ABSTRACT: Brazil and South Africa, countries with economies in transition, are characterized by a dual agrarian structure co-occurring, sometimes, alongside in the same region. Large-scale commercial farming produces crops for export to global markets on the one hand, and small-scale farming, on the other hand, provides for subsistence and local markets. In both systems, maize (Zea mays) is a key crop for these two countries. For the commercial system, maize is a commodity crop while for the small-scale system it is the prime staple crop. In commercial systems, farmers predominantly grow genetically modified (GM) hybrid maize. In small-scale systems, farmers rely on open pollinated varieties (OPVs) and/or landraces and are largely dependent on seed saving systems. The aim of this study was to understand the relationship between transgene expression rates, the resulting concentrations of the transgene product (Bt protein) and its bioactivity in insect pests following transgene flow from GM hybrid maize into non-genetically modified, non-GM near-isogenic maize hybrid (ISO) and OPVs. We modeled segregation patterns and measured cry1Ab transgene expression (mRNA quantification), Cry1Ab protein concentration and pest survival. Two groups of F1, F2 crosses and backcrosses with GM, ISO and OPV maize varieties from Brazil and South Africa were used. Bioassays with the larvae of two lepidopteran maize pest species, Helicoverpa armigera and Spodoptera littoralis, were carried out. Overall, the cry1Ab transgene outcrossed effectively into the genetic backgrounds tested. The cry1Ab transgene was stably expressed in both ISO and OPV genetic backgrounds. Transgene introgression led to consistent, though highly variable, concentrations of Cry1Ab toxins that were similar to those observed in GM parental maize. Most crosses, but not all, suggested the expected Mendelian segregation pattern. Transgene expression rates were significantly higher than expected from purely Mendelian segregation in the South African crosses. In South African materials, ISO and OPV crosses produced significantly lower Cry1Ab concentrations compared to the GM parental maize. The Cry1Ab toxins from crosses were bioactive and induced mortality rates of ?92.19% in H. armigera and ?40.63% in S. littoralis after a period of only 4 days. However, no correlations were observed between the quantitation of mRNA for cry1Ab and the corresponding Cry1Ab protein concentrations, nor between the Cry1Ab concentrations and insect mortality rates across different genetic backgrounds. We therefore suggest that while transcription of the cry1Ab transgene reliably determines the presence of Cry1Ab protein, mRNA levels do not reflect, by themselves, the end Cry1Ab protein concentrations found in the plant. Because predictably high Cry1Ab concentrations are a key prerequisite for effective insect resistance management (IRM) programs, this observation raises questions about the effectiveness of such programs in scenarios with complex crop genetic backgrounds. On the other hand, confirmed bioactivity in all crosses should be expected to impact small farmer's selection behavior, unknowingly favoring the insecticidal trait. This may lead to a fixation of the trait in the wider population, and may influence the genetic diversity of varieties maintained by small-scale farmers.
