Ranking matrices as operational tools for the environmental risk assessment of genetically modified crops on non-target organisms

Ecological Indicators 36 (2014) 367–381

Contents lists available at ScienceDirect

Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind

Ranking matrices as operational tools for the environmental risk assessment of genetically modified crops on non-target organisms Angelika Hilbeck a,∗ , Gabriele Weiss a , Bernadette Oehen a,b , Jörg Römbke c , Stephan Jänsch c , Hanka Teichmann d , Andreas Lang e , Mathias Otto d , Beatrix Tappeser d a

Ecostrat GmbH, Friedensallee 21, 15834 Rangsdorf, Germany Research Institute of Organic Agriculture (FiBL), Ackerstrasse, 5070 Frick, Switzerland c ECT Oekotoxikologie GmbH, Böttgerstrasse 2-14, 65439 Flörsheim, Germany d Federal Agency for Nature Conservation (BfN), Konstantinstrasse 110, 53179 Bonn, Germany e University of Basel, Environmental Geosciences, Bernoullistrasse 30, 4056 Basel, Switzerland b

a r t i c l e

i n f o

Article history: Received 20 July 2012 Received in revised form 25 January 2013 Accepted 15 July 2013 Keywords: Bt toxin TC 1507 maize Bacillus thuringiensis Genetically modified plant Bt maize Environmental risk assessment Selection procedure Non-target organisms Ecotoxicological test species

a b s t r a c t For the operationalization of the structured, stepwise selection procedure for non-target testing organisms integrated into the new EFSA guidelines for environmental risk assessment of GM plants practical tools – i.e. ranking matrices – were developed. These tools – some of them are new and some are refined from older ones – were tested using the GM case crop of TC 1507 maize. The selection procedure consists of six steps. The strategy builds on identifying the important ecological functions for the particular cropping system and compiling a species lists according to their ecological functions and presence in the specific receiving environments. Subsequently, the species numbers are reduced in a systematic, stepwise fashion to a relevant and practical number of testing organisms and/or processes. Four ecological functional categories were selected: herbivory, pollination, natural enemies and soil organisms/processes. Based on these categories, the relevant species were chosen and subjected to the selection steps. Out of a total of 33 herbivores, 73 pollinators/pollen feeders, 48 natural enemies and 77 soil organisms/processes we started with in Step 1, 15 herbivores, 10 pollinators 17 natural enemy species and 9 soil organisms/processes were selected as relevant and suited for a testing program at the end of the selection procedure in Step 4. Although the ranking tools will continue to need further refinement, we could demonstrate that this procedure allows to swiftly select the most important suite of species and processes from a large number of organisms. This expert-driven process increases ecological realism and transparency in risk assessment and tailors it to the particular receiving environment, thus, overcoming important deficiencies of the current approach that has attracted persistent criticism. We recommend balancing ecological requirements with practicability criteria and realism in the test strategy. At present, the ranking is abundance-oriented and, thus, excludes rare and/or endangered species that are sensitive to disturbances. We suggest additional selection criteria to strengthen nature conservation and off-field aspects. © 2013 Elsevier Ltd. All rights reserved.

1. Background and purpose

Abbreviations: EFSA, European Food Safety Authority; ERA, environmental risk assessment; EU, European Union; GM, genetically modified; GMO, genetically modified organism; HR, herbicide resistant; IOBC, International Organization for Biological Control; IR, insect resistance; PAT, phosphinothricin-N-acetyltransferase; NTO, non-target organism; EF, ecological function. ∗ Corresponding author at: Institute of Integrative Biology, Plant Ecology, Universitätsstrasse 16 ETH Zentrum, CHN, CH-8092 Zurich, Switzerland. Tel.: +41 44 632 43 22; fax: +41 44 632 12 15. E-mail address: [email protected] (A. Hilbeck). 1470-160X/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ecolind.2013.07.016

According to the regulation in the European Union (EU), an environmental risk assessment (ERA) for the approval of GM crop plants must be carried out on a case-by-case basis following the principles and recommendations given in Annex II of the Directive 2001/18/EC. In Annex II, a case is defined as a combination of the crop plant (its biology, ecology and agronomy), the novel trait relating to its intended effect and phenotypic characteristics of the GM plant, and the receiving environment related to the intended use of the GM plants. Currently, the identification and characterization of potential adverse effects for the ERA of GM crops on non-target organisms (NTOs) is mainly derived from tests with isolated novel

368

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381

2. Methods

Case definition

Crop biology / Novel trait (intended effect) / receiving environment (intended use)

Step 1: Which functional groups are exposed?

Functional groups Step 2: Ranking of species and functions

Species selection

Part 1: Ecology

Potential species (1 ... n) Step 3: Exposure pathways

Relevant species (many) Part 2: Step 4: For which relevant species reproducible test results can be expected? Practicability Test species

(managable number) Step 5: Development of adverse effects scenarios

Methods selection Test methods

Step 6: Formulating adverse effects scenarios as testable hypotheses and recommendation of relevant experimental protocols

Practical testing Fig. 1. Selection procedure scheme for non-target organisms as published by Hilbeck et al. (2008, 2011).

proteins and testing procedures that closely follow the testing strategy developed for pesticides. This approach has been criticized (Andow and Hilbeck, 2004; Hilbeck et al., 2011) because it presumes that most potential adverse effects of the GM plant on NTOs can be extrapolated from testing of an isolated bacteria-produced novel compound. This and other criticism (e.g. Spök et al., 2004; Dolezel et al., 2011) resulted in a request of the European Commission to the European Food Safety Authority (EFSA) to review its guidelines for ERA of GM organisms with special emphasis on the further development of the risk assessment for NTOs (see EFSA, 2010a, Background). In a preceding project, an improved ERA concept for testing nontarget effects of GM plants was proposed (Hilbeck et al., 2011). This concept is based on a procedure for the selection of the most relevant testing organisms and the development of appropriate testing methods (Fig. 1). This procedure builds on the outcomes of an initiative of scientists engaged in the international ‘GMO ERA Project’ run by the global working group ‘Transgenic Organisms in IPM (Integrated Pest Management) and Biocontrol’ under the auspices of the IOBC (International Organization for Biological Control) (Hilbeck and Andow, 2004; Hilbeck et al., 2006; Andow et al., 2008). The ERA concept from this group focuses on the whole GM plant instead of the isolated toxin (i.e. transgene product) only and tailors the ERA to the case definition as laid out in Directive 2001/18/EC. By doing so, the selection procedure from Hilbeck et al. (2011) aims to identify and select testing species from the receiving environment. This selection procedure (Fig. 1) has been integrated at least in part into the revised guidelines for ERA of GMO by EFSA (EFSA, 2010a,b) and will become part of the EU regulation upon final adoption. However, the implementation of the improved ERA concept still lacks operational tools regarding how, in practical terms, the selection of organisms and methods can be done in a systematic, uniform and transparent fashion. Here, we introduce such practical tools for the proposed selection procedure and test them with the herbicide resistant Bt-maize TC 1507. The main objectives of this project were: i) to develop a set of detailed guidance and selection matrices to facilitate the ranking of the species in Steps 1 through 4 (see Fig. 1) of the selection procedure, and ii) to test these matricesbased ranking tools in the case example of Bt/HR maize in a 3-day expert workshop. In this article, we report about the outcomes of these two objectives.

In the following, we briefly describe the conceptual steps of the improved ERA concept and the tools specifically developed for their operationalization. One guidance table and one matrix are new and applied first time to this case example, two matrices have been further refined that had been developed in earlier projects. 2.1. Selection procedure An effective way to understand the role of biodiversity is through ecological functions (EFs) that the diversity of organisms execute. The use of EFs allows focussing on the identification of possible adverse effects and subsequently testing of the relevant species and critical ecological processes. Identifying important EFs, thus, helps to limit the number of organisms that must be tested to those that are ecologically relevant. Choosing a functional approach to the selection of relevant test organisms is warranted when knowledge about species is limited and incomplete, as e.g. for soil organisms. The strategy is to compile species lists according to their EFs and their presence in the specific cultivation region(s) (i.e. the receiving environments). Subsequently, the species number is reduced in a systematic, transparent, and stepwise fashion to a relevant and practical number of test organisms and/or processes (Hilbeck et al., 2008, 2011). This procedure consists of six steps (Fig. 1). The first four steps that were applied to our case examples are briefly described below. 2.1.1. Selection procedure part 1 – ecology Step 1–Concept: Identify important ecological functions for GM cropping system. EFs are identified that may be impacted by the GM crop in the given cropping system and receiving environment. The identification is based on biodiversity services such as pollination or biological control delivered by NTOs or certain ecological processes. The importance of these functions and services may vary with different crop species, recombinant traits and regions. Pollination, for instance, is of critical importance for insect-pollinated plants like oilseed rape but to a lesser degree for wind-pollinated crops like maize. However, impacts on nearby habitats like field margins or hedgerows and the interaction/influence between in field and off field living species on EF should be taken into account too. Tool for operationalization. A new guidance table was developed that allowed for the systematic selection of the most important EFs required for sustainable production in the given agricultural setting of the GM plant (Table 1). The table is structured on the basis of the elements describing the ‘Case GMO’: (i) The biology of the crop and its agronomic requirements for production. (ii) The novel trait relating to the intended effect. (iii) The receiving environments relating to the intended use. The developed guidance table is still work-in-progress and should be refined with each further round of application. If archived and used for later applications of the same crop, they can be complemented and lead to a detailed and widely re-usable tool for a specific crop species. Step 2 – Matching the ecological functions (EFs) with non-target organisms or ecological processes. For the most important EFs identified in Step 1, the species known to execute that function are assigned to these functions. Some organisms may contribute to more than one function (e.g. ladybeetles may be natural enemies and pollen feeders/pollinators). However, for many Efs, we do not know all contributing species. For example, many important EFs in soils are carried out by an unknown number of unidentified

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381

369

Table 1 Guidance table for Step 1: case-specific selection of important and possibly affected environmental functions as part of the ERA of GMOs: case study for BT/HR maize 1507. Main criteria I Crop biology Harvested product?

Symbiosis with nitrogen-fixing microbes? Sensitive growth stage?

Characteristics

Associated ecological function/agricultural practise

Affected organisms/processes

Kernels Whole corn cob (Corn Cob Mix – CCM) Entire above ground plant parts

Harvest of kernels/cobs until early November. Harvest of CCM beginning mid September. Harvest silage maize/green maize beginning mid September. Harvest with combine or chopper —

All organisms that feed on plant. Spread of transgene product to other ecosystems (incl. water bodies, forests, urban areas) during harvest

Pre-emergence application of broad spectrum herbicides. Sometimes herbicide application postermegence with selective herbicides. Corresponding broadspectrum herbicides (e.g. glyphosate) in conjunction with HR maize. Scaring of birds with various methods (visual with e.g. scare crows, chemical, acustic) Spring when cold and wet: flexible seeding date Summer drought: irrigation

Flora in and around of arable field

No; Nitrogen fertilizers used regularly Weak competitors against weeds during early plant development stages. (Bird) crow feeding after germination

Sensitive growing conditions?

Sensitive to diseases?

Coldness in early growth phases. Soil temperature <8◦ C Drought and coldness in later developmental stages Yes, highly sensitive

Reproduction type? generative, vegetative? Pollination type?

Cross pollination. Wind pollination, no volunteers, no vegetative reproduction

Production of huge amounts of wind distributed pollen

Input routes of transgenic plant parts and transgene products?

Pollen

Wind pollination, pollen collecting insects

Seed coating with pesticides

Leaves, stems and roots, kernels Degradation processes, root exudates Birds, small mammals

—

Birds

—

Non-target organisms of seed coating substances and antagonists (predators and parasites/parasitoids). If seed coating systemic: herbivores and polle feeders, pollinators, pollen collectors and antagonists Pollen feeder, pollinators, pollen collectors. Herbivores, soils organisms feeding by chance Pollen feeders, pollinators, pollen collectors. “Random” pollen consumption. Herbivores, soil organisms including antagonists. Degradation processes of spread material and excretions in agricultural areas (forests, pastures. . .) Root settling macro- and microfauna and flora Soil organisms Pollen feeders/pollen collectors and “random” pollen feeders (herbivores, soil organisms) incl. their antagonists Soil organisms

Plant excretions/exudates possibly containing transgenes, transgene products or metabolites? Plant parts/residues possibly containing transgenes, transgene products or metabolites before harvest?

Root exudates (Bt- and PATcontaining) Unclear whether in phloem or xylem All plant parts incl. pollen contain transgenes and Bt toxin in different concentrations. Pat-gene/protein produced in green plant parts. Non germinated seeds

Plant residues possibly containing transgenes, transgene products or meta bolites after harvest?

Residues of all plants parts, seeds and kernels

Degradation of plant material, maize often remains (standing) until spring time without soil cover on bare soil

Soil organisms, herbivores and their antagonists

Accumulation over time?

Yes, some amount. Residues of plant material and seeds, maize grains. Degree of degradation depends on climate, soil Depending on climate: in Germany, seeds are not expected to survive winter cold. Seeds germinate up to 10 years, but normally no persistent seed bank. In warm winters: some volunteers may occur No, annual species

Degradation of plant material, maize often remains (standing) until spring time without soil cover on bare soil

Detrivores and microorganisms – ecological soil processes

Pollen

Pollen during pollination period by wind, bees etc. During harvest or transportation loss of plant material incl. maize grains/seeds,. After harvest wind, rain, erosion, animals into adjacent terrestrial ecosystems and aquatic ecosystems (ditches, ponds, lakes, rivulets. . .)

Pollen feeders/pollen collectors, “random” pollen feeders incl. antagonists. Soil organisms/detrivores, herbivores and antagonists. Detrivores, microorganisms in soils, aquatic fauna microorganisms

Seed bank possible? If yes: short term or long term persistent?

Whole plants or plant parts survive/regenerate vegetatively? Whole plants or plant parts expected to spread or to be spread into the field margins or beyond? If yes: vectors?

Plant material, seeds/maize grains

Rhizosphere, mycorrhiza

Pollen deposition in the field and in agrarian landscape Degradation processes of plant residues incl. pollen in soil

370

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381

Table 1 (Continued) Main criteria

Characteristics

Associated ecological function/agricultural practise

Affected organisms/processes

Crop-associated valued species?

Pollen in summer important food source for honey bees in intensively farmed regions (little alternative food sources)

Honey bees often fed in addition to food they find in environment, until forest honey can be collected (Exudates from aphids on firs and spruces)

Pollen fedder/pollen collectors and their antagonists

Bt Cry 1F Bt-toxin from B. thuringiensis ssp. aizawai, in all plant parts. PATprotein (Phosphinothricin Acetyltransferase) used as marker gene, conveys resistance against broad spectrum herbicide glufosinate, produced in all green plant parts Unknown

ECB–rarely controlled in maize, occasionally in kernel maize sprayed with insecticides Occasionally weed control postemergence

Other butterflies and moths and their antagonists

Unknown

Research need

Degradation products of novel proteins in living plant. Degradation products of incroporated herbicide if applied. Degradation products of novel proteins and of herbicide (if applied) in dead plant material Pest control of: Ostrinia nubilealis (and Sesamia nonagrioides. HR primarily as marker gene–possibly weed control postemergence

Hardly known Primary metabolite of Glufosinate is Acetyl-Phosphinotricin, other metabolites in HR plants hardly known

Herbivores, their antagonists Soil organisms

Simple preventive plant protection HR primarily as marker gene–possibly weed control postemergence

Not necessarily but possibly if HR trait is used for weed management purposes No

Use of glufosinate postemergence

Other herbivores and associated food chain (antagonists) Field associated flora and their associated food chain, non-target organisms Field associated flora and associated food chain, non-target organisms, soil organisms None

II. Transgenic plant–intended effect Novel transgene product expressed? If yes, which?

Any plant meta bolites eliminated or significantly reduced? Metabolite significantly increased?

Intended effect?

Application of corresponding chemical required? If yes, which? Antibiotic resistance gene present?

III. Receiving environment–intended use? a. Region Landscape structure? Many areas, no specific requirements Climate type? Number of potential different production regions?

b. Farming system How many crop cycles? Intended/anticipated scale of release? Replacing other crops? loss, shift, addition? Expanding agricultural production zones? what degree? Cropping system? Large to small, subsistence Farming practise? Pest management type?

None

No special needs

Flora associated with arable field and their associated food chain incl. predators and parasitoids

Research need

Different ecosystems/receiving environments affected

Atlantic and continental climate Large distribution in Germany. Kernel maize in southern Germany. Otherwise silage maize (up to 750 m above sea level)

Frost protection. Irrigation Increasing production area, use of maize for agrofuels. Production without rotation when sufficient nitrogen fertilizer applied

One harvest per year

In rotation. Without rotation when sufficient nitrogen fertilizer applied Land use

Ecosystem in intensively used fields Increasing spread of production possible Ecosystem in intensively used fields Ecosystem in intensively used fields

Several pesticidal chemicals are used, seed coating, selective herbicides, insecticides. Instead of insecticides egg parasitoids are frequently used (Trichogramma sp.) or Bt sprays (Dipel, Delfin) No till production possible?

Reduction of chemical use probably meaningless: in silage maize ECB hardly a problem and typically not treated. Replaces or reduces other methodlogies such as egg parasitoid use or Bt sprays. Evolution of glufosinate-resistant weeds

Spread of transgenic material into other environmental compartments (forests, pastures, water bodies, etc.)

Different other ecosystems, erosion hazard on slopes

Production likely large scale–commodity production Possible, depends on development of use for agrofuels Possible, depends on development of use for agrofuels Extensively to intensively Intensively, feed production organic production possible Replacing insecticide treatment Possibly replacement of selective herbicides by broad spectrum herbicide based on glufosinate (or glyphosate for Monsanto’s Bt maize events)

III. Environment–where will the new product be used? c. Soil type undemanding Soil type? undemanding Organic matter content? High to low? Yes, when broad spectrum herbicide Prone for soil erosion? sprayed pre-emergence. Soil remains bare for fairly long time

Ecosystem in intensively used fields

Table 2 Matrix I for the ‘Herbivores group’. Ranking of spatio-temporal coincidence with maize (Part A), Ranking of trophic relationship to maize (Part B) and Total Ranking. Species or taxon

Part B

A+B

Geographic distribution

Habitat specialization

Abundance

Phenology: nontarget species perspective

Phenology: crop perspective

Linkage and association with the crop

Ranking of Part IA (Mean)

Potential of nontarget species to become a pest

Significance as vector (diseases)

Significance for other ecological functions

Significance as food (prey/host) for natural

Significance associated with other crops

Significance in natural and seminatural habitats

Ranking of Part IB (Mean)

Total Ranking (Mean)

1 1

1 2

3 2

1 1

1 2

3 2

1.67 1.67

3 2

3 3

3 3

1 1

1 1

1 1

2.00 1.83

1.83 1.75

1

1

1

1

1

2

1.17

2

3

3

1

1

1

1.83

1.50

1

1

1

1

2

2

1.33

2

3

3

1

1

1

1.83

1.58

1

2

2

1

2

2

1.67

2

3

3

1

1

1

1.83

1.75

1

1

1

1

1

1

1.00

2

3

3

1

3

2

2.33

1.67

1 ? 2 1

2 2 2 2

3 3 3 3

1 ? 1 1

2 ? 2 2

3 3 3 3

2.00 2.67 2.17 2.00

2 ? 2 3

3 ? 2 3

3 ? 3 2

1 ? 1 1

1 ? 2 1

2 ? 3 2

2.00

2.00

2.17 2.00

2.17 2.00

1

2

3

1

2

3

2.00

3

3

3

1

3

2

2.50

2.25

1

2

2

1

2

3

1.83

1

2

3

1

2

2

1.83

1.83

1

3

1

1

1

3

1.67

3

3

3

1

1

1

2.00

1.83

?

?

?

?

?

?

?

?

?

?

?

?

1 1 3

2 3 1

3 2 3

2 2 1

2 2 2

3 3 1

2.17 2.17 1.83

2 2 2

3 3 3

3 3 3

1 1 1

1 1 1

2 3 2

2.00 2.17 2.00

2.08 2.17 1.92

? 1

? 2

? 3

? 1

? 3

? 2

2.00

? 1

? 3

? 3

? 1

? 3

? 3

2.33

2.17

3 1 1

1 2 2

3 3 3

1 1 2

1 1 1

1 2 3

1.67 1.67 2.00

2 2 3

3 3 3

3 3 3

1 1 1

2 1 1

3 2 2

2.33 2.00 2.17

2.00 1.83 2.08

? ?

? 1

? 3

? 2

? 3

? 2

2.20

? 3

? 3

? 3

? 1

? 3

? 2

2.50

2.35

1

3

3

3

3

3

2.67

3

3

3

1

3

2

2.50

2.58

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381

Hemiptera, Aphidae Aphis fabae Metopolophium dirhodum Rhopalosiphum maidis** Rhopalosiphum padi** Sitobion avenae Hemiptera, Cicadellidae Zyginidia scutellaris** Hemiptera, Miridae Lygus spec.** Megalocoleus spec. Notostira erratica Stenodema calcaratum Stenodema laevigatum Trigonotylus caelestialium** Thysanoptera, Thripidae Frankliniella tenuicornis** Limothrips cerealium Lepidoptera, Noctuidae Agrotis segetum** Autographa gamma** Helicoverpa armigera** Diptera, Chloropidae Elachiptera cornuta Oscinella frit** Coleoptera, Chrysomelidae Diabrotica virgifera** Oulema melanopus** Phyllotreta spec. Coleoptera, Carabidae Zabrus tenebrioides Coeloptera, Elateridae** Orthoptera, Tettigoniidae Tettigonia viridissima**

Part A

371

2.00

microbial organisms (Wall et al., 2012). In that case, we chose to work with the actual processes that they execute, such as nitrogen or phosphorous cycling, or plant litter decomposition. These processes can often be measured fairly reliably and effectively (Cadisch and Giller, 1997). Listed processes are ranked by the criteria explained in more detail below and in Boxes 1 and 2. Only those species that were ranked highest were carried on to step 3. Tool for operationalization – Matrix I. The organisms assigned to the selected EFs are subjected to the ranking procedure (described in 2.1.3. below) based on ecological criteria for which matrix-style tools were developed. In Boxes 1 and 2, the selection criteria are defined that are systematically aligned in Matrix I (Table 2). Matrix I consists of two parts: Matrix IA evaluates the spatio-temporal coincidence of NTOs with the crop (Box 1) and the significance that a potential adverse effect could have on their EFs. Because the trophic relationships of the different functional groups are different, another matrix (Matrix IB) was developed for each group. In Box 2, the criteria of Matrix IB are defined that evaluate the trophic relationship of the NTOs with the crop as an estimate of the functional significance of the NTO for the production system. Matrix IA and IB will provide a preliminary estimate of the likelihood of an adverse effect to occur (Table 2). Step 3 – Exposure analysis and pathways. For the organisms or functions ranked highest in the previous Step 2, an exposure analysis is conducted to determine whether and to what extent these organisms come into contact with the GM plant and/or its transgene products (incl. their metabolites) or will be affected by measures necessary for the intended effect of the GM plant (e.g. application of a herbicide for HR crops). The goal of this step is to differentiate candidate organisms into those that are most likely to be exposed and those unlikely to be exposed and affected. Tool for operationalization – Matrix II. This analysis is casespecific to the GM plant and requires information on the phenotypic pattern of transgene expression and any induced pleiotropic changes in the various parts of the transgenic plant over the whole growing season including the application of the corresponding chemical, where appropriate. Exposure can be bitrophic via exposure to plant parts, residues and secretions that contain the transgene product. Exposure can occur through higher trophic levels via exposure to the transgene product or metabolites or corresponding chemicals in organisms that have been exposed to these (added or altered) compounds. Exposure is not restricted to the field as plant parts and transgene products are mobile (e.g., pollen, nectar, seeds or plant residue), and may be transported off-field and in different environmental compartments (e.g. water bodies). Likewise, NTOs living outside crops may actively immigrate and forage within the field during part of their life cycle. The transgene may move via gene flow (pollen, seed) to other related plants that may subsequently express the transgene and cause exposure. For a Bt toxin, which acts as a gut toxin, exposure must occur via consumption to have any effect. Transgene products or metabolites may also affect NTOs by interacting with other plant compounds or by affecting plant quality. For systematic, transparent selection a Matrix II was developed (Table 3). This matrix presents the ecological criteria for evaluation and is self-explanatory.

? ? ? ? ? ? ? ? ? ? ? ? 1.33 1.00 1 ? 1 ?

?, Gaps of knowledge ** Selected species.

3 ? 1 ?

1 ?

3 1

1

2

1

1.60

?

?

?

?

?

?

2.17 2 ? ? 2 ? ? 1 ? ? 3 ? ? 3 ? ? 2 ? ? 1.83 1.00 1.00 2 ? ? 1 ? ? 3 ? ? 3 ? ?

1 ? ?

2.33 2 2 1 3 3 3 2.17 3 1 1 3 3

Acari, Tetranychidae Tetranychus urticae** 2 Nematoda 1 Ditylenchus dispaci Heterodera avenae 1 Pratylenchus spec. 1 Microorganisms/Pathogenes Helminthosporium ? maydis 1 Puccinia polysora Rhizoctonia spec. 1

A+B

Ranking of Part IB (Mean) Significance in natural and seminatural habitats Significance associated with other crops Significance as food (prey/host) for natural Significance for other ecological functions Significance as vector (diseases)

Part B

Potential of nontarget species to become a pest Ranking of Part IA (Mean) Linkage and association with the crop Phenology: crop perspective Phenology: nontarget species perspective Abundance Habitat specialization Geographic distribution

Part A Species or taxon

Table 2 (Continued)

2.25

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381

Total Ranking (Mean)

372

2.1.2. Selection procedure part 2 – practicability Step 4 – Practicability test. As a new step, practicability criteria regarding the suitability for ecological testing are applied to the organisms ranked highest in Step 3 and carried over to Step 4. The aim is to differentiate between those organisms for which: i) laboratory test yield reproducible results, ii) may only be testable in field trials, or iii) should be included in long-term monitoring (e.g. long life-spans and generation times).

Table 3 Matrix II for the ‘Herbivores group’: Ranking of non-target organisms based on maximum likelihood of exposure. Bitrophic exposure

Background information Life cycle stage with other functions

Growth stage of maize when organism is present (early, mid, reproductive)

Plant tissues/secretions on which the nontarget organism feeds

Plant tissues or secretions fed upon expresses the transgene product

Is this feeding important for the herbivore?

Ranking of bitrophic exposure

Are transgene products/metabolites detectable after feeding on plant

Ranking of possibility of bitrophic exposure

Total Rank

Hemiptera, Aphidae Rhopalosiphum maidis Rhopalosiphum padi

1 1

None None

Early (120d) Middle (60d)

Phloem Phloem

No No

Yes Yes

3 3

3 3

3 3

3 3

Hemiptera, Cicadellidae Zyginidia scutellaris

1

None

Middle (90d)

Mesophyll

Yes

Yes

1

1

1

1

Hemiptera, Miridae Lygus spec. Trigonotylus caelestialium

1 1

None None

Middle (90d) Middle (90d)

Mesophyll Mesophyll

Yes Yes

Yes Yes

1 1

1 1

1 1

1 1

Thysanoptera, Thripidae Frankliniella tenuicornis

1

None

Early (150d)

Mesophyll

Yes

Yes

1

1

1

1

Lepidoptera, Noctuidae Agrotis segetum Autographa gamma Helicoverpa armigera

1 1 1

None None None

Early (21d) Early (21d) Middle (90d)

Roots Leaves, shoot Leaves, cobs

Yes Yes Yes

Yes Yes Yes

1 1 1

3 3 1

1 1 1

1 1 1

Diptera, Chloropidae Oscinella frit

1

None

Early (21d)

Shoot

Yes

Yes

1

3

1

1

Coleoptera, Chrysomelidae Diabrotica virgifera Coeloptera, Elateridae

1 1

None None

Early (150d) Early (150d)

Whole Plant Roots

Yes Yes

Yes Yes

1 1

1 3

1 1

1 1

Coleoptera, Chrysomelidae Oulema melanopus

1

None

Middle (60d)

Leaves, cobs

Yes

Yes

1

3

1

1

Orthoptera, Tettigoniidae Tettigonia viridissima

3

Life cy

Middle (90d)

Leaves

Yes

No

2

3

1

2

Acari, Tetranychidae Tetranychus urticae

1

None

Middle (60d)

Mesophyll

Yes

Yes

1

1

1

1

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381

Life cycle stage with herbivore function

Species or taxon

373

374

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381

Box 1 Geographic distribution The degree of overlap in the geographic distribution of the crop and the nontarget species at the country or region or agro-ecological zone scale (depending on what scale has been chosen for the analysis). 1 = everywhere in the production area of the crop; 2 = only in parts of the production area of the crop; 3 = does not occur or only in small areas of the production area of the crop; ? = unknown. Habitat specialization The degree of association between the non-target species and the crop habitat. The crop habitat is defined as the crop field and its margins and includes all of the species associated with the field and its margins, including the crop, any intercrop and weeds. A habitat specialist occurs only in the crop habitat; a habitat generalist occurs in many other habitats. If a nontarget species has more than 1 generation per year, that generation should be considered that is most closely associated with the crop. 1 = high (a habitat specialist); 2 = medium; 3 = low (a habitat generalist); ? = unknown. Abundance The average or typical density where the species is present. Assessment of abundance requires good field expertise with the sampling methods used to measure density and knowledge of the typical population fluctuations of the species. Density measures can be difficult to compare across species when different sampling methods are used. Moreover, species may be difficult to compare because of vast differences in size and biology. Field expertise is needed to compare the relative densities of such species. 1 = high abundance; 2 = medium abundance; 3 = low abundance; ? = unknown. Phenology Degree of temporal overlap of nontarget species with the crop plant. a) From the non-target species’ perspective: What proportion of the non-target species’ life cycle takes place while the crop is alive? 1 = all life cycle; 2 = half of the life cycle; 3 = parts of the larval or adult stage only; ? = unknown. b) From the crop’ perspective: What proportion of the crop growing cycle is covered by the non-target species life cycle? 1 = all of the crop’s production cycle; 2 = half to one third of the crop’s production cycle; 2 = short period of the crop’s production cycle; ? = unknown. Note: This criteria is for flower visitors of lesser importance, since almost all species only occur during the flowering phase of the crop. For soil organisms and functions, in contrast, the crop’s life cycle has to include the time necessary for degradation of the crop’s residues. For ecological functions: degree of association with relevant plant tissues, parts, residues and secretions. For soil functions, this should include association with roots, plant parts that fall onto the soil (pollen, flowers, residue), plant residue incorporated into the soil, and root exudates. 1 = close association with the crop; 2 = medium association with the crop; 3 = no association with the crop; ? = unknown. Linkage: degree of specialization on crop For species: degree of specialization to a particular food. For herbivores this would be the degree of feeding specialization to the crop (host range). 1 = feeds exclusively on crop plant (monophagous); 2 = feeds on several other host plants (oligophagous); 3 = feeds on many other host plants (polyphagous); ? = unknown. For higher trophic level species this would be the degree of feeding specialization to the prey/host associated with the crop. Linkage might also be called feeding specialization and focuses on trophic relations. It can also specify what lifestage of the non-target species feeds on the crop plant. 1 = feeds exclusively on species that are associated with crop; 2 = feeds also on species that are not associated with crop; 3 = feeds to large degree on species that are not associated with crop; ? = unknown. Note: For parasitoids, this is of lesser meaning because most are typically closely associated with their hosts. Tool for operationalization – Matrix III. By applying the new Matrix III the species list is further reduced to those species that are suitable for testing under laboratory conditions. The given criteria (see Box 3 and Table 4) favor those organisms for which

established protocols and knowledge regarding their suitability for cultivation already exists (e.g. ladybeetles and lacewings for predators). Species ranked highest in Matrix II should be selected: in case these are non-standard test species, their suitability for

Table 4 Matrix III –Practicability of laboratory testing and ranking process. No.

EG herbivores (maize)[1]

EG flower visitors (maize)

EG predators (maize)

EG soils (maize/potato)

1

Rhopalosiphum maidis (Homoptera)

Apis mellifera (Hymenoptera)

Enchytraeus cf. christenseni (Enchytraeidae)

2

Rhopalosiphum padi (Homoptera)

Bombus terrestris (Hymenoptera) Episyrphus balteatus (Syrphidae) Melanostoma mellinum (Syrphidae)

Coccinella septempunctata (Coleoptera) Propylea quatuordecimpunctata (Coleoptera) Harmonia axyridis (Coleoptera) Pterostichus melanarius (Coleoptera)

Protaphorura armata (Collembola) Lycoriella castanescens (Diptera)

Sphaerophoria rueppelli (Syrphidae) Pieris napi (Lepidoptera)

Calathus fuscipes (Coleoptera) Poecilus cupreus (Coleoptera)

Aphelenchus spp. (Nematoda) Rhabditidae (Nematoda)

Inachis io (Lepidoptera) Melanchra persicariae (Lepidoptera) Plutella xylostella (Lepidoptera) Lacanobia oleracea (Lepidoptera) Papilio machaon (Lepidoptera)

Orius minutus (Heteroptera) Aeolothrips spp. (Thysanoptera) Empis nuntia (Diptera) Coenosia intermedia (Diptera) Tachyporus nitidulus (Coleoptera) Atheta triangulum (Coleoptera) Aloconota gregaria (Coleoptera) Liogluta nitidula (Coleoptera) Philonthus cognatus (Coleoptera) Erigone atra (Araneae) Oedothorax apicatus (Araneae)

destruents incl. facultative pathogens (enzymatic activity, soil respiration) microbial community structure

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 **

**

Zyginidia scutellaris (Homoptera) Trigonotylus coelestialium** (Heteroptera) Lygus ssp. (Heteroptera) Frankliniella tenuicornis** (Thysanoptera) Oscinella frit (Diptera) Diabrotica virgifera (Coleoptera) Oulema melanopus** (Coleoptera) Elateridae ssp.** (Coleoptera) Agrotis segetum (Lepidoptera) Autographa gamma (Lepidoptera) Helicoverpa armigera (Lepidoptera) Tettigonia viridissima (Orthoptera) Tetranychus urticae** (Acari)

Selected species.

Aporrectodea caliginosa (Lumbricidae)

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381

375

Box 2 Herbivores/pathogens Potential of non-target species to become a pest Definition of, potential to become a pest: the capacity of a herbivore population to strongly increase in density and cause damaging effects that lead to economical, social or ecological consequences. Factors to be considered include: Degree of damage during different seasons? Constant or sporadic damage? Number of generations in crop (more generations – higher population densities – higher likelihood for damage) Significance as vector (diseases) Are diseases transmitted by non-target organism? Significance for other ecological functions Is non-target organism a seed distributor or destruent of residues? Significance as food (prey/host) for natural enemies Is the non-target organism a prey or host for natural enemies, including those for other crops? Significance associated with other crops Is non-target organism a pest for rotational crops or other crops associated with the GMO? Significance in natural and semi-natural habitats Is the non-target organism closely associated with natural or semi-natural habitats? Flower visitors Significance as pollinator in crop What is the potential of the non-target organism as a pollinator of the crop? Does GMO require pollinators? Significance as pollinator in other crop What is the potential of the non-target organism as pollinator in intercrops, rotational crops or neighboring crops? Significance as pollinator in natural or semi-natural habitats What is the potential of the non-target organism as pollinator of plants in natural or semi-natural habitats? Note: for many species, their potential as pollinators in other habitats is largely unknown. Many species are simply flower visitors but bad pollinators. Significance as food for natural enemies Is the non-target organism a prey or host for natural enemies, including those for other crops? Significance for other crops Is the non-target organism a pest in intercrops, rotational crops or neighboring crops that are associated with the GMO? Predators/parasitoids Significance as natural enemy in crop What is the potential of the non-target organism as a predator/parasitoid in the crop? To be considered: capacity of natural enemy to keep pest population at or below a certain threshold. Significance as natural enemy in other crops? What is the potential of the non-target organism as predator/parasitoid in intercrops, rotational crops or neighboring crops? Significance as food for other natural enemies Is the non-target organism a prey or host for other natural enemies, including those for other crops? Significance as natural enemy in natural or semi-natural habitats How closely is non-target organism as predator/parasitoid associated with natural or semi-natural habitats? Soil organisms/ecological processes Significance of association with important nutrients in crop system Does the non-target organism/process have an important role in one of the core nutrient cycling systems? Significance as indicator for soil quality Has the non-target organism/process been used as an indicator for soil quality? Significance for degradation of organic matter How significant is the non-target organism/process for the degradation of plant residues? Significance for other processes How important is the non-target organism/process for the other ecosystem processes such as bioturbation?

laboratory culturing should be determined. Table 4 lists the agreed criteria of Matrix III and describes the ranking process. In Box 3, each practicability criteria is defined. This selection step is based on four criteria complexes: (i) breeding and keeping, (ii) removal from nature (field catches), (iii) simulation/standardization of exposure pathways and (iv) sensitivity to stressors. In a first set of questions, the feasibility of establishing laboratory cultures is evaluated based on existing knowledge and expertise regarding the biology and ecology of the organism. If laboratory culture is not possible or associated with prohibitive efforts, the feasibility of using field collected organisms is determined. This includes considerations whether or not the selected organism has an assigned protection status and might be threatened in its local abundance if certain proportion of its population is removed from nature. Considerations also include understanding how difficult it might be to find and identify the organism in its natural habitat. Further, different pathways of exposure (including uptake) are evaluated to ensure realistic

testing protocols. In case the testing organism is a natural enemy of higher trophic levels both direct exposure and exposure via trior multitrophic exposure routes should be considered. The possibility of simulating exposure is evaluated. This criteria complex also includes informing about known measurable parameters and their sensitivity. Lastly, the sensitivity to a range of abiotic environmental conditions is evaluated and the sensitivity of the selected non-target organism or ecological process to a wider range of factors including pesticides and other chemicals or biotic factors is estimated. The total ranking is carried out by calculating the median of the scores of all criteria and counting the overall knowledge gaps. 2.1.3. The ranking process The ranking approach followed largely the methodology described by Hilbeck and Andow (2004) and Hilbeck et al. (2008, 2011) but was modified to improve its applicability. Parameters are ranked for each non-target organism/ecological process in relation to the other NTOs/EFs of the list. The rank scores of each

376

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381

Box 3 Keeping and breeding Easy to keep How easy is the organism to maintain in the laboratory or under semi-field conditions? Is regular and standardized food available or can be developed with ‘reasonable’ efforts? Is special taxonomic knowledge necessary? Easy to breed How easy is the organism to breed in laboratory culture or in semi-field conditions? Is mass culture possible? Are the prerequisites for mass breeding given throughout the year? Quick succession of generations How quick is the succession of generations (e.g. per year)? Removal from nature Feasibility of field collections Can organism be collected in nature in reasonable numbers and with reasonable efforts? What difficulties can occur? Identification of non-target organism/ecological process How easily can organism/process be determined in the field. Are special taxonomic guides or tools necessary? Threat of non-target organism through removal from nature Could field collections decrease population densities of the non-target organism (locally)? (See also Matrix I, Part C) Simulation/Standardization of exposure pathways Standardized food supply Describe the food supply of the non-target organism/ecological process and potential difficulties Possibility of simulation of bitrophic exposure via food source Simmulation of a bitrophic exposure pathway Can a direct – bitrophic – exposure of the non-target organism/ecological process to the test stresssor (substance/food/organism) be simulated with the chosen food supply? Simmulation of a tri- or multitrophic exposure pathway Can tri- or multitrophic exposure of the non-target organism/ecological process to the test stressor (substance/food/organism) be simulated with the chosen food supply? Measured parameter Can different parameters be measure during one test run (e.g. LD50, sublethal effects, reproductive effects)? Does knowledge exist regarding the suitability of parameters for testing? Experience/reports Does expertise exist in the science community regarding behavior of the organism in testing conditions? To what degree is expertise documented? Ecological sensitivity Low sensitivity of non-target organism/ecological process to environmental fluctuations? How does organisms react to changes in environmental conditions (e.g. soil characteristics, temperature, humidity, etc.)? Can the organism be tested in different soils (if yes, in which)? Moderate sensitivity of non-target organism/ecological process to a wide range of stressors? How sensitive is non-target organism/ecological process toward different stressors such as transgene products and metabolites, pesticides, plant residues? A broad sensitivity spectrum is more important than a high sensitivity to individual stress factors.

parameter are aggregated to ranks for the different criteria complexes, which then are added up to a final overall rank for the NTO/EF. This determines the transfer of the NTO/EF to the next step of the process. Finally, the results are subjected to a plausibility test, whose findings have to be documented. In case of uncertainty, expert knowledge can override the outcome of the calculated total rank. However, such decisions have to be documented and justified. Ranking of parameters. All information for each parameter is evaluated according to a three-level ranking process: 1 = high; 2 = intermediate; 3 = low. It is further important to identify knowledge gaps and document them in the selection matrix by designating a ‘?’ (=unknown or insufficient knowledge) to a rank. Knowledge gaps can be addressed using a precautionary methodology (worst case scenario) by assigning the highest rank ‘1’ to a ‘?’ and use this rank for calculation in the selection matrix. This will ensure that organisms with large knowledge gaps will not get lost because of that but will come up high in the ranking process and require additional, ‘best qualitative’ judgment regarding their importance either by the present experts or require additional expertise to be brought in. The idea is that the involved experts eventually agree on each rank. However, this may not always be the case depending on the nature of the knowledge gaps, on the expertise of the involved scientists or the overall disagreement among experts. For example, in the predator group, consensus was often and very quickly achieved regarding ranks for Matrix I, e.g. on geographical distribution, phenology or abundance. Agreement turned

out to be more difficult in ranking bi- and tri-trophic exposure of predators (Matrix II). A conservative approach was chosen here, i.e. in case of general uncertainty or differing judgments among experts the higher value was chosen in order to comply with a precautionary approach. Ranking criteria complexes. The aggregated rank for each criteria complex is calculated by summing the individual ranks of each parameter. If there is only descriptive information for a parameter, the experts “translate” this information for a NTO/EF into a rank relative to the others in the list. The experts have to decide how large the proportion of uncertainty/knowledge gaps (‘?’) can be before a parameter is excluded from the selection process. If large knowledge gaps exist and the threshold is too high, such NTOs/EFs will get a high rank (=low score) and many of them may occupy the top ranks of the list. If the threshold is too low, too many potentially appropriate test organisms are excluded. In other test runs of this selection procedure, a 50% proportion of ‘?’ was considered as practicable threshold value. The experts can weigh and adjust any individual parameters as long as every decision is justified and documented. Every expert group determines the threshold rank for the NTO/EF above which it will move to the next step of the selection process. In case of doubt, experts can change the status of a NTO/EF based on their collective knowledge. If at a later time, more knowledge becomes available such decisions can be revised.

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381

Reference system/framework for the ranking procedure. Ranking is done relative to that of the other NTOs/EFs in the list. This means that there is no fixed threshold value for retaining or discarding species, but the applied and exact threshold value will depend on the distribution of the other values. However, sometimes experts may decide to keep a NTO despite a low rank to allow, for example, for more diversity of the spectrum of tested organisms. All decisions and adjustments must be justified and documented for transparency reasons and to allow revisions at any later point of time. 2.2. The GM crop case example – TC 1507 Bt/HR maize TC 1507 maize contains two novel transgenes, cry1F and pat, for insect and herbicide resistance respectively. Both transgenes were introduced into the parental hybrid maize by particle bombardment. The cry1F transgene stems from Bacillus thuringiensis (Bt) var. aizawai producing the insecticidal Bt protein Cry1F. Cry proteins, of which Cry1F is only one, act by binding to specific sites located on the inner lining of the mid gut of susceptible insect species. Following binding, pores are formed that disrupt mid gut ion flow, causing gut paralysis and eventual death due to bacterial sepsis (Vachon et al., 2012). The artificial Cry1F protein expressed in TC 1507 maize aims to protect against Lepidopteran pest species like: european corn borer (Ostrinia nubilalis), fall armyworm (Spodoptera frugiperda), black cutworm (Agrotis ipsilon), and to some extend to corn earworm (Helicoverpa zea) (Dow AgroSciences, 2007). In addition to the cry1F transgene, the pat transgene was introduced into TC 1507 maize. The pat gene encodes the enzyme phosphinothricin-N-acetyltransferase (PAT) based upon the sequence information from the common aerobic soil actinomycete, Streptomyces viridochromogenes (US EPA, 2005). This allows for the use of glufosinate ammonium, the active ingredient in broad spectrum phosphinothricin herbicides, as a weed control option, and as a breeding tool during the selection process. Glufosinate chemically resembles the amino acid glutamate and inhibits an enzyme, called glutamine synthetase, which is involved in the synthesis of glutamine. The glutamine synthetase is also involved in ammonia detoxification. The action of glufosinate results in an increase in concentrations of ammonia in plant tissues, leading to the cessation of photosynthesis resulting in plant withering and death. The PAT enzyme catalyzes the acetylation of phosphinothricin, detoxifying it into an inactive compound. 2.3. Application of developed tools in expert workshop The expert workshop (held in German language) involved an interdisciplinary team of 20 scientists (from academia or public agencies) mainly from Germany but also German-speaking neighboring countries who had either expertise in risk assessment with GM plants or/and were experts in fields ranging from agro-ecology taxonomy, ecology, agricultural engineering and pest management. In the workshop, we applied a number of matrices developed to assist the selection of testing organisms based on the above described conceptual procedure (Fig. 1) to the case example. Two new tools were introduced: the table guiding the identification of the most important ecological functions serving as basic categories for species assignment (Step 1) and a third matrix aiming at selecting those testing organisms that are most suited for laboratory testing or for field testing only. Matrices I and II were further refined. As receiving environment the continental biogeographical region of Europe was chosen according to the concept of specification of biogeographical regions developed in parallel (Jänsch et al., 2011). The working material for the workshop included initial lists of organisms for each functional category, informational material on the GM case crops and the set of matrices for each category based

377

on specified criteria. These matrices will be described in more detail below. 3. Results 3.1. Part 1 – ecology 3.1.1. Outcomes Step 1 – selection of ecological functions through application of guidance table The evaluation of the criteria underlying the selection of the ecological functions are presented in Table 1 and presented for the Bt/HR maize case example. Four functional categories were identified and selected for inclusion into the ERA: Functional category Herbivory (herbivores/pathogens) Pollination (pollen collectors/pollen feeders) Natural enemies (predators/parasitoids) Soils (organisms/ecological processes)

Expert group name ‘Herbivores group’ ‘Flower visitors group’ ‘Predators group’ ‘Soil organisms group’

In previous workshops with other GM case crops, it was found that the above listed ecological functional categories are the four main functions that almost always will turn out to be among the most important ones (see also Hilbeck and Andow, 2004; Hilbeck et al., 2006; Andow et al., 2008). Nevertheless, using the guidance table to identify functional categories is important not only for selecting the most important ecological functional groups but also for prioritization of the relative importance of these groups. For instance, for maize early competition with weeds is a critical limiting factor. For other crops, like for example potatoes, fungal diseases are more important during early stages than for maize. Hence, any novel trait must not unintentionally interfere adversely with these limiting factors for the production of the respective crop. In addition, the novel traits should not adversely impact the diversity and functions of the organism communities of the field margins relevant for both in-field and off-field processes, and data for regulatory ERA ought to demonstrate that. It also informs biosafety testing programs of stacked Bt/HR crop plants to focus on adverse effects of the novel compound in conjunction with the application of the corresponding broad spectrum herbicide, or any other required chemical. For other GM plants, the focus may lay on other aspects leading to a different suite of testing organisms from the functional categories than for Bt maize. Hence, the guidance table serves as a fundamental scoping and structuring exercise and constitutes the entry point for the ERA. 3.1.2. Outcomes Step 2 – species lists and selection of testing organisms based on ecological significance through application of Matrix I The workshop participants of the ‘Herbivores group’ subjected a list of 33 non-target herbivore species and pathogens to this selection step (Table 2). Of these, 15 herbivores and 3 pathogens were selected for the next step of the process. This represented a reduction of over 50% of the original list of potential testing organisms of this functional category at Step 2. The workshop participants of the ‘Flower visitors group’ subjected a total of 73 species to this selection step. Of these, 51 species were selected for the next step of the process. This was a cut of 30%. For the ‘Predators group’, from an in-going total of 47 species, 22 species were passed on to the next Step 3. This was a reduction of 53%. In the workshop, some relevant taxa such as Opiliones and Acari could not be considered due to lack of experts. The ‘Soil organisms group’ subjected a list of 77 organisms and processes to the selection procedure. Of these 77 organisms and processes, 16 were carried over to Step 3 of the selection procedure. Aside of the numerical ranking as laid out in the matrices, additional criteria were applied by the group for their final selection at this step. These criteria included considerations of the trophic

378

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381

level, taxonomic group diversity, type of exposure to ensure a certain diversity of organisms and processes that cover a broad range of trophic levels, taxonomic diversity, and exposure routes. The ‘Soil organisms group’ noted further the existence of large knowledgegaps not only due to missing experts around the table but also due to the current limited state of knowledge about soil organisms and processes in general. In addition, it was realized that those organisms living on or near the soil surface (mainly saprophagous macro-arthropods such as isopods) were not covered in this group nor in one of the others. 3.1.3. Outcomes Step 3 – selection of testing species based on maximum likelihood of exposure through application of Matrix II For the ‘Herbivores group’, all 15 non-target insect herbivores from Matrix I/Step 2 ended up high again as all of them proved to be potentially serious pest herbivores that could become secondary pests on 1507 Bt maize (Table 3). This would constitute an adverse effect that requires testing within a risk assessment scheme. The Bt toxin is expressed in all plant parts throughout the entire growing season, so all herbivorous insects are likely to be exposed throughout their life span feeding on maize. Hence, they were retained as potential testing species in a pre-release testing program for ERA of GM maize. For the ‘Flower visitors group’, of the 51 species from Matrix II, 24 species were selected for the continued selection procedure. These species included two bee species, four beetles, six hoverfly species, five butterfly and seven moth species. For the ‘Predators group’, of the 22 species retained from Matrix I, one species, Trichogramma brassicae, was discarded as being an egg parasitoid and unlikely to be exposed to the toxin within the host egg. The group noted the necessity to differentiate between adult and larval stages of the various species and pointed to the problem, that for predators a clear allocation to a bi- or tritrophic exposure pathway in some cases is not practical. However, the expert group re-introduced another six species from Matrix I, despite of their lower rank. The group agreed to do this in order to allow for a more diverse and comprehensive range of taxa, ecological relevance, body size distribution, prey spectrum as well as habitats and strata. The outcome resulted in 27 species that were referred to Step 4 (Matrix III). They included 13 beetle species (three coccinellids and five carabids, five staphylinids), two true bug species, two hoverfly species, two fly species, one thrips, three hymenopteran parasitoids and four spider species. For the ‘Soil organisms group’, all 16 organisms/processes selected in Matrix I move on to the last selection step using Matrix III (Step 4). This was because exposure was always given since the food source, GM plant residues, remains the main source of new plant material in the field. In addition, when repeated cultivation of the same GM crop occurs in the same field, there may also be overlap with remaining residues from the previous season(s) depending on how fast the plant residues decay. GM maize plant residues and its transgene products have been shown to last well into another production season in temperate regions (Zwahlen et al., 2003). 3.2. Part 2 – practicability 3.2.1. Outcomes Step 4 – selection of testing species based on practicability criteria through exploring applicability of Matrix III For the ‘Herbivores group’, a total of 15 testing species from eight taxonomic orders were selected including two closely related aphid species (taxonomically difficult to distinguish), two true bug species, one thrips species, one fly species, three beetle species, three butterfly species, one grasshopper species and one mite species. This functional group ended up being taxonomically the most diverse list of proposed testing species despite the fact that it was focusing exclusively on non-target but potentially serious

secondary pest herbivores and, thus, only a segment of all herbivores that occur in a field. The ‘Herbivores group’ selected, for example, the spider mite Tetranychus urticae as one of the final testing species. Reasons for this were that spider mites are of medium ecological significance as damaging pests of maize (Matrix I, Table 2) but are known to be greatly exposed to Bt-containing tissue and to ingest Bt toxins to fairly high concentrations (Obrist et al., 2006; see also Matrix II, Table 3). However, T. urticae is assumed not to be affected by the Bt toxin. Hence, this species should be tested whether or not it thrives better on Bt maize than on its non-transgenic parent cultivar and, thus, could become a secondary pest problem on Bt maize. For example, an unwanted pleiotropic effect of the transformation may render the Bt maize cultivar a better food for spider mites because of a slightly altered primary and/or secondary metabolite composition. Evidence for such effects on Bt-maize exists in the literature (Rovenska et al., 2005). This, in turn, may lead to increased spider mite infestation on GM-maize. The possible requirement of additional pesticide applications constitutes an adverse effect that would also counterbalance the initial aim of reducing external pesticide use with the GM plant. A testable research hypothesis would be that spider mite densities on Bt maize are higher than on the isogenic parent cultivar. As T. urticae is easily culturable and has quick generation turnover times, it is an ideal organism for laboratory testing even if it exhibits low sensitivity to fluctuating environmental conditions (Matrix III, Tables 4 and 5). For the ‘Flower visitors group’, a total of 10 testing species from three different taxonomic orders were selected. These consisted of two bee species, three hoverfly species and five butterfly species. Among these, some organisms turned out to be already used commonly as test organism, although no standardized procedures exist yet: Apis mellifera, Bombus terrestris, Episyrphus balteatus, Pieris napi and Inachis io. As a new test species Papilio machaon resulted from the Matrix III selection exercise. According to experts, Papilio machaon is easy to keep and breed in the lab. For example, the peacock butterfly Inachis io was selected because of its common and wide-distributed occurrence in agroecosystems. Larvae of this species feed on nettle plants (Urtica dioica) which often grown along field margins (Gathmann et al., 2006). Larvae of I. io are susceptible to Bt toxins to which they can be exposed via Bt-pollen dusted host plants that occur in and around Bt-maize fields during tasseling (Felke et al., 2010, Lang and Otto, 2010). The larvae can be recorded and monitored fairly easily in the field (Lang et al., 2011), and can be bred in laboratory culture (Felke, 2003). The existing knowledge on ecology and behavior contributed to the value of the species as an indicator species (e.g. Ebert and Rennwald, 1991), and a recently published model predicted adverse effects of Bt maize cultivation on I. io (Holst et al., 2013) For the ‘Predators group’, strictly applying Matrix III resulted in a total of 17 species from five different taxonomic orders for testing. These consisted of 11 beetle species (three ladybird species, three carabid species, five staphylinid species), one true bug species, two hoverfly species, one thrips species, and two spider species. All were considered important and possible to test. However, the conclusion of the ‘Predators group’ was that strictly applying Matrix III will often result in the common arthropods already used for ecotoxicological testing of pesticides. These test organisms can easily be bred in the laboratory and have been recognized as ecologically relevant in the previous selection steps. However, in order to expand the scope of ecologically relevant species, the feasibility of establishing laboratory cultures of the highest-ranking species should also be determined in cases where knowledge gaps exist. Thus, the group suggested developing a modified selection procedure, e.g. by using a decision-tree approach. Such a decision-tree should take into account the functional categories of Matrix II, the type of

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381

379

Table 5 Final list of selected, suggested testing species from all four functional groups based on consensual expert group judgement of practical aspects. No.

EG herbivores (maize)[1]

EG flower visitors (maize)

EG predators (maize)

EG soils (maize/potato)

1

Rhopalosiphum maidis (Homoptera)

Apis mellifera (Hymenoptera)

Enchytraeus cf. christenseni (Enchytraeidae)

2

Rhopalosiphum padi (Homoptera)

Bombus terrestris (Hymenoptera)

3 4

Zyginidia scutellaris (Homoptera) Trigonotylus coelestialium (Heteroptera) Lygus ssp. (Heteroptera) Frankliniella tenuicornis (Thysanoptera) Oscinella frit (Diptera) Diabrotica virgifera (Coleoptera) Oulema melanopus (Coleoptera) Elateridae ssp. (Coleoptera) Agrotis segetum (Lepidoptera) Autographa gamma (Lepidoptera) Helicoverpa armigera (Lepidoptera) Tettigonia viridissima (Orthoptera) Tetranychus urticae (Acari)

Episyrphus balteatus (Syrphidae) Melanostoma mellinum (Syrphidae)

Coccinella septempunctata (Coleoptera) Propylea quatuordecimpunctata (Coleoptera) Harmonia axyridis (Coleoptera) Pterostichus melanarius (Coleoptera)

Sphaerophoria rueppelli (Syrphidae) Pieris napi (Lepidoptera)

Calathus fuscipes (Coleoptera) Poecilus cupreus (Coleoptera)

Aphelenchus spp. (Nematoda) Rhabditidae (Nematoda)

Inachis io (Lepidoptera) Melanchra persicariae (Lepidoptera) Plutella xylostella (Lepidoptera) Lacanob ia oleracea (Lepidoptera) Papilio machaon (Lepidoptera)

Orius minutus (Heteroptera) Aeolothrips spp. (Thysanoptera) Empis nuntia (Diptera) Coenosia intermedia (Diptera) Tachyporus nitidulus (Coleoptera) Atheta triangulum (Coleoptera) Aloconota gregaria (Coleoptera) Liogluta nitidula (Coleoptera) Philonthus cognatus (Coleoptera) Erigone atra (Araneae) Oedothorax apicatus (Araneae)

destruents incl. facultative pathogens (enzymatic activity, soil respiration) microbial community structure

5 6 7 8 9 10 11 12 13 14 15 16 17

experiments (laboratory, mesocosm, semi-field), the assessment of ecological significance and the existence of knowledge gaps. Due to time restraints, this approach was not developed further. For the ‘Soil organisms group’, 9 organisms/processes were determined to be subjected to a testing program. They consisted of 6 taxa from 6 different families (two species of oligochaetes, one springtail species of Onychiuridae, one dipteran fly species of Sciaridae, one genus of the nematod family of Aphelenchoididae and the nematod family of Rhabditidae) and 3 microbial processes. With the exception of microbial processes, none of these are commonly used standard test organisms.

4. Discussion 4.1. The selection procedure The matrix-guided ranking exercise for selecting testing organisms is a proposal how to operationalize the selection procedure now integrated into EFSA’s revised guidelines for ERA in a transparent fashion based on the best available science and knowledge. This lends legitimacy and accountability to ERA right from the start. However, as with all (semi-) quantitative and structured approaches in a field where much of the information rests with experts, the outcome of the procedure can be highly variable and incomplete even for a relativley simple, human-driven, and well described and researched ecosystem–the agro-ecosystem. The outcome can easily feign a level of accuracy that clearly does not exist and should be understood as a comparative approach in such that the relative significance or importance of an organism is judged in comparison to that of the other organisms. Despite these deficits, we argue that the described procedure is the best alternative to avoid to simply adopt the ecotoxicological testing scheme for pesticides that lacks a sufficient ecological focus and transparency. The interdisciplinary approach used was considered a quality and very valuable as it allowed to involve many different types of expertise leading to broader acceptance and legitimacy. The three different matrix types proved to be of varying importance for the functional groups. While for the ‘Herbivores group’, the ‘Predators group’ and the ‘Soil organisms group’ the biggest reduction in the number of specimes happened in Matrix I, for the ‘Flower visitors group’ this was in Matrix II. With regard to the newly developed and first time applied practicability criteria

Aporrectodea caliginosa (Lumbricidae) Protaphorura armata (Collembola) Lycoriella castanescens (Diptera)

in Matrix III, the tendency of too much bias on standard laboratory testing species arose since the knowledge basis for these species is always much better than for any other species. Because this provides no information whether or not another species could be just as easily kept in laboratory culture if tried, we argue to apply practicability criteria with care and a clear focus on the desired goal to select testing organism that have the highest-ranked ecological relevance and exposure in the potential receiving environments. To this end, we developed and first time applied here a table guiding that process. While in all matrices, large numbers of knowledge gaps were identified, they tended to become an impediment to the procedure predominantly in those expert groups that had to deal with different multi-trophic relationships, i.e. the ‘Predators group’ and the ‘Soil organisms group’. However, formalizing the lack of knowledge and devising a mechanism for how to deal with it, helped with this impediment in an efficient and transparent manner. Clearly, highlighting lack of knowledge and ranking it ‘high’ by default is critical in introducing awareness for both the degree and the severity of these knowledge gaps. To our understanding this presents a mechanism to implement precaution and consciously and transparently declaring a cut-off threshold beyond which these species could no longer be dealt with in the analysis. This in turn may help, inspire and rationalize future research projects that may aim to close some of the identified, most critical gaps. Although the procedure allowed us to reduce a large number of species to a much smaller number during this workshop, the outcomes are not final and some issues may need further work and review. The data gathered in the first steps of the selection procedure (Step 1–3) could be transformed and archived into a data base, so it can become available for future runs and applications of this procedure. Much of the information compiled and synthesized at the early steps are not specific to the introduced transgenes or transgenic events and apply to other GM applications of the same plant species although not, of course, in its entirety. If also adopted, the tools for operationalization will greatly help to harmonize the implementation of this new selection procedure incorporated in the revised guidelines for environmental risk assessment by EFSA (EFSA, 2010a,b) 4.2. Uptake of the selection procedure in research and current EU regulation In 2010, EFSA included parts of the selection procedure in their revised guidelines for risk assessment of non-target organisms

380

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381

(EFSA, 2010a, b). Meanwhile, we have seen parts of this selection procedure being also put into practice and developed further by other working groups. Malone et al. (2010) used parts of the concept in conjunction with others to identify and prioritize non-target species that could be used as experimental subjects for examining risk hypotheses relevant for ERA of transgenic lignin-altered trees in New Zealand. They arrived at total of 10 species from one functional group, herbivores, that the authors suggest to be subjected to testing for impacts of these GM trees (Malone et al., 2010). Van Wyk et al. (2007) applied the selection procedure as developed by Hilbeck et al. (2006) to Bt maize in South Africa. Using a checklist compiled by Van Wyk et al. (2007), non-target species again for the ecological function ‘herbivory’ were selected based on the suggested ecological criteria of our methodology. The authors in this case proposed a list of 8 lepidopteran species for risk assessment in particular in connection with the evolution of resistance against the Bt toxin expressed in the Bt maize (Van Wyk et al., 2007). Further, they used the ‘whole plant method’ suggested by Birch et al. (2004) to develop protocols for evaluating the effects of Bt-maize in South Africa on the selected pest Agrotis segetum (Erasmus et al., 2010). In conjunction with the past case examples (Hilbeck et al., 2004, 2006; Andow et al., 2008) and the one reported about here, we know of a total of 6 applications of the selection procedure and one on-going involving a disease resistant GM potato in Norway (Gillund et al., 2011).

4.3. Outlook – nature conservation aspects and off-field habitats The abundance-oriented approach of the presented selection procedure assigns rare and endangered species, a low rank. Hence, these species are under-represented in the current lists although they may be highly sensitive to disturbances. This was realized during the evaluation phase of the project. As a means to stronger include species of conservation concern, we propose to develop an additional matrix based on the following three selection criteria: (i) the protection status by law at the national, European or international level (e.g. Council Directive 92/43/EEC of 21 May 1992 on the conservation of natural habitats and of wild fauna and flora, Council Directive on the conservation of wild birds 79/409/EEC, Council Regulation 338/97 of 9 December 1996 on the protection of species of wild fauna and flora by regulating trade, Germany’s Federal Ordinance on the Conservation of Species of 16 February 2005 amended on 29 July 2009), (ii) the Red List status on regional, national (Binot-Hafke et al., 2012; Ludwig and Schnittler, 2009; Haupt et al., 2009), European (IUCN, 2010a) or international level (IUCN, 2010b) and (iii) the national responsibility for conservation of the species (Gruttke & Ludwig, 2004; Schmeller et al., 2008; Schuldt & Assmann, 2010). Similar proposals were made for the assessment of the environmental impacts of invasive species (EFSA, 2011). The presented selection procedure focuses on species living in the GM crop field and its field margins. However, species living beyond the fields may also be exposed to GMOs or their transgene products. The workshop participants identified the need to integrate off-field species as testing organisms. This may be achieved by complementing the main selection procedure with a selection tool for off-field species. We propose to analyze the results of the main selection procedure in order to identify relevant exposure pathways and important taxonomic groups/ecological processes that are not yet included. In addition, the most frequent natural and semi-natural habitats in the surroundings of GMO fields and their characteristic species composition need to be identified. On this basis taxonomic groups could be included that are underrepresented in the main selection procedure.

Acknowledgements This work was funded by a grant from the German Agency for Nature Conservation/Ministry of the Environment with grant funding number 3507 89 070. We wish to thank all participants of the workshops carried out at BfN in December 2008 and February 2010. We also thank 2 anonymous reviewers for their very valuable comments.

References Andow, D.A., Hilbeck, A., 2004. Science-based risk assessment for non-target effects of transgenic crops. Bioscience 54, 637–649. Andow, D.A., Hilbeck, A., Van Tuat, N. (Eds.), 2008. In: Challenges and Opportunities with Bt Cotton in Vietnam, vol. 4. CABI Publishing, Wallingford, UK. Binot-Hafke, M., Balzer, S., Becker, N., Gruttke, H., Haupt, H., Hofbauer, N., Ludwig, G., Matzke-Hajek, G., Strauch, M. (Eds.), 2012. Rote Liste gefährdeter Tiere, Pflanzen und Pilze Deutschlands. Band 3: Wirbellose Tiere (Teil 1). Münster (Landwirtschaftsverlag) (Naturschutz und Biologische Vielfalt 70 (3): 327-370). Birch, A.N.E., Wheatley, R., Anyango, B., Arpaia, S., Capalbo, D., Getu Degaga, E., Fontes, E., Kalama, P., Lelmen, E., Lövei, G., Melo, I.S., Muyekho, F., Ngi-Song, A., Ochiendo, D., Ogwang, J., Pitelli, R., Sétamou, M., Sithanantham, S., Smith, J., Van Von, N., Songa, J., Sujii, E., Quang Tan, T., Wan, F.H., Hilbeck, A., 2004. Biodiversity and non-target impacts: case study of Bt-maize in Kenya. In: Hilbeck, A., Andow, D.A. (Eds.), Environmental Risk Assessment of Genetically Modified Organisms: vol. 1. A case study of Bt maize in Kenya. CAB International, Wallingford, UK. Cadisch, G., Giller, K.E., 1997. Driven by Nature. In: Plant Litter Quality and Decomposition. CAB International, Wallingford, UK, pp. 409. Council Directive 92/43/EEC of 21 May 1992 on the conservation of natural habitats and of wild fauna and flora. Council Directive on the conservation of wild birds 79/409/EEC. Council Regulation (EC) 338/97 of 9 December 1996 on the protection of species of wild fauna and flora by regulating trade. Dow AgroSciences, 2007. TC 1507 insect-protected maize (corn). Technical Bulletin. Dolezel, M., Miklau, M., Hilbeck, A., Otto, M., Eckerstorfer, M., Heisenberger, A., Tappeser, B., Gaugitsch, H., 2011. Scrutinizing the current practice of the Environmental Risk Assessment of GM maize applications for cultivation in the EU. Environ. Sci. Eur. 23, 33. Ebert, G., Rennwald, E. (Eds.), 1991. Die Schmetterlinge Baden-Württembergs. Tagfalter Band 1–2. Ulmer Verlag, Stuttgart. EC, 2001. Directive 2001/18/EC of the European Parliament and of the Council of 12 March 2001 on the deliberate release into the environment of genetically modified organisms and repealing Council Directive 90/220/EEC – Commission Declaration. OJ L 106, 17.04.01., pp. 1–39. EFSA (European Food Safety Agency), 2010a. Guidance on the environmental risk assessment of genetically modified plants (GMO). EFSA J. 8 (11), 1879. EFSA (European Food Safety Agency), 2010b. Scientific Opinion on the assessment of potential impacts of genetically modified plants on non-target organisms. EFSA J. 8 (11), 1877. EFSA (European Food Safety Agency), 2011. Guidance on the environmental risk assessment of plant pests. EFSA J. 9 (12), 2460, http://dx.doi.org/10. 2903/j.efsa.2011.2460. EPA (US Environmental Protection Agency) – OPP (Office of Pesticide Program), 2005. BRAD (Biopesticide Registration Action Document). In: Bacillus thuringiensis Cry1F corn. Erasmus, A., van Rensburg, J.B.J., van den Berg, J., 2010. Effects of Bt maize on Agrotis segetum (Lepidoptera: Noctuidae): a pest of maize seedlings. Environ. Entomol. 39, 702–706. Felke, M., 2003. Massenzucht von Aglais urticae (LINNAEUS, 1758) und Inachis io (LINNEAEUS, 1758) (Lepidoptera, Nymphalidae) unter freilandnahen Bedingungen. Entomologische Zeitschrift, 211–213. Felke, M., Langenbruch, G.A., Feiertag, S., Kassa, A., 2010. Effect of Bt-176 maize pollen on first instar larvae of the Peacock butterfly (Inachis io) (Lepidoptera; Nymphalidae). Environ. Biosaf. Res. 9, 5–12. Gathmann, A., Wirooks, L., Eckert, J., Schuphan, I., 2006. Spatial distribution of Aglais urticae (L.) in an agricultural landscape: implications for Bt maize risk assessment and post-market monitoring. Environ. Biosaf. Res. 5, 27–36. Germany’s Federal Ordinance on the Conservation of Species of 16 February 2005 amended on 29 July 2009. Gillund, F., Hilbeck, A., Wikmark, O.G., 2011. Genetically Modifi ed Potato with Increased Resistance to P. infestans – Selecting Test Species for Environmental Impact Assessment on Non-Target Organisms (Biosafety Report 2011/05) http://www.genok.com/news cms/2012/february/new-biosafety-reportgenetically-modifi-ed-potato-with-increased-resistance-to-p-infestansselecting-test-species-for-environmental-impact-assessment-on-non-targetorganisms/152 Gruttke, H., Ludwig, G., 2004. Konzept zur Ermittlung der Verantwortlichkeit für die weltweite Erhaltung von Arten mit Vorkommen in Mitteleuropa: Neuerungen, Präzisierungen und Anwendungen. Natur und Landschaft 79, 271–275. Haupt, H., Ludwig, G., Gruttke, H., Binot-Hafke, M., Otto, C., Pauly, A. (Eds.), 2009. Rote Liste gefährdeter Tiere, Pflanzen und Pilze Deutschlands. Band 1:

A. Hilbeck et al. / Ecological Indicators 36 (2014) 367–381 Wirbeltiere. Bundesamt für Naturschutz. Bonn-Bad Godesberg. Naturschutz und Biol., Vielfalt 70 (1), 1–386. Hilbeck, A., Andow, D.A. (Eds.), 2004. Environmental Risk Assessment of Genetically Modified Organisms, vol. 1: A Case Study of Bt Maize in Kenia. CABI Publishing, Oxon, UK. Hilbeck, A., Nelson, K., Andow, D., Underwood, E., 2004. Problem Formulation and Options Assessment (PFOA) to Assess the Ecological Risks Associated with a GM Crop. In: Breckling, B., Verhoeven, R. (Eds.), Risk, Hazard, Damage – Specification of Criteria to Assess Environmental Impact of Genetically Modified Organisms. Naturschutz und Vielfalt 1, Bonn, pp. 131–143. Hilbeck, A., Andow, D.A., Arpaia, S., Birch, A.N.E., Fontes, E.M.G., Lövei, G., Sujii, E.R., Wheatley, R.E., Underwood, E., 2006. Methodology to support non-target and biodiversity risk assessment. In: Hilbeck, A, Andow, D.A., Fontes, E.M.G. (Eds.), Environmental Risk Assessment of Genetically Modified Organisms: vol. 2. Methodologies for Assessing Bt Cotton in Brazil. CAB International, Wallingford, UK. Hilbeck, A., Andow, D.A., Arpaia, S., Birch, A.N.E., Chen, Y., Fontes, E., Fitt, G., Lang, A., Thi Thu Hong Le, G., Lövei, B., Manachini, Thi Thu Cuc Nguyen, Van Huynh Nguyen, Van Tuat Nguyen, Van Lam Pham, Van Toan Pham, C., Pires, E., Sujii, Khuong Lai Trac, E., Underwood, R., Wheatley, L., Wilson, C., Zwahlen, 2008. Non-target and Biological Diversity Risk Assessment. In: Andow, D.A., Hilbeck, A, van Tuat, N. (Eds.), Environmental Risk Assessment of Genetically Modified Organisms: vol. 4. Challenges and opportunities with Bt cotton in Vietnam. CAB International, Wallingford, UK. Hilbeck, A., Meier, M., Römbke, J., Jänsch, S., Teichmann, H., Tappeser, B., 2011. Environmental risk assessment of genetically modified plants – concepts and controversies. Environ. Sci. Eur. 23, 13, http://dx.doi.org/10.1186/2190-4715-23-13. IUCN, 2010a. European Red List, http://ec.europa.eu/environment/nature/ conservation/species/redlist IUCN, 2010b. IUCN Red List of Threatened Species. Version 2010.1, www.iucnredlist.org Jänsch, S., Römbke, J., Hilbeck, A., Weiss, G., Teichmann, H., Tappeser, B., 2011. Assessing the potential risk of transgenic plants for non-target invertebrates in Europe: a review of classification approaches of the receiving environment. BioRisk 6, 19–40, http://dx.doi.org/10.3897/biorisk.6.1334. Lang, A., Otto, M., 2010. A synthesis of laboratory and field studies on the effects of transgenic Bacillus thuringiensis (Bt) maize on non-target Lepidoptera. Entomol. Exp. Appl. 135, 121–134.

381

Lang, A., Dolek, M., Theißen, B., Zapp, A., 2011. Are adult Crambid Snout Moths (Crambinae) and larval stages of Lepidoptera suitable tools for an environmental monitoring of transgenic crops? – Implications of a field test. Insects 2, 400–411. Ludwig, G., Schnittler, M., 2009. Rote Liste der Pflanzen Deutschlands, http://www. bfn.de/fileadmin/MDB/documents/RoteListePflanzen.pdf Malone, L.A., Todd, J.H., Burgess, E.P.J., Walter, C., Wagner, A., Barratt, B.I.P., 2010. Developing risk hypotheses and selecting species for assessing non-target impacts of GM trees with novel traits: The case of altered-lignin pine trees. Environ. Biosaf. Res. 9, 181–198. Obrist, L.B., Dutton, A., Albajes, R., Bigler, F., 2006. Exposure of arthropod predators to Cry1Ab toxin in Bt maize fields. Ecol. Entomol. 31, 143–154. Rovenska, G., Zemek, R., Schmidt, J.E.U., Hilbeck, A., 2005. Effects of transgenic Bteggplant expressing Cry3Bb toxin on host plant preference of Tetranychus urticae and prey preference of its predator Phytoseiulus persimilis (Acari: Tetranychidae, Phytoseiidae). Biol. Control 33, 293–300. Spök, A., Hofer, H., Lehner, P., Valenta, R., Stirn, S., Gaugitsch, H., 2004. Risk assessment of GMO products in the European Union. Umweltbundesamt GmbH, Wien/Vienna, Österreich/Austria. Schmeller, D.S., Gruber, B., Budrys, E., Framstedt, E., Lengyel, S., Henle, K., 2008. National responsibilities in European Species Conservation: a methodological review. Conserv. Biol. 22 (3), 593–601. Schuldt, A., Assmann, T., 2010. Invertebrate diversity and national responsibility for species conservation across Europe – a multi-taxon approach. Biol. Conserv. 143, 2747–2756. Vachon, V., Laprade, R., Schwartz, J.L., 2012. Current models of mode of action of Bacillus thuringiensis insecticidal crystal proteins: a critical review. J. Invert. Pathol. 111, 1–12, http://www.sciencedirect.com/science/article/ pii/S0022201112001358 Van Wyk, A., van den Berg, J., Hamburg, van H., 2007. Selection of non-target Lepidoptera species for ecological risk assessment of Bt maize in South Africa. Afr. Entomol. 15 (2), 356–366. Wall, D.H., Bardgett, R.D., Behan-Pelletier, V., Herrick, J.E., Hefin Jones, T., Ritz, K., Six, J., Strong, D.R., van der Putten, W.H., 2012. Soil Ecology and Ecosystem Services. Oxford University Press, pp. 424. Zwahlen, C., Hilbeck, A., Howald, R., Nentwig, W., 2003. Effects of transgenic Bt corn litter on the earthworm Lumbricus terrestris. Mol. Ecol. 12, 1077–1086.