Mammalian sprouty proteins inhibit cell growth and differentiation by preventing ras activation.

Sprouty was genetically identified as an antagonist of fibroblast growth factor signaling during tracheal branching in Drosophila. In this study, we provide a functional characterization of mammalian Sprouty1 and Sprouty2. Sprouty1 and Sprouty2 inhibited events downstream of multiple receptor tyrosine kinases and regulated both cell proliferation and differentiation. Using NIH3T3 cell lines conditionally expressing Sprouty1 or Sprouty2, we found that these proteins specifically inhibit the Ras/Raf/MAP kinase pathway by preventing Ras activation. In contrast, activation of the phosphatidylinositol 3-kinase pathway was not affected by Sprouty1 or Sprouty2. We further showed that Sprouty1 and Sprouty2 do no prevent the formation of a SNT.Grb2.Sos complex upon fibroblast growth factor stimulation, yet block Ras activation. Taken together, these results establish mammalian Sprouty proteins as important negative regulators of growth factor signaling and suggest that Sprouty proteins act downstream of the Grb2.Sos complex to selectively uncouple growth factor signals from Ras activation and the MAP Kinase pathway.

Sprouty was genetically identified as an antagonist of fibroblast growth factor signaling during tracheal branching in Drosophila. In this study, we provide a functional characterization of mammalian Sprouty1 and Sprouty2. Sprouty1 and Sprouty2 inhibited events downstream of multiple receptor tyrosine kinases and regulated both cell proliferation and differentiation. Using NIH3T3 cell lines conditionally expressing Sprouty1 or Sprouty2, we found that these proteins specifically inhibit the Ras/Raf/MAP kinase pathway by preventing Ras activation. In contrast, activation of the phosphatidylinositol 3-kinase pathway was not affected by Sprouty1 or Sprouty2. We further showed that Sprouty1 and Sprouty2 do no prevent the formation of a SNT⅐Grb2⅐Sos complex upon fibroblast growth factor stimulation, yet block Ras activation. Taken together, these results establish mammalian Sprouty proteins as important negative regulators of growth factor signaling and suggest that Sprouty proteins act downstream of the Grb2⅐Sos complex to selectively uncouple growth factor signals from Ras activation and the MAP Kinase pathway.
Normal development requires precise spatial and temporal regulation of signal transduction pathways involved in cell growth and differentiation. Negative control of growth factor response is achieved both by restriction of the incoming signal itself and induction of counter regulatory mechanisms affecting the propagation of the signal. The expression of many inhibitors are induced by the pathway they eventually antagonize, providing the potential for a tight autoregulation (for a review, see Ref. 1). Recently, sprouty (spry) was identified by genetic studies as such an inhibitor (2).
Spry was originally described as an antagonist of Breathless FGF 1 receptor signaling during tracheal branching in Drosophila Loss of function mutations of spry led to excessive FGF signaling and ectopic branching, whereas engineered overex-pression of spry blocked the branching (2). As other groups reported genetic interactions between spry and several different receptor tyrosine kinases (RTK) in multiple contexts, it became clear that spry was a general inhibitor of RTK signaling during Drosophila development (3)(4)(5)(6). Through a data base search, three human genes were identified with sequence similarity to Drosophila spry (2) and a fourth family member was described in the mouse (7). Mammalian spry genes are expressed in highly restricted patterns in the embryo during early development and in many adult tissues (7)(8)(9). In most tissues, the different family members appear to be co-regulated and their expression shows a close correlation with known sites of FGF signaling. Mammalian Spry proteins may be key regulators of several developmental processes, including lung branching morphogenesis, midbrain and anterior hindbrain patterning, and limb chondrocyte differentiation (8 -10).
Genetic and biochemical analysis performed by Casci et al. (3) suggested that Drosophila Spry negatively regulates the Ras pathway, but the molecular mechanism of this inhibitory activity was not determined (5). All Spry proteins share a unique, highly conserved, cysteine-rich C-terminal domain. This domain was shown to be necessary for the membrane translocation of Spry by a yet unknown mechanism (2,3,11). The N-terminal portion of the Spry proteins is less conserved as it exhibits only 25-37% identity among the different mouse family members. These sequence differences could be responsible for functional divergence among the Spry proteins. In particular, the size difference between Drosophila and much smaller mammalian Spry N-terminal regions is intriguing.
Our laboratory has studied the Wilms Tumor 1 (WT1) gene, a tumor suppressor gene involved in embryonic kidney development, for several years. We performed a representative difference analysis screen to isolate transcriptional target genes of WT1. One of the genes identified was mouse spry1. 2 To model the role of mammalian Spry during development and tumorigenesis, we established stable inducible Spry1 and Spry2 NIH3T3 cell lines. We demonstrated that Spry1 and Spry2 antagonized growth factor signaling by specifically inhibiting the Ras/Raf/MAP kinase pathway. We methodically examined the inhibitory effect of Spry on the different components of the signal transduction cascade and identified the activation of Ras as the target of Spry activity. We showed that Spry1 and Spry2 can inhibit both proliferation of NIH3T3 cells and differentiation of PC12 cells. These results suggest that Spry proteins, by limiting RTK signaling, play an important role in development and growth control.
Cell Culture, Growth Factors, and Transfection Methods-NIH3T3 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% calf serum and transfected with LipofectAMINE Plus (Life Technologies). PC12 cells were grown on tissue culture dishes coated with poly-L-lysine (0.001%, Sigma) in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and 10% horse serum. Tet-off NIH3T3 cells (gift of I. Gelman (18)) were cultured in histidinedeficient Dulbecco's modified Eagle's high glucose completed with 0.5 mM L-histidinol (Sigma), 4 mM sodium bicarbonate, 2 mM L-glutamine, 10% calf serum, and 0.5 g/ml tetracycline. Spry1 and Spry2 Tet-off NIH3T3 cell lines were established by transfection of the Tet-off cells with pTRE-Spry1 or pTRE-Spry2 or an empty vector and selection in 0.5 g/ml G418-sulfate (Roche Molecular Biochemicals). 293T cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and were transfected with Superfect (Qiagen Inc.). All growth factors (mouse NGF 2.5 S, recombinant human PDGF-BB, recombinant murine tumor necrosis factor-␣, recombinant human epidermal growth factor, and recombinant human bFGF) were from Life Technologies.
Cell Growth and Apoptosis Assays-10 6 NIH3T3 cells (100 mm plates) were transfected by pCEV29, pCEV29-Spry1, or pCEV29-Spry2. After 3 days, cells were diluted (1/25) and selected with G418 (0.5 g/ml) for 2 to 3 weeks. Colonies were stained with Giemsa and counted on 5 plates for each vector. To measure DNA synthesis, control or Spry1/2 inducible cells (70,000 cell/well of 24-well plates) were starved for 24 h (0.3% CS with tetracycline) before stimulation with 10% CS for 24 h in the presence or absence of tetracycline. During the last 4 h, cells were labeled with 1 Ci/ml [methyl-3 H]thymidine (PerkinElmer Life Sciences). After 2 phosphate-buffered saline washes, cells were fixed with methanol and lysed with 0.25% SDS, 0.25 M NaOH. The lysates were neutralized with 0.2 M HCl and mixed with scintillation liquid (Scintiverse TM , Fisher). [ 3 H]Thymidine incorporation was measured using a Beckman LS 6500 scintillation counter. Apoptosis was examined with fluorescence microscopy using annexin V-fluorescein isothiocyanate labeling (Annexin V-FLUOS, Roche) in NIH3T3 cells transfected for 24 h with control, spry1, spry2, or bax expression vectors (1 g) along with a vector expressing a red fluorescent protein (100 ng, DsRed, CLONTECH).
PC12 Differentiation Assays-Proliferating PC12 cells were transfected with bicistronic IRES-GFP plasmids using LipofectAMINE 2000 Reagent (Life Technologies). After 48 h, differentiation was induced by dilution of the cells (1/15) and treatment with NGF (50 ng/ml) or bFGF (20 ng/ml) in Dulbecco's modified Eagle's medium containing 1% horse serum, 2 mM L-glutamine. After 3 days, cells were observed with phasecontrast microscopy to check the general efficiency of the differentiation. The differentiation of the transfected cells was examined with fluorescence microscopy to visualize GFP expression. The efficiency of differentiation of the transfected cells was quantified by counting the cells exhibiting neurite outgrowth (twice the diameter of the cell) among the GFP positive cells.
Raf-1 Kinase Assay-The assay was performed with a Raf-1 Immunoprecipitation Kinase Assay kit purchased from Upstate Biotechnology as indicated by the manufacturer.
GST-Raf/Ras Binding Assay-Raf-1, RBD, and agarose was purchased from Upstate Biotechnology and the assay was performed as recommended with 0.5-1 mg of cellular lysate and 10 l of GST-Raf-1 RBD, agarose conjugate (10 g) at 4°C for 30 min. The proteins bound were subjected to SDS-PAGE and immunoblot analysis as described above.
Co-Immunoprecipitations-Cells were lysed (50 mM Tris, pH 7.5, 1% Nonidet P-40, 150 mM NaCl, 5 mM EDTA, and one tablet of Complete TM protease inhibitors (Roche Molecular Biochemicals) for 50 ml) and precleared by centrifugation. Lysates (0.5-2 mg) were incubated with 1-2 g of the precipitating antibody overnight at 4°C with gentle rocking. During the last hour, 50 l of Protein G-or Protein A-agarose beads (50% slurry, Roche Molecular Biochemicals) were added. The beads were collected by centrifugation, washed 3 times with 1 ml of lysis buffer, and boiled in 50 l of Laemmli sample buffer. The immunoprecipitates were fractionated by SDS-PAGE and analyzed by immunoblot as described above.

Endogenous Spry Expression Is Regulated by Growth Factors in NIH3T3
Cells-To characterize mammalian Spry, we chose NIH3T3 cells which have been extensively utilized as a model for cell proliferation, oncogenesis, and growth factor signaling. We first examined the expression of spry1 and spry2 in proliferating NIH3T3 cells (Fig. 1A). Northern blot analysis revealed a major transcript of about 2.5 kilobases for both genes but endogenous spry2 expression was always significantly higher than spry1 expression, which was barely detectable in NIH3T3 cells. We next stimulated NIH3T3 cells with FGF after serum starvation and examined the spry1 and spry2 expression patterns. As presented in Fig. 1B, FGF clearly stimulated spry2 expression. This 3-fold up-regulation was transient (1 to 5 h), occurred without de novo protein synthesis (data not shown) and was maximal 2 h after treatment. Similar results were obtained with PDGF (data not shown). Surprisingly, in the same cells, expression of spry1 was down-regulated (about 50%) by FGF and PDGF ( Fig. 1B and data not shown). These results show that endogenous spry1 and spry2 are differentially regulated by growth factors in NIH3T3, suggesting that they may play specific roles in these cells. In addition, as endogenous spry2 and spry1 expression are rapidly modified upon growth factor treatment, NIH3T3 cells appear to constitute an appropriate model to study the effect of Spry1 and Spry2 on growth factor signaling.
Spry1 and Spry2 Can Inhibit Proliferation of NIH3T3 Cells-As growth factors stimulate proliferation of fibroblasts, we examined the effect of spry overexpression on cell growth using a colony suppression assay. NIH3T3 cells transfected by spry1 or spry2 formed about 50% fewer colonies upon selection than the cells transfected with a control vector (Fig. 2, A and  B). This indicates that Spry1 and Spry2 inhibit either cell growth or induce apoptosis. Using annexin V labeling (early marker for apoptosis (19)), we did not detect significant levels of apoptosis in NIH3T3 cells transiently transfected with spry1 or spry2 (Fig. 2C). To determine how Spry1 and Spry2 inhibit cell growth, we established NIH3T3 stable cell lines in which expression of mouse spry1 or spry2 is induced by the removal of tetracycline from the culture medium (Fig. 2D). Using [ 3 H]thymidine incorporation assays, we compared the DNA synthesis induced by serum in the absence or presence of Spry1 or Spry2. Fig. 2E shows that Spry1, as well as Spry2, markedly reduced (50 to 80%) the level of DNA synthesis induced by serum. Together, these data indicate that expression of Spry proteins reduces cell growth by limiting DNA synthesis, and not by inducing cell death.
Spry1 and Spry2 Can Inhibit Differentiation of PC12 Cells-We next examined the ability of Spry1 and Spry2 to inhibit differentiation induced by growth factors. We chose as a model the PC12 pheochromocytoma cells, in which neurite outgrowth can be induced by NGF or FGF via the Ras pathway (20 -22). Proliferating PC12 cells were transiently transfected with spry1 or spry2 bicistronic GFP vectors, which co-express Spry1 or Spry2 and the GFP and thus allowed the identification of the transfected cells. Two days after transfection, cells were treated by NGF or FGF to trigger differentiation. After 3 days, most of the cells had developed neurites (Fig. 3A). When the transfected cells were specifically visualized by immunofluorescence, we saw a striking difference between the control and the Spry1 or Spry2 expressing cells (Fig. 3B). Indeed, at least 80% of the cells expressing Spry1 or Spry2 were unable to differentiate upon FGF or NGF treatment compared with the control transfected cells (Fig. 3C). The neurites present in spry-transfected cells were shorter and showed reduced branching compared with the control cells. This shows that Spry1 and Spry2 can inhibit differentiation induced by growth factors. However, endogenous expression of neither spry1 nor spry2 could be detected in PC12 cells by Northern blot, even upon FGF, NGF, or epidermal growth factor treatment (data not shown). This suggests that Spry1 and Spry2 are not usually involved in the differentiation or the growth of PC12 cells yet the Spry proteins could affect the consequences of signaling through the NGF/FGF receptors. It is possible that other members of the spry family may be up-regulated by growth factors and control this neuronal differentiation. Nevertheless, this result led us to focus on the NIH3T3 cells for further experiments.
Spry1 and Spry2 Inhibit Nuclear Targets of Growth Factor Signaling-Since we had shown that Spry1 and Spry2 could reduce proliferation of NIH3T3 cells, we used this cell line to determine the molecular mechanism by which Spry proteins exert their inhibitory activity. We first examined their effect on nuclear targets of growth factor signaling. NIH3T3 cells were transfected with a serum response element-luciferase reporter gene which is as a model for growth factor regulation of transcription (reviewed in Refs. 23 and 24). Fig. 4A shows that the expression of the reporter gene is stimulated upon growth factor treatment but in the presence of Spry proteins, this stimulation is blunted, especially in the case of Spry2. When this assay was repeated with a NF-B response element-luciferase reporter gene, which can be stimulated by growth factor treatment through the Ras/Raf/MAP kinase pathway (25), expression was also inhibited by Spry proteins (Fig. 4B). In contrast, the stimulation of this reporter gene by tumor necrosis factor-␣ treatment, which is mediated by the TRAF/TRAD pathway, independently of Ras (26), was not significantly affected by the Spry proteins. Fig. 4C indicates that the spry constructs were indeed expressed in the conditions used for the reporter assays.
To reinforce the biological significance of these results, we examined the expression of the endogenous c-fos gene in the Spry1 and Spry2 inducible cell lines. c-fos is an immediate early gene whose expression is induced by growth factors and is necessary for progression through the G 1 phase of the cell cycle and subsequent cell proliferation (23,24). As expected, c-fos expression was rapidly induced upon serum or FGF treatment (Fig. 4D). However, in the presence of Spry1 or Spry2, c-fos expression was significantly blocked. Compared with the control NIH3T3 Tet-off cell line, c-fos expression stimulated by serum was reduced by about 30%. More strikingly, c-fos expression induced by FGF was decreased by 70 (Spry1) to 80% (Spry2) (Fig. 4D). Thus, this set of experiments shows that mammalian Spry1 and Spry2 are able to inhibit the transcriptional events mediated by growth factor signaling and the induction of a gene required for DNA synthesis and cell division.
Spry1 and Spry2 Specifically Inhibit the Erk1/2 MAP Kinase Pathway-We used the Spry inducible cell lines to examine the effect of Spry1 and Spry2 on the Ras/Raf/MAP kinase and the phosphatidylinositol 3-kinase (PI 3-kinase) pathways, which are two major pathways mediating growth factor signaling in NIH3T3 cells (for review, see Ref. 27). In the presence of Spry1 or Spry2, FGF or PDGF-mediated activation of the Erk1/2 MAP kinases, as visualized by phospho-specific antibodies, was strikingly inhibited (Fig. 5, A and B). No such effect was found upon tetracycline withdrawal in the control cell line (Fig. 5, A  and B). A time course of FGF stimulation showed that in the presence of Spry2 or Spry1, the stimulation of Erk1/2 was also delayed, with the peak of phosphorylated MAP kinases accumulation occurring at 10 min rather than 3 min after stimula- FIG. 1. Expression of spry1 and spry2 in NIH3T3 cells. The expression of spry1 and spry2 was examined by Northern blot. The blots were successively hybridized with the indicated probes. Quantification was performed by PhosphorImager analysis and similar results were obtained in three independent experiments. A, spry1 and spry2 expression in nonconfluent, proliferating NIH3T3 cells (10% CS). Both Spry1 and Spry2 blots were exposed for 4 days at Ϫ80°C with a Kodak BioMax TransScreen (Kodak, Rochester, NJ). B, spry1 and spry2 expression in bFGF-treated NIH3T3 cells. The cells were serum starved (0.2% CS) for 20 h and then stimulated with 20 ng/ml bFGF (in 0.2% CS) for the indicated time before lysis and RNA extraction. The Spry2 blot was exposed for 1 day and the Spry1 blot for 4 days to allow easier comparison of the expression profiles. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. tion ( Fig. 5C and data not shown for Spry1). Furthermore, MAP kinase activation was less sustained.
In contrast, PDGF-mediated activation of the key protein of the PI 3-kinase pathway, the serine/threonine kinase Akt, was not inhibited in the presence of Spry proteins (Fig. 5D). Moreover, activation of a downstream target of Akt, the serine/ threonine Gsk-3␤, was also not affected (data not shown). A time course of PDGF stimulation indicated that Spry proteins did not inhibit Akt activation at any specific time point (Fig.  5E). These results show that mammalian Spry1 and Spry2 do not inhibit the PI 3-kinase pathway but specifically inhibit the FIG. 2. Spry1 and Spry2 can inhibit proliferation of NIH3T3 cells and block DNA synthesis. A and B, colony suppression assay. NIH3T3 cells were transfected by an empty expression vector (pCEV29), pCEV29-Spry1 or pCEV29-Spry2 for 3 days before dilution and selection for 2 weeks in G418. Colonies were stained with Giemsa and counted. Representative plates are shown (A) and the data are summarized in the table (B). The number of colonies obtained in three independent experiments was averaged and compared. The results are also presented as average ratio between the number of colonies obtained with pCEV29-Spry1 or pCEV29-Spry2 and the number of colonies obtained with pCEV29, this being set at 100%. The S.D. represents the standard deviation of the ratio observed between three independent experiments. C, apoptosis was detected by annexin V-fluorescein isothiocyanate labeling in NIH3T3 cells transfected by bax but not by spry1 or spry2. Transfected cells were identified by co-transfection of a vector encoding a red fluorescent protein. D, expression of Spry1 and Spry2 upon tetracycline removal in the NIH3T3 Tet-off inducible cells was analyzed by immunoblot. The cells were serum starved (0.2% CS for 24 h with tetracycline) before stimulation with 10% CS in the presence (ϩTet) or absence (ϪTet) of tetracycline for 24 h. The anti-Flag tag antibody was used to visualize Spry1 and Spry2. E, [ 3 H]thymidine incorporation assay. Triplicate plates of control (Ct), Spry1 or Spry2 NIH3T3 inducible cells were treated as described in D and [ 3 H]thymidine was added during the last 4 h of serum stimulation. A representative experiment is shown but equivalent results were obtained in independent experiments and with different Spry clones (data not shown). To facilitate comparison, the average serum stimulation (the ratio between the amount of [ 3 H]thymidine incorporated with 10% versus 0.2% serum) in the presence of tetracycline was set at 100% for each clone. The average serum stimulation in the absence of tetracycline was calculated and compared with the one obtained in the presence of tetracycline. Error bars correspond to the standard deviation (in %) observed among the triplicates. Erk1/2 MAP kinase pathway. They also imply that in the presence of Spry proteins, some signals downstream of the RTK can be propagated.

Spry1 and Spry2
Prevent Ras Activation-We next determined the effect of Spry proteins on the stepwise activation of the different components of the MAP kinase pathway upon growth factor treatment. We examined the activation of the dual-specificity MAP kinase kinases Mek1/2 which phosphorylate Erk1/2, stimulating their activity in response to growth factor treatment. As seen in Fig. 6A, Mek1/2 phosphorylation The membranes were successively incubated with the indicated antibodies. Each experiment was repeated at least two times and intensities of the signal were measured using NIH Image software. A-C, the blot was incubated with an antibody directed against phosphorylated Erk1/2 (P-Erk1/2), an antibody directed against Erk1/2 (Erk1/2) and an antibody directed against the Flag tag to check Spry1 or Spry2 expression upon tetracycline removal (Spry1/2). D and E, the membrane was incubated with an antibody directed against phosphorylated Akt (Ser 473 , P-Akt), an antibody directed against Akt (Akt) and an antibody directed against the Flag tag (Spry1/2). Phosphorylation of Erk1/2 was also checked in the same extracts and was found to be decreased upon Spry expression. was dramatically reduced in the presence of Spry1 or Spry2, indicating that the signal was blocked upstream of Mek1/2, possibly at the level of the MAP kinase kinase kinase Raf-1. Therefore, we directly examined Raf-1 kinase activity. Fig. 6B shows that in the presence of Spry2, the ability of immunoprecipitated Raf-1 from FGF-stimulated cells to phosphorylate a recombinant GST-Mek1 protein was greatly reduced. This indicates that Raf-1 is inactive in the presence of Spry proteins and suggests that Spry act upstream of Raf-1.
The steps leading to Raf-1 activation are not fully understood but require binding to Ras (for a review, see Ref. 27). Therefore, we examined the binding of Raf-1 to Ras in the Spry cell lines (28). Fig. 6C shows that the amount of endogenous Ha-Ras bound by recombinant GST-Raf-1 Ras-binding domain (RBD) fusion protein upon FGF treatment was greatly reduced in the presence of Spry1 or Spry2. This could be due either to a direct inhibition of the Ras/Raf-1 interaction or to an inhibition of Ras activation, since only activated, GTP-bound Ras, can bind Raf-1. To discriminate between these two possibilities, we examined the binding of a constitutively active form of Ras (Ha-Ras R12) to Raf-1 in the presence of Spry1 or Spry2. We performed the same GST-Raf-1/Ras binding assay with lysates of NIH3T3 cells transfected with Ras and Spry1 or Spry2 expression vectors. As presented in Fig. 6D, expression of Spry1 or Spry2 reduced the binding of wild type Ha-Ras (induced by FGF) to GST-Raf-1, but had no significant effect on the binding of the constitutively active Ha-Ras R12. This suggests that Spry proteins do not interfere directly with the binding of Ras to Raf-1, but rather inhibit the activation of Ras.
Finally, we examined the phosphorylation of the adaptor FRS2/SNT-1, which is the primary substrate of the activated FGF receptor (29,30) and thus reflects the activity of the RTK. Fig. 6E graphically shows that the ability of the FGF receptor to mediate phosphorylation of FRS2/SNT-1 after FGF stimulation was not inhibited in the presence of Spry proteins while at the same time, the activation of the Erk1/2 MAP kinases was almost completely blocked. This indicates that the ability of the FGF receptor to phosphorylate a downstream substrate and transduce the growth factor signal is intact and that Spry proteins act downstream of the receptor to inhibit signaling.
Spry Proteins Do Not Prevent the Recruitment of Sos to the RTK Signaling Complex-Since Spry expression did not inhibit the ability of a RTK to generate a signal, we hypothesized that the ability of Spry to inhibit Ras activation would involve its binding to one, or several, more downstream component(s) of the Ras pathway at the inner surface of the plasma membrane. We attempted to co-immunoprecipitate Spry1 or Spry2 with several putative partners from lysates of co-transfected FIG. 6. Spry1 and Spry2 act upstream of Ras. A-C and E, control (Ct), Spry1 or Spry2 NIH3T3 inducible cells were serum starved (0.3% CS for 24 h) in the presence (ϩTet) or absence (ϪTet) of tetracycline before stimulation with bFGF (20 ng/ml in A and 50 ng/ml in B, C, and E). After stimulation (15 min in A and 2 min in B, C, and E), the cells were lysed and proteins extracted. Phosphorylation of Erk1/2 was always checked and found to be decreased upon Spry expression. Each experiment was repeated at least twice with similar results being obtained and intensities of the signal were measured using NIH Image software. A, proteins were analyzed by successive immunoblotting with an antibody directed against phosphorylated Mek1/2 (Ser 217-221 , P-Mek1/2), an antibody directed against Mek1/2 (Mek1/2), and an antibody directed against the Flag tag to check Spry1/2 expression (Spry1/2). B, cell extracts were immunoprecipitated with an antibody directed against Raf-1 and the precipitates were incubated with recombinant GST-Mek1 and GST-Erk2 in the presence of [␥-32 P]ATP. Phosphorylated GST-Mek1 and GST-Erk2 were visualized by a combination of SDS-PAGE and autoradiography (kinase Assay, 3/4 of the reaction). An aliquot of the kinase reaction (one-fourth) was analyzed by successive immunoblotting with an antibody recognizing Raf-1 and an antibody recognizing GST. Expression of Spry1 and Spry2 in the whole cell lysates was checked with an antibody directed against the Flag tag (Spry1/2). C, cells were lysed, proteins extracted, and subjected to a GST-Raf-1 RBD binding assay. Bound proteins were analyzed by successive immunoblotting with an antibody directed against Ha-Ras (Ras) and an antibody directed against GST (GST-Raf-1). In parallel, the whole cell lysates were immunoblotted with an antibody directed against Ras and an antibody directed against the Flag tag (Spry1/2). D, NIH3T3 cells were transfected with the indicated plasmids (Ras plasmids, 1 g; Spry plasmids, 3 g). 24 h before lysis, cells were serum starved in 0.2% CS. Cells transfected with wild-type Ras were stimulated with 20 ng/ml bFGF (left, lanes 2-4) 15 min before lysis. Protein extracts were subjected to a GST-Raf-1 RBD binding assay followed by Western blot analysis as described in C. Transfected wild-type Ras was detected using an antibody directed against the AU5 tag and transfected Ras R12 was detected with an antibody directed against the Myc tag. E, proteins were incubated with an FRS2/SNT-1 antibody and the immunoprecipitates were analyzed by immunoblotting using an antibody directed against phosphotyrosine (P-Tyr). MAP kinase activation in the cell extracts was also examined by immunoblotting with an antibody directed against phosphorylated Erk1/2 (P-Erk1/2). NIH3T3 cells or from the Spry inducible cell lines. We were not able to find any physical interaction between Spry1 or Spry2 and endogenous Ha-Ras, Raf-1, and Ras GAP (31) (data not shown). We also failed to find any interaction between recombinant GST-Ras or GST-Raf-1 proteins and Spry proteins from the inducible cell lines (data not shown). Finally, we tested Grb2, the adaptor protein which allows the recruitment of the GDP/GTP exchange factor Sos upon stimulation (32). As shown of Fig. 7A, the Grb2 and Spry2 proteins were readily co-immunoprecipitated from lysates of co-transfected NIH3T3 cells. Surprisingly, in a simultaneous experiment under the same conditions, we could not co-immunoprecipitate the Grb2 and Spry1 proteins (Fig. 7B).
Since Spry2 co-immunoprecipitated with Grb2, we asked whether Spry proteins could inhibit the formation of Grb2⅐Sos complexes and/or the recruitment of Sos to the activated receptor. Under our conditions Grb2⅐Sos complexes were readily detected after FGF stimulation (Fig. 7C, bottom) and such complexes were unaffected by the presence of Spry proteins, despite the fact that Spry2 (but not Spry1) could be found in Sos immunoprecipitates (Fig. 7C, middle right). In addition, a 90-kDa tyrosine-phosphorylated protein, previously documented by several groups to be the membrane-anchored adaptor FRS2/SNT-1 (29,30,33), was also co-immunoprecipitated with Sos (Fig. 7C, top). Thus, Spry proteins do not inhibit the formation of the SNT⅐Grb2⅐Sos complexes. This implies that signaling from the FGF receptor to Sos and the recruitment of Sos to the membrane is not blocked by Spry proteins. On close examination of Fig. 7C (middle), a subtle but reproducible change in Sos was noted. Upon FGF stimulation, Sos migrated through the gel at a slightly reduced mobility consistent with a previously described hyperphosphorylation (32). This alteration in Sos mobility was almost completely abolished in the presence of Spry1 and was reduced in the presence of Spry2. This may be another reflection of the loss of Erk1/2 MAP kinases or other kinase activities downstream of Ras in the presence of Spry. DISCUSSION In this study, we methodically examined the consequences of Spry expression on signaling by the RTK using a biochemical approach. We showed that mammalian Spry proteins like Drosophila Spry, are able to antagonize a wide range of RTK. The fact that Spry1 and Spry2 were able to inhibit both proliferation of NIH3T3 cells and differentiation of PC12 cells indicates that the inhibitory effect of mammalian Spry proteins could affect cell fate depending on the cellular context. Thus, similarly to the factors it antagonizes, Spry activity does not seem to be specific to a biological response, but rather depends on the cellular environment. Interestingly, Spry proteins affect not only the strength of the signal but also its duration. This is of importance, since the duration of Erk1/2 activation was shown to be critical for cell fate determination, such as PC12 differentiation (34).
Growth factors were proposed to regulate Spry activity at different levels. First, growth factors can stimulate the relocalization of Spry proteins from the cytoplasm to the inner membrane of the cell (11). Second, spry expression is up-regulated by the pathway that it antagonizes, yielding a negative feedback loop. In agreement with this model, growth factor treatment of NIH3T3 cells induced an immediate up-regulation of spry2 expression. Unexpectedly, a down-regulation of spry1 expression, parallel to the up-regulation of spry2 expression, was observed upon FGF or PDGF treatment. Similar results were recently obtained in endothelial cells (35). These data indicate that the different Spry genes are not uniformly regulated and suggest that the different family members may not always be functionally equivalent. It also corroborated the fact that in some tissues the expression patterns of the Spry family members do not overlap (8). 2 Thus, the individual Spry genes may be regulated by specific combinations of factors to allow optimal control of signaling. The regulatory logic of spry1 down-regulation in NIH3T3 cells is currently unclear, as our assays did not allow us to see any consistent differences be- FIG. 7. Interaction of Spry proteins with Grb2 and Sos. A and B, NIH3T3 cells were transfected for a total of 48 h with the indicated plasmids. After lysis, proteins were incubated with an antibody directed against the Myc tag and the immunoprecipitates were analyzed by Western blot (IP Myc). An antibody directed against the Myc tag was used to detect Grb2 (Grb2) and an antibody directed against the Flag tag to detect Spry1 or Spry2. Expression from the vectors transfected was checked with the same antibodies in whole cell extracts (Lysates). Experiments were performed in parallel and repeated three times. C, control (Ct), Spry1 or Spry2 NIH3T3 inducible cells were serum starved (0.3% CS for 24 h) in the presence (ϩTet) or absence (ϪTet) of tetracycline before 3 min stimulation with 50 ng/ml bFGF (ϩ). After stimulation, the cells were lysed and proteins extracted. Proteins were incubated with a Sos antibody and the immunoprecipitates (IP Sos) were analyzed by immunoblotting using an antibody directed against phosphotyrosine (P-Tyr, which detected FRS2/SNT at 90 kDa), Sos, Flag (to detect Spry proteins), or Grb2. In parallel, the whole cell lysates (Lys) were immunoblotted with the same set of antibodies. Phosphorylation of Erk1/2 was also checked in the same extracts and was found to be decreased upon Spry expression (data not shown). tween Spry1 and Spry2 activities. This may reflect a functional redundancy due to overexpression, or could be real as Spry1 and Spry2 are the most similar members of the family. Therefore, it would be interesting to examine the activities of Spry3 and Spry4, as these proteins are more divergent in their Nterminal domains.
The fact that mammalian and Drosophila Spry inhibit signaling through multiple receptors suggested that intracellular machinery common to signaling by the RTK is being affected. In the presence of Spry proteins, transcription of c-fos in response to FGF or serum was decreased. Induction by FGF was more dramatically affected. It should be noted that growth factors induction of c-fos expression is mainly mediated by the Ras pathway, whereas serum induction is in part independent of Ras (reviewed in Ref. 24). Transient co-transfection assays performed with a NF-B reporter gene showed that Spry inhibited the stimulation by FGF or PDGF, but not by tumor necrosis factor-␣, also indicating the specificity of Spry proteins toward the Ras pathway. However, Spry proteins do not globally affect RTK signaling, as we showed that they inhibit the Ras/Raf/MAP kinase pathway but not the PI 3-kinase pathway. Since many cell survival signals through RTK are mediated by the PI 3-kinase pathway (reviewed in Ref. 36), its maintenance correlates with the fact that Spry suppressed cell growth, but did not lead to an increase in apoptosis. The induction of Spry in response to RTK activation might be a way of selectively altering the read out of the growth factor signal and might be adaptive during inductive interactions in development. The inhibition of the MAP kinase pathway and maintenance of other pathways including that of PI 3-kinase might allow the target cells to cease proliferation, yet maintain their viability.
Genetic analysis and in vitro assays suggested that Drosophila Spry was an inhibitor of the Ras pathway but the molecular mechanism of the inhibitory effect remained controversial (3,5). Our finding that the phosphorylation of Akt was still induced by PDGF in the presence of Spry implied that binding of the ligand to the RTK still allowed the generation of second messengers and that the kinase activity of the receptor remained intact. This notion was confirmed by the fact that Spry expression did not affect the ability of the FGF receptor to mediate phosphorylation of one of its major substrates, FRS2/ SNT-1. In contrast, we showed that Spry1 and Spry2 inhibited the activation of Ras, Raf-1, Mek1/2, and Erk1/2. The binding of Raf-1 to wild type, but not constitutively activated Ras, was reduced upon Spry expression. Furthermore, Spry proteins could not be shown to physically interact with Ras or Raf-1. Therefore, we propose that Spry proteins prevent the formation, or maintenance, of activated GTP-Ras. Since the kinase activity of RTK is not affected by the presence of Spry, our data indicate that Spry uncouples RTK from the chain of events that lead to the activation of Ras. Our model supports one of the genetic analysis performed in Drosophila (3) which showed that halving the dose of spry increased signaling by an activated RTK but not an activated Ras and thus indicates the functional conservation of the Spry family of proteins.
The exact mechanism of Spry action at the RTK/Ras interface needs to be clarified. Our experiments restrict Spry activity to a very narrow window. We found that Spry2 could form a complex in vivo with the adapter Grb2, which is also a property of the ancestral Drosophila protein (3), and Sos. This potentially could have prevented the coupling of the receptor to Sos. However, we showed that Spry2 did not impair the formation of the SNT⅐Grb2⅐Sos complex upon FGF stimulation. In addition, Spry1 did not immunoprecipitate with Grb2 or Sos. Together, this argues that Spry/Grb2 interaction is neither conserved, nor essential for Spry inhibitory effect. Our results indicate that Spry proteins do not block Ras activation by preventing Sos recruitment to the activated receptor as previously proposed (3). Therefore, Spry proteins potentially act downstream of the recruitment of the SNT⅐Grb2⅐Sos complex. Two possible modes of action may be proposed (Fig. 8). First, Spry proteins may inhibit the nucleotide exchange factor activity of Sos or limit its access to Ras. Second, Spry proteins may recruit a Ras GTPase activating protein (GAP) to the cell membrane or stimulate its activity to deactivate Ras. Neither Spry1 nor Spry2 could be demonstrated to bind to Ras GAP, which tends to argue against the deactivation model. In contrast, Casci et al. (3) found that Drosophila Spry could bind in vitro to the fly counterpart of Ras GAP. However, neither the ability of Drosophila Spry to bind in vivo to the fly Grb2 and Ras GAP, nor the functional consequences of these bindings have been addressed yet and may not be directly linked to its inhibitory effect.
As restriction of signaling is critical for proper patterning throughout development, the dual role played by Spry proteins in proliferation and differentiation raises the possibility that Spry malfunctions are responsible for developmental defects, as well as neoplasia. Our isolation of spry1 as a gene downstream of the WT1 tumor suppressor is consistent with this notion. Future studies should reveal the detailed molecular mechanism of the inhibitory effects of Spry as well as a possible role for these proteins in development and diseases.