Phosphorylation of Tyrosine 801 of Vascular Endothelial Growth Factor Receptor-2 Is Necessary for Akt-dependent Endothelial Nitric-oxide Synthase Activation and Nitric Oxide Release from Endothelial Cells*

Vascular endothelial growth factor (VEGF)-stimulated nitric oxide (NO) release from endothelial cells is mediated through the activation of VEGF receptor-2 (VEGFR-2). Herein, we have attempted to determine which autophosphorylated tyrosine residue on the VEGFR-2 is essential for VEGF-mediated endothelial nitric-oxide synthase (eNOS) activation and NO production from endothelial cells. Tyrosine residues 801, 1175, and 1214 of the VEGFR-2 were mutated to phenylalanine, and the mutated receptors were analyzed for their ability to stimulate NO production. We show, both in COS-7 cells cotransfected with the VEGFR-2 mutants and eNOS and in bovine aortic endothelial cells, that the Y801F-VEGFR-2 mutant is unable to stimulate NO synthesis and eNOS activation in contrast to the wild type, Y1175F-VEGFR-2, and Y1214F-VEGFR-2. However, the Y801F mutant retains the capacity to activate phospholipase C-γ in contrast to the Y1175F-VEGFR-2. Interestingly, the Y801F-VEGFR-2, in contrast to the wild type receptor, does not fully activate phosphatidylinositol 3-kinase or recruit the p85 subunit upon receptor activation. This results in a complete incapacity of the Y801F-VEGFR-2 to stimulate Akt activation and eNOS phosphorylation on serine 1179 in endothelial cells. In addition, constitutive activation of Akt or a phosphomimetic mutant of eNOS (S1179D) fully rescues the inability of the Y801F-VEGFR-2 to induce NO release. Finally, we generated an antibody that specifically recognizes the phosphorylated form of tyrosine 801 of the VEGFR-2 and demonstrate that this residue is actively phosphorylated in response to VEGF stimulation of endothelial cells. We thus conclude that autophosphorylation of tyrosine residue 801 of the VEGFR-2 is essential for VEGF-stimulated NO production from endothelial cells, and this is primarily accomplished via the activation of phosphatidylinositol 3-kinase and Akt signaling to eNOS.

Vascular endothelial growth factor receptor-2 (VEGFR 3 -2/ Flk-1/KDR) is mainly responsible for the biological effects of VEGF in endothelial cells (1)(2)(3)(4). Other transmembrane tyrosine kinase receptors, such as VEGFR-1 and -3, have also been shown to transduce the intracellular signals of VEGF (1,5). The role of VEGFR-1 appears, however, to be mostly evident during embryonic angiogenesis, and VEGFR-3 signaling is restricted to lymphatic endothelial cells (5). Gene inactivation studies in mice have revealed an essential role for VEGFR-2 in vasculogenesis, since VEGFR-2 Ϫ/Ϫ animals die between embryonic day 8.5 and 9.5 due to a lack of vascular development similar to the phenotype observed in VEGF null mice (6,7).
Signal transduction of the VEGFR-2 in endothelial cells leads to the endothelial responses that are characteristic of proangiogenic factors. VEGF-dependent DNA synthesis and proliferation of endothelial cells are strongly dependent on the activation of extracellular signal-regulated kinases 1 and 2 by VEGFR-2 (2, 8 -10). Endothelial cell migration stimulated by VEGF involves both the phosphatidylinositol 3-kinase (PI3K) and the p38 mitogen-activated protein kinase pathways (11)(12)(13)(14). Activation of endothelial cell survival or antiapoptotic signaling by VEGF participates in the formation of new blood vessels and in the maintenance of their integrity; VEGF inhibits endothelial cell apoptosis via the PI3K/Akt pathway (15).
VEGF-stimulated nitric oxide (NO) release from endothelial cells also depends on the activation of VEGFR-2. NO is an important endothelial mediator that intervenes at most stages of VEGF-initiated cellular responses: proliferation, survival, migration, and increase in vascular permeability (11, 16 -20). Moreover, eNOS-deficient mice exhibit reduced responses to VEGF both in terms of new blood vessel formation and increase in vascular permeability (21). VEGF stimulation of endothelial cells activates at least two signaling pathways that converge toward NO production. VEGF stimulates intracellular calcium mobilization in endothelial cells through the activation of phospholipase C-␥ (PLC-␥) (22). Simultaneously, the PI3K-dependent activation of Akt by VEGFR-2 is responsible for eNOS phosphorylation on serine 1179 (1177 in human eNOS), which leads to increased eNOS activity and NO release (23,24). Interestingly, the relative contribution of each pathway in VEGFR-2-stimulated NO release has not been directly investigated.
As for other receptor tyrosine kinases, phosphorylation of several tyrosine residues in the intracellular domain of the VEGFR-2 has been shown to be essential for the recruitment and activation of Src homology 2-bearing proteins implicated in intracellular signal transmission. Many of these ligand-induced autophosphorylated tyrosines have been directly identified through systematic phosphomapping of activated receptors or have been proposed to be phosphorylated based on their essential role in the activation of the defined signaling pathway by VEGF. Some of the proposed phosphorylated tyrosines residues are 801, 951, 996, 1008, 1054, 1059, 1175, and 1214 (8,(25)(26)(27)(28). Although some slight discrepancies are present in the literature on the implication of certain residues in VEGFR-2 signaling, a consensus emerged on the role of some. Tyrosine residues 1054 and 1059 in the kinase insert domain seem to be needed for maximal intrinsic VEGFR-2 kinase activity (29,30). Other phosphorylated residues have been linked to the association and activation of Src homology 2-containing adaptors, as is the case for Tyr 951 and the VRAP/TSad adaptor and Tyr 1175 for Shb and Sck adaptors (25,(31)(32)(33)(34). Phosphorylation of Tyr 1175 has been linked to PLC-␥ activation, cell proliferation, and more recently to vasculogenesis in mice (8,27,35,36). Phosphorylation of Tyr 1214 is needed for p38 mitogen-activated protein kinase-dependent actin remodeling and cellular migration (37). Finally, phosphorylation of Tyr 801 has been linked to PI3K association to the VEGFR-2 and its activation (28).
Herein, we have attempted to determine which phosphorylated tyrosine on the VEGFR-2 is essential for VEGF-mediated NO production from endothelial cells. Since eNOS activation is the result of at least two converging intracellular signals (PLC-␥ and PI3K), we also investigated which of these pathways, upstream of eNOS, are essential for its activation. Our results show, both in a reconstituted COS-7 cell system and in bovine aortic endothelial cells (BAEC), that NO synthesis induced by the mutant Y1175F and Y1214F-VEGFR-2 is similar to that of the wild type receptor. In contrast, mutation of tyrosine 801 results in a complete inhibition of VEGFR-2-induced NO synthesis. Furthermore, the Y801F-VEGFR-2, in contrast to the wild type receptor, cannot activate the PI3K/Akt signaling pathway and induce eNOS phosphorylation on serine 1179. However, the Y801F mutant can still activate PLC-␥ in contrast to the Y1175F-VEGFR-2. Finally, we demonstrate, using a phosphospecific antibody, that tyrosine 801 of the VEGFR-2 is actively phosphorylated in response to VEGF stimulation of endothelial cells. We thus conclude that autophosphorylation of the tyrosine residue 801 of the VEGFR-2 is essential for VEGF-stimulated NO production from endothelial cells, and this is primarily accomplished via the activation of PI3K and Akt signaling to eNOS.

EXPERIMENTAL PROCEDURES
Cell Culture-COS-7 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (Invitrogen), 2.0 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. Bovine aortic endothelial cells (BAEC) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (HyClone), 2.0 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. Endothelial cells were used at passage 7 or 8. For VEGF stimulations, cells were starved, 6 h or overnight, in Dulbecco's modified Eagle's medium supplemented with 2.0 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. Recombinant human VEGF obtained from the BRB Preclinical Repository of the NCI-Frederick Cancer Research and Development Center was used for cell stimulation throughout this study.
Nitric Oxide Release-To measure the amount of NO released from cells, samples of culture medium were taken and processed for the measurement of nitrite (NO 2 Ϫ ), the stable breakdown product of NO in aqueous solution, by NO-specific chemiluminescence using a NO analyzer (Ionics Instruments) as described previously (40). For the measurement of accumulated NO production in transfected COS-7 and BAEC, cell medium was taken 48 h post-transfection and first subjected to ethanol precipitation of proteins. VEGF-stimulated NO production from BAEC was performed 48 h post-transfection on serum-starved cells. The medium was processed for the measurement of nitrite following a 30-min VEGF (40 ng/ml) stimulation.
The rabbit anti-phospho-Tyr 801 -VEGFR-2 antibody was raised against a keyhole limpet hemocyanin-conjugated syn-thetic phosphopeptide (GpYLSIVMDPDELPLDEC, where pY represents phosphotyrosine; Sigma), and the obtained serum was affinity-purified using the SulfoLink kit (Pierce). The rabbit antiserum was first passed through a column linked to the corresponding unphosphorylated peptide (GYLSIVMDPDEL-PLDEC), and the flow-through of this step passed through a second column linked to the above phosphorylated peptide. Antibody specificity was verified following the purification steps of the serum using dot blot against the phosphorylated and unphosphorylated peptides.
Immunoprecipitations and Immunoblotting-Cells were solubilized with a lysis buffer containing 1% Nonidet P-40, 50 mM Tris-HCl, 125 mM NaCl, 0.1 mM EDTA, 0.1 mM EGTA, 0.1% SDS, 0.1% deoxycholic acid, 20 mM sodium fluoride, 1 mM sodium pyrophosphate, 1 mM sodium orthovanadate, and protease inhibitor mixture (Roche Applied Science). Soluble proteins were quantified by DC protein assay (Bio-Rad), and equal protein amounts were boiled in SDS sample buffer, separated by SDS-PAGE, and transferred onto a nitrocellulose membrane (Hybond ECL; GE Healthcare). For immunoprecipitations, soluble proteins were incubated with the primary antibody (2 g) at 4°C for 2 h. Protein A-Sepharose (Sigma) (50 l of a 50% slurry) was then added and incubated for an additional 1 h. The immune complexes were precipitated by centrifugation, washed three times with lysis buffer, boiled in SDS sample buffer, separated by SDS-PAGE, and transferred to nitrocellulose membranes. For Western blotting, antibody detection was performed by a chemiluminescence-based detection system (ECL; GE Healthcare) or by a LI-COR Odyssey infrared imaging system (LI-COR Biosciences) using Alexa 680-or Alexa 800labeled secondary antibodies (Invitrogen).
Immunofluorescence-BAEC were cultured on 0.1% gelatincoated coverslips. Serum-starved cells were stimulated for 5 min with VEGF (20 ng/ml). Cells were fixed for 20 min in phosphate-buffered saline containing 3.5% paraformaldehyde. Cells were rinsed with phosphate-buffered saline and permeabilized with 0.1% Triton in phosphate-buffered saline for 5 min. Fixed cells were blocked with 1% bovine serum albumin and then incubated for 1 h with the primary antibody (4.4 g/ml) in 0.1% bovine serum albumin in PBS (rabbit anti-PO 4 Ϫ -Tyr 801 -VEGFR-2). Bound primary antibody was visualized following 1 h of incubation using Alexa Fluor 568-labeled goat anti-rabbit antibody (1:500) (Molecular Probes). Mounted coverslips were observed using a Zeiss LSM 510 imaging system.
Statistical Analysis-Data were analyzed by analysis of variance, followed by Bonferroni's post hoc test. A p value less than 0.05 was considered as statistically significant.

VEGFR-2 Tyrosine
Mutants-VEGF-stimulated NO release from endothelial cells requires the autophosphorylation on tyrosine residues of the VEGFR-2. In order to identify the phosphorylated tyrosine residues of the VEGFR-2, essential for the activation of intracellular signaling pathways leading to eNOS activation and NO production, we generated punctual substitutions on the receptor. Tyrosine residues 801, 1175, and 1214, proposed in the literature as being potential autophosphorylation sites of the VEGFR-2 receptor, were changed to phenylala-nine (8,27,28,(35)(36)(37)41). In addition, a kinase-inactive receptor that has a lysine to arginine mutation at residue 868 in the ATP binding site of the tyrosine kinase domain was generated as a negative control. To confirm that the substitution did not alter the intrinsic kinase ability of the receptor, the level of total tyrosine phosphorylation of mutants versus wild type (WT) receptors was monitored. Cell lysates from COS-7 cells expressing either the wild type or mutant receptors (K868R, Y801F, Y1175F, Y1214F) were immunoprecipitated using an anti-VEGFR-2 antibody, and levels of tyrosine phosphorylation were evaluated by Western blot with a pan-anti-phosphotyrosine antibody (4G10). All VEGFR-2 constructs were expressed, and autophosphorylation on tyrosine was detected for all of the tyrosine substitution mutants and wild type receptors in contrast to the tyrosine kinasedead receptor K868R (Fig. 1A). This indicates that the overall tyrosine kinase activity of the tyrosine-mutated receptor was not affected by the mutations. Ϫ -Tyr antibody, and the same membrane was then stripped and reprobed with anti-VEGFR-2 antibody to confirm protein levels. B, COS-7 cells were transfected with eNOS in the absence or in presence of the WT or mutant VEGFR-2 as indicated. Following a 24-h incubation, samples of culture medium were subjected to NO quantification as described under "Experimental Procedures." Expression levels of the transfected proteins were monitored by Western blotting using anti-VEGFR-2 and anti-eNOS antibodies. To allow statistical analysis and for comparison of multiple experiments, the relative increase in nitrite compared with eNOS alone is expressed. These results represent the mean of four experiments performed in triplicates. *, p Ͻ 0.01 and not significant (n/s) versus WT VEGFR-2-expressing cells.
Tyrosine 801 of the VEGFR-2 Is Essential for Stimulation of NO Release-We then sought to identify the tyrosine residues of the VEGFR-2 whose phosphorylation is essential for the activation of eNOS. First, we used a transfected COS-7 cells system to compare summarily the capacity of the various VEGFR-2 mutants (K868R, Y801F, Y1175F, or Y1214F) to stimulate NO release. Following a 24-h incubation, samples of culture medium from COS-7 cells, transfected with either eNOS alone or eNOS in combination with the wild type or the mutated receptors, were quantified for the amounts of accumulated NO 2 Ϫ . Nitrites are the major nitric oxide oxidation byproduct in aqueous solution (42). First, higher levels of NO 2 Ϫ were detected in the medium of COS-7 cells expressing eNOS than in the medium of untransfected cells (Fig. 1B, columns 1 and 2). Moreover, the cotransfection of wild type VEGFR-2 receptor with eNOS induced a significant augmentation in the production of NO compared with the production of cells expressing the enzyme alone (Fig. 1B, columns 2 and 3). In contrast, expression of the kinase-inactive receptor (K868R) did not produce this effect (Fig. 1B, column 4). These results are consistent with the fact that the VEGFR-2 receptor has the capacity of inducing NO production (43) and show that this COS-7 heterologous expression system is suitable for monitoring of eNOS activation by the VEGFR-2. Interestingly, mutation of the VEGFR-2 tyrosine 801 to phenylalanine resulted in decreased NO production compared with the wild type VEGFR-2. The extent of NO synthesis induced by the Y801F mutant was similar to the NO released from cells transfected with the kinase-inactive receptor K868R or eNOS alone, meaning that the Y801F mutation totally abolished the VEGFR-2stimulated NO production. In contrast, mutation of tyrosines 1175 and 1214 did not significantly affect the amounts of NO released by the transfected cells when compared with the amounts released from the wild type VEGFR-2-transfected cells. Thus, these results suggest an essential role for tyrosine 801 of the VEGFR-2 in the stimulation of eNOS-dependent NO release. The slight, but not significant, reduction in NO release by the 1175 and 1214 mutants compared with wild type receptor may point toward a secondary role for these residues in eNOS signaling, in sharp contrast to Tyr 801 , which is shown to be essential. In addition, we cotransfected with eNOS a Y1059F-VEGFR-2 mutant that has been previously shown to have altered overall tyrosine kinase activity (results not shown) (29,30), and the amounts of NO released from these transfected cells were similar to the kinase-dead K868R and Y801F mutants (results not shown).
To confirm the involvement of Tyr 801 in signal transduction leading to eNOS activation, we characterized the impact of this mutation on VEGF-induced NO generation in an endothelial cell model. BAEC were transiently transfected with either an empty vector (pcDNA3.1), wild type, or Y801F or Y1175F mutated VEGFR-2. Transfected BAEC were serum-starved for 6 h and stimulated for 30 min with VEGF (40 ng/ml), and medium samples were collected for the quantification of the amount of nitrite present. Fig. 2 shows the VEGF-stimulated increase in NO levels from the transfected BAEC. As expected, stimulation of the endogenous VEGFR-2 receptor in untransfected cells (mock) resulted in increased basal NO production.
Transfection of the WT VEGFR-2 induced a marked increase in VEGFR-2 protein levels, as revealed by Western blotting, and no change was observed on eNOS protein levels (Fig. 2, bottom). The overexpression of the VEGFR-2 resulted in a slight increase, following VEGF stimulation, in the amounts of nitrite accumulated in the culture medium compared with untransfected cells (Fig. 2). Similarly, the mutation of tyrosine 1175 did not affect the VEGF-stimulated increase in NO production when compared with cells overexpressing the wild type receptor (Fig. 2, columns 2 and 4). In contrast, VEGF stimulation of BAEC overexpressing the mutant Y801F receptor led to a significant decrease in NO production compared with cells overexpressing the wild type and Y1175F receptors (Fig. 2, column  3). Interestingly, NO released from cells expressing the Y801F mutant was less than from cells only expressing endogenous VEGFR-2 (Fig. 2, compare columns 1 and 3). Based on this result, it seems that the mutant Y801F receptor possesses the ability to act as a dominant negative on the endogenous VEGFR-2 receptor and thus reduce eNOS activation by VEGF. Overall, these results allow us to confirm the tyrosine 801 residue of the VEGFR-2 as being essential for VEGF-dependent eNOS activation and NO synthesis from endothelial cells.
Phosphorylation of PLC-␥ by the VEGFR-2 Mutants-As previously mentioned, the activation of eNOS by the VEGFR-2 is regulated mainly by two intracellular signaling cascades: the PI3K/Akt pathway and the increase in intracellular free calcium initiated by the activation of PLC-␥. We first analyzed the activation of PLC-␥ by the VEGFR-2 by monitoring the phospho- rylation level of PLC-␥ by Western blotting COS-7 cell lysates expressing the wild type or mutant Y801F, Y1175F, or Y1214F receptors. As expected, expression of a wild type VEGFR-2 receptor in COS-7 cells results in endogenous PLC-␥ phosphorylation (Fig. 3A, middle). In contrast, the expression of the kinase-inactive receptor K868R did not induce PLC-␥ phosphorylation. Expression of the mutant Y801F and Y1214F receptors increased PLC-␥ tyrosine phosphorylation to levels similar to those induced by the wild type VEGFR-2. Conversely, the Y1175F mutation prevented the VEGFR-2-dependent PLC-␥ phosphorylation, since COS-7 cells expressing this mutant displayed levels of PLC-␥ phosphorylation similar to those induced by the K868R inactive kinase mutant and that were significantly different from the wild type receptor. These results were confirmed by the assessment of the VEGFR-2-dependent phosphorylation of PLC-␥ in endothelial cells. BAEC overexpressing the WT VEGFR-2 receptor or the mutants Y801F and Y1175F were stimulated with VEGF (40 ng/ml; 5-15 min), and PLC-␥ activation was monitored via the levels of tyrosine 783 phosphorylation. Fig. 3B shows that VEGF stimulation of transfected BAEC, overexpressing the wild type VEGFR-2 receptor, resulted in a significant increase in PLC-␥ phosphorylation. As seen above for the transfected COS-7 cells (Fig. 3A), the Y801F mutant was still able to mediate a significant VEGFdependent increase in PLC-␥ phosphorylation at 5 and 15 min of stimulation similarly to the increase induced by the wild type receptor. In contrast, similar increases in the phosphorylation of PLC-␥ in response to VEGF stimulation were not observed in BAEC overexpressing the Y1175F-VEGFR-2. The residual PLC-␥ phosphorylation in these Y1175F-VEGFR-2-transfected BAEC comes from the endogenous wild type VEGFR-2, since this phosphorylation level of PLC-␥ is identical to the levels observed in VEGF-stimulated control untransfected BAEC (not shown). These results confirm the previously published data demonstrating the essential role of tyrosine 1175 phosphorylation in the activation of PLC-␥ by VEGF (8,27,35). In addition, our results also demonstrate that phosphorylation on tyrosine 801 of the VEGFR-2 is dispensable for the activating phosphorylation of PLC-␥.
Tyrosine 801 Is a Mediator for VEGFR-2-stimulated PI3K, Akt, and eNOS Phosphorylation-First, we monitored if the Y801F mutant could induce tyrosine phosphorylation of the regulatory subunit of PI3K, p85. COS-7 cells were cotransfected with FLAG-tagged p85 in the absence or in presence of WT or Y801F-VEGFR-2. Cell lysates were immunoprecipitated for the FLAG epitope, and tyrosine phosphorylation of p85 was monitored by Western blotting. Fig. 4A shows that expression of the WT VEGFR-2 induced a marked increase in the tyrosine phosphorylation levels of p85 when compared with cells expressing FLAG-p85 alone. In contrast, the Y801F mutant only induced 40% of the phosphorylation of p85 when compared with the WT VEGFR-2 (p Ͻ 0.05), indicating that tyrosine 801 is necessary for the full activation of PI3K by the VEGFR-2. We also monitored the phosphorylation-dependent association between p85 and the VEGFR-2. FLAG-p85 was thus immunoprecipitated from lysates of COS-7 cells expressing either the WT, Y801F, or K868R-VEGFR-2. Fig. 4B shows that the WT VEGFR-2 was present in the FLAG immunoprecipitate only if FLAG-p85 was cotransfected with the receptor, indicating a specific association between the two proteins (Fig. 4B,  lanes 1 and 2). Interestingly, when compared with the WT receptor, the Y801F-VEGFR-2 displayed a significantly reduced association with p85 (Fig. 4B, lanes 2 and 3). As expected, no association between the kinase-inactive receptor, K868R-VEGFR-2, and p85 was observed. Overall, these results are in agreement with the previously proposed participation of tyrosine 801 phosphorylation (799 in mouse) in the activation and docking of p85 to the VEGFR-2 (28) but suggest, however, that other tyrosine residues may also partially participate in the full activation of PI3K by the VEGFR-2.
The VEGF-dependent activation of Akt was then monitored in BAEC. As above, the WT and Y801F-VEGFR-2 were overexpressed in BAEC, and the activation of Akt, following VEGF stimulation, was monitored via serine 473 phosphorylation (Fig. 5A). As expected, VEGF stimulation of the endogenous VEGFR-2 receptor in untransfected cells resulted in Akt phosphorylation. This phosphorylation was significantly increased in BAEC overexpressing the wild type VEGFR-2 (Fig. 5A). However, the levels of VEGF-dependent Akt phosphorylation in BAEC overexpressing the Y801F mutant were similar to those observed in cells expressing only the endogenous receptor and significantly reduced when compared with cells overexpressing the WT VEGFR-2. This shows that the Y801F mutant is incapable of inducing Akt phosphorylation, since its overexpression does not result in an increase in Akt activation over the levels observed in untransfected cells.
We then monitored, by Western blotting, the Akt phosphorylation site on eNOS, serine 1179, in response to VEGF stimulation of BAEC (Fig. 5B). BAEC, either transfected with an empty vector or overexpressing the wild type and the mutant Y801F-VEGFR-2, were stimulated with VEGF (40 ng/ml; 5 min). Stimulation of control untransfected BAEC resulted, as expected, in eNOS phosphorylation on Ser 1179 , and this phosphorylation was significantly increased in cells overexpressing the WT VEGFR-2. BAEC overexpressing the Y801F mutant displayed eNOS phosphorylation levels, in response to VEGF stimulation, that were identical to those seen in mock-transfected cells and significantly different from levels observed in WT VEGFR-2 overexpressing. We can thus conclude that, as for Akt phosphorylation, the Y801F mutant is incapable of inducing eNOS phosphorylation on Ser 1179 , since its overexpression does not result in increased eNOS phosphorylation to levels above those seen in untransfected cells expressing only the endogenous receptor. Phosphorylation of the VEGFR-2 on tyrosine 801 is essential for the VEGF-dependent phosphorylation/activation of Akt and its substrate eNOS in endothelial cells.
eNOS Phosphorylation by Akt Restores NO Release-To further demonstrate that the inability of the Y801F mutant to stimulate NO release is due to its incapacity to induce eNOS phosphorylation by Akt, we verified if artificial activation of 1) Akt and 2) eNOS would rescue the phenotype induced by the mutation of tyrosine 801 of the VEGFR-2. Fig. 6A shows that the cotransfection of eNOS and the wild type VEGFR-2 in COS-7 cells resulted in increased NO production compared with eNOS expressed alone (white columns). As previously shown, the Y801F-VEGFR-2 mutant did not increase eNOSdependent NO production. Transfection of a constitutively activated form of Akt, myristoylated-Akt (myr-Akt), along with eNOS markedly increased NO production over eNOS expressed alone (Fig. 6A, black columns). Cotransfection of the  WT VEGFR-2 with eNOS and myr-Akt further increased NO release when compared with the levels of NO detected in eNOS-and myr-Akt-transfected cells. Interestingly, although the Y801F mutant did not potentiate NO release from eNOSexpressing cells, cotransfection of this mutant with eNOS and myr-Akt significantly increased NO production to levels seen in cells expressing the wild type VEGFR-2, eNOS, and myr-Akt. This suggests that the VEGFR-2-independent activation of Akt signaling reestablishes the capacity of the Y801F mutant to induce eNOS activation.
In a similar manner, we investigated if constitutively activated eNOS, using an aspartate substitution at residue Ser 1179 , which acts as a phosphomimetic, could rescue the Y801F-VEGFR-2 NO release defect (23,24,44). Transfection of COS-7 cells with wild type eNOS and the wild type or mutant (Y801F or Y1175F) VEGFR-2 gave similar results as previously (Fig.  1B); the wild type and Y1175F-VEGFR-2 both induced an increase in NO production when compared with cells expressing eNOS alone, whereas the Y801F mutation abolished this VEGFR-2-stimulated increase in NO synthesis (Fig. 6B, white  columns). However, when the different forms of the receptor (WT, Y801F, and Y1175F) were transfected in combination with the nonphosphorylatable S1179A mutant of eNOS, the increases in NO synthesis induced by the VEGFR-2 were not observable (Fig. 6B, black columns). As previously reported, elimination of the Akt phosphorylation site on eNOS has the ability to abolish the VEGFR-2-dependent stimulation of NO production (23,24). Furthermore, when the constitutively active S1179D-eNOS was transfected alone into COS-7 cells, the amount of NO released was, as expected, greater than that released by WT eNOS (Fig. 6B, gray columns) (23,24,44). Moreover, cotransfection of S1179D-eNOS with the WT VEGFR-2 resulted in a further increase in NO release, indicating that signaling events other than serine 1179 phosphorylation on eNOS contribute to VEGFR-2-dependent NO release. Cotransfection of the Y1175F-VEGFR-2 with S1179D-eNOS yielded similar amounts of NO when compared with the WT VEGFR-2. Interestingly, the Y801F-VEGFR-2 was able to stimulate NO production from S1179D-eNOS to levels identical to those of the wild type receptor. This clearly suggests that restoration of the phosphorylation of eNOS on the Ser 1179 Akt site allows for a complete rescue of the capacity of the Y801F-VEGFR-2 to stimulate NO release. This also confirms that the incapacity of the Y801F mutant to stimulate eNOS activation is solely due to its inability to induce eNOS phosphorylation on Ser 1179 .
Tyrosine 801 Is Actively Phosphorylated in Response to VEGF Stimulation-Finally, to demonstrate that tyrosine 801 of the VEGFR-2 is phosphorylated following activation of the receptor by VEGF and confirm that it plays a role in VEGF signaling, we have generated an antibody that recognizes the phosphorylated form of the receptor at tyrosine 801. Tyrosine residues 956, 1175, and 1214 of the VEGFR-2 have been directly identified as major autophosphorylation sites following VEGF stimulation (8,25). Although tyrosine 801 has been proposed as a VEGFR-2 activation site for the PI3K signaling pathway (28) and now eNOS activation, direct demonstration of its phosphorylation has never been achieved. To verify the specificity of our antibody for the phosphorylation status of Tyr 801 , we first performed dot blot analysis using the phosphorylated (GpYLSIVM-DPDELPLDEC) immunogenic peptides and nonphosphorylated (GYLSIVMDPDELPLDEC) version. The affinity-purified serum clearly recognized the phosphorylated version of the peptide with higher affinity than the nonphosphorylated version (Fig.   FIGURE 6. Constitutive phosphorylation of eNOS by Akt restores NO release stimulated by Y801F-VEGFR-2. A, COS-7 cells were transfected with either eNOS alone or in the presence of the various VEGFR-2 constructs (white bars). eNOS was also transfected with myr-Akt in the absence or presence of the indicated VEGFR-2 constructs (black bars). Samples of culture medium from the transfected cells were taken following a 24-h accumulation and were then subjected to nitrite quantification as described under "Experimental Procedures." Protein expression was verified by Western blot (wb) of cellular lysates using anti-VEGFR-2, anti-eNOS, and anti-Akt antibodies. For comparison between experiments, NO levels were normalized over the mean production of cells transfected with eNOS alone. *, p Ͻ 0.05 versus eNOS alone; †, p Ͻ 0.05 versus eNOS and myr-Akt-transfected cells. These results are the mean of at least three different experiments in triplicate. B, WT eNOS (white bars), S1179A eNOS (black bars), or S1179D eNOS (gray bars) were transfected in COS-7 cells either alone or in combination with the different VEGFR-2 constructs, as indicated. Following a 24-h accumulation, samples of culture medium were subjected to nitrite quantification, and protein expression levels were verified by Western blotting as in A. *, p Ͻ 0.05 versus S1179D eNOS-transfected cells. These results are the average of at least three experiments.
7A). Next, we transfected COS-7 cells with the wild type and mutant VEGFR-2 receptors (K868R, Y801F, Y1175F). Cell lysates were immunoprecipitated for the VEGFR-2 and Western blotted with the tyrosine 801-phosphospecific antibody. The antibody specifically detected phosphorylation in the wild type and the mutant Y1175F-VEGFR-2 but failed to recognize the kinase-inactive receptor K868R and minimally detected the Y801F-VEGFR-2 (Fig. 7B). In addition, we controlled for the levels of total tyrosine phosphorylation using the pan-phosphotyrosine antibody (4G10) and found, as shown in Fig. 1A, that the overall phosphorylation of the mutant receptor Y801F was similar to that of the mutant Y1175F and wild type receptors (Fig. 7B). The minimal phosphorylation detected by this antibody in the Y801F mutant is possibly due to residual nonspecific detection of other highly phosphorylated tyrosine residues on overexpressed receptor. It is, however, clear that this anti-body displays a marked specificity for the phosphorylated form of Tyr 801 .
To demonstrate that this residue on the VEGFR-2 is phosphorylated in endothelial cells in response to VEGF stimulation, BAEC were stimulated with VEGF (40 ng/ml) for 0, 3, 5, 15, or 30 min, and the receptor was immunoprecipitated and subjected to a Western blot using the phosphospecific Tyr 801 antibody. Fig. 7C shows an increase in Tyr 801 phosphorylation in response to VEGF in endothelial cells that reached maximal phosphorylation at 5 min of stimulation (Fig. 7C, top). The kinetics of tyrosine 801 phosphorylation followed the overall autophosphorylation of the VEGFR-2 as revealed by reprobing the membrane with the panphosphotyrosine antibody (4G10) (Fig. 7C, middle). As previously reported, VEGF stimulation of endothelial cells induced the down-regulation of the VEGFR-2, as revealed by reprobing of the immunoprecipitate with an anti-VEGFR-2 antibody (Fig. 7C, bottom) (38). Finally, we wanted to determine if this antibody could detect Tyr 801 phosphorylation in vivo in endothelial cells. Thus, we performed immunofluorescence studies on VEGF-stimulated and fixed BAEC. Phosphorylation of Tyr 801 of the VEGFR-2 was visualized by confocal microscopy. Fig.  7D shows that, in control serumstarved BAEC, the cellular levels of Tyr 801 phosphorylation are low, since the detected fluorescence is slightly above the levels seen in cells incubated with the secondary antibody alone. This is consistent with the results presented in Fig. 7C, where Tyr 801 and overall phosphorylation of the VEGFR-2 is undetectable in unstimulated cells. VEGF stimulation of BAEC for 5 min induced a marked increase in the fluorescence signal, revealing a punctate distribution of the signal. This intracellular phosphorylation of tyrosine 801 of the VEGFR-2 is in line with the recent reports of an important contribution of the intracellular endosomal compartment in VEGFR-2 signaling (45,46). The Tyr 801 phosphorylation staining is very similar to Tyr 1175 phosphorylation recently reported by Lampugnani et al. (46). This phosphospecific antibody allows us to clearly demonstrate that the tyrosine residue 801 of the VEGFR-2 is an important autophosphorylation site of this receptor following VEGF stimulation and, thus, can potentially participate in the VEGFR-2-in- FIGURE 7. Tyrosine residue 801 of the VEGFR-2 is phosphorylated following VEGF stimulation. A, the purified tyrosine 801 phosphospecific antibody (1:1000 dilution; 0.2 g/ml) was used for dot blot analysis of increasing amounts (0 -100 ng) of the phosphorylated immunogenic peptides and the nonphosphorylated version used for affinity purification of the serum. B, the tyrosine 801-phosphospecific antibody was used for the Western blotting of VEGFR-2 immunoprecipitated from COS-7 cell lysates transfected with either an empty vector, the WT VEGFR-2, or the mutant receptors (K868R, Y801F, or Y1175F). The membrane was subsequently stripped and reprobed with the pan-anti-PO 4 Ϫ -Tyr antibody followed by the anti-VEGFR-2 antibody. C, serumstarved BAEC were stimulated with 40 ng/ml VEGF for 0, 3, 5, 15, and 30 min, and cell lysates were subjected to immunoprecipitation using anti-VEGFR-2 or nonimmune (NI) antibodies. Immunoprecipitates were then subjected to Western blot using the tyrosine 801-phosphospecific antibody followed by the pan-anti-PO 4 Ϫ -Tyr and anti-VEGFR-2 antibodies. D, BAEC were stimulated with 20 ng/ml VEGF for 0 (control) or 5 min, fixed, permeabilized, and incubated in the presence or absence of the tyrosine 801-phosphospecific antibody. Immunocomplexes were detected using a secondary antibody labeled with Alexa-568 and were visualized by confocal laser microscopy. These results are representative of three independent experiments. duced cellular response, leading to eNOS activation and NO release from endothelial cells.

DISCUSSION
The results presented herein show that 1) tyrosine 801 of the VEGFR-2 is necessary for VEGF-stimulated NO release, 2) Tyr 801 is required for maximal phosphorylation and recruitment of the p85 subunit of PI3K to the VEGFR-2, 3) both Akt and eNOS phosphorylation, following VEGFR-2 activation, require Tyr 801 , 4) the incapacity of the Y801F-VEGFR-2 to stimulate NO release can be rescued solely by the artificial phosphorylation of eNOS on the Akt site, Ser 1179 , and finally 5) tyrosine 801 on the VEGFR-2 is dynamically phosphorylated in endothelial cells in response to VEGF stimulation. Overall, we demonstrate that Akt-mediated phosphorylation of eNOS on Ser 1179 following VEGFR-2 activation is an essential event for NO release, and this is dependent on the phosphorylation of tyrosine 801 of the VEGFR-2.
Our results identify eNOS phosphorylation on serine 1179 by Akt as an obligatory event for eNOS activation and NO production in response to VEGF and suggest that other VEGF signaling pathways, such as PLC-␥ activation, are not primary events for eNOS activation. Mutation of Tyr 801 to phenylalanine completely abrogates VEGFR-2-stimulated NO release, both in a reconstituted COS-7 cell system and in endothelial cells. Furthermore, we show that the Y801F mutant is incapable of inducing Akt phosphorylation on Ser 473 and eNOS phosphorylation on Ser 1179 , suggesting that the incapacity of the mutated receptor to induce NO release is due to these incomplete signaling events. Interestingly, the Y801F VEGFR-2 mutant is fully capable of promoting PLC-␥ phosphorylation in contrast to the Y1175F mutant further, indicating that eNOS phosphorylation by Akt is a prerequisite for NO release. Other signaling pathways are indisputably involved in VEGF-stimulated NO production, highlighted here by the fact that the VEGFR-2 is still capable of potentiating NO release from the constitutively activated S1179D-eNOS (Fig. 6B). Increases in intracellular calcium and PLC-␥ activation have clearly been involved in VEGF-stimulated NO release (22). However, our results now allow us to uncouple these two pathways and demonstrate that, even in the context of an activated PLC-␥, VEGFR-2-stimulated NO release does not occur if phosphorylation of eNOS on Ser 1179 by Akt following Tyr 801 phosphorylation of the receptor is not fulfilled. One could imagine that in the absence of an elevation of the intracellular levels of calcium by PLC-␥, VEGF-induced phosphorylation of Ser 1179 of eNOS is sufficient to promote NO release similarly to fluid shear stress. Our results may also indicate that the late ("calciumindependent") phase of eNOS activation by VEGF predominates, in contrast to the early phase (calcium-dependent), in the overall amounts of NO released by endothelial cells following VEGF stimulation (47,48).
While this manuscript was in review, an interesting study by Ahmad et al. (49) reported that mutation of tyrosine 951 of the VEGFR-2 results in reduced stimulated NO release. Using the chimeric EGF-VEGF receptor, they demonstrate that phosphorylation of tyrosine 951 of the VEGFR-2 is involved in eNOS regulation through indirect activation of Akt by PLC-␥. This mutation results in a reduction by about 60% of the VEGFR-2stimulated NO release. Because of this incomplete inhibition of the stimulated NO release, the authors conclude that other tyrosine residues are involved in eNOS activation by the VEGFR-2. Indeed, our study now demonstrates that phosphorylation of Tyr 801 is a major player for this signaling pathway via direct Akt phosphorylation, since its mutation results in a complete inhibition of eNOS phosphorylation and activation. Interestingly, the same study by Ahmad et al. (49) reveals that the tyrosine 794 residue of the VEGFR-1 is essential for eNOS activation by this receptor, which they demonstrate for the first time capable of promoting NO release by VEGF. Moreover, they demonstrate that Tyr 794 of the VEGFR-1 mediates eNOS activation through Akt phosphorylation. This is interesting, because tyrosine 794 of the VEGFR-1 is analogous to tyrosine 801 of the VEGFR-2. Up to now, tyrosine 801 of the VEGFR-2 has not been investigated for its role in eNOS activation. Together, our two studies demonstrate that both VEGF receptors utilize a common signaling mechanism to stimulate NO production from endothelial cells and that eNOS phosphorylation by Akt is a central and mandatory event for eNOS-mediated NO release.
Phosphorylation of tyrosine 801 of the VEGFR-2 has previously been reported as being involved in PI3K activation (28). We show that the mutation of Tyr 801 to Phe of the VEGFR-2 markedly reduces p85 phosphorylation and recruitment; however, both phenomena are not completely abolished. This suggests that the phosphorylation of other tyrosine residues does participate in the recruitment and activation of p85 to the receptor perhaps through adaptor proteins, such as TSAd, known to interact with phosphorylated Tyr 951 , or the recently identified Gab1 (25,31,50). Nonetheless, our results show that a reduction in the activation levels of the p85 subunit of PI3K by the VEGFR-2 leads to a complete disruption of the downstream signaling events, namely Akt and eNOS, resulting in the inhibition of NO release.
Phosphorylation of tyrosine 1175 of the VEGFR-2 has been shown to be essential for embryonic vasculogenesis in mice. Indeed, mice carrying a phenylalanine substitution of this residue die in utero because of a failure of endothelial cell proliferation and differentiation similarly to the VEGFR-2 null mice (36). Our study reveals that phosphorylation of this residue is not essential for NO release, which is in line with a predominant role for NO in physiologic and pathologic VEGF-driven angiogenesis in adults, suggesting that the phosphorylation of the two residues (801 and 1175) might participate in different components of VEGF-mediated vessel formation.
In summary, we have identified phosphorylation of tyrosine 801 of the VEGFR-2 as the earliest intracellular event essential for VEGF signal transduction to eNOS via PI3K/Akt. We also reveal that eNOS phosphorylation on Ser 1179 is a crucial event that is essential for VEGF-stimulated release of an important modulator of VEGF effects on endothelial cells, nitric oxide.
Acknowledgments-We thank Dr. Jean-Francois Côté for critical reading of the manuscript and Daniela Baggio for secretarial assistance.