Truth: GM genes can escape into the environment by horizontal gene transfer with potentially serious consequences

Myth at a glance

Horizontal gene transfer (HGT) is the movement of genetic material between unrelated species through a mechanism other than reproduction.

It is often claimed that HGT from GM crops into bacteria, animals, or humans is unlikely or of no consequence. But independent scientists have warned that GM genes could escape from GM crops into other organisms through HGT.

HGT from plants into other plants or animals does appear to be a low-frequency event.

However, the routes of HGT that are most likely to occur are DNA uptake by bacteria in the environment or the digestive tract. There is good evidence that the latter has already happened in the intestinal bacteria of humans who eat GM soy.

Other scenarios involving HGT by the pathogenic bacterium A. tumefaciens or by viruses are less probable. But given the wide distribution of GM crops and their intended use over decades, even low probabilities translate into a high likelihood that HGT events will occur. It is just a matter of time.

The negative impacts and risks associated with HGT must be taken into account in considering the overall biosafety of any GM crop.

Most GM contamination incidents occur through cross-pollination, contamination of seed stocks, or failure to segregate GM from non-GM crops after harvest. But for years, scientists have warned that GM genes could also escape from GM crops into other organisms through what is known as horizontal gene transfer (HGT). HGT is the movement of genetic material between individuals through a mechanism other than reproduction. Those individuals could be of the same or a different species. Reproduction, in contrast, is known as vertical gene transfer because the genes are passed down through the generations from parent to offspring within a species or closely related species.

Based on very limited experimental data, HGT from plants into bacteria or multicellular organisms (plants, animals, or fungi) is believed to be rare, although HGT is acknowledged to occur frequently between different species of bacteria and more rarely between higher species by certain mechanisms. The EU-supported website GMO Compass states that HGT from plants to bacteria “can only be demonstrated under optimized laboratory conditions.”1

Gijs Kleter, a member of the European Food Safety Authority’s (EFSA) GMO Panel and for some years an affiliate of the GMO industry-funded group ILSI,2 is among those who have argued that if HGT occurs from commercialized GM plants into gut bacteria, this is unlikely to pose a risk to health.3

There are several mechanisms through which HGT can occur, some of which are more likely than others. HGT via some of these mechanisms occurs easily and frequently in nature. The consequences of HGT from GM crops are potentially serious, yet have not been adequately taken into account by regulators.

The basic mechanisms by which HGT could occur are:

  • Uptake of GM DNA by bacteria
  • Uptake of GM DNA from the digestive tract into the tissues of the organism
  • Transmission of GM DNA via pathogenic bacteria, such as Agrobacterium tumefaciens. The capacity of A. tumefaciens to introduce foreign DNA into plants is often exploited by genetic engineers to introduce GM genes into plants
  • Gene transfer by viruses.

The following sections outline these mechanisms and provide a perspective on the frequency at which these events can occur, as well as their potential impacts.

DNA uptake by bacteria

Bacteria are promiscuous. They are always exchanging DNA and taking up DNA from their environment. Some of this environmentally acquired DNA can be incorporated into their genome and may be expressed. There are two scenarios in which DNA uptake by bacteria could result in HGT of GM genes.

The first scenario is the transfer of GM DNA from GM food into intestinal bacteria. DNA from a GM plant is released into the intestinal tract of the consumer during digestion. Contrary to frequent claims, GM DNA is not always broken down in digestion and can survive in sufficiently large fragments to contain intact genes that are potentially biologically active (see Myths 3.6, 3.10).

Bacteria of many different species are present in the digestive tract, some of which can take up DNA from their environment and incorporate it into their own DNA. In the case of GMOs, this could be problematic. For example, if the GM plant contained a gene for antibiotic resistance, the bacterium could incorporate that antibiotic resistance gene into its genome and thereby become resistant to the antibiotic. If the bacteria in question happened to be pathogenic (disease-causing), this process would create an antibiotic-resistant pathogen – a “superbug”.

The transfer of GM genes from food to intestinal bacteria has been documented in a study on humans. The study found that the intestinal bacteria of a person whose diet included soy carried sequences unique to the GM soy that was part of their diet.4

The second scenario in which DNA uptake by bacteria could result in HGT of GM genes is the transfer of GM DNA to soil and aquatic bacteria. Cultivation of transgenic crops leads to the degradation of GM plant material in the environment, liberating GM genes into soil and bodies of water. Every cubic centimetre of soil contains thousands of different species of bacteria, only a small percentage of which have been identified and characterized. Bacteria are abundant in bodies of water, as well. Some soil bacteria are known to take up free DNA that may be present in the soil, incorporating the DNA into their genomes.5 This could result in the transfer of GM genes to natural soil bacterial populations. Based on limited currently available data, this type of event is thought to be extremely rare.6 However, it has been shown that GM DNA can persist in soil at detectable levels for at least a year,7 increasing the likelihood of HGT.

We only know the identities and characteristics of a small fraction of the soil bacteria that could potentially take up GM DNA from their environment.5 Furthermore, if the uptake of a GM gene, for example for antibiotic resistance, were to give the bacterium a survival or growth advantage, this would allow it to outcompete other bacterial strains in the presence of antibiotics widely used in agriculture and medicine. Therefore, this initial rare event could escalate and result in significant environmental and health outcomes.8

DNA uptake during digestion of GM foods

A study in mice demonstrated that foreign DNA present in food can be transferred from the digestive tract to the bloodstream of animals that eat the food. This foreign DNA was also found in white blood cells and in the cells of many other tissues of the mice.9 In another investigation, foreign DNA in a diet fed to pregnant mice was found in the organs of their foetuses and newborn offspring. The foreign DNA was believed to have reached the foetus through the placenta.10

It has also been shown that GM DNA in animal feed can be taken up in the organs of the animals that eat it and can be detected in the meat and fish that people eat.11,12,13,14,15
Most of the GM DNA in food is fragmented before it reaches the blood or tissues, so any genes present would not be able to express and reprogramme the host organism’s cells. However, a few copies of GM DNA large enough to contain the sequence of a full and functional gene are likely to be present in the digestive tract and can be taken up into the blood at low frequency. A study in humans (not involving GM foods) showed that meal-derived DNA fragments large enough to carry complete genes entered the circulatory system. In some of the human blood samples studied the relative concentration of plant DNA was higher than the human DNA. The researchers were even able to identify individual plant varieties eaten by the human subjects from the DNA sequences present in the blood.16

Once the GM DNA, potentially carrying genes, is in the blood, it can then be transported to the cells of the body’s tissues or organs.9 When taken up by a cell, a GM gene could be integrated into the DNA of the cell, causing either direct mutation of a host gene function or reprogramming the host cell to produce the protein for which that GM gene codes, or both.

At present, this scenario is speculative. Although it is possible to detect GM DNA in the tissues of animals that consume GM feed, no research has been published that shows that the GM DNA is integrated and expressed in the tissues of those organisms. Neither has it been shown that the relatively small amount of GM DNA in the GM gene unit is in itself more dangerous than the large quantities of non-GM DNA found in the tissues. While toxic effects have been found from feeding GM diets to animals, it is not likely that GM DNA in itself is the culprit. The culprit is far more likely to be the novel proteins and downstream small molecule metabolites produced by the GM DNA and the overall GM transformation process, and/or the pesticides engineered into or applied to the GM crop.

If expression of the GM DNA in the tissues of animals and humans that eat GM foods did occur, it would most likely not occur frequently. In order to find out whether such expression events do occur, it would be necessary to conduct very large-scale studies – though finding a suitable experimental design would be challenging. Although such events may be infrequent, the widespread and long-term consumption of GMOs by humans and animals could mean that even infrequent events have important biosafety consequences.

Horizontal gene transfer by Agrobacterium tumefaciens

Agrobacterium tumefaciens (A. tumefaciens) is a pathogenic soil bacterium often used to introduce GM genes into plants.

The introduction of GM genes into plants by infection with A. tumefaciens is carried out by exploiting a Ti plasmid – a small circular molecule of DNA that is naturally found in A. tumefaciens. When A. tumefaciens infects a plant, the Ti plasmid is introduced into the plant cells. Parts of the Ti plasmid may then insert themselves into the DNA of the plant and result in plant tumours, called crown gall.

Genetic engineers have adapted this natural but pathogenic process in order to introduce foreign DNA into plants and thereby produce GM crops. First, the naturally occurring genes of the Ti plasmid in the region that can insert into host plant cell DNA are removed and replaced with the GM gene of choice. The now genetically modified Ti plasmid is then introduced into A. tumefaciens, which in turn is used to infect plant cells. Once inside the plant cell, some of the genetically modified Ti plasmid can insert into the host plant cells’ DNA, thereby permanently altering the genetic makeup of the infected cells.

Although A. tumefaciens is a convenient way of introducing new genes into plants, it can also serve as a vehicle for HGT from the GM plant to other species. This can happen via two mechanisms.

First, residual A. tumefaciens carried in a GM plant could infect plants of other species, thereby carrying the GM gene(s) from the intentionally genetically modified plant into other plants. A. tumefaciens can serve as a vehicle for HGT to hundreds of species of plants, since it can infect a wide range of plant species.

The second mechanism creates the risk that A. tumefaciens could pass GM genes on to an even wider range of species, including, but not limited to, plants. It consists of certain types of fungi functioning as intermediate hosts in the transfer of transgenes from GM A. tumefaciens to other organisms.

A study (Knight and colleagues, 2010) found that under conditions found in nature, A. tumefaciens introduced DNA into a species of disease-causing fungi that is known to infect plants. The study also found that GM DNA sequences in the A. tumefaciens were incorporated into the DNA of the fungi. In other words, the A. tumefaciens was genetically engineering the fungi.17

The authors concluded that in cases where a GM plant is infected with fungi, A. tumefaciens in the GM plant could infect the fungi, introducing GM genes into the fungi.17 Many fungi have a wide host range and could therefore pass the GM genes onto a range of other plants.

Genetic engineers had previously assumed that A. tumefaciens only infects plants. But this study showed that it can infect fungi, a different class of organism. The study stated, “A. tumefaciens may be able to [genetically] transform non-plant organisms such as fungi in nature, the implications of which are unknown.”17 The authors pointed out17 that A. tumefaciens is already known to genetically modify human cells in the laboratory.18

One of the study’s co-authors, Andy Bailey, a plant pathologist at the University of Bristol, UK, said, “Our work raises the question of whether [A. tumefaciens’s] host range is wider than we had thought – maybe it’s not confined only to plants after all.”19

The implications of this research are that GM gene(s), once introduced by A. tumefaciens into a GM crop and released into the environment, could then be introduced into an organism outside the plant kingdom – in this case, a fungus – and genetically modify it. This would be an uncontrolled and uncontrollable process, with unpredictable consequences.

Implications of horizontal gene transfer through A. tumefaciens

Could A. tumefaciens transfer GM genes from a GM plant to another organism under realistic farming conditions? The answer depends on whether any A. tumefaciens carrying GM genes remains in the GM crop that is planted in open fields. Genetic engineers use antibiotics to try to remove the A. tumefaciens from the GM plant after the initial GM transformation process is complete in the laboratory. But this process has been found to be unreliable and incomplete:

  • A study on GM brassicas, potatoes and blackberries found that the use of three antibiotics failed to completely remove A. tumefaciens. Instead, the A. tumefaciens contamination levels increased from 12 to 16 weeks after the GM transformation process and the A. tumefaciens was still detected six months after transformation.20
  • A study on GM conifers found that residual A. tumefaciens remained in the trees 12 months after the genetic transformation but were not detected after this time in the same plants.21

However, these experiments only examined the first GM plant clones. In the GM development process, such GM clones go through a long process of back-crossing and propagation with the best-performing non-GM or GM plant relatives in order to try to produce a GM plant that performs well in the field and expresses the desired traits. The important question is whether A. tumefaciens carrying GM genes survives this back-crossing and propagation process and remains in the final GM plant that is commercialized.

To the best of our knowledge there have been no studies to assess whether any A. tumefaciens remains in the final commercialized GM plant. This question should be answered before a GM variety is commercialized, in order to avoid unwanted consequences that could be caused by residual A. tumefaciens in the final GM plant. Examples of consequences that should be excluded are the transfer of insecticidal properties to bacteria, or of herbicide tolerance to other crops or wild plants. The study by Knight and colleagues (2010) discussed above shows that the introduction of GM genes into crop plants could have consequences to organisms outside the plant kingdom, through the mechanism of infection by fungi carrying A. tumefaciens, which in turn carry GM genes.17

The consequences of such HGT for human and animal health and the environment are not predictable based on current knowledge, but are potentially serious. The health and environmental risk assessment for any GM variety must demonstrate that the GM plants have been completely cleared of GM A. tumefaciens before they are approved for commercialization.

Gene transfer by viruses

Viruses are efficient at transferring genes from one organism to another and in effect are able to carry out HGT. Scientists have made use of this capacity to create viral gene transfer vectors that are frequently used in research to introduce GM genes into other organisms. Such vectors based on plant viruses have also been developed to generate GM crops, though no crops produced with this approach have been commercialised to date.22,23

The viral vectors that are used to generate GM crops are designed to prevent the uncontrolled transfer of genetic material. However, because the long time period during which virally engineered crops would be propagated in the environment, and the large numbers of humans and livestock that would be exposed to this GM genetic material, there is a real, though small, risk that unintended modifications could occur that could lead to virus-mediated HGT – with unpredictable effects.

Another potential risk of virus-mediated HGT comes from GM crops engineered to contain a virus gene, in particular those carrying information for a viral “coat” protein. This is done in an attempt to make the crop resist infection and damage by the “wild” virus from which the viral GM gene was derived. However, it has been suggested that if a GM crop containing a viral gene of this type was infected by the wild virus, this may result in exchange of genetic material between the GM viral gene in the plant and the infecting virus, through a process known as recombination. This can potentially result in a new, more potent (“virulent”) strain of virus.24,25

The reasons for these concerns are as follows:

  • The GM viral gene will be present in every cell of the crop. As a result, the large-scale cultivation of such a viral GM gene-containing crop will result in an extremely high concentration of particular viral genes in fields.
  • It has been suggested that this provides an unprecedented opportunity for genetic recombination events to take place between an infecting virus and GM viral genes in the crop, thereby increasing the risk of new, mutated, and potentially more virulent strains of virus being produced.25 Such viral mutation with increased virulence has been shown to occur under laboratory conditions.26,27

To date only two GM crops engineered with genes from viruses have been commercialized: a variety of squash grown in the US28 and Mexico,29 and papaya cultivated in Hawaii.30 There are no reports of any investigations to see if any new viral strains have arisen by recombination in these two crops. Interestingly, and quite unexpectedly, although the GM squash was resistant to viral infection, it was found to be prone to bacterial wilt disease following attack by beetles.31,32


HGT from plants into other plants or animals appears to be a low-frequency event. The methods of HGT that are most likely to occur are DNA uptake by bacteria in the environment or the digestive tract. There is good evidence that the latter has already happened in the intestinal bacteria of humans who eat GM soy.

Other scenarios involving HGT by A. tumefaciens or by viruses are less probable. However, given the extremely wide distribution of GM crops and their intended use over decades, even low probabilities translate into the likelihood that HGT events could occur.

Therefore the negative impacts and risks associated with HGT must be taken into account in considering the overall biosafety of any GM crop.


  1. GMO Compass. Gene transfer to microorganisms. 2006. Available at:
  2. Then C, Bauer-Panskus A. European Food Safety Authority: A playing field for the biotech industry. Munich, Germany: Testbiotech; 2010. Available at:
  3. Kleter GA, Peijnenburg AACM, Aarts HJM. Health considerations regarding horizontal transfer of microbial transgenes present in genetically modified crops. J Biomed Biotechnol. 2005;2005(4):326-352. doi:10.1155/JBB.2005.326.
  4. Netherwood T, Martin-Orue SM, O’Donnell AG, et al. Assessing the survival of transgenic plant DNA in the human gastrointestinal tract. Nat Biotechnol. 2004;22:204–209. doi:10.1038/nbt934.
  5. Pontiroli A, Simonet P, Frostegard A, Vogel TM, Monier JM. Fate of transgenic plant DNA in the environment. Env Biosaf Res. 2007;6:15-35. doi:10.1051/ebr:2007037.
  6. Brigulla M, Wackernagel W. Molecular aspects of gene transfer and foreign DNA acquisition in prokaryotes with regard to safety issues. Appl Microbiol Biotechnol. 2010;86:1027-41. doi:10.1007/s00253-010-2489-3.
  7. Lerat S, Gulden RH, Hart MM, et al. Quantification and persistence of recombinant DNA of Roundup Ready corn and soybean in rotation. J Agric Food Chem. 2007;55:10226-31. doi:10.1021/jf072457z.
  8. Heinemann JA, Traavik T. Problems in monitoring horizontal gene transfer in field trials of transgenic plants. Nat Biotechnol. 2004;22:1105-9. doi:10.1038/nbt1009.
  9. Schubbert R, Renz D, Schmitz B, Doerfler W. Foreign (M13) DNA ingested by mice reaches peripheral leukocytes, spleen, and liver via the intestinal wall mucosa and can be covalently linked to mouse DNA. Proc Natl Acad Sci USA. 1997;94:961-6.
  10. Schubbert R, Hohlweg U, Renz D, Doerfler W. On the fate of orally ingested foreign DNA in mice: chromosomal association and placental transmission to the fetus. Mol Gen Genet. 1998;259:569-76.
  11. Mazza R, Soave M, Morlacchini M, Piva G, Marocco A. Assessing the transfer of genetically modified DNA from feed to animal tissues. Transgenic Res. 2005;14:775–84. doi:10.1007/s11248-005-0009-5.
  12. Sharma R, Damgaard D, Alexander TW, et al. Detection of transgenic and endogenous plant DNA in digesta and tissues of sheep and pigs fed Roundup Ready canola meal. J Agric Food Chem. 2006;54:1699–1709. doi:10.1021/jf052459o.
  13. Chainark P, Satoh S, Hirono I, Aoki T, Endo M. Availability of genetically modified feed ingredient: investigations of ingested foreign DNA in rainbow trout Oncorhynchus mykiss. Fish Sci. 2008;74:380–390.
  14. Ran T, Mei L, Lei W, Aihua L, Ru H, Jie S. Detection of transgenic DNA in tilapias (Oreochromis niloticus, GIFT strain) fed genetically modified soybeans (Roundup Ready). Aquac Res. 2009;40:1350–1357.
  15. Tudisco R, Mastellone V, Cutrignelli MI, et al. Fate of transgenic DNA and evaluation of metabolic effects in goats fed genetically modified soybean and in their offsprings. Animal. 2010;4:1662–1671. doi:10.1017/S1751731110000728.
  16. Spisak S, Solymosi N, Ittzes P, et al. Complete genes may pass from food to human blood. PLOS ONE. 2013;8(7):e69805.
  17. Knight CJ, Bailey AM, Foster GD. Investigating Agrobacterium-mediated transformation of Verticillium albo-atrum on plant surfaces. PLoS ONE. 2010;5:13684. doi:10.1371/journal.pone.0013684.
  18. Kunik T, Tzfira T, Kapulnik Y, Gafni Y, Dingwall C, Citovsky V. Genetic transformation of HeLa cells by Agrobacterium. Proc Natl Acad Sci USA. 2001;98:1871-6. doi:10.1073/pnas.041327598.
  19. Marshall T. Bacteria spread genes to fungi on plants. Planet Earth Online. Published October 27, 2010.
  20. Barrett C, Cobb E, McNicol R, Lyon G. A risk assessment study of plant genetic transformation using Agrobacterium and implications for analysis of transgenic plants. Plant Cell Tissue Organ Cult. 1997;47:135–144.
  21. Charity JA, Klimaszewska K. Persistence of Agrobacterium tumefaciens in transformed conifers. Env Biosaf Res. 2005;4:167-77.
  22. Gleba Y, Marillonnet S, Klimyuk V. Engineering viral expression vectors for plants: the “full virus” and the “deconstructed virus” strategies. Curr Opin Plant Biol. 2004;7:182-8. doi:10.1016/j.pbi.2004.01.003.
  23. Gleba Y, Klimyuk V, Marillonnet S. Viral vectors for the expression of proteins in plants. Curr Opin Biotechnol. 2007;18:134-41. doi:10.1016/j.copbio.2007.03.002.
  24. Hull R. Detection of risks associated with coat protein transgenics. In: Foster GD, Taylor SC, eds. Methods in Molecular Biology: Plant Virology Protocols: From Virus Isolation to Transgenic Resistance. Vol 81. Totowa, NJ: Humana Press Inc. 1998:574–555.
  25. Kleiner K. Fields of genes. New Sci. 1997. Available at:
  26. Nowak R. Disaster in the making. New Sci. 2001;169(2273):4–5.
  27. Jackson RJ, Ramsay AJ, Christensen CD, Beaton S, Hall DF, Ramshaw IA. Expression of mouse interleukin-4 by a recombinant ectromelia virus suppresses cytolytic lymphocyte responses and overcomes genetic resistance to mousepox. J Virol. 2001;75:1205-10. doi:10.1128/JVI.75.3.1205-1210.2001.
  28. US Department of Agriculture Animal and Plant Health Inspection Service (APHIS). Environmental assessment for Upjohn Company/Asgrow Seed Company petition for determination of non-regulated status for CZW-3 squash. Washington, DC; 1996.
  29. Acatzi A, Magaña J, Moles C, Peña C, Castillo M. Detection and quantification of GM maize varieties in Mexican imports. Mexico City, Mexico; 2008. Available at:
  30. Gonsalves D. Transgenic papaya in Hawaii and beyond. AgBioForum. 2004;7:36–40.
  31. Sasu MA, Ferrari MJ, Du D, Winsor JA, Stephenson AG. Indirect costs of a nontarget pathogen mitigate the direct benefits of a virus-resistant transgene in wild Cucurbita. Proc Natl Acad Sci USA. 2009;106:19067-71. doi:10.1073/pnas.0905106106.
  32. Leonard A. Transgenic squash super-weeds gone wild. Published October 28, 2009.