I post in response to the request for opinions on the Greene and Allison
Science paper and the accompanying Perspectives piece by Falk and Bruening.
First, two opinions of note had previously appeared in this forum and are
available in the bionet archives. The opinion of Gus deZoeten, Allison's
department head, were summarized and posted by me. Roger Hull responded with
an opinion of his.
Second, the 22 April issue of Science has an exchange of letters between Mellon
and Rissler of the Union of Concerned Scientists and Falk and Bruening on this
issue.
Third, here is MHO:
The Greene and Allison paper is only significant when read in the context of
the prevailing paradigms in plant RNA virus recombination. These paradigms
have changed over the years. I find the pattern of changes foreboding
particularly compared with the progress in understanding recombination in the
caulimoviruses, one of my research subjects.
When cross protection of plants by transgenic coat proteins was beginning to be
investigated, the prevailing view was RNA's OF PLANT VIRUSES DO NOT RECOMBINE.
Recombination occurs in caulimoviruses such as CaMV, but that, it was implied,
was not relevant because CaMV virions contain double-stranded DNA, not RNA.
Then, recombination at homologous sequences under selected conditions was
demonstrated for an RNA virus. The view then became RNAs OF PLANT VIRUSES DO
NOT RECOMBINE AT NON-HOMOLOGOUS SEQUENCES. The presence in numerous viruses of
defective interfering (DI) particles soon silenced that paradigm.
Immediately prior to the Greene and Allison paper the view was: even though
plant viral RNAs can recombine, RECOMBINATION BETWEEN A TRANSGENE RNA AND A
VIRAL RNA IS NOT POSSIBLE BECAUSE THE TWO ARE PRODUCED IN THE WRONG PLACES AND
IN THE WRONG AMOUNTS. Work in the Hohns' labs and in Schoelz's lab had already
shown that recombination between inoculated defective CaMV DNA and a transgenic
CaMV DNA fragment was possible. Again, I suppose caulimoviruses were thought
irrelevant even though a significant part of their lifestyle involves RNA
intermediates and an error prone polymerase not unlike the RNA replicases. I
think the Greene and Allison paper did a good job of shooting down this straw
man argument. Therein lies its significance.
Now the view adopted by Falk and Bruening is: even though recombination
between a transgene RNA and an inoculated RNA can occur under selective
conditions, RECOMBINANTS WILL NOT SURVIVE UNDER NON-SELECTIVE CONDITIONS.
Recombination has been observed under non-selective conditions between isolates
of CaMV. Indeed recent phylogenetic analysis of isolate nucleotide sequences
by Kelly Chenault and myself (likely soon to be published in J Mol Evol)
suggests that recombination was fairly frequent in the evolutionary history of
CaMV.
Any one willing to bet what the next major discovery in the RNA virus
recombination field will be? Given the pattern of each of the above paradigms
eventually being slain by the ugly facts, I suggest that survival of non-
selected recombinants of RNA viruses will soon be demonstrated. Certainly,
some existing viruses appear to be products of such recombination. I would
also guess that viral RNA recombination happens frequently; perhaps frequently
enough that most potentially "evil" RNAs have already been tried by selection
and found wanting. If this is the case, then the main risk is: planting a
plant with a viral transgene in a place where the virus from which the
transgene was derived does not exist. A novel, untried combination.
Which of the above two views is correct is not clear to me. It is clear to me,
as it was to deZoeten, that we understand too little about viral nucleic acid
recombination and virus evolution to be able to make reliable pronouncements
about the safety or hazard of the widespread dissemination of such plants.
More combinations need to be tried under a variety of conditions. More
research should be done.
Ulrich Melcher
Biochem & Mol Biol
Oklahoma State Univ
