A few weeks ago Antonio Giraldez’s lab posted a new preprint on BioRxiv. The manuscript describes work by Miguel Moreno-Mateos and colleagues, including collaborators from Jennifer Doudna’s group, on the implementation of Cpf1 genome engineering in zebrafish and Xenopus. Cpf1 is an alternative RNA-guided endonuclease to Cas9, which has a number of distinct features that make it an interesting tool for genome engineering applications. When Miguel’s manuscript popped up in my Twitter feed it immediately caught my attention. I had previously tried to adopt the Cpf1 system for genome editing in Drosophila, but with little success. In work published last year (preprint, paper) we had tried to edit three genes with in total four crRNAs, but only succeeded at one locus. Such a low success rate is uncharacteristic for Cas9, where we have previously shown that 65 out of 66 randomly chosen transgenic Cas9 sgRNAs are active. Moreover, the activity of the one active Cpf1/crRNA complex, targeting the pigmentation gene ebony, was less than impressive and well below what we typically see for transgenic Cas9 experiments. Our overall conclusion from these experiments was that while Cpf1 was active in Drosophila, it seemed to be much less robust than Cas9. Hence we decided to continue to focus our work on Cas9 and the Cpf1 transgenic fly strain we made began an entirely unperpurbed life in the 18C room. Until I read Miguel’s preprint.
He and his colleagues had noticed that the accounts of the use of Cpf1 for genome engineering in the literature were quite heterogenous. While some also struggled to get Cpf1 to work as robustly as Cas9 (e.g. here, here, here), others had much more success (e.g. here, here, here). One systematic difference between the studies that reported high levels of Cpf1 activity and those that were less successful appeared to be the temperature at which these experiments were carried out. While we and others working in plants performed our experiments at 25C, the successful studies used cells cultured at 37C. This gave rise to the hypothesis that temperature is a key factor controlling the activity of Cpf1. Consistent with this idea the Girladez lab found that Cpf1, and in particular Cpf1 from Acidaminococcus sp BV3L6 (AsCpf1), the variant we had used for our experiments in the fly, functioned poorly at 25C, but showed high activity at temperatures above 30C.
After reading the preprint I went straight to the fly room to search for our AsCpf1 transgenic line. We had also made transgenic stocks of different crRNAs and so I could simply set up genetic crosses to test the activity of the Cpf1 system again. This time I set up each cross in duplicate and kept one vial at 25C (as we had done previously) and the other one at 29C. For good measure I also incubated the “high temperture” cross for 6 hours at 34C, when most of the offspring was in the early first instar larval stage (Fig. 1A). Afterwards I returned them to the 29C incubator and reared them to adulthood. When the offspring that was transgenic for both AsCpf1 and the crRNA targeting ebony eclosed from the pupal case it became apparent that the flies that developed at a higher temperature were noticable darker (Fig. 1B). This suggested that more cells had undergone biallelic mutagenesis of ebony, as mutations in this gene lead to darker coloration of the cuticle.
I then tested if the flies reared at higher temperature also passed on more non-functional ebony alleles to their offspring. To make my life easier I did this by phenotypic complementation with an existing ebony allele. I crossed act-AsCpf1/U6:3-t:crRNA-e flies to a common ebony mutant lab stock (w;;TM3/TM6b) and analysed how many of their offspring displayed a dark cuticle, indicating that they had inherited a non-functional ebony allele from their Cpf1 parent. While act-AsCpf1/U6:3-t:crRNA-e flies that developed at 25C passed on non-functional ebony alleles to less than 10% of their offspring, similar to what we had reported last year, this number was much higher when the flies were reared at higher tempertures (Fig.1C, D). Flies that were reared at 29C and had experienced a brief pulse of 34C in early development passed on non-functional ebony alleles to more than 80% of their offspring. This dramatic increase is consistent with the finding reported in the Giraldez lab preprint and suggest that also in Drosophila temperature has a strong influence on the activity of AsCpf1.
So is rearing flies at a higher temperature the key to transform AsCpf1 into a robust genome engineering tool in Drosophila? Maybe, but to answer this question more experiments testing the performance of AsCpf1 at different loci are required. Notably, I also performed the experiment described above with two other crRNAs, targeting the genes yellow and sepia. These crRNAs had previously not resulted in any mutagenesis at the target locus at 25C. These crRNA still did not result in detectable mutant phenotypes in AsCpf1/crRNA animals even when these were raised at the high temperature regime. It remains to be seen whether inactive crRNAs are a common feature of AsCpf1 in Drosophila or if these two crRNAs are outliers. It should be noted, however, that both yellow and sepia are genes that are extremely susceptible to Cas9 mediated mutagenesis.
Another exciting finding reported in the preprint is that Cpf1 from another bacterial species, Lachnospiraceae bacterium ND2006, shows higher activity and is less influenced by temperture in zebrafish than AsCpf1. In our previous work we did not test the LbCpf1 variant, as experiments in human cells showed that AsCpf1 and LbCpf1 have comparable activity. With the new data from the fish in hand we will now proceed to test the activity of LbCpf1 in Drosophila.
I would like to thank the authors of the preprint for making their manuscript available on BioRxiv before formal publication. The novel insights reported in this study are very interesting to us and will likely influence our future work. Making them available as a preprint allows us to start building on their discovery already now and not only once the paper has gone through peer review at a conventional science journal. This really highlights one of most profound advantages of the ongoing adoption of preprints in biology: Making work accessible at an early timepoint speeds up scientific progress. I hope this post also highlights another potential benefit of posting preprints. Namely the ability of the authors to receive feedback on their work from the community. In this case the availability of the preprint allowed me to test one of the central claims of this study and demonstrate that the result is reproducible in another experimental system.