Risks of viral transgenes?

Ulrich Melcher umelcher at BMB-FS1.BIOCHEM.OKSTATE.EDU
Fri Apr 28 09:09:06 EST 1995

     On April 21 and 22, 1995, the Animal and Plant Health Inspection
Service of the Department of Agriculture of the USA held a meeting in which
the risks of widespread release of transgenic plants containing parts of
viral genomes were considered.  The assignment for discussion was to answer
specific questions related to the risks of recombination between viral RNAs
and RNAs expressed from transgenes.  Discussions were held in five smaller
groups and summarized at the end of the meeting.  An official report of the
meeting is being prepared and will be available in about six months.
Meanwhile, here is a very personal and unofficial summary of the sense of
the meeting.  I stress that this is PERSONAL and UNOFFICIAL.

     Genes that may be or have been used in developing virus-resistant
plants include coat protein, movement protein, replicase and other genes.
Some of the resistances require expression of the genes; others may be
mediated by antisense constructs or by the ill understood phenomenon of
co-suppression.  Coat protein-mediated resistance has been demonstrated
effective in a large number of cases.  In a few cases, the protection is
strong, completely preventing infection.  In others, infection is delayed
long enough to avoid crop losses associated with infection.  The presence
of coat proteins in these plants complicates plant and seed certification
programs since these often rely on detection of the viruses by antisera
against the coat proteins.  Resistance afforded by defective movement
proteins slows the initial spread of infection of the virus from which the
gene was derived and closely related viruses.  However, it also protects
the plant against systemic, but not local, infection by viruses from a wide
range of taxa.  Studies of defective movement protein-mediated protection
are in their infancy.  Replicase-mediated resistance is usually stronger
than coat protein-mediated resistance but has a higher requirement for
sequence similarity between the transgene and the challenge virus than does
coat-protein mediated resistance.  Co-suppression also is highly effective
but limited to viruses that are about 90% or more identical in sequence to
the transgene.  Anti-sense-mediated resistance was only mentioned briefly.

     Risks are regarded as the product of the frequency of an event
multiplied by the hazard associated with the event.  Some participants
attempted to focus on the two components of risk separately.  Others argued
based on known risks.  It was generally agreed that recombination occurs
frequently among positive-strand RNA viruses of plants when they coinfect a
plant.  Both homologous recombination and non-homologous recombination are
known to occur.  Such recombinations can be intramolecular or

     Evidence for viral recombination derives from coinoculation
experiments using selective pressure to recognize viable recombinants, PCR
experiments to analyze nucleic acids present in mixed infections, the
presence and generation of defective (often interfering particles), the
deletion of extra material inserted into wild type genomes and sequence
comparisons of related viruses.  Attempts to measure the frequency of these
events was complicated by a lack of an ability to directly measure the
events.  Instead, measures such as the frequency of recombinant viruses in
a population or the fraction of inoculated plants in which a recombination
event occurred were used.  In those systems and viral taxa where
recombinant frequencies have been measured without requiring that the
recombinants be viable or superior to wild type, recombination appeared to
occur frequently.  However, not enough experiments with enough taxa have
been done to make a bold generalization.

     Whether the frequency of generating recombinants would be greater or
less when the recombining molecules were an infecting viral RNA and an RNA
derived from a viral transgene than when they were two infecting viral RNAs
was considered.  On an individual plant basis, the consensus was that the
frequency would be substantially less.  The plant expressing the transgene
produces, in general, lower amounts of viral RNA than are produced in an
infection by a virus.  On the other hand, it was pointed out that during a
natural infection, much of the RNA is not available for recombination,
being coated with proteins (coat proteins, ribosomes, movement proteins), a
situation that may not obtain during transgene expression.  Local RNA
concentrations were generally, though not universally, thought important in
determining the rate of recombination.

     In the case where virus-resistant transgenic plants are deployed
commercially, it was felt difficult to predict the consequences for the
frequency of generating recombinants.  On the one hand, the widespread
deployment will greatly reduce the number of plants in which high
concentrations of viral RNA from natural infection by the targeted virus
occur.  On the other hand, in any given growing season with non-transgenic
plants not all plants in all locations will become infected.  Thus, after
widespread deployment, more plants will contain the viral sequence than
would without such deployment.  An epidemiological model is required to
determine whether deployment increases or decreases the frequency of

     The issue of the hazards of recombinants was hotly debated.  Many
argued that all recombinants are either non-viable or enfeebled relative to
wild type parents.  Others cautioned that some recombinations could result
in recombinants that had different, possibly expanded, host ranges or
vector specificities.  Increased virulence was also a possibility, but was
not discussed much, perhaps due to a poor understanding of the determinants
of virulence of plant viruses.  Most of the individuals holding that
recombinants were non-hazardous also felt that there had not been enough
experimentation addressing this question.  Some data was presented
suggesting that intragenic recombinants between closely related viruses
(80% similarity or greater) were viable.  It was speculated that intragenic
recombinants between more distant viruses would not be viable.  Intergenic
recombinants might tolerate a greater sequence dissimilarity.  A few
examples of field isolated viruses that appear to be viable recombinants
were cited.  The examples included tobraviruses and examples of viruses
that appear to have acquired host sequences.

     Many held that because of the high suspected rate of recombination and
the natural prevalence of mixed infections (well documented), viable
recombinants able to compete with other viruses would already have arisen.
Those recombinants (mentioned above) that have been identified, though
pathogenic, are not particularly noxious viruses.

     Whether the deployment of virus-resistant transgenic plants could
provide novel opportunities for recombination by placing the transgene RNA
in locations where it is never found naturally was briefly considered.
Some virus infections are naturally limited to certain tissues.
Phloem-limited luteoviruses were prominently mentioned.  Some argued that
when a luteovirus transgene is put in a plant such that the RNA is
expressed in almost all tissues, a novel opportunity for recombination
arises.  Others countered that in mixed infections with luteoviruses, the
luteovirus moves out of the phloem via complementation by the movement
functions of the other virus.  Thus, opportunities for recombination have

     Deployment of virus-resistant transgenic plants in areas where the
virus from which the gene was derived does not exist was not well
considered.  In this case the transgene RNA will recombine with RNAs of
other viruses endemic to the area.  The recombinants will be unlike any
that have been previously generated during natural coinfections.  Some of
those may be hazardous.

     In a similar vein, the creation of transgenic plants containing a gene
derived from a virus that does not normally infect that plant species may
also create the opportunity for the generation of recombinants that have
not been made before in nature.  This would occur if a second virus infects
the plant and the two viruses do not share a common host.  Such transgenic
plants may be created since it has been demonstrated that transgenic
tobacco plants expressing a defective TMV movement protein gene resist
systemic infection by a range of viruses from widely different taxa.

     An additional scenario, not directly related to recombination, was
raised in the case of resistance mediated by co-suppression.  Many natural
virus populations exist as quasispecies, a population of variants with a
common consensus sequence.  If this population is broad enough that in the
region of the co-suppression target, some members fall below the 90%
identity putatively required for suppression, then these outlying variants
may survive and adapt to form a new quasispecies, substantially different
from the original.

     Several participants addressed the risks directly without considering
the two components of risk.  One view was that in the extensive history of
world agriculture which includes the movement of plant species into new
continents there is no evidence of the generation of any particularly
monstrous viruses.  The activity represented by the widespread deployment
of virus-resistant transgenic plants will not differ significantly from
what has already occurred in agricultural history.  Another view was that
massive deployment of these transgenic plants represents risks that can not
be adequately addressed by field trials.

     There seemed to be general agreement that, in the U.S., monitoring of
transgenic crops for potentially novel viruses by producers and their
reporting of suspicious infections will be adequate to monitor for
hazardous recombinants.  There was less enthusiasm for the efficacy of
monitoring in the developing world and in other developed countries, such
as the United Kingdom.

     The meeting was attended by about a hundred individuals interested in
plant virology.  They included academic, government and industrial
scientists, other industry representatives and government regulators.
Though the individuals were primarily from the US, representatives of other
countries were present.  The meeting was part of the Clinton-Gore program
to reinvent government.

Ulrich Melcher                 umelcher at bmb-fs1.biochem.okstate.edu
Department of Biochemistry     Tel: 405-744-6210
  and Molecular Biology        FAX: 405-744-7799
Oklahoma State University
Stillwater OK 74078-0454 USA

More information about the Virology mailing list