mRNA processing in mutant zebrafish lines generated by chemical and CRISPR-mediated mutagenesis produces unexpected transcripts that escape nonsense-mediated decay

Abstract

As model organism-based research shifts from forward to reverse genetics approaches, largely due to the ease of genome editing technology, a low frequency of abnormal phenotypes is being observed in lines with mutations predicted to lead to deleterious effects on the encoded protein. In zebrafish, this low frequency is in part explained by compensation by genes of redundant or similar function, often resulting from the additional round of teleost-specific whole genome duplication within vertebrates. Here we offer additional explanations for the low frequency of mutant phenotypes. We analyzed mRNA processing in seven zebrafish lines with mutations expected to disrupt gene function, generated by CRISPR/Cas9 or ENU mutagenesis methods. Five of the seven lines showed evidence of altered mRNA processing: one through a skipped exon that did not lead to a frame shift, one through nonsense-associated splicing that did not lead to a frame shift, and three through the use of cryptic splice sites. These results highlight the need for a methodical analysis of the mRNA produced in mutant lines before making conclusions or embarking on studies that assume loss of function as a result of a given genomic change. Furthermore, recognition of the types of adaptations that can occur may inform the strategies of mutant generation.

Fig 1. An ENU-induced G>T mutation in the 3’ essential splice site (ESS) of i33–34 of abca1b leads to a skipped exon and an early termination signal.

Analysis of cDNA sequence and agarose gel electrophoresis (S1 Fig) indicates that loss of exon 34 (116 bases) causes a frame shift, followed by a short ORF (13 AA) and an early termination codon (shown in red). qPCR studies revealed transcript levels down 3.5-fold in 6-dpf abca1b^{sa18382/sa18382} zebrafish. See S1 Table for sequence spanning mutation and predicted outcome.

Fig 2. An ENU-induced G>A mutation in the 5’ essential splice site (ESS) of i2–3 of slc27a2a leads to a skipped exon.

Analysis of cDNA sequence and agarose gel electrophoresis (S1 Fig) reveals a loss of exon 2 (70 AA) with no frame shift. By qPCR, transcript levels of 6-dpf slc27a2a^{sa30701/sa30701} zebrafish were not found to be different than their siblings. See S1 Table for sequence spanning mutation and predicted outcome.

Fig 3. An ENU-induced G>A mutation in the 3’ essential splice site (ESS) of i29–30 of abca1a leads to use of a nearby cryptic splice site and loss of a single AA.

The base change causes a missed splice acceptor and sequence immediately following the mutated base to be used as a cryptic splice site, as confirmed by analysis of cDNA sequence. A single Serine is lost (boxed in green) and the product remains in frame. By qPCR, transcript levels of 6-dpf abca1a^{sa9624/sa9624} zebrafish were not found to be different than their siblings. See S2 Fig for agarose gel electrophoresis of amplified cDNA and S1 Table for sequence spanning mutation and predicted outcome.

Fig 4. An ENU-induced G>A mutation in the 5’ essential splice site (ESS) of i10–11 of cd36 leads to use of a cryptic splice site and a frame shift.

The base change causes loss of a splice donor and use of a cryptic splice site 3 and 4 bases downstream, as confirmed by analysis of cDNA sequence. The intronic sequence “ATAT” preceding the cryptic splice site is thus incorporated, leading to a frame shift. After the frame shift, 18 AA follow before a stop codon (shown in red) directs early termination and loss of exon 12 (154 AA). By qPCR, transcript levels of 6-dpf cd36^{sa9388/sa9388} zebrafish were not found to be different than their siblings. See S2 Fig for agarose gel electrophoresis of amplified cDNA and S1 Table for sequence spanning mutation and predicted outcome.

Fig 5. An ENU-induced C>T nonsense mutation in exon 2 of creb3l3a leads to a skipped exon.

Analysis of cDNA sequence reveals a loss of exon 2 (38 AA) with no frame shift. Agarose gel electrophoresis of amplified cDNA (from pooled intestines) revealed an additional band in the homozygous mutants matching the expected size of a product with a skipped exon 2. By qPCR, transcript levels of 6-dpf creb3l3a^{sa18218/sa18218} zebrafish were not found to be different than their siblings. See S1 Table for sequence spanning mutation and predicted outcome. Band sizes (number of bases) for the ladder is indicated.

Fig 6. A CRISPR-induced deletion (7 bp) in exon 3 of smyd1a corresponds to use of cryptic splice sites in exon 2 in mutant clones.

A. A 7-bp deletion in exon 3 (boxed in green) is predicted to lead to a frame shift mutation and early termination. Of 20 mutant clones sequenced, all 20 had the expected 7-bp deletion. B. In addition, 3 revealed use of one cryptic splice site and another 3 revealed use of a second cryptic splice site in exon 2, leading to a 13- and 40-bp deletion, respectively. Both deletions result in a frame shift and premature termination codon, shown in red text. 20 randomly selected wt clones did not show alternative splicing. By qPCR, transcript levels of 1- and 2-dpf smyd1a^mb4/mb4 zebrafish were down thirteen-fold compared to wildtype siblings. See S1 Table for sequence spanning mutation and predicted outcome.