CRISPR/Cas9-mediated knockout of prpf31 led to retinal morphological defects. (A) The gene structure of prpf31 and CRISPR/Cas9 target site were shown. DNA sequencing of the corresponding genomic region revealed a 134 bp deletion/18 bp insertion mutation. (B) Genotype validation by DNA electrophoresis. +/+, wild-type sibling; +/−, heterozygotes; −/−, homozygotes. (C) Relative expression of prpf31 mRNA was detected by qPCR at 36 and 48 hpf. (D) The protein level of Prpf31 was decreased in prpf31^−/− embryos at 36 and 48 hpf as detected by western blot. GAPDH was used as an internal control. (E) The morphology of bodies and eyes in prpf31^−/− embryos. No obvious defects were observed until 48 hpf. At 60 hpf, the prpf31^−/− embryos showed microphthalmia, smaller head and curved body axis. At 72 hpf, these phenotypes were further aggravated. (F) Injection of zebrafish prpf31 mRNA could rescue the developmental defects of mutant zebrafish. (G, H) Quantification of the embryonic length and eye size in the wild-type, prpf31^−/− and prpf31^−/− + mRNA rescued embryos at 36, 48, 60 and 72 hpf. n = 20 for each panel.

Deletion of Prpf31 impaired RPCs differentiation. (A) Retinal sections of WT, prpf31^−/− and prpf31^−/− mRNA-rescued embryos were immunostained with Sox2 (marker for RPCs), Islet1 (marker for neuron cells), Zpr1 (marker for cone cells), and Zpr3 (marker for rod cells) antibodies at 48 or 60 hpf. n = 8 for each panel. V, ventral side, D, dorsal side. Scale bar, 100 μm. (B) The distributions of Neurod1:EGFP (specialized neurons, upper panels) and Huc:EGFP (post-mitotic neurons, lower panels) labeled cells in whole-mount retinas from WT and prpf31^−/− transgene zebrafish. The dashed circles shown the eyes and lens respectively. n ≥ 7 for each panel; Scale bar, 100 μm. (C) In situ staining of markers for RPCs (ccnd1, vsx2) and for neural precursors (atoh7, crx1), specialized neurons (neurod1) and mature neurons (tuba1) at 48 and 60 hpf. The accumulation of RPCs and reductions of differentiated neurons are shown. Scale bar, 100 μm.

Increased apoptosis and activated p53 pathway in RPCs of prpf31^−/− zebrafish. (A) TUNEL staining showed numerous apoptotic cells in the retina, brain and posterior segment of spinal cord in prpf31 mutants at 48 hpf. Injection of wild-type prpf31 mRNAs reduced the apoptotic cells to a normal level. Dotted lines indicate the boundary of two images from the same embryo. White arrows, apoptotic signals. Scale bar, 100 μm. (B) The up-regulation of p53 pathway genes in prpf31 mutants at 36 hpf as detected by qPCR. (C) Deletion of p53 in prpf31 mutants significantly reduced cell apoptosis, but could not rescue the differentiation defects of RPCs. Scale bar, 100 μm. (D−F) The quantitative analysis of TUNEL positive cells, Sox2 positive cells and Islet1 positive cells shown in (C). n = 9 for each panel. Scale bar, 100 μm.

Prpf31 deficiency causes abnormal spindle structure and mitotic arrest. (A) Double staining of EDU (S-phase cells) and pH3 (M-phase cells) in the retinal sections of WT siblings and prpf31^−/− mutants at 36 and 48 hpf. The number of M-phase cells were significantly increased in the prpf31^−/− mutants compared with siblings, suggesting that RPCs may be arrested in M phase. White arrows, overlapping signals of EDU and pH3. Scale bars: left, 100 μm; right, 10 μm. (B) Quantification of EDU⁺, pH3⁺ and EDU⁺ / pH3⁺ cells shown in (A). n = 6 for each panel. (C) In vivo imaging of the H2A-mCherry labeled chromosomes showed the mitotic progression of RPCs at 36 and 48 hpf. The time point of nuclear envelope breakdown (NEBD) was set as the start of mitosis. Scale bar, 10 μm. (D, E) Quantification of the time from NEBD to anaphase in RPCs at 36 and 48 hpf. 15 cells from more than five embryos were observed for each group. For abnormally divided cells, the longest observation time was 150 min. Scale bar, 10 μm. (F) The spindle and nuclei of RPCs from mutants and WT siblings were stained with anti-α-tubulin (green) antibody and PI (red), respectively. The different types of spindle anomalies are displayed in the panels (Abnormal 1–4). Scale bar, 10 μm. (G) Quantitative analysis of the RPCs numbers in each of the phases of mitosis in sibling and prpf31 mutant embryos at 36 and 48 hpf.

Accumulation of DNA damage in prpf31 mutant retinas. (A) Immunofluorescence analysis using the anti-γH2AX antibody in siblings and prpf31^−/− retinas at 36, 48, and 60 hpf. Scale bar, 50 μm. (B) Quantitative analysis of the γH2AX positive cells shown in (A). (C) Alkaline comet assay showed increased DNA damage in prpf31^−/− zebrafish at 36 and 48 hpf. Scale bar, 10 μm. White arrows showed DNA with single or double strand breaks. (D) Quantitative results of 100 cells from 6 embryos in each group are shown. White arrows indicate DNA damaged cells. (E) The protein levels of Prpf31, γH2AX and p53 in siblings and prpf31^−/− zebrafish at 36, 48 and 60 hpf were detected by western blot. GAPDH was used to normalize protein loading. The black arrows indicated the corresponding protein bands.

The defective retinal development in prpf31 mutants can be rescued by injection of wild-type but not RP mutant PRPF31 mRNA. (A,B) The retina phenotype of the siblings and mutants injected with human PRPF31 wild-type mRNA or RP mutant mRNA. BF, bright field; Islet1, neurons marker; pH3, M-phase marker; TUNEL, apoptosis marker; phalloidin, labeled the lamination of the retina. (C−F) The statistical data show the numbers of specific positive cells in (A), (B). Note that wild type mRNA can significantly rescued the deficient phenotypes in mutant retina, while p.Lys257fs*277 mutant mRNA has almost no effect, the p.Arg372Glnfs*99 mutation seems to have a negative effect on siblings and mutants at 60 hpf. The number of embryos used in each assay was 6. Scale bar, 100μm.

Prpf31 modulates the alternative splicing of a subset of genes involved in DNA repair and spindle assembly. (A) Significantly overrepresented (P < 0.05, enrichment score ≥ 2) DNA repair and mitosis process- associated biological process among upregulated IR and ES DSEs as determined by GO functional enrichment analysis. The top 10 biological process affected by DS events were shown. Node size and color correspond with enrichment score and log(P) value, respectively. (B) The mRNA levels of some genes were significantly down-regulated among DNA repair and mitosis process in RPCs, as detected by RT-PCR. (C) Western blot showed the protein expression of genes with altered splicing efficiency in (B). (D) Increased proportion of transcripts with IR and/or ES among DNA repair and mitosis process in RPCs, as detected by PCR. Sizes (in bp) for major and minor mRNA isoforms (black lines). Head, tail, template cDNA was obtained from the head parts or tail parts of the embryos respectively. PSI values, marked below the DNA band; the black arrows indicate the primers used in this experiment; Hexagon, There was a premature termination codon (PTC) in the corresponding isoforms. (E) Statistical analysis presented as the mean ± SD of PSI values and ΔPSI in two types tissues at 36 and 48 hpf from three biological replicates. PSI, percent splicing in; ΔPSI, the mutant PSI value subtract the wild-type PSI value was used to evaluate the changed extent in pre-mRNAs splicing efficiency after prpf31 knockout.

Prpf31 deletion is more likely to cause the skipping of exons with shorter length and weaker 5′ splicing site. (A) The lengths of severely skipped exons between mutants and siblings (‘Changed’ group) were shorter when compared with the exons with no significant difference in PSI values (‘Unchanged’ group) or all the RefSeq annotated exons (‘Ctrl’ group). (B) No significant difference in the GC content of exons between the ‘Changed’ and ‘Unchanged’ or ‘Ctrl’ groups. (C) The splice strength of 5′SS but not 3′SS were weaker in the ‘PSI-Down’ group when compared with the exons in ‘Unchanged’ or ‘PSI-Up’ groups. ‘PSI-Up’, exons with increased PSI values in ‘Changed’ group. ‘PSI-Down’, exons with decreased PSI values in ‘Changed’ group. (D) The skipping of exon 4 in nsmce1 was shown based on the RNA-seq data. The splice strength of this 5′SS was much weaker than the average score (8.4) of ‘Unchanged’ group. Several nucleotides (marked in red) were mutated to enhance the splice strength. (E) Two minigenes containing the weak and strong 5′SS were constructed and tested in HEK293 cells. Knockdown of PRPF31 significantly suppressed the splicing of the wild-type but not the enhanced form of nsmce1-minigene. (F) Quantitative analysis of the PSI values from three independent experiments.

A schematic summary of the main findings. Prpf31 is required for the maintenance of RPCs and retinal neurogenesis in zebrafish. In the absence of Prpf31, the splicing of genes related to mitosis and DNA repair is compromised, resulting in decreased expression and abnormal functions of these genes. RPCs could not complete the mitotic progression and deal with DNA damages, and finally undergo apoptosis.

ZFIN is incorporating published figure images and captions as part of an ongoing project. Figures from some publications have not yet been curated, or are not available for display because of copyright restrictions.

ZFIN is incorporating published figure images and captions as part of an ongoing project. Figures from some publications have not yet been curated, or are not available for display because of copyright restrictions.

ZFIN is incorporating published figure images and captions as part of an ongoing project. Figures from some publications have not yet been curated, or are not available for display because of copyright restrictions.

ZFIN is incorporating published figure images and captions as part of an ongoing project. Figures from some publications have not yet been curated, or are not available for display because of copyright restrictions.

ZFIN is incorporating published figure images and captions as part of an ongoing project. Figures from some publications have not yet been curated, or are not available for display because of copyright restrictions.

ZFIN is incorporating published figure images and captions as part of an ongoing project. Figures from some publications have not yet been curated, or are not available for display because of copyright restrictions.

Unillustrated author statements