Abstract
Maintenance of stemness is tightly linked to cell cycle regulation through protein phosphorylation by cyclin‐dependent kinases (CDKs). However, how this process is reversed during differentiation is unknown. We report here that exit from stemness and differentiation of pluripotent cells along the neural lineage are controlled by CDC14, a CDK‐counteracting phosphatase whose function in mammals remains obscure. Lack of the two CDC14 family members, CDC14A and CDC14B, results in deficient development of the neural system in the mouse and impairs neural differentiation from embryonic stem cells (ESCs). Mechanistically, CDC14 directly dephosphorylates specific proline‐directed Ser/Thr residues of undifferentiated embryonic transcription Factor 1 (UTF1) during the exit from stemness, triggering its proteasome‐dependent degradation. Multiomic single‐cell analysis of transcription and chromatin accessibility in differentiating ESCs suggests that increased UTF1 levels in the absence of CDC14 prevent the proper firing of bivalent promoters required for differentiation. CDC14 phosphatases are dispensable for mitotic exit, suggesting that CDC14 phosphatases have evolved to control stemness rather than cell cycle exit and establish the CDK‐CDC14 axis as a critical molecular switch for linking cell cycle regulation and self‐renewal.
Keywords: CDC14 phosphatase, epigenetics, neural differentiation, pluripotent stem cells, UTF1
Subject Categories: Cell Cycle, Development, Stem Cells & Regenerative Medicine
While yeast CDC14 has critical roles in mitotic exit, loss of mouse CDC14A/B does not affect proliferation, but neural differentiation via the epigenetic regulator UTF1.
Introduction
Pluripotent embryonic stem cells (ESCs) derive from the inner cell mass of the blastocyst. They have the capacity to self‐renew and to differentiate into every cell type of the body, but cannot form extra‐embryonic tissues such as the placenta. As ESCs differentiate into somatic cells, they lose these capacities and commit to a single‐cell lineage. This process takes place with a profound restructuration of their chromatin into a more closed configuration that will eventually result in the expression of the lineage specific genes and is also accompanied by important changes in cell cycle. However, the molecular machinery that links cell cycle, differentiation, and epigenetic rearrangements remains obscure.
Recent evidences suggest that the cell cycle machinery links cell proliferation with pluripotency and cell fate specification. Proper progression through the different phases of the cell cycle is controlled by the activation of cyclin‐dependent kinases (CDKs), a family of Ser/Thr kinases activated by cyclin subunits and the sequential phosphorylation of a large number of targets (Morgan, 2007; Malumbres, 2014). Whereas cell cycle CDKs are transiently activated at different cell cycle phases in most somatic cells, CDK1 and CDK2 complexes with cyclin E and A subunits are constitutively active in pluripotent stem cells (reviewed in Liu et al, 2019). CDKs are thought to maintain stemness through multiple mechanisms including control of the PI3K/AKT pathway, regulation of specific stemness factors such as OCT4, SOX2, or NANOG, or the control of multiple epigenetic factors (Liu et al, 2017; Wang et al, 2017; Kim et al, 2018; Michowski et al, 2020). Whereas inactivation or knockdown of these cyclins and CDKs results in the loss of the pluripotent state and triggers differentiation, it is unclear how the status of CDK targets is reversed during exit from stemness in unperturbed conditions.
Recent studies in eukaryotes have revealed that both kinase and phosphatase regulations are essential for the timely and accurate execution of cell cycle transitions and checkpoints (Wurzenberger & Gerlich, 2011; Nilsson, 2019). Mitotic exit, specifically, does not only require CDK inactivation but also the activation of CDK‐counteracting phosphatases and the subsequent removal of mitotic phosphorylations from their substrates. In Saccharomyces cerevisiae, the dual‐specificity phosphatase CDC14 is an essential regulator of both CDK inactivation and dephosphorylation of CDK substrates during mitotic exit (Visintin et al, 1998; Stegmeier & Amon, 2004). However, their relevance during mitotic exit in higher organisms remains unclear due to the existence of contradictory results in different studies (Mocciaro & Schiebel, 2010). Vertebrate cells contain two CDC14 paralogs, CDC14A and CDC14B, with a third family member, CDC14C—the product of a near pseudogene of CDC14B—exclusively found in humans (Rosso et al, 2008; Mocciaro & Schiebel, 2010). Genetic studies in somatic cell lines in which either CDC14A or CDC14B was deleted suggest no defects in cell proliferation or viability (Berdougo et al, 2008; Mocciaro et al, 2010). Cdc14b‐null mice are viable and display an aging‐like phenotype but no defects in the cell cycle (Guillamot et al, 2011; Wei et al, 2011). Knockdown of Cdc14a in Cdc14b‐null somatic cells results in enhanced defects in DNA repair without obvious defects in cell cycle progression (Lin et al, 2015). Recent data in cultured RPE1 cells in which both CDC14A and CDC14B genes have been concomitantly knocked out suggest some defects in ciliogenesis without obvious alterations in cell cycle progression (Partscht et al, 2021).
By generating Cdc14a‐null mice and a novel double Cdc14a and Cdc14b knockout model, we report in this work that CDC14 phosphatases are dispensable for mitotic exit in somatic cells but play a critical role in exit from stemness and differentiation of ESCs. CDC14 phosphatases directly dephosphorylate UTF1, an epigenetic regulator with critical roles in chromatin dynamics and gene transcription in stem cells. In pluripotent cells, UTF1 stability is maintained by phosphorylation of proline‐directed Ser/Thr residues by CDK or MAPK, whereas exit from stemness is driven by CDC14‐dependent dephosphorylation leading to decreased UTF1 stability and proper firing of bivalent promoters and expression of genes required for differentiation.
Results
Lack of CDC14 does not alter cell cycle progression but results in partial perinatal lethality
We first generated a Cdc14a knockout allele by targeting exon 3 with loxP sites using a classical homologous recombination strategy in ESCs (Appendix Fig S1a and b). Removal of this exon results in an open reading frame shift and lack of Cdc14a transcripts (Appendix Fig S1c and d). Lack of CDC14A in Cdc14a(−/−) homozygous mutants did not result in any obvious abnormality during development or adult animals. We therefore generated a double Cdc14a; Cdc14b‐null mouse model by combining the Cdc14a(−) allele with the Cdc14b‐deficient model previously generated in our lab (Guillamot et al, 2011). Similarly to the single mutants, Cdc14a(−/−); Cdc14b(−/−) double knockout (Cdc14 DKO) mice are also viable, although 50% of the Cdc14a(−/−); Cdc14b(−/−) pups died within the first 3 weeks of life (Fig 1A) showing reduced size and mobility (Fig 1B and C). Strikingly, Cdc14 DKO mice that were able to survive the perinatal period presented a similar life expectancy than control littermates (Appendix Fig S1e), without major alterations in blood cell counts (Appendix Fig S1f).
Figure 1. Altered brain development in CDC14‐deficient mice.
- Survival curves of wild type (n = 27) and Cdc14 DKO (n = 51) mice during the first 6 weeks of life (***P < 0.001; Log‐rank, Mantel‐Cox test).
- Representative images of postnatal day 9 Cdc14 DKO and wild‐type littermates.
- Weight of Cdc14 DKO and wild‐type littermates during the first 3 weeks of life (n = 4 per genotype).
- Representative macroscopic pictures of brain dorsal view of wild‐type (WT) and Cdc14 KO mice at postnatal 0, 5 and 9 days (P0, P5 and P9). Scale bar, 2 mm. The bar plots show the quantification of brain width, length, and area, respectively.
- Representative cresyl violet staining images of midsagittal sections of cerebellar vermis of WT and Cdc14 KO mice at P0, P5 and P9. Fissures underdeveloped in mutant samples are labeled with arrowheads. Scale bar, 700 μm. The bar plots show the quantification of the cerebellar vermis area, the number of folia, and the average length of the fissures, respectively.
- Detailed cresyl violet staining images of midsagittal sections of cerebellar vermis of the same samples. The molecular layer (ML), internal granule cell layer (IGL), and external granule cell layer (EGL) are indicated. The width of these layers is shown in the bar plots.
- Representative images of misdigitalhree sections stained with DAPI (blue), Calbindin (green), and GFAP (red) of cerebellar vermis of wild‐type ad Cdc14 DKO mice at 9 days after birth. Scale bar, 50 μm. Plots show the quantification of Purkinje cell dendritic length (average distance of 30 random measurements) and the number of cells per 100 μm from three different fields from three different mice in each plot.
Data information: Bars indicate mean ± SEM. In (A–G), *P > 0.05; **P < 0.01; ***P < 0.001, Student's t‐test with Welch's correction. Non‐significant data (P > 0.05) are not labeled. In (D–F), data show the individual data, average and SEM from three WT and three DKO mice. In (F), each point represents the average value of nine different random measurements across distinctive folia for each mouse.
We initially explored possible defects in cell cycle progression using mouse embryonic fibroblasts (MEFs). No defects were observed in the number of cells entering into S‐phase in either primary or immortalized Cdc14 DKO MEFs released from serum starvation (Appendix Fig S2a). Similarly, no differences were observed in the kinetics of mitotic entry, as determined by cell rounding and chromosome condensation (Appendix Fig S2b) or by the accumulation of the mitotic markers, such as cyclin B1, phospho‐Histone H3 S10, Aurora B, or phospho‐CDK1 (Appendix Fig S2c). Both control and Cdc14 DKO cells spent approximately 50 min in mitosis (49.0 ± 8.3 min vs. 48.4 ± 9.8 min; Appendix Fig S2d), with no differences in mitotic defects as compared with control cultures (Appendix Fig S2e). We also observed normal Spindle Assembly Checkpoint (SAC) proficiency after inducing prolonged mitotic arrest with taxol (duration of mitosis in DKO cells = 342.8 ± 92.7 min vs. 335.9 ± 74.6 min in controls) or nocodazole (580.4 ± 118.4 min vs. 582.2 ± 113.7 min; Appendix Fig S2f). In agreement with these results, mutant and control MEFs proliferated at comparable rates (Appendix Fig S2g and h), indicating that CDC14 protein phosphatases are dispensable for cell cycle progression and normal cell proliferation, at least in embryonic fibroblasts.
Defective neural development and differentiation in the absence of CDC14
Histopathological analysis of mutant and control newborns did not show major architectural defects in most tissues analyzed. However, we did observe alterations in the brain, which, in addition to displaying reduced size (Fig 1D), presented specific structural defects in the cerebellar vermis (Fig 1E). The development of the cerebellum is postnatal; at birth, precursors of Granule cells proliferate in the external granule layer (EGL) and start differentiating. These precursors migrate through the molecular layer (ML) and get to the internal granule layer (IGL), where they reach their terminal differentiation. During the perinatal period, the EGL width decreases and disappears by postnatal day 21 (P21). Cdc14 DKO mice that died during the perinatal period showed enlarged EGL and decreased ML (Fig 1F), with concomitant defects in Purkinje cell numbers and development (Fig 1G). Immunofluorescence analysis of neural markers in these samples suggested specific defects in neural differentiation and developmental stage in mutant samples (Fig EV1A). These observations are in agreement with the fact that human CDC14A and CDC14B phosphatases are mostly expressed in the brain in adult tissues (Fig EV1B).
Figure EV1. Expression of CDC14 in human tissues and normal cell cycle progression in Cdc14 double knockout embryonic stem cells (ESCs).
- Immunofluorescence analysis of the indicated proteins in brain sections from embryonic (E)14.5 (left) or postnatal day (P)9 pups (right) in wild‐type and Cdc14 DKO littermates. Scale bars, 100 μm.
- Single‐cell RNA‐seq analysis of human CDC14A and CDC14B transcripts in human tissues and cell types (data from Protein Atlas; https://www.proteinatlas.org/).
- Representative histograms of EdU staining obtained by flow cytometry in wild‐type (WT, red) and Cdc14 DKO (blue) mESCs.
- Quantification of the percentage of EdU‐positive cells obtained by flow cytometry in wild‐type and Cdc14 DKO mESCs. Six independent experiments were performed with two different clones. Bars are mean ± SEM. Statistic significance was assessed using the two‐tailed Student's t‐test. ns, P > 0.05.
- Cell proliferation curve over 7 days of wild‐type and Cdc14 DKO mESCs, showing no significant differences in cell proliferation. Data are mean + SEM. Four independent experiments were performed with two different clones.
- Representative confocal microscopy images of WT and Cdc14 DKO mESCs stained with EdU (magenta), α‐tubulin (green) and DAPI (blue). The percentage of mitotic (M), EdU‐stained (EdU+), and non‐stained (EdU−) cells was quantified (right). n > 600 cells per genotype. Scale bar, 20 μm.
- Neural progenitors were extracted from the subventricular zone of postnatal day 6 (P6) pups and subjected to EdU‐propidium iodide analysis by cytometry (central panels). The quantification of cell cycle phases in wild‐type (n = 2 independent cultures) and Cdc14 DKO (n = 3) samples is shown in the bar plots to the right. Data are shown as mean ± SD. ns, P > 0.05, Student's t‐test with Welch's correction.
To study neural differentiation in this mutant model, we first generated control and Cdc14‐deficient embryonic stem cells (ESCs) from embryos. Similar to MEFs, mutant ESCs did not display any obvious defects in cell cycle progression or proliferation (Fig EV1C–F). Similarly, we also confirmed that Cdc14 ablation did not alter cell cycle properties of neural progenitors isolated from P6 Cdc14 DKO brains (Fig EV1G). However, whereas wild‐type ESCs efficiently generated neuronal (labeled with TUJ1) and glial (GFAP+) cells, neural differentiation was significantly impaired in Cdc14 DKO cultures (Fig 2A). The induction of several neuronal markers such as Tuj1 and Rbfox1, the astrocytic marker Gfap, and the oligodendrocytic marker Claudin 11 (Clau11) was also impaired in Cdc14‐deficient cultures (Fig 2B). The expression of the neural progenitor protein Nestin, which peaks at earlier stages during differentiation, was also reduced in Cdc14 DKO cells (Fig 2C), indicating that lack of CDC14 affects neural differentiation from early stages.
Figure 2. Deficient neural differentiation in CDC14‐deficient ESCs.
- Confocal microscopy analysis of wild‐type and Cdc14 DKO embryonic stem cells (ESCs) subjected to neural differentiation. Neuronal marker TUJ1 (green), glial marker GFAP (red), and nuclear marker DAPI (blue) are stained and quantified after 18 days in differentiation media. Each dot represents the ratio between TUJ or GFAP and DAPI area in an image containing at least 100 cells from four independent experiments with two different clones. ****P‐value < 0.0001. Student's t‐test. Scale bar 50 μm.
- qPCR analysis of different neuronal and glial markers in WT and Cdc14 DKO ESCs at day 0 or after 6 and 18 days in differentiation media. Bars indicate the average from three biological replicates with technical triplicates per assay. *P < 0.05; Students' t‐test.
- Confocal microscopy analysis of WT and Cdc14 DKO ESC subjected to neural differentiation. Neural progenitor marker Nestin was stained and quantified after 6 days in differentiation media. Each dot represents the ratio between Nestin and DAPI area in an image containing at least 100 cells. ****P < 0.0001; Students' t‐test.
- Heatmap analysis of selected genes obtained from RNA‐seq data in the indicated groups. The color scale indicates relative induction (red) or repression (blue).
- Gene set enrichment analysis (GSEA) of specific pathways during differentiation. NES, Normalized enrichment score.
- Representative GSEA analysis of pathways down‐ or upregulated in Cdc14 DKO cells at day 7 during the differentiation protocol. The Normalized Enrichment Score (NES) value and FDR (q‐value) are shown in each plot.
These results were further validated by performing bulk RNA‐seq analysis of control and Cdc14‐deficient ESCs during the differentiation process. Wild‐type ESCs typically expressed pluripotency markers such as Klf5, Oct4 (Pou5f1), Sox2, or Lifr that decreased during neural differentiation in parallel to the increase in transcripts involved in early differentiation to progenitors or neural cells (e.g. Gbx2, Sox11; Fig 2D). Cdc14‐deficient cells unexpectedly displayed higher levels of pluripotency markers at day 0, whereas the induction of neural differentiation markers was decreased. Pathway analysis confirmed a significant defect in general or neural differentiation, accompanied by a p53‐response in double‐mutant cells (Fig 2E and F; Dataset EV1).
CDC14 is required for the de‐repression of bivalent promoters during differentiation
To investigate the molecular mechanisms by which CDC14 may be controlling cell differentiation, we first performed mass spectrometry analysis of control and Cdc14 DKO ESCs at days 0 and 3 of the neural differentiation protocol (Fig 3A; Dataset EV2). Analysis of total protein levels detected a significant upregulation of the neural differentiation program in wild‐type cells in parallel to downregulation of proteins involved in the translation machinery after 3 days in the differentiation medium (Figs 3B and EV2A; Dataset EV3). The downregulation of the translation machinery was also present in Cdc14 DKO cells during this period, whereas the induction of the neural differentiation transcriptional program was less evident (Fig EV2A and B; Dataset EV3). In fact, Cdc14 DKO cells at day 3 display a significant downregulation of proteins typically induced during neural differentiation in control cells (Fig 3C and D; Dataset EV4). Concomitantly, Cdc14 DKO cells at day 3 displayed upregulation of pluripotency proteins (Fig EV2C), as well as proteins involved in the cellular response to LIF, a cytokine that inhibits cell differentiation in ESCs and the most common factor used to maintain the pluripotency state of ESCs in culture. Enrichment analysis of promoter regions pointed to the pluripotency factors NANOG and OCT4 (POU5F1) as the main common transcription factor that occupies the promoter of genes encoding upregulated proteins (Figs 3E and EV2D; Dataset EV5 and EV6), suggesting that stem cell proteins are not properly downregulated or are even upregulated (e.g. ZSCAN4C; Fig 3B) in Cdc14 DKO cells upon differentiation.
Figure 3. Proteomic analysis of neural differentiation in CDC14 DKO cells.
- Cdc14 DKO and control (two clones each) ESCs were subjected to neural differentiation and total protein was extracted at day 0 (T0) and after 3 days (T3) in differentiation. The plot shows the results from mass spectrometry analysis of total protein at day 0 (x axis) and after 3 days (y axis) in differentiation media. Colored dots are classified as indicated in the labels.
- Comparison of total proteins detected at T3 vs. T0 in wild‐type or CDC14‐deficient cells, showing examples of differentiation proteins among proteins less induced in mutant cells (purple dots) or pluripotency markers among the proteins more induced in mutant cells (red). Gray dots represent the rest of detected proteins that do not fall in any of the highlighted categories.
- Summary of main pathways identified after GSEA analysis in the indicated samples.
- Representative GSEA pathways deregulated in Cdc14 DKO cells after 3 days in the differentiation protocol. NES, Normalized enrichment score.
- Enrichment analysis of transcription factors that interact with the promoter of genes encoding proteins up‐ or down‐regulated in Cdc14 DKO cells before (day 0) or after 3 days on differentiation medium.
- GSEA analysis of proteins regulated by bivalent promoters in Cdc14 DKO cells after 3 days in the differentiation protocol.
Data information: In (D, F), NES, Normalized enrichment score; FDR, false discovery rate.
Figure EV2. Supplementary information to total protein analysis.
- Gene Set Enrichment Analysis (GSEA) of main pathways deregulated in wild‐type or mutant cells at day 3 (T3) vs. day 0 (T0) of the neural differentiation protocol. Data from total protein analysis.
- Representative GSEA analysis of downregulated proteins in Cdc14 DKO vs. WT cells after 3 days in differentiation media, showing significant downregulation of proteins related to neural differentiation.
- Immunoblot of the indicated proteins in lysates extracted in RIPA (whole lysates) or high‐salt RIPA buffer (chromatin‐bound proteins; histone 3 was used as a control for chromatin‐bound proteins). Vimentin and N‐cadherin are induced during differentiation in control cells, whereas the levels of pluripotency proteins decrease. These changes are altered in Cdc14 DKO cells.
- Enrichment in transcription factors in the promoters of genes encoding proteins upregulated or downregulated in wild‐type cells when comparing day 3 vs. day 0 of differentiation.
- GSEA of transcripts controlled by bivalent promoters in the indicated samples. Data obtained by obtained by bulk RNA‐seq analysis.
Data information: In (B, E), NES, normalized enrichment score; FDR, false discovery rate.
Source data are available online for this figure.
Transcription factor enrichment analysis also suggested that genes encoding proteins downregulated in Cdc14 DKO cells at day 3 are controlled by the Neuron‐Restrictive Silencer Factor (NRSF, also known as REST), a factor expressed highly in stem cells and an established global repressor of terminal neuronal genes, as well as by the Polycomb components SUZ12 and EZH2. REST actually mediates the recruitment of Polycomb repressor complexes in mammalian cells (Dietrich et al, 2012), and both REST and Polycomb control bivalent promoters during neural differentiation (McGann et al, 2014). In agreement with these observations, GSEA analysis of proteins downregulated in Cdc14 DKO cells at T3 showed a significant enrichment in genes with MLL2‐dependent bivalent promoters (Fig 3F). These results were also confirmed by GSEA analysis of bulk RNAseq data (Fig EV2E), suggesting that bivalent promoters are not properly de‐repressed during neural differentiation from ESCs in the absence of CDC14 phosphatases.
CDC14 dephosphorylates the epigenetic regulator UTF1
We next performed mass spectrometry analysis of phosphoproteins to identify possible direct effects of CDC14 deficiency. In total, 503 and 699 phospho‐sites were enriched in Cdc14 DKO cells at day 0 and day 3, respectively (Fig 4A; Dataset EV7). Eighty‐seven proteins contained upregulated phospho‐sites both at day 0 and day 3, including molecules involved in cellular damage, DNA repair and p53‐dependent responses, actin cytoskeleton, chromatin organization, and epigenetic regulation of gene expression (Fig EV3A; Dataset EV8). These proteins were enriched in target sites for CDK1/2, checkpoint kinases, and MAP kinases (Dataset EV9).
Figure 4. Altered phosphorylation of epigenetic regulators in Cdc14 DKO cells.
- Cdc14 DKO and control (two clones each) ESCs were subjected to neural differentiation and phospho‐residues were analyzed at day 0 (T0) and after 3 days (T3) in differentiation media by mass spectrometry. The Volcano plots show differentially regulated phospopeptides with residues hyper‐phosphorylated in Cdc14 DKO cells shown in orange. All other peptides are in gray.
- List of proteins with the largest number of hyper‐phosphorylated [S/T]P sites in Cdc14 DKO cells at day 3.
- Predicted structure (Jumper et al, 2021) of UTF1 showing the position of the residues hyper‐phosphorylated in Cdc14 DKO cells at day 0 (left) and day 3 (right) with the corresponding log(fold change) and P‐values when compared with control cultures.
- Western blot analysis of CDC14B‐GFP co‐immunoprecipitations. HeLa cells were transfected with UTF1 and GFP or CDC14B‐GFP. Cytoplasmic and nuclear soluble proteins were extracted with RIPA, and nucleolar and chromatin‐bound proteins were extracted with high salt RIPA. Immunoprecipitations with anti‐GFP antibody were performed. Total protein extracts and immunoprecipitates were loaded and UTF (upper panel) and GFP (lower panel) were blotted.
- Immunoprecipitation of UTF1 upon co‐transfection with CDC14B. Total protein was extracted with high‐salt RIPA and immunoprecipitations with anti‐UTF1 antibody were performed.
- Expression of active CDC14B, but not a phosphatase‐dead (PD) isoform, reduces slow‐mobility bands of UTF1. Vinculin was used as a loading control.
- Quantification of UTF1 phopho‐peptides upon UTF1 immunoprecipitation in cells co‐transfected with GFP or CDC14B‐GFP. Bars indicate the average from two technical replicates (dots).
- In vitro kinase assay of recombinant human UTF1 with ERK2, CDK1‐Cyclin B1, CDK1‐Cyclin A2 or in absence of kinases (C).
- Quantification of recombinant human UTF1 phospho‐peptides by mass spectrometry analysis after treatment in vitro with CDK1‐Cyclin A2, ERK2 or in absence of kinases (C). Bars indicate the average from two technical replicates (dots).
- In vitro phosphatase assay of UTF1. Three different fragments of UTF1 (#1, #3 and #4, see Fig EV5C) were phosphorylated by CDK1‐Cyclin A2 and subsequently exposed to wild‐type or a catalytically inactive form (PD) of CDC14B.
Figure EV3. Phospho‐proteins enriched in Cdc14 DKO cells.
- Pathway analysis of proteins with phospho‐sites significantly enriched both at day 0 and 3 during the differentiation protocol. The inset shows the P‐value of the represented pathways.
- Schematic representation of the UTF domains and amino acids of interest. Ser or Thr sites indicate phospho‐residues enriched in Cdc14 DKO cells. The three residues in red were found downregulated upon overexpression of CDC14B. Lys (K) sites indicate putative sites for ubiquitination. The fragments (Peptide #1–#5) used in this work are also shown with colored bars.
- Amino acid sequence of mouse UTF1 with the S/T residues deregulated in Cdc14 DKO cells.
- Predicted structure of mouse UTF1 showing the position of the residues hyper‐phosphorylated in Cdc14 DKO cells as well as SIAH and SPOP recognition domains.
- Expression of active CDC14B, but not a phosphatase‐dead (PD) isoform, reduces slow‐mobility bands of UTF1 in Phos‐Tag gels. Vinculin was used as a loading control.
- In vitro kinase assay of recombinant human UTF1 or the C‐terminal domain (CTD) of RNA polymerase II with ERK2, or CDK1‐Cyclin A2.
- In vitro phosphatase assay of UTF1. Five different fragments of UTF1 (#1 to #5; see panels B, C) were treated with active CDK1‐CyclinA2 complexes and subsequently exposed to CDC14B phosphatase activity.
CDC14 substrate sites include Ser or Thr residues followed by a Pro located within consensus sequences recognized by CDKs, with some preference for S‐P sites (Gray et al, 2003; Bremmer et al, 2012). Among the 34 proteins with upregulated phospho‐[S/T]‐P sites at day 0 and 3 (Dataset EV10), at least four of them (UTF1, DNMT3L, TET1, and TET2) are directly related to the epigenetic control of gene expression. We focused our attention on undifferentiated transcription factor 1 (UTF1), a protein expressed in ESCs that participates in the repression of differentiation genes (Laskowski & Knoepfler, 2012; Raina et al, 2021). Despite its small size, UTF1 presented the highest number of enriched proline‐directed phospho‐S/T residues in Cdc14 DKO cells, 11 of which were SP sequences (Figs 4B and C and EV3B–D). To analyze their possible interaction, we co‐transfected HeLa cells with UTF1 and CDC14B‐GFP or GFP alone and performed immunoprecipitation with anti‐GFP antibodies in two different cell fractions: RIPA‐soluble fraction (cytoplasmic and nuclear‐soluble proteins) and high‐salt RIPA fraction (nucleolar and chromatin‐bound proteins). We detected UTF1 in CDC14B‐GFP co‐immunoprecipitates in the high‐salt RIPA fraction, but not in control GFP immunoprecipitates (Fig 4D), suggesting a specific interaction between CDC14B and UTF1 in the chromatin‐bound protein fraction. Interestingly, the relative mobility of UTF1 changed when co‐transfected with CDC14B‐GFP but not with GFP alone (Fig 4D and E), and this effect was reversed when using a CDC14B phosphatase‐dead mutant (Fig 4F and EV3E), suggesting that UTF1 might be de‐phosphorylated in a CDC14B‐dependent manner.
To investigate whether these changes in the relative mobility of UTF1 correspond with changes in the phosphorylation status of the protein, we immunoprecipitated UTF1 after co‐transfection with GFP or CDC14B‐GFP and subjected it to mass spectrometry analysis. This analysis detected decreased phosphorylation of UTF1 in at least four residues (Fig 4E and G), three of which (S15, S48, and S54) were also found upregulated in Cdc14 DKO cells (Fig 4C). To directly confirm that UTF1 can be directly dephosphorylated by CDC14 phosphatases, we first performed phosphorylation assays by proline‐directed S/T kinases. Human UTF1 was phosphorylated in vitro by the MAP kinase ERK2, as well as the cell‐cycle kinase CDK1‐Cyclin A2, but not by CDK1‐Cyclin B1 (Fig 4H and EV3F). The change in relative mobility induced by CDK1‐Cyclin A2 was bigger than that induced by ERK2, suggesting that CDK1 might be phosphorylating more residues. The phosphorylation of UTF1 by ERK2 and CDK1‐CycA2 was confirmed by mass spectrometry analysis, which allowed the identification of several sites of phosphorylation (Fig 4I). Finally, coupled in vitro kinase and phosphatase assays suggested that CDC14B phosphatases, but not a catalytically dead isoform, can directly dephosphorylate UTF1 at multiple sites, including at least a region containing several sites (S48 and S54) in the N‐terminal domain (fragment #3; Fig 4J) and an additional fragment in the C‐terminal domain (fragment #5; Fig EV3G).
CDC14 controls UTF1 stability during differentiation of ESCs
CDC14 phosphatases are transcriptionally induced during neural differentiation from ESCs (Fig 5A). We also observed that CDC14B is released from the nucleolus during neural differentiation (Fig 5B), suggesting functional activation of this phosphatase as previously reported in yeast and mammalian cells (Stegmeier & Amon, 2004; Bassermann et al, 2008). UTF1, on the other hand, is a chromatin‐binding protein preferentially expressed in ESCs. The total levels of UTF1 decreased during neural differentiation in control cells, but this downregulation was deficient in Cdc14 DKO cells as detected in the transcriptomic (Appendix Fig S3a), real‐time RT‐PCR (Fig 5A) and proteomic analysis (Appendix Fig S3b). These changes were validated by direct immunodetection of UTF1 in ESCs before (T0) and either 1 day (T1; Fig 5C) or 3 days (T3; Fig 5D and Appendix Fig S3c) after the differentiation stimuli. The fact that the reduction in UTF1 protein (Fig 5C), but not Utf1 mRNA levels (Fig 5A), was already evident after only 1 day of differentiation suggests independent but complementary mechanisms for reducing UTF1 levels during neural differentiation.
Figure 5. Control of UTF1 protein levels by phosphorylation.
- Quantitative RT‐PCR in wild‐type (WT) or Cdc14 DKO ESCs of the indicated transcripts during neural differentiation. n = 3 independent experiments (± S.E.M.); *P < 0.05; **P < 0.01; ***P < 0.001; Student's t‐test.
- Expression of a CDC14‐GFP fusion protein in ESCs before (T0) and after 3 days (T3) in differentiation conditions. The arrowheads indicate the concentration of CDC14‐GFP in the nucleolus before but not after the induction of differentiation. Scale bars, 5 μM. The plot shows the ratio between nucleolar vs. total nuclear URF1 intensity at T0 (n = 26 cells) or T3 (n = 31) in cells from two separate experiments. ***P < 0.001, Student's t‐test with Welch's correction.
- Confocal microscopy analysis of UTF1 intensity in WT and Cdc14 DKO ESCs at time 0 (T0) and after 1 day (T1) in differentiation media. The quantification of UTF1 fluorescence mean intensity (FMI) per cell is shown in the plot to the right. Figures and horizontal bars indicate mean ± SEM (n = 1,269 WT T0 cells, n = 2,726 WT T1, n = 974 Cdc14 DKO T0, n = 1,173 Cdc14 DKO T1). **P < 0.01; ****P < 0.0001; Student's t‐test with Welch's correction.
- Western blot analysis of UTF1 expression in WT and Cdc14 DKO ESCs at time 0 and after 3 days in differentiation media.
- Western blot analysis of UTF1 expression in WT and Cdc14 DKO ESCs in absence or presence of the kinase inhibitors roscovitine (R) or trametinib (T).
- Immunoblot analysis of UTF1 protein stability after transfection of HeLa cells with UTF1 and GFP or CDC14B‐GFP. Cells were treated with cycloheximide (CHX), and total proteins were detected at different time points as indicated. Cyclin B1 and β‐catenin are used as controls.
- Western blot analysis of UTF1 WT or mutants in three (UTF1‐3KR) or five (UTF1‐5KR) lysines in the absence or presence of trametinib (T) or roscovitine (R).
- Immunoblot of UTF1 mutants in SPOP, SIAH, or SPOP + SIAH potential binding sites, as well as S48D and S54D phosphorylation sites. Actin was used as a loading control.
Given that UTF1 was phosphorylated in vitro by ERK2 and CDK1‐Cyclin A2 (Fig 4H), we treated ESCs with the ERK pathway inhibitor trametinib and the CDK inhibitor roscovitine. Inhibition of UTF1‐upstream kinases resulted in decreased levels of endogenous UTF1, suggesting a stabilizing effect of [S/T]P phosphorylations on this protein (Fig 5E). Similarly, CDC14B overexpression led to decreased UTF1 stability in cycloheximide‐treated cells, and this effect was rescued with the addition of the proteasome inhibitor MG132 (Fig 5F). The decrease in UTF1 levels could be partially rescued by the mutation of three or five lysines of the protein (Figs 5G and EV3B), suggesting that this decrease is due to UTF1 degradation, probably dependent on ubiquitination on these lysines. We found two potential binding sites for the E3 ubiquitin ligase adaptor SPOP and one potential binding site for the E3 ubiquitin ligase SIAH (Fig EV3C and D). Mutation of SIAH binding site increased the stability of the protein (Fig 5H), and simultaneous mutation of SPOP and SIAH sites further increased stability, suggesting that both proteins can be involved in UTF1 degradation. In addition, the stability of the protein is increased in a phosphomimic mutant of serines 48 and 54 (Fig 5H), which flank SIAH binding site and were found to be de‐phosphorylated by CDC14B in mass spectrometry analysis (Fig 4G) and are contained in the UTF1 fragment de‐phosphorylated in vitro (Fig 4J). All these data suggest that UTF1 is downregulated during differentiation from ESCs both at the transcriptional level and through CDC14‐mediated dephosphorylation and subsequent proteasome‐dependent degradation.
CDC14 modulates epigenetic and transcriptomic changes during exit from stemness in ESCs
We next used a multiomics approach by combining single‐cell (sc)RNA‐seq and scATAC (Assay for Transposase‐Accessible Chromatin using sequencing)‐seq in the same cell before (T0) and 5 days (T5) on the differentiation protocol (Figs 6A and EV4A). In agreement with the bulk RNA‐seq and proteomics data, lack of CDC14 correlated with deficient neural differentiation and increased presence of pluripotency markers (Figs 6B and EV4B). Clustering using the leiden algorithm (Fig 6C and EV4C and D) and pseudotime analysis suggested that control ESCs cells progressed during the differentiation protocol through a variety of intermediate stages to generate neural progenitors and mature neurons (Figs 6D and E and EV4E–G). Cdc14 DKO ESCs, however, were quite inefficient in generating these populations. Instead, these mutant cells underwent a different path to generate new populations (leiden groups 4, 6, 7, 10) that were almost nonexistent in control cultures subjected to the same differentiation protocol (Figs 6C and EV4E–G). Marker analysis suggested that these CDC14‐mutant groups were enriched in transcripts typically found in earlier stages of embryo development and fetal germ cells (cluster 6), as well as markers of the two‐cell stage such as Zscan4c or Zscan4d (Figs 6F and EV4H). Overall, Cdc14 DKO cells displayed higher expression of a 282‐gene signature (Biase et al, 2014) specific of the embryo two‐cell stage (Fig 6G), which was specifically intense in a particular small cluster (leiden 10) with high pseudotime values (Fig 6F and G), suggesting extensive changes from the cells of origin.
Figure 6. Single‐cell multiomic analysis of neural differentiation in CDC14‐deficient cells.
- Uniform Manifold Approximation and Projection (UMAP) of single‐cell RNA‐seq and ATAC‐seq data in the indicated samples.
- Dot plot analysis of the expression of representative markers for the samples and time points indicated.
- Cell clusters (0–10) generated using the leiden algorithm and represented using the scRNA‐seq or scATAC‐seq UMAP representations. The bar plot on the right shows the relative ratio of wild‐type or Cdc14 DKO cells in each of the clusters.
- ForceAtlas2 (FA) representation of samples (left), leiden clusters (middle), and diffusion pseudotime (right) values.
- scRNA‐seq and scATAC‐seq UMAP representation of pseudotime diffusion values for each cell. A random cell from leiden group 2 (wild‐type cells at day 0) was used as root and pseudotime values were calculated using scRNA‐seq data.
- Representative cell‐type specific markers in each of the leiden clusters, indicating the two major trajectories corresponding to most wild‐type (blue shadow) or mutant (red) cells.
- Expression score for the 282‐transcript signature typical of the two‐cell stage (left) and expression of the two‐cell stage‐specific marker Zscan4c (arrow in the right panel) using the UMAP representation for scRNA‐seq (top) and scATAC‐seq (bottom) data.
Figure EV4. Cluster and pseudotime analysis in single‐cell RNA‐seq data.
- Uniform Manifold Approximation and Projection (UMAP) of single‐cell RNA‐seq and ATAC‐seq data of wild‐type and Cdc14 DKO cells.
- Representative Gene Set Enrichment Analysis (GSEA) analysis of reactome pathways significantly downregulated at day 5 in Cdc14 DKO cells. The normalized enrichment score (NES) and FDR (q‐value) is indicated.
- UMAP representation of of single‐cell RNA‐seq (top) and ATAC‐seq (bottom) data of cell clusters generated using the leiden algorithm.
- Markers with the highest score and lowest normalized P‐value for each of the 11 leiden clusters.
- Diffusion component (DC) representation of the samples and time points, leiden clusters and pseudotime values.
- Separate plots for wild‐type and Cdc14 DKO cells in the ForceAtlas2 (FA) representation.
- Projection plots generated using partition‐based graph abstraction (PAGA) showing leiden clusters (right) and diffusion pseudotime values (left).
- Violin plots showing the expression levels of the indicated transcripts in the different leiden clusters.
To understand the connection between expression and epigenetic changes, we compared the RNA‐seq and ATAC‐seq data at the single‐cell level. This analysis suggested open chromatin in neural development genes upregulated in control, but not CDC14 DKO, cells during the differentiation process (Fig 7A). On the other hand, pluripotency (Nanog; Fig 7B) and two‐cell stage markers, such as Zscan4b, Zscan4c, or Zscan4d (Appendix Fig S4), which were repressed with closed chromatin in wild‐type cultures, displayed open chromatin and enhanced transcript levels in specific Cdc14 DKO cell clusters. Pseudotime analysis of RNA‐seq and ATAC‐seq reads confirmed defective opening of chromatin in neural differentiation genes such as Sox11 and Gbx2 and enhanced chromatin availability in genes related to naïve plutipotency or the two‐cell stage in CDC14‐deficient cells (Fig 7C).
Figure 7. Combined RNA‐seq and ATC‐seq analysis of neural differentiation and pluripotency genes at the single cell level.
- Analysis of the local links between mRNA expression and ATAC peaks in Sox11. The scRNA‐seq (top) and scATAC‐seq (bottom) UMAP representation of Sox11 transcripts is shown to the left. Middle panels show violin plots of the expression of Sox11 in the indicated samples and leiden clusters. The right panels show the local links between the expression of this transcript and ATAC peaks its gene in the different samples and data points, as well as leiden clusters.
- Analysis of Nanog mRNA expression and ATAC‐seq peaks as in (A).
- Plot of mRNA and ATAC reads versus diffusion pseudotime values for the indicated genes in wild‐type (blue) and Cdc14 DKO (red) cells.
Deregulation of bivalent genes and UTF1 targets in CDC14‐deficient cells
UTF1 modulates stemness and differentiation at least partially by regulating the expression of developmental genes controlled by bivalent promoters and Polycomb (PRC2) complexes. However, both overexpression and downregulation of UTF1 in ESCs prevent their differentiation potential (van den Boom et al, 2007; Kooistra et al, 2010). Ablation of UTF1 in control ESCs prevents the proper induction of neural differentiation (Fig EV5A), thus making it difficult to test the rescue effect of lowering UTF1 levels in CDC14‐deficient cells. We therefore interrogated scRNA‐seq and scATAC‐seq data to understand the effect of CDC14 inactivation in bivalent promoters as well as UTF1 targets. Utf1 mRNA was enriched in Cdc14 DKO cells perhaps as a consequence of the upregulation of pluripotency transcription factors that drive Utf1 expression (Nishimoto et al, 1999) in these samples (Fig 8A and B). Following the opposite pattern, neural differentiation was accompanied by the induction of bivalent genes in control cells, and this effect was severely impaired in Cdc14 DKO cells (Fig 8C and D). Similarly, whereas UTF1 targets are typically de‐repressed during the differentiation process (day 5 in wild‐type cells and leiden groups 5, 0, and 9; Fig 8E), UTF1 targets were deregulated in Cdc14 DKO cells, following an inverse pattern to UTF1 transcripts (Fig 8B) and protein levels (Fig 5A; Appendix Fig S3). These targets displayed increased levels at day 0 and defective de‐repression at day 5, a pattern that correlates with the dual effect of UTF1 in repressing its targets while limiting the excessive PRC2‐dependent repression in ESCs (Laskowski & Knoepfler, 2012). Many UTF1 targets specifically induced in differentiating wild‐type cells in different populations (leiden clusters representing neural progenitors and inmature and mature neurons in Fig 8F), were not properly induced in the absence of CDC14. This deregulation was also evident at the chromatin level, as induction of UTF1 targets such as Meis2 or Igf1r (two markers of neuron and glial progenitors), as well as Cdkn1a (a CDK inhibitor that correlates with terminal differentiation) was accompanied by opening of local chromatin in leiden clusters 5, 0, or 9, but not in clusters 7, 6, or 10, representing Cdc14 DKO cells at day 5 during the differentiation process (Fig 8F and G and EV5B and C). All together, these data suggest increased UTF1 levels accompanied by defective de‐repression of bivalent and UTF1 targets, during differentiation of CDC14‐deficient ESCs.
Figure EV5. Analysis of UTF1 and its target and bivalent transcripts during differentiation.
- Utf1 knockout using specific sgRNAs (sgUtf1) prevents neural differentiation in control ESCs. The panel on the left shows the levels of UTF1 protein in sgUtf1 or control sgControl ESCs. A representative image of ESCs colonies lacking UTF1 expression is shown in the middle panel. Scale bars, 50 μm. Plots to the right show the levels of the indicated differentiation markers in control (sgC) or sgUtf1 ESCs.
- Violin plot showing the expression of Igf1r and Cdkn1a in the different samples and data points or leiden clusters.
- Local links showing the correlation between the expression of the neural differentiation marker Igf1r or the cell cycle inhibitor Cdkn1a and their ATAC peaks in the different samples and data points, as well as in leiden clusters.
Figure 8. Deregulation of UTF1‐regulated transcripts and bivalent promoters in CDC14‐deficient cells.
- Uniform Manifold Approximation and Projection (UMAP) of single‐cell RNA‐seq (left) and ATAC‐seq (right) data showing the expression of Utf1.
- Violin plot showing the expression of Utf1 in the different samples and data points or leiden clusters.
- scRNA‐seq and scATAC‐seq UMAP representations showing the expression score of bivalent genes.
- Violin plot showing the expression score of bivalent genes in the different samples and data points or leiden clusters.
- Violin plot showing the expression score of a list of 696 bona‐fide UTF1 targets (Jia et al, 2012) in the different samples and data points, or leiden clusters.
- Dot‐plot of the expression of selected UTF1 targets showing specific induction in leiden clusters 5 and 0 (neural and glial progenitors) and/or cluster 9 (immature and mature neurons).
- Local links showing the correlation between the expression of the neural differentiation marker Meis2 and ATAC peaks in the different samples and data points, or in the leiden clusters. The plots to the right show the expression levels of Meis2 in the different samples and leiden clusters.
Discussion
Tissue differentiation and maintenance are regulated by the coordination between differentiation and proliferation of specific stem cells or progenitors. However, the molecular machinery connecting cell‐cycle progression and differentiation remains mostly unknown. It has been described that the duration of the G1 phase is tightly associated with the pluripotency/differentiation status of the cells, with high CDK activity and short G1 phase being a general feature of uncommitted cell populations during embryogenesis (Dalton, 2015; Liu et al, 2019). Accordingly, during neurogenesis of the ventricular zone, the length of G1 increases from 3 to 13 h, and shortening the length of G1 by ectopic elevation of CDK activity delays neural commitment (Lange et al, 2009). Conversely, if CDK activity is reduced and G1 phase extended, neural differentiation is accelerated, suggesting that in neurogenesis the length of G1 controls the balance between self‐renewal and differentiation. It has been suggested that cells are only susceptible to differentiation signals in G1, and that during this phase there is a transcriptional leakiness of developmental genes that is absent in the rest of cell cycle phases (Singh et al, 2013). Therefore, an extension of the G1 duration would be required to allow a sufficient magnitude and duration of the differentiation transcriptional activation to initiate a sustained developmental program. Recent studies provide mechanistic insights into these observations by suggesting critical roles for CDKs in maintaining pluripotency (Neganova et al, 2014; Liu et al, 2017, 2019). In particular, the mitotic kinase CDK1 directly controls the epigenetic landscape of pluripotent cells through the direct phosphorylating of multiple epigenetic regulators, including MLL2, SETDB1, SUV39H2, EZH2, DOT1L, KDM1A, PHF8, and JMJD1C (Michowski et al, 2020). CDK1 kinase inhibition leads to the coordinated expression of differentiation genes, suggesting that CDK‐dependent phosphorylation is key to maintain pluripotency and prevent differentiation.
In yeast cells, CDK phosphorylation during mitotic exit is mostly reversed by CDC14 and PP2A phosphatases (Stegmeier & Amon, 2004; Touati et al, 2019). Whereas PP2A complexes are also critical for mitotic exit in mammals (Wurzenberger & Gerlich, 2011), the relevance of CDC14 phosphatases in higher eukaryotes remains obscure. Previous studies in vertebrate cell lines with genetic ablation of Cdc14a or Cdc14b suggested some overlapping roles for these two paralogs in DNA repair and ciliogenesis (Mocciaro & Schiebel, 2010; Guillamot et al, 2011; Wei et al, 2011; Lin et al, 2015; Uddin et al, 2019; Partscht et al, 2021). In zebrafish, CDC14A and CDC14B are necessary for proper cilia elongation and/or maintenance, providing the first evidence of functional redundancy between these phosphatases (Clement et al, 2012). The analysis of mouse models with single mutations has also suggested some specific but molecular undefined functions in meiosis or aging (Imtiaz et al, 2018; Wen et al, 2020). Mutations of CDC14A in patients with autosomal‐recessive severe to profound deafness suggest additional roles in hearing (Delmaghani et al, 2016; Imtiaz et al, 2018).
By generating a first model with double ablation of both murine Cdc14 genes, Cdc14a and Cdc14b, we report here that CDC14 activity is not essential for mammalian development or postnatal life. In addition, cell cycle progression and cell proliferation display normal kinetics in the absence of these two phosphatases, at least in RPE1 cells (Partscht et al, 2021), as well as MEFs, murine neural progenitors, and ESCs (this work). Overall, all these data suggest that CDC14 phosphatases have evolved to counteract cell cycle kinases in more specialized cellular processes rather than during mitotic exit, a process in which both PP1 and PP2A play major roles both in fission yeast and mammals (Grallert et al, 2015; Holder et al, 2020). Cdc14 double‐mutant mice develop perinatal defects with altered dynamics of brain development; an observation in agreement with the fact that CDC14A and CDC14B transcripts are specifically enriched in brain and germ cells in humans (https://proteinatlas.org; Rosso et al, 2008). Phospho‐proteomics analysis of differentiating ESCs suggests a variety of epigenetic regulators with increased levels of phosphorylation in the absence of CDC14, including UTF1.
UTF1 is a molecule expressed in early stages of embryonic development (blastocyst and epiblast) and primordial germ cells (Okuda et al, 1998), with critical roles in maintaining the pluripotent state (Takahashi & Yamanaka, 2006; van den Boom et al, 2007; Zhao et al, 2008; Buganim et al, 2012; Laskowski & Knoepfler, 2012; Yang et al, 2014). UTF1 maintains a precise level of bivalent gene expression appropriate for maintaining pluripotency and for proper differentiation of pluripotent cells (van den Boom et al, 2007; Laskowski & Knoepfler, 2012). In silico analysis of downregulated proteins in Cdc14 DKO cells indicates that the main common factor that occupies the corresponding downregulated genes is the PRC2 subunit SUZ12 (Fig 3E). The proper balance of SUZ12 and PRC2 deposition in bivalent promoters is actually controlled by UTF1, and the majority of UTF1 target genes are both bivalent and targets of PRC2 (Jia et al, 2012). Mechanistically, UTF1 is a transient suppressor of lineage‐specific genes with a dual role in preventing excessive binding of the PRC2 complex to bivalent genes while simultaneously facilitating the tagging of mRNA from leaky repression for degradation (Jia et al, 2012). Differentiation requires induction of bivalent promoters, and UTF1 is rapidly repressed upon differentiation of human ESCs (Galonska et al, 2014).
Lack of CDC14 results in a significant increase in phosphorylation of UTF1 (Fig 4), increased protein stability (Fig 5), and defective induction of bivalent genes and UTF1 targets (Fig 7), suggesting that removal of UTF1 is mediated through CDC14‐dependent dephosphorylation and protein destabilization during differentiation of ESCs. Despite their critical roles, viable and fertile Utf1‐null (Nishimoto et al, 2013) or Cdc14a; Cdc14b double‐deficient (this work) mice can be obtained, at least in specific genetic backgrounds, suggesting that the detrimental effects of CDC14 or UTF1 deficiency in cell differentiation can be, at least partially, overcome in vivo. The dual role of UTF1 in bivalent promoters makes it difficult to clearly establish its specific function during differentiation. UTF1 deficiency is able to promote neuronal differentiation in the P19 embryonal carcinoma cell line (Lin et al, 2012). However, both overexpression (Thummer et al, 2010, 2012) and downregulation (van den Boom et al, 2007; Kooistra et al, 2010) of UTF1 interfere with both proliferation and proper differentiation of ESCs. This represents an important technical limitation to assess the exact contribution of UTF1 in Cdc14 DKO ESCs with rescue experiments, since UTF1 silencing hampers cell differentiation already in control pluripotent cells. Thus, to what extent UTF1 deregulation contributes to the phenotype of CDC14B‐deficient cells is unclear. However, overexpression of exogenous UTF1 is known to prevent proper differentiation of ESCs and embryonic carcinoma cells, suggesting a causal role for the defective protein degradation of UTF1 in CDC14‐null cells and defective differentiation. In addition, other epigenetic regulators, such as DNMTL, TET1, or TET2, are increasingly phosphorylated in CDC14‐deficient cells. It is therefore likely that, similarly to the control of stemness by CDK1 (Michowski et al, 2020), CDC14 phosphatases modulate differentiation of pluripotent cells through dephosphorylation of multiple epigenetic targets.
All these data consistently point to a model in which mammalian CDC14 phosphatases are dispensable for mitotic exit but participate in the exit from stemness during neural differentiation. Whether this effect is extensible to differentiation toward other cell lineages is still unknown, although our results suggest that CDC14 is also required for proper differentiation to other lineages such as adipocytes (C. Villarroya‐Beltri, A.F.B. Martins, M. Salazar‐Roa, and M. Malumbres, unpublished data). Previous studies in C. elegans suggested that the cdc‐14 phosphatase is required for the quiescent state of specific precursor cells through the hypophosphorylation and stabilization of the CDK inhibitor CKI‐1 (Roy et al, 2011). Our data suggest that CDC14 phosphatases promotes exit from stemness through the direct dephosphorylation of critical epigenetic regulators such as UTF1, leading to the induction of bivalent genes. These data, together with the new role of CDK1 as a master epigenetic regulator in ESCs (Liu et al, 2019; Michowski et al, 2020), establish a critical link between cell cycle kinase‐phosphatase modules and lineage commitment in pluripotent cells.
Materials and Methods
Genetically modified mouse models
Cdc14a‐deficient mice were generated using a targeting vector (Genebridges) in which exon 3 was flanked by loxP sites and a PGK‐neoR cassette, as described in Fig EV1. Mouse embryonic stem cells (ESCs) V6.4 obtained from a hybrid (129 × C57BL/6J) strain were electroporated with 100 μg of linearized DNA carrying a neomycin resistance gene and flanking exon 3 of the murine Cdc14a locus with loxP sequences. Recombinant ESCs and clones were selected in the presence of G418 (neomycin). Positive recombinant clones were microinjected into C57BL/6J blastocysts, and the resulting positive chimaeric mice Cdc14a(+/loxfrt) were crossed with C57BL6/J mice for transmission of the recombinant allele. Subsequently, the neo cassette was removed by mating with mice expressing the Flp recombinase (Tg.pCAG‐Flp), generating the Cdc14a(lox) conditional KO allele. Germline excision of exon 3 was achieved by additional crosses with CMV‐Cre transgenic mice, which express constitutive and ubiquitous DNA recombinase Cre. Excision of exon 3 of the Cdc14a gene leads to a Cdc14a(−) null allele with a frame shift in the mRNA, generating several new premature stop codons. The Cdc14b(−) allele was reported previously (Guillamot et al, 2011), and the Cdc14b(−/−); Cdc14b(−/−) double‐knockout (Cdc14 DKO) mouse model was obtained by additional crossing between the single KOs. Both sexes were used in this study, and animals were monitored from birth to 1‐year age. Animal procedures were approved by the corresponding Ethics Committee of Animal Experimentation in our institution and local government (Instituto Carlos III and Comunidad de Madrid). The CNIO has earned AAALAC International Accreditation Program since 2016.
Histological and immunohistochemical analysis
For histological analysis, embryos or tissue samples were fixed overnight in 10% buffered formalin (Sigma‐Aldrich). Samples were paraffin‐embedded and cut into 3–5 μm thickness sections that were mounted in superfrost®plus slides and dried overnight. Sections were then stained with hematoxylin and eosin (H&E), cresyl violet acetate (Nissl) or subjected to immunohistochemistry (IHC). For IHC, antigen retrieval was first performed; sections were incubated with 0.2% Triton X‐100 for 15 min at room temperature (RT) for permeabilization and blocked with 2% BSA during 2 h at RT followed by specific primary antibody incubation overnight (ON) at 4°C. The following primary antibodies were used: anti‐calbindin (Sigma, C9848), anti‐GFAP (Merck, AB5541), CTIP2 (Abcam, ab18465), Nestin (Thermo Fisher, MA1‐110), SATB2 (Abcam, ab51502), TBR1 (Abcam, ab31940), and UTF1 (Abcam, ab24273). Secondary antibody incubation was performed for 1 h at RT at a 1:500 dilution. Fluorescent‐conjugated secondary antibodies coupled to different Alexa dyes (488, 594, or 647) were purchased from Molecular Probes (Invitrogen). Nuclear staining was included in the last PBS wash, using DAPI, and finally, the slides were mounted with a permanent mounting medium (Fluormount‐GTM, Thermofisher) for microscopic evaluation.
Image acquisition for H&E and Nissl stainings was performed using a Leica D3000 microscope equipped with 20×/1.42 and 40× NA objective lens. As for the remaining stainings, image acquisition was performed using a Leica laser scanning confocal microscope TCS‐SP5 (AOBS) equipped with an oil immersion objective of 40× (HCX‐PLAPO 1.2 N.A.) and LASAF v2.6. software. Image analysis was performed using ImageJ software.
Cell culture
Mouse embryonic fibroblasts (MEFs; both from male and female embryos) were obtained and cultured using standard protocols. Briefly, at 13.5 days post‐coitum, embryos were extracted from the uterus of pregnant females, isolated from the yolk sac, and the placenta was removed. Embryos, without the liver and head, were finely minced and digested with 0.1% trypsin for 5 min at 37°C. Cultures were maintained in Dulbecco's modified Eagle's medium (DMEM; Lonza) supplemented with 0.1% gentamicin and 10% fetal bovine serum (FBS, Sigma‐Aldrich) at 37°C in a humidified 5% CO2 atmosphere. Cells were grown for two population doublings and then frozen. Once reaching confluence, MEFs were subcultured at a ratio of 1:3. Primary MEFs with more than six passages were never used to perform experiments. MEFs immortalization was achieved by infection of pMEFs with retroviruses expressing the first 121 amino acids of the SV40 T large antigen (T121) followed by hygromycin B (150 μg/ml) selection during 2 days and passaging for several weeks. Immortal MEFs were subcultured at a ratio of 1:6. For feeder cells generation, primary MEFs (passage 3) were treated with mitomycin C (10 μg/ml, Roche) for 2 h and washed three times with PBS. Cells were counted and seeded or frozen for later use.
Mouse Embryonic Stem Cells (ESCs) were cultured over feeder cells and in the presence of ESC medium containing KO‐DMEM (Gibco), 2‐Mercaptoethanol (Invitrogen), non‐essential aminoacids MEM NEAA (Invitrogen), GlutamMAX (Gibco), Penicillin/Streptomycin (5,000 μg/ml, Invitrogen), LIF (Leukemia Inhibitor Factor, ESGRO, Millipore), and 10% Fetal Calf Serum (Hyclone) at 37°C in a humidified 5% CO2 atmosphere. Cell medium was changed every 24 h, and ESCs were split at a ratio of 1:6. Overconfluent ESCs cultures, where colonies touch one another, were persistently avoided as this might lead to unwanted precocious differentiation and poorer stem potential of cells. Prior to sample extraction, ESCs were cultured over 0.1% gelatin‐coated plates or in suspension in non‐adherent Petri dishes in order to remove feeder cells. For generation of Cdc14a and Cdc14b simple and double KO ESCs, different clones of ESC from Cdc14a(lox/lox); Cdc14b(+/+) or Cdc14a(lox/lox); Cdc14b(−/−) were seeded over gelatin and infected with adenovirus expressing recombinase CRE or empty adenovirus; CRE cutting efficiency and generation of Cdc14a null alleles were verified. Adenoviruses expressing CMV‐Empty or CMV‐Cre were obtained from the University of Iowa (Iowa City, IA), and infection was carried out at a 250 MOI for two consecutive rounds 12 h apart in 6‐well plates coated with 0.1% gelatin.
For the analysis of neural progenitors, mouse cerebral cortices (P6) were dissected, gently dissociated manually, and neural progenitors were isolated from the subventricular zone as reported previously (Ferron et al, 2004). Single dissociated cortical cells were cultured as spheres under non‐adherent conditions, and experiments were performed in cultures with < 10 passages.
HEK293T (ATCC #CRL‐11268) and HeLa (ATCC #CRM‐CCL‐2) cells were maintained in DMEM medium (Sigma) supplemented with 10% FBS (Sigma‐Aldrich), 0.1% gentamicin and were grown at 37°C in a humidified 5% CO2 atmosphere. These clones were not further authenticated and were routinely tested for mycoplasma.
Cell transfection and cell cycle studies
Plasmid transfection of cells was performed using Lipofectamine 2000 (Invitrogen) in accordance with the manufacturer's instructions. To synchronize cells in G0, asynchronous MEFs cultures were grown in DMEM (Lonza) supplemented with 0.1% gentamicin and 10% fetal bovine serum (FBS; Sigma) until confluence, rinsed twice with PBS, and were then arrested by serum starvation with DMEM supplemented with 0.1% fetal bovine serum (FBS; Sigma) for 72 h. G0‐synchronized MEFs were then trypsinized and split in complete media containing 10% FBS for the indicated time points. For quantification of S‐phase entry, cells were incubated with EdU (10 μM, Sigma) for 1 h before harvesting, and DNA replication was analyzed by flow cytometry. For mitotic entry, cells were treated with nocodazole 20 h after serum addition, and samples were analyzed by immunoblotting. For live cell imaging analysis, nocodazole (0.8 μM) and taxol (1 μM) were added to the cells immediately before imaging. Cytokinesis and segregation defects were assessed and quantified by visual analysis of time‐lapse images of cells stably expressing a histone‐GFP fusion protein. Defects in chromosome segregation were characterized by asymmetric distribution of the GFP signal between the two daughter cells, and defects in cytokinesis were defined as abnormal delay in the separation between the two daughter cells or regression as an interphasic 4n cells.
For cell cycle analysis by flow cytometry, cells were trypsinized and fixed in cold 70% ethanol at 4°C for at least 12 h. For EdU staining, the classical azide‐based EdU Click‐iT protocol (Invitrogen) was followed plus an additional step for signal amplification using biotin‐azide and streptavidin‐Alexa fluor. For cell cycle profile analysis, samples were incubated in the presence of RNAse A (0.1 mg/ml; Qiagen) at 37°C for 30 min, and DNA was stained with propidium iodide (PI, 50 μg/ml; Sigma Aldrich) and kept ON at 4°C. The following day samples were analyzed with a FACSCanto TM flow cytometer (BD Bioscience), and 10,000 single events were acquired. FlowJo Version 9.6.4 software was used to analyze cell populations (TreeStar, Oregon).
For videomicroscopy, immortal MEFs were infected with lentiviruses containing mH2B‐GFP, and 2 days later, GFP‐positive cells were sorted in an Influx or FACS Aria TM sorter, to generate H2B‐GFP stable cell lines. Synchronous MEFs expressing stable H2B‐mGFP were plated on 8‐well glass‐bottom dishes (Ibidi) at a density of 30,000 cells per well and imaged with a DeltaVision RT imaging system (Olympus IX70/71, Applied Precision) equipped with a Plan Apochromatic 20×/1.42 NA objective lens and maintained at 37°C in a humidified CO2 chamber. Images were acquired every 7 or 10 min, in the absence or presence of drugs, respectively. Data processing and analysis were performed using ImageJ software.
Neural differentiation
For in vitro neural differentiation, we basically followed the protocol described previously (Gaspard et al, 2009). Briefly, ESCs were seeded in gelatin‐coated plates at low density (100,000 cells/p6 well). The next day (day 0), LIF‐containing media was replaced with neural differentiation media: Advanced DMEM‐F12 supplemented with 1× N2 (Gibco, 17502001), 1× B27 (Gibco, 17504044), 2 mM N‐acetylcysteine (Sigma, A9165), 0.6% glucose (Gibco, 15023021), 2 mM L‐glutamine (PAN‐Biotech, P04‐80100), 0.7 U/ml heparin (Sigma, H3149), 5 mM hepes (Gibco, 15630080), 1 mM sodium pyruvate (Gibco, 11360070), 20 ng/ml mEGF (R&D System, 2028‐EG‐200), 20 ng/ml hFGF2 (Miltenyi Biotec, 130‐093‐841). Media was changed every 2 days. After 6 days, cells were disaggregated with TrypLe and seeded at high density on plates or cover‐slips (1 p6 well to 1 p12 well or to 6 cover‐slips) previously coated with laminin (5 μg/ml, Biolamina LN511) in neural differentiation media without growth factors. Media was changed every 2 days, and neural differentiation efficiency was assessed at day 18 by confocal microscopy analysis of TUJ1 and GFAP or by qRT‐PCR with oligonucleotides for specific transcripts (Appendix Table S1).
Bulk RNA analysis
To quantify expression of transcripts, total RNA was isolated using the miRVANA RNA extraction kit (Ambion) according to the manufacturer's instructions. Reverse transcription from 1 μg of total RNA was performed with the M‐MLV retrotranscriptase (Promega) followed by quantitative PCR (qPCR) using SYBR Green PCR Master Mix (Applied Biosystems) in Applied Biosystems ViiA7 real‐time PCR machine. Data were analyzed using the comparative Ct method and are expressed as the ratio between the expression of each gene and the corresponding housekeeping gene, which was used for normalization, according to the specification in the figures. Oligonucleotides used for the amplification of specific genes are listed in Appendix Table S1.
For RNA‐seq analysis, ESCs in LIF‐containing media (T0) or after 2 days (T2) or 7 days (T7) in neural differentiation media were collected, and total RNA was extracted using miRVANA RNA extraction kit (Ambion). In total, 500 ng of total RNA was used. Average sample RNA Integrity Number was 9.8 (range 9.5–10) when assayed on an Agilent 2100 Bioanalyzer. Sequencing libraries were prepared with the “QuantSeq 3′ mRNA‐Seq Library Prep Kit (FWD) for Illumina” (Lexogen, 015) following manufacturer instructions. Library generation is initiated by reverse transcription with oligodT priming, and a second strand synthesis is performed from random primers by a DNA polymerase. Primers from both steps contain Illumina‐compatible sequences. Libraries were completed by PCR, applied to an Illumina flow cell for cluster generation and sequenced on Illumina HiSeq 2500 following manufacturer's protocols. The resulting reads were analyzed with the nextpresso pipeline (Graña et al, 2018). The Gencode vM20 gene annotation for GRCm38 was used. GSEA Preranked was used to perform gene set enrichment analysis for the selected gene signatures on a pre‐ranked gene list, setting 1,000 gene set permutations (Subramanian et al, 2005). Only those gene sets with significant enrichment levels (FDR q‐value < 0.25) were considered.
Immunofluorescence
Cells previously seeded in cover slips were fixed in 4% buffered paraformaldehyde (PFA) for 10 min at RT and incubated with 0.5% Triton X‐100 for 10 min at RT for permeabilization. Cells were then blocked with 2% BSA during 2 h at RT followed by primary antibody incubation ON at 4°C. Secondary antibody incubation was performed for 1 h at RT at a 1:500 dilution. The following antibodies were used: anti‐TUJ1 (Sigma, T8660), anti‐GFAP (Merck, AB5541), anti‐Nestin (Thermo Fisher, MA1‐110), anti‐UTF1 (Abcam, ab24273). Fluorescent‐conjugated secondary antibodies coupled to different Alexa dyes (488, 594, or 647) were purchased from Molecular Probes (Invitrogen). Nuclear staining was included in the last PBS wash, using DAPI, and finally, the slides were mounted with a permanent mounting medium (Fluormount‐GTM, Thermofisher) for microscopic evaluation.
Immunoblotting and immunoprecipitation
Cultured cells were harvested and lysed in RIPA (25 mM Tris–HCl pH 7.5, 150 mM NaCl, 1% NP‐40, 0.5% Na deoxycholate, Complete protease inhibitor cocktail) 15 min at 4°C and centrifuged 10 min at 14,000 g. Supernatant was saved for analysis of cytoplasmic and nuclear‐soluble proteins. When indicated, pellet containing chromatin‐bound and nucleolar proteins was resuspended in high‐salt RIPA (25 mM Tris–HCl pH 7.5, 500 mM NaCl, 1% NP‐40, 0.5% Na deoxycholate, complete protease inhibitor cocktail) incubated for 10 min at 4°C and centrifuged 10 min at 14,000 g. When required, total protein was extracted directly by incubating cells in high‐salt RIPA. Protein extracts were quantified by the BCA method, mixed with loading buffer (350 mM Tris–HCl pH 6.8, 30% glycerol, 10% SDS, 0.6 M DTT, 0.1% bromophenol blue) boiled for 5 min, and subjected to electrophoresis using the standard SDS–PAGE method. Proteins were then transferred to a nitrocellulose membrane (BioRad), blocked for 1 h at RT in TBS 0.1% Tween‐20 containing 5% BSA, and incubated overnight at 4°C with specific primary antibodies. Membranes were washed 10 min in TBS‐Tween and incubated with peroxidase‐conjugated secondary antibodies (Dako, 1:5,000) for 45 min at RT. Finally, the membranes were washed for 5 min three times and developed using enhanced chemiluminiscence reagent (Western Lightning Plus‐ECL; Perkin Elmer). The following primary antibodies were used: anti‐GFP (Roche, 11814460001), anti‐UTF1 (Abcam, ab24273), anti‐Actin (Sigma, A5441), anti‐Cyclin B1 (Abcam, ab72), anti‐AURKB (Abcam, ab2254), anti‐phospho‐CDK1 (Santa Cruz, sc‐7989), anti‐phospho‐histone H3 (Sigma, 07‐424), anti‐nucleolin (Abcam, ab22758).
For immunoprecipitation, protein G Dynabeads (Thermo fisher) were washed in PBS‐tween 0.01%, incubated with anti‐GFP or anti‐UTF antibodies for 1 h at RT, and equilibrated in RIPA buffer. RIPA or high‐salt RIPA protein extracts were incubated with pre‐washed and equilibrated Dynabeads 1 h at 4°C for pre‐clearing. Pre‐cleared protein extracts were incubated with antibody‐coupled Dynabeads for 2 h at 4°C. Dynabeads were washed five times with RIPA buffer, resuspended in loading buffer, and boiled for 5 min. Dynabeads were finally removed, and protein‐containing extracts were loaded for Western blot analysis.
Proteomic studies
ESCs in LIF‐containing media (t0) or after 3 days in neural differentiation media (t3) were collected. Cells were lysed during 10 min at 95°C in 5% SDS, 50 mM TEAB pH 7.55. After cooling, DNA was sheared by 10 min of sonication. Protein concentration was determined by micro BCA using BSA as standard. Then, 200 μg of each sample was digested by means of the Protifi™ S‐Trap™ Mini Spin Column Digestion Protocol. Samples were labeled using TMT® reagent 11‐plex following manufacturer's instructions. Samples were mixed in 1:1 ratios based on total peptide amount, and the final mixture was desalted using a Sep‐Pak C18 cartridge (Waters) and dried prior high pH reverse‐phase HPLC pre‐fractionation. Peptides were pre‐fractionated offline by means of high‐pH reverse‐phase chromatography using an Ultimate 3000 HPLC system equipped with a sample collector. Phosphopeptides were enriched using home‐made TiO2 micro‐columns. LC–MS/MS was done by coupling an UltiMate 3000 RSLCnano LC system to a Q Exactive HF‐X mass spectrometer (Thermo Fisher Scientific). The mass spectrometer was operated in a data‐dependent mode, with an automatic switch between MS and MS/MS scans using a top 12 method (Intensity threshold ≥ 9.3e4, dynamic exclusion of 20 s and excluding charges unsassigned, +1 and ≥ +6). Mass spectra were acquired from 350 to 1,500 m/z with a resolution of 60,000 (200 m/z). Ion peptides were isolated using a 1 Th window and fragmented using higher‐energy collisional dissociation (HCD) with a normalized collision energy of 35. MS/MS spectra were acquired with a fixed first mass of 110 m/z and a resolution of 45,000 (200 m/z). The ion target values were 3e6 for MS (maximum IT of 25 ms) and 1e5 for MS/MS (maximum IT of 86 ms). Raw files were processed with MaxQuant (v1.6.0.16) using the standard settings against a mouse protein database (UniProtKB/TrEMBL, 53,449 sequences) supplemented with contaminants. Results were filtered at 1% FDR (peptide and protein level). Afterward, the “Phospho (STY)Sites.txt” was loaded in ProStaR (Wieczorek et al, 2017) for the phosphosite analysis, and the “proteinGroups.txt” file was loaded for the total protein analysis. The reporter intensity values were used for further statistical analysis. Briefly, a global normalization of log2‐transformed intensities across samples was performed using the LOESS function. Differential analysis was done using the empirical Bayes statistics Limma. Sites or proteins with a P value < 0.05 and a log2 fold change ratio > 0.3 or < −0.3 were defined as deregulated. The FDR was estimated to be below 5%.
Mass spectrometry analysis of immunoprecipitated UTF1
HeLa cells were transfected with UTF1 and CDC14B‐GFP or GFP, and UTF1 was immunoprecipitated as described above. Proteins were eluted twice, using 8 M urea in 100 mM Tris–HCl pH 8.0, and digested by means of the standard FASP protocol. Briefly, proteins were reduced and alkylated (15 mM TCEP, 50 mM CAA, 30 min in the dark, RT) and sequentially digested with Lys‐C (Wako) (200 ng of Lys‐C per sample, o/n at RT) and trypsin (Promega) (200 ng of trypsin per sample, 6 h at 37°C). Resulting peptides were desalted using C18 stage‐tips, speed‐vac dried, and re‐dissolved in 21 μl of 0.5% formic acid. LC‐MS/MS was done by coupling an UltiMate 3000 RSLCnano LC system to a Q Exactive Plus mass spectrometer (Thermo Fisher Scientific). The mass spectrometer was operated in a data‐dependent mode, with an automatic switch between MS and MS/MS scans using a top 15 method (Intensity threshold ≥ 4.5e4, dynamic exclusion of 25 s and excluding charges unassigned, +1 and > +6). Mass spectra were acquired from 350 to 1,500 m/z with a resolution of 70,000 FWHM (200 m/z). Ion peptides were isolated using a 2.0 Th window and fragmented using higher‐energy collisional dissociation (HCD) with a normalized collision energy of 27. MS/MS spectra resolution was set to 35,000 (200 m/z). The ion target values were 3e6 for MS (maximum IT of 25 ms) and 1e5 for MS/MS (maximum IT of 110 ms). Raw files were processed with MaxQuant (v1.6.0.16) using the standard settings against a human protein database (UniProtKB/Swiss‐Prot, 20,373 sequences). Carbamidomethylation of cysteines was set as a fixed modification, whereas oxidation of methionines protein N‐term acetylation and phosphorylation of serines, threnonines, and tyrosines were set as variable modifications. Minimal peptide length was set to seven amino acids, and a maximum of two tryptic missed‐cleavages were allowed. Results were filtered at 0.01 FDR (peptide and protein level). Raw data were imported into Skyline. Label‐free quantification of identified phosphopeptides was performed using the extracted ion chromatogram of the isotopic distribution. Only peaks without interference were used for quantification. Normalization was performed using the intensity of identified UTF1 non‐modified peptides.
In vitro kinase and phosphatase assays
His‐tagged recombinant human UTF1 (My Biosource) or UTF1 fragments (GenScript) were attached to His‐Tag dynabeads (Thermo Fisher) for 30 min. UTF1‐attached dynabeads were washed in kinase buffer (50 mM HEPES pH 7.5, 10 mM MgCl2) and resuspended in kinase buffer with 500 ng of ERK2, CDK1‐cyclin B1, or CDK1‐cyclin A2 in the presence of 50 μM cold ATP and 0.15 μCi of γ(32P)ATP. Samples were incubated 30 min at 30°C. For in vitro de‐phosphorylation, previously γ(32P)‐ATP‐phosphorylated UTF1‐attached beads were washed in phosphatase buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 0.1% Triton X‐100) and resuspended in phosphatase buffer with recombinant PIN1 and CDC14B or phosphatase‐dead CDC14B. Samples were incubated 45 min at 30°C, and reactions were stopped by adding loading buffer and incubating 5 min at 95°C.
Mass spectrometry analysis of in vitro phosphorylated UTF1
UTF1 was phosphorylated in vitro as described above but with cold ATP. Proteins were reduced and alkylated by adding 100 μl of 15 mM TCEP, 50 mM CAA and incubating 1 h in the dark. On‐bead digestion was performed using trypsin (Promega) (200 ng of trypsin per sample, 4 h at 37°C). Resulting peptides were desalted using C18 stage‐tips, speed‐vac dried, and re‐dissolved in 21 μl of 0.5% formic acid. LC‐MS/MS was done by coupling an UltiMate 3000 RSLCnano LC system to Q Exactive HF mass spectrometer (Thermo Fisher Scientific). The mass spectrometer was operated in a data‐dependent mode, with an automatic switch between MS and MS/MS scans using a top 15 method (Intensity threshold ≥ 2.2e4, dynamic exclusion of 10 s and excluding charges +1 and > +6). Mass spectra were acquired from 350 to 1,500 m/z with a resolution of 60,000 FWHM (200 m/z). Ion peptides were isolated using a 2.0 Th window and fragmented using higher‐energy collisional dissociation (HCD) with a normalized collision energy of 27. MS/MS spectra resolution was set to 30,000 (200 m/z). The ion target values were 3e6 for MS (maximum IT of 25 ms) and 1e5 for MS/MS (maximum IT of 45 ms). Data analysis was performed as described for immunoprecipitated UTF1.
Analysis of protein stability
HeLa cells were transfected with UTF1 and CDC14B‐GFP or GFP. Forty‐eight hours later, cells were treated with 40 μg/ml cycloheximide B (Sigma, C7698‐1G) and when indicated with MG132 (40 μM). Cells were lysed at the indicated time points and analyzed by Western blot. For analysis of protein stability in the presence of kinase inhibitors, HeLa cells were transfected with UTF1 and GFP or CDC14B‐GFP. Eight hours later, different kinase inhibitors were added at the indicated concentrations (RO‐3306, 10 μM; Roscovitine, 50 μM) and incubated overnight. When indicated, cells were also treated with MG132 (20 μM).
For the generation of mutant cDNAs, pMXs‐UTF (#13369) was obtained from Addgene. UTF1 mutants were generated by sequential directed mutagenesis using QuikChange II Site‐Directed Mutagenesis kit (Agilent Technologies). SPOP binding sites PVTTS and PLTST were changed into PVNYA and PLRAA, respectively. SIAH binding site SPKKPVSPD was changed into SHQGSRAPD. In the phosphomimic mutant, UTF1 Ser48 and Ser54 were changed into aspartic acid. Primers used are shown in Appendix Table S1. Wild‐type and mutated UTF1 were subcloned in pLVX and pEGFP plasmids. pDEST 3.1 CDC14B‐GFP was generated previously (Guillamot et al, 2011). The UTF1 3KR and 5KR mutants were reported previously (Correa‐Vazquez et al, 2021).
Multiomic RNA‐seq and ATAC‐seq analysis in the same cell
ESCs in LIF‐containing media (T0) or after 5 days (T5) in neural differentiation media were washed twice with PBS, trypsinized, and counted. Dead cells were removed with Dead Cell Removal kit (Miltenyi Biotec). Cells were counted again, and 1 million cells were centrifuged at 300 g 10 min at 4°C in PBS 0.04% BSA. Cells were treated with 0.1 U/μl of DNAse I (Invitrogen) 5 min in ice. Cells were washed in PBS 0.04% BSA and centrifuged 10 min at 300 g twice. Cells were incubated 2 min with 100 μl Lysis Buffer (10 mM Tris–HCl pH 7.5, 10 mM NaCl, 3 mM MgCl2, 0.1% Tween‐20, 0.1% NP40, 0.01% Digitonin, 1% BSA, 1 mM DTT, 1 U/μl RNase inhibitor) in ice for nuclei extraction. After incubation, nuclei were washed in 1 ml of Wash Buffer (10 mM Tris–HCl pH 7.5, 10 mM NaCl, 3 mM MgCl2, 0.1% Tween‐20, 1% BSA, 1 mM DTT, 1 U/μl RNase inhibitor) and centrifuged 5 min at 500 g. Washing was repeated two more times, and nuclei were filtered in 40 μm Flowmi Cell Strainer, counted, and assessed for integrity. Cell nuclei were loaded onto a Chromium Next GEM Chip J (10× Genomics) for a target recovery of ~ 5,000 nuclei per condition and processed into a 10× Chromium Controller as described in the manufacturer's protocol (Chromium Next GEM Single Cell Multiome ATAC + Gene Expression, CG000338 Rev A). Generation of gel beads in emulsion (GEMs), barcoding, pre‐amplification PCR, and ATAC library construction were all performed as recommended by the manufacturer. Combined scRNA‐seq and scATAC‐seq data were analyzed using Loupe (v6, 10× Genomics) and Scanpy (Wolf et al, 2018). Pathway and gene set enrichment analysis was performed using GSEA (Subramanian et al, 2005) and Enrichr (Xie et al, 2021), as well as their python wrap (GSEApy; https://github.com/zqfang/GSEApy/) and Metascape (Zhou et al, 2019).
Statistics and data analysis
All statistical analyses were done using suggested analyses by Graphpad Prism 8 software, as indicated in the corresponsing figure legends. Mouse sample size was selected based on previous studies in our laboratory following the 3R reommendations. All mice were randomly selected from their predetermined conditions for experiments, and no blinding was done in this study.
Author contributions
Carolina Villarroya‐Beltri: Conceptualization; data curation; formal analysis; supervision; validation; investigation; visualization; methodology; writing—original draft; writing—review and editing. Ana Filipa B Martins: Formal analysis; investigation. Alejandro García: Investigation. Daniel Giménez: Software; investigation; visualization. Eduardo Zarzuela: Formal analysis; investigation. Mónica Novo: Investigation. Crsitina del Álamo: Investigation. Jose González‐Martínez: Formal analysis; investigation. Gloria C Bonel‐Pérez: Investigation. Irene Díaz: Investigation. María Guillamot: Formal analysis; investigation. Massimo Chiesa: Investigation. Ana Losada: Resources; formal analysis. Osvaldo Graña Castro: Formal analysis. Meritxell Rovira: Investigation. Javier Muñoz: Formal analysis; investigation; methodology. María Salazar‐Roa: Formal analysis; investigation. Marcos Malumbres: Conceptualization; resources; software; formal analysis; supervision; funding acquisition; investigation; visualization; methodology; writing—original draft; project administration; writing—review and editing.
Disclosure and competing interests statement
The authors declare that they have no conflict of interest.
Supporting information
Appendix
Expanded View Figures PDF
Dataset EV1
Dataset EV2
Dataset EV3
Dataset EV4
Dataset EV5
Dataset EV6
Dataset EV7
Dataset EV8
Dataset EV9
Dataset EV10
Source Data for Expanded View
PDF+
Acknowledgments
We thank the Microscopy, Cytometry, Comparative Pathology, Mouse Facility, and Mouse Genome editing core services of the CNIO for their support. We also thank Eva Porlan (UAM, Madrid) for help with neural cultures, and Dr. Mario García‐Domínguez (CABIMER, Seville) for the UTF1 3KR and 5KR mutants. CV received support from the Juan de la Cierva programme, Ministry of Science and Innovation‐Agencia Estatal de Investigación (MCI‐AEI). DG and MSR were supported by the Fundación Científica de la Asociación Española contra el Cáncer (AECC). AFBM and JGM received predoctoral contracts from Foundation La Caixa and the Ministry of Education of Spain (FPI grant BES‐2016‐077901). This work was supported by grants from the European Commission Seventh Framework Programme (ERA‐NET NEURON8‐Full‐815‐094), AEI‐MICIU/FEDER (RTI2018‐095582‐B‐I00 and RED2018‐102723‐T), and the Personalized Medicine and Nanotechnologies in Lung Cancer (iLUNG) and scCANCER programmes from the Comunidad de Madrid (B2017/BMD‐3884 and Y2020/BIO‐6519) to M.M. CNIO is a Severo Ochoa Center of Excellence (AEI‐MICIU CEX2019‐000891‐S).
The EMBO Journal (2023) 41: e111251
Contributor Information
Carolina Villarroya‐Beltri, Email: cvillarroya@cnio.es.
Marcos Malumbres, Email: malumbres@cnio.es.
Data availability
RNA sequencing data that support the findings of this study have been deposited in the Gene Expression Omnibus under accession code GSE191197. Proteomics data are available at the ProteomeXchange repository under accession number PXD030296. Single‐cell RNA‐seq and ATC‐seq data have been deposited in the Gene Expression Omnibus under accession code GSE196140. Jupyter notebooks and additional code will be made available upon reasonable request.
References
- Bassermann F, Frescas D, Guardavaccaro D, Busino L, Peschiaroli A, Pagano M (2008) The Cdc14B‐Cdh1‐Plk1 axis controls the G2 DNA‐damage‐response checkpoint. Cell 134: 256–267 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Berdougo E, Nachury MV, Jackson PK, Jallepalli PV (2008) The nucleolar phosphatase Cdc14B is dispensable for chromosome segregation and mitotic exit in human cells. Cell Cycle 7: 1184–1190 [DOI] [PubMed] [Google Scholar]
- Biase FH, Cao X, Zhong S (2014) Cell fate inclination within 2‐cell and 4‐cell mouse embryos revealed by single‐cell RNA sequencing. Genome Res 24: 1787–1796 [DOI] [PMC free article] [PubMed] [Google Scholar]
- van den Boom V, Kooistra SM, Boesjes M, Geverts B, Houtsmuller AB, Monzen K, Komuro I, Essers J, Drenth‐Diephuis LJ, Eggen BJ (2007) UTF1 is a chromatin‐associated protein involved in ES cell differentiation. J Cell Biol 178: 913–924 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bremmer SC, Hall H, Martinez JS, Eissler CL, Hinrichsen TH, Rossie S, Parker LL, Hall MC, Charbonneau H (2012) Cdc14 phosphatases preferentially dephosphorylate a subset of cyclin‐dependent kinase (Cdk) sites containing phosphoserine. J Biol Chem 287: 1662–1669 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Buganim Y, Faddah DA, Cheng AW, Itskovich E, Markoulaki S, Ganz K, Klemm SL, van Oudenaarden A, Jaenisch R (2012) Single‐cell expression analyses during cellular reprogramming reveal an early stochastic and a late hierarchic phase. Cell 150: 1209–1222 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Clement A, Solnica‐Krezel L, Gould KL (2012) Functional redundancy between Cdc14 phosphatases in zebrafish ciliogenesis. Dev Dyn 241: 1911–1921 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Correa‐Vazquez JF, Juarez‐Vicente F, Garcia‐Gutierrez P, Barysch SV, Melchior F, Garcia‐Dominguez M (2021) The sumo proteome of proliferating and neuronal‐differentiating cells reveals Utf1 among key sumo targets involved in neurogenesis. Cell Death Dis 12: 305 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dalton S (2015) Linking the cell cycle to cell fate decisions. Trends Cell Biol 25: 592–600 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Delmaghani S, Aghaie A, Bouyacoub Y, El Hachmi H, Bonnet C, Riahi Z, Chardenoux S, Perfettini I, Hardelin JP, Houmeida A et al (2016) Mutations in CDC14A, encoding a protein phosphatase involved in hair cell Ciliogenesis, cause autosomal‐recessive severe to profound deafness. Am J Hum Genet 98: 1266–1270 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dietrich N, Lerdrup M, Landt E, Agrawal‐Singh S, Bak M, Tommerup N, Rappsilber J, Sodersten E, Hansen K (2012) REST‐mediated recruitment of polycomb repressor complexes in mammalian cells. PLoS Genet 8: e1002494 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ferron S, Mira H, Franco S, Cano‐Jaimez M, Bellmunt E, Ramirez C, Farinas I, Blasco MA (2004) Telomere shortening and chromosomal instability abrogates proliferation of adult but not embryonic neural stem cells. Development 131: 4059–4070 [DOI] [PubMed] [Google Scholar]
- Galonska C, Smith ZD, Meissner A (2014) In vivo and in vitro dynamics of undifferentiated embryonic cell transcription factor 1. Stem Cell Rep 2: 245–252 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gaspard N, Bouschet T, Herpoel A, Naeije G, van den Ameele J, Vanderhaeghen P (2009) Generation of cortical neurons from mouse embryonic stem cells. Nat Protoc 4: 1454–1463 [DOI] [PubMed] [Google Scholar]
- Grallert A, Boke E, Hagting A, Hodgson B, Connolly Y, Griffiths JR, Smith DL, Pines J, Hagan IM (2015) A PP1‐PP2A phosphatase relay controls mitotic progression. Nature 517: 94–98 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Graña O, Rubio‐Camarillo M, Fdez‐Riverola F, Pisano DG, Gonzalez‐Peña D (2018) Nextpresso: next generation sequencing expression analysis pipeline. Curr Bioinformatics 13: 583–591 [Google Scholar]
- Gray CH, Good VM, Tonks NK, Barford D (2003) The structure of the cell cycle protein Cdc14 reveals a proline‐directed protein phosphatase. EMBO J 22: 3524–3535 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Guillamot M, Manchado E, Chiesa M, Gomez‐Lopez G, Pisano DG, Sacristan MP, Malumbres M (2011) Cdc14b regulates mammalian RNA polymerase II and represses cell cycle transcription. Sci Rep 1: 189 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Holder J, Mohammed S, Barr FA (2020) Ordered dephosphorylation initiated by the selective proteolysis of cyclin B drives mitotic exit. Elife 9: e59885 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Imtiaz A, Belyantseva IA, Beirl AJ, Fenollar‐Ferrer C, Bashir R, Bukhari I, Bouzid A, Shaukat U, Azaiez H, Booth KT et al (2018) CDC14A phosphatase is essential for hearing and male fertility in mouse and human. Hum Mol Genet 27: 780–798 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jia J, Zheng X, Hu G, Cui K, Zhang J, Zhang A, Jiang H, Lu B, Yates J 3rd, Liu C et al (2012) Regulation of pluripotency and self‐ renewal of ESCs through epigenetic‐threshold modulation and mRNA pruning. Cell 151: 576–589 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, Tunyasuvunakool K, Bates R, Zidek A, Potapenko A et al (2021) Highly accurate protein structure prediction with AlphaFold. Nature 596: 583–589 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kim HJ, Shin J, Lee S, Kim TW, Jang H, Suh MY, Kim JH, Hwang IY, Hwang DS, Cho EJ et al (2018) Cyclin‐dependent kinase 1 activity coordinates the chromatin associated state of Oct4 during cell cycle in embryonic stem cells. Nucleic Acids Res 46: 6544–6560 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kooistra SM, van den Boom V, Thummer RP, Johannes F, Wardenaar R, Tesson BM, Veenhoff LM, Fusetti F, O'Neill LP, Turner BM et al (2010) Undifferentiated embryonic cell transcription factor 1 regulates ESC chromatin organization and gene expression. Stem Cells 28: 1703–1714 [DOI] [PubMed] [Google Scholar]
- Lange C, Huttner WB, Calegari F (2009) Cdk4/cyclinD1 overexpression in neural stem cells shortens G1, delays neurogenesis, and promotes the generation and expansion of basal progenitors. Cell Stem Cell 5: 320–331 [DOI] [PubMed] [Google Scholar]
- Laskowski AI, Knoepfler PS (2012) Utf1: goldilocks for ESC bivalency. Cell Stem Cell 11: 732–734 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lin CH, Yang CH, Chen YR (2012) UTF1 deficiency promotes retinoic acid‐induced neuronal differentiation in P19 embryonal carcinoma cells. Int J Biochem Cell Biol 44: 350–357 [DOI] [PubMed] [Google Scholar]
- Lin H, Ha K, Lu G, Fang X, Cheng R, Zuo Q, Zhang P (2015) Cdc14A and Cdc14B redundantly regulate DNA double‐Strand break repair. Mol Cell Biol 35: 3657–3668 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liu L, Michowski W, Inuzuka H, Shimizu K, Nihira NT, Chick JM, Li N, Geng Y, Meng AY, Ordureau A et al (2017) G1 cyclins link proliferation, pluripotency and differentiation of embryonic stem cells. Nat Cell Biol 19: 177–188 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liu L, Michowski W, Kolodziejczyk A, Sicinski P (2019) The cell cycle in stem cell proliferation, pluripotency and differentiation. Nat Cell Biol 21: 1060–1067 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Malumbres M (2014) Cyclin‐dependent kinases. Genome Biol 15: 122 [DOI] [PMC free article] [PubMed] [Google Scholar]
- McGann JC, Oyer JA, Garg S, Yao H, Liu J, Feng X, Liao L, Yates JR 3rd, Mandel G (2014) Polycomb‐ and REST‐associated histone deacetylases are independent pathways toward a mature neuronal phenotype. Elife 3: e04235 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Michowski W, Chick JM, Chu C, Kolodziejczyk A, Wang Y, Suski JM, Abraham B, Anders L, Day D, Dunkl LM et al (2020) Cdk1 controls global epigenetic landscape in embryonic stem cells. Mol Cell 78: 459–476.e13 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mocciaro A, Schiebel E (2010) Cdc14: a highly conserved family of phosphatases with non‐conserved functions? J Cell Sci 123: 2867–2876 [DOI] [PubMed] [Google Scholar]
- Mocciaro A, Berdougo E, Zeng K, Black E, Vagnarelli P, Earnshaw W, Gillespie D, Jallepalli P, Schiebel E (2010) Vertebrate cells genetically deficient for Cdc14A or Cdc14B retain DNA damage checkpoint proficiency but are impaired in DNA repair. J Cell Biol 189: 631–639 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Morgan DO (2007) The cell cycle: principles of control, p 315. London: New Science Press Ltd; [Google Scholar]
- Neganova I, Tilgner K, Buskin A, Paraskevopoulou I, Atkinson SP, Peberdy D, Passos JF, Lako M (2014) CDK1 plays an important role in the maintenance of pluripotency and genomic stability in human pluripotent stem cells. Cell Death Dis 5: e1508 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Nilsson J (2019) Protein phosphatases in the regulation of mitosis. J Cell Biol 218: 395–409 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Nishimoto M, Fukushima A, Okuda A, Muramatsu M (1999) The gene for the embryonic stem cell coactivator UTF1 carries a regulatory element which selectively interacts with a complex composed of Oct‐3/4 and Sox‐2. Mol Cell Biol 19: 5453–5465 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Nishimoto M, Katano M, Yamagishi T, Hishida T, Kamon M, Suzuki A, Hirasaki M, Nabeshima Y, Nabeshima Y, Katsura Y et al (2013) In vivo function and evolution of the eutherian‐specific pluripotency marker UTF1. PLoS One 8: e68119 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Okuda A, Fukushima A, Nishimoto M, Orimo A, Yamagishi T, Nabeshima Y, Kuro‐o M, Nabeshima Y, Boon K, Keaveney M et al (1998) UTF1, a novel transcriptional coactivator expressed in pluripotent embryonic stem cells and extra‐embryonic cells. EMBO J 17: 2019–2032 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Partscht P, Uddin B, Schiebel E (2021) Human cells lacking CDC14A and CDC14B show differences in ciliogenesis but not in mitotic progression. J Cell Sci 134: jcs255950 [DOI] [PubMed] [Google Scholar]
- Raina K, Dey C, Thool M, Sudhagar S, Thummer RP (2021) An insight into the role of UTF1 in development, stem cells, and cancer. Stem Cell Rev Rep 17: 1280–1293 [DOI] [PubMed] [Google Scholar]
- Rosso L, Marques AC, Weier M, Lambert N, Lambot MA, Vanderhaeghen P, Kaessmann H (2008) Birth and rapid subcellular adaptation of a hominoid‐specific CDC14 protein. PLoS Biol 6: e140 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Roy SH, Clayton JE, Holmen J, Beltz E, Saito RM (2011) Control of Cdc14 activity coordinates cell cycle and development in Caenorhabditis elegans . Mech Dev 128: 317–326 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Singh AM, Chappell J, Trost R, Lin L, Wang T, Tang J, Matlock BK, Weller KP, Wu H, Zhao S et al (2013) Cell‐cycle control of developmentally regulated transcription factors accounts for heterogeneity in human pluripotent cells. Stem Cell Rep 1: 532–544 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Stegmeier F, Amon A (2004) Closing mitosis: the functions of the Cdc14 phosphatase and its regulation. Annu Rev Genet 38: 203–232 [DOI] [PubMed] [Google Scholar]
- Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES et al (2005) Gene set enrichment analysis: a knowledge‐based approach for interpreting genome‐wide expression profiles. Proc Natl Acad Sci USA 102: 15545–15550 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Takahashi K, Yamanaka S (2006) Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 126: 663–676 [DOI] [PubMed] [Google Scholar]
- Thummer RP, Drenth‐Diephuis LJ, Carney KE, Eggen BJ (2010) Functional characterization of single‐nucleotide polymorphisms in the human undifferentiated embryonic‐cell transcription factor 1 gene. DNA Cell Biol 29: 241–248 [DOI] [PubMed] [Google Scholar]
- Thummer RP, Drenth‐Diephuis LJ, Eggen BJ (2012) Constitutive GFP‐UTF1 expression interferes with ES and EC cell differentiation. Stem Cell Res Ther 2: 127 [Google Scholar]
- Touati SA, Hofbauer L, Jones AW, Snijders AP, Kelly G, Uhlmann F (2019) Cdc14 and PP2A phosphatases cooperate to shape phosphoproteome dynamics during mitotic exit. Cell Rep 29: 2105–2119.e4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Uddin B, Partscht P, Chen NP, Neuner A, Weiss M, Hardt R, Jafarpour A, Hessling B, Ruppert T, Lorenz H et al (2019) The human phosphatase CDC14A modulates primary cilium length by regulating centrosomal actin nucleation. EMBO Rep 20: e46544 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Visintin R, Craig K, Hwang ES, Prinz S, Tyers M, Amon A (1998) The phosphatase Cdc14 triggers mitotic exit by reversal of Cdk‐dependent phosphorylation. Mol Cell 2: 709–718 [DOI] [PubMed] [Google Scholar]
- Wang XQ, Lo CM, Chen L, Ngan ES, Xu A, Poon RY (2017) CDK1‐PDK1‐PI3K/Akt signaling pathway regulates embryonic and induced pluripotency. Cell Death Differ 24: 38–48 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wei Z, Peddibhotla S, Lin H, Fang X, Li M, Rosen JM, Zhang P (2011) Early‐onset aging and defective DNA damage response in Cdc14b‐deficient mice. Mol Cell Biol 31: 1470–1477 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wen Z, Zhu H, Zhang A, Lin J, Zhang G, Liu D, Xiao Y, Ye C, Sun D, Wu B et al (2020) Cdc14a has a role in spermatogenesis, sperm maturation and male fertility. Exp Cell Res 395: 112178 [DOI] [PubMed] [Google Scholar]
- Wieczorek S, Combes F, Lazar C, Giai Gianetto Q, Gatto L, Dorffer A, Hesse AM, Coute Y, Ferro M, Bruley C et al (2017) DAPAR & ProStaR: software to perform statistical analyses in quantitative discovery proteomics. Bioinformatics 33: 135–136 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wolf FA, Angerer P, Theis FJ (2018) SCANPY: large‐scale single‐cell gene expression data analysis. Genome Biol 19: 15 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wurzenberger C, Gerlich DW (2011) Phosphatases: providing safe passage through mitotic exit. Nat Rev Mol Cell Biol 12: 469–482 [DOI] [PubMed] [Google Scholar]
- Xie Z, Bailey A, Kuleshov MV, Clarke DJB, Evangelista JE, Jenkins SL, Lachmann A, Wojciechowicz ML, Kropiwnicki E, Jagodnik KM et al (2021) Gene set knowledge discovery with Enrichr. Curr Protoc 1: e90 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yang CS, Chang KY, Rana TM (2014) Genome‐wide functional analysis reveals factors needed at the transition steps of induced reprogramming. Cell Rep 8: 327–337 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhao Y, Yin X, Qin H, Zhu F, Liu H, Yang W, Zhang Q, Xiang C, Hou P, Song Z et al (2008) Two supporting factors greatly improve the efficiency of human iPSC generation. Cell Stem Cell 3: 475–479 [DOI] [PubMed] [Google Scholar]
- Zhou Y, Zhou B, Pache L, Chang M, Khodabakhshi AH, Tanaseichuk O, Benner C, Chanda SK (2019) Metascape provides a biologist‐oriented resource for the analysis of systems‐level datasets. Nat Commun 10: 1523 [DOI] [PMC free article] [PubMed] [Google Scholar]