. The cells undergoing
early apoptosis were positive for annexin V and negative for PI,
whereas the cells undergoing late apoptosis and nonviable
cells, which could undergo apoptosis or necrosis, displayed
both high annexin V and PI labeling. Cell debris was positive
for PI and negative for annexin V. The fluorescence of annex-
in V and PI measured in our assays for ABT-737 and three other
compounds are given in the Supporting Information (part III).
A graphical summary of the results is given in Figure 4. One
can see that upon treatment with the three E compounds at
a concentration of around 10 mm after 24 h, up to 60% cells
were observed to undergo early apoptosis. In addition, the
apoptosis-inducing effect of all three compounds is clearly
dose-dependent. In this assay, YCW-E5 and YCW-E10 are mar-
ginally more potent than YCW-E11. However, they are all at
least 10-fold less potent than ABT-737. This level of difference
is understandable because ABT-737 is a highly potent inhibitor
of Bcl-xL/Bcl-2 (Ki <1 nm). Our compounds still need optimiza-
tion to improve the potency.
YCW-E10
YCW-E11
YCW-E12
YCW-E13
YCW-E14
YCW-E15
YCW-E16
YCW-E17
YCW-E18
YCW-E19
YCW-E20
YCW-E21
46
29
>100
>100
>100
>100
23
>100
>100
>100
41
0.7
>100
0.6
12
0.6
31
2
>100
9
15
4
43
39
21
20
>100
>100
13
>100
>100
17
66
39
88
26
0.3
13
16
12
38
22
14
8
32
13
[a] Compounds YCW-D1–D8 are not listed in this table because they were
all inactive in this assay. [b] The concentration at 50% inhibition of cell
growth (IC50) values were derived from the outcomes of three parallel
measurements.
of each substituent group on the phenyl ring. It seems that an
increase in the dipole of the phenyl moiety with electron-with-
drawing groups, such as CF3, Cl, or NO2, helps to achieve
a tighter binding, especially to Mcl-1. We assume that some fa-
vorable dipole–dipole interactions may exist between this sub-
stituted phenyl moiety and some polar groups on the sur-
rounding residues. However, this type of interaction may not
be as specific and stable as hydrogen bonds.
The second assay detected poly-ADP-ribose polymerase pro-
tein (PARP) cleavage and caspase-9 activation, both of which
are hallmarks of apoptosis through the mitochondrial pathway,
upon treatment of HL-60 cells with the given compounds.
PARP is a major substrate for activated caspases. Our western
blot results showed that the full-length PARP (116 kDa) was de-
graded and the cleaved PARP (89 kDa) increased, relative to
the results with the control group, after treatment with ABT-
737 or the other three given compounds (Figure 5). It was also
observed that the precursor procaspase-9 (47 kDa) was de-
graded and the level of cleaved caspase-9 (35 kDa and 37 kDa)
increased after treatment with the compounds (Figure 5).
The outcomes of the second assay suggested that our com-
pounds are able to induce apoptosis through the mitochondri-
al pathway, like ABT-737 does. Thus, the next assay investigat-
ed the effects of the given compounds on the mitochondrial
Our predicted binding modes also suggest that the linker
moiety, including the five-membered ring and the substituent
group on it, also plays an important role. Firstly, this linker
moiety has to take a bent conformation to allow the ligand
molecule to fill up both the L-site and the R-site (Figure 2a
and b). The tetrazole-2-acetamide moiety on the A/B com-
pounds and the thiophene-2-sulfonamide on the E compounds
are able to fulfill this requirement. In contrast, the linker
moiety on the C compounds is basically a rigid planar frag-
ment; thus, the C compounds cannot fill up the L-site and the
R-site simultaneously with a low-energy binding pose. Second-
ly, the sulfonamide group on the E compounds may form hy-
drogen bonds with His224 and Thr266 on Mcl-1, which further
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1445

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
membrane potential (Dym). The
Dym value was monitored by
using the fluorescent dye Rh123.
It is a cell-permeable cationic
dye, which preferentially parti-
tions into the mitochondria as
a result of their highly negative
Dym value. Mitochondrial mem-
brane depolarization results in
the loss of Rh123 from the mito-
chondria, and a decrease in in-
tracellular fluorescence. The
Rh123 fluorescence intensities
measured on HL-60 cells after
24 h of treatment with the given
compounds are given in the
Supporting Information (part III).
A graphical summary of the re-
sults is given in Figure 6. All of
the three E compounds exhibit-
ed obvious dose-dependent ef-
fects on the Dym value, like ABT-
737 does.
The last assay detected the re-
lease of cytochrome c, which is
another hallmark of apoptosis
through the mitochondria path-
way. In this assay, mitochondria
isolated from untreated HL-60
cells were treated with the given
compounds. Cytochrome c in
the mitochondrial supernatants
and pellets was then detected
by western blot experiments. As
compared with the control
group, the amount of cytochro-
me c in the mitochondrial pellets
was reduced, whereas the
amount of cytochrome c in the
mitochondrial supernatants was
increased upon treatment with
any of the three E compounds
(Figure 7). Again, ABT-737 exhib-
ited a stronger effect than our E
compounds, which is under-
standable when the high poten-
cy of this compound is consid-
ered.
Conclusions
Figure 3. a) Superimposed 15N-HSQC spectra of free Bcl-xL (in green) and Bcl-xL in complex with compound YCW-
E5 (in red). The nine residues with the largest chemical shifts (in W values) are labeled explicitly. b) Normalized W
values of all residues on Bcl-xL upon binding with YCW-E5. c) The known complex structure formed by Bcl-xL with
ABT-737 (PDB entry 2YXJ), in which ABT-737 is shown as a stick model and the backbone of Bcl-xL is shown as
gray ribbons. Locations of the nine residues with the largest chemical shifts upon binding with YCW-E5 are indi-
cated by purple spheres on this structure, which indicates that this compound binds at the same site on Bcl-xL as
ABT-737.
In this study, we aimed at opti-
mizing a lead compound, which
was
originally
discovered
through structure-based virtual
screening, as an inhibitor of anti-
apoptotic Bcl-2 family proteins.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1446

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Figure 7. Cytochrome c in mitochondrial supernatants and pellets detected
in a western blot experiment. Mitochondria isolated from untreated HL-60
cells were incubated with ABT-737 (2 mm) or the three tested compounds
(5 mm) at 378C for 30 min. The mitochondria pellets (Mito.) and supernatants
(Sup.) were then obtained by centrifugation. Porin was used as the mito-
chondria loading control.
Figure 4. Early apoptosis in HL-60 cells induced after 24 h of treatment of
the given compounds. Cells with annexin V and PI double staining were ex-
amined by flow cytometry to detect the occurrence of early apoptosis.
compounds do have micromolar binding affinities to the three
target proteins. However, they are not effective apoptosis in-
ducers on living tumor cells due to poor cell permeability.
Next, we synthesized a second set of 18 compounds by re-
placing the phenyltetrazole moiety with a phenyltriazole
moiety, that is, the C compounds. This set of compounds can
be synthesized neatly, but unfortunately, they do not have the
desired biological activities. We then synthesized the third set
of 39 compounds with a 3-phenylthiophene-2-sulfonamide
moiety, that is, the D/E compounds. Encouragingly, some
compounds among them exhibited both promising binding
affinities to the target proteins and obvious cytotoxicity on
tumor cells. The most potent compounds, that is, those E
compounds with a terminal 3,4-dihydroxyphenyl moiety, have
sub-micromolar binding affinities to Mcl-1 (Ki =0.3–0.4 mm) or
Bcl-2 (Ki ꢀ1 mm). As compared with the lead compound, these
compounds are about 10-fold more potent and also easier to
synthesize. Their selectivity towards Mcl-1 and Bcl-2 is also in-
teresting.
Figure 5. PARP cleavage and caspase-9 activation in HL-60 cells detected in
a western blot experiment. Cells were treated with ABT-737 (1 mm), YCW-E10
(7.5 mm), YCW-E5 (10 mm), and YCW-E11 (15 mm) at 378C for 24 h. Actin was
used as the protein loading control.
Three selected compounds, that is, YCW-E5, YCW-E10, and
YCW-E11, were studied in several apoptosis assays conducted
on HL-60 cells. Our results demonstrated that they are able to
induce cell apoptosis through the mitochondrial pathway. It re-
mains to be confirmed that these compounds induce apopto-
sis through Bcl-2/Mcl-1 inhibition in living cells by appropriate
assays. But, in every experiment that we have conducted,
these compounds caused similar effects to ABT-737, which
suggests that these compounds may target the antiapoptotic
Bcl-2 family proteins in living cells. We propose that the com-
pounds with the 3-phenylthiophene-2-sulfonamide core
moiety are worth further optimization as effective apoptosis in-
ducers.
Figure 6. Percentage of Rh123-negative HL-60 cells after 24 h of treatment
of the given compounds. All of the compounds under test exhibited obvi-
ous, dose-dependent effects on the mitochondrial membrane potential.
Experimental Section
Organic synthesis
This compound contains a phenyltetrazole moiety and a hydra-
zinecarbothioamide moiety and, therefore, represents a struc-
tural scaffold not observed among known Bcl-2 inhibitors. We
first synthesized a set of 25 derivative compounds with the
phenyltetrazole moiety, that is, the A/B compounds. This set of
General: All reagents and starting materials used in the syntheses
were purchased from Lancaster, Acros, and Shanghai Chemical Re-
agent Corp. and were used without further purification unless
specified.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1447

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
All NMR spectra were recorded at RT on a Bruker AV-400 400 MHz
spectrometer. They were calibrated by using residual CHCl3 (dH =
7.26 ppm) and CDCl3 (dC =77.16 ppm) or CH3OH (dH =3.31 ppm)
and CD3OD (dC =49.00 ppm) as internal references. The following
abbreviations are used to designate multiplicities: s: singlet; d:
doublet; t: triplet; q: quartet; m: multiplet; quint: quintet; br:
broad. High-resolution mass spectrometry (HRMS) was performed
on a Bruker APEXIII 7.0 Tesla ESI-FT mass spectrometer at a 4000 V
emitter voltage. The purity of the final compounds was determined
by the analytical HPLC method. An Agilent 1100 series HPLC with
an Agilent Zorbax Exlipse SB-C18 (4.6 mm, 5 mm particle sizes) re-
versed-phase column was used to perform analytical HPLC analy-
ses. The elution buffer was a gradient of H2O/CH3OH.
General procedure for the synthesis of compounds YCW-D1–
D18 (Scheme 4)
Aniline 5 was diazotized by NaNO2 and aq H2SO4 at 08C and then
treated with KI from 08C to RT to give 12. Boronic acid 13 was pro-
duced from 12 through lithium–halogen exchange at ꢂ788C in
THF, followed by borate trapping and then hydrolysis in 1.0m HCl
solution. The reaction of 14 with 15 in CH2Cl2 from 08C to RT gave
16. Suzuki cross-coupling between 13 and 16 catalyzed by
Pd(OAc)2 and Sphos with heating at reflux in THF gave 17. Remov-
al of the Boc group on 17 by CF3COOH in CH2Cl2 at RT gave 18.
Boronic acids 21 were synthesized from aniline 19 following the
same procedure as described for 13. Suzuki cross-coupling be-
tween 21 and 16 and subsequent steps to access 23 were per-
formed as for 18. Detailed procedures for each step are described
in the Supporting Information.
Synthetic protocols and characterization data for intermediates 2–
34 are given in the Supporting Information, along with spectral
data for all of the target compounds (YCW-A1–A13, B1–B12, C1–
C18, D1–D18, and E1–E21). For the synthesis of compounds not
described in the Supporting Information, please provide complete
protocols and missing details (equivs/reaction times/reaction tem-
peratures/yields etc.) to allow other to repeat your work.
A solution of 18 (60 mg) in THF (3 mL) was treated with the appro-
priate aryl isothiocyanate (1.1 equiv). The solution was stirred at RT
until the starting material disappeared. Then, the solvent was re-
moved under reduced pressure. The residue was purified by chro-
matography (EtOAc/petroleum ether 1:1) to obtain compounds
YCW-D1–D8 as solids with yields ranging from 95–99% (Table 3).
The reaction of 23 with the appropriate aryl isothiocyanate under
analogous conditions gave YCW-D9–D18 with yields ranging from
89–99% (Table 3).
General procedure for the synthesis of YCW-A1–A13 and YCW-
B1–B12 (Scheme 2)
Compound 2 was obtained through the reaction of 1 with NaN3 in
the presence of Et3N·HCl as a catalyst with heating at reflux in DMF
for 12 h. Compound 2 was further treated with methyl bromoace-
tate in the presence of NaH as a base with heating at reflux in THF
for 5 h to give 3a and 3b. The reaction of 3a or 3b with
NH2NH2·H2O with heating at reflux in THF for 12 h gave 4a or 4b,
respectively. Detailed procedures for each step are described in the
Supporting Information.
General procedure for the synthesis of compounds YCW-E1–E21
(Scheme 5)
Compound 24 was reduced by 85% N2H4·H2O in the presence of
Pd/C in EtOH at RT to give aniline 25. Compound 25 was diazo-
tized by NaNO2 and aq H2SO4 at 08C and then treated with KI from
08C to RT to give 26. The methyl groups were removed by BBr3 in
CH2Cl2 at RT, and then the diol group was protected with TBS to
give 28. Boronic acid 29 was produced from 28 through lithium–
halogen exchange at ꢂ788C in THF followed by borate trapping.
Suzuki cross-coupling between 29 and 16 catalyzed by Pd(OAc)2
and Sphos was conducted with heating at reflux in THF. The re-
moval of the Boc group by CF3COOH in CH2Cl2 at RT gave 31. Bor-
onic acid 32 was produced from 26 in an analogous way to 29 de-
scribed above. Suzuki cross-coupling between 32 and 16 and then
removal of the Boc group gave 34. Detailed procedures for each
step are described in the Supporting Information.
A solution of 4a in THF (3 mL) was treated with the appropriate
aryl isothiocyanate or isocyanate (1.2 equiv). The solution was
heated at reflux until the starting material disappeared. Then, the
reaction was cooled, concentrated and purified by chromatogra-
phy (EtOAc/petroleum ether 3:1) to obtain compounds YCW-A1–
A13 as solids with yields ranging from 59–89% (Table 1). The reac-
tion of 4b with the appropriate aryl isothiocyanate or isocyanate
under analogous conditions gave YCW-B1–B12 with yields ranging
from 84–91% (Table 1).
General procedure for the synthesis of compounds YCW-C1–C18
(Scheme 3)
A solution of 31 (70 mg) in CH2Cl2 (3 mL) was treated with the ap-
propriate aryl isothiocyanate (1.2 equiv). The solution was stirred at
RT until the starting material disappeared. Then, the solvent was
removed under reduced pressure to obtain the protected inter-
mediate. To remove the TBS groups, the intermediate was dis-
solved in a solution of CH3CN (2 mL) and 40% HF (0.2 mL). The so-
lution was stirred at RT until the intermediate disappeared. The so-
lution was diluted with EtOAc and washed with distd H2O and
then saturated aq NaCl. The organic layer was collected and dried
by anhyd Na2SO4. The residue was filtered and condensed and pu-
rified by chromatography (EtOAc/petroleum ether 1:1) to obtain
compounds YCW-E1–E11 as solids with yields ranging from 28–
90% (Table 4). The reaction of 34 with the appropriate aryl isothio-
cyanate under analogous conditions described for the reaction of
31 (deprotection not required) gave YCW-E12–E21 with yields
ranging from 79–99% (Table 4).
Aniline 5 was diazotized by NaNO2 and aq H2SO4 at 08C and then
treated with NaN3 from 08C to RT to give azide 6 in high yields.
Compound 7 was obtained by the reaction of 6 with ethyl propio-
late in the presence of CuI as the catalyst in CH3CN at RT. Saponifi-
cation of 7 by LiOH in MeOH gave acid 8. EDCI-mediated conden-
sation of 8 with glycinamide hydrochloride in DMF at RT gave
amide 9, which was then treated with PIFA to obtain 10 through
the Hofmann rearrangement. The reaction of 7 with 1,2-ethanedia-
mine and with heating at reflux in MeOH gave 11. Detailed proce-
dures for each step are described in the Supporting Information.
A solution of 10 (50 mg) in THF (3 mL) was treated with the appro-
priate aryl isothiocyanate (1.2 equiv). The solution was stirred at RT
until the starting material disappeared. Then, the solvent was re-
moved under reduced pressure. The residue was purified by chro-
matography (EtOAc/petroleum ether 1:1) to obtain compounds
YCW-C1–C9 with yields ranging from 86–93% (Table 2). The reac-
tion of 11 with the appropriate aryl isothiocyanate under analo-
gous conditions gave YCW-C10–C18 with yields ranging from 76–
92% (Table 2).
In vitro binding assays
An FP-based assay was employed in this study to measure the
binding affinities of small-molecule compounds to three antiapop-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1448

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
totic Bcl-2 family proteins, including Mcl-1, Bcl-xL, and Bcl-2. The
5-FAM-Bid-BH3 peptide, that is, a 26-residue peptide (QEDIIR-
NIARHLAQVGDSMDRSIPPG) with a 5-carboxyfluorescein (5-FAM)
label on its N terminus, was used as the fluorescence tracer in this
assay. In each measurement, the target protein was incubated with
the Bid-BH3 peptide first, and then competitive binding of the
given compound was monitored by FP signal changes upon addi-
tion of the compound in multiple doses. Methods for preparing
the protein samples and detailed experimental settings in the FP
measurement can be found in the Supporting Information (part V).
then purified on Superdex75 (GE Healthcare) in phosphate-buf-
fered saline (PBS) solution.
15N-Heteronuclear single quantum correlation (HSQC) NMR spectra
were recorded on a Bruker DMX 600 MHz NMR spectrometer at
258C with samples containing 200–500 mm 15N-labeled protein. The
ratio of each tested compound and the Bcl-xL protein in the
sample was 1:1. The resulting NMR spectra were processed and an-
alyzed by using the nmrPipe and nmrDraw software (http://spin.
niddk.nih.gov/NMRPipe/). The HSQC spectrum of free Bcl-xL, which
is publicly available from the Biological Magnetic Resonance Data
Bank (http://www.bmrb.wisc.edu/; access number: 18250), was
adopted as the reference for the assignment of chemical shifts. In
order to quantify the chemical shift of each residue on the two-di-
mensional HSQC spectra, the W value was computed by using
Equation (2).
In this study, each compound was initially tested at three concen-
trations of 1, 10, and 50 mm against all three target proteins (Mcl-1,
Bcl-xL, and Bcl-2). At each concentration, the average FP values (in
mP units) of three parallel measurements were used to compute
the inhibition ratio by using Equation (1).
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
mPmax ꢂ mPtest
2
2
W ¼ ð0:2 ꢄ Dd15NÞ þðDd1HÞ
ð2Þ
ð1Þ
Inhibition ½%ꢃ ¼
mPmax ꢂ mPNC
In this study, three compounds, that is, YCW-E5, -E10, and -E11,
were selected to be tested by using the protocols described
above. ABT-737, a well-known Bcl-2 inhibitor that binds to the BH3
binding groove on Bcl-xL, was also tested as a positive reference.
Here, mPmax refers to the FP signal detected when the protein was
incubated with the Bid-BH3 peptide, mPNC refers to the FP signal
detected from the negative control, that is, the Bid-BH3 peptide
alone, and mPtest refers to the FP signal detected when the com-
pound under test was added at a certain concentration. If the com-
pound under test showed obvious dose-dependent competitive
binding to the target protein and it achieved an inhibition ratio
over 50% at 50 mm, it was then further tested at 7 different con-
centrations of 1, 10, and 100 nm and 1, 10, 50, and 100 mm, to
obtain a complete dose-dependent binding curve. The binding
curve was derived through nonlinear fitting by using the GraphPad
Prism software (version 5; http://www.graphpad.com/). The con-
centration of the given compound at which 50% of the bound
peptide was displaced (IC50) was derived from the binding curve.
The competitive inhibition constant of the compound under test
was calculated with a mathematical equation developed by Wang
and co-workers,[44] with the assumption that it formed a binary
complex with the target protein.
Molecular modeling
The binding modes of the lead compound BCL-VS-156 and of
YCW-E5 to Mcl-1 were predicted by molecular modeling. As the
first step, molecular docking was employed to derive a rough bind-
ing mode for each given molecule. The complex structure formed
between human Mcl-1 and the mNoxaB-BH3 peptide (PDB entry
2NLA)[46] was used for this task. The molecular structures of BCL-
VS-156 and YCW-E5 were sketched with the SYBYL software (ver-
sion 8.1; http://www.certara.com/) and then optimized with the
MMFF94 force field. Automatic docking of each molecule to Mcl-
1 was performed by using the GOLD software (version 5.1, Cam-
bridge Crystallographic Data Centre, http://www.ccdc.cam.ac.uk/).
The binding site on Mcl-1 for performing docking was defined as
all amino acid residues within 10 ꢂ from Thr266, a residue roughly
at the center of this binding site. The GOLD software employs a ge-
netic algorithm (GA) to control the docking process. In our study,
the whole population was placed on 5 separate islands with 100
individuals on each island. The total number of GA operations to
be performed was set to 100000. Probabilities for crossover, muta-
tion, and migration operations were set to 95, 95, and 10, respec-
tively. The ChemPLP scoring function was adopted to compute the
binding scores. A total of 30 final binding poses were generated
for the input molecule. These binding poses were clustered with
a root mean square deviation (RMSD) cutoff of 1.5 ꢂ by using the
“rms_analysis” tool included in the GOLD package. Among the
binding poses with the highest binding scores in each cluster, one
was selected as the most reasonable one after visual examination.
15N-Heteronuclear single quantum coherence spectroscopy
The Bcl-xL protein used in this experiment was a special truncated
construction of the full-length protein, with deletion of residues
45–84 on a long loop region and residues 210–233 at the C termi-
nus, which has been reported repeatedly in the literature for solv-
ing Bcl-xL structures. This construction was proven to retain the
biological function of the wild type.[45] An 8ꢁHis tag was added to
the N terminus. This protein was cloned into the pSJ2 vector (a
modified vector based on pET28a) at the EcoRI and XhoI sites, by
using the following oligonucleotides: 5’-TCTCGAATTCATGTCTCA-
GAGCAACCGGGA-3’ and 5’-GGTCCTCGAGTCAGCGTTCCTGGCCC-
TTTCG-3’. It was expressed in Escherichia coli BL21(DE3) cells. Cells
were grown at 378C in M9 medium containing 15N-ammonium
chloride and 1 mm ampicillin to an optical density at 600 nm
(OD600) of 0.6. Protein expression was induced by addition of
0.4 mm isopropyl-b-d-thiogalactopyranoside (IPTG) at 208C for
16 h. Cells were lysed in 25 mm tris(hydroxymethyl)aminomethane
(Tris)-HCl buffer at pH 7.0 containing 100 mm NaCl and
0.1 mgmLꢂ1 phenylmethanesulfonyl fluoride (PMSF). 15N-labeled
His-TEV-Bcl-xL protein was purified from the soluble fraction by
using Ni-NTA resin (Qiagen). The 15N-labeled His-TEV-Bcl-xL protein
was further cleaved by TEV protease (0.5 mgmLꢂ1, 48C, 12 h) to
obtain 15N-labeled Bcl-xL protein. The 15N-labeled Bcl-xL protein was
The selected binding pose of the given molecule was then refined
through molecular dynamics (MD) simulation by using the AMBER
program (version 9, University of California, San Francisco; http://
www.ambermd.org/). The MD simulation was performed in an ex-
plicit water box for 10 ns. To set up the job, the force-field parame-
ters applied to the small-molecule ligand were prepared by apply-
ing the Antechamber module in the AMBER program. Atomic par-
tial charges on the small-molecule ligand were derived with the
RESP method, based on the HF/6-31G* computation results given
by the Gaussian software (version 09, released by Gaussian Inc.;
http://www.gaussian.com/). Atoms on the Mcl-1 protein were as-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1449

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
signed the PARM99 template charges implemented in the AMBER
program. All ionizable residues were set at their default protona-
tion states under a neutral pH value. The complex structure was
soaked in a box of TIP3P water molecules with a margin of 10 ꢂ
along each dimension. An appropriate number of counter ions
were added to neutralize the whole system. After these prepara-
tions, the complex structure was first relaxed by 100 cycles of
steepest descent minimization, followed by 4900 cycles of conju-
gated gradient minimization. In this process, the protein and the
ligand were constrained with a constant of 10.0 kcalmolꢂ1 ꢂ2. After
that, the whole system was further relaxed by 5000 cycles of mini-
mization without constraints. The whole system was then heated
up gradually from 0 to 300 K in 100 ps, followed by 10 ns MD sim-
ulation at a constant temperature of 300 K and a constant pressure
of 1 atm. During simulation, electrostatic interactions were calcu-
lated with the PME algorithm, and the distance cutoff of nonbond-
ed interactions was set as 12 ꢂ. The SHAKE algorithm was applied
to fix the lengths of all chemical bonds connecting hydrogen
atoms. On the final MD trajectory, a total of 10000 snapshots were
retrieved at 1 ps interval. These snapshots were grouped into clus-
ters based on mass-weighted RMSD values with a cutoff value of
1.5 ꢂ. The snapshot closest to the cluster center in the largest clus-
ter was selected as the representative model of the given complex.
The final models of BCL-VS-156 and YCW-E5 binding to Mcl-1 are
shown in Figure 2a and b. More details of these two models are
given in the Supporting Information (part II).
Apoptosis mechanism studies
Cell culture and treatment of compounds: Human HL-60 promye-
locytic leukemia cells (ATCC, Manassas, VA) were cultured in RPMI-
1640 medium (Hyclone) supplemented with 10% (v/v) FBS (Hy-
clone), 100 gmLꢂ1 streptomycin (Gibco), and 100 IUmLꢂ1 penicillin
(Gibco). Cells were incubated under a 5% CO2 humidified atmos-
phere at 378C. Cells were seeded into a 75 cm2 Corning cell culture
flask (for mitochondria isolation) or 6- or 12-well cell culture plates
(for drug stimulation) and maintained by addition of new medium
every day. Cells were plated at a density of 5ꢁ105 mL 24 h prior to
compound treatment. Before mitochondria isolation or drug in-
ducement of normal cells, cell viability was confirmed to be >95%
by using the trypan blue exclusion assay. ABT-737 and other tested
compounds were dissolved in DMSO to a final concentration lower
than 0.1%. The control cells were treated with the same volume of
the vehicle. The medium was cultured in the presence or absence
of multiple concentrations of ABT-737 (0, 0.10, 0.20, 0.50, 1.0, and
2.0 mm) or tested compounds (YCW-E10: 0, 1.0, 2.5, 5.0, 7.5, and
10 mm; YCW-E5 and E11: 0, 1.0, 2.5, 5.0, 10, and 15 mm) for 24 h, re-
spectively. Three parallel experiments were conducted for each
compound under the same conditions.
Flow cytometry analysis: The number of apoptotic cells induced
by the tested compounds was measured by flow cytometry with
an Alexa Fluor 488 Annexin V/Dead Cell Apoptosis Kit. The annex-
in V–PI assay evaluates phospholipids turning over from the inner
to the outer lipid layer of the plasma membrane, an event typically
associated with apoptosis. After being treated with ABT-737 or
target compounds at different concentrations for 24 h, HL-60 cells
were harvested and washed with cold PBS. Cell pellets were sus-
pended in 1ꢁ binding buffer at
a concentration of 1ꢁ
106 cellsmLꢂ1. Annexin V–Alexa Fluor 488 (5 mL) and PI (1 mL; the
final concentration of PI was 1 mgmLꢂ1) were added to cell suspen-
sion (100 mL) and vortexed gently. The stained samples were incu-
bated for 15 min at RT in the dark. An additional portion (400 mL)
1ꢁ binding buffer was added to each tube. Samples were analyzed
by a flow cytometer (FACSAria, BD Biosciences) equipped with BD
FACSDiva software (BD Biosciences) within 1 h. A total of 10000
events were acquired by using the green channel FL1 for Annex-
in V–Alexa Fluor 488 and the red channel FL3 for PI. Data present-
ed in this work were representative of those obtained from at least
three independent experiments done in triplicate.
Cytotoxicity assay
The cytotoxic activities of our compounds were tested by the stan-
dard MTT assay on six cell lines. This assay served as a basic mea-
sure in our study for characterizing the cell permeability of our
compounds. These six cell lines included the human liver tumor
cells 7404 and Hep3B, with normal liver HL-7702 cells as a control,
and the human lung tumor cells A549 and H1299, with human
bronchial epithelial BEAS-2B cells as a control. All cell lines were
originally obtained from the American Type Culture Collection
(ATCC, Manassas, VA) and were cultured in Dulbecco’s modified
Eagle’s medium (DMEM; H1299, Hep3B), RPMI-1640 medium (A549,
HL-7702, and 7404), or Beltsville embryo culture medium (BECM;
BEAS-2B) with 10% fetal bovine serum (FBS) and supplemented
with 1% penicillin/streptomycin (P/S). Cell Counting Kit-8 (CK-04)
was bought from Dojindo (Kumamoto, Japan) for the cytotoxicity
assay. All of the tested compounds were dissolved to three con-
centrations of 0.50, 5.0, and 50 mm, respectively, in which the
DMSO concentration in each final solution was lower than 0.04%.
Cytotoxicity was determined by using Cell Counting Kit-8 (CCK-8,
Dojindo, Japan). In each test, 4000 cells were seeded in each well
on 96-well plates, which were then treated with the given com-
pound at the three final concentrations for 48 h. After treatment,
the CCK-8 solution (10 mL) was added to each well. Plates were in-
cubated for an additional 2–4 h at 378C. The supernatant was care-
fully removed, and DMSO (100 mL) was added to dissolve the for-
mazan crystals. The absorbance at 450 nm was then recorded by
using a SpectraMax-190 microplate reader (Molecular Devices,
USA) to calculate the inhibition rates. If necessary, the concentra-
tion at 50% inhibition (IC50) of each tested compound was derived
through nonlinear fitting of the average inhibition rates read from
three independent wells.
Western blot detection of caspase-9 activation: HL-60 cells were
lysed in lysis buffer (1% Triton X-100, 50 mm Tris (pH 7.4), and
150 mm NaCl, containing protease inhibitors (Complete, Roche))
maintained on ice and vortexed gently every 10 min. The sample
was centrifuged for 15 min at 12000 g and 48C to get rid of cell
debris. The protein concentration was assayed by using bicincho-
ninic acid (Pierce). The cytosolic protein (40 mg) was then subjected
to electrophoresis on 10% sodium dodecylsulfate (SDS) polyacryl-
amide gels, by using a Bio-Rad mini-gel apparatus, and blotted to
polyvinylidene difluoride (PVDF) membranes (Millipore, MA, USA)
through electrotransfer. Membranes were blocked for 1 h with 5%
nonfat dry milk in Tris-buffered saline containing 0.1% Tween-20
(TBST) and then incubated with antibodies (mouse monoclonal
anti-caspase-9 antibodies at a 1:1000 dilution, rabbit monoclonal
anti-PARP antibodies at a 1:1000 dilution, and mouse monoclonal
anti-Actin antibodies at a 1:2000 dilution) overnight at 48C. Primary
antibodies were labeled with goat anti-mouse immunoglobulin G
(IgG) at a 1:5000 dilution or goat anti-rabbit IgG at a 1:3000 dilu-
tion conjugated with horseradish peroxidase (Pierce) and shocked
on a shaking table for 1 h at RT. Immunoreactive bands were de-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1450

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
tected by incubation of the membranes with enhanced chemilumi-
nescence (ECL) reagents.
Keywords: antitumor agents
proteins · inhibitors · lead optimization · protein–protein
· apoptosis · Bcl-2 family
Measurement of the mitochondrial membrane potential (DYm):
The mitochondrial membrane potential (DYm) was monitored by
using the fluorescent dye Rh123. Rh123 was added at a final con-
centration of 1 mm for 30 min at 378C after HL-60 cells had been
treated with drugs at various concentrations and washed with PBS.
Cells were collected and washed twice with PBS. Fluorescence was
performed on a FACSaria flow cytometer by using the FL1 channel
(fluorescein isothiocyanate (FITC) channel).
interactions
[1] J. C. Reed, Nat. Rev. Drug Discovery 2002, 1, 111–121.
[2] S. W. Fesik, Nat. Rev. Cancer 2005, 5, 876–885.
[3] M. H. Andersen, J. C. Becker, P. Straten, Nat. Rev. Drug Discovery 2005, 4,
399–409.
[4] A. Vazquez, E. E. Bond, A. J. Levine, G. L. Bond, Nat. Rev. Drug Discovery
2008, 7, 979–987.
Western blot detection of cytochrome c release: Mitochondria
were freshly isolated from normal HL-60 cells. Briefly, cells (3ꢁ107)
were harvested and washed in ice-cold PBS for each mitochondria
isolation experiment, which was performed by using the Pierce Mi-
tochondria Isolation Kit (Pierce, Rockford, IL, USA) or the Dounce
homogenization method. Aliquots of the mitochondrial prepara-
tion were incubated with ABT-737 at 2 mm or tested compounds at
5 mm, respectively, at 378C for 30 min. After centrifugation at
12000 g for 5 min, the supernatants were analyzed for the pres-
ence of cytochrome c by western blotting. Mitochondrial pellets
were resuspended in lysis buffer and were also analyzed by west-
ern blotting.
[5] T. G. Cotter, Nat. Rev. Cancer 2009, 9, 501–507.
[6] S. Cory, J. M. Adams, Nat. Rev. Cancer 2002, 2, 647–656.
[7] G. Lessene, P. E. Czabotar, P. M. Colman, Nat. Rev. Drug Discovery 2008,
7, 989–1000.
[8] A. Ashkenazi, Nat. Rev. Drug Discovery 2008, 7, 1001–1012.
[9] M. R. Arkin, J. A. Wells, Nat. Rev. Drug Discovery 2004, 3, 301–317.
[10] L. Zhao, J. Chmielewski, Curr. Opin. Struct. Biol. 2005, 15, 31–34.
[11] J. A. Wells, C. L. McClendon, Nature 2007, 450, 1001–1009.
[12] T. Oltersdorf, S. W. Elmore, A. R. Shoemaker, R. C. Armstrong, D. J.
Augeri, B. A. Belli, M. Bruncko, T. L. Deckwerth, J. Dinges, P. J. Hajduk,
M. K. Joseph, S. Kitada, S. J. Korsmeyer, A. R. Kunzer, A. Letai, C. Li, M. J.
Mitten, D. G. Nettesheim, S. Ng, P. M. Nimmer, J. M. O’Connor, A. Oleksi-
jew, A. M. Petros, J. C. Reed, W. Shen, S. K. Tahir, C. B. Thompson, K. J.
Tomaselli, B. Wang, M. D. Wendt, H. Zhang, S. W. Fesik, S. H. Rosenberg,
Nature 2005, 435, 677–681.
[13] C. M. Park, M. Bruncko, J. Adickes, J. Bauch, H. Ding, A. Kunzer, K. C.
Marsh, P. Nimmer, A. R. Shoemaker, X. Song, S. K. Tahir, C. Tse, X. Wang,
M. D. Wendt, X. Yang, H. Zhang, S. W. Fesik, S. H. Rosenberg, S. W.
Elmore, J. Med. Chem. 2008, 51, 6902–6915.
[14] C. Tse, A. R. Shoemaker, J. Adickes, M. G. Anderson, J. Chen, S. Jin, E. F.
Johnson, K. C. Marsh, M. J. Mitten, P. Nimmer, L. Roberts, S. K. Tahir, Y.
Xiao, X. Yang, H. Zhang, S. Fesik, S. H. Rosenberg, S. W. Elmore, Cancer
Res. 2008, 68, 3421–3428.
Reagents and material: The Alexa Fluor 488 Annexin V/Dead Cell
Apoptosis Kit and MitoSOX Red mitochondrial superoxide indicator
(MitoSOX) were purchased from Molecular Probes (Eugene, OR,
USA). Anti-cytochrome c monoclonal antibody (mAb) and anti-
actin mAb were purchased from Santa Cruz Biotechnology (Santa
Cruz, CA, USA). Antiporin mAb was purchased from MitoSciences
(Eugene, OR, USA). Caspase-9 mouse mAb and PARP rabbit mAb
were purchased from Cell Signaling Technology (CST; MA, USA).
Rhodamine 123 (Rh123), trypsin, and sodium succinate were pur-
chased from Sigma (St. Louis, MO, USA). Unless stated otherwise,
other reagents used in this study were of analytical purity and ob-
tained from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, P. R.
China). Mitochondrial buffer (MT buffer) containing 250 mm su-
crose, 10 mm 2-[4-(2-hydroxyethyl)-1-piperazinyl]ethanesulfonic
acid (HEPES), 1 mm ethylene glycol tetraacetic acid (EGTA), 4.2 mm
sodium succinate, and 1 mm potassium dihydrogen phosphate
and adjusted to pH 7.0 with 1m potassium hydroxide, was used
for mitochondrial washing and staining. All of the buffers were fil-
tered through a 0.22 mm filter and used within three weeks.
[15] S. Kitada, M. Leone, S. Sareth, D. Zhai, J. C. Reed, M. Pellecchia, J. Med.
Chem. 2003, 46, 4259–4264.
[16] B. Becattini, S. Kitada, M. Leone, E. Monosov, S. Chandler, D. Zhai, T. J.
Kipps, J. C. Reed, M. Pellecchia, Chem. Biol. 2004, 11, 389–395.
[17] M. Nguyen, R. C. Marcellus, A. Roulston, M. Watson, L. Serfass, S. R. Mur-
thy Madiraju, D. Goulet, J. Viallet, L. Belec, X. Billot, S. Acoca, E. Purisima,
A. Wiegmans, L. Cluse, R. W. Johnstone, P. Beauparlant, G. C. Shore,
Proc. Natl. Acad. Sci. USA 2007, 104, 19512–19517.
[18] J. L. Wang, D. Liu, Z. J. Zhang, S. Shan, X. Han, S. M. Srinivasula, C. M.
Croce, E. S. Alnemri, Z. Huang, Proc. Natl. Acad. Sci. USA 2000, 97, 7124–
7129.
[19] A. Degterev, A. Lugovskoy, M. Cardone, B. Mulley, G. Wagner, T. Mitchi-
son, J. Yuan, Nat. Cell Biol. 2001, 3, 173–182.
[20] G. Wang, Z. Nikolovska-Coleska, C. Y. Yang, R. Wang, G. Tang, J. Guo, S.
Shangary, S. Qiu, W. Gao, D. Yang, J. Meagher, J. Stuckey, K. Krajewski, S.
Jiang, P. P. Roller, H. O. Abaan, Y. Tomita, S. Wang, J. Med. Chem. 2006,
49, 6139–6142.
[21] G. Tang, Z. Nikolovska-Coleska, S. Qiu, C. Y. Yang, J. Guo, S. Wang, J.
Med. Chem. 2008, 51, 717–720.
[22] L. Wang, D. T. Sloper, S. N. Addo, D. Tian, J. W. Slaton, C. Xing, Cancer
Res. 2008, 68, 4377–4383.
[23] J. Wei, S. Kitada, M. F. Rega, J. L. Stebbins, D. Zhai, J. Cellitti, H. Yuan, A.
Emdadi, R. Dahl, Z. Zhang, L. Yang, J. C. Reed, M. Pellecchia, J. Med.
Chem. 2009, 52, 4511–4523.
Supporting Information
The Supporting Information contains additional results of 15N-
HSQC spectroscopy, molecular modeling, and apoptotic mecha-
nism studies. Furthermore, protocols for the chemical synthesis or
intermediary compounds, along with characterization data for all
compounds, and more details of the FP-based binding assay are
also given in the Supporting Information.
[24] J. Wei, J. L. Stebbins, S. Kitada, R. Dash, W. Placzek, M. F. Rega, B. Wu, J.
Cellitti, D. Zhai, L. Yang, R. Dahl, P. B. Fisher, J. C. Reed, M. Pellecchia, J.
Med. Chem. 2010, 53, 4166–4176.
[25] Y. Feng, X. Ding, T. Chen, L. Chen, F. Liu, X. Jia, X. Luo, X. Shen, K. Chen,
H. Jiang, H. Wang, H. Liu, D. Liu, J. Med. Chem. 2010, 53, 3465–3479.
[26] Z. Zhang, G. Wu, F. Xie, T. Song, X. Chang, J. Med. Chem. 2011, 54,
1101–1105.
[27] B. E. Sleebs, P. E. Czabotar, W. J. Fairbrother, W. D. Fairlie, J. A. Flygare, D.
Huang, W. J. A. Kersten, M. F. T. Koehler, G. Lessene, K. Lowes, J. P. Pari-
sot, B. J. Smith, M. L. Smith, A. J. Souers, I. P. Street, H. Yang, J. B. Baell, J.
Med. Chem. 2011, 54, 1914–1926.
Acknowledgements
The authors are grateful for financial support from the Chinese
National Natural Science Foundation (grants no. 81172984,
21072213, 21002117, 21102168, 21102165, and 20921091) and
the 863 High-Tech program from the Chinese Ministry of Science
and Technology (grant no. 2012AA020308).
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1451

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
[28] H. Zhou, A. Aguilar, J. Chen, L. Bai, L. Liu, J. L. Meagher, C. Y. Yang, D.
McEachern, X. Cong, J. A. Stuckey, S. Wang, J. Med. Chem. 2012, 55,
6149–6161.
Ruvolo, T. McQueen, W. D. Schober, J. C. Watt, T. Jiffar, X. Ling, F. C.
Marini, D. Harris, M. Dietrich, Z. Estrov, J. McCubrey, W. S. May, J. C.
Reed, M. Andreeff, Cancer Cell 2006, 10, 375–388.
[29] J. Chen, H. Zhou, A. Aguilar, L. Liu, L. Bai, D. McEachern, C. Y. Yang, J. L.
Meagher, J. A. Stuckey, S. Wang, J. Med. Chem. 2012, 55, 8502–8514.
[30] A. J. Souers, J. D. Leverson, E. R. Boghaert, S. L. Ackler, N. D. Catron, J.
Chen, B. D. Dayton, H. Ding, S. H. Enschede, W. J. Fairbrother, D. C.
Huang, S. G. Hymowitz, S. Jin, S. L. Khaw, P. J. Kovar, L. T. Lam, J. Lee,
H. L. Maecker, K. C. Marsh, K. D. Mason, M. J. Mitten, P. M. Nimmer, A.
Oleksijew, C. H. Park, C. M. Park, D. C. Phillips, A. W. Roberts, D. Sampath,
J. F. Seymour, M. L. Smith, G. M. Sullivan, S. K. Tahir, C. Tse, M. D. Wendt,
Y. Xiao, J. C. Xue, H. Zhang, R. A. Humerickhouse, S. H. Rosenberg, S. W.
Elmore, Nat. Med. 2013, 19, 202–208.
[38] M. F. van Delft, A. H. Wei, K. D. Mason, C. J. Vandenberg, L. Chen, P. E.
Czabotar, S. N. Willis, C. L. Scott, C. L. Day, S. Cory, J. M. Adams, A. W.
Roberts, D. C. Huang, Cancer Cell 2006, 10, 389–399.
[39] C. Skoda, B. M. Erovic, V. Wachek, L. Vormittag, F. Wrba, H. Martinek, G.
Heiduschka, P. Kloimstein, E. Selzer, D. Thurnher, Oncol. Rep. 2008, 19,
1499–1503.
[40] B. Zhou, X. Li, Y. Li, Y. Xu, Z. Zhang, M. Zhou, X. Zhang, Z. Liu, J. Zhou,
C. Cao, B. Yu, R. Wang, ChemMedChem 2011, 6, 904–921.
[41] Y. Xu, M. Zhou, Y. Li, C. Li, Z. Zhang, B. Yu, R. Wang, ChemMedChem
2013, 8, 1345–1352.
[31] A. Friberg, D. Vigil, B. Zhao, R. N. Daniels, J. P. Burke, P. M. Garcia-Bar-
rantes, D. Camper, B. A. Chauder, T. Lee, E. T. Olejniczak, S. W. Fesik, J.
Med. Chem. 2013, 56, 15–30.
[32] A. Aguilar, H. Zhou, J. Chen, L. Liu, L. Bai, D. McEachern, C. Y. Yang, J.
Meagher, J. Stuckey, S. Wang, J. Med. Chem. 2013, 56, 3048–3067.
[33] G. Lessene, P. E. Czabotar, B. E. Sleebs, K. Zobel, K. N. Lowes, J. M.
Adams, J. B. Baell, P. M. Colman, K. Deshayes, W. J. Fairbrother, J. A. Fly-
gare, P. Gibbons, W. J. A. Kersten, S. Kulasegaram, R. M. Moss, J. P. Pari-
sot, B. J. Smith, I. P. Street, H. Yang, D. C. S. Huang, K. G. Watson, Nat.
Chem. Biol. 2013, 9, 390–397.
[42] X. Ding, Y. Li, L. Lv, M. Zhou, L. Han, Z. Zhang, Q. Ba, J. Li, H. Wang, H.
Liu, R. Wang, ChemMedChem 2013, 8, 1986–2014.
[43] H. C. Kolb, M. G. Finn, K. B. Sharpless, Angew. Chem. 2001, 113, 2056–
2075; Angew. Chem. Int. Ed. 2001, 40, 2004–2021.
[44] Z. Nikolovska-Coleska, R. Wang, X. Fang, H. Pan, Y. Tomita, P. Li, P. P.
Roller, K. Krajewski, N. G. Saito, J. A. Stuckey, S. Wang, Anal. Biochem.
2004, 332, 261–273.
[45] S. W. Muchmore, M. Sattler, H. Liang, R. P. Meadows, J. E. Harlan, H. S.
Yoon, D. Nettesheim, B. S. Chang, C. B. Thompson, S. L. Wong, S. L. Ng,
S. W. Fesik, Nature 1996, 381, 335–341.
[34] Y. Tanaka, K. Aikawa, G. Nishida, M. Homma, S. Sogabe, S. Igaki, Y.
Hayano, T. Sameshima, I. Miyahisa, T. Kawamoto, M. Tawada, Y. Imai, M.
Inazuka, N. Cho, Y. Imaeda, T. Ishikawa, J. Med. Chem. 2013, 56, 9635–
9645.
[46] P. E. Czabotar, E. F. Lee, M. F. van Delft, C. L. Day, B. J. Smith, D. C. Huang,
W. D. Fairlie, M. G. Hinds, P. M. Colman, Proc. Natl. Acad. Sci. USA 2007,
104, 6217–6222.
[35] L. Han, R. Wang, ChemMedChem 2013, 8, 1437–1440.
[36] R. J. Youle, A. Strasser, Nat. Rev. Mol. Cell Biol. 2008, 9, 47–59.
[37] M. Konopleva, R. Contractor, T. Tsao, I. Samudio, P. P. Ruvolo, S. Kitada,
X. Deng, D. Zhai, Y. X. Shi, T. Sneed, M. Verhaegen, M. Soengas, V. R.
Received: January 25, 2014
Published online on April 29, 2014
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1436 – 1452 1452