Multitrophic Cry-protein flow in Bt-cotton

Authors: EISENRING, MICHAEL - Agroscope; ROMEIS, JOERG - Agroscope; Naranjo, Steven; MEISSLE, MICHAEL - Agroscope

Submitted to: Agriculture, Ecosystems and Environment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/6/2017
Publication Date: 7/14/2017
Publication URL: http://handle.nal.usda.gov/10113/5763082
Citation: Eisenring, M., Romeis, J., Naranjo, S.E., Meissle, M. 2017. Multitrophic Cry-protein flow in Bt-cotton. Agriculture, Ecosystems and Environment. 247:283-289.

Interpretive Summary: Most genetically engineered (GE) cotton plants grown today produce the insecticidal Cry-proteins Cry1Ac and Cry2Ab (Bt-cotton) which are toxic for specific target pest insects. A major concern is the potential harm of Cry-proteins to non-target arthropods, particularly beneficial predators. This risk is thus assessed prior to the commercialization of any Cry-protein expressing GE plant. An important element of assessment is the exposure of non-target arthropods to plant-produced Cry-proteins. This is usually done in laboratory-bioassays. However, studies reporting how Cry-protein acquire from plants to herbivores to predators in the field are important for the verification and interpretation of these laboratory risk assessments. For Bt-cotton, such field data are lacking. In the current study we sampled cotton plants and arthropods in a Bt-cotton field in Maricopa, AZ over the growing season. Cry-protein concentrations in plants, herbivores, and predators were quantified and compared with laboratory-bioassays in which predators were fed with Cry-protein containing caterpillars. In the field, Cry-protein concentrations strongly decreased from plants to herbivores to predators. The latter mostly contained concentrations below the detection limit. Concentrations in arthropods were mainly affected by feeding mode and to a lesser degree by seasonal variation. Compared to plants, arthropods showed lower Cry2Ab:Cry1Ac ratios indicating that Cry2Ab might be less stable than Cry1Ac. Consistent with field results, Cry-protein concentrations measured in predators in laboratory-bioassays were low or below the detection limit. This indicates that laboratory-based feeding studies represent a realistic exposure scenario and are thus informative for the non-target risk assessment of Bt-cotton. These findings should be useful to adopters of the Bt technology and to regulatory agencies charged with assessing benefits and risks.

Technical Abstract: Although most genetically engineered cotton plants grown today produce the insecticidal Cry-proteins Cry1Ac and Cry2Ab, studies are lacking on multitrophic Cry-protein acquisition in dual-gene cotton fields. Such field data are important for the design and interpretation of laboratory risk assessment studies. We sampled cotton plants and arthropods in a dual-gene cotton field over the growing season. Cry-protein concentrations in plants, herbivores, and predators were quantified and compared with tri-trophic laboratory-bioassays. Cry-protein concentrations strongly decreased with increasing trophic level to values mostly below the detection limit in predators. Concentrations in arthropods were mainly affected by feeding mode and to a lesser degree by seasonal variation. Compared to plants, arthropods showed lower Cry2Ab:Cry1Ac ratios indicating that Cry2Ab might be less stable than Cry1Ac. Likewise, Cry-protein concentrations in laboratory-bioassays strongly decreased with increasing trophic level. This supports the validity of laboratory feeding studies for non-target risk assessment.