Extended Data Fig. 9: UPF1 and HNRNPC regulate the stability of alternative-spliced RNA transcripts during EGF stimulation.

(a) Bars indicate the overlap of significant AS events between EGF-stimulated HNRNPC-KD and UPF1-KD samples. 2568 common AS events (1721 genes) with congruent ∆PSI patterns after HNRNPC-KD and UPF1-KD were selected. (b) Bubble plots indicate -log₁₀(P-value) as determined by gProfileR2⁵³ analysis of top GO terms for molecular functions, cellular components, and regulatory processes of 1721 common HNRNPC-KD/UPF-KD AS genes. (c) Workflow of differential transcript expression analysis of EGF-stimulated HNRNPC- and UPF1-KD data compared to control siRNAs. 844 AS transcripts were found to be commonly differentially expressed (DE) in EGF-treated HNRNPC and UPF1 KD samples. (d) Scatter plots indicate log₂FC values against control siRNA of common DE transcripts in EGF-stimulated HNRNPC-KD and UPF1-KD samples. Purple, light blue, orange, steel blue, and red dots: 844 common DE AS transcripts divided for each AS category; grey dots: 3724 significant DE background transcripts. (e) Lines indicate the cumulative fraction of log₂FC values against control siRNA of common DE transcripts in EGF-stimulated HNRNPC-KD and UPF1-KD samples. Purple, light blue, orange, steel blue, and red lines: 844 DE AS transcripts divided for each AS category; grey lines: 3,724 significant DE background transcripts in HNRNPC and UPF1 knockdown samples. (f) Workflow of mRNA stability analysis using actinomycin D after the knockdown of HNRNPC and UPF1 in EGF-stimulated (1 h) A431 cells. 665 AS transcripts were found to be commonly differentially stable (DS) after actinomycin D in EGF-treated HNRNPC and UPF1 KD samples. (g) Bars indicate knockdown efficiencies of HNRNPC and UPF1 in HNRNPC and UPF1 knockdown A431 cells treated with EGF for 1 h and then with actinomycin D for a maximum of 4 h, as detected by qPCR. Data are mean (n = 2 biologically independent experiments). (h) mRNA stability profiles for HNRNPC and UPF1 knocked down transcripts as a quality check of actinomycin D treatment, determined by qPCR, in HNRNPC and UPF1 knockdown samples with EGF for 1 h and then with actinomycin D for a maximum of 4 h. Data are mean (n = 2 biologically independent experiments). (i) Principal component analysis (PCA) plots of EGF-treated HNRNPC and UPF1 KD samples followed by actinomycin D timecourse. Green dots: 0 h; blue dots: 1 h; purple dots: 3 h; red dots: 4 h actinomycin D timepoints. (j) Scatter plots indicate log₂FC values against control siRNA of common DS transcripts in EGF-stimulated HNRNPC-KD and UPF1-KD samples. Purple, light blue, orange, steel blue, and red dots: 665 common DS AS transcripts divided by each AS category; grey dots: 2,900 significant DS background transcripts. (k) Lines indicate the cumulative fraction of log₂FC values against control siRNA of common DS transcripts in EGF-stimulated HNRNPC-KD and UPF1-KD samples. Purple, light blue, orange, steel blue, and red lines: 665 DS AS transcripts; grey lines: 2900 significant DS background transcripts in HNRNPC and UPF1 knockdown samples. (l) Venn diagram indicates the overlap between the 35 co-bound RI genes by HNRNPC and UPF1 and the genes harbouring transcripts that were differential expressed or stable (DE and DS) after HNRNPC and UPF1 KD in EGF-treated A431 cells. (m) Schematics of the minigene construct for RND3 and RPS9 intron retained regions. (n) Representative image of n = 3 minigene assay biologically independent experiments for RND3 (left) and RPS9 (right) intron-retained regions in EGF-treated (1 h) HNRNPC and UPF1 KD samples. Unspliced and spliced regions are indicated with blue-orange-blue and blue-blue boxes, respectively. (o) Stacked bars indicate the percentage of unspliced and spliced minigene fragments for RND3 and RPS9 intron-retained regions in EGF-treated (1 h) HNRNPC and UPF1 KD samples. Data are mean ± SD; RND3-HNRNPC-KD P-value = 2.2e-5***, RND3-UPF1-KD P-value = 1.9e-5***, RPS9-HNRNPC-KD P-value = 8.0e-4***, RPS9-UPF1-KD P-value = 1.1e-3** – one-way ANOVA with two-sided Dunnett’s multiple comparison tests against control siRNA cells (n = 3 biologically independent experiments). In d and j, Red and black lines and grey bands: fitted robust linear regression with standard error. Statistical comparison of regressions between common DE (d) and DS (j) HNRNPC-KD/UPF1-KD AS transcripts and significant background transcripts was made using a two-sided Chow’s test. In e and k, P-values between common DE (e) or DS (k) HNRNPC-KD/UPF1 AS transcripts and significant background transcripts were calculated by a one-sided Kolmogorov-Smirnov test.

Source Data