Skip to main content

Safety assessment of the deliberate release of two model transgenic crop plants, oilseed rape and sugar beet


The objective is to study of the dispersal of genes via genetic engineering into plants under common agricultural situations in the field. New techniques, assays and probes for the quick and efficient tracking of the transgenes in seeds and plants will be developed.
During the Biotechnology Action Programme (BAP) project Study of gene dispersal from plants by recombinant DNA technology characteristic features of small scale releases were identified. At a small scale (which in many cases is related to the first introduction and early development of a genetically modified plant) the experimenter may still include a number of redundant safeguards in the design and carry out detailed monitoring.

As the product development evolves, the surface and scope of the experiments imply larger areas that do no longer entail some of the earlier protection measures, for example including a broader strip of 5 m of nonmodified control plants around a large area proves to be costly and probably inadequate.

Research has been carried out into the evaluation of pollen dispersal in nonconfined releases. The dynamics of pollen movement was studied from fields that have no additional safety features and are comparable to good agricutural practice. One source field was planted with genetically modified plants, carrying an easily recognizable tracer and catcher fields were distributed in the surrounding area. Posttrial evaluation was carried out in the progeny of the catcher fields and the amount of out crossing combined with pollen transport was calculated and fitted in experimental models.

The first results obtained during 2 years of field trials with different field designs, the specifics of each design as well as the problems encountered during the trial will be discussed in a posttrial evaluation and a design will be proposed that will be the key element in a multilocation evaluation to be carried out.

Prior to the cultivation of transgenic plants by farmers, the risk of releasing such plants in open fields needs to be assessed. Research has been carried out in order to estimate the possibility of transfer of an herbicide resitance gene from a genetically modified Brassica napus to different wild speices by cross hybridization. Production of F1 hybrids by manual pollination involved reciprocal interspecific crosses between a transformed rapeseed resistant to Basta and 5 wild species. Most of the hybrids showed a triploid structure, except for 3 amphidiploids. Chromosome pairing and dertility of the hybrids were assessed. Presence of the gene in the hybrid genome was checked by polymerase chain reaction (PCR) and its expression controlled in vivo by a resistance test. When the gene was present, it was expressed in all but 2 plants.

F1 hybrids were produced by spontaneous pollination, using fertile transgenic rapeseed. Wild species (Sinapis arvensis, Raphanus raphanistrum) were grown in insect free cages in mixture with gransgenic Basta resistant oilseed rape, in the presence of a bee hive. Only S arvensis seeds were analysed for Basta resistance. No hybrid was found.
Spontaneous pollination using male sterile rapeseed involved sowing nontransgenic male sterile rapeseeds in alternative rows with wild species (B adpressa, R raphanistrum). Fertility and chromosome number of the hybrids were assessed.

Backcrosses (BC) on the hybrids from manual pollination have been carried out. Using embryo rescue, number of BC1 obtained ranged from 0.1 to 0.4 plants per 100 cultivated ovaries. No seed was obtained on the amphidiploids after BC, except with B oleracea. Selfing of the B napus x B oleracea amphidiploids gave 8.3 to 9.9 seeds per pod. Chromosome number and fertility of the BC1 were assessed. PCR and in vivo tests revealed that when the gene is present in the BC1, it is expressed. For 1 plant the gene was lost. With respect to spontaneous pollination, hybrids bewee n male sterile or transgeic rapeseed and S arvensis, R raphanistru, B adpressa were grown with the corresponding wild species in insect free cages. 1, 10 and 272 seeds were harvested on the hybrids respectively. Morphology of the BCs obtained was intermediary between the wild and the crop parents.
Hybrids between male sterile rapeseed and wild species (B adpressa, R raphanistrum) were sown on alternative rows with the corresponding wild species. Work is in progress to characterize fertility, chromosome number, presence and expression of the Basta resistance gene on the BC1 obtained.

Accurate prediction of the level of gene expression in transgenic plants is a major requirement for assessing thebiosafety of such plants. The variability in transgene expression between individual transgenic plants carrying the same transgene(s) is enormous and is thought to be caused by the random place of integration of the transgene in the plant genome. This position effect is severely hampering both biosafety analyses of transgenic plants as well as studies into the regulation of gene activity.

Gene expression is thought to involve the formation of chromatin loops that are bound to the nuclear scaffold by scaffold attachment sites (SAR). In animal systems, the addition of SAR sequences has been shown to decrease the variability of gene expression. It is assumed that SAR sequences render the introduced genes less dependent upon the particular place of integration. Recently, plant SAR sequences. Also, a decrease of variability of reporter gene expression was suggested in plant callus. The addition of SAR sequences may therefore contribute to a higher predictability of transgene expression.

Using chicken (A element), a series of plant vectors carrying these SAR sequences around the beta-glucuronidase (GUS) reporter gene have been constructed. To investigate the influence of SAR sequences on gene expression during plant development, the reporter gene was driven by the light regulated promoter from the potato 1haC gene. This gene encodes a protein of the light harvesting complex of Photosystem I. Transgenic tobacco plants were obtained using Agrobacterium mediated leaf disc transformation and were analyzed for GUS activity in the leaves.

The presence of the A element resulted in a significantly higher overall GUS activity, depending on the distance between A element and reporter gene, indicating an enhancer effect of the A element.

No obvious reduction in variability of GUS activity was observed. The distribution of gene expression over the population wa s markedly skewed and different from the normal distribution, so most statistical tests for variance were inappropriate. Using nonparametric tests, the 3 populations were compared with respect to distribution and variance of gene expression. Whereas the distribution of gene expression is significantly altered as a result of the presence of A elements, any reduction of the variability in gene expression was more difficult to approach statistically.

Several experiments and surveys concerning the ecological effect of growing glyphosate tolerant sugar beets (Beta vulgaris) have been conducted during the growth season of 1991 and 1992. The following subjects have been studied:
pollen disperal from sugar beet to wild relatives;
competitive ability of the glyphosate tolerant beet;
observations of Beta maritima on 4 wild habitats describing flowering period and phenological diversity;
analysis of the weed beet population in Danish beet fields, origin, ploidy, phenological diversity and seed production.

Pollen dispersal from Beta vulgaris to wild relatives:
1 pollination experiment with Beta maritima was conducted in a beet field. Red beet (Beta vulgaris variety condivita) was used as a marker plant to simulate outcrossing from a transgenic diploid sugar beet variety. Beta maritima was planted varying distances from a red beet pollen source. 5 smaller experiments have been conducted all with red beet as the pollen source and 3 different wild beet species (Beta atriplicifolia, B macrocarpa and B maritima) as receptor plants.

Competitive ability of the glyphosate tolerant beet:
The competitive ability of an ordinary sugar beet variety was measured relative to 2 other agricultrual crops, barley (Hordeum vulgare), and flax (Linum usitatissimum). The cross between a transgenic sugar beet and a Beta maritima have been tested against the parental types in a field trial.

Observations of Beta maritima on 4 wild habitats, concerning flowering period and phenological diversity:
A morphological survey of Beta maritima was conducted on 5 isolated coast lines on the western part of Sealand, Denmark, to establish knowledge of flowering and genetic diversity. The populations were surveyed 4 times during the growth season.

Analysis of the weed beet population in Danish beet fields, origin, ploidy, phenological diversity and seed production:
During the growth season 5 different sugar beet fields, all severley infest ed with weed beets were observed. The origin of the weed beets was examined by isoenzyme analysis and phenological observations, while ploidy was examined by flow cytometrical methods. All weed beets were diploid.

The transfer of introduced genes to wild relatives is one of the questions raised by the use of genetically modified sugar beet. While the significance of such a gene transfer must be evaluated on a gene by gene basis, a study of gene transfer using a genetic markers can provide useful information on sugar beet pollen movement and hybridization with related species under field conditions.

The first field experiment with flowering transgenic sugar beet was conducted in Belgium in 1991. The objective of the study was to test experimental protocols designed to study sugar beet pollen movement and hybridizations with Beta maritima. The pollen source was a sugar beet breeding line, containing genes for both Roundup tolerance and beta-glucuronidase (GUS) activity. Pollen movement was monitored by the use of single row plots of male sterile Beta vulgaris plants planted downwind form the pollen source, at distance of 25, 50 and 75 m. Mixtures of Beta maritima genotypes were planted in 5 directions at the same distance as the male sterile Beta vulgaris. Plants were harvested individually and analysed. Germination levels were generally low, and 3250 GUS assays were performed in total.

The analyses indicated that expression of GUS in the segregating pollen source was 57%, lower than the expected maximum or 75%, possibly because of hybridization with the nontransgenic segregants. On the CMS plants at distance zero, 48% of the germinating seeds were GUS positive (50% expected). The comparable frequency of outcrossing decreased with distance from the transgenic pollen source, to 14%, 5% and 3% of the germinated seeds at distances of 25, 50 and 75 m respectively. The level of germination of seeds harvested for Beta maritima plants was extremely low. Of more than 15 000 seeds harvested at distances of greater than 25 m, some 1000 seeds germinated. Only 2 of these seeds were confirmed to be GUS positive; 1 located at 25 m from the pollen source and the other at 75 m.

One of the main problems that might arise from the release of genetically modified crops is the accidental integration of the resistant gene into the genome of wild relatives. These plants could then become resistant weeds in the field.

Computer simulations have been used to compare the risks of resistance disperal and gene transfer for a large range of resistant crops differing in their biological behaviour and for 3 resistance genetic determinisms.
To study the effect of each parameter of the model on the short term and the long term, the frequency of resistant weeds was noted in the field and the field margin after 5 years and after equilibrium.
The goal of the model is to estimate what mechanisms and parameters influence most the escape of the gene. The results provide information on the most risky situations and point out the most valuable experiments for risks assessment.
The most intuitive result provided by the model is that the parameter which controls the possibility of hybridization between wild and cultivated plants has an opistatic effect over all ther other parameters. For nearly any set of parameters the resistance will be absent in the wild populationif the hybridization between the 2 species is very low. The effect of the genetic determinism of the resistance, of a genetic linkage between domestication characteristics and the resistance and of biological characterics of the plant were also studied.

The model also points out that complex interactions between parameters may exist. As a consequence, it will be impossible to make predictions for a 'new' resistant crop whose biological behaviour differs from already studied species, until many parameters have been precisely evaluated. For this reason experimental work and modelling both profit from being done simultaneously.

The environment consequence of releasing transgenic plants to unconfined conditions depends on the changes in survival rate, growth behaviour and hybridization possibilites caused by the transformation allowing the transformed plant to invade new habitats. Although this is unlikely it is thought that impoved competitiveness of the transgenic plant may facilitate a change in mixed populations towards the transgenic genotype. An assessment of growth behaviour and competitiveness under a range of growth conditions should quantify the extent to which the transformation has changed important characters.

Sugar beets (Beta vulgaris subspecies vulgaris) are grown in all parts of Denmark, in a total area of approximately 170 000 ha 66 000 ha for sugar production and 104 000 ha for fodder use. On the coastlines of the central parts of Denmark the wild relative of sugar beet, Beta vulgaris subspecies maritima, has its natural habitat. The cultivated sugar beet is not found on these coast areas or any other natural habitats in Denmark. In the cultivated fields wild types are occasionally found. It is believed that these wild types may consist of crossings between the cultivated and the wild genotypes, as subspecies vulgaris and subspecies maritima hybridize easily.

A competition experiment involved 3 Beta subspecies: hybrids between transgenic subspecies vulgaris and nontransformed subspecies maritima, nontransformed subspecies vulgaris and nontranformed subspecies maritima. These were planted in binary combinations in greenhouse and field trials. Each of the 3 species combinations consisted of 11 different density combinations, with 3 monocultures of each subspecies and 5 mixed populations. All combinations were harvested at 4 different harvest times both in the greenhouse and in the field. A competition model that characterizes the yield potential and plant weight at low densities, as well as a relative measure of competitiveness of the genotype compared to that of the 2 other genotypes were used to quanfity the growth behaviour and competitiveness of all 3 subpecies. The experiment concurrently illustrates the dynamics of the above parameters over time.

The environmental consequence of releasing trangenic plants to unconfined conditions depends on the changes in survival rate, growth behaviour and hybridization possibilities caused by the transformation. Survival rate depends on the growth conditions and the competiveness of the plant determine the success as an invader. Fundamental changes in growth behaviour may allow the plant to invade new habitats not formerly occupied by the nontransformed genotype, but more likely, the growth behaviour is only slightly modified and the tranformed plant is limited to the same habitats as the nontransformed genotypes.
An assessment of stress tolerance towards limiting growth factors is important in the understanding of the observed difference in resource uptake and allocation within and between different genotypes in mixed populations. Sugar beet, Beta vulgaris, was used as a model plant, and the behaviour of different subsepecies and hybrids were related to well known crop plants such as barley, Hordeum vulgare, and oilseed rape, Brassica napus.

The following Beta subspecies and hybrids were used to assess stress tolerance: transgenic sugar beet (subspecies vulgaris), nontransformed subspecies vulgaris, nontransformed wild beets (subspecies maritima) and hybrids between transgenic subspecies vulgaris and nontransformed subspecies maritima. The Beta subspecies, barley and oilseed rape were grown under optimal growth conditions in greenhouse. Additional to this optimal treatment all genotypes were grown under suboptimal light conditions and suboptimal temperatures. The experiment included 4 harvest times with 2 weeks intervals. The differences in stress tolerance from this experiment will be related to the growth behaviour and competitiveness of the same species and genotypes.

In 1992 it was not possible to test transgenic plants in natural habitats in Denmark. 2 locations of natural habitats with wild Beta subspecies were used to test the establishment, growth behaviour and survival of transplanted nontransformed Beta subspecies, from the preliminary trial will be used in 1993 to perform experiments including transgenic plants.
The participants assess the entire range of safety issues associated with the deliberate release in the environment of genetically modified plants (GMP), specifically, two model crop species, oilseed rape and sugarbeet.

In addition to earlier defined aspects in a number of BAP-projects - pollen dispersal and outcrossing to wild relatives - descriptions are provided on seed dispersal, population dynamic behavior of GMP, hybridity and viability of outcrosses, the stability of GMP and the evolution of the expression of the transgenes over generations and time.

The main assessments are performed in the field under natural agricultural conditions. Suitable general procedures and techniques are evaluated and the globalization of the defined parameters is achieved in a definition of a safety prediction model, relating laboratory, greenhouse and field data.

Funding Scheme

CSC - Cost-sharing contracts


Plant Genetics Systems NV
22,Jozef Plateaustraat
9000 Gent

Participants (8)

1,Droevendaalsesteeg 1
6700 AA Wageningen
Danisco A/S
4,Højbygård Vej
4960 Holeby
Den Kongelige Veterinær- og Landbohøjskole
2630 Taastrup
Domaine De La Motte
35653 Le Rheu
United Kingdom
Norwich Research Park, Colney
Monsanto Europe SA
270-272,Avenue De Tervuren
1150 Bruxelles
The National Forest and Nature Agency- Ministry of the Environment
2970 Hørsholm
Université de Paris XI (Université Paris-Sud)
Centre D'orsay
91405 Orsay