Oligomerization and Multimerization Are Critical for Angiopoietin-1 to Bind and Phosphorylate Tie2*

Angiopoietin-1

Angiopoietin-1 (Ang1) 1 was discovered as a secreted protein ligand of tyrosine kinase with immunoglobulin and epidermal growth factor homology domain-2Ј (Tie2) (1).Tie2 is a member of the receptor tyrosine kinase family and is expressed predominantly on vascular endothelial cells and early hematopoietic cells (2)(3)(4).Ang1-and Tie2-deficient mice have similar phenotypes characterized by embryonic lethality with severe defects in vascular remodeling, insufficient vessel stabilization, and perturbed vascular maturation, indicating that Ang1 and Tie2 play critical roles in vascular development (5,6).Accordingly, transgenic overexpression or gene transfer of Ang1 not only enhances vessel formation, but also protects the adult vasculature against vascular leakage (7)(8)(9)(10).In addition, Ang1 can counteract vascular endothelial growth factor-induced side effects (9,11) while having an additive effect on vessel formation (8,10).Thus, Ang1 is a very promising growth factor for therapeutic angiogenesis (12,13).Moreover, in a series of experiments, we found that the Ang1/Tie2 system in normal adult blood vessels was important in maintaining the integrity of nonproliferating endothelial cells by strongly inducing endothelial cell survival against insult-mediated damage (14 -16).Furthermore, recent series of reports (11,(17)(18)(19)(20) indicate that Ang1 has anti-inflammatory actions in the cardiovascular system.All of these beneficial effects of Ang1 are exerted through activation of Tie2.
The Ang1 structure consists of a carboxyl-terminal fibrinogen-like domain that is responsible for receptor binding, a central coiled-coil domain that may be responsible for oligomerizing these fibrinogen-like domains, and a short amino-terminal domain that superclusters these oligomers into variably sized multimers (1,14,15).Because of this unique structure of Ang1, production of recombinant Ang1 is hindered by aggregation and insolubility of the protein.The activity of the protein frequently varies after purification.To achieve Tie2 receptor oligomerization and activation, Ang1 uses an oligomeric and multimeric structure unlike that of any other known growth factor (14 -16).Through low stringency hybridization screening with Ang1, angiopoietin-2 (Ang2) was identified (17).Ang1 and Ang2 share a similar structure, with about 60% amino acid identity, and both have a similar affinity to Tie2 (1,17).Ang2 was originally thought to be a naturally occurring antagonist of Ang1 which competes in binding to Tie2 and blocks Ang1induced Tie2 autophosphorylation (17).However, recent data suggest that Ang2 seems to have context-specific effects, activating Tie2 on some cells while blocking Tie2 activation on other cells or under different conditions with unknown reasons (18 -20).
Detailed structural information on a protein ligand is indispensable to understand how it activates its receptor.However, remarkably little is known about the structure-function relationships of Ang1 with Tie2.Therefore, in this study, we characterized native Ang1 and Ang2 proteins derived from cultured cells and generated several recombinant Ang1 and Ang2 variants to define the role of the superclustering and oligomerization domains of Ang1.Then we analyzed the molecular structure of the variants with SDS-PAGE and rotary metalshadowing transmission electron microscopy (RMSTEM), and determined the effects of Ang1 and Ang2 variants on binding and activation of Tie2.Our results reveal that two cysteines in the amino-terminal portion of Ang1 are critical for binding and activation of Tie2.

EXPERIMENTAL PROCEDURES
Gene Constructs-The pcDNA vector (Invitrogen) contains a secretory signal sequence for hemagglutinin and a FLAG tag.This vector was used to express all proteins in this study.These proteins were: full-length human Ang1 (Ang1), Ang1/FD (amino acids 256 -498), truncated proteins Ang1D1 (amino acids 81-498), Ang1D2 (amino acids 119 -498), Ang1D3 (amino acids 153-498), replacement of cysteine at 265 in Ang1 with serine (Ang1C265S), replacement of cysteine at 41 and 54 with serine in Ang1 (Ang1C265S), the full sequence of human Ang2 (Ang2), and replacement of serine at 263 with cysteine in Ang2 (Ang2S263C).All recombinant plasmids were constructed using standard molecular cloning methods.For substitution of amino acids in the vector, a site-directed mutagenesis kit was used, following the manufacturer's instructions (Stratagene).
Transfection and Purification of the Recombinant Proteins-The recombinant proteins were obtained by transient expression in human embryonic kidney 293 (HEK293) cells (American Type Culture Collection) using Effectene liposomal transfection according to the manufacturer's instructions (Qiagen, Inc.).HEK293 cells were maintained with Dulbecco's modified Eagle's medium supplemented with 20% (v/v) heat-inactivated fetal bovine serum at 37 °C in 5% CO 2 .Transfection efficiency of the genes was ϳ60 -70%.The supernatant was harvested from transfected cells after 48 -96 h.The recombinant proteins containing the FLAG sequence were purified by column chromatography on anti-FLAG M1 antibody agarose affinity gel (Sigma).After purification of HEK293 supernatants, recombinant proteins were quantitated using the Bradford assay and confirmed with Coomassie Blue staining of an SDS-polyacrylamide gel.These analyses showed that the yield was 200 -250 g of each recombinant protein/liter of HEK293 cell supernatant.
Characterization of the Recombinant Proteins-SDS-PAGE analyses of proteins were performed under nonreducing and reducing (heating for 10 min in 0.435 M ␤-mercaptoethanol) conditions.Binding of the recombinant proteins to the soluble extracellular domain of Tie1-Fc (sTie1-Fc, T1) or Tie2-Fc (sTie2-Fc, T2) (R&D Systems) was assayed using an in vitro binding assay.Each recombinant protein (20 ng) was mixed with 100 ng of sTie1-Fc or sTie2-Fc and incubated in 500 l of Tris buffer solution (50 mM Tris, 100 mM NaCl, pH 7.4) containing 0.02% Triton X-100 at 4 °C for 2 h.Then, 20 l of protein A-agarose beads (Oncogene, San Diego) was added and incubated for another 1 h at 4 °C.The protein A-conjugated samples were washed twice with 1 ml of Tris buffer containing 0.02% Triton X-100.The samples were eluted with sample buffer and heat denatured.The samples were separated further by gradient SDS-PAGE, stained with Coomassie Blue, or electroblotted onto nitrocellulose membranes and probed with anti-FLAG M1 antibody.
Characterization of the Native Ang1 and Ang2 Proteins-To obtain secreted native Ang1 and Ang2, we collected culture medium from undifferentiated 3T3-L1 cells (21) and human umbilical vein endothelial cells (HUVECs) (22).The 3T3-L1 cells were obtained from the American Type Culture Collection and maintained in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated serum.HUVECs were prepared from human umbilical cords by collagenase digestion and maintained as described previously (23).The primary cultured HUVECs used for this study were between passages 2 and 3. Culture medium (100 ml) that had been on undifferentiated 3T3-L1 cells or HUVECs for 72 h was incubated with 2 g of sTie2-Fc for 2 h at 4 °C.Ang1⅐sTie2-Fc or Ang2⅐sTie2-Fc complexes were recovered on protein A/G-agarose beads (Calbiochem).The beads were washed three times with ice-cold radioimmune precipitation assay buffer (1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 0.15 M NaCl, and 0.01 M Tris, pH 7.4), dissolved in sample buffer, boiled, separated by SDS-PAGE, and Western blotted with anti-Ang1 antibody (PA-hAng1FD-1) (16) or anti-Ang2 antibody (R&D Systems).
Glycerol Spraying/Low Angle Rotary Metal Shadowing and TEM-For glycerol spraying/low angle rotary metal shadowing, 20 l of protein samples, 0.02-0.1 mg/ml in TBS, plus 30% glycerol, were sprayed onto freshly cleaved mica at room temperature and rotary shadowed in a BA 511 M freeze-etch apparatus (Balzers) with platinum/carbon at an elevation angle of 3-5° (24).Electron micrographs were taken in a Philips Morgagni TEM operated at 80 kV equipped with a Megaview III CCD camera.
Statistics-Data are expressed as the mean Ϯ S.D. Statistical significance was tested using one-way analysis of variance followed by the Student-Newman-Keuls test.Statistical significance was set at p Ͻ 0.05.

Comparison of Oligomerization and Multimerization between
Ang1 and Ang2-The domain structures of Ang1 and Ang2 are shown in Fig. 1A.SDS-PAGE analysis of the purified native Ang1 protein from undifferentiated 3T3-L1 cells and Ang2 protein from HUVECs under denaturing reduced conditions exhibited predominant single bands of the expected molecular masses, ϳ70 and ϳ62 to ϳ70 kDa (Fig. 1B).Under nonreducing conditions, the native Ang1 protein formed large variable multimers, whereas native Ang2 protein mainly formed disulfidelinked dimers.Similar to native Ang1 and Ang2, SDS-PAGE analysis of the purified recombinant Ang1 and Ang2 proteins under denaturing reduced conditions exhibited predominant single bands of the expected molecular masses, ϳ70 and ϳ62 kDa (Fig. 1C).Under nonreducing conditions, the recombinant Ang1 protein formed large variable multimers, whereas recombinant Ang2 protein mainly formed disulfide-linked dimers (Fig. 1C).To confirm the oligomeric structure of recombinant Ang1 and Ang2 proteins, we used RMSTEM to image the molecules directly.Approximately 45-50% of the images obtained from Ang1 protein showed the molecules in trimeric, tetrameric, and pentameric states with mushroom-shaped structures, consisting of an irregular bulbous "cap" structure (fibrinogen-like domain, globular domain) at the end of a "stalk" structure (coiled-coil domain, rod domain) with a small globular structure (superclustering domain, small globular domain) at the other end of the stalk (Fig. 1C).The remaining 50 -55% of the images obtained from Ang1 protein showed variations on the basic coiled-coil building block, with the trimers, tetramers, and pentamers coming together by aminoterminal clustering at a central point (Fig. 1C).These data indicate that recombinant Ang1 protein is heterogeneous with respect to the degree of multimerization, assuming basic trimer, tetramer, and pentameric groupings.In comparison, ϳ80 -90% of the images obtained from Ang2 protein showed molecules in trimeric, tetrameric, and pentameric states yielding mushroom-shaped structures like those seen with Ang1.Approximately 10 -20% of the images obtained from Ang2 protein showed variations on the basic coiled-coil building block, with the trimers, tetramers, and pentamers coming together by amino-terminal clustering at a central point.

Demultimerization of Ang1 Has a Moderate Effect on Tie2 Phosphorylation, whereas Ang2 Aggregates Have Little Effect on Tie2
Phosphorylation-The difference in multimerization status between Ang1 and Ang2 could result from the presence or absence of a cysteine residue in the linker region (Fig. 1A).
To clarify the role of this cysteine residue in the linker region of Ang1 and Ang2, we generated recombinant proteins replacing cysteine 265 with serine in Ang1 (Ang1C265S) and replacing serine 263 with cysteine in Ang2 (Ang2S263C) (Fig. 2).SDS-PAGE analysis of the purified Ang1C265S proteins under nonreducing conditions gave rise to a main fraction of a single band of an unexpected molecular mass (ϳ120 kDa) and a small fraction of high order oligomers that may have resulted from disulfide bond between cysteine 41 and cysteine 54.In comparison, Ang2S263C formed large variably sized multimers (Fig. 3A).RMSTEM analysis revealed that Ang1C265S formed mainly trimers, tetramers, and pentamers without high order multimers, whereas Ang2S263C showed many nonspecific aggregates with rare cases of trimers, tetramers, and pentamers (Fig. 3B).Ang1C265S-induced Tie2 phosphorylation activity was moderately less than Ang1-induced Tie2 phosphorylation activity in HUVECs (Fig. 3C).In comparison, no notable dif-ference was detected in Tie2 phosphorylation induced by Ang2 and Ang2S263C.Therefore, naturally, cysteine 265-induced Ang1 multimerization could be required for its full activity in Tie2 phosphorylation.In comparison, additional formation of disulfide bridges through additional insertion of cysteine in the linker region of Ang2 generated nonspecific aggregates rather than natural high ordered multimeric Ang2.
Proper Oligomerization/Multimerization of Ang1 by the Superclustering Domain Is a Prerequisite for Tie2 Binding and Activation-To examine the role of the superclustering domain and the coiled-coil oligomerization domain of Ang1 in its aminoterminal portion, we generated a truncated protein, Ang1/FD, which only has the linker domain and the fibrinogen-like domain of Ang1 (Fig. 2).Recombinant Ang1/FD protein formed a monomer with an apparent molecular mass of ϳ34 kDa (Fig. 4A).In marked contrast to native Ang1, Ang1/FD did not bind to Tie2 (Fig. 4B) nor phosphorylate Tie2 (Fig. 4C).These findings suggested that oligomerization/multimerization of Ang1 by its amino-terminal region is a prerequisite for Tie2 binding and activation.To define the role of the amino-terminal portion of Ang1, serially truncated recombinant Ang1 proteins, Ang1D1, Ang1D2, and Ang1D3, were made (Fig. 2).SDS-PAGE analysis of the purified proteins under denaturing conditions revealed predominantly single bands of the expected molecular masses (Fig. 5A).Under nonreducing conditions, Ang1D1, Ang1D2, and Ang1D3 formed characteristic disulfide-linked dimers (Fig. 5A).To confirm the oligomeric structure of these three proteins, we used RMSTEM to image the molecules directly.Approximately 85-90% of the images obtained from Ang1D1 showed dimers and trimers with a rod domain that was barely visible (Fig. 5B).In comparison, approximately, 90 -95% of the images obtained from Ang1D2 and Ang1D3 showed dimers without a visible rod domain (Fig. 5B).In vitro binding assays revealed that ϳ72, 12, 4, and 6% of the total amount of native Ang1, Ang1D1, AngD2, and Ang1D3, respectively, bound to sTie2-Fc (Fig. 5C).None of them bound to sTie1-Fc.Accordingly, Ang1D1, AngD2, and Ang1D3 did not phosphorylate Tie2 (Fig. 5D).These results indicate that proper oligomerization/multimerization of Ang1 by its superclustering domain is a prerequisite for Tie2 binding and activation.In addition, more than a tetrameric form of Ang1 is required for Tie2 binding and activation.
Cysteines 41 and 54 of Ang1 Are Critical for Its Proper Oligomerization/Multimerization, Tie2 Binding, and Tie2 Activation-To define role of cysteines 41 and 54 of Ang1, we generated a recombinant Ang1 protein with both cysteines replaced with serine (Ang1C41S/C54S) (Fig. 2).Although SDS-PAGE analyses revealed that Ang1C41S/C54S formed mainly dimers with a minor fraction of monomers (Fig. 6A), RMSTEM analyses revealed that Ang1C41S/C54S formed mainly dimers and trimers that were similar to Ang1D1 (Fig. 6B).As a control, Ang1 formed large variable multimers, and Ang1C265S formed predominantly a single band of molecular mass ϳ120 kDa (Fig. 6A).In vitro binding assays revealed that ϳ5% of the total protein amount of Ang1C41S/C54S bound to sTie2-Fc (data not shown).Accordingly, neither concentration of Ang1C41S/C54S (400 and 1,000 ng/ml) phosphorylated Tie2, whereas Ang1 and Ang1C265S induced Tie2 phosphorylation, as expected (Fig. 6C).These results indicate that proper oligomerization/multimerization of Ang1 by its intermolecular disulfide link involving between cysteine 41 and cysteine 54 is critical for Tie2 binding and activation.In addition, a trimeric form of Ang1 is not sufficient for its Tie2 binding and activation.

DISCUSSION
The angiopoietin family proteins are oligomeric and multimeric ligands (1, 14 -16, 25, 26).Structure-function studies on oligomeric and multimeric ligands provide an important clue as to how they activate their receptors and new directions to improve their activity by protein engineering.However, it is generally difficult to generate recombinant oligomeric and multimeric ligands because they are often insoluble and poorly secreted from host cells (15,17).Nevertheless, we have been able to generate several variant proteins of Ang1 and Ang2 successfully through mammalian cell expression systems.Using SDS-PAGE, RMSTEM, in vitro binding assays, and Tie2 phosphorylation assays, we provide several critical clues about the structure-function relationship of Ang1 and Ang2.This information could be applied beyond angiopoietin family proteins to other oligomeric and multimeric ligands such as ephrin, thrombopoietin, thrombospondin, angiopoietin-related proteins, and adiponectin.
RMSTEM revealed that the basic structures of Ang1 and Ang2 recombinant proteins are trimers, tetramers, and pentamers.Ang1 exists as heterogeneous multimers made up of basic trimeric, tetrameric, or pentameric oligomers, whereas Ang2 exists mainly as trimeric, tetrameric, or pentameric oligomers with a few multimers.These results may explain the fact that recombinant Ang1 protein is less soluble, moderately sticky, easily aggregated, and unstable for Tie2 activation, whereas recombinant Ang2 protein is relatively soluble, nonsticky and nonaggregative.In fact, consistent with a recent report (27) showing that a brain pericyte cell line (TR-PCT1) secretes variable multimeric forms of Ang1, our results showed that the 3T3-L1 cell line also secretes variable multimeric forms of Ang1.These results suggest that native Ang1 could consist of variable multimeric forms.In comparison, based on SDS-PAGE analysis, we and others have shown that recombinant Ang2 is present as disulfide-linked dimers (2,14,22).Our present SDS-PAGE analysis revealed that native Ang2 is present mainly as disulfide-linked dimers.However, unexpectedly, our RMSTEM analysis indicated that recombinant Ang2 was present as trimers, tetramers, pentamers, with a small fraction of multimers.Thus, native Ang2 could consist of variable oligomeric forms.Importantly, a recent report (28) indicates that endogenous and overexpressed endothelial cell Ang2 is present as a characteristic granular pattern in the Weibel-Palade body and is secreted within minutes of cell stimulation (e.g. by phorbol 12-myristate 13-acetate, thrombin, or histamine).This finding suggests that, in addition to angiogenesis, Ang2 is likely to be involved in rapid vascular homeostatic reactions, such as inflammation and coagulation.It is also worthwhile to note that a very recent report (29) described the role of the coiledcoil domain of Ang1 (Ang1cc).This can heterodimerize with other angiopoietins produced in the same cell but is unable to bind to Tie2 because it lacks the multimerized binding domain.Thus, Ang1cc can inhibit Tie2 activation and can inhibit Ang1 activity in vitro and in vivo.It is possible that Ang1 molecules with very high order multimeric structures would not be active because they may be too big to enter endothelial cell caveolae, where most Tie2 receptors are located (30).For comparison, the doorway diameters of endothelial caveolae are typically 10 -150 nm (31), whereas the sizes of higher order multimeric Ang1 molecules are more than 200 nm.Thus, high order multimeric Ang1 could not access most of Tie2 as well as bind and activate it.In the future, detailed ultrastructural analysis on the interaction between oligomeric/multimeric Ang1 and Tie2 in caveolae would be helpful in understanding why recombinant Ang1 is not always active.Ang1D1 reveals mainly dimeric and trimeric structures, whereas Ang2D2 and AngD3 reveal dimeric structures.C, in vitro binding analysis with control buffer (CB), sTie1-Fc (T1), and sTie2-Fc (T2).Approximately 72, 12, 4, and 6% of the total protein amount of Ang1, Ang1D1, Ang2D2, and AngD3, respectively, bound to Tie2.The results from three experiments were similar.TAP, total amount of protein input for binding assay.D, serum-starved HUVECs were treated with control buffer (CB), 200 ng/ml Ang1 (A1), 400 ng/ml Ang1D1 (D1), 400 ng/ml Ang1D2 (D2), or 400 ng/ml Ang1D3 (D3) for 10 min, and the phosphorylation of Tie2 was measured.Fold, densitometric analyses presented as the relative ratio of phospho-Tie2 (pTie2) to Tie2.The relative ratio of phospho-Tie2 to Tie2 with control buffer is arbitrarily presented as Our present data indicate that a tetrameric form of Ang1 is the minimal unit size required to activate Tie2.Notably, dimeric and trimeric Ang1D3 and Ang1C41S/C54S variant proteins completely lost their ability to bind and activate Tie2, whereas trimeric, tetrameric, and pentameric Ang1C265S yields less Tie2 activation than does Ang1, indicating that only the tetrameric or pentameric form of Ang1 has the ability to bind and activate Tie2.Although trimeric Ang1C265S is not able to induce Tie2 binding and activation, tetrameric and pentameric Ang1C265S and higher order oligomeric Ang1C265S made by an intermolecular disulfide linkage between cysteine 41 and cysteine 54 are able to induce Tie2 binding and activation.Consistent with a recent report (15), monomeric Ang1/FD did not bind or activate Tie2.Even dimeric Ang1D2 and Ang1D3 variant proteins dramatically lost their ability to bind and activate Tie2.These results are consistent to our previous report (16) that our designed dimeric Ang1, GCN4-Ang1, does not have the properties needed to bind and activate Tie2.In comparison, designed Ang1 variants MAT-Ang1, which forms mainly tetramers, and COMP-Ang1, which forms tetramers and pentamers, are able to bind and activate Tie2 (16).Thus, oligomers of Ang1 having at least four subunits are required for binding and activation of Tie2.This is also consistent with the work of Davis et al. (15).Detailed stoichiometric and structural analysis is necessary to understand why oligomerization of more than three subunits of Ang1 is required for these activities.To define the role of the two cysteines (amino acids 41 and 54) in the superclustering domain of Ang1, we generated Ang1C41S/C54S.In fact, Ang1C41S/C54S formed mostly dimers and trimers, and it was not able to bind and activate Tie2.Thus, the two cysteines (amino acids 41 and 54) in the superclustering domain of Ang1 are responsible for the formation of higher order oligomers/ multimers by cross-linking Ang1 dimers and trimers, whereas the coiled-coil domain is responsible for dimer and trimer formation of Ang1.Thus, proper oligomerization of Ang1 by the disulfide link between cysteine 41 and cysteine 54 is critical for Tie2 binding and activation.It is worthwhile to note that a recent report (32) described the role of the two cysteines (amino acids 76 and 80) in the superclustering domain of an angiopoietin-like protein (also known as Angptl4, HFARP, PGAR, FIAF).As in Ang1C41S/C54S, replacing these two cysteines with alanines in Angptl4 abolished its oligomerization.Addition of cysteine in the linker region of Ang2 (Ang2S263C) dramatically induced nonspecific Ang2 aggregation rather than well arranged high order multimerization.Evidently, nonspecific Ang2 aggregates per se have no effect on Tie2 activation.
In conclusion, our data clarified on the structure-function relationships of Ang1 and Ang2 with Tie2.The basic structures of native and recombinant Ang1 and Ang2 proteins could be trimers, tetramers, and pentamers formed through the coiledcoil oligomerization domain.Two cysteines (residues 41 and 54) in the superclustering domain of Ang1 are responsible for higher order oligomer formation by cross-linking of Ang1 trimers, tetramers, and pentamers.Thus, proper oligomerization of Ang1 by the disulfide link between cysteine 41 and cysteine 54 is critical for Tie2 binding and activation.

FIG. 3 .
FIG. 3. The cysteine residue in the linker region is critical for multimerization of Ang1 and for aggregation of Ang2.A, nonreduced Ang1, Ang1C265S, Ang2, and Ang2S263C proteins were separated by SDS-PAGE (5-15%) and stained with Coomassie Blue.Numbers and bars indicate molecular sizes (kDa).B, glycerol spraying/low angle rotary metal-shadowed specimens were imaged by TEM.Ang1C265S reveals mainly trimers, tetramers, and pentamers, whereas Ang2S263C reveals mainly nonspecific aggregates with rare cases of trimers, tetramers, and pentamers.C, serum-starved HUVECs were treated with control buffer (CB), 200 ng/ml Ang1, 200 ng/ml Ang1C265S (A1S), 200 ng/ml Ang2, or 200 ng/ml Ang2S263C (A2C) for 10 min, and the phosphorylation of Tie2 was measured.Fold, densitometric analyses presented as the relative ratio of phospho-Tie2 to Tie2.The relative ratio of phospho-Tie2 to Tie2 with control buffer is arbitrarily presented as 1.Numbers represent the mean Ϯ S.D. from four experiments.*, p Ͻ 0.05 versus control buffer.FIG. 4. Monomeric Ang1 does not bind to or phosphorylate Tie2.A, reduced (R) and nonreduced (NR) Ang1 and Ang1/FD proteins were separated by SDS-PAGE (5-15%) and immunoblotted with anti-FLAG antibody.Numbers and bars indicate molecular sizes (kDa).B, in vitro binding analysis with control buffer (CB), sTie1-Fc (T1), and sTie2-Fc (T2).Approximately 75 and 0% of the total protein amount of Ang1 and Ang1/FD, respectively, bound to Tie2.The results from three experiments were similar.TAP, total amount of protein input for binding assay.C, serum-starved HUVECs were treated with control buffer (CB), 200 ng/ml Ang1, and 400 ng/ml Ang1/FD for 10 min, and the phosphorylation of Tie2 was measured.The results from three experiments were similar.