Phosphoproteomics identifies a bimodal EPHA2 receptor switch that promotes embryonic stem cell differentiation

Abstract

Embryonic Stem Cell (ESC) differentiation requires complex cell signalling network dynamics, although the key molecular events remain poorly understood. Here, we use phosphoproteomics to identify an FGF4-mediated phosphorylation switch centred upon the key Ephrin receptor EPHA2 in differentiating ESCs. We show that EPHA2 maintains pluripotency and restrains commitment by antagonising ERK1/2 signalling. Upon ESC differentiation, FGF4 utilises a bimodal strategy to disable EPHA2, which is accompanied by transcriptional induction of EFN ligands. Mechanistically, FGF4-ERK1/2-RSK signalling inhibits EPHA2 via Ser/Thr phosphorylation, whilst FGF4-ERK1/2 disrupts a core pluripotency transcriptional circuit required for Epha2 gene expression. This system also operates in mouse and human embryos, where EPHA receptors are enriched in pluripotent cells whilst surrounding lineage-specified trophectoderm expresses EFNA ligands. Our data provide insight into function and regulation of EPH-EFN signalling in ESCs, and suggest that segregated EPH-EFN expression coordinates cell fate with compartmentalisation during early embryonic development.

Fig. 1. FGF4 signalling in mESCs promotes EPHA2 Ser phosphorylation.

a FGF4 is a key signal that promotes differentiation of pluripotent mESCs. b Workflow for phosphoproteomic analysis of FGF4 signalling in Fgf4^−/− mESCs. Volcano plot showing significantly modified phosphosites after stimulation of Fgf4^−/− mESCs with FGF4 for 5 min (c) and 20 min (d). Phosphosites on known FGF4 pathway components are highlighted. e Protein kinase phosphopeptides that are significantly upregulated (>2-fold) on at least one time point (5 or 20 min) compared with control. Data are presented as mean ± SD (n = 3). f Fgf4^−/− mESCs were stimulated with FGF4 for the indicated time, and EPHA2 pS898 and EPHA2 levels determined by immunoblotting. Source data are provided as a Source Data file.

Fig. 2. EPHA2 is critical for EFN ligand responses in mESCs.

a Average protein copy number per cell determined for receptor kinases in mESCs, using quantitative whole-cell proteomics. Data are presented as mean ± SD (n = 3). b Workflow for quantification of EPH–EFN interactions in mESCs by EFN ligand affinity purification mass spectrometry. c Proof-of-principle identification of EPH–EFN interactions by EFN ligand affinity purification. EPHA2 levels were determined by immunoblotting. d Coomassie staining of EFNA1/B1 affinity purification from mESCs. EFNA1, EFNB1 and EPHA2 proteins are indicated. (*) = non-specific band. e Mass-spectrometry analysis of 75–130-kDa region of the Coomassie stained EFNA1/EFNB1 affinity purification shown in (d). Total spectral counts recovered for each EPH receptor family member are indicated. f EFNA1/B1 affinity purification from intact Epha2^+/+ and pooled Epha2^−/− mESCs. Phosphotyrosine (pTyr), EPHA2 and ERK1/2 levels were determined by immunoblotting. Note that the pTyr signal is specific for EPHA2, and is not detected in the absence of EFN ligand. g EPHA2 was immunoprecipitated from Epha2^+/+ and Epha2^−/− mESCs, and pTyr, EPHA2, EFNA1 and ERK1/2 levels determined by immunoblotting. h Epha2^+/+ mESCs were stimulated with EFNA1-expressing Epha2^−/− mESCs for 15 or 40 min. EPHA2 was immunoprecipitated, and pTyr and EPHA2 levels determined by immunoblotting. Source data are provided as a Source Data file.

Fig. 3. EFNA1-activated EPHA2 supports mESC pluripotency and restricts commitment to differentiation by suppressing ERK1/2 signalling.

a Epha2^+/+, Epha2^−/− or Epha2^−/− mESCs (clone C4) stably expressing EPHA2 were cultured in the absence of LIF for 48 h. EPHA2, KLF4, DNMT3B, NANOG and OCT4 levels were determined by immunoblotting. b Epha2^+/+ or Epha2^−/− (clone C4) mESCs were maintained in 2i or differentiated in N2B27 media for 72 or 96 h, respectively, whereupon 10% of cells were replated in 2i. Total alkaline phosphatase staining is represented relative to Epha2^+/+ mESCs. The total number of alkaline phosphatase-positive colonies for Epha2^+/+ and Epha2^−/− mESCs is shown, and also represented relative to Epha2^+/+ mESCs. Data show mean ± SEM (n = 3); statistical significance was determined using unpaired two-sided Student’s t test comparing Epha2^−/− with the Epha2^+/+ control (****P < 0.0001, **P = 0.0016). c Epha2^+/+ or Epha2^−/− (clone C4) mESCs stably expressing EFNA1, along with the respective parental controls, were grown in LIF/FBS, and KLF4, NANOG, EPHA2, EFNA1 and ERK1/2 levels determined by immunoblotting, or EPHA2 immunoprecipitated and pTyr and EPHA2 levels determined by immunoblotting. d Epha2^+/+, Epha2^−/− or Epha2^−/− mESCs stably expressing EPHA2 were differentiated as embryoid bodies for 10 days, and the levels of Fgf5, Brachyury, Mixl and Cer1 mRNA determined by qRT-PCR. Box-and-whisker plots show median, first and third quartiles, and maximum and minimum values. The results shown are for technical replicates from two independent experiments, including three Epha2^−/− clones (n = 3); statistical significance at day 4 was determined using unpaired two-sided Student’s t test comparing each group with the Epha2^+/+ control (ns = not significant, *P = 0.0252, **P = 0.0045, ***P < 0.0001). e Epha2^+/+ mESCs cultured in LIF/FBS were stimulated with 1 μg/ml clustered EFNA1 for the indicated times. ppERK1/2, total ERK1/2, STAT3 pY705 and total STAT3 levels were determined by immunoblotting. EPHA2 was immunoprecipitated, and pTyr and EPHA2 levels determined by immunoblotting. ppERK1/2 signal was quantified; data show mean ± SD (n = 3); statistical significance was determined using one-sample two-sided t test comparing each group with control, theoretical mean = 1 (ns = not significant, 5 min; *P = 0.0467, 30 min; ***P = 0.0005, 45 min; **P = 0.0014, 60 min; **P = 0.0018). f Epha2^+/+ mESCs cultured in LIF/FBS were stimulated with 1 μg/ml clustered EFNA1 for the indicated times. SHP2 was immunoprecipitated, and SHP2 and EPHA2 levels detected by immunoblotting. Source data are provided as a Source Data file.

Fig. 4. FGF4–ERK1/2 signalling inhibits EPHA2 activation and rewires EPH–EFN expression.

a Diagram of potential phosphorylation sites within the EPHA2 S898 motif. Mass spectrometry analysis detects phosphorylation of at least three sites in EPHA2 immunoprecipitated from FGF4-stimulated Fgf4^−/− mESCs (see Supplementary Table 1). b Fgf4^−/− mESCs were treated with 10 μM of the indicated inhibitors for 1 h, and stimulated with FGF4 for 10 min. EPHA2 pS898, EPHA2 and AKT pS473 levels were determined by immunoblotting. c Epha2^−/− mESCs were transfected with either wild type or 5E EPHA2 constructs, and stimulated with 1 μg/ml clustered EFNA1 for 15 min. EPHA2 was immunoprecipitated, and pTyr and EPHA2 levels determined by immunoblotting and quantified. Data show mean ± SD (n = 4). d EPHA2 was immunoprecipitated from EPHA2 WT knock-in (KI) or 5A KI cell lines and pTyr and EPHA2 levels determined by immunoblotting (upper panel). Relative pTyr/EPHA2 signal was quantified (lower panel). Data show mean ± SD (n = 4). e Phase-contrast images of EPHA2 WT KI or 5A KI mESC lines; scale bar = 100 µM. f EPHA2 WT KI or 5A KI cell lines were cultured in LIF/FBS medium for 48 h, and KLF4, NANOG, DNMT3B, OCT4 ppERK1/2 and ERK1/2 levels determined by immunoblotting. g Epha2^+/+ mESCs were differentiated as embryoid bodies for 10 days, and EPH receptor expression determined by qRT-PCR at the indicated time points. Box-and-whisker plots show median, first and third quartiles, and maximum and minimum values of four technical replicates (n = 4). h Epha2^+/+ mESCs were differentiated as embryoid bodies for 10 days, and EFN ligand expression determined by qRT-PCR at the indicated time points. Box-and-whisker plots show median, first and third quartiles, and maximum and minimum values of four technical replicates (n = 4). Epha2 (i) and Efna1 (j) mRNA expression in 2i mESCs undergoing differentiation in N2B27 was determined by qRT-PCR analysis at the indicated time points. Box-and-whisker plots show median, first and third quartiles, and maximum and minimum values of two technical and three biological replicates (n = 3). Source data are provided as a Source Data file.

Fig. 5. OCT4 is required for EPHA2 expression and EFN ligand responses in mESCs.

a Epha2^+/+ mESCs were transfected with the indicated siRNAs, and EPHA2, OCT4 and SOX2 levels determined by immunoblotting. A non-specific band was used as a loading control. b Epha2 and Oct4 mRNA expression was determined by qRT-PCR following transfection of Epha2^+/+ mESCs with control or OCT4 siRNA. Data show mean ± SEM of four technical replicates (n = 4). c Pluripotency transcription factor-binding sites in Epha2 gene regulatory regions were extracted from CODEX mESC ChIP-SEQ data (http://codex.stemcells.cam.ac.uk). d Epha2^+/+ mESCs were transfected with the indicated siRNAs, and stimulated with 1 μg/ml clustered EFNA1 for 15 min. EPHA2 was immunoprecipitated using EPHA2 antibody, and pTyr and EPHA2 levels determined by immunoblotting. e EFNA1/B1 affinity purification from control, dasatinib-treated or siOct4-transfected mESCs. pTyr, EPHA2, OCT4 and ERK1/2 levels were determined by immunoblotting. Source data are provided as a Source Data file.

Fig. 6. Reciprocal expression of EPH receptors and ligands in pluripotent and lineage- specified cells of early mouse and human embryos.

Expression (RPKM) of EPH (a, b) or EFN (c, d) mRNA in the inner cell mass (ICM) or trophectoderm (TE) of 64-cell mouse embryos (a, c) or 5-day human embryos (b, d). a, c n = 33 biologically independent cells for ICM, and n = 28 biologically independent cells for TE. b, d n = 73 biologically independent cells for ICM, and n = 142 biologically independent cells for TE. Box-and-whisker plots show median, first and third quartiles, and maximum and minimum values. e EPHA2 regulation and function in mESCs. In the pluripotent state, OCT4 and other pluripotency factors promote EPHA2 receptor expression, enabling activation by EFNA ligands to support pluripotency by restraining ERK1/2. During differentiation, FGF4 drives ERK1/2–RSK activity to phosphorylate and inhibit EPHA2, whilst ERK1/2 suppresses an OCT4–EPHA2 transcriptional module to disable EPHA2 receptor expression. Source data are provided as a Source Data file.