The Adapter Protein, Grb10, Is a Positive Regulator of Vascular Endothelial Growth Factor Signaling

The adapter protein, Grb10, is a positive regulator of vascular endothelial growth factor signaling

S Giorgetti-Peraldi*,1, J Murdaca1, JC Mas1 and E Van Obberghen1

1INSERM U145, IFR 50, FaculteÂ de MeÂdecine, avenue de Valombrose, 06107 Nice Cedex 2 France

Vascular endothelial growth factor (VEGF) is an expressed only on endothelial cells, Flt-1 (Fms like important regulator of vasculogenesis and angiogenesis. tyrosine kinase) and Flk/KDR (Kinase insert domain Activation of VEGF receptors leads to the recruitment of containing receptor). Both receptors are structurally SH2 containing proteins which link the receptors to the related to the PDGF (Platelet Derived Growth Factor) activation of signaling pathways. Here we report that receptor family (Neufeld et al., 1999). Knock-out Grb10, an adapter protein of which the biological role studies have shown that these receptors have dierent remains unknown, is tyrosine phosphorylated in response biological functions. Indeed, in Flk null mice, to VEGF in endothelial cells (HUVEC) and in 293 cells endothelial and hematopoietic cell development is expressing the VEGF receptor KDR. An intact SH2 impaired. In Flt-1 knock-out mice, there is an apparent domain is required for Grb10 tyrosine phosphorylation in overgrowth of endothelial cells, and blood vessels are response to VEGF, and this phosphorylation is mediated disorganized (Fong et al., 1995; Shalaby et al., 1995). in part through the activation of Src. In HUVEC, VEGF These results and other studies performed by activating increases Grb10 mRNA level. Expression of Grb10 in separately only one receptor suggest that KDR is HUVEC or in KDR expressing 293 cells results in an involved in endothelial cell proliferation and survival, increase in the amount and in the tyrosine phosphoryla- while Flt-1 seems to be required for chemotaxis and tion of KDR. In 293 cells, this is correlated with the procoagulant activity. Although both receptors are activation of signaling molecules, such as MAP kinase. capable of autophosphorylation in vitro (Waltenberger By expressing mutants of Grb10, we found that the et al., 1994), only KDR autophosphorylates in vivo. positive action of Grb10 is independent of its SH2 The deletion of the tyrosine kinase domain of Flt-1 domain. Moreover, these Grb10 eects on KDR seem to does not impair blood vessel formation in homozygous be speci®c since Grb10 has no eect on the insulin mice, suggesting that the kinase domain of Flt-1 is not receptor, and Grb2, another adapter protein, does not required for its action on angiogenesis (Hiratsuka et mimic the eect of Grb10 on KDR. In conclusion, we al., 1998). propose that VEGF up-regulates Grb10 level, which in A series of steps in the VEGF signaling pathway turn increases KDR molecules, suggesting that Grb10 have been elucidated. Activation of VEGF receptors could be involved in a positive feedback loop in VEGF promotes tyrosine phosphorylation of several proteins, signaling. Oncogene (2001) 20, 3959 ± 3968. and recruitment of SH2-containing molecules (Walten- berger et al., 1994). Stimulation of KDR leads to PI-3- Keywords: Grb10; VEGF receptor; KDR; angiogenesis kinase and PKB activation (Gerber et al., 1998; Wu et al., 2000), to MAP kinase activation through PLCg and PKC-dependent pathway (Takahashi et al., 1999), Introduction to nitric oxide synthesis and to FAK tyrosine phosphorylation (Abedi and Zachary, 1997; Shen et Vascular Endothelial Growth Factor (VEGF) is al., 1999). However, many players in VEGF signaling involved in endothelial cell proliferation, microvascular have still to be discovered. Tyrosine kinase receptors permeability, vasodilatation and angiogenesis. Angio- control signaling pathways through association of SH2 genesis, the sprouting of new capillaries from existing containing proteins and tyrosine phosphorylated pro- vasculature, is necessary for vascular development, teins. wound healing, organ regeneration, and has been Grb10 protein contains several interaction motifs, involved in disease states such as diabetic retinopathy, such as a N-terminal proline-rich sequence, a central rheumatoid arthritis and tumor growth (Carmeliet, PH (Pleckstrin homology) domain, a BPS domain 2000; Ferrara and Davis, 1997; Risau, 1997). VEGF (Between PH and SH2) and a C-terminal SH2 domain binds to two speci®c tyrosine kinase receptors (Frantz et al., 1997; He et al., 1998; Ooi et al., 1995). Although Grb10 interacts with tyrosine kinase receptors such as PDGF, EGF, insulin and IGF-I receptors, its function as an adapter molecule remains to be *Correspondence: S Giorgetti-Peraldi Received 23 November 2000; revised 6 April 2001; accepted 9 April discovered (He et al., 1998; Stein et al., 1996). Indeed, 2001 Grb10 binds to signaling molecules such as Raf1, Involvement of Grb10 in VEGF signaling pathway S Giorgetti-Peraldi 3960 MEK1, BCR-Abl, Jak2, NEDD4 (Nantel et al., 1998; the conserved arginine 462 by a lysine. This mutant is Bai et al., 1998; Moutoussamy et al., 1998; Morrione et unable to associate with KDR in yeast two hybrid. All al., 1999), but the consequences of these associations the results show that Grb10 binds to KDR only have not been elucidated. The role of Grb10 in cellular through its SH2 domain. In each condition, the signaling is controversial. According to some authors, amount of expressed proteins was identical as revealed microinjection of the Grb10 SH2 domain decreases by Western blot (data not shown). We conclude that insulin- and IGF-I-mediated mitogenesis (Morrione et association of Grb10 to KDR is dependent on the al., 1997), while in other reports, Grb10 has a positive tyrosine kinase activity of the receptor, and requires a role in PDGF, IGF-I and insulin-induced mitogenesis functional SH2 domain of Grb10. (Wang et al., 1999). In this study, we show that Grb10 is tyrosine Grb10 is tyrosine phosphorylated in response to VEGF phosphorylated in response to VEGF in HUVEC and in 293 cells expressing KDR. Moreover, VEGF We studied whether Grb10 is tyrosine phosphorylated increases Grb10 mRNA in HUVEC. In 293 cells and in response to VEGF. Human endothelial cells in HUVEC, overexpression of Grb10 induces an (HUVEC) were stimulated with VEGF and tyrosine increase in the amount and in the tyrosine phosphor- phosphorylation of endogenous Grb10 was assessed ylation of KDR, coupled to an increase in the VEGF- after immunoprecipitation with an antibody to Grb10. induced MAP kinase activity. We show that Grb10 As shown in Figure 2a, VEGF treatment induces a interacts directly with VEGF receptor KDR, and that three-fold increase in Grb10 tyrosine phosphorylation. the SH2 domain of Grb10 mediates this association. To study the mechanism of this phosphorylation, we Since VEGF stimulates the expression of Grb10 expressed epitope-tagged Grb10-HA and KDR-Myc in mRNA, and Grb10 increases KDR molecules, we 293 cells (that lack endogenous VEGF receptors). Cells would like to suggest that a positive feed-back loop were stimulated with VEGF, and Grb10-HA was could exist. immunoprecipitated with an antibody to the HA epitope followed by an antiphosphotyrosine immunoblot. As shown in Figure 2b, VEGF-induced activation of KDR stimulates Grb10 tyrosine phosphorylation. Results Grb10 phosphorylation is abolished when a SH2 domain de®cient mutant of Grb10 is expressed KDR interacts with Grb10 SH2 domain (Grb10 R462K) (Figure 2c). Although Grb10 binds to several tyrosine kinase Src is known to phosphorylate Grb10 on tyrosine receptors, its biological role remains unknown. To (Langlais et al., 2000). Moreover, Src kinases are elucidate the function of Grb10, we investigated its involved in VEGF signaling pathway (Eliceiri et al., implication in VEGF signaling. 1999). To study whether phosphorylation of Grb10 in VEGF binds to two tyrosine kinase receptors, Flt-1 response to VEGF could be due to Src, we treated and KDR. Since it has been shown that only KDR HUVEC with the Src inhibitor, PP2 prior to VEGF autophosphorylates in endothelial cells, we determined stimulation. We observed that PP2 treatment of whether KDR interacts with Grb10 using the yeast two HUVEC did not aect the VEGF-induced tyrosine hybrid system. KDR cDNA corresponding to the phosphorylation of KDR (Figure 3a). Grb10 was cytoplasmic domain of the receptor was subcloned in immunoprecipitated and analysed by antiphosphotyr- frame with the binding domain of LexA. Grb10 cDNA osine immunoblot (Figure 3b). PP2 treatment of was subcloned in frame with the activation domain of HUVEC induces a 40% decrease in the VEGF induced Gal-4. Yeasts were co-transformed by the two hybrids, tyrosine phosphorylation of Grb10. However, the and the association between the proteins was measured tyrosine phosphorylation of Grb10 is not totally by the transactivation of a reporter gene, LacZ. Results blocked suggesting that both Src and KDR are in Figure 1a, show that pLex-KDR interacts with involved in the tyrosine phosphorylation of Grb10 pAct-Grb10 as compared to a negative control, Raf. after VEGF stimulation. In conclusion, Grb10 is This interaction is dependent on the tyrosine kinase tyrosine phosphorylated in response to VEGF. This activity of the receptor, since the generation of a phosphorylation seems to be due to both an action of kinase-dead mutant (KD) of KDR by replacement of KDR and Src. lysine 868 by alanine, abolishes the interaction of KDR with Grb10 in the yeast two hybrid system. Grb10 mRNA is increased by VEGF To localize the binding site on Grb10, we generated several mutants of pACT-Grb10 (Figure 1b). As Since Grb10 seems to be involved in VEGF signaling, represented, Grb10 possesses a proline-rich region, a we investigated whether VEGF could modulate Grb10 PH domain, a BPS region and a SH2 domain. All the expression. Endothelial cells (HUVEC) were treated deletion mutants of Grb10 were tested for their abilities with VEGF for 2, 4 and 6 h, RNA were extracted, and to bind to pLex-KDR (Figure 1c). The SH2 domain of Northern blot analysis was performed using a KDR or Grb10 binds to KDR, but not the N-terminal region a Grb10 cDNA speci®c probe (Figure 4). We observed (NT) nor the BPS domain. Grb10R462K possesses a that VEGF induces a threefold increase in Grb10 non functional SH2 domain due to the replacement of mRNA in HUVEC.

Oncogene Involvement of Grb10 in VEGF signaling pathway S Giorgetti-Peraldi 3961

Figure 1 Interaction of Grb10 with KDR in the yeast two-hybrid system. (a) L40 yeasts were co-transformed with plasmids encoding pLex-KDR wild type (WT) or kinase dead (KD: K868A) and pACT- Grb10. Interaction was revealed by measuring b- galactosidase activities in lysates. (b) Schematic representation of various Grb10 mutants in pACT vector. (c) L40 yeasts were co- transformed with plasmids encoding pLex-KDR and various Grb10 mutants in pACT vector. Interaction was revealed by measuring the b-galactosidase activities. All the results of b-galactosidase activity are expressed as a percentage of wild type activity and represent the average (+s.e.) of ®ve independent experiments, using three independent transformants

was performed. As observed in Figure 5 (right panel), Ectopic expression of Grb10 increases KDR protein and its tyrosine phosphorylation expression of Grb10 increases the amount of KDR (®vefold). The same result was observed after im- Since Grb10 expression is increased in response to munoprecipitation of KDR-Myc (data not shown). An VEGF, we studied its role on VEGF signaling immunoblot using antibodies to Shc shows that pathway. We transfected Grb10 in 293 cells expressing equivalent amount of proteins have been loaded on KDR. Cells were stimulated with VEGF and whole cell the gel. lysate was analysed by an immunoblot using antibody Since we have observed that Grb10 expression leads to phosphotyrosine (Figure 5, left panel). VEGF to an increase in the basal tyrosine phosphorylation of treatment induced tyrosine phosphorylation of KDR- KDR, we investigated whether this basal phosphoryla- Myc (twofold increase). When Grb10-HA was ex- tion was due to KDR itself. To this end, we generated pressed, two tyrosine phosphorylated proteins ap- a kinase-dead mutant of KDR (KDRKD) by mutation peared, Grb10 and KDR. As already shown in of the ATP binding site (K868A). 293 cells were Figure 2, tyrosine phosphorylation of Grb10 is transfected with KDR or KDRKD in absence or in increased after VEGF stimulation. Moreover, after presence of Grb10, stimulated or not with VEGF, and Grb10 expression, KDR is phosphorylated in absence whole cell lysates were analysed by immunoblots using of VEGF treatment, and its tyrosine phosphorylation antibodies to phosphotyrosine, to Myc, to HA or to is increased (twofold) by VEGF. To verify KDR Shc (Figure 6). We observed that VEGF induces expression level, a Western blot with antibody to Myc tyrosine phosphorylation of KDR (lane 2 compared

Oncogene Involvement of Grb10 in VEGF signaling pathway S Giorgetti-Peraldi 3962

Figure 2 VEGF stimulates Grb10 tyrosine phosphorylation. (a) HUVEC (passage 2) were treated with 20 ng/ml VEGF for 10 min. Cell extracts were subjected to immunoprecipitation with antibody to Grb10. Proteins were revealed by a blot with antibodies to phosphotyrosine. (b) 293 were transfected with pcDNA3 (Mock) or with pcDNA3-KDR-Myc without or with pcDNA3-Grb10-HA. Cells were stimulated with 50 ng/ml VEGF for 10 min and cell extracts were subjected to immunoprecipitation with antibody to HA (12CA5). Proteins were revealed by a Western blot with antibodies to phosphotyrosine and to HA. (c) 293 were transfected with pcDNA3-KDR-Myc with pcDNA3-Grb10-HA or pcDNA-Grb10R462K-HA. Cells were stimulated with 50 ng/ml VEGF for 10 min and cell extracts were subjected to immunoprecipitation with antibody to HA (12CA5). Proteins were revealed by a blot with antibodies to phosphotyrosine and to HA

Figure 3 VEGF- induced tyrosine phosphorylation of Grb10 is mediated by Src. HUVEC were treated with 2 mM PP2 for 30 min prior to VEGF stimulation (50 ng/ml for 5 min). (a) Whole cell lysates were analysed by a blot with antibodies to phosphotyrosine. (b) HUVEC cell lysates were subjected to immunoprecipitation using antibodies to Grb10, and analysed by Western blot with antibodies to phosphotyrosine. Western blots were revealed on Storm 840 (Molecular Dynamics)

Oncogene Involvement of Grb10 in VEGF signaling pathway S Giorgetti-Peraldi 3963 to lane 1), and that Grb10 expression induces an any detectable tyrosine phosphorylation of KDRKD,it increase in the tyrosine phosphorylation and the induces an increase in the amount of KDRKD (lanes 7, amount of KDR (lanes 3, 4 compared to lanes 1, 2). As expected VEGF treatment did not stimulate KDRKD tyrosine phosphorylation (lane 6 compared to lane 5). Although Grb10 expression did not induce

Figure 6 The tyrosine kinase activity of KDR mediates its tyrosine phosphorylation after Grb10 expression. 293 cells were transfected with pcDNA3-KDR-Myc, or pcDNA3-KDRKD in Figure 4 VEGF induces an increase in Grb10 mRNA in absence or in presence of pcDNA3-Grb10-HA. Cells were HUVEC. HUVEC (passage 2) were stimulated for 2, 4 and 6 h stimulated with 50 ng/ml of VEGF for 10 minutes at 378C. with VEGF (20 ng/ml) at 378C. RNA were extracted and Proteins were separated on SDS ± PAGE and revealed by a separated by electrophoresis in formaldehyde-containing agarose Western blot with antibodies to phosphotyrosine, to Myc, to HA gel. After transfer, the membrane was hybridized with a probe or to Shc. Western blots were revealed on Storm 840 (Molecular corresponding to KDR, Grb10 and 18S Dynamics)

Figure 5 Expression of Grb10 increases KDR protein and its tyrosine phosphorylation. 293 cells were transfected with pcDNA- KDR-Myc without or with pcDNA3-Grb10-HA. After stimulation with VEGF (50 ng/ml for 10 min at 378C), proteins were separated by SDS ± PAGE and revealed by a Western blot with antibodies to phosphotyrosine, to Myc (9E10), to HA (12CA5) or to Shc

Oncogene Involvement of Grb10 in VEGF signaling pathway S Giorgetti-Peraldi 3964 8 compared to lanes 5, 6). This result suggests that after Grb10 expression KDR basal tyrosine phosphorylation is due to autophosphorylation. The increase in KDR proteins can be correlated with the activation of signaling pathways, such as MAP kinase. 293 cells were transfected with KDR in absence or in presence of Grb10, and MAP kinase activity was tested using a MBP (myelin basic protein) kinase assay (Figure 7). VEGF increases the ability of MAP kinase to phosphorylate MBP. Expression of Grb10 increases both basal and VEGF-stimulated MAP kinase activity. We were interested to see whether Grb10 expression would aect endogenous KDR in HUVEC cells. Human endothelial cells (HUVEC) were transfected with pcDNA3 vector (mock) or with a vector Figure 8 Expression of Grb10 in HUVEC increases KDR expressing Grb10-HA. Cells were stimulated with expression and phosphorylation. HUVEC (passage 2) were VEGF, and whole cell lysates were analysed by transfected with pcDNA3 (Mock) or pcDNA3-Grb10-HA, and Western blotting with antibodies to phosphotyrosine, stimulated with VEGF (20 ng/ml for 5 min). Proteins were separated on SDS-PAGE and revealed by a blot with antibodies to KDR and to HA. VEGF induces tyrosine to phosphotyrosine, to KDR or to HA. Western blots were phosphorylation of endogenous VEGF receptors revealed on Storm840 Molecular Dynamics (Figure 8). As we have observed in 293 cells (Figure 5), expression of Grb10 in HUVEC increases both the basal and the VEGF-induced tyrosine phosphorylation of endogenous KDR, as well as the total amount of KDR. wild-type, Grb10 R462K (SH2 domain de®cient mutant) or the BPS-SH2 region (Figure 9). Since the BPS region does not bind to KDR (Figure 1c), we Grb10 SH2 domain is not required for the increase in assume that the eect will be due only to the SH2 KDR amount and tyrosine phosphorylation domain. 293 cells were stimulated with VEGF and cell We have shown that the SH2 domain of Grb10 is extracts were analysed by an immunoblot using required to allow Grb10 tyrosine phosphorylation by antibodies to phosphotyrosine, Myc or HA. As we activated KDR. To determine whether the Grb10 SH2 have previously shown, expression of Grb10 induces an domain is also required to promote the positive action increase in phosphorylation (blot anti-phosphotyro- of Grb10, we expressed KDR in presence of Grb10 sine) and the amount of KDR (blot anti-Myc) (lanes 3, 4 compared to lanes 1, 2). The same result is obtained after expression of Grb10 R462K (lanes 9, 10 compared to lanes 1, 2). By contrast, expression of the SH2 domain of Grb10 (Grb10 SH2) does not modify the KDR amount or tyrosine phosphorylation compared to KDR expressing cells (lanes 5, 6 compared to lanes 1, 2). In conclusion, the stimulatory action of Grb10 on KDR expression and phosphorylation does not require a functional SH2 domain.

Specificity of Grb10 action on KDR We investigated whether the stimulatory action of Grb10 on KDR is speci®c. To do so, we studied another SH2 domain containing molecule, Grb2. 293 cells were transfected with KDR alone, or in combination with Grb10 or Grb2 (Figure 10a). After VEGF treatment, cell extracts were analysed by Western blot using antibodies to phosphotyrosine, to Myc, to HA and to Grb2. As previously observed, Grb10 expression increases both the tyrosine phosphorylation and the amount of KDR. This appears to be speci®c since Grb2 does not mimic the eect of Figure 7 Expression of Grb10 stimulates VEGF-induced MAP Grb10 on KDR. Expression of Shc, which contains a kinase activity. 293 were transfected with pcDNA-KDR-Myc and SH2 domain, has the same eect as Grb2, and did not pcDNA3-Grb10-HA. Cells were stimulated with VEGF (50 ng/ml for 10 min at 378C), and MAP kinase was immunoprecipitated modify the amount nor the phosphorylation of KDR and tested for its ability to phosphorylate myelin basic protein (data not shown).

Oncogene Involvement of Grb10 in VEGF signaling pathway S Giorgetti-Peraldi 3965

Figure 9 The SH2 domain of Grb10 is not implicated in the positive regulation of KDR. (a) 293 were transfected with pcDNA3- KDR-Myc, without or with pcDNA3-Grb10-HA, Grb10 R462K-HA or Grb10 SH2-HA. Cells were stimulated with 50 ng/ml of VEGF for 10 minutes at 378C. Proteins were separated on SDS ± PAGE and revealed by a Western blot with antibodies to phosphotyrosine, to Myc or to HA. (b) Quanti®cation of experiments

Figure 10 Speci®city of Grb10 action. (a) 293 were transfected with pcDNA3-KDR-Myc, without or with pcDNA3-Grb10-HA or pcDNA3-Grb2. Cells were stimulated with 50 ng/ml VEGF for 10 min at 378C. Proteins were separated on SDS ± PAGE and revealed by a Western blot with antibodies to phosphotyrosine, to Myc, to HA and to Grb2. (b) 293 were transfected with pcDNA3-IR, without or with pcDNA3-Grb10-HA or pcDNA3-Grb2. Cells were stimulated with 100 nM insulin for 10 min at 378C. Proteins were separated on SDS ± PAGE and revealed by a Western blot with antibodies to phosphotyrosine, to insulin receptor, to HA and to Grb2

Oncogene Involvement of Grb10 in VEGF signaling pathway S Giorgetti-Peraldi 3966 This stimulatory eect of Grb10 seems to be speci®c receptors (Frantz et al., 1997; Stein et al., 1996; Wang for KDR. Since Grb10 has been shown to interact with et al., 1999). The comparison of the sequences other tyrosine kinase receptors, such as the insulin recognized by Grb10 SH2 domain does not provide receptor, we expressed the insulin receptor with Grb10 information concerning its binding speci®city. KDR or Grb2 (Figure 10b). We observed that insulin autophosphorylates on four tyrosine residues, Y951, treatment of cells induces an increase in insulin Y996,Y1054 and Y1059 (Dougher-Vermazen et al., 1994; receptor tyrosine phosphorylation, which is not Wu et al., 2000). However, mutation of several KDR changed when Grb10 or Grb2 are expressed in tyrosine residues, including autophosphorylation sites, combination with the insulin receptor. The amount of did not allow us to identify the binding site of Grb10 insulin receptor was not modi®ed by the expression of on KDR (data not shown). Grb2 or Grb10. Interestingly, ectopic expression of Grb10 in 293 cells induces an increase in KDR molecules and in tyrosine phosphorylation. Importantly, the same ob- Discussion servation is made in HUVEC. This eect is not mediated by the SH2 domain of Grb10 since VEGF is one of the major factors involved in expression of the mutant R462K leads to an increase angiogenesis. This growth factor is acting through in KDR molecules, and expression of the SH2 domain activation of its tyrosine kinase receptors Flt-1 and alone does not induce this increase. Since Grb10 R462K KDR, but the signaling pathways activated by these is not phosphorylated in response to VEGF, this receptors are poorly identi®ed. suggests that the tyrosine phosphorylation of Grb10 is Here, we show that VEGF induces an increase in not required for the action of Grb10 on KDR. The Grb10 mRNA level, and then stimulates tyrosine increase in KDR expression can be correlated with an phosphorylation of endogenous Grb10 in HUVEC. increase in MAP kinase activation. Moreover, this This phosphorylation is dependent on the intact Grb10 eect appears to occur speci®cally on KDR, since SH2 domain, since mutation of the SH2 domain expression of Grb10 does not change the amount nor abolishes Grb10 tyrosine phosphorylation in response the tyrosine phosphorylation of insulin receptors. To to VEGF. Until now, it was known that growth the best of our knowledge, this is the ®rst demonstra- factors, such as EGF or insulin, do not stimulate tion of an action of Grb10 on regulation of the amount tyrosine phosphorylation of Grb10, but rather induce and phosphorylation state of a tyrosine kinase its serine phosphorylation (Dong et al., 1997; Ooi et receptor. al., 1995). Our results show for the ®rst time that a To explain our results, we could hypothesize that growth factor, VEGF, acting through activation of its expression of Grb10 would prevent receptor degra- tyrosine kinase receptor, is able to induce Grb10 dation leading to an increase in receptors at the cell tyrosine phosphorylation. Src is involved in VEGF- surface, by a mechanism which remains to be mediated angiogenesis and has been shown to be identi®ed. A protein ubiquitin ligase NEDD4 required for VEGF-mediated PLCg activation (Eliceiri associates with the SH2 domain of Grb10. However, et al., 1999; He et al., 1999). Moreover, Src it is unlikely that Grb10 prevents KDR degradation phosphorylates Grb10 (Langlais et al., 2000). Here, by recruiting NEDD4, since we have observed that we have found that VEGF-induced tyrosine phosphor- the Grb10 SH2 domain is not required to mediate ylation of Grb10 is partially due to Src. Indeed, the increase in KDR molecules. Expression of Grb10 treatment of cells with a Src inhibitor did not leads also to enhanced tyrosine phosphorylation of completely abolish the tyrosine phosphorylation of KDR in both basal and VEGF-stimulated condi- Grb10. These results suggest that Grb10 is a substrate tions. The basal tyrosine phosphorylation of KDR is for Src, but also a substrate for KDR. The latter would due to the tyrosine kinase activity of the receptor suggest that a direct interaction between KDR and itself, since expression of a kinase-dead KDR Grb10 could occur. Indeed, in the yeast two hybrid abolishes the basal tyrosine phosphorylation of the system, we detected an interaction between Grb10 and receptor after Grb10 expression. We propose that KDR. However, we were unable to show coprecipita- the increased number of KDR molecules at the tion between Grb10 and KDR either in HUVEC or in plasma membrane will lead to receptor dimerization, transfected cells (data not shown). This may be due to which induces receptor activation and autophosphor- limitations inherent to the immunoprecipitation tech- ylation. Moreover, it is possible that the receptors nique. Using the yeast two hybrid system, it has been will not be maintained in an inactivated state, due shown that IRS-1 and Shc associate directly with the to a reduced activity of phosphotyrosine phospha- insulin receptor (Gustafson et al., 1995; Tartare- tases which are in limited amount compared to the Deckert et al., 1995), while it is impossible to co- receptors. immunoprecipitate these proteins. In yeast two hybrid Since VEGF stimulates Grb10 mRNA expression, it system, we show that the interaction between Grb10 is tempting to propose that during neovascularisation, and KDR is dependent on the receptor tyrosine kinase VEGF interacts with its receptors, and activates activity and on the integrity of the SH2 domain of signaling pathways which lead to an increase in Grb10. Grb10 SH2 domain recognizes speci®c tyrosine Grb10 proteins. Grb10 in turn would upregulate phosphorylated sequences in several tyrosine kinase VEGF receptors.

Oncogene Involvement of Grb10 in VEGF signaling pathway S Giorgetti-Peraldi 3967 Taken together, our results suggest the existence of a Human Umbilical Vein Endothelial Cells (HUVEC) were positive feed-back loop in which VEGF induces isolated from umbilical cords by digestion with collagenase increased Grb10 expression, and in which Grb10 (Barbieri et al., 1981), or were obtained from Clonetics increases the level of VEGF receptors. Under physio- (Walkersville, MD, USA). Cells were grown in EBM-2 logical conditions, this loop would provide a way to supplemented with endothelial cell medium (Clonetics, Walkersville, MD, USA). amplify the angiogenic action of VEGF.

Transient transfection Transfection of 293 EBNA were performed by calcium Materials and methods phosphate precipitation (10 mg of DNA/9.5 cm2 dish). Sixteen hours after transfection, the calcium phosphate-DNA Materials precipitates were removed, and cells were incubated in VEGF was purchased from Peprotech (Tebu, France). DMEM containing 5% fetal calf serum (v/v). Before use, Antibodies to MAPK (06-182) and to phosphotyrosine (clone cells were serum-starved for 16 h in DMEM containing 0.2% 4G10) were purchased from UBI (Lake Placid, NY, USA), BSA (w/v). and antibodies to Grb2 (sc-255), Grb10 (sc-1026), and insulin Transfection of HUVEC were performed using Superfect receptor (sc-711-G) were purchased from Santa Cruz (Santa kit according manufacturer's instructions (Qiagen, Courta- Cruz, CA, USA). Antibodies to HA (clone 12CA5) were boeuf, France). obtained from Roche Diagnostics (Meylan, France). Anti- bodies to Shc were obtained from BD Transduction Immunoprecipitation and Western blot Laboratories (Franklin Lakes, NJ, USA). Antibodies to KDR were obtained from Sigma (St. Louis, MO, USA). Serum-starved cells were treated by ligands, chilled to 48C Insulin was a kind gift from Novo Nordisk (Copenhagen, and washed with ice-cold PBS (140 mM NaCl, 3 mM KCl, Denmark). 6mM Na2HPO4,1mM KH2PO4, pH 7.4), and solubilized Synthetic de®ned dropout yeast media lacking the with lysis buer (50 mM HEPES, 150 mM NaCl, 10 mM appropriate amino acids were obtained from BIO101 (La EDTA, 10 mM Na4P2O7, 100 mM NaF, 2 mM Vanadate, Jolla, CA, USA). Oligonucleotides were purchased from Life 0.5 mM PMSF, 10 mg/ml aprotinin, 10 mg/ml leupeptin, Technologies (Gaithersburg, MD, USA), and enzymes were pH 7.4, 1% Triton X-100) for 15 min at 48C. Cell from New England Biolabs (Beverly, MA, USA). PP2 was homogenates were clari®ed by centrifugation at 15 0006g purchased from Calbiochem (La Jolla, CA, USA), and all for 10 min at 48C. Cell lysates were incubated with antibodies chemical reagents were from Sigma (St. Louis, MO, USA). and protein-G Sepharose. The pellets were washed twice with lysis buer. Proteins were separated by SDS ± PAGE and transferred by electroblotting to poly(vinylidene di¯uoride) DNA plasmids membranes (Immobilon PVDF, Millipore, Bedford, MA, KDR cDNA was obtained from Masabami Shibuya USA). The membranes were soaked ®rst in blocking buer (University of Tokyo, Japan). The cytoplasmic region of (20 mM Tris, pH 7.4, 137 mM NaCl, 0.1 % Tween 20) KDR (aa 790 ± 1357) was subcloned into pLexA vector. Full- containing 5% (w/v) bovine serum albumin, and second in length KDR cDNA was subcloned in pcDNA3-MycA vector blocking buer containing speci®c antibodies. After washes, (Invitrogen, San Diego, CA, USA). Grb10 cDNA (nucleotide proteins were detected using horseradish peroxidase-linked 1 ± 1612) was subcloned in pACTII vector. An epitope tag secondary antibodies and enhanced chemiluminescence HA (YPYDVPDYAGPAAYPYDVPDYAGP) was added at (Amersham Pharmacia Biotech, Uppsala, Sweden). When the 3' end of the Grb10 cDNA, and was subcloned in indicated, Western blots were revealed by ECL plus on Storm pcDNA3 vector. 840, Molecular Dynamics (Amersham Pharmacia Biotech) All KDR and Grb10 point mutants were generated by the and analysed by Image Quant software (Amersham Pharma- Quickchange site-directed mutagenesis kit (Stratagene, San cia Biotech). Diego, CA, USA). Sequences of the constructs were checked by sequence analysis. MBP kinase assay Cell lysates were subjected to immunoprecipitation using Yeast strains, transformation and b-galactosidase assay antibodies to MAP kinase. After immunoprecipitation, pellets The yeast strain L40 (MAT a, trp1, leu2, his3, LYS2 ::lexA- were washed with lysis buer, and with HNTG (30 mM His3, URA3 ::lexA-lacZ) were obtained from A Vojtek (Fred HEPES, 30 mM NaCl, 0.1% Triton X-100, 10% glycerol). Hutchinson Cancer Research Center, Seattle, WA, USA). MBP phosphorylation were performed for 20 min at room Plasmid DNA transformations were performed using the temperature in presence of 150 mg/ml of myelin basic protein lithium acetate method. Cotransformants were selected on (MBP), 10 mM MgAc, and [g-32P]ATP (5 mM, 33 Ci/mmol). Trp-, Leu- plates. The transformants were tested for b- The phosphorylation was stopped by spotting an aliquot of galactosidase activity by liquid culture assays using the the samples on phosphocellulose paper ®lters (Whatman P81, chlorophenol red-b-D-galactopyranoside as described by Clifton, NJ, USA). After washes in 1% orthophosphoric Miller (1972). acid, radioactivity associated with the dried paper ®lters was determined by Cerenkov counting. Cell culture Northern blot analysis Human embryonic kidney cells, 293 EBNA, were maintained in culture in Dulbecco's modi®ed Eagle's medium (DMEM) Total RNA was isolated from 100 mm dishes using the Trizol containing 5% (vol/vol) fetal calf serum (Hyclone) and reagent according to the manufacturer's instructions (Life 500 mg/ml Geneticin. Technologies, Inc., Gaithersburg, MD, USA). Ten mg of total

Oncogene Involvement of Grb10 in VEGF signaling pathway S Giorgetti-Peraldi 3968 RNA was denaturated in formamide and formaldehyde, and Acknowledgments separated by electrophoresis in formaldehyde containing We thank Masabami Shibuya for KDR cDNA. We are agarose gel. RNA were transferred to Hybond-N membranes grateful to Pascal Peraldi for critical reading of the (Amersham Pharmacia Biotech, Uppsala, Sweden). Probes manuscript, and to Chantal Filloux and Franck Lafuente were labeled with g-32P-dCTP by random priming using the for technical assistance. JC Mas was supported by a Rediprime kit (Amersham Pharmacia Biotech, Uppsala, fellowship of AcadeÂ mie Nationale de MeÂ decine (Paris, Sweden) and puri®ed with the Probequant kit (Amersham France). This research was supported by funds from Pharmacia Biotech). Hybridizations were performed at 428C INSERM,AssociationpourlaRecherchecontreleCancer in NorthernMax Hybridization buer (Ambion, Inc, Austin, (ARC grant 5492). TX, USA). Membranes were washed in 16SSC, 0.5% SDS and revealed on Storm 840, Molecular Dynamics (Amersham Pharmacia Biotech).

References

Abedi H and Zachary I. (1997). J. Biol. Chem., 272, 15442 ± Morrione A, Plant P, Valentinis B, Staub O, Kumar S, Rotin 15451. D and Baserga R. (1999). J. Biol. Chem., 274, 24094 ± Bai RY, Jahn T, Schrem S, Munzert G, Weidner KM, Wang 24099. JY and Duyster J. (1998). Oncogene, 17, 941 ± 948. Morrione A, Valentinis B, Resnico M, Xu S and Baserga R. Barbieri B, Balconi G, Dejana E and Donati MB. (1981). (1997). J. Biol. Chem., 272, 26382 ± 26387. Proc.Soc.Exp.Biol.Med.,168, 204 ± 207. Moutoussamy S, Renaudie F, Lago F, Kelly PA and Finidori Carmeliet P. (2000). Nat. Med., 6, 389 ± 395. J. (1998). J. Biol. Chem., 273, 15906 ± 15912. Dong LQ, Du H, Porter SG, Kolakowski Jr. LF, Lee AV, Nantel A, Mohammad-Ali K, Sherk J, Posner BI and MandarinoLJ,FanJ,YeeD,LiuFandMandarinoJ. Thomas DY. (1998). J. Biol. Chem., 273, 10475 ± 10484. (1997). J. Biol. Chem., 272, 29104 ± 29112. Neufeld G, Cohen T, Gengrinovitch S and Poltorak Z. Dougher-Vermazen M, Hulmes JD, Bohlen P and Terman (1999). FASEB J., 13, 9 ± 22. BI. (1994). Biochem. Biophys. Res. Commun., 205, 728 ± OoiJ,YajnikV,ImmanuelD,GordonM,MoskowJJ, 738. Buchberg AM and Margolis B. (1995). Oncogene, 10, Eliceiri BP, Paul R, Schwartzberg PL, Hood JD, Leng J and 1621 ± 1630. Cheresh DA. (1999). Mol. Cell., 4, 915 ± 924. Risau W. (1997). Nature, 386, 671 ± 674. Ferrara N and Davis ST. (1997). Endo. Rev., 18, 4±25. Shalaby F, Rossant J, Yamaguchi TP, Gertsenstein M, Wu Fong GH, Rossant J, Gertsenstein M and Breitman ML. XF, Breitman ML and Schuh AC. (1995). Nature, 376, (1995). Nature, 376, 66 ± 70. 62 ± 66. Frantz JD, Giorgetti-Peraldi S, Ottinger EA and Shoelson Shen BQ, Lee DY and Zioncheck TF. (1999). J. Biol. Chem., SE. (1997). J. Biol. Chem., 272, 2659 ± 2667. 274, 33057 ± 33063. Gerber HP, McMurtrey A, Kowalski J, Yan M, Keyt BA, Stein E, Cerretti DP and Daniel TO. (1996). J. Biol. Chem., Dixit V and Ferrara N. (1998). J. Biol. Chem., 273, 271, 23588 ± 23593. 30336 ± 30343. Takahashi T, Ueno H and Shibuya M. (1999). Oncogene, 18, Gustafson TA, He W, Craparo A, Schaub CD and O'Neill 2221 ± 2230. TJ. (1995). Mol. Cell. Biol., 15, 2500 ± 2508. Tartare-Deckert S, Sawka-Verhelle D, Murdaca J and Van He H, Venema VJ, Gu X, Venema RC, Marrero MB and Obberghen E. (1995). J. Biol. Chem., 270, 23456 ± 23460. Caldwell RB. (1999). J. Biol. Chem., 274, 25130 ± 25135. Waltenberger J, Claesson-Welsh L, Siegbahn A, Shibuya M He W, Rose DW, Olefsky JM and Gustafson TA. (1998). and Heldin CH. (1994). J. Biol. Chem., 269, 26988 ± 26995. J. Biol. Chem., 273, 6860 ± 6867. Wang J, Dai H, Yousaf N, Moussaif M, Deng Y, Boufelliga HiratsukaS,MinowaO,KunoJ,NodaTandShibuyaM. A, Swamy OR, Leone ME and Riedel H. (1999). Mol. Cell. (1998). Proc. Natl. Acad. Sci. USA, 95, 9349 ± 9354. Biol., 19, 6217 ± 6228. Langlais P, Dong L, Hu D and Liu F. (2000). Oncogene, 19, Wu LW, Mayo LD, Dunbar JD, Kessler KM, Baerwald MR, 2895 ± 2904. Jae EA, Wang D, Warren RS and Donner DB. (2000). Miller JH. (1972). Cold Spring Harbor Laboratory Press, J. Biol. Chem., 275, 5096 ± 5103. Cold Spring Harbor, NY.

Oncogene