Structure-based design of benzo[e]isoindole-1,3-dione derivatives as selective GSK-3β inhibitors to activate Wnt/β-catenin pathway

Article abstract of DOI:10.1016/j.bioorg.2015.05.009

Deregulation of Wnt/β-catenin pathway is closely related to the pathogenesis of neurodegenerative diseases such as Alzheimer's disease (AD), and glycogen synthase kinase 3β (GSK-3β), the central negative regulator of Wnt pathway, is regarded as an important target for these diseases. Here, we report a series of benzo[e]isoindole-1,3-dione derivatives as selective GSK-3β inhibitors by rational-design and synthesis, which show high selectivity against GSK-3β over Cyclin-dependent kinase 2 (CDK2), and significantly activate the cellular Wnt/β-catenin pathway. The structure-activity relationship of these GSK-3β inhibitors was also explored by in silico molecular docking.

Full text of DOI:10.1016/j.bioorg.2015.05.009

Bioorganic Chemistry 61 (2015) 21–27
Contents lists available at ScienceDirect
Bioorganic Chemistry
journal homepage: www.elsevier.com/locate/bioorg
Structure-based design of benzo[e]isoindole-1,3-dione derivatives
as selective GSK-3b inhibitors to activate Wnt/b-catenin pathway
⇑
Hong Yue, Feng Lu, Chen Shen, Jun-Min Quan
Key Laboratory of Structural Biology, School of Chemical Biology & Biotechnology, Peking University, Shenzhen Graduate School, Shenzhen 518055, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 5 April 2015
Available online 29 May 2015
Deregulation of Wnt/b-catenin pathway is closely related to the pathogenesis of neurodegenerative dis-
eases such as Alzheimer’s disease (AD), and glycogen synthase kinase 3b (GSK-3b), the central negative
regulator of Wnt pathway, is regarded as an important target for these diseases. Here, we report a series
of benzo[e]isoindole-1,3-dione derivatives as selective GSK-3b inhibitors by rational-design and synthe-
sis, which show high selectivity against GSK-3b over Cyclin-dependent kinase 2 (CDK2), and significantly
activate the cellular Wnt/b-catenin pathway. The structure–activity relationship of these GSK-3b inhibi-
tors was also explored by in silico molecular docking.
Keywords:
GSK-3b
Inhibitor
Wnt/b-catenin pathway
Neurodegenerative diseases
CDK2
Ó 2015 Elsevier Inc. All rights reserved.
1. Introduction
[32]. A series of synthetically more feasible derivatives (Fig. 1, com-
pounds 1a and 1b) are also developed to inhibit GSK-3b and acti-
Wnt/b-catenin signaling pathway plays essential roles in
embryonic development and tissue homeostasis [1–5].
Deregulation of Wnt/b-catenin pathway is closely associated with
neurodegenerative diseases such as Alzheimer’s disease [6–8],
Parkinson disease [9,10] and Schizophrenia [11,12]. Glycogen syn-
thase kinase 3b (GSK-3b) acts as the key negative regulator of
Wnt/b-catenin signaling pathway, and is therefore regarded as an
important target for these diseases [13–18].
vate Wnt pathway, but the selectivity of these inhibitors over
CDK2 is relatively low [33], which is similar to that of compound
1. Kinase selectivity profiling has shown that compound 1 is highly
selective against GSK-3b over other kinases except for CDK2 [32].
In this study, we optimized these inhibitors to be highly selective
to GSK-3b over CDK2 by structure-based rational design.
GSK-3b is
a constitutively active serine/threonine protein
2. Results and discussion
kinase belonging to the CMGC protein kinase family that include
cyclin-dependent kinases (CDKs), mitogen-activated protein
kinases (MAP kinases), glycogen synthase kinases (GSK-3) and
CDK-like kinases. The structure of GSK-3b is highly homologous
to that of other CMGC members such as cyclin-dependent kinases
(CDKs) [19–21]. In consistent with this homology, many GSK-3b
inhibitors are also potent inhibitors of CDKs [22–25]. Selectivity
is thus considered as a major issue to develop potent GSK-3b inhi-
bitors as probes for biological functions of GSK-3b and drugs for
the treatment of various diseases [18,26–31].
2.1. Rational design of selective GSK-3b inhibitors
The homology between the ATP sites of GSK-3b and CDK2 is
over 86%, so many GSK-3b inhibitors also potently inhibit CDK2.
For example, compound 1b inhibits GSK-3b and CDK2 with a com-
parable potency. To improve the selectivity, we explored the
regions outside the ATP site based on the crystal structures of
GSK-3b and CDK2, and searched for the potential residues that
can be used to design selective GSK-3b inhibitors. Detailed struc-
tural analysis revealed that the residues located at the helix D
immediately following the hinge region are significantly different
between GSK-3b and CDK2 (Fig. 2). GSK-3b is characterized by pos-
itively charged residues (Arg141 and Arg144) in this region, while
CDK2 is characterized by negatively charged residues (Asp86 and
Asp92). The difference of the electrostatic potential in this region
makes it possible to design selective inhibitors against GSK-3b
but not CDK2. We used compound 1b as the starting point, and
introduced sulfonyl group to interact with the positively charged
During our development of a chemical genetic approach to ana-
lyzing biological systems by using zebrafish assay, we identified a
potent GSK-3b inhibitor (Fig. 1, compound 1) bearing
a
benzo[e]isoindole-1,3-dione core structure from a public library,
which inhibits the eye and forebrain formations of zebrafish
embryos, resembling a typical Wnt overexpression phenotype
⇑
Corresponding author.
E-mail address: quanjm@pkusz.edu.cn (J.-M. Quan).
http://dx.doi.org/10.1016/j.bioorg.2015.05.009
0045-2068/Ó 2015 Elsevier Inc. All rights reserved.

22
H. Yue et al. / Bioorganic Chemistry 61 (2015) 21–27
residues Arg141 and Arg144 at the helix D of GSK-3b. The sulfonyl
group could be linked to the methyl group at the 5-position of the
benzo[e]isoindole-1,3-dione scaffold by 4-amino piperidine group
with the consideration of size and solubility of the linker. Such
modification of compound 1b would improve the binding affinity
with GSK-3b by the favorable electrostatic interactions, while
which might impair the binding affinity with CDK2 due to the
repulsion between the partially negatively charged sulfonyl group
and the negatively charged residues of CDK2.
2.2. Chemistry
The desired benzo[e]isoindole-1,3-dione derivatives 8a–i were
synthesized in modest to good yields as shown in Scheme 1. The
sulfonyl piperidine amine substrates (5a–i) were synthesized by
the condensation of 4-Boc-aminopiperidine and sulfonyl chloride
followed by removing the Boc protecting group. The precursor 6
that has been reported previously [33] served as the starting
material. After bromination with NBS and AIBN, the compound 7
Fig. 1. Chemical structures of GSK-3b inhibitors (1, 1a and 1b) with
benzo[e]isoindole-1,3-dione core structure.
Fig. 2. Designed strategy of selective GSK-3b inhibitors. (a) Structural comparison between GSK-3b (green) and CDK2 (cyan), different residues at the helix D were shown in
stick representation. (b) Interaction between the designed inhibitor and GSK-3b, the hydrogen bonds were shown by blue dash lines. (c) Chemical modifications of the parent
compound 1b to selective GSK-3b inhibitors.
Scheme 1. Synthesis of 8a–i. (i) Boc2O, DCM, Et3N, stirring, rt, 4 h; (ii) TFA, stirring, rt, 10 h; (iii) NBS, AIBN, DCM, stirring, 40 °C; (iv) K2CO3, THF, 0 °C to room temperature,
stirring overnight; and then TFA, stirring, 50 °C, 10 h.

H. Yue et al. / Bioorganic Chemistry 61 (2015) 21–27
23
coupled with the sulfonyl piperidine amine substrates under the
classic SN2 reaction conditions (K2CO3, THF, 0 °C to room temper-
ature). The coupled product was refluxed in trifluoroacetic acid to
remove the protected group of isobutyl benzene (Cumene). The
final products 8a–i were obtained, and characterized by 1H NMR,
13C NMR and ESI-MS spectra.
2.3. Kinase inhibition assay
All the compounds (8a–i) were evaluated for their in vitro inhi-
bitory activity against GSK-3b and CDK2 by using Kinase GloÒ
Luminescent Kinase Assay (Promega) [34–36], and compound 1b
was also tested as reference compound (Table 1). IC50 values of
most of these derivatives against CDK2 were much higher than
100
at 100
1b (IC50 = 1.01 0.14
improved inhibitory activity against GSK-3b (IC50 < 1
l
M, percentage inhibitions of CDK2 were therefore measured
M for all the tested compounds. Compared with compound
M), more than half of the compounds have
M). More
Fig. 3. Plot of relative GSK-3b activity in the presence of 8c, 8d, and 8i (1
function of ATP concentration. The data were fitted to an equation for simple
l
M) as a
l
l
competitive inhibition: GSK-3b activity (% of control) = (Kd-ATP + [ATP]) Â 100
/{Kd-ATP + [ATP] + (Kd-ATP/Ki)[I]}, where [I] = 1 lM. I, inhibitors.
l
strikingly, all the modified derivatives (8a–i) showed decreased
inhibitory activity against CDK2, especially, 8d and 8i did not sig-
compound 1b, the derivatives bearing the sulfonyl group including
8c–f and 8i have modestly improved inhibitory activity against
GSK-3b. The docking structures revealed that these compounds
have favorable electrostatic interactions with the positively
charged surface in the helix D of GSK-3b by the partially negatively
charged sulfonyl oxygen or the terminal functional group (Fig. 4).
Compared with 8d bearing the terminal methyl group, several
derivatives (8a, 8b, 8g, and 8h) bearing larger hydrophobic termi-
nal functional group showed decreased inhibition against GSK-3b,
which might be explained by the unfavorable interactions between
the large hydrophobic functional group of the inhibitors and the
highly polar surface of GSK-3b. In addition, all the derivatives bear-
ing the sulfonyl group have lower inhibitory activity against CDK2
nificantly inhibit CDK2 even at 100
the different electrostatic potential in the helix
l
M. This result indicates that
between
D
GSK-3b and CDK2 can be explored to design highly selective
GSK-3b inhibitors over CDK2.
To further understand the mechanism of GSK-3b inhibited by
these compounds, kinase assays were tested with the three most
potent compounds 8c, 8d and 8i in varied ATP concentrations.
The inhibition of GSK-3b by these three compounds was reduced
with the increasing concentration of ATP, which is consistent with
the direct competitive inhibition with ATP (Fig. 3).
To delineate the structure–activity relationship (SAR) of these
compounds, we applied molecular docking to analyze the interac-
tions between the inhibitors and GSK-3b [37]. Compared with
Table 1
GSK-3b and CDK2 inhibitory activity of compounds 1b and 8a–i.
Entry
Products
R
IC50 for GSK-3b (
l
M)
Inhibition for CDK2 (%) at 100 lM
1
2
1b
8a
—
1.01 0.14
3.73 0.23
67.4 1.2
35.4 9.5
3
4
8b
8c
2.61 0.22
0.31 0.03
8.7 0.9
12.3 6.3
5
6
8d
8e
0.36 0.06
0.53 0.11
3.4 1.5
23.7 3.4
7
8f
0.53 0.03
16.9 3.0
8
9
8g
8h
1.03 0.16
1.53 0.35
40.8 7.5
12.3 4.1
10
8i
0.54 0.09
2.7 1.2

24
H. Yue et al. / Bioorganic Chemistry 61 (2015) 21–27
Fig. 4. Docked structures of 8c, 8d, and 8i with GSK-3b (green), CDK2 (cyan) was also shown for analysis.
Fig. 5. (a) Wnt pathway activation by 1b, 8d, and 8i. (b) Structural comparison of helix D between GSK-3a (Magenta) and GSK-3b (Green). The identical residues around this
region are shown by stick representations.
compared to compound 1b, which is consistent with the
above-mentioned analysis that the repulsion between the partially
negatively charged sulfonyl group and the negatively charged resi-
dues in the helix D of CDK2 might impair the interactions between
them. We note that the docking modeling should be considered as
a qualitative method rather than a quantitative method to analyze
the subtle SAR for these compounds.
benzo[e]isoindole-1,3-dione derivatives as selective GSK-3b inhibi-
tors over CDK2. These activities of the derivatives were evaluated
by the in vitro and cellular assays, and the structure–activity rela-
tionship was analyzed by in silico molecular docking. The most
active and selective compound 8d inhibits GSK-3b at submicromo-
lar concentration (IC50 = 0.34 0.14
lM), while exhibits no signifi-
cant inhibition against CDK2 (3.4% inhibition at 100
lM). Such
selective GSK-3b inhibitors can be used to explore the biological
functions of GSK-3b and also serve as lead candidates for future
drug development.
2.4. Wnt activation assay
To evaluate the optimization of the GSK-3b inhibitors in the cel-
lular level, we tested the activating activity of the most potent and
selective derivatives 8d and 8i on Wnt/b-catenin pathway by using
Dual-Luciferase Report Assay System (Promega) [38,39], com-
pound 1b was also tested for comparison. 1b activates the
4. Experimental
4.1. General
Wnt/b-catenin pathway by about 4 folds at 100
the DMSO control. Notably, 8d activates the Wnt/b-catenin path-
way by about 11 folds at 100 M, and 8i also has significantly
improved activity but to less extent compared with 8d (Fig. 5a).
This result is consistent with their in vitro inhibitory activity
against GSK-3b. Alternatively, these compounds might activate
lM compared with
All the chemicals were purchased commercially and used with-
out further purification, unless otherwise specified. Anhydrous
dichloromethane (DCM), diisopropylamine (DIPA) and triethy-
lamine (Et3N) was distilled from calcium hydride (CaH2). Dried
tetrahydrofuran (THF) were distilled from sodium benzophenone.
Yields refer to isolated pure products after silica gel column chro-
matography. Reactions were monitored by thin-layer chromatog-
raphy (TLC) on silica gel plates (0.25 mm, 60F-254) using acidic
anisaldehyde or basic aqueous potassium permanganate
(KMnO4) and UV light as visualizing method. Tsingdao silica gel
(60, particle size 0.040–0.063 mm) was used for flash column chro-
matography. 1H NMR and 13C NMR spectra were recorded on a
Bruker Avance 300 (1H, 300 MHz; 13C, 75.5 MHz), Bruker Avance
400 (1H, 400 MHz; 13C, 100 MHz) or Bruker Avance 500 (1H,
500 MHz; 13C, 125 MHz) spectrometer. The following abbrevia-
tions were used to explain the multiplicities: s = singlet,
l
the Wnt/b-catenin pathway by inhibiting GSK-3a, a close homolog
of GSK-3b, that shares high similarity in the helix D as that of
GSK-3b (Fig. 5b) [27], and also has similar functions in the
Wnt/b-catenin pathway as that of GSK-3b [40].
3. Conclusion
In summary, detailed structural analysis revealed a difference of
electrostatic potential of the helix D outside the ATP sites between
GSK-3b and CDK2, which was explored to optimize a series of

H. Yue et al. / Bioorganic Chemistry 61 (2015) 21–27
25
d = doublet, t = triplet, q = quartet, m = multiplet, brs = broad. Mass
spectrometric data were obtained using QSTAR Elite mass spec-
trometer. Melting points were measured by X-4B micromelting
point detector.
4.2.5. 1-(4-(trifluoromethyl)phenylsulfonyl) piperidin-4-amine (4e)
4e was obtained as a white solid in 75% yield. 1H NMR
(500 MHz, CDCl3) d 7.88 (d, J = 8.2 Hz, 2H), 7.80 (d, J = 8.2 Hz, 2H),
4.45 (s, 1H), 3.71 (d, J = 8.7 Hz, 2H), 3.41 (s, 1H), 2.50 (t,
J = 11.1 Hz, 2H), 1.98 (d, J = 11.0 Hz, 2H), 1.49 (dd, J = 20.6,
11.1 Hz, 2H), 1.40 (s, 9H); 13C NMR (125 MHz, CDCl3) d 154.94,
140.43, 134.67, 134.41, 128.00, 126.25, 126.22, 124.30, 122.13,
79.68, 47.03, 45.09, 31.78, 28.31. HRMS (+ESI): m/z calcd. for
4.2. General procedure for synthesis of compound 4a–i
To a stirred solution of 2 (0.20 g, 1.00 mmol) in anhydrous DCM
(5 mL) was added Et3N (0.42 mL, 3.00 mmol) at room temperature.
Then 3a–i (3.96 mmol) was added in small portions at room tem-
perature. The reaction was stirred for 4 h at this temperature and
then quenched the reaction with saturated aqueous solution of
NaHCO3. The aqueous layer was extracted with ethyl acetate
(3 Â 10 mL). The combined organic layers were dried over anhy-
drous Na2SO4, filtered, and concentrated. Silica gel flash column
chromatography (EtOAc/hexanes = 1:2) of the residue gave 4a–i
as the product. Dissolved 4a–i (0.34 g, 0.97 mmol) in 25 mL TFA
at room temperature, the reaction was stirred at room temperature
for 10 h. Distilled the TFA under vacuum and then diluted with sat-
urated aqueous solution of NaHCO3. The aqueous layer was
extracted with ethyl acetate (3 Â 10 mL). The combined organic
layers were dried over anhydrous Na2SO4, filtered, and concen-
trated. Silica gel flash column chromatography (EtOAc/hexanes =
1:1) of the residue gave 5a–i as the product.
C
17H23F3N2NaO4S (M + Na)+ 431.1228; found 431.1221. Mp: 227–
229 °C.
4.2.6. 1-(2-nitrophenylsulfonyl) piperidin-4-amine (4f)
4f was obtained as a white solid in 93% yield. 1H NMR
(500 MHz, CDCl3) d 7.89 (d, J = 7.3 Hz, 1H), 7.66 (dd, J = 13.2,
7.0 Hz, 2H), 7.55 (d, J = 7.3 Hz, 1H), 4.69 (d, J = 7.7 Hz, 1H), 3.71
(d, J = 12.4 Hz, 2H), 3.48 (s, 1H), 2.83 (t, J = 11.7 Hz, 2H), 1.93 (d,
J = 12.0 Hz, 2H), 1.50–1.40 (m, 2H), 1.37 (s, 9H); 13C NMR
(125 MHz, CDCl3)
d 155.07, 148.24, 133.79, 131.67, 130.72,
124.09, 79.45, 47.05, 44.90, 31.90, 28.34. HRMS (+ESI): m/z calcd.
for C16H23N3NaO6S (M + Na)+ 408.1205; found 408.1200. Mp:
225–226 °C.
4.2.7. 1-(phenylsulfonyl) piperidin-4-amine (4g)
4g was obtained as a white solid in 95% yield. 1H NMR
(400 MHz, D-DMSO) d 7.69 (d, J = 7.6 Hz, 2H), 7.54 (t, J = 7.2 Hz,
1H), 7.48 (t, J = 7.4 Hz, 2H), 4.65 (d, J = 6.4 Hz, 1H), 3.63 (d,
J = 8.3 Hz, 2H), 3.31 (s, 1H), 2.37 (t, J = 10.0 Hz, 2H), 1.89 (d,
J = 11.7 Hz, 2H), 1.53–1.38 (m, 2H), 1.34 (s, 9H); 13C NMR
(125 MHz, CDCl3) d 155.04, 136.23, 132.79, 129.05, 127.51, 79.35,
77.43, 77.17, 76.92, 47.06, 45.20, 31.65, 28.33. HRMS (+ESI): m/z
calcd. for C16H24N2NaO4S (M + Na)+ 363.1354; found 363.1350.
Mp: 217–218 °C.
4.2.1. 1-tosylpiperidin-4-amine (4a)
4a was obtained as a white solid in 97% yield. 1H NMR
(500 MHz, CDCl3) d 7.55 (d, J = 8.1 Hz, 2H), 7.25 (d, J = 8.0 Hz, 2H),
4.71 (d, J = 7.2 Hz, 1H), 3.59 (s, 2H), 3.29 (s, 1H), 2.35 (s, 5H), 1.87
(d, J = 11.3 Hz, 2H), 1.53–1.37 (m, 2H), 1.32 (s, 9H); 13C NMR
(125 MHz, CDCl3) d 155.05, 143.57, 133.11, 129.66, 127.58, 79.29,
47.04, 45.21, 31.63, 28.32, 21.44. HRMS (+ESI) m/z calcd. for
C
17H26N2NaO4S (M + Na)+ 377.1511; found 377.1506. Mp: 163–
4.2.8. 1-(quinolin-8-ylsulfonyl) piperidin-4-amine (4h)
4h was obtained as a white solid in 80% yield. 1H NMR
(500 MHz, CDCl3) d 9.08–8.95 (m, 1H), 8.51–8.37 (m, 1H), 8.22
(dd, J = 8.3, 1.4 Hz, 1H), 8.01 (d, J = 8.1 Hz, 1H), 7.59 (t, J = 7.8 Hz,
1H), 7.51 (dd, J = 8.3, 4.2 Hz, 1H), 4.48 (s, 1H), 4.01 (d, J = 12.1 Hz,
2H), 3.45 (s, 1H), 2.93 (d, J = 11.8 Hz, 2H), 1.94 (d, J = 10.6 Hz,
2H), 1.82 (s, 1H), 1.44 (dd, J = 11.9, 3.6 Hz, 2H), 1.39 (s, 9H); 13C
NMR (125 MHz, CDCl3) d 155.07, 151.09, 144.21, 137.30, 136.40,
133.35, 132.90, 129.02, 125.49, 122.01, 79.47, 47.62, 45.35, 32.63,
165 °C.
4.2.2. 1-(4-methoxyphenylsulfonyl) piperidin-4-amine (4b)
4b was obtained as a white solid in 87% yield. 1H NMR
(400 MHz, CDCl3) d 7.64 (d, J = 8.7 Hz, 2H), 6.96 (d, J = 8.7 Hz, 2H),
4.54 (s, 1H), 3.83 (s, 3H), 3.62 (d, J = 9.3 Hz, 2H), 3.38 (d,
J = 42.1 Hz, 1H), 2.37 (t, J = 10.7 Hz, 2H), 1.92 (d, J = 9.4 Hz, 2H),
1.52–1.41 (m, 2H), 1.38 (d, J = 13.8 Hz, 9H); 13C NMR (100 MHz,
CDCl3) d 163.02, 155.03, 129.69, 127.64, 114.25, 79.46, 55.62,
47.06, 45.22, 31.69, 28.34. HRMS (+ESI): m/z calcd. for
28.37. HRMS (+ESI): m/z calcd. for
392.1644; found 392.1639. Mp: 210–212 °C.
C
19H26N3O4S (M + H)+
C
17H26N2NaO5S (M + Na)+ 393.1460; found 393.1454. Mp: 165–
4.2.9. 1-(thiophen-2-ylsulfonyl) piperidin-4-amine (4i)
166 °C.
4i was obtained as a white solid in 99% yield. 1H NMR
(500 MHz, CDCl3) d 7.57 (dd, J = 5.0, 1.2 Hz, 1H), 7.46 (dd, J = 3.7,
1.3 Hz, 1H), 7.08 (dd, J = 5.0, 3.8 Hz, 1H), 4.68 (d, J = 7.8 Hz, 1H),
3.62 (d, J = 10.1 Hz, 2H), 3.34 (s, 1H), 2.45 (t, J = 11.0 Hz, 2H), 1.92
(d, J = 10.6 Hz, 2H), 1.55–1.39 (m, 2H), 1.34 (s, 9H); 13C NMR
(125 MHz, CDCl3) d 155.06, 136.56, 132.30, 132.08, 127.65, 79.40,
47.01, 45.27, 31.51, 28.34. HRMS (+ESI): m/z calcd. for
4.2.3. 1-(4-nitrophenylsulfonyl) piperidin-4-amine (4c)
4c was obtained as a pale yellow solid in 98% yield. 1H NMR
(500 MHz, CDCl3) d 8.43–8.34 (m, 2H), 8.01–7.91 (m, 2H), 4.42 (s,
1H), 3.75 (d, J = 10.4 Hz, 2H), 3.42 (s, 1H), 2.54 (s, 2H), 2.09–1.94
(m, 2H), 1.63 (s, 1H), 1.50 (dd, J = 11.9, 3.1 Hz, 2H), 1.41 (s, 9H);
13C NMR (125 MHz, CDCl3) d 154.91, 150.28, 128.64, 124.35,
79.79, 46.98, 45.10, 31.79, 28.32. HRMS (+ESI): m/z calcd. For
C
14H22N2NaO4S2 (M + Na)+ 369.0919; found 369.0912. Mp: 142–
143 °C.
C
16H23N3NaO6S (M + Na)+ 408.1205; found 408.1201. Mp: 228–
229 °C.
4.3. General procedure for synthesis of compound 8a–i
4.2.4. 1-(methylsulfonyl)piperidin-4-amine (4d)
To a stirred suspension of 6 (0.36 g, 0.93 mmol) and NBS (0.17 g,
0.99 mmol) in anhydrous DCM (5 mL) was added AIBN (0.06 g,
0.36 mmol) in one portion under N2 atmosphere at room temper-
ature. The reaction was stirred at 40 °C. TLC analysis showed con-
sumption of 6. The reaction was filtered and concentrated. Silica
gel flash column chromatography (EtOAc/hexanes 1:25) of the
residue gave a white solid 7 (0.37 g, 0.81 mmol, 87%) as the prod-
uct. To a stirred solution of 5a (0.21 mmol) in anhydrous THF
4d was obtained as a colorless oil liquid in 55% yield. 1H NMR
(400 MHz, CDCl3) d 4.55 (d, J = 6.0 Hz, 1H), 3.72 (d, J = 11.7 Hz,
2H), 3.54 (s, 1H), 2.88–2.69 (m, 5H), 2.07–1.93 (m, 2H), 1.52–
1.44 (m, 2H), 1.42 (s, 9H); 13C NMR (100 MHz, CDCl3) d 155.08,
79.68, 47.22, 45.05, 34.88, 31.98, 28.38. HRMS (+ESI): m/z calcd.
for C11H22N2NaO4S (M + Na)+ 301.1198; found 301.1192. Mp:
162–163 °C.

26
H. Yue et al. / Bioorganic Chemistry 61 (2015) 21–27
(1 mL) was added anhydrous K2CO3 (56 mg, 0.40 mmol) at 0 °C.
The resulting mixture was stirred at this temperature for 30 min.
To this mixture was added 7 (50 mg, 0.11 mmol) at 0 °C. The
resulting mixture was stirred at room temperature overnight.
The reaction was quenched with saturated aqueous solution of
NaHCO3. The aqueous layer was extracted with ethyl acetate
(3 Â 10 mL). The combined organic layers were dried over anhy-
drous Na2SO4, filtered, and concentrated. The crude product was
dissolved in 10 mL TFA. The resulting mixture was stirred at
50 °C for 10 h. Distilled the TFA under vacuum and then diluted
with saturated aqueous solution of NaHCO3. The aqueous layer
was extracted with ethyl acetate (3 Â 10 mL). The combined
organic layers were dried over anhydrous Na2SO4, filtered, and
concentrated. Silica gel flash column chromatography
(EtOAc/hexanes 1:2) of the residue gave a yellow solid 8a–i as
the product.
31.70. HRMS (+ESI): m/z calcd. for C21H26N3O6S (M + H)+
448.1542; found 448.1537. Mp: 206–207 °C.
4.3.5. 7,8-dimethoxy-5-((1-(4-
(trifluoromethyl)phenylsulfonyl)piperidin-4-ylamino)methyl)-1H-
benzo[e]isoindole-1,3 (2H)-dione (8e)
8e was obtained as a pale yellow solid in 26% two-steps yield
(iv). 1H NMR (300 MHz, D-DMSO) d 11.06 (s, 1H), 8.07 (s, 1H),
8.00 (d, J = 8.5 Hz, 2H), 7.94 (d, J = 8.4 Hz, 2H), 7.65 (s, 1H), 7.45
(s, 1H), 4.12 (s, 2H), 3.89 (d, J = 4.1 Hz, 6H), 3.47 (d, J = 7.8 Hz,
2H), 2.52–2.54 (m, 3H), 2.00–1.92 (m, 2H), 1.44–1.41 (m, 2H);
13C NMR (75 MHz, D-DMSO) d 171.60, 170.43, 152.11, 151.18,
140.20, 133.32, 132.89, 131.42, 130.10, 129.61, 128.82, 127.07,
127.01, 125.73, 124.88, 124.39, 122.11, 116.53, 104.36, 102.74,
55.99, 55.96, 53.03, 48.32, 44.88, 31.40, 29.49. HRMS (+ESI): m/z
calcd. for C27H27F3N3O6S (M + H)+ 578.1573; found 578.1558. Mp:
219–220 °C.
4.3.1. 7,8-dimethoxy-5-((1-tosylpiperidin-4-ylamino)methyl)-1H-
benzo[e]isoindole-1,3(2H)-dione (8a)
4.3.6. 7,8-dimethoxy-5-((1-(2-nitrophenylsulfonyl)piperidin-4-
ylamino)methyl)-1H-benzo[e]isoindole-1,3(2H)-dione (8f)
8a was obtained as a pale yellow solid in 34% two-steps yield
(iv). 1H NMR (500 MHz, D-DMSO) d 11.01 (s, 1H), 8.01 (s, 1H),
7.59 (d, J = 8.6 Hz, 3H), 7.41 (d, J = 8.6 Hz, 3H), 4.04 (d, J = 27.8 Hz,
2H), 3.89 (s, 3H), 3.87 (s, 3H), 3.41 (d, J = 11.1 Hz, 2H), 2.44–2.48
(m, 3H), 2.38 (s, 3H), 1.98–1.89 (m, 2H), 1.48–1.37 (m, 2H); 13C
8f was obtained as a pale yellow solid in 78% two-steps yield
(iv). 1H NMR (400 MHz, D-DMSO) d 11.10 (s, 1H), 8.11 (s, 1H),
8.98–7.95 (m, 2H), 7.89–7.83 (m, 2H), 7.73 (s, 1H), 7.51 (s, 1H),
4.22 (s, 2H), 3.95 (s, 6H), 3.59 (s, 2H), 2.95–2.85 (m, 2H), 2.52 (s,
1H), 2.04–1.96 (m, 2H), 1.47–1.45 (m, 2H); 13C NMR (100 MHz,
NMR (125 MHz, D-DMSO)
d 171.56, 170.40, 152.08, 151.17,
D-DMSO)
d 171.28, 169.89, 152.62, 152.03, 148.22, 135.39,
143.85, 133.20, 131.40, 130.23, 129.58, 127.88, 124.85, 124.37,
116.48, 104.31, 102.76, 55.97, 55.94, 53.32, 48.45, 44.91, 31.43,
132.90, 131.44, 130.86, 130.10, 129.21, 126.64, 124.71, 124.57,
119.18, 116.15, 104.38, 102.97, 56.37, 56.18, 54.77, 45.16, 44.57,
21.44. HRMS (+ESI): m/z calcd. for
524.1855; found 524.1846. Mp: 208–209 °C.
C
27H30N3O6S (M + H)+
28.17. HRMS (+ESI): m/z calcd. for
555.1550; found 555.1442. Mp: 218–220 °C.
C
26H27N4O8S (M + H)+
4.3.2. 7,8-dimethoxy-5-((1-(4-methoxyphenylsulfonyl)piperidin-4-
ylamino)methyl)-1H-benzo[e]isoindole-1,3(2H)-dione (8b)
4.3.7. 7,8-dimethoxy-5-((1-(phenylsulfonyl)piperidin-4-
ylamino)methyl)-1H-benzo[e]isoindole-1,3(2H)-dione (8g)
8b was obtained as a pale yellow solid in 35% two-steps yield
(iv). 1H NMR (400 MHz, D-DMSO) d 11.28 (s, 1H), 8.16 (s, 1H),
7.84 (s, 1H), 7.74–7.62 (m, 2H), 7.42 (s, 1H), 7.26–7.09 (m, 2H),
4.70 (s, 2H), 3.99 (s, 3H), 3.93 (s, 3H), 3.84 (s, 3H), 3.73 (d,
J = 11.4 Hz, 2H), 2.48(s, 1H), 2.25–2.33 (m, 4H), 1.73–1.71 (m,
2H); 13C NMR (100 MHz, D-DMSO) d 171.24, 169.87, 163.30,
152.62, 152.05, 134.35, 131.42, 130.17, 129.21, 127.30, 126.63,
124.57, 119.13, 115.07, 104.33, 103.01, 67.72, 67.66, 56.39, 56.21,
56.18, 54.89, 45.21, 44.95, 27.92. HRMS (+ESI): m/z calcd. for
8g was obtained as a pale yellow solid in 84% two-steps yield
(iv). 1H NMR (400 MHz, D-DMSO) d 11.00 (s, 1H), 7.94 (s, 1H),
7.70 (dd, J = 13.6, 7.2 Hz, 3H), 7.62 (t, J = 7.4 Hz, 2H), 7.56 (s, 1H),
7.33 (s, 1H), 4.04 (s, 2H), 3.85 (d, J = 6.2 Hz, 6H), 3.45 (d,
J = 11.7 Hz, 2H), 2.54–2.43 (m, 3H), 1.94 (dd, J = 10.8, 6.5 Hz, 2H),
1.45–1.38 (m, 2H); 13C NMR (100 MHz, D-DMSO)
d 171.53,
170.38, 151.97, 151.08, 143.24, 136.11, 133.52, 131.29, 129.81,
129.48, 127.82, 124.77, 124.27, 104.15, 102.66, 55.93, 55.88,
53.38, 48.42, 44.96, 31.40. HRMS (+ESI): m/z calcd. for
C
27H30N3O7S (M + H)+ 540.1804; found 540.1802. Mp: 209–210 °C.
C
26H28N3O6S (M + H)+ 510.1699; found 510.1697. Mp: 235–236 °C.
4.3.3. 7,8-dimethoxy-5-((1-(4-nitrophenylsulfonyl)piperidin-4-
ylamino)methyl)-1H-benzo[e]isoindole-1,3(2H)-dione (8c)
4.3.8. 7,8-dimethoxy-5-((1-(quinolin-8-ylsulfonyl)piperidin-4-
ylamino)methyl)-1H-benzo[e]isoindole-1,3(2H)-dione (8h)
8h was obtained as a pale yellow solid in 75% two-steps yield
(iv). 1H NMR (400 MHz, D-DMSO) d 11.03 (s, 1H), 9.03 (dd,
J = 4.1, 1.5 Hz, 1H), 8.71–8.45 (m, 1H), 8.34 (d, J = 6.5 Hz, 1H),
8.27 (d, J = 8.1 Hz, 1H), 8.04 (d, J = 11.4 Hz, 1H), 7.73 (t, J = 7.8 Hz,
1H), 7.69–7.61 (m, 1H), 7.44 (s, 1H), 4.10 (d, J = 10.5 Hz, 2H), 3.89
(s, 3H), 3.85 (s, 3H), 3.83–3.71 (m, 2H), 2.90 (t, J = 10.6 Hz, 2H),
2.58 (s, 1H), 1.96 (dd, J = 18.9, 12.8 Hz, 2H), 1.40–1.30 (m, 2H);
13C NMR (100 MHz, D-DMSO) d 171.59, 170.43, 152.11, 151.73,
151.17, 143.83, 137.25, 136.93, 134.31, 132.82, 131.43, 129.62,
129.20, 126.19, 124.85, 124.72, 124.38, 122.87, 116.49, 104.40,
102.77, 55.96, 53.94, 48.39, 45.03, 32.35. HRMS (+ESI): m/z calcd.
for C29H29N4O6S (M + H)+ 561.1808; found 561.1804. Mp: 217–
218 °C.
8c was obtained as a pale yellow solid in 42% two-steps yield
(iv). 1H NMR (400 MHz, D-DMSO) d 11.04 (s, 1H), 8.40 (d,
J = 8.6 Hz, 2H), 8.12–7.93 (m, 3H), 7.61 (s, 1H), 7.41 (s, 1H), 4.10
(s, 2H), 3.88 (s, 3H), 3.87 (s, 3H), 3.47 (d, J = 11.6 Hz, 2H), 2.60 (t,
J = 9.7 Hz, 3H), 1.92 (t, J = 13.4 Hz, 2H), 1.46–1.39 (m, 2H); 13C
NMR (100 MHz, D-DMSO)
d 171.56, 170.40, 152.08, 151.15,
150.41, 141.90, 131.36, 129.55, 129.39, 125.10, 124.86, 124.36,
116.47, 104.29, 102.72, 55.99, 55.94, 52.97, 48.26, 46.16, 44.77,
31.27. HRMS (+ESI): m/z calcd. for
555.1550; found 555.1563. Mp: 221–222 °C.
C
26H27N4O8S (M + H)+
4.3.4. 7,8-dimethoxy-5-((1-(methylsulfonyl)piperidin-4-
ylamino)methyl)-1H-benzo[e]isoindole-1,3(2H)-dione (8d)
8d was obtained as a pale yellow solid in 90% two-steps yield
(iv). 1H NMR (300 MHz, D-DMSO) d 11.09 (s, 1H), 8.10 (s, 1H),
7.74 (s, 1H), 7.53 (s, 1H), 4.21 (s, 2H), 3.95 (s, 3H), 3.92 (s, 3H),
3.55–3.42 (m, 2H), 2.85–2.75 (m, 5H), 1.98 (d, J = 7.0 Hz, 2H),
1.48 (d, J = 11.5 Hz, 2H); 13C NMR (75 MHz, D-DMSO) d 171.64,
170.49, 152.16, 151.26, 131.47, 130.11, 129.65, 124.90, 124.42,
116.52, 104.43, 102.78, 56.08, 55.99, 53.66, 48.21, 44.46, 35.57,
4.3.9. 7,8-dimethoxy-5-((1-(thiophen-2-ylsulfonyl)piperidin-4-
ylamino)methyl)-1H-benzo[e]isoindole-1,3(2H)-dione (8i)
8i was obtained as a pale yellow solid in 35% two-steps yield
(iv). 1H NMR (400 MHz, D-DMSO) d 11.29 (s, 1H), 9.32 (s, 2H),
8.17 (s, 1H), 8.07 (d, J = 4.9 Hz, 1H), 7.86 (s, 1H), 7.67 (d,
J = 2.7 Hz, 1H), 7.43 (s, 1H), 7.30 (dd, J = 4.7, 4.0 Hz, 1H), 4.73 (s,

H. Yue et al. / Bioorganic Chemistry 61 (2015) 21–27
27
2H), 4.04 (s, 3H), 3.94 (s, 3H), 3.77 (d, J = 11.8 Hz, 2H), 2.48–2.42
References
(m, 3H), 2.32 (d, J = 11.0 Hz, 2H), 1.80–1.72 (m, 2H); 13C NMR
(100 MHz, D-DMSO) d 171.25, 169.88, 152.64, 152.08, 135.60,
134.49, 133.63, 131.38, 129.21, 128.87, 126.68, 124.58, 119.14,
104.31, 103.01, 67.73, 67.65, 56.43, 56.20, 54.78, 45.21, 45.03,
27.78, 20.44. HRMS (+ESI): m/z calcd. for C24H26N3O6S2 (M + H)+
516.1263; found 516.1258. Mp: 201–202 °C.
[1] R.T. Moon, A.D. Kohn, G.V. De Ferrari, A. Kaykas, Nat. Rev. Genet. 5 (2004) 691–
701.
[2] H. Clevers, Cell 127 (2006) 469–480.
[3] B.T. MacDonald, K. Tamai, X. He, Dev. Cell 17 (2009) 9–26.
[4] A. Kikuchi, H. Yamamoto, A. Sato, Trends Cell Biol. 19 (2009) 119–129.
[5] H. Clevers, R. Nusse, Cell 149 (2012) 1192–1205.
[6] G.V. De Ferrari, N.C. Inestrosa, Brain Res. Rev. 33 (2000) 1–12.
[7] N.C. Inestrosa, L. Varela-Nallar, J. Mol. Cell Biol. 6 (2014) 64–74.
[8] W. Wan, S. Xia, B. Kalionis, L. Liu, Y. Li, BioMed. Res. Int. 2014 (2014) 301575.
[9] D.C. Berwick, K. Harvey, Biochem. Soc. Trans. 40 (2012) 1123–1128.
[10] B. Marchetti, F. L’Episcopo, M.C. Morale, C. Tirolo, N. Testa, S. Caniglia, M.F.
Serapide, S. Pluchino, Eur. J. Neurosci. 37 (10) (2013) 1550–1563.
[11] D. Cotter, R. Kerwin, S. Al-Sarraji, J.P. Brion, A. Chadwich, S. Lovestone, B.
Anderton, I. Everall, Neuroreport 9 (7) (1998) 1379–1383.
[12] N.C. Inestrosa, C. Montecinos-Oliva, M. Fuenzalida, J. Neuroimmune
Pharmacol. 7 (2012) 788–807.
[13] K.K. Singh, Clin. Genet. 83 (6) (2013) 511–517.
[14] X. Zeng, K. Tamai, B. Doble, S. Li, H. Huang, R. Habas, H. Okamura, J. Woodgett,
X. He, Nature 438 (2005) 873–877.
[15] R. Luna-Medina, M. Cortes-Canteli, S. Sanchez-Galiano, J.A. Morales-Garcia, A.
Martinez, A. Santos, A. Perez-Castillo, J. Neurosci. 27 (2007) 5766–5776.
[16] G. Wu, H. Huang, J.G. Abreu, X. He, PLoS One 4 (2009) e4926.
[17] D. Wu, W. Pan, Trends Biochem. Sci. 35 (2010) 161–168.
[18] H. Eldar-Finkelman, A. Martinez, Front. Mol. Neurosci. 4 (2011) 32.
[19] M. Welcker, J. Singer, K.R. Loeb, J. Grim, A. Bloecher, M. Gurien-West, B.E.
Clurman, J.M. Roberts, Mol. Cell 12 (2003) 381–392.
4.4. Kinase inhibition assays
The inhibitory activity of the tested compounds against GSK-3b
and CDK2 was measured by using Kinase-Glo luminescent kinase
assay (Promega). Briefly, recombinant human GSK3b (1
10 l) was mixed with 10 of GSK-3b substrate
(YRRAAVPPSPSLSRHSSPHQpSEDEEE, 100 M), 19
l of 2.5 Â kinase
buffer (50 mM Tris–HCl, pH 7.4, 25 mM MgCl2), 10 l of 10
ATP, 1 l of the indicated amount of chemicals in DMSO to a final
volume of 50 l. Control reaction without kinase was set up. The
mixtures were incubated at 30 °C for 40 min. The enzymatic reac-
tion was stopped with 50 L of Kinase-Glo reagent. Glow-type
luminescence was recorded after 10 min using a TECAN infinite
F200 multimode reader. On the other hand, CDK2 (1 M, 10 l)
was mixed with 10 l of CDK2 substrate (PKTPKKAKKL, 100 M),
19
l of 2.5 Â kinase buffer (50 mM Tris–HCl, pH 7.4, 25 mM
MgCl2), 10 l of 10 M ATP, 1 l of the indicated amount of chem-
icals in DMSO to a final volume of 50 l. Control reaction without
lM,
l
ll
l
l
l
lM
l
l
l
l
l
[20] A. Vulpetti, P. Crivori, A. Cameron, J. Bertrand, M.G. Brasca, R.D. Alessio, P.
Pevarello, J. Chem. Inf. Model 45 (5) (2005) 1282–1290.
[21] I. Buch, D. Fishelovitch, N. London, B. Raveh, H.J. Wolfson, R. Nussinov,
Biochemistry 49 (2010) 10890–10901.
l
l
l
l
l
l
[22] P.M. Fischer, Chem. Biol. 10 (12) (2003) 1144–1146.
l
[23] Y. Mettey, M. Gompel, V. Thomas, M. Garnier, M. Leost, I. Ceballos-Picot, M.
Noble, J. Endicott, J.M. Vierfond, L. Meijer, J. Med. Chem. 46 (2003) 222–236.
[24] T.J. Soos, L. Meijer, P.J. Nelson, Drug News Perspect. 19 (6) (2006) 325–328.
[25] E. Damiens, B. Baratte, D. Marie, G. Eisenbrand, L. Meijer, Oncogene 20 (29)
(2001) 3786–3797.
[26] T. Kramer, B. Schmidt, F. Lo, Int. J. Alzheimers Dis. 2012 (2012) 381029.
[27] L. Meijer, M. Flajolet, P. Greengard, Trends Pharmacol. Sci. 25 (9) (2004) 471–
480.
[28] D. Lesuisse, G. Dutruc-Rosset, G. Tiraboschi, M.K. Dreyer, S. Maignan, A.
Chevalier, F. Halley, P. Bertrand, M.C. Burgevin, D. Quarteronet, T. Rooney,
Bioorg. Med. Chem. Lett. 20 (2010) 1985–1989.
[29] L. Avrahami, A. Licht-Murava, M. Eisenstein, H. Eldar-Finkelman, Biochim.
Biophys. Acta 1834 (2013) 1410–1414.
[30] B. Georgievska, J. Sandin, J. Doherty, A. Mörtberg, J. Neelissen, A. Andersson, S.
Gruber, Y. Nilsson, P. Schött, P. Arvidsson, S. Hellberg, G. Osswald, S. Berg, J.
Fälting, R.V. Bhat, J. Neurochem. 125 (3) (2013) 446–456.
[31] D.S. Patel, P.V. Bharatam, Eur. J. Med. Chem. 43 (5) (2008) 949–957.
[32] H. Zhong, H. Zou, M.V. Semenov, D. Moshinsky, X. He, H. Huang, S. Li, J. Quan, Z.
Yang, S. Lin, Mol. Biosyst. 5 (2009) 1356–1360.
[33] H. Zou, L. Zhou, Y. Li, H. Zhong, Z. Pan, Z. Yang, J. Quan, J. Med. Chem. 53 (2010)
994–1003.
[34] M.I. Rafique, R.B. Jeffrey, MethodsX 1 (2014) 96–101.
[35] J.A. Morales-García, C. Susín, S. Alonso-Gil, D.I. Pérez, V. Palomo, C. Pérez, S.
Conde, A. Santos, C. Gil, A. Martínez, A. Pérez-Castillo, Chem. Neurosci. 4 (2013)
350–360.
kinase was set up. The following steps were the same as the detec-
tion of GSK-3 inhibitory activity.
4.5. b-catenin–TCF reporter gene assay
L cells on 96 well plates were transfected with 200 ng TOPflash
reporter plasmid and 100 ng internal control vectors pRL-SV40
using Lipofectamine TM2000 (Life, Cat. No. 11668019) and
Opti-MEM (Life, Cat. No. 31985-062) according to the manufac-
turer’s instructions. 24 h after transfection the tested compounds
were added to cells at 100 lM concentration for 24 h. Control cells
were DMSO vehicle treated. Then cell extracts were prepared and
examined sequentially for firefly and Renilla luciferase activities
using the Dual-Luciferase Reporter assay system (Promega,
Madison, WI). Firefly luciferase readings were normalized against
Renilla luciferase.
4.6. Molecular docking
[36] A. Baki, A. Bielik, L. Molnár, G. Szendrei, G.M. Keserü, Assay Drug Dev. Technol.
5 (2007) 75–83.
[37] G.M. Morris, D.S. Goodsell, R.S. Halliday, R. Huey, W.E. Hart, R.K. Belew, A.J.
Olson, J. Comput. Chem. 19 (1998) 1639–1662.
[38] L. Rinnab, S.V. Schütz, J. Diesch, E. Schmid, R. Küfer, R.E. Hautmann, K.D.
Spindler, M.V. Cronauer, Neoplasia 10 (6) (2008) 624–634.
[39] M.P. Coghlan, A.A. Culbert, D.A. Cross, S.L. Corcoran, J.W. Yates, N.J. Pearce, O.L.
Rausch, G.J. Murphy, P.S. Carter, C.L. Roxbee, D. Mills, M.J. Brown, D. Haigh,
R.W. Ward, D.G. Smith, K.J. Murray, A.D. Reith, J.C. Holder, Chem. Biol. 7 (2000)
793–803.
Protein coordinates for docking were obtained from the crystal
structure of GSK-3b in complex with indirubin (PDB entry 1UV5)
[41] and CDK2 in complex with indirubin (PDB entry 1E9H) [42].
The indirubin binding sites were used to dock compounds 1b
and 8a–i. Waters and indirubin were removed from the PDB file,
and the polar hydrogen atoms were added to the amino acid resi-
dues before the docking study. Docking was performed using
AutoDock, version 4.0 [37]. All illustrated structures were made
using PyMOL (DeLano Scientific, Palo Alto, CA) [43].
[40] B.W. Doble, S. Patel, G.A. Wood, L.K. Kockeritz, J.R. Woodgett, Dev. Cell 12
(2007) 957–971.
[41] L. Meijer, A.L. Skaltsounis, P. Magiatis, P. Polychronopoulos, M. Knockaert, M.
Leost, X.P. Ryan, C.A. Vonica, A. Brivanlou, R. Dajani, C. Crovace, C. Tarricone, A.
Musacchio, S.M. Roe, L. Pearl, P. Greengard, Chem. Biol. 10 (2003) 1255–1266.
[42] T.G. Davies, P. Tunnah, L. Meijer, D. Marko, G. Eisenbrand, J.A. Endicott, M.E.
Noble, Structure 9 (2001) 389–397.
Acknowledgments
This work was supported by funds from the Ministry of Science
and Technology 2012CB722602 (to J.M.Q.); the NSFC21290183
and NSFC81373326 (to J.M.Q.); Shenzhen government
[43] W.L. DeLano, The PyMOL Molecular Graphics System, DeLano Scientific, Palo
Alto, CA, USA, 2002.
JCYJ20130331144947526,
GJHS20120628101219329
and
GJHZ20120614144733420 (to J.M.Q.).