ÐÓ°ÉÔ­´´

FUNGAL DISEASES are one of the worst threats to cultivated plants. Both
the amateur gardener and the arable farmer know to their cost that fungal
infections can damage or destroy their plants. For a century or more the
urgent economic need to understand and control diseases of crops has driven
the science of plant pathology. But what the scientists had overlooked was
the extent to which wild plants are also vulnerable to fungal infections,
presenting an opportunity to turn fungal diseases to the farmer’s advantage
– as weedkillers. Mycoherbicides, fungi that kill weeds, promise to deliver
what other herbicides cannot: a highly specific treatment aimed at a single
species of weed that will leave other plants untouched.

Pathologists and ecologists now recognise that many wild plants are
afflicted by some extremely damaging diseases. As in crops, the symptoms
of disease – spotty or yellowing leaves and distorted stems, for instance
– may be similar on different plants. But such similarities hide the fact
that many different fungi cause these infections, and that many of them
are highly specific to their hosts. Each plant has its own collection of
pathogens, some of which are unique to it. The realisation that certain
fungi do immense damage to weeds while leaving crops untouched has led to
a change of perspective in plant pathology.

ÐÓ°ÉÔ­´´s in Britain and Europe are now following the example of American
researchers in developing myco herbicides from fungi that are natural pathogens
of the target weed. Development of a mycoherbicide begins in the field,
with a somewhat laborious search for specimens of the target weed showing
symptoms of infection. Back in the laboratory, the next task is to grow
the fungus, first sterilising the surface of the plant tissue, then placing
it onto a culture medium appropriate for the pathogen. If the fungus produces
spores, then it is easy to test its effectiveness as a pathogen by spraying
spores onto detached leaves or seedlings of the host weed. Tests on a range
of other plants will show how specific the fungus is to the intended host.
The screening process is designed to pinpoint those fungal isolates that
are highly host-specific, and which regularly – and rapidly – kill the target
species.

The plants chosen for this ‘specificity screen’ are not picked at random.
A pathogen that infects a particular plant is most likely to attack other
species that are closely related to it. The specificity screen is designed
to include many species from the family to which the weed belongs and a
few species from less closely related groups.

A fungus that shows promise in the preliminary tests may go on to the
next stage of development. First, its specificity to the target weed comes
under still closer scrutiny, using a progressively wider range of test plants,
including wild species as well as crops. This strategy is known as ‘centrifugal’
testing, starting with the target plant and working outward to include an
increasing number of other species.

Specificity alone is not enough to make a good mycoherbicide. The fungus
must satisfy several other criteria. It must grow and produce spores readily
in cultures in industrial conditions, preferably in a liquid medium in a
standard industrial fermenter. This ability is not as common as manufacturers
might wish, even among fungi that readily form spores in laboratory cultures
. Once the identity of the fungus is confirmed, patents must be available
to the developer – which means that no one else must have exploited that
particular fungus before. It is also important that other pesticides (fungicides,
insecticides or herbicides) applied to the crop do not harm the fungus.

All these factors are vitally important, but there is yet another hurdle
to leap before a fungus is ready for development as a weedkiller. All the
earlier tests involved parts of plants or young seedlings. In the field,
the mycoherbicide’s target would be older, much hardier plants. The best
way to ensure that the fungus has a chance to work on these tougher individuals
is to design a suitable formulation in which to apply it – one that lengthens
the life of the spores, for instance, or which increases their stickiness
so that they cling to the leaf long enough to germinate.

With so many criteria to meet, almost all natural isolates fail at one
or other obstacle. Many fall at the first fence, for while they might infect
the weed, they are not destructive enough to control it in the field. The
relatively low virulence of most isolates is no great surprise: a pathogen
that eliminates its hosts is likely to die out itself unless it can survive
in dead tissue, which most of these fungi cannot. What is more surprising
perhaps is that very damaging isolates do exist, but they do not wipe out
their hosts because they generally infect only pockets of the host population.
This highlights the key role of environmental factors in limiting the natural
spread of disease. The production of spores and their transmission between
hosts are highly dependent on suitable weather conditions (see ‘Transports
of blight’, New ÐÓ°ÉÔ­´´, 12 August 1989). These rarely allow severe natural
epidemics to develop.

The usefulness of mycoherbicides depends on the ability to remove these
constraints by inundating the weed with massive doses of spores formulated
so that they are no longer dependent on a narrow range of environmental
conditions. In the US, two mycoherbicides, Collego and Devine, have been
on the market for almost 10 years. Both are targeted at weeds that are not
very widespread: northern joint vetch, the target for Collego, infests rice
and soya bean crops in Arkansas, Louisiana and Georgia, and milkweed vine,
Devine’s target, affects citrus groves in Florida .

The success of these two fungal herbicides has paved the way for new
products aimed at more widespread weeds. One of these, Casst (Alternaria
cassiae), is aimed at two weeds: sicklepod (Cassia obtusifolia) and coffee
senna (C. occidentalis) which damage soya bean and peanut crops throughout
the southern US.

The European Weed Research Society has recently drawn up a ‘Top Ten’
of European weeds that are suitable targets for mycoherbicides. The selection
is based on the extent of the economic damage these weeds do throughout
Europe and, in some cases, because of their resistance to chemical herbicides.
The list includes such well known enemies as chickweed and cleavers. At
Long Ashton Research Station near Bristol, Mike Greaves and his colleagues
are studying two of the Top Ten, cleavers (Galium aparine) and bindweed
(Convolvulus arvensis). They have isolated a number of fungi that show promise
in controlling these two weeds. Researchers at Shell’s Research Laboratories
in Sittingbourne, Kent, are investigating other weeds on the hit list.

One problem in marketing mycoherbicides may be that they tend to produce
a less dramatic-looking and less uniform ‘kill’ than chemical herbicides.
The fungi are no less effective for this; survivors of the treatment are
generally stunted and are unable to compete with the growing crop. But effectiveness
could be further improved if more is made of the interactions between mycoherbicides
and chemical herbicides. Alan Watson, at McGill University in Quebec, Canada,
has shown that while neither the mycoherbicide Colletotrichum coccodes nor
the chemical herbicide thidiazuron totally and reliably controls velvetleaf
(a weed of corn and soya beans), a combination of the two will bring the
pest under control. If chemical and fungal weedkillers act in concert then
mixtures of the two could be a way of reducing the doses of chemical herbicides
and their residues that linger in the environment.

Attack on several fronts

Such synergism also means that some of the weaker fungal pathogens might
be suitable as weedkillers, if they are combined with a chemical. The fact
that a fungus may remain pathogenic in the presence of a chemical herbicide
also means that the fungus might be used alongside chemicals in circumstances
where a whole spectrum of unrelated weeds needs controlling.

Fungi can be combined with other fungi and with insect controls as well
as with chemicals. At Lancaster University, Steven Hallett has found that
when certain fungi are applied in pairs, they achieve together a ‘kill’
that is impossible with either fungus alone. Mycoherbicides are an additional
option in programmes of integrated pest management, particularly in partnership
with insects or other invertebrates. Water hyacinth is a good example of
a pest that succumbs to this approach. Water hyacinth is a serious weed
in watercourses, hindering navigation and eliminating native wildlife. Neither
the fungus Cercospora rodmanii nor insect controls, such as Neochetina weevils,
can curb the plant alone. But in combination the fungus and insects kill
more than 99 per cent of the weed.

Mycoherbicides have several advantages over conventional chemical herbicides.
In developing a new pesticide, an agrochemicals company screens up to 13
000 compounds, more than half of them picked at random. The search for mycoherbicides
is more directed, reducing the costs of development. This could make them
an attractive commercial option. Mycoherbicides could also be used as alternatives
to chemicals where the weed has developed resistance. At least 48 species
of weed are now resistant to the once widely-used triazine herbicides. Mycoherbicides
are also more selective than most chemicals, making them a better choice
for controlling weeds that are close relatives of the crops that they infest,
such as fat hen in crops of sugar beet (both belong to the family Chenopodiaceae).

Mycoherbicides clearly offer a way of reducing the amount of potentially
toxic chemical herbicides entering the environment. But do they themselves
pose any problems? Because the fungi are not animal pathogens they probably
pose little risk to human and animal health. Nevertheless, before any mycoherbicide
receives a licence, it is screened to rule out any hazards, such as allergic
reactions among farmers. A more serious concern is that mycoherbicides might
threaten natural vegetation. So far, however, the development of myco herbicides
has involved only fungi that occur naturally in the environment. An application
of a fungus simply increases the size of the local population of the organism.
Most importantly, unlike many chemical herbicides, mycoherbicides are not
persistent. In the absence of its host weed, the fungal population soon
diminishes.

Fungi stay true to their hosts

Genetic manipulation of fungi might be on the cards in the future, but
long experience with diseases of crops has shown that host-specific, native
pathogens do not normally ‘jump’ to new hosts. Even if native strains were
improved, so that they secreted more of the enzymes or toxins that destroy
host tissues, for example, they are likely to pose little risk. They will
simply be producing more of what they already produce in nature. If fungi
could be genetically engineered to control weeds that are not their normal
host, the position would be rather different. Even with extensive testing
of the complete host range of each fungus, the wisdom of releasing such
organisms would be as debatable as the release of microorganisms modified
for other purposes. What is certain is that increasing numbers of new mycoherbicides
based on native pathogens will appear on the market in the next few years.
What is also certain is that the first products must be beyond reproach.
Any mistakes at this early stage could damage irreparably the growing reputation
of this novel approach to weed control.

* * *

1: FIRST CHOOSE YOUR FUNGUS. . .

MOST of the fungi that are destined to become mycoherbicides are pathogens
that affect leaves and stems. This is mainly because it is easier to spray
a suspension of fungal spores onto shoots than roots. Two more fundamental
reasons for exploiting shoot pathogens is that they produce larger numbers
of spores and they are more specific to particular plants than root pathogens.

These last two characteristics probably evolved because shoot pathogens
had the opportunity to travel long distances from plant to suitable plant,
carried by air currents and water droplets.

In contrast, fungi that infect roots are dispersed only short distances
through the soil and are rarely specific to a single host – they could not
afford to be. Much of their evolutionary ‘effort’ has gone into the development
of fewer and larger ‘resting’ structures that help them to survive in the
soil until any one of a range of new potential hosts comes along. These
types of fungi seem to have much less potential as mycoherbicides.

Taxonomically, mycoherbicidal fungi are filamentous, in most cases belonging
to either the Hyphomycetes or Coelomycetes, subdivisions of the Deuteromycotina
(the imperfect fungi). They rarely, if ever, reproduce sexually.

Industrial interests argue that myco herbicides must be produced in
fermenters, which would rule out fungi such as rusts, which cannot be grown
apart from the living host. However, field trials in the US showed that
when enough spores for inoculation were raised, a rust (Puccinia canaliculata)
very effectively controlled purple nutsedge, particularly when combined
with a chemical herbicide.

We believe that biotrophic fungi, which unlike ‘necrotrophic’ fungi,
do not kill host cells with toxins or enzymes that break down the walls
of the cells, should not be excluded from consideration as potential mycoherbicides.
Biotrophs can be effective where the problem weed is very local and when
only relatively small amounts of spores are required.

At Lancaster we are assessing whether another rust, Puccinia lagenophorae,
in combination with a second fungus, can control groundsel, which is a problem
in tree nurseries and where organic vegetables are grown.

* * *

2: COLLEGO BREAKS NEW GROUND

HE mycoherbicide Collego employs the fungus Colletotrichum gloeosporiodes
aeschynomene. Developed by George Templeton in Fayetteville, Arkansas, with
support from the Upjohn Company, it was first introduced in 1982 to control
northern jointvetch (Aeschynomene virginica), whose black seeds contaminate
harvests of rice and soya beans.

Marketed as a dry formulation consisting of 15 per cent viable spores
and 85 per cent inert ingredients, Collego can be stored for long periods
at room temperatures. Each package contains 757 billion spores, which will
treat about 2.5 hectares.

Farmers simply mix the formulated dry spores with a wetting agent and
add about 250 litres of water. The farmer then sprays the suspension, which
gives about 18 million spores per square metre, from the air at a time when
the crop is well watered and relative humidity is likely to be high for
the following 12 hours.

Within a week or two the vetch plants begin to show lesions that gradually
encircle the stem. Most of the plants die within 5 weeks. A single application
of Collego in the growing season of the crop is all that is needed; with
rice this is when the weed has just emerged above the crop canopy.

A few stunted plants may survive treatment but cannot keep up in competition
with the crop.

Growers have taken enthusiastically to Collego, and the fungus has reduced
the input of pesticides by nearly 500 000 litres since its introduction.

Peter Ayres is secretary of the Biological Control Group of the European
Weed Research Society. He and Nigel Paul are based at the Institute of Environmental
and Biological Sciences, University of Lancaster.