Fig 2: Alyref Binding to Poorly Spliced Introns Regulates Their Splicing In Vivo(A) Binding enrichment of RNA export factors and splicing factor Ptbp1 for indicated RNA regions, calculated and displayed as in Figure 1.(B) Examples of transcripts showing Alyref CLIP tags within introns that are normally inefficiently spliced (dashed boxes).(C) Fraction of RIs, DIs, and other transcript regions segregated by Alyref binding status and called as differentially included upon ALYREF RNAi using DEXSeq (false discovery rate [FDR] 10%, fold change > 1.5).See also Figure S2.

Fig 3: eIF4A3 Is Important for Alyref Binding to Various RNAs and mRNA Export(A) RNA export factors average binding profiles at internal exon-exon junctions. EJC’s position is inferred from Figure S4A and marked by brown arrowheads.(B) Relative changes in Alyref RIP efficiency upon eIF4A3 RNAi in 293T cells. Mean of three independent experiments ± SD. **p < 0.01, ***p < 0.001 (t test).(C) Examples of EJC bound to intronic (dashed boxes) or single-exon RNAs. Also see Figure S4C.(D) eIF4A3 and Alyref binding enrichments over Ctrl iCLIP on indicated RNA types and regions (as in Figure 1).(E) Oligo(dT) FISH on HeLa cells RNAi for the indicated factors. Pictures were taken at the same exposure level. White boxes indicate higher magnification. Scale bar, 10 µm. See also Figures S4 and S5.

Fig 4: Ddx5 is present in the c-fos mRNP. (A) RT-PCR analysis of fully spliced c-fos mRNA associated with ddx5 (RNA-ChIP) after cell cross-linking and IP using either a control antibody (IgG) or ddx5 antibody. Input represents nuclear RNA. MCF-7 were treated with either estradiol (+E2) or vehicle (ethanol, -E2) for 1 h. (B) RT-PCR analysis of fully spliced c-fos mRNA associated with HA-ddx5 (RNA-ChIP) using an inducible MCF-7 cell line expressing HA-ddx5 (+DOX) or not (-DOX). After cross-linking, the IP was performed using an HA antibody. Input represents nuclear RNA. (C) Western blot analysis of ddx5 after IP with either control antibodies (IgG) or antibodies against ALY, or TAP, as indicated. Cellular extracts were treated or not with RNase A before IP. (D) RT-qPCR measuring the unspliced CCND1 pre-mRNA using cellular extracts from MCF-7 transfected with either siCTRL or siddx5 before treatment with either estradiol (+E2) or vehicle (ethanol, -E2) for 1 h. The CCND1 pre-mRNA level value obtained in the control experiment (siCTRL, -E2) was used as the baseline. (E) The fully spliced CCND1 mRNA was measured by RT-qPCR using cytosolic or nuclear extracts from MCF-7 transfected with either siCTRL or siddx5 before treatment with either estradiol (+E2) or vehicle (ethanol, -E2) for 1 h. The N/C ratio obtained in the control experiments (siCTRL) was used as the baseline. Histograms represent the average of at least three independent experiments. Error bars represent S.E.M. (*P < 0.05, ***P < 0.0005). (F) Ddx5 is required for the transcriptional activation of the c-fos gene, the efficient co-transcriptional c-fos RNA splicing and the efficient export of the c-fos mature mRNA. In the absence of ddx5, the c-fos RNA might be spliced after transcription, which disturbs the recruitment of TAP and the c-fos mRNA export.

Fig 5: Depletion of SRSF1 suppresses patient-derived C9ORF72-mediated astrocytic toxicity and motor neuron death.(a) Untransfected HEK293T (UT), control (Ctrl)-RNAi, SRSF1-RNAi or LV-SRSF1-RNAi and FLAG-tagged SRSF1 transfected cells were analysed 72 h post-transfection by immunoblotting using anti-FLAG, anti-ALYREF and loading control anti-a-Tubulin antibodies. (b) SRSF1 and ALYREF transcript levels were quantified by qRT–PCR analysis following normalization to U1 snRNA levels in three biological replicate experiments (mean±s.e.m.; two-way ANOVA with Tukey’s correction for multiple comparisons; N (qRT–PCR reactions)=6). Three control (Ctrl-pat) and three C9ORF72-ALS (C9-ALS-pat) colour-coded patient lines were used. (c) Representative immunofluorescence microscopy images from iAstrocytes and motor neuron co-cultures. iAstrocytes derived from controls (top left image) and C9ORF72-ALS patients (bottom left image) were transduced with Ad-RFP as viral control and cultured with Hb9GFP+ motor neurons (green). C9ORF72-ALS iAstrocytes exhibit toxicity against motor neurons as previously reported. Transduction of control Ad-RFP iAstrocytes with LV-SRSF1-RNAi co-expressing GFP led to consistent survival of Hb9GFP motor neurons on control astrocytes (top right image) and increase in survival of motor neurons on C9ORF72-ALS astrocytes (bottom right image). Arrows point to examples of axons of motor neurons. Scale bar, 50 µm. (d) Same control (Ctrl-pat) and C9ORF72-ALS (C9-ALS-pat) colour-coded patient lines (as in b) are used for quantification of Hb9-GFP+ motor neuron counts in four biological replicates of co-cultures of astrocytes and motor neurons at day 3 (mean±s.e.m.; one-way ANOVA with Tukey’s correction for multiple comparisons; N (Hb9-MNs)=Ctrl-pat154: 567/589/582/500, Ctrl-pat154+SRSF1-RNAi: 620/543/504/349, Ctrl-pat155: 602/610/553/571, Ctrl-pat155+SRSF1-RNAi: 554/584/532/516, Ctrl-pat209: 519/599/584/535, Ctrl-pat209+SRSF1-RNAi: 617/486/425/572, C9-ALS-pat78: 352/279/294/258, C9-ALS-pat78+SRSF1-RNAi: 569/451/398/583, C9-ALS-pat183: 200/188/154/145, C9-ALS-pat183+SRSF1-RNAi: 480/420/380/399, C9-ALS-pat201: 201/243/261/224, C9-ALS-pat201+SRSF1-RNAi: 486/463/444/485). (e) Quantification of nuclear and cytoplasmic sense RNA foci in SRSF1-RNAi-transduced iAstrocytes (mean±s.e.m.; two-way ANOVA; N (cells with 1-5 RNA foci)=C9-ALS-pat78+MOI0: 21, C9-ALS-pat78+MOI5: 20, C9-ALS-pat78+MOI7: 22, C9-ALS-pat183+MOI0: 21, C9-ALS-pat183+MOI5: 24, C9-ALS-pat183+MOI7: 22, C9-ALS-pat201+MOI0: 24, C9-ALS-pat201+MOI5: 23, C9-ALS-pat201+MOI7: 22). >95% of cells with RNA foci presented a total of 5 or fewer foci. Statistical significance of data is indicated as follows: NS: non-significant, P=0.05; *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.