Fig. 1. SMG7 depletion impairs NMD activity.

a Schematic representation of the final steps of nonsense-mediated mRNA decay (NMD). Phosphorylated UPF1 (indicated by the blue sphere) recruits the SMG5-SMG7 heterodimer to the target mRNA (indicated in dark gray), thereby promoting deadenylation via the CCR4-NOT complex. Recruitment of SMG6 to UPF1 results in endonucleolytic cleavage of the target transcript via the activity of the SMG6 PIN domain. The SMG5 PIN domain is catalytically inactive. b Western blot analysis of SMG7 knockout (KO) cell lines (clones 2, 31, and 34) with the anti-SMG7 antibodies AK-133, AK-134, and AK-136 (n = 1); Tubulin serves as control (see “Methods” section and Supplementary Data 6 for antibody details). The region of SMG7 detected by the antibodies is schematically depicted and the crRNA targeting site indicated. Asterisks indicate non-specific bands. c End-point RT-PCR detection of SRSF2 transcript isoforms (top) and quantitative RT-PCR-based detection (qPCR; bottom) of ZFAS1 in the indicated cell lines with or without expression of FLAG-tagged SMG7 as rescue construct. The detected SRSF2 isoforms are indicated on the right, the NMD-inducing included exon is marked in red (e = exon). Relative mRNA levels of SRSF2 isoforms were quantified from bands of agarose gels (n=3 biologically independent samples). The ratio of ZFAS1 to the C1orf43 reference was calculated; data points and means from the qPCRs are plotted as log2 fold change (log2FC) (n = 3 biologically independent samples). The plotted points are color-coded based on cell line (gray = WT; blue = SMG7 KO-2; green = SMG7 KO-34). d Quantitative RT-PCR-based detection (qPCR) of SRSF2 isoforms and ZFAS1 in the indicated cell lines upon treatment with the indicated siRNA. The ratio of NMD isoform to canonical isoform (SRSF2) and ZFAS1 to the C1orf43 reference was calculated; data points and means from the qPCRs are plotted as log2 fold change (log2FC) (n = 3 biologically independent samples). e Read coverage of SRSF2 from SMG7 KO and published SMG7 KD (GEO: GSE86148) RNA-Seq data is shown as Integrative Genomics Viewer (IGV) snapshots. The canonical and NMD-sensitive isoforms are schematically indicated below. Percent spliced in (PSI; from LeafCutter analysis) and mean counts from 4 indicative splice junctions are shown. Differential gene expression (from DESeq2) is depicted as log2 fold change (log2FC) in the last column. f Overlap of upregulated or downregulated premature termination codon (PTC)-containing isoforms between the SMG7 KO or KD RNA-Seq data is shown as UpSet plot. g–i Volcano plots showing the differential transcript usage (via IsoformSwitchAnalyzeR) in various SMG7 depletion RNA-Seq data. Isoforms containing GENCODE (release 33) annotated PTC (red, TRUE), regular stop codons (blue, FALSE), or having no annotated open reading frame (gray, NA) are indicated. The change in isoform fraction (dIF) is plotted against the −log10 adjusted p-value (adj.p-value). Density plots show the distribution of filtered isoforms in respect to the dIF, cutoffs were |dIF|> 0.1 and adj.p-value < 0.05. P-values were calculated by IsoformSwitchAnalyzeR using a DEXSeq-based test and corrected for multiple testing using the Benjamini-Hochberg method.