by Dr. Anil Day
University of Manchester, Unite Kingdom

GENE FLOW FROM CROPS TO PLANTS IN THE ENVIRONMENT

NEW TECHNOLOGIES TO REDUCE THE POLLEN-MEDIATED SPREAD OF TRAIT GENES

The consequences of gene flow from crops to other plants depend on the trait genes transferred to hybrids. Trait genes enhancing the nutritional quality of crops are unlikely to provide an advantage to any wild plants that might acquire them. Other traits such as herbicide resistance will be advantageous if wild plants are sprayed with herbicides to which they are resistant but a resistance gene will be ineffective against other groups of herbicides. The poor performance of hybrids provides one barrier to the escape and establishment of trait genes in the environment. Putting trait genes in plastids rather than nuclei provides an additional barrier to reduce their spread. The chloroplast, which contains the green pigment chlorophyll and carries out photosynthesis, is one of the more familiar forms of the plastid family of organelles.

Most plant genes are located in the nucleus and these are transmitted in equal proportions by sperm and egg cells. By contrast, in most flowering plants the sperm cells from pollen transmit plastids at undetectable or very low frequencies to progeny plants. Plastids are passed on to the next generation by egg cells so that plastid genes are inherited maternally. This means that when trait genes are placed in plastids the pollen route of dispersal into the environment is prevented. The potential advantages of localising trait genes in plastids for improving GM crop design was discussed in detail by a sub-group of the Advisory Committee on Releases to the Environment (ACRE).

TRAIT GENES CAN MOVE FROM PLASTIDS TO THE NUCLEUS

The frequency is low and transposition events are detected in 1 in 16,000 pollen grains. Movement of plastid genes to the nucleus was expected because insertions of plastid genes have been found in chromosomes within the nucleus but the frequency of these events was unknown. Prior to these experiments on tobacco, DNA had been shown to escape from the mitochondrion to the nucleus in bakers’ yeast (Saccharomyces cerevisiae). Mitochondria carry out respiration and are found in animal, plant and fungal cells. Like plastids, mitochondria contain their own genes. In normal yeast, escape of mitochondrial genes to the nucleus takes place at frequencies between one cell out of every hundred thousand dividing cells to one cell out of every million dividing cells [3,4]. In the yeast experimental system escaping DNA has an origin of replication that allows it to replicate as a small circular DNA molecule in the nucleus. Without this origin of replication DNA transposing to the nucleus would be lost unless it can land and accommodate (integrate) itself into one of the yeast chromosomes housed in the nucleus. If escaping DNA does not have an origin of replication, which was the case in the tobacco experiments, this added requirement to integrate into chromosomes might have been expected to further reduce successful transposition to below one in a hundred thousand dividing cells.

The estimated frequency of DNA transposition from mitochondria to the nucleus in normal yeast of one event in every hundred thousand dividing cells is almost ten-fold lower than the frequency of plastid to nucleus transposition estimated at one in 16,000 pollen grains. Clearly tobacco plastids are very different from yeast mitochondria but both systems give transposition frequencies that are within ten-fold of each other. The extra-step of integration into tobacco chromosomes might have been expected to reduce the overall transposition frequency. The relatively high frequency of transposition from the plastid to the nucleus was surprising. But there is a proviso that prevents us drawing firm conclusions on relative transposition rates. It should be stressed that this is a very crude comparison because the yeast frequencies are in events per cell division and the tobacco frequencies are in events per pollen grain. Truly meaningful comparisons will require tobacco figures that are calculated as transposition events per cell division. This will require further work on the timing of transposition, which might take place randomly during plant growth and development or be restricted to particular cell types. Estimates of transposition frequency will also depend on experimental design e.g. the type of gene that is tracked.

TRAIT GENES THAT TRANSPOSE FROM THE PLASTID TO THE NUCLEUS ARE NOT FUNCTIONAL

The experimentally obtained frequency of plastid to nuclear transfer of DNA is useful for risk evaluation. Numbers for individual steps in the pathway leading to escape and establishment of a trait gene in the environment can be combined to assess the overall cumulative risk. The figure of 1 in 16,000 pollen grains is very low but given the global acreage of a crop there is likely to be some pollen transmission of a recently transposed plastid gene. The majority of these transposed genes will not be functional because plastid regulatory elements don’t function in the nucleus. In the experiments carried out by Huang, Ayliffe and Timmis, the plastid gene that was tracked was modified to work well in the nucleus. An adjacent engineered gene that contained plastid regulatory elements was transferred to the nucleus but once integrated into nuclear DNA it became inactive in the 17 plants analysed [2]. This means that a tiny fraction of pollen (1/16,000) will contain transposed plastid genes. Most of these will be inactive. If pollen grains with these inactive transposed genes fertilize other plants the resulting hybrids will contain but not express the transposed genes. Unlike the original crop, which benefits from a functional trait gene in the plastid the hybrids will not gain any advantage from defective nuclear-localised trait genes.

PLASTID-LOCALISED TRAIT GENES ARE MORE LIKELY TO BE DISPERSED BY AN EGG ROUTE THROUGH SEEDS THAN A POLLEN-MEDIATED ROUTE INVOLVING PLASTID TO NUCLEAR TRANSPOSITION

The frequencies at which transposed genes could be activated, due to chance chromosomal integration events in the nucleus, is likely to be substantially less than one in 17. If as many as one in a hundred transposed genes were functional about one in a million pollen grains would contain a functional transposed gene. Given this low frequency, pollen dispersal of a plastid transgene into the environment via this transposition route appears unlikely. Seed transmission of a plastid localised trait gene where crop plants act as female parents is a far more likely dispersal mechanism. However, to further reduce this remote risk of spread through pollen following transposition it might be wise to alter the coding sequences of a plastid gene to prevent its expression in the nucleus.

Putting trait genes into plastids is a very effective method for dramatically reducing their transmission through pollen. Transposition of plastid DNA to the nucleus has far greater consequences for our understanding of the evolution of eukaryotic cells than it does on undermining the containment benefits of localising genes to plastids.

REFERENCES