Control of Resistant Pink Bollworm (Pectinophora gossypiella) by Transgenic Cotton That Produces Bacillus thuringiensis Toxin Cry2Ab

Pink bollworm.Larvae were reared on wheat germ diet as described previously (29). We used six strains of pink bollworm: AZP-RO, AZP-RE, AZP99, APHIS-S, MOV97, and MOV97-R. AZP-RO is a Cry1Ac-resistant strain derived from the previously established resistant AZP-R strain as follows. Pink bollworm collected from 10 cotton fields in 1997 were used to start 10 laboratory strains (29). The AZP-R strain was created by pooling from these field-derived strains 159 individuals that survived on diet with 1, 3.2, or 10 μg of Cry1Ac per ml (32). The F5 progeny of the 159 parents of the AZP-R strain were fed a diet with 10 μg of Cry1Ac per ml. Survivors of this second round of selection were split into two subsets, AZP-RO and AZP-RE. AZP-RO was subsequently selected with 100 μg of Cry1Ac per ml of diet in every odd generation (F7, F9, F11, and so on), while AZP-RE was selected with 100 μg of Cry1Ac per ml of diet in every even generation (F6, F8, F10, and so on). AZP99 is a susceptible strain that was derived independently from the other strains studied. AZP99 was started by pooling 68 pupae and 108 adults from the parental through F3 generations of individuals collected from 15 field sites in Arizona in 1999. APHIS-S is a susceptible strain that had been reared in the laboratory for many years without exposure to toxins. MOV97 was started in 1997 from individuals collected from the Mohave Valley, Arizona. MOV97 was one of the 10 strains from which AZP-R was derived and was reared without exposure to toxins. MOV97-R was created by selecting a subset of the F10 generation of MOV97 with Cry1Ac. Subsequently, MOV97-R was selected on diet with 10 μg of Cry1Ac per ml every even generation (F12, F14, and so on).

Greenhouse bioassays.We tested larvae from the Cry1Ac-resistant AZP-RO strain in two greenhouse bioassays (F21 in experiment 1 and F23 in experiment 2). We obtained seeds from Monsanto for five types of cotton: Deltapine 50 (Delta and Pine Land Co., Scott, Miss.), which is not transgenic; Deltapine 50B, which produces Cry1Ac; line 17180, which produces Cry2Ab; and lines 15813 and 15985, which produce Cry1Ac plus Cry2Ab. Plants were grown and tests were conducted in a greenhouse at the University of Arizona Campus Agricultural Center in Tucson. Plants were grown in 20-liter pots on drip irrigation, with halogen lighting providing a 16-h photophase. Each pot contained one to three cotton plants grown in Premiere High Performance Pro-Mix BX (Premier Horticulture Ltd., Dorval, Quebec, Canada) potting soil. Every 2 weeks, each pot was provided with ca. 16 g of Ironite (Ironite Products Co., Scottsdale, Ariz.), a fertilizer and micronutrient supplement. Plants were infested with pink bollworm eggs 106 days (experiment 1) and 84 days (experiment 2) after planting, when they had produced at least two bolls that were >14 days old (ca. 15-mm diameter) per plant. In both experiments, we used a randomized block design, with each of eight blocks containing one plant that produced Cry1Ac and one plant of each of the other three transgenic lines. In experiment 1, each block also contained a non-B. thuringiensis plant.

Experiment 1 started with infestation on 27 January 2000 and ended 9 March 2000. Experiment 2 started 6 April 2000 and ended 18 May 2000. To infest B. thuringiensis cotton plants, squares (10 by 10 mm) of paper towel (Chicopee, Benson, N.C.) onto which 40 to 50 eggs had been laid by moths in the laboratory were placed under the bracts of selected bolls. Non-B. thuringiensis plants were infested with 10 eggs per boll. We used fewer eggs per boll on non-B. thuringiensis plants because in previous tests we found lower survival on B. thuringiensis plants than on non-B. thuringiensis plants. Eggs on the paper towel squares hatched, and neonate larvae entered the bolls. Seven to 12 days after infestation, we counted entrance holes on all infested bolls. Fourteen days after infestation, the bolls were enclosed in small cages constructed from 167-ml plastic cups fitted with screened lids. A small piece of paper towel was placed in each of these cages to provide a substrate on which emerging larvae pupated. Beginning 21 days after infestation, boll cages were opened twice weekly to remove and record pupae. Forty-two days after infestation, all bolls were removed, taken to the laboratory, and dissected to check for survivors. Survival was estimated as the number of survivors divided by the number of entrance holes times 100%.

From each of the five types of cotton, three to six bolls were freeze-dried and ground. We sent subsamples of the resulting powder in numbered vials to EnviroLogix (Portland, Maine) for toxin testing with immunoassays. With the exception of two bolls sampled from a single plant of line 15813 in experiment 2 in which no Cry2Ab was detected (see Results), all bolls tested had the toxin sets characteristic of their variety or line.

Artificial diet bioassays.We tested larvae from the APHIS-S, AZP-RO, AZP-RE, MOV97, MOV97-R, and AZP99 strains in laboratory bioassays on wheat germ diet (23, 29) containing no B. thuringiensis toxin (control), Cry1Ac, or Cry2Aa. Three sets of tests were done with Cry2Aa. In the first set, started on 9 December 1999, we tested APHIS-S, AZP-RO (F19), MOV97 (F27), and MOV97-R (F24). In the second set, started on 24 August 2000, we tested APHIS-S and AZP-RE (F26). Data for APHIS-S were pooled from the two sets. In the third set, started in May 2000, we tested AZP99 (F2) and AZP-RE (F24). APHIS-S, AZP-RO (F16), MOV97 (F24), MOV97-R (F24), and AZP-RE (F26) were tested with Cry1Ac between June 1999 and March 2000.

The source of Cry1Ac was MVPII (Dow Agrosciences, San Diego, Calif.), which is a liquid formulation containing a hybrid protoxin expressed in and encapsulated by Pseudomonas fluorescens (34, 36). Cry2Aa2 (referred to as Cry2Aa for brevity) was purified from Escherichia coli that produced Cry2Aa protoxin inclusion bodies (21). The Cry2Aa used in diet bioassays and the Cry2Ab in plants (formerly called CryIIA and CryIIB, respectively) have 88% amino acid identity (5). In pairwise comparisons against five lepidopteran pests, the median lethal concentrations of Cry2Aa and Cry2Ab differed by 1.1- to 5.8-fold against any particular pest (5). Here we do not assume that toxicity to pink bollworm is the same for Cry2Aa and Cry2Ab.

In the first two sets of diet bioassays, larvae were tested individually as follows. Diet was shredded into strips and dispensed into 33-ml cups. One neonate was transferred into each cup with a fine brush (23). Cups were held in darkness at 29 ± 2°C. Twenty to 60 larvae from each strain were tested individually at each concentration, with tests performed on at least two dates for each strain and concentration. In the third set, we tested 17 groups of larvae from AZP99 (F2) and 12 groups of larvae from AZP-RE (F24) with 20 to 200 larvae per cup (533 ml) at 27 ± 2°C. Although the density per cup varied widely in the third set, percent survival was not associated with density. After 21 days in all three sets, live fourth-instar larvae and pupae were scored as survivors. To adjust for control mortality, the survival on treated diet was divided by the survival on untreated diet.

Data analysis.In the comparison from experiment 1 between nontransgenic cotton (Deltapine 50) and B. thuringiensis cotton with Cry1Ac (Deltapine 50B), we used multiple regression analysis (27) to determine whether the survival of AZP-RO (arcsine-square root transformed) was affected by the cotton variety, block, or number of entrance holes per boll. To determine whether Cry2Ab affected the survival of AZP-RO in experiments 1 and 2, we contrasted the survival on B. thuringiensis cotton with Cry1Ac (Deltapine 50B) and that on B. thuringiensis cotton in the three lines with Cry2Ab alone or with Cry1Ac plus Cry2Ab (lines 17180, 15813, and 15985). Because survival did not differ between experiments 1 and 2 (see Results), we pooled the data from the two experiments and used a randomization test (25) to check the statistical significance of the effects of Cry2Ab on survival. The survival values (n = 8) were sampled at random with replacement 1,000 times, with the difference between the mean of the randomized survival values on B. thuringiensis cotton with Cry1Ac (n = 2) and the mean of the randomized survival values on the three lines with Cry2Ab alone or with Cry2Ab plus Cry1Ac (n = 6) being calculated each time. Using the distribution of the 1,000 differences calculated, we estimated the probability of obtaining a difference equal to or greater than the actual mean difference observed in the two experiments (i.e., 22.9%). We also used a randomization test with 1,000 repetitions to determine whether the survival of individually tested larvae differed significantly between the two susceptible strains (APHIS-S and MOV97) and the three Cry1Ac-resistant strains (MOV97-R, AZP-RO, and AZP-RE) on artificial diet with 0.32 or 1.0 μg of Cry2Aa per ml. In the final analysis, we used a randomization test to determine whether the survival of larvae in replicated groups on artificial diet with 1.0 μg of Cry2Aa per ml differed significantly between the susceptible AZP99 strain and the resistant AZP-RE strain.

Greenhouse bioassays.Two independent sets of greenhouse bioassays showed that larvae from a pink bollworm strain that survived on B. thuringiensis cotton with Cry1Ac had little or no survival on B. thuringiensis cotton with Cry2Ab alone or with Cry1Ac plus Cry2Ab. In experiment 1, the mean survival of Cry1Ac-resistant pink bollworm larvae from strain AZP-RO was 20.4% on non-B. thuringiensis cotton compared with 22.4% on B. thuringiensis cotton that produced Cry1Ac (Table 1). In the comparison between non-B. thuringiensis cotton and B. thuringiensis cotton with Cry1Ac, survival was not affected by cotton variety (F = 0.017; df = 1, 30; P = 0.90), block (F = 0.98; df = 7, 30; P = 0.46), or number of entrance holes per boll (F = 0.30; df = 1, 30; P = 0.58). These results show that Cry1Ac in the bolls of B. thuringiensis cotton did not kill resistant larvae. In contrast, B. thuringiensis cotton producing Cry2Ab alone or Cry1Ac plus Cry2Ab killed all of the resistant larvae tested in experiment 1 (Table 1).

In experiment 2, the mean survival of Cry1Ac-resistant larvae was 24.0% on B. thuringiensis cotton with Cry1Ac. No larvae survived on B. thuringiensis cotton with Cry2Ab alone or on B. thuringiensis cotton line 15985 with Cry1Ac plus Cry2Ab (Table 1). The mean survival on B. thuringiensis cotton line 15813 with Cry1Ac plus Cry2Ab was 1.8%.

The survival of Cry1Ac-resistant larvae on the four types of B. thuringiensis cotton (Deltapine 50B and lines 17180, 15813, and 15985) did not differ between experiments 1 and 2 (t = 0.035; df = 6; P = 0.97). Pooling results from experiments 1 and 2 showed that the average survival of Cry1Ac-resistant larvae was higher on B. thuringiensis cotton with Cry1Ac (23.2%) than on the three lines of B. thuringiensis cotton with Cry2Ab alone or with Cry1Ac plus Cry2Ab (0.3%) (randomization test, 1,000 repetitions, P = 0.008; see Materials and Methods).

The 11 survivors on B. thuringiensis cotton line 15813 with Cry1Ac plus Cry2Ab in experiment 2 were found on two bolls of a single plant. We were not able to rear progeny from the 11 survivors to determine if they had genetically based resistance to Cry2Ab. However, we did not detect Cry2Ab in two bolls of the single plant from which survivors emerged. Thus, we suspect that an absence or reduced concentration of Cry2Ab enabled survival on the aforementioned plant.

Diet bioassays.Tests in which larvae from six strains of pink bollworm were fed artificial diet showed that resistance to Cry1Ac did not confer strong cross-resistance to Cry2Aa. In diet bioassays where larvae were tested individually (Table 2), the MOV97-R, AZP-RO, and AZP-RE strains of pink bollworm had >90% survival on diet with 10 μg of Cry1Ac per ml, but no larvae survived on diet with 3.2 or 10 μg of Cry2Aa per ml. In addition, when individually tested larvae ate diet with 0.32 μg of Cry2Aa per ml, the average survival did not differ between the three resistant strains (MOV97-R, AZP-RO, and AZP-RE) and the two susceptible strains (APHIS-S and MOV-97) (Table 2; randomization test, 1,000 repetitions, P = 0.137). However, survival at 1 μg of Cry2Aa per ml in individual diet bioassays ranged from 5.1 to 10.0% for the three Cry1Ac-resistant strains, whereas the two susceptible strains had 0% survival (Table 2). At this concentration, the average survival was significantly higher for the three resistant strains than for the two susceptible strains (Table 2; randomization test, 1,000 repetitions, P = 0.027). In diet bioassays where larvae were tested in replicated groups, survival at 1 μg of Cry2Aa per ml was greater for the resistant strain AZP-RE (mean = 2.2%; standard error = 0.5%; n = 12 groups) than for the susceptible strain AZP99 (mean = 0%; standard error = 0%; n = 17 groups) (randomization test, 1,000 repetitions, P < 0.001).