|
Genetically modified organisms (GMOs): The significance
of gene flow through pollen transfer, European Environment Agency,
2002
Executive Summary
In 2000 the EEA established a special project for the European
Parliament, on the dissemination of research results from technologies
characterised by scientific complexity and uncertainty, such as
GMOs and chemicals, and on the use of such results by the public
and their representatives in their governance, including the use
of the precautionary principle. This project is in support of the
EEA duty, added to its regulation in 1999, to `assist the Commission
in the diffusion of information on the results of relevant environmental
research'. In order to access European scientific expertise and
to minimise duplication, the EEA established a partnership with
the European Science Foundation to bring together relevant scientific
evidence. This is the first report from the project. Other reports
will summarise monitoring programmes and exposure data for some
representative chemicals,and the use of consensus conferences and
other methods for involving the public in complex scientific issues.
The project will support the EEA in its work of helping to develop
appropriate monitoring and data sources on the impacts of complex
economic/ environment interactions.
The European Science Foundation (ESF) had already established a
research programme, `Assessing the Impact of GM Plants' in 1999.
This AIGM programme brings together researchers and other scientists
from 10 European countries involved in assessing the environmental
and agronomic impact of GM crops, including studies of gene flow
and dispersal through pollen, hybridisation and gene introgression.
The AIGM programme was invited by the ESF to produce a review of
pollen mediated transgene flow based on recent research by participants
in the AIGM programme as well as from published reports and papers.
(The AIGM programme is briefly described in the appendix).
This report considers the significance of pollen-mediated gene
flow from six major crop types that have been genetically modified
and are close to commercial release in the European Union. Oilseed
rape, sugar beet, potatoes, maize, wheat and barley are reviewed
in detail using recent and current research findings to assess their
potential environmental and agronomic impacts. There is also a short
review on the current status of GM fruit crops in Europe. Each crop
type considered has its own distinctive characteristics of pollen
production, dispersal and potential outcrossing, giving varying
levels of gene flow.
Oilseed rape can be described as a high-risk crop for crop-to-crop
gene flow and from crop to wild relatives. At the farm scale low
levels of gene flow will occur at long distances and thus complete
genetic isolation will be difficult to maintain. This particularly
applies to varieties and lines containing male sterile components,
which will outcross with neighbouring fully fertile GM oilseed rape
at higher frequencies and at greater distances than traditional
varieties. Gene stacking in B. napus has been observed in crops
and it is predicted that plants carrying multiple resistance genes
will become common post-GM release and consequently GM volunteers
may require different herbicide management. Oilseed rape is cross-compatible
with a number of wild relatives and thus the likelihood of gene
flow to these species is high.
Sugar beet can be described as medium to high risk for gene flow
from crop to crop and from crop to wild relatives. Pollen from sugar
beet has been recorded at distances of more than 1 km at relatively
high frequencies. Cross-pollination in root crops is not usually
considered an issue since the crop is harvested before flowering.
However a small proportion of plants in a crop will bolt and transgene
movement between crops may occur in this way. Hybridisation and
introgression between cultivated beet and wild sea beet has been
shown to occur.
Potatoes can be described as a low risk crop for gene flow from
crop to crop and from crop to wild relatives. Cross-pollination
between production crops is not usually considered an issue since
the harvested tuber is not affected by incoming pollen. In true
seed production areas, however, the likelihood of cross-pollination
between adjacent crops leading to contamination is higher. The risk
of gene flow exists if volunteers are allowed to persist in afield
from one crop to the next. Naturally occurring hybridisation and
introgression between potato and its related wild species in Europe
is unlikely.
Maize can be described as a medium to high-risk crop for gene
flow from crop to crop. Evidence suggests that GM maize plants would
cross-pollinate non-GM maize plants up to and beyond their recommended
isolation distance of 200 m. There are no known wild relatives in
Europe with which maize can hybridise.
Wheat can be described as a low risk crop for gene flow from crop
to crop and from crop to wild relatives. Cross-pollination under
field conditions normally involves less than 2 % of all florets
so any outcrossing usually occurs with adjacent plants. Hybrids
formed between wheat and several wild barley and grass species generally
appear to be restricted to the first generation with little evidence
for subsequent introgression due to sterility.
Barley can be described as a low risk crop for gene flow from crop
to crop and from crop to wild relatives. Barley reproduces almost
entirely by self-fertilisation, producing small amounts of pollen
so that most outcrossing occurs between closely adjacent plants.
There are no records of naturally occurring hybrids between barley
and any wild relatives in Europe.
Some fruit crops, such as strawberry, apple, grapevine and plum
have outcrossing and hybridisation tendencies which suggest that
gene flow from GM crops to other crops and to wild relatives is
likely to occur. For raspberry, blackberry and blackcurrant the
likelihood of gene flow is less easy to predict, partly due to lack
of available information.
At present none of these crops has pollen which can be completely
contained. This means that the movement of seed and pollen will
have to be measured and managed much more in the future. Management
systems such as spatial and temporal isolation can be used to minimise
direct gene flow between crops, and to minimise seed bank and volunteer
populations. The use of isolation zones, crop barrier rows and other
vegetation barriers between pollen source and recipient crops can
reduce pollen dispersal, although changing weather and environmental
conditions mean that some long distance pollen dispersal will occur.
Biological containment measures are being developed that require
research in order to determine whether plant reproduction can be
controlled to inhibit gene flow through pollen and/ or seed.
The possible implications of hybridisation and introgression between
crops and wild plant species are so far unclear because it is difficult
to predict how the genetically engineered genes will be expressed
in a related wild species. The fitness of wild plant species containing
introgressed genes from a GM crop will depend on many factors involving
both the genes introgressed and the recipient ecosystem. While it
is important to determine frequencies of hybridisation between crops
and wild relatives, it is more important to determine whether genes
will be introgressed into wild populations and establish at levels
which will have a significant ecological impact.
Evaluations and conclusions
Oilseed rape
The majority of pollen is deposited at very short distances from
the pollen source. Pollen can travel considerable distances by means
of both wind and insects. Low frequencies of cross-pollination have
been recorded at distances of up to 4 km from the source.
While pollen is important in the spatial dispersal of transgenes
from oilseed rape, it has a short life-span and provides little
temporal dispersal. Seed is also very important in the spatial dispersal
of transgenes through dispersal along transport corridors within
and between countries. It also allows GM plants to persist at sites
for several years.
On a farm-scale the current recommended isolation distance of 100
m will maintain cross-pollination levels at below 0.5 % in the majority
of fully fertile crops.
Varieties and lines containing male sterile components will outcross
with neighbouring fully fertile GM oilseed rape at higher frequencies
and at greater distances than was previously thought. Varietal associations
will require considerably greater isolation distances from GM crops
than conventional varieties.
Gene flow will occur to and from volunteer and feral populations
which can act as gene pools carrying over the contamination into
subsequent rape crops.
Gene stacking in volunteers has been observed in GM crops. It is
predicted that plants carrying multiple resistances will become
common once GM herbicide tolerant rape is widely commercialised.
Volunteers may become more difficult to control with herbicide treatments
in certain situations, though the current range of selective herbicides
used in cereal crops is effective in controlling single and multiple
tolerant volunteers.
The risk of hybridisation between oilseed rape and some wild relatives,
particularly B. rapa, B. juncea, B. adpressa, B. oleracea, Raphanus
raphanistrum and Hirschfeldia incana is high. Long term introgression
of transgenes into some of these Brassica species is likely to occur
though the rate and level of introgression will be determined by
the enhanced fitness conferred by the transgene. The creation of
a herbicide tolerant, competent weed is possible. Gene introgression
into other hybridising related species is unlikely since backcross
plants fail to persist due to cytoplasmic incompatibility.
Oilseed rape can be described as a high risk crop for pollen mediated
gene flow from crop to crop and from crop to wild relatives.
Sugar beet
Pollen from sugar beet seed crops is primarily wind dispersed
and has been recorded at distances of more than 1 km at relatively
high frequencies. Appropriate atmospheric conditions combined with
peak pollen release times can account for longer distance dispersal.
Sugar beet flowers are visited by a range of pollinating insect
species.
The current recommended isolation distance for GM beet seed production
of 1000 m may not guarantee the prevention of seed contamination
in the long term.
While pollen is important in the spatial dispersal of transgenes
from sugar beet, it has a short life-span and provides little temporal
dispersal. Seed is very important in the spatial dispersal of transgenes
through dispersal along transport corridors within and between countries.
Also, seed can survive in soil from one beet crop to another causing
contamination of subsequent crops.
Cross-pollination in root crops is not usually considered an issue
since the crop is harvested before flowering. However a small proportion
of plants in a crop will bolt and transgene movement between crops
may occur in this way. GM bolters occurring in a following crop
of conventional beet may pollinate bolters in the current crop and
be taken up with the crop at harvest, causing contamination.
Hybridisation between bolting GM beet and weed beet could lead
to the transfer of GM traits, in which case GM weed beet may become
more difficult to control with chemical treatments.
Hybridisation and introgression between cultivated beet and wild
sea beet has been shown to occur. GM traits could therefore in trogress
into wild beets.
In one population of wild sea beet gene flow from cultivated beet
did not lead to the erosion of genetic diversity of that particular
population. In some cases crop-to-wild gene flow will have limited
evolutionary effect on wild populations. However, certain transgenes
maybe more likely to alter the fitness of hybrid or in trogressed
individuals and change niche relationships between populations.
Sugar beet can be described as a medium to high risk crop for pollen
mediated gene flow from crop to crop (especially seed crops) and
from crop to wild relatives.
Potato
Wind is considered a more important vector than insects in
effecting cross-pollination, though pollen dispersal is generally
very restricted due to high self-fertility. Isolation distances
of 20 m were recommended for experimental releases.
High frequencies of outcrossing at distances up to 1 km may have
been shown to occur in one instance where pollination was by pollen
beetle. However this research was criticised
for having a significant proportion of false positives.
Cross-pollination between GM and non-GM production crops would
not result in the harvested potato tubers becoming transgenic. Furthermore,
the crop is usually sown with seed tubers rather than true seed.
In this situation the introgression of transgenes into non-GM crops
or true seed crops nearby is unlikely.
If GM volunteer tubers, plants and true seed are allowed to persist
after a crop the risk of introduction of GM volunteers into following
conventional crops exists.
In true seed production crops the likelihood of cross-pollination
leading to contamination of neighbouring and subsequent crops is
higher unless effective isolation and crop hygiene is carried out.
Feral plants present little or no risk of acting as either a GM
pollen source or recipient, though research should continue to determine
whether increased feralisation is likely in future GM varieties.
Naturally occurring hybridisation and
in trogression between potato and its related wild species in Europe
is unlikely. Evidence suggests that even if cross-pollination occurred,
post-zygotic barriers would prevent the formation of a viable hybrid.
Potato can be described as a low risk crop for pollen-mediated
gene flowfrom crop to crop and from crop to wild relatives.
Maize
Maize is primarily wind pollinated although there is evidence
to suggest that bees and other insects collect pollen from maize.
The majority of airborne pollen is shown to fall within a short
distance of the pollen source, though outcrossing has been recorded
at up to 800 m. It is predicted that under suitable atmospheric
conditions maize pollen has the potential to travel over much longer
distances.
Incoming maize pollen is rapidly diluted by local pollen so that
cross-pollination occurs mostly in the first few rows. The recommended
isolation distance of 200 m will maintain crop purity at 99 % in
most cases. There is no evidence that any current variety is not
interfertile with another variety, for example cross-pollination
data between maize and sweetcorn exists.
There is no indication that hybridisation between maize and other
European crop species can occur.
Maize has poor survival characteristics as a feral plant in much
of Europe. The crop is incapable of sustained reproduction outside
cultivated areas and is non-invasive of natural habitats.
There are no known wild species in Europe with which maize can
hybridise. Maize can be described as a medium to high risk crop
for pollen mediated gene flow from crop to crop, but low risk for
gene flow to wild species.
Wheat
Wheat is typically self-pollinated and produces small amounts of
pollen with a short viability period. Cross-pollination under field
conditions normally involves less than 2 % of all florets. The combination
of these factors means that any outcrossing normallyoccurs with
adjacent plants.
In some cases strong winds can disperse pollen widely. Cross-pollination
has been recorded at distances of 20 m from the source.
There are no records of naturally occurring hybrids between wheat
and any crop relatives. Hybrids formed between wheat and several
wild barley and grass species generally appear to be restricted
to the Fl generation with little evidence for subsequent introgression
due to sterility.
Wheat can be described as a low risk crop for pollen mediated gene
flow from GM crops to other crops and to wild relatives. However,
GM wheat grown in rotation with conventional wheat could cause some
contamination of the latter if volunteers are allowed to persist.
Barley
Barley reproduces almost entirelyby self-fertiliisation Small
amounts of pollen are produced and the bulk of outcrosses occur
between closely adjacent plants.
Rare cross-pollination events are known to occur at distances up
to 60 m from the source. However, a crop isolation distance of lm
is deemed sufficient in maintaining seed contamination within acceptable
levels for most materials.
Strong hybridisation barriers exist between Hordeum species. There
are no records of naturally occurring hybrids between barley and
any wild relatives in Europe.
Barley can be described as a low risk crop for pollen mediated
gene flow from crop to crop and from crop to wild relatives.
However, GM barley grown in rotation with conventional barley could
cause some contamination of the latter if volunteers are allowed
to persist.
Fruit crops
Strawberry has around five wild relatives distributed throughout
Europe. Hybridisation may occur but hybrid viability appears to
be limited. On the basis of present research transgenic strawberries
are expected to have minimal impacts by gene introgression on wild
flora.
Apple is a heterozygous crop with strong selfincompatibility tendencies.
Hybridisation between apple crops and between apple and some wild
species is possible, though it is difficult to predict how widespread
gene flow and introgression of GM apple crops into wild species
maybe.
Grapevine has very few related wild species with which it could
hybridise, although hybridisation does occur. It is possible that
some gene flow from GM grapevine varieties to conventional varieties
and to wild species will occur.
Plum is a complex species with several subspecies or varieties
being recognised. There are 21 species of Prunus recorded in Europe.
Cultivated plum is a frequent escape and therefore there is a high
likelihood of gene flow from GM varieties to wild types occurring.
The European species of Rubus are placed in 5 sub-genera containing
some 2000 species. Hybridisation events in wild and feral populations
are numerous, but gene flow from cultivated blackberries and raspberries
to wild populations does not seem to occur to any significant degree.
GM Rubus varieties should be monitored prior to release to determine
whether outcrossing to wild species would be more likely to occur.
Cultivated blackcurrant has numerous wild relatives scattered across
parts of Europe, though there are no records of hybridisations.
Gene flow from GM blackcurrants to wild species is unlikelybut cannot
be ruled out.
There is limited information on crop to crop gene flow for the
fruit crops and therefore definite conclusions cannot be made at
present. However, there is some likelihood of gene flow from GM
crops of strawberry, apple, grapevine and plum to other crops occurring.
Crop to crop gene flow in blackberry, raspberry and blackcurrant
is more difficult to predict although the reproductive characteristics
of these species make it a possibility.
Future considerations
and recommendations
Gene flow: Crop to crop
At farm and regional scale gene flow can occur over long distances
and therefore complete genetic purity will be difficult to maintain
within the official isolation distances, for crops such as oilseed
rape, maize and sugar beet. Here we present some recommendations
for farm practice to minimise crop contamination.
i) The current isolation distances should be reviewed for some
crop types and further stringency applied in order for levels of
gene flow, albeit low, to be further reduced. This especially applies
to seed production crops because any genetic impurity will then
be present throughout the life cycle of the standard crop grown
from the contaminated seed, and maybe multiplied to higher levels.
ii) It is now apparent that varieties and lines containing male
sterile components will outcross with neighbouring fully fertile
GM varieties at higher frequencies and at greater distances than
previously thought. Therefore varietal associations such as Synergy
will require considerably greater isolation distances from GM crops
than conventional varieties.
iii) As well as isolation distances, barrier crops could be used
as standard where they are thought to be effective in reducing cross-pollination
levels (see section 10.4.2) and where genetic purity is most essential
(e.g. seed production crops).
iv) Neighbouring farms should inform each other of their planting
intentions in order for appropriate isolation measures to be considered.
v) Gene flow can occur to and from volunteer and feral populations
which act as gene pools carrying over the contamination into subsequent
crops. Management systems should be used to minimise GM seed spread
on a farm and to minimise seed bank and volunteer populations. Allowing
GM volunteer populations to discharge viable seed will cause a large
increase in the burden for following crops (Harding & Harris,
1994) through gene exchange from volunteers to crops, and the possibility
that GM volunteer plants could be harvested with the crop and passed
on to the consumer.
vi) As well as removing any volunteers from a field that has previously
been cropped with a GM type, when sowing conventional types volunteer
contamination can be prevented by taking into account whether GM
crops were grown previously in a field and whether farm practices
were likely to have moved seed to that field from other fields.
vii) The development of GM plants which incorporate biological
methods to restrict the spread of transgenes between crops should
be encouraged (see section 10.3).
Gene flow: Crop to wild relatives
It has been recognised that over time even small amounts
of gene flow can have important effects on evolutionary change (Wright,
1931). Gene flow between crops and their related wild species may
have two potentially harmful consequences: the evolution of increased
invasiveness and persistence and the increased likelihood of extinction
of wild relatives. It is difficult to predict, however, the precise
limits of sexual barriers between individual crop types and their
related species, or the likelihood of hybrids forming and persisting
in agricultural or natural habitats. There are several areas in
which we need to become better informed:
i) The current levels of hybridisation and introgression occurring
between conventional crops and wild species, and the behaviour of
these hybrids. This will determine the factors influencing the extent
of gene flow and the likelihood of transgenes becoming established
in wild populations (Dale, 1992). It will also provide baseline
data against which to assess the possible impacts of transgenes.
Ii) The geographical distributions of GM crop types and any wild
plant species with which the crop is capable of hybridising.
Iii) The fate and consequences of transferred genes in different
species in order to improve understanding of the genetic
and ecological principles involved.
iv) The stability of transgene expression over generations and
in different genetic backgrounds, to determine the extent to which
transgene action and stability can be modified by genetic background,
particularly in taxonomically wide hybrids (Dale, 1994).
v) Test protocols to determine the likely effect of a transgene
in a hybrid, so that on release of a GM crop the site can be surveyed
for wild relatives and a risk assessment undertaken on a case-by-case
basis (until we gain a better understanding of the above points).
Gene flow barriers
As well as identifying the agronomic and environmental risks associated
with the release of GM crops, of a primary concern is the development
of methods to restrict the spread of introduced genes to other crops
and to wild plant populations. Developers of transgenic crops also
want to limit gene escape so that competing companies cannot acquire
unique genetic constructs through pollen dispersal. Here we give
an overview of the various biological and physical barriers to gene
flow that are being researched and developed for possible future
use.
Biological gene flow barriers
A consideration for minimising crop to crop gene flow and
environmental exposure to transgenes is to design and construct
GM plants with improved biosafety characters. This could be achieved,
for example, by preventing or minimising cross-pollination, avoiding
antibiotic resistance marker genes, or switching on inserted genes
only when and where they are needed in the plant. There are three
ways in which reduced exposure of transgenes to the environment
might be accomplished:
i) Avoid or minimise inclusion of superfluous transgenes or sequences
ii) Avoid or minimise superfluous expression of the transgene
iii) Avoid or minimise the dispersal of transgenes in the environment
The emerging technologies for each approach are discussed in more
detail in the discussion paper `Guidance on Best Practice in the
Design of GM Crops' (DETR/ ACRE, 2000). Here we look at (iii) the
methods to avoid or minimise the dispersal of transgenes in the
environment. They include:
Apomixis
Apomixis is the production of seeds without fertilisation, a process
that occurs naturally in many plant species. Transfer of the primary
transgene to neighbouring crops via pollen would be minimal because
plants can be male sterile without compromising seed or fruit production.
Cleistogamy
Cleistogamy occurs naturally in some plant species, a process whereby
self pollination and fertilisation occurs with the flower remaining
unopened so pollen is unlikely to escape from the flower. The adoption
of this process to minimise transgene dispersal would require modifications
to flower design.
Hybridisation barriers
Interspecific hybridisation only occurs between closely related
plant species. Hybridisation between more widely diverged species
is prevented by two main barriers; in terspecific incompatibility
at the stigma surface or within the style which prevents fertilisation,
and post-fertilisation barriers that cause seed abortion. Strengthening
either barrier would potentially prevent hybridisation.
Inhibition of flowering to block floral
development
In recent years the molecular basis of the processes that
control flowering has been determined. Such studies open up the
possibility of manipulating flowering time control genes and blocking
or promoting flowering in a range of species.
Genetically engineered male sterility so
that a plant produces infertile anthers
Pollen development can be prevented by destroying the tapetum of
a developing anther using non-specific nucleases driven by cell-specific
promoters. Nuclease inhibitors can be crossed in to restore pollen
fertility. Recently, several promoters have been developed that
are induced by the application of exogenous chemicals. Such promoters
could be used to control flowering or fertility `restorer genes'
when required. Male sterile flowers can still be pollinated by exogenous
pollen and produce viable seeds.
Seed sterility
This technology enables crops to be genetically modified
so that they produce seed that is incapable of germination, offering
a promising technique for genetic isolation. This means, however,
that the seed cannot be saved and replanted the next season. At
present seed sterility has not been adopted because several aspects
of the technology are unreliable and require further development.
Plastid transformation technology
Daniell et al (1998) have obtained high levels of transgene expression
by inserting herbicide tolerance genes into the tobacco chloroplast
genome. An advantage of this technology is that integration of foreign
DNA into chloroplast DNA can be more precise. Also, chloroplast
transformation technology may limit transgene dispersal through
pollen in crops because chloroplasts are predominantly maternally
inherited in most higher plant species, though there may be some
paternal transfer, and this would have to be examined in each risk
assessment.
Physical gene flow barriers
Isolation zones
An isolation zone is an area between a GM crop and a nearby non-GM
crop that is either de-vegetated (a `barren zone') or planted with
a non insect pollinated crop that would discourage insect pollinators
from leaving the GM crop. Recent research on the effects of isolation
zones, and to what extent increasing the width of the isolation
zone reduces gene flow shows varying results. In experiments with
oilseed rape field trials Morris et al (1994) found that barren
zones less than 8m in width actually increased gene exchange above
the amount observed at comparable distances in continuous crops
of oilseed rape. The barren zone appeared to `reset' the zero point
of the gene dispersal profile to the end of the barren zone. The
same type of effect has also been encountered in maize trials. Because
most outcrossing will occur in the first few rows of the crop nearest
to the pollen source, this in effect means that the border rows
act as `buffers' to the dispersal of contaminating pollen in the
rest of the crop.
The impact of isolation zones on rates of gene flow in insect pollinated
plants is ultimately dependent on their influence on the behaviour
of insect pollinators, and this will vary between crop types and
sites, and with different weather conditions. Further research may
establish how widely these parameters are likely to vary between
sites and whether or not standard isolation zones could be applied
to GM crops. Research must also consider how wide an isolation zone
must be before it deters insects from moving from one crop to another,
and whether a valid option would be to discard the first few outer
rows of the recipient crop as `buffer' rows.
Barrier crops
A barrier crop is a border of non-GM plants of the same crop surrounding
the GM variety that can act as an `absorber' of the GM pollen. The
barrier rows are then destroyed after flowering. Barrier crops appear
to function in a number of ways, primarily in producing masses of
pollen to dilute pollen being introduced from adjacent fields. The
barrier also increases the distance the pollen must travel from
a source to a receptor crop, and foraging insects are likely to
visit the barrier crop at the edge of the field before moving on
to a potential receptor crop. Similarly, a barrier crop around a
receptor crop would mean that insects are likely to make their first
visit in the field to a plant in the barrier.
In their experiments Morris et al (1994) found that barrier crops
had a significant influence on gene escape. The authors suggest
that if a small area of 4 to 8m is only available for containment
methods, the most effective strategy would be to plant the entire
area with a trap crop that could be destroyed before seed set. Many
experimental releases of GM crops in the UK and Europe have included
barrier crops to restrict GM pollen flow from the release site.
Barrier crops are also discussed in section 3.3.4 (sugar beet) and
section 5.5.4 (maize).
|
