Fig 1.
RES complex is essential for early vertebrate development.
(A) Schematic model of the RES complex adapted from Brooks et al. [8]. Rbmx2 (light blue) is the core subunit with an RRM-domain structure. Bud13 (orange) and Snip1 (pink) interact with Rbmx2 (light blue). (B) In situ hybridization showing spatial and temporal expression of RES complex members. (C-E) Gene models of the mutant allele generated using CRISPR-Cas9-nanos. (C) bud13, 7 nt deletion in exon 6 generated a premature stop codon. (D) rbmx2, 16 nt deletion removed exon-intron boundary at exon 2 (exon capital letter, intron lower letter). (E) snip1, 11 nt deletion in exon 1 generated a premature stop codon. f-g) Lateral view of RES complex mutant embryos, their corresponding WT sibling and mutants injected with the cognate mRNA. (F) bud13 mutant at 32 hpf (scale bar: 0.35mm) and at 48 hpf (scale bar: 0.5mm). (G, H) rbmx2 and snip1 mutant at 48 hpf respectively. WT: represent phenotypically wild type sibling from the same mutant fish line.
Fig 2.
RES complex is required during zebrafish brain development.
(A) Acridine orange (ao) staining of zebrafish mutant embryos for bud13 (30–32 hpf), rbmx2 and snip1 (48 hpf). Mutants show a marked degree of cells with nuclear uptake of ao compared to WT sibling most predominantly in the head. WT: represent phenotypically wild type sibling from the same mutant fish line. (B) Maximum intensity projections of individual and merged channels (GFP and dsRed) of 3D confocal images of bud13Δ7/Δ7 and their wild-type siblings (scale bar 20 μm) in transgenic lines that label GABAergic neurons and precursors (Tg[dlx6a-1.4kbdlx5a/ dlx6a:GFP]) and glutamatergic neurons (Tg[vglut:DsRed]). WT: represent phenotypically wild type sibling from the same mutant fish line. (C) Total number of dlx5a/6a:GFP+ cells (GABAergic neurons and precursors) and vglut:DsRed+ cells (glutamatergic neurons) in the forebrain of the bud13Δ7/Δ7 (n = 3) and WT sibling (n = 4) were quantified. *** vglut: P = 2 x 10−4; ***dlx5a/6a: P = 3 x 10−4 (one-way ANOVA).
Fig 3.
RES complex mutants show mild widespread intron mis-splicing.
(A) Biplot illustrating percent intron retention (PIR) in each of the RES mutants and the corresponding phenotypically wild type (WT) siblings. Blue and red dots correspond to introns with higher inclusion in the mutant and WT, respectively, using a cutoff of ∆PIR >15. Insets show exon gene expression levels in the same conditions (see S4 Fig for details). (B) Stacked barplot showing the number of mis-regulated events upon RES loss-of-function. Clearly, most changes are intron retention supporting the role of the RES complex in splicing. (C) Scheme showing how PIR was measured (adapted from Braunschweig et al. [22]; see Methods for details). (D) Euler diagram showing the number of retained introns (∆PIR>5) in the three mutants and the inter-mutant overlaps. (E) RNA-seq read density across the rrp8 gene in bud13, rbmx2 and snip1 mutants and their corresponding phenotypically WT siblings. Intronic signal increases in RES mutants (∆PIR>15) (dark blue) (dotted square box). (F) RT-PCR assays validate the increased retention of an rrp8 intron (from panel E) in bud13, rbmx2 and snip1 mutants compared to the corresponding phenotypically WT siblings.
Fig 4.
Genome-wide analysis of intron retention in RES complex mutants.
(A) Schematic representation of intron features analyzed in (B) and (C). Top: Intron position. “Internal” introns correspond to all introns excluding the first two and last three introns. Bottom: NMD vs non-NMD triggering introns. Introns predicted to cause NMD upon retention introduce a premature termination codon (PTC) further than 50 nucleotides upstream of an exon-exon junction. Introns predicted not to cause NMD (noNMD) may correspond to: (i) last introns, (ii) introns in UTRs or non-coding genes, or (iii) introns that preserved the ORF upon retention (multiple of three nucleotides with no in-frame stop codons). (B) Box plots showing the ∆PIR of last and internal introns in the three different mutant of the RES complex. Only genes with more than 10 introns were considered for the analysis. bud13 *** P = 4.37x10-201; rbmx2 *** P = 2.98x10-202; snip1 *** P = 4.76x10-195 (Wilcoxon rank sum test). (C) Box plots showing the ∆PIR of introns predicted to trigger nonsense mediated decay (NMD) upon retention and those predicted not to trigger NMD (no-NMD) in the three different mutant of the RES complex. bud13 *** P = 5.83x10-208; rbmx2 *** P = 4.94x10-184; snip1 *** P = 1.74x10-183 (Wilcoxon rank sum test).
Fig 5.
Molecular features that determine intron retention in RES are associated with splicing through intron definition (A-G). Box plots showing the median (black solid line) and the distribution of values for multiple intron-exon features of several groups of introns of interest. “RESdep”, all highly retained introns (∆PIR >15) in at least two out the three mutants in the RES complex; “no-NMD”, subset of “RESdep” retained introns that are predicted not to trigger NMD; “NMD”, subset of “RESdep” retained introns predicted to trigger NMD; and “Ctr”, control set of introns with a ∆PIR cutoff < 0.5 in the three RES complex mutants (see Materials and Methods for details). (*** P ≤ 0.001; Mann-Whitney-U test). Branch point (BP) definition: BP Score of best-predicted BP. Scored base on Corvelo et al. [24] (See Materials and Methods for details).
Fig 6.
Logistic regression model can accurately classify RES-dependent and non-dependent introns.
(A) Classification performance of logistic regression models for different data sets of differentially retained vs. Ctr introns. ROC curves are averaged over 10,000 repeated holdout experiments where models have been trained with randomly sampled subsets of 90% (1,268) of the RESdep introns versus 1,268 Ctr introns with 30 features (S3 Table) and Lasso feature selection. Classification performance was estimated using the remaining 10% (141) RESdep introns and 141 randomly sampled Ctr introns. Having held fixed parameters, the same model was used to estimate classification performance with randomly sampled 141 introns from the other RES-dependent data sets, namely: (i) “RESdep (∆PIR>10)” introns from the “RESdep” set with ∆PIR > 10 in all three mutants (871 introns); (ii) “RESdep (NMD)”, introns from the “RESdep” set predicted to trigger NMD when retained (574 introns); (iii) “bud13∆7/∆7”, introns with ∆PIR>15 upon bud13 mutation at 32 hpf (2,363 introns); (iv) “rbmx2∆16/∆16”, introns with ∆PIR>15 upon rbmx2 mutation at 48 hpf (2,186 introns); and (v) “snip1∆11/∆11”, introns with ∆PIR>15 upon snip1 mutation at 48 hpf (2,675 introns). 95% confidence interval of reported average AUCs corresponds to AUC ± 0.001. (B) Capacity of each feature to discriminate between “RESdep” and “Ctr” introns, measured as AUC (average area under ROC curve) when used as the only feature in a one-feature logistic regression model. (C) Schematic of the experiment performed to validate the predicted features, see Material and Methods for details. (D) RES dependent but not RES independent intron was retained in bud13∆7/∆7 mutant as the regression model predicted. The validation experiment was done using two independent biological replicates (see S12A Fig).
Fig 7.
Schematic integrative model of the RES complex function in splicing.
The loss-function of the RES complex induces a global weak defect in splicing, but a strong retention of a subset of introns with the following features: these introns are short, have a higher GC content and flanked by GC-depleted exons, features associated with intron definition splicing mechanism. This splicing defect leads to a brain-related phenotype.