To the editor:

In presenting their justifications for reducing the regulatory burden on transgenic food crops (Nat. Biotechnol. 23, 439–444, 2005), we feel that Strauss and colleagues significantly misrepresent the implications and rationale of our report Genome Scrambling-Myth or Reality? Transformation-Induced Mutations in Transgenic Crop Plants1. Unlike their characterization of our work, we did not specifically “argue for rejection if even a single base pair is changed.” In full, our relevant recommendations were that “transgenic lines containing genomic alterations at the site of transgene insertions be rejected” and that “the insertion of superfluous DNA be considered unacceptable.”

Leaving aside the fact that a single base pair change may result in serious phenotypic consequences, these recommendations are best viewed in context. As documented in the report, thorough analysis reveals that all particle bombardment transgene insertion events include extensive rearrangements or loss of host DNA as well as insertion of superfluous DNA. Furthermore, a large fraction of even apparently simple Agrobacterium tumefaciens–mediated transgene insertion events also result in large-scale host DNA rearrangement or deletion and superfluous DNA insertion2. For example, loss of 76 kbp of host DNA3 and duplication/translocation of up to 40 kbp of host DNA have been reported at T-DNA insertion sites4.

Widespread use of transgenic crops carrying insertion-site mutations of this magnitude will, in our opinion, lead sooner or later to harmful consequences. Nevertheless, detailed inspection has shown that mutations such as these would almost certainly pass unnoticed through both the molecular and phenotypic characterization stages of the regulatory systems of both the European Union and the United States5,6,7,8.

We do agree with Strauss and colleagues that analysis of the phenotype is the one true measure of safety. However, rigorous assessment only at the phenotypic level is time consuming, expensive and, more importantly, of unproven effectiveness9. In this context, our recommendations for the detection and elimination of transformation-induced mutations from commercial crop plants are conceived as a straightforward and effective way to reduce the probability of unexpected deleterious phenotypes arising in transgenic crop plants and of protecting consumers and others from an unnecessary risk.