Structure-activity relationship of chemical defenses from the freshwater plant Micranthemum umbrosum

Article abstract of DOI:10.1016/j.phytochem.2006.05.007

Vascular plants produce a variety of molecules of phenylpropanoid biosynthetic origin, including lignoids. Recent investigations indicated that in freshwater plants, some of these natural products function as chemical defenses against generalist consumers

Full text of DOI:10.1016/j.phytochem.2006.05.007

PHYTOCHEMISTRY
Phytochemistry 67 (2006) 1224–1231
www.elsevier.com/locate/phytochem
Structure–activity relationship of chemical defenses from the
freshwater plant Micranthemum umbrosum
a
a,b,
*
Amy L. Lane , Julia Kubanek
a
School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, GA 30332-0400, USA
b
School of Biology, Georgia Institute of Technology, Atlanta, GA 30332-0230, USA
Received 20 December 2005; received in revised form 14 April 2006
Abstract
Vascular plants produce a variety of molecules of phenylpropanoid biosynthetic origin, including lignoids. Recent investigations indi-
cated that in freshwater plants, some of these natural products function as chemical defenses against generalist consumers such as cray-
fish. Certain structural features are shared among several of these anti-herbivore compounds, including phenolic, methoxy,
methylenedioxy, and lactone functional groups. To test the relative importance of various functional groups in contributing to the feed-
ing deterrence of phenylpropanoid-based natural products, we compared the feeding behavior of crayfish offered artificial diets contain-
ing analogs of elemicin (1) and b-apopicropodophyllin (2), chemical defenses of the freshwater macrophyte Micranthemum umbrosum.
Both allyl and methoxy moieties of 1 contributed to feeding deterrence. Disruption of the lactone moiety of 2 reduced its deterrence.
Finally, feeding assays testing effects of 1 and 2 at multiple concentrations established that these two natural products interact additively
in deterring crayfish feeding.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Keywords: Micranthemum umbrosum; Scrophulariaceae; Antifeedant; Structure–activity relationship; Chemical defense; Plant–herbivore interaction;
Lignan; Elemicin; b-Apopicropodophyllin; Procambarus acutus
1. Introduction
occurring herbivores. Among freshwater macrophytes, 11
shikimate-derived natural products (10 lignoids and one
monomeric phenylpropanoid) and one potential shikimate
metabolite (a p-hydroxybenzyl ester) have been shown to
deter herbivory by generalist crayfish (Bolser et al., 1998;
Wilson et al., 1999; Kubanek et al., 2000, 2001; Parker
et al., 2006), whose feeding behavior can dramatically affect
macrophyte distribution and abundance (Lodge, 1991)
(Fig. 1a). In these studies, isolation of deterrent natural
products was guided by feeding assays such that structur-
ally-related but non-deterrent compounds were not identi-
fied. A comparison of structural features for deterrent and
non-deterrent metabolites of the same biosynthetic class
has therefore not been possible, preventing rigorous analysis
of the structural basis for chemical defense.
Phenylpropanoid-derived natural products, biosynthe-
sized via the shikimate pathway, are widespread among vas-
cular plants (Lewis and Davin, 1999) and have been shown to
exhibit antimitotic, antiviral, insect antifeedant, and root
growth inhibition properties (Loike and Horwitz, 1976; Ela-
kovich and Stevens, 1985; Gnabre et al., 1995; Harmatha
and Nawrot, 2002). Most biological studies related to phe-
nylpropanoid-based natural products have focused on their
potential application in medicine and agriculture. A smaller
number of ecological studies have indicated that some of
these metabolites function as chemical defenses against co-
*
Corresponding author. Address: School of Chemistry and Biochem-
It is expected that there exists a definable relationship
between the structure of phenylpropanoid-based plant nat-
ural products and feeding deterrence, when considering a
istry, Georgia Institute of Technology, Atlanta, GA 30332-0230, USA.
Tel.: +1 404 894 8424; fax: +1 404 385 4440.
E-mail address: julia.kubanek@biology.gatech.edu (J. Kubanek).
0031-9422/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2006.05.007

A.L. Lane, J. Kubanek / Phytochemistry 67 (2006) 1224–1231
1225
Fig. 1. (a) Freshwater plant shikimate-derived metabolites previously demonstrated to deter crayfish feeding. 1 Parker et al. (2006); 2 Kubanek et al. (2000,
2001); 3 Bolser et al. (1998), Wilson et al. (1999). (b) Analogs of b-apopicropodophyllin (2) for which crayfish feeding deterrence was assessed in the current
study.
population or species of herbivores. From assessment of a
group of deterrent lignoids possessing a common carbon
skeleton, Kubanek et al. (2000) suggested that increased
aryl hydroxylation may be associated with increased deter-
rence. By strategic manipulation of the molecular struc-
tures of natural products, quantification of biological
activities, and statistical analyses to test for differences,
insights may be gained into the precise nature of this struc-
ture–activity relationship. Such studies have commonly
been used to explore the pharmacological (e.g., Lee et al.,
1999; Zhang et al., 2004) and agricultural (e.g., Kim and
Mullin, 2003; Morimoto et al., 2003) activities of natural
products. However, few studies have investigated such rela-
tionships in an ecological context (but see Assmann et al.,
2000; Lindel et al., 2000; Silva and Trigo, 2002), and no
previous studies have evaluated this for phenylpropanoid-
derived molecules. Such studies could lead to testable
hypotheses regarding mechanisms of chemoreception and
other physiological responses. Structure–activity relation-
ship studies may also provide insights into the evolution
of chemical defenses: if herbivore pressure is intense and
chemical defenses carry significant costs, one might predict
that plants evolved pathways to produce the most deterrent
compounds, suggesting that natural products should have
greater deterrent potency than unnatural, but structur-
ally-related compounds, and that concentrations of chemi-
cal defenses within plants are likely to be adequate, but not
excessive, for deterring herbivores.
In the current study, we selected elemicin (1) and b-
apopicropodophyllin (2) (Fig. 1a), phenylpropanoid-based

1226
A.L. Lane, J. Kubanek / Phytochemistry 67 (2006) 1224–1231
chemical defenses of the freshwater plant Micranthemum
umbrosum (Parker et al., 2006), as model structures for
the evaluation of structure–activity relationships among
freshwater plant chemical defenses. We obtained eight ana-
logs of 1 and 2 by semi-synthesis or from commercial
sources, and from feeding assay data using the crayfish
Procambarus acutus we generated dose–response curves
to quantify deterrent potency and to predict structural
requirements for deterrence. Additionally, we compared
the deterrent potencies of 1 and 2, and evaluated the inter-
action (additive, antagonistic, or synergistic) between 1 and
2 in the chemical defense of the freshwater plant M.
umbrosum.
order to minimize costs associated with chemical defense.
The experimental EC50 for elemicin (1), 8.3 mM (Fig. 2),
was found to be similar to its natural concentration of
3.2 mM (Parker et al., 2006). In contrast, the natural con-
centration of b-apopicropodophyllin (2), 0.96 mM (Parker
et al., 2006), was nearly 100 times greater than its experi-
mental EC50 value, 0.011 mM (Fig. 2). These data suggest
that 2 is more important than 1 in deterring crayfish from
feeding on M. umbrosum, and may indicate that both
metabolites serve multiple ecological functions, as natural
concentrations of the combined compounds are greater
than that required for feeding deterrence, or that metabo-
lite concentrations are not optimally tuned to herbivore
sensitivity.
2. Results and discussion
2.2. Interaction between elemicin (1) and b-
apopicropodophyllin (2) in the chemical defense of
Micranthemum umbrosum
2.1. Feeding deterrence of natural products elemicin (1) vs.
b-apopicropodophyllin (2)
Previous studies have indicated that some phenylpropa-
noids and lignoids interact synergistically in the deterrence
of agricultural pests (Yamashita and Matsui, 1961). Alter-
natively, antagonistic or additive effects could occur. When
1 and 2 were investigated for combined effects in deterring
crayfish feeding, the experimental logEC50 value for the
combined compounds was not significantly different from
the value obtained from a theoretical additive curve (F test,
p = 0.73; Fig. 2), leading us to reject the hypothesis that
these two natural products behave synergistically or antag-
onistically in deterring crayfish feeding.
Comparison of logEC50 values for the feeding deter-
rence of the phenylpropanoid elemicin (1) and the lignoid
b-apopicropodophyllin (2) against the crayfish Procamba-
rus acutus indicated that 2 is approximately 750 times more
deterrent than 1 (p < 0.0001; Fig. 2). Consistent with this
finding, in an investigation of the insect feeding deterrence
activity of compounds belonging to these two structural
groups, Harmatha and Nawrot (2002) reported that lig-
noids were generally more bioactive than phenylpropanoid
monomers.
According to the optimal defense theory (Rhoades and
Cates, 1976), if natural products evolved to fulfill a specific
ecological function such as chemical defense, one would
expect natural concentrations of these defenses to approx-
imately match the sensitivity of potential consumers, in
2.3. Structure–activity relationship of elemicin (1) analogs
At 12· the natural molar concentration of elemicin (1),
allylbenzene (3) was not significantly deterrent to crayfish
(Fisher’s exact test, p > 0.99; Fig. 3), indicating that substit-
uents on the allylbenzene scaffold are necessary for feeding
deterrence. In contrast, dimethoxy-substituted methyl
eugenol (4) was the most effective deterrent of compounds
tested within this group, with an EC50 value 87% less than
that of 1 (F test, p < 0.001; Fig. 3). The greater potency of 4
relative to trimethoxy-substituted 1 indicates that the third
methoxy substituent on the phenyl ring of 1 undermines
deterrence, which may indicate that steric hindrance of
the bulkier 1 alters the interaction with a P. acutus chemo-
receptor. Although crustacean receptors for detecting plant
chemical defenses have not yet been identified, it seems
likely that taste or odorant chemoreceptors are involved,
given the quick response time (seconds) and lack of appar-
ent injury to crayfish rejecting chemically-defended foods
which might be expected if a deterrent caused a non-recep-
tor-mediated effect such as burning (Lane and Kubanek,
pers. observ.). The enhanced activity (relative to 1) of dime-
thoxy-substituted 4, a metabolite previously isolated from
other vascular plants (e.g., Mata et al., 2004), also indicates
that M. umbrosum has not evolved to produce the metabo-
lite most deterrent to this particular herbivore.
Fig. 2. Effect of compound concentration on crayfish feeding behavior for
M. umbrosum natural products elemicin (1) and b-apopicropodophyllin
(2), and effect of combined doses of 1 and 2; n = 13–24 crayfish for each
data point. Grey arrows denote natural concentrations of 1 and 2 in M.
umbrosum from Parker et al. (2006). In determining the effect of combined
doses, 1 and 2 were added to artificial diets in 1:1 molar ratios of the EC50
values for each compound. The theoretical additive curve was calculated
from best fit dose–response curves developed individually for 1 and 2.

A.L. Lane, J. Kubanek / Phytochemistry 67 (2006) 1224–1231
1227
Fig. 3. Comparison of crayfish feeding deterrence of elemicin (1) and analogs. Different letters indicate treatments differing significantly from each other
(F test, p 6 0.05); bars represent standard error. Replacement of aryl methoxy groups with hydroxy groups resulted in decreased potency, as seen in 4 vs. 5
and 1 vs. 6. Increased aryl substitution was also associated with decreased activity (4 vs. 1; 5 vs. 6), except that 3, with only an allyl substituent, was
completely inactive. Dose–response curves were used to calculate EC50 values for 1 (Fig. 2) and 4–6 (data not shown). For 3 and 7, EC50 values could not
be calculated because these compounds were not deterrent at any concentration tested (see text). For 8, an EC50 value could not be calculated because low
synthetic yield prohibited testing at sufficient concentrations (see text).
In addition to the number of aryl methoxy groups affect-
ing deterrence, the identity of substituents also appeared to
play an important role. The presence of a hydroxy group
para to the allyl substituent as in eugenol (5) and methoxy-
eugenol (6), instead of a methoxy group as in elemicin (1)
and methyl eugenol (4), was associated with substantially
reduced deterrent potency for trisubstituted benzenes 5
vs. 4 (F test, p < 0.001; Fig. 3), and marginally reduced
deterrent potency for tetrasubstituted benzenes 6 vs. 1
(p = 0.17; Fig. 3). Both hydroxy and methoxy substituents
may function as hydrogen bond acceptors in interactions
with crayfish chemoreceptors, whereas a hydroxy substitu-
ent can also act as a hydrogen bond donor, which could
potentially affect deterrent potency by altering the orienta-
tion in which a ligand interacts with a chemoreceptor.
Alternatively, the bulkier, more hydrophobic methoxy sub-
stituent in the para position of 4 may enhance deterrence
(relative to 5) via a better steric fit and/or stronger van
der Waals attractive forces within the chemoreceptor bind-
ing site. The current finding, that aryl hydroxy groups are
associated with weaker crayfish deterrence than are aryl
methoxy groups among monomeric phenylpropanoids, is
contrary to the suggestion of Kubanek et al. (2000) regard-
ing the crayfish deterrence among a group of lignans.
Eugenol (5) was previously demonstrated to cause paral-
ysis in crayfish placed in an aqueous solution of this com-
pound (Ozeki, 1975). Given the structural similarity of
elemicin (1) and 5, it is possible that 1 may also be toxic
at certain concentrations; however, we did not observe
incapacitation or mortality of P. acutus during or after
feeding assays. This may indicate that toxicity is dimin-
ished when consumed as part of a diet rather than
absorbed from surrounding water. If 1 or 5 is indeed toxic
at high doses or following repeated consumption of plants
containing these metabolites, this may have favored evolu-
tion of chemoreception in consumers such as P. acutus, in
order to identify and avoid consuming toxic foods.
Supporting the hypothesis that the allyl group at C-5 of
elemicin (1) is important in feeding deterrence, 1,2,3-tri-
methoxybenzene (7) was palatable at 38 mM, 12· the nat-
ural molar concentration of 1 (Fisher’s exact test p = 0.23).
3-(30,40,50-Trimethoxyphenyl)-1,2-propanediol (8), synthe-
sized from 1, was significantly deterrent at the three con-
centrations at which it was tested, spanning 2.9–8.4 mM
(Fisher’s exact test, p = 0.0006–0.05). Due to limited yield
of synthetic product, a full dose–response curve could not
be established for 8. However, comparison of the deter-
rence of 1 and 8 at 3.8 mM indicated that potency did
not appear to differ significantly between these two com-
pounds (Fisher’s exact test, p > 0.99). Thus, conversion of
the allyl substituent to a diol did not appear to reduce bio-
activity, despite the enhanced bulkiness, polarity, and
hydrogen bonding capacity of this group relative to the
allyl substituent of 1. Although the lack of deterrence of tri-
methoxybenzene (7) indicated that a non-hydrogen substi-
tuent was essential at C-5 for deterrence, it appears that the
P. acutus chemoreceptor for 1 is not highly specific for the
C-5 substituent.
2.4. Structure–activity relationship of b-apopicropodophyllin
(2) analogs
A number of structural features may be expected to
influence the bioactivity of b-apopicropodophyllin (2),
including the unsaturated lactone, trimethoxyphenyl
group, methylenedioxy moiety, and stereochemistry at C-
1. We focused on the role of the lactone moiety in crayfish
feeding behavior, as this group was most amenable to syn-
thetic modification and has been demonstrated to affect
bioactivity in pharmacological studies of analogous lig-
noids including podophyllotoxin (9) (Loike and Horwitz,
1976; Brewer et al., 1979).
Podophyllotoxin (9) was marginally less deterrent than
b-apopicropodophyllin (2) (F test, p = 0.07; Table 1),

1228
A.L. Lane, J. Kubanek / Phytochemistry 67 (2006) 1224–1231
Table 1
the use of dose–response curves and represents the first
investigation of the structure–activity relationship of plant
feeding deterrents against a freshwater herbivore.
Comparison of crayfish feeding deterrence for b-apopicropodophyllin (2)
and analogs
Analog
log[EC50 (mM)] SE
EC50 (mM)
Comparison of dose–response curves for two M. umbro-
sum natural products, elemicin (1) and b-apopicropodo-
phyllin (2), and eight analogs (3–10) of these natural
products indicated that the allyl and methoxy moieties of
1 influenced crayfish feeding behavior, as did the lactone
moiety of 2. The 12-member collection of natural products
previously reported to deter crayfish feeding includes some
molecules with none of these functionalities (Wilson et al.,
1999; Kubanek et al., 2000), suggesting that several differ-
ent chemoreceptive mechanisms are likely involved in cray-
fish feeding deterrence and/or that crayfish species or
populations differ in their responses. Studies of interactions
between these molecules and individual herbivore chemore-
ceptors and manipulation of chemoreception physiology
will be necessary to develop further understanding regard-
ing receptor–ligand interactions involving chemical
defense.
b-Apopicropodophyllin (2)
Podophyllotoxin (9)
b-Apopicropodophyllol (10)
ꢀ1.95 0.08a
0.011
0.020
ꢀ1.70 0.07a
(Non-deterrent at all
concentrations tested)
Dose–response curves were used to calculate EC50 values for 2 (Fig. 2) and
9 (data not shown).
a
logEC50 values for 2 and 9 differ marginally (F test, p = 0.07).
which may have resulted from different lactone conforma-
tions, from the lack of the C-2–C-3 unsaturation in 9, or
from the presence of a hydroxy at C-4 in 9. Whereas 2
was significantly deterrent at 1% of its natural molar con-
centration (Fisher’s exact test, p = 0.001) and nearly
100% deterrent at its natural concentration (p < 0.0001;
Fig. 2), disruption of the lactone by reduction to the diol
b-apopicropodophyllol (10) resulted in a complete lack of
deterrence for 10 at either of these concentrations
(p > 0.99 for both concentrations); 10 was significantly less
potent than 2 (Fisher’s exact test, p < 0.001 for comparison
of deterrence of 2 and 10 at both concentrations). Feeding
assays at higher concentrations were not feasible for 10,
due to insufficient synthetic product yields. This conversion
of the lactone ring in 2 to a diol moiety in 10, while retain-
ing the C2–C3 unsaturation, may have resulted in loss of
activity by a diminished capacity of the more polar diol
to bind with hydrophobic regions of the chemoreceptor,
or by increased conformational flexibility which could dis-
rupt orientation of hydrogen bonding or dipole–dipole
interacting groups between a receptor and ligand.
The role of the methylenedioxy moiety of b-apopicrop-
odophyllin (2) in crayfish feeding deterrence remains
unclear, as attempts at selective synthetic modification of
this group were unsuccessful (data not shown). Kubanek
et al. (2000) reported a series of seven antifeedant lignoids
that did not possess methylenedioxy groups, indicating that
these groups are not essential for crayfish feeding deter-
rence among lignoids (although different, albeit congeneric,
species of crayfish were used in Kubanek et al. (2000) vs.
the current study). In contrast, methylenedioxy moieties
have been suggested to be important in the effectiveness
of insect feeding deterrents (Harmatha and Nawrot, 2002).
3. Experimental
3.1. General experimental procedures
Fisher Scientific ACS grade solvents were used for
extractions, flash CC, and chemical transformations;
Fisher Scientific HPLC grade solvents were used for
HPLC. Compounds 3–7, 9, and chemicals for synthetic
modifications were purchased from Sigma–Aldrich (St.
Louis, MO, USA). HPLC analyses were conducted with
Zorbax RX-SIL (Agilent Technologies) semi-preparative
normal phase columns using a Waters HPLC system
(Waters 515 pump; Waters 2487 dual wavelength absor-
bance detector) with UV absorbance monitored at 220
and 254 nm. 1H, 13C, and two-dimensional inverse-detected
NMR spectroscopic data (COSY, HMQC, HMBC) were
acquired with a Bruker DRX-500 MHz spectrometer. All
NMR experiments were conducted using CDCl3 and refer-
1
enced to residual CHCl3 (7.24 and 77.0 ppm, for H and
13C, respectively). High and low resolution electron impact
(EI) mass spectra were collected with a VG Instruments
70SE spectrometer.
3.2. Isolation of plant chemical defenses (1–2)
2.5. Conclusions
Whole M. umbrosum plants were collected in June, 2003,
from ponds at the Owens and Williams Fish Hatchery
(Hawkinsville, GA, USA) and were frozen until extraction.
A voucher specimen GIT-MICR-001 is stored at the Geor-
gia Institute of Technology. Plant material was shredded in
MeOH and extracted successively with MeOH (2·),
Me2CO (2·), and CH2Cl2 (2·). Extracts were combined, fil-
tered, and concentrated in vacuo. This crude extract was
partitioned between petroleum ether and MeOH/H2O
(9:1); the MeOH/H2O (9:1) portion was further partitioned
Previous investigations aimed at elucidating the molecu-
lar structural requirements for ecological function have
made assessments on the basis of one or a few feeding assay
data points (e.g., Assmann et al., 2000; Kubanek et al.,
2000; Lindel et al., 2000), making it difficult to quantita-
tively compare the sensitivity of consumers to chemical
cues in their food. To our knowledge, this study marks
the first aquatic chemical ecology investigation in which
structure–activity relationships were evaluated through

A.L. Lane, J. Kubanek / Phytochemistry 67 (2006) 1224–1231
1229
between MeOH/H2O (3:2) and CHCl3. TLC Rf values of
authentic samples of 1 and 2 were compared with values
for compounds in liquid–liquid partition fractions, in order
to identify fractions containing these natural products.
b-Apopicropodophyllin (2) was found exclusively in the
CHCl3 fraction, and so this fraction was further separated
by flash CC on silica with gradient elution of hexanes/
EtOAc (9:1) to EtOAc. Finally, 2 (23 mg) was purified by
normal phase silica HPLC with CH2Cl2/Me2CO (49:1) as
the mobile phase. Elemicin (1) was found in both the
CHCl3 and hexanes fractions following liquid–liquid parti-
tioning. Both fractions were separated by flash CC as
described above for 2, except that the hexane-soluble por-
tion was subjected to a gradient mobile phase of hexanes
to EtOAc. Finally, 1 (48 mg) was purified by normal phase
silica HPLC using a hexanes/EtOAc (49:1) mobile phase.
The structures of 1 and 2 were determined by spectroscopic
analysis and verified by comparison with lit. values (Gens-
ler et al., 1971; Achenbach and Frey, 1992). Elemicin (1):
1H NMR (500 MHz, CDCl3): d 6.41 (2H, s, H-4 and H-
6), 5.95 (1H, m, H-8), 5.12 (1H, m, H-9), 5.08 (1H, m, H-
9), 3.85 (6H, s, OMe-1 and -3), 3.82 (3H, s, OMe-2), 3.32
(2H, d, J = 6.5 Hz, H-7). 13C NMR (125 MHz, CDCl3): d
153.1 (C-1 and C-3), 137.2 (C-8), 136.2 (C-5), 135.8 (C-
2), 116.0 (C-9), 105.3 (C-4 and C-6), 60.8 (OCH3-2), 56.1
(OCH3-1,3), 40.5 (C-7). EI (m/z): [M+] calculated for
C12H16O3, 208.10994; found 208.10856. b-Apopicropodo-
H-20 and H-60), 3.94 (1H, m, H-2), 3.84 (6H, s, OMe-30
and -50), 3.81 (3H, s, OMe-40), 3.70 (1H, dd, J = 11.5,
2.5 Hz, H-1), 3.49 (1H, dd, J = 11.5, 6.5 Hz, H-1), 2.74
(1H, dd, J = 13.5, 4.3 Hz, H-3); 2.67 (1H, dd, J = 13.5,
13
8.7 Hz, H-3). C-NMR (125 MHz, CDCl3): d 53.6 (C-30
and C-50), 137.0 (C-10), 133.6 (C-40), 106.4 (C-20 and C-
60), 73.2 (C-2), 66.4 (C-1), 61.1 (OMe-40), 56.4 (OMe-30
and -50), 40.4 (C-3). EI (m/z): [M+] calculated for
C12H18O5, 242.11542; found 242.11564.
3.4. Reduction of b-apopicropodophyllin (2) to b-
apopicropodophyllol (10)
A solution of 2 (0.010 mmol) in CH2Cl2 (100 ll) was
added to dry Et2O (5 ml) suspension of LiAlH4
(0.025 mmol) and stirred 15 h at room temperature (Anj-
anamurthy and Rai, 1987). Aqueous HCl (2 M, 15 ml)
were then added, and the mixture was stirred for 30 min.
The Et2O layer was washed with deionized H2O
(2 · 15 ml) and dried over dry Na2SO4. Et2O was removed
in vacuo, and 10 was purified as a white powder by normal
phase silica HPLC with CH2Cl2/EtOAc (7:3) as mobile
phase. Compound 10 (0.0047 mmol, 47% yield) was identi-
fied by spectroscopic analysis and comparison with pub-
lished data (Anjanamurthy and Rai, 1987). 1H-NMR
(500 MHz, CDCl3): d 6.63 (1H, s, H-5), 6.51 (1H, s, H-
8), 6.38 (2H, s, H-20 and H-60), 5.88 (1H, d, J = 1.5 Hz,
OCH2O), 5.84 (1H, d, J = 1.5 Hz, OCH2O), 4.53 (1H, m,
H-1), 4.33 (2H, m, H-10), 4.26 (1H, dd, J = 12.0, 2.5 Hz,
H-9), 4.10 (1H, dd, J = 12.0, 7.8 Hz, H-9), 3.78 (6H, s,
OMe-30 and -50), 3.76 (3H, s, OMe-40), 3.72 (1H, dd,
J = 21.5, 4.2 Hz, H-4), 3.53 (1H, dd, J = 21.5, 3.6 Hz, H-
1
phyllin (2): H-NMR (500 MHz, CDCl3): d 6.70 (1H, s,
H-5), 6.62 (1H, s, H-8), 6.39 (2H, s, H-20 and H-60), 5.93
(1H, d, J = 4.5 Hz, OCH2O), 5.90 (1H, d, J = 4.5 Hz,
OCH2O), 4.87 (1H, m, H-10,), 4.82 (1H, m, H-1), 4.80
(1H, m, H-10), 3.82 (1H, dd, J = 2.5, 27.9 Hz, H-4), 3.79
(3H, s, OMe-40), 3.78 (6H, s, OMe-30,50), 3.64 (1H, dd,
J = 3.5, 27.9 Hz, H-4). 13C NMR (125 MHz, CDCl3): d
172.2 (C-9), 157.3 (C-3), 153.2 (C-30 and C-50), 147.2 (C-
40), 147.0 (C-6), 138.6 (C-7), 136.9 (C-10), 129.6 (C-5a),
128.1 (C-2), 123.7 (C-8a), 109.5 (C-8), 107.7 (C-5), 105.5
(C-20,60), 101.3 (OCH2O), 71.0 (C-10), 60.7 (OCH3-40),
56.1 (OCH3-30,50), 42.7 (C-1), 29.2 (C-4). EI (m/z): [M+]
calculated for C22H20O7, 396.12090; found 396.12071.
13
4). C-NMR (125 MHz, CDCl3): d 53.2 (C-30 and C-50),
146.1–146.0 (C-40 and C-6), 140.5 (C-7), 136.4 (C-10),
135.3 (C-2), 134.1 (C-3), 130.2 (C-5a), 125.4 (C-8a), 107.8
(C-8), 107.1 (C-5), 104.7 (C-20 and C-60), 100.5 (OCH2O),
62.4 (C-10), 61.0 (C-9), 60.6 (OMe-40), 55.8 (OMe-30 and
-50), 50.0 (C-1), 33.6 (C-4). EI (m/z): [M+] calculated for
C22H24O7, 400.15220; found 400.15173.
3.5. Bioassays and statistical methods
3.3. Oxidation of elemicin (1) to 3-(30,40,50-
trimethoxyphenyl)-1,2-propanediol (8)
Feeding assays were conducted using the omnivorous
crayfish, P. acutus, collected from the wetlands in the Chat-
tahoochee National Recreation Area (Atlanta, GA, USA)
in 2003 and 2004, and identified by comparison of morpho-
logical traits (Hobbs, 1981). Crayfish were housed individ-
ually in 12.5 · 12.5 cm chambers of a recirculating water
table at 23 ꢁC and maintained on a diet of commercial
trout food pellets.
Assays were completed following procedures described
in Parker et al. (2006). Artificial food for each feeding assay
was prepared by suspending 100 mg of a 1:1 mixture of
freeze-dried, ground broccoli and lettuce in Me2CO, to
which was added the test compound (natural product or
synthetic analog) dissolved in Me2CO. The mixture was
shaken and the Me2CO removed by rotary evaporation.
A solution of KMnO4 (0.050 mmol) in deionized H2O
(780 ll) was added to an ice-bath cooled solution of 1
(0.070 mmol) in EtOH/H2O (2:1) and stirred for 3 min.
The reaction mixture was filtered and then partitioned into
Et2O and aqueous layers; TLC indicated presence of the
diol product in the aqueous layer. The aqueous layer was
extracted with n-BuOH (2 · 15 ml). The n-BuOH extract
was concentrated in vacuo and 8 was purified by normal
phase silica HPLC with a mobile phase of hexanes/EtOAc
(3:7). The structure of 8 (0.011 mmol; 20% yield) was deter-
mined by spectroscopic analysis and verified by compari-
son with previous data (Dong et al., 1989; Gonzalez
1
et al., 1991). H-NMR (500 MHz, CDCl3): d 6.43 (2H, s,

1230
A.L. Lane, J. Kubanek / Phytochemistry 67 (2006) 1224–1231
This test food powder was then mixed with sodium alginate
(30 mg) in 1 ml of deionized H2O, and dispensed through a
syringe into a 0.10 M aqueous solution of CaCl2. Test food
was allowed to solidify in this solution for 1 min, then
rinsed with deionized H2O, and cut into test food pellets
ca. 3 mm in length. Test compound concentrations were
recorded as millimoles of compound per ml of food mix-
ture. Control food pellets were prepared in the same way,
including the use of Me2CO as solvent, but without the
addition of test compounds. Feeding assays were con-
ducted by first feeding a crayfish a control food pellet to
confirm that the crayfish was not already satiated, and if
that control pellet was consumed, then offering the crayfish
a test food pellet. If the crayfish consumed the test food
pellet, it was considered accepted. Test food pellets were
considered rejected if a crayfish took the pellet into its
mouth cavity twice and rejected it each time, in which case,
a second control pellet was offered to verify the crayfish did
not reject the test food pellet due to satiation. Feeding
deterrence was recorded as the frequency of 14–23 crayfish
rejecting a test food pellet, but accepting both control pel-
lets; crayfish that refused control pellets were not included
in the analysis. A Fisher’s exact test was applied to test for
significance of feeding deterrence data for each compound
at individual concentrations. A Fisher’s exact test was also
used to compare deterrence of two different compounds at
equal molar concentrations (Zar, 1998).
Dose–response curves were constructed using data from
5–11 feeding assays for each test compound, by plotting the
frequency of crayfish rejecting a test food pellet against the
log of the concentration of the test compound in food pel-
lets. These data were fit to a sigmoidal dose–response curve
with a Hill slope of 1; this Hill slope provided the largest R2
goodness-of-fit value for all data sets. Differences among
dose–response curves for different test compounds were
analyzed by an F test of the logEC50 values for each com-
pound using GraphPad Prism version 4 (Motulsky, 1995).
For compounds which were not significantly deterrent at
any tested concentration, the logEC50 could not be calcu-
lated. The highest concentration tested was 38 mM (12·
the natural concentration of 1) for analogs of 1, and
1.3 mM (1.4· the natural concentration of 2) for analogs
of 2. One compound, 3-(30,40,50-trimethoxyphenyl)-1,2-
propanediol (8) was synthesized in limited yield and so
was tested only up to the EC50 of 1 (8.3 mM).
Acknowledgements
This work was supported by an NSF-IGERT predoc-
toral fellowship to A.L.L. and by a Blanchard Assistant
Professorship from Georgia Tech to J.K. The authors
thank M. Hicks, A. Prusak, and E. Prince for assistance
with crayfish collections; J. Parker, D.O. Collins, and
M.E. Hay for sharing unpublished data and samples of 1
and 2 for identification purposes; J. Parker for macrophyte
collection; L. Gelbaum for assistance in collection of NMR
spectroscopic data; D. Bostwick and C. Sullards for mass
spectral analyses; and S. Padove for comments helpful in
improving this manuscript.
References
Achenbach, H., Frey, D., 1992. Cycloartanes and other terpenoids and
phenylpropanoids from Monocylanthus vignei. Phytochemistry 31,
4263–4274.
Anjanamurthy, C., Rai, K.M.L., 1987. Synthesis of podophyllotoxin and
related analogs. Part 3. Synthesis of b-apopicropodophyllin analogues
with modified lactone ring. Indian J. Chem. 26B, 131–135.
Assmann, M., Lichte, E., Pawlik, J.R., Kock, M., 2000. Chemical defenses
of the Caribbean sponges Agelas wiedenmayeri and Agelas conifera.
Mar. Ecol. Prog. Ser. 207, 255–262.
Bolser, R.C., Hay, M.E., Lindquist, N., Fenical, W., Wilson, D., 1998.
Chemical defenses of freshwater macrophytes against crayfish herbiv-
ory. J. Chem. Ecol. 24, 1639–1658.
Brewer, C.F., Loike, J.D., Horwitz, S.B., Sternlicht, H., Gensler, W.J.,
1979. Conformational analysis of podophyllotoxin and its congeners.
Structure–activity relationship in microtubule assembly. J. Med.
Chem. 22, 215–221.
Dong, X., Mondranondra, I.O., Che, C.T., Fong, H.H.S., Farnsworth,
N.R., 1989. Kmeriol and other aromatic constituents of Kmeria
duperreana. Pharm. Res. 6, 637–640.
Elakovich, S.D., Stevens, K.L., 1985. Phytotoxic properties of nord-
ihydroguaiaretic acid, a lignan from Larrea tridentata (Creosote bush).
J. Chem. Ecol. 11, 27–33.
Gensler, W.J., Ahmed, Q.A., Muljiani, Z., Gatsonis, C.D., 1971.
Ultraviolet irradiation of a-apopicropodophyllin. J. Am. Chem. Soc.
93, 2515–2522.
Gnabre, J.N., Brady, J.N., Clanton, D.J., Ito, Y., Dittmer, J., Bates, R.B.,
Huang, R.C.C., 1995. Inhibition of human immunodeficiency virus
type 1 transcription and replication by DNA sequence-selective plant
lignans. Proc. Natl. Acad. Sci. USA 92, 11239–11243.
Gonzalez, A.G., Barrera, J.B., Arancibia, L., Diaz, J.G., De Paz, P.P.,
1991. Two phenylpropanoids from Todaroa aurea subsp. suaveolens.
Phytochemistry 30, 4189–4190.
Harmatha, J., Nawrot, J., 2002. Insect feeding deterrent activity of lignans
and related phenylpropanoids with a methylenedioxyphenyl (pipero-
nyl) structure moiety. Entomol. Exp. Appl. 104, 51–60.
Hobbs, H.H., 1981. The Crayfishes of Georgia. Smithsonian Institution
Press, Washington, DC.
Kim, J.H., Mullin, C.A., 2003. Antifeedant effects of proteinase inhibitors
on feeding behaviors of adult western corn rootworm (Diabrotica
virgifera virgifera). J. Chem. Ecol. 29, 795–810.
Kubanek, J., Fenical, W., Hay, M.E., Brown, P.J., Lindquist, N., 2000.
Two antifeedant lignans from the freshwater macrophyte Saururus
cernuus. Phytochemistry 54, 281–287.
The interaction of natural products 1 and 2 in affecting
crayfish feeding behavior was assessed using feeding assays
incorporating these compounds at nine different concentra-
tions representing 1:1 ratios of experimentally determined
EC50 values for the two compounds (Luszczki and Czucz-
war, 2003). The resulting feeding response curve was com-
pared to a theoretical additive curve, developed on the
basis of best fit dose response curves developed individually
for 1 and 2 (Tallarida et al., 1997). An F test was applied to
test for significant difference between logEC50 values asso-
ciated with the theoretical additive curve vs. the observed
plot (using GraphPad Prism version 4).
Kubanek, J., Hay, M.E., Brown, P.J., Lindquist, N., Fenical, W., 2001.
Lignoid chemical defenses in the freshwater macrophyte Saururus
cernuus. Chemoecology 11, 1–8.
Lee, I.S., Jung, K.Y., Oh, S.R., Park, S.H., Ahn, K.S., Lee, H.K.,
1999. Structure–activity relationships of lignans from Schisandra

A.L. Lane, J. Kubanek / Phytochemistry 67 (2006) 1224–1231
1231
chinensis as platelet activating factor antagonists. Biol. Pharm. Bull.
22, 265–267.
Ozeki, M., 1975. Effects of eugenol on nerve and muscle in crayfish.
Comp. Biochem. Physiol. 50C, 183–191.
Lewis, N.G., Davin, L.B., 1999. Lignans: biosynthesis and function. In:
Barton, D.H.R., Sir, Nakanishi, K., Meth-Cohn, O. (Eds.), Compre-
hensive Natural Products Chemistry, vol. 1. Elsevier, Amsterdam, pp.
639–712.
Parker, J.D., Collins, D.O., Kubanek, J., Sullards, M.C., Bostwick,
D., Hay, M.E., 2006. Chemical defenses promote persistence of
the aquatic plant Micranthemum umbrosum. J. Chem. Ecol. 32,
815–833.
Lindel, T., Hoffmann, H., Hochgurtel, M., Pawlik, J.R., 2000. Structure–
activity relationship of inhibition of fish feeding by sponge-derived and
synthetic pyrrole-imidazole alkaloids. J. Chem. Ecol. 26, 1477–1496.
Lodge, D.M., 1991. Herbivory on fresh water macrophytes. Aquatic Bot.
41, 195–224.
Loike, J.D., Horwitz, S.B., 1976. Effects of podophyllotoxin and VP-16-
213 on microtubule assembly in vitro and nucleoside transport in
HeLa-cells. Biochemistry 15, 5435–5443.
Luszczki, J.J., Czuczwar, S.J., 2003. Isobolographic and subthreshold
methods in the detection of interactions between oxcarbazepine and
conventional antiepileptics – a comparative study. Epilepsy Res. 56,
27–42.
Mata, R., Morales, I., Perez, O., Rivero-Cruz, I., Acevedo, L., Enriquez-
Mendoza, I., Bye, R., Franzblau, S., Timmermann, B., 2004.
Antimycobacterial compounds from Piper sanctum. J. Nat. Prod. 67,
1961–1968.
Rhoades, D.F., Cates, R.G., 1976. Toward a general theory of plant
antiherbivore chemistry. In: Wallace, J.W., Nansel, R.L. (Eds.),
Recent Advances in Phytochemistry, vol. 10. Plenum Press, NY, pp.
168–213.
Silva, K.L., Trigo, J.R., 2002. Structure–activity relationships of pyrro-
lizidine alkaloids in insect chemical defense against the orb-weaving
spider Nephila clavipes. J. Chem. Ecol. 28, 657–668.
Tallarida, R.J., Kimmel, H.L., Holtzman, S.G., 1997. Theory and
statistics of detecting synergism between two active drugs: cocaine
and buprenorphine. Psychopharmacology 133, 378–382.
Wilson, D.M., Fenical, W., Hay, M., Lindquist, N., Bolser, R., 1999.
Habenariol, a freshwater feeding deterrent from the aquatic orchid
Habenaria repens (Orchidaceae). Phytochemistry 50, 1333–1336.
Yamashita, K., Matsui, M., 1961. Studies on phenolic lactones. Syner-
gistic activities of phenolic lactones. Agric. Biol. Chem. 25, 141–143.
Zar, J.H., 1998. Biostatistical Analysis. Prentice Hall, Englewood Cliffs,
NJ.
Morimoto, M., Tanimoto, K., Nakano, S., Ozaki, T., Nakano, A.,
Komai, K., 2003. Insect antifeedant activity of flavones and chro-
mones against Spodoptera litura. J. Agric. Food Chem. 51, 389–393.
Motulsky, H., 1995. Intuitive Biostatistics. Oxford University Press,
Oxford.
Zhang, H.Z., Kasibhatla, S., Wang, Y., Herich, J., Guastella, J., Tseng,
B., Drewe, J., Cai, S.X., 2004. Discovery, characterization, and SAR of
gambogic acid as a potent apoptosis inducer by a HTS assay. Bioorg.
Med. Chem. 12, 309–317.