A common cellular response to broad splicing perturbations is characterized by metabolic transcript downregulation driven by the Mdm2–p53 axis

ABSTRACT Disruptions in core cellular processes elicit stress responses that drive cell-state changes leading to organismal phenotypes. Perturbations in the splicing machinery cause widespread mis-splicing, resulting in p53-dependent cell-state changes that give rise to cell-type-specific phenotypes and disease. However, a unified framework for how cells respond to splicing perturbations, and how this response manifests itself in nuanced disease phenotypes, has yet to be established. Here, we show that a p53-stabilizing Mdm2 alternative splicing event and the resulting widespread downregulation of metabolic transcripts are common events that arise in response to various splicing perturbations in both cellular and organismal models. Together, our results classify a common cellular response to splicing perturbations, put forth a new mechanism behind the cell-type-specific phenotypes that arise when splicing is broadly disrupted, and lend insight into the pleiotropic nature of the effects of p53 stabilization in disease.

Gene Ontology & Pathway Analysis Gene Ontology Over-Representation Analysis and PANTHER Pathway analysis were completed using WebGestalt (Liao et al., 2019).Background reference sets were compiled from all genes detected in DESeq2, all genes detected in rMATS output, or all genes detected in rMATS Exon Skipping output.For Differential Expression analysis, the top 1000 most significantly down or up-regulated genes were plotted.For the overlap sets, all down or all up-regulated genes were used as input.For rMATS analysis, genes that had at least one significant Alternative splicing or SE event were included.We did not account for multiple AS events in the same gene.Analysis was performed using the Bonferroni multiple test correction, with an FDR cutoff of 0.05.Terms with FDR < 0.05 were sorted on enrichment ratio, and the top 5-10 terms are displayed.Basic settings were set at Min # genes/cat 5; max 2000.

Compilation of gene set lists
Glycolysis GO terms GO:0030388, GO:0006003, GO:0006002, GO:0019255, GO:0051156, GO:0019682, GO:0060096 were combined to encapsulate all genes in pathway.We excluded terms from the resulting list whose role was only designated as regulatory or only designated as "Pentose Phosphate Pathway".This yielded a list of 64 genes we designated as core factors in the glycolysis pathway.Sterol gene list was compiled as a combination of GO:0016125, GO:0008203, GO:0016126, GO:0006695 excluding terms whose role was only designated as regulatory p53 responsive gene list was compiled from previous literature, specifically focusing on genes known to be positively regulated and promoters bound by p53 in mESCs (Bowen et al., 2019;Lee et al., 2010;Li et al., 2012) Zebrafish handling & husbandry Zebrafish were handled according to the vertebrate animal handling protocol.AB (wildtype) zebrafish were used for all experiments.For morpholino experiments, 2-cell-stage zebrafish were injected with 1-2nl 500µM p53-morpholino (GCGCCATTGCTTTGCAAGAATTG, GeneTools PCO-ZebrafishP53-100).Uninjected embryos were used as controls.
At 6 hpf, zebrafish were treated with TG003 or 2-Deoxy-D-Glucose (2DG, Sigma-Aldrich D8375-1G) in fish facility water.Drug was washed out at 24 hpf (18 hours-post-treatment), and no additional drug was added.Water was changed every subsequent day, and fish were collected at 5 dpf for cartilage staining.Fish were fixed, bleached, and stained with Alcian Blue as described previously (Dingerkus and Uhler, 1977;Sakata-Haga et al., 2018).Fish were imaged using a Leica M205 FCA stereo microscope, with the same settings for all samples.Craniofacial structure length was analyzed, blind to treatment groups, on Fiji.
Hybridization Chain Reaction 15 hpf zebrafish embryos were fixed overnight in 4% PFA at 4ºC, and dehydrated in methanol.Embryos were rehydrated, and HCR was carried out per manufacturer instructions (Molecular Instruments) with the sox10 (NM_131875) probe.Samples were imaged on an Olympus FV1200 Laser Scanning Confocal Microscope.     A. KEGG Pathway over-representation analysis of top 1000 genes most significantly upregulated under shRNA1 compared to nt-shRNA.B. KEGG Pathway over-representation analysis of all genes commonly significantly upregulated under both Eftud2-targeting shRNAs as compared to nt-shRNA C. Gene Ontology Biological Process over-representation analysis of all commonly significantly upregulated genes under both shRNAs when compared to nt-shRNA.D. KEGG Pathway analysis of top 1000 genes most significantly downregulated under shRNA1 compared to nt-shRNA.E. KEGG Pathway analysis of all genes commonly significantly downregulated under both Eftud2-targeting shRNAs as compared to nt-shRNA F. Gene Ontology Biological Process over-representation analysis of all genes commonly significantly downregulated under both shRNAs when compared to nt-shRNA. u Fig. S3.Mdm2 alternative splicing in p53-null cells A. Semi-quantitative Reverse Transcription PCR (sqRT-PCR) of Mdm2 under Eftud2 shRNA KD in both p53-proficient and p53-null backgrounds.Quantification of band intensity shown below.B. sqRT-PCR of Mdm2 under 24h 100µM TG003 treatment in both p53-proficient and p53null backgrounds.Quantification of band intensity shown below.
Fig. S5.Additional Gene Ontology and KEGG Over-representation Analysis of Eftud2 KD RNA-Seq Fig. S6.Volcano plots highlighting gene subsets of Eftud2 KD DESeq2A.Differential expression of p53-responsive transcripts in shRNA2, as shown in Fig.3BB.Differential expression of p53-responsive transcripts in shRNA1, as shown in Fig.3CC.Differential expression of sterol transcripts in shRNA2, as shown in Fig.3BD.Differential expression of sterol transcripts in shRNA1, as shown in Fig.3CE.Differential expression of glycolysis transcripts in shRNA2, as shown in Fig.3BF.Differential expression of glycolysis transcripts in shRNA1, as shown in Fig.3CG.Differential expression of glycolysis transcripts under shRNA2 KD, genes with significant alternative splicing events are highlighted in green.There are no glycolytic transcripts significantly alternatively spliced under shRNA1 KD.H. Differential expression of sterol transcripts under shRNA2 KD, transcripts with significant alternative splicing events are highlighted in green.
demonstrating additional measurements taken in the 5 dpf zebrafish ventral skeleton.Meckel's Cartilage Distance is used a proxy for width of the craniofacial complex.B. Measurements and ratios of cartilage lengths in 5 dpf TG003-treated zebrafish and controls.Significance values determined via two-tailed students' t-test.* p < 0.05, ** p < .01,*** p < .001,**** p < .0001~ p NS. C. Cartilage measurements and ratios in 5 dpf p53-morphant and control embryos, treated with DMSO or 15µM TG003.Significance and p-value designations as in B) D. sox10 expression in 15 hpf embryos treated with DMSO or 15µM TG003, as assayed by HCR.Scale bar represents 100µm.Arrows denote regions of interest in the anterior portion of the embryo.E. Measurements and ratios of cartilage lengths in 5 dpf 2DG-treated zebrafish and controls.Significance and p-value designations as in B).

Fig. S8 .
Fig. S8.Additional Characterization of the craniofacial phenotypes in zebrafish

Regulation of RNA stability Posttranscriptional regulation of gene expression All genes alternatively spliced in both shRNAs All genes with skipped exons in both shRNAs All genes with skipped exons in shRNA2 All genes alternatively spliced in shRNA2 A B D C Fig. S1. Gene Ontology analysis of alternatively spliced genes under Eftud2 KD
A Fig. S2.Mdm2 Mutually Exclusive Exon event detected under Eftud2 KD A. Diagram of the mutually exclusive exon event (MXE) detected by rMATs in Mdm2 and resulting isoforms with select protein domains annotated.The MXE event gives rise to isoforms that lack portions of the p53-Binding Domain.p53-BD,p53-Binding Domain; Acidic, Acidic Domain; Zn, Zinc Finger; RING, RING Domain.B.The ratio of Exon 3 inclusion and Exon 4 skipping to Exon 3 skipping and Exon 4 inclusionas detected by rMATs.C. "Inclusion" values of MXE events as detected by rMATS in shRNA1 (x-axis) and nt-shRNA (y-axis).Darker colors indicate more significant p-values.Both Mdm2 MXE events indicated are similar events, defined as separate events by rMATS due to slight differences in the upstream exon.D. "Inclusion" values of MXE events as detected by rMATS in shRNA2 (x-axis) and nt-shRNA (y-axis).Darker colors indicate more significant p-values.Both Mdm2 MXE events indicated are similar events, defined as separate events by rMATS due to slight differences in the upstream exon.Disease Models & Mechanisms: doi:10.1242/dmm.050356:Supplementary information Disease Models & Mechanisms • Supplementary information