By Bruce E. Tabashnik, Aaron J. Gassmann

Genetically engineered crops that produce Bacillus thuringiensis (Bt) toxins
kill some key insect pests and can help to reduce reliance on insecticide
sprays.
Bt crops have grown on more than 200 million ha worldwide since their
commercial introduction in 19961. This widespread use has raised two
pressing questions: "How quickly will insects evolve resistance to Bt
toxins?" and "Will resistance be delayed by planting refuges of non-Bt crops
near Bt crops?" Here we summarize our recent paper that addresses these
questions by analyzing global resistance monitoring data in conjunction with
results from computer simulations of the refuge strategy2.

In brief, most insect pests targeted by Bt crops in Australia, China, Spain,
and the U.S. did not evolve resistance during the first decade the crops
were grown. An exception is the bollworm Helicoverpa zea. Field-evolved
resistance to Cry1Ac, the Bt toxin in first generation transgenic cotton,
was initially documented in some H. zea populations in the southeastern U.S.
during 2003 and 2004, after seven to eight years of exposure to Bt cotton.
Overall, extensive monitoring data for H. zea and five other pests are
consistent with expectations arising from the theory underlying the refuge
strategy, suggesting that non-Bt crop refuges help delay insect resistance
to Bt crops.

We focused on the first generation of Bt crops, which consists almost
entirely of transgenic cotton producing Bt toxin Cry1Ac and transgenic corn
producing Bt toxin Cry1Ab. In nature, both toxins are produced in
crystalline form by Bt bacteria (hence their names start with "Cry").
Initial efficacy of first generation Bt crops against larvae of major
lepidopteran pests was high (>99%) for European corn borer (Ostrinia
nubilalis), tobacco budworm (Heliothis virescens), and pink bollworm
(Pectinophora gossypiella); intermediate (<99%) for bollworm (Helicoverpa
zea) and cotton bollworm (Helicoverpa armigera); and low (<20%) for some
other species such as beet armyworm (Spodoptera exigua). These differences
reflect inherent variation among species in their susceptibility to Cry1Ab
and Cry1Ac rather than evolution of resistance, which is defined below.

Field-evolved resistance to Bt toxins in sprays and transgenic crops

Evolution of resistance to a Bt toxin reflects a genetically-based decrease
in the susceptibility of an insect population to the toxin3. This results
from an increase in the frequency of individuals that have alleles
conferring resistance, which occurs over time when populations are exposed
to the toxin. Many insects harbor genetic variation in their susceptibility
to Bt toxins; more than a dozen species have been selected for Bt resistance
in the laboratory. To document field-evolved resistance, a field population
exposed to one or more Bt toxins must show less susceptibility than
conspecific field populations or lab strains with less exposure to the
toxins.

Susceptibility is usually measured with lab bioassays that expose larvae to
Bt toxins in their food. The most common index of susceptibility is the
LC50, which is the toxin concentration that kills 50% of larvae. The
resistance ratio, which is the LC50 of a field-derived strain divided by the
LC50 of a standard susceptible strain, is often used to gauge resistance.
The higher the resistance ratio, the greater the resistance. Resistance
ratios >10 are most likely to reflect genetically-based decreases in
susceptibility3.

Evolution of resistance to Bt toxins used in sprays has been documented for
field populations of diamondback moth (Plutella xylostella) and greenhouse
populations of cabbage looper (Trichoplusia ni). In these two vegetable
pests, initial documentation of resistance was based on a maximum resistance
ratio of 36 for diamondback moth in Hawaii and 160 for cabbage looper in
British Columbia.

Several experiments with H. zea show that the increased LC50s of Cry1Ac in
lab bioassays are linked with higher survival on plant tissues of Bt cotton
producing Cry1Ac. In one set of experiments by Jackson et al.4, survival on
Bt cotton relative to non-Bt cotton was 10% for a susceptible strain vs. 40%
for a lab-selected strain with a resistance ratio of 100. In independent
experiments with field-selected resistant strains, Luttrell and colleagues
report similar results and conclude that reduced susceptibility to Cry1Ac in
bioassays was "associated with a measurable increase in survival on Bt plant
tissue," and "Colonies collected as surviving larvae on Bt cotton tended to
have reduced susceptibility suggesting that some component of observed field
control problems may be associated with the presence of resistance genes."5
In contrast to the resistance documented for some field populations of H.
zea, similar monitoring efforts have not detected resistance in five other
major pests targeted by Bt crops: H. armigera, H. virescens, O. nubilalis,
P. gossypiella, and Sesamia nonagrioides2.

Field monitoring data vs. predictions from the refuge theory

To determine if the field outcomes documented by monitoring data are
consistent with the theory underlying the refuge strategy, we modeled
resistance evolution in each of the six major pests listed above2. The
refuge strategy is based on population genetics theory positing that refuges
of non-Bt host plants delay evolution of resistance by allowing survival of
susceptible insects. The refuge strategy is mandated in the U.S. and many
other countries where Bt crops are grown. This strategy is expected to be
especially effective when resistance is inherited as a recessive trait and
most resistant adults surviving on Bt crops mate with susceptible adults
from refuges. Under these conditions, the frequency of resistance is not
expected to increase rapidly because Bt crops kill the hybrid progeny
produced by matings between resistant and susceptible adults.

Consistent with the field data, modeling results projected that H. zea would
evolve resistance faster than other pests, primarily because its resistance
to Cry1Ac is dominant rather than recessive. Modeling results also showed
that resistance is expected to evolve faster as refuges constitute a smaller
percentage of the pests' host plants. This projection is consistent with
field data showing that H. zea resistance to Cry1Ac evolved faster in states
with lower refuge percentages.

Implications and conclusions

Even though large, genetically-based decreases in susceptibility to Cry1Ac
are well documented for some field populations of H. zea and increased
control problems have been noted anecdotally6, widespread control failures
have not been reported. We think that several factors contribute to this
pattern. First, even in the few states with documented cases of resistance,
most populations are not resistant. Second, data from greenhouse experiments
suggest that Cry1Ac in Bt cotton kills some resistant larvae, e.g., 60% of
larvae in a strain with a resistance ratio of 1004. Third, insecticide
sprays have been used extensively to control H. zea since the introduction
of Bt cotton. Control achieved with insecticides would mask problems
resulting from resistance to Cry1Ac. Finally, "pyramided" Bt cotton
producing Bt toxins Cry1Ac and Cry2Ab was registered in December 2002 and
planted on >1 million ha in the U.S. in 2006 and 2007. Control of
Cry1Ac-resistant H. zea larvae by Cry2Ab also limits control problems
associated with resistance to Cry1Ac.

The negative effects of resistance to Cry1Ac should decline further as the
acreage of cotton producing only Cry1Ac decreases. This acreage decreased
from 2.5 million ha in 2006 to only 1.3 million ha in 2007. In addition,
Monsanto's registration for Bt cotton with only Cry1Ac is scheduled to
expire in 2009. Cotton producing both Cry1Ac and Cry2Ab toxins could
substantially delay evolution of resistance for pests like H. virescens that
remain susceptible to both toxins. For Cry1Ac-resistant populations of H.
zea, however, the resistance-delaying benefits of pyramiding these two
toxins may not be fully realized.

In general, the second generation of crops genetically engineered for
protection against insects offers a greatly increased diversity of toxins.
The first generation of Bt crops was dominated by plants producing either
Cry1Ab or Cry1Ac, two closely related toxins that kill only caterpillars.
The U.S. EPA website of registered plant-incorporated protectants now lists
commercially available varieties of Bt corn and Bt cotton with 12 different
combinations of one to three Cry toxins that kill caterpillars, beetles, or
both. In the near future, more extensive pyramiding is likely, with plans
for up to six different Bt Cry proteins in single corn plants. Registration
is also expected for transgenic varieties producing another type of Bt toxin
called Vip (vegetative insecticidal protein). Other options for genetically
engineered insect protection include modified Bt toxins specifically
designed to kill insects resistant to native Bt toxins and gene-silencing
technology based on RNA interference.

Looking back on the first generation of Bt crops, we think that their
sustained efficacy against nearly all targeted pest populations exceeds the
expectations of many scientists. An exceptional case involving field-evolved
resistance to Cry1Ac in some populations of H. zea is consistent with the
theory underlying the refuge strategy because this resistance is not
recessive. In other words, the concentration of Cry1Ac in Bt cotton is not
high enough to kill the hybrid offspring produced by matings between
Cry1Ac-susceptible and -resistant adults. In reference to this concept, the
refuge strategy is sometimes called the "high-dose refuge strategy." Before
Bt cotton was commercialized, scientists at the U.S. EPA and elsewhere
reported that the high-dose criterion of the refuge strategy was not met for
Bt cotton with Cry1Ac vs. H. zea. Therefore, the relatively rapid resistance
that occurred in this pest is no surprise. As the second generation of Bt
crops proceeds, we can use systematic analyses of monitoring data from the
first decade to maximize benefits and minimize risks. The results summarized
here suggest that refuges can delay pest resistance to Bt crops, especially
when resistance is recessive and refuges are abundant.

Acknowledgment

This work was supported by NRI, CSREES, USDA grant 2006-35302-17365.

References
1. James C. Global status of commercialized biotech/GM crops: 2007. ISAAA
Brief No. 37, International Service for the Acquisition of Agri-Biotech
Applications, Ithaca, NY, USA (2007)

2. Tabashnik BE, Gassmann AJ, Crowder DW, Carri?re Y. 2008. Insect
resistance to Bt crops: evidence versus theory. Nat. Biotech. 26, 199-202

3. Tabashnik BE. 1994. Evolution of resistance to Bacillus thuringiensis.
Annu. Rev. Entomol. 39, 47-94

4. Jackson RE, Bradley JR Jr., Van Duyn JW. 2004. Performance of feral and
Cry1Ac-selected Helicoverpa zea (Lepidoptera: Noctuidae) strains on
transgenic cottons expressing either one or two Bacillus thuringiensis ssp.
kurstaki proteins under greenhouse conditions. J. Entomol. Sci. 39, 46-55

5. Luttrell RG, Ali MI. 2007. Exploring selection for Bt resistance in
Heliothines: results of laboratory and field studies, in Proceedings of the
2007 Beltwide Cotton Conferences, New Orleans, LA, January 9?12, 2007,
1073-1086 (National Cotton Council of America, Memphis, TN; 2007).

6. James, L. Bollworms feeding on Bt cotton in Arkansas. Delta Farm Press.
(2006) [deltafarmpress.com]

7. Tabashnik BE, Cushing NL, Finson N, Johnson MW. 1990. Field development
of resistance to Bacillus thuringiensis in diamondback moth (Lepidoptera:
Plutellidae). J. Econ. Entomol. 83, 1671-1676

8. Janmaat AF, Myers JH. 2003. Rapid evolution and the cost of resistance to
Bacillus thuringiensis in greenhouse populations of cabbage loopers,
Trichoplusia ni. Proc. Roy. Soc. B 270, 2263-2270

9. Luttrell RG, Wan L, Knighten K. 1999. Variation in susceptibility of
Noctuid (Lepidoptera) larvae attacking cotton and soybean to purified
endotoxin proteins and commercial formulations of Bacillus thuringiensis. J.
Econ. Entomol. 92, 21-32

10. Luttrell RG et al. 2004. Resistance to Bt in Arkansas populations of
cotton bollworm, in Proceedings of the 2004 Beltwide Cotton Conferences, San
Antonio, TX, January 5-9, 2004 (ed. Richter, D.A.) 1373-1383 (National
Cotton Council of America, Memphis, TN; 2004)

11. Ali MI, Luttrell RG, Young SY. 2006. Susceptibilities of Helicoverpa zea
and Heliothis virescens (Lepidoptera: Noctuidae) populations to Cry1Ac
insecticidal protein. J. Econ. Entomol. 99, 164-175

12. Ali MI et al. 2007. Monitoring Bt susceptibilities in Helicoverpa zea
and Heliothis virescens: results of 2006 studies, in Proceedings of the 2007
Beltwide Cotton Conferences, New Orleans, LA, January 9?12, 2007, 1062-1072
(National Cotton Council of America, Memphis, TN; 2007)

CONTACT:
Bruce E. Tabashnik1, Aaron J. Gassmann2, David W. Crowder1 and Yves
Carri?re1

1Department of Entomology, University of Arizona, Tucson, Arizona 85721 USA

2 Department of Entomology, Iowa State University, Ames. Iowa 5001

www.ag.arizona.edu
www.checkbiotech.org