Return to inventory: Completed Consultations on Foods from Genetically Engineered Plant Varieties
Date: January 14, 2009
Subject: High Oleic 305423 Soybean
Keywords: soybean, Glycine max, high oleic 305423 soybean, TREUSTM, OECD unique identifier DP-3Ø5423-1, omega-6 desaturase, FAD2-1, gm-fad2-1, seed-specific silencing, gm-hra, soybean acetolactate synthase
In a submission dated December 27, 2006, Pioneer Hi-Bred International, Inc. (Pioneer), a DuPont Company, submitted to the Food and Drug Administration (FDA) a safety and nutritional assessment of the bioengineered high oleic 305423 soybean line containing the transformation event DP-3Ø5423-1. Pioneer provided additional information on August 31, 2007, January 30, February 19, March 18, November 21, December 12, and December 19, 2008. Pioneer concluded that food and feed derived from the 305423 soybean are as safe and nutritious as food and feed derived from conventional soybean varieties currently being marketed.
2. Intended Effect
The intended effect of the modification in 305423 soybean is to produce soybean seeds with increased levels of monounsaturated fatty acid (oleic) and decreased levels of polyunsaturated fatty acids (linoleic and linolenic). To accomplish this objective, Pioneer inserted a fragment of the soybean microsomal omega-6 desaturase gene (FAD2-1) into soybean. The fragment of the FAD2-1 gene does not code for a protein. Transcription of the gene fragment under the control of a seed-preferred KTi3 promoter acts to silence the expression of the endogenous soybean omega-6 desaturase, which results in an increased level of oleic acid and decreased levels of linoleic and linolenic acids in the soybean seed. A gene (gm-hra) encoding a modified version of the soybean acetolactate synthase (als gene) that confers tolerance to a sulfonylurea herbicide was used as a selectable marker for the transformation.
3. Genetic Modifications and Characterization
3.1. Parental Variety
The publicly available soybean variety "Jack" was used as the recipient in the DNA transformation to create 305423 soybean.
3.2. Transformation DNA and Method
Pioneer used microprojectile bombardment to co-transform secondary plant cell embryos with two purified linear DNA fragments: a 2924 base pair fragment containing the gm-fad2-1 cassette, and the 4512 base pair fragment containing the gm-hra cassette. The gm-fad2-1 cassette includes the promoter region from the soybean Kunitz trypsin inhibitor gene (KTi3), a fragment of the FAD2-1 gene that corresponds to approximately 40% of the middle portion of the FAD2-1 gene, and the 3' untranslated region of the KTi3 gene (KTi3 terminator).
The gm-hra cassette includes a promoter and an intron from the 5'regulatory region of the S-adenosyl-L-methionine synthetase (SAMS) gene from soybean, the gm-hra gene that encodes the GM-HRA protein, and the terminator from the endogenous als gene.(1)
3.3. Characterization, Stability, and Inheritance of the Introduced DNA
In order to characterize the introduced DNA, Pioneer conducted Southern blot and sequencing analyses of the DNA inserted into the 305423 soybean. Pioneer reports that the 305423 soybean contains four genetically linked insertions. The 305423 soybean has multiple intact and partial copies of the gm-fad2-1 cassette that contain, in total, eight copies of the KTi3 promoter, seven copies of the gm-fad2-1 gene fragment, and five copies of the KTi3 terminator. Pioneer states that it appears that multiple copies of the gm-fad2-1 gene fragment are necessary for effective co-suppression of the endogenous gene. One copy of the KTi3 promoter is associated with a small non-functional fragment of plasmid backbone DNA. Pioneer determined that a single intact gm-hra cassette is inserted into the genome of 305423 soybean. Pioneer noted that further analysis of 305423 soybean for plasmid backbone sequence using Southern hybridization showed that 305423 soybean contains neither the hygromycin gene nor the bacterial plasmid origin of replication present on the plasmids from which the gm-fad2-1 cassette and the gm-hra cassette are derived. Pioneer sequenced the inserted DNA and reported that the sequence confirmed the results of the Southern analysis.
Each of the four insertions in 305423 soybean was screened for the presence of open reading frames (ORFs) containing both a start and stop codon that spanned any novel junctions. Pioneer identified two such ORFs. Pioneer reports that neither ORF contains the necessary regulatory elements for transcription. Northern blot analysis detected no transcripts of the ORFs in developing seeds from either 305423 soybean or the control soybean. Pioneer concluded that it is very unlikely that a protein is expressed from either ORF. Pioneer also reports that screening of the ORFs against a database containing known protein toxins(2) and to a database of known allergens(3) showed no biologically significant identities to known protein toxins or allergens. Based upon their analyses, Pioneer concluded that there are no safety concerns resulting from these ORFs.
Pioneer reports that Southern blot analyses across three generations showed that the inserted DNA in 305423 soybean is stably integrated into the genome. The same event-specific hybridization pattern was observed for all but one plant that apparently lost the gm-hra cassette due to a recombination event. Pioneer investigated the frequency of recombination in 305423 soybean by examination of 1000 additional segregating 305423 soybean plants by PCR-based assays and found no other recombinants.
Pioneer conducted chi square analysis of trait inheritance data. They report that the expected segregation ratios were observed in crosses showing the Mendelian inheritance and stability of the introduced trait. Pioneer reports that when both traits were analyzed in the same plants, data confirmed co-segregation of the gm-fad2-1 gene fragment and the gm-hra gene.
4. Introduced Protein – GM-HRA
4.1. Identity, Function, and Characterization
Pioneer describes the GM-HRA protein as a modified version of the endogenous soybean acetolactate synthase (ALS) enzyme and that this modified ALS enzyme confers tolerance to ALS-inhibiting herbicides. ALS enzymes are widely distributed in nature and have been isolated from bacteria, fungi, algae, and plants. ALS-inhibiting herbicides inhibit plant growth by blocking the action of the ALS enzyme thereby inhibiting branched-chain amino acid biosynthesis. The mature GM-HRA protein differs from the endogenous ALS enzyme at two specific amino acids and is responsible for GM-HRA insensitivity to ALS-inhibiting herbicides. The gm-hra gene was used only as a selectable marker for the transformation.
Pioneer characterized 305423 soybean-produced GM-HRA protein using various methodologies(4) and demonstrated its equivalence with Escherichia coli-produced GM-HRA protein. The E. coli-derived GM-HRA protein was used for in vitro and in vivo biochemical and toxicological studies.
The GM-HRA protein levels in 305423 soybean were measured in replicated samples of leaf, root, forage and grain tissues using a quantitative enzyme-linked immunosorbent assay (ELISA). Pioneer reports that the mean GM-HRA protein concentrations in 305423 soybean leaf, root, forage, and grain were 4.0, 0.18, 5.7, and 2.5 nanograms per milligram (ng/mg) tissue dry weight respectively. Pioneer concluded that the above results confirm that the expression of the GM-HRA protein in 305423 soybean is constitutive.
4.2. Assessment of Potential Toxicity and Allergenicity
Pioneer reports the results of a global sequence similarity search of the GM-HRA amino acid sequence against the National Center for Biotechnology Information (NCBI) Protein dataset. The search was conducted using the BLASTP algorithm.(5) GM-HRA showed sequence similarity to other ALS proteins. None of the proteins returned by the search was identified as a toxin. Pioneer concluded that the GM-HRA protein did not share relevant sequence similarities with known protein toxins.
Pioneer conducted an acute oral toxicity study in mice. A single dose of 582 mg per kilogram of body weight (mg/kg bw) of 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 reports that while soybean is one of the major food allergens, none of the identified allergens is a member of the ALS family and the ALS protein from soybean has not been characterized as a soy allergen. Pioneer compared the amino acid sequence of the GM-HRA protein to the amino acid sequences of known allergens in the Food Allergy Research and Resource Program (FARRP, version 6.0) database using the FASTA34 sequence alignment program.(6) None of the identified alignments met or exceeded the threshold of greater than or equal to 35% identity over 80 amino acids and no contiguous stretches of 8 or greater amino acids were shared between the GM-HRA protein and the proteins in the allergen database. Additionally, the GM-HRA protein is not glycosylated.
Pioneer also reports that the GM-HRA protein is rapidly (< 30 seconds) hydrolyzed in simulated gastric fluid (SGF) to fragments of < 3 kDa; and when subjected to simulated intestinal fluid (SIF), the GM-HRA protein, including the low weight molecular fragments seen in SGF, is completely and rapidly (< 2 minutes) hydrolyzed.
Pioneer concluded that the GM-HRA protein is unlikely to be a toxin and is not a potential allergen.
5. Evidence for Silencing of FAD2-1 Gene Expression
Pioneer determined that a 597 base pair gm-fad2-1 gene fragment, which is identical to a portion of the coding sequence of the endogenous soybean microsomal omega-6 desaturase gene (FAD2-1), resulted in the down regulation of expression of the endogenous FAD2-1 gene. Pioneer examined the expression of the endogenous FAD2-1 gene in the leaf and developing seed of ten 305423 soybean plants and five control Jack soybean plants. Northern blots indicated that transcripts of the FAD2-1 gene in developing seeds of 305423 soybeans were greatly diminished when compared to the control soybean. This greatly decreased level of transcription confirms that the endogenous FAD2-1 gene is effectively silenced.
6. Food and Feed Uses of Soybean
Pioneer describes historical and current uses of soybean varieties in food and animal feed. The primary use of commodity soybeans is for soybean meal that is consumed by animals. Raw soybeans contain several antinutritional factors (trypsin inhibitors, urease, and hemagglutinins). Heat treatment is the most common processing method used to minimize the activity of such factors. Soybean oil is the major soybean fraction consumed by humans. Soybean oil accounts for 80% of total United States consumption of oils and fats.
Pioneer states that the 305423 soybean variety is intended to be used for the production of high oleic soybean oil. The oil is intended to be a highly stable vegetable oil suitable for frying applications without the need for hydrogenation which produces trans fatty acids and "hydrogenated flavor." Pioneer states that it is aware of no food or feed uses of current soybean varieties for which the 305423 soybean variety would not be equally suitable.
7. Overview of Compositional Analysis
Pioneer assessed the composition of grain and forage from the 305423 soybean and a null segregant (non-transgenic isoline) control. Pioneer states that the null segregant plants are an appropriate control because they are almost genetically identical to the corresponding 305423 soybean plants with the exception that they do not carry the transgenic DNA. Both the transgenic and control soybeans were grown at six field locations in soybean-growing areas of North America using a randomized complete block design with three replicates at each location. Pioneer also analyzed grain and forage from four different commercial soybean varieties.
Pioneer measured 52 components in grain and 5 in forage. Pioneer used the composition data obtained from the commercial varieties to calculate 99% tolerance intervals with 95% confidence for all measured components. To interpret the composition results for 305423 soybeans, Pioneer used the confidence intervals and established a combined literature range using data from published literature and databases on soybean composition.
Pioneer performed statistical analyses of composition data obtained for 305423 soybean and control soybean using mean values calculated from data aggregated from all tests sites. Pioneer used a linear mixed model analysis of variance (ANOVA). In order to hold the rate of false positive results to 5% or less, Pioneer employed the false discovery rate (FDR) approach (Benjamini and Hochberg (1995) and Westfall et al. (1999)). Pioneer reported the composition data by providing mean values, ranges, FDR-adjusted P-values, unadjusted P-values, tolerance intervals, and literature ranges, as available. Pioneer discussed analytical results in the context of FDR-adjusted P-values. Unless so indicated, statistical analyses using the unadjusted and adjusted P-values are in agreement. Pioneer used a P-value of 0.05 to denote a statistically significant difference between the control and the 305423 soybean.
Pioneer analyzed grain samples for proximates (protein, fat, and ash), acid detergent fiber (ADF), neutral detergent fiber (NDF), fatty acids, amino acids, isoflavones, and antinutrients. Compositional analysis of forage samples included proximates, ADF, and NDF. Table 1 contains the complete list of all measured components.
|Proximates* & Fiber*||Fatty Acids**||Amino Acids||Isoflavones+||Antinutrients|
acid detergent fiber (ADF)
neutral detergent fiber (NDF)
|* Proximates and fiber were measured in both soybean grain and forage. All other components were measured in grain only.
** The levels of eleven additional fatty acids were near or below the lower limit of quantitation.
+ The levels of acetylgenistin, acetyldaidzin, and acetylglycitin were below the limit of quantitation.
7.1 Compositional Analysis of Soybean Grain
Intended Compositional Change – Fatty Acids
Pioneer analyzed the fatty acid composition of the oil extracted from the grain of the 305423 and control soybeans. Pioneer provides the levels of 14 fatty acids (see Table 1) calculated as percentages of the total fatty acids. Pioneer states that the fatty acid analysis confirmed the expected high oleic acid phenotype as shown by a substantial increase in the mean level of oleic acid to 76.5% in 305423 soybean as compared to a mean level of 21% in the control soybean. As expected, the increase in the level of oleic acid was accompanied by a decrease in the level of linoleic acid. The mean level of linoleic acid in the control soybean was 52.5% and that in the 305423 soybean was 3.6%. The level of linolenic acid also decreased since linolenic acid is formed in soybeans directly from linoleic acid. Pioneer also notes that the levels of two minor fatty acids, heptadecanoic and heptadecenoic, increased in the 305423 soybean to 0.8% and 1.2% of the total fatty acids, respectively. Pioneer explains that the increase in the levels of heptadecanoic and heptadecenoic fatty acids is not unexpected because the GM-HRA enzyme most likely results in the increased concentration of 2-ketobutyrate, which is the substrate in the biosynthesis of hepta fatty acids in soybeans. Pioneer also reported changes in the levels of all remaining fatty acids, which are not biologically significant.
Proximates and Amino Acids
Pioneer reports that the mean levels of protein and fiber in grain from the 305423 and control soybean are not statistically significantly different. While the mean level of fat was statistically significantly lower in the 305423 soybean than in the control using the unadjusted P-value, no statistical difference was detected using the FDR-adjusted P-value. The mean level of ash was statistically significantly lower in the 305423 soybean as compared to the control soybean. Mean levels of the proximates and fiber in grain from 305423 and control soybean lines were within the 99% tolerance intervals for the reference varieties and within the combined literature range.
Pioneer reports that there are no statistically significant differences observed in the mean levels of amino acids between 305423 and control soybean grain with the exception of threonine and glutamic acid. The mean levels of these two amino acids were statistically significantly increased in 305423 soybean when the unadjusted P-values were used, but not when the FDR-adjusted P-values were used. All these levels were within the 99% tolerance intervals.
Pioneer analyzed grain from the 305423 and control soybeans for twelve isoflavones, of which nine were quantified (see Table 1). The mean levels for daidzin, malonylgenistin, and malonyldaidzin were statistically significantly increased in the 305423 versus the control soybean. Mean genestin levels were only shown to be statistically significantly increased when the unadjusted P-value was used. Mean values for all the measured isoflavones were within the 99% tolerance intervals.
In both the control and 305423 soybean the mean levels of genistin, glycitin, malonylgenistin, malonylglycitin, and malonyldaidzin were above the combined literature range. Daidzin levels were only above the combined literature range in the 305423 soybean. No literature data were available for the level of glycitein. The values for the other two isoflavones (genistein and daidzein) were within the combined literature range.
Pioneer measured several antinutrients in soybean grain including non-digestible oligosaccharides stachyose and raffinose, lectins, phytic acid, and trypsin inhibitor (see Table 1). No statistically significant differences were observed between the 305423 and control soybean in mean values for raffinose, lectins, and phytic acid. The mean value for trypsin inhibitor was statistically significantly lower for 305423 soybean as compared to the control soybean. The reduction in the mean value of trypsin inhibitor in 305423 soybean was expected as Pioneer reported that the promoter used to drive expression of the FAD2-1 gene, the Kunitz trypsin inhibitor promoter, has been shown to silence the KTi3 gene which encodes the Kunitz trypsin inhibitor. The mean value for stachyose was statistically significantly increased when the unadjusted P-value was used, but not when the FDR-adjusted P-value was used. Mean values for all the measured antinutrients were within the 99% tolerance intervals and within the combined literature ranges.
7.2 Compositional Analysis of Soybean Forage
Pioneer analyzed forage for protein, fat, ash, ADF, and NDF. Pioneer reports that no statistically significant differences were observed between the mean levels of these components in forage from the 305423 and control soybeans with the exception of the level of fat. The mean level of fat was statistically significantly decreased in 305423 using the unadjusted P-value, but not when using the FDR-adjusted P-value. All mean levels were within the 99% tolerance intervals and combined literature ranges, with the exception of NDF. The mean level of NDF in both the control and 305423 soybean was above the combined literature range.
7.3 Endogenous Allergens
Pioneer conducted a study to assess whether the transformation process may have increased the overall allergenicity of 305423 soybean compared to conventional soybean. Using sera from clinically reactive soy allergic patients, Pioneer conducted IgE immunoblot and ELISA studies using protein extracts from 305423 soybean and conventional soybean. Pioneer reports that the SDS-PAGE Coomassie blue-stained protein profiles for 305423 and control soybean extracts appeared to be the same; they are similar in their IgE binding profile, and showed the same IgE binding capacity for 305423 and control soy extracts. Pioneer concluded that the levels of endogenous allergens in and the allergic potential of 305423 soybean are comparable to those in nontransgenic control soybean.
8. Fatty Acids Intake
8.1 Human Diet
Pioneer generated estimates of dietary exposure to various fatty acids from the consumption of soy oil derived from conventional soybeans, as well as 305423 soybeans. Pioneer concluded that the intake of oleic acid would increase, while the intakes of linoleic acid and trans fatty acid would decrease. Pioneer states that based on very conservative intake estimates calculated on the assumption that high oleic soybean oil would replace all soybean oil in commercial applications, the dietary intake of linoleic acid would still fall within the current intake levels. Pioneer also notes that a variety of other oils used by the food industry would provide significant amounts of linoleic acid in the diet. Other noted changes in dietary intakes of fatty acids would result in small increases in the consumption of minor fatty acids, heptadacanoic acid (C17:0), heptadecenoic acid (C17:1), and the (9, 15) isomer of linoleic acid (cis-9, cis-15-octadecadienoic acid).
Pioneer states that the 17-carbon fatty acids, heptadecanoic and heptadecenoic acids, occur at low levels in commonly consumed foods. For example, heptadecanoic acid is commonly found in meat (lamb, beef, pork), butter, and tofu, while heptadecenoic acid is found in foods such as tofu, beef, cheese, and olive oil. Pioneer stated that odd-chain fatty acids such as 17-carbon fatty acids are readily metabolized.
8.2 Animal Diets
When oil is removed from the soybean a defatted meal is generated that is used as a primary protein supplement for animal feed. Pioneer provided examples of swine and poultry diets to demonstrate that the reduced intake of linoleic acid, an essential fatty acid, would not lead to a nutritional deficit for animals consuming feeds containing meal derived from 305423 soybeans. Corn, the major ingredient of animal diets, is the primary source of linoleic acid in such diets. Even though soybean meal derived from 305423 soybeans would have reduced amounts of linoleic acid, the quantity of linoleic acid provided by corn is several-fold above the animals' requirement.
9. Common or Usual Name of the Oil Product
Pioneer concluded that based on the intended increase in the level of oleic acid and decrease in the levels of linoleic and linolenic acids, a new common or usual name is appropriate for the oil from 305423 soybeans to distinguish this oil from the conventional soybean oil as defined in the Food Chemicals Codex (FCC). Pioneer proposed the name "high oleic soybean oil" for the oil that will be produced from 305423 soybeans.
Pioneer has concluded that, with the exception of the intended change in fatty acid composition, the 305423 soybean and the foods and feeds derived from it are not materially different in composition, safety, 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 the 305423 soybean to be complete.
Mary D. Ditto, Ph.D.
(1)The gm-hra cassette also contains three Flp recombinase target sequences; however, these sequences were not used in the development of the 305423 soybean. The presence of these sites does not cause recombination. Recombination requires the presence of the specific Flp recombinase enzyme that is not present in plants.
(2)National Center for Biotechnology Information (NCBI) Protein Dataset release 156.0. The NCBI dataset incorporates all non-redundant entries from all GenBank nucleotide translations along with protein sequences from SWISS-PROT, PIR, PRF, and PDB databases.
(3)Allergen database derived from the Food Allergy Research and Resource Program (FARRP, version 6.0).
(4)Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), Western blot analysis, glycoprotein staining, mass determination of the tryptic peptides by matrix assisted laser desorption ionization mass spectrometry (MALDI-MS), and N-terminal amino acid sequence analysis.
(5)The E-score was set at 1.0 to ensure that proteins even with limited similarity would not be overlooked.
(6)University of Nebraska Allergen Database, version 6, January 2006; www.allergenonline.com.