To the editor:

In the continuing debate over the safety of transgenic plants containing the cauliflower mosaic (pararetro)virus (CaMV) 35S promoter, Cummins, Ho, and Ryan write that pararetroviruses, unlike retroviruses, do not require integration into host chromosomes to complete their life cycle1. Although this statement is formally correct, it does not take into account recent data revealing the presence of integrated pararetroviral DNA in plant chromosomes2,3,4.

Banana streak (pararetro) viral (BSV) sequences, which could recombine to create infectious virus, have been detected at several sites in the banana genome2,3. In another example, approximately 1,000 copies of a previously unidentified tobacco pararetrovirus (TPV) were found integrated into tobacco nuclear DNA4. These TPV copies contained frameshifts and other mutations, and therefore, unlike the BSV case, would be unable to produce infectious virus. Nevertheless, the putative TPV promoter has been shown to be active in transgenic constructs reintroduced into plants (W. Aufsatz, M. Matzke, and A. Matzke, unpublished data). TPV-related sequences were also detected in high-molecular-weight DNA of other solanaceous plants, such as tomato, by Southern blot analysis4.

It is therefore likely that numerous additional examples of integrated pararetroviral sequences will be found in plant chromosomes. Although these results do not reflect on the safety of the CaMV 35S promoter per se, they do demonstrate that viral sequences, including potentially active viral promoters, are not exotic components of edible-plant genomes.