Lackner et al., in an effort to eliminate the need to synthesize multiple, unique homology based templates for Cas9-mediated homologous recombination, report the use of a universal donor template that will insert tags into potentially any genomic site in the human cell lines they tested and so enable reporter tagging of endogenous gene expression. This strategy may have widespread applications.
The advent of CRISPR/Cas9 genome engineering has opened the possibility of mutating any gene in any organism, or so the advertising says. Supplying a mix of a Cas9 (a bacterial endonuclease) and an easily synthesised guide RNA which targets cleavage to a specific genetic locus is most often all that is needed to generate indel (knockout) mutations in the genome through faulty DNA repair. Moreover by addition of a DNA template with homology to the target site, it is possible to generate site-specific insertional mutants of your choice via recombination. These can include exon/reporter fusions for gene regulation studies or (sub) cellular imaging.
When trying to scale up this insertional mutagenesis technique, the need to create unique donor templates carrying homology for each locus targeted greatly increases the complexity and expense of the work.
Lachner et al. have used a donor plasmid template consisting of a tag gene, either nanoluc or turbo GFP, flanked by identical Cas9 cleavage sites from a zebrafish gene (Tia1L). The donor plasmid also transcribes a guide RNA against this Tia1L target, which is not found in the human genome. To complete the transfection cocktail, they include a unique guide RNA for the targeted human genomic location and a source of Cas9 (and for nanoluc transfections, an antibiotic selectable marker plasmid).
The highly surprising finding of the work is that the excised DNA fragment encoding the tag, generated in vivo by Cas9 activity, is inserted into the genomic target site with relatively high efficiency and precision. The integrations appeared to be fairly well defined in that the expected cleavage sites (i.e. PAM-3) were used and in many cases were not subject to further end processing, such that the reading frame of insertion could be accurately predicted. Fascinatingly, their efforts to use identical PCR generated fragments as DNA templates did not function, suggesting that Cas9 mediated tag release is critical for efficient integration for some reason.
The majority of work was performed in haploid human cells which made characterisation of integration sites easier because of their single gene copy nature. In a limited set of experiments, they also demonstrated that the system would function in a diploid cell line, which would open the technique to a much broader research field. They also go on to suggest that the system is scalable to generate collections of lines tagged with a variety of reporters.
Utilizing this approach in insects would require several considerations to be evaluated, mostly related to the scale of work envisioned and the relative efficiency compared to homology based approaches. Even with cell lines, the downstream manual work of isolating the correct clones may limit the number of genes that could be examined at any one time, even using automation. The generic approach described found 20% of GFP expressing clones that contained the expected in frame insertions. How this compares to a homology based approach was not assayed directly, but sampling of the literature suggests that this maybe in the same ballpark (Ratz et al, 2015 Scientific Reports http://dx.doi.org/10.1038/srep09592).
The potential for off target insertions is also a consideration, and would perhaps seem more likely to occur using a generic approach. This aspect was not addressed fully in the manuscript, and would require serious examination in diploid cell lines and whole organisms.
However, if the system is found to be as efficient as homology based methods in vivo, it would considerably simplify vector construction for genome tagging. Moreover, apart from creating gene fusions, the technique could easily be adapted to create marked knockout mutations that would allow simplified fluorescence based selection.
Lackner, D.H, Carre, A., Guzzardo, P. M., Banning, C., Mangena, R., Henley, T., Oberndorfer, S., Gapp, G. V., Nijman, S. M. B., Brummelkamp, T. R., Bürckstümmer , T. 2015. A generic strategy for CRISPR-Cas9-mediated gene tagging. Nature Communications 6, Article number: 10237 doi:10.1038/ncomms10237