3-Thiomorpholin-8-oxo-8H-acenaphtho [1,2-b] pyrrole-9-carbonitrile (S1) derivatives as pan-Bcl-2-inhibitors of Bcl-2, Bcl-xL and Mcl-1

Article abstract of DOI:10.1016/j.bmc.2012.11.008

Based on the binding mode of our previously discovered dual inhibitor of Bcl-2 and Mcl-1, 3-thiomorpholin-8-oxo-8H-acenaphtho[1,2-b]pyrrole-9- carbonitrile (3, S1), a library of 9-substituted 3 derivatives was synthesized to further probe the p4 pocket of the two targets. By NMR, structure-activity relationship study, and site-directed mutation, compound 6d (3-(4- aminophenylthio)-8-oxo-8H-acenaphtho[1,2-b]pyrrole-9-3-phenyl)propylamine) was identified to span p2-p4 pockets of Mcl-1, Bcl-2 and Bcl-x_L, and then exhibited 9- to 35-fold better affinity to the three targets than 3 (IC ₅₀ = 10, 20 and 18 nM, respectively), which led to greater activity in induction of apoptosis in multiple cancer cell lines. Different contribution of p4 pocket to binding Bcl-2 and Mcl-1 was also investigated by plotting the potency and the HAC of the derivatives.

Full text of DOI:10.1016/j.bmc.2012.11.008

Bioorganic & Medicinal Chemistry 21 (2013) 11–20
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
3-Thiomorpholin-8-oxo-8H-acenaphtho [1,2-b] pyrrole-9-carbonitrile (S1)
derivatives as pan-Bcl-2-inhibitors of Bcl-2, Bcl-xL and Mcl-1 q
Ting Song a, , Xiangqian Li a, , Xilong Chang b,c, , Xiaomeng Liang a, Yan Zhao d, Guiye Wu a, Shenghui Xie a,
Pengchen Su a, Zhiyong Wu a, Yingang Feng e, Zhichao Zhang a,
⇑
a State Key Laboratory of Fine Chemicals, School of Chemistry, Dalian University of Technology, Dalian 116012, People’s Republic of China
b State Key Laboratory of Natural and Biomimetic Drugs, Peking University, Beijing 100191, People’s Republic of China
c Qilu Pharmaceutical Co., Ltd, Shandong 250000, People’s Republic of China
d School of Life Science and Technology, Dalian University of Technology, Dalian 116024, People’s Republic of China
e Qingdao Institute of BioEnergy and Bioprocess Technology, Chinese Academy of Sciences, Qingdao 266101, People’s Republic of China
a r t i c l e i n f o
a b s t r a c t
Article history:
Based on the binding mode of our previously discovered dual inhibitor of Bcl-2 and Mcl-1,
3-thiomorpholin-8-oxo-8H-acenaphtho[1,2-b]pyrrole-9-carbonitrile (3, S1), a library of 9-substituted 3
derivatives was synthesized to further probe the p4 pocket of the two targets. By NMR, structure–activity
relationship study, and site-directed mutation, compound 6d (3-(4-aminophenylthio)-8-oxo-8H-
acenaphtho[1,2-b]pyrrole-9-3-phenyl)propylamine) was identified to span p2–p4 pockets of Mcl-1,
Bcl-2 and Bcl-xL, and then exhibited 9- to 35-fold better affinity to the three targets than 3 (IC50 = 10,
20 and 18 nM, respectively), which led to greater activity in induction of apoptosis in multiple cancer cell
lines. Different contribution of p4 pocket to binding Bcl-2 and Mcl-1 was also investigated by plotting the
potency and the HAC of the derivatives.
Received 1 September 2012
Revised 28 October 2012
Accepted 8 November 2012
Available online 21 November 2012
Keywords:
Pan-Bcl-2 inhibitor
Bcl-2
Mcl-1
p4 Pocket
Apoptosis
Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction
and Bcl-xL are the primary focus for the design of BH3 mimetics,5
recent studies have demonstrated that Mcl-1 also plays an
Targeting the interface between proteins has huge therapeutic
potential, but discovering small molecule drugs that disrupt
protein–protein interactions (PPIs) is an enormous challenge.1,2
Some notable successes were achieved with the discovery of BH3
mimetics. These molecules modulate the PPIs between anti-apop-
totic and pro-apoptotic Bcl-2 members by occupying the BH3
groove shared by the two opposite functional groups and then in-
duce cancer cells apoptosis.3,4 Although the BH3 grooves of Bcl-2
important role for cancer cell survival. It is necessary to neutralize
both arms of the anti-apoptotic Bcl-2 family (Bcl-2/Bcl-xL and
Mcl-1) for apoptosis to occur in many cell types.6,7 The most potent
dual inhibitor (Bcl-2 and Bcl-xL) ABT-737, for example, is encoun-
tering resistance due to the inability to bind a more divergent
BH3 groove of Mcl-1 protein.8,9 As such, a promising drug-like
BH3 mimetic should be a ‘pan-Bcl-2 inhibitor’ that can bind at least
Bcl-2 and Mcl-1.3
We have previously reported the discovery of a small-molecule
Bcl-2 inhibitor, 3-thiomorpholin-8-oxo-8H-acenaphtho[1,2-b]pyr-
role-9-carbonitrile (3, S1).10–12 The in vivo apoptosis induction by
3 places it on the list of 19 preclinical Bcl-2 inhibitor antitumor
drugs.3 Further studies have described the biological mechanism
of 3 as an authentic BH3 mimetic and a dual inhibitor that targets
both Bcl-2 and Mcl-1.11 In this study, the binding site of 3 was
further confirmed by NMR and expanded to that of Bcl-xL.
Although the pan-Bcl-2 inhibition of 3 shows advantages over
ABT-737, its affinity toward Bcl-2 (Ki = 310 nM by fluorescence
polarization assay (FPA)) is much less potent than ABT-737
(1 nM).8 ABT-737 is a highly optimized ligand spanning the p2,
p3 and p4 pockets which are sub-active sites in the BH3 groove.
According to the X-ray structure of the Bim BH3 peptide in
Abbreviations: Bcl-2, B-cell lymphoma 2; Mcl-1, myeloid cell leukemia sequence
1; Bcl-xL, B-cell lymphoma x long; Bax, Bcl-2-associated x protein; Bak, Bcl-2
homologous antagonist/killer; BH3, Bcl-2 homology domain 3; SAR, Structure–
activity relationship; FPA, fluorescence polarization assay; IC50, the half maximal
inhibitory concentration; Kd, dissociation constant; HPLC, high-performance liquid
chromatography; ADT, AutoDock tools; ITC, isothermal titration calorimetry; NMR,
nuclear magnetic resonance; ELISA, enzyme-linked immunosorbent assay; CCSP,
combined chemical shift perturbation;
DG, binding free energy; HAC, heavy atom
count; PDB, protein data bank.
q
The 3D structures of human Mcl-1 (PDB ID: 2NLA), Mcl-1 in complex with Bim
(PDB ID: 2PQK) and Bcl-2 (PDB ID: 1GJH) were obtained from the protein bank in
the RCSB.
⇑
Corresponding author. Tel./fax: +86 411 84986032.
E-mail address: zczhang@dlut.edu.cn (Z. Zhang).
These authors contributed equally to this work.
0968-0896/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmc.2012.11.008

12
T. Song et al. / Bioorg. Med. Chem. 21 (2013) 11–20
complex with the Bcl-xL, Bcl-2, and Mcl-1 proteins, four conserved
hydrophobic residues (h1–h4) on one face of Bim helix insert into
four hydrophobic pockets (p1–p4) within the BH3 grooves of all
the three proteins.13 The p1–p4 pockets are 4 sub-active hotspots
for the structure based design of BH3 mimetics.14 By SAR studies,
our compound 3 may span p2 and p3 but lose p4 pocket.12
Here, we designed series of derivates based on the binding
mode of 3 which was further identified by NMR binding study in
the present work. We maintained the p2 occupying and expanded
to the p4. Many efforts were given to fit both Bcl-2 and Mcl-1 to
maintain the dual inhibition. We discovered a new potent pan-
Bcl-2 inhibitor 3-(4-aminophenylthio)-8-oxo-8H-acenaphtho[1,2-
b]pyrrole-9-(3-phenyl)propylamine (6d), which the IC50 value to
Mcl-1, Bcl-2 and Bcl-xL by ELISA was 10, 20 and 18 nM, respec-
tively. The 9- to 35-fold better affinity was achieved for the three
targets than its parent 3. Additionally, we defined the molecular
determinants governing the specificity of ligand binding to the
p4 pocket of Bcl-2 and Mcl-1.
Bcl-2-like proteins. The comparison of high-resolution structure
study of Bcl-xL/Bim and Mcl-1/Bim showed that R263 in Mcl-1 is
somewhat less solvent-exposed than its homologue in Bcl-xL.16
Since Bcl-2 and Bcl-xL have quite similar structure with overall
backbone RMSD (root mean square deviation) only 1.85 Å,17 we
concluded the R263 in Mcl-1 and its homologue in Bcl-2 might ren-
der an obvious difference in this region between the two proteins.
Consistently, our computational modeling studies illustrated that
the R263 of Mcl-1 was less solvent-exposed than R146 of Bcl-2
(Fig. 2a and b). We proposed that if 3 derivatives would like to
maintain the binding mode with R263 in Mcl-1 and R146 in Bcl-
2, efforts should be given to adapt to the difference. Additionally,
a recent molecular dynamics study has reported a greater open-
ness of the p4 binding sites on Mcl-1 than Bcl-2.15 Given these
findings, we inferred that the angle of the BH3 groove of two pro-
teins in the p4 region would be different when a molecule fits into
BH3 groove of Mcl-1 and Bcl-2. Bcl-2 could tolerate a relative open
angle, whereas a closed angle may be favored by Mcl-1. A previ-
ously molecule TM-179 also met our hypothesis. When it bound
to Bcl-2, the 2-hydroxyl was used for hydrogen bound with
R146. But an alternative 3-hydroxyl was used for corresponding
hydrogen bond when bound to Mcl-1.18 However, 3 contains one
hydrogen bond available group. We proposed when the interaction
with R263 or R146 was kept, a flexible linker group between the
core structure of 3 and the p4-occupying group at 9-position
should be chosen for the accommodation by both Bcl-2 and Mcl-1.
Additionally, because Bim BH3 peptide utilized F69 to occupy
p4 pocket, F69 becomes the mimicking goal of newly designed
group to occupy p4.19
2. Results and discussion
2.1. Rationale
We recently identified small-molecule 3 as an authentic BH3
mimetic and a dual inhibitor of Bcl-2 and Mcl-1. Herein, we in-
cluded Bcl-xL in competitive binding test and found that 3 had sim-
ilar binding affinities toward Bcl-xL with Bcl-2 (IC50 = 625 and
710 nM, respectively by ELISA assay) (Table 2 and Table S1). The
broader binding profile confirmed 3 as a pan-Bcl-2 inhibitor. Be-
cause the three-dimensional structure of Bcl-2 is very similar with
that of Bcl-xL but different with Mcl-1, we continuously focused on
BH3 grooves of Bcl-2 and Mcl-1 for lead optimization.
SAR study determined that the carbonyl substitution of 3 binds
near R263 of Mcl-1 and the homology of Bcl-2 termed R146 through
hydrogen bonds. Its 3-position substituent extends into the p2
pocket, whereas the 9-position cyano group points to but does
not access the p4 pocket of the two proteins (Fig. S1). 12 To further
identify the binding mode, here we performed a [15N, 1H] NMR
titration study. In good agreement with the SAR results, the NMR
spectra confirmed that 3 occupied the p2 and p3 pocket. The spectra
of Mcl-1 alone (shown in blue) showed well-dispersed peaks, indic-
ative of a folded and stable protein. Upon 3 addition, NMR spectra
generated many chemical shifts (shown in red), which are indica-
tive of tight binding (Fig. S2a). The combined chemical shift pertur-
bation (CCSP) signals revealed that more than 60% of the residues
perturbed above 0.02 ppm were located in the BH3 binding groove
of the protein (Fig. 1a). Among them, some residues located in p2
and p3 pockets (V253, M250, R263, L246, F270 and K234) experi-
ence average chemical shift changes of at least 0.04 ppm (Fig. 1c).
There were several residues that experienced intensity reductions
associated with line broadening. In particular, R263 and the nearing
V253 had large intensity changes following the addition of 3, to a
point where they were no longer detectable (Fig. S2b). It supported
the formation of the hydrogen bond in this area.
Lastly, the molecular planarity and symmetry of the core of 3
has a general negative impact on solubility. Introduction of addi-
tional groups at 9-position would lead to the disruption of molec-
ular planarity and symmetry, which could improve its solubility.20
Thus, optimizing at 9-position of 3 could improve not only its bind-
ing potency, but perhaps also its solubility.
2.2. Structure–activity relationship
We initially substituted the cyano with amino as a linker group,
whichmayrenderflexibilitytocompoundstoallowthemengagewell
into the p4 pocket of both Bcl-2 and Mcl-1. Meanwhile, amino could
facilitate solubility by forming hydrogen bond with water. We sought
to survey a variety of alkyl group in order to identify those win
appropriate trajectory and length to access p4 pocket, meanwhile
maintaining dual inhibition. Specifically, we examined ethylamine,
n-propylamine, n-butylamine, n-pentylamine, and n-hexylamine
groups, yielding analogues 3-thiomorpholin-8-oxo-8H-acenaphtho
[1,2-b]pyrrole-9-ethylamine (5a), 3-thiomorpholin-8-oxo-8H-ace-
naphtho[1,2-b]pyrrole-9-n-propylamine (5b) 3-thiomorpholin-
8-oxo-8H-acenaphtho[1,2-b]pyrrole-9-n-butylamine
(5c),
3-thiomorpholin-8-oxo-8H-acenaphtho[1,2-b] pyrrole-9-n-pentyl-
amine (5d) and 3-thiomorpholin-8-oxo-8H-acenaphtho[1,2-b]pyr-
role-9-n-hexylamine (5e), respectively (Scheme 1). Different with
parental 3, most of the 9-substituted derivatives have interruption
to FPAby the autofluorescence. Thus, we took enzyme-linked immu-
nosorbent assay (ELISA) to measure their abilities to competitively
displace a Bim-derived peptide from Mcl-1 and Bcl-2, respectively.
Triton X-100 was added as a detergent to prevent the possible aggre-
gation of hydrophobic compounds, as described in the supporting
information. The competitive binding curves of these compounds
to Bcl-2 and Mcl-1 were outlined in Figure 3a and b. R-(À)-Gossypol
was used as positive control.14 Compound 3 was tested for compar-
ison. While weaker binding affinity was found for 5a, better binding
affinitiesthan 3 towardMcl-1and Bcl-2werefoundfor 5b–5e. Apro-
gressive increase in length of the substituent (b < c < d < e) resulted
in a corresponding increase in Mcl-1 and Bcl-2 affinity. Compound
With the aim of accessing the p4 pocket to achieve enhanced
dual inhibition effects, we designed series of 3 derivatives that
the cyano group was replaced by longer and larger groups. Previous
studies have reported that Bcl-2 and Mcl-1 show differences in the
structure of their p4 pockets.15 When we tried to occupy the p4, an
optimization path should be carefully designed to resolve differ-
ences of the p4 between Bcl-2 and Mcl-1 without losing either one.
At the outset, we aimed to maintain the binding mode with
R263 in Mcl-1 and R146 in Bcl-2 constant since the key hydrogen
bound was formed in this region. We then sought to the X-ray
structure of the non-selective peptide Bim (BH3) in complex with

T. Song et al. / Bioorg. Med. Chem. 21 (2013) 11–20
13
Table 1
3,9 Substituted 3 derivatives: name, structure, HAC and binding affinity toward Mcl-1 by ELISA (IC50, nM)
O
R2
N
R1
Compounds
R1
R2
HAC
IC50 (nM)
pKd
D )
G (kcal molÀ1
3
CN
24
27
29
95
79
37
7.02
7.10
7.43
9.65
9.76
N
N
N
N
S
S
S
S
HN
5c
5e
HN
10.21
5h
6a
6b
6c
6d
7a
7b
7c
7d
32
26
29
31
34
27
30
32
35
15
90
7.82
7.04
7.30
7.53
8.00
6.41
6.82
7.11
7.38
10.75
9.68
HN
CN
S
S
S
S
S
S
S
S
NH2
NH2
HN
50
10.03
10.34
10.99
8.81
HN
HN
30
NH2
NH2
10
CN
388
150
78
OCH3
OCH3
OCH3
OCH3
HN
9.38
HN
HN
9.77
42
10.14
O
O
O
O
8a
8b
8c
8d
CN
28
31
33
36
321
141
75
6.49
6.85
7.10
7.41
8.92
9.41
HN
HN
HN
9.75
39
10.18
5e exhibited IC50 values of 37 and 96 nM for Mcl-1 and Bcl-2, respec-
tively, which was nearly 3- and 7-fold more effective than 3
(IC50 = 95 and 715 nM). We suspected that 5b have accessed p4
pocket and 5c, 5d, and 5e expanded inside the pocket.
Because 5b–5e were analogues with straight-chain substituent,
we tried to replace the linear group with bulkier group to probe the
surrounded areas between the arginine residue and p4. As such, we
synthesized 5f and 5g with isopropyl and benzyl substitution,
respectively. By ELISA, we found that 5f and 5g completely lost
studies. As shown in Figure 4a and b, benzyl substituent of 5g
was sterically repulsed by the R146 and Y108 residues of Bcl-2.
By comparison, Mcl-1 could, to a certain degree, accommodate this
substitution. Remarkably, though 5g could bind Mcl-1, its potency
was about sixfold lower than that of 5c (79 nM) and 5d (37 nM),
which have a similar substituent length with it, indicating that
bulkier group in this area contributes less to or adversely affect
binding affinity, especially for Bcl-2. On the basis of these data,
we identified n-propyl, n-butyl, n-pentyl and n-hexyl as the proper
groups to access the p4 because they could fit well within different
BH3-binding grooves of Bcl-2 and Mcl-1.
We also explored the effect of rigid group on the dual inhibition
by substituting the cyano with ester as the linker group. As ex-
pected, the conversion of amino group to a rigid ester negatively
impacted the potency. Ethoxycarbonyl and n-propoxycarbonyl
substitution yielded compounds 4a and 4b (Scheme 1) showed
dramatically decreased affinities than 5c and 5d. No binding affin-
binding affinity to Bcl-2 (>10
Bcl-2 protein. However, 5f retained a weak binding affinity to
Mcl-1 (2.5 M) and 5g showed sub-micromolar affinity
lM), indicative of steric clashes in
l
a
(521 nM). The different selectivity by these two compounds raised
a possibility that the surrounded area at this position of Bcl-2 is
different from that at Mcl-1. We calculated the width of this area
by AutoDock tools, and found that the Bcl-2 R146 and Y108 resi-
dues narrowed the area by 8 Å, compared to the corresponding
Mcl-1 R263 and H224 residues, which was 14.4 Å (Fig. S3). As such,
Bcl-2 could not accommodate 5f and 5g with bulkier group. As for
Mcl-1, it could accommodate the more conformationally re-
strained benzyl (5g) substitution but cannot accept the steric bulky
isopropyl (5f) substitution. It was further confirmed by docking
ities to Mcl-1 (>10
lM) and very weak affinities to Bcl-2 (2.2 and
2.1 M, respectively) were found for them. It is reasonable to attri-
l
bute the loss of affinity to the rigid C–O–C@O, which formed a
dihedral equal to 0° or 180°21 These results suggested a flexible lin-
ker are favorable for binding the two protein targets.

14
T. Song et al. / Bioorg. Med. Chem. 21 (2013) 11–20
Table 2
3,9 Substituted 3 derivatives: name, structure, HAC and binding affinity toward Bcl-2 by ELISA (IC50, nM)
O
R2
N
R1
Compounds
R1
R2
HAC
IC50 (nM)
pKd
D )
G (kcal molÀ1
3
CN
24
27
29
715
198
96
6.15
6.70
7.02
8.44
9.21
9.64
N
N
N
N
S
S
S
S
HN
5c
5e
HN
HN
5h
6a
6b
6c
6d
7a
7b
7c
7d
32
26
29
31
34
27
30
32
35
49
331
190
80
7.31
6.48
6.72
7.10
7.70
6.02
6.45
6.69
7.24
10.04
8.90
9.23
9.75
10.58
8.28
8.86
9.19
9.95
S
S
S
S
S
S
S
S
CN
NH2
NH2
HN
HN
HN
NH2
20
NH2
CN
945
356
206
57
OCH3
OCH3
OCH3
OCH3
HN
HN
HN
O
O
O
O
8a
8b
8c
8d
CN
28
31
33
36
739
322
184
42
6.12
6.49
6.74
7.38
8.42
8.92
H N
HN
9.25
HN
10.14
Having optimized the linker with flexible and linear character,
we now tried to utilize a benzene group to mimic the F69 of
non-selective peptide Bim for occupying p4 pocket. Since 5b, 5c,
5d and 5e have accessed p4 pocket, we added benzene to the
terminus of the alkyl chain of these compounds, yielding 3-thi-
omorpholin-8-oxo-8H-acenaphtho[1,2-b]pyrrole-9-(3-phenyl)pro-
pylamine (5h), 3-thiomorpholin-8-oxo-8H-acenaphtho[1,2-b] pyr
role-9-(3-phenyl)butylamine (5i), 3-thiomorpholin-8-oxo-8H-
acenaphtho[1,2-b]pyrrole-9-(3-phenyl)pentylamine (5j) and 3-thi-
omorpholin-8-oxo-8H-acenaphtho[1,2-b]pyrrole-9-(3-phenyl) hex
ylamine (5k). In ELISA assay, 5h exhibited the most potent binding
affinity to Bcl-2 (49 nM) and Mcl-1(15 nM), which achieved nearly
12- and 6-fold better affinities toward Bcl-2 and Mcl-1, respec-
tively than 5b. Compound 5i also showed about 2- to 3-fold im-
proved affinities than 5c. However, 5j exhibited a significant
In our previous SAR study, aminophenylthio group at the
3-position was identified to fit p2 better than thiomorpholin.
Methoxyphenylthio and isopropylphenoxy were groups that could
fit p2 but not as well as thiomorpholin.12 In order to discover more
potent inhibitors by spanning p2 and p4, and gain insights into the
druggability of the p2 and p4 pockets, we selected 3-(4-aminophe-
nylthio)-8-oxo-8H-acenaphtho[1,2-b]pyrrole-9-carbonitrile (6a),
3-(4-methoxyphenylthio)-8-oxo-8H-acenaphtho[1,2-b]pyrrole-9-
carbonitrile (7a) and 3-(4-isopropylphenoxy)-8-oxo-8H-acenaph-
tho[1,2-b]pyrrole-9-carbonitrile (8a) as starting compounds for
further optimization at 9 position. Series 6 (6a–6d), 7 (7a–7d),
and 8 (8a–8d) were synthesized (Table 1). We found that butyl,
hexyl and 3-phenylpropyl substitution brought about a corre-
sponding increase of binding affinity to Bcl-2 and Mcl-1 in the
three series (Tables 1 and 2), which was in agreement with the
results of 5c, 5e and 5h. Comparably, the rank order of binding
affinities for certain 9-substituted analogues from series 5, 6, 7
and 8 was keeping the same as the rank order of p2-occupying effi-
ciency, which is 6a > 3 > 8a > 7a. We also tested the binding affin-
ities of 3 derivatives toward Bcl-xL and gained the same trend of
improvement (Table S1). Among them, the most potent 6d
displayed IC50 values of 10, 20 and 18 nM for Mcl-1, Bcl-2 and
Bcl-xL, respectively, which is 9–35 times more effective than 3.
decrease of binding affinities to Bcl-2 (1.3
lM) and Mcl-
1(1.1 M) compared to 5d. Compound 5k showed a similar loss
l
of affinity. In agreement with it, docking results of 5h and 5j
showed that 3-phenylpropyl was properly located in the p4 pocket
of Mcl-1, while the 3-phenylpentyl of 5j was repulsed out by the p4
pocket (Fig. 4c and d). Therefore, an alkyl group with three carbons
was the optimal length to mimic the northern part of D67 to F69 of
Bim peptide.

T. Song et al. / Bioorg. Med. Chem. 21 (2013) 11–20
15
0.10
0.10
0.08
0.06
0.04
0.02
0.00
b
R215
a
V216*
G219
V220
R263*
F270
0.08
0.06
0.04
0.02
0.00
M250
V253
R263*
F270*
M250
V253*
K234
L246
K234
L246
180
200
220
240
260
280
300
320
180
200
220
240
260
280
300
320
d
c
R263
V253
M250
R263
V253
R215
V216
M2p502, p3
V220 p4
F270
p2, p3
G219
p4
M231
K234
K234
Figure 1. Amide proton and nitrogen CCSP of Mcl-1 derived from the 1H–15N HSQC spectra after titration with (a) 3 to Mcl-1 and (b) 6b to Mcl-1. Resonances that disappeared
upon binding are shown as gray bars. That was V253, R263 and F270 for 3, and V216 and R263 for 6b. (c, d) Chemical shift perturbations of residues induced by 3 and 6b
mapped onto the structure of Mcl-1. The Mcl-1 residues showing chemical shift changes of 0.03 <
which the cross-peaks disappeared upon binding were colored red.
D
CS < 0.06 were colored yellow, and residues showing with DCS >0.06 or for
b
a
H224
Y108
p2
p4
p2
p4
R146
R263
Figure 2. Surface depiction of the three-dimensional structures of (a) Mcl-1 (PDB: 2PQK) and (b) Bcl-2 (PDB: 1GJH). The p2 and p4 interacting residues L62 and F69 in Bim
complex with Mcl-1 were shown in pink, respectively. Different angles of the BH3 groove of two proteins in the p4 region were shown with arrows.
To confirm the results of the ELISA assay data, we also performed
isothermal titration calorimetry (ITC) assay with 6d and 3 in paral-
lel. In consistent with the improvement in IC50 values, 6d exhibited
a Kd value of 105 nM, which was about ninefold improved than that
of 3 (Kd = 964 nM, Fig. 5).
site-directed mutant on R263A and evaluated the binding affinities
of some representative analogues toward wild-type Mcl-1 and
Mcl-1 R263A mutant, respectively by using ITC. 6d, for example,
showed a 20-fold lower affinity toward R263A mutant (2.34 lM,
Table 3 and Fig. S4) than wild-type Mcl-1. The similar decrease
was found for other compounds (Table 3). It supported that R263
still served as an anchor residue by forming hydrogen bond net-
work with 3, 9 substituted 3 derivatives. Together with the same
rank order of p2-occupying efficiency, we proposed that these
derivatives gained increased affinity by additional occupation of
p4 pocket when maintained the p2 occupation.
2.3. Identification of binding site by site-directed mutagenesis
and NMR
R263 of Mcl-1 and the R146 of Bcl-2 have been determined to
form a hydrogen bond with carbonyl of 3. To examine whether
R263 still played an important role in the binding of newly
To further determine the binding site, we produced uniformly
designed 3,
9 substituted 3 derivatives, we constructed a
15N-labeled Mcl-1 protein and performed NMR titration by 6b,

16
T. Song et al. / Bioorg. Med. Chem. 21 (2013) 11–20
5a, R=CH2CH3
5b, R=CH2CH2CH3
5c, R=CH2(CH2)2CH3
5d, R=CH2(CH2)3CH3
5e, R=CH2(CH2)4CH3
5f, R=CH(CH3)2
O
O
O
O
H
N
R
N
CN
N
O
N
R
4a, R=CH2CH3
4b, R=CH2CH2CH3
5g, R=CH2Ph
N
N
S
N
S
5h, R=CH2(CH2)2Ph
5i, R=CH2(CH2)3Ph
5j, R=CH2(CH2)4Ph
5k, R=CH2(CH2)5Ph
S
4a-b
3 (S1)
5a-k
Scheme 1. Structures of 3, 4a–b, and 5a–k.
110
110
100
90
80
70
60
50
40
30
20
10
0
a
b
100
Gossypol 0.450 0.15
0.715 0.17
Gossypol 0.212 0.03
3
90
3 0.095 0.06
4a 2.234 0.21
4b 2.054 0.26
5a 1.412 0.13
5b 0.565 0.11
5c 0.198 0.04
5d 0.124 0.07
5e 0.096 0.02
5f >10
4a (>10)
80
4b (>10)
70
5a 1.132 0.13
5b 0.087 0.02
5c 0.079 0.01
5d 0.057 0.07
5e 0.037 0.04
5f 2.563 0.25
5g 0.521 0.11
5h 0.015 0.05
5i 0.034 0.03
5j 1.105 0.17
5k 2.107 0.15
60
50
40
30
20
10
0
5g >10
5h 0.049 0.03
5i 0.057 0.01
5j 1.347 0.07
5k 2.374 0.16
-10
-10
-2
0
2
4
-2
0
2
4
Inhibitor Log μM
Inhibitor Log μM
Figure 3. Competitive binding curves of (À)-Gossypol, 3, 4a–b, 5a–k to (a) Mcl-1 and (b) Bcl-2 as determined by ELISA assay.
a
b
d
H224
Y108
R146
R263
c
p4
p4
R263
R263
Figure 4. Predicted binding model of 5g in complex with (a) Bcl-2, (b) Mcl-1, (c) 5h in complex with Mcl-1, and (d) 5j with Mcl-1. The carbon, oxygen, nitrogen atoms were
shown in gray, red and blue. The sulfur atom was shown in yellow and green for Bcl-2 and Mcl-1, respectively.
because 6b is more soluble than 6d in Mcl-1 solution. We
and presence of 6b, respectively. Figure S2c showed an overlay of
the spectra before (blue) and after (red) the addition of 6b,
measured two-dimensional [15N, 1H] NMR spectra in the absence

T. Song et al. / Bioorg. Med. Chem. 21 (2013) 11–20
17
3.40
3.20
3.00
2.80
2.60
a
3.60
b
3.40
3.20
3.00
2.80
2.60
2.40
2.40
2.20
0.00
0.00
-1.00
-1.00
-2.00
-3.00
-2.00
-3.00
-4.00
-5.00
-6.00
-7.00
-4.00
-5.00
-6.00
Mcl-1+ 3
Mcl-1+ 6d
N (stoichimetry)
0.913 0.15
964.2 2.83
-8.27
N (stoichimetry)
0.971 0.08
105.3 1.47
-9.59
Kd(nM)
Kd(M)
Δ G (Kcal/mol)
ΔH (Kcal/mol)
-TΔS (Kcal/mol)
Δ G (Kcal/mol)
ΔH (Kcal/mol)
-TΔS (Kcal/mol)
-6.42 0.65
-1.85
-5.78 1.03
-3.81
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
0.0
0.2
0.4
0.6
0.8 1.0
1.4
1.6
1.8
2.0
Molar Rati1o.2
Molar Ratio
Figure 5. (a) Raw data for titration of 3 and (b) 6d into Mcl-1. Titrations consisted of 12 Â 3
to one injection. Kd value was calculated from integration of the raw data.
l
L injections of compound at 300 lM into Mcl-1 (30 lM). Each peak corresponds
together with close-up views of selected residues that underwent
large chemical shift changes in the presence of 6b (Fig. S2d). Fig-
ure 1b showed a plot of the chemical shift perturbations against
the overall Mcl-1 protein residues. Above 60% residues affected
by the binding of 6b were located in the BH3-binding pocket of
Mcl-1, indicating that 6b binds in the BH3-binding groove of
Mcl-1. When compared with NMR spectra for 3 (Fig. 1a), a same
cluster of residues in p2 including R263, V253, L246, M250, F270
and K234 experienced average chemical shift changes of at least
0.05 ppm (Fig. 1b). Notably, another cluster of residues including
V220, R215, V216 and G219 which experienced at changes of at
least 0.06 ppm emerged after adding 6b. In particular, R263 and
V216 are no longer detectable. Mapping these residues into 3D
structure of Mcl-1 showed their locations within and surrounding
the p2 and p4 pocket (Fig. 1d). So far, we identified that 6b occu-
pied p2 and p4 simultaneously.
p4-occuping groups per atom by plotting the potency against the
HAC in serials 5, 6, 7, and 8. When a certain 3-position substituent
was chosen to occupy p2, groups with an increased size were
substituted at 9-position to occupy p4. As shown in Figure 6a for
Mcl-1, a linear relationship between potency and HAC was found
for compounds 3, 5c, 5e and 5h. When plotted with the other seri-
als, nearly parallel trend lines were found, with an average value
for the slop of 0.16 kcal molÀ1 per heavy atom. These data implied
that the parental core remained unchanged and that the 9-position
substituent was simply added the p4 occupation. Given the con-
stancy of slop in different serials, the same binding efficiency of
0.16 should be directly related to the contribution of p4 in Mcl-1.
As for Bcl-2, co-linear relationships between potency and HAC
were also found in the four serials. However, the average value
for the slop was approximately 0.21 kcal molÀ1 per heavy atom.
It suggested that additional groups to occupy p4 contributed less
affinity to Mcl-1 than Bcl-2. From the aspect of p4 pocket, the bind-
ing site of Bcl-2 might be more druggable than that of Mcl-1.
2.4. Comparing the different contribution of p4 to binding
affinity between Bcl-2 and Mcl-1
2.5. Improved solubility of 3, 9 substituted 3 derivatives
The alanine scanning studies have shown when F69 of Bim was
mutated to alanine, it lost most of the affinity to Bcl-xL. However
the mutant can still bind Mcl-1 in the same range even it lost p4
occupation.22 Additionally, a broader p4 pocket in Mcl-1 than that
in Bcl-2 has been revealed by a molecular dynamics study.15 These
studies suggested that the p4 pocket has less contribution to Mcl-1
binding. In the present study, however, we found binding affinities
to Mcl-1 were progressively improved accompanied with in-
creased p4 occupation. We inferred that p4 occupying was still a
dominant contributor for a molecule to bind to Mcl-1. Notably,
for an optimized compound, it was improved less dramatically
for Mcl-1 (average eightfold) than for Bcl-2 (average 17-fold), such
as compound 5h, 6d, 7d, and 8d. It suggested the contribution of
p4 was different between the two proteins.
Next, we tested the solubility of 3, 6b and 6d in 1% DMSO/H2O
which has been accepted to be orally and intravenously available
for drug treatment in vivo.23 As shown in Tables 4 and 3 was prac-
tically insoluble (0.002 mg/mL) in this solvent, while the solubility
of 6b and 6d could reach 0.027 and 0.018 mg/mL, respectively. The
increase of solubility showed the improved drug-like properties
compared to 3.
2.6.3,9 Substituted 3 derivatives induce apoptosis by binding to
the BH3 groove of Bcl-2 and Mcl-1
We next determined whether the improved binding affinities of
6b and 6d toward both Bcl-2 and Mcl-1 could translate into greater
disruption of Bcl-2/Bax and Mcl-1/Bak complexes in living cells.
We applied multiple cell lines with different levels of Mcl-1 and
To further illustrate the contribution of p4 occupation to Bcl-2
and Mcl-1, respectively, we calculated the contribution of

18
T. Song et al. / Bioorg. Med. Chem. 21 (2013) 11–20
Table 3
Binding affinity of compounds with Mcl-1 and R263A mutant by ITC Assay (Kd,
l
M)
O
R2
N
R1
Compound
R-(À)-Gossypol
3
R1
—
N
R2
—
Mcl-1 (wild-type)
2.324
Mcl-1 (R263A)
—
CN
0.954
28.73
S
HN
6b
6d
S
S
0.512
0.105
12.75
2.34
NH2
NH2
HN
HN
O
8c
0.802
19.15
12
Series5 Y=0.201X+3.708
Table 4
Solubility of 3, 6b and 6d in 1% DMSO/H2O (mg/mL)
Series6 Y=0.213X+3.227
Series7 Y=0.206X+2.679
Series8 Y=0.212X+2.397
Bcl-2
11
10
9
O
R2
N
8
7
R1
6
20
22
24
26
28
30
32
34
36
38
40
Compound
R1
R2
1%DMSO/H2O (mg/mL)
HAC(heavyatomcount)
12
11
10
9
3
CN
0.002
0.027
N
S
S
Series5 Y=0.154X+5.744
HN
Series6 Y=0.163X+5.363
Series7 Y=0.168X+4.319
Series8 Y=0.158X+4.501
Mcl-1
6b
NH2
NH2
S
6d
0.018
HN
the compounds released cytochorme c to cytosome in the same or-
der as the complexes disruption (Fig. 7b and c). These data demon-
strated that 6b and 6d exhibited enhanced apoptosis induction
than 3 through disrupting complexes of Bcl-2 and Mcl-1.
8
Compound 3 has been identified as a pure BH3 mimetic that
kills via Bax/Bak completely. Most of the recently developed nano-
molar inhibitors (six of seven) such as Gossypol (BL-193, AT-101)
are not pure BH3 mimetics and then showed unexpected toxic-
ity.8,24 Next, we determined whether or not 6b and 6d retained
the property of 3 as a pure BH3 mimetic. Bax and Bak was both si-
lenced in MCF-7 cells by shRNA (Fig. S5). Gossypol was used as a
7
18
20
22
24
26
28
30
32
34
36
38
40
HAC(heavyatomcount)
Figure 6. Free energy of binding (in kcal molÀ1) for compounds of series 5, 6, 7 and
8 plotted as a function of the number of heavy atoms in the ligands. See Table 1.
negative control. After 12 h of exposure to 5 lM of compounds,
Bcl-2, that is, human chronic myelogenous leukemia cell line K562,
myeloid leukemia cell line HL-60 and breast cancer cell line MCF-7.
Among them, K562 and HL-60 express the highest level of Mcl-1
and Bcl-2, respectively (Fig. 7a). By co-immunoprecipitation
(Co-IP) assays, we found that 6b and 6d exhibited greater potential
than 3 to disrupt Bcl-2/Bax and Mcl-1/Bak complexes. Meanwhile,
cytochrome c release was examined. In contrast with cytochrome
c release in wild-type MCF-7 cells, no cytochrome c release was
found for 3, 6b and 6d in Bax- and Bak-deleted cells. However,
the same amount of cytochrome c release as that in wild-type cells
was detected for Gossypol (Fig. 7d and e). This strongly supported
that 6b and 6d killed completely dependent on Bax/Bak. The pure

T. Song et al. / Bioorg. Med. Chem. 21 (2013) 11–20
19
a
K562
MCF-7 HL-60
Mcl-1
Bcl-2
b
K562
6b
c
HL-60
DMSO
3
6d
DMSO
Bax
3
6b
6d
Bak
IP: Mcl-1
IP: Bcl-2
Mcl-1
Cyto C
Bcl-2
Cytosolic
fraction
Cytosolic
fraction
Cyto C
MCF-7 / Bax/Bak shRNA
DMSO Gossypol 6b 6d
e
MCF-7
DMSO Gossypol
3
6b
6d
3
Cytosolic
fraction
Cyto C
Cyto C
Cyto C
Cyto C
Membrane
fraction
Figure 7. (a) The levels of Mcl-1 and Bcl-2 protein in K562, MCF-7 and HL-60 cells were examined by Western blot. (b) K562 and (c) HL-60 cells were treated with 5
l
M 3, 6b,
6c or 6d for 12 h. Harvested cells were subjected to immunoprecipitation of Mcl-1 and Bcl-2 followed by western blot analysis for Bak and Bax, respectively. The paralleled
harvested cells were lysed in digitonin lysis buffer. After centrifugation, the supernatant (S-100, cytosolic fraction) were collected and subjected to western blot for
cytochrome c. (d) Wild-type MCF-7 cells and (e) Bax/Bak shRNA transfected cells were treated with 5
detection in S-100 fraction.
lM indicated compounds, respectively, followed by cytochrome c
BH3 mimicking property of these derivatives could render them an
attractive aspect in the future clinical development.
J = 8.0 Hz, 1H), 7.88 (t, J = 8.0 Hz, 1H), 7.63 (d, J = 8.0 Hz, 2H), 7.56
(d, J = 8.0 Hz, 1H), 7.38 (d, J = 8.0 Hz, 2H), 3.89 (m, 2H,
–
NHCH2CH2–), 3.45 (br, 2H), 1.75 (m, 2H, –NH CH2CH2CH2–), 1.41
(m, 2H, –CH2CH2CH2CH3), 0.94 (t, J = 8.0 Hz, 3H). TOF MS (EI+):
C24H21N3OS, (m/z): calcd for 399.0401, found 399.0405. HPLC sys-
tem 1: purity = 99.03%, tR = 15.7 5 min. HPLC system 2: pur-
ity = 99.29%, tR = 5.16 min.
3. Conclusions
We substituted the 9-position of 3 to probe the different p4
pockets in Bcl-2 and Mcl-1. [15N, 1H] NMR, SAR studies and site-di-
rected mutation determined that when the key interaction with
R263 in Mcl-1 and its homologue in Bcl-2 was maintained con-
stant, flexible amino liker made derivatives accepted by the differ-
ent angle of the BH3 groove of two proteins in the p4 region.
Additionally, a broader space between R263 and p4 of Mcl-1 than
that of Bcl-2 was identified by small molecules. Ultimately, the
most potent compound 6d that contained a three-carbon alkyl
group to optimally mimic the northern part of D67 to F69 of Bim
peptide and utilized the terminal benzene to mimic F69, exhibited
9- to 35-fold improvement than 3 with IC50 values of 10, 20 and
18 nM toward Mcl-1, Bcl-2 and Bcl-xL, respectively. Excitingly, its
improved binding affinity also translated to greater potency of dis-
rupting Bcl-2/Bax and Mcl-1/Bak complex in cells, resulting in en-
hanced apoptosis induction in multiple cancer cell lines with
variant Mcl-1 and Bcl-2 expression levels. Together with pure
BH3 mimicking property and improved solubility than 3, 6b and
6d represent novel potent lead drugs for cancer therapy.
Compound 6d. Yield: 19%. 1H NMR (400 M, CDCl3): d 8.57 (d,
J = 8.0 Hz, 1H), 8.47 (d, J = 8.0 Hz, 1H), 8.44 (d, J = 8.0 Hz, 1H),
7.88 (t, J = 8.0 Hz, 1H), 7.63 (m, 4H), 7.56 (d, J = 8.0 Hz, 1H), 7.47
(m, 3H), 7.38 (d, J = 8.0 Hz, 2H), d 6.74 (br, 1H),3.89 (m, 2H,–
NHCH2CH2–), 3.55 (br, 2H), 1.75 (m, –NH CH2CH2CH2–, 2H), 1.41
(m, 2H, –CH2CH2CH2 Ph). TOF MS (EI+): C29H23N3OS, (m/z): calcd
for 461.0588, found 461.0585. HPLC system 1: purity = 98.73%,
tR = 17.56 min. HPLC system 2: purity = 99.29%, tR = 4.06 min.
4.2. ELISA assay
An enzyme-linked immunosorbent assay was reported as previ-
ous. For this assay, BiotinylatedBim peptide (residues 81-106,
biotin-(b)A-(b)A-D-M-R-P-E-I-W–I-A-Q–E-L-R-R-I-G-D-E-F-N-A-Y-
Y-A-R-R-amide, hereafter called biotin-Bim) was diluted to
0.09 lg/mL in SuperBlock blocking buffer in PBS (Pierce Biotech-
nology, Inc, Rockford, IL, catalog # 37515) and incubated for 1.5 h
in 96-well microtiter plates already coated with streptavidin
(Qiagen, Catalog # 15500) to allow the formation of the complex
between Biotin-Bim and streptavidin. All incubations were per-
formed at room temperature unless otherwise noted. Each inhibi-
tor was first dissolved in pure DMSO to obtain a 10 mM stock
solution. Next, serial dilutions of compounds with 0.01% Triton
X-100 were incubated with 20 nM protein in PBS for 1 h. The plates
were washed three times with PBS containing 0.05% Tween-20.
In addition, we showed co-linear relationships of potency
against the HAC by 9-position substituted derivatives with the slop
of 0.16 and 0.21 kcal molÀ1 per heavy atom for Mcl-1 and Bcl-2,
respectively. It suggested that p4 occupying could contribute to
dual Bcl-2 and Mcl-1 inhibitor, but the contribution was less for
Mcl-1 than Bcl-2.
4. Experimental
4.1. Chemistry
The inhibitor and protein mixture (100 lL) were transferred to
the plate containing the biotin-Bim/streptavidin complex and
incubated for 2 h. The plate was then washed as before and mouse
anti-His antibody that conjugated with horseradish peroxidase
(Qiagen, Catalog # 34460) was added into the wells and incubated
for 1 h. The plate was then washed with PBS containing 0.05%
The synthesis of 6b and 6d is shown in Supplementary data
Scheme 3. 6b. Yield: 23%. 1H NMR (400 M, CDCl3): d 8.74 (br,
1H), 8.57 (d, J = 8.0 Hz, 1H), 8.47 (d, J = 8.0 Hz, 1H), 8.44 (d,

20
T. Song et al. / Bioorg. Med. Chem. 21 (2013) 11–20
Tween-20. Finally, TMB (100 lL, Beyotime, Catalog # P0209) was
Supplementary data
added to each well; the enzymatic reaction was stopped after
30 min by addition of H2SO4 (100 L, 2 M). Absorbances were mea-
l
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmc.2012.11.008.
sured with a TECAN GENios (Swiss, TECAN) microplate reader
using a wavelength of 450 nm. Three independent experiments
were performed with each inhibitor to calculate average IC50 value
and standard deviation (SD).
References and notes
1. Sackett, D. L.; Sept, D. Nat. Chem. 2009, 1, 596.
2. Veselovsky, A. V.; Archakov, A. I. Curr. Comput. Aided Drug Des. 2007, 3, 51.
3. Azmi, A. S.; Mohammad, R. M. J. Cell. Physiol. 2009, 218, 13.
4. Zhang, L.; Ming, L.; Yu, J. Drug Resist. Updates 2007, 10, 207.
5. Vogler, M.; Dinsdale, D.; Dyer, M. J.; Cohen, G. M. Cell Death Differ. 2009, 16, 360.
6. Warr, M. R.; Shore, G. C. Curr. Mol. Med. 2008, 8, 138.
7. Bednarek, J.; Wesierska, J. D.; Kilianska, Z. M. Postepy Biochem. 2007, 53, 228.
8. Delft, M. F. V.; Wei, A. H.; Mason, K. D.; Vandenberg, C. J.; Chen, L.; Czabotar, P.
E.; Willis, S. N.; Scott, C. L.; Day, C. L.; Cory, S.; Adams, J. M.; Roberts, A. W.;
Huang, D. C. Cancer Cell 2006, 10, 389.
9. Konopleva, M.; Contractor, R.; Tsao, T.; Samudio, I.; Ruvolo, P. P.; Kitada, S.;
Deng, X.; Zhai, D.; Shi, Y. X.; Sneed, T., et al Cancer Cell 2006, 10, 375.
10. Zhang, Z. C.; Jin, L. J.; Qian, X. H.; Wei, M. J.; Wang, Y. Y.; Wang, J. ChemBioChem
2007, 8, 113.
11. Zhang, Z. C.; Song, T.; Zhang, T. T.; Gao, J.; Wu, G. Y.; An, L. J. Int. J. Cancer 2010,
128, 1724.
12. Zhang, Z. C.; Wu, G. Y.; Xie, F. B.; Song, T.; Chang, X. L. J. Med. Chem. 2011, 54,
1101.
13. Lee, E. F.; Czabotar, P. E.; Smith, B. J.; Deshayes, K.; Zobel, K.; Colman, P. M.;
Fairlie, W. D. Cell Death Differ. 2007, 14, 1711.
4.3. ITC assay
Isothermal titration calorimetry (ITC) was performed using
iTC200 (Microcal). Experiments were performed in 20 mM Tris
pH 8.0, 150 mM NaCl, 1% DMSO at 25 °C. For evaluating Kd value
of 6d compared with that of 4, titrations consisted of 12 Â 3
injections of compound at 300 M into Mcl-1 (30 M). For evaluat-
ing Kd values of analogues toward wild-type Mcl-1 compared with
that toward Mcl-1 R263A, titrations consisted of 16 Â 2.5 L injec-
tions of compound at 500 M into Mcl-1 (50 M). All sample data
obtained after control data corrections were analyzed to fit to a
one-site model. For control ITC experiments, the sample cells were
filled with assay buffer and the compound solution was added. This
process was identical to that for protein samples.
lL
l
l
l
l
l
14. Wang, G.; Nikolovska-Coleska, Z.; Yang, C. Y.; Wang, R.; Tang, G.; Guo, J.;
Shanqary, S.; Qiu, S.; Gao, W.; Yang, D., et al J. Med. Chem. 2006, 49, 6139.
15. Acoca, S.; Cui, Q.; Shore, G. C.; Purisima, E. O. Proteins 2011, 79, 2624.
16. Czabotar, P. E.; Lee, E. F.; Van Delft, M. F.; Day, C.; Smith, B. J.; Huang, D. C. S.
Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 6217.
17. Bruncko, M.; Oost, T. K.; Belli, B. A.; Ding, H.; Joseph, M. K.; Kunzer, A.;
Martineau, D.; McClellan, W. J.; Mitten, M.; Nq, S. C., et al J. Med. Chem. 2007,
50, 641.
18. Tang, G.; Nikolovska-Coleska, Z.; Qiu, S.; Yang, C. Y.; Guo, J.; Wang, S. J. Med.
Chem. 2008, 51, 717.
19. Fire, E.; Gulla, S. V.; Grant, R. A.; Keating, A. E. Protein Sci. 2010, 19, 507.
20. Ballard, P.; Bradbury, R. H.; Harris, C. S.; Hennequin, L. F.; Hickinson, M.; Kettle,
J. G.; Kendrew, J.; Klinowska, T.; Ogilvie, D. J.; Pearson, S. E.; Williams, E. J.;
Wilson, I. Bioorg. Med. Chem. Lett. 2006, 16, 4908.
4.4. NMR spectroscopy
[
15N, 1H] HSQC spectra were recorded for 0.1 mM 15N-labeled
Mcl-1. Compound 3 and 6b stock solution was added at a 4:1 ratio,
respectively. The 1H/15N HSQC spectra were recorded on a Bruker
Advance DRX600 MHz spectrometer, processed by NMRPipe and
visualized by NmrViewJ 8.0. Chemical shift changes were calcu-
lated as ((
residue number.
D
1H ppm)2 + (0.2 Â
D
15N ppm)2)0.5) as a function of the
21. Lopes, S.; Lapinski, L.; Fausto, R. Phys. Chem. Chem. Phys. 2002, 4, 5952.
22. Lee, E. F.; Czabotar, P. E.; Delft, M. F.; Michalak, E. M.; Boyle, M. J.; Willis, S. N.;
Puthalakath, H.; Bouillet, P.; Colman, P. M.; Huang, D. C.; Fairlie, W. D. J. Cell
Biol. 2008, 180, 341.
23. Mui, E. J.; Schiehser, G. A.; Milhous, W. K.; Hsu, H.; Roberts, C. W.; Kirisits, M.;
Muench, S.; Rice, D.; Dubey, J. P.; Fowble, J. W., et al PLoS Negl. Trop. Dis. 2008, 2,
e190.
Acknowledgments
The authors would like to thank Professor Zhongjun Li (State
Key Laboratory of Natural and Biomimetic Drugs, Peking Univer-
sity, Beijing, People’s Republic of China) for the discussion. The
work was partly supported by the Fundamental Research Funds
for the Central Universities (DUT11SM08) and partly supported
by the National Natural Science Foundation of China (81273436).
24. Lei, X.; Chen, Y.; Du, G.; Yu, W.; Wang, X.; Qu, H.; Xia, B.; He, H.; Mao, J.; Zong,
W., et al FASEB J. 2006, 20, 2147.