Gene-edited non-browning mushroom may have wrongly escaped regulatory scrutiny, scientists suggest. Report by Claire Robinson
GMOs claimed to contain no foreign DNA may have been exempted from GMO regulatory processes under false pretences, according to a correspondence article by South Korean scientists published in Nature Biotechnology.
While the article was published back in 2016, it has urgent relevance for today, as some governments are moving to deregulate certain classes of gene-edited GMOs based on the hypothetical – and, it now appears, mostly false – assumption that they do not contain any foreign DNA.
In the article, authors Jungeun Kim and Jin-Soo Kim cite the case of the gene-edited non-browning mushroom. Earlier in the same year, the USDA exempted the mushroom from regulatory oversight on the basis of claims by the developer, Yinong Yang of Penn State University, that it was "transgene-free" and had "no foreign DNA" integrated into its genome.
But Kim and Kim point out that the approved mushroom "may still contain tiny fragments of foreign DNA in its genome". This is due to the way it was produced – in other words, the all-important gene editing process. In the process of CRISPR gene editing, genetically engineered plasmids (circular molecules of DNA derived from bacteria) were used to introduce the genes carrying the information for the protein and guide RNA components of the editing tool into mushroom cells with the aim of disrupting a gene linked with browning. In other words, the plasmids encode the gene-editing tool; the whole construct is sometimes called the vector. The idea is that the plasmids will be present in the mushroom cells only temporarily, but long enough to carry out the intended gene edit before disappearing as the mushroom cells divide and new whole mushrooms are generated from the edited cells.
The developer analysed the genome of the resulting edited mushrooms, using methods called PCR and Southern blotting, and found no evidence of foreign (plasmid) DNA in the genome. In response, the USDA deregulated the gene-edited mushroom.
However, Kim and Kim say that the plasmid gene editing method "can result in cells or organisms that contain small portions of foreign DNA (up to hundreds of base pairs) that are derived from the introduced plasmids".
Kim and Kim state that these small inserted sequences "cannot be detected by PCR or Southern blot analyses", such as the mushroom developer used. In other words, the mushroom developer used the wrong tool for the job. Instead, Kim and Kim recommend using whole genome sequencing to look for unintended insertions of foreign plasmid-derived DNA.
Kim and Kim tested their hypothesis by conducting an experiment in which they gene edited the cells of a cress-like model plant called Arabidopsis, using the same plasmid method used by the non-browning mushroom developer. They looked for plasmid-derived DNA sequences using a DNA sequencing screening method targeted to the edit site (not whole genome sequencing). Sure enough, they found them.
They caution that this method will likely have underestimated the frequency of unintended mutations caused by the editing tool because it cannot spot insertions at genome-wide off-target sites. They cite their previous experimental findings in gene-edited human cells that fragments of plasmid-derived DNA can end up scattered all over the genome.
GMWatch warns that given the likelihood that plasmid fragments will be widespread in the genome of CRISPR edited plants, these will be difficult or impossible to breed out in the subsequent breeding steps typically carried out by GMO developers with the aim of removing unintended mutations.
Governments rush to deregulate based on false assumptions
In the EU the gene-edited mushroom would be classed as having been produced with an SDN-1 or gene disruption type of gene editing. GMO lobbyists, as well as the UK government, claim that SDN-1 should be exempted from the GMO regulations on grounds that it could happen naturally and/or does not entail introducing foreign DNA. Based on the “no foreign DNA” assumption, the Indian government has already deregulated SDN-1 gene-edited plants, as well as gene-edited plants made using a DNA repair template (SDN-2). Also, the Australian regulator has deregulated SDN-1.
Given Kim and Kim’s findings, these arguments and moves appear to be scientifically and technically baseless and open to challenge.
It is shocking that this problem of unintended plasmid-derived foreign DNA in gene-edited plants has been known about for many years. But some governments and GMO lobbyists are still turning a blind eye and going along with the pretence that any GMO that is claimed by the developer to be free from foreign DNA really is so.
A better tool?
Kim and Kim recommend that in the gene editing of plants, developers abandon the use of plasmid-encoded editing tools and instead use preassembled editing tools (the technical term is "preassembled Cas9 or Cpf1 RNPs"). This preassembled Cas or Cpf RNP (ribonucleoprotein) complex consists of protein and the guide RNA. As it lacks any plasmid DNA, introducing preassembled Cas or Cpf RNP complexes into plant cells to carry out the gene edit would reduce or eliminate the potential for unwanted foreign DNA in a crop genome. The authors confirmed this in another experiment on Arabidopsis using these preassembled Cas RNP tools. They found that as expected, the target edit site was clear of plasmid-derived DNA insertions.
The authors declare an interest, in that they have filed patent applications on Cas9 and Cpf1 RNP methods. However, this is no reason to doubt their finding, which is solidly argued and confirmed by experiment.
Who will use the preassembled editing tools?
What about the authors’ recommendation that gene-edited plant developers use preassembled editing tools? As a 2019 review points out, using these tools “can be very costly and laborious”. Moreover, they cannot be used in all plants: “So far, these transgene-independent techniques are only feasible with very few plant species and varieties.”
It is clear from this and other reviews that the preferred and most feasible method for plant gene editing is to use plasmids encoding and expressing the CRISPR-Cas and guide RNA, rather than the preassembled RNP tools.
Given the GMO industry’s predilection for taking the cheap and easy route, both in choice of gene-editing tool and screening method, the vast majority of the gene-edited plants we can expect to arrive on the market are likely to have fragments of plasmids integrated in their genomes.
Kim and Kim say they don't believe that the presence of small plasmid-derived insertions in a plant genome pose any risk to health or the environment – an optimistic but baseless assumption, since no risk research has been done that could answer this question. Depending on where the plasmid DNA fragment or fragments have inserted in the plant genome, gene functions could be disrupted, with unknown consequences to the plant’s biochemistry, composition, and safety.
However, they worry that if foreign DNA sequences were to be found in the genome of a gene-edited crop that becomes commercially available in the future, "it would set back both the regulatory authority and the industry".
Potentially, these GMOs would be subject to legal challenge.
GMWatch warns Indian government
Kim and Kim's warnings are reflected in GMWatch's own. Just days ago and without having seen Kim and Kim’s paper, but using exactly the same reasoning, we criticised the Indian government for deregulating GMOs based on the assumption that they are produced without introducing any foreign DNA. We also recommended that whole genome long-read sequencing be used in order to detect any such unintended insertions of DNA, as well as other types of unintended mutations (DNA damage).
While it's gratifying to have additional scientific backup for our position, we'd gladly trade our opportunities to feel smug in return for regulators doing their job properly.
Dangers not restricted to foreign DNA
Avoiding the unwanted incorporation of foreign DNA in edited genomes by using preassembled Cas9 or Cpf1 RNPs can remove one risk factor, but there are plenty of others. Numerous other types of unintended mutations (DNA damage) are caused by gene editing and its associated processes at both off-target sites and on-target ones (i.e. at the desired editing site). These include large deletions, insertions and rearrangements of DNA. These unintended changes could result in dangerous consequences, as a recent scientific paper pointed out – such as unexpected toxicity or allergenicity of gene-edited plants.
While these unintended and risky effects of gene editing remain the principal reason for opposing deregulation, it is also important to point out that one major assumption of the pro-deregulation lobby – that whole classes of gene-edited organism are free from foreign DNA – has no scientific foundation.