¿ìè¶ÌÊÓƵ

In the long-running drama of agricultural pest control, few insecticides
emerge unequivocally as good guys. Most could be cast as surly outsiders,
lashing out at friend and foe alike and tolerated only because they help
to clear away the pests, an even worse crowd. But one insecticide, a group
of naturally occurring toxins produced by the soil bacterium Bacillus
thuringiensis, is a much more lovable hero.

These B. thuringiensis toxins kill serious pests such as caterpillars,
beetles and fly larvae, while sparing humans, spiders and most beneficial
insects. And hailing as they do from bacteria rather than a chemicals factory,
the toxins are a biotechnologist’s dream. Opportunities abound for manipulating
the genes that encode them. Transfer these genes into crop plants, for instance,
and you are a step closer to producing insecticidal crops – ‘no-spray’ cotton,
potato and corn fit for Utopian farms of the 21st century. Not for nothing
are major agrochemicals and biotechnology companies investing heavily in
developing such transgenic crops.

But now the hero’s image is slipping. A handful of pest populations
have grown resistant to B. thurin-giensis toxins, something many pest-control
experts thought would never happen. Many other pests, including serious
scourges such as the Colorado potato beetle and the tobacco budworm, have
the potential to develop resistance as well, according to the latest laboratory
results. And, true to the form of a classical tragedy, one of the hero’s
greatest strengths, a biological pedigree, may contain the seeds of the
B. thuringiensis toxin’s undoing. Many pest-management experts believe that
putting toxin genes into crop plants could speed up the evolution of a formidable
army of resistant pests.

That would fulfil the gloomiest prophecies of green activists and others
campaigning against genetically engineered crops. But researchers are unwilling
to throw in the towel. Far from it. In laboratories throughout academia
and industry, the race is now on to find a strategy for outwitting evolution
and forestalling the development of resistance to B. thuringiensis toxins.

At stake in this rearguard action is a family of protein toxins with
remarkably selective killer instincts. Some B. thuringiensis toxins kill
only butterfly and moth caterpillars, others only fly and mosquito larvae,
and still others target only beetle larvae. Each toxin is ‘like a surgical
tool for taking out the pest’, says entomologist Fred Gould of North Carolina
State University. ‘It doesn’t disrupt (the pest’s) natural enemies. That’s
a big advantage over the broad-spectrum pesticides.’

In recent years, biotechnology companies have been scouring the world
for bacteria with novel toxins. To date, scientists have identified more
than 50 such toxins, with targets including ants, mites and even parasitic
roundworms and flukes. The amino acid sequences of these toxins differ from
one to the next, but must fall into one of several related molecular families.
No one knows why the bacteria have these toxins. But whatever their natural
function, farmers and foresters have seized on them eagerly. Sales of B.
thuringiensis toxins increased more than fourfold during the 1980s to a
worldwide total of more than $100 million. It is this burgeoning application
that is provoking concern about resistance.

The scenario is all too familiar to users of synthetic chemical pesticides.
Chemists find a new, potent insecticide, and farmers begin to lavish it
on their crops. But among tens of millions of individual insects, a few
will be genetically equipped by chance to survive the spray. Being better
able to reproduce, these hardy individuals will contribute more than their
share of offspring to the next generation, each of which will carry the
resistance genes. As each new generation emerges, the insect population
becomes ever more resistant to the toxin. This pattern has repeated itself
for pest after pest, insecticide after insecticide. Today, more than 500
species of insects show resistance to at least one chemical, and a few ‘superbugs’
have evolved the ability to cope with almost every poison in the farmer’s
arsenal.

Until recently, however, many pest-control experts believed that B.
thuringiensis toxins were different, because most commercial sprays contain
a mixture of several toxic proteins. The odds were minute, they reasoned,
that an insect would be resistant by chance to all the toxins at once. Moreover,
farmers had been spraying B. thuringiensis toxins for more than two decades
without the slightest evidence of resistant insects. But some scientists
have always been sceptical. ‘Whenever there’s a new insecticide, people
think of reasons why it’s impossible for insects to become resistant to
it,’ says Bruce Tabashnik, an entomologist and expert on resistance based
at the University of Hawaii. ‘People like me just assume things are going
to become resistant. I think that’s a safer viewpoint.’

For B. thuringiensis toxin, as for every other insecticide so far,
this scepticism is fast being vindicated. In 1985, the first example of
B. thuringiensis resistance turned up in moths taken from grain storage
bins in the American Midwest. In 1990, Tabashnik and his colleagues found
resistant diamondback moths in Hawaiian cabbage and watercress fields. Since
then, resistant diamondback moths have also come to light in agricultural
fields and greenhouses in Florida, New York, Japan and parts of the Asian
mainland. So far the problem amounts to only a handful of resistant pest
populations in the countryside. But there are other danger signs. Almost
two dozen experiments, involving the breeding of successive generations
of insects exposed to B. thuringiensis toxins in laboratories, have confirmed
that a wide range of insects has the capacity to develop resistance.

Genes of wrath

Recent advances in biotechnology may increase the risk of resistance
in the near future, many experts fear. At present, farmers spray B. thuringiensis
toxins onto their crops, where sunlight robs the toxins of their potency
in just a couple of days. As a result, a single treatment provides only
temporary pest control. Biotechnology companies are testing several strategies
to boost the killing power of the toxins, most notably by transferring toxin
genes from soil bacteria into crop plants. These transgenic plants would
produce their own B. thuringiensis toxins continuously, so the Sun would
no longer bleach them into impotence. Furthermore, plant-produced toxins
could kill even those pests that burrow deep into the plant’s stem or roots,
out of reach of most sprays. The chemicals giant Monsanto, widely regarded
as an industry leader in this field, expects to have genetically engineered
cotton and potato plants that produce B. thuringiensis toxins within three
or four years, and maize a year or two later, says Randy Deaton, a company
researcher.

Yet the same features that make such transgenic plants more efficient
at killing pests in the short term also increase the odds that only pests
with resistance genes will survive to reproduce. This hastens the spread
of resistance genes, and may shorten the useful life of B. thuringiensis
toxins as pest-control agents. ‘If it weren’t for the transgenic plants,
there wouldn’t be such an urgent need to deal with the resistance problem,’
says Tabashnik.

As awareness of this risk spreads, pesticide companies, fearing the
loss of a lucrative product, are paying close attention. Four years ago,
a group of companies involved in selling and developing B. thuringiensis
toxins formed a working group to investigate how best to avoid resistance.
The group’s budget is modest – about $250 000 over four years. But scientists
say it has funded useful research on the molecular and cellular basis of
B. thuringiensis toxicity and resistance. In the future, the working group
hopes to fund experimental tests of strategies for minimising the threat
of resistance.

Critics, however, accuse the group of dragging its feet. ‘It’s the old
style of working – let’s study it to death,’ says Maureen Hinkle, a director
of agricultural policy for the National Audubon Society, a zoology and conservation
organisation. ‘We’re past that. By the time resistance appears, it’s too
late. You’ve got to pay attention to clues that it’s coming, or you’ve lost
the battle. I think it’s scientific suicide to sit back and say, let’s wait
and see what’s going on.’

Yet efforts to design a practical programme of resistance management
for B. thuringiensis rapidly bog down in swampy uncertainty over what exactly
to do. Despite more than 30 years of fighting insect resistance to one insecticide
after another, scientists know little about which tactics work and which
don’t. Most of what they do know comes from theoretical calculations, computer
models and a few artificial laboratory experiments. The conclusions from
such exercises are only as good as the assumptions that underlie them,
says Tabashnik. Virtually everyone complains of a lack of data from real
situations.

‘We can model and we can discuss, and we can try to run lab experiments,
but everyone agrees that you can’t do that sufficiently to come up with
an answer that will make us all comfortable,’ says Wendy Gelernter, an insect
pathologist with Mycogen Corporation in San Diego. ‘It’s almost like you
need to try several alternatives on some limited commercial scale and see
what happens.’

The most straightforward approach to resistance management has always
been simply to hit the pests as hard as possible. If farmers use high enough
doses of insecticide, the reasoning goes, they will kill even somewhat resistant
insects and thus prevent the genes that convey resistance from spreading.
For a similar reason, doctors warn their patients not to stop a course of
antibiotics halfway through, but to keep taking them until all the pills
are gone.

Such a strategy, however, must be executed perfectly if it is to work.
‘Obviously, if there are any survivors of a high-dose strategy, they’re
going to be really resistant,’ says David McAuliffe, a biologist with Du
Pont Agricultural Products. Sceptics question whether any pesticide can
ever wipe out every last insect in a population – and shooting for this
mark and missing, they say, may be one of the quickest routes to producing
insects that are highly resistant to insecticides.

A more promising approach substitutes guile for brute force. By varying
their choice of pest-control methods over time, farmers could force natural
selection to chase a moving target, thus delaying the pests’ adaptation
to any particular control tactic. In the simplest version of this strategy,
farmers could simply alternate between different B. thuringiensis toxins,
giving pests at most a few generations to adapt to one toxin before confronting
them with a different one.

Such a strategy will work only if the pests can’t adapt to all the toxins
simultaneously, however – and this assumption may not always hold true.
B. thuringiensis toxins kill by binding to a receptor molecule in the insect’s
digestive system and causing the gut wall to disintegrate; insects can become
resistant by evolving a differently-shaped receptor that no longer binds
the toxin. Since different toxins often use different receptors, insects
that evolve resistance in this way to one toxin would have no advantage
in coping with a second one. Laboratory selection experiments confirm this,
showing that populations selected for resistance to one B. thuringiensis
toxin usually have little cross-resistance to unrelated toxins.

However, resistance to B. thuring-iensis can develop in other ways as
well. For example, Gould is studying a laboratory population of tobacco
budworm that is resistant to several unrelated B. thuringiensis toxins
simultaneously. How the insects manage that feat still is not known, but
Gould conjectures that enzymes might disable the toxins before they bind
to their receptors. And larvae of the rice moth produce a thick mucous layer
lining their gut which interferes with the binding of toxins. In either
case, alternating toxins might prove useless. Indeed, Tabashnik suggests
that farmers who vary the toxins they spray may simply increase the evolutionary
pressure on the pests to develop general resistance mechanisms of this
sort.

Most pest-control experts favour a more complex approach that uses B.
thuringiensis toxins as one component of an integrated pest management
(IPM) programme, says Susan MacIntosh, a scientist at Novo Nordisk Entotech
in Davis, California, who chairs the working group investigating resistance
to B. thuringiensis toxins. IPM drives down pest populations not with a
single blow from one pesticide, but with many small taps – crop rotation,
natural enemies, altered planting dates to miss periods of peak pest development,
sparing use of B. thuringiensis toxins and other pesticides. Since pests
suffer little mortality from any one source, they are less likely to evolve
resistance, advocates contend.

Plants engineered to produce B. thuringiensis toxins might fit into
such a scheme in several ways. Genetic engineers could design plants that
carry sublethal doses of the toxins, thus merely slowing down the pests’
growth instead of killing them. In Massachusetts’ potato fields, that might
be enough to keep the Colorado potato beetle from squeezing a second, damaging
generation into each summer, says David Ferro, an entomologist at the University
of Massachusetts.

Two other possibilities still on the drawing board involve selectively
activating the genes that encode the toxins. By engineering plants to produce
B. thuringiensis toxins only in certain tissues, scientists could steer
the pests towards less vulnerable parts of the plant. In cotton, for example,
the ability to confine the toxins to the bolls would protect these sensitive
parts and shift pests’ sights to the foliage, where damage has little effect
on fibre yield, says Gould. And researchers at CIBA Seeds in North Carolina
are developing plants whose toxin genes lie dormant until activated by a
chemical spray, allowing farmers to hold the plants’ toxins in reserve until
absolutely necessary.

For the moment, however, experts agree that the most promising approach
to controlling resistance is to grow toxin-free plants alongside the genetically
engineered variety. Refuge areas would allow pest individuals that are
still susceptible to B. thuringiensis toxins to survive and contribute their
genes to the next generation. Such refuge plants should work best in combination
with a high-dose strategy for toxin-treated plants. ‘The ideal is to be
very successful at killing most of the population, and then have a substantial
fraction that you don’t treat at all,’ Tabashnik says. The fact that only
the most resistant pests survive on the high-dose plants – a fatal disadvantage
when that strategy is used alone – is much less relevant when those few
survivors are swamped by a tide of susceptible insects from the refuge.

This ideal is difficult to attain with conventional spray insecticides,
because drift and inaccuracies in spraying blur the boundaries between the
kill zone and the refuge. ‘We don’t have any chemicals out there that can
kill close to 100 per cent of insects and have a refuge at the same time,’
Gould says. As a result, no one knows whether such refuges would work as
well in practice as they do in theory. Now that transgenic plants that produce
B. thuringiensis are becoming available, however, trials should be under
way.

Until then, scientists can only theorise about the best way to design
such refuge areas. How big should the refuge be? No one knows, though the
figure of 10 per cent of toxin-free plants often pops up in scientists’
conversation. Should farmers mix toxin-free plants in the same field as
ones that have been genetically altered or plant a separate refuge adjacent
to their treated field? Again, no one really knows, though logic suggests
that if the refuge is too far from the treated plants, the resistant survivors
will mate among themselves instead of diluting their genes by mating with
susceptible insects from the refuge. For stay-at-home pests such as the
Colorado potato beetle, therefore, refuge plants intermingled with toxin-producers
might be best. On the other hand, pests that roam widely, taking a nibble
here and a nibble there, may require separate plantings to ensure that susceptible
individuals will find a toxin-free diet.

Refuges in the real world pose another problem as well: farmers may
be reluctant to deliberately feed some of their current year’s yield to
pests for the intangible benefit of delaying resistance at some unspecified
time in the future. A farmer who plants 10 per cent of his acreage as a
refuge will have a smaller profit margin than one who plants only five per
cent. Tabashnik suggests that seed companies could sell toxin-producing
plants ready mixed with refuge plants, thus giving farmers no choice in
the matter. Even then, however, a company selling a 5 per cent seed mix
might enjoy a sales advantage over its more altruistic competitor’s 10
per cent mix. ‘I think this particular solution is attractive scientifically
but not practical commercially,’ adds Mycogen’s Gelernter.

Of course, governments could step in to set minimum refuge standards,
as some environmental organisations suggest. Few industry scientists favour
such a move, however. ‘I’m not sure that a dictate by the government would
really improve the situation,’ says industry spokesperson MacIntosh. ‘What
is the best strategy? I can’t tell you that. I don’t think anyone can.’

Besides, leaving refuge design unregulated might lead to trial and
error over a broader range of possible strategies. This real-world experience,
Tabashnik says, might be the quickest way to improve our understanding of
the complex world of resistance management – not to mention safeguarding
the heroic image of a certain group of natural pesticides.

Bob Holmes is a freelance science writer based in California.