Zebrafish have historically been used as a model system for genetic studies, but attempts at CRISPR genome editing in zebrafish have yielded varying results. The activity of the CRISPR enzyme Cas9 has been less reliable in zebrafish than in mammalian systems. However, Hoshijima et al. recently reported a systematic study of different approaches to CRISPR-Cas9 genome editing in zebrafish. They found that the activity of Cas9 and the subsequent editing results can be greatly improved and made highly reliable by following appropriate protocols. Indeed, they have optimized CRISPR genome editing protocols for zebrafish that allow disruption of gene function at nearly 100% efficiency and even demonstrate deletion of genomic DNA more than 50 kb in length.
“This is a significant paper for the zebrafish genetics field. It demonstrates that with CRISPR Alt-R RNP methods, gene deletion functional studies can be performed on developing fish embryos injected with CRISPR reagents at the zygote stage. Biallelic knock-out occurred at such a high rate with minimal mosaicism that functional studies could be done immediately. There was no need to wait for subsequent homozygous breeding. This accelerates the pace of research and discovery.”
–Mark Behlke, Chief Scientific Officer, Integrated DNA Technologies
Hoshijima and colleagues initially tested the different formats of guide RNA molecules to improve the editing efficiency and reliability in zebrafish. These formats include single guide RNA (sgRNA) and two-part, or “duplex,” guide RNA (dgRNA). This two-part guide RNA consists of a complex formed from an annealed target-specific crRNA and a universal tracrRNA, both chemically synthesized. In addition to these formats, other variations in the guide RNA in this study included:
- Presence or absence of chemical modifications on chemically synthesized sgRNA
- Presence or absence of chemical modifications on the crRNA used in the dgRNA format
- Presence of extra (“supernumerary”) guanine (G) residues at the 5′ end of sgRNA, which are added when guides are produced by in vitro transcription (IVT)
- Absence of such supernumerary guanines on sgRNA produced by chemical synthesis
For all formats of guide RNA, the guide RNA was combined with recombinant Cas9 enzyme (Alt-R S.p. Cas9 Nuclease V3) to form ribonucleoproteins (RNPs) and then injected into fertilized zebrafish oocytes. The guide RNAs were designed to target specific genomic sites, which were afterwards analyzed for mutations. Consistently, supernumerary G residues decreased genome editing efficiency. However, both dgRNA-containing RNP (dgRNP) and sgRNA-containing RNP (sgRNP) without supernumerary Gs gave high editing efficiency, with or without chemical modifications. These results could explain some of the erratic results historically reported when researchers employed CRISPR methods with IVT-generated sgRNAs in zebrafish, since these extra G residues are almost always added to improve IVT itself. The research group chose to use chemically modified dgRNA (Alt-R CRISPR-Cas9 crRNA and tracrRNA) as the standard approach for further investigations of editing efficiency and reliability in zebrafish. The researchers tested the capacity of this approach to create null mutations with known phenotypes and to delete an entire gene or large genomic segment.
Hoshijima et al. demonstrated that using dgRNPs, they could accurately produce fish with phenotypes of known null mutants. Additionally, they were able to use 2 dgRNPs simultaneously to target the foxd3 and tfap2a genes. Targeting these redundant genes together removed all body surface pigmentation of most of the fish. The researchers went on to target 2 introns on a single chromosome. By targeting pairs of alb locus introns, the group was able to delete large (>20 kb) gene segments, as phenotypically demonstrated by loss of most eye pigmentation. The researchers also targeted embryos transgenically expressing fli1a:EGFP to see if they could induce known vasculature defects associated with loss of egfl7. They excised the gene by targeting both of its ends (a distance of 36.7 kb) and observed phenotypic abnormalities of the vasculature in over 50% of the fish. The researchers additionally performed similar experiments, deleting large genomic segments, up to 51.8 kb in length. These deletions were shown to be transmissible in the germline.
The authors concluded that by following their approaches and by using chemically synthesized guide RNA, researchers can reliably induce zebrafish to undergo genome editing, resulting in desired phenotypes. They also demonstrated that targeting 2 genomic sites often gave them much better results than targeting only 1 site, because in zebrafish, single point mutations are often phenotypically overcome by redundant genes.