1. Bt Crops Paul Jepson Integrated Plant Protection Center Oregon State University

2. Examples of traits and their associated genetic elements and sources Trait Genetic Element Gene Source Insect resistance Cry1Ab delta-endotoxin Cry1Ac delta-endotoxin Cry3A delta-endotoxin Cry9c delta-endotoxin Protease inhibitor Bacillus thuringiensis subsp. kurstaki B. thuringiensis subsp. kurstaki B. thuringiensis subsp. Tenebrionis B. thuiringiensis subsp. Tolworthi S. tubersosum Source: AgBiosafety, UNL (08/2001) www.agbiosafety.unl.edu

3. Regulatory approvals for Bt transgenic crops ArgAustCanChiJapMexS.AfrUSAEUNethSwiUK CottonEnvirFoodFeedMaizeEnvirFoodFeedPotatoEnvirFoodFeed98989896/8/0198/0198/0196969600/101969696/795/796/795/7/995/6/995/7/997979797/897/897/896/996/9969797979797979795/795/895/895/7/895/6/796/7/895/6/994/6/894/6/897/897/897/8979797/897/897/8 Source: AgBiosafety, UNL (08/2001) www.agbiosafety.unl.edu

4. Areas of concern in the public domain EnvironmentEnvironment Health Consumer rights and labeling Ethics Concerns targeted to the poor and excluded Sustainable vs “industrial” agriculture Source: Conway, G. (2000) Rockerfeller Foundation

5. GMO Environmental Risks and Benefits Risks of invasiveness Risks of invasiveness Non-target organism impacts Non-target organism impacts New viral diseases New viral diseases Reduced pesticide environmental impact Reduced pesticide environmental impact Reduced rate of land conversion Reduced rate of land conversion Soil conservation Soil conservation Phytoremediation Phytoremediation Source: Wolfenbarger & Phifer (2000) Science

6. Risk of invasiveness Steps that may lead to environmental harm Introduction of plant Survival outside cultivation Pollen flow to wild relatives Hybrid formation Reproduction outside cultivation Hybrid survival Hybrid Reproduction Self-sustaining populations Introgression of gene into wild relatives Spread and persistence Economic of environmental harm Source: Wolfenbarger & Phifer (2000) Science Example of this pathway for Canola

7. Investigations of risk of invasiveness Crop Pollen flow to relatives Hybrid formation Hybrid survival Hybrid reproduction Introgression to relatives B. napus Herbicide tolerant Gene flow possible Hybrid formation possible B. napus herbicide tolerant and fertlity restorer Hybrid reproduction possible B. napus herbicide tolerant Pollination of B. campestris, not with S. arvensis campestris hybrid formed Hybrid survives Hybrid reproduces B. napus altered oils B. rapa hybrid germinates Hybrid survival Source: Wolfenbarger & Phifer (2000), Science ? ? ? ?

8. Nature Biotechnology, 2004 22: 642

9. O. sativa-weedy rice hybrids containing herbicide resistant traits Chen, LJ et al 2004, Annals of Botany 93, 67-73

10. O. sativa O. rufipogon Design A Design B Design C Gene flow frequencies from rice to O. rufipogon varied between 3% - 8%, measured by SSR markers Bao- Rong Lu

11. Non-target data considered in latest EPA risk assessment for Bt crops* Larval and adult honeybee Larval and adult honeybee Green lacewing* Green lacewing* Ladybird beetles Ladybird beetles Parasitic Hymenoptera Parasitic Hymenoptera Monarch butterfly* Monarch butterfly* Avian oral toxicity Avian oral toxicity Static renewal acute toxicity, Daphnia Static renewal acute toxicity, Daphnia Corn as food for farmed fish Corn as food for farmed fish Collembola Collembola Earthworms Earthworms * Standard studies based on EPA Subdivision M and/or OPPTS 885 Guidelines OVERALL Very limited evidence for toxic effects*

12. Example of an EPA regulatory action October 15 th, 2001 ‘Biopesticides Registration Action Document’, USEPA OPP October 15 th, 2001 ‘Biopesticides Registration Action Document’, USEPA OPP B.t. Corn and B.t. Cotton B.t. Corn and B.t. Cotton Extended registrations with additional terms and conditions Extended registrations with additional terms and conditions – Non-target insects: field census data required – Monarch long-term exposure to Cry 1 Ab – Chronic avian study

13. Monarch butterfly Risk assessment over a large geographic scale

14. Monarch butterfly research Published in PNAS, 2001 Published in PNAS, 2001 Research addressed Research addressed – Sensitivity to B.t. protein and pollen in the lab – Pollen burden on milkweed in and near corn – Exposure assessment – Effects in the field – Overall risk assessment Corn used more extensively as a habitat by Monarch butterfly than expected Corn used more extensively as a habitat by Monarch butterfly than expected Risks to butterflies of most B.t. corn events low Risks to butterflies of most B.t. corn events low The most toxic event removed from the market-place The most toxic event removed from the market-place

15. Impacts of conventional pesticides on field boundary Lepidoptera Spray drifts onto field boundary vegetationSpray drifts onto field boundary vegetation Low pyrethroid doses have an anti-feedant effectLow pyrethroid doses have an anti-feedant effect Pupae are reduced in sizePupae are reduced in size Adults are smallerAdults are smaller (Cilgi and Jepson, 1995)

16. Butterfly mortality can occur as a result of pesticide drift into field boundaries Longley et al, ’97, Env. Tox. & Chem., 16, 165- 172 e.g. mortality of Pieridae in boundaries exposed to pyrethroid drift

17. Field census data: Natural Enemy Abundance in B.t. and Conventional Cotton Fields Head, G., Freeman, B., Moar, W., Ruberson J., And Turnipseed, S.

18. Spiders in Cotton Fields 0 10 20 30 40 50 60 20-Jun27-Jun04-Jul11-Jul18-Jul25-Jul02-Aug Date Abundance / sample Conventional Bt cotton Source, Head et al

19. Ladybird Beetles in Cotton Fields 0 20 40 60 80 100 120 140 20-Jun27-Jun04-Jul11-Jul18-Jul25-Jul02-Aug Date Abundance / sample Conventional Bt cotton Source, Head et al

20. Farm-scale effects of using genetically engineered crops YIELD NET RETURNS PESTICIDE USE Herbicide tolerant cotton ++0 Herbicide tolerant soybeans +0- B.t. cotton ++- Source, Fernandez- Cornejo,J., McBride, W.D. (2000) USDA ERS

21. Some beneficial invertebrates are locally extirpated by repeated application of conventional pesticides to whole fields e.g. Carabid ground beetles., 30, 696-705 These effects are not detected in short term, within-field experimental regimes They have been seen repeatedly in large-scale multi-field experiments Evidence for scale-dependency in ecological impacts

22. Many non-target species disperse within and between fields and repopulate areas following reductions after pesticide use ground spiders rove beetles ground beetles lacewings parasitoid wasps predacious bugs ballooning spiders hoverflies ladybird beetles ~ Short Range~ Long Range~ Middle Range

23. Transgenic vs. conventional delivery Indirect effects may be more important than direct effects –Specialist natural enemies may be reduced because of profound impacts on target herbivores in all B.t. fields Exposure pathways very different –Most non-target taxa are not exposed to plant- incorporated protectants at all: a feeding pathway is required for toxicity –Certain taxa exposed to PIP’s via diet, for the whole season, and some of these may be susceptible –Exposure is synchronized between fields

24. Spider population modeling Landscape average where 50% of fields sprayed in week 24, or in an increasing range of dates in weeks 20-23, & 25-29 (From Halley et al., 1996, J. Appl. Ecol. ) With short-persistence pesticides, small variation in the range of exposure timings can generate effective refugia within fields that have been sprayed

25. Rapid response of spiders to small variation in spray timing is a function of high dispersal and reproductive rates e.g. (Thomas et al.(2003) J. Appl. Ecol. )

26. Selection of organisms for monitoring Exposure: detritivores, herbivores, predators and parasites Indirect effects: trophic position, diet specialization

27. Conclusions (Part 1) Environmental risk assessment still under development, particularly for risk of gene flow, and for large scale effects Ability to determine higher risk situations improving PIP’s significantly different to pesticides Benefits of pesticide reductions need to be examined Potential use within sustainable development programs not certain: many other factors are important Acceptance of work demonstrating negative impacts has been poor ( J. Ag. & Env. Ethics. (2001) 14: 3028)

28. FIFRA Scientific Advisory Panel Concerns about data submitted by industry Many ‘tier 1’ tests submitted by industry are flawed or incomplete – Presence of toxin not demonstrated in artificial diets for test species – Control mortality so high that it masks possible effects – Some tests (e.g. treated insect eggs), do not expose certain insect predators to the toxin – No statistical analysis of some tests Field evaluation no substitute for tier 1 risk assessment testing Data do not support EPA statement that Bt corn (MON 863) results in less impact on non-target invertebrates than conventional pest management practices Several field studies had no statistical analysis to support them Plot sizes were trivial, reducing likelihood of detecting treatment effects (even highly toxic pesticide effects were nor detectable in some investigations) Statistical power halved in GM crop plots as a result of flawed experimental design

29. Recommendations: Specific Non-target invertebrates: Recommendations: Specific (Made in consultation with USDA APHIS, 2003) Develop database of approved protocols Develop database of approved protocols –Exposure and its validation –Species selection and source –Test protocols –Statistical approaches –GLP QA standards including criteria for non-acceptability Adopt a policy of exploiting the range of internationally available test methods Adopt a policy of exploiting the range of internationally available test methods –Participate in international working groups (SETAC, IOBC, ISO, OECD) Develop decision process for test selection Develop decision process for test selection –Relevant, but narrow range of options –Representativeness of taxon –Taxa that fall within susceptibility range Develop a database of results Develop a database of results Develop clear criteria for test evaluation Develop clear criteria for test evaluation –E.g. for field tests, standards for design, layout, sampling method, taxonomic resolution, statistics etc Adhere to the principles of tier-wise testing Adhere to the principles of tier-wise testing –Triggers between stages –Understand role and limitations of laboratory tests –Exploit semi-field, field, and monitoring studies Initiate development of geographically explicit risk assessment Initiate development of geographically explicit risk assessment –Zones of risk

30. Use of Environmental Impact Quotients to compare pesticide environmental risks in conventional and transgenic cotton

31. Traditional cotton (853 fields)B.t. cotton (1032 fields) MeanS.E.RangeMeanS.E.Range AI (kg/ha)4.100.010-13.42.380.060-25.1 Formulated pesticide (kg/ha) 15.570.260-78.48.580.190-55.0 Traditional cotton (853 fields)B.t. cotton (1032 fields) MeanS.E.RangeMeanS.E.Range Number sprays/crop 11.040.121-276.620.111-17 Mass application rates and spray frequencies in B.t. and traditional cotton

32. Average Pesticide Use/Grower (Kg/ha) Bt-CottonTraditional Cotton Most commodities world-wide are treated with <1 Kg/Ha/year

33. Average Number of Sprays/Grower Bt-Cotton Traditional Cotton

34. A method to measure pesticide environmental impact Rating system used to develop environmental impact quotient (EIQ, Kovach et al., 1992) (1, least toxic, 5, most harmful) Mode of action: non-systemic (1), all herbicides (1), systemic (3) Acute dermal LD50 for rabbits/rats (mg/kg): >2000 (1), 200-2000 (3), <1-200 (5) Long-term health effects: little or none (1), possible (3), definite (5) Plant surface residue half-life: 1-2 weeks (1), 2-4 weeks (3), >4 weeks (5) Soil residue half life: 100 d (5) Toxicity to fish (96h LC50): >10 ppm (1), 1-10 ppm (3), <1 ppm (5) Toxicity to birds (8-day LC50): >1000ppm (1), 100-1000 ppm (3), 1-100 ppm (5) Toxicity to bees: rel. non-toxic (1), mod. toxic (3), highly toxic (5) Toxicity to beneficials: low impact (1), moderate impact (3) severe impact (5) Groundwater and run-off potential: small (1), medium (3), large (5)

35. Calculating the EIQ EIQ = [C[(DT*5)+(DT*P)]+[C*((S+P)/2*SY)+(L)]+[(F*R) +(D*((S+P)/2)*3)+(Z*P*3)+(B*P*5)]}/3 DT= dermal toxicity, C= chronic toxicity, SY= systemicity, F= fish toxicity, L= leaching potential, R= surface loss potential, D= bird toxicity, S= soil half life, Z= bee toxicity, B= beneficial arthropod toxicity, P= plant surface half life EIQ (field use rating)= EIQ *%AI*rate

36. QuotientTraditional cotton (853 fields)Bt cotton (1032 fields) MeanS.E.RangeMeanS.E.Range Field EIQ199.23.870-723.7106.22.510-478.2 Farm worker191.03.870-747.0101.32.440-438.1 Consumer29.50.580-112.115.80.370-69.3 Ecological377.17.310-1374.3201.64.810-1124.5 Aquatic/fish92.21.900-329.445.01.220-322.7 Bird108.62.400-475.657.41.540-312.1 Bee62.31.210-219.235.40.840-333.6 Predator114.12.100-385.063.61.460-368.5 Ground water8.00.150-27.34.280.010-23.5 Terrestrial284.95.460-1045.0156.63.640-801.8 Picker32.10.650-128.318.50.460-100.0 Applicator137.92.900-564.971.71.810-343.0 Environmental impact quotient, based on Kovach et al. (1992)

37. Field Use EIQ Bt-Cotton Traditional Cotton

38. Surface reflectance: Global Vegetation Monitoring Unit, JRC, Ispra Farming systems, FAO/World Bank The sensitivity of different farming systems to disturbance highly variable

39. Conclusions/questions At what stage will we know enough to possibly reduce the requirement for extensive testing of Bt crops? Are we exhibiting dual standards by requiring greater scrutiny of Bt crops, compared with conventional pesticides? Is equivalent scrutiny required at each new location for GM crop adoption?