- Email: [email protected]
One-Step Affinity Purification of Recombinant αvβ3Integrin from Transfected Cells
PROTEIN EXPRESSION AND PURIFICATION ARTICLE NO.
8, 68–74 (1996)
0075
One-Step Affinity Purification of Recombinant avb3 Integrin from Transfected Cells1 Cezary Marcinkiewicz,* Louis A. Rosenthal,† Mariola M. Marcinkiewicz,* Maria Anna Kowalska,* and Stefan Niewiarowski*,2 The Sol Sherry Thrombosis Research Center and *Department of Physiology and †Department of Microbiology and Immunology, Temple University School of Medicine, Philadelphia, Pennsylvania 19140
Received September 21, 1995, and in revised form April 2, 1996
Vitronectin receptor (avb3 integrin) is present on the surface of many types of cells. We describe a simple, fast, and reliable method of purification of recombinant human avb3 from Chinese hamster ovary (CHO) cells transfected with avb3 (VNRC3 cells). The method consists of two steps: lysis of the cells and affinity chromatography of the lysate on a GRGDSPK–Sepharose column. The yield of the procedure was about 79%. The purified receptor migrated as two bands on a silver stained SDS–polyacrylamide gel, corresponding to the av and b3 subunits, and was recognized by monoclonal antibodies directed against av and the avb3 complex, but not by monoclonal antibody specific for the aIIbb3 complex. This receptor also bound to immobilized vitronectin, von Willebrand factor, and echistatin. However, binding to immobilized fibrinogen was not observed. Purified recombinant avb3 demonstrated greater immunoreactivity with LM 609, an avb3 complex-specific monoclonal antibody, than avb3 purified from placenta. As visualized by SDS–polyacrylamide gel electrophoresis, preparations of placenta-derived avb3 contained several contaminating proteins that were not present in preparations of recombinant avb3 purified from the transfected CHO cells. q 1996 Academic Press, Inc.
avb3 belongs to the integrin family of glycoproteins. Different members of this family are expressed on the surface of different types of cells. The predominant role of integrins is the mediation of cell–cell and cell–extracellular matrix interactions (1,2). avb3 , also named vitronectin receptor, can recognize fibrinogen (3), von Willebrand factor (3), thrombospondin (4), fibronectin (5), 1
This investigation was supported by NIH Grant HL 45486. To whom correspondence should be addressed. Fax: (215) 7074003 or 707-2783. 2
osteopontin (6), denatured collagen types I and IV (7), thrombin (8), and laminin (9). All of these ligands express the RGD sequence, which is critical for binding to avb3 . The av subunit contains a light chain (25 kDa) and a heavy chain (115 kDa) bound by an S–S bridge. The light chain has a cytoplasmic domain, hydrophobic transmembrane domain, and extracellular domain (10). The b3 subunit (92 kDa) is a single polypeptide, which has short cytoplasmic and transmembrane domains and a longer extracellular domain (11,12). Disintegrins echistatin and eristostatin are antagonists of integrins (13,14). Echistatin is a strong inhibitor of vitronectin binding to immobilized avb3 , whereas eristostatin is much less active in this system (15). On the other hand, eristostatin is the strongest inhibitor of fibrinogen receptor (aIIbb3 integrin) among disintegrins (16). The selectivity of these disintegrins depends upon the structure of their RGD-loop, determined by amino acids adjacent to RGD sequence (14). Vitronectin receptor (avb3 integrin) is constitutively expressed on endothelial cells, smooth muscle cells, macrophages, osteoclasts, and melanoma cells (17). Since avb3 is involved in angiogenesis, bone resorption, cancer metastasis, and smooth muscle proliferation (17), many areas of research would benefit from the availability of a simple and reliable method for purifying this receptor. Since the earlier report by Pytela et al. (18), vitronectin receptor has been routinely purified from human placenta, and several modifications of the original procedure have been reported (19). In our hands, purification of vitronectin receptor from placenta is a laborious, and not entirely reliable, method that yields low quantities of avb3 still contaminated with other proteins. In addition, the placenta-derived avb3 was contaminated with avb5 . For these reasons, we developed a method for purifying avb3 from Chinese hamster ovary (CHO) cells that express recombinant human avb3 .
68
AID
1046-5928/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
pepa
AP: PEP
PURIFICATION OF avb3 INTEGRIN
MATERIALS AND METHODS
CHO cells were cotransfected with expression constructs containing human av and b3 and a CDM8 vector containing the neomycin resistance gene (20,21). The transfected cells (VNRC3 cells) are the generous gift of Drs. Mark Ginsberg and Joseph Loftus (Scripps Research Institute, La Jolla, CA). A2A9, a monoclonal antibody specific for human aIIbb3 (22), was kindly provided by Dr. J. Bennett (University of Pennsylvania, Philadelphia, PA). LM 609 and LM 142, monoclonal antibodies specific for the human avb3 complex and av , respectively (23), were provided by Dr. D. Cheresh (Scripps Research Institute). The monoclonal antibodies 7E2 and PB1 specific for hamster a5 and b1 , respectively (24), were kindly provided by Dr. R. L. Juliano (University of North Carolina, Chapel Hill, NC). Antihuman vitronectin polyclonal antibody, developed in rabbit, was from Gibco BRL (Gaithersburg, MD). GRGDSPK peptide was synthesized by Dr. B. Jameson (Thomas Jefferson University, Philadelphia, PA) and coupled to CNBr–Sepharose 4B (Pharmacia, Piscataway, NJ) in the proportion 20 mg peptide per 1 ml of slurry. n-Octylglucoside (1-O-n-octyl-b-D-glucopyranoside) and Hanks’ balanced salt solution (HBSS) were acquired from Boehringer-Mannheim (Indianapolis, IN) and Gibco BRL, respectively. Human vitronectin was purchased from Sigma (St. Louis, MO) or from Calbiochem (La Jolla, CA), and human von Willebrand factor was obtained from Hematologic Technologies (Essex Junction, VT). Highly purified human fibrinogen was a gift from Dr. A. Budzynski (Temple University, Philadelphia PA). All other reagents were purchased from Sigma. Disintegrins echistatin and eristostatin were purified from the crude venom of the snakes Echis carinatus and Eristocophis macmahoni, respectively, using two steps of C-18 reverse-phase high-performance liquid chromatography (25). Both preparations were highly homogeneous as determined by SDS–polyacrylamide gel electrophoresis (SDS–PAGE) and N-terminal amino acid sequencing (16). avb3 integrin was purified from VNRC3 cells. Cells were grown in Dulbecco’s modified Eagle medium containing 10% fetal calf serum, nonessential amino acids, glutamine, penicillin, and streptomycin. Cells were detached from plates with calcium- and magnesium-free HBSS and 5 mM EDTA and washed three times with HBSS containing calcium and magnesium. All preparative steps were performed at room temperature. About 2 1 108 cells were lysed in the presence of 8 ml of buffer A: 100 mM n-octylglucoside, 50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 1 mM MgCl2 , 1 mM CaCl2 , and 1 mM PMSF. After removal of insoluble material by centrifugation at 3000g, the extract was concentrated to 1 ml using Centriprep 10 (Amicon, Beverly, MA). The
AID
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
69
GRGDSPK–Sepharose 4B column (1 ml) was equilibrated with buffer B: 50 mM n-octylglucoside, 50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 2 mM MgCl2 , and 0.1 mM CaCl2 . One milliliter of lysed cell extract was divided into two 0.5-ml aliquots. After application of the first aliquot, the column was incubated for 30 min to increase binding of integrin to resin. Then the remaining 0.5-ml aliquot was applied, followed by another 30-min incubation. The column was washed with buffer B and then with a modified buffer B containing 300 mM NaCl. The receptor was eluted with a modified buffer B containing 15 mM EDTA. One-milliliter fractions were collected in vials containing 30 ml of 1 M MgCl2 . Fractions interacting with the anti-avb3 monoclonal antibody LM 609 were combined, concentrated to a volume of about 1 ml, and dialyzed against buffer B. Purified avb3 was stored at 0207C in small aliquots. avb3 from placenta was purified as described earlier (18,19). Briefly, fresh human placenta tissue was homogenized in buffer containing 20 mM Tris–HCl, pH 7.4, 10 mM EDTA, 1 M NaCl, 10 mM sodium azide and 0.3 mM PMSF, and the homogenate was stirred for 48 h at 47C. The supernatant was removed by centrifugation (1 h, 8000g), and the pellet was homogenized again in buffer C: 50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 1 mM MgCl2 , 1 mM MnCl2 , 0.1 mM CaCl2 , 0.5 mM PMSF, 1 mg/ml each of aprotinin, antipain, and leupeptin, and 25 mM b-D-octylglucoside. The homogenate was stirred overnight at 47C, and the supernatant was collected following centrifugation at 8000g. The extraction was repeated once (3 h), and the combined supernatants were centrifuged at 30,000g and stored at 0207C in 100- to 200-ml aliquots. Extracts were thawed and diluted 1:1 with buffer C without PMSF. The extract (200 ml) was applied to a GRGDSPK–Sepharose column equilibrated with buffer C overnight. Fractions containing placental avb3 were eluted from the column and identified as described in the procedure for purification of recombinant avb3 . ELISA plates (EIA/RIA, Costar, Cambridge, MA) were coated with purified avb3 , extracellular matrix proteins, or disintegrins in 0.05 M carbonate/bicarbonate buffer (pH 9.3) and incubated overnight at 47C. The wells were blocked with 5% nonfat milk in PBS/Tween (PBST) buffer. Purified avb3 (or lysate of VNRC3 cells) dissolved in buffer B was added to wells coated with matrix proteins or disintegrins, and the plates were incubated for 30 min at 377C. Binding of murine monoclonal antibodies to the receptor was measured using goat anti-mouse IgG conjugated with alkaline phosphatase, as described previously (26). Binding of vitronectin to immobilized avb3 was measured by the method described previously (15). SDS–PAGE was performed using the Laemmli method (27) under reducing conditions. Bands were visualized by silver staining. Protein concentrations
pepa
AP: PEP
70
MARCINKIEWICZ ET AL. TABLE 1
Efficiency of Purification of avb3 from VNRC3 Cells
Number of CHO cells VNRC3 2.2 1 108 CHO-K1 2.2 1 108
Total protein in lysate (mg)
avb3 in lysate (mg)
Protein in fractions eluted from GRGDSPKcolumn with EDTA (mg)
15.2 { 2.2
0.17 { 0.03
0.14 { .04
79 { 6
15.0 { 2.8
Not detectable
0.05 { 0.01
—
Yield (%)
Note. The amount of avb3 in whole cell lysates from CHO cells transfected with avb3 (VNRC3 cells) and nontransfected CHO-K1 cells was determined by ELISA. avb3 present in cell lysates was allowed to bind to immobilized echistatin (140 ng/well) and was then detected by monoclonal antibody LM 609. The amount of receptor was calculated from standard curve. Cell lysates were then applied to a GRGDSPKcolumn, and the amount of protein in fractions eluted with EDTA was determined. The data represent the mean with standard deviation from three independent experiments.
were determined using the BCA protein assay reagent (Pierce Co. Rockford, IL), with bovine serum albumin serving as the standard. RESULTS
We purified human recombinant avb3 integrin from CHO cells transfected with avb3 (VNRC3 cells) by affinity chromatography on a GRGDSPK–Sepharose column. The purification procedure included two steps: lysis of cells and chromatography of the lysate on an affinity column. This method was based on the observation that avb3 integrin binds to ligands via their RGD sequences. The receptor retained on the column was released by EDTA. The fractions were collected in tubes containing a high excess of Mg2/, which stabilizes avb3 . As a control, an identical procedure was carried out on nontransfected CHO-K1 cells. Table 1 presents the avb3 content in VNRC3 cells and the efficiency of the described purification method. This purification procedure was very efficient, yielding approximately 79% of the avb3 present in the VNRC3 cell lysates. The amount of avb3 in the cell lysates was determined using a ligand capture ELISA assay, in which the receptor that selectively bound to immobilized echistatin was detected by an avb3 complex-specific monoclonal antibody (LM 609). The ligand capture ELISA was used because the content of the receptor in the cell lysates was too low to be detected by conventional ELISA. A standard curve was constructed using known concentrations of previously purified avb3 . The level of avb3 in the VNRC3 cell lysates amounted to 1.1% of the total protein content. Figure 1 shows the ability of each fraction, eluted with buffer B containing 300 mM NaCl and 15 mM EDTA, to react with monoclonal antibody LM 609. Fractions 2–6, which demonstrated the highest absorption at 405 nm in the ELISA assay, were combined and concentrated. This figure also shows that fractions
AID
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
eluted from CHO-K1 cell lysate applied to the same column were not immunoreactive. The placental extract was applied to the GRGDSPK–Sepharose column, and immunoreactive material was eluted in the same fractions as in experiments with lysates of VNRC3 cells. The immunoreactive fractions of both preparations of the receptor were adjusted to the same protein concentrations and then analyzed by SDS– PAGE. Figure 2 compares purity of avb3 preparations obtained from VNRC3 cells (lane A) and from placenta (lane B). Both preparations showed identical bands corresponding to the av heavy chain (Ç110 kDa) and b3 subunit (Ç95 kDa). However, the placental preparation was more heavily contaminated with bands migrating with lower molecular weights. In parallel, the nonimmunoreactive fractions from CHO-K1 cell lysate that were adsorbed and eluted from the GRGDSPK– Sepharose column were collected, adjusted to the same protein concentration as material obtained from VNRC3 cells and from placenta, and analyzed by SDS– PAGE. Figure 2 (lane C) shows that fractions corresponding to av and b3 subunits were absent. On the other hand, this material separated into two bands that migrated with different mobility than a and b3 integrins. Since CHO cells express a5 and b1 integrin (28,29) the preparation obtained from CHO-K1 cells after the GRGDSPK–Sepharose column was examined for interaction with anti-a5 hamster monoclonal antibody (7E2) and anti-b1 hamster monoclonal antibody (PB1) raised in mice. However, these antibodies did not react with CHO-K1 preparation in ELISA assay (data not shown), suggesting that the bands observed in Fig. 2, lane C, did not correspond to hamster a5b1 . Interestingly, these two bands were absent in preparation obtained from transfected cells (line A) and from placenta (line B). Figure 3 compares the reactivity of LM 609 with three preparations eluted from the GRGDSPK–Sepharose column. The avb3 from VNRC3 cells was dramati-
pepa
AP: PEP
PURIFICATION OF avb3 INTEGRIN
71
FIG. 1. Interaction of fractions eluted with EDTA from GRGDSPK–Sepharose column with monoclonal antibody LM 609, specific for avb3 . Fractions (100 ml), after dialysis, were immobilized overnight on ELISA plates in carbonate/bicarbonate buffer, pH 9.3. Plates were blocked by addition of 5% milk in PBST buffer to each well, and then the monoclonal antibody LM 609 was added. After incubation for 60 min at 377C, the plates were washed with PBST buffer, and alkaline phosphatase-conjugated goat anti-mouse IgG was added. Plates were incubated for 60 min at 377C, and then after washing with PBST, the substrate for alkaline phosphatase p-nitrophenyl phosphate (pNPP) was added. After color development, the plates were read using an ELISA reader at 405 nm single wavelength. Absorbance at 405 nm represents spectrophotometric reading for the product of alkaline phosphatase reaction with substrate pNPP. Open circles indicate fractions obtained from VNRC3 cells, and filled circles indicate fractions obtained from control CHO-K1 cells. Error bars represent the standard deviation from three independent experiments.
cally more reactive than the avb3 from placenta. As expected, the preparation from CHO-K1 cells was not reactive. Purified avb3 from VNRC3 cells was recognized by monoclonal antibody LM 142, although not as well as by LM 609 (Fig. 4). A2A9, a monoclonal anti-
body specific for aIIbb3 , did not react with recombinant avb3 (Fig. 4) or with placental preparation (not shown). Studies were performed to examine the ligand specificity of avb3 purified from VNRC3 cells. Figure 5A shows that avb3 bound strongly to immobilized vitronectin and von Willebrand factor; however, it did not bind to immobilized fibrinogen. Purified avb3 bound more strongly to the immobilized echistatin than to eristostatin (Fig. 5B). Figure 6 shows that vitronectin was efficiently bound by immobilized avb3 . DISCUSSION
FIG. 2. SDS polyacrylamide gel electrophoresis purified integrins. Lane A, avb3 purified from VNRC3 cells; lane B, avb3 from placenta; lane C, preparation obtained from CHO-K1 cells. Two-microgram aliquots of each preparation were loaded on the gel, and electrophoresis was performed under reducing conditions using Laemmli method followed by silver staining.
AID
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
We report a simple, short, and reliable method of purification of avb3 integrin from transfected cells. Most investigators use placenta as a source of human vitronectin receptor (18,19). However, this method of purification is laborious, and it did not guarantee the removal of other human integrins present in blood from the final preparation. In the case of transfected cells, this problem is avoided, and the purified avb3 is not contaminated with other human integrins. The absence of contaminating aIIbb3 was demonstrated using monoclonal antibody A2A9, which recognizes the aIIbb3 complex but not avb3 . Our preparations did not react
pepa
AP: PEP
72
MARCINKIEWICZ ET AL.
FIG. 3. Binding of LM 609 monoclonal antibody to avb3 purified from VNRC3 cells and from placenta. The purified receptors were immobilized on ELISA plates, and the experiment was performed as described in the Fig. 1 legend. Open circles show binding of LM 609 to avb3 from VNRC3 cells, filled circles to avb3 from placenta, and open triangles to a preparation from CHO-K1 cells. Error bars represent standard deviation from three independent experiments.
tal preparation. The preparation obtained from control CHO-K1 cells was not immunoreactive with antibodies against human integrins. It also contained three times less protein than the preparation obtained from VNRC3 cells. avb3 from VNRC3 cells did not seem to contain hamster integrins, as demonstrated by SDS–PAGE. It is known that a5b1 , an integrin expressed on CHO cells (28,29), has a much lower affinity for GRGDSPK than does avb3 (18). However, the preparation from CHO-K1 cells did not react with hamster’s anti-a5 and anti-b1 monoclonal antibodies. Therefore, it is likely that the CHO-K1 preparation contains integrins other than a5b1 and avb3 integrins which are also able to bind to immobilized GRGDSPK column. This ‘‘unknown’’ integrin probably has higher affinity to GRGDSPK–Sepharose column than a5b1 but lower than avb3 , because it was not found in the VNRC3 preparation. Other investigators have also confirmed absence of avb3 integrin on CHO cells (28,29). avb3 integrin stably expressed on the surface of CHO cells has the same properties as naturally occurring avb3 on human umbilical vein endothelial cells (HUVEC) (30). Suspension of both cells bind monoclonal antibodies specific for avb3 (LM 609, LM 142 and 7E3) and FITC-echistatin. Echistatin induces the ligand-induced binding site epitope recognized by monoclonal antibody Mab 62 (31) on the b3 subunit in both types of cells (32). Endothelial cells appear to be a major
with A2A9, but they were recognized by monoclonal antibodies LM 609 and LM 142, which recognize the avb3 complex and av , respectively. The preparation from placenta may contain integrin avb5 , which is very difficult to separate from avb3 . A similar amount of avb3 can be purified from one placenta (19) as from 5 1 108 VNRC3 cells. However, the time required to complete the latter purification procedure is significantly shorter. Our purification procedure takes 5–6 h, starting from harvesting the cells to final dialysis. Purification of receptor from placenta takes at least 3 days and requires many additional steps including washing and homogenization of placental tissue, stirring homogenate overnight at 47C, centrifugation at 8000g and then ultracentrifugation at 30,000g. The process of applying the placental extract to the affinity column took 14 h. The application of VNRC3 cell extract took only 1 h since the total amount of applied proteins was much lower than in placental extract. This can be advantageous because nonspecific absorption of other proteins, which may contaminate the final product, is avoided. SDS– PAGE demonstrated that the avb3 from placenta had additional bands representing contaminants, whereas the recombinant preparation had no additional bands and was apparently much more pure. Accordingly, the recombinant preparation showed higher immunoreactivity with monoclonal antibody LM 609 than did the placen-
FIG. 4. Binding of monoclonal antibodies to avb3 immobilized on ELISA plates. The purified receptor was immobilized on ELISA plates and binding of monoclonal antibodies was examined as described in the Fig. 1 legend. Filled circles, binding of LM 609; open circles, binding of LM 142; open triangles, binding of A2A9. Error bars represent the standard deviation from three independent experiments.
AID
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
pepa
AP: PEP
PURIFICATION OF avb3 INTEGRIN
73
FIG. 5. Binding of avb3 to extracellular matrix proteins (A) and to disintegrins (B) immobilized on ELISA plates. The ligands (extracellular matrix proteins or disintegrins) were immobilized overnight on ELISA plates in carbonate/bicarbonate buffer, pH 9.2. Plates were blocked, and the avb3 was added in buffer B. The incubation was performed for 30 min at 377C, and then binding of receptor was examined using LM 609 monoclonal antibody under conditions described in the Fig. 1 legend. Open circles, vitronectin (1 mg/well); filled circles, fibrinogen (1 mg/well); open triangles, von Willebrand factor (1 mg/well), open squares, echistatin (140 ng/well); filled squares, eristostatin (140 ng/ well). Error bars represent standard deviation from three independent experiments.
source of avb3 in placenta. We also succeeded in purifying avb3 from HUVEC (not shown), which may represent an alternate source of this receptor.
Purified, recombinant vitronectin receptor bound to vitronectin and von Willebrand factor but showed no interaction with fibrinogen. Conforti et al. (33) made the same observation with avb3 purified from placenta and found that binding of avb3 to fibrinogen increased in the presence of lipids and Mn2/ ions. Recombinant avb3 also reacted with disintegrins in a manner similar to that of placental avb3 (15). In summary, we described a simple, fast, and highyield method of purification of recombinant avb3 from transfected cells. The purified receptor shows the same immunological properties and ability to interact with ligands as receptor purified from other sources. We believe that our method will make it easier for other investigators to obtain useful quantities of functional, highly purified avb3 . ACKNOWLEDGMENTS The authors greatly appreciate the contribution of Drs. M. H. Ginsberg and J. C. Loftus (Scripps Research Institute), who provided cells transfected with avb3 . They thank Dr. D. A. Cheresh and Dr. J. S. Bennett for providing monoclonal antibodies against human avb3 , av , and aIIbb3 and Dr. R. L. Juliano for providing monoclonal antibodies against hamster a5 and b1 . The authors thank Mr. Adam Mohmand and Dr. Y. Wang for technical assistance in experiments.
FIG. 6. Binding of vitronectin to immobilized avb3 . avb3 was immobilized on ELISA plates as described in the Fig. 1 legend, and then vitronectin (1 mg/sample) in buffer B was added. Incubation was performed for 30 min at 377C, and binding of vitronectin was detected using anti-vitronectin polyclonal antibody as described under Materials and Methods. Error bars represent standard deviation from three independent experiments.
AID
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
REFERENCES 1. Hynes, R. O. (1992) Integrins: Versatility, modulation, and signaling in cell adhesion. Cell 69, 11–25. 2. Smyth, S. S., Joneckis, C. C., and Parise, L. V. (1993) Regulation of vascular integrins. Blood 81, 2827–2834.
pepa
AP: PEP
74
MARCINKIEWICZ ET AL.
3. Cheresh, D. A. (1987) Human endothelial cells synthesize and express an arg-gly-asp directed receptor involved in attachment to fibrinogen and von Willebrand factor. Proc. Natl. Acad. Sci. USA 84, 6471–6475. 4. Sun, X., Skorstengaard, K., and Mosher, D. F. (1992) Disulfides modulate RGD-inhibitable cell adhesive activity of thrombospondin. J. Cell. Biol. 118, 693–671. 5. Charo, I. F., Nannizzi, L., Smith J. W., and Cheresh, D. A. (1990) The vitronectin receptor avb3 binds fibronectin and acts in concert with a5b1 in promoting cellular attachment and spreading on fibronectin. J. Cell. Biol. 111, 2795–2800. 6. Miyauchi, A., Alvarez, J., Greenfield, E. M., Teti, A., Grano, M., Colucc, S., Zambonin- Zallone, A., Ross, F. P., Teitelbaum, S. L., Cheresh, D. A., and Hruska, K. A. (1991) Recognition of osteopontin and related peptides by an avb3 integrin stimulates immediate cell signals in osteoclasts. J. Biol. Chem. 226, 20369– 20374. 7. Clyman, R. I., Mauray, F., and Kramer, R. H. (1992) b1 and b3 integrins have different roles in the adhesion and migration of vascular smooth muscle cells on extracellular matrix. Exp. Cell. Res. 200, 272–284. 8. Bar-Shavit, R., Sabbah, V., Lampugnani, M. G., Marchisio, P. C., Fenton, J. W., Vlodavsky, I., and Dejana, E. (1991) An Arg-Gly-Asp sequence within thrombin promotes endothelial cell adhesion. J. Cell. Biol. 112, 335–344. 9. Kramer, R. H., Cheng, Y.-F., and Clyman, R. (1990) Human microvascular endothelial cells use b1 and b3 integrin receptor complexes to attach to laminin. J. Cell. Biol. 111, 1233–1243. 10. Suzuki, S., Argraves, W. S., Arai, H., Languino, L. R., Pierschbacher, M., and Ruoslahti, E. (1987) Amino acid sequence of the vitronectin receptor alpha subunit and comparative expression of adhesion receptor mRNAs. J. Biol. Chem. 262, 14080–14085. 11. Zimrin, A. B., Eisman, R., Vilaire, G., Schwartz, E., and Bennett, J. S. (1988) Structure of platelet Glycoprotein IIIa: A common subunit for two different membrane receptor. J. Clin. Invest. 81, 1470–1475. 12. Calvete, J. J. (1994) Clues for understanding the structure and function of a prototypic human integrin: The platelet glycoprotein IIb/IIIa complex. Thromb. Haemost. 72, 1–15. 13. Gould, R. J., Polokoff, M. A., Friedman, P. A., Huang, T. F., Cook, J. J., and Niewiarowski, S. (1990) Disintegrins: A family of integrin inhibitory proteins form viper venoms. Proc. Soc. Exper. Biol. Med. 195, 168–171. 14. Niewiarowski, S., McLane, M. A., Kloczewiak, M., and Stewart, G. J. (1994) Disintegrins and other naturally occurring antagonists of platelet fibrinogen receptors. Semin. Hematol. 31, 289– 300. 15. Pfaff, M., McLane, M. A., Beviglia, L., Niewiarowski, S., and Timpl, R. (1994) Comparison of disintegrins with limited variations in the RGD loop in their binding to purified integrins and in cell adhesion inhibition. Cell Adhes.Commun. 2, 491–501. 16. McLane, M. A., Kowalska, M. A., Silver, L., Shattil, S. J., and Niewiarowski, S. (1994) Interaction of disintegrins with the aIIbb3 receptor on resting and activated human platelets. Biochem. J. 301, 429–436. 17. Felding-Habermann, B., and Cheresh, D. A. (1993) Vitronectin and its receptors Curr. Opin. Cell Biol. 5, 864–868. 18. Pytela, R., Pierschbacher, M. D., Aggraves, S., Suzuki, S., and
AID
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
19.
20.
21.
22.
23.
24.
25.
26.
27. 28.
29.
30.
31.
32.
33.
Ruoslahti, E. (1987) Arginine–glycine–aspartic acid adhesion receptors. Methods Enzymol. 144, 475–489. Pfaff, M., Tangemann, K., Muller, B., Gurrath, M., Muller, G., Kessler, H., Timpl, R., and Engel, J. (1994) Selective recognition of cyclic RGD peptides of NMR defined conformation by aIIbb3, avb3, and a5b1 integrins. J. Biol. Chem. 269, 20233–20238. O’Toole, T. E., Loftus, J. C., Plow, E. F., Glass, A., Harper, J. R., and Ginsberg, M. H. (1990) Efficient surface expression of platelet GPIIb-IIIa requires both subunits. Blood 74, 14–18. O’Toole, T. E., Loftus, J. C., Du, X., Glass, A. A., Ruggeri, Z. M., Shattil, S. J., Plow, E. E., and Ginsberg, M. H. (1990) Affinity modulation of the aIIbb3 integrin (platelet GPIIb-IIIa) is an intrinsic property of the receptor. Cell Regul. 1, 883–893. Bennett, J. S., Hoxie, J. A., Leitman, S. F., Vilaire, G., and Cines, D. B. (1983) Inhibition of fibrinogen binding to stimulated human platelets by a monoclonal antibody. Proc. Natl. Acad. Sci. USA 80, 2417–2421. Cheresh, D. A., and Harper, J. R. (1987) Arg-gly-asp recognition by a cell adhesion receptor requires its 130-kDa a subunit. J. Biol. Chem. 262, 1434–1437. Brown, P. J., and Juliano, R. L. (1988) Monoclonal antibodies to distinctive epitopes on the a and b subunits of the fibronectin receptor. Exp. Cell Res. 177, 303–318. Williams, J., Rucinski, B., Holt, J. C., and Niewiarowski, S. (1990) Elegantin and albolabrin purified peptides from viper venoms: Homologies with the RGDS domain of fibrinogen and von Willebrand factor. Biochim. Biophys. Acta 1039, 81–89. Cook, J. J., Trybulec, M., Lasz, E. C., Khan, S., and Niewiarowski, S. (1992) Binding of glycoprotein IIIa–derived peptide 217– 231 to fibrinogen and von Willebrand factors and its inhibition by platelet glycoprotein IIb/IIIa complex. Biochim. Biophys. Acta 1119, 312–321. Laemmli, U. K. (1978) Clevage of structural proteins during the assembly of the head of bacteriophage T4 . Nature 227, 680–685. Schreiner, C. L., Bauer, J. S., Danilov, Y. N., Hussein, S., Sczekan, M., and Juliano, R. L. (1989) Isolation and characterization of Chinese hamster ovary cells variants deficient in the expression of fibronectin receptor. J. Cell Biol. 109, 3157–3167. Takada, Y., Ylanne, J., Mandelman, D., Puzon, W., and Ginsberg, M. (1992) A point mutation of integrin b1 subunit blocks binding of a5b1 to fibronectin and invasion but not recruitment to adhesion plaques. J. Cell Biol. 119, 913–921. Jaffe, E. A., Nachman, R. L., and Minick, C. R. (1973) Culture of human endothelial cells derived from umbilical veins. J. Clin. Invest. 52, 2745–2756. Du, X., Gu, M., Weisel, J. W., Nagaswami, C., Bennett, J. S., Bowditch, R., and Ginsberg, M. H. (1993) Long range propagation of conformational changes in integrin aIIbb3 . J. Biol. Chem. 268, 23087–23092. Juliano, D., Wang, Y., Marcinkiewicz, C., Rosenthal, L. A., Stewart, G. J., and Niewiarowski, S. (1996) Disintegrin interaction with avb3 integrin on human umbilical vein endothelial cells. Expression of ligand induced binding site on b3 subunit. Exp. Cell Res. 219, 571–578, 1996. Conforti, G., Zanetti, A., Pasquali-Ronchetti, I., Quaglino, D., Jr., Neyroz, P., and Dejana, E. (1990) Modulation of vitronectin receptor binding by membrane lipid composition. J. Biol. Chem. 265, 4011–4019.
pepa
AP: PEP
8, 68–74 (1996)
0075
One-Step Affinity Purification of Recombinant avb3 Integrin from Transfected Cells1 Cezary Marcinkiewicz,* Louis A. Rosenthal,† Mariola M. Marcinkiewicz,* Maria Anna Kowalska,* and Stefan Niewiarowski*,2 The Sol Sherry Thrombosis Research Center and *Department of Physiology and †Department of Microbiology and Immunology, Temple University School of Medicine, Philadelphia, Pennsylvania 19140
Received September 21, 1995, and in revised form April 2, 1996
Vitronectin receptor (avb3 integrin) is present on the surface of many types of cells. We describe a simple, fast, and reliable method of purification of recombinant human avb3 from Chinese hamster ovary (CHO) cells transfected with avb3 (VNRC3 cells). The method consists of two steps: lysis of the cells and affinity chromatography of the lysate on a GRGDSPK–Sepharose column. The yield of the procedure was about 79%. The purified receptor migrated as two bands on a silver stained SDS–polyacrylamide gel, corresponding to the av and b3 subunits, and was recognized by monoclonal antibodies directed against av and the avb3 complex, but not by monoclonal antibody specific for the aIIbb3 complex. This receptor also bound to immobilized vitronectin, von Willebrand factor, and echistatin. However, binding to immobilized fibrinogen was not observed. Purified recombinant avb3 demonstrated greater immunoreactivity with LM 609, an avb3 complex-specific monoclonal antibody, than avb3 purified from placenta. As visualized by SDS–polyacrylamide gel electrophoresis, preparations of placenta-derived avb3 contained several contaminating proteins that were not present in preparations of recombinant avb3 purified from the transfected CHO cells. q 1996 Academic Press, Inc.
avb3 belongs to the integrin family of glycoproteins. Different members of this family are expressed on the surface of different types of cells. The predominant role of integrins is the mediation of cell–cell and cell–extracellular matrix interactions (1,2). avb3 , also named vitronectin receptor, can recognize fibrinogen (3), von Willebrand factor (3), thrombospondin (4), fibronectin (5), 1
This investigation was supported by NIH Grant HL 45486. To whom correspondence should be addressed. Fax: (215) 7074003 or 707-2783. 2
osteopontin (6), denatured collagen types I and IV (7), thrombin (8), and laminin (9). All of these ligands express the RGD sequence, which is critical for binding to avb3 . The av subunit contains a light chain (25 kDa) and a heavy chain (115 kDa) bound by an S–S bridge. The light chain has a cytoplasmic domain, hydrophobic transmembrane domain, and extracellular domain (10). The b3 subunit (92 kDa) is a single polypeptide, which has short cytoplasmic and transmembrane domains and a longer extracellular domain (11,12). Disintegrins echistatin and eristostatin are antagonists of integrins (13,14). Echistatin is a strong inhibitor of vitronectin binding to immobilized avb3 , whereas eristostatin is much less active in this system (15). On the other hand, eristostatin is the strongest inhibitor of fibrinogen receptor (aIIbb3 integrin) among disintegrins (16). The selectivity of these disintegrins depends upon the structure of their RGD-loop, determined by amino acids adjacent to RGD sequence (14). Vitronectin receptor (avb3 integrin) is constitutively expressed on endothelial cells, smooth muscle cells, macrophages, osteoclasts, and melanoma cells (17). Since avb3 is involved in angiogenesis, bone resorption, cancer metastasis, and smooth muscle proliferation (17), many areas of research would benefit from the availability of a simple and reliable method for purifying this receptor. Since the earlier report by Pytela et al. (18), vitronectin receptor has been routinely purified from human placenta, and several modifications of the original procedure have been reported (19). In our hands, purification of vitronectin receptor from placenta is a laborious, and not entirely reliable, method that yields low quantities of avb3 still contaminated with other proteins. In addition, the placenta-derived avb3 was contaminated with avb5 . For these reasons, we developed a method for purifying avb3 from Chinese hamster ovary (CHO) cells that express recombinant human avb3 .
68
AID
1046-5928/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
pepa
AP: PEP
PURIFICATION OF avb3 INTEGRIN
MATERIALS AND METHODS
CHO cells were cotransfected with expression constructs containing human av and b3 and a CDM8 vector containing the neomycin resistance gene (20,21). The transfected cells (VNRC3 cells) are the generous gift of Drs. Mark Ginsberg and Joseph Loftus (Scripps Research Institute, La Jolla, CA). A2A9, a monoclonal antibody specific for human aIIbb3 (22), was kindly provided by Dr. J. Bennett (University of Pennsylvania, Philadelphia, PA). LM 609 and LM 142, monoclonal antibodies specific for the human avb3 complex and av , respectively (23), were provided by Dr. D. Cheresh (Scripps Research Institute). The monoclonal antibodies 7E2 and PB1 specific for hamster a5 and b1 , respectively (24), were kindly provided by Dr. R. L. Juliano (University of North Carolina, Chapel Hill, NC). Antihuman vitronectin polyclonal antibody, developed in rabbit, was from Gibco BRL (Gaithersburg, MD). GRGDSPK peptide was synthesized by Dr. B. Jameson (Thomas Jefferson University, Philadelphia, PA) and coupled to CNBr–Sepharose 4B (Pharmacia, Piscataway, NJ) in the proportion 20 mg peptide per 1 ml of slurry. n-Octylglucoside (1-O-n-octyl-b-D-glucopyranoside) and Hanks’ balanced salt solution (HBSS) were acquired from Boehringer-Mannheim (Indianapolis, IN) and Gibco BRL, respectively. Human vitronectin was purchased from Sigma (St. Louis, MO) or from Calbiochem (La Jolla, CA), and human von Willebrand factor was obtained from Hematologic Technologies (Essex Junction, VT). Highly purified human fibrinogen was a gift from Dr. A. Budzynski (Temple University, Philadelphia PA). All other reagents were purchased from Sigma. Disintegrins echistatin and eristostatin were purified from the crude venom of the snakes Echis carinatus and Eristocophis macmahoni, respectively, using two steps of C-18 reverse-phase high-performance liquid chromatography (25). Both preparations were highly homogeneous as determined by SDS–polyacrylamide gel electrophoresis (SDS–PAGE) and N-terminal amino acid sequencing (16). avb3 integrin was purified from VNRC3 cells. Cells were grown in Dulbecco’s modified Eagle medium containing 10% fetal calf serum, nonessential amino acids, glutamine, penicillin, and streptomycin. Cells were detached from plates with calcium- and magnesium-free HBSS and 5 mM EDTA and washed three times with HBSS containing calcium and magnesium. All preparative steps were performed at room temperature. About 2 1 108 cells were lysed in the presence of 8 ml of buffer A: 100 mM n-octylglucoside, 50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 1 mM MgCl2 , 1 mM CaCl2 , and 1 mM PMSF. After removal of insoluble material by centrifugation at 3000g, the extract was concentrated to 1 ml using Centriprep 10 (Amicon, Beverly, MA). The
AID
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
69
GRGDSPK–Sepharose 4B column (1 ml) was equilibrated with buffer B: 50 mM n-octylglucoside, 50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 2 mM MgCl2 , and 0.1 mM CaCl2 . One milliliter of lysed cell extract was divided into two 0.5-ml aliquots. After application of the first aliquot, the column was incubated for 30 min to increase binding of integrin to resin. Then the remaining 0.5-ml aliquot was applied, followed by another 30-min incubation. The column was washed with buffer B and then with a modified buffer B containing 300 mM NaCl. The receptor was eluted with a modified buffer B containing 15 mM EDTA. One-milliliter fractions were collected in vials containing 30 ml of 1 M MgCl2 . Fractions interacting with the anti-avb3 monoclonal antibody LM 609 were combined, concentrated to a volume of about 1 ml, and dialyzed against buffer B. Purified avb3 was stored at 0207C in small aliquots. avb3 from placenta was purified as described earlier (18,19). Briefly, fresh human placenta tissue was homogenized in buffer containing 20 mM Tris–HCl, pH 7.4, 10 mM EDTA, 1 M NaCl, 10 mM sodium azide and 0.3 mM PMSF, and the homogenate was stirred for 48 h at 47C. The supernatant was removed by centrifugation (1 h, 8000g), and the pellet was homogenized again in buffer C: 50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 1 mM MgCl2 , 1 mM MnCl2 , 0.1 mM CaCl2 , 0.5 mM PMSF, 1 mg/ml each of aprotinin, antipain, and leupeptin, and 25 mM b-D-octylglucoside. The homogenate was stirred overnight at 47C, and the supernatant was collected following centrifugation at 8000g. The extraction was repeated once (3 h), and the combined supernatants were centrifuged at 30,000g and stored at 0207C in 100- to 200-ml aliquots. Extracts were thawed and diluted 1:1 with buffer C without PMSF. The extract (200 ml) was applied to a GRGDSPK–Sepharose column equilibrated with buffer C overnight. Fractions containing placental avb3 were eluted from the column and identified as described in the procedure for purification of recombinant avb3 . ELISA plates (EIA/RIA, Costar, Cambridge, MA) were coated with purified avb3 , extracellular matrix proteins, or disintegrins in 0.05 M carbonate/bicarbonate buffer (pH 9.3) and incubated overnight at 47C. The wells were blocked with 5% nonfat milk in PBS/Tween (PBST) buffer. Purified avb3 (or lysate of VNRC3 cells) dissolved in buffer B was added to wells coated with matrix proteins or disintegrins, and the plates were incubated for 30 min at 377C. Binding of murine monoclonal antibodies to the receptor was measured using goat anti-mouse IgG conjugated with alkaline phosphatase, as described previously (26). Binding of vitronectin to immobilized avb3 was measured by the method described previously (15). SDS–PAGE was performed using the Laemmli method (27) under reducing conditions. Bands were visualized by silver staining. Protein concentrations
pepa
AP: PEP
70
MARCINKIEWICZ ET AL. TABLE 1
Efficiency of Purification of avb3 from VNRC3 Cells
Number of CHO cells VNRC3 2.2 1 108 CHO-K1 2.2 1 108
Total protein in lysate (mg)
avb3 in lysate (mg)
Protein in fractions eluted from GRGDSPKcolumn with EDTA (mg)
15.2 { 2.2
0.17 { 0.03
0.14 { .04
79 { 6
15.0 { 2.8
Not detectable
0.05 { 0.01
—
Yield (%)
Note. The amount of avb3 in whole cell lysates from CHO cells transfected with avb3 (VNRC3 cells) and nontransfected CHO-K1 cells was determined by ELISA. avb3 present in cell lysates was allowed to bind to immobilized echistatin (140 ng/well) and was then detected by monoclonal antibody LM 609. The amount of receptor was calculated from standard curve. Cell lysates were then applied to a GRGDSPKcolumn, and the amount of protein in fractions eluted with EDTA was determined. The data represent the mean with standard deviation from three independent experiments.
were determined using the BCA protein assay reagent (Pierce Co. Rockford, IL), with bovine serum albumin serving as the standard. RESULTS
We purified human recombinant avb3 integrin from CHO cells transfected with avb3 (VNRC3 cells) by affinity chromatography on a GRGDSPK–Sepharose column. The purification procedure included two steps: lysis of cells and chromatography of the lysate on an affinity column. This method was based on the observation that avb3 integrin binds to ligands via their RGD sequences. The receptor retained on the column was released by EDTA. The fractions were collected in tubes containing a high excess of Mg2/, which stabilizes avb3 . As a control, an identical procedure was carried out on nontransfected CHO-K1 cells. Table 1 presents the avb3 content in VNRC3 cells and the efficiency of the described purification method. This purification procedure was very efficient, yielding approximately 79% of the avb3 present in the VNRC3 cell lysates. The amount of avb3 in the cell lysates was determined using a ligand capture ELISA assay, in which the receptor that selectively bound to immobilized echistatin was detected by an avb3 complex-specific monoclonal antibody (LM 609). The ligand capture ELISA was used because the content of the receptor in the cell lysates was too low to be detected by conventional ELISA. A standard curve was constructed using known concentrations of previously purified avb3 . The level of avb3 in the VNRC3 cell lysates amounted to 1.1% of the total protein content. Figure 1 shows the ability of each fraction, eluted with buffer B containing 300 mM NaCl and 15 mM EDTA, to react with monoclonal antibody LM 609. Fractions 2–6, which demonstrated the highest absorption at 405 nm in the ELISA assay, were combined and concentrated. This figure also shows that fractions
AID
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
eluted from CHO-K1 cell lysate applied to the same column were not immunoreactive. The placental extract was applied to the GRGDSPK–Sepharose column, and immunoreactive material was eluted in the same fractions as in experiments with lysates of VNRC3 cells. The immunoreactive fractions of both preparations of the receptor were adjusted to the same protein concentrations and then analyzed by SDS– PAGE. Figure 2 compares purity of avb3 preparations obtained from VNRC3 cells (lane A) and from placenta (lane B). Both preparations showed identical bands corresponding to the av heavy chain (Ç110 kDa) and b3 subunit (Ç95 kDa). However, the placental preparation was more heavily contaminated with bands migrating with lower molecular weights. In parallel, the nonimmunoreactive fractions from CHO-K1 cell lysate that were adsorbed and eluted from the GRGDSPK– Sepharose column were collected, adjusted to the same protein concentration as material obtained from VNRC3 cells and from placenta, and analyzed by SDS– PAGE. Figure 2 (lane C) shows that fractions corresponding to av and b3 subunits were absent. On the other hand, this material separated into two bands that migrated with different mobility than a and b3 integrins. Since CHO cells express a5 and b1 integrin (28,29) the preparation obtained from CHO-K1 cells after the GRGDSPK–Sepharose column was examined for interaction with anti-a5 hamster monoclonal antibody (7E2) and anti-b1 hamster monoclonal antibody (PB1) raised in mice. However, these antibodies did not react with CHO-K1 preparation in ELISA assay (data not shown), suggesting that the bands observed in Fig. 2, lane C, did not correspond to hamster a5b1 . Interestingly, these two bands were absent in preparation obtained from transfected cells (line A) and from placenta (line B). Figure 3 compares the reactivity of LM 609 with three preparations eluted from the GRGDSPK–Sepharose column. The avb3 from VNRC3 cells was dramati-
pepa
AP: PEP
PURIFICATION OF avb3 INTEGRIN
71
FIG. 1. Interaction of fractions eluted with EDTA from GRGDSPK–Sepharose column with monoclonal antibody LM 609, specific for avb3 . Fractions (100 ml), after dialysis, were immobilized overnight on ELISA plates in carbonate/bicarbonate buffer, pH 9.3. Plates were blocked by addition of 5% milk in PBST buffer to each well, and then the monoclonal antibody LM 609 was added. After incubation for 60 min at 377C, the plates were washed with PBST buffer, and alkaline phosphatase-conjugated goat anti-mouse IgG was added. Plates were incubated for 60 min at 377C, and then after washing with PBST, the substrate for alkaline phosphatase p-nitrophenyl phosphate (pNPP) was added. After color development, the plates were read using an ELISA reader at 405 nm single wavelength. Absorbance at 405 nm represents spectrophotometric reading for the product of alkaline phosphatase reaction with substrate pNPP. Open circles indicate fractions obtained from VNRC3 cells, and filled circles indicate fractions obtained from control CHO-K1 cells. Error bars represent the standard deviation from three independent experiments.
cally more reactive than the avb3 from placenta. As expected, the preparation from CHO-K1 cells was not reactive. Purified avb3 from VNRC3 cells was recognized by monoclonal antibody LM 142, although not as well as by LM 609 (Fig. 4). A2A9, a monoclonal anti-
body specific for aIIbb3 , did not react with recombinant avb3 (Fig. 4) or with placental preparation (not shown). Studies were performed to examine the ligand specificity of avb3 purified from VNRC3 cells. Figure 5A shows that avb3 bound strongly to immobilized vitronectin and von Willebrand factor; however, it did not bind to immobilized fibrinogen. Purified avb3 bound more strongly to the immobilized echistatin than to eristostatin (Fig. 5B). Figure 6 shows that vitronectin was efficiently bound by immobilized avb3 . DISCUSSION
FIG. 2. SDS polyacrylamide gel electrophoresis purified integrins. Lane A, avb3 purified from VNRC3 cells; lane B, avb3 from placenta; lane C, preparation obtained from CHO-K1 cells. Two-microgram aliquots of each preparation were loaded on the gel, and electrophoresis was performed under reducing conditions using Laemmli method followed by silver staining.
AID
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
We report a simple, short, and reliable method of purification of avb3 integrin from transfected cells. Most investigators use placenta as a source of human vitronectin receptor (18,19). However, this method of purification is laborious, and it did not guarantee the removal of other human integrins present in blood from the final preparation. In the case of transfected cells, this problem is avoided, and the purified avb3 is not contaminated with other human integrins. The absence of contaminating aIIbb3 was demonstrated using monoclonal antibody A2A9, which recognizes the aIIbb3 complex but not avb3 . Our preparations did not react
pepa
AP: PEP
72
MARCINKIEWICZ ET AL.
FIG. 3. Binding of LM 609 monoclonal antibody to avb3 purified from VNRC3 cells and from placenta. The purified receptors were immobilized on ELISA plates, and the experiment was performed as described in the Fig. 1 legend. Open circles show binding of LM 609 to avb3 from VNRC3 cells, filled circles to avb3 from placenta, and open triangles to a preparation from CHO-K1 cells. Error bars represent standard deviation from three independent experiments.
tal preparation. The preparation obtained from control CHO-K1 cells was not immunoreactive with antibodies against human integrins. It also contained three times less protein than the preparation obtained from VNRC3 cells. avb3 from VNRC3 cells did not seem to contain hamster integrins, as demonstrated by SDS–PAGE. It is known that a5b1 , an integrin expressed on CHO cells (28,29), has a much lower affinity for GRGDSPK than does avb3 (18). However, the preparation from CHO-K1 cells did not react with hamster’s anti-a5 and anti-b1 monoclonal antibodies. Therefore, it is likely that the CHO-K1 preparation contains integrins other than a5b1 and avb3 integrins which are also able to bind to immobilized GRGDSPK column. This ‘‘unknown’’ integrin probably has higher affinity to GRGDSPK–Sepharose column than a5b1 but lower than avb3 , because it was not found in the VNRC3 preparation. Other investigators have also confirmed absence of avb3 integrin on CHO cells (28,29). avb3 integrin stably expressed on the surface of CHO cells has the same properties as naturally occurring avb3 on human umbilical vein endothelial cells (HUVEC) (30). Suspension of both cells bind monoclonal antibodies specific for avb3 (LM 609, LM 142 and 7E3) and FITC-echistatin. Echistatin induces the ligand-induced binding site epitope recognized by monoclonal antibody Mab 62 (31) on the b3 subunit in both types of cells (32). Endothelial cells appear to be a major
with A2A9, but they were recognized by monoclonal antibodies LM 609 and LM 142, which recognize the avb3 complex and av , respectively. The preparation from placenta may contain integrin avb5 , which is very difficult to separate from avb3 . A similar amount of avb3 can be purified from one placenta (19) as from 5 1 108 VNRC3 cells. However, the time required to complete the latter purification procedure is significantly shorter. Our purification procedure takes 5–6 h, starting from harvesting the cells to final dialysis. Purification of receptor from placenta takes at least 3 days and requires many additional steps including washing and homogenization of placental tissue, stirring homogenate overnight at 47C, centrifugation at 8000g and then ultracentrifugation at 30,000g. The process of applying the placental extract to the affinity column took 14 h. The application of VNRC3 cell extract took only 1 h since the total amount of applied proteins was much lower than in placental extract. This can be advantageous because nonspecific absorption of other proteins, which may contaminate the final product, is avoided. SDS– PAGE demonstrated that the avb3 from placenta had additional bands representing contaminants, whereas the recombinant preparation had no additional bands and was apparently much more pure. Accordingly, the recombinant preparation showed higher immunoreactivity with monoclonal antibody LM 609 than did the placen-
FIG. 4. Binding of monoclonal antibodies to avb3 immobilized on ELISA plates. The purified receptor was immobilized on ELISA plates and binding of monoclonal antibodies was examined as described in the Fig. 1 legend. Filled circles, binding of LM 609; open circles, binding of LM 142; open triangles, binding of A2A9. Error bars represent the standard deviation from three independent experiments.
AID
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
pepa
AP: PEP
PURIFICATION OF avb3 INTEGRIN
73
FIG. 5. Binding of avb3 to extracellular matrix proteins (A) and to disintegrins (B) immobilized on ELISA plates. The ligands (extracellular matrix proteins or disintegrins) were immobilized overnight on ELISA plates in carbonate/bicarbonate buffer, pH 9.2. Plates were blocked, and the avb3 was added in buffer B. The incubation was performed for 30 min at 377C, and then binding of receptor was examined using LM 609 monoclonal antibody under conditions described in the Fig. 1 legend. Open circles, vitronectin (1 mg/well); filled circles, fibrinogen (1 mg/well); open triangles, von Willebrand factor (1 mg/well), open squares, echistatin (140 ng/well); filled squares, eristostatin (140 ng/ well). Error bars represent standard deviation from three independent experiments.
source of avb3 in placenta. We also succeeded in purifying avb3 from HUVEC (not shown), which may represent an alternate source of this receptor.
Purified, recombinant vitronectin receptor bound to vitronectin and von Willebrand factor but showed no interaction with fibrinogen. Conforti et al. (33) made the same observation with avb3 purified from placenta and found that binding of avb3 to fibrinogen increased in the presence of lipids and Mn2/ ions. Recombinant avb3 also reacted with disintegrins in a manner similar to that of placental avb3 (15). In summary, we described a simple, fast, and highyield method of purification of recombinant avb3 from transfected cells. The purified receptor shows the same immunological properties and ability to interact with ligands as receptor purified from other sources. We believe that our method will make it easier for other investigators to obtain useful quantities of functional, highly purified avb3 . ACKNOWLEDGMENTS The authors greatly appreciate the contribution of Drs. M. H. Ginsberg and J. C. Loftus (Scripps Research Institute), who provided cells transfected with avb3 . They thank Dr. D. A. Cheresh and Dr. J. S. Bennett for providing monoclonal antibodies against human avb3 , av , and aIIbb3 and Dr. R. L. Juliano for providing monoclonal antibodies against hamster a5 and b1 . The authors thank Mr. Adam Mohmand and Dr. Y. Wang for technical assistance in experiments.
FIG. 6. Binding of vitronectin to immobilized avb3 . avb3 was immobilized on ELISA plates as described in the Fig. 1 legend, and then vitronectin (1 mg/sample) in buffer B was added. Incubation was performed for 30 min at 377C, and binding of vitronectin was detected using anti-vitronectin polyclonal antibody as described under Materials and Methods. Error bars represent standard deviation from three independent experiments.
AID
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
REFERENCES 1. Hynes, R. O. (1992) Integrins: Versatility, modulation, and signaling in cell adhesion. Cell 69, 11–25. 2. Smyth, S. S., Joneckis, C. C., and Parise, L. V. (1993) Regulation of vascular integrins. Blood 81, 2827–2834.
pepa
AP: PEP
74
MARCINKIEWICZ ET AL.
3. Cheresh, D. A. (1987) Human endothelial cells synthesize and express an arg-gly-asp directed receptor involved in attachment to fibrinogen and von Willebrand factor. Proc. Natl. Acad. Sci. USA 84, 6471–6475. 4. Sun, X., Skorstengaard, K., and Mosher, D. F. (1992) Disulfides modulate RGD-inhibitable cell adhesive activity of thrombospondin. J. Cell. Biol. 118, 693–671. 5. Charo, I. F., Nannizzi, L., Smith J. W., and Cheresh, D. A. (1990) The vitronectin receptor avb3 binds fibronectin and acts in concert with a5b1 in promoting cellular attachment and spreading on fibronectin. J. Cell. Biol. 111, 2795–2800. 6. Miyauchi, A., Alvarez, J., Greenfield, E. M., Teti, A., Grano, M., Colucc, S., Zambonin- Zallone, A., Ross, F. P., Teitelbaum, S. L., Cheresh, D. A., and Hruska, K. A. (1991) Recognition of osteopontin and related peptides by an avb3 integrin stimulates immediate cell signals in osteoclasts. J. Biol. Chem. 226, 20369– 20374. 7. Clyman, R. I., Mauray, F., and Kramer, R. H. (1992) b1 and b3 integrins have different roles in the adhesion and migration of vascular smooth muscle cells on extracellular matrix. Exp. Cell. Res. 200, 272–284. 8. Bar-Shavit, R., Sabbah, V., Lampugnani, M. G., Marchisio, P. C., Fenton, J. W., Vlodavsky, I., and Dejana, E. (1991) An Arg-Gly-Asp sequence within thrombin promotes endothelial cell adhesion. J. Cell. Biol. 112, 335–344. 9. Kramer, R. H., Cheng, Y.-F., and Clyman, R. (1990) Human microvascular endothelial cells use b1 and b3 integrin receptor complexes to attach to laminin. J. Cell. Biol. 111, 1233–1243. 10. Suzuki, S., Argraves, W. S., Arai, H., Languino, L. R., Pierschbacher, M., and Ruoslahti, E. (1987) Amino acid sequence of the vitronectin receptor alpha subunit and comparative expression of adhesion receptor mRNAs. J. Biol. Chem. 262, 14080–14085. 11. Zimrin, A. B., Eisman, R., Vilaire, G., Schwartz, E., and Bennett, J. S. (1988) Structure of platelet Glycoprotein IIIa: A common subunit for two different membrane receptor. J. Clin. Invest. 81, 1470–1475. 12. Calvete, J. J. (1994) Clues for understanding the structure and function of a prototypic human integrin: The platelet glycoprotein IIb/IIIa complex. Thromb. Haemost. 72, 1–15. 13. Gould, R. J., Polokoff, M. A., Friedman, P. A., Huang, T. F., Cook, J. J., and Niewiarowski, S. (1990) Disintegrins: A family of integrin inhibitory proteins form viper venoms. Proc. Soc. Exper. Biol. Med. 195, 168–171. 14. Niewiarowski, S., McLane, M. A., Kloczewiak, M., and Stewart, G. J. (1994) Disintegrins and other naturally occurring antagonists of platelet fibrinogen receptors. Semin. Hematol. 31, 289– 300. 15. Pfaff, M., McLane, M. A., Beviglia, L., Niewiarowski, S., and Timpl, R. (1994) Comparison of disintegrins with limited variations in the RGD loop in their binding to purified integrins and in cell adhesion inhibition. Cell Adhes.Commun. 2, 491–501. 16. McLane, M. A., Kowalska, M. A., Silver, L., Shattil, S. J., and Niewiarowski, S. (1994) Interaction of disintegrins with the aIIbb3 receptor on resting and activated human platelets. Biochem. J. 301, 429–436. 17. Felding-Habermann, B., and Cheresh, D. A. (1993) Vitronectin and its receptors Curr. Opin. Cell Biol. 5, 864–868. 18. Pytela, R., Pierschbacher, M. D., Aggraves, S., Suzuki, S., and
AID
PEP 0614
/
6q08$$$241
07-18-96 19:26:04
19.
20.
21.
22.
23.
24.
25.
26.
27. 28.
29.
30.
31.
32.
33.
Ruoslahti, E. (1987) Arginine–glycine–aspartic acid adhesion receptors. Methods Enzymol. 144, 475–489. Pfaff, M., Tangemann, K., Muller, B., Gurrath, M., Muller, G., Kessler, H., Timpl, R., and Engel, J. (1994) Selective recognition of cyclic RGD peptides of NMR defined conformation by aIIbb3, avb3, and a5b1 integrins. J. Biol. Chem. 269, 20233–20238. O’Toole, T. E., Loftus, J. C., Plow, E. F., Glass, A., Harper, J. R., and Ginsberg, M. H. (1990) Efficient surface expression of platelet GPIIb-IIIa requires both subunits. Blood 74, 14–18. O’Toole, T. E., Loftus, J. C., Du, X., Glass, A. A., Ruggeri, Z. M., Shattil, S. J., Plow, E. E., and Ginsberg, M. H. (1990) Affinity modulation of the aIIbb3 integrin (platelet GPIIb-IIIa) is an intrinsic property of the receptor. Cell Regul. 1, 883–893. Bennett, J. S., Hoxie, J. A., Leitman, S. F., Vilaire, G., and Cines, D. B. (1983) Inhibition of fibrinogen binding to stimulated human platelets by a monoclonal antibody. Proc. Natl. Acad. Sci. USA 80, 2417–2421. Cheresh, D. A., and Harper, J. R. (1987) Arg-gly-asp recognition by a cell adhesion receptor requires its 130-kDa a subunit. J. Biol. Chem. 262, 1434–1437. Brown, P. J., and Juliano, R. L. (1988) Monoclonal antibodies to distinctive epitopes on the a and b subunits of the fibronectin receptor. Exp. Cell Res. 177, 303–318. Williams, J., Rucinski, B., Holt, J. C., and Niewiarowski, S. (1990) Elegantin and albolabrin purified peptides from viper venoms: Homologies with the RGDS domain of fibrinogen and von Willebrand factor. Biochim. Biophys. Acta 1039, 81–89. Cook, J. J., Trybulec, M., Lasz, E. C., Khan, S., and Niewiarowski, S. (1992) Binding of glycoprotein IIIa–derived peptide 217– 231 to fibrinogen and von Willebrand factors and its inhibition by platelet glycoprotein IIb/IIIa complex. Biochim. Biophys. Acta 1119, 312–321. Laemmli, U. K. (1978) Clevage of structural proteins during the assembly of the head of bacteriophage T4 . Nature 227, 680–685. Schreiner, C. L., Bauer, J. S., Danilov, Y. N., Hussein, S., Sczekan, M., and Juliano, R. L. (1989) Isolation and characterization of Chinese hamster ovary cells variants deficient in the expression of fibronectin receptor. J. Cell Biol. 109, 3157–3167. Takada, Y., Ylanne, J., Mandelman, D., Puzon, W., and Ginsberg, M. (1992) A point mutation of integrin b1 subunit blocks binding of a5b1 to fibronectin and invasion but not recruitment to adhesion plaques. J. Cell Biol. 119, 913–921. Jaffe, E. A., Nachman, R. L., and Minick, C. R. (1973) Culture of human endothelial cells derived from umbilical veins. J. Clin. Invest. 52, 2745–2756. Du, X., Gu, M., Weisel, J. W., Nagaswami, C., Bennett, J. S., Bowditch, R., and Ginsberg, M. H. (1993) Long range propagation of conformational changes in integrin aIIbb3 . J. Biol. Chem. 268, 23087–23092. Juliano, D., Wang, Y., Marcinkiewicz, C., Rosenthal, L. A., Stewart, G. J., and Niewiarowski, S. (1996) Disintegrin interaction with avb3 integrin on human umbilical vein endothelial cells. Expression of ligand induced binding site on b3 subunit. Exp. Cell Res. 219, 571–578, 1996. Conforti, G., Zanetti, A., Pasquali-Ronchetti, I., Quaglino, D., Jr., Neyroz, P., and Dejana, E. (1990) Modulation of vitronectin receptor binding by membrane lipid composition. J. Biol. Chem. 265, 4011–4019.
pepa
AP: PEP