62. The following factors are taken into account when postulating relevant risk scenarios:

  • the proposed dealings, which may be to conduct experiments, develop, produce, breed, propagate, grow, import, transport or dispose of the GMOs, use the GMOs in the course of manufacture of a thing that is not the GMO, and the possession, supply and use of the GMOs in the course of any of these dealings
  • the proposed limits
  • the proposed controls
  • characteristics of the parent organism(s)
  • routes of exposure to the GMOs, the introduced gene(s) and gene product(s)
  • potential effects of the introduced gene(s) and gene product(s) expressed in the GMOs
  • potential exposure to the introduced gene(s) and gene product(s) from other sources in the environment
  • the environment at the site(s) of release
  • agronomic management practices for the GMOs.
63. Three risk scenarios were postulated and evaluated. These scenarios are summarised in Table 2 and more detail of the evaluation of these scenarios is provided later in this Section. In the context of the control measures proposed by the applicant and considering both the short and long term, none of the risk scenarios were identified as giving rise to a risk that could be greater than negligible. Therefore, they did not warrant further detailed assessment.

64. All of the GM safflower lines contain the selectable marker gene hph, and some of the GM safflower lines contain the visual reporter gene gfp. These genes and their products have already been considered in detail in previous RARMPs (for example, DIR 077/2007 for hpt and DIR 096 for gfp) and by other regulators. Further information about these genes can be found in the document Marker genes in GM plants available from the Risk Assessment References page on the OGTR website. Since neither of these genes have been found to pose risks to either people or the environment, their potential effects will not be further assessed for this application.

65. All of the introduced regulatory sequences are derived from common plants, bacteria and viruses. Similar regulatory elements are naturally present in safflowers, and the introduced elements are expected to operate in similar ways to endogenous ones. Therefore, although the transfer of introduced regulatory sequences to other sexually compatible plants could result in unpredictable effects, the impact is not likely to be greater than that arising from transfer of endogenous regulatory elements. Hence, these potential effects will not be further assessed for this application.

66. The potential for horizontal gene transfer (HGT) and any possible adverse outcomes has been reviewed in literature (Keese 2008) as well as assessed in many previous RARMPs. HGT was most recently considered in the RARMP for DIR 108. This and other RARMPs are available on the OGTR website or by contacting the OGTR. No risk greater than negligible was identified due to the rarity of these events and because the gene sequences are already present in the environment and available for transfer via demonstrated natural mechanisms. Therefore, HGT will not be assessed further.

67. The potential for unauthorised activities to lead to an adverse outcome has been considered in previous RARMPs. The Act provides for substantial penalties for non-compliance and unauthorised dealings with GMOs. The Act also requires the Regulator to have regard to the suitability of the applicant to hold a licence prior to the issuing of a licence. These legislative provisions are considered sufficient to minimise risks from unauthorised activities, and no risk greater than negligible was identified in previous RARMPs. Therefore unauthorised activities will not be considered further.

Table 1. Summary of risk scenarios from dealings with GM safflower genetically modified for increased levels of oleic acid

Risk category
Risk scenario
Identified risk?Reason
Pathway that may give rise to harmPotential harm
Section 2.1
Production of a substance toxic or allergenic to people or toxic to other organisms
Exposure to GM plant material containing the introduced silencing constructs or their end productsAllergic reactions in people or toxicity in people and other organisms
No
  • The mechanism of gene silencing, via siRNA, does not lead to expression of any introduced protein.
  • Oleic acid is not toxic or allergenic.
  • Plant material from the GMOs would not be used for human food or animal feed.
  • The limited scale, short duration and other proposed limits and controls minimise exposure of people and other organisms to the GM plant material.
Section 2.2
Weediness of GM safflower plants in the environment
The genetic modifications increase the weediness of the GMOsEnvironmental harms associated with weediness; allergic reactions in people or toxicity in people and other organisms
No
  • The genetic modifications of the GM safflower lines are not expected to alter any characteristics associated with weediness.
  • The limits and controls proposed for the release would minimise spread and persistence of the GM safflower.
Section 2.3
Vertical transfer of genes or genetic elements to sexually compatible plants
Expression of the introduced silencing constructs in safflower plants or weedy related species outside the trialEnvironmental harms associated with weediness; allergic reactions in people or toxicity in people and other organisms
No
  • Safflower does not produce fertile hybrids with weedy related species.
  • The applicant has proposed measures to isolate the trial from other safflower plants, which would minimise pollen-mediated gene transfer.
  • Risk scenarios 1 – 2 associated with the genetic modifications did not constitute identified risks for people or the environment.

2.1 Production of a substance toxic or allergenic to people or toxic to other organisms

Top of page
71. Toxicity is the adverse effect(s) of exposure to a dose of a substance as a result of direct cellular or tissue injury, or through the inhibition of normal physiological processes (Felsot 2000).

72. Allergenicity is the potential of a substance to elicit an immunological reaction following its ingestion, dermal contact or inhalation, which may lead to tissue inflammation and organ dysfunction (Arts et al. 2006).

73. A range of organisms may be exposed directly or indirectly to the introduced genetic constructs or their end products. Workers cultivating the GM safflower would be exposed to all plant parts. Organisms may be exposed directly to GM safflower plants through biotic interactions (vertebrates, invertebrates, symbiotic and/or pathogenic microorganisms), or through contact with dead plant material (soil biota) or indirectly through the food chain.

Risk Scenario 1. Exposure to GM plant material containing introduced silencing constructs or their end products


74. The introduced silencing constructs with fragments of safflower genes, or siRNAs produced by transcription of the silencing constructs, or safflower oil with altered composition could be toxic or allergenic for people, or toxic for other organisms. If humans or other organisms were exposed to the GM plant materials through ingestion, contact or inhalation, this may give rise to detrimental biochemical or physiological effects on the health of these people, animals or micro-organisms.

75. In the context of the proposed dealings, both of the following requirements would have to be met for GM safflower to have any increased toxic or allergenic effect:
  • the genetic modification would have to result in production of toxins or allergens either not present in commercially grown safflower varieties or at higher levels than present in commercially grown safflower varieties, and
  • humans or other organisms would have to be exposed to the GM safflower plants through contact, ingestion or inhalation.
76. The silencing constructs contain fragments of three safflower genes: FATB, FAD2 and another fatty acid biosynthesis gene. These gene sequences are naturally present in non-GM safflower as well. Humans and animals have a long history of safe exposure to non-GM safflower crops.

77. Transcription of the gene fragments in the silencing constructs produces RNA which forms a hairpin structure. This double-stranded RNA enters the RNAi pathway rather than being translated into a protein. Therefore, this gene silencing mechanism does not lead to expression of a novel protein that could potentially be toxic or allergenic.

78. Hairpin RNA transcribed from the silencing constructs is processed into siRNAs. Animals and plants naturally produce thousands of different siRNA molecules and these are consumed by humans whenever they eat plant or animal cells. One paper (Zhang et al. 2011) tracked the metabolic fate of a particular natural miRNA (similar to siRNA) that is produced abundantly in rice and happens to have a near perfect sequence match to a mammalian gene. In a study of mice fed a pure rice meal after fasting, the plant miRNA was detected in mouse livers and was reported to modulate the expression of the matching mammalian gene, reducing levels of the encoded protein in the liver by approximately 50%. The effect on the mouse gene by the plant miRNA was transient and ceased when rice was no longer included in the food intake. However, a recent analysis paper (Petrick et al. 2013) suggests some potential alternate explanations for the findings of the Zhang et al study (2011), and after reviewing a number of other papers in the field concludes that the weight of the evidence does not suggest that miRNAs derived from normal dietary exposure have a meaningful effect on mammalian gene expression.

79. The possibility exists that siRNAs produced in GM safflower lines could modulate expression of human or animal genes, with unknown physiological effects. However, to have any significant effect in people or animals, the GM safflower would need to constitute a large proportion of a large meal. The quantity of rice fed to mice in the Zhang et al study (2011) is equivalent to a human eating approximately 33 kg/day of cooked rice, which does not reflect anticipated dietary exposure levels. Also, the siRNAs would need to be produced at high levels in GM safflower, match a target sequence in a human or animal gene, and be taken up by cells expressing that gene. Mammals do not have genes that are homologous to the safflower fatty acid biosynthesis genes targeted by the introduced silencing constructs. Even if siRNAs were acquired through eating GM safflower and did affect expression of a mammalian gene, it is expected that any effect would be transient as described in Zhang et al (2011).

80. The expected phenotypic difference between GM and non-GM safflower is that GM safflower oil will contain a higher proportion of oleic acid and a lower proportion of saturated or polyunsaturated fatty acids. Oleic acid is part of the normal human diet, as it is a major constituent of vegetable oils and animal fats, and it is not toxic or allergenic.

81. The genetic modifications have the potential to cause unintended effects in several ways including off-target siRNA-mediated silencing of genes expressed in safflower, altered expression of endogenous safflower genes by random insertion of introduced DNA in the genome, and secondary effects arising from altered substrate or product levels in biochemical pathways. Unintended effects might result in adverse outcomes such as toxicity or allergenicity. Unanticipated changes can also be induced in plants by conventional methods of plant breeding (Haslberger 2003). The range of possible unintended effects produced by genetic modification is not likely to be greater than that from accepted traditional breeding techniques (Bradford et al. 2005; Committee on Identifying and Assessing Unintended Effects of Genetically Engineered Foods on Human Health 2004). More detail on potential for unintended effects as a result the process of genetic modification can be found in the document Methods of plant genetic modification available from the Risk Assessment References page on the OGTR website.

82. There is little potential for human ingestion of the GM safflower, as no GM plant material would be used as food. Similarly, livestock would not be intentionally exposed as the GM plant material would not be used as animal feed. The applicant proposes that GM plant materials will only be handled by trained and authorised staff. The short duration (2013-2016) and the small size (up to 3 ha per year) of the proposed field trial will also limit the potential for exposure to the GM plant material.

83. Further features of the GMOs that limit the potential for exposure to GM plant material have been declared confidential commercial information (CCI). The confidential information was made available to the prescribed experts and agencies that were consulted on the RARMP for this application.

84. Conclusion: The potential for harm due to exposure to GM plant material containing the introduced silencing constructs, in the context of the limits and controls proposed by the applicant and considering both the short and long term, is not identified as a risk that could be greater than negligible. Therefore it does not warrant further assessment.

2.2 Weediness of the GM safflower plants in the environment


85. This section addresses the question of whether or not the proposed dealings with the GMOs may lead to harm to human health and safety or the environment as a result of an increased potential for spread and/or persistence due to the genetic modification.

86. All plants have the potential to lead to harm in certain environments. Harms that may arise from a certain plant species in a particular environment include:
  • adverse effects on the health of people and/or animals
  • reduction in the establishment, yield and/or quality of desired plants
  • restriction in the physical movement of people, animals, vehicles, machinery and/or water
  • adverse effects on environmental health, such as adverse changes to strata levels, nutrient levels, fire regime, soil salinity, soil stability, or by providing food and/or shelter to pests, pathogens and/or diseases.
87. For the purpose of this document, plant species causing significant levels of one or more of these harms are called ‘weeds’. A plant species may be weedy in one or more land uses, such as dryland cropping or nature conservation.

88. Characteristics that influence the spread (dispersal of the plant or its genetic material) and persistence (establishment, survival and reproduction) of a plant species impact on the degree of its invasiveness. These characteristics include the ability to establish in competition with other plants, to tolerate standard weed management practices, to reproduce quickly, prolifically and asexually as well as sexually, and to be dispersed over long distances by natural and/or human means. The degree of invasiveness of a plant species in a particular environment gives an indication of the likelihood of its weediness in that environment. In addition to local experience, a history of weediness overseas can be used as an indicator for weediness in Australia (Pheloung et al. 1999).

89. Some baseline information on the weediness of safflower is given in Chapter 1, Section 6. Safflower is fairly slow-growing, with an extended rosette stage following emergence and prior to stem development, during which it is poorly competitive with other plants (Dajue & Mündel 1996). Safflower plants are susceptible to a wide range of herbicides as well as physical weed management practices (GRDC 2010).

Risk Scenario 2. The genetic modifications increase weediness of the GMOs

Top of page
90. In the context of the proposed dealings, in order for the GM safflower plants to become weedy in the environment both of the following conditions would need to be met:
  • GM safflower plants are present outside the limits (locations and/or duration) of the proposed trial; and
  • GM safflower plants are able to establish populations that cause harms associated with weediness.
Presence of GM safflower plants outside the trial limits

91. GM safflower plants could be present outside the trial limits due to survival at the trial sites after completion of the field trial, or due to dispersal of reproductive plant material outside the boundaries of the sites during or after the trial.

92. After completion of the trial, it is possible that whole GM plants could survive at the trial sites, or new volunteer plants could grow from residual seed in the trial sites. The applicant proposes a number of control measures to prevent these eventualities, including:
  • destroying all plant materials not required for testing or future trials
  • promoting germination of residual seed by post-harvest tillage and irrigation
  • post-harvest monitoring of the trial sites for two years and destruction of any volunteer safflower prior to flowering.
93. Typical safflower seed losses during harvest are 3-4%, and up to 50% of these residual seeds are viable (Mcpherson et al. 2009b). Most of these seeds would germinate soon after harvest as safflower seeds have very low dormancy (see Chapter 1, Section 4). It is not expected that the genetic modifications to safflower would affect seed yield, viability or germination. While the fatty acid composition is altered, the total fatty acid content of seeds, and thus their stored energy content, remains constant. GM safflower seeds grown in the greenhouse were reported to germinate and establish at the same rate as non-GM comparators. Likewise, it is not expected that the genetic modifications would affect the ability of the GMOs to survive the control measures listed above.

94. Potential dispersal of reproductive GM plant material outside the site boundaries would be limited to seed or pollen, as safflower does not reproduce vegetatively in the field. Safflower seed heads are resistant to shattering and the seeds lack seed dispersal characteristics such as stickiness, burrs and hooks, which can contribute to seed dispersal via animal fur or feathers (Howe & Smallwood 1982). These seed dispersal characteristics are not expected to be altered in the GMOs. Gene flow via pollen is discussed in Risk Scenario 3.

95. Dispersal of viable seed could occur in a variety of ways including: endozoochory (dispersal through ingestion by animals), through transport of seeds by animals, through movement of seeds by people, or through extremes of weather such as flooding or high winds.

96. Small birds can feed on ripening safflower seed in the head, and cockatoos can chew off safflower plants at the base in order to access the seeds (GRDC 2010). Safflower seeds that have passed through bird digestive systems are no longer viable (Cummings et al. 2008). Individual safflower seeds are smooth and unlikely to adhere to birds, and although entire seed heads are spiny, they are large and firmly attached to the plant. Some Northern Hemisphere birds engage in seed-caching behaviour but it is not known whether Australian birds carry seeds away for later consumption. The applicant proposes to prevent bird access to one trial site by bird netting, and to control bird pressure at the other two trial sites using bird scarers, which would minimise seed dispersal through bird activity.

97. Large animals are generally deterred from grazing on standing safflower by its spines. Safflower seeds are firmly held within their seed heads, which limits their accessibility to rodents. Residual GM seeds post-harvest may attract animal predation, and could be transported and hoarded by rodents. However, the applicant proposes to till the trial sites post-harvest, which should bury the GM seeds. A 10 m monitoring zone around the trial sites would be kept bare or mowed short, so should not attract or harbour rodents. The applicant also intends to monitor for rodents by trapping and to control populations by baiting if necessary. Hence seed dispersal through animal or rodent activity is unlikely.

98. Dispersal of seeds by people dealing with the GMOs would be minimised by cleaning of all equipment prior to removal from the trial sites. All GM plant material would be transported in accordance with the Regulator’s transport guidelines to avoid spillage.

99. Safflower is very resistant to shattering or lodging (Mündel et al. 2004), so seeds are unlikely to be dispersed by wind or via water runoff from irrigation or rainfall prior to harvest. Residual seeds that fall during harvest could be dispersed by water runoff from rainfall or by strong winds. However, the applicant proposes to till the trial sites post-harvest and incorporate GM plant material into the soil. Trial sites would be located at least 50 m away from natural waterways to minimise seed dispersal in the event of flooding. Seeds dispersed by flooding would be unlikely to survive and establish, as safflower is very susceptible to damping off and fungal diseases in wet soil (GRDC 2010).

100. The applicant has proposed that harvested seed not required for further experimentation may be buried as a means of disposal. Burial would be to a depth of at least 1 m, which would restrict access and dispersal by animals and dispersal by wind or water.

Establishment of GM safflower populations that cause harms associated with weediness

101. As summarised in Chapter 1 Section 6, safflower is naturalised throughout Australia, primarily as an agricultural or ruderal weed. In New South Wales agricultural areas, it is classified as a Category 1 weed, indicating that it may be a minor problem but is not considered important enough to warrant control at any location (Groves et al. 2003).

102. The only expected phenotypic difference between GM safflower and non-GM safflower is altered fatty acid composition in the GM safflower oil. In the unlikely event of GM safflower plants establishing themselves beyond trial limits, this trait would not lead populations of GM safflower to cause greater environmental harms associated with weediness, such as reducing establishment of desired plants, restricting physical movement, or adversely affecting environmental health, than would be caused by unmodified safflower.

103. As discussed in Risk Scenario 1, the genetic modifications have the potential to cause unintended effects. Unintended phenotypic changes could lead to increased weediness. However, the range of possible unintended effects produced by genetic modification is not likely to be greater than those caused by traditional breeding techniques (Bradford et al. 2005; Committee on Identifying and Assessing Unintended Effects of Genetically Engineered Foods on Human Health 2004). No unexpected phenotypic changes were observed during growing of GM safflower lines in greenhouses, and a standard condition of a licence for a field trial would be that the applicant must immediately notify the OGTR of any unintended effects of the dealings authorised by the licence.

104. Toxicity and allergenicity of the GM safflower were considered in Risk Scenario 1 and it is unlikely that GM safflower plants would have higher toxicity and/or allergenicity than non-GM safflower.

105. Conclusion: The potential for harm due to the genetic modification increasing the weediness of the GMOs, in the context of the limits and controls proposed by the applicant and considering both the short and long term, is not identified as a risk that could be greater than negligible. Therefore, it does not warrant further assessment.

2.3 Vertical transfer of the genetic elements to sexually compatible plants


106. Vertical gene flow is the transfer of genetic information from an individual organism to its progeny by conventional heredity mechanisms, both asexual and sexual. In flowering plants, pollen dispersal is the main mode of gene flow (Waines & Hegde 2003). For GM crops, vertical gene flow could therefore occur via successful cross-pollination between the crop and neighbouring crops, plants, related weeds or native plants (Glover 2002).

107. It should be noted that vertical gene flow per se is not considered an adverse outcome, but may be a link in a chain of events that may lead to an adverse outcome. For an increased potential for adverse effects to arise as a result of gene flow of the introduced genetic elements from the GM safflower to sexually compatible plants, both of the following steps must occur:
  • transfer of the introduced genetic elements to sexually compatible plants
  • increased potential for adverse effects, such as toxicity, allergenicity or weediness of the recipient plants, due to expression of the introduced genetic elements.
108. As summarised in Chapter 1 Section 4, safflower reproduces by a combination of self-pollination and bee-mediated cross-pollination.

109. As described in Chapter 1 Section 6, interspecific hybridisation between safflower and other species of the Carthamus genus present in Australia is difficult due to various cytogenetic barriers (e.g. varying chromosome number). Hybrids can be obtained under experimental conditions but are sterile (Mayerhofer et al. 2011).

Risk Scenario 3. Expression of the introduced silencing constructs in safflower plants or weedy related species outside the field trial

Top of page
110. If the introduced silencing constructs were transferred and expressed in other safflower plants or related species, the resulting hybrid plants could have increased toxicity or allergenicity to people, toxicity to other organisms, or weediness potential.

111. Expression of the introduced silencing constructs, leading to altered oil composition, is not expected to change the pollination characteristics of the GM safflower compared to non-GM safflower.

112. GM safflower could cross-pollinate plants from other Carthamus species at low levels if these weedy species were present in close proximity to the trial sites and flowered synchronously. Hybrids between GM safflower and Carthamus weeds would be annuals like all Carthamus species and would be sterile (Mayerhofer et al. 2011). The hybrids could therefore only be transient weeds in the immediate environs of the trial sites, and could not lead to long-term transfer of the introduced silencing constructs into weedy Carthamus species populations. Nonetheless, the applicant proposes to inspect the areas within 200 m of the trial sites prior to flowering of GM safflower and to destroy any plants from related species, which would minimise cross-species pollination.

113. GM safflower could cross-pollinate non-GM safflower plants outside the trial if either naturalised safflower or commodity safflower crops were present in proximity to the trial sites. In principle, GM safflower pollen could be widely dispersed, as bees forage over kilometre ranges. However, safflower pollen transported by an insect must compete with the floret’s own pollen to result in outcrossing. Bee-mediated cross-pollination of safflower has low efficiency, as transported safflower pollen is only reported to potentially fertilise the next floret visited by the bee (Cresswell 2010). In contrast, in canola crops pollen collected by a bee at one flower may fertilise up to twenty flowers visited subsequently (Cresswell et al. 2002).

114. Outcrossing rates between adjacent safflower plants in India have been reported to range between 0-59%, depending on cultivar (Singh & Nimbkar 2006). However, the higher range of outcrossing rates may result from physical contact between florets rather than insect-mediated cross-pollination. In a series of experiments in Spain where recipient safflower plants were surrounded by donor safflower plants, but separated by distances of 1-1.5 m to prevent physical contact, average outcrossing rates were 5.7-13.2% (Velasco et al. 2012). In general, commercial safflower varieties in Australia are reported to have less than 10% outcrossing unless bee hives are brought in specifically for the purpose (GRDC 2010). The particular parent cultivars in this proposal are also expected to be predominantly self-pollinating. Additional information on these cultivars has been declared CCI. The confidential information was made available to the prescribed experts and agencies that were consulted on the RARMP for this application.

115. Limited information is available on safflower cross-pollination over distance, or on the efficacy of exclusion distances in preventing hybridisation. For a commercial safflower cultivar studied in Canada, cross-pollination rates were measured as 1.7% at 3 m, approximately 0.01% at 100 m, and no outcrossing in 85,000 tested plants at 300 m (Mcpherson et al. 2009a). However, the experimental design involved four continuous strips of recipient safflower plants 107 m long and 1.6 m wide extending outwards from the donor plot, so the inner recipient plants may have provided a partial barrier against pollen flow to the outermost plants. International guidelines for seed-growers require that crops of certified safflower seed be grown with an exclusion distance of 200 m from other safflower crops, and that crops of basic safflower seed be grown with an exclusion distance of 400 m (OECD 2013). These international guidelines were developed for conditions where pollinators include both bumblebees and honeybees. Bumblebees are reported to be more effective at field-to-field pollination of safflower than honeybees (Cresswell 2010), so long-distance outcrossing rates may be reduced in mainland Australia compared to other countries due to the lack of bumblebees.

116. The applicant proposes to inspect the monitoring and isolation zones within 200 m of the trial sites prior to flowering of GM safflower and to destroy any safflower plants. This would minimise vertical gene flow to any safflower populations within 200 m of the trial site.

117. The applicant also proposes to ensure that no other safflower crops are grown within 400 m of the trial sites. There is a possibility that GM safflower could pollinate dense populations of naturalised safflower growing in the areas between 200 m and 400 m from the trial sites, but as safflower is not a common weed, the number of wild safflower plants present in these areas is likely to be very low. However, significant numbers of wild safflower plants might be present between 200 m and 400 m from the trial sites if the plants were growing as volunteers following planting of a safflower crop in the previous year.

118. Cross-pollination events between the GMOs and other safflower crops or dense populations of wild safflower located slightly further than 400 m from the trial sites are expected to be rare. However, there is no published quantitative information about long-distance safflower gene flow.

119. Even if the introduced genetic material was transferred from the GM plants to other safflower plants, it is unlikely that expression of the silencing complexes would cause harm. As discussed in Risk Scenario 1 and Risk Scenario 2, the trait of increased levels of oleic acid is not expected to produce any toxic and/or allergenic substance or to increase weediness in recipient plants.
Top of page
120. Conclusion: The potential for increased allergenicity in people, toxicity in people and other organisms, or increased weediness due to the expression of the introduced genes or gene sequences in commercial safflower crops or other sexually compatible plants as a result of gene transfer, in the context of the limits and controls proposed by the applicant and considering both the short and long term, is not identified as a risk that could be greater than negligible. Therefore it does not warrant further assessment.