Recent evidence confirms that transgenic DNA does jump species to bacteria and even plant and animals, as some of us had predicted.
by Dr. Mae-Wan Ho, Geneticist, Biophysicist and Director of the not-for-profit Institute of Science in Society and Prof. Joe Cummins
Genetic engineering, horizontal gene transfer, and the emergence of infectious diseases
Genetic engineering creates vast arrays of transgenic DNA that could spread, not only through cross-pollination with the same or related species, but also through the direct uptake of the transgenic DNA by cells of unrelated species, a process called horizontal gene transfer. We have been alerting regulators to Horizontal Gene Transfer - The Hidden Hazards of Genetic Engineering  on many occasions since the late 1990s [2-4] (Genetic Engineering Dream or Nightmare, ISIS Publication) when the regulators and their scientific advisors had denied vehemently that horizontal gene transfer could happen, and assumed mistakenly that transgenic DNA, like all DNA, would be rapidly degraded once out of the cell.
In a review published in 1998 , we presented extensive evidence that DNA persists in all environments and can indeed be taken up by cells of many species throughout the living world. We called for a public enquiry on the extent to which the poorly regulated discharge of transgenic organisms and transgenic nucleic acids into the environment could have been responsible for the increased emergence of new viral and bacterial diseases and antibiotic and drug resistance since genetic engineering began in the mid 1970s. Horizontal gene transfer and recombination is the main route for generating new pathogens and spreading antibiotic and drug resistance, and genetic engineering is nothing if not greatly facilitated horizontal gene transfer and recombination.
Transgenic DNA is different from natural DNA and more likely to spread
Regulators frequently dismiss our concerns with remarks such as “The safety of nucleic acids is widely accepted. Both RNA and DNA are part of all food products that we consume. “  (USDA FONSI for Transgenic Poplars Absurd & Dangerous, SiS 38). The implication is that transgenic DNA (or RNA) is no different from natural nucleic acids, and hence no more likely to spread by horizontal gene transfer.
There is no doubt that transgenic DNA is different from natural DNA; not only does it contain new combinations of genes, but also new synthetic genes that have never existed in billions of years of evolution: new coding sequences, promoters and other non-coding regulatory sequences that boost gene expression to abnormally high levels.
Furthermore, there are indeed reasons to suspect transgenic DNA is more likely to transfer horizontally and recombine than natural DNA (see Box adapted from  Living with the Fluid Genome, ISIS publication), and this has been borne out by accumulating evidence, even though dedicated research is still extremely rare.
|Transgenic DNA more likely to spread horizontally
1. Transgenic DNA is designed to jump into genomes, often through viral or bacterial plasmid vectors that can integrate into genomes.
2. Transgenic DNA tends to be structurally unstable and hence prone to break and rejoin, giving rise to numerous deletions, duplications, and other rearrangements during the transformation process, which spread into the host genome; and this is in part responsible for the instability of transgenic varieties [8, 9] (see Transgenic Lines Unstable hence Illegal and Ineligible for Protection and MON810 Genome Rearranged Again, Stability of All Transgenic Lines in Doubt, SiS 38).
3. The mechanisms that enable transgenic constructs to jump into the genome enable them to jump out again and reinsert at another site or into another genome.
4. The borders of the most commonly used vector for transgenic plants, the T-DNA of Agrobacterium, are recombination hotspots (sites that tend to break and join). In addition, a recombination hotspot is also associated with the cauliflower mosaic virus (CaMV) promoter and many transcription terminators, which means that the whole or parts of the integrated DNA will have an increased propensity for secondary horizontal gene transfer and recombination (see main text).
5. The Agrobacterium vector remaining in transgenic plants may be a vehicle for gene escape and can transfer genes widely to many bacteria as well as into human cells (see main text).
6. Transgenic constructs tend to integrate at recombination hotspots in the genome, which again, would tend to increase the chances that they will disintegrate and transfer horizontally .
7. Transgenic DNA often has other genetic signals, such as the origin of replication left over from the plasmid vector. These are also recombination hotspots, and in addition, can enable the transgenic DNA to be replicated independently as a plasmid that is readily transferred horizontally among bacteria and other cells.
8. The metabolic stress on the host organism due to the continuous over-expression of transgenes linked to aggressive promoters such as the CaMV 35S will also increase the instability of the transgenic DNA, thereby facilitating horizontal gene transfer
9. Transgenic DNA is typically a mosaic of DNA sequences copied from many different species and their genetic parasites; these homologies mean that it will be more prone to recombine with, and successfully transfer to the genomes of many species and their genetic parasites. Homologous recombination typically occurs at one thousand to one million times the frequency of non-homologous recombination, and short homologous sequences could act as anchors for acquiring non-homologous sequences (see main text).
In 1999-2000, we alerted our regulators to the CaMV 35S promoter that is in practically every commercial transgenic crop commercialised. A promoter is a signal necessary for a gene to be expressed, and the CaMV 35S promoter is used with many transgenes. It has a recombination (fragmentation) hotspot that would enhance horizontal transfer of transgenic DNA, making transgenic DNA and transgenic lines unstable [10, 11]. Furthermore, contrary to the then common assumption that the CaMV 35S promoter was only active in plants and plant-like organisms, it is in fact active in species across the living world, animal and human cells included . Consequently, it has the potential to activate dormant viruses and trigger cancer, if it happens to land next to certain cancer-related ‘proto-oncogenes’. Since then, the CaMV 35S promoter was demonstrated to be active in human enterocyte-like cells . And evidence of transgenic instability has also emerged, with the CaMV 35S promoter representing a major breakpoint [14, 15] (Transgenic Lines Proven Unstable, SiS 20) precisely as we had predicted.
Agrobacterium vector a vehicle for gene escape
We have also provided evidence strongly suggesting that the most common method of creating transgenic plants may also serve as a ready route for horizontal gene transfer [16, 17].Agrobacterium tumefaciens, the soil bacterium that causes crown gall disease, has been developed as a major gene transfer vector for making transgenic plants. Foreign genes are typically spliced into the T-DNA - part of a plasmid of A. tumefaciens called Ti (tumour-inducing) – which ends up integrated into the genome of the plant cell that subsequently develops into a tumour.
But further investigations revealed that the process whereby Agrobacterium injects T-DNA into plant cells strongly resembles conjugation, the mating process between bacterial cells.
Conjugation, mediated by certain bacterial plasmids requires a sequence called the origin of transfer (oriT) on the DNA that’s transferred. All the other functions can be supplied from unlinked sources, referred to as ‘trans-acting functions’ (or tra). Thus, ‘disabled’ plasmids, with no trans-acting functions, can nevertheless be transferred by ‘helper’ plasmids that carry genes coding for the trans-acting functions. And that’s the basis of a complicated vector system devised, involving Agrobacterium T-DNA, which has been used for creating numerous transgenic plants.
It soon transpired that the left and right borders of the T-DNA are similar to oriT, and can be replaced by it. Furthermore, the disarmed T-DNA, lacking the trans-acting functions (virulence genes that contribute to disease), can be helped by similar genes belonging to many other pathogenic bacteria. It seems that the trans-kingdom gene transfer of Agrobacterium and the conjugative systems of bacteria are both involved in transporting macromolecules, not just DNA but also protein.
That means transgenic plants created by the T-DNA vector system have a ready route for horizontal gene escape, via Agrobacterium, helped by the ordinary conjugative mechanisms of many other bacteria that cause diseases, which are present in the environment.
In fact, the possibility that Agrobacterium can serve as a vehicle for horizontal gene escape was first raised in 1997 in a study sponsored by the UK Government [18, 19], which found it extremely difficult to get rid of the Agrobacterium in the vector system after transformation. Treatment with an armoury of antibiotics and repeated subculture of the transgenic plants over 13 months failed to get rid of the bacterium. Furthermore, 12.5 percent of the Agrobacterium remaining still contained the binary vector (T-DNA and helper plasmid), and were hence fully capable of transforming other plants.
Agrobacterium not only transfers genes into plant cells; there is possibility for retrotransfer of DNA from the plant cell to Agrobacterium . High rates of gene transfer are associated with the plant root system and the germinating seed, where conjugation is most likely . There, Agrobacterium could multiply and transfer transgenic DNA to other bacteria, as well as to the next crop to be planted. These possibilities have yet to be investigated empirically.
Finally, Agrobacterium attaches to and genetically transforms several human cell lines . In stably transformed HeLa cells (a human cell line derived originally from a cancer patient), the integration of T-DNA occurred at the right border, exactly as would happen when it is transferred into a plant cell genome. This suggests that Agrobacterium transforms human cells by a mechanism similar to that which it uses for transforming plants cells.
The possibility that Agrobacterium is a vehicle for horizontal transfer of transgenic DNA remains unresolved to this day.
Evidence of horizontal transgene transfer to bacteria denied and dismissed
By 1999, there was already evidence that horizontal transfer of transgenic DNA could occur, not only in the laboratory but also in the field . Unfortunately, the researchers were far too cautious as scientists, and ended up denying the prima facie evidence that the transgenic DNA had transferred horizontally from plant to soil bacteria ; whereas, a proper application of the precautionary principle would have resulted in the researchers stressing the possibility that it had occurred could not be dismissed.
High frequencies of horizontal transfer of transgenic plant DNA were demonstrated for soil bacteria, Pseudomonas stutzeri and Acinetobacter sp. when the transgenic plant DNA contained sequence homologies to the bacteria . Again, the authors stressed that the transfer “strictly depends on homologous sequences”, which could give the uninformed a false sense of assurance, forgetting that transgenic constructs contain homologies to many different species of bacteria and viruses, and are therefore capable of engaging in high frequencies of horizontal gene transfer and recombination with all of them .
We drew attention to further evidence of the enhanced horizontal transfer of transgenic DNA in our submissions [26, 27] (Molecular Pharming by Chloroplast Transformation, GM Pharmaceuticals from Common Green Alga, SiS 27) to the regulatory authorities in Hawaii objecting to an intended outdoor large-scale facility for transgenic strains of the alga, Chlamydomonas reinhardtii producing a range of pharmaceutical proteins in chloroplast integrated transgenes, which would greatly increase copies of transgenic DNA per cell. We pointed that DNA not only persists in all environments, but also that transformation by direct uptake of DNA is a major route of horizontal gene transfer among bacteria . The close similarities (homologies) between the transformed chloroplasts in transgenic C. reinhardtii and bacterial genomes is expected to further enhance the frequency of horizontal gene transfer, by up to a billion-fold.
In fact, researchers at the University of Oldenburg in Germany demonstrated that the horizontal transfer of non-homologous DNA also occurs at relatively high frequencies when a homologous DNA ‘anchor sequence’ is present, which can be as short as 99bp . In a review published in 2004, the researchers listed at least 87 species of naturally transformable bacteria , which represent 2 percent of all known species. And transgenic DNA can spread not only via the roots and plant debris, but also via pollen drift into fields that had never cultivated transgenic crops. The authors had even developed a bio-monitoring technique for detecting transgenic DNA based on transformation of a competent strain of bacteria that depends on double cross-over event between the transgenic DNA and the bacterial chromosome, a theoretically much rarer event than a single cross-over. Nevertheless, the bio-monitoring technique is at least as sensitive as a routine polymerase chain reaction (PCR) for detecting minute amounts of DNA, indicating that horizontal transfer of transgenic DNA is not exactly a rare event. This conflicts with their conclusion in the same review that, “each of the many step involved from the release of intact DNA from a plant cell to integration into a prokaryotic genome has such a low probability that a successful transfer event [is] extremely rare.”
Researchers at Cardiff University in the UK have confirmed that horizontal transfer of transgenic DNA occurs at detectable levels using a similar system . Transgene sequences kanamycin resistance (nptII) and green flourescent protein (gfp) were driven by their own bacterial promoters. Recipient bacteria carried a copy of these two genes with deletions in their 3’ ends abolishing marker activity. Successful recombination between the plant transgene and the bacterial genome resulted in restoration of the markers, allowing detection through antibiotic selection and fluorescence. Measurable transformation frequencies were obtained in increasingly complex conditions approaching field conditions. In sterile soil microcosms, transformation was detected using pure plant DNA at 3.6 x 10-8 and in ground leaves at 2.5 x 10-11 transformants per recipient; for non-sterile soil using pure plant DNA, the frequency was 5.5 x 10-11 transformants per recipient.
Evidence has continued to accumulate  Horizontal Gene Transfer Happens - II, ISIS Report) indicating that transgenic DNA in food and feed can transfer into animal and human cells  (DNA in GM Food & Feed, SiS 23). Several studies have documented the survival of DNA in food/feed throughout the intestinal tract in mice and pigs [33, 34 and references therein], in the rumen of sheep , and in the rumen and duodenum of cattle , with varying degrees of sensitivities in PCR methods.
In the only feeding trial in human volunteers , a single meal containing GM soya with about 3 x 1012 copies of the soya genome, the complete 2 266 bp of the epsps transgene was recovered from the colostomy bag in six out of seven ileostomy subjects, though at highly variable levels, ranging from 1011 copies (3.7 percent) in one subject to 105 copies in another. This is a strong indication that DNA is not rapidly broken down in the gastrointestinal tract, confirming earlier results from the same research group. In three of the seven ileostomy subjects, about 1 to 3 per million bacteria cultured from the contents of the colostomy bag were positive for the GM soya transgene, indicating that horizontal transfer of transgenic DNA had occurred, either before the single meal was taken, as claimed, or else as the result of the single GM soya meal, a possibility that cannot be ruled out . Interestingly, no bacteria were found to have taken up non-transgenic soya DNA, suggesting that transgenic DNA may be more successfully transferred for reasons given above.
No transgenic DNA was found in the faeces of any of 12 healthy volunteers tested. Either the remaining DNA has completely broken down by then as claimed by the researchers, or else detectable fragments have all passed into the blood stream from the intestine . The researchers had not checked for the presence of transgenic DNA in the bloodstream. It is already known that food material can reach lymphocytes entering the intestinal wall directly, through Peyer’s patches. And fragments of plant DNA were detected in cow’s peripheral blood lymphocytes . From the blood, the DNA can be transported to and taken up by tissue cells, and this has been known from experiments since the late 1990s. Transgenic DNA and viral DNA fed to mice ended up in cells of several tissues , and when fed to pregnant mice, the DNA crossed the placenta and entered the cells of the foetus and the newborn . DNA from ingested food plants were also taken up into tissue cells .
In summary, the evidence shows that horizontal transfer of transgenic DNA does happen and has happened both in the soil and in the gastrointestinal tract, though many scientists are unable or unwilling to acknowledge this, or else dismiss it by saying it has a “low probability” and is “extremely rare”. But recent evidence shows it has been greatly underestimated.
Evidence that horizontal transfer of transgenic DNA has been greatly underestimatedA team of researchers led by Aurora Rizzi at the University of Milan, Italy, developed a strategy for detecting transformation in the soil bacterium Acinetobacter baylyi BD413 in situ by the expression of green fluorescent protein (gfp) . The transformed bacteria growing on plant tissues can be seen and counted directly by fluorescence microscopy. Using this method, the researchers showed that conventional methods based on cultivating and selecting the transformants on agar-plates containing antibiotics underestimates the frequencies of transformation by at least a hundred-fold.
The agar-plating step destroys the original material, so no information can be obtained on the specific location of the gene transfer events or the interaction between the bacteria and the transforming material, including the donor DNA. So it is not possible to locate the hotspots of transformation in a complex environment such as the soil, or within the plant. Furthermore, conventional plating techniques will only isolate cells that can be cultured (which is probably less than one percent of soil microbes).
The reporter strain is engineered with a green fluorescent protein (gfp) that is not expressed due to a deletion that includes its promoter, and which is present in the transgenic DNA. So when the required piece of transgenic DNA is transferred into the correct position – which it will do because the deletion is flanked by sequences homologous to the transgenic DNA – the gfp will be expressed and lights up the cells under the fluorescence microscope.
Using this technique the researchers located actual transformation events on a membrane filter, and in decaying tobacco leaves and roots. The transformation frequencies measured were at least two orders of magnitude higher than previously recorded by cultivation-based plating.
In decaying leaves, transformed cells were located in the interstices between epidermal cells on the surface of leaves and close to the stoma (pores for gas exchange) at the border with other epidermal cells. Transformed bacteria were also found on root surfaces.
Horizontal transfer to plant and animal genomes may occur at even higher frequencies
While researchers on biosafety have been focussing on horizontal gene transfer from plants to bacteria, evidence is emerging suggesting that genomes of higher plants and animals may be even softer targets for horizontal gene transfer. Transgenic DNA may well be taken up by the cells of other plants, and by animals including humans feeding on the plants. We have been warning of this possibility at least since 2001, when experiments in ‘gene therapy’- making transgenic human cells - were demonstrating how easy it was for transgenic constructs to be taken up by human and animal cells  (SLIPPING THROUGH THE REGULATORY NET: ‘Naked’ and ‘free’ nucleic acids)
Now, information from transgenic Arabidopsis and rice, with sequenced genomes, and the huge amounts of relic viruses, transposons, retroelements, and chloroplast and mitochondrial DNAs found in these and other sequenced genomes are persuading geneticists that  “nuclear genomes of plants, like those of other eukaryotes, are promiscuous in integrating nonhomologous DNA.” “Illegitimate recombination”, a rarity in prokaryotes, is the main route for transgene integration in eukaryotes. In other words, eukaryotic genomes, including the human genome, integrate foreign DNA much more readily than bacterial genomes. We have spelt out what such consequences could be : insertion mutations including cancer, activation of dormant viruses, and recombination with viral sequences in the genome to generate new viruses.
Research in recent years has also uncovered substantial amounts of DNA and RNA circulating in peripheral blood, which are actively secreted by living cells, and fully capable of transforming other cells [45, 46]. The nucleic acids appear to play a role in disease progression and metastasis of cancers. In plants too, foreign and endogenous nucleic acids circulate , apparently acting as intercellular messengers. There is a distinct possibility, therefore, that transgenic DNA could be vectored between plants through insects as well as soil bacteria; and transgenic DNA from food could end up in peripheral blood and gain entry into cells .
A fully referenced version of this report is available here