Currently, several new genetic modification techniques are being discussed to determine whether or not the products derived from them will be regulated as transgenic GMOs. This article by Eric Meunier of the journal inf'OGM allows us to understand some of the potential risks associated with the mere implementation of any genetic modification technique on a plant cell culture.
Ahile genetic modification techniques, whether new or old, are not fully mastered, they can bring certain new characteristics to a living being (such as the ability to tolerate a herbicide), they also unintentionally produce other modifications known as "off-target" because they take place in other parts of the genome than the one initially targeted.
On 7 April 2016, as an echo to the remarks made by Yves Bertheau, former member of the High Council for Biotechnology (HCB), since December 2015, Jean-Christophe Pagès, President of the Scientific Committee (SC) of the HCB, explained to the Parliamentary Office for the Evaluation of Scientific and Technical Choices (OPECST) about one of these new techniques, Crispr/Cas9 , that he "... is not only a member of the HCB, but also a member of its Scientific Committee (SC), which is responsible for the evaluation of scientific and technical choices. we must not forget the difficulties of its application [...] in particular in vivo in animals, since a matrix must be provided and there are currently major problems with vectorization to be able to do this..
In vitro, in culture, on the other hand, it's something much easier and that's why most applications are research applications and possibly applications for which we can reconstitute organisms from in vitro culture, and this does indeed concern certain plants. "... These comments are very disturbing because, on closer examination, the opinion of the HCB SC of 4 February 2016 - which has since become an interim report - does not mention such difficulties in vivo, or such "facilities" for in vitro implementation.
READ IN UP' : CRISPR: Revolution in human history or mega time bomb?
In the different stages of the genetic modification process. We are going to focus on the so-called vectorization stage mentioned by Jean-Christophe Pagès, which consists in bringing into a cell the material intended to generate the desired genetic modification. We will also look at the stages prior to this vectorization phase, which are sources of stress that can induce mutations and epimutations (see box below).
Mutation, epimutation: what are we talking about?
A mutation is commonly defined as a change in the genetic information contained in an organism, whether in the form of DNA or RNA. Mutations are hereditary. They may be "silent", meaning they have no involvement in the metabolism of the body. They can also affect the expression of one or more genes, altering metabolism.
Epimutations belong to the family of mutations affecting the expression of a genetic sequence but which are not due to a modification of the genetic sequence itself. They may, for example, be due to a change in the chemical composition of the DNA building blocks, the nucleotides.
Preparation of cells to be transformed
The first step before being able to introduce material into cells (the vectorization mentioned by J.-C. Pagès) is to prepare the plant cells. Laboratory technicians will break the cell wall, or even remove it completely. The plant cells that have had their cell wall removed, or protoplasts, then become transformable, and biologists can then introduce a variety of tools such as large proteins, RNAs and/or DNAs encoding in these cells. However, as Yves Bertheau points out, this "simple" formation of protoplasts induces mutations and epimutations, a phenomenon abundantly observed in the scientific literature [1].
The simple culture of cells induces mutations.
The second step is to grow these protoplasts. The very fact of culturing cells generates mutations and epimutations. The most surprising thing is that a scientific bibliography shows that the mechanisms underlying the appearance of these mutations and epimutations are still relatively unknown despite decades of use [2]. This phenomenon, known as somaclonal variation, is such that it has been used by seed companies to create the "genetic diversity" necessary to "improve" plants, to use an element of seed companies' language.
The National Seed Industries Group (GNIS) explained as well as "Somaclonal variation is the change observed in some cells after a long cycle of in vitro cultures without regeneration. They are then no longer identical to the mother plant. This variation may be due to a change in the nuclear genome or the genome of cytoplasmic organelles. ".
In other words, the plants obtained from these cells will have different characteristics. The GNIS provides a final interesting clarification: " the resulting changes in characteristics are not very stable and are not always reflected in the regenerated plant, or its progeny ». The reason? Changes that appear (epimutations) can make the obtained mutations disappear [3] ... As Yves Bertheau explains, "... it seems difficult under such conditions to predict what impact this cell culture step may have when implementing a new genetic modification technique ".
Vectorization, I mean...
Cells prepared and cultured, we are finally ready to introduce the material to generate the desired modification. Depending on the techniques, this material may consist of proteins and/or genetic sequences of the RNA or DNA type - a molecule more frequently used in plants - coding (oligonucleotides, plasmids, viruses, etc.). The methods used to penetrate this material into cells simply consist of making holes in the remaining membranes (cytoplasmic and nuclear) of the cell. However, as Yves Bertheau explains, making such holes induces, once again, mutations and epimutations [4]. Above all, the researcher believes that it is impossible to predict a general risk assessment grid. It is necessary to choose between several vectorization techniques, between different types of material, depending on the genetic sequences to be introduced and the target species. In the end, only a case-by-case analysis as foreseen for GMOs makes it possible to assess the potential risks associated with all these unintended effects.
The interim report of the SC of the silent HCB on these mutations
In a paper published in 2011, scientists estimated that 35% of all the unintended effects observed following genetic modification of the rice variety Senia by transgenesis are due to the process of cell transformation itself [5]. The phenomenon is therefore not insignificant.
Surprisingly, and despite the statements of its President to OPECST, the HCB Scientific Committee did not report of these risks in its interim report on the risks associated with new techniques. While the issue of "vectorization" is well addressed in the SC sheets for each technique, it is only a matter of listing the means of implementing a technique and describing how the material is introduced into the cells. Nowhere are the mutations and epimutations that may emerge at each of the stages as we have just seen reported. Since HCB is composed of a committee of scientific experts, it is expected that this committee will discuss and explain, not ignore, these points. All the more so since vectorization - to speak only of what is mentioned in the SC report - does not appear to be completely perfected according to the techniques, the same SC noting for oligonucleotide-directed mutagenesis that "... the HCB is not a vectorized mutagen, but rather a vectorized mutagen...". many molecules or molecular mixtures are being tested to improve vectorization, which works relatively well in vitro but very little in whole organisms (Liang et al., 2002) » [6].
Selecting and regenerating "modified" cells is not without effect.
Because of their low or very low transformation efficiency [7], genetic modification techniques involve downstream selection of the cells that have actually been modified. This selection is done by using markers that make it possible to differentiate the cells: antibiotic or herbicide resistance gene (the cells are then immersed in a bath of antibiotics or herbicides and only the really modified cells survive), a gene that allows the cell to grow in the presence of a molecule that is usually toxic, or a gene that allows the cells to use a carbon source that is not normally metabolisable, or any other marker that will be persistent or eliminated in the end [8]. But the suppression of these selection markers is done by more or less reliable and precise techniques, which can therefore potentially induce cellular toxicities and other chromosomal rearrangements [9], leave fingerprints [10] and produce recombination sites (sites of genetic sequence exchange between two strands of DNA) with unpredictable effects [11]. These techniques are not necessarily possible for all plant species.
From these selected cells, plants must then be regenerated. A new stage of cell culture using a variety of synthetic hormones, which may also induce mutations and epimutations with or without chromosome rearrangements [12].
Whether it is the genetic modification stage itself, the preliminary stages, or the subsequent stages, all induce mutations or epimutations, known as off-target effects. But, one can hear in the corridors that eventually these (epi)mutations would not be present in the commercially available plant. The reason? The next step, called backcrossing, would get rid of all these unintended effects .
The theory behind backcrossing
Let us recall the general principle: a company that wishes to market a genetically modified variety (whether legally considered as GMO or not) will not directly modify the genome of one of the "elite" varieties but that of a more common variety. Once the modification has been obtained, the company will then crossbreed the common genetically modified (GM) variety with the commercially interesting elite variety for the first time. It will then re-cross the progeny with the initial elite variety until these progeny are considered (based on statistical analysis) to be almost similar (referred to as quasi-isogenic variety or negative segregants) to the elite variety. The difference between the elite variety and the new variety is assumed to be "almost" only the modified sequence.
The GNIS explained as well as, at the last crossing, " the share [of the genome from] the elite lineage is 96.88%, then the resulting lineage is considered to be sufficiently close to the elite lineage. The aim is to obtain an isogenic lineage, differing from the elite lineage by only one gene [the one containing the modification]. ». It should be noted here that a percentage of 96.88 % still leaves a lot of room for potential mutations, epimutations and rearrangements for some crops with very large genomes such as wheat: for about 17 billion base pairs constituting the wheat genome, the remaining 3.12% still represents more than 500 million base pairs .
The limits of the theory
The theory behind "cleaning up" off-target effects through backcrossing is based on the assumption that the unintended effects that appeared at earlier stages are sufficiently distant from the region where the desired genetic modification took place. Indeed, the closer these off-target effects are to each other, the greater the likelihood that the unwanted effects will be passed on with the genetic modification at each crossing. From a minimum number of seven, the number of crosses needed could then rise to more than 14 to try to get rid of them.
However, in addition to this problem of the proximity of off-target effects to the modified gene sequence, confidence in this theory is also put into perspective by two other biological phenomena. On the one hand, genetic sequences can exist and evolve into more or less large blocks. Since these blocks are passed on to the offspring as they are, the off-target effects contained in these blocks will then remain with the modified gene sequence at each crossing. On the other hand, certain genetic sequences have the capacity to "impose themselves" during the formation of gametes. Call them "selfish", these sequences will be contained in a greater number of gametes than expected and, from generation to generation, they will always be present in the offspring. If off-target effects appear in such sequences, it will be harder to get rid of them [13].
These elements therefore imply that a case-by-case verification appears necessary. At present, only the legislation on transgenic plants requires such case-by-case verification with the most complete set of analyses possible (with deficiencies, however, let us recall this). Indeed, genome sequencing does not make it possible to answer the question of the presence or absence of off-target effects because of the inherent limitations of this type of analysis (see box below).
Sequencing and associated computer tools? Anything but a guarantee of certainty!
On 7 April, at the hearing organised by the Parliamentary Office for the Evaluation of Scientific and Technical Choices (OPECST), André Choulika, CEO of Cellectis, stated on the subject of off-target effects ". resequencing the entire [plant genome] is really important [...] because in the approval, you are asked for the entire sequence again. ». Except that, on closer inspection, the sequencing results obtained are far from being absolutely reliable.
So-called "new generation" sequencing is today relatively cheap and fast. But there are several "problems" with its implementation, reading and use of the results.
First of all, several steps of the sequencing itself "parasitize" the reliability of the final result. One must know how to extract the DNA correctly, cut it into pieces and then sequence them using various platforms and methods. These platforms and methods are quite different from each other, both in terms of limitations and reliability of results [14].
These results should then be read by trying to put the read sequences back together to reconstruct the entire genome. The resulting sequences are then compared with other sequences that are considered "reference" sequences and stored in databases that already contain errors themselves [15].
These steps introduce significant imprecision into the final results of unintended effects detection and thus risk assessment, with uncertainties about sequences increasing with polyploid genomes or numerous repeated sequences [16]. Moreover, the mutations identified may not ultimately be of the same biological significance [17] ...
Numerous articles summarize the difficulties encountered at each stage, compare the methods, platforms [18] and associated software [19], discuss gold standards and standards to be implemented to make the entire process more reliable [20]. In short, as these authors point out, these are all techniques and stages that are in the process of being improved because they have not reached maturity, are evolving and therefore require a number of verifications for case-by-case evaluations.
According to many researchers, knowing how to deal with the accumulation of very many results (one of those famous "big data"), some with multiple errors, and using them rigorously is one of the challenges of today's molecular biology. Moreover, in the face of the skepticism raised by any sequencing result, the minimal demands of scientific article reviewers are such that more and more researchers are now obliged to present longer sequence results to ensure the seriousness of their raw results [21].
In this article, we have discussed mutations and epimutations among the unintended effects of these modern biotechnologies. It can therefore be observed that claiming to detect and eliminate all the unintended effects of new techniques is more a matter of faith than of established science. Above all, in view of this observation of potential effects, the limits of backcrossing and those of sequencing, one may doubt in good faith the ability of breeders to eliminate all off-target effects and then to detect efficiently and rapidly the remaining mutations and epimotations by sequencing. It is therefore surprising that the Scientific Committee of the High Council for Biotechnology (HCB) is silent on all these points in its first report published in February 2016 on the impacts of new genetic modification techniques.
Eric Meunier, Inf'OGM
1] "Stress induces plant somatic cells to acquire some features of stem cells accompanied by selective chromatin reorganization", Florentin, A. et al (2013), Developmental Dynamics, 242(10), 1121-1133 ;
"Developmental stage specificity and the role of mitochondrial metabolism in the response of Arabidopsis leaves to prolonged mild osmotic stress," Skirycz, A. et al. 2010. Plant Physiology, 152(1), 226-244 ;
"Arabidopsis mesophyll protoplasts: a versatile cell system for transient gene expression analysis", Yoo, S.-D. et al. (2007). Nat. Protocols, 2(7), 1565-1572.
2] "Cell culture-induced gradual and frequent epigenetic reprogramming of invertedly repeated tobacco transgene epialleles", Krizova, K. et al. 2009. Plant Physiology, 149(3), 1493-1504 ;
"Extended metAFLP approach in studies of tissue culture induced variation (TCIV) in triticale", Machczyńska, J. et al, (2014). Molecular Breeding, 34(3), 845-854 ;
"Tissue culture-induced novel epialleles of a Myb transcription factor encoded by pericarp color1 in maize, Rhee, Y. et al. 2010. Genetics, 186(3), 843-855 ;
"Transformation-induced mutations in transgenic plants: analysis and biosafety implications, Wilson, A.K. et al., (2006). Biotechnol Genet Eng Rev, 23(1), 209-238 ;
"A whole-genome analysis of a transgenic rice seed-based edible vaccine against cedar pollen allergy", Kawakatsu, T. et al., (2013) . DNA Research 20, 623-631 ;
"Recent progress in the understanding of tissue culture-induced genome level changes in plants and potential applications", Neelakandan et al. 2012. Plant Cell Reports, 31(4), 597-620
3] "Meiotic transmission of epigenetic changes in the cell-division factor requirement of plant cells", Meins, F. et al. 2003. Development, 130(25), 6201-6208.
4] "Cell biology: delivering tough cargo into cells", Marx, V. (2016). Nat Meth, 13(1), 37-40.
5] "Only half the transcriptomic differences between resistant genetically modified and conventional rice are associated with the transgene", Montero, M. et al, (2011). Plant Biotechnology Journal, 9(6), 693-702.
6] HCB SC Report, page 50
7] "Less is more: strategies to remove marker genes from transgenic plants", Yau, Y.Y. et al, (2013), BMC Biotechnology.
8] "Alternatives to Antibiotic Resistance Marker Genes for In Vitro Selection of Genetically Modified Plants - Scientific Developments, Current Use, Operational Access and Biosafety Considerations", Breyer et al (2014) Critical Reviews in Plant Sciences, Vol 33, Issue 4, 286-330.
"Suitability of non-lethal marker and marker-free systems for development of transgenic crop plants: present status and future prospects", Manimaran et al (2011) Biotechnol Adv. 29(6), 703-14
"Effects of antibiotics on suppression of Agrobacterium tumefaciens and plant regeneration from wheat embryo", Han, S-N. et al, (2004), Journal of Crop Science and Biotechnology 10, 92-98.
9] Some of these systems may persist as extra-chromosomal circular elements for several generations (e.g. wheat).
10] that can be used to identify (e.g. transposon excision fingerprint, recombinations) the transformation event.
11] Biotechnology 13, ibid.
12] "Recent progress in the understanding of tissue culture-induced genome level changes in plants and potential applications", Neelakandan, A.K. et al, (2012), Plant Cell Reports 31, 597-620.
"Regeneration in plants and animals : dedifferentiation, transdifferentiation, or just differentiation ? "Sugimoto, K. et al, (2011), Trends in Cell Biology 21, 212-218.
13] " Distortions de ségrégation et amélioration génétique des plantes (synthèse bibliographique) ", Diouf, F.B.H. et al , (2012), Biotechnologie Agronomie Société Et Environnement, 16(4), 499-508
"Quantitative trait locus mapping can benefit from segregation distortion", Xu, S. (2008), Genetics, 180(4), 2201-2208.
"Genetic map construction and detection of genetic loci underlying segregation distortion in an intraspecific cross of Populus deltoides", Zhou, W et al, (2015), PLoS ONE, 10(5), e0126077
14] "Next generation sequencing technology: Advances and applications", Buermans, H.P.J. et al, (2014), Bichimica and Biophysica Acta (BBA) - Molecular Basis of Disease, 1842(10), 1932-1941.
"Next-generation sequencing platforms, Mardis, E.R. (2013), Annual Review of Analytical Chemistry, 6(1), 287-303
"Applications of next-generation sequencing. Sequencing technologies - the next generation", Metzker, M.L. (2010), Nature Reviews Genetics, 11(1), 31-46.
15] "Next-generation sequence assembly: four stages of data processing and computational challenges", El-Metwally, S. et al, (2013), PLoS Comput Biol 9, e1003345.
"Systematic comparison of variant calling pipelines using gold standard personal exome variants", Hwang, S. et al, (2015), Scientific reports 5, 17875.
"Sequence assembly demystified," Nagarajan, N. et al, (2013), Nat Rev Genet 14, 157-167." Improving the quality of genome, protein sequence, and taxonomy databases: a prerequisite for microbiome meta-omics 2.0", Pible, O. et al, (2015). Proteomics 15, 3418-3423
"Theoretical analysis of mutation hotspots and their DNA sequence context specificity", Rogozin, I.B. et al, (2003), Mutation Research/Reviews in Mutation Research 544, 65-85.
16] "Sequencing technologies and tools for short tandem repeat variation detection", Cao, M.D. et al, (2014), Briefings in Bioinformatics.
17] "Open chromatin reveals the functional maize genome", Rodgers-Melnick, E. et al, (2016). Proceedings of the National Academy of Sciences 113, E3177-E3184
"Evolutionary patterns of genic DNA methylation vary across land plants", Takuno, S. et al, (2016), Nature Plants 2, 15222.
18] "Systematic comparison of variant calling pipelines using gold standard personal exome variants", Hwang, S., et al, (2015), Scientific reports 5, 17875.
"Principles and challenges of genome-wide DNA methylation analysis", Laird, P.W. (2010), Nature Reviews Genetics 11, 191-203.
"Performance comparison of whole-genome sequencing platforms", Lam, H.Y.K. et al (2012), Nat Biotech 30, 78-82.
"Low concordance of multiple variant-calling pipelines: practical implications for exome and genome sequencing. O'Rawe, J. et al, (2013), Genome Medicine 5, 1-18.
19] "Next-generation sequence assembly: four stages of data processing and computational challenges", El-Metwally, S. et al, (2013), PLoS Comput Biol 9, e1003345.
20] "Rapid evaluation and quality control of next generation sequencing data with FaQCs", Lo, C.-C. Et al, (2014), BMC Bioinformatics 15, 1-8
21] "Droplet barcoding for massively parallel single-molecule deep sequencing", Lan, F. et al, (2016), Nat Commun 7
