Editing experiments were performed following methods described in (10 (link)) using in vitro assembled Cas9 ribonucleoprotein complexes and the co-conversion method to isolate edits (12 (link)). Co-conversion uses co-editing of a marker locus (dpy-10) to identify animals derived from germ cells that have received Cas9 and the repair templates, reducing possible experimental noise due to variations in injection quality from animal to animal (10 (link),12 (link)). We used a ∼1/3 ratio of dpy-10/locus of interest crRNAs to maximize the recovery of desired edits among worms edited at the dpy-10 locus (10 (link)). Injection mixes contained 15.5 μM Cas9 protein, 42 μM tracrRNA, 11.8 μM dpy-10 crRNA, 0.4 μM dpy-10 repair ssODN, 29.6 μM locus of interest crRNA(s) and varying concentrations of repair templates (0.1–0.5 μM; Supplementary Table S1). For gtbp-1 replacement (Figure 4K), both 5′ and 3′end crRNAs were used at 22.2 μM each and the tracrRNA concentration was increased to 56.2 μM. Final buffer concentrations in injection mixes were 150 mM KCl, 20 mM HEPES, 1.6 mM Tris, 5% glycerol, pH 7.5–8, except for Figure 2E–G and Figure 4A and B where KCl was at 200 mM and for Figure 4K where Tris was at 2.1 mM. Injection mixes were assembled by mixing the components in the following order: Cas9 protein, KCl, HEPES pH 7.5, crRNAs, tracrRNA, ssODNs, H2O and finally PCR fragments if used.
Each injection mix was injected in the oogenic gonad of ∼20 isogenic and synchronized young adult hermaphrodites (wild-type N2 or meg-3 deletion in Figure 4L). The injected mothers were cloned to individual plates 24 h after injection. Five to six days later, broods with the highest numbers of dpy-10 edits were identified (’jackpot broods’). This step selects for broods derived from hermaphrodites that were injected successfully. For each experiment, dpy-10-edited progenies from at least three independent jackpot broods were screened for edits at the locus of interest. GFP+ edits of gtbp-1, glh-1 and pgl-1 were identified by direct inspection of adult F1 animals for GFP expression in the germline. All other edits were identified by PCR genotyping of F2 cohorts derived from cloned F1s. All edits reported were germline, heritable edits. The majority of edits were recovered in the heterozygous state in F1 progenies, but we also obtained a minority of homozygously edited F1s. These observations show that, as expected, edits are created primarily shortly after injection in the oogenic germline (the site of injection). Occasionally, however, edits are also created on paternal chromosomes, presumably in zygotes shortly after fertilization since all edits were germline edits (inherited by next generation). These observations confirm that homology-dependent repair also occurs in zygotes, using the donor templates or the previously edited maternal allele, as we have observed previously (10 (link)).