By Claire Robinson of GMWatch with technical advice from Dr. Michael Antoniou and Dr. Yves Bertheau
Currently, governments around the world are moving to scrap regulatory safeguards around genetically modified organisms (GMOs) that are claimed to contain no foreign genes or foreign DNA. They choose to believe GMO lobby assertions that GM gene-edited plants and animals don’t contain any foreign genetic material in their genomes and are therefore as “natural” and safe as conventionally bred plants and animals.
But this is false. GM gene-edited organisms pose risks that are not confined to the presence of foreign DNA. And GM gene-edited plants and animals can and do contain foreign genetic material in their genomes, either by intention or inadvertently due to the imprecision and limitations of the gene editing process. Below are some of the ways that foreign DNA can get into the genomes of gene-edited GMOs.
Foreign genes or DNA can be intentionally inserted
In a class of gene-edited organisms sometimes referred to as SDN-3, the outcome is the intentional insertion of a new gene. This is achieved as follows. First, a double-strand cut across the DNA double helix is made by the gene-editing tool via its innate “nuclease” enzymatic activity allowing it to act like a “gene scissors”. Second, a large foreign DNA fragment is inserted into the target organism’s genome at the location of the double-strand DNA break. This DNA fragment can contain a complete gene or other genetic elements (e.g., gene regulatory sequences) and acts as a template for repair of the double-strand DNA break brought about by the gene-editing tool’s nuclease function. In this way, the foreign gene or other types of DNA genetic material is intentionally integrated into the genome of the gene-edited plant, animal, or other organisms to confer a new trait.
Gene editing processes introduce foreign genetic material into cells, which can inadvertently integrate into the genome
The DNA that is introduced into cells to bring about gene editing is produced in, and isolated from, micro-organisms (bacteria). The DNA of the micro-organisms can end up being inadvertently integrated into the genome of the organism that is being gene-edited.
For example, in plant gene editing, plasmids (circular DNA molecules) carrying genes encoding for the editing tool introduced into the plant cells are grown in E. coli bacteria. And the soil bacterium Agrobacterium tumefaciens, which is used in older-style genetic engineering to infect and thus introduce and integrate foreign DNA into plants, is also used in gene editing to deliver the DNA encoding the gene editing tools into plant cells. DNA from plasmids, E. coli, and Agrobacterium can end up being unintentionally inserted into the genome of the gene-edited plant.
In animals, viruses and viral vectors (delivery vehicles) are used to carry the gene-editing tool into the target cells. Foreign DNA from these viral vectors has been found to integrate into gene-edited animals’ genomes.
Below are some specific examples of foreign (including contaminating) DNA from production processes inadvertently integrating into the genomes of gene-edited plants and animals.
Fragments of foreign plasmid DNA can inadvertently integrate into plants’ genomes
At present, the vast majority of CRISPR-mediated plant gene editing is undertaken via the “old-style” transgenic GM technique of introducing plasmids into plant cells, which encode for the gene-editing tool. This means that on the plasmid are genes that encode for the protein and guide RNA elements of the CRISPR/Cas complex. Once inside the plant cells, these genes are expressed, resulting in the assembly of the CRISPR/Cas editing tool, which can then go about its business.
Under these circumstances, the plasmid introduced into the plant cells can fragment, with these fragments being randomly inserted into the plant cell genome.
Intact foreign plasmids can unintendedly integrate into animals’ genomes
Foreign plasmid DNA can unintendedly integrate into the genome of gene-edited animals, albeit by a different mechanism than occurs in plants (above). In plants, fragments of the plasmid encoding for the gene-editing tool can be integrated into the genome. In animals, the whole plasmid can be incorporated.
An example of plasmid integration in animals is the case of Recombinetics’ cattle gene-edited not to grow horns. In these cattle, the insertion of the intended DNA sequences for hornlessness was not “clean”. Unexpectedly, the entire DNA repair template, including the plasmid containing bacterial DNA sequences, was integrated into the genome at one of the two intended sites of the double-strand DNA break brought about by the gene-editing tool. As a result, the cattle’s genomes were found to unexpectedly contain bacterial plasmid DNA sequences encoding for resistance to three different antibiotics.
The developer company had failed to spot the rogue DNA, even going too far as to arrogantly and unscientifically assert that their animals were free from the unintended effects of gene editing.
The foreign DNA was only found by scientists at the US Food and Drug Administration, who performed their own analysis of the cattle’s genomes. That’s precisely the kind of analysis that won’t be done in the absence of regulation of gene editing.
In another example, attempts to gene edit cattle by inserting a male sex-determining gene (“SRY”) – giving rise to only male animals – went dramatically wrong. The gene editors tried to achieve a clean insertion of the SRY gene alone at the intended edit site. But what they got was quite different: Seven copies of the whole bacterial plasmid carrying the SRY gene became incorporated into the genome of the single calf that was born. And the calf was always going to be male, even without the gene editing.
This was not only a failed experiment but another instance of gene editing resulting in unintended foreign (in this case bacterial plasmid) DNA insertion into the genome of the targeted organism.
Foreign bacterial chromosomal DNA from Agrobacterium can inadvertently integrate into the genome
Agrobacterium infection is the most efficient genetic modification transformation technique for plants. That is why it is used in older-style transgenic genetic engineering and is still the most frequently used technique in newer gene editing applications such as CRISPR/Cas.
However, it has been found that DNA fragments up to 18,000 base units in length – large enough to contain whole genes from the Agrobacterium genome – can integrate into the genome of the plant during the genetic transformation process. This clearly demonstrates that functional whole genes of Agrobacterium can be introduced into plants during the gene editing process. The authors of this study rightly concluded that there is “a need for greater scrutiny of transgenic plants for undesired bacterial DNA”.
GMO developers often turn a blind eye to foreign DNA
Developers of gene-edited plants and animals routinely miss or ignore the inadvertent integration of foreign DNA into their products, as they are using inadequate screening methods, as pointed out in a recent systematic literature review.
As a result, foreign plasmid or chromosomal DNA can easily be present in the final marketed product, making it transgenic in nature – that is, containing foreign genetic material.
In addition, the foreign DNA may be functional. That means it can produce one or more novel proteins, with unknown consequences for the consumer, such as toxicity or allergenicity.
GMO developers can backcross gene-edited plants with non-GM elite varieties to reduce or eliminate unintended foreign DNA and other unintended damages of the gene editing process. However, this process is costly and time consuming and nullifies the speed that is often the reason why GMO developers favour gene editing. So increasingly, GMO developers prefer to directly gene edit elite varieties, on the basis that it can reduce or eliminate the need for backcrossing.
Even when backcrossing is done, it cannot purge the genomes of all unintended DNA damage. Again, the lack of backcrossing that aims to speed up the journey to market for gene-edited GMOs will inevitably mean that these products will reach the market containing foreign DNA and other unwanted DNA damage.
That’s especially so in gene-edited plants that are vegetatively propagated (for example, potatoes), where no backcrossing is undertaken. In these cases, all unintended gene editing process-induced DNA damage and the resulting altered biochemistry will be present in the final marketed product, with unknown health consequences for the human or animal consumer.
Claims that using pre-assembled editing tools avoid foreign DNA are false
Advocates of gene editing deregulation sometimes claim that in the future, developers will avoid the insertion of foreign plasmid DNA. Instead, the plan is to use short-lived (to reduce off-target and on-target unintended changes), purified, pre-assembled gene editing tool protein complexes. Such a purified complex is the ribonucleoprotein (RNP) that constitutes CRISPR/Cas. For example, pre-assembled CRISPR/Cas has been used to gene-edit maize.
The problem is, though, that purified proteins (enzymes) have been found to be contaminated with DNA from the organism used to manufacture them during the production process.
This raises the possibility that the purified gene editing tool protein complex, including the CRISPR/Cas RNP produced in bacteria, can also be contaminated with bacterial host DNA. In this case, any contaminating bacterial host DNA will be delivered to cells along with the pre-assembled gene editing tool. This means that this key argument of the pro-deregulation lobby may be invalid from the get-go.
Based on these facts, Dr Yves Bertheau, honorary INRA research director at the Muséum national d’Histoire naturelle in Paris, France and an expert in GMO detection and identification, wrote in a scientific book chapter that even the most intensively purified commercial enzymes can be DNA-contaminated. Because of this, he wrote that it is “disgraceful to assert” that RNP CRISPR/Cas editing tool preparations are necessarily DNA-free and that no foreign sequences can therefore be incorporated into the edited plants’ genome – or, more generally, to claim that new GM techniques (called in the chapter “new breeding techniques” or NBTs) do not leave contaminating DNA in the genomes of edited plants.
The absence of foreign DNA insertion can only unequivocally be determined empirically by direct ultradeep whole genome sequencing of the gene edited organism. Without such an analysis, claims that the genomes of gene-edited GMOs are free from foreign DNA are baseless.
This is especially the case, Dr Bertheau told GMWatch, because “The few researchers using RNPs do not verify that the purification process has been fully effective.”
Foreign DNA from gene editing “kill switches” could contaminate the genome
The “gene scissors” nuclease activity of gene editing tools is very efficient at cutting DNA – sometimes too efficient. Gene editors want to make certain cuts but avoid unintended cuts, so they have sought ways to control the editing tool nuclease introduced into the target cells by using activators or inhibitors. This reduces the tendency of the nuclease to cut again and again, causing multiple double-stranded DNA breaks that cause off-target and on-target unintended changes in the original cells and their descendants.
For this reason, “anti-CRISPR proteins” have been used in CRISPR gene editing applications as gene editing “kill switches” or “brakes” to reduce the duration of action of nucleases in bacterial and mammalian cells, as well as in plants.
Anti-CRISPR proteins are introduced into cells either as plasmids encoding them or as purified proteins manufactured in bacteria. As we have seen above, plasmids can unintentionally integrate into the target cell’s genome, either as a whole or as fragments. And proteins purified from bacteria are known to be contaminated with bacterial host DNA. So in either scenario it is possible that foreign DNA (plasmid, bacterial genome) can become integrated into the target cells’ genome during any gene editing process that is also employing an anti-CRISPR protein strategy. Therefore, the use of anti-CRISPR proteins adds yet another way in which foreign DNA may be inadvertently inserted into the genome of the edited plant or animal.
As a result, gene-edited organisms need to be carefully screened to see if elements of the gene editing “kill switches” or contaminating DNA is unintentionally incorporated into the genome of the edited plant or animal.
Foreign DNA from animal-derived culture media can be inserted into the genome of gene-edited animals
Genomes of gene-edited animals can contain foreign contaminating DNA derived from animal-derived culture components (serum). Animal (and human) cell tissue culture media most frequently contain animal-derived serum as a source of essential growth factors. This animal-derived serum is contaminated with DNA from the source animal (e.g., bovine), which can be inserted into the genome of cells during the gene editing process and so can be present in the genomes of gene-edited animals.
For example, edited mouse genomes were found to acquire bovine or goat DNA. This was traced to the use, in standard culture medium for mouse cells, of foetal calf serum – that is, blood extracted from cows.
Commenting on the implications of the finding, Dr Jonathan Latham, director of the Bioscience Resource Project, said, “Cutting DNA inside cells, regardless of the precise type of gene editing, predisposes genomes to acquire unwanted DNA. The unwanted DNA may come from inside the edited cell, or it may come from the culture medium, or it may come from any biological material added to the culture medium, whether accidentally or on purpose.”
Dr Latham warned that some culture media can contain DNA sequences from viruses. Amongst the DNA sequences found by researchers to have inadvertently inserted into the genomes of gene-edited mice was mouse retrovirus DNA (HIV is a type retrovirus): “Therefore, it is not hard to imagine, for instance, gene-edited animals becoming the breeding stock that leads to the development or spread of novel or unwelcome viruses or mycoplasmas [a type of bacteria].”
GMO developers fail to clean up foreign DNA
GMO developers claim they will check for foreign DNA and select or breed it out. But the evidence shows that they don’t do this and that they aren’t even looking properly. They commonly use inadequate screening methods, which can easily miss the presence of foreign DNA and other unintended effects of gene editing.
These foreign fragments of DNA, both large and small, are already detectable using widely available deep and ultra-deep short- and long-read whole genome sequencing methods. Therefore there is no excuse for GMO developers failing to employ them.
Risks not confined to foreign DNA
In any discussion of foreign DNA in gene-edited GMOs, it’s crucial to bear in mind that the risks of these products are not confined to the introduction of foreign genes or other types of DNA elements.
One set of risks arises from plant tissue culture, which is a necessary part of the process of making all older-style transgenic GM and the vast majority of gene-edited plants. Tissue culture is known to cause many mutations (DNA damage). A study on rice plants (Tang and colleagues, 2018) found that tissue culture causes large-scale, genome-wide unintended mutations. The tissue culture-associated gene editing GM transformation process (involving Agrobacterium insertion) markedly increased the quantity of unintended DNA damage over and above the damage from the tissue culture alone. Taken as a whole, the gene editing process was found to cause far more mutations than were found to occur spontaneously in natural breeding of non-GM rice.
The study also found that the activity of a disabled CRISPR/Cas gene editing tool, considered alone, caused few unintended off-target mutations, but the researchers didn’t look for unintended on-target mutations (at the intended edit site), so we don’t know how many of those occurred.
This is an important omission, because a separate study found that CRISPR gene editing in rice caused large numbers of off-target and on-target mutations, though it is not known whether these stemmed from tissue culture or the gene-editing GM transformation or other elements of the gene editing process. The researchers warned that “Understanding of uncertainties and risks regarding genome editing is necessary and critical before a new global policy for the new biotechnology is established”.
In sum, unintended mutations accumulate from the various stages and components of the gene editing process (tissue culture, the GM transformation process, and gene editing tool activity). Not surprisingly, a significant and ever-growing number of studies have found numerous types of unintended DNA damage caused by gene editing and its associated processes at both off-target and on-target sites. These include large deletions, insertions, and rearrangements of DNA.
In gene-edited plants, these unintended genetic changes could result in marked alterations in gene expression and thus in biochemical composition, potentially including unexpected toxicity or allergenicity, as publications in the scientific literature have pointed out.
Taken together, the risks posed by the unintended effects of gene editing – whether or not they involve the integration of foreign DNA into the genome of the gene-edited organism – are the main reason for opposing deregulation. But it should also be emphasised that a major claim of the pro-deregulation lobby and one on which they hang assertions of safety and naturalness – that whole classes of gene-edited organisms are free from foreign DNA – is entirely baseless and false. Only direct whole genome screening (sequencing) of the gene-edited plant or animal can confirm whether it is truly free from foreign DNA. Anything short of this is an assumption without any evidential basis.
Claire Robinson is with GMWatch and Dr Yves Bertheau is honorary INRA research director at the Muséum national d’Histoire naturelle in Paris, France and an expert in GMO detection and identification. We are grateful to them both for their input into this article. However, any remaining errors are entirely the responsibility of the author.
Notes for this article and links to sources are here: