Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Adhesion receptors of the integrin family are responsible for a wide range of cell-extracellular matrix and cell-cell interactions (Hynes, 1992; Clark and Brugge, 1995). Each integrin consists of noncovalently associated alpha and beta subunits which pair to create heterodimers () with distinct adhesive capabilities. As receptors of extracellular matrix proteins, integrins provide anchorage and convey signals that regulate cell growth, differentiation and migration (Juliano and Haskill, 1993; Sastry and Horowitz, 1993). The integrin _v3 is expressed on endothelial cells, osteoclasts, melanoma and other cell types (Cheresh and Spiro, 1987; Cheresh,1987; Miyuchi et al., 1991; Horton, 1990), where it plays a role in physiological processes that include angiogenesis and tissue repair as well as pathological conditions such as osteoporosis, tumor cell metastasis, adenoviral infections and tumor-induced angiogenesis (Schwartz, 1993; Brunhilde,1992; Davis et al., 1993; Ross et al., 1993; Brooks et al., 1994a,b; Wickham et al., 1993). Recent results which demonstrate a critical role for integrin _v3 in angiogenesis have sparked an interest in this receptor (Brooks et al., 1994a,b). Brooks et al. (1994a,b) have shown that antibody and peptide antagonists of integrin _v3 inhibit angiogenesis on the chick chorioallantoic membrane when introduced intravenously into the chick embryo. Evidence has been presented indicating that antagonists of integrin _v3 inhibit this process by selectively promoting apoptosis of vascular endothelial cells (Brooks et al., 1994a). These findings indicate a key role for integrin _v3 in a signaling event critical for the survival and ultimately differentiation of vascular cells undergoing angiogenesis in vivo. These results also provide evidence that antagonists of integrin _v3 may provide a novel therapeutic approach for the treatment of neoplasia or other diseases characterized by angiogenesis.

Although originally isolated as a receptor for vitronectin, _v3 actually recognizes a broad range of extracellular matrix protein ligands such as vitronectin, fibronectin, fibrinogen, von Willebrand factor, thrombospondin and osteopontin, all of which contain the classical integrin recognition motif, Arg-Gly-Asp (RGD) (Leavesly et al., 1992; Charo et al., 1990). The relaxed specificity of _v3 contrasts sharply with the selectivity of ₅₁ integrin which binds to only fibronectin and fibrinogen (D'Souza et al., 1991; Suehiro et al., 1997). Thus, even though the RGD motif has been firmly established as a key determinant in the recognition of extracellular matrix protein ligands by _v3, _v1 and other integrins, the molecular basis for differences in receptor-ligand specificity remains poorly understood.

Investigation of the role of RGD-interactive adhesion molecules has been facilitated by the identification of small RGD-containing proteins derived from snake venoms termed disintegrins (Gould et al., 1990). Disintegrins are a family of naturally occurring, cysteine-rich, small (5-9 kDa) polypeptides that potently inhibit platelet aggregation and cell adhesion (Gould et al., 1990; Niewiaroski et al., 1994). The biological activity of disintegrins depends on the structure of an RGD-containing loop maintained in an appropriate conformation by disulfide bridges. Because they are relatively small (each is 50-80 amino acids), they provide a unique opportunity to gain insight into the three-dimensional structure of RGD-active proteins and the factors that are important in controlling specificity. Echistatin from the venom of Echis carinatus is the smallest (49 amino acids) member of the family and has been the focus of intense research (Gan et al., 1988).

Echistatin is believed to bind to the _v3 integrin expressed on osteoclasts (Sato et al., 1990, 1994; Fisher, et al., 1993). Sato et al. (1990) have shown that echistatin inhibits both excavation of bone by rat osteoclasts and the release of ³H-proline from prelabeled bone particles by chicken osteoclasts (Sato et al., 1994). Because _v3 is the predominant receptor expressed on osteoclasts, these studies suggested that echistatin can bind to integrin _v3. These activities depend on the RGD domain, because substitutions of the arginine of the RGD sequence of the echistatin with alanine resulted in loss of activity (Fisher et al., 1993). However, detailed biochemical studies on the binding of echistatin to _v3 receptor are lacking. This study provides a quantitative biochemical characterization of the binding of echistatin to integrin _v3. In this study, we show that integrin _v3 receptor binds to echistatin with a high affinity both in its purified form and the form expressed on the cell surface. Echistatin can also inhibit the adhesion of human embryonic kidney cells expressing _v3 receptor to vitronectin, which suggests that it can function as an antagonist of the receptor. We also demonstrate that echistatin binds to _v3 in a nondissociable manner similar to vitronectin.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. Human embryonic kidney (HEK 293) cells were obtained from American Type Culture Collection (CRL 1573). DMEM, L-glutamine, nonessential amino acids, gentamycin and synthetic RGD-containing peptides were purchased from Gibco-BRL (Gaithersburg, MD). Fetal bovine serum was from Hazleton Biologicals (Lenox, KS). Octyl--D-glucopyranoside and Nonidet P-40 were purchased from Sigma Chemical Company (St. Louis, MO). Microlite-2 plates were obtained from Dynatech Corporation (Chantilly, VA). Multiscreen-FB opaque plates (1.0 µm Glass Fiber Type B filter) were from Millipore (Billerica, MA). Falcon Microtest III microtiter plates are from Falcon (Franklin Lakes, NJ). _v3 specific (LM609) monoclonal antibodies, anti-_v5 specific antibodies (mAb 1961) and LM609-coupled to Affi-Gel matrix were purchased from Chemicon International Inc. (Temecula, CA). Anti _v specific monoclonals (12084-018) were from Gibco-BRL, and anti-₃ (550036) and anti-₁ (55034) specific monoclonal antibodies were purchased from Becton Dickinson (Franklin Lakes, NJ). ¹²⁵I-Echistatin labeled by the lactoperoxidase method to a specific activity of 2000 Ci/mmol was from Amersham International (Chicago, IL). Echistatin was purchased from Bachem (Torrence, CA).

Protein purification. _v3 was purified as described by Orlando and Cheresh (1991). Human placenta was cut into 2-cm² pieces and washed with 0.05% Digitonin, 2 mM PMSF, 2 mM CaCl₂ in water for 60 min on ice. The placenta was then extracted by incubation with 100 mM octylglucoside, 2 mM CaCl₂, 1 mM PMSF in PBS for 60 min on ice. The resulting extract was filtered through sterile gauze and centrifuged at 50,000 × g for 30 min. The supernatant was then recirculated over LM609-Affi-Gel column overnight at 4°C. The column was washed with 50 column volumes of 0.1% Nonidet P-40, 2 mM CaCl₂ in PBS, followed by 50 column volumes of 0.01 M NaHOAc, pH 4.5, 0.1% Nonidet P-40, 2 mM CaCl₂, in PBS. _v3 was then eluted with 0.01 M NaHOAc, pH 3.0, 0.1% Nonidet P-40 and 2 mM CaCl₂. The column fractions were rapidly neutralized by collecting the fractions directly into 3.0 M Tris, pH 8.8. Fractions containing the receptor, as judged by SDS-PAGE, were pooled and concentrated against high molecular weight Polyethylene glycol. The receptor preparation was then dialyzed against 0.1% Nonidet P-40, 2 mM CaCl₂ in PBS and stored at 80°C. The identity and purity of the protein was confirmed by Western blot analysis with monoclonal antibodies specific for _v and ₃ subunits.

Solid-phase receptor binding assay. The receptor binding assay was performed as described previously (Orlando and Cheresh, 1991). _v3 was diluted at 500 ng/ml in coating buffer (20 mM Tris, pH 7.4, 150 mM NaCl, 2 mM CaCl₂, 1 mM MgCl₂, 1 mM MnCl₂) and an aliquot of 100 µl/well was added to a 96-well microtiter plate (Microlite-2 from Dynatech) and incubated overnight at 4°C. The plate was washed once with blocking/binding buffer (50 mM Tris, pH 7.4, 100 mM NaCl, 2 mM CaCl₂, 1 mM MgCl₂, 1 mM MnCl₂, 1% bovine serum albumin), and incubated an additional 2 h at room temperature. The plate was rinsed twice with the same buffer and incubated with radiolabeled ligand at the indicated concentrations for 3 h at room temperature. For coincubations, unlabeled competitor was included at the concentrations described. For preincubations, after the 3-h incubation with radiolabeled ligand, the plate was washed three times with blocking/binding buffer and further incubated for indicated times in the presence of either competitor or buffer alone. After an additional three washes, the plates were counted by liquid scintillation method with Top count (Packard, Meriden, CT). When ¹²⁵I-ligand incubations were performed without receptor, no interaction was detected because of nonspecific adsorption with the microtiter well. Nonspecific binding of ligand to the receptor was determined with molar excess (200-fold) of the unlabeled ligand. Each data point is a result of the average of triplicate wells.

Cell lines. HEK-293 cells were maintained in DMEM supplemented with 10% fetal bovine serum (Hazleton, Lenox, KS), 1% glutamine, 1% penicillin and 1% streptomycin (Sigma). Human _v and ₃ cDNAs (3.2 kb and 2.4 kb in length) were isolated from published sequences (Suzuki et al., 1987; Rosa et al., 1988) by the reverse transcriptase-polymerase chain reaction method with human placental poly(A)⁺ RNA as a template. The identity of the cDNAs were confirmed by partial sequencing using Sanger's dideoxy method (Fitzgerald et al., 1987a,b). The _v and ₃ cDNAs were subcloned into the mammalian expression vector pcDNA3 (Clontech, Palo Alto, CA) which contains a CMV promoter and a G418 selectable marker. HEK-293 cells were transfected with equimolar concentrations of _v/pcDNA3 and ₃/pcDNA3 vectors by the calcium phosphate method (Chen and Okayama, 1987). Stable transfectants were obtained after selection in 800 µg/ml of G418 (Gibco-BRL) for 2 weeks and maintained thereafter in 250 µg/ml of G418. Cells expressing high levels of _v3 receptor were identified by FACS with LM609 monoclonal antibodies (Chemicon, Temecula, CA).

FACS analysis. FACS analysis was performed by use of standard protocols. Cells were harvested by ethylenediaminetetraacetic acid (0.02%, Gibco) treatment and washed twice with PBS and resuspended at a concentration of 1 × 10⁶ cells/ml in PBS. Cells were incubated with primary antibodies (1:250 dilution) for 1 h on ice and then washed twice with PBS to remove excess primary antibody. Cells were then incubated with fluorescein isothiocyanate-conjugated rabbit anti-mouse secondary antibody (1:250 dilution, Zymed, San Francisco, CA) for 1 h on ice. Cells were washed twice with PBS and resuspended in PBS for FACS analysis on a Becton Dickinson FACvantage (Mountain View, CA).

Radioligand binding measurements. To determine the affinity of ¹²⁵I-echistatin for _v3 integrin on 293 cells, binding isotherms of the interaction between radiolabeled echistatin and 293 cells were generated. Echistatin radiolabeled by the lactoperoxidase method to a specific activity of 2000 Ci/mmol (Amersham, Chicago, IL) was used. For binding assays, cells were harvested and resuspended (2 × 10 ⁶ cells/ml) in adhesion buffer containing 1× Hanks' balanced salt solution lacking divalent cations, 50 mM HEPES (pH 7.4), 1 mg/ml of bovine serum albumin, .5 mM MnCl₂ and 2 mM CaCl₂. A concentration range of ¹²⁵I-echistatin was added to the A4 (_v3 ve) or the B10 (_v3 +ve) 293 cells in suspension (2 × 10 ⁵ cells/well) in 96-well microtiter plate (Falcon Microtest III) and the mixture was incubated for 2 h by shaking at room temperature. At the end of the incubation period, the cells were filtered with use of Millipore Multiscreen-FB (glass fiber type B) plates which had been pretreated with 100 µl of 0.3% polyethylenimine solution for 2 h. The filters were then washed three times with 100 µl of adhesion buffer. The plates were allowed to dry and the individual filters were punched out and counted in a gamma counter. Nonspecific binding was measured in the presence of 200-fold molar excess of echistatin and was subtracted from the total binding to yield specific binding. Each data point is an average of values from triplicate wells. All measurements were repeated at least three times yielding identical results. Bound ligand was calculated from the specific activity of the ligand and the results are presented as picomoles bound per million cells. Scatchard plots were derived by plotting bound/free ligand against ligand bound (pmol/million cells). The binding affinity (K_d) of echistatin to cell surface bound _v3 is derived from the slope of the plot (Scatchard, 1943). The K_d values were also calculated by analyzing the data with nonlinear regression by use of Graph Pad Prism (version 2) and identical results were obtained by both methods (Munson and Rodbard, 1980).

Cell adhesion measurements. Forty-eight-well plates (Costar, Cambridge, MA) were coated overnight at 4°C with 20 µg/ml of vitronectin in PBS, followed by blocking nonspecific sites with 1% heat-treated BSA in PBS for 2 h at 37°C. HEK-293 cells, grown to confluence in 75-cm² flasks, were harvested with trypsin and resuspended in adhesion buffer (10 mM HEPES, pH 7.4, 140 mM NaCl, 5.4 mM KCl, 5.56 mM glucose, 3% bovine serum albumin, 2 mM CaCl₂, 1 mM MgCl₂, 1 mM MnCl₂) to a density of 5 × 10⁵ cells/ml. Cells were either coincubated with the indicated concentrations of competitor or allowed to adhere to vitronectin-coated wells in the absence of competing ligand for 10 min at 37°C. Unbound cells were then removed by rinsing wells three times with adhesion buffer. Bound cells were quantitated as described previously (Orlando and Cheresh, 1991). Adherent cells were fixed with 3% paraformaldehyde in PBS for 20 min at room temperature. Cells were then rinsed once with 0.1 M borate buffer, pH 8.5, and stained with 1% crystal violet in 0.1 M borate buffer, pH 8.5, for 20 min at room temperature. Wells were rinsed four times with 0.1 M borate buffer, and the dye was solubilized with the addition of 10% acetic acid for 20 min. The color was quantitated by measuring optical densities at 595 nm with a Microtek Plate Reader (Molecular Devices, Sunnyvale, CA).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Binding of echistatin to purified _v3 receptor. To characterize the binding of echistatin to integrin _v3 receptor in detail, we purified the receptor from human placenta with monoclonal antibody (LM609) affinity chromatography as described under "Materials and Methods" (Orlando and Cheresh, 1991). The identity and purity of the receptor was assessed by running the preparation on SDS-polyacrylamide gels followed by Western blot analysis with monoclonal antibodies specific for _v and ₃ subunits (Orlando and Cheresh, 1991). The binding of echistatin to purified _v3 was measured by a solid-phase receptor binding assay as described under "Materials and Methods." As shown in figure 1A, ¹²⁵I-echistatin binds to purified _v3 in a saturable and specific manner, because it is effectively competed by coincubation with cold echistatin. Incubation of _v3 receptor (50 ng) with increasing concentrations of ¹²⁵I-echistatin resulted in a saturable binding. Nonspecific binding was evaluated by carrying out the binding assay in the presence of a 200-fold molar excess of echistatin and was typically less than 10% of the total binding (fig. 1A). Scatchard analysis of the binding data gave a linear fit with a K_d of 0.33 nM and B_max of 750 pmol/mg protein as determined by the ligand computer program (results not shown). As described below, because echistatin binds to _v3 in a nondissociable manner, we can only assign an apparent K_d value. The data were also analyzed by nonlinear regression with Graph Pad Prism and identical results were obtained (Munson and Rodbard, 1980).

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Saturation binding isotherm and competitor concentrations required for half-maximal binding of 125I-echistatin to v3. (A) Saturation binding isotherms of 125I-echistatin binding to v3 receptor were determined in a solid-phase receptor binding assay as described under "Materials and Methods." Integrin v3 purified from human placenta was coated at a concentration of 10 ng/well onto Microlite-2 plates and incubated with various concentrations (0.05-5 nM) of 125I-echistatin for 3 h at room temperature. Bound ligand concentration was determined by solubilizing the counts with boiling 2 N NaOH and was subjected to gamma counting. Nonspecific binding was evaluated by carrying out the binding assay in the presence of 200-fold molar excess of cold echistatin and was subtracted from the total binding to calculate specific binding. , total binding; , specific binding (SP); , nonspecific binding (NSP). Each data point is an average of triplicate measurements in which the error was less than 5% of the total binding. To derive the affinity of the interaction between echistatin and v3, the data shown in panel A were analyzed by nonlinear regression analysis with the Graph Pad Prism program and also according to the method of Scatchard (1943). (B) Purified receptor was coated onto Microlite-2 plates at a concentration of 50 ng/well as described above. 125I-Echistatin was added to the wells to a final concentration of 0.05 nM in binding buffer (50 µl/well) in the presence of competing ligand. Cold unlabeled echistatin (), GRGDSP (), and Gpen RGDSp () peptides dissolved in binding buffer at the concentrations indicated were added to the wells before the addition of radioligand. After a 3-h incubation at room temperature, the wells were washed and radioactivity was determined with Top count (Packard). All measurements were done in triplicate with standard deviations less than 5%.

¹²⁵I-Echistatin binds to _v3 in a nondissociable manner. Previous studies have shown that vitronectin and fibronectin bind to integrin _v3 in a nondissociable manner, whereas the binding of an RGD peptide derived from vitronectin to _v3 is specific but is completely dissociable with a K_d of 9.4 × 10⁷ M (Orlando and Cheresh, 1991). The interaction of _v3 with the ligands vitronectin and fibronectin involves the initial integrin-ligand recognition event, which is RGD dependent and fully dissociable, followed by stabilization of the receptor-ligand complex leading to a nondissociable interaction between these proteins. To determine the nature of interaction between ¹²⁵I-echistatin and _v3 receptor, the radiolabeled ligand was preincubated with the receptor for 3 h, the unbound ligand was removed and the wells were washed with binding buffer, followed by the addition of variable amounts of unlabeled competitors. Under these conditions, the binding of radiolabeled echistatin to _v3 cannot be competed by cold echistatin, linear RGD and cyclic RGD peptides, which indicates that radiolabeled ¹²⁵I-echistatin binds to _v3 in a nondissociable manner (fig. 2A). When 6 mM ethylenediaminetetraacetic acid was added after 10 min of incubation of ¹²⁵I-echistatin with _v3, there was very little dissociation of echistatin from _v3, which suggests that binding is irreversible even with shorter incubation times (data not shown). When the competition for bound ligand was performed for longer periods (up to 40 h), there was minimal competition by the competing RGD peptides (fig. 2B), which suggests that binding of echistatin to _v3 results in a highly stabilized association, similar to vitronectin binding to _v3. However, the bound ligand can be eluted from the plates by hot SDS solution and the complex can be dissociated on SDS-PAGE, which suggests that the nature of binding is noncovalent (data not shown). When the _v3 receptor was saturated with native unlabeled echistatin for 3 h before removing the unbound ligand and replacing it with radiolabeled ligand, there was very little binding of the radioligand, which suggests that the native echistatin binds to the receptor in a nondissociable manner (data not shown).

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Lactoperoxidase labeled 125I-echistatin binds to v3 in a nondissociable manner. (A) The binding of 125I-echistatin to v3 was determined by solid-phase receptor assay as described in the legend to figure1. For coincubation conditions, 125I-echistatin was added to a final concentration of 0.05 nM in the presence of competing peptides at the concentrations indicated. After a 3-h incubation at room temperature, the wells were washed and radioactivity was determined by gamma-counting. [, Echistatin (ECH) coincubation; , linear RGD peptide coincubation; , cRGD peptide coincubation] For preincubation conditions, 125I-echistatin was incubated with the receptor for 3 h at room temperature. Wells were washed thoroughly to remove unbound ligand and incubated for an additional 3 h at room temperature with 100 µl/well of blocking/binding buffer containing the indicated concentrations of competing peptides. (, echistatin preincubation; , linear RGD peptide preincubation; , cRGD peptide preincubation) (B) The dissociability of the lactoperoxidase-labeled 125I-echistatin was determined by competition studies by preincubating the receptor with radiolabeled echistatin for 3 h and removing the unbound ligand, washing with binding buffer and then incubating with either buffer (). or 50 nM of cold echistatin () or 500 nM of linear RGD () or 500 nM of () cyclic RGD peptide for different times. The concentration of bound ligand was determined at various times as described in the legend to figure 1. The amount of labeled ligand bound to v3 was given an arbitrary value of 100%. All points represent triplicate determinations with an error of less than 5%.

Association kinetics of echistatin interaction with _v3 receptor. To further characterize the ligand binding characteristics of _v3 for echistatin, the association rate for echistatin interaction with _v3 was determined. Radiolabeled echistatin binding to _v3 at different time points was determined for several ligand concentrations by the solid-phase receptor binding assay (fig. 3A). Assuming pseudo first-order kinetics, because the amount of unbound ligand far exceeded the amount bound, K_obs versus ligand concentration was plotted and the rate constant was determined by the slope of the plot (fig. 3B). The association rate constant for the binding of echistatin to _v3 was calculated as 6.33 × 10⁵ s¹ M¹. The values determined for vitronectin and VN-peptide were 9.0 × 10³ s¹ M¹ and 1.6 × 10⁴ s¹ M¹, respectively (Orlando and Cheresh, 1991). This result shows that echistatin binds to _v3 at a significantly faster rate than either vitronectin or the vitronectin-derived peptide. It should be pointed out that the VN-derived peptide is a linear 15-mer containing the RGD motif, whereas echistatin folds into a rigid core structure stabilized by four disulfide bridges. The RGD sequence is located in a mobile loop-like structure enabling it to fit into the ligand binding site of the receptor.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Determination of the association rate constant for echistatin binding to v3. (A) To determine the pseudo first-order time courses, 125I-echistatin was incubated with v3 at concentrations of 0.05 nM (), 0.1 nM (), 0.2 nM () and 0.5 nM (). At the indicated times, wells were rinsed and radioactivity determined as described under "Materials and Methods." Nonspecific binding for each time was determined by coincubation of 125I-echistatin in the presence of excess cold echistatin and represented 10% of the total ligand bound. (B) The measurements of the slopes of pseudo first-order time courses (Kobs) versus a range of ligand concentrations were plotted. The slope of this plot represents the association rate constant. The R2 value for this analysis is 0.973.

Generating stable cell lines to study the binding between echistatin and _v3 integrin receptor. To characterize the binding of echistatin to _v3 receptor expressed on the cell surface, we chose the human embryonic kidney 293 cells because they lack the _v3 receptor (Bodary and McLean, 1990; Bodary et al., 1989). These cells express the endogenous _v1 receptor (Bodary and McLean, 1990). Thus the wild-type cells serve as a model for measuring echistatin binding to _v1. To generate a cell line with which we could measure echistatin binding to _v3, the 293 cells were transfected with the pCDNA3 expression vectors carrying human _v and ₃ cDNAs. Stable transfectants obtained after selection with G418 were analyzed by FACS analysis with LM 609 monoclonal antibodies that specifically recognize _v3 integrin receptors. The integrin expression profile of the wild-type and _v3-transfected 293 cells was compared by FACS analysis. As shown in figure 4, these studies confirm that 293 cells do not express _v3 because they lack the ₃ subunit. However, they do express the _v subunit, which associates with the ₁ subunit to form the _v1 receptor (fig. 4C). After transfection with the _v and ₃ cDNAs, the 293 stable cell line (B10) expresses the _v3 heterodimer on the cell surface (Fig. 4D). The stable line B10 displays a 10-fold greater binding of LM 609 antibody than the parental 293 cells or the G418 resistant line (A4) that is negative for _v3 expression (data not shown).

View larger version (K): [in this window] [in a new window]   Fig. 4.   FACS analysis of integrin expression on human embryonic kidney 293 cells. A panel of monoclonal antibodies was used to assess integrin expression on wild-type and v and 3 transfected human kidney 293 cells. Cells were incubated with mouse IgG or with the noted primary antibodies and then with secondary fluorescein thiocyanate-conjugated goat anti-mouse IgG. After extensive washing to remove free antibody, the cells were analyzed by flow cytometry. The expression level of each integrin subunit is indicated by the mean fluorescence intensity. The integrin expression profile of wild-type 293 cells was analyzed with mAb LM609 against v3 integrin receptor (A), 14H4 against v (B), 550036 mAb against 3 (C). After transfection of 293 cells with cDNAs for v and 3 subunits in pCDNA3 expression vector, the G418 resistant cell lines were analyzed with mAb LM609 (D). Cells negative for v3 expression (clone A4) and resistant to G418 showed a profile identical with wild-type 293 cells (not shown).

Binding of echistatin to _v3 expressed on 293 cells. To measure the binding of echistatin to _v3 expressed on the cell surface, _v3-transfected 293 cells (B10) were harvested from tissue culture flasks, placed in suspension and incubated with ¹²⁵I-echistatin for 2 h. G418 resistant line (A4) that is negative for _v3 expression was used as a control in these experiments. The B10 cell line binds a 5- to 10-fold higher amount of radiolabeled echistatin than the _v3 negative A4 clone (fig. 5, A and B). The binding of ¹²⁵I-echistatin to B10 clone is competed by linear and cyclic RGD peptides, but not by the peptide containing the RGE sequence (fig. 5A). Addition of increasing concentrations of radioligand to the cells resulted in a linear increase in the total amount of radioligand bound to the B10 cells (fig. 5B). However, there was no appreciable binding to the _{v3 (ve)} A4 cells, which suggests that the endogenous _v1 receptor binds poorly to echistatin.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Binding of 125I-echistatin to v3 expressed on 293 cells. HEK-293 cells (A4 v3 ve and B10 v3 +ve) were harvested from tissue culture flasks and were resuspended in adhesion buffer containing 1 mM Mn++. 125I-Echistatin (1 nM) was added to the cells in the presence of competitors (A). Cold echistatin (5 nM), linear RGD peptide (500 nM) and linear RGE peptide (500 nM) were added to the cells before the addition of radioligand and the mixture was allowed to incubate with shaking for 2 h at room temperature. Bound ligand was separated from free ligand by filtration through microtiter plates with glass fiber filters at the bottom of the wells, as described under "Materials and Methods." The wells were washed and radioactivity was determined by punching the membranes out and counting by gamma counting. Each point is the average of triplicate data points and the results shown are a representative of at least three experiments. (B) A4 and B10 cells (2 × 105 cells/well) were incubated with increasing concentration of 125I-echistatin for 2 h at room temperature and the bound ligand was estimated as described above.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   A measurement of the binding affinity between 125I-echistatin and v3 expressed on 293 cells. Isotherms of 125I-echistatin binding to v and 3 cDNA transfected 293 cells (clone B10) maintained in suspension were generated. Cells were harvested from tissue culture flasks and were resuspended in adhesion buffer containing 1 mM Mn++. 125I-Echistatin of increasing concentration was added to the cells and the mixture was allowed to incubate with shaking for 2 h at room temperature. Bound ligand was separated from free ligand by filtration as described above. Each point is the average of triplicate data points, and the isotherm shown represents at least three experiments. (, total 125I-echistatin bound to the receptor; , specific binding; , nonspecific binding. To derive the number of receptors expressed on 293 cells, the data were replotted according to the method of Scatchard (1943). Bmax is the x-intercept. The R2 value for this analysis is 0.987.

View larger version (14K): [in this window] [in a new window]   Fig. 7.   Competitor concentrations required for half-maximal binding of 125I-echistatin to v3 expressed on HEK-293 cells. Conditions of 125I-echistatin binding to v3 expressed on 293 cells are exactly as described in the legend to figure 5. 125I-Echistatin was added to the cells to a final concentration of 1 nM in binding buffer (50 µl/well) in the presence of competing unlabeled echistatin (), GRGDSP (linear peptide) () and Gpen RGDSP (cyclic peptide) (), dissolved in binding buffer at the concentrations indicated. After a 2-h incubation at room temperature with shaking, the bound ligand was separated from free ligand by filtration with microtiter plates with glass fiber filters in a vacuum manifold. All measurements were done in triplicate with standard deviations less than 5%.

Echistatin inhibits the adhesion of _v3 expressing 293 cells to vitronectin matrix. A cell adhesion assay was established as an additional way to examine the ligand binding characteristics of cell surface bound _v3. We compared the adhesion of parental 293 cells expressing _v1 receptor versus transfected 293 cells expressing _v3 receptor by allowing these cells to adhere to immobilized vitronectin or BSA. The transfected cells (B-10 cells) adhered very efficiently to vitronectin, but did not adhere to echistatin (result not shown) or BSA (fig. 8A). The inability of echistatin to support the adhesion of cells could be caused by its poor binding to plastic, because peptides and sometimes even polypeptides of more than 10,000 daltons show little or no integrin-binding activity in the adhesion assay. The parental 293 cells adhere very poorly to vitronectin, which suggests that the adhesion of B10 cells to vitronectin is mediated by the transfected _v3 receptor. The adhesion of B-10 cells to vitronectin can be blocked by echistatin, RGD peptides and LM609 antibodies that specifically recognize the _v3 receptor (fig. 8B). Peptides containing the RGE sequence do not compete for adhesion to vitronectin. These results show that echistatin can bind to the _v3 receptor efficiently and acts as an antagonist for this receptor presumably by competing for the binding of vitronectin to the RGD recognition site.

View larger version (31K): [in this window] [in a new window]   Fig. 8.   Echistatin antagonizes the adhesion of 293 cells expressing v3 receptor. (A) The adhesion of wild-type (open bars) and v and 3 transfected (dark bars) 293 cells to BSA and vitronectin was determined as described under "Materials and Methods." Cells were allowed to attach for 10 min at room temperature to 48-well plates coated with BSA (3 mg/ml) or vitronectin (10 µg/ml). Each point is the mean ± S.D. of triplicates. Unbound cells were then removed by rinsing wells three times with adhesion buffer. Bound cells were quantitated as described under "Materials and Methods." (B) The adhesion of v3 transfected 293 cells to vitronectin was challenged with echistatin (10 µM), linear RGD peptide (10 µM), linear RGE peptide (10 µM) and LM609 monoclonal antibodies (10 µg/ml).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

In this study, we have carried out a detailed biochemical characterization of the binding of echistatin to _v3 receptor. We have shown that echistatin binds to purified _v3 with high affinity and that both the native and radiolabeled echistatin bind to _v3 in an irreversible manner. Hence, one can only assign an apparent K_d value by Scatchard analysis. The apparent K_d value agrees closely with the K_i value (0.27 nM) for native echistatin determined by competition experiments. The binding of echistatin to _v3 is competed by linear and cyclic RGD peptides with K_i values of 445 ± 25 nM and 183 ± 17 nM, respectively, which indicates that the binding is through the RGD sequence.

Our studies with stable 293 cells expressing _v3 receptor suggest that echistatin binds poorly to the endogenous _v1 receptor expressed on 293 cells. This conclusion is based on the observation that ¹²⁵I-echistatin binds very poorly to _v3 negative A4 clone. The number of _v1 receptors expressed on 293 cells is about 51,000 (Hu et al., 1995). This number is similar to the number of _v3 receptors (56,000) expressed in the stable cell line B12. Hence the lack of binding of echistatin to 293 cells cannot be explained by the low number of _v1 receptors expressed on 293 cells. The K_i value determined for unlabeled echistatin in competition experiments with 293 cells (0.24 nM) is in agreement with the value obtained with purified receptor (0.27 nM), which suggests that echistatin binds preferentially to _v3 receptor expressed in the stable 293 cells.

Previous studies have shown that ¹²⁵I-vitronectin binds to _v3 in a nondissociable manner. In contrast, ¹²⁵I-VN-derived peptide containing the RGD sequence rapidly dissociates from the receptor into the aqueous phase with a dissociation rate of 1.6 × 10⁴ s¹ (Orlando and Cheresh, 1991). Vitronectin and VN-derived peptide associate with the receptor with an association rate of 9.0 × 10³ s¹ M¹ and 1.7 × 10² s¹ M¹, respectively. Thus, vitronectin has a 50-fold higher association rate than the VN peptide. In contrast, echistatin appears to bind to _v3 at a much faster rate with a rate constant of 6.33 × 10⁵ s¹ M¹. The structure of echistatin in aqueous solution has been determined by nuclear magnetic resonance spectroscopy (Saudek et al., 1991). The protein has been shown to fold in a series of irregular loops to form a rigid core stabilized by four disulfide bridges (Calvete et al., 1992). The RGD sequence is located in a mobile loop at the tip of the hairpin. Apart from the restriction imposed by the strands of the hairpin, the RGD recognition site is very mobile and exposed (Saudek et al., 1991; Brockel et al., 1992). Such behavior is typical of small segments of proteins whose function is fast recognition and fitting (Williams, 1989). The hand-in-glove model describes how the speed is paid for by somewhat reduced specificity (Williams, 1978).

It is worth noting an important biochemical distinction between vitronectin and echistatin. Vitronectin exists as a multimer containing between 12 and 15 moieties per multimer (Bittorf et al., 1993). There is also evidence that multimeric vitronectin is also present in extracellular matrices in vivo. In contrast, echistatin used in these studies is a monomer as determined by mass spectral analysis (data not shown). The multimeric nature of vitronectin results in higher nonspecific binding in solid-phase receptor binding assays. On the other hand, echistatin shows very little nonspecific binding in these assays, and binds to _v3 with high affinity and in an irreversible manner similar to vitronectin. These results suggest that echistatin could be used an alternate ligand for high-throughput screening of _v3 antagonists.

This study demonstrates that even though echistatin is a 49-amino-acid-long peptide, it binds to _v3 receptor in an irreversible manner, similar to vitronectin. It has been shown that the interaction of _v3 with vitronectin involves the initial integrin-ligand recognition event, which is RGD dependent and fully dissociable followed by stabilization of the integrin-ligand complex leading to a nondissociable interaction. These results suggested that the primary event of integrin _v3 substrate recognition involve the binding of the RGD sequence followed by additional molecular interactions resulting in a highly stabilized receptor-ligand association. It is suggested that this stabilized molecular interaction between receptor and ligand may be necessary for the transduction of signals between extracellular matrix and the intracellular compartment. Recently, a hierarchy of molecular responses leading from initial integrin interactions with an extracellular ligand, to trans membrane effects on the localization of cytoskeletal proteins or signaling molecules, to the activation of signaling pathways and to eventual regulation of gene expression, has been described (Miyamoto et al., 1995). Echistatin does not support the adhesion of cells because of poor binding to plastic. However, a method by which cytoskeletal and signaling responses to integrin-ligand interaction with beads coated with a variety of molecules has been described (Miyamoto et al., 1995). It would be interesting to see if the binding of echistatin to _v3 leads to signal transduction events similar to vitronectin binding or induces an apoptotic response.