Biotechnology Consultation Note to the File BNF No. 000114
Return to inventory: Submissions on Bioengineered New Plant Varieties
Biotechnology Consultation - Note to File
Biotechnology Notification File No. 000114
DATE: January 18, 20121
Subject: BPS-CV127-9, herbicide tolerant soybean
Keywords: Soybean, Glycine max, imidazolinone herbicide-resistant, BASF, BPS-CV127-9 soybean, csr1-2 gene, Arabidopsis thaliana, acetohydroxyacid synthase, At-AHAS enzyme, AHAS enzymes, OECD Unique Identifier BPS-CV127-9.
This document summarizes FDA’s evaluation of biotechnology notification file (BNF) No. 000114. In a submission dated January 26, 2009, BASF Plant Science, L.L.C. (BASF) submitted to FDA a safety and nutritional assessment of imidazolinone-tolerant soybean BPS-CV127-9 (CV127 soybean). BASF provided additional information in submissions dated May 29, 2009 and May 10, 2011. FDA evaluated the information in BASF's submissions to ensure that regulatory and safety issues regarding the food or feed derived from the new plant variety have been resolved prior to commercial distribution. In our evaluation, we considered all information provided by the notifier 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 effect in CV127 soybean is to confer tolerance to imidazolinone herbicides. Acetoxhydroxyacid synthases (AHAS) catalyze an early step in the synthesis of branched-chain amino acids. Imidazolinone herbicides bind to AHAS enzymes in plants, thus inhibiting the synthesis of branched-chain amino acids. To confer tolerance to imidazolinone herbicides, BASF introduced into the soybean genome the csr1-2 gene from Arabidopsis thaliana. The csr1-2 gene encodes the large catalytic subunit of AHAS (in general the AHAS-L2 subunit; specifically, At-AHAS-L refers to the large subunit of the A. thaliana AHAS enzyme also referred to as At-AHAS protein or At-AHAS enzyme). The csr1-2 gene introduced into soybeans contains a point mutation3 that prevents binding of imidazolinone herbicides to the At-AHAS-L subunit.
The purpose of this evaluation is to assess whether the developer has introduced a substance requiring premarket approval as a food additive or raised 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 in CV127 soybean are considered pesticidal substances. In its submission to FDA, BASF indicated that it intended to submit regulatory packages to EPA for the use of specific imidazoline herbicides on CV127 soybean.
Genetic Modification and Characterization
BASF used the commercial soybean variety “Conquista” to develop imidazolinone-tolerant CV127 soybean.
Introduced DNA and Transformation Method
BASF used plasmid pAC321 to prepare DNA for the transformation of soybean tissue. The plasmid consisted of the Escherichia coli cloning plasmid pBluescript SK(-) and the transformation fragment containing the csr1-2 gene from A. thaliana with its native promoter and 3’ untranslated region. In addition to the native csr1-2 promoter, the 5’ region upstream of the csr1-2 coding sequence contained the complete coding sequence of the A. thaliana SEC61 gamma protein (AtSEC61γ protein), a transport protein, but was considered to contain little of the promoter. The transformation fragment was isolated from plasmid pAC321 by digestion with the restriction endonuclease PvuII, purified, and used to transform embryonic tissue derived from a single soybean seed. A microprojectile bombardment technique was used to produce transformation events containing the csr1-2 gene. The cells subjected to transformation were subsequently grown in a selective medium containing the imidazolinone herbicide imazapyr. A plant tolerant to imazapyr designated T0 was selected for further development; the transformation event was designated soybean event CV127.
Characteristics, Inheritance, and Stability of the Introduced DNA
BASF described the breeding history of CV127 soybean. The plants were self-crossed beginning from the T0 transformant through successive generations to ultimately obtain the T4 generation. The T4 generation was subsequently backcrossed to the parental line Conquista to obtain CV127 soybean, line 603. The T4 generation and generations F4, F8, and F9 of line 603 were used to characterize the insert and assess its stability across multiple generations. The F8 generation of line 603 was crossed with a commercial line closely related to Conquista. The resulting CV127 soybean, line 127 was used in the trait inheritance study. BASF conducted Southern blot and PCR analyses to characterize the insert. Southern blot analyses of the genomic DNA confirmed that no DNA derived from the backbone of plasmid pAC321 was present in the genome of soybean CV127. Southern blot analysis of the DNA from F4, F8 and F9 generations showed the presence of one intact copy of the csr1-2 gene cassette integrated in a single locus of the CV127 soybean genome. Southern blot analysis also confirmed that the insert was stable across multiple generations. These studies also showed that the organization of the genetic elements present in the insert was the same as in plasmid pAC321. BASF reported that the insert contained a 376 base pair (bp) duplication of a portion of the csr1-2 gene located directly before the 3’ integration point. This duplicated segment created a 501-bp putative open reading frame (ORF) extending into the 3’ flanking sequence. BASF conducted RT-PCR analysis of young leaf tissue to determine whether this ORF was expressed. The analysis showed that expression of this 501-bp ORF was not detected. BASF conducted the trait inheritance study on four generations of CV127, line 127. Each generation was sprayed with imazapyr to analyze the inheritance of imidazolinone tolerance based on the expected and observed segregation ratios. The results of this analysis confirmed that the imidazolinone tolerance trait is conferred by a single functional copy of the csr1-2 gene and that the DNA insert is stably integrated into the CV127 soybean genome across multiple generations.
Introduced Protein – At-AHAS
Identity, Function, and Characterization
BASF states that CV127 soybean contains the csr1-2 gene derived from A. thaliana that encodes the imidazolinone-tolerant At-AHAS-L subunit. The At-AHAS-L subunit interacts with endogenous soybean small regulatory subunit (AHAS-S) to form an enzyme complex that catalyzes the first step in the synthesis of branched-chain essential amino acids, valine, leucine, and isoleucine. The AHAS enzymes occur ubiquitously in plants. Imidazolinone herbicides inhibit the native enzymes, resulting in plant death. BASF states that the csr1-2 gene introduced into soybeans differs from the native Arabidopsis csr1-2 gene by two point mutations. One of these mutations4 prevents binding of imidazolinone herbicides to the At-AHAS-L subunit and confers tolerance to these herbicides; the second mutation has no effect. Neither mutation affects the biosynthetic function of the At-AHAS enzyme. BASF also states that, in addition to the csr1-2 gene, CV127 soybean contains the A. thaliana gene encoding the AtSEC61γ subunit, a protein of the endoplasmic reticulum secretory system.
BASF determined the expression levels of the At-AHAS protein in CV127 soybean by enzyme-linked immunosorbent assay (ELISA) using AHAS-specific antibodies. Due to high amino acid homology between the At-AHAS and the endogenous soybean AHAS enzyme, the antibodies used in the ELISA assay could not distinguish the two enzymes. Consequently, the expression levels in CV127 soybean represent the sum of At-AHAS and native AHAS and are collectively referred to in this section as the levels of the AHAS protein. Samples for analysis were collected from field trials conducted in two growing seasons at multiple locations. All plots of the CV127 soybean were sprayed with imidozolinone herbicide for weed control, whereas all plots of the control soybeans were treated with conventional herbicide. Leaf and grain samples were collected from all trial sites at the V2 (young plants) and R8 (full maturity) stages, respectively. In addition, six whole plants per plot at three developmental stages were collected from two trial sites. Three of the six plants were dissected into plant parts depending on the maturity stage. BASF provided the levels of the AHAS protein on a fresh weight and dry weight basis. The highest levels of the AHAS protein were found in the leaves of CV127 soybean plants collected at V2 growth stage, up to 128 nanograms per gram (ng/g) of fresh tissue or up to 714 ng/g of dry tissue. The corresponding levels of the AHAS protein in the isoline control leaf tissue were lower, as expected. The levels of the AHAS protein decreased in all tissues with the age of the plants. The AHAS levels in all grain samples from CV127 soybean and the isoline control were below the level of quantification (LOQ) of 15 ng AHAS/g dry weight. BASF has also reported that the AHAS protein was not detectable (at the detection limit of 6-9 ng/g dry weight) in processed fractions prepared from the grain of CV127 soybean, including defatted toasted meal, defatted untoasted meal, protein isolate, protein concentrate, and oil.
Assessment of Potential for Toxicity and Allergenicity
To obtain sufficient quantities of the At-AHAS protein for conducting safety studies, the crs1-2 gene was expressed in E. coli. The crs1-2 gene introduced into an E. coli expression system contains the same mutations as the crs1-2 gene expressed in CV127 soybean. BASF conducted comparative studies between the E. coli–expressed and soybean-expressed At-AHAS proteins.5 BASF concluded that both proteins were functionally active and had the expected enzymatic activity; were tolerant to the imidazolinone herbicide imazethapyr; had comparable immunoreactivity when probed with anti-AHAS antibody; and neither protein showed evidence of glycosylation. BASF noted that the results of the amino acid sequence analysis indicated that both the E. coli-produced and the CV127 soybean-produced proteins consisted of the predicted mature At-AHAS 585 amino acid sequence; however, the CV127 soybean-produced protein contained approximately 34 additional amino acids at the N-terminus. The latter sequence should be part of the chloroplast transit peptide (CTP) based on the predicted length of 85 amino acid residues for the CTP in the literature, which should be cleaved off during the transition of the At-AHAS protein to the chloroplast. BASF stated that it is conceivable that the CTP of the At-AHAS protein expressed in soybeans was cleaved off at a different C-terminal amino acid residue (most likely residue 51) rather than the predicted C-terminal residue 85. BASF concluded that the overall results of the comparative studies justify the use of the E. coli-produced At-AHAS protein as an appropriate substitute for the CV127 soybean-produced At-AHAS protein in safety assessment studies of the At-AHAS protein. To assess the potential for toxicity of the At-AHAS protein, BASF conducted an amino acid homology search to known toxins using BLASTP sequence similarity search against non-redundant peptide sequences in the GeneBank database (posted on March 14, 2008). BASF reported that no significant homology was detected between the At-AHAS protein and known protein toxins. BASF also conducted an acute oral toxicity study in mice; no significant test substance-related effects at 5000 milligrams per kg body weight in males or females were observed. BASF concluded that neither the sequence homology study nor the toxicity study in mice revealed any safety concerns regarding the At-AHAS protein. To assess the potential for allergenicity of the At-AHAS protein, BASF used a weight-of -evidence approach recommended by the Codex Alimentarius Commission.6 BASF compared the amino acid sequence of the At-AHAS protein to the sequences of known allergens in the Food Allergy Research and Resource Program (FARRP) Allergen Database (version 8.00). BASF reported that the At-AHAS protein did not show 35% or greater amino acid identity over any 80 amino acid segment and did not share a sequence identity of at least eight consecutive amino acids with allergens listed in the FARRP database. Furthermore, BASF reported that the E. coli-produced At-AHAS protein was rapidly degraded in the simulated gastric fluid (SGF) and simulated intestinal fluid (SIF). Similarly, the At-AHAS and endogenous soybean AHAS proteins present in leaf and grain extracts of CV127 soybean were also rapidly degraded in both SGF and SIF.7 BASF concluded that the At-AHAS protein expressed in CV127 soybean does not possess any attributes of known food allergens. BASF also stated that the source of the csr1-2 gene that encodes the At-AHAS protein expressed in CV127 soybean, A. thaliana, is not a known human or animal pathogen and is not known to produce toxic substances or cause allergic reactions. The At-AHAS protein is structurally and biologically closely related to other AHAS proteins that occur ubiquitously in plants, including crop plants used in food and feed.
Potential for expression of AtSEC61γprotein
BASF states that the DNA inserted into CV127 soybean contains the complete coding sequence of the A. thaliana SEC61γ gene encoding the Arabidopsis SEC61γ (AtSEC61γ) subunit, but lacks its native promoter. The AtSEC61γ protein is a component of a multi-protein secretory complex that is ubiquitous in all eukaryotes. BASF reported that despite a very low probability of expression of the AtSEC61γ gene in CV127 soybean, they produced the AtSEC61γ protein in E. coli and evaluated its safety and its expression levels in CV127 soybean. BASF conducted the amino acid sequence homology study and concluded that the AtSEC61γ protein does not share meaningful homology with allergens and protein toxins and is nearly identical to the native soybean SEC61γ subunit. BASF also reported that the AtSEC61γ protein is rapidly digested in SGF. The AtSEC61γ protein was not detectable using Western blot analysis in leaf and grain extracts of both CV127 soybean and control soybean plants at the limits of detection (LOD) of 5 ng/g and 15 ng/g for leaf and grain tissue, respectively.
Food and Feed Use
Soybean (Glycine max) is grown around the world for a variety of food, feed, and industrial uses. Soybean seeds are processed into oil and meal. 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. A small fraction of soybean meal is further processed into soy flours and soy proteins for a variety of food uses. Traditional foods prepared from soybeans include tofu, miso, soymilk, tempeh, and soy sauce. The preponderance of soybean meal is used in animal feed, primarily in poultry, swine, and beef and dairy cattle diets. Soybean meal is processed in moist heat to inactivate trypsin inhibitors and lectins, which are antinutrients occurring in raw soybeans.
Scope of Analysis
BASF analyzed the composition of grain and forage of CV127 soybean line 127.8 BASF used the F5, F6, and F7 generations of line 127 for composition studies using as controls the corresponding near-isogenic nontransgenic generations lacking the csr1-2 gene (null segregants). BASF also analyzed two conventional soybean varieties typically cultivated in Brazil to establish a range of natural variability in the composition.
BASF reported that grain samples of CV127 soybean and controls were harvested from soybean plants grown in four replicated plots at each of six field sites in Brazil during the 2006/2007 growth season and from four field sites in the 2007 short season. Forage tissues were harvested at the R2 growth stage from four replicated plots at each of six field sites in Brazil during the 2007/2008 season. Forage and grain samples were also harvested from two conventional varieties grown on each site. The CV127 soybean plants were treated with imidazolinone herbicide, whereas the control and conventional varieties were treated with conventional herbicide. For grain analysis, the CV127 soybean plants were also treated with conventional herbicide. As CV127 soybean has been developed for use with imidazolinone herbicides, BASF stated that the conventional herbicide will not be used in the commercial cultivation of CV127 soybean. With the exception of moisture, all results are reported on a dry weight basis. BASF performed statistical analysis on data for CV127 soybean line 127 and control soybean from individual test sites and as well as data aggregated from all test sites using the SAS Version 9.1 analysis of variance. Differences in composition were considered significant at the 0.05 confidence level. In addition, BASF provided the composition of processed fractions (soybean oil, toasted meal, protein isolate, and protein concentrate) obtained from grain of CV127 soybean, and three conventional soybean lines.
Results of Analyses
BASF provided analytical data for the following forage components:
- Proximates: moisture, protein, fat, ash, and carbohydrates (calculated by difference)
- Neutral detergent fiber (NDF) and acid detergent fiber (ADF)
BASF states that no statistically significant differences were observed in levels of moisture, ash, fat, protein, carbohydrates, calories, crude fiber, ADF and NDF in forage samples produced from CV127 soybean compared to levels of these components in the isoline control variety. Where small differences in composition were observed between CV127 soybean and conventional varieties, they were attributed to germplasm differences. BASF concludes that the forage from CV127 soybean is compositionally equivalent and as nutritious as forage from conventional soybean varieties with a long history of safe use in animal feed.
BASF analyzed grain for 70 components (listed below) and provided data obtained for each individual test site and data aggregated from all sites. BASF conducted the statistical analysis of data obtained for CV127 soybean and the control soybean. BASF also provided compositional data for two commercial conventional varieties and, for comparison, data reported in the International Life Sciences Institute (ILSI) Crop Composition Database (version 3.0, 2008). ILSI data were not available for several components, e.g., for some vitamins and phospholipids.
BASF provided analytical data for the following grain components:
- Proximates: moisture, protein, fat, ash, and carbohydrates (calculated by difference)
- Total dietary fiber (TDF), crude fiber, NDF and ADF
- Amino acids (18)
- Fatty acids (C14-C22)
- Minerals (calcium, iron, magnesium, phosphorus, and potassium)
- Vitamins (α-, β-, γ-, δ-, and total tocopherols, vitamins E and B1, and folic acid)
- Isoflavones (daidzin, malonyl-daidzin, daidzein, glycitin, malonyl-glycitein, genistin, malonyl-genistin, and genistein)
- Phospholipids (phosphatidic acid, phosphatidyl ethanolamine, phosphatidic acid, phosphatidyl inositol, and phosphatidyl choline)
- Antinutrients (phytic acid, trypsin inhibitor, lectin, urease, raffinose, and stachyose)
For data from individual sites, BASF reports that the mean levels of some components of CV127 soybean were statistically significantly different from the levels in the control soybean. However, these differences were small and not consistent among the different test sites. BASF also noted minor differences in composition between CV127 soybean and conventional varieties and stated that they are likely due to the natural heterogeneity of soybean varieties grown in Brazil.
BASF states that in the combined data analysis, statistically significant differences in the mean levels of some components were identified.9 Nevertheless, the levels of these components in CV127 soybean treated with imidazolinone herbicide and conventional herbicide were comparable to the levels in the control soybean, as well as to the levels in two conventional varieties. BASF also notes that the mean levels of most components were either within or comparable to the ranges reported by the ILSI. In several instances, the levels of some components were equivalent among all tested soybean varieties but were different from the levels reported by ILSI. BASF stated that these differences most likely reflect characteristics of soybean varieties developed for production under tropical growth conditions in Brazil. BASF concludes that the introduction of the A. thaliana csr1-2 gene, which confers resistance to imidazolinone herbicides, does not impact the nutritional composition of grain produced by CV127 soybean.
BASF provided composition data for processed fractions from CV127 soybean (oil, toasted meal, protein isolate, and protein concentrate), control soybean (the parental variety, Conquista), and two commercial conventional soybean varieties. Soybean grain was produced at four field trial locations during the 2006/2007 growing season and processed into fractions according to the standard soybean processing methods. Toasted meal was analyzed for proximates (moisture, ash, fat, protein, and carbohydrates (calculated by difference), fiber (ADF and NDF), antinutrients (raffinose, stachyose, trypsin inhibitor, urease, and phytic acid), and isoflavones. Protein isolate and concentrate were analyzed for proximates and refined oil was analyzed for the fatty acid composition. All data were statistically analyzed. For comparison, BASF provided data on the composition of processed fractions reported in the literature.
BASF noted statistically significant but small differences in the levels of some components. BASF also noted that the levels of several components were outside the ranges reported in the literature. BASF stated that these differences are most likely due to the characteristics of soybean varieties adapted for cultivation in Brazil. BASF noted that the composition of processed fractions from CV127 soybean is equivalent to the composition of the same processed fractions derived from the control and conventional soybean varieties. BASF concluded that processed fractions from CV127 soybean are appropriate for use in human food and animal feed.
Poultry Feeding Study
BASF stated that they conducted a poultry feeding study using toasted meal from CV127 soybean10, control soybean (the parental variety Conquista), and two commercial soybean lines. All these lines were grown at one field location. The grain from each line was harvested and processed into toasted meal according to standard procedures. The grain and meal samples were tested for the presence of several mycotoxins and pesticides used in cultivation of soybeans. None of the analyzed mycotoxins was detected and pesticides were either not detected or were present below the levels of concern. Five hundred seventy six broiler chickens were used in the study. The broilers were organized into groups; each group was fed a balanced feed containing toasted meal from a soybean line. The results were statistically analyzed using the Dunnett’s Test. BASF reported that there were no statistically significant differences (P>0.05) in the measured parameters (body weight, weight gain, feed intake or feed conversion) between broilers fed feed containing meal from CV127 soybean and meal from the control soybean or one of the commercial varieties (Monsoy 8001). In broilers fed a diet containing meal from the second commercial variety (Coodetec 217), body weight and weight gain were statistically significantly lower compared with broilers fed diets containing meal from CV127 soybean. BASF concludes that soybean meal obtained from CV127 soybean is nutritionally comparable to soybean meal processed from commercial soybean varieties and is appropriate for use in animal feed.
BASF stated that they compared the content of major soybean allergens between CV127 soybean and the parental variety Conquista using proteomic studies. BASF subjected protein extracts from soybean grain to the two-dimensional gel electrophoresis (2D-PAGE). The separated proteins were visualized with silver staining. BASF identified the putative allergen spots by comparison with soybean 2D-protein maps published in the literature or available through a public database (http://oilseedproteomics.missouri.edu). BASF subsequently confirmed the identity of the protein spots via mass spectrometry. BASF also reported that they tested in vitro IgE binding to proteins extracted from CV127 soybean, near-isogenic control line, and three conventional commercial soybean lines using sera from individuals allergic to soybeans. Sera from individuals with no known allergy to soybeans were used as controls. Potential differences in IgE binding among the tested lines were assessed using Western blot and inhibition ELISAs. BASF concluded that no meaningful differences in the levels of endogenous allergens were detected between CV127 soybean, near-isogenic control soybean, and three conventional soybean varieties with a history of safe use in food and feed.
FDA evaluated BASF’s submission to determine whether the developer’s product raises any safety or regulatory issues with respect to the intended modification or with respect to the food itself. Based on the information provided by the company 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. BASF has concluded that, with the exception of the intended modification (imidazolinone herbicide tolerance), its BPS-CV127-9 soybean variety and the foods and feeds derived from it are not materially different in composition or any other relevant parameter from other soybean varieties now grown, marketed, and consumed. At this time, based on BASF’s data and information, the agency considers BASF’s consultation on BPS-CV127-9 soybean to be complete.
Robert I. Merker, Ph.D.
1This memorandum was drafted by Dr. Zofia Olempska-Beer, who retired from the FDA on December 30, 2011.
2BASF referred to this as “AHASL” in its assessment; we have added hyphens for clarity.
3A second point mutation was discovered in the csr1-2 gene in the CV127 soybean genome, in which an arginine residue at position 272 replaced by lysine, does not impact either the enzymatic function of the AHAS enzyme or herbicide tolerance.
4A single nucleotide substitution, resulting in the replacement of an asparagine by serine.
5BASF assayed for 1) molecular weight, 2) immunoreactivity with AHAS-specific antibody, 3) enzymatic activity, the inhibition of the activity by imidazolinone herbicide, and feedback inhibition by branched-chain amino acids, 4) glycosylation, and 5) determination of the amino acid sequence of several peptide fragments derived from purified AHAS protein preparations.
6Guideline for the Conduct of Food Safety Assessment of Foods Derived From Recombinant-DNA Plants; CAC/GL 45-2003.
7BASF also compared the heat stability between the E. coli-expressed and the wild-type AtAHAS proteins and reported that both proteins rapidly lose their enzymatic activity at temperatures above 60°C.
8Line 127 was developed by 1) backcrossing the CV127 T4 generation with the parental line Conquista to obtain line 603; and 2) crossing the F8 generation of line 603 (homozygous for the csr1-2 gene) with a conventional line closely related to Conquista.
9Statistically significant differences were found between CV127 soybean and the isoline control in protein, fat, carbohdrates, ADF, NDF, histidine, oleic, linoleic, linolenic, and behenic acids, iron and phosphorus, folic acidtocopherols and vitamin E, and vitamin B1, phospholipids, stachyose, lectins, and isolflavones, but levels of these components were within ranges of the commercial varieties tested.
10BASF used toasted meal from the F9 generation of CV127 soybean line 603.