PRESENCE OF GLYPHOSATE IN FOOD PRODUCTS IN SOUTH AFRICA OF WHICH MAIZE OR SOYBEAN IS THE PRIMARY CONSTITUENT BJ Koortzen February 2017 PRESENCE OF GLYPHOSATE IN FOOD PRODUCTS IN SOUTH AFRICA OF WHICH MAIZE OR SOYBEAN IS THE PRIMARY CONSTITUENT By BJ Koortzen Dissertation submitted in fulfilment of the requirements for the degree Magister Medical Scientiae (Human Molecular Biology) In the Faculty of Health Sciences Department of Haematology and Cell Biology University of the Free State Supervisor: Prof. CD Viljoen Co-Supervisor: Ms. S Sreenivasan February 2017 Bloemfontein South Africa Declaration DECLARATIONS I. I, Barend Johannes Koortzen declare that the masters research dissertation that I herewith submit at the University of the Free State, is my independent work and that I have not previously submitted it for a qualification at another institution of higher education. II. I, Barend Johannes Koortzen hereby declare that I am aware that the copyright is vested in the University of the Free State. III. I, Barend Johannes Koortzen hereby declare that all royalties as regards intellectual property that was developed during the course of and/or in connection with the study at the University of the Free State will accrue to the University. Barend Johannes Koortzen i Contents CONTENTS DECLARATIONS ........................................................................................................... i ACKNOWLEDGEMENTS ............................................................................................. v LIST OF SCIENTIFIC ABBREVIATIONS AND ACRONYMS.................................... vi LIST OF ABBREVIATIONS AND ACRONYMS FOR ORGANIZATIONS, INSTITUTIONS AND/OR AUTHORITIES ................................................................... ix LIST OF FIGURES ........................................................................................................ x LIST OF TABLES ........................................................................................................ xi PREFACE ................................................................................................................... xiii CHAPTER 1: LITERATURE REVIEW ......................................................................... 1 1.1 Introduction to genetically modified organisms ........................................... 1 1.2 GM crop production worldwide ....................................................................... 2 1.3 Glyphosate use in agriculture ......................................................................... 3 1.4 Glyphosate tolerance levels ............................................................................ 5 1.5 Safety assessment of glyphosate ................................................................... 7 1.6 Detection of glyphosate as a result of agricultural application ................ 13 1.6.1 Detection of glyphosate in HT maize and HT soybean ........................... 13 1.6.2 Glyphosate in processed food products and water ................................. 14 1.6.3 Glyphosate in animal tissue and excretions ............................................ 16 1.7 Conclusion ....................................................................................................... 18 CHAPTER 2: RESEARCH AIM AND METHODOLOGY .......................................... 20 2.1 Rationale .......................................................................................................... 20 2.2 Aim of Study .................................................................................................... 20 2.3 Study Design ................................................................................................... 21 2.4 Product selection and sampling ................................................................... 21 2.4.1 Inclusion criteria ........................................................................................ 23 ii Contents 2.4.2 Exclusion criteria....................................................................................... 23 2.5 Methodology for glyphosate screening ....................................................... 23 2.5.1 Sample preparation and glyphosate determination ................................. 23 2.6 Methodology for DNA screening ................................................................... 24 2.6.1 Sample preparation and DNA extraction ................................................. 24 2.6.2 Gel electrophoresis and fluorometry ........................................................ 25 2.6.3 Real-time PCR screening for the presence of GM HT events in food products .................................................................................................... 26 2.6.4 Real-time PCR quantification of GM HT events detected in food products .................................................................................................................. 26 2.7 Data analysis ................................................................................................... 28 2.8 Compliance to mandatory GM labelling in South Africa ........................... 28 CHAPTER 3: RESULTS AND DISCUSSION FOR GLYPHOSATE CONTENT IN FOOD PRODUCTS ..................................................................................................... 29 3.1 Detection of glyphosate in maize and soybean food products ................ 29 3.2 Detection and quantification of GM HT events in food products ............. 33 3.3 Correlation between percentage GM HT event and level of glyphosate in food products ................................................................................................. 36 3.4 Theoretical exposure to glyphosate through food products in South Africa ................................................................................................................ 42 CHAPTER 4: RESULTS AND DISCUSSION IN TERMS OF COMPLIANCE WITH GM LABELLING IN SOUTH AFRICA ........................................................................ 47 4.1 Analysis of GM HT content in GM labelled food products ........................ 47 CHAPTER 5: LIMITATIONS OF THIS STUDY ......................................................... 53 CHAPTER 6: CONCLUSION ..................................................................................... 54 LIST OF REFERENCES ............................................................................................. 57 SUMMARY .................................................................................................................. 72 iii Contents OPSOMMING .............................................................................................................. 74 APPENDIX A ............................................................................................................... 77 APPENDIX B ............................................................................................................... 81 APPENDIX C ............................................................................................................... 89 iv Acknowledgements ACKNOWLEDGEMENTS The completion and success of this study could not have been possible without the help of the following individuals. For you, I am truly grateful. • Prof CD Viljoen, for the funding and guidance of this project, sharing his knowledge in molecular biology with me, and allowing me to develop into an independent scientist. • Ms S Sreenivasan for her assistance and support, as well as for teaching me everything I had to know regarding Real-time qualitative and quantitative PCR. • National Research Foundation (NRF) for the financial support enabling me to complete the study. • The Department of Haematology and Cell Biology for providing the necessary resources and facilities. • My parents, family and friends for their encouragement, prayers and for always believing in me. • My colleagues and friends at the Department of Haematology and Cell Biology for their support, motivation and friendship during this study. • Lastly to my heavenly Father for always being the firm rock beneath my feet. v List of scientific abbreviations and acronyms LIST OF SCIENTIFIC ABBREVIATIONS AND ACRONYMS oC Degree Celsius % Percentage µg Microgram µL Microlitre ADI Acceptable daily intake AMPA Aminomethylphosphoric acid cAMP Cyclic adenosine monophosphate CDK1 Cyclin-dependant kinase 1 CTAB Cetryltrimethylammonium bromide dd Double distilled DNA Deoxyribonucleic acid EDTA Ethylenediamine tetra acetic acid ELISA Enzyme-linked immunosorbent assay EPSPS 5-Enolpyruvylshikimate-3-phosphate synthase et al. Et alia (and others) g Gram GM Genetically modified GMO Genetically modified organism ha Hectare HCl Hydrochloric acid vi List of scientific abbreviations and acronyms HepG2 Hepatocellular carcinoma cell line HMG High Mobility Group gene HT Herbicide tolerant JAr Human choriocarcinoma cell line kg Kilogram L Litre LOD Limit of detection LOQ Limit of quantification m Metre M Molar MCL Maximum contaminant level mg Milligram mL Millilitre mM Millimolar mm2 Square millimetre MRL Maximum residue limit N Normal NaCl Sodium chloride NaOH Sodium hydroxide ND Not detected ng Nanogram vii List of scientific abbreviations and acronyms nm nanometre NOAEL No observed adverse effect level NT Not tested PCR Polymerase chain reaction pH Percentage hydrogen POEA Polyethoxylated tallow amine ppb Parts per billion ppm Parts per million R2 Coefficient of determination RNAse Ribonuclease rpm Revolutions per minute SB Sodium borate TE Tris EDTA UV Ultraviolet V Volt www World wide web viii List of abbreviations and acronyms for organizations, institutions and/or authorities LIST OF ABBREVIATIONS AND ACRONYMS FOR ORGANIZATIONS, INSTITUTIONS AND/OR AUTHORITIES EFSA European Food Safety Authority EU European Union FAO Food and Agriculture Organisation of the United Nations USA FDA United States of America Food and Drug Administration IARC International Agency for Research on Cancer ISAAA International Service for the Acquisition of Agri-biotech Applications JMPR Joint FAO/WHO Meeting on Pesticide Residue NAMC National Agricultural Marketing Council SA NPIC National Pesticide Information Centre PMRA Pest Management Regulatory Agency PRiF Expert Committee on Pesticide Residues in Food SACPA South African Consumer Protection Act SA DAFF South African Department of Agriculture, Forestry and Fisheries SA DOH South African Department of Health SANBI South African National Biodiversity Institute UK United Kingdom UN United Nations USA United States of America USA EPA United States of America Environmental Protection Agency WHO World Health Organisation ix List of figures LIST OF FIGURES Page Figure 3.1 Negative inverted 1% agarose gel image of GM HT negative samples with visible DNA. 37 Figure 3.2 Negative inverted 1% agarose gel image of GM HT negative samples with little or no visible DNA. 38 x List of tables LIST OF TABLES Page Table 1.1 Maximum residue levels (mg/kg) established for glyphosate. 6 Table 1.2 IARC classification system for pesticides. 9 Table 1.3 Summary of studies indicating the toxicity of glyphosate in formulation. 12 Table 2.1 The food products selected for this study. 22 Table 2.2 Copy number standards used in the Real-time PCR quantification of maize events NK603 and GA21. 27 Table 2.3 Copy number standards used in the Real-time PCR quantification of soybean event GTS40-3-2. 28 Table 3.1 Summary of the glyphosate content in maize food products. 31 Table 3.2 Summary of the glyphosate content in soybean food products. 32 Table 3.3 Summary of the glyphosate content in texturized soy protein and corn-soy blend food products. 32 Table 3.4 Detection and quantification of NK603 and GA21 in maize food products. 34 Table 3.5 Detection and quantification of GTS40-3-2 in soybean food products. 35 xi List of tables Table 3.6 Detection and quantification of NK603 in texturized soy protein and corn-soy blend food products. 35 Table 3.7 Detection and quantification of GTS40-3-2 in texturized soy protein and corn-soy blend food products. 35 Table 3.8 Food products that contained glyphosate but no GM HT event. 39 Table 3.9 Fluorometric determination of DNA concentration in GM HT negative samples. 40 Table 3.10 Food products containing GM HT event GTS40-3-2 below the limit of quantification but containing glyphosate. 41 Table 3.11 Food products containing GM HT event NK603 but with no detectable glyphosate. 42 Table 3.12 Theoretical daily intake of glyphosate through food products. 45 Table 4.1 GM HT quantification and GM label. 49 Table 4.2 GM HT quantification of products labelled “non-GMO” or “GMO-free”. 52 xii Preface PREFACE Genetically modified (GM) crops are extensively planted around the world with more than 181 million hectares cultivated in 2015. GM crops are considered to have made a positive contribution to agriculture since their introduction in 1996, especially in terms of crop management. The major GM crops are canola, cotton, maize and soybean that are predominantly engineered to be insect resistant and/or herbicide tolerant. South Africa is considered a major GM crop producing country and planted approximately 2.3 million hectares of GM crops in 2015. Of these crops, approximately 75% (1.73 million ha) was herbicide tolerant. The major herbicide tolerant crops planted in South Africa are maize (approximately 1.2 million ha) and soybean (approximately 508,000 ha). The predominant herbicide used on herbicide tolerant crops is glyphosate. Currently glyphosate is the most widely used herbicide in the world. In 2015, the World Health Organisation International Agency of Research on Cancer (IARC) changed the classification of glyphosate from “possibly carcinogenic to humans” to “probably carcinogenic to humans”. The IARC report, although extensive and in depth, was highly criticised by the agricultural industry. The findings of the IARC on glyphosate differ from regulatory authorities in Europe and the United States of America as well as other international bodies who consider glyphosate safe. There are several possible reasons why the IARC has reached a different conclusion on the safety of glyphosate compared to other internationally recognised bodies: • The IARC only considered documents on the safety of glyphosate that are available in the public domain. Compared to this, other bodies have also taken proprietary documentation provided by the herbicide developer and not available in the public domain into account. • The IARC considered the safety of glyphosate in formulation, which includes surfactants. Compared to this, regulatory bodies assessed the safety of pure glyphosate and not in formulation. xiii Preface • It should be noted that the safety assessment of glyphosate and glyphosate tolerant crops is evaluated separately and not in combination by regulatory authorities. Since the commercialization of herbicide tolerant crops, independent research has generated previously unknown information regarding the application of glyphosate on these crops: • Glyphosate is present in the grain of herbicide tolerant crops treated with glyphosate. • Glyphosate is not removed from food during processing. • Low concentrations of glyphosate in formulation have been found to have genotoxic effects in mammalian cells in vitro. Considering that maize is a major staple and soybean an important source of protein, the safety of glyphosate is an issue of great importance in South Africa. However, before any informed discussion can take place on the safety of glyphosate, we need to know the extent of its presence in the food chain in South Africa, since South Africa predominantly produces glyphosate tolerant maize and soybean. Thus, the aim of this study was to test food products in South Africa containing maize and/or soybean as a primary constituent for glyphosate. The food products were purchased from all major retail stores based on their ingredient list. For ethical reasons no brand names are mentioned in this dissertation also taking into account that the controversy surrounding the safety of glyphosate remains unresolved. It is important to note that this dissertation does not intend to assess the safety of glyphosate but determine whether glyphosate is present in the South African food chain. Care has been taken to present arguments in this dissertation as scientifically as possible with no intention to motivate either for or against the use of glyphosate. To achieve consistency, glyphosate concentrations (either in mg/kg or mg/L) were converted to mg/kg (by using the density of glyphosate 1.75 kg/L) to make data comparison between different studies easier. xiv Preface This dissertation consists of 6 chapters, including a literature review (chapter 1), research aim and methodology (chapter 2), results and discussion for glyphosate content in food products (chapter 3), results and discussion in terms of compliance with GM labelling in South Africa (chapter 4), limitations of the study (chapter 5), as well as a conclusion (chapter 6). The literature review presents the literature regarding glyphosate, its safety assessment and its presence in HT crops, processed food as well as in animals including humans. Chapter 2 includes the research aim and methodology used. Chapter 3 includes the results and discussion for the level of glyphosate present in the maize and soybean food products. Since the data was available, it was used to determine compliance to GM labelling in terms of the Consumer Protection Act (2008) that mandates GM labelling in South Africa. Chapter 4 includes the results and discussion in terms of compliance with GM labelling requirements in South Africa. Chapter 5 includes the limitations of the study. The final chapter (Chapter 6) presents the conclusions over the presence of glyphosate in the South African food chain as well as compliance to mandatory GM labelling. Following chapter 6 there is a summary of the dissertation. xv Chapter 1: Literature review CHAPTER 1: LITERATURE REVIEW 1.1 Introduction to genetically modified organisms A genetically modified organism (GMO) is classified by the World Health Organisation (WHO) as “an organism wherein genetic material has been reformed in a way that does not occur naturally by mating and/or natural recombination” (WHO, 2002). GMOs are developed through genetic engineering which entails manipulating an organism’s genetic material by inserting one or more genes or DNA sequences, or by altering one or more bases of the organism’s genetic code (Paoletti et al., 1996). A GMO has altered DNA, which can either encode for a new protein or modify the function of an existing protein (Paoletti et al., 1996). Genetic modification allows the addition of new properties or “traits” that are not naturally present in the organism (König et al., 2004). The gene inserted into the GMO, also known as the transgene, forms part of a transgene cassette. The transgene cassette contains a promoter, the gene of interest and a terminator (Robinson et al., 2000; Smale and Kadonaga, 2003; Ralston and Shaw, 2008; Lievens et al., 2015). The DNA sequence of the promoter, gene of interest and the terminator can be sourced from various organisms, including bacteria, viruses, plants, fungi and animals (Chawla, 2002). Two methods are commonly used to insert the transgene cassette into the host cell genome (Hansen and Wright, 1999). The first method is particle bombardment which entails the coating of gold or tungsten nanoparticles with the gene cassette (Klein et al., 1987; Segelken, 2010). The coated nanoparticles are then shot into the nucleus of the host cell at high velocity allowing the transgene to be incorporated into the genome of the cell (Klein et al., 1987; Segelken, 2010; Hansen and Wright, 1999). The second method involves the infection of cultured plant cells with Agrobacterium tumefaciens, a plant pathogen with the inherent ability to transfer a particular DNA segment into the nucleus of the host cell allowing its transcription. The inherent ability of Agrobacterium tumefaciens to infect plant cells has allowed it to be used as a vector for genetic manipulation (De la Riva et al., 1998; Gelvin, 1998; Hansen and Wright, 1999; Gelvin, 2003). The point of insertion of the transgene into the genome of a plant cell is random and each insertion is referred to as an “event” (Nester, 2008; Lievens et al., 2015). 1 Chapter 1: Literature review Genetic engineering is extensively utilized in agriculture, to produce crop varieties with improved agricultural traits (König et al., 2004). The major commercial genetically modified (GM) crop traits include insect resistance and herbicide tolerance (James, 2015). Insect resistant plants are modified to produce an endotoxin that kills insect pests. Herbicide tolerant (HT) plants are engineered to be tolerant to the application of herbicides during the growing season to control weeds. Crops have also been modified for other traits, including disease resistance, drought tolerance and improved nutritional content (Gasser and Fraley, 1989; Uzogara, 2000; Nester, 2008; Lievens et al., 2015). The application of GM crops is considered to have resulted in lower use of pesticides, labour, machinery and fuel (Gouse, 2014). In general, GM crops are considered to have made a positive contribution to agriculture since its introduction in 1996, by reducing insect damage and improving crop management (Qaim, 2010). 1.2 GM crop production worldwide GM crops were first commercialised in 1996 in the United States of America (USA) (Mannion and Morse, 2013). By 2014, 28 countries planted more than 181 million hectares (ha) of GM crops (James, 2014). Currently there are four major GM crops produced commercially: canola, cotton, maize and soybean. Other GM crops include alfalfa, papaya, potato, squash, sugar beet and sweet peppers (James, 2013). The ten major GM crop producing countries include the USA (producing 73.1 million ha of GM crop), Brazil (producing 42.2 million ha of GM crop), Argentina (producing 24.3 million ha of GM crop), India (producing 11.6 million ha of GM crop), Canada (producing 11.6 million ha of GM crop), China (producing 3.9 million ha of GM crop), Paraguay (producing 3.9 million ha of GM crop), Pakistan (producing 2.9 million ha of GM crop), South Africa (producing 2.3 million ha of GM crop) and Uruguay (producing 1.6 million ha of GM crop) (James, 2015). GM crops have been commercially planted in South Africa since 1998 (Du Plessis, 2003; SA DAFF, 2005; SANBI, 2010). In 2015, GM crop production in South Africa amounted to 2.3 million ha making it the ninth biggest GM crop producing country in 2 Chapter 1: Literature review the world (James, 2015). It is estimated that 90% of maize (1.8 million ha), 95% of soybean (508,000 ha) and 100% of cotton (12,000 ha) produced in South Africa is GM (James, 2015). The GM traits approved in South Africa for maize include herbicide tolerance (comprising 15.8% or 284,000 ha of GM maize), insect resistance (comprising 30.5% or 550,000 ha of GM maize) and stacked events containing both traits (herbicide tolerance and insect resistance) (comprising 53.4% or 940,000 ha of GM maize) (James, 2015). The only GM trait approved in South Africa for soybean is herbicide tolerance (comprising 100% or 508,000 ha of GM soybean) and for cotton the only approved trait is insect resistance (comprising 100% or 12,000 ha of GM cotton) (James, 2015; SA DAFF, 2015). In South Africa, herbicide tolerance is the major trait in approximately 75% of the GM crops planted. There are currently two HT events approved for maize in South Africa namely GA21 and NK603 and one for soybean, namely GTS40-3-2 (SA DAFF, 2015). HT crops have the ability to tolerate specific broad-spectrum herbicides including glyphosate, glufosinate and 2,4-dichlorophenoxyacetic acid (2,4D) (Benbrook, 2016). HT crops allow the direct application of herbicide to eliminate weeds during the growing season without causing crop damage (Madsen and Streibig, 2003). Glyphosate is the major herbicide used on HT crops worldwide including South Africa (Bonny, 2016). 1.3 Glyphosate use in agriculture Glyphosate is approved in more than 130 countries and is considered the most widely used herbicide in the world (Dill et al., 2010). Since the introduction of GM HT crops in 1996, global use of glyphosate has increased, with a 10 fold increase recorded in 2012 amounting to approximately 720,000 tonnes compared to only 67,078 tonnes in 1995 (Hilton, 2012; Benbrook, 2016). In 2012, South Africa used approximately 40,775 tonnes of glyphosate, mostly in the application on HT crops (Gouse, 2014). Glyphosate is a broad-spectrum, non-selective, systematic herbicide used to kill weeds (Duke et al., 2003; Dill et al., 2010). When applied at lower concentrations, glyphosate is also used as a desiccant (Duke et al., 2003; IARC, 2015). Crop 3 Chapter 1: Literature review desiccation using glyphosate is common practise in the USA. However, the extent of using glyphosate as a desiccant in South Africa is not known. Glyphosate is generally applied by means of directed spray application (USA EPA, 1993). Upon application, glyphosate is absorbed by the foliage and distributed throughout the entire plant (Duke et al., 2003; Arregui et al., 2004). Glyphosate disrupts the shikimate pathway by inhibiting the enzyme 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS), involved in the synthesis of the essential amino acids phenylalanine, tyrosine and tryptophan, which is critical for plant growth (Padgette et al., 1995; Funke et al., 2006). The inhibition of the EPSPS enzyme results in plant death within a matter of days (Franz et al., 1997). HT crops are genetically engineered to express a glyphosate insensitive EPSPS enzyme allowing direct application of glyphosate to selectively kill weeds without crop damage (Dill et al., 2010). It is estimated that approximately 45% of glyphosate produced worldwide is used in the production of HT crops (Dill et al., 2010). However, the continuous application of glyphosate on HT crops has led to the emergence of resistant weeds (Benbrook, 2012). Currently, approximately 31 weed species worldwide have developed resistance to glyphosate (Reinhardt, 2012). As a result of this, glyphosate is being applied at higher concentrations to combat weed resistance (Benbrook, 2012). Recently, Benbrook (2016) reported that the amount of glyphosate applied to HT soybean in the USA increased from 0.7 kg/ha in 1996 to approximately 1.1 kg/ha in 2014. Benbrook (2016) suggested that the upward trend in glyphosate use will likely continue, increasing the glyphosate levels present in the environment and potentially increasing animal and human exposure to the herbicide. 1.4 Glyphosate tolerance levels The maximum residue level (MRL) is the maximum concentration of pesticide residue legally allowed in food or animal feed based on “good” agricultural practice (EFSA, 2009; EFSA, 2015, Codex Alimentarius, 2015). The main purpose of the MRL is to ensure fair practice in international food trade (FAO, 2013). MRLs are used as a regulatory standard to help monitor whether a pesticide is applied as approved. Pesticide residue in food or feed exceeding the MRL indicates misuse of a chemical 4 Chapter 1: Literature review during agricultural application. The MRL, although not considered a safety standard has a direct influence on the amount of pesticide residue present in the food chain. The MRL for a pesticide is established at a level that ensures that the residue levels in food do not exceed the acceptable daily intake (ADI) for the particular pesticide in a country (FAO, 2013). The MRLs are determined by individual countries, as well as by Codex Alimentarius of the Food and Agriculture Organisation of the United Nations (FAO) and WHO (Codex Alimentarius, 2015). Countries which do not have established MRLs for pesticides may use the MRLs as established by Codex Alimentarius (Table 1.1) (FAO, 2013; Codex Alimentarius, 2015). A number of factors are taken into consideration when the MRL are set for a pesticide such as glyphosate in commodities (FAO, 2006). These factors include the minimum effective dose, the standard application dose rate, the time between harvest and consumption and the climatic conditions affecting pesticide efficacy (FAO, 2006). As a result of this, the MRL for a particular pesticide may differ for different commodities. For example, for a crop like bananas, where glyphosate is unlikely to be used either in weed management or as a desiccant, the MRL is 0.05 mg/kg (Table 1.1) (FAO, 2013). However, for crops like wheat, hay or alfalfa where glyphosate is likely to be used as a desiccant or for weed control, the MRL range from 300 mg/kg to 500 mg/kg (Table 1.1) (FAO, 2013). The South African Department of Agriculture, Forestry and Fisheries (SA DAFF) has established glyphosate MRLs for all applicable commodities in South Africa (Table 1.1) (SA DAFF, 2016). The MRLs established for crops in South Africa is similar to those established by Codex Alimentarius, the European Union (EU) and the USA. 5 Chapter 1: Literature review Table 1.1: Maximum residue levels (mg/kg) established for glyphosate. MRLs for glyphosate as established by Codex Alimentarius (Codex), the United States of America Environmental Protection Agency (USA EPA), the European Food Safety Authority (EFSA) and the South African Department of Agriculture, Forestry and Fisheries (SA DAFF). Codex1 USA EPA2 EFSA3 SA DAFF4 Commodities (mg/kg) (mg/kg) (mg/kg) (mg/kg) Banana 0.05 0.2 0.1 0.1 Beans (dry) 2 5 2 2 Sugar cane 2 2 0.1 0.5 Lentil (dry) 5 8 10 NA Peas (dry) 5 8 10 10 Maize 5 5 1 2 Sunflower seed 7 85 20 20 Soybean (dry) 40 20 20 20 Rape seed 20 20 10 20 Wheat bran, Unprocessed 20 30 10 30 Wheat fodder (dry) 300 NA NA NA Hay of grasses (dry) 500 NA NA NA Alfalfa fodder (dry) 500 NA NA NA 1 Codex Alimentarius (www.fao.org/docrep/009/a0209e/a0209e0d.htm) 2 USA EPA (http://www.epa.gov/opp00001/reregistration/REDs/factheets/0178fact.pdf) 3 EFSA (http://www.efsa.europa.eu/ en/topics/topic/pesticides.htm) 4 SA DAFF (http://www.SA DAFF.gov.za/SA DAFFweb3/Branches/Agricultural-Production-Health- Food-Safety/Food-Safety- Quality-Assurance/Maximum-Residue-Limits) Not applicable (NA): No values have been established for particular crops in these countries 6 Chapter 1: Literature review Since the introduction of GM HT crops regulatory authorities have frequently changed the MRLs for glyphosate as a result of increased application. In 1999, the glyphosate MRL for soybean was raised from 0.1 mg/kg to 20 mg/kg in the USA and Europe. Likewise in 2004, the glyphosate MRL for soybean was raised from 0.2 mg/kg to 10 mg/kg in Brazil (Bøhn et al., 2014). Bøhn et al. (2014) suggested that the MRL adjustments were made in response to actual observed increases in the glyphosate residue detected in GM HT soybean. It has been suggested that the recurrent planting of HT crops and glyphosate application on the same fields without crop rotation or the use of different herbicides has contributed to the emergence of resistant weeds. In an effort to combat weed resistance, higher concentrations of glyphosate are applied to HT crops. The increase in the amount of glyphosate sprayed on HT crops has subsequently resulted in an increase in levels of glyphosate residue in HT grains as well as the environment (Benbrook, 2016). 1.5 Safety assessment of glyphosate The commercial formulation of glyphosate contains surfactants which enhance its herbicidal properties. These surfactants facilitate absorption and increase the degree of rainfastness of the herbicide and ensure that it is not washed off by rain or during irrigation (Duke et al., 2003). The United States of America Environmental Protection Agency (USA EPA) does not require safety testing of the surfactants used in pesticides, since they are considered to have no pesticidal properties (Herzfeld and Sargent, 2012). As a result of this, there are no standards on the composition and safety of the non-pesticidal ingredients of pesticide formulations worldwide (Herzfeld and Sargent, 2012). Several studies, reports and reviews on the safety of pure glyphosate have concluded that it is safe for humans if applied at the correct agricultural concentration (Williams et al., 2000). The acute toxicity of glyphosate and its major metabolite aminomethylphosphonic acid (AMPA) has been tested in animal feeding trials and no adverse effects have been found (Williams et al., 2000; Williams et al., 2012). Furthermore, in 1993, the USA EPA classified both glyphosate and AMPA in Category E, which is described as “Evidence of Non-carcinogenicity”, based on the lack of convincing evidence of carcinogenicity in numerous studies (USA EPA, 1993). In 1994, the WHO reaffirmed the findings that there was no evidence that 7 Chapter 1: Literature review glyphosate and AMPA were harmful to humans and that both glyphosate and AMPA had negligible levels of acute toxicity (WHO, 1994). In contrast to studies on pure glyphosate, several studies on glyphosate in formulation (glyphosate with surfactants) have reported toxicity and carcinogenicity. A study by Gasnier et al. (2009) demonstrated that glyphosate in formulation at 5 ppm (5 mg/kg) had toxic effects and resulted in cell death of human liver, umbilical cord and placental cells within 24 hours of exposure. Furthermore, Gasnier et al. (2009) also indicated that glyphosate in formulation at 0.5 ppm (0.5 mg/kg) caused endocrine disruption in human liver cells within 24 hours of exposure (Table 1.3). A more recent study by Belle et al. (2012) determined that 8 mM (1,300 mg/kg) of glyphosate in formulation inhibited the cell replication of human embryonic cells within 24 hours of exposure (Table 1.3). Belle et al. (2012) concluded that the concentration of glyphosate in formulation used in their study was far below the prescribed concentration of 40 mM recommended for herbicide application during agricultural practice. A study by Koller et al. (2012) found that glyphosate in formulation at 10 mg/L to 20 mg/L (5.7 mg/kg to 11.4 mg/kg), caused DNA damage and at 40 mg/L (22 mg/kg) caused membrane damage and mitochondrial impairment in human epithelial cells (Table 1.3). A study by Young et al. (2015) demonstrated that glyphosate in formulation was cytotoxic to human placenta cells at concentrations ranging from 0.005 mM to 0.008 mM (0.85 mg/kg to 1.35 mg/kg) (Table 1.3). Young et al. (2015) confirmed that the surfactants within the glyphosate formulation had a major effect on the toxicity of the herbicide and demonstrated that glyphosate in formulation exhibited similar toxicity at a concentration of 2000 times lower than pure glyphosate. Furthermore, Belle et al. (2012) suggested that glyphosate on its own should not be considered a herbicide, since without surfactants it is not permeable and cannot be absorbed by plant cells. Williams et al. (2012) argued that safety studies on glyphosate in formulation are irrelevant since glyphosate toxicity is as a result of “surfactants present in the formulation” and not due to glyphosate itself. Viljoen (2013) suggested that the argument of Williams et al. (2012) was “irrelevant, since it is the formulation that is being applied to the plant in practice and it is part of the herbicide complex of chemicals taken up by the plant”. 8 Chapter 1: Literature review In 2015, the International Agency for Research on Cancer (IARC) assessed the carcinogenicity of several pesticides, including glyphosate. The IARC changed its classification of glyphosate from “possibly carcinogenic” to “probably carcinogenic” to humans (Table 1.2) (IARC, 2015). The IARC evaluated approximately 260 documents, reports and studies and re-classified glyphosate as a probable human carcinogen based on: limited evidence of carcinogenicity in humans, sufficient evidence of carcinogenicity in experimental animals and strong evidence of genotoxicity and oxidative stress in both human and animal cells (IARC, 2015). The IARC report concluded that glyphosate causes DNA and chromosomal damage in mammalian and non-mammalian cells at low concentrations (IARC, 2015). Table 1.2: IARC classification system for pesticides. IARC classification system Description Group 1 Carcinogenic to humans Group 2A Probably carcinogenic to humans Group 2B Possibly carcinogenic to humans Not classifiable in terms of human Group 3 carcinogenicity Group 4 Probably not carcinogenic to humans IARC/WHO (http://www.iarc/who.monographs.iarc.fr/ENG/Classification/) The IARC report listed approximately 109 studies indicating DNA and chromosomal damage in mammalian and non-mammalian cells as a result of glyphosate in formulation. A study by Alvarez-Moya et al. (2014) reported that glyphosate in formulation at a concentration of 0.12 mg/L (0.069 mg/kg) caused DNA damage in human lymphocytes (Table 1.3). A further study by Roustan et al. (2014) indicated that glyphosate in formulation induced chromosomal breakage in hamster ovary cells at a concentration of only 0.01 mg/L (0.006 mg/kg) (Table 1.3). Similar results were found in fish by Moreno et al. (2014) indicating that glyphosate in formulations caused 9 Chapter 1: Literature review DNA strand breaks in liver and gill cells at a concentration of 0.058 mg/L (0.033 mg/kg) (Table 1.3). DNA and chromosomal damage leads to an increased mutation rate and subsequently an increased risk for developing cancer. The only cancer in humans with a significant link to glyphosate is Non-Hodgkins lymphoma (IARC, 2015). Case- control studies from the USA and Sweden reported a statistically significant increased risk for Non-Hodgkins lymphoma associated with glyphosate exposure (Hardell et al., 2002; De Roos et al., 2003; Eriksson et al., 2008; Orsi et al., 2009). Animal studies have indicated that continuous exposure to glyphosate resulted in a significant increase in the risk for pancreatic islet cell adenoma, renal tubule carcinoma, hepatocellular adenoma and thyroid C cell adenoma in male and female mice (USA EPA, 1986; USA EPA, 1991). Thus as a result, the IARC has concluded that there is limited evidence of carcinogenicity in humans and convincing evidence of carcinogenicity in animals as a result of exposure to glyphosate. The IARC report has been widely criticised. Agricultural companies have claimed that the IARC misinterpreted or incorrectly weighed some of the data it reviewed before classifying glyphosate as a “probable human carcinogen” (Plume, 2015). In April 2016, a group of 16 scientists reviewed the IARC report and produced a detailed critique of the IARC report on glyphosate (Williams et al., 2016). Williams et al. (2016) indicated that their analysis of existing data did not support the IARC’s conclusion that glyphosate is a “probable human carcinogen”. Furthermore, Williams et al. (2016) stated that their review was consistent with previous regulatory assessments and concluded that “glyphosate is unlikely to pose a carcinogenic risk to humans”. In the “declaration of interest” in Williams et al. (2016), 12 of the 16 authors previously served as consultants, or worked for, the companies producing glyphosate. The critique of the IARC report was also funded by a major glyphosate manufacturer. Since the release of the IARC report several regulatory agencies have reviewed the safety of glyphosate. In 2015, the Pest Management Regulatory Agency (PMRA) stated that there was no evidence that glyphosate posed a health risk at the prescribed agricultural dose. Similarly, the European Food Safety Authority (EFSA) re-evaluated the health risk of glyphosate and concluded that “glyphosate is unlikely to pose a carcinogenic hazard to humans and the evidence does not support classification with regards to its carcinogenic potential” (EFSA, 2015). In May 2016, a summary report 10 Chapter 1: Literature review issued after a joint meeting by the FAO and WHO on pesticide residue concluded that glyphosate is unlikely to pose carcinogenic risk to humans through dietary exposure (JMPR, 2016). It should be noted that the regulatory agencies as well as Williams et al. (2016) used an overall weight of evidence approach to reach conclusions regarding the genotoxicity of glyphosate. Taking into consideration that several studies on glyphosate are industry funded and that the majority of published research on glyphosate safety is based on pure glyphosate and not on glyphosate in formulation, the validity of an overall weight of evidence approach is questionable. Recently, Portier et al. (2016) comprising a group of 94 scientists, reviewed the EFSA evaluation of the IARC report on the safety of glyphosate and concluded that there were serious flaws in the EFSA report. Portier et al. (2016) suggested that EFSA dismissed any association of glyphosate with cancer, without clear explanation or justification and ignored important evidence of genotoxicity. Furthermore, Portier et al. (2016) found it problematic that EFSA based their evaluation on the Renewal Assessment Report (RaR) giving almost no weight to published literature while relying heavily on studies provided by the pesticide industry and not available in the public domain. Portier et al. (2016) criticised EFSA and other regulatory bodies that concluded that glyphosate was safe. Furthermore, Portier et al. (2016) suggested that regulatory authorities are under pressure to conclude that glyphosate is safe since it is the major pesticide used in agriculture worldwide. Portier et al. (2016) concluded that the re-classification of glyphosate as a probable carcinogen by the IARC working group accurately reflected the current results of published scientific literature on glyphosate. Considering the extreme viewpoints on glyphosate safety, it is apparent that more research is required on the safety of this herbicide. 11 Chapter 1: Literature review Table 1.3: Summary of studies indicating the toxicity of glyphosate in formulation Minimum Study Tissue studied concentration of Observation glyphosate Human liver cells Umbilical cord cells 5 mg/kg Cell death Gasnier et al., (2009) Placental cells Human liver cells 0.5 mg/kg Endocrine disruption Human embryonic Inhibition of cell Belle et al., (2012) 1,300 mg/kg cells replication Membrane damage Human epithelial 22 mg/kg Mitochondrial cells impairment Koller et al., (2012) Human epithelial 5.7 to 11.4 mg/kg DNA damage cells Human placental Young et al., (2015) 0.85 to 1.3 mg/kg Cytotoxicity cells Alvarez-Moya et al., Human lymphocytes 0.069 mg/kg DNA damage (2014) Chromosomal Roustan et al., (2014) Hamster ovary cell 0.006 mg/kg breakage Fish liver cells Moreno et al., (2014) 0.033 mg/kg DNA strand breaks Fish gill cells 12 Chapter 1: Literature review 1.6 Detection of glyphosate as a result of agricultural application 1.6.1 Detection of glyphosate in HT maize and HT soybean Glyphosate in formulation is applied to GM HT crops one to three times during the growing season to control weeds (Krüger et al., 2014a). After application, glyphosate is absorbed and distributed to all parts of the HT plant tissue (Duke et al., 2003; Robinson, 2009). When glyphosate was initially applied in agriculture it was not known that it would later be detected in the grain of HT crops. Published data on glyphosate residue in HT crops is sparse (Benbrook, 2016). However, studies have detected glyphosate in HT maize from the USA as well as HT soybean from the USA and Argentina (Reddy et al., 2004). The first study detecting glyphosate residue in HT soybean was by Duke et al. (2003). Results from the study indicated that glyphosate was present in the foliage of HT soybean at levels reaching 3.08 mg/kg. A more recent study by Bøhn et al. (2014) tested commercially grown dry soybean samples in the USA and detected glyphosate in all HT soybean samples, at levels reaching up to 3.3 mg/kg. Then (2013) tested 11 HT soybean samples from Argentina and detected glyphosate at levels of up to 26 mg/kg. Results from Then (2013) indicated that some samples contained glyphosate at levels exceeding the MRLs established for soybean in Argentina (20 mg/kg). Then (2013) suggested that this was as a result of glyphosate being applied at concentrations much higher than recommended, most likely due to increasing weed resistance. Furthermore, Then (2013) concluded that the high residue levels within the HT soybean may have a serious health impact through food consumption and suggested that the MRLs for glyphosate in food should be reduced. In 2005, the FAO summarized the findings from 78 trials done on HT maize produced within the USA. Results indicated that glyphosate in maize fodder was detected at 92 mg/kg which was expected due to higher concentrations being sprayed to desiccate the crop. Compared to this, only 2.2 mg/kg of glyphosate was detected in maize grain (FAO, 2005). Overall, the glyphosate levels in HT maize are considered low with all studies detecting glyphosate at below the established MRL for maize in the USA (5 mg/kg). 13 Chapter 1: Literature review In some plant species such as HT soybean, glyphosate is metabolised to its primary metabolite AMPA. In HT maize, no AMPA has been detected, suggesting that glyphosate is not further metabolised in maize (Reddy et al., 2004; Gomes et al., 2014). AMPA is an active metabolite of glyphosate and exhibits phytotoxic ability by disrupting the chlorophyll biosynthesis within plant cells. Then (2013) suggested that since AMPA was an active metabolite of glyphosate, the sum of glyphosate and AMPA should be calculated to determine the residue level per sample. Duke et al. (2003) detected AMPA in HT soybean at levels reaching up to 25 mg/kg. If this value is combined with the glyphosate level detected by Duke et al. (2003) (3.08 mg/kg), the total residue in HT soybean would reach approximately 28.08 mg/kg, exceeding the MRL established for soybean in the USA (20 mg/kg). Similarly, Then (2013) detected AMPA at levels of up to 47 mg/kg in HT soybean from Argentina. The total glyphosate and AMPA levels detected for some HT soybean samples, reached up to 97.4 mg/kg and five of the 11 samples tested contained residue levels exceeding the MRL established for soybean in Argentina (20 mg/kg). These findings suggest that glyphosate is applied at higher rates than what is indicated by the pesticide producer as a result of weed resistance. In South Africa, weed resistance to glyphosate has not been reported extensively and data on this is sparse. As a result, it is not known if weed resistance has influenced the rate of glyphosate application in South Africa. 1.6.2 Glyphosate in processed food products and water Due to its frequent use in agriculture, glyphosate residue is present in various food products as well as in water. Various countries including South Africa have established routine monitoring of pesticides in food. However, this is mostly aimed at fresh fruit, vegetables and grain (Swanepoel, 2014). Limited research has been conducted on the presence of glyphosate in processed foods. Nonetheless, recent studies have detected glyphosate in various processed food products (McQueen et al., 2012; Swanepoel, 2014; Rubio et al., 2014). In 2014, 15.4% (30 out of 195) of bread samples from the United Kingdom (UK) were found to contain glyphosate at concentrations of up to 0.100 mg/kg (PRIF, 2015). Similarly, a study by Swanepoel (2014) in South Africa confirmed that glyphosate was present in 88.0% (seven out of 14 Chapter 1: Literature review eight) of white bread samples tested but did not specify the concentrations of glyphosate detected. Rubio et al. (2014) also detected glyphosate in 60.0% of honey and 36.0% soy sauce products tested from the USA, with concentrations of up to 0.564 mg/kg detected. In a report by EFSA (2009) it was stated that glyphosate cannot be removed from food by washing, processing or cooking. Although the level of glyphosate detected in studies on processed food products were lower than the established MRLs for food commodities, the MRL should not be considered a safety standard but rather a guideline for “good” agricultural practice. Several studies have confirmed that glyphosate is detected in water in several countries. A report by the WHO (2005) found that glyphosate was present in ground water samples at levels of up to 5.15 mg/L (2.94 mg/kg) and 1.7 mg/L (0.97 mg/kg) in Canada and the USA, respectively. More recently, Battaglin et al. (2014) tested more than 3,700 water samples, sourced from 38 states in the USA and indicated that glyphosate was detectable in 53.0% of the water samples, at concentrations of up to 0.470 mg/L (0.269 mg/kg). However, Battaglin et al. (2014) concluded that the glyphosate levels detected in the water were considered low as they were below the maximum contaminant level established for drinking water in the USA (0.7 mg/L) (USA EPA, 2015). The glyphosate levels detected in food and water are considered safe as it is below the MRL (FAO, 2013). However, it is surprising that the tolerated residual levels for glyphosate differ so greatly for different crops and water. For water, the maximum contaminant level of glyphosate is 0.7 mg/L (0.4 mg/kg) (USA EPA, 2015), yet for food products like soybean and maize much higher MRLs (up to 40 mg/kg for soybean and 5 mg/kg for maize) are considered safe in the USA. Both food and water are essential for human survival and are consumed daily. It is questionable why the glyphosate in water is limited to 0.7 mg/L while in crops concentrations of a more than 10 fold are tolerated. This demonstrates the inconsistency when the MRLs for commodities and maximum contaminant level for water are established. There is currently no published data on the levels of glyphosate in HT maize and HT soybean in South Africa. Since HT maize and HT soybean are extensively used in agriculture in South Africa, it is expected that glyphosate is likely to be present in the 15 Chapter 1: Literature review grain. Furthermore, considering that glyphosate cannot be removed from grain by cooking, washing or processing it is likely that glyphosate would be present in food and feed. There is also no data for the levels of glyphosate in water in South Africa. 1.6.3 Glyphosate in animal tissue and excretions Several studies have tested the fate of glyphosate in animals including humans. Research on laboratory animals has confirmed the absorption of glyphosate within the gastrointestinal tract after being fed glyphosate treated feed. Animal feeding studies have also confirmed the distribution of absorbed glyphosate to the tissue of all major organs including the small intestine, kidneys, liver, heart, lungs, blood and bone (Brewster et al., 1991; Krüger et al., 2014a). A study by Brewster et al. (1991) suggested that approximately 35% to 40% of pure glyphosate, administered to Sprague-Dawley rats via oral intubation, was absorbed in the gastrointestinal tract. Brewster et al. (1991) also confirmed the distribution of absorbed glyphosate within the body and detected peak glyphosate levels in the small intestine, kidneys, liver, blood and bone within six hours after application. While Brewster et al. (1991) did not report on the amount of glyphosate detected in each tissue, they did state that the glyphosate levels in all tissues declined rapidly. Furthermore, they suggested that urine and faeces were important routes for glyphosate excretion (Brewster et al., 1991). A more recent study by Krüger et al. (2014a) also reported the presence of glyphosate in the intestine, liver, muscle, spleen, kidneys and urine of dairy cows and fattening rabbits in Germany. The animals were fed feeds including soy, corn and other grains, treated with different concentrations of glyphosate post harvesting (Krüger et al., 2014a). However, the study did not determine the amount of glyphosate absorbed or the extent of glyphosate excretion. Krüger et al. (2014b) also investigated the reason for the severe malformation in piglets on a Danish pig farm and suggested that glyphosate may be a contributing factor. Tissue samples from piglets born with severe malformations were tested and glyphosate was detected in all samples. Kruger et al. (2014b) concluded that ingested glyphosate is transferred across the placental barrier during sow 16 Chapter 1: Literature review pregnancy, but that further studies are necessary to confirm or exclude the role of glyphosate in piglet birth malformation. Although not proven, the possibility of glyphosate bio-accumulation within animal tissue cannot be ignored. It has been suggested that ingested glyphosate is rapidly excreted from the body by means of urine and faeces (Brewster et al., 1991). However, it is not known whether the glyphosate absorbed in organ tissue follows the same excretion rate. Furthermore, the detection of glyphosate in animal tissue as a result of being fed glyphosate treated feed indicates that it is transferred within the food chain. Studies on the fate of glyphosate in humans are limited to urine testing. Studies have confirmed that glyphosate is detectable in the urine of humans from farming and urban communities in the USA and Europe. One of the earliest studies testing for glyphosate in human urine was done by Acquavella et al. (2004). They tested urine samples of 127 individuals from farms in the USA and reported that sixty percent had detectable levels of glyphosate with concentrations of up to 233 µg/L (0.133 mg/kg) (Acquavella et al., 2004). It was suggested that the glyphosate in the urine samples was as a result of occupational exposure during agricultural application (Acquavella et al., 2004). A similar study by Curwin et al. (2007) analysed the urine samples of individuals from farming communities in USA and as a control group, used urine samples of individuals from non-farming communities. Curwin et al. (2007) reported that glyphosate was detected in the majority of samples (60% of adults and 80% of children) with concentrations of up to 18 µg/L (0.010 mg/kg) and that there was no significant difference in urinary glyphosate concentration between individuals from farming or non-farming communities. The findings of Curwin et al. (2007) suggest that other sources of glyphosate exposure should be considered as non-farming individuals are not exposed to glyphosate as a result of occupational application. Several studies have confirmed the presence of glyphosate in grain (FAO, 2005, Then, 2013, Bøhn et al., 2014), processed food (Rubio et al., 2014; Swanepoel, 2014), animal tissue (Krüger et al., 2014 a/b) and water (Battaglin et al., 2014) which all contribute substantially to the human diet. These findings suggest that diet may contribute a greater role in exposing individuals to glyphosate than initially thought. Furthermore, considering the results from animal studies detecting glyphosate in the tissue of all 17 Chapter 1: Literature review major organs after being fed glyphosate treated feed (Brewster et al., 1991; Krüger et al., 2014a), it can be argued that a similar result may be expected in human tissue. With the recent re-classification of glyphosate as a “probable human carcinogen”, residue within the human body, whether from occupational exposure or dietary intake, is concerning and needs further investigation. 1.7 Conclusion GM crops were introduced for commercial planting in 1996 and are considered to have made a positive contribution to crop management in agriculture over the last two decades (Qaim, 2010). Herbicide tolerance is the predominant trait in all GM crops and accounts for approximately 85% of GM crop production worldwide (James, 2015). South Africa is a major GM crop producing country and 75% of GM crops are herbicide tolerant (James, 2015). The most widely used herbicide in the world is glyphosate, which is also used on GM HT crops (Bonny, 2016). In recent years, several studies have confirmed the absorption of glyphosate by GM HT plants, including maize and soybean, and its distribution to all parts of the plant tissue including the grain (Duke et al., 2003; Robinson, 2009; Bøhn et al., 2014). When GM HT crops were initially commercialized it was not known that glyphosate would be detectable in the grain of the crops treated with the herbicide. The levels of glyphosate detected in the grain of HT maize and HT soybean is generally below the MRLs established for these commodities. However, several recent studies have shown that glyphosate in formulation exhibits genotoxic effects at concentrations similar to what has been detected in HT grain (Alvarez-Moya et al., 2014; Roustan et al., 2014; Young et al., 2015). Several animal feeding studies have confirmed that glyphosate is absorbed and can be detected in all major organs of animals after exposure to glyphosate treated feed (Brewster et al., 1991; Krüger et al., 2014a). Brewster et al. (1991) estimated that approximately 30% to 40% of ingested glyphosate is absorbed in the gastrointestinal tract and is rapidly excreted from the body by means of urine and faeces (Brewster et al., 1991). However, it is not known whether the glyphosate absorbed in organ tissue 18 Chapter 1: Literature review follows the same excretion rate and research is needed to determine whether there is any bio-accumulation of glyphosate in animal tissue. Several studies have detected glyphosate in human urine (Acquavella et al., 2004; Curwin et al., 2007). Since these studies focused mostly on farming communities, it was suggested that the glyphosate levels present in urine was as a result of occupational exposure (Acquavella et al., 2004). However, a recent study investigating the urinary glyphosate concentrations of farming and non-farming households concluded that glyphosate was detected in the urine of both groups at similar concentrations. Furthermore, a limited number of studies have also detected glyphosate at low levels in processed food products as well as in water (Battaglin et al., 2014; Rubio et al., 2014; Swanepoel, 2014). The latter results suggest that there may be other sources of glyphosate exposure either through diet or water intake. In South Africa, maize (in the form of maize meal) is a major staple and soybean an important source of protein. The majority of GM maize (68%) and soybean (100%) grown in South Africa is HT (James, 2015). It is currently not known to what extent glyphosate may be present in the South African food chain, specifically regarding maize and soybean containing food products. Considering the potential uncertainty regarding the safety of glyphosate, it is important to clarify the extent of the presence of glyphosate in South African food products. 19 Chapter 2: Research aim and methodology CHAPTER 2: RESEARCH AIM AND METHODOLOGY 2.1 Rationale Maize is a major staple and soybean an important source of protein in South Africa. It is estimated that 61% of maize and 95% of soybean produced in South Africa is HT (James, 2015). The herbicide most widely used to treat HT crops is glyphosate. Studies have shown that after application, glyphosate is absorbed by HT plants and distributed to all plant tissue including grain (Duke et al., 2003; Then, 2013). Animal studies have demonstrated that glyphosate is absorbed from feed and detected in the tissue of all major organs (Brewster et al., 1991; Krüger et al., 2014a; Krüger et al., 2014b). However, up until 2015, glyphosate was considered safe for humans and the environment. In 2015, the IARC changed its classification of glyphosate from “possibly carcinogenic” to “probably carcinogenic” to humans. The IARC based the re- classification of glyphosate on accumulating evidence showing the potential genotoxic properties of glyphosate (IARC, 2015). It is estimated that approximately 500 g of cooked maize meal is consumed per person daily in poor households in South Africa (Payne, 2011). In addition to this, soybean is an important source of protein and is added to various food products to increase the protein content. It is currently not known to what extent glyphosate is present in food products in South Africa, of which maize or soybean are the primary constituent. Taking into consideration that glyphosate may be potentially carcinogenic and is not removed from food by washing, cooking or processing (EFSA, 2009), it is important to know to what extent glyphosate is present in food products in South Africa. 2.2 Aim of Study The primary aim of this study was to determine whether glyphosate is present in commercially available food products in South Africa that contain maize or soybean as the major constituent. The secondary aim was to detect and quantify the presence of GM HT events in maize (NK603 and GA21) and soybean (GTS40-3-2) food products. A minor aim was to use the available data to evaluate the maize and 20 Chapter 2: Research aim and methodology soybean food products in terms of compliance with the South African Consumer Protection Act (SACPA) regarding GM labelling. 2.3 Study Design This study was performed as an analytical experimental study. Commercially available food products which contained maize or soybean as a primary constituent were identified and purchased from retail stores in South Africa. The enzyme-linked immunosorbent assay (ELISA) was used to test all products for the presence and levels of glyphosate. Furthermore, event specific Real-time polymerase chain reaction (PCR) screening was used to determine whether the products contained GM HT events approved in South Africa for maize and soybean. Food products positive for one or more of the GM HT events were quantified using Real-time PCR in order to determine the percentage GM HT event present in each product. Finally, the data from HT event quantification was used to evaluate whether the products tested were compliant with the South African Consumer Protection Act (2008) in terms of GM labelling. 2.4 Product selection and sampling A total of 81 food products were selected from retail outlets including Pick ‘n Pay, Shoprite, Checkers, Spar, Dischem and Woolworths according to product availability during 2015. Products were selected to include as many different product brands as possible (Table 2.1). During sampling, only products which contained maize and/or soybean as the major constituent in raw or processed form were selected. Products were arranged into three categories using their ingredients list as a guideline: samples containing maize, samples containing soybean and samples containing both maize and soybean as a primary constituent. The texturized soy protein products and corn- soy blends listed both maize and soybean as a primary constituent and were tested for both HT maize and soybean events. For soybean, infant milk and soy flour, only one brand was commercially available. 21 Chapter 2: Research aim and methodology Table 2.1: The food products selected for this study. Number of Sample category products Maize products Maize meal 20 Instant maize meal 2 Beer powder 2 Maize grits 5 Maize rice 3 Polenta 5 Corn flakes 7 Corn chips 10 Maize pasta 3 Total 57 Soybean products Soybeans 1 Soy milk 8 Infants milk 1 Soy flour 1 Total 11 Texturized soybean protein products Texturized soy protein (containing only soybean) 4 Texturized soy protein (containing maize and soybean) 3 Total 7 Corn-soy products Corn-soy blend1 6 Total 6 1 Corn-soy blends are precooked cereals containing milled maize and soybean 22 Chapter 2: Research aim and methodology 2.4.1 Inclusion criteria Food products in which maize and/or soybean was the primary constituent of the products based on the ingredient list. 2.4.2 Exclusion criteria Food products in which maize and/or soybean was not the primary constituent of the products based on the ingredient list. 2.5 Methodology for glyphosate screening 2.5.1 Sample preparation and glyphosate determination Glyphosate was detected and quantified using an ELISA kit according to the manufacturer’s instructions (Abraxis, USA) (Rubio et al., 2003). Samples were homogenized where necessary, using a food blender to a maximum particle size of 2.5 mm2. Sample preparation included the addition of 10 mL of 1N HCl to 1 g of sample, followed by vortexing for 2 minutes. For soy milk, 900 µL of 1N HCl was added to 100 µL of soy milk followed by vortexing for 2 minutes. After vortexing, samples were incubated for 5 minutes at room temperature, thereafter 1 mL of the sample supernatant was retained and centrifuged for 5 minutes at 6,000 rpm. After centrifugation, samples were diluted by adding 40 µL of sample to 4 mL of glyphosate sample diluent (supplied in the kit) followed by vortexing for 20 seconds. Sample derivatization followed by the addition of 1 mL of glyphosate assay buffer (supplied with the kit) and 100 µL of diluted glyphosate derivatization reagent (supplied with the kit) to 250 µL of sample, standards and control. Thereafter, the samples were vortexed for 20 seconds and the mixture incubated for 10 minutes at room temperature. The samples were tested in triplicate and each assay included five standards (0, 0.075, 0.2, 0.5, 1 and 4 parts per billion) and a control (0.75 parts per billion) in duplicate. Each standard/control/sample (50 µL) was added to the microtiter plate, followed by the addition of 50 µL glyphosate antibody solution (supplied in the kit). The plate was 23 Chapter 2: Research aim and methodology incubated on a rotary shaker at 100 rpm at room temperature for 30 minutes. After incubation, 50 µL of glyphosate conjugate solution (supplied with the kit) was added to each well followed by incubation on a rotary shaker at 100 rpm at room temperature for 60 minutes. After incubation, the content of the plate was discarded and each well was washed three times by the addition of 250 µL of 1x wash buffer (supplied in the kit). Colour solution (150 µL) (supplied in the kit) was then added to each well and the plate incubated at room temperature for 20 minutes. After incubation, 100 µL of stop solution (supplied in the kit) was added to each well. The absorbance was read within 15 minutes of adding the stop solution, using the BioTek Synergy HT Plate Reader at 450 nm. Compensation for background noise was done by subtracting the mean optical density of the blank (zero standard) containing only reagent, from each sample reading. The standards were used to generate a Four Parameter Logistic curve (R2 > 0.98) (Appendix C), which was used to determine the glyphosate concentration in each sample. Samples with a mean result above the highest standard were diluted as necessary and the assay repeated. The assay had a limit of detection (LOD) of 0.075 parts per billion (ppb) and samples containing glyphosate at levels below the LOD were considered negative according to the manufacturer’s instructions for analysis. 2.6 Methodology for DNA screening 2.6.1 Sample preparation and DNA extraction Samples were homogenized (where necessary), using a food blender to a maximum particle size of 2.5 mm2. DNA extraction was performed in duplicate from homogenized samples using the cetyltrimethylammonium bromide (CTAB) method with some modification (Lipp et al., 1999). Each extraction included an extraction control containing only reagents to ensure that no reagents were contaminated. DNA was extracted from duplicate 2 g samples by the addition of 10 mL CTAB (pH 8.0) [0.11M CTAB, 0.03M EDTA, 2.8M NaCl, 0.2M Tris] and 30 µL proteinase K [20 mg/mL]. After incubation at 60°C for at least 2 hours, 1.5 mL sample/CTAB solution was added to 50 µL RNase [20 mg/mL] and further incubated at 60°C for 15 minutes. Following this, 900 µL of sample supernatant was added to 600 µL of chloroform and the mixture was vortexed and then centrifuged at 14,000 rpm for 10 minutes. The 24 Chapter 2: Research aim and methodology aqueous phase was retained to which 500 µL isopropanol and 2 µL glycogen was added, followed by incubation for 30 minutes at room temperature to precipitate the DNA. The DNA pellet was retained and washed with 500 µL of 75% ethanol followed by centrifugation at 14,000 rpm for 5 minutes. The DNA was then dissolved in 50 µL of 0.1x TE buffer (pH 8.0) [1M Tris and 0.5M EDTA]. The DNA was further purified using Eurofins GeneScan micro-spin columns according to the manufacturer’s instructions. The extracted DNA was stored at 4°C until used. 2.6.2 Gel electrophoresis and fluorometry Samples negative for an HT event were subjected to agarose gel electrophoresis in order to exclude the possibility of a false negative result due to a failed DNA extraction. Samples that screened positive for GM HT events were not subjected to electrophoresis, but quantified and the quality of extracted DNA was evaluated by the copy number of the High Mobility Group (HMG) (for maize) or Lectin (for soybean) gene. Sample DNA (5 µL) was added to 10 µL of blue dye loading buffer [60% Glycerol, 0.5M EDTA (8.0 pH) and bromophenol blue] in duplicate and gel electrophoresis performed using a 1% agarose gel in sodium borate (SB) buffer (pH 8.0) [10mM NaOH, 30mM boric acid]. Lambda DNA (7 µL) [50 ng/µL] was used to evaluate the size of the extracted DNA. The agarose gel was run at 250V for 15 to 20 minutes, followed by staining in ethidium bromide solution [10 mg/mL ethidium bromide] on a rotary shaker at 50 rpm for 15 minutes. The DNA was visualized and documented under UV light using the GelLogic200 (Kodak) system. The concentration of extracted DNA in products that tested negative for GM HT events but that contained glyphosate was determined fluorometrically using the Qubit dsDNA HS Assay Kit (Invitrogen) according to the manufacturer’s instructions. Two calibration standards, at 0 ng/µL and 10 ng/µL respectively, were prepared by the addition of 10 µL of each standard to 189 µL of Qubit dsDNA HS Buffer and 1 µL of Qubit dsDNA HS Reagent provided in the kit. The concentration of DNA was determined by the addition of 1 µL of extracted DNA to 198 µL of Qubit dsDNA HS Buffer and 1 µL Qubit dsDNA HS Reagent. The mixture was vortexed and centrifuged, followed by 25 Chapter 2: Research aim and methodology incubation at room temperature for 2 minutes before measuring the concentration of DNA using the Qubit fluorometer (Invitrogen). 2.6.3 Real-time PCR screening for the presence of GM HT events in food products The extracted DNA was used to screen samples for the presence of glyphosate tolerant events approved in South Africa for maize and soybean, respectively. Event specific qualitative Real-time PCR detection was used to screen for events NK603 and GA21 in maize and event GTS40-3-2 in soybean. The Real-time PCR reaction mixture consisted of 5 µL DNA and 20 µL of Real-time PCR master mix (7.5 µL oligo mix and 12.5 µL basic mix) (Eurofins GeneScan, Germany). A synthetic PCR target was included in the Real-time PCR master mix reaction as internal inhibition control for each sample. All reactions were performed in duplicate including positive, negative and extraction controls. An inhibition control was included Samples as well as all the positive, negative and extraction controls were subjected to the following thermal cycling conditions: an initial denaturation and enzyme activation step of 95°C for 10 minutes, followed by 45 cycles of 95°C for 15 seconds and 60°C for 90 seconds. The real-time PCR reaction was performed on an Mx3005P Cycler (Agilent Technologies) with a LOD of 10 copies. Fluorescence data was collected during the annealing/elongation step at 60°C. Products were scored for the presence or absence of the events, only if duplicate results were uniform. To minimize the risk of cross- contamination, individual steps including sample preparation, DNA extraction, PCR setup and PCR were performed in separate laboratories. 2.6.4 Real-time PCR quantification of GM HT events detected in food products The GM HT event(s) found to be present in DNA samples was quantified by using the GMO Quant NK603 Corn and GMO Quant GA21 Corn kits for maize products and the GMO Quant RoundupReady Soy kit for soybean products, respectively (Eurofins GeneScan). The kits included real-time PCR master mixes with respective primers and probes, four copy number standards each for target and reference genes (Table 26 Chapter 2: Research aim and methodology 2.2 and 2.3), respectively and a 1% GM certified reference control (ERM-BF415d for NK603, ERM-BF414d for GA21 and ERM-BF410dn for GTS40-3-2) (JRC-IRMM, 2017). The copy number standards were used to generate a standard curve consisting of four data points in duplicate with a minimum correlation of 0.98 and a limit of quantification (LOQ) of 0.05%. The total NK603 and GA21 content in the maize products was quantified in relation to the HMG reference gene. The total GTS40-3-2 content in the soybean products was quantified in relation to the Lectin reference gene. The Real-time PCR thermal cycling conditions was as follows: an initial denaturation and enzyme activation step of 95°C for 10 minutes, followed by 45 cycles of 95°C for 15 seconds and 60°C for 90 seconds. The Real-time PCR reaction mixture consisted of 5 µL DNA and 20 µL of Real-time PCR master mix. Each reaction was performed on the ABI 7500 FAST Real-time PCR system (Applied Biosystems). Two dilutions of each sample in duplicate (1:2 and 1:8 for unprocessed samples; 1:4 and 1:8 for processed samples) were used to test for sample inhibition. In addition to this, a no template control was included with each assay to test for contaminated reagents. Products were considered to be negative for a GM HT event if the maize and soybean references gene was detected at sufficient copy number. Table 2.2: Copy number standards used in the Real-time PCR quantification of maize events NK603 and GA21. HMG NK603 and GA21 Standards (copies/5 µL) (copies/5 µL) 1 81,920 copies 10,240 copies 2 10,240 copies 1,280 copies 3 1,280 copies 160 copies 4 160 copies 40 copies 27 Chapter 2: Research aim and methodology Table 2.3: Copy number standards used in the Real-time PCR quantification of soybean event GTS40-3-2. Lectin GTS40-3-2 Standards (copies/5 µL) (copies/5 µL) 1 86,400 copies 8,600 copies 2 14,400 copies 1,440 copies 3 2,400 copies 240 copies 4 400 copies 40 copies 2.7 Data analysis The mean amount of glyphosate and percentage of GM HT event present in each sample as well as the standard deviation was calculated using Excel 2010 (Windows 8). Linear regression was performed to determine whether a correlation existed between the percentage of GM HT DNA and level of glyphosate in each sample. To achieve consistency glyphosate concentrations (either in mg/kg or mg/L) were converted to mg/kg (by using the density of glyphosate 1.75 kg/L) (Dill et al., 2010) to allow for comparison between different studies as well as to the MRL for glyphosate. 2.8 Compliance to mandatory GM labelling in South Africa Data obtained from the quantification of GM HT events in the maize and soybean food products included in this study was used to check for compliance with regards to GM labelling. Products were evaluated in terms of current GM labelling requirements under the Consumer Protection Act (2008): “genetically modified” (above 5% GM), no GM labelling required (below 5% GM) or “non-GMO” (below 1% GM). This was done by comparing the percentage GM HT event to the GM label in the ingredients list of each product used in this study. 28 Chapter 3: Results and discussion for glyphosate content in food products CHAPTER 3: RESULTS AND DISCUSSION FOR GLYPHOSATE CONTENT IN FOOD PRODUCTS 3.1 Detection of glyphosate in maize and soybean food products Eighty-one off-the-shelf maize and soybean food products were tested for glyphosate. Of these products, 57 indicated maize as a primary constituent, 11 specified soybean as a primary constituent, while 13 contained both maize and soybean as a primary constituent (corn-soy blends and texturized soy protein products). Of all the products tested, 54 (66.70%) contained glyphosate in a range of 27 to 2,257 parts per billion (ppb) (0.027 to 2.26 mg/kg). Of the 57 maize products, 30 (52.63%) contained glyphosate in a range of 27 to 95 ppb (0.027 to 0.095 mg/kg) (Table 3.1). All 11 soybean products contained glyphosate in a range of 27 to 142 ppb (0.027 to 0.142 mg/kg) (Table 3.2). Of the six corn-soy blends, all tested positive for glyphosate in a range of 43 to 65 ppb (0.043 to 065 mg/kg) (Table 3.3). All seven texturized soy protein products tested positive for glyphosate in a range of 41 to 2,257 ppb (0.041 to 2.26 mg/kg) (Table 3.3). These findings confirm that glyphosate is present in South African food products containing maize and soybean. The levels of glyphosate detected in the maize products in this study are low when compared to the MRL for glyphosate in maize in South Africa (2 mg/kg). The highest level of glyphosate detected in the maize products (95 ppb or 0.095 mg/kg) was 4.75% of the MRL (Table 1.1). Similarly, none of the soybean products contained levels of glyphosate above the MRL established for soybeans in South Africa (20 mg/kg) and the highest level detected (142 ppb or 0.14 mg/kg) was 0.71% of the MRL (Table 1.1). The levels of glyphosate detected in the corn-soy blends and texturized soy protein products (that contained maize and soybean) were compared to both the MRLs for maize and soybean respectively, since there is no MRL for glyphosate in composite and/or processed products. The highest level of glyphosate detected in the corn-soy blends (65 ppb or 0.065 mg/kg) was below the MRL for maize and soybean (3.25% of the MRL for glyphosate in maize and 0.33% of the MRL for glyphosate in soybean). The highest level of glyphosate detected in texturized soybean protein (containing both maize and soybean) (2,257 ppb or 2.257 mg/kg) exceeded the MRL in maize but was 29 Chapter 3: Results and discussion for glyphosate content in food products below the MRL in soybean (112.85% of the MRL for glyphosate in maize and 11.30% of the MRL for glyphosate in soybean). No other texturized soy protein products contained levels of glyphosate exceeding the MRLs for maize or soybean. With the exception of one texturized soy protein product, the level of glyphosate in all products included in this study was below the MRLs established for maize and soybean in South Africa. The results from this study suggest that the correct application rate of glyphosate is practiced in South African agriculture. As previously stated in this dissertation, the MRL for a pesticide is based on agricultural practice. The MRL for glyphosate in soybean is 10 fold higher than maize simply due to the fact that HT soybean is treated more frequently with the herbicide during the growing season than HT maize. Over the last two decades the MRL for glyphosate in soybean has been increased several times. Prior to 1999 the MRL for glyphosate in soybean in the USA and Europe was 0.1 mg/kg. In 1999, it was increased 200 fold in the USA and Europe to 20 mg/kg. Bøhn et al. (2014) suggested that the MRL for glyphosate in soybean was increased due to the higher levels of herbicide in this commodity. These changes confirm the inconsistency when establishing the MRL for a pesticide. Although the level of glyphosate detected in the maize and soybean food products in this study was below the MRL, it should be noted that a MRL cannot be considered as an indication of safety. 30 Chapter 3: Results and discussion for glyphosate content in food products Table 3.1: Summary of the glyphosate content in maize food products. Number of samples Range of glyphosate in Range of glyphosate in Sample category Number of Samples testing positive for samples samples glyphosate (ppb) (mg/kg) Maize meal 20 20 27 to 93 0.027 to 0.093 Instant maize meal 2 2 34 to 35 0.034 to 0.035 Beer powder 2 2 40 to 95 0.040 to 0.095 Maize grits 5 0 ND ND Maize rice 3 3 28 to 65 0.028 to 0.065 Polenta 5 0 ND ND Corn flakes 7 0 ND ND Corn chips 10 0 ND ND Maize pasta 3 3 47 to 62 0.047 to 0.062 ND: Not detected 31 Chapter 3: Results and discussion for glyphosate content in food products Table 3.2: Summary of the glyphosate content in soybean food products. Number of samples Range of glyphosate in Range of glyphosate in Sample category Number of samples testing positive for samples samples glyphosate (ppb) (mg/kg) Soy milk 8 8 32 to 142 0.032 to 0.142 Soybeans 1 1 49 0.049 Infants milk 1 1 58 0.058 Soy flour 1 1 27 0.027 Table 3.3: Summary of the glyphosate content in texturized soy protein and corn-soy blend food products. Number of samples Range of glyphosate in Range of glyphosate in Sample category Number of samples testing positive for samples samples glyphosate (ppb) (mg/kg) Texturized soy protein 7 7 195 to 2,257 0.195 to 2.257 Corn-soy blend 6 6 43 to 65 0.043 to 0.065 32 Chapter 3: Results and discussion for glyphosate content in food products 3.2 Detection and quantification of GM HT events in food products The majority of maize and soybean food products (70.37%) contained one or more GM HT event. Of the 57 maize products, 44 (77.19%) tested positive for at least one GM HT maize event (NK603 and/or GA21) (Table 3.4). The most common GM HT maize event was NK603, detected in 44 out of 57 maize products (77.19%) in a range of 0.25% to 100.00%. GM HT maize event GA21 was detected in only one maize product at 0.72% (Table 3.4). The higher incidence of event NK603 compared to event GA21 is not surprising, since event NK603 was introduced for commercial production in South Africa in 2002 whereas event GA21 was only introduced for planting in 2010 (ISAAA, 2015). For soybean, GTS40-3-2 is the only GM HT event approved in South Africa. Of the 11 soybean products included in this study, three (27.27%) tested positive for the GM HT soybean event GTS40-3-2 in a range of 0.07% to 9.57% (Table 3.5). The corn-soy blends and texturized soy protein products were tested, for both maize and soybean events, according to their ingredient list. Of the six corn-soy blends, five (83.33%) tested positive for NK603 and/or GTS40-3-2. Event NK603 was detected in three out of six corn-soy blends in a range of 0.05% to 16.70% (Table 3.6). Event GTS40-3-2 was detected in three out of six corn-soy blends in a range of 0.14% to 48.65% (Table 3.7). Of the seven texturized soy protein products, five (71.43%) tested positive for NK603 and/or GTS40-3-2. Event NK603 was detected in three of the seven texturized soy protein products in a range of 27.75% to 54.38%, while GTS40-3-2 was present in five out of seven texturized soy protein products in a range of 16.75% to 92.09% (Table 3.6 and Table 3.7). Three products contained GM HT events at below the LOQ (0.05%) (Appendix B). No GM HT event was detected in 21 out of 81 food products (25.9%) (Appendix B). 33 Chapter 3: Results and discussion for glyphosate content in food products Table 3.4: Detection and quantification of NK603 and GA21 in maize food products. Number of Number of Number of Percentage range Percentage Number of samples samples samples Sample category of NK603 in range of GA21 samples containing HT containing containing samples in samples events event NK603 event GA21 Maize meal 20 19 19 1 1.28% to 100.00% 0.72 Instant maize meal 2 1 1 0 80.61% ND Beer powder 2 2 2 0 26.24% to 40.47% ND Maize grits 5 5 5 4 16.39% to 55.18% Below LOQ Maize rice 3 3 3 1 9.72% to 67.47% Below LOQ Polenta 5 4 4 0 0.25% to 53.65% ND Corn flakes 7 0 0 0 ND ND Corn chips 10 9 9 0 10.76% to 41.65% ND Maize pasta 3 1 1 0 21.15% ND Below LOQ: refers to a GM HT event being detected, but below 0.05% ND: Not detected 34 Chapter 3: Results and discussion for glyphosate content in food products Table 3.5: Detection and quantification of GTS40-3-2 in soybean food products. Number of samples Percentage range of Sample category Number of samples containing HT event GTS40-3-2 detected GTS40-3-2 Soybeans 1 0 ND Soy milk 8 2 5.21% to 9.57% Infants milk 1 1 0.07% Soy flour 1 0 ND ND: Not detected Table 3.6: Detection and quantification of NK603 in texturized soy protein and corn- soy blend food products. Number of samples Sample category Number of samples containing HT event NK603 detected NK6031 Corn-soy blend 6 3 0.04% to 16.70% Texturized soy 7 3 27.75% to 54.38% protein 1 Event NK603 was the only HT maize event detectable in texturized soy protein products and corn-soy blends tested in this study, while event GA21 was not detected in these products Table 3.7: Detection and quantification of GTS40-3-2 in texturized soy protein and corn-soy blend food products. Number of samples Sample category Number of samples containing HT event GTS40-3-2 detected GTS40-3-2 Corn-soy blend 6 3 0.14% to 48.65% Texturized soy 7 5 16.75% to 92.09% protein 35 Chapter 3: Results and discussion for glyphosate content in food products 3.3 Correlation between percentage GM HT event and level of glyphosate in food products Since the products were tested for both glyphosate and GM HT events, a linear regression was used to evaluate whether a correlation existed between the percentage GM HT event and level of glyphosate. There was no correlation between the percentage GM HT event(s) and the level of glyphosate (R2 = 0.17). Moreover, some of the products that contained glyphosate did not contain detectable GM HT events. It is possible that glyphosate was used as a desiccant on crops that did not contain an HT event (Duke et al., 2003; IARC, 2015). It is also possible that the products that tested negative for GM HT events were false HT negative as a result of processing thereby degrading the DNA. After further analysis, by excluding products that tested negative for GM HT events, still no correlation was observed (R2 = 0.16). Thus, it can be concluded that there is no correlation between the GM HT content and the level of glyphosate in a product. Of the 81 products analysed for glyphosate content, 12 contained glyphosate in a range of 27 ppb to 61 ppb (0.027 mg/kg to 0.061 mg/kg) even though they did not contain a detectable GM HT event (Table 3.8). As previously stated, products that tested negative for GM HT events could have been as a result of DNA degradation during food processing (Bauer et al., 2003). However, the products that tested negative for GM HT events, including an instant maize meal, maize meal, soybeans and soy flour, for which there was detectable DNA after gel electrophoresis, suggest that these products were not false negative due to processing (Figure 3.1). Furthermore, PCR inhibition can also be excluded for samples testing negative for GM HT events, since all products were tested at two dilutions. The remaining samples consisted of two maize pastas, one corn-soy blend, four soy milks and a texturized soy protein product from which little or no visible DNA was evident after gel electrophoresis (Figure 3.2). Considering the sensitivity of agarose gel electrophoresis it is possible that low amounts of degraded DNA were present in these samples and fluorometry was used to determine the concentration of DNA (Table 3.9). The highest concentration of DNA in these samples was 0.449 ng/µL which is low and may explain the negative PCR results. A further three products contained event 36 Chapter 3: Results and discussion for glyphosate content in food products GTS40-3-2 at below the LOQ (0.05%) but had a level of glyphosate in a range of 32 ppb to 40 ppb (0.032 mg/kg to 0.040 mg/kg) (Table 3.10). The Real-time PCR quantification results indicated that the Lectin gene was present in sufficient copy numbers in these samples. This confirmed that these products contained low amounts of the event GTS40-3-2 (Table 3.10). Some products tested positive for a GM HT event(s) but did not contain any glyphosate. It was found that 21 products contained GM HT events in a range of 0.25% to 67.47%, without containing glyphosate (Table 3.11). Since glyphosate is not removed from grain or food by washing, cooking or processing (EFSA, 2009), it is likely that the GM HT crops used to produce these products was not treated with the herbicide. The majority of maize and soybean produced in South Africa is cultivated on dryland. Thus, the emergence of weeds (and hence the need to apply glyphosate) would depend on several factors including rainfall and/or agricultural practice. These results suggest that the application of glyphosate is likely to differ from season to season. Figure 3.1: Negative inverted 1% agarose gel image of GM HT negative samples with visible DNA. The extracted DNA for samples 6 (instant maize meal), 11(maize meal), 64 (soybeans) and 74 (soy flour) in duplicate were resolved on a 1% agarose gel. All samples indicated visible DNA. Lambda DNA [50 ng/µL] (lanes marked M) was used to evaluate the quality of the extracted sample DNA. 37 Chapter 3: Results and discussion for glyphosate content in food products Figure 3.2: Negative inverted 1% agarose gel image of GM HT negative samples with little or no visible DNA. The extracted DNA for samples 56 (maize pasta), 57 (maize pasta), 61 (corn-soy blend), 67 (soy milk), 69 (soy milk), 70 (soy milk), 72 (soy milk) and 75 (texturized soy protein) in duplicate were resolved on a 1% agarose gel. No DNA was visible for samples 56 (maize pasta), 57 (maize pasta), 61 (corn-soy blend) and 70 (soy milk). Samples 67 (soy milk), 69 (soy milk), 72 (soy milk) and 75 (texturized soy protein) indicated smears, suggesting that degraded DNA may be present. Lambda DNA [50 ng/µL] (lanes marked M) was used to evaluate the quality of extracted sample DNA. 38 Chapter 3: Results and discussion for glyphosate content in food products Table 3.8: Food products that contained glyphosate but no GM HT event. Sample Glyphosate residue Glyphosate residue identification Description (ppb) (mg/kg) number 6 Instant maize meal 49 0.049 11 Maize meal 34 0.034 56 Maize pasta 47 0.047 57 Maize pasta 50 0.050 61 Corn-soy blend 61 0.061 64 Soybeans 49 0.049 67 Soy milk 36 0.036 69 Soy milk 52 0.052 70 Soy milk 49 0.049 72 Soy milk 50 0.050 74 Soy flour 27 0.027 Texturized soy 75 41 0.041 protein 39 Chapter 3: Results and discussion for glyphosate content in food products Table 3.9: Fluorometric determination of DNA concentration in GM HT negative samples. Sample identification DNA concentration in Mean DNA number duplicate (ng/µL) concentration (ng/µL) 0.145 56 0.172 0.198 0.129 57 0.142 0.155 0.254 61 0.270 0.285 0.009 67 0.008 0.007 0.005 69 0.007 0.009 0.014 70 0.010 0.005 0.003 72 0.005 0.006 0.453 75 0.449 0.445 40 Chapter 3: Results and discussion for glyphosate content in food products Table 3.10: Food products containing GM HT event GTS40-3-2 below the limit of quantification but containing glyphosate. Sample Copy number of Glyphosate residue Glyphosate residue identification Description GM status reference genes (ppb) (mg/kg) number (HMG and Lectin) 15,464 copies of 66 Soy milk Below LOQ 40 0.040 Lectin 20,654 copies of 68 Soy milk Below LOQ 32 0.032 Lectin 18,402 copies of 79 Texturized soy protein Below LOQ 37 0.037 Lectin 41 Chapter 3: Results and discussion for glyphosate content in food products Table 3.11: Food products containing GM HT event NK603 but with no detectable glyphosate. Sample identification Percentage of GM HT Description number event detected 25 Maize grits 54.63% 26 Maize grits 42.51% 27 Maize grits 55.18% 28 Maize grits 16.39% 29 Maize grits 21.50% 30 Maize grits 67.47% 31 Maize grits 49.41% 32 Maize grits 9.72% 33 Polenta 0.25% 34 Polenta 53.65% 35 Polenta 11.55% 36 Polenta 31.56% 45 Corn chips 11.54% 46 Corn chips 41.65% 47 Corn chips 33.15% 48 Corn chips 35.33% 49 Corn chips 10.76% 50 Corn chips 20.36% 51 Corn chips 17.54% 52 Corn chips 16.35% 53 Corn chips 12.22% 42 Chapter 3: Results and discussion for glyphosate content in food products 3.4 Theoretical exposure to glyphosate through food products in South Africa In this study, glyphosate was detected in 66.70% of the food products analysed. The level of glyphosate in the food products was compared to the acute toxicity level as well as to the no observed adverse effect level (NOAEL) established for glyphosate in experimental animals. Acute toxicity refers to the adverse effects of a substance occurring after a single dose or from multiple doses within 24 hours (Rand, 1995), while the NOAEL refers to the highest dose of a substance at which no adverse or toxic effects are observed (Darota and Engelhardt, 2005). The acute toxicity level for glyphosate in experimental animals is suggested to be approximately 5,000 mg/kg, while the NOAEL is 175 mg/kg (USA EPA, 2002). Thus, the acute toxicity level of glyphosate for a 60 kg individual (considered to be the mean weight of individuals from Africa (Walpole et al., 2012)) will be 300,000 mg, while the NOAEL will be 10,500 mg. Based on these data, the level of glyphosate in the food products tested is considered safe. The safety of food additives or pesticide residue is based on the NOAEL based on animal studies that has been shown to cause no adverse effect (USA FDA, 1993). The NOAEL is then used to determine the acceptable daily intake (ADI) based on an appropriate safety factor (USA FDA, 1993). The ADI refers to the amount of a substance that a consumer can ingest daily over an entire lifetime without any recognizable health risk (Benford, 2000). It is important to note that the NOAEL and ADI is based on the safety of pure glyphosate, which compared to glyphosate in formulation, is not very permeable into the cell. The ADI for glyphosate differs between countries and ranges from 0.3 mg/kg in Europe to 1.75 mg/kg in the USA (FAO, 2013). The ADI for glyphosate in South Africa is 1 mg/kg of bodyweight per day (ACB, 2012). The theoretical intake of glyphosate resulting from the consumption of the maize and soybean products in this study was calculated and compared to the ADI (Table 3.12). The theoretical intake of glyphosate, as a result of consuming the maize products tested in this study, ranged from 0.001 mg to 0.047 mg, which is below the ADI (Table 3.12). For the soybean products, the theoretical intake of glyphosate ranged from 0.003 mg to 0.021 mg that is also below the ADI (Table 3.12). One of the texturized soy protein products contained a level of glyphosate of 2.257 mg/kg, which would amount to a theoretical intake of 0.339 mg (Table 3.12). Thus, the intake of glyphosate 43 Chapter 3: Results and discussion for glyphosate content in food products as a result of the daily consumption of the maize and soybean products tested in this study does not exceed the ADI in South Africa. It is important to consider that maize is the primary staple and soybean an important source of protein in South Africa. A study by Payne (2011) suggested that approximately 500 g of maize meal is consumed daily by adults (age 18 to 65) in poor households. Based on this, it is estimated that 182 kg of maize meal is consumed per person annually and approximately 8,601 kg over the average adult lifespan (age 18 to 65). Considering that the level of glyphosate detected in maize meal, in this study, ranged from 0.027 to 0.093 mg/kg (Table 3.1), the total exposure to glyphosate would be approximately 5 mg to 17 mg annually and approximately 240 mg to 806 mg over the average adult lifespan (age 18 to 65). A recent report by the National Agricultural Marketing Counsel of South Africa (NAMC) indicated that texturized soy protein was the major soybean food product consumed in South Africa, accounting for approximately 52% of the soybean food market (NAMC, 2011). In the absence of an estimation of the daily intake of soybean per adult individual in South Africa, the serving suggestion was used to calculate the maximum potential consumption. Based on this, it is estimated that 54.8 kg of texturized soy protein is in theory consumed per person annually and approximately 2,575.6 kg (54.8 x 47 years) over the adult lifespan (age 18 to 65). Considering that the level of glyphosate detected in texturized soy protein products, in this study, ranged from 0.041 to 2.257 mg/kg (Table 3.3), the total exposure to glyphosate would be approximately 2.2 mg to 123.7 mg annually and approximately 105.6 mg to 5,813.1 mg over the average adult lifespan (age 18 to 65). This study suggests that the consumption of maize and soybean food products does expose South Africans to glyphosate but at low levels. 44 Chapter 3: Results and discussion for glyphosate content in food products Table 3.12: Theoretical daily intake of glyphosate through food products. Range of glyphosate in Maximum daily product Daily glyphosate intake % of SA ADI for a 60kg Product category product consumption (g) (mg) adult (60 mg) (mg/kg) Maize meal 500 1 0.027 to 0.093 0.014 to 0.047 0.023 to 0.078 Beer powder 100 2 0.040 to 0.095 0.004 to 0.01 0.007 to 0.017 Maize rice 50 2 0.028 to 0.065 0.001 to 0.003 0.002 to 0.005 Maize pasta 40 2 0.047 to 0.062 0.002 to 0.002 0.003 to 0.003 Soybean milk 150 2 0.032 to 0.142 0.005 to 0.021 0.008 to 0.035 Soybeans 100 2 0.049 0.005 0.008 Infants Milk 144 2 0.058 0.008 0.013 Soy Flour 100 2 0.027 0.003 0.005 Corn-soy blend 30 2 0.043 to 0.065 0.001 to 0.002 0.002 to 0.003 Texturized soy protein 150 2 0.041 to 2.257 0.006 to 0.339 0.01 to 0.565 1 For maize meal, 500 g per person per day was used as estimated by Payne (2011) 2 The maximum daily intake was estimated from the serving suggestion on each packet 45 Chapter 3: Results and discussion for glyphosate content in food products In recent years, several in vitro studies on human cell lines have reported that glyphosate in formulation exhibits toxicity at low concentrations. These findings include human cell death at 5 mg/kg as well as DNA damage and endocrine disruption at concentrations ranging from 0.006 mg/kg to 0.5 mg/kg (Table 1.3). Taking this into account, the levels of glyphosate detected in the maize and soybean products in this study cannot be summarily considered to have no effect. Of the 54 products containing glyphosate, three contained above 0.5 mg/kg (500 ppb) (Appendix A) that has been determined to cause endocrine disruption in cultured humans cells (Gasnier et al., 2009). A further 12 products contained glyphosate above 0.069 mg/kg (69 ppb) (Appendix A) that has been determined to cause DNA damage in human lymphocytes (Alvarez-Moya et al., 2014). Furthermore, all 54 products that were positive for glyphosate contained above 0.006 mg/kg (6 ppb) (Appendix A), that was determined to cause chromosomal breakage in hamster ovary cells (Roustan et al., 2014). In addition to this, the IARC recently re-classified glyphosate as a “probable human carcinogen”. In conclusion, although considered safe based on the ADI as well as the acute toxicity level and NOAEL, the effect of exposure to glyphosate at low levels through maize meal, consumed daily in South Africa, warrants further research. 46 Chapter 4: Results and discussion in terms of compliance with GM labelling in South Africa CHAPTER 4: RESULTS AND DISCUSSION IN TERMS OF COMPLIANCE WITH GM LABELLING IN SOUTH AFRICA 4.1 Analysis of GM HT content in GM labelled food products The percentage of GM HT event was quantified in GM HT positive products and most of these were labelled in terms of GM content. Since the data was available, it was used to determine compliance to mandatory GM labelling in terms of the Consumer Protection Act (2008) that mandates GM labelling in South Africa. According to section D: 24 of the Act, “any person who produces, supplies, imports or packages any prescribed goods must display on, or in association with the packaging of those goods, a notice in the prescribed manner and form that discloses the presence of any genetically modified ingredients or components of those goods in accordance with applicable regulations” (SACPA, 2009). In South Africa, the application of mandatory GM labelling is as follows (Viljoen and Marx, 2013): • Ingredients in food containing more than 5% GM must be labelled “genetically modified”. • Ingredients in food containing less than 1% GM may voluntarily be labelled “non-GMO”. • Companies may use a cost effective option of labelling food ingredients “may contain genetically modified ingredients” where GM testing is not scientifically practical or feasible. Although the food products were only tested for GM HT events (NK603, GA21 and GTS40-3-20), a product was considered not to comply with the Consumer Protection Act if it was labelled as non-GM but contained more than 1% of a GM HT event or was not labelled to contain GM and contained above 5% of a GM HT event. Of the 81 food products, 57 had a label indicating GM status. Of the 57 labelled products, 28 were labelled “contains GMO”, 14 were labelled “may contain GMO”, five were labelled “non GMO” and ten were labelled “GMO free”. Of the 28 products labelled “contains GMO”, 25 contained a GM HT event above 5%, while 3 contained a GM HT event below 1% (Table 4.1). Of the 14 products labelled “may contain GMO”, 12 contained a GM HT event above 5%, while two products contained a GM HT event below 5% (Table 4.1). Based on the percentage GM in these products, it is evident that mandatory GM 47 Chapter 4: Results and discussion in terms of compliance with GM labelling in South Africa labelling has not had a negative impact on the GM market as initially feared by the agricultural industry (Viljoen and Marx, 2013). In terms of labelling to indicate the absence of GM, five products were labelled “non-GMO” and all of these products contained a GM HT event below 1% (Table 4.2). In addition to this, ten products were labelled “GMO free”, of which four products contained a GM HT event above 5%, one contained a GM HT event above 1%, and one contained a GM HT event below 1% (Table 4.2). “GMO free” is not legally defined by the South African Consumer Protection Act (2008) and is not internationally accepted as it cannot be defined analytically (Viljoen et al., 2006). The term “GMO free” implies 0% GM content and as stated above six products labelled “GMO free” contained a GM HT event which ranged from 0.25% to 53.65% (Table 4.2). Therefore, the use of “GMO free” in these products is misleading and inaccurate. From this, it can be concluded that discerning consumers with a preference for non-GM products will be more likely to get what they expect from a “non GMO” labelled product than from products labelled “GMO free”. Overall, the results from this study indicate that the majority of companies comply with the Consumer Protection Act in terms of GM labelling with only a few incidences of mislabelling. Since only GM HT events were tested in this study it is important to note that insect resistant events may be present (only for maize products) resulting in an underestimation of the GM content. Nonetheless, constant monitoring of compliance to the Consumer Protection Act in terms of GM labelling is important to protect the rights of consumers in South Africa. 48 Chapter 4: Results and discussion in terms of compliance with GM labelling in South Africa Table 4.1: GM HT quantification and GM label. Sample GM related claim Percentage GM identification Sample description on product content number 1 Maize meal May contain GMO 92.34%NK603 2 Maize meal May contain GMO 100.00%NK603 3 Maize meal May contain GMO 45.78%NK603 4 Maize meal Contains GMO 61.60%NK603 5 Maize meal Contains GMO 100.00%NK603 7 Maize meal Contains GMO 59.90%NK603 8 Maize meal Contains GMO 35.80%NK603 9 Maize meal Contains GMO 76.27%NK603 10 Maize meal Contains GMO 32.17%NK603 13 Maize meal Contains GMO 32.55%NK603 14 Maize meal Contains GMO 30.45%NK603 15 Maize meal May contain GMO 6.44%NK603 16 Maize meal May contain GMO 35.97%NK603 17 Maize meal May contain GMO 4.77%NK603 18 Maize meal Contains GMO 24.95%NK603 49 Chapter 4: Results and discussion in terms of compliance with GM labelling in South Africa Table 4.1: (continued) Sample GM related claim Percentage GM identification Sample description on product content number 20 Maize meal Contains GMO 80.61%NK603 19 Maize meal Contains GMO 22.81%NK603 27.07%NK603 and 21 Maize meal Contains GMO 0.72%GA21 22 Maize meal Contains GMO 26.62%NK603 24 Beer powder May contain GMO 40.47%NK603 25 Maize grits May contain GMO 54.63%NK603 26 Maize grits Contains GMO 42.51%NK603 27 Maize grits Contains GMO 55.18%NK603 28 Maize grits Contains GMO 16.39%NK603 29 Maize grits May contain GMO 21.50%NK603 30 Maize grits May contain GMO 67.47%NK603 31 Maize grits Contains GMO 49.41%NK603 32 Maize grits May contain GMO 9.72%NK603 38 Corn flakes Contains GMO ND 42 Corn flakes Contains GMO ND 50 Chapter 4: Results and discussion in terms of compliance with GM labelling in South Africa Table 4.1: (continued) Sample GM related claim Percentage GM identification Sample description on product content number 45 Corn chips Contains GMO 11.54%NK603 46 Corn chips Contains GMO 41.65%NK603 47 Corn chips Contains GMO 33.15%NK603 48 Corn chips Contains GMO 35.33%NK603 49 Corn chips Contains GMO 10.76%NK603 50 Corn chips Contains GMO 20.36%NK603 51 Corn chips Contains GMO 17.54%NK603 52 Corn chips May contain GMO 16.35%NK603 53 Corn chips May contain GMO 12.22%NK603 55 Maize pasta Contains GMO 21.15%NK603 58 Corn-soy blend Contains GMO Below LOQ 63 Corn-soy blend May contain GMO 4.01%GTS40-3-2 Below LOQ: refers to a GM HT event being detected, but below 0.05% ND: Not detected 51 Chapter 4: Results and discussion in terms of compliance with GM labelling in South Africa Table 4.2: GM HT quantification of food products labelled “non-GMO” and “GMO free”. Sample GM related Percentage GM identification Sample description claim content number 11 Maize meal GMO free ND 12 Maize meal GMO free 1.28%NK603 33 Polenta GMO free 0.25%NK603 4 Polenta GMO free 53.65%NK603 56 Maize pasta GMO free ND 59 Corn-soy blend Non-GMO 0.12%NK603 61 Corn-soy blend GMO free ND 64 Soybeans GMO free ND 65 Soy milk GMO free 9.57%GTS40-3-2 68 Soy milk Non-GMO Below LOQ 71 Soy milk GMO free 5.21%GTS40-3-2 72 Soy milk GMO free ND 74 Soybean flour Non-GMO ND 75 Texturized soy protein Non-GMO ND 79 Texturized soy protein Non-GMO Below LOQ Below LOQ: refers to a GM HT event being detected but below 0.05% ND: Not detected 52 Chapter 5: Limitations of this study CHAPTER 5: LIMITATIONS OF THIS STUDY There are inherent limitations associated with this study that should be considered when evaluating the data presented: • The Abraxis Glyphosate Plate Assay is specific for glyphosate and does not detect AMPA which is a metabolic by-product of glyphosate. AMPA has only been detected in HT soybean and is not considered a metabolite in maize. Thus, the amount of glyphosate in soybean may be underestimated due to its metabolic conversion to AMPA. • Since only GM HT events were quantified in this study, it is possible that the GM content of the products may be underestimated. However, this is only applicable to the maize products that may contain insect resistant events. The only event approved for GM soybean in South Africa is GTS40-3-2. • The agricultural application rate of glyphosate may vary from season to season. Considering that most farming in South Africa is on dryland, season to season variation in rainfall will influence the extent of weed emergence and subsequently the application of glyphosate that would affect the amount of herbicide present in food products. • It is not known to what extent glyphosate in food is absorbed through the gastrointestinal tract in humans. It is possible but currently unknown whether glyphosate in a food matrix may be more easily absorbed, in the gastrointestinal tract. However, since this is unknown, it must be considered that without surfactant, the glyphosate may not be so readily absorbed. • Although this study focused on glyphosate in maize and soybean food products, it should be noted that South Africa also produces GM HT cotton. Cotton seed oil is used in various food products. Furthermore, although South Africa does not plant GM HT wheat, it is common practice to desiccate wheat with glyphosate before harvesting. The exposure to glyphosate through cotton seed oil and wheat in food was not determined in this study. 53 Chapter 6: Conclusion CHAPTER 6: CONCLUSION South Africa is considered a major GM crop producing country ranking ninth in the world in 2015. In South Africa, GM HT maize and soybean account for approximately 75% of GM crops planted. These crops are frequently treated with glyphosate during the growing season to kill weeds. Several studies have indicated the presence of glyphosate in HT grain as well as processed food products from major GM crop producing countries including the USA and Argentina. However, to date no such data exists regarding the prevalence of glyphosate in HT grain produced in South Africa. Maize, in the form of maize meal, is considered the major staple food in South Africa, with approximately 500 g consumed daily per adult in poor households, while soybean is an important source of protein. Thus, the aim of this study was to test South African food products containing maize and soybean as a primary constituent, for glyphosate residue. The results from the study have shown that glyphosate was present in the majority of maize and soybean food products (n = 54) in a range of 27 to 2,257 ppb (0.027 mg/kg to 2.257 mg/kg) confirming that South African consumers are exposed to glyphosate through their food. Although the research on the presence of glyphosate in food products is limited, some studies have detected levels of up to 100 ppb (0.1 mg/kg) in bread from the UK and levels of up to 564 ppb (0.564 mg/kg) in soy sauce from the USA. Of the food products tested in this study, 49 contained levels of glyphosate below 100 ppb (0.1 mg/kg), while five contained levels exceeding 100 ppb (one soy milk at 142 ppb and four texturized soy protein products ranging from 195 ppb to 2,257 ppb). Furthermore, 51 products contained levels of glyphosate below 564 ppb (0.564 mg/kg), while three contained levels exceeding 564 ppb (three texturized soy protein products ranging from 921 ppb to 2257 ppb). This confirmed that the levels of glyphosate in South African maize and soybean food products were in a similar range compared to levels detected in food products from the UK and USA. There was no correlation between the percentage GM HT event and the level of glyphosate present in a product. The lack of correlation suggests that although farmers may plant GM HT maize or soybean, they will most likely only spray herbicide 54 Chapter 6: Conclusion if it is necessary. The latter explains why some products did contain glyphosate. Furthermore, glyphosate is also used as a desiccant on conventional crops, hence its presence in non-GM grain. Since environmental factors like rainfall and agricultural practice influence the emergence of weeds, season to season variation of glyphosate application is expected. As a result of this, the level of glyphosate in food will differ from season to season. The level of glyphosate, in the food products analysed in this study, was compared to the MRL for maize and soybean as well as the ADI established for South Africa. The data from this study has determined that the level of glyphosate in the food products was low and did not exceed the MRL or ADI. It should be noted that MRL are not a safety indication but purely serve as a guideline for food trade and “good” agricultural practice. ADI is based on the NOAEL (175 mg/kg) of pure glyphosate which, without surfactants is not very permeable across the cell membrane. In agriculture, glyphosate is applied as part of a formulation which contains surfactants to facilitate the absorption into plant cells. Arguably, glyphosate in formulation at a similar concentration would be more toxic than pure glyphosate. While it is not known what the effect is on absorption, once glyphosate is already in a food matrix, it is arguable that once absorbed into a biological system like a plant cell, transfer into animal and human tissue after consumption may occur more easily. In recent years, research has reported that glyphosate in formulation at concentrations as low as 0.006 mg/kg (6 ppb) and 0.069 mg/kg (69 ppb) can cause DNA and chromosomal damage in animal and human cells lines (Alvarez-Moya et al., 2014; Roustan et al., 2014). Based on these findings, the glyphosate in South African maize and soybean food products are at a level, that may affect the health of consumers. All the glyphosate positive samples in this study contained levels above 0.006 mg/kg (6 ppb) that has been reported to cause chromosomal breakage in animal cells. Of these, 12 samples (five maize meals, one beer powder, two soy milks and five texturized soy protein products) contained levels of glyphosate above 0.069 mg/kg that has been determined to cause DNA damage in human cells. When considering these findings, it can be argued that the daily consumption of a major staple like maize meal chronically exposes consumers to glyphosate, that is considered a “probable human carcinogen” by the IARC, at levels known to be genotoxic to human and animal cells. 55 Chapter 6: Conclusion The percentage GM HT event in each sample was also used to evaluate products in terms of GM labelling requirement in South Africa. The South African Consumer Protection Act mandates the labelling of products containing more than 5% GM (labelled as “contains GMO” or “may contain GMO”), while it makes provision for non- GM labelling in products containing below 1% GM (labelled as “non-GMO”). The majority of products tested in this study (n=54) were labelled in terms of their GM content and complied with the GM labelling law as required by the South African Consumer Protection Act (2008). Most of the products labelled “may contain GMO” or “contains GMO”, with the exception of two cornflakes samples, contained a GM HT event. All the products labelled “non-GMO” contained a GM HT event below 1%. However, of the ten products labelled “GMO free”, four contained a GM HT event above 5%, one contained a GM HT event above 1% and one contained a GM HT event below 1%. “GMO free” is not legally defined in South Africa and the use of the term indicates a lack of analytical understanding as it refers to a 0% GM content. Considering that six of the products labelled “GMO free” contained a GM HT event, the use of the term “GMO free” should be reconsidered as it is interpreted inaccurately and may be misleading to consumers. The findings of this study suggest that the majority of companies comply with the Consumer Protection Act in terms of GM labelling, nonetheless frequent monitoring of food products in terms of GM labelling remain important to protect the rights of South African consumers. This is the first study that has investigated the content of glyphosate in South African food products. The current study has confirmed that glyphosate is present in commercially available maize and soybean food products in South Africa. This does not imply that these food products are unsafe but that South African consumers are exposed daily to glyphosate through their diet. The question of whether the glyphosate in food is safe in the long term, needs to be addressed through future research. It is currently not known what the safety implications of chronic exposure to glyphosate are, even at low doses. 56 List of references LIST OF REFERENCES ACB. (2012). African Centre of Biosafety. How much glyphosate is on your dinner plate? Available from http://acbio.org.za/wp-content_uploads/2014/12/ACB_ Glyphosate_Food_Safety_in_SA_July2012.pdf. Accessed 12-10-2016. Acquavella JF, Alexander BH, Mandel JS, Gustin C, Baker B, Chapman P, Bleeke, M. (2004). Glyphosate biomonitoring for farmers and their families: results from the farm family exposure study. Environmental Health Perspectives 112(3): 321-326. Alvarez-Moya C, Silva MR, Ramírez CV, Gallardo DG, Sánchez RL, Aguirre AC. (2014). Comparison of the in vivo and in vitro genotoxicity of glyphosate isopropylamine salt in three different organisms. Genetics and Molecular Biology 37(1): 105-110. Arregui MC, Lenardón A, Sanchez D, Maitre MI, Scotta R, Enrique S. (2004). Monitoring glyphosate residues in transgenic glyphosate resistant soybean. Pest Management Science 60: 163-166. Battaglin WA, Meyer TM, Kuivila KM, Dietze JE. (2014). Glyphosate and its degradation product AMPA occur frequently and widely in U.S. soils, surface water, groundwater, and precipitation. Journal of the American Water Resources Association 50(2): 275-290. Bauer T, Weller P, Hammes WP, Hertel C. (2003). The effect of processing parameters on DNA degradation in food. European Food Research and Technology 217: 338-342. Bellé R, Marc J, Morales J, Cormier P, Mulner-Lorillon O. (2012). Letter to the editor: toxicity of roundup and glyphosate. Journal of Toxicology and Environmental Health 15: 233-237. 57 List of references Benbrook CM. (2012). Impacts of genetically engineered crops on pesticide use in the U.S. the first sixteen years. Environmental Sciences Europe 24: 24. Benbrook CM. (2016). Trends in glyphosate herbicide use in the United States and globally. Environmental Sciences Europe 28(1): 1-15. Benford D. (2000). The acceptable daily intake: a tool for ensuring food safety. Available at http://ilsi.org/publication/the-acceptable-daily-intake-a-tool-for- ensuring-food-safety. Accessed 08-09-2016. Bøhn T, Cuhra M, Traavik T, Sanden M, Fagan J, Primicerio R. (2014). Compositional differences in soybeans on the market: glyphosate accumulates in roundup ready GM soybeans. Food Chemistry 153: 207-215. Bonny S. (2016). Genetically modified herbicide tolerant crops, weeds, and herbicides: overview and impact. Environmental Management 57(1): 31-48. Brewster DW, Warren J, Hopkins WE. (1991). Metabolism of glyphosate in Sprague- Dawley rats: tissue distribution, identification, and quantitation of glyphosate- derived materials following a single oral dose. Fundamental and Applied Toxicology 17: 43-51. Chawla HS. (2002). Introduction to plant biotechnology. Gene transfer in plants. Science Publishing Group 2: 365-368. Codex Alimentarius. (2015). Codex Alimentarius International Food Standards. Pesticide residue in food and feed: glyphosate. Available at http://www.fao.org/fao-who-codexalimentarius/standards/pestres/pesticide- detail/en/?p_id=158. Accessed 20-05-2016. Curwin BD, Hein MJ, Sanderson WT, Streiley C, Heederik D. (2007). Pesticide dose estimates for children of Iowa farmers and non-farmers. Environmental Research 105: 307-315. 58 List of references Darota MA, Engelhardt JA. (2005). The no-observed-adverse-effect-level in drug safety evaluations: use, issues, and definition(s). Regulatory Toxicology and Pharmacology 42(3): 265-274. De la Riva GA, Gonzalez-Cabrera J, Vazquez-Padron R, Ayra-Pardo C. (1998). Agrobacterium tumefaciens: a natural tool for plant transformation. Electronic Journal of Biotechnology 1(3): 118-133. De Roos AJ, Zahm SH, Cantor KP, Weisenburger DD, Holmes FF, Burmeister LF. (2003). Integrative assessment of multiple pesticides as risk factors for non- Hodgkin’s lymphoma among men. Occupational and Environmental Medicine 60(9): 1-9. Dill GM, Sammons RD, Feng PC, Kohn F, Kretzmer K, Mehrsheikh A. (2010). Chapter 1: Glyphosate: discovery, development, applications, and properties. In: Nandula, VK. (Editor). Glyphosate resistance in crops and weeds: history, development, and management. Hoboken (NJ). Wiley 1-33. Du Plessis J. (2003). Maize production, Republic of South Africa. Available at http://www.arc.agric.za/arc-gci/Fact%20Sheets%20Library/Maize%20Pro- duction.pdf. Accessed 09-03-2015. Duke SO, Rimando AM, Pace PF, Reddy KN, Smeda RJ. (2003). Glyphosate and aminomethylphosphonic acid levels in seeds of glyphosate-treated, glyphosate-resistant soybean. Journal of Agricultural and Food Chemistry 51: 340-344. Duke SB, Powles, SO. (2010). Glyphosate-resistant crops and weeds: now and in the future. Journal of Agro-biotechnology Management and Economics 12(3): 346-357. EFSA. (2009). European Food Safety Authority. Modification of the residue definition of glyphosate in genetically modified maize grain and soybeans, and in products of animal origin. EFSA Journal 7: 1310-1317. 59 List of references EFSA. (2015). European Food Safety Authority. Conclusion on the peer review of the pesticide risk assessment of the active substance glyphosate. EFSA Journal 13: 4302. Eriksson M, Hardell L, Carlberg M, Akerman M. (2008). Pesticide exposure as risk factor for non-Hodgkin lymphoma including histopathological subgroup analysis. International Journal of Cancer 123(7): 1657-1663. FAO. (2005). Food and Agricultural Organization of the United Nations. Glyphosate (158). Available at http://www.fao.org/fileadmin/templates/agphome/docu ments/Pests-Pesticides/JMPR/Evaluation97/Glypho.PDF. Accessed 23-02- 2015. FAO. (2006). Food and Agricultural Organization of the United Nations. International code of conduct on the distribution and use of pesticides: guidelines on efficacy evaluation for the registration of plant protection products. Available at http://www.fao.org/fileadmin/templates/agphome/documents/Pests_Pest- icides/Code/Efficacy.pdf. Accessed 09-09-2016. FAO. (2013). Food and Agricultural Organization of the United Nations. FAO specifications and evaluations for agricultural pesticides: glyphosate. Available at http://www.fao.org/fileadmin/templates/agphome/documents/ Pests_Pesticides/Specs /Glypho_2014.pdf. Accessed 20-05-2015. Franz JE, Mao MK, Sikorski JA. (1997). Glyphosate: a unique global herbicide. American Chemical Society 4: 65-97. Funke T., Han H, Healy-Fried ML, Fischer M, Schönbrunn E. (2006). Molecular basis for the herbicide resistance of roundup ready crops. National Academy of Sciences of the United States of America 103: 13010-13015. 60 List of references Gasnier C, Dumont C, Benachour N, Clair E, Chagnon M. (2009). Glyphosate-based herbicides are toxic and endocrine disruptors in human cell lines. Toxicology Journal 262: 184-191. Gasser CS, Fraley RT. (1989). Genetically engineering plants for crop improvement. Science 244(4910): 1293-1299. Gelvin SB. (1998). The introduction and expression of transgenes in plants. Current Opinion in Biotechnology 9: 227-232. Gelvin SB. (2003). Agrobacterium-mediated plant transformation: the biology behind the “gene-jockeying” tool. Microbiology and Molecular Biology Reviews 67(1): 16-37. Gomes MP, Smedbol E, Chalifour A, Hènault-Ethier L, Labrecque M, Lepage L, Lucotte M, Juneau P. (2014). Alteration of plant physiology by glyphosate and it’s by product aminomethylphosphonic acid: an overview. Journal of Experimental Botany 65(17): 2-13. Gouse M. (2014). Assessing the value of glyphosate in the South African agricultural sector. Available at http://www.africabio.com/value-of-glyphosate-in-sa- agriculture/. Accessed 18-09-2016. Hansen G, Wright MS. (1999). Recent advances in the transformation of plants. Trends in Plant Science 4(6): 226-231. Hardell L, Eriksson M, Nordstrom M. (2002). Exposure to pesticides as risk factor for non-Hodgkin’s lymphoma and Hairy cell leukemia: pooled analysis of two Swedish case-control studies. Leukemia and Lymphoma 43(5): 1043-1049. 61 List of references Health Canada. (2015). Health Canada – Pest Management Regulatory Agency. Proposed re-evaluation decision PRVD2015-01, glyphosate. Available at http://www.hc-sc.gc.ca/cps-spc/pest/part/consultations/_prvd2015-01/prvd- 2015-01-eng.php. Accessed 17-03-2016. Herzfeld D, Sargent K. (2012). Private pesticide applicator safety education manager. Available at http://www.extension.umn.edu/agriculture/pesticide-safety /ppatmanual.html. Accessed 16-08-2016. Hilton CW. (2012). Monsanto and the global glyphosate market: case study. Available at http://www.wiglafjournal.com/pricing/2012/06/monsanto-the-global- glyphosatemarket-case-study/. Accessed 28-10-2015. Hoppe HW. (2013). Determination of glyphosate residues in human urine samples from 18 European countries. Available at https://www.foeeurope.org/sites/ default/f iles/glyphosate_studyresults_june12.pdf. Accessed 17-02-2016. IARC. (2015). International Agency for Research on Cancer. Carcinogenicity of tetrachlorvinphos, parathion, malathion, diazinon, and glyphosate. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans 112: 1-2. ISAAA. (2015). International Service for the Acquisition of Agri-Biotech Application. GM approval database. Available at http://www.isaaa.org/gmapprovaldata- base/event/default.asp?EventID. Accessed 15-08-2015. James C. (2013). Global status of commercialized biotech/GM crops: 2013. ISAAA Brief No. 43. Available at http://www.isaaa.org/resources/publications/briefs/ 43/executivesummary/default.asp. Accessed 17-03-2015. James C. (2014). Global status of commercialized biotech/GM crops: 2014. ISAAA Brief No. 49. Available at http://www.isaaa.org/resources/publications/briefs/ 49/executivesummary/default.asp. Accessed 17-03-2015. 62 List of references James C. (2015). Global status of commercialized biotech/GM crops: 2015. ISAAA Brief No. 51. Available at http://www.isaaa.org/resources/publications/briefs/ 51/ executivesummary/default.asp. Accessed 13-06-2016. JMPR. (2016). Joint FAO/WHO Meeting on Pesticide Residues. Summary report. Available at http://www.who.int/foodsafety/jmprsummary2016.pdf. Accessed 14-09-2016. JRC-IRMM. (2017). European Commission’s Joint Research Centre, Institute of Reference Materials and Measurements. Reference materials 2017. Available at https://ec.europa.eu/jrc/sites/jrcsh/files/rm_catalogue_170101_. pdf. Accessed 30-01-2017. Klein TM, Wolf ED, Wu R, Sanford CJ. (1987). High-velocity micro-projectiles for delivering nucleic acids into living cells. Nature 327: 70-73. Koller VJ, Fürhacker M, Nersesyan A, Mišík M, Eisenbauer M, Knasmueller S. (2012). Cytotoxic and DNA-damaging properties of glyphosate and Roundup in human-derived buccal epithelial cells. Archives of Toxicology 86(5): 805-813. König A, Cockburn A, Crevel RWR, Debruyne E, Grafstroem R, Hammerling U, Kimber I, Knudsen I, Kuiper HA, Peijnenburg AA, Penninks AH, Poulsen M, Schauzu M. (2004). Assessment of the safety of foods derived from genetically modified (GM) crops. Food and Chemical Toxicology 42(7): 1047- 1088. Krüger M, Schledorn P, Schrödl W, Hoppe H, Lutz W, Shehata AA. (2014a). Detection of glyphosate residues in animals and humans. Journal of Environmental and Analytical Toxicology 4(2): 1-5. Krüger M, Schrödl W, Pedersen IB, Schledorn P, Shehata AA. (2014b). Detection of glyphosate in malformed piglets. Journal of Environmental and Analytical Toxicology 4(5): 1-2. 63 List of references Lievens A, Petrillo M, Querci M, Patak A. (2015). Genetically modified animals: options and issues for traceability and enforcement. Trends in Food, Science and Technology 44(2): 159-176. Lipp M, Brodmann P, Pietsch K, Pauwels J, Anklam E. (1999). IUPAC collaborative trial study of a method to detect genetically modified soy beans and maize in dried powder. Journal of AOAC International 82(4): 923-928. Madsen KH, Streibig JC. (2003). Benefits and risks of the use of herbicide-resistant crops. Available at http://www.fao.org/docrep/006/y5031e/y5031e0i.htm. Accessed 20-04-2015. Mannion AM, Morse S. (2013). GM crops 1996-2012: a review of agronomic, environmental and socio-economic impacts. Available at www.surrey.ac.uk/ ces/files/pdf/04-13%20MorseMannionGM%20Cropspdf4 Accessed 20-05- 2015. McQueen H, Callan AC, Hinwood AL. (2012). Estimating maternal and prenatal exposure to glyphosate in the community setting. International Journal of Hygiene and Environmental Health 215(6): 570-576. Moreno NC, Sofia SH, Martinez CB. (2014). Genotoxic effects of the herbicide Roundup Transorb and its active ingredient glyphosate on the fish Prochilodus lineatus. Environmental Toxicology and Pharmacology 37(1): 448-454. NAMC. (2011). National Agricultural Marketing Council. The South African soybean value chain. Available at http://www.namc.co.za/upload/all%20reports/NAM C%20Soybean%20Industry%20and%20Competitiveness%20Study%20%20 %20June% 202011.pdf. Accessed 26-05-2016. NAMC. (2014). National Agricultural Marketing Council. Food price monitor. Available at http://www.namc.co.za/upload/food_price_monitoring/NAMC20-Food 20Price%20Monitor-February-202015.pdf. Accessed 26-04-2015. 64 List of references Nester E. (2008). Agrobacterium: the natural genetic engineer: 100 years later. Available at http://www.apsnet.org/publications/apsnetfeatures/Pages/Agro- bacterium.aspx. Accessed 12-02-2015. Niemann L, Sieke C, Pfeil R, Solecki R. (2015). A critical review of glyphosate findings in human urine samples and comparison with the exposure of operators and consumers. Journal of Consumer Protection and Food Safety 10: 3-12. Orgeron AJ. (2012). Sugarcane growth, sucrose content, and yield response to the ripeners glyphosate and trinexapacethyl. Available at http://etd.lsu.edu/ docs/available/etd-11072012-134904/unrestricted/Orgeron_diss.pdf. Accessed 24-05-2016. Orsi L, Delabre L, Monnereau A, Delval P, Berthou C, Fenaux P. (2009). Occupational exposure to pesticides and lymphoid neoplasms among men: results of a French case-control study. Occupational and Environmental Medicine 66(5): 291-298. Padgette SR, Kolacz KH, Delannay X, Re DB, La Vallee BJ, Tinius CN, Rhodes WK, Otero YI, Barry GF, Eichholz DA. (1995). Development, identification, and characterization of a glyphosate-tolerant soybean line. Crop Science 35: 1451-1461. Paoletti MG, Pimente D. (1996). Genetic engineering in agriculture and the environment: assessing risks and benefits. BioScience 46(9): 665-673. Parker R., Boydston R. (2005). Weed management and herbicide performance during drought conditions. Available at http://cru.cahe.wsu.edu/CEPublic- ations/em 4929e/em4929e.pdf. Accessed 13-06-2016. Payne T. (2011). The amazing whiteness of local staples. Mail and Guardian, 11 November 2011. Available at http://mg.co.za/article/2011-12-09-food-prices- pummel-the-poor. Accessed 24-08-2016. 65 List of references Plume K. (2015). Scientists assembled for Monsanto say herbicide not carcinogenic, disputing WHO report. Available at http://www.reuters.com/article/us- monsanto-herbicide-glyphosate. Accessed 11-04-2016. Portier CJ, Armstrong BK, Baguley BC. et al., (2016). Differences in the carcinogenic evaluation of glyphosate between the International Agency for Research on Cancer (IARC) and the European Food Safety Authority (EFSA). Journal of Epidemiology and Community Health 70(8): 741-745. PRIF. (2015). Expert Committee on Pesticide Residue in Food. Annual report for 2015. Available at https://www.gov.uk/government/publications/expert-committee- on-pesticide-residues-in-food-prif-annual-report-for-2015. Accessed 10-10- 2016. Qaim M. (2010). Benefits of genetically modified crops for the poor: household income, nutrition, and health. New Biotechnology 27(5): 552-557. Ralston A, Shaw K. (2008). Gene expression regulates cell differentiation. Nature Education 1(1): 127-129. Rand GM. (Ed.). (1995). Fundamentals of aquatic toxicology: effects, environmental fate, and risk assessment. Boca Raton, Florida: CRC press. Reddy KN, Zablotowicz RM. (2004). Impact of glyphosate on the symbiosis with glyphosate-resistant transgenic soybean. Journal of Environmental Quality 33(3): 825-831. 66 List of references Reinhardt C. (2012). South African herbicide resistance initiative (SAHRI): website for communication on weed resistance issues and promotion of ‘best agricultural practices’. Available at http://www.agrisa.co.za/wp-content/ uploads/2014/11/Press-release-for-17-Nov-website-launch _1.pdf. Accessed 17-03-2016. Robinson S, Scott N, Gackle A. (2000). Gene technology and future foods. Asia Pacific Journal of Clinical Nutrition 9: 113-118. Robinson M. (2009). Application and effective use of herbicides. Available at http://www.earlscliffe.com/use_of_herbicides.htm. Accessed 09-09-2016. Roustan A, Aye M, De Meo M, Di Giorgio C. (2014). Genotoxicity of mixtures of glyphosate and atrazine and their environmental transformation products before and after photoactivation. Chemosphere 108: 93-100. Rubio F, Veldhuis LJ, Clegg BS, Fleeker JR, Hall JC. (2003). Comparison of a direct ELISA and HPLC method for glyphosate determination in water. Journal of Agricultural and Food Chemistry 51: 691-696. Rubio F, Guo E, Kamp L. (2014). Survey of glyphosate residues in honey, corn and soy products. Journal of Environmental and Analytical Toxicology 5(1): 2-8. SACPA. (2009). South African Consumer Protection Act, 2008. Government Gazette. 526(31286): 3-186. SA DAFF. (2005). South African Department of Agriculture, Forestry and Fisheries. Understanding genetically modified organisms (GMO’s). Available at http://www. nda.agric.za. Accessed 17-02-2015. SA DAFF. (2015). South African Department of Agriculture, Forestry and Fisheries. Trends in the agricultural sector 2015. Available at http://www.econo- statistics.co.za/TrendsinSAAgriculture15.pdf. Accessed 09-03-2015. 67 List of references SA DAFF. (2016). South African Department of Agriculture, Forestry and Fisheries. Food safety and quality assurance. Available at http://www.SASA DAFF.gov.za/SA DAFFweb3/Branches/Agricultural-Production-Health-Food- Safety/Food-Safety-Qualt -Assurance/Maximum-Residue-Limits. Accessed 15-06-2016. SANBI. (2010). South African National Biodiversity Institute. Monitoring the environmental impacts of GM maize in South Africa. Available at http://www .sanbi.org/sites/default/files/documens/documents/sanbimaizereportlr.pdf. Accessed 17-03-2015. Segelken R. (2010). Biologist invents gun shooting cells with DNA. Available at https://www.gproxx.com/https://web.archive.org/web/20100426054730/http:/ /www.cals.cornell.edu/cals/public/businesspartnerships/made/gene- shotgun.cfm. Accessed 12-02-2015. Smale ST, Kadonaga J. (2003). The RNA polymerase II core promoter. Annual Review of Biochemistry 72: 449-479. Swanepoel S. (2014). GM contamination, cartels and collusion in South Africa’s bread industry. Available at http://www.acbio.org.za/images/stories/dmdoc- uments/GM-Bread-201405.pdf. Accessed 20-05-2015. Székács A, Darvas B. (2012). Forty years with glyphosate. In: Hasaneen MNAEG, (Editor). Herbicides properties, synthesis and control of weeds. Croatia: InTech, pp. 247–84. Available at http://cdn.intechweb.org/ pdfs/25624.pdf. Accessed 28-08-2016. Then C. (2013). High levels of residues from spraying with glyphosate found in soybeans in Argentina. Available at http://www.testbiotech.de/en/node/926. Accessed 17-02-2016. 68 List of references USA EPA. (1986). United States of America Environmental Protection Agency. Glyphosate, additional histopathological evaluations of kidneys in the chronic feeding study of glyphosate in mice. Available at http://www.epa .gov/pesticides/chemicallsearch/chemical/foia/cleared-reviews. Accessed 21-04-2016. USA EPA. (1991). United States of America Environmental Protection Agency. Second peer review of glyphosate. Available at http://www.epa.gov/pestic ides/chemicalsearch/chemicall/foia/cleared-reviews/reviews/103601/103601 -265.pdf. Accessed 21-04-2016. USA EPA. (1993). United States of America Environmental Protection Agency. RED facts: glyphosate. Available at https://archive.epa.gov/pesticides/reregis- tration/web/pdf/0178fact.pdf. Accessed 12-04-2016. USA EPA. (2002). United States of America Environmental Protection Agency. Glyphosate in/on pasture, rangeland grass, roundup ready wheat and non- grass animal feeds. Health effects division risk assessment. Available at https://www3.epa.gov/pesticides/chem_search/cleared_reviews/csr_PC- 103601_20-Feb-02_a.pdf. Accessed 24-06-2016. USA EPA. (2015). United States of America Environmental Protection Agency. EDSP weight of evidence conclusions on the tier 1 screening assays for the list 1 chemicals. Available at https://www.epa.gov/sites/production/files/ 2015- 06/documents/glyphosate417300_2015-06-29_txr0057175.pdf. Accessed 26-04-2016. USA FDA. (1993). United States of America Food and Drug Administration. Redbook II: Agency review of toxicology information in petitions for direct food additives and colour additives used in food. Available at http://www.fda.gov /downloads/Food/GuidanceRegulation/GuidanceDocumentsRegulatoryInfor mation/IngredientsAdditivesGRASPackaging/UCM078724.pdf. Accessed 03- 02-2017 69 List of references Uzogara SG. (2000). The impact of genetic modification on human foods in the 21st century: a review. Biotechnology Advances 18: 179-206. Viljoen CD, Dajee BK, Botha GM. (2006). Detection of GMO in food in South Africa: Implications of GMO labelling. African Journal of Biotechnology 5(2): 73-82. Viljoen CD. (2013). Food and chemical toxicology: Letter to the editor. Food and Chemical Toxicology 59: 809-810. Viljoen CD, Marx GM. (2013). The implications for mandatory GM labelling under the Consumer Protection Act in South Africa. Food control 31(2): 387-391. Whigham DK, Stoller EW. (1979). Soybean desiccation by paraquat, glyphosate, and ametryn to accelerate harvest. Agronomy Journal 71(4): 630-633. Williams GM, Kroes R, Munro IC. (2000). Safety evaluation and risk assessment of the herbicide Roundup and its active ingredient, glyphosate, for humans. Regulatory Toxicology and Pharmacology 31: 117-165. Williams AL, Watson RE, De Sesso JM. (2012). Developmental and reproductive outcomes in humans and animals after glyphosate exposure: a critical analysis. Journal of Toxicology and Environmental Health 15: 39-96. Williams GM, Aardema M, Acquavella J, Berry C, Brusick D, Burns M, De Camargo JLV, Garabrant D, Greim HA, Kier LD, Kirkland JD, Marsh G, Solomon KR, Sorahan T, Roberts A, Weed DL. (2016). A review of the carcinogenic potential of glyphosate by four independent expert panels and comparison to the IARC assessment. Critical Reviews in Toxicology 46(1): 3-20. WHO. (1994) World Health Organization. The WHO recommended classification of pesticides by hazard and guidelines to classification 1994. Available at http://www.inchem.org/documents/ehc/ehc/ehc159.htm. Accessed 12-02- 2015. 70 List of references WHO. (2002). World Health Organization. 20 questions on genetically modified foods. Available at http://www.who.int/foodsafety/publications/biotech/20- questions/ en/indexhtml. Accessed 12-02-2015. WHO. (2014). World Health Organization. Frequently asked questions on genetically modified foods. Available at http://www.who.int/foodsafety/areas_work/food technology/Freguently_asked_questions_on_gm_foods.pdf. Accessed 05- 09-2015. Young F, Ho D, Glynn D, Edwards V. (2015). Endocrine disruption and cytotoxicity of glyphosate and roundup in human JAr cells in vitro. Integrative Pharmacology, Toxicology and Geno-toxicology 1(1): 12-19. 71 Summary SUMMARY South Africa is considered a major GM crop producing country. The predominant GM trait in maize and soybean in South Africa is herbicide tolerance. Herbicide tolerant (HT) crops allow for the application of herbicide during the growing season to selectively kill weeds without damaging the crop. Glyphosate is the most widely used herbicide on HT crops in South Africa and the world. Glyphosate is absorbed by HT crops after application. Studies have detected levels of up to 2.2 mg/kg in HT maize and 26 mg/kg in HT soybean. Glyphosate is not removed from grain by washing, cooking or processing. As a result of this, glyphosate can also be detected in processed food products. It has also been found that glyphosate can be detected in animal tissue and urine after exposure to the herbicide through feed. Similarly, glyphosate has also been detected in the urine of humans, either as a result of occupational exposure, through diet and/or water. Pure glyphosate is considered safe by regulatory authorities and international bodies. However, recent studies have found that glyphosate in formulation at low concentrations results in endocrine disruption as well as DNA and chromosomal damage. As a result of this, the IARC re-classified glyphosate as a “probable human carcinogen”. Glyphosate is used on HT maize, a major staple food, and HT soybean, an important source of protein, in South Africa. Thus, the aim of this study was to determine whether glyphosate is present in commercially available South African food products containing maize and/or soybean as a primary constituent. The majority of food products tested in this study contained glyphosate. The level of glyphosate ranged from 0.027 mg/kg to 2.257 mg/kg that is below the MRL and ADI established in South Africa. However, recent studies have shown that glyphosate in formulation is genotoxic at the levels found in maize and soybean containing foods in South Africa. The results from this study found that the level of glyphosate in food products in South Africa is comparable to the limited number of studies from the UK and USA. This study is unique to other published studies since it focused on food 72 Summary products likely to be consumed daily. This study has confirmed that South Africans are exposed daily to low levels of glyphosate through food products. The GM HT events in the food products were quantified in order to explain the variation in the levels of glyphosate. It was determined that 57 products contained one or more GM HT event in a range of 0.25% to 100%. However, there was no correlation between the level of glyphosate and percentage GM HT event in the products, even when GM HT negative samples were excluded from the analysis. This suggests that either glyphosate is not applied to some GM HT crops, when weed control is not required, or that the herbicide is applied to non-GM crops as a desiccant prior to harvesting. Most of the food products used in this study were labelled in terms of GM content. The percentage GM HT event(s) in the food products was used to determine compliance to mandatory GM labelling in South Africa since the data was available. Results indicated that the majority of companies in South Africa are compliant with the Consumer Protection Act (2008) concerning GM labelling. However, most of the products labelled “GMO free”, did not comply with the expectation of discerning consumers. To conclude, this is the first study to investigate the extent of glyphosate in South African food products. This study has confirmed that South African consumers are exposed to low levels of glyphosate as a result of consuming maize and soybean food products. The level of glyphosate detected in the food products although considered low, is at a concentration reported to cause genotoxic effects at cellular level by in vitro studies. It is currently not known what the safety implications are of chronic exposure to glyphosate, through the consumption of a staple like maize meal and soybean in South Africa. The question of whether glyphosate in food is safe in the long term needs to be addressed through future research. Keywords: Glyphosate, ELISA, GM crops, GM HT event detection, food products, GM labelling, herbicide tolerance, South Africa 73 Opsomming OPSOMMING Suid-Afrika is werêldwyd bekend as die negende grootste produsent van GM gewasse en verbou hoofsaaklik GM mielies, sojabone en katoen. Onkruiddoder toleransie is die grootste GM eienskap in Suid-Afrika en kom voor in ongeveer 75% van alle GM gewasse wat geplant word. Onkruiddoder tolerante (OT) gewasse laat die direkte toediening van onkruiddoder toe, sonder om skade aan die gewas aan te rig. Die gewildste onkruiddoder in Suid-Afrika asook werêldwyd is glifosaat en word hoofsaaklik gebruik vir toediening op OT gewasse. Verskeie studies het getoon dat glifosaat geabsorbeer word deur OT gewasse na toediening. Glifosaat vlakke van tot 2.2 mg/kg is waargeneem in OT mielies en tot 26 mg/kg in OT sojabone. Glifosaat kan nie verwyder word van graan deur dit te was, gaar te maak of te prosesseer nie. Die resultaat hiervan is dat glifosaat gevind word in geprosesseerde voedsel produkte. Verskeie studies het aangedui dat glifosaat geabsorbeer word en voorkom in die weefsel, uriene en fesus van diere wat gevoer is met glifosaat behandelde grane. Glifosaat is ook gevind in menslike uriene, hoofsaaklik as gevolg van kontak met die onkruiddodder tydens landbou praktyk of deur die inname van glifosaat bevattende voedsel en water. Verskeie internasionale owerhede beskou suiwer glifosaat as veilig vir mense en diere. Onlangse studies het egter bevind dat glifosaat in formulasie toksies is teen lae konsentrasies. Die bevindinge sluit endokriene ontwrigting asook DNA en chromosomale skade in wat voorgekom het in beide mens en dierselle. In 2015 het die Internasionale Agentskap vir Navorsing oor Kanker (IARC) glifosaat geherklassifiseer as “waarskynlik karsinogenies vir mense”. OT mielies in die vorm van mieliemeel is ‘n stapelvoedsel in Suid-Afrika en OT sojabone is n belangrike bron van proteїene. Beide hierdie gewasse word ekstensief behandel met glifosaat gedurende die groeiseisoen. Die doel van hierdie studie was dus, om te bepaal of glifosaat voorkom in kommersieël beskikbare Suid-Afrikaanse voedsel produkte, wat mielies of sojabone as die primêre bestanddeel bevat. Glifosaat kom voor in die meerderheid van voedsel produkte getoets in hierdie studie met vlakke wat wissel van 0.027 mg/kg tot 2.257 mg/kg. Die vlakke van glifosaat in 74 Opsomming die voedsel produkte was onder die toegelate limiete (MRL en ADI) vasgetel vir hierdie onkruiddoder in Suid-Afrika. Onlangse studies dui egter aan dat glifosaat in formulasie toksies is teen konsentrasies gelykstaande aan wat voorkom in Suid-Afrikaanse mielie en sojaboon voedsel produkte. Die resultate van hierdie studie dui aan dat die konsentrasies glifosaat gevind in Suid-Afrikaanse voedsel produkte vergelykbaar is met die vlakke gevind in voedsel produkte van die Verenigde Koninkryk (VK) en die Verenigde State van Amerika (VSA). Hierdie studie is egter uniek omdat dit gefokus het op stapelvoedsel produkte wat daagliks geëet word. Die studie bevestig dat Suid- Afrikaners daagliks blootgestel word aan lae vlakke glifosaat deur hulle dieet. Reële-tyd PKR was gebruik om die persentasie OT geen te bepaal wat voorkom in elke produk. Hierdie resultate het aangedui dat 57 uit die 81 produkte ‘n OT geen bevat met vlakke wat wissel van 0.25% tot 100%. Geen korrelasie is egter gevind tussen die persentasie OT geen en die vlak van glifosaat in die produkte nie. ‘n Verdere ondersoek is gedoen na al die produkte wat negtief getoets het vir n OT geen verwyder is, maar steeds is geen korrelasie gevind nie. Die bevindinge stel voor dat glifosaat nie noodwendig toegedien word tot alle OT gewasse nie. Dit kan verduidelik hoekom sommige voedsel produkte positief was vir n GM OT geen, maar geen glifosaat bevat nie. Dit word beskou as algemene landbou praktyk on nie-OT gewasse te spuit met glifosaat, om hierdie gewasse egalig uit te droog voor dit geoes word. Die bogenoemde verduidelik hoekom sommige produkte glifosaat bevat maar vry is van ‘n GM OT geen. Die voedsel produkte was ook ondersoek in terme van GM etikettering aangesien die persentasie GM OT geen per produk reeds bepaal was en meeste produkte ‘n GM etiket bevat het. Die resultate van hierdie studie bevestig dat die meerderheid van maatskappye wat GM bevattende produkte produseer, voldoen aan die Wet op Verbruikersbeskerming ingestel ten opsigte van GM etikettering in Suid-Afrika. Die resultate dui egter aan dat sommige produkte gemerk “GMO vry”, wel GM OT gene bevat en dus kontrasteer met wat hul etiket aandui. As gevolg hiervan mag die “GMO vry” handelsmerk verwarring skep, spesifiek by individue wat nie-GM produkte verkies. Per opsomming, hiedie is die eerste studie wat ondersoek ingestel het ten opsigte van die voorkoms van glifosaat in die Suid-Afrikaanse voedselketting. Die studie het 75 Opsomming bewys dat Suid Afrikaners daagliks bloot gestel word aan lae konsentrasies glifosaat deur die inname van mielie en sojaboon voedsel produkte. Onlangse studies het egter bevind, dat glifosaat in formulasie by soortgelyke konsentrasies soos gevind in die voedsel produkte getoets in hierdie studie, toksies is vir beide mens en dier. Dit is tans onbekend wat die gesondheids implikasies mag wees van kroniese blootstelling tot ‘n lae konsentrasie glifosaat en moet dus in toekomstige studies ondersoek word. Sluitelwoorde: Glifosaat, ELISA, GM gewasse, voedsel produkte, GM etikettering, onkruiddoder toleransie, Suid-Afrika 76 Appendix A APPENDIX A A summary of the detected glyphosate level, standard deviation and limit of quantification of all products tested. Sample Mean of Sample Standard identification Glyphosate description Deviation (ppb) number detected (ppb) 1 Maize meal 46.52 0.55 2 Maize meal 82.91 2.52 3 Instant maize meal 34.58 0.74 4 Maize meal 41.99 1.48 5 Maize meal 59.71 4.55 6 Instant maize meal 49.83 3.67 7 Maize meal 47.81 2.96 8 Maize meal 43.96 2.92 9 Maize meal 81.98 0.59 10 Maize meal 26.90 2.94 11 Maize meal 33.85 1.91 12 Maize meal 43.85 3.52 13 Maize meal 91.64 3.29 14 Maize meal 32.71 0.94 15 Maize meal 63.39 3.98 16 Maize meal 45.63 5.61 17 Maize meal 68.06 5.35 18 Maize meal 92.83 5.26 19 Maize meal 48.27 1.10 20 Maize meal 35.47 0.89 21 Maize meal 43.86 8.55 22 Maize meal 34.33 3.89 23 Beer powder 94.79 3.38 24 Beer powder 40.19 0.18 77 Appendix A (continued) Sample Average of Sample Standard identification Glyphosate description deviation (ppb) number detected (ppb) 25 Maize grits ND ND 26 Maize grits ND ND 27 Maize grits ND ND 28 Maize grits ND ND 29 Maize grits ND ND 30 Maize rice 40.98 0.98 31 Maize rice 28.34 2.61 32 Maize rice 65.42 4.59 33 Polenta ND ND 34 Polenta ND ND 35 Polenta ND ND 36 Polenta ND ND 37 Polenta ND ND 38 Corn flakes ND ND 39 Corn flakes ND ND 40 Corn flakes ND ND 41 Corn flakes ND ND 42 Corn flakes ND ND 43 Corn flakes ND ND 44 Corn flakes ND ND 45 Corn chips ND ND 46 Corn chips ND ND 47 Corn chips ND ND 48 Corn chips ND ND 49 Corn chips ND ND 50 Corn chips ND ND 78 Appendix A (continued) Sample Average of Sample Standard identification Glyphosate description Deviation (ppb) number detected (ppb) 51 Corn chips ND ND 52 Corn chips ND ND 53 Corn chips ND ND 54 Corn chips ND ND 55 Maize Pasta 61.65 6.10 56 Maize Pasta 46.76 3.92 57 Maize Pasta 50.33 5.91 58 Corn-soy blend 59.34 5.94 59 Corn-soy blend 43.76 1.18 60 Corn-soy blend 52.73 4.92 61 Corn-soy blend 61.36 1.52 62 Corn-soy blend 64.54 10.57 63 Corn-soy blend 43.12 2.07 64 Soybean 48.92 1.25 65 Soy milk 142.18 8.98 66 Soy milk 39.75 0.68 67 Soy milk 35.69 4.38 68 Soy milk 31.71 3.96 69 Soy milk 51.84 2.48 70 Soy milk 49.25 12.45 71 Soy milk 81.20 4.27 72 Soy milk 50.37 7.89 73 Infants milk 57.88 5.64 74 Soy flour 27.14 1.03 Texturized soy 75 40.60 4.19 protein Texturized soy 76 921.33 3.83 protein Texturized soy 77 2257.04 53.02 protein 79 Appendix A (continued) Sample Average of Sample Standard identification glyphosate description deviation (ppb) number detected (ppb) Texturized soy 78 78.11 5.72 protein Texturized soy 79 36.94 0.28 protein Texturized soy 80 1977.67 77.05 protein Texturized soy 81 195.07 12.77 protein ND: Not detected 80 Appendix B APPENDIX B A summary of the detected HT events, standard deviation and limit of quantification for the maize products. Sample HT event Percentage Percentage Sample Limit of HT event Limit of identification NK603 standard 1 standard description quantification GA21 detected quantification number detected deviation deviation 1 Maize meal 92.34% 6.26% 0.05% Below LOQ Below LOQ 0.07% 2 Maize meal 100.00% 2.26% 0.06% ND ND ND Instant maize 3 45.78% 4.41% 1.34% Below LOQ Below LOQ 1.41% meal 4 Maize meal 61.60% 1.10% 0.06% ND ND ND 5 Maize meal 100.00% 3.73% 0.05% Below LOQ Below LOQ 0.07% Instant maize 6 ND ND ND ND ND ND meal 7 Maize meal 59.90% 5.51% 0.04% ND ND ND 8 Maize meal 35.80% 0.87% 0.11% ND ND ND 9 Maize meal 76.27% 5.56% 0.08% Below LOQ Below LOQ 0.09% 10 Maize meal 32.17% 0.82% 0.06% ND ND ND 11 Maize meal ND ND ND ND ND ND 81 Appendix B Maize products (continued) Sample HT event Percentage Percentage Sample Limit of HT event Limit of identification NK603 standard standard description quantification GA21 detected quantification number detected deviation deviation 12 Maize meal 1.28% 0.01% 0.03% Below LOQ Below LOQ 0.07% 13 Maize meal 32.55% 0.58% 0.07% ND ND ND 14 Maize meal 30.45% 1.51% 0.08% ND ND ND 15 Maize meal 6.44% 0.52% 0.11% ND ND ND 16 Maize meal 35.97% 1.48% 0.05% ND ND ND 17 Maize meal 4.77% 0.69% 0.03% Below LOQ Below LOQ 0.05% 18 Maize meal 24.95% 0.48% 0.05% ND ND ND 19 Maize meal 22.81% 1.37% 0.07% ND ND ND 20 Maize meal 80.61% 3.14% 0.83% ND ND ND 21 Maize meal 27.07% 1.88% 0.03% 0.72% 0.03% 0.09% 22 Maize meal 26.62% 1.48% 0.03% ND ND ND 23 Beer powder 26.24% 0.63% 7.62% ND ND ND 82 Appendix B Maize products (continued) Sample HT event Percentage Percentage Sample Limit of HT event Limit of identification NK603 standard standard description quantification GA21 detected quantification number detected deviation deviation 24 Beer powder 40.47% 0.71% 0.02% ND ND ND 25 Maize grits 54.63% 3.35% 0.03% ND ND ND 26 Maize grits 42.51% 2.31% 0.03% Below LOQ Below LOQ 0.16% 27 Maize grits 55.18% 1.04% 0.07% Below LOQ Below LOQ 0.16% 28 Maize grits 16.39% 1.02% 0.05% Below LOQ Below LOQ 0.11% 29 Maize grits 21.50% 2.01% 0.06% Below LOQ Below LOQ 0.16% 30 Maize rice 67.47% 1.79% 0.07% ND ND ND 31 Maize rice 49.41% 0.99% 0.08% ND ND ND 32 Maize rice 9.72% 0.71% 0.03% Below LOQ Below LOQ 0.09% 33 Polenta 0.25% 0.01% 0.03% ND ND ND 34 Polenta 53.65% 0.34% 0.03% ND ND ND 35 Polenta 11.55% 0.39% 0.04% ND ND ND 83 Appendix B Maize products (continued) Sample HT event Percentage Percentage Sample Limit of HT event Limit of identification NK603 standard standard description quantification GA21 detected quantification number detected deviation deviation 36 Polenta 31.56% 0.30% 0.03% ND ND ND 37 Polenta ND ND ND ND ND ND 38 Corn flakes ND ND ND ND ND ND 39 Corn flakes ND ND ND ND ND ND 40 Corn flakes ND ND ND ND ND ND 41 Corn flakes ND ND ND ND ND ND 42 Corn flakes ND ND ND ND ND ND 43 Corn flakes ND ND ND ND ND ND 44 Corn flakes ND ND ND ND ND ND 45 Corn chips 11.54% 0.83% 0.05% ND ND ND 46 Corn chips 41.65% 3.69% 0.02% ND ND ND 47 Corn chips 33.15% 2.23% 0.05% ND ND ND 48 Corn chips 35.33% 0.97% 0.05% ND ND ND 84 Appendix B Maize products (continued) Sample HT event Percentage Percentage Sample Limit of HT event Limit of identification NK603 standard standard description quantification GA21 detected quantification number detected deviation deviation 49 Corn chips 10.76% 1.68% 0.05% ND ND ND 50 Corn chips 20.36% 2.82% 0.08% ND ND ND 51 Corn chips 17.54% 0.51% 0.05% ND ND ND 52 Corn chips 16.35% 0.62% 0.05% ND ND ND 53 Corn chips 12.22% 0.27% 0.07% ND ND ND 54 Corn chips ND ND ND ND ND ND 55 Maize pasta 21.15% 1.11% 0.02% ND ND ND 56 Maize pasta ND ND ND ND ND ND 57 Maize pasta ND ND ND ND ND ND ND: Not detected 1 The limit of quantification is based on the copy number of the reference gene HMG in maize and Lectin in soybean 85 Appendix B A summary of the detected HT events, standard deviation and limit of quantification for the corn-soy blends. Sample HT event Percentage HT event Percentage HT event Percentage Sample Limit of Limit of Limit of identification NK603 standard 1 GA21 standard GTS40-3-2 standard description quantification quantification quantification number detected deviation detected deviation detected deviation Corn-soy 58 0.04% 0.01 0.04% ND ND ND ND ND ND blend Corn-soy 59 0.12% 0.01% 0.08% ND ND ND Below LOQ Below LOQ 0.07% blend Corn-soy 60 ND ND ND ND ND ND 0.14% 0.01% 0.07% blend Corn-soy 61 ND ND ND ND ND ND ND ND ND blend Corn-soy 62 16.70% 0.28% 0.02% ND ND ND 48.65% 1.73% 0.40% blend Corn-soy 63 ND ND ND ND ND ND 4.01% 0.11% 0.07% blend ND: Not detected 1 Limit of quantification is based on the copy number of the reference gene HGM in maize and Lectin in soybean 86 Appendix B A summary of the detected HT events, standard deviation and limit of quantification for the soybean products. Sample HT event Percentage Sample Limit of identification GTS40-3-2 standard description quantification1 number detected deviation 64 Soybean ND ND ND 65 Soy milk 9.57% 0.76% 0.07% 66 Soy milk Below LOQ Below LOQ 0.07% 67 Soy milk ND ND ND 68 Soy milk Below LOQ Below LOQ 0.05% 69 Soy milk ND ND ND 70 Soy milk ND ND ND 71 Soy milk 5.21% 1.31% 0.10% 72 Soy milk ND ND ND 73 Infants milk 0.07% 0.01% 0.05% 74 Soy flour ND ND ND ND: Not detected 1 Limit of quantification is based on the copy number of the reference gene HMG for maize and Lectin for soybean 87 Appendix B A summary of the detected HT events, standard deviation and limit of quantification for the texturized soy protein products. Sample HT event Percentage HT event Percentage HT event Percentage Sample Limit of Limit of Limit of identification NK603 standard 1 GA21 standard GTS40-3-2 standard description quantification quantification quantification number detected deviation detected deviation detected deviation Texturized 75 ND ND ND ND ND ND ND ND ND soy protein Texturized 76 27.75% 1.00% 0.33% ND ND ND 84.70% 2.02% 1.61% soy protein Texturized 77 43.22% 2.22% 0.19% ND ND ND 92.09% 1.26% 1.09% soy protein Texturized 78 ND ND ND ND ND ND 16.75% 0.32% 0.12% soy protein Texturized 79 ND ND ND ND ND ND Below LOQ Below LOQ 0.06% soy protein Texturized 80 ND ND ND ND ND ND 78.10% 3.16% 0.01% soy protein Texturized 81 54.38% 2.19% 0.27% ND ND ND 56.55% 2.38% 6.88% soy protein ND: Not detected 1 Limit of quantification is based on the copy number of the reference gene HMG for maize and Lectin for soybean 88 Appendix C APPENDIX C Four Parameter Logistic 1.2 1 0.8 R² = 0.99 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 3 3.5 4 Concentration Figure A: Example of the Four Parameter Logistic curve generated to determine the glyphosate concentration of the maize and soybean food products in parts per billion (ppb). The %B/B0 was calculated for the standards by dividing the mean absorbance value of each the standard by the mean absorbance value of the zero standard. The %B/B0 calculated for the five standards (red) were used to generate a Four Parameter Logistic curve with a R² of > 0.98. The %B/B0 calculated for each sample (mean absorbance of sample/mean absorbance of zero standard) was plotted on the Four Parameter Logistic curve to determine the glyphosate concentration present in the samples in ppb. 89 %B/Bo Appendix C 38 38 NK603 HMG 28 28 NK603 R² = 0.99 HMG R² = 0.99 18 18 10 1000 100000 10 1000 100000 Copy number Copy number 38 38 GA21 HMG 28 28 GA21 R2 = 0.99 HMG R2 = 0.99 18 18 10 1000 100000 10 1000 100000 Copy number Copy number 38 38 GTS40-3-2 Lectin 28 28 GTS40-3-2 R² = 0.99 Lectin R² = 0.99 18 18 10 1000 100000 10 1000 100000 Copy number Copy number Figure B: Example of the standard curve generated during quantitative Real- time PCR to determine the percentage of GM HT event NK603, GA21 and GTS40- 3-2 in maize and/or soybean products. Four copy number standards (blue) in duplicate were used to generate a standard curve with a R² of > 0.98. 90 Ct Ct Ct Ct Ct Ct