Activation of Tie2 by angiopoietin-1 and angiopoietin-2 results in their release and receptor internalization

The receptor tyrosine kinase Tie2 is highly expressed in endothelial cells and is crucial for angiogenesis and for vascular maintenance in the embryo and in the adult (Dumont et al., 1992; Dumont et al., 1993; Dumont et al., 1994; Sato et al., 1995; Wong et al., 1997; Jones et al., 2001). The ligands for Tie2 are the angiopoietins, which are comprised of three members: angiopoietin-1 (Ang1) (Davis et al., 1996), Ang2 (Maisonpierre et al., 1997), and the mouse/human interspecies orthologs Ang3/Ang4 respectively (Valenzuela et al., 1999). Ang1 has been identified as the primary activating ligand for Tie2 and it is believed that most of the angiogenic functions of Tie2 can be attributed to the actions of Ang1 (Suri et al., 1996). Binding of Ang1 to the extracellular domain of Tie2 in endothelial cells results in receptor auto-phosphorylation and the activation of several intracellular signaling pathways leading to endothelial cell migration (Witzenbichler et al., 1998; Jones et al., 1999), tube formation (Hayes et al., 1999; Teichert-Kuliszewska et al., 2001), sprouting (Koblizek et al., 1998) and survival (Kwak et al., 1999; Jones et al., 1999). The role of Ang2 has been controversial. Ang2 possesses similar receptor affinity to Ang1 but was initially characterized as a Tie2 antagonist (Maisonpierre et al., 1997). In some studies, the binding of Ang2 did not activate Tie2 and competitively inhibited the binding of Ang1, establishing an agonist-antagonist ligand pair regulating vascular development (Maisonpierre et al., 1997). In other studies, the binding of Ang2 to Tie2 did result in receptor activation (Kim et al., 2000b; Teichert-Kuliszewska et al., 2001) with similar biological outcomes to Ang1, namely, endothelial cell migration (Mochizuki et al., 2002), tube formation (Teichert-Kuliszewska et al., 2001; Mochizuki et al., 2002) and survival (Kim et al., 2000b), thus it was thought that Ang2 may act as a context-dependent antagonist.

Tyrosine kinase receptors are regulated by their cognate ligands (Sorkin and Waters, 1993; Gaborik and Hunyady, 2004). For example, binding of the epidermal growth factor (EGF) to the epidermal growth factor receptor (EGFR) induces rapid internalization of the receptor-ligand complex into the cell followed by their delivery to lysosomes for degradation (Carpenter and Cohen, 1976; Beguinot et al., 1984). Rapid internalization and degradation of activated receptors following ligand binding, a phenomenon referred to as receptor downregulation, has been observed for numerous tyrosine kinase receptors, some examples include the platelet-derived growth factor receptor and the insulin receptor (Kasuga et al., 1981; Keating and Williams, 1987). Receptor downregulation is considered to be a primary mechanism for the attenuation of downstream signal transduction and results in an overall loss of receptors from the cell surface thereby diminishing the ability of the cell to respond to further stimulation (Sorkin and Waters, 1993). It has also been suggested that internalization of activated receptors may be necessary to facilitate downstream signaling and that a proper signaling response may require receptor endocytosis (Sorkin and von Zastrow, 2002; Wiley, 2003).

In our study, we demonstrate that the binding of Ang1 and Ang2 to Tie2 impart different regulatory effects on Tie2 which may ultimately have an impact on the biological output of these two related ligands. These observed differences may provide an important link in further understanding the complex nature of the angiogenic and lymphangiogenic ligands - the angiopoietins.

The regulation of many tyrosine kinase receptors is dependent on ligand concentration (Kasuga et al., 1981) and consequently, on receptor activation where higher levels of receptor activation stimulate faster rates of receptor internalization and degradation. To understand the relationship between the concentration of Ang1 and Ang2, and Tie2 activation, we examined Tie2 tyrosine phosphorylation in response to increasing concentrations of each ligand (Fig. 1). Pre-incubating recombinant Ang1 where a polyhistidine tag has been added at the C-terminus together with an anti-polyhistidine monoclonal antibody enhanced Tie2 tyrosine phosphorylation compared with Ang1 alone presumably by cross-linking Ang1 multimers (Fig. 1A,C). Under these conditions, Ang1 activated Tie2 in a concentration-dependent manner with maximal levels of Tie2 activation observed when the cells were stimulated with 800 ng/ml Ang1 (Fig. 1A). The anti-polyhistidine antibody alone did not induce Tie2 activation (Fig. 1C) and was included in Ang1 stimulations to achieve higher levels of Tie2 activation. Ang2 also activated Tie2 in a concentration-dependent manner (Fig. 1B). Similar to Ang1, maximal activation of Tie2 was observed when the cells were stimulated with 800 ng/ml Ang2 (Fig. 1B). In contrast to Ang1, pre-incubating Ang2 with the anti-polyhistidine antibody did not enhance Tie2 phosphorylation (data not shown). To study the regulation of Tie2 by Ang1 and Ang2, 800 ng/ml of each ligand was used in subsequent experiments because these concentrations produced maximal activation of Tie2.

When tyrosine phosphorylation of Tie2 induced by Ang1 and Ang2 was compared, Tie2 phosphorylation induced by Ang1 was considerably higher compared with Ang2 (Fig. 1C). We examined the phosphorylation of tyrosine residues 1102 and 1108 (mouse 1100 and 1106) on Tie2 in response to Ang1 and Ang2. Phosphorylation on tyrosine residues 1102 and 1108 is important for the initiation of downstream signalling pathways (Jones et al., 1999; Jones et al., 2003). Both Ang1 and Ang2 stimulated phosphorylation of Tyr1102 and 1108 on Tie2 although phosphorylation induced by Ang2 was weaker compared with Ang1 (Fig. 1D).

To examine the effect of Ang1 and Ang2 on the turnover of Tie2 in HUVECs, we determined the half-life of Tie2 by pulse-chase in unstimulated cells, and in cells stimulated with either Ang1 or Ang2. HUVECs were first incubated in media containing [³⁵S]methionine and [³⁵S]cysteine to label newly synthesized Tie2. The cells were then washed and incubated in normal growth medium (with or without ligands) for various times. Upon transfer of the cells to normal growth medium the initial radioactively labelled Tie2 population disappeared over time. This approach allowed us to measure the rate of Tie2 degradation and how this may be modulated by angiopoietin stimulation. Under these conditions, Tie2 degraded exponentially over time (Fig. 2A). Semi-logarithmic plots of Tie2 levels versus time were generated and the slopes of the lines were used to calculate the time required for half of the initial receptor mass to decrease by 50%. The half-life of Tie2 in unstimulated cells was ∼9 hours (range 7.5 to 12 hours), in the presence of Ang1 the half-life was determined to be ∼3 hours (range 1.8 to 4.0 hours), and in the presence of Ang2 the half-life was ∼7 hours (range 4.9 to 8.2 hours) (Fig. 2B). A one-way analysis of variance (ANOVA) followed by the Tukey's HSD Test revealed that the half-life values of Tie2 differ significantly between unstimulated cells and in cells treated with either Ang1 and Ang2 (P<0.05). The rate of Tie2 synthesis was unchanged in the presence of either ligand compared with unstimulated cells (Fig. 3).

The increased rate of degradation of Tie2 in the presence of Ang1 and Ang2 could also be observed in cell lysates by western blot analysis (Fig. 4) further suggesting that the rate of Tie2 synthesis was unable to compensate for the ligand-induced degradation of the receptor and that activation of the Tie2 pathway does not up-regulate its own synthesis.

Upon ligand binding, many tyrosine kinase receptors are rapidly internalized into intracellular compartments (Sorkin and Waters, 1993; Gaborik and Hunyady, 2004). The internalization of Tie2 upon binding Ang1 or Ang2 was indirectly assessed by tagging Tie2 at the cell surface with biotin. Biotinylated Tie2 was isolated with streptavidin conjugated to agarose beads and detected by immunoblotting (Fig. 5). To ensure that only Tie2 at the plasma membrane was being isolated, biotin was stripped from the cell surface and proteins pulled down with streptavidin agarose. After stripping the cell surface, most of the Tie2-specific signal in the immunoblot was lost confirming that Tie2 receptors isolated with streptavidin agarose represented the receptors at the plasma membrane (Fig. 5A).

Internalization of Tie2 in the presence of Ang1 and Ang2 was determined by examining the amount of receptor remaining at the cell surface at various times after the addition of either ligand. HUVECs were incubated with either Ang1 or Ang2 for various times at 37°C followed by biotinylation of cell surface proteins. Cell lysates containing equal amounts of protein were incubated with streptavidin agarose and the amount of Tie2 remaining at the cell surface was determined by immunoblotting for Tie2. In the presence of Ang1, Tie2 was rapidly internalized into the cell evidenced by the reduced amount of Tie2 isolated with streptavidin agarose (Fig. 5B). By contrast, in the presence of Ang2, Tie2 internalization was considerably slower compared with Ang1, with almost all of the receptor still remaining at the cell surface after 90 minutes of stimulation (Fig. 5B). When cell stimulations with Ang1 were performed at 0°C, a temperature where receptor endocytosis is suppressed, the disappearance of Tie2 from the cell surface was inhibited indicating that the reduced amount of Tie2 isolated with streptavidin agarose was a consequence of receptor internalization (data not shown).

Most peptide growth factors that bind to their cognate receptors are co-internalized upon activation and internalization of the receptor. In order to determine whether Ang1 or Ang2 are also internalized together with Tie2, HUVECs were stimulated with iodinated Ang1 or Ang2. Iodination of Ang1 and Ang2 did not disrupt the multimerization state of either ligand as shown by gel electrophoresis and western blot analysis. Under reducing conditions, both native Ang1 and ¹²⁵I-labeled Ang1 (¹²⁵I-Ang1) migrated with a molecular mass of ∼70 kDa, and under non-reducing conditions migrated with a minimum molecular mass of ∼200 kDa suggesting that the minimal multimeric state of Ang1 and ¹²⁵I-Ang1 is a trimer as reported previously for Ang1 (Procopio et al., 1999) (Fig. 6A). Ang2 and ¹²⁵I-Ang2 migrated, under reducing conditions, with a molecular mass of ∼70 kDa, and under non-reducing conditions migrated with a minimum molecular mass of ∼150 kDa, suggesting that the minimal multimeric state of Ang2 and ¹²⁵I-Ang2 is a dimer as reported previously for Ang2 (Procopio et al., 1999) (Fig. 6B).

The binding of ¹²⁵I-Ang1 and ¹²⁵I-Ang2 to HUVECs was performed at 0°C. At this temperature, receptor endocytosis is inhibited thereby allowing ligand binding with minimal receptor internalization. Following ligand binding at 0°C, the cells were washed to remove unbound ligands and then transferred to 37°C to initiate receptor internalization. Immediately upon transfer of the cells to 37°C there was a progressive loss of radioactivity from the cells evident within 5 minutes of incubation (Fig. 6C). This was observed for both ¹²⁵I-Ang1 and ¹²⁵I-Ang2 with the loss of radioactivity from the cells faster when ¹²⁵I-Ang2 was bound (Fig. 6C). When ¹²⁵I-Ang2 was bound, more than half of the initial radioactivity was lost from the cells after 30 minutes of incubation at 37°C, whereas when ¹²⁵I-Ang1 was bound, half of the initial radioactivity was lost after 90 minutes of incubation. Cross-linking ¹²⁵I-Ang1 with the anti-polyhistidine antibody did not significantly influence the rate of release of the ligand from the cell surface and was therefore omitted from the binding studies.

When the culture medium was analyzed for the presence of Ang1 and Ang2 there was a progressive accumulation of radioactivity in the media. This release into the medium occurred more rapidly with ¹²⁵I-Ang2 compared with ¹²⁵I-Ang1. Radioactivity was detected in the medium within 5 minutes at 37°C and continued to increase up to 2 hours of incubation for both ligands (Fig. 6D). When the cells were maintained at 0°C the release of ¹²⁵I-Ang1 was significantly reduced. By contrast, ¹²⁵I-Ang2 had a tendency to be released at 0°C (Table 1).

To determine whether the released radioactivity represented intact Ang1 or Ang2, or possibly a proteolytic fragment of the ligands, the media was collected and incubated with anti-polyhistidine antibodies pre-coupled to protein-G-Sepharose. Following SDS-PAGE under reducing conditions and autoradiography, a band with a molecular mass of ∼70 kDa was immunoprecipitated from the medium collected from cells previously incubated with ¹²⁵I-Ang1 or ¹²⁵I-Ang2 (Fig. 6E,F). The ∼70 kDa proteins isolated from the medium collected from ¹²⁵I-Ang1- and ¹²⁵I-Ang2-bound cells migrated at the same molecular mass as full-length ¹²⁵I-Ang1 and ¹²⁵I-Ang2 respectively, suggesting that the ¹²⁵I-Ang1 and ¹²⁵I-Ang2 released from the cells upon Tie2 internalization were not proteolytically modified upon release from the receptor. As predicted from our earlier studies, there was significantly less ¹²⁵I-Ang1 in the medium from cells incubated at 0°C (Fig. 6E), whereas ¹²⁵I-Ang2 was released into the medium even when the cells were maintained at 0°C (Fig. 6F).

To further investigate the observation that ¹²⁵I-Ang1 and ¹²⁵I-Ang2 remain at the cell surface and may not be internalized into endothelial cells, surface bound ligands were differentiated from internalized ligands by cleaving surface proteins with trypsin. Surface-bound ¹²⁵I-Ang1 and ¹²⁵I-Ang2 were not readily removed by washing the cells with high salt and acid buffers which are commonly used to remove surface ligands from their receptors. By treating HUVECs with trypsin at 0°C, ∼90% of surface-bound ¹²⁵I-Ang1 and ¹²⁵I-Ang2 were removed by this procedure.

To determine whether Ang1 and Ang2 remained at the cell surface, HUVECs were incubated with ¹²⁵I-Ang1 or ¹²⁵I-Ang2 for 30 minutes at 37°C. The cells were washed to remove unbound ligands and the proteins at the cell surface were then cleaved with trypsin. Following digestion of surface proteins by trypsin, 90% of ¹²⁵I-Ang1 and 86% of ¹²⁵I-Ang2 were removed from the cells as evidenced by the loss of radioactivity associated with the cells and the concomitant increase of radioactivity in the supernatant fraction (Table 2). These findings suggest that ¹²⁵I-Ang1 and ¹²⁵I-Ang2 were not internalized into endothelial cells but remained at the cell surface where they were accessible to trypsin. Similar results were obtained when the cells were incubated with ¹²⁵I-Ang1 or ¹²⁵I-Ang2 for 15 minutes, 1 hour, or 2 hours (data not shown).

To ensure that the surface release of the ligands was not a consequence of the iodination, we incubated HUVECs with non-iodinated Ang1 or Ang2 and repeated our release experiment. Western blot analysis using the anti-polyhistidine antibody revealed that Ang1 and Ang2 were released into the media with similar kinetics as that observed with ¹²⁵I-Ang1 and ¹²⁵I-Ang2 (Fig. 6G).

To determine whether Ang1 and Ang2 release was due to shedding of the extracellular domain of Tie2, we measured the amount of Tie2 present in the culture medium of cells after stimulation with either Ang1 or Ang2 (Fig. 7A). Our results reveal that although soluble Tie2 is present in the culture medium of HUVECs as described previously (Reusch et al., 2001), the amount of soluble Tie2 released to the media was not increased by Ang1 or Ang2 stimulation indicating that the release of Ang1 and Ang2 does not involve cleavage of the extracellular domain of Tie2. This observation is consistent with our preliminary experiments suggesting that released Ang1 and Ang2 are biologically active and are capable of rebinding to fresh cells (data not shown).

Blocking Tie2 internalization using chemical inhibitors such as phenylarsine oxide (PAO) or hypertonic sucrose (0.5 M), agents shown to block internalization of a variety of cell surface receptors (Hertel et al., 1985; Faussner et al., 2004), had no effect on the release of Ang1 although treatment of cells with 0.5 M sucrose only partially blocked Tie2 internalization in response to Ang1 (data not shown). Receptor internalization is probably not sufficient for ligand release because Ang2 did not significantly promote Tie2 internalization but was released faster compared with Ang1.

To stimulate intracellular signalling within endothelial cells, we exposed endothelial cells to pervanadate, a powerful protein tyrosine phosphatase inhibitor known to considerably enhance phosphorylation on tyrosine, serine and threonine residues of proteins. Exposure of HUVECs to pervanadate led to marked increases in tyrosine phosphorylation of cellular proteins (data not shown). When endothelial cells were exposed to pervanadate Ang1 was released faster compared with endothelial cells not treated with pervanadate (Fig. 7B). Examination of Tie2 internalization in the presence of Ang1 and pervanadate revealed that although Ang1 induced Tie2 internalization, the addition of pervanadate did not further stimulate Tie2 internalization (Fig. 7C) indicating that receptor internalization alone is not sufficient for Ang1 release. Pervanadate also enhanced Ang2 release but more consistent results were obtained with Ang1 (data not shown). In addition, stimulation of endothelial cells with pervanadate led to increased levels of Tie2 tyrosine phosphorylation (data not shown).

In this study, we demonstrate that in response to Ang1, Tie2 is rapidly internalized and targeted for degradation. By contrast, Ang2 only weakly activated Tie2 and did not significantly stimulate receptor internalization compared with Ang1, while mildly inducing Tie2 degradation. More importantly, we show that the angiopoietins are released from the endothelial cell surface after binding and accumulate in the surrounding medium.

The observation that Ang1 induced rapid internalization and degradation of Tie2 is consistent with studies of other tyrosine kinase receptors, such as the EGFR (Stoscheck and Carpenter, 1984), where degradation of the receptor may serve to turn off signal transduction, a process crucial for maintaining cellular homeostasis. For example, a mutant EGFR that cannot undergo ligand-induced downregulation transforms cells in culture (Wells et al., 1990). Evidence that Tie2 signalling may be regulated in the vasculature comes from studies that have identified constitutively active mutants of Tie2 causing vascular abnormalities known as venous malformations in humans (Vikkula et al., 1996; Calvert et al., 1999). The degradation of Tie2 upon binding Ang1 may serve to regulate the magnitude and duration of Ang1 signal transduction. Consistent with this possibility, Ang2 only weakly activated Tie2 and resulted in a correspondingly reduced rate of Tie2 degradation.

Similar to our study, Hashimoto et al. (Hashimoto et al., 2004) have shown that Tie2 is downregulated in cell lysates in the presence of Ang1 but not Ang2. In our study, the range of half-life values obtained for Tie2 in the presence of Ang2 overlapped with the range of half-life values in unstimulated cells indicating that in some experiments the downregulation of Tie2 by Ang2 may not be as pronounced.

The differences in the rate of Tie2 internalization stimulated by Ang1 and Ang2 might be related to the level of Tie2 activation induced by these ligands, where higher levels of Tie2 activation promoted a more rapid rate of Tie2 internalization. It is possible that owing to the higher multimerization state of Ang1, a larger population of Tie2 receptors were activated and subsequently targeted for degradation. In line with this possibility, rapid downregulation of the EGFR is dependent on the tyrosine kinase activity of the receptor (Wiley et al., 1991). Recently, receptor dimerization has also been shown to be sufficient to induce rapid internalization of the EGFR (Wang et al., 2005). This could also be applied to Tie2 where Ang1, owing to its higher multimerization state would induce corresponding higher multimerization states of Tie2 compared with Ang2. To date, the mechanism(s) regulating Tie2 internalization are not known.

Our study has characterized Ang2 as a partial agonist for Tie2 rather than a complete antagonist because both Ang1 and Ang2 exerted similar effects on Tie2 but with varying potency and kinetics. In addition to inducing phosphorylation of Tyr1102 and Tyr1108 on Tie2, Ang1 and Ang2 activate similar signalling pathways such as the activation of Akt (Kim et al., 2000a; Papapetropoulos et al., 2000; Kim et al., 2000b), although a higher concentration of Ang2 is required (Kim et al., 2000b). Nevertheless, our characterization of Ang2 as a partial agonist does not rule out the possibility that Ang2 could still antagonize the Ang1 signal.

The role of Ang2 in Tie2 signalling has been controversial in the literature with some studies describing Ang2 as a Tie2 antagonist (Maisonpierre et al., 1997) whereas other studies reported the agonist properties of Ang2 (Kim et al., 2000b; Teichert-Kuliszewska et al., 2001). The reason for the controversy among the various studies is not clear. In our study, the activation of Tie2 by Ang2 is weak and concentration dependent. Perhaps the weak induction of Tie2 by Ang2 may be interpreted as background phosphorylation. Another angiopoietin family member, Ang4, elicits similar levels of Tie2 activation as Ang2 (our unpublished observations) but is described in the literature as a Tie2 agonist (Valenzuela et al., 1999).

The observation that Ang1 and Ang2 were not internalized into endothelial cells but instead were released from the cell surface into the surrounding medium was surprising because many ligands have been shown to internalize together with their receptors. Similar to our results, Wang et al. (Wang et al., 2002) have shown that VEGF is released into the surrounding medium after binding to human colonic vascular endothelial cells and that released VEGF can be reused. Ang1 and Ang2, which are released into the medium after binding to endothelial cells, are capable of rebinding to fresh cells (data not shown), suggesting that these ligands may also be recycled or reused by endothelial cells.

Our data have indicated that shedding of the extracellular domain of Tie2, and receptor internalization are not responsible for the release of the angiopoietins but suggests that the mechanism(s) may be dependent in part or can be modulated by intracellular signalling. The observation that pervanadate appears to enhance angiopoietin release suggests that activation of Tie2 together with the activation or inhibition of particular signalling pathways or molecules is responsible for the release of Ang1 and Ang2 from the endothelial cell surface. Although the data obtained with pervanadate may suggest the involvement of a signalling pathway(s), the mechanism(s) involved might be more complex. The mechanism(s) regulating Ang1 and Ang2 release may be different. For example, Ang2 only weakly induces tyrosine phosphorylation of Tie2 but was released faster from endothelial cells compared with Ang1. One possibility is that angiopoietin release may be related more to changes in receptor affinity after binding to these ligands. Further work will be required to completely elucidate the mechanism(s) regulating ligand release.

The affinities of Ang1 and Ang2 for Tie2 were reported to be the same when binding of either ligand to the extracellular domain of Tie2 was examined in vitro (Maisonpierre et al., 1997). The interaction of Ang1 and Ang2 with the endothelial cell surface may be more complex, because several reports have shown that Ang1 can bind to other surface receptors such as Tie1 and integrins (Carlson et al., 2001; Saharinen et al., 2005). Ang2 can also bind to integrins and activate intracellular signalling pathways (Hu et al., 2006). The fact that Ang2 was released faster after binding compared with Ang1 suggests that the interaction between these ligands and the endothelial cell surface may be different. One possibility is that Ang1 may bind with higher affinity to other receptors such as Tie1 and/or integrins and therefore remain on the cell surface longer compared with Ang2. The release of the angiopoietins after binding and activation of Tie2 might make them available to bind to other receptors on endothelial cells and on non-endothelial cells.

The angiopoietins together with VEGF are an emerging class of ligands, which, instead of being co-internalized together with their cognate receptors as described for numerous other receptor-ligand systems, are released. We believe that the ability of ligands to be released after receptor activation will be a consideration when studying endothelial cell biology.

Human umbilical vein endothelial cells (HUVECs) (GlycoTech, Rockville, MD) were grown on gelatinized (2% gelatin, Sigma) tissue culture plates in F12K medium (ATCC) supplemented with 10% fetal bovine serum, penicillin-streptomycin, 0.1 mg/ml heparin sodium, 10 ng/ml EGF, 10 ng/ml VEGF, 5 ng/ml bFGF, and 2 mM L-glutamine, and were used up to passage 10. EA.hy926 endothelial cells were cultured as described previously (Jones et al., 1999). Human recombinant Ang1 and Ang2 containing a polyhistidine tag were obtained from R&D Systems. All cell stimulations were performed in supplemented F12K medium containing varying concentrations of Ang1 or Ang2. For some stimulations with Ang1, recombinant Ang1 was pre-incubated together with an anti-polyhistidine monoclonal antibody (10 μg/ml, R&D Systems) to facilitate cross-linking of Ang1 multimers.

HUVECs were washed twice with ice-cold phosphate buffered saline (PBS) and lysed with RIPA lysis buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% Igepal, 0.5% sodium deoxycholate, 0.1% SDS, 2 mM sodium orthovanadate and protease inhibitors). The cells were incubated on ice for 20 minutes followed by centrifugation at 14,000 g for 5 minutes at 4°C. Tie2 was immunoprecipitated from the supernatant using 2-4 μg of anti-Tie2 antibody C20 (Santa Cruz Biotechnology) that had been pre-coupled to 25 μl protein-A-Sepharose (Amersham Biosciences).

Proteins were resolved on a 7.5% polyacrylamide gel and transferred to PVDF (Perkin Elmer) membranes. The membranes were blocked for 1 hour at room temperature in 5% bovine serum albumin (BSA) prior to detecting phosphotyrosine, or in 5% non-fat dry milk for detecting total Tie2. The membranes were incubated with 1 μg/ml anti-phosphotyrosine antibody 4G10 (Upstate), anti-phospho-Tie2 Tyr1102/1108 antibody (1:5000 dilution, Calbiochem) or 0.5 μg/ml anti-Tie2 monoclonal antibody (BD Biosciences Pharmingen). Proteins were visualized using secondary antibodies conjugated to HRP (Bio-Rad) followed by enhanced chemiluminescence.

The half-life of Tie2 in unstimulated cells and in cells treated with either Ang1 or Ang2 was determined by ³⁵S labelling of proteins as described (Burke and Wiley, 1999). HUVECs were incubated for 16-18 hours at 37°C in DMEM lacking methionine and cysteine (Gibco) supplemented with 10% dialyzed fetal bovine serum (Gibco), penicillin-streptomycin, 2 mM L-glutamine, and 100 μCi/ml [³⁵S]methionine and [³⁵S]cysteine (Tran³⁵S-Label, MP Biomedicals). Cells were washed six times with HUVEC growth medium and incubated in the same medium either alone, or containing 800 ng/ml Ang1 (cross-linked with the anti-polyhistidine monoclonal antibody) or 800 ng/ml Ang2 for various times at 37°C. The cells were lysed and Tie2 immunoprecipitated and resolved on a 7.5% polyacrylamide gel. The gel was dried and exposed to a Molecular Dynamics phosphorimager screen. The intensities of the Tie2 bands were quantified using the Bio-Rad Molecular Imager FX. The intensity of the band at time=0 was designated 100%. The intensities of all other bands were expressed as a percentage of the initial (t=0) value. The natural logarithm of the percentage values was taken and plotted against time. The slope of the line was used to determine the time required for the initial value (100%) to decrease to 50%.

To examine Tie2 synthesis in the presence of Ang1 or Ang2, cells were incubated in DMEM lacking methionine and cysteine as described above, containing 100 μCi/ml [³⁵S]methionine and [³⁵S]cysteine and either 800 ng/ml Ang1 (cross-linked) or 800 ng/ml Ang2 for various times at 37°C. The cells were lysed and Tie2 immunoprecipitated and resolved on a 7.5% polyacrylamide gel and the bands were quantified as above.

HUVECs were incubated for various times with either Ang1 (800 ng/ml, cross-linked) or Ang2 (800 ng/ml) at 37°C. The cells were placed on ice, washed twice with ice-cold PBS and incubated for 15 minutes at 4°C with 0.5 mg/ml biotin (Sulfo-NHS-SS-Biotin, Pierce) in PBS containing Ca²⁺ and Mg²⁺. The biotin solution was removed and the cells were washed once with 25 mM Tris-HCl pH 8.0 to quench non-reacted biotin followed by three washes with cold PBS. The cells were lysed with RIPA lysis buffer and cell lysates containing equal amounts of protein were incubated with 25 μl streptavidin conjugated to agarose (Pierce). The beads were washed three times with 1 ml RIPA lysis buffer, resuspended in 85 μl sample buffer, and boiled for 5 minutes. Supernatants (60 μl) were immunoblotted for Tie2 as described. For control purposes, surface biotin was removed by incubating cells three times in cold stripping solution (50 mM reduced glutathione, 75 mM NaCl, 75 mM NaOH, 1 mM EDTA and 1% BSA) for 10 minutes each at 4°C followed by two washes with cold PBS as described previously (Burke et al., 2001).

Carrier-free human recombinant Ang1 and Ang2 (25 μg each, R&D Systems) were iodinated using IODO-GEN pre-coated iodination tubes (Pierce) and 0.2-0.5 mCi Na[¹²⁵I] (MP Biomedicals). All steps were performed according to the manufacturer's instructions. Free iodine was separated using 10 ml Dextran desalting columns (Pierce).

HUVECs grown to confluency were washed once with PBS and incubated in ice-cold binding medium (F12K medium containing 0.1% BSA) containing either ¹²⁵I-Ang1 or ¹²⁵I-Ang2 for 90 minutes on ice at 4°C to allow ligand binding. Cell monolayers were washed six times (4 ml per wash) with cold binding medium, fresh binding medium was added, and the cells transferred to 37°C. At various times, the cells were removed from the incubator and the medium collected and counted. The cells were lysed with solubilization buffer (0.1 M NaOH, 0.1% SDS) as described previously (Waterman et al., 1998). Radioactivity contained within the media and the cells was counted in a Compugamma CS gamma counter. To differentiate between surface bound or internalized ligands, surface ligands were removed by trypsin. Cell monolayers were placed on ice, washed three times with cold PBS, and incubated with trypsin (1 mg/ml) for 55 minutes on ice at 4°C followed by 5 minute incubation at 37°C. The cells were collected and centrifuged at 1000 g for 5 minutes. The supernatant, which contained digested extracellular proteins including cleaved surface ligands, was removed and counted. The cell pellet consisting of internalized ligand, was resuspended in solubilization buffer and counted.

Ang1 and Ang2 were immunoprecipitated from cell culture media using 2 μg anti-polyhistidine monoclonal antibody (R&D Systems) that was pre-coupled to 25 μl protein-G-Sepharose (Amersham Biosciences). The immunoprecipitates were immunoblotted using 1 μg/ml anti-polyhistidine monoclonal antibody (R&D Systems). Proteins were visualized using goat anti-mouse secondary antibodies conjugated to HRP (Bio-Rad) followed by enhanced chemiluminescence.

HUVECs were grown to 25% confluency in 24-well dishes for 24 hours before treatment with Ang1 (cross-linked) or Ang2. Each treatment was done in triplicate. Following cell stimulations, the culture media was collected and cleared by brief centrifugation. Soluble Tie2 present in the culture media was quantified by the Quantikine Human Tie2 Immunoassay (R&D Systems) according to the manufacturer's instructions.

The authors would like thank Dr Mariana Capurro, Dr Vladimir Lhotak and John Ebos, for helpful suggestions. This work was supported by operating grants from the Canadian Institutes of Health Research and the Heart and Stroke Foundation of Ontario. D.J.D. holds a Canada Research Chair in Angiogenic and Lymphangiogenic Signalling. E.B. is supported by the Ontario Graduate Scholarship in Science and Technology. V.P.K.H.N. is a recipient of the Heart and Stroke Master's Studentship.