Allison Wilson and Jonathan Latham
Many genetic engineers have long resented the regulatory procedures imposed on transgenic crop plants, often arguing that there is no difference between the risks arising from transgenic plants and plants bred using ‘conventional’ methods. A recent proposal calls for complete deregulation of transgenic plants which have only plant DNA inserted into their genomes (Schouten et al., 2006a,b). The term cisgenic has been coined for such plants in order to highlight the origins of the transferred DNA. Other terms for plant-derived transgenes include ‘all-native DNA’ and ‘P-DNA’ (Rommens, 2004).
In a response published in Nature Biotechnology, Schubert and Williams (2006) suggest that this proposed cisgenic/transgenic distinction arises more from political than scientific considerations. They point out that cisgenic plants will be created using the same highly mutagenic plant transformation techniques used to create other transgenic plants (Wilson et al., 2006). Furthermore, according to some definitions, cisgenes, like transgenes, may include anti-sense sequences, sequence changes to elude feedback inhibition and combinations of gene coding and regulatory components from different genes and species (Rommens, 2004). Indeed, the cisgenic plants engineered by Rommens et al. (2004), who claim to have made “the first genetically engineered plants that contain only native DNA”, were produced using Agrobacterium-mediated transformation of potato (via tissue culture and transient selection on kanamycin) and the cisgenes were composed of sense or anti-sense plant DNA from three distinct genes. However, unlike most commercially available transgenic plants, they did not have selectable marker genes or vector backbone sequences integrated into their genomes (Rommens et al., 2004). The absence of superfluous DNA, such as marker gene and vector sequences, is a real technological improvement, albeit one which could have and should have been made already to all transgenic plants.
While categorizing transgenes according to their origins may have merit, changes to risk assessment and regulations need to be based on scientific data not semantics. What is so far missing from the cisgenic debate in Nature Biotechnology is actual experimental data to support or refute the rather nebulous claims made for cisgenic plants (Schouten et al., 2006a,b,c). These include claims that, unlike transgenic plant breeding, ‘cisgenesis does not add an extra trait’ and that there is an ‘equivalence of products resulting from cisgenesis and traditional breeding including mutational breeding’ (Schouten et al., 2006b).
Can such claims be tested experimentally? Although not mentioned by Schouten et al. (2006a,b,c), a series of experiments using the model plant Arabidopsis thaliana have done just that (Bergelson et al., 1996; Purrington and Bergelson, 1997; Bergelson et al., 1998; Bergelson and Purrington, 2002). These experiments both (1) assessed whether introduction of a cisgene introduced unanticipated trait changes and (2) looked for differences between breeding methods by comparing plants where either transgenic breeding or a ‘conventional’ method (chemical mutagenesis) was used to introduce the identical trait into the identical genetic background. These detailed experiments are important because they illustrate the careful methodology needed to test claims made for cisgenic plants. Remarkably, the results showed that trait introduction via a cisgene can result in plants which differ in unanticipated and dramatic ways from their conventionally bred counterparts. Furthermore, the observed differences would likely have important agronomic and ecological implications for commercial varieties.
These experiments were designed to compare the fitness of herbicide-resistant A. thaliana plants created either by a ‘conventional’ plant breeding method or by transgenic methods (Bergelson et al., 1996; Purrington and Bergelson, 1997; Bergelson et al., 1998; Purrington and Bergelson, 2002). Fitness was assessed by determining either total seed number or levels of outcrossing. In the ‘conventionally’ bred plants, the new trait, herbicide resistance, was introduced by EMS mutagenesis. These plants differed from the wild-type parents by a single base pair mutation in the acetolactate synthase gene (Csr-1). Transgenic counterparts were made by cloning the identical Csr-1 allele and introducing it into the herbicide-sensitive wild-type parents. The transferred Csr-1 allele was a cisgene, as it was a genomic clone, with its native promoter and terminator sequences. However, the resulting herbicide-resistant plants were not strictly cisgenic plants, as they were also homozygous for T-DNA border sequences and a selectable marker gene. Therefore, as a control, plants transformed with only the selectable marker gene were also tested. As a further control, multiple independently derived transgenic lines were analysed and all herbicide-resistant lines were multiply backcrossed to the wild-type parent.
The results of these experiments were striking. The levels of outcrossing were higher in all transgenic lines carrying the Csr-1 cisgene as compared to the conventionally bred Csr-1 plants (Bergelson et al., 1998; Bergelson and Purrington, 2002). Outcrossing frequency was 0.34% for conventional Csr-1 plants and 1.3%-12.4% for the four independently derived Csr-1 transgenic lines. As further evidence of differences, when grown under field conditions, both the transgenic and conventional herbicide-resistant plants showed decreased total seed numbers as compared to the herbicide-sensitive wild-type parents. However, when nutrients were added to field-grown plants, only the transgenic plants still showed a fitness decrease (Bergelson et al., 1996; Purrington and Bergelson, 1997). Fitness differences were not present in control plants transformed with marker DNA under any of the conditions tested (Purrington and Bergelson, 1997).
These results therefore do not support the claims made by Schouten et al. (2006b). They show instead that a cisgene can introduce important unanticipated phenotypic changes that are not present in wild-type parents, conventionally bred counterparts or in transgenic controls which lack the cisgene.
This series of experiments stands in stark contrast to most other phenotypic analyses of transgenic plants, where a transgenic plant is only cursorily compared to the non-transgenic parent or to other ‘similar’ varieties. The quality of this particular analysis rests on several experimental factors including:
1. Transgenic plants were compared not only to the parent plant but also to a conventionally bred counterpart.
2. Specific null hypotheses were tested (e.g. fitness does not depend on the method of plant breeding).
3. Phenotypic traits were measured quantitatively and experiments were repeated.
4. Effects of marker genes were controlled for.
5. Multiple independent transgenic lines were analysed.
6. Unlinked mutations introduced via plant breeding methods were removed or minimised by backcrossing.
7. Genetic background effects were controlled for by using the same parent to produce both the conventional and the transgenic lines, and as an experimental control.
Unexplained morphological and biochemical differences are often found between transgenic plants and their non-transgenic parents (e.g. Haslberger, 2003). The experiments carried out using the Csr-1 cisgene suggest that the use of cisgenes can also result in unanticipated consequences between cisgenic plants and their conventionally bred counterparts, and regulators will need to take these into account.
The careful methodology and numerous controls used by Bergelson and Purrington are both an illustration of those needed to precisely identify unanticipated traits arising from transgenic plant breeding methods and a prerequisite for analysis of their molecular origins. However, further experiments, on these and other observations of unanticipated consequences, are still needed to pinpoint the exact molecular mechanisms responsible. Once these mechanisms are understood, regulators can take them into account and plant breeders can alter transformation methods to avoid introducing unanticipated traits. Much basic research still remains to be done if risk assessment and regulatory policy for commercial transgenic plants are to be based on science rather than on politics or untested assumptions. The experiments described above are a useful start, but they are nevertheless just a small fraction of what is needed to determine, on a scientific level, what important differences, if any, exist between transgenic and cisgenic plants and how these differ from their conventional counterparts.
Bergelson J and Purrington CB (2002) Factors affecting the spread of resistant Arabidopsis thaliana populations. In: Genetically Engineered Organisms: assessing environmental and human health affects. Letourneau, DK and Burrows B., eds. CRC Press, Boca Raton. Chapter 2, pp. 17-31.
Bergelson J, Purrington CB, Palm CJ, Lopez-Gutierrez J-C (1996) Costs of resistance: a test using transgenic Arabidopsis thaliana. Proceedings of the Royal Society of London B 263: 1659-1663.
Bergelson J, Purrington CB, Wichmann G (1998). Promiscuity in transgenic plants. Nature 395: 25.
Haslberger AG (2003) Codex Guidelines for GM foods include the analysis of unintended effects. Nature Biotechnology 21(7): 739-741.
Purrington CB and Bergelson J (1997). Fitness consequences of genetically engineered herbicide and antibiotic resistance in Arabidopsis thaliana. Genetics 145: 807-814.
Rommens C (2004) All-native DNA transformation: a new approach to plant genetic engineering. Trends in Plant Science 9(9): 457-464.
Rommens CM, Humara JM, Ye J, Yan H, Richael C, Zhang L, Perry R, Swords K (2004) Crop improvement through modification of the plant’s own genome. Plant Physiology 135: 421-431.
Schouten HJ, Krens FA, Jacobsen E (2006a) Cisgenic plants are similar to traditionally bred plants. EMBO Reports 7(8): 750-753.
Schouten HJ, Krens FA, Jacobsen E (2006b) Do cisgenic plants warrent less stringent oversight? Nature Biotechnology 24(7): 753.
Schouten HJ, Krens FA, Jacobsen E (2006c) ‘Cisgenic’ as a product designation. Nature Biotechnology 24(11): 1331-1333.
Schubert D and Williams D (2006) ‘Cisgenic’ as a product designation. Letter to Nature Biotechnology 24(11): 1327-1329.
Wilson A, Latham J, Steinbrecher R (2006). Transformation-induced mutations in transgenic plants: analysis and biosafety implications. Biotechnology and Genetic Engineering Reviews 23: 209-237.