3-D QSAR analysis of inhibition of murine soluble epoxide hydrolase (MsEH) by benzoylureas, arylureas, and their analogues

Article abstract of DOI:10.1016/S0968-0896(00)00198-X

Two hundred and seventy-one compounds including benzoylureas, arylureas and related compounds were assayed using recombinant murine soluble epoxide hydrolase (MsEH) produced from a baculovirus expression system. Among all the insect growth regulators assayed, 18 benzoylphenylurea congeners showed weak activity against MsEH. Newly synthesized cyclohexylphenylurea, 1-benzyl-3-phenylurea, and 1,3-dibenzylurea analogues were rather potent. The introduction of a methyl group at the para-position of the phenyl ring of cyclohexylphenylurea enhanced the activity 6-fold, though similar substituent effects were not seen for any of the benzoylphenylureas. The activities of these compounds, including several previously reported compounds, such as dicyclohexylurea, diphenylurea, and their related analogues (Morisseau et al., Proc. Natl. Acad. Sci., 1999, 96, 8849), were quantitatively analyzed using comparative molecular field analysis (CoMFA), a three-dimensional quantitative structure-activity relationship (3-D QSAR) method. Both steric and electrostatic factors contributing to variations in the activity were visualized using CoMFA. CoMFA results showed that one side of the cyclohexylurea moiety having a trans-amide conformation (A-ring moiety) is surrounded by large sterically unfavorable fields, while the other side of A-ring moiety and the other cyclohexyl group (B-ring moiety) is encompassed by sterically favored fields. Electrostatically negative fields were scattered around the entire molecule, and a positive field surrounds the carbon of the carbonyl group. Hydrophobic fields were visualized using Kellogg's hydropathic interaction (HINT) in conjunction with CoMFA. Hydrophobically favorable fields appeared beside the 4- and 4'-carbon atoms of the cyclohexyl groups, and hydrophobically unfavorable fields surrounded the urea bridge. The addition of the molecular hydrophobicity, it> /it>, to CoMFA did not improve the correlation significantly. The ligand-binding interactions shown by X-ray crystallographic data were rationalized using the results of the CoMFA and HINT analyses, and the essential physicochemical parameters for the design of new MsEH inhibitors were disclosed. Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(00)00198-X

Bioorganic & Medicinal Chemistry 8 (2000) 2663±2673
3-D QSAR Analysis of Inhibition of Murine Soluble
Epoxide Hydrolase (MsEH) by Benzoylureas, Arylureas,
and their Analogues
Yoshiaki Nakagawa,a,* Craig E. Wheelock,a,b Christophe Morisseau,b
Marvin H. Goodrow,b Bruce G. Hammockb and Bruce D. Hammockb
aDivision of Applied Life Sciences, Graduate School of Agriculture, Kyoto University, Kyoto 606-8502, Japan
bDepartment of Entomology, University of California, Davis, CA 95616, USA
Received 1 June 2000; accepted 18 July 2000
AbstractÐTwo hundred and seventy-one compounds including benzoylureas, arylureas and related compounds were assayed using
recombinant murine soluble epoxide hydrolase (MsEH) produced from a baculovirus expression system. Among all the insect
growth regulators assayed, 18 benzoylphenylurea congeners showed weak activity against MsEH. Newly synthesized cyclohex-
ylphenylurea, 1-benzyl-3-phenylurea, and 1,3-dibenzylurea analogues were rather potent. The introduction of a methyl group at the
para-position of the phenyl ring of cyclohexylphenylurea enhanced the activity 6-fold, though similar substituent e€ects were not
seen for any of the benzoylphenylureas. The activities of these compounds, including several previously reported compounds, such
as dicyclohexylurea, diphenylurea, and their related analogues (Morisseau et al., Proc. Natl. Acad. Sci., 1999, 96, 8849), were
quantitatively analyzed using comparative molecular ®eld analysis (CoMFA), a three-dimensional quantitative structure±activity
relationship (3-D QSAR) method. Both steric and electrostatic factors contributing to variations in the activity were visualized
using CoMFA. CoMFA results showed that one side of the cyclohexylurea moiety having a trans-amide conformation (A-ring
moiety) is surrounded by large sterically unfavorable ®elds, while the other side of A-ring moiety and the other cyclohexyl group (B-
ring moiety) is encompassed by sterically favored ®elds. Electrostatically negative ®elds were scattered around the entire molecule,
and a positive ®eld surrounds the carbon of the carbonyl group. Hydrophobic ®elds were visualized using Kellogg's hydropathic
interaction (HINT) in conjunction with CoMFA. Hydrophobically favorable ®elds appeared beside the 4- and 40-carbon atoms of
the cyclohexyl groups, and hydrophobically unfavorable ®elds surrounded the urea bridge. The addition of the molecular hydro-
phobicity, it> /it>, to CoMFA did not improve the correlation signi®cantly. The ligand-binding interactions shown by X-ray
crystallographic data were rationalized using the results of the CoMFA and HINT analyses, and the essential physicochemical
parameters for the design of new MsEH inhibitors were disclosed. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
important enzymes due to their ability to detoxify
mutagenic, toxic, and carcinogenic xenobiotic epox-
ides.1,2 In addition mEH has been hypothesized to have
an endogenous role in bile acid transport,4,5 while sEH
is potentially involved in the metabolism of oxidized
lipids as autacoids or chemical mediators.6
Epoxide hydrolases (EHs, EC 3.2.2.3) are found in both
animals and plants, and play a role in regulating the
metabolism of hormones and secondary metabolites in
biological systems.1 3 Among various EHs, microsomal
EH (mEH) and soluble EH (sEH) are particularly
We have shown that leukotoxin (linoleic acid epoxide) is
not directly cytotoxic to cells, but that the diol metabo-
lite is toxic in all cells assayed to date, including pul-
monary alveolar epithelia cells.6 Therefore, powerful and
selective inhibitors of sEH could be useful for studying
sEH as pharmaceutical target. To increase our under-
standing of this enzyme, we have recently determined the
X-ray structure of inhibitor bound murine sEH (MsEH)
and proposed an inhibitory mechanism of action.7,8 To
further extend our knowledge of this enzyme, it is
Abbreviations: CPU, N-cyclohexyl-N0-(3-phenyl)propylurea; CIU, N-
cyclohexyl-N0-(4-iodophenyl)urea; CDU, N-cyclohexyl-N0-decylurea;
CoMFA, comparative molecular ®eld analysis; CoMSIA, comparative
similarity indices analysis; 3-D QSAR, three-dimensional quantitative
structure±activity relationship; DCU, dicyclohexylurea; HINT,
hydropathic interaction; MsEH, murine soluble epoxide hydrolase;
SAR, structure±activity relationship.
*Corresponding author. Tel.: +81-75-753-6117; fax: +81-75-753-6123;
e-mail: naka@kais.kyoto-u.ac.jp
0968-0896/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00198-X

2664
Y. Nakagawa et al. / Bioorg. Med. Chem. 8 (2000) 2663±2673
important to model the essential physicochemical factors
involved for inhibitors which demonstrate high activity.
These factors are not only necessary for the basic
understanding of the inhibitory mechanism, but also for
the design of more potent inhibitors.
inhibitors in insects,11 13 and some of them have appli-
cations for agricultural use. These compounds can also
be used in the dairy industry for controlling ¯ies and
mosquitoes. Although benzoylphenylureas inhibit chitin
synthesis, they have no e€ect on insect chitin
synthase.11 13 Di¯ubenzuron, a benzoylphenylurea, eli-
cited several e€ects in cell free systems: inhibition of
calcium uptake14 and the activation of protein phos-
phorylation15,16 in a cell-free preparation of the Amer-
ican cockroach (Periplaneta americana), inhibition of
chitin synthase in brine shrimp,17 and activation of ser-
ine proteases.18 Many hypotheses have been proposed
regarding the mode of action of benzoylphenylureas;
however, the actual mechanism remains unknown.19,20 In
addition, some benzoylphenylurea congeners showed
antitumor activity,21,22 antimitotic e€ects,23,24 and the
stimulation of hematopoesis.25
Recent studies have shown chalcone oxides (I; Fig. 1) to
be potent inhibitors of sEH.9 These compounds contain
a carbonyl group beta to the epoxide and essentially act
as poor substrates, thus their ability to inhibit the
enzyme. However, we have found that dicylohexylurea
(DCU, II; Figure 1) and several urea and carbamate
analogues were much more potent inhibitors of sEH.
This group of inhibitors acts through a transition state
analogue mechanism with Asp333 in the sEH active site
attacking the carbonyl carbon of the urea moiety. We
therefore combined the components of the 2 active
groups of inhibitors to test benzoylphenylureas (III;
Figure 1) as inhibitors. This class of compounds con-
tains the potent urea moiety as well as the carbonyl
group beta to the site of nucleophilic attack as in the
chalcone oxides.
The aim of this study was 3-fold. The ®rst aim was to
determine the e€ect of these insect growth regulators on
MsEH, since they contain the essential components for
inhibitory activity. The second aim was to examine the
structure±activity relationship (SAR) for the inhibition
of MsEH by several ureas, carbamates, and their ana-
logues. Thirdly, we wanted to visualize the ligand±
enzyme interactions using computer graphics techniques
in the hope of being able to predict the essential qualities
necessary for the synthesis of more potent inhibitors.
Essential physicochemical properties were extracted,
including electrostatic, steric, and hydrophobic interac-
tions. We used comparative molecular ®eld analysis
(CoMFA),26 a three-dimensional quantitative struc-
ture±activity relationship (3-D QSAR) technique, as
well as CoMFA combined with hydropathic interaction
(HINT) analysis,27 to model the key interactions of
MsEH. Enzyme assays were performed using recombi-
nant MsEH produced in a baculovirus expression sys-
tem as previously reported.9,28,29 Using this CoMFA
model we can predict compound activity, and design
potent sEH inhibitors, thereby increasing our under-
standing of the enzyme as well as developing new tools
for use as a potential pharmaceutical target.
In addition to the benzoylphenylureas, several of these
compounds have commercial applications, with urea
and carbamate moieties being very common in pesti-
cides and medicines. Several urea containing pesticides
have been shown to be potent inhibitors, including
diuron and siduron,10 as well as the carbamates such as
phenoxycarb and carbaryl. It is therefore important to
have an understanding of how these compounds interact
with their enzymatic target. Benzoylphenylureas and
related compounds are well-known chitin synthesis
Minimized conformations and superpositions between
compounds
Modeling and calculation experiments were executed
using SYBYL (ver. 6.5.3; Tripos Co., St. Louis, MO,
USA). The X-ray crystallographic coordinates of 1-(4-
chlorophenyl)-3-(2,6-di¯uorobenzoyl)urea (18, di¯ubenz-
uron)30 were applied in the construction of the initial
conformations of the urea derivatives and related com-
pounds. The X-ray data were downloaded from Cam-
bridge data base, and the benzoylphenylurea basic
structure is sketched as shown in structure III of Figure 1.
The phenyl moiety of benzoylphenylurea is attached to
the trans-amide moiety and the benzoyl group to the cis-
amide.
The structures of DCU (35) and diphenylurea (44) were
constructed by replacing the benzoyl and phenyl groups
of di¯ubenzuron (18) with the corresponding cyclohexyl
or phenyl groups. To distinguish the two cyclohexyl
Figure 1. Structures of chalcone oxides (I), DCU (II), and benzoyl-
phenylureas (III) derived from the X-ray structure of di¯ubenzuron
obtained from the Cambridge data base.

Y. Nakagawa et al. / Bioorg. Med. Chem. 8 (2000) 2663±2673
2665
groups of DCU the cyclohexyl group attached to the
trans-amide is labeled as the A-ring, and the other
cyclohexyl group as the B-ring, as shown in structure II
of Figure 1. The most favorable torsion angles for the
-C(Ar)-C(Ar)-N-H portion at the aniline moiety (C6H5-
NH-) of diphenylurea (44) and the corresponding cyclo-
hexyl moiety (C6H11-NH-) of dicyclohexylurea (35) were
determined by executing a Grid Search procedure, a
SYBYL submodule which systematically searches and
minimizes every conformation using molecular
mechanics (MM2). The search was performed with an
increment of 30 degrees, thus producing 12 conformers
from which the energetically minimized conformation
had to be determined. The structure of compound 34
was built from compound 35. Other dicyclohexyl analo-
gues (36±39) were constructed by changing the nitrogen
or oxygen atom of compound 35 to the corresponding
atom. Compounds 16, 17 and 40±43 were constructed
from compound 35 by replacing the cyclohexyl group
with the corresponding moieties. Compound 45 was
derived from compound 44 by replacing both N atoms
with C atoms, and compound 46 was constructed from
compound 45 by substituting a C atom with an O atom.
Compounds 49 and 50 were constructed by attaching
the necessary CH3 group to compound 44. The most
favorable torsion angle between the OPh group and the
phenyl ring of compounds 2, 3, and 22 was determined
using a Grid Search as described above. The ®nal con-
formations and the atomic charges of all compounds
were minimized by the semi-empirical molecular orbital
methods PM3 and MNDO, respectively. According to
the X-ray structures of N-cyclohexyl-N0-(3-phenyl)pro-
pylurea (CPU),7 N-cyclohexyl-N0-decylurea (CDU) and
N-cyclohexyl-N0-(4-iodophenyl)urea (CIU)8 each cyclo-
hexyl group is attached to the trans-amide correspond-
ing to the A-ring in structure II (Fig. 1). Therefore, the
cyclohexyl groups of compounds 4, 5, 16, 17, 40, 42, 43
and 51 were placed as the A-ring.
automatically prepared in the CoMFA module. The
electrostatic and steric potential energies at each lattice
point were calculated using Coulombic and Lennard-
Jones potential functions, respectively. The hydro-
phobic e€ects were examined using either HINT terms
or it> /it> values as an additional independent vari-
able. The default setting was used for the calculation of
CoMFA and HINT lattice variables. The correlations
of the biological activity index with the lattice variables
and it> /it> were analyzed by the partial least squares
(PLS) method with a column ®ltering setting of 2 kcal
mol 1. The CoMFA results were represented by the
predictive r2 value which was renamed as q2,31 and
standard errors of cross-validated predictions, spress, as
well as the number of optimum components, m, from
the leave-one-out cross-validation analyses. In addition
the conventional correlation coecient, r, standard
deviation, s, and the weight percentage of the type of the
descriptors participating in the correlation were pro-
duced from the analysis. The results were visualized by
color contour maps, which display both the favorable
and unfavorable ®elds for each electrostatic, steric and
hydrophobic interaction surrounding a set of super-
posed molecules.
Results
Inhibition of sEH
Two hundred and seventy-one substituted benzoylphenyl-
urea analogues including 2,6-di¯uorobenzoyl, 2,6-
dichlorobenzoyl, and 4-chlorophenyl analogues, and
related compounds including thiadiazoles, oxalyldiani-
lides, N,N-dimethyl phenyl carbamates, and dibenzoyl-
hydrazines were assayed against MsEH. Among the
compounds tested, only 18 benzoylphenylureas (18±33)
were signi®cant inhibitors of MsEH. Their pIC50 values
are listed in Table 2, although their activity is only
about 1/1000 that of the potent inhibitors DCU (35)
and dicyclohexylcarbamate (36). As shown in Table 2,
highly pure di¯ubenzuron (18) synthesized for this study
and puri®ed through repeated crystallizations was inac-
tive (pIC50<4.00); however, the industrial grade
obtained from a company demonstrated some activity
(pIC50=4.57), indicating the presence of unknown con-
taminant(s). Other acylurea analogues tested did not
give 50% inhibition, even at concentrations up to
100 mM.
DCU (35) was used as a template for the superposition
of other related compounds using from 5 to 8 atoms and
a root mean square (RMS) ®tting method. All struc-
tures were superimposed as displayed in Tables 1±3, in
which ®tting atoms are shown in bold type. In most cases
the 8 atoms of the urea moiety (-C-NH-CO-NH-C-) were
used for the superposition. In cases where there was no
common ®tting N atom (36±38, 45±48), the corre-
sponding carbon or hetero atom was used. For com-
pounds 34, 36±38, 45±47, 49, 50 and 52, an H atom was
not used in the superpositions due to the lack of an -NH
hydrogen atom. In the case of the thiourea analogues in
compounds 13±15 and 39, an S atom was used instead
of an O atom for the superposition.
Newly synthesized benzoylureas (1±9), having substituted
phenyl, cyclohexyl, cyclododecyl, or benzyl groups as
the aniline moiety, were inactive. However, three com-
pounds 10±12 containing benzyl groups in place of the
benzoyl group were active inhibitors. Two benzylphe-
nylureas (10 and 11) were equipotent to diphenylurea
(44), and a dibenzylurea (12) was slightly lower. By
replacing a phenyl group of diphenylurea with a cyclo-
hexyl group, the activity was enhanced three-fold (16
versus 44). Furthermore, the introduction of a CH3
group at the para-position of the benzene ring increased
the activity about 5 times. Three benzoylthioureas (13±
15) demonstrated some activity.
HINT±CoMFA
The analyses were carried out using the QSAR module
of SYBYL 6.5.3 and HINT 2.30S (eduSoft, Ashland,
VA, USA). The lattice spacing was 2 A, a+1 charge was
used to estimate the electrostatic ®eld, and an sp3 carbon
was used to estimate the steric molecular ®elds. The
required region to accommodate the structures which
were superposed according to the above scheme was

2666
Y. Nakagawa et al. / Bioorg. Med. Chem. 8 (2000) 2663±2673
Table 1. Observed and calculated pIC50 values for inhibition of
MsEH by newly synthesized benzoylureas, benzylureas, and phenyl-
ureas
Table 2. Observed and calculated pIC50 values for inhibition of
MsEH by benzoylphenylureas and their logP values.
pIC50 (M)
No.
Structurea
Obsd. Calcdb logPc HINT
logPd
pIC50 (M)
No.
Yn
Xn
Obsd. Eqn (1) logP HINT
logP
1
<4.00
<4.00
<4.00
<4.00
<4.00
<4.00
<4.00
<4.00
<4.00
5.62
4.51
4.22
3.97
5.82
6.05
5.17
4.94
4.98
5.03
5.22
5.48
6.39
3.30
4.22
3.93
4.84
3.04
3.95
2.87
3.88
2.79
3.37
12.74
12.92
13.63
4.82
5.53
7.79
8.38
4.20
4.79
3.54
4.13
1.55
11.14
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
4-Cl
4-n-C3H7
4-n-C4H9
4-n-C8H17
4-OC6H5
4-CN
2,6-F2
2,6-F2
2,6-F2
2,6-F2
2,6-F2
<4.00a
4.24
4.32
4.11
4.06
4.25
4.14
4.18
3.93
4.04
4.68
4.07
4.37
4.12
4.15
4.17
4.34
4.23
4.23
4.33
4.02
4.26
4.03
3.87
4.15
4.17
4.40
3.98
4.19
4.19
4.57
4.41
3.88b 11.47
4.34c 12.50
4.86c 13.04
6.85c 15.20
4.93c 13.08
3.06c 10.31
4.83c 12.75
4.25c 11.26
4.86c 13.64
5.38c 14.18
6.37c 15.26
3.27c 12.49
2
3
2,6-F2
2,6-F2
4-C6H5
4-Cl
2-N(CH3)2
2,6-Cl2
2,6-Cl2
2,6-Cl2
2,6-Cl2
2,6-Cl2
2,6-Cl2
2,6-Cl2
2,6-Cl2
4
4-n-C3H7
4-n-C4H9
4-n-C6H13
4-OC2H5
4-SO2N(C2H5)2
4-SO2N(CH2)4
2-Cl
5
4.36d
3.60d
6.90
7.09
3.64d 12.61
3.30d 12.61
6
2-NO2
7
aThe value evaluated for industrial grade of compound (TH-6040) is
4.57.
8
bExperimentally measured in the shake-¯ask method (ref 48).
cCited from ref 44.
9
dNewly estimated according to the empirical method (refs 49 and 50).
10
11
12
13
5.52
(5.53)
order to assess the role of hydrophobicity, CoMFA
analyses were executed with the addition of the mole-
cular hydrophobicity parameter logP, giving eqn (2) or
(3). However, this addition did not signi®cantly improve
the correlation as shown in Table 4. Equation (3) was
formulated using 6 components (m), because the opti-
mum component number was determined to be 6 in the
cross-validation analysis. The calculated activity values
from eqn (1) are listed in Tables 1±3 and the relationship
between observed pIC50 values and calculated values is
depicted in Figure 3. Figure 4 shows the overlay of
di¯ubenzuron (18) and DCU (35) in magenta with the
major CoMFA steric and electrostatic potential contour
maps, drawn according to eqn (1). The green areas in
Figure 4(a) indicate ®elds where submolecular bulk is well
accommodated with an increase in the activity, whereas
the yellow indicates ®elds where the submolecular bulk is
unfavorable for activity. The red areas in Figure 4(b)
indicate ®elds where negative electrostatic interaction with
the receptor binding site increases the activity, whereas the
blue areas show ®elds where the opposite is the case.
5.74
5.33
(5.75)
5.27
5.37
(5.20)
4.06
4.35
4.24
14
4.34
1.34
11.14
15
16
17
4.31
6.12
6.89
4.03
2.72
3.13
3.63
11.33
4.16
4.82
6.53
(6.19)
6.75
(6.80)
aFitting atoms used for superpositions are drawn as bold type in each
structure.
bCalculated by eqn (1) in Table 4. Values in parentheses were calcu-
lated by eqn (5) in Table 4.
cEmpirically estimated from the analogous compounds according to
our previous publications (refs 44, 48±50).
As drawn in Figure 4(a), the large sterically unfavorable
®elds (yellow areas) are found above the A-ring and urea
bridge moiety, whereas three green ®elds showing the
sterically favorable ®elds are located around the A-ring
and an additional ®eld is located beside the cyclohexyl
group of the B-ring. In Figure 4(b), negative electro-
static-potential ®elds (red areas) are spread around the
whole molecule and a positive ®eld (blue area) is found
below the carbonyl carbon atom in the urea moiety.
dCalculated by HINT program developed by Kellogg et al. (ref 27).
3-D QSAR
All compounds with pIC50 < 4.00 were not included in
the QSAR analyses due to the lack of their activity in
terms of pIC50. In the following QSAR analyses com-
pound 39 was not included, because it deviated drama-
tically from the correlation equation. Equation (1) in
Table 4 was obtained from the CoMFA analysis of 37
compounds (Tables 1±3). The alignments of the 37 com-
pounds used to derive eqn (1) are shown in Figure 2. In
In order to visualize the partial hydrophobic ®elds in
the CoMFA analysis, we used the HINT procedure27
along with CoMFA.26 The polar proximity used for

Y. Nakagawa et al. / Bioorg. Med. Chem. 8 (2000) 2663±2673
2667
Table 3. Observed and calculated pIC50 values for ureas and their
analogues, and their logP values
the calculation of HINT lattice variables was calculated
using the option ``Via Bond'', which was performed by
the Leo±Hansch method of logP calculation.32,33 How-
ever, some logP values calculated using HINT were
signi®cantly di€erent from the experimentally/empiri-
cally measured logP values or resulted in warning mes-
sages following the HINT calculation. We therefore
omitted these 21 compounds (13±15, 19±33, 38, 44, 49)
from the compound set for eqn (1) in the HINT±
CoMFA analysis. Equation (4) was derived for this small
set of compounds (n=16), and the quality in the cross-
validation analysis was worse than that of eqn (1). The
steric and electrostatic ®elds obtained for this small set of
compounds (n=16) are not greatly di€erent from that
obtained for the 37 compounds in eqn (1). CoMFA±
HINT analysis was performed for this small set of com-
pounds to derive eqn (5), and the hydrophobic ®elds are
visualized in Figure 4(c). The hydrophobically favorable
®elds (orange areas) are found around both cyclohexyl
moieties, and hydrophobically unfavorable ®elds (cyan
blue areas) are located around the urea bridge.
pIC50
No.
Structuresa
Obsd.
Calcdb
logP
HINT
logP
34
35
6.60
7.05
6.66
(6.75)
6.95
4.38
3.51
2.70
2.91
(7.11)
36
37
38
7.30
6.24
5.42
6.77
(7.05)
3.90
3.60
4.69
3.64
3.77
3.78
6.66
(6.34)
5.89
6.77
39
4.00c
3.79
3.45
40
41
42
43
44
5.80
<3.30
7.22
6.30
(5.81)
2.54
1.57
3.96
4.14
3.00
1.82
0.61
5.70
(4.50)
All compounds except for compound 39 were well pre-
dicted (r2=0.955) by eqn (1), as shown in Figure 3. If
compound 39 was included in the CoMFA analysis the
conventional correlation was poor (cross-validated
q2=0.567, spress=0.837, m=3, r2=0.873, s=0.454, F=
77.601). The activity of the benzoylphenylureas (18±33)
listed in Table 2, as well as the newly synthesized inac-
tive benzoylphenylureas (1±3), was roughly predicted
by this QSAR model. However, newly synthesized
benzoylureas (4±7) were not rationalized by this current
mode. Previously reported inactive compounds 47, 48
and 509 were also predicted to have low activity, having
similar potency as the benzoylphenylureas, while com-
pound 41 was not predicted. The use of additional
compounds should serve to increase the accuracy of this
equation. Although the cyclohexyl group for com-
pounds 4 and 5 was placed on the left hand side (A-ring
moiety) in the superposition, it may ¯ip, resulting in the
placement of the cyclohexyl group on the right (B-ring).
7.05
(7.12)
3.37
7.22
6.96
(7.34)
3.63
5.64
5.61
11.52
45
46
47
3.37
3.37
3.33
(3.31)
3.13
3.34
3.28
2.86
4.68
3.63
3.17
(3.40)
<3.30
4.47
4.12
3.96
3.94
48
49
50
51
<3.30
3.71
0.05
2.75
3.20
3.65
2.50
10.33
9.02
<3.30
6.80
6.44
(6.84)
4.60
According to our unpublished data, most of the
thiourea analogues had very little or no activity, but
three thiourea analogues (13±15) were reasonably
included in eqn (1). The activity of compound 39 having
a CˆS group was predicted to be rather high (6.77),
being similar to that (7.05) of DCU (35) however; the
observed activity was 4.00. Compounds 14 and 15 were
utilized to synthesize the thiadiazole analogues whose
52
3.91
3.96
(3.99)
2.68
3.88
aFitting atoms used for superpositions are drawn as bold type in each
structure.
bCalculated by eqn (1) in Table 4. Values in parentheses were calculated by
eqn (5) in Table 4.
cThis value is not included in the QSAR analysis.
Figure 2. The superposition of the 37 compounds used for the CoMFA analysis.

2668
Y. Nakagawa et al. / Bioorg. Med. Chem. 8 (2000) 2663±2673
Table 4. CoMFA correlation statistics for two compounds set (n=16 and 37)
Cross-validation
Conventional correlation
Contribution (%)a
N
q2
spress
m
r2
s
F
EL
ST
H
Eqn
37
37
37
16
16
0.750
0.726
0.745
0.634
0.636
0.650
0.680
0.678
0.938
0.982
4
4
6
4
5
0.955
0.930
0.968
0.983
0.994
0.275
0.345
0.371
0.200
0.126
170.724
105.695
90.234
163.406
330.981
42.6
40.3
41.4
32.6
22.6
57.4
55.6
53.2
67.4
42.2
Ð
4.1b
5.4b
Ð
1
2
3
4
5
35.2
aContribution of electronic (EL), steric (ST), and hydrophobic e€ects (H) in the calculation.
bMolecular hydrophobicity logP was used as the independent parameter in the analysis.
well, and the substituents (including H) at the 4- and 40-
positions extend to the sterically favorable ®elds (Fig.
4(a); the structure of chalcone oxide is not shown).
However, the epoxide oxygen should correspond to the
carbonyl carbon of DCU in this superposition accord-
ing to our previous mechanistic model in which the
nucleophilic attack by the COO- residue of Asp is per-
formed at the a-carbon of the chalcone in the catalytic
site.35 Thus, the A-ring moiety of DCU probably cor-
responds to the left phenyl ring containing the 4-sub-
stituted R group of chalcone oxide.
The superposition of chalcone oxide (observed
pIC50=5.5435) to DCU by overlapping the B-ring moi-
ety of structure II on the benzoyl moiety as described
above gives a similar prediction (calculated pIC50=5.95)
to that obtained from the reversed overlapping in which
the benzoyl moiety was superposed on the A-ring moi-
ety (calculated pIC50=4.98), although the signs of
deviations are di€erent between them. In addition, as
we reported previously, the hydrophobic moieties at the
4-position strengthened the inhibition, while neither
electronic e€ects nor the hydrophobicity of the 40-posi-
tion appeared to in¯uence the inhibitory activity.9 Mul-
lin and Hammock also determined that the inhibition is
dominantly a€ected by hydrophobic terms and modu-
lated by 4-position molecular refractivity and 40-steric, but
not electronic, parameters.38 Hydrophobically favorable
®elds were found to encompass both A- and B-ring moi-
eties as shown in Figure 4(c), which is partially con-
sistent with the QSAR result for chalcone oxides. To
make clear the overlapping between chalcone and DCU,
the CoMFA analysis for the combined set of com-
pounds including the chalcone oxide derivatives as well
as other recently designed compounds is in progress.
Figure 3. Plot of observed activity values for inhibition of murine sEH
versus the calculated values using eqn (1). All values are shown to
inhibit 50% of the enzyme (IC50). (*) Indicates compounds included
in the correlation analysis (n=37); (*) indicates excluded compounds
(industrial grade of di¯ubenzuron 18 and a thiourea 39).
mode of action is similar to that of benzoylphenylur-
eas.34 Although 16 thiadiazole analogues with various
substituents at both benzene rings were assayed, none of
them were active. These results suggest that metabolites
which are derived by the opening of the thiadiazole ring
may have inhibitory e€ects on MsEH. In fact these
compounds are the precursors of their corresponding
benzoylamino thiadiazole analogues, as shown below
(Experimental).
The sterically unfavorable region encompassing the
cyclohexyl moiety of DCU might be due to steric repul-
sion occurring between the inhibitor and amino acid resi-
dues in the active site which contain a benzene ring, such
as Tyr381, Tyr465, or Trp334.7 The hydrophobically
unfavorable region surrounding the urea bridge may
correspond to the region accommodating His523 and
Asp 333. These amino acid residues contain imidazoli-
dine and carboxylic groups which could potentially
form hydrogen bonds with the inhibitor.
Discussion
The mechanism of murine and human sEH inhibition
by chalcone oxide derivatives has been analyzed in
detail through both a kinetic approach and QSAR ana-
lysis.35 The results of QSAR analyses of the substituent
e€ects on the activity of chalcone congeners (I), having
various phenyl substituents, demonstrated that the most
sterically favorable moieties at the 4- and 40-positions
are n-pentyl (or phenyl) and n-propyl groups, respec-
tively. Five bridge atoms (the a, b, and g carbons as well
as the carbons to either side) of chalcone oxide (I:
R=R0=H) were ®tted to the -C-N-C-N-C- backbone of
DCU (35). These two molecules superposed each other
In addition to steric e€ects, hydrophobicity appears to be
important as described above. This trend was evidenced
by several things: the activity of siduron (51), which has

Y. Nakagawa et al. / Bioorg. Med. Chem. 8 (2000) 2663±2673
2669
Figure 4. Stereoviews of CoMFA according to eqn (1) or eqn (5) with di¯ubenzuron (18) and DCU (35) colored with magenta. (a) The contours are
shown to surround regions where a higher steric bulk increases (green) or decreases (yellow) the inhibition. (b) The contours are shown to surround
regions where a positive (blue) and negative (red) electrostatic potential increases the inhibition. (c) The contours are shown to surround the regions
where hydrophobic (orange) and hydrophilic (cyan blue) interactions increase the inhibition. The contour maps were drawn using the favored (80%)
and disfavored (20%) contributions.
a CH3 group on the cyclohexyl ring, was more potent
than that of cyclohexylphenylurea (16). Additionally,
the introduction of a CH3 group on the benzene ring of
cyclohexylphenylurea also resulted in increased activity,
suggesting the importance of hydrophobic e€ects on the
inhibition of sEH. However, the addition of molecular
hydrophobicity (logP) did not signi®cantly improve the
correlation as shown in eqns (2) and (3) of Table 4. In
QSAR studies examining enzyme inhibition and receptor
binding, position-speci®c hydrophobicity is sometimes
an important parameter not only for benzoylphenylur-
eas,36 but for some other compounds as well.33 A pre-
requisite for the successful binding of a small molecule
to a macromolecule, such as a receptor or enzyme, is a
negative Gibbs free energy of binding. The target prop-
erty to be correlated and predicted in a comparative
analysis is a free energy value, and the Gibbs energy
change (ÁG) depends on both the enthalpy change
(ÁH) and the entropy change (ÁS).33,37 Since the addi-
tion of the HINT term, not the molecular hydro-
phobicity logP, improved the correlation as shown in
Table 4, the partial hydrophobic e€ects are probably
related to the entropic contribution between a ligand
molecule and a putative receptor. In fact, it was shown
that an N-cyclohexyl group attached to a trans-amide
structure packs well in the hydrophobic pocket and makes
numerous van der Waals contacts with hydrophobic resi-
dues Val497, Phe406, Phe265, Trp524, and Tyr381.7,8
In eqn (5) (graphically shown in Figure 4(c)), the
hydrophobic interaction at both the A- and B-ring
moieties in structure II appears to be important. In eqns
(4) and (5), 21 compounds were excluded from the set of
compounds used to derive eqn (1), because HINT was
unable to accurately calculate their logP values as listed
in Tables 1±3. Another 3-D QSAR method, comparative
similarity indices analysis (CoMSIA),39 was also applied
in an attempt to determine the hydrophobic e€ects for
the total set of compounds (n=37). However, the results
with CoMSIA were not improved over those of CoMFA
(data not shown). The poor prediction of compounds 4±
9 and 41 is probably due to the fact that the hydro-
phobic factor was not considered in eqn (1). The pre-
dicted value (pIC50=5.70) of compound 41 by eqn (1)
was 16 times higher than that (pIC50=4.50) predicted
by eqn (5), suggesting that both position-speci®c steric
and hydrophobic e€ects as well as electronic e€ects are
important for activity. The prediction of activity for
compounds 4±9 by eqn (5) appears to be meaningless
due to the lack of con®dence in the HINT logP values.
The strength of the hydrophobic interactions has impli-
cations for the energetics of protein folding, substrate
binding, and nucleic acid base stacking.37
The electronic e€ects of the 4-substituents of chalcone
oxides (I) on the enzyme reaction rate were analyzed
using a Hammet plot, in which s+ was applied, and the

2670
Y. Nakagawa et al. / Bioorg. Med. Chem. 8 (2000) 2663±2673
coecient of the sigma term in the correlation analysis
was 0.66. This value indicates the development of a
positive charge and excludes the formation of a true car-
bonium ion, being consistent with previous suggestions.
Electrostatically negative ®elds are located above the
carbonyl oxygen and a positive ®eld is found beneath
the carbonyl carbon atom of the urea, indicating that
the electrostatically positive charge of the carbonyl car-
bon atom which is attacked by COO of Asp of MsEH
is important to the activity.
activity, even though weak activity can be observed
sometimes. According to the 3-D structures, this car-
bonyl group may interfere with the nucleophilic attack
of Asp333 at the urea carbonyl carbon. However, in the
case of the chalcone oxide analogues, the carbonyl oxygen
is ``cis'' to the epoxide oxygen, which enhances the
inhibitory activity.9
In this QSAR model only compound 39, containing a
thiourea structure, is not included in the analysis, even
though three benzoyl thiourea analogues 13±15 were
reasonably explained by this model. In order to further
understand the favorable e€ect of thiourea, the synth-
esis of other thiourea analogues such as N-cyclohexyl-
N0-phenylthiourea and N,N0-diphenylthiourea is in pro-
gress. The activity of industrial grade of di¯ubenzuron
will also be included in the analysis as shown in Figure 3.
However, we have to account for the potential existence
of impurities in commercially acquired samples, since the
presence of trace contaminants can strongly a€ect
enzyme activity. For example, the presence of even 1%
of DCU in the industrial batch of di¯ubenzuron could
explain the noted inhibitory activity (pIC50=4.57),
whereas synthetically prepared pure di¯ubenzuron was
inactive. Based on this 3-D-QSAR model new sEH
inhibitors are being designed, and their synthesis and
bioassay are in progress.
The major steric, electrostatic, and hydrophobic inter-
actions for the inhibition of sEH are summarized in
Figure 5. As shown in Figures 4 and 5 the sterically
favorable ®elds overlap the hydrophobically favorable
®elds, and electrostatically positive and negative ®elds
are accommodated within the hydrophobically unfa-
vorable ®elds. Hansch±Fujita type classical QSAR40
studies would provide further understanding of the
ligand±enzyme interactions surrounding the cyclohexyl
groups. These studies are currently in progress. This will
enable us to separate steric and hydrophobic e€ects as
well as to distinguish electronic e€ects. By examining a
series of cyclohexylurea congeners with various sub-
stituents in o-, m-, and p-positions we should be able to
discern position-speci®c interactions.
In our recent study, the distribution of partial charge in
the reaction coordinate of the opening of the epoxide
ring was demonstrated by analyzing the X-ray struc-
tures of two alkylurea inhibitors, N-cyclohexyl-N0-
decylurea (CDU) and N-cyclohexyl-N0-(4-iodophenyl)-
urea (CIU), complexed with MsEH.8 In these com-
plexes, a constellation of hydrogen bonds that activate
epoxide substrates for hydrolysis was observed, while it
was not seen in the CPU-complex.7 These inhibitors
bind in the sEH active site and directly mimic the inter-
actions consistent with the activation of a substrate
epoxide ring for nucleophilic attack by Asp333. The
urea carbonyl group accepts hydrogen bonds from
Tyr465 and Tyr381. We demonstrated that these two
tyrosine residues serve as general acid catalysts that
activate the substrate epoxide ring for nucleophilic
attack by Asp333. We explained that the activity of
benzoylphenylureas was lowered by lowering the overall
hydrophobicity compared to that of DCU. However,
there must be an additional e€ect responsible for low-
ering the activity of benzoylphenylureas. We believe
that the benzoyl carbonyl group is detrimental to the
The application of this model to the rational design of
new inhibitors gives us several new insights. Since the
substrate binding is governed mostly by van der Waals
contacts made with the residues de®ning the active site,
one strategy would be to maximize the number of
available associations. This observation is supported by
the large sterically and hydrophobically favored ®elds
surrounding the inhibitors (Fig. 4(a) and (c)). We have
synthesized diadamantylurea, which also exhibited high
activity (unpublished data). Therefore new classes of
inhibitors should explore the limits of steric bulk for
favorable activity. The other major site of activity is the
negative potential ®eld around the NH groups of the
urea and carbonyl oxygen. The necessity of at least one
NH group has been well established. However, moieties
which increased the negative potential around these
groups could lead to increased activity. Along the same
theme, maximizing the electronic charge of the carbonyl
carbon is important for the activity.
Other e€ects of benzoylphenylureas on mammalian
systems such as stimulation of hematopoesis,25 anti-
tumor activity in vivo against P388 and L1210 leuke-
mias,21,22 and inhibition of tubulin polymerization and
HL-60 cell growth23,24 may be related to the inhibition
of sEH, and the mode of action for the inhibition of
chitin synthesis.
Conclusion
Among 271 insect growth regulators tested, 18 benzoyl-
phenylureas were active against MsEH. However, these
compounds were very weak when compared with sev-
eral other ureas, such as DCU and urea-containing
Figure 5. Summarized CoMFA±HINT map for the inhibition of sEH.

Y. Nakagawa et al. / Bioorg. Med. Chem. 8 (2000) 2663±2673
2671
herbicides. In the CoMFA results, large sterically
unfavorable ®elds were concentrated above the A-ring
and the urea bridge moiety, and sterically favorable
®elds were mainly found beneath the A- and B-rings. A
series of negative potential ®elds surrounds the whole
molecule and one large positive potential ®eld is located
beneath the carbonyl carbon atom. In this CoMFA
model, only one compound which contained a thione
(CˆS) instead of a carbonyl (CˆO) group was exclu-
ded. The ligand binding modes visualized were con-
sistent with the results obtained from crystallographic
analysis. The 3-D QSAR analysis for an expanded set of
compounds including thioureas, carbamates, and alkyl
amines (unpublished data) will enable us to more pre-
cisely map the ligand±receptor environment. This 3-D
QSAR model will be useful for further probing the
active site and mechanism of sEH. In addition, it will be
a valuable tool for designing the rational synthesis of
new potent sEH inhibitors.
(3 g, 24.7 mmol) and oxalyl chloride (5.77 g, 45 mmol)
for 3 h in dichloroethane (10 mL), and puri®ed by dis-
tillation (65 ꢀC at 1 mm Hg). Compounds (3, 5, 7, 9) were
also obtained from the corresponding anilines and 4-
chlorobenzoyl isocyanate in a similar manner as above.
4-Chlorobenzoyl isocyanate was prepared by re¯uxing 4-
chlorobenzamide (2.0 g, 12.9 mmol) and oxalyl chloride
(2.02 g, 15.9 mmol) for 1.5 h in dichloroethane (10 mL)
and puri®ed by distillation (100 ꢀC at 5 mmHg). Since
benzoyl isocyanate and 4-chlorobenzoyl isocyanate are
unstable, they were used immediately for the following
reactions without determining the yield. All products
were recrystallized from ethanol. Yields for compounds
1±9 ranged from 30 to 70%, and these and all other
yields were not optimized.
Compounds 10 and 11 were prepared from phenyl iso-
cyanate (0.75 g, 6.3 mmol) and p-chlorophenyl isocya-
nate (0.58 g, 3.8 mmol) by re¯uxing with benzylamine in
THF (10 mL) followed by recrystallization of the insol-
uble products from 95% ethanol. Yields were 23% (10)
and 51% (11).
Experimental
Bioassay
Dibenzylurea (12) was prepared from benzylamine
(0.86 g, 7.98mmol) and 1,1-carbonyldiimidazole (1.09g,
6.72mmol). After stirring benzylamine and 1,1-carbonyl-
diimidazole in 95% aqueous ethanol at room temperature
overnight, the reaction a€orded colorless needles (0.18g,
18.8%) without recrystallization. mp 168.5±169.5 ꢀC.
Inhibition of sEH was reported as the IC50 (concentra-
tion of inhibitor which results in 50% inhibition of the
enzyme) according to previous methods.9 Brie¯y,
recombinant mouse sEH, expressed in our baculovirus
system,28,29 was puri®ed from cell lysate by anity
chromatography41 and the purity (> 97% pure) of each
protein was judged by SDS±PAGE. The IC50 determi-
1-(4-Chlorophenyl)-3-(2-¯uorobenzoyl)thiourea (13): 2-
¯uorobenzoyl chloride (1.85 g, 11.7 mmol) was added
dropwise to the ammonium thiocyanate (0.98 g, 12.9
mmol) in anhydrous acetone (7 mL) with stirring, then
re¯uxed for 5 min. 4-Chloroaniline (1.85 g, 14.5 mmol)
in 3.5 mL of acetone was added to the reaction mixture
followed by an additional 5 min of re¯uxing. The mix-
ture was then poured into water, resultant crystals were
collected by ®ltration and recrystallized from ethanol to
yield 2.02 g (56%). Mp 130±131 ꢀC.
nations for each inhibitor employed
a spectro-
photometric assay using racemic 4-nitrophenyl-trans-
2,3-epoxy-3-phenylpropyl carbonate as the substrate in
a polystyrene 96-well microtiter plate.42 In each well,
20 mL of enzyme preparation, 2 mL of inhibitor solution
in DMF (0.05 to 500 mM), and 180 mL of sodium phos-
phate bu€er (0.1M, pH 7.4) containing 0.1 mg/mL of
BSA were combined. This mixture was incubated at
30 ꢀC for 5 min, allowing for maximal inhibition with
little enzyme reactivation, followed by addition of 4 mL
of a 2 mM solution of substrate. Activity was assessed
by measuring the appearance of the 4-nitrophenolate
anion at 405 nm at 30 ꢀC after 1 min (Spectramax 200;
Molecular Device, Inc., Sunnyvale, CA). All assays
were performed in quadruplicate, and the IC50 was
determined by regressions of at least ®ve datum points
for each analysis with a minimum of two points in the
linear region of the curve on either side of the IC50. Each
curve was generated from at least three separate runs. In
at least one run, compounds of a similar potency were
included to ensure rank order. The reciprocal logarithm
value of IC50, pIC50, was used as the activity index.
Compounds 14 and 15 were synthesized from
the appropriate benzoyl isothiocyanates and corre-
sponding 4-substituted benzoyl hydrazines. Benzoyl
isothiocyanate was prepared from the benzoyl chloride
and either ammonium thiocyanate (NH4SCN) or potas-
sium thiocyanate (KSCN) according to the conventional
method.43 2,6-Dimethoxybenzoyl chloride (2.52 g,
12.6 mmol) dissolved in acetone was added dropwise to
KSCN (4.5 g, 22.3 mmol) dissolved in dry benzene
(10 mL). The mixture was stirred at room temperature
until the color changed to reddish yellow, then re¯uxed
for 5 min. After cooling, KCl was removed by ®ltration.
The ®ltrate was puri®ed by distillation (139 ꢀC at
17 mmHg) to a€ord 2,6-dimethoxybenzoyl isothio-
cyanate (1.27 g) in a yield of 45%. A mixture of 4-iodo-
benzoic acid (12.4 g, 50.0 mmol), ethanol (2.9 mL), and
sulfuric acid (0.9 mL) in dichloromethane (15 mL) was
re¯uxed for 10 h. After washing the dichloromethane
layer with water, 5% NaHCO3, and water, successively, it
was dried over anhydrous sodium sulfate. Ethyl 4-iodo-
benzoate was obtained by distillation (133 ꢀC, 8 mmHg).
Yield was 10.22 g (78%). Ethyl 4-iodobenzoate (10.22 g),
Chemicals
Compounds (1, 2, 4, 6, 8) were prepared from the corre-
sponding aniline (1.4±4.7 mmol) and an excess amount of
benzoyl isocyanate by mixing them in tetrahydrofuran
(THF) at room temperature followed by recrystalliza-
tion of the insoluble products from 95% ethanol. Ben-
zoyl isocyanate was prepared by re¯uxing benzamide

2672
Y. Nakagawa et al. / Bioorg. Med. Chem. 8 (2000) 2663±2673
60% hydrazinehydrate (1.85 g), and ethanol (10 mL)
were mixed, then re¯uxed for 8 h. After evaporating the
ethanol, the residue was washed with ice-cold water
followed by recrystallization from ethanol to obtain 4-
iodobenzoylhydrazine (4.37 g, 45.1%). 2,6-Dimethoxy-
benzoyl isothiocyanate (2.11 g) and 4-ethyl benzoate
(2.48 g) were dissolved in benzene (20 mL) and re¯uxed for
2 h. After evaporating the solvent, the residue was washed
with a minimal amount of diethyl ether and water.
Recrystallization was performed from ethanol to a€ord
2,6-dimethoxybenzoyl-4-iodobenzoylthiosemicarbazide
(15; 0.71 g; 69%). Mp 185±186 ꢀC. The corresponding 4-
nitro analogue (14) was synthesized in a similar manner
as above.
Compound 17 was obtained as a colorless crystalline
solid in 81% yield from the slow addition of cyclohexyl
isocyanate (0.63 g, 5.0 mmol) to a vigorously stirred solu-
tion of p-toluidine (0.54 g, 5.0 mmol) in 10mL of hexane,
and compound 16 was prepared from the corresponding
aniline as well. Each structure was characterized by 300
1
MHz or 500 MHz H NMR spectra recorded on the
Bruker AC-300 or ARX-500 and elemental analyses. The
analytical values for C, H and N agreed with calculated
Table 5. Physical data for all newly synthesized compounds
No.a
Melting point (ꢀC)
201±202
Elemental analyses (%)
1H NMR (CDCl3)b (ppm)
1
Found: C-72.95, H-6.80, N-9.45
Calcd: C-73.08, H-6.85, N-9.51
1.34 (s, 9H, t-Bu), 7.36±8.04(m, ArH), 9.41
(s, 1H, NH), 10.84 (s, 1H, NH)
2
3
4
206±207
219±220
160±161
Found: C-72.28, H-4.85, N-8.43
Calcd: C-72.16, H-4.74, N-8.45
7.00±8.02 (m, 14H, ArH), 9.34 (br, 1H, NH),
10.80 (s, 1H, NH)
Found: C-65.49, H-4.12, N-7.64
Calcd: C-65.49, H-4.00, N-7.61
6.98±7.94 (m, 13H, ArH), 9.41 (s, 1H, NH),
10.76 (s, 1H, NH)
Found: C-68.27, H-7.37, N-11.37
Calcd: C-68.28, H-7.30, N-11.27
1.24±2.00 (m, 10H, cHexH), 3.74±3.85
(m, 1H, -N-cHexHc), 7.47±7.92 (m, 5H, ArH),
8.63 (d, 1H, J=7.2 Hz, NH), 8.84 (br, 1H, NH)
5
6
7
222.0±222.5
176.5±177.5
207±208
Found: C-59.89, H-6.10, N-9.98
Calcd.: C-60.15, H-6.06, N-9.99
1.26±1.99 (m, 10H, cHexH), 3.70±3.80
(m, 1H, -N-cHexHc), 7.43±7.93 (m, 4H, ArH),
8.64 (d, 1H, J=7.5 Hz, NH), 9.39 (s, 1H, NH)
Found: C-72.69, H-9.15, N-8.48
Calcd: C-72.54, H-9.19, N-8.50
1.26±1.74 (m, 22H, cDodH), 4.00±4.04
(m, 1H, -N-cDodHd) 7.50±7.88 (m, 5H, ArH),
8.52±8.57 (m, 2H, 2 Â NH)
1.25±1.71 (m, 22H, cDodH), 3.96±4.07
(m, 1H, -N-cDodHd) 7.44±7.95 (m, 4H, ArH),
8.65 (d, 1H, J=8.0 Hz, NH), 9.63 (s, 1H, NH)
Found: C-65.83, H-8.01, N-7.68
Calcd: C-65.85, H-8.13, N-7.80
8
166±167
199±200
168±169
206
Found: C-70.85, H-5.55, N-11.02
Calcd: C-70.92, H-5.60, N-11.02
4.57 (d, 2H, J=5.9, CH2), 7.27±7.93 (m, 10H),
9.10 (br, 1H, NH), 9.20 (br, 1H, NH)
9
Found: C-62.40, H-4.54, N-9.70
Calcd: C-62.37, H-4.45, N-9.71
4.57 (d, 2H, J=5.9 Hz, CH2), 7.27±7.92
(m, 9H, ArH), 9.12 (br, 1H, NH), 9.67 (s, 1H, NH)
10
11
12
13
14
Found: C-74.31, H-6.24, N-12.38
Calcd: C-74.23, H-6.19, N-12.64
4.42 (d, 2H, J=5.8), 5.18 (br, 1H, NH), 6.50
(s, 1H, NH), 7.05±7.34 (m, 10H, ArH)
Found: C-64.50, H-5.03, N-10.74
Calcd: C-64.35, H-4.85, N-11.02
4.42 (d, 2H, J=5.8), 6.13 (br, 1H, NH), 7.17±7.39
(m, 9H, ArH), 8.01 (s, 1H, NH)
168.5±169.5
123±125
198±200
Found: C-74.97, H-6.71, N-11.66
Calcd: C-75.09, H-6.69, N-11.73
4.38 (d, 4H, J=5.8 Hz, CH2), 4.64
(br, 2H, 2ÂNH), 7.26±7.32 (m, 10H, ArH)
7.23±8.12 (m, 8H, Ar), 9.67 (d, 1H, J=14.6,
CS-NH-Ar), 12.55 (s, 1H, CO-NH-CS-)
Found: C-54.55, H-3.38, N-9.09
Calcd: C-3.26, H-54.46, N-9.07
Found: C-50.60, H-3.98, N-13.79
Calcd: C-50.49, H-3.99, N-13.85
3.89 (s, 6H, 2ÂOCH3), 6.61±8.38 (m, 7H, Ar),
8.92 (s, 1H, NH), 10.09 (d, J=6.6 Hz, NH), 13.38
(d, 1H, J=8.0 Hz, NH)
15
16
17
185±186
183.5±184.5
203±204
Found: C-42.12, H-3.24, N-8.62
Calcd: C-42.07, H-3.32, N-8.66
3.87 (s, 6H, 2ÂOCH3), 6.59±7.87 (m, 7H, ArH),
8.84 (s, 1H, NH), 9.96 (d, J=7.2 Hz, NH), 13.33
(d, 1H, J=6.8 Hz, NH)
Found: C-71.33, H-8.46, N-12.79
Calcd: C-71.53, H-8.31, N-12.83
1.09±1.99 (m, 10H, cHexH), 3.60±3.73 (m, 1H,
N-cHexHc), 4.57 (br, 1H, NH), 6.16 (s, 1H, NH),
7.07±7.34 (m, 5H, ArH)
Found: C-72.60, H-8.56, N-12.00
Calcd: C-72.38, H-8.68, N-12.06
1.03±1.98 (m, 10H, cHexH), 2.32 (s, 3H, CH3),
3.60±3.72 (m, 1H, N-cHexHc),4.55 (d, 1H,
J=7.6 Hz, NH), 6.09 (s, 1H, NH), 7.13 (s, 4H, ArH),
aNumbers correspond to those in Table 1.
1
bIf compounds were insoluble in CDCl3, a drop of DMSO-d6 was added, and H NMR was recorded at 300 MHz. For compounds 14 and 15 it is
recorded at 500 MHz.
c-N-cHexH means the CH-proton of cyclohexyl group attached to N atom.
d-N-cDodH means the CH-proton of cyclododecyl group attached to N atom.

Y. Nakagawa et al. / Bioorg. Med. Chem. 8 (2000) 2663±2673
2673
values within an error of Æ0.3%. All compounds syn-
thesized are listed in Table 1. Melting points, NMR
spectral data, and elemental analyses are summarized in
Table 5.
16. Nakagawa, Y.; Ishii, S.; Matsumura, F. Insect Biochem.
Molec. Biol. 1996, 26, 891.
17. Horst, M. J. Biol. Chem. 1981, 256, 1412.
18. Leighton, T.; Marks, E. P.; Leighton, F. Science 1981,
213, 905.
19. Oberlander, H., Silhacek, D. L. In Insecticides with Novel
Mode of Action; Ishaaya, I., Degheele, D., Eds.; Springer:
Berlin, 1998, pp 92±105.
20. Nakagawa, Y. J. Pestic. Sci. 1996, 21, 460 (in Japanese).
21. Nakajima, T.; Masuda, H.; Okamoto, T.; Watanabe, M.;
Yokoyama, K.; Yamada, N.; Fujimoto, S.; Tsukagoshi, S.;
Taguchi, T. Cancer Chemother. Pharmacol. 1991, 28, 351.
22. Okada, H.; Koyanagi, T.; Yamada, N.; Haga, T. Chem.
Pharm. Bull. 1991, 39, 2308.
A hundred and eighty-one benzoylphenylureas (III in
Figure 1) having various substituents at both benzene
rings36,44 47 and sixteen benzoylaminothiadiazole ana-
logues34 with various substituents at both benzene rings
were taken from our stock samples prepared in our pre-
vious studies. The benzoylphenylureas included 114 2,6-
di¯uorobenzoylurea analogues with single- or multiply-
substituted phenyl moieties (Yn), 34 2,6-dichloro-
benzoylurea analogues (III; Xn=2,6-Cl2) with single- or
di-substituted phenyl moieties (Yn), and 33 4-chlor-
ophenylurea appendages with various N-substituted ben-
zoyl moieties. Compounds 34±50 were previously
reported,9 and the data of compounds 51 and 52 as well.13
23. Paull, K. D.; Lin, C. M.; Malspeis, L.; Hamel, E. Cancer
Research 1992, 52, 3892.
24. Ando, N.; Nakajima, T.; Masuda, H.; Kawabata, Y.;
Iwai, M.; Watanabe, M.; Kagitani, Y.; Yamada, N.; Tsuka-
goshi, S. Cancer Chemother. Pharmacol. 1995, 37, 63.
25. Jenkins, V. K. 1992, US Patent 5166180.
26. Cramer, R. D. III; Patterson, D. E.; Bunce, J. D. J. Am.
Chem. Soc. 1988, 110, 5959.
Acknowledgements
27. Kellogg, G. E.; Semus, S. F.; Abraham, D. J. J. Comp.-
Aided Mol. Design 1991, 5, 545.
28. Grant, D. F.; Storms, D. H.; Hammock, B. D. J. Biol.
Chem. 1993, 268, 17628.
29. Beetham, J. K.; Tian, T.; Hammock, B. D. Arch. Biochem.
Biophys. 1993, 305, 197.
30. Cruse, W. B. T. Acta Cryst. 1978, B34, 2904.
31. Baroni, M.; Costantino, G.; Cruciani, G.; Riganelli, D.;
Valigi, R.; Clementi, S. Quant. Struct.-Act. Rel. 1993, 12, 9.
32. Leo, A. Chem. Rev. 1993, 93, 1281.
We thank the following sources for funding: NIEHS
ES02710, NIEHS Superfund Basic Research Program
ES04699, UC Systemwide Ecotoxicology Training Pro-
gram, U.S. EPA Center for Ecological Health Research
CR819658, and the NIEHS Center for Environmental
Health Sciences ES05707.
33. Hansch, C.; Leo, A. Exploring QSAR: Fundamentals and
Applications in Chemistry and Biology. American Chemical
Society: Washington, DC, 1995.
References
34. Nakagawa, Y.; Nishimura, K.; Izumi, K.; Kinoshita, K.;
Kimura, K.; Kurihara, N.; Fujita, T. J. Pestic. Sci. 1996, 21,
195.
35. Morisseau, C.; Du, G.; Newman, J. W.; Hammock, B. D.
Arch. Biochem. Biophys. 1998, 356, 214.
36. Nakagawa, Y.; Izumi, K.; Oikawa, N.; Kurozumi, A.;
Iwamura, H.; Fujita, T. Pestic. Biochem. Physiol. 1991, 40, 12.
37. Sharp, K. A.; Nicholls, A.; Friedman, R.; Honig, B. Bio-
chemistry 1991, 30, 9686.
38. Mullin, C. A.; Hammock, B. D. Arch. Biochem. Biophys.
1982, 216, 423.
39. Klebe, G.; Abraham, U.; Mietzner, T. J. Med. Chem.
1994, 37, 4130.
40. Hansch, C.; Fujita, T. J. Am. Chem. Soc. 1964, 86, 1616.
41. Wixtrom, R. N.; Silva, M. H.; Hammock, B. D. Anal.
Biochem. 1988, 169, 71.
42. Dietze, E. C.; Kuwano, E.; Hammock, B. D. Anal. Bio-
chem. 1994, 216, 176.
43. Douglass, I. B.; Dains, F. B. J. Am. Chem. Soc. 1934, 56,
719.
1. Oesch, F. Xenobiotica 1973, 3, 305.
2. Ota, K.; Hammock, B. D. Science 1980, 207, 1479.
3. Wixtrom, R. N.; Hammock, B. D. In Biochemical Pharma-
cology and Toxicology; Zakim, D., Vessey, D. A., Eds.; Wiley:
New York, 1985; Vol. 1, pp 1±93.
4. von Dippe, P.; Amoui, M.; Alves, C.; Levy, D. Am. J.
Physiol. 1993, 264, G528.
5. von Dippe, P.; Amoui, M.; Stellwagen, R. H.; Levy, D. J.
Biol. Chem. 1996, 271, 18176.
6. Moghaddam, M. F.; Grant, D. F.; Cheek, J. M.; Greene, J.
F.; Williamson, K. C.; Hammock, B. D. Nat. Med. 1997, 3, 562.
7. Argiriadi, M. A.; Morisseau, C.; Hammock, B. D.; Chris-
tianson, D. W. Proc. Natl. Acad. Sci. USA 1999, 96, 10637.
8. Argiriadi, M. A.; Morisseau, C.; Goodrow, M. H.; Dowdy,
D. L.; Hammock, B. D.; Christianson, D. W. J. Biol. Chem.
2000, 275, 15265.
9. Morisseau, C.; Goodrow, M. H.; Dowdy, D.; Zheng, J.;
Greene, J. F.; Sanborn, J. R.; Hammock, B. D. Proc. Natl.
Acad. Sci. USA 1999, 96, 8849.
10. Newman, J. W., Denton, D. L., Morisseau, C., Koger, C.
S., Wheelock, C. E., Hinton, D. E., Hammock, B. D. sub-
mitted to Environmental Health Perspective.
11. Marks, E. P., Leighton, T., Leighton, F. In Insecticide
Mode of Action; Coats, J. R., Ed.; Academic Press: New York,
1982, pp 281±313.
44. Nakagawa, Y.; Kitahara, K.; Nishioka, T.; Iwamura, H.;
Fujita, T. Pestic. Biochem. Physiol. 1984, 21, 309.
45. Nakagawa, Y.; Iwamura, I.; Fujita, T. Pestic. Biochem.
Physiol. 1985, 23, 7.
46. Nakagawa, Y.; Sotomatsu, T.; Irie, K.; Kitahara, K.; Iwa-
mura, H.; Fujita, T. Pestic. Biochem. Physiol. 1987, 27, 143.
47. Nakagawa, Y.; Akagi, T.; Iwamura, H.; Fujita, T. Pestic.
Biochem. Physiol. 1988, 30, 67.
12. Cohen, E.; Casida, J. E. Pestic. Biochem. Physiol. 1980,
13, 129.
13. Mitsui, T.; Nobusawa, C.; Fukami, J. J. Pestic. Sci. 1981,
6, 155.
48. Sotomatsu, T.; Nakagawa, Y.; Fujita, T. Pestic. Biochem.
Physiol. 1987, 27, 156.
14. Nakagawa, Y.; Matsumura, F. Insect Biochem. Molec.
Biol. 1994, 24, 1009.
15. Ishii, S.; Matsumura, F. Insect Biochem. Molec. Biol.
1992, 22, 69.
49. Nakagawa, Y.; Izumi, K.; Oikawa, N.; Sotomatsu, T.; Shi-
gemura, M.; Fujita, T. Environ. Toxicol. Chem. 1992, 11, 901.
50. Sotomatsu, T.; Shigemura, M.; Murata, Y.; Fujita, T. J.
Pharm. Sci. 1993, 82, 776.