Pharmacophore identification, synthesis, and biological evaluation of carboxylated chalcone derivatives as CysLT1 antagonists

Article abstract of DOI:10.1016/j.bmc.2010.06.047

The pharmacophore model (Hypo1) with a well prediction capacity for CysLT₁ antagonists was developed using Catalyst/HypoGen program. Virtual screening against an in-house database consisted of carboxylated chalcones using Hypo1 was performed. Retrieved hits 26a, 26b, 27a, and 27b were synthesized and biological evaluated, the results of which demonstrated that these compounds showed moderate to good CysLT₁ antagonistic activities. This study indicated that the generated model (Hypo1) is a reliable and useful tool in lead optimization for novel CysLT₁ antagonists.

Full text of DOI:10.1016/j.bmc.2010.06.047

Bioorganic & Medicinal Chemistry 18 (2010) 5519–5527
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Pharmacophore identification, synthesis, and biological evaluation
of carboxylated chalcone derivatives as CysLT1 antagonists
Xiaowu Dong a, Li Wang a, Xueqin Huang b, Tao Liu a, Erqing Wei b, Lilin Du a, Bo Yang c, Yongzhou Hu a,
*
a ZJU-ENS Joint Laboratory of Medicinal Chemistry, College of Pharmaceutical Sciences, Zhejiang University, Hangzhou 310058, China
b Department of Pharmacology, School of Medicine, Zhejiang University, Hangzhou 310058, China
c Institute of Pharmacology & Toxicology, College of Pharmaceutical Sciences, Zhejiang University, Hangzhou 310058, China
a r t i c l e i n f o
a b s t r a c t
Article history:
The pharmacophore model (Hypo1) with a well prediction capacity for CysLT1 antagonists was developed
using Catalyst/HypoGen program. Virtual screening against an in-house database consisted of carboxyl-
ated chalcones using Hypo1 was performed. Retrieved hits 26a, 26b, 27a, and 27b were synthesized and
biological evaluated, the results of which demonstrated that these compounds showed moderate to good
CysLT1 antagonistic activities. This study indicated that the generated model (Hypo1) is a reliable and
useful tool in lead optimization for novel CysLT1 antagonists.
Received 22 April 2010
Revised 12 June 2010
Accepted 15 June 2010
Available online 20 June 2010
Keywords:
CysLT1 antagonists
HypoGen
Ó 2010 Elsevier Ltd. All rights reserved.
Pharmacophore model
Carboxylated chalcone
QSAR
1. Introduction
Palomer et al. established two 3D-QSAR models based on series
of quinolines-based CysLT1 antagonists which may act in a similar
Peptidoleukotrienes (LTC4, LTD4, LTE4), also referred as cys-
teine-containing leukotrienes (Cys-LTs), play crucial roles in respi-
ratory diseases (such as asthma and allergic rhinitis)1–3 and other
inflammatory conditions including cardiovascular, gastrointestinal
and immune disorders.4–6 The biological action of peptidoleukotri-
enes can be modulated by intervening biosynthesis or blocking the
CysLT1 receptor, which belongs to the G-protein-coupled receptors
(GPCR) superfamily.7–10
Extensive studies on CysLT1 antagonists have led to the discov-
ery of a wide range of compounds (Fig. 1), such as indoles (e.g.,
Zafirlukast),11 benzamides (e.g., Pranlukast)12 and quinolines
(Montelukast).13 Important structural features for high potency
were concluded from these reported CysLT1 antagonists:14 (i) a
lipophilic anchor, which fits into the lipophilic pocket of the CysLT1
receptor; (ii) a central flat unit, which mimics the triene system of
the natural agonist LTD4; (iii) one or two acidic groups, which
resembles the peptide unit and/or the C1-carboxylic acid of LTD4.
Since the X-ray crystal structure of CysLT1 receptor is not available,
ligand-based modeling studies including 3D-QSAR models have
been widely performed. Terada et al. proposed a 3D-QSAR model
based on the comparison of Pranlukast and some derivatives with
LTE4 and proposed the activation site of the CysLT1 receptor.15
preferential fashion.16 However, none of the QSAR models pro-
posed so far was applied in virtual screen for novel CysLT1
antagonists.
In this study, a pharmacophore-based lead optimization proto-
col was applied to discover novel potent CysLT1 antagonists. At
first, a novel 3D pharmacophore model was proposed based on
different CysLT1 antagonists using Catalyst/HypoGen program
(Catalyst 4.1, Molecular Simulations Inc, San Diego, CA). Then,
pharmacophore-based virtual screen was performed to search an
in-house database consisted of newly designed carboxylated chal-
cones according to the literatures.17 Furthermore, preferable
screened hits were picked up, synthesized and biological evaluated
to validate the reliability of the obtained pharmacophore model.
2. Computational methods
2.1. Data preparation
For the pharmacophore modeling studies, a set of 77 CysLT1
antagonists were selected from the literatures,16–21 and divided
into a training set (25 molecules, Fig. 2) and a test set (52 mole-
cules, Fig. S1, Supplementary data) based on principles of struc-
tural diversity and wide coverage of the activity range (between
0.19 and 55,000 nM). Besides, the 1500 negatives compounds
used in enrichment studies were retrieved from Available
* Corresponding author. Tel./fax: +86 571 888208460.
E-mail address: huyz@zju.edu.cn (Y. Hu).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.06.047

5520
X. Dong et al. / Bioorg. Med. Chem. 18 (2010) 5519–5527
Figure 1. The 2D structures of some typical CysLT1 antagonists.
Figure 2. 2D chemical structures of the 25 molecules forming the training set used to obtain HypoGen pharmacophore hypotheses.

X. Dong et al. / Bioorg. Med. Chem. 18 (2010) 5519–5527
5521
Chemical Directory (ACD) database (Symyx Technologies, Santa
2.3. Pharmacophore validation
Clara, CA) using ‘Random Percent Filter protocol’ by Pipeline Pilot
software (Accelrys, Inc., San Diego, USA).
In order to validate the reliability and accuracy of the generated
3D pharmacophore models, cost analysis, test set activity predic-
tion, Fisher’s test and enrichment factors studies were performed.
Carboxylated chalcone derivatives which were firstly discov-
ered by Zwaagstra et al. and showed preferable CysLT1 antagonis-
tic activities in both in vitro and in vivo assays.17 These prompted
us to design a novel series of carboxylated chalcone derivatives
with improved activities. In the course of novel compounds gen-
eration program, the effect of different substituents on A-ring of
chalcones (e.g., halogen, COOH, hydroxyl) and the influence of dif-
ferent substituent and length of lipophilic chain on B-ring of chal-
cones (e.g., phenyl or alkyl chain) were taken into consideration.
All candidate fragments linked to chalcone skeleton were assem-
bled using Pipeline Pilot student edition (Accelrys, Inc., San Diego,
USA).
2.3.1. Cost analysis
The quality of HypoGen models can be described in terms of
fixed cost, null cost, and total cost, all of which were well defined
by Debnath.23 As a good model, the total cost of any hypothesis
should be close to the fixed cost. If a returned cost (total cost) dif-
fers from the null hypothesis by 40–60 bits, it is highly probable
that the hypothesis has 75–90% chance representing the true cor-
relation of the data (Catalyst 4.10 documentation).
All compounds were utilized to ascertain the energy-minimized
conformations using a CHARMm-like force field.22 Then, conforma-
tional models of each compound were generated using the Poling
Algorithm implemented in Catalyst (Catalyst 4.10 documentation).
A maximum of 250 conformations were generated for every com-
pound to ensure maximum coverage of the conformational space
within an energy threshold of 20.0 kcal/mol.
2.3.2. Test set activity prediction
In addition to estimation of the activity of training set mole-
cules, the pharmacophore model was also used to predict the bio-
activities of new compounds. For such a purpose, 52 CysLT1
antagonists in the test set, also having wide range of activities
(IC50, spanning from 0.21 to 10,500 nM) and structural diversity,
were predicted by the best pharmacophore model.
2.2. Pharmacophore generation
2.3.3. Fisher’s test
Using the CatScrambe module, the molecular spreadsheets of
our training set were modified by arbitrary scrambling of the affin-
ity data. These randomized spreadsheets should yield hypotheses
without statistical significance; otherwise, the original model
was also obtained randomly. To achieve a statistical significance le-
vel of 95%, 19 random spreadsheets were generated.
Based on the conformations for each compound, the HypoGen
algorithm implemented in Catalyst program was employed to
establish pharmacophore models. Three chemical feature types
termed ring aromatic (AR), hydrophobic (HY) and negative ioniz-
able (NI) features were selected for hypotheses generation, since
they could effectively map all critical chemical features of all mol-
ecules. All the parameters were set as default and the uncertainty
value was set to 3.0 for compound activity, representing the ratio
of the uncertainty range of measured biology activity against the
actual activity for each compound. The Catalyst program returned
10 pharmacophore models with top ranking scores.
2.3.4. Enrichment factor study
In the lead-discovery studies, the pharmacophore model should
be useful in identifying active leads in virtual screening. So the
generated models were further validated by enrichment factor
(EF) studies in screening an database which was retrieved
Table 1
Experimental and estimated CysLT1 antagonistic activities of training set based on Hypo1
Compd
Fit
Mapping
CysLT1 activities(IC50
)
Errora
Activity scale(IC50
)
Exp. (nM)
Est. (nM)
Exp.
Est.
Ref.
1b
11.92
10.94
9.93
9.71
9.56
9.61
9.79
10.19
9.36
9.67
9.72
10.03
9.48
8.71
8.91
7.75
7.77
7.76
7.78
7.67
7.80
7.63
7.78
6.72
6.23
[11 6 36 38 24]
[11 6 36 37 24]
*
0.19
0.5
0.8
2
2.4
4.1
6.6
7.3
14
14
17
17
38
130
190
270
500
530
810
1700
1700
2100
4000
34,000
55,000
0.066
0.63
6.4
11
15
13
9
3.6
24
12
10
5
18
110
68
970
920
960
910
1200
870
1300
910
10,000
32,000
ꢀ2.9
1.3
8
5.3
6.2
+++
+++
+++
+++
+++
+++
+++
+++
++
++
++
++
++
++
++
++
++
++
++
+
+++
+++
+++
+++
+++
+++
+++
+++
++
++
++
+++
++
++
++
++
++
++
++
+
++
+
++
18
18
19
19
18
16
17
16
19
17
20
21
16
16
18
21
21
20
20
17
21
17
20
17
21
2
3b
[23 13 10 32
]
*
4
[5 33 35 16]
*
5
[11 6 38 24 ]
[19 30 31 59 *]
[2 17 8 49 *]
[2 15 16 48 *]
[14 8 1 * 25]
*
6b
3.3
1.4
7
8
9
ꢀ2
1.7
10
11b
12
13
14b
15
16
17b
18
19b
20
21b
22
23b
24
25b
[20 1 15 25
[14 22 28 30
[14 22 28 31
]
*
*
ꢀ1.2
ꢀ1.6
ꢀ3.4
ꢀ2.1
ꢀ1.3
ꢀ2.9
3.6
1.8
1.8
1.1
]
]
*
]
[42 13 37 1
[57 18 23 * 3]
[11 6 * 34 24]
[14 22 28 * *]
[30 * 27 31 *]
*
*
]
[
22 28 30
* *
]
[2 27 22
*
*
]
[21 16 31
ꢀ1.5
[14 22 27 * *]
[48 24 31 * *]
[14 28 * 36 *]
[44 27 * 14 *]
*
ꢀ2
+
+
+
+
ꢀ1.6
ꢀ4.4
ꢀ3.2
ꢀ1.7
+
+
[2 14 23 26
]
+
a
Error values is ratio between the estimated and experimental activity, while a negative sign indicates the actual activity is higher than the estimated one.
The compounds were used in enrichment factor studies.
b

5522
X. Dong et al. / Bioorg. Med. Chem. 18 (2010) 5519–5527
randomly from ACD database (1500 molecules) spiked some
known inhibitors (10 molecules, Table 1).
range between Hypo1 and the fixed cost is 12.22, while that be-
tween the null hypothesis and Hypo1 is 77.30 (Table 2), which
shows that Hypo1 has more than 90% probability of correlating
the data. Noticeably, the total cost of Hypo1 is much closer to
the fixed cost than to the null cost. Besides, a high correlation coef-
ficient of 0.956 is observed with RMSD value of 0.901 and the
configuration cost of 15.31, demonstrating that a meaningful phar-
macophore model with good statistical parameters were success-
fully developed.
TP
TP þ FP
N
n
Enrichment factorðEFÞ ¼
ꢁ
;
where TP is the number of true positive compounds, FP is the num-
ber of false positive compounds, N is the number of the total com-
pounds and n is the number of the total positive compounds.
Table 1 and Fig. 4A represent the actual and estimated CysLT1
antagonistic activities (IC50 values) of the 25 training set mole-
cules. Based on their potency of CysLT1 antagonistic activity scale,
these 25 compounds were classified to highly active compounds
(+++, IC50 <10 nM), moderately active compounds (++, 10 nM 6
IC50 <1000 nM) and weakly active compounds (+, IC50 P 1000 nM).
Among the molecules 1–25, all of highly active compounds were
estimated correctly, while only 9.1% of moderately active com-
pounds (compound 12) and 33.3% of weakly active compounds
(compounds 21 and 23) were predicted wrongly. For example,
the CysLT1 antagonistic activities of the most active compound 1
3. Results and discussion
3.1. Pharmacophore model development
The results of pharmacophore hypotheses are presented in
Table 2. The Hypo1, the best pharmacophore model, describes
the topological position of the lipophilic moieties and the acidic
residues (Fig. 3A), known as important structural features for Cys-
LT1 receptor antagonists. It is consisted of five features, namely,
one ring aromatic (RA), one negative ionizable (NI), three hydro-
phobic (HY). The total fixed cost of the run is 100.52, the null cost
is 200.04, and the total cost of the Hypo1 is 112.74. Then, the cost
(IC50: 0.19 nM) and the weakly active compound 25 (IC50:
55,000 nM) could be predicted nicely using Hypo1. In the case of
compound 1, the carboxyl group, aromatic ring and aliphatic chain
mapped well onto Hypo1 (fit values: 11.92, Fig. 3B), and the pre-
Table 2
The results of top 10 hypotheses in HypoGen pharmacophore generation
dicted activity fit well with its actual activity (prediction IC50
:
0.066 nM; actual IC50: 0.19 nM). For molecule 25, three hydropho-
bic features and carboxyl group could not fit well to corresponding
pharmacophore features, and the ring aromatic feature is missed,
leading to a poor predicted activity of compound 25 (fit values:
6.23, Fig. 3C), which was also close to its actual activity (prediction
IC50: 32,000 nM; actual IC50: 55,000 nM).
Hypo
Total cost
D
cost
RMSD
Correl.
1
2
3
4
5
6
7
8
9
112.74
115.69
119.01
120.55
122.52
123.15
124.17
124.35
124.95
125.48
87.30
84.35
81.03
79.49
77.52
76.89
75.87
75.69
75.09
74.56
0.901
1.026
1.062
1.170
1.240
1.299
1.306
1.275
1.366
1.352
0.956
0.942
0.940
0.925
0.915
0.905
0.905
0.910
0.894
0.897
3.2. Pharmacophore model validation
10
3.2.1. Test set prediction
The Hypo1 was further validated by using a test set of 52 com-
pounds which were structurally distinct from those included in the
Null cost of 10 top-scored hypothesis is 200.04, fixed cost value is 100.52, and
configuration cost is 15.31.
Figure 3. The generated best Pharmacophore model (Hypo1) with distance constraints (A); Hypo1 aligned to compound 1 (Ki = 0.19 nM, (B) and 25 (Ki = 55,000 nM, (C).
Pharmacophore features are color-coded (hydrophobic, blue; negative ionizable, dark blue; ring aromatic: orange).

X. Dong et al. / Bioorg. Med. Chem. 18 (2010) 5519–5527
5523
Figure 4. The regression line of actual versus predicted activities by Hypo1 for the training set and test set.
training set. The actual activity versus estimated activity of the 52
3.3. Database searching
compounds in the test is shown in Table S1 as the Supplementary
data, which showed a good correlation of 0.814 between the actual
and the estimated activities (Fig. 4b).
To identify novel CysLT1 antagonists, Hypo1 was used as a
search query against an in-house databases consisted of carboxyl-
ated chalcone derivatives (537 molecules). Interestingly, over
ninety compounds fitted well with features of Hypo1 were pre-
dicted as potential candidates for further exploration (calcd Ki
<300 nM). Then, the synthesis and biological evaluation of four
compounds 26a,b and 27a,b (Fig. 6) for their CysLT1 antagonistic
activities were performed according to their good prediction activ-
ities and synthetic feasibility.
3.2.2. Fisher’s test
With the aid of the CatScramble program, the experimental
activities of compounds in the training set were scrambled ran-
domly, and the resulting training set was used for a HypoGen
run. All parameters were adopted and used in the initial HypoGen
calculation. This procedure was reiterated 19 times. None of the
outcome hypotheses had a lower cost score than the initial hypoth-
esis. Ten lowest total cost values of the resulting 19 hypotheses
were compared with our hypothesis costs. None of the costs of
the scrambled hypothesis was less than Hypo1 (Fig. 5), indicating
that there was a 95% chance for the best hypothesis to represent
a true correlation in the training set activity data.
3.4. Chemistry
The synthetic route for compounds 26 and 27 is outlined in
Scheme 1. Treatment of 4-hydroxybenzaldehyde, appropriate
alkylbromide and potassium carbonate in acetone afforded com-
pounds 29. Claisen–Schmidt condensation of acetophenone 28
with benzaldehyde 29a in presence of 10% potassium hydroxide
in a solvent of EtOH/THF/H2O (5:5:2) for 2–3 days gave chalcone
31, which were further etherized with ethyl bromoacetate to yield
compounds 32. Deprotection of the benzoic acid function of 32
using 5% sodium hydroxide in EtOH/H2O mixture solvent at 40 °C
resulted in carboxylated chalcone 26. Besides, acetophenone 30
were obtained using 4-hydroxybenzoic acid (or 3-hydroxybenzoic
acid) as starting materials, according to the literature.17 Condensa-
tion of acetophenone 30 with appropriate benzaldehyde 29 in
presence of 10% potassium hydroxide in EtOH/THF/H2O (5:5:2)
mixture solvent for even longer time (5–6 days), following by
esterization with ethanol catalyzed by sulfuric acid afforded chal-
cone 33. Etherization of 33 with chloroacetonitrile in presence of
potassium carbonate in acetone gave in 34, which were further
alkaloid hydrolyzed using 5% sodium hydroxide in EtOH/H2O mix-
ture solvent at room temperature to get carboxylated chalcone 27.
3.2.3. Enrichment factor study
The Hypo1 was further validated for picking up active mole-
cules (CysLT1 antagonists) from a database, including 1500 inactive
compounds and 10 known inhibitors (labeled in Table 1). The
Hypo1 picked up the known inhibitors with highly enrichment fac-
tors (18.0, 9.0, 4.5, and 3.3 at 5%, 10%, 20%, and 30%, respectively),
indicating that Hypo1 is a useful and reliable tool for identifying
CysLT1 antagonists in virtual screening.
3.5. CysLT1 antagonistic activities
The in vitro CysLT1 antagonistic potency of the compounds was
evaluated in dU937 cells (highly expressing CysLT1 receptor) pre-
stimulated by LTD4 according to reported methods.24–26 In this bio-
logical evaluation model, PT-PCR and Western blot analysis
showed a higher expression of CysLT1 receptor in dU937 cells.
ATP at 50 l
mol Lꢀ1 significantly elevated [Ca2+]i, and was used as
a positive control for measurement of [Ca2+]i. LTD4 at 0.1 mol Lꢀ1
l
also significantly increased [Ca2+]i, which could be blocked by
Montelukast (a well known CysLT1 antagonist) and compounds
Figure 5. The difference in costs between the HypoGen runs (Hypo1) and the
scrambled runs (RM1–RM19).

5524
X. Dong et al. / Bioorg. Med. Chem. 18 (2010) 5519–5527
Figure 6. The structures of selected carboxylated chalcones 26a,b and 27a,b.
Scheme 1. Synthesis of carboxylated chalcones 26a,b and 27a,b. Reagents and conditions: (a) 10% KOH, ethanol/THF/H2O, rt; (b) BrCH2COOEt, K2CO3, acetone, reflux; (c) 5%
NaOH, EtOH/H2O, 40 °C; (d) ethanol, H2SO4, reflux; (e) ClCH2CN, K2CO3, acetone, reflux; (f) 5% NaOH, EtOH/H2O, rt.
26a,b and 27a,b. The promoted antagonistic effects of tested com-
pounds were observed in a dose-dependent manner with the max-
imal effect observed at 10 lM as shown in Figure 7. The selected
four carboxylated chalcones (26a,b and 27a,b) showed moderate
to good CysLT1 antagonistic activities (Table 3), indicating that
pharmacophore-based CysLT1 antagonists design was a rational
protocol in lead optimization. As expected, the introduction of
carboxyl group and lipophilic chain on the A-ring and B-ring of
chalcone skeleton, respectively, enhanced the CysLT1 receptor
antagonistic activities, which was consisted with what discovered
by other groups.14 Cyanomethoxyl group on A-ring of chalcones
may afford additional interaction point, leading to the best antag-
onistic activities of 27a (EC50 = 0.48 lM), a promising novel lead for
further structural optimization and pharmacological studies.
4. Conclusion
Figure 7. The concentration-dependent blocking activity of compounds on LTD4
(0.1 in dU937 cells curves of some compounds 26a,b and
M)-elevated [Ca2+
27a,b.
l
]
i
In this study, we described a rational strategy for identifying no-
vel CysLT1 antagonists using a pharmacophore-based lead optimi-
zation protocol. The best pharmacophore model (Hypo1) was
developed using Catalyst/HypoGen program and showed good sta-
tistical parameters in validation procedure. Furthermore, four car-
bonylated chalcones with preferable predictive activities were
synthesized and identified as moderate to good CysLT1 antagonists,
indicating that pharmacophore-based CysLT1 antagonists design is
a rational protocol in lead optimization. To the best of our knowl-
edge, the 3D alignment and pharmacophoric features of CysLT1
antagonists and its application in virtual screening were evaluated
for the first time. The most active compound 27a would be a prom-
ising lead for further structural optimization and pharmacological
studies.
5. Experiment
5.1. Chemistry
Melting points were obtained on a B-540 Büchi melting-point
apparatus and are uncorrected. 1H NMR spectra was recorded on

X. Dong et al. / Bioorg. Med. Chem. 18 (2010) 5519–5527
5525
Table 3
(d, 2H, J = 8.0 Hz), 7.89 (d, 1H, J = 16.0 Hz), 7.92 (d, 1H, J = 7.6 Hz),
12.64 (s, 1H, OH). ESI-MS: m/z [M+H]+ 423.
The pharmacological results of synthesized carboxylated chalcones 29a,b and 30a,b
a
Compd
Pred. IC50 (nM)
Exp. EC50
(
lM)
Emax (%)
5.3.2. (E)-1-(4-Fluoro-2-hydroxyphenyl)-3-(4-(4-phenoxybut-
oxy)phenyl)prop-2-en-1-one (31b)
Montelukast
26a
26b
27a
27b
5.4
27
87
150
56
0.072
1.4
1.3
0.48
3.9
70.1
53.1
53.2
76.9
48.8
Reagents: acetophenone 28b (500 mg, 3.25 mmol), benzalde-
hyde 29a (1.04 g, 3.41 mmol). Product: pale yellow solid (0.83 g,
58%), mp 137–139 °C. 1H NMR (CDCl3, 400 MHz, d): 1.94 (m, 4H),
4.03(m, 2H), 4.12 (m, 2H), 6.84 (d, 2H, J = 8.0 Hz), 6.96 (d, 2H,
J = 8.0 Hz), 7.03 (dd, 1H, J = 5.2, 7.6 Hz), 7.22 (d, 2H, J = 8.0 Hz),
7.26 (td, J = 8.0, 1.2 Hz, 1H),7.32 (dd, J = 8.0, 1H, 1.2 Hz), 7.54 (d,
1H, J = 16.0 Hz), 7.80 (d, 2H, J = 8.0 Hz), 7.92 (d, 1H, J = 16.0 Hz),
12.64 (s, 1H, OH). ESI-MS: m/z [M+H]+ 441.
a
Predictive IC50 value was estimated according to Hypo1.
a 400 MHz spectrometer (Brüker AM), and the chemical shifts were
expressed as d values in parts per million (ppm) relative to tetra-
methylsilane (TMS). Mass spectra were recorded on an Esquire-
LC-00075 spectrometer. Reagents and solvents were purchased
from commercial sources. Acetophenone 28a,b were synthesized
according to the literature.17
5.4. General procedure for the synthesis of compounds 32a,b
To a solution of compound 31 and ethyl bromoacetate in ace-
tone (20 mL), potassium carbonate (325.7 mg, 2.36 mmol) was
added. The resulting mixture was refluxed for 5 h and then cooled
to room temperature. The mixture was filtered and the filtrate was
concentrated in vacuo. The residue was purified by flash column
chromatography on silica gel using petroleum ether/ethyl acetate
(10:1, v/v) as eluant to give 32.
5.2. General procedure for the synthesis of benzaldehyde 29a,b
To a stirred solution of 4-hydroxybenzaledhyde and (4-bromo-
butoxy)benzene (or 1-(4-bromobutoxy)-4-chlorobenzene) in ace-
tone (200 mL), potassium carbonate (11.3 g, 82 mmol) was
added. The resulting mixture was refluxed for 5 h, and then cooled
to room temperature. The mixture was filtered and the filtrate was
concentrated in vacuo. The residue was purified by flash column
chromatography on silica gel using petroleum ether/ethyl acetate
(10:1, v/v) as eluant to give 29.
5.4.1. (E)-Ethyl 2-(2-(3-(4-(4-(4-chlorophenoxy)butoxy)phenyl)
acryloyl)phenoxy)acetate (32a)
Reagents: compound 31a (500 mg, 1.18 mmol) and ethyl bro-
moacetate (198 mg, 1.18 mmol). Product: pale yellow solid
(493 mg, 82%), mp: 103–105 °C. 1H NMR (CDCl3, 400 MHz, d):
1.27 (t, 3H, J = 6.0 Hz), 1.97 (m, 4H), 4.00 (m, 2H), 4.06 (m, 2H),
4.65 (s, 2H), 6.80 (d, 2H, J = 8.4 Hz), 6.89 (d, 2H, J = 8.4 Hz), 7.21
(d, 2H, J = 8.4 Hz), 7.36 (m, 2H), 7.37 (d, 1H, J = 16.4 Hz), 7.56 (d,
2H, J = 8.4 Hz), 7.72 (d, 1H, J = 1.2 Hz), 7.98 (d, 1H, J = 16.4 Hz).
ESI-MS: m/z [M+H]+ 509.
5.2.1. 4-(4-(4-Chlorophenoxy)butoxy)benzaldehyde (29a)
Reagents: 4-hydroxybenzaledhyde (5.0 g, 41.0 mmol), 1-(4-
bromobutoxy)-4-chlorobenzene (10.8 g, 41.0 mmol). Product:
white solid (9.6 g, 77%), mp 83–85 °C. 1H NMR (CDCl3, 400 MHz,
d): 2.01 (m, 4H), 4.03 (t, 2H, J = 6.4 Hz), 4.12 (t, 2H, J = 6.4 Hz),
7.28 (d, 2H, J = 8.0 Hz), 7.54 (d, 2H, J = 8.4 Hz), 7.65 (d, 2H,
J = 8.4 Hz), 7.81 (d, 2H, J = 8.0 Hz), 9.83 (s, 1H). ESI-MS: m/z
[M+H]+ 305.
5.4.2. (E)-Ethyl 2-(2-(3-(4-(4-(4-chlorophenoxy)butoxy)phenyl)
acryloyl)-5-fluorophenoxy)acetate (32b)
Reagents: compound 31b (500 mg, 1.14 mmol) and ethyl bro-
moacetate (190 mg, 1.14 mmol). Product: pale yellow solid
(526 mg, 88%), mp: 115–118 °C. 1H NMR (CDCl3, 400 MHz,
d):1.25 (t, 3H, J = 6.0 Hz), 1.94 (m, 4H), 4.02 (m, 2H), 4.05 (m,
2H), 4.65 (s, 2H), 6.95 (d, 2H, J = 8.0 Hz), 7.00 (dd, 1H, J = 4.0,
8.0 Hz), 7.12 (d, 2H, J = 8.0 Hz), 7.21 (d, 2H, J = 8.0 Hz), 7.36 (m,
1H), 7.40 (dd, 1H, J = 2.0, 8.0 Hz), 7.37 (d, 1H, J = 16.4 Hz), 7.67 (d,
2H, J = 8.4 Hz), 7.98 (d, 1H, J = 16.4 Hz) ESI-MS: m/z [M+H]+ 527.
5.2.2. 4-(4-Phenoxybutoxy)benzaldehyde (29b)
Reagents: 4-hydroxybenzaldehyde (5.0 g, 41.0 mmol), 4-
bromobutoxybenzene (9.3 g, 41.0 mmol). Product: pale yellow syr-
up (7.5 g, 68%). 1H NMR (CDCl3, 400 MHz, d): 1.99 (m, 4H), 4.01(t,
2H, J = 6.0 Hz), 4.10 (t, 2H, J = 6.0 Hz), 6.88 (m, 3H), 7.03 (d, 2H,
J = 8.0 Hz),7.22 (t, 2H, J = 8.0), 7.68 (d, 2H, J = 8.0 Hz), 9.82 (s, 1H).
ESI-MS: m/z [M+H]+ 271.
5.3. General procedure for the synthesis of chalcone 31a,b
5.5. General procedure for the synthesis of compounds 26a,b
A solution of acetophenone 28, benzaldehyde 29a, and 10%
potassium hydroxide in a solvent of EtOH/THF/H2O (5:5:2, v/v,
6 mL) was stirred at room temperature for 2–3 days. The reaction
mixture was poured into ice-water, acidified to pH ꢂ5 with 2 N
HCl and extracted with ethyl acetate (20 mL ꢃ three times). The or-
ganic phase was washed with brine, dried over anhydrous Na2SO4,
and then concentrated in vacuo. The residue was purified by col-
umn chromatography on silica gel using petroleum ether/ethyl
acetate (15:1, v/v) as eluant to give 31.
A solution of 32 and 5% sodium hydroxide in EtOH/H2O (1:1, v/
v, 2 mL) was stirred at 40 °C for 3 h, then poured into cold water
and extracted with ethyl acetate (10 mL ꢃ three times). The organ-
ic phase was washed with brine, dried over anhydrous Na2SO4, and
then concentrated in vacuo. The crude product was recrystallized
in ethanol to give expected compound.
5.5.1. (E)-2-(2-(3-(4-(4-(4-Chlorophenoxy)butoxy)phenyl)acry-
loyl)phenoxy)acetic acid (26a)
Reagents: compound 32a (50 mg, 0.1 mmol). Product: pale yel-
low solid (35.4 mg, 75%), mp 132–134 °C. 1H NMR (acetone-d6,
400 M, d): 1.93 (m, 4H), 4.04 (t, 2H, J = 6.4 Hz), 4.11 (t, 2H,
J = 6.4 Hz), 4.87 (s, 2H), 6.90 (d, 2H, J = 8.8 Hz), 6.93 (d, 2H,
J = 8.8 Hz), 7.05 (td, 1H, J = 1.6, 8.4 Hz), 7.11 (d, 1H, J = 8.4 Hz),
7.23 (d, 2H, J = 8.8 Hz), 7.47 (td, 2H, J = 1.6, 8.4 Hz), 7.58 (d, 1H,
J = 16.0 Hz), 7.60 (d, 1H, J = 8.4 Hz), 7.63 (d, 1H, J = 16.0 Hz), 7.67
(d, 2H, J = 8.8 Hz). ESI-MS: m/z [MꢀH]ꢀ 479.
5.3.1. (E)-3-(4-(4-(4-Chlorophenoxy)butoxy)phenyl)-1-(2-hydr-
oxyphenyl)prop-2-en-1-one (31a)
Reagents: acetophenone 28a (500 mg, 3.68 mmol) and benzal-
dehyde 29a (1.17 g, 3.86 mmol). Product: pale yellow solid
(1.01 g, 65%), mp 125–128 °C. 1H NMR (CDCl3, 400 MHz, d): 1.99
(m, 4H), 4.01(m, 2H), 4.10 (m, 2H), 6.81 (d, 2H, J = 8.0 Hz), 6.93
(d, 2H, J = 8.0 Hz), 7.03 (td, 1H, J = 1.6, 7.6 Hz), 7.21 (d, 2H,
J = 8.0 Hz), 7.47 (t, 2H, J = 7.6 Hz), 7.55 (d, 1H, J = 16.0 Hz), 7.63

5526
X. Dong et al. / Bioorg. Med. Chem. 18 (2010) 5519–5527
5.5.2. (E)-2-(2-(3-(4-(4-(4-Chlorophenoxy)butoxy)phenyl)acry-
loyl)-5-fluorophenoxy)acetic acid (26b)
5.7.1. (E)-4-(3-(4-(4-(4-Chlorophenoxy)butoxy)phenyl)acry-
loyl)-3-(cyanomethoxy)benzoic acid (27a)
Reagent: compound 32b (50 mg, 0.095 mmol). Product: pale
yellow solid (32.2 mg, 68%), mp 122–124 °C?1H NMR (acetone-d6,
400 M, d): 1.98 (m, 4H), 4.09 (t, 2H, J = 6.4 Hz), 4.16 (t, 2H,
J = 6.4 Hz), 4.91 (s, 2H), 6.96 (d, 2H, J = 8.4 Hz), 6.98 (d, 2H,
J = 8.0 Hz), 7.21 (dd, 1H, J = 4.0, 8.8 Hz), 7.28 (d, 2H, J = 8.4 Hz),
7.29 (m, 1H), 7.38 (dd, 1H, J = 3.2, 8.8 Hz), 7.65 (d, 1H,
J = 16.0 Hz), 7.68 (d, 1H, J = 16.0 Hz), 7.72 (d, 2H, J = 8.0 Hz). ESI-
MS: m/z [MꢀH]ꢀ 497.
Reagent: compound 36a (100 mg, 0.20 mmol), chloroacetonitri-
le (15.3 mg, 0.20 mmol), potassium carbonate (55.9 mg,
0.41 mmol). Product: pale yellow solid (56.2 mg, 55%), mp 148–
151 °C. 1H NMR (DMSO-d6, d): 1.94 (m, 4H), 4.04 (t, 2H,
J = 6.0 Hz), 4.11 (t, 2H, J = 6.0 Hz), 4.96 (s, 2H), 6.91 (d, 2H,
J = 8.8 Hz), 6.95 (d, 2H, J = 8.4 Hz), 7.23 (d, 2H, J = 8.8 Hz), 7.44 (d,
1H, J = 16.0 Hz), 7.56 (d, 1H, J = 16.0 Hz), 7.59 (d, 1H, J = 8.0 Hz),
7.60 (s, 1H), 7.66 (d, 2H, J = 8.4 Hz), 7.72 (d, 1H, J = 1.6, 8.0 Hz).
ESI-MS: m/z [MꢀH]ꢀ 504.
5.6. General procedure for the synthesis of compounds 33a,b
5.7.2. (E)-4-(Cyanomethoxy)-3-(3-(4-(4-phenoxybutoxy)phenyl)
acryloyl)benzoic acid (27b)
A solution of acetophenone 30, corresponding benzaldehyde 29,
and 10% potassium hydroxide in a solvent of EtOH/THF/H2O (5:5:1,
v/v, 6 mL) was stirred at room temperature for 5–6 days. The mix-
ture was poured into ice-water, acidified to pH ꢂ5 with 2 N HCl,
and extracted with ethyl acetate (20 mL ꢃ three times). The organ-
ic phase was washed with brine, dried over anhydrous Na2SO4, and
then concentrated in vacuo to give pale yellow syrup, to which eth-
anol (10 mL) and sulfuric acid (two drops) was added. The reaction
mixture was refluxed for 4 h. Upon finishing of the reaction, water
was added, and the mixture was extracted with ethyl acetate
(20 mL ꢃ three times), then washed with brine and dried over
anhydrous Na2SO4. Evaporating of the combined organic layers
afforded the crude product as an oily residue which was purified
by flash column chromatography on silica gel using petroleum
ether/ethyl acetate (8:1, v/v) as eluant to give 33.
Reagent: compound 36b (100 mg, 0.22 mmol), chloroacetonitri-
le (16.4 mg, 0.22 mmol), potassium carbonate (60 mg, 0.44 mmol).
Product: pale yellow solid (49.1 mg, 48%), mp 173–176 °C. 1H NMR
(DMSO-d6, d): 1.92 (m, 4H), 4.03 (t, 2H, J = 6.0 Hz), 4.10 (t, 2H,
J = 6.0 Hz), 4.96 (s, 2H), 6.86 (t, 1H, J = 8.8 Hz), 6.95 (d, 2H,
J = 8.4 Hz), 6.98 (d, 2H, J = 8.0 Hz), 7.22 (d, 1H, J = 8.4 Hz), 7.30 (t,
2H, J = 8.8 Hz), 7.56 (d, 1H, J = 16.0 Hz), 7.59 (d, 1H, J = 16.0 Hz),
7.69 (d, 2H, J = 8.4 Hz), 8.05 (dd, 1H, J = 2.0, 8.4 Hz), 8.08 (d, 1H,
J = 8.4 Hz). ESI-MS: m/z [MꢀH]ꢀ 470.
5.8. Pharmacological evaluation of CysLT1 antagonistic
activities
5.8.1. Materials
5.6.1. (E)-Ethyl 4-(3-(4-(4-(4-chlorophenoxy)butoxy)phenyl)
acryloyl)-3-hydroxybenzoate (33a)
RPMI 1640 was purchased from Grand Island Biological Com-
pany (Gibco, USA). Fetal bovine serum was purchased from Shang-
hai Genetimes Technology, Inc. (China), and penicillin,
Reagent: acetophenone 30a (500 mg, 2.4 mmol) and benzalde-
hyde 29a (768 mg, 2.52 mmol). Product: pale yellow syrup
(380 mg, 32%). 1H NMR (CDCl3, 400 MHz, d): 1H NMR (CDCl3,
400 MHz, d): 1.26 (t, 3H, J = 6.0 Hz), 2.01 (m, 4H), 4.00 (t, 2H,
J = 6.0 Hz), 4.02 (t, 2H, J = 6.0 Hz), 4.11 (q, 2H, J = 6.0 Hz), 6.89 (d,
2H, J = 8.0 Hz), 6.93 (d, 2H, J = 8.8 Hz), 7.23 (d, 2H, J = 8.0 Hz),
7.52 (d, 1H, J = 16.0 Hz), 7.56 (dd, 1H, J = 1.2, 7.6 Hz), 7.64 (d, 2H,
J = 8.8 Hz), 7.67 (d, 1H, J = 7.6 Hz), 7.93 (d, 1H, J = 16.0 Hz), 7.98
(d, 1H, J = 1.2 Hz), 12.65 (s, 1H, OH). ESI-MS: m/z [M+H]+ 495.
streptomycin, L-glutamine and pluronic F-127 were purchased
from Invitrogen Co. (USA). LTD4, Fluo3/AM, dimethylsulfoxide,
and Hepes were purchased from Sigma Chem. Co (USA). All salts
for saline and Tris solution were purchased from Shanghai Sangon
Biological Engineering Technology and Services Co., Ltd (China).
Disposable culture flasks, petri dishes, and filters were purchased
from Corning Glass Works (USA). Montelukast was a kind gift from
Merck & Co., Inc. (USA).
5.6.2. (E)-Ethyl 4-hydroxy-3-(3-(4-(4-phenoxybutoxy)phenyl)
acryloyl)benzoate (33b)
5.8.2. Cell culture
U937 cells (ATCC) were routinely cultured into flasks in RPMI
1640 medium supplemented with 10% fetal bovine serum, 2 mM
Reagent: acetophenone 30b (500 mg, 2.40 mmol), benzaldehyde
29b (681 mg, 2.52 mmol). Product: pale yellow syrup (331 mg, 30%).
1H NMR (CDCl3, 400 MHz, d): 1.28 (t, 3H, J = 6.0 Hz), 2.01 (m, 4H),
4.02 (t, 2H, J = 6.0 Hz), 4.04 (t, 2H, J = 6.0 Hz), 4.11 (q, 2H,
J = 6.0 Hz), 6.93 (d, 2H, J = 8.0 Hz), 6.95 (m, 3H), 7.05 (d, 1H,
J = 8.0 Hz), 7.26 (d, 2H, J = 8.0 Hz), 7.62 (d, 1H, J = 16.0 Hz), 7.64 (d,
2H, J = 8.0 Hz), 7.95 (d, 1H, J = 16.0 Hz), 8.11 (dd, 1H, J = 1.2,
8.0 Hz), 8.63 (d, 1H, J = 1.2 Hz), 12.68 (s, 1H, OH). ESI-MS: m/z
[M+H]+ 461.
L-glutamine, 4.5 g/L D-glucose, 10 mM Hepes, 2 g/L NaHCO3,
1 mM sodium pyruvate, 100 U/mL penicillin, and 100 mg/mL
streptomycin at 37 °C (5% CO2) and differentiated for 4–5 days
with 1.3% dimethyl sulfoxide (DMSO).
5.8.3. Determination of cytosolic free Ca2+ levels
Briefly, dU937 Cells were washed twice with a balanced salt
solution (BSS in mM: NaCl 140, KCl 4, MgCl2 1, CaCl2 1.25, HEPES
5, glucose 11, NaH2PO4 1, ascorbic acid 5.7; pH 7.4) and incubated
5.7. General procedure for the synthesis of compounds 27a,b
in 1 mL BSS containing 10 lM Fluo3/AM and 0.03% pluronic F-127
for 45 min at 37 °C. After loading, Fluo3/AM was removed and cells
were washed twice with BSS. Then, cells were centrifuged, resus-
pended, diluted to the concentration of 106 cells/mL, and trans-
ferred to 96 wells plate. Tested compounds were added to the
To a solution of compound 33 and chloroacetonitrile in ace-
tone (20 mL), potassium carbonate was added. The resulting mix-
ture was refluxed for 5 h and then cooled to room temperature.
The mixture was filtered and the filtrate was concentrated in va-
cuo to give 34. And then 5% sodium hydroxide in EtOH/H2O (1:1,
v/v, 2 mL) was added to the residue. The reaction mixture was
stirred at room temperature for 5 h, then poured into cold water
and extracted with ethyl acetate. The organic phase was washed
with brine, dried over anhydrous Na2SO4, and then concentrated
in vacuo. The crude product was recrystallized in ethanol to give
27.
wells with 10 lL at the different concentration and were further
incubated for 30 min at 37 °C to complete the hydrolysis of
Fluo3/AM. Cells were transferred to the FLUOstar OPTIMA (BMG
Germany) and LTD4 was added to the wells at the concentration
of 0.1 l
M. Calcium concentration ([Ca2+]i) were measured as the
fluorescence monitored at 37 °C (485 nm excitation, 525 nm emis-
sion). The half maximum antagonistic concentration (EC50) was de-
fined as the concentration of the compounds that induced 50% of

X. Dong et al. / Bioorg. Med. Chem. 18 (2010) 5519–5527
10. Markham, A.; Faulds, D. Drugs 1998, 56, 251.
5527
maximum inhibition of [Ca2+]i elevated by LTD4 (0.1
lM) and was
11. Jacobs, R. T.; Brown, F. J.; Cronk, L. A.; Aharony, D.; Buckner, C. K.; Kusner, E. J.;
Kirkland, K. M.; Neilson, K. L. J. Med. Chem. 1993, 36, 394.
12. Yu, G. L.; Wei, E. Q.; Wang, M. L.; Zhang, W. P.; Zhang, S. H.; Weng, J. Q.; Chu, L.
S.; Fang, S. H.; Zhou, Y.; Chen, Z.; Zhang, Q.; Zhang, L. H. Brain Res. 2005, 1053,
116.
13. Labelle, M.; Belley, M.; Gareau, Y.; Gauthier, J. Y.; Guay, D.; Gordon, R.;
Grossman, S. G.; Jones, T. R.; Leblanc, Y.; McAuliffe, M.; McFarlane, C.; Masson,
P.; Metters, K. M.; Ouimet, N.; Patrick, D. H.; Piechuta, H.; Rochette, C.; Sawyer,
N.; Xiang, Y. B.; Pickett, C. B.; Ford-Hutchinson, A. W.; Zamboni, R. J.; Young, R.
N. Bioorg. Med. Chem. Lett. 1995, 5, 283.
14. Zwaagstra, M. E.; Schoenmakers, S. H. H. F.; Nederkoorn, P. H. J.; Gelens, E.;
Timmerman, H.; Zhang, M. J. Med. Chem. 1998, 41, 1439.
15. QSAR and Drug Design-New Developments and Applications; Terada, H., Goto, S.,
Hori, H., Taira, Z., Fujita, T., Eds.; Elsevier: Amsterdam, 1995. pp 341–367.
16. Palomer, A.; Pascual, J.; Cabre, F.; Garcia, M. L.; Mauleon, D. J. Med. Chem. 2000,
43, 392.
calculated from the concentration–response curve by nonlinear
regression (curve fit) using GraphPad Prism (Version 4.0).
Acknowledgments
This study was financially supported by the National Natural
Science Foundation of China (No. 30772561) and the National
Key Tech Project for Major Creation of New drugs (No.
2009ZX09501-003).
Supplementary data
17. Zwaagstra, M. E.; Timmerman, H.; Tamura, M.; Tohma, T.; Wada, Y.; Onagi, K.;
Zhang, M. J. Med. Chem. 1997, 40, 1075.
18. Guay, D.; Gauthier, J. Y.; Dufresne, C.; Jones, T. R.; McAuliffe, M.; MacFarlane, C.;
Metters, K. M.; Prasit, P.; Rochette, C.; Roy, P.; Sawyer, N.; Zamboni, R. Bioorg.
Med. Chem. Lett. 1998, 8, 453.
19. Zwaagstra, M. E.; Timmerman, H.; Abdoelgafoer, R. S.; Zhang, M. Eur. J. Med.
Chem. 1996, 31, 861.
20. Zwaagstra, M. E.; Korthouwer, R. E. M.; Timmerman, H.; Zhang, M. Eur. J. Med.
Chem. 1998, 33, 95.
21. Zwaagstra, M. E.; Timmerman, H.; Stolpe, A. C.; Kanter, F. J. J.; Tamura, M.;
Wada, Y.; Zhang, M. J. Med. Chem. 1998, 41, 1428.
22. Weiner, S. J.; Kollman, P. A.; Ngyuen, D. T.; Case, D. A. J. Comput. Chem. 1986, 7,
230.
23. Debnath, A. K. J. Med. Chem. 2002, 45, 41.
24. Capra, V.; Ravasi, S.; Accomazzo, M. R.; Parenti, M.; Rovati, G. E. Biochem.
Pharmacol. 2004, 67, 1569.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2010.06.047.
References and notes
1. Claesson, H. E.; Dahlen, S. E. J. Int. Med. 1999, 245, 205.
2. Samuelsson, B.; Dahlen, S. E.; Lindgren, J. A.; Rouzer, C.; Serhan, C. N. Science
1987, 237, 1171.
3. Lieberman, P. Allergy Asthma Proc. 2009, 30, 345.
4. Folco, G.; Rossoni, G.; Buccellati, C.; Berti, F.; Maclouf, J.; Sala, A. Am. J. Respir.
Crit. Care Med. 2000, 161, S112.
5. Sampson, A. P.; Pizzichini, E.; Bisgaard, H. J. Allergy. Clin. Immunol. 2003, 111,
S49. discussion S59–61.
6. Pergola, C.; Werz, O. Expert Opin. Ther. Patents 2010, 20, 355.
7. Coleman, R. A.; Eglen, R. M.; Jones, R. L.; Narumiya, S.; Shimizu, T.; Smith, W. L.
Adv. Prostaglandin Thromboxane Leukot. Res. 1995, 23, 283.
8. Metters, K. M.; Gareau, Y.; Lord, A.; Rochette, C.; Sawyer, N. J. J. Pharmacol. Exp.
Ther. 1994, 270, 399.
25. Frey, E. A.; Nicholson, D. W.; Metters, K. M. Eur. J. Mol. Pharmacol. 1993, 244,
239.
26. Capra, V.; Accomazzo, M. R.; Ravasi, S.; Parenti, M.; Macchia, M.; Nicosia, S.;
Rovati, G. E. Prostaglandins Other Lipid Mediat 2003, 71, 235.
9. Hamilton, A.; Faiferman, I.; Stober, P.; Watson, R. M.; O’Byrne, P. M. J. Allergy
Clin. Immunol. 1998, 102, 177.