GENE FLOW IN SUGAR BEET PRODUCTION FIELDS
Henri Darmency & Marc Richard-Molard
April, 2008

Concern has grown in Europe over the agricultural and environmental impacts of genetically engineered (GE) crops, especially about gene flow to conventional varieties and wild relatives. Contrary to most crops that are grown for their seed or fruit, sugar beet (Beta vulgaris L.) is grown for its root. Therefore, pollen-mediated gene flow in root production areas is not a concern in the debate on GE and non-GE crop co-existence, including table beet and chard in private gardens. However, pollen flow could be responsible for the admixture of GE and non-GE materials in the seed lots provided to farmers; but it can be easily prevented by breeding and multiplying GE and non-GE seeds in different regions. In contrast, pollen flow to wild relatives could certainly generate agronomic trouble much more easily and quickly than with any other crop.

There is always a small proportion of sugar beet plants that flower, in spite of being a root crop, because of either their sensitivity to vernalization or the presence of a dominant bolting gene. Since wild sea beet, weed beet, and sugar beet are the same botanical species, and since they are allogamous, they can easily produce hybrids and are completely interfertile¹. The progeny of crosses among these beet types is perfectly adapted to field conditions, has no hybrid fitness cost, and therefore can display the favorable traits encoded by the transgenes. In particular, herbicide resistance, which is desired by farmers to significantly reduce the number of herbicide sprays and working hours, could result in the spread of herbicide-resistant weed beets. The weed beet is already a serious problem to sugar beet growers since no herbicide available differentiates between weed beet and sugar beet. Severe infestations can reduce to nothing the sugar beet yield and lead the farmer to stop growing this crop. Therefore, breeding transgenic herbicide-resistant varieties can solve this particular problem, provided that gene flow is contained for a long time.

The question of pollen-mediated gene flow to weed beets in sugar beet root production areas was addressed in a six-year farm-scale study. Since GE sugar beet is not yet grown on commercial fields, this study is the only documented background of field growth for that crop. It was a joint action of governmental research institutes (INRA), professional associations (ITB, for beet; CETIOM, for oilseed rape) and industry (Hilleshog and KWS, for providing the GE beets). The program started in 1995 in two locations: Châlons, in Champagne, and Dijon, in Burgundy, in the north-eastern part of France. Besides the gene flow study, the trial confirmed that the number of herbicide sprays was reduced from 4.1 to 2.5 thanks to the post-emergence, non-selective herbicides used for the GE varieties, without any difference in weed control or yield than with conventional herbicide programs².

Field monitoring
In each location there were four 1-ha adjacent fields grown in rotation with GE sugar beet, GE oilseed rape, conventional wheat and fallow. The sugar beet field was divided in two parts, each sown with a heterozygous herbicide-resistant line: a Roundup Ready glyphosate-resistant line from Hilleshog, and a Liberty Link glufosinate-resistant line from KWS. Each variety was sprayed with its respective herbicide, except for the central lane of the field that was sprayed with various herbicides used on conventional sugar beets in the region. Weed beets were transplanted into the central lane of the field when no local weed beet emerged in the sugar beet field. The monitoring took place between 1996 and 2001. More detail is available in the full publication³.

The number of sugar beet bolters varied widely according to the transgenic line and year, from 0 to 121 per ha, thus providing a good opportunity to study the consequences of pollen flow under a wide range of realistic conditions and amounts of pollen escape. Indeed, the rate of sugar beet bolters of commercial varieties in France during the same period ranged from 0.001 to 0.1%⁴. These flowering plants included vernalized herbicide-resistant sugar beets and susceptible annual hybrids coming from pollination of the seed mother plants by susceptible wild beets surrounding the nursery. Susceptible hybrids and spontaneous weed beets could grow and flower in the central lane of the field, which was treated with selective herbicides. All the bolting plants were mapped, and their seeds were carefully collected and then tested for herbicide resistance in the greenhouse.

Production of herbicide-resistant seeds by sugar beet bolters
On average, 58.2% of seeds produced by resistant sugar beet bolters were resistant. The deviation from the 75% expected from mating among heterozygotes denoted both the contribution of pollen coming from susceptible beets and probably the better viability and fertilizing ability of the pollen of weed beets. This category of plant accounted for 84.8% of the total resistant seed production over the years studied, but, as shown in Figure 1, their importance decreased during the second round of the rotation.

Resistant seedlings also appeared in the progeny of susceptible sugar beet bolters at a mean percentage of 1.7%, and they accounted for 1.2% of the total resistant seed production over the six years under study.

Production of herbicide-resistant seeds by weed beet
On average, 6.2% of the seeds produced by weed beets were resistant, which accounted for 14% of the total resistant seed production over the six years in the two locations. A more detailed analysis has been published³. Some of these resistant seeds were not produced within the sugar beet field, but rather in the fallow field, and their proportion decreased as the distance between the fields increased. Fallow field production represented 0.2% of the total resistant seed production. The largest distance at which a cross was recorded between the GE sugar beet bolters and a weed beet was 112 m. This showed that a foreign pollen grain entering a pollen cloud at low frequency over a weed beet population that grows in a distant fallow field has an effective fertilizing ability. Pollen flow monitoring using male sterile plants within and around the farm scale experiments showed fertilization at 277 m and up to 1172 m^3,5. The distribution curve of the number of fertilized seeds in terms of distance from the pollen source had generally a negative power shape^3,5,6. For instance, the number of resistant seeds recorded in different groups of plants in 1999 in Châlons, up to 120 m away from the resistant bolters, followed the equation N = 14.4 d^-0 ,75, R² = 0.89 (Fig. 2).

Thus, in spite of the low density of GE bolters (8 in one ha in 1999), pollen flow can reach adjacent as well as distant weed beet populations and transfer the herbicide-resistance gene. Within the sugar beet field, there was an average of 3.5% resistant seeds in the seed produced by the susceptible weed beet, accounting for 7.4% of the total resistant seed production over the six years and the two locations. The weed plants that produced resistant offspring were not located at shorter distances from the resistant bolters or at farther distances from susceptible plants than the other weed beets. They flowered simultaneously with other weed plants, but they produced more flowers and 2.4 times more viable seeds than plants that did not produce resistant offspring. Higher production of flowers and higher seed sets could partly explain the ability of those plants to catch more numerous pollen grains and mature more embryos, thus simply having a higher probability of producing resistant offspring. However, if the number of viable seeds per plant somewhat depended on a genetic factor, the consequence would be the propagation of herbicide resistance together with the most reproductive individuals, which would be a very unfavorable conjunction of factors with respect to weed control.

Finally, resistant seeds could also originate from resistant weed beets. In 2000 and 2001, a few resistant weed beets emerged, either spontaneously from the soil seed bank containing seeds left to shed in the same field in 1999, or sowed in the field in order to simulate the creation of a soil seed bank containing herbicide-resistant seeds, as would have occurred if the seeds had not totally been harvested in former years. All these resistant weed beets were heterozygotes and produced, on average, 74% resistant seedlings. This category of plants accounted for 6.4% of the total resistant seed production, but unlike resistant seeds produced by sugar beet bolters, it was concentrated in the last two years, during the second crop rotation (Fig. 1).

Management
At the end of the first round of crop rotation, the cumulated number of seeds released on the farm-scale trial (4 ha x 2 locations) was 222,000, of which 22.3% were herbicide-resistant, representing 0.6 resistant seed per m². However, there was a large variability among fields. Most resistant seeds were produced by resistant sugar beet bolters (see Fig. 1). This result strengthens the urgent need to eradicate all transgenic bolters. On one hand, eradication can be achieved through production of high quality certified seeds of varieties that are not sensitive to vernalization and free of annual hybrids. On the other hand, destruction of bolters should also be pointed out as a compulsory task among farmers' good agronomic practices. Clearly, some of the bad results reported above belong to a worst-case scenario, because most farmers would have reduced the risk of seed release and pollen flow by destroying bolters when they were too numerous.

However, if bolting still occurs, even at very low frequency, and is not destroyed by farmers, or if transgenic volunteer roots grow and flower in crops subjected to the same herbicide or in fallow fields, the transgenes will unavoidably be transmitted to weed beets within a short period of time. Pollen flow from resistant sugar beets to susceptible ones and to weed beets outside the field accounted for 0.2% of resistant seeds in the farm-scale study. Subsequent multiplication of resistant weed beets in the second round of crop rotation accounted for 13.6% of resistant seeds. These seeds would be the source of further multiplication of herbicide resistant weed beets. Therefore, farmers must prevent the constitution of a soil seed bank containing herbicide-resistant seeds. Useful practices, besides destruction of bolters, could include management of fallow fields to control weed beets and change of crop rotation. The effect of various farming systems on gene escape from GE crops to volunteers and weed beets could be anticipated by simulation models fed with basic data on weed beet biology, such as those collected in the farm scale study^7,8.

Henri Darmency
Unité Mixte de Recherche sur la Biologie et la Gestion des Adventices
INRA, BP 86510, 21065 Dijon, France
Darmency@dijon.inra.fr