Insect Bioassays

Proteins were tested for insecticidal activity using neonate Lepidopteran larvae on artificial insect diet. Larvae of H. zea, O. nubilalis, S. frugiperda and P. includens were hatched from eggs obtained from Benzon Research Inc. (Carlisle, PA). The bioassays were conducted in 128-well plastic trays specifically designed for insect bioassays (C-D International, Pitman, NJ). Each well contained 1.5 mL of Multi-species Lepidoptera diet (Southland Products, Lake Village, AR). A 40 µL aliquot of protein sample was delivered by pipette onto the 2 cm² diet surface of each well (20 μL/cm²). Treatment concentrations were calculated as the amount (ng) of protein per square centimeter (cm²) of surface area in the well. The treated trays were held in a fume hood until the liquid on the diet surface had evaporated or was absorbed into the diet. Within a few hours of eclosion, individual larvae were picked up with a moistened camel hair brush and deposited on the treated diet, one larva per well. Sixteen animals were used per treatment. The infested wells were then sealed with adhesive sheets of clear plastic, vented to allow gas exchange (C-D International, Pitman, NJ). Bioassay trays were held at 28 °C, ~40% relative humidity and 16:8 hours light:dark for 5 days, after which the total number of insects exposed to each protein sample, the number of dead insects and number of moribund insects were recorded. Moribund insects were classified as insects that were still alive, did not increase in size over the course of the bioassay, and did not respond to perturbation.

Statistical analysis was carried out using JMP® Pro Version 11 software (SAS Institute Inc., Cary, NC). Lethal concentrations (LC₅₀) of Vip3 proteins were calculated on sum of dead and moribund insects using a generalized linear model utilizing Probit analysis of binomial data. An inverse prediction of LC50 with 95% confidence intervals was calculated based on this model. For determination of growth inhibition, total live insect mass was weighed after 5 days of treatment and normalized for insect number. Sixteen insects were used per treatment and experiments were repeated twice. Insects tested on buffer control diet were compared to Vip3-treated samples. Average weight of insects after Vip3 treatment was modeled using linear mixed models for each insect species respectively:

where Y is the insect average weight observed in each treated well, µ is the overall mean, T is the effect of Vip3 protein at a specific dose, E is the effect of different experimental run, ɛ is the error. The Vip3 protein treatment effect T was modeled as a fixed effect, while experimental run E was treated as a random effect. Box-Cox transformation was applied to correct for non-normality and heterogeneous variances. Average insect weight was compared with null hypothesis of no growth difference between a Vip3 protein treatment and the untreated control (buffer) and the alternative hypothesis of different growth under a Vip3 protein treatment. We used Dunnett’s test to adjust P-values for multiple comparisons.

All data generated and analysed during this study are included in this published article.