A tiered approach to risk assessment of virus resistance traits based on studies with wild brassicas in England.

Abstract

AbstractConcerns about the introduction of genetically modified crops frequently centre on the possibility of gene transfer to wild relatives, resulting either in the disruption of natural patterns of genetic diversity by introgressing into species gene pools or in the addition of traits which may cause wild plants to become more abundant or invasive. This chapter describes a phased approach to the assessment of possible harm in the context of a specific transgenic stress-tolerant trait, namely virus tolerance in inter-breeding species. We assessed the hazard of harm to the 'natural' environment as opposed to agricultural productivity. Our baseline was the distribution and relative abundance in field-grown wild or long-established 'naturalized' Brassica species (B. oleracea, B. nigra and B. rapa) of six viruses, and then glasshouse assessments of components of fitness. Because these observations suggested that generic risk assessment was unlikely to be possible, we focused on the economically significant turnip mosaic virus (TuMV), genus Potyvirus. TuMV is a target for transgenic (capsid-coding sequence-based) approaches to disease management in brassicas as an alternative to natural sources of resistance/tolerance to the virus and because insecticides do not kill the aphid vectors of the virus before they effect inoculation. TuMV was not found in B. rapa growing on the banks of the River Thames in Oxfordshire, UK. Glasshouse tests showed that B. rapa from these populations died within a few days of manual inoculation with some isolates of TuMV, but we found that the pathogenicity of three TuMV isolates from the UK was not uniform. We made crosses in which natural B. rapa lines, genome designation AA, n=10, were the female partners and B. napus, genome designation AACC, 2n=38, were pollen donors, B. napus included untransformed lines and lines that contained a transgenic capsid-coding sequence from a potyvirus. As judged using polymerase chain amplification, 'C' genome transfer frequencies varied from 0 to 84% depending on pollen donor, but there was statistically significant within-population variation among B. rapa, P<0.001 at Culham and P<0.05 at Clifton Hampden, in the efficiency of transgene flow from one B. napus cultivar ('Drakkar'). When manually challenged with TuMV, the transformed B. napus was infectible but the virus was not lethal. In contrast, the untransformed counterparts of these plants were sensitive to the same challenge inocula although two cultivars ('Westar' and 'Drakkar') differed in their absolute infectibility by one of the three isolates of TuMV we assessed. Importantly, when F1 hybrid progeny, identified on the basis of the presence of 'C'-specific sequence and capsid-coding sequence as judged by polymerase chain amplification, were manually challenged with TuMV, these plants tended to be more TuMV tolerant than their maternal parents. Thus, our contained glasshouse-based gene flow and pathogenicity tests, even though done in non-competitive conditions and with an incomplete knowledge of factors regulating the wild populations, provided prima facie evidence of a potential for ecological release from that natural virus constraint following introgression of a resistance trait. The assessments of putative fitness impacts in hybrids between transformed (virus-tolerant) crop plants and wild (virus-sensitive) crop relatives were assembled during an ongoing EU-funded project (VRTP-IMPACT; QLK3-CT-2000-00361).