In the last few decades plant breeding strategies have changed considerably. In earlier times conventional crossings, followed by related selection of plants and backcrossing, were the primary approach. Nowadays molecular biological tools have clearly shown to be an important asset to this basic approach, whether by the application of marker assisted selection by ‘classic’ or more recent methods of mutagenesis, or by genetic modification using recombinant DNA techniques. In the latter case the resulting plant product should be regarded as a genetically modified organism (GMO), according to most legislations in the world. Moreover, in recent years new plant breeding technologies have been developed of which the regulatory status of the resulting product, in terms of GMO or not, is not directly evident: for instance, the European Commission is currently assessing seven new breeding methods as to whether the resulting plant varieties should be considered GM plant varieties, or not.
In general it can be stated that most legislations in relation to new plant varieties have developed a process-based approach, for instance in Europe, Brazil, Australia, and China, where the techniques applied in the basic plant breeding strategy determine the procedure for market approval. In other countries, for instance in the United States (USA) and Canada, there is a product-based legislation, where the characteristics of new plant varieties determine to a large extent the procedure for market approval. It should be noted, however, that in these countries GMOs and derived products usually also have a special status, and will in general be assessed in all cases. It has been shown that, in spite of this fundamental difference in legislative approach, in practice the data requirements for GMOs are very similar in different parts of the world (Kleter et al., 2001; MacDonald, 2012).
In recent years there have been a number of publications from scientific platforms and in the scientific literature that underwrite that, in the light of current and foreseen developments in plant breeding strategies, a product-based legislation is best in line with scientific views (for example ACRE, 2013; Kok et al., 2008; Podevin et al., 2012; House of Commons, 2015) and a science-based approach, but this has only to a limited extent been substantiated by a scientific analysis of plant breeding practices.
The present paper discusses whether the food, feed and environmental safety assessment of all new plant varieties could best be performed according to a process-based or a product-based approach in the light of the latest developments in plant breeding strategies. It is investigated whether the knowledge on current plant breeding strategies warrants the current distinction in genetically modified (GM) versus conventional plant varieties. The possibility to discriminate plant varieties based on new plant breeding technologies is studied to conclude on a process or a product-based food, feed and environmental safety assessment.
Plant breeding techniques – type and frequency of mutations
According to the regulatory status of different plant breeding techniques in the EU, new plant varieties can be distinguished in conventional varieties, genetically modified varieties and varieties that result from ‘new breeding techniques’, based on the breeding techniques used (see Figure 1). To assess the potential risk of new plant varieties in relation to food/feed and environmental safety, it is essential to know whether there are specific breeding approaches that inherently are more likely to result in unintended effects on food, feed and environmental safety. Hazards for food/feed and the environment can be qualified as potential unintended effects that relate to the breeding strategy.
At the same time, it needs to be realised that genetic changes in plant varieties do occur as a result of specific plant breeding applications such as crossing or selection. When we lack knowledge about the process which led to these mutations we call them spontaneous, when human activity was involved we attribute the mutation to this involvement. The occurrence of unintended changes related to individual plant breeding strategies may be assessed by the reported mutation frequencies for the different breeding strategies. In addition to the potential unintended effects related to the breeding strategy, i.e. the genetic alterations as a result of the breeding process that includes both the applied technique and subsequent steps to generate plants, also potential unintended side effects that are related to unforeseen metabolic interactions of the introduced trait, the so-called secondary trait effects (STEs), may occur. These STEs are not included in this overview, as they are linked to the specific expressed gene product and related trait and not to the breeding technique itself. These STEs therefore have to be assessed on a case-by-case basis.
The spontaneous mutation frequency in nature depends on several factors. It is clear that in practice it will be difficult to distinguish spontaneous mutations from mutations that have been induced by the subsequent plant breeding steps. There is limited scientific literature on the topic of spontaneous mutations. Azaiez et al. (2006) concluded that the mutation frequency in microsatellites increases with the number of repeats and the orientation of the repeats, but also that this may differ considerably in comparable microsatellites in different species (Azaiez, et al., 2006). According to Nishant et al. (2009), recent work in a variety of organisms has shown that mutation rate is strongly affected by sequence context, i.e. mutations occur more often in repetitive regions, and that there is considerable variation in mutation frequencies in different genomic regions.
In general, mutations fall into one of three categories: single nucleotide mutations, insertions/deletions and chromosome rearrangements (Ossowski et al., 2010). Single nucleotide mutations can result from exposure of the genome to endogenous and exogenous mutagens (Friedberg et al., 2004; Gady et al., 2009; Cheng et al., 2014). These DNA damage events can occur at high frequency (Nishant, et al., 2009). In the absence of DNA repair systems, the replication of DNA containing damaged bases (e.g., oxidation, deamination, and alkylation) can generate mutational events at a high rate, primarily through the loss of template information and/or the recruitment of low-fidelity DNA polymerases that display error rates as high as 1 per 100–1,000 nucleotides incorporated (McCulloch & Kunkel, 2008). Mutations in the genome can also arise as a result of errors during DNA replication, but these events are rare (McCulloch & Kunkel, 2008). Insertions and deletions can involve single base pairs, entire genes, or larger chromosomal regions (Nishant et al., 2009). Large deletions, insertions and gene duplications are the major cause of genome reshuffling (Filkowski et al., 2004). Mutations occur in coding and non-coding regions and can be broadly classified as lethal, deleterious, neutral, or beneficial based on their fitness effects. Essential genes containing simple sequence repeats within open reading frames are at risk for disruption (Nishant, et al., 2009). Thus, in nature different spontaneous mutations can arise with various mutation frequencies. However, the frequency of spontaneous mutation is very low and difficult to use in plant breeding (Ahloowalia et al., 2004; Jain, 2010).
In the next sections the respective categories will be discussed with respect to the observed types of mutations in new plant varieties, as well as the frequency with which the mutations are observed on the basis of available scientific literature.
Conventional varieties are plants created using conventional breeding techniques, i.e. methods used by plant breeders for the improvement of commercial varieties and where the resulting plants/varieties are not covered by Directive 2001/18/EC (EFSA, 2012). Conventional breeding makes use of genetic variation that is present in other, closely related species in order to enlarge the gene pool of a plant (Van de Wiel et al., 2010). An important group of conventional varieties are plant varieties that have been created using mutagenesis by chemicals or ionising radiation. These mutagenesis approaches cause random mutations in the DNA of crops, this form of plant breeding is often referred to as mutation breeding (Lusser et al., 2011; Van de Wiel et al., 2010). Other examples of conventional breeding techniques are crossing, embryo rescue, and somatic hybridization (cell fusion) of crossable species (Van de Wiel et al., 2010).
Classical plant breeding improves plant varieties through crossing of a well-known and well-performing plant line with a plant line, often a wild relative, which contains an additional, desired trait. When a direct cross between two species is not possible, an intermediate crossing with a third species, which is compatible with both species, may bridge the crossing barrier. By this strategy of indirect crossing, the genomes or segments thereof from the two species can be combined (Van de Wiel, et al., 2010). Potential side effects of conventional breeding strategies by crossing are inversions, deletions and reorganisations of chromosomal fragments. Additionally, a potential side effect of crossing is linkage drag (Kok, et al., 2008; Jacobsen, 2013). Linkage drag is the co-inheritance of an undesirable trait(s) with a gene of interest. Typically, this is a phenomenon that will often be applied during backcross breeding using an exotic donor parent (Peleman & Rouppe van der Voort, 2003). The mutation frequency in the case of conventional crossing is not known.
Embryo rescue is the isolation of immature embryos from developing seeds and their culture in the laboratory to produce hybrid plants (Suslow et al, 2002). It is a type of a tissue culture method and permits sexual hybridisation between plants that show post fertilisation incompatibility (e.g. due to endosperm defects) (Wilson et al., 2004). In this way embryo rescue produces hybrid plants (Van de Wiel et al., 2010). Especially the tissue culture step contributes to observed genetic and epigenetic changes (Filipecki & Malepszy, 2006). The mutagenic potential depends on the method: tissue culture methods based on dedifferentiation lead to increased genetic damage and variance compared to those which do not include a step of dedifferentiation (e.g. vegetative propagation or embryo rescue) (Wilson, et al., 2004). The exact mutation frequency in the case of breeding programs based on embryo rescue in combination with the different types of tissue culture steps, however, is not known.
In the case of somatic hybridisation, cytoplasm and nuclear content of different cells of two (crossable) species are mixed, leading to hybrids with a summation of chromosomes (Schaart & Visser, 2009). The somatic cells are combined by fusing the membranes of the cells (Van de Wiel, et al., 2010). When the parental cells originate from crossable species that are not closely related, the final chromosomal constitution of the hybrids can, in rare cases, show aberrations in number and composition. The mutation frequency in the case of breeding programs based on somatic hybridisation of crossable species is not known.
Finally, mutation breeding aims to result in mutations by treatment of plant parts with ionizing radiation (Ahloowalia & Maluszynski, 2001; Ahloowalia, et al., 2004; Jain, 2010; Shirley et al., 1992; Srinivas & Veerabadhiran, 2010) or chemical mutagens using substances such as sodium azide (Srinivas & Veerabadhiran, 2010; Van de Wiel, et al., 2010), the alkylating agents EMS (Greene, et al., 2003; Jain, 2010; Jander, et al., 2003; Van de Wiel, et al., 2010), methyl methanesulfonate (MMS) (Kovalchuk et al., 2000), or N-ethyl-N-nitrosourea (ENU) (Gichner, 2003). The different agents may have different effects (Kok, et al., 2008). Mutation breeding has been highly successful. The wide use of mutation induction for crop improvement is documented in the FAO/IAEA Mutant Variety Database (http://mvgs.iaea.org/AboutMutantVarities.aspx), which includes more than 3,200 officially released mutant varieties from 214 different plant species in more than 60 countries throughout the world. Over 1,000 mutant varieties of major staple crops have sofar been cultivated on tens of millions of hectares. The resulting mutations can be of diverse nature. Point mutations are more often observed in the case of chemical mutagenesis, whereas chromosome breakage, duplication or elimination is more common when irradiation mutagenesis is applied (Van de Wiel et al., 2010). Mutation frequencies as reported in the scientific literature range considerably, from 20 mutations/~1 kb gene fragment in Arabidopsis (Martin et al., 2009) to 0.75-1.36 mutations/1000kb in EMS-treated tomato populations (Gady et al., 2009; Rigola, et al., 2009; Saito et al., 2009), in the case of chemical mutagenesis. In the case of irradiation mutagenesis the mutation frequencies are largely dependent on the irradiation dosage applied, but can be in the same range as for chemical mutagenesis (e.g. Kovalchuk et al. (2000).
Genetic modified (GM) varieties
Genetic modified (GM) varieties are plants created using techniques of genetic modification that alter the genotype of an organism in a way that does not occur naturally by mating and/or natural recombination (EC, 2001). EU Directive 2001/18/EC specifies these techniques of genetic modification. A type of genetically modified varieties are transgenic varieties that are developed using recombinant DNA techniques for the introduction of genetic information into plant cells that leads to the transmission of the input gene (transgene) to successive generations (EFSA, 2012). Unintentional changes to the genome in the case of genetic modification may occur in addition to the desired effects of the introduction of the new gene(s). These alterations can be caused by various processes and mechanisms, for instance genome disruptions, deletions or rearrangements that may result in the silencing of genes or the creation of new open reading frames, as well as by somaclonal variation, if a cell culture phase is part of the transformation procedure (EFSA, 2012, Bennetzen, 2002; Jiang, et al., 2011; Lusser, et al., 2011; Puchta, 2005; Wicker et al, 2010). The EFSA GMO Panel found that the mutation frequency linked to genetic modification procedures can be variable and needs to be assessed case by case (EFSA, 2012). From scientific literature no general information on the mutation frequency in the case of breeding programs based on genetic modification can be deduced.
New breeding technique varieties
New breeding technique varieties are plants created using new breeding techniques as defined by the EC (Lusser et al., 2012). These techniques are not immediately considered to be conventional breeding techniques and have not yet been categorized as techniques leading to a GMO or not under the current legal definition. Seven new breeding methods are currently being assessed by the European Union as to whether the resulting plant varieties should be considered GM plant varieties or not and, subsequently, whether they should be subject to a specific safety assessment or not. These new breeding techniques include Zinc Finger Nuclease technologies (ZFN-1, ZFN-2, ZFN-3), Oligonucleotide Directed Mutagenesis (ODM), cisgenesis / intragenesis, RNA-dependent DNA methylation (RdDM), grafting on GM rootstocks, reverse breeding, and agrofiltration (encompassing agro-infiltration ‘sensu stricto’, agro-infection and floral dip). Here the different techniques will be discussed in terms of observed mutation frequencies, as far as known, with the exception of grafting, as this is not a molecular biological technique as such, but rather a way to combine GM and non-GM plant parts.
Zinc Finger Nuclease (ZFN) technology is a technique for targeted mutagenesis as a first step in plant breeding. It is a highly specific DNA targeting tool allowing specific changes of a particular nucleotide sequence. ZFN technology is based on the use of ZFNs which are custom-made proteins to cut at specific DNA sequences (Lusser, et al., 2011). In the cell, the ZFN complex recognises the target DNA site and generates a double strand break (DSB) at a specific genomic location. This stimulates native cellular repair processes: homologous recombination and non-homologous end-joining, thus facilitating site-specific mutagenesis. Distinction is made between ZFN-1 (without a repair template) and ZFN-2 (with a repair template) and ZFN-3 (targeted introduction of new genes) techniques (Lusser, et al., 2011). Unintended effects of ZFN-technology are likely nonspecific mutations due to non-specific binding of the ZFNs. ZFNs do not always have the desired sequence specificity and affinity, since not all the ZFNs bind to their cognate DNA triplets in a highly sequence-specific manner. The mutation frequency for the ZFN-1 and -2 techniques reported in different publications vary, but are usually rather low (Lusser, et al., 2011). For example, Wright et al. (2005) found a homologous recombination frequency of 7.8 × 10−4 in tobacco (ZFN-2), and Lloyd et al. (2005) concluded that ZFN can generate mutations at specific sites within the Arabidopsis genome at frequencies as high as 0.2 mutations per target (ZFN-1). In the case of ZFN-3 technology, unexpected effects due to so-called ‘position effects’ are expected to be less in comparison to other methods for genetic modification as the genetic construct is not inserted randomly into the genome, but directional (Lusser, et al., 2011). The insertion frequency for the ZFN-3 approach reported in different publications vary, but are usually also rather low (Lusser, et al., 2011). Cai et al. (2009) showed that designed ZFNs can be used to drive targeted transgene integration into an endogenous tobacco gene locus at a frequency up to 10% of the total events generated (Cai, et al., 2009). Similar techniques, already widely adopted in research, such as transcription activator-like effector nucleases (TALENs), LAGLIDADG homing endonucleases, also termed meganucleases, and clustered, regularly interspaced, short palindromic repeats (CRISPRs) used to generate RNA-guided nucleases, also make use of site-directed mutagenesis and as a result may cause similar types of off-target unintended effects (Curtin et al., 2012; Sander and Young, 2014). The European Food Safety Authority (EFSA) concludes in her opinion in 2012 that this type of methods based on site-directed nucleases can induce off-target changes in the genome of the recipient plant, but that these would be fewer than those occurring with most mutagenesis techniques and be basically of the same types as those produced by conventional breeding techniques (EFSA, 2012).
Oligonucleotide directed mutagenesis (ODM) is a technique that is based on the use of oligonucleotides for the induction of targeted mutations in the plant genome, usually of one or a few adjacent nucleotides (Lusser, et al., 2011). ODM technology is expected to generate fewer unintentional changes or effects than those generated by breeding techniques based on irradiation or chemical mutagenesis. The genetic changes that can be obtained using ODM include the introduction of a new mutation, the reversal of an existing mutation, or the induction of short deletions (Lusser, et al., 2011). With respect to oligonucleotides that interact with DNA, the Dutch advisory body COGEM considers that this can lead to sequence changes although the likelihood of this is low (COGEM, 2006). Ruiter et al. (2003) reported that the frequency of oligonucleotide-directed mutations of a chromosomal gene in oilseed rape was not higher than that of the spontaneous mutation rate at approximately once in every 2×107 (Ruiter, et al., 2003). However, Zhu et al. (1999) found that the overall frequency of site-specific targeting in maize is 10−4, which is considerably higher than the frequency of spontaneous mutation (10−5–10−8, depending on the locus, Drake et al., 1998), and gene targeting by homologous recombination (10−5–10−7) in plant cells. Other reports on tobacco cells (Beetham et al., 1999; Kochevenko & Willmitzer, 2003) showed 10–20 times higher frequency of oligonucleotide-induced mutations than spontaneous mutation.
Cisgenesis aims to form new combinations of genetic material from the species’ own gene pool. Cisgenic constructs based on genes and regulatory elements from the same species, are introduced into the plant using recombinant DNA techniques, but as elements from the same species have been used similar products could also be produced by using conventional breeding approaches (Lusser, et al., 2011; Prins & Kok, 2010; Schaart & Visser, 2009). Potential unintended effects of cisgenesis may, similar to transgenesis, be due to insertional mutagenesis at or by the insertion site(s) (Prins & Kok, 2010). Deletion of host DNA can also occur following insertion. Cisgenesis may lead to, for instance, modified levels of expression of the target gene(s), or to gene silencing. The EFSA GMO Panel found that the mutation frequency linked to cisgenesis is likely to be variable, but there are no data as yet available from scientific literature (EFSA, 2012). Intragenesis is similar to cisgenesis, but here some modifications may be applied to the native sequence, or other regulatory elements have been used derived from the same species’ gene pool. Intragenesis can also use silencing approaches, e.g. on the basis of RNA interference (RNAi), by introducing inverted DNA repeats (Lusser, et al., 2011). Potential unintended side effects of intragenesis are very similar to transgenesis and cisgenesis. In case of intragenic crosses, the intragenic expressed protein is likely to be very similar to the conventional protein, but the small changes in coding sequences and/or regulatory elements may lead to extended changes in the plant’s physiology (Prins & Kok, 2010; Schaart & Visser, 2009). The EFSA GMO Panel found here also that the mutation frequency will be variable and needs to be assessed case by case (EFSA, 2012).
RNA-dependent DNA methylation (RdDM) induces gene silencing of targeted genes via the methylation of promoter sequences. For targeted RdDM, genes encoding RNAs which are homologous to promoter regions are delivered to the plant cells by suitable methods of transformation. These genes result in the production of double stranded RNAs (dsRNAs) which, after processing by specific enzymes, induce methylation of the target promoter sequences thereby inhibiting transcription of the target gene (Lusser, et al., 2011). RdDM plants do not contain foreign DNA sequences and no changes or mutations are made in the nucleotide sequence, but gene expression is modified due to the resulting methylation (Lusser, et al., 2011). Unintended effects could be the loss of silencing of the specific gene in the commercial product, and silencing of genes with homologous promoter sequences (Lusser, et al., 2011). The methylated status can continue for a number of generations following the elimination of the inserted genes. The epigenetic effect is assumed to decrease in subsequent generations and to eventually fade out (Lusser, et al., 2011). It is not clear for how long the effect of gene silencing by RdDM generally remains in the absence of the inducing construct. Silencing, and differences in silencing, have been observed to be transmitted to at least the F3 generation (Lusser et al 2011).
Through reverse breeding, homozygous parental lines are produced from selected heterozygous plants. The reverse breeding technique makes use of a step of genetic modification to suppress meiotic recombination. Subsequently, the obtained homozygous lines are hybridised, in order to reconstitute the original genetic composition of the selected heterozygous plants. In subsequent steps, only non-GM plants are selected. Therefore, the offspring of the selected parental lines would genotypically reproduce the elite heterozygous plant and would not carry any additional genomic change (Lusser, et al., 2011). No unintended effects are expected; the genome of the plant has not been modified. The plants are in principle identical to the original parental lines of the original heterozygous line (the starting material) (COGEM, 2006; Schaart & Visser, 2009), if the selection step for non-GM plants is performed adequately. However, no data could be retrieved from scientific literature to underpin this assumption.
Discussion on process-based versus product-based
The present paper provides an overview of different breeding techniques and the related potential for unintended effects that directly relates to the genetic modification achieved by the respective breeding techniques, as well as the estimated associated frequencies for the different mutations. Table 1 shows an overview of the type of changes (intended and potential side effects) and the likelihood of unintended effects per breeding technique based on literature data. This overview was compiled with the aim to assess whether it is possible to draw a line between groups of techniques, i.e. those that have higher potential for unintended effects and those that have lower potential for unintended effects in the final new plant variety. From this overview, however, it can be concluded that in most cases few or no data are available to assess the type of unintended effects that can be considered to be related to a specific plant breeding techniques, and likewise few or no data are available with relation to the associated mutation frequencies. In those cases where data are available on mutation frequency, the data are fragmentary and often wide-ranging, providing no solid basis for further generalisation on unintended effects and related mutation frequencies for specific plant breeding techniques. Secondly, analysing the reported intended and unintended effects obtained by the plant breeding techniques it appears that genomic changes can be considered to be variable, but similar for all of the breeding techniques, and should be regarded as a continuum rather than associated with specific categories of breeding strategies. For example, for new plant varieties that do not contain any ‘new’ or ‘foreign’ DNA in the resulting offspring, unintended changes are likely to be similar to those obtained by conventional breeding, ODM, ZFN-1 and -2, cisgenesis, reverse breeding, and RdDM (see Table 1). But at the same time all these breeding strategies may in specific cases also lead to larger physiological changes than in specific cases of transgenesis. Unintended side effects can be related to any of the plant breeding techniques mentioned in this overview, and likely also to others that have not been dealt with in this overview, but the effects and associated frequencies of the mutations cannot easily be predicted and insufficient data are available to relate them to specific techniques. Indications are, to the contrary, that potential unintended side effects of a certain plant breeding technique can only be assessed on a case-by-case basis rather than generalising on the basis of method(s) used. The differences between the regulatory view on current plant breeding methods and their associated risks and the related picture as can be deduced from the scientific literature are depicted in Figure 2. Indeed, the breeding technique applied should not determine whether a safety assessment has to be performed and what the nature of the assessment should be, if any, but this should rather be determined by the product obtained.
This should then be reflected in the food, feed and environmental safety assessment of the resulting plant variety. Hazards for food and feed safety as well as for the safety of the environment can be related to, on the one hand, the intended effects of the plant breeding strategy, i.e. the altered or new trait, or can be the result of potential unintended side effects of the selected breeding strategy. When focusing on the unintended effects that may relate to the different breeding techniques, it is shown in this overview that there is no scientific basis to distinguish between categories of different techniques that may have higher or lower potential for unintended side effects. Unintended effects related to the intended effect (i.e. the novel trait) have not been taken into account in the current overview, but as mentioned above should be assessed on a case-by-case and depending on the novelty of the trait.
Current regulations within Europe, but also in many other parts of the world are based on the assumption that some breeding techniques, such as recombinant DNA techniques, require a more rigorous assessment in terms of safety for consumers, animals and the environment, than others. Even in countries that have a product-based approach it is common practice to assess plant products that have been obtained with recombinant-DNA techniques in all cases, leaving products from other techniques to be assessed on a case-by-case basis.
This approach is not supported by scientific evidence. The data available from the scientific literature as provided in the present study only allow for a fully product-based approach, where unintended effects are assessed and evaluated on a case-by-case basis.
The present paper discusses whether the food, feed and environmental safety assessment of all new plant varieties could best be performed according to a process-based or a product-based approach in the light of the latest developments in plant breeding strategies. In Europe so far a process-based approach has been applied, requiring considerably more data for new plant varieties that have been obtained using genetic modification, compared to varieties that are obtained by other means. For another group of new plant breeding techniques it has not yet been decided into which of these two categories they should fall.
From the overview in this study it can be concluded that the breeding techniques used for new plant varieties cannot be traced by the type of genetic changes and, as a consequence, there is no scientific rationale for the technique to be decisive for performing a safety assessment. Rather, it is more logical to base the safety assessment on the new characteristics of the derived new plant variety and the derived food and feed products, in other words to follow a product-based approach. Based on the data from the present study only a truly product-based approach, assessing each new plant variety on its own merits in terms of altered characteristics and related hazards, is justified. This will better serve the purpose of maintaining a safe food and feed supply and minimizing potential adverse effects on the environment; also in the longer term.
Paste your essay in here…