The ETS transcription factor ERF controls the exit from the naïve pluripotent state in a MAPK-dependent manner

Abstract

The naïve epiblast transitions to a pluripotent primed state during embryo implantation. Despite the relevance of the FGF pathway during this period, little is known about the downstream effectors regulating this signaling. Here, we examined the molecular mechanisms coordinating the naïve to primed transition by using inducible ESC to genetically eliminate all RAS proteins. We show that differentiated RAS^KO ESC remain trapped in an intermediate state of pluripotency with naïve-associated features. Elimination of the transcription factor ERF overcomes the developmental blockage of RAS-deficient cells by naïve enhancer decommissioning. Mechanistically, ERF regulates NANOG expression and ensures naïve pluripotency by strengthening naïve transcription factor binding at ESC enhancers. Moreover, ERF negatively regulates the expression of the methyltransferase DNMT3B, which participates in the extinction of the naïve transcriptional program. Collectively, we demonstrated an essential role for ERF controlling the exit from naïve pluripotency in a MAPK-dependent manner during the progression to primed pluripotency.

Fig. 1.. ERF expression correlates with naïve pluripotency markers.

(A) Graph showing mean nuclear fluorescent intensity for ERF in mouse embryos at different days of embryonic (E) development. A.U., arbitrary unit. (B) Immunofluorescence analysis of NANOG and ERF in mouse embryos at E2.75, E3.5, E4.0, and E4.75. Note that ERF is expressed in both ICM and TE. However, at E4.75, ERF and NANOG are mostly down-regulated in epiblast cells. Dashed lines highlight the ICM. 4′,6-Diamidino-2-phenylindole (DAPI) was used to visualize nuclei. Scale bars, 20 μm. See Materials and Methods for information on the number of cells and embryos used in (A) and (B). (C) Graphs showing relative nuclear fluorescence intensity of ERF and NANOG (top plots) and ERF and KLF4 (bottom plots). Every dot represents one single nucleus, and each plot corresponds to an individual E3.5 embryo. Two representative examples are shown but at least 10 embryos were analyzed. (D) Immunofluorescence analysis of 2iL- and FA-treated RAS^lox/lox and RAS^KO ESC and stained for ERF (red) and NANOG (purple). DAPI was used to visualize nuclei. Scale bars, 15 μm. Two independent experiments with two ESC clones per genotype were performed, but one representative is shown.

Fig. 2.. Successful exit from naïve pluripotency requires down-regulation of ERF.

(A) Central confocal optical sections of RAS^lox/lox, ERF^KO, RAS^KO, and RAS^KO; ERF^KO embryonic cell rosettes stained for NANOG (red) and podocalyxin (PDX; green). DAPI was used to visualize nuclei. Scale bars, 10 μm. Two independent experiments with two ESC clones per genotype were performed, but one representative is shown. (B) Graph showing the percentage of embryonic rosettes generating a lumen (PDX⁺) in all genotypes. Two independent experiments with biological replicates are shown, and at least a total of 50 rosettes were counted per sample. ***P < 0.001, Student’s t test. (C) Graph showing mean NANOG nuclear fluorescence intensity per nucleus in embryonic rosettes 48 hours after seeding. One representative experiment is shown, and a total of 70 nuclei from different rosettes were counted per sample. ***P < 0.001, Student’s t test. (D) Flow cytometry analysis of REX1-deGFP reporter ESC from all genotypes in 2iL-ESC and EpiLC after 48 hours of induction with FA. Three independent experiments were performed, but one representative experiment is shown. (E) Graph showing the number of alkaline phosphatase positive colonies in a colony-forming assay using ESC from all genotypes. Two independent experiments are shown with at least two technical replicates. ***P < 0.001, Student’s t test.

Fig. 3.. ERF controls the transition to primed pluripotency in a MAPK-dependent manner.

(A) PCA plot of RNA-seq data of RAS^lox/lox, ERF^KO, RAS^KO, and RAS^KO; ERF^KO ESC cultured under naïve conditions (2iL) or induced to differentiate to EpiLC (FA) for 48 hours. Three replicates per condition are shown. (B) Heatmap generated from RNA-seq data from samples described in (A) showing the average from three replicates. (C) Immunofluorescence analysis of 2iL- and FA-treated RAS^lox/lox and RAS^KO ESC stained for OTX2 (green), ERF (red), and NANOG (purple). DAPI was used to visualize nuclei. Scale bars, 100 μm. Two independent experiments were performed but one representative experiment is shown. (D) PCA plot of RNA-seq datasets (three replicates) showing 2iL- and FA-treated RAS^lox/lox and RAS^KO ESC along with RNA-seq datasets (two replicates) from a time course experiment during EpiLC induction (0, 1, 6, 12, 24, 36, 48, and 72 hours) (20). (E) PCA plot of RNA-seq datasets showing 2iL- and FA-treated RAS^lox/lox and RAS^KO ESC along with RNA-seq datasets from RSC {ESC lines grown in LIM [LIF, IWP2 (WNT inhibitor), and MEKi]}, FSC, fPSC, and XPSC (–10). At least two replicates are shown. (F) Heatmap generated from RNA-seq data from samples described in (E) showing the average from two or three replicates as applicable to each sample.

Fig. 4.. ERF ensures optimal naïve pluripotent TF expression in ESC.

(A) Chromatin immunoprecipitation–sequencing read density plot (RPKM) showing OCT4, SOX2, and NANOG occupancy at 2074 ERF-binding sites at enhancers (pink) or 2074 randomly selected non-ERF bound enhancers (orange). Graphs show quantifications of the TF enrichment in each set of sites. ***P < 0.001, Student’s t test. Data were obtained from (66). (B) Graph showing relative fold change (log₂) expression of the indicated genes by triplicate in RAS^lox/lox and ERF^KO ESC grown in 2iL. Genes showing at least a 50% reduction in expression are considered differentially expressed. ***P < 0.001 and **P < 0.01, Student’s t test. (C) Graph showing the relative fold change (log₂) expression of 15 naïve-associated genes in all genotypes under 2iL conditions. In (B) and (C), data for each gene was normalized to the average across all samples and was obtained from RNA-seq datasets. ***P < 0.001, Student’s t test. (D) Unidimensional PCA plot of RNA-seq data from RAS^lox/lox and ERF^KO ESC (three replicates) cultured under 2iL conditions or differentiated to EpiLC (FA). (E) CUT&RUN read density plot (RPKM) showing NANOG and SOX2 occupancy in the set of 2074 ERF-binding sites at enhancers in RAS^lox/lox (blue) and ERF^KO (green) ESC cultured in 2iL. (F) Genome browser tracks showing NANOG and SOX2 occupancy and RNA-seq read count (RPKM) at the PRDM14 gene in the indicated genotypes. ERF binding profile in RAS^KO ESC is shown. ERF binding sites are highlighted. In (E) and (F), inputs [immunoglobulin G (IgG)] are shown as a reference control.

Fig. 5.. The naïve pluripotent TF network is active in FA-RAS^KO ESC.

(A) Plot showing the percentage of OCT4 binding sites co-occupied by ERF in ESC, EpiLC, and common between ESC and EpiLC. (B) Histogram plots showing fold expression changes (log₂) for genes associated to ERF/OCT4^ESC characterized by differential expression between ESC and EpiLC in RAS^lox/lox and RAS^KO cells. N indicates the total number of genes. Data were obtained from RNA-seq datasets. (C) CUT&RUN read density plot (RPKM) showing H3K27ac (green) and NANOG (red) occupancy in ERF/OCT4^ESC sites in RAS^lox/lox and RAS^KO ESC cultured in 2iL or differentiated to EpiLC (FA). Inputs (IgG) is shown as reference control. (D) Genome browser tracks showing H3K27ac deposition, NANOG, ERF occupancy, and RNA-seq RPKM read count at the KLF4 and NANOG genes in the indicated genotypes. Inputs (IgG) are shown as a reference control. ERF binding sites are highlighted. (E) Graph showing the relative fold change (log₂) expression of the indicated ERF-bound super-enhancer–associated genes in RAS^lox/lox and RAS^KO ESC cultured in 2iL or differentiated to EpiLC (FA). For each gene, data were normalized to the average across all samples. Genes showing at least a 50% reduction in expression are considered differentially expressed. Statistical tests are only shown for RAS^KO ESC. ***P < 0.001, **P < 0.01, and *P < 0.05 Student’s t test. Data shown are averages from triplicates and were obtained from RNA-seq datasets.

Fig. 6.. OTX2 shows promiscuous binding in naïve and primed genes in FA-RAS^KO ESC.

(A) Plot showing the relative fold change (log₂) expression for OTX2 in ESC from all genotypes cultured in 2iL or differentiated to EpiLC (FA). Data were normalized to the average across all samples. ***P < 0.001, Student’s t test. Data shown from triplicates and were obtained from RNA-seq datasets. (B) Genome browser tracks showing SOX2 and NANOG occupancy and RNA-seq RPKM read count at the OTX2 gene in RAS^lox/lox and ERF^KO ESC cultured in 2iL. ERF binding profile in RAS^KO ESC is also shown. ERF binding sites are highlighted. (C) Histogram plots showing fold expression changes (log₂) for genes associated to ERF/OCT4^Common that showed differential expression between ESC and EpiLC in RAS^lox/lox and RAS^KO cells. N indicates the total number of genes. Data were obtained from RNA-seq datasets. (D) CUT&RUN read density plot (RPKM) showing NANOG (red) and OTX2 (purple) occupancy in the indicated ERF/OCT4^Common, ERF/OCT4^ESC, and ERF/OCT4^EpiLC sites in RAS^lox/lox and RAS^KO ESC cultured in 2iL or differentiated to EpiLC (FA). (E) Genome browser tracks showing NANOG and OTX2 occupancy and RNA-seq RPKM read count at the LEFTY1 and LEFTY2 genes in RAS^lox/lox and RAS^KO ESC cultured in 2iL or differentiated to EpiLC (FA). ERF binding sites are highlighted. In (B), (D), and (E), corresponding inputs (IgG) are shown as a reference control.

Fig. 7.. ERF controls the levels of de novo methylation during transition to EpiLC.

(A) Western blot analysis performed in ESC from all genotypes cultured in 2iL or differentiated to EpiLC (FA). Tubulin levels are shown as a loading control. One representative experiment is shown but two independent clones were used. (B) Genome browser tracks showing H3K27ac deposition and OCT4 occupancy at LIN28B in RAS^lox/lox ESC cultured in 2iL or differentiated to EpiLC (FA). ERF binding profile in RAS^KO ESC is also shown. Input (IgG) is shown as a reference control. ERF binding site is highlighted. (C) Graph showing the percentage of methylated CpG sites in ESC from all genotypes cultured in 2iL or differentiated to EpiLC (FA). (D) Genome browser tracks showing the level of methylation at CpG sites in a region of chromosome 12 in ESC from all genotypes cultured in 2iL or differentiated to EpiLC (FA). (E) % CpG methylation averaged across the TSS of all protein-coding mouse genes (top panels) or centered at CpG islands (bottom panels) in ESC from all genotypes cultured in 2iL or differentiated to EpiLC (FA). TSS, transcription start sites. (F) Schematic model showing the dual role of ERF during the naïve to primed pluripotent transition. In the absence of FGF signaling, ERF ensures an optimal level of expression for naïve TFs. Following ERF phosphorylation and gene silencing, ESC transition into EpiLC along with global CpG methylation and silencing of the naïve transcriptional network.