Non-Target Effects of Transgenic Blight-Resistant American
Chestnut (Fagales: Fagaceae) on Insect Herbivores
Author(s) :K. H. Post and D. Parry
Source: Environmental Entomology, 40(4):955-963. 2011.
Published By: Entomological Society of America
DOI: http://dx.doi.org/10.1603/EN10063
URL: http://www.bioone.org/doi/full/10.1603/EN10063
BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the
biological, ecological, and environmental sciences. BioOne provides a sustainable online
platform for over 170 journals and books published by nonprofit societies, associations,
museums, institutions, and presses.
Your use of this PDF, the BioOne Web site, and all posted and associated content indicates
your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.
Usage of BioOne content is strictly limited to personal, educational, and non-commercial
use. Commercial inquiries or rights and permissions requests should be directed to the
individual publisher as copyright holder.
BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers,
academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research.
TRANSGENIC PLANTS & INSECTS
Non-Target Effects of Transgenic Blight-Resistant American Chestnut
(Fagales: Fagaceae) on Insect Herbivores
K. H. POST1,2,3
AND
D. PARRY1
Environ. Entomol. 40(4): 955Ð963 (2011); DOI: 10.1603/EN10063
ABSTRACT American chestnut [Castanea dentata (Marshall) Borkhausen], a canopy dominant
species across wide swaths of eastern North America, was reduced to an understory shrub after
introduction of the blight fungus [Cryphonectria parasitica (Murrill) Barr] in the early 1900s. Restoration of American chestnut by using biotechnology is promising, but the imprecise nature of
transgenesis may inadvertently alter tree phenotype, thus potentially impacting ecologically dependent organisms. We quantiÞed effects of genetic engineering and fungal inoculation of trees on insect
herbivores by using transgenic American chestnuts expressing an oxalate oxidase gene and wild-type
American and Chinese (C. mollissima Blume) chestnuts. Of three generalist folivores bioassayed, only
gypsy moth [Lymantria dispar (L.)] was affected by genetic modiÞcation, exhibiting faster growth
on transgenic than on wild-type chestnuts, whereas growth of polyphemus moth [Antheraea
polyphemus (Cramer)] differed between wild-type species, and fall webworm [Hyphantria cunea
(Drury)] performed equally on all trees. Inoculation of chestnuts with blight fungus had no effect on
the growth of two herbivores assayed (polyphemus moth and fall webworm). Enhanced Þtness of
gypsy moth on genetically modiÞed trees may hinder restoration efforts if this invasive herbivoreÕs
growth is improved because of transgene expression.
KEY WORDS genetic modiÞcation, forest restoration, oxalate oxidase, gypsy moth
The American chestnut, Castanea dentata (Marshall)
Borkhausen, was once a dominant or codominant tree
throughout much of the eastern deciduous forest of
North America (Braun 1950). The introduction of
chestnut blight during the Þrst half of the 20th century
killed ⬇3.5 billion trees (Roane et al. 1986), effectively
removing the species from the overstory. Root sprouting has enabled chestnuts to persist (Paillet 1984, 1988,
2002), though their biomass and function are now
greatly reduced (Ellison et al. 2005).
Chestnut blight is a fungal disease caused by the
exotic pathogen Cryphonectria parasitica (Murrill)
Barr (Barr 1979). After infection by either conidia or
ascospores, bark cankers form on branches or the main
stem (Gravatt 1949, Beattie and Diller 1954, Roane et
al. 1986). In lethal cankers, a mycelial mass penetrates
the inner bark and vascular cambium, encircling, girdling, and eventually killing the limb or trunk (Gravatt
1949, Beattie and Diller 1954, Roane et al. 1986). During the blight pandemic of the early 1900s, large trees
were killed in 3Ð 4 yr (Roane et al. 1986).
Conventional breeding programs have developed a
blight-resistant hybrid variety for reforestation (Burnham 1981, Diskin et al. 2006, Jacobs 2007). SpeciÞcally,
1 Department of Environmental and Forest Biology, SUNY-College
of Environmental Science and Forestry, 1 Forestry Dr, Syracuse, NY
13210.
2 Department of Ecology and Evolutionary Biology, University of
Tennessee, Knoxville, TN 37996.
3 Corresponding author, e-mail: dparry@esf.edu.
a ChineseÐAmerican chestnut hybrid (C. mollissima
Blume ⫻ C. dentata) is crossed with an American
chestnut at least three times before intercrossing. The
most blight-resistant progeny are then propagated for
use in chestnut restoration (Hebard 2005). Unfortunately, American chestnut has signiÞcantly less genetic diversity than the congeners Chinese chestnut
and Chinese chinquapin (C. seguinii Dode) (Huang et
al. 1998), and because current breeding programs use
only a fraction of an already small gene pool (Hebard
2005), backcrossed trees may ultimately be unsuccessful because of their reduced capacity to respond to
current and future environmental stressors (Huang et
al. 1998).
Limitations of conventional breeding have spurred
development of transgenic pathogen-resistant American chestnuts. A promising approach involves transformation with an oxalate oxidase (OxO)-encoding
gene (Polin et al. 2006) derived from the wheat (Triticum aestivum L.) protein germin (Lane et al. 1993).
Oxalate oxidase confers resistance to blight fungus by
1) increasing pH at the canker margins, reducing progression of infections (McCarroll and Thor 1978,
Roane et al. 1986, Dutton and Evans 1996); 2) through
antimicrobial signaling(Wu et al. 1995); and 3)
through defensive signaling (Apostol et al. 1989) properties associated with its breakdown product, H2O2.
Transgenic OxO-expressing American chestnut callus
exposed to oxalate did not experience a reduction in
lignin content when compared with a control, sug-
0046-225X/11/0955Ð0963$04.00/0 䉷 2011 Entomological Society of America
956
ENVIRONMENTAL ENTOMOLOGY
Vol. 40, no. 4
Fig. 1. Layout of the chestnut plot at Lafayette Road Experiment Station. Tree types are as follows: wild-type American
(WTA), transgenic American (TA), and Chinese (Ch) chestnut. ¤Inoculated with C. parasitica strain SG-1 on 4 June 2008.
#
Inoculated with C. parasitica strain EP155 on 4 June 2008.
gesting that this protein may provide protection
against blight development (Welch et al. 2007).
Because genetic engineering is essentially a random
process (Chyi et al. 1986, Mullins et al. 2001), it may
disrupt gene expression (Feldmann and Marks 1987,
Feldmann et al. 1989), leading to changes in phenotype (Dale and McPartlan 1992). Similarly, the pleiotropic effects of transgenes also may change phenotypic expression (Dale and McPartlan 1992, Käppeli
and Auberson 1998). Thus, genetic transformation
may inadvertently affect foliar chemistry and impart
Þtness changes on insect herbivores (e.g., Tiimonen et
al. 2005, Hjältén et al. 2007).
Few studies have investigated impacts of transgenic
plants with modiÞed OxO expression on insect herbivores. In corn (Zea mays L.), three transgenic lines
reduced leaf feeding and stalk tunneling by European
corn borer (Ostrinia nubilalis Hübner), and this insect
exhibited slower rates of growth and development
when fed an artiÞcial diet supplemented with H2O2
(Ramputh et al. 2002). In these corn lines, the secondary metabolite ferulic acid was found responsible
in reducing herbivore Þtness (Mao et al. 2007). Tobacco hornworm [Manduca sexta (L.)] and suckßy
[Tupiocoris notatus (Distant)] performance and preference, respectively, increased on one of three transgenic coyote tobacco (Nicotiana attenuata Torrey ex
Watson) lines with silenced germin-like gene expression, which coincided with lower levels of H2O2 (Lou
and Baldwin 2006). These studies support the potential roles of OxO and H2O2 activity in plantÐ herbivore
interactions.
The primary objective of our study was to determine if American chestnut transformed with a plasmid
(p⌬VspB-OxO) containing an OxO gene driven by a
vascular promoter (event LP28) affects the performance of three generalist insect folivores: gypsy moth
[Lymantria dispar (L.), Lepidoptera: Lymantriidae];
polyphemus moth [Antheraea polyphemus (Cramer),
Lepidoptera: Saturniidae]; and fall webworm [Hyphantria cunea (Drury), Lepidoptera: Arctiidae] that
feed on chestnut in nature. We also assessed whether
inoculation of all chestnut types with blight fungus
affects polyphemus moth and fall webworm growth.
Materials and Methods
Site Description and Plant Material. All experiments were conducted at the State University of New
YorkÕs College of Environmental Science and Forest-
ryÕs Lafayette Road Experiment Station in Syracuse,
NY. Three tree genotypes, wild-type American chestnut, transgenic American chestnut (event LP28), and
Chinese chestnut (ÔDaltonÕ cultivar), were randomly
planted in a 6 ⫻ 6 layout in an open Þeld (Fig. 1).
Wild-type American chestnuts were half-siblings and
came from a different cell line than genetically modiÞed American chestnuts. Chinese chestnuts were also
half-siblings. These trees exhibit a range of resistance
to chestnut blight, with the transgenic American demonstrating signiÞcantly increased resistance over the
wild-type American but less than the Chinese chestnut. Trees were 2 yr old and ranged from 0.6 to 1.9 m
in height (0.6 Ð2.6 cm in diameter) at the beginning of
the study. Given the spacing (2 m) and age of the
chestnuts, root grafting was highly unlikely but was
not assessed to avoid injuring the trees. Greenhouse
grown trees were Þeld planted in 2007 in a mix of
FafardÕs planting mix (Conrad Fafard, Inc., Agawam,
MA) (which contains a small amount of N-P-K fertilizer) and were then watered as needed during dry
periods.
Gypsy Moth Bioassay. We conducted a laboratory
feeding experiment by using fourth-instar gypsy
moths to determine whether caterpillar growth rate
was affected by tree type. Egg masses were collected
from northern red oak (Quercus rubra L.) in Quebec,
Canada, and larvae were reared on standard gypsy
moth diet (Southland Products, Lake Island, AR) in
paper cups at 18⬚C and a photoperiod of 14:10 (L:D)
h before experimentation. The experiment was conducted over a phenologically relevant time period
(early June in the study area) and corresponded with
the presence of fourth-instar gypsy moth in the Þeld.
Only chestnuts with sufÞcient foliage to sustain a feeding trial at the time of sampling were used, thus limiting the bioassay to 20 trees: Þve wild-type American,
six transgenic American, and nine Chinese chestnuts.
To reduce variability in leaf quality because of ontogeny, we restricted leaf sampling to full size leaves
corresponding to a plastochron position of four or
higher. Plastochron indices assign numbers to leaves
based on proximity to the terminal growing tip or bud
allowing standardized comparison between leaves on
different branches (e.g., Lamoreaux et al. 1978, Meyer
and Montgomery 1987). For each tree, two larger or
four smaller-sized leaves without herbivore damage
were clipped at the petiole and put in a labeled zip-loc
bag, placed in a cooler with ice, and transported to the
laboratory.
August 2011
POST AND PARRY: NON-TARGET EFFECTS OF TRANSGENIC AMERICAN CHESTNUT
Paper toweling was placed in each of 40 150-mm ⫻
25-mm petri dishes to absorb excess humidity. Because
mature leaf sizes varied considerably within and between the tree species, we tried to standardize leaf
area offered to larvae by placing one large or two
smaller-sized leaves in a petri dish. Moist paper toweling was wrapped around the petiole of each leaf to
maintain turgor. Before experimentation, leaves were
examined and any inadvertently collected insects removed. One gypsy moth caterpillar was haphazardly
selected, weighed using a microbalance, and randomly
assigned to a labeled dish, and then placed in an
environmental chamber. A second dish using foliage
from the same tree was prepared in an identical manner. Larvae were reared separately so that the growth
rate of individuals could be quantiÞed. Thus, 10, 12,
and 18 dishes (subsamples) were prepared from wildtype American, transgenic American, and Chinese
chestnut foliage, respectively. The caterpillars fed for
3 d at 15.5 h (24⬚C):8.5 h (20⬚C) L:D with no dishes
requiring more leaves during that time. Moistened
paper towels were replaced ⬇24 h and 48 h into the
experiment.
After 72 h, caterpillars were removed from dishes
and individually placed into labeled microtubes,
placed in a zip-loc bag, and frozen before being dried
in an oven at 45⬚C for 6 d. Dried caterpillars were
weighed to determine Þnal dry weights. Initial dry
weights (mg) were calculated using a fresh:dry weight
conversion estimate (y ⫽ 0.1306x ⫺ 0.00002; R2 ⫽
0.962) from a regression of a subset of larvae (n ⫽ 21)
from a previous study (D. P., unpublished data),
where y ⫽ dry mass and x ⫽ wet mass. Relative growth
rate (RGR) for each caterpillar was determined over
this 72-h period (Gordon 1968) and used as the response variable.
Polyphemus Moth Bioassay. A. polyphemus are
large, solitary, summer-feeding native caterpillars. We
conducted a bioassay using fourth-instar polyphemus
moths to investigate the effects of tree type and blight
fungus inoculation on their growth rate at the end of
July. Approximately half of each tree type was randomly inoculated with one of two strains of chestnut
blight fungus, SG-1 or EP155, on 4 June 2008 (Fig. 1).
Inoculations were made through a small wound cut on
one stem with a scalpel. A plug of mycelium (the
fungus on potato dextrose agar medium) was overlaid
and sealed with paraÞlm for one week (Powell et al.
2007). After 3Ð 4 wk, callus tissue formed at the insertion site of the transgenic and Chinese trees, whereas
typical blight symptoms (a sunken canker that envelopes the stem along with leaf wilting and formation of
new sprouts above and below the inoculation site,
respectively) formed on wild-type American chestnuts. For the bioassay, leaves were clipped irrespective of inoculated stems. Thus for multistemmed,
pathogen-injected chestnuts, herbivores consumed
foliage from inoculated stems, noninoculated stems,
or both. Because this feeding trial was conducted
later in the growing season, 11 additional trees and
31 trees in total (three noninoculated, wild-type
American; 5 inoculated, wild-type American; six
957
noninoculated, transgenic American; Þve inoculated, transgenic American; six noninoculated, Chinese; and six inoculated, Chinese chestnuts), were
available for experimentation.
Polyphemus moth eggs were obtained by mating
captive reared females (from Massachusetts and
neighboring New York) with wild males. Larvae were
reared on northern red oak foliage in plastic bins at
ambient temperature before experimentation. For
each tree in the experiment, two to four dishes were
used, with the number of dishes proportional to the
amount of foliage available on each tree. SpeciÞcally,
24, 28, and 40 dishes were from wild-type American,
transgenic American, and Chinese chestnuts, respectively. Of these, 18, 14, and 19 dishes, respectively,
were from pathogen-inoculated trees.
As larval development and molting is not synchronous in this species, freshly molted fourth instars were
removed from the rearing stock as they became available and were randomly assigned to individual trees.
Each caterpillar, regardless of start date (either 24 or
25 July), fed for 72 h at 26⬚C with a photoperiod of
14.5:9.5 (L:D) h. After 48 h, additional foliage was
added as needed. Caterpillars were oven dried for 10 d.
Initial dry weights were determined from a fresh:dry
weight conversion estimate (y ⫽ 0.1313x ⫺ 7.2213;
R2 ⫽ 0.907) derived from a regression of a subset of 14
larvae excluded from this bioassay. Relative growth
rate over this 72-h period was again the response
variable and was determined as above for gypsy moth.
Fall Webworm Bioassay. To further assay potential
effects of tree type and pathogen inoculation on herbivore growth, we used fall webworm, a native, colonial, late-season feeder and conducted a feeding bioassay at the end of August. Larvae were collected from
black cherry (Prunus serotina Ehrhart) in Syracuse,
NY and maintained in the laboratory on cut foliage for
48 h before experimentation. For the bioassay, two
large or four smaller-sized chestnut leaves were obtained from the same trees used in the Polyphemus
moth bioassay, placed in individual zip-loc bags, and
stored with moistened paper towels in two large ziploc bags in a refrigerator at 5⬚C overnight to minimize
desiccation. For each tree, two groups of Þve representative caterpillars of various instars of this colonial
species were placed into separate dishes. In total, 14,
21, and 24 dishes were prepared from wild-type American, transgenic American, and Chinese chestnut foliage, respectively. Of these, 9, 10, and 12 dishes, respectively, were from inoculated trees. Caterpillars
fed for 3 d at 13 h (24⬚C):11 h (20⬚C) L:D. Foliage was
added to one dish at 48 h. Larvae were removed from
each dish, and numbers of living and dead larvae were
recorded and placed in separate, labeled microtubes.
Caterpillars were oven dried for 8 d and weighed in
groups. Initial individual dry weights were determined
by dividing group fresh weights at the outset of the
experiment by the number of caterpillars in a group
and using a fresh:dry weight conversion estimate (y ⫽
0.1383x ⫹ 0.434; R2 ⫽ 0.969) from a subset of 23 larvae
excluded from this bioassay. As with the other insects
assayed, RGR was the response variable but for fall
958
ENVIRONMENTAL ENTOMOLOGY
Vol. 40, no. 4
Fig. 2. Fourth-instar gypsy moth mean (⫾SE) relative
growth rate on detached foliage from three tree types: wildtype American (WTA), transgenic American (TA), and Chinese (Ch) chestnut. Treatments with the same letter are not
signiÞcantly different (P ⬎ 0.05).
webworm, it was the mean value for all living caterpillars in a dish at experiment termination rather than
an individual.
Statistical Analysis. All bioassays used a completely
randomized design with subsampling. Thus for each
experiment, response variables were recorded from
individual petri dishes (subsamples), which were
nested under individual trees (experimental units or
replicates). The gypsy moth bioassay was conducted
before pathogen inoculation of the trees and was analyzed as a one-way analysis of variance (ANOVA)
with three treatments: wild-type American, transgenic
American, and Chinese chestnut. The remaining experiments were performed after trees were injected
with the blight fungus and thus were analyzed as 2 by
3 factorial designs. Pathogen inoculation (noninoculated or inoculated) and tree type were the two treatment factors investigated for these assays. An additional set of eight one-way ANOVAs was performed
for both the polyphemus moth and fall webworm
feeding experiments on all tree types, both individually and combined, but differentiating pathogen
strains, SG-1 and EP155, to ascertain whether C. parasitica strains should be considered separate treatments. For these same two bioassays, RGR variances
between noninoculated and inoculated, SG-1-inoculated, and EP155-inoculated chestnuts for all tree
types, both individually and combined, were tested for
homogeneity using LeveneÕs test (Levene 1960). This
test determined whether herbivore growth response
differed among single and multi-stemmed chestnuts
injected with the blight fungus. An analysis of the
latter two feeding trials incorporating only noninoculated trees was also conducted to exclude the pathogen inoculation treatment factor and allow comparison of tree type, as in the gypsy moth bioassay. An
orthogonal set of contrasts was constructed for all
ANOVAs, and where applicable, individual and main
effect means were separated a posteriori using TukeyÕs
Honestly SigniÞcant Difference (HSD) (Tukey 1949)
and SchefféÕs (Scheffé 1953) tests, respectively. All
statistics were computed using SAS (PROC GLM, SAS
Institute 2002).
Fig. 3. Effect of chestnut type (wild-type American,
WTA; transgenic American, TA; and Chinese, Ch) on fourthinstar polyphemus moth mean (⫾SE) relative growth rate
(A) with all trees included, regardless of C. parasitica inoculation treatment, and (B) only trees without C. parasitica
inoculation. Treatments with the same letter are not significantly different (P ⬎ 0.05).
Results
Gypsy Moth Bioassay. Relative growth rates of
fourth-instar gypsy moth were signiÞcantly affected
by tree type (F ⫽ 5.64; df ⫽ 2, 17; P ⬍ 0.014). Planned
comparisons showed no difference between tree species (F ⫽ 2.59; df ⫽ 1, 17; P ⬍ 0.127), whereas caterpillar growth rate was signiÞcantly higher (16.4%; F ⫽
8.08; df ⫽ 1, 17; P ⬍ 0.012) on transgenic American
compared with wild-type American chestnut leaves
(Fig. 2). Larvae reared on Chinese chestnut foliage
had signiÞcantly lower (13%; P ⬍ 0.05) RGRs than
those on transformed American chestnut leaves but
similar growth rates when compared with individuals
fed nontransformed American chestnut foliage (P ⬎
0.05) using two TukeyÕs HSD tests.
Polyphemus Moth Bioassay. There was no interaction between tree type and inoculation (F ⫽ 0.96; df ⫽
2, 25; P ⬍ 0.297) on fourth-instar polyphemus moth
RGR, thus only main effects were considered. Growth
rate of caterpillars was not signiÞcantly altered by
fungal injection (F ⫽ 0.19; df ⫽ 1, 25; P ⬍ 0.665).
However, the main effect of tree type was signiÞcant
(F ⫽ 4.49; df ⫽ 2, 25; P ⬍ 0.022; Fig. 3A), and a planned
comparison of caterpillar growth rates among species
was highly signiÞcant (F ⫽ 8.87; df ⫽ 1, 25; P ⬍ 0.007).
SpeciÞcally, larvae fed leaves from wild-type American trees had the greatest RGR among the three types,
which was only 3.2% greater than its transgenic counterpart (F ⫽ 0.26; df ⫽ 1, 25; P ⬍ 0.613), according to
August 2011
POST AND PARRY: NON-TARGET EFFECTS OF TRANSGENIC AMERICAN CHESTNUT
Fig. 4. Effect of chestnut type (wild-type American,
WTA; transgenic American, TA; and Chinese, Ch) and C.
parasitica inoculation on fall webworm mean (⫾SE) relative
growth rate. Treatments with the same letter are not significantly different (P ⬎ 0.05).
a planned contrast, but signiÞcantly higher (15.3%; P ⬍
0.05) than Chinese chestnuts using SchefféÕs test.
At the start of the bioassay, 31 trees with two to four
dishes each, for a total of 108 dishes were placed in an
environmental chamber. Because of inadvertent incorporation of 16 third-instars, analysis of an experimental design with both unequal replication and subsampling was performed on 92 sampling units.
Fall Webworm Bioassay. There was no interactive
effect between the two treatment factors (F ⫽ 0.05;
df ⫽ 2, 25; P ⬍ 0.953; Fig. 4), and the main effect of tree
type was not signiÞcant (F ⫽ 1.21; df ⫽ 2, 25; P ⬍
0.317). Inoculation reduced larval growth rates by
⬍1% (F ⫽ 0.04; df ⫽ 1, 25; P ⬍ 0.836).
Effect of Pathogen Strain on Herbivore Growth.
Polyphemus moth growth did not vary among inoculation treatments for nontransformed and transformed
American chestnut and all tree types combined (Table
1). For Chinese chestnuts, pathogen strain signiÞcantly affected growth (F ⫽ 4.48; df ⫽ 2, 9; P ⬍ 0.045).
Relative growth rate on detached foliage was similar
between noninoculated and inoculated trees (F ⫽
1.32; df ⫽ 1, 9; P ⬍ 0.281) but differed between strains
(F ⫽ 8.00; df ⫽ 1, 9; P ⬍ 0.02) according to two planned
contrasts. Using TukeyÕs HSD test, two pairwise comparisons between control chestnuts and trees injected
959
with either SG-1 or EP155 fungus indicated no significant difference (P ⬎ 0.05) in caterpillar growth. Fall
webworm responded similarly among inoculation
treatments for all chestnut types, both individually and
combined (Table 1).
Variance Tests Between Inoculated and Non-Inoculated Chestnuts. The variance of larval RGR did not
signiÞcantly differ (P ⬎ 0.05) between noninoculated
and inoculated, SG-1-inoculated, and EP155-inoculated chestnuts for either polyphemus moth or fall
webworm for all tree types, both individually and
combined (Table 2). Because of a lack of American
chestnuts injected with the SG-1 strain, four tests
could not be conducted.
Bioassays Using Non-Inoculated Chestnuts. Polyphemus moth growth was similar among tree types
(F ⫽ 0.56, df ⫽ 2, 12; P ⬍ 0.589; Fig. 3B). Fall webworm
growth was also not affected by treatment (F ⫽ 1.10;
df ⫽ 2, 12; P ⬍ 0.366).
Discussion
Nontarget effects of noninsecticidal transgenic
plants have not been extensively examined; this is
especially true for trees (but see Tiimonen et al. 2005,
Halpin et al. 2007, Hjältén et al. 2007). In our study, the
effects of American chestnut genetically modiÞed for
blight resistance were variable, increasing the growth
rate of gypsy moth relative to nontransformed American chestnut, but having no effect on polyphemus
moth and fall webworm.
The faster growth of fourth-instar gypsy moth fed
transgenic American chestnut foliage presents a potential problem for restoration efforts. The North
American range of gypsy moth (Sharov et al. 2002)
coincides with much of the area formerly dominated
by American chestnut (Gravatt 1949). Though this
particular transgenic line (LP28) is not currently under consideration for forest restoration (W. A. Powell,
personal communication) enhanced performance of
gypsy moth, a polyphagous, exotic, outbreak species
(Elkinton and Liebhold 1990), on genetically modiÞed chestnut foliage could become a concern because
chestnuts expressing OxO after transformation with a
Table 1. Mean relative growth rate (ⴞSE) of polyphemus moth and fall webworm on chestnut foliage from three tree types with different
C. parasitica inoculation treatments
Herbivore and
tree type
Polyphemus moth
Wild-type American
Transgenic American
Chinese
All
Fall webworm
Wild-type American
Transgenic American
Chinese
All
P
No pathogen
SG-1 strain
EP155 strain
(mg 䡠 mg⫺1 ⫻ d⫺1)
(mg 䡠 mg⫺1 ⫻ d⫺1)
(mg 䡠 mg⫺1 ⫻ d⫺1)
0.38 (0.029)
0.36 (0.015)
0.35 (0.014)ab
0.36 (0.013)
0.37 (0.025)
0.40 (0.022)
0.36 (0.020)a
0.38 (0.016)
0.41 (0.023)
0.38 (0.020)
0.28 (0.022)b
0.36 (0.016)
⬍0.495
⬍0.278
⬍0.045
⬍0.621
0.12 (0.040)
0.13 (0.022)
0.15 (0.023)
0.14 (0.014)
0.12 (0.042)
0.15 (0.035)
0.18 (0.033)
0.16 (0.020)
0.13 (0.040)
0.12 (0.029)
0.12 (0.033)
0.12 (0.018)
⬍0.966
⬍0.823
⬍0.389
⬍0.363
Values within a row followed by no letter are not signiÞcantly different (P ⬎ 0.05).
960
ENVIRONMENTAL ENTOMOLOGY
Table 2. Variance of polyphemus moth and fall webworm relative growth rate (mg 䡠 mgⴚ1 ⴛ dⴚ1)2 by tree type and C. parasitica treatment to determine whether stem type (single vs multistemmed) influenced the effect of pathogen inoculation on
herbivore performance
Herbivore and
tree type
Polyphemus moth
Wild-type American
Transgenic American
Chinese
All
Fall webworm
Wild-type American
Transgenic American
Chinese
All
Control
Inoculated
SG-1
strain
EP155
strain
0.0037
0.0014
0.0018
0.0019
0.0013
0.0008
0.0024
0.0026
0.0019a
⬍0.0001a
0.0006
0.0009
0.0005
0.0014
0.0001
0.0039
0.0027
0.0024
0.0021
0.0024
0.0048
0.0025
0.0051
0.0038
0.0049a
0.0009a
0.0019
0.0024
0.0070
0.0042
0.0074
0.0047
Treatment values within a row are not signiÞcantly different (P ⬎
0.05) from the control.
a
No test was performed due to a lack of replicates.
similar vector are now being used in Þeld trials. Although fourth-instar gypsy moth grown on transgenic
chestnut foliage were 16.4% larger than on nontransgenic foliage, similar changes have been observed for
this species when comparing their performance on
different nontransgenic genotypes of the same tree
species (Hwang and Lindroth 1997, Osier and Lindroth 2004).
Polyphemus moth, a polyphagous, solitary mid- to
late-season feeder, was unaffected by genetic modiÞcation of American chestnut for blight resistance, but
analysis was complicated by fungal inoculation. Inclusion of all trees, both noninoculated and pathogeninoculated, yielded a difference in herbivore growth
rate between wild-type American and Chinese chestnuts, but neither differed from the transgenic chestnut. A similar pattern was observed in another study;
in the presence or absence of phytopathogens, more
adults of the spotted tentiform leafminer [Phyllonorycter blancardella (F.)] emerged from a conventionally bred, fungal-resistant apple (Malus domestica
Borkhausen) cultivar than from a susceptible variety,
but neither differed from transgenic-susceptible and
transgenic-resistant genotypes, though development
times were comparable across all tree types (Vogler et
al. 2010). However, in our study, analysis restricted to
noninoculated chestnuts found no variation in larval
performance among tree types. A signiÞcantly greater
RGR on foliage from Chinese chestnut injected with
SG-1 relative to EP155 blight fungus suggests that in
future studies, pathogen strains should be considered
separate treatments.
Although differing in biological and ecological attributes, all three folivores assayed are free-feeders and
thus members of the same functional guild. Other types
of phytophagous insects may respond differently to these
tree types. In particular, phloem-feeders may exhibit
greater Þtness effects from LP28 chestnuts because a
vascular promoter drives the OxO gene. Changes in herbivore preference or performance on genetically modiÞed plants could result from transgene expression, either through the primary (Vaeck et al. 1987) or
Vol. 40, no. 4
secondary (pleiotropic) functions (Puterka et al. 2002)
of the novel protein. Alternatively, effects on phytophagous insects may be caused by unintended alterations in
plant phenotype from the transformation process (Käppeli and Auberson 1998, Conner and Jacobs 1999,
Latham et al. 2006), such as somaclonal variation (Larkin
and Scowcroft 1981), insertion of multiple intact or rearranged transgene copies (Jones et al. 1987, Jorgensen
et al. 1987), insertional mutagenesis (Feldmann and
Marks 1987, Feldmann et al. 1989), epistatic interactions
between the transgene and an endogenous gene (Napoli
et al. 1990), (trans)gene silencing (Matzke and Matzke
1995), or a position effect (Jones et al. 1985, Nagy et al.
1985). Instances in which herbivore Þtness varies across
multiple transgenic lines, especially when response to
only one event deviates from wild-type plants, would be
more indicative of a transgenic engineering side effect
rather than the expression product itself (Tiimonen et al.
2005, Brodeur-Campbell et al. 2006). As more chestnut
lines become available, we will be able to more conclusively evaluate whether the effects on insect performance are indicative of transgenic OxO-expressing
chestnuts or are speciÞc to the single transformation
event LP28.
Blight resistance levels determined from introduction of C. parasitica to the three chestnut types contrasted with growth patterns exhibited by the three
herbivores. Transgenic American chestnut line LP28
demonstrated enhanced pathogen resistance over
wild-type trees, as indicated by postinoculation survival (W. A. Powell, unpublished data). When restricting analysis to noninoculated chestnuts, lack of correlation between fungal resistance levels and larval
growth rates among trees types for each herbivore
indicates that pathogen resistance alone is a poor predictor of folivore Þtness. Unintended phenotypic effects resulting from tree transformation (Dale and
McPartlan 1992, Käppeli and Auberson 1998) the inclusion of a different species to serve as a positive
control, or both (Powell et al. 2007) may obscure
relationships between resistance level and herbivore
performance.
Foliage selection for the polyphemus moth and fall
webworm bioassays was done independently of fungal
injection site on multi-stemmed plants, thus potentially confounding our analyses of insect growth rate.
Because inoculated stems contained greater concentrations of the blight fungus and because plants respond to pathogen infection both locally and systemically (Fodor et al. 1997, Bonello and Blodgett 2003),
we assessed whether there was greater variation in
herbivore growth on leaves from inoculated, multistemmed trees relative to single-stemmed trees (e.g.,
Rostás et al. 2003). However, across all tree types,
there were no differences in the variances of RGR for
polyphemus moth and fall webworm, respectively,
suggesting that the effects of fungal inoculation and
stem type were negligible.
Traditional breeding, speciÞcally backcrossing
(Burnham 1981), is an alternative to biotechnology for
chestnut restoration and as with genetic engineering,
produces phenotypes that vary in suitability for non-
August 2011
POST AND PARRY: NON-TARGET EFFECTS OF TRANSGENIC AMERICAN CHESTNUT
target herbivores. In one study, gypsy moths grew
faster on Chinese-American hybrid seedlings, whereas
consumption did not differ from control wild-type
seedlings (Rieske et al. 2003). In a subsequent study,
Kellogg et al. (2005) showed that fall webworm consumption and gypsy moth development were similar
on foliage from mature, blight-inoculated American,
hybrid, and Þrst generation backcross chestnuts, but
gypsy moth growth, consumption, and Þnal larval
weights varied with tree type. These results and ours
suggest that conventionally bred and transgenic
blight-resistant trees have a similar range of effects on
herbivores.
Our study showed that gypsy moth Þtness is affected by genetically modiÞed American chestnuts
expressing OxO to enhance blight resistance. The inclusion of only one transgenic line prevents the determination of whether transgene expression or the
inexact gene insertion process is responsible for observed differences. Though LP28 chestnuts are not
being considered further for forest restoration, it is
critical to determine whether this effect is driven by
the transformation event, or if it is an outcome of
product expression. Future studies need to incorporate greater replication, multiple transgenic lines,
more feeding guilds, feeding bioassays spanning larval
development, and chemical and molecular analyses of
foliar tissue to conÞrm or refute these results, determine whether other nontarget insect herbivores could
be affected by transgenic blight-resistant chestnuts,
and elucidate the underlying mechanism(s) involved
in this system before considering genetically modiÞed,
OxO-expressing American chestnuts for forest restoration.
Acknowledgments
We thank A. Garzón, P. Barber, B. Hoven, and M. Phillips
for assistance in the Þeld and laboratory and S. Stehman for
help with statistical analysis. We are deeply thankful to W.
Powell and C. Maynard for providing access to their trees.
This study would not have been possible without the longterm, intensive efforts of W. Powell, C. Maynard, and the rest
of the American chestnut Research and Restoration Program
to develop transgenic American chestnuts resistant to chestnut blight. We thank J. Lundgren and two anonymous reviewers whose thoughtful comments improved this manuscript. The research was supported by the Entomological
Foundation, The LeRoy C. Stegeman Endowment in Invertebrate Ecology, NY Chapter of The American Chestnut
Foundation, and ArborGen LLC.
References Cited
Apostol, I., P. F. Heinstein, and P. S. Low. 1989. Rapid stimulation of an oxidative burst during elicitation of cultured
plant cells. Plant Physiol. 90: 109 Ð116.
Barr, M. E. 1979. The Diaporthales in North America with
emphasis on Gnomonia and its segregants. Mycologia
Memoir 7: 232 pp.
Beattie, R. K., and J. D. Diller. 1954. Fifty years of chestnut
blight in America. J. For. 52: 323Ð329.
Bonello, P., and J. T. Blodgett. 2003. Pinus nigra-Sphaeropsis
sapinea as a model pathosystem to investigate local and
961
systemic effects of fungal infection of pines. Physiol. Mol.
Plant Pathol. 63: 249 Ð261.
Braun, E. L. 1950. Deciduous forests of eastern North
America. Blackiston, Philadelphia, PA.
Brodeur-Campbell, S. E., J. A. Vucetich, D. L. Richter, T. A.
Waite, J. N. Rosemier, and C.-J. Tsai. 2006. Insect herbivory on low-lignin transgenic aspen. Environ. Entomol.
35: 1696 Ð1701.
Burnham, C. R. 1981. Blight-resistant American chestnut:
thereÕs hope. Plant Dis. 65: 459 Ð 460.
Chyi, Y.-S., R. A. Jorgensen, D. Goldstein, S. D. Tanksley, and
F. Loaiza-Figueroa. 1986. Locations and stability in the
Lycopersicon genome. Mol. Gen. Genet. 204: 64 Ð 69.
Conner, A. J., and J.M.E. Jacobs. 1999. Genetic engineering
of crops as potential source of genetic hazard in the
human diet. Mutat. Res. 443: 223Ð234.
Dale, P. J., and H. C. McPartlan. 1992. Field performance of
transgenic potato plants compared with controls regenerated from tuber discs and shoot cuttings. Theor. Appl.
Genet. 84: 585Ð591.
Diskin, M., K. C. Steiner, and F. V. Hebard. 2006. Recovery
of American chestnut characteristics following hybridization and backcross breeding to restore blight-ravaged
Castanea dentata. For. Ecol. Manage. 223: 439 Ð 447.
Dutton, M., and C. Evans. 1996. Oxalate production by fungi: its role in pathogenicity and ecology in the soil environment. Can. J. Microbiol. 42: 881Ð 895.
Elkinton, J. S., and A. M. Liebhold. 1990. Population dynamics of gypsy moth in North America. Annu. Rev.
Entomol. 35: 571Ð96.
Ellison, A. M., M. S. Bank, B. D. Clinton, E. A. Colburn, K.
Elliot, C. R. Ford, D. R. Foster, B. D. Kloeppel, J. D.
Knoepp, G. M. Lovett, et al. 2005. Loss of foundation
species: consequences for the structure and dynamics of
forested ecosystems. Front. Ecol. Environ. 3: 479 Ð 486.
Feldmann, K. A., and M. D. Marks. 1987. Agrobacteriummediated transformation of germinating seeds of Arabidopsis thaliana: a non-tissue culture approach. Mol. Gen.
Genet. 208: 1Ð9.
Feldmann, K. A., M. D. Marks, M. L. Christianson, and R. S.
Quatrano. 1989. A dwarf mutant of Arabidopsis generated by T-DNA insertion mutagenesis. Science 243: 1351Ð
1354.
Fodor, J., G. Gullner, A. L. Ádám, B. Barna, T. Kömives, and
Z. Király. 1997. Local and systemic responses of antioxidants to tobacco mosaic virus infection and to salicylic
acid to tobacco. Plant Physiol. 114: 1443Ð1451.
Gordon, H. T. 1968. Quantitative aspects of insect nutrition.
Zoology 8: 131Ð138.
Gravatt, F. 1949. Chestnut blight in Asia and North America. Unasylva 3: 3Ð7.
Halpin, C., S. C. Thain, E. L. Tilston, E. Guiney, C. Lapierre,
and D. W. Hopkins. 2007. Ecological impacts of trees
with modiÞed lignin. Tree Genet. Genomes 3: 101Ð110.
Hebard, F. 2005. The backcross breeding program of the
American chestnut foundation. J. Am. Chestnut Found.
19: 55Ð78.
Hjältén, J., A. Lindau, A. Wennström, P. Blomberg, J.
Witzell, V. Hurry, and L. Ericson. 2007. Unintentional changes of defence traits in GM trees can inßuence plant-herbivore interactions. Basic Appl. Ecol. 8:
434 Ð 443.
Huang, H., F. Dane, and T. L. Kubisiak. 1998. Allozyme and
RAPD analysis of the genetic diversity and geographic
variation in wild populations of the American chestnut
(Fagaceae). Am. J. Bot. 85: 1013Ð1021.
962
ENVIRONMENTAL ENTOMOLOGY
Hwang, S.-Y., and R. L. Lindroth. 1997. Clonal variation in
foliar chemistry of aspen: effects on gypsy moths and
forest tent caterpillars. Oecologia (Berl.). 111: 99 Ð108.
Jacobs, D. F. 2007. Toward development of silvical strategies for forest restoration of American chestnut (Castanea
dentata) using blight-resistant hybrids. Biol. Conserv. 137:
497Ð506.
Jones, J.D.G., P. Dunsmuir, and J. Bedbrook. 1985. High
level expression of introduced chimaeric genes in regenerated transformed plants. EMBO J. 4: 2411Ð2418.
Jones, J.D.G., D. E. Gilbert, K. L. Grady, and R. A. Jorgensen.
1987. T-DNA structure and gene expression in petunia
plants transformed by Agrobacterium tumefaciens C58 derivatives. Mol. Gen. Genet. 207: 478 Ð 485.
Jorgensen, R., C. Snyder, and J.D.G. Jones. 1987. T-DNA is
organized predominantly in inverted repeat structures in
plants transformed with Agrobacterium tumefaciens C58
derivatives. Mol. Gen. Genet. 207: 471Ð 477.
Käppeli, O., and L. Auberson. 1998. How safe is safe enough
in plant genetic engineering? Trends Plant Sci. 3: 276 Ð
281.
Kellogg, S. K., F. V. Hebard, and L. K. Rieske. 2005. Interactions between a fungal pathogen, foliar properties, and
generalist herbivores. Entomol. Exp. Appl. 117: 209 Ð219.
Lamoreaux, R. J., W. R. Chaney, and K. M. Brown. 1978.
The plastochron index: a review after two decades of use.
Am. J. Bot. 65: 586 Ð593.
Lane, B. G., J. M. Dunwell, J. A. Ray, M. R. Schmitt, and A. C.
Cuming. 1993. Germin, a protein marker of early plant
development, is an oxalate oxidase. J. Biol. Chem. 268:
12239 Ð12242.
Larkin, P. J., and W. R. Scowcroft. 1981. Somaclonal variation Ð a novel source of variation from cell cultures for
plant improvement. Theor. Appl. Genet. 60: 197Ð214.
Latham, J. R., A. K. Wilson, and R. A. Steinbrecher. 2006.
The mutational consequences of plant transformation.
J. Biomed. Biotechnol. 2006: 1Ð7.
Levene, H. 1960. Robust tests for equality of variances, pp.
278Ð292. In I. Olkin, S. G. Ghurye, W. Hoeffding, W. G.
Madow, and H. B. Mann (eds.), Contributions to probability
and statistics. Stanford University Press, Stanford, CA.
Lou, Y., and I. T. Baldwin. 2006. Silencing of a germin-like
gene in Nicotiana attenuata improves performance of native herbivores. Plant Physiol. 140: 1126 Ð1136.
Mao, J., A. J. Burt, A.-I. Ramputh, J. Simmonds, L. Cass, K.
Hubbard, S. Miller, I. Altosaar, and J. T. Arnason. 2007.
Diverted secondary metabolism and improved resistance
to European corn borer (Ostrinia nubilalis) in maize (Zea
mays L.) transformed with wheat oxalate oxidase. J. Agric.
Food Chem. 55: 2582Ð2589.
Matzke, M. A., and A.J.M. Matzke. 1995. How and why do
plants inactivate homologous (trans)genes? Plant
Physiol. 107: 679 Ð 685.
McCarroll, D. R., and E. Thor. 1978. The role of oxalic acid
in the pathogenesis of Endothia parasitica, pp. 60 Ð 63. In
W. L. MacDonald, F. C. Cech, J. Luchock and C. Smith
(eds.), Proceedings of the American chestnut symposium. West Virginia University Books, Morgantown, WV,
4 Ð5 January 1978.
Meyer, G. A., and M. E. Montgomery. 1987. Relationships
between leaf age and the food quality of cottonwood
foliage for the gypsy moth, Lymantria dispar. Oecologia
(Berl.). 72: 527Ð532.
Mullins, E. D., X. Chen, P. Romaine, R. Raina, D. M. Geiser,
and S. King. 2001. Agrobacterium-mediated transformation of Fusarium oxysporum: an efÞcient tool for insertional mutagenesis and gene transfer. Phytopathology. 91:
173Ð180.
Vol. 40, no. 4
Nagy, F., G. Morelli, R. T. Fraley, S. G. Rogers, and N.-H.
Chua. 1985. Photoregulated expression of a pea rbcS
gene in leaves of transgenic plants. EMBO J. 4: 3063Ð3068.
Napoli, C., C. Lemieux, and R. Jorgensen. 1990. Introduction of a chimeric chalcone synthase gene into petunia
results in reversible co-suppression of homologous genes
in trans. Plant Cell 2: 279 Ð289.
Osier, T. L., and R. L. Lindroth. 2004. Long-term effects of
defoliation on quaking aspen in relation to genotype and
nutrient availability: plant growth, phytochemistry and
insect performance. Oecologia (Berl.). 139: 55Ð 65.
Paillet, F. L. 1984. Growth-form and ecology of American
chestnut sprout clones in northeastern Massachusetts.
Bulletin of the Torrey Botanical Club 111: 316 Ð328.
Paillet, F. L. 1988. Character and distribution of American
chestnut sprouts in southern New England woodlands.
Bulletin of the Torrey Botanical Club 115: 32Ð 44.
Paillet, F. L. 2002. Chestnut: history and ecology of a transformed species. J. Biogeogr. 29: 1517Ð1530.
Polin, L. D., H. Liang, R. E. Rothrock, M. Nishii, D. L. Diehl,
A. E. Newhouse, C. J. Nairn, W. A. Powell, and C. A. Maynard. 2006. Agrobacterium-mediated transformation of
American chestnut (Castanea dentata (Marsh.) Borkh.) somatic embryos. Plant Cell Tissue Organ Cult. 84: 69Ð78.
Powell, W. A., P. Morley, M. King, and C. A. Maynard. 2007.
Small stem chestnut blight resistance assay. J. Am. Chestnut Found. 21: 34 Ð38.
Puterka, G. J., C. Bocchetti, P. Dang, R. L. Bell, and R.
Scorza. 2002. Pear transformed with a lytic peptide
gene for disease control affects nontarget organism,
pear psylla (Homoptera: Psyllidae). Hortic. Entomol.
95: 797Ð 802.
Ramputh, A. I., J. T. Arnason, L. Cass, and J. A. Simmonds.
2002. Reduced herbivory of the European corn borer
(Ostrinia nubilalis) on corn transformed with germin, a
wheat oxalate oxidase gene. Plant Sci. 162: 431Ð 440.
Rieske, L. K., C. C. Rhoades, and S. P. Miller. 2003. Foliar
chemistry and gypsy moth, Lymantria dispar (L.), herbivory on pure Am. chestnut, Castanea dentata (Fam:
Fagaceae), and a disease-resistant hybrid. Environ. Entomol. 32: 359 Ð365.
Roane, M. K., G. J. Griffin, and J. R. Elkins. 1986. Chestnut
blight, other Endothia diseases, and the genus Endothia.
APS Press, St. Paul, MN.
Rostás, M., M. Simon, and M. Hilker. 2003. Ecological
cross-effects of induced plant responses towards herbivores and phytopathogenic fungi. Basic Appl. Ecol. 4:
43Ð 62.
SAS Institute. 2002. SAS/STAT userÕs guide, version 9.1.3
for windows. SAS Institute, Cary, NC.
Scheffé, H. 1953. A method for judging all contrasts in the
analysis of variance. Biometrika 40: 87Ð104.
Sharov, A. A., D. Leonard, A. M. Liebhold, E. A. Roberts, and
W. Dickerson. 2002. “Slow the spread”: a national program to contain the gypsy moth. J. For. 100: 30 Ð36.
Tiimonen, H., T. Aronen, T. Laakso, P. Saranpää, V.
Chiang, T. Ylioja, H. Roininen, and H. Häggman. 2005.
Does lignin modiÞcation affect feeding preference or
growth performance of insect herbivores in transgenic
silver birch (Betula pendula Roth)? Planta 222: 699 Ð
708.
Tukey, J. W. 1949. Comparing individual means in the analysis of variance. Biometrics 5: 99 Ð114.
Vaeck, M., A. Reynaerts, H. Höfte, S. Jansens, M. De Beuckeleer, C. Dean, M. Zabeau, M. Van Montagu, and J.
Leemans. 1987. Transgenic plants protected from insect
attack. Nature 328: 33Ð37.
August 2011
POST AND PARRY: NON-TARGET EFFECTS OF TRANSGENIC AMERICAN CHESTNUT
Vogler, U., A. S. Rott, C. Gessler, and S. Dorn. 2010. How
transgenic and classically bred apple genotypes affect
non-target organisms on higher trophic levels. Entomol.
Exp. Appl. 134: 114 Ð121.
Welch, A. J., A. J. Stipanovic, C. A. Maynard, and W. A.
Powell. 2007. The effects of oxalic acid on transgenic
Castanea dentata callus tissue expressing oxalate oxidase.
Plant Sci. 172: 488 Ð 496.
963
Wu, G., B. J. Shortt, E. B. Lawrence, E. B. Levine, K. C.
Fitzsimmons, and D. M. Shah. 1995. Disease resistance
conferred by expression of a gene encoding H2O2-generating glucose oxidase in transgenic potato plants. Plant
Cell 7: 1357Ð1368.
Received 6 March 2010; accepted 15 February 2011.