1. Introduction
In recent years, biotechnologists have begun to employ genome editing methods such as ODM (oligonucleotide-directed mutagenesis), CRISPR/Cas (clustered regularly interspaced short palindromic repeats/CRISPR associated protein), TALEN (transcription activator-like effector nuclease) and ZFN (zinc finger nuclease) to modify the characteristics of organisms important in production of food and feed. At present, two genome-edited plants have been commercialized in North America, a herbicide-tolerant canola variety [
1] and a soy bean variety with modified oil composition [
2].
The regulatory landscape regarding genome-edited crops is currently not well defined globally. Despite the commercialization of the above two genome-edited crops and recent policy changes by the US Department of Agriculture, the US has not formalized a comprehensive position on the regulation of genome editing methods. A handful of other countries have established laws or regulations regarding genome-edited plants [
3]. The European Court of Justice has ruled [
4] that crops modified by directed mutagenesis fall within the scope of Directive 2001/18/EC on the release of GMOs into the environment [
5]. In practice, this means that genome-edited crops are regulated according to that Directive [
5]. Preceding and following this decision, there has been sustained discussion with wide-ranging perspectives regarding the regulation of these methods among EU [
6,
7,
8] and member state [
9,
10,
11,
12] government representatives, within the Convention on Biodiversity [
13,
14], and among representatives of the biotechnology industry [
15,
16] and the academic community [
3,
17,
18,
19,
20,
21,
22,
23].
Although the precision of genome editing and range of off-target and other unintended effects, especially in comparison with random mutagenesis, have been one focus of the discussion regarding regulatory status of genome-edited organisms [
3,
19,
22,
23,
24,
25], a second, prominent focus has been feasibility of developing methods for detecting genome-edited organisms [
3,
7,
10,
25,
26,
27,
28,
29]. One of the arguments put forward to justify regulating genome-edited crops differently from recombinant DNA-based GMOs is that Directive 2001/18/EC requires analysis-based surveillance of GMOs, while, it is claimed, there are many technical and regulatory challenges that make development of GMO regulation-compliant analytical identification and quantitation methods for genome-edited organisms difficult or even impossible [
7,
10,
25,
26,
28,
29]. However, these claims are controversial [
3,
27].
To date, discussions regarding detection and quantification of genome-edited organisms have remained mostly on the theoretical level [
28]. To provide an empirical basis for the discussion, we have begun to explore these questions in the laboratory and have successfully developed a real-time quantitative PCR (qPCR) method for identification and quantitation of the first genome-edited crop commercialized, an herbicide-tolerant canola. This method is fully compliant with regulations for monitoring of GMOs in the EU and similar jurisdictions. A single base pair edit in the canola
AHAS1C gene has rendered the resulting gene-product tolerant to sulfonylurea and imidazolinone herbicides [
30]. We deemed this product a worthy test of the question of whether methods could be developed for identification and quantitation of genome-edited crops, since this product embodies many of the more challenging features that genome-edited crops may present, such as the fact that: (1) the edit consists of a single base pair modification and, (2) while the edit has been carried out on one member of a multigene family (
AHAS1C), (3) a second member of the family (
AHAS3A) carries the same single base pair modification generated via chemical mutagenesis. Thus, it was necessary to design a method capable of distinguishing between these two.
Using standard qPCR methodology, in conjunction with the use of locked nucleic acids (LNAs) in primer design, we were able to develop a high-sensitivity quantitative detection method for the single-nucleotide genome edit described above. This approach is likely to apply to other single-nucleotide edits and, since the scientific community has two decades of experience successfully using standard qPCR methods to quantitatively detect indels and gene insertions, this is an approach to method development that may be applicable to all classes of genome-edited crop for which at least minimal construct information is available. This is an important conclusion, because it establishes a clear path forward for analysis-based regulation of genome-edited organisms that is fully compliant with current GMO regulations in jurisdictions such as the EU [
5,
31,
32] and fully consistent with existing practices, workflows, and expertise available in contemporary GMO analytical laboratories. This paper describes the development and independent validation of this method, which has been placed in the public domain, available for use by all laboratories.
4. Discussion
We present in this paper the first experimental evidence addressing the feasibility of developing GMO regulation-compliant analytical methods for identification and quantitation of genome-edited plant material.
The recent report by ENGL [
28] as well as a number of other articles (for instance [
7,
10,
25,
29]) have raised doubts regarding whether it is possible to create accurate and sensitive PCR methods for detection and quantitation of genome-edited plant materials that meet requirements for testing methods set out in the GMO regulations of the EU and similar jurisdictions. Our work provides a definitive answer to this question, demonstrating that highly sensitive and specific PCR methods that meet GMO regulatory requirements can be developed to detect and quantify edited organisms even when the genome edit consists of only a single base pair alteration, and even in cases of multicopy gene targets. This can be achieved even in complex allotetraploid genomes such as that of canola.
Based on the success with this challenging case of genomic editing, we are optimistic that standard qPCR, augmented with strategies such as the use of LNAs in primer design, will prove capable of delivering quantitative detection methods that meet GMO regulatory standards for all classes of genome-edited organisms, including single-nucleotide edits, as well as the indels and larger inserts for which qPCR methods are well established and in use for quantitative detection of GMOs created using recombinant DNA methods. Factors such as genome size and the possibly constrained sequence context of a given SNV (challenges such as the SNV being imbedded in an AT-rich, repetitive or primer–dimer-forming region) can also influence the sensitivity and specificity of the tests that can be developed for a given genome-edited event. However, a number of strategies are available for optimizing performance of the primer-probe set for a given SNV. In the present case, it was incorporation of an LNA residue into the reverse primer that led to a significant increase in specificity of the AHAS1C SU primer-probe system. Other strategies may be successful in other situations. Adjusting primer position by just a few bases can increase specificity by as much as 7 or 8 Ct, and targeting GC-rich regions as primer sites, incorporating mis-matches into primer sequence and use of minor groove binding for probes are additional strategies that have yielded practical success.
Such methods are compatible with the basic standards of practice, equipment and molecular biological expertise found in most regulatory and commercial GMO laboratories, and could be readily integrated into the analytical routine and infrastructure of these laboratories. They also meet the current ENGL requirements [
42] for GMO detection and quantitation methods.
Because of the complexity of the canola
AHAS gene family, the method that we developed for specific detection of SU canola relied on the presence of two SNVs, one at the site of genome editing, and a second that distinguished the
AHAS1C gene from the
AHAS3A gene. However, in most cases, targeting a single SNV at a defined location in the genome of the genome-edited organism will be sufficient to achieve definitive detection and quantitation of the genome-edited event. This was not the case for SU canola because it carried an
AHAS3C gene with the same G-to-T alteration (brought about by chemical mutagenesis is) that was effected by genome editing in the
AHAS1C gene of SU canola. There may be other infrequent exceptions in which a combination of markers may be needed, depending on the specifics of the gene of interest and its genomic context. In each case, it will be necessary to identify gene and genomic features that are distinctive for the genome-edited event of interest and design the test around those features. It is ideal for a method to focus on one SNV, if possible, since this makes it possible to minimize amplicon length. In the case of SU canola, the amplicon length turned out to be 334 bases, which carries the disadvantage that, for samples in which the DNA has been partially degraded by food processing, sensitivity will be reduced as amplicon length increases [
43].
In the case of deletion or insertion of nucleotides via genome editing techniques, developing definitive qPCR detection methods will be even more straightforward than for single base pair edits. For instance, in the case of the Calyxt high-oleic soy, the TALEN nuclease effected deletions of approximately 4 to 63 base pairs at defined positions within soy fatty acid desaturase genes. Because the length of deletions cannot be precisely controlled using the TALEN method, the exact sequence at the deletion site is randomly determined and therefore unique since the probability of generating, by any mechanism, an identical deletion of exactly the same length is extremely low. Thus, a primer that spans the deletion point will uniquely differentiate that event.
At present, the model that most laboratories employ in GMO surveillance is called the matrix approach, where a series of sequence features common to many GMOs, such as the CaMV P-35S promoter and the T-nos terminator are targeted, along with event-specific targets for GMOs that do not carry common sequences [
44,
45]. The result is a minimal set of PCR targets that, together, are capable of efficiently detecting a broad range of authorized GMOs and known-unauthorized GMOs. To incorporate genome-edited GMOs into this scheme is straightforward: the sequence target or targets for each genome-edited GMO are incorporated into the screening matrix along with target sequences for other GMOs that do not carry common sequences. The number of genetically modified crops lacking common sequences has been increasing in recent years, and genome-edited crops will add to this number. Thus, screening will require an increasing number of primer-probe sets in the future. This could become economically unsustainable, if the field continues to rely on real-time qPCR. However, the use of next-generation sequencing technology’s capacity to carry out massively parallel multiplexed quantitative amplicon sequencing provides a viable alternative, although it will require significant development effort. This approach has already been applied qualitatively in GMO screening [
46]. Fields such as microbiological ecology have already applied this approach quantitatively [
47].
Discussions regarding the ability to develop methods for detection and quantitation of genome-edited crops often focus on the challenge of definitively demonstrating whether a particular genome modification is the result of genome editing or, for instance, chemical mutagenesis. However, a close reading of Directive 2001/18/EC [
5] and Commission Implementing Regulation (EU) No 503/2013 [
31] indicates the capacity for such discrimination is not necessary to meet the requirements for marketplace surveillance of GMOs, established in European Union law:
“The method(s) shall be specific to the transformation event (hereafter referred to as ‘event-specific’) and thus shall only be functional with the genetically modified organism or genetically modified based product considered and shall not be functional if applied to other transformation events already authorized; otherwise the method cannot be applied for unequivocal detection/identification/quantification.”—Section 3.1.C.1 of Annex III to Regulation (EU) No 503/2013
This passage from the EU Commission’s implementing regulation clearly does not require that GMO analytical methods must be able to ascertain the process by which a particular GMO was created. It only requires the analytical method to be able to distinguish the modified event from other authorized events in the marketplace. Language similar to that quoted above from Regulation (EU) No 503/2013 is also found in the ENGL document, Definition of Minimum Performance Requirements for Analytical Methods of GMO Testing [
42]. Based on these authoritative sources, it is clear that methods such as the one reported in this paper for SU canola meet the event specificity requirements for GMO analytical methods under EU regulations.
We based development of the detection and quantitation method for SU canola on information available in the public domain, referenced earlier in this paper. More broadly such information would include GMO-related patents, documents related to governmental regulatory approvals, communications to investors and market information relevant to commercialization. This information is accessible to any interested party so that it can be systematically monitored and used for development of test methods. It is unlikely that a commercial product could enter the global food system without sufficient information being released into the public domain to adequately inform the method development process. In certain jurisdictions, such as the EU, information related to detection methods is available through another route, as well; GMO regulatory law requires commercial entities, seeking market authorization for genetically modified food products, to provide a product-specific detection and quantitation method. They are also required to make available reference material. Since the European Court of Justice has already established that genome-edited products fall within the scope of the EU GMO regulation [
4], methods for detection and quantitation of genome-edited products must be made available by the developer as part of the market authorization process for every genome-edited product.
A final issue raised in discussions regarding methods for detection and quantitation of genome-edited crops is challenges related to detection of unauthorized genome-edited crops. In the past, commodities have been screened for unauthorized GMOs using tests that detect common sequence elements, such as the CaMV P-35S promoter and the T-nos terminator that have been used in the construction of many GMOs. It has been argued that genome-edited products are challenging because they do not carry these common sequences and therefore unauthorized genome-edited crops could not be detected using these broad screening methods [
28]. It should be pointed out that this is not a limitation exclusive to genome-edited products. As discussed at length in a report from ENGL [
28] and elsewhere [
44,
45], it is not difficult, using even recombinant DNA methods, to develop GMOs that are free from common sequence elements. Consequently, many such GMOs have been commercialized, and it is quite possible that there are unapproved GMOs in the marketplace, even now, that have not been detected because they do not carry any common sequences. Thus, unauthorized genome-edited products represent just one new class of unauthorized genetically modified products, among others, that cannot be detected by using the existing screening strategy. Screening for unauthorized GMOs is not, and never has been, an exhaustive process, and the presence of genome-edited products in the commercial food system does not create a new set of circumstances that demands fundamental changes in the regulatory regime for GMOs.
Although our work demonstrates that it may be possible to develop event-specific, GMO regulation-compliant detection methods for virtually any gene-edited organism based on information disclosed by the developer or gathered from the public domain, screening methods capable of detecting whole classes of gene-edited products, such as CRISPR or TALEN, would also be a useful addition to the screening matrix. CRISPR and TALEN modified organisms are often transgenic due to the incorporation into the host genome of the genetic apparatus that synthesizes the sequence- specific nuclease involved. Demorest et al. [
48], as an example, reported that 80% of the TALEN events that they generated were transgenic. Screening methods for CRISPR or TALEN could, for instance, rely on retained transgenic fragments unintentionally left in the final plants’ genomes. There are attempts to select events that do not carry these sequences [
49] or to use strategies that do not require such expression vectors [
50], but there is quite limited empirical evidence at this time demonstrating the success and generalizability of these efforts. We expect that research in coming years will lead to screening methods for at least some, if not all, categories of genome-edited products.
5. Conclusions
We have developed a sensitive, GMO regulation-compliant method for detecting the first genome-edited crop to be commercialized and suggest that it may represent a general approach for detecting genome-edited organisms. This effort was intended to address the concern voiced by several [
7,
10,
25,
26,
28,
29] that it may not be possible to develop quantitative detection methods for genome-edited plant materials that meet GMO regulatory requirements in jurisdictions such as the EU. In particular, they questioned the feasibility of developing quantitative methods that target single-nucleotide edits and that are sufficiently sensitive and specific to meet the requirements of EU law and regulations. The research presented here provides a clear answer to these concerns, showing that straightforward qPCR methods can be developed for single-nucleotide edits (SNVs). This, in conjunction with the fact that the scientific community has been using qPCR to quantitatively detect indels and inserted genes for two decades, indicates that it may be possible to develop qPCR methods for virtually any genome edit.
Our approach does not rely on specialized point mutation detection procedures, but employs straightforward real-time qPCR methods that are compatible with the basic standards of practice, equipment and molecular biological expertise found in most regulatory and commercial GMO testing laboratories. The resulting methods can readily integrate into the analytical routine of the typical regulatory or commercial GMO testing laboratory, including the matrix approach. The SU canola assay has been independently validated and fully meets the requirements for GMO testing methods laid down in European Union law and regulations and has been placed in the public domain, accessible to all laboratories.
This work establishes the basis for developing a test-based strategy for monitoring genome-edited plant products that integrates seamlessly into the same strategy that is used today in the EU and most other countries to monitor and regulate GMOs developed through recombinant DNA methods. Such an approach will deliver the transparency that consumers are increasingly demanding for the food that they provide to their families, and will, if necessary, provide the post-market traceability needed in case of unintended biosafety, environmental or socio-economic impacts.