LIF-dependent survival of embryonic stem cells is regulated by a novel palmitoylated Gab1 signalling protein

Abstract

The cytokine leukaemia inhibitory factor (LIF) promotes self-renewal of mouse embryonic stem cells (ESCs) through activation of the transcription factor Stat3. However, the contribution of other ancillary pathways stimulated by LIF in ESCs, such as the MAPK and PI3K pathways, is less well understood. We show here that naive-type mouse ESCs express high levels of a novel effector of the MAPK and PI3K pathways. This effector is an isoform of the Gab1 (Grb2-associated binder protein 1) adaptor protein that lacks the N-terminal pleckstrin homology (PH) membrane-binding domain. Although not essential for rapid unrestricted growth of ESCs under optimal conditions, the novel Gab1 variant (Gab1β) is required for LIF-mediated cell survival under conditions of limited nutrient availability. This enhanced survival is absolutely dependent upon a latent palmitoylation site that targets Gab1β directly to ESC membranes. These results show that constitutive association of Gab1 with membranes through a novel mechanism promotes LIF-dependent survival of murine ESCs in nutrient-poor conditions.

Keywords: Embryonic stem cells; Gab1; Leukaemia inhibitory factor; Palmitoylation; Stem cell survival.

Fig. 1.

Expression profile of a novel Gab1 variant protein. (A) Gab1 western blot analysis of whole cell lysates from ESCs (ES), embryonal carcinoma (EC), 10T1/2 embryonic fibroblasts (fb) and pooled E12.5 embryo tissues (em). The positions of the 110 and 95 kDa Gab1 proteins are indicated. (B) Schematic of the Gab1 gene locus, showing the location of the transcription start sites for Gab1α and Gab1β, and the exons. (C) Schematic showing the structure of Gab1α and Gab1β proteins. Dark grey and light grey areas are proline rich-regions and the Met binding domain (MBD), respectively. (D) Gab1 and Oct4 western blots of whole cell lysates of ESCs (ES), primary embryonic fibroblasts (fb), pre-iPS cells (pr), iPSCs reprogrammed in 2i medium (2i) or serum/LIF medium (sr) and an epiblast stem cell line (Epi). (E) Quantitative RT-PCR analysis of Gab1β and Gab1α expression in ESCs (ES) transitioned into epiblasts stem cells (Epi) and embryoid bodies (EB). Results represent means±s.d. from one experiment. Expression is normalised relative to the level in ESCs.

Fig. 2.

Gab1β as an ESC signal transducer. (A) Western blot of Gab1 immunoprecipitates from unstimulated ESCs (−), and ESCs stimulated with serum (sr) or LIF (lf) for 10 min, probed with Gab1pY627, Gab1pY307, phosphotyrosine and Gab1 specific antibodies. The dashes indicate position of 100 kDa MW marker. (B) Western blot of SHP2 immunoprecipitates from cells treated as in A probed with antibodies to Gab1 and SHP2. (C) Western blot of cell lysates from wild-type ESCs, and Gab1β knockout ESC line stably transfected with empty vector (−/−), Gab1β, and Gab1α expression vectors, unstimulated or following stimulation with LIF for 15 min. The blot was incubated with antibodies to phospho-Akt, Akt, phospho-Erk, Erk1/2, phospho-Stat3 and Gab1.

Fig. 3.

Function of Gab1β in ESCs. (A) Undifferentiated ESCs identified by alkaline phosphatase activity (pink stain) in two Gab1β heterozygous (HET) and two Gab1β knockout (KO) E14Tg2a cell lines 5 days after plating. Scale bars: 200 µm. (B) Growth curves of Gab1β heterozygous (HET) and Gab1β knockout (KO) E14Tg2a cell lines measured by DNA-dependent fluorescence in live cells. Cells were plated at a density close to that for routine passaging of ESCs (6×10⁴ cells/cm²) and DNA fluorescence was measured daily for 6 days. The data represent the mean±s.e.m. of three independent experiments. Statistical analysis using Student's t-test showed significant differences between the means for the HET and KO lines from day 3 to day 6 (P≤0.005). (C) Growth curves performed as described in B with pooled transfected ESCs: Gab1β heterozygous (HET-mCherry), Gab1β knockout (KO-EGFP) and Gab1β knockout stably expressing Gab1β from a cDNA expression vector (KO-Gab1β). Student's t-test analysis showed significant differences between the KO-EGFP and KO-Gab1β lines from day 4 to day 6 (P<0.0001), but not between HET-mCherry and KO-Gab1β cells.

Fig. 4.

Gab1β regulation of ESC cell cycle and apoptosis. (A) Flow cytometry cell cycle analysis of Gab1β-expressing or non-expressing ESCs. Triplicate ESC samples were collected on day 2 and 3 of culture, fixed and stained using propidium iodide, and analysed by flow cytometry. Representative scans are shown and sub-diploid cell material is indicated with an asterisk. (B) Quantification of cell cycle distribution. Mean±s.d. values of cell cycle phases generated from three independent cultures (Student's t-test, *P<0.005). (C) Flow cytometry of Annexin V staining in day 1 and day 2 ESC cultures. Grey and black bars are values from Gab1β KO and Gab1β-expressing (Gab1β KO+Gab1β cDNA vector) ESCs, respectively (Student's t-test, *P<0.005). (D) Apoptosis, cytotoxicity and viability assays of day 2 cultures of Gab1β KO and Gab1β-expressing ESCs. Values are means±s.d. of four biological replicates (apoptosis and cytotoxicity t-tests, P<0.0001; viability t-test, P<0.005).

Fig. 5.

Gab1β and ESC response to nutrient availability. (A) Growth profiles of Gab1β KO and Gab1β-expressing ESCs treated with 25 nM mTOR inhibitor INK128 (mTORi). Values are means±s.e.m. of three independent experiments. (B,C) Growth of Gab1β KO and Gab1β-expressing ESC cultures supplemented on day 2 with 10 µl PBS, regular DMEM (+) or DMEM (−) lacking glucose, glutamine and sodium pyruvate. Values are means±s.d. of three biological replicates.

Fig. 6.

Gab1β-mediated ESC survival depends on LIF, and is rescued by Mek inhibition. Growth profiles of Gab1β KO and Gab1β-expressing ESCs cultured in the presence or absence of: (A) LIF (±1 µM PD0325901 Mek inhibitor); (B) 2i (1 µM PD0325901 Mek inhibitor, 3 µM CHIR99021 GSK3 inhibitor) +LIF serum free medium; (C) FGFR inhibitor PD173074 (one experiment with quadruplicate biological samples; and (D) 1 µM PD0325901 Mek inhibitor. Data points in A,B and D represent the means±s.e.m. of three independent experiments Data points in graph C represent the means from four biological replicates in one experiment. For all graphs, Student's t-tests showed statistically significant differences between serum+LIF-treated KO and Gab1β-expressing control lines at days 4-6 (P≤0.0001), and between KO control and KO treated (KO+Meki) cells (P≤0.0001).

Fig. 7.

Localisation of Gab1β at the cell membrane. Confocal images of (A) wild-type ESCs, (B) Gab1β knockout ESCs and (C) Gab1β knockout ESCs stably transfected with a Gab1β cDNA expression vector, immunostained with antibodies against Gab1 (red) and counterstained for DNA with DAPI (blue). (D-I) Confocal images of EGFP fluorescence in ESC stably expressing the following Gab1-EGFP fusion proteins: (D) Gab1β-EGFP; (E) Gab1α-EGFP; (F) N-terminal EGFP-Gab1β; (G) Gab1β-EGFP fusion lacking 15 N-terminal amino acids; (H) the 47 N-terminal amino acid peptide of Gab1β fused to EGFP; and (I) 15 N-terminal amino-acid peptide of Gab1β fused to EGFP. White arrows highlight consistently observed intracellular accumulations of Gab1β proteins. Scale bars: 10 µm.

Fig. 8.

The N-terminus of Gab1β directs palmitoylation and is required for membrane localisation and function. (A) Schematic outlining the structure of the Gab1 PH domain and comparison of the N-terminus of Gab1β with corresponding regions of other Gab-related proteins. The arrangement of the seven β-sheets and the α-helix, and relative positions of a putative nuclear localisation sequence (NLS) and PIP3 binding residues are shown. Cysteines within the α-helix are highlighted in bold. (B) EGFP immunoprecipitates from ESCs stably transfected with EGFP or Gab1-EGFP fusion proteins cultured overnight with [³H]palmitate, were fractionated on an SDS protein gel, transferred to a filter and autoradiographed (top) and probed for EGFP protein (bottom). CA3 designates variants in which the three N-terminal cysteines of Gab1β are substituted with alanine. (C) Confocal images of EGFP fluorescence in ESCs transfected with the 47-EGFP, 47CA3-EGFP, PH-domain EGPF and PH-domain CA3-EGPF fusion proteins (counterstained with DAPI). Scale bars: 10 µm. (D) Growth curves of Gab1β knockout (KO) ESCs and Gab1β KO ESCs stably transfected with either Gab1β wild-type or CA3 mutant cDNA expression vectors. The graphs represent the means±s.e.m. obtained from three independent experiments. Student's t-test analysis showed significant differences between the KO and Gab1β-expressing lines from day 4 to day 6 (P<0.0001), but not between KO and KO-Gab1β:CA3m cells.