Fig 1.
EJC components are co-expressed in neurogenesis.
(A) Schematic of embryonic neurogenesis of the dorsal telencephalon. NSC, neural stem cell; IP, intermediate progenitor. (B) Two main questions posed in this study. 1. Does Eif4a3 haploinsufficiency cause microcephaly? 2. Do EJC components regulate common pathways during neurogenesis? (C) qPCR of Magoh, Eif4a3, and Rbm8a mRNA levels in developing neocortices of indicated ages. qPCR was performed using a standard curve, with Magoh relative expression at E10.5 set to 1.0, and all expression levels normalized to Gapdh. (D-I) Immunofluorescence of E10.5 dorsal neocortices for Hoechst (blue), Magoh (D, E), Rbm8a (F, G), and Eif4a3 (H, I). (E, G, I) are high magnification images of D, F, H, respectively. Student’s t test, Error bars, S.D., **, p<0.01, ns = not significant. n = 3 biological replicates each age. Scale bars, D, F, H; 50 μm; E, G, I, 25 μm.
Fig 2.
Eif4a3 is required for embryonic neurogenesis and brain size.
(A) Top, Eif4a3 genomic mouse locus. Middle, targeted allele with 2 loxp sites (black arrowheads), Neo cassette, and 2 FRT sites (white arrowheads). Genotyping primers are indicated. Bottom, the conditional allele following FLP- and Cre-mediated recombination. (B) qPCR quantification of Eif4a3 mRNA levels in E10.5 neocortices, following normalization using Gapdh. Eif4a3 mRNA level of Emx1-Cre samples was set to 1.0. (C) Quantification of Eif4a3 protein levels in E11.5 dorsal cortices by densitometry of western blots, following normalization with α-Tubulin for loading. (D,E) Whole mount E12.5 Emx1-Cre and Emx1-Cre;Eif4a3lox/+ brains. Note the forebrain (dotted lines) is noticeably smaller in the Eif4a3 mutant. (F) Quantification of cortical thickness of E12.5 Emx1-Cre and Emx1-Cre;Eif4a3lox/+ dorsal neocortices. (G-J) 4 different coronal sections from E12.5 Emx1-Cre (G,H) and Emx1-Cre;Eif4a3lox/+ (I,J) neocortices stained for Hoechst (blue), Pax6 (green, G,I) or Tuj1 (green, H,J). (K) Density of Pax6+ cells within 200 μm wide radial columns spanning the E12.5 cortices of indicated genotypes. (L, M) Images of E11.5 Emx1-Cre (L) or Emx1-Cre;Eif4a3lox/+ (M) cortices stained for PH3 (green). (N) Graph depicting percentage of all cells which are PH3-positive for indicated genotypes at E11.5. (O-T) E12.5 coronal sections from Emx1-Cre (O) and Emx1-Cre;Eif4a3lox/+ brains (P-T) stained for Hoechst (blue), CC3 (red), Pax6 (green, Q, S), and Tuj1 (green, R,T). S and T are high-magnification views of Q and R, respectively, as indicated. Arrowheads depict cells co-labeled for apoptotic and cell fate markers. (U-X) Coronal sections of E14.5 Emx1-Cre (U,W) and Emx1-Cre;Eif4a3lox/+ (V,X) cortices stained for Hoechst (white or blue) and Tuj1 (green). W and X are high-magnification images of U and V, respectively as indicated. Red brackets denote cortical thickness. Vent, ventricle. Student’s t test, *, p<0.05, ***, p<0.001. Error bars, S.D. n = 3 biological replicates each. Scale bars, D, E, 1 mm; G-J, L,M,O-R, W, X, 50 μm; S,T, 20 μm; U,V, 200 μm.
Fig 3.
Haploinsufficiency of EJC components causes microcephaly.
(A-D) Images of whole mount brains at P12 from indicated genotypes. Dotted lines denote dorsal cortex. (E) Quantification of relative dorsal cortical area in P12 brains of indicated genotypes. The area of Emx1-Cre brains was set to 1.0. ANOVA with Tukey posthoc, ***, p<0.001, Error bars, S.D. n = 3–4 biological replicates, Scale bar, A-D, 2 mm.
Fig 4.
Transcriptome analyses of E10.5 Magoh, Rbm8a, and Eif4a3 haploinsufficient cortices reveal common downstream pathways.
(A) Diagrammatic overview of RNA sequencing analysis of E10.5 neocortices from indicated genotypes. (B) qPCR showing expression of Magoh, Rbm8a, and Eif4a3 in their respective mutant E10.5 cortices. (C) Heatmaps showing z-score transformed normalized expression for all affected transcripts with an FDR corrected p-value, q< 0.05. Genes and samples were clustered using correlation distance with complete linkage. (D) Scatter plots of transcripts significantly upregulated (green dots) and downregulated (red dots) in E10.5 Emx1-Cre;Magohlox/+, Emx1-Cre;Rbm8alox/+, and Emx1-Cre;Eif4a3lox/+ cortices (q<0.05). (E, F) qRT-PCR validation at E11.5 compared to relative RNA-seq values of Tbr2 (E) and Ngn2 (F) in the indicated genotypes. For RNA-seq and qPCR, each control was normalized to 1.0 and compared to mutants. (G) Graph depicting common KEGG terms identified by GSEA analysis that were significant in all 3 EJC mutants, showing corresponding enrichment score. Student’s t test (B,E,F), Error bars, S.D. *, p<0.05, **, p<0.01, ***, p<0.001.
Fig 5.
Transcriptome analyses of E10.5 Magoh germline haploinsufficient brains identifies alterations in ribosome and p53 signaling pathways.
(A) Diagrammatic overview of RNA sequencing analysis of E10.5 neocortices (dotted lines) from indicated genotypes. (B) Heatmaps showing z-score transformed normalized expression for control and MagohMos2/+. Genes and samples were clustered using correlation distance with complete linkage. (C) Scatter plot of transcripts significantly upregulated (green dots) and downregulated (red dots) in E10.5 MagohMos2/+ cortices (q<0.05), n = 4 biological replicates each. (D) Validation and RNA-seq values for Dclk1 and Tbr2 in indicated E11.5 mutant dorsal neocortices. Controls were normalized to 1.0. (E) Graph depicting top ranked KEGG terms by GSEA analysis in MagohMos2/+ showing corresponding fold enrichment. Student’s t test (D), Error bars, S.D. *, p<0.05, ***, p<0.001.
Fig 6.
Haploinsufficiency for EJC components alters mRNA splicing of splicing regulators.
(A) Bar graph showing alternative splicing events for each mutant relative to the control. (B-D) Clustered column graphs of the distribution of ψ values of all identified intron retention (RI) events in E10.5 dorsal cortices for each EJC mutant, using a threshold of 20 for Bayes factor. Ψ<0 indicates higher probability for the mutant to have intron retention when compared to the control. (E) Top: IGV view of increased Mapk13 intron 6–7 reads in red frame. Primers indicated as arrows. Bottom: RT-PCR showing increased Mapk13 RI isoforms in Emx1-Cre;Rbm8alox/+ E11.5 dorsal cortices compared to the control. (F) Bar graph of common KEGG terms that were significant in all 3 EJC mutants, showing corresponding enrichment score. ANOVA (A), Modified fisher’s exact test (F), *, p<0.05, **, p<0.01, ***, p<0.001.
Fig 7.
Proteomic analysis of E11.5 EJC mutant brains reveals alterations in levels of ribosome-associated proteins and ribonucleoproteins.
(A) Diagrammatic representation of workflow to perform proteomic analysis of E11.5 dorsal cortices. (B, C) Bar graph of all common enriched KEGG terms (B) and top common GO terms (C) among all 3 EJC mutants showing corresponding fold enrichment and P values. (D) STRING network analysis of proteins within the broadest GO category, “Ribonucleoprotein Complex” altered in any of the 3 EJC mutants. Stronger associations are represented by thicker lines, and circles are colored based upon alteration in 1 or more mutants and level of significance. Two networks of splicing regulators and ribosome-associated proteins are detected. Modified fisher’s exact test, *, p<0.05, **, p<0.01, ***, P<0.001.
Fig 8.
EJC haploinsufficiency induces P53 activation.
(A-J) Coronal sections of cortices from E13.5 Emx1-Cre (A), E13.5 Emx1-Cre;Magohlox/+ (B), E12.5 Emx1-Cre (C, G, I), E12.5 Emx1-Cre;Eif4a3lox/+ (D), E11.5 Emx1-Cre (E), E11.5 Emx1-Cre;Rbm8alox/+ (F), and E12.5 Emx1-Cre;Rbm8alox/+ (H,J) embryonic cortices stained for Hoechst (blue), P53 (green), and Pax6 (red), with co-localization indicated in yellow. Sections were demarcated with dotted lines. Each image is representative of at least 3 independent biological samples. Scale bar, A-J, 50 μM.
Fig 9.
Loss of p53 partially rescues microcephaly of Magoh, Rbm8a, and Eif4a3 haploinsufficient mutants.
(A-D, F-I, and K-N) Whole mount brains of E18.5 embryos with indicated genotypes. (E, J, O) Quantification of cortical area in E18.5 embryonic brains with indicated genotypes. Dotted lines demarcate the dorsal cortical areas measured. The surface area of littermate control brains was set to 100. ANOVA with Tukey posthoc, **, p<0.01, ***, p<0.001, NS, not significant. Error bars, S.D. n = 3–9 biological replicates each. Scale bars, A-D, E-I, and K-N, 1 mm.
Fig 10.
Loss of p53 partially rescues neuron number and distribution associated with Rbm8a haploinsufficiency.
(A-D) Coronal sections of E18.5 brains of indicated genotypes stained with Hoechst (white). (E-L) Regions of coronal sections indicated in (A-D, red dotted square) stained for Cux1 (E-H) and Tbr1 (I-L). (M, N) Quantification of Cux1+ (M) and Tbr1+ (N) density within a 250 μm radial column for indicated genotypes. (O, P) Bar graphs depicting density of Tbr1+ neurons in VZ/SVZ (bin 1–2, O) and cortical plate (bin 3–5, P) of indicated genotypes. Bins were quantified as indicated in I. Brackets denote general boundaries of Cux1 and Tbr1 layers. ANOVA with Tukey posthoc *, p<0.05, **, p<0.01, ***, p<0.001, ns, not significant. Error bars, S.D. n = 2–3 biological replicates each. Scale bars, A-L, 50 μm.