The Bimodal Regulation of Epidermal Growth Factor Signaling by Human Sprouty Proteins

The bimodal regulation of epidermal growth factor signaling by human Sprouty proteins

James E. Egan*, Amy B. Hall†, Bogdan A. Yatsula†, and Dafna Bar-Sagi†‡

†Department of Molecular Genetics and Microbiology and *Graduate Program in Molecular Pharmacology, State University of New York, Stony Brook, NY 11794-5222

Communicated by Joseph Schlessinger, Yale University School of Medicine, New Haven, CT, March 6, 2002 (received for review July 22, 2001) Signal transduction through epidermal growth factor receptors inhibitory effects of dSpry on DER signaling, genetic and biochem- (EGFRs) is essential for the growth and development of multicellular ical studies have shown that it also negatively regulates Drosophila organisms. A genetic screen for regulators of EGFR signaling has led tracheal branching by interfering with Branchless fibroblast growth to the identiﬁcation of Sprouty, a cell autonomous inhibitor of EGF factor (FGF) signaling (8). Although in the Drosophila tracheal signaling that is transcriptionally induced by the pathway. However, system dSpry acts non-cell-autonomously to repress FGF signaling the molecular mechanisms by which Sprouty exerts its antagonistic (8), in the Drosophila eye system, dSpry acts cell autonomously to effect remain largely unknown. Here we have used transient expres- antagonize EGFR signaling (4, 5). sion in human cells to investigate the functional properties of human The Drosophila Sprouty gene encodes a 63-kD membrane- Sprouty (hSpry) proteins. Ectopically expressed full-length hSpry1 associated protein with a unique carboxyl-terminal cysteine-rich and hSpry2 induce the potentiation of EGFR-mediated mitogen- domain that is highly conserved in the four vertebrate Sprouty activated protein (MAP) kinase activation. In contrast, truncation homologs (Spry1-4) (9, 10). The biological role of vertebrate mutants of hSpry1 and hSpry2 containing the highly conserved Sprouty seems to be evolutionarily conserved as murine Spry2 has carboxyl-terminal cysteine-rich domain inhibit EGF-induced MAP ki- been shown to interfere with the FGF-mediated processes of lung nase activation. The potentiating effect of the full-length hSpry2 morphogenesis, and murine Spry2 and Spry4 inhibit FGF-mediated proteins on EGF signaling is mediated by the amino-terminal domain bone outgrowth and limb patterning (9–11). Murine Spry4 has also and results from the sequestration of c-Cbl, which in turn leads to the been shown to inhibit vascular endothelial growth factor (VEGF)- inhibition of EGFR ubiquitination and degradation. These results mediated angiogenesis and prevent FGF- or VEGF-induced MAP indicate that hSpry2 can function both as a negative and positive kinase activation (12). In mammalian cells, EGF stimulation in- regulator of EGFR-mediated MAP kinase signaling in a domain- duces the expression of Spry2 and Spry4 (13). However, the role of dependent fashion. A dual function of this kind could provide a Sprouty proteins in EGF-mediated signaling in mammalian cells mechanism for achieving proper balance between the activation and remains unclear. In this study we have investigated the effects of repression of EGFR signaling. human Spry2 (hSpry2) and Spry1 (hSpry1) on EGF-mediated extracellular signal-regulated kinase 2 (ERK2) MAP kinase acti- eceptor tyrosine kinase-mediated signaling pathways are used vation in human cells. We demonstrate that hSpry2 can exert two Rin both invertebrates and vertebrates to control critical aspects opposing effects on EGF-induced ERK2 MAP kinase activation in of organ morphogenesis, patterning, cellular proliferation, and a domain-dependent manner. Whereas the highly conserved cys- differentiation (1). These signaling pathways are tightly regulated by teine-rich domain is able to inhibit EGF-induced MAP kinase positive and negative inputs that ensure the proper cellular fate (2). activation, the amino-terminal domain potentiates EGF-induced The development of the Drosophila compound eye is regulated MAP kinase activation by interfering with EGFR ubiquitination primarily through inductive signaling emanating from the Drosoph- and down-regulation. ila homolog of the epidermal growth factor receptor (EGFR) (DER; ref. 3). Recently, a genetic screen for modifiers of DER Materials and Methods signaling in the Drosophila eye has identified Sprouty (dSpry) as a Cell Culture and Transient Transfection. HeLa cells were cultured in negative modulator of this pathway (4). Loss-of-function dSpry DMEM supplemented with 10% FBS in a humidified incubator mutations were able to overcome the phenotype resulting from the with 5% CO2 at 37°C. Subconfluent HeLa cells were transfected overexpression of another negative regulator of EGFR signaling, with the indicated plasmids, using the standard calcium phosphate Argos, whereas the overexpression of dSpry induced a phenotype method. The precipitate was left on the cells for 10 h, and cells were similar to that exhibited by EGFR loss-of-function mutants (4). In allowed to recover for 12 h in DMEM supplemented with 10% FBS. addition, dSpry function is required for the proper development of Chinese hamster ovary (CHO) cells were cultured in Ham’s F-12 other tissues in Drosophila that require DER signaling, including medium supplemented with 10% FCS. For transfection, CHO cells embryonic chordotonal organ precursors, the wing imaginal disk, were incubated with Fugene6 (Roche Molecular Biochemicals) for midline glia, and ovarian follicle cells (5, 6). Overexpression of 24 h and then serum-deprived for 16 h. dSpry in these tissues mimics a DER loss-of-function phenotype, whereas loss of dSpry leads to a hyperactivation of the DER EGF Stimulation. Cells were serum-deprived for 2 h and then pathway. incubated with EGF stimulation medium [DMEM containing 25 The activation of DER by EGF leads to the stimulation of the mM Hepes (pH 7.4) and 10 ng͞ml recombinant human EGF] at 4°C ͞ Ras mitogen-activated protein (MAP) kinase pathway (7). In for 15 min. The stimulation medium was subsequently removed, CELL BIOLOGY Drosophila, dSpry seems to impinge on the DER pathway upstream of MAP kinase as overexpression of dSpry leads to a decrease in activated MAP kinase (6). Epistasis experiments have indicated Abbreviations: MAP, mitogen-activated protein; ERK2, extracellular signal-regulated ki- that during wing development dSpry is acting downstream of the nase 2; EGF, epidermal growth factor; EGFR, EGF receptor; HA, hemagglutinin; MBP, myelin basic protein; DER, Drosophila homolog of EGFR; CHO, Chinese hamster ovary; IP, immu- receptor and at or below the level of Raf (6). However, genetic noprecipitated͞immunoprecipitation; Wb, Western blotting. analysis of eye development has placed the level of inhibition by ‡To whom reprint requests should be addressed. E-mail: [email protected]. dSpry upstream of Ras (4). The expression of dSpry is induced by The publication costs of this article were defrayed in part by page charge payment. This DER signaling, indicating the involvement of dSpry in the negative article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. feedback control of DER signaling (4–6). In addition to the §1734 solely to indicate this fact.

www.pnas.org͞cgi͞doi͞10.1073͞pnas.052090899 PNAS ͉ April 30, 2002 ͉ vol. 99 ͉ no. 9 ͉ 6041–6046 Downloaded by guest on October 1, 2021 Fig. 1. The expression of hSpry1 and hSpry2 potentiates EGF-induced MAP kinase activation. Serum- deprived HeLa cells coexpressing HA-ERK2 and vector, HA-hSpry2 (A and B), or HA-hSpry1 (C and D) were stimulated with 10 ng͞ml of EGF for the indicated times. MAP kinase activation was measured by an immunocomplex kinase assay with MBP as a substrate. (A and C) Expression levels of hSpry proteins and ERK2 were determined by Western blotting (Wb), using anti-Spry and anti-ERK2 Abs, respectively. (B and D) Quantitation of the kinetics of MAP kinase activation by EGF in the absence (open circles) or presence (closed circles) of hSpry proteins. The amount of phosphorylated MBP was determined with a PhosphorImager and is presented in arbitrary units as the fold activation over unstimulated vector-expressing cells. Shown are the averages of ﬁve independent experiments Ϯ SDs.

and the cells were incubated with DMEM at 37°C for the indicated immunoprecipitations were kindly provided by Joseph Schless- times. To stop stimulation, cells were washed with ice-cold PBS and inger. Human recombinant EGF was obtained from GIBCO. kept on ice. Protein A Sepharose was obtained from Sigma–Aldrich. Mixed monoclonal anti-PLC-␥1 and anti-phosphotyrosine (4G10) were EGFR Ubiquitination Assay. Serum-deprived CHO cells were washed purchased from Upstate Biotechnology (Lake Placid, NY) Ab. twice with cold PBS and lysed in coimmunoprecipitation buffer (50 mM Tris, pH 7.5͞150 mM NaCl͞10% glycerol͞1% Nonidet P-40͞1 Western Blotting and Coimmunoprecipitation. HeLa cells were lysed mM EDTA͞1 mM phenylmethanesulfonyl fluoride͞1mM in 1 ml of RIPA buffer (150 mM NaCl͞10 mM Tris⅐HCl, pH ͞ ␮ ͞ ␮ ͞ ␮ ͞ ͞ ͞ ͞ NaV04 10 g/ml pepstatin 10 g/ml aprotinin 10 g/ml leupep- 7.4 1% Triton X-100 0.5% deoxycholate 0.1% SDS 1mM tin͞10 mM benzamidine). The lysates were clarified at 14,000 ϫ g EDTA͞50 mM NaF), and lysates were incubated with 2 ␮g͞ml of for 15 min and then incubated with anti-EGFR mAb for1hat4°C. the indicated Abs for 45 min, followed by a 45-min incubation with Subsequently, the immune complexes were incubated with protein 60 ␮l of a 50% slurry protein A Sepharose beads at 4°C. The A Sepharose beads for 30 min. The immunoprecipitates were immunocomplexes were resolved by SDS͞PAGE, and the proteins washed 3 times in cold lysis buffer and then resuspended in SDS were transferred to nitrocellulose membranes. Membranes were sample buffer. The samples were resolved by SDS͞PAGE. incubated for 30 min at room temperature in blocking buffer containing 5% BSA. The anti-HA Ab (1:10,000), anti-ERK2 Cloning and Plasmid Construction. cDNAs encoding human Spry1 (1:1,000), anti-T7 (1:10,000), anti-hSpry (1:2,500), antiphosphoty- and Spry2 were isolated by reverse transcription–PCR. Total RNA rosine (1:1000), anti-EGFR (1:200), or anti-PLC-␥1 (1:1000) pri- was prepared from HeLa cells by the LiCl-urea method (14). All mary Abs were added to the blocking buffer, and blots were Sprouty constructs and the ERK2 construct were cloned into a incubated for1hatroom temperature. The blots were washed in hemagglutinin (HA) tag-containing mammalian expression vector blocking buffer containing 5% milk and then incubated with either (pCGN), and Raf-CAAX and HRasV12 were cloned into a T7 goat anti-mouse horseradish peroxidase (HRP)-conjugated tag-containing mammalian expression vector (pCGT), using stan- (1:10,000) or goat anti-rabbit HRP-conjugated (1:5,000) secondary dard PCR and molecular cloning techniques. The human PLC-␥1 Abs. Reactive proteins were detected with an enhanced chemilu- cDNA was kindly provided by M. Mohammadi (New York Uni- minescence reagent (NEN). versity, School of Medicine). Results ERK2 MAP Kinase Assay. ERK2 immunocomplex assay was per- Expression of hSpry1 and hSpry2 Potentiates EGF-Induced MAP Kinase formed as described by using myelin basic protein (MBP) as a Activation. To test the effect of hSpry2 on EGF signaling, we substrate (15). The amount of radiolabeled MBP was quantified investigated the effects of ectopically expressed hSpry2 on EGF- with a PhosphorImager (Molecular Dynamics). induced ERK2 MAP kinase activation, hereafter referred to as MAP kinase. HeLa cells were transiently transfected with HA- Abs and Reagents. Anti-c-Cbl and anti-EGFR for Western blotting tagged ERK2 and HA-tagged hSpry2 constructs, and MAP kinase was purchased from Santa Cruz Biotechnology, anti-ERK2 was activity was measured by an immunocomplex kinase assay with purchased from Cell Signaling (Beverly, MA), anti-T7 was pur- MBP as a substrate. In the absence of hSpry2, MAP kinase chased from Novagen, and the goat anti-mouse and anti-rabbit activation induced by EGF reached a peak 5 min after the addition horseradish peroxidase-conjugated secondary Abs were from ICN. of the growth factor and returned to basal levels within1h(Fig.1 Polyclonal anti-Spry Abs were generated by Cocalico Biologicals A and B). The expression of hSpry2 had no effect on the magnitude (Reamstown, PA) with a fragment of hSpry2 corresponding to of ERK2 activation by EGF. However, a small but reproducible amino acids 1–172. The anti-EGFR and anti-PLC-␥1 Abs used for increase in the duration of EGF-induced MAP kinase activation

6042 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.052090899 Egan et al. Downloaded by guest on October 1, 2021 Fig. 2. The carboxyl-terminal cysteine-rich and the amino-terminal domains of hSpry2 have opposing effects on EGF-induced MAP kinase activation. Serum-deprived HeLa cells coexpressing HA-ERK2 and vector, HA-hSpry2- 172 (A and B),or HA-hSpry-⌬175 (C and D) were stimulated with 10 ng͞ml of EGF for the indicated times. MAP kinase activation was measured by an immunocomplex kinase assay with MBP as a substrate. (A and C) Expression levels of HA-hSpry2-172, HA-hSpry-⌬175, or HA- ERK2 were determined by Wb, using anti- hSpry, anti-HA, and anti-ERK2 Abs, respectively. (B and D) Quantitation of the kinetics of MAP kinase activation by EGF in the absence (open circles) or presence (closed circles) of hSpry proteins. The amount of phosphorylated MBP was determined with a PhosphorImager and is presented in arbitrary units as the fold activation over unstimulated vector- expressing cells. Shown are the averages of ﬁve independent experiments Ϯ standard deviations.

was observed in hSpry2-expressing cells. These observations differ activity is located within the amino terminal. The mutual antago- from the genetic data implicating dSpry in the negative regulation nism between the amino- and carboxyl-terminal domains could of EGF-mediated MAP kinase activation (4, 6). That this differ- explain the lack of an inhibitory effect of the full-length hSpry ence does not reflect a unique aspect of the experimental system proteins on EGF-induced MAP kinase activation. Indeed, the used is indicated by the fact that the lack of an inhibitory effect by inhibitory effect of the cysteine-rich domain on MAP kinase hSpry2 expression on MAP kinase activation was observed in a activation could be suppressed by the coexpression of the amino- number of different cell types including COS1, 293T, and NIH 3T3 terminal domain (data not shown). cells (data not shown). Moreover, the expression of a different isoform of human Sprouty, hSpry1, had the same effect on the The Cysteine-Rich Domain of hSpry2 Acts Upstream of Ras. It has been kinetics of EGF-induced MAP kinase activation as hSpry2 (Fig. 1 shown that during Drosophila compound eye development, dSpry C and D). Thus, the expression of hSpry1 and hSpry2 protein in inhibits MAP kinase activation upstream of Ras and downstream mammalian cells does not interfere with the activation of MAP of the receptor (4). To determine the level at which the cysteine-rich kinase by EGF. domain of hSpry2 is inhibiting EGF-mediated MAP kinase activation, constitutively active forms of Ras (HRasV12) or Raf The Carboxyl-Terminal Cysteine-Rich Domain and the Amino-Terminal (Raf-CAAX) were coexpressed with the cysteine-rich domain of Domain of hSpry2 Have Opposing Effects on EGF-Induced MAP Kinase hSpry2 and MAP kinase activation was measured. As illustrated in Activation. There are two main regions of homology between Fig. 3 A and B, the cysteine-rich domain of hSpry2 was unable to hSpry2 and dSpry proteins. The most conserved region corre- inhibit either the Raf-CAAX- or HRasV12-mediated activation of MAP kinase. Similar results were obtained with the cysteine-rich sponds to a carboxyl-terminal cysteine-rich domain (amino acids domain of hSpry1-⌬177 (data not shown). These data place the 179–301 in hSpry2 and 380–503 in dSpry) where the degree of level of inhibition by the cysteine-rich domain of hSpry2 down- homology is 51%. The second region resides in the amino-terminal stream of the EGFR and upstream of both Raf and Ras. To domain of the proteins and corresponds to amino acids 36–60 in investigate whether hSpry2-⌬175 can influence other downstream hSpry2 and amino acids 179–205 in dSpry (8). To investigate the effectors of EGFR signaling, this construct was assayed for its possible role of these two regions of homology in EGF signaling, ability to effect EGF-mediated activation of PLC-␥1. HeLa cells HA-tagged truncation mutants were constructed that contained were transiently transfected with either vector or hSpry2-⌬175 and either the amino-terminal homology region (residues 1–172, human PLC-␥1. The transfection efficiency under the conditions hSpry2-172) or the cysteine-rich domain (residues 176–315, ⌬ used was greater than 90%. The cells were then stimulated with hSpry2- 175). These constructs were assayed for their ability to EGF, and PLC-␥1 activation was monitored by assessing its tyrosine influence the activation of MAP kinase by EGF. Transiently phosphorylation state by means of immunoprecipitating with anti-

transfected HeLa cells expressing hSpry2-172 displayed an en- PLC-␥1 Ab and immunoblotting with antiphosphotyrosine Ab. The CELL BIOLOGY hancement in the level of MAP kinase activation in response to expression of hSpry2-⌬175 had no effect on EGF-mediated ty- EGF stimulation at all time points measured (Fig. 2 A and B). rosine phosphorylation of PLC-␥1 (Fig. 3C), indicating that the Conversely, when hSpry2-⌬175 was ectopically expressed there was cysteine-rich domain is not acting in a broad, indiscriminate manner a marked attenuation of EGF-induced MAP kinase activation (Fig. to block EGF signaling but rather interferes specifically with 2 C and D). Similar results were obtained after the expression of the EGFR-mediated activation of the Ras͞MAP kinase pathway. cysteine-rich domain of hSpry1 (residues 177–319 hSpry1-⌬177, data not shown). These observations suggest that an inhibitory The Interaction of hSpry2 with c-Cbl Influences its Ability to Nega- activity, with regards to EGFR signaling, resides in the carboxyl tively Modulate EGF-Induced MAP Kinase Activation. Both hSpry2 terminal of hSpry1 and hSpry2, whereas an opposing stimulatory and dSpry have been shown to bind constitutively to the protoon-

Egan et al. PNAS ͉ April 30, 2002 ͉ vol. 99 ͉ no. 9 ͉ 6043 Downloaded by guest on October 1, 2021 Fig. 3. The cysteine-rich domain of hSpry2 acts upstream of Ras. (A) HeLa cells cotransfected with HA-ERK2 in the absence or presence of HA-hSpry-⌬175, and vector, T7-Raf-CAAX, or T7-HRasV12. Transfected cells were serum-deprived for 16 h, and MAP kinase activation was measured by an immunocomplex kinase assay with MBP as a substrate. Expression levels of ectopically expressed proteins were determined by the Wb of cell lysates with the indicated Abs. (B) Quantitation of the level of MAP kinase activity induced by the transfection of vector, T7-Raf-CAAX, or T7-HRasV12 in the absence (shaded bars) or presence of HA-hSpry2-⌬175 (solid bars). The amount of phosphorylated MBP was determined with a Phospho- rImager and is presented in arbitrary units as the fold activation over serum-deprived vector-expressing cells. Shown are the averages of ﬁve independent experiments Ϯ SDs. (C) Serum-starved HeLa cells coexpressing either vector or HA-hSpry2-⌬175 and human PLC-␥1 were stimulated with 10 ng͞ml of EGF for the indicated times. PLC-␥1 was immunoprecipitated (IP) with anti- PLC-␥1Ab and detected by Wb with antiphosphotyrosine Ab (PY). The slower migrating tyrosine-phosphorylated band corresponds to the activated EGFR that was coprecipitated with PLC-␥1. Expression levels of HA-hSpry2-⌬175 or PLC-␥1 were determined by using anti-HA and anti- PLC-␥1 Abs, respectively. The results shown are from a single experiment that is representative of four independent experiments.

cogene c-Cbl. This interaction is mediated by the amino-terminal lack endogenous EGFR, were cotransfected with the EGFR and conserved region of hSpry2 and dSpry (residues 36–53 in hSpry2 HA-ubiquitin with or without hSpry2. The EGFR was immuno- and 179–205 in dSpry) and the RING finger domain of c-Cbl (16). precipitated, and ubiquitination was assessed by immunoblotting c-Cbl functions as a negative regulator of receptor tyrosine kinases with anti-HA Ab. Consistent with earlier findings (17), the ubiq- by targeting activated receptors for ubiquitination and lysosomal uitination of EGFR was enhanced after addition of EGF (Fig. 5A). degradation (17, 18). To investigate the functional significance of The expression of hSpry2 induced a marked inhibitory effect on hSpry2 binding to c-Cbl, we constructed an hSpry2 deletion mutant EGFR ubiquitination. In contrast, the hSpry2-⌬36-53 deletion (hSpry2-⌬36-53) that lacks the c-Cbl binding domain. In agreement mutant that does not interact with c-Cbl had no effect on EGFR with an earlier study (16), this deletion mutant was defective in ubiquitination (Fig. 5B), indicating that the antagonistic effect of binding to ectopically expressed c-Cbl (Fig. 4A) as well as endog- hSpry2 on EGFR ubiquitination is a result of its interaction with enous c-Cbl (data not shown). In contrast to the full-length hSpry2, c-Cbl. The inhibitory effect of hSpry2 on EGFR ubiquitination was the hSpry2-⌬36-53 construct was able to inhibit MAP kinase accompanied by the inhibition of receptor degradation, which could activation by approximately 50% (Fig. 4 B and C). These results be clearly observed in the presence of overexpressed c-Cbl (Fig. suggest that the interaction of hSpry2 with c-Cbl interferes with the 5C). The interference of hSpry2 with ligand-induced EGFR down- ability of hSpry2 to inhibit EGF signaling. Furthermore, these regulation provides a mechanistic explanation for the inability of observations illustrate that the negative regulatory function of the full-length hSpry2 to exert an inhibitory effect on EGF-mediated cysteine-rich domain of hSpry2 can be exerted in the context of the MAP kinase activation and for the enhancement of EGF-mediated full-length protein. MAP kinase activation induced by the expression of hSpry2 amino- terminal region. hSpry2 Inhibits EGFR Ubiquitination. The targeting of activated EGFR for ubiquitination by c-Cbl is mediated by its RING finger Discussion domain that acts as an E3 protein ubiquitin ligase (19–21). Because EGFR-mediated signaling is critical for the growth and develop- hSpry2 has been shown to bind directly to the RING finger domain ment of multicellular organisms (22). Genetic studies in Drosophila of c-Cbl it would be predicted to interfere with c-Cbl-dependent suggest that dSpry negatively regulates EGFR signaling pathways in EGFR ubiquitination. To test this idea, we investigated the effects multiple tissues by interfering with EGF-mediated Ras͞MAP of hSpry2 expression on EGFR ubiquitination. CHO cells, which kinase activation (4, 6). However, the precise mechanisms by which

6044 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.052090899 Egan et al. Downloaded by guest on October 1, 2021 of dSpry, they are in agreement with a recent study documenting the lack of an inhibitory effect of either mSpry1 or mSpry2 on EGF-induced MAP kinase activation in endothelial cells (24). In contrast to the full-length forms of hSpry1 and hSpry2, their truncated forms encompassing the carboxyl-terminal cysteine-rich domain exerts a pronounced inhibitory effect on EGF-mediated MAP kinase activation. The cysteine-rich domain comprises the most highly conserved region between vertebrate and invertebrate Sprouty proteins (8) and has been implicated in the membrane targeting of both Drosophila and human Sprouty proteins (4, 25). Therefore, it is likely that the capacity of this domain to interfere with EGF-induced MAP kinase activation in mammalian cells represents a conserved inhibitory mechanism. This possibility is also supported by our observation that, like dSpry, the cysteine-rich domain of hSpry2 acts downstream of the EGFR and upstream of Ras. Moreover, the cysteine-rich domain is necessary for the inhibition of cellular migration in response to serum, EGF, FGF, and platelet-derived growth factor by hSpry2 (23), and the cysteine- rich domain of Xenopus Sprouty is sufficient to block FGF- mediated PLC-␥ activation (26). Taken together, these findings indicate that the inhibitory activity of Sprouty on receptor tyrosine kinase-signaling proteins is conferred by the highly conserved cysteine-rich domain. Although the molecular basis for the inhibitory effect of the cysteine-rich domain remains to be determined, it is noteworthy that this effect is exerted only in the context of EGF-induced MAP kinase activation, suggesting that the cysteine- rich domain of hSpry proteins targets regulatory events that specifically control the coupling of activated EGFR with the Ras͞ MAP kinase signaling pathway. The amino-terminal domains of Sprouty proteins are highly divergent not only between the invertebrate and vertebrate proteins, but also within the four vertebrate proteins themselves (9). Our data with the amino-terminal domain of hSpry2 have demonstrated that this region is a positive modulator of EGF-dependent MAP kinase activation. This effect results from the inhibition of EGFR ubiquitination and down-regulation. That the potentiating effect of the amino-terminal domain of hSpry2 on EGF-induced MAP kinase activation is antagonistic to the inhibitory effect of the cysteine-rich domain suggests that the negative regulation of EGFR signaling by Sprouty might be best accomplished when receptor levels are limiting. In support of this idea, it has been shown that in Drosophila the EGFR signaling defect induced by dSpry can be rescued by the overexpression of the EGFR (4). We have demonstrated that the potentiating effects of hSpry2 on EGF-induced MAP kinase activation can be attributed to its Fig. 4. The interaction of hSpry2 with c-Cbl inﬂuences its ability to negatively interaction with c-Cbl. Because this interaction occurs through the modulate EGF-induced MAP kinase activation. (A) HeLa cells transfected with binding of a small amino-terminal region of hSpry2 to the RING HA-c-Cbl and vector, HA-hSpry2, HA-hSpry2-172, or HA-hSpry2-⌬36-53 were finger domain of c-Cbl (16), it is presumed to interfere with the E3 lysed and analyzed by immunoprecipitation with anti-c-Cbl Ab (IP) and Wb with enzymatic activity of the RING finger. The c-Cbl binding region of the indicated Abs. (B) HeLa cells transfected with HA-ERK2 and either vector or hSpry is conserved in dSpry, and the latter has been shown to also HA-hSpry2-⌬36-53 were stimulated with 10 ng͞ml of EGF for the indicated times. interact with Drosophila Cbl (dCbl) (16). Yet genetically, dSpry MAP kinase activation was measured by an immunocomplex kinase assay with ͞ MBP as a substrate. Expression levels of HA-hSpry2-⌬36-53 and HA-ERK2 were seems to function solely as a negative regulator of EGFR MAP determined by Wb, using anti-Spry or anti-ERK2 Abs, respectively. (C) Quantita- kinase signaling. What could account for the apparent differences tion of the kinetics of MAP kinase activation by EGF in the absence (open circles) between the effects of Cbl binding on the activities of mammalian or presence (closed circles) of HA-hSpry2-⌬36-53 protein. The amount of phos- and Drosophila Sprouty proteins? One possibility is that in the phorylated MBP was determined with a PhosphorImager and is presented in EGFR-dependent developmental processes that are influenced by arbitrary units as the fold activation over unstimulated vector-expressing cells. Sprouty, the level of EGFR activity is limiting even in the presence Ϯ Shown are the averages of three independent experiments SDs. of dSpry. Indeed, the influence of dCbl on EGFR signaling is

extremely sensitive to the dosage of genes that participate in this CELL BIOLOGY pathway, and, on a wild-type background, overexpression of dCbl dSpry inhibits EGFR signaling are poorly understood. In mamma- has no visible effect on EGFR signaling in the eye (27, 28). Another lian cells, murine Sprouty proteins (mSpry) and hSpry2 have been possibility is that the interaction of Cbl with Sprouty is differentially shown to inhibit EGF-induced cell migration and proliferation (23, regulated in Drosophila and mammalian cells. The diverse se- 24), but the biochemical basis for these effects is unknown. In the quences in the amino-terminal domains of mammalian and Dro- present study we have found that ectopic expression of either sophila Sprouty proteins may direct different protein–protein in- hSpry1 or hSpry2 did not interfere with EGF-mediated activation teractions or posttranslational modifications that in turn could of MAP kinase in mammalian cells. Although these observations influence Cbl binding. Finally, the functional consequences of seem to be contradictory to the genetic evidence for the function changes in receptor ubiquitination might be different in Drosophila

Egan et al. PNAS ͉ April 30, 2002 ͉ vol. 99 ͉ no. 9 ͉ 6045 Downloaded by guest on October 1, 2021 Fig. 5. hSpry inhibits EGFR ubiquitination and degradation through its c-Cbl binding domain. (A and B) CHO cells were transiently transfected with expression vectors encoding the EGFR, HA-ubiquitin, and either HA-hSpry2 (A) or HA-hSpry2-⌬36-53 (B). Forty-eight hours after transfection, the cells were incubated with or without EGF (10 ng͞ml) for 10 min. Subsequently, the cells were lysed and analyzed by IP and Wb with the indicated Abs. (C) CHO cells were transiently transfected with expression vectors encoding EGFR, HA-c-Cbl, and HA-hSpry2. The cells were stimulated with EGF as described in A and B and analyzed by IP and Wb with the indicated Abs.

and mammalian systems. Whereas in mammalian cells c-Cbl- EGFR signaling in mammalian cells would induce the expression of mediated ubiquitination promotes EGFR down-regulation (17, 29), Sprouty proteins, which could act initially as inhibitors and subse- a recent study has demonstrated a role for deubiquitination in quently as potentiators of EGFR signaling. The nature of the effects stimulating receptor internalization in Drosophila (30). In this study, exerted by Sprouty and their sequential order would depend on the the Drosophila Fat facet protein, which is a member of the UBP availability of proteins that interact with Sprouty and regulate its class of deubiquitinating enzymes and plays a critical role in function. In this manner, cells can regain the capacity to be activated Drosophila eye development, has been found to activate endocytosis by EGF while Sprouty is present. This mechanism would be of (30). Biochemical analysis of the effects of dSpry on the ubiquiti- particular importance in biological systems where there is a re- nation and endocytic sorting of the EGFR in Drosophila will be quirement for the repeated use of EGF signaling. required to distinguish between these possibilities. Our observations indicate that, depending on its interactions with We thank all members of the laboratory for helpful advice and discussions c-Cbl, hSpry2 can function as a positive or negative regulator of and James Keller for the generation of anti-hSpry Abs. A.B.H. is supported EGFR͞MAP kinase signaling. A dual function of this kind could by National Institutes of Health Training Grant T32DK07521-14. D.B.S. be used to set a temporal limit to the inhibitory effect of Sprouty acknowledges support by National Institutes of Health Grants CA55360 and on EGF signaling. For example, by analogy to dSpry, high levels of CA28146.

1. Schlessinger, J. (2000) Cell 103, 211–225. 17. Levkowitz, G., Waterman, H., Zamir, E., Kam, Z., Oved, S., Langdon, W. Y., Beguinot, L., 2. Freeman, M. (2000) Nature (London) 408, 313–319. Geiger, B. & Yarden, Y. (1998) Genes Dev. 12, 3663–3674. 3. Freeman, M. (1997) Development (Cambridge, U.K.) 124, 261–270. 18. Miyake, S., Lupher, M. L., Jr., Druker, B. & Band, H. (1998) Proc. Natl. Acad. Sci. USA 95, 4. Casci, T., Vinos, J. & Freeman, M. (1999) Cell 96, 655–665. 7927–7932. 5. Kramer, S., Okabe, M., Hacohen, N., Krasnow, M. & Hiromi, Y. (1999) Development 19. Joazeiro, C. A. n. P., Wing, S. S., Huang, H.-k., Leverson, J. D., Hunter, T. & Liu, Y.-C. (Cambridge, U.K.) 126, 2515–2525. (1999) Science 286, 309–312. 6. Reich, A., Sapir, A. & Shilo, B.-Z. (1999) Development (Cambridge, U.K.) 126, 4139–4147. 20. Yokouchi, M., Kondo, T., Houghton, A., Bartkiewicz, M., Horne, W. C., Zhang, H., 7. Chang, H., Karim, F., O’Neill, E., Rebay, I., Solomon, N., Therrien, M., Wassarman, D., Yoshimura, A. & Baron, R. (1999) J. Biol. Chem. 274, 31707–31712. Wolff, T. & Rubin, G. (1994) Cold Spring Harbor Symp. Quant. Biol. 59, 147–153. 21. Waterman, H., Levkowitz, G., Alroy, I. & Yarden, Y. (1999) J. Biol. Chem. 274, 22151– 8. Hacohen, N., Kramer, S., Sutherland, D., Hiromi, Y. & Krasnow, M. (1998) Cell 92, 253–263. 22154. 9. Minowada, G., Jarvis, L., Chi, C. L., Neubuser, A., Sun, X., Hacohen, N., Krasnow, M. & 22. Moghal, N. & Sternberg, P. W. (1999) Curr. Opin. Cell Biol. 11, 190–196. Martin, G. R. (1999) Development (Cambridge, U.K.) 126, 4465–4475. 23. Yigzaw, Y., Cartin, L., Pierre, S., Scholich, K. & Patel, T. B. (2001) J. Biol. Chem. 276, 10. Tefft, J. D., Lee, M., Smith, S., Leinwand, M., Zhao, J., Bringas, P., Crowe, D. L. & 22742–22747. Warburton, D. (1999) Curr. Biol. 9, 219–222. 24. Impagnatiello, M.-A., Weitzer, S., Gannon, G., Compagni, A., Cotten, M. & Christofori, G. 11. Mailleux, A., Tefft, D., Ndiaye, D., Itoh, N., Thiery, J. P., Warburton, D. & Bellusci, S. (2001) 152, Mech. Dev. 102, 81–94. (2001) J. Cell Biol. 1087–1098. 12. Lee, S. H., Schloss, D. J., Jarvis, L., Krasnow, M. A. & Swain, J. L. (2001) J. Biol. Chem. 25. Lim, J., Wong, E. S. M., Ong, S. H., Yusoff, P., Low, B. C. & Guy, G. R. (2000) J. Biol. Chem. 276, 4128–4133. 275, 32837–32845. 13. Sasaki, A., Taketomi, T., Wakioka, T., Kato, R. & Yoshimura, A. (2001) J. Biol. Chem. 276, 26. Nutt, S., Dingwell, K., Holt, C. & Amaya, E. (2001) Genes Dev. 15, 1152–1166. 36804–36808. 27. Pai, L.-M., Barcelo, G. & Schupbach, T. (2000) Cell 103, 51–61. 14. Sambrook, J., Fritsch, E. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual 28. Meisner, H., Daga, A., Buxton, J., Fernandez, B., Chawla, A., Banerjee, U. & Czech, M. (Cold Spring Harbor Lab. Press, Plainview, NY). (1997) Mol. Cell. Biol. 17, 2217–2225. 15. Walsh, A. B. & Bar-Sagi, D. (2001) J. Biol. Chem. 276, 15609–15615. 29. Lill, N. L., Douillard, P., Awwad, R. A., Ota, S., Lupher, M. L., Jr., Miyake, S., 16. Wong, E. S. M., Lim, J., Low, B. C., Chen, Q. & Guy, G. R. (2001) J. Biol. Chem. 276, Meissner-Lula, N., Hsu, V. W. & Band, H. (2000) J. Biol. Chem. 275, 367–377. 5866–5875. 30. Cadavid, A., Ginzel, A. & Fischer, J. (2000) Development (Cambridge, U.K.) 127, 1727–1736.

6046 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.052090899 Egan et al. Downloaded by guest on October 1, 2021