Genetically modified (GM) crops were first commercialized in the mid-1990s and currently are planted on over 90% of corn, cotton and soybean acres in the United States (USDA-NASS 2017). GM crop adoption continues to increase globally, due to their economic and sustainability benefits (Anderson et al. 2016; ISAAA 2016). Most commercial GM crops containing insect protection traits currently rely on genes derived from Bacillus thuringiensis (Bt) to provide selective protection against economically important pests. The safety of Bt as a source of insecticidal genes for GM crops is well established (Box 1). Bt was initially developed as a microbial pesticide spray and has a history of safe use in agriculture when applied, as intended, on food and feed crops (US-EPA 1998). Bt is ubiquitous in the environment (Schnepf et al. 1998), non-toxic to mammals and does not have pathogenic or allergenic properties (US-EPA 1998).
Pseudomonads are rod-shaped, aerobic, gram-negative bacteria. Certain Pseudomonas species have previously been reported to have entomopathogenic properties and represent a promising source of insecticidal genes for use in GM crops (Kupferschmied et al. 2013). A gene, ipd072Aa, from Pseudomonas chlororaphis, which encodes the IPD072Aa protein, has recently been reported to confer protection against certain coleopteran pests when expressed in maize plants (Schellenberger et al. 2016).
The safety assessment framework for GM crops is well established and has been adopted globally to evaluate a variety of trait types, including those for insect protection (Codex Alimentarius Commission 2009; EFSA 2006; FAO/WHO 1991). The assessment includes, in part, an evaluation of each introduced trait, including its source organism, for potential adverse pathogenic, toxic and allergenic effects (Delaney et al. 2008). This paper provides an assessment of the safety of P. chlororaphis as a gene source for GM crops. Like Bt, certain species of Pseudomonas including P. chlororaphis are ubiquitous in the environment, have a history of safe use in agriculture as seed treatments, foliar-applied biopesticides and as a gene source for GM crops, and lack known pathogenic, toxic or allergenic properties. This information supports, in part, the safety assessment of potential traits, such as IPD072Aa, derived from this source organism.
Ubiquity in the environment
The genus Pseudomonas has been well studied and is estimated to contain over 100 species and 10 sub-species (Gomila et al. 2015; Peix et al. 2009). Sequence analysis of conserved housekeeping genes has provided information on the phylogenetic relatedness of Pseudomonas species within the genus (Anzai et al. 2000; Garrity et al. 2005; Gomila et al. 2015; Moore et al. 2006). Pseudomonas species have been classified into 7 groups: P. syringae, P. chlororaphis, P. fluorescens, P. putida, P. stutzeri, P. aeruginosa and P. pertucinogena (Anzai et al. 2000; Fig. 1). P. chlororaphis contains four subspecies: P. chlororaphis subsp. aurantiaca, P. chlororaphis subsp. aureofaciens, P. chlororaphis subsp. chlororaphis and P. chlororaphis subsp. piscium (Burr et al. 2010).
Most Pseudomonas species, including P. chlororaphis, are ubiquitous in the environment, have widespread distribution in soil and water (Peix et al. 2009) and perform a range of economic services and ecological functions. Some Pseudomonas species inhabit the rhizosphere, are associated with plant roots and provide benefits to the plant by competing with soil-borne plant pathogens and protecting against fungal pests (Anderson and Kim 2018; Kupferschmied et al. 2013; Mauchline and Malone 2017). P. chlororaphis, specifically, has been reported to promote plant growth, stimulate microbial communities and protect plants by producing compounds (e.g., phenazine-type antibiotics, hydrogen cyanide, chitinases and proteases) that inhibit fungal growth (EFSA 2015b), insects and nematodes (Anderson and Kim 2018). Other Pseudomonas species protect plants by preventing colonization by deleterious microorganisms (Mendes et al. 2011).
Certain Pseudomonas species have been utilized in a variety of applications, including the biological control of phytopathogens (Walsh et al. 2001), promotion of plant growth (Mercado-Blanco and Bakker 2007), phosphate solubilization (Rodríguez and Fraga 1999) and bioremediation of organic compounds (Moore et al. 2006; Peix et al. 2009). Many Pseudomonas species have a history of safe use in agriculture and other sectors (EFSA 2015b; Montie 1998). For example, certain Pseudomonas species are entomopathogenic and are being utilized as biopesticides to provide plant protection against insect pests. Insecticidal toxins in the genome of P. entomophilia have been identified (Luiu et al. 2013), and P. fluorescens has been shown to exert insecticidal activity against aphids, termites and other agricultural pests (Kupferschmied et al. 2013). Similarly, other species of Pseudomonas, including P. chlororaphis, P. protegens and P. aeruginosa, have demonstrated insecticidal activity (see Table 2 of Kupferschmied et al. 2013). Because of their role in plant protection and defense, P. chlororaphis and other Pseudomonas species with biopesticidal activity are being marketed for use as seed-treatment and foliar-applied biopesticides or as gene donors for GM crops (Kupferschmied et al. 2013).
History of safe use in agriculture
Pseudomonas-based biopesticides and plant protection products
Several biopesticide products containing Pseudomonas species that provide protection against fungal pathogens and diseases have been developed and assessed for their safety (Table 1). For example, two strains of P. syringae (ESC-l0 and ESC-11) have been shown to control post-harvest mold contamination on certain fruits, and dry rot and silver scurf on potatoes during storage (US-EPA 1999b, 2001a, 2009b). The products developed with these strains emphasize the long history of safe use of Pseudomonas-based biopesticides, as they were first registered with the United States Environmental Protection Agency (US-EPA) in 1990 and 1996 (US-EPA 2017b). Over the past 30 years, several additional Pseudomonas-based biopesticides and plant protection products have been registered with the US-EPA or approved by the European Food Safety Authority (EFSA); this further demonstrates the long history of safety (Table 1). For example, Pseudomonas sp. DSMZ 13134, which is closely related to P. fluorescens, has been shown to provide protection against fungal diseases in vegetables and flowers (Buddrus-Schiemann et al. 2010; EFSA 2012), and P. aureofaciens strain Tx-1 has been shown to provide protection against fungal pathogens on golf course turf (US-EPA 1999a, 2000). P. chlororaphis strain AFS009 is being leveraged to provide protection against a range of soil-borne fungal pathogens (AgBiome 2017; US-EPA 2017a), and other strains of P. chlororaphis (strains MA 342 and 63-28) have been shown to control fungal pathogens in cereals (EFSA 2017; Johnsson et al. 1998), as well as in greenhouse ornamentals and vegetable crops (US-EPA 2001b, d). In addition to fungal protection, Pseudomonas-based products are used to protect plants against frost damage. For example, P. syringae is known to protect plant leaves from frost through ice nucleation (Hirano and Upper 2000), and a non-frost-forming strain of P. fluorescens (strain A506) is being used to reduce frost damage on fruit and vegetable crops (Nufarm Americas Inc. 2012; US-EPA 1992b). The same strain of P. fluorescens is also being used to suppress pathogenic bacterial growth (e.g., fire blight and russet inducing bacteria) on apple and pear crops (Nufarm Americas Inc. 2012; US-EPA 1992b), whereas P. fluorescens strain D7 is being used to suppress growth of certain invasive grass species (US-EPA 2014).
As part of the registration requirements of biopesticide products, environmental and human health risk assessments are conducted prior to commercialization (US-EPA 2017c). The US-EPA concluded that these Pseudomonas strains are low risk, therefore these strains were granted exemptions from the requirement for a tolerance (40 CFR Parts 180.1114, 180.1145, 180.1212, 180.1304, 180.1326 and 180.1341). The human health and environmental safety of P. chlororaphis strain 63-28 and P. aureofaciens strain Tx-1 have been reviewed by the US-EPA. Both strains were determined to have no toxicity or human health concerns (US-EPA 2000, 2001d). Similarly, the human health and environmental safety of P. chlororaphis strains MA 342 and DSMZ 13134 have been reviewed by the European Commission (EC 2002; Velivelli et al. 2014) and EFSA (2012, 2017). For strain MA 342, the European Commission acknowledged that there were no signs of toxicity or pathogenicity based on a rat acute oral study, and P. chlororaphis is unlikely to grow at mammalian body temperature (EC 2002); EFSA recommended additional studies to finalize the risk assessment (EFSA 2017). For DSMZ 13134, EFSA concluded that this strain of P. chlororaphis is unlikely to cause toxicity or pathogenicity via oral exposure based on clinical and other experimental data (EFSA 2012).
Pseudomonas syringae strains ESC-10 and ESC-11 and P. fluorescens strain A506 were registered with the US-EPA in the early 1990s. According to the US-EPA, these strains of P. syringae pose low risk to humans or birds because they do not survive at temperatures above 32 °C, and they do not cause adverse effects in mammals when ingested, inhaled or applied topically (US-EPA 2009b). Similarly, P. fluorescens is ubiquitous in the environment, is not generally considered to be a human or animal pathogen (US-EPA 1992a) and is not expected to have adverse ecological effects on avian wildlife, aquatic organisms, non-target insects, mammalian systems or endangered species (US-EPA 1992a, 2009a).
Pseudomonas species and related species as a gene source for GM Crops
Certain Pseudomonas species and related species have also served as gene sources for genetically modified crops (Table 1). The GM crop products developed with these strains also emphasize the long history of safe use of Pseudomonas species as gene donors, as the first GM crop containing a gene from P. chlororaphis was deregulated by the United States Department of Agriculture (USDA) in 1995 (USDA-APHIS 1995, 2017). Event 8338 tomato (OECD Unique Identifier CGN-89322-3) was developed by Monsanto (Monsanto Company 1995). These GM tomatoes contain a gene from P. chlororaphis that encodes the 1-amino-cyclopropane-1-carboxylic acid deaminase (ACCd) enzyme, which has been shown to delay ripening when expressed in tomato plants by reducing ethylene production.
Similarly, in 2013 and 2014, the USDA deregulated four herbicide tolerant GM soybean and cotton varieties that were developed with genes from P. fluorescens and Delftia acidovorans (USDA-APHIS 2017). The gene from P. fluorescens encodes the hydroxyphenylpyruvate dioxygenase (HPPD) protein, which has been demonstrated to confer tolerance to isoxaflutole (IFT) herbicides when expressed in plants. Bayer CropScience developed herbicide tolerant soybean event FG72 (OECD Unique Identifier MST-FGØ72-2) using the HPPD W366 gene from P. fluorescens strain A32 (USDA-APHIS 2013). The gene from D. acidovorans has been demonstrated to confer tolerance to aryloxyalkanoate herbicides by expression of the aryloxyalkanoate dioxygenase-12 (AAD-12) protein. Herbicide tolerance traits have been developed using the aad-12 gene from D. acidovorans in soybean and cotton by Dow AgroSciences LLC [OECD Unique Identifiers DAS-44406-6; DAS-68416-4 and DAS-81910-7 (USDA-APHIS 2014a, b, 2015), respectively]. Delftia acidovorans was previously classified as Pseudomonas acidovorans and Comamonas acidovorans, before being reclassified recently as Delftia (Dow AgroSciences 2010; Tamaoka et al. 1987). The safety of both P. fluorescens and D. acidovorans as a gene sources for GM crops has been assessed by several regulatory authorities [for example, EFSA (2015a), FSANZ (2013), USDA-APHIS (2013) and CFIA (2013), FSANZ (2014), Health Canada (2014), USDA-APHIS (2014c), respectively]. Based on this and other evidence, GM soybean containing the gene from P. fluorescens and the GM soybean and cotton events containing the gene from D. acidovorans have been approved by several regulatory authorities globally (ISAAA 2018).
Pathogenic, toxic or allergenic properties
As previously reviewed by Leuschner et al. (2010), regulatory authorities in the US and Europe concluded that P. chlororaphis strains used for plant protection purposes pose no health concerns for humans (EC 2002; US-EPA 2001d). Additionally, P. chlororaphis was previously reviewed by EFSA using a Qualified Presumption of Safety (QPS) Approach (EFSA 2015b), which included a thorough assessment of the species’ life history characteristics, commercial uses and safety concerns. The thorough review of P. chlororaphis safety resulted in a general consensus that it is non-pathogenic to humans and livestock because of its inability to grow and proliferate at mammalian body temperatures (EC 2002). Based on this weight of evidence, P. chlororaphis was determined to be safe for biocontrol applications (Chen et al. 2015).
While there have been a few reports where P. chlororaphis has been isolated from animals with disease or illness (for example, Hatai et al. 1975), these reports are rare and there has been no causal link to clinical illness (EC 2002; EFSA 2015b). As part of the QPS evaluation, microorganisms are considered within the context that they are “deliberately introduced in the food chain either directly or as a source of additive or food enzyme” (Leuschner et al. 2010). The QPS assessment does not consider the organism’s safety for use as a gene source for GM crops, therefore the utility of this QPS assessment is limited to applications where the organism is either used directly or as a source of additive or food enzyme in food and feed applications. The QPS assessment for P. chlororaphis noted that it may produce secondary metabolites (for example, rhamnolipids and phenazine compounds) (EFSA 2015b). However, the potential for a gene source to produce a secondary metabolite like rhamolipids or phenazine compounds does not indicate inherent risk for the GM crop. Secondary metabolites like rhamolipids or phenazine compounds are synthesized through complex biochemical pathways involving multiple genes. For example, rhamnolipids biosynthesis occurs in sequential reactions catalyzed by RhlA, RhlB and RhlC proteins [under the control of the rhlA, rhlB and rhlC genes, respectively (Gunther et al. 2005; Reis et al. 2011)]. Biosynthesis of phenazine compounds is controlled by phz genes (Dowling and O’Gara 1994). The safety of the specific gene inserted into the plant and gene products is assessed as part of the safety assessment of GM crops, and there is no evidence to suggest that the ipd072Aa gene from P. chlororaphis is involved in the biosynthesis of secondary metabolites like rhamnolipids or phenazine compounds.
Phylogenetic relatedness to known human and plant pathogens
There is currently a robust understanding of the phylogenetic relatedness within the genus Pseudomonas (Anzai et al. 2000; Burr et al. 2010; Garrity et al. 2005; Gomila et al. 2015; Moore et al. 2006). The Pseudomonas genus does contain some well-recognized plant and human pathogens, including P. aeruginosa and P. syringae (Peix et al. 2009). Therefore, the phylogenetic relatedness of pathogenic Pseudomonas species and other Pseudomonas species intended for agricultural applications should be considered before potential use. P. aeruginosa is a gram-negative, aerobic bacterium that is relatively ubiquitous in the environment and can be found in soil and water, as well as on the surface of plants. P. aeruginosa is well recognized as both a plant pathogen and an opportunistic human pathogen that can cause respiratory infection in immunocompromised patients (Sadikot et al. 2005). The pathogenicity of P. aeruginosa is thought to be related to virulence factors carried by pathogenicity islands. For example, the pathogenicity islands PAPI-1 and PAPI-2 have been linked to the virulence of P. aeruginosa. It has been confirmed that P. chlororaphis does not contain virulence factors and shares no genomic homology with these known pathogenicity islands (Chen et al. 2015). P. aeruginosa is phylogenetically distant from P. chlororaphis (Anzai et al. 2000; EC 2002; Fig. 1).
The pathogenicity of P. syringae to plants is well understood. The taxonomy of the species is separated into pathovars, each distinguishable based on the primary host plant(s) and carbon source(s) they utilize for growth (Garrity et al. 2005). The plant pathogenicity of P. syringae is based on an array of phytotoxins that produce disease symptoms. For example, P. syringae pathovar syringae disrupts the plasma membrane in host plants via production of syringomycins, syringopeptins and syringotoxins. P. syringae is phylogenetically distant from P. chlororaphis (Anzai et al. 2000; Fig. 1). Additionally, it has been confirmed that P. chlororaphis does not contain the genes that code for the biosynthesis of these or other phytotoxins or exoenzymes (cellulases, pectinases, pectin lyases) that compromise plant cell walls (EFSA 2015b).
While it is important to consider phylogenetic relatedness to known pathogens, identifying a pathogen in the same genus as a potential source donor for a GM crop does not indicate inherent risk. Many species share phylogenetic relatedness with known pathogens without being pathogenic themselves. For example, the phylogenetic relatedness of species belonging to the Bacillus genus has been published previously based on 16S rRNA gene sequences (see Fig. 2 in Alcaraz et al. 2010). While Bt shares distant phylogenetic relatedness with a few pathogens (e.g., Bacillus anthracis; Alcaraz et al. 2010), it has a long history of safe use as a biopesticide and as a gene source for GM crops (US-EPA 1998, 2001c). Similarly, the phylogenetic relatedness of species belonging to the Streptomyces genus has been published previously based on 16S rRNA gene sequences (see Fig. 1 in Kämpfer 2006). Very few species of Streptomyces are human, animal or plant pathogens (Kämpfer 2006). For example, Streptomyces scabiei is a well-known plant pathogen associated with potato scab (Zhang et al. 2016), and Streptomyces somaliensis is a human pathogen that causes deep tissue and bone infections (Kirby et al. 2012). Even though phylogenetically related to these pathogens, the safety of Streptomyces viridochromogenes as a gene source for GM crops is well established (OECD 2007).