BioControl 48: 611–636, 2003. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

FORUM

Assessing the risks of resistant transgenic plants on entomophagous : Bt-maize expressing Cry1Ab as a case study

Anna DUTTON, Jörg ROMEIS and Franz BIGLER* Swiss Federal Research Station for Agroecology and Agriculture (FAL), Reckenholzstr. 191, 8046 Zurich, Switzerland ∗ Author for correspondence; e-mail: [email protected]

Received 19 December 2002; accepted in revised form 15 September 2003

Abstract. One of the primary concerns related to the adoption of insect resistant trans- genic plants in the environment is the detrimental effect that these may pose on non-target organisms, including entomophagous arthropods (parasitoids and predators) which have an important function in regulating pests. Despite the fact that regulatory bodies require informa- tion regarding the potential risk of releasing transgenic plants in the environment, to date, no specific protocols have been designed for assessing the risks of insect resistant transgenic crops on entomophagous arthropods. Here a framework for risk assessment is proposed to evaluate the effects of insect resistant plants on entomophagous arthropods. Using maize expressing the Bacillus thuringiensis gene which codes for the Cry1Ab toxin, we illustrate the procedure necessary for assessing the risks. As a first step, it is required to determine which entomo- phagous arthropods play a major role in regulating maize pests, and which may be at risk. Because the risk which transgenic plants pose to entomophagous arthropods depends on both, their exposure, and their sensitivity to the insecticidal protein, it is essential to determine, as a second step, if and at what level organisms are exposed to the transgene compound. Exposure will be associated with the feeding behaviour of phytophagous and entomophagous arthropods together with the tissue and cell specific temporal and spatial expression of the insecticidal protein. For those organisms which could potentially be exposed to the insecticidal protein, sensitivity tests, as a third step, should be performed to assess toxicity. The testing procedure and the type of tests which should be adopted to quantify the effects of insect resistant plants on natural enemies are subsequently illustrated. Taking the green lacewing Chrysoperla carnea as an example, we propose a procedure on how to perform tests and give evidence that Bt-maize poses no risk to this predator.

Key words: Chrysoperla carnea, exposure, genetically engineered crops, parasitoids, preda- tors, risk assessment, test procedure, tiered system, toxicity

Introduction

To date, the only insect resistant transgenic plants that are commercially available are those expressing genes which code for Bacillus thuringiensis 612 ANNA DUTTON ET AL.

(Bt) toxins. The amount of area being cultivated with these crops is rapidly increasing (James, 2002), and other genes coding for new Bt-toxins, lectins, proteinase or α-amylase inhibitors, and other insecticidal products have been successfully engineered in plants (Schuler et al., 1998; Jouanin et al., 1998). Some of these plants are being tested at the field scale, such as peas (Pisum sativum) expressing the gene coding for common bean α-amylase inhib- itors (αAIs) (Morton et al., 2000). Moreover, in 2002, over 200 applications for release permits to conduct field tests with 11 different transgenic crops expressing various insect resistant genes were notified in the USA alone (ISB, 2002), indicating that the adoption of various transgenic plants is likely to increase. Similar to other plant protection technology, insect resistant transgenic plants bear risks and benefits to the environment (NAS, 2002). The primary ecological concerns to the release of transgenic plants include those related to their possible invasiveness in ecosystems, out-crossing, horizontal gene transfer, development of pest resistance, and effects on non-target organisms (Conner et al., 2003). Effects of GM plants on non-target entomophagous arthropods (predators and parasitoids) have been a major concern as these organisms often play an important role in natural pest regulation, and are considered to be of economic value. Moreover, this group of organisms may be a good indicator of potential ecological impacts of transgenic plants as they belong to the third trophic level in the food chain (Groot and Dicke, 2002). Although we are aware that other non-target arthropods such as herbivores, pollen feeders (bees), soil arthropods as well as other organisms including birds, mammals, and fish could be affected by transgenic plants, here we shall limit ourselves to the assessment of the risks of transgenic plants on entomophagous arthropods. Despite the fact that regulatory bodies in different countries require a detailed environmental risk assessment for the release of transgenic plants (Nap et al., 2003), it is often debated what and how to measure (Conner et al., 2003). This is distinctively the case for the assessment of non-target organisms as these include a large number of . Moreover, these organ- isms can potentially be affected by insect-resistant transgenic plants through various ways (Schuler et al., 1999; Groot and Dicke, 2002). For example, the impacts on non-target entomophagous arthropods can be due to direct toxic effects through exposure to the insecticidal protein, indirect effects via reduc- tion in prey/host quantity and/or quality, or indirect effects due to unintended changes of plant properties (chemical or physical) caused by the insertion of a new gene (pleotropic effects or insertional mutagenesis). Although this complexity can make testing and assessment difficult, uncertainty can be BT-MAIZE EXPRESSING CRY1AB AS A CASE STUDY 613 minimised by selecting appropriate species, and by conducting suitable tests to produce meaningful crop specific results. Despite the fact that regulatory agencies require data concerning the effects of GM crops on entomophagous arthropods, to date there is no standard procedure indicating what should be assessed and how tests should be conducted. Recently two different sequential approaches have been suggested for selecting and testing the effects of Bt-maize (Schmitz et al., 2003) and protease inhibitor-expressing crops (Cowgill and Atkinson, 2003) on non-target herbivorous arthropods. However, there is until now no sequen- tial approach to risk assessment of insect resistant transgenic plants on non-target entomophagous arthropods. In this article, we first give an overview of the current requirements to assess the risk of transgenic plants on non-target organisms for the commer- cialisation of GM plants in the US and the EU, and its shortcomings. Second, taking Bt-maize expressing the cry1Ab gene as a model system, we demon- strate the sequential steps that need to be taken in order to make a selection of entomophagous arthropods which could be at risk, and present a testing procedure for assessing the effects. Our purpose is to provide a conceptual testing framework for the selection of organisms and apply an existing testing procedure (tiered system) for assessing the effects. Moreover, we provide a review of the literature concerning the effects of Bt-maize on beneficial arthropods and make a detailed evaluation on the risks of Bt-maize on the green lacewing Chrysoperla carnea (Stephens) (Neuroptera: Chrysopidae) which has been the only predatory insect shown to be affected by Bt-maize in laboratory and greenhouse conditions (Hilbeck et al., 1998a, b; Dutton et al., 2002).

Regulatory status of test requirements for non-target organisms and their shortcomings

In the US, according to the National Academy of Sciences, the trait of a GM plant should be the focus of the risk assessment, and not the process by which it has been produced (NAS, 2000). As a result the Environmental Protection Agency (EPA), requires that risk assessment tests are performed on non-target organisms to assess toxicity of so called plant-incorporated protectants. This is in contrast to EU regulatory procedures which consider GM plants as something new, for which both direct (toxic) and indirect effects should be assessed prior to their commercial release (Directive 2001/18/EC) (EC, 2001). Since in the US the major concern regarding the potential effects of GM plants expressing insecticidal proteins on non-target organisms is that related 614 ANNA DUTTON ET AL. to the toxicity of the protein contained in the plant, test requirements on non-target organisms remained similar to those required for plant protection products. This means testing the effects on a set of representative such as bees, ladybirds and lacewings (NAS, 2002; EPA, 2002) using a tiered testing system in which the 1st tier reflects a maximum hazard approach. Organisms are exposed to the pure insecticidal protein at a range of doses and concentrations 10 to 100 times higher than those expected in the environment (EPA, 2000). Since these tests do not provide any information on the possible non-intended indirect effects that could occur once the plants are released in the field, the EPA often requires information from field studies (EPA, 2003). Field studies, however, have often the drawback of being expensive and lack appropriate statistical power. In the EU, a more extensive procedure is required for the commercial- ization of transgenic plants. The assessment should consider the potential immediate and delayed effects resulting from direct and indirect interac- tions of the GM plant with non-target organisms. The assessment should include effects on organisms which interact with the target pest, and impact on population levels of competitors, herbivores, symbionts, parasites and patho- gens (Directive 90/220/EEC). However, there is no indication as to which organisms should be selected and tested nor on the type of tests that should be conducted to make the assessment. Although monitoring has been considered an important part of the risk assessment procedure for the commercialization of GM plants in the EU, to date there is also no protocol to be followed for monitoring studies (Saeglitz and Bartsch, 2003). These two contrasting regulatory systems, one which requires the testing of a few non-target organisms that often are selected irrespective of the GM crop (EPA, 2003), and the other which requires the assessment of an infinite number of organisms, lead us to propose a scheme which could be used for the selection of relevant predators and parasitoids, in the specific crop, which may be exposed to the insecticidal compound and which could be affected. Applying the already established tiered testing system used for evaluation of plant protection products, we suggest some modifications for the type of tests that should be conducted. The proposed tests aim to take into consider- ation and measure both direct and indirect effects of GM plants on beneficial arthropods.

Principles of a tiered testing procedure for transgenic plants

For ecotoxicological evaluation of plant protection products, arthropods including bees, predators, parasitoids and soil organisms are required to be tested in both the US (EPA 712-C-96-280, February 1996) and in the EU BT-MAIZE EXPRESSING CRY1AB AS A CASE STUDY 615

Figure 1. Sequential test procedure for assessing the effects of insect resistant transgenic plants on entomophagous arthropods. 1based on results, additional species might need to be selected for testing, 2based on amenability, semi-field tests can be skipped.

(Directive91/414/EEC). A tiered testing approach has been developed and is generally used for assessing the effects of pesticides on these different organisms. A similar approach is recommended for assessing insect resistant transgenic plants (Figure 1). The procedure starts with laboratory tests (1st tier), followed by semi-field (2nd tier) and field (3rd tier) tests (Barrett et al., 1994; Hassan, 1998; Candolfi et al., 2000; Candolfi et al., 2001). However, given that the mode of exposure of arthropods to the insecticidal protein expressed by the plant differs to contact sprayed insecticides, the fact that non-target organisms may continuously be exposed to the insecticidal compound, and the possible unintended changes on plant quality due to the transformation, some adaptations to the testing procedure presently used for pesticides are suggested for the assessment of insect resistant transgenic plants.

Selection of entomophagous species to test

As a first step in the tiered testing procedure, species to be tested need to be selected (Figure 1). According to test guidelines that have been developed by the Office of Prevention, Pesticides and Toxic Substances (EPA), three species should be tested representing at least two of the following groups – parasitic dipterans, predaceous hemipterans, predaceous cole- 616 ANNA DUTTON ET AL. opterans, predaceous mites, predaceous neuropterans, parasitic hymenop- terans (OPPTS 885.4340). For EU registration procedures of plant protection products, risk assessment tests have to be conducted with four to six arthro- pods. However, recently it has been proposed that only two sensitive indicator species (Aphidius rhopalosiphi and Typhlodromus pyri) are used. These two species were selected over the course of several years during which at least 23 entomophagous species and 95 pesticides have been evaluated (Candolfi et al., 2001). For assessing transgenic plants, it is suggested that the selection of species is done according to their economic and ecological importance for natural pest regulation in the target crop, and according to the likelihood of exposure to the insecticidal protein. Since exposure depends on several factors, including feeding behaviour of both herbivores and entomophagous arthropods, availability of prey, and expression of the insecticidal protein in the plant, previous knowledge and information is required about the agro- ecosystem and transgenic plant in order to select relevant arthropods that could be at risk. The choice of arthropods should also include a wide range of taxonomic groups. Knowledge on the specificity of the insecticidal protein may also assist in the selection of organisms to test. Although it is important to select arthropods that are amenable and available for laboratory and semi- field testing, species which are not relevant in the specific crop or not exposed to the toxin should not be selected. We recognise that non-target herbivorous arthropods feeding on the trans- genic crop may also be affected (lethally or sub-lethally). These effects may lead to a reduction in populations which may affect organisms in higher trophic levels. For registration of pesticide products, tests on non-target herbivorous arthropods are generally not required, despite the fact that most insecticides affect to some extent non-target herbivores that occur in the crop. By adopting the same principle, one can argue that risk assessment for non-target herbivores should not be required for registration of insect resistant transgenic plants. However, given the season-long expression of insecticidal proteins in transgenic plants it would be useful to have informa- tion on the effects of transgenic plants on some non-target herbivores. From an ecological perspective, and for integrated pest management strategies, such information would be valuable.

Risk assessment – evaluating effects and trigger values

In the EU, a hazard quotient (HQ) approach for 1st tier tests will soon substitute the beneficial capacity (E) value (Overmeer and van Zon, 1982) which is presently used for evaluating effects of plant protection products on entomophagous arthropods (Candolfi et al., 2001). At present E values BT-MAIZE EXPRESSING CRY1AB AS A CASE STUDY 617 from four to six organisms, which are selected according to the crop in which the product will be used, are calculated according to the effect of the product on the survival and reproduction of the organisms (Barrett et al., 1994). Depending on the degree at which plant protection products reduce the beneficial capacity of natural enemies, they are classified in one of the four IOBC (International Organisation for Biological and Integrated Control of Noxious and Plants) categories relative to their toxicity when applied on an inert surface. A threshold scale for determining the level of risk has been established for laboratory tests, with E values lower than 30 repre- senting no risks, values between 30–79 representing low risk, 80–99 medium, and above 99 high risk (Hassan, 1992). In contrast, the newly proposed HQ approach, which shall substitute E values, takes into account the crop specific application rate of pesticides for in-field and drift rates together with the LR50 (lethal rate causing 50% mortality) generated with two indicator species. HQ threshold value that exceed two, require further testing and/or that additional organisms are assessed, or that a risk management program is established (Candolfi et al., 2001). Since exposure of entomophagous arthropods to insecticidal proteins expressed in transgenic plants differs from sprayed pesticides, and tests for assessing effects of transgenic plants on natural enemies have not yet been standardised, we suggest that life table parameters including survival, devel- opment and reproduction of the selected organisms are assessed. Statistical differences (Type I error levels set at 5%) between treatment and control (e.g. organisms fed with Bt containing and non-Bt containing diets) should lead to the decision to perform further laboratory tests (behavioural, and/or physiolo- gical studies) or proceed to higher tier tests (see below). It is advisable that statistical tests are accompanied by a power analysis to avoid committing type II errors (Steidl et al., 1997; Marvier, 2002). As information is acquired, the most sensitive and relevant parameters to measure, and threshold values, may be established in a similar manner as it has been done with plant protection products.

Laboratory 1st tier tests

In 1st tier tests, worst case toxicity (sensitivity) studies should be performed with the selected species using both the pure insecticidal protein and trans- genic plant material. Dose-response test should be performed using pure insecticidal protein incorporated into appropriate artificial diets and fed to the selected species. Concentrations should be chosen to determine the LC50 (lethal concentration causing 50% mortality). In cases where the substance is not enough toxic to kill 50%, the No-Observed-Effect-Concentration (NOEC) can be used. The NOEC concentration should be at least 100 times above the 618 ANNA DUTTON ET AL. maximum exposure of the organisms to the compound under field conditions. Provided that little is known on the consumption of toxin that a specific organism may ingest under field conditions, the highest amounts of toxin measured in the plant tissue (environmental concentration) should be used as a base for calculating the NOEL, and tests conducted with 10–100 times the expected environmental concentration (EPA, 2000). When appropriate artificial diets for the entomophagous arthropod to test are not available, dose response tests could be conducted by providing the test species with prey or host which have been reared on artificial diet into which the pure insecticidal protein has been incorporated at various concentrations. Since pure insecticidal proteins and proteins expressed in the plant may not be identical, and possible interactions between toxicity and reduced food quality may result in effects, tests using transgenic plant material should also be performed. Natural enemies should be exposed to the toxin in a manner representing the worst case situation. This would include feeding the natural enemy exclusively with a suitable herbivore, which is known to ingest the toxin when reared on the transgenic plant. The use of a suitable herbivore is recommended to avoid testing effects which are not related to the transgenic plant. However, a susceptible herbivore could be selected in order to include potential indirect effects. If no effects are observed in worst case 1st tier tests, no further testing is needed for that given species. If effects are observed with the pure insecticidal protein and with plant material, 2nd tier tests are required. If the species is highly sensitive it is advisable to test a closely related second species in the 1st tier. If effects are observed with either the pure toxin or the plant material it would be necessary to either go to the next tier or to do additional laboratory behavioural studies to investigate the likelihood and degree of exposure of the organism to the toxin and/or physiological studies to understand the toxicity mechanisms of the compound in order to characterise the risk.

Semi-field 2nd tier tests

In contrast to 1st tier tests, where the primary goal is to examine whether entomophagous species are sensitive to the toxin when exposed continuously to the insecticidal protein, the objective of the 2nd tier test is to assess toxicity at exposure levels representing more closely the field situation. These tests should allow undisturbed movement and searching behaviour of both ento- mophagous and phytophagous species on whole plants. Tests should involve single species, where only one prey species is offered. The type of endpoint (lethal or sub-lethal effects) to be assessed will depend on findings of the previous 1st tier laboratory test. BT-MAIZE EXPRESSING CRY1AB AS A CASE STUDY 619

For specific questions and organisms it may be advisable to skip the 2nd tier tests and to perform field tests. This would be the case when appropriate 2nd tier testing methods are not available or when insects are not amenable for such tests.

Field 3rd tier tests

Field trials for regulatory purposes are indicated only if the question of acceptable risks can not be addressed in lower tier tests. Field experiments should go beyond the mere collection and identification of non-target species. They should provide more ecological information (i.e. prey populations in the field and exposure of predators to contaminated prey) and answer questions related to the effects observed in laboratory and semi-field tests. Field tests should be targeted to those species where effects in lower tiers were observed. The effect of GM plants on selected arthropod populations and defined communities should be the major questions addressed in field experiments. However, it should be noted that effects identified in field trials are caused by a multitude of biotic, abiotic, direct and indirect factors that make it extremely difficult to assign population changes of a species or a functional group of organisms to a single factor. Valuable information on interactions and ecosystem response may be gained, but data are often difficult to analyse and to interpret due to the high complexity and to spatial and temporal variation. Field studies are not the first choice to answer the questions of regulatory testing of transgenic crops within reasonable time. Using power analysis, results of a five year Farm-Scale-Evaluations (FSE) study in England, showed that the planned replication should consist of at least 60 pairs of fields (half- field experimental units) over three years to provide sufficient information from which valid significant differences can be drawn (Perry et al., 2003). Thus, field tests could be performed as part of an obligatory monitoring program of commercially grown transgenic crops but not necessarily for registration purposes.

Case study – Bt-maize

To assess the risks of any insect resistant transgenic plant on non-target arthropods, as a first step, it is necessary to identify arthropods that belong to the crop in the specific region into which the transgenic plant will be introduced. Here we shall limit the assessment to entomophagous arthropods belonging to the maize agroecosystem in central and western Europe. To identify entomophagous arthropods that could potentially be exposed to the 620 ANNA DUTTON ET AL. toxin, and to assess the toxicity of Bt-maize to the exposed arthropods, knowl- edge and information is required on (a) phytophagous and entomophagous arthropod species in maize, (b) expression of the Bt-toxin in the plant, (c) ingestion of the toxin by both phytophagous and entomophagous arthropods.

Non-target arthropods in the maize system

Phytophagous arthropods

The most frequently reported maize herbivores in western and central Europe are listed in Table 1. It is evident that the European corn borer, aphids, wire- worms together with the frit fly are the predominant species. Some of these species appear important only regionally as is the case with spider mites (Tetranychus sp.) in Spain (Eizaguirre and Albejas, 1987) or the corn weevil (Tanymecus dilaticollis) in Bulgaria/Hungary (Gerginov, 1987). Although the list does not include all species feeding on maize plants, it provides an overview of the most abundant species which are of primary economic importance. Moreover, these species contribute as major prey and hosts for entomophagous arthropods.

Entomophagous arthropods

Entomophagous arthropods in the maize crop represent a diverse group, which encompasses a wide ecological and taxonomic spectrum with the classes of Insecta and Arachnida. Considering the available literature on maize, the most frequently cited predators and parasitoids in central and western Europe are listed in Table 2. Although the list is not comprehensive, it provides an overview of the groups which are likely to play a role in regulating pests in maize. Indeed, it demonstrates that several species of Coccinellidae, Carabidae, Syrphidae, Chrysopidae, Anthocoridae, as well as various parasitoids of aphids and the European corn borer are regularly present in maize and their function as pest regulators is repeatedly reported. However, there is a lack and need of information regarding the specific parasitoid species attacking other common maize herbivores. Given that natural enemies to be considered in the risk assessment should be selected according to their importance in pest regulation and should include a wide range of species which belong to different taxonomic groups (Figure 1), we have chosen form Table 2 the seven most commonly occurring natural enemies recorded in the literature. These include Coccinella septem- punctata, Bembidion spp., Episyrphus balteatus, Aphidoletes aphidimyza, Chrysoperla carnea, Orius spp.,andAphidius spp. BT-MAIZE EXPRESSING CRY1AB AS A CASE STUDY 621

Table 1. Most frequently reported phytophagous arthropods in maize in central and western Europe

Herbivores Common names References

Lepidoptera Ostrinia nubilalis European corn borer 1, 2, 3, 4, 5, 7, 8, 9, 10, 13 Agrotis ipsilon black cutworm 1, 7, 9 Sesamia nonagriodes pink stem borer 2

Hemiptera: Aphididae 4, 12 Metopolophium dirhodum rose-grainaphid 1,2,3,5,6,7,11 Rhopalosiphum padi bird cherry-oat aphid 1, 2, 3, 5, 6, 7, 9, 10, 11 Sitobion avenae Englishgrainaphid 1,2,3,5,6,7,11

Coleoptera: Elateridae Agriotes spp. wireworm or click beetle 1, 2, 3, 5, 7, 8, 9

Coleoptera: Chrysomelidae Oulema melanopa cereal leaf beetle 1, 3, 10

Coleoptera: Curculionidae Tanymecus dilaticollis corn weevil 3, 9

Diptera: Chloropidae Oscinella frit fritfly 1,5,7,8,9,10

Thysanoptera: Thripidae 5 Frankinella spp. thrips 10 Haplothrips aculeatus rice aculeated thrips 4

Arachnida: Tetranychidae Tetranychus urticae two-spotted spider mite 2, 10 1, Hurle et al. (1996); 2, Eizaguirre and Albajes (1987); 3, Gerginov (1987); 4, Kot and Bilewicz-Pawinska (1987); 5, Longauerova (1987); 6, Plewka and Pankanin-Franczyk (1989); 7, Häni et al. (1987); 8, Anglade (1975); 9, Dolinka (1975); 10, Schmitz and Bartsch (2001); 11, Cabanettes (1985); 12, della Giustina et al. (1987); 13, Reh (1985).

Expression of Bt-toxin in maize

Expression of Bt-toxins in maize is often cited in the literature to be constitutive, meaning that expression occurs in all tissues at all times. This is misleading since different promoters have been used for the various commer- cial maize hybrids (Table 3) and these different hybrids have been shown to express different amounts of toxin in different plant tissues (Table 4). For example, three of the four commercial Bt-maize constructs (Bt11, MON810 622 ANNA DUTTON ET AL.

Table 2. Commonly occuring entomophagous arthropods (predators and parasitoids) in maize in central and western Europe

Natural enemies References Natural enemies References

Polyphagous predators Coleoptera: Coccinellidae Dermaptera: Forticulidae Coccinella septempunctata 1, 2, 3, 4, 5, 6, 7, 8, 9, 17 Forficula auricularia 3, 4, 5 Propylaea Dermaptera: Labiduridae quatuordecimpunctata 2, 3, 4, 5, 7, 8, 9, 17 Labidura riparia 5 Adalia bipunctata 3, 4, 6, 7, 8, 9 Adonia variegata 5, 7, 17 Aranea 3, 5, 6, 17 Coccinella quinquepunctata 1, 7, 8 Coleoptera:Carabidae 1 Thysanoptera: Aeolothripidae Dermetrias atricapillus 5 Aelothrips intermedius 18 Harpalus spp. 5, 10 Bembidion spp. 5, 9, 10, 11 Aphid specific parasitoids Poecillus cupreus 5, 9 Hymenoptera: Aphidiidae 1 Agonum dorsale 5, 9, 11 Aphidius spp. 2, 3, 8, 9 Trechus spp. 9 Praon volucre 2, 3, 8 Platysma vulgare 9 Ephedrus plagiator 2, 3 Coleoptera: Staphylinidae 1 Hymenoptera: Aphelinidae Tachyporus obtusus 6 Aphelinus sp. 2 Oxytelus rugosus 6 Paederidus litoralis 6 Corn borer specific parasitoids Hymenoptera: Diptera: Syrphidae 1, 2, 6, 16 Trichogrammatidae Episyrphus balteatus 3, 4, 5, 7, 9 Trichogramma spp. 1 Melanostoma balteatus 7 Hymenoptera: Ichneumonidae Sphaerophoria scripta 7 Sinophorus turionus 12 Syrphus corollae 7, 9 Eriborus terebrans 13 melanostoma 7, 9 Hymenoptera: Braconidae Metasyrphus corollae 7 Microgaster tibialis 14 Melanostoma mellinum 7 Diptera: Cecidomyiidae Diptera: Tachinidae 1 Aphidoletes aphidimyza 1, 2, 3, 5, 9 Lydella thomsoni 15

Neuroptera: Chrysopidae 9 Chrysoperla carnea 1, 3, 4, 5, 6, 7, 8 Chrysoperla perla 8 Chrysoperla phyllochroma 8

Heteroptera: Anthocoridae Orius spp. 1, 2, 3, 4, 5, 9, 17 Anthocoris nemorum 6 Anthocoris limbatus 6 Heteroptera: Nabidae 18 Nabis provencalis 5 Nabis limbatus 6

1, Bailly et al. (1984); 2, della Giustina et al. (1987); 3, Katz (1993); 4, Attia (1985); 5, Asin and Pons (1998); 6, Reh (1980); 7, Kokubo (1986); 8, Plewka and Pankanin-Tranczyk (1989); 9, Fougeroux (1984); 10, Lövei and Szentkiralyi´ (1984); 11, Häni et al. (1987); 12, Caga´nˇ and Bokor (1998); 13, Bokor and Caga´nˇ (1999a); 14, Bokor and Caga´nˇ (1999b); 15, Caga´nˇ et al. (1999); 16, Visnyovszky and Racz´ (1989); 17, Kiss et al. (2003). BT-MAIZE EXPRESSING CRY1AB AS A CASE STUDY 623

Table 3. Commercial maize constructs expressing Bacillus thuringiensis genes

Constructs Company Cry protein Promoter

Event176* Novartis/Mycogen Cry1Ab PEPC + pollen specific promoter Bt11 Novartis Cry1Ab CaMV 35S MON810 Monsanto Cry1Ab CaMV 35S CBH-351 AgrEvo/PGS Cry9C CaMV 35S

*No longer registered. and CBH-351) express the toxin under the CaMV 35S promoter (Table 3). In these constructs Bt-toxin expression is higher in leaves and roots compared to pollen (Table 4). In contrast Event 176 contains two promoters (PEPC and a pollen specific promoter), and expression of the Bt-toxin is highest in leaves and pollen, but low in roots. Expression of Bt-toxins at the cellular level is often unknown in commer- cially available transgenic plants, and this is important to know as various piercing-sucking arthropods feed on specific cells. Various methods could be used for providing such information. For example, immunohystochemical tests have been used to detect other proteins expressed by genetically engi- neered plants expressing different genes (Battraw and Hall, 1990; Narváez- Vasquez et al., 1992; Rao et al., 1998). A microcapillary cell content extrac- tion technique (Brandt et al., 1999), together with immunological assays have also been used for the detection of gene expression in specific cells and for quantifying transgene protein at the cellular level. Applying the microcapil- lary technique Raps et al. (2001) demonstrated that Bt-toxin is not present in the phloem sap of two maize constructs (Event176 and Bt11). However, care must be taken not to assume that for other plant species or other maize constructs this observation holds true. For example, in different varieties of transgenic rice containing different promoters (including CaMV 35S), results suggest that Bt-toxin is transported in phloem, as honeydew of phloem feeding insects was found to contain the toxin (Bernal et al., 2002).

Ingestion of Bt-toxin by arthropods and exposure in maize

Exposure of phytophagous arthropods to Bt-toxin

Considering the feeding behaviour of herbivores, and the expression of Bt- toxin in maize (i.e. Bt11), it can be established that all major plant chewing herbivores ingest the toxin upon feeding on Bt-maize (Table 5) given that plant parts fed upon contain the toxin (Table 4). In contrast, exposure to the 624 ANNA DUTTON ET AL.

Table 4. Expression of Bt-toxin in various plant parts of commercial maize constructs

Construct Leaf Root Pollen Stalk Seed/kern Source ------µg/gfreshweight------

Event176 2.8 - 4.4 0.08 7.1 0.08 0.05 EPA, 2000 Event176 2.1 - 3.3 0.08 5.0–11.0 0.08 0.05 Fearing et al., 1997

Bt11 3.3 2.2–37.0* <90** nd 1.4 EPA, 2000 Bt11 4.8–28.9 nd nd nd nd CFIA, 2000

MON810 7.9 -10.3 nd <90** nd 0.19–0.39 EPA, 2000 MON810 7.9–10.3 nd nd nd 0.19–0.39 CFIA, 2000

CBH-351 44*** 25.9*** 0.24*** 2.8*** 18.6*** EPA, 2000 CBH-351 24.3–39.0** nd nd 84.8** 1.8** Jansens et al., 1997 CBH-351 225.0 nd nd 18.0 12.0 Jansens et al., 1997

*Values represent ng/mg protein. **Values represent ng/g dry weight. ***Values represent µg/g dry weight. nd = not detected. toxin differs among the different piercing-sucking herbivores. Aphids do not ingest the toxin when either reared on Bt11, Event176 (Raps et al., 2001; Dutton et al., 2002) or Mon810 (Head et al., 2001). Spider mites and thrips which feed on mesophyll cells ingest the toxin when reared on Bt11 (Dutton et al., 2002; Obrist et al., in prep.).

Exposure of entomophagous arthropods to Bt-toxin

Compared to herbivores which are exposed to Bt-toxin through direct feeding on plant material, entomophagous arthropods can be exposed to the toxin through direct or indirect routes. Three routes of exposure have been identi- fied for entomophagous arthropods and these include (a) feeding on ‘contam- inated’ prey/hosts or parasitising ‘contaminated’ hosts, (b) direct feeding on plant substrates (pollen, nectar or plant sap) and (c) feeding on honeydew excreted from homopterans (indirect plant feeding). For commercialised Bt- maize constructs (Mon810, Bt11, Event 176) this last route of exposure does not exist, as Bt-toxin is not ingested or excreted by aphids. However, this holds true only for these maize constructs on which only two aphid species (R. padi and R. maidis) have been tested (Head et al., 2001; Raps et al., 2001; Dutton et al., 2002). Considering the feeding range of the predominant entomophagous arthro- pods in maize in central and western Europe, together with the expression of the Bt-toxin, we can establish that from the seven most frequently occurring predator and parasitoid species, only five can potentially ingest the Bt-toxin, BT-MAIZE EXPRESSING CRY1AB AS A CASE STUDY 625

Table 5. Site and mode of feeding, and ingestion of Bt-toxin of most frequently reported phytophagous herbivores in maize in central and western Europe

Herbivores Feeding site Mode of feeding Ingestion of the Bt-toxin

Cutworms Stalk and leaf tissue Chewing Yes Aphids Phloem sap Piercing-sucking No1 Wireworm Roots, stalk and leaf tissue Chewing Yes Ceareal leaf beetle Leaf tissue Chewing Yes Corn weevil Leaf tissue Chewing Yes Frit fly Leaf tissue Rasping-sucking Yes Thrips Epidermis/mesophyll Piercing-sucking Yes Spider mites Mesophyll Piercing-sucking Yes

1No Bt-toxin detected in phloem (Raps et al., 2001). based on our knowledge on their feeding behaviour (Table 6). Parasitoids belonging to the genus Aphidius, and the predator A. aphidimyza are unlikely to be exposed to the Bt-toxin, as these species feed primarily on aphids, and as adults they feed on nectar and/or honeydew, which do not contain the toxin. According to our risk assessment procedure, we suggest that 1st tier tests for registration of Bt-maize in central and western Europe should be conducted with Coccinella septempunctata, Bembidion spp., Episyrphus balteatus, Chrysoperla carnea and Orius spp., the five species that can be exposed to the toxin.

Toxicity of Bt-maize for entomophagous arthropods

Various laboratory studies have been published with regard to assessment of the sensitivity of entomophagous arthropods to the Cry1Ab toxin produced by maize (Table 7). From the five suggested species to test (see above) only two have been assessed to date (C. carnea and Orius spp). In addition, tests have been performed with Coleomegilla maculata, a Coccinellidae not occurring in Europe. The methods used for assessing the effects of Bt-maize or the Cry1Ab toxin on the arthropods differed from one study to another. In fact, none of the studies measured if and how much of the Bt-toxin was ingested by the entomophagous arthropods, and how the quantities provided relate to the exposure levels in the field. According to our knowledge on the mode of exposure we can assume that ingestion took place in all but one study (Lozzia et al., 1998) in which C. carnea was fed with aphids reared on Bt-maize. 626 ANNA DUTTON ET AL. Kulp et al. (1989); 6 . (1984); Bailly et al 5 Digilio and Pennacchio (1989). 11 Polgar (1984); 10 Sutherland et al. (1999); 4 Corey et al. (1998); 9 LarvaAdult Aphids, eggs, mites, thrips, larvae*Larva Aphids, eggs, mites, thrips, larvae*Adult Pollen Larvae*, eggs, aphids, thrips, mitesLarva Pollen Larvae*, eggs, aphids, thrips, mitesAdult AphidsLarvaAdult Honeydew AphidsLarva Yes Adult Larvae*, eggs, aphids, thrips, mitesNymph Yes Adult Larvae*, eggs, aphids, thrips, mites Larvae*, eggs, aphids, thrips, mites Pollen, sapLarva Yes Pollen,Adult sap Yes Aphids Honeydew Pollen Honeydew Honeydew Yes Yes Yes Honeydew Pollen Yes No Honeydew Honeydew No No Yes Honeydew No No Holopainen and Helenius (1992); 3 2 , 1 6 5 , 8 , 4 7 Sheldon and Macleod (1971); 8 11 Triltsch (1999); , 3 2 10 spp. 9 spp. Food sources for larvae and adults of the most commonly documented entomophagous arthropods in maize in central and western spp. Natural enemiesPredators Coccinella septempunctata Stage Prey/host Plant material Other Exposure Bembidion Episyrphus balteatus Aphidoletes aphidimyza Chrysoperla carnea Orius Parasitoid Aphidius Hodek (1996); Bay et al. (1993); Table 6. Europe and possible exposure to Bt-toxin 1 7 *Lepidopteran, coleopteran and dipteran larvae. BT-MAIZE EXPRESSING CRY1AB AS A CASE STUDY 627

Table 7. Laboratory sensitivity tests of entomophagous arthropods to Bt-toxin (Cry1Ab) expressed in maize

Natural enemies Toxin offered in the form of Effects Reference

Coleoptera: Coccinelidae Coleomegilla maculata Maize pollen None Pilcher et al., 1997

Heteroptera: Anthocoridae Orius insidiosus Maize pollen None Pilcher et al., 1997 Maize silk None Al-Deeb et al., 2001 Orius majusculus Thrips reared on Bt-maize None Zwahlen et al., 2000

Neuroptera: Chrysopidae Chrysoperla carnea Bt-maize pollen and lepidopteran eggs None Pilcher et al., 1997 Cry1Ab incorporated in artificial diet Yes1 Hilbeck et al., 1998a Lepidopteran larvae reared on Bt-maize Yes1 Dutton et al., 2002; Hilbeck et al., 1998b Spider mites reared on Bt-maize None Dutton et al., 2002 Aphids reared on Bt-maize None Dutton et al., 2002; Lozzia et al., 1998 Cry1Ab incorporated in sugar None Romeis et al., 2004 solution

1Effects observed in mortality and development.

Results from all studies show that none of the tested predators were affected by the Cry1Ab Bt-toxin. This was with the exception of the green lacewing C. carnea (Table 7). Although a decrease of the European corn borer in Bt-maize fields has been shown to lead to a reduction in parasitoids of the target pest (Wold et al., 2001; Bourguet et al., 2002), this is to be expected when using Bt-maize or any other pest control method.

Assessing effects of Bt-maize on the predator Chrysoperla carnea

Various laboratory tests have been performed to determine the effects of Bt-maize on C. carnea (Table 7). Results obtained vary according to how the tests were conducted. With our present knowledge of C. carnea and of Bt-toxin expressed in maize, we here perform a systematic stepwise evalu- ation providing advice on how to carry out appropriate tests and to obtain conclusive evidence on whether or not Bt-maize poses a risk to this predator. As a first step in the assessment (Figure 2) potential routes of exposure to the transgene product must be determined. Based on the feeding behaviour of both the predator and prey and the expression of the toxin in maize, it was previously established that this predator is potentially exposed to the Bt-toxin 628 ANNA DUTTON ET AL. . Chrysoperla carnea Assessing the risks of Bt-maize on the predator Figure 2. BT-MAIZE EXPRESSING CRY1AB AS A CASE STUDY 629

(Table 6), by feeding either on spider mites or on insect larvae (coelopteran, lepidopteran or dipteran). However, exposure is not likely to occur through aphids, often the primary prey, or by feeding on arthropod eggs. Given that different field environments offer C. carnea a variety of prey, it is not easy to establish the likelihood or amount of toxin that this predator would ingest. If spider mites are the only prey available, amounts of Bt-toxin ingested could be substantial, as these have been found to contain high concentrations of Bt when reared on transgenic maize. Less amounts of Bt-toxin would be ingested by C. carnea when only feeding on lepidopteran larvae (Dutton et al., 2002). For C. carnea adults, which feed on pollen and sugar sources, exposure to Bt-toxin would be possible only through the consumption of pollen (Table 6). However, exposure would depend on the movement of adults in and outside the crop, the duration of pollen shed together with the persistence of pollen, and the toxin content in pollen for the maize variety grown. GiventhatinBt-maizeC. carnea larvae and adults are potentially exposed to the Bt-toxin, susceptibility of the predator towards the toxin should be assessed. According to the suggested first tier test (Figure 1), toxicity assess- ment should be performed using a dose response test (to establish the LD50), or in the case where no toxicity is noted, using the NOEC. The predator should be fed with the insecticidal protein by incorporating into an artificial diet. However, because the pure toxin may not be equivalent to that expressed by the plant, an assessment should also be performed using plant material. A suitable herbivore which is reared on the transgenic plant should be used as prey. If effects are observed in these initial tests, behavioural test to determine exposure, and/or physiological tests to establish mode of toxicity would be advisable before 2nd or 3rd tier tests are performed (Figure 1). Feeding C. carnea larvae with the Cry1Ab toxin, either incorporated in a complex diet (100 µg/ml) or through the Lepidoptera prey herbivore, Spodop- tera littoralis, reared on Bt-maize, caused negative effects on the survival and development of the predator (Hilbeck et al., 1998a, b; Table 7). Considering that Cry1Ab is a Lepidoptera-specific toxin, the result that C. carnea is sensi- tive was surprising. Moreover, the work has received criticism as unrealistic amount of toxin was provided to the predator through the artificial diet and because the predator was fed with S. littoralis, which is a sub-optimal prey. Although these results would lead to 2nd tier tests, further 1st tier tests were performed to assess toxicity of Cry1Ab on C. carnea, and behavioural tests to determine likelihood of exposure. Assessment of toxicity has been conducted by using another herbivore, namely spider mites which were reared on Bt-maize and fed to C. carnea larvae. No negative effects on development, survival or weight of C. carnea were observed, even though the amount of the toxin measured in spider mites 630 ANNA DUTTON ET AL. reared on Bt-maize was three times higher than that measured in lepidopteran larvae (S. littoralis) (Dutton et al., 2002). In addition, toxicity was assessed by feeding C. carnea larvae with a sucrose solution in which the toxin was incorporated at a concentration of up to 1000 µg/ml (Romeis et al., 2004). Entomophagous arthropods including chrysopids are known to use carbohydrate sources to prolong their life-span when prey is scarce (Limburg and Rosenheim, 2001). No differences were observed on the survival of the first instar chrysopid larvae when fed the sucrose solution into which the toxin was incorporated, compared to larvae fed with the pure sucrose solution. Given that sucrose does not provide nutrients for further development, the predator was provided with a protein rich food source (Ephestia kuehniella eggs) after six days of feeding only on the Bt-sucrose solution. No detrimental effects on the development of chrysopids were observed, reflecting that Bt-toxin had no effect on nutrient assimilation of C. carnea. The amount of toxin ingested by C. carnea larvae was quantified using enzyme-linked immunosorbent assay. Amounts were shown to be higher than those observed when C. carnea was fed solely on S. littoralis larvae reared on Bt-maize, reflecting that amounts were above those encountered in the field (Romeis et al., 2004). To assess exposure of C. carnea to Bt-toxin, behavioural feeding prefer- ence bioassay were conducted. C. carnea larvae were given a choice between lepidopteran larvae (S. littoralis) containing the toxin and aphids (Meier and Hilbeck, 2001). A clear aphid feeding preference of C. carnea larvae was observed, independent on whether the prey was reared on Bt-maize or not, suggesting that lepidopteran larvae are not a preferred prey for this predator. Additional behavioural tests were performed to determine how successful C. carnea is in capturing S. littoralis on maize plants. More than 70% of 1st instar S. littoralis larvae encountered by 3rd instar C. carnea escaped predation (Dutton, unpublished). Considering the results on exposure and toxicity of the Cry1Ab toxin on C. carnea we can conclude that this predator is not affected when feeding on either aphids or spider mites kept on Bt-maize. The only route of exposure where effects might be observed is when predators feed on lepidopteran larvae. These effects might be attributed to the fact that lepidopteran larvae are themselves affected by the toxin. However, feeding of C.carnea on only lepidopteran larvae is unlikely to occur as they are not a preferred prey, and often escape predation. This leads us to suggest that no further assessment in higher tier tests should be required. However, several field studies assessing the effects of Bt-maize have been requested by regulatory agencies and results confirm our observations that Bt-maize poses no risk to chrysopids (Orr and BT-MAIZE EXPRESSING CRY1AB AS A CASE STUDY 631

Landis, 1997; Pilcher et al., 1997b; Wold et al., 2001; Bourguet et al., 2002; Candolfi et al., 2004). The presented case study with Bt-maize and the thorough evaluation on the exposure and the sensitivity of C. carnea to the Cry1Ab toxin have demonstrated that a tiered testing approach is an appropriate way to evaluate effects of genetically modified insecticidal plants for registration purposes. This concept can guide regulators and investigators to a systematic step- by-step and case-by-case procedure in which the relevant questions will be answered with accuracy and efficiency. Using such an approach will not only provide valid data for an effective evaluation, but it will also give valuable information on mechanisms (exposure, toxicity) causing the effects, informa- tion which is crucial for performing a comprehensive risk assessment for the registration of GM plants.

Acknowledgements

This work was supported be the Swiss National Science Foundation project number 5002-57820. We thank Ellen Hütter for assembling some of the information presented and anonymous reviewers for their comments.

References

Al-Deeb, M.A., G.E. Wilds and R. Higgins, 2001. No effect of Bacillus thuringiensis corn and Bacillus thuringiensis on the predator Orius insidiosus (Hemiptera: Anthocoridae). Environ. Entomol. 30: 625–629. Anglade, P., 1975. Corn pest management in western Europe exemplified by French systems. International Project on Ostrinia nubilalis, Rep. 2: 42–46. Asín, L. and X. Pons, 1998. Aphid predators in maize fields. IOBC Bull. 21(8): 163–170. Attia, B.M., 1985. Ökologische Beziehungen zwischen Maisblattläusen, polyphagen Präda- toren und dem Maiszünsler Ostrina nubilalis Hbn. Ph.D. Thesis, Institut für Phytomedizin der Unversität Hohenheim, Germany. 96 pp. Bailly, R., G. Dubois, A. Fougeroux, J.P Gendrier et al., 1984. Les Auxiliaires: Enemies naturels des revageurs des cultures. Association de Coordination Technique Agricole, Paris, France. 64 pp. Barrett, K., N. Grandy, E.G. Harrison and S. Hassan, 1994. Guidance document on regulatory testing procedures for pesticides with non-target arthropods. In: P. Ommen (ed), SETAC, Brussels. 50 pp. Battraw, M.J. and T.C. Hall, 1990. Histochemical analysis of CaMV 35S promoter-ß- glucuronidase gene expression in transgenic rice plants. Plant Mol. Bio. 15: 527–538. Bay, T., M. Hommes and H.P. Plate, 1993. Die Florfliege Chrysoperla carnea (Stephens): Überblick über Systematik, Verbreitung, Biologie, Zucht und Anwendung. Mitteilungen aus der Biologischen Bundesanstalt für Land- und Forstwirtschaft, Heft 288, Berlin. 175 pp. 632 ANNA DUTTON ET AL.

Bernal, C.C., R.M. Aguda and M.B. Cohen, 2002. Effect of rice lines transformed with Bacillus thuringiensis toxin genes on the brown planthopper and its predator Cyrtorhinus lividipennis. Entomol. Exp. Appl. 102: 21–28. Bokor, P. and L. Cagán,ˇ 1999a. Occurrence and bionomics of Erioborus terebrans (Graven- horst) (Hymenoptera: Ichneumonidae), a parasitoid of the European corn borer, Ostrinia nubilalis Hbn. (Lepidoptera: Pyralidae), in Central Europe. Plant Protect. Sci. 35: 17–22. Bokor, P. and L. Cagán,ˇ 1999b. Phenology, basic biology and parasitism of Microgaster tibialis (Hymenoptera, Braconidae), a parasitoid of the European corn borer, Ostrinia nubilalis, in Central Europe. Bio. Bratislava 54(5): 567–572. Bourguet, D., J. Chaufaux, A. Micoud, M. Delos, B. Naibo, F. Bombarde, G. Marque, N. Eychenne and C. Pagliari, 2002. Ostrinia nubilalis parasitism and the field abundance of non-target insects in transgenic Bacillus thuringiensis corn (Zea mays). Environ. Biosafety Res. 1: 49–60. Brandt, S., J. Kehr, C. Walz, A. Immlau, L. Willmitzer and J. Fisahn, 1999. A rapid method for detection of plant gene transcripts for single epidermal, mesophyll and companion cells in intact leaves. Plant J. 20: 245–250. Cabanettes, J.P., 1985. Mais: Bilan phytosanitaire. Phytoma 366: 15–16. Cagán,ˇ L. and P. Bokor, 1998. Sinophorus turionus Ratz, the parasitoid of the European corn borer, Ostrinia nubilalis Hbn in Slovakia, Czech Republic and Southwestern Poland. Acta Pythopathol. Entomol. Hung. 33: 435–445. Cagán,ˇ L., T. Turlings, P. Bokor and S. Dorn, 1999. Lydella thompsoni Herting (Dipt., Tachin- idae), a parasitoid of the European corn borer, Ostrinia nubilalis Hbn. (Lep., Pyralidae) in Slovakia, Czech Republic and south-western Poland. J. Appl. Ent. 123: 577–583. Candolfi, M., F. Bigler, P. Cambell, U. Heimbach, R. Schmuck et al., 2000. Principles for regu- latory testing and interpretation of semi-field and field studies with non-target arthropds. J. Pest Sci. 73: 141–147. Candolfi, M.P., K.L. Barrett, P.J. Cambell, R. Forster, N. Grandy, M-C. Huet, G. Lewis, P.A. Oomen, R. Schmuck and H. Vogt, 2001. Guidance document on regulatory testing and risk assessment procedures for plant protection products with non-target arthropods. ESCORT 2 Workshop held in Wageningen, The Netherlands. SETAC, Pensacola, FL, USA. 46 pp. Candolfi, M.P., K. Brown, C. Grimm, B. Reber and H. Schmidli, 2004. A faunistic approach to assess potential side-effects of genetically modified Bt-corn on non-target arthropods under field conditions. Biocontrol Science and Technology 14 (in press). CFIA, 2000. Decision document DD96-12; Determination of environmental safety of Northrup King seeds’ European corn borer (ECB) resistant corn (Zea mays L.). Available free at http://www.cfia-acia.agr.ca/english/plaveg/pbo/dd9612e.shtml. Conner, A.J, T.R. Glare and J-P. Nap, 2003. The release of genetically modified crops into the environment. Part II: Overview of ecological risk assessment. Plant J. 33: 19–46. Corey, D., S. Kambhampati and G. Wilde, 1998. Electrophoretic analysis of Orius insidiosus (Hemiptera: Anthocoridae) feeding habitats in field corn. J. Kansas Entomol. Society 71: 11–17. Cowgill, S.E. and H.J. Atkinson, 2003. A sequential approach to risk assessment of transgenic plants expressing protease inhibitors: effects on nontarget herbivorous insects. Transgenic Res. 12: 439–449. Digilio, M.C. and F. Pennacchio, 1989. Quantitative analysis of some biological parameters characterising an Italian population of Aphidius ervi Haliday (Hymenoptera, Braconidae, Aphidiinae). B. Lab. Entomol. Agri. Filippo-Silvestri 46: 59–74. Dolinka, B., 1975. Pest control in corn on the Danube plain illustrated by the Hungarian examples. International Project on Ostrinia nubilalis, Rep. 2: 35–41. BT-MAIZE EXPRESSING CRY1AB AS A CASE STUDY 633

Dutton, A., H. Klein, J. Romeis and F. Bigler, 2002. Uptake of Bt-toxin by herbivores feeding on transgenic maize and consequences for the predator Chrysoperla carnea. Ecol. Entomol. 27: 441–447. EC, 2001. Directive 2001/18/EC. Available free at: http://biosafety.ihe.be/PDF/2001_18.pdf. Eizaguirre, M. and R. Albajes, 1987. Pest status of maize arthropod pests in the northeast of Spain. In: F. Szentkiralyi (ed), International Symposium on Maize Arthropods, Gödöllö, Hungary. Plant protection Institute of the Hungarian Academy of Sciences, p. 18. EPA (Environmental Protection Agency), 2000. Bt plant pesticides biopesticides registration action document. II Science assessment: Product Characterization. Available free at: http://www.epa.gov/scipoly sap/2000/october/brad2_scienceassessment.pdf. EPA (Environmental Protection Agency), 2002. EPA’s Regulation of biotechnology for use in pest management (735-F-02-003). Available free at: http://www.epa.gov/pesticides/ biopesticides/reg_og_biotech/epategofbiotech.htm. EPA (Environmental Protection Agency), 2003. Event MON863 Bt Cry3Bb1 corn biopesti- cide registration action document. Available free at: http:www.epa.gov/pesticides/ biopesticides/ingredients/tech_docs/brad_006484.htm. Fearing, P.L., D. Brown, D. Vlachos, M. Meghji and L. Privalle, 1997. Quantitative analysis of Cry1A (b) expression in Bt maiz plants, and silage and stability of expression over successive generations. Mol. Breed. 3: 169–176. Fougeroux, A., 1984. Les insectes prédateurs et parasites de pucerons en culture de blé et de maïs. Phytoma June. pp. 35–39. Gerginov, L., 1987. Maize pests in Bulgaria and pest control. In: F. Szentkiralyi (ed), Inter- national Symposium on Maize Arthropods, Plant protection Institute of the Hungarian Academy of Sciences, Gödöllö, Hungary, p. 21. della Giustina W., P. Deriu and P. Foessel, 1987. Role of specific natural enemies in the control of maize aphid populations in the Paris area, preliminary results. SROP/WPRS Bull. 10(1): 12–22. Groot, A.T. and M. Dicke, 2002. Insect-resistant transgenic plants in a multi-trophic context. Plant J. 31: 387–406. Häni, F., G. Popow, H. Reinhard, A. Schwarz, K. Tanner and M. Vorlet, 1987. Integrierter Pflanzenschutz im Ackerbau. Verlag Landwirtschaftliche Lehrmittelzentrale, Zollikofen, Switzerland, 333 pp. Hassan, S.A. 1992. Meeting of the workshop group Pesticides and Beneficial Organisms, University of Southampton, UK September, 1991. IOBC/WPRS Bulletin XV(3): 1–3. Hassan, S.A., 1998. Introduction to standard characteristics of test methods. In: P.T. Haskell and P. McEwen (eds), Ecotoxicology: Pesticides and beneficial organisms.Kluwer Academic Publishers, Dordrecht, NL. pp. 55–68. Head, G., C.R. Brown, M.E. Groth and J.J. Duan, 2001. Cry1Ab protein levels in phyto- phagous insects feeding on transgenic corn: implications for secondary exposure risk assessment. Entomol. Exp. Appl. 99: 37–45. Hilbeck, A., W.J. Moar, M. Pusztai-Carey, A. Filippini and F. Bigler, 1998a. Toxicity of Bacillus thuringiensis Cry1Ab toxin to the predator Chrysoperla carnea (Neuroptera: Chrysopidae). Environ. Entomol. 27: 1255–1263. Hilbeck, A., M. Baumgartner, P.M. Fried and F. Bigler, 1998b. Effects of transgenic Bacillus thuringiensis corn-fed prey on mortality and development time of immature Chrysoperla carnea (Neuroptera: Chrysopidae). Environ. Entomol. 27: 480–487. Hodek, I., 1996. Food Relationships In: I. Hodek and A. Honek (eds), Ecology of Coccinel- lidae. Kluwer Academic Publisher, London, UK. pp 143–238. 634 ANNA DUTTON ET AL.

Holopainen, J.K. and J. Helenius, 1992. Gut content of ground beetles (Col., Carabidae), and activity of these and other epigeal predators during an outbreak of Rhopalosiphum padi (Hom., Aphidae). Acta Agr. Scand. B-S. P. 42: 57–61. Hurle, K., M. Lechner and K. König, 1996. Mais: Unkräuter – Schädlinge – Krankheiten. Verlag Th. Mann, Gelsenkirchen, Germany. pp. 1–116. ISB, 2002. Information systems for Biotechnology. Field test releases in the US. Available free at http://www.isb.vt.edu/cfdocs/fieldtests1.cfm. James, C., 2002. Global status of commercialised transgenic crops: 2002. ISAAA Briefs, no. 27. EuroCenter, Northwich, UK. pp. 1–24. Jansens S., A. van Vliet, C. Dickburt, L.P.C. Buysse, B. Saey, A. De Wulf, V. Gossele, A.G.E. Paez and M. Peferoen, 1997. Transgenic corn expressing a Cry9C insecticidal protein from Bacillus thuringiensis protected from European corn borer damage. Crop Sci. 37: 1616–1624. Jouanin, L., M. Bonade-Bottino, C. Girard, G. Morrot and M. Giband, 1998. Transgenic plants for insect resistance. Plant Sci. 131: 1–11. Katz, P., 1993. Analyse der Populationsdynamik von Maisblattläuse. Ph.D Thesis, Institut für Phytomedizin der Universität Hohenheim, Germany. 151 pp. Kiss, J., F. Szentkirályi, F. Tóth, Á. Szénási, F. Kádár, K. Árpás, D. Szekeres and C.R. Edwards, 2003. Bt-corn: Impact on non-target and adjusting to local IPM systems. In: T. Lelley, E. Balázs and M. Tepfer (eds), Ecological Impact of GMO Dissemination in Agro-Ecosystems, Facultas Verlags- und Buchhandels AG, Austria. pp. 157–172. Kokubu, H., 1986. Migration rates, in situ reproduction, and flight characteristics of aphido- phagous insects (Chrysopidae, Coccinellidae, Syrphidae) in corn fields. Ph.D. Thesis, Institute of Zoology, University of Basel, Switzerland. 184 pp. Kot, J. and T. Bilewicz-Pawinska, 1987. From studies of maize entomofauna in Warsaw region. In: F. Szentkiralyi (ed), International Symposium on Maize Arthropods,Plant protection Institute of the Hungarian Academy of Sciences, Gödöllö, Hungary. p. 34. Kulp, D., M. Formann, M. Hommes and H.P. Plate, 1989. Die räuberische Gallmücke Aphido- lotes aphidimyza (Rondani) (Diptera: Cecidomyiidae): Ein bedeutender Blattlausprädator- Nachschlagewerk zur Systematik, Verbreitung, Biologie, Zucht und Anwendung. Mitteilungen aus der Biologischen Bundesanstalt für Land- und Forstwirtschaft, Heft 250, Berlin. 126 pp. Limburg, D.D. and J.A. Rosenheim, 2001. Extrafloral nectar consumption and its influence on survival and development of an omnivorous predator, larval Chrysoperla plorabunda (Neuroptera: Chrysopidae). Environ. Entomol. 30: 595–604. Longauerova, J., 1987. Some problems with arthropods on maize in Czechoslovakia. In: F. Szentkiralyi (ed), International Symposium on Maize Arthropods, Plant protection Institute of the Hungarian Academy of Sciences, Gödöllö, Hungary. p. 36. Lövei, G. L. and F. Szentkirályi, 1984. Carabids climbing maize plants. Appl. J. Entomol. 97: 107–110. Lozzia, G.C., C. Furlanis, B. Manachini and I.E. Rigamonti, 1998. Effects of Bt corn on Rhopalosiphum padi L. (Rhynchota Aphididae) and on its predator Chrysoperla carnea Stephen (Neuroptera Chrysopidae). Boll. Zool. agr. Bachic. 30: 153–164. Marvier, M., 2002. Improving risk assessment for nontarget safety of transgneic crops. Ecol. Appl. 12(4): 1119–1124. Meier, M.S. and A. Hilbeck, 2001. Influence of transgenic Bacillus thuringiensis corn-fed prey and prey preference of immature Chrysoperla carnea (Neuroptera: Chrysopidae). Basic Appl. Ecol. 2: 35–44. BT-MAIZE EXPRESSING CRY1AB AS A CASE STUDY 635

Morton, R.L., H.E. Schroeder, K.S. Bateman, M.J. Chrispeels, E. Armstrong and T.J.V. Higgins, 2000. Bean alpha-amylase anhibitor 1 in transgenic peas (Pisum sativum) provides complete protection from pea weevil (Bruchus pisorum) under field conditions. P. Natl. Acad. Sci. 97: 3820–3825. Nap, J-P., P.L.J. Metz, M. Escaler and A.J. Conner, 2003. The release of genetically modified crops into the environment. Part I: Overview of current status and regulations. Plant J. 33: 1–18. Narváez-Vasquez, J., M.L. Orozco-Cardenas and C.A. Ryan, 1992. Differential expression of a chimeric CaMV-tomato proteinase inhibitor I gene in leaves of transformed nightshade, tobacco and alfalfa plants. Plant Mol. Biol. 20: 1149–1157. NAS (National Academy of Sciences), 2000. Genetically modified pest-protected plants: Science and regulation. Washington, DC: National Academy Press. Available free at http://www.napedu/books/0309069300/html/. NAS (National Academy of Sciences), 2002. Environmental effects of transgenic plants: The scope and adequacy of regulation. Washington, DC: National Academy Press. Available free at http://books.nap.edu/books/0309082633/html/index/html. Orr, D.B. and D.A. Landis, 1997. Oviposition of European corn borer (Lepidoptera: Pyralidae) and impact of natural enemy populations in transgenic versus isogenic corn. J. Econ. Entomol. 90: 905–909. Overmeer, W.P.J. and A.Q. van Zon, 1982. A standardised method for testing side-effects of pesticides on the predacious mite Amplyseius potentillae German (Acarina: Phytosdiidae). Entomophaga 27: 357–364. Perry, J.N., P. Rothery, S.J. Clark, M.S. Heard and C. Hawes, 2003. Design, analysis, and statistical power of the Farm-Scale Evaluations of genetically modified herbizide-tolerant crops. J. Appl. Eco. 40: 17–31. Pilcher, C.D., J.J. Obrycki, M.E. Rice and L.C. Lewis, 1997. Preimaginal development, survival and field abundance of insect predators on transgenic Bacillus thuringiensis corn. Environ. Entomol. 26: 446–454. Plewka, T. and M. Pankanin-Franczyk, 1989. Aphids and aphidophages on maize in central Poland. Acta Phytopathol. Acad. Sci. Hungar. 24: 169–171. Polgar, L., 1984. The role of aphid parasites (Hymenoptera: Aphidiidae) in the maize ecosystem. Acta Phytopathol. Acad. Sci. Hungar. 17: 139–145. Rao, K.V., K.S. Rathore, T.K. Hodges, X. Fu, E. Stoger, D. Sudhakar, S. Williams, P. Christou, M. Bharathi, D.P. Bown, K.S. Powell, J. Spence, A.M.R. Gatehouse and J.A. Gatehouse, 1998. Expression of snowdrop lectin (GNA) in transgenic rice plants confers resistance to rice brown planthopper. Plant J. 15: 469–477. Raps, A., J. Kehr, P. Gugerli, W.J. Moar and F. Bigler, 2001. Detection of Cry 1Ab in phloem sap of Bacillus thuringiensis corn and in the selected herbivores Rhopalosiphum padi (Homoptera: Aphidae) and Spodoptera littoralis (Lepidoptera: Noctuidae). Mol. Ecol. 10: 525–533. Reh, P., 1980 Vergleigende Untersuchungen über die jahreszeitlichen Veränderung der Fauna von Maisbeständen in der Oberrheinischen Riefebene und der Filderebene, unter beson- derer Berücksichtigung räuberichen Gegenspieler des Maiszünsler Ostrinia nubilalis. Diplomarbeit, Hohenheim, Germany. Reh, P., 1985. Untersuchungen zur Populationsdynamik des Maiszünslers Ostrinia nubilalis Hbn. Ph.D. Thesis, Institut für Phytomedizin der Universität Hohenheim, Germany. 131 pp. 636 ANNA DUTTON ET AL.

Romeis, J., A. Dutton and F. Bigler, 2004. Bacillus thuringiensis toxin (Cry1Ab) has no direct toxic effect on larvae of the green lacewing Chrysoperla carnea. J. Insect Physiol. (in press). Saeglist, C. and D. Baetsch, 2003. Regulatory and associated political issues with respect to Bt transgenic maize in the European union. J. Invertebr. Pathol. 83: 107–109. Schmitz, G. and D. Bartsch, 2001. Biozoenotische Untersuchungen in Maisfeldern bei Bonn und Aachen. Mitteilungen der Deutschen Gesellschaft für allgemeine und Angewandte Entomologie 13: 615–618. Schmitz, G., D. Bartsch and P. Pretscher, 2003. Selection of relevant non-target herbivores for monitoring the environmental effects of Bt-maize pollen. Environ. Biosafety Res.2: 117–132. Schuler, T.H., G.M. Poppy, B.R. Kerry and I. Denholm, 1998. Insect-resistant transgenic plants. Trends Biotech. 16: 168–175. Schuler, T.H., G.M. Poppy, B.R. Kerry and I. Denholm, 1999. Potential side effects of insect- resistant transgenic plants on arthropod natural enemies. Trends in Biotechnology 12: 210– 216. Sheldon, J.K. and E.G. MacLeod, 1971. Studies on the biology of the Chrysopidae II. The feeding behavior of the adult of Chrysoperla carnea (Neuroptera). Psyche 78: 107–121. Steidl, R.J., J.P. Hayes and E. Schauber, 1997. Statistical power analysis in wildlife research. J. Wildlife Manage. 61: 270–279. Sutherland, J.P., M.S. Sullivan and G.M. Poppy, 1999. The influence of floral character on the foraging behaviour of the hoverfly, Episyrphus balteatus. Entomol. Exp. Appl. 93: 157–164. Triltsh, H., 1999. Food remains in the guts of Coccinella septempunctata (Coleoptera: Coccinellidae) adults and larvae. Europ. J. Entomol. 96: 355–364. Visnyovszky, E. and V. Rácz, 1989. Investigation of Syrphids in maize stands. Acta Phyto- pathol. Entomol. Hungar. 24: 219–223. Wold, S.J., E.C. Burkness, W.D. Hutchison and R.C. Venette, 2001. In-field monitoring of beneficial insect populations in transgenic corn expressing a Bacillus thuringiensis toxin. J. Entomol. Sci. 36: 177–187. Zwahlen, C., W. Nentwig, F. Bigler and A. Hilbeck, 2000. Tritrophic interactions of transgenic Bacillus thuringiensis corn-fed Anaphothrips obscurus (Thysanoptera: Thripidae) with the predator Orius majusculus (Heteroptera: Anthocoridae). Environ. Entomol. 29: 846–850.