Biotechnology Consultation Note to the File BNF No. 000108
Return to inventory: Completed Consultations on Foods from Genetically Engineered Plant Varieties
Date: September 21, 2007
Subject: Biotechnology Notification File (BNF) BNF 000108; GAT4601 and GM-HRA proteins; Soybean Transformation Event 356043
Keywords: Soybean, Glycine max (L.) Merr., Jack cultivar, transformation event 356043, herbicide tolerance, glyphosate, glyphosate acetyltransferase, GAT, GAT4601 protein, gat4601 gene, Bacillus licheniformis, acetolactate synthase (or acetohydroxyacid synthase or AHAS), ALS-inhibiting herbicides, GM-HRA protein, gm-hra gene
In a submission dated November 16, 2006, Pioneer Hi-Bred International, Inc. (Pioneer), a DuPont company, submitted to FDA a safety and nutritional assessment of genetically engineered dual herbicide tolerant soybean, transformation event 356043 (hereafter referred to as 356043 soybean). Pioneer submitted additional information dated December 4, 2006; May 29, 2007; June 5, 2007; August 22, 2007; August 27, 2007; August 30, 2007; and September 14, 2007. Pioneer concluded that food and feed derived from 356043 soybeans are as safe and nutritious as food and feed derived from conventional soybeans.
2. Intended Effect
The intended effect of the modification is to confer tolerance to both glyphosate and acetolactate synthase (ALS)-inhibiting herbicides. To accomplish this objective, Pioneer introduced a glyphosate N-acetyltransferase (gat4601) gene, derived from the sequences of three gat genes from Bacillus licheniformis (B. licheniformis), and a modified acetolactate synthase (gm-hra) gene, derived from the soybean gm-als gene, into the 'Jack' soybean variety. The GAT4601 protein, encoded by the gat4601 gene, confers tolerance to glyphosate; the GM-HRA protein, encoded by the gm-hra gene, confers tolerance to the ALS-inhibiting class of herbicides.
3. Genetic Modifications and Characterizations
3.1 Parental Variety
Pioneer used the Jack cultivar as the parental variety for transformation. The Jack cultivar is a publicly available variety that is homozygous at nearly all loci.
3.2 Transformation Plasmid and Method
Pioneer constructed plasmid PHP20163 that contains two expression cassettes and an antibiotic (hygromycin) resistance marker gene. From this plasmid, a linear fragment was excised using two restriction enzymes Asc I and Not I. This linear fragment, called PHP20163A, was used for transformation and is therefore referred to as PHP20163A transformation fragment. The PHP20163A transformation fragment contains two expression cassettes, gat4601 gene cassette and the gm-hra gene cassette, but does not contain the antibiotic resistance marker gene. The details of each of these two gene cassettes are discussed below.
- The gat4601 gene cassette contains the coding sequence of the gat4601 gene, which encodes the GAT4601 protein. It is a synthetic gene derived from three gat genes from B. licheniformis. In planta expression of the gat4601 gene is under the control of two regulatory elements, the SCP1 promoter and the TMV omega 5'-UTR. The SCP1 promoter is a synthetic constitutive promoter containing a portion of the CaMV 35S promoter and the Rsyn7-Syn II Core synthetic consensus promoter. The TMV omega 5'-UTR is the omega 5' untranslated leader of the Tobacco Mosaic Virus. The SCP1 promoter drives the transcription of the gat4601 gene.The TMV omega 5'-UTR, which is located downstream from the SCP1 promoter, enhances translation. Termination of transcription of the gat4601 gene is under the control of the pinII terminator, the 3' terminator sequence from the proteinase inhibitor II gene of Solanum tuberosum.
- The gm-hra gene cassette contains the coding sequence of the gm-hra gene, which encodes the GM-HRA protein. It is a modified version of the endogenous soybean acetolactate synthase gene (gm-als), which encodes the GM-ALS I protein.1 Compared to the GM-ALS I protein, the GM-HRA protein sequence contains two amino acid substitutions important for tolerance to the ALS-inhibiting class of herbicides, and five additional N-terminal amino acids derived from the translation of 15 nucleotides of the gm-als 5' untranslated region. In planta expression of the gm-hra gene is controlled by the promoter derived from the S-adenosyl-L-methionine synthetase (SAMS) gene from soybean, and an intron that interrupts the SAMS 5' untranslated region (5'-UTR). Termination of transcription of the gm-hra gene is under the control of the native soybean acetolactate synthase terminator (gm-als terminator).
Soybean embryonic somatic cultures, derived from explants from small, immature soybean seeds of the Jack cultivar, were transformed by particle bombardment with microscopic gold particles coated with the purified PHP20163A DNA fragment. Following transformation, the soybean tissue was transferred to liquid culture flasks and allowed to recover for seven days before being transferred to a selective medium supplemented with chlorsulfuron - a member of the ALS-inhibiting class of herbicides. Only soybean cells that had inherited the gm-hra transgene would be expected to grow in the selective medium. Chlorsulfuron-tolerant green embyrogenic tissue was excised after several weeks and subsequently regenerated (T0 soybean plants).
3.3 Characterization, Inheritance, and Stability of the Introduced DNA
Pioneer summarized the results of the genomic DNA blot (Southern) analyses used to characterize the introduced DNA in 356043 soybean. Pioneer states that 356043 soybean contains one insert with a single, intact copy of the transformation fragment PHP20163A containing the gat4601 gene and gm-hra gene cassettes. Pioneer also conducted Southern blot analysis using a plasmid backbone-specific probe, and confirmed that plasmid backbone sequences, such as those coding for the hygromycin antibiotic resistance gene, were not present in 356043 soybean.
Pioneer studied the inheritance of these traits in five generations (T1, F2, F3, BC1F2, and C2F2). From the Chi-square analysis of the trait inheritance data, Pioneer concluded that the inheritance of the integrated insert with the gat4601 and the gm-hra genes is consistent with the Mendelian inheritance pattern of a single locus.
Pioneer assessed the stability of the insert in three generations using Southern analyses. The single Bgl II site present in the PHP20163A fragment is expected to generate a unique event-specific DNA hybridization pattern. Pioneer did not detect changes to the hybridization pattern among the tested generations of the 356043 soybean (the selfed T4 and T5 generations and the traditionally-bred F3 generation). Given the absence of pattern changes, Pioneer concluded that the insertion of the PHP20163A DNA fragment in 356043 soybean is stable.
4. Introduced Proteins
4.1 Identity and Function
Pioneer noted that the 356043 soybean was genetically engineered to express two proteins, the glyphosate N-acetyltransferase GAT4601 protein and the acetolactate synthase GM-HRA protein, that render the transgenic plant tolerant to two classes of herbicides.
The gat4601 gene is derived from the sequences of three gat genes from the common soil bacterium B. licheniformis, a gram positive saprophytic bacterium that is widespread in nature. The GAT4601 protein is 84% homologous to each of the three native GAT proteins from which it was derived. The GAT4601 protein is 146 amino acids in length and has an approximate molecular mass of 17 kDa. There are 23 or 24 amino acid changes (15 or 16 of which are conservative) in the GAT4601 protein, depending on which of the three original native B. licheniformis GAT proteins is used for comparison. The GAT4601 protein acetylates (and inactivates) glyphosate more efficiently than the native B. licheniformis enzymes.
The GM-HRA protein is a modified version of the native soybean acetolactate synthase protein GM-ALS I. The native ALS protein catalyzes the first common step in the biosynthesis of the essential branched-chain amino acids isoleucine, leucine and valine. ALS-inhibiting herbicides function by inhibiting branched-chain amino acid biosynthesis. The mature, modified GM-HRA protein is 604 amino acids in length with a predicted molecular mass of 65 kDa. It is tolerant to ALS-inhibiting herbicides and differs from GM-ALS I protein by two amino acids, alanine substituted for proline and leucine substituted for tryptophan. The locations of these two substitutions are identical to the locations of commonly found natural tolerance mutations reported in scientific literature.
4.2 Expression Levels and Protein Characterization
Pioneer measured the levels of the GAT4601 and GM-HRA proteins in the grain, forage, and root tissue samples collected from plants grown at six field locations in North America. Three replicated samples per tissue per location were collected for 356043 soybean and one sample per tissue per location was collected for control soybean. Pioneer reported the mean protein levels across six locations in 356043 soybean forage, root, and grain. The mean levels for the GAT4601 protein in forage, root, and grain were 1.6, 1.6, and 0.24 nanograms per milligram (ng/mg) of tissue (dry weight), respectively. The mean protein levels for the GM-HRA protein in forage, root, and grain were 27, 3.2, and 0.91 ng/mg of tissue (dry weight), respectively. Neither the GAT4601 nor the GM-HRA proteins were detected in nontransgenic Jack cultivar (control soybean) tissues sampled from the six locations. Pioneer concluded that the above results confirm that expression of the GAT4601 and GM-HRA proteins in 356043 soybean is constitutive.
Pioneer also measured the levels of GAT4601 and GM-HRA proteins in toasted soybean meal and hulls and reported that the levels of both proteins were found to be below the limit of quantitation (LOQ).
Pioneer stated that due to (1) low transgenic protein expression levels in planta and (2) difficulties involved in isolation and purification of the transgenic proteins from plant tissues, the GAT4601 and GM-HRA proteins were produced and purified from a heterologous Escherichia coli (E. coli) protein expression system. To confirm the identity and equivalency of the E. coli-produced and soybean-produced GAT4601 and GM-HRA proteins, Pioneer used the following physiochemical methods:
- Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (to confirm purity and molecular weight)
- Western blot (to confirm immunoreactivity and molecular weight)
- Glycoprotein staining (to determine glycosylation)
- Matrix assisted laser desorption ionization mass spectroscopy (MALDI-MS) (to confirm identity)
- N-terminal sequencing (to confirm identity)
Based on these studies, Pioneer concluded that the E. coli-produced and soybean-produced GAT4601 and GM-HRA proteins were equivalent. The E. coli-produced GAT4601 and GM-HRA proteins were subsequently used for in vitro and in vivo biochemical and toxicological studies.
5. Safety Assessment of Introduced Proteins
To assess potential allergenicity of the GAT4601 protein,2 Pioneer used a weight of evidence approach:
- when the amino acid sequence of the GAT4601 protein is compared to the amino acid sequences of known allergens in the FARRP6 database (University of Nebraska Allergen Database, Version 6, January 2006; www.allergenonline.com) using the FASTA34 sequence alignment program, the identified alignments did not exceed the 35% threshold within the 80 amino acid windows and that there were no eight or greater contiguous identical amino acid matches between the GAT4601 and known allergens;
- when subjected to simulated gastric fluid (SGF), the GAT4601 protein is rapidly (< 30 seconds) hydrolyzed to fragments of < 3 kDa; and when subjected to simulated intestinal fluid (SIF), the GAT4601 protein, including the low molecular weight fragments seen in SGF, is completely and rapidly (< 2 minutes) hydrolyzed;
- the GAT4601 protein is not glycosylated;
- the DNA donor organism, B. licheniformis, has a history of safe use in the food industry; and
- there have been some reports of specific detergent enzymes (alpha-amylases and serine proteases) derived from Bacillus species, including B. licheniformis, producing occupational asthma in detergent industry workers. However, the GAT4601 protein is structurally and functionally distinct from enzymes derived from B. licheniformis reported to cause occupational asthma.
Pioneer concluded that the GAT4601 protein is not likely to cause allergic reactions.
To assess potential toxicity, Pioneer conducted an amino acid homology search of the GAT4601 amino acid sequence against the sequences of known protein toxins in the National Center for Biotechnology Information (NCBI) protein database containing all entries from GenBank nucleotide translations and protein sequences from SWISS-PROT, PIR, PRF, and PDB databases. Pioneer concluded that the GAT4601 protein did not share relevant sequence similarities with known protein toxins and was therefore unlikely to be a toxin itself.
Pioneer also conducted an acute oral toxicity study in mice. A single dose of 1680 milligrams per kilogram of bodyweight (mg/kg bw) of E. coli-produced and purified GAT4601 protein was administered by oral gavage to five male and five female mice. No clinical symptoms of toxicity, body weight loss, gross organ lesions or mortality were observed. Pioneer concluded that the results of this study show that the GAT4601 protein is not acutely toxic.
Pioneer also noted that the GAT4601 protein retains the characteristics found in other N-acetyltransferases that are ubiquitous in plants and microorganisms. Although GAT4601 is a synthetic protein, it is 84% identical and 94-95% similar at the amino acid level to the translated protein sequences of each of the three original B. licheniformis gat alleles from which gat4601 was derived. Pioneer reported that due to its ubiquitous presence as spores in soil and dust, B. licheniformis is widely known as a contaminant of food but is not associated with any adverse effects.
Pioneer concluded that the GAT4601 protein is not toxic.
5.2 GM-HRA Protein
To assess potential allergenicity of the GM-HRA protein, Pioneer used a weight of evidence approach:
- when the amino acid sequence of the GM-HRA protein is compared to the amino acid sequences of known allergens in the FARRP6 database (University of Nebraska Allergen Database, version 6, January 2006; www.allergenonline.com) using the FASTA34 sequence alignment program, the identified alignments did not exceed the 35% threshold within the 80 amino acid windows and that there were no 8 or greater contiguous identical amino acid matches between the GM-HRA and known allergens;
- when subjected to SGF, the GM-HRA protein is rapidly (< 30 seconds) hydrolyzed in SGF to fragments of < 3 kDa; and when subjected to SIF, the GM-HRA protein, including the low weight molecular fragments seen in SGF, is completely and rapidly (< 2 minutes) hydrolyzed;
- the GM-HRA protein is not glycosylated; and,
- although soybean is one of the eight most commonly allergenic foods, none of the identified allergens is a member of the ALS family, and the ALS protein from soybean has not been identified as a soy allergen.
Pioneer concluded that the GM-HRA protein is not likely to cause allergic reactions.
To assess potential toxicity, Pioneer conducted an amino acid homology search of the GM-HRA amino acid sequence, using the BLASTP algorithm, against the sequences in the NCBI protein database containing all entries from GenBank nucleotide translations and protein sequences from SWISS-PROT, PIR, PRF, and PDB databases. Pioneer concluded that the GM-HRA protein did not share relevant sequence similarities with known protein toxins and was therefore unlikely to be a toxin itself.
Pioneer also conducted an acute oral toxicity study in mice. A single dose of 582 mg/kg bw of E. coli-produced and purified GM-HRA protein was administered by oral gavage to five male and five female mice. No clinical symptoms of toxicity, body weight loss, gross organ lesions or mortality were observed. Pioneer concluded that the result of this study shows that the GM-HRA protein does not cause acute toxicity.
Pioneer concluded that the GM-HRA protein is not toxic.
6. Food and Feed Use
Pioneer states that 356043 soybean will be grown for the same uses as currently commercialized soybean varieties in the USA. Approximately 97% of soybean meal generated from soybean grain is used for animal feed and approximately 3% is used in pet food. Overall human consumption of soybean products is low with approximately 94% consumed as oil. Soy protein products are also added to a number of meat, dairy, bakery and cereal products as protein extenders.3
7. Compositional Analysis
Pioneer analyzed soybean forage for proximates (protein, fat, and ash), acid detergent fiber (ADF) and neutral detergent fiber (NDF). Pioneer analyzed soybean seeds for proximates, fiber, fatty acids, total amino acids, acetylated amino acids (N-acetylaspartate (NAA) and N-acetylglutamate (NAG)), free amino acids, isoflavones, and antinutrients. The components measured in seeds are listed in Table 1.
|Proximates & Fiber||Fatty Acids1||Total Amino Acids and Acetylated Amino Acids||Free Amino Acids and Other Amino Compounds2||Isoflavones 3||Anti-nutrients|
|1 The levels of the following fatty acids were below the limit of quantitation: caprylic (8:0), capric (10:0), lauric (12:0), myristoleic (14:1), pentadecanoic (15:0), pentadecenoic (15:1), γ-linolenic (18:3), eicosatrienoic (20:3), arachidonic (20:4), and erucic (22:1).
2 The amino acids taurine, hydroxyl-L-proline, cysteine, and L-cystine were not detected.
3 The levels of the isoflavones acetylgenistin, acetyldaidzin, glycitein and acetylglyitin were below the limit of quantitation.
7.1 Testing Strategy
Pioneer analyzed the composition of the forage and seed tissues collected from the T-5 generation of transgenic 356043 soybean and the non-transgenic parental isoline Jack, which served as a control line. Both lines were grown in 2005 at six field locations in soybean-growing areas in the United States (U.S.) and Canada. In a separate experiment, forage and seed tissue were collected from four different conventional (nontransgenic) commercial soybean lines grown in 2005 at six field locations. In both experiments, soybean plants were grown using a randomized complete block design of two-row plots with three replicates at each location.
Pioneer used compositional data derived for the four conventional reference lines to calculate tolerance intervals that contain 99% of the values obtained for each component with 95% confidence. Pioneer states that the compositional analysis of the reference varieties helps to establish the normal variation in the levels of the measured components. Pioneer subsequently compared the compositional data obtained for 356043 soybean to the 99% tolerance interval and to the combined range of values for each analyte available from published literature and on-line databases.4
Pioneer performed the statistical analysis on compositional data obtained for 356043 soybean and control soybean. Pioneer used a linear mixed model to account for the design effects of location and blocks within location. In order to manage the false discovery rate, Pioneer employed the false discovery rate (FDR) approach introduced by Benjamini and Hohberg (Benjamini and Hohberg, 1995; Westfall et al., 1999). A statistically significant difference between the mean level of each component in 356043 soybean and control soybean was established at the FDR-adjusted P-value < 0.05. Pioneer reported analytical data for each measured component as mean values from all locations. Pioneer also provided ranges, P-values, FDR-adjusted P-values, tolerance intervals, and available combined literature ranges.
7.2 Forage Analysis
Pioneer assessed the composition of soybean forage by measuring proximates (protein, fat, and ash) and fiber (ADF and NDF). No statistically significant differences between 356043 soybean and control soybean were observed in the mean levels of proximates and fiber when analyzed using either the conventional statistical approach or the FDR approach. All mean levels were also within the 99% tolerance intervals and combined literature ranges.
7.3 Seed Analysis
Pioneer assessed the composition of soybean seeds by measuring components listed in Table 1. The results reported by Pioneer are summarized below.
7.3.1 Proximates and Fiber
Pioneer measured the levels of proximates, ADF and NDF in soybean seeds. No statistically significant differences between 356043 soybean and control soybean were observed in the mean levels of protein, ash, and ADF. Based on the conventional statistical approach, statistically significant differences were observed in the mean levels of fat and NDF between the 356043 soybean and the control soybean. However, based on the FDR approach, these differences were not statistically significant. All mean levels were within the 99% tolerance intervals and combined literature ranges.
7.3.2 Fatty Acids
Pioneer measured the levels of fatty acids in soybean seeds (Table 1). No statistically significant differences between 356043 soybean and control soybean were observed in the mean levels of palmitoleic acid, stearic acid, arachidic acid, eicosenoic acid and behenic acid. Statistically significant differences between 356043 soybean and control soybean were observed in the mean levels of seven fatty acids, based on analysis by both the conventional and FDR statistical approaches. The mean levels were observed to be lower in 356043 soybean than in control soybean for myristic acid, palmitic acid, and linoleic acid. The mean levels were observed to be higher in 356043 soybean than in control soybean for oleic acid, heptadecanoic acid, heptadecenoic acid, and linolenic acid. However, these mean levels, with the exception of the mean levels of two minor fatty acids, heptadecanoic (C17:0) and heptadecenoic (C17:1), were within the 99% tolerance intervals and combined literature ranges. The mean levels of the two 17-carbon fatty acids in 356043 soybean were above the 99% tolerance intervals and the combined literature values.
Pioneer assessed the biological significance of increased levels of C17:0 and C17:1 in the human diet. Pioneer stated that C17:0 is found in commonly consumed foods. C17:0 is found in cooked lamb, butter, tofu (prepared with nigari), and pre-cooked beef and sausage. C17:1 is found in tofu (prepared with nigari), raw and pan-broiled ground beef, and olive oil. Pioneer conducted an exposure assessment for humans for these two fatty acids to compare the estimates of dietary intake of C17:0 and C17:1 with and without exposure from 356043 soybean and estimated that commercialization of 356043 soybean, namely as processed oil, may increase the dietary exposure to C17:0 and C17:1 in the U.S. population slightly above current levels.5 Pioneer also described odd chain fatty acid metabolism and concluded that C17:0 and C17:1 should be readily metabolized. On the basis of its comparison of dietary intake with and without exposure from 356043 soybean and its review of odd chain fatty acid metabolism, Pioneer concluded that no safety issues are expected as a result of the estimated increase in the human exposure to C17:0 and C17:1.
Pioneer considered whether commercialization of 356043 soybean would result in increased levels of C17:0 and C17:1 in the animal diet. Pioneer stated that fatty acids, including C17:0 and C17:1, are found in the oil fraction of processed soybeans and that, although soybean oil may be used in animal feeds, the fat/oil added to animal feeds typically comes from sources other than soybean oil. Rather, the primary soybean-derived product used in animal feeds is defatted, toasted soybean meal, which contains negligible amounts of fatty acids in general, and C17:0 and C17:1 in particular. On this basis, Pioneer concluded that no increase in animal exposure to C17:0 and C17:1 is expected with the commercialization of 356043 soybean.
7.3.3 Amino Acids
Pioneer measured the levels of total amino acids, two acetylated amino acids (NAA and NAG), and free amino acids in soybean seeds (Table 1).
Total amino acids:
Pioneer noted that the amino acids asparagine and glutamine were converted to aspartic and glutamic acids, respectively, in the course of the amino acid analysis. Consequently, the levels of asparagine and glutamine are included in those of aspartic acid (aspartate) and glutamic acid (glutamate). No statistically significant differences between 356043 soybean and control soybean were observed in the mean levels of amino acids, with the exception of the mean levels of arginine and alanine. Based on the conventional statistical approach, the mean levels of arginine and alanine are both statistically significantly higher in 356043 soybean than in control soybean. However, based on the FDR approach, these differences are not statistically significant. Pioneer noted that the mean levels of all amino acids were within the 99% tolerance interval and/or combined literature ranges with the exception of lysine in 356043 soybean. The level of lysine in 356043 soybean was above the 99% tolerance interval and the range reported in the literature. The level of lysine in the control soybean was also above the combined literature range. Pioneer concluded that any noted differences in the levels of amino acids were biologically insignificant.
Acetylated amino acids:
According to a recent publication cited by Pioneer, native GAT enzymes from B. licheniformis are able to acetylate several amino acids under in vitro conditions. Because the GAT4601 protein is derived from B. licheniformis GAT enzymes, Pioneer tested the affinity of GAT4601 protein toward amino acid substrates. Of twenty-one amino acids surveyed by Pioneer, only L-aspartate, L-glutamate, L-serine, glycine, and L-threonine indicated low but measurable affinity to GAT4601. Pioneer was able to calculate the catalytic efficiency of GAT4601 on L-glutamate and L-aspartate, which was approximately 3% of that on glyphosate; however, Pioneer states that the affinity of GAT4601 towards L-serine, glycine, and L-threonine was so low that catalytic efficiencies could not be calculated.
Pioneer measured N-acetylaspartate (NAA) and N-acetylglutamate (NAG) in soybean seed and found that the mean levels of these amino acids were statistically significantly higher in the 356043 soybean than in the control soybeans and were also above the upper limits of tolerance intervals derived for commercial soybean varieties. The mean levels of NAA and NAG in 356043 soybean were 0.580 mg/g dry weight (dw) and 0.0116 mg/g dw, respectively. For comparison, the mean levels of NAA and NAG in the control soybeans were 0.00252 mg/g dw and 0.00153 mg/g dw, respectively. Pioneer calculated that NAA and NAG make up 0.14% of the total amino acids in 356043 soybean.
Pioneer assessed the biological significance of the increased levels of NAA and NAG in the human diet. Pioneer analyzed the levels of NAA and NAG in several different foods that contain high concentrations of aspartic acid and glutamic acid.6 NAA was found to be present in a variety of foods, including autolyzed yeast, chicken bouillon (vegan), eggs, and ground turkey, chicken and beef; NAG was found to be present in autolyzed yeast, dried egg powder, and ground turkey and beef. Based on this information, Pioneer concluded that NAA and NAG are components of common foods in the human diet. Pioneer conducted an exposure assessment for humans for these two acetylated amino acids to compare the estimates of dietary intake of NAA and NAG with and without exposure from 356043 soybean. Pioneer estimated that commercialization of 356043 soybean may increase the dietary exposure to NAA and NAG in the U.S. population above current levels of exposure.7 However, deacetylase enzymes that deacetylate acetylated amino acids are widely distributed throughout the human body. For example, aspartoacylase which is present in multiple tissues has been shown to deacetylate NAA. The wide distribution of deacetylases in the human body led Pioneer to conclude that NAA and NAG will be deacetylated as a routine part of digestion or metabolism and that the estimated increase in exposure to NAA and NAG do not pose a food safety concern.
Pioneer considered the increases in dietary exposure to NAA from commercialization of 356043 Soybean on individuals with Canavan Disease (CD). Individuals with CD do not express a functioning aspartoacylase protein and are unable to deacetylate NAA. However, in CD the principal source of NAA is endogenous production in the brain. The exposure to NAA from consumption of 356043 soybean is negligible compared to the amount of NAA produced in the brain. Consequently, Pioneer concludes that there will be no adverse impact to individuals with CD from the potential increase in dietary NAA as a result of consuming 356043 soybeans, and that 356043 soybeans do not pose a safety concern for persons with CD.
Pioneer assessed the biological significance of increased levels of NAA and NAG in the poultry diet since poultry consume about 50% of the soybean meal consumed in the U.S. Pioneer cites the results of a 42-day broiler chicken feeding study. Pioneer calculated exposure to NAA and NAG from diets containing processed fractions (meal, hulls, and oil) derived from 356043 soybean. The average dietary exposure to NAA was 24.4 mg/kg bw and to NAG was 1.3 mg/kg bw. No statistically significant differences were noted in mortality, weight gain, feed efficiency and organ and carcass yield variables between broilers consuming diets produced with 356043 soybeans and those consuming diets produced with the control soybeans. All these variables also fell within the tolerance intervals derived from broilers fed diets containing the nontransgenic commercial soybean reference varieties. Pioneer concluded that no safety issues are expected as a result of the estimated increase in exposure to NAA and NAG in the poultry diet.
Pioneer also considered the effect of the increased levels of NAA and NAG in swine diets where about 26% of the consumed soybean meal is used. Pioneer calculated that the level of exposure to NAA and NAG would be much lower in swine diets than that calculated for poultry diets where no effects were seen.
Pioneer further concludes that, as dairy and beef cattle diets contain lower amounts of soybean fractions and as acetylated proteins are readily metabolized and have been safely consumed by animals, no safety issues are expected as a result of the estimated increase in the animal exposure to NAA and NAG.
Free amino acids:
Pioneer states that more than 99% of the total amino acids in soybeans are incorporated into proteins. The remaining fraction (<1%) occurs as free amino acids. The free amino acid pool in soybeans includes L-isomers of amino acids that comprise proteins as well as several amino acids that are not normally incorporated into proteins. To clarify whether low levels of acetylation of aspartate and glutamate affected the composition of the free amino acid pool in 356043 soybean, Pioneer compared individual free amino acids levels in 356043 soybean and control soybean. L-cystine, L-cysteine, hydroxyl-L-proline, and taurine were not detected. Two non-amino acid compounds (ethanolamine and ammonia) were measured along with the amino acids. There were no statistically significant differences in the mean levels of the measured amino acids between 356043 soybean and control soybean. Furthermore, all mean levels were within the 99% tolerance intervals. Pioneer concluded that the free amino acid pools in 356043 soybean and the control soybean were comparable.
Pioneer stated that there are three basic types of isoflavones in soybeans: daidzein, genistein, and glycitein. Each of these forms (known as free forms or aglucones) can also exist in three conjugate forms: glucoside, acetylglucoside, and malonylglucoside. Consequently, twelve forms of isoflavones can occur in soybeans. Pioneer reported that the levels of acetylgenistin, acetyldaidzin, glycitein, and acetylglycitin were below the limit of quantitation in all soybean samples. Based on the conventional statistical approach, the mean levels of daidzin, malonyldaidzin, glycitin, and malonylglycitin were statistically significantly higher in 356043 soybean than in control soybean; however, based on the FDR approach, only the level of malonyldaidzin was statistically significantly higher in 356043 soybean than in control soybean. Finally, although the mean levels of several of these isoflavones were outside the combined literature ranges, all were within the 99% tolerance intervals.
Antinutrients that occur in soybean seeds include non-digestible carbohydrates stachyose and raffinose, lectins, phytic acid, and trypsin inhibitor (Table 1). Pioneer reported that the mean levels of these antinutrients, with the exception of trypsin inhibitor, were comparable in 356043 soybean and control soybean. The mean level of trypsin inhibitor was statistically significantly lower in 356043 soybean than in control soybean. However, the mean levels of the antinutrients measured, including trypsin inhibitor, were within the 99% tolerance intervals and combined literature ranges.
7.3.6 Endogenous Allergens
To assess whether the transformation process may have increased the overall allergenicity of 356043 soybean compared to conventional soybean, Pioneer conducted both one dimensional IgE immunoblot analysis (SDS-PAGE) and IgE-binding enzyme linked immunosorbent assay (ELISA) inhibition analysis using protein extracts from 356043 soybean and control soybean. The sera used in these assays were obtained from ten clinically reactive soy-allergic patients; the sera were pooled, but not diluted, for the assays. Pioneer states that the protein profiles obtained from SDS-PAGE for 356043 soybean and control soybean extracts appeared the same. Pioneer states that the ELISA inhibition data showed the same IgE binding profiles between 356043 soybean and control soybean protein extracts. From these experiments, Pioneer concluded that the allergic potential of the 356043 soybean is no greater than that of the control soybean.
Pioneer has concluded that its dual herbicide tolerant soybean variety, 356043 soybean, and the foods and feeds derived from it are not materially different in safety, composition, or any other relevant parameter from soybeans now grown, marketed, and consumed. At this time, based on Pioneer's data and information, the agency considers Pioneer's consultation on 356043 soybean to be complete.
Carrie Hendrickson, Ph.D.
(1)In its submission, Pioneer uses the terms GM-ALS and GM-ALS I interchangeably when referring to the soybean endogenous acetolactate synthase protein. For the purpose of consistency, this memorandum refers to the protein using the term GM-ALS I.
(2)The GAT4601 protein is the subject of a New Protein Consultation, NPC 00003. NPC 00003 is incorporated by reference into BNF 108.
(3)OECD, 2001, ENV/JM/MONO(2001)15, p1-30.
(4)Pioneer's combined literature ranges are derived from the following sources: OECD, 2001; ILSI, 2006; Taylor et al., 1995; and Kim et al., 2005. In an email dated June 5, 2007, Pioneer confirmed that the version of the ILSI database used for the combined literature ranges is Version 3.0 (released in April 10, 2006) that contains additional data on soybean composition not contained in Versions 1.0 or 2.0.
(5)Pioneer used the Dietary Exposure Evaluation Model (DEEM) and the Food Commodity Intake Database (FCID) (Exponent, Inc., Washington, D.C.) to estimate the amount of C17:0 and C17:1 fatty acids consumed in the human diet with (356043 soybean scenario) and without (baseline scenario) exposure from 356043 soybean. Pioneer estimated exposure to C17:0 and C17:1 fatty acids for the U.S. population and one population subgroup (children of 1-2 years of age). In additional information submitted by Pioneer on August 27, 2007, Pioneer assumed for its 356043 soybean scenario, that 100 percent of the commodity soybeans grown in the U.S. would be 356043 soybean. Pioneer's estimated mean and 90th percentile exposure to C17:0 for the US population was 4.0 milligrams/kilogram of bodyweight per day (mg/kg bw/d) and 8.0 mg/kg bw/d for the baseline scenario compared to 4.7 mg/kg bw/d and 9.6 mg/kg bw/d for the 356043 soybean scenario. Pioneer's estimated mean and 90th percentile exposure to C17:1 for the US population was 1.2 mg/kg bw/d and 2.4 mg/kg bw/d for the baseline scenario compared to 1.7 mg/kg bw/d and 3.5 mg/kg bw/d for the 356043 soybean scenario.
(6) Pioneer selected high aspartic acid and glutamic acid foods using the United States Department of Agriculture Nutrient Database for Standard Reference (Release 19; 2006).
(7) Pioneer used DEEM in conjunction with FCID (Exponent, Inc., Washington, D.C.) to estimate the amount of NAA and NAG consumed in the human diet with (356043 soybean scenario) and without (baseline scenario) exposure from 356043 soybean. Pioneer estimated exposure to NAA and NAG for the U.S. population as well as several populations subgroups, including children of 1-2 years of age. In additional information submitted by Pioneer on August 27, 2007, Pioneer assumed for its 356043 soybean scenario that 100 percent of the commodity soybeans grown in the US would be 356043 soybean. Pioneer's estimated mean and 90th percentile exposure to NAA for the US population was 0.0026 mg/kg bw/d and 0.0058 mg/kg bw/d for the baseline scenario compared to 0.0114 mg/kg bw/d and 0.0214 mg/kg bw/d for the 356043 soybean scenario. Pioneer's estimated mean and 90th percentile exposure to NAG for the US population was 0.0019 mg/kg bw/d and 0.0043 mg/kg bw/d for the baseline scenario compared to 0.0047 mg/kg bw/d and 0.0092 mg/kg bw/d for the 356043 soybean scenario.