Fig 2: PUMILIO, but not RBMX, binding to NORAD is necessary for genome stability.(A) Representative DNA FISH images for chromosome 7 (green) and 20 (red) showing examples of cells with modal (2 n) and non-modal (2n ± 1) chromosome numbers. Arrowheads indicate chromosome gain or loss. (B) HCT116 CRISPRi cells stably expressing the indicated AAVS1 knock-in construct were infected with lentivirus expressing control or endogenous NORAD-targeting sgRNA. Aneuploidy was assayed 18–21 days later using DNA FISH for chromosome 7 and 20, and the frequency of interphase cells exhibiting a non-modal (2 n) chromosome number was scored. Replicates represent two independently-derived AAVS1 knock-in and sgRNA-infected cell lines. 200 nuclei were scored per sample. The dotted line denotes the highest level of background aneuploidy observed in sgControl-infected cells. ns, not significant; *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001, chi-square test comparing sgNORAD to sgControl for each replicate. (C) Representative images of anaphase cells with normal or abnormal (arrowheads) chromosome segregation in DAPI-stained HCT116 CRISPRi cells. (D) The frequency of mitotic cells exhibiting chromosome segregation defects was determined in both biological replicates of each cell population (100 anaphase cells assayed per sample). The dotted line denotes the highest percentage of chromosome segregation defects observed in sgControl-infected cells. ns, not significant; *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001, chi-square test comparing all sgNORAD replicate 1 samples to sgControl GFP replicate 1 and all sgNORAD replicate 2 samples to sgControl GFP replicate 2.

Fig 3: Generation and stable expression of NORAD constructs.(A) Schematic depicting wild-type or mutant NORAD constructs. NORAD sequence conservation in mammals (UCSC Genome Browser Hg38 PhastCons track) highlights the strong conservation of the region of NORAD harboring PREs (arrowheads). PREmut contains 18 UGU to ACA mutations in PREs (gray arrowheads); 5' deletion (5' del) lacks the RBMX binding site (nt 1–898) (Munschauer et al., 2018); 5' fragment (5' frag) spans the RBMX binding site (nt 33–898); ND4 construct represents the most conserved segment of NORAD (nt 2494–3156). (B) Schematic depicting insertion of constructs into the AAVS1/PPP1R12C locus using TALENs. (C) qRT-PCR analysis of expression of each NORAD construct in HCT116 CRISPRi cells after infection with control or endogenous NORAD-targeting sgRNAs. Expression was normalized to endogenous NORAD level, represented by expression in AAVS1-GFP cells infected with sgControl (replicate 1 samples normalized to sgControl AAVS1-GFP replicate 1; replicate 2 samples normalized to sgControl AAVS1-GFP replicate 2). The data in the left graph were generated with a primer pair in ND4 that does not amplify the 5' fragment, while the right graph used primers at the NORAD 5' end. Replicates represent two independently-derived AAVS1 knock-in and sgRNA-infected cell lines. Values normalized to GAPDH expression. n = 3 technical replicates per sample.

Fig 4: Nuclear PWARSN promotes RBMX degradation via the ubiquitination‐proteasome pathway. (A) The interaction between PWARSN and RBMX was confirmed by RNA pull‐down and western blotting. (B) PWARSN interaction with endogenous RBMX in HK‐2 cells was confirmed by RNA immunoprecipitation (RIP) analysis. (C) FISH images showing the co‐localization of PWARSN (red) and RBMX (green) in HK‐2 cells. Scale bar, 20 μm. (D–E) RBMX mRNA (D) and protein (E) levels in PWARSN‐overexpressing and PWARSN‐knockout HK‐2 cells by qRT‐PCR and western blotting. (F) The level of RBMX protein was detected in HG‐treated and PWARSN‐knockout HK‐2 cells incubated with the protein synthesis inhibitor cycloheximide (CHX, 100 μg/ml). (G) The level of RBMX protein was detected in NG‐treated and PWARSN‐overexpressing HK‐2 cells incubated with MG‐132 (10 nmol/ml). (H‐I) Cell lysates were immunoprecipitated with antibody against RBMX and analysed by immunoblotting with ubiquitin (Ub)‐specific antibody and anti‐RBMX antibody. Bottom: Input from cell lysates. (J) Schematic of the interaction between PWARSN and RBMX created using HNADOCK. The binding regions are shown in two different nucleotide sequences of PWARSN. (K) RNA pull down showing the interaction of sequentially deleted PWARSN mutants with RBMX in vitro. (L) The protein levels of RBMX and TXNIP in HK‐2 cells transfected with PWARSN and mutant plasmid of PWARSN interacting with RBMX. Data are presented as the means ± SD, and significance was determined using unpaired Student's t‐test (B) for two groups and one‐way ANOVA (D) for multiple groups (*p < 0.05; **p < 0.01; ***p < 0.001; NS, not significant).

Fig 5: BFA reduces expression of SRSF9 and Tra2-ß, and enhances SMN exon 7 ESE-binding activity of hnRNP M and PSF. (A) SMA fibroblasts were treated with BFA (0.1 and 1 µM) for 24 h, and total cell lysates were prepared and subjected to Western blotting with anti-SRSF1, -SRSF9, -Tra2-ß and -hnRNP A1 antibodies. (B) The amounts of SRSF1, SRSF9, Tra2-ß and hnRNP A1 proteins in panel a were quantified using ImageJ software. The relative amount of each protein was calculated by normalizing to the value from DMSO-treated SMA cells and presented as a percentage. The mean values and standard deviation were determined from two independent experiments. (C) The SMN2 exon 7 splicing effected by siRNA-mediated knockdown of splicing factors was analyzed by RT-PCR. siRNAs targeting GFP mRNA and scrambled siRNA were used as controls. GAPDH mRNA was also analyzed as a control. (D) The PCR products of FL-SMN and SMN?7 mRNAs in panel c were quantified using the ImageJ software. The relative amount of each PCR product was determined by normalizing to the value from GFP siRNA-transfected SMA cells, and are presented as a percentage. (E) 293T cells were treated with DMSO or BFA (0.1 µM) for 24 h, and the cell lysates were prepared and subjected to an RNA affinity assay. Cell lysates were incubated with biotinylated Exon 7-ESE RNAs, and ESE-bound proteins were precipitated by biotin-streptavidin bead interaction. Precipitated proteins were analyzed by Western blotting with anti-PSF, -hnRNP M, -Tra2-ß, -hnRNP G and -SRSF9 antibodies. Histone H3 was also analyzed as a negative control.