Structure-activity relationships of phenylpropanoids as antifeedants for the pine weevil Hylobius abietis

Article abstract of DOI:10.1007/s10886-008-9435-1

Ethyl cinnamate has been isolated from the bark of Pinus contorta in the search for antifeedants for the pine weevil, Hylobius abietis. Based on this lead compound, a number of structurally related compounds were synthesized and tested. The usability of the Topliss scheme, a flow diagram previously used in numerous structure-activity relationship (SAR) studies, was evaluated in an attempt to find the most potent antifeedants. The scheme was initially followed stepwise; subsequently, all compounds found in the scheme were compared. In total, 51 phenylpropanoids were tested and analyzed for SARs by using arguments from the field of medicinal chemistry (rational drug design). Individual Hansch parameters based on hydrophobicity, steric, and electronic properties were examined. The effects of position and numbers of substituents on the aromatic ring, the effects of conjugation in the molecules, and the effects of the properties of the parent alcohol part of the esters were also evaluated. It proved difficult to find strong SARs derived from single physicochemical descriptors, but our study led to numerous new, potent, phenylpropanoid antifeedants for the pine weevil. Among the most potent were methyl 3-phenylpropanoates monosubstituted with chloro, fluoro, or methyl groups and the 3,4-dichlorinated methyl 3-phenylpropanoate.

Full text of DOI:10.1007/s10886-008-9435-1

J Chem Ecol (2008) 34:339–352
DOI 10.1007/s10886-008-9435-1
Structure–Activity Relationships of Phenylpropanoids
as Antifeedants for the Pine Weevil Hylobius abietis
B. Bohman & G. Nordlander & H. Nordenhem &
K. Sunnerheim & A.-K. Borg-Karlson & C. R. Unelius
Received: 23 July 2007 /Revised: 19 December 2007 /Accepted: 16 January 2008 /Published online: 29 February 2008
#
Springer Science + Business Media, LLC 2008
Abstract Ethyl cinnamate has been isolated from the bark
of Pinus contorta in the search for antifeedants for the pine
weevil, Hylobius abietis. Based on this lead compound, a
number of structurally related compounds were synthesized
and tested. The usability of the Topliss scheme, a flow
diagram previously used in numerous structure–activity
relationship (SAR) studies, was evaluated in an attempt to
find the most potent antifeedants. The scheme was initially
followed stepwise; subsequently, all compounds found in
the scheme were compared. In total, 51 phenylpropanoids
were tested and analyzed for SARs by using arguments
from the field of medicinal chemistry (rational drug
design). Individual Hansch parameters based on hydropho-
bicity, steric, and electronic properties were examined. The
effects of position and numbers of substituents on the
aromatic ring, the effects of conjugation in the molecules,
and the effects of the properties of the parent alcohol part of
the esters were also evaluated. It proved difficult to find
strong SARs derived from single physicochemical descrip-
tors, but our study led to numerous new, potent, phenyl-
propanoid antifeedants for the pine weevil. Among the
most potent were methyl 3-phenylpropanoates monosub-
stituted with chloro, fluoro, or methyl groups and the 3,4-
dichlorinated methyl 3-phenylpropanoate.
.
.
.
Keywords Antifeedant Pine weevil Hylobius abietis
.
.
.
Structure–activity SAR Phenylpropanoid
.
Phenylpropanoate Phenylacrylate
Introduction
The pine weevil Hylobius abietis (L.) is a severe pest in
areas of Europe where clear-cutting of conifer forests with
subsequent replanting is practiced (Långström and Day
2004) because newly planted seedlings are frequently killed
by the feeding of adult weevils on the stem bark (Day et al.
2004). There is a strong demand in Sweden and other
European countries for new methods that prevent this pest
from damaging forest plantations without the use of
pesticides (Långström and Day 2004). Pine weevils
walking on the ground locate the seedlings by responding
to both olfactory and visual stimuli (Björklund et al. 2005).
Previously, it has been shown that lodgepole pine, Pinus
contorta Douglas ex Loudon, is not as severely attacked as
Scots pine, P. sylvestris L. (Bratt et al. 2001). This
observation led to the hypothesis that the bark of P.
contorta contains secondary metabolites that act as natural
antifeedants against H. abietis. This was confirmed in
laboratory bioassays where bark extracts from P. contorta
deterred feeding more than bark extracts from P. sylvestris.
Bioassay-guided fractionation of P. contorta bark led to the
:
B. Bohman C. R. Unelius (*)
School of Pure and Applied Natural Sciences,
University of Kalmar,
SE-391 82 Kalmar, Sweden
e-mail: rikard.unelius@hik.se
:
G. Nordlander H. Nordenhem
Department of Ecology,
Swedish University of Agricultural Sciences,
P.O. Box 7044, SE-750 07 Uppsala, Sweden
K. Sunnerheim
Department of Natural Sciences, Mid Sweden University,
SE-851 70 Sundsvall, Sweden
A.-K. Borg-Karlson
KTH Chemistry, Organic Chemistry,
Ecological Chemistry Group, Royal Institute of Technology,
SE-100 44 Stockholm, Sweden

340
J Chem Ecol (2008) 34:339–352
isolation and identification of two compounds, ethyl
cinnamate and ethyl 2,3-dibromo-3-phenylpropanoate,
which possessed antifeedant activity (Bratt et al. 2001).
Ethyl cinnamate and structurally related esters are more
potent antifeedants than the corresponding carboxylic acids
(Sunnerheim et al. 2007). The results of an investigation of
benzoate derivatives as antifeedants for H. abietis indicated
that even small changes in the structure of a derivative might
induce dramatic effects on the antifeedant activity (Unelius
et al. 2006). Therefore, it is important to understand the
relation between the substitution pattern of the aromatic ring
and also the antifeedant effect in phenylpropanoids.
When looking for new bioactive compounds, it might
not be feasible to test a large number of structures to find
the most potent one. Under these circumstances, it is
beneficial to test compounds for activity as they are
synthesized and to use the results to select the next analog
to be tested. These ideas are applied in the Topliss approach
(Topliss 1972), which we used as a tool in search for the
phenylpropanoid with the highest antifeedant effect. A
Topliss scheme is a flow diagram designed such that the
optimal substituent can be found expediently (Topliss
1977), and it gives recommendations on how to proceed
after each compound has been tested. The substituents in
the Topliss scheme have been chosen based on physico-
chemical properties in combination with ease of synthesis.
The test approach drawn up by Topliss is a modified and
simplified quantitative SAR (QSAR) method developed
from the Hansch method for structure–activity correlations.
In the Hansch method, mathematical functions are used to
correlate biological activity to chemical structure. For
instance, the biological activities of a number of substituted
aromatic compounds on bacteria, insects, and mammals
have been examined and correlated to the properties of the
substituents (Hansch and Fujita 1963). Depending on the
structural configurations of the receptor active sites,
different factors may affect efficiency. In this study, the
relationships between individual Hansch parameters and
potency of antifeedants against the pine weevil are
presented.
The overall goal of the present study was to find new
effective antifeedants related to ethyl cinnamate that could
be used for conifer seedling protection. Throughout this
process, we wanted to evaluate the Topliss approach for
finding more potent antifeedants by using a low number of
syntheses and biological tests. Individual Hansch parame-
ters were also evaluated, as were the number, pattern, and
properties of substituents on the aromatic ring. Further-
more, a number of 3-phenylpropanoates and 3-phenyl-
acrylates were compared in search of a potential effect of
conjugation. In addition, methyl, ethyl, propyl, and butyl
esters of selected 3-phenylpropanoic and 3-phenylacrylic
acids were included to determine the importance of the parent
alcohol part of the esters. Finally, we compared the results
from structural alterations on the antifeedant activity of the
phenylpropanoids with the results from a similar previous
study on benzoic acid derivatives (Unelius et al. 2006) in
search for similarities and differences in selectivity.
Methods and Materials
Collection and maintenance of weevils Both sexes of H.
abietis were collected during spring migration at a saw mill
in southern Sweden where they landed in large numbers.
After collection, weevils were stored in darkness at 10°C
and provided with fresh branches of Scots pine, Pinus
sylvestris, as food. These storage conditions interrupted the
reproductive development so that females did not begin
ovipositing until about a week after they had been
transferred to the experimental conditions, i.e., a light
regime of L18/D6 at 22°C. This transfer was made about
10 d before the insects were used in the following bioassay.
Laboratory bioassay Compounds were tested for their
antifeedant effect on H. abietis by means of a two-choice
laboratory bioassay (Bratt et al. 2001; Legrand et al. 2004;
Borg-Karlson et al. 2006; Unelius et al. 2006). For each
test, 40 pine weevils (20 females+20 males) were starved
for 24 hr before the test period. Each weevil was placed in a
Petri dish (142 mm diameter) and provided with a Scots
pine twig (prepared as described below) placed on a
moistened filter paper. One day before the test, the twigs
were wrapped in aluminum foil, and two holes (diameter
5 mm; 25 mm apart) were punched in the foil with sharp-
edged metal rings. After the removal of the aluminum foil
inside the rings, one of the two areas was treated with
100 μl of a 50- or 5-mM methyl acetate solution of the
compound to be tested, and the other was treated with the
same amount of pure solvent as control. When the solvent
had evaporated on the following day, the metal rings were
removed and the bioassay started. After 24 hr, the
proportion of treated bark and control area of each test
To further explore the effects of structural changes in the
aromatic rings of compounds such as phenylpropanoids, the
number of substituents and the substitution pattern on the ring
can be examined and evaluated in relation to activity, i.e.,
antifeedant efficacy. Certain positions are expected to be
beneficial for a strong interaction to a target receptor. In
addition to monosubstituted analogs, compounds with two or
three substituents can also be tested and evaluated. As
biological activity is potentially dependent upon the electronic
properties of the entire molecule, it is important to investigate
whether acrylates or propanoates are the superior antifeedants.
The choice of parent alcohol for the ester is also likely to be
important (Unelius et al. 2006).

J Chem Ecol (2008) 34:339–352
341
twig that had been consumed was recorded. There was
generally no significant difference in response between the
sexes, and the data presented have been pooled.
Method B: Preparation of esters by reaction of the
carboxylic acid with dicyclohexylcarbodiimide
(DCC), 4-methylaminopyridin (DMAP), and
methanol in CH2Cl2. A typical example is the
synthesis of methyl-3-(4-(dimethylamino)phe-
nyl)acrylate: 3-(4-(dimethylamino)phenyl)
acrylic acid (3.15 mmol) and DCC (3.67 mmol)
were dissolved in methanol (35 ml). A solution
of DMAP (1.23 mmol) in CH2Cl2 (5 ml) was
added, and the reaction mixture was refluxed for
2 hr, concentrated with a rotary evaporator,
dissolved in CH2Cl2, filtered, washed with
water, and dried over MgSO4. Evaporation of
the solvent and purification by silica gel
chromatography yielded the product (2.44 mmol,
77%) as yellow crystals.
Method C: Preparation of propanoates by catalytic hydroge-
nation of acrylate analogs (Vogel 1989). A typical
example is the formation of methyl 3-(4-
(dimethylamino)phenyl)propanoate: Methyl 3-
(4-(dimethylamino)phenyl)acrylate (1.32 mmol)
was dissolved in methanol (30 ml). To the
solution, 39 mg Pd/C (10%) were added, and the
reaction mixture was set under an atmosphere of
hydrogen. When the reaction was completed
after 45 min, the catalyst was filtered off, and the
solvent evaporated. Methyl 3-(4-(dimethylamino)
phenyl)propanoate was obtained (0.83 mmol,
63%) as slightly yellow crystals.
Method D: Preparation of propanoates by O-alkylation of
hydroxyl-substituted analogs (Legrand et al.
2004): Isopropyl 3-(4-methoxyphenyl)propa-
noate was obtained by reacting isopropyl 3-(4-
hydroxyphenyl)propanoate with methyl iodide
and potassium carbonate in acetone. Methyl 3-
(4-butoxyphenyl)propanoate was obtained by
reacting methyl 3-(4-hydroxyphenyl)propa-
noate with potassium hydroxide and butyl
iodide. Methyl 3-(3,4-dimethoxyphenyl)prop-
anoate was obtained by reacting methyl 3-(3,4-
dihydroxyphenyl)propanoate with sodium
hydride and methyl iodide in THF according
to the standard procedure (Vogel 1989).
Method E: Preparation of 3-phenylacrylates by Knoevenagel
reaction (Harwood et al. 1998) followed by
acid-catalyzed esterification. A typical exam-
ple is the synthesis of methyl 3-(4-ethyl-
phenyl)acrylate: Malonic acid (10.4 mmol)
was dissolved in pyridine (3 ml) with heating.
4-Ethylbenzaldehyde (10.2 mmol) and a cata-
lytic amount of piperidine (10 drops) were
added while stirring. The reaction mixture was
refluxed until the production of carbon dioxide
The effects of the various treatments are described by
two alternative antifeedant indices (AFI) (Blaney et al.
1984) with the general formula: 100×(C−T)/(C+T). The
two variants are: (1) antifeedant index, area (AFIa) where C
represents the mean area of the control surfaces consumed
and T represents the mean area of the treated surfaces
consumed and (2) antifeedant index, number (AFIn) where
C represents the number of the control surfaces with
feeding scars and T represents the number of the treated
surfaces with feeding scars.
Thus, AFIa is a measure that captures the reduction in
feeding, whereas AFIn is a measure of the frequency of
complete inhibition of the initiation of feeding. The two
indices were fairly well correlated, but AFIa tended to be
higher than AFIn because the antifeedant substances
generally affected both the initiation of feeding and the
amount of bark consumed if feeding commenced. For both
indices, positive values (up to a maximum of 100) reflect an
antifeedant effect, whereas negative ones (down to a
minimum of −100) indicate a stimulant effect.
Test chemicals The origins of the compounds tested are
given in Tables 1, 2, and 3. Final purities ranged from 96%
to 99%. When necessary, compounds were purified by
preparative chromatography (Baeckström et al. 1987) or
flash chromatography on silica gel (Merck 60, 0.040–
0.063 mm, Darmstadt, Germany). The tested esters were
synthesized by methods A–F (see “Synthesis” below),
purchased from Sigma-Aldrich Sweden AB, Sweden or
obtained from previous work by H. Erdtman and T. Norin
at KTH Chemistry, Organic Chemistry, Stockholm. All
cinnamic acid derivatives had the (E)-configuration.
Synthesis Commercial phenylpropanoic and phenylacrylic
acids were either esterified by using methods A or B or the
esters were formed according to methods C–F.
Method A: Acid-catalyzed esterification from corresponding
commercial phenylpropanoic or -phenylacrylic
acids. A typical procedure is given for the
synthesis of methyl 3-(4-methylphenyl)acrylate:
3-(4-methylphenyl)acrylic acid (1.2 mmol) was
dissolved in methanol (11 ml). Concentrated
sulfuric acid (three drops) was added. The
mixture was refluxed for 4 hr, concentrated with
a rotary evaporator, diluted with water, and
extracted with ethyl ether. The ether phase was
washed with Na2CO3 (aq), dried over MgSO4,
and evaporated to give white crystals
(1.15 mmol, 93%).

342
J Chem Ecol (2008) 34:339–352
Table 1 Structures of mono-
substituted phenylpropanoids
and physical properties -
Hansch parameter values for
substituents on the aromatic
rings (Hansch et al. 1973;
Topliss 1972; Swain and
Lupton 1968)
Entry
1
Structure
Name of compound
σ
π
F
R
MR MW
O
a
a
a
a
a
0
1
0
0
0
0
OMe
Methyl 3-phenylpropanoate
O
OMe
2
3
0.23 0.71 0.41 -0.15 6.03 35
N/A 0.71 0.41 -0.15 6.03 35
0.37 0.71 0.41 -0.15 6.03 35
0.60 1.42 0.82 -0.30 12.1 71
0.54 0.88 0.38 0.19 5.02 69
0.54 0.88 0.38 0.19 5.02 69
0.23 0.86 0.44 -0.17 8.88 80
0.06 0.14 0.43 -0.34 0.92 19
-0.27 -0.02 0.26 -0.51 7.87 31
N/A -0.02 0.26 -0.51 7.87 31
0.12 -0.02 0.26 -0.51 7.87 31
N/A 0.56 -0.04 -0.13 5.65 15
-0.07 0.56 -0.04 -0.13 5.65 15
-0.17 0.56 -0.04 -0.13 5.65 15
-0.17 0.56 -0.04 -0.13 5.65 15
Methyl 3-(4-chlorophenyl)propanoate
Methyl 3-(2-chlorophenyl)propanoate
Methyl 3-(3-chlorophenyl)propanoate
Methyl 3-(3,4-dichlorophenyl)propanoate
Cl
Cl
O
Cl
OMe
O
4
OMe
O
Cl
Cl
5
OMe
O
OMe
b
a
6
Methyl 3-(4-trifluoromethylphenyl)propanoate
F3C
F3C
O
OMeMethyl 3-(4-trifluoromethylphenyl)acrylate
7
O
a
a
c
a
a
a
a
a
a
8
OMe
Methyl 3-(4-bromophenyl)propanoate
Methyl 3-(4-fluorophenyl)propanoate
Methyl 3-(4-methoxyphenyl)propanoate
Methyl 3-(2-methoxyphenyl)propanoate
Methyl 3-(3-methoxyphenyl)propanoate
Methyl 3-(2-methylphenyl)propanoate
Methyl 3-(3-methylphenyl)propanoate
Methyl 3-(4-methylphenyl)propanoate
Methyl 3-(4-methylphenyl)acrylate
Br
O
9
OMe
F
O
OMe
10
11
12
13
14
15
MeO
O
MeO
OMe
O
MeO
OMe
O
OMe
O
OMe
O
OMe
O
16
OMe
a
prepared from the corresponding carboxylic acids by refluxing in the alcohol with H2SO4 as a catalyst: Method A
b prepared by catalytic hydrogenation of acrylates: Method C.
c
obtained from previous work by H. Erdtman and T. Norin at the Dep. of Organic Chemistry, KTH, Stockholm.
abated. HCl (aq, 2 M, 10 ml) was added, and
the precipitate formed was filtered off by
Method F: Methyl 3-(4-acetyloxyphenyl)propanoate was
prepared from methyl 3-(4-hydroxyphenyl)
propanoate according to the standard proce-
dure (Vogel 1989).
suction and washed with HCl (aq, 2 M,
10 ml), water (10 ml), and hexane (10 ml) and
dried in vacuo to yield methyl 3-(4-ethyl-
phenyl)acrylic acid (7.8 mmol, 76%). The
carboxylic acid was esterified according to
method A.
All reactions were monitored by thin layer chromatography.
The spectroscopic data of the products were analyzed and
compared with literature data.

J Chem Ecol (2008) 34:339–352
Table 1 (continued)
343
Entry Structural formula
Compound
σ
π
F
R
MR MW
O
d
b
b
b
e
e
f
17
18
19
20
21
22
23
24
25
26
27
-0.15 1.02 -0.05 -0.10 10.3 29
-0.15 1.53 -0.05 -0.10 15.0 43
-0.15 1.53 -0.05 -0.10 15.0 43
0.78 -0.28 0.67 0.16 7.36 46
OMe
Methyl 3-(4-ethylphenyl)acrylate
Methyl 3-(4-isopropylphenyl)propanoate
Methyl 3-(4-isopropylphenyl)acrylate
Ethyl 3-(4-nitrophenyl)propanoate
Methyl 3-(4-aminophenyl)propanoate
O
OMe
O
OMe
O
O
NO2
H2N
O
OMe
-0,66 -1,23
0 -0.68 5.42 16
O
O
OMe Methyl 3-(4-dimethylaminophenyl)propanoate
-0.83 0.18 0.10 -0.92 15.6 44
-0.83 0.18 0.10 -0.92 15.6 44
N
N
OMe
Methyl 3-(4-dimethylaminophenyl)acrylate
Methyl 3-(4-aminophenyl)acrylate
O
b
b
OMe
-0,66 -1,23
0 -0.68 5.42 16
H2N
O
OMe
O
0.78 -0.28 0.67 0.16 7.36 46
-0.32 1.55 0.25 -0.55 21.7 73
-0.37 -0.67 0.29 -0.64 2.85 17
0.31 -0.64 0.41 -0.07 12.5 59
Methyl 3-(4-nitrophenyl)acrylate
O2N
g
OMe
Methyl 3-(4-butoxyphenyl)propanoate
Methyl 3-(4-hydroxyphenyl)propanoate
Methyl 3-(4-acetyloxyphenyl)propanoate
O
O
h
i
OMe
HO
O
OMe
28
AcO
d
e
f
prepared via a Knoevenagel condensation: Method E.
prepared by catalytic hydrogenation of cinnamates: Method C.
prepared from the carboxylic acid by reaction with DCC, DMAP and methanol in CH2Cl2: Method B.
prepared by O-alkylation of hydroxy-substituted analogs: Method D.
purchased from SigmaAldrich Co, Sweden.
g
h
i
prepared by O-acetylation of hydroxy-substituted analogs: Method F.
Spectroscopy H NMR (400 or 250 MHz) and C NMR (100
or 63 MHz) spectra were recorded on Varian 400, Bruker
400, or Bruker 250 instruments using the solvent signals,
CDCl3 or CD3OD, as internal standards.
Hansch parameters The Hansch equation (Eq. 1) contains
a number of parameters that reflect different properties of
substituents on aromatic rings (Hansch and Deutsch 1966),
such as hydrophobicity (π), electronic effects (σ, F, and R)
and steric factors (molar refractivity [MR] and molecular
weight [MW]). The relation between steric parameters can
be found in the Lorentz–Lorentz equation (Eq. 2). One
expression for the relationship between electronic effects is
also shown (Eq. 3).
The Topliss scheme The first substance in the Topliss
scheme possesses an unsubstituted aromatic ring (or a
phenylic compound with one or more substituents not being
changed over the time of the study) (Topliss 1972). In the
modified Topliss scheme used, we started with methyl 3-
phenylpropanoate (Fig. 1). The analogs were tested at
50 mM concentration in methyl acetate with the solvent as
control treatment. As the AFI values for many substances
were close to 100 at 50 mM, new tests were conducted at
5 mM concentration. The Topliss scheme was used
continuously in the dynamic experimental plan, i.e., after
the synthesis of each compound and after the evaluation of
the test results from bioassays, the scheme was consulted
for the choice of compound next to be synthesized and
tested. Finally, analogs from all positions of the scheme
were synthesized and tested to critically evaluate the
method.
2
Logð1=CÞ ¼ k1ðlog πÞ þk2 log π þ k3σ þ k4MR þ k5 ð1Þ
À
Áꢀ À
Á Â
ꢀ
Ã
MR ¼ MW n2 ꢀ 1 d n2 þ 2 cm3 mol
ð2Þ
ð3Þ
R ¼ s ꢀ k6F
C is the concentration required for biological activity, k1–k6
are the system-dependent constants, n is the refraction
index, and d is the density.
In this study, the hydrophobic parameter used was π,
derived from log P values (Hansch et al. 1973). Three

344
J Chem Ecol (2008) 34:339–352
Table 2 Structures of disub-
stituted and trisubstituted phe-
nylpropanoids and physical
properties - Hansch parameter
values for substituents on the
aromatic rings (Hansch et al.
1973; Topliss 1972; Swain and
Lupton 1968)
Entry Structural formula
Compound
σ
π
F
R
MR MW
O
MeO
MeO
a
a
a
a
b
c
b
a
a
a
a
a
b
a
29
30
31
32
33
34
35
36
37
38
39
40
41
42
N/A -0.04 0.52 -1.02 15.7 62
N/A -0.04 0.52 -1.02 15.7 62
-0.15 -0.04 0.52 -1.02 15.7 62
0.24 -0.04 0.52 -1.02 15.7 62
N/A -0.04 0.52 -1.02 15.7 62
-0.15 -0.04 0.52 -1.02 15.7 62
0.24 -0.04 0.52 -1.02 15.7 62
0.12 0.84 0.70 -0.68 16.8 111
-0.25 -0.69 0.55 -1.15 10.7 48
-0.03 -0.06 0.78 -1.53 23.6 93
N/A 1.12 -0.08 -0.26 11.3 30
-0.24 1.12 -0.08 -0.26 11.3 30
0.66 1.59 0.79 0.04 11.0 104
0.66 1.59 0.79 0.04 11.0 104
OMe
OMe
Methyl 3-(2,3-dimethoxyphenyl)acrylate
Methyl 3-(2,4-dimethoxyphenyl)acrylate
O
MeO
MeO
O
O
MeO
OMe
OMe
Methyl 3-(3,4-dimethoxyphenyl)acrylate
MeO
MeO
Methyl 3-(3,5-dimethoxyphenyl)acrylate
MeO
MeO
MeO
O
O
OMe
OMe
Methyl 3-(2,3-dimethoxyphenyl)propanoate
Methyl 3-(3,4-dimethoxyphenyl)propanoate
Methyl 3-(3,5-dimethoxyphenyl)propanoate
Methyl (3-bromo-4-methoxyphenyl)propanoate
Methyl 3-(4-hydroxy-3-methoxyphenyl)propanoate
Methyl 3-(3,4,5-trimethoxyphenyl)propanoate
Methyl 3-(2,4-dimethylphenyl)propanoate
Methyl 3-(3,4-dimethylphenyl)propanoate
Methyl 3-(4-chloro-3-trifluoromethylphenyl)propanoate
Methyl 3-(4-chloro-3-trifluoromethylphenyl)acrylate
MeO
O
MeO
MeO
OMe
MeO
O
O
Br
OMe
OMe
MeO
MeO
HO
MeO
O
OMe
MeO
MeO
O
O
OMe
OMe
O
F3C
OMe
Cl
O
F3C
OMe
Cl
a
b
c
prepared from the corresponding carboxylic acids by refluxing in the alcohol with H2SO4 as a catalyst: Method A.
prepared by catalytic hydrogenation of cinnamates: Method C.
prepared by O-alkylation of hydroxy-substituted analogs: Method D.
Table 3 Structures of phenyl-
propanoids (variations in the
parent alcohol parts) and phys-
ical properties - Hansch pa-
rameter values for substituents
on the aromatic rings (Hansch
et al. 1973; Topliss 1972;
Entry Structural formula
Compound
σ
π
F
R
MR MW
O
O
O
a
a
a
a
a
a
b
a
a
43
44
45
46
47
48
49
50
51
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
O
O
O
Ethyl cinnamate
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Propyl cinnamate
Isopropyl cinnamate
Swain and Lupton 1968)
O
O
Butyl cinnamate
O
O
O
O
2-Butyl cinnamate
Ethyl 3-phenylpropanoate
O
O
-0.27-0.02 0.26 -0.51 7.87 31
-0.17 0.56 -0.04 -0.13 5.65 15
0.23 0.71 0.41 -0.15 6.03 35
Isopropyl 3-(4-methoxyphenyl)propanoate
Ethyl 3-(4-methylphenyl)propanoate
Ethyl 3-(4-chlorophenyl)propanoate
MeO
O
O
O
O
Cl
H
O
O
H
a
b
prepared from the corresponding carboxylic acids by refluxing in the alcohol with H2SO4 (catalyst): Method A.
prepared by O-alkylation of hydroxyl-substituted analogs: Method D.

J Chem Ecol (2008) 34:339–352
345
H
4-Cl
4-OMe
3-Cl
4-Me
3,4-Cl2
4-CF3 / 4-Br
4-NO2
3-Cl
4-N(Me)2
4-NH2
3-Cl
3-Me
3,4-Me2
4-CF3 / 4-Br
4-NO2
3-CF3, 4-Cl
4-NH2
2-Cl / 2-Me / 2-OMe
4-NO2
4-F
Fig. 1 The complete modified Topliss scheme used in this study.
Substituents on the aromatic ring are shown for each analog. From
each “parent” compound (from the top of the scheme) the choice of
the next analog is shown, depending on the result of the previous test.
If the potency increased for the latest tested compound, the next
compound to be tested is the one to the right in the next lower level of
the scheme, i.e., first, the unsubstituted compound and the 4-chloro-
substituted compound is tested, and if the latter compound is more
active, the 3,4-dichloro analog is tested next. If the activity is equally
high, the analog straight below is tested, i.e., if the 4-Cl analog is
equally active as the 4-H, the 4-Me analog is tested next. If the activity
is lower, the analog below to the left is tested, i.e., if the 4-Cl analog is
less active than the unsubstituted compound, the 4-OMe analog is
tested next. Our results directed us to conduct the test series indicated
by shaded, bold boxes. The results (AFIn) are shown in circle
diagrams below each analog (filled black circle AFIn=100, empty
gray circle AFIn=0; all tests carried out at 50 mM)
different electronic parameters were applied: σ, F, and R.
The widely used σ parameter has the disadvantage that the
resonance effect contributes for metasubstituted aromatic
rings (σ values for orthosubstituted compounds have not
been included), whereas for parasubstituted systems, the
field factors affect the values of σ (Swain and Lupton
1968). For the parameters F and R, the field effects (F) and
resonance effects (R) are separated. The steric parameters
used to measure the “bulk” of the molecules were
molecular weight (MW) and molar refractivity (MR). MR
differs from MW in having an electronic contribution, as it
is directly proportional to polarizability (Eq. 2; Hansch
et al. 1973). The values of all parameters used in this study
are available in the literature (Hansch and Deutsch 1966;
Swain and Lupton 1968 and Hansch et al 1973).
hydrophobic and electron-withdrawing 4-chloro compound
(entry 2, Table 1), which was more active than the
unsubstituted compound. The next analog to be tested
according to the scheme was the 3,4-dichloro compound
(entry 5, Table 1), which was even more hydrophobic, and
the ring was more electron-poor. This analog showed
similar or slightly lower activity. The 4-CF3 analog (entry
6, Table 1), which did not have a metasubstituent and was
less lipophilic than 3,4-Cl but more lipophilic than 4-Cl,
was tested next and was less active than the previous one.
The 4-Br analog (entry 8, Table 1), an alternative to 4-CF3,
also showed low activity. Finally, the analog with the less
lipophilic but more electron-withdrawing 4-NO2 substituent,
can be found in the bottom of scheme 1. Ethyl 3-(4-
nitrophenyl)propanoate (entry 20, Table 1) was tested instead
of methyl 3-(4-nitrophenyl)propanoate but the results can be
compared with the results from ethyl 3-phenylpropanoate
(entry 48, Table 3), and it is obvious that nitro analogs have
relatively low activities (Fig. 2a and Table 1).
Results
Evaluation of the Topliss method for finding potent methyl
3-phenylpropanoate antifeedants To determine fully the
usability of this approach, the scheme was initially followed
step by step (as was intended in Topliss’ method); but in a
second phase of the study, all compounds found in all
branches of the scheme were obtained and compared.
The investigation started with methyl 3-phenylpropanoate
(entry 1, Table 1). The first analog to be tested was the more
To find out if the Topliss method generated the optimal
antifeedant, we decided to investigate all three branches in
parallel.
Following the right branch of the Topliss scheme, all
compounds except methyl 3(4-chloro-3-trifluoromethyl-
phenyl)propanoate (entry 41, Table 2) were tested, as the
scheme was followed systematically. This compound was
less active than the analogs substituted with 4-Cl (entry 2,

346
J Chem Ecol (2008) 34:339–352
Table 1), 3,4-Cl (entry 5, Table 1), and 4-Br (entry 8,
Table 1) but more active than the 4-CF3 (entry 6, Table 1)
and 4-NO2 (entry 20, Table 1) analogs (Fig. 2a).
Effects of unsaturation and conjugation To understand the
importance of conjugation to the carbonyl group in the
phenylpropanoids, several substituted methyl 3-phenylpro-
panoates were compared with their corresponding 3-
phenylacrylates (Fig. 4). Most of the pairs of conjugated
(3-phenylacrylates) and nonconjugated compounds (3-phe-
nylpropanoates) had similar activities (Fig. 4). However,
when we compared the activities of the analogs with the
most electron-withdrawing and electron-donating groups,
we found a trend. The activities of the compounds with
electron-withdrawing groups (4-CF3 and 4-NO2) (entries 6–
7, 20, 25, Table 1) were favored by conjugation, whereas
the activity of compounds with electron-donating substitu-
ents (4-N(CH3)2 and 4-NH2) (entries 21–24, Table 1)
decreased for the conjugated analogs (Fig. 4b).
In the center branch, the 4-methyl (entry 15, Table 1) and
3,4-dimethyl analogs (entry 40, Table 2) were tested next.
Further down this branch, steric factors were examined by
testing the 2- and 3-substituted analogs, i.e., the 3-chloro
(entry 4, Table 1) and the 2-chloro (entry 3, Table 1), 2-
methyl (entry 13, Table 1) and 2-methoxy (entry 11,
Table 1) analogs. These compounds generally had lower
activity than the 4-substituted analogs but the electronic
factors did not seem to greatly affect the activity. The 4-F
analog (entry 9, Table 1) has similar values of the Hansch
parameters but showed higher potency than the parent
unsubstituted methyl 3-phenylpropanoate (Fig. 2b).
For the left branch of the scheme, the choices after the
electron-donating 4-methoxy analog (entry 10, Table 1)
were other more electron-donating compounds: the 4-N
(Me)2 (entry 22, Table 1) and the 4-NH2 (entry 21, Table 1)
analogs. The dimethylamino compound had similar potency
compared with the 4-methoxy compound, whereas the
amino compound was much less potent (Fig. 2c). As a
hypothesis, the polarity of the amino group is the cause of
the low antifeedant activity.
To be able to better compare the most potent antifee-
dants, which all had AFI values close to 100, additional
tests were carried out at 5 mM concentration. The results
from these tests showed a stronger decrease in activity at
the lower concentration for the methoxy analogs than for
the analogs with chloro, bromo, fluoro, and methyl
substituents (Fig. 3).
As potent antifeedants were found in all branches of the
Topliss scheme and not preferentially among the analogs
obtained when the proposed route was followed, the Topliss
approach was abandoned. Consequently, we continued our
survey of the SARs of phenylpropanoids by using other
methods.
Comparison of disubstituted and trisubstituted compounds We
also investigated whether it was possible to improve the
activity by adding more substituents to the aromatic ring
and by optimizing the substitution pattern. Consequently,
several compounds with two or three substituents were
tested (Fig. 3b). Different substitution patterns were
compared for dimethoxy- and trimethoxy-substituted
(entries 33–35, 38, Table 2) methyl 3-phenylpropanoates.
For the dimethoxy-substituted compounds, there was a
strong effect favoring 2,3- and 3,5-substitution over 3,4-
and 3,4,5-substitution. The 2,3- and 3,5-disubstituted
methoxy analogs were slightly less active compared with
the monosubstituted methoxy analogs (Fig. 3b). For
disubstituted methyl analogs (entries 39–40, Table 2), the
2,4-analog was slightly more potent than the 3,4-analog,
but the difference was minor. The disubstituted methyl
compounds were approximately as potent as the monosub-
stituted analogs (Fig. 3c). Furthermore, methyl 3-(4-ethyl-
phenyl)acrylate (entry 17, Table 1) was tested to establish
whether the larger and more hydrophobic ethyl group alone
could enhance the activity, but this compound was less
active than the ones previously tested.
Effect of the position of substituents on the aromatic ring To
see whether substituents on the aromatic ring should be
situated in the ortho, meta, or para position for optimal
activity, tests were performed for all monosubstituted
isomers of the methyl (entries 13–15, Table 1), chloro
(entries 2–4, Table 1), and methoxy (entries 10–12, Table 1)
analogs (Fig. 3a). For the methyl-substituted compounds, the
difference in position did not greatly affect the potency,
neither on AFIn nor AFIa. For the monochlorinated
compounds, there was a larger difference in activity in the
order para>meta>ortho. For the methoxy analogs, the
metasubstituted and parasubstituted compounds showed
similar potency in all tests, whereas it is interesting to note
that the activity of the orthosubstituted analog was almost nil
at 5 mM (Fig. 3a).
The 3,4-dichlorinated analog (entry 5, Table 1) was tested
and compared with the monohalogenated analogs. It
showed similar potency as the 3-chloro, (entry 4, Table 1)
but slightly less potency than the 4-chloro compound (entry
2, Table 1).
A series of dimethoxy-substituted 3-phenylacrylates
(entries 29–32, Table 2) were also tested for the influence
of substitution pattern on activity. The order of reactivity
between combinations of substituents was 2,3>2,4≫3,5>3,4
(Fig. 4a). These results corresponded well with the results
from the dimethoxyphenylpropanoates above.
Activity of phenylpropanoates with mixed substituents -
Three analogs with mixed substituents were added to the

J Chem Ecol (2008) 34:339–352
347
a
b
100
90
80
70
60
50
40
30
20
10
0
100
90
80
70
60
50
40
30
20
10
0
3-CF3,4-
Cl
4-NO2*
(20)
4-Br
(8)
4-Cl
(2)
4-CF3
(6)
4-H
(1)
3,4-Cl
(5)
4-H 4-Cl 4-Me 3,4-Me 3-Cl 3-Me 2-Cl 2-Me 2-OMe 4-NO2* 4-F
(1) (2) (15) (40) (4) (14) (3) (13) (11)
(20) (9)
(41)
Compound (substituent and entry)
Compound (substituent and entry)
c
d
100
90
80
70
60
50
40
30
20
10
0
100
90
80
70
60
50
40
30
20
10
0
2,4-Me
(39)
4-Me
(15)
2-Me
(13)
4-Br
(8)
3,4-Cl
(5)
4-Cl
(2)
4-OMe
(10)
3-Cl
(4)
4-N(Me)2
(22)
4-NH2
(21)
4-Cl
(2)
4-F
(9)
Compound (substituent and entry)
Compound (substituent and entry)
Fig. 2 AFI for substituted 3-phenylpropanoates in the Topliss
scheme. AFIn is represented by dark columns and AFIa by light
columns. Substituent(s) and entry number (in parentheses) are given
beneath the x-axis: a the right branch (50 mM concentration), ethyl 3-
(4-nitrophenyl)propanoate is denoted by an asterisk; b the center
branch (50 mM concentration), ethyl 3-(4-nitrophenyl)propanoate is
denoted by an asterisk; c the left branch (50 mM concentration); d
AFIn for the most potent compounds from the Topliss scheme tested at
50 mM (dark columns) and 5 mM concentrations (dotted columns)
a
b
100
90
80
70
60
50
40
30
20
10
0
3-OMe
(12)
2-OMe 3,4-OMe 3,4,5-OMe 2,3-OMe 3,5-OMe
4-OMe
(10)
4-Me 3-Me 2-Me 4-Cl 3-Cl 2-Cl 4-OMe
3-OMe 2-OMe
(12) (11)
(11)
(34)
(38)
(33)
(35)
(15) (14) (13)
(2)
(4)
(3) (10)
Compound (substituent and entry)
Compound (substituent and entry)
c
d
100
100
90
80
70
60
50
40
30
20
10
0
90
80
70
60
50
40
30
20
10
0
4-Cl
(2)
3-Cl
(4)
2-Cl
(3)
3,4-Cl
(5)
4-Br
(8)
4-Me
(15)
3-Me
(14)
2-Me
(13)
2,4-Me
(39)
3,4-Me
(40)
4-Et**
(17)
Compound (substituent and entry)
Compound (substituent and entry)
Fig. 3 AFI for substituted 3-phenylpropanoates. AFIn is represented
by dark columns and AFIa by light columns. Substituent(s) and entry
number (in parentheses) are given beneath the x-axis: a methyl-,
chloro-, and methoxy-substituted analogs (5 mM concentration, dotted
columns); b methoxy-substituted analogs (50 mM concentration); c
methyl-substituted analogs (50 mM concentration), methyl 3-(4-
ethylphenyl)acrylate is denoted by two asterisks; d chloro- and
bromo-substituted analogs (at 50 mM concentration)

348
J Chem Ecol (2008) 34:339–352
a
b
100
90
80
70
60
50
40
30
20
10
0
100
90
80
70
60
50
40
30
20
10
0
4-NH2
(24+21)
4-Cl, 3-CF3
(42+41)
4-N(Me)2 4-CH(Me)2
(23+22) (19+18)
4-NO2*
(25+20)
4-Me
3,4-OMe
(31+34)
2,4-OMe**
(30)
2,3-OMe
(29+33)
3,5-OMe
(32+35)
4-CF3
(7+6)
(16+15)
Compound (substituent and entry)
Compound (substituent and entry)
Fig. 4 AFI for methyl 3-phenylacrylates (acr; striped columns) and
methyl 3-phenylpropanoates (pro; plain columns) at 50 mM concen-
tration. AFIn is represented by dark columns and AFIa by light
columns. Substituent(s) and entry number (in parentheses) are given
beneath the x-axis. Each group of compounds with identical
substitution pattern are shown in the following order: AFIn(acr),
AFIa(acr), AFIn(pro), and AFIa(pro). Ethyl 3-(4-nitrophenyl)propa-
noate tested instead of methyl 3-(4-nitrophenyl)propanoate is denoted
by an asterisk; 3-phenylpropanoate analog not tested is denoted by
two asterisks. a Methoxy-substituted analogs; b analogs with
substituents other than methoxy
test series. These were the 3-bromo-4-methoxy (entry 36,
Table 2), 4-chloro-3-trifluoromethyl (entry 41, Table 2), and
4-hydroxy-3-methoxy analogs (entry 37, Table 2). The first
compound was less active than the corresponding mono-
substituted analogs ((3-Br, 4-OCH3) vs. 4-Br and 4-OCH3).
The (3-CF3, 4-Cl) analog was less potent than 4-Cl but
more potent than 4-CF3, and the (3-OCH3, 4-OH) analog
was a more active antifeedant than 4-OH (Fig. 5a).
Although not all possible variations were tested, the
conclusion from our test series was that no beneficial
effects could be found by mixing different substituents on
the aromatic ring.
Influences by Hansch parameters By comparing Hansch
parameters in relation to antifeedant activity for a series of
substituted methyl 3-phenylpropanoates, the effect of each
parameter was analyzed. The aim was to find the
importance of each parameter and thus the properties of
the substituents that affect antifeedant activity. In total, 29
methyl 3-phenylpropanoates and 11 3-phenylacrylates were
compared.
The Hansch parameters that were examined for correla-
tions to antifeedant activity were σ, π, F, R, MR, and MW
(Tables 1, 2, and 3). For the electronic parameter σ and the
field effect (F), there was no apparent correlation to
antifeedant potency. A similar result was found after
analyses of the steric parameters (MR and MW), i.e., small
and light molecules as well as bulky and heavy compounds
showed good antifeedant activity. For the lipophilicity
parameter (π), there was a correlation with antifeedant
potency. If π was negative (π<−0.1), the activity was, on
average, lower (Fig. 6a). The resonance factor R also
showed a correlation to AFI: No compound with good
antifeedant potency had R<−0.5 (Fig. 6b). For 3-phenyl-
acrylates, the results were similar.
Effect of the parent alcohol part of the esters on antifeedant
activity of phenylpropanoids The methyl (entry 10, Table 1)
and isopropyl (entry 49, Table 3) esters of 3-(4-methox-
yphenyl)propanoate and five alkyl cinnamates (ethyl-, n-
propyl-, isopropyl-, n-butyl-, and 2-butylcinnamate) (entries
43–47, Table 3) were compared to elucidate the effect of
the parent alcohol of the esters. All compounds were tested
at 50 mM concentration. Isopropyl cinnamate was the most
potent of the cinnamates (Fig. 5b) while methyl 3-(4-
methoxyphenyl)propanoate was more potent than the
corresponding isopropyl ester analog (Fig. 5c). The butyl
esters were much less active in all tests, and therefore, no
esters derived from higher alcohols were tested.
Furthermore, the methyl and ethyl esters of 3-phenyl-
propanoic acid, (entry 1, Table 1, vs. entry 48, Table 3), 3-
(4-methylphenyl)propanoic acid (entry 15, Table 1, vs.
entry 50, Table 3) and 3-(4-chlorophenyl)propanoic acid
(entry 2, Table 1, vs. entry 51, Table 3) were compared.
Apparently, no major difference in activity between methyl
and ethyl esters were detectable except for the unsubstituted
3-phenylpropanoate where the methyl ester was consider-
ably less active compared to the ethyl ester (Fig. 5c).
Comparison with benzoate antifeedants In a previous
study, we measured the AFI of a large number of benzoates
(Unelius et al. 2006). Methoxy-substituted benzoates
showed high antifeedant activity. Therefore, the following
analogous phenylpropanoids were tested in addition to the
compounds already discussed: Methyl 3-(4-butoxyphenyl)
propanoate, (entry 26, Table 1), methyl 3-(4-hydroxyphenyl)
propanoate (entry 27, Table 1), methyl 3-(4-acetyloxy-
phenyl)propanoate (entry 28, Table 1), and methyl 3-(4-
hydroxy-3-methoxyphenyl)propanoate (entry 37, Table 2).
In agreement with the results from the benzoate study, the
methoxy analogs (for example, 3-(4-methoxyphenyl)prop-

J Chem Ecol (2008) 34:339–352
349
a
b
100
90
80
70
60
50
40
30
20
10
0
100
90
80
70
60
50
40
30
20
10
0
4-Br 4-OMe 4-CF3 4-Cl
(8) (10) (6) (2)
4-OH 3-OMe 3-Br,4-
4-Cl,3- 4-OH,3-
n-Propyl
iso-Propyl
cinnamate cinnamate cinnamate cinnamate
(44) (45) (46) (47)
n-Butyl
2-Butyl
Ethyl cinnamate
(43)
(27)
(12)
OMe
(36)
CF3
(41)
OMe
(37)
Compound (substituent and entry)
Compound (substituent and entry)
c
d
100
100
90
80
70
60
50
40
30
20
10
0
90
80
70
60
50
40
30
20
10
0
4-BuO
(26)
4-OH
(27)
4-AcO 4-OH, 3-OMe 4-OMe
(28) (37) (10)
Me
(1)
Et Me, 4-0Me Isopropyl, Me, 4-Me Et, 4-Me Me, 4-Cl Et, 4-Cl
(51)
(48)
(10)
4-0Me
(49)
(15)
(50)
(2)
Compound (substituent and entry)
Compound (parent alcohol moiety, substituent, entry)
Fig. 5 AFI for 3-phenylpropanoids (at 50 mM concentration). AFIn is
represented by dark columns and AFIa by light columns. Substituent
(s) and entry number (in parentheses) are given beneath the x-axis: a
methyl 3-phenylpropanoates with mono and mixed substituents; b
cinnamates with different parent alcohol moieties; c 3-phenylpropa-
noates with different parent alcohol moieties; d derivatives of methyl
3-(4-hydroxyphenyl)propanoate
anoate; entry 10, Table 1) were most effective (Fig. 5d),
but, in contradiction to the results from the benzoate study,
the tested methyl- or halogen-substituted phenylpropanoids
had similar or even higher antifeedant activity than the
methoxy analogs.
SAR study. Nevertheless, the Topliss approach may be a
useful tool in other projects within the field of chemical
ecology where an effective semiochemical is to be found in
a limited time and at low costs.
To correlate the structures of the phenylpropanoids to their
antifeedant activity for H. abietis was difficult. Apart from
the result that highly polar substituents were unfavorable for
activity, it was difficult to discriminate any particular factors
that affected the antifeedant activity. The position of
substituents on the aromatic ring does not appear to be
critical for the antifeedant activity–a number of 2-, 3-, and 4-
substituted analogs had similar potency. The most potent
compounds tested were methyl 3-phenylpropanoates mono-
substituted with chloro, fluoro, or methyl groups, and the
3,4-dichlorinated methyl 3-phenylpropanoate.
Discussion
When initially applying the Topliss approach in our search
for potent antifeedants, the most active compounds hap-
pened to appear early in the scheme, so the use of the
remaining scheme became redundant. Normal rational
design was used, therefore, during the reminder of this
a
b
100
80
60
40
20
0
100
80
60
40
20
0
-1,5
-1
-0,5
0
0,5
1
1,5
2
-2
-1,5
-1
-0,5
0
0,5
R
π
Fig. 6 AFIn for methyl 3-phenylpropanoates (at 50 mM concentration) vs. values of various Hansch parameters: a AFIn vs. the hydrophobicity
parameter π; b AFIn vs. the steric parameter R

350
J Chem Ecol (2008) 34:339–352
In a comparison of 3-phenylpropanoates and acrylates,
A similar relationship was found for R where compounds
having R<−0.5 showed low AFIn and AFIa values, i.e., a
relatively high resonance potential seems to be necessary.
The steric parameters, MR and MW, showed no correlation
to antifeedant activity. When the steric factors for each
position (ortho, meta, para) on the aromatic ring were
correlated to the AFIs, no general trend was found, but the
number of compounds tested was limited (10 para, 3 meta
and 3 orthomonosubstituted compounds). These results may
be indicative of a multireceptor response that causes the
antifeedant effect, since small as well as bulky, and light as
well as heavy compounds all showed similar antifeedant
potency. This multireceptor hypothesis is supported by the
fact that most other parameters also did not correlate to AFI.
In the relatively few studies available on the responses of
other organisms to phenylpropanoids, there are both simi-
larities and dissimilarities with our results for H. abietis.
That esters of phenylpropanoids are more active than the
corresponding alcohols and acids was shown both in the
present study and in studies of derivatives of cinnamic acid
as bird repellents (Jacubas et al. 1992; Watkins et al. 1999).
An example of a compound with diverging effects between
organisms is ethyl 3-(4-nitrophenyl)acrylate, which showed
no antifeedant activity against H. abietis although it has
previously been found to be an oviposition deterrent for the
onion fly Delia antique (Meigen) (Cowles et al. 1990).
Another related compound, cinnamaldehyde, has been
proved to act as an antifeedant for grain storage insects
(Huang and Ho 1998). A comparison of the phenyl-
propanoid vs. the benzoate antifeedants of H. abietis
(Unelius et al. 2006) reveals a number of similarities:
Carboxylic acids are generally less effective antifeedants
than the corresponding esters, which also is in agreement
with the results from the study on H. abietis by Eriksson
(2006). Furthermore, the length of the alkyl chain in the
parent alcohol moiety of the esters should be short, and
methoxy-substituted aromatics are more potent than hy-
droxy-substituted in both classes of substances. On the
other hand, the prerequisites of the substitution pattern for
good antifeedant activity seem to differ between benzoates
and phenylpropanoids. For example, the best benzoate is
the 2,4-dimethoxy analog, whereas the corresponding 2,4-
dimethoxy-phenylpropanoids have rather low activity, and
the most effective propanoids, the halogen- or methyl-
substituted compounds, have very low AFIs as benzoate
analogs.
the lower flexibility caused by the double bond in the
acrylates may be an important factor that affects interac-
tions with a receptor. Nevertheless, we focused on the
electronic effects in our analyses of conjugated vs.
nonconjugated analogs. In compounds with electron-with-
drawing substituents on the aromatic ring, the antifeedant
effect increased with conjugation, and for compounds with
electron-donating substituents, the effect decreased for the
conjugated analogs. Electron richness in the aromatic ring
of the phenylpropanoids may be important for high activity.
Alternatively, or additionally, the carbonyl moiety should
be deprived of electron density. Eriksson (2006) studied the
antifeedant activity of 3-phenylpropenals and 3-phenyl-
propanals against H. abietis and concluded that conjugation
does not seem to be critical for activity.
In the comparison between monosubstituted and disub-
stituted compounds, the disubstituted methyl analogs were
slightly less potent than the monosubstituted. The decreased
potency cannot be correlated to a negative effect of the
increased bulkiness of the compounds as this effect was not
seen for the chloro analogs. The magnitude of difference
observed (more than tenfold difference) between the
disubstituted and trisubstituted methoxy analogs is intrigu-
ing. The 3,4- and 3,4,5-substituted saturated compounds
were much less active than the 2,3- and 3,5-substituted,
which is in agreement with the results from saturated
analogs of cinnamic aldehyde (Eriksson 2006). When we
compared the antifeedant effect of conjugated acrylate
analogs (3,4-OCH3, 2,4-OCH3, 2,3-OCH3, and 3,5-
OCH3), the results differed considerably. In this case, the
2,3- and 2,4-analogs were more potent than the 3,4- and
3,5-analogs, which is contradictory to the results for the
corresponding phenylpropenals (Eriksson 2006). Our
results indicate that there was an effect by conjugation.
For the compounds with mixed substituents, the (3-Br,
4-OCH3) analog was less potent than the monosubstituted
compounds, which implies that the increased bulk lowered
activity. In contrast, the (3-CF3, 4-Cl) compound was more
potent than its monosubstituted analogs.
When esters with different alcohol parts were compared,
methyl, ethyl, and isopropyl esters showed potency within
the same range, whereas the butyl esters tested showed
lower potency. One explanation could be that the bulkiness
of the butyl analogs caused the lower antifeedant activity.
The investigation of the relation of the various Hansch
parameters to antifeedant activity revealed a correlation to π
and R, respectively. For π, a negative value (π<−0.1) gave
a lower AFI. All eight compounds with a more negative F
had AFIa<60 at 50 mM concentration. This result was
expected as too highly hydrophilic compounds have not
been very active in previous investigations of antifeedants
for H. abietis (Unelius et al. 2006; Sunnerheim et al. 2007).
Among the 51 compounds tested in this study, a large
number of potent antifeedants were found. A majority of the
compounds tested was more effective than the lead compound,
ethyl cinnamate, which has been identified in the inner bark of
Pinus contorta (Bratt et al. 2001). The most potent
compounds were methyl 3-phenylpropanoates monosubsti-
tuted with chloro, fluoro, or methyl groups on the aromatic

J Chem Ecol (2008) 34:339–352
351
one and 2,6-dimethyl-2,7-octadien-4-ol. Acta Chem. Scand
B41:442–447.
ring, and the 3,4-dichlorinated methyl 3-phenylpropanoate.
These compounds are good candidates for further studies in
the laboratory and in the field. A pool of over 20 highly active
phenylpropanoid compounds provides many possibilities to
develop mixtures of compounds that could potentially be even
more effective than the single compounds tested so far. Field
assays that evaluate chemical and physical properties, such as
volatility and stability to resist water and UV radiation, are
required to establish the effectiveness of these compounds for
conifer seedling protection under field conditions.
BJÖRKLUND, N., NORDLANDER, G., and BYLUND, H. 2005. Olfactory
and visual stimuli used in orientation to conifer seedlings by the
pine weevil, Hylobius abietis. Physiol. Entomol 30:225–231.
BLANEY, W. M., SIMMONDS, M. S. J., EVANS, S. V., and FELLOWS, L.
E. 1984. The role of the secondary plant compound 2,5-
dihydroxymethyl 3,4-dihydroxypyrrolidine as a feeding inhibitor
for insects. Entomol. Exp. Appl 36:209–216.
BORG-KARLSON, A.-K., NORDLANDER, G., MUDALIGE, A., NORDEN-
HEM, H., and UNELIUS, C. R. 2006. Antifeedants in the feces of
the pine weevil Hylobius abietis: Identification and biological
activity. J. Chem. Ecol 32:943–957.
Our findings are probably a consequence of effects from
the stimuli of several taste and odor receptors processed by
the insect brain. Earlier investigations (Wibe et al. 1996,
1997, 1998) have described the response of numerous
receptors in the antennae of H. abietis, of which some are
highly specific and some more generally tuned to olfactory
stimuli, within groups of monoterpenes and aromatic
compounds. Thus, it is likely that several receptor types
are activated during feeding and that each type is tuned to
molecules with different structural properties. The results
from a study of structural analogs of naphthoquionones as
feedants/antifeedants for the larvae of the Mexican bean
beetle, Epilachna varivestis, are as intriguing as ours. A
complex interplay of electronic, steric, electrochemical, and
positional requirements was found that affected the feeding
response (Weissenberg et al. 1997).
BRATT, K., SUNNERHEIM, K., NORDENHEM, H., NORDLANDER, G., and
LÅNGSTRÖM, B. 2001. Pine weevil (Hylobius abietis) antifee-
dants from lodgepole pine (Pinus contorta). J. Chem. Ecol
27:2253–2262.
COWLES, R. S., MILLER, J. R., HOLLINGSWORTH, R. M., ABDEL-AAL,
M. T., SZURDOKI, F., BAUER, K., and MATOLCSY, G. 1990.
Cinnamyl derivatives and monoterpenoids as non-specific ovipo-
sitional deterrents for the onion fly. J. Chem. Ecol 16:2401–2428.
DAY, K. R., NORDLANDER, G., KENIS, M., and HALLDÓRSON, G.
2004. General biology and life cycles of bark weevils, Chapter
14, Bark and Wood Boring Insects in Living Trees in Europe, A
Synthesis. pp. 331–349, in F. Lieutier, K. R. Day, A. Battisti, J. C
Grégoire, and H. F. Evans (eds.). Kluwer, Dordrecht.
ERIKSSON, C. 2006, Isolation, synthesis and structure–activity rela-
tionships of antifeedants against the pine weevil, Hylobius
abietis. Ph.D. Dissertation, KTH, Stockholm, Sweden. ISBN:
91-7178-301-6.
HANSCH, C., and FUJITA, T. 1963. ρ–σ–π Analysis. A method for the
correlation of biological activity and chemical structure. J. Am.
Chem. Soc 86:1616–1626.
The ecological background to why feeding of H. abietis is
affected by derivatives of phenylpropanoates is still largely
unknown. Although it is possible that ethyl cinnamate is
partly responsible for the fact that H. abietis feeds less on
the bark of P. contorta than of P. sylvestris (Bratt et al.
2001), this is at least not specifically a substance used by P.
contorta for defense against this insect, as the tree and the
insect species have their natural distributions in different
parts of the world. However, phenylpropanoates may be of
more general importance in plant–insect interactions than
presently acknowledged, as indicated by the frequently
strong antifeedant responses in the pine weevil.
HANSCH, C., and DEUTSCH, E. W. 1966. The structure–activity
relationship in amides inhibiting photosynthesis. Biochim. Bio-
phys. Acta 112:381–391.
HANSCH, C., LEO, A., UNGER, S. H., KI, H. K., NIKAITANI, D., and
LIEN, E. J. 1973. “Aromatic” substituent constants for structure–
activity correlations. J. Med. Chem 16:1207–1216.
HARWOOD, L. M., MOODY, C. J., and PERCY, J. M. 1998. Experimental
Organic Chemistry. 2nd edn. Blackwell Science, Oxford.
HUANG, Y., and HO, S. H. 1998. Toxicity and antifeedant activities of
cinnamaldehyde against the grain storage insects, Tribolium
castaneum (Herbst) and Sitophilus zeamais Motsch. J. Stored
Prod. Res 34:11–17.
JACUBAS, W. J., SHAH, P. S., MASON, J. R., and NORMAN, D. M.
1992. Avian repellency of coniferyl and cinnamyl derivatives.
Ecol. Appl 2:147–156.
LÅNGSTRÖM, B., and DAY, K. R. 2004. Damage, control and
management of weevil pests, especially Hylobius abietis, Chapter
19, Bark and Wood Boring Insects in Living Trees in Europe, A
Synthesis. pp. 415–444, in F. Lieutier, K. R. Day, A. Battisti, J. C
Grégoire, and H. F. Evans (eds.). Kluwer, Dordrecht.
LEGRAND, S., NORDLANDER, G., NORDENHEM, H., BORG-KARLSON,
A.-K., and UNELIUS, C. R. 2004. Hydroxy-methoxybenzoic
methyl esters: synthesis and antifeedant activity on the pine
weevil, Hylobius abietis. Z. Naturforsch 59b:829–835.
SUNNERHEIM, K., NORDQVIST, A., NORDLANDER, G., BORG-KARLSON,
A.-K., UNELIUS, C. R., BOHMAN, B., NORDENHEM, H., HELLQV-
IST, C., and KARLÉN, A. 2007. Quantitative structure–activity
relationships of pine weevil antifeedants, a multivariate approach.
J. Agric. Food Chem 55:9365–9372.
Acknowledgments We thank The Swedish Research Council for
Environment, Agricultural Sciences and Spatial Planning, The
Swedish Hylobius Research Program, and The University of Kalmar
for supporting this study. Henning Henschel, Abtin Daghigi, Meral
Sari, Veronica Siljehav, Konstantin Dubrovinski, Tonja Freisolt, and
Magnus Besev were helpful in the syntheses of the test compounds
and Claes Hellqvist in processing the bioassay data.
References
SWAIN, C. G., and LUPTON, E. C. 1968. Field and resonance
components of substituent effects. J. Am. Chem. Soc 90:4328–4337.
TOPLISS, J. G. 1972. Utilization of operational schemes for analog
synthesis in drug design. J. Med. Chem 15:1006–1011.
BAECKSTRÖM, P., STRIDH, K., LI, L., and NORIN, T. 1987. Claisen
rearrangements with mesityl oxide dimethyl ketal. Synthesis of
ipsdienone, E- and Z-ocimenone, 2,6-dimethyl-2,7-octadien-4-

352
J Chem Ecol (2008) 34:339–352
TOPLISS, J. G. 1977. A manual method for applying the Hansch
approach to drug design. J. Med. Chem 20:463–469.
UNELIUS, C. R., NORDLANDER, G., NORDENHEM, H., HELLQVIST, C.,
LEGRAND, S., and BORG-KARLSON, A.-K. 2006. Structure–activity
relationships of benzoic acid derivatives as antifeedants for the pine
weevil, Hylobius abietis. J. Chem. Ecol 32:2191–2203.
Effect of substituent and ring changes in naturally occuring
naphthoquinones on the feeding response of larvae of the mexican
bean beetle, Epilachna varivestis. J. Chem. Ecol 23:3–18.
WIBE, A., BORG-KARLSON, A.-K., NORIN, T., and MUSTAPARTA, H.
1996. Identification of plant volatiles activating the same receptor
neurons in the pine weevil, Hylobius abietis. Entomol. Exp. Appl
80:39–42.
WIBE, A., BORG-KARLSON, A.-K., NORIN, T., and MUSTAPARTA, H.
1997. Head-space plant volatiles eliciting responses in receptor
neurons of the pine weevil (Hylobius abietis L.). J. Comp.
Physiol 180:585–595.
VOGEL, A. I. 1989. Textbook of Practical Organic Chemistry. 5th edn.
Prentice Hall, Upper Saddle River, NJ.
WATKINS, R. W., LUMLEY, J. A., GILL, E. L., BISHOP, J. D., LANGTON,
S. D., MACNICOLL, A. D., PRICE, N. R., and DREW, M. G. B.
1999. Quantitative structure–activity relationships (QSAR) of
cinnamic acid bird repellents. J. Chem. Ecol 25:2825–2845.
WEISSENBERG, M., MEISNER, J., KLEIN, M., SCHAEFFLER, I.,
ELIYAHU, M., SCHMUTTERER, H., and ASCHER, K. R. S. 1997.
WIBE, A., BORG-KARLSON, A.-K., PERSSON, M., NORIN, T., and
MUSTAPARTA, H. 1998. Enantioselectivity of receptor neurons in
the pine weevil (Hylobius abietis). J. Chem. Ecol 24:273–287.