Genetically Engineered Fungus Bites Back at the Crops It's Meant to Save

New Scientist (UK)
September 28, 2002
Bob Holmes

FOR the first time, a fungus has been genetically modified to be more deadly
to the weed it blights. The snag is that the GM fungus kills crop plants as

While the modified fungus will not be released as a result of the findings,
the case shows how genetic modification can have unintended consequences. It
is also proof, were any needed, that biotechnology could be harnessed to
create weapons that attack crops (see "'Act now' plea on bioterror threat").
The fungus was modified to attack velvetleaf weed (Abutilon theophrasti). As
it is a close relative of cotton, most weedkillers that target it destroy
the crop as well as the weed. "Herbicides don't work, and that's where you
have to head in with biocontrol," says Jonathan Gressel, a plant
physiologist at the Weizmann Institute of Science in Rehovot, Israel.

In theory, diseases are ideal for biocontrol, because many infect just one
plant species. The US, for example, is testing funguses that target coca
plants or opium poppies. But diseases such as the fungus that causes
anthracnose in velvetleaf tend not to make good killers, as any that
eradicates its host is itself doomed.

So Gressel's team decided to give the anthracnose fungus (Colletotrichum
coccodes) a killer punch by adding a gene for a toxin from another fungus,
Fusarium. The modified fungus was indeed much more lethal to velvetleaf
seedlings in greenhouse experiments, they report in Nature Biotechnology
(DOI:10.1038/nbt743). "This puts it over the brink to something that would
be useful," says Gressel.

But the enhanced fungus also killed off tomato and tobacco seedlings,
neither of which would normally be affected by the anthracnose fungus. This
is exactly the kind of unexpected consequence of a genetic modification that
opponents of GM have been warning about, although the case also shows that
such effects can be detected at an early stage. "This business of putting in
a toxin raises a red flag," says Jane Rissler of the Union of Concerned
Scientists in Washington DC.

To allay these fears, Gressel suggests further "fail-safe" modifications to
any such fungus before it is tested outside a sealed greenhouse. Removing
the genes for sexual reproduction would prevent it passing on any added
virulence genes to related fungi that attack other plants. And removing the
genes for spore formation would prevent it spreading via the air, and ensure
it died out completely each winter.

Even these measures, however, may not be foolproof. For example, the fungus
may survive without spores, especially in moist tropical regions, says plant
pathologist Alan Watson at McGill University in Montreal.

There may be other, safer ways to boost the killing power of biocontrol
agents. Watson has patented a mixture of anthracnose fungi and low doses of
ordinary herbicides for weed control. The herbicide interferes with the
weed's normal defences against disease, allowing the fungus to get the upper
hand. It has yet to be tested on cotton fields, however. Bob Holmes