Summary of methods for generating stable and transient zebrafish models of genetic human diseases
| Technique | Mechanism | Required components | Advantages for modelling rare genetic diseases in zebrafish | Limitations for modelling rare genetic diseases in zebrafish |
| Stable models | ||||
| ENU-mediated mutagenesis | Random mutagenesis | Database of mutants that have already been generated by ENU-mediated random mutagenesis | If LOF mutants for genes of interest are available, this abrogates need to generate a new model | Mutants for genes of interest are not always available. Outcrossing is often required to generate a stable model. |
| Retroviral-mediated insertional mutagenesis | Random mutagenesis | Database of mutants that have already been generated by retroviral-mediated insertional mutagenesis | If LOF mutants for genes of interest are available, this abrogates need to generate a new model | Mutants for genes of interest are not always available. Outcrossing is often required to generate a stable model. |
| ZFNs+NHEJ-mediated repair | Creates double-stranded DNA break at target site, resulting in repair by NHEJ | Multiple DNA-binding zinc finger peptides (which each recognise 3 bp of target DNA) fused to FokI nuclease domain | Enables targeted frameshift mutations to be introduced in candidate genes of interest | Tailored protein component needs to be generated for each genomic target. Outcrossing is often required to generate a stable model. |
| TALENs+NHEJ-mediated repair | Creates double-stranded DNA break at target site, resulting in repair by NHEJ | Customisable DNA-binding domain (peptide-based) fused to FokI nuclease domain | Enables targeted frameshift mutations to be introduced in candidate genes of interest | Tailored protein component needs to be generated for each genomic target. Outcrossing is often required to generate a stable model. |
| CRISPR/Cas9+NHEJ-mediated repair | Creates double-stranded DNA break at target site, resulting in repair by NHEJ | Specific 20 nt guide RNA complementary to target site+Cas9 endonuclease | Enables targeted frameshift mutations to be introduced in candidate genes of interest. gRNAs can easily be designed for different targets. LOF models can be efficiently generated through NHEJ-mediated repair | Some off-target effects are possible but can be minimised through appropriate gRNA design. Outcrossing is often required to generate a stable model. |
| ZFNs, TALENs or CRISPR/Cas9+HDR-mediated repair | Creates double-stranded DNA break at target site. Simultaneous addition of DNA repair template results in HDR and incorporation of specific sequences or mutations of interest | Zinc finger peptides or customisable peptide-based DNA-binding domain fused to FokI nuclease domain, or specific ~20 nt guide RNA complementary to target site+Cas9 endonuclease (+DNA repair template containing sequence of interest) | Allows knock-in of specific mutations of interest (most commonly via CRISPR/Cas9) | Currently less efficient than NHEJ-mediated LOF model generation. Outcrossing is often required to generate a stable model. |
| CRISPR/Cas9-mediated base-editing | Deaminates cytidine or adenine bases at genomic target site, resulting in single base-pair substitutions | Specific ~20 nt guide RNA complementary to target site+catalytically inactive Cas9 (dCas9), fused to cytidine or adenine deaminase enzyme | Allows introduction of disease-relevant missense mutations arising due to C-T or A-G base substitutions | Some base-pair substitutions cannot be modelled using this approach. Evidence of efficacy in zebrafish is limited. Specificity for target site needs to be established. |
| Transient models | ||||
| Morpholino knockdown | Blocks mRNA translation or splicing (post-translational) | Synthetic 25 bp oligonucleotide | Allows for rapid examination of LOF phenotypes. Could be used to rapidly obtain evidence to support causality of LOF candidate variants | Effects are short-lasting. May be associated with significant off-target effects. Cannot be used to model gain-of-function or patient-specific mutations. |
| CRISPR/dCas9 (CRISPR interference) | Blocks transcription (and can be coupled to transcriptional activators or repressors to further control gene dosage) | Specific ~20 nt gRNA complementary to target site+dCas9 (which can be fused to a transcriptional activator or repressor) | Can be used to model both gain and LOF phenotypes. Has the potential to be used on a large scale | Currently not widely used in zebrafish. Specificity for target site needs to be established. |
Cas9, CRISPR-associated protein 9; CRISPR, clustered regularly interspaced repeats; dCas9, catalytically inactive Cas9; ENU, N-ethyl-N-nitrosourea; HDR, homology-directed repair; LOF, loss-of-function; NHEJ, non-homologous end joining; TALENs, transcription activation-like effector nucleases; ZFNs, zinc finger nucleases.