MEC1109.fm Page 2035 Saturday, November 11, 2000 2:47 PM

Molecular Ecology (2000) 9, 2035 – 2040

Monitoring gene flow from transgenic sugar beet using cytoplasmic male-sterile bait plants

Blackwell Science, Ltd

C H R I S T I A N E S A E G L I T Z , * M AT T H I A S P O H L † and D E T L E F B A RT S C H * *Department of Biology V, Ecology, Ecochemistry and Ecotoxicology, Aachen University of Technology — RWTH Aachen, Worringerweg 1, 52056 Aachen, Germany, †TÜV-Hannover/Sachsen-Anhalt eV, Division of Energy and Systems Technology, Services in Biotechnology, Am TÜV 1, 30519 Hannover, Germany

Abstract One of the most discussed environmental effects associated with the use of transgenic plants is the flow of genes to plants in the environment. The flow of genes may occur through pollen since it is the reproductive system that is designed for gene movement. Pollen-mediated gene escape is hard to control in mating plants. Pollen from a wind pollinator can move over distances of more than 1000 m. To investigate the efficiency of transgenic pollen movement under realistic environmental conditions, the use of bait plants might be an effective tool. In this study, cytoplasmic male-sterile (CMS) sugar beets were tested with regard to their potential for monitoring transgene flow. As the pollen source, transgenic sugar beets were used that express recombinant DNA encoding viral (beet necrotic yellow vein virus) resistance, and antibiotic (kanamycin) and herbicide (glufosinate) tolerance genes. In a field trial, the effectiveness of a hemp (Cannabis sativa) stripe containment strategy was tested by measuring the frequency of pollinated CMS bait plants placed at different distances and directions from a transgenic pollen source. The results demonstrated the ineffectiveness of the containment strategy. Physiological and molecular tests confirmed the escape and production of transgenic offspring more than 200 m behind the hemp containment. Since absolute containment is unlikely to be effective, the CMS–bait plant detection system is a useful tool for other monitoring purposes. Keywords: Beta vulgaris, biosafety, containment, cytoplasmic male-sterile bait plant, gene escape, PCR marker, transgenic sugar beet Received 23 March 2000; revision received 18 July 2000; accepted 20 July 2000

Introduction The use of molecular techniques in plant breeding offers the possibility of developing transgenic plants containing traits of agronomic importance. In addition to the benefits of this technique, the ecological implications of the spread and loss of control of transgenic plants require investigation (Tiedje et al. 1989). One concern is the hybridization of transgenic cultivars with wild relatives and the successive introgression of transgenic traits into the gene pool of wild plant populations (Ellstrand & Hoffman 1990). Gene flow is very likely to occur through pollen dispersal since reproductive organs are intended to create gene

Correspondence: Christiane Saeglitz. Fax: +49 241 8888 182; E-mail: [email protected] © 2000 Blackwell Science Ltd

movement and therefore crossing of transgenic plants with wild plants is hard to control. Pollen from windpollinators can easily move over more than 1000 m (Barocka 1985), and, if the transgenes confer a selective advantage, they could then change the invasive character of the recipient plant. There are known examples from classical breeding indicating that gene flow contributes to the development of new weedy forms (Ellstrand et al. 1999; Marvier et al. 1999). On the other hand, gene flow may also contribute to the decrease of genetic diversity, if the transgene acts as a ‘pollutant’ by decreasing population fitness. After the hybridization and introgression of unfavourable genes into natural populations, a decline of the effective population number might lead to a loss of rare alleles. Hybridization between existing crops and their wild relatives is quite common, 12 of the 13 most important crops worldwide grow sympatrically with

MEC1109.fm Page 2036 Saturday, November 11, 2000 2:47 PM

2036 C . S A E G L I T Z E T A L . crossable relatives somewhere in their cultivation area (Ellstrand et al. 1999). One interesting case is the potential for gene escape from sugar beet (Beta vulgaris ssp. vulgaris var. altissima) to its wild relative sea beet (Beta vulgaris ssp. maritima, Bartsch et al. 1999). Although sugar beets are biennial plants that are harvested before reaching the reproductive stage, early flowering is unavoidable through occasional vernalization at low spring temperatures and genetic factors inducing flowering during cultivation (Boudry et al. 1993; Bartsch et al. 2001). Pollen grains are small and produced in large quantities, which makes them hard to control. Since transgenic sugar beet plants and conventional cultivars hybridize with wild beet (Bartsch & Pohl-Orf 1996), the monitoring of pollen-mediated gene flow is an important task of biosafety research. There has long been a need to prevent sugar beet seed production from unintended pollen flow. Sugar beet breeders have used hemp stripes of Cannabis sativa L. for small-scale seed production plots of 1 × 3 m (Barocka 1985), whereby they protect their sugar beet breeding lines from unintentional pollination with other (cultivar) sources. Hemp is mostly effective against wind-mediated pollen flow because of its height (4 m) and sticky leaf surface. The hemp containment strategy was also applied for the field release at the Aachen University of Technology (RWTH) in 1998 to limit the pollen escape from flowering transgenic sugar beet in larger plots (25 × 25 m). The first goal of this study was to test the suitability of cytoplasmic male-sterile (CMS) mother plants for monitoring transgenic pollen flow. These plants are usually used as pollen acceptors for the large-scale seed production of sugar beet. Cytoplasmic male sterility is a maternally inherited trait characterized by the inability of the plant to produce functional pollen (Owen 1945). Since CMS plants are not able to self-pollinate, the male parent for the offspring must be fertile. Analyses of the offspring can provide data on gene transfer into the environment. The second goal was to test the effectiveness of containment strategies that aim to limit gene escape via pollen. In order to investigate the correlation between the wind direction and pollen spread, the bait plants were placed at different positions surrounding the pollen source.

(Velten et al. 1984), and the bar gene (phosphinothricintolerance gene) from Streptomyces hygroscopicus mediating tolerance to the herbicide LIBERTY® (Aventis Crop Science, Lyon, France) (Thompson et al. 1987). All pollen donor plants were homozygous for the transgenes and fertile. Since the transgenic alleles are genetically dominant, heterozygous F1 hybrid offspring from pollination of non-transgenic plants were expected to be phenotypically herbicide-, virus- and antibiotic-tolerant. The plants used as bait were CMS. All plants were provided by the German breeding company KWS Saat AG (Einbeck, Germany).

Field trial design The field design is described in Fig. 1. The background of the overall test was a series of over-wintering experiments with transgenic sugar beet as described by Pohl-Orf et al. (1999). The CMS bait plant monitoring experiment was carried out during the 1998 growing season (May– September) at a field site in Aachen, Germany. Fifty-six bait plants were positioned surrounding a field site containing 30 flowering transgenic plants as pollen donor source: 12 bait plants were placed 9 m from the pollen source but inside a hemp containment strip, 12 plants were placed just outside the hemp barrier, and the remaining bait plants were placed in eight different directions (north, north-west, west, south-west, south, south-east, east, northeast) at distances of 50, 100, 200 and 300 m from the transgenic donor plants. They were planted in 12 litre pots. After the end of flowering, plants were brought into a greenhouse for seed ripening. This strategy was used in order to avoid seed loss on the field.

Screening progeny for herbicide resistance After harvest at the end of the vegetation period in September, seeds were air-dried for 4 weeks, sown in a greenhouse and screened for transgenic attributes. The frequency of gene flow from the transgenic plants was calculated as the percentage of individuals surviving herbicide treatment in comparison to the total amount of seedlings (originating from a single bait plant). The herbicide treatment was an application of a 1% (v/v) solution of the commercially available LIBERTY® to the plants.

Material and methods Screening progeny at the DNA level Plant material The transgenic modification of the pollen donor plants was as described previously by Mannerlöf et al. (1996). The plants carried the cDNA for the coat protein of beet necrotic yellow vein virus (cpBNYVV gene) under the control of the CaMV 35S promotor (Meulewater et al. 1989), the nptII gene as a tolerance marker against kanamycin

The individual herbicide tolerance was confirmed at the DNA [polymerase chain reaction (PCR) ] and protein (ELISA) levels. Plants that survived the herbicide treatment were examined for the presence of transgenic DNA sequences (cpBNYVV, nptII, bar) by PCR. Approximately 0.5 g leaf tissue (primary leaves) of young seedlings were frozen immediately after harvest in liquid nitrogen and © 2000 Blackwell Science Ltd, Molecular Ecology, 9, 2035 – 2040

MEC1109.fm Page 2037 Saturday, November 11, 2000 2:47 PM

T R A N S G E N I C P O L L E N M O N I T O R I N G 2037 Fig. 1 The experimental design consisted of a central plot of 30 transgenic sugar beets (X) surrounded by 12 non-transgenic CMS plants (numbers 1–12). A hemp containment is planted around the inner field. In the outer field, 12 cm bait plants were planted next to the hemp strip (numbers 13–24) and four CMS plants were planted at each direction of the compass as shown and at different distances from the central plot (50, 100, 200, 300 m). The percentage of transgenic progeny at each position is given. n.d., not detected (lost to animal feeding). A source of non-transgenic pollen by seed production from 60 sea beet plants is located near point W50 (Z).

Table 1 Wind direction during the flowering season

Month

Wind direction

Number of days

Average wind speed (Beaufort scale)

July

South-west North-west East

23 5 0

3 1.5–2.5 —

South-west North-west East

14 8 0

3 2–3 —

August

pulverized. Genomic DNA was extracted from leaf tissue using the salt-extraction procedure of Aljanabi & Martinez (1997). PCR reactions were performed according to the procedure outlined by Saiki et al. (1988). Primers and amplification conditions are given in Table 1. Each reaction contained approximately 50 ng of template DNA. The PCR amplification was performed in 25 µL reactions with 35 cycles. DNA from non-transgenic plants served as a © 2000 Blackwell Science Ltd, Molecular Ecology, 9, 2035–2040

negative control. The parental transgenic lines were used as positive control. PCR products were separated in a 1.5% agarose gel electrophoresis and documented after ethidium bromide staining.

Screening progeny at the protein level Expression of the cp gene in the progeny of the recipient plants was quantified by ELISA as reported by Kœnig et al. (1987) using a detection kit from Loewe Biochemica (Sauerlach, Germany). BNYVV coat protein concentration in leaves was analysed by crushing 1 g of primary leaves in a 1:10 ratio with sample buffer. Each extract was tested twice.

Results Seed production of bait plants Although incapable of self-fertilization, the bait plants produced a significant amount of seed offspring. The

MEC1109.fm Page 2038 Saturday, November 11, 2000 2:47 PM

2038 C . S A E G L I T Z E T A L . Fig. 2 PCR amplification of the cp gene of each position with transgenic progeny (see Fig. 1). Outer lanes: M, Eco47I ladder; N, negative controls, P, positive control; other lanes, individual offspring from bait plants according to Fig 1.

Primer sequences

Table 2 Primer sequences and resulting amplification products

PCR product length (bp)

Gene sequence

cp1 cp2

5′-CTA TTG TCC GGG TGG ACT GGT TCT-3′ 5′-ATG ACA TGG AAG GAT ATG TCA CAT-3′

547

cp gene

npt1 npt2

5′-GTG-GAG-AGG-CTA-TTC-GGC-TA-3′ 5′-CCA-CCA-TGA-TAT-TCG-GCA-AG-3′

550

nptII gene

bar1 bar2

5′-TCA GAT CTC GGT GAC GGG CAG G-3′ 5′-GTC AAC CAC TAC ATC GAG ACA A-3′

480

bar gene

mean number of seed produced was approximately 1300 per plant (range 10 – 5900). Ten of the bait plants (5, 12, 14, 17, 19, 23, 24, SW50, S200, NW50) (according to Fig. 1) were lost to animal feeding before seed harvest.

Seed analyses for transgenic attributes In the offspring, 20% of the seedlings were tolerant to phosphinothricin application, indicating them as potential transgenic offspring. This was confirmed by PCR analysis for the presence of cpBNYVV, nptII and bar genes (Fig. 2). In addition, the translation of DNA markers was detected by ELISA at the protein level. All plants of the transgenic progeny expressed the coat protein gene of BNYVV at significant levels.

Monitoring the out-crossing frequency at the Aachen field site in 1999 Analysis of the bait plant offspring demonstrated dispersal of transgenic pollen inside and outside the hemp containment stripes. This was shown by PCR of the transgenic offspring of 31 of the 45 bait plants (Fig. 2). The progeny from 14 (31%) of bait plants were not pollinated by any of the genetically modified pollen donors. As expected, out-crossing was highest within the hemp containment, with rates up to 80% on the eastern side of the pollen source. These data correlated with the principle wind direction from west to east (see Table 2). Significant out-crossing was observed outside the hemp isolation area, with rates between 0.5% on the west © 2000 Blackwell Science Ltd, Molecular Ecology, 9, 2035 – 2040

MEC1109.fm Page 2039 Saturday, November 11, 2000 2:47 PM

T R A N S G E N I C P O L L E N M O N I T O R I N G 2039 side and 40% on the east at a distance of 200 m from the transgenic pollen donor plants. However, seeds collected from plants within the hemp containment area were not completely transgenic, indicating that non-transgenic pollen from flowering plants outside the boundary passed the hemp containment to the inner side and pollinated the bait plants.

Discussion The use of CMS bait plants proved to be a useful monitoring instrument for in vivo detection of gene escape through wind-distributed pollen. Seeds produced on CMS bait plants revealed out-crossing of transgenic traits. The advantages of using bait plants over physical traps such as sticky surfaces, which only indicate the potential distribution of pollen grains, include detecting the viability of the pollen. A further advantage of the bait plant system is the opportunity to characterize the resulting offspring. Since self-pollination of the bait plants was excluded, out-crossing events were much more frequent than would be the case in investigations with fertile trap plants. This effect may help detect rare pollination events that would be hidden under natural conditions. Therefore, this experiment does not mirror the natural conditions in the field where transgenic pollen has to compete with the pollen produced by the potential recipient plant. The data represented in Fig. 1 also demonstrate that the background pollen flow was much higher than the pollen flow from the transgenic source, since most of the offspring seeds were non-transgenic. Inflow of airborne beet pollen is a common phenomenon (Meier & Artschwager 1938; Archimowitsch 1949; Darmency 1996). The nearest sugar beet field without bolter control was more than 1 km from the test site. At this distance, gene flow rates are less than 1% based on the experience of sugar beet breeders, who apply a 1 km isolation distance between seed production of two different cultivar lines (Barocka 1985). The most likely sources for pollen contamination of the bait plants were three small plots (3 × 5 m each) of 60 sea beet plants (B. vulgaris ssp. maritima) 50 m west of the transgenic beet pollen donors (near bait plant W50; Fig. 1). These plants were surrounded by an additional 5 m strip of Cannabis containment. Every developing sea beet flower was supposed to be either bagged in paper bags for seed production or cut off every second day. However, we cannot exclude pollen escape from these plants. As we demonstrated, transgenic pollen can escape and contaminate bait plants despite Cannabis containment stripes. Non-transgenic pollen from other sources is obviously no exception from this phenomenon. This was also demonstrated by other authors (Hokanson et al. 1997). It is only legitimate to report transgene flow as a percentage if the CMS bait plants had fully set seed. Future © 2000 Blackwell Science Ltd, Molecular Ecology, 9, 2035–2040

studies should take into account that lack of pollen leads to an over-estimation of transgene flow rates. The outcrossing pattern of transgenes correlated with the main wind direction (Fig. 1, Table 1), reflecting the known windmediated pollination of beet (Tyldesley 1977). Few reports are available for insect pollination (Archimowitsch 1949; Dark 1971; Free et al. 1975), leaving open the possibility that pollination of bait plants could also have been mediated by insects. Bees were occasionally seen on the plants during the experiment. This kind of pollination could have contributed to the transgene dispersal near the pollen source (e.g. directly behind the hemp stripe). Despite the loss of bait plants by animal feeding, this did not affect the principle conclusion of the experiment. Future experiments should use physical protection to prevent these losses. Pollen represents an ecologically important vector for the spread of genetically modified genes. According to Ellstrand & Hoffman (1990), pollen can act as a route of escape of foreign genes into the wider environment. Although beet pollen flow can be observed over more than 1 km (Van Raamsdonk & Schouten 1997), in the present investigation the distribution was limited to 200 m. The transgenic pollen also has to compete with background pollen flow from different sources. As expected, all bait plants near the pollen source inside the hemp containment had transgenic offspring. However, even outside the hemp, trap plants produced some transgenic seeds, demonstrating the limited suitability of hemp stripes for genetic containment. It was clearly demonstrated that the transgenes were detectable at the DNA, protein and physiological level, which was also reported by Bartsch & Pohl-Orf (1996). Since the herbicide tolerance test is much easier and costeffective than the PCR and ELISA proof, this physiological screen offers advantages for large-scale monitoring. The presence of all three engineered genes in the genome of the progeny of the bait plants was detected by PCR. All the plants that survived the herbicide application proved to be transgenic. This indicates that all three foreign genes were linked and co-inherited. By planting bait plants in movable pots, several unfavourable planting conditions (e.g. stony ground) could be overcome. The bait plants could be used just for the period when the donor plants flower and then be removed to a controlled environment (e.g. a greenhouse). When the maturation process takes place in a greenhouse and not on the field, the chance of transgenic seed dispersal is limited. One disadvantage of the bait plant method is the higher expenditure of time for the analyses of the offspring than with the direct investigation of pollen grains. The DNA content of pollen from a physical trap can be investigated immediately by the use of molecular techniques, which offers analysis results some hours after trapping (Petersen et al. 1996). By using bait plants, the offspring data are not available before seed ripening, germination and analysis.

MEC1109.fm Page 2040 Saturday, November 11, 2000 2:47 PM

2040 C . S A E G L I T Z E T A L . We conclude that the bait plant method is an excellent method for monitoring transgenic pollen escape. Moreover, the results of the present study showed that the containment of flowering transgenic sugar beet with 5 m hemp stripes did not prevent the spread of transgenic pollen. Thus hemp containment stripes are a suitable and wellestablished method for isolation in conventional seed production only to prevent the in-crossing of foreign pollen at a reasonable threshold level.

Acknowledgements We thank the following for assistance during this project: Claudia Morak for her technical assistance, Marc Lehnen and Nico Esser for their help in the field experiment (RWTH Aachen), and Ingolf Schuphan for supporting our project. Many thanks to Bob Kosier and Neil J. Emans for helpful comments. This study was supported by the German Ministry of Science and Technology (grants 0310532 and 0310785).

References Aljanabi SM, Martinez I (1997) Universal and rapid salt-extraction of high quality genomic DNA for PCR-based techniques. Nucleic Acids Research, 25, 4692 – 4693. Archimowitsch A (1949) Control of pollination in sugarbeet. Botany Review, 15, 613 – 618. Barocka KH (1985) Zucker- und Futterrüben. In: Lehrbuch der Pflanzenzüchtung Landwirtschaftlicher Kulturformen, Volume 2 (eds Fischbeck G, Plarre W, Schuster W), pp. 245–287. Spezieller Teil Paul Parey Verlag, Berlin/Hamburg. Bartsch D, Pohl-Orf M (1996) Ecological aspects of transgenic sugar beet: transfer and expression of herbicide resistance in hybrids with wild beets. Euphytica, 91, 55–58. Bartsch D, Lehnen M, Clegg J, Pohl-Orf M, Schuphan I, Ellstrand NC (1999) Impact of gene flow from cultivated beet on genetic diversity of wild sea beet populations. Molecular Ecology, 8, 1733 –1741. Bartsch D, Brand U, Morak C, Pohl-Orf M, Schuphan I (2001) Biosafety aspects of genetically engineered virus resistant plants: performance of hybrids between transgenic sugar beet and Swiss chard as a model for gene flow consequences to wild beet. Ecological Applications, 11 (1), in press. Boudry P, Mörchen M, Saumitou-Laprade P, Vernet P, Van Dijk H (1993) The origin and evolution of weed beets: consequences for the breeding and release of herbicide-resistant transgenic sugar-beets. Theoretical and Applied Genetics, 87, 471–478. Dark SOS (1971) Experiments on the cross-pollination of sugar beet in the field. Journal of the National Institute of Agricultural Botany, 12, 242 – 266. Darmency H (1996) Gene flow between crops and weeds: risk for new herbicide resistant weeds? In: Weed and Crop Resistance to Herbicides (ed. Deprado, R), pp. 239 – 248. Kluwer Academic Press, Dordrecht, The Netherlands. Ellstrand NC, Hoffman CA (1990) Hybridization as an avenue of escape for engineered genes. Biosciene, 40, 438–442. Ellstrand NC, Prentice HC, Hancock JF (1999) Gene flow and

introgression from domesticated plants into their wild relatives. Annual Review of Ecology and Systematics, 30, 539 – 563. Free JB, Williams IH, Longden PC, Johnson MG (1975) Insect pollination of sugar beet (Beta vulgaris) seed crops. Annals of Applied Biology, 81, 127–134. Hokanson SC, Hancock JF, Grumet R (1997) Direct comparison of pollen-mediated movement of native and engineered genes. Euphytica, 96, 397–403. Kœnig R, Burgermeister W, Lesemann DE (1987) Methods for detection and identification of beet necrotic yellow vein virus. In: Proceedings of the 50th Winter Congress of the Institute International de Recherche Betteravieres, pp. 17–22. Brussels, Belgium. Mannerlöf M, Lennerfors B, Tenning P (1996) Reduced titer of BNYVV in transgenic sugar beets expressing the BNYVV coat protein. Euphytica, 90, 293–299. Marvier M, Meir E, Kareiva P (1999) How do the design of monitoring and control strategies affect the chance of detecting and containing transgenic weeds? In: Methods for Risk Assessment of Transgenic Plants III: Ecological Risks and Prospects of Transgenic Plants. Where do we go from here? (eds Ammann K, Jacot Y, Simonsen V, Kjellsson G), pp. 109 –122. Birkhäuser Verlag, Basel. Meier FC, Artschwager E (1938) Airplane collections of sugar beet pollen. Science, 88, 507–508. Meulewater F, Soetaert P, van Emmelo H (1989) Structural analysis of the coat protein gene in different BNYVV isolates. Medelingen Faculteit Landbouwwetenschap Riijksuniversiteit Gent, 54, 465–468. Owen FW (1945) Cytoplasmatically inherited male sterility in sugar beets. Journal of Agriculture Research, 71, 423 – 440. Petersen G, Johansen B, Seberg O (1996) PCR and sequencing from a single pollen grain. Plant Molecular Biology, 31, 189 –191. Pohl-Orf M, Brand U, Drießen S, et al. (1999) Overwintering of genetically modified sugar beet, Beta vulgaris var. altissima Döll, as a source for dispersal of transgenic pollen. Euphytica, 108, 181–186. Saiki RK, Gelfand DH, Stoffel S, et al. (1988) Primer directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science, 239, 487–491. Thompson CJ, Mova NR, Tizard R, et al. (1987) Characterization of the herbicide-resistance gene bar from Streptomyces hygroscopicus. EMBO Journal, 3, 2723–2730. Tiedje JMR, Colwell RL, Grossman YI, et al. (1989) The planned introduction of genetically engineered organisms: ecological considerations and recommendations. Ecology, 70, 298 – 315. Tyldesley JB (1977) Out-crossing in sugar-beet due to airborne pollen. Agricultural Meteorology, 19, 463 – 469. Van Raamsdonk LWD, Schouten HJ (1997) Gene flow and establishment of transgenes in natural plant populations. Acta Botanica Neer Landica, 46, 69–84. Velten J, Velten L, Hain R, Schell J (1984) Isolation of a dual plant promoter fragment from the Ti plasmid of Agrobacterium tumefaciens. EMBO Journal, 3, 2723–2730.

C. Saeglitz is a plant and insect ecologist. D. Bartsch is a plant ecologist with a general interest in plant invasiveness and the ecological behaviour of genetically modified species. His working group for applied biosafety research on transgenic organisms includes a molecular biologist (M. Pohl).

© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 2035 – 2040