1. Introduction
Transgenic crops that express insecticidal toxins from the bacterium
Bacillus thuringiensis (Berliner) (Bt) provide valuable pest management options for pests of field crops worldwide [
1,
2,
3]. Positive socio-economic and environmental impacts of Bt crop adoption have been reported since commercial release in 1996 [
4,
5]. In the United States (U.S.), the Bt technology provides control of the major cotton pests
Chloridea virescens (Fabricius) (Lepidoptera: Noctuidae),
Pectinophora gossypiella (Saunders) (Lepidoptera: Gelichiidae), and
Helicoverpa zea (Boddie) (Lepidoptera: Noctuidae) [
6,
7]. Bt cotton also improves the management of other lepidopteran pests, such as
Spodoptera exigua (Hübner),
Trichoplusia ni (Hübner),
Spodoptera frugiperda (J.E. Smith),
S. eridania (Stoll), and
Chrysodeixis includens (Walker) [
8,
9,
10,
11]. From 1996–2003, commercial Bt cotton in the U.S. was limited to events that expressed the Cry1Ac toxin [
12]. Second generation dual-gene Bt cotton that expressed the toxins Cry1Ac+Cry2Ab, Cry1Ac+Cry1F, and Cry1Ab+Cry2Ae became available in 2003 and were widely adopted [
7,
13]. The goal of these second-generation cotton events is to reduce the risk of resistance evolution by targeting unique and independent target sites. Since 2014 the third generation of cotton events with triple Bt traits became available, including those expressing Cry1Ac+Cry1F+Vip3Aa19, Cry1Ac+Cry2Ab+Vip3Aa19, and Cry1Ab+Cry2Ae+Vip3Aa19. The Cry and Vip families are produced during different stages of the
B. thuringiensis life cycle [
14]. Vip3A shares no sequence homology with any known Bt Cry toxins. The amount of shared sequence homology between Bt toxins is an important indicator of the risk that the two Bt toxins will share binding sites on the midgut of the insect, predisposing the toxins to cross-resistance. Although the two toxin classes are thought to have a similar mode of action against the target insects, they have different receptors in the insect midgut [
14,
15].
The tobacco budworm,
C. virescens, is highly susceptible to most commercial Cry Bt toxins, and Bt cotton usually reaches a high-dose condition, killing almost all heterozygotes for Bt resistance [
8,
16], even for single toxin events. However, the cotton bollworm,
H. zea is less susceptible to Cry toxins expressed in cotton and corn, which, therefore, do not satisfy high-dose criteria [
17]. Toxicological bioassays performed with populations of
H. zea from the southeastern U.S., including populations from the Florida Panhandle, have indicated a decrease in susceptibility of
H. zea populations to Cry1Ab, Cry1Ac, Cry1A.105, and Cry2Ab, but not to Vip3Aa [
5,
18,
19,
20]. However, a major resistance allele conferring high levels of Vip3Aa resistance in a field-derived strain of
H. zea in Texas has been recently reported [
21].
Field-evolved resistance in target pests is a threat to the success of the Bt technology, which may lead to control failures [
22] and the need for traditional insecticides for supplemental control [
23].
Helicoverpa zea exhibits a sequence of host crop utilization based on the temporal dynamics of the southeastern U.S. agricultural landscape. In general, the first generation of this pest feeds on Bt field corn, cultivated during the spring to the beginning of summer. Corn is a major host plant of
H. zea. However, this pest has a low impact on the yield of field corn when it is planted early in the season [
24]. Later in the summer, subsequent generations of
H. zea disperse from corn to cotton, which becomes the prevalent Bt crop in the southeastern U.S. agricultural landscape until the end of the crop season. Thus, corn serves as a source of
H. zea populations, and if these source populations develop on Bt corn, a dispersion of Bt pre-exposed survivors from corn to cotton is likely. Larval feeding in both crops producing the very same or similar Bt toxin exerts continuous selection pressure and raises concerns about the selection of resistant populations [
13].
The United States Environmental Protection Agency (EPA) has listed the knowledge of pest biology and ecology as key elements in formulating a Bt insect resistance management (IRM) programs [
25]. While complete studies documenting the reproductive potential of
H. zea populations feeding on dual- or triple-toxin Bt cotton are scarce, Cry1Ab Bt corn, for which is only moderately toxic, might reduce
H. zea growth potential [
26,
27]. Developing life tables for
H. zea on dual and triple-gene Bt cotton events allow the documentation of key fitness parameters, such as survival, development time, fertility, and population growth [
28]. Fitness components and rates of population growth (i.e., demographic performance), if available, contribute to developing models to predict the rate of resistance evolution in target pests or to comparatively assess different resistance management practices, such as the use of structured and/or natural refuges [
22,
29,
30,
31].
Helicoverpa zea populations from the Florida Panhandle represent valuable resources for documenting the fitness components and demographic performance of lepidopteran pests targeted by Bt toxins in cotton. The region is in the Gulf Coastal Plain of the southeastern U.S., an ecological transition zone between temperate and subtropical climates. In the region,
H. zea populations can overwinter and disperse throughout the growing season [
32,
33]. The overwintering survival of
H. zea could be a carry-over source of Bt resistance alleles for other regions and between seasons [
13]. Furthermore, the region has a distinctive regional landscape consisting of natural vegetation, forests, and field crops. Cotton is cultivated in large areas (approximately 50,000 hectares), with 100% adoption of Bt cultivars [
34,
35]. In this study, we report individual and population fitness of representative
H. zea larvae challenged with dual- and triple-toxin Bt cotton technologies, information that contributes to the development and validation of resistance management recommendations.
3. Discussion
In this study, the life-history traits and demographic performance of
H. zea from the Florida Panhandle varied among cotton cultivars and field populations, indicating differences among the cultivars in the efficacy against
H. zea and the current population susceptibility to the Bt toxins. Gassmann et al. (2009) suggest that survival, developmental time, and body weight are key individual fitness components [
30]. Here, the effects on immature insect fitness components associated with both population and cotton cultivar translated to negative effects on the growth potential of
H. zea. Although the non-Bt cotton cultivar used as control is not isoline of the Bt cultivars, the differences in the effects on
H. zea life-history between the Bt and non-Bt cotton documented here are likely associated with the expression of Bt toxins in each cotton cultivar tissue. Cotton plants are rich in terpenoid compounds, which may function as a barrier against herbivores impairing growth/development and/or behavioral traits. However, the cotton plant has been modified during domestication and breeding for high yield and quality, including low gossypol oil in cottonseeds, which may have lessened the content of anti-herbivory secondary metabolites (such as gossypol) [
36,
37,
38]. Information on secondary compounds in the cotton cultivars used in the present study was not available. However, other studies comparing non-Bt cotton cultivars have shown low or no change in noctuid life-history [
39].
The triple-gene Bt cotton that expressed Vip3Aa (Cry1Ac+Cry1F+Vip3Aa19, Cry1Ac+Cry2Ab+Vip3Aa19, or Cry1Ab+Cry2Ae+Vip3Aa19) caused 100% larval mortality in all
H. zea populations, which reinforces the high efficacy of this toxin for
H. zea control [
18,
22]. These results suggest a low frequency of resistant alleles to Vip3A in
H. zea populations tested. Data from laboratory and field in the U.S. consistently indicate high efficacy of the Vip3A against
H. zea [
6,
40,
41], and the debate has been focused on whether the trait meets the high dose definition [
42]. Due to the relatively recent adoption of Vip3A toxins in commercial cultivars and limited insect sample size (40-130 individuals) in the present study, it would be unlikely to detect resistance to Vip3A at its current low frequency in the field. Vip toxins show limited amino-acid sequence homology with Cry toxins and cause pore formation with unique properties, thus, having a low risk for cross-resistance between them [
14,
43,
44]. Cry toxins co-expressed in some Bt cotton cultivars have reportedly low impact in some
H. zea populations [
19,
45], which compromises the pyramid of Bt genes. Our study demonstrates that life-history traits (body weight, survival, and development time) of insects from Escambia, Santa Rosa, and Jackson populations were negatively affected by the dual-gene Bt cotton cultivars that expressed Cry toxins.
Life-history traits were more affected by Cry1Ab+Cry2Ae and Cry1Ac+Cry2Ab, while Cry1Ac+Cry1F caused fewer negative impacts given the relative lack of toxicity that Cry1F has on the
H. zea larvae and the widespread Cry1Ac resistance [
19]. Significant mortality from Cry1Ac+Cry1F was observed only in the Santa Rosa population. Cry toxins have been expressed in Bt cotton cultivars since its first commercial release, and the first report of
H. zea Cry1Ac resistance was documented 15 years later in the U.S., and recently the widespread resistance to Cry2Ab [
6,
20,
46]. These may be a contributing factor to the considerable rates of larval survival on Cry1Ac+Cry2Ab and Cry1Ab+Cry2Ae cultivars. It also confirms the Cry1A and possibly Cry2A resistance alleles occurrence at high frequencies in the
H. zea populations tested, which does not mean that the dual gene Bt cotton cultivars lost the benefit on
H. zea management completely. Overall, our data on life-history traits (survival rates, body weight, developmental time) are consistent with previous reports that Cry1Ac+Cry1F affects
H. zea larvae less than Cry1Ac+Cry2Ab, Cry1Ab+Cry2Ae, and cultivars expressing Vip3Aa [
5,
7,
19].
Sublethal effects of Bt toxins on
H. zea, as indicated by reduced body weight and the prolonged larval development, may have implications for pest management. Delayed larval development and low body weight are expected to increase the likelihood of exposure to other mortality factors. For example, early-instar larvae are unable to bore into the cotton bolls [
47]. Consequently, they may be more exposed to insecticide applications and vulnerable to natural enemies [
48]. Moreover, slow larval growth tends to increase the intervals for insecticide applications, which should target the most vulnerable stage of smaller larvae (about 1 cm) [
49].
Cotton is the last summer crop to be planted in the Florida Panhandle region, remaining for a longer period than other crops in the agricultural landscape before the fallow season. The longer larval development time of
H. zea when feeding on Cry1Ac+Cry2Ab, associated with infestations during mid- and late season, could expose larvae to shorter days and decreasing temperatures, factors that regulate insect diapause [
50,
51]. The Florida Panhandle is considered a “hybrid zone” of populations of noctuids, such as
S. frugiperda, which flies from south Florida and Texas to the northern U.S. [
52,
53]. Diapausing and migration of
H. zea populations from the Florida Panhandle may contribute to infestations in corn and cotton, North to 40 N latitude, where
H. zea cannot permanently survive [
32].
The pupal viability of
H. zea was similar among different cotton cultivars and populations. In contrast,
H. zea pupal weight varied when feeding on different cotton cultivars. Pupal weight is often correlated with fecundity [
54], although this correlation might be affected by several other factors [
55]. In our study, the heavier pupal weight was linked with higher fecundity, which agrees with reports for other noctuids, such as
H. armigera [
56].
Cumulative effects on specific life-history traits of
H. zea (i.e., larval survival and development time) impact the population growth potential on Bt cotton cultivars. The life table parameters indicated that
H. zea populations tested are expected to grow when feeding on Cry1Ac+Cry1F or Cry1Ac+Cry2Ab Bt cotton, but with reduced growth rates (i.e.,
R0,
rm). Overall, insects feeding on Cry1Ac+Cry1F and Cry1Ac+Cry2Ab are expected to generate 30% and 42% fewer individuals per day compared to non-Bt cotton, respectively. The growth potential of
H. zea from Jackson county population was similar when feeding on non-Bt or on Cry1Ac+Cry1F, which could result in a higher number of exposed offspring [
12]. However, the insects of the Jackson county had lower growth rates on non-Bt cotton compared with Escambia and Santa Rosa populations, indicating the presence of fitness costs [
57] when they do not feed on Cry1Ac+Cry1F cotton. In a theoretical scenario where only Cry1Ac+Cry1F cotton is cultivated in the Florida Panhandle, the
H. zea population from Jackson county is expected to produce in one generation 10–27% more females per female than the populations from Escambia and Santa Rosa. These differences between
H. zea populations reinforce that resistance may develop because of local selection [
58,
59].
During a period of over 23 years, in which commercialized Bt crops have been used in the U.S., IRM programs have relied on models to predict how quickly resistance to Bt may occur in different scenarios [
60,
61]. Although our study was performed in the laboratory and conclusions about field-evolved resistance are limited, this is one of the most complete life table studies of
H. zea in Bt cotton available in the literature. The information provided in this study (life-history traits and life table parameters) can contribute to the refinement of predictive models and delayed resistance to important Bt toxins, such as Vip3Aa [
22]. Our results reinforce the need for region-specific knowledge of target pests of Bt technology when designing IRM programs [
60].
Helicoverpa zea has a high dispersal capacity and reproductive biology, which leads to extensive gene flow [
60,
61]. However, fitness components and their variability across environments should be taken into consideration in the simulation of predictive models [
60].
In conclusion, this paper has quantified the dual and triple-gene Bt cotton effect on the life-history and demographic performance of three populations of H. zea from the Florida Panhandle. Triple-gene Bt cotton caused 100% larval mortality in all populations tested, indicating the value of Vip3Aa toxin on H. zea management in the region. Despite resistance, dual-gene Bt cotton containing Cry1A and Cry2A toxins significantly affected the fitness and demographic growth of the three populations of H. zea evaluated. However, the magnitude of the effect on the life-history, and consequently on the life table parameters of H. zea in a landscape containing 100% Bt cotton varied. Interaction between H. zea populations (Escambia, Santa Rosa, and Jackson counties) and cotton cultivars (Bt and non-Bt) was detected. These findings improve our understanding of how data on demographic growth rates of target pests to Bt technology matters and fill a gap by providing region-specific information when developing IRM programs. The results of this study also provide valuable parameters for the refinement of models to better predict the risk of resistance evolution and validate resistance management strategies, including refuge recommendations.