Classical and three-dimensional QSAR for the inhibition of [ 3H]ponasterone a binding by diacylhydrazine-type ecdysone agonists to insect Sf-9 cells

Article abstract of DOI:10.1016/j.bmc.2004.11.004

The activity of 52 diacylhydrazine congeners was evaluated by measuring the inhibition of the incorporation of [³H]ponasterone A into intact Sf-9 cells. Eleven compounds were newly synthesized in this study. Results showed that the substitution of the 2-CH₃ or 3-OCH₃ moiety of methoxyfenozide with other groups or the removal of either group was unfavorable to the activity. The activity was quantitatively analyzed using both classical QSAR (Hansch-Fujita) and three-dimensional QSAR methods (comparative molecular field analysis, CoMFA). Sterically favorable fields were observed at the 3- and 4-positions of the benzene ring opposite from the t-butyl group (B-ring), and a sterically unfavorable field was evidenced at the 2-position. Another sterically unfavorable field developed surrounding the favorable field observed at the 4-position of the B-ring. Electrostatically negative fields were observed near the CO moiety, above the benzene ring, and at the 4-position of the B-ring. The optimum hydrophobicity of compounds in terms of their log P values was calculated to be ≈4.1. Results of the three dimensional structure-activity relationship analyses were consistent with those obtained from the previously reported classical QSAR for 2-chlorobenzoyl analogs containing various para-substituents. The high activity of potent insecticides such as tebufenozide and chromafenozide were rationalized by CoMFA. Thus, this CoMFA result will be useful in the design of new compounds and in understanding the molecular mechanism of the ligand-receptor interactions.

Full text of DOI:10.1016/j.bmc.2004.11.004

Bioorganic & Medicinal Chemistry 13 (2005) 1333–1340
Classical and three-dimensional QSAR for the inhibition
of [3H]ponasterone A binding by diacylhydrazine-type ecdysone
agonists to insect Sf-9 cells
Yoshiaki Nakagawa,* Kaoru Takahashi, Hidetoshi Kishikawa, Takehiko Ogura,
Chieka Minakuchi and Hisashi Miyagawa
Division of Applied Life Sciences, Graduate School of Agriculture, Kyoto University, Kyoto 606-8502, Japan
Received 15 October 2004; revised 4 November 2004; accepted 5 November 2004
Available online 23 November 2004
Abstract—The activity of 52diacylhydrazine congeners was evaluated by measuring the inhibition of the incorporation of
[3H]ponasterone A into intact Sf-9 cells. Eleven compounds were newly synthesized in this study. Results showed that the substi-
tution of the 2-CH3 or 3-OCH3 moiety of methoxyfenozide with other groups or the removal of either group was unfavorable
to the activity. The activity was quantitatively analyzed using both classical QSAR (Hansch-Fujita) and three-dimensional QSAR
methods (comparative molecular field analysis, CoMFA). Sterically favorable fields were observed at the 3- and 4-positions of the
benzene ring opposite from the t-butyl group (B-ring), and a sterically unfavorable field was evidenced at the 2-position. Another
sterically unfavorable field developed surrounding the favorable field observed at the 4-position of the B-ring. Electrostatically nega-
tive fields were observed near the CO moiety, above the benzene ring, and at the 4-position of the B-ring. The optimum hydro-
phobicity of compounds in terms of their logP values was calculated to be ꢀ4.1. Results of the three dimensional structure–
activity relationship analyses were consistent with those obtained from the previously reported classical QSAR for 2-chlorobenzoyl
analogs containing various para-substituents. The high activity of potent insecticides such as tebufenozide and chromafenozide were
rationalized by CoMFA. Thus, this CoMFA result will be useful in the design of new compounds and in understanding the mole-
cular mechanism of the ligand–receptor interactions.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
have more potent activity, compared to tebufenozide.
The former two dibenzoylhydrazines commonly have
multiple substitution patterns on the benzene ring oppo-
site from the t-butyl group (B-ring), resulting in physico-
chemical properties that are considered to be essential
to compound potency.
Diacylhydrazines are non-steroidal molting hormone
agonists that have insecticidal activity.1–7 To date four
compounds, tebufenozide,8 methoxyfenozide,9 chroma-
fenozide,10 and halofenozide11 are currently used as
insecticides. Although a steroid compound, 20-hydroxy-
ecdysone (20E), is commonly utilized in insects as the
molting hormone, selective toxicity is observed among
different insect orders for these diacylhydrazine-type
compounds. Tebufenozide, methoxyfenozide, and chro-
mafenozide are very potent against Lepidoptera, but
weak or inactive against other insect orders such as Dip-
tera and Coleoptera; while halofenozide is effective on
Coleoptera.12 Among the Lepidoptera-selective diacyl-
hydrazines, methoxyfenozide,9 and chromafenozide13
The hormonal activity of diacylhydrazines is due to their
interaction with the ecdysone receptor.14–16 The binding
affinity of methoxyfenozide to receptor proteins was
reported to be higher than that of tebufenozide against
Plodia interpunctella (Pyralidae), which was consistent
with the higher insecticidal activity of methoxyfeno-
zide.9 However, for the Chilo suppressalis (Pyralidae)
ecdysone receptor, the binding affinity of methoxyfeno-
zide and chromafenozide to the receptor proteins was at
the same level as that of tebufenozide, and no significant
difference was observed.16 The binding affinity of meth-
oxyfenozide to the ecdysone receptor proteins of the
Sf-9 cell line, which was established from Spodoptera
frugiperda (Noctuidae), was three times less potent than
Keywords: Ecdysone agonists; Structure–activity relationships; Diacyl-
hydrazines; QSAR.
*
Corresponding author. Tel.: +81 75 753 6117; fax: +81 75 753
6123; e-mail: naka@kais.kyoto-u.ac.jp
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.11.004

1334
Y. Nakagawa et al. / Bioorg. Med. Chem. 13 (2005) 1333–1340
t-Bu
N
ticular attention was directed to the tentatively unfavo-
t-Bu
N
H
N
H
N
rable effects of 2,3-disubstitution in the B-ring of ligands
on the Sf-9 ecdysone receptor. For this purpose, several
methoxyfenozide congeners were newly synthesized to
increase the variety of compounds with this type of sub-
stitution. The activity was evaluated by measuring the
inhibition of [3H]ponasterone A ([3H]PoA) binding to
intact cells,14,18,19 and the activity was quantitatively
analyzed using the Hansch-Fujita classical QSAR20
and a three dimensional QSAR method, a comparative
molecular field analysis (CoMFA).21
A
C
O
C
O
A
B
C
O
C
O
RB
Yn Xn
Xn
(I)
(II)
Figure 1. Structures of diacylhydrazines.
tebufenozide.14 Thus, it is evident that the effects of the
physicochemical properties of dibenzoylhydrazines on
the receptor binding depend on the origin of the recep-
tor, even among Lepidoptera.
Since the interaction of a compound with the ecdysone
receptor is the most important key for the expression
of hormonal activity, structural factors that affect the
ligand–receptor interaction are of great concern.
Previously, we measured the activity of a set of di-
benzoylhydrazines with various substituents at the
para-position of the B-ring using intact Sf-9 cells, and
quantitatively analyzed the substituent effects on the
receptor binding in Sf-9 cells.17 In this study, an ex-
panded set of compounds including methoxyfenozide,
chromafenozide, and alkanoyl analogs (II, Fig. 1) were
prepared to extensively examine the structure–activity
relationship for their receptor binding in Sf-9 cells. Par-
2. Experimental
2.1. Compounds
Compounds 2, 7–11, 15–17, 41, and 52 listed in Table 1
were newly synthesized according to the conventional
method as shown below.2,3,22 The chemical structures
were confirmed by NMR and elemental analyses. The
analytical values for C, H, and N in the elemental ana-
lysis agreed with the calculated values within 0.3% as
shown in Table 2. The melting points of newly synthe-
sized compounds are listed in Table 2. Chromafenozide
Table 1. Inhibition of [3H]ponasterone A binding by diacylhydrazine analogs in Sf-9 cells
No
Compounds
Binding activity [pIC50 (M)]
Eq. 2
t-Bu
A
B
C
O
N
N
H
C
O
Yn
Xn
Xn
Yn
He
Obsda
Calcdb
Dc
LogPd
1
H
6.44f
6.10
6.73
7.58
6.62
6.49
6.80
7.00
0.34
ꢁ0.25
0.06
2.45g,h
3.37
3.51k
2.04g,h
3.28g,h
3.39g,h
3.98
2
H
4-Cl (RH-0345)i
4-Et
6.48 0.18 (2)
7.64j
3
H
4
2-OCH3
3-Cl
H
6.57 0.09 (2)
6.56 0.14 (2)
7.27 0.23 (2)
6.86 0.11 (2)
7.36 0.02(2)
6.80 0.05 (2)
7.09 0.04 (2)
6.60 0.04 (2)
8.81f
ꢁ0.05
0.07
0.47
5
H
6
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
2-Cl
H
7
2-CH3
3-CH3
3-OH
3-OCH3
3-OEt
4-Etl
ꢁ0.14
ꢁ0.11
ꢁ0.15
0.26
8
7.47
3.98
3.34
9
6.95
6.83
6.49
8.63
7.17
8.44
7.17
7.30
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
3.75
4.28
0.11
0.18
4.39k
4-n-Bu
7.08j
8.46f
ꢁ0.09
5.39k
m
k
2-CH3–3-OCH3
2-CH3–3-OEt
2-CH3–3-OH
2,3-(CH3)2
0.023.93
7.25 0.19 (2)
7.01 0.04 (2)
7.92 0.12(2)
8.78 0.11 (3)
7.08k
0.08
ꢁ0.29
0.13
4.78
3.84
7.79
4.43
2-CH3–3,4-(CH2CH2CH2O)n
8.66
6.87
6.25
7.15
7.12
7.10
7.28
6.99
7.31
7.58
7.05
6.320..7180
5.99
7.45
7.87
0.124.80
H
0.21
ꢁ0.21
ꢁ0.05
ꢁ0.24
ꢁ0.15
0.01
2.59g,h
1.99o
2.91g,o
3.11g,o
2.81o
2-Cl
2-Cl
2-NO2
2-CH3
3-CH3
3-OCH3
4-F
6.04j
7.10 0.08 (2)
6.88 0.03 (2)
6.95 0.00 (2)
7.29p
2-Cl
2-Cl
2-Cl
2-Cl
2.87g,o
3.51g,o
3.73o
4-Cl
4-Br
6.89p
ꢁ0.10
0.38
2-Cl
2-Cl
7.69p
4-I
7.30p
ꢁ0.28
ꢁ0.09
2
3.96o
2-Cl
2-Cl
4-CF3
4-NO2
4-CN
4-CH3
4-Et
6.96p
3.68o
6.42p
o
2-Cl
2-Cl
5.74p
ꢁ0.25
ꢁ0.17
2.44o
7.28p
3.15g,o
o
2-Cl
7.89p
0.023.59

Y. Nakagawa et al. / Bioorg. Med. Chem. 13 (2005) 1333–1340
1335
Table 1 (continued)
No
Compounds
Binding activity [pIC50 (M)]
Eq. 2
t-Bu
A
B
C
O
N
N
H
C
O
Yn
Xn
Xn
Yn
Obsda
Calcdb
Dc
LogPd
33
34
35
36
37
38
39
40
41
42
43
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
4-nPr
4-iPr
7.64p
7.86
7.79
8.05
7.95
7.28
6.720..8429
5.67
6.21
8.52
8.30
7.53
ꢁ0.22
0.03
0.22
4.06o
4.11o
4.60p
4.48o
7.82p
4-n-Bu
4-tBu
4-Ph
8.27p
7.66j
ꢁ0.29
ꢁ2.11
2
5.17p
4.49o
o
4-OCH3
7.21p
4-SO2CH3
4-COCH3
2-CH3, 3-OCH3
2,3-CH3
5.54p
ꢁ0.13
0.16
1.46q
2.42q
3.46
3.40o
2.35o
6.37p
8.48 0.01 (2)
8.24 0.08 (2)
7.52j
ꢁ0.04
ꢁ0.06
ꢁ0.01
2,6-F2
t-Bu
N
A
C
O
N
H
C
O
R
B
Xn
Xn
RB
44
45
46
47
48
49
50
51
52
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
3,5-(CH3)2
2-Cl
n-Pr
n-Bu
5.92 0.06 (2)
6.79 0.07 (2)
7.44 0.17 (2)
7.49 0.08 (2)
7.62 0.03 (2)
7.88 0.10 (3)
6.70 0.01 (2)
6.07 0.08 (2)
5.75 0.16 (3)
6.19
ꢁ0.27
0.12
0.03
2.60k
3.13k
3.66k
6.67
7.41
7.19
n-Pentyl
i-Pentyl
n-Hexyl
i-Hexyl
n-Heptyl
n-Nonyl
n-Hexyl
0.30
3.53k
k
7.48
0.14
4.19
8.03
6.94
6.11
7.94
ꢁ0.15
ꢁ0.24
ꢁ0.04
ꢁ2.19
4.06k
4.72k
5.78k
3.50
a Unless noted measured in this study. With the mean standard deviation for the number of replications indicated in parentheses.
b Calculated by Eq. 2.
c Difference between observed and calculated values.
d Unless noted evaluated in this study.
e RH-5849.
f Ref. 27.
g Experimentally measured.
h Ref. 3.
i Halofenozide.
j Ref. 19.
k Ref. 14.
l Tebufenozide (RH-5992).
m Methoxyfenozide (RH-2485).
n Chromafenozide (ANS-118).
o Ref. 2.
p Ref. 17.
q Ref. 5.
(18) was a generous gift from Sankyo Agro Co., Ltd and
Nippon Kayaku Co., Ltd, and other compounds were
from our stock and are described in previous laboratory
publications. The experimental procedures for the syn-
thesis of N-3,5-dimethylbenzoyl-N-t-butyl-N0-2-methyl-
3-hydroxybenzoylhydrazine are shown below.
drous dichloromethane (3.6mL). This benzoyl chloride
in dichloromethane containing triethylamine 1.2mL
(8.6mmol) was added dropwise to N-t-butyl-N-(3,5-
dimethylbenzoyl)hydrazine 0.48g (2.2mmol) dissolved
in anhydrous dichloromethane (3.6mL) under the ice-
cold condition, and stirred overnight at room tempera-
ture. The reaction mixture was washed with aqueous
1M HCl and saline, which was submitted to the follow-
ing reaction without purification. To this dichlorometh-
ane solution containing N-t-butyl-N-(3,5-dimethyl-
benzoyl)-N0-(3-acetoxy-2-methylbenzoyl)hydrazine was
added 1M NaOH (34mL) and stirred overnight at room
temperature. After adding 3M HCl (15mL) to the aque-
ous layer, the mixture was extracted with ethyl acetate.
The organic layer was washed with saline and dried over
2.2. N-t-Butyl-N-3,5-dimethylbenzoyl-N0-2-methyl-3-
hydroxybenzoylhydrazine
3-Acetoxy-2-methylbenzoic acid 0.40g (2.1mmol) in
thionyl chloride 0.4mL (5.5mmol) was refluxed at
80ꢁC for 5h. Excess thionyl chloride was removed
azeotropically with toluene to afford crude 3-acetoxy-
2-methylbenzoyl chloride which was dissolved in anhy-

1336
Y. Nakagawa et al. / Bioorg. Med. Chem. 13 (2005) 1333–1340
Table 2. Melting points and elemental analysis values of newly
synthesized compounds
2.4. QSAR analyses
Noa
Mp (ꢁC)
Elemental analysis values
The Hansch-Fujita QSAR analysis was performed using
QREG system ver.2.05.25 LogP (P: partition coefficient
in 1-octanol/water system) values of newly synthesized
compounds were calculated by CLOG P program
MacLogP Ver.4.0 (BioByte Corp., Claremont, CA,
USA).26 ClogP values of representative compounds 1,
6, and 19 were 2.48, 3.48, and 2.76, which are close to
the respective measured values. Other logP values were
cited from our previous reports.2,3,17,22,27 In all equa-
tions n is the number of compounds used for the regres-
sion analyses, s is the standard deviation, r is the
correlation coefficient, and the values in parentheses
are 95% confidence intervals of the regression coefficient
and intercept.
2
153–155
Found: C-65.27, H-5.89, N-8.33
Calcd: C-65.35, H-5.79, N-8.47
7
163–164
162–165
218–219
Found: C-74.57, H-7.75, N-8.26
Calcd: C-74.52, H-7.74, N-8.28
8
Found: C-74.49, H-7.73, N-8.35
Calcd: C-74.52, H-7.74, N-8.28
9
Found: C-70.38, H-7.11, N-8.17
Calcd: C-70.56, H-7.11, N-8.23
10
11
15
16
17
41
52
159–162Found: C-704.9, 8N, -H7.-970.2
Calcd: C-71.16, H-7.39, N-7.90
133–134
167–169
214–216
175–176
Found: C-71.86, H-7.68, N-7.87
Calcd: C-71.71, H-7.66, N-7.60
All computations for CoMFA were performed with the
molecular modeling software package SYBYL ver 6.91
(Tripos Co., St. Louis, MO, USA). Structures were
generated by modifying the structure of RH-5849 (1),
whose X-ray structure has been previously reported.28
The geometry of all structures are fully optimized with
PM3 and charges were calculated by AM1. All diacyl-
hydrazines are automatically aligned using the SYBYL
module (Align Database), in which common skeletal
chain C–C–N–N–C–C was used as a template. The
molecules were superposed in the lattice space of
Found: C-72.29, H-7.98, N-7.36
Calcd: C-72.22, H-7.91, N-7.32
Found: C-71.19, H-7.30, N-7.93
Calcd: C-71.16, H-7.39, N-7.90
Found: C-75.02, H-8.26, N-7.99
Calcd: C-74.97, H-8.01, N-7.95
171–172Found: C-63.80, H-6.13, N-7.49
Calcd: C-64.08, H-6.18, N-7.47
117–118
Found: C-63.71, H-7.84, N-8.21
Calcd: C-63.80, H-8.03, N-8.27
˚
˚
˚
23.78A · 21.36A · 25.84A (X = ꢁ8.72to 15.06,
Y = ꢁ9.85 to 11.51, Z = ꢁ7.40 to 8.44). The superposi-
tion of all 52compounds is shown in Figure 2. CoM-
FA was executed using the SYBYL QSAR module. The
a Corresponds to the compound number given in Table 1.
3
˚
anhydrous MgSO4, then the solvent was evaporated.
The residue was recrystallized from ethyl acetate to af-
ford colorless plate 0.50g (1.4mmol). Yield 66%.
NMR dH (CDCl3): 1.59 (9H, s), 1.76 (3H, s), 2.29 (6H,
s), 6.11 (1H, dd, J = 7.5, 1.0Hz), 6.75 (1H, dd, J = 8.0,
1.1Hz), 6.88 (1H, dd, J = 7.7Hz, 7.9Hz), 7.08 (3H, s).
NMR dC (CDCl3): 12.1, 21.3, 28.2, 62.4, 117.4, 118.6,
123.8, 125.5, 127.1, 132.1, 136.4, 138.8, 138.9, 157.0,
171.3, 176. Other new dibenzoylhydrazine analogs were
synthesized similarly.
lattice spacing was 2A, and a + 1 charge and a SP car-
bon was used as probes to estimate the electrostatic
and steric fields, respectively. The electrostatic and
steric potential energies at each lattice point were cal-
culated using Coulombic and Lennard–Jones potential
functions, respectively. The hydrophobic effect was
evaluated using logP and (logP)2 as the lattice inde-
pendent external descriptors. The correlation analyses
were analyzed with the partial least squares (PLS)
method.29,30 We initially selected the number of latent
variables in the set from the leave-one-out method
setting the column filtering at 2kcal/mol, and then
performed the conventional analysis using optimum
component number, m. CoMFA results were repre-
sented by the conventional equations with statistic val-
ues of the leave-one-out cross-validated correlation
coefficient, q2, the cross-validated standard error, Scv,
the conventional correlation coefficient, r, and the
standard deviation, s. The relative contribution (%)
of descriptors to the correlation equations was also
shown. The results were visualized by contour maps
using connected lattice points having an equivalent
coefficient level for each molecular field.
2.3. Bioassay
The inhibition of the binding of [3H]PoA (ARC Inc,
Carlsbad, CA, USA) to Sf-9 cells was examined accord-
ing to previously reported methods.14,18 In brief, 400lL
of cell suspension (4 · 106 cells/mL) was incubated with
1lL of DMSO solution of the test compound and
2lL of the 70% ethanol solution of [3H]PoA (0.5lM,
ca. 60,000dpm) for 30min at 25ꢁC. The reaction mix-
ture was immediately filtered through a glass filter
(GF/F) and washed three times with water (1mL). The
radioactivity collected in the filter was counted with a
liquid scintillation counter (LSC) in 3mL of Aquasol-2
(Packard Instrument Co., Meriden, CT, USA). The con-
centration-response curve for the inhibition of the
[3H]PoA binding was drawn for each compound. The
concentration required to give 50% inhibition (IC50)
was determined by probit analysis,23,24 and its reciprocal
logarithm of IC50, pIC50, was used as the index of the
activity.
3. Results
3.1. Inhibition of the binding of [3H]PonA to Sf-9 cells
The activity values of 52diacylhydrazine congeners in
terms of pIC50 (M) are listed in Table 1. Chromafenozide

Y. Nakagawa et al. / Bioorg. Med. Chem. 13 (2005) 1333–1340
1337
Figure 2. Stereoview of the superpositions of all 52compounds.
(18) was equipotent to tebufenozide (12) with the highest
activity, whereas methoxyfenozide (14) was half as po-
tent as tebufenozide. By replacing the 3-OCH3 group
on the B-ring of methoxyfenozide with H (7), OH (16),
or OC2H5 (15) the activity decreased 16–30-fold, but this
decrease was only 3.5-fold for compound 17 in which
the 3-OCH3 was substituted with 3-CH3. Thus, the ef-
fects of meta-substituents on the enhancement of the
activity were in the following order, OCH3 (14) > CH3
(17) > OEt (15) > OH (16) P H (7). This sequence was
greatly different from that observed for mono meta-sub-
stituted compounds, being decreased in the order of
CH3 (8) > H (6) > OCH3 (10) > OH (9) > OEt (11).
true even though the compound with 3,5-dimethyl-sub-
stituents on the A-ring and 4-Et on the B-ring (tebufe-
nozide, 12) was still 10 times more potent than that
with 2-Cl on the A-ring and 4-Et on B-ring (32). With
respect to the compounds with heptanoyl (RB = n-hexyl)
instead of benzoyl as the B-ring moiety, the activity of 2-
Cl-A-ring type compound 52 was about 100 times less
potent than the corresponding 3,5-dimethyl-A-ring type
compound 48. The effects of other substituents (H, 2-
CH3, 3-CH3, 3-OCH3, 2-CH3–3-OCH3, 2,3-diCH3) were
similar between 3,5-dimethyl- and 2-Cl-A-ring type
compounds.
3.2. QSAR analyses
Interestingly, the effects of structural changes in the B-
ring moiety on the activity are different between the
compound groups having 2-Cl-type A ring and 3,5-di-
methyl-type A ring, as shown in Figure 3. Typically,
the introduction of a n-Bu group on the B-ring is 10
times more effective in compounds containing a 2-Cl
on the A-ring (19 vs 35) than in compounds containing
3,5-dimethyl groups (6 vs 13). This observation holds
Previous classical QSAR analyses for the activity of di-
acylhydrazines indicated that the molecular hydrophobi-
city in terms of logP is essential for obtaining significant
correlation equations.17 Therefore, we began the analy-
sis using logP to formulate the initial significant correla-
tion (Eq. 1), in which compounds 37 and 52 are omitted.
Compound 37 was omitted in our previous QSAR anal-
ysis,17 and compound 52 was exceptionally weak as
show in Table 1 and Figure 3.
2
pIC50 ¼ 2:557ðÆ1:132Þ log P ꢁ 0:302ðÆ0:156Þðlog PÞ
þ 2:151ðÆ1:999Þ;
2-CH3-3-OCH3
8.5
n ¼ 50; s ¼ 0:588; r2 ¼ 0:410;
F 2;47 ¼ 16:318; log Popt; ¼ 4:23:
ð1Þ
4-n-Bu
2,3-(CH3)2
4-Et
7.5
As shown in Eq. 1, 40% of the total variation of activity
of 50 diacylhydrazines was explained by logP and its
squared term. The optimum logP value, logPopt, was
evaluated to be 4.23 from Eq. 1. Since steric and elec-
tronic effects participate to vary the activity in addition
to the hydrophobicity, the addition of other physico-
chemical parameters in Eq. 1 should improve the corre-
lation. However, since compounds 44–51 contain
alkanoyl instead of benzoyl moieties, it is not easy to ap-
ply the classical QSAR method to the combined set of
these compounds. Therefore, CoMFA was used to ex-
plore electronic and steric effects of diacylhydrazines to
derive Eq. 2.
2-CH3
H
3-CH3
3-OCH3
6.5
5.5
Acyl C6
6.5
7.5
8.5
pIC50 [3,5-(CH3)2 type]
Figure 3. Comparison of the activity values between 3,5-dimethyl and
2-Cl A-ring series of compounds.

1338
Y. Nakagawa et al. / Bioorg. Med. Chem. 13 (2005) 1333–1340
Figure 4. Stereoviews of the CoMFA models with chromafenozide. (A) The contours are shown to surround regions where a higher steric bulk
increases (green) or decreases (yellow) the binding. (B) The contours are shown to surround regions where a positive (blue) and a negative (red)
electrostatic potential increases the binding.
2
maps in Figure 4. Sterically favorable fields (green) ap-
peared above the 3- and 4-positions of the B-ring, and
sterically unfavorable fields (yellow) appeared covering
pIC50 ¼ 1:620 log P ꢁ 0:197ðlog PÞ
þ ½CoMFAꢂ þ 2:891
n ¼ 50; s ¼ 0:214; r2 ¼ 0:928
ðq2 ¼ 0:593; Scv ¼ 0:510; F 6;43 ¼ 92:755; m ¼ 6Þ
Steric ¼ 35:3%;
Electrostatic ¼ 18:4%; log P ¼ 24:6%;
ðlog PÞ ¼ 21:7%; log Popt ¼ 4:11:
The logPopt value (=4.11) estimated in Eq. 2 was close
ð2Þ
2-position as shown in Figure 4A. Another sterically
unfavorable field developed surrounding the favorable
field observed at the 4-position of B-ring. A large posi-
tive electrostatic-potential network (blue) appeared sur-
rounding the 3,4-positions of the B-ring, and negative
potential networks (red) appeared near the C@O moiety
and above the B-ring as shown in Figure 4-B. The rela-
tionship between observed and calculated values is
shown in Figure 5.
2
to that in Eq. 1. Views of the steric and electrostatic
CoMFA were shown as STDEV COEFF contour
*
4. Discussion
9
8
7
Previously we reported the classical-QSAR for the activ-
ity of a set of dibenzoylhydrazine congeners with vari-
ous para-substituents at the B-ring against Sf-9 cells.17
The activity of 17 para-substituted compounds (I:
Xn = 2-Cl in Fig. 1) was quantitatively analyzed with
electronic, steric, and hydrophobic substituent parame-
ters to formulate Eq. 3.
pIC50 ¼ 0:607 log P ꢁ 0:822r ꢁ 0:367DB1 þ 5:502
n ¼ 17; s ¼ 0:243; r2 ¼ 0:917; F 3;13 ¼ 46:026:
ð3Þ
52
8
6
5
37
5
6
7
9
In Eq. 3, r is the regular HammettÕs value representing
the electron withdrawal property.31 B1 is the STERI-
MOL parameter which represents the minimum width
of substituents in the projection on the plane perpendi-
cular to the axis connecting the a-atom of the substitu-
ents with the rest of the molecule, and D means that
Calculated
Figure 5. Graphical representation of observed pIC50 values versus
pIC50 values calculated with Eq. 2 for 50 compounds (closed circles).
Open circles are omitted compounds (37 and 52) in the correlation
analysis.

Y. Nakagawa et al. / Bioorg. Med. Chem. 13 (2005) 1333–1340
1339
the reference of the B1 value was shifted to that of H.32
The equation indicated that the electron-donating para-
substituents at the B-ring moiety were favorable for
activity. The electron donating effect from the B-ring is
likely to increase the electron density of the carbonyl
oxygen, and therefore, we propose that one of the driv-
ing forces of the ligand binding could the electron trans-
fer from the carbonyl oxygen to the acidic site of the
receptor.
2-Cl type compound 35 was 4.60. The former is 10 times
more hydrophobic than the optimum logP, whereas the
latter was closer to the optimum logP value. In fact, the
binding activity of compound 13 was about 10 times
lower than that of compound 35. By contrast, 3,5-di-
methyl-type compound 12 containing an Et group on
the B-ring has a logP value of 4.39, very close to the
optimum value. In the case of the corresponding 2-Cl
type compound 32, the logP value (3.59) was fairly
low, contributing to the reduction in the activity.
The result of the 3D-QSAR analysis in the present study
using an expanded set of compounds was consistent
with previously reported results. As shown in Figure 3,
the electrostatically positive (blue) fields appear sur-
rounding the 3- and 4-positions of the B-ring, showing
that the presence of electron-donating substituents on
these positions is favorable for activity. Furthermore,
the appearance of the electrostatically negative field near
the C@O group and above the B-ring strongly supports
our proposed mechanism of ligand–receptor interaction.
According to this CoMFA study, 50 compounds are
fairly well predicted as shown in Table 1 and Figure 5,
whereas omitted compounds (37 and 52) were both pre-
dicted to be 100 times more potent. The receptor bind-
ing modes of these compounds may not be same as
that of ordinary dibenzoylhydrazine-type compounds.
The 3D-QSAR result demonstrated that the presence of
a substituent at the 2-position on the B-ring is essentially
unfavorable for receptor binding due to steric effects,
although one of the commonly used insecticides, meth-
oxyfenozide, contains a CH3 group at this position. In
fact, methoxyfenozide has reduced receptor binding
activity, compared to tebufenozide, which contains no
substituent at the corresponding position. Chromafenoz-
ide, however, is equipotent to tebufenozide, in spite of
the presence of 2-CH3 group on the B-ring, which is lo-
cated in a sterically unfavorable region near the ortho-
position of the B-ring. This observation is probably
due to counterbalancing by the oxygen atom in the chro-
mane ring located in the electrostatically negative region.
Previously we reported that the binding affinity of 10
ecdysone agonists to Sf-9 cells is linearly correlated with
the hormonal activity in cultured integument of C. sup-
pressalis, in which both steroidal and non-steroidal
compounds were considered.14,15 Here we examined
the relationship between these two activities for an ex-
panded set of ecdysone agonists (n = 25) including ecdy-
steroids as shown in Figure 6, and still obtained a strong
linear correlation. Some scatter is certainly recognized in
this correlation, which is probably due to the small dif-
ference of the binding pockets of receptors between S.
frugiperda and C. suppressalis. However, the overall
similarity in QSAR provides useful findings for design-
ing new potent compounds against various Lepidoptera
insects.
In Eq. 2 the squared logP term became significant,
although it was insignificant in Eq. 3. This observation
is most likely due to the fact that a number of com-
pounds with supra-optimum molecular hydrophobicity
were included in the 50 compounds used to formulate
Eq. 2. In fact, the squared logP term was significant
in the QSAR of the hormonal activity against C. sup-
pressalis as indicated in Eq. 4,27 although the biological
activity is different from the binding activity used in Eqs.
(1)–(3). LogP values of diacylhydrazine congeners used
in Eq. 4 vary from 2.60 to 5.78.
2
pEC50 ¼ 7:779 log P ꢁ 0:775ðlog PÞ ꢁ 0:893DA
ꢁ 0:609DB ꢁ 3:227
ð4Þ
n ¼ 23; s ¼ 0:557; r2 ¼ 0:863;
In conclusion, the binding affinity of various diacyl-
hydrazines was correlated with CoMFA steric and
F 4;18 ¼ 28:291; log Popt ¼ 5:15:
9
8
7
6
5
The logPopt value (5.15) estimated in Eq. 4 was higher
than that (4.11) evaluated for the activity to Sf-9 cells.
In Eq. 4, DA and DB express the length of the A-ring
and B-ring moieties of the diacylhydrazine congeners,
respectively.
The existence of the optimum molecular hydrophobicity
in the ligand binding is likely to account for different
substituent effects at the B-ring between 3,5-dimethyl
and 2-Cl type compounds. Since the hydrophobicity val-
ues of these two types of A-ring moieties are consider-
ably different, the introduction of the same substituent
into the B-ring may have a non-linear effect on the
activity as a result of the change in the total molecular
hydrophobicity. For example, the logP value of 3,5-di-
methyl-type compound 13 containing a n-Bu group on
the B-ring was 5.39, whereas that of the corresponding
4
5
6
7
8
9
pIC50(M)
Figure 6. Relationship between the binding (pIC50) of ecdysone
agonists to Sf-9 cells and their molting hormonal activity (pEC50)
measured in the cultured integument of C. suppressalis.

1340
Y. Nakagawa et al. / Bioorg. Med. Chem. 13 (2005) 1333–1340
10. Sawada, Y.; Yanai, T.; Nakagawa, H.; Tsukamoto, Y.;
Tamagawa, Y.; Yokoi, S.; Yanagi, M.; Toya, T.; Sugizaki,
H.; Kato, Y.; Shirakura, H.; Watanabe, T.; Yajima, Y.;
Kodama, S.; Masui, A. Pest Manag. Sci. 2003, 59, 49.
11. Dhadialla, T. S.; Carlson, G. R.; Le, D. P. Annu. Rev.
Entomol. 1998, 43, 545.
electrostatic parameters as well as the molecular hydro-
phobicity parameters. This CoMFA result was consist-
ent with the previous classical QSAR result17 and
reasonably predicted the potent analogs such as tebufen-
ozide, chromafenozide, and methoxyfenozide. We also
demonstrated that QSARs will be useful to verify the
ligand–receptor binding mode.
12. Cowles, R. S.; Villani, M. G. J. Econ. Entomol. 1996, 89,
1556.
13. Sawada, Y.; Yanai, T.; Nakagawa, H.; Tsukamoto, Y.;
Yokoi, S.; Yanagi, M.; Toya, T.; Sugizaki, H.; Kato, Y.;
Shirakura, H.; Watanabe, T.; Yajima, Y.; Kodama, S.;
Masui, A. Pest Manag. Sci. 2003, 59, 36.
Acknowledgements
14. Nakagawa, Y.; Minakuchi, C.; Ueno, T. Steroids 2000, 65,
537.
15. Watanabe, B.; Nakagawa, Y.; Miyagawa, H. J. Pestic.
Sci. 2003, 28, 188.
16. Minakuchi, C.; Nakagawa, Y.; Kamimura, M.; Miya-
gawa, H. Eur. J. Biochem. 2003, 270, 4095.
17. Ogura, T.; Nakagawa, Y.; Minakuchi, C.; Miyagawa, H.
J. Pestic. Sci., in press.
18. Nakagawa, Y.; Minakuchi, C.; Takahashi, K.; Ueno, T.
Insect Biochem. Mol. Biol. 2002, 32, 175.
19. Minakuchi, C.; Nakagawa, Y.; Miyagawa, H. J. Pestic.
Sci. 2003, 28, 55.
20. Hansch, C.; Fujita, T. J. Am. Chem. Soc. 1964, 86, 1616.
21. Cramer, R. D., III; Patterson, D. E.; Bunce, J. D. J. Am.
Chem. Soc. 1988, 110, 5959.
We are thankful to Dr. Craig Wheelock for carefully
reviewing this manuscript. We express our sincere
thanks to Sankyo Agro, Co. Ltd and Nippon Kayaku
Co., Ltd for the gift of chromafenozide and Mr. Masa-
hiro Kawakatsu for his skillful cooperation in synthetic
experiments. We also thank to Mr. Ryoichi Takami for
his advice in the setting of the computer system. A part
of this study was performed in the RI center of Kyoto
University. This investigation was supported, in part
by a Grant-in-Aid for Scientific Research from Ministry
of Education, Culture, Sports, Science, and Technology
(formally Ministry of Education, Science, and Culture)
of Japan (09660117, 10161207).
22. Nakagawa, Y.; Hattori, K.; Shimizu, B.; Akamatsu, M.;
Miyagawa, H.; Ueno, T. Pestic. Sci. 1998, 53, 267.
23. Finney, D. J. Probit Analysis; Cambridge University Press:
Cambridge, 1952.
References and notes
1. Wing, K. D.; Slawecki, R. A.; Carlson, G. R. Science
1988, 241, 470.
2. Oikawa, N.; Nakagawa, Y.; Nishimura, K.; Ueno, T.;
Fujita, T. Pestic. Biochem. Physiol. 1994, 48, 135.
3. Oikawa, N.; Nakagawa, Y.; Nishimura, K.; Ueno, T.;
Fujita, T. Pestic. Sci. 1994, 41, 139.
4. Smagge, G.; Nakagawa, Y.; Carton, B.; Mourad, A. K.;
Fujita, T.; Tirry, L. Arch. Insect Biochem. Physiol. 1999,
41, 42 .
5. Nakagawa, Y.; Smagghe, G.; Kugimiya, S.; Hattori, K.;
Ueno, T.; Tirry, L.; Fujita, T. Pestic. Sci. 1999, 55, 909.
6. Nakagawa, Y.; Smagghe, G.; Van Paemel, M.; Tirry, L.;
Fujita, T. Pest Manag. Sci. 2001, 57, 858.
7. Nakagawa, Y.; Smagghe, G.; Tirry, L.; Fujita, T. Pest
Manag. Sci. 2002, 58, 131.
8. Hsu, A. C.-T.; Fujimoto, T. T.; Dhadialla, T. S. Phyto-
chemicals for Pest Control; American Chemical Society:
Washington, DC, 1997; pp 206–219.
9. Carlson, G. R.; Dhadialla, T. S.; Hunter, R.; Jansson, R.
K.; Jany, C. S.; Lidert, Z.; Slawecki, R. A. Pest Manag.
Sci. 2001, 57, 115.
24. Sakuma, M. Appl. Entmol. Zool. 1998, 33, 339.
25. Asao, M.; Shimizu, R.; Nakao, K.; Fujita, T. Japan
Chemistry Program Exchange; Society of Computer
Chemistry: Japan, 1997.
26. Leo, A. J. Chem. Rev. 1993, 93, 1281.
27. Nakagawa, Y.; Hattori, K.; Minakuchi, C.; Kugimiya, S.;
Ueno, T. Steroids 2000, 65, 117.
28. Nakagawa, Y.; Shimizu, B.; Oikawa, N.; Akamatsu, M.;
Nishimura, K.; Kurihara, N.; Ueno, T.; Fujita, T.
Classical and Three-Dimensional QSAR in Agrochemistry;
American Chemical Society: Washington, DC, 1995; pp
288–301.
29. Wold, S.; Ruhe, A.; Wold, H.; Dunn, W. J., III SIAM J.
Sci. Stat. Comput. 1984, 5, 735.
30. Cramer, R. D., III Perspect. Drug Discovery Des. 1993, 1,
269.
31. Hansch, C.; Leo, A. J. Exploring QSAR: Fundamentals
and Applications in Chemistry and Biology; American
Chemical Society: Washington DC, 1995.
32. Verloop, A.; Hoogenstraaten, W.; Tipker, J. Drug Design;
Academic: New York, 1976; pp 165–201.