Novel Food Information - Herbicide Tolerant Soybean SYHT0H2

Health Canada has notified Syngenta Seeds Canada Inc. and Bayer CropScience Inc. that it has no objection to the sale of food derived from Herbicide Tolerant Soybean SYHT0H2. The Department conducted a comprehensive assessment of this soybean event according to its Guidelines for the Safety Assessment of Novel Foods. These Guidelines are based upon internationally accepted principles for establishing the safety of foods with novel traits.

Background:

The following provides a summary of the notification from Syngenta Seeds Canada and Bayer CropScience and the evaluation by Heath Canada. This document contains no confidential business information.

1. Introduction

Syngenta Seeds Canada and Bayer CropScience developed Herbicide Tolerant Soybean SYHT0H2 using recombinant DNA techniques to introduce the p-hydroxyphenylpyruvate dioxygenase (HPPD) coding sequence (avhppd-03) derived from Avena sativa (oat). Herbicides that inhibit HPPD have a lower binding affinity for the expressed AvHPPD-03 protein, which acts in place of the native soybean HPPD in the tyrosine catabolic pathway to provide the desired herbicide tolerance. This event was also modified to introduce two copies of the coding sequence for phosphinothricin acetyl-transferase (pat) from Streptomyces viridochromogenes strain Tü494. The expressed PAT protein acetylates L-phosphinothricin, the active compound in glufosinate herbicides, rendering it inert.

The safety assessment performed by Food Directorate evaluators was conducted according to Health Canada's Guidelines for the Safety Assessment of Novel Foods. These Guidelines are based on harmonization efforts with other regulatory authorities and reflect international guidance documents in this area (e.g., Codex Alimentarius). The assessment considered: how Herbicide Tolerant Soybean SYHT0H2 was developed; how the composition and nutritional quality of Herbicide Tolerant Soybean SYHT0H2 compared to non-modified varieties; and what the potential is for Herbicide Tolerant Soybean SYHT0H2 to be toxic or cause allergic reactions. Syngenta and Bayer CropScience have provided data which demonstrates that Herbicide Tolerant Soybean SYHT0H2is as safe and of the same nutritional quality as traditional soybean varieties used as food in Canada.

The Food Directorate has a legislated responsibility for pre-market assessment of novel foods and novel food ingredients as detailed in Division 28 of Part B of the Food and Drug Regulations (Novel Foods). Foods derived from Herbicide Tolerant Soybean SYHT0H2 are considered novel foods under the following part of the definition of novel foods: "c) a food that is derived from a plant, animal or microorganism that has been genetically modified such that

1. the plant, animal or microorganism exhibits characteristics that were not previously observed in that plant, animal or microorganism."

2. Development of the Modified Plant

The petitioners have provided information describing the methods used to develop Herbicide Tolerant Soybean SYHT0H2 (hereafter referred to as SYHT0H2) and molecular biology data that characterize the genetic change that confers herbicide tolerance. SYHT0H2 was produced using Agrobacterium tumefaciens (A. tumefaciens) mediated transformation of soybean variety 'Jack' with the transformation vector pSYN15954. This transformation vector was constructed to contain a single transfer DNA (T-DNA) region, which contained three gene expression cassettes.

The first expression cassette contains the coding region that confers tolerance to HPPD herbicides. This coding region produces an HPPD protein derived from Avena sativa (oat). The gene, avhppd-03, is a synthetic copy of the native oat hppd coding sequence that has been codon optimized for expression in soybean. This coding sequence results in an HPPD protein that is 99.7% identical to the wild-type oat protein.

The second and third expression cassettes each contain the coding region that confers tolerance to glufosinate ammonium herbicides. Two versions of the pat coding region were used in the development of SYHT0H2, pat-03-01 and pat-03-02. These coding regions produce the PAT protein from Streptomyces viridochromogenes strain Tü494. The native coding sequence was codon optimized for soybeans, and in the case of pat-03-02 was altered to remove restriction sites. Both versions of the coding sequence produce a PAT protein 100% identical to the wild-type. The petitioners have indicated that two copies of the pat coding sequences were included to ensure an acceptable level of tolerance to the herbicides.

Maturing soybean pods were harvested from greenhouse grown plants, sterilized with diluted bleach solution and rinsed. Immature seeds were excised from the pods, sterilized and rinsed. The explants were then prepared from these immature seeds, infected with A. tumefaciens containing the pSYN15954 plasmid and incubated for 30 to 210 minutes. Excess A. tumefaciens was then removed from the explants via aspiration, and the explants were then moved to non-selective co-culturing medium plates. These plates, containing the explants and any remaining A. tumefaciens were co-cultured in the dark for 4 days. After 4 days, the explants were removed from the plates and placed on regeneration medium supplements with antibiotics (ticarcillin, cefotaxime and vancomycin) to kill the remaining A. tumefaciens. These plates were then incubated for an additional 7 days in the dark. The explants were then transferred to cell-culture medium containing the same antibiotic mixture and a glufosinate-ammonium concentration low enough to permit optimal shoot growth in plants containing the pat gene. The regenerative plantlets that survived this selection were tested via real-time polymerase chain reaction (rtPCR) for the presence of the avhppd-03 and pat genes and the absence of the plasmid backbone spec gene. All plantlets positive for these conditions were then transferred to the greenhouse for seed setting. These plantlets were considered to be the T0 plants.

The T0 plants were field tested for tolerance to mesotrione (HPPD inhibiting herbicide) and glufosinate-ammonium herbicides and agronomic performance. From this testing, SYHT0H2 was selected as a lead candidate for development and use in performing regulatory studies. The petitioners have provided a complete breeding tree for SYHT0H2. As of the date of submission, the petitioner had generated up to the T8 generation via self-pollination. For the regulatory studies, the petitioners have identified that the T4 generation was used in molecular characterization of SYHT0H2, with the exception of the stability analysis, where T4, T5 and T6 were used in testing. The petitioners also state that for all studies the parental line "Jack" was used as the non-isogenic control.

3. Characterization of the Modified Plant

Using 9 overlapping PCR primer pairs, the petitioners generated PCR fragments covering the entire length of the insert. These fragments were cloned and then sequenced. The 9 fragments were then aligned to generate the final consensus sequence for the entire insert, which was compared to the pSYN15954 T-DNA sequence. From this sequence analysis, it was determined that SYHT0H2 contains a single insert consisting of two inverted and truncated copies of the T-DNA. The inverted copies are centred around the RB region, which is truncated in both copies. In addition to the truncations, two additional DNA sequences have been identified within the DNA insert. The first sequence, of 44 bp, resides between the two inverted copies and has some similarity to the avhppd-03 coding sequence. The second insert, of 17 bp, is located in the 35S promoter of the 3' pat-03-01 cassette.

Based on the sequencing analysis SYHT0H2 soybeans contain a single insert comprised of two complete copies of the pat-03-02 cassette, a single complete copy of the pat-03-01 cassette, a second truncated copy of the pat-03-01 cassette, and a single, truncated copy of the avhppd-03 cassette.

The petitioners conducted further sequence analysis to assess the flanking regions in the soybean genome. This analysis demonstrated that 15 bp of genomic DNA present in "Jack" soybeans, had been deleted from SYHT0H2 soybean. Furthermore, this analysis demonstrated a 7 bp insertion in the flanking genomic DNA. These deletions and insertions have been shown to occur through the use of A. tumefaciens transformations, likely as a result of double stranded break repair. Similar occurrences have been observed and reported in the literature as a result of Agrobacterium mediated transformations.

In addition to sequencing, the petitioners used Southern blots to confirm the number of copies of each functional element within the insert. To do this, the petitioners used a variety of restriction enzymes and six different probes, each specific for one or more functional elements in the T-DNA. The Southern blot data presented by the petitioners confirmed the expected copy number of each functional element, as predicted by the sequence analysis, further confirming a single truncated repeat insertion of the T-DNA in SYHT0H2.

Generational stability of the single insert was determined across three generations of SYHT0H2 (T4, T5, and T6). Southern blot analysis was presented for each generation confirming the presence of the T-DNA cassette and its stability across generations. Similarly, the petitioners performed the same methodology to demonstrate a lack of plasmid backbone sequences across generations. The evidence presented confirmed a lack of backbone sequences.

The petitioners undertook a bioinformatic analysis of the insertion site to determine whether the transformation had interrupted a known soybean gene. To do this, the genomic flanking sequences were screened for similarities to known sequences found in the public databases. This screen was completed on the 1000 base pairs found flanking both the 5' and 3' ends of the insert. From the analysis, the majority of the matches were considered to be the result of random occurrence. For those matches that were less likely to be the result of random occurrence, the alignments were predominately with unannotated regions. For those few alignments which were with annotated sequences, the region of similarity was distant from the insert junction and in no cases did the similarity correspond across the insert. Based on this analysis, it was concluded that the insert did not interrupt any known soybean genes.

The petitioners completed an analysis of any potential opening reading frames created by the insertion into the genomic DNA. Bioinformatic analysis was used to assess the similarity between any potential reading frames and databases of known toxins and allergens. Open Reading Frames (ORFs) spanning the 5' flanking sequence DNA-inserted DNA junction, and 3' flanking sequence DNA-inserted DNA junction were translated from putative start codons (ATG) to putative stop codons (TAG, TAA and TGA). This approach identified 47 potential ORFs (excluding the five expected ORFs) within the insert and 1 ORF within the 5' insert junction. The 48 theoretical proteins produced by the putative ORFs were compared with a database of all the toxins listed in the NCBI Entrez Protein database. No significant similarity to any proteins found in the toxin database were identified. Therefore, it is not considered likely that the theoretical proteins produced by the putative ORFs would be a toxin.

Additionally, the petitioners used the recommended bioinformatics approach to assess the putative ORFs for potential allergenicity. Using the FAARP database, the petitioners looked for sequence identity of 35% or greater in a sliding window of 80 amino acids. This was followed by a search for identity between regions of 8 a.a. Neither of these searches found any similarity with proteins in the database. So, it is not considered likely that the theoretical proteins produced by the putative ORFs would be an allergen. Further to this work, the petitioners also conducted an analysis of the theoretical ORFs proximity to known promoters and the genetic context of the start codon. Of the identified ORFs, bioinformatic analysis ruled out potential expression of all but a single theoretical ORF. The remaining ORF had already been analysed and determined not to have any sequence similarity to known toxins or allergens.

The petitioners have conducted the toxicological assessment using AvHPPD-03 and PAT proteins expressed in and purified from Escherichia coli. To ensure that the results of the toxicological studies are applicable to the proteins expressed in SYHT0H2 soybean, equivalence studies (i.e., SDS-PAGE, western blot analysis, glycoprotein staining, MALDI-TOF MS, N-terminal amino acid sequence analysis and specific enzymatic activity) were conducted to confirm that the proteins produced in E. coli used for toxicology studies are representative of the proteins produced in the modified soybean. Based on the results of these studies, the proteins were determined to be equivalent with respect to their physical properties, immunological staining properties and sequencing.

4. Product Information

Herbicide Tolerant Soybean SYHT0H2 differs from conventional soybean by the insertion of two novel genes, avhppd-03 and pat, and their associated regulatory elements. The insertion of these genes results in the expression of two novel proteins for this soybean variety: AvHPPD-03 and PAT. The expression of AvHPPD-03 confers tolerance to HPPD herbicides. The expression of PAT in SYHT0H2 confers tolerance to glufosinate ammonium herbicides.

The AvHPPD-03 protein functions in soybeans in place of the native soybean HPPD as part of the tyrosine catabolic pathway. HPPD enzymes are found in nearly all aerobic forms of life and catalyze the conversion of 4-hydroxyphenylpyruvate (HPP) to homogentisic acid (HGA), the precursor to plastoquinone and tocopherol (Vitamin E) biosynthesis. This biosynthetic pathway, including HPPD, is found in nearly all aerobic organisms, including plants, animals and bacteria, and is important in both photosynthesis and cellular metabolism via the citric acid cycle. The widespread presence of HPPD in plants and animals, and within the food supply, supports the overall safety of this coding sequence.

PAT is a highly specific enzyme for acetylation of glufosinate ammonium herbicides and has been shown not to acetylate glutamine, the closest structural analogue, or other L-amino acids. PAT belongs to a class of acetyltransferases common in plants and animals and shares structural and functional properties with other acetyltransferases common in the human diet. Numerous previously reviewed genetically modified crops have used PAT as the source of glufosinate herbicide tolerance, and there has not been any evidence, to date, suggesting that it is unsafe for human consumption.

5. Dietary Exposure

SYHT0H2 soybean is expected to be used in applications similar to those of other soybean varieties. Soybeans are the largest source of vegetable oil worldwide. Refined, bleached, and deodorized soybean can be further processed to produce cooking oils, shortening, margarine, mayonnaise, salad dressings, and a wide variety of products that are either based entirely on fats and oils or contain fat or oil as a principal ingredient. Soy protein isolate is also used in such foods as soups, sauce bases, energy bars, nutritional beverages, infant formula, and dairy replacements. As the uses of SYHT0H2 soybean are not expected to differ, no increase in dietary exposure to soy is expected to occur.

The petitioners have provided data to demonstrate the level of expression of both AvHPPD-03 and PAT proteins in SYHT0H2. This study used plant samples from four field sites planted in the growing season in the major soybean growing locations in Argentina. Each site was planted with two replicates of test soybeans and one of control soybeans. One of the plots was sprayed with mesotrione and glufosinate once during growth. All plots were grown according to local agronomic practices; this included the use of a variety of other commercially available herbicides, fungicides and insecticides. Protein quantification was conducted via validated ELISA in leaves (at four growth stages), roots (at two growth stages), forage and seed. For each sampled plant part, and at each growth stage, 5 test samples per plot plus two control samples were collected. Analysis of the near-isogenic control line confirmed the specificity of the ELISA to AvHPPD-03, not the native soybean HPPD.

Protein quantities for the tissues were calculated on a microgram (µg) per gram (g) dry weight (dwt) basis. The mean AvHPPD-03 protein concentration across locations for seed, the food component, was 8.18 (0.62-28.30) µg/g dwt. Additionally, the petitioners provided the mean (range) values for leaves (at V4, V8, V10 and R6 growth stages), roots (at V8 and R6 growth stages) and forage. The AvHPPD-03 concentration for each of the plant parts at each growth stage was, 242.00 (20.23 - 585.46), 212.98 (53.77 - 386.15), 165.14 (55.96 - 302.90), 105.32 (16.94 - 255.30), 79.49 (15.43 - 201.47), 22.50 (1.50 - 69.95) and 79.66 (16.76 -164.01) µg/g dwt, respectively.

Protein quantities for the tissues were calculated on a microgram (µg) per gram (g) dry weight (dwt) basis. The mean PAT protein concentration across locations for seed, the food component, was 2.70 (0.07-14.85) µg/g dwt. Additionally, the petitioners provided the mean (range) values for leaves (at V4, V8, V10 and R6 growth stages), roots (at V8 and R6 growth stages) and forage. The PAT concentration for each of the plant parts at each growth stage was, 52.21 (0.89 - 167.97), 23.00 (2.04 - 83.43), 38.23 (4.77 - 115.86), 29.41 (0.77 - 101.58), 21.16 (0.33 - 46.07), 9.12 (0.32 - 29.35) and 19.17 (1.12 -60.91) µg/g dwt, respectively.

6. Nutrition

The nutrient data for this submission was obtained from test (SYHT0H2) with and without herbicide treatment and control soybean grown in 2010 at 8 USA locations in a randomized complete block design with four plots for test, control and reference varieties. All field trial experiments used to test SYHT0H2 soybean were acceptable. All analyses of test, control and reference varieties were done using approved scientific and appropriate statistical methods.

Test, control and reference varieties were analyzed for nutrients and anti-nutrients, in seed, as follows: Proximates: moisture, crude protein, crude fat, ash, carbohydrates; plus acid detergent fibre (ADF), neutral detergent fibre (NDF); Minerals: calcium, iron, magnesium, phosphorus, potassium; Fatty Acids: C8 to C22; Amino Acids: aspartic acid, threonine, serine, glutamic acid, proline, glycine, alanine, cysteine, valine, methionine, isoleucine, leucine, tyrosine, phenylalanine, histidine, lysine, arginine, and tryptophan; Vitamins: folic acid, vitamin A (beta-carotene), vitamin B1, vitamin B2, vitamin E (α-tocopherol, β-tocopherol, γ-tocopherol, δ-tocopherol, α-tocotrienol, β-tocotrienol, γ-tocotrienol, δ-tocotrienol), Vitamin K; Anti-Nutrients: lectin, phytic acid, raffinose, stachyose, and trypsin inhibitor, Other Substances: daidzein, glycitein, genistein.

For combined locations, statistical differences (SYHT0H2 vs. control) were noted in 25 analytes as follows: acid detergent fibre, iron, potassium, palmitic acid, stearic acid, oleic acid, linoleic acid, linolenic acid, arachidic acid, behenic acid, aspartic acid, threonine, serine, glutamic acid, proline, alanine, leucine, tyrosine, phenylalanine, lysine, histidine, arginine, α-tocopherol, γ-tocopherol, and δ-tocopherol. For nutrients where a statistical difference was determined, all mean values were within reference and/or literature ranges.

7. Toxicology

Health Canada has previously approved the expression of S. viridochromogenes PAT protein in soybean, cotton, sugarbeet, canola and maize lines intended for human consumption. The safety of PAT protein is well established and as such, the consumption of PAT protein from SYHT0H2 soybeans is not expected to pose a toxicological or allergenic human health concern.

SYHT0H2 produces a transgenic AvHPPD-03 enzyme which shares 99.7% amino acid identity with native HPPD from oat. Oat is a cereal grain that has a history of safe food use by Canadians. It is commercially available as a processed commodity in porridge, breakfast cereal, cookies and snack bars. Incidental exposure to HPPD protein from oats in the diet is not associated with any known toxicity, allergenicity or anti-nutrient effect. Thus, it is unlikely that AvHPPD-03 protein expressed in SYHT0H2 will have a toxic, allergenic or anti-nutrient effect when consumed.

Microbially-derived AvHPPD-03 did not produce any adverse effects when given as an acute oral dose to mice. These results demonstrate that ingestion of AvHPPD-03 would not result in adverse health effects at doses up to 2 g AvHPPD-03/kg b.w. (the highest dose tested).

The petitioners provided the results of an in silico search which compared the amino acid sequences of AvHPPD-03 to amino acid sequences of known toxins. The results of the search determined that AvHPPD-03 has significant sequence similarity to bacterial hemolytic proteins which have HPPD-like function. The petitioners have shown that SYHT0H2 soybean plants do not exhibit any phenotypic or compositional changes that would suggest that AvHPPD-03 has unintended properties (e.g., HPPD enzymatic pathway disruptions). Further, the acute animal feeding study did not show any indication of hemolysis in treated mice. Finally, hemolysis in humans occurs due to the accumulation and polymerization of a HPPD pathway metabolite, homogentisic acid (HGA), in the blood; however this has only been observed under very specific conditions, such as blood-borne bacterial infection or hereditary disorder. These conditions cannot be met through the consumption of AvHPPD-03 in SYHT0H2 soybeans. AvHPPD-03 in SYHT0H2 would not be expected to pose any greater risk of hemolysis in humans than incidental exposure to endogenous HPPD in oat, to which humans have a history of safe exposure. Because oat, the source organism for the AvHPPD-03 sequence, has been implicated as a potential source of proteins that are avoided by individuals with celiac disease, the in silico search also compared the amino acid sequence of AvHPPD-03 to amino acid sequences of putative or known allergens, including allergens listed in the FARRP AllergenOnline Celiac Database (2013). The results of the search determined that AvHPPD-03 protein did not share significant sequence similarity with putative or known allergens, including proteins associated with celiac disease. Although the safety of oats in a gluten-free diet for individuals with celiac disease has historically been an issue of debate, Health Canada (2007)^{Footnote 1} concluded, based on an extensive review of the scientific literature, that the majority of people with celiac disease can tolerate pure oats, i.e., oats uncontaminated with other cereal grains known to cause celiac disease.

The petitioners provided the results of in vitro simulated gastric and intestinal fluid assays, which demonstrated that microbially-derived AvHPPD-03 protein was rapidly digested (within 5 minutes). As such, ingested AvHPPD-03 is unlikely to gain access to the general circulatory system and therefore unlikely to exert a systemic toxic or allergenic effect. The results of a heat stability assay showed that microbially-derived AvHPPD-03 was denatured and inactivated at temperatures commonly involved in soybean processing (> 65 ºC). This finding suggests that consumers would not be exposed to intact or active protein.

The expression level of AvHPPD-03 protein in SYHT0H2 soybean seed is considered very low (0.002% total protein in SYHT0H2 soybean seeds d.w.). The small quantity of AvHPPD-03 expressed in SYHT0H2 soybean would not be expected to survive digestion or food processing and so exposure to active AvHPPD-03 would be considered negligible. The estimated intake in the subpopulation which consumes the greatest amount of soy protein (on a w/w basis), infants 0-2 months of age (95th percentile, eater's only), is expected to be very low (110.36 µg AvHPPD-03 proteins and 36.42 µg PAT proteins per kg b.w./day). The margin of exposure (MOE) between the predicted levels of exposure and the no-observable-effect level (NOEL) in the acute oral toxicity studies in mice was about 5.5 x 105 for AvHPPD-03 protein. The MOE is considered sufficiently large to account for any uncertainties in the toxicological database and exposure estimates.

A quantitative mass spectrometry assay was performed to quantify the concentration of 12 known endogenous soybean allergens in trypsin-treated protein extracts from SYHT0H2. The resulting concentrations were compared to those of the parental control and 17 reference soybean varieties. Results of this experiment suggest that the levels of the 12 assayed allergens, as well as the total allergen content, in SYHT0H2 were within the natural range normally observed in non-transgenic soybeans. SYHT0H2 soybeans are unlikely to be any more allergenic than non-transgenic soybeans that are currently available to consumers.

Based on the low levels of expected exposure, their rapid degradation when exposed to simulated gastric fluid, denaturation at high temperatures and the absence of similarity to known toxins and allergens, no toxicity or allergenicity concerns with SYHT0H2 were identified.

Conclusion:

Health Canada's review of the information presented in support of the food use of Herbicide Tolerant Soybean SYHT0H2 concluded that derived food products do not raise concerns related to safety. Health Canada is of the opinion that Herbicide Tolerant Soybean SYHT0H2 is similar to regular conventional commodity soybean in terms of being an acceptable food source.

Health Canada's opinion deals only with the human food use of Herbicide Tolerant Soybean SYHT0H2. Issues related to the environmental safety of Herbicide Tolerant Soybean SYHT0H2 in Canada and its use as livestock feed have been addressed separately through existing regulatory processes in the Canadian Food Inspection Agency.

This Novel Food Information document has been prepared to summarize the opinion regarding the subject product provided by the Food Directorate, Health Products and Food Branch, Health Canada. This opinion is based upon the comprehensive review of information submitted by the petitioner according to the Guidelines for the Safety Assessment of Novel Foods.

For further information, please contact:

Novel Foods Section
Food Directorate
Health Products and Food Branch
Health Canada, PL2204A1
251 Frederick Banting Driveway
Ottawa, Ontario K1A 0K9
novelfoods-alimentsnouveaux@hc-sc.gc.ca