Biotechnology Consultation Note to the File BNF No. 000122
Return to inventory: Completed Consultations on Foods from Genetically Engineered Plant Varieties
Biotechnology Consultation - Note to the File
Biotechnology Notification File BNF No. 000122
July 12, 2012
Event FG72, herbicide tolerant soybean
Soybean; Glycine max; modified 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS); 2mepsps, 2mEPSPS, modified p-hydroxyphenylpyruvate dioxygenase (HPPD W336), hppdPfW336, herbicide tolerance, glyphosate, isoxaflutole, IFT, OECD Unique Identifier MST-FGØ72-2
This document summarizes our evaluation of biotechnology notification file (BNF) No. 000122. In a submission dated December 3, 2009, Bayer CropScience (Bayer) and M.S. Technologies LLC (M.S.) submitted to the Food and Drug Administration (FDA) a safety and nutritional assessment of bioengineered herbicide tolerant soybean, transformation event FG72 (hereafter referred to as FG72 soybean). Bayer and M.S. provided additional information on September 13, 2011, and May 3, 2013. FDA evaluated the information in Bayer and M.S.’s submissions to ensure that regulatory and safety issues regarding human food and animal feed from the new plant variety have been resolved prior to commercial distribution.
In our evaluation of BNF 000122, we considered all information provided by Bayer and M.S. as well as publicly available information and information in the agency’s files. Here, we discuss the outcome of the consultation, but do not intend to restate the information provided in the final consultation in its entirety.
The intended technical effects of the modification in FG72 soybean are to confer tolerance to the herbicides glyphosate and p-hydroxyphenylpyruvate dioxygenase (HPPD) inhibitors, such as isoxaflutole. To accomplish this objective, Bayer and M.S. introduced into the soybean genome two genes 2mepsps and hppdPfW336, which encode modified 5-enolpyruvylshikimate 3-phosphate synthase (EPSPS) protein from corn and modified p-hydroxyphenylpyruvate dioxygenase from Pseudomonas fluorescens (designated HPPD W3361), respectively. Incorporating resistance to both glyphosate- and isoxaflutole-containing herbicides provides soybean growers with additional options for weed control.
The purpose of this evaluation is to assess whether the developer has introduced a substance requiring premarket approval as a food additive or use of the new plant variety in food or animal feed raises other regulatory issues with respect to the Federal Food, Drug, and Cosmetic Act (FD&C Act).
The Environmental Protection Agency (EPA) regulates herbicides under the FD&C Act and the Federal Insecticide, Fungicide, and Rodenticide Act. Under EPA regulations, the herbicide residues and metabolic by-products in FG72 soybean, resulting from the detoxification of the applied herbicides by the expression products, are considered pesticide residues.
Genetic Modification and Characterization
Bayer and M.S. transformed the recipient line “Jack” to obtain FG72 soybean.
Introduced DNA and Transformation Method
Bayer and M.S. describe plasmid pSF10. The Sal I fragment from the plasmid contains two expression cassettes, the 2mepsps expression cassette and hppdPfW336 expression cassette. The 2mepsps expression cassette contains the coding sequence for the modified 2mepsps gene derived from Zea mays under the control of the following components: (1) the Ph4a748 promoter region from the gene for histone H4 from Arabidopsis thaliana, (2) 1 h3At the first intron of gene II of the histone H3.III variant of A. thaliana, (3) the optimized TPotp C transit peptide coding sequence derived from the RuBisCO small subunit genes of Z. mays and Helianthus annuus, and (4) the 3' untranslated region encoding a polyadenylation signal from the A. thaliana histone H4 gene (histonAt). The hppdPfW336 expression cassette contains the coding sequence for the modified hppd gene from P. fluorescens strain A32 under the control of the following genetic elements: (1) a variant Ph4a748 promoter designated Ph4a748 ABBC (with an internal duplication), (2) the 5'tev leader sequence from the tobacco etch virus, the TPotp Y transit peptide coding sequence for an optimized transit peptide derived from sequences from the RuBisCO small subunit genes of Z. mays and H. annuus, and (3) the 3' terminator region from nopaline synthase from Agrobacterium tumefaciens. FG72 soybean was generated using direct gene transfer of a purified Sal I fragment of pSF10 into Jack embryo tissue.
Characteristics, Inheritance, and Stability of the Introduced DNA
Bayer and M.S. characterized the insert in FG72 soybean using restriction enzyme digestion of genomic DNA followed by Southern blot analysis. Bayer and M.S. conclude that results of this analysis demonstrate that two complete copies of the Sal I fragment from pSF10 plasmid integrated at a single locus in the soybean genome, flanked on the 5' end by two partial 3' histonAt elements. Bayer and M.S. also indicate that a segment of soybean genomic DNA was excised and translocated downstream of the 3' end of the insertion point. Bayer and M.S. note that the Southern blot analysis indicates that the integrated copies of the fragment are aligned from head-to-tail, with no interruptions from one to the next. Bayer and M.S. further note that Southern blot analyses using probes designed to detect plasmid sequences of pSF10 confirmed that no plasmid vector backbone sequences were present in FG72 soybean. Therefore, the bla gene from pSF10 was not inserted into the soybean genome.
Bayer and M.S. state that results from the Southern blot analyses were confirmed by PCR-based DNA sequencing of the flanking genomic regions and translocated fragment, which were compared to the non-transformed genome of Jack. Bayer and M.S. note that this comparison demonstrated that the flanking sequences surrounding the inserted fragment from FG72 soybean matched the organization of the soybean genome.
Bayer and M.S. assessed the stability of the DNA insert across three generations through Southern blot analysis. Bayer and M.S. state that these data demonstrated that FG72 soybean contained two copies of the 2mepsps-hppdPfW336 expression cassettes, which are stably inherited across multiple generations. Bayer and M.S. also assessed heritability and stability of the introduced DNA in FG72 soybean through analyses of segregation data as determined through tissue sensitivity to glyphosate. Bayer and M.S. state that Chi-square analysis demonstrated no statistically significant difference between the observed and expected phenotypic segregation ratios. This supports the conclusion that the duplicated 2mepsps-hppdPfW336 region contained in FG72 soybean is located at a single locus and is inherited according to Mendelian principles.
Bayer and M.S. performed bioinformatic analyses using genomic DNA sequence data to assess the potential for new open reading frames (ORF) that may encode a biologically active putative polypeptide as the result of insertion of the introduced DNA. Bayer and M.S. conclude that this analysis confirmed that the transgenic DNA insertion neither created novel open reading frames nor interrupted any genomic ORF.
Identity and Function of Introduced Proteins
FG72 soybean was genetically engineered to express modified EPSPS (2mEPSPS) protein, which confers resistance to the herbicide glyphosate. Bayer and M.S. note that the gene encoding the 2mEPSPS protein was derived through site-directed mutagenesis of a wild type epsps gene isolated from maize. Two point mutations resulted in the double mutant 2mepsps gene. EPSPS proteins, which are key enzymes in the biosynthesis of aromatic amino acids, have been previously evaluated by FDA in many other biotechnology consultations.2 The safety of the 2mEPSPS protein was described in two previous biotechnology consultations (BNF 51 and 109).
FG72 soybean was also genetically engineered to express the modified HPPD protein, which confers resistance to the herbicide isoxaflutole. Bayer and M.S. notes that the hppd gene was isolated from the P. fluorescens strain A32 and modified to contain a single point mutation by site-directed mutagenesis. The modified HPPD designated HPPD W336, contains a single amino acid substitution at position 336 to improve tolerance to herbicides. The HPPD W336 protein was the subject of an early food safety evaluation3 submitted to FDA in 2009.4 HPPD proteins are critical for plant survival; their reaction product homogenistate is the aromatic precursor of tocopherol and plastoquinone, which are essential to photosynthetic systems. HPPD enzyme inhibition also results in the disruption of the biosynthesis of carotenoids, which further destabilizes photosynthesis and leads to bleaching of the foliage and death of the plant.
Bayer and M.S. discuss analyses5 used to confirm the identity of the 2mEPSPS and HPPD W336 proteins introduced into FG72 soybean with 2mEPSPS and HPPD W336 proteins expressed in Escherichia coli. Bayer and M.S. report that these analyses confirm that 2m EPSPS and HPPD W336 proteins from FG72 soybean are equivalent to 2mEPSPS and HPPD W336 proteins expressed in E. coli, respectively. Bayer and M.S. note that evidence supports that the bacterial and the plant produced 2mEPSPS and HPPD W336 proteins lacked N-terminal methionine.
Protein Expression Levels
Bayer and M.S. conducted field expression studies of the 2mEPSPS and HPPD W336 proteins in FG72 soybean. Protein levels were determined by enzyme-linked immunosorbent assay (ELISA) conducted on several soybean tissue types at differing stages of development from samples collected from 10 field locations in the United States (U.S.) in 2008.6 Bayer and M.S. report mean concentrations of 2mEPSPS protein in soybean tissues ranging from 150 to 668 micrograms per gram (μg/g) (dry weight) in seed and leaf (growth stage V8), respectively. For HPPD W336 protein, Bayer and M.S. report mean concentrations ranging from 0.94 to 38.4 μg/g (dry weight) in seed and leaf (V4), respectively.
Potential for Toxicity of the Introduced Proteins
Bayer and M.S. provide bioinformatic analyses for the 2mEPSPS and HPPD W336 proteins using standard methods, including FASTA algorithm comparison of 2mEPSPS and HPPD W336 proteins against known protein toxins that are in the Uniprot_Swissprot, Uniprot_TrEMBL, PDB, DAD and GenPept databases. Bayer and M.S. state that these analyses revealed no relevant sequence alignments between 2mEPSPS and known protein toxins.
Bayer and M.S. report that an identity of 54% was observed between HPPD W336 and the VLLY protein from the bacteria Vibrio vulnificus, a known pathogen. An identity of approximately 50% was also seen for a similar protein, LLY, derived from Legionella pneumophila, also a pathogen. Bayer and M.S. report that both VLLY and LLY are suspected of functioning as hemolysins, and also have HPPD activity. Bayer and M.S. discuss this finding, noting that proteins with HPPD activity in plants, animals and bacteria share a high degree of homology in certain conserved domains. Bayer and M.S. note these conserved domains in LLY and VLLY correspond to the homology detected in HPPD W336. Bayer and M.S. further cite the scientific literature noting that while LLY and VLLY expression were required for hemolytic activity in vivo, there is no evidence of direct hemolytic activity by these proteins. Bayer and M.S. suggest that the proteins may have been incorrectly annotated as hemolysins.7 Bayer and M.S. performed a direct hemolysis assays with a suspension of human erythrocytes and found no lysis under conditions that allowed lysis by a positive control. Bayer and M.S. conclude that HPPD W336 is similar to other HPPD proteins, which are not toxic, and as a cutoff value for homology to toxins has not been established, the homology detected is unlikely to be toxicologically significant.
Bayer and M.S. conclude that no sequence similarity exists between 2mEPSPS and HPPD W336 protein and any known toxin or other biologically active proteins that would be harmful to human or animal health.
Bayer and M.S. also describe results from separately conducted acute toxicity studies in mice on 2mEPSPS and HPPD W336 proteins. Female OF1 mice were administered a single dose by oral gavage of 2000 milligrams per kilogram of body weight (mg/kg bw) of either 2mEPSPS protein produced in E. coli or bovine serum albumin as a control. Separately, female OF1 mice were administered a single dose of 2000 mg/kg bw of E.coli-produced HPPD W36 protein or bovine serum albumin. Bayer and M.S. report that no treatment-related events were observed in either study, as determined by survival, clinical observations, or gross pathology. Bayer and M.S. conclude that the 2mEPSPS and HPPD W336 proteins are unlikely to have acute toxic effects in humans or animals.
Potential for Allergenicity of the Introduced Protein
Bayer and M.S. provide analyses of the potential for allergenicity of the 2mEPSPS and HPPD W336 proteins. Bayer and M.S. assessed the structural similarity of 2mEPSPS and HPPD W336 proteins to known protein allergens through standard bioinformatic methods using the public allergen database AllergenOnline (www.allergenonline.com); release 8, 2008 for 2mEPSPS protein, release 9.2, 2009 for HPPD W336 protein). Bayer and M.S. report that no sequence alignment met or exceeded the threshold of 35% identity over 80 amino acids and there were no contiguous stretches of 8 or greater amino acids shared between the 2mEPSPS or HPPD W336 proteins and proteins in the allergen database.
Bayer and M.S. describe a series of in vitro assessments of the stability of E. coli-produced 2mEPSPS and HPPD W336 proteins in simulated gastric fluid. Bayer and M.S. state that these studies indicate that 2mEPSPS and HPPD W336 proteins are rapidly digested. Bayer and M.S. also discuss the lack of potential for glycosylation and the heat lability of 2mEPSPS and HPPD W336 proteins as further evidence supporting their conclusions that these proteins are unlikely to be allergenic.
Food & Feed Use
Bayer and M.S. review the historical uses of soybean (Glycine max), noting that soybeans account for approximately 90% of the total U.S. oilseed production and that processed soybeans are the largest source of protein in animal feed. Bayer and M.S. note that in the U.S., nearly all soybeans are crushed to extract oil from the meal. A small amount of whole soybeans are used for seed, roasted for snacks or on-farm feed. Soybean oil is rich in polyunsaturated fatty acids and is commonly used as a salad and cooking oil and in the production of food ingredients. The majority of soybean meal is used in animal feed, primarily in poultry, swine, and beef and dairy cattle diets, while a small fraction of soybean meal is further processed into soybean flours and soybean proteins for a variety of food uses.
Scope of Analysis
Bayer and M.S. report data on sixty-seven components, including key nutrients, anti-nutrients and isoflavones, in grain or forage derived from FG72 soybean and compared them with the parent “Jack” non-transgenic control variety (control). Two FG72 soybean treatment groups were utilized in the study, a treated group that received both glyphosate and isoxaflutole (treated) and a second FG72 soybean group that did not receive these herbicides (not treated).8 Commercial reference soybean varieties were also grown alongside the three test crops for comparison.
Study Design - Compositional Analyses
Bayer and M.S. state that grain was obtained from soybean grown in 10 locations in soybean growing regions in the U.S. during the 2008 growing season. At each location, three replicates of the treated and untreated FG72 soybean and control and three commercial reference soybean varieties were grown using a randomized block design. Bayer and M.S. analyzed the data using analysis of variance methods comparing compositional data combined from all sites from treated and untreated FG72 soybean and the control. Statistical differences in composition were considered significant at the p < 0.05 level. When a statistically significant difference was detected between FG72 soybean and the control, Bayer and M.S. assessed whether the difference was biologically meaningful9 with regard to food and feed safety. This assessment included analysis of the variation across locations and comparison of the values for FG72 soybean with the range observed from the reference varieties, as well as values in the published literature.10,11 Bayer considered component means that fell within the commercial or literature reference ranges to be within the normal variation for commercial soybeans.
Results of analyses
Compositional Analysis of Soybean Forage
Bayer and M.S. report the results of compositional analysis for seven components in soybean forage (moisture, crude protein, crude fat, ash, carbohydrates (by calculation), acid detergent fiber (ADF), and neutral detergent fiber (NDF). Bayer and M.S. state that no significant numerical or biological differences were observed in the analysis of components from FG72 soybean when compared to the control and reference varieties.12 Bayer and M.S. conclude that this analysis supports the conclusion that forage from FG72 soybean is compositionally equivalent to conventional soybean forage.
Compositional Analysis of Soybean Grain
Bayer and M.S. analyzed FG72 soybean and control grain samples for proximates (moisture, crude protein, crude fat, ash, carbohydrates (by calculation)) and fiber (ADF and NDF), 24 fatty acids (C8-C24), 18 amino acids, 10 vitamins, and 6 minerals.
Bayer and M.S. report that statistically significant differences in fat, ash, carbohydrates, palmitic (16:0), stearic (18:0), arachidic (20:0), behenic (C22:0), lignoceric (C24:0), oleic (C18:1), eicosenoic (C20:1), linoleic (C18:2), and α-linolenic (18:3) acids, α-tocopherol, γ-tocopherol, total tocopherols, calcium, potassium, magnesium, and iron were obtained when FG72 soybean, treated or not treated, was compared with the control. In the cases where statistically significant differences were observed, Bayer and M.S. note that mean values for FG72 soybean for each of these components fell within the calculated ranges established from the commercial varieties grown concurrently. Bayer and M.S. state that compositional components identified as statistically different between FG72 soybean and the control were within their natural variability in soybean. Bayer and M.S. thus conclude that these differences are not biologically meaningful to food and feed safety and nutrition.
Anti-Nutrients and Isoflavones
Bayer and M.S. report that FG72 soybean and control samples were analyzed for anti-nutrients (phytic acid, raffinose, stachyose, lectin, and trypsin inhibitor) and isoflavones (daidzin, genistin, glycitin, daidzein, genistein, and glycitein). Statistically significant differences in raffinose, lectin, genistin, glycitin, and total isoflavones were observed in one or both of the FG72 soybean treatment groups when compared to control. Bayer and M.S. note that the mean values for these components observed for FG72 soybean were within the range of natural variability reported in the literature. Bayer and M.S. thus conclude that the observed differences are not biologically meaningful to food and feed safety and nutrition.
Summary of Compositional Analyses
Bayer and M.S. conclude that the differences in the levels of the components described above are not considered to be biologically meaningful to food and feed safety and nutrition. Bayer and M.S. state that the compositional analysis supports the conclusion that FG72 soybean is compositionally equivalent to conventional soybean varieties.
Poultry Feeding Study
Bayer and M.S. provide summary data from a 42-day broiler chicken feeding study comparing three different diets containing 20 percent toasted soybean meal derived from either: (1) FG72 soybean, (2) Jack soybean, or (3) a commercial variety. Bayer and M.S. state that growth rate and health parameters were similar across the three groups. Bayer and M.S. state that this study confirmed the nutritional wholesomeness of FG72 soybean.
Endogenous Allergens in Soybeans
Bayer and M.S. also discuss the effects of FG72 soybean on expression of endogenous allergens. Bayer and M.S. compared endogenous allergen protein expression in FG72 soybean with the expression in the non-transgenic Jack soybean variety and observed no significant changes in the concentrations of known allergenic proteins. Bayer and M.S. also tested reactions in Western blot analyses of extracted proteins from FG72 soybean, the control, and three commercial non-transgenic soybean lines with the sera from seven patients known to be allergic to soybeans. Responses to FG72 soybean were comparable to those of the other soybean varieties. Bayer and M.S. conclude that levels of endogenous allergens in FG72 and soy-allergic serum antibody binding to extracted proteins from FG72 soybean were comparable to those of other soybean varieties.
FDA evaluated Bayer and M.S.’s submission to determine whether FG72 soybean raises any safety or regulatory issues with respect to the intended modifications or with respect to the food and feed itself. Based on the information provided by the companies and other information available to the agency, FDA did not identify any safety or regulatory issues under the FD&C Act that would require further evaluation at this time.
Bayer and M.S. have concluded that their glyphosate- and isoxaflutole-tolerant soybean variety, FG72 soybean, and the foods and feeds derived from it are as safe as conventional soybean varieties and, with the exception of the herbicide tolerance traits, are not materially different in composition or any other relevant parameter from other soybean varieties now grown, marketed, and consumed in the U.S. At this time, based on Bayer and M.S.’s data and information, the agency considers Bayer and M.S.’s consultation on FG72 soybean to be complete.
Richard E. Bonnette
1Bayer submitted its evaluation of the potential for allergenicity and toxicity of the HPPD 336 protein, which FDA designated as New Protein Consultation No. NPC 000010 under FDA's Guidance to Industry: "Recommendations for the Early Food Safety Evaluation of New Non-Pesticidal Proteins Produced by New Plant Varieties Intended for Food Use." FDA responded that it had no questions regarding Bayer's conclusions.
2BNF 1, 20, 26, 35, 51, 56, 71, 77, 79, 80, 84, 90, 97, 98, 104, and 109 (http://www.fda.gov/bioconinventory)
3Under FDA's Guidance for Industry: Recommendations for the Early Food Safety Evaluation of New Non-Pesticidal Proteins Produced by New Plant Varieties Intended for Food Use (available at http://www.fda.gov/npcguidance)
4NPC 000010 (available at http://www.fda.gov/npcinventory)
5These analyses included N-terminal amino acid sequence analyses through mass spectrometry, western blot analyses and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and enzyme activity.
6Seed, leaf (V4, V6, and V8), stem (V4 and V8), and root (V4 and V8).
7Bayer and M.S. note that the LLY protein can convert tyrosine in culture medium to homogenistic acid, which then oxidizes and polymerizes into a beta-hemolytic melanin-like component. Bayer and M.S. consider that other over-expressed HPPD proteins could react similarly.
8Bayer notes that pre-emergence, soil-applied herbicides pendimethalin (with each treatment group at each location) and metolachlor (in one field trial) were used in the study.
9FDA considers “biologically meaningful” to connote that there is a notable impact on dietary requirements, or an impact that otherwise would affect human or animal health.
10Bayer used version 3.0 (accessed on September 7, 2007) of the International Life Sciences Institute-Crop Composition Database (ILSI-CCD) for its comparative analysis. The database is maintained by ILSI and can be accessed at http://www.cropcomposition.org/.
12Bayer and M.S. compared calculated means and standard deviations for components measured in forage.