Fig 2: Structural and functional deficiency of paraspeckles in FUS ΔNLS lines. a Interaction of FUS with SFPQ and NONO is reduced in FUS ΔNLS lines as revealed by proximity ligation assay (PLA). PLA was performed in a heterozygous (ΔNLS2_het) and a homozygous (ΔNLS1_ho) lines; FUS KO cells were used as a negative control. Representative images and quantification (number of single interactions (dots) per cell (foci per cell)) are shown. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (one-way ANOVA with Holm-Sidak test). b Extractability of NEAT1_2 is increased in FUS ΔNLS lines. NEAT1_2 extractability was analysed by determining its levels in QIAzol-lysed heated versus non-heated samples (“fold extraction”) by qRT-PCR. Note near-complete NEAT1_2 extractability in FUS KO cells (fold extraction ~ 1). See also Additional file 1: Figure S4B. N = 3 per line. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (one-way ANOVA). c NEAT1_1 accumulates in soluble nuclear extract (SNE) in FUS ΔNLS lines. Left, representative PCR (non-saturated conditions, 26 cycles); right, qRT-PCR analysis. A primer pair located immediately upstream NEAT1_1 polyA-tail (NEAT1 pA) was used to quantify NEAT1_1 in cDNA of polyadenylated RNA. Note that NEAT1_2 which is not polyadenylated is undetectable under these conditions. *p < 0.05, **p < 0.01 (one-way ANOVA with Holm-Sidak test). d NEAT1 displays diffuse distribution in poly(I:C)-stimulated ΔNLS_het lines. Cells were analysed 8 h after poly(I:C) transfection by NEAT1 RNA-FISH. Representative images and quantification of the fraction of cells with diffuse NEAT1 distribution are shown. e Paraspeckle-regulated miRNAs are decreased in FUS ΔNLS lines. Levels of six mature miRNAs produced from pri-miR17~92 were measured by qRT-PCR separately for heterozygous and homozygous FUS ΔNLS lines, and combined average values were plotted. *p < 0.05 (Mann-Whitney U-test). Combined data for three heterozygous and three homozygous lines are referred as “het pooled” and “ho pooled”, respectively. In a and d, numbers of cells analysed are indicated within each bar. Scale bars, 10 μm

Fig 3: FUS/Pur-alpha physical interaction. (a) GST, GST-fused C-terminal region of wild-type FUS (FUS_Ct_WT), and GST C-terminal domain of multimutated FUS (FUS_Ct_MM) were used as baits in affinity purification experiments from a rat brain Triton X-100 extract, in the presence or absence of RNAse. Affinity-purified material retained by the GST fusion proteins was resolved by SDS-PAGE and processed by western blotting with anti-Pur-alpha antibody (top). The same volume of eluted material analyzed by western blotting was separated on a different SDS-PAGE and stained with Coomassie blue to verify equal loading of the different GST fusion baits (bottom). 1 : 500 of total brain extract and 1:50 of proteins retained from each column were loaded on the gel. SM, starting material. (b) Interaction between the same GST fusion proteins utilized in (a) and in vitro-translated HA-tagged Pur-alpha was tested by pull-down in the presence or absence of RNAse. HA Pur-alpha bound to the GST fusion proteins was resolved by SDS-PAGE and analyzed by western blotting with anti-HA antibody. GST fusion proteins used in the pull-down assay were resolved by SDS-PAGE and stained with Coomassie blue (bottom). 1 : 50 of reticulocyte extract exploited in the pull down and 1:3 of proteins retained from each column were loaded on the gel. (c) Protein extracts from HeLa cells expressing HA-Pur-alpha on its own (Control), or together with FUSWT, FUS carrying single mutations (R521G, R522G, R524S, or P525L), or FUSMM were incubated with or without RNAse and immunoprecipitated with anti-Flag antibody. Retained proteins were separated by SDS-PAGE and analyzed by western blotting with anti-HA and anti-Flag antibodies. 1 : 50 of total cell extract utilized for each immunoprecipitation and 1 : 3 of bound proteins were loaded on the gel. SM, starting material; IP, immunoprecipitate

Fig 4: p53 signaling and FUS-CHOP expression in FUS-CHOP-driven primary tumor cell lines. (a) Mouse background and tumor generation methods for primary tumors. (b) FUS-CHOP expression in primary mouse tumors and cell lines. (c) p53 and p21 expression 1 hour post-10 Gy irradiation in primary mouse tumors generated in different mouse backgrounds. 3T3 cells are mouse fusion negative controls, and 3T3-FC are mouse fusion positive controls for p53 signaling. (d) mRNA expression of p53 target genes 1 hour after treatment with 10 Gy ionizing radiation in NIH-3T3 and FUS-CHOP-driven tumor cell lines.

Fig 5: AltFUS potentiate FUS recruitment to stress granules AImages by confocal microscopy of altFUS (Flag‐tagged—white), FUS (GFP‐tagged—green) and TIA‐1 (red) signals in HeLa cells transfected with the bicistronic GFP‐FUS(F lag )‐R495x or the monocistronic GFP‐FUS(Ø‐ flag )‐R495x constructs, or co‐transfected with the monocistronic GFP‐FUS(Ø‐F lag )‐R495x and altFUS‐Flag constructs (representative images from n = 3). The white scale bar corresponds to 10 μm.B–DImages by confocal microscopy of altFUS(Flag‐tagged—white), FUS (GFP‐tagged—green) and TIA‐1 (red) signals in HeLa cells transfected with the bicistronic constructs (B), monocistronic constructs (C) or co‐transfected with altFUS‐Flag and the monocistronic constructs (D) of 6 ALS‐associated mutants: FUS‐G156E, FUS‐K510E, FUS‐Q519x, FUS‐Q519I‐fs527x, FUS‐R521C and FUS‐P525L (representative images from n = 3). The white scale bar corresponds to 10 μm.