When transgenic plants are cultivated near wild species that are sexually compatible with the crop, gene flow between the crop and wild plants is possible. A resultant concern is that transgene flow and transgene introgression within wild populations could have unintended ecological consequences. In order to begin testing for these potential effects, it is necessary to locate and monitor wild populations into which transgenes have escaped. Empirical data on transgene escapes is just beginning to emerge in the scientific literature, and in the November, 2006, issue of Molecular Ecology we presented the first evidence for establishment of transgenic plants in wild populations within the USA.¹ The case involved glyphosate-resistant creeping bentgrass (Agrostis stolonifera L.) plants expressing CP4 EPSPS (5-enolpyruvylshikimate-3-phosphate synthase gene from Agrobacterium spp. strain CP4) transgenes that were found in non-agronomic habitats outside of experimental test plots in central Oregon.

In 2003 under USDA APHIS permit issued to the Scotts Company, flowering of approximately 162 ha of Roundup Ready creeping bentgrass (event ASR368) occurred for the first time in Jefferson County Oregon, USA, within a 4453 ha agricultural bentgrass control area established by the Oregon Department of Agriculture (Fig. 1).

Following the initial ASR368 flowering event, we documented hybridization of Agrostis plants by viable transgenic pollen as far as 21 km beyond the perimeter of the bentgrass control area in central Oregon during 2003.³ In that study, seeds were collected from panicles of sentinel and resident plants that had been placed or were naturally growing outside the bentgrass control area. Collected seeds were germinated in a greenhouse and seedlings were sprayed with the herbicide glyphosate. Survivors of herbicide treatment were tested for production of CP4 EPSPS protein using a lateral flow test strip, and additional molecular tests (PCR and sequencing) were run to confirm the presence of the transgene. Field surveys to assess the establishment of wild transgenic Agrostis pollen-mediated hybrids were not conducted during the initial study. However, in a separate survey conducted by Mallory-Smith et al.,⁴ numerous CP4 EPSPS positive Agrostis stolonifera plants were found in agronomic settings inside the bentgrass control area at locations where either no Agrostis plants were detected in the previous year, or where they had been removed. The volunteers were presumed to be GE seed progeny due to their growth in plowed fields of other crops, in open disturbed spaces, or along irrigation canals near ASR368 fields. The research recently published in Molecular Ecology expanded on our previous work and was driven by our interest in determining whether or not transgenic plants could become established in the environment in non-agronomic habitats. The paper thus focused on locating and phylogenetically identifying transgenic Agrostis plants which became established in wild populations within one year of the initial 2003 test production season.¹

The objective of our field studies was to locate transgenic plants established by either crop X wild hybridization or by crop seed dispersal in wild (native and naturalized) Agrostis populations outside the control area. Leaf samples from 50 or more plants at each location were combined and tested for presence of CP4 EPSPS protein using TraitChek tests (Strategic Diagnostics). Bulk samples that tested positive were sub-sampled in the field to identify specific CP4 EPSPS positive plants. In the laboratory, we utilized PCR and sequencing based approaches to confirm the presence of the transgene in individual samples. Proprietary constraints by Scotts and Monsanto limited our access to ASR368 plants and their potential hybrid progeny. We thus used molecular systematic methods to characterize and identify Agrostis hybrids. To detect F₁ interspecific hybrids among wild transgenics, their parentage was determined by comparing nuclear ribosomal ITS1-5.8S-ITS2 (ITS) and maternally inherited chloroplast trnK intron maturase (matK) DNA sequences.

Fifty-five Agrostis spp. populations were located during field surveys of publicly accessible lands in the study area. Of the species that are sexually compatible with the glyphosate-resistant creeping bentgrass crop, those present in the sampled populations included A. stolonifera, A. gigantea, and A. exarata. A total of 20,400 plant tissue samples were collected from these species for analysis with TraitChek kits for the CP4 EPSPS protein. Approximately 0.04% (9/20,400) of plant tissue samples tested positive for the protein. The nine positive plants that were distributed between six of the surveyed populations were identified as A. stolonifera based on morphology.

Spatial distribution of positive plants was consistent with wind movement as the primary physical mechanism for transport of both pollen and seeds from the ASR368 crop fields to resident populations. Seven CP4 EPSPS positive plants were distributed south and southeast of the control area in the direction of the prevailing winds, while two were established 0.2 km from the southwestern border. All seven plants to the south and southeast were found in mesic habitats specifically preferred by A. stolonifera. Three of these plants were located along banks of main irrigation canals, three were at a pond, and one was found along the bank of a small irrigation canal. The two plants found at the greatest distances outside the control area (3.7 km and 3.8 km to the southeast, respectively) were located along a canal in the Crooked River National Grasslands. The two transgenic plants to the southwest were not associated with a waterway, but were located on a roadside. Resident A. stolonifera were present within all populations where CP4 EPSPS positive plants were found, except the two in closest proximity to the control district. The only Agrostis species located at these two sites was A. gigantea (Fig. 2).

For all nine plants that tested positive for CP4 EPSPS protein in the field, the presence of the engineered construct encoding the protein was confirmed in the laboratory by PCR and DNA sequencing. Comparisons of ITS and matK nucleotide variations between the escaped transgenic plants and various Agrostis reference taxa indicated that both the paternal and maternal parents of the wild transgenic plants were A. stolonifera. While our analyses of ITS and matK DNA sequence data have the capacity to identify recently formed interspecific Agrostis hybrids, none were present in the nine wild transgenic plants that were found established in non-agronomic habitats.

Reichman et al. ¹ present the first evidence for escape of transgenic plants from an engineered crop into wild plant populations within the USA. The distribution and parentage of the wild transgenic plants suggest that six of the plants established in the wild resulted from pollen-mediated gene flow to wild A. stolonifera plants and that three came from dispersed GM seeds. As expected, transgenics were generally found at sites in the direction of prevailing winds. Evidence for feral ASR368 comes from three plants established at two sites nearest the control area where there was an absence of sympatric non-cultivated A. stolonifera plants. These three plants may have resulted from crop seeds that were dispersed by various mechanisms; wind, water, wildlife and/or mechanical means (e.g., vehicles).

Our results demonstrate that even short-term production can result in transgene flow and transgene establishment within compatible wild populations at multiple locations that provide suitable habitat. In this example, transgenic A. stolonifera plants became established up to several kilometers away from the crop source fields after only a single growing season, and the escapes were due to movement of both pollen and crop seeds. When such establishment involves or leads to the formation of hybrids with either full or partial fertility, then transgene introgression into wild populations through backcrossing becomes possible.

Because CP4 EPSPS makes plants resistant to glyphosate herbicide, application or drift of this product could favor continued persistence of the wild plants that now have this new gene. Plant reproduction and further spread could cause the herbicide-resistant plants to persist in wild populations even without additional herbicide use. The long-term fate and ecological impacts of CP4 EPSPS transgenes within wild Agrostis populations in central Oregon remain to be determined.

In August 2006, the US District Court for the District of Hawaii considered allegations that the US Department of Agriculture had violated the National Environmental Policy Act (NEPA) and the Endangered Species Act (ESA) by granting permits for limited field tests of genetically engineered (GE) corn and sugarcane. The judge decided that the USDA’s Animal and Plant Health Inspection Service (APHIS) had violated the ESA by failing to obtain information about any endangered species and critical habitats in the regions of the proposed tests. Turning to the alleged NEPA violations, the judge said that his review of APHIS’ records revealed no evidence of an Environmental Assessment, an Environmental Impact Statement, or an explanation as to why neither study had been required before granting the permits. The judge granted the plaintiffs summary judgment on claims that APHIS’s approval of the permits had violated the ESA and NEPA.

Following the court’s verdict, a spokesperson for APHIS announced that the agency was devising a comprehensive programmatic environmental impact statement to address concerns about its oversight of GE crops. In two cases decided this year, federal court judges prodded the agency to accelerate an overhaul of procedures.

On February 5, Judge Henry H. Kennedy, Jr., of the US District Court for the District of Columbia concluded ruminations about the regulation of GE creeping bentgrass. The bentgrass had been engineered to tolerate glyphosate, the active ingredient in the herbicide Roundup^®. GE bentgrass could be used for lawns, athletic fields, and on golf courses. Over the years, efforts to develop Roundup Ready grasses have inspired concern that the gene conferring glyphosate tolerance might spread through reproduction with sexually compatible wild relatives and then persist in the environment.

The plaintiffs alleged that APHIS had acted arbitrarily and capriciously when it denied their petition to list GE bentgrass and glyphosate-tolerant Kentucky bluegrass as noxious weeds pursuant to the Plant Protection Act. APHIS had concluded that no biological basis existed for treating glyphosate-resistant strains of bentgrass and bluegrass differently from their nonresistant counterparts. APHIS then determined whether the plant species warranted quarantine pest status, because the plant is either "new or not known to be widely prevalent." Since neither Kentucky bluegrass nor creeping bentgrass fits this criterion, APHIS concluded that listing was not warranted.

The judge agreed with plaintiffs that the "new or not known to be widely prevalent" standard – borrowed from international agreements – is not a required consideration for the review of a noxious weed petition. APHIS’ insistence on this criterion, the judge decided, was arbitrary and capricious.

The judge vacated the denial of the noxious weed petition and sent the petition back to APHIS. "Congress’s intent in passing the Plant Protection Act (PPA)," the judge wrote, "was plainly to provide for regulation of all dangerous noxious weeds, whether new or old, or whether prevalent or not." The judge also cautioned that APHIS cannot supply its decision without providing a reasoned explanation, informed by sound science.

In a second allegation, plaintiffs claimed that APHIS had failed to comply with its own regulations when it granted field test permits for GE bentgrass. They argued that the agency had approved the GE bentgrass field trials without considering evidence that the plant is a weed in the areas of proposed release, and the agency had failed to make any type of localized weediness determination.

APHIS countered that it had complied with its regulations, which include a state-agency notification process for field trials. If local or state authorities consider the plant to be a weed, then APHIS does as well. If state authorities don’t consider the plant to be a weed, then APHIS doesn’t either.

Although the judge voiced concern that "APHIS has essentially ceded to state authorities the task of considering whether a given organism is a weed in the area of release," courts must give great deference to an agency’s interpretation of its rules. The judge granted summary judgment in favor of defendants.

The plaintiffs also claimed that APHIS had violated NEPA when it failed to determine whether the GE bentgrass field trials qualified as exempt from the agency’s obligation to conduct an Environmental Assessment (EA) or an Environmental Impact Statement (EIS). The judge said that the record contained no findings that the field trials fell under this exemption. Yet such findings were unnecessary. Any field test of GE organisms permitted by APHIS pursuant to the Plant Protection Act inherently falls under the "confined field release" NEPA exemption.

Falling within an exemption does not end the story, however. Even when APHIS has determined that an action falls under one of the NEPA exemptions, it still must determine whether an exception to the exemption applies. APHIS must prepare an EA or EIS if a confined field release of GE organisms has the potential to significantly affect the quality of the human environment. For example, the confined release may involve a new species or organism, or novel modifications of an organism that raise new issues. The judge could not find any evidence in the record that APHIS had considered these aspects of the proposed GE bentgrass field tests.

Judge Kennedy found substantial evidence that the field tests may have had the potential to significantly affect the quality of the human environment, and that the tests may have involved, at the least, novel modifications or new organisms that raised new environmental issues. APHIS’ apparent failure to consider these possibilities, Judge Kennedy decided, manifested arbitrary and capricious agency action and violates NEPA.

The judge granted summary judgment in plaintiffs’ favor, and he enjoined APHIS from processing any permit for a plant pest or potential plant pest without inquiring whether the NEPA exception applies to the permit and whether an environmental assessment should be prepared.

On February 13, Charles R. Breyer, a judge in the US District Court for the Northern District of California, decided another case that focused on APHIS’ procedures. This time, APHIS had taken the step of drafting an Environmental Assessment.

In May 2003, Monsanto Company submitted a petition that requested nonregulated status for GE Roundup Ready alfalfa. APHIS prepared an Environmental Assessment and accepted comments from the public about the EA and Monsanto’s petition for deregulation.

Many objected to deregulation of the GE plant. One of the main objections focused on the possibility that bee pollination could spread the glyphosate tolerance gene from GE alfalfa to conventional alfalfa, organically-grown alfalfa, or wild populations of alfalfa. Genetic contamination would affect US markets for organic and conventional products, as well as foreign markets. Seventy-five percent of exported US alfalfa heads to Japan, a country that bans the import of glyphosate tolerant alfalfa. Commentators also objected that deregulation of the GE alfalfa with the affiliated increase in Roundup use could boost the development of glyphosate-resistant weeds.

In June 2005, APHIS issued a Finding of No Significant Impact (FONSI) and approved Monsanto’s deregulation petition. APHIS concluded that it would be "up to the individual organic seed or hay grower to institute those procedures that will assure" that their crops will not include any GE alfalfa. By using reasonable quality control, the agency decided, "it is highly unlikely that the level of glyphosate tolerant alfalfa will exceed 1% in conventional alfalfa hay." This level of contamination would not bar the product from Japan.

While APHIS agreed that the deregulation of the GE alfalfa could lead to the development of additional glyphosate-resistant weeds, the agency did not see this impact as significant. After all, weed species have developed resistance to every widely used herbicide. Alternate herbicides and good stewardship could afford a defense against this potential problem, the agency assured.

Alfalfa growers, the Sierra Club, and other farmer and consumer associations alleged that the USDA’s deregulation of GE alfalfa violated NEPA. They contended that the introduction of the GE alfalfa would pass on the glyphosate tolerance gene to natural alfalfa, a significant environmental impact.

In the judge’s view, APHIS had effectively concluded that, whatever the likelihood of gene transmission, the impact would be insignificant, because organic and conventional farmers bore the responsibility to ensure that such contamination did not occur. Judge Breyer could find no evidence that APHIS had investigated whether farmers could actually protect their crops from genetic contamination.

APHIS could have approved the petition with a geographic limitation to isolate GE alfalfa, but it did not. "APHIS’s rejection of this option," the judge wrote, "without making any inquiry into the extent of likely gene transmission from genetically engineered seed crops to non-engineered seed crops is arbitrary and capricious."

The judge did not care for APHIS’ conclusion about the effect of GE alfalfa on exports. He could find no support in the EA or the FONSI for APHIS’ conclusion that gene transmission is highly unlikely to occur with the application of reasonable quality control. Judge Breyer also decided that the plaintiffs had raised substantial questions about the extent to which the GE alfalfa would contribute to the development of Roundup-resistant weeds, and about methods farmers use to control the resistant weeds.

The judge decided that APHIS had failed to take a "hard look" at the potential environmental impacts of its deregulation decision, as required by NEPA. He granted plaintiffs’ motion for summary judgment on its NEPA claim that APHIS must prepare an EIS.

Judge Breyer ordered the parties to submit a proposed remedy by the end of February. This did not happen. On March 2, the plaintiffs filed a request for permanent injunction against deregulation of the GE alfalfa before APHIS performed an environmental review in an EIS.

On March 12, a federal judge in San Francisco issued a court decision barring the sale and use of Roundup Ready (alfalfa tolerant to glyphosate herbicide) alfalfa after March 30. U.S. District Court Judge Charles Breyer granted the injunction at the request of a group of organic forage growers and environmental and consumer activists, who claimed Roundup Ready alfalfa could be harmful to the environment and the economy. Breyer’s ruling marked the first time a U.S. Department of Agriculture approval for a genetically modified seed product was overturned by a federal court. An April 27 hearing will determine whether the injunction becomes permanent.

Farmers who’ve already planted Roundup Ready alfalfa are not affected by the court decision, points out Larry Nees, state seed administrator for the Office of the Indiana State Chemist. "The injunction that’s been filed does not impact any continued used, harvest or sale of Roundup Ready forage," he said. "It’s important to note that the decision of this case was not focused on the safety of Roundup Ready alfalfa and Roundup Ready alfalfa seed. The district court that issued the injunction, and other regulatory agencies like the USDA, all agree that it poses no harm to humans and/or livestock. It’s just an issue of technicality as to how this was originally approved by the USDA and whether all the steps were taken to make sure that there was no impact on the organic growers and the conventional alfalfa growers in certain areas of the country."

Said Jerry Steiner, executive vice president for Monsanto, about the injuction: We are hopeful that a reasoned approach in this matter will address questions about the regulatory approval process for Roundup Ready alfalfa while maintaining farmer access to this beneficial technology. The extensive regulatory dossier for Roundup Ready alfalfa, combined with farmer stewardship agreements, provides a robust and responsible approach to managing the environmental questions raised by the plaintiffs in this case.

Monsanto, Forage Genetics International (a seed developer and subsidiary of Land O’Lakes Inc.), and several farmers were granted intervenor status in this case on March 8. Plaintiffs, defendants and intervenors can participate in oral arguments for this case on April 27.

The court has already accepted the fact that Roundup Ready alfalfa poses no harm to humans and livestock. As part of its regulatory filing for Roundup Ready alfalfa in April 2004, Monsanto provided USDA with an extensive dossier that addresses a variety of environmental, stewardship, and crop management considerations. Other regulatory agencies around the world, including Canada and Japan, have confirmed the environmental safety of Roundup Ready alfalfa.

In some parts of the country, the March 30 planting deadline does not leave enough time to plant Roundup Ready alfalfa that has been purchased. "We don’t plant alfalfa until the middle of May," said Dale Scheps, who operates a 145-cow dairy farm in Almena, Wisc. Scheps planted 35 acres of Roundup Ready alfalfa in 2006 and had already purchased enough seed to plant another 35 acres in 2007. "It’s a major setback to have this technology taken away from us," Scheps said. "It will needlessly drive up our feed costs because we will have to replace superior quality hay."

Hay & Forage Grower (hayandforage.com), the only national publication devoted exclusively to alfalfa and other forage crops, reported this spring that if Roundup Ready alfalfa is held off the market for an extended period, an already tight supply of conventional alfalfa seed could get tighter. Some growers who bought Roundup Ready alfalfa for spring planting will switch to conventional varieties, while others might turn to other crops instead.

Because Roundup Ready alfalfa was introduced just two years ago and costs more than twice as much than conventional alfalfa seed, it represents a minute share of the U.S. alfalfa market. Of the 22 million acres of alfalfa grown in the United States last year, it’s estimated that only about 200,000 acres were Roundup Ready – about 0.01 percent of the total. Thus, it is a small slice of business for technology provider Monsanto and most seed companies that sell it.

There are distinct advantages to planting Roundup Ready alfalfa. Plant scientists Kevin Bradley and Robert Kallenbach at the University of Missouri-Columbia point out that one of the clear advantages of this technology is in its broad spectrum weed control, including troublesome broadleaf weeds like curly dock, musk, bull, and Canada thistle, horsenettle, and dandelion. The application window is longer, with more effective control of many weeds than standard herbicides, and there is less risk of crop injury compared to other commercial herbicides.

Another advantage they see with the Roundup Ready technology is with spring establishment. Often, spring-established alfalfa is more difficult from a weed management standpoint. This is because many summer annual weeds emerge throughout April and May into newly seeded stands that have little to no canopy. To complicate this issue further, only a few conventional herbicide options are available for application on these newly seeded stands. However, they point out that the technology fee alone for Roundup Ready alfalfa costs about an additional $2.50 per pound of seed planted. Depending on alfalfa variety and seeding rate, this is an additional $125 per bag of alfalfa seed purchased.

Caffeine (1,3,7–trimethylxanthine) is an alkaloid compound that acts as a central nervous system stimulant in humans and is likely the world’s most popular psychoactive substance. Caffeine is generally found in the beans, leaves, and fruits of over 60 plants, where it acts as a natural pesticide, paralyzing and killing certain phytophagous insects and repelling slugs and snails. In the last fifteen years, considerable advances have been made in the genetic transformation of coffee plants. Researchers have been able to transform coffee plants with genes for insect resistance and herbicide tolerance, engineer decaffeinated coffee, and control coffee fruit maturation.

Herbivory accounts for approximately 37% loss in world agriculture production. Due to its natural antiherbivory function, caffeine production within food crops may provide one useful means for protecting important crops. Research indicates the reproductive potential (ovary length and egg number) of lepidopterans is significantly reduced in insects fed leaves treated with caffeine and the related compound theophylline, found in tea.

The Nara Institute of Science and Technology in Japan recently reported research on the development of caffeine-producing transgenic tobacco plants tolerant to tobacco cutworms (Spodoptera litura). Previously, the researchers isolated genes encoding three distinct N-methyltransferases and demonstrated in vitro production of the recombinant enzymes responsible for caffeine yield. They also published a review of the metabolic engineering of the caffeine biosynthetic pathway utilizing both gene silencing and over-expression approaches. The application of this research supported further efforts to employ transgenic caffeine-expressing plants as insect repellents.

Caffeine is synthesized from precursor xanthosine in three methylation steps and a ribose elimination step. The researchers first cloned three coffee cDNAs, designated CaXMT1, CaMXMT1, and CaDXMT1, encoding the three N-methyltransferases found in the caffeine biosynthetic pathway. Using xanthosine as substrate together with E. coli crude extracts, bacterially-expressed proteins were subjected to in vitro reaction. The subsequent production of caffeine demonstrated that these proteins are involved in caffeine synthesis in plants and also that bacterial extracts contain the 7-methyl xanthosine nucleosidase activity required for the methylation steps.

Next the researchers generated caffeine-producing transgenic plants in an attempt to confer stress tolerance. To do this, leaf discs from tobacco (Nicotiana tabacum cv Xanthi), which is naturally devoid of caffeine, were transformed via Agrobacterium by introducing a multigene transfer vector, pBIN-NMT777, containing the three coffee N-methyltransferase genes, CaXMT1, CaMXMT2, and CaDXMT1.

Fifteen transgenic plantlets were obtained and confirmed to express all three N-methyl transferase genes. HPLC analyses revealed that caffeine level varied, depending on the developmental stage of the plants. During vegetative growth and before flower bud formation, the caffeine content was low (0.2 μg/g fresh weight), and thereafter the caffeine content increased in small, young (3μg/g fresh weight) and large, aged (4.3μg/g fresh weight) leaves. Up to 6μg/g fresh weight of caffeine was noted in one line. No caffeine was detected in wild plants.

The caffeine-producing transgenic plants were then investigated to detect a potential chemical herbivory defense effect conferred by caffeine by observing the feeding behavior of the tobacco cutworm caterpillar. The tests indicated that insects positively avoided the transgenic material and that a caffeine level as low as 0.4μg/g fresh weight was sufficient to confer repelling effects. In further studies, caterpillars fed on caffeine-containing leaves in a no choice environment grew normally until reaching the pupating stage. The researchers concluded that though caffeine may not be a powerful pest control agent for agricultural applications, the production of caffeine in transgenic plants can be used as a promising approach in an overall pest control strategy, by directly acting as a repellent and indirectly reducing insect fecundity.

In conclusion, because damage to agricultural crops due to insects is considerably high, pest control assumes great significance. Several chemical pesticides, biocontrol agents, and genetic transformation technologies have been developed to control the incidence of pest infestation, among which utilization of caffeine appears to be a novel and safe method, as demonstrated in the present study in tobacco. Using the tobacco plant model, the team is currently attempting to develop transgenic insect tolerant Chinese cabbage (Brassica campestris) and chrysanthemum. Results from these experiments will further validate this strategy in pest control.

GENERATION OF MARKER-FREE TRANSGENIC PLANTS
Mihály Kondrák, Ingrid M. van der Meer, and Zsófia Bánfalvi

Transgenic technologies have enormous potential to improve crops of interest in a relatively precise way. However, the methods to introduce foreign DNA in a plant cell, either by Agrobacterium, microinjection, particle gun, or protoplast transformation, are relatively inefficient. For identifying those cells that have integrated the DNA into their genome, a selectable marker gene is co-introduced with the gene of interest. Approximately fifty different selection systems have been developed over the past several years. Despite the large number of systems, marker genes that confer resistance to the antibiotics kanamycin (nptII) and hygromycin (hpt) or the herbicide phosphinothricin (bar) have been used in most plant research and crop development techniques. Selection markers are not required in mature plants, especially when they are grown in fields. The European Union suggests avoiding the use of selectable markers in genetically engineered crops, and the ultimate goal is to introduce as few foreign sequences, in addition to the gene of interest, as possible. Moreover, the generation of marker-free transgenic plants responds not only to public concerns over the safety of genetically engineered (GE) crops, but supports multiple transformation cycles for transgene pyramiding.

De Vetten et al.¹ reported transformation of potato without the use of selectable markers. The best results were obtained with the potato variety Karnico, using the Agrobacterium tumefaciens strain AGL0, which exhibits extremely high transformation efficiency because it contains a DNA region originating from a super virulent A. tumefaciens strain. In this experiment, approximately 5000 regenerated shoots were isolated and analyzed by PCR. Transgenic lines were obtained with an average frequency of 4.5%. However, vector backbone sequences, which are as undesirable and unacceptable as selectable markers, were transferred along with the gene of interest in 60 out of the 99 transgenic lines. Moreover, only 10 vector-free lines contained a single T-DNA insertion, which is another important criterion for commercialization of GM crops.

The simplest marker elimination strategy is the co-transformation of genes of interest with selectable marker genes followed by the segregation of the separate genes through sexual crosses. Co-transformation has been accomplished in a number of ways, including co-inoculation of plant cells with two Agrobacterium strains, each containing a simple binary vector, dual binary vector systems, and modified two-border Agrobacterium transformation vectors.²

The isopentenyl transferase (ipt) gene that leads to cytokinine overproduction and results in transgenic shoots with abnormal shoot morphology can also be used as a selectable marker. In this case, appearance of normal-looking plants emerging from abnormal tissues indicates excision of the ipt gene, resulting in marker-free plants. Ipt selection was combined with a plant-derived T-DNA-like P-DNA fragment and used to generate marker- and backbone-free potato lines in a dual binary vector system with negative selection provided by codA against nptII marker gene integration. CodA is a conditionally lethal dominant gene encoding an enzyme that converts the non-toxic 5-fluorocytosine to cytotoxic 5-fluorouracil. Using this highly efficient way of selection hundreds of marker- and backbone-free Ranger Russet potato plants displaying reduced expression of a tuber-specific polyphenol oxidase gene were produced by Rommens et al.³

Efficient transformation systems using ‘shooter’ mutant Agrobacterium strains are also reported.⁴ These strains possess defective auxin-synthesis genes, but carry an intact ipt gene on the T-DNA of their own Ti plasmid that results in proliferation of transgenic cells and differentiation of adventitious shoots. Using a ‘shooter’ strain, regeneration on growth regulator-free media only occurs after successful infection of the plant tissues by agrobacteria. Furthermore, in a ‘shooter’ mutant / binary vector experiment, more than 60% of the transgenic lines proved to be ipt-free.⁴ Thus this system is a potentially useful alternative for marker-free gene transfer.

This strategy is based on the use of site-specific recombinases, under the control of inducible promoters, to excise the marker genes. Successful use of the Cre/lox, FLP/FRT, or R/Rs systems has been reported in different plant species in which Cre, FLP, and R are the recombinases, and lox, FRT, and Rs are the recombination sites, respectively.⁵

Recently, a binary vector designated PROGMO was constructed to assess the potential of the Zygosaccharomyces rouxii R/Rs recombination system for generating marker- and vector backbone-free transgenic plants with high transgene expression and low copy number insertion.⁶ The PROGMO vector utilizes a constitutively expressed plant-adapted R recombinase and a codA-nptII bi-functional, positive/negative selectable marker gene. It carries only the right border (RB) of T-DNA and consequently the whole plasmid will be inserted as one long T-DNA into the plant genome. The Rs recognition sites are located at certain positions such that recombinase enzyme activity will recombine and delete both the bi-functional marker genes as well as the backbone of the binary vector, leaving only the gene of interest flanked by a copy of Rs and RB (Fig. 1).

The efficiency of PROGMO transformation was tested by introduction of the β-glucuronidase (GUS) reporter gene into potato. It was shown that after 21 days of positive selection and using 300 mgl^-1 5-fluorocytosine for negative selection (Fig. 2), 29% of regenerated shoots carried only the GUS gene flanked by a copy of Rs and RB.

The PROGMO vector approach is simple and might be widely applicable for the production of marker- and backbone-free transgenic plants of many crop species.