Selective Binding of VEGF to One of the Three Vascular Endothelial Growth Factor Receptors of Vascular Endothelial Cells

VEGF and VEGF are vascular endothelial growth factor splice variants that promote the proliferation of endothelial cells and angiogenesis. VEGF contains the 44 additional amino acids encoded by exon 7 of the VEGF gene. These amino acids confer upon VEGF a heparin binding capability which VEGF lacks. I-VEGF bound to three vascular endothelial growth factor (VEGF) receptors on endothelial cells, while I-VEGF bound selectively only to the flk-1 VEGF receptor which corresponds to the larger of the three VEGF receptors. The binding of I-VEGF to flk-1 was not affected by the removal of cell surface heparan sulfates or by heparin. Both VEGF and VEGF inhibited the binding of I-VEGF to a soluble extracellular domain of the flk-1 VEGF receptor in the absence of heparin. However, heparin potentiated the inhibitory effect of VEGF by 2-3-fold. These results contrast with previous observations which have indicated that the binding of I-VEGF to the flk-1 receptor is strongly dependent on heparin-like molecules. Further experiments showed that the receptor binding ability of VEGF is susceptible to oxidative damage caused by oxidants such as HO or chloramine-T. VEGF was also damaged by oxidants but to a lesser extent. Heparin or cell surface heparan sulfates restored the flk-1 binding ability of damaged VEGF but not the receptor binding ability of damaged VEGF. These observations suggest that alternative splicing can generate a diversity in growth factor signaling by determining receptor recognition patterns. They also indicate that the heparin binding ability of VEGF may enable the restoration of damaged VEGF function in processes such as inflammation or wound healing.

to the flk-1 receptor is strongly dependent on heparin-like molecules. Further experiments showed that the receptor binding ability of VEGF 165 is susceptible to oxidative damage caused by oxidants such as H 2 O 2 or chloramine-T. VEGF 121 was also damaged by oxidants but to a lesser extent. Heparin or cell surface heparan sulfates restored the flk-1 binding ability of damaged VEGF 165 but not the receptor binding ability of damaged VEGF 121 . These observations suggest that alternative splicing can generate a diversity in growth factor signaling by determining receptor recognition patterns. They also indicate that the heparin binding ability of VEGF 165 may enable the restoration of damaged VEGF 165 function in processes such as inflammation or wound healing.
Four vascular endothelial growth factor (VEGF) 1 forms are produced by alternative splicing from the VEGF gene (1)(2)(3). The 121-amino acid form (VEGF 121 ) lacks a heparin binding ability, while VEGF 165 , VEGF 189 , and VEGF 206 bind efficiently to heparin. All forms are mitogenic to vascular endothelial cells and induce permeabilization of blood vessels (4,5). VEGF 165 induces angiogenesis in vivo (6) and plays a central role in the process of tumor angiogenesis (7-9). Binding and cross-linking experiments have shown that VEGF 165 binds to three VEGF receptors on the cell surface of vascular endothelial cells (10 -12). Two VEGF receptors have recently been identified and cloned (13)(14)(15)(16). These include the mouse flk-1 receptor and its human homologue KDR, and the human flt-1 receptor. The flk-1 receptor was shown to transduce a VEGF 165 mitogenic signal (9,17), while activation of the flt-1 receptor does not seem to result in a similar mitogenic response (16,18).
The binding of 125 I-VEGF 165 to various VEGF receptors and the effects of heparin on the binding have been characterized extensively (10 -12, 19), while the interaction of VEGF 121 with the VEGF receptors of vascular endothelial cells has not yet been studied. We show that VEGF 121 binds selectively to the larger of the three VEGF receptors of human umbilical veinderived endothelial cells (HUE) and that this receptor probably corresponds to the human homologue of the flk-1 VEGF receptor. We also show that both VEGF 165 and VEGF 121 are susceptible to oxidative damage and that heparin restores the receptor binding ability of damaged VEGF 165 , but not the receptor binding ability of damaged VEGF 121 .

EXPERIMENTAL PROCEDURES
Materials-VEGF 165 and VEGF 121 were produced using the baculovirus system as described for VEGF 165 (20). VEGF 165 and VEGF 121 were purified as described previously (19,21). The flk-1/SEAP soluble receptor was produced as described (19). Anti-flk-1 and anti-flt-1 antibodies directed against peptides derived from the intracellular domains of the respective receptors were purchased from Santa Cruz Biotechnology Inc. Anti-alkaline phosphatase antibodies were purchased from Dako. Tissue culture media and reagents were from Biological Industries (Beth-Haemek Biological Industries, Kibbutz Beth Haenek, Israel) and tissue culture plasticware from Nunc. Disuccinimidyl suberate was from Pierce Chemical Co., 125 I-sodium from DuPont NEN, and heparin-Sepharose from Pharmacia. Recombinant bacterial heparinase-1 was kindly provided by Dr. Zimermann (Ibex Technologies, Montreal, Canada). High molecular mass protein markers were obtained from Bio-Rad. Porcine mucosa-derived heparin (H-7005) was purchased from Sigma as were all of the other chemicals.
Cell Culture-HUE cells were grown in gelatin-coated dishes in M-199 medium supplemented with 20% fetal calf serum, 4 mM glutamine, antibiotics, and 1 ng/ml bFGF which was added to the cells every other day.
Binding and Cross-linking Experiments-Iodination of human recombinant VEGF 165 or VEGF 121 was performed using either the chloramine-T method or the IODOGEN method with similar results as described previously (10,21,22). However, while 125 I-VEGF 165 was separated from free iodine using a heparin-Sepharose column as described (10), 125 I-VEGF 121 was separated from free iodine using size exclusion chromatography on Sephadex G-25 (21). The specific activities of the 125 I-VEGF 165 and the 125 I-VEGF 121 were about 10 5 cpm/ng.
The binding and the cross-linking of 125 I-VEGF 165 to endothelial cells was done as described previously (10,11), and the binding of 125 I-VEGF 121 was done similarly. Nonspecific binding was determined in the presence of 1-2 g/ml unlabeled VEGF. The level of nonspecific binding ranged between 10 and 20% of the total binding. The binding and cross-linking of 125 I-VEGF 121 to a flk-1/SEAP fusion protein containing the extracellular domain of the flk-1 receptor was done as described previously for 125 I-VEGF 165 (19,23).
Immunoprecipitation Experiments-125 I-VEGF 165 (5 ng/ml) or 125 I-VEGF 121 (10 ng/ml) were bound and cross-linked to confluent HUE cells grown in 6-cm dishes. The cells were lysed with 1 ml of lysis buffer containing 1% Nonidet P-40, 0.5% deoxycholate, and protease inhibitors. The lysates were centrifuged briefly to remove insoluble debris, and aliquots containing 125 I-VEGF-receptor complexes were taken for immunoprecipitation using various antibodies. Cell lysates were precleared by a 1-h incubation at 4°C with Sepharose CL-4B. The cell lysates were then incubated for 1 h at 4°C with 1 g/ml concentrations of various antibodies. Antibody-125 I-VEGF-receptor complexes were precipitated for 1 h at 4°C using protein G-Sepharose. The beads were washed 3 times with lysis buffer, and 125 I-VEGF-receptor complexes were then detached from the beads by boiling in SDS-PAGE sample buffer. 125 I-VEGF-receptor complexes were separated on a 6% gel and visualized as described.
Inactivation of VEGF by Oxidation-VEGF 121 or VEGF 165 (2 g) in 100 mM sodium phosphate buffer (pH 7.2) were incubated for 1 min with 0.1 mg/ml chloramine-T or 0.1-1% H 2 O 2 in a final reaction volume of 40 l. At the end of the incubation, sodium metabisulfite (5 l) was added to a final concentration of 0.5 mg/ml. After 2 min, bovine serum albumin was added to a final concentration of 1 mg/ml.

125
I-VEGF 121 Binds to One VEGF Receptor on HUE Cells, and the Binding Is Not Affected by Heparin-like Molecules-VEGF 121 and VEGF 165 were produced and purified from the conditioned medium of recombinant baculovirus-infected Sf9 cells and iodinated as described (19,21). The purified labeled VEGF forms (Fig. 1) were used further in binding and crosslinking experiments. When a saturating concentration of 125 I-VEGF 121 (20 ng/ml) was bound and cross-linked to HUE cells, only one 125 I-VEGF 121 -receptor complex of about 225 kDa could be detected. The formation of the 225-kDa 125 I-VEGF 121 -receptor complex was completely inhibited when the binding of 125 I-VEGF 121 was performed in the presence of either 2 g/ml unlabeled VEGF 121 or 2 g/ml unlabeled VEGF 165 (Fig. 2, lanes  2 and 3), indicating that this receptor binds both VEGF 121 and VEGF 165 . Addition of exogenous heparin (1 g/ml) to the binding reaction, or removal of cell surface-associated heparan sulfates using heparinase-1, did not affect significantly the formation of the 125 I-VEGF 121 -receptor complex (Fig. 2, lanes 4 and 5, respectively). Additional 125 I-VEGF 121 -receptor complexes were not detected when heparin was added to the binding reaction. In contrast, when 5 ng/ml 125 I-VEGF 165 were bound to the cells in the presence of 1 g/ml heparin, two smaller 125 I-VEGF 165 -receptor complexes were observed (Fig. 2, compare lane 7 to lane 8). 125 I-VEGF 165 also bound to the two smaller VEGF receptors in the absence of added exogenous heparin provided that higher 125 I-VEGF 165 concentrations were used (11). In addition, larger 125 I-VEGF 165 containing cross-linked complexes of about ϳ400 kDa were observed (Fig. 2, lane 8). Such high molecular weight complexes were observed by us in the past (23), and they may represent dimerized receptors. Competition with 2 g/ml VEGF 121 inhibited the formation of the 225-kDa 125 I-VEGF 165 -receptor complex but did not affect the formation of the two smaller 125 I-VEGF 165 -receptor complexes (Fig. 2, lane 10, and Fig. 3). The formation of the two smaller 125 I-VEGF 165 -receptor complexes could not be inhibited significantly even when VEGF 121 concentrations as high as 20 g/ml were added to a binding reaction in which 125 I-VEGF 165 was bound to the cells in the presence of heparin (not shown). It therefore seems that VEGF 121 is not able to bind to the two smaller VEGF receptors, or alternatively, that the affinity of VEGF 121 to these receptors is much lower than that of VEGF 165 . The experiments therefore suggest that VEGF 121 and VEGF 165 differ not only with regard to their heparin binding ability, but also in their ability to recognize various species of VEGF receptors.
The binding of 125 I-VEGF 121 to the larger VEGF receptor of the endothelial cells was not affected by the removal of cell surface heparin-like molecules (Fig. 2, lane 5). This was perhaps to be expected as 125 I-VEGF 121 does not bind to heparin. This observation also indicates that heparin-like molecules do not affect the VEGF 121 binding ability of the larger VEGF receptor of the endothelial cells. However, not all of the VEGF receptors that are capable of 125 I-VEGF 121 binding behave similarly. The 125 I-VEGF 121 binding ability of the VEGF receptors of YU-ZAZ6 melanoma cells is inhibited upon the removal of cell surface heparin-like molecules by heparinase digestion or by the addition of exogenous heparin, suggesting that heparin-like molecules can modulate the VEGF 121 binding ability of the YU-ZAZ6 VEGF receptors (21).
Antibodies Directed against the flk-1 VEGF Receptor Immunoprecipitate 125 I-VEGF 121 -Receptor Complexes-In order to identify the 125 I-VEGF-receptor complexes seen in the crosslinking experiments, immunoprecipitation experiments were performed. The larger 225-kDa 125 I-VEGF 165 -or 125 I-VEGF 121labeled receptor was specifically immunoprecipitated by antibodies that recognize the intracellular domain of the flt-1 receptor (Fig. 4A, lane 4, and Fig. 4B, lane 2, respectively). In contrast, none of the labeled complexes could be precipitated by antibodies that recognize the intracellular domain of the flk-1 receptor or by an irrelevant antibody (Fig. 4). The inability to immunoprecipitate 125 I-VEGF-receptor complexes with antiflt-1 antibodies was expected since the mRNA encoding the flt-1 VEGF receptor is not very abundant in HUE cells (21). The two smaller 125 I-VEGF 165 -receptor complexes were not precipitated by any of these antibodies and may represent either novel VEGF receptors or truncated versions of known VEGF receptors which are not recognized by the antibodies used.

I-VEGF 121 Binds to a Soluble Fusion Protein Containing the Extracellular Domain of flk-1, and the Binding Is Inhibited by VEGF 121 and VEGF 165 Even in the Absence of Heparin-
The 225-kDa 125 I-VEGF 121 -receptor complex of HUE cells appears to contain the KDR/flk-1 receptor. To verify that 125 I-VEGF 121 can indeed bind to the flk-1 VEGF receptor, the interaction of VEGF 121 and VEGF 165 with a soluble fusion protein containing the entire extracellular domain of flk-1 fused to soluble alkaline phosphatase (flk-1/SEAP) was examined (Fig. 5, A-C) (19). This soluble fusion protein was adsorbed to ELISA dishes coated with an antibody to alkaline phosphatase and used in quantitative binding assays. As expected, 125 I-VEGF 121 bound to flk-1/SEAP, and the binding was inhibited by unlabeled VEGF 121 (Fig. 5B). The binding was effective even in the absence of exogenous heparin, and the addition of increasing concentrations of heparin did not affect the binding (not shown). However, to our surprise, we have found that unlabeled VEGF 165 was able to compete with 125 I-VEGF 121 for binding to the flk-1/SEAP fusion protein in the absence of exogenous heparin (Fig. 5A). Addition of heparin modulated the concentration at which a half-maximal displacement of bound 125 I-VEGF 121 was observed, shifting it to VEGF 165 concentrations 2-3-fold lower than those seen in the absence of heparin (not shown). Similar results were obtained when similar binding experiments were performed using heparinasedigested HUE cells (Fig. 5D). These observations are not in agreement with experiments which have indicated that unless  (19). These experiments also disagree with observations which have shown that the binding of 125 I-VEGF 165 to cell surface receptors on vascular endothelial cells requires cell surface heparan sulfates (Fig. 2, lanes 11 and 12) (11).
These experiments imply that the receptor binding ability of VEGF 165 may be impaired during iodination. Subsequent experiments have indicated that the flk-1/SEAP binding ability of VEGF 165 is sensitive to oxidants. VEGF 165 damaged by oxidizing agents such as chloramine-T or H 2 O 2 was not able to compete with 125 I-VEGF 121 for binding to flk-1/SEAP in the absence of heparin (Fig. 5, A and C). However, the ability to compete with 125 I-VEGF 121 for binding to flk-1/SEAP was partially restored by the addition of 1 g/ml heparin to the binding reaction (Fig. 5, A and C). Similar results were also obtained in analogous experiments performed with heparinase-treated HUE cells (not shown). Oxidized VEGF 121 also lost some of its ability to compete with 125 I-VEGF 121 for binding to flk-1/SEAP, although VEGF 121 seemed to be somewhat more resistant than VEGF 165 to oxidation (Fig. 5, B and C). However, in contrast to VEGF 165 , the addition of heparin did not restore the flk-1/ SEAP binding ability of damaged VEGF 121 (Fig. 5, B and C).
The relative insensitivity of VEGF 121 to oxidative damage may explain why 125 I-VEGF 121 , in contrast to 125 I-VEGF 165 , is still able to bind to flk-1/SEAP and to the flk-1 receptor of the vascular endothelial cells in the absence of heparin-like molecules.
The potentiating effect that heparin exerts on the receptor binding ability of untreated VEGF 165 could reflect oxidative damage sustained before or during VEGF 165 purification. Alternatively, it could mean that heparin has a real ability to potentiate the binding of undamaged VEGF 165 to the flk-1 VEGF receptor and perhaps to other types of VEGF receptors as well. Our results imply that the heparin binding ability of VEGF 165 may be required under conditions in which oxidizing agents and free radicals are produced. Such conditions can be encountered in biological processes such as wound healing, hypoxia-induced angiogenesis, or inflammation, processes in which VEGF 165 was shown to play an important role (24 -28). Under such conditions, cell surface heparin-like molecules could restore the activity of damaged VEGF 165 molecules. This restorative function of heparin-like molecules could be of critical importance under conditions in which the initial concentration of VEGF 165 is low to begin with. Heparin-like molecules are also able to restore the activity of damaged bFGF and aFGF FIG. 5. Heparin restores the receptor binding ability of oxidized VEGF 165 , but not the receptor binding ability of oxidized VEGF 121 . A, 125 I-VEGF 121 (25 ng/ml) was bound to ELISA dishes coated with flk-1/SEAP fusion protein in a final volume of 50 l as described under "Experimental Procedures." The binding was performed in the presence (छ) or absence (Ⅺ, ࡗ) of heparin (1 g/ml) and in the presence of increasing concentrations of either untreated VEGF 165 (Ⅺ) or VEGF 165 which was treated with chloramine-T as described under "Experimental Procedures" (ࡗ, छ). The binding was performed for 2 h at room temperature, after which the dishes were washed 3 times with buffer containing 0.1% Tween 20 as described. Bound 125 I-VEGF 121 was solubilized using 0.5 N NaOH, and aliquots were counted in a ␥-counter. 100% of 125 I-VEGF 121 binding corresponds to 5500 cpm. B, 125 I-VEGF 121 (25 ng/ml) was bound to ELISA dishes coated with flk-1/SEAP fusion protein in a final volume of 50 l as described. The binding was performed in the presence (छ) or absence (Ⅺ, ࡗ) of heparin (1 g/ml) and in the presence of increasing concentrations of either untreated VEGF 121 (□) or VEGF 121 which was treated with chloramine-T as described (ࡗ, छ). The binding was performed as described under A. 100% of 125 I-VEGF 121 binding corresponds to 10,000 cpm. C, 125 I-VEGF 121 (25 ng/ml) was bound to ELISA dishes coated with flk-1/SEAP fusion protein as described. The binding was performed in the presence (hatched columns) or in the absence (empty columns) of 1 g/ml heparin. The VEGF 121 and VEGF 165 concentration used for competition was 1 g/ml. VEGF 165 and VEGF 121 were treated or not with H 2 O 2 (1%) as described under "Experimental Procedures." D, 125 I-VEGF 121 (20 ng/ml) was bound and cross-linked to confluent heparinase 1-digested HUE cells grown in 10-cm dishes in the absence (lanes 1 and 3) or in the presence (lanes 2 and 4) of 1 g/ml heparin. Unlabeled VEGF 165 was added to a final concentration of 2 g/ml to some of the binding reactions (lanes 3 and 4). The heparinase digestion and the visualization of cross-linked complexes were done as described in Fig. 2. and to protect them from inactivation by heat and oxidation (29,30). The protective and restorative effects of heparin could perhaps account for some of the opposing conclusions that were obtained in experiments designed to assess the importance of heparin-like molecules in the interaction of bFGF with FGF receptors (31)(32)(33). Such a restorative effect would be harder to detect in the case of bFGF since an active bFGF homologue lacking a heparin binding ability (like VEGF 121 ) is unavailable.
In conclusion, our experiments indicate that both VEGF 121 and VEGF 165 bind to the 180-kDa VEGF receptor of HUE cells forming 220 -230-kDa complexes after covalent cross-linking. However, only VEGF 165 is capable of binding to the two smaller VEGF receptors of the endothelial cells. To the best of our knowledge, this is the first time that splice variants of a growth factor are found to differ in receptor recognition patterns. We have also performed immunoprecipitation experiments which indicate that the 220-kDa 125 I-VEGF-receptor complex contains the KDR/flk-1 VEGF receptor. Competition experiments using a soluble fusion protein containing the extracellular domain of flk-1 have also revealed that heparin is not essential for the binding of VEGF 121 or VEGF 165 to flk-1 receptors. These experiments also suggest that the ability to bind heparin-like molecules may help to preserve the biological function of VEGF 165 under conditions in which oxidants and free radicals are produced.