Fitness cost of resistance to Bt cotton linked with increased gossypol content in pink bollworm larvae.
ABSTRACT Fitness costs of resistance to Bacillus thuringiensis (Bt) crops occur in the absence of Bt toxins, when individuals with resistance alleles are less fit than individuals without resistance alleles. As costs of Bt resistance are common, refuges of non-Bt host plants can delay resistance not only by providing susceptible individuals to mate with resistant individuals, but also by selecting against resistance. Because costs typically vary across host plants, refuges with host plants that magnify costs or make them less recessive could enhance resistance management. Limited understanding of the physiological mechanisms causing fitness costs, however, hampers attempts to increase costs. In several major cotton pests including pink bollworm (Pectinophora gossypiella), resistance to Cry1Ac cotton is associated with mutations altering cadherin proteins that bind this toxin in susceptible larvae. Here we report that the concentration of gossypol, a cotton defensive chemical, was higher in pink bollworm larvae with cadherin resistance alleles than in larvae lacking such alleles. Adding gossypol to the larval diet decreased larval weight and survival, and increased the fitness cost affecting larval growth, but not survival. Across cadherin genotypes, the cost affecting larval growth increased as the gossypol concentration of larvae increased. These results suggest that increased accumulation of plant defensive chemicals may contribute to fitness costs associated with resistance to Bt toxins.
Article: Areawide suppression of European corn borer with Bt maize reaps savings to non-Bt maize growers.[show abstract] [hide abstract]
ABSTRACT: Transgenic maize engineered to express insecticidal proteins from the bacterium Bacillus thuringiensis (Bt) has become widely adopted in U.S. agriculture. In 2009, Bt maize was planted on more than 22.2 million hectares, constituting 63% of the U.S. crop. Using statistical analysis of per capita growth rate estimates, we found that areawide suppression of the primary pest Ostrinia nubilalis (European corn borer) is associated with Bt maize use. Cumulative benefits over 14 years are an estimated $3.2 billion for maize growers in Illinois, Minnesota, and Wisconsin, with more than $2.4 billion of this total accruing to non-Bt maize growers. Comparable estimates for Iowa and Nebraska are $3.6 billion in total, with $1.9 billion for non-Bt maize growers. These results affirm theoretical predictions of pest population suppression and highlight economic incentives for growers to maintain non-Bt maize refugia for sustainable insect resistance management.Science 10/2010; 330(6001):222-5. · 31.20 Impact Factor
[show abstract] [hide abstract]
ABSTRACT: Evolution of insect resistance threatens the continued success of transgenic crops producing Bacillus thuringiensis (Bt) toxins that kill pests. The approach used most widely to delay insect resistance to Bt crops is the refuge strategy, which requires refuges of host plants without Bt toxins near Bt crops to promote survival of susceptible pests. However, large-scale tests of the refuge strategy have been problematic. Analysis of more than a decade of global monitoring data reveals that the frequency of resistance alleles has increased substantially in some field populations of Helicoverpa zea, but not in five other major pests in Australia, China, Spain and the United States. The resistance of H. zea to Bt toxin Cry1Ac in transgenic cotton has not caused widespread crop failures, in part because other tactics augment control of this pest. The field outcomes documented with monitoring data are consistent with the theory underlying the refuge strategy, suggesting that refuges have helped to delay resistance.Nature Biotechnology 03/2008; 26(2):199-202. · 29.50 Impact Factor
[show abstract] [hide abstract]
ABSTRACT: Transgenic crops producing Bacillus thuringiensis (Bt) toxins for insect pest control have been successful, but their efficacy is reduced when pests evolve resistance. Here we review the definition of field-evolved resistance, the relationship between resistance and field control problems, the theory underlying strategies for delaying resistance, and resistance monitoring methods. We also analyze resistance monitoring data from five continents reported in 41 studies that evaluate responses of field populations of 11 lepidopteran pests to four Bt toxins produced by Bt corn and cotton. After more than a decade since initial commercialization of Bt crops, most target pest populations remain susceptible, whereas field-evolved resistance has been documented in some populations of three noctuid moth species: Spodoptera frugiperda (J. E. Smith) to Cry1F in Bt corn in Puerto Rico, Busseola fusca (Fuller) to CrylAb in Bt corn in South Africa, and Helicoverpa zea (Boddie) to CrylAc and Cry2Ab in Bt cotton in the southeastern United States. Field outcomes are consistent with predictions from theory, suggesting that factors delaying resistance include recessive inheritance of resistance, abundant refuges of non-Bt host plants, and two-toxin Bt crops deployed separately from one-toxin Bt crops. The insights gained from systematic analyses of resistance monitoring data may help to enhance the durability of transgenic insecticidal crops. We recommend continued use of the longstanding definition of resistance cited here and encourage discussions about which regulatory actions, if any, should be triggered by specific data on the magnitude, distribution, and impact of field-evolved resistance.Journal of Economic Entomology 12/2009; 102(6):2011-25. · 1.70 Impact Factor
Fitness Cost of Resistance to Bt Cotton Linked with
Increased Gossypol Content in Pink Bollworm Larvae
Jennifer L. Williams1, Christa Ellers-Kirk1, Robert G. Orth2, Aaron J. Gassmann1,3, Graham Head2, Bruce E.
Tabashnik1, Yves Carrie `re1*
1Department of Entomology, University of Arizona, Tucson, Arizona, United States of America, 2Monsanto LLC, St. Louis, Missouri, United States of America,
3Department of Entomology, Iowa State University, Ames, Iowa, United States of America
Fitness costs of resistance to Bacillus thuringiensis (Bt) crops occur in the absence of Bt toxins, when individuals with
resistance alleles are less fit than individuals without resistance alleles. As costs of Bt resistance are common, refuges of non-
Bt host plants can delay resistance not only by providing susceptible individuals to mate with resistant individuals, but also
by selecting against resistance. Because costs typically vary across host plants, refuges with host plants that magnify costs or
make them less recessive could enhance resistance management. Limited understanding of the physiological mechanisms
causing fitness costs, however, hampers attempts to increase costs. In several major cotton pests including pink bollworm
(Pectinophora gossypiella), resistance to Cry1Ac cotton is associated with mutations altering cadherin proteins that bind this
toxin in susceptible larvae. Here we report that the concentration of gossypol, a cotton defensive chemical, was higher in
pink bollworm larvae with cadherin resistance alleles than in larvae lacking such alleles. Adding gossypol to the larval diet
decreased larval weight and survival, and increased the fitness cost affecting larval growth, but not survival. Across cadherin
genotypes, the cost affecting larval growth increased as the gossypol concentration of larvae increased. These results
suggest that increased accumulation of plant defensive chemicals may contribute to fitness costs associated with resistance
to Bt toxins.
Citation: Williams JL, Ellers-Kirk C, Orth RG, Gassmann AJ, Head G, et al. (2011) Fitness Cost of Resistance to Bt Cotton Linked with Increased Gossypol Content in
Pink Bollworm Larvae. PLoS ONE 6(6): e21863. doi:10.1371/journal.pone.0021863
Editor: Guy Smagghe, Ghent University, Belgium
Received February 7, 2011; Accepted June 11, 2011; Published June 30, 2011
Copyright: ? 2011 Williams et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This research was partly funded by USDA Biotechnology Risk Assessment Research Grant 2003-04371 (http://www.csrees.usda.gov/fo/
biotechnologyriskassessment.cfm). No additional external funding received for this study. The funders had no role in study design, data collection and analysis,
decision to publish, or preparation of the manuscript.
Competing Interests: BE Tabashnik received support for research that is not related to this publication from the following sources: Cotton Foundation, Cotton
Inc., National Cotton Council, Monsanto, and Dow AgroSciences. Tabashnik is also co-author of a patent application on engineering modified Bt toxins to counter
pest resistance, which is related to research described by Soberon et al. (2007, Science 318: 1640-1642). Robert G. Orth and Graham Head are employed by a
commercial company, Monsanto LLC, which provides transgenic crops for the control of pests. This does not alter the authors’ adherence to all the PLoS ONE
policies on sharing data and materials.
* E-mail: firstname.lastname@example.org
Corn and cotton engineered to produce insecticidal proteins
from Bacillus thuringiensis (Bt) can increase agricultural profitability
while reducing reliance on insecticide sprays [1,2]. However, field-
evolved resistance to toxins in Bt crops, which has been reported in
several species of major insect pests, threatens these benefits [3–7].
Analysis of global monitoring data suggests that host plants that do
not make Bt toxins and grow near Bt crops can reduce the risk of
resistance [2–4]. Such non-Bt plant ‘‘refuges’’ can provide many
susceptible individuals to mate with the rare resistant individuals
surviving on Bt crops, yielding hybrid offspring. Refuges are
expected to delay resistance most effectively when resistance is
inherited as a recessive trait, so the hybrid offspring are killed on
Bt crops [4,8,9].
Fitness costs of resistance occur in the absence of Bt toxins
through pleiotropic effects that reduce fitness of individuals
carrying resistance alleles relative to susceptible individuals that
lack such alleles . Because costs of Bt resistance are common,
refuges not only delay resistance by providing susceptible
individuals to mate with resistant individuals, but also by selecting
against resistance alleles [10–14]. Costs are modulated by
variation in environmental conditions, including host plants,
competition, overwintering, and natural enemies . According-
ly, refuges that magnify costs or make them less recessive could
enhance resistance management [9,10,14]. Although dozens of
studies have documented fitness costs of Bt resistance that affect
many life history traits, little is known about the physiological
mechanisms that cause such costs . In particular, the
mechanisms underlying variation in costs among host plants are
not well understood, which hampers attempts to identify or create
refuge plants that magnify costs.
Knowledge of the molecular and genetic basis of resistance to Bt
toxins is essential for understanding what causes fitness costs and
why such costs vary among host plants. In pink bollworm,
Pectinophora gossypiella, and two other major lepidopteran pests of
cotton, mutations in genes encoding cadherin proteins that bind Bt
toxins are associated with resistance to Bt toxins [15–17]. In pink
bollworm, three mutant alleles (r1, r2, and r3) linked with
resistance to Bt toxin Cry1Ac encode incomplete versions of a
cadherin protein that binds Cry1Ac in susceptible larvae [16,18].
The normal role of Bt toxin-binding cadherin proteins, which
occur in the larval midgut, remains unclear. They may affect the
morphology of microvilli on the apical surface of midgut cells
PLoS ONE | www.plosone.org1 June 2011 | Volume 6 | Issue 6 | e21863
through several processes, including enhancement of cell adhesion
and guidance of cell differentiation [19–22]. If cadherin mutations
conferring resistance to Bt toxins interfere with midgut functional
or structural integrity, such mutations could cause fitness costs by
increasing the absorption and concentration of plant defensive
chemicals in insect larvae.
Here we tested the hypothesis that fitness costs of resistance to
Bt cotton in pink bollworm are associated with increased
concentration in larvae of a plant defensive chemical, gossypol.
Gossypol is a polyphenolic aldehyde from cotton (Gossypium spp.)
that is toxic to many insects and plant pathogens [23–25].
Although the mechanisms responsible for gossypol toxicity to
insects remain unknown, gossypol in the hemolymph may
permeate cells and cause toxic effects on many life history traits.
We reported previously that adding gossypol to the larval diet of
pink bollworm reduced performance and increased fitness costs
, but we did not measure the concentration of gossypol in
larvae or identify their cadherin genotype. The results reported
here confirm that gossypol reduces larval performance. We
discovered that gossypol concentration was higher in larvae with
cadherin resistance alleles than in larvae without such alleles. The
results also show that across larval cadherin genotypes, increased
gossypol concentration was associated with higher fitness costs.
Effects of Cadherin Genotype on Larval Gossypol
As expected, gossypol was not detected in any larvae from
control diet without gossypol, either from the related susceptible
strain (MOV97H1-S) and resistant strain (MOV97-H1R) in
experiment one, or from the hybrid strain (MOV97-H3) in
experiment two (Table 1, n=10). In contrast, across the two
experiments, gossypol was detected in 93% of larvae from
gossypol-treated diet (n=175).
Across both experiments, the proportion of larvae from gossypol
diet in which we detected no gossypol was significantly higher for
susceptible (ss) larvae (10/47=0.21) than for larvae with either one
r allele (rs) or two r alleles (rr) (3/128=0.02) (Fisher’s exact test,
P,0.0001). For larvae fed gossypol diet in both experiments,
gossypol concentration was higher in rs and rr larvae than in ss
larvae (Fig. 1 and Table 1). In experiment one, gossypol
concentration was 4.8 times higher in rr (2.67 mg/g) than ss
(0.56 mg/g) (t=4.58, df=81, P,0.0001). In experiment two,
relative to the gossypol content of ss (0.41 mg/g), gossypol content
was 13.6 times higher in rr (5.60 mg/g) and 4.4 times higher in rs
(1.80 mg/g). Gossypol content was significantly higher in rr than ss
(t=6.35, df=84, one-sided P,0.0001) and in rs than ss (t=7.55,
df=84, one-sided P,0.0001), but did not differ significantly
between rs and rr (t=1.67, df=84, P=0.098). Because gossypol
content of rs differed from ss but not rr, inheritance of this trait was
not recessive. In both experiments, gossypol concentration differed
between each rs or rr genotypes and ss (Fig. 1A, one-sided P
values,0.024; Fig. 1B, one-sided P values,0.0001).
Effects of Gossypol and Cadherin Genotype on Larval
In both experiments, larval weight was lower on gossypol diet
than on control diet. Also, in both experiments, the reduction in
larval weight on gossypol diet relative to control diet was greater
for rs and rr larvae than for ss larvae. In experiment one on control
diet, weight was lower in rr (28.3 mg) than ss (30.5 mg) (t=5.04,
df=1028, P,0.0001), showing a fitness cost without gossypol. On
gossypol diet, weight was also lower in rr (24.9 mg) than ss
(28.8 mg) (t=7.81, df=1028, one-sided P,0.0001). The coeffi-
cient of the linear contrast with associated 95% confidence interval
was 20.037 (20.49, 20.25) on control diet and 20.67 (20.081,
20.053) on gossypol diet. As each coefficient lies outside the
confidence interval for the other, costs were significantly higher on
gossypol than control diet (P,0.05). Larval weight was generally
lower on gossypol diet than control diet (Fig 2A, t=26.17, one-
sided P,0.0001). Relative to ss, weight differed more between
gossypol diet and control diet in r1r1 (t=23.52, one-sided
P,0.00025), but not in the other genotypes (one-sided P
values.0.08) (Fig. 1A).
In experiment two on control diet, weight of rr (29.0 mg) did not
differ significantly from ss (29.9 mg) (t=0.80, df=482, P=0.42),
nor did weight of rs (29.0 mg) differ from ss (t=0.96, df=482,
P=0.34). Thus, no cost affecting weight was seen on control diet
in this experiment. On gossypol diet, however, weight was lower
Table 1. Mean larval weight (mg) and gossypol concentration
(mg/g dry weight) on gossypol and control diet in cadherin
genotypes from experiment one (MOV97-H1S and MOV97-
H1R) and experiment two (MOV97-H3).
Experiment 1: MOV97-H1S and MOV97-H1R
Controlss30.5 (0.3) 31101
r3r3 28.6 (0.5)11301
r1r3 27.9 (0.6)10001
r1r1 28.6 (0.5)6501
rr (averaged)28.3 (0.4) 303
Gossypol ss28.8 (0.3) 2320.56 (0.08) 25
r3r326.0 (0.5) 851.06 (0.14)25
r1r325.1 (0.6)802.09 (0.32)23
r1r123.7 (0.7) 504.87 (1.15)12
rr (averaged)24.9 (0.6)32.67 (1.14)3
Experiment 2: MOV97-H3
Control ss29.9 (0.7)6101
r3s27.3 (0.6) 8601
r3r3 28.1 (1.0)3301
r1r330.3 (0.9) 3201
r1r128.7 (1.0)24 01
rs (averaged) 29.0 (0.7)202
rr (averaged) 29.0 (0.9)303
Gossypol ss28.8 (0.8)560.41 (0.08)22
r3s25.7 (0.7) 771.43 (0.08) 31
r1s24.7 (1.1) 332.17 (0.16) 14
r3r3 25.4 (1.2)28 2.06 (0.24)12
r1r3 31.0 (1.5)21 3.52 (0.76)9
r1r1 20.2 (5.8)411.21 (3.80)2
rs (averaged) 25.2 (0.9)2 1.80 (0.37)2
rr (averaged) 25.5 (2.8)3 5.60 (2.84)3
1For each combination of genotype and diet, standard errors are reported in
2For each combination of genotype and diet, number of larvae (n) are reported.
3The means for rr are the average across r3r3, r1r3, and r1r1; for rs these are the
average across r3s and r1s.
Fitness Cost of Resistance to Bt Cotton
PLoS ONE | www.plosone.org2 June 2011 | Volume 6 | Issue 6 | e21863
for rr (25.5 mg) than ss (28.8 mg) (t=3.33, df=482, one-sided
P=0.00045), which indicates a cost. Furthermore, weight was
lower for rs (25.2 mg) than ss (t=3.93, df=482, one-sided
P,0.0001), but weight of rs and rr did not differ significantly
(t=0.53, df=482, P=0.59). Thus, gossypol induced a non-
recessive cost affecting larval weight. Larval weight was lower on
gossypol diet than control diet (Fig. 2B, t=25.11, one-sided
P,0.0001). Relative to ss, weight differed more between gossypol
diet and control diet in r1r1 (t=23.18, one-sided P=0.008) and
r1s (t=22.96, one-sided P=0.002), but not in the other genotypes
(one-sided P values.0.14) (Fig. 1B).
Pooling rr and rs genotypes across experiments, the mean cost
(6SE) affecting weight was significantly different from zero both
on control diet (5.361.7%; t=3.04, df=7, P=0.019) and
gossypol diet (12.963.7%; t=3.52, df=7, one-sided P=0.0048).
The increase in cost from control to gossypol diet was statistically
significant (paired t-test, t=1.98, df=7, one-sided P=0.044).
Across the six larval genotypes tested in experiments one and two,
a positive association occurred between gossypol content and the
decrease in larval weight on gossypol diet compared to control diet
(Fig. 3, t=3.23, df=8, one-sided P=0.0060). The association
between gossypol content and decreased weight was also
Figure 1. Gossypol concentration in pink bollworm cadherin genotypes. Mean gossypol concentration (6SE, log [x+0.001] transformed) in
larvae from (A) MOV97-H1S and MOV97-H1R and (B) MOV97-H3 fed on gossypol diet.
Fitness Cost of Resistance to Bt Cotton
PLoS ONE | www.plosone.org3June 2011 | Volume 6 | Issue 6 | e21863
significant when data were analyzed separately for experiment one
(P=0.0015) and two (P=0.046). This suggests that increased
gossypol concentration in larvae with one or two r alleles
compared to ss increased the cost affecting larval growth.
Effects of Gossypol and Cadherin Genotype on Survival
Similar to the results with larval weight, larval survival was
lower on gossypol diet than on control diet in both experiments. In
experiment one, survival was lower on gossypol diet (27.9%) than
on control diet (36.8%) (one-sided P,0.0001). On control diet,
survival of ss (38.9%) did not differ significantly from survival of rr
(34.8%) (Table 1, P=0.10). On gossypol diet, survival of ss (29.0%)
also did not differ significantly from survival of rr (26.9%) (one-
sided P=0.18). Thus, gossypol did not increase the survival cost.
The difference in genotype frequency on gossypol diet relative to
control diet did not differ significantly between ss and r1r1, r1r3, or
r3r3 (Fig. 4 A, one-sided P values.0.34), which suggests that
gossypol did not reduce survival of any of the rr genotypes relative
In experiment two, survival was lower on gossypol diet (27.4%)
than on control diet (34.4%) (one-sided P=0.0014). We could not
test directly for a fitness cost in experiment two (see Methods), but
we could determine if the frequency of rs and rr relative to ss was
reduced more on gossypol diet than on control diet. Compared to
Figure 2. Weight of pink bollworm cadherin genotypes on control and gossypol diet. Mean weight (6SE, log transformed) of larvae from
(A) MOV97-H1S and MOV97-H1R and (B) MOV97-H3.
Fitness Cost of Resistance to Bt Cotton
PLoS ONE | www.plosone.org4June 2011 | Volume 6 | Issue 6 | e21863
control diet, the frequency of rr (vs. ss) was marginally reduced on
gossypol diet (one-sided P=0.057), although the frequency of rs
remained similar on control and gossypol diet (one-sided P=0.45).
The reduction in frequency on gossypol diet relative to control diet
was significantly greater in r1r1 than ss (Fig. 4 B, x2=11.73, one-
sided P=0.0003), but not in other genotypes with r alleles (all one-
sided P values.0.15). This suggests that gossypol reduced survival
of r1r1 relative to ss.
We focused here on the effects of the cotton defensive
compound gossypol on fitness costs by feeding Bt-resistant and
susceptible pink bollworm larvae on diet without Bt toxin that
either contained or lacked gossypol. We found that when pink
bollworm larvae ate diet containing gossypol, the concentration of
gossypol was higher in larvae with cadherin alleles conferring
resistance to Cry1Ac than in susceptible larvae lacking such alleles
(Fig. 1). In both experiments, costs affecting larval growth rate, as
indicated by weight of 14-day-old larvae, were magnified on
gossypol diet (Table 1; Fig. 2). In experiment two, larval gossypol
concentration was non-recessive; it was higher in rs larvae than in
ss larvae (Fig. 1). In addition, in this experiment, costs affecting
growth were absent on control diet but dominant on gossypol diet
(Table 1). These results show that higher larval gossypol content
was associated with a non-recessive cost affecting larval growth.
This is potentially important because non-recessive costs select
more strongly against resistance than recessive costs [9–14].
When results from both experiments were considered, costs
affecting growth were significantly higher on gossypol diet (12.9%)
than on control diet (5.3%). Furthermore, a significant, positive
association across genotypes occurred between larval gossypol
concentration and the decrease in weight on gossypol diet relative
to control diet (Fig. 3). These results show that cadherin mutations
conferring resistance to Cry1Ac cotton in pink bollworm were
associated with higher larval gossypol concentrations and a higher
fitness cost affecting larval growth.
The strains MOV97-H1S and MOV97-H1R and MOV97-H3
had a common origin and contained similar r and s alleles, but
differed in their rearing history. Costs reducing larval growth on
control diet were present in MOV97-H1S and MOV97-H1R but
not in MOV97-H3 (Fig. 2). Also, relative to ss, the frequency of
r1r1 decreased more on gossypol than control diet in MOV97-H3
but not in MOV97-H1S and MOV97-H1R (Fig. 4). Such
differences in the response of the same genotypes from different
strains indicate that variation in genetic background affected costs.
The r1 allele has a deletion of 24 base pairs resulting in two amino
acid substitutions and the elimination of eight amino acids. The r3
allele has a deletion of 126 base pairs resulting in the omission of
42 amino acids . Across all of the strains studied, we found that
effects of gossypol in diet on larval gossypol concentration and
growth were greater for r1 than r3 (Fig. 1, 2 and 3).
The results reported here confirm results from a previous study
of fitness costs in three strains of pink bollworm derived from
MOV97: a susceptible strain, a Cry1Ac-resistant strain, and their
F1 progeny . In the previous study, larval gossypol
concentration was not measured and cadherin genotypes were
not identified. However, similar to the results reported here,
gossypol in diet significantly increased the cost affecting larval
developmental time, but did not increase the cost affecting survival
In general, costs of Bt resistance are significantly higher on
plants than on diet . Compared with the mean costs from
many studies , the cost affecting growth on control diet seen
here (5.3%) is virtually identical to the mean cost affecting
development time on diet (5.0%), while the cost affecting growth
on gossypol diet seen here (12.9%) is somewhat higher than the
mean cost affecting developmental time on plants (8.9%).
Although it was hypothesized that fitness costs are higher on
host plants with low suitability [26,27], available data show no
Figure 3. Association between gossypol concentration and difference in weight on gossypol diet relative to control diet. Closed
circles show data from MOV97-H1S and MOV97-H1R; open circles show data from MOV97-H3.
Fitness Cost of Resistance to Bt Cotton
PLoS ONE | www.plosone.org5June 2011 | Volume 6 | Issue 6 | e21863
consistent relationship between costs of Bt resistance and host
plant suitability . In Trichoplusia ni, costs were negatively
associated with host plant suitability, supporting this hypothesis
. In Plutella xylostella, however, the cost affecting development
time was greater on the least suitable host plant, but the survival
cost was higher on the most suitable host in one of the two P.
xylostella populations investigated . Furthermore in H. armigera,
cotton, pigeon pea, and sorghum were equally suitable to ss
individuals, but costs were larger or less recessive on sorghum and
cotton than pigeon pea . In P. gossypiella, survival was
significantly greater on the cotton cultivar DP50 with high
gossypol content than on TX53 with low gossypol content, but
the cost affecting survival was recessive and of similar magnitude
on both cultivars . High content of defensive chemicals other
than gossypol could have suppressed survival on TX53, however,
and these chemicals could have compensated for effects of high
gossypol content in DP50.
In the studies cited above, among-host variation in nutrient
availability and defensive chemicals was not well characterized,
and the molecular basis of resistance was known only in pink
bollworm. Resistance to Bt toxins is associated with mutations
affecting ABC transporter, aminopeptidase-N, and cadherin
proteins that act as toxin receptors in the midgut, as well as with
proteases that convert Bt protoxins to activated toxins [30,31].
Mutations affecting aminopeptidase-N and proteases, which are
two major digestive proteases, could impair protein digestion .
Mutations altering ABC transporters could contribute in reducing
export of toxic compounds by midgut epithelial cells to the gut
lumen , while cadherin mutations could contribute in
increasing concentrations of plant defensive chemicals in larvae,
as suggested by the results here. Accordingly, low availability of
nutrients could be the main factor increasing costs in insects with
aminopeptidase-N- and protease-mediated resistance, as it would
be harder to compensate for digestive deficiencies when nutrient
availability is low rather than high. In contrast, high concentration
of defensive chemicals could be the most important reason for an
increase in costs in insects with ABC transporter- and cadherin-
mediated resistance. Therefore, future studies on variation in costs
across host plants may benefit from a more mechanistic approach.
Functional studies of larvae with various cadherin genotypes are
needed to determine the effects of cadherin mutations on midgut
structure and permeability to plant defensive chemicals. To better
understand how host plants affect costs, it will also be necessary to
evaluate the relationships among defensive compounds, nutrient
availability, and fitness when Bt-susceptible and -resistant insects
develop on their natural host plants. Cotton was recently
genetically engineered to produce plants with little gossypol in
seeds but normal levels in stems and leaves . As pink bollworm
larvae primarily feed on seeds, these transgenic cultivars provide
an ideal system to further assess the interaction between cadherin-
based resistance to Bt cotton, absorption of defensive compounds
and expression of fitness costs.
Materials and Methods
We used three pink bollworm strains derived from strain
MOV97, which was established from a single collection in
Mohave Valley, Arizona in 1997 . At the locus encoding a
cadherin protein that binds Bt toxin Cry1Ac, MOV97 had two
alleles (r1 and r3) that confer recessive resistance to Bt toxin
Cry1Ac and transgenic cotton that produces Cry1Ac [16,34]. At
this locus, MOV97 also had s alleles that confer susceptibility to
Cry1Ac [16,34]. We conducted two independent experiments: In
experiment one, we used two strains that had a similar genetic
background, yet one was resistant (MOV97-H1R) and contained
only the r1 and r3 alleles, while the other was susceptible (MOV97-
H1S) and contained only s alleles. In experiment two, we used a
hybrid strain (MOV97-H3) in which the r1, r3 and s alleles
segregated at the cadherin locus. We refer to individuals with any
two r alleles (r1r1, r3r3, or r1r3) as rr, and to individuals with one r
allele (r1s and r3s) as rs.
MOV97-H1R, MOV97-H1S and MOV97-H3 were derived
from the hybrid strain MOV97-H1. MOV97-H1 was produced by
crossing a resistant strain (MOV97-R) and a susceptible strain
(MOV97-S) that had been derived from MOV97 . MOV97-
H1R was produced by exposing larvae of the F18 generation of
MOV97-H1 to a diagnostic concentration of Cry1Ac in synthetic
diet (10 mg toxin per ml diet) that allows survival of only rr
individuals [35,36]. MOV97-H1S was initiated by screening
mated pairs of MOV97-H1 (F17) with PCR to find pairs of ss
individuals . Ten mated pairs with either rr or ss individuals
were caged individually to initiate MOV97-H1R and MOV97-
H1S. From each of the 10 mated pairs, 12 pupae were collected
(120 per strain). The adult population size was 378 per strain in
the F2 generation and 1200 in the following generations .
MOV97-H3 was created by crossing insects from MOV97-H1
(F22), which were predominantly ss and rs, with a subset of insects
Figure 4. Frequency of cadherin genotypes on control diet
(grey bars) and gossypol diet (black bars). (A) MOV97-H1S and
MOV97-H1R and (B) MOV97-H3. For each type of diet, frequency for
each genotype was calculated as the number of survivors of that
genotype divided by total number of survivors.
Fitness Cost of Resistance to Bt Cotton
PLoS ONE | www.plosone.org6 June 2011 | Volume 6 | Issue 6 | e21863
from MOV97-H1 (F21) that had been selected with the diagnostic
concentration of Cry1Ac and thus were resistant . Adult
population size in MOV97-H1 and MOV97-H3 was.1000 in
Experiment one was conducted in September 2005 with F4
larvae of MOV97-H1S and MOV97-H1R. Experiment two was
done in November 2006 with F5 larvae of MOV97-H3. The
parental strain MOV97-H1 had been reared for 17–18 genera-
tions on non-Bt diet before MOV97-H1S and MOV97-H1R were
created, and 21–22 generations before MOV97-H3 was created.
Also, before experiments were done, each of these three strains
had been reared for either four or five generations on non-Bt diet.
This rearing reduced linkage disequilibrium between cadherin
alleles and other alleles that may affect fitness .
To investigate the effects of gossypol, newly hatched larvae were
placed on wheat germ diet  with gossypol (gossypol diet) or
without gossypol (control diet). For gossypol diet, we added
gossypol (95% in acetic acid crystal, Sigma-Aldrich) in 1 ml of
hexane to achieve a concentration of 0.1% fresh weight (0.1 g
gossypol/100g fresh weight) . This is equivalent to 0.6% dry
weight, a concentration within the range that occurs naturally in
seeds of Gossypium spp. [29,39,40]. For control diet, we added the
same amount of hexane without gossypol. We also made red diet
without gossypol by adding red calco dye (Bio-Serv, Frenchtown,
NJ, USA) to control diet at a concentration of 0.010% . We
used the red diet to monitor voiding of gossypol diet from the
larval gut (see below).
Effect of Gossypol Diet on Costs
We randomly assigned neonate larvae of the susceptible strain
(MOV97-H1S, n=800), the resistant strain (MOV97-H1R,
n=800) and the hybrid strain (MOV97-H3, n=400) to diet with
or without gossypol. We reared larvae in trays with 16 wells
(15 mm deep, 3 ml in volume). We put 1.5 g of diet and one
neonate in each well. Each tray was sealed with a transparent
cover with holes for ventilation. Trays containing one of the two
diet types were randomly allocated to shelves in a growth chamber
maintained at 2762uC with ambient relative humidity and a
photoperiod of 14:10 (L:D) h.
After feeding for 14 days, survivors from both diet types were
placed on red diet for 1.5 h to void their gut of the undyed diet.
Preliminary work showed that 1.5 h was sufficient to void the gut
of diet. After 1.5 h on red diet, each larva was scored for mortality
and weighed. The top half of the head capsule of each larva was
excised with a razor blade such that no hemolymph was lost, and
preserved in 100% EtOH for subsequent genotyping. Larvae were
then freeze dried with liquid nitrogen, placed in glass extraction
vials (KIMAX brand sample vials, borosilicate glass, with PTFE-
lined screw cap 4 ml), and stored at 280 C until gossypol was
extracted and quantified.
We extracted DNA from individual larvae using the protocol from
Morin et al.  with slight modifications . We determined larval
cadherin genotype using PCR with primer sets that selectively amplify
each of the four types of cadherin alleles (r1, r2, r3, and s) . As
expected, no r2 alleles were found in any of the larvae genotyped.
We measured gossypol concentration in individual larvae by
creating an aniline Schiff’s base and quantifying the resulting
dianilino-gossypol complex with high pressure chromatography
coupled with a triple quadrupole mass spectrometer . We used
an external calibration curve and an internal standard. We
assumed extraction efficiency was 100%. The lowest detectable
concentration was 0.025 mg/g. Values below this limit are
reported as zero.
In experiment one, a total of 89 larvae from MOV97-H1S and
MOV97-H1R were sent to Monsanto labs (Creve Coeur, MO) for
analysis. These comprised a random sample of 25 MOV97-H1S
larvae fed on gossypol diet, a random sample of 60 MOV97-H1R
larvae fed on gossypol diet, one MOV97-H1S larva fed on control
diet selected at random from ss larvae previously identified with
PCR, and one of each of r1r1, r1r3 and r3r3 from MOV97-H1R
fed on control diet and selected randomly from larvae previously
identified with PCR. In the second experiment, a total of 96
MOV97-H3 larvae were sent to Monsanto labs. These comprised
90 randomly selected larvae that had fed on gossypol diet, and one
larva from each genotype (i.e., ss, r1s, r3s, r1r1, r1r3 and r3r3) fed
on control diet and selected randomly from larvae identified
previously with PCR. Larvae were shipped overnight on dry ice
and remained frozen for the duration of the shipment. Gossypol
analyses were conducted without knowledge of genotype or
experimental diet, as the shipped samples were identified only
by a number.
We used statistical analyses to test our main hypotheses: 1) for
larvae fed gossypol diet, gossypol concentration is lower in ss larvae
than in rr and rs larvae; 2) gossypol reduces performance of all
larvae as indicated by lower larval weight or lower survival; 3) the
extent of reduction in larval performance on gossypol diet relative
to control diet is greater for rr and rs larvae than for ss larvae. As
each of these a priori hypotheses specifies the direction of the
difference between groups, we report one-tailed P values for all
tests of these hypotheses.
In each experiment, one-way ANOVA followed by linear
combinations of means (hereafter linear contrasts) were used to
assess whether gossypol concentration (mg/g dry weight, trans-
formed log x+0.001) differed among the ss, rs, and rr genotypes.
Multiple regression with indicator variables for each genotype with
r allele (s) was further used to compare gossypol concentration
between r1r1, r1r3, r1r1 and ss in experiment one and between r1s,
r3s, r1r1, r1r3, r3r3 and ss in experiment 2.
In each experiment, two-way ANOVA followed by linear
contrasts was used to evaluate whether larval weight (fresh weight
in mg, log transformed) differed among the ss, rs, and rr genotypes.
In experiment one, costs on control and gossypol diet were
compared by assessing whether each contrast coefficient (i.e., ss vs.
rr) lied outside the 95% confidence interval for the other .
Multiple regression with indicator variables for each genotypes
with r allele (s) and gossypol diet was further used to assess the
effects of diet (gossypol versus control diet), genotype, and the
interaction between these factors on larval weight (fresh weight in
mg, log transformed) in each experiment. A significant negative
coefficient associated with the interaction between a particular
genotype and diet indicated that larval weight of this genotype was
reduced more (relative to ss) on diet with gossypol than on control
We measured costs affecting larval weight of a genotype with r
alleles on a particular diet as: cost in %=100%6([mean weight
(genotype) – mean weight ss] / mean weight ss). As costs did not
differ between experiments on either diet (2-sample t-test, P
values.0.15), we pooled genotypes from both experiment and
used a one-sample t-test to assess whether costs were significantly
Fitness Cost of Resistance to Bt Cotton
PLoS ONE | www.plosone.org7June 2011 | Volume 6 | Issue 6 | e21863
different from zero on control and gossypol diet. We also used a
paired t-test to evaluate whether costs were significantly greater on
gossypol than control diet.
For each genotype, we calculated the percentage reduction in
weight from control to gossypol diet as: 100%6([mean weight on
control diet – mean weight on gossypol diet]/mean weight on
control diet). We used a covariance analysis to test for effects of
experiment, gossypol concentration in a genotype, and the
interaction between these factors on the change in weight from
control to gossypol diet. Because the effects of experiment
(P=0.61) and interaction (P=0.93) were not significant, we
pooled data and used linear regression to assess the association
between gossypol content and change in weight from control to
gossypol diet. We also used linear regression to evaluate the
association between gossypol content and change in weight from
control to gossypol diet separately for experiment one and two.
In both experiments, we used a Fisher’s exact test to evaluate
whether the proportion of larval survival differed between gossypol
and control diet. To test for fitness costs affecting survival in
experiment one, we pooled the rr genotypes and used a Fisher’s
exact test to determine if the proportion of survival of ss and rr
differed on gossypol or control diet. We could not measure survival
of genotypes in experiment two because we did not know the
initial number of individuals of each genotype, only the number of
survivors of each genotype. Nevertheless, in both experiments, we
assessed whether the frequency of rr or rs decline more from
control to gossypol diet than the frequency of ss. In experiment
two, we used a Fisher’s exact test to compare the frequency of rr
(three genotypes pooled) and rs (two genotypes pooled) to the
frequency of ss on each diet (gossypol and control). In both
experiments, we also used log-linear regression to assess whether
frequency decreased more on gossypol than control diet in each
genotype with r allele (s) than is ss. Explanatory variables in the
regression model were indicator variables for diet, genotype, and
the interaction between these factors, while the response variable
was the number of individuals of each genotype surviving on the
diet types. A significant negative coefficient associated with the
interaction between a particular genotype and diet indicated that
frequency of this genotype was reduced more (relative to ss) on diet
with gossypol than on control diet . Statistical analyses were
performed in JMP .
Conceived and designed the experiments: JW YC. Performed the
experiments: JW CE-K RO AG. Analyzed the data: JW BT YC.
Contributed reagents/materials/analysis tools: RO GH BT YC. Wrote the
paper: BT YC.
1. Hutchison WD, Burkness EC, Moon RD, Leslie T, Fleischer SJ, et al. (2010)
Areawide suppression of European corn borer with Bt maize reaps savings to
non-Bt maize growers. Science 330: 222–225.
2. Committee on the Impact of Biotechnology on Farm-Level Economics and
Sustainability, National Research Council (2010) The impact of genetically
engineered crops on farm sustainability in the United States. Washington, DC:
The National Academies Press. 235 p.
3. Tabashnik BE, Gassmann AJ, Crowder DW, Carrie `re Y (2008) Insect resistance
to Bt crops: evidence versus theory. Nat Biotechnol 26: 199–202.
4. Tabashnik BE, Van Rensburg JBJ, Carrie `re Y (2009) Field-evolved insect
resistance to Bt crops: definition, theory and data. J Econ Entomol 102:
5. Bagla P (2010) Hardy cotton-munching pests are latest blow to GM crops.
Science 327: 1439.
6. Downes S, Parker T, Mahon R (2010) Incipient resistance of Helicoverpa punctigera
to the Cry2Ab Bt toxin in Bollgard II cotton. PloS ONE 5(9): e12567.
7. Storer NP, Babcock JM, Schlenz M, Meade T, Thompson GD, et al. (2010)
Discovery and characterization of field resistance to Bt maize: Spodoptera frugiperda
(Lepidoptera: Noctuidae) in Puerto Rico. J Econ Entomol 103: 1031–1038.
8. Onstad DW (2008) Insect resistance management: Biology, economics, and
prediction San Diego: Elselvier. 305 p.
9. Carrie `re Y, Crowder DW, Tabashnik BE (2010) Evolutionary ecology of
adaptation to Bt crops. Evol Appl 3: 561–573.
10. Gassmann AJ, Carrie `re Y, Tabashnik BE (2009) Fitness costs of insect resistance
to Bacillus thuringiensis. Annu Rev Entomol 54: 147–163.
11. Carrie `re Y, Tabashnik BE (2001) Reversing insect adaptation to transgenic
insecticidal plants. Proc R Soc Lond B 268: 1475–1480.
12. Tabashnik BE, Dennehey TJ, Carrie `re Y (2005) Delayed resistance to trangenic
cotton in pink bollworm. Proc Natl Acad Sci 43: 15389–15393.
13. Gould F, Cohen MB, Bentur JS, Kennedy GG, Van Duyn J (2006) Impact of
small fitness costs on pest adaptation to crop varieties with multiple toxins: A
heuristic model. J Econ Entomol 99: 2091–2099.
14. Crowder DW, Carrie `re Y (2009) Comparing the refuge strategy for managing
the evolution of insect resistance under different reproductive strategies. J Theor
Biol 261: 423–430.
15. Gahan LJ, Gould F, Heckel DG (2001) Identification of a gene associated with
Bt resistance in Heliothis virescens. Science 293: 857–860.
16. Morin S, Biggs RW, Sisterson MS, Shriver L, Ellers-Kirk C, et al. (2003) Three
cadherin alleles associated with resistance to Bacillus thuringiensis in pink
bollworm. Proc Natl Acad Sci 100: 5004–5009.
17. Xu XJ, Yu LY, Wu YD (2005) Disruption of a cadherin gene associated with
resistance to Cry1Ac delta-endotoxin of Bacillus thuringiensis in Helicoverpa armigera.
Appl Environ Microbiol 71: 948–954.
18. Fabrick JA, Tabashnik BE (2007) Binding of Bacillus thuringiensis toxin Cry1Ac to
multiple sites of cadherin in pink bollworm. Insect Biochem Mol Biol 37: 97–106.
19. Schlichting K, Wilsch-Bra ¨uninger, Demontis F, Dahmann C (2006) Cadherin
Cad99C is required for normal microvilli morphology in Drosophila follicle cells.
J Cell Sci 119: 1184–1195.
20. Hill E, Broadbent ID, Chothia C, Pettitt J (2001) Cadherin superfamily proteins
in Caenorhabditis elegans and Drosophila melanogaster. J Mol Biol 305: 1011–1024.
21. Midboe EG, Mehmet C, Bulla LA Jr. (2003) Expression of a midgut-specific
cadherin BT-R1 during the development of Manduca sexta larva. Comp Biochem
Physiol B 135: 125–137.
22. Carrie `re Y, Showalter AM, Fabrick JA, Sollome J, Ellers-Kirk C, et al. (2009)
Cadherin gene expression and effects of Bt resistance on sperm transfer in pink
bollworm. J Insect Physiol 55: 1058–1064.
23. Carrie `re Y, Ellers-Kirk C, Biggs R, Dennehy TJ, Tabashnik BE (2004) Effects of
gossypol on fitness costs associated with resistance to Bt cotton in pink bollworm.
J Econ Entomol 97: 1710–1718.
24. Mao YB, Cai WJ, Wang JW, Hong GJ, Tao XY, et al. (2007) Silencing a cotton
bollworm P450 monooxygenase gene by plant-mediated RNAi impairs larval
tolerance of gossypol. Nat Biotechnol 25: 1307–1313.
25. Stipanovic RD, Lo ´pez JD, Dowd MK, Puckhaber LS, Duke SE (2008) Effect of
racemic (+)- and (–) –gossypol on survival and development of Heliothis virescens
larvae. Physiol Ecol 37: 1081–1085.
26. Janmaat AF, Myers J (2005) The cost of resistance to Bacillus thuringiensis varies
with the host plant of Trichoplusia ni. Proc Roy Soc B 272: 1031–8.
27. Raymond B, Sayyed AH, Wright DJ (2007) Host plant and population
determine the fitness cost to Bacillus thuringiensis. Biol Lett 3: 82–5.
28. Bird LJ, Akhurst RJ (2007) Effects of host plant species on fitness costs of Bt
resistance in Helicoverpa armigera (Lepidoptera:Noctuidae). Biol Contr 40:
29. Carrie `re Y, Ellers-Kirk C, Biggs R, Degain B, Holley D, et al. (2005) Effects of
cotton cultivar on fitness costs associated with resistance of pink bollworm
(Lepidopters: Gelechiidae) to Bt cotton. J Econ Entomol 98: 947–54.
30. Bravo A, Sobero ´n M (2008) How to cope with insect resistance to Bt toxins?
Trends Biotechnol 26: 573–579.
31. Gahan LJ, Pauchet Y, Vogel H, Heckel DG (2010) An ABC transporter
mutation is correlated with insect resistance to Bacillus thuringiensis Cry1Ac toxin.
PLoS Genet 6(12): e1001248. doi:10.1371/journal.pgen.1001248.
32. Sunilkumar G, Campbell LM, Puckhaber L, Stipanovic RD, Rathore KS (2006)
Engineering cottonseed for use in human nutrition by tissue-specific reduction of
toxic gossypol. Proc Natl Acad Sci 103: 18054–18059.
33. Tabashnik BE, Patin AL, Dennehy TJ, Liu YB, Carrie `re Y, et al. (2000)
Frequency of resistance to Bacillus thuringiensis in field populations of pink
bollworm. Proc Natl Acad Sci 97: 12980–12984.
34. Carrie `re Y, Ellers Kirk C, Biggs RW, Nyboer ME, Unnithan GC, et al. (2006)
Cadherin-based resistance to Bt cotton in hybrid strains of pink bollworm: fitness
costs and incomplete resistance. J Econ Entomol 99: 1925–1935.
35. Tabashnik BE, Biggs RW, Higginson DM, Unnithan DC, Unnithan GC, et al.
(2005) Association between resistance to Bt cotton and cadherin genotype in
pink bollworm. J Econ Entomol 98: 635–644.
36. Gassmann AJ, Stock SP, Carrie `re Y, Tabashnik BE (2006) Effect of
entomopathogenic nematodes on the fitness cost of resistance to Bt toxin
Cry1Ac in pink bollworm (Lepidoptera: Gelechiidae). J Econ Entomol 99:
Fitness Cost of Resistance to Bt Cotton
PLoS ONE | www.plosone.org8June 2011 | Volume 6 | Issue 6 | e21863
37. Falconer DS (1981) Introduction to quantitative genetics. New York: Longman.
38. Adkinsson PL, Vaderzant ES, Bull DL, Allision WE (1960) A wheat germ
medium for rearing the pink bollworm. J Econ Entomol 53: 759–762.
39. Shaver TN, Parrott WL (1970) Relationship of larval age to toxicity of gossypol
to bollworms, tobacco budworms, and pink bollworms. J Econ Entomol 63:
40. Hron RJ, Kim HL, Calhoun MC, Fisher GS (1999) Determination of (+)-, (–)-,
and total gossypol in cottonseed by high-performance liquid chromatography.
J Am Oil Chem Soc 76: 1351–1355.
41. Graham HM, Mangum CL (1971) Larval diets containing dyes for tagging pink
bollworm moths internally. J Econ Entomol 64: 376–379.
42. Higginson DM, Nyboer ME, Biggs RW, Morin S, Tabashnik BE, et al. (2005)
Evolutionary trade-offs of insect resistance to Bacillus thuringiensis crops: fitness
costs affecting paternity. Evolution 59: 915–920.
43. Morin S, Carrie `re Y, Dennehy TJ, Brown JK, Tabashnik BE (2004) DNA-based
detection of Bt resistance alleles in pink bollworm. Insect Biochem Mol Biol 34:
44. Orth RG, Head G, Mierkowski M (2007) Determining larval host plant use by a
polyphagous lepidopteran through analysis of adult moths for plant secondary
metabolites. J Chem Ecol 33: 1131–1148.
45. Ramsey FL, Schafer DW (2002) The statistical sleuth: a course in methods of
data analysis. Pacific Grove: Duxbury. 742 p.
46. SAS (2008) JMP 8.0. Cary: SAS Institute Inc..
Fitness Cost of Resistance to Bt Cotton
PLoS ONE | www.plosone.org9June 2011 | Volume 6 | Issue 6 | e21863