This snapshot, taken on 21/12/2003, shows web content selected for preservation by The National Archives. External links, forms and search boxes may not work in archived websites.

Department for Environment, Food & Rural Affairs

Guidance on Best Practice in the Design of Genetically Modified Crops

Comments on this guidance note should be sent before 21 December 2000 to The Secretary, ACRE Best Practice Sub-group, Floor 3/H11, Ashdown House, London, SW1E 6DE.

Contents

Summary

Technologies are available to produce GM plants that contain only minimal genetic modification. Insertion of the smallest quantities of DNA required to obtain the desired trait will help to simplify risk assessment and further reduce uncertainty. More challenging, although the technology is under development, will be the integrated use of risk management traits to biologically contain transgenes and their products and so avoid or minimise environmental exposure. The reliability of these systems is testable. These approaches, which include the regulation of processes such as seed fertility and flowering to minimise pollen dispersal will need to satisfy the requirements imposed by breeders, seed producers and agricultural practice. Where novel technologies have been developed, intellectual property rights may restrict access and have a large impact on how widely they are employed.

Section 1: Introduction and aims of this guidance

1.1 The release of genetically modified organisms[1] (GMOs) in Europe is controlled by Directive 90/220/EEC[2]. The Directive provides a European Community-wide regime so that no GMOs may be released or marketed in the Europe without consent from the regulatory authorities. Applicants for release consent must supply a dossier of prescribed information[3] about the GMO, which includes a detailed risk assessment of the possible impact on human health and the environment. Before consent is issued the regulators must be satisfied that all appropriate measures have been taken to avoid adverse effects on human health and the environment.

1.2 In the UK, the Advisory Committee on Releases to the Environment (ACRE) critically reviews all applications to release and market GMOs and advises Ministers[4] on the risks. Consent will be issued only if ACRE considers that a proposed release will be safe. More details about the regulatory framework and the role of ACRE can be found at Annex I.

1.3 In December 1999, ACRE decided that it should offer guidance on best practice in the design of GM plants that are intended for release to the environment. A technical subgroup was established to consider this in detail; the ACRE subgroup on best practice in GM crop design. The minutes of its meetings can be found at: www.defra.gov.uk/environment/acre/bestprac/index.htm.

1.4 The aim of the subgroup was to consider how the design and construction of GM plants could be used to further improve their safety and/or to simplify the risk assessment. For example by preventing or minimising cross-pollination, avoiding antibiotic resistance marker genes or switching on inserted genes only when and where they are needed in the plant. The output of the subgroup's deliberations is this guidance document.

1.5 This guidance is particularly aimed at practitioners developing GM plants for commercial use. It is intended to be proactive and looks at practices that could be applied both now and in the future. Therefore it will be subject to periodic review. The advice is based on experience gained from past applications to market GM crops in Europe, knowledge of emerging technologies and direct consultation. It establishes some general principles of best practice and reviews technologies that might enable these principles to be applied in the construction of the next generations of GM crops.

1.6 The guidance is not intended to be prescriptive or to promote specific technologies. Nor are approaches listed here meant to be definitive or exhaustive. Application of these technologies is neither a requirement nor a prescription for the successful passage of products through the regulatory system. ACRE will continue to consider applications for consent to release and market GM crops case-by-case and on the basis of safety.

1.7 This Guidance document is set out in three main sections. Section 2 explains the basic principles of best practice in the design of GM plants and sets these within the context of risk assessment. Section 3 explores the issues involved with putting best practice into practical use. Whereas Section 4 sets out the technical details and associated enabling technologies.

Section 2: The philosophy of base practice in the design of GM plants

Uncertainty and the Precautionary Principle

2.1 Safety assessments of GMOs are made on the basis of the best scientific evidence available at the time. When knowledge is limited in a specific area of risk assessment, uncertainty may lead to disagreement on the risks and acceptability of particular modifications. This necessarily impedes the progress of new products through a science-based regulatory system. Nowhere is this better demonstrated in agricultural biotechnology than over the issue of the use of antibiotic resistance marker genes in GM plants.

2.2 The precautionary principle can be invoked when there is significant uncertainty. For example, the UK has voted against commercial release of GM plants that contain specific antibiotic resistance marker genes. This is despite the fact that there is no scientific evidence that demonstrates transfer of functional genes from plant material to bacteria in the environment. But, it is the view of some that there could be serious consequences for the effective use of antibiotic therapeutics if such transfer were to occur and be selected for. In the absence of clear scientific advice on this issue, regulatory authorities have taken a precautionary approach. This has led to the commissioning of research in an effort to resolve these concerns.

2.3 Long-term and indirect effects of new technologies always present a challenge to safety assessment. Post-release monitoring provides a mechanism for the early detection of any adverse effects. The challenge for scientific committees such as ACRE, applicants and regulators is to identify the key areas of uncertainty and design methods to monitor these effectively. The requirement for applicants to consider post-market monitoring is a key amendment to Directive 90/220/EEC. In this respect, detection methods and unique identifiers may be an important consideration in risk assessments[5].

Environmental Risk Assessment

2.4 Risk assessment underpins the decision making process for granting consents to release and market GMOs, since the sole basis for making decisions to grant consent is safety. Environmental safety assessments typically include consideration of the following potential hazards:

2.5 Harm may result if hazards are realised. Risk assessment evaluates the likelihood of realisation and what the consequences will be; risk is therefore a product of these two quantities.

2.6 The risks of GM crops approved for the market-place have been assessed to be low or negligible because the plants do not possess inherent characteristics that could cause damage to human health or the environment. GM plants in this category are typically those that are not persistent, are incapable of gene transfer to other plants or are unlikely to cause damage to the environment because of the limited nature of the novel traits that they express.

2.7 Assessing the risks of releasing GM plants in the near future may be more complex. Plants with several transgenic traits (stacked traits) or traits that could confer advantages to wild plants, or modifications that produce pharmaceuticals, are being developed.

2.8 What are the options for risk assessment and management in the future? Carefully collected and analysed ecological data will certainly be required in some cases and post-release monitoring will become increasingly important. But are there any other approaches to reducing risk?

2.9 A unifying principle that should be considered throughout the risk assessment is the connection between hazard (source), pathway, receptor, and effect. The intrinsic properties of any activity, substance or organism that may cause harm are termed hazards. It is vitally important that connectivity between these four (the exposure) can be shown. If any of these elements is missing, then the risk can be treated as zero and the risk assessment need go no further. Consider the theoretical example of a GM plant that produces a gene product which, when added to sugar solutions and fed to bees in laboratory tests, is toxic. This GM plant may be a hazard to bees in the environment but they will only be at risk if the hazard is realised. For example, if the plant does not flower, or the gene product is not present in pollen or nectar, bees will not be exposed. Similarly, if the plant flowers but susceptible bees do not visit the flowers because the flower is unattractive to them they will not eat the pollen or nectar. Formulation of the problem is key to risk assessment. An example of this is summarised in Figure 1.

Figure 1 Problem Formulation

Hazard
|		Bee Toxin
|
Pathway
|		Bee Visits Flower
|
Receptor
|		Bee Eats Pollen
|
EffectBee Mortality

2.10 Excision of antibiotic resistance marker genes after they have served their function in the early stages of the GM plant modification process would avoid the contentious issues over their presence in crops in the field, feed or food. This would be an example of where a potential hazard is removed.

2.11 A principle objective of best practice in GM plant design is to reduce exposure and therefore reduce risk from transgenes and their products. There are three ways of accomplishing this:

i. Avoid or minimise inclusion of superfluous transgenes or sequences
ii. Avoid or minimise superfluous expression of the transgene
iii. Avoid or minimise the dispersal of transgenes in the environment

2.12 This approach simplifies risk assessment and post-release monitoring. For example, where environmental exposure is avoided by removal of unnecessary transgenes, risk assessment involving detailed knowledge of natural ecosystems, is unnecessary. Where a risk management trait is deliberately used to avoid environmental exposure, the risk assessment will focus on the efficacy of the risk management trait.

What are the prospects for the concept of best practice?

2.13 As we improve our understanding of plant molecular genetics, the technology for genetic modification of plants should become increasingly precise and more predictable in its outcome. This is in contrast to the complexity of predicting environmental impacts and their significance.

2.14 The next two sections consider best practice and emerging GM strategies that could contribute to the design of genetically modified plants to minimise environmental exposure to transgenes and their products. Consumers desire to separate GM and non-GM supply chains may be a key driving force behind the adoption of bio-containment techniques.

Section 3: Best practice philosophy - in practice

Minimise Extraneous DNA

3.1 It is acknowledged that plant species and cultivars can respond to transformation differently, so what is offered here is the most desirable outcome rather than a prescription of the mechanism by which this might be achieved. Thus, the ultimate aim is to obtain a plant with a new trait but with the minimum amount of inserted DNA, all extraneous sequences including marker genes having been eliminated or avoided.

3.2 There are a number of reasons to aim to produce transgenic plants with as little extraneous DNA as possible:

3.3 However, there are times when the presence of genetic material other than the transgene of interest may be an advantage, such as a gene conferring a seed sterility trait to provide biological containment of the primary transgene, or a sequence to assist traceability. In which case any such gene should be linked closely to the primary transgene to avoid genetic segregation.

3.4 Thought needs to be applied to the final transformed plant product at the start of the process of vector construction. If a marker gene is present in the T-DNA, then mechanisms to allow subsequent excision could be included. Alternatively, co-transformation of the transgene of interest and a selectable marker could be used such that subsequent meiotic segregation could remove the marker gene. If transformation of plant cells by direct gene transformation technology such as biolisitics or electroporation is possible, then it may be a simple step to remove the unnecessary genes from the DNA beforehand. For herbicide tolerance traits, the gene of interest can be selected directly, but, for other transgenes, a selection system is needed unless the transformation frequency is high enough to identify transformants easily without selection.

3.5 It is also advisable at an early stage of screening transformants, to discard those with unwanted vector sequences, such as those from outside the T-DNA especially plasmid replication origins and antibiotic resistance genes. Polymerase Chain Reaction (PCR) studies carried out with appropriate positive and negative controls, allow much larger numbers of transformants to be screened than was the case in the past. The structure of the inserted DNA will influence the ease with which it can be characterised. This may in itself be viewed as an important "trait" so that when comparing, say, 100 plant transformants, those with the desired trait and a minimum of extraneous DNA would be the ones picked for further testing. In future, technical advances may allow direct selection of transformants without the need for selectable marker genes.

3.6 Transgenes can be integrated into chloroplast DNA by homologous recombination. In this way the precise location of the gene can be controlled. Because of the specificity of the integration event fewer duplications or illegitimate insertions occur. The high copy number of chloroplasts allows high level expression to be achieved and the implications of this in the risk assessment will need to be considered.

Minimise Transgene Dispersal in the Environment

3.7 Plant species have different breeding systems so it is only possible to define the most desirable outcome rather than a prescription of the mechanism by which it might be achieved. Thus, the ultimate aim is to obtain (where appropriate) a plant with a new trait that is biologically contained.

3.8 It is important to consider whether or not it is necessary to release plants into the environment. For some crops that produce high value pharmaceutical products, physical containment (e.g. glasshouses) might be just as effective.

3.9 Where environmental release is necessary, thought should be given to which plant species is the most appropriate recipient of the transgene. For example, the expression of pharmaceutical genes in crop plants or plants with wild relatives may not be desirable, choosing a recipient plant species so as to exploit natural sexual incompatibility is a sensible starting point.

3.10 Exploiting flowering-time difference between plants might also offer a simple method of genetic isolation. The most obvious approach would be to use existing varieties with different flowering times. Another possibility is to use varieties, or breed purpose-made varieties, that are unattractive to insects by altering their flower colour, shape, scent or the production of nectar or pollen. Key genes controlling floral development have recently been isolated and their manipulation in crops may enable these processes to be controlled. Such transgenic plants would also have the advantage of reducing exposure to non-target visiting insects.

3.11 For most plants, pollen and seed are the main agents of gene dispersal. Transgenic plants that cannot produce pollen already exist and have been developed to facilitate hybrid seed production. The production of transgenic plants that produce sterile seed is also feasible and this technology has been developed as a gene protection system to secure intellectual property rights. Both systems could also be used for risk management purposes. The benefit of linking a trait gene to a sterility gene to arrest pollen or seed development is that the frequency of both genes declines in subsequent populations as strong selection against them occurs. This happens because plants that inherit these genes do not produce pollen or seed.

3.12 There is a range of potential methods to prevent the escape of transgenes from crops to the wider environment, but solutions are needed to meet problems caused by certain practical constraints, such as the requirements of plant breeders and seed producers. The life histories of specific crop species must also be considered with respect to normal agricultural management practice. For example, the complete elimination of flowering is acceptable in vegetable crops and forage grasses during the 'cropping stage', but unless this trait is switchable in some way, the production of seed by the seed producer, or the breeding of new varieties by the plant breeder, is difficult or impossible. In crops where the harvest is a fruit or a seed, given that most crop species are self-pollinating, the production of pollen, by at least the majority of flowers, is essential (but see Section 4.6a). Most major crops fall into this category.

3.13 Plastid transformation is a mechanism to reduce gene dispersal, but there may be constraints to the use of this technique (see Section 4.10).

Minimise Unnecessary Transgene Expression

3.14 The ultimate aim is to obtain a plant with a new trait but with no unnecessary transgene expression. Optimum best practice would result in constructs in which the introduced genes are expressed in specific plant tissues and only at the time they are needed. This would make the environmental risk assessment easier and minimise unnecessary exposure to non-target organisms.

3.15 Control of gene expression is technically possible. The importance of this approach is that gene products can be excluded from specific tissues such as pollen or fruit. In addition, regulated expression might provide genetic switches to manipulate risk management traits as required.

3.16 The use of promoters that are induced by chemicals, for example, offers the potential to regulate or control the fertility of a crop. Such systems could be manipulated so that crops that do not produce pollen (male sterility) are the norm and fertility is restored by treatment with a specific chemical. Thus, breeders and seed producers can carry out their work with the plant variety, but equally, farmers can use the same variety in the sterile phase, minimising any potential risks to the environment.

Section 4: Enabling technologies

4.1 This section builds on Section 3 and briefly sets out enabling technologies and their associated technical details to illustrate how, in practical terms, best practice guidelines might be adopted in designing GM crops.

Alternative Markers to Antibiotic Resistance Genes

4.2 Plant transformation is an inefficient process and only a small proportion of the targeted cells take up and integrate the recombinant DNA molecules. Selectable marker genes allow selection of transformed cells. Many selectable markers in common use encode resistance to antibiotics although a number of alternative selection systems are available. Possible alternatives include reporter genes; genes that confer resistance to cytotoxic agents and genes that confer an ability to utilise compounds that are normally inaccessible.

4.3 It is also important to consider whether the alternatives to antibiotic resistance markers could bring their own risk assessment issues. For example, could the use of these genes cause adverse change in plant metabolism that could harm people, livestock or the environment. Another potential problem with alternative approaches (but not related directly to risk) is that of intellectual property rights. Intellectual property is connected with most marker gene systems and could restrict the widespread commercial adoption of the best systems.

4.4 The following sub-sections set out briefly some possible alternative approaches to using antibiotic resistance marker genes. In each case an indication is given whether this technology is available now or whether it is still largely just a concept but which could reasonably be expected to be available in the future.

4.4a Reporter genes (available now)
Reporter gene technology might be used more frequently in future because many of these marker genes are not subject to intellectual property rights – they are widely applicable now. The safety of these gene products is often well characterised. For instance, although there is minimal risk associated with the use of the gusA (uidA) gene that encodes beta glucuronidase (GUS), risk assessment would still be required for a plant carrying this gene. Various detection systems are available for selection of the reporter genes, these include fluorescent cell sorting or direct selection of expression. These might be feasible for some applications.

4.4b Resistance to cytotoxic agents (available now and requiring future development)
These selectable marker genes function by encoding a protein that can modify a chemical from a toxic to non-toxic form, thereby allowing the cell with the marker to grow in the presence of the chemical. There are several limitations associated with these genes, which do not always work in all plants. There may be unknown or unpredictable effects in different plants and so the applicability of each would have to be considered on a case-by-case basis.

4.4c Herbicide tolerance (HT) genes (available now)
The presence of HT markers may be undesirable in certain transgenic crops. The management of these crops, to avoid gene stacking of different herbicide tolerances, is important. The presence of a HT trait used only for the selection of the GM plant and not intended to be used as an agronomic trait, might tempt growers to use the HT trait inappropriately.

4.4d Auxotrophic markers (available now and requiring future development)
Auxotrophic markers either complement mutants that are deficient in a corresponding essential growth factor, such as the ability to synthesise an amino acid e.g. histidine, or enable normal cells to utilise an otherwise inaccessible nutrient e.g. mannose.

Bacterial plasmids have been constructed that carry a copy of the HIS3 gene that facilitates marker gene selection by complementation following transfer to an Escherichia coli strain that does not normally produce histidine. The HIS3 gene is very unlikely to pose a risk to human health or the environment and might have a useful role in reducing the current reliance of industry on antibiotic resistance marker genes. However, auxotrophic markers are not, at present, effective substitutes for antibiotic marker genes for maintaining plasmids for plant transformation mediated by Agrobacterium. Auxotrophic complementation is not a reliable genetic procedure in this bacterium because such mutations may be difficult to identify and complementation is ambiguous and therefore hard to detect.

Phosphomannose isomerase (PMI) is a novel plant selectable marker. The enzyme is not present in most plant cells. It catalyses the reversible interconversion of mannose 6-phosphate and fructose 6-phosphate and expression of this marker gene enables plant cells to utilise mannose. The risk assessment will have to include a consideration of the potential ecological impacts of plants that have this capability and any potentially toxic or allergenic changes in plant metabolism that may arise as a result of this genetic modification.

Removal of Extraneous DNA

4.5 Gene excision systems offer a way of removing unnecessary genetic elements once they have been integrated into transgenic plants. In some situations this may be essential. Currently, unwanted marker genes can be removed in three different ways:

4.5a Co-transformation (available now and requiring future development)
Co-transformation can be carried out in the presence of two binary Ti vectors during Agrobacterium tumefaciens - mediated transformation such that one vector contains the gene of interest and the other carries a marker gene. A proportion of the resulting transformants will contain both genes inserted at distinct loci, which can be later separated during meiosis. Although extra effort is needed to generate sufficient transformants for screening and selection of the ones containing both genes on separable loci, this system is effective. The current limitation is that only about 25% of transformants have separate T-DNA loci which can be segregated during meiosis. So there is scope for improvement.

However, higher co-transformation frequencies can be attained. This is achieved by using a binary vector, harbouring a marker or reporter gene, in Agrobacterium rhizogenes and scoring for the development of hairy roots from inoculated explants on phytohormone-free medium (a Ri trait) and expression of the reporter gene (a binary Ti trait). In this way co-transformation frequencies of 60% are possible. The virtues of this approach are that it is relatively simple technically and is not subject to intellectual property rights other than those already existing. The limitations are that it cannot be used with crops that have no meiosis or where the generation time is very long – e.g. fruit trees.

4.5b Transposon-mediated repositioning of marker genes (available now and requiring future development)
Transposable elements are sequences that have the capacity to move from one genomic location to another. Here, the vector is designed in such a way that the marker gene is transposed to a genetically unlinked site (by the activity of an introduced transposase enzyme) allowing selectable marker-free plants to be identified in the segregating progeny. However, the frequency of excision is fairly low and the screening required can, in practice, be a limiting factor. There may be rearrangements at the site of transposition. Therefore rigorous molecular data will be required to define the site of insertion, confirm the absence of unwanted sequence and that rearrangement have not occurred.

4.5c Site-specific recombination mediated excision of marker genes (available now and requiring future development)
Highly specific systems have been described that allow the removal of the marker gene by site specific recombination after transformation. All are simple two-component systems involving a recombinase enzyme that acts in trans to catalyse recombination between two short, specific DNA sequences. With the CRE/lox system, selectable markers are flanked by a pair of 34bp lox sites specifically recognised by CRE recombinase enzyme that catalyses recombination between two 34bp lox sequences. This results in marker deletion. The transformation vector can be designed so that the marker gene and the cre gene are next to each other and flanked by a pair of lox sites so that both can be excised together. To control the system, the recombinase must be under the control of a tightly regulated, inducible promoter. An alternative approach would be to introduce the recombinase into a separate line, cross it with the line containing the marker gene and primary transgene and then select progeny that only contain the primary transgene.

Risk assessment should include consideration of any allergenic or toxicological properties of the recombinase protein (if still present), molecular analysis to confirm the removal of the marker gene and potential unintended recombinase-mediated rearrangements (also see section (4.12a).

Control of Flowering and Fertility in Crop Plants to Minimise Transgene Dispersal in the Environment

4.6 Broadly, solutions fall into the following five categories:

4.6a Apomixis (requiring future development)
Apomixis, the production of seeds without fertilisation, occurs naturally in many plant species, some of which are close relatives of crop plants. When seed are produced apomictically the embryo, or embryo and endosperm, develop without a requirement for fertilisation. Transfer of the primary transgene to neighbouring crops via pollen would be minimal, as plants can be male sterile (do not produce pollen) without compromising seed or fruit production. The processes underlying apomixis can already be genetically manipulated.

4.6b Cleistogamy (requiring future development)
In some plant species, flowers are produced that develop normally, but fail to open. Consequently, self-pollination occurs, but pollen is unlikely to escape from the flower. This solution would require modifications to flower design. Several floral genes have recently been identified which could be manipulated to this end.

4.6c Hybidisation barriers (requiring future development)
Inter-specific hybridisation only occurs between closely related plant species. Two main barriers prevent hybridisation between more widely diverged species - inter-specific incompatibility at the stigma surface or within the style, which prevents fertilisation, and post-fertilisation barriers that cause seed abortion, usually through failures in endosperm development. Strengthening either barrier would potentially prevent hybridisation. Currently, the most tractable solution is probably post-fertilisation modification since the molecular basis of regulation is understood.

4.6d Inhibition of flowering to block floral development (requiring future development)
In recent years, the molecular basis of the processes that control flowering has been determined in Arabidopsis and other species such as Antirrhinum spp. These studies open up the possibility of manipulating flowering-time control genes and blocking or promoting flowering in a range of species.

4.6e Genetically engineering male sterility so that a plant produces infertile anthers (available now and requiring future development)
It has already been demonstrated that pollen development can be prevented by destroying the tapetum of a developing anther in both dicotyledonous (and to a lesser extent monocotyledonous) crops using non-specific nucleases driven by cell-specific promoters. Nuclease inhibitors can be crossed in to restore pollen fertility. Recently, several promoters have been developed that are induced by the application of exogenous chemicals. There is now the possibility to use such promoters to control flowering or fertility 'restorer genes', when required. However, this is a one-way barrier – the crop can still be fertilised by pollen to produce a hybrid and this should be taken into account in the risk assessment. The ability to manipulate fertility in this way may enable the constraints of breeders, seed producers and growers to be reconciled with environmental considerations.

Seed Sterility (emerging technology)

4.7 Genetic systems have been devised that enable crops to be genetically modified so that they produce seed that is incapable of germination (sterile seed). One consequence of this is that the seed cannot be saved and replanted the following season. It is beyond the scope of this guidance to consider the socio-economic and ethical debate raised by this capability.

4.8 This technology offers a promising technique for genetic isolation. However it has not been adopted for a number of reasons. To be reliable the GM trait of interest needs to be physically linked to a sterility gene to prevent gene escape, or a recombination event will disrupt the linkage. The likelihood of this occurring is probably low, predictable and verifiable, and its importance in risk assessment will depend upon the nature of the trait gene.

4.9 The benefit of linking a transgene to a sterility gene is that the frequencies of both decline in the population simply because of selection against the sterility gene due to the fact that plants that inherit these genes do not produce viable seed. The ecological significance of this for wild sexually compatible relatives of a GM crop would depend upon the background frequency of legitimate pollination and seed set. Where there is low gene flow, seed sterility will have an insignificant effect on population growth. However, there may be situations in which a recipient plant is very rare, or where gene flow was high, such that reduced fecundity could significantly affect population viability.

Plastid Transformation Technology (available now and requiring further development)

4.10 Transformation of plastids is an emerging technology, and at present the methodologies have not been applied successfully to all plant types. The technology offers two potential advantages over transformation of the nuclear genome. First, integration of foreign DNA into chloroplast DNA can be more precise, producing single inserts with few rearrangements. As the complete nucleotide sequences of sixteen chloroplast genomes have already been determined, it is possible to target precisely the site of transgene insertion. Secondly, chloroplast transformation technology may limit dispersal through pollen in crops. Amongst higher plants, plastids are predominantly maternally inherited although there may be some paternal transfer. The contribution of pollen to plastid transfer would have to be examined for each species as part of a risk assessment.

4.11 Once homoplasty has been achieved, gene expression appears to be stable. Stability of transgene expression is part of the risk assessment for any plant going through the regulatory system. The aadA gene (conferring streptinomycin/spectinomycin resistance) has been shown to be an effective marker gene in this system. Alternatives are in the pipeline and are likely to be available in the near future. Furthermore, it is technically feasible to remove marker genes once they have fulfilled their function.

Strategies to Minimise Transgene Expression

4.12 There are a number of approaches to minimise the expression of transgenes when it is not required.

4.12a Gene excision systems (requires future development)
An excision system might be used to remove unnecessary genes in a GM plant under field conditions in a tissue specific manner. The excision system would be controlled by a tightly regulated promoter that might, for example, be activated during fruit or flower development leading to the exclusion of the gene in those tissue types.

This could be used at an early stage to remove marker genes, or just before flower/fruit development of a plant containing a transgene, if the resulting fruit is to be consumed. For excision to take place at a later stage the process would have to be near 100% efficient, or specific acceptability levels of non-excision would have to be set. Another problem may result if the excised gene were to reintegrate at another site. The technology has still to be optimised, but it shows promise for future genetic containment. If ACRE received an application for the release of plants based on this system, the application would have to demonstrate in the risk assessment – and backed up by appropriate data - that the genes had been eliminated from the genome.

Depending on the construct, the recombinase gene may remain in the plant line and result in recombination at other sites in the genome, a possibility that carries uncertainty about its subsequent effects. Therefore, it may be desirable to remove of the introduced recombinase. The risk assessment will have to consider the possibility of less than 100% excision efficiency and possible rearrangements and their effects.

4.12b Introns and chloroplast editing sequences (available now)
Introns typical of plant genes, and editing sequences peculiar to chloroplast genes, are not processed in bacteria. In the unlikely event of environmental gene transfer from GM-plants to bacteria the inclusion of introns in the construct would prevent their expression in the recipient, if genetic containment systems were required due to the nature of the GM trait.

4.12c Chemically inducible promoters (available now and require future development)
Chemically inducible gene expression has interesting possibilities within the context of best practice minimisation of gene expression. The reliability of these systems will need to be demonstrated so that for example, fertility is not restored by freak environmental conditions in the field.

Annex I: The legal framework for decision making on the release and marketing of GMOs in the EU

Introduction

The purpose of this annex is to outline the legal framework against which decisions are made about the release and marketing of genetically modified organisms (GMOs) in the UK and EU. Releases and marketing of GMOs can only take place in the EU with the explicit consent of the regulatory authorities. Under the legislation, decisions on the granting of consents are based solely on the safety of the proposed GMO release. The decision making process is underpinned by risk assessment.

European Framework

The release and marketing of GMOs is controlled in the EU under Council Directive 90/220/EEC[6]. The Directive defines a GMO as "an organism in which the genetic material has been altered in a way that does not occur naturally by mating and / or natural recombination". The Directive requires Member States to ensure that all appropriate measures are taken to avoid adverse effects on human health and the environment that might arise from the release or marketing of GMOs.

The Directive sets out two regulatory regimes; Part B for controlling releases for research and development, and Part C for placing GMOs on the market in the EU. Essentially, the Directive requires any person, before undertaking a release of a GMO or placing a GMO on the EU market, to submit a notification to the competent authority of the Member State within whose territory the release is to take place, or where the GMO is to be placed on the market for the first time. The key difference between Part B and Part C is that for research and development releases, decisions are made by individual Member States, whereas for placing GMO products on the market, decisions are made by all Member States. Part C decision making often necessitates a voting procedure to resolve differences in opinion on risk, and hence delays in decision making if the Commission, lead Member State and the notifier fail to find an acceptable way forward.

Research and Development Releases (Part B releases)

Any person intending to carry out a research and development release of a GMO must first submit a Part B notification to the competent authority of the Member State in which the release is to take place. The notification must include a technical dossier supplying information necessary for evaluating foreseeable risks that the GMO(s) may pose to human health and the environment.

The competent authority of the Member State that receives the notification examines it for compliance with the Directive and evaluates the risks posed by the release. The information provided must be sufficient to enable a decision to be made, but further information and clarification can be sought at any time. A decision must be made within 90 days of receiving the notification, during which period comments may be received from other Member States who are informed of the notification soon after its receipt. The decision on whether or not to grant consent is made on the basis of safety to human health and the environment. No other criteria are considered in the decision making process. The notifier may proceed with the release only after receipt of written consent of the competent authority, and in conformity with any conditions required in the consent.

Marketing GMOs in the EU (Part C releases)

A Part C notification for placing a GMO on the EU market must be submitted to the competent authority of the Member State where the product is to be placed on the market first.

The competent authority then reviews the notification and forms an opinion, which must be within 90 days of receipt. If that opinion is favourable, the lead competent authority will forward the dossier to the European Commission. The Commission circulates the dossier to the other 14 Member States who are given 60 days to evaluate the application in detail, taking into account the particular health and environmental safety issues unique to their territories. If no objections are made, the lead competent authority issues the marketing consent, which applies throughout the European Community.

If, however, another Member State objects to the GMO being placed on the market, then the Directive provides for committee procedures ('Article 21 Committee Procedure') to come to a resolution. The European Commission drafts a decision that reflects the concerns raised by Member States, on which a vote is taken using the qualified majority voting procedure. If a resolution is not possible, it falls to the Council of Environment Ministers to decide. If the Council fails to decide within three months, the Commission can adopt its proposal.

United Kingdom Legislation

In the UK, Directive 90/220/EEC is implemented by Part VI of the Environmental Protection Act 1990 and the Genetically Modified Organisms (Deliberate Release) Regulations 1992[7], as amended in 1995[8]8 and 1997[9]. Consents for releases in England are signed on behalf of DETR and MAFF Ministers, acting jointly, with the agreement of the Health and Safety Executive. Consents for releases in Wales or Scotland are signed on behalf of the appropriate Ministers of the devolved administrations, again with the agreement of the Health and Safety Executive.

In making decisions on releases and marketing of GMOs, the Government is advised by the statutory Advisory Committee on Releases to the Environment (ACRE), established under Part VI of the Environmental Protection Act 1990. ACRE consists of a number of independent experts (currently thirteen), each specialising in a particular discipline, such as plant ecology, toxicology, farming, entomology, and microbiology. The joint Secretariat of ACRE is provided by the Department of the Environment Transport and the Regions (DETR), and all interested Government Departments attend meetings, together with a representative of the Statutory Nature Conservation bodies.

ACRE's main statutory role is to advise the Government on the safety of proposed releases and marketing of GMOs. Consent applications, are evaluated critically by experts on ACRE, and only if the risks of the proposed release or marketing of a GMO are considered to be low will the Committee advise that consent may be issued. . When making decisions in issuing a consent, the Ministers will also take account of safety issues raised by experts in other Government departments, the Statutory Nature Conservation Agencies and the general public.

Recent Changes to Directive 90/220/EEC

In 1998, negotiations started on the amendment of Directive 90/220/EEC. Many procedural changes have been made to the Directive. Among these are two new annexes, one on Risk Assessment (Annex 2) and another on Post-Market Monitoring (Annex 7). Until December 1998, the EU competent authorities had not agreed a harmonised approach to risk assessment. Therefore, it was important that some clear principles that reflect current best practice in EU Member States for risk assessment were formalised.

Annex II: Glossary of terms used in this guidance

Agrobacterium tumefaciens
A naturally occurring soil bacterium that is capable of inserting DNA (genetic information) into plants. Used in agricultural biotechnology to carry transgenes into plants.

Biological containment
Where biological properties of an organism (its traits) limit an organisms contact with the environment. This includes intrinsic barriers to horizontal gene transfer.

Chloroplast
An organelle found in the green tissues of plants - chloroplasts give plant tissues their green colour - which contain their own genetic material (genome). Up to 10,000 plastids can be present in a given plant cell.

Genetically modified organism (GMO)
An organism in which the genetic material has been altered in a way that does not occur naturally by mating and / or natural recombination.

Gene expression
A gene is active (switched on and expressed) when it produces a gene product (protein or RNA) and inactive (switched off) when it does not produce any products.

Hazard
Any characteristic (biological, chemical or physical) of the genetically modified organism which, in particular circumstances, could lead to harm.

Homoplasty
A plant containing a uniform population of transformed plastid genomes.

Hybrid system
A plant breeding system to facilitate hybrid seed variety development in plant breeding programmes. Typically comprising two components: one line which does not produce pollen and another line which restores fertility in hybrid progeny.

Hybridisation barrier
Natural or artificial barriers between plants which restrict the transmission of genes (gene flow) between plants.

Intron
An intervening sequence of DNA found in some plant genes, which must usually be removed before gene expression can take place.

Maternal inheritance
Genes transmitted to the next generation via ovum and not pollen.

Male sterile plant
A plant which cannot produce pollen.

Marker gene
A gene which facilitates the identification of organisms which have taken up recombinant DNA molecules.

Plant transformation
The process by which transgenes are inserted into plants. Biolistics and Agrobacterium tumefaciens – mediated transformation are two examples of plant transformation methods in current use.

Plastid
Plastids are plant specific organelles, such as chloroplasts, which carry their own DNA genomes.

Promoter
A DNA sequence which regulates gene expression.

Recombinant DNA technology
A range of biochemical techniques which enable the precise cutting and joining of DNA molecules (genes) at will in a test-tube and subsequent introduction into organisms.

Reporter gene
A class of marker gene where the product reacts with a chemical to produce a detectable coloured compound, fluoresces, or emits light and enables a tagged trait gene to be identified.

Recombinase
An enzyme that can specifically recombine DNA sequences e.g. CRE enzyme.

Risk
A combination of the probability, or frequency, of occurrence of a defined hazard and the magnitude of the consequences of the occurrence.

Risk Assessment
A process of evaluating risks involving hazard identification, estimating likelihood of its occurrence and magnitude of the consequences.

Risk management
Measures taken to minimise risk e.g. removal of flowers to eliminate cross pollination of nearby crops.

Risk management traits
Use of traits to limit an organisms contact with the environment (see biological containment).

Sexually compatible plants
Plants which can exchange genetic material by sexual hybridisation involving pollination, fertilisation and resulting in seed development.

Sequencing
The determination of the basic structure (nucleotide sequence) of DNA or RNA molecules.

Transgene
A gene introduced into an organism (usually originating from a different organism) using recombinant DNA technology.

Target organisms/Non-target organisms
Target organisms are organisms such as, for example, agronomic pests against which a genetically modified plant has been modified to destroy. Non-target organisms are organisms that are not intentionally targeted.

T-DNA
Part of the Ti plasmid, which is transferred from the soil bacterium Agrobacterium tumefaciens to a recipient plant cell. A natural process that has been utilised to introduce transgenes into plants to genetically modify them.

Transposon
A genetic element found in nature able to replicate and insert a copy at a new location in the genome.

Unique genetic identifiers
Genetic sequences used to identify and trace genetically modified organisms and their transgenes.

Vector
A Recombinant DNA molecule (e.g. a plasmid) designed to manipulate and introduce transgenes into recipient organisms.

1: The term "genetically modified organism" used in this guidance refers to the definition used in section 106 of the Environmental Protection Act 1990.
2: In the UK, Directive 90/220/EEC is implemented by the Genetically Modified Organisms (Deliberate Release) Regulations 1992 (amended 1995 and 1997), and Part VI of the Environmental Protection Act 1990.
3: Annex II of Directive 90/220/EEC.
4: UK Government and Devolved Administrations of Scotland, Wales and Northern Ireland.
5: Unique identifiers and detection methods have a key role to play in identity preservation of GM and Non GM supply chains. The Consumers demand for choice is likely to be the main driver for this technology rather than the need to monitor for safety reasons.
6: Council Directive of 23 April 1990 on the deliberate release into the environment of genetically modified organisms (90/220/EEC); OJ N0 L117/15.
7: The Regulation & Control of the Deliberate Release of Genetically Modified Organisms; DoE/ACRE Guidance Note 1.
8: Guidance to the Genetically Modified Organisms (Deliberate Release) Regulations 1995; DoE/ACRE Guidance Note 7.
9: Guidance to the Genetically Modified Organisms (Deliberate Release and Risk Assessment - Amendment) Regulations 1997; DoE/ACRE Guidance Note 10.

Published 23 October 2000
Sub-group on Best Practice in GM crop Design Index
ACRE Index
Environmental Protection Index
Defra Home Page