Field experiments were conducted in Chile and western Canada to measure short-distance (0 to 100 m) outcrossing from transgenic safflower (Carthamus tinctorius L.) intended for plant molecular farming to non-transgenic commodity safflower of the same variety. The transgenic safflower used as the pollen source was transformed with a construct for seed-specific expression of a high-value protein and constitutive expression of a gene conferring resistance to the broad-spectrum herbicide glufosinate. Progeny of non-transgenic plants grown in plots adjacent to the transgenic pollen source were screened for glufosinate resistance to measure outcrossing frequency. Outcrossing frequency differed among locations: values closest to the transgenic pollen source (0 to 3 m) ranged from 0.48 to 1.67% and rapidly declined to between 0.0024 to 0.03% at distances of 50 to 100 m. At each location, outcrossing frequency was spatially heterogeneous, indicating insects or wind moved pollen asymmetrically. A power analysis assuming a binomial distribution and a range of alpha values (type 1 error) was conducted to estimate an upper and lower confidence interval for the probable transgenic seed frequency in each sample. This facilitated interpretation when large numbers of seeds were screened from the outcrossing experiments and no transgenic seeds were found. This study should aid regulators and the plant molecular farming industry in developing confinement strategies to mitigate pollen mediated gene flow from transgenic to non-transgenic safflower.

Horizontal gene transfer (HGT) is an important mechanism by which bacteria recombine and acquire novel genes and functions. Risk scenarios where novel plant transgenes transfer horizontally into bacteria have been addressed in numerical theoretical assessments and experimental studies. A key outcome of these studies has been that the frequencies of such inter-domain transfer are very low, if occurring at all, suggesting that such transfers would not occur at a level that is biologically significant. The relationship between transfer frequencies and the subsequent selection or genetic drift of transgene carrying bacteria often remains unresolved in these studies and assessments. Here we present a stochastic model to better understand the initial establishment and population dynamics of rare bacterial transformants carrying horizontally acquired (trans)genes. The following key parameters are considered: initial transformant numbers, strength of selection, bacterial population size and bacterial generations (time). We find that the initial number of transformants is important for the subsequent persistence of transformants only in the range of 1 to ~50 independent HGT events. Our simulations show that transformant populations under a wide range of HGT rates and selection coefficients undergo stochastic developments where they persist at low frequencies for up to several years (at frequencies that are below detection using available field sampling methodology), after which they eventually may go to fixation. Stochastic variability may thus play a crucial but disregarded role in the design of field monitoring strategies e.g. in biosafety assessments. We also estimate the time required for transformants to reach 0.0002% prevalence in a bacterial population, a threshold that allows experimental detection of transgene carrying bacteria through sampling of the larger bacterial populations.

This is the report of the first open field release of genetically modified microorganisms (GMMs) in Italy. It covers ten years of monitoring, and follows in-field GMM dynamics from strain release to disappearance below detection limits, as well as assessment of impact on resident microorganisms. The bacteria released belong to the nitrogen fixing legume endosymbiont Rhizobium leguminosarum bv. viciae, and were engineered with non-agronomically-proficient traits, in order to assess their behavior and fate without GMM-specific positive feedback from the plant. A DNA cassette containing mercury resistance and ß-galactosidase genes was introduced in either plasmid-borne or chromosomally integrated versions, in order to test the resulting strain stability. A synthetic promoter was used to drive the lacZ gene, conferring high catabolic activity to the GMM. Two different wild-type Rhizobium backgrounds were tested, comparing a non-indigenous vs. an indigenous, highly competitive strain. The latter had much greater persistence, since it was able to survive and establish at technically detectable levels for over four years after release. Selection factors, such as reiterated presence of the plant host, or lactose substrate supply, enhanced long-term survival to different extents. The lactose treatment showed that even a single trophic supplementation can surpass the benefits of symbiotic interaction for a period of several years. Concerning impact, the GMMs did not alter substantially the other soil community general microbiota. However, there were some significant differences in microbiota as a consequence of the Rhizobium inoculation. This effect was observed with either the WT or GMM, and was more evident in the release of the indigenous Rhizobium. Moreover, as the indigenous GMM had its parental, dominant wild-type in the same soil, it was possible to evaluate to what extent the GMM version could result in parent displacement (“self-impact”), and how much the two rhizobia would additively contribute to nodulation.

Auchenorrhyncha (planthoppers and leafhoppers) are herbivorous organisms that can ingest Cry proteins from genetically engineered Bt-crops depending on their feeding behaviour. Consequently, they might be directly affected by non-target Bt-protein action and more importantly serve as a source of Cry protein exposure to beneficial predatory arthropods. During a three year field study, we surveyed the community of Auchenorrhyncha in Diabrotica-resistant Bt-maize DKc5143-Bt (event MON88017), its near-isogenic line and two conventional hybrids using sweep netting and custom made sticky traps. Zyginidia scutellaris (Herrich-Schäffer) (Hemiptera: Cicadellidae) represented more than 60% of all captured individuals, indicating that it is the dominant leafhopper within the maize community. The statistical analysis of Z. scutellaris data using confidence intervals for the ratios of mean abundance showed no consistent differences between the Bt-maize and the near-isogenic cultivar, indicating no negative impact of event MON88017. The two conventional hybrids Benicia and DK315 exhibited differences in terms of Z. scutellaris densities, which were greater than those observed between MON88017 and the near-isogenic line, but also not consistent over the years. Six more species accounted for an additional 39% of all captured specimens, while ten more species were found only as single individuals and can be considered vagrants from neighbouring habitats. These results inform future field work on the non-target impact of Bt-maize on this group of arthropods and monitoring approaches to assess biological control function by surveying herbivore communities.

The objective of this study was to evaluate the effect of insect larval feeding on the fate and genetic transformability of recombinant DNA from a transplastomic plant. Leaves of tobacco plants with an aadA antibiotic resistance gene inserted into their chloroplast genome were incubated with larvae of the tobacco hornworm Manduca sexta (Lepidoptera). The specifically designed Acinetobacter strain BD413 pBAB2 was chosen to analyze the functional integrity of the aadA transgene for natural transformation after gut passages. No gene transfer was detected after simultaneous feeding of leaves and the Acinetobacter BD413 pBAB2 as a recipient, even though 15% of ingested Acinetobacter BD413 cells could be recovered as viable cells from feces 6 h after feeding. Results with real-time PCR indicated that an average of 98.2 to 99.99% of the aadA gene was degraded during the gut passage, but the range in the number of aadA genes in feces of larvae fed with transplastomic leaves was enormous, varying from 5 × 106 to 1 × 109 copies.g-1. DNA extracted from feces of larvae fed with transplastomic leaves was still able to transform externally added competent Acinetobacter BD413 pBAB2in vitro. Transformation frequencies with concentrated feces DNA were in the same range as those found with leaves (10-4–10-6 transformants per recipient) or purified plasmid DNA (10-3–10-7). The presence of functionally intact DNA was also qualitatively observed after incubation of 30 mg freshly shed feces directly with competent Acinetobacter BD413 pBAB2, demonstrating that aadA genes in feces have a potential to undergo further horizontal gene transfer under environmental conditions.

To provide estimates of horizontal gene transfer from transgenic crops to indigenous soil bacteria, transformation frequencies were obtained for naturally transformable Acinetobacter baylyi BD413 using a chromosomally integrated plant transgene. The transgene comprised sequences for two phenotypic markers: kanamycin resistance (npt II) and green fluorescent protein (gfp), expressed from their own bacterial promoters. Recipient bacteria carried a copy of these two genes, with deletions in their 3'-termini abolishing the marker activity, these genes were integrated into a 16S rRNA gene in the bacterial chromosomal genome or carried on a broad host range plasmid. Successful recombination between the plant transgene and the bacterial genome resulted in restoration of the markers, allowing detection through antibiotic selection and fluorescence. Transformation parameters of increasing complexity, without any enrichment steps, were used to approach the field conditions, while still obtaining measurable transformation frequencies. In pure culture filter experiments, transformation was detected using ground, chopped and whole leaves, as well as whole sterile seedlings, and ground roots. In sterile soil microcosms, transformation was detected using pure plant DNA (3.6 × 10-8 transformants per recipient) and ground leaves (2.5 × 10-11). Transformation was also detected for the first time in non-sterile soil using pure plant DNA (5.5 × 10-11). Since the same constructs were used throughout, these data allow predictions of even more complex environmental systems where measurable frequencies are not easily obtainable.

Cassava is an important subsistence crop grown only in the tropics, and represents a major source of calories for many people in developing countries. Improvements in the areas of resistance to insects and viral diseases, enhanced nutritional qualities, reduced cyanogenic content and modified starch characteristics are urgently needed. Traditional breeding is hampered by the nature of the crop, which has a high degree of heterozygosity, irregular flowering, and poor seed set. Biotechnology has the potential to enhance crop improvement efforts, and genetic engineering techniques for cassava have thus been developed over the past decade. Selectable and scorable markers are critical to efficient transformation technology, and must be evaluated for biosafety, as well as efficiency and cost-effectiveness. In order to facilitate research planning and regulatory submission, the literature on biosafety aspects of the selectable and scorable markers currently used in cassava biotechnology is surveyed. The source, mode of action and current use of each marker gene is described. The potential for toxicity, allergenicity, pleiotropic effects, horizontal gene transfer, and the impact of these on food or feed safety and environmental safety is evaluated. Based on extensive information, the selectable marker genes nptII, hpt, bar/pat, and manA, and the scorable marker gene uidA, all have little risk in terms of biosafety. These appear to represent the safest options for use in cassava biotechnology available at this time.

When growing different transgenic herbicide-resistant oilseed rape cultivars side by side, seeds with multiple herbicide resistance can arise, possibly causing problems for the management of volunteer plants. Large-scale field experiments were performed in the years 1999/2000 and 2000/2001 in order to investigate the frequencies and the consequences of the transfer of herbicide resistance genes from transgenic oilseed rape to cultivars grown on neighboring agricultural fields. Transgenic oilseed rape with resistance to glufosinate-ammonium (LibertyLink, LL) and with glyphosate resistance (RoundupReady, RR), respectively, was sown in adjacent 0.5 ha plots, surrounded by about 8 ha non-transgenic oilseed rape. The plots and the field were either in direct contact (0.5 m gap width) or they were separated by 10 m of fallow land. Seed samples taken during harvest in the transgenic plots at different distances were investigated for progeny with resistance to the respective other herbicide. It was found that outcrossing frequencies were reduced to different extents by a 10 m isolation distance. In addition to pollen-mediated transgene flow as a result of outcrossing, we found considerable seed-mediated gene flow by adventitious dispersal of transgenic seeds through the harvesting machine. Volunteer plants with double herbicide resistance emerging in the transgenic plots after harvest were selected by suitable applications of the complementary herbicides Basta® and Roundup Ultra®. In both years, double-resistant volunteers were largely restricted to the inner edges of the plots. Expression analysis under controlled laboratory conditions of double-resistant plants generated by manual crosses revealed stability of transgene expression even at elevated temperatures. Greenhouse tests with double-resistant oilseed rape plants gave no indication that the sensitivity to a range of different herbicides is changed as compared to non-transgenic oilseed rape.

To obtain information on differences between the metabolic pathways of the herbicide glufosinate (trade names: BASTA®, LIBERTY®) in non-transgenic, glufosinate-sensitive plants and in transgenic, glufosinate-resistant plants, the metabolism of 14C-labeled glufosinate and its enantiomers L- and D-glufosinate was studied using cell cultures of oilseed rape and corn. Transformation of glufosinate in both sensitive and transgenic rape cells remained at a low rate of about 3-10% in contrast to corn cells, where 20% was transformed in sensitive and 43% in transgenic cells after 14 days of incubation, the rest remaining as unchanged glufosinate. In sensitive rape and corn cells the main metabolite was 4-methylphosphinico-2-oxo-butanoic acid (PPO) with 7.3 and 16.4%, respectively, together with low amounts of 3-methylphosphinicopropionic acid (MPP), 4-methylphosphinico-2-hydroxybutanoic acid (MHB), 4-methylphosphinicobutanoic acid (MPB) and 2-methylphosphinicoacetic acid (MPA). An additional metabolite formed in transgenic cell cultures was 2-acetamido-4-methylbutanoic acid (N-acetyl-L-glufosinate, NGA), which was formed at rates of 3.2% in rape and 16.1% in corn. A further minor metabolite, not yet identified, was detected in both cell types. The liberation of 0.2% 14CO2 indicates further metabolic steps prior to a limited mineralization in plant cell cultures. L-glufosinate was transformed into the same metabolites as the glufosinate racemate. D-glufosinate was not metabolized.

The reliable use of purified Cry1Ab protein standards is a prerequisite for ecological studies and resistance monitoring programs of Cry1Ab-expressing transgenic corn. In this study the stability and activity of different Cry1Ab protein batches expressed in and purified from Escherichia coli were determined during two-year storage at different temperature conditions (4 °C, –20 °C, and –80 °C). SDS-Polyacrylamide gel electrophoresis showed degradation of the protein stored at 4 °C over four months, whereas no difference in the band intensity of the Cry1Ab proteins stored at –20 °C and –80 °C was observed. Bioassays with neonate larvae of Ostrinianubilalis indicated that the biological activity of Cry1Ab varied from batch to batch, depending on the production process. Cry1Ab protein stored at 4 °C for four months showed a significantly decreasing activity measured as median lethal concentration (LC50), whereas the protein activity declined less than 11-fold after two years storage at –20 °C. When stored at –80 °C the toxin activity remained relatively stable for at least 30 months, as indicated by low LC50 values of 7–10 ng Cry1Ab per cm2 diet. These experiments demonstrate that appropriate long-term storage conditions of Cry1Ab protein standards are crucial for resistance monitoring programs of Bt corn, and storage at –80 °C is recommended.

Impact assessments of virus resistance transgene introgression into wild, free-living populations are important for determining whether these transgenes present a risk to agriculture or the environment. Transgenic virus-resistant Cucurbita pepo ssp. ovifera var. ovifera L. (squash) cultivars have been commercialized, and may be cultivated in close proximity to cross-compatible wild, free-living relatives (C. pepo subsp. pepo vars. ozarkana and texana). Therefore, the potential impact of these virus resistance transgenes was studied by surveying the incidence and fluctuations of virus infection (as assayed by ELISA), virus symptoms (which may not be seen in an infected plant) and population size in forty-three free-living C. pepo populations in Illinois, Missouri, Arkansas, Mississippi, Louisiana, and Texas. Ten of these populations were studied over three consecutive seasons. Depending on the year, 61% to 78% percent of the populations had at least one individual infected by at CMV, ZYMV or WMV2, but the median incidence of infection within populations was 13%. The observed infection level in free-living populations was consistent with levels defined as “low” in field plot experiments conducted by others, leading to the conclusion that transgenic virus resistance should not provide a significant fitness advantage to the free-living populations examined. Viral symptoms were detected in only 2% of plants observed, indicating that severity of viral infection was low. CMV, ZYMV, and WMV2 were not the only viruses infecting these populations, further reducing the likelihood that resistance to these viruses would release populations from constraints imposed by virus diseases.

The objectives of the present study were (1) to investigate the qualitative composition of rhizodeposits leached from soils cropped with non-transgenic and genetically modified (GM) potatoes, and disclose if there were GM-specific modifications in potato rhizodeposition, and (2) to compare these results with conventional bulk parameters of microbial activity in soil. We have raised potatoes from a non-transgenic line (Solanum tuberosum L. cv. Désirée) and three GM lines, which expressed a gene for the resistance to kanamycin (DLH 9000) and a gene for T4 lysozyme (DL10 and DL12). A sandy soil placed in 340 cm3-“CombiSart” containers was used, from which the rhizodeposit was leached after a six-week growth period. The freeze-dried leachates were analyzed by pyrolysis-field ionization mass spectrometry (Py-FIMS). The Py-FI mass spectra gave detailed molecular-chemical information about the composition of leachates, indicating that the potato growth generally altered the composition of the soil solution. Moreover, a principal component analysis of the mass spectra showed differences between the leachates from the non-transgenic parent line and the GM potatoes as well as among the latter group. However, these differences in molecular composition could not be assigned to the release of T4-lysozyme into soil. Dehydrogenase activity and substrate-induced soil respiration as more common bulk parameters of soil microbial activity failed to disclose any significant effects of the various potatoes grown. The limitations of the described rhizodeposit leaching and analysis for risk assessment of GM potato cropping under field conditions are discussed critically. However, it could be concluded that the Py-FI mass spectrometric “fingerprint” can be developed as a fast, comprehensive, highly sensitive and reproducible analytical approach to discern any effects GM-crops may exert on soil ecological parameters.

Modeling pollen dispersal to predict cross-pollination is of great importance for the ongoing discussion of adventitious presence of genetically modified material in food and feed. Two different modeling approaches for pollen dispersal were used to simulate two years of data for the rate of cross-pollination of non-GM maize (Zea mays (L.)) fields by pollen from a central 1 ha transgenic field. The models combine the processes of wind pollen dispersal (transport) and pollen competition. Both models used for the simulation of pollen dispersal were Lagrangian approaches: a stochastic particle Lagrange model and a Lagrangian transfer function model. Both modeling approaches proved to be appropriate for the simulation of the cross-pollination rates. However, model performance differed significantly between years. We considered different complexity in meteorological input data. Predictions compare well with experimental results for all simplification steps, except that systematic deviations occurred when only main wind direction was used. Concluding, it can be pointed out that both models might be adapted to other pollen dispersal experiments of different crops and plot sizes, when wind direction statistics are available. However, calibration of certain model parameters is necessary.

Since genetically modified (GM) lactic acid bacteria (LAB) might be released in open environments for future nutritional and medical applications, the purpose of this study was to determine an upper limit for the horizontal gene transfer (HGT) in the digestive tract (DT) from Lactococcus lactis carrying heterologous genes (lux genes encoding a bacterial luciferase) to Enterococcus faecalis. Two enterococcal wide host-range conjugative model systems were used: (i) a system composed of a mobilizable plasmid containing the heterologous lux genes and a native conjugative helper plasmid; and (ii) a Tn916-lux transposon. Both systems were tested under the most transfer-prone conditions, i.e. germfree mice mono-associated with the recipient E. faecalis. No transfer was observed with the transposon system. Transfers of the mobilizable plasmid carrying heterologous genes were below 102 transconjugants per g of faeces for a single donor dose and reached between 103 and 104 transconjugants per g of faeces when continuous inoculation of the donor strain was used. Once established in mice, transconjugants persisted at low levels in the mouse DT.

Natural transformation is assumed to be the most likely mechanism by which DNA from transgenic plants could be horizontally transferred to bacteria. In order to determine the occurrence of naturally transformable bacteria amongst bulk and rhizosphere soil bacteria, different transformation strategies were employed using either plasmid DNA (IncQ plasmids pSM1890 and pSM1885, conferring GFP, Smr, Gmr and GFP, Smr, Tcr, respectively) or genomic DNA from rhizosphere isolates, which were chromosomally tagged with mini-Tn5~(GFP, Tcr, as transforming DNA. Transformation assays were done in microtiter plates (262 isolates and pSM1890 or pSM1885), on filters (i) with rhizosphere bacterial community mixed with pSM1890 or pSM1885, (ii) with 24 rhizosphere or soil bacterial isolates mixed with genomic DNA of the corresponding mini-Tn5-tagged strains, and in the rhizosphere of tobacco plants inoculated with rifampicin-resistant bacterial isolates and genomic DNA of the corresponding mini-Tn5-tagged strains added. One transformant colony was obtained when Brevundimonas vesicularis was transformed with genomic DNA of the corresponding mini-Tn5-tagged strain. Attempts to reproduce this result were unsuccessful. With this single exception, transformants were neither detected in the collection of isolates nor in the rhizosphere bacterial community. Acinetobacter baylyi BD413 used as a positive control showed drastically reduced transformation frequencies with plasmid pSM1890 as transforming DNA when mixed with the rhizosphere pellet. All transformants were characterized by BOX-PCR fingerprints, and three different BOX patterns were revealed. Sequencing the 16S rRNA gene showed that all transformants could be assigned to Acinetobacter sp. Since transformants were only observed in the positive control, the introduced BD413 either underwent genomic rearrangements, or competence of the Acinetobacter population present in the rhizosphere was stimulated by the introduction of BD413. The various transformation assays performed indicate that the proportion of rhizosphere or bulk soil bacteria which are naturally transformable is negligibly low.

Risk-assessment studies of virus-resistant transgenic plants (VRTPs) focussing on recombination of a plant virus with a transgenic sequence of a different virus should include a comparison of recombination frequencies between viruses in double-infected non-transgenic plants with those observed in singly infected transgenic plants to estimate recombination incidence in VRTPs. In this study, the occurrence of recombination events was investigated in non-transgenic plants double-infected with two different potyviruses, as well as in potyviral genomes in singly infected transgenic plants expressing potyvirus sequences. Different potyviruses, namely Potato virus A (PVA), Tobacco vein mottling virus (TVMV), two strains of Potato virus Y (PVY-O, PVY-H) and two strains of Plum pox virus (PPV-NAT, PPV-SK68), were used in three combinations for double infection of a common host. Furthermore, transgenic plants expressing either potyviral coat protein (CP), helicase (CI) or polymerase (NIb) coding sequences (PPV-NAT-CP, PVY-CI, PVY-NIb) were singly-infected with a heterologous potyvirus, which was not targeted by the respective transgenic resistance. To identify recombinant potyviral sequences, a sensitive RT-PCR was developed to detect up to one recombinant molecule out of 106 parental molecules. In 304 mixed infected non-transgenic plants, 92 mixed and 164 single infected transgenic plants screened for recombinant sequences no recombinant potyviral sequence was found. These results indicate that recombination events between different potyviruses in mixed infections and between a potyvirus infecting a potyvirus-resistant transgenic plant are likely to be very infrequent.

With the extensive adoption of transgenic crops, an understanding of transgene flow isessential to manage gene flow to non-GM crops. Thus, a flexible and accurate numericalmodel is required to assess gene flow through pollen dispersal. A three-dimensionalatmospheric model combined with a diffusion transport model would be a useful tool forpredicting pollen dispersal since it would be flexible enough to incorporate the effectsof factors such as the spatial arrangement of crop combinations, land use, topography,windbreaks, and buildings. We applied such a model to field measurements of gene flowbetween two adjacent maize (Zea mays) cultivars, with suppression effectsdue to windbreaks, in an experimental cornfield in Japan. This combined model reproducedthe measured cross-pollination distribution quite well in the case of maize plots withplant windbreaks slightly taller than the maize and without windbreaks, but the modelunderestimated the effect of a 6-m-tall windbreak net beyond 25 m from the donor pollensource on cross-pollination. The underestimation was most probably due to the problem ofassimilated wind data. The model showed that the 6-m-tall windbreak and the plant windbreak suppressed average cross-pollination rate by about 60% and 30%, respectively.Half-tall and coarser mesh windbreak net suppressed cross-pollination rates by 40% byreducing the swirl of donor pollen by reduced wind speed.

Over the last decade, plant virus-based vectors have been developed and successfully exploited for high-yield production of heterologous proteins in plants. However, widespread application of recombinant viruses raises concerns about possible risks to the environment. One of the primary safety issues that must be considered is the uncontrolled spread of the genetically engineered virus from experimental plants to susceptible weeds or crops. Using a movement-deficient Potato virus X (PVX)-based transient gene expression vector which harbors the β-glucuronidase (gus) gene, we established a plant viral expression system that provides containment of the recombinant virus and allows for safe and efficient protein production. By deletion of the viral 25k movement protein gene, systemic spread of the modified virus in non-transgenic Nicotiana benthamiana plants was successfully inhibited. In transgenic N. benthamiana plants expressing the 25K viral movement protein, this deficiency was complemented, thus resulting in systemic infection with the movement-deficient virus. While no differences in virus spread and accumulation were observed compared to infection caused by wild-type PVX in non-transgenic plants, the movement protein transgenic plants exhibited none of the normal symptoms of viral infection. Several biosafety aspects were investigated including the potential for recombination between the defective virus and the movement protein transgene, as well as complementation effects in non-transgenic plants doubly infected with the defective and the wild-type virus. Furthermore, the applicability of the safety system for the production of heterologous proteins was evaluated with gus as a model gene. With respect to the stability of the gus insert and the expression level of the GUS protein, there were no differences between the novel system developed and the conventional PVX-based expression system.

Maize diversity is widespread in Mexico and it has been stewarded by campesinos in small communities until the present. With the arrival of transgenic maize, the objective of this study is to analyze possible scenarios that could result if genetically modified maize were not regulated and openly available in Mexico. By applying a simple logistic model based on the conditions of maize production in Mexico, the dispersion of transgenic maize in different situations within fields of farmers is described. In traditional open systems of freely exchanged seed within communities it is concluded that the most likely outcome of GM maize release is the incorporation of transgenes in the genome of Mexican germplasm and possibly in that of teosinte.