DISCUSSION
Our results show that the CRISPR-Cas9 system can be used to induce DSBs in all three species studied here, when desired, after transformation of cells with the same plasmid in which the appropriate gRNA sequence is cloned. All species here are haploid, and since we did not provide an intact template for homologous repair of the cut-site in the ADE2 gene, repair must occur through NHEJ. Our red/white screening strategy, in parallel with monitoring of cell death on plates and in time-course experiments, allows us to estimate the efficiency of cut and repair in each species. Repair by faithful NHEJ results in a regenerated, intact, cut-site, and thus cutting by CRISPR-Cas9 can occur once more. If cells do not die from this, through cell-cycle checkpoint arrest, or through degradation of chromosomal DNA at some point, they can undergo a futile cut-ligate-recut cycle (Maroc and Fairhead, 2019). Escaping this potential futile cycle can occur through plasmid mutation/rearrangement, or through plasmid loss, accompanied by integration of the URA3 gene (whose presence is selected in the medium), and finally through events of repair by unfaithful NHEJ that modify the cut-site recognition sequence. In our experiments targeting the ADE2 gene, we observe a very high percentage of red colonies in survivors, indicating a very efficient chromosomal cutting followed by unfaithful NHEJ.
In S. cerevisiae , a CRISPR-Cas9-created chromosomal cut, in a haploid strain in a unique region is lethal for most cells (Yarrington et al., 2018). In C. bracarensis and C. nivariensis , even though survival is higher than in the model yeast, only 17%-34% of cells survive continuous induction on plates. In temporary liquid medium induction, we observe that survival drops with increasing exposure and that long exposure results in survival rates similar to induction on plates. The percentage of cells repaired by unfaithful NHEJ (red colonies) increases with exposure time, reflecting either the time course of cutting by CRISPR-Cas9, or the prevalence of faithful NHEJ at the beginning of exposure, gradually replaced by unfaithful NHEJ events which are resistant to further cutting.
It is remarkable that C. glabrata has a much higher survival rate than the other two Candida . We have previously determined that cells of C. glabrata undergo futile cycles of cut-religation and that the cell-cycle chekpoint must be “leaky” (Maroc and Fairhead, 2019). The situation may well be different in the other two otherCandida studied here. Another, non exclusive, reason for lower survival may be that unfaithful NHEJ is much less efficient in the other two species than in C. glabrata , preventing the creation of non-cuttable ade2 - alleles in the former. On the other hand, perhaps the cell cycle checkpoints are less ”leaky” in C. bracarensis and C. nivariensis , and cells with unrepaired DSBs die more frequently.
Sequencing of the new junctions created by this type of NHEJ, i.e. from red colonies, shows that cells have repaired the DSB with varying amounts of rearrangements, that differ between species. There is more variation in the types of unfaithful NHEJ repair events in C. bracarensis and C. nivariensis than in C. glabrata , where survival is also much higher. Perhaps an efficient and reproducible unfaithful NHEJ mechanism in C. glabrata allows this survival rate, accompanied by leakiness of checkpoints. This may signify different adaptation mechanisms, and remains to be investigated thoroughly.