Synthesis, biological evaluation and 3D-QSAR studies of imidazolidine-2,4-dione derivatives as novel protein tyrosine phosphatase 1B inhibitors

Article abstract of DOI:10.1016/j.ejmech.2015.08.037

Protein tyrosine phosphatase 1B (PTP1B) plays a vital role in the regulation of insulin sensitivity and dephosphorylation of the insulin receptor, so PTP1B inhibitors may be potential agents to treat type 2 diabetes. In this work, a series of novel imidazolidine-2,4-dione derivatives were designed, synthesized and assayed for their PTP1B inhibitory activities. These compounds exhibited potent activities with IC₅₀ values at 0.57-172 μM. A 3D-QSAR study using CoMFA and CoMSIA techniques was carried out to explore structure activity relationship of these molecules. The CoMSIA model was more predictive with q² Combining double low line 0.777, r² Combining double low line 0.999, SEE Combining double low line 0.013 and r²_pred Combining double low line 0.836, while the CoMFA model gave q² Combining double low line 0.543, r² Combining double low line 0.998, SEE Combining double low line 0.029 and r²_pred Combining double low line 0.754. The contour maps derived from the best CoMFA and CoMSIA models combined with docking analysis provided good insights into the structural features relevant to the bioactivity, and could be used in the molecular design of novel imidazolidine-2,4-dione derivatives.

Full text of DOI:10.1016/j.ejmech.2015.08.037

European Journal of Medicinal Chemistry 103 (2015) 91e104
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Synthesis, biological evaluation and 3D-QSAR studies of
imidazolidine-2,4-dione derivatives as novel protein tyrosine
phosphatase 1B inhibitors
,
,
Mei-Yan Wang a 1, Yuan-Yuan Jin a 1, Hui-Yu Wei b, Li-Song Zhang a, Su-Xia Sun a,
Xiu-Bo Chen b, Wei-Li Dong a, Wei-Ren Xu c, Xian-Chao Cheng a *, Run-Ling Wang a
,
, **
a Tianjin Key Laboratory on Technologies Enabling Development of Clinical Therapeutics and Diagnostics (Theranostics), School of Pharmacy, Tianjin
Medical University, Tianjin 300070, China
b Tianjin Medical University Eye Hospital, Tianjin 300384, China
c Tianjin Key Laboratory of Molecular Design and Drug Discovery, Tianjin Institute of Pharmaceutical Research, Tianjin 300193, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Protein tyrosine phosphatase 1B (PTP1B) plays a vital role in the regulation of insulin sensitivity and
dephosphorylation of the insulin receptor, so PTP1B inhibitors may be potential agents to treat type 2
diabetes. In this work, a series of novel imidazolidine-2,4-dione derivatives were designed, synthesized
and assayed for their PTP1B inhibitory activities. These compounds exhibited potent activities with IC50
values at 0.57e172 mM. A 3D-QSAR study using CoMFA and CoMSIA techniques was carried out to explore
structure activity relationship of these molecules. The CoMSIA model was more predictive with
Received 4 January 2015
Received in revised form
6 July 2015
Accepted 17 August 2015
Available online 20 August 2015
q2 ¼ 0.777, r2 ¼ 0.999, SEE ¼ 0.013 and r2
¼ 0.836, while the CoMFA model gave q2 ¼ 0.543,
pred
Keywords:
PTP1B
Imidazolidine-2,4-dione
Synthesis
CoMFA/CoMSIA
Docking
r2 ¼ 0.998, SEE ¼ 0.029 and r2
¼ 0.754. The contour maps derived from the best CoMFA and CoMSIA
pred
models combined with docking analysis provided good insights into the structural features relevant to
the bioactivity, and could be used in the molecular design of novel imidazolidine-2,4-dione derivatives.
© 2015 Elsevier Masson SAS. All rights reserved.
1. Introduction
ertiprotafib, developed by American Home Products, is the first
drug candidate entering PhaseⅡclinical trials. Unfortunately, it was
Type 2 diabetes is associated with a variety of vascular com-
plications resulting from insulin resistance of the human organism
[1]. Over 300 million people are now suffering from type 2 diabetes
and the incidence of it is still on the rise [2e4]. Therefore, devel-
opment of specific insulin-sensitizing drugs is pressing.
withdrawn as a result of undesirable side-effects. Then, ISIS-113715,
an antisense inhibitor of the PTP1B gene, is the second one reaching
PhaseⅡclinical trials [13]. However, despite intensive efforts and
great progress [14e24], a safe and efficient PTP1B inhibitor is still a
big challenge for medicinal chemists.
PTP1B plays as a negative regulator of insulin signaling path-
ways by dephosphorylating insulin receptor (IR) and IR substrate
[5e10]. Studies on PTP1B knockout mice displayed enhanced in-
sulin sensitivity and resistance to high-fat diet induced obesity
[11,12], which suggested that PTP1B might be a promising target for
type 2 diabetes therapy.
During the last decade, glitazones, which shared a common
molecular skeleton of 2,4-thiazolidinediones (TZDs), were
approved by USA FDA as anti-diabetic drugs. These TZD drugs
produced the anti-diabetic effects by agonizing peroxisome
proliferator-activated receptor
g. Interestingly, some TZD de-
rivatives also exhibited significant PTP1B inhibitory activities in
both experiments performed by Ottana et al. [25,26] and Bhattarai
et al. [27]. According to their research, the substituents at N3 po-
sition of TZD scaffold could mimic pTyr group to interact with the
receptor [28], and the lipophilic arylidene group at C5 position of
TZD scaffold could improve the stability of receptor-inhibitor
binding by hydrophobic interaction. Based on the bioisosteric
principle, the TZD scaffold was replaced by imidazolidine-2,4-dione
Among all potential inhibitors of PTP1B known so far,
* Corresponding author.
** Corresponding author.
E-mail addresses: chengxianchao@aliyun.com (X.-C. Cheng), wangrunling@
tijmu.edu.cn (R.-L. Wang).
1
These authors contributed equally to this work.
http://dx.doi.org/10.1016/j.ejmech.2015.08.037
0223-5234/© 2015 Elsevier Masson SAS. All rights reserved.

92
M.-Y. Wang et al. / European Journal of Medicinal Chemistry 103 (2015) 91e104
scaffold, which was then substituted by various benzyl group or
alkyl group at N3 position (R1), benzyl group at N1 position (R2) and
para-substituted benzylidene group at C5 position (R3) (Fig. 1).
These compounds were synthesized and assayed for their PTP1B
inhibitory activities, hoping to get similar or better PTP1B inhibi-
tory activities.
Recently, three-dimensional quantitative structure-activity
relationship (3D-QSAR) technique has been widely employed to
correlate structural descriptors from a cluster of related molecules
to their known bioactivity. In terms of PTP1B inhibitors, the 3D-
QSAR study of 2-(oxalylamino) benzoic acid analogs [29,30],
benzofuran and benzothiophene biphenyls analogs [29], pyridazine
analogs [31], and 2,4-thiazolidinedione [32] have been reported.
However, the 3D-QSAR research about imidazolidine-2,4-dione
derivatives as new PTP1B inhibitors has not been reported till now.
In this paper, the synthesis, biological evaluation and structure
activity relationship of novel imidazolidine-2,4-dione derivatives
as PTP1B inhibitors were reported. The CoMFA/CoMSIA analysis
and docking studies were also performed to investigate
compounds-target binding details.
groups in these two positions (4ae4e, 5a, 5ce5f, 10a, 10fe10g, and
13a) appeared to be more active than the ones with relatively small
substituents in either of the two places (6ae6c, 6f, 9a, 10b, 11b, 11e,
12ae12b, and 13b). Besides, some molecules without substituents
or with an aryl group in part R1 exhibited better bioactivities
(4ae4e, 5a, 5ce5f, 9b, 10a, 10de10g, 13a, and 13c). From the above,
the conceivable SARs of PTP1B inhibitors observed only from
Table 1 is that compounds with hydrophobic, aromatic and bulky
substituents in parts R2 and R3, and with H or an aromatic group in
R1 would be more likely to acquire higher activities (4ae4e, 5a,
5ce5f, 10a, 10de10g, and 13a). The hypothesis will be tested in the
following 3D-QSAR study.
In this paper, CoMFA/CoMSIA and docking studies were per-
formed to investigate the relationship between biological activities
and chemical structures.
3.2. 3D-QSAR analysis
3.2.1. CoMFA/CoMSIA statistical analysis
The CoMFA/CoMSIA models were constructed using 32 imida-
zolidine-2,4-dione derivatives as the training set. This model was
used to predict activities of the test set containing 7 compounds.
The statistical data for the best CoMFA/CoMSIA models were
summarized in Table 2.
2. Chemistry
The synthetic routes of all the 39 target compounds were shown
in Schemes 1e4.
Using 4-hydroxybenzaldehyde and hydantoin as the starting
material, compounds 4aee, 5aef and 6aeh were synthesized via
Knoevenagel condensation and alkylation with various haloge-
nated compounds (Scheme 1).
As shown in Scheme 2, N1-benzyl hydantoin was reacted with
methyl bromoacetate using K2CO3 as the catalyst to produce 8a.
Compound 8a was further reacted with 4-hydroxybenzaldehyde to
yield compound 9a. On one hand, 9a was reacted with halogenated
compounds using K2CO3 as the catalyst to afford the compounds
10ae10c, which were further hydrolyzed to obtain corresponding
11ae11c. On the other hand, 9a was directly hydrolyzed to get the
target compound 12a, which was further reacted with halogenated
compounds to afford the compounds 13a and 13b. In a similar way,
the target compounds 9b, 10de10g, 11de11f, 12b and 13c were
synthesized (Schemes 3 and 4).
Table 2 showed that the Leave-One-Out (LOO) cross-validated q2
of the best CoMFA model was 0.543 with 8 optimal components.
The q2 value was correlated to the predictability of the 3D-QSAR
model and 0.543 (q2 > 0.5) signified a good predictive ability of the
CoMFA model. Non-cross-validated coefficient r2 of 0.998 with
standard error of 0.029 and F value of 277.689 suggested that the
linearity of the model was pretty good. The steric field descriptor
was 0.427 while the electrostatic descriptor was 0.573, which
suggested a greater contribution of electrostatic field than steric
field to PTP1B inhibitory activity in CoMFA study.
Table 2 clearly displayed that the best CoMSIA model covered
five descriptor fields: steric, electrostatic, hydrophobic, hydrogen-
bond donor, and hydrogen-bond acceptor. The relative contribu-
tion of the five fields was found to be 0.101 (steric), 0.231 (elec-
trostatic), 0.237 (hydrophobic), 0.225 (hydrogen-bond acceptor)
and 0.206 (hydrogen-bond donor), respectively, indicating that
compared with steric, the other four fields were more important to
biological activity. PLS regression of CoMSIA model showed a q2
value of 0.777 with 11 optimal components, non-cross-validated r2
value of 0.999, the standard error of estimate of 0.013 and F value of
344.740. The high q2 value represented a good predictive ability of
the CoMSIA model.
3. Results and discussion
3.1. Biological evaluation
Table 1 listed the PTP1B inhibitory activities of the 39 imida-
zolidine-2,4-dione derivatives. From Table 1, we could clearly see
that most of these molecules exhibited mild inhibitory activities
The predictive ability of the CoMFA and CoMSIA models were
assessed by the test set of 7 imidazolidine-2,4-dione derivatives,
and the results including predicted activities, experimental activ-
ities and their residues were listed in Table 3. The predicted pIC50
values were in good agreement with the experimental data within
a statistically tolerable error range. The predictive correlation co-
against human PTP1B with IC50 values at about 0.57e172
Compound 5e showed the most potent PTP1B inhibitory activity
with the IC50 value at 0.57 M.
mM.
m
We can see from Table 1, most molecules were substituted in
part R2 and R3. However, the most active molecule 5e
(R2 ¼ R3 ¼ 2,4-di-Cl-PhCH2) and 10d (R2 ¼ PhCH2, R3 ¼ 4-CH3OOC-
PhCH2) and other molecules with larger hydrophobic aromatic
efficient r2
was 0.754 and 0.836 for the CoMFA and CoMSIA
pred
models, respectively, and the correlations between the predicted
values and the experimental values for both models were depicted
in Fig. 2.
The above results demonstrated that the constructed 3D-QSAR
models were statistically significant and reliable, which could be
used to design novel imidazolidine-2,4-dione derivatives with
improved PTP1B inhibitory activity.
3.2.2. CoMFA/CoMSIA contour plots
The contour plots of 3D-QSAR analysis were shown in Figs. 3 and
4. The visualization of the results of both CoMFA and CoMSIA
models was performed using StDev*Coeff (the standard deviation
Fig. 1. Design of imidazolidine-2,4-dione derivatives as PTP1B inhibitors.

M.-Y. Wang et al. / European Journal of Medicinal Chemistry 103 (2015) 91e104
93
Scheme 1. Reagents and conditions: (a) piperidine; (b) K2CO3, acetone, ethyl bromoacetate or NaH, DMF, organochlorine compounds.
Scheme 2. Reagents and conditions: (a) K2CO3, DMF, methyl bromoacetate; (b) 4-hydroxybenzaldehyde, piperidine, HOAc, toluene; (c) K2CO3, DMF, halogenated compounds; (d)
Sodium hydroxide solution (1 mol/L), methanol; (e) Sodium hydroxide solution (1 mol/L), methanol; (f) K2CO3, DMF/acetone, halogenated compounds.
Scheme 3. Reagents and conditions: (a) K2CO3, DMF, methyl 4-(bromomethyl) benzoate; (b) 4-hydroxybenzaldehyde, piperidine, HOAc, toluene; (c) K2CO3, DMF, halogenated
compounds; (d) Sodium hydroxide solution (1 mol/L), methanol; (e) Sodium hydroxide solution (1 mol/L), methanol; (f) K2CO3, DMF, methyl bromoacetate.
and the coefficient) [33,34] mapping option contoured by contri-
bution. Favored and disfavored levels of the contour maps by
contribution were set at 80% and 20%, respectively. The most active
compound 5e was embedded into the maps to illustrate its affinity
for the steric, electrostatic, hydrophobic, hydrogen-bond acceptor
and hydrogen-bond donor regions of inhibitors.

94
M.-Y. Wang et al. / European Journal of Medicinal Chemistry 103 (2015) 91e104
Scheme 4. Reagents and conditions: (a) K2CO3, DMF, ethyl 2-bromoacetate; (b) 4-hydroxybenzaldehyde, piperidine, HOAc, toluene; (c) K2CO3, DMF, benzyl chloride.
Table 1
Structure and PTP1B inhibitory activities of imidazolidine-2,4-dione derivatives.
No.
R1
R2
R3
IC50
(
m
M)
4a
4b
4c
4d
4e
5a
5b
5c
5d
5e
5f
6a
2-Cl-PhCH2
PhCOCH2
4-CH3O-PhCH2
2,4-di-Cl-PhCH2
3,4-di-Cl-PhCH2
H
H
H
H
H
H
H
2-Cl-PhCH2
PhCOCH2
4-CH3O-PhCH2
2,4-di-Cl-PhCH2
3,4-di-Cl-PhCH2
PhCH2
CH3CH2OOCCH2
4-Cl-PhCH2
3,4-di-Cl-PhCH2
2,4-di-Cl-PhCH2
4-CH3OPhCH2
H
2-Cl-PhCH2
PhCOCH2
4-CH3O-PhCH2
2,4-di-Cl-PhCH2
3,4-di-Cl-PhCH2
PhCH2
CH3CH2OOCCH2
4-Cl-PhCH2
3,4-di-Cl-PhCH2
2,4-di-Cl-PhCH2
4-CH3OPhCH2
PhCH2
3.5
11
7.4
6.6
4.1
10
133
3.3
1.3
0.57
3.4
170
6b
6c
6d
H
H
H
H
H
H
H
H
H
H
H
H
H
H
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
PhCH2
CH3CH2OOCCH2
PhCOCH2
172
155
15
1.7
122
14
2-Cl-PhCH2
2,4-di-Cl-PhCH2
4-Cl-PhCH2
PhCH2CH2CH2
3,4-di-Cl-PhCH2
H
6e
6fa
6ga
6h
11
136
9a
9b
CH3OOCCH2
4-CH3OOCPhCH2
CH3OOCCH2
CH3OOCCH2
CH3OOCCH2
4-CH3OOCPhCH2
4-CH3OOCPhCH2
4-CH3OOCPhCH2
CH3CH2OOCCH2
HOOCCH2
H
8.1
6.8
114
110
0.96
9.7
5.1
11
103
122
113
89
103
96
142
104
5.4
10a
10b
10c
10da
10ea
10f
10ga
11a
11b
11c
11d
11e
11f
12a
12b
13aa
13b
13ca
4-CH3OOCPhCH2
CH3OOCCH2
PhCH2
4-CH3OOCPhCH2
CH3OOCCH2
PhCH2
PhCH2
4-HOOCPhCH2
HOOCCH2
PhCH2
4-HOOCPhCH2
HOOCCH2
PhCH2
H
HOOCCH2
HOOCCH2
4-HOOCPhCH2
4-HOOCPhCH2
4-HOOCPhCH2
HOOCCH2
4-HOOCPhCH2
4-CH3OOCPhCH2OOCCH2
CH3OOCCH2OOCCH2
4-CH3OOCCH2OOCPhCH2
H
PhCH2
PhCH2
PhCH2
4-CH3OOCPhCH2
CH3OOCCH2
CH3OOCCH2
101
22
a
Compounds selected for the test set.
The CoMFA contour plot representing the contribution of steric
field was depicted in Fig. 3a. The steric contours were shown in
yellow and green, with green regions indicating a steric contribu-
tion to activity, while yellow regions indicating a steric hindrance to
potency. As shown in Fig. 3a, a large green contour surrounded
para-position of the aromatic ring at the C5 position of the imida-
zolidine-2,4-dione ring, indicating that increased steric bulk was
preferred in the position to improve the activity. This could be
confirmed by the fact that compounds containing a bulkier aro-
matic group at R3 (4ae4e, 5ce5f, 6e, 10a and 10f) had higher

M.-Y. Wang et al. / European Journal of Medicinal Chemistry 103 (2015) 91e104
95
Table 2
implying that large substituents was positive around this position.
This may be a reason why compounds 4ae4e, 5ce5f, 6e, 10a and
10f showed better potency than compounds without substituents
at corresponding position (6ae6c). Some yellow regions located
near the C2, N3 and C4 positions of the imidazolidine-2,4-dione ring
indicated that large substituents stretching into the domains
bounded by four yellow contours might decrease the activity,
which was in harmony with the fact that compounds 10c, 11a, 11b
(Fig. 5a), 11c, 11d and 11f (Fig. 5b) displayed low activities. Inter-
estingly, though compound 10d was substituted by a bulky aro-
matic group (4-CH3OOCPhCH2) at R1, the steric conformation of the
large substituent avoided the yellow regions which were associated
with unfavorable steric interactions, and hence compound 10d still
exhibited satisfactory activity (Fig. 6b). That also explained why
compounds 4ae4e, 10a, 10f, 10g and 13a still showed good activ-
ities in spite of with a large R1 substituent.
The CoMFA electrostatic interactions were displayed as blue and
red contours, with the blue regions signifying areas of electropos-
itive potential while red regions indicating areas of electronegative
potential, and the results were illustrated in Fig. 3b. The electro-
static contour plots were discovered to be governed by red con-
tours, suggesting a preference for electronegative substituents. The
aromatic group of 5e at N1-position of the imidazolidine-2,4-dione
ring was found to be embedded in a large red region and the other
two red contours were observed near the benzene ring at C5 po-
sition and the nitrogen atom at N3-position of the imidazolidine-
2,4-dione nucleus, respectively. This might be a powerful expla-
nation to the high biological activity of 5e. The other three small red
areas close to para-position of the benzene ring linked to oxygen
atom of common substructure of 39 imidazolidine-2,4-dione de-
rivatives indicated that negatively charged substituents at this
position was favored. This was supported by the fact that 10a and
10d with 4-CH3OOC-PhCH2 substituent at corresponding place
possessed high potency. Also a blue contour favorable positively
charged substitution was observed above aromatic nucleus in R3
part, suggesting that electropositive substituents could be designed
at this place to improve the PTP1B inhibitory potency of imidazo-
lidine-2,4-dione derivatives.
The superiority of CoMSIA analysis is that Gaussian function is
used to define the distance dependence. The influence of various
properties contributing to the potency can be divided into spatial
locations to identify the decisive areas to bioactivity and these can
be visualized as contour maps. As a result, the CoMSIA contour
maps, derived from using steric, electrostatic, hydrophobic,
hydrogen bond donor, and hydrogen bond acceptor fields, were
displayed in Fig. 4. CoMSIA steric and electrostatic contours were
more or less similar to those of CoMFA, which were not displayed
here.
Fig. 4a represented the CoMSIA hydrophobic field contribution
and the yellow regions were associated with increased bioactivity.
A major yellow polyhedron was discovered overlapping the plane
of the 2,4-dichloro-substituted phenyl ring in part R3 of compound
5e, and another yellow region appeared around the phenyl ring in
part R2 of compound 5e. The third large yellow polyhedron was
under the imidazolidine-2,4-dione ring. Hydrophobic bulks were
preferred at the above mentioned positions, which might explain
why compound 5e possessed the most potent activity and com-
pounds with aromatic substituents at these positions showed
better activities (4ae4e, 5ce5f, 10ce10f and 13a). The correlations
were consistent with the results of steric contours.
The PLS statistic results of CoMFA and CoMSIA models.
Statistical Parameters
CoMFA
CoMSIA
Cross-validated (LOO)
q2a
0.543
8
0.777
11
ONCb
Non-cross-validated
r2c
0.998
0.029
277.689
0.754
0.999
0.013
344.740
0.836
SEEd
F valuee
f
r2
pred
Field distribution
Steric
Electrostatic
Hydrophobic
Hydrogen bond donor
Hydrogen bond acceptor
0.427
0.573
e
0.101
0.231
0.237
0.206
0.171
e
e
a
LOO-cross-validated correlation coefficient.
Optimal components.
Non-cross-validated correlation coefficient.
Standard error of estimate.
F-ratio of F-statistic for the analysis.
Predictive correlation coefficient.
b
c
d
e
f
Table 3
Experimental versus predicted activities of training and test set molecules according
to the CoMFA and CoMSIA models.
No.
Experimental activity (pIC50
)
CoMFA
Predict Residue Predict Residue
ꢀ0.517 ꢀ0.022 ꢀ0.529 ꢀ0.010
CoMSIA
4a
4b
4c
4d
4e
5a
5b
5c
5d
5e
5f
6a
ꢀ0.539
ꢀ1.045
ꢀ0.872
ꢀ0.820
ꢀ0.614
ꢀ1.006
ꢀ2.123
ꢀ0.520
ꢀ0.127
0.244
ꢀ0.528
ꢀ2.230
ꢀ2.236
ꢀ2.191
ꢀ1.182
ꢀ0.218
ꢀ2.085
ꢀ1.145
ꢀ1.042
ꢀ2.135
ꢀ0.911
ꢀ0.833
ꢀ2.057
ꢀ2.040
0.016
ꢀ1.047 0.002 ꢀ1.055 0.010
ꢀ0.869 ꢀ0.003
ꢀ0.829 0.009
ꢀ0.613 ꢀ0.001
0.011
ꢀ0.871 ꢀ0.001
ꢀ0.835
ꢀ0.630
ꢀ1.056
0.015
0.016
0.050
ꢀ1.017
ꢀ2.099 ꢀ0.024
ꢀ0.519 ꢀ0.001
ꢀ2.114 ꢀ0.009
ꢀ0.512 ꢀ0.008
ꢀ0.141
0.014
ꢀ0.138
0.011
0.294 ꢀ0.050
ꢀ0.510 ꢀ0.018
ꢀ2.214 ꢀ0.016
ꢀ2.223 ꢀ0.013
ꢀ2.174 ꢀ0.017
ꢀ1.179 ꢀ0.003
0.255 ꢀ0.011
ꢀ0.539
ꢀ2.232
ꢀ2.244
0.011
0.002
0.008
6b
6c
6d
6e
ꢀ2.190 ꢀ0.001
ꢀ1.149 ꢀ0.033
ꢀ0.265
ꢀ2.117
ꢀ1.612
0.047
0.032
0.467
ꢀ0.242
0.024
6fa
6ga
6h
9a
ꢀ2.076 ꢀ0.009
ꢀ1.505 0.360
ꢀ1.038 ꢀ0.004
ꢀ1.040 ꢀ0.002
ꢀ2.165
ꢀ0.917
ꢀ0.835
0.030
0.006
0.002
ꢀ2.148
ꢀ0.911
0.013
0.000
9b
10a
10b
10c
10da
ꢀ0.822 ꢀ0.011
ꢀ2.053 ꢀ0.004
ꢀ2.035 ꢀ0.005
0.029 ꢀ0.013
ꢀ2.042 ꢀ0.015
ꢀ2.053
0.031 ꢀ0.015
ꢀ1.547 0.559
ꢀ0.684 ꢀ0.021
ꢀ1.582 0.556
0.013
10ea ꢀ0.988
ꢀ1.362
0.374
10f
ꢀ0.705
ꢀ0.696 ꢀ0.009
10ga ꢀ1.026
ꢀ1.564
ꢀ2.016
ꢀ2.096
ꢀ2.057
ꢀ1.953
ꢀ2.014
ꢀ1.986
0.538
0.004
0.010
0.004
0.004
0.002
0.002
11a
11b
11c
11d
11e
11f
ꢀ2.012
ꢀ2.086
ꢀ2.053
ꢀ1.949
ꢀ2.012
ꢀ1.984
ꢀ2.152
ꢀ2.016
ꢀ1.969 ꢀ0.043
ꢀ2.102
ꢀ2.083
ꢀ1.986
ꢀ2.023
0.016
0.030
0.037
0.011
ꢀ1.957 ꢀ0.027
ꢀ2.114 ꢀ0.038
12a
12b
ꢀ2.134 ꢀ0.018
ꢀ2.025
ꢀ1.517
ꢀ2.017
0.009
0.785
0.014
ꢀ2.013 ꢀ0.003
13aa ꢀ0.732
ꢀ0.986
ꢀ2.004
0.254
0.001
13b
ꢀ2.003
ꢀ1.342
13ca
ꢀ1.298 ꢀ0.044
ꢀ1.132 ꢀ0.210
a
Compounds taken for the test set.
The CoMSIA hydrogen bond interactions were illustrated in
Fig. 4b. Hydrogen bond donor field was depicted by cyan and purple
contours. Cyan contours signified regions favoring hydrogen bond
donors on ligands while purple contours indicated regions where
hydrogen bond donor substituents on ligands were unfavorable for
activities than molecules without a substituent at this position (9a,
12a or 12b). Two other small green contours were found around the
phenyl ring at N1-position of the imidazolidine-2,4-dione nucleus,

96
M.-Y. Wang et al. / European Journal of Medicinal Chemistry 103 (2015) 91e104
Fig. 2. CoMFA/CoMSIA predicted versus experimental pIC50 values. (a) Fitted predictions for the training set. (b) Fitted predictions for the test set.
Fig. 3. CoMFA contour plots for steric field (a) and electrostatic field (b) with the most active compound 5e as the reference molecule. (a) Sterically favored and disfavored regions
are shown in green and yellow, respectively. (b) Electropositive favored and electronegative favored regions are shown in blue and red, respectively. (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 4. CoMSIA contour plots for hydrophobic field (a) and H-bond interaction field (b) with the most active compound 5e as the reference molecule. (a) Yellow polyhedra indicate
regions where hydrophobic substituents are favored. (b) H-bond donors on ligand favored and disfavored are illustrated by cyan contours and purple contours, respectively, and H-
bond acceptors on ligand favored are shown as magenta contours. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article.)

M.-Y. Wang et al. / European Journal of Medicinal Chemistry 103 (2015) 91e104
97
Fig. 5. CoMFA steric contours for less potent compounds: (a) compound 11b, (b) compound 11f.
Fig. 6. CoMFA steric contours for highly potent compounds: (a) compound 5e, (b) compound 10d.
bioactivity. Besides, the magenta contours indicated hydrogen bond
acceptor substituents on ligands were advantageous to activity. A
cyan polyhedron was found near the nitrogen atom of the imida-
zolidine-2,4-dione ring, implying that hydrogen bond donor on
inhibitors here was beneficial to activity. The presence of strong H-
bond donor in compound 5e was responsible for its high potency.
Also, compounds with H-bond donors near these contours could
observe a good range of activity (5ce5f and 6e). However, a purple
polyhedron appeared closely to the benzene ring of R2 part of 5e,
indicating that hydrogen bond donors were not favored here. This
was consistent with the interpretation of hydrophobic contours
that hydrophilic substituents at this position decreased the pIC50
values. On the other side, a large magenta region surrounded two
carbonyl oxygens at the C2 and C4 positions of the imidazolidine-
2,4-dione ring, indicating that strong H-bond acceptors were
crucial to the biological activity. This might powerfully support the
importance of imidazolidine-2,4-dione scaffold to the high PTP1B
affinity.
the second phosphotyrosine binding site. On the other hand, the
aryl group at R1 of compound 10d also perfectly matched with the
second binding site. This may be a reason why compounds 5e and
10d showed high potency. As a common substructure, imidazoli-
dine-2,4-dione ring of compound 5e formed intense interactions
with binding site, including hydrogen bonds with Arg254, Arg24
and Gly259, respectively, and non-bonded interactions (Fig. 7b).
However, compound 10d only formed weaker interactions (van der
Waals interactions, hydrophobic interactions, etc.). That might lead
to lower bioactivity of compound 10d than compound 5e.
4. Conclusion
In summary, a series of imidazolidine-2,4-dione derivatives was
synthesized and evaluated for the PTP1B inhibitory activities. A 3D-
QSAR study using CoMFA and CoMSIA methods was applied to
study the structure-activity relationship and the results were as
follows: (1) large groups at the C5 position of imidazolidine-2,4-
dione ring was beneficial to bioactivity; (2) a hydrophobic, elec-
tronegative and bulky substitutes at N1 position of imidazolidine-
2,4-dione nucleus was necessary to activity; (3) two carbonyl ox-
ygens at the C2 and C4 positions of the imidazolidine-2,4-dione ring
was crucial to the biological activity. The above results were
completely consistent with the subsequent docking analysis. The
combined 3D-QSAR and docking studies could provide useful in-
formation about the action mechanism between ligand and target,
and guide further modification of imidazolidine-2,4-dione
derivatives.
3.3. Docking analysis
Molecular docking could offer more insight into understanding
the protein-inhibitor interactions [35e37]. Combined with docking
results, we could further understand the important regions of
imidazolidine-2,4-dione derivatives. The most active compound 5e
was found to bind in the same pocket and in the similar confor-
mation of PTP1B receptor (PDB code: 2QBQ) that was occupied by
the co-crystallized ligand 4B3, as shown in Fig. 7, which made our
molecular docking in this paper more conceivable.
The structure of PTP1B consists of two binding sites: the cata-
lytic active site (Cys215-Arg221) and the second phosphotyrosine
binding site (Arg24, Arg254, Met258, Gly259, Gln262 and Gln266)
[38]. Fig. 7a showed that large aromatic group at C5 position of
imidazolidine-2,4-dione ring of compounds 5e and 10d entered the
deep and narrow catalytic pocket. This was consistent with the 3D-
QSAR contour that a bulky hydrophobic group was beneficial to
bioactivity at this place. On one hand, the aromatic ring at N1-po-
sition of imidazolidine-2,4-dione nucleus of compound 5e entered
5. Experimental section
5.1. Chemistry
All the reagents were purchased from commercial suppliers and
were used without further purification unless otherwise indicated.
Reaction courses were monitored by TLC on silica gel pre-coated
F254 Merck plates. 1H NMR spectra was taken on a Bruker spec-
trometer using TMS as the internal standard, and DMSO-d6 was

98
M.-Y. Wang et al. / European Journal of Medicinal Chemistry 103 (2015) 91e104
Fig. 7. Docking of compound 5e (green), 10d (yellow) and co-crystallized ligand 4B3 (cyan) with PTP1B. (a) Surface representation of PTP1B in complex with 5e, 10d and 4B3. (b)
Docking interactions of compound 5e, 10d and 4B3 with PTP1B active site. The H-bond interactions between PTP1B and 5e were shown as red dashed lines. (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of this article.)
used as the solvent. Mass spectra were measured on an Agilent
1100 series.
5.1.1.2. Ethyl 2-(5-(4-(2-ethoxy-2-oxoethoxy)benzylidene)-2,4-
dioxoimidazolidin-1-yl) acetate(5b). A yellow solid with the
following characteristics: 20.91% yield; m.p. 158e160 ꢁC; 1H NMR
(DMSO-d6, 400 Hz):
d
10.858 (s, 1H, N3H); 7.635 (d, 2H, J ¼ 9.2 Hz,
AreH); 6.984 (d, 2H, J ¼ 8.8 Hz, AreH); 6.876 (s, 1H, ]CH); 4.849 (s,
2H, AreOeCH2); 4.304 (s, 2H, N1eCH2); 4.162 (q, 4H, J ¼ 7.2 Hz,
OCH2); 1.217 (m, 6H, CH3); MS (m/z): 377.2 (M þ 1).
5.1.1. General procedure for the synthesis of 4ae4e, 5ae5f and
6ae6h
4-hydroxybenzaldehyde (50 mmol) and hydantoin (55 mmol)
were dissolved in 10 ml piperidine. The mixture was heated to
130 ꢁC for 0.5 h, and TLC analysis indicated that the reaction was
complete. Water (200 ml) was added to the mixture after it was
cooled to the room temperature. Then, stirring until the mixture
was dissolved and filtrating to remove the undissolved substance.
Subsequently, 20 ml10 N HCl was dropwise added to the filtrate and
the precipitate was filtrated, washed with water and dried to give
compound 3.
5.1.1.3. Ethyl
2-(4-((2,5-dioxoimidazolidin-4-ylidene)methyl)phe-
noxy)acetate(6b). A white solid with the following characteristics:
11.27% yield; m.p. 229e230 ꢁC; 1H NMR (DMSO-d6, 400 Hz):
d
10.727 (s, 1H, N1H); 9.923 (s, 1H, N3H); 7.526 (d, 2H, J ¼ 8.8 Hz,
AreH); 6.811 (d, 2H, J ¼ 8.4 Hz, AreH); 6.544 (s, 1H, ]CH); 4.293 (s,
2H, AreOeCH2); 4.158 (q, 4H, J ¼ 7.2 Hz, OCH2); 1.210 (m, 3H, CH3);
MS (m/z): 291.2 (M þ 1).
A mixture of 5-(4-hydroxybenzylidene)imidazolidine-2,4-dione
(3) (20 mmol), ethyl 2-bromoacetate (20 mmol), K2CO3 (20 mmol)
and acetone (120 ml) was refluxed for 24 h. After the reaction was
complete detected by TLC analysis, the mixture was filtrated to
remove the inorganic salt and concentrated under vacuo. The
resulting residue was purified by column chromatography to give
5b and 6b.
5.1.1.4. 1,3-bis(2-chlorobenzyl)-5-(4-((2-chlorobenzyl)oxy)benzyli-
dene)imidazolidine-2,4-dione(4a). A white solid with the following
characteristics: 24.94% yield; m.p. 134e136 ꢁC; 1H NMR (DMSO-d6,
400 Hz):
d 7.475e7.515 (m, 3H, AreH); 7.331e7.391 (m, 4H, AreH);
7.205 (q, 3H, J ¼ 7.2 Hz, AreH); 7.011 (t, 3H, J ¼ 8.4 Hz, AreH); 6.860
(t, 3H, J ¼ 8.8 Hz, AreH); 5.735 (s, 1H, ]CH); 5.123 (s, 2H, OCH2);
4.840 (s, 2H, N1eCH2); 4.718 (s, 2H, N3eCH2); MS (m/z): 577.0
(M þ 1).
5-(4-hydroxybenzylidene)imidazolidine-2,4-dione
(3)
(2 mmol), organochlorine compound (1.83 mmol) and NaH
(6.47 mmol) were added in 8.6 ml DMF. The mixture was heated to
70 ꢁC for 4 h, and TLC analysis showed that the reaction was
complete. The mixture was extracted by ethyl acetate, and the
combined organic layer was washed with saturated brine, dried
over anhydrous MgSO4, filtered, and concentrated. The resulting
residue was purified by silica gel column chromatography to give
the target compounds 4ae4e, 5a, 5ce5f, 6a and 6ce6h.
5.1.1.5. 5-(4-(2-oxo-2-phenylethoxy)benzylidene)-1,3-bis(2-oxo-2-
phenylethyl)imidazolidine-2,4-dione(4b). A white solid with the
following characteristics: 6.71% yield; m.p. 153e155 ꢁC; 1H NMR
(DMSO-d6, 400 Hz):
d
8.093 (t, 2H, J ¼ 7.2 Hz, AreH); 7.947 (t, 2H,
J ¼ 7.2 Hz, AreH); 7.556e7.755 (m, 9H, AreH); 7.493 (t, 2H,
J ¼ 7.6 Hz, AreH); 7.203 (d, 2H, J ¼ 8.8 Hz, AreH); 6.854 (s, 1H, ]
CH); 6.784 (d, 2H, J ¼ 8.8 Hz, AreH); 5.404 (s, 2H, COCH2); 5.243 (s,
2H, COCH2); 5.112 (d, 4H, OCH2); MS (m/z): 559.0 (M þ 1).
5.1.1.1. 5-(4-hydroxybenzylidene)imidazolidine-2,4-dione(3). A yel-
low solid with the following characteristics: 79.5% yield; m.p.
5.1.1.6. 1,3-bis(4-methoxybenzyl)-5-(4-((4-methoxybenzyl)oxy)ben-
zylidene)imidazolidine-2,4-dione(4c). A white solid with the
following characteristics: 34% yield; m.p. 129e131 ꢁC; 1H NMR
316e318 ꢁC; 1H NMR (DMSO-d6, 400 Hz):
d 11.089 (s, 1H, N1H);
10.299 (s, 1H, N3H); 9.856 (s, 1H, OH); 7.469 (d, 2H, J ¼ 8.4 Hz,
AreH); 6.767 (q, 2H, J ¼ 8.4 Hz, AreH); 6.346 (s, 1H, Ar-CH); MS (m/
z): 205.0 (M þ 1).
(DMSO-d6, 400 Hz):
d
7.384 (d, 2H, J ¼ 8.8 Hz, AreH); 7.212 (t, 4H,
J ¼ 8.8 Hz, AreH); 6.943 (m, 6H, AreH); 6.722 (s, 1H, ]CH); 6.619

M.-Y. Wang et al. / European Journal of Medicinal Chemistry 103 (2015) 91e104
99
(d, 2H, J ¼ 8.4 Hz, AreH); 6.451 (d, 2H, J ¼ 8.4 Hz, AreH); 5.074 (s,
2H, OCH2); 4.654 (d, 4H, J ¼ 6.8 Hz, NCH2); 3.732 (s, 6H, OCH3);
3.643 (s, 3H, OCH3); MS (m/z): 565.2 (M þ 1).
5.1.1.15. 5-(4-((2-chlorobenzyl)oxy)benzylidene)imidazolidine-2,4-
dione(6d). A white solid with the following characteristics: 12%
yield; m.p. 265e268 ꢁC; 1H NMR (DMSO-d6, 400 Hz):
d 11.135 (s, 1H,
N3H); 10.428 (s, 1H, N1H); 7.376e7.604 (m, 6H, AreH); 7.046 (d, 2H,
AreH); 6.381 (s, 1H, ]CH); 5.201 (s, 2H, OCH2); MS (m/z): 329.0
(M þ 1).
5.1.1.7. 1,3-bis(2,4-dichlorobenzyl)-5-(4-((2,4-dichlorobenzyl)oxy)
benzylidene)imidazolidine-2,4-dione(4d). A white solid with the
following characteristics: 18.86% yield; m.p. 168e169 ꢁC; 1H NMR
(DMSO-d6,400 Hz):
2H, AreOeCH2); 4.769 (s, 2H, N1eCH2); 4.630 (s, 2H, N3eCH2); MS
(m/z): 682.9 (M þ 1).
d 6.769e7.673 (m, 14H, AreH, ]CH); 5.094 (s,
5.1.1.16. 5-(4-((2,4-dichlorobenzyl)oxy)benzylidene)imidazolidine-
2,4-dione(6e). A yellow solid with the following characteristics:
12% yield; m.p. 241e242 ꢁC; 1H NMR (DMSO-d6, 400 Hz):
d 10.719
(s,1H, N3H); 9.905 (s,1H, N1H); 7.648 (d,1H, J ¼ 2.0 Hz, AreH); 7.518
(t, 2H, AreH); 7.400 (q, 1H, J ¼ 6.0 Hz, AreH); 7.248 (d, 1H,
J ¼ 6.0 Hz, AreH); 6.797 (d, 2H, J ¼ 8.4 Hz, AreH); 6.523 (s, 1H, ]
CH); 4.706 (s, 2H, OCH2); MS (m/z): 362.0 (Mꢀ1).
5.1.1.8. 1,3-bis(3,4-dichlorobenzyl)-5-(4-((3,4-dichlorobenzyl)oxy)
benzylidene)imidazolidine-2,4-dione(4e). A pale yellow solid with
the following characteristics: 23.7% yield; m.p. 139e140 ꢁC; 1H
NMR (DMSO-d6, 400 Hz):
d 7.596e7.719 (m, 4H, AreH); 7.452 (q,
1H, J ¼ 6.4 Hz, AreH); 7.320 (d, 2H, J ¼ 8.0 Hz, AreH); 7.104 (d, 2H,
J ¼ 8.4 Hz, AreH); 6.922 (d, 2H, J ¼ 8.8 Hz, AreH); 6.753 (s, 2H,
AreH); 5.738 (s, 1H, ]CH); 5.149 (s, 2H, NCH2); 4.743 (s, 2H, NCH2);
4.699 (s, 2H, OCH2); MS (m/z): 696.0 (M þ NH4).
5.1.1.17. 5-(4-((4-chlorobenzyl)oxy)benzylidene)imidazolidine-2,4-
dione(6f). A white solid with the following characteristics: 14%
yield; m.p. 255e258 ꢁC; 1H NMR (DMSO-d6, 400 Hz):
d 10.654 (s,
1H, N3H); 9.895 (s, 1H, N1H); 7.499 (s, 2H, J ¼ 8.8 Hz, AreH);
7.381e7.415 (m, 2H, AreH); 7.303 (d, 2H, J ¼ 8.4 Hz, AreH); 6.786
(d, 2H, J ¼ 8.8 Hz, AreH); 6.502 (s, 1H, ]CH); 4.641 (s, 2H, OCH2);
MS (m/z): 329.0 (M þ 1).
5.1.1.9. 1-benzyl-5-(4-(benzyloxy)benzylidene)imidazolidine-2,4-
dione(5a). A white solid with the following characteristics: 26%
yield; m.p. 231e233 ꢁC; 1H NMR (DMSO-d6, 400 Hz):
d 10.740 (s,1H,
N3H); 7.614 (d, 2H, J ¼ 8.8 Hz, AreH); 7.449 (d, 2H, J ¼ 7.2 Hz, AreH);
7.379 (q, 2H, J ¼ 7.6 Hz, AreH); 7.270e7.342 (m, 6H, AreH); 7.043 (d,
2H, J ¼ 8.8 Hz, AreH); 6.540 (s, 1H, ]CH); 5.145 (d, 2H, J ¼ 8.4 Hz,
OCH2); 4.654 (s, 2H, CH2); MS (m/z): 383.0 (Mꢀ1).
5.1.1.18. 5-(4-(3-phenylpropoxy)benzylidene)imidazolidine-2,4-
dione(6g). A pale yellow solid with the following characteristics:
12% yield; m.p. 246e249 ꢁC; 1H NMR (DMSO-d6, 400 Hz):
d 10.519
(s, 1H, N3H); 9.864 (s, 1H, N1H); 7.485 (d, 2H, J ¼ 8.4 Hz, AreH);
7.143e7.284 (m, 5H, AreH); 6.773 (t, 2H, J ¼ 7.2 Hz, AreH); 6.453 (s,
1H, ]CH); 3.486 (t, 2H, J ¼ 7.2 Hz, OCH2); 2.586 (t, 2H, J ¼ 7.2 Hz,
PhCH2); 1.823e1.896 (m, 2H, PhCH2CH2); MS (m/z): 323 (M þ 1).
5.1.1.10. 1-(4-chlorobenzyl)-5-(4-((4-chlorobenzyl)oxy)benzylidene)
imidazolidine-2,4-dione(5c). A pale yellow solid with the following
characteristics: 27% yield; m.p. 167e168 ꢁC; 1H NMR (DMSO-d6,
400 Hz):
d 9.867 (s, 1H, N3H); 7.310e7.434 (m, 9H, AreH);
6.680e6.734 (m, 2H, AreH); 6.429 (s, 1H, ]CH); 4.938 (s, 2H,
ArCH2); 4.720 (t, 2H, J ¼ 6.4 Hz, OCH2); MS (m/z): 452.0 (Mꢀ1).
5.1.1.19. 5-(4-((3,4-dichlorobenzyl)oxy)benzylidene)imidazolidine-
2,4-dione(6h). A yellow solid with the following characteristics: 9%
yield; m.p. 236e238 ꢁC; 1H NMR (DMSO-d6, 400 Hz):
d 10.671 (s,
5.1.1.11. 1-(3,4-dichlorobenzyl)-5-(4-((3,4-dichlorobenzyl)oxy)benzy-
lidene)imidazolidine-2,4-dione(5d). A white solid with the
following characteristics: 49% yield; m.p. 242e244 ꢁC; 1H NMR
1H, N3H); 9.909 (s, 1H, N1H); 7.576 (q, 2H, J ¼ 7.6 Hz, AreH); 7.511
(d, 2H, J ¼ 8.4 Hz, AreH); 7.269 (q, 1H, J ¼ 6.0 Hz, AreH); 6.790 (d,
1H, J ¼ 8.4 Hz, AreH); 6.511 (s, 1H, ]CH); 4.658 (s, 2H, OCH2); MS
(m/z): 360.9 (Mꢀ1).
(DMSO-d6, 400 Hz):
d
9.876 (s, 1H, N3H); 7.841 (d, 2H, J ¼ 8.4 Hz,
AreH); 7.595e7.660 (m, 4H, AreH); 7.305e7.358 (m, 2H, AreH);
6.730 (d, 2H, J ¼ 8.8 Hz, AreH); 6.426 (s, 1H, ]CH); 4.945 (s, 2H,
AreOeCH2); 4.725 (s, 2H, N1eCH2); MS (m/z): 521.0 (Mꢀ1).
5.1.2. General procedure for the synthesis of 9a and 9b
A mixture of benzyl hydantoin (4 mmol), methyl bromoacetate
(or methyl 4-(bromomethyl) benzoate) (4 mmol) and K2CO3
(4 mmol) in 30 ml DMF was heated to 85 ꢁC for 3 h. After
completion of the reaction as monitored by TLC, The mixture was
filtrated to remove the solid. The filtrate was concentrated under
vacuo, extracted by ethyl acetate and dried over anhydrous MgSO4.
The resulting residue was purified using silica gel column chro-
matography with petroleum ether: ethyl acetate ¼ 8:1e4:1 to give
methyl 2-(3-benzyl-2,5-dioxoimidazolidin-1-yl)acetate (8a) (or
methyl 4-((3-benzyl-2,5-dioxoimidazolidin-1-yl)methyl)benzoate
(8b)). Subsequently, the mixture of compound 8a (or 8b)
(3.2 mmol) and 4-hydroxybenzaldehyde (3.2 mmol) in 25 ml of
toluene was treated with 2 ml piperidine and 2 ml acetic acid. The
reaction mixture was heated at reflux for 4.5 h and then concen-
trated via rotary evaporator to afford the crude product. The
resulting residue was purified by silica gel column chromatography
to give 9a (or 9b).
5.1.1.12. 1-(2,4-dichlorobenzyl)-5-(4-((2,4-dichlorobenzyl)oxy)ben-
zylidene)imidazolidine-2,4-dione(5e). A pale yellow solid with the
following characteristics: 38% yield; m.p. 261e262 ꢁC; 1H NMR
(DMSO-d6, 400 Hz):
d
9.888 (s, 1H, N3H); 7.840 (d, 2H, J ¼ 8.8 Hz,
AreH); 7.669 (q, 2H, J ¼ 12.0 Hz, AreH); 7.308e7.429 (m, 4H,
AreH); 6.724 (d, 2H, J ¼ 8.8 Hz, AreH); 6.361 (s, 1H, ]CH); 4.969 (s,
2H, AreOeCH2); 4.777 (s, 2H, N1eCH2); MS (m/z): 519.0 (Mꢀ1).
5.1.1.13. 5-(4-(benzyloxy)benzylidene)imidazolidine-2,4-dione(6a).
A pale gray solid with the following characteristics: 9% yield; m.p.
285e288 ꢁC; 1H NMR (DMSO-d6, 400 Hz):
d 10.638 (s, 1H, N3H);
9.886 (s, 1H, N1H); 7.500 (d, 2H, J ¼ 8.8 Hz, AreH); 7.267e7.353 (q,
4H, AreH); 6.789 (d, 2H, J ¼ 8.8 Hz, AreH); 6.505 (s, 1H, ]CH);
4.651 (s, 2H, CH2); MS (m/z): 293.0 (Mꢀ1).
5.1.1.14. 5-(4-(2-oxo-2-phenylethoxy)benzylidene)imidazolidine-2,4-
dione(6c). A yellow solid with the following characteristics: 12%
yield; m.p. 251e252 ꢁC; 1H NMR (DMSO-d6, 400 Hz):
d
10.730 (s,
5.1.2.1. Methyl 2-(3-benzyl-2,5-dioxoimidazolidin-1-yl)acetate(8a).
1H, N3H); 9.919 (s, 1H, N1H); 8.068 (d, 2H, J ¼ 7.2 Hz, AreH); 7.724
(t, 1H, J ¼ 7.6 Hz, AreH); 7.528e7.606 (m, 4H, AreH); 6.813 (d, 2H,
J ¼ 8.8 Hz, AreH); 6.548 (s, 1H, ]CH); 5.096 (s, 2H, OCH2); MS(m/
z): 323.2 (M þ 1).
A colorless oil with the following characteristics: 77.41% yield; 1H
NMR (DMSO-d6, 400 MHz):
d 7.379e7.268 (5H, m, AreH); 4.534
(2H, s, NeCH2); 4.246 (2H, s, NeCH2); 4.018 (2H, s, NeCH2); 3.691
(3H, s, OCH3); MS (m/z): 280.1 (M þ NH4).

100
M.-Y. Wang et al. / European Journal of Medicinal Chemistry 103 (2015) 91e104
5.1.2.2. Methyl 4-((3-benzyl-2,5-dioxoimidazolidin-1-yl)methyl)ben-
75.83% yield; m.p. 175.6e176.4 ꢁC; 1H NMR (DMSO-d6, 400 MHz):
zoate(8b). A colorless oil with the following characteristics: 80.25%
d
8.000 (2H, d, J ¼ 8.00 Hz, AreH); 7.882 (2H, d, J ¼ 8.80 Hz, AreH);
yield; 1H NMR (DMSO-d6, 400 MHz):
d
7.941 (2H, t, J ¼ 8.40 Hz,
7.511 (2H, d, J ¼ 8.00 Hz, AreH); 7.470 (2H, d, J ¼ 8.00 Hz, AreH);
7.364 (2H, d, J ¼ 7.20 Hz, AreH); 7.333e7.460 (5H, m, AreH); 6.911
(2H, d, J ¼ 8.00 Hz, AreH); 6.189 (1H, s, ]CH); 5.126 (2H, s, NeCH2);
AreH); 7.721 (2H, t, J ¼ 8.00 Hz, AreH); 7.388e7.684 (5H, m, AreH);
4.467 (2H, s, NeCH2); 4.529 (2H, s, NeCH2); 3.992 (2H, s, NeCH2);
3.885 (3H, s, OCH3); MS (m/z): 356.1 (M þ NH4).
d
4.927 (2H, s, NeCH2); 4.861 (2H, s, OeCH2); 3.926 (3H, s, OCH3);
3.906 (3H, s, OCH3); MS (m/z): 608.3 (M þ NH4).
5.1.2.3. (E)-Methyl 2-(3-benzyl-4-(4-hydroxybenzylidene)-2,5-
dioxoimidazolidin-1-yl)acetate(9a). A yellow solid with the
following characteristics: 27.04% yield; m.p. 205.3e205.8 ꢁC; 1H
5.1.3.5. (E)-Methyl 4-((3-benzyl-4-(4-(2-methoxy-2-oxoethoxy)ben-
zylidene)-2,5-dioxoimidazolidin-1-yl)methyl)benzoate(10e). A white
solid with the following characteristics: 66.87% yield; m.p.
NMR (DMSO-d6, 400 MHz):
d 9.956 (1H, s, AreOH); 7.834 (2H, d,
J ¼ 8.64 Hz, AreH); 7.392e7.281 (5H, m, AreH); 6.746 (2H, d,
J ¼ 8.64 Hz, AreH); 6.515 (1H, s, ]CH); 4.988 (2H, s, NeCH2); 4.411
(2H, s, NeCH2); 3.716 (3H, s, OCH3); MS (m/z): 367.1 (M þ 1).
161.7e162.6 ꢁC; 1H NMR (DMSO-d6, 400 MHz):
d 7.610(2H, d,
J ¼ 8.00 Hz, AreH); 7.352 (2H, d, J ¼ 1.60 Hz, AreH); 7.326 (2H, d,
J ¼ 7.20 Hz, AreH); 7.255e7.290 (5H, m, AreH); 6.855 (2H, d,
J ¼ 8.80 Hz, AreH); 6.182 (1H, s, ]CH); 4.927 (2H, s, NeCH2); 4.861
(2H, s, NeCH2); 4.633 (2H, s, CH2); 3.905 (3H, s, OCH3); 3.712 (3H, s,
OCH3); MS (m/z): 532.2 (M þ NH4).
5.1.2.4. (E)-Methyl 4-((3-benzyl-4-(4-hydroxybenzylidene)-2,5-
dioxoimidazolidin-1-yl)methyl)benzoate(9b). A yellow solid with
the following characteristics: 25.28% yield; m.p. 197.5e197.9 ꢁC; 1H
NMR (DMSO-d6, 400 MHz):
d
9.913 (1H, s, AreOH); 7.955 (2H, d,
5.1.3.6. (E)-Methyl 4-((3-benzyl-4-(4-(benzyloxy)benzylidene)-2,5-
dioxoimidazolidin-1-yl)methyl)benzoate(10f). A white solid with
the following characteristics: 82.74% yield; m.p. 154.4e154.9 ꢁC; 1H
J ¼ 8.00 Hz, AreH); 7.836 (2H, d, J ¼ 8.80 Hz, AreH); 7.356e7.477
(5H, m, AreH); 7.293 (2H, d, J ¼ 8.00 Hz, AreH); 6.731 (2H, d,
J ¼ 8.40 Hz, AreH); 6.469 (1H, s, ]CH); 4.970 (2H, s, NeCH2); 4.817
(2H, s, NeCH2); 3.847 (3H, s, OCH3); MS (m/z): 460.2 (M þ NH4).
NMR (DMSO-d6, 400 MHz):
d
7.427 (2H, d, J ¼ 10.0 Hz, AreH); 7.400
(2H, d, J ¼ 8.40 Hz, AreH); 7.343e7.381 (5H, m, AreH); 7.277e7.340
(5H, m, AreH); 7.340 (2H, d, J ¼ 4.28 Hz, AreH); 6.517 (1H, s, ]CH);
5.127 (2H, s, NeCH2); 4.972 (2H, s, NeCH2); 4.813 (2H, s, CH2); 3.846
(3H, s, OCH3); MS (m/z): 550.2 (M þ NH4).
5.1.3. General procedure for the synthesis of 10a-10f
A mixture of 9a (or 9b) (2 mmol), methyl 4-(bromomethyl)
benzoate (or methyl bromoacetate or benzyl chloride) (2 mmol)
and K2CO3 (2 mmol) in 50 ml DMF was stirred at 85 ꢁC for 3 h, and
TLC analysis indicated the reaction was complete. Then, the mixture
was filtrated and the filtrate was concentrated under vacuo,
extracted by ethyl acetate and dried over anhydrous MgSO4. The
resulting residue was purified by silica gel column chromatography
with petroleum ether: ethyl acetate ¼ 8:1e4:1 to give 10ae10c (or
10de10f).
5.1.4. General procedure for the synthesis of 11a-11f
A mixture of 10a-10f (1 mmol), 15 ml NaOH (1 mol/L) and 15 ml
methanol was refluxed for 3.5 h. After completion of the reaction as
monitored by TLC, the pH of the mixture was regulated to 3e5 by
hydrochloric acid (1 mol/L). Then, the mixture was extracted by
ethyl acetate, dried over anhydrous MgSO4 and evaporated to give
11a-11f.
5.1.3.1. (E)-Methyl 4-((4-((3-benzyl-1-(2-methoxy-2-oxoethyl)-2,5-
dioxoimidazolidin-4-ylidene)methyl)phenoxy)methyl)benzoate(10a).
A white solid with the following characteristics: 70.42% yield; m.p.
5.1. 4.1. (E)-4-((4-((3-benzyl-1-(carboxymethyl)-2, 5-
dioxoimidazolidin-4-ylidene)methyl)phenoxy)methyl)benzoic
acid(11a). A white solid with the following characteristics: 23.05%
145.5e146.2 ꢁC; 1H NMR (DMSO-d6, 400 MHz):
d
7.930 (4H, t,
yield; m.p. 155.6e156.3 ꢁC; 1H NMR (DMSO-d6, 400 MHz):
d 12.632
J ¼ 9.44 Hz, AreH); 7.583 (4H, d, J ¼ 7.84 Hz, AreH); 7.372e7.011
(5H, m, AreH); 6.571 (1H, s, ]CH); 5.241 (2H, s, NeCH2); 4.993 (2H,
s, CH2); 4.408 (2H, s, NeCH2); 3.852 (3H, s, OCH3); 3.714 (3H, s,
OCH3); MS (m/z): 532.2 (M þ NH4).
(2H, s, COOH); 7.975 (2H, d, J ¼ 7.96 Hz, AreH); 7.570e7.506 (5H, m,
AreH); 7.090 (4H, t, J ¼ 7.46 Hz, AreH); 6.910 (2H, d, J ¼ 8.52 Hz,
AreH); 6.625 (1H, s, ]CH); 5.229 (2H, s, NeCH2); 4.594 (2H, s,
CH2); 4.323 (2H, s, NeCH2); MS (m/z): 505.0 (M þ NH4).
5.1.3.2. (E)-Methyl 2-(4-((3-benzyl-1-(2-methoxy-2-oxoethyl)-2,5-
dioxoimidazolidin-4-ylidene)methyl)phenoxy)acetate(10b). A white
solid with the following characteristics: 69.33% yield; m.p.
5 .1. 4 . 2 . ( E ) - 2 - ( 4 - ( ( 3 - b e n z yl - 1 - ( c a r b o x y m e t h yl ) - 2 , 5 -
dioxoimidazolidin-4-ylidene)methyl)phenoxy)acetic
A white solid with the following characteristics: 20.43% yield; m.p.
132.6e133.2 ꢁC; 1H NMR (DMSO-d6, 400 MHz):
12.632 (2H, s,
acid(11b).
136.9e137.4 ꢁC; 1H NMR (DMSO-d6, 400 MHz):
d
7.888 (2H, d,
d
J ¼ 8.44 Hz, AreH); 7.393e7.283 (5H, m, AreH); 6.928 (2H, d,
J ¼ 8.44 Hz, AreH); 6.5739 (1H, s, ]CH); 4.994 (2H, s, NeCH2);
4.836 (2H, s, CH2); 4.408 (2H, s, NeCH2); 3.7141 (3H, s, OCH3);
3.6901 (3H, s, OCH3); MS (m/z): 360.9 (Mꢀ1).
COOH); 7.509 (2H, d, J ¼ 8.44 Hz. AreH); 7.098 (5H, m, AreH); 6.800
(2H, d, J ¼ 8.60 Hz, AreH); 6.628 (1H, s, ]CH); 4.695 (2H, s,
NeCH2); 4.570 (2H, s, NeCH2); 4.350e4.001 (2H, m, OeCH2); MS
(m/z): 409.1 (Mꢀ1).
5.1.3.3. (E)-Methyl 2-(3-benzyl-4-(4-(benzyloxy)benzylidene)-2,5-
dioxoimidazolidin-1-yl)acetate(10c). A yellow solid with the
following characteristics: 79.86% yield; m.p. 112.7e113.9 ꢁC; 1H
5.1.4.3. (E)-2-(3-benzyl-4-(4-(benzyloxy)benzylidene)-2,5-
dioxoimidazolidin-1-yl)acetic acid(11c). A white solid with the
following characteristics: 61.32% yield; m.p. 173.9e174.3 ꢁC; 1H
NMR (DMSO-d6, 400 MHz):
d
7.404 (2H, d, J ¼ 0.80 Hz, AreH);
NMR (DMSO-d6, 400 MHz): d 13.249 (1H, s, COOH), 7.910 (2H, d,
7.359e7.385 (5H, m, AreH); 7.325e7.360 (5H, m, AreH); 7.010 (2H,
d, J ¼ 8.80 Hz, AreH); 6.565 (1H, s, ]CH); 5.131 (2H, s, NeCH2);
4.994 (2H, s, OeCH2); 4.413 (2H, s, NeCH2); 3.371 (3H, s, OCH3); MS
(m/z): 474.2 (M þ NH4).
J ¼ 8.84 Hz, AreH); 7.367e7.451 (5H, m, AreH); 7.280e7.343 (5H,
m, AreH); 7.108 (2H, d, J ¼ 8.80 Hz, AreH); 6.546 (1H, s, ]CH);
5.132 (2H, s, NeCH2); 4.987 (2H, s, OeCH2); 4.272 (2H, s, NeCH2);
MS (m/z): 460.2 (M þ NH4).
5.1.3.4. (E)-Methyl 4-((4-((3-benzyl-1-(4-(methoxycarbonyl)benzyl)-
2,5-dioxoimidazolidin-4-ylidene)methyl)phenoxy)methyl)benzoa-
te(10d). A pale yellow solid with the following characteristics:
5.1.4.4. (E)-4-((4-((3-benzyl-1-(4-carboxybenzyl)-2,5-
dioxoimidazolidin-4-ylidene)methyl)phenoxy)methyl)benzoic
acid(11d). A white solid with the following characteristics: 27.45%

M.-Y. Wang et al. / European Journal of Medicinal Chemistry 103 (2015) 91e104
101
yield; m.p. 184.3e185.0 ꢁC; 1H NMR (DMSO-d6, 400 MHz):
d
12.819
NMR (DMSO-d6, 400 MHz):
d
7.778 (2H, t, J ¼ 8.80 Hz, AreH);
(2H, s, COOH); 7.977 (2H, d, J ¼ 8.40 Hz, AreH); 7.957 (2H, d,
J ¼ 6.00 Hz, AreH); 7.885 (2H, d, J ¼ 8.80 Hz, AreH); 7.762 (2H, d,
J ¼ 8.00 Hz, AreH); 7.580 (2H, d, J ¼ 8.00 Hz, AreH); 7.188e7.432
(5H, m, AreH); 7.187 (2H, d, J ¼ 8.00 Hz, AreH); 6.189 (1H, s, ]CH);
5.335 (2H, s, NeCH2); 5.216 (2H, s, NeCH2); 5.181 (2H, s, OeCH2).
d
7.777 (2H, t, J ¼ 8.80 Hz, AreH); 7.476 (2H, d, J ¼ 8.40 Hz, AreH);
7.410 (2H, d, J ¼ 8.00 Hz, AreH); 7.410 (2H, d, J ¼ 8.00 Hz, AreH);
7.260e7.332 (5H, m, AreH); 8.901 (2H, d, J ¼ 9.20 Hz, AreH); 6.207
(1H, s, ]CH); 5.264 (2H, s, NeCH2); 5.130 (2H, s, CH2); 4.952 (2H, s,
NeCH2); 4.497 (2H, s, CH2); 3.922 (3H, s, OCH3); 3.917 (3H, s, OCH3);
MS (m/z): 666.2 (M þ NH4).
5.1.4.5. (E)-4-((3-benzyl-4-(4-(carboxymethoxy)benzylidene)-2,5-
dioxoimidazolidin-1-yl)methyl)benzoic acid(11e). A white solid with
the following characteristics: 35.47% yield; m.p. 144.6e145.3 ꢁC; 1H
5.1.6.2. (E)-2-Methoxy-2-oxoethyl 2-(3-benzyl-4-(4-(2-methoxy-2-
oxoethoxy)benzylidene)-2,5-dioxoimidazolidin-1-yl)acetate(13b).
A white solid with the following characteristics: 68.91% yield; m.p.
NMR (DMSO-d6, 400 MHz):
d 12.985 (2H, s, COOH); 7.926 (2H, d,
J ¼ 8.40 Hz, AreH); 7.894 (2H, d, J ¼ 8.80 Hz, AreH); 7.410 (2H, d,
J ¼ 8.40 Hz, AreH); 7.362e7.372 (5H, m, AreH); 6.892 (2H, d,
J ¼ 8.80 Hz, AreH); 6.524 (1H, s, ]CH); 4.974 (2H, s, NeCH2); 4.805
(2H, s, NeCH2); 4.699 (2H, s, CH2); MS (m/z): 504.1 (M þ NH4).
96.1e96.8 ꢁC; 1H NMR (DMSO-d6, 400 MHz):
d 7.893 (2H, d,
J ¼ 8.40 Hz, AreH); 7.283e7.356 (5H, m, AreH); 6.933 (2H, d,
J ¼ 8.40 Hz, AreH); 6.578 (1H, s, ]CH); 4.999 (2H, s, NeCH2); 4.838
(2H, s, OCH2); 4.834 (2H, s, OCH2); 4.505 (2H, s, NeCH2); 3.690 (3H,
s, OCH3); 3.682 (3H, s, OCH3); MS (m/z): 514.2 (M þ NH4).
5.1.4.6. (E)-4-((3-benzyl-4-(4-(benzyloxy)benzylidene)-2,5-
dioxoimidazolidin-1-yl)methyl)benzoic acid(11f). A white solid with
the following characteristics: 65.65% yield; m.p. 173.7e173.9 ꢁC; 1H
5.1.6.3. (E)-2-Methoxy-2-oxoethyl 4-((3-benzyl-4-(4-(2-methoxy-2-
oxoethoxy)benzylidene)-2,5-dioxoimidazolidin-1-yl)methyl)benzoa-
te(13c). A white solid with the following characteristics: 72.34%
NMR (DMSO-d6, 400 MHz):
d 12.742 (1H, s, COOH), 7.762 (2H, d,
J ¼ 8.00 Hz, AreH); 7.486 (2H, d, J ¼ 9.60 Hz, AreH); 7.423e7.473
(5H, m, AreH); 7.342e7.409 (5H, m, AreH); 7.197 (2H, d, J ¼ 4.4
0 Hz, AreH); 6.899 (1H, s, ]CH); 5.122 (2H, s, NeCH2); 4.729 (2H, s,
NeCH2); 4.339 (2H, s, CH2); MS (m/z): 537.2 (M þ NH4).
yield; m.p. 154.6e155.3 ꢁC; 1H NMR (DMSO-d6, 400 MHz):
d 7.298
(2H, d, J ¼ 6.00 Hz, AreH); 7.268 (2H, d, J ¼ 7.80 Hz, AreH); 7.111
(2H, d, J ¼ 8.00 Hz, AreH); 6.839e7.864 (5H, m, AreH); 6.578 (2H, d,
J ¼ 6.80 Hz, AreH); 6.187 (1H, s, ]CH); 4.938 (2H, s, OeCH2); 4.862
(2H, s, NeCH2); 4.768 (2H, s, NeCH2); 4.633 (2H, s, OeCH2); 3.793
(3H, s, OCH3); 3.785 (3H, s, OCH3); MS (m/z): 590.1 (M þ NH4).
5.1.5. General procedure for the synthesis of 12a and 12b
A mixture of 9a (or 9b) (1 mmol), 15 ml NaOH (1 mol/L) and
15 ml methanol was heated to 60 ꢁC for 3.5 h. After completion of
the reaction as monitored by TLC, the pH of the mixture was
regulated to 3e5 by hydrochloric acid (1 mol/L). The solid obtained
(12a or 12b) was filtered.
5.1.7. General procedure for the synthesis of 10g
According to the synthesis of 10a-10f, a mixture of benzyl
hydantoin (4 mmol), ethyl 2-bromoacetate (4 mmol) and K2CO3
(4 mmol) in 30 ml DMF was heated to 85 ꢁC for 3 h to produce ethyl
2-(3-benzyl-2,5-dioxoimidazolidin-1-yl)acetate (8c). Then, the
compound 8c was used to react with 4-hydroxybenzaldehyde to
5.1.5.1. (E)-2-(3-benzyl-4-(4-hydroxybenzylidene)-2,5-
dioxoimidazolidin-1-yl)acetic acid(12a). A yellow solid with the
following characteristics: 68.32% yield; m.p. 192.5e193.4 ꢁC; 1H
afford
ethyl
2-(3-benzyl-4-(4-hydroxybenzylidene)-2,5-
dioxoimidazolidin-1-yl)acetate (9c), which reacted with benzyl
chloride to afford the target compound 10g.
NMR (DMSO-d6, 400 MHz):
d 13.234 (1H, s, COOH); 9.894 (1H, s,
AreOH); 7.823 (2H, d, J ¼ 8.72 Hz, AreH); 7.362e7.276 (5H, m,
AreH); 6.737 (2H, d, J ¼ 8.72 Hz, AreH); 6.485 (1H, s, ]CH); 4.978
(2H, s, NeCH2); 4.268 (2H, s, NeCH2); MS (m/z): 370.2 (M þ NH4).
5.1.7.1. (E)-Ethyl 2-(3-benzyl-4-(4-(benzyloxy)benzylidene)-2,5-
dioxoimidazolidin-1-yl)acetate(10g). A white solid with the
following characteristics: 84.21% yield; m.p. 107.2e107.9 ꢁC; 1H
5.1.5.2. (E)-4-((3-benzyl-4-(4-hydroxybenzylidene)-2,5-
dioxoimidazolidin-1-yl)methyl)benzoic acid(12b). A yellow solid
with the following characteristics: 62.37% yield; m.p.
NMR (DMSO-d6, 400 MHz):
d
7.906 (2H, d, J ¼ 8.80 Hz, AreH);
7.386e7.450 (5H, m, AreH); 7.282e7.335 (5H, m, AreH); 7.009 (2H,
d, J ¼ 8.8 0 Hz, AreH); 6.565 (1H, s, ]CH); 5.133 (2H, s, NeCH2);
4.994 (2H, s, OeCH2); 4.384 (2H, s, NeCH2); 4.173 (2H, q,
J ¼ 6.80 Hz, OCH2e); 1.216 (3H, t, J ¼ 7.20 Hz, eCH3); MS (m/z):
488.3 (M þ NH4).
197.9e198.7 ꢁC; 1H NMR (DMSO-d6, 400 MHz):
d 12.833 (1H, s,
COOH);
d
10.022 (1H, s, AreOH); 7.932 (2H, d, J ¼ 8.00 Hz, AreH);
7.835 (2H, d, J ¼ 8.40 Hz, AreH); 7.755 (2H, d, J ¼ 8.00 Hz, AreH);
7.356e7.465 (5H, m, AreH); 6.731 (2H, d, J ¼ 8.40 Hz, AreH); 6.653
(1H, s, ]CH); 4.970 (2H, s, NeCH2); 4.808 (2H, s, NeCH2); MS (m/z):
446.1 (M þ NH4).
5.2. Biological assay
The PTP1B activity was assayed using pNPP (Para-Nitro Phenyl
Phosphate) as the substrate in 96-well plate. pNPP can be hy-
drolyzed by PTP1B to give para-nitrophenol. Para-nitrophenol will
convert into para-nitrophenolate (pNP), which is an intense yel-
low compound and can be measured with a spectrophotometer
after stopping solution was added. The regents used in this bio-
logical assay were as follows: human recombinant PTP1B,
expressed in E coli and purified by Ni-NTA affinity chromatog-
raphy in our lab; buffer, consisting of 50 mM citrate (pH 6.0), 0.1 M
NaCl, 1 mM EDTA and 1 mM dithiothreitol (DTT); substrate, 2 mM
pNPP in buffer; stopping solution, 0.2 M sodium hydroxide. PTP1B
and tested compounds with multifarious concentrations mixed in
5.1.6. General procedure for the synthesis of 13ae13c
12a (or 12b) (1 mmol), methyl 4-(bromomethyl) benzoate (or
methyl bromoacetate) (4 mmol) and K2CO3 (4 mmol) were dis-
solved in 50 ml DMF. The mixture was stirred at 75 ꢁC for 3 h, and
TLC examination showed the reaction was complete. Then, the
mixture was concentrated, extracted by distilled water, ethyl ace-
tate and saturated brine, and dried over anhydrous MgSO4. The
resulting residue was purified by column chromatography with
petroleum ether: ethyl acetate ¼ 8:1e4:1 to give 13a or 13b (or
13c).
5.1.6.1. (E)-Methyl
4-((4-((3-benzyl-1-(2-((4-(methoxycarbonyl)
50
replaced by an equivalent volume of buffer in blank. After pre-
incubation for 15 min at room temperature, 50 L substrate was
added and incubated at 37 ꢁC for 30 min, then stopping the
mL buffer were added to each well of a 96-well plate. PTP1B was
benzyl)oxy)-2-oxoethyl)-2,5-dioxoimidazolidin-4-ylidene)methyl)
phenoxy)methyl)benzoate(13a). A pale yellow solid with the
following characteristics: 59.21% yield; m.p. 179.8e180.6 ꢁC; 1H
m

102
M.-Y. Wang et al. / European Journal of Medicinal Chemistry 103 (2015) 91e104
reaction by adding 10
m
L 0.2 M sodium hydroxide and chilled on
energies and Electrostatic (Coulombic potential) fields. The
computed field energies were truncated to 30 kcal/mol for both
Steric and Electrostatic fields. CoMSIA calculates steric, electro-
static, hydrophobic, hydrogen-bond donor and acceptor fields
with the same lattice box and probe atom used for the CoMFA
ice quickly and the absorption of plate was measured at 405 nm
against blank. The following relation was used: OD405 ¼ ε$b$
[pNP], the molar absorption coefficient of pNP was
17,800 Mꢀ1 cmꢀ1, b was the light path (cm), and [pNP] was the
pNP concentration. IC50 values were calculated by analyzing the
data using OriginPro 7.5 software.
calculations. The attenuation factor
a, which determines the
steepness of the Gaussian function, was assigned a default value of
0.3, and an optimal value normally ranging from 0.2 to 0.4
[50e53].
5.3. 3D-QSAR model
5.3.1. Dataset
5.3.4. Partial least squares (PLS) analysis
Table 1 listed the structures and biological data of imidazoli-
dine-2,4-dione derivatives, which were used to construct 3D-QSAR
(CoMFA and CoMSIA) [39,40] models. Selections of test set mole-
cules were made by considering the fact that test set molecules
represent biological and structural diversity of training set mole-
cules. Besides, the molecules of test set represent 20% of training
set, which is regarded as a good percentage to validate a molecular
model [41e46]. Based on above principle, the dataset was divided
into training set and test set comprising 32 and 7 compounds,
respectively. For QSAR analysis, IC50 values were converted to pIC50
values according to the following equation:
Partial least square was used to linearly correlate the CoMFA/
CoMSIA fields to the bioactivity. Cross-validation was performed by
the Leave-One-Out (LOO) procedure to determine the optimum
number of components (ONC). Based on ONC, the non-cross-
validation was performed to give the conventional correlation co-
efficients (r2), standard error, and the F ratio. The predictive cor-
relation coefficient ðr2predÞ was calculated using test set to further
verify the predictive ability of the model.
5.4. Molecular docking
pIC50 ¼ ꢀlog IC50
The Surflex-Dock [54e57] module in the SYBYL-X 1.1.1 molec-
ular modeling package [47] was used to undertake molecular
docking for all studied compounds with target. The 3D structure of
PTP1B was taken from the Brookhaven Protein Databank (PDB
code: 2QBQ). Firstly, the compounds were prepared via the ligand
structure preparation module “sanitize”, which involved filling
valences, checking for correctness, standardizing, removing dupli-
cates and producing only one molecule of input structure [58,59].
Then, the receptor PTP1B was prepared. The hydrogen and missing
atoms were added, waters were removed and the energy was
minimized to obtain the docking receptor [60]. Finally, Surflex-
docking guided by protomol “Ligand mode” [61] was carried out.
In this mode, the receptor binding pocket was defined by the co-
ordinate space of the extracted co-crystallized ligand [62e64]. The
threshold and bloat parameters with value of 0.50 and 0 Å were
established to determine the extent of the binding pocket [65].
During the docking process, the ligand generated various confor-
mations to fit the receptor, in which the top 20 were displayed in
the result list according to their docking score [66]. The ligand
conformation with the highest total score was analyzed for the
interactions with receptor PTP1B.
5.3.2. Structural optimization and alignment
The CoMFA/CoMSIA studies were performed on a Dell com-
puter workstation with the Linux operating system using
the SYBYL-X 1.1.1 software package [47]. Energy minimizations
were performed using Gasteiger-huckel charge and Tripos force
field with a distance-dependent dielectric and the Powell conju-
gate gradient algorithm with convergence criterion of 0.05 kcal/
mol [45]. Subsequently, all structures were aligned by the com-
mon substructure using “molecular alignment” [48,49] with the
most active compound 5e was chosen as a template molecule
(Fig. 8).
5.3.3. CoMFA/CoMSIA interaction energy calculation
For aligned training set, CoMFA steric and electrostatic fields
were calculated at each lattice point with grid spacing value
of 2.0 Å in all Cartesian directions. A default sp3 C probe atom
with a Van-Der Waals radius of 1.52 Å and a charge of 1.0
was used to generate Steric (Lennard-Jones 6e12 potential) field
Fig. 8. (a) Structure of template molecule 5e used for molecular alignment of imidazolidine-2,4-dione derivatives. (b) Molecular alignment.

M.-Y. Wang et al. / European Journal of Medicinal Chemistry 103 (2015) 91e104
5137e5149.
103
Acknowledgments
[26] R. Ottana, R. Maccari, R. Ciurleo, P. Paoli, M. Jacomelli, G. Manao, G. Camici,
C. Laggner, T. Langer, 5-Arylidene-2-phenylimino-4-thiazolidinones as PTP1B
and LMW-PTP inhibitors, Bioorg. Med. Chem. 17 (2009) 1928e1937.
[27] B.R. Bhattarai, B. Kafle, J.-S. Hwang, D. Khadka, S.-M. Lee, J.-S. Kang, S.W. Ham,
I.-O. Han, H. Park, H. Cho, Thiazolidinedione derivatives as PTP1B inhibitors
with antihyperglycemic and antiobesity effects, Bioorg. Med. Chem. Lett. 19
(2009) 6161e6165.
This study was supported by the National Natural Science
Foundation of China (Grant No. 81273361 and 21202120), China
Postdoctoral Science Foundation funded project (2014T70222).
[28] M. Sarmiento, L. Wu, Y.F. Keng, L. Song, Z. Luo, Z. Huang, G.Z. Wu, A.K. Yuan,
Z.Y. Zhang, Structure-based discovery of small molecule inhibitors targeted to
protein tyrosine phosphatase 1B, J. Med. Chem. 43 (2000) 146e155.
[29] Y. Cheng, M. Zhou, C.-H. Tung, M. Ji, F. Zhang, Studies on two types of PTP1B
inhibitors for the treatment of type 2 diabetes: hologram QSAR for OBA and
BBB analogues, Bioorg. Med. Chem. Lett. 20 (2010) 3329e3337.
[30] M. Zhou, M. Ji, Molecular docking and 3D-QSAR on 2-(oxalylamino) benzoic
acid and its analogues as protein tyrosine phosphatase 1B inhibitors, Bioorg.
Med. Chem. Lett. 15 (2005) 5521e5525.
References
[1] J.M. Chehade, A.D. Mooradian, A rational approach to drug therapy of type 2
diabetes mellitus, Drugs 60 (2000) 95e113.
[2] C.R. Chittleborough, J.F. Grant, P.J. Phillips, A.W. Taylor, The increasing prev-
alence of diabetes in South Australia: the relationship with population ageing
and obesity, Public Health 121 (2007) 92e99.
[3] L. Blonde, Current challenges in diabetes management, Clin. Cornerstone 7
(Suppl. 3) (2005) S6eS17.
[4] S.J. Patankar, P.C. Jurs, Classification of inhibitors of protein tyrosine phos-
phatase 1B using molecular structure based descriptors, J. Chem. Inf. Comput.
Sci. 43 (2003) 885e899.
[5] J. Tobin, S. Tam, Recent advances in the development of small molecule in-
hibitors of PTP1B for the treatment of insulin resistance and type 2 diabetes,
Curr. Opin. Drug Discov. Dev. 5 (2002) 500e512.
[6] G. Liu, J.M. Trevillyan, Protein tyrosine phosphatase 1B as a target for the
treatment of impaired glucose tolerance and type II diabetes, Curr. Opin.
Investig. Drugs 3 (2002) 1608e1616.
[31] P.C. Nair, M.E. Sobhia, CoMFA based de novo design of pyridazine analogs as
PTP1B inhibitors, J. Mol. Graph. Model. 26 (2007) 117e123.
[32] S. Thareja, S. Aggarwal, T.R. Bhardwaj, M. Kumar, Self-organizing molecular
field analysis of 2,4-thiazolidinediones: a 3D-QSAR model for the develop-
ment of human PTP1B inhibitors, Eur. J. Med. Chem. 45 (2010) 2537e2546.
[33] A.O. Ajala, C.O. Okoro, CoMFA and CoMSIA studies on fluorinated hexahy-
dropyrimidine derivatives, Bioorg. Med. Chem. Lett. 21 (2011) 7392e7398.
[34] M.A. Khanfar, D.T. Youssef, K.A. El Sayed, 3D-QSAR studies of latrunculin-
based actin polymerization inhibitors using CoMFA and CoMSIA approaches,
Eur. J. Med. Chem. 45 (2010) 3662e3668.
[35] R.B. Huang, Q.S. Du, C.H. Wang, K.C. Chou, An in-depth analysis of the bio-
logical functional studies based on the NMR M2 channel structure of influenza
A virus, Biochem. Biophys. Res. Commun. 377 (2008) 1243e1247.
[36] S.Q. Wang, Q.S. Du, R.B. Huang, D.W. Zhang, K.C. Chou, Insights from inves-
tigating the interaction of oseltamivir (Tamiflu) with neuraminidase of the
2009 H1N1 swine flu virus, Biochem. Biophys. Res. Commun. 386 (2009)
432e436.
[37] Q.H. Liao, Q.Z. Gao, J. Wei, K.C. Chou, Docking and molecular dynamics study
on the inhibitory activity of novel inhibitors on epidermal growth factor re-
ceptor (EGFR), Med. Chem. 7 (2011) 24e31.
[38] D.P. Wilson, Z.K. Wan, W.X. Xu, S.J. Kirincich, B.C. Follows, D. Joseph-McCarthy,
K. Foreman, A. Moretto, J. Wu, M. Zhu, E. Binnun, Y.L. Zhang, M. Tam, D.V. Erbe,
J. Tobin, X. Xu, L. Leung, A. Shilling, S.Y. Tam, T.S. Mansour, J. Lee, Structure-
based optimization of protein tyrosine phosphatase 1B inhibitors: from the
active site to the second phosphotyrosine binding site, J. Med. Chem. 50
(2007) 4681e4698.
[39] R.D. Cramer, D.E. Patterson, J.D. Bunce, Comparative molecular field analysis
(CoMFA). 1. Effect of shape on binding of steroids to carrier proteins, J. Am.
Chem. Soc. 110 (1988) 5959e5967.
[7] T.O. Johnson, J. Ermolieff, M.R. Jirousek, Protein tyrosine phosphatase 1B in-
hibitors for diabetes, Nat. Rev. Drug Discov. 1 (2002) 696e709.
[8] T. Hunter, Protein kinases and phosphatases: the yin and yang of protein
phosphorylation and signaling, Cell 80 (1995) 225e236.
[9] J.L. Evans, B. Jallal, Protein tyrosine phosphatases: their role in insulin action
and potential as drug targets, Expert Opin. Invest. Drugs 8 (1999) 139e160.
[10] A. Ostman, F.D. Bohmer, Regulation of receptor tyrosine kinase signaling by
protein tyrosine phosphatases, Trends Cell Biol. 11 (2001) 258e266.
[11] M. Elchebly, P. Payette, E. Michaliszyn, W. Cromlish, S. Collins, A.L. Loy,
D. Normandin, A. Cheng, J. Himms-Hagen, C.-C. Chan, Increased insulin
sensitivity and obesity resistance in mice lacking the protein tyrosine
phosphatase-1B gene, Science 283 (1999) 1544e1548.
[12] L.D. Klaman, O. Boss, O.D. Peroni, J.K. Kim, J.L. Martino, J.M. Zabolotny,
N. Moghal, M. Lubkin, Y.B. Kim, A.H. Sharpe, A. Stricker-Krongrad,
G.I. Shulman, B.G. Neel, B.B. Kahn, Increased energy expenditure, decreased
adiposity, and tissue-specific insulin sensitivity in protein-tyrosine phospha-
tase 1B-deficient mice, Mol. Cell. Biol. 20 (2000) 5479e5489.
[13] G. Liu, Technology evaluation: ISIS-113715, Isis, Curr. Opin. Mol. Ther. 6
(2004) 331e336.
[40] G. Klebe, U. Abraham, T. Mietzner, Molecular similarity indices in a compar-
ative analysis (CoMSIA) of drug molecules to correlate and predict their bio-
logical activity, J. Med. Chem. 37 (1994) 4130e4146.
[14] S. Koren, I.G. Fantus, Inhibition of the protein tyrosine phosphatase PTP1B:
potential therapy for obesity, insulin resistance and type-2 diabetes mellitus,
Best Pract. Res. Clin. Endocrinol. Metab. 21 (2007) 621e640.
[41] E. Cichero, S. Cesarini, P. Fossa, A. Spallarossa, L. Mosti, Thiocarbamates as
non-nucleoside HIV-1 reverse transcriptase inhibitors: docking-based CoMFA
and CoMSIA analyses, Eur. J. Med. Chem. 44 (2009) 2059e2070.
[42] S.Y. Liao, L. Qian, T.F. Miao, H.L. Lu, K.C. Zheng, CoMFA and docking studies of
2-phenylindole derivatives with anticancer activity, Eur. J. Med. Chem. 44
(2009) 2822e2827.
[43] J. Perez-Villanueva, J.L. Medina-Franco, T.R. Caulfield, A. Hernandez-Campos,
F. Hernandez-Luis, L. Yepez-Mulia, R. Castillo, Comparative molecular field
analysis (CoMFA) and comparative molecular similarity indices analysis
(CoMSIA) of some benzimidazole derivatives with trichomonicidal activity,
Eur. J. Med. Chem. 46 (2011) 3499e3508.
[44] C. Gueto, J. Torres, R. Vivas-Reyes, CoMFA, LeapFrog and blind docking studies
on sulfonanilide derivatives acting as selective aromatase expression regula-
tors, Eur. J. Med. Chem. 44 (2009) 3445e3451.
[45] S. Pirhadi, J.B. Ghasemi, 3D-QSAR analysis of human immunodeficiency virus
entry-1 inhibitors by CoMFA and CoMSIA, Eur. J. Med. Chem. 45 (2010)
4897e4903.
[46] S.R. Patel, R. Gangwal, A.T. Sangamwar, R. Jain, Synthesis, biological evaluation
and 3D-QSAR study of hydrazide, semicarbazide and thiosemicarbazide de-
rivatives of 4-(adamantan-1-yl)quinoline as anti-tuberculosis agents, Eur. J.
Med. Chem. 85 (2014) 255e267.
[47] R.W. Homer, J. Swanson, R.J. Jilek, T. Hurst, R.D. Clark, SYBYL line notation
(SLN): a single notation to represent chemical structures, queries, reactions,
and virtual libraries, J. Chem. Inf. Model. 48 (2008) 2294e2307.
[48] P. Aparoy, G.K. Suresh, K. Kumar Reddy, P. Reddanna, CoMFA and CoMSIA
studies on 5-hydroxyindole-3-carboxylate derivatives as 5-lipoxygenase in-
hibitors: generation of homology model and docking studies, Bioorg. Med.
Chem. Lett. 21 (2011) 456e462.
[49] P. Lu, X. Wei, R. Zhang, CoMFA and CoMSIA studies on HIV-1 attachment
inhibitors, Eur. J. Med. Chem. 45 (2010) 1792e1798.
[50] K.C. Chou, D.Q. Wei, W.Z. Zhong, Binding mechanism of coronavirus main
proteinase with ligands and its implication to drug design against SARS,
Biochem. Biophys. Res. Commun. 308 (2003) 148e151.
[15] S. Lee, Q. Wang, Recent development of small molecular specific inhibitor of
protein tyrosine phosphatase 1B, Med. Res. Rev. 27 (2007) 553e573.
[16] S. Zhang, Z.-Y. Zhang, PTP1B as a drug target: recent developments in PTP1B
inhibitor discovery, Drug Discov. Today 12 (2007) 373e381.
[17] Z.-Q. Liu, T. Liu, C. Chen, M.-Y. Li, Z.-Y. Wang, R.-s. Chen, G.-x. Wei, X.-y. Wang,
D.-Q. Luo, Fumosorinone, a novel PTP1B inhibitor, activates insulin signaling
in insulin-resistance HepG2 cells and shows anti-diabetic effect in diabetic
KKAy mice, Toxicol. Appl. Pharmacol.
[18] H.B. He, L.X. Gao, Q.F. Deng, W.P. Ma, C.L. Tang, W.W. Qiu, J. Tang, J.Y. Li, J. Li,
F. Yang, Synthesis and biological evaluation of 4,4-dimethyl lithocholic acid
derivatives as novel inhibitors of protein tyrosine phosphatase 1B, Bioorg.
Med. Chem. Lett. 22 (2012) 7237e7242.
[19] M.V. Reddy, C. Ghadiyaram, S.K. Panigrahi, S. Hosahalli, L.N. Mangamoori,
Diphenylether derivative as selective inhibitor of protein tyrosine phospha-
tase 1B (PTP1B) over t-cell protein tyrosine phosphatase (TCPTP) identified
through virtual screening, Mini Rev. Med. Chem. 13 (2013) 1602e1606.
[20] A. Tsuchiya, T. Kanno, T. Nishizaki, Stearic acid serves as a potent inhibitor of
protein tyrosine phosphatase 1B, Cell. Physiol. Biochem. 32 (2013)
1451e1459.
[21] M. Ito, S. Fukuda, S. Sakata, H. Morinaga, T. Ohta, Pharmacological effects of
JTT-551, a novel protein tyrosine phosphatase 1B inhibitor, in diet-induced
obesity mice, J. Diabetes Res. 2014 (2014) 680348.
[22] Y. Li, Y. Yu, K. Jin, L. Gao, T. Luo, L. Sheng, X. Shao, J. Li, Synthesis and biological
evaluation of novel thiadiazole amides as potent Cdc25B and PTP1B inhibitors,
Bioorg. Med. Chem. Lett. 24 (2014) 4125e4128.
[23] A. Maeda, K. Kai, M. Ishii, T. Ishii, M. Akagawa, Safranal, a novel protein
tyrosine phosphatase 1B inhibitor, activates insulin signaling in C2C12 myo-
tubes and improves glucose tolerance in diabetic KK-Ay mice, Mol. Nutr. Food
Res. 58 (2014) 1177e1189.
[24] Z. Qin, N.R. Pandey, X. Zhou, C.A. Stewart, A. Hari, H. Huang, A.F.R. Stewart,
J.M. Brunel, H.-H. Chen, Functional properties of Claramine: a novel PTP1B
inhibitor and insulin-mimetic compound, Biochem. Biophys. Res. Commun.
458 (2015) 21e27.
[51] Q. Du, S. Wang, D. Wei, S. Sirois, K.C. Chou, Molecular modeling and chemical
modification for finding peptide inhibitor against severe acute respiratory
syndrome coronavirus main proteinase, Anal. Biochem. 337 (2005) 262e270.
[25] R. Maccari, P. Paoli, R. Ottana, M. Jacomelli, R. Ciurleo, G. Manao, T. Steindl,
T. Langer, M.G. Vigorita, G. Camici, 5-Arylidene-2,4-thiazolidinediones as in-
hibitors of protein tyrosine phosphatases, Bioorg. Med. Chem. 15 (2007)

104
M.-Y. Wang et al. / European Journal of Medicinal Chemistry 103 (2015) 91e104
[52] Q.S. Du, R.B. Huang, C.H. Wang, X.M. Li, K.C. Chou, Energetic analysis of the
two controversial drug binding sites of the M2 proton channel in influenza A
virus, J. Theor. Biol. 259 (2009) 159e164.
[53] H. Wei, C.H. Wang, Q.S. Du, J. Meng, K.C. Chou, Investigation into adamantane-
based M2 inhibitors with FB-QSAR, Med. Chem. 5 (2009) 305e317.
[54] R. Spitzer, A.N. Jain, Surflex-Dock: docking benchmarks and real-world
application, J. Comput. Aided Mol. Des. 26 (2012) 687e699.
[55] A.N. Jain, Surflex-Dock 2.1: robust performance from ligand energetic
modeling, ring flexibility, and knowledge-based search, J. Comput. Aided Mol.
Des. 21 (2007) 281e306.
[56] M.A. Miteva, W.H. Lee, M.O. Montes, B.O. Villoutreix, Fast structure-based
virtual ligand screening combining FRED, DOCK, and Surflex, J. Med. Chem.
48 (2005) 6012e6022.
enrichments, J. Comput. Aided Mol. Des. 27 (2013) 221e234.
[60] J. Xu, Y.F. Ma, W.S. Liu, Y. Tang, M.Y. Tan, Preparation, crystal structures and
luminescent properties of terbium and europium complexes with a new
amino-alkenone type ligand, J. Fluoresc. 21 (2011) 35e42.
[61] J. Ruppert, W. Welch, A.N. Jain, Automatic identification and representation of
protein binding sites for molecular docking, Protein Sci. 6 (1997) 524e533.
[62] J. Zhang, C.H. Luan, K.C. Chou, G.V. Johnson, Identification of the N-terminal
functional domains of Cdk5 by molecular truncation and computer modeling,
Proteins 48 (2002) 447e453.
[63] J.F. Wang, D.Q. Wei, L. Li, S.Y. Zheng, Y.X. Li, K.C. Chou, 3D structure modeling
of cytochrome P450 2C19 and its implication for personalized drug design,
Biochem. Biophys. Res. Commun. 355 (2007) 513e519.
[64] R.M. Pielak, J.R. Schnell, J.J. Chou, Mechanism of drug inhibition and drug
resistance of influenza A M2 channel, Proc. Natl. Acad. Sci. U. S. A. 106 (2009)
7379e7384.
[57] A.N. Jain, Surflex: fully automatic flexible molecular docking using a molecular
similarity-based search engine, J. Med. Chem. 46 (2003) 499e511.
[58] S.S. Rozenel, J.B. Kerr, J. Arnold, Metal complexes of Co, Ni and Cu with the
pincer ligand HN(CH2CH2P(i)Pr2)2: preparation, characterization and elec-
trochemistry, Dalton Trans. 40 (2011) 10397e10405.
[65] M. Lape, C. Elam, S. Paula, Comparison of current docking tools for the
simulation of inhibitor binding by the transmembrane domain of the sarco/
endoplasmic reticulum calcium ATPase, Biophys. Chem. 150 (2010) 88e97.
[66] R.D. Clark, A. Strizhev, J.M. Leonard, J.F. Blake, J.B. Matthew, Consensus scoring
for ligand/protein interactions, J. Mol. Graph. Model. 20 (2002) 281e295.
[59] G.M. Sastry, M. Adzhigirey, T. Day, R. Annabhimoju, W. Sherman, Protein and
ligand preparation: parameters, protocols, and influence on virtual screening