Gene ontology (GO) analysis and characterization of SERBP1-interacting proteins.

A) SERBP1-interacting factors were identified via pull-down analysis in 293T cells. GO enrichment analysis (Biological Processes, Molecular Function, and Cellular Component) was performed with ShinyGO (27). A selection of the most relevant GO terms is represented. FE = fold enrichment; FDR = false discovery rate. B) Visualization of associated GO-enriched terms in three main regulatory branches: translation, ribosome biogenesis, and RNA processing. Sum and circle size represent the count of SERBP1 interactors assigned to each GO term. C) Representation of top enriched protein domains according to PFAM (42) among SERBP1 interactors. Only PFAM domains with counts of 5 or more and significant fold enrichment (FE) and false discovery rate (FDR) were included. The occurrence of each domain within the total human proteome (ALL) is listed as a reference. D) Occurrence of intrinsically disordered proteins (IDPs) (45) among SERBP1 interactors according to PFAM (“derived” = automatic entries from databases; “curated” = manually checked). E) Distribution of identified SERBP1-interacting proteins in reference to the presence of RGG boxes (50) and binding to G-quadruplexes (G4s) (46, 47). Data used to generate the figures are in Tables S1, S4 and S6.

SERBP1 is implicated in cell division.

A) SERBP1-associated proteins and their presence in specific cellular structures relevant to mitosis. B) Western blot showing SERBP1 knockdown in U251 and U343 cells. C) Multinucleated cells in U251 and U343 control and SERBP1 knockdown (KD) cells after treatment with paclitaxel, an agent that causes cell cycle arrest. SERBP1 KD increased the number of multinucleated nuclei. Data are shown as means of counts of 1000 cells ± standard deviation (in triplicate) and statistical significance was determined by Student’s t-test. P<0.001. D) Aspect of cells exposed to paclitaxel. On the left staining with anti-α-Tubulin; in the middle staining with DAPI showing an increased number of multinucleated cells after siSERBP1 KD. 100X magnification for detailed visualization of a single multinucleated cell. E) Plot shows the results of a cell viability screening (291 drugs) performed in U343 control vs. U343 SERBP1 over-expressing (OE) cells. Red dots correspond to drugs whose impact on cell viability was significantly different between U343 control and U343 SERBP1 OE. F) Highlights of the screening showing cell cycle/division and DNA damage/replication inhibitors whose impact on cell viability was smaller in U343 SERBP1 OE in comparison to control. Datasets used to prepare the figure and detailed analysis are in Table S1 and S7.

SERBP1 influences splicing.

A) SERBP1-interacting factors are preferentially associated with splicing. Networks show connections between splicing factors identified as SERBP1 interactors. B) SERBP1 knockdown in U251 cells produced major changes in splicing. Pie chart displays the distribution of splicing events affected by SERBP1 knockdown according to their type: SE, exon skipping; RI, intron retention; MXE, multiple exclusive exons; A5SS, alternative 5’ splice sites; A3SS, alternative 3’ splice site. C) Bar graphs showing the percentage of splicing events affected by SERBP1 knockdown with evidence of SERBP1 binding sites close (<100nt) to regulated splice sites. D) hnRNPU was identified as a potential partner of SERBP1 in splicing regulation. SERBP1 and hnRNPU display high expression correlation in normal and tumor tissues. E) Venn diagram shows a strong overlap between the splicing events affected by SERBP1 and hnRNPU knockdowns in U251 cells. F) Sashimi plots showing examples of splicing events affected by both SERBP1 and hnRNPU knockdowns in U251 cells. The red arrows indicate affected exons. Datasets used to prepare the figure and detailed analyses are in Table S8.

SERBP1 interacts with PARP1 and influences PARylation.

A) Results of IP-western in U251 cells with control and anti-SERBP1 antibodies confirm SERBP1 interaction with PARP1, NCL and SYNCRIP. SERBP1 and PARP1 strong association is corroborated by their high expression correlation in different studies (B) and similar expression profiles during cortex development according to Cortecon (142). (C) . D) PARylation sites observed in SERBP1 protein according to (62, 63). E) PARP1 ADP-ribosylates SERBP1 in vitro. Purified recombinant PARP1 (0.1 µM) and SERBP1 (0.6 µM) were combined in a reaction with or without sheared salmon sperm DNA (sssDNA) (100 ng/µL) and NAD+ (100 µM) as indicated. The reaction products were analyzed by SDS-PAGE with silver staining (left) and Western blotting for PAR (right). Uncropped gels are shown. F) Venn diagram shows that majority of SERBP1-associated factors get PARylated and/or bind PAR (6264). G) Increase of PARylation levels in 293T and U251 GBM cells after H2O2 treatment. H) SBP-SERBP1 and Flag-PARP1 were co-transfected into 293T cells. Cells were treated with H2O2 to induce PARylation and a pull-down experiment with streptavidin beads was performed. Western analysis showed increased SERBP1 association with PARP1 in cells treated with H2O2. SBP-SERBP1-His detected by His antibody; Flag-PARP1 detected by Flag antibody. I) SBP-SERBP1 was transfected into 293T cells. Cells were treated with DMSO or 10μM PJ34 (PARP inhibitor) for two hours and a pull-down experiment with streptavidin beads was performed. Western analysis showed a decrease in SERBP1 association with PARP1 and GPBP1. J) SERBP1 transgenic expression (mGreen-SERBP1) in 293T cells increased the levels of PARylated proteins as indicated by PAR-detecting agent. Datasets used to prepare the figure and detailed analyses are in Tables S1, S5 and S6.

Shared SERBP1 and PARP1 interactors.

A) Venn diagram shows the overlap between PARP1 (66) and SERBP1 interactomes. Analysis of shared interactors indicated that the majority of them are PAR binding (64) and/or get PARylated (62, 63). B) Selection of enriched GO terms (biological processes) related to SERBP1-PARP1 shared interactors according to ShinyGO (27). FE = fold enrichment; FDR = false discovery rate. C) Network showing shared SERBP1 and PARP1 interactors implicated in ribosome biogenesis. D) Proposed SERBP1-PARP1 feedback model; SERBP1 function and association with partner proteins is modulated by PARylation while SERBP1 influences PARP activity. Datasets used to prepare the figure and detailed analyses are in Tables S1 and S6.

SERBP1-associated proteins are present in membraneless organelles.

A) Different types of membraneless organelles in a cell and examples of SERBP1-associated proteins present in each structure. B) Membraneless organelles identified in the GO enrichment analysis (cellular component) of SERBP1 interactors (72). FE = fold enrichment; FDR = false discovery rate. C) Bar graph showing the distribution of SERBP1-associated proteins in different membraneless organelles. D) Venn diagram representing the distribution of SERBP1-associated proteins present in membraneless organelles in respect to G4 binding (46, 47), PAR binding (64), and PARylation (62, 63). Datasets used to prepare the figure and detailed analyses are in Tables S1, S6 and S9.

SERBP1 is present in Tau aggregates.

A) Representative co-immunofluorescence of SERBP1 and Phospho-Tau (Thr231) in AD brain tissues. Merged channel is represented. DAPI was used to stain nuclei. Magnification 20x and white scale bar: 50 µm. Three different Insets selected from merged channels are represented in zoomed images as merged, SERBP1 (green) and Phospho-Tau (red). B) Representative PLA of SERBP1 vs. pTau in AD and aged-matched control brains (magnification 40x and white scale bar: 50µm). C) Percentage of positive area to PLA fluorescence in region of interests in AD and control brains (Ctr vs AD, *** p<0.001, paired t-test). D) Western blot showing SERBP1 expression in normal and AD brains and presence of oligomers. GAPDH and Tau oligomers immunoblot are shown. E) SERBP1 relative density in Ctr and AD brains, quantified as function of GAPDH (from immunoblot in D). (Ctr vs. AD, * p<0.05, paired t-test). F) Venn diagram shows overlap between proteins often identified in Tau aggregates and SERBP1 associated proteins identified in this study (top). Gene Ontology (GO) enrichment analysis (Biological Processes) using ShinyGO (27) indicated that SERBP1-associated factors that are also present in Tau aggregates are implicated in ribosome biogenesis, translation, splicing and telome maintenance. FE = fold enrichment; FDR = false discovery rate. Datasets used to prepare the figure and detailed analyses are in Tables S1 and S9.

SERBP1 association with PARP1 in Alzheimer’s brains.

A) Representative co-immunofluorescence of SERBP1 and PARP1 in AD and age-matched control brain tissues. PARP1 and SERBP1 are represented in gray while merged images represent PARP1 (red), SERBP1 (green). DAPI was used to stain nuclei (blue). Magnification 20x and white scale bar: 100 µm. Each frame has a zoomed inset representing the detailed distribution of each target. B) Pearson Coefficient (PCC) of co-localizing SERBP1 and PARP1 between cells of age-matched control and AD brains (CTR vs. AD, **** p<0.001 paired t-test). C) Fluorescence intensity profiles of PARP1 (red), SERBP1 (green), and DAPI (blue) in representative cells from age-matched control and AD brains. Distance is represented in pixels and fluorescence intensity as Grey value obtained using ImageJ FIJI software. D) Representative PLA of SERBP1 vs. PARP1 in AD and aged-matched control brains (magnification 40x and white scale bar: 50µm). Percentage of positive area to PLA fluorescence in region of interests in AD and control brains (Ctr vs AD, ** p<0.01, paired t-test).

Table 1: Brain tissues analyzed in this study from diseased and age-matched non-demented control subjects are summarized.