Cell fate decisions are specified by the dynamic ERK interactome

Abstract

Extracellular signal-regulated kinase (ERK) controls fundamental cellular functions, including cell fate decisions. In PC12, cells shifting ERK activation from transient to sustained induces neuronal differentiation. As ERK associates with both regulators and effectors, we hypothesized that the mechanisms underlying the switch could be revealed by assessing the dynamic changes in ERK-interacting proteins that specifically occur under differentiation conditions. Using quantitative proteomics, we identified 284 ERK-interacting proteins. Upon induction of differentiation, 60 proteins changed their binding to ERK, including many proteins that were not known to participate in differentiation. We functionally characterized a subset, showing that they regulate the pathway at several levels and by different mechanisms, including signal duration, ERK localization, feedback, crosstalk with the Akt pathway and differential interaction and phosphorylation of transcription factors. Integrating these data with a mathematical model confirmed that ERK dynamics and differentiation are regulated by distributed control mechanisms rather than by a single master switch.

Figure 1

Identification of dynamic ERK1 interactions. (a) Workflow of the SILAC experiments. PC12 cells were grown in media supplemented with ‘light’, ‘heavy’ and ‘super heavy’ arginine and lysine isotopes. The coloured isotope numbers highlight changes in residues compared with serum-starved cells. Endogenous ERK1 was immunoprecipitated from serum-starved PC12 cells stimulated with EGF (20 ng ml⁻¹) or NGF (100 ng ml⁻¹). Proteins were eluted from the antibody, combined, separated by SDS–PAGE and analysed by mass spectrometry. IP, immunoprecipitate. LC-MS/MS, liquid chromatography-tandem mass spectrometry. (b, c) Heatmaps of growth factor-induced changes of proteins bound to ERK after 5 min (b) and 30 min (c) of stimulation, based on mass spectroscopy data. (d) Confirmation of selected dynamic interactions by co-immunoprecipitation and western blot analysis. PC12 cells were starved overnight and stimulated with EGF or NGF as indicated. Left panel, ERK1 immunoprecipitates were subjected to western blotting with the indicated antibodies. Asterisk indicates an unspecific band. Complete blots are shown in Supplementary Information, Fig. S17. Middle panel, total lysates from the same cells were blotted showing that protein expression remained constant during the experiment. Right panel, dynamic changes in the association to ERK1, as shown from quantified mass spectrometry data.

Figure 2

NF1 controls Ras activity and signal duration. (a) PC12 cells were transfected with HA–H-Ras, serum-starved overnight and stimulated as indicated. After treatment with a chemical crosslinker, cells were lysed and HA–H-Ras immunoprecipitates (IP) were subjected to western blotting with the indicated antibodies. (b) ERK1 immunoprecipitates were prepared from PC12 cells treated with EGF (for 5 min) or EGF and U0126 (10 μM pre-incubation for 1 h), as indicated, and were examined for associated NF1. (c) NF1 immunoprecipitates from PC12 cells treated as above were blotted with an antibody specific to phospho-Thr-Pro (pTP). (d) PC12 cells were serum-starved overnight and stimulated with EGF or NGF. Activated, GTP-loaded Ras was precipitated with GST–RBD beads and subjected to western blotting with an anti-pan Ras antibody. The bar graph shows the mean ± s.e.m. of three independent experiments. (e) PC12 cells were transfected with control or NF1 siRNA, starved overnight and stimulated with EGF. GTP-loaded Ras was precipitated with GST–RBD beads and subjected to western blotting with a pan Ras antibody. Lysates were blotted for NF1, and total Ras and ERK were used as loading controls. (f) PC12 cells were transfected with control (−) and NF1 (+) siRNA and treated as above. Lysates were blotted for MEK and ERK phosphorylation, and total Raf-1 was used as a loading control. (g) PC12 cells were co-transfected with an eGFP expression vector plus control or NF1 siRNA. Serum-starved cells were stimulated with EGF or NGF. After 24 h, fluorescent eGFP-expressing cells with protrusions longer than two cell diameters were counted as differentiated. Error bars show mean ± s.e.m. of three independent assays.

Figure 3

Regulation of ERK localization in PC12 cells by PEA-15. (a) Top panel, cells were transfected with control and siRNAs and PEA-15 levels were examined 24 h later. Bottom panel, nuclear and cytosolic fractions were prepared from siRNA-transfected, serum-starved PC12 cells and subjected to western blotting with the indicated antibodies. To assess the quality of the fractionation, lysates were blotted for proteins with known nuclear (Lamin A) or cytosolic (MEK) localization. (b) Cells were co-transfected with an eGFP expression vector plus control or PEA-15 siRNA #3. Differentiation was assessed as in Fig. 2g. (c) Serum-starved cells were treated with 10 μM of PtdIns(3)K inhibitor LY294002 (LY) or DMSO (Control) for 1 h before stimulation with NGF for 15 min. Nuclear and cytosolic fractions were blotted with the indicated antibodies. (d) Cells were transfected with Flag–PEA-15, starved overnight and incubated with LY294002 or DMSO (Control) for 1 h before stimulation with EGF or NGF for 15 min. Lysates were blotted with the indicated antibodies. (e) Cells were transfected with Flag–PEA-15 (WT) or the Flag–PEA-15^S116A mutant and serum-starved overnight. Cells were pretreated with LY294002 (+) or DMSO (−) for 1 h and stimulated with NGF for 15 min. Flag immunoprecipitates (IPs) were blotted for associated endogenous ERK. (f) Cells were transfected with expression plasmids for eGFP, PEA-15, PEA-15^S116A or a vector control, and assessed for differentiation as in Fig. 2g. Expression of PEA-15 constructs was monitored by blotting with an anti-PEA-15 antibody. Error bars show the mean ± s.e.m. of three independent experiments. (g) Cells were transfected with eGFP, serum-starved overnight, incubated with 10 μM of DMSO or LY294002 for 1 h, and subsequently stimulated with EGF or NGF for 24 h (left panel). Differentiation was scored as described in Fig. 2g. LY294002 inhibits Akt activation, as shown by western blotting of Akt phosphorylation (right panel). LY, LY294002.

Figure 4

Differential regulation of transcriptional repressors by EGF and NGF in PC12 cells. (a) Serum-starved cells were stimulated with EGF or NGF, as indicated. Cytoplasmic and nuclear fractions were subjected to western blotting with the indicated antibodies. (b) Cells were co-transfected with eGFP and a control plasmid, wild-type ERF or a seven serine to alanine mutant (ERF ^1-7A), and examined for differentiation as in Fig. 2g. Equal expression of the ERF constructs was assessed for by western blotting. (c) Endogenous TRPS1 was immunoprecipitated from EGF- or NGF-treated PC12 cells and subjected to western blotting with the PXpSP antibody, which detects ERK phosphorylation sites, and a TRPS1 antibody was used as a loading control. (d) Cells were transfected with TRPS1–Flag, serum-starved overnight, then incubated for 1 h with DMSO or U0126 (U0) and subsequently stimulated with NGF (100 ng ml⁻¹) for 5 min. TRPS1–Flag immunoprecipitates were blotted with the indicated antibodies. (e) Cells were transfected with wild-type TRPS1–Flag or the S229A,S1097A mutant (SS-AA). Cells were starved overnight and stimulated with NGF for 15 min. TRPS1–Flag immunoprecipitates were blotted with the indicated antibodies. (f) Cells were co-transfected with expression vectors for eGFP and a control plasmid, TRPS1–Flag or the TRPS1^SS-AA mutant. Differentiation was assessed as in Fig. 2g. Error bars show the mean ± s.e.m. of four independent experiments. (g) Cells were co-transfected with reporter vectors for GATA-firefly-luciferase, Renilla-luciferase and control plasmid, and expression vectors for TRPS1–Flag or the TRPS1^SS-AA mutant and starved or treated with NGF (100 ng ml⁻¹) for 18 h. The firefly luciferase output was normalized to Renilla luciferase. Error bars show the mean ± s.e.m. of three independent experiments.

Figure 5

Kinetic mathematical model of the ERK pathway in PC12 cells. (a) Topology scheme of the model. Asterisks indicate active proteins. (b, c) Sensitivity analysis of EGF (b) and NGF (c) treatment. Raw sensitivity coefficients were calculated by making 1% changes in model parameters, and then calculating the change in either the amount of whole cell ERK* at 30 min after stimulation, or the time integral of the nuclear ERK* profile. These raw coefficients were normalized by dividing by the baseline value (no parameter perturbation), and then absolute values were taken. (d) Cooperation between NF1- and PEA-15-knockdown. Cells were co-transfected with an eGFP expression vector plus control or NF1 and PEA-15 siRNA. Serum-starved cells were treated with EGF or NGF for 24 h, and scored for differentiation as in Fig. 2g.