Research ArticleCell Biology

ERK-Mediated Phosphorylation of Fibroblast Growth Factor Receptor 1 on Ser777 Inhibits Signaling

See allHide authors and affiliations

Science Signaling  12 Feb 2013:
Vol. 6, Issue 262, pp. ra11
DOI: 10.1126/scisignal.2003087

Abstract

Fibroblast growth factor 1 (FGF1) controls cellular activities through the activation of specific cell-surface FGF receptors (FGFRs). Transphosphorylation of tyrosine residues in the kinase domain of FGFRs leads to activation of intracellular signaling cascades, including those mediated by mitogen-activated protein kinases (MAPKs). FGFRs also contain a serine-rich C-terminal tail. We identified a regulatory mechanism of FGFR signaling involving phosphorylation of Ser777 in the C-terminal region of FGFR1 by the MAPKs extracellular signal–regulated kinase 1 (ERK1) and ERK2. Prevention of the phosphorylation of Ser777 in FGFR1 or mutation of Ser777 to alanine enhanced FGF-stimulated receptor tyrosine phosphorylation and increased cell proliferation, cell migration, and axonal growth. A form of FGFR1 with a phosphomimetic mutation at Ser777 exhibited reduced signaling. Activation of MAPKs by other receptor tyrosine kinases also resulted in phosphorylation of Ser777 in FGFR1, thereby enabling crosstalk regulation of FGFR activity by other signaling pathways. Our data reveal a negative feedback mechanism that controls FGF signaling and thereby protects the cell from excessive activation of FGFR.

Introduction

Detailed understanding of the regulatory mechanisms of receptor tyrosine kinases, including fibroblast growth factor (FGF) receptors (FGFRs), is an important goal of current biomedical research because these transmembrane proteins are commonly involved in many kinds of human cancers and other pathological conditions, including skeletal and olfactory syndromes and metabolic disorders. The FGFR family consists of four closely related receptors (FGFR1 to FGFR4) (1) that play important roles in diverse physiological processes, such as cell proliferation, differentiation, migration, and survival (2). FGFRs have a conserved structure composed of three extracellular immunoglobulin (Ig)–like domains (D1 to D3), an acidic box, a transmembrane domain, and a well-conserved intracellular tyrosine kinase domain. The second and third Ig-like domains of the FGFRs are crucial for ligand binding, whereas the first Ig-like domain plays a role in receptor autoinhibition (3). Apart from the fact that one of the tyrosine residues (Tyr766 in FGFR1) in the C-terminal tail of the receptor serves as a binding site for phospholipase C γ (PLC-γ) (4), very little is known about the biological function of the disordered C-terminal region of the FGFRs, which is composed of about 50 to 56 amino acid residues.

FGFRs are activated by 18 high-affinity extracellular ligands of the FGF family. The binding of FGF to FGFR results in dimerization of the receptor, which, in turn, activates the tyrosine kinase domain of the receptor by transphosphorylation. Phosphorylation of tyrosine residues in the receptor is a precisely ordered sequential event that includes seven tyrosine residues (Tyr463, Tyr583, Tyr585, Tyr653, Tyr654, Tyr730, and Tyr766) in the case of FGFR1 (5, 6). The activated receptor then phosphorylates multiple intracellular proteins, including FGFR substrate 2α (FRS2α) and PLC-γ (7). Phosphorylated FRS2α forms two specific binding sites for Src homology 2 (SH2) domain–containing protein tyrosine phosphatase-2 (SHP2) and four binding sites for growth factor receptor–bound protein 2 (GRB2). GRB2 is constitutively bound through its SH3 domains to the protein son of sevenless (SOS) and to GRB2-associated binding protein 1 (GAB1), and these proteins constitute a signaling complex that activates the Ras–mitogen-activated protein kinase (MAPK) and phosphoinositide 3-kinase (PI3K)–Akt signaling pathways (7). Several other signaling pathways, including those mediated by the MAPKs p38 and c-Jun N-terminal kinase (JNK) and signal transducer and activator of transcription (STAT) proteins, are stimulated by activated FGFR (8, 9).

The precise regulation of FGFR signaling is critical for cell fate determination because it may influence the proliferation potential of the cell and, as a consequence, result in carcinogenesis and tumor progression. However, in contrast to the well-researched mechanisms of FGFR activation, those mechanisms ensuring receptor deactivation are only poorly understood (8). Attenuation of the signal generated by an activated FGFR can be achieved by endocytosis and lysosomal degradation of the receptor (10). In addition, a negative feedback mechanism leading to the silencing of FGF-induced signals is mediated by several proteins, such as MAPK phosphatase 3 (MKP3), Sprouty proteins, and SEF (similar expression to FGF) family members, which modulate receptor signaling at different steps of the signal transduction pathway (8, 1113). Furthermore, FGFR deactivation also depends on MAPK signaling and threonine phosphorylation of FRS2α (14). FGF stimulates the phosphorylation of at least eight threonine residues of FRS2α, which is mediated by extracellular signal–regulated kinase 1 (ERK1) and ERK2 (collectively known as ERK1/2). Phosphorylation of these threonine residues effectively reduces the extent of the tyrosine phosphorylation of FRS2α, thereby decreasing the recruitment of GRB2 and attenuating FGFR signaling (14).

Here, we found that FGF1 stimulated the direct ERK1/2-mediated phosphorylation of a specific serine residue (Ser777) in the C-terminal region of FGFR. Phosphorylation of this residue reduced the extent of tyrosine phosphorylation in the kinase domain of the receptor, thus inhibiting activation of the receptor through a negative feedback route. Our work demonstrates a previously uncharacterized mechanism for the control of FGFR1 activity, which could potentially protect cells from excessive intracellular signaling.

Results

FGFR1 is directly phosphorylated on Ser777 by ERK1 and ERK2

The C-terminal region of FGFR1, which consists of amino acid residues 755 to 822, is rich in serines (Fig. 1A); therefore, we considered the possibility that the signaling potential of FGFR might be regulated by phosphorylation of specific serines. Analysis of the amino acid sequence of FGFR1 by the Scansite server (http://scansite.mit.edu) revealed the presence of a strong consensus sequence in the C-terminal region of the receptor for the MAPK ERK1, with a predicted phosphorylation site at Ser777 and a binding site at Pro781 (15). Furthermore, Ser777 is conserved throughout the human FGFR family and also in different species, ranging from zebrafish and newt to frog, rat, mouse, and human (Fig. 1A).

Fig. 1

Phosphorylation of FGFR1 by MAPKs. (A) ClustalX2 alignment of the C-terminal regions of FGFRs from various species. Numbers refer to the amino acid numbering used for human FGFR1. Ser777 of FGFR1 and the corresponding serines of other FGFRs are indicated by asterisks. (B) In vitro phosphorylation of the C-terminal tail of recombinant FGFR1 by MAPKs. The GST-tagged C-terminal (Ct) region of wild-type FGFR1 (Ct-FGFR1-WT), FGFR1 S777A mutant (Ct-FGFR1-S777A), FGFR1 S777D mutant (Ct-FGFR1-S777D), or GST alone was subjected to phosphorylation reactions with active recombinant ERK1 or ERK2. Samples were analyzed by autoradiography (upper panel) and Coomassie blue staining (lower panel). (C) MAPK-mediated in vitro phosphorylation of full-length FGFR1 immunoprecipitated (IP) from U2OSR1, U2OSR1 S777A, or U2OSR1 S777D cells. Immunoprecipitated receptors were subjected to phosphorylation reactions with active recombinant ERK1 or ERK2. Samples were analyzed by autoradiography (upper panel) and Western blotting (IB) with anti-FGFR1 antibody (lower panel). (D) In vitro coimmunoprecipitation of ERK1 and the recombinant C-terminal tail of FGFR1. GST-tagged Ct-FGFR1-WT, its mutants, or GST alone was incubated with recombinant ERK1. Protein complexes were immunoprecipitated with an anti-GST antibody and analyzed by Western blotting. (E) Binding of recombinant ERK1 to full-length FGFR1 immunoprecipitated from U2OSR1 cells. FGFR1 immunoprecipitates were incubated with active His-tagged ERK1, and complexes were analyzed by Western blotting. Data are representative of three independent experiments.

To examine whether ERK1 and ERK2 directly phosphorylated FGFR1 at Ser777, we performed an in vitro phosphorylation assay with recombinant kinases, [γ-33P]-labeled adenosine triphosphate (ATP), and the recombinant glutathione S-transferase (GST)–tagged C-terminal tail of FGFR1 or of its mutants as substrates. Wild-type FGFR1 was phosphorylated by active forms of ERK1 and ERK2, whereas two FGFR1 proteins with mutations at the critical serine, S777A and S777D, were not phosphorylated (Fig. 1B). We used recombinant p38α MAPK, which phosphorylates FGFR1 (16), as a control. When we used the same number of kinase units of p38α as those of ERK1/2, we observed that the phosphorylation of FGFR1 by the ERKs was more substantial than that by p38α (Fig. 1B). A radiolabeled band of p38α was detected only upon long exposure (fig. S1A). As a control, we used an excess of GST alone, which was very weakly phosphorylated (Fig. 1B). In addition, we performed phosphorylation reactions with a kinase unrelated to the MAPK pathway, the serine and threonine kinase Akt, which was unable to phosphorylate the C-terminal tail of FGFR1 (fig. S1A). Because we intended to block the ERKs in cellular experiments with specific inhibitors of MEK [mitogen-activated or extracellular signal–regulated protein kinase kinase (MAPKK)], the kinase directly upstream of ERK1/2, we also investigated whether MEK phosphorylated the C-tail of FGFR1. In experiments with active MEK1, we did not observe any band that corresponded to phosphorylated FGFR1 (pFGFR1) (fig. S1B).

To confirm that the ERKs phosphorylated full-length FGFR1 and that this phosphorylation occurred specifically on Ser777, we performed in vitro phosphorylation assays with recombinant kinases and FGFR1 that was immunoprecipitated from U2OS cells stably expressing FGFR1 or its mutants. The phosphorylation reaction was performed in the presence of the specific FGFR tyrosine kinase inhibitor PD173084 to prevent receptor autophosphorylation. These experiments showed that both ERK1 and ERK2 efficiently phosphorylated wild-type FGFR1, but not FGFR1 proteins mutated at Ser777 (Fig. 1C). These results indicated that the MAPKs ERK1 and ERK2, but not MEK, directly and specifically phosphorylated FGFR1 at Ser777 and, furthermore, that Ser777 was likely the only site in FGFR1 that was phosphorylated by these MAPKs.

ERK1 binds to FGFR1

To test whether ERK1 bound directly to the C-terminal tail of FGFR1, we performed a pull-down experiment with active recombinant His-tagged ERK1, the GST-tagged recombinant C-terminal region of FGFR1 or of its mutants, and Dynabeads with immobilized anti-GST antibodies. With an anti-MAPK antibody, we found that only the GST-tagged C-terminal tail of wild-type FGFR1 was able to pull down ERK1, in contrast to the S777A and S777D mutant receptors (Fig. 1D). As controls for the unspecific binding of ERK1 to GST, we used GST alone and GST-tagged FGF1, but in neither case did we detect any bands with the anti-MAPK antibody (Fig. 1D).

We also analyzed the binding of recombinant His-tagged ERK1 to full-length FGFR1 that was immunoprecipitated from U2OSR1 cells with anti-FGFR1 antibodies conjugated to Dynabeads. With the anti-His antibody, we detected ERK1 in a complex with FGFR1 from cells (Fig. 1E). As a negative control, we used a nonspecific antibody for the immunoprecipitations. Both experiments revealed that ERK1 bound directly to FGFR1 (Fig. 1, D and E). Moreover, in addition to being the phosphorylation site for ERK1/2, Ser777 of FGFR1 was also necessary for binding of the receptor to ERK1.

Blocking MAPK activity inhibits the phosphorylation of FGFR1 Ser777

To study the involvement of ERK1/2 in the phosphorylation of FGFR1 Ser777 in cells, we generated a phospho-specific antibody against phosphorylated Ser777 (pS777-FGFR1) and used two different cell lines: U2OS cells that were stably transfected to express FGFR1 (U2OSR1 cells) and NIH3T3 cells, which have endogenous FGFR1. In experiments with specific antibodies against pS777-FGFR1, we observed that upon treatment with FGF1, Ser777 was effectively phosphorylated in cells. Moreover, when ERK1/2 were inactivated in the presence of U0126 or of SL327, another MEK1/2 inhibitor, we did not detect phosphorylation of Ser777 (Fig. 2A).

Fig. 2

FGFR1 Ser777 phosphorylation depends on the activity of MAPKs. (A) Western blotting analysis of serum-starved U2OSR1 or NIH3T3 cells that were not pretreated or were pretreated with MEK1/2 inhibitors (20 μM U0126 or 1 μM SL327) and then were stimulated with FGF1 for the indicated times. Data are representative of five independent experiments. (B) Confocal microscopy images of serum-starved U2OSR1 cells that were left untreated or were incubated with Cy3-FGF1 for 30 min before being incubated with anti–pS777-FGFR1 and anti-EEA1 antibodies. As indicated, some cells were also incubated with U0126 before the addition of FGF1 and throughout the experiment. Scale bars, 5 μm. (C) Confocal microscopy images of U2OSR1 cells that were transfected with the indicated constructs, starved of serum, and then incubated in the presence or absence of Cy3-FGF1 for 30 min before being incubated with anti–pS777-FGFR1 and anti–hemagglutinin (HA) antibodies. As indicated, some cells were also incubated with U0126. DN-MAPKK, dominant-negative MAPKK; CA-MAPKK, constitutively active MAPKK. Scale bars, 5 μm. Confocal images are representative of four independent experiments per condition.

In addition, we performed confocal microscopy analysis of U2OSR1 cells with antibody against pS777-FGFR1. The cells were starved of serum for 4 hours and then were stimulated with Cy3-labeled FGF1 (Cy3-FGF1) for 20 min. Subsequently, the cells were fixed and stained with antibody against pS777-FGFR1 and with an antibody against early endosomal antigen 1 (EEA1) to detect endosomes (17). We observed staining of pSer777 in the presence of FGF1 (Fig. 2B, upper panel), but not in unstimulated cells (Fig. 2B, middle panel). Moreover, pSer777 colocalized with both Cy3-FGF1 and the endosomal marker EEA1, indicating that the staining was specific and intracellular. We also performed the same experiment in the presence of the specific MEK inhibitor U0126, and we could not detect any pSer777 staining (Fig. 2B, lower panel). There were no substantial differences in the colocalization patterns of Cy3-FGF1 and EEA1 in the presence or absence of U0126 (Fig. 2B), indicating that the cellular localization of the FGF1-FGFR1 complex was unaffected by inhibition of MEK.

To further investigate the role of ERK1/2 in the phosphorylation of FGFR1 on Ser777, we used expression constructs (previously characterized by Pouyssegur and co-workers) that either positively or negatively regulate MAPK activities (18, 19). This group reported that a construct encoding MEK1 with mutations of its two phosphorylation sites, Ser218 and Ser222, to aspartic acid residues (MEK1-SS/DD) activates MAPKs constitutively, in contrast to the dominant-negative form of MEK1 (MEK1-S222A) in which Ser222 was mutated to alanine. Upon expression of the dominant-negative mutant of MEK1 (MEK1-S222A) and stimulation of the cells with FGF1, we could not detect pS777-FGFR1 (Fig. 2C). Conversely, when the constitutively active variant of MEK1 (MEK1-SS/DD) was expressed in U2OSR1 cells in the absence of FGF1, we observed staining of pS777-FGFR1 that was undetectable in the presence of U0126 (Fig. 2C), suggesting that MAPKs are likely to be directly involved in the serine phosphorylation of FGFR1.

The tyrosine phosphorylation of FGFR1 is enhanced by blocking the activities of ERK1 and ERK2

To examine the potential impact of ERK1/2 on FGFR1 activity, we studied the effect of ERK inhibition on FGF1-induced signaling in NIH3T3 and U2OSR1 cells. In both cell lines, the FGF1-induced tyrosine phosphorylation of FGFR1 (on Tyr653 and Tyr654) as well as the phosphorylation of downstream signaling molecules was augmented and more sustained in the presence of either U0126 or SL327 as compared to untreated cells (Fig. 3A). With antibodies against pS777-FGFR1, we observed that the kinetics of Ser777 phosphorylation was similar to that of the tyrosine phosphorylation of FGFR1. Inactivation of ERK1/2 abrogated Ser777 phosphorylation, as was seen earlier. To ensure complete blocking of ERK1/2 activity, we used 20 μM U0126 in all of our experiments, even though 5 μM U0126 was sufficient to inhibit the phosphorylation (and activation) of ERK1/2 (fig. S2). Experiments in U2OSR1 and NIH3T3 cells were performed in the presence of cycloheximide (10 μg/ml) or brefeldin A (2 μg/ml) to prevent the appearance of newly synthesized receptors. No substantial differences in signaling were observed between cells treated with either of these inhibitors (fig. S3). The MEK inhibitors U0126 and SL327, which prevent the activation of ERK1/2, also brought about a shift in the electrophoretic mobility of pFRS2α (at Tyr196) and enhanced the abundance of pFRS2α compared to that in control cells (Fig. 3A). This effect was caused by a lack of ERK-mediated phosphorylation of FRS2α on eight threonine residues, which in turn reduces the extent of Tyr196 phosphorylation (14).

Fig. 3

Effect of blocking ERK1/2 activity on FGFR1 signaling. (A) Western blotting analysis of serum-starved U2OSR1 or NIH3T3 cells pretreated with 20 μM U0126 or 1 μM SL327 and then stimulated with FGF1 in the presence of brefeldin A (2 μg/ml) for the indicated times. Quantification of bands corresponding to pFGFR1 (Tyr653/Tyr654) is presented in the graphs and expressed as a fraction of the maximal response in the absence of inhibitor. Data are means ± SD of three independent experiments. (B) Western blotting analysis of U2OSR1 cells that were transfected with GRB2-specific siRNA or a nontargeting siRNA control (scr), starved of serum, and stimulated with FGF1 for the indicated times. The efficiency of knockdown was assessed in every experiment by Western blotting analysis with an anti-GRB2 antibody. Quantification of bands corresponding to pFGFR1 (Tyr653/Tyr654) is presented in the graph and expressed as a fraction of the maximal response in cells transfected with control siRNA. Data are means ± SD of three independent experiments. *P < 0.05, **P < 0.01, statistically significant differences between treatments and control.

To test further whether inhibition of ERK1/2 was the direct cause of the enhanced and more sustained tyrosine phosphorylation of FGFR1 that we observed, we attempted to specifically knock down ERK1 and ERK2 in U2OSR1 cells with validated short interfering RNA (siRNA) oligonucleotides. Unfortunately, despite achieving an efficient reduction in the abundance of both proteins, we were left with trace amounts of ERK1/2 that were highly active upon phosphorylation by upstream kinases (fig. S4A). Therefore, we decided to use siRNA specific for GRB2, an adaptor protein recruited by activated FRS2α that is responsible for Ras activation and further activation of the MAPK pathway (20). We used a scrambled siRNA pool as a control. The GRB2-specific siRNA efficiently knocked down the abundance of GRB2, which resulted in an almost complete deactivation of ERK1/2 (Fig. 3B). As expected, depletion of GRB2 had an outcome similar to that when pharmacological inhibition of MEK was performed. We observed that the extent of tyrosine phosphorylation of FGFR1 was prolonged and slightly more intense in GRB2-deficient cells compared to that in control cells (Fig. 3B and fig. S4B). However, because of residual activity of the ERKs, the observed effects were less pronounced than those caused by MEK inhibitors. In addition, we found that when GRB2 was depleted, the extent of phosphorylation of FRS2α was reduced compared to that in cells transfected with control siRNA.

Blocking serine and threonine phosphatases reduces the extent of tyrosine phosphorylation of FGFR1

To further elucidate the role of Ser777 phosphorylation in FGFR1 signaling, we stimulated serum-starved U2OSR1 and NIH3T3 cells that had been treated overnight with 150 nM okadaic acid [a protein serine and threonine phosphatase (1 and 2A) inhibitor] with FGF1. In U2OSR1 cells, the activation and tyrosine phosphorylation of FGFR1, as detected with an anti-pFGFR (Tyr653/Tyr654) antibody, in response to FGF1 were substantially reduced when Ser777 of FGFR1 was already phosphorylated because of phosphatase inhibition (Fig. 4A). In the case of NIH3T3 cells, we detected almost no tyrosine phosphorylation and, as a consequence, almost no cellular response upon pretreatment of the cells with okadaic acid (Fig. 4A). Together, these findings suggest that the phosphorylation status of Ser777 influences the extent of cellular response to FGF1.

Fig. 4

Interplay between Ser777 phosphorylation and tyrosine phosphorylation of FGFR1. (A) Effect of blocking serine and threonine phosphatases on the tyrosine phosphorylation status of FGFR1. Western blotting analysis of serum-starved U2OSR1 and NIH3T3 cells that were untreated or were incubated overnight with 150 nM okadaic acid before being stimulated with FGF1 in the presence of cycloheximide (10 μg/ml) for the indicated times. Quantification of bands corresponding to pFGFR1 (Tyr653/Tyr654) is presented in the graphs and is expressed as a fraction of the maximal response in the absence of okadaic acid. Data are means ± SD of three independent experiments. ***P < 0.001, statistically significant differences between cells treated with okadaic acid and control cells. (B) Effect of mutation of Ser777 on the tyrosine phosphorylation of FGFR1. Western blotting analysis of serum-starved U2OSR1, U2OSR1 S777A, or U2OSR1 S777D cells that were not pretreated or were pretreated with 20 μM U0126 before being stimulated with FGF1 in the presence of brefeldin A (2 μg/ml) for the indicated times. Quantification of the intensities of the bands corresponding to pFGFR1 (Tyr653/Tyr654) that were divided by the intensities of the bands corresponding to total FGFR1 protein is presented in the graphs and expressed as a fraction of the maximal response in WT U2OSR1 cells. Data are means ± SD of three independent experiments. Asterisks indicate statistically significant differences between U2OSR1 cells in the absence of U0126 (left panel) or U2OSR1 cells in the presence of U0126 (right panel) and all others: *P < 0.05, **P < 0.01.

FGF1-induced tyrosine phosphorylation of FGFR1 is prolonged in U2OS cells expressing the FGFR1 S777A mutant

We showed that when Ser777 phosphorylation was reduced or blocked by treatment with different MEK inhibitors, the extent of tyrosine phosphorylation of FGFR1 stimulated by FGF1 was enhanced and prolonged compared to that in control cells. Conversely, inhibition of serine and threonine phosphatases by okadaic acid and the subsequent enhancement of Ser777 phosphorylation resulted in decreased FGF1-dependent tyrosine phosphorylation of FGFR1. To clarify the relationship between Ser777 and tyrosine phosphorylation, we generated two stable U2OS cell lines expressing either the S777A or S777D mutants of FGFR1 (U2OSR1 S777A and U2OSR1 S777D, respectively). The S777A mutant FGFR1 cannot be phosphorylated, whereas the S777D variant FGFR1 mimics the constitutively phosphorylated receptor. With these cell lines, it was possible for us to compare the biological responses of FGFR1 at different phosphorylation states in the same cellular system and under similar conditions.

We found that in U2OSR1 S777A cells, the duration of FGF1-induced tyrosine phosphorylation of FGFR was prolonged compared to that in control U2OSR1 cells (Fig. 4B). Conversely, in U2OSR1 S777D cells, the signal from the activated receptor was weaker and lasted for a shorter time (Fig. 4B). As expected, the MEK inhibitor U0126 did not substantially increase the extent of tyrosine phosphorylation of FGFR1, neither in U2OSR1 S777A nor in U2OSR1 S777D cells. However, the tyrosine phosphorylation of FRS2α was increased and prolonged in the presence of U0126 regardless of the identity of the amino acid residue at position 777 of FGFR1 (Fig. 4B). These results suggest that the phosphorylation status of Ser777 determines the extent and duration of the tyrosine phosphorylation of FGFR1.

U2OSR1 S777A cells exhibit an enhanced mitogenic response and increased cell migration

As observed earlier, we found substantial differences in FGF1-induced signaling between the phosphorylation-incompetent and phosphomimetic Ser777 mutants of FGFR1. To investigate the biological consequences of Ser777 phosphorylation, we studied the ability of U2OS cells stably expressing either the S777A or S777D variants of FGFR1 (in comparison to control U2OSR1 cells) to proliferate in response to FGF1. We found that cells with the S777A mutant proliferated more efficiently than did cells with the wild-type receptor, whereas the proliferation potential of cells with the S777D variant was substantially reduced (Fig. 5A). Similar results were obtained by analyzing additional clones stably expressing wild-type FGFR1 or the S777A or S777D variants (fig. S5A). These data suggest that the inhibition of receptor tyrosine phosphorylation caused by phosphorylation of FGFR1 Ser777 is a crucial element in controlling FGF1 signaling and cell proliferation.

Fig. 5

Effects of blocking the phosphorylation of FGFR1 Ser777 on cellular functions. (A) FGF1-induced proliferation of U2OSR1, U2OSR1 S777A, and U2OSR1 S777D cells. AlamarBlue reagent was added to serum-starved cells that were stimulated with FGF1 for the indicated times, and fluorescence corresponding to the number of cells was measured. Data are means ± SEM of three independent experiments. (B) Effect of mutation of Ser777 on cell migration. FGF1-stimulated U2OS, U2OSR1, U2OSR1 S777A, and U2OSR1 S777D cells were subjected to time-lapse, live-cell imaging. Values are given as means ± SEM of three independent experiments in each of which 30 to 50 cells were analyzed. (C) Effect of mutation of Ser777 on axon growth and neuronal morphology of adult DRG neurons. Total axonal length, maximal distance of the longest axon, and the number of branch points per cell were measured in FGF1-stimulated DRG neurons cotransfected with plasmid encoding EGFP and plasmids expressing the indicated proteins. Data are means ± SEM of three independent experiments with a total number of >46 neurons analyzed per group. Representative images of the neuronal morphology of adult DRG neurons before and after treatment with FGF1 for 24 hours are shown. *P < 0.05, **P < 0.01, ***P < 0.001.

To test another functional consequence of the phosphorylation status of FGFR1 Ser777, we examined the motility of untransfected U2OS cells and that of U2OS cells stably expressing wild-type FGFR1 or its mutants. Time-lapse, live-cell imaging upon stimulation with FGF1 revealed that U2OS cells expressing wild-type FGFR1 (U2OSR1 cells) migrated faster than did untransfected U2OS cells, indicating the role of FGFR1 in cell migration (Fig. 5B). The velocity of U2OSR1 S777A cells was substantially increased compared to that of U2OSR1 cells, which was in contrast to that of U2OSR1 S777D cells, which migrated more slowly than did U2OSR1 cells (Fig. 5B). Because we obtained similar results when we investigated independent cell clones stably expressing FGFR1 mutants, we concluded that the results were not cell clone–dependent (fig. S5B). The observed effect of the phosphomimetic mutation demonstrates that phosphorylation on Ser777 of FGFR1 is important in switching off multiple signal transduction pathways activated by FGF1.

The S777A mutation enhances axonal growth in dorsal root ganglion neurons

FGF signaling has been implicated in several processes during development and regeneration of the nervous system, such as neural induction, patterning, axon guidance, and synapse formation (21). To further elucidate the physiological relevance of the phosphorylation of Ser777, we examined axonal growth and branching in adult dorsal root ganglion (DRG) neurons overexpressing wild-type FGFR1 or the S777A or S777D mutants. DRG neuron culture is a well-characterized system in which to investigate the mechanisms of neuritogenesis, and it has previously been reported that FGFR1 signaling in DRG neurons enhances elongative axon growth, but not branching (22). We found that the total axonal length of adult DRG neurons and the maximal distance of the longest axon (a parameter for elongative axon growth) were substantially increased in neurons overexpressing the S777A mutant compared to those of neurons expressing wild-type FGFR1 or the S777D mutant (Fig. 5C). In contrast, the number of branch points was not enhanced by the S777A variant (Fig. 5C). Thus, the elongating properties of FGFR1 S777A in neurons provide evidence for an enhanced regenerative capacity of adult neurons that is directed by the phosphorylation status of Ser777. These findings are in accordance with previous results demonstrating that enhanced FGFR1 signaling leads to increased axon growth (22, 23).

Epidermal growth factor or serum stimulates FGFR1 Ser777 phosphorylation

Because other stimuli can activate MAPKs in the absence of FGF-induced signaling, we tested whether FGFR1 could be phosphorylated on Ser777 in the absence of its activation by FGF1. We stimulated cells with epidermal growth factor (EGF) or serum because both stimuli activate MAPKs in an FGF1-independent manner and they do not stimulate the tyrosine phosphorylation of either FGFR1 or FRS2α. Through Western blotting analysis, we found that both EGF and serum efficiently stimulated the phosphorylation of FGFR1 Ser777 in U2OSR1 cells and NIH3T3 cells (Fig. 6, A and B, lane 2). Addition of the MEK inhibitor U0126 abrogated FGFR1 Ser777 phosphorylation (Fig. 6, A and B). We also pretreated both cell lines with serum or EGF for 30 min before stimulating them with FGF1. When used, the MEK inhibitor U0126 or the FGFR inhibitor PD173074 was added simultaneously with EGF or serum and kept throughout the experiment. Phosphorylation of Ser777 stimulated by EGF or serum was again completely blocked in the presence of U0126 (Fig. 6, A and B). We also observed a strong shift in the electrophoretic mobility of FRS2α under such conditions, as previously described by Lax et al. (14). Treatment of cells with the FGFR inhibitor PD173074 in the absence of EGF or serum inhibited Ser777 phosphorylation (Fig. 6, A and B); however, when the cells were stimulated with EGF or serum, we detected a band corresponding to pS777-FGFR1 even in the presence of PD173074. Although we detected marked ERK activation in response to EGF, we could not detect any phosphorylated EGF receptor (EGFR) in U2OSR1 cells by Western blotting analysis. Therefore, we decided to repeat the experiment in U2OSR1 cells that were transiently transfected with plasmid encoding the EGFR (Fig. 6A, U2OSR1 cells, lanes 7 to 12). Again, we observed substantial FGFR1 Ser777 phosphorylation in response to EGF (Fig. 6A, U2OSR1 cells, lane 8), but we did not detect bands that corresponded to pS777-FGFR1 in the presence of the MEK inhibitor when cells were pretreated with EGF and then stimulated with FGF1 (Fig. 6A, U2OSR1 cells, lane 11). In addition, we analyzed the effect of EGF on FGFR1 Ser777 phosphorylation by confocal microscopy in serum-starved U2OSR1 cells. Similarly to the Western blotting data, we observed staining of pS777-FGFR1 in cells stimulated with EGF in the absence of FGF1 (Fig. 6C).

Fig. 6

EGF and serum stimulate the phosphorylation of FGFR1 Ser777. (A and B) Western blotting analysis of serum-starved U2OSR1 cells, U2OSR1 cells transiently transfected with plasmid expressing EGFR, and NIH3T3 cells that were stimulated with (A) FGF1 or EGF or (B) FGF1 or serum for 15 min. Where both stimuli are indicated, cells were pretreated with serum or EGF for 30 min and then subsequently treated with FGF1 for 15 min. U0126 (20 μM) and PD173074 (100 nM) were added as indicated. Data are representative of two independent experiments. (C) Confocal microscopy images of serum-starved U2OSR1 cells stimulated with Cy3-FGF1 or EGF for 30 min. Scale bars, 5 μm. Confocal images are representative of four independent experiments per condition.

We also monitored the kinetics of FGFR activation (through its tyrosine phosphorylation) and downstream signaling in response to FGF1 after pretreatment with EGF. Under these conditions, FGFR1 Ser777 was already phosphorylated before FGF1 was added to the cells. We found that cells pretreated with EGF before the addition of FGF1 responded more weakly than did cells stimulated with FGF1 alone: tyrosine phosphorylation of FGFR1 and phosphorylation of its downstream signaling molecules were reduced and less prolonged (Fig. 7A). We obtained similar results by confocal microscopy analysis. Serum-starved U2OSR1 cells were treated either with a combination of EGF and FGF1 or with FGF1 alone after a 30-min pretreatment with EGF, and then the cells were stained with anti-pFGFR (Tyr653/Tyr654) antibody. Consistent with the Western blotting data, we observed a substantial decrease (~25%) in the abundance of pFGFR in response to FGF1 upon pretreatment of the cells with EGF (Fig. 7B). These experiments demonstrated that EGF-induced stimulation of MAPKs leads to the phosphorylation of FGFR1 on Ser777, which results in reduced FGFR1 activation upon treatment with FGF1. On the basis of these findings, we suggest that both FGF1-dependent and FGF1-independent MAPK activation inhibit FGFR1 signaling through Ser777 phosphorylation of the receptor.

Fig. 7

Pretreatment with EGF reduces subsequent cell responses to FGF1. (A) Western blotting analysis of serum-starved U2OSR1 and NIH3T3 cells pretreated for 30 min with EGF and then stimulated with FGF1 for the indicated times. Quantification of the intensities of the bands corresponding to pFGFR1 (Tyr653/Tyr654), FRS2α, and ERK1/2 is presented in the bar graphs and is expressed as the fraction of the maximal response in cells that were not treated with EGF. Data are means ± range of two independent experiments. (B) Confocal microscopy images of serum-starved U2OSR1 cells incubated with either EGF or FGF1 alone or after a 30-min pretreatment with EGF before being incubated with an anti–pTyr-FGFR antibody. Scale bars, 5 μm. Quantification of pTyr-FGFR staining is presented in the graph. Eighty to 100 cells were quantified per experiment for each condition. Data are means ± SEM of three independent experiments. ***P < 0.001.

The number of serine residues correlates positively with organism complexity

Because the unordered C-terminal region of FGFR1 is very rich in serines, they are likely to be of physiological importance (24, 25). Tan et al. showed that the number of genomically encoded tyrosine residues (and thus of potentially phosphorylated tyrosines) decreases with increasing species complexity in metazoans (26). Therefore, we calculated the numbers of tyrosine, serine, and threonine residues in the cytoplasmic region of FGFR1 in different species. We observed that the number of tyrosines decreases with the number of different cell types, in contrast to serines (as well as serines and threonines combined), whose number correlates positively with the complexity of the organism (fig. S6 and Fig. 8A). This indicates an important role of serines (and threonines) in more complex organisms, probably through regulating signaling.

Fig. 8

Mechanisms of ERK-mediated regulation of FGFR signaling and evolutionary analysis of the serine content in the C-terminal tail of FGFR1. (A) Correlation between the expansion in the number of serines in the cytoplasmic region of FGFR1 and organism complexity. The number of different cell types correlates positively and statistically significantly (R2 = 0.817) with the extent of the serine content of FGFR1. (B) Schematic representation of negative feedback attenuating FGFR1 signaling. FGF1-induced tyrosine phosphorylation of FGFR1 leads to activation of FRS2α, which is followed by GRB2- and SOS-mediated activation of Ras and ERK1/2. ERK1/2 can also be activated by different pathways independent of FGFR, for example, in response to serum or EGF. Activated ERKs phosphorylate FRS2α on threonines (14) and FGFR1 on Ser777, which results in reduced tyrosine phosphorylation of both FRS2α and FGFR1 and consequent attenuation of FGFR signaling.

Discussion

Enhanced FGFR signaling is of critical importance in many human cancers (2); therefore, it is essential to understand the system that modulates receptor activity. Here, we described a previously uncharacterized mechanism for the regulation of FGFR1 signaling on the basis of the activity of the MAPKs ERK1 and ERK2 (Fig. 8B). Because the C-terminal regions of the FGFRs are very rich in serine residues of unknown function, we analyzed potential phosphorylation sites in this region and found that Ser777 in the C-terminal tail of FGFR1 is a specific phosphorylation site targeted by ERK1 and ERK2. We also showed that this phosphorylation event takes place in cells in response to FGF1-dependent and FGF1-independent activation of ERKs. Furthermore, the phosphorylation status of FGFR1 Ser777 influenced the tyrosine phosphorylation (and activation) of the receptor and its downstream signaling. We found that phosphorylation of Ser777 decreased FGFR1 activity and its ability to propagate mitogenic signals. Analyzing the migratory response, we observed a correlation between the phosphorylation status of FGFR1 Ser777 and the ability of cells to migrate, which could be important in many types of cancer. In addition, we showed that the phosphorylation status of Ser777 influenced long-distance axon extension, which is required during nerve regeneration. Our results indicate that the lack of Ser777 phosphorylation of FGFR1 has biological implications.

It was previously shown that the MAPK pathway can function as a negative feedback amplifier (27). Several feedback loops for ERKs involving Raf, SOS, and FRS2α have been implicated in the attenuation of FGF-induced signaling. Inhibition of FRS2α activity is achieved by its phosphorylation by ERKs at eight threonine residues in response to different stimuli. A mutant FRS2α deficient in ERK-mediated phosphorylation displays increased tyrosine phosphorylation (14). Here, we showed that the ERKs, in a similar manner, inhibited FGFR1 tyrosine phosphorylation and activity through direct phosphorylation of Ser777 of FGFR1. It is not clear how the phosphorylation of threonines in FRS2α inhibits its ability to be tyrosine phosphorylated. It is possible that upon its threonine phosphorylation, FRS2α becomes a poor substrate for the receptor kinase or that its associations with subcellular compartments or components are altered. Similar mechanisms might apply for the inhibition of FGFR1. We suggest that the phosphorylated Ser777 could act as a binding site for tyrosine phosphatases responsible for receptor inactivation; alternatively, ERK-mediated phosphorylation might cause local conformational changes within the receptor that make FGFR1 a better substrate for tyrosine dephosphorylation. It is also possible that the phosphate group could disrupt interactions between FGFR1 and its binding partners, resulting in attenuation of receptor signaling.

Because ERKs can play a dual role in FGFR1 signal transduction by inducing both mitogenic signaling and negative regulation, it seems that their spatial and temporal dynamics might be crucial for FGF1-induced biological activities. We observed that the maximal stimulation of ERK1/2 was reached by 5 min after stimulation with FGF1, whereas Ser777 phosphorylation attained its maximal extent ~10 min later. This suggests that ERKs are therefore early signaling molecules that activate their downstream targets before the activity of FGFR1 is inhibited by pSer777.

An alternatively spliced variant of FGFR2 (FGFR2 IIIb), which has a shortened C-terminal tail, is found in certain human cancers. Cha et al. found that deletion of the C-terminal sequence of FGFR2 IIIb contributed to the transforming activity of the receptor (28). They suggested a dual mechanism of enhanced transformation caused by aberrant receptor recycling and persistent FRS2-dependent signaling. Furthermore, they correlated this effect with deletion of Tyr770 and Leu773 in the C-terminal tail of FGFR2. However, it is possible that the transforming effect of truncated FGFR2 IIIb could be due to the deletion of Ser781, which corresponds to Ser777 in FGFR1, which could result in evasion from the MAPK-mediated negative feedback mechanism that we have described here.

We showed that even in the absence of FGF1, ERKs activated by EGF, serum, or a constitutively active variant of MEK1 phosphorylated Ser777 of FGFR1. Pretreatment of cells with EGF, which led to phosphorylation of Ser777, reduced the subsequent response of the cells to FGF1. These data confirm the multifarious interplay of different signaling pathways acting through the regulatory mechanism of FGFR1 signaling that we described. Direct serine phosphorylation of FGFR1 controls the extent of tyrosine phosphorylation within the kinase domain of the receptor and thereby its downstream signaling events and biological functions. This mechanism operates in addition to other negative feedback mechanisms that control FGFR1 activity and ensure specific and precise signaling. The phosphorylation of Ser777 in FGFR1 after activation of ERKs protects the cell not only from ligand-independent receptor activation but also from simultaneous activation by several different stimuli. It is possible that this type of regulation also applies to other signaling receptors and functions as a mechanism to control inappropriate signaling upon exposure to multiple growth factors. Previous phosphorylation of Ser777 in the absence of an FGFR1 ligand desensitizes the cells and therefore regulates the magnitude of the response to subsequent FGF. Moreover, in the absence of an FGFR1 ligand, phosphorylation of Ser777 by other stimuli may protect the cell from inadvertent activation of FGFR1 signaling by ligand-independent dimerization of receptor molecules (14, 29). This mechanism might also be relevant in the context of cancer, in which FGFRs are overactive and pSer777 may play a tumor-suppressive role.

Indeed, FGFR1 is not the only receptor that is phosphorylated by MAPKs. ERKs interact with and phosphorylate other transmembrane proteins, including EGFR and the netrin receptor deleted in colorectal carcinoma (DCC) (30, 31). The cytoplasmic domain of DCC that is phosphorylated by ERK2 is unstructured, similarly to the C-terminal tail of FGFR1. The exact role of DCC phosphorylation by MAPKs is unknown; nevertheless, the link between ERK signaling and DCC signaling suggests that ERK-mediated serine and threonine modification is of physiological importance (31).

We suggest that Ser777 and probably other serines in the C-terminal tail of FGFR1 act as molecular switches and modulate the primary response upon receptor stimulation. That the number of tyrosine residues decreases with organism complexity [as described by Tan et al. (26)] suggests that this may be a result of evolutionary selection to eliminate phosphorylation events that could lead to uncontrolled or unspecific signaling. We observed a negative correlation between the number of cell types and the number of tyrosines in the cytoplasmic region of FGFR1, in contrast to a positive correlation in the case of number of serines. The reason for this could be that conserved phosphorylated tyrosine sites tend to be located in ordered protein domains, in contrast to phosphorylated serines, which are located in disordered regions (such as the C terminus of FGFR1), which evolve rapidly (25, 26). The increased number of serines and the combined number of serines and threonines in FGFR1s of higher organisms suggest that, simultaneously with the gain of new cell types, additional mechanisms of signaling control were developed.

In summary, our study reveals a previously uncharacterized regulatory mechanism of FGFR1 signaling that involves ERK-dependent phosphorylation of Ser777 in the receptor. We found that the C-terminal tail of FGFR1 was a direct substrate for the MAPKs ERK1 and ERK2, and that these kinases regulated their own signaling by decreasing the activity of the receptor. This negative feedback loop based on Ser777 phosphorylation of FGFR1 is part of a complex system controlling proliferation signals, which in excess may lead to damaging events, including the development of cancer. A better understanding of the regulatory mechanisms of FGFR1, which are often disrupted in oncogenic malignancies, should ultimately lead to more efficient drug discovery and design.

Materials and Methods

Antibodies and reagents

The following primary antibodies were used: rabbit anti-MAPK (ERK1/2, p44/p42), mouse anti-pMAPK (ERK1/2, p44/p42) (Thr202/Tyr204), rabbit anti–pAkt (Ser473), rabbit anti-FGFR1, mouse anti–pFGFR (Tyr653/Tyr654), rabbit anti–pFRS2α (Tyr196), and rabbit anti-GRB2 were obtained from Cell Signaling Technology; rabbit anti–pPLC-γ (Tyr783), rabbit anti-FRS2α, mouse anti-GST, rabbit anti-GST, and rabbit anti-FLAG were purchased from Santa Cruz Biotechnology; mouse anti–pp38 MAPK (Thr180/Tyr182), mouse anti-HSP90, mouse anti-EEA1, and mouse anti–EGFR (activated form) were from BD Transduction Laboratories; mouse anti–FLAG M2 was from Sigma-Aldrich; mouse anti-Myc was obtained from Upstate Biotechnology; and mouse anti-HA.11 and mouse anti–6-His were from Nordic Biosite. Specific anti–pFGFR1 (Ser777) (pS777-FGFR1) antibody was made by GenScript with the following phosphorylated peptide: CSMPLDQYpSPSFPDTR. The antibody was purified with the phosphorylated peptide and by cross-adsorption to the corresponding nonphosphorylated peptide. Horseradish peroxidase–conjugated and fluorescently labeled secondary antibodies were from Jackson ImmunoResearch Laboratories. Heparin–Sepharose CL-6B affinity resin was from Amersham. Mowiol, brefeldin A, okadaic acid, PD173074, and the MEK1/2 inhibitor SL327 were from Calbiochem. Cycloheximide, heparin, and U0126 were from Sigma-Aldrich. Protease inhibitor cocktail tablets (EDTA-free, complete) were obtained from Roche Diagnostics. FGF1 was labeled with Cy3-maleimide (GE Healthcare) according to the manufacturer’s procedures. Hoechst 33342, AlamarBlue, Dynabeads anti-mouse IgG, and Dynabeads Protein G were purchased from Invitrogen. All other chemicals were obtained from Sigma-Aldrich.

Cell lines and bacterial strains

NIH3T3 cells were grown in Quantum 333 medium (PAA Laboratories) supplemented with 2% bovine serum (Gibco), penicillin (100 U/ml), and streptomycin (100 μg/ml). DOTAP liposomal transfection reagent (Roche Diagnostics) was used according to the manufacturer’s protocol to obtain U2OS cells stably expressing either the S777A or S777D mutants of FGFR1. Clones were selected with geneticin (G-418; 1 mg/ml) (Invitrogen). Clones presented here were chosen on the basis of their receptor abundance as analyzed by immunofluorescence and Western blotting. U2OS cells stably expressing FGFR1 were described previously (10). These cells were propagated in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, penicillin (100 U/ml), and streptomycin (100 μg/ml). In addition, geneticin (0.2 to 1 mg/ml) was added to the growth medium of stably transfected U2OS cells. Adult rat DRGs were dissected, collected in ice-cold RPMI medium supplemented with antibiotic-antimycotic [penicillin (100 U/ml), streptomycin (100 μg/ml), amphotericin B (0.25 μg/ml); Gibco], and cultured as described previously (22). DRG cultures were maintained in TNB 100 medium (Biochrom) supplemented with protein-lipid complex (Biochrom) and antibiotic-antimycotic. For expression of FGF1 and GST–C-tail FGFR1 constructs, the Escherichia coli strain Bl21(DE3)pLysS from New England Biolabs was used.

Plasmids

The plasmids pcDNA3-FGFR1, pcDNA3–FGFR1 S777A, and pcDNA3–FGFR1 S777D have been described previously (16, 32). A construct encoding the 68 C-terminal amino acid residues of FGFR1 (residues 755 to 822) in the plasmid pGEX was used for expression of the C-terminal tail of FGFR1 fused to the C terminus of GST (16). The FGF1 construct consisted of a truncated form (residues 21 to 154) of human FGF1 in the pET-3c vector (33). pECE-HA-MEK1-SS/DD and pECE-HA-MEK1-S222A plasmids were a gift from J. Pouyssegur (Institute of Developmental Biology and Cancer Research, University of Nice) (18, 19). The pEGFR plasmid was provided by H. Stenmark (Institute for Cancer Research, The Norwegian Radium Hospital) and was originally a gift from A. Sorkin (Department of Pharmacology, School of Medicine, University of Colorado, Denver).

siRNA oligonucleotides and cell transfection

siRNA oligos specific for GRB2 (targeting sequence: 5-CAUGUUUCCCCGCAAUUAUTT-3) were purchased from Dharmacon RNA Technologies. siRNAs specific for ERK1 (p44, targeting sequence: 5-CTCCCTGACCCGTCTAATATA-3) and ERK2 (p42, targeting sequence: 5-AATGACATTATTCGAGCACCA-3) were purchased from Qiagen. ON-TARGETplus siCONTROL siRNA was purchased from Dharmacon RNA Technologies. siRNA knockdown was performed with the DharmaFECT transfection reagent according to the manufacturer’s protocol. Experiments were performed 72 hours after transfection. Transient expression of the different constructs was performed by transfecting cells with plasmid DNA with FuGENE 6 transfection reagent (Roche Diagnostics) according to the manufacturer’s protocol. Softened DRGs were transfected by electroporation with the Amaxa nucleofector (Lonza) using program O-003 and the Rat Neuron Nucleofector solution (Lonza).

Recombinant proteins

The active human recombinant MAPKs ERK1 (p44) and ERK2 (p42) were purchased from Calbiochem. Active human MEK1 was obtained from Chemicon International, whereas inactive MEK1 was from Upstate. Recombinant human p38α MAPK was from R&D Systems. Recombinant FGF1 was produced in E. coli as described previously (33). Recombinant EGF and MBP (myelin basic protein) were obtained from Sigma-Aldrich. Recombinant fusion proteins of the C-terminal region of FGFR1 with GST (GST–C-tail FGFR1) and of its mutants were produced in E. coli and purified with glutathione-Sepharose (Amersham Biosciences) according to standard procedures.

Analysis of signaling cascades

Serum-starved cells were stimulated with FGF1 (20 ng/ml) in the presence of heparin (10 U/ml) or EGF (20 ng/ml) or with 10% serum in the presence or absence of the indicated inhibitors for different times. Inhibitors were added 15 min before stimulation and were kept throughout the experiments, which were performed in the presence or absence of cycloheximide (10 μg/ml) or brefeldin A (2 μg/ml). The cells were lysed with SDS sample buffer, scraped, and sonicated. Total cell lysates were resolved by SDS–polyacrylamide gel electrophoresis (SDS-PAGE), transferred onto Immobilon-P membranes, and subjected to Western blotting analysis. After analysis of the first set of proteins, the membranes were stripped and incubated with antibodies against different targets. ImageQuant version 5 was used to quantify the intensities of the bands of interest. In the case of quantification of the intensity of bands corresponding to pFGFR1 from U2OSR1, U2OSR1 S777A, and U2OSR1 S777D cells, the intensities of the bands corresponding to pFGFR1 were divided by the intensities of the bands corresponding to total FGFR1 and then were normalized to the maximal response of U2OSR1 wild-type cells.

In vitro phosphorylation of the recombinant C-tail of FGFR1

In vitro phosphorylation experiments were performed with recombinant proteins. One microgram of a fusion protein was incubated with kinases and [γ-33P]ATP (40 μCi/ml) in reaction buffer [25 mM Hepes (pH 7.5), 20 mM MgCl2, 1 mM Na2MO3, 20 mM sodium β-glycerophosphate, 1 mM dithiothreitol, 5 mM EGTA] at 30°C for 30 min. As a control, 2 μg of GST was used. The reaction was stopped by precipitation with trichloroacetic acid for 30 min on ice. The samples were then centrifuged, washed twice with cold acetone, and dissolved in sample buffer. The proteins were analyzed by SDS-PAGE, Western blotting, and autoradiography, and then the blot was stained with Coomassie blue.

In vitro coimmunoprecipitation of ERK1 with the recombinant C-tail of FGFR1

Human recombinant His-tagged ERK1 (100 ng) was incubated for 2 hours at 4°C with 500 ng of GST-tagged recombinant fusion proteins of the C-terminal tail of FGFR1 or of its mutants or with GST alone or with GST-tagged FGF1 (controls) in phosphate-buffered saline (PBS) supplemented with 0.1% (w/v) bovine serum albumin (BSA) and protease inhibitors. Complexes were then pelleted with mouse anti-GST antibodies immobilized on Dynabeads anti-mouse IgG. The Dynabeads were washed six times with binding buffer, and then proteins were eluted in sample buffer and analyzed by SDS-PAGE and Western blotting.

FGFR1 immunoprecipitation

FGFR1 was immunoprecipitated from lysates of U2OS cells stably expressing FGFR1 or its mutants with anti-FGFR1 antibodies immobilized to Dynabeads Protein G. The immunoprecipitates were washed three times with lysis buffer [100 mM NaCl, 10 mM tris-HCl (pH 7.4), 5 mM EDTA, 1% Triton X-100] and once with PBS supplemented with 0.1% (w/v) Tween (washing buffer). To analyze the binding of ERK1 to full-length FGFR1, FGFR1 immunoprecipitates were incubated with 50 ng of active recombinant His-tagged ERK1 at 4°C for 90 min. Complexes were then washed with washing buffer, eluted in samples buffer, and analyzed by SDS-PAGE and Western blotting. To test the phosphorylation of full-length FGFR1, FGFR1 immunoprecipitates were washed with high-salt buffer (washing buffer with 1 M NaCl) and incubated for 15 min with 100 nM PD17034 before performing the in vitro phosphorylation assay as described earlier for the recombinant C-terminal tail of FGFR1.

Laser scanning confocal microscopy

Cells grown on coverslips were starved for 4 hours and then incubated for 30 min at 37°C with Cy3-FGF1 (100 ng/ml) in the presence of heparin (20 U/ml) or with EGF (100 ng/ml). The cells were fixed in 4% formaldehyde solution and permeabilized with 0.1% Triton X-100. The cells were then incubated with primary antibodies for 20 min, washed, and then incubated with secondary antibodies coupled to a fluorophore for 20 min before mounting in Mowiol. The cells were examined with a Zeiss LSM 510 META confocal microscope (Carl Zeiss). Images were prepared with Zeiss LSM Image Browser version 3.2 (Carl Zeiss) and CorelDRAW11. Quantifications were performed with ImageJ software.

Proliferation assay

Serum-starved U2OS cells stably expressing FGFR1 or its mutants were treated with FGF1 (100 ng/ml) in the presence of heparin (10 U/ml) for 48, 72, and 96 hours. At each time point, AlamarBlue reagent was added to cells, and 3 hours later, the fluorescence of the reduced form of the dye was measured with an EnVision multimode plate reader (PerkinElmer). The fluorescence signal reflecting the number of cells in each experiment was normalized to the number of U2OSR1 cells at time point 0.

Time-lapse live-cell imaging and cell migration tracking

Serum-starved U2OS cells and U2OS cells stably expressing FGFR1 or its mutants were plated on Hi-Q4 culture dishes (ibidi Integrated BioDiagnostics) and observed with a BioStation IM Live Cell Recorder (Nikon Instruments Inc.) with a 20× phase contrast objective at 37°C and 5% CO2 in humidified air. In all experiments, cells were stimulated with Cy3-labeled FGF1 (200 ng/ml) and heparin (20 U/ml). Image acquisition was performed every 10 min over a period of 8 hours. Images were analyzed with ImageJ software with Manual Tracking and Chemotaxis and Migration Tool (ibidi GmbH) plugins. The velocity of cells was calculated for the whole time frame.

Analysis of axon growth

For measurement of axonal growth, DRG neuron cultures transfected with pcDNA3-FGFR1, pcDNA3–FGFR1 S777A, or pcDNA3–FGFR1 S777D together with plasmid encoding EGFP were analyzed by inverted fluorescence microscopy with a Zeiss Axiovert 100 microscope equipped with a SPOT RT digital camera. Transfected neurons were analyzed 48 and 72 hours after transfection before and after a 24-hour treatment with FGF1 (100 ng/ml) in the presence of heparin (10 U/ml). MetaMorph (Visitron Systems) morphometry software was applied to measure the maximal distance of the longest axon, the total axonal length, and the number of branch points. All morphologically intact neurons in each dish that had a maximal distance of the longest axon of ≥100 μm were analyzed.

Analysis of the relationship between organism complexity and the expansion in the numbers of tyrosines, serines, and threonines in the cytoplasmic region of FGFR1

The numbers of serines, threonines, and tyrosines within the sequence of the cytoplasmic region of FGFR1 were counted in several representative species of metazoans. Squares of Pearson correlation coefficients (R2) were calculated for the relation between the numbers of the amino acids and organism complexity [measured as the number of cell types (26)].

Statistical analysis

For statistical analysis, one-way analysis of variance (ANOVA) with Tukey’s posttest was applied. P < 0.05 was considered statistically significant.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/6/262/ra11/DC1

Fig. S1. In vitro phosphorylation assay with the recombinant C-terminal tail of FGFR1 and Akt or MEK1.

Fig. S2. Effect of U0126 on ERK1/2 activity.

Fig. S3. Effect of MEK inhibitors on FGFR1 activity in the presence or absence of brefeldin A or cycloheximide.

Fig. S4. The effect of siRNA-mediated knockdown of ERK1/2 and GRB2 on FGFR1 activity.

Fig. S5. Effect of the phosphorylation status of FGFR1 Ser777 on cell proliferation and migration in additional clones of stably transfected U2OS cells.

Fig. S6. Correlation between organism complexity and tyrosine, serine, and threonine contents in the cytoplasmic region of FGFR1 proteins.

References and Notes

Acknowledgments: We thank H. Stenmark for critical reading of the manuscript and A. Engen, A.-M. G. Pedersen, A. G. Bergersen, and A. Kubiak for skillful technical assistance. Funding: M.Z. was a fellow of The Research Council of Norway and E.M.H. was a fellow of The Norwegian Cancer Society. J.W. was supported by the South-Eastern Norway Regional Health Authority Fellowships. This work was supported by the Polish Ministry of Science and Higher Education (grant 0627/IP1/2011/71). Author contributions: M.Z. designed the study and performed in vitro phosphorylation experiments, all of the signaling experiments, and the cell proliferation study; E.M.H. and J.W. performed laser scanning confocal microscopy; B.N.-W. conducted immunoprecipitation and in vitro kinase assays; A.O. performed cell migration experiments; B.H. performed the analysis of axonal growth; E.M.H. and Y.J. prepared the stable cell lines; A.W. contributed to the design of the study and supervised the project; M.Z., A.W., J.W., and J.O. contributed to data analysis; M.Z. wrote the manuscript; and A.W., J.W., E.M.H., A.O., and J.O. edited the manuscript. Competing interests: The authors declare that they have no competing interests.
View Abstract

Stay Connected to Science Signaling

Navigate This Article