Identification of FAM53C as a positive regulator of cell cycle progression in G1.

A. Strategy to identify new cell cycle regulators using the DepMap database. See Supplementary Table S1 for the complete list of the 38 genes and associated DepMap data. B. Schematic representation of key factors in the G1/S machinery and scores in the screen for selected factors. CDKi, CDK inhibitors; RBBP, RB binding protein. C. Representation of the number of genes with different overlap scores in the DepMap analysis. RB1 and FAM53C are indicated; the maximal overlap score was 13 for the CDK2 (not shown on graph). The cut-off overlap score value for top candidates was set at 3 (top 5%). See Supplementary Table S2 for complete data. D. Immunoblot analysis for FAM53C in knock-down (siFAM53C) RPE-1 cells compared to controls (siCtrl) at 24 h and 48 h. HSP90 serves as a loading control. Molecular weights are indicated in kDa. *, non-specific signal. E. Cell cycle analysis by BrdU/PI staining in control and FAM53C knock-down RPE-1 cells (N=4). F. Population growth analysis (cell counts) in control and FAM53C knock-down RPE-1 cells (N=3-5). G. Representative (n=3) immunoblot analysis for FAM53C overexpression (HA-tagged FAM53C) compared to control cells (with GFP expression) in RPE-1 cells. GAPDH serves as a loading control. Molecular weights are indicated in kDa. H. Cell cycle analysis by BrdU/PI staining in control and FAM53C-overexpressing RPE-1 cells (n=4). I. Population growth analysis in control and FAM53C-overexpressing RPE-=1 cells (n=3-5). J. Fraction of cells in G1 from BrdU/PI staining in control (wild-type, WT) and RB knockout (KO) RPE-1 cells, with or without FAM53C knock-down (N=4). K. Fraction of cells in G1 from BrdU/PI staining in control (GFP) and FAM53C-overexpressing RPE-1 cells, with or without palbociclib treatment (N=3). L. Cartoon placing FAM53C in the G1/S transition of the cell cycle. P-values for (EH), (J), and (K) were calculated by paired t-test. P-values for (F) and (I) were calculated by mixed-model ANOVA followed by post-hoc paired two-tail t-test.

The FAM53C interactome identifies cell cycle factors.

A. Cartoon representation of the AP/MS experiment to determine the FAM53C interactome. B. Top cell cycle interactors from the AP/MS experiment. See Supplementary Tables S3 and S4 for the complete list. C. CytoTRACE analysis integrating the results of the AP-MS experiment with public databases. Solid lines indicate p ≤ 0.05, and line thickness correlates to pNSAF from AP-MS experiment. D. Biolayer interferometry assay to measure the binding of recombinant FAM53C to DYRK1A-coated streptavidin (SA) sensors. Association begins a 0 seconds. Dotted line indicates start of dissociation phase. Black solid line indicates portion of the curve that was analyzed for on and off rates. Reported rate constants are from data fitting of observed on and off rates and the equilibrium constant was determined from the steady-state response analysis.

FAM53C can inhibit DYRK1A function in cells

A. In vitro phosphorylation assay. Recombinantly expressed Cyclin D1 (1 µM) was phosphorylated by DYRK1A (200 nM) for 15 min alone or in the presence of increasing amounts of FAM53C. FAM53C concentration was 5 µM (lane #3) or 1, 2.5, 5, and 10 µM (lanes #5-8). Lane #1 contains DYRK1A without substrate. The DYRK1A kinase inhibitor Harmine was added at 25 µM. Representative experiment of N=3 experiments. B. Quantification of immunoassays for Cyclin D1 protein levels relative to the loading control HSP90 in FAM53C knock-down (siFAM53C) RPE-1 cells compared to controls (siCtrl) at 24 h (N=3). C. Quantification of immunoassays for p21 protein levels relative to the loading control HSP90 in FAM53C knock-down (siFAM53C) RPE-1 cells compared to controls (siCtrl) at 24 h (N=3). D. Quantification of immunoassays for Cyclin D1 protein levels relative to the loading control HSP90 in RPE-1 cells expressing HA-FAM53C compared to GFP controls (N=3). E. Quantification of immunoassays for p21 protein levels relative to the loading control HSP90 in RPE-1 cells expressing HA-FAM53C compared to GFP controls (N=3). F. Quantification of immunoassays for Cyclin D1 protein levels in control and FAM53C knock-down RPE-1 cells treated with or without different concentrations of the SM13797 DYRK1Ai (N=3 per concentration, 48 hours of knock-down and treatment). G. Fraction of FAM53C knock-down (siFAM53C) RPE-1 cells in G1 compared to controls (siCtrl), with or without DYRK1Ai treatment (N=5). P-values for (B), (C), (D), and (E) were calculated by paired t-test. P-value for (F) was calculated by 2-way ANOVA.

FAM53C knock-down activates p53

A. Fold-change analysis of the p53 target genes in FAM53C knock-down cells compared to controls (RNA-seq data). Black dot: P-value>0.05. B. Immunoassay for p21 in p53 wild-type (WT) and knockout (KO) RPE-1 cells, with (siFAM53C) or without (siCtrl) FAM53C knock-down. HSP90 serves as a loading control (N=1). Values on the bottom represent the ratio between the signal for p21 and the signal for HSP90 in the same lane. C. Fraction of cells in G1 in RPE-1 cells, with or without FAM53C knock-down, with or without p21 knock-down (p21 is encoded by CDKN2A), with or without treatment with the DYRK1A inhibitor (DYRK1Ai) (N=3). DMSO is a control for DYRK1Ai. Note that the G1 arrest observed upon FAM53C knock-down is still present in cells with p21 knock-down and DYRK1Ai treatment (ns, not significant). D. Fraction of FAM53C knock-down (siFAM53C) p53 knockout RPE-1 cells in G1 compared to controls (siCtrl), with or without DYRK1Ai treatment (N=4). Note that the G1 arrest observed upon FAM53C knock-down is absent in cells with p53 knockout and DYRK1Ai treatment. E. Cell counts for p53 knockout RPE-1 cells, with or without FAM53C knock-down, with or without DYRK1Ai treatment (N=2). Note that cells with FAM53C knock-down and DYRK1Ai treatment do not proliferate despite the decrease in G1 fraction in (D). F. Fraction of FAM53C knock-down (siFAM53C) HCT-116 cells in G1 compared to controls (siCtrl), with or without DYRK1Ai treatment (N=4). Note that the G1 arrest observed upon FAM53C knock-down is absent in cells with p53 knockout and DYRK1Ai treatment. P-values for (C), (D), and (F) were calculated by paired t-test.

Loss of FAM53C impairs the development of human cortical organoids.

A. Cartoon of the differentiation protocol from human induced pluripotent stem cells (iPSCs) to human cortical organoids (hCOs). B. Representative images of wild-type and knockout hCOs. Scale bar, 500 µm. C. Quantification of (B). D. Quantification of Edu-positive cells in wild-type and knockout hCOs. E. Representative immunoblot analysis of wild-type (WT) and FAM53C knockout (KO) hCOs. β-actin serves as a loading control. F. Quantification of (E) for DYRK1A relative to β-actin. G. Quantification of (E) for phospho-Cyclin D1 relative to Cyclin D1 levels. H. Quantification of (E) for p21 levels relative to β-actin levels. P-values for (C), (D), (F), (G), and (H) calculated by paired t-test.

Fam53C knockout mice are viable and display limited phenotypes

A. Cartoon of the Fam53C mutant allele, with deletion of the major coding exon (exon 4) (not to scale). B. Genotypes of mouse pups at weaning from Fam53C+/- crosses at the Stanford facility. WT: wild-type; Het, heterozygous mutant mice; Mut, homozygous mutant mice. C. Body weight analysis at weaning for mice generated by Fam53C+/- crosses. D. Measure of the latency to first transition into a dark chamber (Light-Dark test) for control (n=1998 males and n=2037 females – including historical controls) and Fam53C knockout mice (n=6 males and n=7 females). Note the extremely different sizes in the cohorts of control and mutant mice, which may affect the results of the statistical analysis. P-values for (B) calculated by a Chi-squared test. P-values for (C) calculated by unpaired t tests. The details of the Linear Mixed Model framework for statistical analysis for (D) are available on the IMPC web site.

Identification of FAM53C as a positive regulator of cell cycle progression in G1.

A. Representative example of flow cytometry analysis for BrdU/PI staining of RPE-1 cells with control siRNAs and FAM53C siRNAs (related to Figure 1E). B. Fraction of cells in G1 from BrdU/PI staining in RPE-1 cells, with or without FAM53C knock-down, using two individual siRNAs (N=3). C. Representative example of flow cytometry analysis for Annexin V/PI staining of RPE-1 cells with control siRNAs and FAM53C siRNAs. The double negative cells are the live cells. D. Quantification of (C) (N=3). E. Cell cycle analysis by BrdU/PI staining in control and FAM53C knock-down U2OS cells (N=3). F. Cell cycle analysis by BrdU/PI staining in control and FAM53C knock-down A549 cells (N=3). G. Representative example of flow cytometry analysis for BrdU/PI staining of RPE-1 cells with GFP or FAM53C overexpression. H. Representative example of flow cytometry analysis for BrdU/PI staining of RPE-1 cells in control and RB knockout RPE-1 cells as in Figure 1J. I. Representative example of flow cytometry analysis for BrdU/PI staining of RPE-1 cells in control and FAM53C-overexpressing RPE-1 cells as in Figure 1K. P-values for (B), (D), (E), and (F) were calculated by t-test.

The FAM53C interactome identifies cell cycle factors.

A. Silver stain of protein fractions in the AP/MS experiment. Sin: S-bead input (cleavage of protein from the GFP-beads before binding to s-beads); SFT: S-bead flowthrough (solution after binding to S-beads). El1, El2: elution 1 and elution 2 from S-beads.

FAM53C can inhibit DYRK1A function in cells

A. Recombinantly expressed LIN52 (10 µM) was phosphorylated by DYRK1A (400 nM) for 30 min alone or in the presence of increasing amounts of FAM53C. FAM53C concentration was 1.6, 5, 15 µM (lanes #4-6). Lane #1 contains DYRK1A without substrate. Representative experiment of N=2 experiments. B. Representative immunoassay for Cyclin D1 protein levels relative to the loading control HSP90 in FAM53C knock-down (siFAM53C) RPE-1 cells compared to controls (siCtrl) at 24 h (Related to Figure 3B). C. Representative immunoassay for p21 protein levels relative to the loading control HSP90 in FAM53C knock-down (siFAM53C) RPE-1 cells compared to controls (siCtrl) at 24 h (Related to Figure 3C). D. Representative immunoassay for Cyclin D1 protein levels relative to the loading control HSP90 in RPE-1 cells expressing HA-FAM53C compared to GFP controls (Related to Figure 3D). E. Representative immunoassay for p21 protein levels relative to the loading control HSP90 in RPE-1 cells expressing HA-FAM53C compared to GFP controls (Related to Figure 3E). F. SM13797 biochemical IC50 values determined using the ThermoFisher LanthaScreen platform for CLK4 or Z’LYTE™ platform for the other kinases. G. List of kinases inhibited at 90% or more by SM13797 (1 µM) using ThermoFisher SelectScreen service and showing IC50 within 20-fold of that of DYRK1A. IC50 determination for DYRK1B, DYRK2 and DYRK3 were performed by ThermoFisher SelectScreen service. H. Cellular target engagement profile of SM13797 against CLK/DYRK family members. Target engagement assay IC50 values were determined using the Promega NanoBRET® TE Intracellular Kinase Assay platform in transiently transfected HEK293T cells. I. Representative immunoassay for Cyclin D1 protein levels in RPE-1 cells treated with or without different concentration of the SM13797 DYRK1Ai (related to Figure 3F). J. Representative immunoassay for p21 protein levels in RPE-1 cells treated with or without different concentration of the SM13797 DYRK1Ai. K. Quantification of immunoassays for p21 protein levels in RPE-1 cells treated with or without different concentration of the SM13797 DYRK1Ai, as in (I) (N=3 per concentration). P-value for (K) was calculated by 2-way ANOVA.

FAM53C knock-down activates p53

A. Volcano plot from the RNA-seq data comparing control and FAM53C knock-down RPE-1 cells (48 h after siRNA transfection). See Supplementary Table S5. B. Fold-change analysis of the 38 genes selected in the initial DepMap screen (Supplementary Table S1) in FAM53C knock-down cells compared to controls (from the RNA-seq data). Pink dots: P-value<0.05. C. Immunoassay for p21 (encoded by CDKN2A) in RPE-1 cells treated with DMSO or the DYRK1Ai, with or without FAM53C knock-down (48 hours of knock-down and treatment). HSP90 serves as a loading control (N=3). D. Quantification of (C) (N=3). E. Representative example of flow cytometry analysis for PI staining of p53 knockout (KO) RPE-1 cells treated with DMSO or the DYRK1Ai, with or without FAM53C knock-down. F. Fraction of cells in G1 in control (wild-type, WT) and p53 knockout (KO) RPE-1 cells, with or without FAM53C knock-down (N=3), as in (E). Note that a G1 arrest is still present in p53 knockout cells with the FAM53C knock-down. G. Representative example of flow cytometry analysis for PI staining of p53 wild-type (WT) RPE-1 cells treated with DMSO or the DYRK1Ai, with or without FAM53C knock-down. H. Fraction of cells in G1 in p53 WT RPE-1 cells as in (G) (N=5). I. Immunoblot for cleaved caspase 3 (CC3) in p53 wild-type and knockout RPE-1 cells, with or without FAM53C knock-down, and with or without DYRK1Ai treatment. HSP90 serves as a loading control (N=1). I. Fraction of cells in G1 in RPE-1 as in (H). Note that cells arrest in G1 upon FAM53C knock-down and that this arrest is not rescued by the p21 knock-down (ns, not significant). J. Correlation for the gene effects of CCND1 and Tp53 (coding for Cyclin D1 and p53, respectively) in the DepMap CRISPR dataset. Four cell lines are highlighted, three derivatives of RPE-1 cells and HCT-116. K. Representative example of flow cytometry analysis for PI staining of HCT-116 cells, with or without FAM53C knock-down, with or without DYRK1Ai treatment, as in Figure 4F. L. Cartoon placing FAM53C in the G1/S transition of the cell cycle upstream of the RB and p53 pathways. P-values for (F) and (H) were calculated by paired t-test.

Loss of FAM53C impairs the development of human cortical organoids.

A. Cartoon of the FAM53C human gene, with the location of the sgRNAs used to knockout the gene in iPSCs (not to scale). B. Sequence histogram showing gene inactivation upon CRISPR/Cas9-mediated deletion. C. Sequence histogram showing no changes in the sequence of possible off target genes. D. Representative images of wild-type and knockout iPSC colonies. Scale bar, 200 µm. E. Representative images of immunofluorescence staining of pluripotency markers in wild-type and knockout iPSCs. Scale bar, 50 µm. F. Representative example of flow cytometry analysis for EdU/DAPI staining of hCOs, as in Figure 4E.

Fam53C knockout mice are viable and display limited phenotypes

A. Genotypes of mouse pups at weaning from Fam53C+/- crosses at the IMPC facility. B. Body weight analysis of a large cohort of historical controls and Fam53C-/- mice. C. Histological analyses (sections stained with hematoxylin and eosin, H&E) for the indicated tissues from mice generated by Fam53C+/- crosses. Representative images for one wild-type and one knockout female are shown out of two females and two males analyzed (age, 18 months). Scale bar, 2 mm. D. Survival analysis for N=23 Fam53C-/- mice (7 males and 15 females) and N=16 wild-type controls (7 males and 9 females) from heterozygous crosses. The analysis was stopped at 18 months.