Developing a method for laboratory biosafety testing impacts of transgenic plants on non-target insects

A series of laboratory experiments was carried out with the main objective of contributing to the development of biosafety testing methods for transgenic plants or plants produced by other novel technologies. The adult stage of a New Zealand native weevil, Nicaeana cervina (Broun) (Coleoptera: Curculionidae: Entiminae), was used as a test insect as this has become well-adapted to introduced pasture species, feeding on both ryegrass (Lolium perenne L. (Poaceae)) and white clover (Trifolium repens L. (Fabaceae)). Ryegrass plants used in experiments were infected with strains of the symbiotic endophytic fungi Neotyphodium lolii (Latch, M.J. Chr. & Samuels) and Epichloe festucae Leuchtm., Schardl & M.R. Siegel and included a transgenic endophyte, which was modified to eliminate production of the insect deterrent, peramine. Nicaeana cervina was exposed to these endophyte-infected plants along with endophyte-free ryegrass. Weevils were also exposed to white clover with high or low levels of cyanogens, and to white clover selections bred conventionally to contain high or low levels of crude protein. These bioassays were carried out sometimes over a series of time intervals, and weevil survival, feeding, and reproductive performance variables were measured. Nicaeana cervina survival was generally high, although feeding and oviposition were variable between experiments despite high replication in some cases. A summary of significant differences obtained across all experiments suggested that data on the proportion of weevils feeding and amounts of leaf material consumed were more often significantly different than treatment effects on survival or reproductive performance. A positive relationship between the proportion of weevils feeding and those ovipositing was found, suggesting that a 50% reduction in feeding would result in a 28% reduction in the proportion of weevils ovipositing. Power analysis using the range of values recorded for the variables measured indicated that for proportions of weevils surviving, feeding, or fecund, about 35-42 replicates would be required for a 40% difference in values. For measures of leaf consumption and numbers of eggs per female, fewer replicates were required for 95% power when these values were at the higher end of the range than when they were at the lower end of the observed range. The usefulness of the case study test insect and risk assessment protocols used for biosafety testing are discussed.