No abstract

Transgenic plants that produce pesticidal proteins will release these proteins into the soil when these plants are incorporated into the soil by tillage or as leaf litter. Little is known about the fate and persistence of transgenic plant pesticidal products in the soil. We used a model system of transgenic cotton that produces Bacillus thuringlensis var. kurstaki δ-endotoxin (Btk toxin) to evaluate the persistence of transgenic pesticides in soil. Purified Btk toxin or transgenic cotton leaves containing Btk toxin were added to soil in five different microcosm experiments in concentrations ranging from 1 to 1600 ng Btk toxin/g soil. The concentration of the extractable Btk toxin was measured for up to 140 days. An initial rapid decline in extractable toxin concentration in the first 14 days, followed by a slower decline, was observed in four of the five experiments. At the end of the experiments, Btk toxin from transgenic plant tissue was undetectable (less than 0.1% of starting concentration) in two of the microcosm experiments and at 3, 16, and 35% of the original amounts in the other experiments. In addition, experiments using γ-irradiated sterilized soil indicated that the observed decline in extractable toxin concentration was due largely to biotic degradation rather than to physical adsorption by the soil.Key words: transgenic plants, Bacillus thuringlensis toxin, risk assessment.

Transgenic plants that produce pesticidal proteins have the potential to release these products into the environment when the plants are incorporated into soil. This could result in novel exposure of soil organisms to these pesticidal proteins. There is a lack of knowledge about the fate and persistence of transgenic pesticidal products in the soil. A model system of transgenic cotton, which produces Bacillus thuringiensis kurstakiδ‐endotoxin (Bt toxin), was used to address this issue. Methods were developed to quantify Btk toxin in soil and soil/plant litter by extraction of the Btk toxin with an aqueous buffer and quantification by ELISA. The highest recovery of Btk toxin from soil was obtained with a high salt, high pH buffer. In addition, for certain soil types, addition of a non‐ionic detergent, Tween‐20, was needed for optimal recovery. Recovery of Btk toxin from soil ranged from 60% for a low clay content, low organic matter soil to 27% for a high clay content, high organic matter soil. The limit of detection of this method is 0.5 ng of extractable toxin per g dry weight soil. The method was shown to be useful in tracking over time the persistence of both purified and transgenic Btk toxin in laboratory experiments.

Methods were developed to monitor persistence of genomic DNA in decaying plants in the field. As a model, we used recombinant neomycin phosphotransferase II (rNPT‐II) marker genes present in genetically engineered plants. Polymerase chain reaction (PCR) primers were designed, complementary to 20‐bp sequences of the nopaline synthase promoter in a transgenic tobacco and the cauliflower mosaic virus 35S promoter in a transgenic potato. The PCR reverse primer was complementary to a 20‐bp sequence of the N‐terminal NPT‐II coding region. The PCR protocol allowed for quantification of as few as 10 rNPT‐II genes per reaction. We analysed rNPT‐II marker gene amounts in samples obtained from two field experiments performed at different locations in Oregon. In transgenic tobacco leaves, buried at 10 cm depth in a field plot in Corvallis, marker DNA amount dropped to 0.36% during the first 14 days and was detectable for 77 days at a final level of 0.06% of the initial amount. Monitoring of residual potato plant litter, from the soil surface of a test field in Hermiston, was performed for 137 days. After 84 days marker gene amounts dropped to 2.74% (leaf and stem) and 0.50% (tuber) of the initially detected amount. At the final sample date 1.98% (leaf and stem) and 0.19% (tuber) were detectable. These results represent the first quantitative analysis of plant DNA stability under field conditions and indicate that a proportion of the plant genomic DNA may persist in the field for several months.

Summary 1.A ®eld study using transgenic plants with associated recombinant micro-organisms was conducted to assess the potential eects of genetically engineered organisms on soil ecosystems. Three genotypes of alfalfa plants (parental, transgenic aamylase-producing and transgenic lignin peroxidase-producing) were planted in an agricultural ®eld plot. Immediately prior to planting, the roots of the alfalfa plants were left uninoculated or were inoculated with a wild-type strain (PC), a recombinant strain with antibiotic resistances (RMB7201), or a recombinant strain with antibiotic resistances and enhanced nitrogen-®xation capability (RMBPC-2), of Sinorhizobium meliloti. 2. Analyses of the alfalfa plants and ®eld plot soil were made over two growing seasons and included: metabolic ®ngerprints and DNA ®ngerprints of soil bacterial communities; soil microbial respiration; population counts of indigenous soil bacteria, fungi, nematodes, protozoa and micro-arthropods; identi®cation of nematodes and micro-arthropods; plant shoot weight and chemistries; and soil chemistries and enzyme activities. 3. The lignin peroxidase transgenic plants had signi®cantly lower shoot weight, and higher nitrogen and phosphorus content, than the parental or transgenic amylase plants. Distinct metabolic ®ngerprints, based on patterns of substrate utilization in Biolog plates, were exhibited by the soil bacterial communities associated with the three alfalfa genotypes, and those for the lignin peroxidase plants were the most unique. Signi®cantly higher population levels of culturable, aerobic sporeforming and cellulose-utilizing bacteria, lower activity of the soil enzymes dehydrogenase and alkaline phosphatase, and higher soil pH levels, were also associated with the lignin peroxidase transgenic plants. Signi®cantly higher population levels of culturable, aerobic spore-forming bacteria were also measured in the treatments containing the recombinant RMBPC-2 S. meliloti. 4. Population levels of protozoa, nematodes and micro-arthropods, DNA ®nger-prints of indigenous soil bacteria, and rates of microbial substrate-induced respiration were not signi®cantly aected by the transgenic alfalfa and recombinant S. meliloti treatments. 5. These results suggest that the genetically engineered organisms caused detectable