24. The GM canola proposed for release is the product of conventional breeding between InVigor® canola approved for commercial release under DIR 021/2002, held by
Bayer, and Roundup Ready® canola approved for commercial release under DIR 020/2002, held by Monsanto Australia Ltd (Monsanto). These risk assessments are
available by contacting the OGTR. Information from these assessments is summarised below, and new information included where available.
25. Seven elite GM canola lines (T45, Topas 19/2, MS1, MS8, RF1, RF2 and RF3) were authorised for commercial release under licence DIR 021/2002. All seven GM canola lines contain a gene conferring tolerance to the herbicide glufosinate ammonium (Table 1 and Table 2). In addition, lines MS1, MS8, RF1, RF2 and RF3 contain genes comprising a hybrid breeding system. Lines Topas 19/2, MS1, RF1 and RF2 also contain an antibiotic resistance gene.
26. The MS and RF lines, and hybrids derived from MS x RF crosses, are covered by the registered trade name InVigor® canola. The hybrid derived from the cross between MS8 and RF3 lines is licensed as InVigor® Hybrid canola for release in Australia under DIR 021/2002. The other lines approved under DIR 021/2002 (T45, Topas 19/2, MS1, RF1 and RF2) were not intended for commercial release in Australia.
27. Roundup Ready® canola has been genetically modified by transformation event GT73 to express two genes conferring tolerance to the herbicide glyphosate (Table 1 and Table 2).
Table 1 The genes introduced into the parental GM canola lines
| GM canola line | Glufosinate ammonium tolerance | Glyphosate tolerance | Hybrid breeding system | Antibiotic resistance |
|---|---|---|---|---|
| GT73 | - | Cp4 epsps and goxv247 | - | - |
| T45 | pat | - | - | - |
| Topas 19/2 | pat (2 copies) | - | nptll (2 copies) | |
| MS1 | bar | - | barnase | nptll |
| MS8 | bar | - | barnase | - |
| RF1 | bar | - | barstar | nptll |
| RF2 | bar | - | barstar | nptll |
| RF3 | bar | - | barstar (2 copies) | - |
Table 2 Genetic elements and their origin
| Gene (source) | Protein produced | Protein function | Promoter (source) | Terminator (source) | Additional elements (source) |
|---|---|---|---|---|---|
| cp4 epsps (Agrobacteriem sp. strain CP4) | CP4 EPSPS | tolerance to the herbicide glyphosate | P-CMoVb (ligwort mosaic virus) | E9 3' (Pisum sativum) | AEPSPS/CTP2 (Arabidopsis thaliana) |
| goxv247(Agrobacteriem sp. strain CP4) | glyphosate oxidoreductase | tolerance to the herbicide glyphosate | PCMoVb (ligwort mosaic virus) | E9 3' (Pisum sativum) | SSU1A/CTP1(Arabidopsis thaliana) |
| bar (Streptomyces hygroscopicus | phosphinothricin acetyl transferase | tolerance to the herbicide glufosinate ammonium | PSsuAra(Arabidopsis thaliana) | 3' g7 (Agrobacterium tumefaciens) | - |
| pat (Streptomyces viridochromogenes | phosphinothricin acetyl transferase | tolerance to the herbicide glufosinate | P-35S (Cauliflower mosaic virus) | T-35S (Cauliflower mosaic virus) | - |
| barnase (Bacillus amyloliquefaciens) | Barnase (RNase) | Male sterility | PTA29 (Nicotiana Tabacum) | 3'-nos (Agrobacterium tumefaciens) | - |
| barstar (Bacillus amyloliquefaciens) | Barstar (RNase inhibitor) | Restoration of fertility | PTA29 (Nicotiana tabacum) | 3'-nos (Agrobacterium tumefaciens) | - |
| nptII (Escherichia coli) | neomycin phosphotransferase | resistance to antibiotics such as kanamycin and neomycin (selectable marker) | P-nos (Agrobacterium tumefaciens) | 3'-ocs (Agrobacterium tumefaciens) | - |
Top of page
5.1 The introduced genes, their encoded proteins and their associated effects
5.1.1 Hybrid breeding system genes and their encoded proteins28. Traditional plant breeding selects for plants with agronomically valuable characteristics. However, repetitive self-pollination of desirable lines can produce progeny that display lowered fitness or vigour when compared to their out-crossing counterparts, a phenomenon termed inbreeding depression. By contrast, when crosses are made between genetically distinct parents, the progeny often outperform the parental lines and are said to display hybrid vigour. Hybrid vigour is commercially advantageous, but ensuring a hybrid cross is technically difficult to achieve, especially when working with species that have both male and female floral organs borne on the same flower and are predominantly self-fertilising, such as canola.
29. To facilitate the production of hybrid canola plants, Bayer has developed a hybrid breeding system that is conferred by expression of the barnase and barstar genes derived from the common soil bacterium Bacillus amyloliquefaciens. Barnase encodes a ~12kD ribonuclease (RNase) called BARNASE, and barstar encodes a ~10kD RNase inhibitor protein, BARSTAR, which specifically binds to BARNASE and suppresses its activity (Hartley 1988; Hartley 1989).
30. RNases are commonly found in nature and collectively their function is to degrade the messenger ribonucleic acid (mRNA) that allows genetic information to be translated into protein production. This turnover of mRNA is important for regulating the activity of genes. In B. amyloliquefaciens, the BARNASE enzyme is secreted extracellularly as a defence mechanism where it degrades the ribonucleic acid of competing organisms. BARSTAR accumulates intracellularly to protect the host cell from the destructive properties of its own ribonuclease enzyme.
31. In the GM canola lines MS1 and MS8, barnase is controlled by a promoter that directs gene expression solely within the tapetal cell layer of the anthers. This results in localised degradation of ribonucleic acid within the tapetal cells prior to microspore development and prevents the production of pollen (Mariani et al. 1990; De Block & De Bouwer 1993). The resulting plants are male-sterile (MS) and can only be fertilised by the pollen of another plant, thereby ensuring the production of hybrid progeny.
32. To reverse the effects of barnase expression, GM canola lines have also been generated that contain the barstar gene. The introduced barstar gene in GM canola lines RF1, RF2 and RF3, is under the control of the same tapetum-specific promoter. Expression of barstar has no effect on pollen development and GM canola plants have a normal appearance and viable pollen (Mariani et al. 1992). When a GM line containing barnase is crossed with a GM line containing barstar, progeny that inherit both genes display completely normal fertility due to the specific inhibition of BARNASE activity by BARSTAR (Mariani et al. 1992). For this reason, GM lines modified with the barstar gene expressed from a tapetum-specific promoter are designated as restorers of fertility (RF).
33. Control of fertility by expression of the barnase and barstar genes is the basis of InVigor® canola hybrids derived from MS x RF crosses, which display hybrid vigour resulting in increased yields over the parental varieties. InVigor® Hybrid canola resulting from the cross between MS8 and RF3 was approved for commercial release under licence DIR 021/2002.
5.1.2 Herbicide tolerance genes and their encoded proteins
Glufosinate ammonium tolerance
34. Glufosinate ammonium is the active ingredient in a number of proprietary broad-spectrum herbicides that have been registered for use in Australia, including Basta®, Finale® and Liberty®. These herbicides function by inhibiting the plant enzyme glutamine synthetase, which is a key enzyme involved in plant nitrogen metabolism. In the absence of glutamine synthetase activity, ammonia accumulates in plant tissues causing inhibition of amino acid biosynthesis, inhibition of photosynthesis and rapid death of the plant (Evstigneeva et al. 2003).
35. The herbicidal component of glufosinate ammonium is the L-isoform of phosphinothricin (PPT). PPT is a component of the antibiotic bialaphos, which is produced naturally by the soil bacteria Streptomyces hygroscopicus and Streptomyces viridochromogenes. To avoid the toxicity associated with biaphalos production, S. hygroscopicus and S. viridochromogenes express the biaphalos resistance genes bar and pat, respectively (Murakami et al. 1986; Thompson et al. 1987; Wohlleben et al. 1988; Strauch et al. 1988). Both the bar and pat genes encode phosphinothricin acetyl transferase (PAT), an enzyme that acetylates the free amino groups of PPT with high affinity and specificity to render it inactive (Wohlleben et al. 1988; Droge-Laser et al. 1994; OECD 1999b).
The bar and pat genes and their encoded proteins
36. Each of the GM canola lines authorised under licence DIR 021/2002 was modified for tolerance to glufosinate ammonium by the introduction of either the bar gene from S. hygroscopicus or the pat gene from S. viridochromogenes. The bar and pat genes are very similar with an overall identity of 87% at the nucleotide sequence level. Both genes encode PAT proteins of 183 amino acids with 85% amino acid sequence identity, comparable molecular weights (~22 kDa) and similar substrate affinity and biochemical activity (Wehrmann et al. 1996). In fact, the PAT proteins encoded by bar and pat are so similar as to be functionally equivalent for the purpose of conferring tolerance to glufosinate ammonium (Wehrmann et al. 1996; OECD 1999b).
37. The DNA sequences of both the pat and bar genes introduced into the GM canola lines approved under DIR 021/2002 were modified for plant-preferred codon usage to ensure optimal expression in canola (European Scientific Committee on Plants 1998a; European Scientific Committee on Plants 1998b; EFSA 2008).
38. The PAT protein produced from the pat gene in GM canola lines T45 and Topas 19/2 has exactly the same amino acid sequence as the native protein from S. viridochromogenes (European Scientific Committee on Plants 1998a; OECD 1999b).
39. The bar gene introduced into the MS and RF GM canola lines was modified by the substitution of the N terminal two codons of the bacterial gene, GTG and AGC, with the codons ATG and GAC, respectively (OECD 1999b; Japanese Biosafety Clearing House 2007). The modification from GTG to ATG does not result in an amino acid change, but serine changes to aspartic acid in the modification from AGC to GAC. However, the function of the PAT gene with this single amino acid substitution remains unchanged (Japanese Biosafety Clearing House 2007).
Glyphosate tolerance
40. Glyphosate is the active ingredient in a number of broad-spectrum systemic herbicides that have been approved for use in Australia and was first marketed as the proprietary herbicide Roundup®The herbicidal activity of glyphosate is derived from its ability to inhibit the function of 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS), a key enzyme involved in the shikimate biosynthetic pathway present in all plants, bacteria and fungi.
41. The shikimate pathway enables biosynthesis of aromatic compounds from carbohydrate precursors in a series of seven biosynthetic steps. The penultimate step in the pathway is the condensation of shikimate 3-phosphate and phosphoenol pyruvate to form 5-enolpyruvylshikimate 3-phosphate, a reaction catalysed by EPSPS (reviewed by Herrmann & Weaver 1999). Glyphosate competes with phosphoenol pyruvate for binding to the complex formed between EPSPS and shikimate 3-phosphate. Upon glyphosate binding, the EPSPS:shikimate 3-phosphate complex is very stable and has a slow reversal rate, effectively terminating the shikimate pathway prematurely and preventing biosynthesis of essential aromatic compounds required for plant growth and development, including the amino acids phenylalanine, tyrosine and tryptophan (Dill 2005).
42. Two main approaches have been utilised to generate GM plants that are tolerant to glyphosate-based herbicides: introduction of genes that encode proteins capable of detoxifying the glyphosate molecule; and introduction of genes that encode EPSPS enzymes with reduced affinity for glyphosate (Dill 2005).
43. Roundup Ready® canola line GT73, approved under DIR 020/2002, was modified to contain both a glyphosate detoxifying enzyme, encoded by the goxv247 gene, and an EPSPS protein with naturally reduced affinity for glyphosate, encoded by the cp4 epsps gene.
The goxv247 gene and its encoded protein
44. The goxv247 gene introduced into Roundup Ready® canola GT73 was isolated from the common soil bacterium Ochrobactrum anthropi strain LBAA (formerly Achromobacter sp.). It encodes a glyphosate oxidoreductase (GOX) enzyme that inactivates glyphosate by converting it into aminomethylphosphonic acid (AMPA) and glyoxylate (Pipke & Amrhein 1988; Duke 2010). Glyoxylate is a common plant metabolite and AMPA is degraded by several microorganisms (ANZFA 2000).
45. The goxv247 gene encodes a single polypeptide of 431 amino acids with a molecular mass of 46.1 kD. This gene is a variant of the O. anthropi gox gene and has improved affinity for glyphosate and therefore degrades the herbicide more efficiently. The DNA sequence of goxv247 was modified for plant-preferred codon usage. The goxv247 gene varies from the gox gene by only 5 nucleotides, and the variant GOXv247 protein is 99% identical to the native GOX enzyme, differing by 3 amino acids.
The cp4 epsps gene and its encoded protein
46. The cp4 epsps gene introduced into Roundup Ready® canola GT73 was isolated from the soil bacterium Agrobacterium sp. strain CP4 and encodes an EPSPS protein with naturally reduced affinity for glyphosate relative to endogenous plant EPSPS enzymes. The presence of CP4 EPSPS in Roundup Ready® canola allows the plants to complete the shikimate pathway even in the presence of glyphosate.
47. The cp4 epsps gene encodes a protein of 47.6 kD consisting of a single polypeptide of 455 amino acids. The nucleotide sequence of the bacterial cp4 epsps gene was modified for plant-preferred codon usage, but these nucleotide substitutions did not alter the amino acid sequence of the encoded protein.
5.1.3 Toxicity/allergenicity of the proteins encoded by the introduced genes
BARNASE and BARSTAR proteins
48. The barnase and barstar genes that comprise the hybrid breeding system were derived from the common soil bacterium, B. amyloliquefaciens. This bacterium is used commercially as a source of industrial enzyme production, particularly á-amylase, and is also used in the food industry for brewing and bread-making. Although some Bacillus species have been implicated as the causal agents of human diseases, B. amyloliquefaciens is not known to be allergenic or pathogenic towards humans.
49. GM InVigor® canola lines incorporating the MS and RF hybrid breeding system have been approved for limited and controlled release under DIRs 010/2001, 032/2002, 057/2004, 069/2006 and 104, and for commercial release under DIR 021/2002. Therefore, the toxicity and allergenicity of the BARNASE and BARSTAR proteins have been previously assessed by the Regulator and the assessments concluded that they are unlikely to be toxic or allergenic.
50. No sequence homology was found between BARNASE or BARSTAR and known toxins or allergens (see DIR 069/2006; Van den Bulcke 1997). Further bioinformatic studies using updated databases have confirmed these results (EFSA 2009b). BARNASE and BARSTAR do not have characteristics typical of known protein allergens (Van den Bulcke 1997) and no matches with known IgE epitopes were found (Kleter & Peijnenburg 2002). Both proteins are rapidly degraded in simulated gastric juices (0.32% pepsin and acidic pH) with complete protein degradation within five minutes (Van den Bulcke 1997), showing that these proteins would not survive in the digestive tract.
51. Feeding studies in animals using seed from GM canola lines expressing barnase and barstar are discussed in Section 5.4.5.
52. Food derived from GM InVigor® canola lines MS1, MS8, RF1, RF2 and RF3, which expresses BARNASE and BARSTAR proteins, has been considered safe for human consumption by FSANZ (ANZFA 2001b), and InVigor® canola has been grown commercially in North America since 1995 without reports of toxicity or allergenicity associated with the introduced genes.
53. BARNASE degrades ribonucleic acid into its component ribonucleotides. Ribonucleotides are ubiquitous in nature and are not considered toxic or allergenic. BARSTAR does not possess enzymatic activity but, instead, exerts its action by binding to the BARNASE enzyme to form an inactive complex. Therefore, the products of the enzymatic reactions catalysed by the novel proteins are also unlikely to be toxic or allergenic.
PAT protein
Top of page
54. The bar and pat genes were obtained from the common soil bacteria S. hygroscopicus and S. viridochromogenes, respectively. These species of Streptomyces are saprophytic, soil-borne microbes that are not considered pathogens of plants, humans or other animals (OECD 1999b).
55. The bar and pat genes have both been used extensively in the production of GM plants as selectable markers in the laboratory or to provide herbicide tolerance in the field. Consequently, PAT proteins have been used in several GM plants approved by the Regulator for limited and controlled release (for example, see DIR 71/2006, DIR 86/2008 and DIR 100). In addition, GM canola (DIR 021/2002), GM Liberty Link® cotton (DIR 062/2005) and GM WideStrike® cotton (DIR 091) expressing the bar or pat genes have been approved for commercial release. Therefore, the toxicity and allergenicity of PAT proteins have been previously assessed by the Regulator and the assessments concluded that they are unlikely to be toxic or allergenic.
56. A review of published literature and experimental studies was used to evaluate the safety of the PAT proteins encoded by the pat and bar genes (Herouet et al. 2005). The authors concluded that there is a reasonable certainty of no harm resulting from including the PAT proteins in human food or animal feed.
57. No sequence homology has been found between PAT and any known toxic or allergenic proteins (Van den Bulcke 1997; Herouet et al. 2005; EFSA 2009b). The PAT proteins do not possess any of the characteristics associated with food allergens and they are not stable in simulated gastric or intestinal fluid conditions (Wehrmann et al. 1996; OECD 1999b; ANZFA 2001b; Herouet et al. 2005) hence the potential for the PAT protein to be a food allergen is minimal (EPA 1997b). In addition, PAT proteins are inactivated by heat, low pH and during processing of canola (Wehrmann et al. 1996; EPA 1997b; European Scientific Committee on Plants 1998a; OECD 1999b).
58. There is no evidence that the PAT proteins encoded by the bar and pat genes are toxic to either humans or other animals. The potential for PAT to be toxic has been addressed via acute toxicity studies, as detailed in the RARMP for DIR 021/2002. In summary, 14-day acute oral toxicity studies in mice and rats found no treatment-related significant effects (Merriman 1996; Bremmer & Leist 1996). A more recent study found no toxic effect in mice after acute intravenous administration of the PAT proteins at up to 10 mg/kg body weight (Herouet et al. 2005).
59. Feeding studies in animals using seed from GM canola lines containing PAT proteins are discussed in Section 5.4.5.
60. FSANZ has approved the use of food derived from GM plants containing either the bar or pat gene, including GM canola, cotton, maize, rice and soybean, concluding that the PAT proteins are not toxic (ANZFA 2001a; ANZFA 2001b; FSANZ 2004; FSANZ 2005c; FSANZ 2008).
61. A number of international regulatory bodies have also assessed the PAT proteins expressed in GM plants as safe. These include the United Stated Food and Drug Administration (FDA 1996; FDA 1997; FDA 1998), Health Canada (Health Canada 1997; Health Canada 1999b), the Canadian Food Inspection Agency (Canadian Food Inspection Agency 1995c; Canadian Food Inspection Agency 1996), the European Commission (European Scientific Committee on Plants 1998a) and the European Food Safety Authority (EFSA 2008; EFSA 2009b). The United States Environmental Protection Agency has determined that PAT, and the genetic material necessary for its production, is exempt from the requirement to establish a maximum permissible level for residues in plants (EPA 1997b).
GOXv247 and CP4 EPSPS proteins
62. The goxv247 gene is derived from O. anthropi strain LBAA (formerly Achromobacter sp.), a bacterium commonly found in the soil. The goxv247 gene encodes the GOXv247 protein that differs from the original O. anthropi enzyme by three amino acids.
63. O. anthropi is an opportunistic human pathogen (Alnor et al. 1994; Mahmood et al. 2000). However, the gox gene represents a very small proportion of the pathogen genome and is not, in itself, infectious or pathogenic. The bacterial GOX protein is highly specific for its substrate, glyphosate (OECD 1999a), hence it is unlikely to be involved in human pathogenesis.
64. The cp4 epsps gene is derived from another common soil bacteria, Agrobacterium sp. strain CP4 (Padgette et al. 1995), which is widespread in the environment and can be found on plant produce (especially raw vegetables). The CP4 EPSPS protein is functionally and structurally similar to EPSPS proteins naturally present in canola and in human food and animal feed derived from other plant and microbial sources (Nair et al. 2002).
65. CP4 EPSPS has been used extensively in GM plants as a selectable marker or a source of field resistance to glyphosate herbicides. Consequently, the Regulator has approved several GM plants expressing cp4 epsps for limited and controlled release (for example, see DIR 074/2007, pima cotton; DIR 082, perennial ryegrass and tall fescue; and DIR 101, cotton). The Regulator has also approved GM cotton lines expressing cp4 epsps for commercial release under licences DIR 012/2002, DIR 023/2003, DIR 059/2005 and DIR 066/2006. Both the GOXv247 and CP4 EPSPS proteins are present in GM canola approved for limited and controlled release under DIR 011/2001 and DIR 104, and for commercial release under DIR 020/2002. Therefore, the toxicity and allergenicity of GM plants expressing the GOXv247 and CP4 EPSPS proteins has been previously assessed by the Regulator and the assessments concluded that they are unlikely to be toxic or allergenic.
66. The amino acid sequences of both CP4 EPSPS (Mitsky 1993; Harrison et al. 1996) and GOX (Astwood 1995) were compared to the amino acid sequences of known protein toxins and allergens and no significant homology was found. Further bioinformatic studies using updated databases have confirmed that the GOXv247 and CP4 EPSPS proteins do not share any similarity with any known toxins or allergens (EFSA 2009d). The GOXv247 and CP4 EPSPS proteins are readily inactivated by heat and rapidly degraded by simulated mammalian digestive conditions (Harrison et al. 1996; OECD 1999a; Chang et al. 2003).
67. Acute oral toxicity studies using CP4 EPSPS and GOXv247 proteins produced by bacterial expression systems are described in the RARMP for DIR 020/2002. In summary, high doses of the CP4 EPSPS and GOXv247 proteins fed to mice had no adverse effects on food consumption, survival, body weight or gross pathology (Naylor 1994a; Harrison et al. 1996).
68. Feeding studies in animals using seed from GM canola lines containing the CP4 EPSPS and GOXv247 proteins are discussed in Section 5.4.5.
69. Food from Roundup Ready® canola has been approved for human consumption by FSANZ (ANZFA 2000). Food derived from GM soybean, cotton, sugarbeet, maize and lucerne lines that express the cp4 epsps gene have also been considered safe for human consumption by FSANZ (FSANZ 2005a; FSANZ 2005b; FSANZ 2006; FSANZ 2007).
70. A number of international regulatory bodies have also assessed Roundup Ready® canola GT73 with regard to toxicity and allergenicity. These include the United States Environmental Protection Agency (EPA 1996; EPA 1997a), the United Stated Food and Drug Administration (FDA 1995), Health Canada (Health Canada 1999a), the Canadian Food Inspection Agency (Canadian Food Inspection Agency 1995b) and the European Food Safety Authority (EFSA 2009d). These agencies have concluded that the presence of EPSPS and GOX proteins in food does not pose a significant toxicity or allergenicity risk. The EPA considers these proteins as inert ingredients (EPA 1996; EPA 1997a).
Toxicity of herbicide metabolites
71. The potential toxicity of herbicide metabolites is considered by the the Australian Pesticides and Veterinary Medicines Authority (APVMA) in its assessment of a new use pattern for particular herbicides, in this case glyphosate and glufosinate ammonium on InVigor® x Roundup Ready® canola. The issue is summarised below.
Glyphosate metabolites
72. There is no difference in the metabolic fate of glyphosate in non-GM canola and in GM canola expressing goxv247 and cp4 epsps. In the case of CP4 EPSPS, no new metabolic products are formed as the only difference from the native enzyme is the reduced affinity for glyphosate (OECD 1999a).
73. In glyphosate-sensitive plants very little of the glyphosate that is applied would be broken-down. The presence of the GOXv247 protein confers glyphosate tolerance by increasing the rate of breakdown of glyphosate to glyoxylate and aminomethylphosphonic acid (AMPA). Glyoxylate is a common metabolite in plants and forms part of the biochemical pathway that allows synthesis of carbohydrates from fat (the glyoxylate cycle).
74. AMPA is the most frequently detected metabolite of glyphosate in soil, water and plants (Reddy et al. 2008). Despite the faster breakdown of glyphosate, AMPA does not accumulate to higher levels in GM canola expressing GOX than in soybean that does not contain an introduced gox gene (Nandula et al. 2007; Duke 2010). AMPA is either non-selectively bound to natural plant constituents, conjugated with naturally occurring organic acids to give trace level secondary metabolites, or further degraded to one-carbon fragments that are incorporated into a variety of natural products and plant constituents (FAO & WHO 1998b).
75. Glyphosate and AMPA have similar toxicological profiles and both exhibit low toxicity (EPA 1997a; Williams et al. 2000; WHO 2005), although AMPA was shown to be genotoxic (able to change DNA) in a recent study using a very sensitive test (Manas et al. 2009). The APVMA sets maximum residue limits (MRLs) for agricultural and veterinary chemicals in agricultural produce, particularly produce entering the food chain. MRLs are set to reflect the legal use of a chemical and to ensure a safe food supply, and are set well below the level that would be harmful. The residue definition for glyphosate includes the metabolite AMPA (APVMA 2011).
Glufosinate ammonium metabolites
76. The herbicide glufosinate ammonium is comprised of a racemic (equal) mixture of the L- and D- enantiomers. The L- enantiomer is the active constituent and acts by inhibiting the enzyme glutamine synthetase. D-glufosinate ammonium does not exhibit herbicidal activity and is not metabolised by plants (Ruhland et al. 2002).
77. The PAT enzyme, encoded by either the bar or pat gene, inactivates the L-isomer of glufosinate ammonium by acetylating it to N-acetyl- L- glufosinate ammonium (NAG), which does not inhibit glutamine synthetase (Droge-Laser et al. 1994; OECD 2002). This metabolite is not found in non-GM plants.
78. The metabolism of glufosinate ammonium in tolerant GM plants and in non-GM (non-tolerant) plants has been reviewed (FAO & WHO 1998a; OECD 2002). In non-GM plants the metabolism of glufosinate ammonium is low to non-existent because of plant death due to the herbicidal activity. However, some metabolism does occur (Muller et al. 2001) and is different to that in GM plants expressing the PAT protein (Droge et al. 1992).
79. Two pathways for the metabolism of glufosinate ammonium in non-GM plants have been identified. The first step, common to both pathways, is the rapid deamination of L-phosphinothricin to the unstable intermediate 4-methylphosphonico-2-oxo-butanoic acid (PPO). PPO is then metabolised to either:
- 3-methyl-phosphinico-propionic acid (MPP, sometimes referred to as 3-hydroxy-methyl phosphinoyl-propionic acid) which may be further converted to 2-methyl-phosphinico-acetic acid (MPA); or
- 4-methylphosphonico-2-hydroxy-butanoic acid (MHB), which may be further converted to 4-methylphosphonico-butanoic acid (MPB), a final and stable product (Droge-Laser et al. 1994; Ruhland et al. 2002; Ruhland et al. 2004).
81. The metabolism of glufosinate ammonium has been investigated in GM herbicide-tolerant canola, maize, tomato, soybean and sugar beet (FAO & WHO 1998a; OECD 2002). The major residue present in the GM crops after glufosinate ammonium herbicide application was NAG, with lower concentrations of glufosinate ammonium and MPP. Studies using cell cultures of GM canola gave similar results, with NAG being the major metabolite (Ruhland et al. 2002).
82. Both NAG and MPP are less toxic than glufosinate ammonium, which itself has low toxicity (OECD 1999b; OECD 2002; EFSA 2005).
5.1.5 The antibiotic resistance marker gene (nptII) and its encoded protein
83. The GM canola lines Topas 19/2, MS1, RF1 and RF2 authorised under DIR 021/2002 contain the antibiotic resistance marker gene neomycin phosphotransferase type II (nptII).
84. The nptII gene, encoding the enzyme neomycin phosphotransferase, was derived from the common gut bacterium Escherichia coli and confers resistance to antibiotics such as kanamycin and neomycin on GM plant cells. The nptII gene was used during initial development of the GM plants in the laboratory to select plant cells containing the introduced genes.
85. The nptII gene is used extensively as a selectable marker in the production of GM plants (Miki & McHugh 2004). As discussed in previous DIR RARMPs, and in more detail in the RARMPs for DIR 070/2006 and DIR 074/2007, regulatory agencies in Australia and in other countries have assessed the use of the nptII gene in GM plants as not posing a risk to human or animal health or to the environment. A recent detailed evaluation of nptII in terms of human safety by the European Food Safety Authority concluded that the use of the nptII gene as a selectable marker in GM plants (and derived food or feed) does not pose a risk to human or animal health or to the environment (EFSA 2009a).
5.2 The regulatory sequences
86. Promoters are DNA sequences that are required in order to allow RNA polymerase to bind and initiate correct transcription. Also required for gene expression in plants is an mRNA termination region, including a polyadenylation signal. Information on the promoters, terminators and other regulatory genetic elements used to control expression of the introduced genes in the parental GM canola lines are listed in Table 2 (above) and described below.5.2.1 Regulatory sequences for the expression of the introduced bar gene
87. Expression of bar is controlled by the plant promoter PSsuAra from the Arabidopsis thaliana ats1A gene, which encodes a ribulose-1,5-bisphosphate carboxylase small subunit (rbcS) peptide (Krebbers et al. 1988). The PSsuAra promoter directs gene expression predominantly in green plant tissues (Krebbers et al. 1988; De Almeida et al. 1989).
88. The mRNA terminator for the bar gene is derived from the 3’ non-translated region of the T-DNA gene 7 (3’g7) of Agrobacterium tumefaciens (Dhaese et al. 1983).
Top of page
5.2.2 Regulatory sequences for the expression of the introduced pat gene
89. The pat gene is controlled by the constitutive 35S promoter and 35S terminator from Cauliflower mosaic virus (CaMV) (Odell et al. 1985) in both lines T45 and Topas 19/2. The CaMV 35S promoter has been used extensively in plant transformation studies (Sunilkumar et al. 2002; Squires et al. 2007). The 35S terminator has also been widely used in GM plants (Mitsuhara et al. 1996).
5.2.3 Regulatory sequences for the expression of the introduced barnase and barstar genes
90. The barnase and barstar genes are controlled by PTA29, a 1.5 kb promoter fragment derived from the tobacco (Nicotiana tabacum) TA29 gene (Goldberg 1988; Seurinck et al. 1990). TA29 is expressed specifically in the tapetal cells of tobacco anthers (Koltunow et al. 1990) and anther-specific expression was reproduced when the PTA29 promoter was used to drive transgene expression in tobacco and canola (Mariani et al. 1990; De Block & De Bouwer 1993).
91. As discussed in Section 5.1.1, expression of the barnase and barstar genes in GM InVigor canola lines occurs only in the tapetum cell layer of the pollen sac during anther development, resulting in production of cytotoxic RNase, and inactivation of the same RNase activity, respectively (Mariani et al. 1990; Mariani et al. 1992; De Block & De Bouwer 1993).
92. For both genes, the terminators are derived from the 3’ non-translated region of the nopaline synthase gene (3’ nos) from A. tumefaciens (Depicker et al. 1982). The nos terminator has been used in a wide variety of constructs for plant genetic modification (Reiting et al. 2007).
5.2.4 Regulatory sequences for the expression of the introduced cp4 epsps and goxv247 genes
93. Expression of cp4 epsps and goxv247 is driven by the Figwort mosaic virus (FMV) promoter P-CMoVb (Richins et al. 1987; Gowda et al. 1989; Sanger et al. 1990). P-CMoVb is a constitutive promoter which directs gene expression in all plant parts (Sanger et al. 1990; Maiti et al. 1997). The P-CMoVb promoter is thought to be equivalent to the 35S promoter from CaMV, despite low sequence conservation overall between these two promoters. This conclusion was reached because the two promoters occupy similar positions in their respective viral genomes, both increase in strength with increasing sequence length, and the core promoters have significant sequence homology (Sanger et al. 1990).
94. For both genes, the terminators are derived from the 3’ untranslated region of the E9 gene (E9 3’) from Pisum sativum, which encodes a ribulose-1,5-bisphosphate carboxylase/oxygenase small subunit (rbcS) peptide (Coruzzi et al. 1984; Morelli et al. 1985).
95. The cp4 epsps and goxv247 genes are each fused to a chloroplast transit peptide sequence to target the proteins to the chloroplasts (the site of aromatic amino acid biosynthesis). Transit peptides occur naturally in plants and function to direct proteins into specific organelles. In plants, EPSPS is synthesised as a pre-protein containing a transit peptide by free cytoplasmic ribosomes. The pre-protein is transported into the chloroplast stroma where the transit peptide is cleaved and rapidly degraded leaving the mature enzyme (Bartlett et al. 1982; della-Cioppa et al. 1986).
96. The cp4 epsps gene is fused to a chloroplast transit peptide from the A. thaliana epsps gene (AEPSPS/CTP2) (Klee et al. 1987). The goxv247 gene is fused to a chloroplast transit peptide from an A. thaliana gene encoding an rbcS peptide (SSU1A/CTP1) (Krebbers et al. 1988).
5.2.5 Regulatory sequences for the expression of the introduced nptII gene
97. Expression of the nptII gene in GM canola lines Topas 19/2, MS1, RF1 and RF2 is controlled by the nopaline synthase promoter (P-nos) from A. tumefaciens (Bevan et al. 1983). The terminator is derived from the 3’ non-translated region of the octopine synthase gene (3’ ocs) from A. tumefaciens (Dhaese et al. 1983).
5.3 Method of genetic modification
98. The GM canola proposed for release is the product of conventional breeding between InVigor® canola lines approved for commercial release under DIR 021/2002 and Roundup Ready® canola line GT73 approved for commercial release under DIR 020/2002.99. All of the parental GM canola lines were generated by Agrobacterium-mediated transformation using the plasmids described in Table 3 (della-Cioppa et al. 1987; De Block et al. 1989; FAO & WHO 1998a). A. tumefaciens is a soil bacterium that causes gall formation on a wide range of plant species. The gall is induced by transfer of hormone-producing genes from the bacterial cell into the plant genome. The genes are carried on an extrachromosomal, circular DNA molecule found within the bacterial cell called a Tumour-inducing (Ti) plasmid. During the infection process, only a section of the Ti plasmid known as the Transfer DNA (T-DNA) is transferred to the plant.
100. Molecular biologists have studied the infection and T-DNA transfer process of A. tumefaciens for many years and have used this natural process to facilitate genetic modification of plants. A. tumefaciens Ti plasmids have been produced that lack the genes responsible for tumour formation (disarmed plasmids) and instead enable genes of interest to be inserted between the T-DNA border sequences. When used to infect plants, A. tumefaciens cells carrying such plasmids cannot produce a tumour but will transfer the T-DNA sequence carrying the genes of interest into the plant cell where they stably integrate into the plant genome (Bevan 1984; Klee & Rogers 1989).
Table 3 List of plasmids used to generate the parental GM canola lines
| Plasmid name | GM canola line | Introduced Genetic Elements |
|---|---|---|
| pHOE4/Ac(II) | T45 | - 35S promoter from CaMV - pat gene from S. viridochromogenes - 35S terminator from CaMV |
| pOCA18/Ac | Topas 19/2 | - P-nos promoter from A. tumefaciens - nptII gene from E. coli - 3’-ocs terminator from A. tumefaciens - colE1 origin of replication from E. coli* - 35S promoter from CaMV - pat gene from S. viridochromogenes - 35S terminator from CaMV - cos site from bacteriophage lambda* |
| pTTM8RE | MS1 | - 3’-ocs terminator from A. tumefaciens - nptII gene from E. coli - P-nos promoter from A. tumefaciens - PTA29 promoter from N. tabacum - barnase gene from B. amyloliquefaciens - 3’-nos from A. tumefaciens - PSsuAra promoter from A. thaliana - bar gene from S. hygroscopicus - 3’ g7 terminator from A. tumefaciens |
| pTHW107 | MS8 | - PTA29 promoter from N. tabacum - barnase gene from B. amyloliquefaciens - 3’-nos from A. tumefaciens - PSsuAra promoter from A. thaliana - bar gene from S. hygroscopicus - 3’ g7 from A. tumefaciens |
| pTVE74RE | RF1 and RF2 | - 3’-ocs terminator from A. tumefaciens - nptII gene from E. coli - P-nos promoter from A. tumefaciens - PTA29 promoter from N. tabacum - barnase gene from B. amyloliquefaciens - 3’-nos from A. tumefaciens - PSsuAra promoter from A. thaliana - bar gene from S. hygroscopicus - 3’ g7 terminator from A. tumefaciens |
| pTHW118 | RF3 | - PTA29 promoter from N. tabacum - barstar gene from B. amyloliquefaciens - 3’-nos from A. tumefaciens - PSsuAra promoter from A. thaliana - bar gene from S. hygroscopicus - 3’ g7 from A. tumefaciens |
| PV-BNGT04 | Roundup Ready® canola GT73 | - P-CMoVb promoter from FMV - SSU1A/CTP1 sequence from A. thaliana - goxv247 gene from O. anthropi - 3’ E9 from P. sativum - P-CMoVb promoter from FMV - CTP2 sequence of the epsps gene from A. thaliana - cp4 epsps gene from Agrobacterium strain CP4 - 3’ E9 from P. sativum |
* The colE1 and cos sequences are of non-eukaryotic origin and will not function in the plant
5.4 Toxicity/allergenicity of the parental GM canola lines
101. The toxicity of the parental GM canola lines to people and to other organisms, including insects, birds, mice, rabbits, kangaroos and grazing livestock, was considered in the RARMPs for DIRs 020/2002 and 021/2002. The safety of feed produced from the parental GM canola lines for livestock was also considered. The Regulator concluded that the parental GM canola lines are as safe as non-GM canola. These assessments, plus new or updated information, are summarised below.Top of page
5.4.1 Toxicity/allergenicity to humans
102. Canola oil is the only fraction used in human food. Due to the extensive processing applied during canola oil extraction and refinement, no protein, including any novel proteins, would be expected to be detected in canola oil (ANZFA 2001b). Therefore, oil derived from the GM canola proposed for release would not contain any of the novel proteins.
103. Food derived from all of the parent lines used to generate the GM canola proposed for release has been approved for human consumption in Australia (ANZFA 2000; ANZFA 2001b) and other countries (see Section 5.1.3). These approvals also cover the GM InVigor® x Roundup Ready® canola proposed for release.
104. People could be exposed to pollen containing the introduced genes, either through occupational exposure or in honey. Canola is commonly utilised as a source of nectar and pollen for commercial honey production by honeybees. However, only low amounts of canola pollen are present in honey. The percentage dry weight of canola pollen per wet weight of honey that is produced from hives placed in canola fields is only 0.2 % (Hornitzky & Ghalayini 2006). If the honey is sieved or filtered the pollen content is further reduced (discussed in Malone 2002).
105. The introduced proteins are expressed only at low levels in plant tissues. No expression of the bar, barnase, barstar or nptII genes has been detected in pollen from the InVigor® canola parental lines (see Chapter 1, Section 6.2.2). Therefore, the level of exposure of people to the introduced proteins in pollen would be extremely low. Most importantly, none of the introduced proteins are toxic or allergenic, and the introduced genes were all isolated from common bacteria, that are widespread and prevalent in the environment (see Section 5.1.3).
5.4.2 Toxicity to animals, including livestock
106. Canola meal is produced as a by-product during the extraction of oil from canola seed. It is a significant component of livestock feed in Australia and a rich source of protein for livestock. Unprocessed canola seed can also be used directly as animal feed. In addition, canola can be used as a dual-purpose crop in Australia, whereby it is used for forage prior to seed production (Kirkegaard et al. 2008).
107. The production of canola meal involves a number of processes, including seed flaking, heating, mechanical crushing to remove oil, solvent extraction of oil, desolventising and toasting of the meal. Toasted canola meal is the most common fraction used as animal feed, although some meal (20%) is physically extracted without added heat. A small amount (5%) of canola meal available in Australia is from cold-pressed seed (Mailer 2004).
108. As discussed in Section 4, glucosinolates and erucic acid are naturally occurring toxicants in canola seed. Glucosinolates remain in the canola meal after oil extraction while erucic acid is removed with the oil fraction during processing of the seed. Industry standards require canola meal to contain less than 30 µmoles g-1 of glucosinolates. Compositional analyses demonstrate that the levels of erucic acid and glucosinolates in Roundup Ready® and InVigor® canola lines are below standard levels and do not vary significantly from their parental cultivars or other commercially available canola.
109. The introduced genes were all isolated from common soil bacteria that are widespread and prevalent in the environment. The barnase and barstar genes are not expressed in the seeds or leaves of InVigor® canola, therefore livestock would not be exposed to these proteins. The nptII, pat, cp4 epsps and goxv247 genes are only expressed at low levels in GM canola seed and/or leaves. The amount of each protein is further reduced during processing of the seed that results in the production of meal (ANZFA 2000; ANZFA 2001b).
110. The PAT, CP4 EPSPS and GOXv247 proteins are not toxic, even at high doses, as demonstrated by acute oral toxicity studies in animals (see Section 5.1.3). While the assessment of the toxicity of the herbicide metabolites to non-target organisms is the responsibility of the APVMA, the major metabolites of glufosinate ammonium and glyphosate are also not toxic (see Section 5.1.4). The composition of the parental GM canola lines does not differ significantly from non-GM canola (see Section 6.2.4) other than by the presence of the introduced proteins, and feeding studies on a range of organisms demonstrate that there are no anti-nutritional effects of the genetic modifications in the parental GM canola lines (see Section 5.4.5).
111. The parental GM canola lines have been assessed and approved for use in animal feed by regulatory agencies in Europe, Canada and the USA (FDA 1995; Canadian Food Inspection Agency 1995b; Canadian Food Inspection Agency 1996; FDA 1996; FDA 1997; FDA 1998; European Scientific Committee on Plants 1998b; EFSA 2004). Roundup Ready® canola, glufosinate ammonium tolerant canola and/or InVigor® hybrid lines have been approved for use in animal feed since 1995 and there have been no reports of adverse effects to livestock fed these GM canola lines.
5.4.3 Toxicity to honey bees
112. As honey bees are a major pollinator of canola, the potential effects of the genetic modifications in the parental GM canola lines on honey bees were considered in detail in the RARMPs for DIR 020/2002 and 021/2002. Studies cited in these documents did not find any negative impacts on bees foraging on Roundup Ready® canola, InVigor® canola, or other GM glufosinate ammonium tolerant canola plants (USDA-APHIS 1999a; Canadian Food Inspection Agency 1995a; Canadian Food Inspection Agency 1995c; USDA-APHIS 1998; European Scientific Committee on Plants 1998a; European Scientific Committee on Plants 1998b; Malone & Pham-Delegue 2001; Malone 2002; Pham-Delegue et al. 2002).
113. Two more recent studies have shown reduced abundance of bees in GM herbicide tolerant canola compared to non-GM canola (Haughton et al. 2003; Morandin & Winston 2005). In both studies, the authors propose that the differences were an indirect result of herbicide treatments that effectively reduced weed numbers and diversity in the GM fields, consequently reducing forage for bees.
114. A number of regulatory agencies have assessed whether the parental GM canola lines have any increased toxicity to non-target organisms as a result of the genetic modifications. In its assessments of Roundup Ready® canola and GM canola lines MS8 and RF3, the USDA-APHIS determined that the GM canola lines would not harm threatened or endangered species or other organisms, such as bees, that are beneficial to agriculture (USDA-APHIS 1999a; USDA-APHIS 1999b; USDA-APHIS 1999c). The Canadian Food Inspection Agency (CFIA) concluded that the unconfined release of Roundup Ready® canola and GM canola lines MS8 and RF3 would not result in altered impacts on interacting organisms, and that their potential impact on biodiversity is equivalent to that of currently commercialised canola varieties (Canadian Food Inspection Agency 1995b; Canadian Food Inspection Agency 1996).
5.4.4 Toxicity to soil microbes
115. Several studies have investigated the effects of growing GM glyphosate tolerant canola or GM glufosinate ammonium tolerant canola on soil microbes. These studies were described in detail in the RARMPs prepared for DIR 020/2002 and 021/2002. Slightly altered microbial communities in the rhizosphere of GM canola plants have been reported. These differences were minor and generally not sustained after removal of the GM plants (Dunfield & Germida 2001; Gyamfi et al. 2002; Dunfield & Germida 2003).
116. Recent studies have confirmed the lack of permanent effects on soil biota by GM glyphosate tolerant crops. For example, no permanent effects on soil biota were observed in a series of experiments designed to estimate the effect of glyphosate tolerant soybean and maize, and their management, on the abundance of detritivorous soil biota and crop litter decomposition (Powell et al. 2009). While significant effects were observed in a few of the measured groups, in most cases the effects were only observed in the first year of the study and were not consistent across sample dates or across the four study years. The most frequent effect of the glyphosate tolerant herbicide system was a transient shift toward more fungal biomass relative to bacterial. The genetic modification in the soybean and maize had little effect on litter decomposition, however the use of glyphosate did reduce decomposition of surface (but not buried) litter.
117. In a field experiment conducted at six sites in Canada, repeated plantings of glyphosate tolerant wheat and glyphosate tolerant canola grown in rotation had only minor and inconsistent effects on soil microorganisms over a wide range of growing conditions and crop management regimes (Lupwayi et al. 2007). As is the case for many studies that show an effect of herbicide resistant cropping systems on microbial communities, the effects of the glyphosate tolerance trait and the herbicide applications were not separated in this study. Application of herbicides can affect proportions of soil microbes (for example, see Becker et al. 2001; Gyamfi et al. 2002; Kremer & Means 2009; Mijangos et al. 2009).
118. Crop type (GM or non-GM) made no difference to the abundance or structure of microbial communities in a study designed to separate the effects of GM glyphosate tolerant maize from the use of glyphosate on denitrifying bacteria and fungi (Hart et al. 2009). The GM maize in this study expressed the cp4 epsps gene, and the authors note that the use of a protein derived from a common soil bacterium may affect soil microbial communities less than modifications that introduce novel proteins into the soil. The genes for herbicide tolerance and a hybrid breeding system in this DIR 108 application were all isolated from common soil bacteria.
5.4.5 Feeding Studies
119. Several feeding studies have been undertaken with the parent lines used to generate the GM canola proposed for release. Data from these studies were submitted in conjunction with the applications for licences DIR 020/2002 and 021/2002, and fully assessed in the RARMPs for these licences. A brief summary of these studies, along with new or updated information, is provided below.
InVigor® canola
120. Two feeding studies were conducted in rabbits to investigate the nutritive value of canola seed of hybrids derived from crosses of MS1 x RF1 (ANZFA 2001b) and MS8 x RF3 (Maertens et al. 1996). No significant differences in feed intake, feed efficiency, weight gain or final weight of the rabbits were observed between the GM canola diet and the non-GM canola diet, indicating that the nutritional value of the GM hybrid canola was comparable to the non-GM parental line (ANZFA 2001b).
121. Similarly, in a study of canaries fed seed from either MS1 x RF1 hybrids or non-GM canola, no differences in food consumption, behaviour or body weight were observed between the GM and non-GM diets (Canadian Food Inspection Agency 1995c).
122. One feeding study involving broiler chickens fed seed from GM canola line Topas 19/2 was described in the RARMP for DIR 021/2002. There were no differences between the chickens fed Topas 19/2 canola seed and those fed non-GM canola seed for any of the measured parameters, including body weight, body weight gain, feed intake, mortality rate and carcass characteristics at post-mortem (Leeson 1999).
123. Subsequently, another 42-day feeding study in broiler chickens has been reported (EFSA 2009b). This study was carried out on 420 male broiler chickens, which were divided into three groups and fed diets containing 10% GM canola hybrid MS8 x RF3 that had been either treated with glufosinate ammonium or untreated, or a diet containing 10% non-GM canola. No significant differences were observed in any of the parameters measured (animal health, survival, feed intake, weight gain, feed conversion and carcass and muscle weight), showing that MS8 x RF3 GM hybrid canola is nutritionally equivalent to non-GM canola (EFSA 2009b).
Roundup Ready® canola
124. Broiler chickens were used to compare diets containing Roundup Ready® canola GT73, the parental non-GM canola line, and six commercially available canola lines (Taylor et al. 2004; Stanisiewski et al. 2002). Values obtained for a range of parameters were similar across the diets demonstrating that Roundup Ready® canola GT73 is as nutritious as non-GM canola.
125. Similarly, feeding studies in bobwhite quail chicks (Campbell et al. 1993; Campbell & Beavers 1994), trout (Brown et al. 2003), lambs (Stanford et al. 2002; Stanford et al. 2003) and pigs (Aalhus et al. 2003; Caine et al. 2007) found no significant differences between animals fed Roundup Ready® canola GT73 containing diets and control diets, supporting the conclusion that Roundup Ready® canola meal is nutritionally equivalent to non-GM canola meal (EFSA 2009d).
126. Three one-month feeding studies were conducted on rats (Naylor 1994b; Nickson & Hammond 2002). No changes attributable to the genetic modification were observed. FSANZ thoroughly considered these studies in its assessment of Roundup Ready® canola GT73 before reaching the conclusion that ‘oil derived from glyphosate-tolerant canola GT73 is as safe for human consumption as oil from other commercial canola varieties’ (ANZFA 2000).
5.5 Weediness of the parental GM canola lines
127. The risk of the genetic modifications in the parental GM canola lines making them more invasive or persistent than non-GM canola in Australia was assessed in the RARMPs for licences DIR 020/2002 and 021/2002. The Regulator concluded that the parental GM canola lines are no more invasive or persistent than non-GM canola. A brief summary of this assessment, along with new or updated information, is provided below.5.5.1 Spread and persistence in the environment
128. Although conventional canola has a number of weedy characteristics, it is a poor competitor and is not invasive. Canola is not a significant weed in habitats outside agricultural areas and does not pose a serious threat to the environment and biodiversity. The risk that the Roundup Ready® or InVigor® canola will be more likely to spread and persist in the environment and cause more harm to the environment than non-GM canola is negligible.
129. There is no evidence to show that the introduced genes increase the potential weediness of the plants. The germination, seed dormancy and fitness traits such as sensitivity to other herbicides, disease resistance, stress adaptation and competitiveness for Roundup Ready® or InVigor® canola fall within the range of non-GM open-pollinated and hybrid canola varieties.
130. The hybrid vigour displayed in InVigor® canola hybrids is not a function of the genetic modification that can be transferred as a single trait, but is a result of breeding two genetically distinct parents. In general, hybrid vigour manifested in the F1 generation declines in subsequent generations (Falconer & Mackay 1996).
131. InVigor canola hybrids have displayed yield increases of 10-20% over non-GM open pollinated varieties in Australia and greater than 20% in Canada (Clayton et al. 1999; Zand & Beckie 2002; Bayer CropScience 2003; Harker et al. 2003). However, the superior seedling emergence and increased seed numbers (Clayton et al. 1999; Bayer CropScience 2003; Harker et al. 2003) does not lead to the expected increase in volunteers in commercial fields in Canada (Beckie & Owen 2007) or in trials in the UK, due to greater uniformity in ripening (Crawley et al. 1993; Sweet 1999; MacDonald & Kuntz 2000). Volunteers of herbicide resistant hybrids are no more invasive of agricultural or disturbed habitats than volunteers of herbicide resistant open pollinated canola (Beckie & Owen 2007; Warwick et al. 2009). Data obtained in Australia indicate that the vigour exhibited by InVigor canola hybrids falls within the range of vigour exhibited by non-GM hybrid and open pollinated varieties of canola grown commercially (see DIR 021/2002).
132. GM herbicide tolerant canola has no altered weedy or invasiveness potential (Hall et al. 2005; Warwick et al. 2009). The genetic modifications do not provide Roundup Ready® or InVigor® canola with an ecological advantage over conventional canola except in the presence of glyphosate or glufosinate ammonium, respectively. Glyphosate is widely used for weed control in broad-acre agriculture, horticulture and other weed management situations. Glufosinate ammonium is not registered for use in any broad-acre crop except on Bayer’s GM InVigor® canola and GM Liberty Link® cotton varieties. It is used in viticulture and horticulture but is rarely used in non-agricultural areas.
133. Roundup Ready® and InVigor® canola are only tolerant to glyphosate or glufosinate ammonium, respectively, and their susceptibility to other herbicides is no different to non-GM canola. GM canola volunteers can be managed and controlled using alternative herbicides assessed and approved by the APVMA as well as other non-chemical management practices in the same manner as non-GM canola volunteers. The impact of such changes is considered to be primarily an agricultural production issue with a potential economic impact.
Top of page
Agricultural environments
134. The risk that Roundup Ready® or InVigor® canola will be more invasive or persistent in the agricultural environment than non-GM canola and result in a more detrimental environmental impact was assessed as negligible.
135. Non-GM canola is primarily dispersed by human activities (harvest, transport) (Agrisearch 2001; Crawley & Brown 2004; von der Lippe & Kowarik 2007) and this would be the case with Roundup Ready® or InVigor® canola.
136. Canola seed can be dispersed by grazing animals (eg sheep, Stanton et al. 2003) or wild birds (Twigg et al. 2008; Woodgate et al. 2011). Wind may move plant material from windrows. This material is generally caught in the next windrow or trapped by the remaining stubble, but can on occasions be moved over greater distances and cross boundary fences. There is no evidence that dispersal of seed would be different for GM canola.
137. Volunteer canola (non-GM and GM) represents a weed of agricultural production systems (Legere et al. 2001; Beckie et al. 2001; Martens 2001; Simard & Legere 2001; Simard et al. 2002). There are no differences between Roundup Ready® or InVigor® canola and non-GM canola with respect to the intrinsic characteristics contributing to spread and persistence, such as seed production, shattering or dormancy, and competitiveness. Roundup Ready and InVigor® canola varieties have been grown commercially in Canada since the mid-1990’s and there is no indication that they are more intrinsically persistent than non-GM canola (Derksen et al. 1999; Norris et al. 1999; MacDonald & Kuntz 2000; Crawley et al. 2001).
138. Non-GM canola can display secondary dormancy and can persist for several years as an agricultural weed, particularly as volunteers following canola crops resulting from harvest losses (Lutman 1993; Pekrun et al. 1998; Gruber et al. 2005; Harker et al. 2006; Gruber et al. 2008; Gruber et al. 2010). This appears to apply equally to glyphosate or glufosinate ammonium tolerant canola (Fredshavn & Poulsen 1996; Norris et al. 1999; Simard et al. 2002; Salisbury 2002c; Beckie & Owen 2007). Gulden et al (2000) found no significant differences between dormancy of Roundup Ready® canola or other herbicide tolerant canola, including InVigor® cultivars, and non-GM canola, but did find significant differences between varieties, indicating that the parental genotype is an important factor in the degree of dormancy (Gulden et al. 2000).
139. Roundup Ready® and InVigor® canola only have a survival advantage in the presence of glyphosate or glufosinate ammonium, respectively. Studies of glufosinate ammonium tolerant canola lines and non-GM cultivars grown in monoculture or in a mixture with barley showed no differences in competitive ability (Poulsen et al. 1999). Another study showed that glufosinate ammonium tolerant oilseed-rape showed significantly lower seedling establishment when compared with non-GM canola lines in six out of twelve cases and significantly higher in two cases (Crawley et al. 2001).
140. Glyphosate is commonly used in broad-acre cropping for pre-emergent weed control prior to planting. Glyphosate would not be effective in controlling canola volunteers in situations where Roundup Ready® canola had been grown previously. The presence of Roundup Ready® canola volunteers in agricultural or disturbed habitats has implications for the choice of herbicide(s) in situations where glyphosate is the principal weed control strategy.
141. Roundup Ready® and InVigor® canola are as susceptible to all other herbicides except glyphosate or glufosinate ammonium, respectively, as non-GM canola. The GM canola volunteers can be controlled by using the variety of other herbicides assessed and approved by the APVMA as well as non-chemical management methods currently used to control non-GM canola.
Non-cropped disturbed habitats
142. Canola is found in low densities in non-cropped disturbed situations, such as grassy road verges (MacDonald & Kuntz 2000; Norton 2003). The available evidence supports the conclusion that the GM canola lines approved for release under DIRs 020/2002 and 021/2002 pose no greater weed threat than non-GM canola in non-cropped disturbed habitats.
143. Due to its primary colonising nature, canola can take advantage of disturbed land (Salisbury 2002c). Canola plants are often observed growing near transport routes and at field margins (Agrisearch 2001; Crawley & Brown 2004; von der Lippe & Kowarik 2007; Nishizawa et al. 2009). In Australia and Canada, roadside canola populations are thought to be reliant on re-supply of seed from seed spillage during harvest and transport operations rather than forming self-sustaining weed populations (Salisbury 2002c; Gulden et al. 2008). However, canola is a poor competitor and will be displaced unless the habitats are disturbed on a regular basis (OECD 1997; Beckie et al. 2001; Salisbury 2002c). Herbicide tolerant crops in general are not considered noxious weeds and have not been more invasive in disturbed areas (Beckie et al. 2006; Beckie & Owen 2007; Warwick et al. 2009).
144. The Conservation Council of Western Australia recently published a survey of roadside canola plants conducted by the Conservation Council (WA) Citizen Science Program, Esperance Local Environmental Action Forum (LEAF) and GM Cropwatch (accessed 9 November 2011). The survey was conducted in September 2011 to determine the frequency and distribution of GM Roundup Ready® canola plants in the Esperance region of WA after one year of commercial production. Two GM positive plants were detected among 190 canola plants collected and tested, representing 1.05%7. The area sown to GM canola was around 8% of the total canola crop in WA in 2010 (DAFWA 2010).
145. Roundup Ready® and InVigor® canola volunteers occurring in disturbed environments will not have any competitive advantage over conventional canola in the absence of glyphosate or glufosinate ammonium selection (Wilkinson et al. 1995; Senior & Dale 2002; Warwick et al. 2009).
146. Glufosinate ammonium is registered for use in commercial and industrial areas, rights-of-way and other non-agricultural areas under the trade names Basta and Finale, but it is not widely used for weed control by local councils and Road and Rail authorities (Dignam 2001).
147. Glyphosate is widely used in weed control operations in disturbed environments such as roadsides. However, while glyphosate is very effective in controlling grasses, it does not always achieve complete control of established broadleaf weeds. A mixture of herbicides (commonly referred to as ‘spiking’) may be used to ensure complete control of broadleaf weeds. Management of Roundup Ready® canola in roadsides and other disturbed habitats can be achieved by the variety of management strategies available, including a range of alternative herbicides to glyphosate, tank mixing of other herbicides with glyphosate, and non-chemical management methods such as mowing, cultivation, burning and grazing.
Undisturbed environments
148. Canola is considered a weed of agricultural and disturbed habitats, but is of minor significance to natural ecosystems (Groves et al. 2003). The genotypes used for commercial canola cultivation are bred for maximum production in managed environments in which optimal water and nutrient availability is ensured. In natural environments where water and nutrient availability are limited, canola is considered a poor competitor compared with native species (Hall et al. 2005; Oram et al. 2005; Canadian Food Inspection Agency 2007). When roadsides were surveyed for the presence of GM canola, it was only found in the 5 m closest to the edge of the road, but not further away from the road (Crawley & Brown 2004).
149. The available evidence supports the conclusion that the GM canola lines approved for release under DIRs 020/2002 and 021/2002 pose no greater weed threat resulting in adverse impacts on the environment than non-GM canola in undisturbed natural habitats. GM herbicide tolerant crops in general are not considered noxious weeds and have not been more invasive in natural ecosystems (Beckie et al. 2006; Beckie & Owen 2007; Warwick et al. 2009).
150. Roundup Ready® and InVigor® canola do not have any competitive advantage in the absence of glyphosate or glufosinate ammonium, respectively. Even if glufosinate ammonium and/or glyphosate tolerant canola did establish in undisturbed natural habitats, they would be unlikely to persist because of their poor competitiveness.
151. Where herbicides are used to control weeds in undisturbed environments glyphosate is frequently used, but removal is normally by spot spraying, not broadcast spraying, and if Roundup Ready® canola did occur in these environments it could be effectively controlled using other herbicides and non-chemical management techniques.
5.5.2 Weed risk assessment of parental GM canola lines
152. A weed risk assessment of non-GM canola based on the National Post-Border Weed Risk Management Protocol is included in the reference document “The Biology of Brassica napus L. (canola) (see Appendix 1, OGTR 2011) and summarised in Section 4.2.2 above. The genetic modifications in the GM parental lines do not alter the ratings for invasiveness or impact in any of the land uses where canola primarily occurs, namely, dryland and irrigated agricultural areas, and highly disturbed areas such as roadsides. The property of herbicide tolerance (either to glufosinate ammonium for InVigor® canola or to glyphosate for Roundup Ready® canola) could affect the plant’s tolerance to average weed management practices. However, as discussed above, all of the parental GM canola lines remain susceptible to alternative herbicides, as well as standard agronomic and mechanical management practices.
153. These conclusions are consistent with the RARMPs prepared for DIR 020/2002 and DIR 021/2002, which assessed the risk of increased weediness from commercial release of these GMOs as negligible when compared to non-GM canola.
Top of page
5.5.3 Herbicide resistance
154. There is some potential for development of herbicide-resistant weeds if the parental GM canola lines and their corresponding herbicides are used inappropriately. The repetitious use of a single herbicide, or herbicide group1, increases the likelihood of selecting weeds that have developed herbicide resistance through natural mechanisms (Gressel 2002). Integrated weed management practices help to avoid selection of resistant weed biotypes (CropLife Australia 2011).
155. Herbicide resistance comes under the regulatory oversight of the APVMA. The APVMA has primary regulatory responsibility for agricultural chemicals in Australia. The APVMA operates the national system that evaluates, registers and regulates agricultural and veterinary chemical products. Any changes to a product that is already on the market must also be referred to the APVMA.
156. The development of resistance to glufosinate ammonium and glyphosate herbicides would have implications for the choice of herbicide(s) available for weed control operations in agriculture and elsewhere. The APVMA assesses all herbicides used in Australia and sets their conditions of use.
157. Glyphosate has historically been considered a low risk herbicide for the development of herbicide resistance because its mode of action imposes genetic and biochemical constraints associated with potential mechanisms of resistance (Jasieniuk 1995; Bradshaw et al. 1997) and the frequency of mutations that impart glyphosate tolerance in plants is lower than for other herbicides (Weersink et al. 2005). However, the recent intensive use of glyphosate across large areas has resulted in several reports of glyphosate-tolerant weed species (Powles et al. 1998; Pratley et al. 1999; Neve et al. 2004; Powles & Preston 2006; Yu et al. 2006; Green et al. 2008).
158. Among others, these weeds include: Lolium rigidum (rigid ryegrass) in Australia; Conyza bonariensis (hairy fleabane) in South Africa and North America; Eluesine indica (goosegrass) in Malaysia; Lolium multiflorum (Italian ryegrass) in Chile; Plantago lanceolata (Buckhorn plantain) in South Africa; and Cyperus esculentus (yellow nutsedge), Commelina benghalensis (tropical spiderwort), Ipomoea spp. (morning glory) and Acalypha (wild buckwheat) in North America (Powles & Preston 2006; Green et al. 2008; Heap 2011).
159. A review in 2008 found no reports of glufosinate ammonium tolerant weeds (Green et al. 2008). Since then, there has only been one report of a glufosinate ammonium tolerant weed (E. indica in Malaysia in 2009, Heap 2011).
160. Stacking of multiple herbicide tolerant traits, such as in the InVigor® x Roundup Ready® canola proposed for release, increases the number of herbicide mixture options with multiple modes of action (Green et al. 2008). This could reduce the selective pressure on weed populations that occurs when a single herbicide is used exclusively.
161. Crop Management Plans have been developed separately by Bayer CropScience and Monsanto for InVigor® and Roundup Ready® canola, respectively. These CMPs are required to be followed by canola growers when growing either InVigor® canola, Roundup Ready® canola or InVigor® x Roundup Ready® canola. The CMPs address issues such as minimising and managing canola volunteers in crops following GM herbicide tolerant canola in a rotation, and minimising the development of herbicide tolerant weeds.
5.6 Potential for gene transfer from the parental GM canola lines
162. The potential for gene transfer from the parental GM canola lines to other sexually compatible plants (including other herbicide tolerant canola crops) was assessed in the RARMPs for licences DIR 020/2002 and 021/2002. A brief summary of this assessment, along with any new or updated information, is provided below.163. Any transfer of the barnase gene to other sexually compatible plants will not have any negative environmental impacts because it will only result in male sterility and not confer any selective advantage in terms of weediness or persistence. The fertility restorer gene (barstar) would have no impact on a plant’s phenotype apart from restoring male fertility for a portion of the progeny of a cross with a plant containing the male sterile gene. Therefore, only the potential for transfer of the herbicide tolerance traits is discussed below.
5.6.1 Gene transfer to other canola crops
164. Canola is predominantly self-pollinating with average inter-plant outcrossing rates of 30%. Outcrossing frequencies are highest in the first 10 m of the recipient fields, and rates decline with distance (Husken & Dietz-Pfeilstetter 2007). In a commercial situation, where different canola crops may be grown in adjacent fields, outcrossing is likely to occur beyond 10 m of the field borders. Cross pollination between canola lines is inevitable given sufficient proximity and exposure. There was no indication that the genetic modifications of the parental GM canola lines would increase the rate of outcrossing.
165. If Roundup Ready® or InVigor® canola is grown in close proximity to other canola crops there is a high likelihood of some outcrossing resulting in herbicide tolerant volunteers in adjacent fields where GM herbicide tolerant canola has not been grown. However, the overall frequency of hybridisation will be low and the number of resultant herbicide tolerant volunteers would be reduced by the vast majority of hybrid seeds being harvested along with the crop. Such volunteers would pose the same negligible risk to human health and safety and the environment as the parental GM canola, as assessed in DIR 020/2002 and DIR021/2002.
166. The possibility of gene transfer from Roundup Ready® or InVigor® canola crops would make the management of canola volunteers more complex and have implications for the choice of herbicide(s) selected for control operations, not only for growers of GM herbicide tolerant canola, but also for growers of other canola varieties. However, as discussed previously, volunteers can be readily controlled by alternative herbicide and non-chemical management practices currently used to control canola volunteers.
Gene transfer to herbicide tolerant canola
167. The ‘stacking’ of multiple herbicide tolerance traits through outcrossing between the two GM herbicide tolerant canolas and non-GM herbicide tolerant canola varieties could also occur at a low frequency, and would have implications for herbicide choices for the control of canola volunteers. In 2005–2006, approximately 75% of the canola crop in Australia comprised non-GM imidazolinone tolerant (Clearfield®) and triazine tolerant (TT) varieties (Norton & Roush 2007).
168. Note that because the triazine tolerance trait in TT canola is maternally inherited, and so cannot be spread by pollen movement, stacking of the glyphosate or glufosinate ammonium tolerance traits will only occur in the direction of Roundup Ready® or InVigor® canola to TT canola, and not vice versa.
169. Hybridisation between the existing non-GM herbicide-tolerant canola varieties, InVigor® canola and Roundup Ready® canola could result in accumulation or ‘stacking’ of genes for tolerance to up to four different herbicide groups within the same plant. However, development of canola plants with all four herbicide tolerance traits would only be expected to occur at an extremely low frequency because it would require at least three separate hybridisation events (two crosses between different pairs of herbicide tolerant canolas and a cross between the progeny of these).
170. Attention to volunteer management, proper crop rotation and herbicide management practices should limit the frequency of productive hybridisation between different herbicide tolerant canola varieties and hence the development of multiple herbicide tolerant canola in Australia (Rieger et al. 2001; Downey 1999; Salisbury 2002c). If multiple-herbicide tolerant canola plants were to occur, they are unlikely to be more invasive or persistent than non-herbicide tolerant canola plants and could be controlled by other herbicides or other agricultural practices.
5.6.2 Gene transfer to other sexually compatible species
171. Canola can cross with other B. napus groups or subspecies (including vegetable forms), B. oleracea, B. juncea and B. rapa under natural conditions. Naturally occurring hybrids between B. napus and R. raphanistrum, H. incana and S. arvensis have also been reported at very low frequencies (Salisbury 2002b; Warwick et al. 2009). All of these species are naturalized in Australia and weedy forms are known to be present (Groves et al. 2003). B. juncea, H. incana, R. raphanistrum and S. arvensis are problematic weeds in commercial canola growing regions of Australia. Therefore, it is likely that some or all of these sexually compatible species may be found growing at or near sites where the parental GM canola lines are grown. Hybridisation requires synchronicity of flowering between the parental GM canola lines and sexually compatible species to enable cross-pollination and gene flow to occur.
172. The RARMPs prepared for DIR 020/2002 and 021/2002 assessed the risks associated with gene flow from the parental GM canola lines to B. rapa, H. incana, R. raphanistrum and S. arvensis as very low, while the risks associated with gene flow to B. napus vegetables and forage rape, B. oleracea or B. juncea were assessed as negligible.
173. B. napus vegetables or forage are generally harvested or used for forage before flowering. B. napus vegetable seed production crops are isolated from other B. napus vegetable or canola crops to prevent outcrossing. Of the other sexually compatible Brassica species, hybridization offurs most readily between canola and B. rapa. Hybrids are often observed when the two species are grown in close proximity (Simard et al. 2006) and the transfer of traits from commercially grown canola to wild populations of B. rapa has been observed in Canada (Warwick et al. 2003). Warwick et al. (2008) showed that a herbicide tolerance trait from a commercial canola crop was transferred to, and stably maintained in, a wild B. rapa population for at least six years. The trait persisted despite the fact that the corresponding herbicide had not been applied during this period and, hence, no selective pressure had been applied.
174. The research of Warwick et al. (2008) illustrates that, if plants are growing in close proximity with synchronous or overlapping flowering periods, gene flow to sexually compatible species can occur. However, all interspecific hybrids have reduced fertility and low seed set due to the genetic barriers that exist (Jorgensen & Andersen 1994; Jorgensen et al. 1998; Salisbury 2002a; Warwick et al. 2003; Salisbury 2006). With the exception of the relatively productive interspecific hybridisation that occurs between Brassica species that contain the A genome (B. napus, B. juncea and B. rapa), most other interspecific hybridisation events occur at very low frequency.
175. Gene transfer from the parental GM canola lines to brassicaceous weeds would have implications for the choice of herbicide(s) for control of brassicaceous weeds. Glyphosate or glufosinate ammonium tolerant hybrids can be effectively controlled using a range of alternative herbicides and other non-chemical management techniques currently used for the control of Brassicaceous weeds. In addition, glyphosate or glufosinate ammonium would not be used for weed control in or adjacent to paddocks where Roundup Ready® or InVigor® canola has been grown because it would be ineffective in controlling the GM herbicide tolerant canola volunteers. Measures taken to control GM herbicide tolerant canola volunteers would also eliminate any herbicide tolerant hybrids.
Top of page
1Herbicides are classified into groups based on their mode of action. All herbicide product labels must display the mode of action group. This enables users to rotate among herbicides with different modes of action to delay the development of herbicide tolerance in weeds.